24.02/operator_list.html - ml/armnn - Gitiles

 <!-- HTML header for doxygen 1.8.17-->
 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "https://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 1.8.17"/>
 <meta name="viewport" content="width=device-width, initial-scale=1"/>
 <title>Arm NN: Arm NN Operators</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>
 <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="customdoxygen.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="ArmNN" src="Arm_NN_horizontal_blue.png" style="max-width: 15rem; margin-top: .5rem; margin-left 13px"/>
   <td id="projectalign" style="padding-left: 0.9em;">
    <div id="projectname">
    &#160;<span id="projectnumber">24.02</span>
    </div>
   </td>
  </tr>
  </tbody>
 </table>
 </div>
 <!-- end header part -->
 <!-- Generated by Doxygen 1.8.17 -->
 <script type="text/javascript">
 /* @license magnet:?xt=urn:btih:cf05388f2679ee054f2beb29a391d25f4e673ac3&amp;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&amp;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&amp;dn=gpl-2.0.txt GPL-v2 */
 $(document).ready(function(){initNavTree('operator_list.html',''); initResizable(); });
 /* @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="PageDoc"><div class="header">
   <div class="headertitle">
 <div class="title">Arm NN Operators </div>  </div>
 </div><!--header-->
 <div class="contents">
 <div class="textblock"><h1><a class="anchor" id="S5_1_operator_list"></a>
 Arm NN Operators</h1>
 <p>Arm NN supports operators that are listed in below table.</p>
 <p>Arm NN supports a wide list of data-types. The main data-types that the Machine Learning functions support are the following: </p><ul>
 <li>
 <b>BFLOAT16:</b> 16-bit non-standard brain floating point </li>
 <li>
 <b>QASYMMU8:</b> 8-bit unsigned asymmetric quantized </li>
 <li>
 <b>QASYMMS8:</b> 8-bit signed asymmetric quantized </li>
 <li>
 <b>QUANTIZEDSYMM8PERAXIS:</b> 8-bit signed symmetric quantized </li>
 <li>
 <b>QSYMMS8:</b> 8-bit signed symmetric quantized </li>
 <li>
 <b>QSYMMS16:</b> 16-bit signed symmetric quantized </li>
 <li>
 <b>FLOAT32:</b> 32-bit single precision floating point </li>
 <li>
 <b>FLOAT16:</b> 16-bit half precision floating point </li>
 <li>
 <b>SIGNED32:</b> 32-bit signed integer </li>
 <li>
 <b>BOOLEAN:</b> 8-bit unsigned char </li>
 <li>
 <b>All:</b> Agnostic to any specific data type </li>
 </ul>
 <p>Arm NN supports the following data layouts (fast changing dimension from right to left): </p><ul>
 <li>
 <b>NHWC:</b> Layout where channels are in the fastest changing dimension </li>
 <li>
 <b>NCHW:</b> Layout where width is in the fastest changing dimension </li>
 <li>
 <b>All:</b> Agnostic to any specific data layout </li>
 </ul>
 <p>where N = batches, C = channels, H = height, W = width</p>
 <a class="anchor" id="multi_row"></a>
 <table class="doxtable">
 <caption></caption>
 <tr>
 <th>Operator </th><th>Description </th><th>Equivalent Android NNAPI Operator </th><th>Backends </th><th>Data Layouts </th><th>Data Types </th></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_abs_layer.html">AbsLayer</a> </td><td rowspan="3"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform absolute operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ABS </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_activation_layer.html" title="This layer represents an activation operation with the specified activation function.">ActivationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to simulate an activation layer with the specified activation function. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ABS </li>
 <li>
 ANEURALNETWORKS_ELU </li>
 <li>
 ANEURALNETWORKS_HARD_SWISH </li>
 <li>
 ANEURALNETWORKS_LOGISTIC </li>
 <li>
 ANEURALNETWORKS_PRELU </li>
 <li>
 ANEURALNETWORKS_RELU </li>
 <li>
 ANEURALNETWORKS_RELU1 </li>
 <li>
 ANEURALNETWORKS_RELU6 </li>
 <li>
 ANEURALNETWORKS_SQRT </li>
 <li>
 ANEURALNETWORKS_TANH </li>
 <li>
 GELU </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_addition_layer.html" title="This layer represents an addition operation.">AdditionLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to add 2 tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ADD </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_arg_min_max_layer.html" title="This layer represents a ArgMinMax operation.">ArgMinMaxLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to calculate the index of the minimum or maximum values in a tensor based on an axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ARGMAX </li>
 <li>
 ANEURALNETWORKS_ARGMIN </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>SIGNED64 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_batch_mat_mul_layer.html">BatchMatMulLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform batch matrix multiplication. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_batch_normalization_layer.html" title="This layer represents a batch normalization operation.">BatchNormalizationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform batch normalization. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_batch_to_space_nd_layer.html" title="This layer represents a BatchToSpaceNd operation.">BatchToSpaceNdLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform a batch to space transformation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_BATCH_TO_SPACE_ND </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_broadcast_to_layer.html">BroadcastToLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to broadcast a tensor to a given size. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>N/A </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>N/A </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>N/A </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_cast_layer.html" title="This layer represents a cast operation.">CastLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to cast a tensor to a type. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_CAST </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>SIGNED64 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>SIGNED64 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>SIGNED64 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_channel_shuffle_layer.html">ChannelShuffleLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to reorganize the channels of a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_CHANNEL_SHUFFLE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_comparison_layer.html" title="This layer represents a comparison operation.">ComparisonLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compare 2 tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_EQUAL </li>
