chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Support NVFP4 quantized dense models on AMD CDNA2/CDNA3 GPUs (#7302)

Browse Source 
Co-authored-by: HAI <hixiao@gmail.com>
Co-authored-by: Sai Enduri <saimanas.enduri@amd.com>
This commit is contained in:
Haohui Mai
2025-07-18 19:59:39 -07:00
committed by GitHub
parent 3964b352c3
commit d918ab7985
7 changed files with 361 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
python/sglang/srt/layers/linear.py
View File

@@ -53,6 +53,7 @@ WEIGHT_LOADER_V2_SUPPORTED = [
"ModelOptFp8LinearMethod",
"ModelOptFp4LinearMethod",
"IPEXAWQLinearMethod",
"PetitNvFp4LinearMethod",
]
_is_cpu = is_cpu()