Stable Diffusion - Using Wildcards in Fooocus
HTML-код
- Опубликовано: 7 июн 2024
- Master Fooocus and Stable Diffusion for Creative Image Generation! 🎨 In this Stable Diffusion tutorial learn how to use and easy ways to create wildcards in Fooocus. Ideal for digital artists and creators looking to add some randomness to your image generations. Unlock your creativity with Fooocus and Stable Diffusion! #FooocusTutorial #StableDiffusion #DigitalArt #ImageManipulation #CreativeTools
This is part of my ongoing series of videos on installing and using Fooocus for Stable Diffusion. Fooocus is a great interface for using Stable Diffusion and is very easy to get started. It also will run on systems with only 4GB VRAM.
If you find the video useful and would like to tip, RUclips does have the Super Thanks option. Those are greatly appreciated so I can have my coffee!!
☕ BUY ME A COFFEE: ko-fi.com/kleebztech
Also I am a verified creator with Brave and do accept BAT for users of the Brave browser.
Download Fooocus: github.com/lllyasviel/Fooocus
How to install Fooocus video: • Stable Diffusion - Dow...
Fooocus Beginners Guide: • Super Easy Stable Diff...
Fooocus Upscales and Variations: • Stable Diffusion - Foo...
Checkpoints and LoRAs the Basics: • Stable Diffusion - Che...
Fooocus Poses and FaceSwap - Image Prompts: • Stable Diffusion - Pos...
Zoom with Outpainting: • Stable Diffusion - Zoo...
Inpainting with Fooocus - Don't Regenerate, Fix!: • Stable Diffusion - Inp...
Background Replacement with Fooocus: • Stable Diffusion - Bac...
00:00 Intro
00:25 How to Use Wildcards
01:55 Multiple Wildcards
03:12 Nested Wildcards
04:22 Wildcard File Location
06:30 Easy Ways to Create Wildcards - Наука
If you find the video useful and would like to tip you can buy me some electricity for all these image generations. It is greatly appreciated! ko-fi.com/kleebztech
I just discovered the wildcard function of fooocus thanks to this vid.
Some new updates to wildcards here. Nothing major but makes them a little more useful. ruclips.net/video/RX6kCZ-OPn4/видео.htmlsi=MV9v-53141TStMzz&t=145
@@KLEEBZTECH Thanks again
Nice, I'm really enjoying taking a break from comfy, and this has been a breath of fresh air without pulling my hair out or thinking too hard.
Yup. I went from Comfy to this for most of what I do.
I love randomness! It makes it always a surprise what will generate. Thanks for the wildcard guide. 👍
Surprise is the name of the game with AI. lol.
That's great! Thank you very much!
You are welcome!
LEgend thanks as always mate! :)
nice tutorial, keep the Fooocus content coming. btw, there's a way to add/edit the aspect ratios by editing the config file, but be sure to make a copy first!
Plan on it! Yes I actually plan on doing a video on editing the config file. But trying to compile all the edits that can be made before I do. I have personally only edited a few things myself. Probably one of the next couple videos.
learned my new thing for the day, thanks!
Excellent!
Thanks man
You're welcome!
Hello, I found it very helpful to watch the various videos on Fooocus AI. Do you have any ideas on how to solve the issues with artifacts on the hands? Is there anything that can be done through this interface and not others, like Automatic1111, which I find to be too confusing? Thank you, keep up the good work!
ah nice, didn't know Fooocus had wildcards. Too bad it doesn't autocomplete as you type them like it does in A1111 with the dynamic prompts and tagcomplete extensions
Cool, Can i suggest you find a solution to wrongly placed handles and knobs when you create an image of a kitchen.
I will experiment with that over the next few days and see if I can suggest any tips.
This information is very useful in certain kind of task. If I work with Fooocus using Colab, how can I save the text file in the wildcard folder?
Sadly I have very little experience with running in Colab. Maybe someone else might know. Also might want to check the discussion section on the Github page.
👍👍👍👍
By the way, if I choose two Styles, does Fooocus mix them or does it randomly select one of them again?
It mixes them. Which is why if you use photographic styles and try to do other images like comics you will have a difficult time.
@@KLEEBZTECH if different styles Fooocus do random choice too(. It's lottery!
i need prompt? bro
Hi, there’s no way o adding multiple word prompts with commas as wildcards like:
Red hair, long hair, curly hair
Pink hair, short hair, straight hair
So the red hair will always be long and curly and the pink hair always short and straight?
Yes you can even add whole prompts to each line. Whatever is on that line will be added.
Thanks for the video. Kleebz, did you try to make Fooocus create pictures from several wildcards consecutively?
I have Wildcard1: Item1 (new line) Item 2
Wildcard2: ItemA (new line) ItemB.
