Stable Diffusion AI: The Ultimate Beginner’s Guide to Text to Image Transformation

Want to learn Stable Diffusion AI? This beginner’s guide is for newbies with zero experience with Stable Diffusion or other AI image generators. You will get an overview of Stable Diffusion and some basic useful tips.

And you will get a chance to try out Stable Diffusion right on this page. It’s the best way to learn!

This is the first part of the beginner’s guide series.
Read part 2: Prompt building.
Read part 3: Inpainting.
Read part 4: Models.

What is Stable Diffusion?

Stable Diffusion AI is a latent diffusion model for generating AI images. The images can be photorealistic, like those captured by a camera, or in an artistic style as if produced by a professional artist.

The best part is that it is free – you can run it on your PC.

How to use Stable Diffusion AI?

You need to give it a prompt that describes an image. For example:

gingerbread house, diorama, in focus, white background, toast , crunch cereal

Stable Diffusion turns this prompt into images like the ones below.

You can generate as many variations as you want from the same prompt.

What’s the advantage of Stable Diffusion?

There are similar text-to-image generation services like DALLE and MidJourney. Why Stable Diffusion? The advantages of Stable Diffusion AI are

  • Open-source: Many enthusiasts have created free tools and models.
  • Designed for low-power computers: It’s free or cheap to run.

Is Stable Diffusion AI Free?

Stable Diffusion is free to use when running on your own Windows or Mac machines. An online service will likely cost you a modest fee because someone needs to provide you the hardware to run on.

Try the Stable Diffusion online demo

The best way to understand Stable Diffusion is to try it out yourself.

Try the Stable Diffusion image generator below. Here are the four easy steps.

  1. Close your eyes.
  2. Imagine an image you want to make.
  3. Describe the image in words as much detail as possible. (For the best result, be sure to cover the subject and the background and use a lot of describing words)
  4. Write it in the prompt input box below.

You can leave the negative prompt unchanged.

Here’s a list of simple examples of prompts you can try.

a cute Siberian cat running on a beach

a cyborg in style of van Gogh

french-bulldog warrior on a field, digital art, attractive, beautiful, intricate details, detailed face, hyper-detailed closed eyes,zorro eye mask, artstation, ambient light

Switch the model to see the effect. You will learn about models in the last part of this guide.

  • Stable Diffusion v1.5: The official base model. Versatile in all styles.
  • Realistic Vision v2.0: Excel in generating photo-style realistic images.
  • Anything v3.0: Anime style.

You may have noticed that the images could be hit or miss. Don’t worry; there are many ways you can improve the images. Read on.

What Can Stable Diffusion Do?

1. Generate images from text

The most basic usage of Stable Diffusion is text-to-image (txt2img). Here are some examples of images you can generate with Stable Diffusion.

Anime style

Stable Diffusion beginner's guide

Photorealistic style

Learn how to generate realistic people and realistic street humans.

Landscape

Fantasy

realistic princess.

Artistic style

Animals

Learn how to generate animals.

2. Generate an image from another image

Image-to-image (img2img) transforms one image to another using Stable Diffusion AI.

Below is an example of transforming my drawing of an apple into a photo-realistic one. (Tutorial)

Image-to-image generates an image base on an input image and a prompt.

3. Photo Editing

You can use inpainting to regenerate part of an AI or real image. This is the same as Photoshop’s new generative fill function, but free.

4. Make videos

There are two main ways to make videos with Stable Diffusion: (1) from a text prompt and (2) from another video.

Deforum is a popular way to make a video from a text prompt. You have probably seen one of them on social media. It looks like this.

The second way is to stylize a video using Stable Diffusion. See the video-to-video tutorial.

Original
Animation

This is a more advanced topic. It is best to master text-to-image and image-to-image before diving into it.

How do you use Stable Diffusion AI?

Online generator

For absolute beginners, I recommend using the free online generator above or other online services. You can start generating without the hassle of setting things up.

Advanced GUI

The downside of free online generators is that the functionalities are pretty limited.

