CODEWITHSUNDEEP
tensorflow
python
python

Reputation: 55

tf.reduce_mean prints strange results when reduction_indices is specified

tf.reduce_mean command is used to take mean along specified axis. But if don't specify axis and use reduction indices zero then it prints middle row. Why is it so? How this reduction indices effects output values? If I print sum.eval() command why it prints [[4,5,6]]? I also go through the documentation but it doesn't give any explanation for reduction_indices.Please clear my concept about this command.

a = tf.range(0,12)
b = tf.reshape(a,(4,3))
sum = tf.reduce_mean(b,reduction_indices=[0],keep_dims=True)   
sess = tf.InteractiveSession()
print (b.eval())
print (b.get_shape())
print (sum.eval())
print (sum.get_shape())

Upvotes: 0

Views: 334

Answers (1)

f4.
f4.

Reputation: 3852

reduction_indices behaves like axis.

As per the documentation of tf.reduce_mean:

reduction_indices: The old (deprecated) name for axis.

https://www.tensorflow.org/api_docs/python/tf/reduce_mean

4, 5 and 6 are just the mean of each column

 (0+3+6+9)/4 = 4
(1+4+7+10)/4 = 5
(2+5+8+11)/4 = 6

Upvotes: 0

Related Questions