Source code of the paper Meta-Learning with Fewer Tasks through Task Interpolation
If you find this repository useful in your research, please cite the following paper:
@inproceedings{yao2022meta,
title={Meta-Learning with Fewer Tasks through Task Interpolation},
author={Yao, Huaxiu and Zhang, Linjun and Finn, Chelsea},
booktitle={Proceeding of the 10th International Conference on Learning Representations},
year={2022}
}
We have put the related datasets in Google Drive
- python 3.*
- Pytorch 1.10+
- Bio folder is for NCI and Metabolism
- Image_NLS is for miniImagenet-S, ISIC and DermNet-S
- Please check run.sh in each folder for running scripts
Code of MLTI with ProtoNet is released. Code of MLTI with MAML coming soon.