SMPL-X

To convert SMPL meshes (in .ply or .obj format) to SMPL+H parameters, use the transfer_model module with the smpl2smplh.yaml configuration.

Note: This command must be run from the top directory of the repository. Update the path in the .yaml config to point to your SMPL meshes.

python -m transfer_model --exp-cfg config_files/smpl2smplh.yaml

You can install the smplx package via PyPI or by cloning the repository and using setup.py.

Option 1: PyPI (Recommended)

pip install smplx[all]

Option 2: From Source

git clone https://github.com/vchoutas/smplx
python setup.py install

The model parameters are not included in the repository and must be downloaded separately after registering on the respective project websites:

• SMPL-X: https://smpl-x.is.tue.mpg.de
• SMPL+H: http://mano.is.tue.mpg.de
• SMPL (Male/Female): http://smpl.is.tue.mpg.de
• SMPL (Gender Neutral): http://smplify.is.tue.mpg.de

If you are working in a Python 2 virtual environment with Chumpy installed, you can use tools/clean_ch.py to remove any Chumpy objects from your model data files.

python tools/clean_ch.py --input-models path-to-models/*.pkl --output-folder output-folder

To use the SMPL-H version required by AMASS (which uses 16 betas), follow these steps to merge the specific model versions:

1. Download the zip folder from "Models & Code" and extract it to obtain the mano_v1_2 folder.
2. Download the zip folder from "Extended SMPL+H model" and extract it to obtain the smplh folder.
3. Run the merge script using the paths from the extracted folders.
4. Copy the resulting .pkl file to your smplx_models/smplh/ directory.

$ git clone https://github.com/vchoutas/smplx.git
$ cd smplx
$ python tools/merge_smplh_mano.py \
--smplh-fn /path/to/smplh/female/model.npz \
--mano-left-fn /path/to/mano_v1_2/models/MANO_LEFT.pkl \
--mano-right-fn /path/to/mano_v1_2/models/MANO_RIGHT.pkl \
--output-folder /path/to/smplh/merged

cp /path/to/smplh/merged/model.pkl /path/to/smplx_models/smplh/SMPLH_FEMALE.pkl

To convert SMPL-X meshes (in .ply or .obj format) to SMPL parameters, use the main.py script with the smplx2smpl.yaml configuration.

Important Constraints:

• Update the path in smplx2smpl.yaml to point to your SMPL-X meshes.
• When generating the input SMPL-X meshes, do not use hand and face parameters, as SMPL cannot model them and this information will be lost.

python main.py --exp-cfg config_files/smplx2smpl.yaml

To use the PyTorch SMPL-H module, you must first merge the SMPL-H and MANO parameters into a single file. This requires having the SMPL-H files and the MANO files (left and right) available locally.

python tools/merge_smplh_mano.py --smplh-fn SMPLH_FOLDER/SMPLH_GENDER.pkl \
 --mano-left-fn MANO_FOLDER/MANO_LEFT.pkl \
 --mano-right-fn MANO_FOLDER/MANO_RIGHT.pkl \
 --output-folder OUTPUT_FOLDER

To convert SMPL-X meshes (in .ply or .obj format) to SMPL+H parameters, use the transfer_model module with the smplx2smplh.yaml configuration.

Important Constraints:

• Update the path in the .yaml config to point to your SMPL-X meshes.
• When generating the input SMPL-X meshes, do not use jaw pose and expression parameters.

python -m transfer_model --exp-cfg config_files/smplx2smplh.yaml

To convert SMPL+H meshes (in .ply or .obj format) to SMPL-X parameters, use the transfer_model module with the smplh2smplx.yaml configuration.

Note: This command must be run from the top directory of the repository. Update the path in the .yaml config to point to your SMPL+H meshes.

python -m transfer_model --exp-cfg config_files/smplh2smplx.yaml

To use the model parameter transfer tools, follow these steps:

1. Install mesh.
2. Clone the SMPL-X repository:

   git clone https://github.com/vchoutas/smplx.git

3. Install Python dependencies:

   pip install -r requirements.txt

4. Install the Torch Trust Region optimizer.
5. Install the following additional packages:
   • loguru
   • open3d
   • omegaconf

To convert SMPL+H meshes (in .ply or .obj format) to SMPL parameters, use the transfer_model module with the smplh2smpl.yaml configuration.

Note: This command must be run from the top directory of the repository. Using this direction results in the loss of hand information.

python -m transfer_model --exp-cfg config_files/smplh2smpl.yaml