CODEWITHSUNDEEP
pythonmachine-learningkeras
Tom Leung
Tom Leung

Reputation: 374

TypeError when using Keras Lambda layer

I am learning Keras and trying to build my layers, so I build a simple Lambda layer which just calculates the mean of the inputs.

TypeError occurred in line 4 when I add this layer to a sequential model. Here is my code:

mean_layer = keras.layers.Lambda(lambda x: tf.reduce_mean(x))
model = keras.models.Sequential()
model.add(keras.layers.InputLayer(input_shape=[10]))
model.add(mean_layer())
model.summary()

TypeError: __call__() missing 1 required positional argument: 'inputs'

However, if I use the functional model, the thing goes well and I get the expected output.

input = keras.layers.Input(shape=[10])
output = mean_layer(input)
model = keras.models.Model(inputs=[input], outputs=[output])
model.summary()

Did I build the model in the wrong way when I use the sequential API? Thanks

Upvotes: 0

Views: 206

Answers (1)

Dr. Snoopy
Dr. Snoopy

Reputation: 56357

Yes you are doing it wrong, it should be like this:

model.add(mean_layer)

Upvotes: 1

Related Questions