| <!-- HTML header for doxygen 1.8.15--> |
| <!-- Remember to use version doxygen 1.8.15 +--> |
| <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> |
| <html xmlns="http://www.w3.org/1999/xhtml"> |
| <head> |
| <meta http-equiv="Content-Type" content="text/xhtml;charset=UTF-8"/> |
| <meta http-equiv="X-UA-Compatible" content="IE=9"/> |
| <meta name="generator" content="Doxygen 1.8.15"/> |
| <meta name="robots" content="NOINDEX, NOFOLLOW" /> <!-- Prevent indexing by search engines --> |
| <title>Compute Library: src/core/NEON/kernels/NEFuseBatchNormalizationKernel.cpp Source File</title> |
| <link href="tabs.css" rel="stylesheet" type="text/css"/> |
| <script type="text/javascript" src="jquery.js"></script> |
| <script type="text/javascript" src="dynsections.js"></script> |
| <link href="navtree.css" rel="stylesheet" type="text/css"/> |
| <script type="text/javascript" src="resize.js"></script> |
| <script type="text/javascript" src="navtreedata.js"></script> |
| <script type="text/javascript" src="navtree.js"></script> |
| <script type="text/javascript"> |
| /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ |
| $(document).ready(initResizable); |
| /* @license-end */</script> |
| <link href="search/search.css" rel="stylesheet" type="text/css"/> |
| <script type="text/javascript" src="search/searchdata.js"></script> |
| <script type="text/javascript" src="search/search.js"></script> |
| <script type="text/x-mathjax-config"> |
| MathJax.Hub.Config({ |
| extensions: ["tex2jax.js"], |
| jax: ["input/TeX","output/HTML-CSS"], |
| }); |
| </script><script type="text/javascript" async="async" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js"></script> |
| <link href="doxygen.css" rel="stylesheet" type="text/css" /> |
| <link href="stylesheet.css" rel="stylesheet" type="text/css"/> |
| </head> |
| <body> |
| <div id="top"><!-- do not remove this div, it is closed by doxygen! --> |
| <div id="titlearea"> |
| <table cellspacing="0" cellpadding="0"> |
| <tbody> |
| <tr style="height: 56px;"> |
| <img alt="Compute Library" src="https://raw.githubusercontent.com/ARM-software/ComputeLibrary/gh-pages/ACL_logo.png" style="max-width: 100%;margin-top: 15px;margin-left: 10px"/> |
| <td style="padding-left: 0.5em;"> |
| <div id="projectname"> |
|  <span id="projectnumber">19.08</span> |
| </div> |
| </td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| <!-- end header part --> |
| <!-- Generated by Doxygen 1.8.15 --> |
| <script type="text/javascript"> |
| /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ |
| var searchBox = new SearchBox("searchBox", "search",false,'Search'); |
| /* @license-end */ |
| </script> |
| <script type="text/javascript" src="menudata.js"></script> |
| <script type="text/javascript" src="menu.js"></script> |
| <script type="text/javascript"> |
| /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ |
| $(function() { |
| initMenu('',true,false,'search.php','Search'); |
| $(document).ready(function() { init_search(); }); |
| }); |
| /* @license-end */</script> |
| <div id="main-nav"></div> |
| </div><!-- top --> |
| <div id="side-nav" class="ui-resizable side-nav-resizable"> |
| <div id="nav-tree"> |
| <div id="nav-tree-contents"> |
| <div id="nav-sync" class="sync"></div> |
| </div> |
| </div> |
| <div id="splitbar" style="-moz-user-select:none;" |
| class="ui-resizable-handle"> |
| </div> |
| </div> |
| <script type="text/javascript"> |
| /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&dn=gpl-2.0.txt GPL-v2 */ |
| $(document).ready(function(){initNavTree('_n_e_fuse_batch_normalization_kernel_8cpp_source.xhtml','');}); |
| /* @license-end */ |
| </script> |
| <div id="doc-content"> |
| <!-- window showing the filter options --> |
| <div id="MSearchSelectWindow" |
| onmouseover="return searchBox.OnSearchSelectShow()" |
| onmouseout="return searchBox.OnSearchSelectHide()" |
| onkeydown="return searchBox.OnSearchSelectKey(event)"> |
| </div> |
| |
| <!-- iframe showing the search results (closed by default) --> |
| <div id="MSearchResultsWindow"> |
| <iframe src="javascript:void(0)" frameborder="0" |
| name="MSearchResults" id="MSearchResults"> |
| </iframe> |
| </div> |
| |
| <div class="header"> |
| <div class="headertitle"> |
| <div class="title">NEFuseBatchNormalizationKernel.cpp</div> </div> |
| </div><!--header--> |
| <div class="contents"> |
| <a href="_n_e_fuse_batch_normalization_kernel_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 (c) 2018-2019 ARM Limited.</span></div><div class="line"><a name="l00003"></a><span class="lineno"> 3</span> <span class="comment"> *</span></div><div class="line"><a name="l00004"></a><span class="lineno"> 4</span> <span class="comment"> * SPDX-License-Identifier: MIT</span></div><div class="line"><a name="l00005"></a><span class="lineno"> 5</span> <span class="comment"> *</span></div><div class="line"><a name="l00006"></a><span class="lineno"> 6</span> <span class="comment"> * Permission is hereby granted, free of charge, to any person obtaining a copy</span></div><div class="line"><a name="l00007"></a><span class="lineno"> 7</span> <span class="comment"> * of this software and associated documentation files (the "Software"), to</span></div><div class="line"><a name="l00008"></a><span class="lineno"> 8</span> <span class="comment"> * deal in the Software without restriction, including without limitation the</span></div><div class="line"><a name="l00009"></a><span class="lineno"> 9</span> <span class="comment"> * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or</span></div><div class="line"><a name="l00010"></a><span class="lineno"> 10</span> <span class="comment"> * sell copies of the Software, and to permit persons to whom the Software is</span></div><div class="line"><a name="l00011"></a><span class="lineno"> 11</span> <span class="comment"> * furnished to do so, subject to the following conditions:</span></div><div class="line"><a name="l00012"></a><span class="lineno"> 12</span> <span class="comment"> *</span></div><div class="line"><a name="l00013"></a><span class="lineno"> 13</span> <span class="comment"> * The above copyright notice and this permission notice shall be included in all</span></div><div class="line"><a name="l00014"></a><span class="lineno"> 14</span> <span class="comment"> * copies or substantial portions of the Software.</span></div><div class="line"><a name="l00015"></a><span class="lineno"> 15</span> <span class="comment"> *</span></div><div class="line"><a name="l00016"></a><span class="lineno"> 16</span> <span class="comment"> * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR</span></div><div class="line"><a name="l00017"></a><span class="lineno"> 17</span> <span class="comment"> * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,</span></div><div class="line"><a name="l00018"></a><span class="lineno"> 18</span> <span class="comment"> * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE</span></div><div class="line"><a name="l00019"></a><span class="lineno"> 19</span> <span class="comment"> * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER</span></div><div class="line"><a name="l00020"></a><span class="lineno"> 20</span> <span class="comment"> * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,</span></div><div class="line"><a name="l00021"></a><span class="lineno"> 21</span> <span class="comment"> * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE</span></div><div class="line"><a name="l00022"></a><span class="lineno"> 22</span> <span class="comment"> * SOFTWARE.</span></div><div class="line"><a name="l00023"></a><span class="lineno"> 23</span> <span class="comment"> */</span></div><div class="line"><a name="l00024"></a><span class="lineno"> 24</span> <span class="preprocessor">#include "<a class="code" href="_n_e_fuse_batch_normalization_kernel_8h.xhtml">arm_compute/core/NEON/kernels/NEFuseBatchNormalizationKernel.h</a>"</span></div><div class="line"><a name="l00025"></a><span class="lineno"> 25</span> </div><div class="line"><a name="l00026"></a><span class="lineno"> 26</span> <span class="preprocessor">#include "<a class="code" href="_c_p_p_2_validate_8h.xhtml">arm_compute/core/CPP/Validate.h</a>"</span></div><div class="line"><a name="l00027"></a><span class="lineno"> 27</span> <span class="preprocessor">#include "<a class="code" href="arm__compute_2core_2_helpers_8h.xhtml">arm_compute/core/Helpers.h</a>"</span></div><div class="line"><a name="l00028"></a><span class="lineno"> 28</span> <span class="preprocessor">#include "<a class="code" href="_i_tensor_8h.xhtml">arm_compute/core/ITensor.h</a>"</span></div><div class="line"><a name="l00029"></a><span class="lineno"> 29</span> <span class="preprocessor">#include "<a class="code" href="wrapper_8h.xhtml">arm_compute/core/NEON/wrapper/wrapper.h</a>"</span></div><div class="line"><a name="l00030"></a><span class="lineno"> 30</span> <span class="preprocessor">#include "<a class="code" href="_tensor_info_8h.xhtml">arm_compute/core/TensorInfo.h</a>"</span></div><div class="line"><a name="l00031"></a><span class="lineno"> 31</span> <span class="preprocessor">#include "<a class="code" href="arm__compute_2core_2_utils_8h.xhtml">arm_compute/core/Utils.h</a>"</span></div><div class="line"><a name="l00032"></a><span class="lineno"> 32</span> <span class="preprocessor">#include "<a class="code" href="_validate_8h.xhtml">arm_compute/core/Validate.h</a>"</span></div><div class="line"><a name="l00033"></a><span class="lineno"> 33</span> <span class="preprocessor">#include "<a class="code" href="_window_8h.xhtml">arm_compute/core/Window.h</a>"</span></div><div class="line"><a name="l00034"></a><span class="lineno"> 34</span> <span class="preprocessor">#include "<a class="code" href="_toolchain_support_8h.xhtml">support/ToolchainSupport.h</a>"</span></div><div class="line"><a name="l00035"></a><span class="lineno"> 35</span> </div><div class="line"><a name="l00036"></a><span class="lineno"> 36</span> <span class="preprocessor">#include "<a class="code" href="utils_2_type_printer_8h.xhtml">utils/TypePrinter.h</a>"</span></div><div class="line"><a name="l00037"></a><span class="lineno"> 37</span> <span class="preprocessor">#include <map></span></div><div class="line"><a name="l00038"></a><span class="lineno"> 38</span> </div><div class="line"><a name="l00039"></a><span class="lineno"> 39</span> <span class="keyword">namespace </span><a class="code" href="namespacearm__compute.xhtml">arm_compute</a></div><div class="line"><a name="l00040"></a><span class="lineno"> 40</span> {</div><div class="line"><a name="l00041"></a><span class="lineno"> 41</span> <span class="keyword">namespace</span></div><div class="line"><a name="l00042"></a><span class="lineno"> 42</span> {</div><div class="line"><a name="l00043"></a><span class="lineno"> 43</span> Status validate_arguments(<span class="keyword">const</span> ITensorInfo *input_weights, <span class="keyword">const</span> ITensorInfo *bn_mean, <span class="keyword">const</span> ITensorInfo *bn_var,</div><div class="line"><a name="l00044"></a><span class="lineno"> 44</span>  <span class="keyword">const</span> ITensorInfo *fused_weights, <span class="keyword">const</span> ITensorInfo *fused_bias,</div><div class="line"><a name="l00045"></a><span class="lineno"> 45</span>  <span class="keyword">const</span> ITensorInfo *input_bias, <span class="keyword">const</span> ITensorInfo *bn_beta, <span class="keyword">const</span> ITensorInfo *bn_gamma,</div><div class="line"><a name="l00046"></a><span class="lineno"> 46</span>  <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <a class="code" href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0ca">FuseBatchNormalizationType</a> fbn_type)</div><div class="line"><a name="l00047"></a><span class="lineno"> 47</span> {</div><div class="line"><a name="l00048"></a><span class="lineno"> 48</span>  <a class="code" href="_error_8h.xhtml#a6dc630a6ae9cc063b3924bcea8dee9d6">ARM_COMPUTE_UNUSED</a>(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>);</div><div class="line"><a name="l00049"></a><span class="lineno"> 49</span>  <a class="code" href="_validate_8h.xhtml#aff911654521523937ff24372a870b89f">ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR</a>(input_weights, bn_mean, bn_var);</div><div class="line"><a name="l00050"></a><span class="lineno"> 50</span>  <a class="code" href="_c_p_p_2_validate_8h.xhtml#ad2633f3560322e1f8d926949dec1b730">ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED</a>(input_weights);</div><div class="line"><a name="l00051"></a><span class="lineno"> 51</span>  <a class="code" href="_validate_8h.xhtml#ae7eed178dac535c6e727061b1f5bc6eb">ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN</a>(input_weights, 