This repository contains a semantic segmentation model trained on the FloodNet dataset. The model is capable of classifying each pixel in an image into one of 10 predefined classes. The project utilizes the Dice Score as the evaluation metric.