Tensorflow – Converting Tensors to Numpy Arrays

You can convert a tensor to a numpy array with ease by using the tensor.numpy() method or the np.array() function.

Example:

import numpy as np
import tensorflow as tf

# Create demo tensors
ts1 = tf.constant([
                   [1, 2, 3],
                   [4, 5, 6]
], dtype=tf.int16)
ts2 = tf.constant([0.2, 1.1, 3.3, 4.4], dtype=tf.float32)

# Using numpy() method
arr1 = ts1.numpy()

# Using np.array() function
arr2 = np.array(ts2)

print(type(arr1))
print(type(arr2))

Output:

<class 'numpy.ndarray'>
<class 'numpy.ndarray'>

Hope this helps!