1, <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a56d8353718e6fdc78b8d69078a2cdb94">DataType::F16</a>, <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">DataType::F32</a>);</div><div class="line"><a name="l00052"></a><span class="lineno"> 52</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(bn_mean, bn_var);</div><div class="line"><a name="l00053"></a><span class="lineno"> 53</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, bn_mean, bn_var);</div><div class="line"><a name="l00054"></a><span class="lineno"> 54</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(input_bias == <span class="keyword">nullptr</span> && fused_bias == <span class="keyword">nullptr</span>);</div><div class="line"><a name="l00055"></a><span class="lineno"> 55</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(bn_mean->num_dimensions() > 1);</div><div class="line"><a name="l00056"></a><span class="lineno"> 56</span> </div><div class="line"><a name="l00057"></a><span class="lineno"> 57</span>  <span class="keywordflow">if</span>(fbn_type == <a class="code" href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0caaf7fb587bad912d390b57a5d6e7ccc55d">FuseBatchNormalizationType::CONVOLUTION</a>)</div><div class="line"><a name="l00058"></a><span class="lineno"> 58</span>  {</div><div class="line"><a name="l00059"></a><span class="lineno"> 59</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(input_weights->dimension(3) != bn_mean->dimension(0));</div><div class="line"><a name="l00060"></a><span class="lineno"> 60</span>  }</div><div class="line"><a name="l00061"></a><span class="lineno"> 61</span>  <span class="keywordflow">else</span></div><div class="line"><a name="l00062"></a><span class="lineno"> 62</span>  {</div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span>  <span class="keyword">const</span> <span class="keywordtype">size_t</span> channel_idx = <a class="code" href="namespacearm__compute.xhtml#a46e938020a3ac8c926d0590b7fe957db">get_data_layout_dimension_index</a>(input_weights->data_layout(), <a class="code" href="namespacearm__compute.xhtml#a74ce3f7420453d3446218ff3b7453e02af52e9c50a060add65a035429b2a22229">DataLayoutDimension::CHANNEL</a>);</div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(input_weights->dimension(channel_idx) != bn_mean->dimension(0));</div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span>  }</div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span>  <span class="comment">// Validate bias</span></div><div class="line"><a name="l00067"></a><span class="lineno"> 67</span>  <span class="keywordflow">if</span>(input_bias != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00068"></a><span class="lineno"> 68</span>  {</div><div class="line"><a name="l00069"></a><span class="lineno"> 69</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(bn_mean, input_bias);</div><div class="line"><a name="l00070"></a><span class="lineno"> 70</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, input_bias);</div><div class="line"><a name="l00071"></a><span class="lineno"> 71</span>  }</div><div class="line"><a name="l00072"></a><span class="lineno"> 72</span>  <span class="comment">// Validate beta</span></div><div class="line"><a name="l00073"></a><span class="lineno"> 73</span>  <span class="keywordflow">if</span>(bn_beta != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00074"></a><span class="lineno"> 74</span>  {</div><div class="line"><a name="l00075"></a><span class="lineno"> 75</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(bn_mean, bn_beta);</div><div class="line"><a name="l00076"></a><span class="lineno"> 76</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, bn_beta);</div><div class="line"><a name="l00077"></a><span class="lineno"> 77</span>  }</div><div class="line"><a name="l00078"></a><span class="lineno"> 78</span>  <span class="comment">// Validate gamma</span></div><div class="line"><a name="l00079"></a><span class="lineno"> 79</span>  <span class="keywordflow">if</span>(bn_gamma != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00080"></a><span class="lineno"> 80</span>  {</div><div class="line"><a name="l00081"></a><span class="lineno"> 81</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(bn_mean, bn_gamma);</div><div class="line"><a name="l00082"></a><span class="lineno"> 82</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, bn_gamma);</div><div class="line"><a name="l00083"></a><span class="lineno"> 83</span>  }</div><div class="line"><a name="l00084"></a><span class="lineno"> 84</span> </div><div class="line"><a name="l00085"></a><span class="lineno"> 85</span>  <span class="comment">// Validate output weights</span></div><div class="line"><a name="l00086"></a><span class="lineno"> 86</span>  <span class="keywordflow">if</span>(fused_weights != <span class="keyword">nullptr</span> && fused_weights->total_size() != 0)</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>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(input_weights, fused_weights);</div><div class="line"><a name="l00089"></a><span class="lineno"> 89</span>  <a class="code" href="_validate_8h.xhtml#abdb9168800c70e5e2c4c020a3b905738">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT</a>(input_weights, fused_weights);</div><div class="line"><a name="l00090"></a><span class="lineno"> 90</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, fused_weights);</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>  <span class="comment">// Validate output bias</span></div><div class="line"><a name="l00093"></a><span class="lineno"> 93</span>  <span class="keywordflow">if</span>(fused_bias != <span class="keyword">nullptr</span> && fused_bias->total_size() != 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>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(bn_mean, fused_bias);</div><div class="line"><a name="l00096"></a><span class="lineno"> 96</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(input_weights, fused_bias);</div><div class="line"><a name="l00097"></a><span class="lineno"> 97</span>  }</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="keywordflow">return</span> Status{};</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> </div><div class="line"><a name="l00102"></a><span class="lineno"> 102</span> <span class="keyword">template</span> <<span class="keyword">typename</span> VectorType></div><div class="line"><a name="l00103"></a><span class="lineno"> 103</span> <span class="keywordtype">void</span> fused_batch_normalization_conv(<span class="keyword">const</span> ITensor *conv_weights, <span class="keyword">const</span> ITensor *conv_bias, ITensor *fused_weights, ITensor *fused_bias,</div><div class="line"><a name="l00104"></a><span class="lineno"> 104</span>  <span class="keyword">const</span> ITensor *bn_mean, <span class="keyword">const</span> ITensor *bn_var, <span class="keyword">const</span> ITensor *bn_beta, <span class="keyword">const</span> ITensor *bn_gamma, <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <span class="keyword">const</span> Window &window)</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>  <span class="keyword">using</span> ScalarType = <span class="keyword">typename</span> VectorType::scalar_type;</div><div class="line"><a name="l00107"></a><span class="lineno"> 107</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> size = 16 / conv_weights->info()->element_size();</div><div class="line"><a name="l00108"></a><span class="lineno"> 108</span>  <span class="keyword">using</span> ExactTagType = <span class="keyword">typename</span> VectorType::tag_type;</div><div class="line"><a name="l00109"></a><span class="lineno"> 109</span> </div><div class="line"><a name="l00110"></a><span class="lineno"> 110</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_weights = (fused_weights == <span class="keyword">nullptr</span>) || (fused_weights == conv_weights);</div><div class="line"><a name="l00111"></a><span class="lineno"> 111</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_bias = (fused_bias == <span class="keyword">nullptr</span>) || (conv_bias != <span class="keyword">nullptr</span> && fused_bias == conv_bias);</div><div class="line"><a name="l00112"></a><span class="lineno"> 112</span> </div><div class="line"><a name="l00113"></a><span class="lineno"> 113</span>  <span class="comment">// Set build options</span></div><div class="line"><a name="l00114"></a><span class="lineno"> 114</span>  Window win = window;</div><div class="line"><a name="l00115"></a><span class="lineno"> 115</span>  win.set(<a class="code" href="classarm__compute_1_1_window.xhtml#aa96e81276ee4f87ab386cd05a5539a7d">Window::DimX</a>, Window::Dimension(0, 1, 1));</div><div class="line"><a name="l00116"></a><span class="lineno"> 116</span> </div><div class="line"><a name="l00117"></a><span class="lineno"> 117</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> window_step_x = size;</div><div class="line"><a name="l00118"></a><span class="lineno"> 118</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_start_x = static_cast<int>(window.x().start());</div><div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_end_x = static_cast<int>(window.x().end());</div><div class="line"><a name="l00120"></a><span class="lineno"> 120</span> </div><div class="line"><a name="l00121"></a><span class="lineno"> 121</span>  Iterator conv_w_in(conv_weights, win);</div><div class="line"><a name="l00122"></a><span class="lineno"> 122</span>  Iterator conv_w_out(run_in_place_weights ? conv_weights : fused_weights, win);</div><div class="line"><a name="l00123"></a><span class="lineno"> 123</span> </div><div class="line"><a name="l00124"></a><span class="lineno"> 124</span>  <span class="keyword">const</span> <span class="keyword">auto</span> conv_bias_in = (conv_bias != <span class="keyword">nullptr</span> ? reinterpret_cast<ScalarType *>(conv_bias->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>);</div><div class="line"><a name="l00125"></a><span class="lineno"> 125</span>  <span class="keyword">auto</span> conv_bias_out = (run_in_place_bias ? conv_bias_in : reinterpret_cast<ScalarType *>(fused_bias->ptr_to_element(Coordinates(0, 0))));</div><div class="line"><a name="l00126"></a><span class="lineno"> 126</span> </div><div class="line"><a name="l00127"></a><span class="lineno"> 127</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_mean = reinterpret_cast<const ScalarType *>(bn_mean->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00128"></a><span class="lineno"> 128</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_var = reinterpret_cast<const ScalarType *>(bn_var->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00129"></a><span class="lineno"> 129</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_gamma = (bn_gamma != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_gamma->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</div><div class="line"><a name="l00130"></a><span class="lineno"> 130</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_beta = (bn_beta != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_beta->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</div><div class="line"><a name="l00131"></a><span class="lineno"> 131</span> </div><div class="line"><a name="l00132"></a><span class="lineno"> 132</span>  <span class="keyword">auto</span> mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00133"></a><span class="lineno"> 133</span>  <span class="keyword">auto</span> var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00134"></a><span class="lineno"> 134</span>  <span class="keyword">auto</span> gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(1), ExactTagType{});</div><div class="line"><a name="l00135"></a><span class="lineno"> 135</span>  <span class="keyword">auto</span> beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00136"></a><span class="lineno"> 136</span>  <span class="keyword">auto</span> rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00137"></a><span class="lineno"> 137</span>  <span class="keyword">const</span> <span class="keyword">auto</span> epsilon_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>), ExactTagType{});</div><div class="line"><a name="l00138"></a><span class="lineno"> 138</span> </div><div class="line"><a name="l00139"></a><span class="lineno"> 139</span>  <span class="keyword">auto</span> mean = ScalarType(0.0);</div><div class="line"><a name="l00140"></a><span class="lineno"> 140</span>  <span class="keyword">auto</span> var = ScalarType(0.0);</div><div class="line"><a name="l00141"></a><span class="lineno"> 141</span>  <span class="keyword">auto</span> gamma = ScalarType(1.0);</div><div class="line"><a name="l00142"></a><span class="lineno"> 142</span>  <span class="keyword">auto</span> beta = ScalarType(0.0);</div><div class="line"><a name="l00143"></a><span class="lineno"> 143</span>  <span class="keyword">auto</span> conv_bias_in_scalar = ScalarType(0.0);</div><div class="line"><a name="l00144"></a><span class="lineno"> 144</span>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(win, [&](<span class="keyword">const</span> Coordinates & <span class="keywordtype">id</span>)</div><div class="line"><a name="l00145"></a><span class="lineno"> 145</span>  {</div><div class="line"><a name="l00146"></a><span class="lineno"> 146</span>  var = input_var[<span class="keywordtype">id</span>[3]];</div><div class="line"><a name="l00147"></a><span class="lineno"> 147</span>  <span class="keywordflow">if</span>(input_gamma != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00148"></a><span class="lineno"> 148</span>  {</div><div class="line"><a name="l00149"></a><span class="lineno"> 149</span>  gamma = input_gamma[<span class="keywordtype">id</span>[3]];</div><div class="line"><a name="l00150"></a><span class="lineno"> 150</span>  }</div><div class="line"><a name="l00151"></a><span class="lineno"> 151</span> </div><div class="line"><a name="l00152"></a><span class="lineno"> 152</span>  <span class="keywordflow">if</span>((<span class="keywordtype">id</span>[0] == 0) && (<span class="keywordtype">id</span>[1] == 0) && (<span class="keywordtype">id</span>[2] == 0))</div><div class="line"><a name="l00153"></a><span class="lineno"> 153</span>  {</div><div class="line"><a name="l00154"></a><span class="lineno"> 154</span>  <span class="keywordflow">if</span>(input_beta != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00155"></a><span class="lineno"> 155</span>  {</div><div class="line"><a name="l00156"></a><span class="lineno"> 156</span>  beta = input_beta[<span class="keywordtype">id</span>[3]];</div><div class="line"><a name="l00157"></a><span class="lineno"> 157</span>  beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(beta, ExactTagType{});</div><div class="line"><a name="l00158"></a><span class="lineno"> 158</span>  }</div><div class="line"><a name="l00159"></a><span class="lineno"> 159</span> </div><div class="line"><a name="l00160"></a><span class="lineno"> 160</span>  <span class="comment">// Construct vectors</span></div><div class="line"><a name="l00161"></a><span class="lineno"> 161</span>  mean = input_mean[<span class="keywordtype">id</span>[3]];</div><div class="line"><a name="l00162"></a><span class="lineno"> 162</span>  mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(mean, ExactTagType{});</div><div class="line"><a name="l00163"></a><span class="lineno"> 163</span> </div><div class="line"><a name="l00164"></a><span class="lineno"> 164</span>  <span class="keywordflow">if</span>(conv_bias_in != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00165"></a><span class="lineno"> 165</span>  {</div><div class="line"><a name="l00166"></a><span class="lineno"> 166</span>  conv_bias_in_scalar = conv_bias_in[<span class="keywordtype">id</span>[3]];</div><div class="line"><a name="l00167"></a><span class="lineno"> 167</span>  }</div><div class="line"><a name="l00168"></a><span class="lineno"> 168</span>  <span class="keyword">auto</span> conv_bias_tmp_scalar = (conv_bias_in_scalar - mean) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>));</div><div class="line"><a name="l00169"></a><span class="lineno"> 169</span>  conv_bias_out[<span class="keywordtype">id</span>[3]] = (conv_bias_tmp_scalar * gamma) + beta;</div><div class="line"><a name="l00170"></a><span class="lineno"> 170</span>  }</div><div class="line"><a name="l00171"></a><span class="lineno"> 171</span> </div><div class="line"><a name="l00172"></a><span class="lineno"> 172</span>  <span class="keywordtype">int</span> x = window_start_x;</div><div class="line"><a name="l00173"></a><span class="lineno"> 173</span>  <span class="keyword">auto</span> conv_w_in_ptr = reinterpret_cast<const ScalarType *>(conv_w_in.ptr());</div><div class="line"><a name="l00174"></a><span class="lineno"> 174</span>  <span class="keyword">auto</span> conv_w_out_ptr = reinterpret_cast<ScalarType *>(conv_w_out.ptr());</div><div class="line"><a name="l00175"></a><span class="lineno"> 175</span>  var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(var, ExactTagType{});</div><div class="line"><a name="l00176"></a><span class="lineno"> 176</span>  gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(gamma, ExactTagType{});</div><div class="line"><a name="l00177"></a><span class="lineno"> 177</span>  rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a569988c8e1a2619532b8f4d0394d5d79">wrapper::vinvsqrt</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(var_vec, epsilon_vec));</div><div class="line"><a name="l00178"></a><span class="lineno"> 178</span> </div><div class="line"><a name="l00179"></a><span class="lineno"> 179</span>  <span class="keywordflow">for</span>(; x <= (window_end_x - window_step_x); x += window_step_x)</div><div class="line"><a name="l00180"></a><span class="lineno"> 180</span>  {</div><div class="line"><a name="l00181"></a><span class="lineno"> 181</span>  <span class="keyword">auto</span> wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(conv_w_in_ptr + x);</div><div class="line"><a name="l00182"></a><span class="lineno"> 182</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, rvar_vec);</div><div class="line"><a name="l00183"></a><span class="lineno"> 183</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, gamma_vec);</div><div class="line"><a name="l00184"></a><span class="lineno"> 184</span> </div><div class="line"><a name="l00185"></a><span class="lineno"> 185</span>  <span class="comment">// Store results</span></div><div class="line"><a name="l00186"></a><span class="lineno"> 186</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(conv_w_out_ptr + x, wn);</div><div class="line"><a name="l00187"></a><span class="lineno"> 187</span>  }</div><div class="line"><a name="l00188"></a><span class="lineno"> 188</span> </div><div class="line"><a name="l00189"></a><span class="lineno"> 189</span>  <span class="comment">// Compute left-over elements</span></div><div class="line"><a name="l00190"></a><span class="lineno"> 190</span>  <span class="keywordflow">for</span>(; x < window_end_x; ++x)</div><div class="line"><a name="l00191"></a><span class="lineno"> 191</span>  {</div><div class="line"><a name="l00192"></a><span class="lineno"> 192</span>  *(conv_w_out_ptr + x) = *(conv_w_in_ptr + x) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>)) * gamma;</div><div class="line"><a name="l00193"></a><span class="lineno"> 193</span>  }</div><div class="line"><a name="l00194"></a><span class="lineno"> 194</span>  },</div><div class="line"><a name="l00195"></a><span class="lineno"> 195</span>  conv_w_in, conv_w_out);</div><div class="line"><a name="l00196"></a><span class="lineno"> 196</span> }</div><div class="line"><a name="l00197"></a><span class="lineno"> 197</span> </div><div class="line"><a name="l00198"></a><span class="lineno"> 198</span> <span class="keyword">template</span> <<span class="keyword">typename</span> VectorType></div><div class="line"><a name="l00199"></a><span class="lineno"> 199</span> <span class="keywordtype">void</span> fused_batch_normalization_dwc_nhwc(<span class="keyword">const</span> ITensor *dwc_weights, <span class="keyword">const</span> ITensor *dwc_bias, ITensor *fused_weights, ITensor *fused_bias,</div><div class="line"><a name="l00200"></a><span class="lineno"> 200</span>  <span class="keyword">const</span> ITensor *bn_mean, <span class="keyword">const</span> ITensor *bn_var, <span class="keyword">const</span> ITensor *bn_beta, <span class="keyword">const</span> ITensor *bn_gamma, <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <span class="keyword">const</span> Window &window)</div><div class="line"><a name="l00201"></a><span class="lineno"> 201</span> {</div><div class="line"><a name="l00202"></a><span class="lineno"> 202</span>  <span class="keyword">using</span> ScalarType = <span class="keyword">typename</span> VectorType::scalar_type;</div><div class="line"><a name="l00203"></a><span class="lineno"> 203</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> size = 16 / dwc_weights->info()->element_size();</div><div class="line"><a name="l00204"></a><span class="lineno"> 204</span>  <span class="keyword">using</span> ExactTagType = <span class="keyword">typename</span> VectorType::tag_type;</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>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_weights = (fused_weights == <span class="keyword">nullptr</span>) || (fused_weights == dwc_weights);</div><div class="line"><a name="l00207"></a><span class="lineno"> 207</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_bias = (fused_bias == <span class="keyword">nullptr</span>) || (dwc_bias != <span class="keyword">nullptr</span> && fused_bias == dwc_bias);</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>  <span class="comment">// Set build options</span></div><div class="line"><a name="l00210"></a><span class="lineno"> 210</span>  Window win = window;</div><div class="line"><a name="l00211"></a><span class="lineno"> 211</span>  win.set(<a class="code" href="classarm__compute_1_1_window.xhtml#aa96e81276ee4f87ab386cd05a5539a7d">Window::DimX</a>, Window::Dimension(0, 1, 1));</div><div class="line"><a name="l00212"></a><span class="lineno"> 212</span> </div><div class="line"><a name="l00213"></a><span class="lineno"> 213</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> window_step_x = size;</div><div class="line"><a name="l00214"></a><span class="lineno"> 214</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_start_x = static_cast<int>(window.x().start());</div><div class="line"><a name="l00215"></a><span class="lineno"> 215</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_end_x = static_cast<int>(window.x().end());</div><div class="line"><a name="l00216"></a><span class="lineno"> 216</span> </div><div class="line"><a name="l00217"></a><span class="lineno"> 217</span>  Iterator dwc_w_in(dwc_weights, win);</div><div class="line"><a name="l00218"></a><span class="lineno"> 218</span>  Iterator dwc_w_out(run_in_place_weights ? dwc_weights : fused_weights, win);</div><div class="line"><a name="l00219"></a><span class="lineno"> 219</span> </div><div class="line"><a name="l00220"></a><span class="lineno"> 220</span>  <span class="keyword">const</span> <span class="keyword">auto</span> dwc_bias_in = (dwc_bias != <span class="keyword">nullptr</span> ? reinterpret_cast<ScalarType *>(dwc_bias->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>);</div><div class="line"><a name="l00221"></a><span class="lineno"> 221</span>  <span class="keyword">auto</span> dwc_bias_out = (run_in_place_bias ? dwc_bias_in : reinterpret_cast<ScalarType *>(fused_bias->ptr_to_element(Coordinates(0, 0))));</div><div class="line"><a name="l00222"></a><span class="lineno"> 222</span> </div><div class="line"><a name="l00223"></a><span class="lineno"> 223</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_mean = reinterpret_cast<const ScalarType *>(bn_mean->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00224"></a><span class="lineno"> 224</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_var = reinterpret_cast<const ScalarType *>(bn_var->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00225"></a><span class="lineno"> 225</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_gamma = (bn_gamma != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_gamma->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</div><div class="line"><a name="l00226"></a><span class="lineno"> 226</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_beta = (bn_beta != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_beta->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</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="keyword">auto</span> mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00229"></a><span class="lineno"> 229</span>  <span class="keyword">auto</span> var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00230"></a><span class="lineno"> 230</span>  <span class="keyword">auto</span> gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(1), ExactTagType{});</div><div class="line"><a name="l00231"></a><span class="lineno"> 231</span>  <span class="keyword">auto</span> beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00232"></a><span class="lineno"> 232</span>  <span class="keyword">auto</span> rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00233"></a><span class="lineno"> 233</span>  <span class="keyword">auto</span> dwc_bias_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00234"></a><span class="lineno"> 234</span>  <span class="keyword">const</span> <span class="keyword">auto</span> epsilon_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>), ExactTagType{});</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>  <span class="keyword">auto</span> gamma = ScalarType(1.0);</div><div class="line"><a name="l00237"></a><span class="lineno"> 237</span>  <span class="keyword">auto</span> beta = ScalarType(0.0);</div><div class="line"><a name="l00238"></a><span class="lineno"> 238</span>  <span class="keyword">auto</span> dwc_bias_in_scalar = ScalarType(0);</div><div class="line"><a name="l00239"></a><span class="lineno"> 239</span> </div><div class="line"><a name="l00240"></a><span class="lineno"> 240</span>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(win, [&](<span class="keyword">const</span> Coordinates & <span class="keywordtype">id</span>)</div><div class="line"><a name="l00241"></a><span class="lineno"> 241</span>  {</div><div class="line"><a name="l00242"></a><span class="lineno"> 242</span>  <span class="keywordtype">int</span> x = window_start_x;</div><div class="line"><a name="l00243"></a><span class="lineno"> 243</span>  <span class="keywordflow">for</span>(; x <= (window_end_x - window_step_x); x += window_step_x)</div><div class="line"><a name="l00244"></a><span class="lineno"> 244</span>  {</div><div class="line"><a name="l00245"></a><span class="lineno"> 245</span>  var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(input_var + x);</div><div class="line"><a name="l00246"></a><span class="lineno"> 246</span>  <span class="keywordflow">if</span>(input_gamma != <span class="keyword">nullptr</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>  gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(input_gamma + x);</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> </div><div class="line"><a name="l00251"></a><span class="lineno"> 251</span>  <span class="keywordflow">if</span>((<span class="keywordtype">id</span>[2] == 0) && (<span class="keywordtype">id</span>[1] == 0))</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>  mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(input_mean + x);</div><div class="line"><a name="l00254"></a><span class="lineno"> 254</span> </div><div class="line"><a name="l00255"></a><span class="lineno"> 255</span>  <span class="comment">// Construct vectors</span></div><div class="line"><a name="l00256"></a><span class="lineno"> 256</span>  <span class="keywordflow">if</span>(input_beta != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00257"></a><span class="lineno"> 257</span>  {</div><div class="line"><a name="l00258"></a><span class="lineno"> 258</span>  beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(input_beta + x);</div><div class="line"><a name="l00259"></a><span class="lineno"> 259</span>  }</div><div class="line"><a name="l00260"></a><span class="lineno"> 260</span> </div><div class="line"><a name="l00261"></a><span class="lineno"> 261</span>  <span class="keywordflow">if</span>(dwc_bias_in != <span class="keyword">nullptr</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>  dwc_bias_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(dwc_bias_in + x);</div><div class="line"><a name="l00264"></a><span class="lineno"> 264</span>  }</div><div class="line"><a name="l00265"></a><span class="lineno"> 265</span> </div><div class="line"><a name="l00266"></a><span class="lineno"> 266</span>  <span class="keyword">auto</span> dwc_bias_tmp_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(dwc_bias_vec, mean_vec), <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a569988c8e1a2619532b8f4d0394d5d79">wrapper::vinvsqrt</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(var_vec, epsilon_vec)));</div><div class="line"><a name="l00267"></a><span class="lineno"> 267</span>  dwc_bias_tmp_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(dwc_bias_tmp_vec, gamma_vec), beta_vec);</div><div class="line"><a name="l00268"></a><span class="lineno"> 268</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(dwc_bias_out + x, dwc_bias_tmp_vec);</div><div class="line"><a name="l00269"></a><span class="lineno"> 269</span>  }</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>  <span class="keyword">auto</span> dwc_w_in_ptr = reinterpret_cast<const ScalarType *>(dwc_w_in.ptr());</div><div class="line"><a name="l00272"></a><span class="lineno"> 272</span>  <span class="keyword">auto</span> dwc_w_out_ptr = reinterpret_cast<ScalarType *>(dwc_w_out.ptr());</div><div class="line"><a name="l00273"></a><span class="lineno"> 273</span> </div><div class="line"><a name="l00274"></a><span class="lineno"> 274</span>  <span class="keyword">auto</span> wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(dwc_w_in_ptr + x);</div><div class="line"><a name="l00275"></a><span class="lineno"> 275</span>  rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a569988c8e1a2619532b8f4d0394d5d79">wrapper::vinvsqrt</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(var_vec, epsilon_vec));</div><div class="line"><a name="l00276"></a><span class="lineno"> 276</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, rvar_vec);</div><div class="line"><a name="l00277"></a><span class="lineno"> 277</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, gamma_vec);</div><div class="line"><a name="l00278"></a><span class="lineno"> 278</span> </div><div class="line"><a name="l00279"></a><span class="lineno"> 279</span>  <span class="comment">// Store results</span></div><div class="line"><a name="l00280"></a><span class="lineno"> 280</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(dwc_w_out_ptr + x, wn);</div><div class="line"><a name="l00281"></a><span class="lineno"> 281</span>  }</div><div class="line"><a name="l00282"></a><span class="lineno"> 282</span> </div><div class="line"><a name="l00283"></a><span class="lineno"> 283</span>  <span class="comment">// Compute left-over elements</span></div><div class="line"><a name="l00284"></a><span class="lineno"> 284</span>  <span class="keywordflow">for</span>(; x < window_end_x; ++x)</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>  <span class="keyword">auto</span> var = input_var[x];</div><div class="line"><a name="l00287"></a><span class="lineno"> 287</span>  <span class="keywordflow">if</span>(input_gamma != <span class="keyword">nullptr</span>)</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>  gamma = input_gamma[x];</div><div class="line"><a name="l00290"></a><span class="lineno"> 290</span>  }</div><div class="line"><a name="l00291"></a><span class="lineno"> 291</span> </div><div class="line"><a name="l00292"></a><span class="lineno"> 292</span>  <span class="keywordflow">if</span>(<span class="keywordtype">id</span>[2] == 0 && <span class="keywordtype">id</span>[1] == 0)</div><div class="line"><a name="l00293"></a><span class="lineno"> 293</span>  {</div><div class="line"><a name="l00294"></a><span class="lineno"> 294</span>  <span class="keyword">auto</span> mean = input_mean[x];</div><div class="line"><a name="l00295"></a><span class="lineno"> 295</span>  <span class="keywordflow">if</span>(input_beta != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00296"></a><span class="lineno"> 296</span>  {</div><div class="line"><a name="l00297"></a><span class="lineno"> 297</span>  beta = input_beta[x];</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="keywordflow">if</span>(dwc_bias_in != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00300"></a><span class="lineno"> 300</span>  {</div><div class="line"><a name="l00301"></a><span class="lineno"> 301</span>  dwc_bias_in_scalar = dwc_bias_in[x];</div><div class="line"><a name="l00302"></a><span class="lineno"> 302</span>  }</div><div class="line"><a name="l00303"></a><span class="lineno"> 303</span> </div><div class="line"><a name="l00304"></a><span class="lineno"> 304</span>  <span class="keyword">auto</span> dwc_bias_tmp_scalar = (dwc_bias_in_scalar - mean) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>));</div><div class="line"><a name="l00305"></a><span class="lineno"> 305</span>  dwc_bias_out[x] = (dwc_bias_tmp_scalar * gamma) + beta;</div><div class="line"><a name="l00306"></a><span class="lineno"> 306</span>  }</div><div class="line"><a name="l00307"></a><span class="lineno"> 307</span> </div><div class="line"><a name="l00308"></a><span class="lineno"> 308</span>  <span class="keyword">const</span> <span class="keyword">auto</span> dwc_w_in_ptr = reinterpret_cast<const ScalarType *>(dwc_w_in.ptr());</div><div class="line"><a name="l00309"></a><span class="lineno"> 309</span>  <span class="keyword">auto</span> dwc_w_out_ptr = reinterpret_cast<ScalarType *>(dwc_w_out.ptr());</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>  *(dwc_w_out_ptr + x) = *(dwc_w_in_ptr + x) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>)) * gamma;</div><div class="line"><a name="l00312"></a><span class="lineno"> 312</span>  }</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>  dwc_w_in, dwc_w_out);</div><div class="line"><a name="l00315"></a><span class="lineno"> 315</span> }</div><div class="line"><a name="l00316"></a><span class="lineno"> 316</span> </div><div class="line"><a name="l00317"></a><span class="lineno"> 317</span> <span class="keyword">template</span> <<span class="keyword">typename</span> VectorType></div><div class="line"><a name="l00318"></a><span class="lineno"> 318</span> <span class="keywordtype">void</span> fused_batch_normalization_dwc_nchw(<span class="keyword">const</span> ITensor *dwc_weights, <span class="keyword">const</span> ITensor *dwc_bias, ITensor *fused_weights, ITensor *fused_bias,</div><div class="line"><a name="l00319"></a><span class="lineno"> 319</span>  <span class="keyword">const</span> ITensor *bn_mean, <span class="keyword">const</span> ITensor *bn_var, <span class="keyword">const</span> ITensor *bn_beta, <span class="keyword">const</span> ITensor *bn_gamma, <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <span class="keyword">const</span> Window &window)</div><div class="line"><a name="l00320"></a><span class="lineno"> 320</span> {</div><div class="line"><a name="l00321"></a><span class="lineno"> 321</span>  <span class="keyword">using</span> ScalarType = <span class="keyword">typename</span> VectorType::scalar_type;</div><div class="line"><a name="l00322"></a><span class="lineno"> 322</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> size = 16 / dwc_weights->info()->element_size();</div><div class="line"><a name="l00323"></a><span class="lineno"> 323</span>  <span class="keyword">using</span> ExactTagType = <span class="keyword">typename</span> VectorType::tag_type;</div><div class="line"><a name="l00324"></a><span class="lineno"> 324</span> </div><div class="line"><a name="l00325"></a><span class="lineno"> 325</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_weights = (fused_weights == <span class="keyword">nullptr</span>) || (fused_weights == dwc_weights);</div><div class="line"><a name="l00326"></a><span class="lineno"> 326</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> run_in_place_bias = (fused_bias == <span class="keyword">nullptr</span>) || (dwc_bias != <span class="keyword">nullptr</span> && fused_bias == dwc_bias);</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>  <span class="comment">// Set build options</span></div><div class="line"><a name="l00329"></a><span class="lineno"> 329</span>  Window win = window;</div><div class="line"><a name="l00330"></a><span class="lineno"> 330</span>  win.set(<a class="code" href="classarm__compute_1_1_window.xhtml#aa96e81276ee4f87ab386cd05a5539a7d">Window::DimX</a>, Window::Dimension(0, 1, 1));</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>  <span class="keyword">const</span> <span class="keywordtype">int</span> window_step_x = size;</div><div class="line"><a name="l00333"></a><span class="lineno"> 333</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_start_x = static_cast<int>(window.x().start());</div><div class="line"><a name="l00334"></a><span class="lineno"> 334</span>  <span class="keyword">const</span> <span class="keyword">auto</span> window_end_x = static_cast<int>(window.x().end());</div><div class="line"><a name="l00335"></a><span class="lineno"> 335</span> </div><div class="line"><a name="l00336"></a><span class="lineno"> 336</span>  Iterator dwc_w_in(dwc_weights, win);</div><div class="line"><a name="l00337"></a><span class="lineno"> 337</span>  Iterator dwc_w_out(run_in_place_weights ? dwc_weights : fused_weights, win);</div><div class="line"><a name="l00338"></a><span class="lineno"> 338</span> </div><div class="line"><a name="l00339"></a><span class="lineno"> 339</span>  <span class="keyword">const</span> <span class="keyword">auto</span> dwc_bias_in = (dwc_bias != <span class="keyword">nullptr</span> ? reinterpret_cast<ScalarType *>(dwc_bias->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>);</div><div class="line"><a name="l00340"></a><span class="lineno"> 340</span>  <span class="keyword">auto</span> dwc_bias_out = (run_in_place_bias ? dwc_bias_in : reinterpret_cast<ScalarType *>(fused_bias->ptr_to_element(Coordinates(0, 0))));</div><div class="line"><a name="l00341"></a><span class="lineno"> 341</span> </div><div class="line"><a name="l00342"></a><span class="lineno"> 342</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_mean = reinterpret_cast<const ScalarType *>(bn_mean->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00343"></a><span class="lineno"> 343</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_var = reinterpret_cast<const ScalarType *>(bn_var->ptr_to_element(Coordinates(0, 0)));</div><div class="line"><a name="l00344"></a><span class="lineno"> 344</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_gamma = (bn_gamma != