How to Build a Diffusion Model - An Introduction

Fractal Analytics

How to Build a Diffusion Model - An Introduction

Fractal Analytics Academy

Akshesh Shah

Instructors: Fractal Analytics Academy

Access provided by The MasterMind Cooperative

Gain insight into a topic and learn the fundamentals.

Intermediate level

Recommended experience

1 week to complete

at 10 hours a week

Flexible schedule

Learn at your own pace

Gain insight into a topic and learn the fundamentals.

Intermediate level

Recommended experience

1 week to complete

at 10 hours a week

Flexible schedule

Learn at your own pace

What you'll learn

Explain the core concepts of diffusion models and their role in the generative AI landscape.
Design diffusion models from scratch using effective training strategies.
Build text-to-image generation systems leveraging advanced techniques.
Evaluate model performance using real-world metrics.

Skills you'll gain

Tools you'll learn

Details to know

Shareable certificate

Add to your LinkedIn profile

Assessments

7 assignments

Taught in English

See how employees at top companies are mastering in-demand skills

Learn more about Coursera for Business

logos of Petrobras, TATA, Danone, Capgemini, P&G and L'Oreal

There are 3 modules in this course

Explore the fascinating world of generative models with a deep focus on diffusion models for high-quality image generation. You’ll begin by mastering the core principles of diffusion and then advance to the architectures that power modern text-to-image systems. Learn how these models transform random noise into stunning visuals through forward and reverse processes, and discover optimization techniques using loss functions and training strategies.

Explore the fundamentals of deep learning and generative models. Understand the diffusion process, its types, and applications in AI.

What's included

7 videos3 readings3 assignments

7 videos Total 25 minutes

Course Introduction 3 minutes
Fundamentals of Deep Learning 5 minutes
Introduction to Generative Models 3 minutes
Overview of Diffusion Process 3 minutes
Why Diffusion Models? 4 minutes
Types of Diffusion Models 3 minutes
Applications and Use-cases 4 minutes

3 readings Total 45 minutes

Course Syllabus 5 minutes
Diffusion Explainer 10 minutes
Recent Buzz and Advancements about Diffusion Models 30 minutes

3 assignments Total 35 minutes

Diffusion 101: A Primer on Generative Models 15 minutes
Understanding Diffusion Models 10 minutes
Starting with Diffusion Models 10 minutes

Learn the architecture and mechanics of diffusion models. Dive into forward/reverse passes, loss functions, and training strategies.

What's included

29 videos5 readings4 assignments5 ungraded labs

29 videos Total 138 minutes

Steps in Building a Diffusion Model 4 minutes
Architecture Design - Forward Pass Part-1 5 minutes
Architecture Design - Forward Pass Part-2 4 minutes
Architecture Design - Reverse Pass Part-1 4 minutes
Architecture Design - Reverse Pass Part-2 4 minutes
Tailoring the Architecture to Data Types 4 minutes
Handling Multimodal Architectures 4 minutes
Text Data Preparation 4 minutes
Image Data Preprocessing 5 minutes
Preparing the Dataset 3 minutes
Advance Techniques for Image Data Preparation 5 minutes
Data Integration and Fusion 5 minutes
Data Preprocessing 2 minutes
Training and Optimization 6 minutes
Building the Model 3 minutes
Learning Rate Scheduling 5 minutes
Regularization Techniques 5 minutes
Understanding Loss Functions 5 minutes
Optimization Algorithms 5 minutes
Gradient Clipping 5 minutes
Hyperparameter Tuning 6 minutes
Training the Model 5 minutes
Importance of Model Evaluation 6 minutes
Basic Evaluation Metrics for Diffusion Models 6 minutes
Qualitative Evaluation Metrics for Diffusion Models 6 minutes
Evaluating the Model 3 minutes
Bootstrapping Validation Technique 5 minutes
Validation using Ensemble Methods 5 minutes
Classical Validation Technique 6 minutes

5 readings Total 65 minutes

Flow-based Architecture 10 minutes
Data Preparation 10 minutes
Data Splitting and Partitioning 15 minutes
Batch Training vs Online Training 15 minutes
Understanding Validation Strategies 15 minutes

4 assignments Total 45 minutes

Transforming Data into Art 10 minutes
Mastering the Fundamentals 10 minutes
Performance Proofs: Validating Diffusion Models 15 minutes
Review and Refinement 10 minutes

5 ungraded labs Total 225 minutes

How to Prepare Data Set 45 minutes
How to Handle Data Preprocessing 45 minutes
Building Your Diffusion Model 45 minutes
Training Your Model 45 minutes
Model Evaluation 45 minutes

Build end-to-end text-to-image systems. Cover data preparation, model construction, training, evaluation, and hands-on labs.

What's included

1 video3 readings1 peer review

Instructors

Fractal Analytics Academy

Fractal Analytics Academy

Fractal Analytics

23 Courses 124,832 learners

Offered by

Fractal Analytics

Why people choose Coursera for their career

Felipe M.

Learner since 2018

"To be able to take courses at my own pace and rhythm has been an amazing experience. I can learn whenever it fits my schedule and mood."

Jennifer J.

Learner since 2020

"I directly applied the concepts and skills I learned from my courses to an exciting new project at work."

Larry W.

Learner since 2021

"When I need courses on topics that my university doesn't offer, Coursera is one of the best places to go."

Chaitanya A.

"Learning isn't just about being better at your job: it's so much more than that. Coursera allows me to learn without limits."

Explore more from Data Science

Coursera
Fine-tuning Image Models with Diffusion
Course
Google Cloud
Introduction to Image Generation
Course
DeepLearning.AI
How Diffusion Models Work
Project
Microsoft
Core generative models and techniques
Course