soobinseo / tacotron-pytorch Goto Github PK

Hi soobinseo,
I am new to TTS. I found that there were two versions of tacotron implemented by tensorflow and pytorch. I think you have already tested your pytorch version. Would you please share some speech wav
file results synthesised by this project?

Hi I'm running into the below error - has anyone else encountered this and know a fix?

File ".../tacotron/text/init.py", line 4, in
from Tacotron.text import cleaners
ModuleNotFoundError: No module named 'Tacotron'

Many thanks in advance.

How would you plot the spectrogram?
I tried to create an image from inside of inv_spectrogram(spectrogram) definition after the # Convert back to linear line but still it is black image generated of (1024, 1000) dimension.
I am looking for such methodology because I want to convert an image [of spectrogram] back into audio wav file.
Can you please help for that?

how to import inflect?

Because my computer training speed is too slow, I want to seek the parameters of training and continue training. thank you

Such a wonderful work, I learned a lot from it. thanks @soobinseo
I have encountered a problem when running the program. My GPU has 12G memory, but when the batch size larger than one, out of memory error will occur.
I tried to use inplace=True for all relu activation function, but it seems useless. all the remaining hyperparameter stay the same with the origin provided.
Any suggestions for this problem, Thanks!

Hello!When I run the code, I meet with this error (0,)ValueError: operands could not be broadcast together with shapes (1,1024) (0,). Could you help me solve it?

Hi there, thx for this repo, really !! Do you mind to open source your repo with a corressponding LICENSE ?