Skip to content

tum-vision/dca

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

This repository contains the source code for the NeurIPS 2022 paper "Deep Combinatorial Aggregation" by Yuesong Shen and Daniel Cremers.

dca

[paper] [Arxiv link]

Usage

Dependencies

This implementation has been tested with

  • python 3.9.7
  • pytorch 1.10.0 (cuda 11.3)
  • torchvision 0.11.1

It should work with Python >= 3.6 and latest pytorch / torchvision.

Setup

  1. Prepare a python environment with the above dependencies.

  2. Install the dca package locally by running

pip install -e .

in this folder.

  1. Checkout the Python scripts in folder "experiments" to run the respective experiments:
  • cifar10: In-domain CIFAR-10 experiments;
  • svhn: In-domain SVHN experiments;
  • cifar10c: Distributional shift experiments on CIFAR-10-C;
  • ood: Out of distribution experiments;

Both the CIFAR-10-C and the OOD experiments require pre-trained models on CIFAR-10.

Citation

@InProceedings{shen2022dca,
title={Deep Combinatorial Aggregation},
author={Yuesong Shen and Daniel Cremers},
booktitle = {NeurIPS},
year = {2022}
}

About

Source code for the NeurIPS 2022 paper "Deep Combinatorial Aggregation"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages