swm-dmc-expert-policies

Expert policy checkpoints for DeepMind Control Suite environments, from the galilai-group/stable-worldmodel project.

Files

Each environment folder contains:

File	Description
expert_policy.zip	Stable-Baselines3 SAC policy checkpoint
vec_normalize.pkl	Observation normalization statistics

Environments included: cheetah

Usage

from stable_worldmodel.envs.dmcontrol import ExpertPolicy
policy = ExpertPolicy(
    ckpt_path="cheetah/expert_policy/expert_policy.zip",
    vec_normalize_path="cheetah/expert_policy/vec_normalize.pkl",
    device="cpu",
)

See CVlization stable_worldmodel_dmc example for a complete runnable Docker example.

Source

• Upstream repo: https://github.com/galilai-group/stable-worldmodel

Citation

If you use these artifacts, please cite the original work:

@misc{maes_lelidec2026swm-1,
      title={stable-worldmodel-v1: Reproducible World Modeling Research and Evaluation},
      author={Lucas Maes and Quentin Le Lidec and Dan Haramati and
              Nassim Massaudi and Damien Scieur and Yann LeCun and
              Randall Balestriero},
      year={2026},
      eprint={2602.08968},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2602.08968},
}