Google Git
Sign in
apache / mxnet-site / 5d2e45b9c41410156d7d50982069a8e0f3991694 / . / versions / 1.5.0 / _modules / mxnet / module / base_module.html
blob: d9e04a625d427748f8fd6bcb5cf10087143ec9f9 [file] [log] [blame]
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="utf-8"/>
<meta content="IE=edge" http-equiv="X-UA-Compatible"/>
<meta content="width=device-width, initial-scale=1" name="viewport"/>
<meta content="mxnet.module.base_module" property="og:title">
<meta content="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/og-logo.png" property="og:image">
<meta content="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/og-logo.png" property="og:image:secure_url">
<meta content="mxnet.module.base_module" property="og:description"/>
<title>mxnet.module.base_module — mxnet documentation</title>
<link crossorigin="anonymous" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/css/bootstrap.min.css" integrity="sha384-1q8mTJOASx8j1Au+a5WDVnPi2lkFfwwEAa8hDDdjZlpLegxhjVME1fgjWPGmkzs7" rel="stylesheet"/>
<link href="https://maxcdn.bootstrapcdn.com/font-awesome/4.5.0/css/font-awesome.min.css" rel="stylesheet"/>
<link href="../../../_static/basic.css" rel="stylesheet" type="text/css">
<link href="../../../_static/pygments.css" rel="stylesheet" type="text/css">
<link href="../../../_static/mxnet.css" rel="stylesheet" type="text/css"/>
<script type="text/javascript">
var DOCUMENTATION_OPTIONS = {
URL_ROOT: '../../../',
VERSION: '',
COLLAPSE_INDEX: false,
FILE_SUFFIX: '.html',
HAS_SOURCE: true,
SOURCELINK_SUFFIX: '.txt'
};
</script>
<script src="https://code.jquery.com/jquery-1.11.1.min.js" type="text/javascript"></script>
<script src="../../../_static/underscore.js" type="text/javascript"></script>
<script src="../../../_static/searchtools_custom.js" type="text/javascript"></script>
<script src="../../../_static/doctools.js" type="text/javascript"></script>
<script src="../../../_static/selectlang.js" type="text/javascript"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML" type="text/javascript"></script>
<script type="text/javascript"> jQuery(function() { Search.loadIndex("/versions/1.5.0/searchindex.js"); Search.init();}); </script>
<script>
(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new
Date();a=s.createElement(o),
m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
})(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
ga('create', 'UA-96378503-1', 'auto');
ga('send', 'pageview');
</script>
<!-- -->
<!-- <script type="text/javascript" src="../../../_static/jquery.js"></script> -->
<!-- -->
<!-- <script type="text/javascript" src="../../../_static/underscore.js"></script> -->
<!-- -->
<!-- <script type="text/javascript" src="../../../_static/doctools.js"></script> -->
<!-- -->
<!-- <script type="text/javascript" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.0/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script> -->
<!-- -->
<link href="../../../genindex.html" rel="index" title="Index">
<link href="../../../search.html" rel="search" title="Search"/>
<link href="../../index.html" rel="up" title="Module code"/>
<link href="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet-icon.png" rel="icon" type="image/png"/>
</link></link></link></meta></meta></meta></head>
<body background="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet-background-compressed.jpeg" role="document">
<div class="content-block"><div class="navbar navbar-fixed-top">
<div class="container" id="navContainer">
<div class="innder" id="header-inner">
<h1 id="logo-wrap">
<a href="../../../" id="logo"><img src="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/mxnet_logo.png"/></a>
</h1>
<nav class="nav-bar" id="main-nav">
<a class="main-nav-link" href="/versions/1.5.0/install/index.html">Install</a>
<span id="dropdown-menu-position-anchor">
<a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Gluon <span class="caret"></span></a>
<ul class="dropdown-menu navbar-menu" id="package-dropdown-menu">
<li><a class="main-nav-link" href="/versions/1.5.0/tutorials/gluon/gluon.html">About</a></li>
<li><a class="main-nav-link" href="https://www.d2l.ai/">Dive into Deep Learning</a></li>
<li><a class="main-nav-link" href="https://gluon-cv.mxnet.io">GluonCV Toolkit</a></li>
<li><a class="main-nav-link" href="https://gluon-nlp.mxnet.io/">GluonNLP Toolkit</a></li>
</ul>
</span>
<span id="dropdown-menu-position-anchor">
<a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">API <span class="caret"></span></a>
<ul class="dropdown-menu navbar-menu" id="package-dropdown-menu">
<li><a class="main-nav-link" href="/versions/1.5.0/api/python/index.html">Python</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/c++/index.html">C++</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/clojure/index.html">Clojure</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/java/index.html">Java</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/julia/index.html">Julia</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/perl/index.html">Perl</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/r/index.html">R</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/scala/index.html">Scala</a></li>
</ul>
</span>
<span id="dropdown-menu-position-anchor-docs">
<a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Docs <span class="caret"></span></a>
<ul class="dropdown-menu navbar-menu" id="package-dropdown-menu-docs">
<li><a class="main-nav-link" href="/versions/1.5.0/faq/index.html">FAQ</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/tutorials/index.html">Tutorials</a>
<li><a class="main-nav-link" href="https://github.com/apache/incubator-mxnet/tree/1.5.0/example">Examples</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/architecture/index.html">Architecture</a></li>
<li><a class="main-nav-link" href="https://cwiki.apache.org/confluence/display/MXNET/Apache+MXNet+Home">Developer Wiki</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/model_zoo/index.html">Model Zoo</a></li>
<li><a class="main-nav-link" href="https://github.com/onnx/onnx-mxnet">ONNX</a></li>
</li></ul>
</span>
<span id="dropdown-menu-position-anchor-community">
<a aria-expanded="true" aria-haspopup="true" class="main-nav-link dropdown-toggle" data-toggle="dropdown" href="#" role="button">Community <span class="caret"></span></a>
<ul class="dropdown-menu navbar-menu" id="package-dropdown-menu-community">
<li><a class="main-nav-link" href="http://discuss.mxnet.io">Forum</a></li>
<li><a class="main-nav-link" href="https://github.com/apache/incubator-mxnet/tree/1.5.0">Github</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/community/contribute.html">Contribute</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/community/ecosystem.html">Ecosystem</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/community/powered_by.html">Powered By</a></li>
</ul>
</span>
<span id="dropdown-menu-position-anchor-version" style="position: relative"><a href="#" class="main-nav-link dropdown-toggle" data-toggle="dropdown" role="button" aria-haspopup="true" aria-expanded="true">1.5.0<span class="caret"></span></a><ul id="package-dropdown-menu" class="dropdown-menu"><li><a href="/">master</a></li><li><a href="/versions/1.7.0/">1.7.0</a></li><li><a href=/versions/1.6.0/>1.6.0</a></li><li><a href=/versions/1.5.0/>1.5.0</a></li><li><a href=/versions/1.4.1/>1.4.1</a></li><li><a href=/versions/1.3.1/>1.3.1</a></li><li><a href=/versions/1.2.1/>1.2.1</a></li><li><a href=/versions/1.1.0/>1.1.0</a></li><li><a href=/versions/1.0.0/>1.0.0</a></li><li><a href=/versions/0.12.1/>0.12.1</a></li><li><a href=/versions/0.11.0/>0.11.0</a></li></ul></span></nav>
<script> function getRootPath(){ return "../../../" } </script>
<div class="burgerIcon dropdown">
<a class="dropdown-toggle" data-toggle="dropdown" href="#" role="button">☰</a>
<ul class="dropdown-menu" id="burgerMenu">
<li><a href="/versions/1.5.0/install/index.html">Install</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/tutorials/index.html">Tutorials</a></li>
<li class="dropdown-submenu dropdown">
<a aria-expanded="true" aria-haspopup="true" class="dropdown-toggle burger-link" data-toggle="dropdown" href="#" tabindex="-1">Gluon</a>
<ul class="dropdown-menu navbar-menu" id="package-dropdown-menu">
<li><a class="main-nav-link" href="/versions/1.5.0/tutorials/gluon/gluon.html">About</a></li>
<li><a class="main-nav-link" href="http://gluon.mxnet.io">The Straight Dope (Tutorials)</a></li>
<li><a class="main-nav-link" href="https://gluon-cv.mxnet.io">GluonCV Toolkit</a></li>
<li><a class="main-nav-link" href="https://gluon-nlp.mxnet.io/">GluonNLP Toolkit</a></li>
</ul>
</li>
<li class="dropdown-submenu">
<a aria-expanded="true" aria-haspopup="true" class="dropdown-toggle burger-link" data-toggle="dropdown" href="#" tabindex="-1">API</a>
<ul class="dropdown-menu">
<li><a class="main-nav-link" href="/versions/1.5.0/api/python/index.html">Python</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/c++/index.html">C++</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/clojure/index.html">Clojure</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/java/index.html">Java</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/julia/index.html">Julia</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/perl/index.html">Perl</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/r/index.html">R</a></li>
<li><a class="main-nav-link" href="/versions/1.5.0/api/scala/index.html">Scala</a></li>
</ul>
</li>
<li class="dropdown-submenu">
<a aria-expanded="true" aria-haspopup="true" class="dropdown-toggle burger-link" data-toggle="dropdown" href="#" tabindex="-1">Docs</a>
<ul class="dropdown-menu">
<li><a href="/versions/1.5.0/faq/index.html" tabindex="-1">FAQ</a></li>
<li><a href="/versions/1.5.0/tutorials/index.html" tabindex="-1">Tutorials</a></li>
<li><a href="https://github.com/apache/incubator-mxnet/tree/1.5.0/example" tabindex="-1">Examples</a></li>
<li><a href="/versions/1.5.0/architecture/index.html" tabindex="-1">Architecture</a></li>
<li><a href="https://cwiki.apache.org/confluence/display/MXNET/Apache+MXNet+Home" tabindex="-1">Developer Wiki</a></li>
<li><a href="/versions/1.5.0/model_zoo/index.html" tabindex="-1">Gluon Model Zoo</a></li>
<li><a href="https://github.com/onnx/onnx-mxnet" tabindex="-1">ONNX</a></li>
</ul>
</li>
<li class="dropdown-submenu dropdown">
<a aria-haspopup="true" class="dropdown-toggle burger-link" data-toggle="dropdown" href="#" role="button" tabindex="-1">Community</a>
<ul class="dropdown-menu">
<li><a href="http://discuss.mxnet.io" tabindex="-1">Forum</a></li>
<li><a href="https://github.com/apache/incubator-mxnet/tree/1.5.0" tabindex="-1">Github</a></li>
<li><a href="/versions/1.5.0/community/contribute.html" tabindex="-1">Contribute</a></li>
<li><a href="/versions/1.5.0/community/ecosystem.html" tabindex="-1">Ecosystem</a></li>
<li><a href="/versions/1.5.0/community/powered_by.html" tabindex="-1">Powered By</a></li>
</ul>
</li>
<li id="dropdown-menu-position-anchor-version-mobile" class="dropdown-submenu" style="position: relative"><a href="#" tabindex="-1">1.5.0</a><ul class="dropdown-menu"><li><a tabindex="-1" href=/>master</a></li><li><a tabindex="-1" href=/versions/1.6.0/>1.6.0</a></li><li><a tabindex="-1" href=/versions/1.5.0/>1.5.0</a></li><li><a tabindex="-1" href=/versions/1.4.1/>1.4.1</a></li><li><a tabindex="-1" href=/versions/1.3.1/>1.3.1</a></li><li><a tabindex="-1" href=/versions/1.2.1/>1.2.1</a></li><li><a tabindex="-1" href=/versions/1.1.0/>1.1.0</a></li><li><a tabindex="-1" href=/versions/1.0.0/>1.0.0</a></li><li><a tabindex="-1" href=/versions/0.12.1/>0.12.1</a></li><li><a tabindex="-1" href=/versions/0.11.0/>0.11.0</a></li></ul></li></ul>
</div>
<div class="plusIcon dropdown">
<a class="dropdown-toggle" data-toggle="dropdown" href="#" role="button"><span aria-hidden="true" class="glyphicon glyphicon-plus"></span></a>
<ul class="dropdown-menu dropdown-menu-right" id="plusMenu"></ul>
</div>
<div id="search-input-wrap">
<form action="../../../search.html" autocomplete="off" class="" method="get" role="search">
<div class="form-group inner-addon left-addon">
<i class="glyphicon glyphicon-search"></i>
<input class="form-control" name="q" placeholder="Search" type="text"/>
</div>
<input name="check_keywords" type="hidden" value="yes">
<input name="area" type="hidden" value="default"/>
</input></form>
<div id="search-preview"></div>
</div>
<div id="searchIcon">
<span aria-hidden="true" class="glyphicon glyphicon-search"></span>
</div>
<!-- <div id="lang-select-wrap"> -->
<!-- <label id="lang-select-label"> -->
<!-- <\!-- <i class="fa fa-globe"></i> -\-> -->
<!-- <span></span> -->
<!-- </label> -->
<!-- <select id="lang-select"> -->
<!-- <option value="en">Eng</option> -->
<!-- <option value="zh">中文</option> -->
<!-- </select> -->
<!-- </div> -->
<!-- <a id="mobile-nav-toggle">
<span class="mobile-nav-toggle-bar"></span>
<span class="mobile-nav-toggle-bar"></span>
<span class="mobile-nav-toggle-bar"></span>
</a> -->
</div>
</div>
</div>
<script type="text/javascript">
$('body').css('background', 'white');
</script>
<div class="container">
<div class="row">
<div aria-label="main navigation" class="sphinxsidebar leftsidebar" role="navigation">
<div class="sphinxsidebarwrapper">
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../../api/index.html">MXNet APIs</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../architecture/index.html">MXNet Architecture</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../community/index.html">MXNet Community</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../faq/index.html">MXNet FAQ</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../gluon/index.html">About Gluon</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../install/index.html">Installing MXNet</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../install/index.html#nvidia-jetson-tx-family">Nvidia Jetson TX family</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../install/index.html#source-download">Source Download</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../model_zoo/index.html">MXNet Model Zoo</a></li>
<li class="toctree-l1"><a class="reference internal" href="../../../tutorials/index.html">Tutorials</a></li>
</ul>
</div>
</div>
<div class="content">
<div class="page-tracker"></div>
<h1>Source code for mxnet.module.base_module</h1><div class="highlight"><pre>
<span></span><span class="c1"># Licensed to the Apache Software Foundation (ASF) under one</span>
<span class="c1"># or more contributor license agreements. See the NOTICE file</span>
<span class="c1"># distributed with this work for additional information</span>
<span class="c1"># regarding copyright ownership. The ASF licenses this file</span>
<span class="c1"># to you under the Apache License, Version 2.0 (the</span>
<span class="c1"># "License"); you may not use this file except in compliance</span>
<span class="c1"># with the License. You may obtain a copy of the License at</span>
<span class="c1">#</span>
<span class="c1"># http://www.apache.org/licenses/LICENSE-2.0</span>
<span class="c1">#</span>
<span class="c1"># Unless required by applicable law or agreed to in writing,</span>
<span class="c1"># software distributed under the License is distributed on an</span>
<span class="c1"># "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY</span>
<span class="c1"># KIND, either express or implied. See the License for the</span>
<span class="c1"># specific language governing permissions and limitations</span>
<span class="c1"># under the License.</span>
<span class="c1"># pylint: disable=fixme, too-many-arguments, too-many-locals, no-else-raise</span>
<span class="c1"># pylint: disable=too-many-public-methods, too-many-branches, too-many-lines</span>
<span class="sd">"""`BaseModule` defines an API for modules."""</span>
<span class="kn">import</span> <span class="nn">time</span>
<span class="kn">import</span> <span class="nn">logging</span>
<span class="kn">import</span> <span class="nn">warnings</span>
<span class="kn">import</span> <span class="nn">numpy</span> <span class="k">as</span> <span class="nn">np</span>
<span class="kn">from</span> <span class="nn">..</span> <span class="k">import</span> <span class="n">metric</span>
<span class="kn">from</span> <span class="nn">..</span> <span class="k">import</span> <span class="n">ndarray</span>
<span class="kn">from</span> <span class="nn">..context</span> <span class="k">import</span> <span class="n">cpu</span>
<span class="kn">from</span> <span class="nn">..model</span> <span class="k">import</span> <span class="n">BatchEndParam</span>
<span class="kn">from</span> <span class="nn">..initializer</span> <span class="k">import</span> <span class="n">Uniform</span>
<span class="kn">from</span> <span class="nn">..io</span> <span class="k">import</span> <span class="n">DataDesc</span><span class="p">,</span> <span class="n">DataIter</span><span class="p">,</span> <span class="n">DataBatch</span>
<span class="kn">from</span> <span class="nn">..base</span> <span class="k">import</span> <span class="n">_as_list</span>
<span class="k">def</span> <span class="nf">_check_input_names</span><span class="p">(</span><span class="n">symbol</span><span class="p">,</span> <span class="n">names</span><span class="p">,</span> <span class="n">typename</span><span class="p">,</span> <span class="n">throw</span><span class="p">):</span>
<span class="sd">"""Check that all input names are in symbol's arguments."""</span>
<span class="n">args</span> <span class="o">=</span> <span class="n">symbol</span><span class="o">.</span><span class="n">list_arguments</span><span class="p">()</span>
<span class="k">for</span> <span class="n">name</span> <span class="ow">in</span> <span class="n">names</span><span class="p">:</span>
<span class="k">if</span> <span class="n">name</span> <span class="ow">in</span> <span class="n">args</span><span class="p">:</span>
<span class="k">continue</span>
<span class="n">candidates</span> <span class="o">=</span> <span class="p">[</span><span class="n">arg</span> <span class="k">for</span> <span class="n">arg</span> <span class="ow">in</span> <span class="n">args</span> <span class="k">if</span>
<span class="ow">not</span> <span class="n">arg</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'_weight'</span><span class="p">)</span> <span class="ow">and</span>
<span class="ow">not</span> <span class="n">arg</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'_bias'</span><span class="p">)</span> <span class="ow">and</span>
<span class="ow">not</span> <span class="n">arg</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'_gamma'</span><span class="p">)</span> <span class="ow">and</span>
<span class="ow">not</span> <span class="n">arg</span><span class="o">.</span><span class="n">endswith</span><span class="p">(</span><span class="s1">'_beta'</span><span class="p">)]</span>
<span class="n">msg</span> <span class="o">=</span> <span class="s2">"</span><span class="se">\033</span><span class="s2">[91mYou created Module with Module(..., </span><span class="si">%s</span><span class="s2">_names=</span><span class="si">%s</span><span class="s2">) but "</span> \
<span class="s2">"input with name '</span><span class="si">%s</span><span class="s2">' is not found in symbol.list_arguments(). "</span> \
<span class="s2">"Did you mean one of:</span><span class="se">\n\t</span><span class="si">%s</span><span class="se">\033</span><span class="s2">[0m"</span><span class="o">%</span><span class="p">(</span>
<span class="n">typename</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="n">names</span><span class="p">),</span> <span class="n">name</span><span class="p">,</span> <span class="s1">'</span><span class="se">\n\t</span><span class="s1">'</span><span class="o">.</span><span class="n">join</span><span class="p">(</span><span class="n">candidates</span><span class="p">))</span>
<span class="k">if</span> <span class="n">throw</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="n">msg</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="n">msg</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">_check_names_match</span><span class="p">(</span><span class="n">data_names</span><span class="p">,</span> <span class="n">data_shapes</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="n">throw</span><span class="p">):</span>
<span class="sd">"""Check that input names matches input data descriptors."""</span>
<span class="n">actual</span> <span class="o">=</span> <span class="p">[</span><span class="n">x</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">data_shapes</span><span class="p">]</span>
<span class="k">if</span> <span class="nb">sorted</span><span class="p">(</span><span class="n">data_names</span><span class="p">)</span> <span class="o">!=</span> <span class="nb">sorted</span><span class="p">(</span><span class="n">actual</span><span class="p">):</span>
<span class="n">msg</span> <span class="o">=</span> <span class="s2">"Data provided by </span><span class="si">%s</span><span class="s2">_shapes don't match names specified by </span><span class="si">%s</span><span class="s2">_names (</span><span class="si">%s</span><span class="s2"> vs. </span><span class="si">%s</span><span class="s2">)"</span><span class="o">%</span><span class="p">(</span>
<span class="n">name</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="nb">str</span><span class="p">(</span><span class="n">data_shapes</span><span class="p">),</span> <span class="nb">str</span><span class="p">(</span><span class="n">data_names</span><span class="p">))</span>
<span class="k">if</span> <span class="n">throw</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="n">msg</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="n">msg</span><span class="p">)</span>
<span class="k">def</span> <span class="nf">_parse_data_desc</span><span class="p">(</span><span class="n">data_names</span><span class="p">,</span> <span class="n">label_names</span><span class="p">,</span> <span class="n">data_shapes</span><span class="p">,</span> <span class="n">label_shapes</span><span class="p">):</span>
<span class="sd">"""parse data_attrs into DataDesc format and check that names match"""</span>
<span class="n">data_shapes</span> <span class="o">=</span> <span class="p">[</span><span class="n">x</span> <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">DataDesc</span><span class="p">)</span> <span class="k">else</span> <span class="n">DataDesc</span><span class="p">(</span><span class="o">*</span><span class="n">x</span><span class="p">)</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">data_shapes</span><span class="p">]</span>
<span class="n">_check_names_match</span><span class="p">(</span><span class="n">data_names</span><span class="p">,</span> <span class="n">data_shapes</span><span class="p">,</span> <span class="s1">'data'</span><span class="p">,</span> <span class="kc">True</span><span class="p">)</span>
<span class="k">if</span> <span class="n">label_shapes</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">label_shapes</span> <span class="o">=</span> <span class="p">[</span><span class="n">x</span> <span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">x</span><span class="p">,</span> <span class="n">DataDesc</span><span class="p">)</span> <span class="k">else</span> <span class="n">DataDesc</span><span class="p">(</span><span class="o">*</span><span class="n">x</span><span class="p">)</span> <span class="k">for</span> <span class="n">x</span> <span class="ow">in</span> <span class="n">label_shapes</span><span class="p">]</span>
<span class="n">_check_names_match</span><span class="p">(</span><span class="n">label_names</span><span class="p">,</span> <span class="n">label_shapes</span><span class="p">,</span> <span class="s1">'label'</span><span class="p">,</span> <span class="kc">False</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="n">_check_names_match</span><span class="p">(</span><span class="n">label_names</span><span class="p">,</span> <span class="p">[],</span> <span class="s1">'label'</span><span class="p">,</span> <span class="kc">False</span><span class="p">)</span>
<span class="k">return</span> <span class="n">data_shapes</span><span class="p">,</span> <span class="n">label_shapes</span>
<div class="viewcode-block" id="BaseModule"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule">[docs]</a><span class="k">class</span> <span class="nc">BaseModule</span><span class="p">(</span><span class="nb">object</span><span class="p">):</span>
<span class="sd">"""The base class of a module.</span>
<span class="sd"> A module represents a computation component. One can think of module as a computation machine.</span>
<span class="sd"> A module can execute forward and backward passes and update parameters in a model.</span>
<span class="sd"> We aim to make the APIs easy to use, especially in the case when we need to use the imperative</span>
<span class="sd"> API to work with multiple modules (e.g. stochastic depth network).</span>
<span class="sd"> A module has several states:</span>
<span class="sd"> - Initial state: Memory is not allocated yet, so the module is not ready for computation yet.</span>
<span class="sd"> - Binded: Shapes for inputs, outputs, and parameters are all known, memory has been allocated,</span>
<span class="sd"> and the module is ready for computation.</span>
<span class="sd"> - Parameters are initialized: For modules with parameters, doing computation before</span>
<span class="sd"> initializing the parameters might result in undefined outputs.</span>
<span class="sd"> - Optimizer is installed: An optimizer can be installed to a module. After this, the parameters</span>
<span class="sd"> of the module can be updated according to the optimizer after gradients are computed</span>
<span class="sd"> (forward-backward).</span>
<span class="sd"> In order for a module to interact with others, it must be able to report the</span>
<span class="sd"> following information in its initial state (before binding):</span>
<span class="sd"> - `data_names`: list of type string indicating the names of the required input data.</span>
<span class="sd"> - `output_names`: list of type string indicating the names of the required outputs.</span>
<span class="sd"> After binding, a module should be able to report the following richer information:</span>
<span class="sd"> - state information</span>
<span class="sd"> - `binded`: `bool`, indicates whether the memory buffers needed for computation</span>
<span class="sd"> have been allocated.</span>
<span class="sd"> - `for_training`: whether the module is bound for training.</span>
