| { |
| "op": "ai.onnx.BatchNormalization", |
| "fixtureArrays": { |
| "ort_positive_single_channel_7x7_exact_default_epsilon_input_x": [0.329876, -0.287158, -0.411425, 0.473621, 0.18156, -0.170596, -0.329516, -0.170733, -0.121664, 0.4372, -0.485668, 0.218049, -0.360263, 0.107016, 0.45358, 0.325056, 0.15995, 0.098852, -0.283453, -0.373051, 0.257542, 0.0614853, -0.0592363, 0.434488, -0.0179583, 0.398374, -0.451602, -0.132009, -0.174468, -0.0247169, 0.418897, -0.47159, -0.131925, 0.470943, 0.118357, 0.155664, 0.370062, -0.279229, 0.240311, -0.451034, 0.249178, -0.294496, 0.13683, -0.0806475, -0.309849, -0.450604, -0.28048, -0.420197, -0.433369] |
| }, |
| "cases": [ |
| { |
| "name": "dispatch_cliff_nc_inference_rank2", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { "dtype": "float32", "shape": [16776961, 1], "data": { "kind": "linspace", "start": -2.0, "end": 2.0 } }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "constant", "value": 1.5 } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "constant", "value": 0.25 } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "constant", "value": 0.1 } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "constant", "value": 0.5 } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [16776961, 1], "tolerance": 0.0001 } } |
| }, |
| { |
| "name": "nchw_inference", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 3, 2, 2], |
| "data": { "kind": "values", "values": [1.0, 2.0, 3.0, 4.0, -1.0, -2.0, -3.0, -4.0, 2.0, 4.0, 6.0, 8.0] } |
| }, |
| "scale": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.0, 0.5, 2.0] } }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 1.0, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [2.5, -2.5, 5.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.25, 1.25, 5.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 3, 2, 2], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "f32_subnormal_variance_epsilon_zero_gpu_gap", |
| "skipGpu": { |
| "category": "permanent", |
| "reason": "Portable WGSL floating-point semantics do not guarantee preservation of the subnormal values required by this fixture. Backend evidence: WebGPU/Metal flushes subnormals to zero in f32; the subnormal variance collapses to zero so inverseSqrt yields Infinity instead of a finite value." |
| }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.SpatialNoBatch_2", |
| "notes": "Valid epsilon=0 edge: subnormal input variance and tiny normal centered values should produce finite normalized outputs, not infinities." |
| }, |
| "attrs": { "epsilon": 0 }, |
| "inputs": { |
| "x": { "dtype": "float32", "shape": [1, 1, 2], "data": { "kind": "values", "values": [1e-20, -1e-20] } }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.0] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1e-40] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 2], "tolerance": 0.00001 } } |
| }, |
| { |
| "name": "f32_subnormal_scale_rank3_gpu_gap", |
| "skipGpu": { |
| "category": "permanent", |
| "reason": "Portable WGSL floating-point semantics do not guarantee preservation of the subnormal values required by this fixture. Backend evidence: WebGPU/Metal flushes subnormals to zero in f32; the subnormal affine scale collapses to zero, losing the tiny output it should preserve." |
| }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCase", |
| "notes": "Subnormal scale is a valid affine parameter; inference output should preserve the tiny normalized values." |
| }, |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { "dtype": "float32", "shape": [1, 1, 3], "data": { "kind": "values", "values": [-1.0, 0.0, 2.0] } }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1e-40] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 3], "tolerance": 1e-44 } } |
| }, |
| { |
| "name": "f32_subnormal_scale_rank4_vec4_gpu_gap", |
| "skipGpu": { |
| "category": "permanent", |
| "reason": "Portable WGSL floating-point semantics do not guarantee preservation of the subnormal values required by this fixture. Backend evidence: WebGPU/Metal flushes subnormals to zero in f32; the subnormal affine scale collapses to zero (vec4 path)." |
| }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCase", |
| "notes": "Vec4 inference companion: subnormal scale should not collapse an otherwise ordinary normalized channel to zero." |
| }, |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 1, 2, 2], |
| "data": { "kind": "values", "values": [-1.0, 0.0, 1.0, 2.0] } |
| }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-2e-40] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 2, 2], "tolerance": 1e-44 } } |
