[Bug]: No module named librosa

[w013nad]

Your current environment

The output of `python collect_env.py`

Your output of `python collect_env.py` here

🐛 Describe the bug

We're running vllm 0.7.3 latest docker container. It looks like librosa is not installed by default in the dockerfile which means that we can't do transcription(unless we install it manually)

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[DarkLight1337]

This is intentional. See the notes in https://docs.vllm.ai/en/latest/deployment/docker.html

[w013nad]

Disappointing but fair