TensorFlow Error: Dimensions must be equal but are 6558 and 6562 for '{{node mean_squared_error/SquaredDifference}}

Question

I am working on a 1D Convolutional Neural Network for the classification of malware types. I am currently running into a problem that I cannot seem to figure out.

Here is the need to know info on the error:

    File "C:\Users\Admin\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.9_qbz5n2kfra8p0\LocalCache\local-packages\Python39\site-packages\keras\losses.py", line 1329, in mean_squared_error
        return backend.mean(tf.math.squared_difference(y_pred, y_true), axis=-1)

    ValueError: Dimensions must be equal, but are 6561 and 6562 for '{{node mean_squared_error/SquaredDifference}} = SquaredDifference[T=DT_FLOAT](sequential/output_layer/BiasAdd, IteratorGetNext:1)' with input shapes: [1,6561,13], [6562,1].

I will put my code following that:

X, _ = d.get_encoded_data()
y = d.get_y()
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=1)

LAYERS = 52 # arbitrary number for now
EPOCHS = 100 # arbitrary

# TODO: define all the testing data

def getModel():
    # all three layers are arbitrarily chosen, don't know input shape
    return keras.Sequential(
        [   
            layers.Conv1D(32, 4, padding='same', activation='relu', name="layer1", input_shape=X_train.shape[1:]),
            layers.Dropout(.3),
            layers.Conv1D(32, 2, activation='relu'),
            layers.Dense(LAYERS / 4, activation='relu', name='layer3'),
            layers.Dropout(.3),
            layers.Dense(13, name='output_layer')
        ])

def plot(history, epochs, metric, type):
    loss_train = history.history[metric]
    epochs = range(1,epochs)
    plt.plot(epochs, loss_train, 'g', label=metric)
    plt.title(type, metric)
    plt.xlabel('Epochs')
    plt.ylabel(metric)
    plt.legend()
    plt.show()

if __name__ == '__main__':
    X_train = X_train.reshape(1, 6562, 50)
    y_train = y_train.reshape(1, 6562, 1)
    
    X_train = X_train.astype(np.float32)
    y_train = y_train.astype(np.float32)
    
    train_dataset = tf.data.Dataset.from_tensor_slices((X_train, y_train))
    
    model = getModel()
    optimizer = keras.optimizers.Adam(learning_rate=0.001)
   
    model.compile(optimizer, 'mse', metrics=[keras.metrics.Accuracy(),
                                                keras.metrics.Precision(), 
                                                keras.metrics.Recall()])

    model.build()
    model.summary()
    history = model.fit(
        train_dataset,
        epochs=EPOCHS
    )
   
    for metric in history.history:
        plot(history, EPOCHS, metric, "Training")
   
    print("Evaluate on test data")
    results = model.evaluate(X_test, y_test)
    print("test loss, test acc:", results)

The main issue here is in the getModel() function and the fit function. My input sizes are (6562, 50) for the X_train and (6562,) for the y_train. Why am I getting this error?

For your info, d.get_encoded_data gets the data as it says, but the _ represents the y. I want y to be one hot encoded tho, which is the line following that. I can't seem to find good tutorials on Conv1D that help at all. Any help is greatly appreciated.