ACT
Action Chunking with Transformers, the imitation-learning policy introduced with ALOHA by Stanford, UC Berkeley and Meta. It runs as a single-task baseline on RoboTwin 2.0 and RoboDojo.
Stanford / UC Berkeley et al. · Apr 23, 2023 · PolicyReport an error
0.5%Dojo Sim · #62 of 71
15.7%RoboTwin 2 · #32 of 33
2Benchmarks reported
Apr 23, 2023Released
Benchmark results2
RoboDojo (sim)Simulation
- Official leaderboard ↗Jul 6, 2026· single-task
#62 of 710.5%
- Official leaderboard ↗Aug 4, 2025· single-task SFT· n = 10000
#32 of 3315.7%
Not reported on 13 benchmarks
Wrong or missing number? Report an error