Multinomial#

  • Domain: ai.onnx

  • Since version: 22

Generate a tensor of samples from a multinomial distribution according to the probabilities of each of the possible outcomes.

Inputs

  • input (T1): Input tensor with shape [batch_size, class_size], where class_size is the number of all possible outcomes. Each value along the axis zero represents the unnormalized log-probability of each corresponding outcome in a batch.

Outputs

  • output (T2): Output tensor with shape [batch_size, sample_size], where sample_size is the number of times to sample. Each value along the axis zero represents the outcome of the corresponding sample in a batch.

Attributes

  • dtype (int): (Optional) The data type for the elements of the output tensor, if not specified, we will use int32.

  • sample_size (int): Number of times to sample.

  • seed (float): (Optional) Seed to the random generator, if not specified we will auto generate one.

Type Constraints

  • T1: Constrain input types to float tensors. Allowed types: tensor(bfloat16), tensor(double), tensor(float), tensor(float16).

  • T2: Constrain output types to integral tensors. Allowed types: tensor(int32), tensor(int64).

Examples#

test_cc_multinomial

Node:
  Multinomial(x) -> (y)

Inputs:
  x: shape=(2, 3), dtype=float32
    [[ 0.,  0.,  0.],
     [-5., -5.,  5.]]

Outputs:
  y: shape=(2, 1), dtype=int32
    [[1],
     [2]]

test_cc_multinomial_int64

Node:
  Multinomial(x) -> (y)
  Attributes:
    sample_size = 4
    dtype = 7

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

Outputs:
  y: shape=(1, 4), dtype=int64
    [[1, 2, 2, 2]]

test_cc_multinomial_seeded

Node:
  Multinomial(x) -> (y)
  Attributes:
    sample_size = 5
    seed = 42.0

Inputs:
  x: shape=(2, 4), dtype=float32
    [[1.    , 2.    , 3.    , 4.    ],
     [0.5   , 0.25  , 0.125 , 0.0625]]

Outputs:
  y: shape=(2, 5), dtype=int32
    [[3, 2, 3, 3, 3],
     [0, 1, 1, 0, 2]]

Differences with previous version (7)#

SchemaDiff: Multinomial (domain 'ai.onnx')

  • old version: 7

  • new version: 22

  • breaking: no

Type constraints:

  • changed ‘T1’: added types: [‘tensor(bfloat16)’]

Version History#