Forest fires pose a significant threat to ecosystems, wildlife, and human lives. Given their devastating consequences, the ability to predict forest fires plays a crucial role in mitigating their impact. Traditional methods have long been employed to forecast these fires, but with the advent of machine learning, we now have the potential to enhance our predictive capabilities significantly. In this workshop, we will walk through how to perform image analysis to predict potential forest fire likelihoods based on regions of known forest fires acquired via the MODIS (Moderate Resolution Imaging Spectroradiometer) dataset. We will use the Intel® Extension for PyTorch*, powered by oneAPI, to optimize and accelerate PyTorch-based model finetuning, whereby the pre-trained ResNet model is adapted to aerial photos. We will walk through utility functions, a trainer class, a model class, and a metrics class. Lastly, we will create a final confusion matrix to review the model accuracy.