EngineX-Iluvatar/enginex-vllm-bi100-qwen36
16
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Add reasoning parser mechanism + qwen3 parser + bugfixes

Browse Source
This commit is contained in:
Lu Xinlong
2026-06-10 18:22:29 +08:00
parent 4ab36b51d5
commit 365da18436
9 changed files with 3821 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
qwen3_6_scripts/patch_ops.sh
View File

@@ -58,3 +58,14 @@ python3 ./patch_xformers_sdpa_seq.py
# Use at server start: --tool-call-parser qwen3_coder --enable-auto-tool-choice
cp ./qwen3coder_tool_parser.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/openai/tool_parsers/
python3 ./patch_vllm_tool_parser.py
# --- reasoning parser: Qwen3 <think>...</think> split ------------------------
# Adds --reasoning-parser qwen3 support.
# Routes thinking tokens to reasoning_content, rest to content in the delta.
# Works together with --tool-call-parser qwen3_coder (think → tool call flow).
cp -r ./reasoning /usr/local/corex/lib/python3/dist-packages/vllm/
cp ./protocol.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/openai/protocol.py
cp ./cli_args.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/openai/cli_args.py
cp ./serving_chat.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/openai/serving_chat.py
cp ./api_server.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/openai/api_server.py
cp ./chat_utils.py /usr/local/corex/lib/python3/dist-packages/vllm/entrypoints/chat_utils.py