For this project, I made a cats and dog classifier using Convolution Neural Network. Specifically I used:
Conv2D
Layer - Kernel size of 3x3 andrelu
as activation function. In total 3 such layers were used.MaxPooling2D
- Max-pooling is performed over a 2 × 2 pixel window. In total 3 such layers were used.Dropout
Layer - A dropout of 20% is applied with a larger dropout rate of 50% applied after the fully connected layer in the classifier part of the model.
After this I used Flatten
and Dense
layer to classify the input in two available classes (Cats and Dogs).
For training and validation part I used: https://www.kaggle.com/chetankv/dogs-cats-images.
And for testing part a random dataset containing cats and dogs images in no specific order.