Scalable Behavior Cloning with Open Data, Training, and Evaluation

ABC: a fully open-source stack for bimanual manipulation with behavior cloning.

Scalable Behavior Cloning with Open Data, Training, and Evaluation Scalable Behavior Cloning with Open Data, Training, and Evaluation Arthur Allshire 1* , Himanshu Gaurav Singh 1* , Ritvik Singh 1* , Adam Rashid 2* , Hongsuk Choi 1* , David McAllister 1* Justin Yu 1,4 , Yiyuan Chen 4 , Huang Huang 4 , Pieter Abbeel 1,3 , Xi Chen 3 , Rocky Duan 3 , Phillip Isola 2 , Jitendra Malik 1,3 , Fred Shentu 1,4 , Guanya Shi 3,5 , Philipp Wu 4 , Angjoo Kanazawa 1,3 1 UC Berkeley 2 MIT 3 Amazon FAR 4 XDOF 5 Carnegie Mellon University * Core Contributor; work performed during an internship at Amazon FAR. 0

related reading

• State of Robot Learning, December 2025vedder.io
• A Careful Examination of Large Behavior Models for Multitask Dexterous Manipulationtoyotaresearchinstitute.github.io
• Generalist - GEN-0 / Embodied Foundation Models That Scale with Physical Interactiongeneralistai.com
• The First Fully General Computer Action Model | blogsi.inc
• Ch. 21 - Imitation Learningunderactuated.mit.edu
• EgoScaleresearch.nvidia.com
• [2304.13705] Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardwarear5iv.labs.arxiv.org
• e5b5c402bb7bd5e60bede6961d6fe39e-Paper-Conference.pdfproceedings.iclr.cc
• Open X-Embodiment: Robotic Learning Datasets and RT-X Modelsarxiv.org
• LDA-1B: Scaling Latent Dynamics Action Model via Universal Embodied Data Ingestionpku-epic.github.io
• GR00T N1: An Open Foundation Model for Generalist Humanoid Robotsarxiv.org
• ALOHA Unleashed: A Simple Recipe for Robot Dexterityaloha-unleashed.github.io