| }, |
| { |
| "name": "nc_inference_rank2", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3], |
| "data": { "kind": "values", "values": [1.0, 2.0, 3.0, 4.0, 5.0, 6.0] } |
| }, |
| "scale": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.0, 0.5, 2.0] } }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 1.0, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [2.5, 3.5, 4.5] } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [2.25, 2.25, 2.25] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "nc_inference_rank2_vec4", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 8], |
| "data": { "kind": "fillFloat32", "sinStep": 0.17, "cosStep": 0.29 } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [1.0, 0.5, 2.0, -1.0, 0.25, 1.5, -0.75, 0.8] } |
| }, |
| "b": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [0.0, 1.0, -1.0, 0.25, 0.5, -0.5, 0.75, -0.25] } |
| }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [0.2, -0.3, 0.4, -0.5, 0.1, -0.2, 0.3, -0.4] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [1.0, 0.75, 1.25, 2.0, 0.5, 1.5, 0.9, 1.1] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 8], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ncl_inference_rank3", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3, 4], |
| "data": { |
| "kind": "values", |
| "values": [1.0, 2.0, 3.0, 4.0, -1.0, -2.0, -3.0, -4.0, 2.0, 4.0, 6.0, 8.0, 0.5, 1.5, 2.5, 3.5, -0.5, -1.5, -2.5, -3.5, 3.0, 5.0, 7.0, 9.0] |
| } |
| }, |
| "scale": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.0, 0.5, 2.0] } }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 1.0, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [2.0, -2.5, 5.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.25, 1.25, 5.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3, 4], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ncdhw_inference_rank5", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 2, 2, 2, 2], |
| "data": { |
| "kind": "values", |
| "values": [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, -1.0, -2.0, -3.0, -4.0, -5.0, -6.0, -7.0, -8.0] |
| } |
| }, |
| "scale": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [1.0, 0.5] } }, |
| "b": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.0, 1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [4.5, -4.5] } }, |
| "inputVar": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [5.25, 5.25] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 2, 2, 2, 2], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_zero_variance_large_epsilon", |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCase", |
| "notes": "Project regression adapted from ORT BatchNormalization positive cases to stress zero variance with a large epsilon." |
| }, |
| "attrs": { "epsilon": 0.25 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 2, 2, 2], |
| "data": { "kind": "values", "values": [1.0, 1.25, 0.75, 2.0, -3.0, -2.5, -4.0, -3.5] } |
| }, |
| "scale": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [1.0, -2.0] } }, |
| "b": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.5, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [1.0, -3.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [2], "data": { "kind": "constant", "value": 0.0 } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 2, 2, 2], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_positive_single_channel_7x7_generated", |
| "attrs": { "epsilon": 0.00001 }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCase", |
| "notes": "Uses ORT's single-channel rank-4 shape and parameter values; X is generated deterministically." |
| }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 1, 7, 7], |
| "data": { "kind": "fillFloat32", "sinStep": 0.07, "cosStep": 0.19, "scale": 0.75 } |
| }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.589433] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-0.384622] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-2.45673] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.37998] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 7, 7], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_pytorch_rank3_4x5x3_generated", |
| "attrs": { "epsilon": 0.00001 }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.BatchNorm1d_3d_Pytorch", |
| "notes": "Uses ORT's rank-3 Pytorch-style shape and parameter values; X is generated deterministically." |
| }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [4, 5, 3], |
