Section: Deep Learning•ID #48

Stochastic Gradient Descent (SGD)

Step 1: TheoryTheory Article

SGD Optimization in Neural Nets

Study the core concepts and background material for this learning item.

Step 2: PracticeKaggle Workbook

Stochastic Gradient Descent Exercise

Solve the challenges and apply your knowledge interactively.

Medium20 min estimated study

Overview

Train networks using mini-batch gradient updates.

Learning Objectives

  • Differentiate batch sizes
  • Tuned SGD learning rates

Prerequisites

Locked

Tracking Control

Completion Reward+200 XP
Study Checklist
Studied theory resource
Completed practice exercise