在剛開始AI繪圖時(約半年前)我曾有產過一張遠坂凜，這次想說來比較一下這段時間產圖的差距。
以下舊版:
| ((masterpiece)), (((best quality))), ((ultra-detailed)), ((illustration)), ((disheveled hair)), ((frills)), (1 girl), (solo), dynamic angle, beautiful detailed eyes, (((Rin Tohsaka))), (school uniform) Negative prompt: owres, 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,missing fingers,bad hands,missing arms, long neck, Humpbacked, pointed ears Steps: 30, Sampler: Euler a, CFG scale: 5.5, Seed: 2064431814, Size: 512x768, Model hash: 6569e224, Eta: 0.68, Clip skip: 2, ENSD: 31337 |
新版參數:
| best quality, ultra-detailed, dynamic angle, upper body, (from side:0.8), looking at viewer, 1 girl, (Rin Tohsaka:1.2), beautiful detailed blue eyes, blush, (disheveled hair:1.2), red sweater, white cross, pull hair, simple background, Negative prompt: (EasyNegative:0.7), text, signature, watermark, username, nsfw, lace, bare shoulders, Steps: 20, Sampler: DPM++ 2M Karras, CFG scale: 6, Seed: 1886310203, Size: 512x768, Model hash: 7f96a1a9ca, Model: Anythingv5V3_v5PrtRE, Denoising strength: 0.55, Clip skip: 2, Hires upscale: 2, Hires upscaler: R-ESRGAN 4x+ Anime6B, Version: v1.3.2 |
新舊版參數編寫方式差非常多。同樣在沒有修圖重繪、大量重roll、Lora和Controlnet的前提下，解析度、眼睛精緻度、光影效果都提升了，但身體姿勢和髮飾效果新版反而比較糟，看來還有許多待加強的部分。
我的其他作品請參考: Pixiv連結
達人
