DeepSeek/DeepGEMM
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepGEMM synced 2025-06-26 23:15:49 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Unify kwargs usages

Browse Source
This commit is contained in:
Chenggang Zhao
2025-05-15 16:53:52 +08:00
parent 350989eef3
commit 3b412f458a
6 changed files with 14 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
deep_gemm/jit_kernels/wgrad_gemm.py
View File

@@ -108,9 +108,9 @@ def wgrad_gemm_fp8_fp8_fp32_nt(lhs: Tuple[torch.Tensor, torch.Tensor],
}
# Generate, build and run the kernel
code = FP8WGradGemmRuntime.generate(**kwargs)
code = FP8WGradGemmRuntime.generate(kwargs)
runtime = build('wgrad_gemm_fp8_fp8_fp32_nt', code, FP8WGradGemmRuntime, kwargs)
runtime(**kwargs)
runtime(kwargs)
def k_grouped_wgrad_gemm_fp8_fp8_fp32_nt(lhs: Tuple[torch.Tensor, torch.Tensor],