To avoid any conflict with your existing Python setup, it is suggested to work in a virtual environment with virtualenv. To install virtualenv:
pip install --upgrade virtualenvCreate a virtual environment, activate it and install the requirements in requirements.txt.
virtualenv venv
source venv/bin/activate
pip install -r requirements.txtOmniglot, miniImageNet, and tieredImageNet will be downloaded automatically upon runnning the scripts, with the help of pytorch-meta. To download meta-dataset, follow the following steps:
- Download ILSVRC2012 (by creating an account here and downloading
ILSVRC2012.tar) and Cu_birds2012 (downloading fromhttp://www.vision.caltech.edu/visipedia-data/CUB-200-2011/CUB_200_2011.tgz) separately. - Run
sbatch scripts/download_meta_dataset/install_meta_dataset_parallel.shto download and prune all datasets in a parallel fashion. Note that, due to memory constraints, we are saving ILSVRC2012 in a seperate directory from the other 9 datasets. While running the scripts to train the model, make sure to copy all the datasets to a single directory. - That's it, you are good to go! :)
For the few-shot-regression setting, we test on Sinusoid, Sinusoid & Line, and Harmonic dataset. Furthermore, these datasets are toy examples and require no downloads.
We have carefully organized our codes under scripts.
Our model can be trained on all samplers in a parallel fashion as follows:
sbatch scripts/MAML/Train/train_maml_<dataset>_all_samplers.shSimilarly, our models can be tested on a fixed set of tasks in a parallel fashion as follows:
sbatch scripts/MAML/Test/test_maml_<dataset>_all_samplers.shAll our codes were run with 1 GPUs, other than CNAPs, which has been run with 2. Furthermore, all our codes have been optimized to be run on less than 30Gb RAM, including our experiments on Meta-Dataset!
To compute results along with statistics, run:
python -m src.analysis.py <path_to_task_json> -O <path_to_output_json>If you find our codes useful, do consider citing our paper:
@misc{kumar2022effect,
title={The Effect of Diversity in Meta-Learning},
author={Ramnath Kumar and Tristan Deleu and Yoshua Bengio},
year={2022},
eprint={2201.11775},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
Our repository makes use of various open-source codes. If you find the respective codes useful, do cite their respective papers:
@article{snell2017prototypical,
title={Prototypical networks for few-shot learning},
author={Snell, Jake and Swersky, Kevin and Zemel, Richard S},
journal={arXiv preprint arXiv:1703.05175},
year={2017}
}
@article{vinyals2016matching,
title={Matching networks for one shot learning},
author={Vinyals, Oriol and Blundell, Charles and Lillicrap, Timothy and Wierstra, Daan and others},
journal={Advances in neural information processing systems},
volume={29},
pages={3630--3638},
year={2016}
}
@inproceedings{finn2017model,
title={Model-agnostic meta-learning for fast adaptation of deep networks},
author={Finn, Chelsea and Abbeel, Pieter and Levine, Sergey},
booktitle={International Conference on Machine Learning},
pages={1126--1135},
year={2017},
organization={PMLR}
}
@article{nichol2018first,
title={On first-order meta-learning algorithms},
author={Nichol, Alex and Achiam, Joshua and Schulman, John},
journal={arXiv preprint arXiv:1803.02999},
year={2018}
}
@article{requeima2019fast,
title={Fast and flexible multi-task classification using conditional neural adaptive processes},
author={Requeima, James and Gordon, Jonathan and Bronskill, John and Nowozin, Sebastian and Turner, Richard E},
journal={Advances in Neural Information Processing Systems},
volume={32},
pages={7959--7970},
year={2019}
}
@inproceedings{lee2019meta,
title={Meta-learning with differentiable convex optimization},
author={Lee, Kwonjoon and Maji, Subhransu and Ravichandran, Avinash and Soatto, Stefano},
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
pages={10657--10665},
year={2019}
}
@inproceedings{lake2011one,
title={One shot learning of simple visual concepts},
author={Lake, Brenden and Salakhutdinov, Ruslan and Gross, Jason and Tenenbaum, Joshua},
booktitle={Proceedings of the annual meeting of the cognitive science society},
volume={33},
number={33},
year={2011}
}
@article{ravi2016optimization,
title={Optimization as a model for few-shot learning},
author={Ravi, Sachin and Larochelle, Hugo},
year={2016}
}
@article{ren2018meta,
title={Meta-learning for semi-supervised few-shot classification},
author={Ren, Mengye and Triantafillou, Eleni and Ravi, Sachin and Snell, Jake and Swersky, Kevin and Tenenbaum, Joshua B and Larochelle, Hugo and Zemel, Richard S},
journal={arXiv preprint arXiv:1803.00676},
year={2018}
}
@article{triantafillou2019meta,
title={Meta-dataset: A dataset of datasets for learning to learn from few examples},
author={Triantafillou, Eleni and Zhu, Tyler and Dumoulin, Vincent and Lamblin, Pascal and Evci, Utku and Xu, Kelvin and Goroshin, Ross and Gelada, Carles and Swersky, Kevin and Manzagol, Pierre-Antoine and others},
journal={arXiv preprint arXiv:1903.03096},
year={2019}
}
@inproceedings{finn2017model,
title={Model-agnostic meta-learning for fast adaptation of deep networks},
author={Finn, Chelsea and Abbeel, Pieter and Levine, Sergey},
booktitle={International Conference on Machine Learning},
pages={1126--1135},
year={2017},
organization={PMLR}
}
@article{finn2018probabilistic,
title={Probabilistic model-agnostic meta-learning},
author={Finn, Chelsea and Xu, Kelvin and Levine, Sergey},
journal={arXiv preprint arXiv:1806.02817},
year={2018}
}
@article{lacoste2018uncertainty,
title={Uncertainty in multitask transfer learning},
author={Lacoste, Alexandre and Oreshkin, Boris and Chung, Wonchang and Boquet, Thomas and Rostamzadeh, Negar and Krueger, David},
journal={arXiv preprint arXiv:1806.07528},
year={2018}
}
@misc{deleu2019torchmeta,
title={{Torchmeta: A Meta-Learning library for PyTorch}},
author={Deleu, Tristan and W\"urfl, Tobias and Samiei, Mandana and Cohen, Joseph Paul and Bengio, Yoshua},
year={2019},
url={https://arxiv.org/abs/1909.06576},
note={Available at: https://github.com/tristandeleu/pytorch-meta}
}