CivArchive
    Flux Upscaling Generalist - v1.0
    NSFW
    Preview 44947813
    Preview 44947836
    Preview 44947891
    Preview 44947923
    Preview 44947964

    V1 - General Upscaler + reskin
    The Upscale Generalist works for pretty much all image types. It uses a converted FP8 Version of the JasperAi Upscale Controlnet (for faster infernece) as conditioning combined with SDUltimate upscale. There are two different versions of the workflow. One is a general upscaler that works with every image style while V2 is optimized for upscaling faces, focusing on a more realistic skin texture. Jump with key 1,2,3 between each upscaling pass and control the strength via the silder. With key 1 u can bypass the groups. Pass 1 and Two are good for general upscaling, Pass 3 is made for retexturing skin - it uses @denrakeiw new skin LoRa for this process. All settings that I prefixed in this wf work pretty good for me but pls feel free to adjust it to your needs.

    Big Shoutout to @denrakeiw for the nice jam session

    You can find all files here
    Base Model: https://civarchive.com/models/647237?modelVersionId=724149
    CNUP: https://civarchive.com/models/1028044/flux-controlnet-upscale-fp8
    Skin Lora: https://civarchive.com/models/1018511/flux-skin-detailer-realistic-skin-dora
    Upscaling Models: https://drive.google.com/drive/folders/11ZmP-Wp96dfcDMUZq0gu1m8kUYOS9tWU?usp=sharing

    V2 - Regional Detailer/Upscale

    With V2 of the Upscaling Generalist u are able to upcale more detail in precise areas. Use segfromer or sam2 for auto masking and define the region that should get a more attention. The Noiseblend passes blend a noise layer on top of the given mask, like a region specific nosie injection to get more regional depth and detail.

    Jump with key 1,2,3 between each upscaling pass and control the strength via the silder. With key 1 u can bypass the groups. Pass 1 and Two are good for general upscaling, Pass 3 is made for retexturing skin - it uses denrakeiw new skin LoRa for this process.

    For less hallucinations use low denoise values like 0.13-0.08 and the multiplysigma between 0.5-0.2. To achieve the best possible results you should work iterativ, always go a bit lower per pass. Here a quick example pass1denoise=0.13 pass2denoise=0.099

    All settings that I prefixed in this wf work pretty good for me but pls feel free to adjust it to your needs.

    Have fun and enjoy.

    S/O to denrakeiw for the Idea, check his profile he is a amazing creator. https://civarchive.com/user/denrakeiw

    add this instead of the upscaling model:
    CN-Inpaint FP8: https://civarchive.com/images/45719781


    Tipps for segfromere installation

    First step Acticavte the wf, and create the segfromer folder in your comfy model directory, dont miss to use the right name for the folder as written in the error or follow the hf link for the right name.

    Find all Model files here ->

    https://huggingface.co/chflame163/ComfyUI_LayerStyle/tree/main/ComfyUI/models/segformer_b2_clothes

    Then download all files and place them in the folder.

    Refresh the wf and start a new run

    Description

    FAQ

    Workflows
    Flux.1 D

    Looks like we don't have an active mirror for this file right now.

    CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.

    Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.

    Details

    Downloads
    447
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    4/24/2025
    Updated
    5/1/2026
    Deleted
    4/24/2025

    Files

    fluxUpscaling_v10.zip

    Mirrors

    CivitAI (1 mirrors)

    fluxUpscaling_v10.zip

    Mirrors

    CivitAI (1 mirrors)

    fluxUpscaling_v10.zip

    Mirrors

    CivitAI (1 mirrors)

    fluxUpscaling_v10.zip

    Mirrors

    CivitAI (1 mirrors)