Skip to content

Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #26694

Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM.

Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #26694

Annotations

1 error and 10 warnings

Lint C++

failed Feb 28, 2024 in 22m 47s