SPB Git forge

spb/ai-atlas

Public
41commits 1branches 0releases
4.6 MBsize
maindefault branch
12 days agolast push
HTML 77.2% TypeScript 10.5% Python 9.6% JavaScript 2.5%
588.6 KB · 626 lines
Raw Blame History
1<?xml version="1.0" encoding="UTF-8"?>2<feed xmlns="http://www.w3.org/2005/Atom" xmlns:media="http://search.yahoo.com/mrss/" xml:lang="en-US">3  <id>tag:github.com,2008:https://github.com/vllm-project/vllm/releases</id>4  <link type="text/html" rel="alternate" href="https://github.com/vllm-project/vllm/releases"/>5  <link type="application/atom+xml" rel="self" href="https://github.com/vllm-project/vllm/releases.atom"/>6  <title>Release notes from vllm</title>7  <updated>2026-09-11T08:30:05Z</updated>8  <entry>9    <id>tag:github.com,2008:Repository/599547518/proto-v0.1.0</id>10    <updated>2026-09-11T08:30:05Z</updated>11    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/proto-v0.1.0"/>12    <title>proto-v0.1.0</title>13    <content type="html">&lt;p&gt;vllm-proto 0.1.0&lt;/p&gt;</content>14    <author>15      <name>BugenZhao</name>16    </author>17    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/25862682?s=60&amp;v=4"/>18  </entry>19  <entry>20    <id>tag:github.com,2008:Repository/599547518/v0.29.0</id>21    <updated>2026-09-10T23:53:36Z</updated>22    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0"/>23    <title>v0.29.0</title>24    <content type="html">&lt;h1&gt;v0.29.0&lt;/h1&gt;25&lt;h2&gt;Highlights&lt;/h2&gt;26&lt;p&gt;This release features 594 commits from 277 contributors (91 new)!&lt;/p&gt;27&lt;ul&gt;28&lt;li&gt;&lt;strong&gt;Model Runner V2 is now the default for all models&lt;/strong&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5209436023&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53183&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53183/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53183&quot;&gt;#53183&lt;/a&gt;), completing the rollout that began with pooling models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4859338343&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48290&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48290/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48290&quot;&gt;#48290&lt;/a&gt;). MRV2 also gained CUDA graph memory profiling for KV cache auto-sizing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5217002273&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53306&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53306/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53306&quot;&gt;#53306&lt;/a&gt;), batch-sharded sampling that cuts per-step logits memory by 1/TP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5023038565&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50465&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50465/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50465&quot;&gt;#50465&lt;/a&gt;), prompt embeds (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4468165522&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/42963&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/42963/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/42963&quot;&gt;#42963&lt;/a&gt;), &lt;code&gt;extract_hidden_states&lt;/code&gt; speculation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4976435154&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49811&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49811/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49811&quot;&gt;#49811&lt;/a&gt;), padded FULL cudagraph dispatch for uniform decode under spec decode (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5224268841&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53407&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53407/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53407&quot;&gt;#53407&lt;/a&gt;), and DP-sync skipping before EAGLE/MTP draft prefill (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5243116045&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53694&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53694/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53694&quot;&gt;#53694&lt;/a&gt;). MRV1 remains in use for a few ROCm models and features MRV2 does not yet support.&lt;/li&gt;29&lt;li&gt;&lt;strong&gt;New models&lt;/strong&gt;: Hy4-preview, Tencent&#39;s 770B/49B-active MoE with Gated DeepSeek Sparse Attention and native MTP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5275259831&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54160&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54160/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54160&quot;&gt;#54160&lt;/a&gt;); Qwen3.8-Flash-Next with BF16/FP8/NVFP4 and MTP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5257259389&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53896&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53896/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53896&quot;&gt;#53896&lt;/a&gt;); GraniteSWA and GraniteMoeSWA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179178405&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52706&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52706/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52706&quot;&gt;#52706&lt;/a&gt;); NemotronH_Omni_Reasoning_V3 with MTP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5191749387&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52929&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52929/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52929&quot;&gt;#52929&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5204637894&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53121&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53121/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53121&quot;&gt;#53121&lt;/a&gt;); Kimi K3 NVFP4 checkpoints (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5205567349&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53132&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53132/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53132&quot;&gt;#53132&lt;/a&gt;).&lt;/li&gt;30&lt;li&gt;&lt;strong&gt;Kimi-K3 and DeepSeek V4 performance&lt;/strong&gt;: fused MXFP4 top-k finalization in the K3 latent tail (about 5% E2E latency, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5206732861&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53152&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53152/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53152&quot;&gt;#53152&lt;/a&gt;), K3 Mamba metadata preparation in one Triton launch (6.6-7.6x kernel speedup, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5155366912&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52388&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52388/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52388&quot;&gt;#52388&lt;/a&gt;), tuned Hopper low-latency GEMM (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5270636056&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54088&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54088/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54088&quot;&gt;#54088&lt;/a&gt;) now also dispatched on SM100 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231554402&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53534&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53534/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53534&quot;&gt;#53534&lt;/a&gt;) and used for &lt;code&gt;eh_proj&lt;/code&gt; (12.9-25.2% kernel speedup, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5260560316&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53942&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53942/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53942&quot;&gt;#53942&lt;/a&gt;), GEMM-RS extended to GEMM-AR (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5200092730&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53053&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53053/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53053&quot;&gt;#53053&lt;/a&gt;), MLA gate merged into the QKV-A projection (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5265157591&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54015&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54015/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54015&quot;&gt;#54015&lt;/a&gt;), K3 DCP with DSpark (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5143408851&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52188&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52188/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52188&quot;&gt;#52188&lt;/a&gt;) and DCP partial prefix cache hits (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5024902326&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50493&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50493/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50493&quot;&gt;#50493&lt;/a&gt;); DeepSeek V4 shared experts fused into MegaMoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199333771&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53040&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53040/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53040&quot;&gt;#53040&lt;/a&gt;), adaptive top-k width re-landed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185770562&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52823&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52823/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52823&quot;&gt;#52823&lt;/a&gt;), a native SwiGLU clamp kernel for Humming MoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5242692210&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53685&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53685/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53685&quot;&gt;#53685&lt;/a&gt;), and an opt-in FlashInfer &lt;code&gt;moe_ep&lt;/code&gt; expert backend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4961919843&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49636&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49636/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49636&quot;&gt;#49636&lt;/a&gt;).&lt;/li&gt;31&lt;li&gt;&lt;strong&gt;Speculative decoding&lt;/strong&gt;: per-request acceptance stats in OpenAI API responses via &lt;code&gt;--per-request-spec-decode-metrics&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4908287242&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48915&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48915/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48915&quot;&gt;#48915&lt;/a&gt;), adaptive verification extended to logprobs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5147428527&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52242&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52242/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52242&quot;&gt;#52242&lt;/a&gt;), SM100 sparse MLA for GLM-5.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183188300&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52783&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52783/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52783&quot;&gt;#52783&lt;/a&gt;) and DeepSeek V4 on SM90 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183962079&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52795&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52795/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52795&quot;&gt;#52795&lt;/a&gt;), Qwen3-Omni DSpark drafts (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5167704960&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52560&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52560/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52560&quot;&gt;#52560&lt;/a&gt;), PLaMo3 EAGLE-3/DFlash (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5280514396&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54239&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54239/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54239&quot;&gt;#54239&lt;/a&gt;), and DFlash2 loading from speculators format (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5250815518&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53797&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53797/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53797&quot;&gt;#53797&lt;/a&gt;).&lt;/li&gt;32&lt;li&gt;&lt;strong&gt;RL weight sync&lt;/strong&gt;: a new &lt;code&gt;sharded_rdt&lt;/code&gt; P2P backend where each worker pulls only its TP/EP slice over NIXL or Ray Direct Transport (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4498962307&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43375&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43375/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43375&quot;&gt;#43375&lt;/a&gt;), rank-local IPC weight updates (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163032127&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52497&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52497/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52497&quot;&gt;#52497&lt;/a&gt;), sparse checkpoint-coordinate updates through native weight loaders (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040743562&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50723&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50723/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50723&quot;&gt;#50723&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247441451&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53751&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53751/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53751&quot;&gt;#53751&lt;/a&gt;), and routed expert loading for gpt-oss (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145232591&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52209&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52209/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52209&quot;&gt;#52209&lt;/a&gt;).&lt;/li&gt;33&lt;li&gt;&lt;strong&gt;Mamba prefix caching&lt;/strong&gt;: internal prefill checkpoints deliver a 9%-25% TTFT improvement (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183735872&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52789&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52789/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52789&quot;&gt;#52789&lt;/a&gt;); &lt;code&gt;prefix_cache_retention_interval&lt;/code&gt; is now a CLI argument defaulting to 0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145771808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52216&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52216/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52216&quot;&gt;#52216&lt;/a&gt;), with dense retention automatically restored for hybrid models using EAGLE/MTP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5377316976&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/55760&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/55760/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/55760&quot;&gt;#55760&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5384232024&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/55861&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/55861/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/55861&quot;&gt;#55861&lt;/a&gt;).&lt;/li&gt;34&lt;li&gt;&lt;strong&gt;New defaults&lt;/strong&gt;: FlashInfer all-reduce enabled by default for TP CUDA groups, opt out with &lt;code&gt;VLLM_ALLREDUCE_USE_FLASHINFER=0&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5196781600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52998&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52998/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52998&quot;&gt;#52998&lt;/a&gt;); prefix-cache &lt;code&gt;NONE_HASH&lt;/code&gt; is deterministic by default so distributed KV cache users no longer need to pin &lt;code&gt;PYTHONHASHSEED&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124402404&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51875&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51875/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51875&quot;&gt;#51875&lt;/a&gt;); new &lt;code&gt;--max-num-queued-reqs&lt;/code&gt; / &lt;code&gt;--max-num-queued-tokens&lt;/code&gt; admission-control flags (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4948101157&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49445&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49445/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49445&quot;&gt;#49445&lt;/a&gt;).&lt;/li&gt;35&lt;li&gt;&lt;strong&gt;Breaking changes&lt;/strong&gt;: ten deprecated model architectures removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5237599903&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53608&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53608/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53608&quot;&gt;#53608&lt;/a&gt;); FlexOlmo, Olmo3 and Hunyuan V1/VL migrated to the Transformers modeling backend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5238189101&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53615&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53615/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53615&quot;&gt;#53615&lt;/a&gt;); PyAV video decoder backend removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5279922155&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54231&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54231/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54231&quot;&gt;#54231&lt;/a&gt;); &lt;code&gt;python -m vllm.entrypoints.openai.api_server&lt;/code&gt; deprecated in favor of &lt;code&gt;vllm serve&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5139471549&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52131&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52131/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52131&quot;&gt;#52131&lt;/a&gt;); &lt;code&gt;VLLM_TEST_FORCE_FP8_MARLIN&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142686969&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52182&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52182/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52182&quot;&gt;#52182&lt;/a&gt;) and &lt;code&gt;VLLM_ROCM_USE_AITER_FP4_ASM_GEMM&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5205961808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53141&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53141/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53141&quot;&gt;#53141&lt;/a&gt;) removed.&lt;/li&gt;36&lt;/ul&gt;37&lt;h2&gt;&lt;g-emoji class=&quot;g-emoji&quot; alias=&quot;warning&quot;&gt;⚠️&lt;/g-emoji&gt; Model Runner V1 Deprecation&lt;/h2&gt;38&lt;p&gt;Now that Model Runner V2 is used by default, we are considering Model Runner V1 deprecated and are targeting v0.32 for its removal. We do not intend to accept any more MRV1-specific improvements or optimizations.&lt;/p&gt;39&lt;p&gt;Some &lt;a href=&quot;https://github.com/vllm-project/vllm/issues/47172&quot; data-hovercard-type=&quot;issue&quot; data-hovercard-url=&quot;/vllm-project/vllm/issues/47172/hovercard&quot;&gt;features&lt;/a&gt; are not yet supported in MRV2 but we are planning for these gaps to be closed within the next 2-3 weeks. These include sequence parallelism, dual-batch overlap, elastic expert parallellism, custom logits processors and certain speculative decoding methods. For now, vLLM will still fall back to use MRV1 if any of these features are configured.&lt;/p&gt;40&lt;h2&gt;Release Artifacts&lt;/h2&gt;41&lt;h3&gt;Python Wheels&lt;/h3&gt;42&lt;table&gt;43&lt;thead&gt;44&lt;tr&gt;45&lt;th&gt;Platform&lt;/th&gt;46&lt;th&gt;Install&lt;/th&gt;47&lt;/tr&gt;48&lt;/thead&gt;49&lt;tbody&gt;50&lt;tr&gt;51&lt;td&gt;PyPI (CUDA 13.0)&lt;/td&gt;52&lt;td&gt;&lt;code&gt;pip install vllm&lt;/code&gt;&lt;/td&gt;53&lt;/tr&gt;54&lt;tr&gt;55&lt;td&gt;PyPI (CUDA 13.0, uv)&lt;/td&gt;56&lt;td&gt;&lt;code&gt;uv pip install vllm --torch-backend=auto&lt;/code&gt;&lt;/td&gt;57&lt;/tr&gt;58&lt;tr&gt;59&lt;td&gt;ROCm&lt;/td&gt;60&lt;td&gt;&lt;code&gt;pip install vllm --extra-index-url https://wheels.vllm.ai/rocm/0.29.0/rocm723&lt;/code&gt;&lt;/td&gt;61&lt;/tr&gt;62&lt;tr&gt;63&lt;td&gt;XPU&lt;/td&gt;64&lt;td&gt;&lt;code&gt;uv pip install vllm --extra-index-url https://wheels.vllm.ai/0.29.0/xpu --extra-index-url https://download.pytorch.org/whl/xpu --index-strategy unsafe-best-match&lt;/code&gt;&lt;/td&gt;65&lt;/tr&gt;66&lt;/tbody&gt;67&lt;/table&gt;68&lt;h3&gt;Docker Images&lt;/h3&gt;69&lt;table&gt;70&lt;thead&gt;71&lt;tr&gt;72&lt;th&gt;Platform&lt;/th&gt;73&lt;th&gt;Docker Image&lt;/th&gt;74&lt;/tr&gt;75&lt;/thead&gt;76&lt;tbody&gt;77&lt;tr&gt;78&lt;td&gt;CUDA 13.0 (Default)&lt;/td&gt;79&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.29.0&lt;/code&gt; (&lt;code&gt;v0.29.0-cu130&lt;/code&gt; also works)&lt;/td&gt;80&lt;/tr&gt;81&lt;tr&gt;82&lt;td&gt;CUDA 12.9&lt;/td&gt;83&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.29.0-cu129&lt;/code&gt;&lt;/td&gt;84&lt;/tr&gt;85&lt;tr&gt;86&lt;td&gt;CUDA 13.0 + Ubuntu 24.04&lt;/td&gt;87&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.29.0-ubuntu2404&lt;/code&gt;&lt;/td&gt;88&lt;/tr&gt;89&lt;tr&gt;90&lt;td&gt;CUDA 12.9 + Ubuntu 24.04&lt;/td&gt;91&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.29.0-cu129-ubuntu2404&lt;/code&gt;&lt;/td&gt;92&lt;/tr&gt;93&lt;tr&gt;94&lt;td&gt;ROCm&lt;/td&gt;95&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-rocm:v0.29.0&lt;/code&gt;&lt;/td&gt;96&lt;/tr&gt;97&lt;tr&gt;98&lt;td&gt;CPU&lt;/td&gt;99&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-cpu:v0.29.0&lt;/code&gt;&lt;/td&gt;100&lt;/tr&gt;101&lt;tr&gt;102&lt;td&gt;XPU&lt;/td&gt;103&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-xpu:v0.29.0&lt;/code&gt;&lt;/td&gt;104&lt;/tr&gt;105&lt;/tbody&gt;106&lt;/table&gt;107&lt;h3&gt;Other Artifacts&lt;/h3&gt;108&lt;p&gt;Pre-built release artifacts are available in the &lt;strong&gt;Assets&lt;/strong&gt; section at the bottom of this page, including:&lt;/p&gt;109&lt;ul&gt;110&lt;li&gt;Source distribution tarball&lt;/li&gt;111&lt;li&gt;CUDA 12.9 Python wheels for x86_64 and arm64&lt;/li&gt;112&lt;li&gt;CUDA 13.0 Python wheels for x86_64 and arm64&lt;/li&gt;113&lt;li&gt;CPU Python wheels for x86_64, arm64, and macOS&lt;/li&gt;114&lt;li&gt;XPU Python wheel for x86_64&lt;/li&gt;115&lt;/ul&gt;116&lt;h2&gt;Model Support&lt;/h2&gt;117&lt;ul&gt;118&lt;li&gt;&lt;strong&gt;New models&lt;/strong&gt;: Hy4-preview (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5275259831&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54160&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54160/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54160&quot;&gt;#54160&lt;/a&gt;), Qwen3.8-Flash-Next (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5257259389&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53896&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53896/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53896&quot;&gt;#53896&lt;/a&gt;), GraniteSWA and GraniteMoeSWA via the existing Granite implementation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179178405&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52706&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52706/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52706&quot;&gt;#52706&lt;/a&gt;), NemotronH_Omni_Reasoning_V3 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5191749387&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52929&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52929/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52929&quot;&gt;#52929&lt;/a&gt;) with MTP for Nemotron VL models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5204637894&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53121&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53121/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53121&quot;&gt;#53121&lt;/a&gt;), bidirectional attention for DeepSeek-backbone embedding models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5193245192&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52948&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52948/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52948&quot;&gt;#52948&lt;/a&gt;), FP8 ModernBERT (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5202873727&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53101&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53101/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53101&quot;&gt;#53101&lt;/a&gt;), and Kimi K3 NVFP4 checkpoints (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5205567349&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53132&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53132/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53132&quot;&gt;#53132&lt;/a&gt;).&lt;/li&gt;119&lt;li&gt;&lt;strong&gt;Transformers modeling backend&lt;/strong&gt;: FlexOlmo, Olmo3 and Hunyuan V1/VL migrated off native implementations (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5238189101&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53615&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53615/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53615&quot;&gt;#53615&lt;/a&gt;), multimodal path hardened (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5121124453&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51827&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51827/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51827&quot;&gt;#51827&lt;/a&gt;), and &lt;code&gt;RMSNormFuser.fuse&lt;/code&gt; performance fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5182262756&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52766&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52766/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52766&quot;&gt;#52766&lt;/a&gt;).&lt;/li&gt;120&lt;li&gt;&lt;strong&gt;Weight loading&lt;/strong&gt;: weight tying now inspects the checkpoint so a real &lt;code&gt;lm_head&lt;/code&gt; is loaded even when the config claims tied embeddings (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109666279&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51665&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51665/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51665&quot;&gt;#51665&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5208221370&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53170&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53170/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53170&quot;&gt;#53170&lt;/a&gt;).&lt;/li&gt;121&lt;li&gt;&lt;strong&gt;LoRA&lt;/strong&gt;: DeepSeek V4 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5220828681&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53361&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53361/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53361&quot;&gt;#53361&lt;/a&gt;), Qwen3-Omni multimodal LoRA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183318748&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52786&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52786/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52786&quot;&gt;#52786&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5233194725&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53557&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53557/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53557&quot;&gt;#53557&lt;/a&gt;), tower/connector LoRA for LLaVA-NeXT (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4974910691&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49788&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49788/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49788&quot;&gt;#49788&lt;/a&gt;) and LFM2-VL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5097242709&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51498&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51498/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51498&quot;&gt;#51498&lt;/a&gt;), plus fixes for Muse-Glimmer (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5230858374&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53513&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53513/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53513&quot;&gt;#53513&lt;/a&gt;), Qwen3.5 embedding modules (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4902578074&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48850&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48850/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48850&quot;&gt;#48850&lt;/a&gt;), partial LoRA on Qwen3.5/3.6 GatedDeltaNet (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4812313997&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47640&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47640/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47640&quot;&gt;#47640&lt;/a&gt;), int32 overflow in punica kernels at long context (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199007710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53034&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53034/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53034&quot;&gt;#53034&lt;/a&gt;), false target matches on unsupported module types (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5150483821&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52313&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52313/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52313&quot;&gt;#52313&lt;/a&gt;), and base-layer/routed-expert prefix ordering (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5166563133&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52552&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52552/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52552&quot;&gt;#52552&lt;/a&gt;).&lt;/li&gt;122&lt;li&gt;&lt;strong&gt;Multimodal performance&lt;/strong&gt;: ViT full CUDA graph for Idefics3 and SmolVLM (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4811392083&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47625&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47625/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47625&quot;&gt;#47625&lt;/a&gt;), packed encoder attention for Pixtral (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142898516&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52185&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52185/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52185&quot;&gt;#52185&lt;/a&gt;), fused Kimi vision Q/K RoPE kernel (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5017612796&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50400&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50400/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50400&quot;&gt;#50400&lt;/a&gt;), Dots3 NOTE runtime (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5230913424&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53517&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53517/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53517&quot;&gt;#53517&lt;/a&gt;) and Omni encoder optimizations (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5227026612&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53460&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53460/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53460&quot;&gt;#53460&lt;/a&gt;), MM tensors no longer broadcast to workers for prefix-cache-covered items (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5134759141&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52041&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52041/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52041&quot;&gt;#52041&lt;/a&gt;), redundant placeholder scans skipped (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5191315493&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52925&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52925/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52925&quot;&gt;#52925&lt;/a&gt;), common token sequences cached (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5233476466&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53560&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53560/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53560&quot;&gt;#53560&lt;/a&gt;), async media resolved concurrently across modalities (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5298486219&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54537&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54537/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54537&quot;&gt;#54537&lt;/a&gt;), and H2D copies pinned to avoid stream stalls (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5224446151&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53412&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53412/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53412&quot;&gt;#53412&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5283293547&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54292&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54292/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54292&quot;&gt;#54292&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5283530752&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54299&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54299/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54299&quot;&gt;#54299&lt;/a&gt;).&lt;/li&gt;123&lt;li&gt;&lt;strong&gt;Multimodal inputs&lt;/strong&gt;: video embeds accepted by the Python frontend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5280853540&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54242&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54242/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54242&quot;&gt;#54242&lt;/a&gt;), modality-scoped &lt;code&gt;mm_processor_kwargs&lt;/code&gt; honored (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5251698126&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53808&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53808/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53808&quot;&gt;#53808&lt;/a&gt;), Qwen3-VL profiling honors &lt;code&gt;cap_pixels_per_frame&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5289377865&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54380&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54380/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54380&quot;&gt;#54380&lt;/a&gt;), video frame sampling respects MP4 edit-list trims (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4882623102&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48608&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48608/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48608&quot;&gt;#48608&lt;/a&gt;), oversized items skip the MM processor cache instead of crashing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5198011038&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53016&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53016/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53016&quot;&gt;#53016&lt;/a&gt;), encoder cache entries survive until their last use (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5283029292&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54284&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54284/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54284&quot;&gt;#54284&lt;/a&gt;), and repeated multimodal requests with the SHM cache no longer terminate the engine (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5326935622&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54994&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54994/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54994&quot;&gt;#54994&lt;/a&gt;).&lt;/li&gt;124&lt;li&gt;&lt;strong&gt;Correctness&lt;/strong&gt;: PaliGemma stale image scaling removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178143135&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52692&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52692/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52692&quot;&gt;#52692&lt;/a&gt;), Mistral3 placeholder grid with processor size overrides (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5189161450&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52874&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52874/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52874&quot;&gt;#52874&lt;/a&gt;), MiniCPM-o on Transformers v5 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5296466801&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54501&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54501/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54501&quot;&gt;#54501&lt;/a&gt;), JinaVL cache order (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5232825783&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53553&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53553/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53553&quot;&gt;#53553&lt;/a&gt;), mixed CLIP/SigLIP pooling batches (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5207528366&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53165&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53165/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53165&quot;&gt;#53165&lt;/a&gt;), XD-RoPE models on prefix-cache hits (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5226658887&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53456&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53456/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53456&quot;&gt;#53456&lt;/a&gt;), Qwen3-Omni audio encoder with non-divisible TP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5049295618&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50858&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50858/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50858&quot;&gt;#50858&lt;/a&gt;), GLM-5.2 no longer uses dense MHA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163559243&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52512&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52512/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52512&quot;&gt;#52512&lt;/a&gt;), Moondream3 MoE all-reduce (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5274944053&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54152&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54152/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54152&quot;&gt;#54152&lt;/a&gt;), deepseek-vl2 config defaults (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084211010&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51302&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51302/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51302&quot;&gt;#51302&lt;/a&gt;), HY-V3 compressed-tensors ignore matching (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4888766885&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48682&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48682/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48682&quot;&gt;#48682&lt;/a&gt;), MiniMax-M3 FP8 query allocation under CUDA graph replay (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5075534328&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51203&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51203/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51203&quot;&gt;#51203&lt;/a&gt;), and GraniteMoeHybrid quantized expert loading (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5267929189&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54052&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54052/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54052&quot;&gt;#54052&lt;/a&gt;).&lt;/li&gt;125&lt;/ul&gt;126&lt;h2&gt;Engine Core&lt;/h2&gt;127&lt;ul&gt;128&lt;li&gt;&lt;strong&gt;Model Runner V2&lt;/strong&gt;: default for all models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5209436023&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53183&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53183/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53183&quot;&gt;#53183&lt;/a&gt;) and pooling models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4859338343&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48290&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48290/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48290&quot;&gt;#48290&lt;/a&gt;), CUDA graph memory reservation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5217002273&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53306&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53306/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53306&quot;&gt;#53306&lt;/a&gt;), batch-sharded sampling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5023038565&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50465&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50465/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50465&quot;&gt;#50465&lt;/a&gt;), prompt embeds (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4468165522&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/42963&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/42963/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/42963&quot;&gt;#42963&lt;/a&gt;), &lt;code&gt;extract_hidden_states&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4976435154&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49811&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49811/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49811&quot;&gt;#49811&lt;/a&gt;), padded FULL cudagraph dispatch (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5224268841&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53407&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53407/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53407&quot;&gt;#53407&lt;/a&gt;), DP-sync skipping for drafts (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5243116045&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53694&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53694/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53694&quot;&gt;#53694&lt;/a&gt;), decoupled draft/target gumbel noise streams (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5282981569&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54282&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54282/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54282&quot;&gt;#54282&lt;/a&gt;), encoder-only path split out (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5208829726&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53176&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53176/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53176&quot;&gt;#53176&lt;/a&gt;), and memory released correctly on shutdown and sleep (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5230786234&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53508&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53508/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53508&quot;&gt;#53508&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5280947114&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54246&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54246/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54246&quot;&gt;#54246&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5275336219&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54162&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54162/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54162&quot;&gt;#54162&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5261657229&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53955&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53955/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53955&quot;&gt;#53955&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5242587074&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53682&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53682/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53682&quot;&gt;#53682&lt;/a&gt;).