Skip to content

Latest commit

 

History

History
25 lines (20 loc) · 664 Bytes

File metadata and controls

25 lines (20 loc) · 664 Bytes

Snake AI with Reinforcement Learning

Teaching myself how to train AI to play Snake Game

How to Run

python agent.py

Demo

Running the app:

Gameplay

Tech Stacks

  • Python/Pygame (Game)
  • Agent (for ML training)
  • Pytorch (Model)

Lesson & Challenges

  • Working with Pytorch and training a model is pretty challenging given my skillset (for now).
  • Training can take super long (which is expected).
  • Agent logic is fairly straight forward for a snake game so implementation was pretty easy.
  • Learning Pytorch framework takes times and I love it.