CUDA-equivalent tensor-core acceleration for Apple Silicon. C-ABI kernel library wrapping simdgroup_matrix (M1+) and mpp::tensor_ops (M5+): GEMM, FlashAttention, Conv2D, Q4_0/Q8_0 quantized inference, GGUF reader, full transformer training kernels. One binary, M1 → M5.
metal objective-c-plus-plus gemm mlx mixed-precision apple-silicon llm-inference flash-attention gguf cuda-alternative simdgroup-matrix tensor-ops flashattention2
-
Updated
Aug 7, 2026 - Python