ZeyuLing/Motius-MotionGPT3-HumanML3D

<h1 align="center">MotionGPT3 Model Card</h1>

<p align="center"> <strong>A continuous-latent bimodal motion-language model, packaged for HumanML3D motion captioning.</strong> </p>

<p align="center"> <a href="https://arxiv.org/abs/2506.24086">Paper</a> | <a href="https://motiongpt3.github.io/">Project Page</a> | <a href="https://github.com/OpenMotionLab/MotionGPT3">Original GitHub</a> | <a href="https://huggingface.co/ZeyuLing/Motius-MotionGPT3-HumanML3D">Motius Checkpoint</a> </p>

MotionGPT3 separates text and motion processing into modality-specific branches with shared attention. Unlike tokenized motion-language models, it represents motion in a continuous VAE latent space. The Motius artifact packages the final official multi-task checkpoint and all model/tokenizer configuration required by Pipeline.from_pretrained.

Release Snapshot

Usage

import numpy as np
from motius.pipelines.motiongpt3 import MotionGPT3Pipeline

pipe = MotionGPT3Pipeline.from_pretrained(
    "ZeyuLing/Motius-MotionGPT3-HumanML3D",
    bundle_kwargs={"device": "cuda"},
)
motion = np.load("sample.npy")  # denormalized HumanML3D-263
caption = pipe.infer_m2t([motion], lengths=[len(motion)])[0]

M2T Evaluation

Motius Components

Citation

@misc{zhu2025motiongpt3,
  title={MotionGPT3: Human Motion as a Second Modality},
  author={Zhu, Bingfan and Jiang, Biao and Wang, Sunyi and Tang, Shixiang and Chen, Tao and Luo, Linjie and Zheng, Youyi and Chen, Xin},
  year={2025},
  eprint={2506.24086},
  archivePrefix={arXiv}
}

Direct Loading

from motius import Pipeline

pipeline = Pipeline.from_pretrained("ZeyuLing/Motius-MotionGPT3-HumanML3D")