Google Git
Sign in
apache / mxnet-site / 6054ba467ff2ccce12789e38fd5604ffad8abd17 / . / versions / 1.8.0 / api / python / docs / api / module / index.html
blob: 58ffd1cbb6b8533ec9d3450604ed3450cbb9c0cf [file] [log] [blame]
<!DOCTYPE html>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta charset="utf-8" />
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no">
<meta http-equiv="x-ua-compatible" content="ie=edge">
<style>
.dropdown {
position: relative;
display: inline-block;
}
.dropdown-content {
display: none;
position: absolute;
background-color: #f9f9f9;
min-width: 160px;
box-shadow: 0px 8px 16px 0px rgba(0,0,0,0.2);
padding: 12px 16px;
z-index: 1;
text-align: left;
}
.dropdown:hover .dropdown-content {
display: block;
}
.dropdown-option:hover {
color: #FF4500;
}
.dropdown-option-active {
color: #FF4500;
font-weight: lighter;
}
.dropdown-option {
color: #000000;
font-weight: lighter;
}
.dropdown-header {
color: #FFFFFF;
display: inline-flex;
}
.dropdown-caret {
width: 18px;
}
.dropdown-caret-path {
fill: #FFFFFF;
}
</style>
<title>mxnet.module &#8212; Apache MXNet documentation</title>
<link rel="stylesheet" href="../../_static/basic.css" type="text/css" />
<link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
<link rel="stylesheet" type="text/css" href="../../_static/mxnet.css" />
<link rel="stylesheet" href="../../_static/material-design-lite-1.3.0/material.blue-deep_orange.min.css" type="text/css" />
<link rel="stylesheet" href="../../_static/sphinx_materialdesign_theme.css" type="text/css" />
<link rel="stylesheet" href="../../_static/fontawesome/all.css" type="text/css" />
<link rel="stylesheet" href="../../_static/fonts.css" type="text/css" />
<link rel="stylesheet" href="../../_static/feedback.css" type="text/css" />
<script id="documentation_options" data-url_root="../../" src="../../_static/documentation_options.js"></script>
<script src="../../_static/jquery.js"></script>
<script src="../../_static/underscore.js"></script>
<script src="../../_static/doctools.js"></script>
<script src="../../_static/language_data.js"></script>
<script src="../../_static/google_analytics.js"></script>
<script src="../../_static/autodoc.js"></script>
<script crossorigin="anonymous" integrity="sha256-Ae2Vz/4ePdIu6ZyI/5ZGsYnb+m0JlOmKPjt6XZ9JJkA=" src="https://cdnjs.cloudflare.com/ajax/libs/require.js/2.3.4/require.min.js"></script>
<script async="async" src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.5/latest.js?config=TeX-AMS-MML_HTMLorMML"></script>
<script type="text/x-mathjax-config">MathJax.Hub.Config({"tex2jax": {"inlineMath": [["$", "$"], ["\\(", "\\)"]], "processEscapes": true, "ignoreClass": "document", "processClass": "math|output_area"}})</script>
<link rel="shortcut icon" href="../../_static/mxnet-icon.png"/>
<link rel="index" title="Index" href="../../genindex.html" />
<link rel="search" title="Search" href="../../search.html" />
<link rel="next" title="mxnet.contrib" href="../contrib/index.html" />
<link rel="prev" title="symbol.sparse" href="../symbol/sparse/index.html" />
</head>
<body><header class="site-header" role="banner">
<div class="wrapper">
<a class="site-title" rel="author" href="/versions/1.8.0/"><img
src="../../_static/mxnet_logo.png" class="site-header-logo"></a>
<nav class="site-nav">
<input type="checkbox" id="nav-trigger" class="nav-trigger"/>
<label for="nav-trigger">
<span class="menu-icon">
<svg viewBox="0 0 18 15" width="18px" height="15px">
<path d="M18,1.484c0,0.82-0.665,1.484-1.484,1.484H1.484C0.665,2.969,0,2.304,0,1.484l0,0C0,0.665,0.665,0,1.484,0 h15.032C17.335,0,18,0.665,18,1.484L18,1.484z M18,7.516C18,8.335,17.335,9,16.516,9H1.484C0.665,9,0,8.335,0,7.516l0,0 c0-0.82,0.665-1.484,1.484-1.484h15.032C17.335,6.031,18,6.696,18,7.516L18,7.516z M18,13.516C18,14.335,17.335,15,16.516,15H1.484 C0.665,15,0,14.335,0,13.516l0,0c0-0.82,0.665-1.483,1.484-1.483h15.032C17.335,12.031,18,12.695,18,13.516L18,13.516z"/>
</svg>
</span>
</label>
<div class="trigger">
<a class="page-link" href="/versions/1.8.0/get_started">Get Started</a>
<a class="page-link" href="/versions/1.8.0/blog">Blog</a>
<a class="page-link" href="/versions/1.8.0/features">Features</a>
<a class="page-link" href="/versions/1.8.0/ecosystem">Ecosystem</a>
<a class="page-link page-current" href="/versions/1.8.0/api">Docs & Tutorials</a>
<a class="page-link" href="https://github.com/apache/incubator-mxnet">GitHub</a>
<div class="dropdown">
<span class="dropdown-header">1.8.0
<svg class="dropdown-caret" viewBox="0 0 32 32" class="icon icon-caret-bottom" aria-hidden="true"><path class="dropdown-caret-path" d="M24 11.305l-7.997 11.39L8 11.305z"></path></svg>
</span>
<div class="dropdown-content">
<a class="dropdown-option" href="/">master</a><br>
<a class="dropdown-option-active" href="/versions/1.8.0/">1.8.0</a><br>
<a class="dropdown-option" href="/versions/1.7.0/">1.7.0</a><br>
<a class="dropdown-option" href="/versions/1.6.0/">1.6.0</a><br>
<a class="dropdown-option" href="/versions/1.5.0/">1.5.0</a><br>
<a class="dropdown-option" href="/versions/1.4.1/">1.4.1</a><br>
<a class="dropdown-option" href="/versions/1.3.1/">1.3.1</a><br>
<a class="dropdown-option" href="/versions/1.2.1/">1.2.1</a><br>
<a class="dropdown-option" href="/versions/1.1.0/">1.1.0</a><br>
<a class="dropdown-option" href="/versions/1.0.0/">1.0.0</a><br>
<a class="dropdown-option" href="/versions/0.12.1/">0.12.1</a><br>
<a class="dropdown-option" href="/versions/0.11.0/">0.11.0</a>
</div>
</div>
</div>
</nav>
</div>
</header>
<div class="mdl-layout mdl-js-layout mdl-layout--fixed-header mdl-layout--fixed-drawer"><header class="mdl-layout__header mdl-layout__header--waterfall ">
<div class="mdl-layout__header-row">
<nav class="mdl-navigation breadcrumb">
<a class="mdl-navigation__link" href="../index.html">Python API</a><i class="material-icons">navigate_next</i>
<a class="mdl-navigation__link is-active">mxnet.module</a>
</nav>
<div class="mdl-layout-spacer"></div>
<nav class="mdl-navigation">
<form class="form-inline pull-sm-right" action="../../search.html" method="get">
<div class="mdl-textfield mdl-js-textfield mdl-textfield--expandable mdl-textfield--floating-label mdl-textfield--align-right">
<label id="quick-search-icon" class="mdl-button mdl-js-button mdl-button--icon" for="waterfall-exp">
<i class="material-icons">search</i>
</label>
<div class="mdl-textfield__expandable-holder">
<input class="mdl-textfield__input" type="text" name="q" id="waterfall-exp" placeholder="Search" />
<input type="hidden" name="check_keywords" value="yes" />
<input type="hidden" name="area" value="default" />
</div>
</div>
<div class="mdl-tooltip" data-mdl-for="quick-search-icon">
Quick search
</div>
</form>
<a id="button-show-source"
class="mdl-button mdl-js-button mdl-button--icon"
href="../../_sources/api/module/index.rst" rel="nofollow">
<i class="material-icons">code</i>
</a>
<div class="mdl-tooltip" data-mdl-for="button-show-source">
Show Source
</div>
</nav>
</div>
<div class="mdl-layout__header-row header-links">
<div class="mdl-layout-spacer"></div>
<nav class="mdl-navigation">
</nav>
</div>
</header><header class="mdl-layout__drawer">
<div class="globaltoc">
<span class="mdl-layout-title toc">Table Of Contents</span>
<nav class="mdl-navigation">
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../../tutorials/index.html">Python Tutorials</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/getting-started/index.html">Getting Started</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/crash-course/index.html">Crash Course</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/1-ndarray.html">Manipulate data with <code class="docutils literal notranslate"><span class="pre">ndarray</span></code></a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/2-nn.html">Create a neural network</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/3-autograd.html">Automatic differentiation with <code class="docutils literal notranslate"><span class="pre">autograd</span></code></a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/4-train.html">Train the neural network</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/5-predict.html">Predict with a pre-trained model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/6-use_gpus.html">Use GPUs</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/to-mxnet/index.html">Moving to MXNet from Other Frameworks</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/to-mxnet/pytorch.html">PyTorch vs Apache MXNet</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/gluon_from_experiment_to_deployment.html">Gluon: from experiment to deployment</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/logistic_regression_explained.html">Logistic regression explained</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/image/mnist.html">MNIST</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/packages/index.html">Packages</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/autograd/index.html">Automatic Differentiation</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/gluon/index.html">Gluon</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/index.html">Blocks</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/custom-layer.html">Custom Layers</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/custom_layer_beginners.html">Customer Layers (Beginners)</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/hybridize.html">Hybridize</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/init.html">Initialization</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/naming.html">Parameter and Block Naming</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/nn.html">Layers and Blocks</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/parameters.html">Parameter Management</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/save_load_params.html">Saving and Loading Gluon Models</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/activations/activations.html">Activation Blocks</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/data/index.html">Data Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html">Image Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Spatial-Augmentation">Spatial Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Color-Augmentation">Color Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Composed-Augmentations">Composed Augmentations</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html">Gluon <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s and <code class="docutils literal notranslate"><span class="pre">DataLoader</span></code></a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Using-own-data-with-included-Datasets">Using own data with included <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Using-own-data-with-custom-Datasets">Using own data with custom <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Appendix:-Upgrading-from-Module-DataIter-to-Gluon-DataLoader">Appendix: Upgrading from Module <code class="docutils literal notranslate"><span class="pre">DataIter</span></code> to Gluon <code class="docutils literal notranslate"><span class="pre">DataLoader</span></code></a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/image/index.html">Image Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/image-augmentation.html">Image Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/info_gan.html">Image similarity search with InfoGAN</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/mnist.html">Handwritten Digit Recognition</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/pretrained_models.html">Using pre-trained models in MXNet</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/loss/index.html">Losses</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/custom-loss.html">Custom Loss Blocks</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/kl_divergence.html">Kullback-Leibler (KL) Divergence</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/loss.html">Loss functions</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/text/index.html">Text Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/text/gnmt.html">Google Neural Machine Translation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/text/transformer.html">Machine Translation with Transformer</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/training/index.html">Training</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/fit_api_tutorial.html">MXNet Gluon Fit API</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/trainer.html">Trainer</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/index.html">Learning Rates</a><ul>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_finder.html">Learning Rate Finder</a></li>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_schedules.html">Learning Rate Schedules</a></li>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_schedules_advanced.html">Advanced Learning Rate Schedules</a></li>
</ul>
</li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/normalization/index.html">Normalization Blocks</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/kvstore/index.html">KVStore</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/kvstore/kvstore.html">Distributed Key-Value Store</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/ndarray/index.html">NDArray</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/01-ndarray-intro.html">An Intro: Manipulate Data the MXNet Way with NDArray</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/02-ndarray-operations.html">NDArray Operations</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/03-ndarray-contexts.html">NDArray Contexts</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/gotchas_numpy_in_mxnet.html">Gotchas using NumPy in Apache MXNet</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/index.html">Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/csr.html">CSRNDArray - NDArray in Compressed Sparse Row Storage Format</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/row_sparse.html">RowSparseNDArray - NDArray for Sparse Gradient Updates</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/train.html">Train a Linear Regression Model with Sparse Symbols</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/train_gluon.html">Sparse NDArrays with Gluon</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/onnx/index.html">ONNX</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/fine_tuning_gluon.html">Fine-tuning an ONNX model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/inference_on_onnx_model.html">Running inference on MXNet/Gluon from an ONNX model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/super_resolution.html">Importing an ONNX model into MXNet</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/deploy/export/onnx.html">Export ONNX Models</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/optimizer/index.html">Optimizers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/viz/index.html">Visualization</a><ul>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/visualize_graph">Visualize networks</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/performance/index.html">Performance</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/performance/compression/index.html">Compression</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/compression/int8.html">Deploy with int-8</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/float16">Float16</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/gradient_compression">Gradient Compression</a></li>
<li class="toctree-l4"><a class="reference external" href="https://gluon-cv.mxnet.io/build/examples_deployment/int8_inference.html">GluonCV with Quantized Models</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/performance/backend/index.html">Accelerated Backend Tools</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/index.html">Intel MKL-DNN</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/mkldnn_quantization.html">Quantize with MKL-DNN backend</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/mkldnn_readme.html">Install MXNet with MKL-DNN</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/tensorrt/index.html">TensorRT</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/tensorrt/tensorrt.html">Optimizing Deep Learning Computation Graphs with TensorRT</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/tvm.html">Use TVM</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/profiler.html">Profiling MXNet Models</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/amp.html">Using AMP: Automatic Mixed Precision</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/deploy/index.html">Deployment</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/export/index.html">Export</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/export/onnx.html">Exporting to ONNX format</a></li>
<li class="toctree-l4"><a class="reference external" href="https://gluon-cv.mxnet.io/build/examples_deployment/export_network.html">Export Gluon CV Models</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/blocks/save_load_params.html">Save / Load Parameters</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/inference/index.html">Inference</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/cpp.html">Deploy into C++</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/image_classification_jetson.html">Image Classication using pretrained ResNet-50 model on Jetson module</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/scala.html">Deploy into a Java or Scala Environment</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/wine_detector.html">Real-time Object Detection with MXNet On The Raspberry Pi</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/index.html">Run on AWS</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/use_ec2.html">Run on an EC2 Instance</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/use_sagemaker.html">Run on Amazon SageMaker</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/cloud.html">MXNet on the Cloud</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/extend/index.html">Extend</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/extend/custom_layer.html">Custom Layers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/extend/customop.html">Custom Numpy Operators</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/faq/new_op">New Operator Creation</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/faq/add_op_in_backend">New Operator in MXNet Backend</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l1 current"><a class="reference internal" href="../index.html">Python API</a><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="../ndarray/index.html">mxnet.ndarray</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/ndarray.html">ndarray</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/contrib/index.html">ndarray.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/image/index.html">ndarray.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/linalg/index.html">ndarray.linalg</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/op/index.html">ndarray.op</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/random/index.html">ndarray.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/register/index.html">ndarray.register</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/sparse/index.html">ndarray.sparse</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/utils/index.html">ndarray.utils</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../gluon/index.html">mxnet.gluon</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../gluon/block.html">gluon.Block</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/hybrid_block.html">gluon.HybridBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/symbol_block.html">gluon.SymbolBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/constant.html">gluon.Constant</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/parameter.html">gluon.Parameter</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/parameter_dict.html">gluon.ParameterDict</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/trainer.html">gluon.Trainer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/contrib/index.html">gluon.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/data/index.html">gluon.data</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../gluon/data/vision/index.html">data.vision</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../gluon/data/vision/datasets/index.html">vision.datasets</a></li>
<li class="toctree-l5"><a class="reference internal" href="../gluon/data/vision/transforms/index.html">vision.transforms</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/loss/index.html">gluon.loss</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/model_zoo/index.html">gluon.model_zoo.vision</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/nn/index.html">gluon.nn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/rnn/index.html">gluon.rnn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/utils/index.html">gluon.utils</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../autograd/index.html">mxnet.autograd</a></li>
<li class="toctree-l2"><a class="reference internal" href="../initializer/index.html">mxnet.initializer</a></li>
<li class="toctree-l2"><a class="reference internal" href="../optimizer/index.html">mxnet.optimizer</a></li>
<li class="toctree-l2"><a class="reference internal" href="../lr_scheduler/index.html">mxnet.lr_scheduler</a></li>
<li class="toctree-l2"><a class="reference internal" href="../metric/index.html">mxnet.metric</a></li>
<li class="toctree-l2"><a class="reference internal" href="../kvstore/index.html">mxnet.kvstore</a></li>
<li class="toctree-l2"><a class="reference internal" href="../symbol/index.html">mxnet.symbol</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../symbol/symbol.html">symbol</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/contrib/index.html">symbol.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/image/index.html">symbol.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/linalg/index.html">symbol.linalg</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/op/index.html">symbol.op</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/random/index.html">symbol.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/register/index.html">symbol.register</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/sparse/index.html">symbol.sparse</a></li>
</ul>
</li>
<li class="toctree-l2 current"><a class="current reference internal" href="#">mxnet.module</a></li>
<li class="toctree-l2"><a class="reference internal" href="../contrib/index.html">mxnet.contrib</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../contrib/autograd/index.html">contrib.autograd</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/io/index.html">contrib.io</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/ndarray/index.html">contrib.ndarray</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/onnx/index.html">contrib.onnx</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/quantization/index.html">contrib.quantization</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/symbol/index.html">contrib.symbol</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/tensorboard/index.html">contrib.tensorboard</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/tensorrt/index.html">contrib.tensorrt</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/text/index.html">contrib.text</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../mxnet/index.html">mxnet</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/attribute/index.html">mxnet.attribute</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/base/index.html">mxnet.base</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/callback/index.html">mxnet.callback</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/context/index.html">mxnet.context</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/engine/index.html">mxnet.engine</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/executor/index.html">mxnet.executor</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/executor_manager/index.html">mxnet.executor_manager</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/image/index.html">mxnet.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/io/index.html">mxnet.io</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/kvstore_server/index.html">mxnet.kvstore_server</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/libinfo/index.html">mxnet.libinfo</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/log/index.html">mxnet.log</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/model/index.html">mxnet.model</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/monitor/index.html">mxnet.monitor</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/name/index.html">mxnet.name</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/notebook/index.html">mxnet.notebook</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/operator/index.html">mxnet.operator</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/profiler/index.html">mxnet.profiler</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/random/index.html">mxnet.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/recordio/index.html">mxnet.recordio</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/registry/index.html">mxnet.registry</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/rtc/index.html">mxnet.rtc</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/runtime/index.html">mxnet.runtime</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/test_utils/index.html">mxnet.test_utils</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/torch/index.html">mxnet.torch</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/util/index.html">mxnet.util</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/visualization/index.html">mxnet.visualization</a></li>
</ul>
</li>
</ul>
</li>
</ul>
</nav>
</div>
</header>
<main class="mdl-layout__content" tabIndex="0">
<script type="text/javascript" src="../../_static/sphinx_materialdesign_theme.js "></script>
<script type="text/javascript" src="../../_static/feedback.js"></script>
<header class="mdl-layout__drawer">
<div class="globaltoc">
<span class="mdl-layout-title toc">Table Of Contents</span>
<nav class="mdl-navigation">
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../../tutorials/index.html">Python Tutorials</a><ul>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/getting-started/index.html">Getting Started</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/crash-course/index.html">Crash Course</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/1-ndarray.html">Manipulate data with <code class="docutils literal notranslate"><span class="pre">ndarray</span></code></a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/2-nn.html">Create a neural network</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/3-autograd.html">Automatic differentiation with <code class="docutils literal notranslate"><span class="pre">autograd</span></code></a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/4-train.html">Train the neural network</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/5-predict.html">Predict with a pre-trained model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/crash-course/6-use_gpus.html">Use GPUs</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/to-mxnet/index.html">Moving to MXNet from Other Frameworks</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/getting-started/to-mxnet/pytorch.html">PyTorch vs Apache MXNet</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/gluon_from_experiment_to_deployment.html">Gluon: from experiment to deployment</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/getting-started/logistic_regression_explained.html">Logistic regression explained</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/image/mnist.html">MNIST</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/packages/index.html">Packages</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/autograd/index.html">Automatic Differentiation</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/gluon/index.html">Gluon</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/index.html">Blocks</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/custom-layer.html">Custom Layers</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/custom_layer_beginners.html">Customer Layers (Beginners)</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/hybridize.html">Hybridize</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/init.html">Initialization</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/naming.html">Parameter and Block Naming</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/nn.html">Layers and Blocks</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/parameters.html">Parameter Management</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/save_load_params.html">Saving and Loading Gluon Models</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/blocks/activations/activations.html">Activation Blocks</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/data/index.html">Data Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html">Image Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Spatial-Augmentation">Spatial Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Color-Augmentation">Color Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/data_augmentation.html#Composed-Augmentations">Composed Augmentations</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html">Gluon <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s and <code class="docutils literal notranslate"><span class="pre">DataLoader</span></code></a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Using-own-data-with-included-Datasets">Using own data with included <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Using-own-data-with-custom-Datasets">Using own data with custom <code class="docutils literal notranslate"><span class="pre">Dataset</span></code>s</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/data/datasets.html#Appendix:-Upgrading-from-Module-DataIter-to-Gluon-DataLoader">Appendix: Upgrading from Module <code class="docutils literal notranslate"><span class="pre">DataIter</span></code> to Gluon <code class="docutils literal notranslate"><span class="pre">DataLoader</span></code></a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/image/index.html">Image Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/image-augmentation.html">Image Augmentation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/info_gan.html">Image similarity search with InfoGAN</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/mnist.html">Handwritten Digit Recognition</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/image/pretrained_models.html">Using pre-trained models in MXNet</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/loss/index.html">Losses</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/custom-loss.html">Custom Loss Blocks</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/kl_divergence.html">Kullback-Leibler (KL) Divergence</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/loss/loss.html">Loss functions</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/text/index.html">Text Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/text/gnmt.html">Google Neural Machine Translation</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/text/transformer.html">Machine Translation with Transformer</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/gluon/training/index.html">Training</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/fit_api_tutorial.html">MXNet Gluon Fit API</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/trainer.html">Trainer</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/index.html">Learning Rates</a><ul>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_finder.html">Learning Rate Finder</a></li>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_schedules.html">Learning Rate Schedules</a></li>
<li class="toctree-l6"><a class="reference internal" href="../../tutorials/packages/gluon/training/learning_rates/learning_rate_schedules_advanced.html">Advanced Learning Rate Schedules</a></li>
</ul>
</li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/gluon/training/normalization/index.html">Normalization Blocks</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/kvstore/index.html">KVStore</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/kvstore/kvstore.html">Distributed Key-Value Store</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/ndarray/index.html">NDArray</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/01-ndarray-intro.html">An Intro: Manipulate Data the MXNet Way with NDArray</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/02-ndarray-operations.html">NDArray Operations</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/03-ndarray-contexts.html">NDArray Contexts</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/gotchas_numpy_in_mxnet.html">Gotchas using NumPy in Apache MXNet</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/index.html">Tutorials</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/csr.html">CSRNDArray - NDArray in Compressed Sparse Row Storage Format</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/row_sparse.html">RowSparseNDArray - NDArray for Sparse Gradient Updates</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/train.html">Train a Linear Regression Model with Sparse Symbols</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/packages/ndarray/sparse/train_gluon.html">Sparse NDArrays with Gluon</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/onnx/index.html">ONNX</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/fine_tuning_gluon.html">Fine-tuning an ONNX model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/inference_on_onnx_model.html">Running inference on MXNet/Gluon from an ONNX model</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/packages/onnx/super_resolution.html">Importing an ONNX model into MXNet</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/deploy/export/onnx.html">Export ONNX Models</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/optimizer/index.html">Optimizers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/packages/viz/index.html">Visualization</a><ul>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/visualize_graph">Visualize networks</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/performance/index.html">Performance</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/performance/compression/index.html">Compression</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/compression/int8.html">Deploy with int-8</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/float16">Float16</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/faq/gradient_compression">Gradient Compression</a></li>
<li class="toctree-l4"><a class="reference external" href="https://gluon-cv.mxnet.io/build/examples_deployment/int8_inference.html">GluonCV with Quantized Models</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/performance/backend/index.html">Accelerated Backend Tools</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/index.html">Intel MKL-DNN</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/mkldnn_quantization.html">Quantize with MKL-DNN backend</a></li>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/mkldnn/mkldnn_readme.html">Install MXNet with MKL-DNN</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/tensorrt/index.html">TensorRT</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../../tutorials/performance/backend/tensorrt/tensorrt.html">Optimizing Deep Learning Computation Graphs with TensorRT</a></li>
</ul>
</li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/tvm.html">Use TVM</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/profiler.html">Profiling MXNet Models</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/performance/backend/amp.html">Using AMP: Automatic Mixed Precision</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/deploy/index.html">Deployment</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/export/index.html">Export</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/export/onnx.html">Exporting to ONNX format</a></li>
<li class="toctree-l4"><a class="reference external" href="https://gluon-cv.mxnet.io/build/examples_deployment/export_network.html">Export Gluon CV Models</a></li>
<li class="toctree-l4"><a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/blocks/save_load_params.html">Save / Load Parameters</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/inference/index.html">Inference</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/cpp.html">Deploy into C++</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/image_classification_jetson.html">Image Classication using pretrained ResNet-50 model on Jetson module</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/scala.html">Deploy into a Java or Scala Environment</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/inference/wine_detector.html">Real-time Object Detection with MXNet On The Raspberry Pi</a></li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/index.html">Run on AWS</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/use_ec2.html">Run on an EC2 Instance</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/use_sagemaker.html">Run on Amazon SageMaker</a></li>
<li class="toctree-l4"><a class="reference internal" href="../../tutorials/deploy/run-on-aws/cloud.html">MXNet on the Cloud</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../../tutorials/extend/index.html">Extend</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/extend/custom_layer.html">Custom Layers</a></li>
<li class="toctree-l3"><a class="reference internal" href="../../tutorials/extend/customop.html">Custom Numpy Operators</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/faq/new_op">New Operator Creation</a></li>
<li class="toctree-l3"><a class="reference external" href="https://mxnet.apache.org/api/faq/add_op_in_backend">New Operator in MXNet Backend</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l1 current"><a class="reference internal" href="../index.html">Python API</a><ul class="current">
<li class="toctree-l2"><a class="reference internal" href="../ndarray/index.html">mxnet.ndarray</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/ndarray.html">ndarray</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/contrib/index.html">ndarray.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/image/index.html">ndarray.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/linalg/index.html">ndarray.linalg</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/op/index.html">ndarray.op</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/random/index.html">ndarray.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/register/index.html">ndarray.register</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/sparse/index.html">ndarray.sparse</a></li>
<li class="toctree-l3"><a class="reference internal" href="../ndarray/utils/index.html">ndarray.utils</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../gluon/index.html">mxnet.gluon</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../gluon/block.html">gluon.Block</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/hybrid_block.html">gluon.HybridBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/symbol_block.html">gluon.SymbolBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/constant.html">gluon.Constant</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/parameter.html">gluon.Parameter</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/parameter_dict.html">gluon.ParameterDict</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/trainer.html">gluon.Trainer</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/contrib/index.html">gluon.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/data/index.html">gluon.data</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../gluon/data/vision/index.html">data.vision</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../gluon/data/vision/datasets/index.html">vision.datasets</a></li>
<li class="toctree-l5"><a class="reference internal" href="../gluon/data/vision/transforms/index.html">vision.transforms</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/loss/index.html">gluon.loss</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/model_zoo/index.html">gluon.model_zoo.vision</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/nn/index.html">gluon.nn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/rnn/index.html">gluon.rnn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../gluon/utils/index.html">gluon.utils</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../autograd/index.html">mxnet.autograd</a></li>
<li class="toctree-l2"><a class="reference internal" href="../initializer/index.html">mxnet.initializer</a></li>
<li class="toctree-l2"><a class="reference internal" href="../optimizer/index.html">mxnet.optimizer</a></li>
<li class="toctree-l2"><a class="reference internal" href="../lr_scheduler/index.html">mxnet.lr_scheduler</a></li>
<li class="toctree-l2"><a class="reference internal" href="../metric/index.html">mxnet.metric</a></li>
<li class="toctree-l2"><a class="reference internal" href="../kvstore/index.html">mxnet.kvstore</a></li>
<li class="toctree-l2"><a class="reference internal" href="../symbol/index.html">mxnet.symbol</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../symbol/symbol.html">symbol</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/contrib/index.html">symbol.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/image/index.html">symbol.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/linalg/index.html">symbol.linalg</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/op/index.html">symbol.op</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/random/index.html">symbol.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/register/index.html">symbol.register</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol/sparse/index.html">symbol.sparse</a></li>
</ul>
</li>
<li class="toctree-l2 current"><a class="current reference internal" href="#">mxnet.module</a></li>
<li class="toctree-l2"><a class="reference internal" href="../contrib/index.html">mxnet.contrib</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../contrib/autograd/index.html">contrib.autograd</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/io/index.html">contrib.io</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/ndarray/index.html">contrib.ndarray</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/onnx/index.html">contrib.onnx</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/quantization/index.html">contrib.quantization</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/symbol/index.html">contrib.symbol</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/tensorboard/index.html">contrib.tensorboard</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/tensorrt/index.html">contrib.tensorrt</a></li>
<li class="toctree-l3"><a class="reference internal" href="../contrib/text/index.html">contrib.text</a></li>
</ul>
</li>
<li class="toctree-l2"><a class="reference internal" href="../mxnet/index.html">mxnet</a><ul>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/attribute/index.html">mxnet.attribute</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/base/index.html">mxnet.base</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/callback/index.html">mxnet.callback</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/context/index.html">mxnet.context</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/engine/index.html">mxnet.engine</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/executor/index.html">mxnet.executor</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/executor_manager/index.html">mxnet.executor_manager</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/image/index.html">mxnet.image</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/io/index.html">mxnet.io</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/kvstore_server/index.html">mxnet.kvstore_server</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/libinfo/index.html">mxnet.libinfo</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/log/index.html">mxnet.log</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/model/index.html">mxnet.model</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/monitor/index.html">mxnet.monitor</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/name/index.html">mxnet.name</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/notebook/index.html">mxnet.notebook</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/operator/index.html">mxnet.operator</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/profiler/index.html">mxnet.profiler</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/random/index.html">mxnet.random</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/recordio/index.html">mxnet.recordio</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/registry/index.html">mxnet.registry</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/rtc/index.html">mxnet.rtc</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/runtime/index.html">mxnet.runtime</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/test_utils/index.html">mxnet.test_utils</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/torch/index.html">mxnet.torch</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/util/index.html">mxnet.util</a></li>
<li class="toctree-l3"><a class="reference internal" href="../mxnet/visualization/index.html">mxnet.visualization</a></li>
</ul>
</li>
</ul>
</li>
</ul>
</nav>
</div>
</header>
<div class="document">
<div class="page-content" role="main">
<div class="section" id="module-mxnet.module">
<span id="mxnet-module"></span><h1>mxnet.module<a class="headerlink" href="#module-mxnet.module" title="Permalink to this headline"></a></h1>
<p>A module is like a FeedForward model. But we would like to make it
easier to compose, similar to Torch modules.</p>
<p><strong>Classes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule" title="mxnet.module.BaseModule"><code class="xref py py-obj docutils literal notranslate"><span class="pre">BaseModule</span></code></a>([logger])</p></td>
<td><p>The base class of a module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule" title="mxnet.module.BucketingModule"><code class="xref py py-obj docutils literal notranslate"><span class="pre">BucketingModule</span></code></a>(sym_gen[, …])</p></td>
<td><p>This module helps to deal efficiently with varying-length inputs.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Module</span></code></a>(symbol[, data_names, label_names, …])</p></td>
<td><p>Module is a basic module that wrap a <cite>Symbol</cite>.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule" title="mxnet.module.PythonLossModule"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PythonLossModule</span></code></a>([name, data_names, …])</p></td>
<td><p>A convenient module class that implements many of the module APIs as empty functions.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule" title="mxnet.module.PythonModule"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PythonModule</span></code></a>(data_names, label_names, …[, …])</p></td>
<td><p>A convenient module class that implements many of the module APIs as empty functions.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule" title="mxnet.module.SequentialModule"><code class="xref py py-obj docutils literal notranslate"><span class="pre">SequentialModule</span></code></a>([logger])</p></td>
<td><p>A SequentialModule is a container module that can chain multiple modules together.</p></td>
</tr>
</tbody>
</table>
<dl class="class">
<dt id="mxnet.module.BaseModule">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">BaseModule</code><span class="sig-paren">(</span><em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">object</span></code></p>
<p>The base class of a module.</p>
<p>A module represents a computation component. One can think of module as a computation machine.
A module can execute forward and backward passes and update parameters in a model.