 <li>
 ANEURALNETWORKS_GREATER </li>
 <li>
 ANEURALNETWORKS_GREATER_EQUAL </li>
 <li>
 ANEURALNETWORKS_LESS </li>
 <li>
 ANEURALNETWORKS_LESS_EQUAL </li>
 <li>
 ANEURALNETWORKS_NOT_EQUAL </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_concat_layer.html" title="This layer represents a merge operation.">ConcatLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to concatenate tensors along a given axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_CONCATENATION </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_constant_layer.html" title="A layer that the constant data can be bound to.">ConstantLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to provide a constant tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_convert_fp16_to_fp32_layer.html" title="This layer converts data type Float 16 to Float 32.">ConvertFp16ToFp32Layer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to convert Float16 tensor to Float32 tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_convert_fp32_to_fp16_layer.html" title="This layer converts data type Float 32 to Float 16.">ConvertFp32ToFp16Layer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to convert Float32 tensor to Float16 tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_convolution2d_layer.html" title="This layer represents a convolution 2d operation.">Convolution2dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compute a convolution operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_CONV_2D </li>
 <li>
 ANEURALNETWORKS_GROUPED_CONV_2D </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_convolution3d_layer.html" title="This layer represents a convolution 3d operation.">Convolution3dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compute a 3D convolution operation. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 NDHWC </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td><td><ul>
 <li>
 N/A </li>
 </ul>
 </td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_debug_layer.html" title="This layer visualizes the data flowing through the network.">DebugLayer</a> </td><td rowspan="1" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to print out inter layer tensor information. </td><td rowspan="1"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_depth_to_space_layer.html" title="This layer represents a DepthToSpace operation.">DepthToSpaceLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Depth to Space transformation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_DEPTH_TO_SPACE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_depthwise_convolution2d_layer.html" title="This layer represents a depthwise convolution 2d operation.">DepthwiseConvolution2dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compute a deconvolution or transpose convolution. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_DEPTHWISE_CONV_2D </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_dequantize_layer.html" title="This layer dequantizes the input tensor.">DequantizeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to dequantize the values in a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_DEQUANTIZE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="2"><a class="el" href="classarmnn_1_1_detection_post_process_layer.html" title="This layer represents a detection postprocess operator.">DetectionPostProcessLayer</a> </td><td rowspan="2" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to generate the detection output based on center size encoded boxes, class prediction and anchors by doing non maximum suppression (NMS). </td><td rowspan="2"><ul>
 <li>
 ANEURALNETWORKS_DETECTION_POSTPROCESSING </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_division_layer.html" title="This layer represents a division operation.">DivisionLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to divide 2 tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_DIV </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_elementwise_base_layer.html" title="NOTE: this is an abstract class to encapsulate the element wise operations, it does not implement: st...">ElementwiseBaseLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Add - Div - Max - Min - Mul operations. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ADD </li>
 <li>
 ANEURALNETWORKS_DIV </li>
 <li>
 ANEURALNETWORKS_MAXIMUM </li>
 <li>
 ANEURALNETWORKS_MINIMUM </li>
 <li>
 ANEURALNETWORKS_MUL </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_elementwise_binary_layer.html" title="This layer represents a elementwiseBinary operation.">ElementwiseBinaryLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Power and Square Difference operations. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_POW </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_elementwise_unary_layer.html" title="This layer represents a elementwiseUnary operation.">ElementwiseUnaryLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Rsqrt - Exp - Neg - Log - Abs - Sin - Sqrt - Ceil operations. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_ABS </li>
 <li>
 ANEURALNETWORKS_EXP </li>
 <li>
 ANEURALNETWORKS_LOG </li>
 <li>
 ANEURALNETWORKS_NEG </li>
 <li>
 ANEURALNETWORKS_RSQRT </li>
 <li>
 ANEURALNETWORKS_SIN </li>
 <li>
