Rerun - Visualize your data effortlessly

Rerun is an open-source platform designed to simplify the handling and visualization of multimodal data for spatial and embodied AI. With its intuitive setup and powerful logging features, you can quickly analyze and visualize your data.

How It Works

Rerun operates on a time-aware entity component system that helps streamline the modeling of data for effective storage and visualization. Here’s how it works:

• Model: Utilize the Rerun SDK to structure your data and log it. The seamless integration ensures you can write data to storage or a live viewer effortlessly.
• View: Leverage Rerun’s fast multimodal viewer to visualize both live and recorded data interactively. Enjoy features like time travel through your visuals for better analysis.
• Build: Customize your visualizations via code or directly in the UI. The flexibility allows you to integrate visual tools into various applications seamlessly.
• Query: Implement powerful query APIs to pull time-aligned datasets from robotics recordings, transforming complex logs into usable data for further analysis.

Usage

Getting started with Rerun is straightforward and requires minimal setup. Follow these steps for installation and initial usage:

1. Install the SDK: Begin by installing the Rerun SDK using pip. The command is simple:

   pip install rerun-sdk
   

2. Initialize your application: Start by importing the SDK in your Python script and initializing it with your application name:

   import rerun as rr
   rr.init('my_data_generating_application')
   

3. Connect to the viewer: Set up a connection with the remote viewer to visualize your logs:

   rr.connect()
   

4. Log data: Use the SDK's logging functions to record spatial data, image data, and any errors:

   rr.log('points', rr.Points3D(positions))
   rr.log('camera', rr.Transform3D(pos, rot))
   rr.log('camera/image', rr.Image(tensor))
   

5. View data: After logging, open the viewer to interact with the data you have recorded. The viewer allows for seamless visualization and debugging for better understanding.
6. Analyze data: Use query APIs to extract aligned datasets for further analysis—making your data management efficient and effective.

By following these basic steps, you can start visualizing and analyzing your multimodal data within seconds.

Research and Development

Utilize Rerun to visualize and analyze data from experiments, helping to refine models through clear visual feedback.

Machine Learning Model Training

Integrate Rerun into your model training pipeline to track training metrics and visualize outcomes effectively.

Robotic Systems Testing

Employ Rerun to log and visualize sensor data from robotic systems, ensuring clear identification of areas for improvement.

Prototype Debugging

Use Rerun to visualize prototype data and pinpoint issues, leading to faster iteration cycles.

Data Analysis and Reporting

Communicate findings effectively by visualizing complex data through Rerun’s easy-to-understand interfaces.

Education and Training

Implement Rerun as a teaching tool to help students visualize and comprehend AI and data handling concepts.

Features

• Open Source: Rerun is dual licensed under MIT and Apache 2, allowing free access to its powerful features for everyone.
• Multimodal Visualization: The tool provides powerful and interactive visualizations for both live and recorded data, making analysis intuitive.
• Flexible SDK: Users can easily log data and integrate with their systems using a simple and effective SDK.
• Time-Aware Modeling: Rerun's data model facilitates time-based analysis, ideal for experimental evaluations.
• Easy Setup: Installation requires minimal effort, enabling users to get started in seconds.
• Comprehensive Query APIs: Developers can extract aligned datasets efficiently, aiding in deeper analysis across various applications.

FAQ

1. What is Rerun?

Rerun is an open-source platform designed to handle and visualize multimodal data specifically for spatial and embodied AI.

2. How can I install Rerun?

You can install Rerun simply by running pip install rerun-sdk in your Python environment.

3. Is Rerun suitable for beginners?

Yes, Rerun is designed for ease of use, making it accessible for both beginners and experienced developers.

4. What types of data can I visualize with Rerun?

Rerun supports various types of multimodal data including 3D points, images, and sensor data.

5. How does the Rerun viewer work?

The Rerun viewer allows for interactive visualization of both live and recorded data, enabling time travel for deeper insights.

6. Is there a commercial version of Rerun?

Yes, Rerun is in development for a commercial version focused on data management at scale.

7. What are the main features of Rerun?

Key features include powerful visualizations, a flexible SDK, open-source access, and comprehensive query APIs.

8. Can I customize my visualizations in Rerun?

Absolutely, you can build layouts and customize visualizations through code or the UI.