Skip to content

Loss almost does not change #7

Description

@Oguzhanercan

Hi, I am trying to experience different losses. I Have implement a face similarity loss and disabled all of other losses. But The loss almost does not changed (%1 decreased). At my method, I am looking at the cosine similarity between the reference image and generated image like RectifID(https://github.com/feifeiobama/RectifID). Do you have any experiments with that or do you have any suggestions about it?

Face recognition model: Arcface Resnet200
Loss: (1 - Cosine simlarity)
Lr: 1 to 30 (I tried different settings on this)
SD Model: Pixart

Activity

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

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions