TEG 1.0.0rc1

Reinforcement Learning Environments for train RL agents

TEG

TEG is a straightforward environment for Reinforcement Learning that enables the training of RL agents for a robot manipulator. It's based on the Gymnasium and Mujoco.

Installation

This project use python 3.7+

You can install it by using pip

pip install TEG

Or manually cloning the github repository

git clone https://github.com/Alexfm101/TEG.git 
cd TEG
python -m pip install -e .

Example

TEG environment are simple Python env classes to allow an AI agent to interact with them very simple. Here's an example:

from TEG.envs.UR5_v0 import UR5Env_v0

env = UR5Env_v0(simulation_frames=5, torque_control= 0.01, distance_threshold=0.05)

def main():
    for episode in range(5):
        print("episode {}".format(episode))
        env.reset()

        for t in range(1000):
            action = env.action_space.sample()
            observation, reward, done, info = env.step(action)
            
            
            if done:
                print("Episode finished after {} timesteps".format(t+1))
                break

    return env.robot, env.sim

if __name__ == '__main__':
    main()

🧾 License

The Apache 2.0 License