1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
|
//
// Copyright © 2024 Arm Ltd and Contributors. All rights reserved.
// SPDX-License-Identifier: MIT
//
#pragma once
#include "TestUtils.hpp"
#include <armnn_delegate.hpp>
#include <DelegateTestInterpreter.hpp>
#include <tensorflow/lite/version.h>
namespace
{
std::vector<char> CreateScatterNdTfLiteModel(tflite::TensorType tensorType,
const std::vector<int32_t>& indicesShape,
const std::vector<int32_t>& updatesShape,
const std::vector<int32_t>& shapeShape,
const std::vector<int32_t>& outputShape,
const std::vector<int32_t>& shapeData,
float quantScale = 1.0f,
int quantOffset = 0)
{
using namespace tflite;
flatbuffers::FlatBufferBuilder flatBufferBuilder;
std::vector<flatbuffers::Offset<tflite::Buffer>> buffers;
buffers.push_back(CreateBuffer(flatBufferBuilder));
buffers.push_back(CreateBuffer(flatBufferBuilder)); // indices
buffers.push_back(CreateBuffer(flatBufferBuilder)); // updates
buffers.push_back(CreateBuffer(flatBufferBuilder,
flatBufferBuilder.CreateVector(reinterpret_cast<const uint8_t*>(shapeData.data()),
sizeof(int32_t) * shapeData.size())));
buffers.push_back(CreateBuffer(flatBufferBuilder)); // output
auto quantizationParameters =
CreateQuantizationParameters(flatBufferBuilder,
0,
0,
flatBufferBuilder.CreateVector<float>({ quantScale }),
flatBufferBuilder.CreateVector<int64_t>({ quantOffset }));
std::array<flatbuffers::Offset<Tensor>, 4> tensors;
tensors[0] = CreateTensor(flatBufferBuilder,
flatBufferBuilder.CreateVector<int32_t>(indicesShape.data(),
indicesShape.size()),
TensorType_INT32,
1,
flatBufferBuilder.CreateString("indices_tensor"),
quantizationParameters);
tensors[1] = CreateTensor(flatBufferBuilder,
flatBufferBuilder.CreateVector<int32_t>(updatesShape.data(),
updatesShape.size()),
tensorType,
2,
flatBufferBuilder.CreateString("updates_tensor"),
quantizationParameters);
tensors[2] = CreateTensor(flatBufferBuilder,
flatBufferBuilder.CreateVector<int32_t>(shapeShape.data(),
shapeShape.size()),
TensorType_INT32,
3,
flatBufferBuilder.CreateString("shape_tensor"),
quantizationParameters);
tensors[3] = CreateTensor(flatBufferBuilder,
flatBufferBuilder.CreateVector<int32_t>(outputShape.data(),
outputShape.size()),
tensorType,
4,
flatBufferBuilder.CreateString("output_tensor"),
quantizationParameters);
// Create Operator
tflite::BuiltinOptions operatorBuiltinOptionsType = tflite::BuiltinOptions_ScatterNdOptions;
flatbuffers::Offset<void> operatorBuiltinOptions = CreateScatterNdOptions(flatBufferBuilder).Union();
const std::vector<int> operatorInputs { 0, 1, 2 };
const std::vector<int> operatorOutputs { 3 };
flatbuffers::Offset<Operator> scatterNdOperator =
CreateOperator(flatBufferBuilder,
0,
flatBufferBuilder.CreateVector<int32_t>(operatorInputs.data(), operatorInputs.size()),
flatBufferBuilder.CreateVector<int32_t>(operatorOutputs.data(), operatorOutputs.size()),
operatorBuiltinOptionsType,
operatorBuiltinOptions);
const std::vector<int> subgraphInputs{ 0, 1, 2 };
const std::vector<int> subgraphOutputs{ 3 };
flatbuffers::Offset <SubGraph> subgraph =
CreateSubGraph(flatBufferBuilder,
flatBufferBuilder.CreateVector(tensors.data(), tensors.size()),
