Official implementation of Representation Learning via Consistent Assignment of Views over Random Partitions (CARP)

Thirty-seventh Conference on Neural Information Processing Systems

Important links

• Project website
• arXiv
• NeurIPS procedings

Download pre-trained models

Make sure to download CARP's pretraining files and place them in /pretrained/carp/ folder.

Checkpoints can be downloaded here.

Performance

Method	Epochs	Multicrop	Top-1	k-NN	URL
CARP	100	2x224 + 6x96	72.5	63.5	CARP-100ep
CARP	200	2x224 + 6x96	74.2	66.5	CARP-100ep
CARP	400	2x224	73.0	67.6	CARP-400ep
CARP	400	2x224 + 6x96	75.3	67.7	CARP-400ep

Running evaluations

To run evaluations, ensure you have a proper Python environment with PyTorch 2.0 and other dependencies.

Go to specific evaluation folders (such as knn or kmeans) for examples of how to run each.

Reference

Please, cite this work as:

@inproceedings{
  Silva2023,
  title={Representation Learning via Consistent Assignment of Views over Random Partitions},
  author={Silva, Thalles and Ram\'irez Rivera, Ad\'in},
  booktitle={Thirty-seventh Conference on Neural Information Processing Systems ({NeurIPS})},
  year={2023},
  url={https://openreview.net/forum?id=fem6BIJkdv}
}