Combined with the bottom mold of whiteXL realistic, incorporating some excellent models. After cleaning the fixed real features of realistic, it can now stably output various styles like base, real, cartoon, and fantasy.

I actually tested the generalization and it works well. It can be used as a base model for specialized model training, or as a base model for training Lora in a specified direction. The results are both good

XLbase is an excellent model, and better development in the future lies in the establishment of a community and the continuous iteration of developers.

I hope you like it

结合了whiteXLrealistic的底模，融合了一些优秀的模型。把realistic固定的真实特征清洗了一遍，现在可以像base一样稳定输出各种风格，真实的，卡通的，幻想的。

实际测了一遍泛化比较好，可以作为底模进行特化模型训练，也可以作为lora的底模训练指定方向的lora，效果都不错

XLbase是一个优秀的模型，后续更好的开发在于社群的建立，以及各位开发者不断的迭代。

希望你们喜欢