/Doom-RL

Implementation of Deep Reinforcement Learning algorithms. Methods are applied to a self-learning video game agent in the DOOM videogame.

Primary LanguagePython

Learning to kill demons.

Implemented methods can be found in different branch of the repository:

  • Policy gradient (master)
  • A2C (a2c)
  • A3C (a3c)
  • PPO (ppo)
  • PPO set up with transfer learning approach (PPO_transfer_learn)
  • DDDQN (ddqnper)

Video | Report