| // META: title=validation tests for WebNN API reduction operation |
| // META: global=window |
| // META: variant=?cpu |
| // META: variant=?gpu |
| // META: variant=?npu |
| // META: script=../resources/utils_validation.js |
| |
| 'use strict'; |
| |
| const kReductionOperators = [ |
| 'reduceL1', |
| 'reduceL2', |
| 'reduceLogSum', |
| 'reduceLogSumExp', |
| 'reduceMax', |
| 'reduceMean', |
| 'reduceMin', |
| 'reduceProduct', |
| 'reduceSum', |
| 'reduceSumSquare', |
| ]; |
| |
| const label = 'reduce_op_xxx'; |
| |
| const allReductionOperatorsTests = [ |
| { |
| name: '[reduce] Test reduce with keepDimensions=true.', |
| input: {dataType: 'float32', shape: [1, 3, 4, 4]}, |
| options: { |
| keepDimensions: true, |
| }, |
| output: {dataType: 'float32', shape: [1, 1, 1, 1]} |
| }, |
| { |
| name: '[reduce] Test reduce with axes=[0, 1] and keep_dimensions=false.', |
| input: {dataType: 'float32', shape: [1, 3, 5, 5]}, |
| options: {axes: [0, 1]}, |
| output: {dataType: 'float32', shape: [5, 5]} |
| }, |
| { |
| name: '[reduce] Throw if a value in axes is out of range of [0, N-1].', |
| input: {dataType: 'float32', shape: [1, 2, 5, 5]}, |
| options: { |
| axes: [4], |
| label: label, |
| }, |
| }, |
| { |
| name: '[reduce] Throw if the two values are same in axes sequence.', |
| input: {dataType: 'float32', shape: [1, 2, 5, 5]}, |
| options: { |
| axes: [0, 1, 1], |
| label: label, |
| }, |
| }, |
| ]; |
| |
| function runReductionTests(operatorName, tests) { |
| tests.forEach(test => { |
| promise_test(async t => { |
| const builder = new MLGraphBuilder(context); |
| const input = builder.input('input', test.input); |
| |
| if (test.output) { |
| const output = builder[operatorName](input, test.options); |
| assert_equals(output.dataType, test.output.dataType); |
| assert_array_equals(output.shape, test.output.shape); |
| } else { |
| const regrexp = new RegExp('\\[' + label + '\\]'); |
| assert_throws_with_label( |
| () => builder[operatorName](input, test.options), regrexp); |
| } |
| }, test.name.replace('[reduce]', `[${operatorName}]`)); |
| }); |
| } |
| |
| kReductionOperators.forEach((operatorName) => { |
| validateInputFromAnotherBuilder(operatorName); |
| runReductionTests(operatorName, allReductionOperatorsTests); |
| }); |
| |
| kReductionOperators.forEach((operatorName) => { |
| promise_test(async t => { |
| for (let dataType of allWebNNOperandDataTypes) { |
| if (!context.opSupportLimits().input.dataTypes.includes(dataType)) { |
| continue; |
| } |
| const builder = new MLGraphBuilder(context); |
| const input = builder.input(`input`, {dataType, shape: shape3D}); |
| if (context.opSupportLimits()[operatorName].input.dataTypes.includes( |
| dataType)) { |
| const output = builder[operatorName](input); |
| assert_equals(output.dataType, dataType); |
| assert_array_equals(output.shape, []); |
| } else { |
| assert_throws_js(TypeError, () => builder[operatorName](input)); |
| } |
| } |
| }, `[${operatorName}] Test reduce with all of the data types.`); |
| }); |