Command Palette
Search for a command to run...
ROCKET-2: 3D Game Zero-Shot Transfer
An error occurred in the Server Components render. The specific message is omitted in production builds to avoid leaking sensitive details. A digest property is included on this error instance which may provide additional details about the nature of the error.
Failed to load notebook details1. Tutorial Introduction

ROCKET-2 is a cross-view target alignment visual motion control framework launched by the CraftJarvis team on March 21, 2025. It focuses on solving complex task control problems in robotics. Through innovative multi-view target alignment technology, this project significantly improves the generalization ability and controllability of visual motion policies in dynamic environments. Related research papers are available. ROCKET-2: Steering Visuomotor Policy via Cross-View Goal Alignment .
This tutorial uses a single RTX 5090 card as the resource.
2. Project Examples

3. Operation steps
1. After starting the container, click the API address to enter the Web interface
If "Bad Gateway" is displayed, it means the model is initializing. Since the model is large, please wait about 1-2 minutes and refresh the page.

2. After entering the webpage, you can start a conversation with the model
How to use
- Go to Tutorial to view the tutorial guide

- Enter Customize Environment and select the environment you want to load

- Enter Launch Rocket loading environment

- Enter Specify Goal and select the target point and interaction method

- Enter Launch Rocket's Setting Panel and select the model

- Enter Launch Rocket's Control Panel to set the inference steps and perform inference

- Repeat steps 4-6 until the inference process is complete. Then, enter the Record Video mode to create and download the video. The video cannot be played online.

Citation Information
The citation information for this project is as follows:
@article{cai2025rocket,
title={ROCKET-2: Steering Visuomotor Policy via Cross-View Goal Alignment},
author={Cai, Shaofei and Mu, Zhancun and Liu, Anji and Liang, Yitao},
journal={arXiv preprint arXiv:2503.02505},
year={2025}
}
Build AI with AI
From idea to launch — accelerate your AI development with free AI co-coding, out-of-the-box environment and best price of GPUs.