I'm trying to make a proejct for school, a deep q learning project from scratch, obviously in a dev container enviroment, with an Agent class and a NeuralNetwork abstract class.
It use gradient discent for optimization, e-greedy algorithm for exploration vs exploitation.
https://github.com/LorenzoTinfena/deep-q-learning-itt-final-project
sample: https://drive.google.com/file/d/1rpj_q0yWhHk19fWHck_IgzEuirXwNQaJ/view?usp=sharing