If you find the video useful and would like to tip you can buy a cup of electricity for all these image generations. It is greatly appreciated! ko-fi.com/kleebztech
Yes bro! And please can you make a video on how keep clothes same when combining two pictures in image prompts for a different scene but wanna copy clothes of one character in a complex scene? I appreciate you🙏
Hi, it would be interesting if you made a guide like the one you did to generate the face grids of some previous tutorial but to create more complex scenes with more people, for example "a girl and a boy by the river", "a boy and a girl fighting each other in a castle" and then from that generated grid then generate the actual image, unfortunately there are no guides on this, perhaps not necessarily using a grid but also generating scenes from scratch in depth map
So I've got a character made in SD, that I love. But can't get it to do t-poses so I can try to start making a character mesh to import into UE5. Can I use the Focus tool to keep the same character. But get the poses I want?
Good. Very good. Simplicity. I have watched many videos, they are talking about OMG crazy mixes of styles, enable that and this and that will improve some shit dramatically and all of I had is shit. This tutorial is 5/5 stars with no fkin doubt.
I created 4 different pose images and combined them in a basic image editing tool and then use that for reference to get a better grid. I also do cover similar in this video for just the face/head: ua-cam.com/video/MntZa4qLwn8/v-deo.html
When I fully find a decent way I will. It is actually an idea I have been working on but still have not gotten to a good enough result to make a video yet.
Not as well for full body since the resolution is so low. I have done it and is decent with some work and upscaling but the more detail the more things can change when doing that.
We are getting trouble while installing it on windows even having an nvidia gpu on computer so make something that resolves this issue because no one is telling that
Great hacks here. Im new to this. All the sliders at the end are a bit confusing. Ive been on photoshop for 25+ years and living these new tools. Anything i should look out for when creating a faceless character with huge mouth and teeth. These art generators live to stick in eyes and noses. Really frustraiting.
@@KLEEBZTECH Hey thanks, it was just a bit towards the end of the video , you were using the sliders when u ticked a box (ha, i didnt even know they existed). Thanks. Any lora tutorials would be great, someone said tensorart (tensor dot art) might be an option, but again it looks a little baffling, for a first timer ;)
I think one of those options is not necessary anymore the one about the force override of the refiner switch. I've been looking into doing a lora. I did one a while back but not able to do it anymore that way. I will definitely do a video once I find a good option that seems easy for people.
I am considering live but there are potential issues doing that. My system doesn't do well recording while doing some of these things. Plus I've never done anything live so I have to overcome that fear.
I didn't take it as forcing me. No worries. I am considering it soon and probably cover styles. Those are just the 3 formats for images you can have it save as and really would be hard to go into all the details on each one but there is lots of information online if you are curious. PNG is the best to usually use if you don't need to save space. The other two are smaller file sizes but you will be giving up a little quality since they are compressed. As for vector, you can't create actual vector art and have it save to a vector format, you can create images in the style of vector art. You can add "vector art" to the prompt but there is a style called "Simple Vector Art" that I would recommend using. Be sure to disable most of the other default styles since those are geared towards photography. Fooocus V2 is fine to keep enabled. Also may want to try other models like Cheyenne for example. Ones not geared toward photorealism. Also there are LoRAs that might help. There is a logo style as well.
Have you tried running on something like Colab? I am not very familiar with it and only tried it out with ComyUI but it is better than nothing. I may give it a try myself at some point soon.
It sounds like you may be running out of resources. I think there's a way to check that. I don't know if you've looked on the GitHub page and seeing if anybody else has any suggestions. I know I've seen conversations on there in the discussion section about things like that.
If you find the video useful and would like to tip you can buy a cup of electricity for all these image generations. It is greatly appreciated! ko-fi.com/kleebztech
You have some of the best content on consistent characters on UA-cam that I know of. Thanks bro
Thank you for the feedback! Hopefully I can come up with more. 😁
Yes bro!
And please can you make a video on how keep clothes same when combining two pictures in image prompts for a different scene but wanna copy clothes of one character in a complex scene?
I appreciate you🙏
Another awesome video. Thx for the good content you deliver
Thanks!
Awesome contents🎉 i have a question. Can we do something similar to style reference on midjourney v6 using fooocus ?
Not really. But I know it has been discussed. github.com/lllyasviel/Fooocus/discussions/861
I had a bunch of stuff to do so I am way behind. Really look forward to your stuff
Thankfully it is not going anywhere. 😀
Huge thanks :)
A video on Turbo Models would be awesome (all the stuff more advanced than SDXL1.0)
On my list of video ideas. Still messing with the Turbo models.
Useful as always) thanks!
My pleasure!
This is like rocket science. But it makes fun.
great video as always, can you try 2 character in one photo
Something I have on my list.
Thank you so much for a great and helpful tutorial!
Hi, it would be interesting if you made a guide like the one you did to generate the face grids of some previous tutorial but to create more complex scenes with more people, for example "a girl and a boy by the river", "a boy and a girl fighting each other in a castle" and then from that generated grid then generate the actual image, unfortunately there are no guides on this, perhaps not necessarily using a grid but also generating scenes from scratch in depth map
I have been experimenting with ideas for this but have yet to come up with something worth making a video on. But not giving up.
huge thanks, greate video!
