op {
  graph_op_name: "SparseSegmentSum"
  in_arg {
    name: "indices"
    description: <<END
A 1-D tensor. Has same rank as `segment_ids`.
END
  }
  in_arg {
    name: "segment_ids"
    description: <<END
A 1-D tensor. Values should be sorted and can be repeated.
END
  }
  out_arg {
    name: "output"
    description: <<END
Has same shape as data, except for dimension 0 which
has size `k`, the number of segments.
END
  }
  summary: "Computes the sum along sparse segments of a tensor."
  description: <<END
Read
[the section on segmentation](https://tensorflow.org/api_guides/python/math_ops#Segmentation)
for an explanation of segments.

Like `SegmentSum`, but `segment_ids` can have rank less than `data`'s first
dimension, selecting a subset of dimension 0, specified by `indices`.

For example:

```python
c = tf.constant([[1,2,3,4], [-1,-2,-3,-4], [5,6,7,8]])

# Select two rows, one segment.
tf.sparse_segment_sum(c, tf.constant([0, 1]), tf.constant([0, 0]))
# => [[0 0 0 0]]

# Select two rows, two segment.
tf.sparse_segment_sum(c, tf.constant([0, 1]), tf.constant([0, 1]))
# => [[ 1  2  3  4]
#     [-1 -2 -3 -4]]

# Select all rows, two segments.
tf.sparse_segment_sum(c, tf.constant([0, 1, 2]), tf.constant([0, 0, 1]))
# => [[0 0 0 0]
#     [5 6 7 8]]

# Which is equivalent to:
tf.segment_sum(c, tf.constant([0, 0, 1]))
```
END
}