Never Give Up Save

PyTorch implementation of Never Give Up: Learning Directed Exploration Strategies

Project README

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

Open Source Agenda is not affiliated with "Never Give Up" Project. README Source: Coac/never-give-up
Stars
51
Open Issues
1
Last Commit
3 years ago
Repository
License
MIT

Open Source Agenda Badge

Open Source Agenda Rating