MotionGlot: A Multi-Embodied Motion Generation Model

TL;DR

MotionGlot is a versatile motion generation model that adapts language model training principles to produce diverse motions across different embodiments, validated through multiple tasks, datasets, and real-world experiments.

Contribution

We introduce MotionGlot, a novel multi-embodied motion generation model using instruction-tuning inspired by large language models, along with two new motion datasets.

Findings

35.3% average improvement across tasks

Successful adaptation of LLM training principles to motion generation

Validated in real-world hardware experiments

Abstract

This paper introduces MotionGlot, a model that can generate motion across multiple embodiments with different action dimensions, such as quadruped robots and human bodies. By leveraging the well-established training procedures commonly used in large language models (LLMs), we introduce an instruction-tuning template specifically designed for motionrelated tasks. Our approach demonstrates that the principles underlying LLM training can be successfully adapted to learn a wide range of motion generation tasks across multiple embodiments with different action dimensions. We demonstrate the various abilities of MotionGlot on a set of 6 tasks and report an average improvement of 35.3% across tasks. Additionally, we contribute two new datasets: (1) a dataset of expert-controlled quadruped locomotion with approximately 48,000 trajectories paired with direction-based text annotations, and (2) a dataset of over 23,000 situational text prompts for human motion generation tasks. Finally, we conduct hardware experiments to validate the capabilities of our system in real-world applications.