Flower classification using CNNs.
The aim of this exercise is to code a CNN that provides good results on a classification task. The idea is to test different tools and to measure their impact on the provided dataset.
All code can be found inside flower-classification.ipynb. Data used is a subset of popular flower datasets.
This project was developed as an exercise for my Deep Learning class in my master program IMLEX, at Jean Monnet University.
This dataset is constituted by 10 categories of flowers.
It contains 800 color images: 600 (60 per category) are in the training set and 200 (20 per category) in the validation set. The train/validation split is provided.
The images have different sizes, so it is required to resize them to 128x128.