这个模型是基于T-Pony下训练的特蕾西娅的个人角色lora，使用时，激活词为“theresa”,模型训练数据来看，部分时候可能会有点过拟合，但不妨碍模型的表现，如果有过拟合现象，降低模型的权重或者调低CFG都可以解决问题，作者B站账号名：特蕾西娅的岁月人年推荐使用Euler a采样器，CFG默认为6，若出现过拟合状态可降到4或者3，迭代步数跟T-PONY一样是20步左右，也可以视情况更高，CLIP推荐2默认裙子的提示词是三个：cleavage_cutout,dress,white dress

[SDXL] T-Pony  明日方舟 特蕾西娅  Theresa arknights

改为以pony为底模，然后用于t-pony出图，因为t-pony训练中发现会有些许偏差，同时改善了人物头部的训练，角色头部特征相比之前更稳定明显，服饰的话更换相比之前更灵活，效果也更好，受原服饰影响比之前小Change to using Pony as the base model and then use it for t-Pony drawing, as some slight deviations were found during t-Pony training. At the same time, the training of the character's head has been improved, and the characteristics of the character's head are more stable and obvious compared to before. The clothing changes are more flexible than before, and the effect is better. It is less affected by the original clothing than before

theresa_Pony_1.safetensors

v2.0

v1.0

THIS IS A REPRINT!Go to original author to check their other models or to support them. <hr /> 这个模型是基于T-Pony下训练的特蕾西娅的个人角色lora，使用时，激活词为“theresa”,模型训练数据来看，部分时候可能会有点过拟合，但不妨碍模型的表现，如果有过拟合现象，降低模型的权重或者调低CFG都可以解决问题，作者B站账号名：特蕾西娅的岁月人年推荐使用Euler a采样器，CFG默认为7，若出现过拟合状态可降到4或者3，迭代步数跟T-PONY一样是20步左右，也可以视情况更高，CLIP推荐2默认裙子的提示词是三个：cleavage_cutout,dress,white dress

Files with SHA256: 200583aef26041c33d0d328cdd4b6d905b7aa47b508c67987505bf132a7e859b