Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Try using UC_DALLE to train on cifar10 dataset #49

@twq0105

Description

@twq0105

When I use unconditional dalle to train on cifar10 dataset, the loss always fluctuates around 3 and can hardly go down. My settings are consistent with those in ffhq.yaml. Please tell me what is wrong? Or has anyone successfully reduced the loss and achieved better results when reproducing it?

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