search

p0p4k / vits2_pytorch

unofficial vits2-TTS implementation in pytorch
https://arxiv.org/abs/2307.16430
MIT License
465 stars 81 forks source link

ValueError: too many values to unpack (expected 2) #57

Closed KevinWang676 closed 9 months ago

KevinWang676 commented 9 months ago

Hi, I was using cjke_cleaners2 to clean the texts. However, I got the error ValueError: too many values to unpack (expected 2) as shown below

image

But I don't think I got more than 2 values to unpack. Here is the structure of my cleaned texts

image

So I wonder which step could go wrong. Thank you!

p0p4k commented 9 months ago

Can you go to the filter function and before the for statement, print out self.audiopaths_and_text ?

KevinWang676 commented 9 months ago

Thanks for your reply! I printed out self.audiopaths_and_text, but I didn't see why it went wrong. It seems that the audiopaths and text are split correctly. Thanks.

image

KevinWang676 commented 9 months ago

Sorry, this issure is due to some unexpected symbols in my texts such as {, #, and <. Thank you!

p0p4k commented 9 months ago

Good luck! Share your sound quality results after training is done! Happy holidays

KevinWang676 commented 9 months ago

Thank you. Happy holidays! I'm start the training, and I wonder if a custom dataset that contains one-hour speech would work. I think it might be a little bit small.