Skip to content

HongguangZhang/arn-eccv20-master

Repository files navigation

arn-eccv20-master

The evaluation code for ECCV'20 paper "Few-shot Action Recognition via Permutation-invariant Attention"

Pipline of ARN

Requires.

pytorch-1.7.1
numpy
scipy

For miniImagenet training and testing, run following commands.

python eval_arn.py -s $SHOT -d 'hmdb51' -m 'sa_jigsaw'
-s: shot number
-d: dataset
-m: model name

Citation

If you use this code in your research, please cite the following paper.

@inproceedings{zhang2020few,
  title={Few-shot action recognition with permutation-invariant attention},
  author={Zhang, H and Zhang, L and Qi, X and Li, H and Torr, PHS and Koniusz, P},
  booktitle={Proceedings of the European Conference on Computer Vision (ECCV 2020)},
  volume={12350},
  year={2020},
  organization={Springer}
}

About

The code for ECCV'20 paper 'Few-shot Action Recognition via Permutation-invariant Attention'

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages