Course book
Deep Learning with TensorFlow and Keras
Four illustrated chapters, one for each video course. They explain the ideas in original words and work the shapes, parameter counts and calculations by hand.
- Chapter 1TensorFlow from Tensors to Q-LearningAI, ML and neural networks, tensor ranks and shapes, core algorithms with a hidden Markov model worked by hand, Keras networks, CNNs, RNNs and a Q-learning update.
- Chapter 2Image and Text Classifiers in KerasThe Keras workflow, inspecting and scaling image data, predictions with argmax, encoding and padding text, embeddings and pooling, parameter counts, saving models and GPUs.
- Chapter 3Keras in Practice: From Data to Fine-TuningProcessing and scaling data, validation sets, inference and confusion matrices, three ways to save, CNNs, fine-tuning VGG16 and MobileNet, augmentation and parameter counts.
- Chapter 4Classifying Text with Neural NetworksDataset anatomy and splits, loss functions worked by hand, scaling and oversampling, a feedforward network, RNN hidden states by hand, and text models with embeddings and LSTMs.