Skip to content

A question bout mini-batch size #4

@JerryWang22

Description

@JerryWang22

Hi, thank you for sharing you work. I have a question about the parameters setting. Durding training, you set the mini-batch size to 1 and the number of iterations to 6e5. Is there any theory or experiment that proves that such a setting is better than a large mini-batch size and a small number of iterations?

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