Credit to PHC and simple-retarget

Installation

Here we use uv to manage the environment, please make sure you have installed uv first.

./install.sh

Modifications:

• set OMP_NUM_THREADS to 1 for multiprocessing
• simplified the config for retargeting process

Data preparation

Download the smpl model first:

uv run gdown --folder https://drive.google.com/drive/folders/1eSfJma_5VuNqaw_IRE8xVn6UgAmvjoeT\?usp\=drive_link -O ./data/smpl

Under the data/AMASS folder, I have downloaded one motion file for instance:

|-- data
|   |-- AMASS
|       |-- SFU                                 # sub-dataset name
|           |-- 0005                            # subject id
|               |-- 0005_Walking001_poses.npz   # motion file name
|               |-- ...

Of course, you can download the other motion files you like from the AMASS website.

Fitting motion data from AMASS

First we need to fit the smpl shape to the humanoid robot:

uv run scripts/1-fit_smpl_shape.py

It will save the fitted smpl beta parameters in the g1_29dof_shape.npz file.

Then we need to fit the motion data to the humanoid robot:

uv run scripts/3-fit_smpl_motion_mp.py +motion_path=./data/AMASS +target_fps=30 +output_name=<motion_file>

It will save the fitted motion data in the data/g1_29dof/<output_name>.pkl file.

Finally, we can visualize the fitted motion data:

uv run scripts/3-vis_q_mj.py +motion_file=./data/g1_29dof/<output_name>.pkl