How to set parameters in keras to be non-trainable?

Question

I am new to Keras and I am building a model. I want to freeze the weights of the last few layers of the model while training the previous layers. I tried to set the trainable property of the lateral model to be False, but it dosen't seem to work. Here is the code and the model summary:

opt = optimizers.Adam(1e-3)
domain_layers = self._build_domain_regressor()
domain_layers.trainble = False
feature_extrator = self._build_common()
img_inputs = Input(shape=(160, 160, 3))
conv_out = feature_extrator(img_inputs)
domain_label = domain_layers(conv_out)
self.domain_regressor = Model(img_inputs, domain_label)
self.domain_regressor.compile(optimizer = opt, loss='binary_crossentropy', metrics=['accuracy'])
self.domain_regressor.summary()

The model summary: model summary.

As you can see, model_1 is trainable. But according to the code, it is set to be non-trainable.

Answer 1

You can simple assign a boolean value to the layer property trainable.

model.layers[n].trainable = False

You can visualize which layer is trainable:

for l in model.layers:
    print(l.name, l.trainable)

You can pass it by the model definition too:

frozen_layer = Dense(32, trainable=False)

From Keras documentation:

To "freeze" a layer means to exclude it from training, i.e. its weights will never be updated. This is useful in the context of fine-tuning a model, or using fixed embeddings for a text input.
You can pass a trainable argument (boolean) to a layer constructor to set a layer to be non-trainable. Additionally, you can set the trainable property of a layer to True or False after instantiation. For this to take effect, you will need to call compile() on your model after modifying the trainable property.