7

I am using ubuntu with python 3 and keras over tensorflow, I am trying to create a model using transfer learning from a pre trained keras model as explained here:

I am using the following code 

import numpy as np
from keras.applications import vgg16, inception_v3, resnet50, mobilenet
from keras import Model

a = np.random.rand(1, 224, 224, 3) + 0.001
a = mobilenet.preprocess_input(a)

mobilenet_model = mobilenet.MobileNet(weights='imagenet')

mobilenet_model.summary()
inputLayer = mobilenet_model.input

m = Model(input=inputLayer, outputs=mobilenet_model.get_layer("conv_pw_13_relu")(inputLayer))
m.set_weights(mobilenet_model.get_weights()[:len(m.get_weights())])
p = m.predict(a)
print(np.std(p), np.mean(p))
print(p.shape)

The output of the layer I am using is always an array of zeros, Should I load the weight to p that i am creating in order for the pre trained model to actually work?

thebeancounter
  • 4,261
  • 8
  • 61
  • 109

2 Answers2

14

There is a difference between layers and the outputs of those layers in Keras. You can think of layers as representing a computation and the outputs as the results of those computation. When you instantiate a Model object, it expects the results of a computation as it's output, instead of the computation itself, hence the error. To fix it, you can pass the output of the layer to the Model constructor:

import numpy as np
from keras.applications import vgg16, inception_v3, resnet50, mobilenet
from keras import Model

a = np.random.rand(24, 224, 224, 3)
a = mobilenet.preprocess_input(a)

mobilenet_model = mobilenet.MobileNet(weights='imagenet')

mobilenet_model.summary()

model_output = mobilenet_model.get_layer("conv_pw_13_relu").output
m = Model(inputs=mobilenet_model.input, outputs=model_output)
print(m.predict(a))
Agost Biro
  • 2,709
  • 1
  • 20
  • 33
2

In order to acess the output of an intermediate layer in a Keras model, Keras Provides different ways.

In your case You can take output of the layer you want like this

model_out = mobilenet_model.get_layer("layer_you_want").output
m = Model(input=inputLayer, outputs=model_out)

For more details regarding this and other methods available take a look at this documentation

Sreeram TP
  • 11,346
  • 7
  • 54
  • 108