Wrong shape output with conv1D layer

I'm trying some experiments with a small autoencoder defined like this:

input_layer = Input(shape=input_shape)

x = Conv1D(8, kernel_size=5, activation='relu', padding='same')(input_layer)
x = BatchNormalization()(x)
encoded = Conv1D(4, kernel_size=5, activation='relu', padding='same')(x)

x = Conv1D(8, kernel_size=5, activation='relu', padding='same')(encoded)
x = BatchNormalization()(x)

decoded = Conv1D(WINDOW_SIZE, kernel_size=5, activation='relu', padding='same')(x)

input_shape = 8192 and autoencoder.summary() returns:

 Layer (type)                Output Shape              Param #
=================================================================
 input_1 (InputLayer)        [(None, 8192, 1)]         0

 conv1d (Conv1D)             (None, 8192, 8)           48

 batch_normalization (BatchN  (None, 8192, 8)          32
 ormalization)

 conv1d_1 (Conv1D)           (None, 8192, 4)           164

 conv1d_2 (Conv1D)           (None, 8192, 8)           168

 batch_normalization_1 (Batc  (None, 8192, 8)          32
 hNormalization)

 conv1d_3 (Conv1D)           (None, 8192, 8192)        335872

=================================================================
Total params: 336,316
Trainable params: 336,284
Non-trainable params: 32

Basically, I want an autoencoder that takes a vector of 8192 values and try to predict an another vector of 8192 values.

but when I do an inference like that:

predictions = autoencoder.predict(batch_data, batch_size=BATCH_SIZE)

I've that shape:

>>> predictions[0].shape
(8192,8192)

What I'm doing wrong ? what can I change to get a vector with (8192) ?

Maybe something like this?

import tensorflow as tf

input_layer = tf.keras.layers.Input(shape=(8192, 1))

x = tf.keras.layers.Conv1D(8, kernel_size=5, activation='relu', padding='same')(input_layer)
x = tf.keras.layers.BatchNormalization()(x)
encoded = tf.keras.layers.Conv1D(4, kernel_size=5, activation='relu', padding='same')(x)

x = tf.keras.layers.Conv1D(8, kernel_size=5, activation='relu', padding='same')(encoded)
x = tf.keras.layers.BatchNormalization()(x)

decoded = tf.keras.layers.Conv1D(1, kernel_size=5, activation='relu', padding='same')(x)
model = tf.keras.Model(input_layer, decoded)
model.summary()

Model: "model_1"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 input_2 (InputLayer)        [(None, 8192, 1)]         0         
                                                                 
 conv1d_4 (Conv1D)           (None, 8192, 8)           48        
                                                                 
 batch_normalization_2 (Batc  (None, 8192, 8)          32        
 hNormalization)                                                 
                                                                 
 conv1d_5 (Conv1D)           (None, 8192, 4)           164       
                                                                 
 conv1d_6 (Conv1D)           (None, 8192, 8)           168       
                                                                 
 batch_normalization_3 (Batc  (None, 8192, 8)          32        
 hNormalization)                                                 
                                                                 
 conv1d_7 (Conv1D)           (None, 8192, 1)           41        
                                                                 
=================================================================
Total params: 485
Trainable params: 453
Non-trainable params: 32
_________________________________________________________________