Deep Reinforcement Learning For Flappy Bird
| Project Name | Stars | Downloads | Repos Using This | Packages Using This | Most Recent Commit | Total Releases | Latest Release | Open Issues | License | Language |
|---|---|---|---|---|---|---|---|---|---|---|
| Drl Flappybird | 536 | 3 years ago | 5 | Python | ||||||
| Playing Flappy Bird Using Deep Reinforcement Learning (Based on Deep Q Learning DQN using Tensorflow) | ||||||||||
| Reinforcement Learning Algorithms | 407 | 2 years ago | 4 | Python | ||||||
| This repository contains most of pytorch implementation based classic deep reinforcement learning algorithms, including - DQN, DDQN, Dueling Network, DDPG, SAC, A2C, PPO, TRPO. (More algorithms are still in progress) | ||||||||||
| Flappy Bird Deep Q Learning Pytorch | 407 | 2 years ago | 4 | mit | Python | |||||
| Deep Q-learning for playing flappy bird game | ||||||||||
| Deeplearningforfun | 335 | 7 months ago | mit | Scala | ||||||
| Implementation of some interesting ideas of deeplearning. | ||||||||||
| Flappybirdlearning | 44 | 7 years ago | JavaScript | |||||||
| Use deep learning to auto play flappy bird | ||||||||||
| Flappybird Es | 31 | 6 years ago | 1 | Python | ||||||
| An AI agent Learning to play Flappy Bird using Evolution Strategies and deep learning models. | ||||||||||
| Playing Custom Games Using Deep Learning | 22 | 4 years ago | Python | |||||||
| Implementation of Google's paper on playing atari games using deep learning in python. | ||||||||||
| Neuroevolution Flappy Bird | 15 | 5 years ago | mit | Jupyter Notebook | ||||||
| A comparison between humans, neuroevolution and multilayer perceptrons playing Flapy Bird implemented in Python | ||||||||||
| Deep Q Learning Flappy Bird | 6 | 7 years ago | mit | JavaScript | ||||||
| Flappybird Nueralnetwork | 4 | 6 years ago | Python | |||||||
| Using Deep Q Network to play Flappy Bird | ||||||||||
Deep Reinforcement Learning for Flappy Bird
This implementation explores Deep-Q-Learning with the core idea taken from the nature paper titled "Human-level control through deep reinforcement learning" link : http://www.nature.com/nature/journal/v518/n7540/full/nature14236.html?foxtrotcallback=true
The ideas for implementing Priority Sampling for the flappy bird game was taken from the paper "PRIORITIZED EXPERIENCE REPLAY" link : https://arxiv.org/pdf/1511.05952.pdf
The ideas for implementing Double-Deep-Q-Learning was taken from the paper "Deep Reinforcement Learning with Double Q-Learning" link : https://arxiv.org/pdf/1509.06461.pdf
Sincere gratitude to Yan Lau Pau from where the wrapper for the flappy bird pygame was taken and the ideas for the initial base implementation of Deep Reinforcement Learning using Keras. (https://yanpanlau.github.io/2016/07/10/FlappyBird-Keras.html)
Big thanks to the online blog nervana for assistance in helping to understand Deep Q Learning better. link : https://www.intelnervana.com/demystifying-deep-reinforcement-learning/
Instructions for running the application
- Clone the complete git directory. The code in Double DQN and Proirity sampling is meant for training the flappy bird. An example command to run the code would be:
python (filename.py)
-
The DQN.pycan be used in run mode and train mode. Use Train or Run to pass as arguements from the command line.
-
All implementations in training mode give random actions to the flappy bird for the first 3000 iterations in order to populate the experience replay memory, beyond which training starts.
Example command for running : python DQN.py Run
