RuntimeError: Calculated padded input size per channel: (4 x 4). Kernel size: (5 x 5). Kernel size can't be greater than actual input size

Dear all,

I know this error is not new somehow but I’m stuck though I follow some suggestions with those similar errors.
The network is trained with Inception Model V3 and then fully connected with the following sequential classifier.
Before training with Inception V3, trained with ResNet and FC with the same architecture classifier. And it worked.
I’m trying to implement the ensemble NN.
Welcome every suggestion and thanks in advance.

Your input is too small for the model and I believe Inception models were trained with inputs of [299, 299]. You might this need to increase the spatially dimension of your input or try to decrease the pooling in the model.

1 Like