本LoRA主要针对《世界计划：缤纷舞台》的镜音连形象进行训练，还原了游戏卡面的绘图风格，对游戏中镜音连在各个团体时的服装做到了80%~90%的还原。注意：该LoRA仅适用于Anima底模，无法在Illustrious底模上使用。This LoRA is mainly trained on Kagamine Len from Project SEKAI COLORFUL STAGE!. It restores the game's card illustration style and achieves 80%~90% restoration of each group's costumes in the game.Note: This LoRA only works with Anima base models and is incompatible with Illustrious base models.触发词/Trigger Word1boy, Len pjsk,male, male focus, solo, blonde hair, spiked hair, short ponytail, aqua eyes, 服装对应关键词/Costume Keywordsvs outfit, (white and yellow sailor shirt,black shorts, leg warmers, yellow-white shoes,)ln outfit, (studded choker, black choker, black jacket, yellow plaid hoodie ,white shirt,pink tie, black shoes,)mmj outfit, (mini hat, short gloves, white elaborate ringmaster outfit,)vbs1 outfit, (gray puffy vest, blue shirt, dark cargo pants, brown boots,)vbs2 outfit, (black cropped jacket, checkered shirt, striped tie, beige shorts, black platform boots,)ws outfit, (vibrant orange ringmaster outfit)25ji outfit, (white ruffled shirt, black shorts, mismatched socks,)使用方法/Usage Guide温馨提示：Anima类模型使用普通标签+自然语言的组合方式效果最佳，仅单独使用其中一种方式也可生效，但生成效果会大幅下降。在提示词的普通标签部分直接使用对应服装的关键词，同时在自然语言描述部分简单概括服装细节，可自行输入也可参考上方括号内容。 补充说明：镜音连VS团公式服 (vs outfit)还原度不稳定，生成该服装时，可按需将触发词从「Len pjsk」替换为通用词「kagamine len」；除公式服外，其余服装不建议更换触发词。Tip: Anima series models achieve the best generation results with a combination of tags and natural language descriptions. Using only one of them is feasible, but the output quality will be greatly reduced.Use the corresponding costume keywords directly in the tag section of your prompt, and briefly describe the costume details in the natural language part. You can customize the description or refer to the content in parentheses above. Note: The reduction effect of Len's VS official outfit (vs outfit) is unstable. You can replace the trigger word from "Len pjsk" to the general word "kagamine len" when generating this outfit. For all other costumes, this replacement is not recommended.

Files with SHA256: 04984f6bc8ba47611062c327d6e71a4ff9caaed331e3f9e4d159bb234880edc4