This repo contains the code for a research project done in Harvard CS282r under the supervision of Dr. Ben Adlam titled "Finding Sparse, Trainable, and Accurate Subnetworks at Initialization".

A report detailing our work can be found here: https://drive.google.com/file/d/1ECOEXFoEhdP-RyygbKoVXa8SrGqGhADB/view?usp=sharing