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
This def will help me with my content creation moving forward. Wish I had found your channel much earlier. Naturally subbed and once I do get some traction on my stuff (if any) I pledge to share % of it as an appreciation cuz you deserve more than just a cup of coffee. Kudos.
Another fantastic tutorial! Ive got to tell ya, it seems like, whatever project I'm working on at any given moment (that requires a specific image from fooocus) your tutorials always seem to pop up at the perfect time. I'm currently working on a video clip with 2 characters talking, and this tutorial is Exactly what I needed..Today! How do I make a donation?
Excellent! I am glad I didn't wait a couple more days to do it then. lol. You can donate at ko-fi.com/kleebztech or TH-cam does offer the Super Thanks option. Whichever way is easier for you. Very much appreciated since the ad revenue on TH-cam is not that great for this stuff.
Great information in your vids as usual. Request: How about a video on how to make 16-bit character sprite sheets for video gaming? It often takes a couple similar sprites to create the illusion of walking, attacking, etc. Seems like an area that AI would be strong in.
I have tried similar stuff in the past with little luck besides the occasional lucky roll. But it is something I will think about since I have figured out a bunch of stuff since I have tried.
Hello, first I want to thank you for the excellent content available on your channel. I'm a graphic artist and in my opinion, Fooocus is one of the best AI tools to date. Do you know if there is a forecast for a new Fooocus update? I hope the app can improve towards effective character consistency. Making a single piece of art with the help of AI is fine... the problem is reproducing the same color pattern, scenery, characters and accessories... I've already tested several techniques: LoRA, text prompt, image prompt... ( these same tools combined) however, creating consistent sequences is unproductive, uncertain and frustrating.
Wait so inpainting on fooocus could be affected by prya and image prompt image if you set it up beforehand? Alright, very enlightening thank you thumbs up
why do you remove the background if you just need a layer with only the text? just take the image, add a layer, position the text, remove the image and voilà
One question: how do you arrange your Input Image windows like that? I only have a single one and a bunch of tabs... I've poked around but can't find anything to rescale them. Any suggestions?
@@KLEEBZTECH My panels/windows are shown under each other, rather than in a 2x2 grid (sorry I didn't make that clear at first). And yeah, regular Fooocus.
So you have all 4 but they are in a single column on top of each other? I would assume it has to do with display resolution and possibly the size of text/page zoom. If I use the browser zoom (CTRL-/+) I get them to form into a single column.
@@KLEEBZTECH thanks. My screen is set at 3840x2160 so it's probably something to do with that. It's not a big deal, I was just more intrigued as to why your display looked so much easier to work with...
Quick q: When I do this, all it does is remove whatever I have masked and fill in the spot with the background. Not sure why. But the background removed looks awesome. Any thought?
I just tested with Colab and had no issues with styles. Do you mean that you don't notice the styles in the results or it won't actually let you choose them? Do need to remember that it is usually best to disable other styles when using some since they can fight each other.
I use the same one that is linked on the Github page. So you are saying if you disable all other styles and choose something obvious like "SAI Craft Clay" in styles and just have a prompt of "a dog" it gives you something like a photo and not something made of clay? github.com/lllyasviel/Fooocus?tab=readme-ov-file#colab
Excellent tips. Midjourney are working on 3d they just hired a load of experts. Im gathering all these other companies will follow suit. Imagine being able to pan and zoom around this image. I was wondering if i was going to create an army of insects with a human look. Lotr meets bugs life. Wot u think best model to download. Any suggestions greatly appreciated :)
As for which model to use... I guess that would partly depend on the style you are looking for. Animated or more realistic. I honestly find JuggernautXL to be a good model to start with and try for most things. Anthropomorphic LoRAs or models I have not really seen much of but many checkpoints can do them.
@@KLEEBZTECH ah thanks. Sum great videos uploaded. Do appreciate the comments. I was looking for noble medieval warriors with maybe human ish bodies and insect head and limbs?. I'm getting some ok results in MJ but gonna have to use your method to get them fighting and posing. Thanks once again ;).
Yeah anthropomorphic will probably be the big term to use. Also of course hybrid as well. I have done a decent amount of that sort of thing myself. Getting them consistent might be a challenge.
thanks，i have a Clothing mannequins,white fake head and hand foot,how to replace them to real head hand foot.keep the same pose.inpaint not work well,can we get pose editor? sorry my bad english.
