Adding cuda kernel (optimized for sm80) for block-wise 4b quantized float 16 GEMM. #18619
Azure Pipelines / Windows CPU CI Pipeline (x64_release_dnnl build_x64_release)
succeeded
Feb 29, 2024 in 1h 8m 13s
x64_release_dnnl build_x64_release succeeded
Loading