[Feature request] Text to Speach

[HyunjunA]

Name of the feature
In general, the feature you want added should be supported by HuggingFace's transformers library:

• If requesting a model, it must be listed here.
• If requesting a pipeline, it must be listed here.
• If requesting a task, it must be listed here.

Model: SpeechT5_TTS (Text-to-Speech)
The model is available here on HuggingFace's platform.

Reason for request
Why is it important that we add this feature? What is your intended use case? Remember, we are more likely to add support for models/pipelines/tasks that are popular (e.g., many downloads), or contain functionality that does not exist (e.g., new input type).

Usefulness: The SpeechT5_TTS model is designed to convert text into spoken audio. This model could serve multiple purposes across different sectors, including but not limited to education, automation, and accessibility.

Popularity: As a Microsoft model, it is backed by significant research and development, making it one of the more robust and versatile options available for text-to-speech.

New Functionality: Though text-to-speech is not a new technology, the advanced capabilities of this model could offer more natural and clear speech, which is especially valuable in applications where voice clarity and natural intonation are important.

Additional context
Add any other context or screenshots about the feature request here.

[xenova]

Hi there 👋 See #59 and #279 for existing feature requests. We are currently waiting for Optimum to support exporting speecht5 (and bark) to ONNX. Perhaps @fxmarty can provide an update?

[fxmarty]

Thank you, given the interest I could add the support this week & do a release. Which architecture are you interested in priority?

[HyunjunA]

I want to use this model: microsoft/speecht5_tts.

[aiascq]

I want to use this model:bark