@@KLEEBZTECH I have some dummy model props wearing clothes, some models have heads, and some don't. How can I use a pose editor tool similar to A11111 to position bones. I am trying to use paint to apply to the face, hands, and feet of the model's props. I need to create many images to have a usable one. 我有一些穿着衣服的假人模特道具，有的道具模特有头，有的没有，怎么用类似A11111里的pose editor 工具来摆骨骼。 我尝试着用inpaint来涂抹模特道具的脸，手，脚，要出好多图才能有一张能用的。
I cannot wait to get hands on it. This is exactly what I need so much. I have even more ideas to add on this tut. There is a branch of Fooocus there called Deefooocus, which is claimed to be faster and can do magic on clothes change with ease.
I have tried Defooocus but not sure it is any faster. As for the clothing I assume you are taking about the ability to create masks automatically. I use mashb1t's 1-Up Edition usually which also includes that but also has some things that Defooocus does not.
I have used with several checkpoint models without issues. could you describe the type of results (or non results) you are getting? Maybe we can figure out the issue. I will do a little experimenting once I have a better understanding of what you are getting. @@fishpickles1377
So after some more testing with previously generated images it does appear that it can be very hit or miss. I think it has to do with things not easily matching up with the whole scene and the AI matching things with the inpainting.
Are things out of alignment? If so then the settings may not be correct. First the reference images should be the same aspect ratio and when inpainting you do need to change the inpaint respective field.
The process for making comics needs to be alot more simplified and made to be All-in-one like KITA IA . You still have to go here ;there ; and every where just to make one image . To make a 20 comic book this way would take you forever
People can use whatever tools they prefer. I am just trying to show how to mix multiple characters with AI and using Fooocus. But many techniques can be used elsewhere.
@@KLEEBZTECH These are good techniques and thank you. But it still takes too long. To create one image doesn't matter what tools you use. Today. They need to be improved.
Since Fooocus is just seems like a beginners level for Automatic1111 or Forge, where would we find the same "inpaint" and "prompt" extensions in Automatic1111? Help the noobs that have outgrown Fooocus but are still learning Automatic1111.
I do not agree that it is for just beginners. They do make it easier for beginners but there is a ton it can do once you learn how to use it. I don't use A1111/Forge much these days so really can't say much about the inpainting since Fooocus has the best inpainting. I can do most of what I need in Fooocus so rarely find the need to use others. I hated spending forever tweaking things to get images.
@@KLEEBZTECH I feel your pain. But I was going nuts trying to master four different GUI's at once. I settled on Automatic1111 as the most advanced short of ComfeUI.
I use Fooocus and ComfyUI only for the most part these days. Can't stand A1111 interface. Just a cluttered mess and for most of what I do it is not needed. Comfy I am okay with since the node system I am familiar with from other software. But will admit I use that rarely as well since I can get most things from Fooocus.
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
Thanks!
Thank you very much as well! These make my day.
You're welcome. Keep up the great work. These tutorials are helping me a lot, as Fooocus can be So annoying/unpredictable most of the time
AI unpredictable? lol. You can say that again.
Thanks! Super helpful to learn new Fooocus techniques.
Very welcome! And thank you very much for the donation!
Danke!
And thank you as well!
Thanks
Welcome. And thank you very very much for the donation!
Great ! You are a Fooocus master ! Thanks !
This def will help me with my content creation moving forward. Wish I had found your channel much earlier. Naturally subbed and once I do get some traction on my stuff (if any) I pledge to share % of it as an appreciation cuz you deserve more than just a cup of coffee. Kudos.
Thanks! Good luck with what you are working on.
Sweet tutorial ❤ love it
Thank you! Cheers!
Great video. Excellent explanation!! Thanks
You're welcome!
Great tutorial and workflow. Thanks!
You're very welcome!
Many thanks, very interesting
As always, super straight and useful🤸 💥
As always - Great!
Thanks again!
Excellent video, I needed it, thank you God bless you in his glory
Glad it was helpful!
VERY helpful video!
Thank you VERY much!
Another fantastic tutorial!
Ive got to tell ya, it seems like, whatever project I'm working on at any given moment (that requires a specific image from fooocus) your tutorials always seem to pop up at the perfect time.
I'm currently working on a video clip with 2 characters talking, and this tutorial is Exactly what I needed..Today!
