Option for pre-processing JSUT data

NVIDIA / DeepLearningExamples

State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and performance on enterprise-grade infrastructure.

13.66k stars 3.25k forks source link

Option for pre-processing JSUT data #370

Closed Arafat4341 closed 4 years ago

Arafat4341 commented 4 years ago

Hello! I want to train this implementation of tacotron-2 on jsut data. Is it possible?

ghost commented 4 years ago

Hi @Arafat4341, we haven't tested it yet but you would need a transliteration cleaner that would transform the Japanese alphabet to ASCII. Since there are a few requests about JSUT dataset, we're planning to enhance our reposotory with the possibility to train on non-English datasets. Stay tuned!