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_gamma->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</div><div class="line"><a name="l00345"></a><span class="lineno"> 345</span>  <span class="keyword">const</span> <span class="keyword">auto</span> input_beta = (bn_beta != <span class="keyword">nullptr</span>) ? reinterpret_cast<const ScalarType *>(bn_beta->ptr_to_element(Coordinates(0, 0))) : <span class="keyword">nullptr</span>;</div><div class="line"><a name="l00346"></a><span class="lineno"> 346</span> </div><div class="line"><a name="l00347"></a><span class="lineno"> 347</span>  <span class="keyword">auto</span> mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00348"></a><span class="lineno"> 348</span>  <span class="keyword">auto</span> var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00349"></a><span class="lineno"> 349</span>  <span class="keyword">auto</span> gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(1), ExactTagType{});</div><div class="line"><a name="l00350"></a><span class="lineno"> 350</span>  <span class="keyword">auto</span> beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00351"></a><span class="lineno"> 351</span>  <span class="keyword">auto</span> rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(0), ExactTagType{});</div><div class="line"><a name="l00352"></a><span class="lineno"> 352</span>  <span class="keyword">const</span> <span class="keyword">auto</span> epsilon_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>), ExactTagType{});</div><div class="line"><a name="l00353"></a><span class="lineno"> 353</span> </div><div class="line"><a name="l00354"></a><span class="lineno"> 354</span>  <span class="keyword">auto</span> mean = ScalarType(0.0);</div><div class="line"><a name="l00355"></a><span class="lineno"> 355</span>  <span class="keyword">auto</span> var = ScalarType(0.0);</div><div class="line"><a name="l00356"></a><span class="lineno"> 356</span>  <span class="keyword">auto</span> gamma = ScalarType(1.0);</div><div class="line"><a name="l00357"></a><span class="lineno"> 357</span>  <span class="keyword">auto</span> beta = ScalarType(0.0);</div><div class="line"><a name="l00358"></a><span class="lineno"> 358</span>  <span class="keyword">auto</span> dwc_bias_in_scalar = ScalarType(0.0);</div><div class="line"><a name="l00359"></a><span class="lineno"> 359</span>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(win, [&](<span class="keyword">const</span> Coordinates & <span class="keywordtype">id</span>)</div><div class="line"><a name="l00360"></a><span class="lineno"> 360</span>  {</div><div class="line"><a name="l00361"></a><span class="lineno"> 361</span>  var = input_var[<span class="keywordtype">id</span>[2]];</div><div class="line"><a name="l00362"></a><span class="lineno"> 362</span>  <span class="keywordflow">if</span>(input_gamma != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00363"></a><span class="lineno"> 363</span>  {</div><div class="line"><a name="l00364"></a><span class="lineno"> 364</span>  gamma = input_gamma[<span class="keywordtype">id</span>[2]];</div><div class="line"><a name="l00365"></a><span class="lineno"> 365</span>  }</div><div class="line"><a name="l00366"></a><span class="lineno"> 366</span> </div><div class="line"><a name="l00367"></a><span class="lineno"> 367</span>  <span class="keywordflow">if</span>(<span class="keywordtype">id</span>[1] == 0)</div><div class="line"><a name="l00368"></a><span class="lineno"> 368</span>  {</div><div class="line"><a name="l00369"></a><span class="lineno"> 369</span>  mean = input_mean[<span class="keywordtype">id</span>[2]];</div><div class="line"><a name="l00370"></a><span class="lineno"> 370</span> </div><div class="line"><a name="l00371"></a><span class="lineno"> 371</span>  <span class="comment">// Construct vectors</span></div><div class="line"><a name="l00372"></a><span class="lineno"> 372</span>  mean_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(mean, ExactTagType{});</div><div class="line"><a name="l00373"></a><span class="lineno"> 373</span>  <span class="keywordflow">if</span>(input_beta != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00374"></a><span class="lineno"> 374</span>  {</div><div class="line"><a name="l00375"></a><span class="lineno"> 375</span>  beta = input_beta[<span class="keywordtype">id</span>[2]];</div><div class="line"><a name="l00376"></a><span class="lineno"> 376</span>  beta_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(beta, ExactTagType{});</div><div class="line"><a name="l00377"></a><span class="lineno"> 377</span>  }</div><div class="line"><a name="l00378"></a><span class="lineno"> 378</span> </div><div class="line"><a name="l00379"></a><span class="lineno"> 379</span>  <span class="keywordflow">if</span>(dwc_bias_in != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00380"></a><span class="lineno"> 380</span>  {</div><div class="line"><a name="l00381"></a><span class="lineno"> 381</span>  dwc_bias_in_scalar = dwc_bias_in[<span class="keywordtype">id</span>[2]];</div><div class="line"><a name="l00382"></a><span class="lineno"> 382</span>  }</div><div class="line"><a name="l00383"></a><span class="lineno"> 383</span> </div><div class="line"><a name="l00384"></a><span class="lineno"> 384</span>  <span class="keyword">auto</span> dwc_bias_tmp_scalar = (dwc_bias_in_scalar - mean) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>));</div><div class="line"><a name="l00385"></a><span class="lineno"> 385</span>  dwc_bias_out[<span class="keywordtype">id</span>[2]] = (dwc_bias_tmp_scalar * gamma) + beta;</div><div class="line"><a name="l00386"></a><span class="lineno"> 386</span>  }</div><div class="line"><a name="l00387"></a><span class="lineno"> 387</span> </div><div class="line"><a name="l00388"></a><span class="lineno"> 388</span>  <span class="keywordtype">int</span> x = window_start_x;</div><div class="line"><a name="l00389"></a><span class="lineno"> 389</span>  <span class="keyword">auto</span> dwc_w_in_ptr = reinterpret_cast<const ScalarType *>(dwc_w_in.ptr());</div><div class="line"><a name="l00390"></a><span class="lineno"> 390</span>  <span class="keyword">auto</span> dwc_w_out_ptr = reinterpret_cast<ScalarType *>(dwc_w_out.ptr());</div><div class="line"><a name="l00391"></a><span class="lineno"> 391</span>  var_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(var, ExactTagType{});</div><div class="line"><a name="l00392"></a><span class="lineno"> 392</span>  gamma_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(gamma, ExactTagType{});</div><div class="line"><a name="l00393"></a><span class="lineno"> 393</span>  rvar_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a569988c8e1a2619532b8f4d0394d5d79">wrapper::vinvsqrt</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(var_vec, epsilon_vec));</div><div class="line"><a name="l00394"></a><span class="lineno"> 394</span> </div><div class="line"><a name="l00395"></a><span class="lineno"> 395</span>  <span class="keywordflow">for</span>(; x <= (window_end_x - window_step_x); x += window_step_x)</div><div class="line"><a name="l00396"></a><span class="lineno"> 396</span>  {</div><div class="line"><a name="l00397"></a><span class="lineno"> 397</span>  <span class="keyword">auto</span> wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(dwc_w_in_ptr + x);</div><div class="line"><a name="l00398"></a><span class="lineno"> 398</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, rvar_vec);</div><div class="line"><a name="l00399"></a><span class="lineno"> 399</span>  wn = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(wn, gamma_vec);</div><div class="line"><a name="l00400"></a><span class="lineno"> 400</span> </div><div class="line"><a name="l00401"></a><span class="lineno"> 401</span>  <span class="comment">// Store results</span></div><div class="line"><a name="l00402"></a><span class="lineno"> 402</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(dwc_w_out_ptr + x, wn);</div><div class="line"><a name="l00403"></a><span class="lineno"> 403</span>  }</div><div class="line"><a name="l00404"></a><span class="lineno"> 404</span> </div><div class="line"><a name="l00405"></a><span class="lineno"> 405</span>  <span class="comment">// Compute left-over elements</span></div><div class="line"><a name="l00406"></a><span class="lineno"> 406</span>  <span class="keywordflow">for</span>(; x < window_end_x; ++x)</div><div class="line"><a name="l00407"></a><span class="lineno"> 407</span>  {</div><div class="line"><a name="l00408"></a><span class="lineno"> 408</span>  *(dwc_w_out_ptr + x) = *(dwc_w_in_ptr + x) / std::sqrt(var + ScalarType(<a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>)) * gamma;</div><div class="line"><a name="l00409"></a><span class="lineno"> 409</span>  }</div><div class="line"><a name="l00410"></a><span class="lineno"> 410</span>  },</div><div class="line"><a name="l00411"></a><span class="lineno"> 411</span>  dwc_w_in, dwc_w_out);</div><div class="line"><a name="l00412"></a><span class="lineno"> 412</span> }</div><div class="line"><a name="l00413"></a><span class="lineno"> 413</span> </div><div class="line"><a name="l00414"></a><span class="lineno"> 414</span> } <span class="comment">// namespace</span></div><div class="line"><a name="l00415"></a><span class="lineno"> 415</span> </div><div class="line"><a name="l00416"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#ac254b18a929efcdd8742b3467ca5c8ff"> 416</a></span> <a class="code" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#ac254b18a929efcdd8742b3467ca5c8ff">NEFuseBatchNormalizationKernel::NEFuseBatchNormalizationKernel</a>()</div><div class="line"><a name="l00417"></a><span class="lineno"> 417</span>  : _input_weights(nullptr), _input_bias(nullptr), _bn_mean(nullptr), _bn_var(nullptr), _bn_gamma(nullptr), _bn_beta(nullptr), _fused_weights(nullptr), _fused_bias(nullptr), _epsilon(),</div><div class="line"><a name="l00418"></a><span class="lineno"> 418</span>  _run_in_place_weights(false), _run_in_place_bias(false), _func(nullptr)</div><div class="line"><a name="l00419"></a><span class="lineno"> 419</span> {</div><div class="line"><a name="l00420"></a><span class="lineno"> 420</span> }</div><div class="line"><a name="l00421"></a><span class="lineno"> 421</span> </div><div class="line"><a name="l00422"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a245ec81a11c260c325ec481c41b15d96"> 422</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a245ec81a11c260c325ec481c41b15d96">NEFuseBatchNormalizationKernel::configure</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *input_weights, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *bn_mean, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *bn_var,</div><div class="line"><a name="l00423"></a><span class="lineno"> 423</span>  <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *fused_weights, <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *fused_bias,</div><div class="line"><a name="l00424"></a><span class="lineno"> 424</span>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *input_bias, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *bn_beta, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *bn_gamma,</div><div