Source code

Revision control

Copy as Markdown

Other Tools

Test Info: Warnings

// META: title=test WebNN API leakyRelu operation
// META: global=window,dedicatedworker
// META: variant=?cpu
// META: variant=?gpu
// META: variant=?npu
// META: script=../resources/utils.js
// META: timeout=long
'use strict';
// Calculate the leaky version of rectified linear function on the input tensor
// element-wise. The calculation follows the expression
// max(0, x) + alpha * min(0, x).
//
// dictionary MLLeakyReluOptions {
// double alpha = 0.01;
// };
//
// MLOperand leakyRelu(
// MLOperand input, optional MLLeakyReluOptions options = {});
const leakyReluTests = [
{
'name': 'leakyRelu float32 1D constant tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [24], dataType: 'float32'},
'constant': true
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 0D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [-19.053640365600586],
'descriptor': {shape: [], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [-0.19053640961647034],
'descriptor': {shape: [], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 1D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [24], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 2D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 3D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 4D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [1, 2, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [1, 2, 3, 4], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 5D tensor default options',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [1, 2, 1, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-0.19053640961647034, 50.77590560913086, -0.695496678352356,
-0.8057432770729065, -0.9040110111236572, 76.02884674072266,
66.33873748779297, -0.8410186767578125, -0.1719101369380951,
-0.8747624158859253, -0.0341646634042263, -0.2277235984802246,
-0.02509489096701145, 18.933284759521484, 98.61402893066406,
55.3392333984375, -0.33178603649139404, -0.4603901207447052,
-0.6147925853729248, 64.26514434814453, 21.469341278076172,
-0.31514689326286316, -0.4127694368362427, -0.6559529304504395
],
'descriptor': {shape: [1, 2, 1, 3, 4], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 1D tensor negative options.alpha',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [24], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [
{'input': 'leakyReluInput'},
{'options': {'alpha': -97.70109193608776}}
],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
1861.5615234375, 50.77590560913086, 6795.07861328125,
7872.19970703125, 8832.2861328125, 76.02884674072266,
66.33873748779297, 8216.8447265625, 1679.580810546875,
8546.5244140625, 333.7925109863281, 2224.884521484375,
245.17982482910156, 18.933284759521484, 98.61402893066406,
55.3392333984375, 3241.5859375, 4498.06201171875,
6006.5908203125, 64.26514434814453, 21.469341278076172,
3079.019775390625, 4032.802490234375, 6408.73193359375
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 2D tensor positive options.alpha',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [
{'input': 'leakyReluInput'},
{'options': {'alpha': 35.799162942273234}}
],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
-682.1043701171875, 50.77590560913086, -2489.81982421875,
-2884.493408203125, -3236.28369140625, 76.02884674072266,
66.33873748779297, -3010.776611328125, -615.4238891601562,
-3131.576416015625, -122.306640625, -815.2314453125,
-89.83760833740234, 18.933284759521484, 98.61402893066406,
55.3392333984375, -1187.7662353515625, -1648.158203125,
-2200.906005859375, 64.26514434814453, 21.469341278076172,
-1128.1995849609375, -1477.6800537109375, -2348.256591796875
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
}
}
},
{
'name': 'leakyRelu float32 5D tensor options.alpha=0.0',
'graph': {
'inputs': {
'leakyReluInput': {
'data': [
-19.053640365600586, 50.77590560913086, -69.54966735839844,
-80.57432556152344, -90.4011001586914, 76.02884674072266,
66.33873748779297, -84.10186767578125, -17.19101333618164,
-87.47624206542969, -3.416466474533081, -22.77235984802246,
-2.509489059448242, 18.933284759521484, 98.61402893066406,
55.3392333984375, -33.17860412597656, -46.03901290893555,
-61.47925567626953, 64.26514434814453, 21.469341278076172,
-31.514690399169922, -41.27694320678711, -65.59529113769531
],
'descriptor': {shape: [1, 2, 1, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'leakyRelu',
'arguments': [{'input': 'leakyReluInput'}, {'options': {'alpha': 0}}],
'outputs': 'leakyReluOutput'
}],
'expectedOutputs': {
'leakyReluOutput': {
'data': [
0,
50.77590560913086,
0,
0,
0,
76.02884674072266,
66.33873748779297,
0,
0,
0,
0,
0,
0,
18.933284759521484,
98.61402893066406,
55.3392333984375,
0,
0,
0,
64.26514434814453,
21.469341278076172,
0,
0,
0
],
'descriptor': {shape: [1, 2, 1, 3, 4], dataType: 'float32'}
}
}
}
}
];
if (navigator.ml) {
leakyReluTests.forEach((test) => {
webnn_conformance_test(buildAndExecuteGraph, getPrecisionTolerance, test);
});
} else {
test(() => assert_implements(navigator.ml, 'missing navigator.ml'));
}