&lt;/li&gt;129&lt;li&gt;&lt;strong&gt;Speculative decoding&lt;/strong&gt;: per-request acceptance stats (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4908287242&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48915&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48915/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48915&quot;&gt;#48915&lt;/a&gt;), adaptive verification with logprobs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5147428527&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52242&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52242/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52242&quot;&gt;#52242&lt;/a&gt;), on SM100 sparse MLA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183188300&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52783&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52783/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52783&quot;&gt;#52783&lt;/a&gt;) and DSv4 + SM90 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183962079&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52795&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52795/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52795&quot;&gt;#52795&lt;/a&gt;), varlen trtllm-gen decode (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5140749483&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52157&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52157/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52157&quot;&gt;#52157&lt;/a&gt;), FlashInfer MLA for DSpark drafting under DCP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5282657237&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54277&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54277/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54277&quot;&gt;#54277&lt;/a&gt;), fused GDN MTP for all Qwen head ratios (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5165287951&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52539&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52539/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52539&quot;&gt;#52539&lt;/a&gt;), widest uniform decode batch captured by default (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5024571653&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50488&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50488/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50488&quot;&gt;#50488&lt;/a&gt;) with memory-safe graph sizes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5291610036&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54418&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54418/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54418&quot;&gt;#54418&lt;/a&gt;), and fixes for short_conv/LFM2 targets (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5009940371&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50272&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50272/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50272&quot;&gt;#50272&lt;/a&gt;), speculators-format config overrides (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4426318207&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/42376&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/42376/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/42376&quot;&gt;#42376&lt;/a&gt;), DFlash draft RoPE layout (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5288486337&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54373&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54373/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54373&quot;&gt;#54373&lt;/a&gt;), draft models with a different GQA ratio (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5197159965&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53002&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53002/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53002&quot;&gt;#53002&lt;/a&gt;), draft models with a larger hidden size under TP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5143949093&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52193&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52193/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52193&quot;&gt;#52193&lt;/a&gt;), DSpark backend inheritance scoped to DeepSeek V4 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185034825&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52809&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52809/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52809&quot;&gt;#52809&lt;/a&gt;), generic &lt;code&gt;DSparkDraftModel&lt;/code&gt; configs for Qwen3 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5144136494&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52197&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52197/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52197&quot;&gt;#52197&lt;/a&gt;), and Gemma4 MTP under CUDA graphs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5256705004&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53884&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53884/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53884&quot;&gt;#53884&lt;/a&gt;).&lt;/li&gt;130&lt;li&gt;&lt;strong&gt;Prefix caching &amp;amp; scheduling&lt;/strong&gt;: Mamba internal prefill checkpoints (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5183735872&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52789&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52789/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52789&quot;&gt;#52789&lt;/a&gt;), &lt;code&gt;prefix_cache_retention_interval&lt;/code&gt; argument (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145771808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52216&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52216/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52216&quot;&gt;#52216&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5377316976&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/55760&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/55760/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/55760&quot;&gt;#55760&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5384232024&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/55861&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/55861/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/55861&quot;&gt;#55861&lt;/a&gt;), deterministic &lt;code&gt;NONE_HASH&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124402404&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51875&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51875/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51875&quot;&gt;#51875&lt;/a&gt;), queue admission control (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4948101157&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49445&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49445/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49445&quot;&gt;#49445&lt;/a&gt;), KV null block reserved when validating &lt;code&gt;max_model_len&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4783990461&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47272&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47272/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47272&quot;&gt;#47272&lt;/a&gt;), spec decode no longer padded up to &lt;code&gt;max_model_len&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5262211619&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53962&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53962/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53962&quot;&gt;#53962&lt;/a&gt;), and negative external block allocation prevented (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179259792&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52707&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52707/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52707&quot;&gt;#52707&lt;/a&gt;).&lt;/li&gt;131&lt;li&gt;&lt;strong&gt;RL workflows&lt;/strong&gt;: &lt;code&gt;sharded_rdt&lt;/code&gt; P2P weight sync (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4498962307&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43375&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43375/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43375&quot;&gt;#43375&lt;/a&gt;), rank-local IPC updates (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163032127&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52497&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52497/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52497&quot;&gt;#52497&lt;/a&gt;), sparse checkpoint updates (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040743562&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50723&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50723/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50723&quot;&gt;#50723&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247441451&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53751&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53751/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53751&quot;&gt;#53751&lt;/a&gt;), gpt-oss routed expert loading (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145232591&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52209&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52209/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52209&quot;&gt;#52209&lt;/a&gt;), stable DeepSeek V4 mHC broadcast buffers across weight sync (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5172562231&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52626&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52626/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52626&quot;&gt;#52626&lt;/a&gt;), and packed weight transfer stream reuse capping reserved-memory waste (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5193773321&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52951&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52951/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52951&quot;&gt;#52951&lt;/a&gt;).&lt;/li&gt;132&lt;li&gt;&lt;strong&gt;Determinism&lt;/strong&gt;: &lt;code&gt;trace_decode_token_ids&lt;/code&gt; for deterministic decode replay (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4741800584&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46701&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46701/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46701&quot;&gt;#46701&lt;/a&gt;), per-arch tuned batch-invariant matmul configs (about 3x decode kernels on RTX 4090D/H20, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5212728093&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53247&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53247/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53247&quot;&gt;#53247&lt;/a&gt;), Blackwell autotuning with 33.6% E2E latency reduction (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5240357352&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53649&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53649/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53649&quot;&gt;#53649&lt;/a&gt;), deterministic MoE combine under DP+EP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4664004847&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45683&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45683/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45683&quot;&gt;#45683&lt;/a&gt;), and &lt;code&gt;fuse_allreduce_rms&lt;/code&gt; disabled under batch invariance (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5083618168&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51292&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51292/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51292&quot;&gt;#51292&lt;/a&gt;).&lt;/li&gt;133&lt;li&gt;&lt;strong&gt;Kernels&lt;/strong&gt;: fused embedding kernel (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5242321219&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53677&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53677/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53677&quot;&gt;#53677&lt;/a&gt;), FA4 re-enabled for head_size=256 on Blackwell (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5195404163&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52980&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52980/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52980&quot;&gt;#52980&lt;/a&gt;), vectorized sparse MLA mask loads (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145820471&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52217&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52217/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52217&quot;&gt;#52217&lt;/a&gt;), masked MHA prefill for GLM-5 head dimensions (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5249992008&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53785&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53785/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53785&quot;&gt;#53785&lt;/a&gt;), GLM-5.2 sparse MLA Q concatenation fused with head padding (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5256552481&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53878&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53878/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53878&quot;&gt;#53878&lt;/a&gt;), fused QK-norm + partial MRoPE + gate for Qwen3.6 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5177222479&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52676&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52676/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52676&quot;&gt;#52676&lt;/a&gt;), FlashInfer CuTeDSL BF16 low-latency GEMM as opt-in &lt;code&gt;--linear-backend flashinfer_cutedsl&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5029607342&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50572&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50572/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50572&quot;&gt;#50572&lt;/a&gt;), replicated embedding and norm fusion for DSV3 flat models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4872775914&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48484&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48484/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48484&quot;&gt;#48484&lt;/a&gt;), standardized fused shared-expert selection (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5111501584&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51695&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51695/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51695&quot;&gt;#51695&lt;/a&gt;), GPT-OSS MoE topk metadata reuse (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4652747210&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45457&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45457/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45457&quot;&gt;#45457&lt;/a&gt;), tuned cooperative topk (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5222178122&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53382&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53382/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53382&quot;&gt;#53382&lt;/a&gt;), and tuned FP8 fused_moe for Qwen3.5 on L40S (+7%, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252641623&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53819&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53819/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53819&quot;&gt;#53819&lt;/a&gt;).&lt;/li&gt;134&lt;li&gt;&lt;strong&gt;Robustness&lt;/strong&gt;: KV cache layout standardized under a &lt;code&gt;KVCacheLayout&lt;/code&gt; enum (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5113191409&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51718&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51718/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51718&quot;&gt;#51718&lt;/a&gt;), JIT warmup provider registry (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5002247365&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50174&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50174/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50174&quot;&gt;#50174&lt;/a&gt;), &lt;code&gt;--cpu-offload-params&lt;/code&gt; now reaches vision/audio towers (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5204596603&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53120&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53120/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53120&quot;&gt;#53120&lt;/a&gt;), attention backend probe failures no longer crash init (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5112014582&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51703&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51703/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51703&quot;&gt;#51703&lt;/a&gt;), FlashInfer XQA falls back on unsupported head_dim (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5203887298&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53111&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53111/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53111&quot;&gt;#53111&lt;/a&gt;), FlashInfer prefill LSE normalized before merging to fix prefix-cache logits divergence (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5184047381&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52796&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52796/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52796&quot;&gt;#52796&lt;/a&gt;), seed preserved when a batch mixes seeded and unseeded requests (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124104450&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51866&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51866/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51866&quot;&gt;#51866&lt;/a&gt;), startup thread allocation accounts for local DP workers (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5155009604&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52385&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52385/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52385&quot;&gt;#52385&lt;/a&gt;), int32 overflow fixes in fused SiLU block quant (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5224370355&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53409&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53409/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53409&quot;&gt;#53409&lt;/a&gt;) and LoRA kernels (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199007710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53034&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53034/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53034&quot;&gt;#53034&lt;/a&gt;), a shared-memory race in fused groupwise RMSNorm quantization (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5272351015&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54111&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54111/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54111&quot;&gt;#54111&lt;/a&gt;), BLHNC addressing for FlashInfer sparse MLA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5293970400&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54465&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54465/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54465&quot;&gt;#54465&lt;/a&gt;), Mamba state copy race (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040991926&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50729&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50729/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50729&quot;&gt;#50729&lt;/a&gt;), and a &lt;code&gt;start_profile&lt;/code&gt; no-op after auto-stop (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5121873512&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51839&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51839/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51839&quot;&gt;#51839&lt;/a&gt;).&lt;/li&gt;135&lt;/ul&gt;136&lt;h2&gt;Hardware &amp;amp; Performance&lt;/h2&gt;137&lt;ul&gt;138&lt;li&gt;&lt;strong&gt;NVIDIA&lt;/strong&gt;: DeepSeek V3.2 / GLM-5.2 DSA routed to the optimized CUDA path on all GPUs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5188564287&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52861&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52861/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52861&quot;&gt;#52861&lt;/a&gt;), PCP for DSv3.2 sparse MLA (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5135271047&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52046&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52046/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52046&quot;&gt;#52046&lt;/a&gt;), cuBLAS out_dtype router GEMM on all archs including GB10 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5267649250&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54048&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54048/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54048&quot;&gt;#54048&lt;/a&gt;), FlashInfer all-reduce tuning on SM103 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5218112707&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53318&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53318/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53318&quot;&gt;#53318&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5237365487&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53606&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53606/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53606&quot;&gt;#53606&lt;/a&gt;), FA4 hdim256 on SM100 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5195404163&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52980&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52980/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52980&quot;&gt;#52980&lt;/a&gt;), SM120 sparse MLA fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5090252678&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51395&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51395/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51395&quot;&gt;#51395&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5234791846&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53574&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53574/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53574&quot;&gt;#53574&lt;/a&gt;), DeepSeek V3.2 fused kernel grids hardened for 65k+ token launches (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5154772546&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52381&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52381/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52381&quot;&gt;#52381&lt;/a&gt;), FlashMLA sparse decode workspace fix (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247833448&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53755&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53755/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53755&quot;&gt;#53755&lt;/a&gt;), MNNVL Lamport corruption fix (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5196888523&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53000&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53000/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53000&quot;&gt;#53000&lt;/a&gt;), and opt-in Rubin Docker builds for CUDA 13.4/13.5 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5225808497&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53443&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53443/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53443&quot;&gt;#53443&lt;/a&gt;).&lt;/li&gt;139&lt;li&gt;&lt;strong&gt;AMD ROCm&lt;/strong&gt;: dual-stream decode with hipgraphs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5134061672&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52033&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52033/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52033&quot;&gt;#52033&lt;/a&gt;), W4A4 preshuffled asm GEMM by default (+15% throughput on Llama-3.3-70B MXFP4, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5205961808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53141&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53141/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53141&quot;&gt;#53141&lt;/a&gt;), ROCr/CLR update fixing graph replay segfaults with up to about 20% TPOT improvement (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5244809469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53712&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53712/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53712&quot;&gt;#53712&lt;/a&gt;), fused KDA decode on MI325X (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5149417773&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52293&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52293/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52293&quot;&gt;#52293&lt;/a&gt;), FULL cudagraphs for AITER MLA spec decode (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5072151307&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51171&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51171/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51171&quot;&gt;#51171&lt;/a&gt;), FP8 asm MLA prefill for non-divisor head counts (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5061802707&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51040&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51040/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51040&quot;&gt;#51040&lt;/a&gt;), DCP causal multi-token verification (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5112328984&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51705&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51705/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51705&quot;&gt;#51705&lt;/a&gt;) and prefix cache hits for Kimi-K3 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5236835602&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53598&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53598/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53598&quot;&gt;#53598&lt;/a&gt;), AITER PA gluon decode for MiniMax-M3 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5187682710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52849&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52849/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52849&quot;&gt;#52849&lt;/a&gt;), DeepSeek-V4 fusions for mHC/RMSNorm (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5180504751&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52737&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52737/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52737&quot;&gt;#52737&lt;/a&gt;), C4A top-k (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5189358891&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52882&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52882/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52882&quot;&gt;#52882&lt;/a&gt;), C4 compressor GEMMs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5253661833&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53838&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53838/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53838&quot;&gt;#53838&lt;/a&gt;) and SWA q/kv norm + FP8 quant (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231886266&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53540&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53540/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53540&quot;&gt;#53540&lt;/a&gt;), fused shared experts for block-FP8 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5202501068&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53097&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53097/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53097&quot;&gt;#53097&lt;/a&gt;), CPU offload on ROCm 7.13+ (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4472960194&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43018&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43018/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43018&quot;&gt;#43018&lt;/a&gt;), TheRock 7.14 preview docker (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4983770111&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49925&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49925/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49925&quot;&gt;#49925&lt;/a&gt;), int4/int8 quantization fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5138599752&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52112&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52112/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52112&quot;&gt;#52112&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4915714074&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48998&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48998/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48998&quot;&gt;#48998&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5107635909&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51632&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51632/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51632&quot;&gt;#51632&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5203868298&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53110&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53110/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53110&quot;&gt;#53110&lt;/a&gt;), CUDA graphs captured on the current stream (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252521303&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53818&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53818/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53818&quot;&gt;#53818&lt;/a&gt;), AITER metadata preserved across graph replay (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252708600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53821&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53821/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53821&quot;&gt;#53821&lt;/a&gt;), and improved ROCm detection under WSL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4160497140&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/38434&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/38434/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/38434&quot;&gt;#38434&lt;/a&gt;).&lt;/li&gt;140&lt;li&gt;&lt;strong&gt;Intel XPU&lt;/strong&gt;: INC int4 W4A8 linear backend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5025615988&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50501&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50501/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50501&quot;&gt;#50501&lt;/a&gt;), AutoRound MXFP8 MoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5079981233&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51248&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51248/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51248&quot;&gt;#51248&lt;/a&gt;), EC connector KV offloading (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4956097859&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49532&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49532/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49532&quot;&gt;#49532&lt;/a&gt;), and fixes for HunyuanOCR XD-RoPE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142330814&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52174&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52174/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52174&quot;&gt;#52174&lt;/a&gt;), sparse-MLA metadata sync (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5136054941&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52066&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52066/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52066&quot;&gt;#52066&lt;/a&gt;), Mamba state pointer overflow (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4844992467&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48109&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48109/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48109&quot;&gt;#48109&lt;/a&gt;), oneCCL warm-up at world size 1 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5155455899&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52389&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52389/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52389&quot;&gt;#52389&lt;/a&gt;), and MRoPE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5210099570&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53201&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53201/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53201&quot;&gt;#53201&lt;/a&gt;).&lt;/li&gt;141&lt;li&gt;&lt;strong&gt;CPU&lt;/strong&gt;: AMX high-performance MLA backend for DeepSeek V2/V3/R1 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5171924819&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52616&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52616/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52616&quot;&gt;#52616&lt;/a&gt;) with MLA now running end-to-end (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5095464600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51471&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51471/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51471&quot;&gt;#51471&lt;/a&gt;), FP16/BF16 persisted GDN state on AMX (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5143816100&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52191&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52191/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52191&quot;&gt;#52191&lt;/a&gt;), Int8 MoE through zentorch on AMD Zen (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4610349549&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44834&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44834/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44834&quot;&gt;#44834&lt;/a&gt;), Voxtral support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5259223723&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53921&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53921/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53921&quot;&gt;#53921&lt;/a&gt;), C++ causal_conv1d GDN on non-AMX AVX-512BF16 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4966314300&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49688&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49688/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49688&quot;&gt;#49688&lt;/a&gt;), and assorted fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5267272744&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54042&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54042/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54042&quot;&gt;#54042&lt;/a&gt;).&lt;/li&gt;142&lt;/ul&gt;143&lt;h2&gt;Large Scale Serving&lt;/h2&gt;144&lt;ul&gt;145&lt;li&gt;&lt;strong&gt;Context parallelism&lt;/strong&gt;: Kimi-K3 DCP with DSpark (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5143408851&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52188&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52188/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52188&quot;&gt;#52188&lt;/a&gt;) and partial prefix cache hits (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5024902326&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50493&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50493/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50493&quot;&gt;#50493&lt;/a&gt;), FlashInfer native CP for MLA decode (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5265035147&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54012&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54012/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54012&quot;&gt;#54012&lt;/a&gt;), FlashMLA sparse DCP on Hopper with MTP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4727193891&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46514&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46514/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46514&quot;&gt;#46514&lt;/a&gt;), NIXL P/D DCP for MLA models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5032346193&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50611&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50611/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50611&quot;&gt;#50611&lt;/a&gt;), PCP for DSv3.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5135271047&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52046&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52046/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52046&quot;&gt;#52046&lt;/a&gt;) and NIXL PCP producers (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5182876833&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52779&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52779/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52779&quot;&gt;#52779&lt;/a&gt;), &lt;code&gt;--dcp-q-replicate&lt;/code&gt; with query replication default-on for GLM sparse attention (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5016269892&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50382&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50382/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50382&quot;&gt;#50382&lt;/a&gt;), DCP fused attention fix for DeepSeek-V3.2 / GLM-5.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4989595181&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50005&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50005/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50005&quot;&gt;#50005&lt;/a&gt;), sparse MLA metadata and kernel block sizes under DCP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5154614070&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52377&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52377/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52377&quot;&gt;#52377&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5061299849&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51031&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51031/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51031&quot;&gt;#51031&lt;/a&gt;), PCP PIECEWISE cudagraph fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5255923997&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53869&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53869/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53869&quot;&gt;#53869&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5230910486&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53515&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53515/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53515&quot;&gt;#53515&lt;/a&gt;), and PCP compatibility checks delegated to the PCP manager so plugins can enable GQA+PCP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5254301913&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53853&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53853/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53853&quot;&gt;#53853&lt;/a&gt;).&lt;/li&gt;146&lt;li&gt;&lt;strong&gt;Elastic EP&lt;/strong&gt;: reduced eager-mode reconfiguration downtime (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5125332483&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51885&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51885/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51885&quot;&gt;#51885&lt;/a&gt;), AOT cache reuse preserved during scaling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5221788715&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53378&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53378/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53378&quot;&gt;#53378&lt;/a&gt;), and scale-below-minimum rejected (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178794791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52702&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52702/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52702&quot;&gt;#52702&lt;/a&gt;).&lt;/li&gt;147&lt;li&gt;&lt;strong&gt;MoE communication&lt;/strong&gt;: DeepEP v2 receiver CPU overhead (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5067252911&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51114&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51114/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51114&quot;&gt;#51114&lt;/a&gt;) and MXFP8 activation scale dispatch (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5090849670&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51398&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51398/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51398&quot;&gt;#51398&lt;/a&gt;), FlashInfer one-sided All2All refinements (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5127258176&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51924&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51924/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51924&quot;&gt;#51924&lt;/a&gt;), DeepEP v2 fixes for &lt;code&gt;--enforce-eager&lt;/code&gt; startup (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5120810925&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51824&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51824/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51824&quot;&gt;#51824&lt;/a&gt;) and the decode/cudagraph path (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5173468412&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52632&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52632/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52632&quot;&gt;#52632&lt;/a&gt;), NCCL&amp;gt;=2.31 heap overflow fix (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5197472419&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53008&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53008/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53008&quot;&gt;#53008&lt;/a&gt;), cross-node MNNVL all-reduce gated by capability (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5213350847&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53253&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53253/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53253&quot;&gt;#53253&lt;/a&gt;), and MNNVL all-reduce buffers sized for DSpark (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5055170821&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50932&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50932/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50932&quot;&gt;#50932&lt;/a&gt;).&lt;/li&gt;148&lt;li&gt;&lt;strong&gt;KV connectors&lt;/strong&gt;: Mooncake Store decode KV saving via &lt;code&gt;save_decode_cache&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5160954650&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52466&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52466/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52466&quot;&gt;#52466&lt;/a&gt;) and hybrid DCP prefix caching (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5218350653&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53324&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53324/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53324&quot;&gt;#53324&lt;/a&gt;), externally transferable KV cache group identification (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5249690529&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53779&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53779/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53779&quot;&gt;#53779&lt;/a&gt;), async KV loads deferred past forward launch (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5219004789&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53333&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53333/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53333&quot;&gt;#53333&lt;/a&gt;), &lt;code&gt;kv_transfer_params&lt;/code&gt; for &lt;code&gt;/inference/v1/generate&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4446567128&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/42644&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/42644/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/42644&quot;&gt;#42644&lt;/a&gt;), MoRIIO shared KV region registration after the layout refactor (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5243691876&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53698&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53698/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53698&quot;&gt;#53698&lt;/a&gt;), and Mamba fixes for Mooncake (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5087157013&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51362&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51362/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51362&quot;&gt;#51362&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5086979993&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51358&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51358/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51358&quot;&gt;#51358&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5241480015&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53663&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53663/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53663&quot;&gt;#53663&lt;/a&gt;) and NIXL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231165743&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53523&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53523/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53523&quot;&gt;#53523&lt;/a&gt;).&lt;/li&gt;149&lt;li&gt;&lt;strong&gt;KV offloading&lt;/strong&gt;: EC offloading connector driven by CUDA events (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4989030901&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49994&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49994/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49994&quot;&gt;#49994&lt;/a&gt;), ownership in KV cache events (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5136187894&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52067&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52067/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52067&quot;&gt;#52067&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5136206962&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52068&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52068/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52068&quot;&gt;#52068&lt;/a&gt;), P2P tier request-level offload (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5190792789&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52912&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52912/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52912&quot;&gt;#52912&lt;/a&gt;) and abort handling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5168113159&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52571&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52571/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52571&quot;&gt;#52571&lt;/a&gt;), &lt;code&gt;/dev/shm&lt;/code&gt; leak on crash fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5170095777&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52596&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52596/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52596&quot;&gt;#52596&lt;/a&gt;), CPU-&amp;gt;GPU loads ordered against compute stream (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5038239407&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50696&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50696/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50696&quot;&gt;#50696&lt;/a&gt;), &lt;code&gt;store_threshold&lt;/code&gt; counting fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5146372859&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52227&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52227/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52227&quot;&gt;#52227&lt;/a&gt;), in-flight primary keys cascaded (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5218859213&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53329&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53329/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53329&quot;&gt;#53329&lt;/a&gt;), and padded GPU cache storage handled (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5265467470&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54021&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54021/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54021&quot;&gt;#54021&lt;/a&gt;).&lt;/li&gt;150&lt;li&gt;&lt;strong&gt;Data/pipeline parallel&lt;/strong&gt;: PP silent corruption fix (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5324060236&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54962&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54962/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54962&quot;&gt;#54962&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4935159122&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49274&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49274/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49274&quot;&gt;#49274&lt;/a&gt;), DP coordinator wake handling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5095868711&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51481&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51481/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51481&quot;&gt;#51481&lt;/a&gt;), device sync on pause (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5190917705&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52914&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52914/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52914&quot;&gt;#52914&lt;/a&gt;), TCPStore port fixes for Ray (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5241656210&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53666&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53666/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53666&quot;&gt;#53666&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5056901305&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50969&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50969/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50969&quot;&gt;#50969&lt;/a&gt;), DP supervisor inheriting the uvicorn config (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5161531915&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52473&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52473/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52473&quot;&gt;#52473&lt;/a&gt;), EPD encoder round-robin fix (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5162595398&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52491&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52491/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52491&quot;&gt;#52491&lt;/a&gt;), and producer-only EC config normalization (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5241165547&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53656&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53656/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53656&quot;&gt;#53656&lt;/a&gt;).&lt;/li&gt;151&lt;/ul&gt;152&lt;h2&gt;Quantization&lt;/h2&gt;153&lt;ul&gt;154&lt;li&gt;&lt;strong&gt;New backends&lt;/strong&gt;: FlashInfer TRT-LLM MXFP8 linear (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5144692620&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52204&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52204/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52204&quot;&gt;#52204&lt;/a&gt;), b12x FP4 MoE for SM120/SM121 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5132887932&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52018&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52018/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52018&quot;&gt;#52018&lt;/a&gt;), AutoRound block-wise FP8 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4794241679&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47434&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47434/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47434&quot;&gt;#47434&lt;/a&gt;), Humming MoE with MXFP4 weights + block-FP8 activations (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5085703484&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51332&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51332/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51332&quot;&gt;#51332&lt;/a&gt;), and Humming for compressed-tensors WNA16 MoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4908628525&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48918&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48918/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48918&quot;&gt;#48918&lt;/a&gt;).&lt;/li&gt;155&lt;li&gt;&lt;strong&gt;Fixes&lt;/strong&gt;: weight-only NVFP4 checkpoints routed through W4A16 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5292310719&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54427&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54427/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54427&quot;&gt;#54427&lt;/a&gt;), compressed-tensors block FP8 with Marlin (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5194512373&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52966&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52966/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52966&quot;&gt;#52966&lt;/a&gt;) and int8 grouped WNA16 MoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5132036711&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52002&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52002/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52002&quot;&gt;#52002&lt;/a&gt;), OCP MX mxfp6 activation emulation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178960258&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52704&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52704/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52704&quot;&gt;#52704&lt;/a&gt;), MXFP8 FlashInfer path guarded on availability (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5174715929&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52648&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52648/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52648&quot;&gt;#52648&lt;/a&gt;), and Humming activation aliasing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5268166858&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54056&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54056/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54056&quot;&gt;#54056&lt;/a&gt;).&lt;/li&gt;156&lt;/ul&gt;157&lt;h2&gt;API &amp;amp; Frontend&lt;/h2&gt;158&lt;ul&gt;159&lt;li&gt;&lt;strong&gt;New endpoints &amp;amp; options&lt;/strong&gt;: &lt;code&gt;/v1/messages/render&lt;/code&gt; for the Anthropic Messages API (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4672953156&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45803&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45803/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45803&quot;&gt;#45803&lt;/a&gt;), &lt;code&gt;/cohere/v2/chat/render&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5211362691&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53219&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53219/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53219&quot;&gt;#53219&lt;/a&gt;), SSE keep-alive comments for idle streams (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5061493469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51034&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51034/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51034&quot;&gt;#51034&lt;/a&gt;), per-request spec decode metrics (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4908287242&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48915&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48915/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48915&quot;&gt;#48915&lt;/a&gt;), video embeds input (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5280853540&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54242&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54242/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54242&quot;&gt;#54242&lt;/a&gt;), &lt;code&gt;--max-num-queued-reqs&lt;/code&gt; / &lt;code&gt;--max-num-queued-tokens&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4948101157&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49445&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49445/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49445&quot;&gt;#49445&lt;/a&gt;), and pooling requests can set padding (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5070709133&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51157&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51157/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51157&quot;&gt;#51157&lt;/a&gt;).&lt;/li&gt;160&lt;li&gt;&lt;strong&gt;Anthropic &amp;amp; Cohere&lt;/strong&gt;: &lt;code&gt;vllm_xargs&lt;/code&gt; forwarded to sampling params (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5217217428&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53308&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53308/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53308&quot;&gt;#53308&lt;/a&gt;), &lt;code&gt;stop_sequence&lt;/code&gt; stop reason reported (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4673469165&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45807&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45807/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45807&quot;&gt;#45807&lt;/a&gt;), Cohere citation/tool 500s fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142417361&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52175&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52175/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52175&quot;&gt;#52175&lt;/a&gt;), and stop string limit applied (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247377261&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53750&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53750/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53750&quot;&gt;#53750&lt;/a&gt;).&lt;/li&gt;161&lt;li&gt;&lt;strong&gt;OpenAI compatibility&lt;/strong&gt;: &lt;code&gt;logprobs=-1&lt;/code&gt; in Completions (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4702354344&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46175&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46175/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46175&quot;&gt;#46175&lt;/a&gt;), streamed logprob offsets with echo (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4825254121&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47815&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47815/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47815&quot;&gt;#47815&lt;/a&gt;), batched chat echo (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5164651710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52529&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52529/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52529&quot;&gt;#52529&lt;/a&gt;), all choices returned from &lt;code&gt;/inference/v1/generate&lt;/code&gt; with n&amp;gt;1 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5156330156&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52399&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52399/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52399&quot;&gt;#52399&lt;/a&gt;), &lt;code&gt;cache_salt&lt;/code&gt; forwarded for content parts (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5284869290&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54315&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54315/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54315&quot;&gt;#54315&lt;/a&gt;), &lt;code&gt;stop_token_ids&lt;/code&gt; validated against vocab (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5277040609&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54196&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54196/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54196&quot;&gt;#54196&lt;/a&gt;), and 4xx for client errors in &lt;code&gt;/detokenize&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5172198842&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52622&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52622/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52622&quot;&gt;#52622&lt;/a&gt;), malformed namespace tools (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5248350117&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53763&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53763/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53763&quot;&gt;#53763&lt;/a&gt;), malformed base64 audio (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247065549&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53744&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53744/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53744&quot;&gt;#53744&lt;/a&gt;), and unknown chat roles in DeepSeek encoders (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5201157754&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53071&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53071/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53071&quot;&gt;#53071&lt;/a&gt;); tool-call arguments in replayed history are parsed defensively (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4909019190&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48922&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48922/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48922&quot;&gt;#48922&lt;/a&gt;), empty &lt;code&gt;FlatLogprobs&lt;/code&gt; slices are handled (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5244170627&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53704&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53704/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53704&quot;&gt;#53704&lt;/a&gt;), and empty bad-word tokenizations are rejected (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5225478969&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53433&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53433/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53433&quot;&gt;#53433&lt;/a&gt;).&lt;/li&gt;162&lt;li&gt;&lt;strong&gt;Structured output &amp;amp; parsers&lt;/strong&gt;: reasoning-end detection scoped to the current turn (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5270715291&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54089&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54089/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54089&quot;&gt;#54089&lt;/a&gt;), terminal grammars stop under &lt;code&gt;min_tokens&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5279444949&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54218&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54218/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54218&quot;&gt;#54218&lt;/a&gt;), unsupported pattern+length schemas rejected (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4989352366&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49996&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49996/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49996&quot;&gt;#49996&lt;/a&gt;), MistralCommonBackend tokenizers (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179888512&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52720&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52720/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52720&quot;&gt;#52720&lt;/a&gt;), XGrammar termination in batches (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5184712897&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52805&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52805/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52805&quot;&gt;#52805&lt;/a&gt;), spurious FSM errors after speculative reasoning end (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199654396&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53046&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53046/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53046&quot;&gt;#53046&lt;/a&gt;), shared parser engine adapters (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5186240533&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52830&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52830/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52830&quot;&gt;#52830&lt;/a&gt;), Gemma4 parenthesized tool calls (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5241193116&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53657&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53657/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53657&quot;&gt;#53657&lt;/a&gt;) and &lt;code&gt;enable_thinking&lt;/code&gt; default (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5157922888&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52430&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52430/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52430&quot;&gt;#52430&lt;/a&gt;), HY-V3 parallel calls in one delta (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5262459954&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53965&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53965/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53965&quot;&gt;#53965&lt;/a&gt;), GPT-OSS Harmony strict grammar (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5146121681&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52222&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52222/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52222&quot;&gt;#52222&lt;/a&gt;), Kimi K3 reserved markers excluded from response text (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5189523224&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52889&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52889/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52889&quot;&gt;#52889&lt;/a&gt;), and unused request-local reasoners skipped (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5168474415&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52573&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52573/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52573&quot;&gt;#52573&lt;/a&gt;).&lt;/li&gt;163&lt;li&gt;&lt;strong&gt;Rust frontend&lt;/strong&gt;: HY3 unified parser with local XGrammar structural tags (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5200108726&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53054&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53054/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53054&quot;&gt;#53054&lt;/a&gt;), gRPC audio/video inputs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5248097964&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53760&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53760/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53760&quot;&gt;#53760&lt;/a&gt;), gRPC LoRA lifecycle control (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5187034854&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52840&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52840/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52840&quot;&gt;#52840&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5133870428&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52031&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52031/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52031&quot;&gt;#52031&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5247940651&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53756&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53756/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53756&quot;&gt;#53756&lt;/a&gt;), &lt;code&gt;--generation-config vllm&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199448971&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53044&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53044/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53044&quot;&gt;#53044&lt;/a&gt;), &lt;code&gt;truncate_prompt_tokens&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4881046433&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48584&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48584/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48584&quot;&gt;#48584&lt;/a&gt;), OpenAI edge-case alignment (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5211336411&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53218&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53218/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53218&quot;&gt;#53218&lt;/a&gt;), optimized SSE hot path (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5085325841&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51321&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51321/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51321&quot;&gt;#51321&lt;/a&gt;), pure-Rust &lt;code&gt;protox&lt;/code&gt; replacing &lt;code&gt;protoc&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5189704536&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52892&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52892/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52892&quot;&gt;#52892&lt;/a&gt;), RL world-size reporting (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5210206507&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53204&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53204/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53204&quot;&gt;#53204&lt;/a&gt;) and routed expert prompt offsets (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178920917&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52703&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52703/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52703&quot;&gt;#52703&lt;/a&gt;), and fixes for GLM-5.2 template parity (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5092455028&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51426&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51426/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51426&quot;&gt;#51426&lt;/a&gt;), Qwen parser auto-detection (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5072051228&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51169&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51169/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51169&quot;&gt;#51169&lt;/a&gt;), Kimi K3 &lt;code&gt;reasoning_effort=&quot;none&quot;&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199401760&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53043&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53043/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53043&quot;&gt;#53043&lt;/a&gt;), &lt;code&gt;n &amp;gt; 1&lt;/code&gt; rejection on &lt;code&gt;/inference/v1/generate&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5187627320&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52844&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52844/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52844&quot;&gt;#52844&lt;/a&gt;), and the &lt;code&gt;LogprobsTensors&lt;/code&gt; wire schema (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5260458913&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53939&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53939/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53939&quot;&gt;#53939&lt;/a&gt;).&lt;/li&gt;164&lt;li&gt;&lt;strong&gt;Pooling&lt;/strong&gt;: BGE-M3 throughput (+3.13%, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5227627215&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53464&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53464/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53464&quot;&gt;#53464&lt;/a&gt;) and task validation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5120627199&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51823&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51823/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51823&quot;&gt;#51823&lt;/a&gt;), prompts truncated before padding (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5287568175&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54364&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54364/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54364&quot;&gt;#54364&lt;/a&gt;), parallel arrays truncated with the prompt (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5291138665&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54407&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54407/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54407&quot;&gt;#54407&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5296990904&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54509&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54509/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54509&quot;&gt;#54509&lt;/a&gt;), and batched input throughput in &lt;code&gt;vllm bench serve&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5210690624&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53213&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53213/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53213&quot;&gt;#53213&lt;/a&gt;).&lt;/li&gt;165&lt;li&gt;&lt;strong&gt;CLI &amp;amp; tooling&lt;/strong&gt;: &lt;code&gt;vllm launch&lt;/code&gt; runs the serve argument checks (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185940058&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52825&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52825/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52825&quot;&gt;#52825&lt;/a&gt;), &lt;code&gt;run_batch.py&lt;/code&gt; moved out of the openai folder (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5230582851&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53500&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53500/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53500&quot;&gt;#53500&lt;/a&gt;) with its upload retry fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5031185594&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50588&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50588/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50588&quot;&gt;#50588&lt;/a&gt;), &lt;code&gt;vllm bench&lt;/code&gt; warns on a warm prefix cache for random runs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5259210791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53920&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53920/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53920&quot;&gt;#53920&lt;/a&gt;) and restores multimodal datasets on the &lt;code&gt;vllm&lt;/code&gt; throughput backend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5141922176&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52168&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52168/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52168&quot;&gt;#52168&lt;/a&gt;), and the vLLM recipes tool gained sweep recommendations and alias parsing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5218372543&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53325&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53325/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53325&quot;&gt;#53325&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5260815323&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53946&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53946/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53946&quot;&gt;#53946&lt;/a&gt;).&lt;/li&gt;166&lt;/ul&gt;167&lt;h2&gt;Security&lt;/h2&gt;168&lt;ul&gt;169&lt;li&gt;&lt;code&gt;cache_salt&lt;/code&gt; length bounded to 1024 to prevent scheduler CPU exhaustion (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5286769014&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54353&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54353/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54353&quot;&gt;#54353&lt;/a&gt;).&lt;/li&gt;170&lt;li&gt;Oversized media rejected before full download (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5125680263&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51896&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51896/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51896&quot;&gt;#51896&lt;/a&gt;); &lt;code&gt;VLLM_MAX_AUDIO_CLIP_FILESIZE_MB&lt;/code&gt; enforced on all audio paths (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5233508812&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53561&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53561/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53561&quot;&gt;#53561&lt;/a&gt;).&lt;/li&gt;171&lt;li&gt;Decoder prompt-length validation enforced for processors that skip the check (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4732705502&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46588&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46588/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46588&quot;&gt;#46588&lt;/a&gt;); PyNvVideoCodec decoder slot limit bypass fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5139393086&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52126&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52126/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52126&quot;&gt;#52126&lt;/a&gt;).&lt;/li&gt;172&lt;li&gt;&lt;code&gt;api_key&lt;/code&gt; and &lt;code&gt;hf_token&lt;/code&gt; redacted from startup logs, compile cache factors, and the Rust frontend launch log (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5164415222&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52523&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52523/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52523&quot;&gt;#52523&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5238741378&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53625&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53625/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53625&quot;&gt;#53625&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5246943808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53738&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53738/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53738&quot;&gt;#53738&lt;/a&gt;).&lt;/li&gt;173&lt;/ul&gt;174&lt;h2&gt;Dependencies&lt;/h2&gt;175&lt;ul&gt;176&lt;li&gt;FlashInfer 0.6.18 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5284825458&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54313&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54313/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54313&quot;&gt;#54313&lt;/a&gt;), huggingface-hub 1.28.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5184048784&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52797&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52797/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52797&quot;&gt;#52797&lt;/a&gt;), tpu-inference v0.28.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5265461395&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54020&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54020/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54020&quot;&gt;#54020&lt;/a&gt;), NIXL 1.3.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117694715&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51777&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51777/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51777&quot;&gt;#51777&lt;/a&gt;).&lt;/li&gt;177&lt;li&gt;InstantTensor added to CUDA dependencies; the loader is not enabled by default (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5184105735&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52801&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52801/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52801&quot;&gt;#52801&lt;/a&gt;). CuPy constraint relaxed to exclude only 14.1.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4569052177&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44284&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44284/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44284&quot;&gt;#44284&lt;/a&gt;).&lt;/li&gt;178&lt;li&gt;ROCm base image: ROCr/CLR update (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5244809469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53712&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53712/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53712&quot;&gt;#53712&lt;/a&gt;), rocprofiler-sdk 1.3.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5209421339&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53182&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53182/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53182&quot;&gt;#53182&lt;/a&gt;), TheRock 7.14 preview (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4983770111&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49925&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49925/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49925&quot;&gt;#49925&lt;/a&gt;), LMCache connector packages (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5076186627&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51208&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51208/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51208&quot;&gt;#51208&lt;/a&gt;).&lt;/li&gt;179&lt;li&gt;XPU: vllm_xpu_kernels 0.1.14.1 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5277684434&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54203&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54203/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54203&quot;&gt;#54203&lt;/a&gt;), UCX install updated (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252425369&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53817&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53817/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53817&quot;&gt;#53817&lt;/a&gt;).&lt;/li&gt;180&lt;li&gt;Build fails closed when the selected precompiled CUDA variant is unavailable (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5165593048&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52545&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52545/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52545&quot;&gt;#52545&lt;/a&gt;).&lt;/li&gt;181&lt;/ul&gt;182&lt;h2&gt;Breaking Changes &amp;amp; Deprecations&lt;/h2&gt;183&lt;ul&gt;184&lt;li&gt;Ten deprecated architectures removed: Arctic, Chameleon, Cheers, Fairseq2Llama, FireRedLID, GritLM, HCXVision, MPT, the &lt;code&gt;RWForCausalLM&lt;/code&gt; and &lt;code&gt;StableLMEpochForCausalLM&lt;/code&gt; aliases, and &lt;code&gt;PrithviGeoSpatialMAE&lt;/code&gt; (superseded by &lt;code&gt;Terratorch&lt;/code&gt;) (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5237599903&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53608&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53608/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53608&quot;&gt;#53608&lt;/a&gt;).&lt;/li&gt;185&lt;li&gt;FlexOlmo, Olmo3, Hunyuan V1 and Hunyuan VL are now served through the Transformers modeling backend (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5238189101&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53615&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53615/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53615&quot;&gt;#53615&lt;/a&gt;).&lt;/li&gt;186&lt;li&gt;PyAV video decoder backend removed; use OpenCV or Torchcodec (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5279922155&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54231&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54231/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54231&quot;&gt;#54231&lt;/a&gt;).&lt;/li&gt;187&lt;li&gt;&lt;code&gt;python -m vllm.entrypoints.openai.api_server&lt;/code&gt; is deprecated; use &lt;code&gt;vllm serve&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5139471549&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52131&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52131/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52131&quot;&gt;#52131&lt;/a&gt;).&lt;/li&gt;188&lt;li&gt;Model Runner V2 is the default runner for all models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5209436023&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53183&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53183/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53183&quot;&gt;#53183&lt;/a&gt;).&lt;/li&gt;189&lt;li&gt;&lt;code&gt;prefix_cache_retention_interval&lt;/code&gt; default changed from dense to 0 for SWA/SSM models; the env var is deprecated in favor of the argument (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145771808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52216&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52216/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52216&quot;&gt;#52216&lt;/a&gt;).&lt;/li&gt;190&lt;li&gt;FlashInfer all-reduce enabled by default (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5196781600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52998&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52998/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52998&quot;&gt;#52998&lt;/a&gt;).