Extracting a kernel from a trained network
One way to interpret models is to examine the properties of the kernels in the convolutional layers. In this exercise, you will extract one of the kernels from a convolutional neural network with weights that you saved in a hdf5 file.
This exercise is part of the course
Image Modeling with Keras
Exercise instructions
- Load the weights into the model from the file
weights.hdf5. - Get the first convolutional layer in the model from the
layersattribute. - Use the
.get_weights()method to extract the weights from this layer.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Load the weights into the model
model.____('weights.hdf5')
# Get the first convolutional layer from the model
c1 = model.____[0]
# Get the weights of the first convolutional layer
weights1 = c1.____()
# Pull out the first channel of the first kernel in the first layer
kernel = weights1[0][...,0, 0]
print(kernel)