How do you create and manipulate Tensors in TensorFlow? — TensorFlow Interview Question #6

In TensorFlow, tensors are the fundamental building blocks of a computational graph. They are used to represent the data that flows through the graph during computation. Here’s how to create and manipulate tensors in TensorFlow:

Creating Tensors: You can create tensors in TensorFlow using various methods, such as tf.constant(), tf.Variable(), and tf.placeholder(). tf.constant() creates a tensor with a fixed value that cannot be changed. For example, the following code creates a constant tensor with the value [1, 2, 3]:

    import tensorflow as tf
    
    tensor_const = tf.constant([1, 2, 3])

tf.Variable() creates a tensor with an initial value that can be changed during computation. For example, the following code creates a variable tensor with the value [1, 2, 3]:

    tensor_var = tf.Variable([1, 2, 3])

tf.placeholder() creates a tensor that acts as a placeholder for data that will be fed into the graph during computation. For example, the following code creates a placeholder tensor with the shape [None, 3], where None represents a variable batch size:

    tensor_placeholder = tf.placeholder(tf.float32, shape=[None, 3])

Manipulating Tensors: You can manipulate tensors in TensorFlow using various operations, such as tf.add(), tf.multiply(), and tf.reshape(). tf.add() performs element-wise addition on two tensors. For example, the following code adds two tensors element-wise:

    tensor_a = tf.constant([1, 2, 3])
    tensor_b = tf.constant([4, 5, 6])
    tensor_add = tf.add(tensor_a, tensor_b)  # [5, 7, 9]

tf.multiply() performs element-wise multiplication on two tensors. For example, the following code multiplies two tensors element-wise:

    tensor_c = tf.constant([2, 4, 6])
    tensor_d = tf.constant([3, 5, 7])
    tensor_mul = tf.multiply(tensor_c, tensor_d)  # [6, 20, 42]

tf.reshape() changes the shape of a tensor. For example, the following code reshapes a tensor from shape [2, 3] to shape [3, 2]:

    tensor_e = tf.constant([[1, 2, 3], [4, 5, 6]])
    tensor_reshape = tf.reshape(tensor_e, [3, 2])  # [[1, 2], [3, 4], [5, 6]]

These are just a few examples of how to create and manipulate tensors in TensorFlow. The TensorFlow API provides a wide range of operations and functions for working with tensors, making it a powerful tool for building and training machine learning models.