# GenM$^3$: Generative Pretrained Multi-path Motion Model for Text Conditional Human Motion Generation

Source: [GenM$^3$: Generative Pretrained Multi-path Motion Model for Text Conditional Human Motion Generation](<https://www.sota2.com/research/paper/68fa7a441e209131cea820d5>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T06:58:49.434Z

Authors: Junyu Shi, Lijiang Liu, Yong Sun, Zhiyuan Zhang, Jinni Zhou, Qiang Nie

Published: 2025-03-19T05:56:52.000Z

## Abstract

Scaling up motion datasets is crucial to enhance motion generation capabilities. However, training on large-scale multi-source datasets introduces data heterogeneity challenges due to variations in motion content. To address this, we propose Generative Pretrained Multi-path Motion Model (GenM\\(^3\\)), a comprehensive framework designed to learn unified motion representations. GenM\\(^3\\) comprises two components: 1) a Multi-Expert VQ-VAE (MEVQ-VAE) that adapts to different dataset distributions to learn a unified discrete motion representation, and 2) a Multi-path Motion Transformer (MMT) that improves intra-modal representations by using separate modality-specific pathways, each with densely activated experts to accommodate variations within that modality, and improves inter-modal alignment by the text-motion shared pathway. To enable large-scale training, we integrate and unify 11 high-quality motion datasets (approximately 220 hours of motion data) and augment it with textual annotations (nearly 10,000 motion sequences labeled by a large language model and 300+ by human experts). After training on our integrated dataset, GenM\\(^3\\) achieves a state-of-the-art FID of 0.035 on the HumanML3D benchmark, surpassing state-of-the-art methods by a large margin. It also demonstrates strong zero-shot generalization on IDEA400 dataset, highlighting its effectiveness and adaptability across diverse motion scenarios.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Text-to-motion generation on HumanML3D 1 (test)](<https://www.sota2.com/research/sota/text-to-motion-generation-on-humanml3d-1-test.md>): Text-to-motion generation; HumanML3D 1 (test); R-Precision (Top 1): 0.511.

- [Text-to-motion generation on HumanML3D (test)](<https://www.sota2.com/research/sota/text-to-motion-generation-on-humanml3d-test.md>): Text-to-motion generation; HumanML3D (test); FID: 0.046.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2503.14919>)

[PDF](<https://arxiv.org/pdf/2503.14919.pdf>)