to keep close to design try any of these - or put in Neg to reduce chance
may need to add 3d, or realistic, in neg for 2d/anime - suggest weight between - 0.9 - 0.6
PonyXL - TW is Gamoraxl, - may update to v6 later which should help accuracy
iLLXL - TW is GamraXL, - has little to no 2d training - may need use of NEG for 2d styles or less weight - may update to v2 at a later date
For appearance
Gamora, red eyes, multicolored hair,
green skin, or green skinned female, - if needed
makeup, or black facepaint,
hair over one eye, may mess with adetailers
for outfit
black and white bodysuit, armor, white boots, white gloves,
pouch, may randomly spawn - Neg to reduce
variations of suit may occur
no clue how well spawning holding sword, will work
Description
FAQ
Comments (9)
I'm aware we all can do the same content, making it a diverse choice -- am not complaining btu I HAD TO LAUGH because i've been CHICKEN to train her for two months now -- have her dataset tagged and everything XD and You beat me to it (mind you mine are i never train mine the way I should XD)
I usually check before I plan to make a lora, only reasons I do it anyway is if I already started or if people ask. in this case I've seen only the MCU version so I was like "ok why not?"
and if people didnt/couldnt make their own version of the same lora half the loras on this site be removed/pointless
@True_Might I do want to try yours regardless, I really enjoy seeing the different syles people do - mine's gonna be weird but i'm gonna not change the data and just see what it does cause it's got all MCU and Game and comics mixed in XD -- I tend to do that and get yacked at
@True_Might And even if you hadn't check it. I mean, you can try literally what you want even if its done or not. Diversity is always good. :)
@LDWorksDavid Well that's why True Might just inspired me to ACTUALLY train mine, because I think they'll go hand in hand, because the video game is in my data, it just doesn't understand it AS well, or i just don't prompt well XD -- I like mixing both my training and other people's loras when i go to eventually play not just test -- I used to do that for testing but i remembered using too many additives confuses people :3
@duskfallcrew Ay, nothin wrong with a bit of friendly competition. Multiple options for the same thing, people get to choose based on preference and quality, or maybe one works better for their preferred checkpoint than the other. Choices and variety are the spice of life.
well all I meant was atleast, is that id rather focus on loras that's missing than add more (variation) of the same, I don't actively look at a lora of a character and be like yeah I wanna do that one as well. But I don't see anything wrong with inspiration or competition. all it does for me is make where I don't need to bother making a lora if its already decent enough.
im thinking of the least popular or obscure characters 1st because its obvious there's going to be like 10+ loras for well known characters. my endgame is to stop making loras at some point because nearly everyone is available. I've already finished all the characters I wanted several weeks ago im just grabbing any strays or whoever is requesting at this point.
I always try to do a LORA with concepts that I don't see here but if I create a concept that already existed here, 0 problems. But it's true that with all we have that can train it's difficult not to find new things.
Can I suggest Pony model next? It performs amazingly