CODEWITHSUNDEEP
pythonmachine-learningkeraskeras-layervgg-net
wellido
wellido

Reputation: 45

How to add a Lambda layer as an input layer to an existing model in Keras?

I have a task to add a image preprocessing layer to a Keras model, so after I loaded a Keras model, I want to add a new input layer for this model.

I found I can use Lambda layer to preprocess the image data. The layer code is:

def vgg16preprocessing(x):
    mean_tensor = K.backend.variable([125.307, 122.95, 113.865], name="mean")
    std_tensor = K.backend.constant([62.9932, 62.0887, 66.7048], name="std_tensor")
    result = (x - mean_tensor) / (std_tensor)
    return K.backend.reshape(result, (-1, 32, 32, 3))
preproc_layer = K.layers.Lambda(vgg16preprocessing, output_shape=(32, 32, 3), input_shape=(32, 32, 3))

But I don't know how to add this layer in the front of my model. I found this answer, but I can't add the layer in the keras.layers.Input().

Is there any ways to set the Lambda layer as a new input layer?

Upvotes: 3

Views: 1006

Answers (1)

today
today

Reputation: 33410

You can use the VGG16 model and apply it on the output of Lambda layer:

vgg = VGG16(...)

input_img = Input(shape=...)
preproc_img = Lambda(vgg16preprocessing)(input_img)
output = vgg(preproc_img)

model = Model(input_img, output)

Upvotes: 2

Related Questions