Google Git
Sign in
apache / mxnet-site / 6054ba467ff2ccce12789e38fd5604ffad8abd17 / . / versions / 1.8.0 / api / python / docs / api / gluon / contrib / index.html
blob: 64091f0f3b633136cec2ee65b6a3fa58188e6f5a [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>gluon.contrib &#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="gluon.data" href="../data/index.html" />
<link rel="prev" title="gluon.Trainer" href="../trainer.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" href="../index.html">mxnet.gluon</a><i class="material-icons">navigate_next</i>
<a class="mdl-navigation__link is-active">gluon.contrib</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/gluon/contrib/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 current"><a class="reference internal" href="../index.html">mxnet.gluon</a><ul class="current">
<li class="toctree-l3"><a class="reference internal" href="../block.html">gluon.Block</a></li>
<li class="toctree-l3"><a class="reference internal" href="../hybrid_block.html">gluon.HybridBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol_block.html">gluon.SymbolBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../constant.html">gluon.Constant</a></li>
<li class="toctree-l3"><a class="reference internal" href="../parameter.html">gluon.Parameter</a></li>
<li class="toctree-l3"><a class="reference internal" href="../parameter_dict.html">gluon.ParameterDict</a></li>
<li class="toctree-l3"><a class="reference internal" href="../trainer.html">gluon.Trainer</a></li>
<li class="toctree-l3 current"><a class="current reference internal" href="#">gluon.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../data/index.html">gluon.data</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../data/vision/index.html">data.vision</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../data/vision/datasets/index.html">vision.datasets</a></li>
<li class="toctree-l5"><a class="reference internal" href="../data/vision/transforms/index.html">vision.transforms</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../loss/index.html">gluon.loss</a></li>
<li class="toctree-l3"><a class="reference internal" href="../model_zoo/index.html">gluon.model_zoo.vision</a></li>
<li class="toctree-l3"><a class="reference internal" href="../nn/index.html">gluon.nn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../rnn/index.html">gluon.rnn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../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"><a class="reference internal" href="../../module/index.html">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 current"><a class="reference internal" href="../index.html">mxnet.gluon</a><ul class="current">
<li class="toctree-l3"><a class="reference internal" href="../block.html">gluon.Block</a></li>
<li class="toctree-l3"><a class="reference internal" href="../hybrid_block.html">gluon.HybridBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../symbol_block.html">gluon.SymbolBlock</a></li>
<li class="toctree-l3"><a class="reference internal" href="../constant.html">gluon.Constant</a></li>
<li class="toctree-l3"><a class="reference internal" href="../parameter.html">gluon.Parameter</a></li>
<li class="toctree-l3"><a class="reference internal" href="../parameter_dict.html">gluon.ParameterDict</a></li>
<li class="toctree-l3"><a class="reference internal" href="../trainer.html">gluon.Trainer</a></li>
<li class="toctree-l3 current"><a class="current reference internal" href="#">gluon.contrib</a></li>
<li class="toctree-l3"><a class="reference internal" href="../data/index.html">gluon.data</a><ul>
<li class="toctree-l4"><a class="reference internal" href="../data/vision/index.html">data.vision</a><ul>
<li class="toctree-l5"><a class="reference internal" href="../data/vision/datasets/index.html">vision.datasets</a></li>
<li class="toctree-l5"><a class="reference internal" href="../data/vision/transforms/index.html">vision.transforms</a></li>
</ul>
</li>
</ul>
</li>
<li class="toctree-l3"><a class="reference internal" href="../loss/index.html">gluon.loss</a></li>
<li class="toctree-l3"><a class="reference internal" href="../model_zoo/index.html">gluon.model_zoo.vision</a></li>
<li class="toctree-l3"><a class="reference internal" href="../nn/index.html">gluon.nn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../rnn/index.html">gluon.rnn</a></li>
<li class="toctree-l3"><a class="reference internal" href="../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"><a class="reference internal" href="../../module/index.html">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="gluon-contrib">
<h1>gluon.contrib<a class="headerlink" href="#gluon-contrib" title="Permalink to this headline"></a></h1>
<p>This document lists the contrib APIs in Gluon:</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="#module-mxnet.gluon.contrib" title="mxnet.gluon.contrib"><code class="xref py py-obj docutils literal notranslate"><span class="pre">mxnet.gluon.contrib</span></code></a></p></td>
<td><p>Contrib neural network module.</p></td>
</tr>
</tbody>
</table>
<p>The <cite>Gluon Contrib</cite> API, defined in the <cite>gluon.contrib</cite> package, provides
many useful experimental APIs for new features.
