101 <a href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml">Go to the documentation of this file.</a><div class="fragment"><div class="line"><a name="l00001"></a><span class="lineno"> 1</span> <span class="comment">//</span></div><div class="line"><a name="l00002"></a><span class="lineno"> 2</span> <span class="comment">// Copyright © 2017 Arm Ltd. All rights reserved.</span></div><div class="line"><a name="l00003"></a><span class="lineno"> 3</span> <span class="comment">// SPDX-License-Identifier: MIT</span></div><div class="line"><a name="l00004"></a><span class="lineno"> 4</span> <span class="comment">//</span></div><div class="line"><a name="l00005"></a><span class="lineno"> 5</span> </div><div class="line"><a name="l00006"></a><span class="lineno"> 6</span> <span class="preprocessor">#include <boost/test/unit_test.hpp></span></div><div class="line"><a name="l00007"></a><span class="lineno"> 7</span> <span class="preprocessor">#include "<a class="code" href="_parser_flatbuffers_fixture_8hpp.xhtml">ParserFlatbuffersFixture.hpp</a>"</span></div><div class="line"><a name="l00008"></a><span class="lineno"> 8</span> <span class="preprocessor">#include "../TfLiteParser.hpp"</span></div><div class="line"><a name="l00009"></a><span class="lineno"> 9</span> <span class="preprocessor">#include <sstream></span></div><div class="line"><a name="l00010"></a><span class="lineno"> 10</span> </div><div class="line"><a name="l00011"></a><span class="lineno"> 11</span> <a class="code" href="_output_shape_of_squeeze_8cpp.xhtml#ae3a6cb217a792718f2bd0e8f45e3ca9e">BOOST_AUTO_TEST_SUITE</a>(TensorflowLiteParser)</div><div class="line"><a name="l00012"></a><span class="lineno"> 12</span> </div><div class="line"><a name="l00013"></a><span class="lineno"> 13</span> <span class="keyword">struct </span>SimpleConv2DFixture : <span class="keyword">public</span> <a class="code" href="struct_parser_flatbuffers_fixture.xhtml">ParserFlatbuffersFixture</a></div><div class="line"><a name="l00014"></a><span class="lineno"> 14</span> {</div><div class="line"><a name="l00015"></a><span class="lineno"> 15</span>  <span class="keyword">explicit</span> SimpleConv2DFixture()</div><div class="line"><a name="l00016"></a><span class="lineno"> 16</span>  {</div><div class="line"><a name="l00017"></a><span class="lineno"> 17</span>  <a class="code" href="struct_parser_flatbuffers_fixture.xhtml#a803c86dca3acef653c1cc481a27be7a9">m_JsonString</a> = R<span class="stringliteral">"(</span></div><div class="line"><a name="l00018"></a><span class="lineno"> 18</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00019"></a><span class="lineno"> 19</span> <span class="stringliteral"> "version": 3,</span></div><div class="line"><a name="l00020"></a><span class="lineno"> 20</span> <span class="stringliteral"> "operator_codes": [ { "builtin_code": "CONV_2D" } ],</span></div><div class="line"><a name="l00021"></a><span class="lineno"> 21</span> <span class="stringliteral"> "subgraphs": [ {</span></div><div class="line"><a name="l00022"></a><span class="lineno"> 22</span> <span class="stringliteral"> "tensors": [</span></div><div class="line"><a name="l00023"></a><span class="lineno"> 23</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00024"></a><span class="lineno"> 24</span> <span class="stringliteral"> "shape": [ 1, 3, 3, 1 ],</span></div><div class="line"><a name="l00025"></a><span class="lineno"> 25</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00026"></a><span class="lineno"> 26</span> <span class="stringliteral"> "buffer": 0,</span></div><div class="line"><a name="l00027"></a><span class="lineno"> 27</span> <span class="stringliteral"> "name": "inputTensor",</span></div><div class="line"><a name="l00028"></a><span class="lineno"> 28</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00029"></a><span class="lineno"> 29</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00030"></a><span class="lineno"> 30</span> <span class="stringliteral"> "max": [ 255.0 ],</span></div><div class="line"><a name="l00031"></a><span class="lineno"> 31</span> <span class="stringliteral"> "scale": [ 1.0 ],</span></div><div class="line"><a name="l00032"></a><span class="lineno"> 32</span> <span class="stringliteral"> "zero_point": [ 0 ],</span></div><div class="line"><a name="l00033"></a><span class="lineno"> 33</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00034"></a><span class="lineno"> 34</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00035"></a><span class="lineno"> 35</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00036"></a><span class="lineno"> 36</span> <span class="stringliteral"> "shape": [ 1, 1, 1, 1 ],</span></div><div class="line"><a name="l00037"></a><span class="lineno"> 37</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00038"></a><span class="lineno"> 38</span> <span class="stringliteral"> "buffer": 1,</span></div><div class="line"><a name="l00039"></a><span class="lineno"> 39</span> <span class="stringliteral"> "name": "outputTensor",</span></div><div class="line"><a name="l00040"></a><span class="lineno"> 40</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00041"></a><span class="lineno"> 41</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00042"></a><span class="lineno"> 42</span> <span class="stringliteral"> "max": [ 511.0 ],</span></div><div class="line"><a name="l00043"></a><span class="lineno"> 43</span> <span class="stringliteral"> "scale": [ 2.0 ],</span></div><div class="line"><a name="l00044"></a><span class="lineno"> 44</span> <span class="stringliteral"> "zero_point": [ 0 ],</span></div><div class="line"><a