<span class="sd"> - `params_initialized`: `bool`, indicates whether the parameters of this module</span>
<span class="sd"> have been initialized.</span>
<span class="sd"> - `optimizer_initialized`: `bool`, indicates whether an optimizer is defined</span>
<span class="sd"> and initialized.</span>
<span class="sd"> - `inputs_need_grad`: `bool`, indicates whether gradients with respect to the</span>
<span class="sd"> input data are needed. Might be useful when implementing composition of modules.</span>
<span class="sd"> - input/output information</span>
<span class="sd"> - `data_shapes`: a list of `(name, shape)`. In theory, since the memory is allocated,</span>
<span class="sd"> we could directly provide the data arrays. But in the case of data parallelism,</span>
<span class="sd"> the data arrays might not be of the same shape as viewed from the external world.</span>
<span class="sd"> - `label_shapes`: a list of `(name, shape)`. This might be `[]` if the module does</span>
<span class="sd"> not need labels (e.g. it does not contains a loss function at the top), or a module</span>
<span class="sd"> is not bound for training.</span>
<span class="sd"> - `output_shapes`: a list of `(name, shape)` for outputs of the module.</span>
<span class="sd"> - parameters (for modules with parameters)</span>
<span class="sd"> - `get_params()`: return a tuple `(arg_params, aux_params)`. Each of those</span>
<span class="sd"> is a dictionary of name to ``NDArray`` mapping. Those `NDArray` always lives on</span>
<span class="sd"> CPU. The actual parameters used for computing might live on other devices (GPUs),</span>
<span class="sd"> this function will retrieve (a copy of) the latest parameters.</span>
<span class="sd"> - ``set_params(arg_params, aux_params)``: assign parameters to the devices</span>
<span class="sd"> doing the computation.</span>
<span class="sd"> - ``init_params(...)``: a more flexible interface to assign or initialize the parameters.</span>
<span class="sd"> - setup</span>
<span class="sd"> - `bind()`: prepare environment for computation.</span>
<span class="sd"> - `init_optimizer()`: install optimizer for parameter updating.</span>
<span class="sd"> - `prepare()`: prepare the module based on the current data batch.</span>
<span class="sd"> - computation</span>
<span class="sd"> - `forward(data_batch)`: forward operation.</span>
<span class="sd"> - `backward(out_grads=None)`: backward operation.</span>
<span class="sd"> - `update()`: update parameters according to installed optimizer.</span>
<span class="sd"> - `get_outputs()`: get outputs of the previous forward operation.</span>
<span class="sd"> - `get_input_grads()`: get the gradients with respect to the inputs computed</span>
<span class="sd"> in the previous backward operation.</span>
<span class="sd"> - `update_metric(metric, labels, pre_sliced=False)`: update performance metric</span>
<span class="sd"> for the previous forward</span>
<span class="sd"> computed results.</span>
<span class="sd"> - other properties (mostly for backward compatibility)</span>
<span class="sd"> - `symbol`: the underlying symbolic graph for this module (if any)</span>
<span class="sd"> This property is not necessarily constant. For example, for `BucketingModule`,</span>
<span class="sd"> this property is simply the *current* symbol being used. For other modules,</span>
<span class="sd"> this value might not be well defined.</span>
<span class="sd"> When those intermediate-level API are implemented properly, the following</span>
<span class="sd"> high-level API will be automatically available for a module:</span>
<span class="sd"> - `fit`: train the module parameters on a data set.</span>
<span class="sd"> - `predict`: run prediction on a data set and collect outputs.</span>
<span class="sd"> - `score`: run prediction on a data set and evaluate performance.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of creating a mxnet module.</span>
<span class="sd"> >>> import mxnet as mx</span>
<span class="sd"> >>> data = mx.symbol.Variable('data')</span>
<span class="sd"> >>> fc1 = mx.symbol.FullyConnected(data, name='fc1', num_hidden=128)</span>
<span class="sd"> >>> act1 = mx.symbol.Activation(fc1, name='relu1', act_type="relu")</span>
<span class="sd"> >>> fc2 = mx.symbol.FullyConnected(act1, name = 'fc2', num_hidden = 64)</span>
<span class="sd"> >>> act2 = mx.symbol.Activation(fc2, name='relu2', act_type="relu")</span>
<span class="sd"> >>> fc3 = mx.symbol.FullyConnected(act2, name='fc3', num_hidden=10)</span>
<span class="sd"> >>> out = mx.symbol.SoftmaxOutput(fc3, name = 'softmax')</span>
<span class="sd"> >>> mod = mx.mod.Module(out)</span>
<span class="sd"> """</span>
<span class="k">def</span> <span class="nf">__init__</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">logger</span><span class="o">=</span><span class="n">logging</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">logger</span> <span class="o">=</span> <span class="n">logger</span>
<span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="o">=</span> <span class="kc">False</span>
<span class="bp">self</span><span class="o">.</span><span class="n">for_training</span> <span class="o">=</span> <span class="kc">False</span>
<span class="bp">self</span><span class="o">.</span><span class="n">inputs_need_grad</span> <span class="o">=</span> <span class="kc">False</span>
<span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span> <span class="o">=</span> <span class="kc">False</span>
<span class="bp">self</span><span class="o">.</span><span class="n">optimizer_initialized</span> <span class="o">=</span> <span class="kc">False</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_symbol</span> <span class="o">=</span> <span class="kc">None</span>
<span class="bp">self</span><span class="o">.</span><span class="n">_total_exec_bytes</span> <span class="o">=</span> <span class="mi">0</span>
<span class="c1">################################################################################</span>
<span class="c1"># High Level API</span>
<span class="c1">################################################################################</span>
<div class="viewcode-block" id="BaseModule.forward_backward"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.forward_backward">[docs]</a> <span class="k">def</span> <span class="nf">forward_backward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data_batch</span><span class="p">):</span>
<span class="sd">"""A convenient function that calls both ``forward`` and ``backward``."""</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">data_batch</span><span class="p">,</span> <span class="n">is_train</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.score"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.score">[docs]</a> <span class="k">def</span> <span class="nf">score</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">eval_data</span><span class="p">,</span> <span class="n">eval_metric</span><span class="p">,</span> <span class="n">num_batch</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">batch_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">score_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">reset</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">epoch</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Runs prediction on ``eval_data`` and evaluates the performance according to</span>
<span class="sd"> the given ``eval_metric``.</span>
<span class="sd"> Checkout `Module Tutorial </tutorials/basic/module.html>`_ to see</span>
<span class="sd"> a end-to-end use-case.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> eval_data : DataIter</span>
<span class="sd"> Evaluation data to run prediction on.</span>
<span class="sd"> eval_metric : EvalMetric or list of EvalMetrics</span>
<span class="sd"> Evaluation metric to use.</span>
<span class="sd"> num_batch : int</span>
<span class="sd"> Number of batches to run. Defaults to ``None``, indicating run until the `DataIter`</span>
<span class="sd"> finishes.</span>
<span class="sd"> batch_end_callback : function</span>
<span class="sd"> Could also be a list of functions.</span>
<span class="sd"> reset : bool</span>
<span class="sd"> Defaults to ``True``. Indicates whether we should reset `eval_data` before starting</span>
<span class="sd"> evaluating.</span>
<span class="sd"> epoch : int</span>
<span class="sd"> Defaults to 0. For compatibility, this will be passed to callbacks (if any).</span>
<span class="sd"> During training, this will correspond to the training epoch number.</span>
<span class="sd"> sparse_row_id_fn : A callback function</span>
<span class="sd"> The function takes `data_batch` as an input and returns a dict of</span>
<span class="sd"> str -> NDArray. The resulting dict is used for pulling row_sparse</span>
<span class="sd"> parameters from the kvstore, where the str key is the name of the param,</span>
<span class="sd"> and the value is the row id of the param to pull.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of using score for prediction.</span>
<span class="sd"> >>> # Evaluate accuracy on val_dataiter</span>
<span class="sd"> >>> metric = mx.metric.Accuracy()</span>
<span class="sd"> >>> mod.score(val_dataiter, metric)</span>
<span class="sd"> >>> mod.score(val_dataiter, ['mse', 'acc'])</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span>
<span class="k">if</span> <span class="n">reset</span><span class="p">:</span>
<span class="n">eval_data</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
<span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span> <span class="n">metric</span><span class="o">.</span><span class="n">EvalMetric</span><span class="p">):</span>
<span class="n">eval_metric</span> <span class="o">=</span> <span class="n">metric</span><span class="o">.</span><span class="n">create</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">)</span>
<span class="n">eval_metric</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
<span class="n">actual_num_batch</span> <span class="o">=</span> <span class="mi">0</span>
<span class="k">for</span> <span class="n">nbatch</span><span class="p">,</span> <span class="n">eval_batch</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">eval_data</span><span class="p">):</span>
<span class="k">if</span> <span class="n">num_batch</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="n">nbatch</span> <span class="o">==</span> <span class="n">num_batch</span><span class="p">:</span>
<span class="k">break</span>
<span class="bp">self</span><span class="o">.</span><span class="n">prepare</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="n">sparse_row_id_fn</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">is_train</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="nb">list</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">update_metric</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span> <span class="p">[</span><span class="n">eb</span><span class="o">.</span><span class="n">label</span> <span class="k">for</span> <span class="n">eb</span> <span class="ow">in</span> <span class="n">eval_batch</span><span class="p">],</span> <span class="n">pre_sliced</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">update_metric</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span> <span class="n">eval_batch</span><span class="o">.</span><span class="n">label</span><span class="p">)</span>
<span class="k">if</span> <span class="n">batch_end_callback</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">batch_end_params</span> <span class="o">=</span> <span class="n">BatchEndParam</span><span class="p">(</span><span class="n">epoch</span><span class="o">=</span><span class="n">epoch</span><span class="p">,</span>
<span class="n">nbatch</span><span class="o">=</span><span class="n">nbatch</span><span class="p">,</span>
<span class="n">eval_metric</span><span class="o">=</span><span class="n">eval_metric</span><span class="p">,</span>
<span class="nb">locals</span><span class="o">=</span><span class="nb">locals</span><span class="p">())</span>
<span class="k">for</span> <span class="n">callback</span> <span class="ow">in</span> <span class="n">_as_list</span><span class="p">(</span><span class="n">batch_end_callback</span><span class="p">):</span>
<span class="n">callback</span><span class="p">(</span><span class="n">batch_end_params</span><span class="p">)</span>
<span class="n">actual_num_batch</span> <span class="o">+=</span> <span class="mi">1</span>
<span class="k">if</span> <span class="n">score_end_callback</span><span class="p">:</span>
<span class="n">params</span> <span class="o">=</span> <span class="n">BatchEndParam</span><span class="p">(</span><span class="n">epoch</span><span class="o">=</span><span class="n">epoch</span><span class="p">,</span>
<span class="n">nbatch</span><span class="o">=</span><span class="n">actual_num_batch</span><span class="p">,</span>
<span class="n">eval_metric</span><span class="o">=</span><span class="n">eval_metric</span><span class="p">,</span>
<span class="nb">locals</span><span class="o">=</span><span class="nb">locals</span><span class="p">())</span>
<span class="k">for</span> <span class="n">callback</span> <span class="ow">in</span> <span class="n">_as_list</span><span class="p">(</span><span class="n">score_end_callback</span><span class="p">):</span>