| "data": { "kind": "fillFloat32", "sinStep": 0.11, "cosStep": 0.05, "scale": 1.25 } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [5], |
| "data": { "kind": "values", "values": [0.36102, 0.592982, 0.808513, 0.0531484, 0.0960613] } |
| }, |
| "b": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputMean": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputVar": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 1.0 } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [4, 5, 3], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_positive_single_channel_7x7_exact_default_epsilon", |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCaseDefaultEpsilon" |
| }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 1, 7, 7], |
| "data": { |
| "kind": "values", |
| "values": { "$ref": "#/fixtureArrays/ort_positive_single_channel_7x7_exact_default_epsilon_input_x" } |
| } |
| }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.589433] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-0.384622] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-2.45673] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.37998] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 7, 7], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_pytorch_rank3_4x5x3_exact", |
| "attrs": { "epsilon": 0.00001 }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.BatchNorm1d_3d_Pytorch" |
| }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [4, 5, 3], |
| "data": { |
| "kind": "values", |
| "values": [0.948241, 1.23591, -0.39321, 1.4254, -0.730771, 0.439872, 0.0265089, 0.8748, -0.197505, 0.962646, 0.421469, 1.94512, 0.234179, -0.931897, -0.214905, -0.982965, -0.495436, 0.81949, -0.796605, -0.758605, 0.665557, 0.0909539, 1.10448, 1.91214, -1.97433, -2.26429, -0.384419, -0.226564, 0.230568, 0.533968, -1.31382, -0.156257, 0.532323, -0.16714, 0.971087, 0.600249, 0.858778, 0.423108, -0.414433, -1.17608, 0.673753, 0.278517, -2.19044, -0.161453, 1.17092, -0.155138, -0.094729, 0.19479, -1.17344, -0.213813, 0.118659, -2.39525, 0.257687, 0.784609, 0.297942, 1.10277, -1.58026, 0.197625, 0.0432784, 1.12924] |
| } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [5], |
| "data": { "kind": "values", "values": [0.36102, 0.592982, 0.808513, 0.0531484, 0.0960613] } |
| }, |
| "b": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputMean": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputVar": { "dtype": "float32", "shape": [5], "data": { "kind": "constant", "value": 1.0 } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [4, 5, 3], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_positive_single_channel_5d_exact", |
| "attrs": { "epsilon": 0.00001 }, |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.PositiveTestCase_5D" |
| }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 1, 7, 7, 1], |
| "data": { |
| "kind": "values", |
| "values": { "$ref": "#/fixtureArrays/ort_positive_single_channel_7x7_exact_default_epsilon_input_x" } |
| } |
| }, |
| "scale": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [0.589433] } }, |
| "b": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-0.384622] } }, |
| "inputMean": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [-2.45673] } }, |
| "inputVar": { "dtype": "float32", "shape": [1], "data": { "kind": "values", "values": [1.37998] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 1, 7, 7, 1], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_pytorch_rank4_2x3x6x6_generated", |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.BatchNorm2d_Pytorch", |
| "notes": "Uses the ORT PyTorch rank-4 shape and normalization parameters with deterministic fixture data to avoid committing the large literal tensor." |
| }, |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3, 6, 6], |
| "data": { "kind": "fillFloat32", "scale": 1.4, "sinStep": 0.017, "cosStep": 0.031 } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.736494, 0.580251, 0.374834] } |
| }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 0.0, 0.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 0.0, 0.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.0, 1.0, 1.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3, 6, 6], "tolerance": 0.00001 } } |
| }, |
| { |
| "name": "ort_pytorch_rank5_2x3x4x4x4_generated", |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.BatchNorm3d_Pytorch", |
| "notes": "Uses the ORT PyTorch rank-5 shape and normalization parameters with deterministic fixture data to avoid committing the large literal tensor." |
| }, |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3, 4, 4, 4], |
