Convert TensorFlow Tensor data type in Python

Python TensorFlow Basic: Exercise-15 with Solution

Write a Python program that converts a TensorFlow tensor from one data type to another (e.g., from int32 to float64).

Sample Solution:

Python Code:

import tensorflow as tf

# Create a TensorFlow tensor with int32 data type
# Tensors are multi-dimensional arrays with a uniform type (called a dtype ).
tensor_int32 = tf.constant([100, 200, 300], dtype=tf.int32)

# Convert the tensor to float64 data type
tensor_float64 = tf.cast(tensor_int32, dtype=tf.float64)

# Print the original and converted tensors
print("Original Tensor (int32):", tensor_int32.numpy())
print("Converted Tensor (float64):", tensor_float64.numpy())

Output:

Original Tensor (int32): [100 200 300]
Converted Tensor (float64): [100. 200. 300.]

Explanation:

In the exercise above -

• Import TensorFlow as tf.
• Create a TensorFlow tensor tensor_int32 with data type int32 using tf.constant().
• Use tf.cast() to convert the tensor_int32 tensor to the float64 data type, storing the result in tensor_float64.
• Finally, we print both the original and the converted tensor.

Python Code Editor:

Previous: Specified TensorFlow Data Type in Python.
Next: Create and print TensorFlow string tensor in Python.