| <!-- 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/NESoftmaxLayerKernel.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">20.02.1</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_softmax_layer_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">NESoftmaxLayerKernel.cpp</div> </div> |
| </div><!--header--> |
| <div class="contents"> |
| <a href="_n_e_softmax_layer_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) 2017-2019 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_softmax_layer_kernel_8h.xhtml">arm_compute/core/NEON/kernels/NESoftmaxLayerKernel.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="_access_window_static_8h.xhtml">arm_compute/core/AccessWindowStatic.h</a>"</span></div><div class="line"><a name="l00027"></a><span class="lineno"> 27</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="l00028"></a><span class="lineno"> 28</span> <span class="preprocessor">#include "<a class="code" href="_error_8h.xhtml">arm_compute/core/Error.h</a>"</span></div><div class="line"><a name="l00029"></a><span class="lineno"> 29</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="l00030"></a><span class="lineno"> 30</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="l00031"></a><span class="lineno"> 31</span> <span class="preprocessor">#include "<a class="code" href="_n_e_fixed_point_8h.xhtml">arm_compute/core/NEON/NEFixedPoint.h</a>"</span></div><div class="line"><a name="l00032"></a><span class="lineno"> 32</span> <span class="preprocessor">#include "<a class="code" href="_n_e_math_8h.xhtml">arm_compute/core/NEON/NEMath.h</a>"</span></div><div class="line"><a name="l00033"></a><span class="lineno"> 33</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="l00034"></a><span class="lineno"> 34</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="l00035"></a><span class="lineno"> 35</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="l00036"></a><span class="lineno"> 36</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="l00037"></a><span class="lineno"> 37</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="l00038"></a><span class="lineno"> 38</span> <span class="preprocessor">#include "<a class="code" href="_saturate_cast_8h.xhtml">arm_compute/core/utils/misc/SaturateCast.h</a>"</span></div><div class="line"><a name="l00039"></a><span class="lineno"> 39</span> </div><div class="line"><a name="l00040"></a><span class="lineno"> 40</span> <span class="preprocessor">#include <algorithm></span></div><div class="line"><a name="l00041"></a><span class="lineno"> 41</span> <span class="preprocessor">#include <arm_neon.h></span></div><div class="line"><a name="l00042"></a><span class="lineno"> 42</span> <span class="preprocessor">#include <cfloat></span></div><div class="line"><a name="l00043"></a><span class="lineno"> 43</span> <span class="preprocessor">#include <functional></span></div><div class="line"><a name="l00044"></a><span class="lineno"> 44</span> </div><div class="line"><a name="l00045"></a><span class="lineno"> 45</span> <span class="keyword">namespace </span><a class="code" href="namespacearm__compute.xhtml">arm_compute</a></div><div class="line"><a name="l00046"></a><span class="lineno"> 46</span> {</div><div class="line"><a name="l00047"></a><span class="lineno"> 47</span> <span class="keyword">template</span> <<span class="keyword">typename</span> <span class="keywordtype">float</span>_vec_type, <span class="keyword">typename</span> <span class="keywordtype">int</span>_vec_type></div><div class="line"><a name="l00048"></a><span class="lineno"> 48</span> int_vec_type <a class="code" href="namespacearm__compute.xhtml#a17bff173004242d8f44b5b663177f9c0">convert_float_to_int</a>(<span class="keyword">const</span> float_vec_type &in);</div><div class="line"><a name="l00049"></a><span class="lineno"> 49</span> </div><div class="line"><a name="l00050"></a><span class="lineno"> 50</span> <span class="keyword">template</span> <<span class="keyword">typename</span> <span class="keywordtype">float</span>_vec_type, <span class="keyword">typename</span> <span class="keywordtype">int</span>_vec_type></div><div class="line"><a name="l00051"></a><span class="lineno"> 51</span> float_vec_type <a class="code" href="namespacearm__compute.xhtml#ad339c829071ae619193300cb501195ac">convert_int_to_float</a>(<span class="keyword">const</span> int_vec_type &in);</div><div class="line"><a name="l00052"></a><span class="lineno"> 52</span> </div><div class="line"><a name="l00053"></a><span class="lineno"> 53</span> <span class="keyword">template</span> <></div><div class="line"><a name="l00054"></a><span class="lineno"><a class="line" href="namespacearm__compute.xhtml#ad63fda57a59d2dcc82eb13599581fcc0"> 54</a></span> uint8x16_t <a class="code" href="namespacearm__compute.xhtml#ad63fda57a59d2dcc82eb13599581fcc0">convert_float_to_int<float32x4x4_t, uint8x16_t></a>(<span class="keyword">const</span> float32x4x4_t &in)</div><div class="line"><a name="l00055"></a><span class="lineno"> 55</span> {</div><div class="line"><a name="l00056"></a><span class="lineno"> 56</span>  uint8x16_t out;</div><div class="line"><a name="l00057"></a><span class="lineno"> 57</span>  <a class="code" href="namespacearm__compute.xhtml#a4ee01955c2dd8c14a0d107a36cb62618">convert_float32x4x4_to_uint8x16</a>(in, out);</div><div class="line"><a name="l00058"></a><span class="lineno"> 58</span>  <span class="keywordflow">return</span> out;</div><div class="line"><a name="l00059"></a><span class="lineno"> 59</span> }</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="keyword">template</span> <></div><div class="line"><a name="l00062"></a><span class="lineno"><a class="line" href="namespacearm__compute.xhtml#a6048f3cbd52aea7ce3a693ce485393be"> 62</a></span> int8x16_t <a class="code" href="namespacearm__compute.xhtml#a6048f3cbd52aea7ce3a693ce485393be">convert_float_to_int<float32x4x4_t, int8x16_t></a>(<span class="keyword">const</span> float32x4x4_t &in)</div><div class="line"><a name="l00063"></a><span class="lineno"> 63</span> {</div><div class="line"><a name="l00064"></a><span class="lineno"> 64</span>  int8x16_t out;</div><div class="line"><a name="l00065"></a><span class="lineno"> 65</span>  <a class="code" href="namespacearm__compute.xhtml#a9c673ade3003d5d4e28ebf5875795dc4">convert_float32x4x4_to_int8x16</a>(in, out);</div><div class="line"><a name="l00066"></a><span class="lineno"> 66</span>  <span class="keywordflow">return</span> out;</div><div class="line"><a name="l00067"></a><span class="lineno"> 67</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> <span class="keyword">template</span> <></div><div class="line"><a name="l00070"></a><span class="lineno"><a class="line" href="namespacearm__compute.xhtml#a8677ddae33f84bb0e84e995c54bf6ba3"> 70</a></span> float32x4x4_t <a class="code" href="namespacearm__compute.xhtml#a8677ddae33f84bb0e84e995c54bf6ba3">convert_int_to_float<float32x4x4_t, uint8x16_t></a>(<span class="keyword">const</span> uint8x16_t &in)</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="keywordflow">return</span> <a class="code" href="namespacearm__compute.xhtml#a1b362fecc52ac83bcfbb0850549edce4">convert_uint8x16_to_float32x4x4</a>(in);</div><div class="line"><a name="l00073"></a><span class="lineno"> 73</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> <span class="keyword">template</span> <></div><div class="line"><a name="l00076"></a><span class="lineno"><a class="line" href="namespacearm__compute.xhtml#a9075aa895fe4b2368e3c0cfae26ebf0b"> 76</a></span> float32x4x4_t <a class="code" href="namespacearm__compute.xhtml#a9075aa895fe4b2368e3c0cfae26ebf0b">convert_int_to_float<float32x4x4_t, int8x16_t></a>(<span class="keyword">const</span> int8x16_t &in)</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="keywordflow">return</span> <a class="code" href="namespacearm__compute.xhtml#a4a20779ec9b583a9468b4fa60575d52e">convert_int8x16_to_float32x4x4</a>(in);</div><div class="line"><a name="l00079"></a><span class="lineno"> 79</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> <span class="keyword">namespace</span></div><div class="line"><a name="l00082"></a><span class="lineno"> 82</span> {</div><div class="line"><a name="l00083"></a><span class="lineno"> 83</span> Status validate_arguments_logits_1d_max(<span class="keyword">const</span> ITensorInfo &<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <span class="keyword">const</span> ITensorInfo &output)</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>  <a class="code" href="_c_p_p_2_validate_8h.xhtml#ad2633f3560322e1f8d926949dec1b730">ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>);</div><div class="line"><a name="l00086"></a><span class="lineno"> 86</span>  <a class="code" href="_validate_8h.xhtml#ae7eed178dac535c6e727061b1f5bc6eb">ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, 1, <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924">DataType::QASYMM8</a>, <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9">DataType::QASYMM8_SIGNED</a>, <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="l00087"></a><span class="lineno"> 87</span> </div><div class="line"><a name="l00088"></a><span class="lineno"> 88</span>  <span class="comment">// Validate in case of configured output</span></div><div class="line"><a name="l00089"></a><span class="lineno"> 89</span>  <span class="keywordflow">if</span>(output.total_size() != 0)</div><div class="line"><a name="l00090"></a><span class="lineno"> 90</span>  {</div><div class="line"><a name="l00091"></a><span class="lineno"> 91</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &output);</div><div class="line"><a name="l00092"></a><span class="lineno"> 92</span>  <a class="code" href="_validate_8h.xhtml#aba910b683652be1f65437ef37a9da2a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &output);</div><div class="line"><a name="l00093"></a><span class="lineno"> 93</span>  <a class="code" href="_validate_8h.xhtml#a1da797d2762c1cdbb73bfc83136c3a38">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS</a>(output.tensor_shape(), TensorShape(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.tensor_shape()).<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a1ec8dd7ea96aec0e4068497f0b5e17b2">set</a>(0, 1));</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> </div><div class="line"><a name="l00096"></a><span class="lineno"> 96</span>  <span class="keywordflow">return</span> Status{};</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> std::pair<Status, Window> validate_and_configure_window_logits_1d_max(ITensorInfo &<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, ITensorInfo &output)</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>  <span class="comment">// Softmax across the x dimension</span></div><div class="line"><a name="l00102"></a><span class="lineno"> 102</span>  <span class="keyword">const</span> TensorShape <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#ab1806bf0c5a41f674fb9d2dc6af644f5">output_shape</a> = TensorShape(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.tensor_shape()).<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a1ec8dd7ea96aec0e4068497f0b5e17b2">set</a>(0, 1);</div><div class="line"><a name="l00103"></a><span class="lineno"> 103</span>  <span class="comment">// Output auto initialization if not yet initialized</span></div><div class="line"><a name="l00104"></a><span class="lineno"> 104</span>  <a class="code" href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">auto_init_if_empty</a>(output, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#ab1806bf0c5a41f674fb9d2dc6af644f5">output_shape</a>, 1, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type(), <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.quantization_info());</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="comment">// Configure kernel window</span></div><div class="line"><a name="l00107"></a><span class="lineno"> 107</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> input_width = <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().shape.x();</div><div class="line"><a name="l00108"></a><span class="lineno"> 108</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a> = 16<a class="code" href="namespacearm__compute_1_1utils_1_1cast.xhtml#a2ea3d1fc01a3a442900249ca182ffa5e">U</a> / <a class="code" href="namespacearm__compute.xhtml#abb7e0f23a4f2e63f39433f158dad47ab">data_size_from_type</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type());</div><div class="line"><a name="l00109"></a><span class="lineno"> 109</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> num_elems_read_per_iteration = <a class="code" href="namespacearm__compute.xhtml#ab237a0a375cf382d52b61653248d3d4a">ceil_to_multiple</a>(input_width, <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a>);</div><div class="line"><a name="l00110"></a><span class="lineno"> 110</span> </div><div class="line"><a name="l00111"></a><span class="lineno"> 111</span>  <span class="keyword">const</span> ValidRegion out_valid_region(ValidRegion(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region()).<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a1ec8dd7ea96aec0e4068497f0b5e17b2">set</a>(0, 0, 1));</div><div class="line"><a name="l00112"></a><span class="lineno"> 112</span>  output.set_valid_region(out_valid_region);</div><div class="line"><a name="l00113"></a><span class="lineno"> 113</span> </div><div class="line"><a name="l00114"></a><span class="lineno"> 114</span>  Window win = <a class="code" href="namespacearm__compute.xhtml#ab7980fa5ee693e3282a76da047a1c3b5">calculate_max_window</a>(output);</div><div class="line"><a name="l00115"></a><span class="lineno"> 115</span> </div><div class="line"><a name="l00116"></a><span class="lineno"> 116</span>  AccessWindowHorizontal input_access(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().anchor.x(), num_elems_read_per_iteration);</div><div class="line"><a name="l00117"></a><span class="lineno"> 117</span>  AccessWindowHorizontal output_access(&output, 0, 1);</div><div class="line"><a name="l00118"></a><span class="lineno"> 118</span> </div><div class="line"><a name="l00119"></a><span class="lineno"> 119</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> window_changed = <a class="code" href="namespacearm__compute.xhtml#afc4bd8e872567d9c4c57d89eb0bb3da1">update_window_and_padding</a>(win, input_access, output_access);</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>  <span class="keyword">const</span> Status err = (window_changed) ? <a class="code" href="_error_8h.xhtml#af1b8ff8eb557a2ad11272f1505f45d34">ARM_COMPUTE_CREATE_ERROR</a>(<a class="code" href="namespacearm__compute.xhtml#a59e56af19e754a6aa26a612ebf91d05fa62be47fdd89da032cf78dfce82239579">ErrorCode::RUNTIME_ERROR</a>, <span class="stringliteral">"Insufficient Padding!"</span>) : Status{};</div><div class="line"><a name="l00122"></a><span class="lineno"> 122</span>  <span class="keywordflow">return</span> std::make_pair(err, 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> </div><div class="line"><a name="l00125"></a><span class="lineno"> 125</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T></div><div class="line"><a name="l00126"></a><span class="lineno"> 126</span> <span class="keywordtype">void</span> logits_1d_max(<span class="keyword">const</span> ITensor &in, ITensor &out, <span class="keyword">const</span> Window &window)</div><div class="line"><a name="l00127"></a><span class="lineno"> 127</span> {</div><div class="line"><a name="l00128"></a><span class="lineno"> 128</span>  <span class="keyword">const</span> <span class="keyword">auto</span> start_x = in.info()->valid_region().anchor.x();</div><div class="line"><a name="l00129"></a><span class="lineno"> 129</span>  <span class="keyword">const</span> <span class="keywordtype">size_t</span> input_width = in.info()->valid_region().shape.x();</div><div class="line"><a name="l00130"></a><span class="lineno"> 130</span> <span class="comment"></span></div><div class="line"><a name="l00131"></a><span class="lineno"> 131</span> <span class="comment"> /** NEON vector tag type. */</span></div><div class="line"><a name="l00132"></a><span class="lineno"> 132</span>  <span class="keyword">using</span> ExactTagType = <span class="keyword">typename</span> wrapper::traits::neon_bitvector_tag_t<T, wrapper::traits::BitWidth::W128>;</div><div class="line"><a name="l00133"></a><span class="lineno"> 133</span> </div><div class="line"><a name="l00134"></a><span class="lineno"> 134</span>  Iterator <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>(&in, window);</div><div class="line"><a name="l00135"></a><span class="lineno"> 135</span>  Iterator output(&out, window);</div><div class="line"><a name="l00136"></a><span class="lineno"> 136</span> </div><div class="line"><a name="l00137"></a><span class="lineno"> 137</span>  constexpr <span class="keywordtype">int</span> window_step_x = 16 / <span class="keyword">sizeof</span>(T);</div><div class="line"><a name="l00138"></a><span class="lineno"> 138</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> sum_stages = log2(window_step_x / 2);</div><div class="line"><a name="l00139"></a><span class="lineno"> 139</span>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(window, [&](<span class="keyword">const</span> Coordinates &)</div><div class="line"><a name="l00140"></a><span class="lineno"> 140</span>  {</div><div class="line"><a name="l00141"></a><span class="lineno"> 141</span>  <span class="comment">// Get pointers</span></div><div class="line"><a name="l00142"></a><span class="lineno"> 142</span>  <span class="keyword">const</span> <span class="keyword">auto</span> in_ptr = reinterpret_cast<const T *>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.ptr()) + start_x;</div><div class="line"><a name="l00143"></a><span class="lineno"> 143</span>  <span class="keyword">const</span> <span class="keyword">auto</span> out_ptr = reinterpret_cast<T *>(output.ptr());</div><div class="line"><a name="l00144"></a><span class="lineno"> 144</span> </div><div class="line"><a name="l00145"></a><span class="lineno"> 145</span>  <span class="comment">// Init max value</span></div><div class="line"><a name="l00146"></a><span class="lineno"> 146</span>  <span class="keyword">auto</span> vec_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(support::cpp11::lowest<T>(), ExactTagType{});</div><div class="line"><a name="l00147"></a><span class="lineno"> 147</span> </div><div class="line"><a name="l00148"></a><span class="lineno"> 148</span>  <span class="comment">// Loop over input row</span></div><div class="line"><a name="l00149"></a><span class="lineno"> 149</span>  <span class="keywordflow">for</span>(<span class="keyword">const</span> T *it = in_ptr; it < (in_ptr + input_width); it += window_step_x)</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>  <span class="keyword">const</span> <span class="keyword">auto</span> current_value = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(it);</div><div class="line"><a name="l00152"></a><span class="lineno"> 152</span>  vec_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a8d1d14000c959f1146a8a64e1dafae5f">wrapper::vmax</a>(vec_max, current_value);</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> </div><div class="line"><a name="l00155"></a><span class="lineno"> 155</span>  <span class="keyword">auto</span> carry_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ab799fd9cd5fc5989593e69c55bc400f9">wrapper::vpmax</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a95ee388aa7c5bccab918235dc538a6b3">wrapper::vgethigh</a>(vec_max), <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a2902775707bc7bf7d6da1bda1cc15783">wrapper::vgetlow</a>(vec_max));</div><div class="line"><a name="l00156"></a><span class="lineno"> 156</span> </div><div class="line"><a name="l00157"></a><span class="lineno"> 157</span>  <span class="keywordflow">for</span>(<span class="keywordtype">int</span> i = 0; i < sum_stages; ++i)</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>  carry_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ab799fd9cd5fc5989593e69c55bc400f9">wrapper::vpmax</a>(carry_max, carry_max);</div><div class="line"><a name="l00160"></a><span class="lineno"> 160</span>  }</div><div class="line"><a name="l00161"></a><span class="lineno"> 161</span>  <span class="keyword">const</span> T max_val = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#aa16ace001ab8287faa46d6962f369219">wrapper::vgetlane</a>(carry_max, 0);</div><div class="line"><a name="l00162"></a><span class="lineno"> 162</span>  *out_ptr = max_val;</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>  <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, output);</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> } <span class="comment">// namespace</span></div><div class="line"><a name="l00167"></a><span class="lineno"> 167</span> </div><div class="line"><a name="l00168"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a7dcfcda37359d31c2510ad865f4ca051"> 168</a></span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a7dcfcda37359d31c2510ad865f4ca051">NELogits1DMaxKernel::NELogits1DMaxKernel</a>()</div><div class="line"><a name="l00169"></a><span class="lineno"> 169</span>  : _func(nullptr), _border_size()</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> </div><div class="line"><a name="l00173"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a423f9a45a52983b4de5e2b347f4369c7"> 173</a></span> <a class="code" href="structarm__compute_1_1_border_size.xhtml">BorderSize</a> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a423f9a45a52983b4de5e2b347f4369c7">NELogits1DMaxKernel::border_size</a>()<span class="keyword"> const</span></div><div class="line"><a name="l00174"></a><span class="lineno"> 174</span> <span class="keyword"></span>{</div><div class="line"><a name="l00175"></a><span class="lineno"> 175</span>  <span class="keywordflow">return</span> _border_size;</div><div class="line"><a name="l00176"></a><span class="lineno"> 176</span> }</div><div class="line"><a name="l00177"></a><span class="lineno"> 177</span> </div><div class="line"><a name="l00178"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a83a344e60eb7db895953a942abf16628"> 178</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a83a344e60eb7db895953a942abf16628">NELogits1DMaxKernel::configure</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *output)</div><div class="line"><a name="l00179"></a><span class="lineno"> 179</span> {</div><div class="line"><a name="l00180"></a><span class="lineno"> 180</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, output);</div><div class="line"><a name="l00181"></a><span class="lineno"> 181</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>());</div><div class="line"><a name="l00182"></a><span class="lineno"> 182</span>  <span class="comment">// Perform validation step</span></div><div class="line"><a name="l00183"></a><span class="lineno"> 183</span>  <a class="code" href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a>(validate_arguments_logits_1d_max(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), *output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>()));</div><div class="line"><a name="l00184"></a><span class="lineno"> 184</span>  <span class="comment">// Configure kernel window</span></div><div class="line"><a name="l00185"></a><span class="lineno"> 185</span>  <span class="keyword">auto</span> win_config = validate_and_configure_window_logits_1d_max(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), *output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>());</div><div class="line"><a name="l00186"></a><span class="lineno"> 186</span>  <a class="code" href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a>(win_config.first);</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>  <span class="keywordflow">switch</span>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info()->data_type())</div><div class="line"><a name="l00189"></a><span class="lineno"> 189</span>  {</div><div class="line"><a name="l00190"></a><span class="lineno"> 190</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924">DataType::QASYMM8</a>:</div><div class="line"><a name="l00191"></a><span class="lineno"> 191</span>  _func = &logits_1d_max<qasymm8_t>;</div><div class="line"><a name="l00192"></a><span class="lineno"> 192</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00193"></a><span class="lineno"> 193</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9">DataType::QASYMM8_SIGNED</a>:</div><div class="line"><a name="l00194"></a><span class="lineno"> 194</span>  _func = &logits_1d_max<qasymm8_signed_t>;</div><div class="line"><a name="l00195"></a><span class="lineno"> 195</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00196"></a><span class="lineno"> 196</span> <span class="preprocessor">#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC</span></div><div class="line"><a name="l00197"></a><span class="lineno"> 197</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a56d8353718e6fdc78b8d69078a2cdb94">DataType::F16</a>:</div><div class="line"><a name="l00198"></a><span class="lineno"> 198</span>  _func = &logits_1d_max<float16_t>;</div><div class="line"><a name="l00199"></a><span class="lineno"> 199</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00200"></a><span class="lineno"> 200</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="l00201"></a><span class="lineno"> 201</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">DataType::F32</a>:</div><div class="line"><a name="l00202"></a><span class="lineno"> 202</span>  _func = &logits_1d_max<float>;</div><div class="line"><a name="l00203"></a><span class="lineno"> 203</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00204"></a><span class="lineno"> 204</span>  <span class="keywordflow">default</span>:</div><div class="line"><a name="l00205"></a><span class="lineno"> 205</span>  <a class="code" href="_error_8h.xhtml#a7cf8d8b669b8f7b05680230be30d60f4">ARM_COMPUTE_ERROR</a>(<span class="stringliteral">"Unsupported data type."</span>);</div><div class="line"><a name="l00206"></a><span class="lineno"> 206</span>  }</div><div class="line"><a name="l00207"></a><span class="lineno"> 207</span> </div><div class="line"><a name="l00208"></a><span class="lineno"> 208</span>  _input = <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>;</div><div class="line"><a name="l00209"></a><span class="lineno"> 209</span>  _output = output;</div><div class="line"><a name="l00210"></a><span class="lineno"> 210</span> </div><div class="line"><a name="l00211"></a><span class="lineno"> 211</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> input_width = <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info()->valid_region().shape.x();</div><div class="line"><a name="l00212"></a><span class="lineno"> 212</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a> = 16<a class="code" href="namespacearm__compute.xhtml#a1ce9b523fd4f3b5bbcadcd796183455aa4c614360da93c0a041b22e537de151eb">U</a> / <a class="code" href="namespacearm__compute.xhtml#abb7e0f23a4f2e63f39433f158dad47ab">data_size_from_type</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info()->data_type());</div><div class="line"><a name="l00213"></a><span class="lineno"> 213</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> num_elems_read_per_iteration = <a class="code" href="namespacearm__compute.xhtml#ab237a0a375cf382d52b61653248d3d4a">ceil_to_multiple</a>(input_width, <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a>);</div><div class="line"><a name="l00214"></a><span class="lineno"> 214</span> </div><div class="line"><a name="l00215"></a><span class="lineno"> 215</span>  _border_size = <a class="code" href="structarm__compute_1_1_border_size.xhtml">BorderSize</a>(0, num_elems_read_per_iteration - input_width, 0, 0);</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>  INEKernel::configure(win_config.second);</div><div class="line"><a name="l00218"></a><span class="lineno"> 218</span> }</div><div class="line"><a name="l00219"></a><span class="lineno"> 219</span> </div><div class="line"><a name="l00220"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a968b23a6ef327fcfb5b99d58e3fbe883"> 220</a></span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a968b23a6ef327fcfb5b99d58e3fbe883">NELogits1DMaxKernel::validate</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *output)</div><div class="line"><a name="l00221"></a><span class="lineno"> 221</span> {</div><div class="line"><a name="l00222"></a><span class="lineno"> 222</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, output);</div><div class="line"><a name="l00223"></a><span class="lineno"> 223</span> </div><div class="line"><a name="l00224"></a><span class="lineno"> 224</span>  <a class="code" href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a>(validate_arguments_logits_1d_max(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, *output));</div><div class="line"><a name="l00225"></a><span class="lineno"> 225</span>  <a class="code" href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a>(validate_and_configure_window_logits_1d_max(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->clone(), *output-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>()).first);</div><div class="line"><a name="l00226"></a><span class="lineno"> 226</span> </div><div