Mean#

  • Domain: ai.onnx

  • Since version: 13

Element-wise mean of each of the input tensors (with Numpy-style broadcasting support). All inputs and outputs must have the same data type.

This operator supports multidirectional (i.e., Numpy-style) broadcasting; for more details please check the doc.

Inputs

  • data_0 (T): List of tensors for mean.

Outputs

  • mean (T): Output tensor.

Type Constraints

  • T: Constrain input and output types to float tensors. Allowed types: tensor(bfloat16), tensor(double), tensor(float), tensor(float16).

Examples#

test_cc_mean_bcast

Node:
  Mean(data_0, data_1) -> (result)

Inputs:
  data_0: shape=(2, 2), dtype=float32
    [[1., 2.],
     [3., 4.]]
  data_1: shape=(), dtype=float32
    10.

Outputs:
  result: shape=(2, 2), dtype=float32
    [[5.5, 6. ],
     [6.5, 7. ]]

test_mean_example

Node:
  Mean(data_0, data_1, data_2) -> (result)

Inputs:
  data_0: shape=(3,), dtype=float32
    [3., 0., 2.]
  data_1: shape=(3,), dtype=float32
    [1., 3., 4.]
  data_2: shape=(3,), dtype=float32
    [2., 6., 6.]

Outputs:
  result: shape=(3,), dtype=float32
    [2., 3., 4.]

test_mean_one_input

Node:
  Mean(data_0) -> (result)

Inputs:
  data_0: shape=(3,), dtype=float32
    [3., 0., 2.]

Outputs:
  result: shape=(3,), dtype=float32
    [3., 0., 2.]

test_mean_two_inputs

Node:
  Mean(data_0, data_1) -> (result)

Inputs:
  data_0: shape=(3,), dtype=float32
    [3., 0., 2.]
  data_1: shape=(3,), dtype=float32
    [1., 3., 4.]

Outputs:
  result: shape=(3,), dtype=float32
    [2. , 1.5, 3. ]

Differences with previous version (8)#

SchemaDiff: Mean (domain 'ai.onnx')

  • old version: 8

  • new version: 13

  • breaking: no

Type constraints:

  • changed ‘T’: added types: [‘tensor(bfloat16)’]

Version History#