Common TensorFlow data types in Python

Python TensorFlow Basic: Exercise-13 with Solution

Write a Python code to list some common TensorFlow data types available.

Sample Solution:

Python Code:

import tensorflow as tf

# List of common TensorFlow data types
common_data_types = [
    tf.bfloat16,    # 16-bit bfloat (brain floating point).
    tf.bool,        # Boolean. 
    tf.complex128,  # 128-bit complex.
    tf.complex64,   # 64-bit complex.
    tf.double,      # 64-bit (double precision) floating-point.
    tf.float16,     # 16-bit (half precision) floating-point.
    tf.float32,     # 32-bit (single precision) floating-point.
    tf.float64,     # 64-bit (double precision) floating-point.
    tf.half,        # 16-bit (half precision) floating-point.
    tf.int8,        # Signed 8-bit integer.
    tf.int16,       # Signed 16-bit integer.
    tf.int32,       # Signed 32-bit integer.
    tf.int64,       # Signed 64-bit integer.
    tf.qint32,      # Signed quantized 32-bit integer.
    tf.qint8,       # Signed quantized 8-bit integer.
    tf.quint16,     # Unsigned quantized 16-bit integer.
    tf.quint8,      # Unsigned quantized 8-bit integer.
    tf.resource,    # Handle to a mutable, dynamically allocated resource.
    tf.uint8,       # Unsigned 8-bit (byte) integer.
    tf.uint16,      # Unsigned 16-bit (word) integer.
    tf.uint32,      # Unsigned 32-bit (dword) integer.
    tf.uint64,      # Unsigned 64-bit (qword) integer.
    tf.string,      # Variable-length string, represented as byte array.
    tf.variant      # Data of arbitrary type (known at runtime).  
]

# Print the data types and their names
for dtype in common_data_types:
    print(f"{dtype}: {tf.as_dtype(dtype).name}")

Output:

<dtype: 'bfloat16'>: bfloat16
<dtype: 'bool'>: bool
<dtype: 'complex128'>: complex128
<dtype: 'complex64'>: complex64
<dtype: 'float64'>: float64
<dtype: 'float16'>: float16
<dtype: 'float32'>: float32
<dtype: 'float64'>: float64
<dtype: 'float16'>: float16
<dtype: 'int8'>: int8
<dtype: 'int16'>: int16
<dtype: 'int32'>: int32
<dtype: 'int64'>: int64
<dtype: 'qint32'>: qint32
<dtype: 'qint8'>: qint8
<dtype: 'quint16'>: quint16
<dtype: 'quint8'>: quint8
<dtype: 'resource'>: resource
<dtype: 'uint8'>: uint8
<dtype: 'uint16'>: uint16
<dtype: 'uint32'>: uint32
<dtype: 'uint64'>: uint64
<dtype: 'string'>: string
<dtype: 'variant'>: variant

Explanation:

In the exercise above -

• Import TensorFlow as tf.
• Create a list named common_data_types containing various common TensorFlow data types, such as floating-point types (e.g., tf.float32), integer types (e.g., tf.int64), unsigned integer types (e.g., tf.uint8), boolean (tf.bool), and string (tf.string) data types.
• Use a for loop to iterate through the list and print each data type along with its name using tf.as_dtype(dtype).name.

Python Code Editor:

Previous: Updating TensorFlow variables in Python.
Next: Specified TensorFlow Data Type in Python.