How do I make a donation?
Excellent! I am glad I didn't wait a couple more days to do it then. lol. You can donate at ko-fi.com/kleebztech or TH-cam does offer the Super Thanks option. Whichever way is easier for you. Very much appreciated since the ad revenue on TH-cam is not that great for this stuff.
Any demo for your clip result? would love to see it :)
Great job 😊
👌👌👌
Thanks once again
You're welcome 😊
best Tutorials ever !
Great information in your vids as usual.
Request: How about a video on how to make 16-bit character sprite sheets for video gaming? It often takes a couple similar sprites to create the illusion of walking, attacking, etc. Seems like an area that AI would be strong in.
I have tried similar stuff in the past with little luck besides the occasional lucky roll. But it is something I will think about since I have figured out a bunch of stuff since I have tried.
@@KLEEBZTECH Cheers. You're the man
Thanks a lot. But why remove background??
Just for fun 😉
Hello, first I want to thank you for the excellent content available on your channel. I'm a graphic artist and in my opinion, Fooocus is one of the best AI tools to date. Do you know if there is a forecast for a new Fooocus update? I hope the app can improve towards effective character consistency. Making a single piece of art with the help of AI is fine... the problem is reproducing the same color pattern, scenery, characters and accessories... I've already tested several techniques: LoRA, text prompt, image prompt... ( these same tools combined) however, creating consistent sequences is unproductive, uncertain and frustrating.
awesome, thx
You bet!
Wait so inpainting on fooocus could be affected by prya and image prompt image if you set it up beforehand? Alright, very enlightening thank you thumbs up
Yup. Also how I do text effects. I have a couple videos on that.
why do you remove the background if you just need a layer with only the text? just take the image, add a layer, position the text, remove the image and voilà
That will work.
One question: how do you arrange your Input Image windows like that? I only have a single one and a bunch of tabs... I've poked around but can't find anything to rescale them. Any suggestions?
If you go to the Image Prompt tab in the Input Image section you don't have a grid of 4? Are you using the regular Fooocus or a fork?
@@KLEEBZTECH My panels/windows are shown under each other, rather than in a 2x2 grid (sorry I didn't make that clear at first). And yeah, regular Fooocus.
So you have all 4 but they are in a single column on top of each other? I would assume it has to do with display resolution and possibly the size of text/page zoom. If I use the browser zoom (CTRL-/+) I get them to form into a single column.
@@KLEEBZTECH thanks. My screen is set at 3840x2160 so it's probably something to do with that. It's not a big deal, I was just more intrigued as to why your display looked so much easier to work with...
@@KLEEBZTECH 90% zoom and I get the 2x2 grid. Thanks for clearing that up!
Wow !!
Quick q: When I do this, all it does is remove whatever I have masked and fill in the spot with the background. Not sure why. But the background removed looks awesome. Any thought?
Have you enabled image prompt and inpainting in the debug menu?
I test on my google and its not makeing the styles, looks like not working on google colab anymore
I just tested with Colab and had no issues with styles. Do you mean that you don't notice the styles in the results or it won't actually let you choose them? Do need to remember that it is usually best to disable other styles when using some since they can fight each other.
@@KLEEBZTECH no, its alway render as defult, can you send me the link for the fooocuseon collab?
I use the same one that is linked on the Github page. So you are saying if you disable all other styles and choose something obvious like "SAI Craft Clay" in styles and just have a prompt of "a dog" it gives you something like a photo and not something made of clay? github.com/lllyasviel/Fooocus?tab=readme-ov-file#colab
Excellent tips.
Midjourney are working on 3d they just hired a load of experts. Im gathering all these other companies will follow suit. Imagine being able to pan and zoom around this image.
I was wondering if i was going to create an army of insects with a human look. Lotr meets bugs life. Wot u think best model to download. Any suggestions greatly appreciated :)
As for which model to use... I guess that would partly depend on the style you are looking for. Animated or more realistic. I honestly find JuggernautXL to be a good model to start with and try for most things. Anthropomorphic LoRAs or models I have not really seen much of but many checkpoints can do them.
@@KLEEBZTECH ah thanks. Sum great videos uploaded. Do appreciate the comments.
I was looking for noble medieval warriors with maybe human ish bodies and insect head and limbs?.