This is a place for the community to try out the new features,
so that feature contributors can receive feedback.</p>
<div class="admonition warning">
<p class="admonition-title">Warning</p>
<p>This package contains experimental APIs and may change in the near future.</p>
</div>
<p>In the rest of this document, we list routines provided by the <cite>gluon.contrib</cite> package.</p>
<div class="section" id="neural-network">
<h2>Neural Network<a class="headerlink" href="#neural-network" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.nn.Concurrent" title="mxnet.gluon.contrib.nn.Concurrent"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Concurrent</span></code></a></p></td>
<td><p>Lays <cite>Block</cite> s concurrently.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.HybridConcurrent" title="mxnet.gluon.contrib.nn.HybridConcurrent"><code class="xref py py-obj docutils literal notranslate"><span class="pre">HybridConcurrent</span></code></a></p></td>
<td><p>Lays <cite>HybridBlock</cite> s concurrently.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.Identity" title="mxnet.gluon.contrib.nn.Identity"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Identity</span></code></a></p></td>
<td><p>Block that passes through the input directly.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.SparseEmbedding" title="mxnet.gluon.contrib.nn.SparseEmbedding"><code class="xref py py-obj docutils literal notranslate"><span class="pre">SparseEmbedding</span></code></a></p></td>
<td><p>Turns non-negative integers (indexes/tokens) into dense vectors of fixed size.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.SyncBatchNorm" title="mxnet.gluon.contrib.nn.SyncBatchNorm"><code class="xref py py-obj docutils literal notranslate"><span class="pre">SyncBatchNorm</span></code></a></p></td>
<td><p>Cross-GPU Synchronized Batch normalization (SyncBN)</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.PixelShuffle1D" title="mxnet.gluon.contrib.nn.PixelShuffle1D"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PixelShuffle1D</span></code></a></p></td>
<td><p>Pixel-shuffle layer for upsampling in 1 dimension.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.PixelShuffle2D" title="mxnet.gluon.contrib.nn.PixelShuffle2D"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PixelShuffle2D</span></code></a></p></td>
<td><p>Pixel-shuffle layer for upsampling in 2 dimensions.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.nn.PixelShuffle3D" title="mxnet.gluon.contrib.nn.PixelShuffle3D"><code class="xref py py-obj docutils literal notranslate"><span class="pre">PixelShuffle3D</span></code></a></p></td>
<td><p>Pixel-shuffle layer for upsampling in 3 dimensions.</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="convolutional-neural-network">
<h2>Convolutional Neural Network<a class="headerlink" href="#convolutional-neural-network" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.cnn.DeformableConvolution" title="mxnet.gluon.contrib.cnn.DeformableConvolution"><code class="xref py py-obj docutils literal notranslate"><span class="pre">DeformableConvolution</span></code></a></p></td>
<td><p>2-D Deformable Convolution v_1 (Dai, 2017).</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="recurrent-neural-network">
<h2>Recurrent Neural Network<a class="headerlink" href="#recurrent-neural-network" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.rnn.VariationalDropoutCell" title="mxnet.gluon.contrib.rnn.VariationalDropoutCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">VariationalDropoutCell</span></code></a></p></td>
<td><p>Applies Variational Dropout on base cell.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv1DRNNCell" title="mxnet.gluon.contrib.rnn.Conv1DRNNCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv1DRNNCell</span></code></a></p></td>
<td><p>1D Convolutional RNN cell.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv2DRNNCell" title="mxnet.gluon.contrib.rnn.Conv2DRNNCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv2DRNNCell</span></code></a></p></td>
<td><p>2D Convolutional RNN cell.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv3DRNNCell" title="mxnet.gluon.contrib.rnn.Conv3DRNNCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv3DRNNCell</span></code></a></p></td>
<td><p>3D Convolutional RNN cells</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv1DLSTMCell" title="mxnet.gluon.contrib.rnn.Conv1DLSTMCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv1DLSTMCell</span></code></a></p></td>
<td><p>1D Convolutional LSTM network cell.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv2DLSTMCell" title="mxnet.gluon.contrib.rnn.Conv2DLSTMCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv2DLSTMCell</span></code></a></p></td>
<td><p>2D Convolutional LSTM network cell.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv3DLSTMCell" title="mxnet.gluon.contrib.rnn.Conv3DLSTMCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv3DLSTMCell</span></code></a></p></td>
<td><p>3D Convolutional LSTM network cell.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv1DGRUCell" title="mxnet.gluon.contrib.rnn.Conv1DGRUCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv1DGRUCell</span></code></a></p></td>
<td><p>1D Convolutional Gated Rectified Unit (GRU) network cell.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv2DGRUCell" title="mxnet.gluon.contrib.rnn.Conv2DGRUCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv2DGRUCell</span></code></a></p></td>
<td><p>2D Convolutional Gated Rectified Unit (GRU) network cell.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.Conv3DGRUCell" title="mxnet.gluon.contrib.rnn.Conv3DGRUCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Conv3DGRUCell</span></code></a></p></td>
<td><p>3D Convolutional Gated Rectified Unit (GRU) network cell.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.rnn.LSTMPCell" title="mxnet.gluon.contrib.rnn.LSTMPCell"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LSTMPCell</span></code></a></p></td>
<td><p>Long-Short Term Memory Projected (LSTMP) network cell.</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="data">
<h2>Data<a class="headerlink" href="#data" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.data.sampler.IntervalSampler" title="mxnet.gluon.contrib.data.sampler.IntervalSampler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">IntervalSampler</span></code></a></p></td>
<td><p>Samples elements from [0, length) at fixed intervals.</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="text-dataset">
<h2>Text Dataset<a class="headerlink" href="#text-dataset" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.data.text.WikiText2" title="mxnet.gluon.contrib.data.text.WikiText2"><code class="xref py py-obj docutils literal notranslate"><span class="pre">WikiText2</span></code></a></p></td>
<td><p>WikiText-2 word-level dataset for language modeling, from Salesforce research.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.data.text.WikiText103" title="mxnet.gluon.contrib.data.text.WikiText103"><code class="xref py py-obj docutils literal notranslate"><span class="pre">WikiText103</span></code></a></p></td>
<td><p>WikiText-103 word-level dataset for language modeling, from Salesforce research.</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="estimator">
<h2>Estimator<a class="headerlink" href="#estimator" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.estimator.Estimator" title="mxnet.gluon.contrib.estimator.Estimator"><code class="xref py py-obj docutils literal notranslate"><span class="pre">Estimator</span></code></a></p></td>
<td><p>Estimator Class for easy model training</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="event-handler">
<h2>Event Handler<a class="headerlink" href="#event-handler" title="Permalink to this headline"></a></h2>
<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.gluon.contrib.estimator.StoppingHandler" title="mxnet.gluon.contrib.estimator.StoppingHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">StoppingHandler</span></code></a></p></td>
<td><p>Stop conditions to stop training Stop training if maximum number of batches or epochs reached.</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.MetricHandler" title="mxnet.gluon.contrib.estimator.MetricHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">MetricHandler</span></code></a></p></td>
<td><p>Metric Handler that update metric values at batch end</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="mxnet.gluon.contrib.estimator.ValidationHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">ValidationHandler</span></code></a></p></td>
<td><p>Validation Handler that evaluate model on validation dataset</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.LoggingHandler" title="mxnet.gluon.contrib.estimator.LoggingHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">LoggingHandler</span></code></a></p></td>
<td><p>Basic Logging Handler that applies to every Gluon estimator by default.</p></td>
</tr>
<tr class="row-odd"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="mxnet.gluon.contrib.estimator.CheckpointHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">CheckpointHandler</span></code></a></p></td>
<td><p>Save the model after user define period</p></td>
</tr>
<tr class="row-even"><td><p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.EarlyStoppingHandler" title="mxnet.gluon.contrib.estimator.EarlyStoppingHandler"><code class="xref py py-obj docutils literal notranslate"><span class="pre">EarlyStoppingHandler</span></code></a></p></td>
<td><p>Early stop training if monitored value is not improving</p></td>
</tr>
</tbody>
</table>
</div>
<div class="section" id="module-mxnet.gluon.contrib">
<span id="api-reference"></span><h2>API Reference<a class="headerlink" href="#module-mxnet.gluon.contrib" title="Permalink to this headline"></a></h2>
<p>Contrib neural network module.</p>
<span class="target" id="module-mxnet.gluon.contrib.nn"></span><p>Contributed neural network modules.</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.Concurrent">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">Concurrent</code><span class="sig-paren">(</span><em class="sig-param">axis=-1</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#Concurrent"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.Concurrent" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.nn.basic_layers.Sequential</span></code></p>
<p>Lays <cite>Block</cite> s concurrently.</p>
