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

把merge CNN into FC是完全等价的吗? #1

Closed
Fengwills opened this issue May 11, 2021 · 3 comments
Closed

把merge CNN into FC是完全等价的吗? #1

Fengwills opened this issue May 11, 2021 · 3 comments

Comments

@Fengwills
Copy link

把merge CNN into FC是完全等价的吗?
还是像formulation部分说的一种输入输出维度相同的替换?
如果是完全等价,可否在训练阶段也去掉CNN?

@DingXiaoH
Copy link
Owner

DingXiaoH commented May 11, 2021 via email

@Fengwills
Copy link
Author

是完全等价的,但我不知道第二句的意思是什么。训练阶段去掉,性能就会变差,这在实验的第一节已经展示了。 @.*** 发件人: will 发送时间: 2021-05-11 11:25 收件人: DingXiaoH/RepMLP 抄送: Subscribed 主题: [DingXiaoH/RepMLP] 把merge CNN into FC是完全等价的吗? (#1) 把merge CNN into FC是完全等价的吗? 还是像formulation部分说的一种输入输出维度相同的替换? 如果是完全等价,可否在训练阶段也去掉CNN? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.

论文中formulation 部分说的CONV替换成MMUL,其实只是输入输出为度相同,过程并不相同吧。如果完全等价的话,按理说,在训练阶段去掉也可以吧。实验第一节是指在训练阶段也去掉了CNN?完全不加local perceptron这部分?只保留global perceptron 和 partition perceptron?

@DingXiaoH
Copy link
Owner

DingXiaoH commented May 21, 2021 via email

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