PLEASE READ THE VERSION NOTES
FinePorn Turbo NSFW (Krea 2)
A Krea 2 Turbo checkpoint looking for the balance between three things: NSFW capability, prompt adherence and a casual smartphone look. The idea is a crisp, casual image, not a studio render. This is still the starting point, expect a lot of updates.
⚙️ Recommended settings (V6)
er_sde + beta (euler + simple also works)
CFG 1, 10 to 12 steps
Go a bit above the standard sizes, around 1.5 MP: 832x1216 to 1040x1520 or 1248x1824, 1024x1024 to 1280x1280 or 1536x1536
Merged models:
LoveInTheAir by mqblbhai
Krea 2 Turbo Official
FinePorn V5
Baked LoRAs:
Realistic Snapshot by MonkeyForever
SNOFS (Sex, Nudes, Other Fun Stuff) by Ashen3
Detailed Emotions and Expressions by DerekG2
💡 Notes
I recommend starting your prompts with something like "This is an amateur photo taken from smartphone, casual photo"
You don't need a face detailer or a second pass. With the right config, some patience and a bit of seed luck you can get great stuff.
⚠️ Disclaimer
I didn't train this. My job here is merging and baking weights to my own taste, for my prompt style. All merged models and baked LoRAs are credited in the version notes.
ABOUT PROMPTS
You, prompt scammer. You are free to scam my prompt without mentioning or giving the credit, like the majority does here. i won't reveal incredible prompts anymore, so you can keep reusing my old prompts.
Description
Removed realistic snapshot
Added Lenovo at low weight
lowered NSFW V4 weights
added a FP8 version
Changed the internal settings to bake the loras
FAQ
Comments (33)
Fixed seems to be working better, more prompt adherence and better image quality, less unwanted stuffs and also the fp8 is available.
hoping your feedback
Interesting that a prompt like "low quality" helps get rid of that "AI look" a little bit.. I'm giving this model a try.
yes hahaha, that's because the captioning of some loras uses low quality for smarphone-like photos
@simonishere oh, makes sense then.. but yeah, that's a nice prompt to have to combat the "AI perfection" ;)
@ArtiOtaku that's exactly what i want to achieve with this model haha
When I load the int8 version, the filename is the same as the fp8 version.
I don't know if that's a mistake on Civitai's end, but I just wanted to let you know in case you can change it.
By the way, great model!
the int8 has a '_INT8' added on the end, but maybe civit is breaking it, i see it well tbh, but thank your for pointing it and thanks for your comment! :)
Best checkpoint!
Thank you!!
Hello! Interesting model. In many cases, you don't even need to use a censorship-removal LoRA since the model handles it well on its own. Thanks for sharing!
Thank you bud, that's my main target, a model wich can handle several nsfw concepts without needing of loras
Hey ma dude great work, could you by any chance share the weights (intensity) of the LoRAs you merged, privately? If not, no worries!
NVM found it in metadata!
V3.1 has better color, but its skin detail falls short of V2; V2 remains the best in terms of realism.
No idea why, V3 doesn't follow my prompts at all, but V2 is awesome, my go to. Thank you
this looks to be the best checkpoint for krea identity edit: between all checkpoints i tried this is the one that most reliably modifies the scene without changing the character too much
Wow i didn't know it, thanks for the info and thanks for your feedback <33
which version? v3+fix?
@kronos1959777 i've been finding v2's results slightly higher quality but both are great
New version coming on weekend!! i found newer combinations wich gives better results and better variations, maybe v4, maybe 3.2. :).
also in my plans is included creating a well-done article to show how to achieve very long prompts to work well and give unique results, this will take more time
thank you for your work bro! any chance you can release a raw version of it for use with turbo lora?
@q346bq364b3446 Sure, i guess i can do that
Great news, I'll be looking forward to it. Is there a future bf16 version planned?
Go Ahead, still v2 is my go to model for realism, V3 falls off, probably it would be better with better prompting, but i am so bad in it. so an article about prompting is very welcome.
weekend
@laizen1997 haha here in the western still being friday, i'm preparing the model
Works really well with the identity lora and Krea 2 img edit workflows. er_sde beta at 8-10 steps gets mid to great results.
how do i use this fix in comfyui? not a lora right? is it better just to use the workflows ? new to the merged checkpoints what folder does it go in? same as diffusion?
this is a checkpoint, same as the original checkpoint, not a lora, just use the checkpoint on any workflow, it works well
nice
Is it possible to download fp16/bf16 of your wonderful model somewhere?
Unfortunately I can't use quantized versions on my setup yet.
New version tomorrow will have all quants
@simonishere nvfp4 version planned?