I need to get Filter size, Activation function, padding, striding values of each layer from the pre-trained model which is compiled by NNVM compiler. When I parse the JSON, I get the layer names and when I tried to parse parameter file, I am getting the weights of every layer in the form of a matrix. I can load the json and h5 file to a model and iterate over the layers to get the info using the “get_config()” command . But is that the only way to do that?
