SPB Git

spb/forge

Public MIT

Forge — LLM training from scratch in pure C++20 + Metal on Apple Silicon.

History
NameLast commitUpdated
..
cpu Add wave 3: DeepSeek-V3-style MoE routing, all config-selected
metal Add wave 3: DeepSeek-V3-style MoE routing, all config-selected
ops.cpp Add wave 3: DeepSeek-V3-style MoE routing, all config-selected
ops.h Add wave 3: DeepSeek-V3-style MoE routing, all config-selected

Clone forge

Anonymous clone and fetch. Pushing requires the owner's personal access token.

ZIP tar.gz snapshot of 4a8c52702b3372b339b272f1207588fbb0bf9fba
Files · 4a8c52702b3372b339b272f1207588fbb0bf9fba
Menu

Theme