Max#

Domain: ai.onnx
Since version: 13

Element-wise max 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 max.

Outputs

max (T): Output tensor.

Type Constraints

T: Constrain input and output types to numeric tensors. Allowed types: tensor(bfloat16), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8).

Examples#

test_cc_max_bcast

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

Inputs:
  data_0: shape=(2, 2), dtype=float32
    [[1., 2.],
     [3., 4.]]
  data_1: shape=(), dtype=float32
    2.5

Outputs:
  result: shape=(2, 2), dtype=float32
    [[2.5, 2.5],
     [3. , 4. ]]

test_cc_shape_inference_resize_tile

Node:
  Max(tile_out, zeros_scalar) -> (output)

Inputs:
  tile_out: shape=(10, 6), dtype=float32
    [[ 1.,  2.,  3.,  4.,  5.,  6.],
     [ 7.,  8.,  9., 10., 11., 12.],
     [13., 14., 15., 16., 17., 18.],
     [19., 20., 21., 22., 23., 24.],
     [25., 26., 27., 28., 29., 30.],
     [31., 32., 33., 34., 35., 36.],
     [37., 38., 39., 40., 41., 42.],
     [43., 44., 45., 46., 47., 48.],
     [49., 50., 51., 52., 53., 54.],
     [55., 56., 57., 58., 59., 60.]]

Outputs:
  output: shape=(10, 6), dtype=float32
    [[ 1.,  3.,  5.,  1.,  3.,  5.],
     [13., 15., 17., 13., 15., 17.],
     [25., 27., 29., 25., 27., 29.],
     [37., 39., 41., 37., 39., 41.],
     [49., 51., 53., 49., 51., 53.],
     [ 1.,  3.,  5.,  1.,  3.,  5.],
     [13., 15., 17., 13., 15., 17.],
     [25., 27., 29., 25., 27., 29.],
     [37., 39., 41., 37., 39., 41.],
     [49., 51., 53., 49., 51., 53.]]

test_max_example

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

Inputs:
  data_0: shape=(3,), dtype=float32
    [3., 2., 1.]
  data_1: shape=(3,), dtype=float32
    [1., 4., 4.]
  data_2: shape=(3,), dtype=float32
    [2., 5., 3.]

Outputs:
  result: shape=(3,), dtype=float32
    [3., 5., 4.]

test_max_float16

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

Inputs:
  data_0: shape=(3,), dtype=float16
    [1., 4., 3.]
  data_1: shape=(3,), dtype=float16
    [3., 2., 5.]

Outputs:
  result: shape=(3,), dtype=float16
    [3., 4., 5.]

test_max_float32

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

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

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

test_max_float64

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

Inputs:
  data_0: shape=(3,), dtype=float64
    [3., 2., 1.]
  data_1: shape=(3,), dtype=float64
    [1., 4., 4.]

Outputs:
  result: shape=(3,), dtype=float64
    [3., 4., 4.]

test_max_int16

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

Inputs:
  data_0: shape=(3,), dtype=int16
    [3, 2, 1]
  data_1: shape=(3,), dtype=int16
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=int16
    [3, 4, 4]

test_max_int32

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

Inputs:
  data_0: shape=(3,), dtype=int32
    [3, 2, 1]
  data_1: shape=(3,), dtype=int32
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=int32
    [3, 4, 4]

test_max_int64

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

Inputs:
  data_0: shape=(3,), dtype=int64
    [3, 2, 1]
  data_1: shape=(3,), dtype=int64
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=int64
    [3, 4, 4]

test_max_int8

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

Inputs:
  data_0: shape=(3,), dtype=int8
    [3, 2, 1]
  data_1: shape=(3,), dtype=int8
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=int8
    [3, 4, 4]

test_max_one_input

Node:
  Max(data_0) -> (result)

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

Outputs:
  result: shape=(3,), dtype=float32
    [3., 2., 1.]

test_max_two_inputs

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

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

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

test_max_uint16

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

Inputs:
  data_0: shape=(3,), dtype=uint16
    [3, 2, 1]
  data_1: shape=(3,), dtype=uint16
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=uint16
    [3, 4, 4]

test_max_uint32

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

Inputs:
  data_0: shape=(3,), dtype=uint32
    [3, 2, 1]
  data_1: shape=(3,), dtype=uint32
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=uint32
    [3, 4, 4]

test_max_uint64

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

Inputs:
  data_0: shape=(3,), dtype=uint64
    [3, 2, 1]
  data_1: shape=(3,), dtype=uint64
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=uint64
    [3, 4, 4]

test_max_uint8

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

Inputs:
  data_0: shape=(3,), dtype=uint8
    [3, 2, 1]
  data_1: shape=(3,), dtype=uint8
    [1, 4, 4]

Outputs:
  result: shape=(3,), dtype=uint8
    [3, 4, 4]

Differences with previous version (12)#

SchemaDiff: Max (domain 'ai.onnx')

old version: 12
new version: 13
breaking: no

Type constraints:

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

Max#

Examples#

Differences with previous version (12)#

Version History#