How to apply different kernels to each example in a batch when using convolution?

vision
12

Thank you for the reply. You are right. I checked it on the tensors with random binary values. It became zero