Source code

Revision control

Copy as Markdown

Other Tools

Test Info:

// META: title=test WebNN API element-wise erf 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 error function of the input tensor, element-wise.
//
// MLOperand erf(MLOperand input);
const getErfPrecisionTolerance = (graphResources) => {
const toleranceValueDict = {float32: 1 / 1024, float16: 1 / 512};
const expectedDataType =
getExpectedDataTypeOfSingleOutput(graphResources.expectedOutputs);
return {metricType: 'ATOL', value: toleranceValueDict[expectedDataType]};
};
const erfTests = [
{
'name': 'erf float32 0D scalar',
'graph': {
'inputs': {
'erfInput': {
'data': [-0.004352752584964037],
'descriptor': {shape: [], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [-0.004911554511636496],
'descriptor': {shape: [], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 1D constant tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [24], dataType: 'float32'},
'constant': true
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 1D tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [24], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [24], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 2D tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [4, 6], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 3D tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [2, 3, 4], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 4D tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [2, 2, 2, 3], dataType: 'float32'}
}
}
}
},
{
'name': 'erf float32 5D tensor',
'graph': {
'inputs': {
'erfInput': {
'data': [
-0.5883062481880188, -0.5323766469955444, 0.1279870569705963,
0.5772181749343872, 0.904577910900116, -0.49856746196746826,
0.9843886494636536, 0.21642258763313293, 0.25235170125961304,
-0.1491669863462448, 0.9136939644813538, -0.5736439824104309,
0.7404468059539795, 0.5202203392982483, -0.16549505293369293,
0.8005731701850891, 0.4001320004463196, -0.7489422559738159,
-0.8946002721786499, -0.3920256197452545, 0.6854220628738403,
-0.6744900345802307, -0.439808189868927, 0.2709762454032898
],
'descriptor': {shape: [2, 1, 4, 1, 3], dataType: 'float32'}
}
},
'operators': [{
'name': 'erf',
'arguments': [{'input': 'erfInput'}],
'outputs': 'erfOutput'
}],
'expectedOutputs': {
'erfOutput': {
'data': [
-0.5945858955383301, -0.5484867095947266, 0.14363321661949158,
0.5856771469116211, 0.7991966009140015, -0.5192402005195618,
0.8361185789108276, 0.2404468059539795, 0.27881762385368347,
-0.16707675158977509, 0.8036975860595703, -0.5827808976173401,
0.7049695253372192, 0.5380884408950806, -0.1850501298904419,
0.742441713809967, 0.4285193383693695, -0.710474967956543,
-0.7941843867301941, -0.4207003712654114, 0.6676210165023804,
-0.6598520874977112, -0.46604686975479126, 0.29844197630882263
],
'descriptor': {shape: [2, 1, 4, 1, 3], dataType: 'float32'}
}
}
}
}
];
if (navigator.ml) {
erfTests.forEach((test) => {
webnn_conformance_test(
buildAndExecuteGraph, getErfPrecisionTolerance, test);
});
} else {
test(() => assert_implements(navigator.ml, 'missing navigator.ml'));
}