Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
- Download and preprocess the HumanML3D dataset under the dataset/ directory.
- Convert the raw HumanML3D data into our global representation format.
(The conversion script will be released soon.)
python -m train.train_sfcontrol --endeffector_conditioned --endeffector_only --endeffector_selection_scheme random_joints --head --diffusion_steps 50 --latent_dim 128 --lambda_ric_pos 0.1 --lambda_ric_vel 10.0 --lambda_ric_fc 10.0 --dataset humanml --data_rep glo_jointspython -m train.train_sfcontrol --endeffector_conditioned --endeffector_selection_scheme all_endeffectors --head --diffusion_steps 50 --latent_dim 512 --dataset humanml --data_rep glo_rootIf you find our work helpful, please consider citing:
@misc{hwang2025motionsynthesissparseflexible,
title={Motion Synthesis with Sparse and Flexible Keyjoint Control},
author={Inwoo Hwang and Jinseok Bae and Donggeun Lim and Young Min Kim},
year={2025},
eprint={2503.15557},
archivePrefix={arXiv},
primaryClass={cs.GR},
url={https://arxiv.org/abs/2503.15557}
}
We sincerely thank the open-source projects that our code builds upon and draws inspiration from: CondMDI, GMD and MDM.