CivArchive
    Wan 2.2 FP8 T2V Image Only - v1.0
    Preview 118904584
    Preview 118904591
    Preview 118904594
    Preview 118904648

    Wan 2.2 FP8 T2V for Image Generation

    I use this because I like how wan 2.2 generates people. I use res3s sampler and Beta57 scheduler. With lightning 4 step lora I usually do 6-8 steps for quality or 4 steps for speed. I usually use the FP8 14B base model, or swap it out depending on what kind of images I want.

    Description

    Simple base image generation workflow

    FAQ

    Comments (4)

    roger34341976Jan 27, 2026
    CivitAI

    Hi,

    This is excellent!

    When using lora's should we use the LOW lora?

    Also I guess this could be used to generate an image for an I2V workflow?

    jesse_i666
    Author
    Jan 27, 2026

    Thanks, and yes, use LOW loras. This workflow is generally the starting point of my I2V workflow. Sometimes I'll use some other model, but Wan 2.2 is my go-to most of the time.

    jetpisces619Feb 9, 2026
    CivitAI

    Thank you so much. Impressive quality! I recommend it to everyone!!!

    jesse_i666
    Author
    Feb 9, 2026

    Thanks, glad you like it.

    Workflows
    Wan Video 2.2 T2V-A14B

    Details

    Downloads
    319
    Platform
    CivitAI
    Platform Status
    Available
    Created
    1/27/2026
    Updated
    6/29/2026
    Deleted
    -

    Files

    wan22FP8T2VImageOnly_v10.zip

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)