 ANEURALNETWORKS_SQRT </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_fake_quantization_layer.html" title="This layer represents a fake quantization operation.">FakeQuantizationLayer</a> </td><td rowspan="1" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to quantize float values and dequantize afterwards. The current implementation does not dequantize the values. </td><td rowspan="1"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_fill_layer.html" title="This layer represents a fill operation.">FillLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to set the values of a tensor with a given value. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_FILL </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_floor_layer.html" title="This layer represents a floor operation.">FloorLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to round the value to the lowest whole number. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_FLOOR </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_fully_connected_layer.html" title="This layer represents a fully connected operation.">FullyConnectedLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform a fully connected / dense operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_FULLY_CONNECTED </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_gather_layer.html" title="This layer represents a Gather operator.">GatherLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform the gather operation along the chosen axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_GATHER </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_gather_nd_layer.html" title="This layer represents a GatherNd operator.">GatherNdLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform the gatherNd operation. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_input_layer.html" title="A layer user-provided data can be bound to (e.g. inputs, outputs).">InputLayer</a> </td><td rowspan="1" style="width:200px;">Special layer used to provide input data to the computational network. </td><td rowspan="1"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>All </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_instance_normalization_layer.html" title="This layer represents an instance normalization operation.">InstanceNormalizationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an instance normalization on a given axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_INSTANCE_NORMALIZATION </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_l2_normalization_layer.html" title="This layer represents a L2 normalization operation.">L2NormalizationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an L2 normalization on a given axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_L2_NORMALIZATION </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_log_softmax_layer.html" title="This layer represents a log softmax operation.">LogSoftmaxLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform the log softmax activations given logits. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_logical_binary_layer.html" title="This layer represents a Logical Binary operation.">LogicalBinaryLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Logical AND - Logical NOT - Logical OR operations. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_LOGICAL_AND </li>
 <li>
 ANEURALNETWORKS_LOGICAL_NOT </li>
 <li>
 ANEURALNETWORKS_LOGICAL_OR </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_lstm_layer.html" title="This layer represents a LSTM operation.">LstmLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform a single time step in a Long Short-Term Memory (LSTM) operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_LSTM </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_map_layer.html" title="This layer represents a memory copy operation.">MapLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform map operation on tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_maximum_layer.html" title="This layer represents a maximum operation.">MaximumLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an elementwise maximum of two tensors. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_mean_layer.html" title="This layer represents a mean operation.">MeanLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform reduce mean operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_MEAN </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_mem_copy_layer.html" title="This layer represents a memory copy operation.">MemCopyLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform memory copy operation. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_mem_import_layer.html" title="This layer represents a memory import operation.">MemImportLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform memory import operation. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_merge_layer.html" title="This layer dequantizes the input tensor.">MergeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to concatenate tensors along a given axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_CONCATENATION </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_minimum_layer.html" title="This layer represents a minimum operation.">MinimumLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an elementwise minimum of two tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_MINIMUM </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_multiplication_layer.html" title="This layer represents a multiplication operation.">MultiplicationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an elementwise multiplication of two tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_MUL </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_normalization_layer.html" title="This layer represents a normalization operation.">NormalizationLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compute normalization operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_LOCAL_RESPONSE_NORMALIZATION </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_output_layer.html" title="A layer user-provided data can be bound to (e.g. inputs, outputs).">OutputLayer</a> </td><td rowspan="1" style="width:200px;">A special layer providing access to a user supplied buffer into which the output of a network can be written. </td><td rowspan="1"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>All </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_pad_layer.html" title="This layer represents a pad operation.">PadLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to pad a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_PAD </li>
 <li>
 ANEURALNETWORKS_PAD_V2 </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_permute_layer.html" title="This layer represents a permutation operation.">PermuteLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to transpose an ND tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_TRANSPOSE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_pooling2d_layer.html" title="This layer represents a pooling 2d operation.">Pooling2dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform 2D pooling with the specified pooling operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_AVERAGE_POOL_2D </li>
 <li>
 ANEURALNETWORKS_L2_POOL_2D </li>
 <li>
 ANEURALNETWORKS_MAX_POOL_2D </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_pooling3d_layer.html" title="This layer represents a pooling 3d operation.">Pooling3dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform 3D pooling with the specified pooling operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_AVERAGE_POOL_3D </li>
 <li>
 ANEURALNETWORKS_L2_POOL_3D </li>
 <li>
 ANEURALNETWORKS_MAX_POOL_3D </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 NDHWC </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NA </li>
 </ul>
 </td><td></td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NDHWC </li>
 </ul>
 </td><td></td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_pre_compiled_layer.html">PreCompiledLayer</a> </td><td rowspan="1" style="width:200px;">Opaque layer provided by a backend which provides an executable representation of a subgraph from the original network. </td><td rowspan="1"><ul>
 <li>
 </li>
 </ul>
 <br  />
  N/A  </td><td>N/A </td><td>N/A </td><td>N/A </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_prelu_layer.html">PreluLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to compute the activation layer with the PRELU activation function. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_PRELU </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_q_lstm_layer.html" title="This layer represents a QLstm operation.">QLstmLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform quantized LSTM (Long Short-Term Memory) operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_QUANTIZED_LSTM </li>
 <li>
 ANEURALNETWORKS_QUANTIZED_16BIT_LSTM </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_quantize_layer.html">QuantizeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform quantization operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_QUANTIZE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMM16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMM16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_quantized_lstm_layer.html" title="This layer represents a QuantizedLstm operation.">QuantizedLstmLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform quantized LSTM (Long Short-Term Memory) operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_QUANTIZED_LSTM </li>
 <li>
 ANEURALNETWORKS_QUANTIZED_16BIT_LSTM </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_rank_layer.html">RankLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform a rank operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_RANK </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_reduce_layer.html" title="This layer represents a reduction operation.">ReduceLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform reduce with the following operations - ARG_IDX_MAX: Index of the max value - ARG_IDX_MIN: Index of the min value - MEAN_SUM: Mean of sum - PROD: Product - SUM_SQUARE: Sum of squares - SUM: Sum - MIN: Min - MAX: Max </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_REDUCE_MAX </li>
 <li>
 ANEURALNETWORKS_REDUCE_MIN </li>
 <li>
 ANEURALNETWORKS_REDUCE_SUM </li>
 <li>
 ANEURALNETWORKS_REDUCE_PROD </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_reshape_layer.html" title="This layer represents a reshape operation.">ReshapeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to reshape a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_RESHAPE </li>
 <li>
 ANEURALNETWORKS_SQUEEZE </li>
 <li>
 ANEURALNETWORKS_EXPAND_DIMS </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>BOOLEAN </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_resize_layer.html" title="This layer represents a resize operation.">ResizeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform resize of a tensor using one of the interpolation methods: - Bilinear - Nearest Neighbor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_RESIZE_BILINEAR </li>
 <li>
 ANEURALNETWORKS_RESIZE_NEAREST_NEIGHBOR </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_reverse_v2_layer.html" title="This layer represents a ReverseV2 operation.">ReverseV2Layer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform reverse of a tensor. </td><td rowspan="3"><ul>
 <li>
 NA </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_rsqrt_layer.html">RsqrtLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform Rsqrt operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_RSQRT </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_shape_layer.html">ShapeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to return the shape of the input tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_slice_layer.html">SliceLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform tensor slicing. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_SLICE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_softmax_layer.html" title="This layer represents a softmax operation.">SoftmaxLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform softmax, log-softmax operation over the specified axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_LOG_SOFTMAX </li>