name="l00045"></a><span class="lineno"> 45</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00046"></a><span class="lineno"> 46</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00047"></a><span class="lineno"> 47</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00048"></a><span class="lineno"> 48</span> <span class="stringliteral"> "shape": [ 1, 3, 3, 1 ],</span></div><div class="line"><a name="l00049"></a><span class="lineno"> 49</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00050"></a><span class="lineno"> 50</span> <span class="stringliteral"> "buffer": 2,</span></div><div class="line"><a name="l00051"></a><span class="lineno"> 51</span> <span class="stringliteral"> "name": "filterTensor",</span></div><div class="line"><a name="l00052"></a><span class="lineno"> 52</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00053"></a><span class="lineno"> 53</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00054"></a><span class="lineno"> 54</span> <span class="stringliteral"> "max": [ 255.0 ],</span></div><div class="line"><a name="l00055"></a><span class="lineno"> 55</span> <span class="stringliteral"> "scale": [ 1.0 ],</span></div><div class="line"><a name="l00056"></a><span class="lineno"> 56</span> <span class="stringliteral"> "zero_point": [ 0 ],</span></div><div class="line"><a name="l00057"></a><span class="lineno"> 57</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00058"></a><span class="lineno"> 58</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00059"></a><span class="lineno"> 59</span> <span class="stringliteral"> ],</span></div><div class="line"><a name="l00060"></a><span class="lineno"> 60</span> <span class="stringliteral"> "inputs": [ 0 ],</span></div><div class="line"><a name="l00061"></a><span class="lineno"> 61</span> <span class="stringliteral"> "outputs": [ 1 ],</span></div><div class="line"><a name="l00062"></a><span class="lineno"> 62</span> <span class="stringliteral"> "operators": [</span></div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span> <span class="stringliteral"> "opcode_index": 0,</span></div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span> <span class="stringliteral"> "inputs": [ 0, 2 ],</span></div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span> <span class="stringliteral"> "outputs": [ 1 ],</span></div><div class="line"><a name="l00067"></a><span class="lineno"> 67</span> <span class="stringliteral"> "builtin_options_type": "Conv2DOptions",</span></div><div class="line"><a name="l00068"></a><span class="lineno"> 68</span> <span class="stringliteral"> "builtin_options": {</span></div><div class="line"><a name="l00069"></a><span class="lineno"> 69</span> <span class="stringliteral"> "padding": "VALID",</span></div><div class="line"><a name="l00070"></a><span class="lineno"> 70</span> <span class="stringliteral"> "stride_w": 1,</span></div><div class="line"><a name="l00071"></a><span class="lineno"> 71</span> <span class="stringliteral"> "stride_h": 1,</span></div><div class="line"><a name="l00072"></a><span class="lineno"> 72</span> <span class="stringliteral"> "fused_activation_function": "NONE"</span></div><div class="line"><a name="l00073"></a><span class="lineno"> 73</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00074"></a><span class="lineno"> 74</span> <span class="stringliteral"> "custom_options_format": "FLEXBUFFERS"</span></div><div class="line"><a name="l00075"></a><span class="lineno"> 75</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00076"></a><span class="lineno"> 76</span> <span class="stringliteral"> ],</span></div><div class="line"><a name="l00077"></a><span class="lineno"> 77</span> <span class="stringliteral"> } ],</span></div><div class="line"><a name="l00078"></a><span class="lineno"> 78</span> <span class="stringliteral"> "buffers" : [</span></div><div class="line"><a name="l00079"></a><span class="lineno"> 79</span> <span class="stringliteral"> { },</span></div><div class="line"><a name="l00080"></a><span class="lineno"> 80</span> <span class="stringliteral"> { },</span></div><div class="line"><a name="l00081"></a><span class="lineno"> 81</span> <span class="stringliteral"> { "data": [ 2,1,0, 6,2,1, 4,1,2 ], },</span></div><div class="line"><a name="l00082"></a><span class="lineno"> 82</span> <span class="stringliteral"> { },</span></div><div class="line"><a name="l00083"></a><span class="lineno"> 83</span> <span class="stringliteral"> ]</span></div><div class="line"><a name="l00084"></a><span class="lineno"> 84</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00085"></a><span class="lineno"> 85</span> <span class="stringliteral"> )";</span></div><div class="line"><a name="l00086"></a><span class="lineno"> 86</span> <span class="stringliteral"> <a class="code" href="struct_parser_flatbuffers_fixture.xhtml#a2bb4ea256fbbf6d53068ca93bb4bc95c">SetupSingleInputSingleOutput</a>(</span><span class="stringliteral">"inputTensor"</span>, <span class="stringliteral">"outputTensor"</span>);</div><div class="line"><a name="l00087"></a><span class="lineno"> 87</span>  }</div><div class="line"><a name="l00088"></a><span class="lineno"> 88</span> };</div><div class="line"><a name="l00089"></a><span class="lineno"> 89</span> </div><div class="line"><a name="l00090"></a><span class="lineno"><a class="line" href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml#afcc317a537dfa8ab47071d5c464bad43"> 90</a></span> <a class="code" href="armnn_onnx_parser_2test_2_conv2_d_8cpp.xhtml#aae03717a46f9e4b8ad98831cc73687ce">BOOST_FIXTURE_TEST_CASE</a>( ParseSimpleConv2D, SimpleConv2DFixture )</div><div class="line"><a