Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

关于深层网络的迁移问题 #2

Open
zzh-ecnu opened this issue Jan 6, 2021 · 1 comment
Open

关于深层网络的迁移问题 #2

zzh-ecnu opened this issue Jan 6, 2021 · 1 comment

Comments

@zzh-ecnu
Copy link

zzh-ecnu commented Jan 6, 2021

你好~

我看到在README里有人脸识别的结果,包括resnet50和resnet101,但是code里好像是一个5层CNN。

所以想请教一下对于resnet的深层网络,怎么将DGR迁移过去,需要做出哪些改变?

谢谢~

@RenMin1991
Copy link
Owner

RenMin1991 commented Jan 6, 2021

需要在以下三个方面进行出调整:

  1. 调整Graph模块的插入位置,对于resnet而言,可以在最后一个(single-scale)或三个(multi-scale)block之后插入Graph模块;
  2. 调整Graph的节点数量,其所插入的feature map尺寸越大,所需要的节点数量越多,例如当前feature map的尺寸为 H X W,可以将节点数量设置为 2*(H*W)^0.5附近的数值。
  3. 调整Graph模块的初始化参数,我们在训练中发现模型性能会在一定程度上受到初始化参数的影响,具体可以在 model/model_singlescale.py 中的 GraphGen 类的 _init_params 函数中进行设置。
    希望有所帮助!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants