We have updated our quality standards found under our uploading guidelines.
This was done mostly because modern AI generations are of much better quality now than they have been in the past, and we would like for this archive to focus on displaying the best work, not just any and all work. Please take a moment to review the guidelines before uploading going forward now. Older submissions will be exempt for this at this point in time, but keep in mind in the far future we might clean those up as quality continues to improve.
Dimot
MemberIt's... gorgeous.
I often wonder how carefully tweaked the AI generation must be and how much touch up is needed... It's just too good!
Wait a minute... I don't even know the slightest bit of what's going on behind the scenes!?!
Tell us your secrets!
oldhroft
MemberThanks! But there is nothing too fancy. Since both characters have their own LoRAs - I genned them separately, then bashed images together in Krita and inpainted the seams with some additional details. It's not the only way to do it, but I found it's the quickest method for such case. The gens are done on Bananastrike model with very simple prompts, I try to stay below 75 tokens if I can.
Used prompt for Shadow image as an example
anthro female, [sergal:velari:0.25], full-length portrait,
black body, white spots, white (cocktail:0.0) dress, slim, small breasts, (neck tuft:1.1), fluffy, shy, blush, happy, smile, green sclera, looking at viewer, beak, digitigrade, 3 toes,
mucknagabe, <lora:velari_v1.0_compass:0.7>
Negative prompt: muscular, [worst | bad] quality, [deformed | disfigured], poorly drawn, [bad | wrong] anatomy, [extra | missing] limb
Steps: 12, Sampler: Euler a, Schedule type: SGM Uniform, CFG scale: 2.5, Model hash: 446d23fb70, Model: saxl_bananastrikeXL_v10
Dimot
MemberππΏπ
The secrets are known!
Top shelf respect awarded.
Login to respond Β»