| "data": { "kind": "fillFloat32", "scale": 1.1, "sinStep": 0.013, "cosStep": 0.019 } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.241661, 0.960798, 0.474727] } |
| }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 0.0, 0.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [0.0, 0.0, 0.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "values", "values": [1.0, 1.0, 1.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3, 4, 4, 4], "tolerance": 0.00001 } } |
| }, |
| { |
| "name": "onnx_backend_batchnorm_epsilon", |
| "attrs": { "epsilon": 0.009999999776482582 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3, 4, 5], |
| "data": { |
| "kind": "values", |
| "values": [0.40746182203292847, 1.3439544439315796, -0.8182209730148315, 0.08270993828773499, -1.2910584211349487, -0.6611042022705078, -1.1801910400390625, 0.19764263927936554, 0.4138999879360199, 1.1973220109939575, 1.88335382938385, 0.7142238020896912, 2.2843334674835205, 1.5641025304794312, 0.6111037135124207, -0.8773633241653442, -1.6210875511169434, -0.5816730260848999, -0.5378339290618896, -1.5560237169265747, -0.05446484312415123, -1.8112788200378418, -0.6311752200126648, -0.9281591773033142, 1.4907219409942627, 0.19549933075904846, -0.4716043472290039, 1.8123546838760376, -2.294137477874756, 0.6512093544006348, -1.1304965019226074, -0.777346670627594, 1.115938425064087, 1.3394529819488525, -1.7674336433410645, 0.42441245913505554, 1.0893090963363647, -0.38418567180633545, 0.6322014331817627, -0.5496559143066406, 0.5211257338523865, 0.10834954679012299, 0.26166847348213196, -0.9147553443908691, 0.8582378029823303, 0.09433428943157196, -1.4859038591384888, -1.9005842208862305, -1.1375792026519775, -1.7620388269424438, -0.28862321376800537, 1.0479822158813477, 0.24995754659175873, 0.046904463320970535, -1.032243013381958, 0.4031856954097748, -0.6840592622756958, 1.262322187423706, -2.005556583404541, -0.33203038573265076, -0.2961004078388214, -2.218360662460327, -0.18350288271903992, 0.3923080563545227, 0.241634801030159, 0.10393591225147247, -0.8295711874961853, 0.4927593767642975, 0.09011279046535492, -0.9975675344467163, -0.80003821849823, 0.2070755809545517, 0.5234630107879639, -0.6993948221206665, 0.9137058258056641, -0.6727848052978516, 0.1333245038986206, 0.42689600586891174, -0.01284939143806696, -0.35224831104278564, 0.8194665908813477, 0.5219877362251282, 1.1972599029541016, -0.3824862241744995, 0.6916618943214417, 0.3538850247859955, 1.047585368156433, -0.4238962233066559, -3.514768123626709, -1.3431566953659058, 1.4255061149597168, 0.2285820096731186, -0.257663756608963, 0.0503707155585289, -1.3802108764648438, -0.26167207956314087, -0.1793796867132187, -0.6927706003189087, 1.137826919555664, -0.1691572517156601, -0.7639136910438538, -0.49807310104370117, -0.3628911077976227, 0.2639603018760681, -0.6296418905258179, -0.47225841879844666, -1.5133610963821411, 1.1076246500015259, 0.1762387454509735, -0.940353512763977, 0.9295943379402161, -1.0627949237823486, -0.8864062428474426, 1.92134690284729, -0.4597805142402649, -1.0890344381332397, 0.9841172695159912, -1.1592062711715698, -0.43653708696365356, 1.0092445611953735] |
| } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.7133895754814148, -0.7280577421188354, 0.8395164608955383] } |
| }, |
| "b": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [1.2390209436416626, -1.784803867340088, -0.7961858510971069] } |
| }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [-1.4005413055419922, -0.18435057997703552, -1.391193151473999] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.044612299650907516, 0.7997958660125732, 0.07695644348859787] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3, 4, 5] } }, |
| "provenance": { "source": "cmake/external/onnx/onnx/backend/test/data/node/test_batchnorm_epsilon" } |
| }, |
| { |
| "name": "onnx_backend_batchnorm_example", |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [2, 3, 4, 5], |
| "data": { |
| "kind": "values", |