<p>This block feeds its input to all children blocks, and
produce the output by concatenating all the children blocks’ outputs
on the specified axis.</p>
<p>Example:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">net</span> <span class="o">=</span> <span class="n">Concurrent</span><span class="p">()</span>
<span class="c1"># use net&#39;s name_scope to give children blocks appropriate names.</span>
<span class="k">with</span> <span class="n">net</span><span class="o">.</span><span class="n">name_scope</span><span class="p">():</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">activation</span><span class="o">=</span><span class="s1">&#39;relu&#39;</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">Identity</span><span class="p">())</span>
</pre></div>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>axis</strong> (<em>int</em><em>, </em><em>default -1</em>) – The axis on which to concatenate the outputs.</p>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.Concurrent.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#Concurrent.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.Concurrent.forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to implement forward computation using <code class="xref py py-class docutils literal notranslate"><span class="pre">NDArray</span></code>. Only
accepts positional arguments.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>*args</strong> (<em>list of NDArray</em>) – Input tensors.</p>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.HybridConcurrent">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">HybridConcurrent</code><span class="sig-paren">(</span><em class="sig-param">axis=-1</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#HybridConcurrent"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.HybridConcurrent" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.nn.basic_layers.HybridSequential</span></code></p>
<p>Lays <cite>HybridBlock</cite> s concurrently.</p>
<p>This block feeds its input to all children blocks, and
produce the output by concatenating all the children blocks’ outputs
on the specified axis.</p>
<p>Example:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">net</span> <span class="o">=</span> <span class="n">HybridConcurrent</span><span class="p">()</span>
<span class="c1"># use net&#39;s name_scope to give children blocks appropriate names.</span>
<span class="k">with</span> <span class="n">net</span><span class="o">.</span><span class="n">name_scope</span><span class="p">():</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">activation</span><span class="o">=</span><span class="s1">&#39;relu&#39;</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">Identity</span><span class="p">())</span>
</pre></div>
</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>axis</strong> (<em>int</em><em>, </em><em>default -1</em>) – The axis on which to concatenate the outputs.</p>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.HybridConcurrent.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#HybridConcurrent.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.HybridConcurrent.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.Identity">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">Identity</code><span class="sig-paren">(</span><em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#Identity"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.Identity" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>Block that passes through the input directly.</p>
<p>This block can be used in conjunction with HybridConcurrent
block for residual connection.</p>
<p>Example:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">net</span> <span class="o">=</span> <span class="n">HybridConcurrent</span><span class="p">()</span>
<span class="c1"># use net&#39;s name_scope to give child Blocks appropriate names.</span>
<span class="k">with</span> <span class="n">net</span><span class="o">.</span><span class="n">name_scope</span><span class="p">():</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">10</span><span class="p">,</span> <span class="n">activation</span><span class="o">=</span><span class="s1">&#39;relu&#39;</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">nn</span><span class="o">.</span><span class="n">Dense</span><span class="p">(</span><span class="mi">20</span><span class="p">))</span>
<span class="n">net</span><span class="o">.</span><span class="n">add</span><span class="p">(</span><span class="n">Identity</span><span class="p">())</span>
</pre></div>
</div>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.Identity.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#Identity.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.Identity.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.SparseEmbedding">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">SparseEmbedding</code><span class="sig-paren">(</span><em class="sig-param">input_dim</em>, <em class="sig-param">output_dim</em>, <em class="sig-param">dtype='float32'</em>, <em class="sig-param">weight_initializer=None</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#SparseEmbedding"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.SparseEmbedding" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.Block</span></code></p>
<p>Turns non-negative integers (indexes/tokens) into dense vectors
of fixed size. eg. [4, 20] -&gt; [[0.25, 0.1], [0.6, -0.2]]</p>
<p>This SparseBlock is designed for distributed training with extremely large
input dimension. Both weight and gradient w.r.t. weight are <cite>RowSparseNDArray</cite>.</p>
<p>Note: if <cite>sparse_grad</cite> is set to True, the gradient w.r.t weight will be
sparse. Only a subset of optimizers support sparse gradients, including SGD, AdaGrad
and Adam. By default lazy updates is turned on, which may perform differently
from standard updates. For more details, please check the Optimization API at:
<a class="reference external" href="https://mxnet.incubator.apache.org/api/python/optimization/optimization.html">https://mxnet.incubator.apache.org/api/python/optimization/optimization.html</a></p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_dim</strong> (<em>int</em>) – Size of the vocabulary, i.e. maximum integer index + 1.</p></li>
<li><p><strong>output_dim</strong> (<em>int</em>) – Dimension of the dense embedding.</p></li>
<li><p><strong>dtype</strong> (<em>str</em><em> or </em><em>np.dtype</em><em>, </em><em>default 'float32'</em>) – Data type of output embeddings.</p></li>
<li><p><strong>weight_initializer</strong> (<a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the <cite>embeddings</cite> matrix.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: (N-1)-D tensor with shape: <cite>(x1, x2, …, xN-1)</cite>.</p></li>
</ul>
</p></li>
<li><p><strong>Output</strong> – <ul>
<li><p><strong>out</strong>: N-D tensor with shape: <cite>(x1, x2, …, xN-1, output_dim)</cite>.</p></li>
</ul>
</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.SparseEmbedding.forward">
<code class="sig-name descname">forward</code><span class="sig-paren">(</span><em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#SparseEmbedding.forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.SparseEmbedding.forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to implement forward computation using <code class="xref py py-class docutils literal notranslate"><span class="pre">NDArray</span></code>. Only
accepts positional arguments.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><p><strong>*args</strong> (<em>list of NDArray</em>) – Input tensors.</p>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.SyncBatchNorm">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">SyncBatchNorm</code><span class="sig-paren">(</span><em class="sig-param">in_channels=0</em>, <em class="sig-param">num_devices=None</em>, <em class="sig-param">momentum=0.9</em>, <em class="sig-param">epsilon=1e-05</em>, <em class="sig-param">center=True</em>, <em class="sig-param">scale=True</em>, <em class="sig-param">use_global_stats=False</em>, <em class="sig-param">beta_initializer='zeros'</em>, <em class="sig-param">gamma_initializer='ones'</em>, <em class="sig-param">running_mean_initializer='zeros'</em>, <em class="sig-param">running_variance_initializer='ones'</em>, <em class="sig-param">**kwargs</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#SyncBatchNorm"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.SyncBatchNorm" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.nn.basic_layers.BatchNorm</span></code></p>
<p>Cross-GPU Synchronized Batch normalization (SyncBN)</p>
<p>Standard BN <a class="footnote-reference brackets" href="#id3" id="id1">1</a> implementation only normalize the data within each device.
SyncBN normalizes the input within the whole mini-batch.
We follow the implementation described in the paper <a class="footnote-reference brackets" href="#id4" id="id2">2</a>.</p>
<p>Note: Current implementation of SyncBN does not support FP16 training.
For FP16 inference, use standard nn.BatchNorm instead of SyncBN.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>in_channels</strong> (<em>int</em><em>, </em><em>default 0</em>) – Number of channels (feature maps) in input data. If not specified,
initialization will be deferred to the first time <cite>forward</cite> is called
and <cite>in_channels</cite> will be inferred from the shape of input data.</p></li>
<li><p><strong>num_devices</strong> (<em>int</em><em>, </em><em>default number of visible GPUs</em>) – </p></li>
<li><p><strong>momentum</strong> (<em>float</em><em>, </em><em>default 0.9</em>) – Momentum for the moving average.</p></li>
<li><p><strong>epsilon</strong> (<em>float</em><em>, </em><em>default 1e-5</em>) – Small float added to variance to avoid dividing by zero.</p></li>
<li><p><strong>center</strong> (<em>bool</em><em>, </em><em>default True</em>) – If True, add offset of <cite>beta</cite> to normalized tensor.
If False, <cite>beta</cite> is ignored.</p></li>
<li><p><strong>scale</strong> (<em>bool</em><em>, </em><em>default True</em>) – If True, multiply by <cite>gamma</cite>. If False, <cite>gamma</cite> is not used.
When the next layer is linear (also e.g. <cite>nn.relu</cite>),
this can be disabled since the scaling
will be done by the next layer.</p></li>
<li><p><strong>use_global_stats</strong> (<em>bool</em><em>, </em><em>default False</em>) – If True, use global moving statistics instead of local batch-norm. This will force
change batch-norm into a scale shift operator.
