tf.concat(
values,
axis, # concat_dim
name='concat'
)
axis = 0 means concatenating in 1st dim.
axis = 1 means concatenating in 2nd dim.
axis = n-1 means concatenating in n th dim.
import tensorflow as tf
t1 = [[1, 2, 3], [4, 5, 6]]
t2 = [[7, 8, 9], [10, 11, 12]]
sess = tf.Session()
a1 = tf.concat([t1, t2], 0) # concatenate in 1st dim, row
a2 = tf.concat([t1, t2], 1) # concatenate in 2nd dim, column
a3 = tf.concat([t1, t2], 2) # concatenate in 3rd dim, which does not exist.
print(sess.run(c1))
print("\n")
print(sess.run(c2))
[[ 1 2 3]
[ 4 5 6]
[ 7 8 9]
[10 11 12]]
[[ 1 2 3 7 8 9]
[ 4 5 6 10 11 12]]
a3 results in Error as expected here:
ValueError: Shape must be at least rank 3 but is rank 2 for 'concat_14' (op: 'ConcatV2') with input shapes: [2,3], [2,3], [].
Similarly, numpy.concatenate() results the same as tf.concat()
import numpy as np
t1 = [[1, 2, 3], [4, 5, 6]]
t2 = [[7, 8, 9], [10, 11, 12]]
b1 = np.concatenate((t1, t2), 0)
b2 = np.concatenate((t1, t2), 1)
print(b1)
print("\n")
print(b2)
[[ 1 2 3]
[ 4 5 6]
[ 7 8 9]
[10 11 12]]
[[ 1 2 3 7 8 9]
[ 4 5 6 10 11 12]]
