GOLDENJUICE

There's two versions of this lora, same dataset. minor changes on tags

what changes ?

second version is more simpler to apply the lora to the image since besides the (goldenjuice) activation u are pretty much ready to go (and put /poseb/ to get more nice results, /poosea/ dont seems to work i dont know why BUT SOMETIMES WORK, but you know, be advised) ((/posea/ should be frontal closeups and /pose/ rear closeups))

this one i did using the first version with the following prompt : "(goldenjuice) (peeing) perfect ???? picture of supergirl peeing (multiple views) <lora:GOLDENJUICE:1>"

this is using version 2 with simple tags with the following prompt : "(goldenjuice) perfect ???? picture of supergirl (multiple views) <lora:GOLDENJUICE2.1:1>" version two also is way more try/error since i guess my model dont like when you try to generate an image with two things happening at onece (or just me who dont know how configurate tags, test both versions if u can, and i would love any feedback to help me with this tricky thing

>>BEWARE LONG POST OF NOOB LEARNING STUFF<<

OKAY GUYS im still on this pretty journey of learning and trying my best to make things work as they should be ! soon i will release the version 2 of my first post, but since that one got way more complex than i thought i would be i decided to work on other lora i was making, which is the one you are looking now, GOLDENJUICE. just like my first post consist in closeups that in the same time shows the person aside, this time is that with girls peeing ! 😎 i only tested on the "anythingv3" (that is not anythingv3 but has that name for some reason) and ???????? really nice results

STILL TRYING TO FIGURE OUT HOW THOSE "tags" SHOULD WORK, i did two versions of this lora, one i didnt change the tags that was generated, and the second i changed them a little bit, so it could be more flexible on the prompt BUT THINGS HAPPENED and i think still dont know how to make an lora that can have "customizable" prompts ): (YET)

>>AND KEEP IN MIND EVEN WITH THE BEST PROMPTS (THE PICS I POSTED) I CAN STILL GET WEIRD OUTPUTS LIKE sometimes you only gonna get the close up, without the person silhouette, im aware of that and have no idea how to fix at the moment ): but like 80% of times u should be good

ANYWAY the first version, you will need lot of prompts to make work, but is more flexible when it comes on how you want (front view, back view) the second version is way more simple to make work since i changed the tags, BUT this way it got harder to get front view close ups so in second version is almost SURE that you gonna get an rear view from the close up (i not complaning) so thats it, pick your poison try it and SHOW ME YOUR RESULTS I WANNA SEE IT