search

r9y9 / deepvoice3_pytorch

PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models
https://r9y9.github.io/deepvoice3_pytorch/
Other
1.97k stars 485 forks source link

Which folder to give as input for training? #171

Closed harirawat closed 5 years ago

harirawat commented 5 years ago

I want to fine tune the pre-trained VCTK model for my own voice containing 400 audio samples. I ran preprocess.py which generated mel and spec files for my audio files.

I used the VCTK pre-trained model and the folder containing mel and spec files as input. The generated audio files are not good.

My question is that is it the folder containing mel and spec files that we have to give as input folder to train.py or something else.

harirawat commented 5 years ago

I gave folder containing mel files, spec files and train.txt as input folder for train.txt. The generated output is adapting my voice but nothing can be heard clearly. My train.txt looks as shown below.

spec-p225_1.npy|mel-p225_1.npy|106|C:/Users/abc/Downloads/deepvoice3_pytorch-master/deepvoice3_pytorch-master/data/out/p225_1.txt

ignore_recognition_level=2 in hparams.py.

@r9y9 @engiecat Do you have any thoughts on this?

stale[bot] commented 5 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.