class="line"><a name="l00425"></a><span class="lineno"> 425</span>  <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <a class="code" href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0ca">FuseBatchNormalizationType</a> fbn_type)</div><div class="line"><a name="l00426"></a><span class="lineno"> 426</span> {</div><div class="line"><a name="l00427"></a><span class="lineno"> 427</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(input_weights, bn_mean, bn_var);</div><div class="line"><a name="l00428"></a><span class="lineno"> 428</span> </div><div class="line"><a name="l00429"></a><span class="lineno"> 429</span>  _input_weights = input_weights;</div><div class="line"><a name="l00430"></a><span class="lineno"> 430</span>  _input_bias = input_bias;</div><div class="line"><a name="l00431"></a><span class="lineno"> 431</span>  _bn_mean = bn_mean;</div><div class="line"><a name="l00432"></a><span class="lineno"> 432</span>  _bn_var = bn_var;</div><div class="line"><a name="l00433"></a><span class="lineno"> 433</span>  _bn_beta = bn_beta;</div><div class="line"><a name="l00434"></a><span class="lineno"> 434</span>  _bn_gamma = bn_gamma;</div><div class="line"><a name="l00435"></a><span class="lineno"> 435</span>  _fused_weights = fused_weights;</div><div class="line"><a name="l00436"></a><span class="lineno"> 436</span>  _fused_bias = fused_bias;</div><div class="line"><a name="l00437"></a><span class="lineno"> 437</span>  _epsilon = <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>;</div><div class="line"><a name="l00438"></a><span class="lineno"> 438</span> </div><div class="line"><a name="l00439"></a><span class="lineno"> 439</span>  _run_in_place_weights = (fused_weights == <span class="keyword">nullptr</span>) || (fused_weights == input_weights);</div><div class="line"><a name="l00440"></a><span class="lineno"> 440</span>  _run_in_place_bias = (fused_bias == <span class="keyword">nullptr</span>) || (input_bias != <span class="keyword">nullptr</span> && fused_bias == input_bias);</div><div class="line"><a name="l00441"></a><span class="lineno"> 441</span> </div><div class="line"><a name="l00442"></a><span class="lineno"> 442</span>  <span class="comment">// Auto initialize outputs</span></div><div class="line"><a name="l00443"></a><span class="lineno"> 443</span>  <span class="keywordflow">if</span>(_fused_weights != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00444"></a><span class="lineno"> 444</span>  {</div><div class="line"><a name="l00445"></a><span class="lineno"> 445</span>  <span class="comment">// Output tensor auto initialization if not yet initialized</span></div><div class="line"><a name="l00446"></a><span class="lineno"> 446</span>  <a class="code" href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">auto_init_if_empty</a>(*_fused_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), *_input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>());</div><div class="line"><a name="l00447"></a><span class="lineno"> 447</span>  fused_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a9586081a29fceb532ab270bd843abee6">set_valid_region</a>(input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a74dfd07380a290c34fe7c8e065029b95">valid_region</a>());</div><div class="line"><a name="l00448"></a><span class="lineno"> 448</span>  }</div><div class="line"><a name="l00449"></a><span class="lineno"> 449</span>  <span class="keywordflow">if</span>(_fused_bias != <span class="keyword">nullptr</span>)</div><div class="line"><a name="l00450"></a><span class="lineno"> 450</span>  {</div><div class="line"><a name="l00451"></a><span class="lineno"> 451</span>  <span class="comment">// Output tensor auto initialization if not yet initialized</span></div><div class="line"><a name="l00452"></a><span class="lineno"> 452</span>  <a class="code" href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">auto_init_if_empty</a>(*_fused_bias-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), *_bn_mean-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>());</div><div class="line"><a name="l00453"></a><span class="lineno"> 453</span>  _fused_bias-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a9586081a29fceb532ab270bd843abee6">set_valid_region</a>(bn_mean-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a74dfd07380a290c34fe7c8e065029b95">valid_region</a>());</div><div class="line"><a name="l00454"></a><span class="lineno"> 454</span>  }</div><div class="line"><a name="l00455"></a><span class="lineno"> 455</span> </div><div class="line"><a name="l00456"></a><span class="lineno"> 456</span>  <span class="comment">// Validate arguments</span></div><div class="line"><a name="l00457"></a><span class="lineno"> 457</span>  <a class="code" href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a>(validate_arguments(input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), bn_mean-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), bn_var-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(),</div><div class="line"><a name="l00458"></a><span class="lineno"> 458</span>  (fused_weights != <span class="keyword">nullptr</span>) ? fused_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>() : <span class="keyword">nullptr</span>,</div><div class="line"><a name="l00459"></a><span class="lineno"> 459</span>  (fused_bias != <span class="keyword">nullptr</span>) ? fused_bias-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>() : <span class="keyword">nullptr</span>,</div><div class="line"><a name="l00460"></a><span class="lineno"> 460</span>  (input_bias != <span class="keyword">nullptr</span>) ? input_bias-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>() : <span class="keyword">nullptr</span>,</div><div class="line"><a name="l00461"></a><span class="lineno"> 461</span>  (bn_beta != <span class="keyword">nullptr</span>) ? bn_beta-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>() : <span class="keyword">nullptr</span>,</div><div class="line"><a name="l00462"></a><span class="lineno"> 462</span>  (bn_gamma != <span class="keyword">nullptr</span>) ? bn_gamma-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>() : <span class="keyword">nullptr</span>,</div><div class="line"><a name="l00463"></a><span class="lineno"> 463</span>  <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, fbn_type));</div><div class="line"><a name="l00464"></a><span class="lineno"> 464</span> </div><div class="line"><a name="l00465"></a><span class="lineno"> 465</span>  <span class="comment">// Configure kernel window</span></div><div class="line"><a name="l00466"></a><span class="lineno"> 466</span>  <a class="code" href="classarm__compute_1_1_window.xhtml">Window</a> win = <a class="code" href="namespacearm__compute.xhtml#ab7980fa5ee693e3282a76da047a1c3b5">calculate_max_window</a>(*input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>());</div><div class="line"><a name="l00467"></a><span class="lineno"> 467</span>  INEKernel::configure(win);</div><div class="line"><a name="l00468"></a><span class="lineno"> 468</span> </div><div class="line"><a name="l00469"></a><span class="lineno"> 469</span>  <span class="comment">// Configure function</span></div><div class="line"><a name="l00470"></a><span class="lineno"> 470</span>  <span class="keyword">static</span> std::map<std::string, FuseBatchNormFunction *> map_function =</div><div class="line"><a name="l00471"></a><span class="lineno"> 471</span>  {</div><div class="line"><a name="l00472"></a><span class="lineno"> 472</span>  { <span class="stringliteral">"fused_batch_normalization_conv_NHWC_F32"</span>, &fused_batch_normalization_conv<wrapper::traits::neon_vector<float, 4>> },</div><div class="line"><a name="l00473"></a><span class="lineno"> 473</span>  { <span class="stringliteral">"fused_batch_normalization_conv_NCHW_F32"</span>, &fused_batch_normalization_conv<wrapper::traits::neon_vector<float, 4>> },</div><div class="line"><a name="l00474"></a><span class="lineno"> 474</span>  { <span class="stringliteral">"fused_batch_normalization_dwc_NHWC_F32"</span>, &fused_batch_normalization_dwc_nhwc<wrapper::traits::neon_vector<float, 4>> },</div><div class="line"><a name="l00475"></a><span class="lineno"> 475</span>  { <span class="stringliteral">"fused_batch_normalization_dwc_NCHW_F32"</span>, &fused_batch_normalization_dwc_nchw<wrapper::traits::neon_vector<float, 4>> },</div><div class="line"><a name="l00476"></a><span class="lineno"> 476</span> <span class="preprocessor">#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC</span></div><div class="line"><a name="l00477"></a><span class="lineno"> 477</span>  { <span class="stringliteral">"fused_batch_normalization_conv_NHWC_F16"</span>, &fused_batch_normalization_conv<wrapper::traits::neon_vector<float16_t, 8>> },</div><div class="line"><a name="l00478"></a><span class="lineno"> 478</span>  { <span class="stringliteral">"fused_batch_normalization_conv_NCHW_F16"</span>, &fused_batch_normalization_conv<wrapper::traits::neon_vector<float16_t, 8>> },</div><div class="line"><a name="l00479"></a><span class="lineno"> 479</span>  { <span class="stringliteral">"fused_batch_normalization_dwc_NHWC_F16"</span>, &fused_batch_normalization_dwc_nhwc<wrapper::traits::neon_vector<float16_t, 8>> },</div><div class="line"><a name="l00480"></a><span class="lineno"> 480</span>  { <span class="stringliteral">"fused_batch_normalization_dwc_NCHW_F16"</span>, &fused_batch_normalization_dwc_nchw<wrapper::traits::neon_vector<float16_t, 8>> },</div><div class="line"><a name="l00481"></a><span class="lineno"> 481</span> <span class="preprocessor">#endif </span><span class="comment">/* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */</span><span class="preprocessor"></span></div><div class="line"><a name="l00482"></a><span class="lineno"> 482</span>  };</div><div class="line"><a name="l00483"></a><span class="lineno"> 483</span> </div><div class="line"><a name="l00484"></a><span class="lineno"> 484</span>  std::string function_to_call(<span class="stringliteral">"fused_batch_normalization_"</span>);</div><div class="line"><a name="l00485"></a><span class="lineno"> 485</span>  function_to_call += fbn_type == <a class="code" href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0caaf7fb587bad912d390b57a5d6e7ccc55d">FuseBatchNormalizationType::CONVOLUTION</a> ? <span class="stringliteral">"conv_"</span> : <span class="stringliteral">"dwc_"</span>;</div><div class="line"><a name="l00486"></a><span class="lineno"> 486</span>  function_to_call += <a class="code" href="namespacearm__compute.xhtml#a56b9e25c2e8731ca5488e7b3ccd66f58">string_from_data_layout</a>(_input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a367b5090ab432bc7de2c32369e087ab1">data_layout</a>());</div><div class="line"><a name="l00487"></a><span class="lineno"> 487</span>  function_to_call += <span class="stringliteral">"_"</span>;</div><div class="line"><a name="l00488"></a><span class="lineno"> 488</span>  function_to_call += <a class="code" href="namespacearm__compute.xhtml#a142b55a483cadf4e1068a1a09a55e8e9">string_from_data_type</a>(_input_weights-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()-><a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a7cfb31af63202568efef5214acfbf3ba">data_type</a>());</div><div class="line"><a name="l00489"></a><span class="lineno"> 489</span> </div><div class="line"><a name="l00490"></a><span class="lineno"> 490</span>  <span class="keyword">auto</span> it = map_function.find(function_to_call);</div><div class="line"><a name="l00491"></a><span class="lineno"> 491</span> </div><div class="line"><a name="l00492"></a><span class="lineno"> 492</span>  <span class="keywordflow">if</span>(it != map_function.end())</div><div class="line"><a name="l00493"></a><span class="lineno"> 493</span>  {</div><div class="line"><a name="l00494"></a><span class="lineno"> 494</span>  _func = it->second;</div><div class="line"><a name="l00495"></a><span class="lineno"> 495</span>  }</div><div class="line"><a name="l00496"></a><span class="lineno"> 496</span> }</div><div class="line"><a name="l00497"></a><span class="lineno"> 497</span> </div><div class="line"><a name="l00498"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a0edd3b29f8bd62c14d3c08483f0234d8"> 498</a></span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a> <a class="code" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a0edd3b29f8bd62c14d3c08483f0234d8">NEFuseBatchNormalizationKernel::validate</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *input_weights, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *bn_mean, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *bn_var,</div><div