CivArchive
    Artius 2-1 - V1.0 VAE
    Preview 378440Preview 378457Preview 378456Preview 378455Preview 378454Preview 378453Preview 378452Preview 378451Preview 378449Preview 378448Preview 378447Preview 378446Preview 378445Preview 378444Preview 378443Preview 378442Preview 378441

    Explore your imagination!
    Artuis is a versatile and flexible model, it allows you to generate crisp, high-resolution images:
    - Realistic portraits
    - N_S_F_W
    - Stylized Art
    - Logos
    - Game assets
    - Texture
    - People, animals, food photos and more
    - Product concept
    - Any genre: Realistic, Fantasy, Sci-Fi, Cyberpunk, Anime and more

    If you liked my model, please leave a review, it helps other users to find this model and see its features!

    This is SD 2.1 based model, also I have SD 1.5 based version: https://civarchive.com/models/47691/artiusv1-5

    The model works best with the following samplers:
    Euler a
    DPM2
    DPM++ 2M Karras

    Optimal settings:
    Clip skip 1
    Eta noise seed delta: 31337
    DPM++ 2M Karras
    25 Steps
    CFG Scale 5-6
    Resolution from 768 (768, 1024, 1280)

    The model is prompt sensitive, so it depends on your imagination.
    I've generated some demo images, you can study the prompt and make your own based on them.

    To make unique faces, mix faces that the model knows: [person1 : person2 : 0.5]

    Description

    Checkpoint
    SD 2.1 768

    Details

    Downloads
    5,545
    Platform
    CivitAI
    Platform Status
    Available
    Created
    4/1/2023
    Updated
    9/27/2025
    Deleted
    -

    Files

    artius21_v10VAE.safetensors

    Mirrors

    Huggingface (1 mirrors)
    CivitAI (1 mirrors)

    artius21_v10VAE.yaml

    Mirrors

    CivitAI (83 mirrors)