young-jo-choi/23clwithoutrm

Code Implemetation

This document provides instructions for setting up and running the CIFAR-100 dataset implementation using the provided main.py script.

Requirements

The following packages are required to run the script:

torch==1.12.0+cu102
torchvision==0.13.0+cu102
timm==0.9.2
continuum==1.2.4
easydict
yaml

Implementation

To run the script with the default settings, use the following command:

python main.py --initial-classes=0 --incremental-classes=10 --gpu=0

--initial-classes: Number of initial classes in the dataset.
--incremental-classes: Number of classes to add incrementally.
--gpu: GPU id to use for training. Default is 0.

Inference

If you want to see the results, check the acc.ipynb.

Note that this code is heavily based on the great codebase of DyTox.

@inproceedings{douillard2021dytox,
  title     = {DyTox: Transformers for Continual Learning with DYnamic TOken eXpansion},
  author    = {Douillard, Arthur and Ram\'e, Alexandre and Couairon, Guillaume and Cord, Matthieu},
  booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2022}
}