name="l00091"></a><span class="lineno"> 91</span> {</div><div class="line"><a name="l00092"></a><span class="lineno"> 92</span>  RunTest<4, armnn::DataType::QAsymmU8>(</div><div class="line"><a name="l00093"></a><span class="lineno"> 93</span>  0,</div><div class="line"><a name="l00094"></a><span class="lineno"> 94</span>  {</div><div class="line"><a name="l00095"></a><span class="lineno"> 95</span>  1, 2, 3,</div><div class="line"><a name="l00096"></a><span class="lineno"> 96</span>  4, 5, 6,</div><div class="line"><a name="l00097"></a><span class="lineno"> 97</span>  7, 8, 9,</div><div class="line"><a name="l00098"></a><span class="lineno"> 98</span>  },</div><div class="line"><a name="l00099"></a><span class="lineno"> 99</span>  <span class="comment">// because of the output scaling we need to take half of the values</span></div><div class="line"><a name="l00100"></a><span class="lineno"> 100</span>  {</div><div class="line"><a name="l00101"></a><span class="lineno"> 101</span>  (1*2 + 2*1 + 3*0 +</div><div class="line"><a name="l00102"></a><span class="lineno"> 102</span>  4*6 + 5*2 + 6*1 +</div><div class="line"><a name="l00103"></a><span class="lineno"> 103</span>  7*4 + 8*1 + 9*2) /2</div><div class="line"><a name="l00104"></a><span class="lineno"> 104</span>  });</div><div class="line"><a name="l00105"></a><span class="lineno"> 105</span> }</div><div class="line"><a name="l00106"></a><span class="lineno"> 106</span> </div><div class="line"><a name="l00107"></a><span class="lineno"> 107</span> <span class="keyword">struct </span>Conv2DWithBiasesFixture : <span class="keyword">public</span> <a class="code" href="struct_parser_flatbuffers_fixture.xhtml">ParserFlatbuffersFixture</a></div><div class="line"><a name="l00108"></a><span class="lineno"> 108</span> {</div><div class="line"><a name="l00109"></a><span class="lineno"> 109</span>  <span class="keyword">explicit</span> Conv2DWithBiasesFixture(<span class="keyword">const</span> std::string & inputShape,</div><div class="line"><a name="l00110"></a><span class="lineno"> 110</span>  <span class="keyword">const</span> std::string & outputShape,</div><div class="line"><a name="l00111"></a><span class="lineno"> 111</span>  <span class="keyword">const</span> std::string & filterShape,</div><div class="line"><a name="l00112"></a><span class="lineno"> 112</span>  <span class="keyword">const</span> std::string & filterData,</div><div class="line"><a name="l00113"></a><span class="lineno"> 113</span>  <span class="keyword">const</span> std::string & biasShape,</div><div class="line"><a name="l00114"></a><span class="lineno"> 114</span>  <span class="keyword">const</span> std::string & biasData,</div><div class="line"><a name="l00115"></a><span class="lineno"> 115</span>  <span class="keyword">const</span> std::string & strides,</div><div class="line"><a name="l00116"></a><span class="lineno"> 116</span>  <span class="keyword">const</span> std::string & activation=<span class="stringliteral">"NONE"</span>,</div><div class="line"><a name="l00117"></a><span class="lineno"> 117</span>  <span class="keyword">const</span> std::string & filterScale=<span class="stringliteral">"1.0"</span>,</div><div class="line"><a name="l00118"></a><span class="lineno"> 118</span>  <span class="keyword">const</span> std::string & filterZeroPoint=<span class="stringliteral">"0"</span>,</div><div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  <span class="keyword">const</span> std::string & outputScale=<span class="stringliteral">"2.0"</span>,</div><div class="line"><a name="l00120"></a><span class="lineno"> 120</span>  <span class="keyword">const</span> std::string & outputZeroPoint=<span class="stringliteral">"0"</span>)</div><div class="line"><a name="l00121"></a><span class="lineno"> 121</span>  {</div><div class="line"><a name="l00122"></a><span class="lineno"> 122</span>  m_JsonString = R<span class="stringliteral">"(</span></div><div class="line"><a name="l00123"></a><span class="lineno"> 123</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00124"></a><span class="lineno"> 124</span> <span class="stringliteral"> "version": 3,</span></div><div class="line"><a name="l00125"></a><span class="lineno"> 125</span> <span class="stringliteral"> "operator_codes": [ { "builtin_code": "CONV_2D" } ],</span></div><div class="line"><a name="l00126"></a><span class="lineno"> 126</span> <span class="stringliteral"> "subgraphs": [ {</span></div><div class="line"><a name="l00127"></a><span class="lineno"> 127</span> <span class="stringliteral"> "tensors": [</span></div><div class="line"><a name="l00128"></a><span class="lineno"> 128</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00129"></a><span class="lineno"> 129</span> <span class="stringliteral"> "shape": )" + inputShape + R</span><span class="stringliteral">"(,</span></div><div class="line"><a name="l00130"></a><span class="lineno"> 130</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00131"></a><span class="lineno"> 131</span> <span class="stringliteral"> "buffer": 0,</span></div><div class="line"><a name="l00132"></a><span class="lineno"> 132</span> <span class="stringliteral"> "name": "inputTensor",</span></div><div class="line"><a name="l00133"></a><span class="lineno"> 133</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00134"></a><span class="lineno"> 134</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00135"></a><span class="lineno"> 135</span> <span class="stringliteral"> "max": [ 255.0 ],</span></div><div class="line"><a name="l00136"></a><span class="lineno"> 136</span> <span class="stringliteral"> "scale": [ 1.0 ],</span></div><div