This repository showcases the visual patterns that Convolutional Neural Networks (CNN) filters respond to, implemented in the PyTorch Deep Learning Framework. The inspiration for this project comes from François Chollet's insightful post, "Visualizing what convnets learn".
In the field of image classification, it is essential to understand what kind of visual patterns neural networks, particularly CNNs, learn during training. This repository uses the powerful ResNet50V2 model, pre-trained on the ImageNet dataset, to demonstrate the visual patterns that it has learned to recognize.
To visualize the patterns that CNN filters respond to, you can run the provided Jupyter Notebook. These files will guide you through the process of loading the pre-trained ResNet50V2 model and applying it to input images to observe filter responses.