FastSpeech2

To reproduce best model you should run

git clone https://github.com/ivan7022/FastSpeech2.git

cd FastSpeech2

./setup.sh

After that you can sequentially run reproduce.ipynb cells to train model.

All experiment logs and speech synthesys examples are here