LLM挂载&部署

cd /data01/downloadModel/Qwen    
#单卡  
vllm serve Qwen3-0.6B --port 8000 --enable-reasoning --reasoning-parser deepseek_r1   

#多卡跑Qwen3-0.6B  
vllm serve Qwen3-0.6B --port 8000 --max-model-len 4096 --tensor-parallel-size 2 --max-num-seqs 4 --enable-reasoning --reasoning-parser deepseek_r1   

#多卡跑Qwen2.5-14B(32B)-Instruct  
vllm serve Qwen2.5-32B-Instruct --port 8000 --tensor-parallel-size 4 --max-num-seqs 4 

#单卡跑Qwen2.5-VL-7B-Instruct 
vllm serve Qwen2.5-VL-7B-Instruct --dtype half --port 6006 --limit_mm_per_prompt image=4 --max_model_len 8784