DreamMachines/actuator_unboxing_20h_345_dp

actuator_unboxing/dp_20h_345

Diffusion Policy — visuomotor control as a generative diffusion process. See the paper.

Trained with LeDream (a fork of LeRobot).

Training

• Dataset: DreamMachines/actuator_unboxing_20h_345
• Step (this revision): 200000
• Batch size: 64
• Total steps target: 200000
• Optimizer: adam (lr 0.0001)

Available revisions

Every training checkpoint is uploaded as its own git revision; the latest is also on main. Load a specific step with:

from lerobot.policies.factory import make_policy
from lerobot.configs.policies import PreTrainedConfig

cfg = PreTrainedConfig.from_pretrained("DreamMachines/actuator_unboxing_20h_345_dp", revision="v002000")
policy = make_policy(cfg)

Available: v020000, v040000, v060000, v080000, v100000, v120000, v140000, v160000, v180000, v200000

Quick eval on a real DK1

make eval MODEL=DreamMachines/actuator_unboxing_20h_345_dp

See docs/source/policy_pi05_README.md for π₀.₅-specific notes.