shoutingisfun Style Lora (Illustrious)

详情

模型描述

请先阅读“关于此版本”部分中的版本详情和建议。

LoRA,基于毛茸茸艺术家 shoutingisfun(又名 concretestaircase)在 Civitai 上训练

请使用以下链接支持该艺术家:

https://twitter.com/shoutingisfun

https://derpibooru.org/tags/artist-colon-shoutingisfun

https://bsky.app/profile/shoutingisfun.bsky.social

https://ko-fi.com/shoutingisfun

https://subchan.org/ab/res/2.html

https://shoutingisfun.neocities.org

https://mlpg.co/art/res/9696.html

https://mlpg.co/art/res/16668.html

https://comicfury.com/profile.php?username=Shoutingisfun

  1. 始终使用标签 by shoutingisfun

  2. 此 LoRA 基于 Illustrious 模型

  3. 尽管使用了 Illustrious 基础模型,但在 Civitai 上使用其他模型时,图像中可能出现不一致,尤其是在 眼睛和手部。如在 Civitai 上使用,请搭配一些 Embedding / Face Fix 以获得更佳效果。对于使用 ComfyUI 的用户,建议使用 Hirefix / DDetailer 来改善角色面部图像质量。

  4. 在 Civitai 上调整 LoRA 权重由您自行决定

  5. 为在 Civitai 上更好地使用 LoRA,我建议使用 VAE (SDXL VAE) 以减少 手部和眼睛 细节的错误。

  6. 如果您希望生成彩色图像,请将 monochrome, greyscale, sketch, comic, silent comic, blank speech bubble, speech bubble 标签加入 负面提示,因为此 LoRA 是基于黑白图像或单一调色板训练的。

  7. 请先检查与您选择图像兼容的模型,某些模型会影响 LoRA 在最终图像生成中的效果,尤其针对:毛茸茸、人类、写实图像等类型。因此,同一模型未必适用于所有图像类型。

  8. 若您希望强制图像生成器使用您选择的 LoRA 权重,请在正向提示中添加以下标签:

Illustrious (v1.0)

<lora:shoutingisfun_Style_Lora_Illustrious:1>

注意:LoRA 名称末尾的数字(1)表示 LoRA 权重。在图像生成器中我们有权重滑块,但若想“强制”使用,请在正向提示中使用与生成器滑块相同的数值,例如:

Illustrious (v1.0)

Civitai 滑块 = 1

正向提示: <lora:shoutingisfun_Style_Lora_Illustrious:1>
  1. 我添加了与艺术家最知名角色相关的标签。若希望生成接近该角色的图像,请勿忘记输入角色名称。以下是每个角色的正向提示示例

Queenie

by shoutingisfun, monochrome, greyscale, sketch, queenie (shoutingisfun), furry, furry female, anthro, animal ears, animal nose, snout, pokemon, generation 1 pokemon, nidoqueen, pokemon (species), tail, spiked tail, spikes, spikes (anatomy), claws, slightly chubby, voluptuous, 1girl, solo, looking at viewer, smile, black background, simple background, gradient background, dotted background, cowboy shot, white outline

Gena

by shoutingisfun, gena (shoutingisfun), furry, furry female, anthro, pokemon, gengar, generation 1 pokemon, pokemon (species), spirit, purple fur, body fur, red eyes, colored sclera, narrowed eyes, shortstack, slightly chubby, plump, wide hips, thick thighs, naughty face, grin, big breasts, 1girl, solo, dark nipples, belly, navel, plump labia, completely nude, pokemon (creature), looking at viewer, full body, (dark theme:1.4), night, forest, nature, trees, grass

Tank

by shoutingisfun, tank (shoutingisfun), furry, furry female, anthro, animal ears, animal nose, snout, pokemon, bovine, generation 2 pokemon, miltank, pokemon (species), animal print, bovine horns, cow horn, cow print, black fur, pink fur, body fur, two tone fur, purple eyes, black nose, overweight female, wide hips, teats, udders, 1girl, solo, standing, full body, completely nude, looking at viewer, smile, teeth, white background, simple background

注意:艺术家使用了多个广为人知的毛茸茸角色(卡通、电影、漫画等),但在训练中我保持了 LoRA 的灵活性和可混合性,以便与其他 LoRA 结合使用。我的建议是:使用特定角色的 LoRA,以获得更满意的结果。

以下为初始配置,但其他配置完全由您自行决定:

ILLUSTRIOUS

模型

World Furry Fusion - v1.0(标准版)

LoRA

shoutingisfun Style Lora v1.0 (illustrious) - 0.5 ~ 1.0 (建议使用 1)

高级设置

CFG Scale - 3.0 ~ 9.0 (建议使用 5)

采样器 - DDIM / Euler a / Euler / Heun / DPM2 / DPM++ 2M Karras / Uni PC (建议使用 Euler a)

步数 - 20 ~ 60 (建议使用 30)

Clip Skip - 1/2

VAE - SDXL VAE

正向提示

masterpiece, best quality, amazing quality, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery, depth of field, scenery, volumetric lighting, by shoutingisfun, furry, furry female, anthro, blaziken, generation 3 pokemon, pokemon, pokemon (species), looking back, looking at viewer, red fur, body fur, blonde hair, long hair, yellow sclera, blue eyes, 1girl, solo, huge breasts, smile, closed mouth, blush, blue shirt, short sleeves, crop top, underboob, bottomless, navel, stomach, tail, pokemon (creature), half-closed eyes, looking at viewer, white outline, white border, outside border, standing, hands behind back, yellow background, simple background, dotted background, gradient background

负向提示

bad quality, worst quality, worst detail, lowres, jpeg artifacts, signature, watermark, username, artist name, monochrome, greyscale, sketch, comic, silent comic, blank speech bubble, speech bubble

通用建议:

  • 如有可能,在使用此 LoRA 时请注明作者 shoutingisfun 的署名。

  • 请勿出售或商业使用此 LoRA。

  • 请勿移除由 LoRA 生成图像的元数据。

此模型生成的图像

未找到图像。