Skip to content
RobothecaBeta
Log in

ACT + ALOHA

  • Software
  • Humanoids and robot learning
  • Framework
  • Runs on hardware you can build

ACT is an imitation-learning policy that predicts action chunks with transformers, with simulated tasks for training and evaluating it in sim or on real robots.

  • Stale

Source on GitHub: github.com/tonyzhaozh/act (opens in a new tab). Facts read at commit 742c753 (opens in a new tab).

Licenses

Code
MIT (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit 742c753 (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

Labels

  • StaleNo commit in 12 months: the newest is from 28 Jan 2024.

What it does

  • Implements the ACT policy, adapted from DETR, for learning robot behaviour from demonstrations.
  • Includes two simulated environments, Transfer Cube and Bimanual Insertion, with joint-space and end-effector control.
  • Provides scripts to record scripted demonstration episodes, train ACT, and evaluate the trained policy.
  • Supports temporal ensembling during evaluation to smooth actions.
  • Can be used on real hardware when combined with the separate ALOHA project.
  • Includes a script to save videos of episodes from an HDF5 dataset.

Written by Robotheca from the project's README (opens in a new tab) at commit 742c753.

Getting started

Difficulty
Intermediate

Maintenance

Status
Stale: no commit in 12 months
Newest commit
28 Jan 2024 on main
Latest release
None published on GitHub

Checked on GitHub on 16 Oct 2026.

Links

Robots on Robotheca that use it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.