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

[Misc] [MXFP8] Drop sm100 mxfp8 warning (#21881)

Browse Source
This commit is contained in:
Ziang Li
2026-04-11 11:11:28 +00:00
committed by GitHub
parent 61a62c6503
commit 78043d4448
1 changed files with 4 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/configs/model_config.py
+4 -2
View File
@@ -1071,8 +1071,10 @@ class ModelConfig:
f"supported in ROCm." f"supported in ROCm."
) )
if self.quantization not in optimized_quantization_methods: if self.quantization not in optimized_quantization_methods:
# Don't warn for MXFP4 on SM100 since it has optimized kernels # Don't warn for MXFP4/MXFP8 on SM100 since they have optimized kernels
if not (self.quantization == "mxfp4" and is_sm100_supported()): if not (
self.quantization in ["mxfp4", "mxfp8"] and is_sm100_supported()
):
logger.warning( logger.warning(
"%s quantization is not fully " "%s quantization is not fully "
"optimized yet. The speed can be slower than " "optimized yet. The speed can be slower than "