Added fa/en reza-and-ibrahim model.
#22
by
mah92
- opened
Persian/English tts model, able to read both languages clearly.
Dataset by RezaHedayatfar(persian) and IbrahimWalk(english).
Trained by Ali Mahmoudi.
Special thanks to @MasoudS0ft, Amirreza Ramezani (
@brightening-eyes
), and Mr. Jafari (Khaneh Noor Iranian Basir)
mah92
changed pull request title from
Upload 6 files
to Added fa/en reza-and-ibrahim model.
mah92
changed pull request status to
closed
mah92
changed pull request status to
open
please hold the merge request. Although the model is ok when testing usin the piper itself, it fails when using sherpa-onnx. I need to check the problem.
It seems the problem persists for other piper voices models like en-lessac-medium. I have made an issue
https://github.com/k2-fsa/sherpa-onnx/issues/1637
Problem solved, needed to add meta-data for sherpa, it was not actually a problem.
Ready to merge!
I manually added the voice in both fa_IR and en_US. Thank you!
synesthesiam
changed pull request status to
closed