| "values": [1.764052391052246, 0.40015721321105957, 0.978738009929657, 2.2408931255340576, 1.8675580024719238, -0.9772778749465942, 0.9500884413719177, -0.15135720372200012, -0.10321885347366333, 0.4105985164642334, 0.14404356479644775, 1.4542734622955322, 0.7610377073287964, 0.12167501449584961, 0.44386324286460876, 0.3336743414402008, 1.4940791130065918, -0.2051582634449005, 0.3130677044391632, -0.8540957570075989, -2.5529897212982178, 0.653618574142456, 0.8644362092018127, -0.7421650290489197, 2.269754648208618, -1.4543657302856445, 0.04575851559638977, -0.18718385696411133, 1.5327792167663574, 1.4693588018417358, 0.154947429895401, 0.37816253304481506, -0.8877857327461243, -1.980796456336975, -0.34791216254234314, 0.15634897351264954, 1.2302906513214111, 1.202379822731018, -0.38732680678367615, -0.302302747964859, -1.0485529899597168, -1.420017957687378, -1.7062702178955078, 1.950775384902954, -0.5096521973609924, -0.4380742907524109, -1.2527953386306763, 0.7774903774261475, -1.6138978004455566, -0.21274028718471527, -0.8954665660858154, 0.38690251111984253, -0.5108051300048828, -1.18063223361969, -0.02818222902715206, 0.4283318817615509, 0.06651721894741058, 0.30247190594673157, -0.6343221068382263, -0.3627411723136902, -0.6724604368209839, -0.35955315828323364, -0.8131462931632996, -1.7262825965881348, 0.17742614448070526, -0.4017809331417084, -1.630198359489441, 0.46278226375579834, -0.9072983860969543, 0.05194539576768875, 0.7290905714035034, 0.12898291647434235, 1.1394007205963135, -1.234825849533081, 0.4023416340351105, -0.6848101019859314, -0.8707971572875977, -0.5788496732711792, -0.3115525245666504, 0.056165341287851334, -1.1651498079299927, 0.9008265137672424, 0.4656624495983124, -1.5362436771392822, 1.4882521629333496, 1.895889163017273, 1.1787796020507812, -0.1799248307943344, -1.0707526206970215, 1.0544517040252686, -0.4031769335269928, 1.222445011138916, 0.2082749754190445, 0.9766390323638916, 0.3563663959503174, 0.7065731883049011, 0.01050002034753561, 1.7858705520629883, 0.12691208720207214, 0.4019893705844879, 1.8831506967544556, -1.3477590084075928, -1.2704850435256958, 0.969396710395813, -1.1731233596801758, 1.9436211585998535, -0.4136189818382263, -0.747454822063446, 1.922942042350769, 1.4805147647857666, 1.8675589561462402, 0.9060446619987488, -0.8612256646156311, 1.910064935684204, -0.26800337433815, 0.8024563789367676, 0.9472519755363464, -0.15501008927822113, 0.6140793561935425, 0.922206699848175] |
| } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.37642553448677063, -1.0994007587432861, 0.29823818802833557] } |
| }, |
| "b": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [1.3263858556747437, -0.694567859172821, -0.14963454008102417] } |
| }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [-0.4351535439491272, 1.8492637872695923, 0.6722947359085083] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [3], |
| "data": { "kind": "values", "values": [0.9755215048789978, 0.855803370475769, 0.011714084073901176] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [2, 3, 4, 5] } }, |
| "provenance": { "source": "cmake/external/onnx/onnx/backend/test/data/node/test_batchnorm_example" } |
| }, |
| { |
| "name": "ort_fp16_inference_rank4", |
| "provenance": { |
| "source": "onnxruntime/test/providers/cpu/nn/batch_norm_op_test.cc", |
| "test": "BatchNormTest.BatchNorm2d_fp16", |
| "notes": "Compact rank-4 fp16 inference projection of ORT's BatchNorm2d_fp16 coverage." |
| }, |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float16", |
| "shape": [1, 2, 2, 2], |
| "data": { "kind": "values", "values": [1.0, 2.0, 3.0, 4.0, -1.0, 0.0, 1.0, 2.0] } |
| }, |
| "scale": { "dtype": "float16", "shape": [2], "data": { "kind": "values", "values": [1.5, -2.0] } }, |
| "b": { "dtype": "float16", "shape": [2], "data": { "kind": "values", "values": [0.25, 1.0] } }, |
| "inputMean": { "dtype": "float16", "shape": [2], "data": { "kind": "values", "values": [2.5, 0.5] } }, |
| "inputVar": { "dtype": "float16", "shape": [2], "data": { "kind": "values", "values": [1.25, 1.25] } } |
| }, |
| "outputs": { "y": { "dtype": "float16", "shape": [1, 2, 2, 2], "tolerance": 0.002 } } |
| }, |
| { |
| "name": "empty_zero_dim", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { "dtype": "float32", "shape": [0, 3, 2, 2], "data": { "kind": "values", "values": [] } }, |
| "scale": { "dtype": "float32", "shape": [3], "data": { "kind": "constant", "value": 1.0 } }, |
| "b": { "dtype": "float32", "shape": [3], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputMean": { "dtype": "float32", "shape": [3], "data": { "kind": "constant", "value": 0.0 } }, |
| "inputVar": { "dtype": "float32", "shape": [3], "data": { "kind": "constant", "value": 1.0 } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [0, 3, 2, 2], "tolerance": 0 } } |
| }, |
| { |
| "name": "empty_zero_dim_nc_vec4_rank2", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { "dtype": "float32", "shape": [0, 8], "data": { "kind": "values", "values": [] } }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [1.0, 0.5, 2.0, -1.0, 0.25, 1.5, -0.75, 0.8] } |
| }, |