<span class="n">callback</span><span class="p">(</span><span class="n">params</span><span class="p">)</span>
<span class="k">return</span> <span class="n">eval_metric</span><span class="o">.</span><span class="n">get_name_value</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.iter_predict"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.iter_predict">[docs]</a> <span class="k">def</span> <span class="nf">iter_predict</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">eval_data</span><span class="p">,</span> <span class="n">num_batch</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">reset</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Iterates over predictions.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> for pred, i_batch, batch in module.iter_predict(eval_data):</span>
<span class="sd"> ... # pred is a list of outputs from the module</span>
<span class="sd"> ... # i_batch is a integer</span>
<span class="sd"> ... # batch is the data batch from the data iterator</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> eval_data : DataIter</span>
<span class="sd"> Evaluation data to run prediction on.</span>
<span class="sd"> num_batch : int</span>
<span class="sd"> Default is ``None``, indicating running all the batches in the data iterator.</span>
<span class="sd"> reset : bool</span>
<span class="sd"> Default is ``True``, indicating whether we should reset the data iter before start</span>
<span class="sd"> doing prediction.</span>
<span class="sd"> sparse_row_id_fn : A callback function</span>
<span class="sd"> The function takes `data_batch` as an input and returns a dict of</span>
<span class="sd"> str -> NDArray. The resulting dict is used for pulling row_sparse</span>
<span class="sd"> parameters from the kvstore, where the str key is the name of the param,</span>
<span class="sd"> and the value is the row id of the param to pull.</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span>
<span class="k">if</span> <span class="n">reset</span><span class="p">:</span>
<span class="n">eval_data</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
<span class="k">for</span> <span class="n">nbatch</span><span class="p">,</span> <span class="n">eval_batch</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">eval_data</span><span class="p">):</span>
<span class="k">if</span> <span class="n">num_batch</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="n">nbatch</span> <span class="o">==</span> <span class="n">num_batch</span><span class="p">:</span>
<span class="k">break</span>
<span class="bp">self</span><span class="o">.</span><span class="n">prepare</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="n">sparse_row_id_fn</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">is_train</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="n">pad</span> <span class="o">=</span> <span class="n">eval_batch</span><span class="o">.</span><span class="n">pad</span>
<span class="n">outputs</span> <span class="o">=</span> <span class="p">[</span><span class="n">out</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="n">pad</span><span class="p">]</span> <span class="k">for</span> <span class="n">out</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()]</span>
<span class="k">yield</span> <span class="p">(</span><span class="n">outputs</span><span class="p">,</span> <span class="n">nbatch</span><span class="p">,</span> <span class="n">eval_batch</span><span class="p">)</span></div>
<div class="viewcode-block" id="BaseModule.predict"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.predict">[docs]</a> <span class="k">def</span> <span class="nf">predict</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">eval_data</span><span class="p">,</span> <span class="n">num_batch</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">merge_batches</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">reset</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
<span class="n">always_output_list</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Runs prediction and collects the outputs.</span>
<span class="sd"> When `merge_batches` is ``True`` (by default), the return value will be a list</span>
<span class="sd"> ``[out1, out2, out3]``, where each element is formed by concatenating the outputs for</span>
<span class="sd"> all the mini-batches. When `always_output_list` is ``False`` (as by default),</span>
<span class="sd"> then in the case of a single output, `out1` is returned instead of ``[out1]``.</span>
<span class="sd"> When `merge_batches` is ``False``, the return value will be a nested list like</span>
<span class="sd"> ``[[out1_batch1, out2_batch1], [out1_batch2], ...]``. This mode is useful because</span>
<span class="sd"> in some cases (e.g. bucketing), the module does not necessarily produce the same</span>
<span class="sd"> number of outputs.</span>
<span class="sd"> The objects in the results have type `NDArray`. If you need to work with a numpy array,</span>
<span class="sd"> just call ``.asnumpy()`` on each `NDArray`.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> eval_data : DataIter or NDArray or numpy array</span>
<span class="sd"> Evaluation data to run prediction on.</span>
<span class="sd"> num_batch : int</span>
<span class="sd"> Defaults to ``None``, indicates running all the batches in the data iterator.</span>
<span class="sd"> merge_batches : bool</span>
<span class="sd"> Defaults to ``True``, see above for return values.</span>
<span class="sd"> reset : bool</span>
<span class="sd"> Defaults to ``True``, indicates whether we should reset the data iter before</span>
<span class="sd"> doing prediction.</span>
<span class="sd"> always_output_list : bool</span>
<span class="sd"> Defaults to ``False``, see above for return values.</span>
<span class="sd"> sparse_row_id_fn : A callback function</span>
<span class="sd"> The function takes `data_batch` as an input and returns a dict of</span>
<span class="sd"> str -> NDArray. The resulting dict is used for pulling row_sparse</span>
<span class="sd"> parameters from the kvstore, where the str key is the name of the param,</span>
<span class="sd"> and the value is the row id of the param to pull.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> list of NDArray or list of list of NDArray</span>
<span class="sd"> Prediction results.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of using `predict` for prediction.</span>
<span class="sd"> >>> # Predict on the first 10 batches of val_dataiter</span>
<span class="sd"> >>> mod.predict(eval_data=val_dataiter, num_batch=10)</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span>
<span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_data</span><span class="p">,</span> <span class="p">(</span><span class="n">ndarray</span><span class="o">.</span><span class="n">NDArray</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">)):</span>
<span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_data</span><span class="p">,</span> <span class="n">np</span><span class="o">.</span><span class="n">ndarray</span><span class="p">):</span>
<span class="n">eval_data</span> <span class="o">=</span> <span class="n">ndarray</span><span class="o">.</span><span class="n">array</span><span class="p">(</span><span class="n">eval_data</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">DataBatch</span><span class="p">([</span><span class="n">eval_data</span><span class="p">]))</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span>
<span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_data</span><span class="p">,</span> <span class="n">DataIter</span><span class="p">):</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s1">'eval_data must be of type NDArray or DataIter'</span><span class="p">)</span>
<span class="k">if</span> <span class="n">reset</span><span class="p">:</span>
<span class="n">eval_data</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
<span class="n">output_list</span> <span class="o">=</span> <span class="p">[]</span>
<span class="k">for</span> <span class="n">nbatch</span><span class="p">,</span> <span class="n">eval_batch</span> <span class="ow">in</span> <span class="nb">enumerate</span><span class="p">(</span><span class="n">eval_data</span><span class="p">):</span>
<span class="k">if</span> <span class="n">num_batch</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span> <span class="ow">and</span> <span class="n">nbatch</span> <span class="o">==</span> <span class="n">num_batch</span><span class="p">:</span>
<span class="k">break</span>
<span class="bp">self</span><span class="o">.</span><span class="n">prepare</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="n">sparse_row_id_fn</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">eval_batch</span><span class="p">,</span> <span class="n">is_train</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="n">pad</span> <span class="o">=</span> <span class="n">eval_batch</span><span class="o">.</span><span class="n">pad</span>
<span class="n">outputs</span> <span class="o">=</span> <span class="p">[</span><span class="n">out</span><span class="p">[</span><span class="mi">0</span><span class="p">:</span><span class="n">out</span><span class="o">.</span><span class="n">shape</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="n">pad</span><span class="p">]</span><span class="o">.</span><span class="n">copy</span><span class="p">()</span> <span class="k">for</span> <span class="n">out</span> <span class="ow">in</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_outputs</span><span class="p">()]</span>
<span class="n">output_list</span><span class="o">.</span><span class="n">append</span><span class="p">(</span><span class="n">outputs</span><span class="p">)</span>
<span class="k">if</span> <span class="nb">len</span><span class="p">(</span><span class="n">output_list</span><span class="p">)</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
<span class="k">return</span> <span class="n">output_list</span>
<span class="k">if</span> <span class="n">merge_batches</span><span class="p">:</span>
<span class="n">num_outputs</span> <span class="o">=</span> <span class="nb">len</span><span class="p">(</span><span class="n">output_list</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span>
<span class="k">for</span> <span class="n">out</span> <span class="ow">in</span> <span class="n">output_list</span><span class="p">:</span>
<span class="k">assert</span> <span class="nb">len</span><span class="p">(</span><span class="n">out</span><span class="p">)</span> <span class="o">==</span> <span class="n">num_outputs</span><span class="p">,</span> \
<span class="s1">'Cannot merge batches, as num of outputs is not the same '</span> <span class="o">+</span> \
<span class="s1">'in mini-batches. Maybe bucketing is used?'</span>
<span class="n">output_list2</span> <span class="o">=</span> <span class="p">[</span><span class="n">ndarray</span><span class="o">.</span><span class="n">concatenate</span><span class="p">([</span><span class="n">out</span><span class="p">[</span><span class="n">i</span><span class="p">]</span> <span class="k">for</span> <span class="n">out</span> <span class="ow">in</span> <span class="n">output_list</span><span class="p">])</span>
<span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">num_outputs</span><span class="p">)]</span>
<span class="k">if</span> <span class="n">num_outputs</span> <span class="o">==</span> <span class="mi">1</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">always_output_list</span><span class="p">:</span>
<span class="k">return</span> <span class="n">output_list2</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span>
<span class="k">return</span> <span class="n">output_list2</span>
<span class="k">return</span> <span class="n">output_list</span></div>
<div class="viewcode-block" id="BaseModule.fit"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.fit">[docs]</a> <span class="k">def</span> <span class="nf">fit</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">train_data</span><span class="p">,</span> <span class="n">eval_data</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">eval_metric</span><span class="o">=</span><span class="s1">'acc'</span><span class="p">,</span>
<span class="n">epoch_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">batch_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">kvstore</span><span class="o">=</span><span class="s1">'local'</span><span class="p">,</span>
<span class="n">optimizer</span><span class="o">=</span><span class="s1">'sgd'</span><span class="p">,</span> <span class="n">optimizer_params</span><span class="o">=</span><span class="p">((</span><span class="s1">'learning_rate'</span><span class="p">,</span> <span class="mf">0.01</span><span class="p">),),</span>
<span class="n">eval_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">eval_batch_end_callback</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">initializer</span><span class="o">=</span><span class="n">Uniform</span><span class="p">(</span><span class="mf">0.01</span><span class="p">),</span>
<span class="n">arg_params</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">aux_params</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">allow_missing</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span>
