This work is about creating a DQN agent, able to learn and well perform in the Atari2600 Pong, using Reinforcement Learning and Deep Q-Network. The code is developed on a jupyter notebook in python while the game environment is provided by OpenAi called "OpenAi Gym".
The repository also contains a brief report of the work.
The code might contain errors!!
The training of the agent has been carried out on a Virtual Private Server (VPS) with 4 Ampere Altra CPU cores (based on the ARM Neoverse N1 architecture) and 24GB RAM on Oracle Cloud for 3 days and for a total number of 1200 episodes and then stopped manually.
