lxy

VLLM 0.7.2 can start the model normally, but there is no output when simulating a request using Curl, it blocks!