First, download and set up the repo:
We provide an environment.yml file that can be used to create a Conda environment.
conda env create -f environment.yml
conda activate DiTtorchrun --nnodes=1 --nproc_per_node=N train_label.py --model DiT-L/4 --data-path /path/to/ISIC/trainpython generation.py --model DiT-L/4 --image-size 256 --ckpt /path/to/model.pt@article{munia2025class,
title={Class-N-Diff: Classification-Induced Diffusion Model Can Make Fair Skin Cancer Diagnosis},
author={Munia, Nusrat and Imran, Abdullah},
journal={arXiv preprint arXiv:2510.16887},
year={2025}
}