Skip to content

NAT可以使用其他结构代替吗? #31

Description

@VincentZhao1992

目前正在尝试onnx导出,但是存在精度损失过大问题,在最后导出的onnx模型中,推理结果与pytorch版本差距较大,尤其在其他目标的行为预测prediction上,prediction的第一个通道损失看起来还可以接受,但在后面的目标上,几乎无法接受,在逐层特征分析时,发现NAT这个结构中,特征损失达到了无法接受的程度,那么我想请问一下,NAT可以使用其他结构代替吗?

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

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions