tf.math.accumulate_n

Returns the element-wise sum of a list of tensors. (deprecated)

tf.math.accumulate_n(
    inputs, shape=None, tensor_dtype=None, name=None
)

Optionally, pass shape and tensor_dtype for shape and type checking, otherwise, these are inferred.

For example:

a = tf.constant([[1, 2], [3, 4]])
b = tf.constant([[5, 0], [0, 6]])
tf.math.accumulate_n([a, b, a]).numpy()
array([[ 7, 4],
       [ 6, 14]], dtype=int32)

# Explicitly pass shape and type
tf.math.accumulate_n(
    [a, b, a], shape=[2, 2], tensor_dtype=tf.int32).numpy()
array([[ 7,  4],
       [ 6, 14]], dtype=int32)

Note: The input must be a list or tuple. This function does not handle IndexedSlices

See Also:

• tf.reduce_sum(inputs, axis=0) - This performe the same mathematical operation, but tf.add_n may be more efficient because it sums the tensors directly. reduce_sum on the other hand calls tf.convert_to_tensor on the list of tensors, unncessairly stacking them into a single tensor before summing.
• tf.add_n - This is another python wrapper for the same Op. It has nearly identical functionality.