CivArchive
    Kantai Collection (Kancolle) Submarines | I-168 I-58 I-8 I-19 I-401 艦これ 潜水艦 イムヤ 伊168 ゴーヤ 伊58 はち 伊8 イク 伊19 しおい 伊401 - Goya-v1
    NSFW
    Preview 902301
    Preview 902302
    Preview 902304
    Preview 902305
    Preview 902376

    This is a set of LoRA for generating the submarine kanmusu from Kantai Collection (Kancolle). Currently, I-168, I-58, I-8, I-19 and I-401 are supported, but I hope to cover them all one day. For now, I'm making individual character models but I want to create combined models eventually as well.

    Next sub: I-26

    Recommended Weights

    Weights around 0.7-1.0 recommended, start at 1 and scale back if it's too strong.

    For best effect use with highres fix.

    Recommended Block Weights

    Use with LoRA Block Weight extension for Automatic1111 WebUI. It's a powerful tool that allows you to tune what parts of the LoRA apply.

    Mainly character features

    OUTD:1,0,0,0,0,0,0,0,1,1,1,1,0,0,0,0,0

    Character features + stylization

    OUTALL:1,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1

    Posing and character features (affects background as well)

    INS_MIDD:1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0

    Suggested prompts

    See "About this version" for each model

    Suggested models

    Trained on NAI (final-pruned.ckpt [89d59c3dde]) so it should be compatible with most SD1.5 anime models. However I've noticed some have a propensity to turn sukumizu into bikinis so you may need to adjust prompts for those.

    I mainly use Anything-V5-prt-RE, it's very versatile and easy to work with.

    TIs

    You don't really need any of these but they're helpful

    negative_hand-neg https://civarchive.com/models/56519?modelVersionId=60938

    bad-hands-5 https://huggingface.co/yesyeahvh/bad-hands-5/tree/main

    verybadimagenegative https://civarchive.com/models/11772?modelVersionId=15556

    easynegativev2 https://huggingface.co/gsdf/Counterfeit-V3.0/tree/main/embedding

    supercha-negv2-p1 supercha-negv2-p2 https://civarchive.com/models/56033/supercha-neg

    Description

    LoRA model for Goya, or I-58, from Kantai Collection.

    Recommended Weights: weights around 0.7-1.0 recommended, start at 1 and scale back if it's too strong.

    Suggested prompts

    Default outfit (serafuku over old sukumizu)

    goya, one-piece swimsuit, sailor collar, serafuku, (swimsuit under clothes), short hair, pink hair, ahoge, hair accessory, pink eyes, small breasts

    Things to try

    torpedo, torn swimsuit, school swimsuit, old school swimsuit, submerged, underwater, freediving

    Recommended Block Weights

    Use with LoRA Block Weight extension for Automatic1111 WebUI. It's a powerful tool that allows you to tune what parts of the LoRA apply.

    Mainly character features

    OUTD:1,0,0,0,0,0,0,0,1,1,1,1,0,0,0,0,0

    Character features + stylization

    OUTALL:1,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1

    Posing and character features (affects background as well)

    INS_MIDD:1,1,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0

    Suggested models

    Trained on NAI (final-pruned.ckpt [89d59c3dde]) so it should be compatible with most SD1.5 anime models. However I've noticed some have a propencity to turn sukumizu into bikinis so you may need to adjust prompts for those.

    I mainly use Anything-V5-prt-RE, it's very versatile and easy to work with.

    FAQ

    LORA
    SD 1.5

    Details

    Downloads
    808
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/25/2023
    Updated
    4/23/2026
    Deleted
    -
    Trigger Words:
    goya, one-piece swimsuit, sailor collar, serafuku, (swimsuit under clothes), short hair, pink hair, ahoge, hair accessory, pink eyes, small breasts

    Files

    goya_nai_1-60.safetensors

    Mirrors

    CivitAI (1 mirrors)