How to install scales for Convolution2D?

Question

I would like to set the weight of the Convolution2D layer:

 conv = Convolution2D(conv_out_size, window_size, embedding_size, border_mode='same', activation='relu', weights=weights, name='conv_{:d}'.format(i))(in_x) 

but I'm not sure what is expected here. I tried a few things, but most of the time I get

 ValueError: You called `set_weights(weights)` on layer "conv_0" with a weight list of length 1, but the layer was expecting 2 weights. 

Not sure what that means.

Answer 1

You must pass the numpy array to your convolutional layer using the set_weights method.

Remember that the weight of the convolutional layer is not only the weight of each individual filter, but also the offset . Therefore, if you want to set the weight, you need to add an extra dimension.

For example, if you want to install a 1x3x3 filter with all weights of zero, except for the central element, you should do this:

 w = np.asarray([ [[[ [0,0,0], [0,2,0], [0,0,0] ]]] ]) 

And then install it.

For the code you could run:

 #!/usr/bin/env python # -*- coding: utf-8 -*- from __future__ import print_function import numpy as np np.random.seed(1234) from keras.layers import Input from keras.layers.convolutional import Convolution2D from keras.models import Model print("Building Model...") inp = Input(shape=(1,None,None)) output = Convolution2D(1, 3, 3, border_mode='same', init='normal',bias=False)(inp) model_network = Model(input=inp, output=output) print("Weights before change:") print (model_network.layers[1].get_weights()) w = np.asarray([ [[[ [0,0,0], [0,2,0], [0,0,0] ]]] ]) input_mat = np.asarray([ [[ [1.,2.,3.], [4.,5.,6.], [7.,8.,9.] ]] ]) model_network.layers[1].set_weights(w) print("Weights after change:") print(model_network.layers[1].get_weights()) print("Input:") print(input_mat) print("Output:") print(model_network.predict(input_mat)) 

Try changing the central element in the convolutional placeholder (2 in the example).

What the code does:

Build the model first.

 inp = Input(shape=(1,None,None)) output = Convolution2D(1, 3, 3, border_mode='same', init='normal',bias=False)(inp) model_network = Model(input=inp, output=output) 

Print the original weights (initialized by normal distribution, init = 'normal')

 print (model_network.layers[1].get_weights()) 

Create the desired weight tensor w and some input input_mat

 w = np.asarray([ [[[ [0,0,0], [0,2,0], [0,0,0] ]]] ]) input_mat = np.asarray([ [[ [1.,2.,3.], [4.,5.,6.], [7.,8.,9.] ]] ]) 

install the scales and print them

 model_network.layers[1].set_weights(w) print("Weights after change:") print(model_network.layers[1].get_weights()) 

Finally, use it to generate prediction output (prediction will automatically compile your model)

 print(model_network.predict(input_mat)) 

Result:

 Using Theano backend. Building Model... Weights before change: [array([[[[ 0.02357176, -0.05954878, 0.07163535], [-0.01563259, -0.03602944, 0.04435815], [ 0.04297942, -0.03182618, 0.00078482]]]], dtype=float32)] Weights after change: [array([[[[ 0., 0., 0.], [ 0., 2., 0.], [ 0., 0., 0.]]]], dtype=float32)] Input: [[[[ 1. 2. 3.] [ 4. 5. 6.] [ 7. 8. 9.]]]] Output: [[[[ 2. 4. 6.] [ 8. 10. 12.] [ 14. 16. 18.]]]]