Skip to content

devanshjain7/dqn-cartpole

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 

Repository files navigation

Deep RL

Overview

  • This project aims at implementing and evaluating Q-Learning on a simple control task - the CartPole environment.
  • The Q-learning algorithm and the nuts and bolts of a Deep Q-Network have been coded up and can be referred from dqn.py.
  • You do not require a GPU (and hence, AWS) for this task; the DQN can be trained easily within a few minutes on a decent CPU.
  • While each of the runs won't take much time, depending on your implementation, you may have to play around with the parameters like learning rate, epsilon-annealing, batch size, hidden layer sizes, etc. which could take up the bulk of your time.
  • The performance of the Q-network using a different set of hyperparameters is summarized in the report.

Installation

  • Install TensorFlow for your machine following the documentation. We recommend using the Virtualenv installation since it provides a virtual and isolated Python environment, incapable of interfering with or being affected by other Python programs, versions, and libraries being used by your other projects on your machine.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages