# Python – tensorflow.math.top_k()

• Last Updated : 16 Jun, 2020

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

top_k()  is used to find top k largest entries for the last dimension(along each row for matrices).

Syntax:  tensorflow.math.top_k(input, k, sorted, name)

Parameter:

• input: It’s the input Tensor with 1 or more dimensions.
• k(optional): It’s is 0-D tensor with default value 0.
• sorted(optional): If it’s set to true returned elements will be sorted. Default is True.
• name(optional): It defines the name for the operation.

Returns:

• values:  k largest elements along each last dimensional slice.
• indices: indices of values within the last dimension of input.

Example 1:

## Python3

 `# importing the library ` `import` `tensorflow as tf ` ` `  `# Initializing the input tensor ` `a ``=` `tf.constant([``7``, ``2``, ``3``, ``9``, ``5``], dtype ``=` `tf.float64) ` ` `  `# Printing the input tensor ` `print``(``'a: '``, a) ` ` `  `# Calculating result ` `res ``=` `tf.math.top_k(a) ` ` `  `# Printing the result ` `print``(``'Result: '``, res) `

Output:

```a:  tf.Tensor([7. 2. 3. 9. 5.], shape=(5, ), dtype=float64)
Result:  TopKV2(values=<tf.Tensor: shape=(1, ), dtype=float64, numpy=array([9.])>,
indices=<tf.Tensor: shape=(1, ), dtype=int32, numpy=array([3], dtype=int32)>)

```

Example 2:

## Python3

 `# importing the library ` `import` `tensorflow as tf ` ` `  `# Initializing the input tensor ` `a ``=` `tf.constant([[``7``, ``2``, ``3``], [ ``9``, ``5``, ``7``]], dtype ``=` `tf.float64) ` ` `  `# Printing the input tensor ` `print``(``'a: '``, a) ` ` `  `# Calculating result ` `res ``=` `tf.math.top_k(a, k ``=` `2``) ` ` `  `# Printing the result ` `print``(``'Result: '``, res) `

Output:

```a:  tf.Tensor(
[[7. 2. 3.]
[9. 5. 7.]], shape=(2, 3), dtype=float64)
Result:  TopKV2(values=<tf.Tensor: shape=(2, 2), dtype=float64, numpy=
array([[7., 3.],
[9., 7.]])>, indices=<tf.Tensor: shape=(2, 2), dtype=int32, numpy=
array([[0, 2],
[0, 2]], dtype=int32)>)
```

My Personal Notes arrow_drop_up
Related Articles