Repository for SP 2023 CSCI 567's Course Project.
Our course project is to compete in a Machine Learning competition: DrivenData's Richter's Predictor: Modeling Earthquake Damage.
You can find out more about the competition here: https://www.drivendata.org/competitions/57/nepal-earthquake/page/135/
Our submission achieves an F-1 Micro score of 0.7531, placing us at rank 60 in the world at the time of writing: May 5, 2023.
To run our code, download all the csv files found at the URL below and place them in your /data directory.
https://www.drivendata.org/competitions/57/nepal-earthquake/data/
Library requirements and their versions can be found at the head of the SOTA notebook.
Pretrain autoencoder model can be found here