search

as-ideas / ForwardTacotron

⏩ Generating speech in a single forward pass without any attention!
https://as-ideas.github.io/ForwardTacotron/
MIT License
578 stars 113 forks source link

Advice on prepping datasets other than LJspeech? #6

Closed scripples closed 4 years ago

scripples commented 4 years ago

Hi, I'm trying to prep my own dataset to train on the ForwardTacotron model--could you give any insight as to what train_tacotron.py or train_forward.py is expecting in terms of training data organization? Like, the old NVIDIA TT2 repo expects two text files formatted in a certain way and a path to the WAV files in the arguments. Is there something similar for this repo?

cschaefer26 commented 4 years ago

Hey, just prepare your dataset in the LJSpeech format:

|- dataset_folder/
|   |- metadata.csv
|   |- wav/
|       |- file1.wav
|       |- ...

If the language differs from English, make sure you set the correct language in the hparams.py file:

language = 'fr'
tts_cleaner_name = 'basic_cleaners'

Then just follow the steps from the README with preprocessing the folder, everything should be done automatically including splitting of the dataset into train/val etc.

I updated the README to be clearer on this. Best of luck!

scripples commented 4 years ago

Thanks for your help! I'm looking forward to giving it a try.

ghost commented 4 years ago

Where find list is supported languages?

prajwaljpj commented 3 years ago

@paklau99988 you can find the list of languages from here