RightAiknow/act_1arm_2cam_banana_pick_v1

Model Card for ACT/BananaPick

Action Chunking Transformer Policy (as per Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware) trained for banana pick-and-place on a 1-arm SO-101 robot.

How to Get Started with the Model

See the IB-Robot project (particularly the inference_service) for instructions on how to load and deploy this model with ROS 2.

To load the model directly in Python:

from lerobot.common.policies.act.modeling_act import ACTPolicy

policy = ACTPolicy.from_pretrained("RightAiknow/act_1arm_2cam_banana_pick_v1")

Training Details

• Policy: ACT (Action Chunking with Transformers)
• Dataset: RightAiknow/1arm_2cam_banana_pick_v1
• Robot: 1-arm SO-101
• Cameras: top, wrist (480x640)
• Training steps: 500,000
• Checkpoint: step_364000
• Vision backbone: ResNet18
• Chunk size: 100
• Batch size: 16

Citation

@software{ib_robot,
  title = {IB-Robot: Intelligence Boom Robot},
  url = {https://gitcode.com/openeuler/IB_Robot},
  license = {Apache-2.0}
}