class="line"><a name="l00137"></a><span class="lineno"> 137</span> <span class="stringliteral"> "zero_point": [ 0 ],</span></div><div class="line"><a name="l00138"></a><span class="lineno"> 138</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00139"></a><span class="lineno"> 139</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00140"></a><span class="lineno"> 140</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00141"></a><span class="lineno"> 141</span> <span class="stringliteral"> "shape": )" + outputShape + R</span><span class="stringliteral">"(,</span></div><div class="line"><a name="l00142"></a><span class="lineno"> 142</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00143"></a><span class="lineno"> 143</span> <span class="stringliteral"> "buffer": 1,</span></div><div class="line"><a name="l00144"></a><span class="lineno"> 144</span> <span class="stringliteral"> "name": "outputTensor",</span></div><div class="line"><a name="l00145"></a><span class="lineno"> 145</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00146"></a><span class="lineno"> 146</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00147"></a><span class="lineno"> 147</span> <span class="stringliteral"> "max": [ 511.0 ],</span></div><div class="line"><a name="l00148"></a><span class="lineno"> 148</span> <span class="stringliteral"> "scale": [ )" + outputScale + R</span><span class="stringliteral">"( ],</span></div><div class="line"><a name="l00149"></a><span class="lineno"> 149</span> <span class="stringliteral"> "zero_point": [ )" + outputZeroPoint + R</span><span class="stringliteral">"( ],</span></div><div class="line"><a name="l00150"></a><span class="lineno"> 150</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00151"></a><span class="lineno"> 151</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00152"></a><span class="lineno"> 152</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00153"></a><span class="lineno"> 153</span> <span class="stringliteral"> "shape": )" + filterShape + R</span><span class="stringliteral">"( ,</span></div><div class="line"><a name="l00154"></a><span class="lineno"> 154</span> <span class="stringliteral"> "type": "UINT8",</span></div><div class="line"><a name="l00155"></a><span class="lineno"> 155</span> <span class="stringliteral"> "buffer": 2,</span></div><div class="line"><a name="l00156"></a><span class="lineno"> 156</span> <span class="stringliteral"> "name": "filterTensor",</span></div><div class="line"><a name="l00157"></a><span class="lineno"> 157</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00158"></a><span class="lineno"> 158</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00159"></a><span class="lineno"> 159</span> <span class="stringliteral"> "max": [ 255.0 ],</span></div><div class="line"><a name="l00160"></a><span class="lineno"> 160</span> <span class="stringliteral"> "scale": [ )" + filterScale + R</span><span class="stringliteral">"( ],</span></div><div class="line"><a name="l00161"></a><span class="lineno"> 161</span> <span class="stringliteral"> "zero_point": [ )" + filterZeroPoint + R</span><span class="stringliteral">"( ],</span></div><div class="line"><a name="l00162"></a><span class="lineno"> 162</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00163"></a><span class="lineno"> 163</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00164"></a><span class="lineno"> 164</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00165"></a><span class="lineno"> 165</span> <span class="stringliteral"> "shape": )" + biasShape + R</span><span class="stringliteral">"( ,</span></div><div class="line"><a name="l00166"></a><span class="lineno"> 166</span> <span class="stringliteral"> "type": "INT32",</span></div><div class="line"><a name="l00167"></a><span class="lineno"> 167</span> <span class="stringliteral"> "buffer": 3,</span></div><div class="line"><a name="l00168"></a><span class="lineno"> 168</span> <span class="stringliteral"> "name": "biasTensor",</span></div><div class="line"><a name="l00169"></a><span class="lineno"> 169</span> <span class="stringliteral"> "quantization": {</span></div><div class="line"><a name="l00170"></a><span class="lineno"> 170</span> <span class="stringliteral"> "min": [ 0.0 ],</span></div><div class="line"><a name="l00171"></a><span class="lineno"> 171</span> <span class="stringliteral"> "max": [ 255.0 ],</span></div><div class="line"><a name="l00172"></a><span class="lineno"> 172</span> <span class="stringliteral"> "scale": [ 1.0 ],</span></div><div class="line"><a name="l00173"></a><span class="lineno"> 173</span> <span class="stringliteral"> "zero_point": [ 0 ],</span></div><div class="line"><a name="l00174"></a><span class="lineno"> 174</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00175"></a><span class="lineno"> 175</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00176"></a><span class="lineno"> 176</span> <span class="stringliteral"> ],</span></div><div class="line"><a name="l00177"></a><span class="lineno"> 177</span> <span class="stringliteral"> "inputs": [ 0 ],</span></div><div class="line"><a name="l00178"></a><span class="lineno"> 178</span> <span class="stringliteral"> "outputs": [ 1 ],</span></div><div class="line"><a name="l00179"></a><span class="lineno"> 179</span> <span class="stringliteral"> "operators": [</span></div><div class="line"><a name="l00180"></a><span class="lineno"> 180</span> <span class="stringliteral"> {</span></div><div class="line"><a name="l00181"></a><span class="lineno"> 181</span> <span class="stringliteral"> "opcode_index": 0,</span></div><div class="line"><a name="l00182"></a><span