CivArchive
    anima [int8] - qwen3 0.6b
    NSFW
    Preview 136070839

    Int8convrot, converted from https://huggingface.co/circlestone-labs/Anima/blob/main/split_files/diffusion_models/anima-base-v1.0.safetensors

    Int8convrot qwen3 0.6b, mainly for fun. Because why not.

    qwen3 0.6b (b): 200MiB smaller than previous version. Embeddings also quantized to int8.

    Description

    FAQ

    Comments (6)

    rift1000Jul 7, 2026
    CivitAI

    is this convrot?

    reakaakasky
    Author
    Jul 7, 2026· 2 reactions

    of course

    KazuhiraJul 8, 2026
    CivitAI

    I just use this in my workflow but this actually slower than the original model. How is that?

    alter2611Jul 8, 2026· 1 reaction

    In my case the culprit was old CUDA (12.x). After upgrading to pytorch /w CUDA 13+ it works faster.

    KazuhiraJul 8, 2026· 2 reactions

    @alter2611 Thanks. CUDA 13 solves the problem.

    aiRabbit0Jul 9, 2026
    CivitAI

    Can it be used with the base model, or only in int8?

    Checkpoint
    Other
    by reakaakasky
    Download (Beta) View on CivitAI
    base model

    Details

    Downloads
    306
    Platform
    CivitAI
    Platform Status
    Available
    Created
    7/8/2026
    Updated
    8/25/2026
    Deleted
    -

    Files

    animaInt8_qwen306b.safetensors

    Size:
    718.30 MB
    SHA256:
    67b1d933e28e76c1c86a9627fbf96116893bf56692310f2cc416a2a8d0a5e504

    Mirrors

    CivitAI (1 mirrors)