Fooocus : How does it compare to other Stable Diffusion WebUI

Learning how to use stable diffusion can be a game-changer for beginners looking to create stunning AI-generated images. Unfortunately, Stable Diffusion often gets mistaken as difficult, technical, and less user-friendly with its Stable Diffusion WebUI. With this kind of reputation, new users often just opt for another AI generator like Midjourney or Dall-E where users can just focus on the prompt wording. Fortunately, the Stable Diffusion Community is quick to respond to this problem. Fooocus, which starts to gain popularity last month, is efficient, much simpler, and easier to use, even easier than Midjourney and without the need to use it in Discord. The open-source version also does not need heavy hardware requirements at 4GB Nvidia GPU memory (4GB VRAM) and 8GB system memory (8GB RAM). Fooocus can be a good alternative for beginners to try Stable Diffusion if you are not ready for heavy customization like Automatic1111 & ComfyUI.

Let’s explore how Fooocus compares to Automatic1111.

Fooocus Interface

As you can see, Fooocus has a much simpler interface than A1111. Fooocus remained focused (no pun intended) on generating images, while A1111 brings the flexibility to add as many extensions to create fully customizable images, animations, and Stable video diffusion.

Customisation

Fooocus offers plenty of customization options, mainly: Faceswap, Inpainting, and Outpainting.

Fooocus Prompt Engineering

Fooocus developers use an offline GPT-2 based prompt processing engine and lots of sampling improvements. This way, image generation will always produce gorgeous results, regardless of whether your prompt is as brief as “pink horse” or as long as 1,000 words.

Fooocus also has a tool called Image Prompt, which allows users to upload an image and have it generate another image comparable to the given image. Image Prompt might be quite useful if you find tough poses or want to include a specific color scheme in the image production process.

Image Input:

Image Output:

Aspect Ratio

While Automatic1111 mostly use 512×512, Fooocus have extensive list of image size for users to choose from. However, the default setting for Aspect Ratio is 1024×1024.

Fooocus Model & Style

Unlike Automatic1111 & ComfyUI where users can download and install various models easily, Fooocus currently only support SDXL models. However, DiffusionHub users can enjoy downloading models using model downloader at our Automatic1111 feature and then change instance to Fooocus and select the desired model.

Fooocus also makes it easy for Stable Diffusion newbie to generate images. You don’t need to engineer a prompt to achieve certain styles like anime, cinematic, realistic, you name it. Simply just click the Style Tab and choose which style you want to get.

Describe

The Describe function in Input Image guesses a prompt of an image. It is similar to the Interrogate CLIP button in Automatic1111.

Conclusion

Fooocus was originally designed to reinvent the Midjourney experience. Often the case with Midjourney is the lack of User Interface as a result of the use of Discord, a paid service that is subject to censorship.. However, Midjourney does generate good image with simple prompt.

With many adjustments in Fooocus happening over the last few months, it is reasonable to conclude that Fooocus has caught up to Midjourney. This alternative Stable Diffusion WebUI is easy to use, fast, and reliable with the ability to generate high-quality images. Comparing Fooocus to Automatic1111 is a little one-sided because Automatic1111 supports a wide range of extensions for ultimate customization. DiffusionHub now supports Fooocus for easy generation, Let us know what you think!