青髪&紫髪カーディガンさん その1
2026/4/13使用したAI Stable Diffusion XL
年齢制限 全年齢
AIイラスト生成で最近好きになったモデルの Flowing Light Rendering v10 でIRIAM配信者さんのアバターを基に桜並木を通学する水色のカーディガン制服姿をAI画像生成しました
ポージングが固まりがちになってしまったので、ソコは今後の研究に生かそうと思いました
プロンプト
1girl, solo, looking at viewer, (dynamic pose:1.2), dynamic angle, (leaning forward:1.25), (waving, hand up:1.1), (hand behind back:1.1), cherry blossoms, cherry blossom tree, falling petals, outdoors, day, spring, sunlight, dappled sunlight, sunbeam, depth of field, blurry background, (rim lighting:1.15), backlighting, cinematic lighting, highly detailed background, BREAK ((split-color hair:1.3), ((purple hair:1.2), (dark blue hair:1.2)):1.3), (two-tone hair:1.2), (long hair:1.2), (long wavy hair:1.3), bangs, (white hair:-1), (pink hair:-1), (blue eyes:1.2), japanese_girl, (medium breasts:1.3), (curvy:1.3), cleavage, collarbone, (voluptuous:1.2), (shiny skin, gleaming skin:1.4), smile, open mouth, blush, BREAK (school uniform:1.3), (light blue open cardigan:1.2), collared shirt, (see-through white shirt:0.4), (light blue ribbon, striped ribbon, neck ribbon:1.1), (white skirt, plaid skirt, pleated skirt:1.2), (thighhighs, white thighhighs, zettai ryouiki:1.2), ((8k, masterpiece ,best quality ,ultra-detailed)), game cg, masterpiece, year 2025, newest, (perfect face:1.1025), (perfect hands:1.1025), (perfect legs:1.1025), (perfect arms:1.1025), (perfect fingers:1.1025), <lora:xl_more_art-full_v1:0.6> <lora:add-detail-xl:1> <lora:Fixhands_anime_bdsqlsz_V1:1>
ネガティブプロンプト
(worst quality, low quality, normal quality, bad quality:1.4), (lowres, jpeg artifacts, compression noise:1.2), (bad anatomy, deformed, disfigured, malformed limbs:1.4), (bad hands, poorly drawn hands, missing fingers, extra digits, fused fingers, too many fingers, webbed fingers:1.4), (poorly drawn face, ugly, asymmetrical face, cloned face:1.3), bad proportions, gross proportions, long neck, extra arms, extra legs, disjointed, text, signature, watermark, username, logo, 3d, 3d render, cgi, photorealistic, realistic, cartoon, sketch, monochrome, grayscale, multiple views, comic, oversaturated, blurry, out of focus, depth of field artifacts, aliasing, bad perspective, 3rd party watermark, artist name
Steps
27
Scale
7
Seed
1881821410
Sampler
Euler a
Noise
0.7
Strength
コメント

 
zanteiban(ざん) zanteiban(ざん) 2023年6月より活動開始 88投稿22フォロー27フォロワー 艦これ、ウマ娘、配信者さんのAI画像生成をしています AI生成環境はGoogle Colab、Stable Diffusion Web UI Forgeを使っています AI画像生成は始めて半年くらいです、皆さんと仲良く出来れば嬉しいです