Edresson/TTS-Conv

A TensorFlow Implementation of DCTTS

DCTTS is introduced in Efficiently Trainable Text-to-Speech System Based on Deep Convolutional Networks with Guided Attention.

Requirements

• NumPy >= 1.11.1
• TensorFlow >= 1.3 (Note that the API of tf.contrib.layers.layer_norm has changed since 1.3)
• librosa
• tqdm
• matplotlib
• scipy

Data

I train Portuguese models with

1. TTS-Portuguese Corpus

Training

• STEP 0. Download TTS-Portuguese Corpus or prepare your own data.
• STEP 1. Run python prepro.py.
• STEP 2. Run python train.py 1 for training Text2Mel.
• STEP 3. Run python train.py 2 for training SSRN.

You can do STEP 2 and 3 at the same time, if you have more than one gpu card.

Training Curves

Attention Plot

Sample Synthesis

I generate speech samples based on phonetically balanced sentences as the original paper does. It is already included in the repo.

• Run synthesize.py and check the files in samples.

Generated Samples

| Dataset | Samples | | :-------------| | TTS-Portuguese Corpus with Text |2115k|

| TTS-Portuguese Corpus with Phoneme |1734k|

Online TTS demo

A notebook supposed to be executed on https://colab.research.google.com is available:

• [DCTTS Text demo](https://colab.research.google.com/drive/1GwC1hp-gbuNC-_fk3Bm7k2kj6kG6SRsz

Pretrained Model for Models

TTS-Portuguese Corpus with Text Download this.

TTS-Portuguese Corpus with Phoneme Download this.

Notes

• The changes not described in the paper were inspired by the repository: dc_tts

• The paper didn't mention normalization, but without normalization I couldn't get it to work. So I added layer normalization.

  • The paper didn't mention dropouts. So I added 0.05 for all layers.

• The paper fixed the learning rate to 0.001, but it didn't work for me. So I decayed it.

• This implementation is inspired by the repository: dc_tts