how to change layer shape in keras?

user_11069811 0
Viewed 453

I have a decoder that has output (28,28,1)

I am trying that the output of a decoder has (32,32,3)

how can i achieve that?

i am using colab with gpu t4

latent_dim = 20

encoder_inputs = keras.Input(shape=(28, 28, 1))
x = layers.Conv2D(32, 3, activation="relu", strides=2, padding="same")(encoder_inputs)
x = layers.Conv2D(64, 3, activation="relu", strides=2, padding="same")(x)
x = layers.Flatten()(x)
x = layers.Dense(16, activation="relu")(x)
z_mean = layers.Dense(latent_dim, name="z_mean")(x)
z_log_var = layers.Dense(latent_dim, name="z_log_var")(x)
z = Sampling()([z_mean, z_log_var])
encoder = keras.Model(encoder_inputs, [z_mean, z_log_var, z], name="encoder")
encoder.summary()





latent_inputs = keras.Input(shape=(latent_dim,))
x = layers.Dense(7 * 7 * 64, activation="relu")(latent_inputs)
x = layers.Reshape((7, 7, 64))(x)
x = layers.Conv2DTranspose(64, 3, activation="relu", strides=2, padding="same")(x)
x = layers.Conv2DTranspose(32, 3, activation="relu", strides=2, padding="same")(x)
decoder_outputs = layers.Conv2DTranspose(1, 3, activation="sigmoid", padding="same")(x)
decoder = keras.Model(latent_inputs, decoder_outputs, name="decoder")
decoder.summary()


vae = VAE(encoder, decoder)
vae.compile(optimizer=keras.optimizers.Adam())
vae.fit(mnist_digits, epochs=30, batch_size=128)

summary decoder

how change (28,28,1) to (32,32,3)?

how can i achieve that?

Model: "decoder"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_6 (InputLayer)         [(None, 20)]              0         
_________________________________________________________________
dense_5 (Dense)              (None, 3136)              65856     
_________________________________________________________________
reshape_3 (Reshape)          (None, 7, 7, 64)          0         
_________________________________________________________________
conv2d_transpose_9 (Conv2DTr (None, 14, 14, 64)        36928     
_________________________________________________________________
conv2d_transpose_10 (Conv2DT (None, 28, 28, 32)        18464     
_________________________________________________________________
conv2d_transpose_11 (Conv2DT (None, 28, 28, 3)         867       
=================================================================
Total params: 122,115
Trainable params: 122,115
Non-trainable params: 0
_________________________________________________________________
keras tensorflow deep-learning machine-learning python
1 Answers

in your decoder, you can simply change this layers:

x = layers.Dense(7 * 7 * 64, activation="relu")(latent_inputs)
x = layers.Reshape((7, 7, 64))(x)

into

x = layers.Dense(8 * 8 * 64, activation="relu")(latent_inputs)
x = layers.Reshape((8, 8, 64))(x)

remember also to set 3 output channels in the last decoder layer.

Full example:

latent_dim = 20

class Sampling(layers.Layer):
    """Uses (z_mean, z_log_var) to sample z, the vector encoding a digit."""

    def call(self, inputs):
        z_mean, z_log_var = inputs
        batch = tf.shape(z_mean)[0]
        dim = tf.shape(z_mean)[1]
        epsilon = tf.keras.backend.random_normal(shape=(batch, dim))
        return z_mean + tf.exp(0.5 * z_log_var) * epsilon

encoder_inputs = Input(shape=(32, 32, 3))
x = layers.Conv2D(32, 3, activation="relu", strides=2, padding="same")(encoder_inputs)
x = layers.Conv2D(64, 3, activation="relu", strides=2, padding="same")(x)
x = layers.Flatten()(x)
x = layers.Dense(16, activation="relu")(x)
z_mean = layers.Dense(latent_dim, name="z_mean")(x)
z_log_var = layers.Dense(latent_dim, name="z_log_var")(x)
z = Sampling()([z_mean, z_log_var])
encoder = Model(encoder_inputs, [z_mean, z_log_var, z], name="encoder")

latent_inputs = Input(shape=(latent_dim,))
x = layers.Dense(8 * 8 * 64, activation="relu")(latent_inputs)
x = layers.Reshape((8, 8, 64))(x)
x = layers.Conv2DTranspose(64, 3, activation="relu", strides=2, padding="same")(x)
x = layers.Conv2DTranspose(32, 3, activation="relu", strides=2, padding="same")(x)
decoder_outputs = layers.Conv2DTranspose(3, 3, activation="sigmoid", padding="same")(x)
decoder = Model(latent_inputs, decoder_outputs, name="decoder")
user_103750490
Related