Introduction to Stable Diffusion for Developers & Designers

Name: Introduction to Stable Diffusion for Developers & Designers
Rating: 4.3 (54 reviews)

Expertly crafted course for Technical and Non-Technical audience Introducing GenerativeAI Text2Image Stable Diffusion

Created byMemexAI Technology, Amir Nagri

Last updated 4/2023

English

What you'll learn

Understand the Stable Diffusion Technology in-depth - diffusion models, de-noising iterative process, text2image, image2text CLIP technology and origins
Understand Prompt Engineering - Prompt Anatomy, Negative Prompts, Prompt Building using Iterative + Experimentation Approach, Prompt Examples and Resources
Understand Stable Diffusion Parameters and Fine Tuning - Classifier Free Guidance (CFG), Steps, Sampler, Seed, Image Dimensions
Understand Image2image Technology using Stable Diffusion - Image2image, Style Fusion, In-Painting including Add, Remove and Replace, Out-painting, Upscaling
Generate Images and Artwork using Stable Diffusion given a brief, 10+ hands-on guided labs, 15+ briefs, generating over 100+ images during the complete course

Course content

5 sections • 33 lectures • 1h 30m total length

1.1 - Welcome1:17
In this lecture video, i give you a quick background about me, my experience with Generative AI, Stable Diffusion and AI Startup based on Stable Diffusion. I also give a quick background story of how this video series came about.
1.2 - Why should you be interested in GenerativeAI Technology Stable Diffusion?2:24
In this video, I cover a few of the most popular examples of successful products built using Stable Diffusion and their success stories. Also touch upon how innovative, creative and powerful this technology is, yet so simple and accessible that indiehackers and solopreneurs are creating creative applications using it and competing with large corporations.
1.3 - Introduction to Stable Diffusion - Curriculum2:29
In this video lecture, I cover the course curriculum. This course is designed as an introductory course targeted to both technical and non-technical audience. We do dive deeper, but at the level of more than a non-technical AI Artist, but less than a fully technical Data Scientist.

The course curriculum starts with the understanding of diffusion model and its architecture. Then throughout the course, each of the feature and capabilities of Stable Diffusion AI model is tied back to this architecture. This allows you to understand the technology and its capabilities in and out.

The course curriculum covers Prompt Engineering in-depth, Parameter Tuning and Image2image with all its variants. Check out the course outline to get a good idea of how the course is structured.
1.4 - Stable Diffusion Course Series3:05
Covering Stable Diffusion in one-single course is going to be overwhelming. I will also not be able to target the non-technical audience if I make a single super technical course. So I have divided it into logical chunks.

In the intermediate course series, we use the advance Automatic1111 web UI to do the advance functionalities it allows us to do. We cover the Automatic1111 web UI in full depth. We have special focus on covering Controlnet, its architecture and multiple use cases it enables. Other special focus is on model fine tuning, understanding it in depth, the various options we have for model fine tuning - textual inversion, controlnet, lora, dreambooth, text2image, checkpoint merger, and the pros and cons of each of the approach.
In the advance course, we finally use python to hand-code whatever we have accomplished using the PlaygroundAI and Automatic1111 web UI. This enables us to build our own applications and deploy it to stateful and serverless GPUs.
In the application course, I demo 15+ applications possible through Stable Diffusion technology, also provide the complete source code. This gives you deep insights into developing your own creative applications.
1.5 - Tools and Setup - PlaygroundAI0:34
In this video, I cover the tools we will use in this course. We use PlaygroundAI. Later in the lectures, I tell you why I chose PlaygroundAI over other web UIs.
1.6 - Tools and Setup - Miro2:15
In this video, I introduce you to the command center for the course built using miro. This board has the complete bird eye view of all the activities and resources needed for this course.
1.7 - Tools and Setup - Lab Guides using tango.us2:13
In this video, I introduce you to tango.us. This tool allows users to create a step-by-step lab guides that are perfect for a hands-on course like ours. Each step contains a screenshot of the relevant screen, and the text input needed for that step. I tell you the basics of the lab guide, how to navigate, how to find the inputs and controls needed for that step, how to find the inputs to paste in the input box etc.
All labs will be driven using the Tango lab guides.

2.1 - What is Stable Diffusion?2:26
In this video, I briefly explain the history of Stable Diffusion and how it came to be. I also dive into the diffusion model architecture and mechanism. Also explain the iterative de-noising process using visual animated course material. I connect back various input parameters that you are familiar with to the Stable Diffusion AI model architecture.
2.2 - PlaygroundAI5:08
In this video, I explain the main components of PlaygroundAI, and how each of those inputs connect back to the Stable Diffusion model architecture that we covered in previous lecture. I explain why I chose PlaygroundAI over other Web UIs.
2.3 - PlaygroundAI - Lab Guide Demo4:06
In this video, I kick off the image generation using PlaygroundAI. I also demo how to use the tango lab guide. After this video lecture, you should be set to use the lab guides and try it out by yourselves. Feel free to come back to this video lecture if you have any confusion on using the lab guides.
2.4 - Prompt Engineering - Introduction0:41
In this video, I kick off the Prompt Engineering section and various lectures that we are going to cover as part of this course.
2.5 Prompt Engineering - CLIP or Image2Text1:13
In this video, I cover how text2image models like Stable Diffusion came about to be, and how they are related to image2text models like CLIP. I also touch upon what is the core or our goal with prompt engineering.
2.6 - Prompt Engineering - Anatomy0:55
In this video, I cover Prompt Anatomy, the discovered structure for a good prompt. I cover one of the example of Prompt Anatomy, and how we can follow this structure for the rest of the course.
2.7 - Prompt Engineering - Iterative Process1:43
In this video, I cover the iterative process of building a prompt. I cover why the iterative process is needed, and how to go about it.
2.8 - Prompt Engineering - Iterative Process Demo11:42
In this video, I demo the iterative process covered in the previous lecture.
2.9 - Prompt Engineering - Negative Prompt2:23
In this video, I cover the negative prompts, briefly touch upon how they work, and provide mechanism to use the negative prompts in PlaygroundAI and elsewhere.
2.10 - Prompt Engineering - PlaygroundAI Filters1:05
In this video, I cover the Filters feature of PlaygroundAI and its underlying mechanism.
2.11 - Prompt Engineering - Miscellaneous2:19
In this video, I cover all the miscellaneous but important concepts related to Prompt Engineering, that are so brief that they don't require their own dedicated video.
2.12 - Prompt Engineering - Labs Overview1:20
In this video, I give you an overview of the labs associated with this section on Prompt Engineering.

4.1 - Image2image - Introduction1:05
In this video, we will dive into the underlying mechanism of image2image. We will cover how you trick Stable Diffusion model by passing your own starting image, rather than a random noisy image to influence the output. Using this mechanism as base, we will explore different techniques popular in image2image.
4.2 - Image2image - Create Variants2:03
In this video, I will cover how you can do creative exploration using the Create Variants feature of Stable Diffusion, also cover how this feature works under the hood.
4.3 - Image2image - Style Fusion1:19
In this video, I will cover how you can fuse different styles in the same image using the Style Fusion technique of Image2image.
4.4 - Image2image - Variants + Style Fusion Demo4:35
In this video, I demo the Create Variants and Style Fusion feature, and you can combine both to generate your artwork.
4.5 - Image2image - Doodle to Painting Demo8:27
In this video, I will cover how you can convert your doodle into read-to-use painting, and how this feature works under the hood.
4.6 - In-painting - Object add, remove, replace3:54
In this video, we cover in-painting, how in-painting works under the hood, and how you can use this technique to add, remove or replace objects in your artwork.
4.7 - Edit Image or Instruct Pix2pix0:36
In this video, I cover instruct pix2pix, the underlying paper, and how you can use command like syntax to edit your images.
4.8 - Out-painting Demo5:48
In this video, I cover outpainting, and how you can use PlaygroundAI to try out outpainting and create large artworks.
4.9 - PlaygroundAI - Face Restoration and Upscale 4x1:50
In this video, I cover the post-processing steps available in PlaygroundAI, and how you can use them to finalize your image for production use.
4.10 - Image2image - Lab Overview0:48
In this video, I give an overview of the lab section of image2image, various lab guides available, and sample output that I got when I followed those lab guides.

Requirements

The course is designed for both technical and non-technical learners. No programming experience or any experience with Stable Diffusion is required

Description

Welcome to Introduction to Stable Diffusion Course.

Quick intro about the trainer, I have 17+ years of experience in IT, with last few years at the leadership positions of Decacorns and one of the largest product companies in South Asia. Past 2 years, I have immersed myself in GenerativeAI, Large Language Models, and for past 6 months in Stable Diffusion.

In the past 6 months, I have won various hackathons themed on GenerativeAI and Stable Diffusion. I have launched a GenerativeAI startup, as well as delivered an enterprise grade hands-on online workshop on Stable Diffusion. I have compiled all my learnings, specially last 2 years on GenerativeAI, and past 6 months on Stable Diffusion in this online course.

In this course, you are going to learn in-depth, hands-on on the topic of Stable Diffusion. This course is expertly crafted, and targeted at both technical and non-technical audience. We dive deep into the topic at more than an AI Artist level, but less than a Data Scientist level. You will understand the ins-and-outs of Stable Diffusion technology, and will be able to tie back the controls and parameters of how the AI model works under the hood.

We cover main applications of Stable Diffusion, including generating art, prompt engineering, negative prompts, iterative prompt engineering with experimentation, parameter tuning - cfg, steps, seed, dimensions, sampler, image2image applications including - variations, doodle to art, in-painting (addition, removal, replacement), out-painting, instruct pix2pix, face restoration and upscaling.

You are going to have 10+ lab guides, generating 100+ art works during the duration of the course, gaining 100% confidence to start producing production quality artwork and image assets by using the tips and techniques introduced in this course.

The course also gives you pointer and guidance on exploring the technology deeper and gain even deeper understanding.

Best wishes and Happy Learning.

Who this course is for:

Course is designed for solopreneurs and indiehackers who are planning to launch their own GenerativeAI startup using Stable Diffusion technology
Course is for anyone curious about Technology underlying Text2Image tools like Stable Diffusion, Dall-E, Midjourney
Course is for Designers who want to get started with GenerativeAI Text2Image technology and create production quality artworks Stable Diffusion and similar technologies
Course is designed for Technical Audience like Developers, Machine Learning Engineers, Data Scientist who are looking to be introduced to GenerativeAI Text2Image technology like Stable Diffusion

Introduction to Stable Diffusion for Developers & Designers

What you'll learn

Explore related topics

Course content

Introduction7 lectures • 14min

Prompt Engineering12 lectures • 35min

Parameter Tuning2 lectures • 9min

Image2image10 lectures • 30min

Conclusion2 lectures • 2min

Requirements

Description

Who this course is for: