Sqrt#

  • Domain: ai.onnx

  • Since version: 13

Square root takes one input data (Tensor ) and produces one output data (Tensor ) where the square root is, y = x^0.5, is applied to the tensor elementwise. If x is negative, then it will return NaN.

Inputs

  • X (T): Input tensor

Outputs

  • Y (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_sqrt

Node:
  Sqrt(x) -> (y)

Inputs:
  x: shape=(2, 3), dtype=float32
    [[0.  , 0.25, 1.  ],
     [2.  , 4.  , 9.  ]]

Outputs:
  y: shape=(2, 3), dtype=float32
    [[0.       , 0.5      , 1.       ],
     [1.4142135, 2.       , 3.       ]]

test_cc_sqrt_bfloat16

Node:
  Sqrt(x) -> (y)

Inputs:
  x: shape=(2, 3), dtype=bfloat16
    [[0, 0.25, 1],
     [2.25, 4, 9]]

Outputs:
  y: shape=(2, 3), dtype=bfloat16
    [[0, 0.5, 1],
     [1.5, 2, 3]]

test_cc_sqrt_float16

Node:
  Sqrt(x) -> (y)

Inputs:
  x: shape=(2, 3), dtype=float16
    [[0.  , 0.25, 1.  ],
     [2.25, 4.  , 9.  ]]

Outputs:
  y: shape=(2, 3), dtype=float16
    [[0. , 0.5, 1. ],
     [1.5, 2. , 3. ]]

test_sqrt

Node:
  Sqrt(x) -> (y)

Inputs:
  x: shape=(3, 4, 5), dtype=float32
    [[[0.5665616 , 0.7457818 , 0.97100276, 0.4443592 , 0.4442647 ],
      [0.7628944 , 0.87734866, 0.5230672 , 0.2855087 , 0.79399663],
      [0.40414217, 0.60542035, 0.4549379 , 0.530079  , 0.4359654 ],
      [0.16703498, 0.64533466, 0.8153506 , 0.681705  , 0.88432455]],

     [[0.06596019, 0.08141465, 0.49587995, 0.12310889, 0.28691137],
      [0.04790118, 0.5155199 , 0.7137708 , 0.04374827, 0.9977479 ],
      [0.5978522 , 0.5865951 , 0.39716926, 0.43898672, 0.25327316],
      [0.5297574 , 0.5436708 , 0.74823534, 0.8180335 , 0.66857344]],

     [[0.86182827, 0.70817924, 0.23636054, 0.6562355 , 0.86890894],
      [0.8393313 , 0.32443324, 0.15917933, 0.893939  , 0.91269016],
      [0.30185628, 0.13158494, 0.32568365, 0.9401739 , 0.39366215],
      [0.08773083, 0.60704976, 0.1554775 , 0.9583321 , 0.8756153 ]]]

Outputs:
  y: shape=(3, 4, 5), dtype=float32
    [[[0.75270283, 0.8635866 , 0.9853947 , 0.66660273, 0.66653186],
      [0.87343824, 0.93666893, 0.7232338 , 0.5343301 , 0.8910649 ],
      [0.63572174, 0.7780876 , 0.67449087, 0.72806525, 0.6602768 ],
      [0.40869913, 0.80332726, 0.90296763, 0.82565427, 0.94038534]],

     [[0.25682718, 0.28533253, 0.70418745, 0.35086876, 0.5356411 ],
      [0.21886338, 0.71799713, 0.8448496 , 0.20916088, 0.9988733 ],
      [0.77320904, 0.76589495, 0.6302137 , 0.6625607 , 0.5032625 ],
      [0.72784436, 0.73734033, 0.86500597, 0.904452  , 0.81766343]],

     [[0.92834705, 0.84153384, 0.48616925, 0.8100836 , 0.93215287],
      [0.9161503 , 0.5695904 , 0.39897284, 0.9454835 , 0.9553482 ],
      [0.5494145 , 0.3627464 , 0.570687  , 0.96962565, 0.627425  ],
      [0.2961939 , 0.779134  , 0.39430633, 0.9789444 , 0.93574315]]]

test_sqrt_example

Node:
  Sqrt(x) -> (y)

Inputs:
  x: shape=(3,), dtype=float32
    [1., 4., 9.]

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

Differences with previous version (6)#

SchemaDiff: Sqrt (domain 'ai.onnx')

  • old version: 6

  • new version: 13

  • breaking: no

Type constraints:

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

Version History#