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

26是否可以替换成其他更小数值 #70

Open
laocan opened this issue Sep 23, 2022 · 2 comments
Open

26是否可以替换成其他更小数值 #70

laocan opened this issue Sep 23, 2022 · 2 comments

Comments

@laocan
Copy link

laocan commented Sep 23, 2022

看res2net的layer1部分:26*4 = 104 ==conv1x1==> 256
想减小通道数,把256 改成 64
请问,可以把26改成13,或者更小吗?会有什么不好的影响吗

比如下面:
① 134 = 52 ==conv1x1==> 64
② 7
4 = 28 ==conv1x1==> 64
谢谢

@gasvn
Copy link
Member

gasvn commented Sep 27, 2022

通道数可以减少,但是一般情况下由于参数量下降性能也会相应的下降。

@laocan
Copy link
Author

laocan commented Sep 27, 2022

多谢回复~
尝试改成 7,发现训练的曲线跟13和26基本吻合。
可能是我的场景下,特征比较容易提取。
请问,有对比过不同通道数对性能指标的影响吗?有可以参考的论文吗?
谢谢~

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