class="line"><a name="l00227"></a><span class="lineno"> 227</span>  <span class="keywordflow">return</span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a>{};</div><div class="line"><a name="l00228"></a><span class="lineno"> 228</span> }</div><div class="line"><a name="l00229"></a><span class="lineno"> 229</span> </div><div class="line"><a name="l00230"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82"> 230</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">NELogits1DMaxKernel::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="l00231"></a><span class="lineno"> 231</span> {</div><div class="line"><a name="l00232"></a><span class="lineno"> 232</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="l00233"></a><span class="lineno"> 233</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="l00234"></a><span class="lineno"> 234</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">INEKernel::window</a>(), <a class="code" href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">window</a>);</div><div class="line"><a name="l00235"></a><span class="lineno"> 235</span>  <a class="code" href="_error_8h.xhtml#a54a6080c9f4df1f908e57a9bbb46f5da">ARM_COMPUTE_ERROR_ON</a>(_func == <span class="keyword">nullptr</span>);</div><div class="line"><a name="l00236"></a><span class="lineno"> 236</span> </div><div class="line"><a name="l00237"></a><span class="lineno"> 237</span>  (*_func)(*_input, *_output, <a class="code" href="classarm__compute_1_1_i_kernel.xhtml#ad34a46f53686c12a5c5e717cc9617fb6">window</a>);</div><div class="line"><a name="l00238"></a><span class="lineno"> 238</span> }</div><div class="line"><a name="l00239"></a><span class="lineno"> 239</span> </div><div class="line"><a name="l00240"></a><span class="lineno"> 240</span> <span class="keyword">namespace</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> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a> validate_arguments_logits_softmax(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> &<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> &max,</div><div class="line"><a name="l00243"></a><span class="lineno"> 243</span>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> &output, <span class="keyword">const</span> <span class="keywordtype">float</span> beta, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> &tmp, <span class="keywordtype">bool</span> is_log)</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>  <a class="code" href="_error_8h.xhtml#a6dc630a6ae9cc063b3924bcea8dee9d6">ARM_COMPUTE_UNUSED</a>(beta);</div><div class="line"><a name="l00246"></a><span class="lineno"> 246</span>  <span class="comment">// Check input</span></div><div class="line"><a name="l00247"></a><span class="lineno"> 247</span>  <a class="code" href="_c_p_p_2_validate_8h.xhtml#ad2633f3560322e1f8d926949dec1b730">ARM_COMPUTE_RETURN_ERROR_ON_CPU_F16_UNSUPPORTED</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>);</div><div class="line"><a name="l00248"></a><span class="lineno"> 248</span>  <a class="code" href="_validate_8h.xhtml#ae7eed178dac535c6e727061b1f5bc6eb">ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, 1, <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924">DataType::QASYMM8</a>, <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9">DataType::QASYMM8_SIGNED</a>, <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="l00249"></a><span class="lineno"> 249</span> </div><div class="line"><a name="l00250"></a><span class="lineno"> 250</span>  <span class="keyword">const</span> <span class="keywordtype">bool</span> is_quantized_asymmetric = <a class="code" href="namespacearm__compute.xhtml#a14f46283f316e7f0fad301d5c1507e9f">is_data_type_quantized_asymmetric</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type());</div><div class="line"><a name="l00251"></a><span class="lineno"> 251</span> </div><div class="line"><a name="l00252"></a><span class="lineno"> 252</span>  <span class="comment">// Check max</span></div><div class="line"><a name="l00253"></a><span class="lineno"> 253</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &max);</div><div class="line"><a name="l00254"></a><span class="lineno"> 254</span>  <a class="code" href="_validate_8h.xhtml#a1da797d2762c1cdbb73bfc83136c3a38">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS</a>(<a class="code" href="classarm__compute_1_1_tensor_shape.xhtml">TensorShape</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.tensor_shape()).<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a1ec8dd7ea96aec0e4068497f0b5e17b2">set</a>(0, 1), max.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a7c66505457d00ece3aa4b34cab80757d">tensor_shape</a>());</div><div class="line"><a name="l00255"></a><span class="lineno"> 255</span>  <a class="code" href="_validate_8h.xhtml#aba910b683652be1f65437ef37a9da2a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &max);</div><div class="line"><a name="l00256"></a><span class="lineno"> 256</span> </div><div class="line"><a name="l00257"></a><span class="lineno"> 257</span>  <span class="comment">// Check output if configured</span></div><div class="line"><a name="l00258"></a><span class="lineno"> 258</span>  <span class="keywordflow">if</span>(output.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a18064e0011c3869d884653e9e7c47b66">total_size</a>() != 0)</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>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_quantization_info.xhtml">QuantizationInfo</a> output_quantization = is_quantized_asymmetric ? <a class="code" href="namespacearm__compute.xhtml#a1dcccbff6075b2fad30ebd2a0da6b995">arm_compute::get_softmax_output_quantization_info</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type(), is_log) : output.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a3f3e1a3200223e6a304a533b1016e749">quantization_info</a>();</div><div class="line"><a name="l00261"></a><span class="lineno"> 261</span>  <a class="code" href="_validate_8h.xhtml#a8f3ff7da485ff7e75dab07baadf5b4bd">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &output);</div><div class="line"><a name="l00262"></a><span class="lineno"> 262</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &output);</div><div class="line"><a name="l00263"></a><span class="lineno"> 263</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(output.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a3f3e1a3200223e6a304a533b1016e749">quantization_info</a>() != output_quantization);</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="comment">// Check tmp if configured</span></div><div class="line"><a name="l00267"></a><span class="lineno"> 267</span>  <span class="keywordflow">if</span>(tmp.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a18064e0011c3869d884653e9e7c47b66">total_size</a>() != 0)</div><div class="line"><a name="l00268"></a><span class="lineno"> 268</span>  {</div><div class="line"><a name="l00269"></a><span class="lineno"> 269</span>  <span class="keyword">const</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6">DataType</a> tmp_data_type = is_quantized_asymmetric ? <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">DataType::F32</a> : <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type();</div><div class="line"><a name="l00270"></a><span class="lineno"> 270</span>  <a class="code" href="_error_8h.xhtml#a206d6e247e0957ac3dee45d27756fc25">ARM_COMPUTE_RETURN_ERROR_ON</a>(tmp.<a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml#a7cfb31af63202568efef5214acfbf3ba">data_type</a>() != tmp_data_type);</div><div class="line"><a name="l00271"></a><span class="lineno"> 271</span>  <span class="comment">// We could potentially reduce tmp memory if we could predict or make an assumption</span></div><div class="line"><a name="l00272"></a><span class="lineno"> 272</span>  <span class="comment">// on the maximum number of threads that will run in parallel.</span></div><div class="line"><a name="l00273"></a><span class="lineno"> 273</span>  <a class="code" href="_validate_8h.xhtml#a27e4638546c88b8916f967e6e54480a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES</a>(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, &tmp);</div><div class="line"><a name="l00274"></a><span class="lineno"> 274</span>  }</div><div class="line"><a name="l00275"></a><span class="lineno"> 275</span> </div><div class="line"><a name="l00276"></a><span class="lineno"> 276</span>  <span class="keywordflow">return</span> Status{};</div><div class="line"><a name="l00277"></a><span class="lineno"> 277</span> }</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> std::pair<Status, Window> validate_and_configure_window_logits_softmax(ITensorInfo &<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, ITensorInfo &max,</div><div class="line"><a name="l00280"></a><span class="lineno"> 280</span>  ITensorInfo &output, ITensorInfo &tmp, <span class="keywordtype">bool</span> is_log)</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>  <span class="keyword">const</span> <span class="keywordtype">bool</span> is_quantized_asymmetric = <a class="code" href="namespacearm__compute.xhtml#a14f46283f316e7f0fad301d5c1507e9f">is_data_type_quantized_asymmetric</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type());</div><div class="line"><a name="l00283"></a><span class="lineno"> 283</span> </div><div class="line"><a name="l00284"></a><span class="lineno"> 284</span>  <span class="comment">// Output auto initialization if not yet initialized</span></div><div class="line"><a name="l00285"></a><span class="lineno"> 285</span>  <span class="keyword">const</span> QuantizationInfo output_quantization = is_quantized_asymmetric ? <a class="code" href="namespacearm__compute.xhtml#a1dcccbff6075b2fad30ebd2a0da6b995">arm_compute::get_softmax_output_quantization_info</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type(), is_log) : output.quantization_info();</div><div class="line"><a name="l00286"></a><span class="lineno"> 286</span>  <a class="code" href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">auto_init_if_empty</a>(output, TensorInfo(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>).set_quantization_info(output_quantization).reset_padding());</div><div class="line"><a name="l00287"></a><span class="lineno"> 287</span> </div><div class="line"><a name="l00288"></a><span class="lineno"> 288</span>  <span class="comment">// Tmp auto initialization if not yet initialized</span></div><div class="line"><a name="l00289"></a><span class="lineno"> 289</span>  <span class="keyword">const</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6">DataType</a> tmp_data_type = is_quantized_asymmetric ? <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">DataType::F32</a> : <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.data_type();</div><div class="line"><a name="l00290"></a><span class="lineno"> 290</span>  <a class="code" href="namespacearm__compute.xhtml#a47be6fa38308d0003c25b60b7dbc45ce">auto_init_if_empty</a>(tmp, TensorInfo(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>).set_data_type(tmp_data_type).reset_padding());</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="keyword">const</span> <span class="keywordtype">int</span> input_width = <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().shape.x();</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>  Window win = <a class="code" href="namespacearm__compute.xhtml#ab7980fa5ee693e3282a76da047a1c3b5">calculate_max_window</a>(max);</div><div class="line"><a name="l00295"></a><span class="lineno"> 295</span> </div><div class="line"><a name="l00296"></a><span class="lineno"> 296</span>  AccessWindowHorizontal input_access(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().anchor.x(), input_width);</div><div class="line"><a name="l00297"></a><span class="lineno"> 297</span>  AccessWindowHorizontal max_access(&<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, 0, 1);</div><div class="line"><a name="l00298"></a><span class="lineno"> 298</span>  AccessWindowHorizontal output_access(&output, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().anchor.x(), input_width);</div><div class="line"><a name="l00299"></a><span class="lineno"> 299</span>  AccessWindowHorizontal tmp_access(&tmp, <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region().anchor.x(), input_width);</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>  <span class="keyword">const</span> <span class="keywordtype">bool</span> window_changed = <a class="code" href="namespacearm__compute.xhtml#afc4bd8e872567d9c4c57d89eb0bb3da1">update_window_and_padding</a>(win, input_access, max_access, output_access, tmp_access);</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>  output.set_valid_region(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>.valid_region());</div><div class="line"><a name="l00304"></a><span class="lineno"> 304</span> </div><div class="line"><a name="l00305"></a><span class="lineno"> 305</span>  <span class="keyword">const</span> Status err = (window_changed) ? <a class="code" href="_error_8h.xhtml#af1b8ff8eb557a2ad11272f1505f45d34">ARM_COMPUTE_CREATE_ERROR</a>(<a class="code" href="namespacearm__compute.xhtml#a59e56af19e754a6aa26a612ebf91d05fa62be47fdd89da032cf78dfce82239579">ErrorCode::RUNTIME_ERROR</a>, <span class="stringliteral">"Insufficient Padding!"</span>) : Status{};</div><div class="line"><a name="l00306"></a><span class="lineno"> 306</span>  <span class="keywordflow">return</span> std::make_pair(err, win);</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> </div><div class="line"><a name="l00309"></a><span class="lineno"> 309</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keywordtype">bool</span> is_log></div><div class="line"><a name="l00310"></a><span class="lineno"> 310</span> <span class="keywordtype">void</span> logits_1d_softmax_qasymm8(<span class="keyword">const</span> ITensor &in, <span class="keyword">const</span> ITensor &max, <span class="keywordtype">void</span> *<span class="keyword">const</span> tmp, ITensor &out, <span class="keyword">const</span> <span class="keywordtype">float</span> beta, <span class="keyword">const</span> Window &window)</div><div class="line"><a name="l00311"></a><span class="lineno"> 311</span> {</div><div class="line"><a name="l00312"></a><span class="lineno"> 312</span>  static_assert(std::is_same<T, qasymm8_t>::value</div><div class="line"><a name="l00313"></a><span class="lineno"> 313</span>  || std::is_same<T, qasymm8_signed_t>::value,</div><div class="line"><a name="l00314"></a><span class="lineno"> 314</span>  <span class="stringliteral">"quantized type should be either qasymm8_t or qasymm8_signed_t."