How to convert a tensor into a numpy array when using Tensorflow with Python bindings?

Any tensor returned by Session.run or eval is a NumPy array.


>>>print(type(tf.Session().run(tf.constant([1,2,3]))))
This post was edited by Raji Reddy A at June 11, 2019 4:30 PM IST

To convert back from tensor to numpy array you can simply run .eval() on the transformed tensor.

You need to:

encode the image tensor in some format (jpeg, png) to binary tensor
evaluate (run) the binary tensor in a session
turn the binary to stream
feed to PIL image
(optional) displaythe image with matplotlib
Code:

import tensorflow as tf
import matplotlib.pyplot as plt
import PIL

...

image_tensor =
jpeg_bin_tensor = tf.image.encode_jpeg(image_tensor)

with tf.Session() as sess:
# display encoded back to image data
jpeg_bin = sess.run(jpeg_bin_tensor)
jpeg_str = StringIO.StringIO(jpeg_bin)
jpeg_image = PIL.Image.open(jpeg_str)
plt.imshow(jpeg_image)
This worked for me. You can try it in a ipython notebook. Just don't forget to add the following line:

%matplotlib inline

The Eager Execution of the TensorFlow library can be used to convert a tensor to a NumPy array in Python. With Eager Execution, the behavior of the operations of TensorFlow library changes, and the operations execute immediately. We can also perform NumPy operations on Tensor objects with Eager Execution. The Tensor.numpy() function converts the Tensor to a NumPy array in Python. In TensorFlow 2.0, the Eager Execution is enabled by default. So, this approach works best for the TensorFlow version 2.0. See the following code example.

import tensorflow as tf
tensor = tf.constant([[1,2,3],[4,5,6],[7,8,9]])
print("Tensor = ",tensor)
array = tensor.numpy()
print("Array = ",array)​