Pytorch implementation of Deepmind's WaveRNN model from Efficient Neural Audio Synthesis
the code from https://github.com/fatchord/WaveRNN
It's early days so there are just a couple of toy experiments online for now. The next couple of notebooks will be mainly concerned with speeding up the training (currently @ 0.5 batches/second) and implementing local conditioning.
Disclaimer I do not represent or work for Deepmind/Google.