Chrysolites Tech Academy

Video by freeCodeCamp.org · streamed from YouTube. All rights remain with the creator.

Lesson 3: Keras with TensorFlow

About this lesson

This course from deeplizard teaches deep learning with Keras, the neural network interface that is built into TensorFlow, and assumes basic programming skills and some experience with Python. It begins with an introduction to Keras and then shows how to process data into the form a neural network expects, before building a simple artificial neural network and training it. You learn how to use a validation set to watch for overfitting during training, how to run inference on new data and judge the results, and the different ways a model can be saved and loaded again later. The next part moves to images with convolutional neural networks. It covers how image data is organised and processed, how a convolutional network is built and trained from scratch, and how the trained network is used for inference. The course then introduces fine-tuning and transfer learning, starting with the well-known VGG16 model, fine-tuning it for a new image task and using the result for predictions. It continues with MobileNet, a much smaller network designed to run efficiently, and fine-tunes it on a small subset of data. The final topics cover data augmentation, which creates extra training images from existing ones, how Keras assigns labels to images, how to achieve reproducible results from run to run, and how to count the learnable parameters in a network. Afterwards, read Chapter 3 of the course book, which counts parameters for fine-tuning by hand.

Install Chrysolites Tech Academy

Android / Chrome / Edge: open the browser menu and choose Install app or Add to Home screen.

iPhone / iPad (Safari): tap the Share button, then Add to Home Screen.