Skip to content

jbeck9/RNN-RL

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

11 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

RNN-RL

RNN+Transformer based reinforcement learning. Currently implemented for solving procedurally generated knapsack problems (https://en.wikipedia.org/wiki/Knapsack_problem)

Features:

  • Transformer-based input encoder.
  • GRU combines encodings with action history, using hidden layer for RL state representation.
  • Pytorch only DQL implementation. With memory buffer and double-Q implementation.
  • Bayesian Q-value output. Use MDN representation of Q-value for a richer understanding of expected reward. Could be used for custom exploration or inference strategies.

Documentation and code-cleaning work in progress.

About

RNN based reinforcement learning

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages