Skip to content

请问训练5个epoch后损失在1左右,单从图像级精度来看取得了不错的效果,为什么损失再没收敛的情况下,能达到这样的精度呢? #51

Open

Description

@Sword987
Image

Activity

  1. caoyunkang commented on Nov 1, 2025

    @caoyunkang
    Owner

    零样本异常检测其实不需要过分的拟合,不然存在过拟合风险。这一点也可以从我们的模型设计一般不会添加太多额外的参数看出 -- 过多的参数,会增大模型的表征能力,以至于过拟合风险增加。

  2. QinPengfei-star commented on Nov 13, 2025

    @QinPengfei-star

    Image为什么我复现出来的结果和你这个结果相差这么多呢,哭T^T,有解决办法吗

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions