Machine Learning in Elixir

Chapter 8Stop Reinventing the Wheel

In the previous chapter, you implemented a new type of neural network: the convolutional neural network. You built and trained a convolutional neural network to accurately classify images of cats and dogs. After applying a few additional model training tricks, such as data augmentation and dropout, you were able to train a model to classify an image as a cat or a dog with 87% accuracy. While 87% accuracy is great, you can do even better still—with minimal changes to your training pipeline—by leveraging the power of pre-trained models and transfer learning. In this chapter, you’ll discover what transfer learning is, when it’s necessary, and how to perform transfer learning with Axon.