README.md

metadata

license: apache-2.0

A repository that contains example json files that can be used for different applications of the LeRobot code base.

Current available configs:

• env_config_so100.json: real robot environment configuration to be used to teleoperate, record dataset and replay a dataset on the real robot with the lerobot/scripts/rl/gym_manipulator.py script.
• train_config_hilserl_so100.json: training config on the real robot with the HILSerl RL framework in LeRobot.
• gym_hil_env.json: simulated environment configuration for the gym_hil mujoco based env.
• train_gym_hil_env.json: training configuration for gym_hil with the HILSerl RL framework in LeRobot.
• reward_classifier_train_config.json: configuration to train a reward classifier with LeRobot.