<span class="n">force_rebind</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">force_init</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">begin_epoch</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">num_epoch</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">validation_metric</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">monitor</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Trains the module parameters.</span>
<span class="sd"> Checkout `Module Tutorial </tutorials/basic/module.html>`_ to see</span>
<span class="sd"> a end-to-end use-case.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> train_data : DataIter</span>
<span class="sd"> Train DataIter.</span>
<span class="sd"> eval_data : DataIter</span>
<span class="sd"> If not ``None``, will be used as validation set and the performance</span>
<span class="sd"> after each epoch will be evaluated.</span>
<span class="sd"> eval_metric : str or EvalMetric</span>
<span class="sd"> Defaults to 'accuracy'. The performance measure used to display during training.</span>
<span class="sd"> Other possible predefined metrics are:</span>
<span class="sd"> 'ce' (CrossEntropy), 'f1', 'mae', 'mse', 'rmse', 'top_k_accuracy'.</span>
<span class="sd"> epoch_end_callback : function or list of functions</span>
<span class="sd"> Each callback will be called with the current `epoch`, `symbol`, `arg_params`</span>
<span class="sd"> and `aux_params`.</span>
<span class="sd"> batch_end_callback : function or list of function</span>
<span class="sd"> Each callback will be called with a `BatchEndParam`.</span>
<span class="sd"> kvstore : str or KVStore</span>
<span class="sd"> Defaults to 'local'.</span>
<span class="sd"> optimizer : str or Optimizer</span>
<span class="sd"> Defaults to 'sgd'.</span>
<span class="sd"> optimizer_params : dict</span>
<span class="sd"> Defaults to ``(('learning_rate', 0.01),)``. The parameters for</span>
<span class="sd"> the optimizer constructor.</span>
<span class="sd"> The default value is not a dict, just to avoid pylint warning on dangerous</span>
<span class="sd"> default values.</span>
<span class="sd"> eval_end_callback : function or list of function</span>
<span class="sd"> These will be called at the end of each full evaluation, with the metrics over</span>
<span class="sd"> the entire evaluation set.</span>
<span class="sd"> eval_batch_end_callback : function or list of function</span>
<span class="sd"> These will be called at the end of each mini-batch during evaluation.</span>
<span class="sd"> initializer : Initializer</span>
<span class="sd"> The initializer is called to initialize the module parameters when they are</span>
<span class="sd"> not already initialized.</span>
<span class="sd"> arg_params : dict</span>
<span class="sd"> Defaults to ``None``, if not ``None``, should be existing parameters from a trained</span>
<span class="sd"> model or loaded from a checkpoint (previously saved model). In this case,</span>
<span class="sd"> the value here will be used to initialize the module parameters, unless they</span>
<span class="sd"> are already initialized by the user via a call to `init_params` or `fit`.</span>
<span class="sd"> `arg_params` has a higher priority than `initializer`.</span>
<span class="sd"> aux_params : dict</span>
<span class="sd"> Defaults to ``None``. Similar to `arg_params`, except for auxiliary states.</span>
<span class="sd"> allow_missing : bool</span>
<span class="sd"> Defaults to ``False``. Indicates whether to allow missing parameters when `arg_params`</span>
<span class="sd"> and `aux_params` are not ``None``. If this is ``True``, then the missing parameters</span>
<span class="sd"> will be initialized via the `initializer`.</span>
<span class="sd"> force_rebind : bool</span>
<span class="sd"> Defaults to ``False``. Whether to force rebinding the executors if already bound.</span>
<span class="sd"> force_init : bool</span>
<span class="sd"> Defaults to ``False``. Indicates whether to force initialization even if the</span>
<span class="sd"> parameters are already initialized.</span>
<span class="sd"> begin_epoch : int</span>
<span class="sd"> Defaults to 0. Indicates the starting epoch. Usually, if resumed from a</span>
<span class="sd"> checkpoint saved at a previous training phase at epoch N, then this value should be</span>
<span class="sd"> N+1.</span>
<span class="sd"> num_epoch : int</span>
<span class="sd"> Number of epochs for training.</span>
<span class="sd"> sparse_row_id_fn : A callback function</span>
<span class="sd"> The function takes `data_batch` as an input and returns a dict of</span>
<span class="sd"> str -> NDArray. The resulting dict is used for pulling row_sparse</span>
<span class="sd"> parameters from the kvstore, where the str key is the name of the param,</span>
<span class="sd"> and the value is the row id of the param to pull.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of using fit for training.</span>
<span class="sd"> >>> # Assume training dataIter and validation dataIter are ready</span>
<span class="sd"> >>> # Assume loading a previously checkpointed model</span>
<span class="sd"> >>> sym, arg_params, aux_params = mx.model.load_checkpoint(model_prefix, 3)</span>
<span class="sd"> >>> mod.fit(train_data=train_dataiter, eval_data=val_dataiter, optimizer='sgd',</span>
<span class="sd"> ... optimizer_params={'learning_rate':0.01, 'momentum': 0.9},</span>
<span class="sd"> ... arg_params=arg_params, aux_params=aux_params,</span>
<span class="sd"> ... eval_metric='acc', num_epoch=10, begin_epoch=3)</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="n">num_epoch</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">,</span> <span class="s1">'please specify number of epochs'</span>
<span class="bp">self</span><span class="o">.</span><span class="n">bind</span><span class="p">(</span><span class="n">data_shapes</span><span class="o">=</span><span class="n">train_data</span><span class="o">.</span><span class="n">provide_data</span><span class="p">,</span> <span class="n">label_shapes</span><span class="o">=</span><span class="n">train_data</span><span class="o">.</span><span class="n">provide_label</span><span class="p">,</span>
<span class="n">for_training</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span> <span class="n">force_rebind</span><span class="o">=</span><span class="n">force_rebind</span><span class="p">)</span>
<span class="k">if</span> <span class="n">monitor</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">install_monitor</span><span class="p">(</span><span class="n">monitor</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">init_params</span><span class="p">(</span><span class="n">initializer</span><span class="o">=</span><span class="n">initializer</span><span class="p">,</span> <span class="n">arg_params</span><span class="o">=</span><span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="o">=</span><span class="n">aux_params</span><span class="p">,</span>
<span class="n">allow_missing</span><span class="o">=</span><span class="n">allow_missing</span><span class="p">,</span> <span class="n">force_init</span><span class="o">=</span><span class="n">force_init</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">init_optimizer</span><span class="p">(</span><span class="n">kvstore</span><span class="o">=</span><span class="n">kvstore</span><span class="p">,</span> <span class="n">optimizer</span><span class="o">=</span><span class="n">optimizer</span><span class="p">,</span>
<span class="n">optimizer_params</span><span class="o">=</span><span class="n">optimizer_params</span><span class="p">)</span>
<span class="k">if</span> <span class="n">validation_metric</span> <span class="ow">is</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">validation_metric</span> <span class="o">=</span> <span class="n">eval_metric</span>
<span class="k">if</span> <span class="ow">not</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span> <span class="n">metric</span><span class="o">.</span><span class="n">EvalMetric</span><span class="p">):</span>
<span class="n">eval_metric</span> <span class="o">=</span> <span class="n">metric</span><span class="o">.</span><span class="n">create</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">)</span>
<span class="c1">################################################################################</span>
<span class="c1"># training loop</span>
<span class="c1">################################################################################</span>
<span class="k">for</span> <span class="n">epoch</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">begin_epoch</span><span class="p">,</span> <span class="n">num_epoch</span><span class="p">):</span>
<span class="n">tic</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span>
<span class="n">eval_metric</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
<span class="n">nbatch</span> <span class="o">=</span> <span class="mi">0</span>
<span class="n">data_iter</span> <span class="o">=</span> <span class="nb">iter</span><span class="p">(</span><span class="n">train_data</span><span class="p">)</span>
<span class="n">end_of_batch</span> <span class="o">=</span> <span class="kc">False</span>
<span class="n">next_data_batch</span> <span class="o">=</span> <span class="nb">next</span><span class="p">(</span><span class="n">data_iter</span><span class="p">)</span>
<span class="k">while</span> <span class="ow">not</span> <span class="n">end_of_batch</span><span class="p">:</span>
<span class="n">data_batch</span> <span class="o">=</span> <span class="n">next_data_batch</span>
<span class="k">if</span> <span class="n">monitor</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">monitor</span><span class="o">.</span><span class="n">tic</span><span class="p">()</span>
<span class="bp">self</span><span class="o">.</span><span class="n">forward_backward</span><span class="p">(</span><span class="n">data_batch</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">update</span><span class="p">()</span>
<span class="k">if</span> <span class="nb">isinstance</span><span class="p">(</span><span class="n">data_batch</span><span class="p">,</span> <span class="nb">list</span><span class="p">):</span>
<span class="bp">self</span><span class="o">.</span><span class="n">update_metric</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span>
<span class="p">[</span><span class="n">db</span><span class="o">.</span><span class="n">label</span> <span class="k">for</span> <span class="n">db</span> <span class="ow">in</span> <span class="n">data_batch</span><span class="p">],</span>
<span class="n">pre_sliced</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
<span class="k">else</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">update_metric</span><span class="p">(</span><span class="n">eval_metric</span><span class="p">,</span> <span class="n">data_batch</span><span class="o">.</span><span class="n">label</span><span class="p">)</span>
<span class="k">try</span><span class="p">:</span>
<span class="c1"># pre fetch next batch</span>
<span class="n">next_data_batch</span> <span class="o">=</span> <span class="nb">next</span><span class="p">(</span><span class="n">data_iter</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">prepare</span><span class="p">(</span><span class="n">next_data_batch</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="n">sparse_row_id_fn</span><span class="p">)</span>
<span class="k">except</span> <span class="ne">StopIteration</span><span class="p">:</span>
<span class="n">end_of_batch</span> <span class="o">=</span> <span class="kc">True</span>
<span class="k">if</span> <span class="n">monitor</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">monitor</span><span class="o">.</span><span class="n">toc_print</span><span class="p">()</span>
<span class="k">if</span> <span class="n">end_of_batch</span><span class="p">:</span>
<span class="n">eval_name_vals</span> <span class="o">=</span> <span class="n">eval_metric</span><span class="o">.</span><span class="n">get_global_name_value</span><span class="p">()</span>
<span class="k">if</span> <span class="n">batch_end_callback</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">batch_end_params</span> <span class="o">=</span> <span class="n">BatchEndParam</span><span class="p">(</span><span class="n">epoch</span><span class="o">=</span><span class="n">epoch</span><span class="p">,</span> <span class="n">nbatch</span><span class="o">=</span><span class="n">nbatch</span><span class="p">,</span>
<span class="n">eval_metric</span><span class="o">=</span><span class="n">eval_metric</span><span class="p">,</span>
<span class="nb">locals</span><span class="o">=</span><span class="nb">locals</span><span class="p">())</span>
<span class="k">for</span> <span class="n">callback</span> <span class="ow">in</span> <span class="n">_as_list</span><span class="p">(</span><span class="n">batch_end_callback</span><span class="p">):</span>
<span class="n">callback</span><span class="p">(</span><span class="n">batch_end_params</span><span class="p">)</span>
<span class="n">nbatch</span> <span class="o">+=</span> <span class="mi">1</span>
<span class="c1"># one epoch of training is finished</span>
<span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">val</span> <span class="ow">in</span> <span class="n">eval_name_vals</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Epoch[</span><span class="si">%d</span><span class="s1">] Train-</span><span class="si">%s</span><span class="s1">=</span><span class="si">%f</span><span class="s1">'</span><span class="p">,</span> <span class="n">epoch</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="n">val</span><span class="p">)</span>
<span class="n">toc</span> <span class="o">=</span> <span class="n">time</span><span class="o">.</span><span class="n">time</span><span class="p">()</span>
<span class="bp">self</span><span class="o">.</span><span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Epoch[</span><span class="si">%d</span><span class="s1">] Time cost=</span><span class="si">%.3f</span><span class="s1">'</span><span class="p">,</span> <span class="n">epoch</span><span class="p">,</span> <span class="p">(</span><span class="n">toc</span><span class="o">-</span><span class="n">tic</span><span class="p">))</span>
<span class="c1"># sync aux params across devices</span>
<span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_params</span><span class="p">()</span>
<span class="bp">self</span><span class="o">.</span><span class="n">set_params</span><span class="p">(</span><span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="p">)</span>
<span class="k">if</span> <span class="n">epoch_end_callback</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="k">for</span> <span class="n">callback</span> <span class="ow">in</span> <span class="n">_as_list</span><span class="p">(</span><span class="n">epoch_end_callback</span><span class="p">):</span>
<span class="n">callback</span><span class="p">(</span><span class="n">epoch</span><span class="p">,</span> <span class="bp">self</span><span class="o">.</span><span class="n">symbol</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="p">)</span>
<span class="c1">#----------------------------------------</span>
<span class="c1"># evaluation on validation set</span>
<span class="k">if</span> <span class="n">eval_data</span><span class="p">:</span>
<span class="n">res</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">eval_data</span><span class="p">,</span> <span class="n">validation_metric</span><span class="p">,</span>
<span class="n">score_end_callback</span><span class="o">=</span><span class="n">eval_end_callback</span><span class="p">,</span>
<span class="n">batch_end_callback</span><span class="o">=</span><span class="n">eval_batch_end_callback</span><span class="p">,</span> <span class="n">epoch</span><span class="o">=</span><span class="n">epoch</span><span class="p">)</span>
<span class="c1">#TODO: pull this into default</span>
<span class="k">for</span> <span class="n">name</span><span class="p">,</span> <span class="n">val</span> <span class="ow">in</span> <span class="n">res</span><span class="p">:</span>
<span class="bp">self</span><span class="o">.</span><span class="n">logger</span><span class="o">.</span><span class="n">info</span><span class="p">(</span><span class="s1">'Epoch[</span><span class="si">%d</span><span class="s1">] Validation-</span><span class="si">%s</span><span class="s1">=</span><span class="si">%f</span><span class="s1">'</span><span class="p">,</span> <span class="n">epoch</span><span class="p">,</span> <span class="n">name</span><span class="p">,</span> <span class="n">val</span><span class="p">)</span>
<span class="c1"># end of 1 epoch, reset the data-iter for another epoch</span>
<span class="n">train_data</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span></div>
<span class="c1">################################################################################</span>
<span class="c1"># Symbol information</span>
<span class="c1">################################################################################</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">data_names</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""A list of names for data required by this module."""</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">output_names</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""A list of names for the outputs of this module."""</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span>
<span class="c1">################################################################################</span>
<span class="c1"># Input/Output information</span>
<span class="c1">################################################################################</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">data_shapes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""A list of (name, shape) pairs specifying the data inputs to this module."""</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">label_shapes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""A list of (name, shape) pairs specifying the label inputs to this module.</span>
<span class="sd"> If this module does not accept labels -- either it is a module without loss</span>
<span class="sd"> function, or it is not bound for training, then this should return an empty</span>
<span class="sd"> list ``[]``.</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">output_shapes</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""A list of (name, shape) pairs specifying the outputs of this module."""</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span>
<span class="c1">################################################################################</span>
<span class="c1"># Parameters of a module</span>
<span class="c1">################################################################################</span>
<div class="viewcode-block" id="BaseModule.get_params"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.get_params">[docs]</a> <span class="k">def</span> <span class="nf">get_params</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""Gets parameters, those are potentially copies of the the actual parameters used</span>
<span class="sd"> to do computation on the device.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> ``(arg_params, aux_params)``</span>
<span class="sd"> A pair of dictionaries each mapping parameter names to NDArray values.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of getting module parameters.</span>
<span class="sd"> >>> print mod.get_params()</span>
<span class="sd"> ({'fc2_weight': <NDArray 64x128 @cpu(0)>, 'fc1_weight': <NDArray 128x100 @cpu(0)>,</span>
<span class="sd"> 'fc3_bias': <NDArray 10 @cpu(0)>, 'fc3_weight': <NDArray 10x64 @cpu(0)>,</span>
<span class="sd"> 'fc2_bias': <NDArray 64 @cpu(0)>, 'fc1_bias': <NDArray 128 @cpu(0)>}, {})</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.init_params"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.init_params">[docs]</a> <span class="k">def</span> <span class="nf">init_params</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">initializer</span><span class="o">=</span><span class="n">Uniform</span><span class="p">(</span><span class="mf">0.01</span><span class="p">),</span> <span class="n">arg_params</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">aux_params</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">allow_missing</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">force_init</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">allow_extra</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""Initializes the parameters and auxiliary states.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> initializer : Initializer</span>
<span class="sd"> Called to initialize parameters if needed.</span>
<span class="sd"> arg_params : dict</span>
<span class="sd"> If not ``None``, should be a dictionary of existing `arg_params`. Initialization</span>
<span class="sd"> will be copied from that.</span>
<span class="sd"> aux_params : dict</span>
<span class="sd"> If not ``None``, should be a dictionary of existing `aux_params`. Initialization</span>
<span class="sd"> will be copied from that.</span>
<span class="sd"> allow_missing : bool</span>
<span class="sd"> If ``True``, params could contain missing values, and the initializer will be</span>
<span class="sd"> called to fill those missing params.</span>
<span class="sd"> force_init : bool</span>
<span class="sd"> If ``True``, `force_init` will force re-initialize even if already initialized.</span>
<span class="sd"> allow_extra : boolean, optional</span>
<span class="sd"> Whether allow extra parameters that are not needed by symbol.</span>
<span class="sd"> If this is True, no error will be thrown when arg_params or aux_params</span>
<span class="sd"> contain extra parameters that is not needed by the executor.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of initializing module parameters.</span>
<span class="sd"> >>> mod.init_params()</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.set_params"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.set_params">[docs]</a> <span class="k">def</span> <span class="nf">set_params</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="p">,</span> <span class="n">allow_missing</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">force_init</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
<span class="n">allow_extra</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""Assigns parameter and aux state values.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> arg_params : dict</span>
<span class="sd"> Dictionary of name to value (`NDArray`) mapping.</span>
<span class="sd"> aux_params : dict</span>
<span class="sd"> Dictionary of name to value (`NDArray`) mapping.</span>
<span class="sd"> allow_missing : bool</span>
<span class="sd"> If ``True``, params could contain missing values, and the initializer will be</span>
<span class="sd"> called to fill those missing params.</span>
<span class="sd"> force_init : bool</span>
<span class="sd"> If ``True``, will force re-initialize even if already initialized.</span>
<span class="sd"> allow_extra : boolean, optional</span>
<span class="sd"> Whether allow extra parameters that are not needed by symbol.</span>
<span class="sd"> If this is True, no error will be thrown when arg_params or aux_params</span>
<span class="sd"> contain extra parameters that is not needed by the executor.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of setting module parameters.</span>
<span class="sd"> >>> sym, arg_params, aux_params = mx.model.load_checkpoint(model_prefix, n_epoch_load)</span>
<span class="sd"> >>> mod.set_params(arg_params=arg_params, aux_params=aux_params)</span>
<span class="sd"> """</span>
<span class="bp">self</span><span class="o">.</span><span class="n">init_params</span><span class="p">(</span><span class="n">initializer</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">arg_params</span><span class="o">=</span><span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="o">=</span><span class="n">aux_params</span><span class="p">,</span>
<span class="n">allow_missing</span><span class="o">=</span><span class="n">allow_missing</span><span class="p">,</span> <span class="n">force_init</span><span class="o">=</span><span class="n">force_init</span><span class="p">,</span>
<span class="n">allow_extra</span><span class="o">=</span><span class="n">allow_extra</span><span class="p">)</span></div>
<div class="viewcode-block" id="BaseModule.save_params"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.save_params">[docs]</a> <span class="k">def</span> <span class="nf">save_params</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">fname</span><span class="p">):</span>
<span class="sd">"""Saves model parameters to file.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> fname : str</span>
<span class="sd"> Path to output param file.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of saving module parameters.</span>
<span class="sd"> >>> mod.save_params('myfile')</span>
<span class="sd"> """</span>
<span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="bp">self</span><span class="o">.</span><span class="n">get_params</span><span class="p">()</span>
<span class="n">save_dict</span> <span class="o">=</span> <span class="p">{(</span><span class="s1">'arg:</span><span class="si">%s</span><span class="s1">'</span> <span class="o">%</span> <span class="n">k</span><span class="p">)</span> <span class="p">:</span> <span class="n">v</span><span class="o">.</span><span class="n">as_in_context</span><span class="p">(</span><span class="n">cpu</span><span class="p">())</span> <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">arg_params</span><span class="o">.</span><span class="n">items</span><span class="p">()}</span>
<span class="n">save_dict</span><span class="o">.</span><span class="n">update</span><span class="p">({(</span><span class="s1">'aux:</span><span class="si">%s</span><span class="s1">'</span> <span class="o">%</span> <span class="n">k</span><span class="p">)</span> <span class="p">:</span> <span class="n">v</span><span class="o">.</span><span class="n">as_in_context</span><span class="p">(</span><span class="n">cpu</span><span class="p">())</span> <span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">v</span> <span class="ow">in</span> <span class="n">aux_params</span><span class="o">.</span><span class="n">items</span><span class="p">()})</span>
