raymondyeh07 / equivariance_discovery

[AISTATS2022] Equivariance Discovery by Learned Parameter-Sharing

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Equivariance Discovery by Learned Parameter-Sharing

AISTATS 2022

Raymond A. Yeh1 , Yuan-Ting Hu, Mark Hasegawa-Johnson, Alexander G. Schwing
Toyota Technological Institute at Chicago1
University of Illinois at Urbana-Champaign

Overview

This repository contains code for Equivariance Discovery by Learned Parameter-Sharing (AISTATS 2022).

If you used this code or found it helpful, please consider citing the following paper:

@inproceedings{YehAISTATS2022,
               author = {R.~A. Yeh and Y.-T. Hu and M. Hasegawa-Johnson and A.~G. Schwing},
               title = {Equivariance Discovery by Learned Parameter-Sharing},
               booktitle = {Proc. AISTATS},
               year = {2022},
}

Setup

To install the dependencies, run the following

conda create - n discover python = 3.7
conda activate discover
conda install conda-build
cd equivariance_discovery
conda install pytorch torchvision torchaudio cudatoolkit = 10.1 - c pytorch
pip install - r requirements.txt
conda develop .

Experiments

Gaussian Vectors with Shared Means (Sec. 5.1)

The following commands run the experiments in Fig. 2, 3, 4, 5.

cd projects/GaussianSharing
python experiments/exp1.py
python experiments/exp2.py
python experiments/exp3.py
python experiments/exp4.py

Recovering Permutation Invariance (Sec. 5.2)

The following commands run the experiments in Fig. 7

cd projects/PermutationSharing
python experiments/exp1.py
python experiments/exp2.py

About

[AISTATS2022] Equivariance Discovery by Learned Parameter-Sharing

License:MIT License


Languages

Language:Python 100.0%