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

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

See the guide: Math > Reduction

Returns the element-wise sum of a list of tensors.

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

NOTE: This operation is not differentiable and cannot be used if inputs depend on trainable variables. Please use tf.add_n for such cases.

For example:

# tensor 'a' is [[1, 2], [3, 4]]
# tensor `b` is [[5, 0], [0, 6]]
tf.accumulate_n([a, b, a]) ==> [[7, 4], [6, 14]]

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


  • inputs: A list of Tensor objects, each with same shape and type.
  • shape: Shape of elements of inputs.
  • tensor_dtype: The type of inputs.
  • name: A name for the operation (optional).


A Tensor of same shape and type as the elements of inputs.


  • ValueError: If inputs don't all have same shape and dtype or the shape cannot be inferred.

Defined in tensorflow/python/ops/