CivArchive
    Rosalina | Character Lora 1077 - v1.0
    NSFW

    Lora trained on Rosalina from Mario.

    Crown is semi random can be called using crown probably even tho its pruned out.

    Trained on NAI, gen'd on AOM2 nsfw
    I'm tired so I didn't try other outfits but I'd assume they would work xD


    Disclaimer:

    The following is a parody and is not affiliated with, endorsed by, or in any way officially connected with Nintendo. This parody is intended to be a parody on Nintendo characters and is protected by fair use.

    The use of a Nintendo character in this parody is transformative and adds significant value. This parody is not a substitute for the original work and does not compete in any way with Nintendo's use of their character.

    Nintendo is the rightful owner of all intellectual property associated with their characters, including the character depicted in this parody. The use of this character in this parody is not intended to harm or damage Nintendo or their intellectual property in any way.

    Description

    FAQ

    Comments (6)

    kaisatomiMar 6, 2023· 1 reaction
    CivitAI

    Is next Daisy?

    Numeratic
    Author
    Mar 6, 2023

    princess king boo is next, then daisy.

    helloborororoMar 7, 2023
    CivitAI

    nice model!

    could you teach me how you learned? (learning rate,Number of images you used...etc)

    Numeratic
    Author
    Mar 8, 2023· 3 reactions

    1077 images pruned character traits.
    Basically dim 8, 1 repeat, 30 epochs, just pick whichever comes out best.

    Aim for 300-600+ images depends on what u want to make.

    If u cant get 300-600 just increase repeats then to get close to it.

    https://github.com/derrian-distro/LoRA_Easy_Training_Scripts/releases/tag/installers-v5
    Bunch of default settings:
    "caption_dropout_rate": null,

    "caption_dropout_every_n_epochs": null,

    "caption_tag_dropout_rate": null,

    "net_dim": 8,

    "alpha": 4.0,

    "scheduler": "cosine_with_restarts",

    "cosine_restarts": 1,

    "scheduler_power": 1,

    "warmup_lr_ratio": null,

    "learning_rate": 0.0001,

    "text_encoder_lr": null,

    "unet_lr": null,

    "num_workers": 1,

    "persistent_workers": true,

    "batch_size": 2,

    "num_epochs": 40,

    "save_at_n_epochs": 1,

    "shuffle_captions": false,

    "keep_tokens": null,

    "max_steps": null,

    "tag_occurrence_txt_file": false,

    "train_resolution": 512,

    "min_bucket_resolution": 320,

    "max_bucket_resolution": 960,

    "lora_model_for_resume": null,

    "save_state": false,

    "load_previous_save_state": null,

    "training_comment": null,

    "unet_only": false,

    "text_only": false,

    "reg_img_folder": null,

    "clip_skip": 2,

    "test_seed": 23,

    "prior_loss_weight": 1,

    "gradient_checkpointing": false,

    "gradient_acc_steps": null,

    "mixed_precision": "fp16",

    "save_precision": "fp16",

    "save_as": "safetensors",

    "caption_extension": ".txt",

    "max_clip_token_length": 150,

    "buckets": true,

    "xformers": true,

    "use_8bit_adam": true,

    "cache_latents": true,

    "color_aug": false,

    "flip_aug": false,

    "vae": null,

    "no_meta": false,

    "log_dir": null,

    "v2": false,

    "v_parameterization": false

    helloborororoMar 9, 2023

    @Numeratic Wow, thank you very much for the very thorough explanation great teacher!

    nicholerigsby884Aug 14, 2024

    I haven't gotten into training loras yet but I'm looking into doing just that here soon. So I'm saving this because it seems like legitimately good advice, so thanks in advance for helping me as well!

    LORA
    SD 1.5

    Details

    Downloads
    7,895
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/6/2023
    Updated
    5/13/2026
    Deleted
    -
    Trigger Words:
    <lora:ros1-000020:1>

    Available On (1 platform)

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