A Brief Walkthrough of FooocusUI: Tips and Tricks to Enhance Your Experience

Introduction

Fooocus is super easy to use! In the default mode, just type your prompt and hit Generate—it's that simple! (Press Ctrl+Enter on Windows or Cmd+Enter on Mac.) Get ready to watch your ideas come to life in seconds! Whether you're creating art, marketing materials, or just having fun, Fooocus makes it quick and enjoyable. Let your creativity flow and see what magic you can create!

Prompt:

monkey king, snow, moon, mountain range

• Prompt Engineering
• Advanced Settings
  • Preset
  • Performance
  • Aspect Ratio
  • Add Custom Aspect Ratio
  • Image Number
  • Output Format
  • Negative Prompt
  • Seed (aka Random)
• Consistent Results
  • Steps to Create an Image Prompt
  • Let's Create Results with Same Looks

Prompt Engineering

In above example, the images are clearly different from one another, and that's largely due to the lack of specific instructions in the prompts. A more detailed prompt can guide the model to produce results that align closely with your vision.

Let's try to expand our prompt:

Create a scene with the Monkey King standing on a snow-covered mountain range under a full moon. The moonlight illuminates the snow, highlighting the Monkey King’s golden staff. Capture his playful yet powerful presence against the dramatic peaks and soft clouds in the night sky

Result:

Now do you feel the difference? The lack of clear instructions led to unexpected results. However, when we use a bit more detail in our prompts, the images start to look more similar and aligned with our vision.

This little experiment shows how important it is to be specific—detailed prompts can guide Fooocus to produce the cohesive results you’re aiming for. So, don’t hesitate to add those extra details and watch your ideas come to life!

Still confused? Check out this guide (toddler level) 👉 Importance of Prompts in Stable Diffusion?

Advanced Settings

Selecting the Advanced checkbox at the bottom of the UI opens up the advanced settings menu.

Preset

Fooocus offers built-in presets, and for now, let's stick with the default option to keep things simple and effective.

anime	realistic	sai

Performance

Choose from a range of performance settings to match your needs: Quality, Speed, Extreme Speed, Lightning, and Hyper-SD. Select the one that fits your project best!

Hyper-SD	Extreme Speed	Quality

Aspect Ratio

You have the flexibility to use any resolution you like. Adjust the aspect ratio to create images that perfectly fit your vision.

Add Custom Aspect Ratio

Want even more control? You can add custom resolutions in the modules/flags.py file. Simply go to the line containing the variable sdxl_aspect_ratios, and below it, add a new line like this:

sdxl_aspect_ratios.append("width*height")

Example: if you wanna add 1440p resolution then

sdxl_aspect_ratios.append("2560*1440")`

This way, you can tailor the aspect ratios to suit your specific needs!

Image Number

Specify the number of images you want to generate for each prompt. More images mean more creative options to choose from!

Output Format

Select your preferred output format for the generated images. Whether you need JPEG, PNG, or something else, Fooocus has you covered.

Negative Prompt

A negative prompt is like giving instructions on what to avoid. It helps guide the model by specifying elements you don’t want. For example, if you want to steer clear of anything inappropriate, you can add terms like "nudity, nsfw, sexual themes."

Seed (aka Random)

By default, Fooocus shows this option as Random. This means a new seed is generated each time, creating unique results. A "seed" is a numerical value that initializes the random number generator for image creation. If you want consistent and reproducible results, you can set a specific seed number for your runs. This way, you can recreate that perfect image whenever you need it!

Consistent Results

Fooocus also offers a fantastic array of options to help you achieve consistent results. One exciting feature is the ability to mix text and image prompts for truly incredible outcomes!

Steps to Create an Image Prompt

1. Click on Input Image: Start by selecting the image you want to use.
2. Go to Image Prompt: Navigate to the image prompt settings.
3. Turn on Advanced: Enable the Advanced option at the bottom of the dialog.
4. Choose Your Image: Select the image you want to base your results on.
5. Select Faceswap: This will allow you to blend features effectively.
6. Use the Default Config: The default settings work great for most cases!

Now, when you type your text prompt and hit generate, the resulting face will resemble the face from your image prompt, infused with details drawn from your text. It’s a powerful way to bring your ideas to life with striking detail!

Let's Create Results with Same Looks

Text Prompt 1:

samurai, cherry blossom

Text Prompt 2:

drinking black coffee, cozy environment, black formal dress

Text Prompt 3:

indian saree, fireworks in sky, night sky

Base image	Text Prompt 1	Text Prompt 2	Text Prompt 3

I adjusted the StopAt value to 0.95 and set the weight to 0.9 because my base image had green eyes, which can be tricky for Stable Diffusion to replicate. After making these changes, I noticed a significant improvement—Stable Diffusion started capturing the eye features much more accurately!

These tweaks really helped enhance the results, allowing for finer details and a closer match to the original image. It’s amazing how small adjustments can lead to such a big difference in the final output!