Expressive Whole-Body Control for
Stompy Humanoid Robots

This is the code adopted from expressive-humanoid and make it work for Stompy from kscale labs.

Installation

conda create -n humanoid python=3.8
conda activate humanoid
cd
pip3 install torch==1.10.0+cu113 torchvision==0.11.1+cu113 torchaudio==0.10.0+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
git clone git@github.com:chengxuxin/expressive_humanoid.git
cd expressive_humanoid
# Download the Isaac Gym binaries from https://developer.nvidia.com/isaac-gym 
cd isaacgym/python && pip install -e .
cd ~/expressive_humanoid/rsl_rl && pip install -e .
cd ~/expressive_humanoid/legged_gym && pip install -e .
pip install "numpy<1.24" pydelatin wandb tqdm opencv-python ipdb pyfqmr flask dill gdown

Next install fbx. Follow the instructions here.

Prepare dataset

1. Download from here and extract the zip file to ASE/ase/poselib/data/cmu_fbx_all that contains all .fbx files.

2. Gnerate .yaml file for the motions you want to use.

cd ASE/ase/poselib
python parse_cmu_mocap_all.py

This step is not mandatory because the .yaml file is already generated. But if you want to add more motions, you can use this script to generate the .yaml file.

3. Import motions

cd ASE/ase/poselib
python fbx_importer_all.py

This will import all motions in CMU Mocap dataset into ASE/ase/poselib/data/npy.

The dataset is quite large and importing takes a lot of effort, you can directly download the already generated dataset here

4. Retarget motions

cd ASE/ase/poselib
mkdir pkl retarget_npy
python retarget_motion_stompy_all.py

This will retarget all motions in ASE/ase/poselib/data/npy to ASE/ase/poselib/data/retarget_npy.

5. Gnerate keybody positions

This step will require running simulation to extract more precise key body positions.

cd legged_gym/legged_gym/scripts
python train.py debug --task stompy_view --motion_name motions_debug.yaml --debug

Train for 1 iteration and kill the program to have a dummy model to load.

python play.py debug --task stompy_view --motion_name motions_debug.yaml

It is recommended to use motions_autogen_all.yaml at the first time, so that later if you have a subset it is not neccessary to regenerate keybody positions. This will generate keybody positions to ASE/ase/poselib/data/retarget_npy.

Set wandb asset: you can login your own wandb account with arg --entity WANDB_ENTITY

Usage

To train a new policy

python train.py xxx-xx-some_descriptions_of_run --device cuda:0 --entity WANDB_ENTITY

xxx-xx is usually an id like 000-01. motion_type and motion_name are defined in legged_gym/legged_gym/envs/stompy/stompy_mimic_config.py. They can be also given as arguments. Can set default WANDB_ENTITY in legged_gym/legged_gym/utils/helpers.py.

To play a policy

python play.py xxx-xx --task stompy_mimic_eval --record_video

No need to write the full experimentt id. The parser will auto match runs with first 6 strings (xxx-xx). So better make sure you don't reuse xxx-xx. Delay is added after 8k iters. If you want to play after 8k, add --delay.

To play with example pretrained models, find expid of the pretrained model in log/stompy like 060-40,

python play.py 060-40 --task stompy_mimic_eval --delay --motion_name motions_debug.yaml

Try to press + or - to see different motions. The motion name will be printed on terminal. motions_debug.yaml is a small subset of motions for debugging and contains some representative motions.

5. Save models for deployment:

python save_jit.py --exptid xxx-xx

This will save the models in legged_gym/logs/parkour_new/xxx-xx/traced/.

Viewer Usage

Can be used in both IsaacGym and web viewer.

• ALT + Mouse Left + Drag Mouse: move view.
• [ ]: switch to next/prev robot.
• Space: pause/unpause.
• F: switch between free camera and following camera.

IsaacGym viewer specific

• +: next motion in yaml.
• -: prev motion in yaml.
• r: reset the motion to the beginning.
• ]: camera focus on next env.
• [: camera focus on prev env.

Arguments

• --exptid: string, can be xxx-xx-WHATEVER, xxx-xx is typically numbers only. WHATEVER is the description of the run.
• --device: can be cuda:0, cpu, etc.
• --delay: whether add delay or not.
• --checkpoint: the specific checkpoint you want to load. If not specified load the latest one.
• --resume: resume from another checkpoint, used together with --resumeid.
• --seed: random seed.
• --no_wandb: no wandb logging.
• --entity: specify wandb entity
• --web: used for playing on headless machines. It will forward a port with vscode and you can visualize seemlessly in vscode with your idle gpu or cpu. Live Preview vscode extension required, otherwise you can view it in any browser.
• --motion_name: e.g. 07_04 or motions_all.yaml. If motions_all.yaml is used, motion_type should be yaml.
• --motion_type: single or yaml
• --fix_base: fix the base of the robot.

For more arguments, refer legged_gym/utils/helpers.py.

Errors

If you observe errors with libpython3.8.so.1.0, you can try the following:

export LD_LIBRARY_PATH=PATH_TO_YOUR_ENV/lib:$LD_LIBRARY_PATH

Acknowledgement

We derive the retargetting code from ASE.

The original expressive humanoid code is from expressive-humanoid