If you like my models and want to support my work, please subscribe to my Patreon! For only $3.99 a month, you'll get access to all my works and exclusive content. Your support means the world to me and helps me keep creating more high-quality models! ❤️

https://www.patreon.com/c/AIBeauty565

EasonIllustrious_Style_10 V1

Checkpoint type: Illustrious

training checkpoint: illustriousXL_v01.safetensors

example:

breasts,1boy,hetero,sex,blonde hair,long hair,blue eyes,ass,bare shoulders,open mouth,large breasts,torn clothes,elbow gloves,clothed female nude male,sex from behind,gloves,hair over one eye,arched back,nude,blush,arm held back,thighs,solo focus,sleeveless,arm grab,indoors,torn bodysuit,white gloves,ass ripple,skin tight,clothed sex,bodysuit,bouncing breasts,shirt,

Negative prompt:

ugly,sketch,monochrome,greyscale,car,logo,letter,signature,logo,text,patreon logo,letters,name tag,patreon username,

Steps: 30, Sampler: Euler a, Schedule type: Automatic, CFG scale: 7, Seed: 1648115519, Size: 1296x936, Model hash: afc8b022eb, Model: EasonIllustrious_Style_10.fp16, Denoising strength: 0, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer prompt: ilvendetta, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 25.3.0, Hires upscale: 1.6, Hires steps: 20, Hires upscaler: 4x-UltraSharp, Version: v1.10.1

现在应该玩IL模型的人不多了，这个是个人训练的一个不错的风格的6万步的LORA进行融合。准备自己在X上出一些patreon的作品。

There probably aren't many people working with IL models these days. This is a well-trained 60,000-step LoRa model with a good style, which I've merged together. I'm planning to release some Patreon works on X.

-

融合了我喜欢两个model,其中一个眼睛描绘的不错，尤其是下蹲的动作，但是面部和身体显得不太立体。而第二个模型虽然面部立体，但是生成图片又显得过于真实。融合了这个模型我试了一下，感觉手部描绘要比原先的模型好一些.(astranime_V6+fantexi_V08). 有些重要参数在英文版说明里，黄色字体你需要特别注意，你可以顺便来我推特: https://twitter.com/EasonnLi

本模型融合2D 2.5D模型精粹，使之有趋向2.5D的二维漫画，对下蹲（squatting）和上半身(upper body)以及丝袜(black stocking)小美女等贴身衣物给力，尤其是眼睛的刻画，一定要开启after detail插件和Hires upscaler: 4x-UltraSharp插件

I merged two models I favorited. One of them does a good job of depicting eyes, especially in squatting poses, but the facial features and body don't appear very three-dimensional. While the second model excels at creating three-dimensional facial features, the generated images seem overly realistic. After merged these models, I tried it out and found that the depiction of hands is better than in the original model. (astranime_V6+fantexi_V08). welcome to my twitter: https://twitter.com/EasonnLi

This model integrates the essence of 2D and 2.5D models, giving it a tendency towards 2.5D two-dimensional comics. It excels particularly in depicting squatting, upper body, and close-fitting garments such as black stockings on petite girls. The portrayal of eyes is especially noteworthy. It is imperative to activate the "after detail" plugin and the "Hires upscaler: 4x-UltraSharp" plugin for optimal results.

prompt: standing, legs, sexy pose, upper body, (masterpiece, best quality:1.2), solo, 1girl, shimohira reika, expressionless, looking at viewer, black bodysuit, skin tight, underboob, elbow gloves, black gloves, more_details, cyberpunk city, black eyes, rock, black shoes
Negative prompt: verybadimagenegative_v1.0-6400,EasyNegative,(bad hands, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped),(worst quality:1.4),(low quality:1.4),(normal quality:1.4),jpeg artifacts,signature,watermark,username,(blurry:1.4), abs
Steps: 25,

Sampler: Euler a,

CFG scale: 7,

Seed: 1168988765,

Size: 512x768,

Model hash: 189479ae24,

VAE hash: c6a580b13a, VAE: vae-ft-mse-840000-ema-pruned.ckpt,

Denoising strength: 0.4, Clip skip: 2, ENSD: 31337,

ADetailer model: face_yolov8n.pt, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 24.3.0,

Hires upscale: 2.1,

Hires upscaler: 4x-UltraSharp, AddNet Enabled: True, AddNet Module 1: LoRA, AddNet Model 1: gantz_shimohira-10(d8de85dcbff9),

AddNet Weight A 1: 0.85,

TI hashes: "easynegative: c74b4e810b03", Version: v1.8.0

私は良いと思われる2つのモデルを組み合わせました。1つは特に目の描写が優れており、特にしゃがんでいるポーズが良いですが、顔や体が立体的ではなく見えます。2つ目のモデルは、立体的な顔の特徴を作成するのに優れていますが、生成される画像が過剰にリアルに見えます。これら2つのモデルを組み合わせた後、試してみたところ、手の描写が元のモデルよりも良いと感じました。my twitter: https://twitter.com/EasonnLi

このモデルは、2Dと2.5Dのモデルの精華を融合し、2.5Dの二次元漫画に近づけています。特に、しゃがんだ姿勢や上半身、そしてブラックストッキングなどの密着した衣類を着用した小柄な女性の表現に優れています。特に目の表現が注目に値します。最適な結果を得るためには、「after detail」プラグインと「Hires upscaler: 4x-UltraSharp」プラグインを活性化することが不可欠です。