Please initialize `TimeDistributed` layer with a `tf.keras.layers.Layer` instance. You passed: (input)

Package:

keras

52268

Exception Class:

ValueError

Raise code

""" ses:
    ValueError: If not initialized with a `tf.keras.layers.Layer` instance.
  """

  def __init__(self, layer, **kwargs):
    if not isinstance(layer, Layer):
      raise ValueError(
          'Please initialize `TimeDistributed` layer with a '
          '`tf.keras.layers.Layer` instance. You passed: {input}'.format(
              input=layer))
    super(TimeDistributed, self).__init__(layer, **kwargs)
    self.supports_masking = True

    # It is safe to use the fast, reshape-based approach with all of our
    #

Links to the raise (1)

https://github.com/keras-team/keras/blob/4a978914d2298db2c79baa4012af5ceff4a4e203/keras/layers/wrappers.py#L120

Ways to fix

Summary:

This exception arises when creating an instance of TimeDistributed. The init function takes in a required Layer object. A check is performed to assert that the passed object is indeed a layer object. The exception occurs when calling TimeDistributed from keras.layers.wrappers. The best solution for getting it to work is to use tf.keras.layer.TimeDistributed. Using it will give the expected results without the errors.

Code to Reproduce the Error (Wrong):

from keras.layers.wrappers import TimeDistributed
import tensorflow as tf

conv_2d_layer = tf.keras.layers.Conv2D(64, (3, 3))
output = TimeDistributed(conv_2d_layer)

Error Message:

ValueError                                Traceback (most recent call last)
<ipython-input-59-1f812a7853d4> in <module>()
      3 
      4 conv_2d_layer = tf.keras.layers.Conv2D(64, (3, 3))
----> 5 output = TimeDistributed(conv_2d_layer)

/usr/local/lib/python3.7/dist-packages/keras/layers/wrappers.py in __init__(self, layer, **kwargs)
    121           'Please initialize `TimeDistributed` layer with a '
    122           '`tf.keras.layers.Layer` instance. You passed: {input}'.format(
--> 123               input=layer))
    124     super(TimeDistributed, self).__init__(layer, **kwargs)
    125     self.supports_masking = True

ValueError: Please initialize `TimeDistributed` layer with a `tf.keras.layers.Layer` instance. You passed: <tensorflow.python.keras.layers.convolutional.Conv2D object at 0x7fbe43dceb50>

Working Version (Right):

from keras.layers.wrappers import TimeDistributed
import tensorflow as tf

conv_2d_layer = tf.keras.layers.Conv2D(64, (3, 3))
outputs = tf.keras.layers.TimeDistributed(conv_2d_layer)