We aim to make the APIs easy to use, especially in the case when we need to use the imperative
API to work with multiple modules (e.g. stochastic depth network).</p>
<p>A module has several states:</p>
<ul class="simple">
<li><p>Initial state: Memory is not allocated yet, so the module is not ready for computation yet.</p></li>
<li><p>Binded: Shapes for inputs, outputs, and parameters are all known, memory has been allocated,
and the module is ready for computation.</p></li>
<li><p>Parameters are initialized: For modules with parameters, doing computation before
initializing the parameters might result in undefined outputs.</p></li>
<li><p>Optimizer is installed: An optimizer can be installed to a module. After this, the parameters
of the module can be updated according to the optimizer after gradients are computed
(forward-backward).</p></li>
</ul>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.backward" title="mxnet.module.BaseModule.backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">backward</span></code></a>([out_grads])</p></td>
<td><p>Backward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.bind" title="mxnet.module.BaseModule.bind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">bind</span></code></a>(data_shapes[, label_shapes, …])</p></td>
<td><p>Binds the symbols to construct executors.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.fit" title="mxnet.module.BaseModule.fit"><code class="xref py py-obj docutils literal notranslate"><span class="pre">fit</span></code></a>(train_data[, eval_data, eval_metric, …])</p></td>
<td><p>Trains the module parameters.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.forward" title="mxnet.module.BaseModule.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(data_batch[, is_train])</p></td>
<td><p>Forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.forward_backward" title="mxnet.module.BaseModule.forward_backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward_backward</span></code></a>(data_batch)</p></td>
<td><p>A convenient function that calls both <code class="docutils literal notranslate"><span class="pre">forward</span></code> and <code class="docutils literal notranslate"><span class="pre">backward</span></code>.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.get_input_grads" title="mxnet.module.BaseModule.get_input_grads"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_input_grads</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets the gradients to the inputs, computed in the previous backward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.get_outputs" title="mxnet.module.BaseModule.get_outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_outputs</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets outputs of the previous forward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.get_params" title="mxnet.module.BaseModule.get_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_params</span></code></a>()</p></td>
<td><p>Gets parameters, those are potentially copies of the actual parameters used to do computation on the device.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.get_states" title="mxnet.module.BaseModule.get_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_states</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets states from all devices</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.init_optimizer" title="mxnet.module.BaseModule.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>([kvstore, optimizer, …])</p></td>
<td><p>Installs and initializes optimizers, as well as initialize kvstore for</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.init_params" title="mxnet.module.BaseModule.init_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_params</span></code></a>([initializer, arg_params, …])</p></td>
<td><p>Initializes the parameters and auxiliary states.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.install_monitor" title="mxnet.module.BaseModule.install_monitor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">install_monitor</span></code></a>(mon)</p></td>
<td><p>Installs monitor on all executors.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.iter_predict" title="mxnet.module.BaseModule.iter_predict"><code class="xref py py-obj docutils literal notranslate"><span class="pre">iter_predict</span></code></a>(eval_data[, num_batch, reset, …])</p></td>
<td><p>Iterates over predictions.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.load_params" title="mxnet.module.BaseModule.load_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_params</span></code></a>(fname)</p></td>
<td><p>Loads model parameters from file.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.predict" title="mxnet.module.BaseModule.predict"><code class="xref py py-obj docutils literal notranslate"><span class="pre">predict</span></code></a>(eval_data[, num_batch, …])</p></td>
<td><p>Runs prediction and collects the outputs.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.prepare" title="mxnet.module.BaseModule.prepare"><code class="xref py py-obj docutils literal notranslate"><span class="pre">prepare</span></code></a>(data_batch[, sparse_row_id_fn])</p></td>
<td><p>Prepares the module for processing a data batch.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.save_params" title="mxnet.module.BaseModule.save_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save_params</span></code></a>(fname)</p></td>
<td><p>Saves model parameters to file.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.score" title="mxnet.module.BaseModule.score"><code class="xref py py-obj docutils literal notranslate"><span class="pre">score</span></code></a>(eval_data, eval_metric[, num_batch, …])</p></td>
<td><p>Runs prediction on <code class="docutils literal notranslate"><span class="pre">eval_data</span></code> and evaluates the performance according to the given <code class="docutils literal notranslate"><span class="pre">eval_metric</span></code>.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.set_params" title="mxnet.module.BaseModule.set_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_params</span></code></a>(arg_params, aux_params[, …])</p></td>
<td><p>Assigns parameter and aux state values.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.set_states" title="mxnet.module.BaseModule.set_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_states</span></code></a>([states, value])</p></td>
<td><p>Sets value for states.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.update" title="mxnet.module.BaseModule.update"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update</span></code></a>()</p></td>
<td><p>Updates parameters according to the installed optimizer and the gradients computed in the previous forward-backward batch.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.update_metric" title="mxnet.module.BaseModule.update_metric"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_metric</span></code></a>(eval_metric, labels[, pre_sliced])</p></td>
<td><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p></td>
</tr>
</tbody>
</table>
<p><strong>Attributes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.data_names" title="mxnet.module.BaseModule.data_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_names</span></code></a></p></td>
<td><p>A list of names for data required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.data_shapes" title="mxnet.module.BaseModule.data_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the data inputs to this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.label_shapes" title="mxnet.module.BaseModule.label_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the label inputs to this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.output_names" title="mxnet.module.BaseModule.output_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_names</span></code></a></p></td>
<td><p>A list of names for the outputs of this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.output_shapes" title="mxnet.module.BaseModule.output_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the outputs of this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BaseModule.symbol" title="mxnet.module.BaseModule.symbol"><code class="xref py py-obj docutils literal notranslate"><span class="pre">symbol</span></code></a></p></td>
<td><p>Gets the symbol associated with this module.</p></td>
</tr>
</tbody>
</table>
<p>In order for a module to interact with others, it must be able to report the
following information in its initial state (before binding):</p>
<ul class="simple">
<li><p><cite>data_names</cite>: list of type string indicating the names of the required input data.</p></li>
<li><p><cite>output_names</cite>: list of type string indicating the names of the required outputs.</p></li>
</ul>
<p>After binding, a module should be able to report the following richer information:</p>
<ul class="simple">
<li><dl class="simple">
<dt>state information</dt><dd><ul>
<li><p><cite>binded</cite>: <cite>bool</cite>, indicates whether the memory buffers needed for computation
have been allocated.</p></li>
<li><p><cite>for_training</cite>: whether the module is bound for training.</p></li>
<li><p><cite>params_initialized</cite>: <cite>bool</cite>, indicates whether the parameters of this module
have been initialized.</p></li>
<li><p><cite>optimizer_initialized</cite>: <cite>bool</cite>, indicates whether an optimizer is defined
and initialized.</p></li>
<li><p><cite>inputs_need_grad</cite>: <cite>bool</cite>, indicates whether gradients with respect to the
input data are needed. Might be useful when implementing composition of modules.</p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>input/output information</dt><dd><ul>
<li><p><cite>data_shapes</cite>: a list of <cite>(name, shape)</cite>. In theory, since the memory is allocated,
we could directly provide the data arrays. But in the case of data parallelism,
the data arrays might not be of the same shape as viewed from the external world.</p></li>
<li><p><cite>label_shapes</cite>: a list of <cite>(name, shape)</cite>. This might be <cite>[]</cite> if the module does
not need labels (e.g. it does not contains a loss function at the top), or a module
is not bound for training.</p></li>
<li><p><cite>output_shapes</cite>: a list of <cite>(name, shape)</cite> for outputs of the module.</p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>parameters (for modules with parameters)</dt><dd><ul>
<li><p><cite>get_params()</cite>: return a tuple <cite>(arg_params, aux_params)</cite>. Each of those
is a dictionary of name to <code class="docutils literal notranslate"><span class="pre">NDArray</span></code> mapping. Those <cite>NDArray</cite> always lives on
CPU. The actual parameters used for computing might live on other devices (GPUs),
this function will retrieve (a copy of) the latest parameters.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">set_params(arg_params,</span> <span class="pre">aux_params)</span></code>: assign parameters to the devices
doing the computation.</p></li>
<li><p><code class="docutils literal notranslate"><span class="pre">init_params(...)</span></code>: a more flexible interface to assign or initialize the parameters.</p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>setup</dt><dd><ul>
<li><p><cite>bind()</cite>: prepare environment for computation.</p></li>
<li><p><cite>init_optimizer()</cite>: install optimizer for parameter updating.</p></li>
<li><p><cite>prepare()</cite>: prepare the module based on the current data batch.</p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>computation</dt><dd><ul>
<li><p><cite>forward(data_batch)</cite>: forward operation.</p></li>
<li><p><cite>backward(out_grads=None)</cite>: backward operation.</p></li>
<li><p><cite>update()</cite>: update parameters according to installed optimizer.</p></li>
<li><p><cite>get_outputs()</cite>: get outputs of the previous forward operation.</p></li>
<li><p><cite>get_input_grads()</cite>: get the gradients with respect to the inputs computed
in the previous backward operation.</p></li>
<li><p><cite>update_metric(metric, labels, pre_sliced=False)</cite>: update performance metric
for the previous forward
computed results.</p></li>
</ul>
</dd>
</dl>
</li>
<li><dl class="simple">
<dt>other properties (mostly for backward compatibility)</dt><dd><ul>
<li><p><cite>symbol</cite>: the underlying symbolic graph for this module (if any)
This property is not necessarily constant. For example, for <cite>BucketingModule</cite>,
this property is simply the <em>current</em> symbol being used. For other modules,
this value might not be well defined.</p></li>
</ul>
</dd>
</dl>
</li>
</ul>
<p>When those intermediate-level API are implemented properly, the following
high-level API will be automatically available for a module:</p>
<ul class="simple">
<li><p><cite>fit</cite>: train the module parameters on a data set.</p></li>
<li><p><cite>predict</cite>: run prediction on a data set and collect outputs.</p></li>
<li><p><cite>score</cite>: run prediction on a data set and evaluate performance.</p></li>
</ul>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of creating a mxnet module.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="kn">import</span> <span class="nn">mxnet</span> <span class="k">as</span> <span class="nn">mx</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">data</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">Variable</span><span class="p">(</span><span class="s1">&#39;data&#39;</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">fc1</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">FullyConnected</span><span class="p">(</span><span class="n">data</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">&#39;fc1&#39;</span><span class="p">,</span> <span class="n">num_hidden</span><span class="o">=</span><span class="mi">128</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">act1</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">Activation</span><span class="p">(</span><span class="n">fc1</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">&#39;relu1&#39;</span><span class="p">,</span> <span class="n">act_type</span><span class="o">=</span><span class="s2">&quot;relu&quot;</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">fc2</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">FullyConnected</span><span class="p">(</span><span class="n">act1</span><span class="p">,</span> <span class="n">name</span> <span class="o">=</span> <span class="s1">&#39;fc2&#39;</span><span class="p">,</span> <span class="n">num_hidden</span> <span class="o">=</span> <span class="mi">64</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">act2</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">Activation</span><span class="p">(</span><span class="n">fc2</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">&#39;relu2&#39;</span><span class="p">,</span> <span class="n">act_type</span><span class="o">=</span><span class="s2">&quot;relu&quot;</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">fc3</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">FullyConnected</span><span class="p">(</span><span class="n">act2</span><span class="p">,</span> <span class="n">name</span><span class="o">=</span><span class="s1">&#39;fc3&#39;</span><span class="p">,</span> <span class="n">num_hidden</span><span class="o">=</span><span class="mi">10</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">out</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">symbol</span><span class="o">.</span><span class="n">SoftmaxOutput</span><span class="p">(</span><span class="n">fc3</span><span class="p">,</span> <span class="n">name</span> <span class="o">=</span> <span class="s1">&#39;softmax&#39;</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">mod</span><span class="o">.</span><span class="n">Module</span><span class="p">(</span><span class="n">out</span><span class="p">)</span>
</pre></div>
</div>
<dl class="method">
<dt id="mxnet.module.BaseModule.backward">
<code class="sig-name descname">backward</code><span class="sig-paren">(</span><em class="sig-param">out_grads=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.backward" title="Permalink to this definition"></a></dt>
<dd><p>Backward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>out_grads</strong> (<a class="reference internal" href="../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a><em> or </em><em>list of NDArray</em><em>, </em><em>optional</em>) – Gradient on the outputs to be propagated back.
This parameter is only needed when bind is called
on outputs that are not a loss function.</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of backward computation.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</span><span class="o">.</span><span class="n">get_input_grads</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[[ 1.10182791e-05 5.12257748e-06 4.01927764e-06 8.32566820e-06</span>
<span class="go"> -1.59775993e-06 7.24269375e-06 7.28067835e-06 -1.65902311e-05</span>
<span class="go"> 5.46342608e-06 8.44196393e-07]</span>
<span class="go"> ...]]</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.bind">
<code class="sig-name descname">bind</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em>, <em class="sig-param">for_training=True</em>, <em class="sig-param">inputs_need_grad=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">shared_module=None</em>, <em class="sig-param">grad_req='write'</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.bind"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.bind" title="Permalink to this definition"></a></dt>
<dd><p>Binds the symbols to construct executors. This is necessary before one
can perform computation with the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>) or </em><em>DataDesc objects</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_data</span></code>. Can also be a list of
(data name, data shape).</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>) or </em><em>DataDesc objects</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_label</span></code>. Can also be a list of
(label name, label shape).</p></li>
<li><p><strong>for_training</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. Whether the executors should be bind for training.</p></li>
<li><p><strong>inputs_need_grad</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. Whether the gradients to the input data need to be computed.
Typically this is not needed. But this might be needed when implementing composition
of modules.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. This function does nothing if the executors are already
bound. But with this <code class="docutils literal notranslate"><span class="pre">True</span></code>, the executors will be forced to rebind.</p></li>
<li><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><em>Module</em></a>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>. This is used in bucketing. When not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the shared module
essentially corresponds to a different bucket – a module with different symbol
but with the same sets of parameters (e.g. unrolled RNNs with different lengths).</p></li>
<li><p><strong>grad_req</strong> (<em>str</em><em>, </em><em>list of str</em><em>, </em><em>dict of str to str</em>) – Requirement for gradient accumulation. Can be ‘write’, ‘add’, or ‘null’
(default to ‘write’).
Can be specified globally (str) or for each argument (list, dict).</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of binding symbols.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</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="p">[(</span><span class="s1">&#39;data&#39;</span><span class="p">,</span> <span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">10</span><span class="p">,</span> <span class="mi">10</span><span class="p">))])</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Assume train_iter is already created.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</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_iter</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_iter</span><span class="o">.</span><span class="n">provide_label</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.data_names">
<em class="property">property </em><code class="sig-name descname">data_names</code><a class="headerlink" href="#mxnet.module.BaseModule.data_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for data required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.data_shapes">
<em class="property">property </em><code class="sig-name descname">data_shapes</code><a class="headerlink" href="#mxnet.module.BaseModule.data_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the data inputs to this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.fit">
<code class="sig-name descname">fit</code><span class="sig-paren">(</span><em class="sig-param">train_data</em>, <em class="sig-param">eval_data=None</em>, <em class="sig-param">eval_metric='acc'</em>, <em class="sig-param">epoch_end_callback=None</em>, <em class="sig-param">batch_end_callback=None</em>, <em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">eval_end_callback=None</em>, <em class="sig-param">eval_batch_end_callback=None</em>, <em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">begin_epoch=0</em>, <em class="sig-param">num_epoch=None</em>, <em class="sig-param">validation_metric=None</em>, <em class="sig-param">monitor=None</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.fit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.fit" title="Permalink to this definition"></a></dt>
<dd><p>Trains the module parameters.</p>
<p>Checkout <a class="reference external" href="https://mxnet.apache.org/api/python/tutorials/packages/module/index.html">Module Tutorial</a>
to see an end-to-end use-case.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>train_data</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataIter" title="mxnet.io.DataIter"><em>DataIter</em></a>) – Train DataIter.</p></li>
<li><p><strong>eval_data</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataIter" title="mxnet.io.DataIter"><em>DataIter</em></a>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, will be used as validation set and the performance
after each epoch will be evaluated.</p></li>
<li><p><strong>eval_metric</strong> (<em>str</em><em> or </em><a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – Defaults to ‘accuracy’. The performance measure used to display during training.
Other possible predefined metrics are:
‘ce’ (CrossEntropy), ‘f1’, ‘mae’, ‘mse’, ‘rmse’, ‘top_k_accuracy’.</p></li>
<li><p><strong>epoch_end_callback</strong> (<em>function</em><em> or </em><em>list of functions</em>) – Each callback will be called with the current <cite>epoch</cite>, <cite>symbol</cite>, <cite>arg_params</cite>
and <cite>aux_params</cite>.</p></li>
<li><p><strong>batch_end_callback</strong> (<em>function</em><em> or </em><em>list of function</em>) – Each callback will be called with a <cite>BatchEndParam</cite>.</p></li>
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Defaults to ‘local’.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Defaults to ‘sgd’.</p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">(('learning_rate',</span> <span class="pre">0.01),)</span></code>. The parameters for
the optimizer constructor.
The default value is not a dict, just to avoid pylint warning on dangerous
default values.</p></li>
<li><p><strong>eval_end_callback</strong> (<em>function</em><em> or </em><em>list of function</em>) – These will be called at the end of each full evaluation, with the metrics over
the entire evaluation set.</p></li>
<li><p><strong>eval_batch_end_callback</strong> (<em>function</em><em> or </em><em>list of function</em>) – These will be called at the end of each mini-batch during evaluation.</p></li>
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – The initializer is called to initialize the module parameters when they are
not already initialized.</p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, if not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be existing parameters from a trained
model or loaded from a checkpoint (previously saved model). In this case,
the value here will be used to initialize the module parameters, unless they
are already initialized by the user via a call to <cite>init_params</cite> or <cite>fit</cite>.
<cite>arg_params</cite> has a higher priority than <cite>initializer</cite>.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>. Similar to <cite>arg_params</cite>, except for auxiliary states.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>. Indicates whether to allow missing parameters when <cite>arg_params</cite>
and <cite>aux_params</cite> are not <code class="docutils literal notranslate"><span class="pre">None</span></code>. If this is <code class="docutils literal notranslate"><span class="pre">True</span></code>, then the missing parameters
will be initialized via the <cite>initializer</cite>.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>. Whether to force rebinding the executors if already bound.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>. Indicates whether to force initialization even if the
parameters are already initialized.</p></li>
<li><p><strong>begin_epoch</strong> (<em>int</em>) – Defaults to 0. Indicates the starting epoch. Usually, if resumed from a
checkpoint saved at a previous training phase at epoch N, then this value should be
N+1.</p></li>
<li><p><strong>num_epoch</strong> (<em>int</em>) – Number of epochs for training.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of using fit for training.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Assume training dataIter and validation dataIter are ready</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Assume loading a previously checkpointed model</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">sym</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_checkpoint</span><span class="p">(</span><span class="n">model_prefix</span><span class="p">,</span> <span class="mi">3</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">fit</span><span class="p">(</span><span class="n">train_data</span><span class="o">=</span><span class="n">train_dataiter</span><span class="p">,</span> <span class="n">eval_data</span><span class="o">=</span><span class="n">val_dataiter</span><span class="p">,</span> <span class="n">optimizer</span><span class="o">=</span><span class="s1">&#39;sgd&#39;</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">optimizer_params</span><span class="o">=</span><span class="p">{</span><span class="s1">&#39;learning_rate&#39;</span><span class="p">:</span><span class="mf">0.01</span><span class="p">,</span> <span class="s1">&#39;momentum&#39;</span><span class="p">:</span> <span class="mf">0.9</span><span class="p">},</span>
<span class="gp">... </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="gp">... </span> <span class="n">eval_metric</span><span class="o">=</span><span class="s1">&#39;acc&#39;</span><span class="p">,</span> <span class="n">num_epoch</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">begin_epoch</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">is_train=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.forward" title="Permalink to this definition"></a></dt>
<dd><p>Forward computation. It supports data batches with different shapes, such as
different batch sizes or different image sizes.
If reshaping of data batch relates to modification of symbol or module, such as
changing image layout ordering or switching from training to predicting, module
rebinding is required.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – Could be anything with similar API implemented.</p></li>
<li><p><strong>is_train</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>, which means <cite>is_train</cite> takes the value of <code class="docutils literal notranslate"><span class="pre">self.for_training</span></code>.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="kn">import</span> <span class="nn">mxnet</span> <span class="k">as</span> <span class="nn">mx</span>
<span class="gp">&gt;&gt;&gt; </span><span class="kn">from</span> <span class="nn">collections</span> <span class="kn">import</span> <span class="n">namedtuple</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">Batch</span> <span class="o">=</span> <span class="n">namedtuple</span><span class="p">(</span><span class="s1">&#39;Batch&#39;</span><span class="p">,</span> <span class="p">[</span><span class="s1">&#39;data&#39;</span><span class="p">])</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">data</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">sym</span><span class="o">.</span><span class="n">Variable</span><span class="p">(</span><span class="s1">&#39;data&#39;</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">out</span> <span class="o">=</span> <span class="n">data</span> <span class="o">*</span> <span class="mi">2</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">mod</span><span class="o">.</span><span class="n">Module</span><span class="p">(</span><span class="n">symbol</span><span class="o">=</span><span class="n">out</span><span class="p">,</span> <span class="n">label_names</span><span class="o">=</span><span class="kc">None</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</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="p">[(</span><span class="s1">&#39;data&#39;</span><span class="p">,</span> <span class="p">(</span><span class="mi">1</span><span class="p">,</span> <span class="mi">10</span><span class="p">))])</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">init_params</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">data1</span> <span class="o">=</span> <span class="p">[</span><span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">1</span><span class="p">,</span> <span class="mi">10</span><span class="p">))]</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">Batch</span><span class="p">(</span><span class="n">data1</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</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="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[ 2. 2. 2. 2. 2. 2. 2. 2. 2. 2.]]</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Forward with data batch of different shape</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">data2</span> <span class="o">=</span> <span class="p">[</span><span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">ones</span><span class="p">((</span><span class="mi">3</span><span class="p">,</span> <span class="mi">5</span><span class="p">))]</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">forward</span><span class="p">(</span><span class="n">Batch</span><span class="p">(</span><span class="n">data2</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</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="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[ 2. 2. 2. 2. 2.]</span>
<span class="go"> [ 2. 2. 2. 2. 2.]</span>
<span class="go"> [ 2. 2. 2. 2. 2.]]</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.forward_backward">
<code class="sig-name descname">forward_backward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.forward_backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.forward_backward" title="Permalink to this definition"></a></dt>
<dd><p>A convenient function that calls both <code class="docutils literal notranslate"><span class="pre">forward</span></code> and <code class="docutils literal notranslate"><span class="pre">backward</span></code>.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.get_input_grads">
<code class="sig-name descname">get_input_grads</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.get_input_grads"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.get_input_grads" title="Permalink to this definition"></a></dt>
<dd><p>Gets the gradients to the inputs, computed in the previous backward computation.</p>
<p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[grad1,</span> <span class="pre">grad2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[grad1_dev1,</span> <span class="pre">grad1_dev2],</span> <span class="pre">[grad2_dev1,</span> <span class="pre">grad2_dev2]]</span></code>. All the output
elements have type <cite>NDArray</cite>. When <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">False</span></code>, those <cite>NDArray</cite>
instances might live on different devices.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the gradients
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicates that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>Input gradients.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of getting input gradients.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</span><span class="o">.</span><span class="n">get_input_grads</span><span class="p">()[</span><span class="mi">0</span><span class="p">]</span><span class="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[[ 1.10182791e-05 5.12257748e-06 4.01927764e-06 8.32566820e-06</span>
<span class="go"> -1.59775993e-06 7.24269375e-06 7.28067835e-06 -1.65902311e-05</span>
<span class="go"> 5.46342608e-06 8.44196393e-07]</span>
<span class="go"> ...]]</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.get_outputs">
<code class="sig-name descname">get_outputs</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.get_outputs"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.get_outputs" title="Permalink to this definition"></a></dt>
<dd><p>Gets outputs of the previous forward computation.</p>
<p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>. Otherwise,
it returns out put of form <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>.
