A detailed high-definition depiction of a cyborg elf wearing a braided flower headdress, standing confidently with hands on hips amidst flowers under a starry night sky in a retro anime style.
A gorgeous Latina woman wearing detailed chest plate armor and a hooded tunic, sitting against a stone wall in a medieval keep during golden hour with cinematic god rays and vibrant lighting.
Sweaty fitness girl with auburn hair and marine blue eyes wearing sportswear stretching on a mat in a gym with dumbbells in the background.
A smiling dark-skinned girl with wavy black hair wearing an elegant white and gold dress, sitting in a colorful tulip field during golden hour.

推奨ネガティブプロンプト

(worst quality:2),(low quality:2),(normal quality:2),lowres,bad anatomy,(text, font, logo, copyright, watermark)

(worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, nsfw, (monochrome), (grayscale), (text, font, logo, copyright, watermark:1), EasyNegativeV2

推奨パラメータ

samplers

DPM++ 2M Karras

steps

26 - 28

cfg

7.5

clip skip

2

resolution

512x720, 512x736, 528x760

vae

baked in VAE

other models

amour_bold_v3.5 (80fc83c345)

推奨ハイレゾパラメータ

upscaler

4x- UniScaleV2_moderate, 4x-UniScaleV2_Moderate

upscale

1.8 - 4

steps

15 - 27

denoising strength

0.46 - 0.5

ヒント

Use outline, monochrome, line art, grayscale, black and white to effectively trigger black and white images with the model.

Use with caution; prompt crafting is important as the model can sometimes be stubborn especially with black and white images.

Use hire.fix to improve image quality after generation.

If necessary, use inpaint after upscaling to polish images further.

Adetailer tool can be used to fix up faces and is highly recommended.

クリエイタースポンサー

Hello thank you for taking an interest in my model ♥

I decided to separate bold from regular amour - updated and will upload soon

my mango for making this model was because I mainly wanted to make something for graphic novels and comics that had a manhwa feel to it. V2 was never released but I am releasing it now : D (be sure to check out V3 if you want a more retro vibe)

V3.5

Yeaaaaaa I did not like how v3 did not look manhwa enough, I loved it but I had to amplify the husbandos…. so I decided to do a hotfix.

Pretty balanced hopefully 👍 (maybe not :p)

(also I like the detail on this one betta)

V3

I gave it a more dated look to it, it still has bold lines, a bit more simple. a bit different from v1 and v2 but amour is mixed in.

( this version is very different from V2 so be sure to check out V2 if you like more details, sexy men and less simple )

-Baked in VAE

-Pruned

V2

A bit more stable than v1 and composition is much better. (this version makes some shmexy ass males 🥵🥵🥵 )

-Baked in VAE

-Pruned

V1

Like the title says, it is more BOLD, bolder lines and can make manga/manhwa style stuff. (I made it to use in my own comics and stuff)

*be sure to use: outline, monochrome, line art, grayscale, black and white to trigger black and white (although sometimes some images will already come out black and white, sometimes it can be stubborn, prompt matters a lot with this model, it ain’t perfect, but when used correctly it can make bangers. (USE WITH CAUTION)

-Baked in VAE

-Pruned

Tools used:

EasynegativeV2 - https://civitai.com/models/100191?modelVersionId=107234

Easynegative - https://civitai.com/models/7808/easynegative

Deep Negative V1.x - https://civitai.com/models/4629?modelVersionId=5637 

[ these work too ]

Badhandv4 - https://civitai.com/models/16993?modelVersionId=20068

Bad-Hands-5 - https://civitai.com/models/116230/bad-hands-5

Fast Negative Embedding - https://civitai.com/models/71961?modelVersionId=94057

Adetailer - https://github.com/Bing-su/adetailer - use to fix up faces, this is a miracle tool

((please use hire.fix to improve quality))

Upscaler - https://mega.nz/folder/zZhA1KoD#ds2nmgDNV4hfFpNSfEqN6Q – 4x- UniScaleV2_moderate

If necessary be sure to use inpaint after upscaling to polish things up even more,

It is best to use these in negative or somthing similar: (worst quality:2), (low quality:2), (normal quality:2), lowres, bad anatomy, (text, font, logo, copyright, watermark) { negative embedding }

前の画像
style of Andreas Achenbach [SDXL] 102 - v1.0
次の画像
Tangbohu Detail Tweaker LoRA (唐伯虎细节调整LoRA) - v1.0

モデル詳細

モデルタイプ

Checkpoint

ベースモデル

SD 1.5

モデルバージョン

v3.5

モデルハッシュ

ba9c62d23f

作成者

ディスカッション

コメントを残すには log in してください。

「amour bold - v3.5」による画像

アニメ画像

女性画像

男性画像

マンファ画像

スタイル画像