This repository contains the source code used for performing the experiments described in this story: Transfer Learning in Image Classification: how much training data do we really need?
In Image Classification problems, how much training data do we actually need when adopting a Transfer Learning approach?
In this repo, I try to answer this question by applying two Transfer Learning techniques (Feature Extraction and Fine-Tuning) for addressing a simple Image Classification task, varying the number of examples on which the models are trained to see how the lack of data affects the effectiveness of the adopted approaches.