Created by GeomanticArts and please consider supporting the trainer at geoloras.com
Ankha. From Animal Crossing.
Example images all made with AOM2-7thV3_C_40 at weight 1
See the type chart for your desired level. Mix and match, try things out. Use one or many of these. Results can be quite unexpected if you do not.
'Ankha (Animal Crossing), snake hair ornament, white dress' - Example prompt
LoRA metadata has all the tags used.
The trainer's twitter can be found at https://twitter.com/GeomanticArts
Trained on NAI.
Description
FAQ
Comments (11)
Oh my god. You gave her a nose.
It looks a lot weirder without a nose
For some reason when I try to load up this lora, it shows up as <lora:lora:1> instead of the correct name, even though the file is named correctly. Anyone else having this issue with Automatic1111?
What do you mean 'shows up as'? Where does it appear as lora:lora:1?
@RockHardArt In the prompt text box
@RockHardArt Seems theres an update in Automatic1111 where it loads up the lora using metadata in the lora file. You must have left the data blank, or it had "lora" by default.
I just came across another lora file that tries to load as "Addams1"
I wanna tell you, not having that lil (i) with the prompt in the example images is a real annoying thing. Sometimes, you need that initial starting point to really figure out how to work a lora well.
I figured that with the example prompt it the description it would work just fine. For instance the second image just has the prompt
'highres, detailed, soft lighting, outdoors, desert, wrapped in bandages, white dress, ankha [animal crossing], snake hair ornament, furry, 1girl, smug expression, hearts"
which seems pretty expected. Is it giving issues?
@RockRockRock Not exactly giving problems, but this has a wide range of bodytypes it can do, so it may be cool to know what to do to get each and every one of them exactly, which is what that (i) is GREAT for.
@antimatter_john I did actually think the chart there was helpful with the range of bodytypes. I used to think that if the LoRA didn't work without me explicitly telling people how to use it, it was just a bad training. Since in my mind if it took explaining how to work it, the training settings just needed to be changed. But now more people have been asking to just have the prompts to use so that has been different on the new uploads and I just include the png info.
Name of this lora is <lora:AnkhaAnimalCrossing1:1> and not <lora:ankha:1> or whatever there is in the sample images below.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.