</span>);</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>  <span class="keyword">const</span> <span class="keywordtype">int</span> start_x = in.info()->valid_region().anchor.x();</div><div class="line"><a name="l00317"></a><span class="lineno"> 317</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> input_width = in.info()->valid_region().shape.x();</div><div class="line"><a name="l00318"></a><span class="lineno"> 318</span> </div><div class="line"><a name="l00319"></a><span class="lineno"> 319</span>  <span class="keyword">const</span> <span class="keywordtype">float</span> scale_beta = -beta * in.info()->quantization_info().uniform().scale;</div><div class="line"><a name="l00320"></a><span class="lineno"> 320</span>  <span class="keyword">const</span> <span class="keyword">auto</span> scale_beta_vec = vdupq_n_f32(scale_beta);</div><div class="line"><a name="l00321"></a><span class="lineno"> 321</span> </div><div class="line"><a name="l00322"></a><span class="lineno"> 322</span>  Iterator in_it(&in, window);</div><div class="line"><a name="l00323"></a><span class="lineno"> 323</span>  Iterator max_it(&max, window);</div><div class="line"><a name="l00324"></a><span class="lineno"> 324</span>  Iterator out_it(&out, window);</div><div class="line"><a name="l00325"></a><span class="lineno"> 325</span>  constexpr <span class="keywordtype">int</span> vec_size = 16;</div><div class="line"><a name="l00326"></a><span class="lineno"> 326</span> </div><div class="line"><a name="l00327"></a><span class="lineno"> 327</span>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(window, [&](<span class="keyword">const</span> Coordinates &)</div><div class="line"><a name="l00328"></a><span class="lineno"> 328</span>  {</div><div class="line"><a name="l00329"></a><span class="lineno"> 329</span>  <span class="comment">/* Get pointers */</span></div><div class="line"><a name="l00330"></a><span class="lineno"> 330</span>  <span class="keyword">const</span> <span class="keyword">auto</span> in_ptr = reinterpret_cast<const T *>(in_it.ptr()) + start_x;</div><div class="line"><a name="l00331"></a><span class="lineno"> 331</span>  <span class="keyword">const</span> <span class="keyword">auto</span> out_ptr = reinterpret_cast<T *>(out_it.ptr()) + start_x;</div><div class="line"><a name="l00332"></a><span class="lineno"> 332</span>  <span class="keyword">const</span> <span class="keyword">auto</span> tmp_ptr = reinterpret_cast<float *>(tmp);</div><div class="line"><a name="l00333"></a><span class="lineno"> 333</span> </div><div class="line"><a name="l00334"></a><span class="lineno"> 334</span>  <span class="keywordtype">float</span> <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>{};</div><div class="line"><a name="l00335"></a><span class="lineno"> 335</span>  <span class="keywordtype">float</span> sum_inversed{};</div><div class="line"><a name="l00336"></a><span class="lineno"> 336</span> </div><div class="line"><a name="l00337"></a><span class="lineno"> 337</span>  <span class="comment">/* Compute exponentials and sum */</span></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="comment">/* Get max value */</span></div><div class="line"><a name="l00340"></a><span class="lineno"> 340</span>  <span class="keyword">const</span> <span class="keyword">auto</span> max_val = *reinterpret_cast<const T *>(max_it.ptr());</div><div class="line"><a name="l00341"></a><span class="lineno"> 341</span>  <span class="keyword">const</span> <span class="keyword">auto</span> vec_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(max_val, wrapper::traits::vector_128_tag{});</div><div class="line"><a name="l00342"></a><span class="lineno"> 342</span> </div><div class="line"><a name="l00343"></a><span class="lineno"> 343</span>  <span class="comment">/* Init sum to zero */</span></div><div class="line"><a name="l00344"></a><span class="lineno"> 344</span>  float32x4x4_t vec_sum =</div><div class="line"><a name="l00345"></a><span class="lineno"> 345</span>  {</div><div class="line"><a name="l00346"></a><span class="lineno"> 346</span>  vdupq_n_f32(0.f),</div><div class="line"><a name="l00347"></a><span class="lineno"> 347</span>  vdupq_n_f32(0.f),</div><div class="line"><a name="l00348"></a><span class="lineno"> 348</span>  vdupq_n_f32(0.f),</div><div class="line"><a name="l00349"></a><span class="lineno"> 349</span>  vdupq_n_f32(0.f),</div><div class="line"><a name="l00350"></a><span class="lineno"> 350</span>  };</div><div class="line"><a name="l00351"></a><span class="lineno"> 351</span> </div><div class="line"><a name="l00352"></a><span class="lineno"> 352</span>  <span class="comment">/* Loop over row and compute exponentials and sum */</span></div><div class="line"><a name="l00353"></a><span class="lineno"> 353</span>  <span class="keywordtype">int</span> x = 0;</div><div class="line"><a name="l00354"></a><span class="lineno"> 354</span>  <span class="keywordflow">for</span>(; x <= (input_width - vec_size); x += vec_size)</div><div class="line"><a name="l00355"></a><span class="lineno"> 355</span>  {</div><div class="line"><a name="l00356"></a><span class="lineno"> 356</span>  <span class="keyword">auto</span> vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(in_ptr + x);</div><div class="line"><a name="l00357"></a><span class="lineno"> 357</span>  vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(vec_max, vec_elements);</div><div class="line"><a name="l00358"></a><span class="lineno"> 358</span>  <span class="keyword">auto</span> vec_elements_flt = convert_int_to_float<float32x4x4_t>(vec_elements);</div><div class="line"><a name="l00359"></a><span class="lineno"> 359</span> </div><div class="line"><a name="l00360"></a><span class="lineno"> 360</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00361"></a><span class="lineno"> 361</span>  {</div><div class="line"><a name="l00362"></a><span class="lineno"> 362</span>  vec_elements_flt.val[0] = vmulq_f32(vec_elements_flt.val[0], scale_beta_vec);</div><div class="line"><a name="l00363"></a><span class="lineno"> 363</span>  vec_elements_flt.val[1] = vmulq_f32(vec_elements_flt.val[1], scale_beta_vec);</div><div class="line"><a name="l00364"></a><span class="lineno"> 364</span>  vec_elements_flt.val[2] = vmulq_f32(vec_elements_flt.val[2], scale_beta_vec);</div><div class="line"><a name="l00365"></a><span class="lineno"> 365</span>  vec_elements_flt.val[3] = vmulq_f32(vec_elements_flt.val[3], scale_beta_vec);</div><div class="line"><a name="l00366"></a><span class="lineno"> 366</span>  vec_sum.val[0] = vaddq_f32(vec_sum.val[0], <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vec_elements_flt.val[0]));</div><div class="line"><a name="l00367"></a><span class="lineno"> 367</span>  vec_sum.val[1] = vaddq_f32(vec_sum.val[1], <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vec_elements_flt.val[1]));</div><div class="line"><a name="l00368"></a><span class="lineno"> 368</span>  vec_sum.val[2] = vaddq_f32(vec_sum.val[2], <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vec_elements_flt.val[2]));</div><div class="line"><a name="l00369"></a><span class="lineno"> 369</span>  vec_sum.val[3] = vaddq_f32(vec_sum.val[3], <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vec_elements_flt.val[3]));</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="keywordflow">else</span></div><div class="line"><a name="l00372"></a><span class="lineno"> 372</span>  {</div><div class="line"><a name="l00373"></a><span class="lineno"> 373</span>  vec_elements_flt.val[0] = <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vmulq_f32(vec_elements_flt.val[0], scale_beta_vec));</div><div class="line"><a name="l00374"></a><span class="lineno"> 374</span>  vec_elements_flt.val[1] = <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vmulq_f32(vec_elements_flt.val[1], scale_beta_vec));</div><div class="line"><a name="l00375"></a><span class="lineno"> 375</span>  vec_elements_flt.val[2] = <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vmulq_f32(vec_elements_flt.val[2], scale_beta_vec));</div><div class="line"><a name="l00376"></a><span class="lineno"> 376</span>  vec_elements_flt.val[3] = <a class="code" href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">vexpq_f32</a>(vmulq_f32(vec_elements_flt.val[3], scale_beta_vec));</div><div class="line"><a name="l00377"></a><span class="lineno"> 377</span>  vec_sum.val[0] = vaddq_f32(vec_sum.val[0], vec_elements_flt.val[0]);</div><div class="line"><a name="l00378"></a><span class="lineno"> 378</span>  vec_sum.val[1] = vaddq_f32(vec_sum.val[1], vec_elements_flt.val[1]);</div><div class="line"><a name="l00379"></a><span class="lineno"> 379</span>  vec_sum.val[2] = vaddq_f32(vec_sum.val[2], vec_elements_flt.val[2]);</div><div class="line"><a name="l00380"></a><span class="lineno"> 380</span>  vec_sum.val[3] = vaddq_f32(vec_sum.val[3], vec_elements_flt.val[3]);</div><div class="line"><a name="l00381"></a><span class="lineno"> 381</span>  }</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>  vst4q_f32(tmp_ptr + x, vec_elements_flt);</div><div class="line"><a name="l00384"></a><span class="lineno"> 384</span>  }</div><div class="line"><a name="l00385"></a><span class="lineno"> 385</span> </div><div class="line"><a name="l00386"></a><span class="lineno"> 386</span>  <span class="comment">/* Reduce sum */</span></div><div class="line"><a name="l00387"></a><span class="lineno"> 387</span>  <span class="keyword">const</span> <span class="keyword">auto</span> sum_16_byte = vaddq_f32(vaddq_f32(vec_sum.val[0], vec_sum.val[1]), vaddq_f32(vec_sum.val[2], vec_sum.val[3]));</div><div class="line"><a name="l00388"></a><span class="lineno"> 388</span>  <span class="keyword">auto</span> sum_res = vpadd_f32(vget_high_f32(sum_16_byte), vget_low_f32(sum_16_byte));</div><div class="line"><a name="l00389"></a><span class="lineno"> 389</span>  sum_res = vpadd_f32(sum_res, sum_res);</div><div class="line"><a name="l00390"></a><span class="lineno"> 390</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#aa16ace001ab8287faa46d6962f369219">wrapper::vgetlane</a>(sum_res, 0);</div><div class="line"><a name="l00391"></a><span class="lineno"> 391</span> </div><div class="line"><a name="l00392"></a><span class="lineno"> 392</span>  <span class="comment">/* Run remaining elements */</span></div><div class="line"><a name="l00393"></a><span class="lineno"> 393</span>  <span class="keywordflow">for</span>(; x < input_width; ++x)</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="keywordtype">float</span> element{};</div><div class="line"><a name="l00396"></a><span class="lineno"> 396</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00397"></a><span class="lineno"> 397</span>  {</div><div class="line"><a name="l00398"></a><span class="lineno"> 398</span>  element = (max_val - in_ptr[x]) * scale_beta;</div><div class="line"><a name="l00399"></a><span class="lineno"> 399</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> += std::exp(element);</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="keywordflow">else</span></div><div class="line"><a name="l00402"></a><span class="lineno"> 402</span>  {</div><div class="line"><a name="l00403"></a><span class="lineno"> 403</span>  element = std::exp((max_val - in_ptr[x]) * scale_beta);</div><div class="line"><a name="l00404"></a><span class="lineno"> 404</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> += element;</div><div class="line"><a name="l00405"></a><span class="lineno"> 405</span>  }</div><div class="line"><a name="l00406"></a><span class="lineno"> 406</span> </div><div class="line"><a name="l00407"></a><span class="lineno"> 407</span>  tmp_ptr[x] = element;</div><div class="line"><a name="l00408"></a><span class="lineno"> 408</span>  }</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>  <span class="keywordflow">if</span>(!is_log)</div><div class="line"><a name="l00411"></a><span class="lineno"> 411</span>  {</div><div class="line"><a name="l00412"></a><span class="lineno"> 412</span>  sum_inversed = 256.f / <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>;</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>  }</div><div class="line"><a name="l00415"></a><span class="lineno"> 415</span> </div><div class="line"><a name="l00416"></a><span class="lineno"> 416</span>  <span class="comment">/* Normalize exponentials */</span></div><div class="line"><a name="l00417"></a><span class="lineno"> 417</span>  {</div><div class="line"><a name="l00418"></a><span class="lineno"> 418</span>  constexpr <span class="keywordtype">bool</span> is_qasymm8_signed = std::is_same<T, qasymm8_signed_t>::value;</div><div class="line"><a name="l00419"></a><span class="lineno"> 419</span>  <span class="comment">/* Loop over row and compute softmax */</span></div><div class="line"><a name="l00420"></a><span class="lineno"> 420</span>  <span class="keywordtype">int</span> x = 0;</div><div class="line"><a name="l00421"></a><span class="lineno"> 421</span>  <span class="keywordflow">for</span>(; x <= (input_width - vec_size); x += vec_size)</div><div class="line"><a name="l00422"></a><span class="lineno"> 422</span>  {</div><div class="line"><a name="l00423"></a><span class="lineno"> 423</span>  <span class="keyword">using</span> int_vec_type = wrapper::traits::neon_vector_t<T, 16>;</div><div class="line"><a name="l00424"></a><span class="lineno"> 424</span>  float32x4x4_t vec_in = vld4q_f32(tmp_ptr + x);</div><div class="line"><a name="l00425"></a><span class="lineno"> 425</span>  int_vec_type normalized_value{};</div><div class="line"><a name="l00426"></a><span class="lineno"> 426</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00427"></a><span class="lineno"> 427</span>  {</div><div class="line"><a name="l00428"></a><span class="lineno"> 428</span>  <span class="keyword">const</span> float32x4x4_t <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a306c2d0ce612405d1d4ee732534dc543">sub</a> =</div><div class="line"><a name="l00429"></a><span class="lineno"> 429</span>  {</div><div class="line"><a name="l00430"></a><span class="lineno"> 430</span>  vsubq_f32(vec_in.val[0], vdupq_n_f32(<a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>)),</div><div class="line"><a name="l00431"></a><span class="lineno"> 431</span>  vsubq_f32(vec_in.val[1], vdupq_n_f32(<a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>)),</div><div class="line"><a name="l00432"></a><span class="lineno"> 432</span>  vsubq_f32(vec_in.val[2], vdupq_n_f32(<a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>)),</div><div class="line"><a name="l00433"></a><span class="lineno"> 433</span>  vsubq_f32(vec_in.val[3], vdupq_n_f32(<a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>)),</div><div class="line"><a name="l00434"></a><span class="lineno"> 434</span>  };</div><div