训练 CNN：ValueError：没有为任何变量提供梯度

Question

I want to train a model with the following layers:

embedding_dim = 80

model = tf.keras.Sequential()
model.add(tf.keras.layers.Conv2D(64, 3, padding='same', activation='relu', input_shape=(50, 120, 3)))
model.add(tf.keras.layers.MaxPool2D(padding='same'))

model.add(tf.keras.layers.Conv2D(128, 3, padding='same', activation='relu'))
model.add(tf.keras.layers.MaxPool2D(padding='same'))

model.add(tf.keras.layers.Conv2D(256, 3, padding='same', activation='relu'))

model.add(tf.keras.layers.Conv2D(256, 3, padding='same', activation='relu'))

model.add(tf.keras.layers.Conv2D(512, 3, padding='same', activation='relu'))

model.add(tf.keras.layers.Conv2D(512,  2, strides=(2, 4), activation='relu'))

model.add(tf.keras.layers.Conv2D(512,  3, activation='relu'))

model.add(tf.keras.layers.Lambda(add_timing_signal_nd))

model.add(tf.keras.layers.Flatten())
model.add(tf.keras.layers.Dense(embedding_dim))

After, I run

model.compile(optimizer='adam', loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True, reduction='none'), metrics=['accuracy'])

model.fit(image_dataset, epochs=10, validation_data=val_dataset)

I get the following error

ValueError: No gradients provided for any variable: ['conv2d/kernel:0', 'conv2d/bias:0', 'conv2d_1/kernel:0', 'conv2d_1/bias:0', 'conv2d_2/kernel:0', 'conv2d_2/bias:0', 'conv2d_3/kernel:0', 'conv2d_3/bias:0', 'conv2d_4/kernel:0', 'conv2d_4/bias:0', 'conv2d_5/kernel:0', 'conv2d_5/bias:0', 'conv2d_6/kernel:0', 'conv2d_6/bias:0', 'dense/kernel:0', 'dense/bias:0'].

For further context, add_timing_signa_nd is defined as the following

def add_timing_signal_nd(x, min_timescale=1.0, max_timescale=1.0e4):
    """
    Args:
        x: a Tensor with shape [batch, d1 ... dn, channels]
        min_timescale: a float
        max_timescale: a float
    Returns:
        a Tensor the same shape as x.
    """
    static_shape = x.get_shape().as_list()
    num_dims = len(static_shape) - 2
    channels = tf.shape(x)[-1]
    num_timescales = channels // (num_dims * 2)
    log_timescale_increment = (
            math.log(float(max_timescale) / float(min_timescale)) /
            (tf.cast(num_timescales, dtype=tf.float32) - 1))
    inv_timescales = min_timescale * tf.exp(
            tf.cast(tf.range(num_timescales), dtype=tf.float32) * -log_timescale_increment)
    for dim in xrange(num_dims):
        length = tf.shape(x)[dim + 1]
        position = tf.cast(tf.range(length), dtype=tf.float32)
        scaled_time = tf.expand_dims(position, 1) * tf.expand_dims(
                inv_timescales, 0)
        signal = tf.concat([tf.sin(scaled_time), tf.cos(scaled_time)], axis=1)
        prepad = dim * 2 * num_timescales
        postpad = channels - (dim + 1) * 2 * num_timescales
        signal = tf.pad(signal, [[0, 0], [prepad, postpad]])
        for _ in xrange(1 + dim):
            signal = tf.expand_dims(signal, 0)
        for _ in xrange(num_dims - 1 - dim):
            signal = tf.expand_dims(signal, -2)
        x += signal
    return x

If it helps by input sizes are the following

(3, 50, 120, 3)
(3, 50, 120, 3)
(3, 50, 120, 3)
(1, 50, 120, 3)

Also, I'm planning to extract the weights after training to use them in another problem.

Thanks in advance!

Answer 1

I guess the image_dataset or val_dataset is not proper. Follow your code, I simulated some data（include labels） to train and it could run properly.

image_dataset = np.random.uniform(0, 1, (3000, 50, 120, 3))
image_dataset_y = np.random.uniform(0, embedding_dim, (3000,)).astype(np.int)
val_dataset = np.random.uniform(0, 1, (300, 50, 120, 3))
val_dataset_y = np.random.uniform(0, embedding_dim, (300,)).astype(np.int)
model.fit(image_dataset, image_dataset_y , batch_size=30, epochs=10, validation_data=(val_dataset, val_dataset_y))