Reinforcement Learning library for public, for now, it only supports TensorFlow.

pip install rockrl

RL algorithms are implemented to support gymnasium==0.29.1 version. Main requirements are that:

• env.reset() would return state and info states.
• env.step(action) would return state, reward, terminated, truncated, info states.

• PPO (Discrete and Continuous)

• RockRL/tensorflow/examples/ppo/LunarLander-v2/LunarLander-v2.py is an example of using PPO to solve LunarLander-v2 (Discrete) environment.
• RockRL/tensorflow/examples/ppo/BipedalWalker-v3/BipedalWalker-v3.py is an example of using PPO to solve BipedalWalker-v2 (Continuous) environment.
• RockRL/tensorflow/examples/ppo/BipedalWalkerHardcore-v3/BipedalWalkerHardcore-v3.py is an example of using PPO to solve BipedalWalker-v3 Hardcore (Continuous) environment.