This repo contains the continuation of a proof of concept for a Machine Learning directed study course. The focus of the project is predicting fuel-type layers from satellite images. I've recently worked on automating some of the scripting using Approaching Almost Any Machine Learning Problem as a reference.
- You'll need the data... which I'm not at liberty to distribute (sorry)
- Next, you'll need to take that raw data and generate some training files in the form of csv (Description of config.py)
python src/create_folds.py --label <the label you want to create a model for> --folds <number of k folds (stratified)>