Scalable Behavior Cloning with Open Data, Training, and Evaluation
2026-06-25 • Robotics
Robotics
AI summaryⓘ
The authors present ABC, an open-source system designed to teach robots how to perform tasks by copying human behavior, called behavior cloning. They created ABC-130K, a huge dataset of robot teleoperation with over 3,500 hours of recordings across 195 different tasks. They also share tools like hardware designs, training setups, and simulations to help others train and test robotic models. Their research compares different model approaches and shows these robots can do complex tasks like folding boxes and taking credit cards out of wallets. Overall, their goal is to provide a shared resource for researchers to build better robot skills together.
Behavior CloningTeleoperationDatasetDiffusion TransformersVision-Language-Action ModelsSimulationDexterous ManipulationCo-trainingRoboticsOpen-source
Authors
Arthur Allshire, Himanshu Gaurav Singh, Ritvik Singh, Adam Rashid, Hongsuk Choi, David McAllister, Justin Yu, Yiyuan Chen, Huang Huang, Pieter Abbeel, Xi Chen, Rocky Duan, Phillip Isola, Jitendra Malik, Fred Shentu, Guanya Shi, Philipp Wu, Angjoo Kanazawa
Abstract
We introduce ABC, a fully open-source stack for manipulation with behavior cloning. At its core is ABC-130K: the largest open-source teleoperation dataset to date, featuring 3,500 hours of data spanning over 130K episodes across 195 diverse tasks. Furthermore, we open-source our accessible hardware setup, training infrastructure, and simulation pipeline. We also release 400 hours of sim-teleop data and provide a co-training recipe that produces correlated simulation and real-world evaluation, offering a reliable proxy for ablating model-design and training decisions before costly real-world evaluation. We explore various training recipes and compare common architectural choices for Diffusion Transformers (DiT) and Vision-Language-Action (VLA) models, grounding our findings in real-world evaluations. The resulting policies successfully execute dexterous tasks such as box folding and extracting credit cards from wallets. By providing a reproducible toolkit, we aim to place researchers on an equal footing, establishing the necessary foundation to learn the ABCs of Behavior Cloning together as a community.