A mallard version of ASE (https://xbpeng.github.io/projects/ASE/index.html)
Download Isaac Gym from the website, then follow the installation instructions.
Once Isaac Gym is installed, install the external dependencies for this repo:
pip install -r requirements.txt
First, an AWD model can be trained to imitate a dataset of motions clips using the following command:
python awd/run.py --task DucklingAMP --cfg_env awd/data/cfg/go_bdx/duckling.yaml --cfg_train awd/data/cfg/go_bdx/train/awd_duckling.yaml --motion_file awd/data/motions/go_bdx/walk_forward.txt --headless
--motion_file can be used to specify a dataset of motion clips that the model should imitate.
The task GoBDX will train a model to imitate a dataset of motion clips.
Over the course of training, the latest checkpoint Checkpoint.pth will be regularly saved to output/,
along with a Tensorboard log. --headless is used to disable visualizations. If you want to view the
simulation, simply remove this flag. To test a trained model, use the following command:
python awd/run.py --test --task DucklingAMP --num_envs 8 --cfg_env awd/data/cfg/go_bdx/duckling.yaml --cfg_train awd/data/cfg/go_bdx/train/awd_duckling.yaml --motion_file awd/data/motions/go_bdx/walk_forward.txt --checkpoint [path_to_awd_checkpoint]
You can also test the robustness of the model with --task GoBDXPerturb, which will throw projectiles at the character.
We also provide an implementation of Adversarial Motion Priors (https://xbpeng.github.io/projects/AMP/index.html). A model can be trained to imitate a given reference motion using the following command:
python awd/run.py --task DucklingAMP --cfg_env awd/data/cfg/go_bdx/train/amp_duckling.yaml --cfg_train awd/data/cfg/go_bdx/train/amp_duckling.yaml --motion_file awd/data/motions/go_bdx/walk_forward.txt --headless
The trained model can then be tested with:
python awd/run.py --test --task DucklingAMP --num_envs 16 --cfg_env awd/data/cfg/go_bdx/train/amp_duckling.yaml --cfg_train awd/data/cfg/go_bdx/train/amp_duckling.yaml --motion_file awd/data/motions/go_bdx/walk_forward.txt --checkpoint [path_to_amp_checkpoint]
Motion clips are located in awd/data/motions/. Individual motion clips are stored as .json files. Motion datasets are specified by .yaml files, which contains a list of motion clips to be included in the dataset. Motion clips can be visualized with the following command:
python awd/run.py --test --task DucklingViewMotion --num_envs 2 --cfg_env awd/data/cfg/go_bdx/duckling.yaml --cfg_train awd/data/cfg/go_bdx/train/amp_duckling.yaml --motion_file awd/data/motions/go_bdx/walk_forward.json
python awd/run.py --test --task DucklingViewMotion --num_envs 2 --cfg_env awd/data/cfg/mini_bdx/duckling.yaml --cfg_train awd/data/cfg/mini_bdx/train/amp_duckling.yaml --motion_file awd/data/motions/mini_bdx/walk_forward.json
--motion_file can be used to visualize a single motion clip .npy or a motion dataset .yaml.
See this
python gait_playground.py
python gait_playground.py --mini
python view_urdf.py awd/data/assets/go_bdx/go_bdx.urdf
python view_urdf.py awd/data/assets/mini_bdx/urdf/bdx.urdf
python view_urdf.py awd/data/assets/go_bdx/go_bdx.urdf --frames left_foot right_foot trunk head
python view_urdf.py awd/data/assets/mini_bdx/urdf/bdx.urdf --frames left_foot right_foot trunk head