Skip to content

DPO的梯度系数推导不够严格？ #38

New issue
New issue
Open
Open
DPO的梯度系数推导不够严格？#38
@liyy96

Description

@liyy96
liyy96
opened on Feb 25, 2025

sigmoid内的求和号，应该不能被提取出来，和后面的token求和项融合？

Image

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