class="line"><a name="l00435"></a><span class="lineno"> 435</span>  normalized_value = convert_float_to_int<float32x4x4_t, int_vec_type>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a306c2d0ce612405d1d4ee732534dc543">sub</a>);</div><div class="line"><a name="l00436"></a><span class="lineno"> 436</span>  }</div><div class="line"><a name="l00437"></a><span class="lineno"> 437</span>  <span class="keywordflow">else</span></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>  float32x4x4_t mul =</div><div class="line"><a name="l00440"></a><span class="lineno"> 440</span>  {</div><div class="line"><a name="l00441"></a><span class="lineno"> 441</span>  vmulq_f32(vec_in.val[0], vdupq_n_f32(sum_inversed)),</div><div class="line"><a name="l00442"></a><span class="lineno"> 442</span>  vmulq_f32(vec_in.val[1], vdupq_n_f32(sum_inversed)),</div><div class="line"><a name="l00443"></a><span class="lineno"> 443</span>  vmulq_f32(vec_in.val[2], vdupq_n_f32(sum_inversed)),</div><div class="line"><a name="l00444"></a><span class="lineno"> 444</span>  vmulq_f32(vec_in.val[3], vdupq_n_f32(sum_inversed)),</div><div class="line"><a name="l00445"></a><span class="lineno"> 445</span>  };</div><div class="line"><a name="l00446"></a><span class="lineno"> 446</span> </div><div class="line"><a name="l00447"></a><span class="lineno"> 447</span>  <span class="keywordflow">if</span>(is_qasymm8_signed)</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="keyword">const</span> <span class="keyword">auto</span> offset_vec = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(128.f, wrapper::traits::vector_128_tag{});</div><div class="line"><a name="l00450"></a><span class="lineno"> 450</span>  mul.val[0] = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(mul.val[0], offset_vec);</div><div class="line"><a name="l00451"></a><span class="lineno"> 451</span>  mul.val[1] = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(mul.val[1], offset_vec);</div><div class="line"><a name="l00452"></a><span class="lineno"> 452</span>  mul.val[2] = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(mul.val[2], offset_vec);</div><div class="line"><a name="l00453"></a><span class="lineno"> 453</span>  mul.val[3] = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(mul.val[3], offset_vec);</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>  normalized_value = convert_float_to_int<float32x4x4_t, int_vec_type>(mul);</div><div class="line"><a name="l00457"></a><span class="lineno"> 457</span>  }</div><div class="line"><a name="l00458"></a><span class="lineno"> 458</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(out_ptr + x, normalized_value);</div><div class="line"><a name="l00459"></a><span class="lineno"> 459</span>  }</div><div class="line"><a name="l00460"></a><span class="lineno"> 460</span>  <span class="comment">/* Run remaining elements */</span></div><div class="line"><a name="l00461"></a><span class="lineno"> 461</span>  <span class="keywordflow">for</span>(; x < input_width; ++x)</div><div class="line"><a name="l00462"></a><span class="lineno"> 462</span>  {</div><div class="line"><a name="l00463"></a><span class="lineno"> 463</span>  <span class="keywordflow">if</span>(is_log)</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>  out_ptr[x] = utils::cast::saturate_cast<T>(tmp_ptr[x] - <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>);</div><div class="line"><a name="l00466"></a><span class="lineno"> 466</span>  }</div><div class="line"><a name="l00467"></a><span class="lineno"> 467</span>  <span class="keywordflow">else</span></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>  out_ptr[x] = utils::cast::saturate_cast<T>((tmp_ptr[x] * sum_inversed) - (is_qasymm8_signed ? 128.f : 0));</div><div class="line"><a name="l00470"></a><span class="lineno"> 470</span>  }</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>  }</div><div class="line"><a name="l00473"></a><span class="lineno"> 473</span>  },</div><div class="line"><a name="l00474"></a><span class="lineno"> 474</span>  in_it, max_it, out_it);</div><div class="line"><a name="l00475"></a><span class="lineno"> 475</span> }</div><div class="line"><a name="l00476"></a><span class="lineno"> 476</span> </div><div class="line"><a name="l00477"></a><span class="lineno"> 477</span> <span class="keyword">template</span> <<span class="keyword">typename</span> T, <span class="keywordtype">bool</span> is_log = false></div><div class="line"><a name="l00478"></a><span class="lineno"> 478</span> <span class="keywordtype">void</span> logits_1d_softmax_float(<span class="keyword">const</span> ITensor &in, <span class="keyword">const</span> ITensor &max, <span class="keywordtype">void</span> *<span class="keyword">const</span> tmp,</div><div class="line"><a name="l00479"></a><span class="lineno"> 479</span>  ITensor &out, <span class="keyword">const</span> <span class="keywordtype">float</span> beta, <span class="keyword">const</span> Window &window)</div><div class="line"><a name="l00480"></a><span class="lineno"> 480</span> {</div><div class="line"><a name="l00481"></a><span class="lineno"> 481</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> start_x = in.info()->valid_region().anchor.x();</div><div class="line"><a name="l00482"></a><span class="lineno"> 482</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> input_width = in.info()->valid_region().shape.x();</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>  Iterator in_it(&in, window);</div><div class="line"><a name="l00485"></a><span class="lineno"> 485</span>  Iterator max_it(&max, window);</div><div class="line"><a name="l00486"></a><span class="lineno"> 486</span>  Iterator out_it(&out, window);</div><div class="line"><a name="l00487"></a><span class="lineno"> 487</span> <span class="comment"></span></div><div class="line"><a name="l00488"></a><span class="lineno"> 488</span> <span class="comment"> /** NEON vector tag type. */</span></div><div class="line"><a name="l00489"></a><span class="lineno"> 489</span>  <span class="keyword">using</span> ExactTagType = <span class="keyword">typename</span> wrapper::traits::neon_bitvector_tag_t<T, wrapper::traits::BitWidth::W128>;</div><div class="line"><a name="l00490"></a><span class="lineno"> 490</span> </div><div class="line"><a name="l00491"></a><span class="lineno"> 491</span>  constexpr <span class="keywordtype">int</span> vec_size = 16 / <span class="keyword">sizeof</span>(T);</div><div class="line"><a name="l00492"></a><span class="lineno"> 492</span>  <span class="keyword">const</span> <span class="keywordtype">int</span> sum_stages = log2(vec_size / 2);</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>  <a class="code" href="namespacearm__compute.xhtml#a5002bf7ec46d52971f9526e94172cfee">execute_window_loop</a>(window, [&](<span class="keyword">const</span> Coordinates &)</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>  <span class="comment">/* Get pointers */</span></div><div class="line"><a name="l00497"></a><span class="lineno"> 497</span>  <span class="keyword">const</span> <span class="keyword">auto</span> in_ptr = reinterpret_cast<const T *>(in_it.ptr()) + start_x;</div><div class="line"><a name="l00498"></a><span class="lineno"> 498</span>  <span class="keyword">const</span> <span class="keyword">auto</span> out_ptr = reinterpret_cast<T *>(out_it.ptr()) + start_x;</div><div class="line"><a name="l00499"></a><span class="lineno"> 499</span>  <span class="keyword">const</span> <span class="keyword">auto</span> tmp_ptr = reinterpret_cast<T *>(tmp);</div><div class="line"><a name="l00500"></a><span class="lineno"> 500</span> </div><div class="line"><a name="l00501"></a><span class="lineno"> 501</span>  T <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>{};</div><div class="line"><a name="l00502"></a><span class="lineno"> 502</span>  T sum_inversed{};</div><div class="line"><a name="l00503"></a><span class="lineno"> 503</span> </div><div class="line"><a name="l00504"></a><span class="lineno"> 504</span>  <span class="comment">/* Compute exponentials and sum */</span></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>  <span class="comment">/* Get max value */</span></div><div class="line"><a name="l00507"></a><span class="lineno"> 507</span>  <span class="keyword">const</span> <span class="keyword">auto</span> max_val = *reinterpret_cast<const T *>(max_it.ptr());</div><div class="line"><a name="l00508"></a><span class="lineno"> 508</span>  <span class="keyword">const</span> <span class="keyword">auto</span> vec_max = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(max_val, ExactTagType{});</div><div class="line"><a name="l00509"></a><span class="lineno"> 509</span> </div><div class="line"><a name="l00510"></a><span class="lineno"> 510</span>  <span class="comment">/* Init sum to zero */</span></div><div class="line"><a name="l00511"></a><span class="lineno"> 511</span>  <span class="keyword">auto</span> vec_sum = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(0), ExactTagType{});</div><div class="line"><a name="l00512"></a><span class="lineno"> 512</span> </div><div class="line"><a name="l00513"></a><span class="lineno"> 513</span>  <span class="comment">/* Loop over row and compute exponentials and sum */</span></div><div class="line"><a name="l00514"></a><span class="lineno"> 514</span>  <span class="keywordtype">int</span> x = 0;</div><div class="line"><a name="l00515"></a><span class="lineno"> 515</span>  <span class="keywordflow">for</span>(; x <= (input_width - vec_size); x += vec_size)</div><div class="line"><a name="l00516"></a><span class="lineno"> 516</span>  {</div><div class="line"><a name="l00517"></a><span class="lineno"> 517</span>  <span class="keyword">auto</span> vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(in_ptr + x);</div><div class="line"><a name="l00518"></a><span class="lineno"> 518</span>  vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(vec_elements, vec_max);</div><div class="line"><a name="l00519"></a><span class="lineno"> 519</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00520"></a><span class="lineno"> 520</span>  {</div><div class="line"><a name="l00521"></a><span class="lineno"> 521</span>  vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(vec_elements, <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(beta), ExactTagType{}));</div><div class="line"><a name="l00522"></a><span class="lineno"> 522</span>  vec_sum = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(vec_sum, <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a3d35fe714076b40450ba6ba870fae322">wrapper::vexpq</a>(vec_elements));</div><div class="line"><a name="l00523"></a><span class="lineno"> 523</span>  }</div><div class="line"><a name="l00524"></a><span class="lineno"> 524</span>  <span class="keywordflow">else</span></div><div class="line"><a name="l00525"></a><span class="lineno"> 525</span>  {</div><div class="line"><a name="l00526"></a><span class="lineno"> 526</span>  vec_elements = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a3d35fe714076b40450ba6ba870fae322">wrapper::vexpq</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(vec_elements, <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(beta), ExactTagType{})));</div><div class="line"><a name="l00527"></a><span class="lineno"> 527</span>  vec_sum = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a1894e825a225f3b2013f594cbffdae73">wrapper::vadd</a>(vec_sum, vec_elements);</div><div class="line"><a name="l00528"></a><span class="lineno"> 528</span>  }</div><div class="line"><a name="l00529"></a><span class="lineno"> 529</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(tmp_ptr + x, vec_elements);</div><div class="line"><a name="l00530"></a><span class="lineno"> 530</span>  }</div><div class="line"><a name="l00531"></a><span class="lineno"> 531</span> </div><div class="line"><a name="l00532"></a><span class="lineno"> 532</span>  <span class="comment">/* Reduce sum */</span></div><div class="line"><a name="l00533"></a><span class="lineno"> 533</span>  <span class="keyword">auto</span> sum_res = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a475ee3c8ef70a2e4f18effd1d527a63a">wrapper::vpadd</a>(<a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a95ee388aa7c5bccab918235dc538a6b3">wrapper::vgethigh</a>(vec_sum), <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a2902775707bc7bf7d6da1bda1cc15783">wrapper::vgetlow</a>(vec_sum));</div><div class="line"><a name="l00534"></a><span class="lineno"> 534</span>  <span class="keywordflow">for</span>(<span class="keywordtype">int</span> i = 0; i < sum_stages; ++i)</div><div class="line"><a name="l00535"></a><span class="lineno"> 535</span>  {</div><div class="line"><a name="l00536"></a><span class="lineno"> 536</span>  sum_res = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a475ee3c8ef70a2e4f18effd1d527a63a">wrapper::vpadd</a>(sum_res, sum_res);</div><div class="line"><a name="l00537"></a><span class="lineno"> 537</span>  }</div><div class="line"><a name="l00538"></a><span class="lineno"> 538</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#aa16ace001ab8287faa46d6962f369219">wrapper::vgetlane</a>(sum_res, 0);</div><div class="line"><a name="l00539"></a><span class="lineno"> 539</span> </div><div class="line"><a name="l00540"></a><span class="lineno"> 540</span>  <span class="comment">/* Run remaining elements */</span></div><div class="line"><a name="l00541"></a><span class="lineno"> 541</span>  <span class="keywordflow">for</span>(; x < input_width; ++x)</div><div class="line"><a name="l00542"></a><span class="lineno"> 542</span>  {</div><div class="line"><a name="l00543"></a><span class="lineno"> 543</span>  T element{};</div><div class="line"><a name="l00544"></a><span class="lineno"> 544</span> </div><div class="line"><a name="l00545"></a><span class="lineno"> 545</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00546"></a><span class="lineno"> 546</span>  {</div><div class="line"><a name="l00547"></a><span class="lineno"> 547</span>  element = (in_ptr[x] - max_val) * beta;</div><div class="line"><a name="l00548"></a><span class="lineno"> 548</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> += std::exp(element);</div><div class="line"><a name="l00549"></a><span class="lineno"> 549</span>  }</div><div class="line"><a name="l00550"></a><span class="lineno"> 550</span>  <span class="keywordflow">else</span></div><div class="line"><a name="l00551"></a><span class="lineno"> 551</span>  {</div><div class="line"><a name="l00552"></a><span class="lineno"> 552</span>  element = std::exp((in_ptr[x] - max_val) * beta);</div><div class="line"><a name="l00553"></a><span class="lineno"> 553</span>  <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a> += element;</div><div class="line"><a name="l00554"></a><span class="lineno"> 554</span>  }</div><div class="line"><a name="l00555"></a><span class="lineno"> 555</span>  tmp_ptr[x] = element;</div><div class="line"><a name="l00556"></a><span class="lineno"> 556</span>  }</div><div class="line"><a name="l00557"></a><span class="lineno"> 557</span> </div><div class="line"><a name="l00558"></a><span class="lineno"> 558</span>  <span class="keywordflow">if</span>(!is_log)</div><div class="line"><a name="l00559"></a><span class="lineno"> 559</span>  {</div><div class="line"><a name="l00560"></a><span class="lineno"> 560</span>  sum_inversed = T(1) / <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>;</div><div class="line"><a name="l00561"></a><span class="lineno"> 561</span>  }</div><div class="line"><a name="l00562"></a><span class="lineno"> 562</span>  }</div><div class="line"><a name="l00563"></a><span class="lineno"> 563</span> </div><div class="line"><a name="l00564"></a><span class="lineno"> 564</span>  <span class="comment">/* Normalize exponentials */</span></div><div class="line"><a name="l00565"></a><span class="lineno"> 565</span>  {</div><div class="line"><a name="l00566"></a><span class="lineno"> 566</span>  <span class="comment">/* Loop over row and compute softmax */</span></div><div class="line"><a name="l00567"></a><span class="lineno"> 567</span>  <span class="keywordtype">int</span> x = 0;</div><div class="line"><a name="l00568"></a><span class="lineno"> 568</span>  <span class="keywordflow">for</span>(; x <= (input_width - vec_size); x += vec_size)</div><div class="line"><a name="l00569"></a><span class="lineno"> 569</span>  {</div><div class="line"><a name="l00570"></a><span class="lineno"> 570</span>  <span class="keyword">auto</span> vec_in = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a77f54eded7fef436d3a4f21ad5a00da6">wrapper::vloadq</a>(tmp_ptr + x);</div><div class="line"><a name="l00571"></a><span class="lineno"> 571</span>  <span class="keyword">auto</span> normalized_value = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(0), ExactTagType{});</div><div class="line"><a name="l00572"></a><span class="lineno"> 572</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00573"></a><span class="lineno"> 573</span>  {</div><div class="line"><a name="l00574"></a><span class="lineno"> 574</span>  normalized_value = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#acae4ec59246f93845fa78120d83d7c1e">wrapper::vsub</a>(vec_in, <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(<a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>), ExactTagType{}));</div><div class="line"><a name="l00575"></a><span class="lineno"> 575</span>  }</div><div class="line"><a name="l00576"></a><span class="lineno"> 576</span>  <span class="keywordflow">else</span></div><div class="line"><a name="l00577"></a><span class="lineno"> 577</span>  {</div><div class="line"><a name="l00578"></a><span class="lineno"> 578</span>  normalized_value = <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a66c8dd218a4863a14fec363e3dd13dc4">wrapper::vmul</a>(vec_in, <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#a39e87435be178fba49b76f49426ef873">wrapper::vdup_n</a>(static_cast<T>(sum_inversed), ExactTagType{}));</div><div class="line"><a name="l00579"></a><span class="lineno"> 579</span>  }</div><div class="line"><a name="l00580"></a><span class="lineno"> 580</span>  <a class="code" href="namespacearm__compute_1_1wrapper.xhtml#ae7943ea9c1f74dc72c62d4cc3966a459">wrapper::vstore</a>(out_ptr + x, normalized_value);</div><div class="line"><a name="l00581"></a><span class="lineno"> 581</span>  }</div><div class="line"><a name="l00582"></a><span class="lineno"> 582</span>  <span class="comment">/* Run remaining elements */</span></div><div class="line"><a name="l00583"></a><span class="lineno"> 583</span>  <span class="keywordflow">for</span>(; x < input_width; ++x)</div><div class="line"><a name="l00584"></a><span class="lineno"> 584</span>  {</div><div class="line"><a name="l00585"></a><span class="lineno"> 585</span>  <span class="keywordflow">if</span>(is_log)</div><div class="line"><a name="l00586"></a><span class="lineno"> 586</span>  {</div><div class="line"><a name="l00587"></a><span class="lineno"> 587</span>  out_ptr[x] = tmp_ptr[x] - <a class="code" href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a>;</div><div class="line"><a name="l00588"></a><span class="lineno"> 588</span>  }</div><div class="line"><a name="l00589"></a><span class="lineno"> 589</span>  <span class="keywordflow">else</span></div><div class="line"><a name="l00590"></a><span class="lineno"> 590</span>  {</div><div class="line"><a name="l00591"></a><span class="lineno"> 591</span>  out_ptr[x] = tmp_ptr[x] * sum_inversed;</div><div class="line"><a name="l00592"></a><span class="lineno"> 592</span>  }</div><div class="line"><a name="l00593"></a><span class="lineno"> 593</span>  }</div><div class="line"><a name="l00594"></a><span class="lineno"> 594</span>  }</div><div class="line"><a name="l00595"></a><span class="lineno"> 595</span>  },</div><div class="line"><a name="l00596"></a><span class="lineno"> 596</span>  in_it, max_it, out_it);</div><div class="line"><a name="l00597"></a><span class="lineno"> 597</span> }</div><div class="line"><a name="l00598"></a><span class="lineno"> 598</span> } <span class="comment">// namespace</span></div><div class="line"><a name="l00599"></a><span class="lineno"> 599</span> </div><div class="line"><a name="l00600"></a><span class="lineno"> 600</span> <span class="keyword">template</span> <<span class="keywordtype">bool</span> IS_LOG></div><div class="line"><a name="l00601"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a83bde72fd3fc3d67311f78315eafd23c"> 601</a></span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a83bde72fd3fc3d67311f78315eafd23c">NELogits1DSoftmaxKernel<IS_LOG>::NELogits1DSoftmaxKernel</a>()</div><div class="line"><a name="l00602"></a><span class="lineno"> 602</span>  : _func(nullptr), _input(nullptr), _max(nullptr), _output(nullptr), _beta(1.0f), _tmp(nullptr)</div><div class="line"><a name="l00603"></a><span class="lineno"> 603</span> {</div><div class="line"><a name="l00604"></a><span class="lineno"> 604</span> }</div><div class="line"><a name="l00605"></a><span class="lineno"> 605</span> </div><div class="line"><a name="l00606"></a><span class="lineno"> 606</span> <span class="keyword">template</span> <<span class="keywordtype">bool</span> IS_LOG></div><div class="line"><a name="l00607"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ad7a5f63cd7550fd2712d6c0e1e677419"> 607</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ad7a5f63cd7550fd2712d6c0e1e677419">NELogits1DSoftmaxKernel<IS_LOG>::configure</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *max, <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *output, <span class="keyword">const</span> <span class="keywordtype">float</span> beta, <a class="code" href="classarm__compute_1_1_i_tensor.xhtml">ITensor</a> *tmp)</div><div class="line"><a name="l00608"></a><span class="lineno"> 608</span> {</div><div class="line"><a name="l00609"></a><span class="lineno"> 609</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, max, output, tmp);</div><div class="line"><a name="l00610"></a><span class="lineno"> 610</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), max-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), tmp-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>());</div><div class="line"><a name="l00611"></a><span class="lineno"> 611</span>  <span class="comment">// Perform validation step</span></div><div class="line"><a name="l00612"></a><span class="lineno"> 612</span>  <a class="code" href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a>(validate_arguments_logits_softmax(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), *max-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), *output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), beta, *tmp-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), IS_LOG));</div><div class="line"><a name="l00613"></a><span class="lineno"> 613</span>  <span class="comment">// Configure kernel window</span></div><div class="line"><a name="l00614"></a><span class="lineno"> 614</span>  <span class="keyword">auto</span> win_config = validate_and_configure_window_logits_softmax(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info(), *max-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), *output-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), *tmp-><a class="code" href="classarm__compute_1_1_i_tensor.xhtml#a0e95dc1e53c361348314873b168ae237">info</a>(), IS_LOG);</div><div class="line"><a name="l00615"></a><span class="lineno"> 615</span>  <a class="code" href="_error_8h.xhtml#a938dcd406ce611ef5345ad2531cdb948">ARM_COMPUTE_ERROR_THROW_ON</a>(win_config.first);</div><div class="line"><a name="l00616"></a><span class="lineno"> 616</span> </div><div class="line"><a name="l00617"></a><span class="lineno"> 617</span>  <span class="keywordflow">switch</span>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->info()->data_type())</div><div class="line"><a name="l00618"></a><span class="lineno"> 618</span>  {</div><div class="line"><a name="l00619"></a><span class="lineno"> 619</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924">DataType::QASYMM8</a>:</div><div class="line"><a name="l00620"></a><span class="lineno"> 620</span>  _func = &logits_1d_softmax_qasymm8<qasymm8_t, IS_LOG>;</div><div class="line"><a name="l00621"></a><span class="lineno"> 621</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00622"></a><span class="lineno"> 622</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9">DataType::QASYMM8_SIGNED</a>:</div><div class="line"><a name="l00623"></a><span class="lineno"> 623</span>  _func = &logits_1d_softmax_qasymm8<qasymm8_signed_t, IS_LOG>;</div><div class="line"><a name="l00624"></a><span class="lineno"> 624</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00625"></a><span class="lineno"> 625</span> <span class="preprocessor">#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC</span></div><div class="line"><a name="l00626"></a><span class="lineno"> 626</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a56d8353718e6fdc78b8d69078a2cdb94">DataType::F16</a>:</div><div class="line"><a name="l00627"></a><span class="lineno"> 627</span>  _func = &logits_1d_softmax_float<float16_t, IS_LOG>;</div><div class="line"><a name="l00628"></a><span class="lineno"> 628</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00629"></a><span class="lineno"> 629</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="l00630"></a><span class="lineno"> 630</span>  <span class="keywordflow">case</span> <a class="code" href="namespacearm__compute.xhtml#ab4e88c89b3b7ea1735996cc4def22d58a44ad4ef5a76e6aa6fb3e3fa079a54fda">DataType::F32</a>:</div><div class="line"><a name="l00631"></a><span class="lineno"> 631</span>  _func = &logits_1d_softmax_float<float, IS_LOG>;</div><div class="line"><a name="l00632"></a><span class="lineno"> 632</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00633"></a><span class="lineno"> 633</span>  <span class="keywordflow">default</span>:</div><div class="line"><a name="l00634"></a><span class="lineno"> 634</span>  <a class="code" href="_error_8h.xhtml#a7cf8d8b669b8f7b05680230be30d60f4">ARM_COMPUTE_ERROR</a>(<span class="stringliteral">"Unsupported data type."</span>);</div><div class="line"><a name="l00635"></a><span class="lineno"> 635</span>  <span class="keywordflow">break</span>;</div><div class="line"><a name="l00636"></a><span class="lineno"> 636</span>  }</div><div class="line"><a name="l00637"></a><span class="lineno"> 637</span> </div><div class="line"><a name="l00638"></a><span class="lineno"> 638</span>  _input = <a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>;</div><div class="line"><a name="l00639"></a><span class="lineno"> 639</span>  _max = max;</div><div class="line"><a name="l00640"></a><span class="lineno"> 640</span>  _output = output;</div><div class="line"><a name="l00641"></a><span class="lineno"> 641</span>  _beta = beta;</div><div class="line"><a name="l00642"></a><span class="lineno"> 642</span>  _tmp = tmp;</div><div class="line"><a name="l00643"></a><span class="lineno"> 643</span> </div><div class="line"><a name="l00644"></a><span class="lineno"> 644</span>  INEKernel::configure(win_config.second);</div><div class="line"><a name="l00645"></a><span class="lineno"> 645</span> }</div><div class="line"><a name="l00646"></a><span class="lineno"> 646</span> </div><div class="line"><a name="l00647"></a><span class="lineno"> 647</span> <span class="keyword">template</span> <<span class="keywordtype">bool</span> IS_LOG></div><div class="line"><a name="l00648"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ae9a64d99f08581b961e1ac5a5fc46af8"> 648</a></span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ae9a64d99f08581b961e1ac5a5fc46af8">NELogits1DSoftmaxKernel<IS_LOG>::validate</a>(<span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *max,</div><div class="line"><a name="l00649"></a><span class="lineno"> 649</span>  <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *output, <span class="keyword">const</span> <span class="keywordtype">float</span> beta, <span class="keyword">const</span> <a class="code" href="classarm__compute_1_1_i_tensor_info.xhtml">ITensorInfo</a> *tmp)</div><div class="line"><a name="l00650"></a><span class="lineno"> 650</span> {</div><div class="line"><a name="l00651"></a><span class="lineno"> 651</span>  <a class="code" href="_validate_8h.xhtml#a921b705e9e3e0fe928928447869e62a5">ARM_COMPUTE_ERROR_ON_NULLPTR</a>(<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, max, output, tmp);</div><div class="line"><a name="l00652"></a><span class="lineno"> 652</span> </div><div class="line"><a name="l00653"></a><span class="lineno"> 653</span>  <a class="code" href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a>(validate_arguments_logits_softmax(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>, *max, *output, beta, *tmp, IS_LOG));</div><div class="line"><a name="l00654"></a><span class="lineno"> 654</span>  <a class="code" href="_error_8h.xhtml#a8a1e1c105f0bdaf37db408c7cfcb77a4">ARM_COMPUTE_RETURN_ON_ERROR</a>(validate_and_configure_window_logits_softmax(*<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">input</a>->clone(), *max-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>(), *output-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>(), *tmp-><a class="code" href="classarm__compute_1_1misc_1_1_i_cloneable.xhtml#a4d10e5012a872e7f78f2b539b673049d">clone</a>(), IS_LOG).first);</div><div