Dueling Wizards

All about a game called "Dueling Wizards," where an RL model controls a character named Zephyra against an opponent named Aurion. This environment simulates the game's rules and dynamics, including health, energy levels, actions, cooldowns, and rewards, allowing the RL model to learn through interactions and improve its performance. The code also includes a training loop that uses the Stable Baselines3 library to train the RL model (Zephyra) against a pre-defined opponent logic module. The training process involves learning the optimal strategies for Zephyra by adjusting its policy based on rewards received, aiming to win the game against Aurion. Finally, the code incorporates a curriculum learning approach, progressively increasing the complexity of the opponent logic module to challenge Zephyra and encourage it to develop more sophisticated strategies.