← Leaderboard

UniAct

VLA
Tsinghua AIR et al. · Jan 17, 2025

Small cross-embodiment VLA that learns a universal action space shared across robots, decoded by lightweight embodiment-specific heads. Its LIBERO results are published only as a chart, so it has no ranked scores.

Type
VLA
Parameters
500M
Backbone
LLaVA-OneVision-0.5B
Weights
Open
License
–
Commercial use
Unclear
Released
Jan 17, 2025
Robots
WidowX, Franka Panda (sim), AIRBOT
SO-101
–
LeRobot
–
Inputs
–
VRAM
–

Benchmark results

0 of 13 benchmarks

No benchmark results recorded yet.

Not reported on 13 benchmarks
Wrong or missing number? Report an error