All the output elements have type <cite>NDArray</cite>. When <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">False</span></code>,
those <cite>NDArray</cite> instances might live on different devices.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicates that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>Output</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of <cite>NDArray</cite> or list of list of <cite>NDArray</cite>.</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of getting forward output.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</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="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[ 0.09999977 0.10000153 0.10000716 0.10000195 0.09999853 0.09999743</span>
<span class="go"> 0.10000272 0.10000113 0.09999088 0.09999888]]</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.get_params">
<code class="sig-name descname">get_params</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.get_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.get_params" title="Permalink to this definition"></a></dt>
<dd><p>Gets parameters, those are potentially copies of the actual parameters used
to do computation on the device.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A pair of dictionaries each mapping parameter names to NDArray values.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p><code class="docutils literal notranslate"><span class="pre">(arg_params,</span> <span class="pre">aux_params)</span></code></p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of getting module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</span><span class="o">.</span><span class="n">get_params</span><span class="p">()</span>
<span class="go">({&#39;fc2_weight&#39;: &lt;NDArray 64x128 @cpu(0)&gt;, &#39;fc1_weight&#39;: &lt;NDArray 128x100 @cpu(0)&gt;,</span>
<span class="go">&#39;fc3_bias&#39;: &lt;NDArray 10 @cpu(0)&gt;, &#39;fc3_weight&#39;: &lt;NDArray 10x64 @cpu(0)&gt;,</span>
<span class="go">&#39;fc2_bias&#39;: &lt;NDArray 64 @cpu(0)&gt;, &#39;fc1_bias&#39;: &lt;NDArray 128 @cpu(0)&gt;}, {})</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.get_states">
<code class="sig-name descname">get_states</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.get_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.get_states" title="Permalink to this definition"></a></dt>
<dd><p>Gets states from all devices</p>
<p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, returns output of form <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>.
Otherwise, it returns output of the form
<code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>.
All output elements are <cite>NDArray</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the states
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicates that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p></p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>A list of <code class="docutils literal notranslate"><span class="pre">NDArray</span></code> or a list of list of <code class="docutils literal notranslate"><span class="pre">NDArray</span></code>.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.init_optimizer">
<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">force_init=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.init_optimizer" title="Permalink to this definition"></a></dt>
<dd><dl class="simple">
<dt>Installs and initializes optimizers, as well as initialize kvstore for</dt><dd><p>distributed training</p>
</dd>
</dl>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Defaults to <cite>‘local’</cite>.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Defaults to <cite>‘sgd’</cite>.</p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">(('learning_rate',</span> <span class="pre">0.01),)</span></code>. The default value is not a dictionary,
just to avoid pylint warning of dangerous default values.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>, indicates whether to force re-initializing an optimizer
if it is already installed.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of initializing optimizer.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">init_optimizer</span><span class="p">(</span><span class="n">optimizer</span><span class="o">=</span><span class="s1">&#39;sgd&#39;</span><span class="p">,</span> <span class="n">optimizer_params</span><span class="o">=</span><span class="p">((</span><span class="s1">&#39;learning_rate&#39;</span><span class="p">,</span> <span class="mf">0.005</span><span class="p">),))</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.init_params">
<code class="sig-name descname">init_params</code><span class="sig-paren">(</span><em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.init_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.init_params" title="Permalink to this definition"></a></dt>
<dd><p>Initializes the parameters and auxiliary states.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Called to initialize parameters if needed.</p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing <cite>arg_params</cite>. Initialization
will be copied from that.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing <cite>aux_params</cite>. Initialization
will be copied from that.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, <cite>force_init</cite> will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of initializing module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">init_params</span><span class="p">()</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.install_monitor">
<code class="sig-name descname">install_monitor</code><span class="sig-paren">(</span><em class="sig-param">mon</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.install_monitor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.install_monitor" title="Permalink to this definition"></a></dt>
<dd><p>Installs monitor on all executors.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.iter_predict">
<code class="sig-name descname">iter_predict</code><span class="sig-paren">(</span><em class="sig-param">eval_data</em>, <em class="sig-param">num_batch=None</em>, <em class="sig-param">reset=True</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.iter_predict"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.iter_predict" title="Permalink to this definition"></a></dt>
<dd><p>Iterates over predictions.</p>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="k">for</span> <span class="n">pred</span><span class="p">,</span> <span class="n">i_batch</span><span class="p">,</span> <span class="n">batch</span> <span class="ow">in</span> <span class="n">module</span><span class="o">.</span><span class="n">iter_predict</span><span class="p">(</span><span class="n">eval_data</span><span class="p">):</span>
<span class="gp">... </span> <span class="c1"># pred is a list of outputs from the module</span>
<span class="gp">... </span> <span class="c1"># i_batch is a integer</span>
<span class="gp">... </span> <span class="c1"># batch is the data batch from the data iterator</span>
</pre></div>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_data</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataIter" title="mxnet.io.DataIter"><em>DataIter</em></a>) – Evaluation data to run prediction on.</p></li>
<li><p><strong>num_batch</strong> (<em>int</em>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating running all the batches in the data iterator.</p></li>
<li><p><strong>reset</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>, indicating whether we should reset the data iter before start
doing prediction.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.label_shapes">
<em class="property">property </em><code class="sig-name descname">label_shapes</code><a class="headerlink" href="#mxnet.module.BaseModule.label_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the label inputs to this module.
If this module does not accept labels – either it is a module without loss
function, or it is not bound for training, then this should return an empty
list <code class="docutils literal notranslate"><span class="pre">[]</span></code>.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.load_params">
<code class="sig-name descname">load_params</code><span class="sig-paren">(</span><em class="sig-param">fname</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.load_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.load_params" title="Permalink to this definition"></a></dt>
<dd><p>Loads model parameters from file.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>fname</strong> (<em>str</em>) – Path to input param file.</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of loading module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">load_params</span><span class="p">(</span><span class="s1">&#39;myfile&#39;</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.output_names">
<em class="property">property </em><code class="sig-name descname">output_names</code><a class="headerlink" href="#mxnet.module.BaseModule.output_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.output_shapes">
<em class="property">property </em><code class="sig-name descname">output_shapes</code><a class="headerlink" href="#mxnet.module.BaseModule.output_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.predict">
<code class="sig-name descname">predict</code><span class="sig-paren">(</span><em class="sig-param">eval_data</em>, <em class="sig-param">num_batch=None</em>, <em class="sig-param">merge_batches=True</em>, <em class="sig-param">reset=True</em>, <em class="sig-param">always_output_list=False</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.predict"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.predict" title="Permalink to this definition"></a></dt>
<dd><p>Runs prediction and collects the outputs.</p>
<p>When <cite>merge_batches</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code> (by default), the return value will be a list
<code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2,</span> <span class="pre">out3]</span></code>, where each element is formed by concatenating the outputs for
all the mini-batches. When <cite>always_output_list</cite> is <code class="docutils literal notranslate"><span class="pre">False</span></code> (as by default),
then in the case of a single output, <cite>out1</cite> is returned instead of <code class="docutils literal notranslate"><span class="pre">[out1]</span></code>.</p>
<p>When <cite>merge_batches</cite> is <code class="docutils literal notranslate"><span class="pre">False</span></code>, the return value will be a nested list like
<code class="docutils literal notranslate"><span class="pre">[[out1_batch1,</span> <span class="pre">out2_batch1],</span> <span class="pre">[out1_batch2],</span> <span class="pre">...]</span></code>. This mode is useful because
in some cases (e.g. bucketing), the module does not necessarily produce the same
number of outputs.</p>
<p>The objects in the results have type <cite>NDArray</cite>. If you need to work with a numpy array,
just call <code class="docutils literal notranslate"><span class="pre">.asnumpy()</span></code> on each <cite>NDArray</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_data</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataIter" title="mxnet.io.DataIter"><em>DataIter</em></a><em> or </em><a class="reference internal" href="../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a><em> or </em><em>numpy array</em>) – Evaluation data to run prediction on.</p></li>
<li><p><strong>num_batch</strong> (<em>int</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicates running all the batches in the data iterator.</p></li>
<li><p><strong>merge_batches</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>, see above for return values.</p></li>
<li><p><strong>reset</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>, indicates whether we should reset the data iter before
doing prediction.</p></li>
<li><p><strong>always_output_list</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>, see above for return values.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>Prediction results.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of using `predict` for prediction.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Predict on the first 10 batches of val_dataiter</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">predict</span><span class="p">(</span><span class="n">eval_data</span><span class="o">=</span><span class="n">val_dataiter</span><span class="p">,</span> <span class="n">num_batch</span><span class="o">=</span><span class="mi">10</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.prepare">
<code class="sig-name descname">prepare</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.prepare"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.prepare" title="Permalink to this definition"></a></dt>
<dd><p>Prepares the module for processing a data batch.</p>
<p>Usually involves switching bucket and reshaping.
For modules that contain <cite>row_sparse</cite> parameters in KVStore,
it prepares the <cite>row_sparse</cite> parameters based on the sparse_row_id_fn.</p>
<p>When KVStore is used to update parameters for multi-device or multi-machine training,
a copy of the parameters are stored in KVStore. Note that for <cite>row_sparse</cite> parameters,
the <cite>update()</cite> updates the copy of parameters in KVStore, but doesn’t broadcast
the updated parameters to all devices / machines. The <cite>prepare</cite> function is used to
broadcast <cite>row_sparse</cite> parameters with the next batch of data.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – The current batch of data for forward computation.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.save_params">
<code class="sig-name descname">save_params</code><span class="sig-paren">(</span><em class="sig-param">fname</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.save_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.save_params" title="Permalink to this definition"></a></dt>
<dd><p>Saves model parameters to file.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>fname</strong> (<em>str</em>) – Path to output param file.</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of saving module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">save_params</span><span class="p">(</span><span class="s1">&#39;myfile&#39;</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.score">
<code class="sig-name descname">score</code><span class="sig-paren">(</span><em class="sig-param">eval_data</em>, <em class="sig-param">eval_metric</em>, <em class="sig-param">num_batch=None</em>, <em class="sig-param">batch_end_callback=None</em>, <em class="sig-param">score_end_callback=None</em>, <em class="sig-param">reset=True</em>, <em class="sig-param">epoch=0</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.score"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.score" title="Permalink to this definition"></a></dt>
<dd><p>Runs prediction on <code class="docutils literal notranslate"><span class="pre">eval_data</span></code> and evaluates the performance according to
the given <code class="docutils literal notranslate"><span class="pre">eval_metric</span></code>.</p>
<p>Checkout <a class="reference external" href="https://mxnet.apache.org/api/python/tutorials/packages/module/index.html">Module Tutorial</a>
to see an end-to-end use-case.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_data</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataIter" title="mxnet.io.DataIter"><em>DataIter</em></a>) – Evaluation data to run prediction on.</p></li>
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a><em> or </em><em>list of EvalMetrics</em>) – Evaluation metric to use.</p></li>
<li><p><strong>num_batch</strong> (<em>int</em>) – Number of batches to run. Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating run until the <cite>DataIter</cite>
finishes.</p></li>
<li><p><strong>batch_end_callback</strong> (<em>function</em>) – Could also be a list of functions.</p></li>
<li><p><strong>reset</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. Indicates whether we should reset <cite>eval_data</cite> before starting
evaluating.</p></li>
<li><p><strong>epoch</strong> (<em>int</em>) – Defaults to 0. For compatibility, this will be passed to callbacks (if any).
During training, this will correspond to the training epoch number.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of using score for prediction.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="c1"># Evaluate accuracy on val_dataiter</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">metric</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">metric</span><span class="o">.</span><span class="n">Accuracy</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">val_dataiter</span><span class="p">,</span> <span class="n">metric</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">score</span><span class="p">(</span><span class="n">val_dataiter</span><span class="p">,</span> <span class="p">[</span><span class="s1">&#39;mse&#39;</span><span class="p">,</span> <span class="s1">&#39;acc&#39;</span><span class="p">])</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.set_params">
<code class="sig-name descname">set_params</code><span class="sig-paren">(</span><em class="sig-param">arg_params</em>, <em class="sig-param">aux_params</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=True</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.set_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.set_params" title="Permalink to this definition"></a></dt>
<dd><p>Assigns parameter and aux state values.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>arg_params</strong> (<em>dict</em>) – Dictionary of name to value (<cite>NDArray</cite>) mapping.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Dictionary of name to value (<cite>NDArray</cite>) mapping.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of setting module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">sym</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_checkpoint</span><span class="p">(</span><span class="n">model_prefix</span><span class="p">,</span> <span class="n">n_epoch_load</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">set_params</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>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.set_states">
<code class="sig-name descname">set_states</code><span class="sig-paren">(</span><em class="sig-param">states=None</em>, <em class="sig-param">value=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.set_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.set_states" title="Permalink to this definition"></a></dt>
<dd><p>Sets value for states. Only one of states &amp; value can be specified.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>states</strong> (<em>list of list of NDArray</em>) – Source states arrays formatted like
<code class="docutils literal notranslate"><span class="pre">[[state1_dev1,</span> <span class="pre">state1_dev2],</span> <span class="pre">[state2_dev1,</span> <span class="pre">state2_dev2]]</span></code>.</p></li>
<li><p><strong>value</strong> (<em>number</em>) – A single scalar value for all state arrays.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.symbol">
<em class="property">property </em><code class="sig-name descname">symbol</code><a class="headerlink" href="#mxnet.module.BaseModule.symbol" title="Permalink to this definition"></a></dt>
<dd><p>Gets the symbol associated with this module.</p>
<p>Except for <cite>Module</cite>, for other types of modules (e.g. <cite>BucketingModule</cite>), this
property might not be a constant throughout its life time. Some modules might
not even be associated with any symbols.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.update">
<code class="sig-name descname">update</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.update"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.update" title="Permalink to this definition"></a></dt>
<dd><p>Updates parameters according to the installed optimizer and the gradients computed
in the previous forward-backward batch.</p>
<p>When KVStore is used to update parameters for multi-device or multi-machine training,
a copy of the parameters are stored in KVStore. Note that for <cite>row_sparse</cite> parameters,
this function does update the copy of parameters in KVStore, but doesn’t broadcast the
updated parameters to all devices / machines. Please call <cite>prepare</cite> to broadcast
<cite>row_sparse</cite> parameters with the next batch of data.</p>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of updating module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</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="s1">&#39;local&#39;</span><span class="p">,</span> <span class="n">optimizer</span><span class="o">=</span><span class="s1">&#39;sgd&#39;</span><span class="p">,</span>
<span class="gp">... </span> <span class="n">optimizer_params</span><span class="o">=</span><span class="p">((</span><span class="s1">&#39;learning_rate&#39;</span><span class="p">,</span> <span class="mf">0.01</span><span class="p">),</span> <span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">backward</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">update</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">print</span> <span class="n">mod</span><span class="o">.</span><span class="n">get_params</span><span class="p">()[</span><span class="mi">0</span><span class="p">][</span><span class="s1">&#39;fc3_weight&#39;</span><span class="p">]</span><span class="o">.</span><span class="n">asnumpy</span><span class="p">()</span>
<span class="go">[[ 5.86930104e-03 5.28078526e-03 -8.88729654e-03 -1.08308345e-03</span>
<span class="go"> 6.13054074e-03 4.27560415e-03 1.53817423e-03 4.62131854e-03</span>
<span class="go"> 4.69872449e-03 -2.42400169e-03 9.94111411e-04 1.12386420e-03</span>
<span class="go"> ...]]</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BaseModule.update_metric">
<code class="sig-name descname">update_metric</code><span class="sig-paren">(</span><em class="sig-param">eval_metric</em>, <em class="sig-param">labels</em>, <em class="sig-param">pre_sliced=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/base_module.html#BaseModule.update_metric"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BaseModule.update_metric" title="Permalink to this definition"></a></dt>
<dd><p>Evaluates and accumulates evaluation metric on outputs of the last forward
computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – Evaluation metric to use.</p></li>
<li><p><strong>labels</strong> (list of NDArray if <cite>pre_sliced</cite> parameter is set to <cite>False</cite>,) – list of lists of NDArray otherwise. Typically <cite>data_batch.label</cite>.</p></li>
<li><p><strong>pre_sliced</strong> (<em>bool</em>) – Whether the labels are already sliced per device (default: False).</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of updating evaluation metric.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</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="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">update_metric</span><span class="p">(</span><span class="n">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>
</pre></div>
</div>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.module.BucketingModule">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">BucketingModule</code><span class="sig-paren">(</span><em class="sig-param">sym_gen</em>, <em class="sig-param">default_bucket_key=None</em>, <em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em>, <em class="sig-param">context=cpu(0)</em>, <em class="sig-param">work_load_list=None</em>, <em class="sig-param">fixed_param_names=None</em>, <em class="sig-param">state_names=None</em>, <em class="sig-param">group2ctxs=None</em>, <em class="sig-param">compression_params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.module.base_module.BaseModule</span></code></p>
<p>This module helps to deal efficiently with varying-length inputs.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>sym_gen</strong> (<em>function</em>) – A function when called with a bucket key, returns a triple
<code class="docutils literal notranslate"><span class="pre">(symbol,</span> <span class="pre">data_names,</span> <span class="pre">label_names)</span></code>.</p></li>
<li><p><strong>default_bucket_key</strong> (<em>str</em><em> (or </em><em>any python object</em><em>)</em>) – The key for the default bucket.</p></li>
<li><p><strong>logger</strong> (<em>Logger</em>) – </p></li>
<li><p><strong>context</strong> (<a class="reference internal" href="../mxnet/context/index.html#mxnet.context.Context" title="mxnet.context.Context"><em>Context</em></a><em> or </em><em>list of Context</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">mx.cpu()</span></code></p></li>
<li><p><strong>work_load_list</strong> (<em>list of number</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating uniform workload.</p></li>
<li><p><strong>fixed_param_names</strong> (<em>list of str</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating no network parameters are fixed.</p></li>
<li><p><strong>state_names</strong> (<em>list of str</em>) – States are similar to data and label, but not provided by data iterator.
