Source code
Revision control
Copy as Markdown
Other Tools
Test Info:
- This WPT test may be referenced by the following Test IDs:
- /webnn/conformance_tests/pow.https.any.html?cpu - WPT Dashboard Interop Dashboard
- /webnn/conformance_tests/pow.https.any.html?gpu - WPT Dashboard Interop Dashboard
- /webnn/conformance_tests/pow.https.any.html?npu - WPT Dashboard Interop Dashboard
- /webnn/conformance_tests/pow.https.any.worker.html?cpu - WPT Dashboard Interop Dashboard
- /webnn/conformance_tests/pow.https.any.worker.html?gpu - WPT Dashboard Interop Dashboard
- /webnn/conformance_tests/pow.https.any.worker.html?npu - WPT Dashboard Interop Dashboard
// META: title=test WebNN API element-wise pow operation
// META: global=window,dedicatedworker
// META: variant=?cpu
// META: variant=?gpu
// META: variant=?npu
// META: script=../resources/utils.js
// META: timeout=long
'use strict';
// Compute the element-wise binary power of the two input tensors.
// MLOperand pow(MLOperand a, MLOperand b);
const getPowPrecisionTolerance = (graphResources) => {
const toleranceValueDict = {float32: 32, float16: 2};
const expectedDataType =
getExpectedDataTypeOfSingleOutput(graphResources.expectedOutputs);
return {metricType: 'ULP', value: toleranceValueDict[expectedDataType]};
};
const powTests = [
{
'name':
'pow float32 constant 1D base tensor and 1D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [24], dataType: 'float32'},
'constant': true
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [24], dataType: 'float32'},
'constant': true
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'pow float32 1D base tensor and 1D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [24], dataType: 'float32'}
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [24], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'pow float32 2D base tensor and 2D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
}
}
},
{
'name': 'pow float32 3D base tensor and 3D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
}
}
},
{
'name': 'pow float32 4D base tensor and 4D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name': 'pow float32 5D base tensor and 5D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 2, 1, 2, 3], dataType: 'float32'}
},
'inputB': {
'data': [
1, 6, -7, 7, -2, 1, 4, -10, -2, -5, -2, -10,
-8, -7, -1, -3, -9, 6, -6, 7, -5, -5, 7, -6
],
'descriptor': {shape: [2, 2, 1, 2, 3], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
17.846010208129883,
6.316321332633379e-8,
-1.0973203501407625e-7,
21800822,
0.0033234376460313797,
11.862250328063965,
80273.3359375,
0.00005692423656000756,
0.12908191978931427,
-0.0000020299064544815337,
0.005799346603453159,
3880.540283203125,
2.7385585465999895e-10,
8.223764069725803e-8,
0.06714485585689545,
-0.008288968354463577,
-4.1750155416186985e-11,
1210.5478515625,
3.8231124932508465e-8,
-1667.201416015625,
-0.16149713099002838,
-0.00015812950732652098,
485079424,
3.584487018315485e-8
],
'descriptor': {shape: [2, 2, 1, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name':
'pow (square) float32 4D base tensor and broadcastable 0D integer exponent scalar',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'},
'constant': true
},
'inputB': {'data': [2], 'descriptor': {shape: [], dataType: 'float32'}}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
318.4800720214844, 0.00398249039426446, 97.38143157958984,
124.94144439697266, 300.8932800292969, 140.71298217773438,
283.32550048828125, 7.062208652496338, 7.747018814086914,
189.23854064941406, 172.43321228027344, 0.19152326881885529,
245.8212890625, 105.74638366699219, 221.806640625,
24.41554069519043, 202.5444793701172, 10.657630920410156,
296.84356689453125, 8.328523635864258, 2.073643445968628,
33.14323043823242, 303.1522521972656, 303.2897033691406
