How do I turn a numpy array into a tensor in "Tensorflow"?

Question

I tried to test some learning networks after I completed training with a tensorflow.

But my test image is [512 512 1] data of channel 1 in 512 horizontal and 512 vertical pixels.

I changed the image data to a numpy array.

The tensor network should be [? 512 512 1] It looks like this.

How do I convert a numpy array to a tensor? ([512 512 1] -> [? 512 512 1])

A message says ... ValueError: Cannot feed value of shape (512, 512, 1) for Tensor 'images:0', which has shape '(?, 512, 512, 1)' — ddjfjfj djfiejdn
– ddjfjfj djfiejdn, Commented Nov 9, 2018 at 9:38

Lau · Accepted Answer · 2018-11-09 09:59:22Z

4

You just have to append one dimension

arr = your_image # [512, 512, 1]
new_arr = np.expand_dims(arr, 0)

tensor = tf.convert_to_tensor(new_arr)

Now you can use feed dict or something else.

answered Nov 9, 2018 at 9:59

Lau

1,48511 silver badges33 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

thsnks you very very much!