Instead they are initialized to 0 and can be set by set_states()</p></li>
<li><p><strong>group2ctxs</strong> (<em>dict of str to context</em><em> or </em><em>list of context</em><em>,</em>) – or list of dict of str to context
Default is <cite>None</cite>. Mapping the <cite>ctx_group</cite> attribute to the context assignment.</p></li>
<li><p><strong>compression_params</strong> (<em>dict</em>) – Specifies type of gradient compression and additional arguments depending
on the type of compression being used. For example, 2bit compression requires a threshold.
Arguments would then be {‘type’:’2bit’, ‘threshold’:0.5}
See mxnet.KVStore.set_gradient_compression method for more details on gradient compression.</p></li>
</ul>
</dd>
</dl>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.backward" title="mxnet.module.BucketingModule.backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">backward</span></code></a>([out_grads])</p></td>
<td><p>Backward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.bind" title="mxnet.module.BucketingModule.bind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">bind</span></code></a>(data_shapes[, label_shapes, …])</p></td>
<td><p>Binding for a <cite>BucketingModule</cite> means setting up the buckets and binding the executor for the default bucket key.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.forward" title="mxnet.module.BucketingModule.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(data_batch[, is_train])</p></td>
<td><p>Forward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.get_input_grads" title="mxnet.module.BucketingModule.get_input_grads"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_input_grads</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets the gradients with respect to the inputs of the module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.get_outputs" title="mxnet.module.BucketingModule.get_outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_outputs</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets outputs from a previous forward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.get_params" title="mxnet.module.BucketingModule.get_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_params</span></code></a>()</p></td>
<td><p>Gets current parameters.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.get_states" title="mxnet.module.BucketingModule.get_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_states</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets states from all devices.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.init_optimizer" title="mxnet.module.BucketingModule.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>([kvstore, optimizer, …])</p></td>
<td><p>Installs and initializes optimizers.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.init_params" title="mxnet.module.BucketingModule.init_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_params</span></code></a>([initializer, arg_params, …])</p></td>
<td><p>Initializes parameters.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.install_monitor" title="mxnet.module.BucketingModule.install_monitor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">install_monitor</span></code></a>(mon)</p></td>
<td><p>Installs monitor on all executors</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.load" title="mxnet.module.BucketingModule.load"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code></a>(prefix, epoch[, sym_gen, …])</p></td>
<td><p>Creates a model from previously saved checkpoint.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.load_dict" title="mxnet.module.BucketingModule.load_dict"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_dict</span></code></a>([sym_dict, sym_gen, …])</p></td>
<td><p>Creates a model from a dict mapping bucket_key to symbols and shared arg_params and aux_params.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.prepare" title="mxnet.module.BucketingModule.prepare"><code class="xref py py-obj docutils literal notranslate"><span class="pre">prepare</span></code></a>(data_batch[, sparse_row_id_fn])</p></td>
<td><p>Prepares the module for processing a data batch.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.save_checkpoint" title="mxnet.module.BucketingModule.save_checkpoint"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save_checkpoint</span></code></a>(prefix, epoch[, remove_amp_cast])</p></td>
<td><p>Saves current progress to checkpoint for all buckets in BucketingModule Use <cite>mx.callback.module_checkpoint</cite> as <cite>epoch_end_callback</cite> to save during training.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.set_params" title="mxnet.module.BucketingModule.set_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_params</span></code></a>(arg_params, aux_params[, …])</p></td>
<td><p>Assigns parameters and aux state values.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.set_states" title="mxnet.module.BucketingModule.set_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_states</span></code></a>([states, value])</p></td>
<td><p>Sets value for states.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.switch_bucket" title="mxnet.module.BucketingModule.switch_bucket"><code class="xref py py-obj docutils literal notranslate"><span class="pre">switch_bucket</span></code></a>(bucket_key, data_shapes[, …])</p></td>
<td><p>Switches to a different bucket.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.update" title="mxnet.module.BucketingModule.update"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update</span></code></a>()</p></td>
<td><p>Updates parameters according to installed optimizer and the gradient computed in the previous forward-backward cycle.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.update_metric" title="mxnet.module.BucketingModule.update_metric"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_metric</span></code></a>(eval_metric, labels[, pre_sliced])</p></td>
<td><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p></td>
</tr>
</tbody>
</table>
<p><strong>Attributes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.data_names" title="mxnet.module.BucketingModule.data_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_names</span></code></a></p></td>
<td><p>A list of names for data required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.data_shapes" title="mxnet.module.BucketingModule.data_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_shapes</span></code></a></p></td>
<td><p>Get data shapes.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.label_shapes" title="mxnet.module.BucketingModule.label_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_shapes</span></code></a></p></td>
<td><p>Get label shapes.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.output_names" title="mxnet.module.BucketingModule.output_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_names</span></code></a></p></td>
<td><p>A list of names for the outputs of this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.output_shapes" title="mxnet.module.BucketingModule.output_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_shapes</span></code></a></p></td>
<td><p>Gets output shapes.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.BucketingModule.symbol" title="mxnet.module.BucketingModule.symbol"><code class="xref py py-obj docutils literal notranslate"><span class="pre">symbol</span></code></a></p></td>
<td><p>The symbol of the current bucket being used.</p></td>
</tr>
</tbody>
</table>
<dl class="method">
<dt id="mxnet.module.BucketingModule.backward">
<code class="sig-name descname">backward</code><span class="sig-paren">(</span><em class="sig-param">out_grads=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.backward" title="Permalink to this definition"></a></dt>
<dd><p>Backward computation.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.bind">
<code class="sig-name descname">bind</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em>, <em class="sig-param">for_training=True</em>, <em class="sig-param">inputs_need_grad=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">shared_module=None</em>, <em class="sig-param">grad_req='write'</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.bind"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.bind" title="Permalink to this definition"></a></dt>
<dd><p>Binding for a <cite>BucketingModule</cite> means setting up the buckets and binding the
executor for the default bucket key. Executors corresponding to other keys are
bound afterwards with <cite>switch_bucket</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – This should correspond to the symbol for the default bucket.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – This should correspond to the symbol for the default bucket.</p></li>
<li><p><strong>for_training</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>.</p></li>
<li><p><strong>inputs_need_grad</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>.</p></li>
<li><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.BucketingModule" title="mxnet.module.BucketingModule"><em>BucketingModule</em></a>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>. This value is currently not used.</p></li>
<li><p><strong>grad_req</strong> (<em>str</em><em>, </em><em>list of str</em><em>, </em><em>dict of str to str</em>) – Requirement for gradient accumulation. Can be ‘write’, ‘add’, or ‘null’
(default to ‘write’).
Can be specified globally (str) or for each argument (list, dict).</p></li>
<li><p><strong>bucket_key</strong> (<em>str</em><em> (or </em><em>any python object</em><em>)</em>) – bucket key for binding. by default use the default_bucket_key</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.data_names">
<em class="property">property </em><code class="sig-name descname">data_names</code><a class="headerlink" href="#mxnet.module.BucketingModule.data_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for data required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.data_shapes">
<em class="property">property </em><code class="sig-name descname">data_shapes</code><a class="headerlink" href="#mxnet.module.BucketingModule.data_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Get data shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p></p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">is_train=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.forward" title="Permalink to this definition"></a></dt>
<dd><p>Forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – </p></li>
<li><p><strong>is_train</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>, in which case <cite>is_train</cite> is take as <code class="docutils literal notranslate"><span class="pre">self.for_training</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.get_input_grads">
<code class="sig-name descname">get_input_grads</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.get_input_grads"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.get_input_grads" title="Permalink to this definition"></a></dt>
<dd><p>Gets the gradients with respect to the inputs of the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[grad1,</span> <span class="pre">grad2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[grad1_dev1,</span> <span class="pre">grad1_dev2],</span> <span class="pre">[grad2_dev1,</span> <span class="pre">grad2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArrays or list of list of NDArrays</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.get_outputs">
<code class="sig-name descname">get_outputs</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.get_outputs"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.get_outputs" title="Permalink to this definition"></a></dt>
<dd><p>Gets outputs from a previous forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>. All the output
elements are numpy arrays.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of numpy arrays or list of list of numpy arrays</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.get_params">
<code class="sig-name descname">get_params</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.get_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.get_params" title="Permalink to this definition"></a></dt>
<dd><p>Gets current parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A pair of dictionaries each mapping parameter names to NDArray values.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p><cite>(arg_params, aux_params)</cite></p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.get_states">
<code class="sig-name descname">get_states</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.get_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.get_states" title="Permalink to this definition"></a></dt>
<dd><p>Gets states from all devices.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <cite>True</cite>. In the case when data-parallelism is used, the states
will be collected from multiple devices. A <cite>True</cite> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArrays or list of list of NDArrays</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.init_optimizer">
<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">force_init=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.init_optimizer" title="Permalink to this definition"></a></dt>
<dd><p>Installs and initializes optimizers.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Defaults to <cite>‘local’</cite>.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Defaults to <cite>‘sgd’</cite></p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Defaults to <cite>((‘learning_rate’, 0.01),)</cite>. The default value is not a dictionary,
just to avoid pylint warning of dangerous default values.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>, indicating whether we should force re-initializing the
optimizer in the case an optimizer is already installed.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.init_params">
<code class="sig-name descname">init_params</code><span class="sig-paren">(</span><em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.init_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.init_params" title="Permalink to this definition"></a></dt>
<dd><p>Initializes parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – </p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>. Existing parameters. This has higher priority
than <cite>initializer</cite>.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">None</span></code>. Existing auxiliary states. This has higher priority
than <cite>initializer</cite>.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – Allow missing values in <cite>arg_params</cite> and <cite>aux_params</cite> (if not <code class="docutils literal notranslate"><span class="pre">None</span></code>).
In this case, missing values will be filled with <cite>initializer</cite>.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">False</span></code>.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.install_monitor">
<code class="sig-name descname">install_monitor</code><span class="sig-paren">(</span><em class="sig-param">mon</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.install_monitor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.install_monitor" title="Permalink to this definition"></a></dt>
<dd><p>Installs monitor on all executors</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.label_shapes">
<em class="property">property </em><code class="sig-name descname">label_shapes</code><a class="headerlink" href="#mxnet.module.BucketingModule.label_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Get label shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>The return value could be <code class="docutils literal notranslate"><span class="pre">None</span></code> if the module does not need labels,
or if the module is not bound for training (in this case, label information
is not available).</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.load">
<em class="property">static </em><code class="sig-name descname">load</code><span class="sig-paren">(</span><em class="sig-param">prefix</em>, <em class="sig-param">epoch</em>, <em class="sig-param">sym_gen=None</em>, <em class="sig-param">default_bucket_key=None</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.load"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.load" title="Permalink to this definition"></a></dt>
<dd><p>Creates a model from previously saved checkpoint.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>prefix</strong> (<em>str</em>) – path prefix of saved model files. You should have
“prefix-symbol.json”, “prefix-xxxx.params”, and
optionally “prefix-xxxx.states”, where xxxx is the
epoch number.</p></li>
<li><p><strong>epoch</strong> (<em>int</em>) – epoch to load.</p></li>
<li><p><strong>sym_gen</strong> (<em>function</em>) – A function when called with a bucket key, returns a triple
<code class="docutils literal notranslate"><span class="pre">(symbol,</span> <span class="pre">data_names,</span> <span class="pre">label_names)</span></code>.
provide sym_gen which was used when saving bucketing module.</p></li>
<li><p><strong>logger</strong> (<em>Logger</em>) – Default is <cite>logging</cite>.</p></li>
<li><p><strong>context</strong> (<a class="reference internal" href="../mxnet/context/index.html#mxnet.context.Context" title="mxnet.context.Context"><em>Context</em></a><em> or </em><em>list of Context</em>) – Default is <code class="docutils literal notranslate"><span class="pre">cpu()</span></code>.</p></li>
<li><p><strong>work_load_list</strong> (<em>list of number</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating uniform workload.</p></li>
<li><p><strong>fixed_param_names</strong> (<em>list of str</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating no network parameters are fixed.</p></li>
<li><p><strong>state_names</strong> (<em>list of str</em>) – States are similar to data and label, but not provided by data iterator.
Instead they are initialized to 0 and can be set by set_states()</p></li>
<li><p><strong>group2ctxs</strong> (<em>dict of str to context</em><em> or </em><em>list of context</em><em>,</em>) – or list of dict of str to context
Default is <cite>None</cite>. Mapping the <cite>ctx_group</cite> attribute to the context assignment.</p></li>
<li><p><strong>compression_params</strong> (<em>dict</em>) – Specifies type of gradient compression and additional arguments depending
on the type of compression being used. For example, 2bit compression requires a threshold.
Arguments would then be {‘type’:’2bit’, ‘threshold’:0.5}
See mxnet.KVStore.set_gradient_compression method for more details on gradient compression.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.load_dict">
<em class="property">static </em><code class="sig-name descname">load_dict</code><span class="sig-paren">(</span><em class="sig-param">sym_dict=None</em>, <em class="sig-param">sym_gen=None</em>, <em class="sig-param">default_bucket_key=None</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.load_dict"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.load_dict" title="Permalink to this definition"></a></dt>
<dd><p>Creates a model from a dict mapping bucket_key to symbols and shared arg_params
and aux_params.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>sym_dict</strong> (<em>dict mapping bucket_key to symbol</em>) – Dict mapping bucket key to symbol</p></li>
<li><p><strong>sym_gen</strong> (<em>function</em>) – A function when called with a bucket key, returns a triple
<code class="docutils literal notranslate"><span class="pre">(symbol,</span> <span class="pre">data_names,</span> <span class="pre">label_names)</span></code>.
provide sym_gen which was used when saving bucketing module.</p></li>
<li><p><strong>default_bucket_key</strong> (<em>str</em><em> (or </em><em>any python object</em><em>)</em>) – The key for the default bucket.</p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – Required for loading the BucketingModule.
Dict of name to parameter ndarrays.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Required for loading the BucketingModule.
Dict of name to auxiliary state ndarrays.</p></li>
<li><p><strong>logger</strong> (<em>Logger</em>) – Default is <cite>logging</cite>.</p></li>
<li><p><strong>context</strong> (<a class="reference internal" href="../mxnet/context/index.html#mxnet.context.Context" title="mxnet.context.Context"><em>Context</em></a><em> or </em><em>list of Context</em>) – Default is <code class="docutils literal notranslate"><span class="pre">cpu()</span></code>.</p></li>
<li><p><strong>work_load_list</strong> (<em>list of number</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating uniform workload.</p></li>
<li><p><strong>fixed_param_names</strong> (<em>list of str</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating no network parameters are fixed.</p></li>
<li><p><strong>state_names</strong> (<em>list of str</em>) – States are similar to data and label, but not provided by data iterator.
Instead they are initialized to 0 and can be set by set_states()</p></li>
<li><p><strong>group2ctxs</strong> (<em>dict of str to context</em><em> or </em><em>list of context</em><em>,</em>) – or list of dict of str to context
Default is <cite>None</cite>. Mapping the <cite>ctx_group</cite> attribute to the context assignment.</p></li>
<li><p><strong>compression_params</strong> (<em>dict</em>) – Specifies type of gradient compression and additional arguments depending
on the type of compression being used. For example, 2bit compression requires a threshold.
Arguments would then be {‘type’:’2bit’, ‘threshold’:0.5}
See mxnet.KVStore.set_gradient_compression method for more details on gradient compression.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.output_names">
<em class="property">property </em><code class="sig-name descname">output_names</code><a class="headerlink" href="#mxnet.module.BucketingModule.output_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.output_shapes">
<em class="property">property </em><code class="sig-name descname">output_shapes</code><a class="headerlink" href="#mxnet.module.BucketingModule.output_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets output shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p></p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.prepare">
<code class="sig-name descname">prepare</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.prepare"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.prepare" title="Permalink to this definition"></a></dt>
<dd><p>Prepares the module for processing a data batch.</p>
<p>Usually involves switching bucket and reshaping.
