Home > Software engineering >  How to remove list of elements from a Tensorflow tensor
How to remove list of elements from a Tensorflow tensor

Time:12-21

For the following tensor:

 <tf.Tensor: shape=(2, 10, 6), dtype=int64, numpy=
 array([[[  3,  16,  43,  10,   7, 431],
         [  3,   2,   6,   5,   7,   2],
         [  3,  37,   5,   7,   2,  12],
         [  3,   2,  11,   5,   7,   2],
         [  3,   2,   6,  18,  14, 195],
         [  3,   2,   6,   5,   7, 195],
         [  3,   2,   6,   5,   7,   9],
         [  3,   2,  11,   7,   2,  12],
         [  3,  16,  52,  92, 177, 923],
         [  3,   9,  43,  10,   7,   9]],
 
        [[  3,   2,  22, 495, 230,   4],
         [  3,   2,  22,   5, 102, 122],
         [  3,   2,  22,   5, 102, 230],
         [  3,   2,  22,   5,  70, 908],
         [  3,   2,  22,   5,  70, 450],
         [  3,   2,  22,   5,  70, 122],
         [  3,   2,  22,   5,  70, 122],
         [  3,   2,  22,   5,  70, 230],
         [  3,   2,  22,  70,  34, 470],
         [  3,   2,  22, 855, 450,   4]]], dtype=int64)>)

I want to remove the last list [ 3, 2, 22, 855, 450, 4] in the tensor. I tried with tf.unstack but it didn't work.

CodePudding user response:

you can try below to remove the last list from the tensor:

sliced_tensor = tf.slice(tensor, [0, 0, 0], [2, 9, 6]) 

CodePudding user response:

try this instead

new_tensor = tf.slice(tensor, [0,0,0], [2,9,6], [1,1,1])
  • Related