Cat-And-Dog

Introduction
Data for this project from cat and dog competition on Kagle, which I download directly to Google Colab.
Using SeparableConv2D, BatchNormalization, GlobalAveragePooling2D, and data augmentation builds a classifier whose accuracy is 0.96.

Sample Images

Augmentation Images

Evaluation

Test accuracy: 0.963 Test loss: 0.0958