class="line"><a name="l00499"></a><span class="lineno"> 499</span>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *fused_weights, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *fused_bias,</div><div class="line"><a name="l00500"></a><span class="lineno"> 500</span>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *input_bias, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *bn_beta, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *bn_gamma,</div><div class="line"><a name="l00501"></a><span class="lineno"> 501</span>  <span class="keywordtype">float</span> <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, <a class="code" href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0ca">FuseBatchNormalizationType</a> fbn_type)</div><div class="line"><a name="l00502"></a><span class="lineno"> 502</span> {</div><div class="line"><a name="l00503"></a><span class="lineno"> 503</span>  <a class="code" href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a>(validate_arguments(input_weights, bn_mean, bn_var, fused_weights, fused_bias, input_bias, bn_beta, bn_gamma, <a class="code" href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a>, fbn_type));</div><div class="line"><a name="l00504"></a><span class="lineno"> 504</span>  <span class="keywordflow">return</span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a>{};</div><div class="line"><a name="l00505"></a><span class="lineno"> 505</span> }</div><div class="line"><a name="l00506"></a><span class="lineno"> 506</span> </div><div class="line"><a name="l00507"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82"> 507</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">NEFuseBatchNormalizationKernel::run</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_window.xhtml">Window</a> &window, <span class="keyword">const</span> <a class="code" href="structarm__compute_1_1_thread_info.xhtml">ThreadInfo</a> &<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a4f4125dba5283887b34f889b1c615c0c">info</a>)</div><div class="line"><a name="l00508"></a><span class="lineno"> 508</span> {</div><div class="line"><a name="l00509"></a><span class="lineno"> 509</span>  <a class="code" href="_error_8h.xhtml#a6dc630a6ae9cc063b3924bcea8dee9d6">ARM_COMPUTE_UNUSED</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a4f4125dba5283887b34f889b1c615c0c">info</a>);</div><div class="line"><a name="l00510"></a><span class="lineno"> 510</span>  <a class="code" href="_validate_8h.xhtml#a1b35b0d258183cf9ef36adf684d0b88c">ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL</a>(<span class="keyword">this</span>);</div><div class="line"><a name="l00511"></a><span class="lineno"> 511</span>  <a class="code" href="_validate_8h.xhtml#a6eb9ce82815fe429250189da7592ba75">ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW</a>(<a class="code" href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">IKernel::window</a>(), <a class="code" href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">window</a>);</div><div class="line"><a name="l00512"></a><span class="lineno"> 512</span>  (*_func)(_input_weights, _input_bias, _fused_weights, _fused_bias, _bn_mean, _bn_var, _bn_beta, _bn_gamma, _epsilon, <a class="code" href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">window</a>);</div><div class="line"><a name="l00513"></a><span class="lineno"> 513</span> }</div><div class="line"><a name="l00514"></a><span class="lineno"> 514</span> } <span class="comment">// namespace arm_compute</span></div><div class="ttc" id="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel_xhtml_a0edd3b29f8bd62c14d3c08483f0234d8"><div class="ttname"><a href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a0edd3b29f8bd62c14d3c08483f0234d8">arm_compute::NEFuseBatchNormalizationKernel::validate</a></div><div class="ttdeci">static Status validate(const ITensorInfo *input_weights, const ITensorInfo *bn_mean, const ITensorInfo *bn_var, const ITensorInfo *fused_weights, const ITensorInfo *fused_bias, const ITensorInfo *input_bias=nullptr, const ITensorInfo *bn_beta=nullptr, const ITensorInfo *bn_gamma=nullptr, float epsilon=0.001f, FuseBatchNormalizationType fbn_type=FuseBatchNormalizationType::CONVOLUTION)</div><div class="ttdoc">Static function to check if given info will lead to a valid configuration of NEFuseBatchNormalization...</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_fuse_batch_normalization_kernel_8cpp_source.xhtml#l00498">NEFuseBatchNormalizationKernel.cpp:498</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_kernel_xhtml_ad34a46f53686c12a5c5e717cc9617fb6"><div class="ttname"><a href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">arm_compute::IKernel::window</a></div><div class="ttdeci">const Window & window() const</div><div class="ttdoc">The maximum window the kernel can be executed on.</div><div class="ttdef"><b>Definition:</b> <a href="_i_kernel_8cpp_source.xhtml#l00028">IKernel.cpp:28</a></div></div> |
| <div class="ttc" id="_toolchain_support_8h_xhtml"><div class="ttname"><a href="_toolchain_support_8h.xhtml">ToolchainSupport.h</a></div></div> |
| <div class="ttc" id="_i_tensor_8h_xhtml"><div class="ttname"><a href="_i_tensor_8h.xhtml">ITensor.h</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_abdb9168800c70e5e2c4c020a3b905738"><div class="ttname"><a href="_validate_8h.xhtml#abdb9168800c70e5e2c4c020a3b905738">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00494">Validate.h:494</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a8f3ff7da485ff7e75dab07baadf5b4bd"><div class="ttname"><a href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00545">Validate.h:545</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a569988c8e1a2619532b8f4d0394d5d79"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a569988c8e1a2619532b8f4d0394d5d79">arm_compute::wrapper::vinvsqrt</a></div><div class="ttdeci">float32x2_t vinvsqrt(const float32x2_t &a)</div><div class="ttdef"><b>Definition:</b> <a href="invsqrt_8h_source.xhtml#l00040">invsqrt.h:40</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a77f54eded7fef436d3a4f21ad5a00da6"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">arm_compute::wrapper::vloadq</a></div><div class="ttdeci">uint8x16_t vloadq(const uint8_t *ptr)</div><div class="ttdef"><b>Definition:</b> <a href="load_8h_source.xhtml#l00058">load.h:58</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a8a1e1c105f0bdaf37db408c7cfcb77a4"><div class="ttname"><a href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ON_ERROR(status)</div><div class="ttdoc">Checks if a status contains an error and returns it.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00193">Error.h:193</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml_a7cfb31af63202568efef5214acfbf3ba"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a7cfb31af63202568efef5214acfbf3ba">arm_compute::ITensorInfo::data_type</a></div><div class="ttdeci">virtual DataType data_type() const =0</div><div class="ttdoc">Data type used for each element of the tensor.</div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a1894e825a225f3b2013f594cbffdae73"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">arm_compute::wrapper::vadd</a></div><div class="ttdeci">uint8x8_t vadd(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="intrinsics_2add_8h_source.xhtml#l00039">add.h:39</a></div></div> |
| <div class="ttc" id="_asymm_helpers_8cpp_xhtml_a552dc3787d7ea1675f3e4e8993501d58"><div class="ttname"><a href="_asymm_helpers_8cpp.xhtml#a552dc3787d7ea1675f3e4e8993501d58">epsilon</a></div><div class="ttdeci">constexpr float epsilon</div><div class="ttdef"><b>Definition:</b> <a href="_asymm_helpers_8cpp_source.xhtml#l00033">AsymmHelpers.cpp:33</a></div></div> |
| <div class="ttc" id="_window_8h_xhtml"><div class="ttname"><a href="_window_8h.xhtml">Window.h</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_ae7eed178dac535c6e727061b1f5bc6eb"><div class="ttname"><a href="_validate_8h.xhtml#ae7eed178dac535c6e727061b1f5bc6eb">ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00791">Validate.h:791</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda"><div class="ttname"><a href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">arm_compute::Format::F32</a></div><div class="ttdoc">1 channel, 1 F32 per channel</div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_afb7e3dd8a833840aaaf618bd43ead0caaf7fb587bad912d390b57a5d6e7ccc55d"><div class="ttname"><a href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0caaf7fb587bad912d390b57a5d6e7ccc55d">arm_compute::FuseBatchNormalizationType::CONVOLUTION</a></div><div class="ttdoc">For Convolution weights.</div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml">arm_compute::ITensorInfo</a></div><div class="ttdoc">Store the tensor's metadata.</div><div class="ttdef"><b>Definition:</b> <a href="_i_tensor_info_8h_source.xhtml#l00040">ITensorInfo.h:40</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a938dcd406ce611ef5345ad2531cdb948"><div class="ttname"><a href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR_THROW_ON(status)</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00327">Error.h:327</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_acae4ec59246f93845fa78120d83d7c1e"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">arm_compute::wrapper::vsub</a></div><div class="ttdeci">uint8x8_t vsub(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="sub_8h_source.xhtml#l00039">sub.h:39</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_status_xhtml"><div class="ttname"><a href="classarm__compute_1_1_status.xhtml">arm_compute::Status</a></div><div class="ttdoc">Status class.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00052">Error.h:52</a></div></div> |
| <div class="ttc" id="_tensor_info_8h_xhtml"><div class="ttname"><a href="_tensor_info_8h.xhtml">TensorInfo.h</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a206d6e247e0957ac3dee45d27756fc25"><div class="ttname"><a href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON(cond)</div><div class="ttdoc">If the condition is true, an error is returned.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00244">Error.h:244</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_xhtml"><div class="ttname"><a href="classarm__compute_1_1_i_tensor.xhtml">arm_compute::ITensor</a></div><div class="ttdoc">Interface for NEON tensor.</div><div class="ttdef"><b>Definition:</b> <a href="_i_tensor_8h_source.xhtml#l00036">ITensor.h:36</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ab7980fa5ee693e3282a76da047a1c3b5"><div class="ttname"><a href="namespacearm__compute.xhtml#ab7980fa5ee693e3282a76da047a1c3b5">arm_compute::calculate_max_window</a></div><div class="ttdeci">Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps=Steps(), bool skip_border=false, BorderSize border_size=BorderSize())</div><div class="ttdoc">Calculate the maximum window for a given tensor shape and border setting.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_helpers_8cpp_source.xhtml#l00028">Helpers.cpp:28</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml"><div class="ttname"><a href="namespacearm__compute.xhtml">arm_compute</a></div><div class="ttdoc">Copyright (c) 2017-2018 ARM Limited.</div><div class="ttdef"><b>Definition:</b> <a href="00__introduction_8dox_source.xhtml#l00024">00_introduction.dox:24</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml_a9586081a29fceb532ab270bd843abee6"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a9586081a29fceb532ab270bd843abee6">arm_compute::ITensorInfo::set_valid_region</a></div><div class="ttdeci">virtual void set_valid_region(const ValidRegion &valid_region)=0</div><div class="ttdoc">Set the valid region of the tensor.</div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a47be6fa38308d0003c25b60b7dbc45ce"><div class="ttname"><a href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">arm_compute::auto_init_if_empty</a></div><div class="ttdeci">bool auto_init_if_empty(ITensorInfo &info, const TensorShape &shape, int num_channels, DataType data_type, QuantizationInfo quantization_info=QuantizationInfo())</div><div class="ttdoc">Auto initialize the tensor info (shape, number of channels and data type) if the current assignment i...</div><div class="ttdef"><b>Definition:</b> <a href="_helpers_8inl_source.xhtml#l00201">Helpers.inl:201</a></div></div> |
