chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

[Feature][NPU]: add runtime support for GPTQ-quantized MoE models (#16364)

Browse Source 
Co-authored-by: GuoYechang <52730608+GuoYechang@users.noreply.github.com>
Co-authored-by: root <root@localhost.localdomain>
This commit is contained in:
YeChang Guo
2026-03-04 16:02:19 +03:00
committed by GitHub
co-authored by GuoYechang root
parent c2b66d320d
commit 6910c1b281
4 changed files with 370 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/layers/linear.py
+2
View File
@@ -64,6 +64,8 @@ WEIGHT_LOADER_V2_SUPPORTED = [
"GPTQMarlin24LinearMethod",
"TPUInt8LinearMethod",
"GPTQLinearMethod",
"GPTQLinearAscendMethod",
"GPTQMoEAscendMethod",
"FBGEMMFp8LinearMethod",
"GPTQLinearAscendMethod",
"ModelOptFp8LinearMethod",