Error: "POST /generate HTTP/1.1" 404 Not Found when running Locust tool against vLLM model server

[Edwinhr716]

Whenever running the locust benchmarking tool against a vLLM model server, I get this error

POST /generate HTTP/1.1" 404 Not Found

After some investigation, it looks like it was caused by #694, which switched the entrypoint from vllm.entrypoints.api_server to vllm.entrypoints.openai.api_server. the openAI entrypoint does not have a /generate endpoint, which breaks the locust benchmarking tool.

Looks like it was called out here #694 (comment) but never addressed