[Feature request] Server api to return available model names and speaker id; load and unload downloaded model

idiap / coqui-ai-TTS

🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production

https://coqui-tts.readthedocs.io

Mozilla Public License 2.0

211 stars 13 forks source link

[Feature request] Server api to return available model names and speaker id; load and unload downloaded model #37

Open chigkim opened 1 month ago

chigkim commented 1 month ago

Right now, you can request server to return audio by fetching

http://localhost:5002/api/tts?text={text}&speaker_id={speaker}

Can we also have api points for the server to return available model names and speaker ids (for multi speakers) as well as load and unload downloaded model?

Thanks for your consideration!

eginhard commented 1 month ago

Seems sensible, at least listing models/languages/speakers should be straightforward. I won't implement this myself since we don't use this server, but would merge a PR.