<span class="n">ndarray</span><span class="o">.</span><span class="n">save</span><span class="p">(</span><span class="n">fname</span><span class="p">,</span> <span class="n">save_dict</span><span class="p">)</span></div>
<div class="viewcode-block" id="BaseModule.load_params"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.load_params">[docs]</a> <span class="k">def</span> <span class="nf">load_params</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">fname</span><span class="p">):</span>
<span class="sd">"""Loads model parameters from file.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> fname : str</span>
<span class="sd"> Path to input param file.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of loading module parameters.</span>
<span class="sd"> >>> mod.load_params('myfile')</span>
<span class="sd"> """</span>
<span class="n">save_dict</span> <span class="o">=</span> <span class="n">ndarray</span><span class="o">.</span><span class="n">load</span><span class="p">(</span><span class="n">fname</span><span class="p">)</span>
<span class="n">arg_params</span> <span class="o">=</span> <span class="p">{}</span>
<span class="n">aux_params</span> <span class="o">=</span> <span class="p">{}</span>
<span class="k">for</span> <span class="n">k</span><span class="p">,</span> <span class="n">value</span> <span class="ow">in</span> <span class="n">save_dict</span><span class="o">.</span><span class="n">items</span><span class="p">():</span>
<span class="n">arg_type</span><span class="p">,</span> <span class="n">name</span> <span class="o">=</span> <span class="n">k</span><span class="o">.</span><span class="n">split</span><span class="p">(</span><span class="s1">':'</span><span class="p">,</span> <span class="mi">1</span><span class="p">)</span>
<span class="k">if</span> <span class="n">arg_type</span> <span class="o">==</span> <span class="s1">'arg'</span><span class="p">:</span>
<span class="n">arg_params</span><span class="p">[</span><span class="n">name</span><span class="p">]</span> <span class="o">=</span> <span class="n">value</span>
<span class="k">elif</span> <span class="n">arg_type</span> <span class="o">==</span> <span class="s1">'aux'</span><span class="p">:</span>
<span class="n">aux_params</span><span class="p">[</span><span class="n">name</span><span class="p">]</span> <span class="o">=</span> <span class="n">value</span>
<span class="k">else</span><span class="p">:</span>
<span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Invalid param file "</span> <span class="o">+</span> <span class="n">fname</span><span class="p">)</span>
<span class="bp">self</span><span class="o">.</span><span class="n">set_params</span><span class="p">(</span><span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span><span class="p">)</span></div>
<div class="viewcode-block" id="BaseModule.get_states"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.get_states">[docs]</a> <span class="k">def</span> <span class="nf">get_states</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">merge_multi_context</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
<span class="sd">"""Gets states from all devices</span>
<span class="sd"> If `merge_multi_context` is ``True``, returns output of form ``[out1, out2]``.</span>
<span class="sd"> Otherwise, it returns output of the form</span>
<span class="sd"> ``[[out1_dev1, out1_dev2], [out2_dev1, out2_dev2]]``.</span>
<span class="sd"> All output elements are `NDArray`.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> merge_multi_context : bool</span>
<span class="sd"> Defaults to ``True``. In the case when data-parallelism is used, the states</span>
<span class="sd"> will be collected from multiple devices. A ``True`` value indicates that we</span>
<span class="sd"> should merge the collected results so that they look like from a single</span>
<span class="sd"> executor.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> A list of ``NDArray`` or a list of list of ``NDArray``.</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span>
<span class="k">assert</span> <span class="ow">not</span> <span class="n">merge_multi_context</span>
<span class="k">return</span> <span class="p">[]</span></div>
<div class="viewcode-block" id="BaseModule.set_states"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.set_states">[docs]</a> <span class="k">def</span> <span class="nf">set_states</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">states</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">value</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Sets value for states. Only one of states &amp; value can be specified.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> states : list of list of NDArray</span>
<span class="sd"> Source states arrays formatted like</span>
<span class="sd"> ``[[state1_dev1, state1_dev2], [state2_dev1, state2_dev2]]``.</span>
<span class="sd"> value : number</span>
<span class="sd"> A single scalar value for all state arrays.</span>
<span class="sd"> """</span>
<span class="k">assert</span> <span class="bp">self</span><span class="o">.</span><span class="n">binded</span> <span class="ow">and</span> <span class="bp">self</span><span class="o">.</span><span class="n">params_initialized</span>
<span class="k">assert</span> <span class="ow">not</span> <span class="n">states</span> <span class="ow">and</span> <span class="ow">not</span> <span class="n">value</span></div>
<div class="viewcode-block" id="BaseModule.install_monitor"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.install_monitor">[docs]</a> <span class="k">def</span> <span class="nf">install_monitor</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">mon</span><span class="p">):</span>
<span class="sd">"""Installs monitor on all executors."""</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<span class="c1">################################################################################</span>
<span class="c1"># Computations</span>
<span class="c1">################################################################################</span>
<span class="c1"># pylint: disable=unused-argument</span>
<div class="viewcode-block" id="BaseModule.prepare"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.prepare">[docs]</a> <span class="k">def</span> <span class="nf">prepare</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data_batch</span><span class="p">,</span> <span class="n">sparse_row_id_fn</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">'''Prepares the module for processing a data batch.</span>
<span class="sd"> Usually involves switching bucket and reshaping.</span>
<span class="sd"> For modules that contain `row_sparse` parameters in KVStore,</span>
<span class="sd"> it prepares the `row_sparse` parameters based on the sparse_row_id_fn.</span>
<span class="sd"> When KVStore is used to update parameters for multi-device or multi-machine training,</span>
<span class="sd"> a copy of the parameters are stored in KVStore. Note that for `row_sparse` parameters,</span>
<span class="sd"> the `update()` updates the copy of parameters in KVStore, but doesn't broadcast</span>
<span class="sd"> the updated parameters to all devices / machines. The `prepare` function is used to</span>
<span class="sd"> broadcast `row_sparse` parameters with the next batch of data.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> data_batch : DataBatch</span>
<span class="sd"> The current batch of data for forward computation.</span>
<span class="sd"> sparse_row_id_fn : A callback function</span>
<span class="sd"> The function takes `data_batch` as an input and returns a dict of</span>
<span class="sd"> str -> NDArray. The resulting dict is used for pulling row_sparse</span>
<span class="sd"> parameters from the kvstore, where the str key is the name of the param,</span>
<span class="sd"> and the value is the row id of the param to pull.</span>
<span class="sd"> '''</span>
<span class="k">if</span> <span class="n">sparse_row_id_fn</span> <span class="ow">is</span> <span class="ow">not</span> <span class="kc">None</span><span class="p">:</span>
<span class="n">warnings</span><span class="o">.</span><span class="n">warn</span><span class="p">(</span><span class="ne">UserWarning</span><span class="p">(</span><span class="s2">"sparse_row_id_fn is not invoked for BaseModule."</span><span class="p">))</span></div>
<span class="c1"># pylint: enable=unused-argument</span>
<div class="viewcode-block" id="BaseModule.forward"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.forward">[docs]</a> <span class="k">def</span> <span class="nf">forward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data_batch</span><span class="p">,</span> <span class="n">is_train</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Forward computation. It supports data batches with different shapes, such as</span>
<span class="sd"> different batch sizes or different image sizes.</span>
<span class="sd"> If reshaping of data batch relates to modification of symbol or module, such as</span>
<span class="sd"> changing image layout ordering or switching from training to predicting, module</span>
<span class="sd"> rebinding is required.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> data_batch : DataBatch</span>
<span class="sd"> Could be anything with similar API implemented.</span>
<span class="sd"> is_train : bool</span>
<span class="sd"> Default is ``None``, which means `is_train` takes the value of ``self.for_training``.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> import mxnet as mx</span>
<span class="sd"> >>> from collections import namedtuple</span>
<span class="sd"> >>> Batch = namedtuple('Batch', ['data'])</span>
<span class="sd"> >>> data = mx.sym.Variable('data')</span>
<span class="sd"> >>> out = data * 2</span>
<span class="sd"> >>> mod = mx.mod.Module(symbol=out, label_names=None)</span>
<span class="sd"> >>> mod.bind(data_shapes=[('data', (1, 10))])</span>
<span class="sd"> >>> mod.init_params()</span>
<span class="sd"> >>> data1 = [mx.nd.ones((1, 10))]</span>
<span class="sd"> >>> mod.forward(Batch(data1))</span>
<span class="sd"> >>> print mod.get_outputs()[0].asnumpy()</span>
<span class="sd"> [[ 2. 2. 2. 2. 2. 2. 2. 2. 2. 2.]]</span>
<span class="sd"> >>> # Forward with data batch of different shape</span>
<span class="sd"> >>> data2 = [mx.nd.ones((3, 5))]</span>
<span class="sd"> >>> mod.forward(Batch(data2))</span>
<span class="sd"> >>> print mod.get_outputs()[0].asnumpy()</span>
<span class="sd"> [[ 2. 2. 2. 2. 2.]</span>
<span class="sd"> [ 2. 2. 2. 2. 2.]</span>
<span class="sd"> [ 2. 2. 2. 2. 2.]]</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.backward"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.backward">[docs]</a> <span class="k">def</span> <span class="nf">backward</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">out_grads</span><span class="o">=</span><span class="kc">None</span><span class="p">):</span>
<span class="sd">"""Backward computation.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> out_grads : NDArray or list of NDArray, optional</span>
<span class="sd"> Gradient on the outputs to be propagated back.</span>
<span class="sd"> This parameter is only needed when bind is called</span>
<span class="sd"> on outputs that are not a loss function.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of backward computation.</span>
<span class="sd"> >>> mod.backward()</span>
<span class="sd"> >>> print mod.get_input_grads()[0].asnumpy()</span>
<span class="sd"> [[[ 1.10182791e-05 5.12257748e-06 4.01927764e-06 8.32566820e-06</span>
<span class="sd"> -1.59775993e-06 7.24269375e-06 7.28067835e-06 -1.65902311e-05</span>
<span class="sd"> 5.46342608e-06 8.44196393e-07]</span>
<span class="sd"> ...]]</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.get_outputs"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.get_outputs">[docs]</a> <span class="k">def</span> <span class="nf">get_outputs</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">merge_multi_context</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
<span class="sd">"""Gets outputs of the previous forward computation.</span>
<span class="sd"> If `merge_multi_context` is ``True``, it is like ``[out1, out2]``. Otherwise,</span>
<span class="sd"> it returns out put of form ``[[out1_dev1, out1_dev2], [out2_dev1, out2_dev2]]``.</span>
<span class="sd"> All the output elements have type `NDArray`. When `merge_multi_context` is ``False``,</span>
<span class="sd"> those `NDArray` instances might live on different devices.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> merge_multi_context : bool</span>
<span class="sd"> Defaults to ``True``. In the case when data-parallelism is used, the outputs</span>
<span class="sd"> will be collected from multiple devices. A ``True`` value indicates that we</span>
<span class="sd"> should merge the collected results so that they look like from a single</span>
<span class="sd"> executor.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> list of `NDArray` or list of list of `NDArray`.</span>
<span class="sd"> Output</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of getting forward output.</span>
<span class="sd"> >>> print mod.get_outputs()[0].asnumpy()</span>
<span class="sd"> [[ 0.09999977 0.10000153 0.10000716 0.10000195 0.09999853 0.09999743</span>
