b:botstest.Submit a resource ↗

Models / Research

dexwm · facebookresearch

Official code and data from DexWM ("World Models Can Leverage Human Videos for Dexterous Manipulation").

world models
Publisher repository ↗
dexwm · facebookresearch resource page screenshot
Resource screenshot · captured 2026-10-10 · visually reviewed

Overview

Official code and data from DexWM ("World Models Can Leverage Human Videos for Dexterous Manipulation").

Review observation/action interfaces, checkpoints, robot embodiments and adaptation requirements. Repository code licensing does not automatically license weights or training data.

Project information

Maintainer
facebookresearch
Primary language
Python
Repository code license
Not specified
Model-weight license
Not recorded
Dataset license
Not recorded
Hardware design license
Not recorded
Maintenance
Not archived in the last recorded metadata
Last observed repository push
2026-06-23

Code licenses do not automatically cover model weights, datasets or hardware designs. Unrecorded license details require confirmation with the publisher.

Documentation sections

  • DexWM: World Models for Learning Dexterous Hand-Object Interactions from Human Videos <br><sub>Official PyTorch Implementation</sub>
  • [Paper](https://arxiv.org/pdf/2512.13644) | [Project Page](https://raktimgg.github.io/dexwm/) | [Data (RoboCasa Random)](https://huggingface.co/datasets/facebook/dexwm)
  • Setup
  • Data
  • Training
  • Pre-Train on EgoDex and DROID

Browse this resource’s categories

Models / World models ↗

Explore related collections

Browse Models →

More in Models