class="line"><a name="l00655"></a><span class="lineno"> 655</span> </div><div class="line"><a name="l00656"></a><span class="lineno"> 656</span>  <span class="keywordflow">return</span> <a class="code" href="classarm__compute_1_1_status.xhtml">Status</a>{};</div><div class="line"><a name="l00657"></a><span class="lineno"> 657</span> }</div><div class="line"><a name="l00658"></a><span class="lineno"> 658</span> </div><div class="line"><a name="l00659"></a><span class="lineno"> 659</span> <span class="keyword">template</span> <<span class="keywordtype">bool</span> IS_LOG></div><div class="line"><a name="l00660"></a><span class="lineno"><a class="line" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82"> 660</a></span> <span class="keywordtype">void</span> <a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">NELogits1DSoftmaxKernel<IS_LOG>::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="l00661"></a><span class="lineno"> 661</span> {</div><div class="line"><a name="l00662"></a><span class="lineno"> 662</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="l00663"></a><span class="lineno"> 663</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="l00664"></a><span class="lineno"> 664</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">INEKernel::window</a>(), window);</div><div class="line"><a name="l00665"></a><span class="lineno"> 665</span> </div><div class="line"><a name="l00666"></a><span class="lineno"> 666</span>  <span class="keyword">const</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a> = _input->info()->valid_region().shape.x();</div><div class="line"><a name="l00667"></a><span class="lineno"> 667</span>  <span class="keyword">const</span> <span class="keywordtype">unsigned</span> <span class="keywordtype">int</span> tmp_size_for_thread = _tmp->info()->element_size() * <a class="code" href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a>;</div><div class="line"><a name="l00668"></a><span class="lineno"> 668</span> </div><div class="line"><a name="l00669"></a><span class="lineno"> 669</span>  <a class="code" href="_error_8h.xhtml#a54a6080c9f4df1f908e57a9bbb46f5da">ARM_COMPUTE_ERROR_ON</a>(_tmp->info()->total_size() < (<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a4f4125dba5283887b34f889b1c615c0c">info</a>.num_threads * tmp_size_for_thread));</div><div class="line"><a name="l00670"></a><span class="lineno"> 670</span> </div><div class="line"><a name="l00671"></a><span class="lineno"> 671</span>  <span class="keywordtype">void</span> *tmp_for_thread = _tmp->buffer() + (<a class="code" href="namespacearm__compute_1_1test_1_1validation.xhtml#a4f4125dba5283887b34f889b1c615c0c">info</a>.thread_id * tmp_size_for_thread);</div><div class="line"><a name="l00672"></a><span class="lineno"> 672</span> </div><div class="line"><a name="l00673"></a><span class="lineno"> 673</span>  (*_func)(*_input, *_max, tmp_for_thread, *_output, _beta, window);</div><div class="line"><a name="l00674"></a><span class="lineno"> 674</span> }</div><div class="line"><a name="l00675"></a><span class="lineno"> 675</span> </div><div class="line"><a name="l00676"></a><span class="lineno"> 676</span> <span class="keyword">template</span> <span class="keyword">class </span><a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml">NELogits1DSoftmaxKernel<true></a>;</div><div class="line"><a name="l00677"></a><span class="lineno"> 677</span> <span class="keyword">template</span> <span class="keyword">class </span><a class="code" href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml">NELogits1DSoftmaxKernel<false></a>;</div><div class="line"><a name="l00678"></a><span class="lineno"> 678</span> </div><div class="line"><a name="l00679"></a><span class="lineno"> 679</span> } <span class="comment">// namespace arm_compute</span></div><div class="ttc" id="_n_e_softmax_layer_kernel_8h_xhtml"><div class="ttname"><a href="_n_e_softmax_layer_kernel_8h.xhtml">NESoftmaxLayerKernel.h</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="classarm__compute_1_1_tensor_shape_xhtml"><div class="ttname"><a href="classarm__compute_1_1_tensor_shape.xhtml">arm_compute::TensorShape</a></div><div class="ttdoc">Shape of a tensor.</div><div class="ttdef"><b>Definition:</b> <a href="_tensor_shape_8h_source.xhtml#l00039">TensorShape.h:39</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="structarm__compute_1_1_border_size_xhtml"><div class="ttname"><a href="structarm__compute_1_1_border_size.xhtml">arm_compute::BorderSize</a></div><div class="ttdoc">Container for 2D border size.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_types_8h_source.xhtml#l00269">Types.h:269</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_xhtml_a6048f3cbd52aea7ce3a693ce485393be"><div class="ttname"><a href="namespacearm__compute.xhtml#a6048f3cbd52aea7ce3a693ce485393be">arm_compute::convert_float_to_int< float32x4x4_t, int8x16_t ></a></div><div class="ttdeci">int8x16_t convert_float_to_int< float32x4x4_t, int8x16_t >(const float32x4x4_t &in)</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00062">NESoftmaxLayerKernel.cpp:62</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a7cf8d8b669b8f7b05680230be30d60f4"><div class="ttname"><a href="_error_8h.xhtml#a7cf8d8b669b8f7b05680230be30d60f4">ARM_COMPUTE_ERROR</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR(msg)</div><div class="ttdoc">Print the given message then throw an std::runtime_error.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00352">Error.h:352</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a9075aa895fe4b2368e3c0cfae26ebf0b"><div class="ttname"><a href="namespacearm__compute.xhtml#a9075aa895fe4b2368e3c0cfae26ebf0b">arm_compute::convert_int_to_float< float32x4x4_t, int8x16_t ></a></div><div class="ttdeci">float32x4x4_t convert_int_to_float< float32x4x4_t, int8x16_t >(const int8x16_t &in)</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00076">NESoftmaxLayerKernel.cpp:76</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="reduction__operation_8cl_xhtml_ab0df00f5333da51860deb93deb44a782"><div class="ttname"><a href="reduction__operation_8cl.xhtml#ab0df00f5333da51860deb93deb44a782">sum</a></div><div class="ttdeci">DATA_TYPE sum(__global const DATA_TYPE *input)</div><div class="ttdoc">Calculate sum of a vector.</div><div class="ttdef"><b>Definition:</b> <a href="reduction__operation_8cl_source.xhtml#l00066">reduction_operation.cl:66</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#l00204">Error.h:204</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="_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#l00792">Validate.h:792</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="classarm__compute_1_1_n_e_logits1_d_max_kernel_xhtml_a7dcfcda37359d31c2510ad865f4ca051"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a7dcfcda37359d31c2510ad865f4ca051">arm_compute::NELogits1DMaxKernel::NELogits1DMaxKernel</a></div><div class="ttdeci">NELogits1DMaxKernel()</div><div class="ttdoc">Default constructor.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00168">NESoftmaxLayerKernel.cpp:168</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a4a20779ec9b583a9468b4fa60575d52e"><div class="ttname"><a href="namespacearm__compute.xhtml#a4a20779ec9b583a9468b4fa60575d52e">arm_compute::convert_int8x16_to_float32x4x4</a></div><div class="ttdeci">float32x4x4_t convert_int8x16_to_float32x4x4(const int8x16_t &in)</div><div class="ttdoc">Converts from int8x16 to float32x4x4_t.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_math_8inl_source.xhtml#l00334">NEMath.inl:334</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a1b362fecc52ac83bcfbb0850549edce4"><div class="ttname"><a href="namespacearm__compute.xhtml#a1b362fecc52ac83bcfbb0850549edce4">arm_compute::convert_uint8x16_to_float32x4x4</a></div><div class="ttdeci">float32x4x4_t convert_uint8x16_to_float32x4x4(const uint8x16_t &in)</div><div class="ttdoc">Converts from uint8x16 to float32x4x4_t.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_math_8inl_source.xhtml#l00320">NEMath.inl:320</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml_a54a6080c9f4df1f908e57a9bbb46f5da"><div class="ttname"><a href="_error_8h.xhtml#a54a6080c9f4df1f908e57a9bbb46f5da">ARM_COMPUTE_ERROR_ON</a></div><div class="ttdeci">#define ARM_COMPUTE_ERROR_ON(cond)</div><div class="ttdoc">If the condition is true then an error message is printed and an exception thrown.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00466">Error.h:466</a></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="namespacearm__compute_xhtml_a1dcccbff6075b2fad30ebd2a0da6b995"><div class="ttname"><a href="namespacearm__compute.xhtml#a1dcccbff6075b2fad30ebd2a0da6b995">arm_compute::get_softmax_output_quantization_info</a></div><div class="ttdeci">QuantizationInfo get_softmax_output_quantization_info(DataType input_type, bool is_log)</div><div class="ttdoc">Returns output quantization information for softmax layer.</div><div class="ttdef"><b>Definition:</b> <a href="src_2core_2_utils_8cpp_source.xhtml#l00445">Utils.cpp:445</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#l00455">Error.h:455</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="namespacearm__compute_xhtml_a17bff173004242d8f44b5b663177f9c0"><div class="ttname"><a href="namespacearm__compute.xhtml#a17bff173004242d8f44b5b663177f9c0">arm_compute::convert_float_to_int</a></div><div class="ttdeci">int_vec_type convert_float_to_int(const float_vec_type &in)</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#l00296">Error.h:296</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) 2017-2020 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="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#l00202">Helpers.inl:202</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_n_e_logits1_d_max_kernel_xhtml_a112b35dd205c62ea6ed1447ef226da82"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">arm_compute::NELogits1DMaxKernel::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_softmax_layer_kernel_8cpp_source.xhtml#l00230">NESoftmaxLayerKernel.cpp:230</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1test_1_1validation_xhtml_a8fcf2ddd9a1d58b1b280f5c0aed71845"><div class="ttname"><a href="namespacearm__compute_1_1test_1_1validation.xhtml#a8fcf2ddd9a1d58b1b280f5c0aed71845">arm_compute::test::validation::input</a></div><div class="ttdeci">auto input</div><div class="ttdef"><b>Definition:</b> <a href="_c_l_2_l_s_t_m_layer_quantized_8cpp_source.xhtml#l00487">LSTMLayerQuantized.cpp:487</a></div></div> |
| <div class="ttc" id="_saturate_cast_8h_xhtml"><div class="ttname"><a href="_saturate_cast_8h.xhtml">SaturateCast.h</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_softmax_kernel_xhtml_a83bde72fd3fc3d67311f78315eafd23c"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a83bde72fd3fc3d67311f78315eafd23c">arm_compute::NELogits1DSoftmaxKernel::NELogits1DSoftmaxKernel</a></div><div class="ttdeci">NELogits1DSoftmaxKernel()</div><div class="ttdoc">Default constructor.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00601">NESoftmaxLayerKernel.cpp:601</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_softmax_kernel_xhtml_ae9a64d99f08581b961e1ac5a5fc46af8"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ae9a64d99f08581b961e1ac5a5fc46af8">arm_compute::NELogits1DSoftmaxKernel::validate</a></div><div class="ttdeci">static Status validate(const ITensorInfo *input, const ITensorInfo *max, const ITensorInfo *output, const float beta, const ITensorInfo *tmp)</div><div class="ttdoc">Static function to check if given info will lead to a valid configuration of NELogits1DSoftmaxKernel.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00648">NESoftmaxLayerKernel.cpp:648</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="classarm__compute_1_1_quantization_info_xhtml"><div class="ttname"><a href="classarm__compute_1_1_quantization_info.xhtml">arm_compute::QuantizationInfo</a></div><div class="ttdoc">Quantization information.</div><div class="ttdef"><b>Definition:</b> <a href="_quantization_info_8h_source.xhtml#l00069">QuantizationInfo.h:69</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a475ee3c8ef70a2e4f18effd1d527a63a"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a475ee3c8ef70a2e4f18effd1d527a63a">arm_compute::wrapper::vpadd</a></div><div class="ttdeci">uint8x8_t vpadd(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="intrinsics_2add_8h_source.xhtml#l00187">add.h:187</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_softmax_kernel_xhtml_ad7a5f63cd7550fd2712d6c0e1e677419"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#ad7a5f63cd7550fd2712d6c0e1e677419">arm_compute::NELogits1DSoftmaxKernel::configure</a></div><div class="ttdeci">void configure(const ITensor *input, const ITensor *max, ITensor *output, const float beta, ITensor *tmp)</div><div class="ttdoc">Set the input and output tensors.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00607">NESoftmaxLayerKernel.cpp:607</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_aa16ace001ab8287faa46d6962f369219"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#aa16ace001ab8287faa46d6962f369219">arm_compute::wrapper::vgetlane</a></div><div class="ttdeci">uint8_t vgetlane(const uint8x8_t vector, const unsigned int lane)</div><div class="ttdef"><b>Definition:</b> <a href="getlane_8h_source.xhtml#l00091">getlane.h:91</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_afc4bd8e872567d9c4c57d89eb0bb3da1"><div class="ttname"><a href="namespacearm__compute.xhtml#afc4bd8e872567d9c4c57d89eb0bb3da1">arm_compute::update_window_and_padding</a></div><div class="ttdeci">bool update_window_and_padding(Window &win, Ts &&... patterns)</div><div class="ttdoc">Update window and padding size for each of the access patterns.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_helpers_8h_source.xhtml#l00402">Helpers.h:402</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#l00152">Error.h:152</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="namespacearm__compute_xhtml_a1ce9b523fd4f3b5bbcadcd796183455aa4c614360da93c0a041b22e537de151eb"><div class="ttname"><a href="namespacearm__compute.xhtml#a1ce9b523fd4f3b5bbcadcd796183455aa4c614360da93c0a041b22e537de151eb">arm_compute::Channel::U</a></div><div class="ttdoc">Cb/U channel.</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_i_tensor_info_xhtml_a7c66505457d00ece3aa4b34cab80757d"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a7c66505457d00ece3aa4b34cab80757d">arm_compute::ITensorInfo::tensor_shape</a></div><div class="ttdeci">virtual const TensorShape & tensor_shape() const =0</div><div class="ttdoc">Size for each dimension of the tensor.