SPB Git

spb/forge Public MIT

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

C++ 61.2% C 23% Python 7.6% TeX 7.2% CMake 1.1%
NameLast commitUpdated
adamw.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
attention.metal Add architecture-variant waves 1+2: train... 5 days ago
cross_entropy.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
elementwise.metal Add wave 3: DeepSeek-V3-style MoE routing, all config-selected 5 days ago
embedding.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
flash_attention_mma.metal Add architecture-variant waves 1+2: train... 5 days ago
flash_attention.metal Add architecture-variant waves 1+2: train... 5 days ago
layernorm.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
matmul_mpp.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
matmul_simd.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
matmul.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
moe.metal Add wave 3: DeepSeek-V3-style MoE routing, all config-selected 5 days ago
precision_probe.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago
quant.metal Add QAT, MoE, and architecture-variant knobs — all config-selected 5 days ago
softmax.metal Forge: LLM training from scratch in C++20 + Metal on Apple Silicon 10 days ago