Matching Networks for One Shot Learning

This repo provides a Pytorch implementation fo the Matching Networks for One Shot Learning paper.

Installation of pytorch

The experiments needs installing Pytorch

Data

For the Omniglot dataset the download of the dataset is automatic. For the miniImageNet you need to download the ImageNet dataset and execute the script utils.create_miniImagenet.py changing the lines:

pathImageNet = '<path_to_downloaded_ImageNet>/ILSVRC2012_img_train'
pathminiImageNet = '<path_to_save_MiniImageNet>/miniImagenet/'

And also change the main file option.py line or pass it by command line arguments:

parser.add_argument('--dataroot', type=str, default='<path_to_save_MiniImageNet>/miniImagenet/',help='path to dataset')

Installation

$ pip install -r requirements.txt
$ python mainOmniglot.py `#Code for OmniGlot`
$ python mainMiniImageNet.py `#Code for miniImageNet`

Acknowledgements

Special thanks to https://github.com/zergylord and https://github.com/AntreasAntoniou for their Matching Networks implementation. I intend to use some parts for this implementation. More details at https://github.com/zergylord/oneshot and https://github.com/AntreasAntoniou/MatchingNetworks

Cite

@inproceedings{vinyals2016matching,
  title={Matching networks for one shot learning},
  author={Vinyals, Oriol and Blundell, Charles and Lillicrap, Tim and Wierstra, Daan and others},
  booktitle={Advances in Neural Information Processing Systems},
  pages={3630--3638},
  year={2016}
}

Authors

• Albert Berenguel (@aberenguel) Webpage