Skip to content
RobothecaBeta
Log in

Faster-WAM

  • AI model
  • Humanoids and robot learning

Faster-WAM is an open-source world action model for robot manipulation that reuses future representations during inference to speed up action prediction while staying robust under distribution shifts.

Made by HUST Vision Lab, China (checked on github.com/hustvl (opens in a new tab)).

Source on GitHub: github.com/hustvl/FasterWAM (opens in a new tab). Facts read at commit 8366781 (opens in a new tab).

Licenses

Code
Apache-2.0 (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (Apache-2.0) at commit 8366781 (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

What it does

  • Computes future representations once and selectively reuses them during action denoising, reducing redundant video-action interaction.
  • Reports state-of-the-art in-distribution performance and robust out-of-distribution generalization on simulated and real-world manipulation.
  • Substantially reduces inference latency compared with the FastWAM and JointWAM baselines kept in the repository.
  • Builds on the Wan2.2-TI2V-5B video model, with scripts for ActionDiT initialization, text embedding caching, and training.
  • Provides released checkpoints for LIBERO and RoboTwin, plus evaluation scripts for LIBERO, LIBERO-Plus, and RoboTwin.
  • Includes a script to measure inference latency across the FasterWAM, FastWAM, and JointWAM models.

Written by Robotheca from the project's README (opens in a new tab) at commit 8366781.

Getting started

Difficulty
Advanced

Maintenance

Status
Active: a commit in the last 90 days
Newest commit
31 Aug 2026 on main
Latest release
None published on GitHub

Checked on GitHub on 16 Oct 2026.

Links

Robots that can run 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 AI models as a reference: nothing is hosted or run here, and every fact links to where it was read.