[Feature] Implement TTS

[TheAlex25]

Feature Request

Add an optional text-to-speech (TTS) toggle switch to the GPT4All front-end, allowing users to enable TTS on LLM outputs. This feature would utilize the Dia TTS model, a 1.6B parameter open-weights model created by Nari Labs under the Apache 2.0 license.

Benefits:

**Enhanced user experience:** Users can choose to listen to LLM outputs instead of reading them.
**Accessibility:** TTS functionality can assist users with visual impairments or reading difficulties.
**Convenience:** Users can multitask while receiving audio feedback from the LLM.

Implementation:
Integrate the Dia TTS model into the GPT4All front-end.
Add a toggle switch in the settings or output section to enable/disable TTS.
When TTS is enabled, the LLM output would be passed to the Dia TTS model for audio generation.

Use Case:
Users can enable TTS when interacting with the LLM for tasks like:

Receiving step-by-step instructions or tutorials
Listening to generated stories or creative writing
Getting audio feedback on their inputs or queries

This feature would expand the usability and accessibility of the GPT4All front-end, providing users with more flexibility and options for interacting with LLMs.

I am not affiliated with Nari Labs, but saw this on r/LocalLLaMA.

Link: https://github.com/nari-labs/dia?tab=readme-ov-file