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

[fix] Fix mxfp4 weight loading bug with TP sharding in GPT-OSS (#9433)

Browse Source 
Signed-off-by: Hao Lu <14827759+hlu1@users.noreply.github.com>
Signed-off-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
Co-authored-by: Xinyuan Tong <xinyuantong.cs@gmail.com>
This commit is contained in:
hlu1
2025-08-21 03:50:51 -07:00
committed by GitHub
parent e85cb1ce9d
commit dae9a80f43
2 changed files with 11 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
python/sglang/srt/entrypoints/openai/protocol.py
View File

@@ -737,8 +737,8 @@ class ResponsesRequest(BaseModel):
else:
max_tokens = default_max_tokens
# Avoid exceed the context length by minus 1 token
max_tokens -= 1
# Avoid exceed the context length by minus 2 token
max_tokens -= 2
# Get parameters with defaults
temperature = self.temperature