Skip to content

surajitdb/PowerlawLSTM

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 

Repository files navigation

PowerLawLSTM

Implementing power law gated LSTM (pLSTM) on copy task with Pytorch.

The code is implemented in the paper titled "Slower is Better: Revisiting the Forgetting Mechanism in LSTM for Slower Information Decay", submitted to ICML2021

dataset.py

Generating training and testing datasets for the copy task. The T for the task could be flexibly adjusted to generate corresponding sequences for the task.

experiment.py

Running the copy task experiment including the model settings, training, validation and testing procedure etc.

model.py

Including LSTM and pLSTM model classes for learning copy task.

metric.py

Metric (calculating the accuracy in this case) for evaluating the model performance on the copy task

CopyTask_run.ipynb

The main file for running the task using Jupyter notebook. Please install/import Pytorch and other required libraries along with the python files included in this zip file to run the copy task using LSTM/pLSTM.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 72.7%
  • Jupyter Notebook 27.3%