We are building a model for doing breed classification of a dog by just looking into its image and tracking the accuracy this model might be able to achieve. With modern machine learning frameworks like TensorFlow and pre-trained models for image recognition the task could be solved with a pretty good accuracy without applying too many efforts and spending too much time and resources. I am going to share the end-to-end process of building dog breed classifier using TensorFlow. We are going through image preprocessing with specific techniques (e.g. whitening, equalisation and possibly modification of image size), data augmentation (mirroring, cropping…), CNN from scratch, CNN using InceptionResNet architecture, and Transfer Learning (InceptionResNet, VGG19, Xception, Resnet50, InceptionV3). The data used for the application comes from Stanford Dogs Dataset. This project could be found on Github