Applications of Fast.ai Pretrained Models in Image Classification Problem

Abstract

Since the inception of Transformers and GPTs, artificial intelligence has proliferated. Fast.ai is among the cutting-edge libraries that are leading new advancements in the field. To harness the power of fast.ai and other advancements in the field, we set out to try and evaluate the practicality of the fast.ai library. To achieve this, we choose a given use case for artificial intelligence and then set out to fulfill said use case by leveraging fast.ai. We created three image classification models through fast.ai and then made an application that used those models. The use case we chose was a local wildlife fauna and flora classifier. The results from training the models were models with meager error rates, and these models had little to no data engineering. Key Terms ¾ computer vision, deep learning, fast.ai, image classification.