Implementation of DQN improvements leading to Rainbow.
Deep Q-Networks (DQN): https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/1_DQN.ipynb
Double DQN (DDQN): https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/2_Double_DQN.ipynb
Prioritized Experience Replay (PER): https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/3_PER.ipynb
Dueling DQN: https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/4_Dueling_DQN.ipynb
Multi-Step Learning: https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/5_Multi_Step_Learning.ipynb
Distributional RL: https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/6_Distributional_RL.ipynb
7Noisy Nets: https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/7_Noisy_Nets.ipynb
Rainbow: https://nbviewer.jupyter.org/github/ridouaneg/rainbow_rl/blob/main/8_Rainbow.ipynb