CivArchive
    Animation original style noobv prediction AgelessBeta(动画原画风格_noob_v预测_ageless_测试版) - v1.0
    Preview 42707104Preview 42707126Preview 42707217Preview 42707266

    Model name: Animation original style noobv prediction AgelessBeta

    Model features:

    • Style: This model focuses on generating images that have the style of an animated original painting, emphasizing the simplicity of the lines.

    Background: Models are good at generating simple backgrounds, such as solid colored backgrounds (especially white backgrounds), to make characters stand out.

    • Applicable scene: It is suitable for scenes that need to quickly generate animation original drawing style images, such as concept design, character design, illustration creation, etc.

    Training data:

    • Data volume: The model uses tens of thousands of images for regularization training, and thousands of additional high-quality animation original style images for targeted training.

    Training parameters:

    • Training rounds (EPOchs): 40

    • Other parameters: [lokr, 5e-4]

    Suggestions for use:

    • Prompt words: It is recommended to use the following prompt words to guide the model to produce a more expected image: Style: color trace background: simple background, white background

    • Other suggestions: The noob common prompt word format is recommended.

    • Recommended usage weight: 0.5-0.9

    Model use requires v prediction, if not recommended watch video: noobai V Prediction tutorial (sdwebui) (comfyui) : [A video to teach you to use V prediction model,NoobAI Xiaobai tutorial - Bilibili] https://b23.tv/dqvsdr3

    Model version: V1.0

    Update log:

    •V1.0 (2024/11/26): Model release

    Disclaimer:

    • This model is for study and research purposes only and should not be used for any commercial purposes.

    Contact information:

    •[qq: 1002026893]

    • I leave a trace here


    模型名称: 动画原画风格_noob_v预测_ageless_测试版

    模型特点:

    •风格: 该模型专注于生成具有动画原画风格的图像,强调线条简洁。

    •背景: 模型擅长生成简单背景,例如纯色背景(尤其是白色背景),使角色更加突出。

    •适用场景: 适用于需要快速生成动画原画风格图像的场景,例如概念设计、角色设计、插画创作等。

    训练数据:

    •数据量: 模型使用了数万张图像进行正则化训练,并额外使用了数千张高质量的动画原画风格图像进行针对性训练。

    训练参数:

    •训练轮数 (Epoch): 40

    •其他参数: [lokr,5e-4]

    使用建议:

    •提示词: 建议使用以下提示词来引导模型生成更符合预期的图像:风格: color trace (彩色描线)背景: simple background (简单背景), white background (白色背景)

    •其他建议: 推荐使用noob常用提示词格式。

    •建议使用权重:0.5-0.9

    模型使用需要v预测,如不会使用建议观看视频:noobai的V预测教程(sdwebui)(comfyui): 【一个视频教会你用V预测模型,NoobAI小白向教程-哔哩哔哩】https://b23.tv/dqvsdr3

    模型版本: V1.0

    更新日志:

    •V1.0 (2024/11/26): 模型发布

    免责声明:

    •该模型仅供学习和研究使用,请勿用于任何商业用途。

    联系方式:

    •[qq:1002026893]

    •我在此处留下一个痕迹

    Description

    This version has some sampling problems in the training, when the vertical picture does not match, it is recommended to reduce the weight of the model

    这个版本在训练中出现了一些采样上的问题,当竖图出现画面不协调建议降低模型权重使用

    LoCon
    SDXL 1.0

    Details

    Downloads
    363
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/29/2024
    Updated
    9/30/2025
    Deleted
    -
    Trigger Words:
    color trace
    simple background
    white background

    Files

    原画风格动画.safetensors

    Mirrors