With the prompt: '[[Wildcard1], [Wildcard2]]' and enabled option 'Read wildcards in order', Fooocus jumps from 1 wildcard to another, and I get as a result:
Picture1: Item1
Picture2: ItemA
Picture3: Item2
Picture4: ItemB
How can I make Fooocus first go through the whole Wildcard1 and only then go to Wildcard2?
I didn't manage to find out
Using Wildcard3 with __Wildcard1__ (new line) __Wildcard2__ gave the same result
Well.... I am not sure I fully understand your comment but I will mention the wildcards have changed since this video by adding the reading them in order. But there are also arrays. Using the [[option1, option2]] is an array and would be red as option1 for the first image generated and option2 for the second one. You can put wildcards in an array but it would still do those in order. You want it to go through the whole wildcard1 first and then wildcard2 after? First question would be why not just combine the two wildcard lists into one? That is the easiest way off the top of my head but also not sure if that is exactly what you mean.
@@KLEEBZTECH Yes, you understood me right: whole wildcard1, then whole wildcard2.
The reason I'm looking for an optimization of this process is because I run Fooocus to generate hundreds of pictures at one set.
I have wildcard1, 2, 3, 4, 5... on different topics, and I don't want final images to be mixed in the output folder, 'cause it takes time to sort them. And I like to keep wildcards separate to keep files ordered on my drive
Right now my brain can't think of a way of doing it but if I do I will let you know.
After consulting with ChatGPT on the subject I don't think there is a way of doing what you are trying to do with separate wildcard files. Several things work against that but one is the fact that it would have no way of knowing when you get to the end of the first wildcard list to move on to the next even if there was a way to do the rest.
@@KLEEBZTECH thanks for taking the time trying to find the solution Kleebz.
I wouldn't think of using Chatgpt though, since I wouldn't know how to explain to Chatgpt how Fooocus works))
Thanks Kleebz! I appreciate your help =)
Thank you for providing us with a tutorial on the variety and possibilities of what we can achieve in Fooocus.
However, is there any practical use for the wildcard trick, rather than simply specifying the exact subject we want?
For me, it seems that it would only slow down the workflow when someone desires less randomness and more precision.
Depends on what your are doing. If someone wants to try out different variations of something this can really be useful. I often generate 30+ images of things and use them to try out different ideas. Then I can narrow down and be more specific.
@@KLEEBZTECH Thanks for the tutorial. Do you know if it is possible to substitute words one by one instead of randomly?
I think you will see support for arrays coming very soon which would be what you are looking for. github.com/lllyasviel/Fooocus/pull/1503
How do you tell it the color you want to use from the list file?
You don't. The point of a wild card is to randomly put one in there although now you can do them in order as well. If you want a specific color just put that in your prompt.
@@KLEEBZTECH yeah I played around and this works. __haircuts___, short (((Afro))) puts an Afro on subject.
But you don't need to put the wildcard in the prompt if you want an afro haircut. Just put that in the prompt. Otherwise you will get a random one and then you needed to add the extra weight to overcome the other one from the wildcard.
Can wildcards be added to styles?
I have not actually tried. I have a feeling they may not be able to be based upon how Focus does things but I'm completely going off of memory. But definitely worth trying.
@@KLEEBZTECH They definitely don't work the normal way of just typing the name with 4 underscores. It seems to either make something completely random or make something based on the wildcard name.
I'm guessing it's either not implemented yet or you need to add more than just the underscores. It would be super cool if it worked since I like making wildcards with full prompts and then nesting them multiple times for complete randomness lol.
I think the wildcards are handled before the styles are added which would mean it won't work. @@alt666
My dream is one day we could have apparelswap to swap the clothes
I think that sort of thing will be possible soon.
how i can have always the same face?
Faceswap is one way to help with that. ruclips.net/video/2hW0M-Y1R6k/видео.htmlsi=qLvoZnZ5bM1Ro1u8
The choice is absolutely random?
It is random. But I have seen that there may be support for arrays soon which would allow you to have it go through a list in order.
@@KLEEBZTECH If you specify the same token _x_ twice in one prompt, will it drop twice by chance?
If you are referring to using the same wildcard twice, it will pick a different one for each. Of course you could still get the same one. For example in the video I used __color__ twice in a prompt and it picked the same color for both.
when i select log, i get: {"detail":"File not allowed: ...
Not sure with the limited info but have you changed where the output folder is? What OS?
o sorry. windows11. i've placed a custom output location in E: (same volume) in config file @@KLEEBZTECH
And where do you see that error? Is it in the browser window? Did you have any issues if the output folder was in the original location?
yea in the browser window. I hadnt tried it before so dont know. Gonna post it in github thanks!@@KLEEBZTECH
Is it possible to share the actual folder path you have for the output location? I have a theory.