The Intel Image Classification task is a multi-class image classification problem. The dataset contains 24,940 images of 6 different categories: buildings, forest, glacier, mountain, sea, and street. The goal of the task is to build a deep learning model that can classify these images accurately.
We will be training a densenet121 architecture to achieve state of art results. Densenet 121 is a neural network architecture that was introduced in the paper Densly Convolutional Neural Networks in 2016. For more information about the densenet architecture you can refer the paper at https://arxiv.org/abs/1608.06993. The intel image classification dataset can be accessed from the link https://www.kaggle.com/datasets/puneet6060/intel-image-classification.