If False, use local batch-norm.</p></li>
<li><p><strong>beta_initializer</strong> (str or <cite>Initializer</cite>, default ‘zeros’) – Initializer for the beta weight.</p></li>
<li><p><strong>gamma_initializer</strong> (str or <cite>Initializer</cite>, default ‘ones’) – Initializer for the gamma weight.</p></li>
<li><p><strong>running_mean_initializer</strong> (str or <cite>Initializer</cite>, default ‘zeros’) – Initializer for the running mean.</p></li>
<li><p><strong>running_variance_initializer</strong> (str or <cite>Initializer</cite>, default ‘ones’) – Initializer for the running variance.</p></li>
</ul>
</dd>
</dl>
<dl>
<dt>Inputs:</dt><dd><ul class="simple">
<li><p><strong>data</strong>: input tensor with arbitrary shape.</p></li>
</ul>
</dd>
<dt>Outputs:</dt><dd><ul class="simple">
<li><p><strong>out</strong>: output tensor with the same shape as <cite>data</cite>.</p></li>
</ul>
</dd>
<dt>Reference:</dt><dd><dl class="footnote brackets">
<dt class="label" id="id3"><span class="brackets"><a class="fn-backref" href="#id1">1</a></span></dt>
<dd><p>Ioffe, Sergey, and Christian Szegedy. “Batch normalization: Accelerating deep network training by reducing internal covariate shift.” <em>ICML 2015</em></p>
</dd>
<dt class="label" id="id4"><span class="brackets"><a class="fn-backref" href="#id2">2</a></span></dt>
<dd><p>Hang Zhang, Kristin Dana, Jianping Shi, Zhongyue Zhang, Xiaogang Wang, Ambrish Tyagi, and Amit Agrawal. “Context Encoding for Semantic Segmentation.” <em>CVPR 2018</em></p>
</dd>
</dl>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.SyncBatchNorm.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em>, <em class="sig-param">gamma</em>, <em class="sig-param">beta</em>, <em class="sig-param">running_mean</em>, <em class="sig-param">running_var</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#SyncBatchNorm.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.SyncBatchNorm.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle1D">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">PixelShuffle1D</code><span class="sig-paren">(</span><em class="sig-param">factor</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle1D"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle1D" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>Pixel-shuffle layer for upsampling in 1 dimension.</p>
<p>Pixel-shuffling is the operation of taking groups of values along
the <em>channel</em> dimension and regrouping them into blocks of pixels
along the <code class="docutils literal notranslate"><span class="pre">W</span></code> dimension, thereby effectively multiplying that dimension
by a constant factor in size.</p>
<p>For example, a feature map of shape <span class="math notranslate nohighlight">\((fC, W)\)</span> is reshaped
into <span class="math notranslate nohighlight">\((C, fW)\)</span> by forming little value groups of size <span class="math notranslate nohighlight">\(f\)</span>
and arranging them in a grid of size <span class="math notranslate nohighlight">\(W\)</span>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>factor</strong> (<em>int</em><em> or </em><em>1-tuple of int</em>) – Upsampling factor, applied to the <code class="docutils literal notranslate"><span class="pre">W</span></code> dimension.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">f*C,</span> <span class="pre">W)</span></code>.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">C,</span> <span class="pre">W*f)</span></code>.</p></li>
</ul>
</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="n">pxshuf</span> <span class="o">=</span> <span class="n">PixelShuffle1D</span><span class="p">(</span><span class="mi">2</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="mi">1</span><span class="p">,</span> <span class="mi">8</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">pxshuf</span><span class="p">(</span><span class="n">x</span><span class="p">)</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1, 4, 6)</span>
</pre></div>
</div>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle1D.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle1D.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle1D.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Perform pixel-shuffling on the input.</p>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle2D">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">PixelShuffle2D</code><span class="sig-paren">(</span><em class="sig-param">factor</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle2D"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle2D" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>Pixel-shuffle layer for upsampling in 2 dimensions.</p>
<p>Pixel-shuffling is the operation of taking groups of values along
the <em>channel</em> dimension and regrouping them into blocks of pixels
along the <code class="docutils literal notranslate"><span class="pre">H</span></code> and <code class="docutils literal notranslate"><span class="pre">W</span></code> dimensions, thereby effectively multiplying
those dimensions by a constant factor in size.</p>
<p>For example, a feature map of shape <span class="math notranslate nohighlight">\((f^2 C, H, W)\)</span> is reshaped
