EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

fix SUPPORT_CUTLASS_BLOCK_FP8 flag (#4640)

Browse Source
This commit is contained in:
Cheng Wan
2025-03-21 00:45:07 -04:00
committed by GitHub
parent ad4e58bf67
commit 7b5fc71972
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
python/sglang/srt/layers/quantization/fp8_utils.py
View File

@@ -82,7 +82,7 @@ def normalize_e4m3fn_to_e4m3fnuz(
def cutlass_block_fp8_supported() -> bool: def cutlass_block_fp8_supported() -> bool:
if get_bool_env_var("SUPPORT_CUTLASS_BLOCK_FP8"): if not get_bool_env_var("SUPPORT_CUTLASS_BLOCK_FP8"):
return False return False
if _is_cuda: if _is_cuda:
major, minor = torch.cuda.get_device_capability() major, minor = torch.cuda.get_device_capability()