FastUMI 100K: Advancing Data Driven Robotic Manipulation with a Large Scale UMI Style Dataset [paper] [dataset] Overview FastUMI 100K is a large scale, high quality UMI style dataset designed for data driven robotic manipulation learning. Featuring over 100K+ demonstration trajectories across 54 diverse tasks and hundreds of object types, the dataset provides multi view wrist mounted fisheye images and high frequency end effector states. To facilitate seamless integration with modern robot learning frameworks, all data has been standardized into the LeRobot v2.1 format. Installation To use the FastUMI 100K dataset, you need to install the official Hugging Face LeRobot library. 1. Create and activate a conda environment 2. Install LeRobot (For advanced usage or compiling from source, please refer to the official LeRobot repository.) Dataset Structure The dataset is recorded at 20 FPS . All states and actions are provided in the relative end effector frame. The dataset is naturally divided into two main configurations: Single Arm Configuration Used for tasks such as take items out of drawer and wash clothes . Observation Images: Primary view camera (720, 1280, 3) Observation State /…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy