Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #25468
Triggered via pull request
February 28, 2024 21:01
chenfucn
Status
Success
Total duration
1h 5m 24s
Artifacts
–