基于vllm的千问3/2.5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-Instruct
vllm serve Qwen2.5-14B-Instruct --port 8000 --tensor-parallel-size 4 --max-num-seqs 4 

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{
    "model": "Qwen3-0.6B",
    "messages": [
      {"role": "user", "content": "请介绍一下昇腾。"}
    ],
    "temperature": 0.7
  }'

import requests
url = "http://localhost:8000/v1/chat/completions"
headers = {"Content-Type": "application/json"}
data = {
    "model": "Qwen3-0.6B",
    "messages": [
        {"role": "user", "content": "请介绍一下昇腾。"}
    ],
    "temperature": 0.7,
}
response = requests.post(url, headers=headers, json=data)
print(response.json())

cd /data01/downloadModel/Qwen3/examples/demo

python cli_demo.py