Kim Dokja Omniscient Reader's Viewpoint

详情

模型描述

📘 LoRA指南:金独子——全知读者视角

🧍‍♂️ 角色类型:沉静、黑发的男性主角
📌 触发词1boy, kim_dokja, black_hair
🎯 适用场景:写实动漫风格渲染,戏剧性或内省氛围


🔧 推荐生成参数

参数
采样步数 25+(推荐:28–35)
CFG比例 5–7
采样器 DPM++ 2M Karras / DPM++ SDE Karras
分辨率 768x11521024x1536(竖版)
高分辨率修复 ✅(隐空间 → R-ESRGAN 4x+ Anime6B)
ADetailer ✅(面部 + 双手,以获得锐利细节)

正面质量提示词

masterpiece, best quality, amazing quality, very aesthetic, absurdres, score_9, score_8_up, extremely detailed, film grain, (cinematic:1.3), volumetric lighting, rim lighting, photoshoot, (EyesHD:1.2), (4k, 8k, Ultra HD), ultra-detailed, sharp focus, depth of field, ray tracing, best lighting, beautiful face, beautiful eyes, detailed face, detailed background, dramatic lighting, solo, standing, black school uniform, pensive expression, city ruins, moody, high contrast lighting, perfect composition, highly detailed eyes, flowing hair


负面提示词

EasyNegativecore_6, worst quality, low quality, lowres, bad anatomy, bad hands, extra digits, missing fingers, cropped, signature, watermark, deformed, mutated, disfigured, blurry, sketch, cartoon, manga, comic, text, jpeg artifacts, watermark, username, reference, extra limbs, unnatural face, bad perspective, bad proportions, long body, glitch, ugly, conjoined, AI-generated, bad lighting, out of frame, multiple views, amputee, 2koma, 4koma, logo, monster

此模型生成的图像

未找到图像。