This course is a complete introduction to the nearly magical art of designing images by the use of generative AI.
This course is a complete introduction to the nearly magical art of designing images by the use of generative AI.
Stable diffusion is one of the most powerful AI tools released by Stability AI and it provides a thorough basis for learning about generative AI generally, but also it can be used, with sufficient skill, in a production environment.
The course includes the following
•An Introduction to Stable Diffusion
•A guide to Installing Stable Diffusion using an Nvidia Graphics Card
•Understand the user interface
•Understanding Key Features
Key concepts learned include prompt construction, evalution and optimization.
Stable Diffusion is a latent diffusion model, a kind of deep generative neural network. Its code and model weights have been released publicly, and it can run on most consumer hardware equipped with a modest GPU with at least
The course explores options for users with less powerful equipment.
Stable Diffusion is entirely free and open-source, with no restrictions on commercial use. It is the most flexible AI image generator that you can even train your own models based on your own dataset to get it to generate exactly the kind of images you want
Students also learn where to get valuable resources like 3rd party checkpoints and models which can be used to improve the workflow and to provide creative freedom.
Stable Diffusion is a deep learning, text-to-image model that is primarily used to generate detailed images conditioned on text descriptions. It can also be applied to other tasks such as inpainting, outpainting, and generating image-to-image translations guided by a text prompt1. The main use cases of stable diffusion include:
Text-to-Image: the classic application where you enter a text prompt and Stable Diffusion generates a corresponding image.
Image-to-Image: tweak an existing image. You provide an image and a prompt and SD uses your image and tweaks it towards the prompt.
Inpainting: tweak an existing image only at specific masked parts.
Outpainting: add to an existing image at the border of it.
Understand course structure and content and core course aims.
What you need to complete the course.
The version of Stable Diffusion used in this course requires a discrete graphics processor.
To follow along with installation steps an Nvidia graphics card with at least 8GB of VRAM is required. I would recommend an entry level RTX 30 series Nvidia graphics card, the newer RTX 40 series Nvidia graphics cards, however earlier editions with 8GB of VRAM will also provide optimal performance.
Students who already understand how to set up Nvidia low VRAM installations or AMD installations of AUTOMATIC1111 can skip this part.
Towards the end of the course are some advices on how to set up and run on a low VRAM system.
This is the core learning for anyone starting Stable Diffusion.
By the end you will understand the origins of Stable Diffusion, the capabilites of this artificial intelligence
The strengths and weaknesses of the Latent Diffusion model
The advantages of AUTOMATIC1111
Students will learn how to install AUTOMATIC1111 and Xformers
Learners will be able to install an AUTOMATIC1111 instance of Stable Diffusion in a PC with an Nvidia GPU with 8GB or more of VRAM
Learners will be able to install the important Xformers Accelerator from Meta Research
The Dreamshaper model which is only a 2GB download may be used instead of the Runway checkpoint.
This is a very recent update showing a new way of Installing Automatic1111's WebUI in a standalone version which increases compatibility with other user interfaces for Stable Diffusion - for instance ComfyUI.
This is optional, the course can be completed without it, but learners may prefer to test out both installations to see which suits requirements best.
Core ideas to understand the functioning of Stable Diffusion including a preliminary look at prompts.
This lecture provides a deep understanding of the interplay between different parts of the user interface. This is introductory, but covers most of the aspects we shall utilize in this course.
Understand the role samplers / sampling methods in image generation. Samplers are a core parameter in operating stable diffusion.
Understanding the 'Maths' section in Stable Diffusion. Don't worry this is not a math lecture.
How to analyse, assess and optimize output, and a very brief outline of some of the advanced features in Stable Diffusion
A deep dive into some more advance methods and criteria in prompt construction.
Students will learn how to use English language instructions and prompt commands to produce optimal and reliable image outputs.
The benefits of using custom models are profound. This video examines how to access checkpoint and textual inversion models and how to use them with Stable Diffusion.
Automatic1111 can be used to install Controlnets, IPAdapters and T2I adapters.
This is a sketch outline on how these can be installed and deployed within Automatic 1111. This requires at a minimum version 1.6 of Automatic1111.
Generative Fill is a term used by Adobe to describe the process of using generative AI to remove or replace an object in Photoshop. This is possible in stable diffusion and we look at the process in this lecture.
Automatic1111 provides several options for working with Graphics Cards with less than 8GB of Memory. There may be times when you would want to use optimizations like this even when you have 8GB of VRAM as it may nonetheless prove a bottleneck with more complex workflows.. What are these methods and how well do they work?
Learn how to manually update your installation of Stable Diffusion.
This video demonstrates a technique for upgrading the installation Automatic1111 of Xformers. This isn't an essential technique. Some students have reported issues after upgrades, so you may want to skip this until you are more sure of your abilities to fix any issues you may encounter in your installation.
OpenCourser helps millions of learners each year. People visit us to learn workspace skills, ace their exams, and nurture their curiosity.
Our extensive catalog contains over 50,000 courses and twice as many books. Browse by search, by topic, or even by career interests. We'll match you to the right resources quickly.
Find this site helpful? Tell a friend about us.
We're supported by our community of learners. When you purchase or subscribe to courses and programs or purchase books, we may earn a commission from our partners.
Your purchases help us maintain our catalog and keep our servers humming without ads.
Thank you for supporting OpenCourser.