Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #24534
Triggered via pull request
January 30, 2024 18:20
chenfucn
Status
Cancelled
Total duration
16m 56s
Artifacts
–
Annotations
2 errors
|
Onnxruntime-TVM
Canceling since a higher priority waiting request for 'Linux_CI-cfu_kernel' exists
|
|
Onnxruntime-TVM
The operation was canceled.
|