class="lineno"> 182</span> <span class="stringliteral"> "inputs": [ 0, 2, 3 ],</span></div><div class="line"><a name="l00183"></a><span class="lineno"> 183</span> <span class="stringliteral"> "outputs": [ 1 ],</span></div><div class="line"><a name="l00184"></a><span class="lineno"> 184</span> <span class="stringliteral"> "builtin_options_type": "Conv2DOptions",</span></div><div class="line"><a name="l00185"></a><span class="lineno"> 185</span> <span class="stringliteral"> "builtin_options": {</span></div><div class="line"><a name="l00186"></a><span class="lineno"> 186</span> <span class="stringliteral"> "padding": "SAME",</span></div><div class="line"><a name="l00187"></a><span class="lineno"> 187</span> <span class="stringliteral"> "stride_w": )" + strides + R</span><span class="stringliteral">"(,</span></div><div class="line"><a name="l00188"></a><span class="lineno"> 188</span> <span class="stringliteral"> "stride_h": )" + strides + R</span><span class="stringliteral">"(,</span></div><div class="line"><a name="l00189"></a><span class="lineno"> 189</span> <span class="stringliteral"> "fused_activation_function": )" + activation + R</span><span class="stringliteral">"(</span></div><div class="line"><a name="l00190"></a><span class="lineno"> 190</span> <span class="stringliteral"> },</span></div><div class="line"><a name="l00191"></a><span class="lineno"> 191</span> <span class="stringliteral"> "custom_options_format": "FLEXBUFFERS"</span></div><div class="line"><a name="l00192"></a><span class="lineno"> 192</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00193"></a><span class="lineno"> 193</span> <span class="stringliteral"> ],</span></div><div class="line"><a name="l00194"></a><span class="lineno"> 194</span> <span class="stringliteral"> } ],</span></div><div class="line"><a name="l00195"></a><span class="lineno"> 195</span> <span class="stringliteral"> "buffers" : [</span></div><div class="line"><a name="l00196"></a><span class="lineno"> 196</span> <span class="stringliteral"> { },</span></div><div class="line"><a name="l00197"></a><span class="lineno"> 197</span> <span class="stringliteral"> { },</span></div><div class="line"><a name="l00198"></a><span class="lineno"> 198</span> <span class="stringliteral"> { "data": )" + filterData + R</span><span class="stringliteral">"(, },</span></div><div class="line"><a name="l00199"></a><span class="lineno"> 199</span> <span class="stringliteral"> { "data": )" + biasData + R</span><span class="stringliteral">"(, },</span></div><div class="line"><a name="l00200"></a><span class="lineno"> 200</span> <span class="stringliteral"> ]</span></div><div class="line"><a name="l00201"></a><span class="lineno"> 201</span> <span class="stringliteral"> }</span></div><div class="line"><a name="l00202"></a><span class="lineno"> 202</span> <span class="stringliteral"> )";</span></div><div class="line"><a name="l00203"></a><span class="lineno"> 203</span> <span class="stringliteral"> SetupSingleInputSingleOutput(</span><span class="stringliteral">"inputTensor"</span>, <span class="stringliteral">"outputTensor"</span>);</div><div class="line"><a name="l00204"></a><span class="lineno"> 204</span>  }</div><div class="line"><a name="l00205"></a><span class="lineno"> 205</span> };</div><div class="line"><a name="l00206"></a><span class="lineno"> 206</span> </div><div class="line"><a name="l00207"></a><span class="lineno"> 207</span> <span class="keyword">struct </span>SimpleConv2DWithBiasesFixture : Conv2DWithBiasesFixture</div><div class="line"><a name="l00208"></a><span class="lineno"> 208</span> {</div><div class="line"><a name="l00209"></a><span class="lineno"> 209</span>  SimpleConv2DWithBiasesFixture()</div><div class="line"><a name="l00210"></a><span class="lineno"> 210</span>  : Conv2DWithBiasesFixture(<span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// inputShape</span></div><div class="line"><a name="l00211"></a><span class="lineno"> 211</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// outputShape</span></div><div class="line"><a name="l00212"></a><span class="lineno"> 212</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// filterShape</span></div><div class="line"><a name="l00213"></a><span class="lineno"> 213</span>  <span class="stringliteral">"[ 2,1, 0,6 ]"</span>, <span class="comment">// filterData</span></div><div class="line"><a name="l00214"></a><span class="lineno"> 214</span>  <span class="stringliteral">"[ 1 ]"</span>, <span class="comment">// biasShape</span></div><div class="line"><a name="l00215"></a><span class="lineno"> 215</span>  <span class="stringliteral">"[ 10, 0, 0, 0 ]"</span>, <span class="comment">// biasData</span></div><div class="line"><a name="l00216"></a><span class="lineno"> 216</span>  <span class="stringliteral">"1"</span>) <span class="comment">// stride w and h</span></div><div class="line"><a name="l00217"></a><span class="lineno"> 217</span>  {}</div><div class="line"><a name="l00218"></a><span class="lineno"> 218</span> };</div><div class="line"><a name="l00219"></a><span class="lineno"> 219</span> </div><div class="line"><a name="l00220"></a><span class="lineno"><a class="line" href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml#a1f8274795936acc6692b9e18a2d7f7a0"> 220</a></span> <a class="code" href="armnn_onnx_parser_2test_2_conv2_d_8cpp.xhtml#aae03717a46f9e4b8ad98831cc73687ce">BOOST_FIXTURE_TEST_CASE</a>( ParseConv2DWithBias, SimpleConv2DWithBiasesFixture )</div><div class="line"><a name="l00221"></a><span class="lineno"> 221</span> {</div><div class="line"><a name="l00222"></a><span class="lineno"> 222</span>  