For modules that contain <cite>row_sparse</cite> parameters in KVStore,
it prepares the <cite>row_sparse</cite> parameters based on the sparse_row_id_fn.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – The current batch of data for forward computation.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.save_checkpoint">
<code class="sig-name descname">save_checkpoint</code><span class="sig-paren">(</span><em class="sig-param">prefix</em>, <em class="sig-param">epoch</em>, <em class="sig-param">remove_amp_cast=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.save_checkpoint"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.save_checkpoint" title="Permalink to this definition"></a></dt>
<dd><p>Saves current progress to checkpoint for all buckets in BucketingModule
Use <cite>mx.callback.module_checkpoint</cite> as <cite>epoch_end_callback</cite> to save during training.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>prefix</strong> (<em>str</em>) – The file prefix to checkpoint to.</p></li>
<li><p><strong>epoch</strong> (<em>int</em>) – The current epoch number.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.set_params">
<code class="sig-name descname">set_params</code><span class="sig-paren">(</span><em class="sig-param">arg_params</em>, <em class="sig-param">aux_params</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=True</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.set_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.set_params" title="Permalink to this definition"></a></dt>
<dd><p>Assigns parameters and aux state values.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>arg_params</strong> (<em>dict</em>) – Dictionary of name to value (<cite>NDArray</cite>) mapping.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Dictionary of name to value (<cite>NDArray</cite>) mapping.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If true, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If true, will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of setting module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">sym</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_checkpoint</span><span class="p">(</span><span class="n">model_prefix</span><span class="p">,</span> <span class="n">n_epoch_load</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">set_params</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>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.set_states">
<code class="sig-name descname">set_states</code><span class="sig-paren">(</span><em class="sig-param">states=None</em>, <em class="sig-param">value=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.set_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.set_states" title="Permalink to this definition"></a></dt>
<dd><p>Sets value for states. Only one of states &amp; values can be specified.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>states</strong> (<em>list of list of NDArrays</em>) – Source states arrays formatted like <code class="docutils literal notranslate"><span class="pre">[[state1_dev1,</span> <span class="pre">state1_dev2],</span>
<span class="pre">[state2_dev1,</span> <span class="pre">state2_dev2]]</span></code>.</p></li>
<li><p><strong>value</strong> (<em>number</em>) – A single scalar value for all state arrays.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.switch_bucket">
<code class="sig-name descname">switch_bucket</code><span class="sig-paren">(</span><em class="sig-param">bucket_key</em>, <em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.switch_bucket"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.switch_bucket" title="Permalink to this definition"></a></dt>
<dd><p>Switches to a different bucket. This will change <code class="docutils literal notranslate"><span class="pre">self.curr_module</span></code>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>bucket_key</strong> (<em>str</em><em> (or </em><em>any python object</em><em>)</em>) – The key of the target bucket.</p></li>
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically <code class="docutils literal notranslate"><span class="pre">data_batch.provide_data</span></code>.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically <code class="docutils literal notranslate"><span class="pre">data_batch.provide_label</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.symbol">
<em class="property">property </em><code class="sig-name descname">symbol</code><a class="headerlink" href="#mxnet.module.BucketingModule.symbol" title="Permalink to this definition"></a></dt>
<dd><p>The symbol of the current bucket being used.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.update">
<code class="sig-name descname">update</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.update"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.update" title="Permalink to this definition"></a></dt>
<dd><p>Updates parameters according to installed optimizer and the gradient computed
in the previous forward-backward cycle.</p>
<p>When KVStore is used to update parameters for multi-device or multi-machine training,
a copy of the parameters are stored in KVStore. Note that for <cite>row_sparse</cite> parameters,
this function does update the copy of parameters in KVStore, but doesn’t broadcast the
updated parameters to all devices / machines. Please call <cite>prepare</cite> to broadcast
<cite>row_sparse</cite> parameters with the next batch of data.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.BucketingModule.update_metric">
<code class="sig-name descname">update_metric</code><span class="sig-paren">(</span><em class="sig-param">eval_metric</em>, <em class="sig-param">labels</em>, <em class="sig-param">pre_sliced=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/bucketing_module.html#BucketingModule.update_metric"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.BucketingModule.update_metric" title="Permalink to this definition"></a></dt>
<dd><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – </p></li>
<li><p><strong>labels</strong> (<em>list of NDArray</em>) – Typically <code class="docutils literal notranslate"><span class="pre">data_batch.label</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.module.Module">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">Module</code><span class="sig-paren">(</span><em class="sig-param">symbol</em>, <em class="sig-param">data_names=('data'</em>, <em class="sig-param">)</em>, <em class="sig-param">label_names=('softmax_label'</em>, <em class="sig-param">)</em>, <em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em>, <em class="sig-param">context=cpu(0)</em>, <em class="sig-param">work_load_list=None</em>, <em class="sig-param">fixed_param_names=None</em>, <em class="sig-param">state_names=None</em>, <em class="sig-param">group2ctxs=None</em>, <em class="sig-param">compression_params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.module.base_module.BaseModule</span></code></p>
<p>Module is a basic module that wrap a <cite>Symbol</cite>. It is functionally the same
as the <cite>FeedForward</cite> model, except under the module API.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>symbol</strong> (<a class="reference internal" href="../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a>) – </p></li>
<li><p><strong>data_names</strong> (<em>list of str</em>) – Defaults to <cite>(‘data’)</cite> for a typical model used in image classification.</p></li>
<li><p><strong>label_names</strong> (<em>list of str</em>) – Defaults to <cite>(‘softmax_label’)</cite> for a typical model used in image
classification.</p></li>
<li><p><strong>logger</strong> (<em>Logger</em>) – Defaults to <cite>logging</cite>.</p></li>
<li><p><strong>context</strong> (<a class="reference internal" href="../mxnet/context/index.html#mxnet.context.Context" title="mxnet.context.Context"><em>Context</em></a><em> or </em><em>list of Context</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">mx.cpu()</span></code>.</p></li>
<li><p><strong>work_load_list</strong> (<em>list of number</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating uniform workload.</p></li>
<li><p><strong>fixed_param_names</strong> (<em>list of str</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating no network parameters are fixed.</p></li>
<li><p><strong>state_names</strong> (<em>list of str</em>) – states are similar to data and label, but not provided by data iterator.
Instead they are initialized to 0 and can be set by <cite>set_states()</cite>.</p></li>
<li><p><strong>group2ctxs</strong> (<em>dict of str to context</em><em> or </em><em>list of context</em><em>,</em>) – or list of dict of str to context
Default is <cite>None</cite>. Mapping the <cite>ctx_group</cite> attribute to the context assignment.</p></li>
<li><p><strong>compression_params</strong> (<em>dict</em>) – Specifies type of gradient compression and additional arguments depending
on the type of compression being used. For example, 2bit compression requires a threshold.
Arguments would then be {‘type’:’2bit’, ‘threshold’:0.5}
See mxnet.KVStore.set_gradient_compression method for more details on gradient compression.</p></li>
</ul>
</dd>
</dl>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.backward" title="mxnet.module.Module.backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">backward</span></code></a>([out_grads])</p></td>
<td><p>Backward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.bind" title="mxnet.module.Module.bind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">bind</span></code></a>(data_shapes[, label_shapes, …])</p></td>
<td><p>Binds the symbols to construct executors.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.borrow_optimizer" title="mxnet.module.Module.borrow_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">borrow_optimizer</span></code></a>(shared_module)</p></td>
<td><p>Borrows optimizer from a shared module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.forward" title="mxnet.module.Module.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(data_batch[, is_train])</p></td>
<td><p>Forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.get_input_grads" title="mxnet.module.Module.get_input_grads"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_input_grads</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets the gradients with respect to the inputs of the module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.get_outputs" title="mxnet.module.Module.get_outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_outputs</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets outputs of the previous forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.get_params" title="mxnet.module.Module.get_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_params</span></code></a>()</p></td>
<td><p>Gets current parameters.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.get_states" title="mxnet.module.Module.get_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_states</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets states from all devices.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.init_optimizer" title="mxnet.module.Module.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>([kvstore, optimizer, …])</p></td>
<td><p>Installs and initializes optimizers.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.init_params" title="mxnet.module.Module.init_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_params</span></code></a>([initializer, arg_params, …])</p></td>
<td><p>Initializes the parameters and auxiliary states.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.install_monitor" title="mxnet.module.Module.install_monitor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">install_monitor</span></code></a>(mon)</p></td>
<td><p>Installs monitor on all executors.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.load" title="mxnet.module.Module.load"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load</span></code></a>(prefix, epoch[, load_optimizer_states])</p></td>
<td><p>Creates a model from previously saved checkpoint.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.load_optimizer_states" title="mxnet.module.Module.load_optimizer_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">load_optimizer_states</span></code></a>(fname)</p></td>
<td><p>Loads optimizer (updater) state from a file.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.prepare" title="mxnet.module.Module.prepare"><code class="xref py py-obj docutils literal notranslate"><span class="pre">prepare</span></code></a>(data_batch[, sparse_row_id_fn])</p></td>
<td><p>Prepares the module for processing a data batch.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.reshape" title="mxnet.module.Module.reshape"><code class="xref py py-obj docutils literal notranslate"><span class="pre">reshape</span></code></a>(data_shapes[, label_shapes])</p></td>
<td><p>Reshapes the module for new input shapes.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.save_checkpoint" title="mxnet.module.Module.save_checkpoint"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save_checkpoint</span></code></a>(prefix, epoch[, …])</p></td>
<td><p>Saves current progress to checkpoint.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.save_optimizer_states" title="mxnet.module.Module.save_optimizer_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">save_optimizer_states</span></code></a>(fname)</p></td>
<td><p>Saves optimizer (updater) state to a file.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.set_params" title="mxnet.module.Module.set_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_params</span></code></a>(arg_params, aux_params[, …])</p></td>
<td><p>Assigns parameter and aux state values.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.set_states" title="mxnet.module.Module.set_states"><code class="xref py py-obj docutils literal notranslate"><span class="pre">set_states</span></code></a>([states, value])</p></td>
<td><p>Sets value for states.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.update" title="mxnet.module.Module.update"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update</span></code></a>()</p></td>
<td><p>Updates parameters according to the installed optimizer and the gradients computed in the previous forward-backward batch.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.update_metric" title="mxnet.module.Module.update_metric"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_metric</span></code></a>(eval_metric, labels[, pre_sliced])</p></td>
<td><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p></td>
</tr>
</tbody>
</table>
<p><strong>Attributes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.data_names" title="mxnet.module.Module.data_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_names</span></code></a></p></td>
<td><p>A list of names for data required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.data_shapes" title="mxnet.module.Module.data_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_shapes</span></code></a></p></td>
<td><p>Gets data shapes.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.label_names" title="mxnet.module.Module.label_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_names</span></code></a></p></td>
<td><p>A list of names for labels required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.label_shapes" title="mxnet.module.Module.label_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_shapes</span></code></a></p></td>
<td><p>Gets label shapes.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.Module.output_names" title="mxnet.module.Module.output_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_names</span></code></a></p></td>
<td><p>A list of names for the outputs of this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.Module.output_shapes" title="mxnet.module.Module.output_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_shapes</span></code></a></p></td>
<td><p>Gets output shapes.</p></td>
</tr>
</tbody>
</table>
<dl class="method">
<dt id="mxnet.module.Module.backward">
<code class="sig-name descname">backward</code><span class="sig-paren">(</span><em class="sig-param">out_grads=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.backward" title="Permalink to this definition"></a></dt>
<dd><p>Backward computation.</p>
<div class="admonition seealso">
<p class="admonition-title">See also</p>
<p><a class="reference internal" href="#mxnet.module.BaseModule.backward" title="mxnet.module.BaseModule.backward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">BaseModule.backward()</span></code></a></p>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>out_grads</strong> (<a class="reference internal" href="../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a><em> or </em><em>list of NDArray</em><em>, </em><em>optional</em>) – Gradient on the outputs to be propagated back.
This parameter is only needed when bind is called
on outputs that are not a loss function.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.bind">
<code class="sig-name descname">bind</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em>, <em class="sig-param">for_training=True</em>, <em class="sig-param">inputs_need_grad=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">shared_module=None</em>, <em class="sig-param">grad_req='write'</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.bind"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.bind" title="Permalink to this definition"></a></dt>
<dd><p>Binds the symbols to construct executors. This is necessary before one
can perform computation with the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_data</span></code>.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_label</span></code>.</p></li>
<li><p><strong>for_training</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. Whether the executors should be bound for training.</p></li>
<li><p><strong>inputs_need_grad</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. Whether the gradients to the input data need to be computed.
Typically this is not needed. But this might be needed when implementing composition
of modules.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. This function does nothing if the executors are already
bound. But with this <code class="docutils literal notranslate"><span class="pre">True</span></code>, the executors will be forced to rebind.</p></li>
<li><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><em>Module</em></a>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>. This is used in bucketing. When not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the shared module
essentially corresponds to a different bucket – a module with different symbol
but with the same sets of parameters (e.g. unrolled RNNs with different lengths).</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.borrow_optimizer">
<code class="sig-name descname">borrow_optimizer</code><span class="sig-paren">(</span><em class="sig-param">shared_module</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.borrow_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.borrow_optimizer" title="Permalink to this definition"></a></dt>
<dd><p>Borrows optimizer from a shared module. Used in bucketing, where exactly the same
optimizer (esp. kvstore) is used.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><em>Module</em></a>) – </p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.data_names">
<em class="property">property </em><code class="sig-name descname">data_names</code><a class="headerlink" href="#mxnet.module.Module.data_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for data required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.data_shapes">
<em class="property">property </em><code class="sig-name descname">data_shapes</code><a class="headerlink" href="#mxnet.module.Module.data_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets data shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p></p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">is_train=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.forward" title="Permalink to this definition"></a></dt>
<dd><p>Forward computation. It supports data batches with different shapes, such as
different batch sizes or different image sizes.
If reshaping of data batch relates to modification of symbol or module, such as
changing image layout ordering or switching from training to predicting, module
rebinding is required.</p>
<div class="admonition seealso">
<p class="admonition-title">See also</p>
<p><a class="reference internal" href="#mxnet.module.BaseModule.forward" title="mxnet.module.BaseModule.forward"><code class="xref py py-meth docutils literal notranslate"><span class="pre">BaseModule.forward()</span></code></a></p>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – Could be anything with similar API implemented.</p></li>
<li><p><strong>is_train</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>, which means <code class="docutils literal notranslate"><span class="pre">is_train</span></code> takes the value of <code class="docutils literal notranslate"><span class="pre">self.for_training</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.get_input_grads">
<code class="sig-name descname">get_input_grads</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.get_input_grads"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.get_input_grads" title="Permalink to this definition"></a></dt>
<dd><p>Gets the gradients with respect to the inputs of the module.</p>
<p>If <code class="docutils literal notranslate"><span class="pre">merge_multi_context</span></code> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[grad1,</span> <span class="pre">grad2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[grad1_dev1,</span> <span class="pre">grad1_dev2],</span> <span class="pre">[grad2_dev1,</span> <span class="pre">grad2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>Input gradients</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.get_outputs">
<code class="sig-name descname">get_outputs</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.get_outputs"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.get_outputs" title="Permalink to this definition"></a></dt>
<dd><p>Gets outputs of the previous forward computation.</p>
<p>If <code class="docutils literal notranslate"><span class="pre">merge_multi_context</span></code> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>. When <cite>merge_multi_context</cite> is <cite>False</cite>, those <cite>NDArray</cite>
might live on different devices.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>Output.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.get_params">
<code class="sig-name descname">get_params</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.get_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.get_params" title="Permalink to this definition"></a></dt>
<dd><p>Gets current parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A pair of dictionaries each mapping parameter names to NDArray values.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p><cite>(arg_params, aux_params)</cite></p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.get_states">
<code class="sig-name descname">get_states</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.get_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.get_states" title="Permalink to this definition"></a></dt>
<dd><p>Gets states from all devices.</p>
<p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span> <span class="pre">out2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span> <span class="pre">out2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the states
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>States</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.init_optimizer">
<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">force_init=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.init_optimizer" title="Permalink to this definition"></a></dt>
<dd><p>Installs and initializes optimizers.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Default <cite>‘local’</cite>.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Default <cite>‘sgd’</cite></p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Default <cite>((‘learning_rate’, 0.01),)</cite>. The default value is not a dictionary,
just to avoid pylint warning of dangerous default values.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Default <code class="docutils literal notranslate"><span class="pre">False</span></code>, indicating whether we should force re-initializing the
optimizer in the case an optimizer is already installed.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.init_params">
<code class="sig-name descname">init_params</code><span class="sig-paren">(</span><em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.init_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.init_params" title="Permalink to this definition"></a></dt>
<dd><p>Initializes the parameters and auxiliary states.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Called to initialize parameters if needed.</p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing arg_params. Initialization
will be copied from that.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing aux_params. Initialization
will be copied from that.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.install_monitor">
<code class="sig-name descname">install_monitor</code><span class="sig-paren">(</span><em class="sig-param">mon</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.install_monitor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.install_monitor" title="Permalink to this definition"></a></dt>
<dd><p>Installs monitor on all executors.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.label_names">
<em class="property">property </em><code class="sig-name descname">label_names</code><a class="headerlink" href="#mxnet.module.Module.label_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for labels required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.label_shapes">
<em class="property">property </em><code class="sig-name descname">label_shapes</code><a class="headerlink" href="#mxnet.module.Module.label_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets label shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>The return value could be <code class="docutils literal notranslate"><span class="pre">None</span></code> if
the module does not need labels, or if the module is not bound for
training (in this case, label information is not available).</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.load">
<em class="property">static </em><code class="sig-name descname">load</code><span class="sig-paren">(</span><em class="sig-param">prefix</em>, <em class="sig-param">epoch</em>, <em class="sig-param">load_optimizer_states=False</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.load"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.load" title="Permalink to this definition"></a></dt>
<dd><p>Creates a model from previously saved checkpoint.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>prefix</strong> (<em>str</em>) – path prefix of saved model files. You should have
“prefix-symbol.json”, “prefix-xxxx.params”, and
optionally “prefix-xxxx.states”, where xxxx is the
epoch number.</p></li>
<li><p><strong>epoch</strong> (<em>int</em>) – epoch to load.</p></li>
<li><p><strong>load_optimizer_states</strong> (<em>bool</em>) – whether to load optimizer states. Checkpoint needs
to have been made with save_optimizer_states=True.</p></li>
<li><p><strong>data_names</strong> (<em>list of str</em>) – Default is <cite>(‘data’)</cite> for a typical model used in image classification.</p></li>
<li><p><strong>label_names</strong> (<em>list of str</em>) – Default is <cite>(‘softmax_label’)</cite> for a typical model used in image
classification.</p></li>
<li><p><strong>logger</strong> (<em>Logger</em>) – Default is <cite>logging</cite>.</p></li>
<li><p><strong>context</strong> (<a class="reference internal" href="../mxnet/context/index.html#mxnet.context.Context" title="mxnet.context.Context"><em>Context</em></a><em> or </em><em>list of Context</em>) – Default is <code class="docutils literal notranslate"><span class="pre">cpu()</span></code>.</p></li>
<li><p><strong>work_load_list</strong> (<em>list of number</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating uniform workload.</p></li>
<li><p><strong>fixed_param_names</strong> (<em>list of str</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>, indicating no network parameters are fixed.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.load_optimizer_states">
<code class="sig-name descname">load_optimizer_states</code><span class="sig-paren">(</span><em class="sig-param">fname</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.load_optimizer_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.load_optimizer_states" title="Permalink to this definition"></a></dt>
<dd><p>Loads optimizer (updater) state from a file.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>fname</strong> (<em>str</em>) – Path to input states file.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.output_names">
<em class="property">property </em><code class="sig-name descname">output_names</code><a class="headerlink" href="#mxnet.module.Module.output_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.output_shapes">
<em class="property">property </em><code class="sig-name descname">output_shapes</code><a class="headerlink" href="#mxnet.module.Module.output_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets output shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p></p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>A list of <cite>(name, shape)</cite> pairs.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.prepare">
<code class="sig-name descname">prepare</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">sparse_row_id_fn=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.prepare"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.prepare" title="Permalink to this definition"></a></dt>
<dd><p>Prepares the module for processing a data batch.</p>
<p>Usually involves switching bucket and reshaping.
