Sign inGet started
Dataset · Hugging Face

bx6d/r2s2r-blue-cube-to-bowl-expanded

xArm6 blue cube to bowl: Expanded Layouts and Robot Starts 1,000 training + 100 validation + 200 test episodes, complete and validated. Task: "Pick up the blue cube and place it in the bowl." There are 1,242,579 action timesteps and 1,243,879 two-camera observation pairs. This dataset uses the original visual style, not v7. It contains synthetic MuJoCo actuator-executed observations, not real robot recordings or ideal IK states. Table/wrist RGB images are 224×224 at 20 Hz. State… See the full description on the dataset page: https://huggingface.co/datasets/bx6d/r2s2r-blue-cube-to-bowl-expanded

Author / organizationbx6d
Downloads0 (as reported by Hugging Face)
Likes0
Created23 October 2026
Last updated at source23 October 2026
roboticsimitation-learningpi0.5mujocosyntheticxarm6
Embed this entry
<iframe src="https://silytics.si/embed/item/fa1d1e86-9a50-4c8a-8a24-47fef562be00" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

More from bx6d

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 23 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.