How get output value in array from each layer Keras

Question

I am a beginner in python and Keras. I am using Keras with tensorflow backend. I want to get value in array from each layer (hidden and output layer) in Keras. How can i do it?

This is my sequential model

def baseline_model():
# create model
model = Sequential()
model.add(Conv2D(32, (5, 5), input_shape=(1, 28, 28), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.2))
model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dense(num_classes, activation=tempsigmoid))
# Compile model
model.compile(loss='mse', optimizer='adam', metrics=['accuracy'])
return model

# build the model
model = baseline_model()

I had tried using this code

hidden_layer = model.layers[4].output
print(hidden_layer)

but the result in tensor

Tensor("dense_1/Relu:0", shape=(?, 128), dtype=float32)

Leevo · Accepted Answer · 2019-05-15 07:50:34Z

2

To extract the i-th layer of a Neural Network you can use Keras functions. Let's assume that you are training a Model on some data df:

from tensorflow.keras import backend as K

# create a Keras function to get i-th layer
get_layer_output = K.function(inputs = Model.layers[0].input, outputs = Model.layers[i].output)

# extract output
layer_output = get_layer_output(df)

You can find a practical application here. Hope this helps, otherwise let me know.

answered May 15, 2019 at 7:50

Leevo

1,7533 gold badges21 silver badges36 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

astri Over a year ago

thankyou for the response. I tried and got error message AttributeError: 'Conv2D' object has no attribute 'inputs' . Could it be because my model is convolutional ?

Leevo Over a year ago

Reading your error message... did you write inputs instead of input?

Timbus Calin Over a year ago

This is the right way, should be marked as final solution.

Leevo Over a year ago

@astri if it worked, please consider accepting the answer as solved

astri Over a year ago

@Leevo thank you , i have revised my code but it still get error TypeError: inputs` to a TensorFlow backend function should be a list or tuple.` I am figuring it now

|

Thibault Bacqueyrisses · Accepted Answer · 2019-05-15 08:51:13Z

0

You can do the same thing but save the data with that way :

hidden_layer = model.layers[4].output
hidden_layer = hidden_layer.eval(session=tf.Session())
print(hidden_layer)

That will save directly the Tensor data in the hidden_layer variable.

answered May 15, 2019 at 8:51

Thibault Bacqueyrisses

2,3511 gold badge9 silver badges18 bronze badges

Collectives™ on Stack Overflow

How get output value in array from each layer Keras

2 Answers 2

7 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

7 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related