I can print the network and it also shows the layers, however it doesn't give me other information. summary(model, ...) needs the size of input to the network. But I'm not sure how to define the size of a dictionary and this method seems not suitable for a dictionary input. I tried to use torchinfohttps://github.com/TylerYep/torchinfo, but they require passing the dict as a list of args or a dict of kwargs to the forward() function.
How did you manage to summarize it, when the network takes a dictionary as input? Or do you have any other methods to get the total number of parameters of the network?
Hi,
I am trying to summarize the network as the following form to have an overview of its architecture and number of parameters:
I can print the network and it also shows the layers, however it doesn't give me other information.
summary(model, ...)
needs the size of input to the network. But I'm not sure how to define the size of a dictionary and this method seems not suitable for a dictionary input. I tried to usetorchinfo
https://github.com/TylerYep/torchinfo, but they require passing the dict as a list of args or a dict of kwargs to the forward() function. How did you manage to summarize it, when the network takes a dictionary as input? Or do you have any other methods to get the total number of parameters of the network?