johnlime / spring_creatures_experiment

A virtual creature environment trained with NEAT.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Spring Creatures

An virtual creature environment trained with NEAT made for CS/CSYS 352: Evolutionary Computation by Prof. Nick Cheney at UVM.

Uses NEAT-Python library.

Environment Settings

Create Anaconda environment.

conda env create -f conda_env/mac-11-6-1.yml
conda activate spring_creatures

Add working directory to PYTHONPATH.

export PYTHONPATH=$PWD

Run sanity check to see if the Spring Creatures environment runs.

python run/run_sanity.py

Visualize and play with the Spring Creatures environment using Pygame.

python run/run_pygame.py

Run NEAT on Spring Creatures

Run evolutionary algorithm on the Spring Creatures environment (overwrites winner file).

python run/run_train.py

Run and visualize the final solution.

python run/run_test.py

Video Demonstration

Video

Collaborators

NEAT-Python implementation
marszzibros

About

A virtual creature environment trained with NEAT.


Languages

Language:Python 100.0%