Train RL in Simulation

This guide explains how to use the gym_hil simulation environments as an alternative to real robots when working with the LeRobot framework for Human-In-the-Loop (HIL) reinforcement learning.

gym_hil is a package that provides Gymnasium-compatible simulation environments specifically designed for Human-In-the-Loop reinforcement learning. These environments allow you to:

Currently, the main environment is a Franka Panda robot simulation based on MuJoCo, with tasks like picking up a cube.

Installation

First, install the gym_hil package within the LeRobot environment:

pip install -e ".[hilserl]"

What do I need?

Configuration

To use gym_hil with LeRobot, you need to create a configuration file. An example is provided here. Key configuration sections include:

Environment Type and Task

{
    "type": "hil",
    "name": "franka_sim",
    "task": "PandaPickCubeGamepad-v0",
    "device": "cuda"
}

Available tasks:

Gym Wrappers Configuration

"wrapper": {
    "gripper_penalty": -0.02,
    "control_time_s": 15.0,
    "use_gripper": true,
    "fixed_reset_joint_positions": [0.0, 0.195, 0.0, -2.43, 0.0, 2.62, 0.785],
    "end_effector_step_sizes": {
        "x": 0.025,
        "y": 0.025,
        "z": 0.025
    },
    "control_mode": "gamepad"
    }

Important parameters:

Running with HIL RL of LeRobot

Basic Usage

To run the environment, set mode to null:

python lerobot/scripts/rl/gym_manipulator.py --config_path path/to/gym_hil_env.json

Recording a Dataset

To collect a dataset, set the mode to record whilst defining the repo_id and number of episodes to record:

python lerobot/scripts/rl/gym_manipulator.py --config_path path/to/gym_hil_env.json

Training a Policy

To train a policy, checkout the configuration example available here and run the actor and learner servers:

python lerobot/scripts/rl/actor.py --config_path path/to/train_gym_hil_env.json

In a different terminal, run the learner server:

python lerobot/scripts/rl/learner.py --config_path path/to/train_gym_hil_env.json

The simulation environment provides a safe and repeatable way to develop and test your Human-In-the-Loop reinforcement learning components before deploying to real robots.

Congrats 🎉, you have finished this tutorial!

If you have any questions or need help, please reach out on Discord.

Paper citation:

@article{luo2024precise,
  title={Precise and Dexterous Robotic Manipulation via Human-in-the-Loop Reinforcement Learning},
  author={Luo, Jianlan and Xu, Charles and Wu, Jeffrey and Levine, Sergey},
  journal={arXiv preprint arXiv:2410.21845},
  year={2024}
}
< > Update on GitHub