Revert "Add online serving to Stable Audio Diffusion and introduce v1/audio/generate endpoint"

[hsliuustc0106]

Reverts #1255

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 5241588121

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Initialize speech handler state in pure diffusion mode

This pure-diffusion initialization path returns without defining app.state.openai_serving_speech, but Omnispeech() later does a direct attribute read (request.app.state.openai_serving_speech) before the route-level None checks run. In pure diffusion deployments, calling /v1/audio/speech or /v1/audio/voices will therefore raise AttributeError and return a 500 instead of the expected “not supported” response; set openai_serving_speech = None in this branch (or switch Omnispeech to getattr) to preserve graceful handling.

Useful? React with 👍 / 👎.

[ekagra-ranjan]

I am still in the process of seeing why it fails but I see a fix already: https://github.com/vllm-project/vllm-omni/pull/1771/changes

I think this works? The addition of _Unsupported might failed a CI but it highlighted the access of undefined variable serving_models.base_model_paths being accessed here. The fix PR adds the property here

[ekagra-ranjan]

I now understand why it happened and the fix.

The issue comes from _Unsupported being added to _DiffusionServingModels. It was added because earlier OpenAIServing accessing variables not defined in _DiffusionServingModels class with self.model_config = self.models.model_config which would fail even though its not used later on. Hence, I added _Unsupported so that such assignment dont fail during init but loudly if accessed. This failure during init still fails on vllm v0.16 but is fixed in vllm v0.17

However, there are other checks in vllm-omni that rely on hasattr() to see if _DiffusionServingModels has a variable and avoids using undefined variables. These checks fail because hasattr() calls getattr() and _Unsupported is truthy.

But now the recent vllm code assigns these variables in OpenAIServing using self.engine_client instead of self.models so the _Unsupported way can be safely removed. Will do and raise PR.