Conditional Motion In-Betweening (CMIB)
Official implementation of paper: Conditional Motion In-betweeening.
Paper(arXiv) | Project Page | YouTube
| in-betweening | pose-conditioned |
|---|---|
 |
 |
| walk | jump | dance |
|---|---|---|
 |
 |
 |
Environments
This repo is tested on following environment:
- Ubuntu 20.04
- Python >= 3.7
- PyTorch == 1.10.1
- Cuda V11.3.109
Install
-
Follow
LAFAN1dataset's installation guide. You need to install git lfs first before cloning the dataset repo. -
Run LAFAN1's
evaluate.pyto unzip and validate it. (Installnumpyfirst if you don't have it)$ pip install numpy $ python ubisoft-laforge-animation-dataset/evaluate.py
With this, you will have unpacked LAFAN dataset under
ubisoft-laforge-animation-datasetfolder. -
Install appropriate
pytorchversion depending on your device(CPU/GPU), then install packages listed inrequirements.txt. .
Trained Weights
You can download trained weights from here.
Train from Scratch
Trining script is trainer.py.
python trainer.py \
--processed_data_dir="processed_data_80/" \
--window=90 \
--batch_size=32 \
--epochs=5000 \
--device=0 \
--entity=cmib_exp \
--exp_name="cmib_80" \
--save_interval=50 \
--learning_rate=0.0001 \
--loss_cond_weight=1.5 \
--loss_pos_weight=0.05 \
--loss_rot_weight=2.0 \
--from_idx=9 \
--target_idx=88 \
--interpolation='slerp'
Inference
You can use run_cmib.py for inference. Please refer to help page of run_cmib.py for more details.
python run_cmib.py --help
Reference
- LAFAN1 Dataset
@article{harvey2020robust, author = {Félix G. Harvey and Mike Yurick and Derek Nowrouzezahrai and Christopher Pal}, title = {Robust Motion In-Betweening}, booktitle = {ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH)}, publisher = {ACM}, volume = {39}, number = {4}, year = {2020} }
Citation
@misc{kim2022conditional,
title={Conditional Motion In-betweening},
author={Jihoon Kim and Taehyun Byun and Seungyoun Shin and Jungdam Won and Sungjoon Choi},
year={2022},
eprint={2202.04307},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
134 Dec 23, 2022
131 Dec 31, 2022
31 Jul 22, 2022
22 Nov 08, 2022
276 Dec 31, 2022
7.4k Jan 04, 2023
109 Nov 16, 2022
1 Jan 19, 2022
47 Dec 16, 2022
83 Oct 13, 2022
789 Jan 04, 2023
8 Jun 07, 2022
160 Jan 04, 2023
96 Dec 22, 2022
33 Oct 31, 2022
5 Mar 04, 2022
13 Nov 04, 2022
27 Dec 16, 2022
19 Nov 20, 2022
958 Jan 03, 2023