flatBufferBuilder.CreateVector<int32_t>(subgraphInputs.data(), subgraphInputs.size()),
flatBufferBuilder.CreateVector<int32_t>(subgraphOutputs.data(), subgraphOutputs.size()),
flatBufferBuilder.CreateVector(&scatterNdOperator, 1));
flatbuffers::Offset <flatbuffers::String> modelDescription =
flatBufferBuilder.CreateString("ArmnnDelegate: ScatterNd Operator Model");
flatbuffers::Offset <OperatorCode> opCode = CreateOperatorCode(flatBufferBuilder,
tflite::BuiltinOperator_SCATTER_ND);
flatbuffers::Offset <Model> flatbufferModel =
CreateModel(flatBufferBuilder,
TFLITE_SCHEMA_VERSION,
flatBufferBuilder.CreateVector(&opCode, 1),
flatBufferBuilder.CreateVector(&subgraph, 1),
modelDescription,
flatBufferBuilder.CreateVector(buffers.data(), buffers.size()));
flatBufferBuilder.Finish(flatbufferModel, armnnDelegate::FILE_IDENTIFIER);
return std::vector<char>(flatBufferBuilder.GetBufferPointer(),
flatBufferBuilder.GetBufferPointer() + flatBufferBuilder.GetSize());
}
template<typename T>
void ScatterNdTestImpl(tflite::TensorType tensorType,
std::vector<int32_t>& indicesShape,
std::vector<int32_t>& indicesValues,
std::vector<int32_t>& updatesShape,
std::vector<T>& updatesValues,
std::vector<int32_t>& shapeShape,
std::vector<int32_t>& shapeValue,
std::vector<int32_t>& expectedOutputShape,
std::vector<T>& expectedOutputValues,
const std::vector<armnn::BackendId>& backends = {},
float quantScale = 1.0f,
int quantOffset = 0)
{
using namespace delegateTestInterpreter;
std::vector<char> modelBuffer = CreateScatterNdTfLiteModel(tensorType,
indicesShape,
updatesShape,
shapeShape,
expectedOutputShape,
shapeValue,
quantScale,
quantOffset);
// Setup interpreter with just TFLite Runtime.
auto tfLiteInterpreter = DelegateTestInterpreter(modelBuffer);
CHECK(tfLiteInterpreter.AllocateTensors() == kTfLiteOk);
CHECK(tfLiteInterpreter.FillInputTensor<int32_t>(indicesValues, 0) == kTfLiteOk);
CHECK(tfLiteInterpreter.FillInputTensor<T>(updatesValues, 1) == kTfLiteOk);
CHECK(tfLiteInterpreter.FillInputTensor<int32_t>(shapeValue, 2) == kTfLiteOk);
CHECK(tfLiteInterpreter.Invoke() == kTfLiteOk);
std::vector<T> tfLiteOutputValues = tfLiteInterpreter.GetOutputResult<T>(0);
std::vector<int32_t> tfLiteOutputShape = tfLiteInterpreter.GetOutputShape(0);
// Setup interpreter with Arm NN Delegate applied.
auto armnnInterpreter = DelegateTestInterpreter(modelBuffer, CaptureAvailableBackends(backends));
CHECK(armnnInterpreter.AllocateTensors() == kTfLiteOk);
CHECK(armnnInterpreter.FillInputTensor<int32_t>(indicesValues, 0) == kTfLiteOk);
CHECK(armnnInterpreter.FillInputTensor<T>(updatesValues, 1) == kTfLiteOk);
CHECK(armnnInterpreter.FillInputTensor<int32_t>(shapeValue, 2) == kTfLiteOk);
CHECK(armnnInterpreter.Invoke() == kTfLiteOk);
std::vector<T> armnnOutputValues = armnnInterpreter.GetOutputResult<T>(0);
std::vector<int32_t> armnnOutputShape = armnnInterpreter.GetOutputShape(0);
armnnDelegate::CompareOutputData<T>(tfLiteOutputValues, armnnOutputValues, expectedOutputValues);
armnnDelegate::CompareOutputShape(tfLiteOutputShape, armnnOutputShape, expectedOutputShape);
tfLiteInterpreter.Cleanup();
armnnInterpreter.Cleanup();
}
} // anonymous namespace
|
