Skip to content
/ WReN Public

A Pytorch implementation of "Measuring abstract reasoning in neural networks" in ICML 2018 by DeepMind

Notifications You must be signed in to change notification settings

Fen9/WReN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Wild-Relation-Network

The repo is a PyTorch implementation of Wild Relational Network (WReN) introduced in DeepMind's Measuring abstract reasoning in neural networks (ICML 2018).

Dependencies

Important

  • PyTorch (0.4.1)
  • TensorBoardX (and Tensorboard)

See requirements.txt for other dependencies.

Usage

Run

python main.py --model <WReN/CNN_MLP/Resnet50_MLP/LSTM> --img_size <input image size> --path <path to your dataset>

Accuracy Plot

The following figure shows the WReN performance we got using the hyper-parameters in the paper.

AccPlot

About

A Pytorch implementation of "Measuring abstract reasoning in neural networks" in ICML 2018 by DeepMind

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages