100 Days of Deep Learning
Day 1 - The history of neural networks
Day 4 - Regression and Classification using MLPs
Day 5 - Using the Keras Sequential API
Day 6 - Using the Keras Functional API
Day 7 - Building Dynamic Models with the Keras Subclassing API
Day 8 - Saving and Restoring Models
Day 9 - Using Callbacks in Keras
Day 10 - Using TensorBoard for Visualization
Day 11 - Fine-Tuning Neural Network Hyperparameters
Day 12 - Vanishing and Exploding Gradients
Day 13 - Glorot & He Initialization
Day 14 - Nonsaturating Activation Functions
Day 17 - Reusing Pretrained Layers
Day 18 - Momentum Optimization
Day 19 - Nesterov Accelerated Gradient
Day 20 - Adaptive Learning Rates
Day 21 - Learning Rate Scheduling
Day 22 - Preventing Overfitting using Regularization
Day 23 - Custom Losses in TensorFlow
Day 24 - Custom Activation Functions, Initializers, Regularizers and Constraints
Day 26 - Custom Layers in TensorFlow
Day 27 - Custom Models, Losses and Metrics in TensorFlow
Day 28 - Using Autodiff in TensorFlow
Day 29 - Custom Training Loop in TensorFlow
References
- Géron, Aurélien. Hands-On Machine Learning with Scikit-Learn and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems. United States, O'Reilly Media, 2017.