In your paper, the kernel size (3x7x7) in the 'conv1' layer of ResNeXt-101 and ResNet-10 are different from that (7x7x7) in your code. Could you explain it? Whether the method in your code could get a higher performance?
In addition, I did not find your single-time activation function in your code. Could u point it out?
Hi,
Thank you in advance for attention.
In your paper, the kernel size (3x7x7) in the 'conv1' layer of ResNeXt-101 and ResNet-10 are different from that (7x7x7) in your code. Could you explain it? Whether the method in your code could get a higher performance?
In addition, I did not find your single-time activation function in your code. Could u point it out?
Looking forward to your reply.