I try to convert the variable type to Tensor("strided_slice_1:0", shape=(24, 24, 3), dtype=float32) to a NumPy matrix.
image=images_train[1]
with tf.Session() as sess:
array1 = sess.run(image)
But the program is not executed after this.
Add a comment
|
1 Answer
Eager Execution is enabled by default in TF 2.0, so you need to call .numpy() on the Tensor object.
import tensorflow as tf
import numpy as np
print(tf.__version__)
#tf.executing_eagerly()
def tensor_to_array(array_value):
return array_value.numpy()
a = tf.constant([5.0,6.0,7.0])
b = tf.constant([6.0,8.0,9.0])
c = a * b
print(c)
print(tensor_to_array(c))
print(type(tensor_to_array(c)))
Output:
2.2.0
tf.Tensor([30. 48. 63.], shape=(3,), dtype=float32)
[30. 48. 63.]
<class 'numpy.ndarray'>
If Eager Execution is disabled, you can build a graph and then run it through tf.compat.v1.Session in TF 2.x and tf.Session in TF 1.x
%tensorflow_version 1.x
import tensorflow as tf
print(tf.__version__)
with tf.Session() as sess:
a = tf.constant([5.0,6.0,7.0])
b = tf.constant([6.0,8.0,9.0])
c = a * b
print(c)
print(sess.run(c))
print(type(sess.run(c)))
Output:
1.15.2
Tensor("mul_1:0", shape=(3,), dtype=float32)
[30. 48. 63.]
<class 'numpy.ndarray'>