CivArchive
    Preview 34159713
    Preview 34159715
    Preview 34159706
    Preview 34159711

    Excited to release a new quality control LoRA for NoobAI XL v-pred!

    It took a month to develop this LoRA and package it into an easy-to-use format.

    NoobAI XL is an excellent model, but suffers from a variety of issues - wherein you have to use smoothing, quality and detail LoRAs or fine-tunes to get high quality results.

    This LoRA significantly improves coherency of the output latents and prompt obedience using a variety of techniques while retaining all model knowledge.

    If the effect is too strong, start with 0.5 LoRA strength, which will fix many issues without affecting the original tonality as much.

    Effect:

    For reference this is the target image:

    Source: https://civarchive.com/images/46921580

    You can see the LoRA brings out the prompting intention more clearly, and has significant details. The stength:0 image is very far from the intended result, unless you try many different seeds.

    Control LoRA Collection

    A collection of various control LoRAs for use in gens, merges and fine tunes.

    Model Fix series

    NAI v-pred fix: fixes issues with NoobAI XL v-pred models

    Detail series

    SDXL Enhance: controls the details and compositional elements of SDXL-base models.

    CFG ScaleBoost: boost the CFG scale of your gens using a LORA without burning out your latents.

    Original rainbow messageL

    Announcing the release of the CFG Scale Boost LoRA! After much research I've discovered a method which allows you to pseudo-control the CFG scale of your gens using a special LoRA. It is listed under Pony but will work with every SDXL model (including Illustrious XL)! It has the biggest effect on the backgrounds, where it will fill them out and make the scenes feel more complete without frying your gens. Second, it has a detailer-like effect on the main subject, but not always! Experimentation is key. I recommend 0.3-0.7 strength. Released under EA as many sleepless nights were dedicated to this project, which I originally wasn't going to release publicly. Works with negative values as well, allowing you to control the composition and detail of your art to a fine degree, something not possible with typical detail LoRAs.

    Realism series

    Update: Finally added a Pony version!

    Original description:

    This LoRA will make your model either 2D/anime <-> 3D/photorealistic depending on how to use it.

    For example <lora:realistic:-1.0> (negative values) will strengthen the 2D style and <lore:realistic:1.0> (positive values) will strengthen the photo style.

    Potential uses:

    • Adjust the style of your model slightly to make it look more original

    • Make a realistic model more realistic (such as when the skin is too smooth/plastic)

    • Make a 2D/anime model more stylised by making it more 2D

    • Combined with the detail LoRA (linked in the resources) you can reduce details in your anime model and make it look more 2D to create a unique cell shaded style

    This LoRA works well with blended models. It is neutral and won’t overwrite the style of your gens.

    Tip: depending on what model you’re using will determine what range of values will be suitable. 1.0 might be too much for some models and not enough for others, for example. This is especially true of negative values.

    Discord

    Description

    FAQ

    Comments (15)

    daskmasterOct 12, 2024· 2 reactions
    CivitAI

    This worked so SD1.5?

    StagnationOct 12, 2024· 5 reactions
    CivitAI

    Confirmed that it works on Illustrious-XL-based models too, nice one.

    DevilLadyOct 18, 2024· 4 reactions
    CivitAI

    Holy shit thtat cfg boost actually works! Backgrounds look way better

    MerothOct 27, 2024· 8 reactions
    CivitAI

    that cfg boost is amazing!!!

    RickMsNov 5, 2024· 7 reactions
    CivitAI

    This is really good, but depending on the prompt it can turn male to female

    lingondricka580Nov 23, 2024· 3 reactions
    CivitAI

    Amazing work

    GlennRadarsNov 25, 2024· 1 reaction
    CivitAI

    ruin lora character face...how to fix it?

    BigboyblazikenDec 12, 2024· 9 reactions
    CivitAI

    Ive used this a bit with pony models, but it seems to work just as well in Illustrious ones. It seems to do what it says, and different to similar loras, affect the cfg more directly, while some others seem to add details in other ways.

    I tried around a bit with my fav IL model and 2 other detailer loras now:

    https://civitai.com/models/827184/wai-nsfw-illustrious-sdxl

    https://civitai.com/models/510261/spo-sdxl4k-p10eplorawebui

    https://civitai.com/models/1001945/detail-slider-lora-or-illustrious-xl

    Now even this just by itself can improve an image, BUT not always. Ive had generations where i liked it actually better without this, so just try around a bit. The main thing tho is, you probably have to reduce the cfg from what you use normally, even tho i already use only 3-4 on my gens (all models so far seem to have worked well with that for me, even tho i see others use higher values like 6-7, which almost fries my gens), when using this one i have to go down to 2-3. Using it is not as extreme as changing the cfg directly, but you can still easily make it too much.

    Im still in testing, but using all the loras ive listed above together, i actually seem to get pretty good results by just leaving cfg at 3, sometimes 2, and leaving all the loras on 1, sometimes like 3 for the detail slider one cause its build for that.

    This might not work for all models, maybe not even all webUIs, i use forge, and it seems to handle these things pretty well. Try around what fits best for you, but baseline seems to be, it works, and it works with other detailer loras that work in different ways, sometimes great.

    non_trivialJan 7, 2025· 2 reactions
    CivitAI

    Works fantastically with Illustrious models (kudos!), but thanks to CivitAI's weird policies, I can't credit the model as a resource if I have any Illustrious models also being credited. I believe this is because the base model is set to Pony. This also means it can't be used in the online generator on Illustrious models.

    Could the base model be changed to SDXL 1.0?

    waikittyJan 7, 2025· 8 reactions
    CivitAI

    What kind of crack is in this lora!! so good

    4559169Jan 12, 2025· 4 reactions
    CivitAI

    i really love the pony version of this, the only issues i have with it is it will change face and eye shape too much, so img to img with low denoise of .24 or less works best for that, the other issue is it tends to add extra hip and rib bones. still one of the best realism loras though!

    disparaging_wordFeb 3, 2025· 10 reactions
    CivitAI

    this lora changed my life!

    jbsbiomeMar 2, 2025· 3 reactions
    CivitAI

    sometimes it seems to ignore prompt words with a weight after them, like blue:1.5 unless they are in parenthesis ala (blue:1.5) but i'm not entirely sure if that's happening

    Mago000Apr 17, 2025· 9 reactions
    CivitAI

    Bro cooked with this one!

    aviel08Jun 27, 2025· 8 reactions
    CivitAI

    I just wanted to thank you, I use these 2 loras everyday!!! it really brings each generation to life.

    LORA
    SDXL 1.0

    Details

    Downloads
    7,320
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    10/12/2024
    Updated
    4/27/2026
    Deleted
    4/27/2026