Autoencoder Conv1D Wrong shape

Question

I am trying to train a model based on a number of txt. file with 2623 lines of numbers. The model is as follows:

input_img = Input(shape=(2623,1), name='input')

x = Conv1D(32, 3, activation='relu', padding='same', use_bias=False)(input_img)
x = BatchNormalization(axis=-1)(x)
x = MaxPooling1D(2, padding='same')(x)
x = Conv1D(16, 3, activation='relu', padding='same', use_bias=False, input_shape=(2623,1))(x)
x = BatchNormalization(axis=-1)(x)
x = MaxPooling1D(2, padding='same')(x)
x = Conv1D(16,3, activation='relu', padding='same', use_bias=False, input_shape=(2623,1))(x)
x = BatchNormalization(axis=-1)(x)
encoded = MaxPooling1D(2, padding='same')(x)


x = Conv1D(16,3, activation='relu', padding='same', use_bias=False, input_shape=(2623,1))(encoded)
x = BatchNormalization(axis=-1)(x)
x = UpSampling1D(2)(x)
x = Conv1D(16,3, activation='relu', padding='same', use_bias=False, input_shape=(2623,1))(x)
x = BatchNormalization(axis=-1)(x)
x = UpSampling1D(2)(x)
x = Conv1D(32, 3, activation='relu', padding='same', use_bias=False, input_shape=(2623,1))(x)  #input_shape=(30, 1))
x = BatchNormalization(axis=-1)(x)  
x = UpSampling1D(2)(x)
decoded = Conv1D(3, 3, activation='sigmoid', padding='same', use_bias=False,input_shape=(2623,1))(x)

autoencoder = Model(input_img, decoded)
autoencoder.compile(optimizer='rmsprop', loss='binary_crossentropy')


x = Input(shape=(16, 300), name="input")
h = x
h = Conv1D(filters=300, kernel_size=16,
             activation="relu", padding='same', name='Conv1',input_shape=(2623,1))(h)
h = MaxPooling1D(pool_size=16, name='Maxpool1')(h)

Whenever i try to fit the model with the following code i get an error saying: Error when checking target: expected conv1d_283 to have shape (2624, 1) but got array with shape (2623, 1)

I tried different padding and setting the input space smaller but it does not seem to work. I do not get the error because whatever i do in the end i have an error for different dimensions. Any ideas?

history = autoencoder.fit(train, train,
                epochs=100,
                batch_size=256,
                shuffle=True,
                validation_data=(test, test))

Check your train dimensions.They must have shape (2624,1).Or, change input_shape according to your train shape — George
– George, Commented Oct 5, 2018 at 11:34
my train shape is (1020, 2623) which basically is 1020 txt files with 2623 lines each so it is the same i guess — powersiganos powersiganos
– powersiganos powersiganos, Commented Oct 5, 2018 at 11:35
Can you print and tell the shape of encoded? Also there are multiple Conv1Ds. Check which of them is giving the error. — betelgeuse
– betelgeuse, Commented Oct 5, 2018 at 11:36
Hmm running encoded.shape i get: TensorShape([Dimension(None), Dimension(328), Dimension(16)]) — powersiganos powersiganos
– powersiganos powersiganos, Commented Oct 5, 2018 at 11:38
Decoded.shape() gives: TensorShape([Dimension(None), Dimension(2624), Dimension(3)]) on the other hand. I think this is the problem — powersiganos powersiganos
– powersiganos powersiganos, Commented Oct 5, 2018 at 11:40

desertnaut · Accepted Answer · 2018-10-05 15:00:07Z

1

Thank you for the help. In the end adding a croplayer in the end did the trick, i.e:

x = Cropping1D(cropping=(0, 1))(x) # Crop nothing from input but crop 1 elemnt from the end

edited Oct 5, 2018 at 15:00

desertnaut

60.9k32 gold badges158 silver badges184 bronze badges

answered Oct 5, 2018 at 13:17

powersiganos powersiganos

3310 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Autoencoder Conv1D Wrong shape

1 Answer 1

Comments

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Related