CivArchive
    WAN2.2 14B GGUF (I2V,T2V) Low-Vram Workflow - v1.0
    NSFW
    Preview 91194243
    Preview 91194241
    Preview 91194242
    Preview 91194244

    Coming Soon

    • Infinite talk

    • MMaudio

    • Lastframefirstframe

    • Ultimate Version

    Supported Workflows

    • Wan2.2 T2V (Text‑to‑Video) 14B 

    • Wan2.2 I2V (Image‑to‑Video) 14B

    • Wan2.2 TI2V‑5B (Hybrid Text/Image‑to‑Video) (old version)


    Model Downloads + VAE

    *these are conversions from the original models to run on less VRAM.

    All these GGUF conversions are done at the moment by:

    https://huggingface.co/city96

    https://huggingface.co/bullerwins

    https://huggingface.co/QuantStack

    *If you cant find the model you are looking for check out there profiles!


    Additional Required Files

    https://huggingface.co/lightx2v

    Description

    FAQ

    Comments (5)

    kreslFeb 2, 2026
    CivitAI

    getting error "mat1 and mat2 shapes cannot be multiplied (77x768 and 4096x3072)"

    The_frizzy1
    Author
    Feb 4, 2026

    thats ususally when youre using a incompatible vae with a incompatible clip or model its a mis match error you should be able to see the node that is the problem and then look what part of the model is mismatch and find another model to download.

    isabellasmithx0x0777Mar 14, 2026
    CivitAI

    The frames get tinted. My guess is it has to do with the vae tiled decode. I've tried changing the values. In fact i even replaced tiled vae decode with the standard one. The tinting persists

    The_frizzy1
    Author
    Apr 9, 2026

    Yeah there were defenetly some issues with it I have done a new update with way better Ksampler settings planning audio very soon aswell sort of like LTX2.3 .

    The_frizzy1
    Author
    Apr 9, 2026
    CivitAI

    Quick Start Guide

    -Install the missing nodes via manager

    -Download missing models (Upscaler)

    -Download Custom models (see List)

    -Put the models in the right folders

    (If you dont have make)

    --GGUF - models/diffusion_models

    --Vae - models/vae

    --clip vision - models/clip_vision

    --Text Encoder - models/text_encoder

    --Lora - models/lora

    -Set the resoloution

    -Set the Amount of frames

    - Optional upload Image

    - Run Workflow

    Workflows
    Wan Video 14B i2v 720p

    Details

    Downloads
    2,817
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/28/2025
    Updated
    8/3/2026
    Deleted
    -

    Files