The official PyTorch implementation of the paper "Dance Like a Chicken: Low-Rank Stylization for Human Motion Diffusion".

Please visit our webpage for more details.

If you find this code useful in your research, please cite:

  @misc{sawdayee2025chicken,
    title={Dance Like a Chicken: Low-Rank Stylization for Human Motion Diffusion}, 
    author={Haim Sawdayee and Chuan Guo and Guy Tevet and Bing Zhou and Jian Wang and Amit H. Bermano},
    year={2025},
    eprint={2503.19557},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
    url={https://arxiv.org/abs/2503.19557}, 
}

This code was tested on Ubuntu 18.04.5 LTS and requires:

• Python 3.7
• conda3 or miniconda3
• CUDA capable GPU (one is enough)

Install ffmpeg (if not already installed):

sudo apt update
sudo apt install ffmpeg

For windows use this instead.

Setup conda env:

conda env create -f environment.yml
conda activate lora_mdm
python -m spacy download en_core_web_sm
pip install git+https://github.com/openai/CLIP.git

Download dependencies:

bash prepare/download_smpl_files.sh
bash prepare/download_glove.sh
bash prepare/download_t2m_evaluators.sh

Download the model(s) you wish to use from this Drive, then unzip and place them in ./save/.

python -m sample.generate  --lora_finetune --model_path save/mdm/model000500000.pt --styles Chicken --num_samples 10 --num_repetitions 3 --output_dir 'save/out/' 

python -m sample.generate  --lora_finetune --model_path save/mdm/model000500000.pt --styles Chicken --input_text ./assets/example_text_prompts.txt --output_dir 'save/out/' 

python -m sample.generate  --lora_finetune --model_path save/mdm/model000500000.pt --styles Chicken --output_dir 'save/out/' --text_prompt "A person is kicking in sks style."

Notes:

• --styles can be either a style name or a path
• Text prompts should end with 'in sks style.'

You may also define:

• --device id.
• --seed to sample different prompts.
• --motion_length (text-to-motion only) in seconds (maximum is 9.8[sec]).

Running those will get you:

• results.npy file with text prompts and xyz positions of the generated animation
• sample##_rep##.mp4 - a stick figure animation for each generated motion.

It will look something like this:

You can stop here, or render the SMPL mesh using the following script.

To create SMPL mesh per frame run:

python -m visualize.render_mesh --input_path /path/to/mp4/stick/figure/file

This script outputs:

• sample##_rep##_smpl_params.npy - SMPL parameters (thetas, root translations, vertices and faces)
• sample##_rep##_obj - Mesh per frame in .obj format.

Notes:

• The .obj can be integrated into Blender/Maya/3DS-MAX and rendered using them.
• This script is running SMPLify and needs GPU as well (can be specified with the --device flag).
• Important - Do not change the original .mp4 path before running the script.

Notes for 3d makers:

• You have two ways to animate the sequence:
  1. Use the SMPL add-on and the theta parameters saved to sample##_rep##_smpl_params.npy (we always use beta=0 and the gender-neutral model).
  2. A more straightforward way is using the mesh data itself. All meshes have the same topology (SMPL), so you just need to keyframe vertex locations. Since the OBJs are not preserving vertices order, we also save this data to the sample##_rep##_smpl_params.npy file for your convenience.

python -m train.train_mdm --save_dir save/mdm --dataset humanml --diffusion_steps 100 --arch trans_dec --text_encoder_type bert --mask_frames --lambda_prior_preserv 0.0

```shell
python -m train.train_mdm --save_dir save/lora/Chicken --num_steps 4000 --diffusion_steps 1000 --dataset 100style --arch trans_dec --text_encoder_type bert --starting_checkpoint save/mdm/model000500000.pt --styles Chicken --lora_finetune --mask_frames
```
* For evaluation use `--lambda_prior_preserv 0.25`

• Use --device to define GPU id.
• Use --arch to choose one of the architectures reported in the paper {trans_enc, trans_dec} (trans_enc is default).
• Add --train_platform_type {ClearmlPlatform, TensorboardPlatform} to track results with either ClearML or Tensorboard.
• Add --eval_during_training to run a short (90 minutes) evaluation for each saved checkpoint. This will slow down training but will give you better monitoring.

python  -m eval.eval_lora_mdm --model_path  save/mdm/model000500000.pt --lora_finetune --lora_rank 5 --arch trans_dec --text_encoder_type bert --classifier_style_group All

This code is standing on the shoulders of giants. We want to thank the following contributors that our code is based on:

MDM, lora-pytorch, SMooDi, MotionCLIP, text-to-motion, actor, joints2smpl, MoDi.

This code is distributed under an MIT LICENSE.

Note that our code depends on other libraries, including CLIP, SMPL, SMPL-X, PyTorch3D, and uses datasets that each have their own respective licenses that must also be followed.