&lt;/li&gt;191&lt;li&gt;&lt;code&gt;VLLM_TEST_FORCE_FP8_MARLIN&lt;/code&gt; removed in favor of &lt;code&gt;--linear-backend&lt;/code&gt; / &lt;code&gt;--moe-backend&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142686969&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52182&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52182/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52182&quot;&gt;#52182&lt;/a&gt;); &lt;code&gt;VLLM_ROCM_USE_AITER_FP4_ASM_GEMM&lt;/code&gt; removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5205961808&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53141&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53141/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53141&quot;&gt;#53141&lt;/a&gt;); dead &lt;code&gt;--attention-config.use_prefill_decode_attention&lt;/code&gt; removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5167066764&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52557&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52557/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52557&quot;&gt;#52557&lt;/a&gt;); other long-deprecated parameters cleaned up (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5233285489&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53559&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53559/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53559&quot;&gt;#53559&lt;/a&gt;).&lt;/li&gt;192&lt;/ul&gt;193&lt;h2&gt;New Contributors&lt;/h2&gt;194&lt;ul&gt;195&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/030611/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/030611&quot;&gt;@030611&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5120627199&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51823&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51823/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51823&quot;&gt;#51823&lt;/a&gt;&lt;/li&gt;196&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/92hyungjun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/92hyungjun&quot;&gt;@92hyungjun&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4783990461&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47272&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47272/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47272&quot;&gt;#47272&lt;/a&gt;&lt;/li&gt;197&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ActiveSky/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ActiveSky&quot;&gt;@ActiveSky&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178143135&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52692&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52692/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52692&quot;&gt;#52692&lt;/a&gt;&lt;/li&gt;198&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/adisivaprasad/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/adisivaprasad&quot;&gt;@adisivaprasad&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5254304672&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53854&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53854/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53854&quot;&gt;#53854&lt;/a&gt;&lt;/li&gt;199&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Agoni-02/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Agoni-02&quot;&gt;@Agoni-02&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4902578074&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48850&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48850/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48850&quot;&gt;#48850&lt;/a&gt;&lt;/li&gt;200&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AmitMY/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AmitMY&quot;&gt;@AmitMY&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4882623102&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48608&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48608/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48608&quot;&gt;#48608&lt;/a&gt;&lt;/li&gt;201&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Andy365-365/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Andy365-365&quot;&gt;@Andy365-365&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5164415222&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52523&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52523/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52523&quot;&gt;#52523&lt;/a&gt;&lt;/li&gt;202&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/andyluo7/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/andyluo7&quot;&gt;@andyluo7&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252708600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53821&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53821/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53821&quot;&gt;#53821&lt;/a&gt;&lt;/li&gt;203&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AnkitNakhawa/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AnkitNakhawa&quot;&gt;@AnkitNakhawa&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5162595398&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52491&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52491/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52491&quot;&gt;#52491&lt;/a&gt;&lt;/li&gt;204&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anmolgupt/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anmolgupt&quot;&gt;@anmolgupt&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5189161450&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52874&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52874/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52874&quot;&gt;#52874&lt;/a&gt;&lt;/li&gt;205&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bobboli/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bobboli&quot;&gt;@bobboli&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5127258176&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51924&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51924/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51924&quot;&gt;#51924&lt;/a&gt;&lt;/li&gt;206&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/brianosaurus/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/brianosaurus&quot;&gt;@brianosaurus&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5167066764&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52557&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52557/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52557&quot;&gt;#52557&lt;/a&gt;&lt;/li&gt;207&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CalvinXKY/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CalvinXKY&quot;&gt;@CalvinXKY&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5049295618&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50858&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50858/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50858&quot;&gt;#50858&lt;/a&gt;&lt;/li&gt;208&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/canlahlah/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/canlahlah&quot;&gt;@canlahlah&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5224370355&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53409&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53409/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53409&quot;&gt;#53409&lt;/a&gt;&lt;/li&gt;209&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CherryLemon/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CherryLemon&quot;&gt;@CherryLemon&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5256506704&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53877&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53877/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53877&quot;&gt;#53877&lt;/a&gt;&lt;/li&gt;210&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CHIPMUNK-T0T/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CHIPMUNK-T0T&quot;&gt;@CHIPMUNK-T0T&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4811392083&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47625&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47625/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47625&quot;&gt;#47625&lt;/a&gt;&lt;/li&gt;211&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cogniera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cogniera&quot;&gt;@cogniera&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5253709794&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53839&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53839/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53839&quot;&gt;#53839&lt;/a&gt;&lt;/li&gt;212&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cr-zhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cr-zhao&quot;&gt;@cr-zhao&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5155009604&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52385&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52385/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52385&quot;&gt;#52385&lt;/a&gt;&lt;/li&gt;213&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/daviswer/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/daviswer&quot;&gt;@daviswer&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179178405&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52706&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52706/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52706&quot;&gt;#52706&lt;/a&gt;&lt;/li&gt;214&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/DCoEngine/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/DCoEngine&quot;&gt;@DCoEngine&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4888766885&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48682&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48682/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48682&quot;&gt;#48682&lt;/a&gt;&lt;/li&gt;215&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dineshchitlangia/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dineshchitlangia&quot;&gt;@dineshchitlangia&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4966314300&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49688&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49688/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49688&quot;&gt;#49688&lt;/a&gt;&lt;/li&gt;216&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dkrisman/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dkrisman&quot;&gt;@dkrisman&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5289377865&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54380&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54380/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54380&quot;&gt;#54380&lt;/a&gt;&lt;/li&gt;217&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmvevents/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmvevents&quot;&gt;@dmvevents&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5173468412&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52632&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52632/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52632&quot;&gt;#52632&lt;/a&gt;&lt;/li&gt;218&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Edge-Explorer/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Edge-Explorer&quot;&gt;@Edge-Explorer&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5225511425&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53435&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53435/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53435&quot;&gt;#53435&lt;/a&gt;&lt;/li&gt;219&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/eilamc14/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/eilamc14&quot;&gt;@eilamc14&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4812313997&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47640&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47640/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47640&quot;&gt;#47640&lt;/a&gt;&lt;/li&gt;220&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/eligotts/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/eligotts&quot;&gt;@eligotts&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5284869290&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54315&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54315/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54315&quot;&gt;#54315&lt;/a&gt;&lt;/li&gt;221&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Eoin-Houstoun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Eoin-Houstoun&quot;&gt;@Eoin-Houstoun&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5112014582&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51703&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51703/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51703&quot;&gt;#51703&lt;/a&gt;&lt;/li&gt;222&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/floatlibai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/floatlibai&quot;&gt;@floatlibai&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5140155788&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52144&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52144/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52144&quot;&gt;#52144&lt;/a&gt;&lt;/li&gt;223&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fuzzifikation/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fuzzifikation&quot;&gt;@fuzzifikation&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5061493469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51034&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51034/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51034&quot;&gt;#51034&lt;/a&gt;&lt;/li&gt;224&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hagaikwa-redhat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hagaikwa-redhat&quot;&gt;@hagaikwa-redhat&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5211678040&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53230&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53230/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53230&quot;&gt;#53230&lt;/a&gt;&lt;/li&gt;225&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/haoyangqian/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/haoyangqian&quot;&gt;@haoyangqian&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5177840622&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52690&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52690/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52690&quot;&gt;#52690&lt;/a&gt;&lt;/li&gt;226&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Hert4/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Hert4&quot;&gt;@Hert4&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5070709133&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51157&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51157/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51157&quot;&gt;#51157&lt;/a&gt;&lt;/li&gt;227&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ima-helikoptaaa/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ima-helikoptaaa&quot;&gt;@ima-helikoptaaa&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5292310719&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/54427&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/54427/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/54427&quot;&gt;#54427&lt;/a&gt;&lt;/li&gt;228&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jbyczkow/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jbyczkow&quot;&gt;@jbyczkow&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142330814&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52174&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52174/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52174&quot;&gt;#52174&lt;/a&gt;&lt;/li&gt;229&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JC-ut0/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JC-ut0&quot;&gt;@JC-ut0&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5201157754&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53071&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53071/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53071&quot;&gt;#53071&lt;/a&gt;&lt;/li&gt;230&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jiahaoliang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jiahaoliang&quot;&gt;@jiahaoliang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5080909185&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51262&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51262/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51262&quot;&gt;#51262&lt;/a&gt;&lt;/li&gt;231&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JiataiWang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JiataiWang&quot;&gt;@JiataiWang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5232825783&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53553&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53553/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53553&quot;&gt;#53553&lt;/a&gt;&lt;/li&gt;232&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jl9876/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jl9876&quot;&gt;@jl9876&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5079981233&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51248&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51248/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51248&quot;&gt;#51248&lt;/a&gt;&lt;/li&gt;233&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JulianZJN/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JulianZJN&quot;&gt;@JulianZJN&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5213350847&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53253&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53253/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53253&quot;&gt;#53253&lt;/a&gt;&lt;/li&gt;234&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jungjiyu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jungjiyu&quot;&gt;@jungjiyu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5260458913&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53939&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53939/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53939&quot;&gt;#53939&lt;/a&gt;&lt;/li&gt;235&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/kyleliang-nv/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/kyleliang-nv&quot;&gt;@kyleliang-nv&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5075534328&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51203&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51203/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51203&quot;&gt;#51203&lt;/a&gt;&lt;/li&gt;236&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LH-and-FPGA/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LH-and-FPGA&quot;&gt;@LH-and-FPGA&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5174715929&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52648&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52648/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52648&quot;&gt;#52648&lt;/a&gt;&lt;/li&gt;237&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/li-ukumar/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/li-ukumar&quot;&gt;@li-ukumar&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5168113159&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52571&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52571/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52571&quot;&gt;#52571&lt;/a&gt;&lt;/li&gt;238&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LioEinaudi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LioEinaudi&quot;&gt;@LioEinaudi&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5212728093&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53247&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53247/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53247&quot;&gt;#53247&lt;/a&gt;&lt;/li&gt;239&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LironKesem/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LironKesem&quot;&gt;@LironKesem&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5259223723&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53921&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53921/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53921&quot;&gt;#53921&lt;/a&gt;&lt;/li&gt;240&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Lossfull/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Lossfull&quot;&gt;@Lossfull&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5193245192&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52948&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52948/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52948&quot;&gt;#52948&lt;/a&gt;&lt;/li&gt;241&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lxy-alexander/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lxy-alexander&quot;&gt;@lxy-alexander&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5157922888&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52430&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52430/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52430&quot;&gt;#52430&lt;/a&gt;&lt;/li&gt;242&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lxyxinyi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lxyxinyi&quot;&gt;@lxyxinyi&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5046146281&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50809&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50809/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50809&quot;&gt;#50809&lt;/a&gt;&lt;/li&gt;243&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/machero/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/machero&quot;&gt;@machero&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231323743&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53531&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53531/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53531&quot;&gt;#53531&lt;/a&gt;&lt;/li&gt;244&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/matthewkotila/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/matthewkotila&quot;&gt;@matthewkotila&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4908287242&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48915&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48915/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48915&quot;&gt;#48915&lt;/a&gt;&lt;/li&gt;245&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mhoqueanik/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mhoqueanik&quot;&gt;@mhoqueanik&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4961919843&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49636&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49636/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49636&quot;&gt;#49636&lt;/a&gt;&lt;/li&gt;246&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mhuzaifa3/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mhuzaifa3&quot;&gt;@mhuzaifa3&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5238741378&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53625&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53625/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53625&quot;&gt;#53625&lt;/a&gt;&lt;/li&gt;247&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/minjang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/minjang&quot;&gt;@minjang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231289454&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53530&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53530/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53530&quot;&gt;#53530&lt;/a&gt;&lt;/li&gt;248&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/MKQuantum/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/MKQuantum&quot;&gt;@MKQuantum&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124104450&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51866&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51866/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51866&quot;&gt;#51866&lt;/a&gt;&lt;/li&gt;249&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/new-TonyWang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/new-TonyWang&quot;&gt;@new-TonyWang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5244170627&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53704&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53704/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53704&quot;&gt;#53704&lt;/a&gt;&lt;/li&gt;250&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/nicholaskh-ai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/nicholaskh-ai&quot;&gt;@nicholaskh-ai&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5252641623&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53819&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53819/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53819&quot;&gt;#53819&lt;/a&gt;&lt;/li&gt;251&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/oliverholworthy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/oliverholworthy&quot;&gt;@oliverholworthy&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142898516&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52185&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52185/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52185&quot;&gt;#52185&lt;/a&gt;&lt;/li&gt;252&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/prakharPant/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/prakharPant&quot;&gt;@prakharPant&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5180954309&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52743&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52743/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52743&quot;&gt;#52743&lt;/a&gt;&lt;/li&gt;253&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Prudhvivuda/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Prudhvivuda&quot;&gt;@Prudhvivuda&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5198011038&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53016&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53016/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53016&quot;&gt;#53016&lt;/a&gt;&lt;/li&gt;254&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rajathpi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rajathpi&quot;&gt;@rajathpi&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5172198842&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52622&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52622/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52622&quot;&gt;#52622&lt;/a&gt;&lt;/li&gt;255&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ray24777/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ray24777&quot;&gt;@ray24777&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5204596603&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53120&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53120/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53120&quot;&gt;#53120&lt;/a&gt;&lt;/li&gt;256&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/roachsinai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/roachsinai&quot;&gt;@roachsinai&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5231273628&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53528&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53528/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53528&quot;&gt;#53528&lt;/a&gt;&lt;/li&gt;257&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RookieCoder-Camera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RookieCoder-Camera&quot;&gt;@RookieCoder-Camera&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4935159122&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49274&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49274/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49274&quot;&gt;#49274&lt;/a&gt;&lt;/li&gt;258&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sandeep-maddipatla/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sandeep-maddipatla&quot;&gt;@sandeep-maddipatla&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117694715&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51777&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51777/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51777&quot;&gt;#51777&lt;/a&gt;&lt;/li&gt;259&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/seonjinn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/seonjinn&quot;&gt;@seonjinn&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5144692620&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52204&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52204/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52204&quot;&gt;#52204&lt;/a&gt;&lt;/li&gt;260&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ShengleiFu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ShengleiFu&quot;&gt;@ShengleiFu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5240472066&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53650&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53650/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53650&quot;&gt;#53650&lt;/a&gt;&lt;/li&gt;261&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shepark/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shepark&quot;&gt;@shepark&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084211010&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51302&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51302/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51302&quot;&gt;#51302&lt;/a&gt;&lt;/li&gt;262&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shijuzhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shijuzhao&quot;&gt;@shijuzhao&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4664004847&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45683&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45683/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45683&quot;&gt;#45683&lt;/a&gt;&lt;/li&gt;263&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shipiyouniao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shipiyouniao&quot;&gt;@shipiyouniao&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5130003345&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51979&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51979/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51979&quot;&gt;#51979&lt;/a&gt;&lt;/li&gt;264&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ShuaiShao93/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ShuaiShao93&quot;&gt;@ShuaiShao93&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5199007710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53034&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53034/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53034&quot;&gt;#53034&lt;/a&gt;&lt;/li&gt;265&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/simon-veitner-redhat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/simon-veitner-redhat&quot;&gt;@simon-veitner-redhat&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5195404163&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52980&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52980/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52980&quot;&gt;#52980&lt;/a&gt;&lt;/li&gt;266&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sseanliu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sseanliu&quot;&gt;@sseanliu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5134759141&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52041&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52041/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52041&quot;&gt;#52041&lt;/a&gt;&lt;/li&gt;267&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/studioego/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/studioego&quot;&gt;@studioego&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5155455899&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52389&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52389/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52389&quot;&gt;#52389&lt;/a&gt;&lt;/li&gt;268&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tanchao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tanchao&quot;&gt;@tanchao&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5003831860&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50191&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50191/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50191&quot;&gt;#50191&lt;/a&gt;&lt;/li&gt;269&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thanhpt1110/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thanhpt1110&quot;&gt;@thanhpt1110&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5179888512&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52720&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52720/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52720&quot;&gt;#52720&lt;/a&gt;&lt;/li&gt;270&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/theamalsebastian/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/theamalsebastian&quot;&gt;@theamalsebastian&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5169450657&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52588&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52588/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52588&quot;&gt;#52588&lt;/a&gt;&lt;/li&gt;271&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thunguo/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thunguo&quot;&gt;@thunguo&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5202873727&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53101&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53101/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53101&quot;&gt;#53101&lt;/a&gt;&lt;/li&gt;272&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tolleybot/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tolleybot&quot;&gt;@tolleybot&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5083618168&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51292&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51292/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51292&quot;&gt;#51292&lt;/a&gt;&lt;/li&gt;273&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tommy-asai-sonarsource/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tommy-asai-sonarsource&quot;&gt;@tommy-asai-sonarsource&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5090252678&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51395&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51395/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51395&quot;&gt;#51395&lt;/a&gt;&lt;/li&gt;274&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tthakkal/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tthakkal&quot;&gt;@tthakkal&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5025615988&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50501&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50501/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50501&quot;&gt;#50501&lt;/a&gt;&lt;/li&gt;275&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ukannika/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ukannika&quot;&gt;@ukannika&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5187682710&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52849&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52849/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52849&quot;&gt;#52849&lt;/a&gt;&lt;/li&gt;276&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/VBS2004/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/VBS2004&quot;&gt;@VBS2004&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4909019190&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48922&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48922/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48922&quot;&gt;#48922&lt;/a&gt;&lt;/li&gt;277&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wyettzeng/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wyettzeng&quot;&gt;@wyettzeng&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145232591&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52209&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52209/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52209&quot;&gt;#52209&lt;/a&gt;&lt;/li&gt;278&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Xuan-1998/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Xuan-1998&quot;&gt;@Xuan-1998&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5197472419&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53008&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53008/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53008&quot;&gt;#53008&lt;/a&gt;&lt;/li&gt;279&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/y0hnn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/y0hnn&quot;&gt;@y0hnn&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5132036711&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52002&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52002/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52002&quot;&gt;#52002&lt;/a&gt;&lt;/li&gt;280&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Yiqin-17/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Yiqin-17&quot;&gt;@Yiqin-17&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5191315493&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52925&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52925/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52925&quot;&gt;#52925&lt;/a&gt;&lt;/li&gt;281&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/YukioZzz/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/YukioZzz&quot;&gt;@YukioZzz&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5112328984&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51705&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51705/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51705&quot;&gt;#51705&lt;/a&gt;&lt;/li&gt;282&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ZHIHANCHEN03/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ZHIHANCHEN03&quot;&gt;@ZHIHANCHEN03&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5225466209&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/53432&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/53432/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/53432&quot;&gt;#53432&lt;/a&gt;&lt;/li&gt;283&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Zhou248/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Zhou248&quot;&gt;@Zhou248&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5167704960&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52560&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52560/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52560&quot;&gt;#52560&lt;/a&gt;&lt;/li&gt;284&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zllion/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zllion&quot;&gt;@zllion&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4741800584&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46701&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46701/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46701&quot;&gt;#46701&lt;/a&gt;&lt;/li&gt;285&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zwischenraum/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zwischenraum&quot;&gt;@zwischenraum&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5009940371&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50272&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50272/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50272&quot;&gt;#50272&lt;/a&gt;&lt;/li&gt;286&lt;/ul&gt;287&lt;h2&gt;Contributors&lt;/h2&gt;288&lt;p&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AndreasKaratzas/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AndreasKaratzas&quot;&gt;@AndreasKaratzas&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/khluu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/khluu&quot;&gt;@khluu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mgoin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mgoin&quot;&gt;@mgoin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/taneem-ibrahim/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/taneem-ibrahim&quot;&gt;@taneem-ibrahim&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yewentao256/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yewentao256&quot;&gt;@yewentao256&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/njhill/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/njhill&quot;&gt;@njhill&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/BugenZhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/BugenZhao&quot;&gt;@BugenZhao&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hmellor/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hmellor&quot;&gt;@hmellor&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/GirasoleY/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/GirasoleY&quot;&gt;@GirasoleY&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/DarkLight1337/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/DarkLight1337&quot;&gt;@DarkLight1337&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mayuyuace/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mayuyuace&quot;&gt;@mayuyuace&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/WoosukKwon/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/WoosukKwon&quot;&gt;@WoosukKwon&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jeejeelee/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jeejeelee&quot;&gt;@jeejeelee&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/noooop/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/noooop&quot;&gt;@noooop&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jperezdealgaba/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jperezdealgaba&quot;&gt;@jperezdealgaba&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LucasWilkinson/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LucasWilkinson&quot;&gt;@LucasWilkinson&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wzhao18/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wzhao18&quot;&gt;@wzhao18&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/stefankoncarevic/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/stefankoncarevic&quot;&gt;@stefankoncarevic&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gau-nernst/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gau-nernst&quot;&gt;@gau-nernst&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/okorzh-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/okorzh-amd&quot;&gt;@okorzh-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ZJY0516/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ZJY0516&quot;&gt;@ZJY0516&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/HollowMan6/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/HollowMan6&quot;&gt;@HollowMan6&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/connorcarpenter15/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/connorcarpenter15&quot;&gt;@connorcarpenter15&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gcanlin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gcanlin&quot;&gt;@gcanlin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/aoshen02/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/aoshen02&quot;&gt;@aoshen02&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