license: cc-by-nc-nd-4.0
task_categories:
- robotics
tags:
- LeRobot
- robot arms
- so-101 arms
- so-101
- imitation learning
size_categories:
- n<1K
So-101 Dataset
The dataset comprises 20 episodes of high-quality teleoperated robotic arm recordings at ~30 FPS. It is a compact, curated set captured using multi-view cameras (front, wrist, top) within the calibrated robot studio of the SO-101 arm, leveraging the LeRobot framework. Specifically designed for imitation learning and model training on the LeRobot platform, this dataset demonstrates basic manipulation and pick-and-place tasks with structured LeRobot v2.1 annotations.
By utilizing this robotic dataset, researchers and developers can advance evaluation of pretrained models, address issues in setups, and explore calibrations for robotics applications facing real-world tasks. - Get the data
Its core purpose is to support imitation learning, model training, and the evaluation of pretrained models. This compact set of 20 teleoperated robotic arm recordings is designed for imitation learning, model training, and training models on the lerobot platform, demonstrating pick-and-place and basic manipulation tasks with structured LeRobot v2.1 annotations across robots, arms, platforms, frameworks, and tools.
💵 Buy the Dataset: This is a limited preview of the data. To access the full dataset, please contact us at https://unidata.pro to discuss your requirements and pricing options.
Researchers can leverage this dataset to optimize training on the lerobot platform, tackle issues in robotic frameworks, and build robust models for robots and arms.