| <div class="ttc" id="_c_p_p_2_validate_8h_xhtml_ad2633f3560322e1f8d926949dec1b730"><div class="ttname"><a href="_c_p_p_2_validate_8h.xhtml#ad2633f3560322e1f8d926949dec1b730">ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED(tensor)</div><div class="ttdef"><b>Definition:</b> <a href="_c_p_p_2_validate_8h_source.xhtml#l00071">Validate.h:71</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ab4e88c89b3b7ea1735996cc4def22d58a56d8353718e6fdc78b8d69078a2cdb94"><div class="ttname"><a href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a56d8353718e6fdc78b8d69078a2cdb94">arm_compute::Format::F16</a></div><div class="ttdoc">1 channel, 1 F16 per channel</div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml_a74dfd07380a290c34fe7c8e065029b95"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a74dfd07380a290c34fe7c8e065029b95">arm_compute::ITensorInfo::valid_region</a></div><div class="ttdeci">virtual ValidRegion valid_region() const =0</div><div class="ttdoc">Valid region of the tensor.</div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel_xhtml_a112b35dd205c62ea6ed1447ef226da82"><div class="ttname"><a href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">arm_compute::NEFuseBatchNormalizationKernel::run</a></div><div class="ttdeci">void run(const Window &window, const ThreadInfo &info) override</div><div class="ttdoc">Execute the kernel on the passed window.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_fuse_batch_normalization_kernel_8cpp_source.xhtml#l00507">NEFuseBatchNormalizationKernel.cpp:507</a></div></div> |
| <div class="ttc" id="arm__compute_2core_2_utils_8h_xhtml"><div class="ttname"><a href="arm__compute_2core_2_utils_8h.xhtml">Utils.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_afb7e3dd8a833840aaaf618bd43ead0ca"><div class="ttname"><a href="namespacearm__compute.xhtml#afb7e3dd8a833840aaaf618bd43ead0ca">arm_compute::FuseBatchNormalizationType</a></div><div class="ttdeci">FuseBatchNormalizationType</div><div class="ttdoc">Available FuseBatchNormalizationType.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_types_8h_source.xhtml#l00148">Types.h:148</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a142b55a483cadf4e1068a1a09a55e8e9"><div class="ttname"><a href="namespacearm__compute.xhtml#a142b55a483cadf4e1068a1a09a55e8e9">arm_compute::string_from_data_type</a></div><div class="ttdeci">const std::string & string_from_data_type(DataType dt)</div><div class="ttdoc">Convert a data type identity into a string.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_utils_8cpp_source.xhtml#l00144">Utils.cpp:144</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_window_xhtml_aa96e81276ee4f87ab386cd05a5539a7d"><div class="ttname"><a href="classarm__compute_1_1_window.xhtml#aa96e81276ee4f87ab386cd05a5539a7d">arm_compute::Window::DimX</a></div><div class="ttdeci">static constexpr size_t DimX</div><div class="ttdoc">Alias for dimension 0 also known as X dimension.</div><div class="ttdef"><b>Definition:</b> <a href="_window_8h_source.xhtml#l00043">Window.h:43</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a6dc630a6ae9cc063b3924bcea8dee9d6"><div class="ttname"><a href="_error_8h.xhtml#a6dc630a6ae9cc063b3924bcea8dee9d6">ARM_COMPUTE_UNUSED</a></div><div class="ttdeci">#define ARM_COMPUTE_UNUSED(...)</div><div class="ttdoc">To avoid unused variables warnings.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00160">Error.h:160</a></div></div> |
| <div class="ttc" id="_c_p_p_2_validate_8h_xhtml"><div class="ttname"><a href="_c_p_p_2_validate_8h.xhtml">Validate.h</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a27e4638546c88b8916f967e6e54480a9"><div class="ttname"><a href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00443">Validate.h:443</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel_xhtml_a245ec81a11c260c325ec481c41b15d96"><div class="ttname"><a href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#a245ec81a11c260c325ec481c41b15d96">arm_compute::NEFuseBatchNormalizationKernel::configure</a></div><div class="ttdeci">void configure(const ITensor *input_weights, const ITensor *bn_mean, const ITensor *bn_var, ITensor *fused_weights, ITensor *fused_bias, const ITensor *input_bias=nullptr, const ITensor *bn_beta=nullptr, const ITensor *bn_gamma=nullptr, float epsilon=0.001f, FuseBatchNormalizationType fbn_type=FuseBatchNormalizationType::CONVOLUTION)</div><div class="ttdoc">Set the source, destination of the kernel.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_fuse_batch_normalization_kernel_8cpp_source.xhtml#l00422">NEFuseBatchNormalizationKernel.cpp:422</a></div></div> |
| <div class="ttc" id="utils_2_type_printer_8h_xhtml"><div class="ttname"><a href="utils_2_type_printer_8h.xhtml">TypePrinter.h</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1misc_1_1_i_cloneable_xhtml_a4d10e5012a872e7f78f2b539b673049d"><div class="ttname"><a href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">arm_compute::misc::ICloneable::clone</a></div><div class="ttdeci">virtual std::unique_ptr< T > clone() const =0</div><div class="ttdoc">Provide a clone of the current object of class T.</div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_xhtml_a0e95dc1e53c361348314873b168ae237"><div class="ttname"><a href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">arm_compute::ITensor::info</a></div><div class="ttdeci">virtual ITensorInfo * info() const =0</div><div class="ttdoc">Interface to be implemented by the child class to return the tensor's metadata.</div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a74ce3f7420453d3446218ff3b7453e02af52e9c50a060add65a035429b2a22229"><div class="ttname"><a href="namespacearm__compute.xhtml#a74ce3f7420453d3446218ff3b7453e02af52e9c50a060add65a035429b2a22229">arm_compute::DataLayoutDimension::CHANNEL</a></div><div class="ttdoc">channel</div></div> |
| <div class="ttc" id="_validate_8h_xhtml_aff911654521523937ff24372a870b89f"><div class="ttname"><a href="_validate_8h.xhtml#aff911654521523937ff24372a870b89f">ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00163">Validate.h:163</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a921b705e9e3e0fe928928447869e62a5"><div class="ttname"><a href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00161">Validate.h:161</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a66c8dd218a4863a14fec363e3dd13dc4"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">arm_compute::wrapper::vmul</a></div><div class="ttdeci">uint8x8_t vmul(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="mul_8h_source.xhtml#l00039">mul.h:39</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a56b9e25c2e8731ca5488e7b3ccd66f58"><div class="ttname"><a href="namespacearm__compute.xhtml#a56b9e25c2e8731ca5488e7b3ccd66f58">arm_compute::string_from_data_layout</a></div><div class="ttdeci">const std::string & string_from_data_layout(DataLayout dl)</div><div class="ttdoc">Convert a data layout identity into a string.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_utils_8cpp_source.xhtml#l00132">Utils.cpp:132</a></div></div> |
| <div class="ttc" id="structarm__compute_1_1_thread_info_xhtml"><div class="ttname"><a href="structarm__compute_1_1_thread_info.xhtml">arm_compute::ThreadInfo</a></div><div class="ttdoc">Information about executing thread and CPU.</div><div class="ttdef"><b>Definition:</b> <a href="_c_p_p_types_8h_source.xhtml#l00225">CPPTypes.h:225</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel_xhtml_ac254b18a929efcdd8742b3467ca5c8ff"><div class="ttname"><a href="classarm__compute_1_1_n_e_fuse_batch_normalization_kernel.xhtml#ac254b18a929efcdd8742b3467ca5c8ff">arm_compute::NEFuseBatchNormalizationKernel::NEFuseBatchNormalizationKernel</a></div><div class="ttdeci">NEFuseBatchNormalizationKernel()</div><div class="ttdoc">Default constructor.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_fuse_batch_normalization_kernel_8cpp_source.xhtml#l00416">NEFuseBatchNormalizationKernel.cpp:416</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_ae7943ea9c1f74dc72c62d4cc3966a459"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">arm_compute::wrapper::vstore</a></div><div class="ttdeci">void vstore(uint8_t *ptr, uint8x8_t val)</div><div class="ttdef"><b>Definition:</b> <a href="store_8h_source.xhtml#l00039">store.h:39</a></div></div> |
| <div class="ttc" id="arm__compute_2core_2_helpers_8h_xhtml"><div class="ttname"><a href="arm__compute_2core_2_helpers_8h.xhtml">Helpers.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a39e87435be178fba49b76f49426ef873"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">arm_compute::wrapper::vdup_n</a></div><div class="ttdeci">uint8x8_t vdup_n(uint8_t value, traits::vector_64_tag)</div><div class="ttdef"><b>Definition:</b> <a href="dup__n_8h_source.xhtml#l00041">dup_n.h:41</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a5002bf7ec46d52971f9526e94172cfee"><div class="ttname"><a href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">arm_compute::execute_window_loop</a></div><div class="ttdeci">void execute_window_loop(const Window &w, L &&lambda_function, Ts &&... iterators)</div><div class="ttdoc">Iterate through the passed window, automatically adjusting the iterators and calling the lambda_funct...</div><div class="ttdef"><b>Definition:</b> <a href="_helpers_8inl_source.xhtml#l00122">Helpers.inl:122</a></div></div> |
| <div class="ttc" id="_n_e_fuse_batch_normalization_kernel_8h_xhtml"><div class="ttname"><a href="_n_e_fuse_batch_normalization_kernel_8h.xhtml">NEFuseBatchNormalizationKernel.h</a></div></div> |
| <div class="ttc" id="wrapper_8h_xhtml"><div class="ttname"><a href="wrapper_8h.xhtml">wrapper.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a46e938020a3ac8c926d0590b7fe957db"><div class="ttname"><a href="namespacearm__compute.xhtml#a46e938020a3ac8c926d0590b7fe957db">arm_compute::get_data_layout_dimension_index</a></div><div class="ttdeci">size_t get_data_layout_dimension_index(const DataLayout data_layout, const DataLayoutDimension data_layout_dimension)</div><div class="ttdoc">Get the index of the given dimension.</div><div class="ttdef"><b>Definition:</b> <a href="_helpers_8inl_source.xhtml#l00326">Helpers.inl:326</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1test_1_1validation_xhtml_a4f4125dba5283887b34f889b1c615c0c"><div class="ttname"><a href="namespacearm__compute_1_1test_1_1validation.xhtml#a4f4125dba5283887b34f889b1c615c0c">arm_compute::test::validation::info</a></div><div class="ttdeci">info</div><div class="ttdef"><b>Definition:</b> <a href="validation_2_c_l_2_convolution_layer_8cpp_source.xhtml#l00174">ConvolutionLayer.cpp:174</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a6eb9ce82815fe429250189da7592ba75"><div class="ttname"><a href="_validate_8h.xhtml#a6eb9ce82815fe429250189da7592ba75">ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00205">Validate.h:205</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_window_xhtml"><div class="ttname"><a href="classarm__compute_1_1_window.xhtml">arm_compute::Window</a></div><div class="ttdoc">Describe a multidimensional execution window.</div><div class="ttdef"><b>Definition:</b> <a href="_window_8h_source.xhtml#l00039">Window.h:39</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a1b35b0d258183cf9ef36adf684d0b88c"><div class="ttname"><a href="_validate_8h.xhtml#a1b35b0d258183cf9ef36adf684d0b88c">ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00940">Validate.h:940</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml"><div class="ttname"><a href="_validate_8h.xhtml">Validate.h</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml_a367b5090ab432bc7de2c32369e087ab1"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a367b5090ab432bc7de2c32369e087ab1">arm_compute::ITensorInfo::data_layout</a></div><div class="ttdeci">virtual DataLayout data_layout() const =0</div><div class="ttdoc">Get the data layout of the tensor.</div></div> |
| </div><!-- fragment --></div><!-- contents --> |
| </div><!-- doc-content --> |
| <!-- start footer part --> |
| <div id="nav-path" class="navpath"><!-- id is needed for treeview function! --> |
| <ul> |
| <li class="navelem"><a class="el" href="dir_68267d1309a1af8e8297ef4c3efbcdba.xhtml">src</a></li><li class="navelem"><a class="el" href="dir_aebb8dcc11953d78e620bbef0b9e2183.xhtml">core</a></li><li class="navelem"><a class="el" href="dir_d91e1966cc57df4f49d1e07be7697713.xhtml">NEON</a></li><li class="navelem"><a class="el" href="dir_1ce938f580a6f304b99685e2c1dd19db.xhtml">kernels</a></li><li class="navelem"><a class="el" href="_n_e_fuse_batch_normalization_kernel_8cpp.xhtml">NEFuseBatchNormalizationKernel.cpp</a></li> |
| <li class="footer">Generated on Mon Sep 2 2019 11:47:24 for Compute Library by |
| <a href="http://www.doxygen.org/index.html"> |
| <img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.8.15 </li> |
| </ul> |
| </div> |
| </body> |
| </html> |
