Gives a “Hey Arnold” look to characters.
Place the file in your models/loras/ folder and add <lora:Hey_Arnold_style_illustrious_by_dom83-10:0.9> to your prompt.
This is a first version. (Well, my third try actually.) It captured the style of the characters but it can't generate the characters from the series, except a few ones, and it did not capture the style of the backgrounds. I will keep trying to train a better LoRa.
My training pictures are very bad quality and the result suffers from this. I can't find better quality screenshots of the characters. If you can, please help.
The model I used for my example pictures is zukiCuteILL_v25 ( this link is dead, you can download it from this other website : zukiCuteILL_v25 , or try this new version instead: Zuki New Cute )
Description
FAQ
Comments (13)
amazing ! by the way i have screenshots from the show, i just have to find it somewhere on a hard drive!! but this is tvrip too in low resolution. Which kind of characters/ background you need ?
Oh thanks for offering your help. I'd take any good quality screenshot showing any character alone with no part hidden by an object, and any background with no character.
@Dom83 all screenshots are in 640x480 TVrip format. Unfortunately most of them has been lost in my cloud storage that i no longer have. Fortunately i do it again last year so i have screenshot again, almost 750. It is pretty focused on female characters (Lila,Olga,Helga and Rhonda.) but still there are background only, items, male characters, minor characters or even a little full body yes
@JohnB79 Good. Please zip it and upload it somewhere like mediafire or dropbox. I'll see which pictures are better than mine to improve the quality of my dataset and model. I remember I had very few good pictures of Nadine and Sheena for example. Are there other series you took many screenshots of?
@Dom83 i sent you link by pm. I take screenshots from every show or movie i can watch so yes there are others lol
@JohnB79 By any chance do you happen to have good screenshots of Braceface, Martin Matin or Shuriken School in your vast collection? I want to train models on these but the episodes I have are in 320x240.
@Dom83 I don't have them. Still for Braceface, I have episodes in 480p, but no screenshots yet. 🤔
I realllyyyyy love your loras!!!! <3
What checkpoint model did you use to generate the previews?
zukiCuteILL_v25 : https://civitai.com/models/1000708?modelVersionId=1185703
edit: it seems to have been deleted for some reason :(
@Dom83 Seaart . ai has most of the removed safetensors / Checkpoints as long as the word isn't censored then you have to search the censored as a word number of "?"'s so if the word is taboo and 7 letters long like, big hairy "Monster", its Big Hairy ???????, or say its feral "fox", = feral ???.. So its there if you know the baseline name but add ?'s to search the archives. Even logged in.
Here ya go bud https://www.seaart.ai/models/detail/3f7d2924be9aa904d5529ca3a43ee227
Also Fair warning I can see 100% why they deleted it that's generating some rather sus content but to be fair Any Checkpoint can sadly .. Thats just the price of images and size being built on scale and position
There is also this if you want to stay more local https://civitai.com/models/1437224?modelVersionId=1687032
@XX0XX Thanks. I updated the link.
@Dom83 No prob :D
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.