Someone finally made a really, really easy way to use Stable Diffusion. It's even easier than MidJourney now. No Discord is necessary and of course no paid subscription. It's just a text box, a generate button, and you get incredible images made on your own gaming PC. And the best thing, it's built on the brand new Stable Diffusion XL. So the resolution and quality are unmatched. Hey, I'm Albert and today I'll introduce you to Fooocus. Not sure how to pronounce that. The new UI for Stable Diffusion just released by Ilyasviel. If you have a bit of knowledge about Stable Diffusion, that name might already be very dear to you, because they contributed significantly to ControlNet. Let's get right into it. Installation is super simple. Visit this page on GitHub. Link is in the video description. Make sure your PC meets these requirements and has around 15 gigs free. Click here to download a compressed file. Wait for it to download. and extract it to a folder of your choice. Next, double-click this run.bat file. Now it will automatically download the two checkpoints of SDXL, which are the heart of this. They're pretty big, so it might take a while. Once that's done, it will launch the UI in your web browser. If it doesn't, visit this URL. As you can see, the user interface of Fooocus is super simple. You can enter your prompt here, and press Generate. Let's do so. How about "portrait of a victorian-era dragon wearing a suit?" Don't write more than that, because that's how simple a prompt can be in Fooocus. Great, you see it works really well. But now, let's check out these advanced settings. Click here. Up here is Performance. You have two choices, Speed and Quality. Let's switch to Quality. Next up are aspect ratios, so the resolution of width and height. Let's do a vertical image, like 832 by 1216. Image number is self-explanatory. How many pics you want? The higher, the longer it will take, of course. Seed you can leave alone, unless you want to create the same or a similar image again, in which case you enter that image's seed here. Negative prompt is where you enter what you don't want to see in your pic. For example, if something keeps showing up that you want to get rid of in your results, write it in here. Style is my favorite tab in Fooocus. It kind of reveals how this UI works in the background. The prompt we entered is only part of what is passed on to the model. The style you select here is added and processed with your prompt, so we don't have to worry about prompting a ton of text each time we want to generate. People assume that this is what MidJourney does to prompts that make the results there so great, and this is the first time that that sort of automation is available to us in SD. You can see the default is cinematic, which is why our first result looked the way it did. So let's now pick the retro game style and generate again. Cool. And now finally, let's check out the Advanced tab in the Advanced menu. Here you can pick which model to use, in case you've downloaded a different one, as well as various LORAs. You can get those on sites like civitai.com, which I explain in more detail in my other videos. Those are all the features of Fooocus, which is what makes it such a great introduction to Stable Diffusion, but it can also save you a ton of time even if you're a more experienced user. I'm sure I'll be using it often to demonstrate SD or just quickly generate some concept images without thinking too much about prompting. If this was a little too simple for you, check out my recent tutorial on how to install the Auto1111 UI, and be sure to leave a like, subscribe, and keep an eye out for future videos. In addition to tutorials, I'll be keeping you up to date on the AI film production company I'm building, Storybook Studios. Thanks for watching, I'm Albert Bozesan, and I hope you have fun with Stable Diffusion.
