CivArchive
    SamsungCam UltraReal - v1.0 Qwen
    NSFW
    Preview 103501646
    Preview 103501616
    Preview 103501911
    Preview 103501179
    Preview 103501638
    Preview 103501737
    Preview 103501936
    Preview 103507050
    Preview 103507083
    Preview 103508788
    Preview 103535928
    Preview 103536073
    Preview 103536340
    Preview 103740395
    Preview 103740557
    Preview 105519466
    Preview 106338131
    Preview 106385378
    Preview 106719309
    Preview 107960105

    --Qwen2512
    Trained with the best my technique. Best use with NiceGirls lora (if u need better female). Also works good with other loras + works good with some nsfw loras

    --Qwen
    I have prepared a version for Qwen. It looks quite promising; however, work on its enhancement is still ongoing. Examples with girls i generated with my lora NiceGirls

    P.S: yeah, I changed epoch in models manager, cause 7k steps gave overtrained result, and in process of testing I understood that 3k is enough


    🗺️ Backstory

    Lately I kept getting DMs like “When will you drop an iPhone‑style LoRA?” Problem: iPhone LoRA already lives on Civitai—and, well, I shoot on Samsung. SamsungCam UltraReal captures the S‑series look you can’t fake with generic “phone” prompts

    After some tests i noticed that the best to use with UltraReal Fine-tune or UltraReal Abliterated with 60 steps (yeah, so much steps gives pretty good details, also text usually looks much better).
    Use 0.7-1.0 weight (depends on artifacts in scene)


    🎨 What it brings

    • Vivid, high‑contrast colors

    • Smartphone softness – just enough in‑camera smoothing to feel “real phone,” not plastic CGI.

    • Highlight‑happy HDR – clipped whites around street lamps for that over‑exposed‑yet‑cinematic pop.

    • Less bokeh, more context


    🤝 NSFWloras-friendly
    Tested with NippleDiffusion and MysticXXX (For some reasons is not visible in Suggested Resources)

    You can try to generate in 2MP for more realism and details, but be ready that in some scenes hands will be broken. But still looks very good imo in usual 1MP.
    In most scenes removes blurred background (sometimes it could be still blurred). Also improves details a lot even on default flux.dev (cause mostly i train all loras for my ultrareal fine-tune checkpoint

    P.S: yeah, first version is slightly overtrained, but i glad with a result. v2.0 already WIP

    Description

    FAQ

    Comments (24)

    9401290Oct 4, 2025· 3 reactions
    CivitAI

    Will this one be ported to Wan2.2?

    Danrisi
    Author
    Oct 4, 2025· 1 reaction

    Maybe. But I noticed everything I train for wan still looks like default with some minor changes

    flo11ok874Oct 5, 2025· 8 reactions
    CivitAI

    Qwen Lora works also fine with Lightning-4steps v2 Lora.

    3193975562yu312Oct 6, 2025
    CivitAI

    This is really great. I'm a Chinese user and I'm very eager to add some Chinese characters. May I ask if we will continue to upgrade this aspect in the future?

    Danrisi
    Author
    Oct 6, 2025· 1 reaction

    Hi. I didn't specifically train on hieroglyphs; the model simply draws Chinese well out of the box. With my generation, I wanted to show that I've trained a model that strives to be as flexible and balanced as possible without breaking anything

    abhiraj320Oct 15, 2025
    CivitAI

    You said to load this Lora with UltraReal Fine-Tune.
    But that is based on Flux 1D correct?

    So how can we use it?

    Can you give any workflow for it please on how are you saying?

    abhiraj320Oct 15, 2025
    CivitAI

    What sampler do you use with this one?

    abhiraj320Oct 15, 2025· 1 reaction
    CivitAI

    Can you share workflow here? I tried using this with qwen, but not at all getting the dersired result.

    Can you share workflow link or something for me to check and analyze as there are so many other parameters.

    angelmaderoia274Nov 7, 2025

    Hi, you just need to drag the image from here to the ComfyUI interface (you can split the screen and have Civitai on one half and ComfyUI on the other; this will open the workflow. After that, you'll need to download and configure what you're missing; ask chatgpt for help). You're welcome.

    abhiraj320Oct 15, 2025
    CivitAI

    Do we have to put something in the prompts or something?

    A workflow here shared for the girl generated in the pink dress would be very very great to do learning.

    10398328Nov 15, 2025· 2 reactions

    saaar there's literally a "Trigger Words" section in the "Details" list saar please saar please read saar

    franklynsotelo72838Oct 15, 2025
    CivitAI

    No matter what strength I use between .7 and 1.0, I keep getting banding on the final image. Any suggestions?

    Danrisi
    Author
    Oct 16, 2025· 1 reaction

    yes. do u use .safetensors (fp8 or fp8_scaled)?
    UPD: u are talking about qwen or flux?

    franklynsotelo72838Oct 16, 2025· 1 reaction

    sorry. qwen

    absamplephNov 23, 2025

    @franklynsotelo72838 Have you tried this: https://github.com/spacepxl/ComfyUI-VAE-Utils/tree/main

    MollyAnneOct 22, 2025· 2 reactions
    CivitAI

    WOW!

    Every 11th image something magical happens and it's so shockingly realistic that it almost stops time.

    Thank you

    YinsuTarvoOct 31, 2025· 2 reactions
    CivitAI

    Excellent, your loras are the best! Is there any chance of making versions for Illustrious and Pony? 👌

    Danrisi
    Author
    Oct 31, 2025· 2 reactions

    Thanks so much <3 You literally read my mind about Pony. I was just testing the new v7 today for realistic images, and it looks really promising. So a LoRA for it is definitely in my plans

    z1112Dec 25, 2025· 2 reactions
    CivitAI

    nice

    flo11ok874Jan 4, 2026· 5 reactions
    CivitAI

    If you wonder this Lora works fine with new Qwen 2512

    AnimaXxJan 5, 2026

    It should I've used old Qwen lora with the new Qwen image. It works but you have to lower the value I think will increase the steps. I'm not sure, especially when you have multiple lots. I'm still new to ComfyUI

    temptatiorJan 13, 2026
    CivitAI

    Nice LoRA! I was wondering if this works with Qwen-Image-Edit(2511) aswell, or only for Qwen-Image?

    ackibanMar 10, 2026

    I have same question

    LORA
    Qwen

    Details

    Downloads
    16,217
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/1/2025
    Updated
    5/26/2026
    Deleted
    -

    Files

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.