The successful replication of the PyTorch version of this project

[LJY-M]

thanks @andabi project

Due to the research needs, I tried to replicate the project using PyTorch, and finally successfully implemented it.

This implementation uses the Docker environment: PyTorch/PyTorch: 1.5-CUDA10.1-CUDNN7-Runtime

Probably due to the internal mechanism of TensorPack, using the native data loading and training process of PyTorch accelerated the training speed of the model to the original 9x

Welcome to discuss the PyTorch project. The project address is as follows : PPG_tacotron