This is the model of my latest training ideasIt is very sensitive to tag, overfitting is almost solved, even the training time only needs 2/3 of the originalThe example diagram uses the donko-mix nsfw hard modelIt is recommended to open Hires fix这是我最新训练思路所训练出的模型对tag很敏感，过拟合几乎被解决，甚至训练时间只需要原本的2/3示例图使用的模型为donko-mix nsfw hard，建议打开高清修复另外我会接LoRA模型定制单，有意加群找同名管理员私聊1018450371

Files with SHA256: f01cac92e009aab674f119d7db0f742ec4fe0ee09e928bf1adb9ddbfd1a3c0cb