Week 4: Deep Learning Modules

Modules

Overview: This week introduces you to deep learning with PyTorch and building image classifiers.

Notes

Module 1: PyTorch Fundamentals (2 Days)

Session 1: Introduction to PyTorch and Neural Networks
Session 2: The ML Pipeline and Building Your First Model
Session 3: Activation Functions
Session 4: Working with Tensors
Lab 1: Building a Simple Neural Network
Lab 2: Modeling Non-Linear Patterns with Activation Functions
Lab 3: Tensors: The Core of PyTorch
Assignment: Deeper Regression, Smarter Features

Module 2: Image Classification with PyTorch (1 Day)

Session 1: Data and Model Building
Session 1.5: From Equations to Vectors (Optional)
Session 2: Loss Functions and Optimizers
Session 3: Device Management and Image Classification Setup
Session 4: Training and Evaluating Your Classifier
Lab 1: Building Your First Image Classifier
Assignment: EMNIST Letter Detective

Module 3: Data Management (0.5 Day)

Lab 1: Data Management
Assignment: Building a Robust Data Pipeline

Module 4: Core Neural Network Components (1 Day)

Session 1: Convolutional Neural Networks
Session 1.5: Transfer Learning
Session 2: PyTorch Techniques and Model Inspection
Lab 1: Building a CNN for Nature Classification
Lab 2: Model Debugging, Inspection, and Modularization
Assignment: Overcoming Overfitting: Building a Robust CNN

IMPORTANT Notes to Presenter

Press S for speaker notes.

Press M for outline view or O for slide overview.

Resources