MoMo Framework Enables Dialable Motion Modes in Robotics
July 29, 2026
MoMo introduces a two-stage imitation-learning framework using a spatiotemporal action tokenizer and a behavior-cloning transformer. This allows robots to adapt execution-level motion through a continuous motion-mode condition.
HOW THIS AFFECTS YOU
●
builderYou can implement more fluid and task-specific robot manipulations using this two-stage framework.
●
researcherYou can leverage spatiotemporal action tokenization to improve robot adaptability across diverse tasks.