diffusion-openarm-pick-and-place-ikea-red-block

Diffusion policy trained on the pick-and-place IKEA red block task (single right arm). 100k steps, horizon 24, n_action_steps 16, ResNet-18 vision backbone, absolute joint-space actions.

Trained on openarm-pick-and-place-ikea-red-block.

Summary

Field	Value
Policy	diffusion
Robot	openarm
Task	pick-and-place
State / Action dims	8 / 8
Extra state inputs	—
Cameras	chest, waist, wrist_r (3)

Architecture

Field	Value
n_obs_steps	2
Horizon	24
n_action_steps	16
Vision backbone	resnet18
Action normalization	MIN_MAX

Training

Field	Value
Source dataset	openarm-pick-and-place-ikea-red-block
Total episodes	250
Episodes used (train)	200
Splits	200 / 25 / 25 (train / val / test)
Training steps	100,000
Batch size	16
Optimizer	adam, lr=1e-04
Image augmentation	no
W&B	placing-one-block-r4-fisheye / w5i6jb4t

Dataset

Field	Value
Source dataset	openarm-pick-and-place-ikea-red-block
Total episodes	250
Train / val / test	200 / 25 / 25 (train / val / test)
Trained with	anvil-embodied-ai

Task

Task description not available.

License

apache-2.0