Chatgpt image 2 在 @openart_ai 上

GPT-Image-2 · @doctorwasif · Fri Jul 24

在画廊中打开

分享链接 · https://openimages.ajiang.me/p/2080520917831307671/

Chatgpt image 2 在 @openart_ai 上 提示词 { "identity_ref": { "source": "uploaded photo", "lock…
Chatgpt image 2 在 @openart_ai 上 提示词 { "identity_ref": { "source": "uploaded photo", "lock…
Chatgpt image 2 在 @openart_ai 上 提示词 { "identity_ref": { "source": "uploaded photo", "lock…
Chatgpt image 2 在 @openart_ai 上 提示词 { "identity_ref": { "source": "uploaded photo", "lock…
composition · full-bodyhuman_attributes · long-hairlighting · soft-lightmood · playfulscene · studiosubject_type · character

提示词(原文,逐字保留)

Chatgpt image 2 on @openart_ai prompt { "identity_ref": { "source": "uploaded photo", "lock": ["face", "eyes/nose/mouth", "skin tone", "hair (length/texture/color)", "accessories", "exact outfit"], "rule": "no invented details" }, "composition": {"format": "vertical 4:5/9:16", "figures": ["real subject", "ink-twin doodle"], "layout": "side by side, poses fully visible"}, "real_subject": {"style": "photoreal, full-body", "outfit": "match reference exactly", "expression": "playful, embarrassed, caught-out charm", "pose": "natural, toned-down version"}, "ink_twin": { "style": "rough black sketch/charcoal doodle on wall", "recognizability": "silhouette only (hair/outfit/pose shape)", "avoid": ["realistic 2nd person", "flat shadow", "color cartoon"], "energy": "amplified, chaotic vs real subject", "extras": "optional sparse comic marks (motion lines/star/heart)" }, "pose_rule": { "instruction": "new playful/mischievous shared pose each time", "avoid": ["pointing", "finger-guns", "mirrored/repeated poses"], "relationship": "subject = subtle version; twin = amplified cartoon version, same beat" }, "setting": {"background": "clean minimal wall (white/cream/gray)", "lighting": "soft, even, daylight-style, face visible"}, "mood": "playful, stylish, chaotic-fun, shareable", "negative_prompt": [ "invented appearance", "outfit changes", "twin as real 2nd person", "flat/creepy shadow", "color cartoon twin", "weak resemblance", "clutter", "repeated/stiff/mirrored poses", "text", "watermark", "extra limbs", "bad hands" ] }

中文译文

Chatgpt image 2 在 @openart_ai 上 提示词 { "identity_ref": { "source": "uploaded photo", "lock": ["face", "eyes/nose/mouth", "skin tone", "hair (length/texture/color)", "accessories", "exact outfit"], "rule": "no invented details" }, "composition": {"format": "vertical 4:5/9:16", "figures": ["real subject", "ink-twin doodle"], "layout": "side by side, poses fully visible"}, "real_subject": {"style": "photoreal, full-body", "outfit": "match reference exactly", "expression": "playful, embarrassed, caught-out charm", "pose": "natural, toned-down version"}, "ink_twin": { "style": "rough black sketch/charcoal doodle on wall", "recognizability": "silhouette only (hair/outfit/pose shape)", "avoid": ["realistic 2nd person", "flat shadow", "color cartoon"], "energy": "amplified, chaotic vs real subject", "extras": "optional sparse comic marks (motion lines/star/heart)" }, "pose_rule": { "instruction": "new playful/mischievous shared pose each time", "avoid": ["pointing", "finger-guns", "mirrored/repeated poses"], "relationship": "subject = subtle version; twin = amplified cartoon version, same beat" }, "setting": {"background": "clean minimal wall (white/cream/gray)", "lighting": "soft, even, daylight-style, face visible"}, "mood": "playful, stylish, chaotic-fun, shareable", "negative_prompt": [ "invented appearance", "outfit changes", "twin as real 2nd person", "flat/creepy shadow", "color cartoon twin", "weak resemblance", "clutter", "repeated/stiff/mirrored poses", "text", "watermark", "extra limbs", "bad hands" ] }

来源与署名

原文由 @doctorwasif 发布在 X:查看原推文。 本页逐字保留原文并提供机器翻译的中文解读;版权归原作者所有。

在画廊中浏览

同工具的更多提示词