HandsOn splits evaluation: Core-50 execution vs Transfer-16 zero-demo transfer
stingning · x · 2026-10-27
HandsOn separates demonstrated-task execution from transfer: Core-50 provides demonstrations for policy learning and evaluates execution across 50 tasks; Transfer-16 contains no target-task training demos, with 16 tasks covering 7 asset/relation tasks, 6 skill compositions, and 3 deformable-object tasks. Every task was verified by human operators. Evaluation uses task-specific completion checks (release, stability, or dwell requirements where applicable), and ordered progress is reported so partial completion can be examined separately from full success.
More from Embodied
- Mundo turns a few room photos into an explorable, simulatable 3D world — at 1.5 FPS — mark_k · 2026-10-27
- Zurich Physical AI Hackathon puts attendees in a room with $1.5M of humanoids — ___Mufasaa · 2026-10-27
- Apple responds after iPhone Duo inner displays crack along the hinge — mobileraj · 2026-10-27
- VLA models encode task concepts as curved multi-dimensional geometry, 19-author study finds — ninamiolane · 2026-10-27
- Munro Live Posts Full Cybercab Engineering Deep-Dive Video — yunta_tsai · 2026-10-27
- Tesla's Robotaxi Fleet Keeps Growing in Austin — yunta_tsai · 2026-10-27