went through the lesson 2 video. Was interested in knowing what the ConvNetBuilder class code was doing… was looking into the ConvNetBuilder class code , specifically the constructor class “__init”
May I know what does the below lines of code do:
self.nf = model_features[f] if f in model_features else (num_features(layers)*2)
Is there any reason to multiply the value found by the function num_features by 2 ?
I am guessing its trying to find the unrolled feature vector which is to be connected to the linear layers.