如果大家有喜欢或推荐的风格，请留言，合适的话有时间会加到后续版本。仅业余爱好折腾，如有版权问题请联系！If there is any fantacy styles you like, please leave your comments here. Would add it to the future version if possible. This Lora is only a hobby toss, if have any copyright issues, please contact me!基于MJ精致图片训练的梦幻童话风格Lora，可以用较短的提示语生成相应风格的梦幻图片。欢迎试玩！Some fairy tale styles trained on midjouney picutures. We could generated specific style picture with a short prompts. Just try and enjoy!<ol><li>paper cut. keyword: paper-cut [craft/art painting]. recommended weight:~0.7.</li><li>super cute animal. keyword: fairy tale. recommended weight:~0.7.</li><li>delicate scene/CG rendering. keyword: delicate scene. recommended weight:~0.7.</li><li>rime. keyword: rime. recommended weight:~0.7.</li><li>crystal flower. keyword: transparent/crystal . recommended weight:0.3-1.0.</li><li>others. keyword: close-up display/ animie 1gir/fantacy painting.</li></ol>2023/5/10 Update V2 version: included papercut styles based on 60 incremental papercut pictures.<ol><li>two styles: craft and art paiting. key words: paper-cut [craft/art painting].</li><li>with/without photo-frame, key words: photo-frame/no-frame.</li><li>with/without mask, key words: [shape] mask/no-mask.</li></ol>2023/5/3 Update v1s version: enlarge texture embedding dimension and training batch size in AIStudio Platform(<a target="_blank" rel="ugc" href="https://aistudio.baidu.com">https://aistudio.baidu.com</a>) providing free GPU resources. Tag expression is more accurate and picture seems more clear and transparent.

Midjourney Fairy Tale(梦幻童话)

This version enlarge texture embedding dimension and training batch size. Tag expression is more accurate and picture seems more clear and transparent.

Files with SHA256: 2d2557e4bfa52435be1958cf916eee2053d0103118a6e494a3620a1ca78b1e16