how to run IDC on my own cifar10-vgg16

[yuningkang]

sir ,we are trying to use the vgg16 model of cifar10 to run the idc apporch. The convolutional layer of this model requires padding and two convolutional layers are adjacent. We adjusted it according to your padding method for lenet5 like this

However, because the two convolutional layers are adjacent, the dimensions of self.Y and R do not match when the convolutional layer calculates R_norm

this is our vgg16 model one geogel drive:
https://drive.google.com/file/d/1bjzJxwWBGewZxdoiUkX9YR8r1rUoxjj6/view?usp=sharing

This is our running command, ``python run.py --dataset cifar10 -L -1 --model neural_networks/cifar10_vgg16 -RN 10
please help us to solve this problem

[hfeniser]

Hi @yuningkang! I am not sure if I understood clearly. I believe, even though the two convolutional layers are adjacent this padding method should work. Could you give us more information?