Port of Phased LSTM to Pytorch

This is a quick port of Phased LSTM to Pytorch. Original is here: https://github.com/dannyneil/public_plstm

Rule of Thumb

In general, if you are using ~1000 timesteps or more in your input sequence, or you have asynchronously sampled data, you can benefit from PLSTM.

To run

To run, simply execute:

python main.py

This will train an LSTM and Phased LSTM on the asynchronous sin wave classification task.

Requirements

Requires Pytorch. Currently tested on torch==1.8.0.