Hugging Face's logo

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

Sign Up
Back to feed
PHBJT 
posted an update Sep 17
Post
1712
Bringing Open-Source Text-to-Speech to French! 🗣️🇫🇷

Hugging Face's Parler TTS mini can now speak French! 🇫🇷🎉
You can try it here: PHBJT/french_parler_tts

Key highlights:
Transform the English TTS model to speak French 🇬🇧➡️🇫🇷
Fully open source (code, weights, and datasets) 🛠️
It can be replicated for every language 🌍

Read more about it in this article: https://huggingface.co/blog/PHBJT/french-parler-tts

Special thanks to FlexAI and their dedicated team for providing the computing power that made this possible and of course to all of the Parler TTS community 🤗
aidystark
•
edited Sep 22

Hi.. this is amazing thanks for this
I want to use some advice from you for an endangered/low resource Language (35hrs) which is not too bad for such class of Language
Is it a really stringent that it must be meet this criteria ( minimum: 100 hours of audio (up to 1000 hours for optimal results))
I don't hope for perfection for a start and how do I handle cases where my tokenizer is not captured in FlanT5. Thanks

·
PHBJT

Hi! 35 hours might work (some people have reported success with 48 hours and >50k steps, if I recall correctly), but you will probably not get very good results.
Regarding the tokenizer and phonemizer, it would help if you could tell me which language you are targeting.

In this post