Use a more advanced GUI (Graphical User Interface) if you’ve outgrown them. A whole array of tools are at your disposal. To name a few:

AUTOMATIC1111 is a popular choice. See the Quick Start Guide for setting up the Google Colab cloud server. Running it on your PC is also a good option if you have the right PC. See install guides for Windows and Mac.

How to build a good prompt?

There’s a lot to learn to craft a good prompt. But the basic is to describe your subject in as much detail as possible. Make sure to include powerful keywords to define the style.

Using a prompt generator is a great way to learn a step-by-step process and important keywords. It is essential for beginners to learn a set of powerful keywords and their expected effects. This is like learning vocabulary for a new language. You can also find a short list of keywords and notes here.

A shortcut to generating high-quality images is to reuse existing prompts. Head to the prompt collection, pick an image you like, and steal the prompt! The downside is that you may not understand why it generates high-quality images. Read the notes and change the prompt to see the effect.

Alternatively, use image collection sites like Playground AI. Pick an image you like and remix the prompt. But it could be like finding a needle in a haystack for a high-quality prompt.

Treat the prompt as a starting point. Modify to suit your needs.

Rules of thumb for building good prompts

Two rules: (1) Be detailed and specific, and (2) use powerful keywords.

Be detailed and specific

Although AI advances in leaps and bounds, Stable Diffusion still cannot read your mind. You need to describe your image in as much detail as possible.

Let’s say you want to generate a picture of a woman in a street scene. A simplistic prompt

a woman on street

gives you an image like this:

Well, you may not want the generate a grandma, but this technically matches your prompt. You cannot blame Stable Diffusion…

So instead, you should write more.

a young lady, brown eyes, highlights in hair, smile, wearing stylish business casual attire, sitting outside, quiet city street, rim lighting

See the drastic difference. So work on your prompt-building skills!

Use powerful keywords

Some keywords are more powerful than others. Examples are

  • Celebrity names (e.g. Emma Watson)
  • Artist names (e.g. van Gogh)
  • Art medium (e.g. illustration, painting, photograph)

Using them carefully can steer the image in the direction you want.

You can learn more about prompt building and example keywords in the basics of building prompts.

Want to cheat? Like doing homework, you can use ChatGPT to generate prompts!

What are those parameters, and should I change them?

Most online generators allow you to change a limited set of parameters. Below are some important ones:

  • Image size: The size of the output image. The standard size is 512×512 pixels. Changing it to portrait or landscape size can have a big impact on the image. For example, use portrait size to generate a full-body image.
  • Sampling steps: Use at least 20 steps. Increase if you see a blurry image.
  • CFG scale: Typical value is 7. Increase if you want the image to follow the prompt more.
  • Seed value: -1 generates a random image. Specify a value if you want the same image.

See recommendations for other settings.

How many images should I generate?

You should always generate multiple images when testing a prompt.

I generate 2-4 images at a time when making big changes to the prompt so that I can speed up the search. I would generate 4 at a time when making small changes to increase the chance of seeing something usable.

Some prompt only works half of the time or less. So don’t write off a prompt based on one image.

Common ways to fix defects in images

When you see stunning AI images shared on social media, there’s a good chance they have undergone a series of post-processing steps. We will go over some of them in this section.

Face Restoration

Left: Original images. Right: After face restoration.

It’s well-known in the AI artist community that Stable Diffusion is not good at generating faces. Very often, the faces generated have artifacts.

We often use image AI models that are trained for restoring faces, for example, CodeFormer, which AUTOMATIC1111 GUI has built-in support. See how to turn it on.

Do you know there’s an update to v1.4 and v1.5 models to fix eyes? Check out how to install a VAE.

Fixing small artifacts with inpainting

It is difficult to get the image you want on the first try. A better approach is to generate an image with good composition. Then repair the defects with inpainting.

Below is an example of an image before and after inpainting. Using the original prompt for inpainting works 90% of the time.

Left: Original image with defects. Right: The face and arm are fixed by inpainting.

There are other techniques to fix things. Read more about fixing common issues.

What are custom models?

The official models released by Stability AI and their partners are called base models. Some examples of base models are Stable Diffusion 1.4, 1.5, 2.0, and 2.1.

Custom models are trained from the base models. Currently, most of the models are trained from v1.4 or v1.5. They are trained with additional data for generating images of particular styles or objects.

