mohamednabil500/dreamer-breakout-expert

-- tags: - reinforcement-learning - model-based-rl - dreamer - breakout - atari - gymnasium

Dreamer Agent for Breakout

This is a Model-Based Reinforcement Learning agent using the Dreamer architecture trained on Atari Breakout. Unlike model-free algorithms (like PPO), Dreamer learns a world model to predict future states and rewards, allowing it to "dream" and learn behaviors within its own internal simulation.

Training Details

• Algorithm: Dreamer (Model-Based RL)
• Environment: BreakoutNoFrameskip-v4
• Framework: PyTorch/TensorFlow (Dreamer implementation)
• Key Components: Recurrent State-Space Model (RSSM), Actor, and Critic.

Performance

• Behavior: Known for high sample efficiency and learning latent representations of the game environment.