into <span class="math notranslate nohighlight">\((C, fH, fW)\)</span> by forming little <span class="math notranslate nohighlight">\(f \times f\)</span> blocks
of pixels and arranging them in an <span class="math notranslate nohighlight">\(H \times W\)</span> grid.</p>
<p>Pixel-shuffling together with regular convolution is an alternative,
learnable way of upsampling an image by arbitrary factors. It is reported
to help overcome checkerboard artifacts that are common in upsampling with
transposed convolutions (also called deconvolutions). See the paper
<a class="reference external" href="https://arxiv.org/abs/1609.05158">Real-Time Single Image and Video Super-Resolution Using an Efficient
Sub-Pixel Convolutional Neural Network</a>
for further details.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>factor</strong> (<em>int</em><em> or </em><em>2-tuple of int</em>) – Upsampling factors, applied to the <code class="docutils literal notranslate"><span class="pre">H</span></code> and <code class="docutils literal notranslate"><span class="pre">W</span></code> dimensions,
in that order.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">f1*f2*C,</span> <span class="pre">H,</span> <span class="pre">W)</span></code>.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">C,</span> <span class="pre">H*f1,</span> <span class="pre">W*f2)</span></code>.</p></li>
</ul>
</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="n">pxshuf</span> <span class="o">=</span> <span class="n">PixelShuffle2D</span><span class="p">((</span><span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="mi">1</span><span class="p">,</span> <span class="mi">12</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">pxshuf</span><span class="p">(</span><span class="n">x</span><span class="p">)</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1, 2, 6, 15)</span>
</pre></div>
</div>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle2D.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle2D.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle2D.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Perform pixel-shuffling on the input.</p>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle3D">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.nn.</code><code class="sig-name descname">PixelShuffle3D</code><span class="sig-paren">(</span><em class="sig-param">factor</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle3D"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle3D" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>Pixel-shuffle layer for upsampling in 3 dimensions.</p>
<p>Pixel-shuffling (or voxel-shuffling in 3D) is the operation of taking
groups of values along the <em>channel</em> dimension and regrouping them into
blocks of voxels along the <code class="docutils literal notranslate"><span class="pre">D</span></code>, <code class="docutils literal notranslate"><span class="pre">H</span></code> and <code class="docutils literal notranslate"><span class="pre">W</span></code> dimensions, thereby
effectively multiplying those dimensions by a constant factor in size.</p>
<p>For example, a feature map of shape <span class="math notranslate nohighlight">\((f^3 C, D, H, W)\)</span> is reshaped
into <span class="math notranslate nohighlight">\((C, fD, fH, fW)\)</span> by forming little <span class="math notranslate nohighlight">\(f \times f \times f\)</span>
blocks of voxels and arranging them in a <span class="math notranslate nohighlight">\(D \times H \times W\)</span> grid.</p>
<p>Pixel-shuffling together with regular convolution is an alternative,
learnable way of upsampling an image by arbitrary factors. It is reported
to help overcome checkerboard artifacts that are common in upsampling with
transposed convolutions (also called deconvolutions). See the paper
<a class="reference external" href="https://arxiv.org/abs/1609.05158">Real-Time Single Image and Video Super-Resolution Using an Efficient
Sub-Pixel Convolutional Neural Network</a>
for further details.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>factor</strong> (<em>int</em><em> or </em><em>3-tuple of int</em>) – Upsampling factors, applied to the <code class="docutils literal notranslate"><span class="pre">D</span></code>, <code class="docutils literal notranslate"><span class="pre">H</span></code> and <code class="docutils literal notranslate"><span class="pre">W</span></code>
dimensions, in that order.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">f1*f2*f3*C,</span> <span class="pre">D,</span> <span class="pre">H,</span> <span class="pre">W)</span></code>.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: Tensor of shape <code class="docutils literal notranslate"><span class="pre">(N,</span> <span class="pre">C,</span> <span class="pre">D*f1,</span> <span class="pre">H*f2,</span> <span class="pre">W*f3)</span></code>.</p></li>
</ul>
