TTS icon indicating copy to clipboard operation
TTS copied to clipboard

added-custom-duration-for-vits

Open p0p4k opened this issue 1 year ago β€’ 2 comments

  • added a self explanatory vits phoneme duration changer.
  • it can work for any model that uses duration predictor actually.
  • notebook shows example for a single sentence, but can be extended to multiple sentences, will add more examples and notes with detailed explanations if users request for it.

p0p4k avatar Apr 15 '23 09:04 p0p4k

@erogol hello, hope you having a good day! The checks are failing cause of premium coqui studio models, nothing else. Thank you!

p0p4k avatar Apr 17 '23 14:04 p0p4k

Hey thanks for the PR and sorry for the late response. Can you some tests for the intended usage of the change?

erogol avatar Apr 26 '23 13:04 erogol

@erogol There is no change in the main body of the code. This notebook is just kind of a tutorial for extracting durations and modifying them as well as mixing multiple speaker embeddings.

That is my another GitHub account by the way πŸ˜„

p0p4k avatar May 15 '23 01:05 p0p4k

I think this would make good content but I'd rather keep it separate. Notebooks are not ready to test and they create more annoyance than help after a while when there are compatibility issues.

Maybe you can create a nice blog post with that?

erogol avatar May 15 '23 23:05 erogol

Okay, nice blog post on its way. Soon.

p0p4k avatar May 16 '23 02:05 p0p4k

Hi @p0p4k , may I ask where is the blog? Many thanks! I'd like to implement something similar in tortoise tts, is that possible?

kexul avatar Jan 08 '24 12:01 kexul

@kexul the notebooks are self explanatory. If any doubts, shoot me a dm on discord (p0p4k)

p0p4k avatar Jan 09 '24 02:01 p0p4k