Skip to content

A TensorFlow distributed implementation of DeepMind's PathNet architecture for RL training with OpenAI Gym agents. Modified from jaesik817's implementation.

Notifications You must be signed in to change notification settings

adit-chandra/pathnet-gym

Repository files navigation

PathNet for OpenAI Gym Environments

Overview

A TensorFlow distributed implementation of DeepMind's PathNet architecture. Implementation is currently specific to training OpenAI Gym agents. Modified from jaesik817's implementation.

Work in Progress!

Network has shown promising performance for select Atari environments. Continuing to develop on Doom environment and planning to expand soon compatibility with other environments.

-Adit

About

A TensorFlow distributed implementation of DeepMind's PathNet architecture for RL training with OpenAI Gym agents. Modified from jaesik817's implementation.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published