習作：中ノ尾ガール
ソレイユ14世
失敗：LoRA学習、苦戦中！ ボスケテ！！
漫画からスタイルLoRAを作るコツを誰か教えてクレメンス！！（TT） バストショット中心に30枚ほどで学習してるんだけど、一発では上手く行かないもの？ ブラッシュアップするならどうやればイイの？ 漫画の画像だけでなく、こうやって生成した画像も使ったりする？
Parameter anzeigen
- Ursprüngliche Eingabeaufforderung
- (masterpiece:1.5), (best quality:1.5), (very aesthetic:1.2), (absurdres:1.2), (detailed background), newest, intricate details, ultra detailed, 8K, 1girl, solo,, nakanoo_kei (style),, beautiful girl, beautiful face, cheerful girl, 4-5 heads high, well-proportioned, curvaceous, (black hair:1.4), straight hair, (shiny hair:1.2), (glossy hair:1.05), (detailed hair:1.2), (thin hair:1.2), (dark brown eyes:1.2), (detailed eyes:1.4), catchlight, (detailed eyebrows:1.3), (smooth skin), (detailed skin), (skin texture), matte skin, smile, (blush:0.8), (looking at viewer:1.3), posing, the pose, dynamic pose, serafuku, detailed clothes, classroom, indoors, (cowboy shot:1.4), (from side:0.3),, <lora:triggerxxxholicstyleplease:1>
- Größe
- 768 x 1280
- Sampling-Schritte
- 28
- Sampling-Methode
- Euler a
- CFG scale
- 5.5
- Negativ
- (bad quality,worst quality,worst detail,bad, low quality, bad anatomy:1.2), (polydactyly, extra fingers, bad hands, bad hand, bad finger, bad fingers, missing finger, split finger, split fingers, extra digits, fused arms, fused hands:1.2), monochrome, greyscale, black and white, well,
Gesichtskorrektur
Hohe Auflösung
Prompt-Assistent
Verwendetes Modell und LoRA
1 private Modelle verwendet