Welcome!
So I've got a character made in SD, that I love. But can't get it to do t-poses so I can try to start making a character mesh to import into UE5. Can I use the Focus tool to keep the same character. But get the poses I want?
can you do one for photo realism? i went half through the tutorial and then figured out that it works mostly with comic book
Good. Very good. Simplicity. I have watched many videos, they are talking about OMG crazy mixes of styles, enable that and this and that will improve some shit dramatically and all of I had is shit. This tutorial is 5/5 stars with no fkin doubt.
Thanks! As with something I mentioned in my latest video, start out simple and work your way up. Often we overcomplicate things when we don't need to.
I could not get the grid like you showed. How to get headshots with front, left and right side of head for passport photos?
I created 4 different pose images and combined them in a basic image editing tool and then use that for reference to get a better grid. I also do cover similar in this video for just the face/head: ua-cam.com/video/MntZa4qLwn8/v-deo.html
@@KLEEBZTECH Thanks. Please let us know as soon as SD 3 is available!
How can I contact you for some freelance work?
I have seen Lycoris or something, something like LORA. Is that any good?
They are supposed to be a little better than LoRAs and both can be useful for consistent characters especially if you can train your own.
Thnks bro for effort keep going if you can make video about creation of children storybook
When I fully find a decent way I will. It is actually an idea I have been working on but still have not gotten to a good enough result to make a video yet.
tnks bro @@KLEEBZTECH
I need this type on the architectural side, which means I want four side elevations of a same house in same style of a reference Image
❤
No luck there for me yet.
Does this method works fine with realistic models?
Not as well for full body since the resolution is so low. I have done it and is decent with some work and upscaling but the more detail the more things can change when doing that.
We are getting trouble while installing it on windows even having an nvidia gpu on computer so make something that resolves this issue because no one is telling that
error occurred help about that please
What error?
@@KLEEBZTECH It opens for some seconds and then stops working a notification of error appears
What does it say for an error in the CMD window? Or are you using through something like Colab? Need to know what the actual error is.
Great hacks here. Im new to this. All the sliders at the end are a bit confusing.
Ive been on photoshop for 25+ years and living these new tools.
Anything i should look out for when creating a faceless character with huge mouth and teeth.
These art generators live to stick in eyes and noses. Really frustraiting.
I will try to experiment with that tomorrow. Might need to use inpainting... And any questions about sliders just ask. I may or may not know.
@@KLEEBZTECH Hey thanks, it was just a bit towards the end of the video , you were using the sliders when u ticked a box (ha, i didnt even know they existed).
Thanks.
Any lora tutorials would be great, someone said tensorart (tensor dot art) might be an option, but again it looks a little baffling, for a first timer ;)
I think one of those options is not necessary anymore the one about the force override of the refiner switch. I've been looking into doing a lora. I did one a while back but not able to do it anymore that way. I will definitely do a video once I find a good option that seems easy for people.
The one I did was for my goat. He's no longer with us so I'm happy I created it . The way I did it is no longer an option sadly.
Go live and share your information with us if you don't mind
I am considering live but there are potential issues doing that. My system doesn't do well recording while doing some of these things. Plus I've never done anything live so I have to overcome that fear.
@@KLEEBZTECH I apologize for forcing you, I requested the live because your videos are full of good information... Good luck
@KLEEBZTECH I want ask u what the difference between png jpeg webp in fooocus and could I make vector for logos or characters ?
I didn't take it as forcing me. No worries. I am considering it soon and probably cover styles. Those are just the 3 formats for images you can have it save as and really would be hard to go into all the details on each one but there is lots of information online if you are curious. PNG is the best to usually use if you don't need to save space. The other two are smaller file sizes but you will be giving up a little quality since they are compressed. As for vector, you can't create actual vector art and have it save to a vector format, you can create images in the style of vector art. You can add "vector art" to the prompt but there is a style called "Simple Vector Art" that I would recommend using. Be sure to disable most of the other default styles since those are geared towards photography. Fooocus V2 is fine to keep enabled. Also may want to try other models like Cheyenne for example. Ones not geared toward photorealism. Also there are LoRAs that might help. There is a logo style as well.
Just to add you may want to disable Fooocus V2 as well for vector stuff. Try with and with to see the results.
It seems like everyone has a powerful PC! 😢
Have you tried running on something like Colab? I am not very familiar with it and only tried it out with ComyUI but it is better than nothing. I may give it a try myself at some point soon.
@@KLEEBZTECH GPU disconnecting problem with task, like faceswap in inpainting, and 2 to 3 images in image prompt!
It sounds like you may be running out of resources. I think there's a way to check that. I don't know if you've looked on the GitHub page and seeing if anybody else has any suggestions. I know I've seen conversations on there in the discussion section about things like that.
@@KLEEBZTECH thank u!!!
@@KLEEBZTECH It's being said that Google Colab restricts heavy tasks like inpainting and image prompts, which causes the GPU to disconnect.