dp-cupbasket-cap1k

LeRobot diffusion policy (ROBOTIS lerobot-cyclo fork, 288 M parameters) for the bimanual cup handover onto a basket with the ROBOTIS FFW SG2 (SG2 + CAD grippers, calibrated wrist cameras), trained on sim demos.

Files at the repo root = step 100k (DiffusionPolicy.from_pretrained("padiredd/dp-cupbasket-cap1k")); step-10000/, step-20000/, step-30000/, step-40000/, step-50000/, step-60000/, step-70000/, step-80000/, step-90000/ = intermediate checkpoints (every 10k steps).

data	pick-cup-basket-cap1k-3cam (not on HF yet): 1,011 successful sim demos of the bimanual cup handover onto an upside-down woven basket (ROBOTIS FFW SG2 + CAD grippers in Isaac Lab), generated by a code-as-policy program (data_collection/cup_basket/programs/cup_handover_v3.py: left pinches the cup rim, carries it to the real demos' handover pose, right grasps the body, left releases, right places it upright on the basket; per-episode random pinch tilt/azimuth/depth, pre-close, grasp height, place offset/yaw and timing); measured cup 6.25 in tall, 3.25 in basket; cup and basket positions from the real demos' ranges; failed episodes left out; first 60 parts (seeds 60001+) of the cup2k_v3 collection
sampling	uniform over all frames (LeRobot default EpisodeAwareSampler)
inputs / outputs	head + left wrist + right wrist cameras, each 240x240, + 22-dim joint state; action = 16 joint targets, 15 Hz
image augmentation	heavy, random order, up to 5 of: brightness 0.6-1.4, contrast 0.6-1.4, saturation 0.3-1.7, hue +-0.1, sharpness 0.3-2.0, affine +-10 deg / 10 % / 0.9-1.1, perspective 0.2, blur, random erasing, grayscale
state noise	none
recipe	100k steps, batch 32, ResNet18 per camera, Adam 1e-4, cosine, DDPM 100 steps, horizon 64 / 32 executed, n_obs_steps 1, seed 1000
hardware	one NVIDIA RTX 4090 24 GB (lab225), alone

Task string: "pick up the blue cup with the left hand, hand it to the right hand, and place it on the basket". Evaluation results are added once they are run.