Compute-Efficient Active Learning - Official PyTorch Implementation

Prequisites:

Tested using Python 3.8 Use requirements.txt to install necessary dependencies.

Running the Code

Example on MNIST:

python3 src/main_train.py --dataset mnist --strategy entropy --subsample-size 5000 --subsample-unlabeled

Example on CIFAR10:

python3 src/main_train.py --dataset cifar10 --strategy entropy --subsample-size 10000 --subsample-unlabeled

Citation

If you use our code in your research, or find our work helpful, please consider citing us with the bibtex below:

@inproceedings{
n{\'e}meth2023computeefficient,
title={Compute-Efficient Active Learning},
author={G{\'a}bor N{\'e}meth and Tamas Matuszka},
booktitle={NeurIPS 2023 Workshop on Adaptive Experimental Design and Active Learning in the Real World},
year={2023},
url={https://openreview.net/forum?id=G6ujG6LaKV}
}

More details about our work can be found on the paper and poster.

License

This project is licensed under the Attribution-NonCommercial-ShareAlike 4.0 International license.

Please see the LICENSE.txt file for more details about the license terms and conditions.