/Python/When training= false still dropout

When training= false still dropout

from keras import layers
from keras import models
from keras import backend as K
import numpy as np

inp = layers.Input(shape=(10,))
out = layers.Dropout(0.5)(inp)

model = models.Model(inp, out)
model.layers[-1].trainable = False  # set dropout layer as non-trainable
model.compile(optimizer='adam', loss='mse') # IMPORTANT: we must always compile model after changing `trainable` attribute

# create a custom backend function so that we can control the learning phase
func = K.function(model.inputs + [K.learning_phase()], model.outputs)

x = np.ones((1,10))
# learning phase = 1, i.e. training mode
print(func([x, 1]))
# the output will be:
[array([[2., 2., 2., 0., 0., 2., 2., 2., 0., 0.]], dtype=float32)]
# as you can see some of the neurons have been dropped

# now set learning phase = 0, i.e test mode
print(func([x, 0]))
# the output will be:
[array([[1., 1., 1., 1., 1., 1., 1., 1., 1., 1.]], dtype=float32)]
# unsurprisingly, no neurons have been dropped in test phase

Source: stackoverflow.com

Similar codes

4 votes124 views3 answers
ArrayField in Django
Python
-3 votes134 views1 answer
Django socketio web chat example
Python Django
1 vote119 views1 answer
Python - Comment chiffrer une boucle
Python
1 vote127 views1 answer
How to iteratively create a grid within a bigger grid in python
Python
13 votes130 views6 answers
Remove special characters from string python
Python
7 votes105 views2 answers
Python prime check
Python
1 vote106 views3 answers
Pandas add column from list
Python
0 vote111 views1 answer
Static Assets in Django
Python
17 votes114 views2 answers
WARNING: You are using pip version 19.2.3, however version 21.2.4 is available.
Python
1 vote107 views2 answers
Types of all columns pandas
Python