Yumi (Sayaka) from Ape Escape (Saru Getchu)
(Ah, the good old random name changes)
ORIGINAL red hair, twintails, green eyes, hair bow, pink hoodie, short sleeves, wristband, white skirt, bike shorts, socks, sneakers
COWBOY red hair, twintails, green eyes, camisole, blue scarf, denim shorts, blue shorts, cowboy hat, chaps, pink gloves, belt
CHINESE red hair, hair bun, green eyes, bun cover, chinese clothes, clothing cutout, ribbon, detached sleeves, white pants
ARABIAN red hair, twintails, green eyes, earrings, arabian clothes, harem pants, pointy hat, forehead jewel, vest, neck ring, armlet, bracelet, jewelry
BLONDE blonde hair, twintails, green eyes, pink dress, tiara, short sleeves, gloves, wrist cuffs, ribbon, heart, short sleeves, wings, feathers, white pantyhose
FANTASY red hair, twin braids, green eyes, witch hat, white cloak, black pantyhose, boots
NINJA red hair, twintails, green eyes, headband, hair ornament, pink scarf, ninja, japanese clothes, sleeveless, obi, white shorts, thighhighs, tabi, sandals, fingerless gloves
This lora as a bounty for NanashiAnon Creator Profile | Civitai
Description
FAQ
Comments (15)
it's so funny preview LUL
A little request. Could you add the trigger words to the model description pls? It's easier for save them
The lora works without trigger words. The tags used for each one of the costumes, are listed in the model description. So, if you want one of the costumes you have to put all those tags (or some of them) on the prompt.
Now, I'm aware this makes prompting more tedious while compared to loras with trigger words, "costume1" is shorter than "red hair, twintails, green eyes, hair bow, pink hoodie, short sleeves, wristband, white skirt, bike shorts, socks, sneakers" but this way makes it more easy to combine different pieces from the different costumes. Without trigger words is more easy to combine and mix pieces in the lora and when combined with other loras.
@neclordx I mean, if you can add that list of tags for each costume
@Konoko ??? Sorry, the tag list for each costume is on the description already. Or may be its not showing up? Im seeing them. May be they are not visible for everyone else?
@neclordx Haha, maybe I'm confusing you. In the model version section, there is an input where you can add the trigger words of a lora (the information that appears in the right sidebar). There you can add each costume tags set (I recommend you 1 set per tag, you can see an example of what I mean in this lora: https://civitai.com/models/17922/ninomae-inanis-hololive-5-outfits), with this if someone downloads the lora with a helper or automatic script (or has a civitai data extractor), they can get the trigger words (or in this case the costume's tags) automatically from the API information.
oh, Ok. Good to know XD I added them (I think)
@neclordx Thank you so much c:
There's a missing E in the title, which hinders finding in search.
UOH
Can you do an Illustrious version (with a trigger word)?
IF I ever did it, and its a big IF because the number of lora I want to do keep piling up and the time I have to do them is reducing (bacause life happens, no big stories here) it would not have a trigger word.
What's the point of a trigger word anyway? Honest question here, after a year playing with AI I still never understood what is the point of it, if you are using a lora I understand you want to use that lora, I dont see the point in needing a keywork to make the lora you put already work.
What's the trigger word for? I ask for curiosity, what is exactly the applicable purpose?
@neclordx You technically already have a trigger word, it's just "red hair" and/or "twintails", which causes a bunch of uneeded issues. As for why give her a dedicated word
1: There's no downside to using one.
2: Using hair color as trigger makes it way worse for monochrome images and some more far from norm styles.
3: You can actually put other characters in another's outfits by using the trigger word as a negative and using the other prompts in positive.
4: You can weight the trigger itself, which has uses for troubleshooting combining LoRAs and effects distinct from weighting the LoRA
5: Illustrious handles multiple characters at once better, but to even try it it needs to be able to call in the single word.