Model Card for VQ-BeT/PushT

VQ-BeT (as per Behavior Generation with Latent Actions) trained for the PushT environment from gym-pusht.

How to Get Started with the Model

See the LeRobot library (particularly the evaluation script) for instructions on how to load and evaluate this model.

The model was trained using this command:

python lerobot/scripts/train.py \
  policy=vqbet \
  env=pusht dataset_repo_id=lerobot/pusht \
  wandb.enable=true \
  device=cuda

This took about 7 hours to train on an Nvida A6000.

	Number of Parameters
RGB Encoder	11.2M
Remaining VQ-BeT Parts	26.3M

The model was evaluated on the PushT environment from gym-pusht. There are two evaluation metrics on a per-episode basis:

Maximum overlap with target (seen as eval/avg_max_reward in the charts above). This ranges in [0, 1].
Success: whether or not the maximum overlap is at least 95%.

Here are the metrics for 500 episodes worth of evaluation.

	Ours
Average max. overlap ratio for 500 episodes	0.887
Success rate for 500 episodes (%)	66.0

The results of each of the individual rollouts may be found in eval_info.json.