 <li>
 ANEURALNETWORKS_SOFTMAX </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_space_to_batch_nd_layer.html" title="This layer represents a SpaceToBatchNd operation.">SpaceToBatchNdLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to divide spatial dimensions of the tensor into a grid of blocks and interleaves these blocks with the batch dimension. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_SPACE_TO_BATCH_ND </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_space_to_depth_layer.html" title="This layer represents a SpaceToDepth operation.">SpaceToDepthLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to rearrange blocks of spatial data into depth. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_SPACE_TO_DEPTH </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_splitter_layer.html" title="This layer represents a split operation.">SplitterLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to split a tensor along a given axis. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_SPLIT </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_stack_layer.html" title="This layer represents a stack operation.">StackLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to stack tensors along an axis. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="1"><a class="el" href="classarmnn_1_1_stand_in_layer.html" title="This layer represents an unknown operation in the input graph.">StandInLayer</a> </td><td rowspan="1" style="width:200px;">A layer to represent "unknown" or "unsupported" operations in the input graph. It has a configurable number of input and output slots and an optional name. </td><td rowspan="1"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>N/A </td><td>N/A </td><td>N/A </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_strided_slice_layer.html" title="This layer represents a strided slice operation.">StridedSliceLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to extract a strided slice of a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_STRIDED_SLICE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_subtraction_layer.html" title="This layer represents a subtraction operation.">SubtractionLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform an elementwise subtract of 2 tensors. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_SUB </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_tile_layer.html">TileLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to construct a tensor by repeating in tiles a given tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_TILE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMM8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_transpose_convolution2d_layer.html" title="This layer represents a 2D transpose convolution operation.">TransposeConvolution2dLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform 2D transpose convolution (deconvolution) operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_TRANSPOSE_CONV_2D </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>SIGNED32 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QUANTIZEDSYMM8PERAXIS </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_transpose_layer.html" title="This layer represents a transpose operation.">TransposeLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to transpose a tensor. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_TRANSPOSE </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>BFLOAT16 </td></tr>
 <tr>
 <td>FLOAT16 </td></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 <tr>
 <td>QASYMMU8 </td></tr>
 <tr>
 <td>QSYMMS16 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3">UnidirectionalSquenceLstmLayer </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform unidirectional sequence LSTM operation. </td><td rowspan="3"><ul>
 <li>
 ANEURALNETWORKS_UNIDIRECTIONAL_SEQUENCE_LSTM </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th>Input Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 <table class="doxtable">
 <tr>
 <th>Weight Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 <tr>
 <td>QASYMMS8 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th>Input Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 <table class="doxtable">
 <tr>
 <th>Weight Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th>Input Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 <table class="doxtable">
 <tr>
 <th>Weight Types </th></tr>
 <tr>
 <td>FLOAT32 </td></tr>
 </table>
 </td></tr>
 <tr>
 <td rowspan="3"><a class="el" href="classarmnn_1_1_unmap_layer.html" title="This layer represents a memory copy operation.">UnmapLayer</a> </td><td rowspan="3" style="width:200px;"><a class="el" href="classarmnn_1_1_layer.html">Layer</a> to perform unmap operation on tensor. </td><td rowspan="3"><ul>
 <li>
 N/A </li>
 </ul>
 </td><td>CpuRef </td><td><ul>
 <li>
 All </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>CpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 <tr>
 <td>GpuAcc </td><td><ul>
 <li>
 NHWC </li>
 <li>
 NCHW </li>
 </ul>
 </td><td><table class="doxtable">
 <tr>
 <th></th></tr>
 <tr>
 <td>All </td></tr>
 </table>
 </td></tr>
 </table>
 </div></div><!-- contents -->
 </div><!-- PageDoc -->
 </div><!-- doc-content -->
 <!-- start footer part -->
 <div id="nav-path" class="navpath"><!-- id is needed for treeview function! -->
   <ul>
     <li class="footer">Generated on Wed Feb 14 2024 16:36:20 for Arm NN by
     <a href="http://www.doxygen.org/index.html">
     <img class="footer" src="doxygen.png" alt="doxygen"/></a> 1.8.17 </li>
   </ul>
 </div>
 </body>
 </html>