Vits models for Persian throw error or generates unintelligible output #3667
Unanswered
karim23657
asked this question in
General Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I use TTS v.0.21.1 .![Hugging Face Spaces](https://camo.githubusercontent.com/4edb3d4bab64d943f251cd4d010fb4e22e361648b56dcf4d03c0e32d794f742f/68747470733a2f2f63646e2e737461746963616c6c792e696f2f67682f6b6172696d32333635372f626c6f676d6174657269616c732f6d61696e2f6173736574732f68662e737667)
I trained some Persian vits models : https://github.com/karim23657/Persian-tts-coqui
And here's Hugging Face demo for them:
The Hugging Face demo works very well without any errors. However, when I test them using the TTS Python API with sentences containing punctuation, it generates unintelligible output. Here is a Colab notebook with the output audio files. :![Open In Colab](https://camo.githubusercontent.com/f5e0d0538a9c2972b5d413e0ace04cecd8efd828d133133933dfffec282a4e1b/68747470733a2f2f636f6c61622e72657365617263682e676f6f676c652e636f6d2f6173736574732f636f6c61622d62616467652e737667)
in the notebook also i see a new error
AttributeError: 'TTS' object has no attribute 'is_multi_lingual'
And also I tested it on windows with bellow code:
Output:
sp.mp4
Similar issue here : karim23657/Persian-tts-coqui#36
Beta Was this translation helpful? Give feedback.
All reactions