This is an Illustrious LoRA model trained on the character Hikawa Hina from BanG Dream!.
The trigger word is hikawa hina
The model was trained on the anime-style base model wai-IllustriousSDXL_v170
The training dataset consists of 136 images, sourced from anime screenshots of BanG Dream! Season 2 and Season 3（26 images）, as well as official card art (79 old-style cards + 31 new-style cards). The vast majority of these images have plain/transparent backgrounds.
During training, the images were tagged with bangdreamanime, oldcard, and newcard respectively. However, these tags appear to have no significant impact on the generated images
这是一个基于《BanG Dream!》角色冰川日菜训练的 illustrious LoRA模型。
触发词是hikawa hina
模型在二次元大模型 wai-IllustriousSDXL_v170上训练
训练集是136张分别来源于Bangdream第二季和第三季的动漫截图（26张）和官方卡面（79张旧画风+31新画风），其中绝大多数是无背景卡面
训练时分别标注了bangdreamanime，oldcard，newcard，但似乎对生成的图片没有影响
If you have any questions, suggestions, or would like to discuss this model, feel free to leave a comment or send me a DM, and If you spot any bugs or areas that could be improved while using this model, please let me know so I can update it accordingly.
如有问题或想讨论，可以留言或私信联系我，如果在使用过程中发现有任何值得注意的问题和可以改进的地方，请告诉我以更新这个LoRA
Description
Comments (2)
You need provide trigger words not just for the character but her outfits and other details. Not everyone wants to load it all up on ComfyUI or can do that.
key features like her hairstyle, eye color, have already been fused into the primary trigger word during training, you don't need to add extra tags for her appearance; clothing is not the primary focus of this LoRA. The school uniform is the only outfit that appears frequently and in the training dataset, use haneoka_school_uniform