For modules that contain <cite>row_sparse</cite> parameters in KVStore,
it prepares the <cite>row_sparse</cite> parameters based on the sparse_row_id_fn.</p>
<p>When KVStore is used to update parameters for multi-device or multi-machine training,
a copy of the parameters are stored in KVStore. Note that for <cite>row_sparse</cite> parameters,
the <cite>update()</cite> updates the copy of parameters in KVStore, but doesn’t broadcast
the updated parameters to all devices / machines. The <cite>prepare</cite> function is used to
broadcast <cite>row_sparse</cite> parameters with the next batch of data.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – The current batch of data for forward computation.</p></li>
<li><p><strong>sparse_row_id_fn</strong> (<em>A callback function</em>) – The function takes <cite>data_batch</cite> as an input and returns a dict of
str -&gt; NDArray. The resulting dict is used for pulling row_sparse
parameters from the kvstore, where the str key is the name of the param,
and the value is the row id of the param to pull.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.reshape">
<code class="sig-name descname">reshape</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.reshape"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.reshape" title="Permalink to this definition"></a></dt>
<dd><p>Reshapes the module for new input shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_data</span></code>.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_label</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.save_checkpoint">
<code class="sig-name descname">save_checkpoint</code><span class="sig-paren">(</span><em class="sig-param">prefix</em>, <em class="sig-param">epoch</em>, <em class="sig-param">save_optimizer_states=False</em>, <em class="sig-param">remove_amp_cast=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.save_checkpoint"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.save_checkpoint" title="Permalink to this definition"></a></dt>
<dd><p>Saves current progress to checkpoint.
Use <cite>mx.callback.module_checkpoint</cite> as <cite>epoch_end_callback</cite> to save during training.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>prefix</strong> (<em>str</em>) – The file prefix to checkpoint to.</p></li>
<li><p><strong>epoch</strong> (<em>int</em>) – The current epoch number.</p></li>
<li><p><strong>save_optimizer_states</strong> (<em>bool</em>) – Whether to save optimizer states to continue training.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.save_optimizer_states">
<code class="sig-name descname">save_optimizer_states</code><span class="sig-paren">(</span><em class="sig-param">fname</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.save_optimizer_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.save_optimizer_states" title="Permalink to this definition"></a></dt>
<dd><p>Saves optimizer (updater) state to a file.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>fname</strong> (<em>str</em>) – Path to output states file.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.set_params">
<code class="sig-name descname">set_params</code><span class="sig-paren">(</span><em class="sig-param">arg_params</em>, <em class="sig-param">aux_params</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=True</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.set_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.set_params" title="Permalink to this definition"></a></dt>
<dd><p>Assigns parameter and aux state values.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>arg_params</strong> (<em>dict</em>) – Dictionary of name to <cite>NDArray</cite>.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Dictionary of name to <cite>NDArray</cite>.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of setting module parameters.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">sym</span><span class="p">,</span> <span class="n">arg_params</span><span class="p">,</span> <span class="n">aux_params</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">model</span><span class="o">.</span><span class="n">load_checkpoint</span><span class="p">(</span><span class="n">model_prefix</span><span class="p">,</span> <span class="n">n_epoch_load</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">mod</span><span class="o">.</span><span class="n">set_params</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>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.set_states">
<code class="sig-name descname">set_states</code><span class="sig-paren">(</span><em class="sig-param">states=None</em>, <em class="sig-param">value=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.set_states"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.set_states" title="Permalink to this definition"></a></dt>
<dd><p>Sets value for states. Only one of the states &amp; value can be specified.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>states</strong> (<em>list of list of NDArrays</em>) – source states arrays formatted like <code class="docutils literal notranslate"><span class="pre">[[state1_dev1,</span> <span class="pre">state1_dev2],</span>
<span class="pre">[state2_dev1,</span> <span class="pre">state2_dev2]]</span></code>.</p></li>
<li><p><strong>value</strong> (<em>number</em>) – a single scalar value for all state arrays.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.update">
<code class="sig-name descname">update</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.update"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.update" title="Permalink to this definition"></a></dt>
<dd><p>Updates parameters according to the installed optimizer and the gradients computed
in the previous forward-backward batch.</p>
<p>When KVStore is used to update parameters for multi-device or multi-machine training,
a copy of the parameters are stored in KVStore. Note that for <cite>row_sparse</cite> parameters,
this function does update the copy of parameters in KVStore, but doesn’t broadcast the
updated parameters to all devices / machines. Please call <cite>prepare</cite> to broadcast
<cite>row_sparse</cite> parameters with the next batch of data.</p>
<div class="admonition seealso">
<p class="admonition-title">See also</p>
<p><a class="reference internal" href="#mxnet.module.BaseModule.update" title="mxnet.module.BaseModule.update"><code class="xref py py-meth docutils literal notranslate"><span class="pre">BaseModule.update()</span></code></a></p>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.Module.update_metric">
<code class="sig-name descname">update_metric</code><span class="sig-paren">(</span><em class="sig-param">eval_metric</em>, <em class="sig-param">labels</em>, <em class="sig-param">pre_sliced=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/module.html#Module.update_metric"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.Module.update_metric" title="Permalink to this definition"></a></dt>
<dd><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p>
<div class="admonition seealso">
<p class="admonition-title">See also</p>
<p><a class="reference internal" href="#mxnet.module.BaseModule.update_metric" title="mxnet.module.BaseModule.update_metric"><code class="xref py py-meth docutils literal notranslate"><span class="pre">BaseModule.update_metric()</span></code></a></p>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – Evaluation metric to use.</p></li>
<li><p><strong>labels</strong> (list of NDArray if <cite>pre_sliced</cite> parameter is set to <cite>False</cite>,) – list of lists of NDArray otherwise. Typically <cite>data_batch.label</cite>.</p></li>
<li><p><strong>pre_sliced</strong> (<em>bool</em>) – Whether the labels are already sliced per device (default: False).</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.module.PythonLossModule">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">PythonLossModule</code><span class="sig-paren">(</span><em class="sig-param">name='pyloss'</em>, <em class="sig-param">data_names=('data'</em>, <em class="sig-param">)</em>, <em class="sig-param">label_names=('softmax_label'</em>, <em class="sig-param">)</em>, <em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em>, <em class="sig-param">grad_func=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.module.python_module.PythonModule</span></code></p>
<p>A convenient module class that implements many of the module APIs as
empty functions.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>name</strong> (<em>str</em>) – Names of the module. The outputs will be named <cite>[name + ‘_output’]</cite>.</p></li>
<li><p><strong>data_names</strong> (<em>list of str</em>) – Defaults to <code class="docutils literal notranslate"><span class="pre">['data']</span></code>. Names of the data expected by this module.
Should be a list of only one name.</p></li>
<li><p><strong>label_names</strong> (<em>list of str</em>) – Default <code class="docutils literal notranslate"><span class="pre">['softmax_label']</span></code>. Names of the labels expected by the module.
Should be a list of only one name.</p></li>
<li><p><strong>grad_func</strong> (<em>function</em>) – Optional. If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a function that takes <cite>scores</cite>
and <cite>labels</cite>, both of type <cite>NDArray</cite>, and return the gradients with
respect to the scores according to this loss function. The return
value could be a numpy array or an <cite>NDArray</cite>.</p></li>
</ul>
</dd>
</dl>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule.backward" title="mxnet.module.PythonLossModule.backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">backward</span></code></a>([out_grads])</p></td>
<td><p>Backward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule.forward" title="mxnet.module.PythonLossModule.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(data_batch[, is_train])</p></td>
<td><p>Forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule.get_input_grads" title="mxnet.module.PythonLossModule.get_input_grads"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_input_grads</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets the gradients to the inputs, computed in the previous backward computation.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule.get_outputs" title="mxnet.module.PythonLossModule.get_outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_outputs</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets outputs of the previous forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonLossModule.install_monitor" title="mxnet.module.PythonLossModule.install_monitor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">install_monitor</span></code></a>(mon)</p></td>
<td><p>Installs monitor on all executors.</p></td>
</tr>
</tbody>
</table>
<dl class="method">
<dt id="mxnet.module.PythonLossModule.backward">
<code class="sig-name descname">backward</code><span class="sig-paren">(</span><em class="sig-param">out_grads=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule.backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule.backward" title="Permalink to this definition"></a></dt>
<dd><p>Backward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>out_grads</strong> (<a class="reference internal" href="../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a><em> or </em><em>list of NDArray</em><em>, </em><em>optional</em>) – Gradient on the outputs to be propagated back.
This parameter is only needed when bind is called
on outputs that are not a loss function.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonLossModule.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">is_train=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule.forward" title="Permalink to this definition"></a></dt>
<dd><p>Forward computation. Here we do nothing but to keep a reference to
the scores and the labels so that we can do backward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – Could be anything with similar API implemented.</p></li>
<li><p><strong>is_train</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>, which means <cite>is_train</cite> takes the value of <code class="docutils literal notranslate"><span class="pre">self.for_training</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonLossModule.get_input_grads">
<code class="sig-name descname">get_input_grads</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule.get_input_grads"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule.get_input_grads" title="Permalink to this definition"></a></dt>
<dd><p>Gets the gradients to the inputs, computed in the previous backward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Should always be <code class="docutils literal notranslate"><span class="pre">True</span></code> because we do not use multiple context for computation.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonLossModule.get_outputs">
<code class="sig-name descname">get_outputs</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule.get_outputs"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule.get_outputs" title="Permalink to this definition"></a></dt>
<dd><p>Gets outputs of the previous forward computation. As a output loss module,
we treat the inputs to this module as scores, and simply return them.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Should always be <code class="docutils literal notranslate"><span class="pre">True</span></code>, because we do not use multiple contexts for computing.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonLossModule.install_monitor">
<code class="sig-name descname">install_monitor</code><span class="sig-paren">(</span><em class="sig-param">mon</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonLossModule.install_monitor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonLossModule.install_monitor" title="Permalink to this definition"></a></dt>
<dd><p>Installs monitor on all executors.</p>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.module.PythonModule">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">PythonModule</code><span class="sig-paren">(</span><em class="sig-param">data_names</em>, <em class="sig-param">label_names</em>, <em class="sig-param">output_names</em>, <em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.module.base_module.BaseModule</span></code></p>
<p>A convenient module class that implements many of the module APIs as
empty functions.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_names</strong> (<em>list of str</em>) – Names of the data expected by the module.</p></li>
<li><p><strong>label_names</strong> (<em>list of str</em>) – Names of the labels expected by the module. Could be <code class="docutils literal notranslate"><span class="pre">None</span></code> if the
module does not need labels.</p></li>
<li><p><strong>output_names</strong> (<em>list of str</em>) – Names of the outputs.</p></li>
</ul>
</dd>
</dl>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.bind" title="mxnet.module.PythonModule.bind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">bind</span></code></a>(data_shapes[, label_shapes, …])</p></td>
<td><p>Binds the symbols to construct executors.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.get_params" title="mxnet.module.PythonModule.get_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_params</span></code></a>()</p></td>
<td><p>Gets parameters, those are potentially copies of the actual parameters used to do computation on the device.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.init_optimizer" title="mxnet.module.PythonModule.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>([kvstore, optimizer, …])</p></td>
<td><p>Installs and initializes optimizers.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.init_params" title="mxnet.module.PythonModule.init_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_params</span></code></a>([initializer, arg_params, …])</p></td>
<td><p>Initializes the parameters and auxiliary states.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.update" title="mxnet.module.PythonModule.update"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update</span></code></a>()</p></td>
<td><p>Updates parameters according to the installed optimizer and the gradients computed in the previous forward-backward batch.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.update_metric" title="mxnet.module.PythonModule.update_metric"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_metric</span></code></a>(eval_metric, labels[, pre_sliced])</p></td>
<td><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p></td>
</tr>
</tbody>
</table>
<p><strong>Attributes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.data_names" title="mxnet.module.PythonModule.data_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_names</span></code></a></p></td>
<td><p>A list of names for data required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.data_shapes" title="mxnet.module.PythonModule.data_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the data inputs to this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.label_shapes" title="mxnet.module.PythonModule.label_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the label inputs to this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.output_names" title="mxnet.module.PythonModule.output_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_names</span></code></a></p></td>
<td><p>A list of names for the outputs of this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.PythonModule.output_shapes" title="mxnet.module.PythonModule.output_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_shapes</span></code></a></p></td>
<td><p>A list of (name, shape) pairs specifying the outputs of this module.</p></td>
</tr>
</tbody>
</table>
<dl class="method">
<dt id="mxnet.module.PythonModule.bind">
<code class="sig-name descname">bind</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em>, <em class="sig-param">for_training=True</em>, <em class="sig-param">inputs_need_grad=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">shared_module=None</em>, <em class="sig-param">grad_req='write'</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.bind"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.bind" title="Permalink to this definition"></a></dt>
<dd><p>Binds the symbols to construct executors. This is necessary before one
can perform computation with the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_data</span></code>.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <code class="docutils literal notranslate"><span class="pre">data_iter.provide_label</span></code>.</p></li>
<li><p><strong>for_training</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. Whether the executors should be bind for training.</p></li>
<li><p><strong>inputs_need_grad</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. Whether the gradients to the input data need to be computed.
Typically this is not needed. But this might be needed when implementing composition
of modules.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. This function does nothing if the executors are already
bound. But with this <code class="docutils literal notranslate"><span class="pre">True</span></code>, the executors will be forced to rebind.</p></li>
<li><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><em>Module</em></a>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>. This is used in bucketing. When not <code class="docutils literal notranslate"><span class="pre">None</span></code>, the shared module
essentially corresponds to a different bucket – a module with different symbol
but with the same sets of parameters (e.g. unrolled RNNs with different lengths).</p></li>
<li><p><strong>grad_req</strong> (<em>str</em><em>, </em><em>list of str</em><em>, </em><em>dict of str to str</em>) – Requirement for gradient accumulation. Can be ‘write’, ‘add’, or ‘null’
(default to ‘write’).
Can be specified globally (str) or for each argument (list, dict).</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.data_names">
<em class="property">property </em><code class="sig-name descname">data_names</code><a class="headerlink" href="#mxnet.module.PythonModule.data_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for data required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.data_shapes">
<em class="property">property </em><code class="sig-name descname">data_shapes</code><a class="headerlink" href="#mxnet.module.PythonModule.data_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the data inputs to this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.get_params">
<code class="sig-name descname">get_params</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.get_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.get_params" title="Permalink to this definition"></a></dt>
<dd><p>Gets parameters, those are potentially copies of the actual parameters used
to do computation on the device. Subclass should override this method if contains
parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p></p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p><code class="docutils literal notranslate"><span class="pre">({},</span> <span class="pre">{})</span></code>, a pair of empty dict.</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.init_optimizer">
<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">force_init=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.init_optimizer" title="Permalink to this definition"></a></dt>
<dd><p>Installs and initializes optimizers. By default we do nothing. Subclass should
override this method if needed.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Default <cite>‘local’</cite>.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Default <cite>‘sgd’</cite></p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Default <cite>((‘learning_rate’, 0.01),)</cite>. The default value is not a dictionary,
just to avoid pylint warning of dangerous default values.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Default <cite>False</cite>, indicating whether we should force re-initializing the
optimizer in the case an optimizer is already installed.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.init_params">
<code class="sig-name descname">init_params</code><span class="sig-paren">(</span><em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.init_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.init_params" title="Permalink to this definition"></a></dt>
<dd><p>Initializes the parameters and auxiliary states. By default this function
does nothing. Subclass should override this method if contains parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Called to initialize parameters if needed.</p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing <cite>arg_params</cite>. Initialization
will be copied from that.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – If not <code class="docutils literal notranslate"><span class="pre">None</span></code>, should be a dictionary of existing <cite>aux_params</cite>. Initialization
will be copied from that.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, params could contain missing values, and the initializer will be
called to fill those missing params.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – If <code class="docutils literal notranslate"><span class="pre">True</span></code>, will force re-initialize even if already initialized.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.label_shapes">
<em class="property">property </em><code class="sig-name descname">label_shapes</code><a class="headerlink" href="#mxnet.module.PythonModule.label_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the label inputs to this module.
If this module does not accept labels – either it is a module without loss
function, or it is not bound for training, then this should return an empty
list <code class="docutils literal notranslate"><span class="pre">[]`</span></code>.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.output_names">
<em class="property">property </em><code class="sig-name descname">output_names</code><a class="headerlink" href="#mxnet.module.PythonModule.output_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.output_shapes">
<em class="property">property </em><code class="sig-name descname">output_shapes</code><a class="headerlink" href="#mxnet.module.PythonModule.output_shapes" title="Permalink to this definition"></a></dt>
<dd><p>A list of (name, shape) pairs specifying the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.update">
<code class="sig-name descname">update</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.update"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.update" title="Permalink to this definition"></a></dt>
<dd><p>Updates parameters according to the installed optimizer and the gradients computed
in the previous forward-backward batch. Currently we do nothing here. Subclass should
override this method if contains parameters.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.PythonModule.update_metric">
<code class="sig-name descname">update_metric</code><span class="sig-paren">(</span><em class="sig-param">eval_metric</em>, <em class="sig-param">labels</em>, <em class="sig-param">pre_sliced=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/python_module.html#PythonModule.update_metric"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.PythonModule.update_metric" title="Permalink to this definition"></a></dt>
<dd><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.
