Wherewolf 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。

基于毛茸茸艺术家 Wherewolf 在 Civitai 上训练的 LoRA

请通过以下链接支持该艺术家:

https://www.pixiv.net/en/users/4255153

https://subscribestar.adult/wherewolf

https://wherewolf.fanbox.cc/

  1. 始终使用标签 by wherewolf

  2. 此 LoRA 基于 Pony V6 版本 / Illustrious

  3. 尽管使用了 Pony V6 基础模型,但在 Civitai 上使用其他模型时,图像中某些部分(尤其是 眼睛和手部)可能出现不一致。建议使用某些 Embedding / Face Fix 以获得更好效果。对于使用 ComfyUI 的用户,最佳效果可通过 Hirefix / DDetailer 改善角色面部图像实现。

  4. 不要忘记使用与 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 上调整 LoRA 权重由您自行决定。

  2. 如果您希望生成无任何类型审查的图像,请在 负面提示 中添加以下标签:censored, mosaic censoring, bar censor, heart censor, out-of-frame censoring, black censor,因为部分由该 LoRA 生成的图像原作者绘制时包含审查元素。

  3. 为在 Civitai 上更好地使用 LoRA,我建议使用 VAE (SDXL VAE) 和 Embedding (Pony PDXL 负面嵌入),以减少手部与眼睛细节的错误。

  4. 请先检查与您选择的图像兼容的模型,某些模型会影响 LoRA 在最终图像生成中的效果,尤其对于:毛茸茸、人类、写实图像 等类型。因此,同一模型并不总是适用于所有图像类型。

  5. 若要强制图像生成器使用您选择的 LoRA 权重,请在正向提示中使用以下标签:

Illustrious (v1.0)

<lora:Wherewolf_Style_Lora_Illustrious:0.8>

Pony (v1.0)

<lora:Wherewolf_Style_Lora:0.8>

注意:LoRA 名称末尾的数字(0.8)表示 LoRA 权重。在图像生成器中我们有滑块控制权重,但若您希望“强制”使用特定权重,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 = 0.8

正向提示: <lora:Wherewolf_Style_Lora_Illustrious:0.8>

Pony (v1.0)

Civitai 滑块 = 0.8

正向提示: <lora:Wherewolf_Style_Lora:0.8>

注意:该艺术家使用了多个知名毛茸茸角色(卡通、电影、漫画等),但在训练中我使 LoRA 保持灵活,可与其他 LoRA 混合使用。我建议您使用特定角色的 LoRA,以获得更满意的最终结果。

  1. 我添加了与艺术家最知名角色相关的标签。如果您希望生成接近某角色的图像,请不要忘记输入角色名称。以下为每个角色的 正向提示示例

Lucy

by wherewolf, lucy, furry, furry female, anthro, domestic dog, nordic sled dog, canid, canine, canis, husky, spitz, black fur, black body, black arms, black ears, black tail, blue eyes, eyebrows, gloves (marking), multicolored body, multicolored fur, markings, tuft, wide hips, thick thighs, inner ear fluff, body fur, two tone arms, two tone body, two tone fur, two tone legs, two tone tail, white arms, white belly, white body, white breasts, white countershading, white eyebrows, white fingers, white hair, white hands, white inner ear, white inner ear fluff, white legs, white tail, collar

Cooney

by wherewolf, cooney, furry, furry female, anthro, raccoon, procyonid, black body, black fur, brown body, brown fur, eyebrow through hair, eyebrows, green eyes, markings, multicolored body, multicolored fur, red hair, ring (marking), ringed tail, striped markings, striped tail, tail markings, wide hips, thick thighs, thighhighs, black thighhighs, collar

以下为初始配置,但其他配置方式由您自行决定:

ILLUSTRIOUS

模型

Hoseki - LustrousMix [IllustriousXL] - Illustrious v1

LoRA

Wherewolf Style Lora - v1.0 (illustrious) - 0.5 ~ 1.0 (建议使用 0.8)

高级设置

CFG Scale - 3.0 ~ 9.0 (建议使用 6)

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 50)

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 wherewolf, furry, furry female, anthro, 1girl, rabbit girl, rabbit tail, animal ears, rabbit ears, solo, tail, hair ornament, black hair, street, night, lamppost, bag, shorts, red eyes, carrot, blush, black shorts, glasses, carrot hair ornament, dolphin shorts, breasts, closed mouth, holding bag, border, holding, full body, hairclip, waving, ass, food-themed hair ornament, looking at viewer, short shorts, night, hair between eyes, standing, shopping bag, short hair, bangs, sandals, tank top, medium hair, medium breasts, looking back, looking to the side, white skin, black-framed eyewear, outdoors, colored skin, white fur

负向提示

tears, cry, bad quality, worst quality, worst detail, bad quality, worst quality, worst detail, sketch, censor, multiple tails, modern, recent, old, oldest, graphic, cartoon, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured, long body, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, very displeasing, (worst quality, bad quality:1.2), bad anatomy, sketch, jpeg artifacts, signature, watermark, username, simple background, conjoined, ai-generated, worst quality, low quality, off-topic, old, early, chromatic aberration, abstract, jpeg artifacts, scan artifacts, logo

PONY

模型

Nova Furry XL - v6.0

LoRA

Wherewolf Style Lora - v1.0 (Pony) - 0.5 ~ 1.0 (建议使用 0.8)

高级设置

CFG Scale - 3.0 ~ 9.0 (建议使用 6)

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

score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, source_furry, BREAK, furry, anthro, detailed face eyes and fur, by wherewolf, renamon, digimon, digimon (species), small breasts, 1girl, solo, tail, blush, animal ears, blue eyes, black slera, fox ears, fox tail, pussy, green background, gradient background, fox girl, navel, colored sclera, cleft of venus, body fur, sweat, nude, nipples, animal nose, hands on hips, white fur, two-tone fur, detached sleeves, cowboy shot, yellow fur, purple sleeves, :o, ?, question mark

负向提示

zPDXL3, 3D, 2D, censored, mosaic censoring, bar censor, heart censor, out-of-frame censoring, black censor, blurry, jpeg artifacts, username, watermark, signature, normal quality, worst quality, large head, low quality, text, error, missing fingers, extra digits, fewer digits, bad eye, cry, tears

由于此模型基于 Pony Diffusion,负向提示中无需过多设置,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:

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

  • 请勿出售或商业性使用此 LoRA;如有可能,请通过您的 Patreon 帮助原作者。

  • 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。