<span class="sd"> 0.10000272 0.10000113 0.09999088 0.09999888]]</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.get_input_grads"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.get_input_grads">[docs]</a> <span class="k">def</span> <span class="nf">get_input_grads</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">merge_multi_context</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
<span class="sd">"""Gets the gradients to the inputs, computed in the previous backward computation.</span>
<span class="sd"> If `merge_multi_context` is ``True``, it is like ``[grad1, grad2]``. Otherwise, it</span>
<span class="sd"> is like ``[[grad1_dev1, grad1_dev2], [grad2_dev1, grad2_dev2]]``. All the output</span>
<span class="sd"> elements have type `NDArray`. When `merge_multi_context` is ``False``, those `NDArray`</span>
<span class="sd"> instances might live on different devices.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> merge_multi_context : bool</span>
<span class="sd"> Defaults to ``True``. In the case when data-parallelism is used, the gradients</span>
<span class="sd"> will be collected from multiple devices. A ``True`` value indicates that we</span>
<span class="sd"> should merge the collected results so that they look like from a single</span>
<span class="sd"> executor.</span>
<span class="sd"> Returns</span>
<span class="sd"> -------</span>
<span class="sd"> list of NDArray or list of list of NDArray</span>
<span class="sd"> Input gradients.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of getting input gradients.</span>
<span class="sd"> >>> print mod.get_input_grads()[0].asnumpy()</span>
<span class="sd"> [[[ 1.10182791e-05 5.12257748e-06 4.01927764e-06 8.32566820e-06</span>
<span class="sd"> -1.59775993e-06 7.24269375e-06 7.28067835e-06 -1.65902311e-05</span>
<span class="sd"> 5.46342608e-06 8.44196393e-07]</span>
<span class="sd"> ...]]</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.update"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.update">[docs]</a> <span class="k">def</span> <span class="nf">update</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""Updates parameters according to the installed optimizer and the gradients computed</span>
<span class="sd"> in the previous forward-backward batch.</span>
<span class="sd"> When KVStore is used to update parameters for multi-device or multi-machine training,</span>
<span class="sd"> a copy of the parameters are stored in KVStore. Note that for `row_sparse` parameters,</span>
<span class="sd"> this function does update the copy of parameters in KVStore, but doesn't broadcast the</span>
<span class="sd"> updated parameters to all devices / machines. Please call `prepare` to broadcast</span>
<span class="sd"> `row_sparse` parameters with the next batch of data.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of updating module parameters.</span>
<span class="sd"> >>> mod.init_optimizer(kvstore='local', optimizer='sgd',</span>
<span class="sd"> ... optimizer_params=(('learning_rate', 0.01), ))</span>
<span class="sd"> >>> mod.backward()</span>
<span class="sd"> >>> mod.update()</span>
<span class="sd"> >>> print mod.get_params()[0]['fc3_weight'].asnumpy()</span>
<span class="sd"> [[ 5.86930104e-03 5.28078526e-03 -8.88729654e-03 -1.08308345e-03</span>
<span class="sd"> 6.13054074e-03 4.27560415e-03 1.53817423e-03 4.62131854e-03</span>
<span class="sd"> 4.69872449e-03 -2.42400169e-03 9.94111411e-04 1.12386420e-03</span>
<span class="sd"> ...]]</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.update_metric"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.update_metric">[docs]</a> <span class="k">def</span> <span class="nf">update_metric</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">eval_metric</span><span class="p">,</span> <span class="n">labels</span><span class="p">,</span> <span class="n">pre_sliced</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""Evaluates and accumulates evaluation metric on outputs of the last forward</span>
<span class="sd"> computation.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> eval_metric : EvalMetric</span>
<span class="sd"> Evaluation metric to use.</span>
<span class="sd"> labels : list of NDArray if `pre_sliced` parameter is set to `False`,</span>
<span class="sd"> list of lists of NDArray otherwise. Typically `data_batch.label`.</span>
<span class="sd"> pre_sliced: bool</span>
<span class="sd"> Whether the labels are already sliced per device (default: False).</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of updating evaluation metric.</span>
<span class="sd"> >>> mod.forward(data_batch)</span>
<span class="sd"> >>> mod.update_metric(metric, data_batch.label)</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<span class="c1">################################################################################</span>
<span class="c1"># module setup</span>
<span class="c1">################################################################################</span>
<div class="viewcode-block" id="BaseModule.bind"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.bind">[docs]</a> <span class="k">def</span> <span class="nf">bind</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">data_shapes</span><span class="p">,</span> <span class="n">label_shapes</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span> <span class="n">for_training</span><span class="o">=</span><span class="kc">True</span><span class="p">,</span>
<span class="n">inputs_need_grad</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">force_rebind</span><span class="o">=</span><span class="kc">False</span><span class="p">,</span> <span class="n">shared_module</span><span class="o">=</span><span class="kc">None</span><span class="p">,</span>
<span class="n">grad_req</span><span class="o">=</span><span class="s1">'write'</span><span class="p">):</span>
<span class="sd">"""Binds the symbols to construct executors. This is necessary before one</span>
<span class="sd"> can perform computation with the module.</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> data_shapes : list of (str, tuple) or DataDesc objects</span>
<span class="sd"> Typically is ``data_iter.provide_data``. Can also be a list of</span>
<span class="sd"> (data name, data shape).</span>
<span class="sd"> label_shapes : list of (str, tuple) or DataDesc objects</span>
<span class="sd"> Typically is ``data_iter.provide_label``. Can also be a list of</span>
<span class="sd"> (label name, label shape).</span>
<span class="sd"> for_training : bool</span>
<span class="sd"> Default is ``True``. Whether the executors should be bind for training.</span>
<span class="sd"> inputs_need_grad : bool</span>
<span class="sd"> Default is ``False``. Whether the gradients to the input data need to be computed.</span>
<span class="sd"> Typically this is not needed. But this might be needed when implementing composition</span>
<span class="sd"> of modules.</span>
<span class="sd"> force_rebind : bool</span>
<span class="sd"> Default is ``False``. This function does nothing if the executors are already</span>
<span class="sd"> bound. But with this ``True``, the executors will be forced to rebind.</span>
<span class="sd"> shared_module : Module</span>
<span class="sd"> Default is ``None``. This is used in bucketing. When not ``None``, the shared module</span>
<span class="sd"> essentially corresponds to a different bucket -- a module with different symbol</span>
<span class="sd"> but with the same sets of parameters (e.g. unrolled RNNs with different lengths).</span>
<span class="sd"> grad_req : str, list of str, dict of str to str</span>
<span class="sd"> Requirement for gradient accumulation. Can be 'write', 'add', or 'null'</span>
<span class="sd"> (default to 'write').</span>
<span class="sd"> Can be specified globally (str) or for each argument (list, dict).</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of binding symbols.</span>
<span class="sd"> >>> mod.bind(data_shapes=[('data', (1, 10, 10))])</span>
<span class="sd"> >>> # Assume train_iter is already created.</span>
<span class="sd"> >>> mod.bind(data_shapes=train_iter.provide_data, label_shapes=train_iter.provide_label)</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<div class="viewcode-block" id="BaseModule.init_optimizer"><a class="viewcode-back" href="../../../api/python/module/module.html#mxnet.module.BaseModule.init_optimizer">[docs]</a> <span class="k">def</span> <span class="nf">init_optimizer</span><span class="p">(</span><span class="bp">self</span><span class="p">,</span> <span class="n">kvstore</span><span class="o">=</span><span class="s1">'local'</span><span class="p">,</span> <span class="n">optimizer</span><span class="o">=</span><span class="s1">'sgd'</span><span class="p">,</span>
<span class="n">optimizer_params</span><span class="o">=</span><span class="p">((</span><span class="s1">'learning_rate'</span><span class="p">,</span> <span class="mf">0.01</span><span class="p">),),</span> <span class="n">force_init</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
<span class="sd">"""Installs and initializes optimizers, as well as initialize kvstore for</span>
<span class="sd"> distributed training</span>
<span class="sd"> Parameters</span>
<span class="sd"> ----------</span>
<span class="sd"> kvstore : str or KVStore</span>
<span class="sd"> Defaults to `'local'`.</span>
<span class="sd"> optimizer : str or Optimizer</span>
<span class="sd"> Defaults to `'sgd'`.</span>
<span class="sd"> optimizer_params : dict</span>
<span class="sd"> Defaults to ``(('learning_rate', 0.01),)``. The default value is not a dictionary,</span>
<span class="sd"> just to avoid pylint warning of dangerous default values.</span>
<span class="sd"> force_init : bool</span>
<span class="sd"> Defaults to ``False``, indicates whether to force re-initializing an optimizer</span>
<span class="sd"> if it is already installed.</span>
<span class="sd"> Examples</span>
<span class="sd"> --------</span>
<span class="sd"> >>> # An example of initializing optimizer.</span>
<span class="sd"> >>> mod.init_optimizer(optimizer='sgd', optimizer_params=(('learning_rate', 0.005),))</span>
<span class="sd"> """</span>
<span class="k">raise</span> <span class="ne">NotImplementedError</span><span class="p">()</span></div>
<span class="c1">################################################################################</span>
<span class="c1"># misc</span>
<span class="c1">################################################################################</span>
<span class="nd">@property</span>
<span class="k">def</span> <span class="nf">symbol</span><span class="p">(</span><span class="bp">self</span><span class="p">):</span>
<span class="sd">"""Gets the symbol associated with this module.</span>
<span class="sd"> Except for `Module`, for other types of modules (e.g. `BucketingModule`), this</span>
<span class="sd"> property might not be a constant throughout its life time. Some modules might</span>
<span class="sd"> not even be associated with any symbols.</span>
<span class="sd"> """</span>
<span class="k">return</span> <span class="bp">self</span><span class="o">.</span><span class="n">_symbol</span></div>
</pre></div>
</div>
</div>
<div aria-label="main navigation" class="sphinxsidebar rightsidebar" role="navigation">
<div class="sphinxsidebarwrapper">
</div>
</div>
</div><div class="footer">
<div class="section-disclaimer">
<div class="container">
<div>
<img height="60" src="https://raw.githubusercontent.com/dmlc/web-data/master/mxnet/image/apache_incubator_logo.png"/>
<p>
Apache MXNet is an effort undergoing incubation at The Apache Software Foundation (ASF), <strong>sponsored by the <i>Apache Incubator</i></strong>. Incubation is required of all newly accepted projects until a further review indicates that the infrastructure, communications, and decision making process have stabilized in a manner consistent with other successful ASF projects. While incubation status is not necessarily a reflection of the completeness or stability of the code, it does indicate that the project has yet to be fully endorsed by the ASF.
</p>
<p>
"Copyright © 2017-2018, The Apache Software Foundation
Apache MXNet, MXNet, Apache, the Apache feather, and the Apache MXNet project logo are either registered trademarks or trademarks of the Apache Software Foundation."
</p>
</div>
</div>
</div>
</div> <!-- pagename != index -->
</div>
<script crossorigin="anonymous" integrity="sha384-0mSbJDEHialfmuBBQP6A4Qrprq5OVfW37PRR3j5ELqxss1yVqOtnepnHVP9aJ7xS" src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.6/js/bootstrap.min.js"></script>
<script src="../../../_static/js/sidebar.js" type="text/javascript"></script>
<script src="../../../_static/js/search.js" type="text/javascript"></script>
<script src="../../../_static/js/navbar.js" type="text/javascript"></script>
<script src="../../../_static/js/clipboard.min.js" type="text/javascript"></script>
<script src="../../../_static/js/copycode.js" type="text/javascript"></script>
<script src="../../../_static/js/page.js" type="text/javascript"></script>
<script src="../../../_static/js/docversion.js" type="text/javascript"></script>
<script type="text/javascript">
$('body').ready(function () {
$('body').css('visibility', 'visible');
});
</script>
</body>
</html>