This lora allows you to remove the censorship of Flux on this middle finger gesture.
Check the exemples prompts to trigger.
Any buzz donations will be appreciated 🥰
Description
Light version !
FAQ
Comments (17)
This will come in extremely useful when I need to communicate succinctly with HR at my workplace.
Could someone please post a proper prompt? I only get 👆 or ✌️ but no 🖕
"raw photography of Mario doing the middle finger at viewer in a random place, highly detailed, photographic style, shot with a 50mm lens, natural lighting, in high resolution 8K"
yeah maybe we can use this to do a handsign language generator. Maybe it could generate hands in real time and translate real videos in real time and actually help people. :O
I propose we use a thicc version of Belle delphine in bikini (with big boobs) as an actor for this translator :D
ha cool ça faisait un moment
can it do both hands?
yes it can i have tried and it worked perfectly!
Great work my dear friend, u really should keep releasing more loras like this that focus on removing the censorship from FLUX DEV!! BTW sometimes this lora affect the person lora face that i'm using but i did mess around with an extension called Block Lora weight from github that let me play freely with the weights of your lora so that it dosen't affect the character lora that i'm using, Great Job again! :)
Thanks for your comment.
You are right, you must play with weight, especialy with others loras !
Where do you use this block lora?
@elfheeim Probably in ComfyUI !
can u share the workflow? all those loras affect the my person lora face, if theres a way to use other loras with my portrait lora without affecting face, that would be fantastic.
Idk if its just me, when I use this, subtle grid covers a whole image. zoom in, you can see them. I also tried other middle finger loras, they all behave like this one. so weird. btw i'm using v4. generation resolution is 1280x1920.
Hello,
Maybe it come from your specific resolution, I never notice this problème.
Can you see it in showcase image ? or Generated by users bellow ?
@Isulion Hello i'm sorry. the problem wasn't your lora. it was flux thing. several peeps already experienced those stuff. at 2mp the grid barely appears. with 3mp it is much more likely to happen.
@fluxxes Thank you for the feed back !
@Isulion Hello its me again, can I ask how do you prepare dataset and prompting to train this type of lora? since i've been training only celebrity face loras Idk how to do others. my usual process goes like this. I prepare 1024x images with faces and throw them at Gemini for training prompting to make txt files. say if u wanna make a lora forming a heart shape with hands, how do u prepare dataset and prompting? I got bunch of image sets of a person forming a heath shape with their hands. Should I remove faces from image sets to prevent this lora from affecting other face lora? and for training prompting, should I type everything about images except for hands forming a heart shape?