4090模型切换

 

 

 

 

8B:

vllm serve /home/xurui/.cache/modelscope/hub/models/Qwen/Qwen3-8B \

--served-model-name qwen3-8b \

--host 0.0.0.0 \

--port 5001 \

--enable-lora \

--lora-modules xiaohui=/home/xurui/nanwang/qwen3_8b_xiaohui fgpt=/home/xurui/nanwang/qwen3_8b_fgpt \

--max-loras 2 \

--max-lora-rank 64 \

--gpu-memory-utilization 0.9 \

--max-model-len 25600 \

--dtype auto \

--enable-auto-tool-choice \

--tool-call-parser qwen3_coder

 

4B:

vllm serve /home/xurui/.cache/modelscope/hub/models/Qwen/Qwen3-4B-Instruct-2507 \

--served-model-name qwen3-4b \

--host 0.0.0.0 \

--port 5001 \

--enable-lora \

--lora-modules xiaohui=/home/xurui/nanwang/qwen3_4b_xiaohui fgpt=/home/xurui/nanwang/qwen3_4b_fgpt \

--max-loras 2 \

--max-lora-rank 64 \

--gpu-memory-utilization 0.9 \

--max-model-len 25600 \

--dtype auto \

--enable-auto-tool-choice \

--tool-call-parser qwen3_coder