| // META: title=validation tests for WebNN API prelu operation |
| // META: global=window |
| // META: variant=?cpu |
| // META: variant=?gpu |
| // META: variant=?npu |
| // META: script=../resources/utils_validation.js |
| |
| 'use strict'; |
| |
| validateTwoInputsFromMultipleBuilders('prelu'); |
| |
| const tests = [ |
| { |
| name: |
| '[prelu] Test slope\'s shape = [3, 2, 5] which is the same as input\'s shape.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'float32', shape: [3, 2, 5]}, |
| output: {dataType: 'float32', shape: [3, 2, 5]}, |
| }, |
| { |
| name: |
| '[prelu] Test slope\'s shape = [5] which is unidirectionally broadcastable to input\'s shape.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'float32', shape: [5]}, |
| output: {dataType: 'float32', shape: [3, 2, 5]}, |
| }, |
| { |
| name: |
| '[prelu] Test slope\'s shape = [] which is unidirectionally broadcastable to input\'s shape.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'float32', shape: []}, |
| output: {dataType: 'float32', shape: [3, 2, 5]}, |
| }, |
| { |
| name: |
| '[prelu] Test slope\'s shape = [2, 5] which is unidirectionally broadcastable to input\'s shape.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'float32', shape: [2, 5]}, |
| output: {dataType: 'float32', shape: [3, 2, 5]}, |
| }, |
| { |
| name: |
| '[prelu] Throw if the shape of slope is not broadcastable to the shape of input.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'float32', shape: [2]}, |
| }, |
| { |
| name: |
| '[prelu] Throw if the data type of slope does not match the data type of input.', |
| input: {dataType: 'float32', shape: [3, 2, 5]}, |
| slope: {dataType: 'int32', shape: [3, 2, 5]}, |
| }, |
| ]; |
| |
| tests.forEach( |
| test => promise_test(async t => { |
| const builder = new MLGraphBuilder(context); |
| const input = builder.input('input', test.input); |
| const slope = builder.input('input', test.slope); |
| if (test.output) { |
| const output = builder.prelu(input, slope); |
| assert_equals(output.dataType, test.output.dataType); |
| assert_array_equals(output.shape, test.output.shape); |
| } else { |
| const label = 'prelu_123'; |
| const options = {label}; |
| const regrexp = new RegExp('\\[' + label + '\\]'); |
| assert_throws_with_label( |
| () => builder.prelu(input, slope, options), regrexp); |
| } |
| }, test.name)); |
| |
| promise_test(async t => { |
| for (let dataType of allWebNNOperandDataTypes) { |
| if (!context.opSupportLimits().input.dataTypes.includes(dataType)) { |
| continue; |
| } |
| const builder = new MLGraphBuilder(context); |
| const shape = [1]; |
| const input = builder.input(`input`, {dataType, shape}); |
| if (context.opSupportLimits().prelu.input.dataTypes.includes(dataType)) { |
| const output = builder.prelu(input, input); |
| assert_equals(output.dataType, dataType); |
| assert_array_equals(output.shape, shape); |
| } else { |
| assert_throws_js(TypeError, () => builder.prelu(input, input)); |
| } |
| } |
| }, `[prelu] Test prelu with all of the data types.`); |