IMPORTANT: 3 out of 9 SFW example images under autoflag. Please be patient while they return. :)
Only trained on screenshots from the movie. It doesn't replicate his jacket very well but it does understand him with any other clothing a lot better.
Description
FAQ
Comments (5)
This one seems particularly prone to being hit by queue. 2/3 (edit 2 of 5. Pics 3 and 4 of post.) images I made got hit by it. It's weird because I'm used to that misflagging everything.
Oh, I wonder what might be the reason for that. Maybe it's some kind of hidden restriction by the mods? In any case, you can also use the model more freely here: https://pixai.art/en/model/1956737216058186881
@CoffeeDrop I'm pretty confident it's ultimately the posted images getting the "male child", or "child" tag (or "female child", "aged down" and a few others not applicable to this) when the rating+tag thing gets them. The problem is it's so wildly prone to false positives while not noticing (PG) images that probably warrant one of those tags. I know twin braids and large eyes like to get images slapped with it (only thing I've seen consistently generate a false positive).
Head size relative to body might be a primary factor. Your images that were held have the whole body to work with. Ultimately never going to understand what the rating+tagger bases things off of. At least it isn't handing out PG-13 to every image with a skirt anymore
@NanashiAnon Damn. As if the censorship at CivitAI wasn't crazy enough, it continues to get worse.
@CoffeeDrop This system has been in place for a while, through several tagger infrastructures, and its always been crazy inconsistent. I suspect it's more for (and hopefully better at determining) realistic images. Going through the queue seems to strip the offending tags entirely, regardless of if it's a false positive, which suggests they're not looking for cartoons with that tag.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.








