CivArchive
    NSFW MASTER FLUX fp16 (lora merged with flux1-dev fp16 saved in multiple precisions and formats) - v1.2 Q8_0 gguf
    NSFW
    Preview 68999204
    Preview 68999114
    Preview 68999042
    Preview 68999067
    Preview 68999223
    Preview 68999257

    original content based on model definition by @Defozo. this model will hopefully allow users to experience the fp16 model content on lower hardware spec'd computers.

    This is a merged model of a base model flux1-dev fp16 and NSFW MASTER FLUX LoRA : https://civarchive.com/models/667086?modelVersionId=746602

    Version 1.2 has added PussyDiffusion - Flux Lora merged in to improve vagina detail:

    https://civarchive.com/models/983498/pussydiffusion-flux

    Version 1.2 hyper 8step models have the content of version 1.2 plus ByteDance Hyper-FLUX Acceleration LoRA has been merged in allowing images to be generated in 8 (sometimes 9 or 10) steps:

    https://civarchive.com/models/691446?modelVersionId=774008

    NOTE: When ByteDance Hyper-FLUX Acceleration LoRA is merged into a model and the model is converted into fp8 precision, the conversion process screws-up the hyper 8step acceleration blocks from the lora. This causes the model to behave like a standard (non-hyper) fp8 model. The good news is it can still be included in the image prompt and still get the benefit of hyper 8step with fp8 model.

    If you like this content, try a version of my other model series:

    https://civarchive.com/models/967270?modelVersionId=1171422


    To
    convert and merge, I used scripts from:
    https://github.com/kohya-ss/sd-scripts/

    To convert to gguf images, I used scripts from:

    https://github.com/Zuntan03/EasyForge/

    Using --ratios 0.8 in the merging command for version 1.0 and --ratios 0.65 for version 1.1

    requires ae, clip_l and (t5xxl_fp8_e4m3fn or t5xxl_fp16) vae/text encoders

    creates fairly good nsfw content. no key words. tested with base flux image generation settings.

    NOTE: if you're not interested in clothing, use the version 1.0 model. it does a much better job with female anatomy. version 1.1 sacrifices quality to work better with clothed women.

    Description

    use t5xxl_fp8_e4m3fn text encoder to reduce vram usage.

    content of version 1.1 merged with pussydiffusion to add more realistic vagina.

    notes from lora owner:

    Trigger words (that sometimes Flux does not follow):

    • bush pussy - Shaved pussy/vagina with hair on top / landing strip

    • butterfly pussy - Pussy/vagina with big lips

    • hairy pussy - Hairy pussy/vagina, simple as that

    • shaved pussy - Shaved pussy/vagina, in the style of cameltoe (innie)

    • asian pussy - Dark lips pussy/vagina on a white skin woman

    • pink pussy - Pink delicate pussy/vagina

    Adding "showing her ass from the back" or something similar to show a rear view may work, as I trained rear view too for every trigger above.

    You can combine more than one. Inpainting them to improve the result is a good option.

    FAQ

    Comments (29)

    q1925615622918Apr 27, 2025
    CivitAI

    Does anyone know why ComfyUI can't detect this model? I've already put it in the corresponding folder.

    tedbiv
    Author
    Apr 27, 2025

    did you restart comfy? do you have other gguf models in the folder it can recognize?

    q1925615622918Apr 27, 2025

    @tedbivI just found out that the loader can recognize it, but it doesn't show up in the model library.

    q1925615622918Apr 27, 2025

    @tedbiv Is it because compyui itself does not support gguf?

    tedbiv
    Author
    Apr 27, 2025

    @q1925615622918 i don't understand. do you mean it doesn't show up in the dropdown list when you try to select model from load gguf model loader node

    tedbiv
    Author
    Apr 27, 2025

    @q1925615622918 no, you need a specific gguf model loader. go to node list and type in 'model loader' see if gguf type shows up. it won't load with regular unet loader.

    tedbiv
    Author
    Apr 27, 2025

    'unet loader gguf' is the name

    q1925615622918Apr 27, 2025

    At first, I couldn't find it in the model library of ComfyUI. Later, I found out that it could be selected in the gguf loader.However, the model library of ComfyUI still does not display this model.

    tedbiv
    Author
    Apr 27, 2025

    @q1925615622918 are you able to load and run it? if so, it's probably something in the metadata comfy doesn't like.

    q1925615622918Apr 28, 2025

    yes,i can run it.Thanks for your reply.

    tedbiv
    Author
    Apr 28, 2025

    @q1925615622918 np. hope you enjoy it.

    zaphodyMay 10, 2025· 4 reactions
    CivitAI

    I've noticed that when using the v1.2 Q6_K gguf version, my prompts for nudity are almost universally honored. However, when I switch to v1.2 Q8_0 gguf, the same prompts won't produce any nudity at all, in fact I can't get any nudity out of that precision/format. Is anyone else experiencing this?

    tedbiv
    Author
    May 10, 2025

    you are correct. i just downloaded the model to test it... it looks like a base flux1.dev content. i will upload a new file. thanks for catching this... i'm surprised no one else noticed this...

    tedbiv
    Author
    May 10, 2025· 3 reactions

    new model uploaded. try it see if it's better. sorry about that...

    zaphodyMay 11, 2025· 1 reaction

    Awesome, much better. Thanks!

    x0_io_0xJul 28, 2025
    CivitAI

    whats better 4K_S or 4K_0 ?

    tedbiv
    Author
    Jul 28, 2025

    Don't know. Never tried 4k0 . I mostly created at request of others.

    hatponytw272Jan 4, 2026· 1 reaction
    CivitAI

    I've finally found a model that meets my requirements; it's really great.

    tedbiv
    Author
    Jan 4, 2026

    glad you like it. you might like one of my 'babes and boobs' series

    Nazuna_VampiJan 5, 2026
    CivitAI

    What's the sampler and scheduler for the v1.2 fp8? The images on the model don't load any workflow in ComfyUI.

    tedbiv
    Author
    Jan 5, 2026

    these images were created in forge. Steps: 20-30, Sampler: Euler, Schedule type: Simple, CFG scale: 1, Distilled CFG Scale: 3.5

    tedbiv
    Author
    Jan 5, 2026

    png info for sample image -

    raw photo, a stunningly beautiful 20 yr old supermodel european women in miniskirt, full body shot, (Fantasy), magical, mythical, medieval, colorful, realistic, detailed, textured, skin, hair, eyes, by Alex Huguet, Mike Hill, Ian Spriggs, JaeCheol Park, Marek Denko <lora:Disney-Studios-Flux:0.2> <lora:midjourney_whisper_innocent_eyes_v01:0.5>
    Steps: 20, Sampler: Euler, Schedule type: Beta, CFG scale: 1, Distilled CFG Scale: 3.5, Seed: 1942410610, Size: 1024x1280, Model hash: ceb8d1e9dc, Model: nsfw-master-v12-Q8_0, Lora hashes: "Disney-Studios-Flux: f53d3f2f901e, midjourney_whisper_innocent_eyes_v01: 4c2fbf17aad2", Beta schedule alpha: 0.6, Beta schedule beta: 0.6, Version: f2.0.1v1.10.1-previous-659-gc055f2d4, Diffusion in Low Bits: Automatic (fp16 LoRA), Module 1: ae, Module 2: clip_l, Module 3: t5xxl_fp8_e4m3fn

    Nazuna_VampiJan 6, 2026· 1 reaction

    @tedbiv I see, thank you.

    david469Jan 16, 2026
    CivitAI

    You say 1.0 is superior to 1.1 for NSFW. Where does 1.2 fit in that? Better/worse than 1.0 for NSFW?

    TIA

    tedbiv
    Author
    Jan 16, 2026

    v1.0 to v1.1 = change nfsw master flux lora merge ratio from 0.8 to 0.65 to prevent nipple and vagina from bleeding through clothing. these are base flux1.dev base model merged with 'only' nsfw master flux lora. for >= v1.2, additional loras were merged to increase detail for breasts and vulvas. read 'about this version' to see what was added. v1.0 was the original posters base mix ratio, converted to smaller models for others. all other versions were my additions/changes. if you're interested in female nsfw content, you might be interested in one of these mixes - https://civitai.com/models/967270/flux-babes-and-boobs. different mixes of herbs and spices to tantalize your taste buds... enjoy.

    marcomentre792Jan 24, 2026
    CivitAI

    Is it possible to think at a flux kontext version? Can you do it?

    tedbiv
    Author
    Jan 24, 2026

    i don't know... are flux1.dev loras compatible with flux kontext?

    tedbiv
    Author
    Jan 24, 2026

    i'm not sure why you need it...? i just used a very nsfw image with base flux kontext workflow and it copied breasts and vulvas perfectly with no loras.

    danteAIApr 22, 2026
    CivitAI

    you have any workflow for v1.0 fp8?? i have some errors...

    Checkpoint
    Flux.1 D

    Details

    Downloads
    8,770
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    4/9/2025
    Updated
    4/22/2026
    Deleted
    4/16/2026

    Files