`test/backends/test_openai_vllm.py` should be marked with the same heavy gpu use flag.

[planetf1]

I believe test/backends/test_openai_vllm.py should be marked with the same heavy gpu use flag.

I think the other changes make sense and appear to be working.

Originally posted by @jakelorocco in #605 (review)

[planetf1]

Given the removal of vllm in #781 -- closing

[jakelorocco]

We are keeping test_openai_vllm.py to ensure our OpenAI backend maintains compatibility with vLLM servers. But I think it's fine to close this issue either way.