Fooocus - The Fast And Easy Ui For Stable Diffusion - Sdxl Ready! Only 6gb Vram

get better pictures faster hello my friends how are you doing today I'm gonna show you a UI that is quick it is simplified it is concentrated just on having fun with AI and on top of that it uses much less resources let's get started so here's a quick look on how this UI is looking and as you can see it is super simple you have the preview of what you're rendering down here is the prompt you're rendering and then you have a generate button so you are focused on your art of course there is more settings you have here input image you can do in painting you can do out painting you also have advanced settings but all of that is hidden so that you can stay focused on playing with amazing prompts now here's how to install that and that is also super simple we are on the GitHub page and when you scroll down a little bit you come here to the windows install section there's a link here for a seven set file and this can be unpacked for example with the renderar but also with other tools inside of that you're gonna find these two folders and the file as you can see from one of the folders this has an embedded python this will run right from the folder it's portable also the Run bet file here is starting Python and then running the UI there is nothing to install however as you can see down here this is going to download the sdxl base 1.009 model and also the refiner model for that and you need that because opening this UI is always checking if these two files are there if you already have them you can link to them you don't need to download them again another thing that is important here to know is when you use the in painting for the first time this is also going to download its own in paint Focus whose model you can of course also use your own in paint models if you want to here is a little advice for you if you download the folders and you already have automatic 1111 or any kind of other UI where you have already downloaded a lot of models you want to go here into the focus folder in the unpacked folder that you've just downloaded and in there you're going to find the user path config.txt file now in that file you're going to find a list here the first one says model file path and you simply copy the location of where your models are now there's one thing that's a little bit strange and specific about this you can see that here you have these double lines and you need them in your direction too otherwise it's not going to work you can see here that I'm using the location from automatic 1111 model stable diffusion so that I have all of my models to choose from inside of focus after that you're simply going to save the text file and then you want to double click on run bad so that this is starting the UI once this is done you're ending up in the UI I have already showed you and here is the thing this is super simple as you can see and most of the time the idea is that you are only writing the simple prompt down here and click on generate because focus is doing a lot of magic in the background for you now the magic can be found by clicking down here on Advanced this is going to open up a new section here on the right side where you can choose between the speed of the rendering and the quality of the rendering no specific settings this will do the job for you below that you have different resolutions of ratios no sliders just the resolutions you can select how many images you want to render at a time and down here if you would like to you can also also write a negative prompt but focus is going to create its own negative prompts for you so this is not really a necessity now here's the next thing if you go to the style tab you're going to find a long list pre-created for you the styling words for the positive and negative prompt to give you amazing results in that specific style if you don't use that this will automatically have the focus version 2 style and the default slightly cinematic style selected for you which means that often the image are going to come out as photorealistic cinematic images but of course you can also create other styles so here for example I have created an old Chinese man in a cinematic photographic style really beautiful then I selected a different style and as you can see this is more of a digital art style looks very beautiful here we have more of a pop r art style also very very nice in the next image we have our Chinese old guy in an art new of style and I find that also very beautiful next we have a very cute rendition as a low poly version of the old bearded Chinese guy very very cute and last but not least we haven't as a legendary neon Punk Samurai this one I really love the most now just to remind you all of these images are using the simple prompt old man sad long beard and then in round brackets Samurai clothing and the only thing I do to change these Styles is to select the styles from over here on the right side where there is tons and tons of styles to choose from on top of that you also have the advanced tab in here you can select the model you want to use from your model list in this case I'm selecting chargernaut XL you can also select an sdxl refiner in this case I have none because this model doesn't need a refiner and down here you can also select loras now in that case the UI itself comes with the offset example Laura for stxl and this is also pre-selected for you and with the slider over here you can also of course select the weight of that specific Lora down here it says advanced again you can click on that and there is a little extra section down here for sampling sharpness so you can have this slider adjusted to your choice experiment with that and then also down here you have document and when you click on that link this will bring you to a documentation of the UI so this is as simple as it can be when we close the advanced section you can also see we have an input image section now this has multiple functions as you can see on the right side and again this is doing most of the work and the settings for you you have different choices you can select an image here as an input and then create variations of that image in a subtle way or in a strong way as variations here I have an input image of a cowboy and then next you can see a subtle variation which is pretty close to the original and here we have a strong variation of that image and then you can do an upscale by 1.52 and then also a fast version of two you don't have to do any kind of settings for that and if you have rendered an image up here like we have with our neon Punk Samurai you can simply grab that image and drag it down here to use it as an input for either the upscaling or for creating variations with that image and as you can also see we have here a second part for in painting and the out painting beta when you have an image already loaded in the upscale or variation Tab and you click on in painting this should be automatically loaded into the in painting input if it is not loaded you can again either drag it down here or load it from a folder now for rendering the in painting you can select on the right side the size of your brush and then you simply Mark what you want to in paint write the prompt up here and click on generate this will generate the in painting for you and then also below that you have here several selections for out painting So on the left the right top or bottom if you want to extend the images to the side now I have to say this is in beta and the results are not great as you can see here we have here original image with the cowboy and here we have an extended image it's kind of good but not really so this still needs some work to get you better results I have to say I really enjoy this Sim Amplified UI because it makes you much more focused on creating your art rather than playing around with a million different settings and extensions now one thing I'm kind of missing here is the ability to nail down the seat if you want to play around with different prompt variations and I invite you to try it out because it is super simple to use it is portable and you will have a lot of fun with that oh you're still here so uh This is the End screen there's other stuff you can watch like this or that's really cool and yeah I hope I see you soon leave a like if you haven't yet and well um yeah I wish you good weekend

Channel: Olivio Sarikas

Views: 60,209

Keywords: oliviosarikas, olivio sarikas, olivio tutorials, Fooocus, Fooocus ui, Fooocus AI, stable diffusion, stable diffusion tutorial, stable diffusion install, stable diffusion img2img, A1111 alternative, Midjourney alternative, Midjourney local, run local AI, ai art, sdxl 1.0, ai art lookbook, ai art tutorial, generative ai tutorial, ai art video, generative ai explained, stable diffusion video, ai artist

Id: zIhODzEVZqg

Length: 9min 40sec (580 seconds)

Published: Sat Sep 30 2023