openEuler/IB_Robot_ACT_banana_pick_distill

Model Card for ACT/BananaPick (Distilled)

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. This checkpoint is a distilled model obtained via knowledge distillation.

Model Structure

├── pytorch_model/                # PyTorch 权重 (用于 GPU 推理)
│   ├── config.json
│   ├── model.safetensors
│   ├── train_config.json
│   ├── policy_preprocessor.json
│   ├── policy_postprocessor.json
│   ├── policy_preprocessor_step_3_normalizer_processor.safetensors
│   └── policy_postprocessor_step_0_unnormalizer_processor.safetensors

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 (weights under pytorch_model/):

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

policy = ACTPolicy.from_pretrained("openEuler/IB_Robot_ACT_banana_pick_distill", subfolder="pytorch_model")

Training Details

This model was trained via knowledge distillation (kd: true) within the IB-Robot framework.

• Policy: ACT (Action Chunking with Transformers)
• Training method: Knowledge distillation
• Robot: 1-arm SO-101
• Cameras: top, wrist (480x640)
• Action dim: 6
• Training steps: 500,000 (configured)
• Checkpoint: step_160000
• Chunk size: 100
• Batch size: 32
• Vision backbone: ResNet18
• Optimizer: AdamW (lr=1e-5, weight_decay=1e-4)
• Latent dim (VAE): 32
• Dim model: 1024

Citation

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