[Feature request] Language Support ("Hindi") missing in XTTS on local machine.

[ghost]

*Title

On XTTS site (i.e https://huggingface.co/coqui/XTTS-v2), Hindi language is already supported and it's working perfect on demo site. But, the problem is when I install XTTS on my local machine, error appears that 'hi' is not supported.

Language hi is not supported. Supported languages are ['en', 'es', 'fr', 'de', 'it', 'pt', 'pl', 'tr', 'ru', 'nl', 'cs', 'ar', 'zh-cn', 'hu', 'ko', 'ja']

Possible solution might be adding 'hi' index in languages. Since afaik, XTTS on local machine import the dataset as on the online demo website. Also, for fine tuning XTTS (used colab notebook from this video - https://www.youtube.com/watch?v=8tpDiiouGxc), Hindi language option isn't available as a option. It would be helpful to add support plus able to fine tune XTTS for that language.

Thanks!

[eginhard]

Thanks for spotting this. I've enabled Hindi in the fine-tuning demo and tokenizer in our fork: https://github.com/idiap/coqui-ai-TTS

[ghost]

Hi, thanks for the fork repo link, I noticed keyerror='hi' type of errors when I tried to fine tune model, however I've added some 'hi' tags/attributes in some function located in tokenizer.py and now it's works perfectly.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. You might also look our discussion channels.

[jiten634]

Hi, thanks for the fork repo link, I noticed keyerror='hi' type of errors when I tried to fine tune model, however I've added some 'hi' tags/attributes in some function located in tokenizer.py and now it's works perfectly.

Can you share the files, where you have added hindi tags/att ? because I am still stuck at there. Please help... I am facing the same issue in xtts v1.1

[ghost]

Install the Coqui TTS, then,
Download the selected tokenizer.py and put it in TTS/tts/layers/xtts folder
https://drive.google.com/file/d/1cAW8b-JTttyVLST3_PIYGJbvs4gQQMzc/view?usp=drivesdk

[jiten634]

Install the Coqui TTS, then, Download the selected tokenizer.py and put it in TTS/tts/layers/xtts folder https://drive.google.com/file/d/1cAW8b-JTttyVLST3_PIYGJbvs4gQQMzc/view?usp=drivesdk

Thanks for the link. I am also facing one more problem in XTTS V2, when I am making hindi voice, in the end of the audio file, sometimes it cut the word, or sometimes it speak a different hallucination type voice. is there any solution for this?

[ghost]

It's the known issue of XTTS V2, there's no fix for this issue because as far as I know XTTS is proprietary TTS engine(not open source).