Flow me step by step

First Step

Install all the required packages:

  1. gym[atari]
  2. For windows user,you can install gym[atari] using the following command:

    pip install --no-index -f https://github.com/Kojoley/atari-py/releases atari_py
    pip install git+https://github.com/Kojoley/atari-py.git

    Thanks for Nikita Kniazev's solution on stackoverflow.

  3. pickle
  4. TensorFlow

Second Step

Third Step

Using your own parameters to train the model:

Number of hidden layer neurons:

Batch size to update:

Learning rate:

Discount factor for reward:

Decay factor for RMSProp leaky sum of grad^2:


Copy this line of code and run it at the position where script stored:

Resume from last training

You can stop training whenever you want and resume last training use:

python rlPong.py resume

CONTACT

Find some bugs? Please let us know!

Worcester Polytechnic Institute
100 Institute Road
Worcester, MA | 01609-2280
Hang Qi: hqi3@wpi.edu
Jin Huang: jhuang2@wpi.edu
Yuhan Liu: yliu26@wpi.edu
Yuxiang Bao: charlesbao95@outlook.com
;