/linitra24/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/linitra24&quot;&gt;@linitra24&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mhuzaifa3/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mhuzaifa3&quot;&gt;@mhuzaifa3&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Isotr0py/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Isotr0py&quot;&gt;@Isotr0py&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Etelis/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Etelis&quot;&gt;@Etelis&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/MatthewBonanni/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/MatthewBonanni&quot;&gt;@MatthewBonanni&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/TheEpicDolphin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/TheEpicDolphin&quot;&gt;@TheEpicDolphin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zupengwang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zupengwang&quot;&gt;@zupengwang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/atalman/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/atalman&quot;&gt;@atalman&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/NickLucche/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/NickLucche&quot;&gt;@NickLucche&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rasmith/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rasmith&quot;&gt;@rasmith&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/elvircrn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/elvircrn&quot;&gt;@elvircrn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JasonKeyiL/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JasonKeyiL&quot;&gt;@JasonKeyiL&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Hotragn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Hotragn&quot;&gt;@Hotragn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/waizuichougou/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/waizuichougou&quot;&gt;@waizuichougou&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pmanczak/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pmanczak&quot;&gt;@pmanczak&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/divakar-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/divakar-amd&quot;&gt;@divakar-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/he-yufeng/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/he-yufeng&quot;&gt;@he-yufeng&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xuebwang-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xuebwang-amd&quot;&gt;@xuebwang-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/qgallouedec/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/qgallouedec&quot;&gt;@qgallouedec&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/esmeetu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/esmeetu&quot;&gt;@esmeetu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ZeldaHuang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ZeldaHuang&quot;&gt;@ZeldaHuang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Prudhvivuda/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Prudhvivuda&quot;&gt;@Prudhvivuda&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/chaunceyjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/chaunceyjiang&quot;&gt;@chaunceyjiang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zxd1997066/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zxd1997066&quot;&gt;@zxd1997066&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/louie-tsai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/louie-tsai&quot;&gt;@louie-tsai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/andyxning/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/andyxning&quot;&gt;@andyxning&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pisceskkk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pisceskkk&quot;&gt;@pisceskkk&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tjtanaa/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tjtanaa&quot;&gt;@tjtanaa&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/qli88/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/qli88&quot;&gt;@qli88&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tlrmchlsmth/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tlrmchlsmth&quot;&gt;@tlrmchlsmth&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LopezCastroRoberto/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LopezCastroRoberto&quot;&gt;@LopezCastroRoberto&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/micah-wil/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/micah-wil&quot;&gt;@micah-wil&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hongxiayang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hongxiayang&quot;&gt;@hongxiayang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/BabyDrangoner/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/BabyDrangoner&quot;&gt;@BabyDrangoner&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yimdev/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yimdev&quot;&gt;@yimdev&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mganczarenko/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mganczarenko&quot;&gt;@mganczarenko&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/SageMoore/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/SageMoore&quot;&gt;@SageMoore&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/biswapanda/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/biswapanda&quot;&gt;@biswapanda&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sfeng33/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sfeng33&quot;&gt;@sfeng33&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/russellb/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/russellb&quot;&gt;@russellb&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/itayalroy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/itayalroy&quot;&gt;@itayalroy&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bigPYJ1151/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bigPYJ1151&quot;&gt;@bigPYJ1151&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/drakosha/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/drakosha&quot;&gt;@drakosha&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/qwerqwerqwe8688-jpg/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/qwerqwerqwe8688-jpg&quot;&gt;@qwerqwerqwe8688-jpg&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/akii96/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/akii96&quot;&gt;@akii96&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shen-shanshan/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shen-shanshan&quot;&gt;@shen-shanshan&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/meiyeh123/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/meiyeh123&quot;&gt;@meiyeh123&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/reidliu41/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/reidliu41&quot;&gt;@reidliu41&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yma11/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yma11&quot;&gt;@yma11&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/frgossen/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/frgossen&quot;&gt;@frgossen&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lukealonso/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lukealonso&quot;&gt;@lukealonso&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tanchao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tanchao&quot;&gt;@tanchao&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Fangzhou-Ai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Fangzhou-Ai&quot;&gt;@Fangzhou-Ai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vllm-agent/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vllm-agent&quot;&gt;@vllm-agent&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/djramic/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/djramic&quot;&gt;@djramic&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/andrewbcohere/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/andrewbcohere&quot;&gt;@andrewbcohere&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zyongye/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zyongye&quot;&gt;@zyongye&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gty111/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gty111&quot;&gt;@gty111&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ShuoleiWang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ShuoleiWang&quot;&gt;@ShuoleiWang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ZHIHANCHEN03/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ZHIHANCHEN03&quot;&gt;@ZHIHANCHEN03&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/KurodaKanbei/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/KurodaKanbei&quot;&gt;@KurodaKanbei&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/alexeldeib/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/alexeldeib&quot;&gt;@alexeldeib&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/khushali9/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/khushali9&quot;&gt;@khushali9&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xiaohuguo2023/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xiaohuguo2023&quot;&gt;@xiaohuguo2023&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hungnnvidia/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hungnnvidia&quot;&gt;@hungnnvidia&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thisjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thisjiang&quot;&gt;@thisjiang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/arpera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/arpera&quot;&gt;@arpera&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ShengleiFu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ShengleiFu&quot;&gt;@ShengleiFu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zhenwei-intel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zhenwei-intel&quot;&gt;@zhenwei-intel&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jungjiyu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jungjiyu&quot;&gt;@jungjiyu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yzong-rh/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yzong-rh&quot;&gt;@yzong-rh&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/charlifu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/charlifu&quot;&gt;@charlifu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hclsys/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hclsys&quot;&gt;@hclsys&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Ronald1995/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Ronald1995&quot;&gt;@Ronald1995&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mkhazraee/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mkhazraee&quot;&gt;@mkhazraee&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/YukioZzz/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/YukioZzz&quot;&gt;@YukioZzz&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/theamalsebastian/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/theamalsebastian&quot;&gt;@theamalsebastian&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/030611/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/030611&quot;&gt;@030611&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cr-zhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cr-zhao&quot;&gt;@cr-zhao&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jbyczkow/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jbyczkow&quot;&gt;@jbyczkow&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tommy-asai-sonarsource/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tommy-asai-sonarsource&quot;&gt;@tommy-asai-sonarsource&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jikunshang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jikunshang&quot;&gt;@jikunshang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bobboli/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bobboli&quot;&gt;@bobboli&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LH-and-FPGA/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LH-and-FPGA&quot;&gt;@LH-and-FPGA&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/SayHelloToWorld/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/SayHelloToWorld&quot;&gt;@SayHelloToWorld&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/kkt-cohere/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/kkt-cohere&quot;&gt;@kkt-cohere&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/floatlibai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/floatlibai&quot;&gt;@floatlibai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rajathpi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rajathpi&quot;&gt;@rajathpi&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lxy-alexander/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lxy-alexander&quot;&gt;@lxy-alexander&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gangula-karthik/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gangula-karthik&quot;&gt;@gangula-karthik&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ActiveSky/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ActiveSky&quot;&gt;@ActiveSky&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fxmarty-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fxmarty-amd&quot;&gt;@fxmarty-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sstamenk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sstamenk&quot;&gt;@sstamenk&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mpashkovskii/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mpashkovskii&quot;&gt;@mpashkovskii&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LiuYinfeng01/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LiuYinfeng01&quot;&gt;@LiuYinfeng01&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/chaojun-zhang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/chaojun-zhang&quot;&gt;@chaojun-zhang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Oxygen56/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Oxygen56&quot;&gt;@Oxygen56&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/daviswer/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/daviswer&quot;&gt;@daviswer&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vineethsaivs/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vineethsaivs&quot;&gt;@vineethsaivs&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Agoni-02/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Agoni-02&quot;&gt;@Agoni-02&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Andy365-365/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Andy365-365&quot;&gt;@Andy365-365&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wangxiyuan/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wangxiyuan&quot;&gt;@wangxiyuan&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/haoyangqian/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/haoyangqian&quot;&gt;@haoyangqian&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Naveassaf/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Naveassaf&quot;&gt;@Naveassaf&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/eilamc14/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/eilamc14&quot;&gt;@eilamc14&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sseanliu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sseanliu&quot;&gt;@sseanliu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/y0hnn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/y0hnn&quot;&gt;@y0hnn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dineshchitlangia/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dineshchitlangia&quot;&gt;@dineshchitlangia&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yiliu30/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yiliu30&quot;&gt;@yiliu30&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Lossfull/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Lossfull&quot;&gt;@Lossfull&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lxyxinyi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lxyxinyi&quot;&gt;@lxyxinyi&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vanshbhatia-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vanshbhatia-amd&quot;&gt;@vanshbhatia-amd&lt;/a&gt;, @jcotant-inferact, @chengy-sysu, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/92hyungjun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/92hyungjun&quot;&gt;@92hyungjun&lt;/a&gt;, @xyang16, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AmitMY/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AmitMY&quot;&gt;@AmitMY&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zllion/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zllion&quot;&gt;@zllion&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AnkitNakhawa/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AnkitNakhawa&quot;&gt;@AnkitNakhawa&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmvevents/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmvevents&quot;&gt;@dmvevents&lt;/a&gt;, @lengrongfu, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JC-ut0/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JC-ut0&quot;&gt;@JC-ut0&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Eoin-Houstoun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Eoin-Houstoun&quot;&gt;@Eoin-Houstoun&lt;/a&gt;, @sagearc, @wjabbour, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sandeep-maddipatla/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sandeep-maddipatla&quot;&gt;@sandeep-maddipatla&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/seonjinn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/seonjinn&quot;&gt;@seonjinn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Yiqin-17/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Yiqin-17&quot;&gt;@Yiqin-17&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/kyleliang-nv/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/kyleliang-nv&quot;&gt;@kyleliang-nv&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thanhpt1110/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thanhpt1110&quot;&gt;@thanhpt1110&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/MKQuantum/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/MKQuantum&quot;&gt;@MKQuantum&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/matthewkotila/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/matthewkotila&quot;&gt;@matthewkotila&lt;/a&gt;, @vhagor, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tthakkal/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tthakkal&quot;&gt;@tthakkal&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ShuaiShao93/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ShuaiShao93&quot;&gt;@ShuaiShao93&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hagaikwa-redhat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hagaikwa-redhat&quot;&gt;@hagaikwa-redhat&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shijuzhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shijuzhao&quot;&gt;@shijuzhao&lt;/a&gt;, @elwhyjay, @Gregory-Pereira, @ErenAta16, @morrison-turnansky, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/DCoEngine/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/DCoEngine&quot;&gt;@DCoEngine&lt;/a&gt;, @SoluMilken, @xianbaoqian, @kliuae, @nascheme, @waynehacking8, @stecasta, @vMaroon, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zwischenraum/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zwischenraum&quot;&gt;@zwischenraum&lt;/a&gt;, @Rohan138, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Zhou248/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Zhou248&quot;&gt;@Zhou248&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/brianosaurus/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/brianosaurus&quot;&gt;@brianosaurus&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anmolgupt/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anmolgupt&quot;&gt;@anmolgupt&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wyettzeng/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wyettzeng&quot;&gt;@wyettzeng&lt;/a&gt;, @hao-aaron, @frank-suwen, @danisereb, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fuzzifikation/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fuzzifikation&quot;&gt;@fuzzifikation&lt;/a&gt;, @KernelClint, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thunguo/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thunguo&quot;&gt;@thunguo&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/studioego/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/studioego&quot;&gt;@studioego&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shipiyouniao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shipiyouniao&quot;&gt;@shipiyouniao&lt;/a&gt;, @cjackal, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LioEinaudi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LioEinaudi&quot;&gt;@LioEinaudi&lt;/a&gt;, @guan404ming, @libinta, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mhoqueanik/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mhoqueanik&quot;&gt;@mhoqueanik&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/minjang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/minjang&quot;&gt;@minjang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Edge-Explorer/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Edge-Explorer&quot;&gt;@Edge-Explorer&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shepark/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shepark&quot;&gt;@shepark&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JiataiWang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JiataiWang&quot;&gt;@JiataiWang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jiahaoliang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jiahaoliang&quot;&gt;@jiahaoliang&lt;/a&gt;, @therealnaveenkamal, @ColinZ22, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tolleybot/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tolleybot&quot;&gt;@tolleybot&lt;/a&gt;, @almogtavor, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/simon-veitner-redhat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/simon-veitner-redhat&quot;&gt;@simon-veitner-redhat&lt;/a&gt;, @hyeongyun0916, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/nicholaskh-ai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/nicholaskh-ai&quot;&gt;@nicholaskh-ai&lt;/a&gt;, @mawong-amd, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/adisivaprasad/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/adisivaprasad&quot;&gt;@adisivaprasad&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ray24777/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ray24777&quot;&gt;@ray24777&lt;/a&gt;, @robertgshaw2-redhat, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/new-TonyWang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/new-TonyWang&quot;&gt;@new-TonyWang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/oliverholworthy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/oliverholworthy&quot;&gt;@oliverholworthy&lt;/a&gt;, @fanxingran, @lucianommartins, @hallerite, @avininjamay8, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/VBS2004/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/VBS2004&quot;&gt;@VBS2004&lt;/a&gt;, @omerpaz95, @Sunt-ing, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Hert4/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Hert4&quot;&gt;@Hert4&lt;/a&gt;, @wangshangsam, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cogniera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cogniera&quot;&gt;@cogniera&lt;/a&gt;, @jeffreywang88, @Xarbirus, @fynnsu, @jiahanc, @zixi-qi, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/prakharPant/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/prakharPant&quot;&gt;@prakharPant&lt;/a&gt;, @haic0, @maobaolong, @xinyu-intel, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CHIPMUNK-T0T/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CHIPMUNK-T0T&quot;&gt;@CHIPMUNK-T0T&lt;/a&gt;, @hangy-amd, @afriedri, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/canlahlah/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/canlahlah&quot;&gt;@canlahlah&lt;/a&gt;, @Alnusjaponica, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ukannika/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ukannika&quot;&gt;@ukannika&lt;/a&gt;, @pranavthakur0-0, @zzaebok, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Xuan-1998/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Xuan-1998&quot;&gt;@Xuan-1998&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JulianZJN/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JulianZJN&quot;&gt;@JulianZJN&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/roachsinai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/roachsinai&quot;&gt;@roachsinai&lt;/a&gt;, @ivanium, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CalvinXKY/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CalvinXKY&quot;&gt;@CalvinXKY&lt;/a&gt;, @yu-xin-c, @rchalamala, @QwertyJack, @Kaif10, @yudigege86, @yiz-liu, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dkrisman/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dkrisman&quot;&gt;@dkrisman&lt;/a&gt;, @simondanielsson, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/machero/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/machero&quot;&gt;@machero&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/CherryLemon/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/CherryLemon&quot;&gt;@CherryLemon&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ima-helikoptaaa/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ima-helikoptaaa&quot;&gt;@ima-helikoptaaa&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RookieCoder-Camera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RookieCoder-Camera&quot;&gt;@RookieCoder-Camera&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LironKesem/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LironKesem&quot;&gt;@LironKesem&lt;/a&gt;, @peakcrosser7, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/eligotts/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/eligotts&quot;&gt;@eligotts&lt;/a&gt;, @foraxe, @zufangzhu, @tianmu-li, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jl9876/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jl9876&quot;&gt;@jl9876&lt;/a&gt;, @lucifer1004, @liranschour, @maithilijoshi20, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/li-ukumar/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/li-ukumar&quot;&gt;@li-ukumar&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/andyluo7/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/andyluo7&quot;&gt;@andyluo7&lt;/a&gt;, @Zhenzhong1, @faaany, @joerowell, @juhi10071998, @ganeshr10, @Priyjain-amd, @SubSir, @luyixiao95, @djw8605, @codex&lt;/p&gt;</content>289    <author>290      <name>khluu</name>291    </author>292    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/51931015?s=60&amp;v=4"/>293  </entry>294  <entry>295    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc6</id>296    <updated>2026-09-08T08:34:20Z</updated>297    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc6"/>298    <title>v0.29.0rc6</title>299    <content type="html">&lt;p&gt;[Bugfix][Core] Apply dense prefix cache default to hybrid models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;1693327998&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/55&quot; data-hovercard-type=&quot;issue&quot; data-hovercard-url=&quot;/vllm-project/vllm/issues/55/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/issues/55&quot;&gt;#55&lt;/a&gt;…&lt;/p&gt;</content>300    <author>301      <name>ZJY0516</name>302    </author>303    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/86695626?s=60&amp;v=4"/>304  </entry>305  <entry>306    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc5</id>307    <updated>2026-09-08T08:20:33Z</updated>308    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc5"/>309    <title>v0.29.0rc5</title>310    <content type="html">&lt;p&gt;[Core] Default prefix_cache_retention_interval to dense for Mamba + E…&lt;/p&gt;</content>311    <author>312      <name>ZJY0516</name>313    </author>314    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/86695626?s=60&amp;v=4"/>315  </entry>316  <entry>317    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc4</id>318    <updated>2026-09-04T22:58:29Z</updated>319    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc4"/>320    <title>v0.29.0rc4: [Bugfix] Avoid sync in TRT-LLM ragged prefill</title>321    <content type="html">&lt;p&gt;Generated-by: Codex &lt;a href=&quot;mailto:codex@openai.com&quot;&gt;codex@openai.com&lt;/a&gt;&lt;/p&gt;322&lt;p&gt;Signed-off-by: Codex &lt;a href=&quot;mailto:codex@openai.com&quot;&gt;codex@openai.com&lt;/a&gt;&lt;/p&gt;</content>323    <author>324      <name>codex</name>325    </author>326    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/267193182?s=60&amp;v=4"/>327  </entry>328  <entry>329    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc3</id>330    <updated>2026-09-04T00:52:54Z</updated>331    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc3"/>332    <title>v0.29.0rc3</title>333    <content type="html">&lt;p&gt;[CI] Remove deleted nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF1…&lt;/p&gt;</content>334    <author>335      <name>khluu</name>336    </author>337    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/51931015?s=60&amp;v=4"/>338  </entry>339  <entry>340    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc2</id>341    <updated>2026-09-03T09:41:12Z</updated>342    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc2"/>343    <title>v0.29.0rc2</title>344    <content type="html">&lt;p&gt;[Bugfix][Multimodal] Handle prefix-covered items in SHM worker cache …&lt;/p&gt;</content>345    <author>346      <name>waizuichougou</name>347    </author>348    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/105572299?s=60&amp;v=4"/>349  </entry>350  <entry>351    <id>tag:github.com,2008:Repository/599547518/v0.29.0rc1</id>352    <updated>2026-09-02T02:11:30Z</updated>353    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.29.0rc1"/>354    <title>v0.29.0rc1: [Bugfix] Handle padded routes in CUTLASS MoE permutations (#54747)</title>355    <content type="html">&lt;p&gt;Signed-off-by: Kevin Luu &lt;a href=&quot;mailto:51931015+khluu@users.noreply.github.com&quot;&gt;51931015+khluu@users.noreply.github.com&lt;/a&gt;&lt;br&gt;356Signed-off-by: Yongye Zhu &lt;a href=&quot;mailto:zyy1102000@gmail.com&quot;&gt;zyy1102000@gmail.com&lt;/a&gt;&lt;br&gt;357Co-authored-by: Codex &lt;a href=&quot;mailto:codex@openai.com&quot;&gt;codex@openai.com&lt;/a&gt;&lt;br&gt;358Co-authored-by: Yongye Zhu &lt;a href=&quot;mailto:zyy1102000@gmail.com&quot;&gt;zyy1102000@gmail.com&lt;/a&gt;&lt;/p&gt;</content>359    <author>360      <name>khluu</name>361    </author>362    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/51931015?s=60&amp;v=4"/>363  </entry>364  <entry>365    <id>tag:github.com,2008:Repository/599547518/v0.28.1rc0</id>366    <updated>2026-08-27T06:58:23Z</updated>367    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.28.1rc0"/>368    <title>v0.28.1rc0</title>369    <content type="html">&lt;p&gt;[Tools][Recipes] Improve sweep recommendations and short-alias parsin…&lt;/p&gt;</content>370    <author>371      <name>louie-tsai</name>372    </author>373    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/21761437?s=60&amp;v=4"/>374  </entry>375  <entry>376    <id>tag:github.com,2008:Repository/599547518/v0.28.0</id>377    <updated>2026-08-26T10:17:05Z</updated>378    <link rel="alternate" type="text/html" href="https://github.com/vllm-project/vllm/releases/tag/v0.28.0"/>379    <title>v0.28.0</title>380    <content type="html">&lt;h1&gt;v0.28.0&lt;/h1&gt;381&lt;h2&gt;Highlights&lt;/h2&gt;382&lt;p&gt;This release features 584 commits from 270 contributors (76 new)!&lt;/p&gt;383&lt;ul&gt;384&lt;li&gt;&lt;strong&gt;Kimi-K3 performance push&lt;/strong&gt;: a major optimization effort for Kimi-K3 across the stack — Decode Context Parallel (DCP) support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5024323514&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50484&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50484/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50484&quot;&gt;#50484&lt;/a&gt;), fused FlashKDA decode and prefill kernels (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5035155331&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50654&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50654/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50654&quot;&gt;#50654&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084819738&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51311&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51311/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51311&quot;&gt;#51311&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5160378588&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52458&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52458/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52458&quot;&gt;#52458&lt;/a&gt;), SiTU activation support for MegaMoE (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5026036754&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50510&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50510/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50510&quot;&gt;#50510&lt;/a&gt;), GEMM-RS for sequence parallelism (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5136777880&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52079&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52079/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52079&quot;&gt;#52079&lt;/a&gt;), combined all-gathers with 1.5~3x kernel-level speedup (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5064176312&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51070&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51070/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51070&quot;&gt;#51070&lt;/a&gt;), an adaptive speculative token budget delivering ~60% better DSpark TTFT (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5113682718&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51725&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51725/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51725&quot;&gt;#51725&lt;/a&gt;), and optional shared-expert sharding saving ~17 GiB of memory per GPU (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5053646772&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50912&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50912/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50912&quot;&gt;#50912&lt;/a&gt;). Kimi-K3 also now runs on ROCm with the V2 model runner (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5108926237&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51653&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51653/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51653&quot;&gt;#51653&lt;/a&gt;).&lt;/li&gt;385&lt;li&gt;&lt;strong&gt;DeepSeek V4&lt;/strong&gt;: sparse MLA now works end-to-end for plain decode, MTP, and DSpark speculative decoding (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5101531400&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51538&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51538/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51538&quot;&gt;#51538&lt;/a&gt;), joined by AMD Quark NVFP4 support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4835520784&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47972&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47972/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47972&quot;&gt;#47972&lt;/a&gt;), reasoning-effort prompts and mappings (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5030980529&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50580&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50580/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50580&quot;&gt;#50580&lt;/a&gt;), sparse top-k metadata kernel optimizations (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5137310673&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52084&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52084/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52084&quot;&gt;#52084&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5129184600&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51967&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51967/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51967&quot;&gt;#51967&lt;/a&gt;), narrowed eager CUDA graph regions (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5092896769&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51430&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51430/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51430&quot;&gt;#51430&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5156348462&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52401&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52401/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52401&quot;&gt;#52401&lt;/a&gt;), and ROCm enablement on gfx11 and gfx950 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4766317415&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47017&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47017/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47017&quot;&gt;#47017&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145545577&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52212&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52212/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52212&quot;&gt;#52212&lt;/a&gt;).&lt;/li&gt;386&lt;li&gt;&lt;strong&gt;Speculative decoding advances&lt;/strong&gt;: DFlash2 with local convolution and a candidate selector (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185508751&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52816&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52816/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52816&quot;&gt;#52816&lt;/a&gt;), DSpark confidence-scheduled verification (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4824929811&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47808&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47808/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47808&quot;&gt;#47808&lt;/a&gt;), and async scheduling auto-enabled for draft models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4862124280&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48341&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48341/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48341&quot;&gt;#48341&lt;/a&gt;).&lt;/li&gt;387&lt;li&gt;&lt;strong&gt;Model Runner V2 maturation&lt;/strong&gt;: E/P/D disaggregation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4155353913&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/38390&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/38390/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/38390&quot;&gt;#38390&lt;/a&gt;), weight offloading (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5091635999&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51413&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51413/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51413&quot;&gt;#51413&lt;/a&gt;), multi-layer MTP KV cache support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4993780584&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50062&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50062/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50062&quot;&gt;#50062&lt;/a&gt;), encoder CUDA graphs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4979046200&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49852&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49852/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49852&quot;&gt;#49852&lt;/a&gt;), decoder token-wise pooling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5055136007&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50931&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50931/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50931&quot;&gt;#50931&lt;/a&gt;) plus Transformers pooling models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5157751117&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52425&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52425/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52425&quot;&gt;#52425&lt;/a&gt;), attention-free models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5154157605&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52374&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52374/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52374&quot;&gt;#52374&lt;/a&gt;), and &lt;code&gt;thinking_token_budget&lt;/code&gt; support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4744261944&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46727&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46727/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46727&quot;&gt;#46727&lt;/a&gt;).&lt;/li&gt;388&lt;li&gt;&lt;strong&gt;Tiered KV cache offloading&lt;/strong&gt;: disk offloading support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4963653413&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49644&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49644/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49644&quot;&gt;#49644&lt;/a&gt;), out-of-tree secondary tier managers via &lt;code&gt;module_path&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5058868734&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51007&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51007/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51007&quot;&gt;#51007&lt;/a&gt;), partial secondary-tier load results (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5013391634&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50321&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50321/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50321&quot;&gt;#50321&lt;/a&gt;), tiering metrics (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4898117346&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48798&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48798/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48798&quot;&gt;#48798&lt;/a&gt;), and a canonical CPU layout for parallelism-agnostic offload (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4867574663&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48414&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48414/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48414&quot;&gt;#48414&lt;/a&gt;).