Python – tensorflow.ensure_shape()

TensorFlow is open-source Python library designed by Google to develop Machine Learning models and deep learning  neural networks. 

ensure_shape() is used to update and check the shape of Tensor.

Syntax: tensorflow.ensure_shape( x, shape, name)

Parameters:

• x: It is input Tensor.
• shape: It is TensorShape which represents the shape of input Tensor.
• name(optional): It defines the name for the operation.

Returns:

It returns a Tensor same as x or raise  tf.errors.InvalidArgumentError if shapes are incompatible.

Example 1:

Python3

# Importing the library
import tensorflow as tf

# Initializing the input
x = tf.constant([[2, 3, 6], [4, 8, 15]])

# Printing the input
print('x:', x)

# Calculating result
res = tf.ensure_shape(x, (2, 3))

# Printing the result
print('res: ', res)

Output:

x: tf.Tensor(
[[ 2  3  6]
 [ 4  8 15]], shape=(2, 3), dtype=int32)
res:  tf.Tensor(
[[ 2  3  6]
 [ 4  8 15]], shape=(2, 3), dtype=int32)

Example 2: In this example shape is incompatible with shape of x so error is raised.

Python3

# Importing the library
import tensorflow as tf

# Initializing the input
x = tf.constant([[2, 3, 6], [4, 8, 15]])

# Printing the input
print('x:', x)

# Calculating result
res = tf.ensure_shape(x, (2, 4))

# Printing the result
print('res: ', res)

Output:

x: tf.Tensor(
[[ 2  3  6]
 [ 4  8 15]], shape=(2, 3), dtype=int32)

---------------------------------------------------------------------------

InvalidArgumentError                      Traceback (most recent call last)

<ipython-input-1-ab1be364fadb> in <module>()
      9 
     10 # Calculating result
---> 11 res = tf.ensure_shape(x, (2, 4))
     12 
     13 # Printing the result

3 frames

/usr/local/lib/python3.6/dist-packages/six.py in raise_from(value, from_value)

InvalidArgumentError: Shape of tensor dummy_input [2, 3] is not compatible with expected shape [2, 4]. [Op:EnsureShape]