CyberRealistic XL v2.2真实摄影模型

使用介绍

v2.2模型得到了进一步的完善和重新定义。

虽然花了点时间，但 Cyber​​Realistic 的 SDXL 版本已经问世。
此模型的标准是使用简单和多功能性。因此，没有复杂的提示（如果您愿意，也可以）；只需释放您的想象力即可。尽情享受游戏乐趣！

请放心，Cyber​​Realistic 和 Cyber​​Realistic Classic（基于 1.5）都将继续像往常一样获得更新和支持。

个人设置

步骤：> 25
采样器：DPM++ 3M SDE / DPM++ 2M Karras / Euler A
VAE：已在
HiRes中烘焙：4x_NickelbackFS_72000_G/4xNMKD-Siax_200k

作品参照

Photo portrait of a 21 y.o. Dutch woman, defiant, pensive, freckles, long dark straight blonde hair, blue eyes
Negative prompt: (worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad art:1.4)
Steps: 30, CFG scale: 5, Sampler: DPM++ SDE, Seed: 3175431270, Size: 768x1024, Model: CyberRealisticXLPlay_V2.2-VAE, Version: v1.10.1, Model hash: 779ed5cd37, Schedule type: Karras

A analogue photo of a young man with messy hair and sunglasses, sunlit alleyway, Rome, Punk tee shirt, relaxed vibe, classic film grain, muted colors, nostalgic atmosphere, light leaks, color fading, vignetting, shot on kodak portra 400, expired film
Negative prompt: (worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad art:1.4)
Steps: 30, CFG scale: 5, Sampler: DPM++ SDE, Seed: 4294910326, Size: 768x1024, Model: CyberRealisticXLPlay_V2.2-VAE, Version: v1.10.1, Model hash: 779ed5cd37, Schedule type: Karras

gleaming red Ford Gran Torino, front view, Chicago avenue, bright headlights, foggy morning, 4K, highly detailed, (grainy 35mm photo)
Negative prompt: illustration, unrealistic, painting, drawing, sketch, cartoon, anime, manga, render, watermark, signature, label, (cg, 3d, worst quality, low quality)
Steps: 60, CFG scale: 4, Sampler: dpmpp_3m_sde_gpu, Seed: 860758098274473000, VAE: Default (model), Size: 832x1216, Model: cyberrealisticXL_v22, Version: Fooocus v2.5.3, Scheduler: karras, Sharpness: 2, Model hash: 779ed5cd37, Raw prompt: [object Object], Performance: Quality, ADM Guidance: [object Object], Lora weights: [object Object], Raw negative prompt: [object Object], Clip skip: 2