Skip to content

Deepest-Project/agent57_from_ngu_pytorch

Repository files navigation

Never give up

PyTorch implementation of Never Give Up: Learning Directed Exploration Strategies [Paper]

Only the episodic curiosity with embedding network has been implemented.

Install

Tested with Python 3.7.9

pip install -r requirements.txt

Train

python train.py

Results

Results on 5x5 gym-maze

reward

Credits

R2D2 base is from here by g6ling

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •  

Languages