CivArchive
    LoRA Stack Weight Optimizer - Perfect Blends - Optimizer V1.1
    Preview 123571669

    Add as many LoRAs as you want.The optimizer automatically balances their weights to prevent conflicts.

    • Automatic LoRA balancing

    • Hires Fix included

    • Create Perfectly Balanced LoRA Stacks

    Required Custom Nodes

    1. ComfyUI-Easy-Use
    https://github.com/yolain/ComfyUI-Easy-Use

    2. ComfyUI-Impact-Pack
    https://github.com/ltdrdata/ComfyUI-Impact-Pack

    3. ComfyUI-mxToolkit
    https://github.com/Smirnov75/ComfyUI-mxToolkit

    4. ComfyUI-Image-Saver
    https://github.com/giriss/comfy-image-saver

    5. ComfyUI-LoRA-Optimizer
    https://github.com/ethanfel/ComfyUI-LoRA-Optimizer

    Description

    FAQ

    Comments (9)

    BoerboerMar 8, 2026
    CivitAI

    Having used the optimizer node for a bit I can say its one of the more exiting bits of Comfyui tech I've seen in a while. Takes away a lot of guesswork and more often than not produces cleaner pics if you're the type that uses multiple style loras. Pairing it with the lora stacker node and merge saver are both brilliant ideas. For those using I can also recommend outputting the "analysis_report" into a text node to see what it does under the hood, super interesting.

    nu3DecMar 8, 2026
    CivitAI

    very good, but my 4090 sent my a message Allocation on device This error means you ran out of memory on your GPU
    10loras stack is very hard for it >_<

    LatteLeopard
    Author
    Mar 11, 2026

    There's setting to control vram usage I believe. I also tested this on a 4090. But I did 8 max

    masieu9991191Mar 9, 2026· 5 reactions
    CivitAI

    Will this be available for ForgeNeo? O//w//O

    LatteLeopard
    Author
    Mar 9, 2026

    To be honest. Im not even sure what Forge Neo is. Becuase im more of a model merger than someone who actually uses text2image. Im not up to date on the latest applications.

    Alot of the tools for merging still are built for A1111 or ComfyUI

    masieu9991191Mar 10, 2026

    @LatteLeopard Ahhh, I see.
    ForgeNeo (now called Forge Classic) is the continuous version of A1111, it supports Anima, ZI, Wan, Flux model, and more ^w^
    My brain cell can't understand the nodes in comfyUI, so ForgeNeo is my only choice >w<
    Here is the link to it: https://github.com/Haoming02/sd-webui-forge-classic

    LatteLeopard
    Author
    Mar 11, 2026· 3 reactions

    @masieu9991191 ah. I see what you mean. And I feel the same about nodes. I had chatgpt help me set this up.

    Id have to figure out how to convert the GitHub project to a1111. I'm not sure if my skills even with chatgpt could do that.

    shoes22Mar 12, 2026
    CivitAI

    your link no. 3 is not working
    should have been https://github.com/Smirnov75/ComfyUI-mxToolkit

    LatteLeopard
    Author
    Mar 16, 2026

    oh shit. Thanks!

    Workflows
    Illustrious

    Details

    Downloads
    244
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/10/2026
    Updated
    8/13/2026
    Deleted
    -

    Files

    loraStackWeightOptimizer_optimizerV11.zip

    loraStackWeightOptimizer_optimizerV11.zip