https://github.com/arthurdouillard/incremental_learning.pytorch
A collection of incremental learning paper implementations including PODNet (ECCV20) and Ghost (CVPR-W21).
https://github.com/arthurdouillard/incremental_learning.pytorch
continual-learning deep-learning incremental-learning lifelong-learning pytorch research
Last synced: 10 days ago
JSON representation
A collection of incremental learning paper implementations including PODNet (ECCV20) and Ghost (CVPR-W21).
- Host: GitHub
- URL: https://github.com/arthurdouillard/incremental_learning.pytorch
- Owner: arthurdouillard
- License: mit
- Created: 2019-04-02T19:57:06.000Z (about 6 years ago)
- Default Branch: master
- Last Pushed: 2023-10-03T21:36:06.000Z (over 1 year ago)
- Last Synced: 2025-03-29T13:09:03.752Z (17 days ago)
- Topics: continual-learning, deep-learning, incremental-learning, lifelong-learning, pytorch, research
- Language: Python
- Homepage:
- Size: 6.03 MB
- Stars: 398
- Watchers: 14
- Forks: 59
- Open Issues: 8
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
- Authors: AUTHORS
Awesome Lists containing this project
- awesome-machine-learning-resources - **[Code Collection
README
# Incremental Learners for Continual Learning
Repository storing some my public works done during my PhD thesis (2019-).
You will find in there both known implementation (iCaRL, etc.) but also all my papers.
You can find the list of the latter on my [Google Scholar](https://scholar.google.com/citations?user=snwgZBIAAAAJ&hl=en).
My work on continual segmentation can be found [here](https://github.com/arthurdouillard/CVPR2021_PLOP) and on continual data loaders [here](https://github.com/Continvvm/continuum).
## Structures
Every model must inherit `inclearn.models.base.IncrementalLearner`.
# PODNet: Pooled Outputs Distillation for Small-Tasks Incremental Learning
[](https://arxiv.org/abs/2004.13513)
[](https://www.youtube.com/watch?v=SWFO1_lTcR8)
If you use this paper/code in your research, please consider citing us:
```
@inproceedings{douillard2020podnet,
title={PODNet: Pooled Outputs Distillation for Small-Tasks Incremental Learning},
author={Douillard, Arthur and Cord, Matthieu and Ollion, Charles and Robert, Thomas and Valle, Eduardo},
booktitle={Proceedings of the IEEE European Conference on Computer Vision (ECCV)},
year={2020}
}
```
To run experiments on CIFAR100 with three different class orders, with the challenging
setting of 50 steps:
```bash
python3 -minclearn --options options/podnet/podnet_cnn_cifar100.yaml options/data/cifar100_3orders.yaml \
--initial-increment 50 --increment 1 --fixed-memory \
--device --label podnet_cnn_cifar100_50steps \
--data-path
```
Likewise, for ImageNet100:
```bash
python3 -minclearn --options options/podnet/podnet_cnn_imagenet100.yaml options/data/imagenet100_1order.yaml \
--initial-increment 50 --increment 1 --fixed-memory \
--device --label podnet_cnn_imagenet100_50steps \
--data-path
```
And ImageNet1000:
Likewise, for ImageNet100:
```bash
python3 -minclearn --options options/podnet/podnet_cnn_imagenet100.yaml options/data/imagenet1000_1order.yaml \
--initial-increment 500 --increment 50 --fixed-memory --memory-size 20000 \
--device --label podnet_cnn_imagenet1000_10steps \
--data-path
```
Furthermore several options files are available to reproduce the ablations showcased
in the paper. Please see the directory `./options/podnet/ablations/`.
# Insight From the Future for Continual Learning
[](https://arxiv.org/abs/2006.13748)
If you use this paper/code in your research, please consider citing us:
```
@inproceedings{douillard2020ghost,
title={Insight From the Future for Continual Learning},
author={Arthur Douillard and Eduardo Valle and Charles Ollion and Thomas Robert and Matthieu Cord},
booktitle={arXiv preprint library},
year={2020}
}
```
The code is still very dirty, I'll clean it later. Forgive me.