V2.5 for ZIT is trained on base and plays along well with other loras on ZIT. Improvement focused on accurate female genitalia.
Added version trained on Z-Image "Base". All samples are genned on Z-Image Turbo, where the main benefit is that it seems much more resilient on stacking loras and has less of an effect on overwriting features or look from other loras.
Added version trained on ZIT, the effect is pretty subdued but cuts away from the cinematic color with very little impact on detail on my tests, also steer models preference bit away from asian women. Works without trigger phrase, but was trained with "Low quality cellphone photo of". For accurate representation of genitals, look for some other lora.
Added WAN 2.2 T2V models, trained on the same dataset as V1. All gens done with Lightning 4 steps loras @480p, so likely there's possibility of improved quality by improving the workflow.
V1.5 is more consistent and degrades Flux quality less, but the nude selfie might not be as authentic or perform as good as V1.0. Art nude dataset was added and nude selfie dataset was expanded considerably.
Flux lora trained on non-specific set of 50 or so images of amateur nudes, mostly selfies and mirror selfies. Knows "nude selfie" and to lesser degree breast sizes and body shapes, prompt following for body sizes is limited but it's definitely not as stuck on the model body as base Flux.
Anatomy is not perfect by any means.
Description
Added bunch of similar images to the dataset and also similar sized art nudes dataset.
FAQ
Comments (11)
With all due respect, I think it's stupid to put the title Amateur nudes in civitai and nudes_3k_prodigy in LoRa, because after a while you don't know where that LoRa came from.
Fair. You can also rename files on your computer if it bothers you greatly.
@konflikti Yes, I agree. I downloaded this lora and I appreciate you sharing it, but I wanted to look at what the prompts were like and I went crazy trying to find which one was the Lora because the name didn't match. nothing more.
Can't agree more, but I think civitai can force some naming structure on file names, it will make things organized, for example: LORA-XL-ID#-CARTOON-STYLE-CREATOR-NAME.safetensors "while ID# is placed in a box in the download page underneath trigger words and hash, this will make it easy for us to offline download LORA pages and easily match the downloaded page with the file name for LORA or model.
@konflikti Yes, people can each fix your arbitrary and lacking context file naming.... OR....
I completely agree. I can't count how many times I've downloaded a LoRa only to look at it later and think 'What is this even for?' Even if it's clear at first, after some time it becomes impossible to identify. I really wish the post title and filename would at least match - that way I could cross-reference my download history to figure it out.
I rename each lora to match the name on the description, then I keep an excel sheet with all the important stuff like recommended strength and keywords. I'm also weird.
ha, look at LoRa named: 22WPJH484SFVBSE4J2ZDEZQ8H0 :)
Yes, the model is appreciated (thank-you) - but it is frustrating.
Indeed we can rename them, but especially for those of us who use tools in Comfy to download, organize, apply images, etc. not naming the file accordingly can cause a lot of unnecessary work and confusion.
Nice model though, thanks again! :D
@tupu You could use a node that auto directs back to the civitai page like Power Lora Loader (rgthree) or you could make a table to keep record. These people are not obliged to hold your hand or deal with your skill issues.
If you're incapable of organizing yourself then you have choices, the most obvious would be to not use it. You could then take your whining and bitching to a hug box like reddit where users could console you and fortify your shortcomings with compassion.
Amateur Nudes as a title would also include males. So, the title should be Amateur Female Nudes.