[LoRA] Cryomancer
模型简介：无论如何，寒冷从未困扰过我。
我本以为这会是一个简单的问题，但事实并非如此。除了图像数据的稀缺性之外，冰魔法和力量有许多不同的形式，例如碎片、尖刺、雪、冰雹等。因此，人工智能很难学习和区分它。最后，我把它分成两种形式。一种是松散的粉状雪魔法，另一种是更坚固的冰魔法。只需记住使用相应的标签即可。此外，该模型最适合半现实和 2D 风格，因为这是用于训练它的大部分数据。
重量：0.65 - 0.85
Trigger：
· ice-flurry, snow
· ice-shard, ice, crystal
附加标签：magic, cryokinesis, dynamic pose, fighting stance, casting spell
p/s：如果您要发表评论，请考虑分享您生成的图像。我喜欢看您的作品。
给我买杯咖啡。☕
https://www.buymeacoffee.com/samirfan90
~~~~~~~~~~~~~~~~~~~~~~~o(*￣︶￣*)o~~~~~~~~~~~~~~~~~~~~~~~
主图提示词和参数：
Prompt: <realistic, ((masterpiece)), ((best quality)), (detailed), cinematic, dynamic lighting, soft shadow, detailed background, professional photography, depth of field, intricate, detailed face, subsurface scattering, realistic hair, realistic eyes, muscular, manly, photo of a handsome man, ice-shard, snow, magic, dynamic pose, indoors, cape, braid hair,
Negative prompt: city, urban, 1girl, feminine face, low quality face, deformed face, deformed penis, 3d rendering, cgi, cartoon, lowres, worst quality, grayscale, monochrome, watermark, bad-hands-5, verybadimagenegative_v1.3, ng_deepnegative_v1_75t,
Steps: 35, Size: 512x768, Seed: 3160612616, Model: HomoVeritas_v1.2, Version: v1.5.1, Sampler: DPM++ 2M SDE Karras, CFG scale: 7, Clip skip: 2, Model hash: c315907e03, "bad-hands-5: aa7651be154c, Hires upscale: 2, AddNet Enabled: True, AddNet Model 1: Concept - Cryomancer(79e2142120b8), Hires upscaler: R-ESRGAN 4x+, ADetailer model: face_yolov8n.pt, AddNet Module 1: LoRA, Face restoration: CodeFormer, ADetailer version: 23.8.0, AddNet Weight A 1: 0.7, AddNet Weight B 1: 0.7, Denoising strength: 0.35, ADetailer mask blur: 4, ADetailer confidence: 0.3, ADetailer dilate/erode: 4, ng_deepnegative_v1_75t: 54e7e4826d53", ADetailer inpaint padding: 32, verybadimagenegative_v1.3: d70463f87042, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True
页:
[1]