RunTest<4, armnn::DataType::QAsymmU8>(</div><div class="line"><a name="l00223"></a><span class="lineno"> 223</span>  0,</div><div class="line"><a name="l00224"></a><span class="lineno"> 224</span>  {</div><div class="line"><a name="l00225"></a><span class="lineno"> 225</span>  1, 2,</div><div class="line"><a name="l00226"></a><span class="lineno"> 226</span>  3, 4,</div><div class="line"><a name="l00227"></a><span class="lineno"> 227</span>  },</div><div class="line"><a name="l00228"></a><span class="lineno"> 228</span>  <span class="comment">// because of the output scaling we need to take half of the values</span></div><div class="line"><a name="l00229"></a><span class="lineno"> 229</span>  {</div><div class="line"><a name="l00230"></a><span class="lineno"> 230</span>  (1*2 + 2*1 + 3*0 + 4*6 + 10)/2,</div><div class="line"><a name="l00231"></a><span class="lineno"> 231</span>  (2*2 + 0*1 + 4*0 + 0*6 + 10)/2,</div><div class="line"><a name="l00232"></a><span class="lineno"> 232</span>  (3*2 + 4*1 + 0*0 + 0*6 + 10)/2,</div><div class="line"><a name="l00233"></a><span class="lineno"> 233</span>  (4*2 + 0*1 + 0*0 + 0*6 + 10)/2</div><div class="line"><a name="l00234"></a><span class="lineno"> 234</span>  });</div><div class="line"><a name="l00235"></a><span class="lineno"> 235</span> }</div><div class="line"><a name="l00236"></a><span class="lineno"> 236</span> </div><div class="line"><a name="l00237"></a><span class="lineno"> 237</span> <span class="keyword">struct </span>Conv2DShapeTestFixture : Conv2DWithBiasesFixture</div><div class="line"><a name="l00238"></a><span class="lineno"> 238</span> {</div><div class="line"><a name="l00239"></a><span class="lineno"> 239</span>  <span class="keyword">static</span> std::string GenerateInts(<span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> n)</div><div class="line"><a name="l00240"></a><span class="lineno"> 240</span>  {</div><div class="line"><a name="l00241"></a><span class="lineno"> 241</span>  std::stringstream ss;</div><div class="line"><a name="l00242"></a><span class="lineno"> 242</span>  ss << <span class="stringliteral">" [ "</span>;</div><div class="line"><a name="l00243"></a><span class="lineno"> 243</span>  <span class="keywordflow">for</span>( <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> i=0; i<n; ++i ) {</div><div class="line"><a name="l00244"></a><span class="lineno"> 244</span>  <span class="keywordflow">if</span> (i > 0 )</div><div class="line"><a name="l00245"></a><span class="lineno"> 245</span>  {</div><div class="line"><a name="l00246"></a><span class="lineno"> 246</span>  ss << <span class="stringliteral">" , "</span>;</div><div class="line"><a name="l00247"></a><span class="lineno"> 247</span>  }</div><div class="line"><a name="l00248"></a><span class="lineno"> 248</span>  ss << <span class="stringliteral">" "</span> << (i%256);</div><div class="line"><a name="l00249"></a><span class="lineno"> 249</span>  }</div><div class="line"><a name="l00250"></a><span class="lineno"> 250</span>  ss << <span class="stringliteral">" ] "</span>;</div><div class="line"><a name="l00251"></a><span class="lineno"> 251</span>  <span class="keywordflow">return</span> ss.str();</div><div class="line"><a name="l00252"></a><span class="lineno"> 252</span>  }</div><div class="line"><a name="l00253"></a><span class="lineno"> 253</span> </div><div class="line"><a name="l00254"></a><span class="lineno"> 254</span>  Conv2DShapeTestFixture()</div><div class="line"><a name="l00255"></a><span class="lineno"> 255</span>  : Conv2DWithBiasesFixture(<span class="stringliteral">"[ 1, 224, 224, 3 ]"</span>, <span class="comment">// inputShape</span></div><div class="line"><a name="l00256"></a><span class="lineno"> 256</span>  <span class="stringliteral">"[ 1, 112, 112, 32 ]"</span>, <span class="comment">// outputShape</span></div><div class="line"><a name="l00257"></a><span class="lineno"> 257</span>  <span class="stringliteral">"[ 32, 3, 3, 3 ]"</span>, <span class="comment">// filterShape</span></div><div class="line"><a name="l00258"></a><span class="lineno"> 258</span>  GenerateInts(32*3*3*3), <span class="comment">// filterData</span></div><div class="line"><a name="l00259"></a><span class="lineno"> 259</span>  <span class="stringliteral">"[ 32 ]"</span>, <span class="comment">// biasShape</span></div><div class="line"><a name="l00260"></a><span class="lineno"> 260</span>  GenerateInts(32*4), <span class="comment">// biasData</span></div><div class="line"><a name="l00261"></a><span class="lineno"> 261</span>  <span class="stringliteral">"2"</span>) <span class="comment">// stride w and h</span></div><div class="line"><a name="l00262"></a><span class="lineno"> 262</span>  {}</div><div class="line"><a name="l00263"></a><span class="lineno"> 263</span> };</div><div class="line"><a name="l00264"></a><span class="lineno"> 264</span> </div><div class="line"><a name="l00265"></a><span class="lineno"><a class="line" href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml#a37807b86838def29aa4a342b5ee11a1e"> 265</a></span> <a class="code" href="armnn_onnx_parser_2test_2_conv2_d_8cpp.xhtml#aae03717a46f9e4b8ad98831cc73687ce">BOOST_FIXTURE_TEST_CASE</a>( ParseConv2D_112x112_out, Conv2DShapeTestFixture )</div><div class="line"><a name="l00266"></a><span class="lineno"> 266</span> {</div><div class="line"><a name="l00267"></a><span class="lineno"> 267</span> }</div><div class="line"><a name="l00268"></a><span class="lineno"> 268</span> </div><div class="line"><a name="l00269"></a><span class="lineno"> 269</span> <span class="keyword">struct </span>ReluConv2DWithBiasesFixture : Conv2DWithBiasesFixture</div><div