</div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ab237a0a375cf382d52b61653248d3d4a"><div class="ttname"><a href="namespacearm__compute.xhtml#ab237a0a375cf382d52b61653248d3d4a">arm_compute::ceil_to_multiple</a></div><div class="ttdeci">auto ceil_to_multiple(S value, T divisor) -> decltype(((value+divisor - 1)/divisor) *divisor)</div><div class="ttdoc">Computes the smallest number larger or equal to value that is a multiple of divisor.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_utils_8h_source.xhtml#l00066">Utils.h:66</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml_a1da797d2762c1cdbb73bfc83136c3a38"><div class="ttname"><a href="_validate_8h.xhtml#a1da797d2762c1cdbb73bfc83136c3a38">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00288">Validate.h:288</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924"><div class="ttname"><a href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6af14462d71aa842202c3e4b272c7ec924">arm_compute::DataType::QASYMM8</a></div><div class="ttdoc">quantized, asymmetric fixed-point 8-bit number unsigned</div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a9c673ade3003d5d4e28ebf5875795dc4"><div class="ttname"><a href="namespacearm__compute.xhtml#a9c673ade3003d5d4e28ebf5875795dc4">arm_compute::convert_float32x4x4_to_int8x16</a></div><div class="ttdeci">void convert_float32x4x4_to_int8x16(const float32x4x4_t &in, int8x16_t &out)</div><div class="ttdoc">Converts from float32x4x4_t to just one int8x16_t.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_math_8inl_source.xhtml#l00367">NEMath.inl:367</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ad63fda57a59d2dcc82eb13599581fcc0"><div class="ttname"><a href="namespacearm__compute.xhtml#ad63fda57a59d2dcc82eb13599581fcc0">arm_compute::convert_float_to_int< float32x4x4_t, uint8x16_t ></a></div><div class="ttdeci">uint8x16_t convert_float_to_int< float32x4x4_t, uint8x16_t >(const float32x4x4_t &in)</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00054">NESoftmaxLayerKernel.cpp:54</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1utils_1_1cast_xhtml_a2ea3d1fc01a3a442900249ca182ffa5e"><div class="ttname"><a href="namespacearm__compute_1_1utils_1_1cast.xhtml#a2ea3d1fc01a3a442900249ca182ffa5e">arm_compute::utils::cast::U</a></div><div class="ttdeci">U</div><div class="ttdef"><b>Definition:</b> <a href="_saturate_cast_8h_source.xhtml#l00057">SaturateCast.h:57</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_abb7e0f23a4f2e63f39433f158dad47ab"><div class="ttname"><a href="namespacearm__compute.xhtml#abb7e0f23a4f2e63f39433f158dad47ab">arm_compute::data_size_from_type</a></div><div class="ttdeci">size_t data_size_from_type(DataType data_type)</div><div class="ttdoc">The size in bytes of the data type.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_utils_8h_source.xhtml#l00109">Utils.h:109</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a2902775707bc7bf7d6da1bda1cc15783"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a2902775707bc7bf7d6da1bda1cc15783">arm_compute::wrapper::vgetlow</a></div><div class="ttdeci">uint8x8_t vgetlow(const uint8x16_t val)</div><div class="ttdef"><b>Definition:</b> <a href="getlow_8h_source.xhtml#l00039">getlow.h:39</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_max_kernel_xhtml_a83a344e60eb7db895953a942abf16628"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a83a344e60eb7db895953a942abf16628">arm_compute::NELogits1DMaxKernel::configure</a></div><div class="ttdeci">void configure(const ITensor *input, ITensor *output)</div><div class="ttdoc">Set the input and output tensors.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00178">NESoftmaxLayerKernel.cpp:178</a></div></div> |
| <div class="ttc" id="_error_8h_xhtml"><div class="ttname"><a href="_error_8h.xhtml">Error.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a59e56af19e754a6aa26a612ebf91d05fa62be47fdd89da032cf78dfce82239579"><div class="ttname"><a href="namespacearm__compute.xhtml#a59e56af19e754a6aa26a612ebf91d05fa62be47fdd89da032cf78dfce82239579">arm_compute::ErrorCode::RUNTIME_ERROR</a></div><div class="ttdoc">Generic runtime error.</div></div> |
| <div class="ttc" id="classarm__compute_1_1_i_tensor_info_xhtml_a3f3e1a3200223e6a304a533b1016e749"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a3f3e1a3200223e6a304a533b1016e749">arm_compute::ITensorInfo::quantization_info</a></div><div class="ttdeci">virtual QuantizationInfo quantization_info() const =0</div><div class="ttdoc">Get the quantization settings (scale and offset) of the tensor.</div></div> |
| <div class="ttc" id="_validate_8h_xhtml_aba910b683652be1f65437ef37a9da2a9"><div class="ttname"><a href="_validate_8h.xhtml#aba910b683652be1f65437ef37a9da2a9">ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO</a></div><div class="ttdeci">#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(...)</div><div class="ttdef"><b>Definition:</b> <a href="_validate_8h_source.xhtml#l00610">Validate.h:610</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a14f46283f316e7f0fad301d5c1507e9f"><div class="ttname"><a href="namespacearm__compute.xhtml#a14f46283f316e7f0fad301d5c1507e9f">arm_compute::is_data_type_quantized_asymmetric</a></div><div class="ttdeci">bool is_data_type_quantized_asymmetric(DataType dt)</div><div class="ttdoc">Check if a given data type is of asymmetric quantized type.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_utils_8h_source.xhtml#l01139">Utils.h:1139</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a95ee388aa7c5bccab918235dc538a6b3"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a95ee388aa7c5bccab918235dc538a6b3">arm_compute::wrapper::vgethigh</a></div><div class="ttdeci">uint8x8_t vgethigh(const uint8x16_t val)</div><div class="ttdef"><b>Definition:</b> <a href="gethigh_8h_source.xhtml#l00039">gethigh.h:39</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="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_i_tensor_info_xhtml_a18064e0011c3869d884653e9e7c47b66"><div class="ttname"><a href="classarm__compute_1_1_i_tensor_info.xhtml#a18064e0011c3869d884653e9e7c47b66">arm_compute::ITensorInfo::total_size</a></div><div class="ttdeci">virtual size_t total_size() const =0</div><div class="ttdoc">Returns the total size of the tensor in bytes.</div></div> |
| <div class="ttc" id="_error_8h_xhtml_af1b8ff8eb557a2ad11272f1505f45d34"><div class="ttname"><a href="_error_8h.xhtml#af1b8ff8eb557a2ad11272f1505f45d34">ARM_COMPUTE_CREATE_ERROR</a></div><div class="ttdeci">#define ARM_COMPUTE_CREATE_ERROR(error_code, msg)</div><div class="ttdoc">Creates an error with a given message.</div><div class="ttdef"><b>Definition:</b> <a href="_error_8h_source.xhtml#l00159">Error.h:159</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ad339c829071ae619193300cb501195ac"><div class="ttname"><a href="namespacearm__compute.xhtml#ad339c829071ae619193300cb501195ac">arm_compute::convert_int_to_float</a></div><div class="ttdeci">float_vec_type convert_int_to_float(const int_vec_type &in)</div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_max_kernel_xhtml_a423f9a45a52983b4de5e2b347f4369c7"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a423f9a45a52983b4de5e2b347f4369c7">arm_compute::NELogits1DMaxKernel::border_size</a></div><div class="ttdeci">BorderSize border_size() const override</div><div class="ttdoc">The size of the border for that kernel.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00173">NESoftmaxLayerKernel.cpp:173</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_softmax_kernel_xhtml"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml">arm_compute::NELogits1DSoftmaxKernel</a></div><div class="ttdoc">Interface for softmax computation for QASYMM8 with pre-computed max.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8h_source.xhtml#l00073">NESoftmaxLayerKernel.h:73</a></div></div> |
| <div class="ttc" id="_n_e_math_8h_xhtml"><div class="ttname"><a href="_n_e_math_8h.xhtml">NEMath.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ab55486f1153a4c5f507d4923b59fc6a0"><div class="ttname"><a href="namespacearm__compute.xhtml#ab55486f1153a4c5f507d4923b59fc6a0">arm_compute::vexpq_f32</a></div><div class="ttdeci">float32x4_t vexpq_f32(float32x4_t x)</div><div class="ttdoc">Calculate exponential.</div></div> |
| <div class="ttc" id="namespacearm__compute_1_1test_1_1validation_xhtml_a306c2d0ce612405d1d4ee732534dc543"><div class="ttname"><a href="namespacearm__compute_1_1test_1_1validation.xhtml#a306c2d0ce612405d1d4ee732534dc543">arm_compute::test::validation::sub</a></div><div class="ttdeci">CLArithmeticSubtraction sub</div><div class="ttdef"><b>Definition:</b> <a href="_c_l_2_arithmetic_subtraction_8cpp_source.xhtml#l00226">ArithmeticSubtraction.cpp:226</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a8677ddae33f84bb0e84e995c54bf6ba3"><div class="ttname"><a href="namespacearm__compute.xhtml#a8677ddae33f84bb0e84e995c54bf6ba3">arm_compute::convert_int_to_float< float32x4x4_t, uint8x16_t ></a></div><div class="ttdeci">float32x4x4_t convert_int_to_float< float32x4x4_t, uint8x16_t >(const uint8x16_t &in)</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00070">NESoftmaxLayerKernel.cpp:70</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="namespacearm__compute_1_1test_1_1validation_xhtml_ab1806bf0c5a41f674fb9d2dc6af644f5"><div class="ttname"><a href="namespacearm__compute_1_1test_1_1validation.xhtml#ab1806bf0c5a41f674fb9d2dc6af644f5">arm_compute::test::validation::output_shape</a></div><div class="ttdeci">output_shape</div><div class="ttdef"><b>Definition:</b> <a href="_c_l_2_convolution_layer_8cpp_source.xhtml#l00182">ConvolutionLayer.cpp:182</a></div></div> |
| <div class="ttc" id="_c_l_im2_col_kernel_8cpp_xhtml_a4e45c1f5e4280813a78a77dda71d8799"><div class="ttname"><a href="_c_l_im2_col_kernel_8cpp.xhtml#a4e45c1f5e4280813a78a77dda71d8799">num_elems_processed_per_iteration</a></div><div class="ttdeci">unsigned int num_elems_processed_per_iteration</div><div class="ttdef"><b>Definition:</b> <a href="_c_l_im2_col_kernel_8cpp_source.xhtml#l00054">CLIm2ColKernel.cpp:54</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#l00123">Helpers.inl:123</a></div></div> |
| <div class="ttc" id="_access_window_static_8h_xhtml"><div class="ttname"><a href="_access_window_static_8h.xhtml">AccessWindowStatic.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_a4ee01955c2dd8c14a0d107a36cb62618"><div class="ttname"><a href="namespacearm__compute.xhtml#a4ee01955c2dd8c14a0d107a36cb62618">arm_compute::convert_float32x4x4_to_uint8x16</a></div><div class="ttdeci">void convert_float32x4x4_to_uint8x16(const float32x4x4_t &in, uint8x16_t &out)</div><div class="ttdoc">Converts from two float32x4x4_t to just one uint8x16_t.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_math_8inl_source.xhtml#l00358">NEMath.inl:358</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9"><div class="ttname"><a href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6a329f5d0c4b0c80e3474951d2c4435dd9">arm_compute::DataType::QASYMM8_SIGNED</a></div><div class="ttdoc">quantized, asymmetric fixed-point 8-bit number signed</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="classarm__compute_1_1_n_e_logits1_d_max_kernel_xhtml_a968b23a6ef327fcfb5b99d58e3fbe883"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_max_kernel.xhtml#a968b23a6ef327fcfb5b99d58e3fbe883">arm_compute::NELogits1DMaxKernel::validate</a></div><div class="ttdeci">static Status validate(const ITensorInfo *input, const ITensorInfo *output)</div><div class="ttdoc">Static function to check if given info will lead to a valid configuration of NELogits1DMaxKernel.</div><div class="ttdef"><b>Definition:</b> <a href="_n_e_softmax_layer_kernel_8cpp_source.xhtml#l00220">NESoftmaxLayerKernel.cpp:220</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_ab799fd9cd5fc5989593e69c55bc400f9"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#ab799fd9cd5fc5989593e69c55bc400f9">arm_compute::wrapper::vpmax</a></div><div class="ttdeci">uint8x8_t vpmax(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="pmax_8h_source.xhtml#l00039">pmax.h:39</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1test_1_1validation_xhtml_a1ec8dd7ea96aec0e4068497f0b5e17b2"><div class="ttname"><a href="namespacearm__compute_1_1test_1_1validation.xhtml#a1ec8dd7ea96aec0e4068497f0b5e17b2">arm_compute::test::validation::set</a></div><div class="ttdeci">dst_shape set(0, output_wh.first)</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="_c_l_2_convolution_layer_8cpp_source.xhtml#l00182">ConvolutionLayer.cpp:182</a></div></div> |
| <div class="ttc" id="classarm__compute_1_1_n_e_logits1_d_softmax_kernel_xhtml_a112b35dd205c62ea6ed1447ef226da82"><div class="ttname"><a href="classarm__compute_1_1_n_e_logits1_d_softmax_kernel.xhtml#a112b35dd205c62ea6ed1447ef226da82">arm_compute::NELogits1DSoftmaxKernel::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_softmax_layer_kernel_8cpp_source.xhtml#l00660">NESoftmaxLayerKernel.cpp:660</a></div></div> |
| <div class="ttc" id="namespacearm__compute_xhtml_ad8ed01ff3ff33333d8e19db4d2818bb6"><div class="ttname"><a href="namespacearm__compute.xhtml#ad8ed01ff3ff33333d8e19db4d2818bb6">arm_compute::DataType</a></div><div class="ttdeci">DataType</div><div class="ttdoc">Available data types.</div><div class="ttdef"><b>Definition:</b> <a href="arm__compute_2core_2_types_8h_source.xhtml#l00075">Types.h:75</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="_n_e_fixed_point_8h_xhtml"><div class="ttname"><a href="_n_e_fixed_point_8h.xhtml">NEFixedPoint.h</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a8d1d14000c959f1146a8a64e1dafae5f"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a8d1d14000c959f1146a8a64e1dafae5f">arm_compute::wrapper::vmax</a></div><div class="ttdeci">uint8x8_t vmax(const uint8x8_t &a, const uint8x8_t &b)</div><div class="ttdef"><b>Definition:</b> <a href="max_8h_source.xhtml#l00039">max.h:39</a></div></div> |
| <div class="ttc" id="namespacearm__compute_1_1wrapper_xhtml_a3d35fe714076b40450ba6ba870fae322"><div class="ttname"><a href="namespacearm__compute_1_1wrapper.xhtml#a3d35fe714076b40450ba6ba870fae322">arm_compute::wrapper::vexpq</a></div><div class="ttdeci">float32x4_t vexpq(const float32x4_t &a)</div><div class="ttdef"><b>Definition:</b> <a href="exp_8h_source.xhtml#l00040">exp.h:40</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#l00941">Validate.h:941</a></div></div> |
| <div class="ttc" id="_validate_8h_xhtml"><div class="ttname"><a href="_validate_8h.xhtml">Validate.h</a></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_softmax_layer_kernel_8cpp.xhtml">NESoftmaxLayerKernel.cpp</a></li> |
| <li class="footer">Generated on Thu Mar 5 2020 16:07:02 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> |
