SligarTheTiger Style Lora (Illustrious / Pony)

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。

如果您使用Pony版本,建议优先使用更新且稳定的2.0版本。

LoRA,由Civitai基于毛茸茸艺术家SligarTheTiger训练

也被称为Sligar / Dubmare

请通过以下链接支持该艺术家:

https://www.patreon.com/SligarTheTiger

https://www.furaffinity.net/user/sligarthetiger/

https://sligarthetiger.newgrounds.com/

https://x.com/dubmare

https://www.pixiv.net/en/users/33557664

https://inkbunny.net/SligarTheTiger

https://www.redbubble.com/people/dubmare/shop?asc=u

  1. 始终使用标签 by sligarthetiger

  2. 此LoRA基于Pony V6版本 / Illustrious

  3. 尽管使用了Pony V6基础模型,但在某些其他模型中生成图像时可能存在不一致,尤其是在眼睛和手部。若在Civitai上使用,请添加一些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。

  4. 别忘了使用与Pony模型相关的标签

score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up,
  1. 在Civitai中LoRA权重调整由您自行决定

  2. 如果您希望生成穿着高跟鞋的角色图像,请在正向提示词中添加标签:high heels, platform heels, platform footwear

  3. 如果您希望生成不穿高跟鞋的角色图像,请在负向提示词中添加标签:high heels, platform heels, platform footwear;若想生成赤脚图像,请在正向提示词中添加标签:feet

  4. 根据艺术家的风格,角色身体大多具有光泽效果。如需去除光泽,建议在负向提示词中添加标签:shiny body, glistening body, shiny fur, glistening fur,以强制生成无光泽角色。但我不保证100%在所有图像中均有效,请根据您的需求灵活使用。

  5. 请先检查与您选择的图像风格兼容的模型,某些模型可能影响LoRA在最终图像生成中的效果,例如:毛茸茸、人类、写实图像等。因此,同一模型未必适用于所有类型的图像。

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

Illustrious

<lora:SligarTheTiger_Style_Lora_Illustrious:0.8>

Pony (v1.0)

<lora:SligarTheTiger_Style_Lora:1>

Pony (v2.0)

<lora:SligarTheTiger_Style_Lora_v2:0.8>

注意:LoRA名称末尾的数字(如0.8)代表LoRA权重。在图像生成器中,我们通常使用滑块调节权重,但若想“强制”使用特定权重,请在正向提示词中使用与生成器滑块相同的数值,例如:

Illustrious

Civitai滑块 = 0.8

正向提示词: <lora:SligarTheTiger_Style_Lora_Illustrious:0.8>

Pony (v1.0)

Civitai滑块 = 1

正向提示词: <lora:SligarTheTiger_Style_Lora:1>

Pony (v2.0)

Civitai滑块 = 0.8

正向提示词: <lora:SligarTheTiger_Style_Lora_v2:0.8>

为获得2.0版本(Pony)的最佳效果,由于艺术家的独特风格及训练时AI对人脸检测的困难,我建议使用ADetailer(TensorArt)/ Face Fix(Civitai)。若您使用本地图像生成器(ComfyUI / A1111 / Forge),我也推荐以下设置:

Face Fix(Civitai)

- 工作流:Image-to-Image Face Fix
- 去噪强度:0.4
- VAE:SDXL VAE

超分辨率(ComfyUI / A1111 / Forge)

- 超分模型:4x_foolhardy_Remacri / R-ESRGAN 4x+
- 高分辨率步数:20
- 去噪强度:0.5

ADetailer(TensorArt)

- 检测模型置信度阈值:0.4
- 修补蒙版模糊:8
- 去噪强度:0.4
- 仅修补蒙版区域:勾选(true)

  1. 我添加了与艺术家最著名角色相关的标签。如需生成接近特定角色的图像,请务必输入角色名称。以下为每个角色的正向提示词示例

Dubmare

by sligarthetiger, dubmare, wolf, canid, canine, canis, furry, furry female, anthro, black body, black fur, big butt, huge butt, thick thighs, wide hips, big breasts, high heels, platform heels, platform footwear, striped body, striped fur, stripes, voluptuous, black skin, tail, nipples, solo, pussy, dark nipples, abs, navel, large breasts, muscular, muscular female, muscular anthro

