How to port the model to PyTorch?
Created by: ankitmundada
Is it possible to port this model to pytorch. I am planning to use it Cuda 9 and cudnn 7, which is not supported by PaddlePaddle. I can think of following things at an upper level. What other steps should be there?
- Create a network similar to https://github.com/PaddlePaddle/DeepSpeech/blob/develop/model_utils/network.py
- Load params from param.tgz
Has anyone tried doing this before?