search

pytorch / audio

Data manipulation and transformation for audio signal processing, powered by PyTorch
https://pytorch.org/audio
BSD 2-Clause "Simplified" License
2.54k stars 652 forks source link

how do i train the model for different language? #2118

Open FarhadGazi opened 2 years ago

FarhadGazi commented 2 years ago

hello there, I want to run this tts with Bangla language. I have the dataset on my google drive . Can you tell me if I run your model in colab where should I update?

mthrok commented 2 years ago

Hi @FarhadGazi

Sorry for the late response. To train a TTS model like the Tacotron2 example, then you need to train multiple models. One for grapheme to phoneme (G2P) models, one for embedding to spectrogram (the actual Tacotron2 part).

For training tacotron2 please refer to https://github.com/pytorch/audio/tree/main/examples/pipeline_tacotron2. For G2P models, currently torchaudio does not have a model nor training script, so please consult the dedicated projects, like deep-phonemizer https://github.com/as-ideas/DeepPhonemizer.

Alternatively, https://github.com/coqui-ai/TTS is a project specialized for TTS and very active, so you might get a better help there.

FarhadGazi commented 2 years ago

thank you for your kind response.

On Tue, Jan 18, 2022 at 11:03 PM moto @.***> wrote:

Hi @FarhadGazi https://github.com/FarhadGazi

Sorry for the late response. To train a TTS model like the Tacotron2 example, then you need to train multiple models. One for grapheme to phoneme (G2P) models, one for embedding to spectrogram (the actual Tacotron2 part).

For training tacotron2 please refer to https://github.com/pytorch/audio/tree/main/examples/pipeline_tacotron2. For G2P models, currently torchaudio does not have a model nor training script, so please consult the dedicated projects, like deep-phonemizer https://github.com/as-ideas/DeepPhonemizer.

Alternatively, https://github.com/coqui-ai/TTS is a project specialized for TTS and very active, so you might get a better help there.

— Reply to this email directly, view it on GitHub https://github.com/pytorch/audio/issues/2118#issuecomment-1015623203, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANUU7W5VVINZJ4EYUDAKRRDUWWMO7ANCNFSM5LAHOLYA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you were mentioned.Message ID: @.***>