duobinji/vllm repository overview

Tag cpu: docker run -d --name vllm duobinji/vllm:cpu

镜像里已经包含Qwen2.5-0.5B-Instruct模型，可直接使用下面命令测试。 #查询当前运行的模型 curl http://localhost:8000/v1/models⁠

#调用当前模型推理 curl http://localhost:8000/v1/chat/completions⁠
-H "Content-Type: application/json"
-d '{ "model": "/workspace/vllm/models/Qwen2.5-0.5B-Instruct", "messages": [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "你好,你知道牛顿第二力学定律么？给我解释一下！"} ] }'