Can I use more network layers？

[quqxui]

Hi, thanks for your job and code,
I have a question about that the number of network layer. I noticed you only using one conv2d and one FC, as everyone knows, deep model can get better performance. I wonder why you don't use more convolution and FC layers? Doesn't the deep model help performance?

Hope your reply, thanks again.