CivArchive
    Sam_MOSS - Turbo v1.0
    NSFW
    Preview 146051424
    Preview 146051427
    Preview 146051435
    Preview 146051467
    Preview 146051468
    Preview 146051472
    Preview 146051476
    Preview 146051480
    Preview 146051485
    Preview 146051493
    Preview 146051494
    Preview 146051506
    Preview 146051513
    Preview 146051514
    Preview 146051521
    Preview 146051522
    Preview 146051525
    Preview 146051526
    Preview 146051529
    Preview 146051732

    An attempt at creating a "turbo" version; it works with 8 or more steps.

    The int8 version was taken as the basis, there will be no bf16 version!

    Description

    FAQ

    Comments (4)

    physisOct 18, 2026
    CivitAI

    Can you share your workflow? sampler / scheduler / etc?

    toya_san
    Author
    Oct 18, 2026

    Uploaded it; go ahead and download it, enjoy!

    would_rOct 18, 2026· 1 reaction

    @toya_san Why is that extension needed? It works as is with native nodes, how did you make the model? Did you include darask0/MOSS-Image_4-step_adapter_lora?
    e: It seems you put something else because darask0's adapter seems to be biased toward realism while your model isn't.

    toya_san
    Author
    Oct 18, 2026

    @would_r It’s a blend of the darask0 LoRA, lodestones, and another LoRA, all mixed in varying proportions. You’re right—that wasn’t the original intention, but something shifted after saving the model, causing it to behave the way it does now. I tested it out, and the results seemed interesting, so I decided to release it. If there are future versions, they’ll be even better.

    Checkpoint
    MOSS Image
    by toya_san
    Download (Beta) View on CivitAI
    base model

    Details

    Downloads
    245
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/18/2026
    Updated
    11/1/2026
    Deleted
    -

    Files

    flux2-vae.safetensors

    Size:
    320.64 MB
    SHA256:
    868fe7b343cc8f3a19dbcfcafbc3d5f888802be3f89bd81b65b3621a066ce8f3

    Mirrors

    HuggingFace (162 mirrors)
    CivitAI (14 mirrors)
    ModelScope (11 mirrors)
    ModelScope CN (4 mirrors)

    samMOSS_turboV10.safetensors

    Size:
    5.04 GB
    SHA256:
    83dc7273e12321602cec8cd07db3ef4733c9029234f2b96f2184666f25285c89

    Mirrors

    CivitAI (1 mirrors)

    samMOSS_turboV10.json

    Size:
    11.91 KB
    SHA256:
    72d24c714caf5e13c9bd7dad4eb673070218c689b8e076b386768a342065c7af

    Mirrors

    CivitAI (1 mirrors)