Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #18619
Azure Pipelines / iOS CI Pipeline
succeeded
Feb 29, 2024 in 1h 21m 19s
Build #20240228.27 succeeded
Loading