HuslyoRealismXL 全能摄影大模型作为基于 SDXL 架构深度训练的模型，凭借精准的光影算法与细腻的细节捕捉能力，在多元场景中展现卓越实力。无论是壮阔风景的层次晕染、人像肌肤的质感呈现，还是动物神态的灵动还原，都散发自然真实的视觉张力。尤其在微距摄影领域，对花蕊纹路、昆虫羽翼等微观细节的刻画堪称惊艳，让每帧画面都兼具摄影质感与艺术温度。
经过大量的实验、合并和测试，我发布了我的第一个完整自定义大模型：
HuslyoRealism XL： SDXL 的摄影第一现实主义大模型，针对自然光、柔软皮肤和 NF/SW现实主义进行了调整。
这是一个以摄影为先的模型，能够处理 SFW 和 NF/SW 内容，尤其注重 NF/SW 人群。在开发过程中，我创建了 12 个不同版本的 HuslyoRealism，每个版本生成超过 2800 张图片，以测试稳定性和一致性。此版本是迄今为止最平衡的版本，尽管仍处于实验阶段。
推荐设置
- Sampler: DPM++ 2M SDE Karras or DPM++ SDE Karras
- Steps: 30-32
- Clip Skip: 2
- CFG Scale: 4-6
- Resolution: 768×1152, 648×1152, 864×1152, or any SDXL resolution you prefer.
- Eyes/Face: Use FaceDetailer or Adetailer with face_yolov8n.pt and eye detailer with PitEyeDetailer-v2-seg.pt
- Negative: cartoon, illustration, anime, painting, CGI, 3D render, low quality, watermark, logo, label, deformed eyes, deformed iris
在这个版本中，脸部和眼睛的识别可能会有点不准确，我会在 V2 中解决这个问题。话虽如此，我在图库中发布了数百个成功的示例，每个示例都包含完整的提示和设置。请仔细查看这些示例，看看哪些有效。
模型说明
- 非常适合 SFW 和 NF/SW 图像，而且还非常适合风景（风景或肖像）。
- 与我的Touch of Realism LoRA和本系列中的其他 LoRA混合，这意味着您不需要加载那些额外的 LoRA（尽管您可以使用它们来放大效果）。
- 注重光滑的肤色、自然光线和摄影构图。
- 对 film grain, grainy texture, photo of…提示的反应非常好，并且在macro photography-style close-ups.方面表现出色。
- 与其他一些 SDXL 模型相比，其表面会稍微柔和一些，但真实感会增强。
- 我会继续开发新版本。如果这个版本不符合你的目标，我强烈建议你换用Analog Madness XL3（我最喜欢的通用型号）或Lustify（NF/SW 最佳性能）。
效果良好的提示建议和关键词
此模型对以下提示片段和样式反应尤为良好。将它们组合使用可获得最佳效果：
• foreground blur, soft foreground, realistic depth falloff, cinematic shallow DOF, lens softness, bokeh in the distance
• natural light from the side, sunlight through the window, catching the sunlight, sun-kissed skin, golden hour light, dusty sunlight
• soft focus edges, ISO noise effect, grainy texture, realistic grain, soft skin texture with visible grain
• retro selfie style, analog photo look, vintage photo vibe, nostalgic atmosphere, faded look, vintage lighting, retro color grading
• natural textures, clay and wood tones, sunlit minimalist room, photorealistic beige mood, earthy palette, neutral styling, beige interior, Scandinavian light, calm aesthetic
• 适用于 photo of [subject], macro shot of, close-up portrait of, 以及类似的基于摄影的措辞
请随意混合或组合它们。该模型旨在奖励摄影措辞和自然光描述。
作品参照
editorial shoot, strobe effect, 20-year-old Belgian woman, fair skin, chin-length blunt bob with soft waves hair, brown eyes, soft oval face, slightly flushed cheeks, mouth closed framed between two hanging tapestries, head tilted slightly, strands of hair caught in fabric texture, strong depth blur around foreground patterns <lora:Touch_of_Realism_SDXL_V2:0.6> touch-of-realismV2 <lora:Touch_of_Vintage_Glow:1> touch-of-vintage-glow
Negative prompt: cartoon, illustration, anime, painting, CGI, 3D render, low quality, watermark, logo, label, teeth, mouth open
Steps: 32, CFG scale: 5, Sampler: DPM++ 2M SDE
a photo of an alpine lake with reflections of snowy mountain peaks and clear blue skies <lora:Touch_of_Realism_SDXL_V2:0.6> touch-of-realismV2 <lora:Touch_of_Grain:0.6> touch-of-grain, grain, film grain
Negative prompt: cartoon, illustration, anime, painting, CGI, 3D render, low quality, watermark, logo, label
Steps: 32, CFG scale: 4, Sampler: DPM++ 2M SDE, Seed: 2975688507, Size: 768x1152, Model: HuslyoRealismXL_v1_Beta.fp16, Version: v1.10.1, Model hash: 9bd7483e0b, Schedule type: Karras, ADetailer model: face_yolov8n.pt, ADetailer version: 25.3.0, ADetailer CFG scale: 1.0, ADetailer CLIP skip: 1, ADetailer mask blur: 4, ADetailer model 2nd: mediapipe_face_mesh_eyes_only, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer mask blur 2nd: 4, ADetailer confidence 2nd: 0.3, ADetailer inpaint padding: 32, ADetailer mask only top k: 2, ADetailer dilate erode 2nd: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding 2nd: 32, ADetailer denoising strength 2nd: 0.4, ADetailer use separate CFG scale: True, ADetailer use separate CLIP skip: True, ADetailer inpaint only masked 2nd: True, ADetailer method to decide top k masks: Area, Clip skip: 2
realistic image of an fluffy Orange cat, sitting calmly at the front of the bed, it's front paws are neatly crossed and resting on top of the bed sheets, facing directly towards the camera, detailed fur catching in the backlit by morning light, grainy film texture, <lora:spo_sdxl_10ep_4k-data_lora_webui:0.65>
Negative prompt: cartoon, illustration, anime, painting, CGI, 3D render, low quality, watermark, logo, label, deformed eyes, deformed iris
Steps: 30, CFG scale: 4.6, Sampler: DPM++ 2M SDE, Seed: 3471751523, Size: 832x1216, Model: huslyorealismxl_v10, Version: f2.0.1v1.10.1-previous-669-gdfdcbab6, Model hash: 9bd7483e0b, Hires upscale: 2, Original Size: 832x1216, Schedule type: Automatic, Hires Module 1: Use same choices, Hires upscaler: 4x_UniversalUpscalerV2-Sharp_101000_G, ADetailer model: person_yolov8n-seg.pt, Hires CFG Scale: 1, ADetailer prompt: [object Object], ADetailer version: 25.3.0, Denoising strength: 0.35, ADetailer mask blur: 4, Hires schedule type: Exponential, ADetailer confidence: 0.3, ADetailer dilate erode: 4, ADetailer inpaint padding: 32, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, Clip skip: 2
