biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[gpt-oss] tool parser supports for /chat/completions [1/n] (#22386)

Browse Source 
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
Co-authored-by: Simon Mo <simon.mo@hey.com>
This commit is contained in:
Aaron Pham
2025-09-04 23:39:12 -04:00
committed by GitHub
parent 65e038931d
commit c29fb540ff
8 changed files with 573 additions and 63 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/model_executor/models/config.py
View File

@@ -256,7 +256,7 @@ class GptOssForCausalLMConfig(VerifyAndUpdateConfig):
def verify_and_update_config(vllm_config: "VllmConfig") -> None:
decoding_config = vllm_config.decoding_config
if decoding_config.reasoning_backend == "":
decoding_config.reasoning_backend = "GptOss"
decoding_config.reasoning_backend = "openai_gptoss"
# Increase the max capture size from 512 to 1024 for performance.
# NOTE(woosuk): This will increase the number of CUDA graphs