# VITRA

**Repository Path**: mirrors_microsoft/VITRA

## Basic Information

- **Project Name**: VITRA
- **Description**: VITRA: Scalable Vision-Language-Action Model Pretraining for Robotic Manipulation with Real-Life Human Activity Videos
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2025-10-27
- **Last Updated**: 2025-11-15

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

<h1 align="center"><span
    style="font-family: 'Courier New', Courier, monospace; font-size: 115%;"><span style="font-size: 130%;">V</span>ITRA</span>:<br><span
    style="font-size:2.22rem;">Scalable Vision-Language-Action Model Pretraining for Robotic Manipulation with Real-Life Human Activity Videos
    </span></h1>
<p align="center"><a href="https://arxiv.org/abs/2510.21571"><img src='https://img.shields.io/badge/arXiv-Paper-red?logo=arxiv&logoColor=white' alt='arXiv'></a>
<a href='https://microsoft.github.io/VITRA/'><img src='https://img.shields.io/badge/Project_Page-Website-green?logo=googlechrome&logoColor=white' alt='Project Page'></a>
<a href=''><img src='https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Model-blue'></a>
<a href=''><img src='https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Data-yellow'></a>
</p>
<p align="center"><img src="assets/teaser.jpg" width="100%"></p>

<span style="font-family: 'Courier New', Courier, monospace; font-size: 115%;"><span style="font-size: 130%;">V</span>ITRA</span> is a novel approach for pretraining robotic manipulation Vision-Language-Action (VLA) models using a large corpus of unscripted real-life video recordings of human hand activities. Treating human hand as dexterous robot end-effector, we show that "in-the-wild" egocentric human videos without any annotations can be transformed into data formats fully aligned with existing robotic V-L-A training data in terms of task granularity and labels.


## ⏩ Updates
**Data, code, and pretrained models will be released no later than 11/30/2025. Please stay tuned.**