CivArchive
    WhiteSpace - WhiteSpace Prism
    NSFW
    Preview 150274
    Preview 150285
    Preview 150284
    Preview 150283
    Preview 150282
    Preview 150281
    Preview 150280
    Preview 150279
    Preview 150278
    Preview 150277
    Preview 150276
    Preview 150275

    WhiteSpace SD Model


    It's called Whitespace, but it's not really whitespace at all. I was trying to add a more colorful and mysterious feel to the Pastelmix line of models.

    Some prompts that work well with the models include the following.

    "prism, glowing, glitter, particles, bloom, extra lights, cinematic lights, glowing white particles, +a"


    There are three types of models (WhiteSpace Prism, Quasar, and Black), each with slightly different characteristics.

    First of all, the Prism model is the base model for the other models and has the most stylish and distinctive look. It has a jewel-like design and unusual light effects. Recommended setting #1 is the best fit.

    The Quasar model takes the Prism and mixes it with the Chilloutmix and Ultracolor models to create an unusual semi-photorealistic model. It's still an unusual color, but it's a bit more like the other semi-photorealistic models. This model can be a bit unstable. Recommended setting #2 works best.

    The Black model was created by mixing Prism with a little bit of Helltaker lora, Maplestory lora, and World Flipper lora to create a bold and strong style. Enjoy the cartoon-like feel with strong lines and contrast. Recommended setting #2 works best.

    A detailed comparison of each model is shown in the image below.
    (WSB = Black, WSQ = Quasar, WSP = Prism)



    Based on the stable diffusion WebUI, the recommended settings are:

    Recommended setting #1 (oldQPrompt):

    - Prompt: masterpiece, best quility, high_quality, high_resolution, masterpiece*portait,

    - Negative prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name, ugly, duplicate, morbid, mutilated, tranny, trans, transsexual, hermaphrodite, extra fingers, mutated hands, poorly drawn hands, poorly drawn face, mutation, deformed, ugly, blurry, bad anatomy, bad proportions, extra limbs, cloned face, disfigured. more than 2 nipples. ugly, extra limbs, bad anatomy, gross proportions, malformed limbs, missing arms, missing legs, extra arms, extra legs, mutated hands, fused fingers, too many fingers, long neck

    - Sampler: DPM++ 2M Karras or DPM++ SDE Karras

    - Upscaler (optional): 4x-UltraSharp

    - CFG Scale : 6~7

    -Sampling steps: 20

    Recommended setting #2 (newQPrompt):

    - Prompt: (finely detailed, distinct image, ultra resolution, extremely amazing detailed, awesome detailed), subsurface scattering, best ratio four finger and one thumb, masterpiece, best quility, high_quality,

    - Negative prompts: lowers, (worst quality, low quality:1.4), EasyNegative, text,error, jpeg artifacts, signature, watermark, username, artist name

    - Sampler: DPM++ 2M Karras or DPM++ SDE Karras (Black model shows good quality even with Euler a)

    - Upscaler (optional): 4x-UltraSharp

    - CFG Scale : 6~7

    -Sampling steps: 20~26

    These are the settings in my case and I don't see any problems not following them, but in my case the above settings have the highest quality.

    Description

    W H I T E S P A C E P R I S M

    V1 0.1.0 Changelogs

    - First Version!

    - The following models were used for the merge:

    pastelmix x 0.25 + anything-v4.5 x 0.5 + Counterfeit-V2.5 x 0.25 = WS1

    WS1 x 0.5 + qwerty_v2 x 0.5 = WS2

    WS2 x 0.25 + HyperBombMix-fp16-no-ema x 0.5 + schoolAnime_schoolAnime x 0.25 = WS3

    WS3 x 0.25 + colormixed x 0.5 + nullV22_nullV22 x 0.25 = WS4(=WhiteSpace Prism V1, WSP)

    - This version includes the following files:
    WS4.safetensors(also .ckpt) : Full checkpoint file of This model
    WS4_Pruned.safetensors(also .ckpt) : no-ema + fp16

    FAQ

    Comments (4)

    Bunger_BeesechurgerFeb 25, 2023· 1 reaction
    CivitAI

    This is a great model! Would you be able to train on top of it using some danbooru images tagged "spot_color"? Could add a different level of contrast and pop to colors on top of what's already there

    Kyanite0029
    Author
    Feb 25, 2023

    I've never done model training on my own computer, so I don't know if it's possible, but I'll give it a try when I get some time. Thanks for the great feedback!

    @KN029 Forgot to mention, but I think merging this with the nou artist lora would also make some real cool stuff

    Kyanite0029
    Author
    Feb 26, 2023

    @Bunger_Beesechurger I tried it and it's great!!! I got a deeper color contrast and some interesting shapes.

    Checkpoint
    SD 1.5

    Details

    Downloads
    1,538
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/25/2023
    Updated
    5/13/2026
    Deleted
    -
    Trigger Words:
    prism
    glow
    glowing
    glitter
    particles
    bloom
    extra lights
    cinematic lights
    colorful
    glowing white particles
    glowing eyes
    prism hair
    prism eyes

    Files

    whitespace_Prism.ckpt

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    whitespace_Prism.safetensors

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    whitespace_Prism.ckpt

    Mirrors

    HuggingFace (1 mirrors)
    CivitAI (1 mirrors)

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.