BSSEquinox Illus 半写实大模型

BSSEquinox Illus 半写实大模型，是一款依托 Illustrious 优质模型架构与训练体系深度研发的专业生成模型。它在 Illustrious 原有技术底座上，针对半写实风格创作需求进行了专项优化，不仅保留了基础模型在细节刻画、色彩还原上的优势，还通过海量半写实风格图像数据的二次训练，大幅提升了对人物肌理、场景质感、光影层次的精准把控能力。

Equinox是一款灵活的半写实 XL 模型，专为创意控制和适应性而设计。Equinox
以 Illustrious-XL 为基础，提供自然的渲染效果、富有表现力的解剖结构、干净的灯光和卓越的快速理解能力，不受特定视觉风格的束缚。

v2.0 中的新功能

输出质量

• 更清晰、更干净— 改进的细节渲染和整体清晰度
• 精致的解剖学——更好的比例和自然的形态
• 平衡照明——减少刺眼的对比度和伪影

提示与控制

• 更灵活的提示——更好地适应不同的概念和风格
• 改进的快速跟随功能——更紧密地响应您的指令
• 双模式现实主义：
  • 添加“realistic” →更现实的结果
  • 省略“realistic” →更柔和的半现实主义/动漫风格的图像

分辨率和性能

• 高分辨率提升— 高达1536px的出色效果

个人笔记

我对 Equinox 的发展感到非常自豪，并计划继续进一步发展它。

推荐设置

• 采样器： Euler A (Karras) / DPM++ 2M SDE (Karras)
• 步骤： 25–40
• CFG： 4-6
• 片段跳过： 2
• VAE：Baked-in
• 原始分辨率： 832×1216 / 896×1152 / 1216×832 / 1024×1536
• LoRA支持：是
• ControlNet 支持：是

提示指南

正面提示示例：

masterpiece, best quality, amazing quality, realistic, <-{optional}

否定提示示例：

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

亮点

• 半现实的视觉输出，具有清晰的渲染和微妙的细节
• 通过 Illustrious‑XL 主干网实现高级提示理解
• 适应不同的风格 — — 从电影风格到绘画风格
• 与LoRAs、ControlNet、Refiners等完全兼容

使用提示：

CLIP SKIP: 2

作品参照

masterpiece, best quality, amazing quality, BSS_RSGRL, 1girl, solo, looking at viewer, short hair, shirt, brown hair, hair ornament, bow, brown eyes, jacket, school uniform, white shirt, collared shirt, hairclip, medium hair, bowtie, blurry, black jacket, eyelashes, lips, blurry background, thick eyebrows, portrait, pink bow, hair behind ear, pink bowtie, photorealistic
Negative prompt: bad quality,worst quality,worst detail,sketch,censor,
Steps: 25, CFG scale: 5, Sampler: Euler a, Seed: 229845110713917, Model: BSS_Equinox_v2_00002__00001_, width: 1024, height: 1536

masterpiece, best quality, amazing quality, 1girl, solo, breasts, short hair, dress, bare shoulders, medium breasts, closed mouth, upper body, white hair, hairband, sleeveless, mole, blurry, black dress, lips, wet, depth of field, blurry background, turtleneck, phone, cellphone, black hairband, wet clothes, mole under mouth, facing viewer, smartphone, rain, water drop, blindfold, wet hair, covered eyes, black blindfold, yorha no. 2 type b
Negative prompt: bad quality,worst quality,worst detail,sketch,censor,
Steps: 25, CFG scale: 5, Sampler: Euler a, Seed: 229845110713917, Model: BSS_Equinox_v2_00002__00001_, width: 1024, height: 1536

masterpiece, best quality, amazing quality, 1girl, solo, breasts, looking at viewer, blush, smile, open mouth, short hair, brown hair, black hair, red eyes, dress, ribbon, cleavage, bare shoulders, medium breasts, collarbone, upper body, sidelocks, outdoors, sky, sleeveless, choker, day, fang, cloud, white dress, sleeveless dress, dutch angle, black choker, half-closed eyes, hair tubes, sunlight, cloudy sky, short hair with long locks, backlighting, sunset, tress ribbon, spaghetti strap, sundress
Negative prompt: bad quality, worst quality, worst detail, sketch, censor
Steps: 25, CFG scale: 5, Sampler: Euler a, Seed: 628159884703990