Skip to content

Instantly share code, notes, and snippets.

@bwallace
Last active August 16, 2016 20:35
Show Gist options
  • Save bwallace/25d34f65ec4ea8f8e540eeb2edc27176 to your computer and use it in GitHub Desktop.
Save bwallace/25d34f65ec4ea8f8e540eeb2edc27176 to your computer and use it in GitHub Desktop.
input = Input(shape=(2,))
probs = Dense(2, activation='softmax', name='probs')(input)
probs = Dropout(1e-100)(probs)
model = Model(input=input, output=probs)
model.compile(optimizer='sgd', loss='categorical_crossentropy', metrics=['accuracy'])
from keras.utils.np_utils import to_categorical
X, y = np.array([[1, 2], [3, 4]]), to_categorical([1, 0])
model.fit(X, y, validation_data=[X, y])
probas = model.get_layer('probs')
import keras.backend as K
f = K.function(inputs=model.inputs + [K.learning_phase()], outputs=[probas.output])
# test mode = 0, train mode = 1
f(inputs=[X, 1])
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment