Do Pytorch models have concepts similar to top and bottom in caffe structure files

1

Does the Pytorch model have top and bottom concepts similar to those in the caffe structure file, and if so, how do you get the

2

I don’t think it works the same.
I implemented caffe models in pytorch, and while caffe use top and bottom to define layers, this is defined in the forward function in pytorch. So you need to know how data is passed to layers to know what are top and bottoms.

3

How do you implement caffe on pytorch, and how do you define the input and output relationships between layers

4

I display the caffe model using : Netscope
And then create the pytorch model myself and tadaaa.

output = layer(input)

simple as that