TensorFlow 2 AttributeError: 'tuple' object has no attribute 'layer'

Question

I was following a tutorial to use Keras pre-trained models but I am getting this error:

AttributeError: 'tuple' object has no attribute 'layer'

I am running my code on Google Colab, so I tried to downgrade TensorFlow to version 1.x, but the issue was not solved. So how can I solve this?

Here is my code fragment:

resnet = ResNet152V2(
             include_top=False,
             weights="imagenet",
             input_shape=(40,40,3)
)

for layer in resnet.layers:
    layer.trainable = False

x = resnet.output
x = Flatten()(x)
prediction = Dense(4, activation='softmax')(x)

model = Model(inputs=resnet.input, outputs=prediction)

model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

Answer 1

What dependencies did you import?

Had the same problem just now, and I fixed it by making sure that my dependencies were consistent. Take note that tensorflow.keras is different from importing from keras.