first attempt to brew nsfw capable model:
and yes, I am aware to issues with two or or more or just irregular poses...
Description
nsfw, naked, nude encouraging words
best setup that worked for me so far was using UNI as a sampler, non basic number of steps, personally I use 44-46, cfg=7, can vary from 5-11, depending on results, other samplers do work fine- I never was looking for speed but for better results. and this is always a "work in progress" thing
people feel offended when offered an opportunity to test it, so my answer is always the same: you welcome.
FAQ
Comments (10)
what merges went into this?
multiple merges from various sources, most of it just most successful in my opinion, and be merges themselves...goal was to max out erotic esthetics, understanding of basic concepts. I see no difference using this model for generation. if you trying to create something out of this model I would suggest to use the trained sources instead
@BzzzDarklord .... and those models were called? It's kind of common curtsy to list them, and it also helps people understand what to expect out of your model.
@TheP3NGU1N -it would be impossible to recall the names, the weights, the order,- so just leave it be
@BzzzDarklord that's not at all sus that you can't list any... feels like you've probably used checkpoints you werent allowed to.. but whatever. Totally a hard pass now.
@TheP3NGU1N I think you just misbehave by texting this stuff, and really-really got yourself confused me with someone else. get a life and be happy @TheP3NGU1N I don't own you personally anything sus me or not its your own problem
Welp I researched what is in this.. And since you can't remember (apparently) this is what is in your model:
Promissing_Bareskin_V5_.safetensors
Promising_v1_00001_.safetensors
Promissing_Merge_int.safetensors
Promissing_BareSkin_v1_.safetensors
formulaxlXLComfyui_v20Pruned.safetensors
Now next time someone asks.. you know.
@TheP3NGU1N Hope this helps you 'a lot'
I'm not sure why people get so upset not knowing what went into merges.. They aren't gatekeeping it like large corporations (Naiv3 for one). They're kind enough to share it. I don't know if you've made merges, like actually seriously sat down and did merge into merge, into weights by weights and what not, but you really do lose track of things cause there's no real 'method' to it, you mix things together, see if it works, add in something you think if that works, etc etc. Doing so many of that AND having to keep track of everything you merge makes it no longer fun to experiment but it becomes a chore. Also, especially XL merges, you start to run out of storage fast, so you usually delete the ones you've already merged to make room for more.
The way I look at it, is, try it out, if it performs better than your own merges or your favorite checkpoints use it and be grateful for it, if it doesn't then use whichever one performs better or gets your job done. It is nice to know the details, but it is not necessary to get rowdy if they are not provided.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.