Subclass should override this method if needed.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – </p></li>
<li><p><strong>labels</strong> (<em>list of NDArray</em>) – Typically <code class="docutils literal notranslate"><span class="pre">data_batch.label</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.module.SequentialModule">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.module.</code><code class="sig-name descname">SequentialModule</code><span class="sig-paren">(</span><em class="sig-param">logger=&lt;module 'logging' from '/work/conda_env/lib/python3.9/logging/__init__.py'&gt;</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.module.base_module.BaseModule</span></code></p>
<p>A SequentialModule is a container module that can chain multiple modules together.</p>
<div class="admonition note">
<p class="admonition-title">Note</p>
<p>Building a computation graph with this kind of imperative container is less
flexible and less efficient than the symbolic graph. So, this should be only used as a
handy utility.</p>
</div>
<p><strong>Methods</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.add" title="mxnet.module.SequentialModule.add"><code class="xref py py-obj docutils literal notranslate"><span class="pre">add</span></code></a>(module, **kwargs)</p></td>
<td><p>Add a module to the chain.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.backward" title="mxnet.module.SequentialModule.backward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">backward</span></code></a>([out_grads])</p></td>
<td><p>Backward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.bind" title="mxnet.module.SequentialModule.bind"><code class="xref py py-obj docutils literal notranslate"><span class="pre">bind</span></code></a>(data_shapes[, label_shapes, …])</p></td>
<td><p>Binds the symbols to construct executors.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.forward" title="mxnet.module.SequentialModule.forward"><code class="xref py py-obj docutils literal notranslate"><span class="pre">forward</span></code></a>(data_batch[, is_train])</p></td>
<td><p>Forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.get_input_grads" title="mxnet.module.SequentialModule.get_input_grads"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_input_grads</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets the gradients with respect to the inputs of the module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.get_outputs" title="mxnet.module.SequentialModule.get_outputs"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_outputs</span></code></a>([merge_multi_context])</p></td>
<td><p>Gets outputs from a previous forward computation.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.get_params" title="mxnet.module.SequentialModule.get_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">get_params</span></code></a>()</p></td>
<td><p>Gets current parameters.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.init_optimizer" title="mxnet.module.SequentialModule.init_optimizer"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_optimizer</span></code></a>([kvstore, optimizer, …])</p></td>
<td><p>Installs and initializes optimizers.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.init_params" title="mxnet.module.SequentialModule.init_params"><code class="xref py py-obj docutils literal notranslate"><span class="pre">init_params</span></code></a>([initializer, arg_params, …])</p></td>
<td><p>Initializes parameters.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.install_monitor" title="mxnet.module.SequentialModule.install_monitor"><code class="xref py py-obj docutils literal notranslate"><span class="pre">install_monitor</span></code></a>(mon)</p></td>
<td><p>Installs monitor on all executors.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.update" title="mxnet.module.SequentialModule.update"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update</span></code></a>()</p></td>
<td><p>Updates parameters according to installed optimizer and the gradient computed in the previous forward-backward cycle.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.update_metric" title="mxnet.module.SequentialModule.update_metric"><code class="xref py py-obj docutils literal notranslate"><span class="pre">update_metric</span></code></a>(eval_metric, labels[, pre_sliced])</p></td>
<td><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p></td>
</tr>
</tbody>
</table>
<p><strong>Attributes</strong></p>
<table class="longtable docutils align-default">
<colgroup>
<col style="width: 10%" />
<col style="width: 90%" />
</colgroup>
<tbody>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.data_names" title="mxnet.module.SequentialModule.data_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_names</span></code></a></p></td>
<td><p>A list of names for data required by this module.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.data_shapes" title="mxnet.module.SequentialModule.data_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">data_shapes</span></code></a></p></td>
<td><p>Gets data shapes.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.label_shapes" title="mxnet.module.SequentialModule.label_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">label_shapes</span></code></a></p></td>
<td><p>Gets label shapes.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.output_names" title="mxnet.module.SequentialModule.output_names"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_names</span></code></a></p></td>
<td><p>A list of names for the outputs of this module.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.module.SequentialModule.output_shapes" title="mxnet.module.SequentialModule.output_shapes"><code class="xref py py-obj docutils literal notranslate"><span class="pre">output_shapes</span></code></a></p></td>
<td><p>Gets output shapes.</p></td>
</tr>
</tbody>
</table>
<dl class="method">
<dt id="mxnet.module.SequentialModule.add">
<code class="sig-name descname">add</code><span class="sig-paren">(</span><em class="sig-param">module</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.add"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.add" title="Permalink to this definition"></a></dt>
<dd><p>Add a module to the chain.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>module</strong> (<a class="reference internal" href="#mxnet.module.BaseModule" title="mxnet.module.BaseModule"><em>BaseModule</em></a>) – The new module to add.</p></li>
<li><p><strong>kwargs</strong> (<code class="docutils literal notranslate"><span class="pre">**keywords</span></code>) – <p>All the keyword arguments are saved as meta information
for the added module. The currently known meta includes</p>
<ul>
<li><dl class="simple">
<dt><cite>take_labels</cite>: indicating whether the module expect to</dt><dd><p>take labels when doing computation. Note any module in
the chain can take labels (not necessarily only the top
most one), and they all take the same labels passed
from the original data batch for the <cite>SequentialModule</cite>.</p>
</dd>
</dl>
</li>
</ul>
</p></li>
</ul>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>This function returns <cite>self</cite> to allow us to easily chain a
series of <cite>add</cite> calls.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>self</p>
</dd>
</dl>
<p class="rubric">Examples</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="c1"># An example of addinging two modules to a chain.</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">seq_mod</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">mod</span><span class="o">.</span><span class="n">SequentialModule</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">seq_mod</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mod1</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">seq_mod</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">mod2</span><span class="p">)</span>
</pre></div>
</div>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.backward">
<code class="sig-name descname">backward</code><span class="sig-paren">(</span><em class="sig-param">out_grads=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.backward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.backward" title="Permalink to this definition"></a></dt>
<dd><p>Backward computation.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.bind">
<code class="sig-name descname">bind</code><span class="sig-paren">(</span><em class="sig-param">data_shapes</em>, <em class="sig-param">label_shapes=None</em>, <em class="sig-param">for_training=True</em>, <em class="sig-param">inputs_need_grad=False</em>, <em class="sig-param">force_rebind=False</em>, <em class="sig-param">shared_module=None</em>, <em class="sig-param">grad_req='write'</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.bind"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.bind" title="Permalink to this definition"></a></dt>
<dd><p>Binds the symbols to construct executors. This is necessary before one
can perform computation with the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <cite>data_iter.provide_data</cite>.</p></li>
<li><p><strong>label_shapes</strong> (<em>list of</em><em> (</em><em>str</em><em>, </em><em>tuple</em><em>)</em>) – Typically is <cite>data_iter.provide_label</cite>.</p></li>
<li><p><strong>for_training</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. Whether the executors should be bind for training.</p></li>
<li><p><strong>inputs_need_grad</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. Whether the gradients to the input data need to be computed.
Typically this is not needed. But this might be needed when implementing composition
of modules.</p></li>
<li><p><strong>force_rebind</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">False</span></code>. This function does nothing if the executors are already
bound. But with this <code class="docutils literal notranslate"><span class="pre">True</span></code>, the executors will be forced to rebind.</p></li>
<li><p><strong>shared_module</strong> (<a class="reference internal" href="#mxnet.module.Module" title="mxnet.module.Module"><em>Module</em></a>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>. Currently shared module is not supported for <cite>SequentialModule</cite>.</p></li>
<li><p><strong>grad_req</strong> (<em>str</em><em>, </em><em>list of str</em><em>, </em><em>dict of str to str</em>) – Requirement for gradient accumulation. Can be ‘write’, ‘add’, or ‘null’
(default to ‘write’).
Can be specified globally (str) or for each argument (list, dict).</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.data_names">
<em class="property">property </em><code class="sig-name descname">data_names</code><a class="headerlink" href="#mxnet.module.SequentialModule.data_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for data required by this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.data_shapes">
<em class="property">property </em><code class="sig-name descname">data_shapes</code><a class="headerlink" href="#mxnet.module.SequentialModule.data_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets data shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A list of <cite>(name, shape)</cite> pairs. The data shapes of the first module
is the data shape of a <cite>SequentialModule</cite>.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>list</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">data_batch</em>, <em class="sig-param">is_train=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.forward" title="Permalink to this definition"></a></dt>
<dd><p>Forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>data_batch</strong> (<a class="reference internal" href="../mxnet/io/index.html#mxnet.io.DataBatch" title="mxnet.io.DataBatch"><em>DataBatch</em></a>) – </p></li>
<li><p><strong>is_train</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">None</span></code>, in which case <cite>is_train</cite> is take as <code class="docutils literal notranslate"><span class="pre">self.for_training</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.get_input_grads">
<code class="sig-name descname">get_input_grads</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.get_input_grads"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.get_input_grads" title="Permalink to this definition"></a></dt>
<dd><p>Gets the gradients with respect to the inputs of the module.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[grad1,</span> <span class="pre">grad2]</span></code>. Otherwise, it
is like <code class="docutils literal notranslate"><span class="pre">[[grad1_dev1,</span> <span class="pre">grad1_dev2],</span> <span class="pre">[grad2_dev1,</span> <span class="pre">grad2_dev2]]</span></code>. All the output
elements are <cite>NDArray</cite>.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArrays or list of list of NDArrays</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.get_outputs">
<code class="sig-name descname">get_outputs</code><span class="sig-paren">(</span><em class="sig-param">merge_multi_context=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.get_outputs"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.get_outputs" title="Permalink to this definition"></a></dt>
<dd><p>Gets outputs from a previous forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>merge_multi_context</strong> (<em>bool</em>) – Default is <code class="docutils literal notranslate"><span class="pre">True</span></code>. In the case when data-parallelism is used, the outputs
will be collected from multiple devices. A <code class="docutils literal notranslate"><span class="pre">True</span></code> value indicate that we
should merge the collected results so that they look like from a single
executor.</p>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p>If <cite>merge_multi_context</cite> is <code class="docutils literal notranslate"><span class="pre">True</span></code>, it is like <code class="docutils literal notranslate"><span class="pre">[out1,</span>
<span class="pre">out2]</span></code>. Otherwise, it is like <code class="docutils literal notranslate"><span class="pre">[[out1_dev1,</span> <span class="pre">out1_dev2],</span> <span class="pre">[out2_dev1,</span>
<span class="pre">out2_dev2]]</span></code>. All the output elements are numpy arrays.</p>
</dd>
<dt class="field-odd">Return type</dt>
<dd class="field-odd"><p>list of NDArray or list of list of NDArray</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.get_params">
<code class="sig-name descname">get_params</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.get_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.get_params" title="Permalink to this definition"></a></dt>
<dd><p>Gets current parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A pair of dictionaries each mapping parameter names to NDArray values. This
is a merged dictionary of all the parameters in the modules.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>(arg_params, aux_params)</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.init_optimizer">
<code class="sig-name descname">init_optimizer</code><span class="sig-paren">(</span><em class="sig-param">kvstore='local'</em>, <em class="sig-param">optimizer='sgd'</em>, <em class="sig-param">optimizer_params=(('learning_rate'</em>, <em class="sig-param">0.01)</em>, <em class="sig-param">)</em>, <em class="sig-param">force_init=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.init_optimizer"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.init_optimizer" title="Permalink to this definition"></a></dt>
<dd><p>Installs and initializes optimizers.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>kvstore</strong> (<em>str</em><em> or </em><em>KVStore</em>) – Default <cite>‘local’</cite>.</p></li>
<li><p><strong>optimizer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../optimizer/index.html#mxnet.optimizer.Optimizer" title="mxnet.optimizer.Optimizer"><em>Optimizer</em></a>) – Default <cite>‘sgd’</cite></p></li>
<li><p><strong>optimizer_params</strong> (<em>dict</em>) – Default <code class="docutils literal notranslate"><span class="pre">(('learning_rate',</span> <span class="pre">0.01),)</span></code>. The default value is not a dictionary,
just to avoid pylint warning of dangerous default values.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Default <code class="docutils literal notranslate"><span class="pre">False</span></code>, indicating whether we should force re-initializing the
optimizer in the case an optimizer is already installed.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.init_params">
<code class="sig-name descname">init_params</code><span class="sig-paren">(</span><em class="sig-param">initializer=&lt;mxnet.initializer.Uniform object&gt;</em>, <em class="sig-param">arg_params=None</em>, <em class="sig-param">aux_params=None</em>, <em class="sig-param">allow_missing=False</em>, <em class="sig-param">force_init=False</em>, <em class="sig-param">allow_extra=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.init_params"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.init_params" title="Permalink to this definition"></a></dt>
<dd><p>Initializes parameters.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>initializer</strong> (<a class="reference internal" href="../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – </p></li>
<li><p><strong>arg_params</strong> (<em>dict</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>. Existing parameters. This has higher priority
than <cite>initializer</cite>.</p></li>
<li><p><strong>aux_params</strong> (<em>dict</em>) – Default <code class="docutils literal notranslate"><span class="pre">None</span></code>. Existing auxiliary states. This has higher priority
than <cite>initializer</cite>.</p></li>
<li><p><strong>allow_missing</strong> (<em>bool</em>) – Allow missing values in <cite>arg_params</cite> and <cite>aux_params</cite> (if not <code class="docutils literal notranslate"><span class="pre">None</span></code>).
In this case, missing values will be filled with <cite>initializer</cite>.</p></li>
<li><p><strong>force_init</strong> (<em>bool</em>) – Default <code class="docutils literal notranslate"><span class="pre">False</span></code>.</p></li>
<li><p><strong>allow_extra</strong> (<em>boolean</em><em>, </em><em>optional</em>) – Whether allow extra parameters that are not needed by symbol.
If this is True, no error will be thrown when arg_params or aux_params
contain extra parameters that is not needed by the executor.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.install_monitor">
<code class="sig-name descname">install_monitor</code><span class="sig-paren">(</span><em class="sig-param">mon</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.install_monitor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.install_monitor" title="Permalink to this definition"></a></dt>
<dd><p>Installs monitor on all executors.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.label_shapes">
<em class="property">property </em><code class="sig-name descname">label_shapes</code><a class="headerlink" href="#mxnet.module.SequentialModule.label_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets label shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A list of <cite>(name, shape)</cite> pairs. The return value could be <cite>None</cite> if
the module does not need labels, or if the module is not bound for
training (in this case, label information is not available).</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>list</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.output_names">
<em class="property">property </em><code class="sig-name descname">output_names</code><a class="headerlink" href="#mxnet.module.SequentialModule.output_names" title="Permalink to this definition"></a></dt>
<dd><p>A list of names for the outputs of this module.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.output_shapes">
<em class="property">property </em><code class="sig-name descname">output_shapes</code><a class="headerlink" href="#mxnet.module.SequentialModule.output_shapes" title="Permalink to this definition"></a></dt>
<dd><p>Gets output shapes.</p>
<dl class="field-list simple">
<dt class="field-odd">Returns</dt>
<dd class="field-odd"><p>A list of <cite>(name, shape)</cite> pairs. The output shapes of the last
module is the output shape of a <cite>SequentialModule</cite>.</p>
</dd>
<dt class="field-even">Return type</dt>
<dd class="field-even"><p>list</p>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.update">
<code class="sig-name descname">update</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.update"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.update" title="Permalink to this definition"></a></dt>
<dd><p>Updates parameters according to installed optimizer and the gradient computed
in the previous forward-backward cycle.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.module.SequentialModule.update_metric">
<code class="sig-name descname">update_metric</code><span class="sig-paren">(</span><em class="sig-param">eval_metric</em>, <em class="sig-param">labels</em>, <em class="sig-param">pre_sliced=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../_modules/mxnet/module/sequential_module.html#SequentialModule.update_metric"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.module.SequentialModule.update_metric" title="Permalink to this definition"></a></dt>
<dd><p>Evaluates and accumulates evaluation metric on outputs of the last forward computation.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>eval_metric</strong> (<a class="reference internal" href="../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – </p></li>
<li><p><strong>labels</strong> (<em>list of NDArray</em>) – Typically <code class="docutils literal notranslate"><span class="pre">data_batch.label</span></code>.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
</div>
<hr class="feedback-hr-top" />
<div class="feedback-container">
<div class="feedback-question">Did this page help you?</div>
<div class="feedback-answer-container">
<div class="feedback-answer yes-link" data-response="yes">Yes</div>
<div class="feedback-answer no-link" data-response="no">No</div>
</div>
<div class="feedback-thank-you">Thanks for your feedback!</div>
</div>
<hr class="feedback-hr-bottom" />
</div>
<div class="side-doc-outline">
<div class="side-doc-outline--content">
</div>
</div>
<div class="clearer"></div>
</div><div class="pagenation">
<a id="button-prev" href="../symbol/sparse/index.html" class="mdl-button mdl-js-button mdl-js-ripple-effect mdl-button--colored" role="botton" accesskey="P">
<i class="pagenation-arrow-L fas fa-arrow-left fa-lg"></i>
<div class="pagenation-text">
<span class="pagenation-direction">Previous</span>
<div>symbol.sparse</div>
</div>
</a>
<a id="button-next" href="../contrib/index.html" class="mdl-button mdl-js-button mdl-js-ripple-effect mdl-button--colored" role="botton" accesskey="N">
<i class="pagenation-arrow-R fas fa-arrow-right fa-lg"></i>
<div class="pagenation-text">
<span class="pagenation-direction">Next</span>
<div>mxnet.contrib</div>
</div>
</a>
</div>
<footer class="site-footer h-card">
<div class="wrapper">
<div class="row">
<div class="col-4">
<h4 class="footer-category-title">Resources</h4>
<ul class="contact-list">
<li><a class="u-email" href="mailto:dev@mxnet.apache.org">Dev list</a></li>
<li><a class="u-email" href="mailto:user@mxnet.apache.org">User mailing list</a></li>
<li><a href="https://cwiki.apache.org/confluence/display/MXNET/Apache+MXNet+Home">Developer Wiki</a></li>
<li><a href="https://issues.apache.org/jira/projects/MXNET/issues">Jira Tracker</a></li>
<li><a href="https://github.com/apache/incubator-mxnet/labels/Roadmap">Github Roadmap</a></li>
<li><a href="https://discuss.mxnet.io">MXNet Discuss forum</a></li>
<li><a href="/community/contribute">Contribute To MXNet</a></li>
</ul>
</div>
<div class="col-4"><ul class="social-media-list"><li><a href="https://github.com/apache/incubator-mxnet"><svg class="svg-icon"><use xlink:href="../../_static/minima-social-icons.svg#github"></use></svg> <span class="username">apache/incubator-mxnet</span></a></li><li><a href="https://www.twitter.com/apachemxnet"><svg class="svg-icon"><use xlink:href="../../_static/minima-social-icons.svg#twitter"></use></svg> <span class="username">apachemxnet</span></a></li><li><a href="https://youtube.com/apachemxnet"><svg class="svg-icon"><use xlink:href="../../_static/minima-social-icons.svg#youtube"></use></svg> <span class="username">apachemxnet</span></a></li></ul>
</div>
<div class="col-4 footer-text">
<p>A flexible and efficient library for deep learning.</p>
</div>
</div>
</div>
</footer>
<footer class="site-footer2">
<div class="wrapper">
<div class="row">
<div class="col-3">
<img src="../../_static/apache_incubator_logo.png" class="footer-logo col-2">
</div>
<div class="footer-bottom-warning col-9">
<p>Apache MXNet is an effort undergoing incubation at The Apache Software Foundation (ASF), <span style="font-weight:bold">sponsored by the <i>Apache Incubator</i></span>. 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>
</footer>
</body>
</html>