https://github.com/berlino/memnn

Implementation of Memory Network in pyTorch
https://github.com/berlino/memnn

Last synced: 8 days ago
JSON representation

Implementation of Memory Network in pyTorch

• Host: GitHub
• URL: https://github.com/berlino/memnn
• Owner: berlino
• Created: 2017-05-23T01:27:00.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2020-10-30T15:21:57.000Z (about 4 years ago)
• Last Synced: 2023-04-08T06:47:36.809Z (over 1 year ago)
• Language: Python
• Homepage:
• Size: 11.7 KB
• Stars: 35
• Watchers: 3
• Forks: 9
• Open Issues: 3
• Metadata Files:
  • Readme: readme.md

Awesome Lists containing this project

README

        
# Implementation of Memory Network in pyTorch

This repo including implementations of End-to-End Memory Network and Key-Value Memory Network

Dataset used here is WikiMovie. Note that the preprocessing scripts are borrowed from [original repo](https://github.com/facebook/MemNN/tree/master/KVmemnn)

# Code (mainly in pytorch)

* config.py: some global configurations for data dir and training detail

* gen\_dict.py: generate dict

* data\_loader.py: load preprocessed data

* model.py:  model for training similarities and readers

* train\_\*.py: training script for experimental model

* plot.py: plot the graph for report

# Run

* run setup\_processed\_data.sh to get the processed data

* run gen\_dict.py to generate the dictionary file

* run gen\_sim\_data.py to generate the data for training similarity function

* run the train\_mlp.py to train the similarity function

* run gen\_reader\_data.py to generate the data for reader

* run train\_lstm.py, train\_memory\_network.py, train\_kv\_mm.py to train the corresponding model