Thanks for you code, it helps me to learn how to emotion recognition.
But when I read your code,there is something I can't understand.
In the paper inside your repo,I find that it use a "valid" padding so the image shape is
[4848]->[4444]->[2222]->.....->3072
while in your code it is that:
[4848]->[4848]->[2424]->......->3072
I can't understand this.Can you tell me why?Thanks very much
Thanks for you code, it helps me to learn how to emotion recognition. But when I read your code,there is something I can't understand. In the paper inside your repo,I find that it use a "valid" padding so the image shape is [4848]->[4444]->[2222]->.....->3072 while in your code it is that: [4848]->[4848]->[2424]->......->3072 I can't understand this.Can you tell me why?Thanks very much