Is there any way to compute the number of trainable parameters from code in a deep CNN network?

Hello

I am using a six layer compact CNN model for classification after intantiating the layers and training data to trainNetwork(). I want to calculate the number of trainable parameters in this network. Something similar to the below in pytorch:

model=net() 
for parameter in model.parameters():
    print(parameter)
    

Is there any way avaible in MATLAB?

Answer 1

[Harsh]

Hi @krishna Chauhan,

You can calculate the number of trainable parameters in your Convolutional Neural Network (CNN) model using the “analyzeNetwork” function, which provides a detailed layer-by-layer breakdown, including the count of learnable parameters for each layer.

Please refer to the following documentation for more information regarding “analyzeNetwork” function - https://www.mathworks.com/help/deeplearning/ref/analyzenetwork.html