class="line"><a name="l00270"></a><span class="lineno"> 270</span> {</div><div class="line"><a name="l00271"></a><span class="lineno"> 271</span>  ReluConv2DWithBiasesFixture()</div><div class="line"><a name="l00272"></a><span class="lineno"> 272</span>  : Conv2DWithBiasesFixture(<span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// inputShape</span></div><div class="line"><a name="l00273"></a><span class="lineno"> 273</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// outputShape</span></div><div class="line"><a name="l00274"></a><span class="lineno"> 274</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// filterShape</span></div><div class="line"><a name="l00275"></a><span class="lineno"> 275</span>  <span class="stringliteral">"[ 2,1, 0,6 ]"</span>, <span class="comment">// filterData</span></div><div class="line"><a name="l00276"></a><span class="lineno"> 276</span>  <span class="stringliteral">"[ 1 ]"</span>, <span class="comment">// biasShape</span></div><div class="line"><a name="l00277"></a><span class="lineno"> 277</span>  <span class="stringliteral">"[ 16, 0, 0, 0 ]"</span>, <span class="comment">// biasData</span></div><div class="line"><a name="l00278"></a><span class="lineno"> 278</span>  <span class="stringliteral">"1"</span>, <span class="comment">// stride w and h</span></div><div class="line"><a name="l00279"></a><span class="lineno"> 279</span>  <span class="stringliteral">"RELU"</span>, <span class="comment">// activation</span></div><div class="line"><a name="l00280"></a><span class="lineno"> 280</span>  <span class="stringliteral">"1.0"</span>, <span class="comment">// filter scale</span></div><div class="line"><a name="l00281"></a><span class="lineno"> 281</span>  <span class="stringliteral">"4"</span>, <span class="comment">// filter zero point</span></div><div class="line"><a name="l00282"></a><span class="lineno"> 282</span>  <span class="stringliteral">"2.0"</span>, <span class="comment">// output scale</span></div><div class="line"><a name="l00283"></a><span class="lineno"> 283</span>  <span class="stringliteral">"20"</span>) <span class="comment">// output zero point</span></div><div class="line"><a name="l00284"></a><span class="lineno"> 284</span>  {}</div><div class="line"><a name="l00285"></a><span class="lineno"> 285</span> };</div><div class="line"><a name="l00286"></a><span class="lineno"> 286</span> </div><div class="line"><a name="l00287"></a><span class="lineno"><a class="line" href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml#aeae31b1eb962892a52153424119c87e9"> 287</a></span> <a class="code" href="armnn_onnx_parser_2test_2_conv2_d_8cpp.xhtml#aae03717a46f9e4b8ad98831cc73687ce">BOOST_FIXTURE_TEST_CASE</a>( ParseConv2DAndReluWithBias, ReluConv2DWithBiasesFixture )</div><div class="line"><a name="l00288"></a><span class="lineno"> 288</span> {</div><div class="line"><a name="l00289"></a><span class="lineno"> 289</span>  uint8_t bias = 16;</div><div class="line"><a name="l00290"></a><span class="lineno"> 290</span>  uint8_t outZero = 20;</div><div class="line"><a name="l00291"></a><span class="lineno"> 291</span>  uint8_t fz = 4; <span class="comment">// filter zero point</span></div><div class="line"><a name="l00292"></a><span class="lineno"> 292</span> </div><div class="line"><a name="l00293"></a><span class="lineno"> 293</span>  RunTest<4, armnn::DataType::QAsymmU8>(</div><div class="line"><a name="l00294"></a><span class="lineno"> 294</span>  0,</div><div class="line"><a name="l00295"></a><span class="lineno"> 295</span>  {</div><div class="line"><a name="l00296"></a><span class="lineno"> 296</span>  1, 2,</div><div class="line"><a name="l00297"></a><span class="lineno"> 297</span>  4, 8,</div><div class="line"><a name="l00298"></a><span class="lineno"> 298</span>  },</div><div class="line"><a name="l00299"></a><span class="lineno"> 299</span>  <span class="comment">// factors to consider:</span></div><div class="line"><a name="l00300"></a><span class="lineno"> 300</span>  <span class="comment">// - the filter zero point is non zero, hence the (x-fz)</span></div><div class="line"><a name="l00301"></a><span class="lineno"> 301</span>  <span class="comment">// - the output scale is 2 hence the /2</span></div><div class="line"><a name="l00302"></a><span class="lineno"> 302</span>  <span class="comment">// - output zero point is non zero, hence the +outZero</span></div><div class="line"><a name="l00303"></a><span class="lineno"> 303</span>  <span class="comment">// - RELU cuts negative values and then we add the output zero point</span></div><div class="line"><a name="l00304"></a><span class="lineno"> 304</span>  {</div><div class="line"><a name="l00305"></a><span class="lineno"> 305</span>  std::max(outZero, static_cast<uint8_t>((1*(2-fz) + 2*(1-fz) + 4*(0-fz) + 8*(6-fz) + bias)/2 + outZero)),</div><div class="line"><a name="l00306"></a><span class="lineno"> 306</span>  std::max(outZero, static_cast<uint8_t>((2*(2-fz) + 0*(1-fz) + 8*(0-fz) + 0*(6-fz) + bias)/2 + outZero)),</div><div class="line"><a name="l00307"></a><span class="lineno"> 307</span>  std::max(outZero, static_cast<uint8_t>((4*(2-fz) + 8*(1-fz) + 0*(0-fz) + 0*(6-fz) + bias)/2 + outZero)),</div><div class="line"><a name="l00308"></a><span class="lineno"> 308</span>  std::max(outZero, static_cast<uint8_t>((8*(2-fz) + 0*(1-fz) + 0*(0-fz) + 0*(6-fz) + bias)/2 + outZero))</div><div class="line"><a name="l00309"></a><span class="lineno"> 309</span>  });</div><div class="line"><a name="l00310"></a><span class="lineno"> 310</span> }</div><div class="line"><a name="l00311"></a><span class="lineno"> 311</span> </div><div