Not Just Object, But State: Compositional Incremental Learning without Forgetting (NeurIPS 2024)

Hey there!

This is PyTorch code for the NeurIPS 2024 paper:
Not Just Object, But State: Compositional Incremental Learning without Forgetting
Yanyi Zhang, Binglin Qiu, Qi Jia, Yu Liu, Ran He
NeurIPS 2024, the Thirty-Eighth Annual Conference on Neural Information Processing Systems
[arXiv]

Environment

The system I used and tested in

• Ubuntu 20.04.4 LTS
• Slurm 21.08.1
• NVIDIA GeForce RTX 3090
• Python 3.8

Usage

First, clone the repository locally:

git clone https://github.com/Yanyi-Zhang/CompILer
cd CompILer

Then, install the packages below:

pytorch==1.12.1
torchvision==0.13.1
timm==0.6.7
pillow==9.2.0
matplotlib==3.5.3

Data preparation

The propsoed datasets Split-Clothing and Split-UT-Zappos can be download from here and here respectively.

Instructions on running CompILer

bash run.sh

The model will be trained on 5 tasks Split-UT-Zappos, 10 tasks Split-UT-Zappos, and Split-Clothing sequentially. After each training, inference will be conducted automatically.

Citation

If you found our work useful for your research, please cite our work:

@INPROCEEDINGS{Yanyi_2024_NeurIPS,
  author={Zhang, Yanyi and Qiu, Binglin and Jia, Qi and Liu, Yu and He, Ran},
  booktitle={Annual Conference on Neural Information Processing Systems (NeurIPS)}, 
  title={Not Just Object, But State: Compositional Incremental Learning without Forgetting}, 
  year={2024}
  }

Feel free to contact us: yanyi.zhang{at}mail{dot}dlut{dot}edu{dot}cn

Acknowledgment

We thank the following repos providing helpful components/functions in our work.

• L2P
• CODA-Prompt
• CGE