Ivan Vovk

Implementation of WaveGrad high-fidelity vocoder from Google Brain in PyTorch.

ivanvovk

Implementation of "Duration Informed Attention Network for Multimodal Synthesis" paper in PyTorch.

ivanvovk

Compressed version of Tacotron 2 using Tensor Train + Waveglow.

ivanvovk

Controllable Face Generation via pretrained Conditional Adversarial Latent Autoencoder (ALAE)

ivanvovk

Multilingual text processing API for cleaning, IPA word phonemization, tokenization, translating into sequence of character IDs for easy stacking with neural Text-to-Speech models.

ivanvovk