模型简介: 添加更多细节 - 细节增强器/调整器 你喜欢我做的事情吗?考虑在Patreon上支持我🅿️ 或者随时请我喝杯咖啡☕ 我已经使用 CyberAIchemist 的(很棒的)Detail Tweaker 一段时间了,我真的很好奇。我真的只是想知道我是否能够做类似的事情。 我做这件事很开心,而且我认为结果非常棒。我肯定会把它融入到我今后的大部分工作中。 您应该根据自己的喜好,在 0.5 和 1 之间使用这个值。当然,您也可以使用低于 0.5 的值来获得更微妙的效果。 我还发现,有时负权重会产生一些有趣的结果。请参阅示例以了解我的意思。 如何在 auto1111 中使用 LoRA: 更新 webui(使用git pull 类似此处或重新下载) 将文件复制到stable-diffusion-webui/models/lora 像本视频中一样选择您的 LoRA 确保改变权重(默认情况下,它:1通常太高) ~~~~~~~~~~~~~~~~~~~~~~~o(* ̄︶ ̄*)o~~~~~~~~~~~~~~~~~~~~~~~ 主图提示词和参数: Prompt: best quality,masterpiece,highly detailed,ultra-detailed, <lora:neg4all_bdsqlsz_V3.5:-1> 1girl, braid, upper body, (Gauntlets),An armor made of metal gloves <lora:ActionFigure_last:1> Action Figure, , plastic collectable action figure, collectable toy action figure Negative prompt: lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, artist name,by bad-artist-anime by bad-artist,(monochrome, greyscale, BW:1.4), (7 fingers:1.9)( 6 fingers:1.9) (twisted fingers:1.9) (too long fingers:1.3) (3 arms:1.9) (3 legs:1.9) negative_hand-neg,,, Steps: 20, VAE: vae-ft-mse-840000-ema-pruned.ckpt, Size: 512x640, Seed: 722249461, Model: UnstableinkdreamV8.0-photo-V2, Version: v1.6.0, Sampler: DPM++ 2M Karras, AntiBurn: Count=3;, CFG mode: Half Cosine Up, VAE hash: c6a580b13a, CFG scale: 20, Clip skip: 2, Mimic mode: Half Cosine Up, Model hash: bd67d37dac, bad-artist: 2d356134903e, Hires steps: 12, Mimic scale: 7, Hires upscale: 2, Hires upscaler: 4x-AnimeSharp, Interpolate Phi: 1, "bad-artist-anime: 5f7bea88750c, ActionFigure_last: 2a724a6b1b3c", CFG scale minimum: 3.5, negative_hand-neg: 73b524a2da12", Denoising strength: 0.55, Scaling Startpoint: MEAN, Mimic scale minimum: 7, Token merging ratio: 0.4, Variability Measure: AD, Threshold percentile: 100, "neg4all_bdsqlsz_V3.5: b1b5db66e2f6, Token merging ratio hr: 0.4, Separate Feature Channels: True, Dynamic thresholding enabled: True |
1人表示很赞