This repository contains the implementation of the paper: Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics.

To train the EBM on the rotated swiss roll dataset, you can run:

python ebm/train.py --config-name=particle_ebm_swiss_roll \
    trainer.e_l_step_size=0.005 \
    trainer.g_l_step_size=0.9

Please refer to appendix D.1 of the paper for more details on hyperparameters.

To train the particle EBM on images, you can run:

data="celeba64"  # or svhn, cifar10
python ebm/train.py --config-name=particle_ebm_image \
    logging.use_wandb=true \
    image_common=$data

Please refer to appendix D.2 of the paper for more details on hyperparameters.

@article{marks2025learninglatentenergybasedmodels,
      title={Learning Latent Energy-Based Models via Interacting Particle Langevin Dynamics}, 
      author={Joanna Marks and Tim Y. J. Wang and O. Deniz Akyildiz},
      year={2025},
      journal={arXiv preprint arXiv:2510.12311}
}