The Keras mobilenet_v2 model underwent training using over 10,000 images sourced from Kaggle's dog breed identification dataset, comprising 120 different classes. Its purpose was to accurately identify the breed of a given dog. For evaluation purposes, the model utilized the Multi Class Log Loss method, resulting in a score of 0.0167 after being trained on the complete dataset after preventing overfitting.
For each unique breed there's roughly more than 80 pictures and labels according to them.
