CivArchive
    Trio - TripleLatents for SDXL & IPAdapterV2 - v4
    NSFW
    Preview 9046219Preview 9046248Preview 9046322Preview 9046267Preview 9046334Preview 9046344Preview 9046359Preview 9046358Preview 9046367Preview 9046437Preview 9046563

    July update: Trio-Tpose: Master 2D Puppet Animation

    All other workflows in Packs:

    (scroll down for details)
    v7 video:

    v6 video:


    We have a new IP Adapter since the end of March 2024.


    current version: v7-Pack
    (includes previous packs)

    • TrioAdvanced - uses the advanced IPAdapterV2 apply node

    • TrioCombine - how to use ipadaptor encoder to mix embeds

    • TrioNoise - noise as image negative input

    • TrioStyle-InstantID - InstantID (face target) feature with Style transfer

    • TrioWeighted - demonstrates how to adjust image input weights

    • TrioZero - instant lora method, use small dataset as input images

    Bonus Workflow:

    -HD, versions use iterative upscaling


    V6 (mostly SDXL Style workflows)

    Trio Basic - the most basic triple latent, with SDXL IPadaptor style.

    Trio Standard - The same as Basic, but introducing IPAdaptor Embeds in the workflow.

    Trio Stack - allows 3 input images to be used with the Style transfer.

    Trio Magic - used magic prompt with auto negative prompt, aspect sizer & 3 latent passes with iterative upscaler using remacri.

    " -HD " variants add iterative upscaler
    " -plus " variants add magic prompt

    more will be added in the next pack.
    v4 is just an early version of Trio Magic

    Description

    Workflows
    SDXL 1.0

    Details

    Downloads
    62
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/3/2024
    Updated
    9/27/2025
    Deleted
    -

    Files

    trioTriplelatentsFor_v4.zip

    Mirrors

    CivitAI (1 mirrors)