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name':
'pow (sqrt) float32 4D positive base tensor and broadcastable 0D integer exponent scalar',
'graph': {
'inputs': {
'inputA': {
'data': [
1.418652057647705, 19.384845733642578, 12.983916282653809,
2.4603159427642822, 7.818154811859131, 6.94444465637207,
12.183951377868652, 17.912473678588867, 11.356568336486816,
8.924248695373535, 17.636823654174805, 11.49622917175293,
18.516279220581055, 2.2580490112304688, 2.231948137283325,
13.629855155944824, 17.54841423034668, 0.5390734076499939,
5.891367435455322, 0.12803149223327637, 19.654495239257812,
3.4122724533081055, 4.945034980773926, 4.437101364135742
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'},
'constant': true
},
'inputB':
{'data': [0.5], 'descriptor': {shape: [], dataType: 'float32'}}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
1.19107186794281, 4.402822494506836, 3.6033201217651367,
1.5685393810272217, 2.7960963249206543, 2.6352314949035645,
3.490551710128784, 4.23231315612793, 3.369950771331787,
2.9873480796813965, 4.199621677398682, 3.3906090259552,
4.3030548095703125, 1.5026806592941284, 1.4939706325531006,
3.6918632984161377, 4.189082622528076, 0.7342162132263184,
2.4272139072418213, 0.35781487822532654, 4.4333391189575195,
1.847233772277832, 2.223743438720703, 2.106442928314209
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name':
'pow float32 4D base tensor and broadcastable 2D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
},
'inputB': {
'data': [5, -10, -10, 7, -7, -9],
'descriptor': {shape: [2, 3], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
1810113, 998220038144, 1.1418765932802444e-10,
21800822, -2.11619832768406e-9, 2.1502860603206386e-10,
-1351182.875, 0.00005692423656000756, 0.000035836616007145494,
-93225256, 1.4853429597394552e-8, -1698.2552490234375,
-947433.5, 7.562621362477984e-11, 1.8626330946375225e-12,
-71917.1015625, -8.45626324519344e-9, 0.00002374253199377563,
1518165.5, 0.00002495513399480842, 0.026081321761012077,
-209595.46875, 2.0615180673644318e-9, 6.786416914539295e-12
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name':
'pow float32 4D base tensor and broadcastable 3D integer exponent tensor',
'graph': {
'inputs': {
'inputA': {
'data': [
17.846010208129883, -0.0631069764494896, -9.868203163146973,
11.17772102355957, -17.346275329589844, 11.862250328063965,
-16.832275390625, 2.6574816703796387, -2.783346652984619,
-13.756400108337402, 13.131382942199707, -0.4376337230205536,
-15.678689002990723, 10.283306121826172, 14.893174171447754,
-4.941208362579346, -14.231812477111816, 3.2646026611328125,
17.229148864746094, -2.885918140411377, -1.4400150775909424,
-5.757015705108643, 17.41126823425293, 17.41521453857422
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
},
'inputB': {
'data': [-5, -10, 9, -6],
'descriptor': {shape: [2, 2, 1], dataType: 'float32'}
}
},
'operators': [{
'name': 'pow',
'arguments': [{'a': 'inputA'}, {'b': 'inputB'}],
'outputs': 'output'
}],
'expectedOutputs': {
'output': {
'data': [
5.524516950572433e-7,
-999109.625,
-0.000010685862434911542,
3.284485530774539e-11,
4.0545030440680696e-13,
1.81271334748212e-11,
-108463955968,
6610.47265625,
-10025.4921875,
1.4756086841316574e-7,
1.9504606996179064e-7,
142.34274291992188,
-0.0000010554830396358739,
0.000008696333679836243,
0.0000013647832020069472,
1.1525726506533829e-7,
2.9335795945217846e-12,
0.000007272717084561009,
133774827520,
-13885.326171875,
-26.625843048095703,
0.000027467271138448268,
3.5893645389251105e-8,
3.584487018315485e-8
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
}
];
if (navigator.ml) {
powTests.forEach((test) => {
webnn_conformance_test(
buildAndExecuteGraph, getPowPrecisionTolerance, test);
});
} else {
test(() => assert_implements(navigator.ml, 'missing navigator.ml'));
}