class="line"><a name="l00312"></a><span class="lineno"> 312</span> <span class="keyword">struct </span>Relu6Conv2DWithBiasesFixture : Conv2DWithBiasesFixture</div><div class="line"><a name="l00313"></a><span class="lineno"> 313</span> {</div><div class="line"><a name="l00314"></a><span class="lineno"> 314</span>  Relu6Conv2DWithBiasesFixture()</div><div class="line"><a name="l00315"></a><span class="lineno"> 315</span>  : Conv2DWithBiasesFixture(<span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// inputShape</span></div><div class="line"><a name="l00316"></a><span class="lineno"> 316</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// outputShape</span></div><div class="line"><a name="l00317"></a><span class="lineno"> 317</span>  <span class="stringliteral">"[ 1, 2, 2, 1 ]"</span>, <span class="comment">// filterShape</span></div><div class="line"><a name="l00318"></a><span class="lineno"> 318</span>  <span class="stringliteral">"[ 2,1, 0,6 ]"</span>, <span class="comment">// filterData</span></div><div class="line"><a name="l00319"></a><span class="lineno"> 319</span>  <span class="stringliteral">"[ 1 ]"</span>, <span class="comment">// biasShape</span></div><div class="line"><a name="l00320"></a><span class="lineno"> 320</span>  <span class="stringliteral">"[ 0, 0, 0, 0 ]"</span>, <span class="comment">// biasData</span></div><div class="line"><a name="l00321"></a><span class="lineno"> 321</span>  <span class="stringliteral">"1"</span>, <span class="comment">// stride w and h</span></div><div class="line"><a name="l00322"></a><span class="lineno"> 322</span>  <span class="stringliteral">"RELU6"</span>, <span class="comment">// activation</span></div><div class="line"><a name="l00323"></a><span class="lineno"> 323</span>  <span class="stringliteral">"1.0"</span>, <span class="comment">// filter scale</span></div><div class="line"><a name="l00324"></a><span class="lineno"> 324</span>  <span class="stringliteral">"0"</span>, <span class="comment">// filter zero point</span></div><div class="line"><a name="l00325"></a><span class="lineno"> 325</span>  <span class="stringliteral">"2.0"</span>, <span class="comment">// output scale</span></div><div class="line"><a name="l00326"></a><span class="lineno"> 326</span>  <span class="stringliteral">"0"</span>) <span class="comment">// output zero point</span></div><div class="line"><a name="l00327"></a><span class="lineno"> 327</span>  {}</div><div class="line"><a name="l00328"></a><span class="lineno"> 328</span> };</div><div class="line"><a name="l00329"></a><span class="lineno"> 329</span> </div><div class="line"><a name="l00330"></a><span class="lineno"><a class="line" href="armnn_tf_lite_parser_2test_2_conv2_d_8cpp.xhtml#a16517509d65fc4291eb40bcc285c03fd"> 330</a></span> <a class="code" href="armnn_onnx_parser_2test_2_conv2_d_8cpp.xhtml#aae03717a46f9e4b8ad98831cc73687ce">BOOST_FIXTURE_TEST_CASE</a>( ParseConv2DAndRelu6WithBias, Relu6Conv2DWithBiasesFixture )</div><div class="line"><a name="l00331"></a><span class="lineno"> 331</span> {</div><div class="line"><a name="l00332"></a><span class="lineno"> 332</span>  uint8_t relu6Min = 6 / 2; <span class="comment">// divide by output scale</span></div><div class="line"><a name="l00333"></a><span class="lineno"> 333</span> </div><div class="line"><a name="l00334"></a><span class="lineno"> 334</span>  RunTest<4, armnn::DataType::QAsymmU8>(</div><div class="line"><a name="l00335"></a><span class="lineno"> 335</span>  0,</div><div class="line"><a name="l00336"></a><span class="lineno"> 336</span>  {</div><div class="line"><a name="l00337"></a><span class="lineno"> 337</span>  1, 2,</div><div class="line"><a name="l00338"></a><span class="lineno"> 338</span>  4, 1,</div><div class="line"><a name="l00339"></a><span class="lineno"> 339</span>  },</div><div class="line"><a name="l00340"></a><span class="lineno"> 340</span>  <span class="comment">// factors to consider:</span></div><div class="line"><a name="l00341"></a><span class="lineno"> 341</span>  <span class="comment">// - the output scale is 2 hence the /2</span></div><div class="line"><a name="l00342"></a><span class="lineno"> 342</span>  <span class="comment">// - RELU6 cuts output values at +6</span></div><div class="line"><a name="l00343"></a><span class="lineno"> 343</span>  {</div><div class="line"><a name="l00344"></a><span class="lineno"> 344</span>  std::min(relu6Min, static_cast<uint8_t>((1*2 + 2*1 + 4*0 + 1*6)/2)),</div><div class="line"><a name="l00345"></a><span class="lineno"> 345</span>  std::min(relu6Min, static_cast<uint8_t>((2*2 + 0*1 + 1*0 + 0*6)/2)),</div><div class="line"><a name="l00346"></a><span class="lineno"> 346</span>  std::min(relu6Min, static_cast<uint8_t>((4*2 + 1*1 + 0*0 + 0*6)/2)),</div><div class="line"><a name="l00347"></a><span class="lineno"> 347</span>  std::min(relu6Min, static_cast<uint8_t>((1*2 + 0*1 + 0*0 + 0*6)/2))</div><div class="line"><a name="l00348"></a><span class="lineno"> 348</span>  });</div><div class="line"><a name="l00349"></a><span class="lineno"> 349</span> }</div><div class="line"><a name="l00350"></a><span class="lineno"> 350</span> </div><div class="line"><a name="l00351"></a><span class="lineno"> 351</span> <a class="code" href="_profiler_tests_8cpp.xhtml#af7f71af5c6c124222dd1c42c5df892f4">BOOST_AUTO_TEST_SUITE_END</a>()</div><div class="ttc" id="_output_shape_of_squeeze_8cpp_xhtml_ae3a6cb217a792718f2bd0e8f45e3ca9e"><div class="ttname"><a href="_output_shape_of_squeeze_8cpp.xhtml#ae3a6cb217a792718f2bd0e8f45e3ca9e">BOOST_AUTO_TEST_SUITE</a></div><div class="ttdeci">BOOST_AUTO_TEST_SUITE(TensorflowLiteParser)</div></div>