I'm getting some ok results in MJ but gonna have to use your method to get them fighting and posing. Thanks once again ;).
Yeah anthropomorphic will probably be the big term to use. Also of course hybrid as well. I have done a decent amount of that sort of thing myself. Getting them consistent might be a challenge.
thanks，i have a Clothing mannequins,white fake head and hand foot,how to replace them to real head hand foot.keep the same pose.inpaint not work well,can we get pose editor? sorry my bad english.
I will admit I am not sure of your question. If you want to ask in another language and then I will use Google translate that might work better.
@@KLEEBZTECH I have some dummy model props wearing clothes, some models have heads, and some don't. How can I use a pose editor tool similar to A11111 to position bones.
I am trying to use paint to apply to the face, hands, and feet of the model's props. I need to create many images to have a usable one.
我有一些穿着衣服的假人模特道具，有的道具模特有头，有的没有，怎么用类似A11111里的pose editor 工具来摆骨骼。
我尝试着用inpaint来涂抹模特道具的脸，手，脚，要出好多图才能有一张能用的。
Fooocus doesn't have a pose editor which is why I use posemyart and how I show in the video. th-cam.com/video/8USLRg4ZCgE/w-d-xo.html
I cannot wait to get hands on it. This is exactly what I need so much. I have even more ideas to add on this tut. There is a branch of Fooocus there called Deefooocus, which is claimed to be faster and can do magic on clothes change with ease.
I have tried Defooocus but not sure it is any faster. As for the clothing I assume you are taking about the ability to create masks automatically. I use mashb1t's 1-Up Edition usually which also includes that but also has some things that Defooocus does not.
😍
More controlnet more fun, Fooooous need ip adapter...
Have you had any luck inpainting these poses into existing images? I've been trying, but it doesn't seem to want to work for me.
Are we talking about previously generated images or real photos?
@@KLEEBZTECH Previously generated. I'm wondering if it's because i'm using juggernaut.
I have used with several checkpoint models without issues. could you describe the type of results (or non results) you are getting? Maybe we can figure out the issue. I will do a little experimenting once I have a better understanding of what you are getting. @@fishpickles1377
So after some more testing with previously generated images it does appear that it can be very hit or miss. I think it has to do with things not easily matching up with the whole scene and the AI matching things with the inpainting.
It does it but sometimes does not look the greatest.
I really got weird results with this. Can anyone help me? When inserting the image as a reference, it gets the character all messed up
Are things out of alignment? If so then the settings may not be correct. First the reference images should be the same aspect ratio and when inpainting you do need to change the inpaint respective field.
@@KLEEBZTECH I did all of that, but the character I get is all messed up. Please help :(
The process for making comics needs to be alot more simplified and made to be All-in-one like KITA IA . You still have to go here ;there ; and every where just to make one image . To make a 20 comic book this way would take you forever
People can use whatever tools they prefer. I am just trying to show how to mix multiple characters with AI and using Fooocus. But many techniques can be used elsewhere.
@@KLEEBZTECH These are good techniques and thank you. But it still takes too long. To create one image doesn't matter what tools you use. Today. They need to be improved.
In the 'old days' it took even longer because they had to use tools like pencils and brushes :)
@Art0691p That's right.That's why the technology needs to keep Evolving, because time is money.
@@dwainmorris7854 How dispiriting. Time spent creating is enjoyment for me. Not thinking about the money. That came later.
Since Fooocus is just seems like a beginners level for Automatic1111 or Forge, where would we find the same "inpaint" and "prompt" extensions in Automatic1111? Help the noobs that have outgrown Fooocus but are still learning Automatic1111.
I do not agree that it is for just beginners. They do make it easier for beginners but there is a ton it can do once you learn how to use it. I don't use A1111/Forge much these days so really can't say much about the inpainting since Fooocus has the best inpainting. I can do most of what I need in Fooocus so rarely find the need to use others. I hated spending forever tweaking things to get images.
@@KLEEBZTECH I feel your pain. But I was going nuts trying to master four different GUI's at once. I settled on Automatic1111 as the most advanced short of ComfeUI.
I use Fooocus and ComfyUI only for the most part these days. Can't stand A1111 interface. Just a cluttered mess and for most of what I do it is not needed. Comfy I am okay with since the node system I am familiar with from other software. But will admit I use that rarely as well since I can get most things from Fooocus.