Sim2Sim

This section explains how to deploy policies trained in Gym to other simulators, avoiding overfitting to Gym-specific characteristics.

Run the following command to launch the MuJoCo simulation:

cd ~/Documents/pnd_rl_gym
python deploy/deploy_mujoco/deploy_mujoco.py adam_lite_12dof.yaml

Replace Network Model

Default model path: deploy/pre_train/{robot}/motion.pt

User-trained models are saved at logs/adam_lite_12dof/exported/policies/policy_lstm_1.pt.
You only need to replace policy_path in the .yaml configuration file.