Switches to multilingual Elevenlabs #181
Conversation
Since there's no setting available that defines a model for Elevenlabs, this commit switches the default model to eleven_multilingual_v2, at the time of this writing the latest available text-to-speech model.
- Required for dependencies
- Text only for now
- Ensure max_tokens is set, as without it, the GPT-4V API calls only return clipped responses.
* fix/remove-cogent-links: Remove links to cogentapps in fork
There was a problem hiding this comment.
hei, is it here I can change the model for using only the model "Eleven Turbo v2.5" ? my voice is getting a bit danish when I used the eleven multilingual v2. Wondering if you could change the files for me and host the software also ? can offcourse pay. [email protected]
There was a problem hiding this comment.
Hi,
several notes on your message: First, I don't recommend you to pay a foreign person to host stuff like that privately - how would you know whom to trust or not. Second, this software is pretty unmaintained. If hosting (or if you find someone who does it), you'd better stick with the more maintained fork here:
https://github.com/jp-ipu/chat-with-gpt
Or one of those very far developed frontends. I know of big-agi and librechat, they have a hosted instance available, can also be self-hosted and have an ElevenLabs integration as well.
Since there's no setting available that defines a model for Elevenlabs, this commit switches the default model to eleven_multilingual_v2,
at the time of this writing the latest available text-to-speech model.