Skip to content

Dataset · OpenMOSS-Team

Robotic manipulation in an omnimodal context

OpenMOSS-Team/OmniAction-LIBERO

Data for proactive robotic manipulation with audio and multimodal context.

What's inside

The dataset is related to RoboOmni, an approach to proactive robotic manipulation in an omnimodal context. The card specifies robotics, any-to-any, and audio-to-audio tasks in English. The project accompanies a paper accepted at ICLR 2026; exact volume is not specified in metadata.

What you can do with it

  • Train a VLA model for robotic manipulation
  • Evaluate an any-to-any audio-multimodal system
  • Fine-tune a proactive robot behavior model
Tasks
Robotics, Multimodal, Audio → audio
Languages
English
Downloads this month
7.5K
Likes
70

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("OpenMOSS-Team/OmniAction-LIBERO")