Only the sky is the limit when it comes to custom models. It can be anime style, Disney style, or the style of another AI. You name it.

Below is a comparison of 5 different models.

Images generated by 5 different models.

It is also easy to merge two models to create a style in between.

Which model should I use?

Stick with the base models if you are starting out. There are pretty to learn and play with to keep you busy for months.

The three main groups of base models are v1, v2 and SDXL.

  • v1 models are 1.4 and 1.5.
  • v2 models are 2.0 and 2.1.
  • SDXL is still not officially released yet. It is now in SDXL 0.9 version for research use only.

You may think you should start with the newer v2 models. People are still trying to figure out how to use the v2 models. Images from v2 are not necessarily better than v1’s.

SDXL shows promise but it is still not for prime time yet.

I recommend using the v1.5 model if you are new to Stable Diffusion.

How to train a new model?

An advantage of using Stable Diffusion is that you have total control of the model. You can create your own model with a unique style if you want. Two main ways to train models: (1) Dreambooth and (2) embedding.

Dreambooth is considered more powerful because it fine-tunes the weight of the whole model. Embeddings leave the model untouched but find keywords to describe the new subject or style.

You can experiment with the Colab notebook in the dreambooth article.

Negative prompts

You put what you want to see in the prompt. You put what you don’t want to see in the negative prompt. Not all Stable Diffusion services support negative prompts. But it is valuable for v1 models and a must for v2 models. It doesn’t hurt for a beginner to use a universal negative prompt. Read more about negative prompts:

How to make large prints with Stable Diffusion?

Stable Diffusion’s native resolution is 512×512 pixels for v1 models. You should NOT generate images with width and height that deviates too much from 512 pixels. Use the following size settings to generate the initial image.

  • Landscape image: Set the height to 512 pixels. Set the width to higher, e.g. 768 pixels (2:3 aspect ratio)
  • Portrait image: Set the width to 512 pixels. Set the height to higher, e.g. 768 pixels (3:2 aspect ratio)

If you set the initial width and height too high, you will see duplicate subjects.

The next step is to upscale the image. The free AUTOMATIC1111 GUI comes with some popular AI upscalers.

How to control image composition?

Stable Diffusion technology is rapidly improving. There are a few ways.

Image-to-image

You can ask Stable Diffusion to roughly follow an input image when generating a new one. It’s called image-to-image. Below is an example of using an input image of an eagle to generate a dragon. The composition of the output image follows the input.

Input image
Output image

ControlNet

ControlNet similarly uses an input image to direct the output. But it can extract specific information, for example, human poses. Below is an example of using ControlNet to copy a human pose from the input image.

Input image
Output image

In addition to human poses, ControlNet can extract other information, such as outlines.

Regional prompting

You can specify prompts for certain parts of images using an extension called Regional Prompter. This technique is very helpful for drawing objects only in certain parts of the image.

Below is an example of placing a wolf at the bottom left corner and skulls at the bottom right corner.

Read the Regional Prompter tutorial to learn more to use it.

Depth-to-image

Depth-to-image is another way to control composition through an input image. It can detect the foreground and the background of the input image. The output image will follow the same foreground and background. Below is an example.

Input image
Output image

Generating specific subjects

Realistic people

You can use Stable Diffusion to generate photo-style realistic people. Let’s see some samples.

It comes down to using the right prompt and special model trained to produce photo-style realistic humans. Learn more in the tutorial for generating realistic people.

Animals

Animals are popular subjects among Stable Diffusion users.

Here are some samples.

Read the tutorial for generating animals to learn how to.

What is unstable diffusion?

Unstable Diffusion is a company that develops Stable Diffusion models for AI porn. They made headlines when their Kickstarter fundraising campaign got shut down. So far, they have not released any models publicly.

The company is not related to Stability AI, the company that released Stable Diffusion AI.

Next Step

So you have completed the first tutorial of the Beginner’s Guide! Check out the rest of them.

This is part 1 of the beginner’s guide series.
Read part 2: Prompt building.
Read part 3: Inpainting.
Read part 4: Models.

Leave a Reply

Your email address will not be published. Required fields are marked *