&lt;/li&gt;389&lt;li&gt;&lt;strong&gt;Rust frontend &amp;amp; gRPC&lt;/strong&gt;: a standalone renderer (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5011355404&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50289&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50289/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50289&quot;&gt;#50289&lt;/a&gt;), multimodal image inference over gRPC (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5015640495&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50368&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50368/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50368&quot;&gt;#50368&lt;/a&gt;), explicit data-parallel rank routing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5072919309&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51178&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51178/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51178&quot;&gt;#51178&lt;/a&gt;), and RL lifecycle control (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084945361&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51316&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51316/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51316&quot;&gt;#51316&lt;/a&gt;), with protobuf schemas now published to Buf (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5082805927&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51276&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51276/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51276&quot;&gt;#51276&lt;/a&gt;).&lt;/li&gt;390&lt;li&gt;&lt;strong&gt;New defaults&lt;/strong&gt;: &lt;code&gt;max_num_batched_tokens&lt;/code&gt; raised from 8192 to 16384 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5113757558&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51726&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51726/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51726&quot;&gt;#51726&lt;/a&gt;), prefix caching enabled by default for Mamba models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5057990179&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50991&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50991/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50991&quot;&gt;#50991&lt;/a&gt;), and the Blackwell CUDA graph capture default raised to 1024 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4944760977&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49390&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49390/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49390&quot;&gt;#49390&lt;/a&gt;).&lt;/li&gt;391&lt;li&gt;&lt;strong&gt;Breaking changes&lt;/strong&gt;: bitsandbytes support migrated to an out-of-tree plugin (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4512000173&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43529&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43529/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43529&quot;&gt;#43529&lt;/a&gt;); Transformers bumped to 5.15.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109844757&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51668&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51668/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51668&quot;&gt;#51668&lt;/a&gt;); the deprecated &lt;code&gt;calculate_kv_scales&lt;/code&gt; runtime KV scale calculation was removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4944675936&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49389&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49389/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49389&quot;&gt;#49389&lt;/a&gt;); &lt;code&gt;override_attention_dtype&lt;/code&gt; was removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4888820267&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48684&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48684/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48684&quot;&gt;#48684&lt;/a&gt;).&lt;/li&gt;392&lt;/ul&gt;393&lt;h2&gt;Release Artifacts&lt;/h2&gt;394&lt;h3&gt;Python Wheels&lt;/h3&gt;395&lt;table&gt;396&lt;thead&gt;397&lt;tr&gt;398&lt;th&gt;Platform&lt;/th&gt;399&lt;th&gt;Install&lt;/th&gt;400&lt;/tr&gt;401&lt;/thead&gt;402&lt;tbody&gt;403&lt;tr&gt;404&lt;td&gt;PyPI (CUDA 13.0)&lt;/td&gt;405&lt;td&gt;&lt;code&gt;pip install vllm&lt;/code&gt;&lt;/td&gt;406&lt;/tr&gt;407&lt;tr&gt;408&lt;td&gt;PyPI (CUDA 13.0, uv)&lt;/td&gt;409&lt;td&gt;&lt;code&gt;uv pip install vllm --torch-backend=auto&lt;/code&gt;&lt;/td&gt;410&lt;/tr&gt;411&lt;tr&gt;412&lt;td&gt;ROCm&lt;/td&gt;413&lt;td&gt;&lt;code&gt;pip install vllm --extra-index-url https://wheels.vllm.ai/rocm/0.28.0/rocm722&lt;/code&gt;&lt;/td&gt;414&lt;/tr&gt;415&lt;/tbody&gt;416&lt;/table&gt;417&lt;h3&gt;Docker Images&lt;/h3&gt;418&lt;table&gt;419&lt;thead&gt;420&lt;tr&gt;421&lt;th&gt;Platform&lt;/th&gt;422&lt;th&gt;Docker Image&lt;/th&gt;423&lt;/tr&gt;424&lt;/thead&gt;425&lt;tbody&gt;426&lt;tr&gt;427&lt;td&gt;CUDA 13.0 (Default)&lt;/td&gt;428&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.28.0&lt;/code&gt; (&lt;code&gt;v0.28.0-cu130&lt;/code&gt; also works)&lt;/td&gt;429&lt;/tr&gt;430&lt;tr&gt;431&lt;td&gt;CUDA 12.9&lt;/td&gt;432&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.28.0-cu129&lt;/code&gt;&lt;/td&gt;433&lt;/tr&gt;434&lt;tr&gt;435&lt;td&gt;CUDA 13.0 + Ubuntu 24.04&lt;/td&gt;436&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.28.0-ubuntu2404&lt;/code&gt;&lt;/td&gt;437&lt;/tr&gt;438&lt;tr&gt;439&lt;td&gt;CUDA 12.9 + Ubuntu 24.04&lt;/td&gt;440&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai:v0.28.0-cu129-ubuntu2404&lt;/code&gt;&lt;/td&gt;441&lt;/tr&gt;442&lt;tr&gt;443&lt;td&gt;ROCm&lt;/td&gt;444&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-rocm:v0.28.0&lt;/code&gt;&lt;/td&gt;445&lt;/tr&gt;446&lt;tr&gt;447&lt;td&gt;CPU&lt;/td&gt;448&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-cpu:v0.28.0&lt;/code&gt;&lt;/td&gt;449&lt;/tr&gt;450&lt;tr&gt;451&lt;td&gt;XPU&lt;/td&gt;452&lt;td&gt;&lt;code&gt;docker pull vllm/vllm-openai-xpu:v0.28.0&lt;/code&gt;&lt;/td&gt;453&lt;/tr&gt;454&lt;/tbody&gt;455&lt;/table&gt;456&lt;h3&gt;Other Artifacts&lt;/h3&gt;457&lt;p&gt;Pre-built release artifacts are available in the &lt;strong&gt;Assets&lt;/strong&gt; section at the bottom of this page, including:&lt;/p&gt;458&lt;ul&gt;459&lt;li&gt;Source distribution tarball&lt;/li&gt;460&lt;li&gt;CUDA 12.9 Python wheels for x86_64 and arm64&lt;/li&gt;461&lt;li&gt;CUDA 13.0 Python wheels for x86_64 and arm64&lt;/li&gt;462&lt;li&gt;CPU Python wheels for x86_64, arm64, and macOS&lt;/li&gt;463&lt;/ul&gt;464&lt;h2&gt;Model Support&lt;/h2&gt;465&lt;ul&gt;466&lt;li&gt;&lt;strong&gt;New models&lt;/strong&gt;: Muse Glimmer (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109068083&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51655&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51655/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51655&quot;&gt;#51655&lt;/a&gt;), Ling 3.0 Flash with BF16, MTP, and parser support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5062200231&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51045&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51045/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51045&quot;&gt;#51045&lt;/a&gt;) plus an FP8 variant (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5081088003&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51265&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51265/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51265&quot;&gt;#51265&lt;/a&gt;) and hybrid MXFP4 routed experts (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5138714046&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52114&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52114/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52114&quot;&gt;#52114&lt;/a&gt;), Dots3 NOTE native multimodal support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5080636421&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51255&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51255/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51255&quot;&gt;#51255&lt;/a&gt;), and Interns2mobius (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5070167356&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51149&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51149/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51149&quot;&gt;#51149&lt;/a&gt;).&lt;/li&gt;467&lt;li&gt;&lt;strong&gt;Qwen&lt;/strong&gt;: Qwen3.8 enabled on AMD ROCm (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4994163972&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50068&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50068/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50068&quot;&gt;#50068&lt;/a&gt;), fused CUDA post-conv MTP decode kernel for Qwen3.5 GDN (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5110251224&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51674&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51674/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51674&quot;&gt;#51674&lt;/a&gt;), GDN gates aligned with speculative tokens (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5119924903&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51812&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51812/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51812&quot;&gt;#51812&lt;/a&gt;), and Qwen3.5 fixes for text-only checkpoints (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5041533191&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50734&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50734/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50734&quot;&gt;#50734&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5015166584&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50355&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50355/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50355&quot;&gt;#50355&lt;/a&gt;).&lt;/li&gt;468&lt;li&gt;&lt;strong&gt;Transformers modeling backend&lt;/strong&gt;: MLA support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4854954561&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48250&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48250/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48250&quot;&gt;#48250&lt;/a&gt;), hardware-agnostic model definition (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4949852429&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49458&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49458/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49458&quot;&gt;#49458&lt;/a&gt;), fully generalized input embedding handling (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5079925625&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51247&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51247/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51247&quot;&gt;#51247&lt;/a&gt;), logit softcapping (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142232127&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52173&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52173/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52173&quot;&gt;#52173&lt;/a&gt;), and a hardened multimodal path (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5091535399&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51408&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51408/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51408&quot;&gt;#51408&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109234791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51657&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51657/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51657&quot;&gt;#51657&lt;/a&gt;).&lt;/li&gt;469&lt;li&gt;&lt;strong&gt;LoRA&lt;/strong&gt;: vision tower LoRA for Gemma4 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4447596402&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/42662&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/42662/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/42662&quot;&gt;#42662&lt;/a&gt;), tower/connector LoRA for Keye (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117831469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51780&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51780/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51780&quot;&gt;#51780&lt;/a&gt;) and Ultravox (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4852532476&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48215&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48215/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48215&quot;&gt;#48215&lt;/a&gt;).&lt;/li&gt;470&lt;li&gt;&lt;strong&gt;Vision encoders&lt;/strong&gt;: ViT full CUDA graph for Kimi-K2.5 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5054946690&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50929&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50929/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50929&quot;&gt;#50929&lt;/a&gt;) and Ernie-4.5-VL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4638377271&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45254&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45254/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45254&quot;&gt;#45254&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5094710843&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51461&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51461/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51461&quot;&gt;#51461&lt;/a&gt;), torch.compile for the Qwen3-VL encoder (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4280907120&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/40116&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/40116/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/40116&quot;&gt;#40116&lt;/a&gt;), and long-blocking H2D copies avoided in ViT (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5122128314&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51841&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51841/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51841&quot;&gt;#51841&lt;/a&gt;).&lt;/li&gt;471&lt;li&gt;&lt;strong&gt;MoE&lt;/strong&gt;: extended EPLB support for Mistral Large 3 and additional MoE backends (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4862575204&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48355&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48355/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48355&quot;&gt;#48355&lt;/a&gt;), CuTe DSL skinny GEMM extended to GLM-5.2 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4974961267&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49791&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49791/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49791&quot;&gt;#49791&lt;/a&gt;).&lt;/li&gt;472&lt;li&gt;&lt;strong&gt;Speculative decoding coverage&lt;/strong&gt;: EAGLE3 support declared on KimiLinear (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5142174474&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52171&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52171/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52171&quot;&gt;#52171&lt;/a&gt;), Qwen3.6 dSpark acceptance coverage (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084816525&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51310&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51310/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51310&quot;&gt;#51310&lt;/a&gt;).&lt;/li&gt;473&lt;li&gt;&lt;strong&gt;Correctness&lt;/strong&gt;: MiniMax-M3 NVFP4 inference (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4909402263&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48929&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48929/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48929&quot;&gt;#48929&lt;/a&gt;) and compressed-tensors FP8 MoE SwiGLU params (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4753355344&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46845&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46845/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46845&quot;&gt;#46845&lt;/a&gt;), Gemma3n/Gemma4 variable-length audio batch padding (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5056517829&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50958&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50958/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50958&quot;&gt;#50958&lt;/a&gt;), Gemma 4 compatibility with the upcoming Transformers version (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4975118381&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49797&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49797/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49797&quot;&gt;#49797&lt;/a&gt;), and a Qwen3-Omni crash on video without an audio track (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4867742380&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48420&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48420/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48420&quot;&gt;#48420&lt;/a&gt;).&lt;/li&gt;474&lt;li&gt;&lt;strong&gt;Multimodal performance&lt;/strong&gt;: fused on-device multimodal preprocess normalization (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5018037064&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50411&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50411/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50411&quot;&gt;#50411&lt;/a&gt;), faster placeholder and token-match scanning (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040490475&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50716&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50716/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50716&quot;&gt;#50716&lt;/a&gt;), and repeated prompt-update scans avoided (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117550492&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51774&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51774/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51774&quot;&gt;#51774&lt;/a&gt;).&lt;/li&gt;475&lt;/ul&gt;476&lt;h2&gt;Engine Core&lt;/h2&gt;477&lt;ul&gt;478&lt;li&gt;&lt;strong&gt;Speculative decoding&lt;/strong&gt;: DSpark confidence-scheduled verification (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4824929811&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47808&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47808/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47808&quot;&gt;#47808&lt;/a&gt;), top-k DSpark Markov projection (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4986937996&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49969&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49969/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49969&quot;&gt;#49969&lt;/a&gt;), DFlash2 with local convolution and a candidate selector (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185508751&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52816&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52816/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52816&quot;&gt;#52816&lt;/a&gt;), async scheduling auto-enabled for draft models (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4862124280&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48341&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48341/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48341&quot;&gt;#48341&lt;/a&gt;), fused MTP trailing all-reduce with local-argmax draft tokens (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4974961517&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49793&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49793/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49793&quot;&gt;#49793&lt;/a&gt;), and an adaptive budget for speculative scheduled input tokens (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5113682718&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51725&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51725/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51725&quot;&gt;#51725&lt;/a&gt;).&lt;/li&gt;479&lt;li&gt;&lt;strong&gt;KV cache &amp;amp; scheduling&lt;/strong&gt;: per-request scheduling for MLA chunked context (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5032406796&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50613&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50613/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50613&quot;&gt;#50613&lt;/a&gt;), partial-tail prefix reuse with fine-grained prefix matching (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5025949938&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50507&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50507/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50507&quot;&gt;#50507&lt;/a&gt;), backend-published KV packing in the KV-cache layout refactor (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5106577224&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51612&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51612/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51612&quot;&gt;#51612&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5112161404&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51704&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51704/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51704&quot;&gt;#51704&lt;/a&gt;), LIFO free-block reuse order restored when prefix caching is off (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5095892290&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51482&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51482/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51482&quot;&gt;#51482&lt;/a&gt;), and silent request skipping in priority scheduling fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4929919653&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49206&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49206/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49206&quot;&gt;#49206&lt;/a&gt;).&lt;/li&gt;480&lt;li&gt;&lt;strong&gt;Performance&lt;/strong&gt;: continued elimination of GPU&amp;lt;-&amp;gt;CPU syncs on the execution path (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5094482379&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51458&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51458/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51458&quot;&gt;#51458&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5114866861&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51738&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51738/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51738&quot;&gt;#51738&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5153411613&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52369&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52369/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52369&quot;&gt;#52369&lt;/a&gt;) now guarded by a CI sync check (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4478832887&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43107&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43107/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43107&quot;&gt;#43107&lt;/a&gt;), new JIT warmup infrastructure with predicate filtering (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4937589477&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49315&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49315/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49315&quot;&gt;#49315&lt;/a&gt;), the top-k/top-p Triton sampler launched with 8 warps (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5098414548&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51507&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51507/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51507&quot;&gt;#51507&lt;/a&gt;), detokenization skipped in offline beam search (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5013860355&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50333&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50333/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50333&quot;&gt;#50333&lt;/a&gt;), Mask Replay (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4957847501&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49577&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49577/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49577&quot;&gt;#49577&lt;/a&gt;), optimized long-context MLA cache gathers (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5114956422&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51739&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51739/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51739&quot;&gt;#51739&lt;/a&gt;), and HF revisions resolved to a commit hash once per model load (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4988652213&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49990&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49990/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49990&quot;&gt;#49990&lt;/a&gt;).&lt;/li&gt;481&lt;li&gt;&lt;strong&gt;Hybrid/Mamba&lt;/strong&gt;: prefix caching on by default (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5057990179&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50991&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50991/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50991&quot;&gt;#50991&lt;/a&gt;), the final part of the Mamba attention module refactor (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4611638222&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44857&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44857/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44857&quot;&gt;#44857&lt;/a&gt;), 3D-grid tiling of the state-copy Triton kernels (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4947786164&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49436&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49436/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49436&quot;&gt;#49436&lt;/a&gt;), and Mamba alignment applied before encoder caps (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5106204681&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51603&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51603/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51603&quot;&gt;#51603&lt;/a&gt;).&lt;/li&gt;482&lt;li&gt;&lt;strong&gt;RL workflows&lt;/strong&gt;: stateful trainer send over NCCL and sparse NCCL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5052865253&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50902&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50902/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50902&quot;&gt;#50902&lt;/a&gt;), &lt;code&gt;CuMemAllocator.discard()&lt;/code&gt; for tag-selective GPU memory release (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163599130&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52514&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52514/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52514&quot;&gt;#52514&lt;/a&gt;), level-2 sleep/wake/reload fixed with LoRA enabled (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4270869719&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/39935&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/39935/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/39935&quot;&gt;#39935&lt;/a&gt;), and rewritten weight-transfer docs with standardized examples (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5114163681&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51729&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51729/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51729&quot;&gt;#51729&lt;/a&gt;).&lt;/li&gt;483&lt;li&gt;&lt;strong&gt;Startup robustness&lt;/strong&gt;: file:// rendezvous for single-node executors eliminates startup port races (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5058472637&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50999&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50999/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50999&quot;&gt;#50999&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5108869919&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51652&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51652/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51652&quot;&gt;#51652&lt;/a&gt;), frontend processes are watched during engine startup (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4501773174&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43417&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43417/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43417&quot;&gt;#43417&lt;/a&gt;), a &lt;code&gt;get_open_port()&lt;/code&gt; livelock on DP-reserved ports was fixed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5056737065&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50965&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50965/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50965&quot;&gt;#50965&lt;/a&gt;), and NVML is no longer re-initialized on every device-capability check (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5017351693&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50393&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50393/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50393&quot;&gt;#50393&lt;/a&gt;).&lt;/li&gt;484&lt;/ul&gt;485&lt;h2&gt;Hardware &amp;amp; Performance&lt;/h2&gt;486&lt;ul&gt;487&lt;li&gt;&lt;strong&gt;NVIDIA&lt;/strong&gt;: FlashInfer XQA decode support on SM12x (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4968310920&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49718&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49718/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49718&quot;&gt;#49718&lt;/a&gt;), a CuTeDSL fused query kernel on SM100 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4974961399&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49792&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49792/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49792&quot;&gt;#49792&lt;/a&gt;), programmatic dependent launch for the DSA decode kernels (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5007213321&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50230&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50230/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50230&quot;&gt;#50230&lt;/a&gt;), the native DSA decode path for MTP=3 on SM90 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5141736925&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52164&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52164/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52164&quot;&gt;#52164&lt;/a&gt;), GB10 fused-MoE FP8 tuning configs (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163209275&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52502&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52502/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52502&quot;&gt;#52502&lt;/a&gt;), and B12X dense linear backends (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5132880482&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52016&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52016/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52016&quot;&gt;#52016&lt;/a&gt;).&lt;/li&gt;488&lt;li&gt;&lt;strong&gt;AMD ROCm&lt;/strong&gt;: torch 2.12 / triton 3.7 stack bump (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5032123371&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50607&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50607/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50607&quot;&gt;#50607&lt;/a&gt;), AITER and FP8 inference enabled on GFX120x (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4518781910&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43615&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43615/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43615&quot;&gt;#43615&lt;/a&gt;), DeepSeek-V4 on gfx11 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4766317415&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47017&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47017/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47017&quot;&gt;#47017&lt;/a&gt;), optimized Triton sparse-MLA decode on gfx950 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5145545577&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52212&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52212/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52212&quot;&gt;#52212&lt;/a&gt;), FlyDSL decode-attention kernel for 4-bit TurboQuant KV cache (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4830518406&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47896&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47896/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47896&quot;&gt;#47896&lt;/a&gt;) and an fp8 MQA logits kernel on gfx942 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4956444620&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49544&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49544/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49544&quot;&gt;#49544&lt;/a&gt;), a fused Kimi-K3 KDA decode kernel (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5035155331&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50654&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50654/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50654&quot;&gt;#50654&lt;/a&gt;), fused bf16→fp32 router GEMM (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5009779919&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50268&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50268/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50268&quot;&gt;#50268&lt;/a&gt;), pinned memory on supported WSL2 kernels (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4997388876&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50126&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50126/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50126&quot;&gt;#50126&lt;/a&gt;), and preshuffled sparse indexing for 16-token blocks (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5076855828&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51216&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51216/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51216&quot;&gt;#51216&lt;/a&gt;).&lt;/li&gt;489&lt;li&gt;&lt;strong&gt;Intel XPU&lt;/strong&gt;: a torch linear backend including blockwise GEMM (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4964968726&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49664&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49664/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49664&quot;&gt;#49664&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5047328210&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50826&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50826/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50826&quot;&gt;#50826&lt;/a&gt;), MXFP8 linear weights for the INC DeepSeek V4 model (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4871724476&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48476&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48476/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48476&quot;&gt;#48476&lt;/a&gt;), async-scheduling PP sampled-token broadcast overlapped with compute (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5108748890&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51650&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51650/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51650&quot;&gt;#51650&lt;/a&gt;), an XPU wheel added to the release pipeline (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5138581112&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52108&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52108/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52108&quot;&gt;#52108&lt;/a&gt;), tuned Mamba SSU configs for Arc Pro B70 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5026981323&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50534&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50534/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50534&quot;&gt;#50534&lt;/a&gt;), and UVA weight offloading fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117276306&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51770&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51770/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51770&quot;&gt;#51770&lt;/a&gt;).&lt;/li&gt;490&lt;li&gt;&lt;strong&gt;CPU&lt;/strong&gt;: an MLA backend so DeepSeek-V2/V3 can run on CPU (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4948910396&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49453&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49453/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49453&quot;&gt;#49453&lt;/a&gt;), a triton-cpu wheel (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5137623585&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52092&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52092/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52092&quot;&gt;#52092&lt;/a&gt;), GPTQ and AWQ enabled on s390x (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5070079354&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51148&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51148/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51148&quot;&gt;#51148&lt;/a&gt;) along with tcmalloc (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5048108687&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50841&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50841/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50841&quot;&gt;#50841&lt;/a&gt;), BF16 MoE routed through zentorch on AMD (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4562633711&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44201&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44201/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44201&quot;&gt;#44201&lt;/a&gt;), an unquantized MoE backend for Power (VSX) (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5107432976&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51624&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51624/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51624&quot;&gt;#51624&lt;/a&gt;), unquantized MoE migrated to the modular-kernel experts structure (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4997915103&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50133&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50133/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50133&quot;&gt;#50133&lt;/a&gt;), and the MXFP4 block scale folded in 2 instructions instead of 4 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5104914644&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51583&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51583/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51583&quot;&gt;#51583&lt;/a&gt;).&lt;/li&gt;491&lt;/ul&gt;492&lt;h2&gt;Large Scale Serving&lt;/h2&gt;493&lt;ul&gt;494&lt;li&gt;&lt;strong&gt;E/P/D disaggregation&lt;/strong&gt;: Model Runner V2 E/P/D support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4155353913&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/38390&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/38390/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/38390&quot;&gt;#38390&lt;/a&gt;), duplicate image preprocessing removed with GPU-side preprocessing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5017201855&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50390&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50390/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50390&quot;&gt;#50390&lt;/a&gt;), KV consumers may omit multimodal embeddings (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5178516449&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52697&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52697/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52697&quot;&gt;#52697&lt;/a&gt;), encoder-instance requests kept alive until their images are encoded (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5010138054&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50275&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50275/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50275&quot;&gt;#50275&lt;/a&gt;), and EC connector scheduler/worker metadata plumbing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4958264375&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49579&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49579/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49579&quot;&gt;#49579&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4958842553&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49585&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49585/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49585&quot;&gt;#49585&lt;/a&gt;).&lt;/li&gt;495&lt;li&gt;&lt;strong&gt;KV offloading&lt;/strong&gt;: disk offloading for SimpleCPUOffloadConnector (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4963653413&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49644&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49644/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49644&quot;&gt;#49644&lt;/a&gt;), out-of-tree secondary tier managers via &lt;code&gt;module_path&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5058868734&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51007&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51007/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51007&quot;&gt;#51007&lt;/a&gt;), partial secondary-tier load results (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5013391634&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50321&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50321/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50321&quot;&gt;#50321&lt;/a&gt;), tiering offloading metrics (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4898117346&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48798&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48798/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48798&quot;&gt;#48798&lt;/a&gt;), data-parallel topology exposed to offloading backends (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124610407&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51879&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51879/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51879&quot;&gt;#51879&lt;/a&gt;), a canonical CPU layout for parallelism-agnostic offload (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4867574663&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48414&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48414/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48414&quot;&gt;#48414&lt;/a&gt;), and quadratic ARC batch eviction avoided (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5058051344&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50992&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50992/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50992&quot;&gt;#50992&lt;/a&gt;).&lt;/li&gt;496&lt;li&gt;&lt;strong&gt;Mooncake&lt;/strong&gt;: store group semantics (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4619061887&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44956&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44956/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44956&quot;&gt;#44956&lt;/a&gt;), tenant ID support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4842736295&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48069&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48069/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48069&quot;&gt;#48069&lt;/a&gt;), and official wheels in the Docker image (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5063707796&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51067&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51067/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51067&quot;&gt;#51067&lt;/a&gt;).