JointNET is a deep learning model designed to predict active inflammation in sacroiliac joints using radiographs. Developed using a dataset of 1,537 grade 0 SIJs, the model showcases superior accuracy compared to human observers. This repository contains the code used in the development and validation of JointNET.