TRI-ML

TRI-ML / raiden

Public

RAIDEN is a toolkit for YAM robots that streamlines calibration, coordinated bimanual data collection, and dataset conversion.

13
0
100% credibility
Found Mar 24, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Raiden is a complete toolkit for recording synchronized camera videos and robot arm movements from YAM bimanual systems, converting them into formats ready for training AI robot policies.

How It Works

1
🔍 Discover Raiden

You hear about Raiden, the easy way to teach robot arms new skills by recording their movements with cameras.

2
🔌 Connect your gear

Plug in your robot arms and cameras — everything gets recognized automatically.

3
📐 Calibrate with a board

Hold up a checkerboard pattern and let the system learn how cameras see the world.

4
🤖 Teach by example

Guide the arms through tasks like picking objects — it records video and motion smoothly.

5
✨ Review your lessons

Browse recordings in a simple screen, mark which ones worked perfectly, and fix any mistakes.

6
🎥 Turn into training data

One click extracts clear images, depth maps, and poses ready for AI learning.

🚀 Ready to train robots

Your dataset is perfect — feed it to AI models and watch your robots learn new skills!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is raiden?

Raiden is a Python toolkit for YAM robot arms that streamlines the full data pipeline: camera calibration, bimanual teleoperation, multi-camera recording, and dataset conversion to policy-ready formats. It handles coordinated bimanual setups with leader-follower control or SpaceMouse, mixing ZED and RealSense cameras, plus depth from TRI Stereo or Fast Foundation Stereo. Forget raiden network github or raiden shogun builds—this raiden delivers robot-ready datasets via simple CLI commands like `rd calibrate`, `rd record`, and `rd convert`.

Why is it gaining traction?

It stands out with automated hand-eye calibration, metadata consoles for labeling demos, and outputs synced frames with extrinsics and joint poses—plug straight into training. Flexible depth backends and visualization via Rerun beat manual scripting, while bimanual IK avoids singularities for smoother collection. Devs grab it for turning raw SVO2/bag files into sharded WebDatasets fast.

Who should use this?

Robotics researchers with YAM dual-arm setups collecting manipulation data for policies. Perfect for teams doing teleop demos, needing precise wrist/scene calibration, or converting coordinated recordings to train vision-based models like raiden vision systems.

Verdict

Worth trying for YAM owners—excellent docs and CLI make it production-ready despite 13 stars and 1.0% credibility score. Early research software from Toyota Research Institute; test safely before heavy hardware use.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.