SCP-1471

by sligarthetiger, scp-1471, scp-1471-a, scp foundation, canid, canine, malo, furry, anthro, furry female, skull, skull head, tail, animal ears, nipples, black hair, black fur, thick thighs, teeth, large breasts, body fur, claws

Sligar

by sligarthetiger, sligar, furry, anthro, furry male, canid, canine, canis, coyote, fox, hybrid, black body, black fur, black eyes, black nose, grey hair, white body, white fur

Guetto Delphox

by sligarthetiger, ghetto delphox, canid, delphox, generation 6, pokemon (species), furry, anthro, furry female, black fur, black body, black nipples, multicolored body, multicolored fur, narrowed eyes, white body, white fur, yellow body, yellow fur, red body, red fur, tail, thick thighs, pokemon (creature), half-closed eyes, plump labia, snout, wide hips, pokemorph, red nose, tuft, voluptuous, black eyes, huge breasts, curvy figure

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),但我在训练中使LoRA保持灵活可混合,以便与其他LoRA搭配使用。我建议您使用针对特定角色的LoRA,以获得更满意的结果。

以下为初始配置,但您可自由调整:

ILLUSTRIOUS

模型

WAI-NSFW-illustrious-SDXL - v10.0

LoRA

SligarTheTiger Style Lora v1.0 (illustrious) - 0.5 ~ 1.0 (推荐 0.8)

高级设置

CFG Scale - 3.0 ~ 9.0 (推荐 7)

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

步数 - 20 ~ 60 (推荐 30)

Clip Skip - 1/2

VAE - SDXL VAE

正向提示词

masterpiece, best quality, amazing quality, very aesthetic, high resolution, ultra-detailed, absurdres, newest, scenery, depth of field, scenery, volumetric lighting, by sligarthetiger, fox, canine, canid, 1girl, furry female, breasts, furry, solo, car, thick thighs, motor vehicle, ground vehicle, pussy, sitting, collar, nude, thighs, tail, claws, animal ears, navel, snout, nipples, huge breasts, full body, large breasts, uncensored, on vehicle, leash, spiked collar, jewelry, sharp fingernails, large areolae

负向提示词

bad quality, worst quality, worst detail, sketch, censor

PONY (v1.0)

模型

Pony Diffusion V6 XL - V6(建议从此开始)

LoRA

SligarTheTiger Style Lora - v1.0 (Pony) - 0.5 ~ 1.0 (推荐 1)

高级设置

CFG Scale - 3.0 ~ 9.0 (推荐 5)

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

步数 - 20 ~ 50 (推荐 30)

Clip Skip - 1/2

VAE - 无

正向提示词

score_9, score_8_up, score_7_up, by sligarthetiger, curvy figure, undertale, toriel, ass, solo, on side, furry, sunglasses, lying, furry female, black lips, ass focus, thighs, couch, thick thighs, 1girl, fishnet top, fishnets, tail, looking at viewer, bottomless, on couch, smile, huge ass, legs together, from behind, full body, looking back, barefoot, feet, shiny body

负向提示词

score_6, score_5_up, score_4_up, zPDXL-neg, 3D, 2D, human, high heels, platform heels, platform footwear

PONY (v2.0)

模型

Ponysauce XL - v1.0

LoRA

SligarTheTiger Style Lora - v2.0 (Pony) - 0.5 ~ 1.0 (推荐 0.8)

高级设置

CFG Scale - 3.0 ~ 9.0 (推荐 7)

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

步数 - 20 ~ 50 (推荐 30)

Clip Skip - 1/2

VAE - SDXL VAE

正向提示词

score_9, score_8_up, score_7_up, by sligarthetiger, cattle, bovid, bovine, ass, solo, furry, tail, thick thighs, tail, red fur, on side, room, window, animal ears, ass focus, animal ears, horns, earrings, lying, furry female, indoors, female focus, 1girl, plant, looking at viewer, sofa

负向提示词

zPDXL3, 3D, 2D, high heels, platform heels, platform footwear

由于此模型基于Pony Diffusion,负向提示词中无需过多设定,因为Pony模型本身非常灵活,但我始终保留 zPDXL-neg 标签(/model/332646/pony-pdxl-negative-embeddings)。

通用建议:

  • 如可能,在使用此LoRA时请注明作者SligarTheTiger的署名。

  • 请勿出售或商业化此LoRA;如有可能,请通过您的Patreon支持艺术家。

  • 请勿删除由LoRA生成图像的元数据。

此模型生成的图像

未找到图像。