Forge — LLM training from scratch in pure C++20 + Metal on Apple Silicon.
| Name | Last commit | Updated |
|---|---|---|
| .. | ||
| 50m-tinystories.forge | Add wave 3: DeepSeek-V3-style MoE routing, all config-selected | |
Forge — LLM training from scratch in pure C++20 + Metal on Apple Silicon.
| Name | Last commit | Updated |
|---|---|---|
| .. | ||
| 50m-tinystories.forge | Add wave 3: DeepSeek-V3-style MoE routing, all config-selected | |