| "b": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [0.0, 1.0, -1.0, 0.25, 0.5, -0.5, 0.75, -0.25] } |
| }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [0.2, -0.3, 0.4, -0.5, 0.1, -0.2, 0.3, -0.4] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [8], |
| "data": { "kind": "values", "values": [1.0, 0.75, 1.25, 2.0, 0.5, 1.5, 0.9, 1.1] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [0, 8], "tolerance": 0 } } |
| }, |
| { |
| "name": "nc_vec4_rank2_multigroup_3x12", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [3, 12], |
| "data": { "kind": "fillFloat32", "sinStep": 0.13, "cosStep": 0.23, "scale": 1.5 } |
| }, |
| "scale": { |
| "dtype": "float32", |
| "shape": [12], |
| "data": { "kind": "values", "values": [1.0, 0.5, 2.0, -1.0, 0.25, 1.5, -0.75, 0.8, 1.2, -0.6, 0.9, -1.3] } |
| }, |
| "b": { |
| "dtype": "float32", |
| "shape": [12], |
| "data": { "kind": "values", "values": [0.0, 1.0, -1.0, 0.25, 0.5, -0.5, 0.75, -0.25, 0.1, -0.1, 0.3, -0.3] } |
| }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [12], |
| "data": { "kind": "values", "values": [0.2, -0.3, 0.4, -0.5, 0.1, -0.2, 0.3, -0.4, 0.05, -0.05, 0.15, -0.15] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [12], |
| "data": { "kind": "values", "values": [1.0, 0.75, 1.25, 2.0, 0.5, 1.5, 0.9, 1.1, 0.6, 1.4, 0.8, 1.2] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [3, 12], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "nchw_inference_vec4_channel_boundary_1x4x1x4", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 4, 1, 4], |
| "data": { |
| "kind": "values", |
| "values": [1.0, 2.0, 3.0, 4.0, -1.0, -2.0, -3.0, -4.0, 10.0, 20.0, 30.0, 40.0, 0.5, 1.5, 2.5, 3.5] |
| } |
| }, |
| "scale": { "dtype": "float32", "shape": [4], "data": { "kind": "values", "values": [1.0, 0.5, 2.0, -1.5] } }, |
| "b": { "dtype": "float32", "shape": [4], "data": { "kind": "values", "values": [0.0, 1.0, -1.0, 0.25] } }, |
| "inputMean": { |
| "dtype": "float32", |
| "shape": [4], |
| "data": { "kind": "values", "values": [2.5, -2.5, 25.0, 2.0] } |
| }, |
| "inputVar": { |
| "dtype": "float32", |
| "shape": [4], |
| "data": { "kind": "values", "values": [1.25, 1.25, 125.0, 1.25] } |
| } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 4, 1, 4], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "ort_f16_inference_near_zero_variance_multichannel", |
| "attrs": { "epsilon": 0.00001 }, |
| "inputs": { |
| "x": { |
| "dtype": "float16", |
| "shape": [1, 3, 2, 2], |
| "data": { |
| "kind": "values", |
| "values": [8.0, 9.0, 10.0, 11.0, -4.0, -3.0, -2.0, -1.0, 100.0, 120.0, 140.0, 160.0] |
| } |
| }, |
| "scale": { "dtype": "float16", "shape": [3], "data": { "kind": "values", "values": [2.0, 1.0, 0.5] } }, |
| "b": { "dtype": "float16", "shape": [3], "data": { "kind": "values", "values": [0.5, -1.0, 2.0] } }, |
| "inputMean": { "dtype": "float16", "shape": [3], "data": { "kind": "values", "values": [9.5, -2.5, 130.0] } }, |
| "inputVar": { "dtype": "float16", "shape": [3], "data": { "kind": "values", "values": [0.01, 1.25, 400.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float16", "shape": [1, 3, 2, 2], "tolerance": 0.01 } } |
| }, |
| { |
| "name": "inference_rank6_spatial_flatten", |
| "attrs": { "epsilon": 0.00001, "training_mode": 0 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 2, 2, 1, 2, 1], |
| "data": { "kind": "values", "values": [1.0, 2.0, 3.0, 4.0, -4.0, -2.0, 0.0, 2.0] } |
| }, |
| "scale": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [2.0, 0.5] } }, |
| "b": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.25, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [2.5, -1.0] } }, |
| "inputVar": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [1.25, 5.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 2, 2, 1, 2, 1], "tolerance": 0.000001 } } |
| }, |
| { |
| "name": "inference_rank8_spatial_flatten", |
| "attrs": { "epsilon": 0.00001, "training_mode": 0 }, |
| "inputs": { |
| "x": { |
| "dtype": "float32", |
| "shape": [1, 2, 2, 1, 2, 1, 2, 2], |
| "data": { "kind": "fillFloat32", "sinStep": 0.13, "cosStep": 0.27 } |
| }, |
| "scale": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [1.5, -0.5] } }, |
| "b": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.25, -1.0] } }, |
| "inputMean": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.1, -0.2] } }, |
| "inputVar": { "dtype": "float32", "shape": [2], "data": { "kind": "values", "values": [0.5, 2.0] } } |
| }, |
| "outputs": { "y": { "dtype": "float32", "shape": [1, 2, 2, 1, 2, 1, 2, 2], "tolerance": 0.000001 } } |
| } |
| ] |
| } |
|
|
