Hi, I have a saved params.pkl file for my trained model but I don’t know the architecture. How can I somehow visualize or find out the network architecture?
The params.pkl file contains a dictionary of model parameter values, and contains no information about the model architecture itself. In some case, if standard model architecture was use and the params were names aptly, one can guess the architecture of the model from the keys of the dictionary. But this is approach is completely based on luck.