This repository consists of jupyter notebooks which show experiments being done on 1-10% of the food vision dataset, to find the optimum model design. This is later progressed into a full model which is trained on the entire dataset, in the final version an augmented layer with pre trained efficientnet weights is used which has the last 5 layers unfrozen and set to trainable.