Simple CNN Classifier

This is a very simple repo for explaining basic concepts about Convolutional Neural Networks (CNNs) to beginners. The example exploits the PyTorch library (https://pytorch.org/) for performing a basic binary classification task on the Kaggle Dogs vs. Cats dataset (https://www.kaggle.com/c/dogs-vs-cats/data).

Preparing the dataset

1. Download the dataset from Kaggle
2. Use the script prepare_dataset.py in order to split train samples in two folders.

Training

Use the script train.py in order to train a CNN for this purpose.

Validation

Use the script eval.py in order to evaluate the CNN performance on this task.