Forward Transformer TTS – a non-autoregressive Transformer for text to speech • Hacker News Board

Forward Transformer TTS – a non-autoregressive Transformer for text to speech

#1

Post by datitran » Tue, Jun 02, 2020, 8:20 AM UTC

Forward Transformer TTS – a non-autoregressive Transformer for text to speech
as-ideas.github.io

Re: Forward Transformer TTS – a non-autoregressive Transformer for text to speech

#2

Post by datitran » Tue, Jun 02, 2020, 8:21 AM UTC

We released the Forward version of our Transformer TTS implementation, a text-to-speech Transformer in TensorFlow 2/Keras. Now the model is more robust, fast and controllable.

- Samples: https://as-ideas.github.io/TransformerTTS/

- Code: https://github.com/as-ideas/TransformerTTS

Re: Forward Transformer TTS – a non-autoregressive Transformer for text to speech

#3

Post by modeless » Tue, Jun 02, 2020, 11:27 PM UTC

datitran wrote: post #2

We released the Forward version of our Transformer TTS implementation, a text-to-speech Transformer in TensorFlow 2/Keras. Now the model is more robust, fast and controllable. - Samples: https://as-ideas.github.io/TransformerTTS/ - Code: https://github.com/as-ideas/TransformerTTS

How fast is it?

Re: Forward Transformer TTS – a non-autoregressive Transformer for text to speech

#4

Post by microtherion » Wed, Jun 03, 2020, 1:08 AM UTC

modeless wrote: post #3

datitran wrote: post #2

We released the Forward version of our Transformer TTS implementation, a text-to-speech Transformer in TensorFlow 2/Keras. Now the model is more robust, fast and controllable. - Samples: https://as-ideas.github.io/TransformerTTS/ - Code: https://github.com/as-ideas/TransformerTTS

How fast is it?

I cannot speak for the authors, but my impression is that the vocoder (WaveRNN) part of current TTS systems takes much more computing than the part that this research addresses, so this may not help all that much.

Re: Forward Transformer TTS – a non-autoregressive Transformer for text to speech

#5

Post by narrationbox » Wed, Jun 03, 2020, 1:48 AM UTC

datitran wrote: post #2

We released the Forward version of our Transformer TTS implementation, a text-to-speech Transformer in TensorFlow 2/Keras. Now the model is more robust, fast and controllable. - Samples: https://as-ideas.github.io/TransformerTTS/ - Code: https://github.com/as-ideas/TransformerTTS

Your work looks good. Code is very clean, well done. What hardware did you use for training? Do you see any potential improvement/application for multispeaker tasks?

Also, what does a publishing house's in-house ML team generally do? Are you building the implementations for peer review replication?