Skip to content

training from scratch problem #23

@andyl-flwls

Description

@andyl-flwls

I have a single A10 for training the model, from the start of the training, it says:

Found NaN, decreased lg_loss_scale to 19.0

is that normal?

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