&lt;/li&gt;497&lt;li&gt;&lt;strong&gt;Connectors&lt;/strong&gt;: transfer mode (push/pull) included in the NIXL compatibility hash (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5032823656&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50620&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50620/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50620&quot;&gt;#50620&lt;/a&gt;), a MoRIIO per-layer READ-completion barrier (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4877993752&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48534&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48534/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48534&quot;&gt;#48534&lt;/a&gt;), 2P2D wide-EP with mori-ep/mori-io at dp=ep=16 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4623803757&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45043&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45043/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45043&quot;&gt;#45043&lt;/a&gt;), and stale remote cleanup in the push connector (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5007459375&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50234&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50234/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50234&quot;&gt;#50234&lt;/a&gt;).&lt;/li&gt;498&lt;li&gt;&lt;strong&gt;Parallelism&lt;/strong&gt;: EPLB balancedness calculation fixed and tested (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5119937614&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51813&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51813/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51813&quot;&gt;#51813&lt;/a&gt;), dense multinode DP rescope (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4930303331&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49212&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49212/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49212&quot;&gt;#49212&lt;/a&gt;).&lt;/li&gt;499&lt;/ul&gt;500&lt;h2&gt;Quantization&lt;/h2&gt;501&lt;ul&gt;502&lt;li&gt;&lt;strong&gt;Online quantization&lt;/strong&gt;: online MXFP4 support (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4940594936&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49347&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49347/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49347&quot;&gt;#49347&lt;/a&gt;), online weight scales shared across TP (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4972474663&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49764&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49764/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49764&quot;&gt;#49764&lt;/a&gt;), precision preserved in online NVFP4 expert packing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4991067162&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50029&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50029/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50029&quot;&gt;#50029&lt;/a&gt;), and the online NVFP4 MoE kernel reused across reloads (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4994606400&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50074&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50074/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50074&quot;&gt;#50074&lt;/a&gt;).&lt;/li&gt;503&lt;li&gt;&lt;strong&gt;NVFP4&lt;/strong&gt;: batch-invariant NVFP4 MoE via CUTLASS (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4296607986&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/40372&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/40372/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/40372&quot;&gt;#40372&lt;/a&gt;), KV 4-over-6 scale search (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4634559370&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45187&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45187/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45187&quot;&gt;#45187&lt;/a&gt;), CuTeDSL MoE with SwiGLU-OAI and ReLU2 activations (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4773510143&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47106&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47106/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47106&quot;&gt;#47106&lt;/a&gt;), and out_dtype matched to the model dtype (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4904369637&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48861&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48861/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48861&quot;&gt;#48861&lt;/a&gt;).&lt;/li&gt;504&lt;li&gt;&lt;strong&gt;New kernels&lt;/strong&gt;: block-wise scaled_mm (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4984081053&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49932&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49932/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49932&quot;&gt;#49932&lt;/a&gt;), DeepSeek-V4 AMD Quark NVFP4 with an emulation kernel (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4835520784&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47972&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47972/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47972&quot;&gt;#47972&lt;/a&gt;).&lt;/li&gt;505&lt;li&gt;&lt;strong&gt;Fixes&lt;/strong&gt;: dynamic INT8 W8A8 MoE config no longer built as W8A16 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5047730226&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50833&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50833/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50833&quot;&gt;#50833&lt;/a&gt;) and a TritonExperts crash (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5091616049&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51411&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51411/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51411&quot;&gt;#51411&lt;/a&gt;), MXFP4 conversion for FlashInfer CUTLASS (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5061729452&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51038&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51038/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51038&quot;&gt;#51038&lt;/a&gt;), fp32 weight scales and per-expert checkpoint mapping for MXFP4 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5091982004&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51419&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51419/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51419&quot;&gt;#51419&lt;/a&gt;), and fused block-scale orientation (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040990227&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50727&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50727/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50727&quot;&gt;#50727&lt;/a&gt;).&lt;/li&gt;506&lt;/ul&gt;507&lt;h2&gt;API &amp;amp; Frontend&lt;/h2&gt;508&lt;ul&gt;509&lt;li&gt;&lt;strong&gt;New capabilities&lt;/strong&gt;: request priority parsed from an HTTP header (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5066001756&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51089&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51089/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51089&quot;&gt;#51089&lt;/a&gt;), session ID plumbing into requests (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4841177349&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48048&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48048/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48048&quot;&gt;#48048&lt;/a&gt;), &lt;code&gt;count_reasoning_tokens&lt;/code&gt; in the streaming parser engine (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4672899048&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45802&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45802/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45802&quot;&gt;#45802&lt;/a&gt;), &lt;code&gt;content_parts&lt;/code&gt; on &lt;code&gt;/inference/v1/generate&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5095831219&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51478&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51478/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51478&quot;&gt;#51478&lt;/a&gt;), &lt;code&gt;model&lt;/code&gt; optional on all &lt;code&gt;/derender&lt;/code&gt; request classes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5094894762&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51463&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51463/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51463&quot;&gt;#51463&lt;/a&gt;), output token IDs logged at DEBUG level (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5137940807&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52098&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52098/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52098&quot;&gt;#52098&lt;/a&gt;), and vLLM Recipes connected to native config-based deployment and benchmarking (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084531279&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51308&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51308/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51308&quot;&gt;#51308&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124601163&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51878&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51878/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51878&quot;&gt;#51878&lt;/a&gt;).&lt;/li&gt;510&lt;li&gt;&lt;strong&gt;Rust frontend&lt;/strong&gt;: a standalone renderer (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5011355404&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50289&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50289/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50289&quot;&gt;#50289&lt;/a&gt;), gRPC multimodal image inference (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5015640495&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50368&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50368/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50368&quot;&gt;#50368&lt;/a&gt;), explicit data-parallel rank routing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5072919309&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51178&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51178/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51178&quot;&gt;#51178&lt;/a&gt;), RL lifecycle control (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5084945361&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51316&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51316/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51316&quot;&gt;#51316&lt;/a&gt;), dynamic tools from developer messages (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5069432556&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51144&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51144/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51144&quot;&gt;#51144&lt;/a&gt;), protobuf schemas published to Buf (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5082805927&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51276&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51276/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51276&quot;&gt;#51276&lt;/a&gt;), and MiniJinja upgraded to 2.22 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5078832228&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51235&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51235/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51235&quot;&gt;#51235&lt;/a&gt;).&lt;/li&gt;511&lt;li&gt;&lt;strong&gt;Anthropic API&lt;/strong&gt;: 4xx returned for client-caused errors on &lt;code&gt;/v1/messages&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5147551906&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52246&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52246/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52246&quot;&gt;#52246&lt;/a&gt;), &lt;code&gt;disable_parallel_tool_use&lt;/code&gt; preserved (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5133246623&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52021&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52021/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52021&quot;&gt;#52021&lt;/a&gt;), and stop sequences bounded (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5131747956&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51997&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51997/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51997&quot;&gt;#51997&lt;/a&gt;).&lt;/li&gt;512&lt;li&gt;&lt;strong&gt;Cohere&lt;/strong&gt;: upstreamed parser fixes (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5131753634&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51998&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51998/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51998&quot;&gt;#51998&lt;/a&gt;), stop sequences reported correctly (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5102649101&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51556&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51556/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51556&quot;&gt;#51556&lt;/a&gt;), and vectorized binary embedding bit-packing (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5148708378&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52277&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52277/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52277&quot;&gt;#52277&lt;/a&gt;).&lt;/li&gt;513&lt;li&gt;&lt;strong&gt;Structured output&lt;/strong&gt;: request stop tokens masked in xgrammar until the grammar terminates (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4931076709&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49227&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49227/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49227&quot;&gt;#49227&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5031582024&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50595&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50595/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50595&quot;&gt;#50595&lt;/a&gt;), NUL bytes rejected in &lt;code&gt;structured_outputs.regex&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5118845323&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51796&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51796/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51796&quot;&gt;#51796&lt;/a&gt;), negative token IDs rejected as out-of-vocabulary (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5118829807&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51795&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51795/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51795&quot;&gt;#51795&lt;/a&gt;), and &lt;code&gt;VLLMValidationError&lt;/code&gt; raised from validators (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5156138521&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52394&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52394/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52394&quot;&gt;#52394&lt;/a&gt;).&lt;/li&gt;514&lt;li&gt;&lt;strong&gt;Robustness&lt;/strong&gt;: uvicorn signal handlers disabled instead of racing them (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5054208715&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50916&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50916/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50916&quot;&gt;#50916&lt;/a&gt;), a consolidated entrypoint exception handler (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5148463554&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52261&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52261/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52261&quot;&gt;#52261&lt;/a&gt;), 400 instead of 500 on non-object JSON bodies (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109037008&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51654&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51654/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51654&quot;&gt;#51654&lt;/a&gt;, &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5164650433&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52528&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52528/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52528&quot;&gt;#52528&lt;/a&gt;), generation inputs bounded before expensive work (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5093882388&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51447&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51447/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51447&quot;&gt;#51447&lt;/a&gt;), and &lt;code&gt;cache_salt&lt;/code&gt; now required to be non-empty (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5046668464&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50816&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50816/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50816&quot;&gt;#50816&lt;/a&gt;).&lt;/li&gt;515&lt;/ul&gt;516&lt;h2&gt;Security&lt;/h2&gt;517&lt;ul&gt;518&lt;li&gt;Fixed a DoS via sample-rate forgery that bypassed the audio decode duration guard (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4985115111&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49948&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49948/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49948&quot;&gt;#49948&lt;/a&gt;); the audio decode duration limit is now also enforced in NanoNemotronVL (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5006702499&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50221&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50221/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50221&quot;&gt;#50221&lt;/a&gt;).&lt;/li&gt;519&lt;li&gt;DeepStream classified as a GPU backend with pixel limits enforced (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5043666962&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50755&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50755/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50755&quot;&gt;#50755&lt;/a&gt;).&lt;/li&gt;520&lt;li&gt;&lt;code&gt;_load_ov2_processor&lt;/code&gt; guarded with &lt;code&gt;resolve_trust_remote_code&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5193797677&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52952&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52952/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52952&quot;&gt;#52952&lt;/a&gt;).&lt;/li&gt;521&lt;li&gt;Documentation now warns that &lt;code&gt;--api-key&lt;/code&gt; does not gate all endpoints (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5131764514&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51999&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51999/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51999&quot;&gt;#51999&lt;/a&gt;).&lt;/li&gt;522&lt;/ul&gt;523&lt;h2&gt;Dependencies&lt;/h2&gt;524&lt;ul&gt;525&lt;li&gt;Transformers 5.15.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109844757&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51668&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51668/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51668&quot;&gt;#51668&lt;/a&gt;), huggingface-hub 1.27.0 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5092131862&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51422&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51422/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51422&quot;&gt;#51422&lt;/a&gt;), fastsafetensors upgrade (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5047371176&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50827&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50827/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50827&quot;&gt;#50827&lt;/a&gt;).&lt;/li&gt;526&lt;li&gt;ROCm: torch 2.12, triton 3.7, torchaudio, torchvision (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5032123371&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50607&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50607/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50607&quot;&gt;#50607&lt;/a&gt;).&lt;/li&gt;527&lt;li&gt;Runtime image upgraded to Ubuntu 24.04, picking up rdma-core &amp;gt; 44 (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5063091059&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51058&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51058/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51058&quot;&gt;#51058&lt;/a&gt;).&lt;/li&gt;528&lt;li&gt;DeepGEMM pinned to the deepseek-ai nv_dev tip (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5134204828&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52035&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52035/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52035&quot;&gt;#52035&lt;/a&gt;), DeepEP pinned by full commit hash (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5133584994&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52028&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52028/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52028&quot;&gt;#52028&lt;/a&gt;), FlashAttention 3 built with the torch stable API (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4960056521&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49599&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49599/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49599&quot;&gt;#49599&lt;/a&gt;).&lt;/li&gt;529&lt;/ul&gt;530&lt;h2&gt;Breaking Changes &amp;amp; Deprecations&lt;/h2&gt;531&lt;ul&gt;532&lt;li&gt;bitsandbytes support is now an out-of-tree plugin (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4512000173&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43529&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43529/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43529&quot;&gt;#43529&lt;/a&gt;).&lt;/li&gt;533&lt;li&gt;The deprecated &lt;code&gt;calculate_kv_scales&lt;/code&gt; runtime KV scale calculation was removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4944675936&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49389&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49389/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49389&quot;&gt;#49389&lt;/a&gt;).&lt;/li&gt;534&lt;li&gt;&lt;code&gt;override_attention_dtype&lt;/code&gt; was removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4888820267&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48684&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48684/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48684&quot;&gt;#48684&lt;/a&gt;).&lt;/li&gt;535&lt;li&gt;&lt;code&gt;reasoning_content&lt;/code&gt; output removal is documented as a breaking client change (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5033215124&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50624&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50624/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50624&quot;&gt;#50624&lt;/a&gt;).&lt;/li&gt;536&lt;li&gt;KV offload tiering metrics renamed from &lt;code&gt;kv_offload_tiering_block_{queries,hits}&lt;/code&gt; to &lt;code&gt;..._chunk_...&lt;/code&gt; (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5185155341&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52812&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52812/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52812&quot;&gt;#52812&lt;/a&gt;).&lt;/li&gt;537&lt;li&gt;MoE legacy code removed (&lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5064749809&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51078&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51078/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51078&quot;&gt;#51078&lt;/a&gt;).&lt;/li&gt;538&lt;/ul&gt;539&lt;h2&gt;New Contributors&lt;/h2&gt;540&lt;ul&gt;541&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/acheamponge/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/acheamponge&quot;&gt;@acheamponge&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4941429215&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49353&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49353/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49353&quot;&gt;#49353&lt;/a&gt;&lt;/li&gt;542&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/acmore/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/acmore&quot;&gt;@acmore&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5080785739&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51259&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51259/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51259&quot;&gt;#51259&lt;/a&gt;&lt;/li&gt;543&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anhtra3889/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anhtra3889&quot;&gt;@anhtra3889&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5058701438&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51002&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51002/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51002&quot;&gt;#51002&lt;/a&gt;&lt;/li&gt;544&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anujbolewar/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anujbolewar&quot;&gt;@anujbolewar&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5049889451&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50867&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50867/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50867&quot;&gt;#50867&lt;/a&gt;&lt;/li&gt;545&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/arthurgao2003/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/arthurgao2003&quot;&gt;@arthurgao2003&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4852532476&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48215&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48215/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48215&quot;&gt;#48215&lt;/a&gt;&lt;/li&gt;546&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/baodii/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/baodii&quot;&gt;@baodii&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5076795320&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51215&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51215/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51215&quot;&gt;#51215&lt;/a&gt;&lt;/li&gt;547&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bitborne/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bitborne&quot;&gt;@bitborne&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4619061887&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/44956&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/44956/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/44956&quot;&gt;#44956&lt;/a&gt;&lt;/li&gt;548&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bohnstingl/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bohnstingl&quot;&gt;@bohnstingl&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4949852429&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49458&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49458/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49458&quot;&gt;#49458&lt;/a&gt;&lt;/li&gt;549&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ccaadaro/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ccaadaro&quot;&gt;@ccaadaro&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5104914644&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51583&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51583/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51583&quot;&gt;#51583&lt;/a&gt;&lt;/li&gt;550&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cmiyai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cmiyai&quot;&gt;@cmiyai&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4754819791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46870&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46870/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46870&quot;&gt;#46870&lt;/a&gt;&lt;/li&gt;551&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/d4l3k/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/d4l3k&quot;&gt;@d4l3k&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5066535102&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51097&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51097/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51097&quot;&gt;#51097&lt;/a&gt;&lt;/li&gt;552&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmai-afk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmai-afk&quot;&gt;@dmai-afk&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5067868281&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51120&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51120/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51120&quot;&gt;#51120&lt;/a&gt;&lt;/li&gt;553&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmholtz/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmholtz&quot;&gt;@dmholtz&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5057306751&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50977&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50977/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50977&quot;&gt;#50977&lt;/a&gt;&lt;/li&gt;554&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/efschu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/efschu&quot;&gt;@efschu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5041533191&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50734&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50734/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50734&quot;&gt;#50734&lt;/a&gt;&lt;/li&gt;555&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fangchenli/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fangchenli&quot;&gt;@fangchenli&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5148708378&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52277&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52277/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52277&quot;&gt;#52277&lt;/a&gt;&lt;/li&gt;556&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fanxingran/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fanxingran&quot;&gt;@fanxingran&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5059044965&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51011&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51011/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51011&quot;&gt;#51011&lt;/a&gt;&lt;/li&gt;557&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fatday/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fatday&quot;&gt;@fatday&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4850038638&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48171&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48171/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48171&quot;&gt;#48171&lt;/a&gt;&lt;/li&gt;558&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fattchris/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fattchris&quot;&gt;@fattchris&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4904369637&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48861&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48861/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48861&quot;&gt;#48861&lt;/a&gt;&lt;/li&gt;559&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fcui-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fcui-amd&quot;&gt;@fcui-amd&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4997388876&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50126&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50126/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50126&quot;&gt;#50126&lt;/a&gt;&lt;/li&gt;560&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fede-kamel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fede-kamel&quot;&gt;@fede-kamel&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5033215124&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50624&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50624/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50624&quot;&gt;#50624&lt;/a&gt;&lt;/li&gt;561&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fxfxfxfxfxfxfxfx/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fxfxfxfxfxfxfxfx&quot;&gt;@fxfxfxfxfxfxfxfx&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4925208326&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49139&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49139/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49139&quot;&gt;#49139&lt;/a&gt;&lt;/li&gt;562&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gabriel-peracio/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gabriel-peracio&quot;&gt;@gabriel-peracio&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5002983791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50183&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50183/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50183&quot;&gt;#50183&lt;/a&gt;&lt;/li&gt;563&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gchinora/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gchinora&quot;&gt;@gchinora&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5092593137&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51427&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51427/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51427&quot;&gt;#51427&lt;/a&gt;&lt;/li&gt;564&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/guanxingithub/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/guanxingithub&quot;&gt;@guanxingithub&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5031207461&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50589&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50589/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50589&quot;&gt;#50589&lt;/a&gt;&lt;/li&gt;565&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/haregali/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/haregali&quot;&gt;@haregali&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5040490475&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50716&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50716/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50716&quot;&gt;#50716&lt;/a&gt;&lt;/li&gt;566&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hsusul/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hsusul&quot;&gt;@hsusul&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4960695993&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49613&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49613/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49613&quot;&gt;#49613&lt;/a&gt;&lt;/li&gt;567&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/iwannagotobed/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/iwannagotobed&quot;&gt;@iwannagotobed&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5109640766&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51664&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51664/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51664&quot;&gt;#51664&lt;/a&gt;&lt;/li&gt;568&lt;li&gt;@jacobzhang22 made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4190902375&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/38771&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/38771/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/38771&quot;&gt;#38771&lt;/a&gt;&lt;/li&gt;569&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jairitAge/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jairitAge&quot;&gt;@jairitAge&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5066645094&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51100&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51100/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51100&quot;&gt;#51100&lt;/a&gt;&lt;/li&gt;570&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jamesETsmith/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jamesETsmith&quot;&gt;@jamesETsmith&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5076855828&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51216&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51216/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51216&quot;&gt;#51216&lt;/a&gt;&lt;/li&gt;571&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jimmy-adams/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jimmy-adams&quot;&gt;@jimmy-adams&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4835520784&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47972&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47972/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47972&quot;&gt;#47972&lt;/a&gt;&lt;/li&gt;572&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jyan-R/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jyan-R&quot;&gt;@jyan-R&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5150426007&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52311&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52311/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52311&quot;&gt;#52311&lt;/a&gt;&lt;/li&gt;573&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Kaif10/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Kaif10&quot;&gt;@Kaif10&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5164650433&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52528&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52528/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52528&quot;&gt;#52528&lt;/a&gt;&lt;/li&gt;574&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/karen-sy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/karen-sy&quot;&gt;@karen-sy&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4841177349&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48048&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48048/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48048&quot;&gt;#48048&lt;/a&gt;&lt;/li&gt;575&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Lin-z-w/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Lin-z-w&quot;&gt;@Lin-z-w&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4842736295&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48069&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48069/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48069&quot;&gt;#48069&lt;/a&gt;&lt;/li&gt;576&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/liushujia122/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/liushujia122&quot;&gt;@liushujia122&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5117831469&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51780&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51780/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51780&quot;&gt;#51780&lt;/a&gt;&lt;/li&gt;577&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Luosuu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Luosuu&quot;&gt;@Luosuu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4896453823&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48789&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48789/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48789&quot;&gt;#48789&lt;/a&gt;&lt;/li&gt;578&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mispa-ms/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mispa-ms&quot;&gt;@mispa-ms&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5157569183&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52419&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52419/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52419&quot;&gt;#52419&lt;/a&gt;&lt;/li&gt;579&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mkhazraee/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mkhazraee&quot;&gt;@mkhazraee&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5013391634&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50321&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50321/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50321&quot;&gt;#50321&lt;/a&gt;&lt;/li&gt;580&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/NVShreyas/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/NVShreyas&quot;&gt;@NVShreyas&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5053519932&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50911&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50911/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50911&quot;&gt;#50911&lt;/a&gt;&lt;/li&gt;581&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pavelzak/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pavelzak&quot;&gt;@pavelzak&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5163209275&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52502&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52502/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52502&quot;&gt;#52502&lt;/a&gt;&lt;/li&gt;582&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/positive666/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/positive666&quot;&gt;@positive666&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5151372357&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52329&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52329/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52329&quot;&gt;#52329&lt;/a&gt;&lt;/li&gt;583&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rajfirke/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rajfirke&quot;&gt;@rajfirke&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5104027805&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51573&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51573/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51573&quot;&gt;#51573&lt;/a&gt;&lt;/li&gt;584&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Rapisurazurite/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Rapisurazurite&quot;&gt;@Rapisurazurite&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5056274756&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50950&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50950/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50950&quot;&gt;#50950&lt;/a&gt;&lt;/li&gt;585&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rchalamala/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rchalamala&quot;&gt;@rchalamala&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5024552383&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50487&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50487/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50487&quot;&gt;#50487&lt;/a&gt;&lt;/li&gt;586&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RobbieJ/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RobbieJ&quot;&gt;@RobbieJ&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4938814574&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49328&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49328/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49328&quot;&gt;#49328&lt;/a&gt;&lt;/li&gt;587&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ruirui6946/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ruirui6946&quot;&gt;@ruirui6946&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5137940807&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52098&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52098/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52098&quot;&gt;#52098&lt;/a&gt;&lt;/li&gt;588&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RyanJHamby/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RyanJHamby&quot;&gt;@RyanJHamby&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4867742380&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/48420&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/48420/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/48420&quot;&gt;#48420&lt;/a&gt;&lt;/li&gt;589&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/samuelkim7/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/samuelkim7&quot;&gt;@samuelkim7&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5013860355&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50333&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50333/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50333&quot;&gt;#50333&lt;/a&gt;&lt;/li&gt;590&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shanewidanagama/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shanewidanagama&quot;&gt;@shanewidanagama&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5125777348&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51901&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51901/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51901&quot;&gt;#51901&lt;/a&gt;&lt;/li&gt;591&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shikamd123/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shikamd123&quot;&gt;@shikamd123&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4623803757&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/45043&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/45043/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/45043&quot;&gt;#45043&lt;/a&gt;&lt;/li&gt;592&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/SilenNaihin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/SilenNaihin&quot;&gt;@SilenNaihin&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4270869719&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/39935&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/39935/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/39935&quot;&gt;#39935&lt;/a&gt;&lt;/li&gt;593&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/skysnow2001/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/skysnow2001&quot;&gt;@skysnow2001&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4518781910&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/43615&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/43615/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/43615&quot;&gt;#43615&lt;/a&gt;&lt;/li&gt;594&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Sundaresan-G/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Sundaresan-G&quot;&gt;@Sundaresan-G&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5026734206&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50526&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50526/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50526&quot;&gt;#50526&lt;/a&gt;&lt;/li&gt;595&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/syedalijaseem/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/syedalijaseem&quot;&gt;@syedalijaseem&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4816318436&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47692&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47692/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47692&quot;&gt;#47692&lt;/a&gt;&lt;/li&gt;596&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/taking-lying-flat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/taking-lying-flat&quot;&gt;@taking-lying-flat&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5089824977&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51391&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51391/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51391&quot;&gt;#51391&lt;/a&gt;&lt;/li&gt;597&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tandixit95/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tandixit95&quot;&gt;@tandixit95&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5022876167&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50462&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50462/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50462&quot;&gt;#50462&lt;/a&gt;&lt;/li&gt;598&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Tejas-Raj01/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Tejas-Raj01&quot;&gt;@Tejas-Raj01&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4929919653&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49206&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49206/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49206&quot;&gt;#49206&lt;/a&gt;&lt;/li&gt;599&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/theminghuang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/theminghuang&quot;&gt;@theminghuang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5095892290&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51482&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51482/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51482&quot;&gt;#51482&lt;/a&gt;&lt;/li&gt;600&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tobymao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tobymao&quot;&gt;@tobymao&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5085087394&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51318&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51318/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51318&quot;&gt;#51318&lt;/a&gt;&lt;/li&gt;601&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/TrainToGPB/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/TrainToGPB&quot;&gt;@TrainToGPB&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5056517829&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50958&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50958/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50958&quot;&gt;#50958&lt;/a&gt;&lt;/li&gt;602&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tzulingk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tzulingk&quot;&gt;@tzulingk&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4931304334&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49230&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49230/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49230&quot;&gt;#49230&lt;/a&gt;&lt;/li&gt;603&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/UgaTheDev/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/UgaTheDev&quot;&gt;@UgaTheDev&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5107513556&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51627&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51627/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51627&quot;&gt;#51627&lt;/a&gt;&lt;/li&gt;604&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/varoudis/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/varoudis&quot;&gt;@varoudis&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5017659562&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50404&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50404/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50404&quot;&gt;#50404&lt;/a&gt;&lt;/li&gt;605&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Vegetog/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Vegetog&quot;&gt;@Vegetog&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4980552220&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/49876&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/49876/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/49876&quot;&gt;#49876&lt;/a&gt;&lt;/li&gt;606&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vitamin-chaos/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vitamin-chaos&quot;&gt;@vitamin-chaos&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4773510143&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/47106&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/47106/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/47106&quot;&gt;#47106&lt;/a&gt;&lt;/li&gt;607&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wangxian001/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wangxian001&quot;&gt;@wangxian001&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5010158133&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50276&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50276/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50276&quot;&gt;#50276&lt;/a&gt;&lt;/li&gt;608&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/WillZZZy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/WillZZZy&quot;&gt;@WillZZZy&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;4746058933&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/46747&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/46747/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/46747&quot;&gt;#46747&lt;/a&gt;&lt;/li&gt;609&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xiaopusun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xiaopusun&quot;&gt;@xiaopusun&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5096858411&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51495&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51495/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51495&quot;&gt;#51495&lt;/a&gt;&lt;/li&gt;610&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xijiaat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xijiaat&quot;&gt;@xijiaat&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5037831541&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50693&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50693/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50693&quot;&gt;#50693&lt;/a&gt;&lt;/li&gt;611&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xudonlyu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xudonlyu&quot;&gt;@xudonlyu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5110787791&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51682&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51682/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51682&quot;&gt;#51682&lt;/a&gt;&lt;/li&gt;612&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yifjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yifjiang&quot;&gt;@yifjiang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5076989189&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51218&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51218/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51218&quot;&gt;#51218&lt;/a&gt;&lt;/li&gt;613&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yu-xin-c/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yu-xin-c&quot;&gt;@yu-xin-c&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5108869919&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51652&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51652/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51652&quot;&gt;#51652&lt;/a&gt;&lt;/li&gt;614&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zcxGGmu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zcxGGmu&quot;&gt;@zcxGGmu&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5042543748&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/50746&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/50746/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/50746&quot;&gt;#50746&lt;/a&gt;&lt;/li&gt;615&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ziqifan617/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ziqifan617&quot;&gt;@ziqifan617&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5124610407&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/51879&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/51879/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/51879&quot;&gt;#51879&lt;/a&gt;&lt;/li&gt;616&lt;li&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zobinHuang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zobinHuang&quot;&gt;@zobinHuang&lt;/a&gt; made their first contribution in &lt;a class=&quot;issue-link js-issue-link&quot; data-error-text=&quot;Failed to load title&quot; data-id=&quot;5141736925&quot; data-permission-text=&quot;Title is private&quot; data-url=&quot;https://github.com/vllm-project/vllm/issues/52164&quot; data-hovercard-type=&quot;pull_request&quot; data-hovercard-url=&quot;/vllm-project/vllm/pull/52164/hovercard&quot; href=&quot;https://github.com/vllm-project/vllm/pull/52164&quot;&gt;#52164&lt;/a&gt;&lt;/li&gt;617&lt;/ul&gt;618&lt;h2&gt;Contributors&lt;/h2&gt;619&lt;p&gt;&lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yewentao256/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yewentao256&quot;&gt;@yewentao256&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/AndreasKaratzas/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/AndreasKaratzas&quot;&gt;@AndreasKaratzas&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/njhill/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/njhill&quot;&gt;@njhill&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mgoin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mgoin&quot;&gt;@mgoin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/aoshen02/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/aoshen02&quot;&gt;@aoshen02&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/khluu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/khluu&quot;&gt;@khluu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hmellor/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hmellor&quot;&gt;@hmellor&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/stefankoncarevic/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/stefankoncarevic&quot;&gt;@stefankoncarevic&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/taneem-ibrahim/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/taneem-ibrahim&quot;&gt;@taneem-ibrahim&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LucasWilkinson/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LucasWilkinson&quot;&gt;@LucasWilkinson&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/chaunceyjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/chaunceyjiang&quot;&gt;@chaunceyjiang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jikunshang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jikunshang&quot;&gt;@jikunshang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zufangzhu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zufangzhu&quot;&gt;@zufangzhu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bigPYJ1151/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bigPYJ1151&quot;&gt;@bigPYJ1151&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/NickLucche/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/NickLucche&quot;&gt;@NickLucche&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/askliar/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/askliar&quot;&gt;@askliar&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fxmarty-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fxmarty-amd&quot;&gt;@fxmarty-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Rohan138/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Rohan138&quot;&gt;@Rohan138&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gty111/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gty111&quot;&gt;@gty111&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zhenwei-intel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zhenwei-intel&quot;&gt;@zhenwei-intel&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Isotr0py/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Isotr0py&quot;&gt;@Isotr0py&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jeejeelee/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jeejeelee&quot;&gt;@jeejeelee&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ZJY0516/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ZJY0516&quot;&gt;@ZJY0516&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wangxiyuan/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wangxiyuan&quot;&gt;@wangxiyuan&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/BugenZhao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/BugenZhao&quot;&gt;@BugenZhao&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yma11/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yma11&quot;&gt;@yma11&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zyongye/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zyongye&quot;&gt;@zyongye&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/DarkLight1337/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/DarkLight1337&quot;&gt;@DarkLight1337&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jperezdealgaba/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jperezdealgaba&quot;&gt;@jperezdealgaba&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zhou9402/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zhou9402&quot;&gt;@zhou9402&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/connorcarpenter15/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/connorcarpenter15&quot;&gt;@connorcarpenter15&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/kliuae/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/kliuae&quot;&gt;@kliuae&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lucifer1004/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lucifer1004&quot;&gt;@lucifer1004&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Alex-ai-future/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Alex-ai-future&quot;&gt;@Alex-ai-future&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/aarushjain29/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/aarushjain29&quot;&gt;@aarushjain29&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/TheEpicDolphin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/TheEpicDolphin&quot;&gt;@TheEpicDolphin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/chaojun-zhang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/chaojun-zhang&quot;&gt;@chaojun-zhang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pmanczak/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pmanczak&quot;&gt;@pmanczak&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/WoosukKwon/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/WoosukKwon&quot;&gt;@WoosukKwon&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/BabyDrangoner/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/BabyDrangoner&quot;&gt;@BabyDrangoner&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/noooop/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/noooop&quot;&gt;@noooop&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/almogtavor/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/almogtavor&quot;&gt;@almogtavor&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hongxiayang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hongxiayang&quot;&gt;@hongxiayang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fuscof-ibm/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fuscof-ibm&quot;&gt;@fuscof-ibm&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gau-nernst/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gau-nernst&quot;&gt;@gau-nernst&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zxd1997066/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zxd1997066&quot;&gt;@zxd1997066&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tlrmchlsmth/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tlrmchlsmth&quot;&gt;@tlrmchlsmth&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/music-dino/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/music-dino&quot;&gt;@music-dino&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zexplorerhj/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zexplorerhj&quot;&gt;@zexplorerhj&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/S1ro1/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/S1ro1&quot;&gt;@S1ro1&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jdebache/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jdebache&quot;&gt;@jdebache&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sfeng33/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sfeng33&quot;&gt;@sfeng33&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mayuyuace/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mayuyuace&quot;&gt;@mayuyuace&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ganeshr10/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ganeshr10&quot;&gt;@ganeshr10&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/benchislett/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/benchislett&quot;&gt;@benchislett&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tzulingk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tzulingk&quot;&gt;@tzulingk&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gcanlin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gcanlin&quot;&gt;@gcanlin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ivanium/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ivanium&quot;&gt;@ivanium&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/divakar-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/divakar-amd&quot;&gt;@divakar-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/R3hankhan123/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/R3hankhan123&quot;&gt;@R3hankhan123&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/sagearc/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/sagearc&quot;&gt;@sagearc&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vhagor/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vhagor&quot;&gt;@vhagor&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ronensc/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ronensc&quot;&gt;@ronensc&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/micah-wil/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/micah-wil&quot;&gt;@micah-wil&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/qyYue1389/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/qyYue1389&quot;&gt;@qyYue1389&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vanshbhatia-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vanshbhatia-amd&quot;&gt;@vanshbhatia-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hao-aaron/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hao-aaron&quot;&gt;@hao-aaron&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/chengy-sysu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/chengy-sysu&quot;&gt;@chengy-sysu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/elvircrn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/elvircrn&quot;&gt;@elvircrn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/taking-lying-flat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/taking-lying-flat&quot;&gt;@taking-lying-flat&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/omerpaz95/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/omerpaz95&quot;&gt;@omerpaz95&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Etelis/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Etelis&quot;&gt;@Etelis&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vllmellm/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vllmellm&quot;&gt;@vllmellm&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/frank-suwen/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/frank-suwen&quot;&gt;@frank-suwen&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/KernelClint/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/KernelClint&quot;&gt;@KernelClint&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Fangzhou-Ai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Fangzhou-Ai&quot;&gt;@Fangzhou-Ai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/louie-tsai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/louie-tsai&quot;&gt;@louie-tsai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/simondanielsson/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/simondanielsson&quot;&gt;@simondanielsson&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/maxyanghu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/maxyanghu&quot;&gt;@maxyanghu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmai-afk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmai-afk&quot;&gt;@dmai-afk&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/KurodaKanbei/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/KurodaKanbei&quot;&gt;@KurodaKanbei&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ziqifan617/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ziqifan617&quot;&gt;@ziqifan617&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bastefaniak/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bastefaniak&quot;&gt;@bastefaniak&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ECMGit/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ECMGit&quot;&gt;@ECMGit&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/haregali/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/haregali&quot;&gt;@haregali&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cmiyai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cmiyai&quot;&gt;@cmiyai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fede-kamel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fede-kamel&quot;&gt;@fede-kamel&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/drakosha/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/drakosha&quot;&gt;@drakosha&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vineethsaivs/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vineethsaivs&quot;&gt;@vineethsaivs&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zcxGGmu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zcxGGmu&quot;&gt;@zcxGGmu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/TQCB/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/TQCB&quot;&gt;@TQCB&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/skysnow2001/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/skysnow2001&quot;&gt;@skysnow2001&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shenoyvvarun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shenoyvvarun&quot;&gt;@shenoyvvarun&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/karen-sy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/karen-sy&quot;&gt;@karen-sy&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fattchris/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fattchris&quot;&gt;@fattchris&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RyanJHamby/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RyanJHamby&quot;&gt;@RyanJHamby&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shikamd123/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shikamd123&quot;&gt;@shikamd123&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/namgyu-youn/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/namgyu-youn&quot;&gt;@namgyu-youn&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zzt93/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zzt93&quot;&gt;@zzt93&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/abmfy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/abmfy&quot;&gt;@abmfy&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/reidliu41/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/reidliu41&quot;&gt;@reidliu41&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Rapisurazurite/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Rapisurazurite&quot;&gt;@Rapisurazurite&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tandixit95/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tandixit95&quot;&gt;@tandixit95&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mganczarenko/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mganczarenko&quot;&gt;@mganczarenko&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yimdev/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yimdev&quot;&gt;@yimdev&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anujbolewar/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anujbolewar&quot;&gt;@anujbolewar&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/LiuYinfeng01/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/LiuYinfeng01&quot;&gt;@LiuYinfeng01&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lk-chen/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lk-chen&quot;&gt;@lk-chen&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/NVShreyas/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/NVShreyas&quot;&gt;@NVShreyas&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/huangzhilin-hzl/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/huangzhilin-hzl&quot;&gt;@huangzhilin-hzl&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/varoudis/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/varoudis&quot;&gt;@varoudis&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Yejing-Lai/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Yejing-Lai&quot;&gt;@Yejing-Lai&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mkhazraee/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mkhazraee&quot;&gt;@mkhazraee&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jzakrzew/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jzakrzew&quot;&gt;@jzakrzew&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/TrainToGPB/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/TrainToGPB&quot;&gt;@TrainToGPB&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/waynehacking8/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/waynehacking8&quot;&gt;@waynehacking8&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zixi-qi/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zixi-qi&quot;&gt;@zixi-qi&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Sundaresan-G/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Sundaresan-G&quot;&gt;@Sundaresan-G&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mindungil/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mindungil&quot;&gt;@mindungil&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bitborne/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bitborne&quot;&gt;@bitborne&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Wauplin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Wauplin&quot;&gt;@Wauplin&lt;/a&gt;, @jacobzhang22, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zhewenl/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zhewenl&quot;&gt;@zhewenl&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bnellnm/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bnellnm&quot;&gt;@bnellnm&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pisceskkk/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pisceskkk&quot;&gt;@pisceskkk&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Lin-z-w/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Lin-z-w&quot;&gt;@Lin-z-w&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gabriel-peracio/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gabriel-peracio&quot;&gt;@gabriel-peracio&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/SilenNaihin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/SilenNaihin&quot;&gt;@SilenNaihin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/baodii/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/baodii&quot;&gt;@baodii&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/YunzhuLu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/YunzhuLu&quot;&gt;@YunzhuLu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xwu-intel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xwu-intel&quot;&gt;@xwu-intel&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/BWAAEEEK/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/BWAAEEEK&quot;&gt;@BWAAEEEK&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/thisjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/thisjiang&quot;&gt;@thisjiang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/maobaolong/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/maobaolong&quot;&gt;@maobaolong&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/anhtra3889/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/anhtra3889&quot;&gt;@anhtra3889&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/JaredforReal/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/JaredforReal&quot;&gt;@JaredforReal&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lvhan028/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lvhan028&quot;&gt;@lvhan028&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xiaolong-intel/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xiaolong-intel&quot;&gt;@xiaolong-intel&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/andyxning/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/andyxning&quot;&gt;@andyxning&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/cleonard530/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/cleonard530&quot;&gt;@cleonard530&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gnovack/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gnovack&quot;&gt;@gnovack&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/MatthewBonanni/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/MatthewBonanni&quot;&gt;@MatthewBonanni&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/wangxian001/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/wangxian001&quot;&gt;@wangxian001&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/lengrongfu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/lengrongfu&quot;&gt;@lengrongfu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Tejas-Raj01/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Tejas-Raj01&quot;&gt;@Tejas-Raj01&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/simon-mo/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/simon-mo&quot;&gt;@simon-mo&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/vitamin-chaos/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/vitamin-chaos&quot;&gt;@vitamin-chaos&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/arpera/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/arpera&quot;&gt;@arpera&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jairitAge/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jairitAge&quot;&gt;@jairitAge&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jimmy-adams/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jimmy-adams&quot;&gt;@jimmy-adams&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ILikeIneine/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ILikeIneine&quot;&gt;@ILikeIneine&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/woosebastian/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/woosebastian&quot;&gt;@woosebastian&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/haic0/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/haic0&quot;&gt;@haic0&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/edwinlim0919/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/edwinlim0919&quot;&gt;@edwinlim0919&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fcui-amd/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fcui-amd&quot;&gt;@fcui-amd&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jhu960213/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jhu960213&quot;&gt;@jhu960213&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jinzhen-lin/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jinzhen-lin&quot;&gt;@jinzhen-lin&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/coltonottley/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/coltonottley&quot;&gt;@coltonottley&lt;/a&gt;, @walterbm, @meenchen, @matteso1, @djramic, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/gchinora/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/gchinora&quot;&gt;@gchinora&lt;/a&gt;, @davidjpyu, @tianmu-li, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xiaopusun/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xiaopusun&quot;&gt;@xiaopusun&lt;/a&gt;, @majunze2001, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Vegetog/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Vegetog&quot;&gt;@Vegetog&lt;/a&gt;, @puririshi98, @janeyx99, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/RobbieJ/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/RobbieJ&quot;&gt;@RobbieJ&lt;/a&gt;, @oonyshch, @thegoldenflow, @Srinivasoo7, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fatday/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fatday&quot;&gt;@fatday&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/acheamponge/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/acheamponge&quot;&gt;@acheamponge&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/efschu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/efschu&quot;&gt;@efschu&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rajfirke/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rajfirke&quot;&gt;@rajfirke&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fanxingran/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fanxingran&quot;&gt;@fanxingran&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xudonlyu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xudonlyu&quot;&gt;@xudonlyu&lt;/a&gt;, @lcskrishna, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/xijiaat/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/xijiaat&quot;&gt;@xijiaat&lt;/a&gt;, @GirasoleY, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/d4l3k/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/d4l3k&quot;&gt;@d4l3k&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/samuelkim7/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/samuelkim7&quot;&gt;@samuelkim7&lt;/a&gt;, @tarukumar, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/acmore/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/acmore&quot;&gt;@acmore&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/theminghuang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/theminghuang&quot;&gt;@theminghuang&lt;/a&gt;, @khushali9, @wzhao18, @Priyjain-amd, @yiz-liu, @lkm2835, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/dmholtz/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/dmholtz&quot;&gt;@dmholtz&lt;/a&gt;, @Dao007forever, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/liushujia122/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/liushujia122&quot;&gt;@liushujia122&lt;/a&gt;, @LopezCastroRoberto, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/UgaTheDev/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/UgaTheDev&quot;&gt;@UgaTheDev&lt;/a&gt;, @tuukkjs, @aditi-amd, @guan404ming, @yiliu30, @zou3519, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Luosuu/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Luosuu&quot;&gt;@Luosuu&lt;/a&gt;, @JoursBleu, @varun-sundar-rabindranath, @mpashkovskii, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yu-xin-c/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yu-xin-c&quot;&gt;@yu-xin-c&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/WillZZZy/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/WillZZZy&quot;&gt;@WillZZZy&lt;/a&gt;, @vrdn-23, @xyang16, @ccrhx4, @tanpinsiang, @russellb, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fxfxfxfxfxfxfxfx/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fxfxfxfxfxfxfxfx&quot;&gt;@fxfxfxfxfxfxfxfx&lt;/a&gt;, @afriedri, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/yifjiang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/yifjiang&quot;&gt;@yifjiang&lt;/a&gt;, @Akashcodes732, @HF-001, @ovidiusm, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/arthurgao2003/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/arthurgao2003&quot;&gt;@arthurgao2003&lt;/a&gt;, @TomerBN-Nvidia, @hotTea123, @vx120, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/bohnstingl/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/bohnstingl&quot;&gt;@bohnstingl&lt;/a&gt;, @qwerqwerqwe8688-jpg, @jasonozuzu-cohere, @vineetatiwari27, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ruirui6946/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ruirui6946&quot;&gt;@ruirui6946&lt;/a&gt;, @linitra24, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/syedalijaseem/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/syedalijaseem&quot;&gt;@syedalijaseem&lt;/a&gt;, @nickus, @yzong-rh, @s3woz, @jhaotingc, @lukealonso, @Jie-Fang, @kzwrime, @xianbaoqian, @velonica0, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/ccaadaro/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/ccaadaro&quot;&gt;@ccaadaro&lt;/a&gt;, @yisustc, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/fangchenli/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/fangchenli&quot;&gt;@fangchenli&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/iwannagotobed/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/iwannagotobed&quot;&gt;@iwannagotobed&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/zobinHuang/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/zobinHuang&quot;&gt;@zobinHuang&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/rchalamala/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/rchalamala&quot;&gt;@rchalamala&lt;/a&gt;, @shanjiaz, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jamesETsmith/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jamesETsmith&quot;&gt;@jamesETsmith&lt;/a&gt;, @stacyroberts, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/guanxingithub/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/guanxingithub&quot;&gt;@guanxingithub&lt;/a&gt;, @biswapanda, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/shanewidanagama/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/shanewidanagama&quot;&gt;@shanewidanagama&lt;/a&gt;, @UranusSeven, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/hsusul/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/hsusul&quot;&gt;@hsusul&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/tobymao/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/tobymao&quot;&gt;@tobymao&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/mispa-ms/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/mispa-ms&quot;&gt;@mispa-ms&lt;/a&gt;, @jeffreywang88, @SayHelloToWorld, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/jyan-R/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/jyan-R&quot;&gt;@jyan-R&lt;/a&gt;, @oops-oom, @shantipriya-amd, @andakai, @akii96, @shen-shanshan, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/Kaif10/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/Kaif10&quot;&gt;@Kaif10&lt;/a&gt;, @yitingdc, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/positive666/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/positive666&quot;&gt;@positive666&lt;/a&gt;, &lt;a class=&quot;user-mention notranslate&quot; data-hovercard-type=&quot;user&quot; data-hovercard-url=&quot;/users/pavelzak/hovercard&quot; data-octo-click=&quot;hovercard-link-click&quot; data-octo-dimensions=&quot;link_type:self&quot; href=&quot;https://github.com/pavelzak&quot;&gt;@pavelzak&lt;/a&gt;, @SubSir, @ywang96&lt;/p&gt;</content>620    <author>621      <name>khluu</name>622    </author>623    <media:thumbnail height="30" width="30" url="https://avatars.githubusercontent.com/u/51931015?s=60&amp;v=4"/>624  </entry>625</feed>626