</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="n">pxshuf</span> <span class="o">=</span> <span class="n">PixelShuffle3D</span><span class="p">((</span><span class="mi">2</span><span class="p">,</span> <span class="mi">3</span><span class="p">,</span> <span class="mi">4</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">x</span> <span class="o">=</span> <span class="n">mx</span><span class="o">.</span><span class="n">nd</span><span class="o">.</span><span class="n">zeros</span><span class="p">((</span><span class="mi">1</span><span class="p">,</span> <span class="mi">48</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="mi">7</span><span class="p">))</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">pxshuf</span><span class="p">(</span><span class="n">x</span><span class="p">)</span><span class="o">.</span><span class="n">shape</span>
<span class="go">(1, 2, 6, 15, 28)</span>
</pre></div>
</div>
<dl class="method">
<dt id="mxnet.gluon.contrib.nn.PixelShuffle3D.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/nn/basic_layers.html#PixelShuffle3D.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.nn.PixelShuffle3D.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Perform pixel-shuffling on the input.</p>
</dd></dl>
</dd></dl>
<span class="target" id="module-mxnet.gluon.contrib.cnn"></span><p>Contrib convolutional neural network module.</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.cnn.DeformableConvolution">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.cnn.</code><code class="sig-name descname">DeformableConvolution</code><span class="sig-paren">(</span><em class="sig-param">channels</em>, <em class="sig-param">kernel_size=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">strides=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">padding=(0</em>, <em class="sig-param">0)</em>, <em class="sig-param">dilation=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">groups=1</em>, <em class="sig-param">num_deformable_group=1</em>, <em class="sig-param">layout='NCHW'</em>, <em class="sig-param">use_bias=True</em>, <em class="sig-param">in_channels=0</em>, <em class="sig-param">activation=None</em>, <em class="sig-param">weight_initializer=None</em>, <em class="sig-param">bias_initializer='zeros'</em>, <em class="sig-param">offset_weight_initializer='zeros'</em>, <em class="sig-param">offset_bias_initializer='zeros'</em>, <em class="sig-param">offset_use_bias=True</em>, <em class="sig-param">op_name='DeformableConvolution'</em>, <em class="sig-param">adj=None</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/cnn/conv_layers.html#DeformableConvolution"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.cnn.DeformableConvolution" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>2-D Deformable Convolution v_1 (Dai, 2017).
Normal Convolution uses sampling points in a regular grid, while the sampling
points of Deformablem Convolution can be offset. The offset is learned with a
separate convolution layer during the training. Both the convolution layer for
generating the output features and the offsets are included in this gluon layer.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>channels</strong> (<em>int</em><em>,</em>) – The dimensionality of the output space
i.e. the number of output channels in the convolution.</p></li>
<li><p><strong>kernel_size</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the dimensions of the convolution window.</p></li>
<li><p><strong>strides</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the strides of the convolution.</p></li>
<li><p><strong>padding</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>0</em><em>,</em><em>0</em><em>)</em><em>)</em>) – If padding is non-zero, then the input is implicitly zero-padded
on both sides for padding number of points.</p></li>
<li><p><strong>dilation</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the dilation rate to use for dilated convolution.</p></li>
<li><p><strong>groups</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 1</em><em>)</em>) – Controls the connections between inputs and outputs.
At groups=1, all inputs are convolved to all outputs.
At groups=2, the operation becomes equivalent to having two convolution
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.</p></li>
<li><p><strong>num_deformable_group</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 1</em><em>)</em>) – Number of deformable group partitions.</p></li>
<li><p><strong>layout</strong> (<em>str</em><em>, </em><em>(</em><em>Default value = NCHW</em><em>)</em>) – Dimension ordering of data and weight. Can be ‘NCW’, ‘NWC’, ‘NCHW’,
‘NHWC’, ‘NCDHW’, ‘NDHWC’, etc. ‘N’, ‘C’, ‘H’, ‘W’, ‘D’ stands for
batch, channel, height, width and depth dimensions respectively.
Convolution is performed over ‘D’, ‘H’, and ‘W’ dimensions.</p></li>
<li><p><strong>use_bias</strong> (<em>bool</em><em>, </em><em>(</em><em>Default value = True</em><em>)</em>) – Whether the layer for generating the output features uses a bias vector.</p></li>
<li><p><strong>in_channels</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 0</em><em>)</em>) – The number of input channels to this layer. If not specified,
initialization will be deferred to the first time <cite>forward</cite> is called
and input channels will be inferred from the shape of input data.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em>, </em><em>(</em><em>Default value = None</em><em>)</em>) – Activation function to use. See <a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a>.
If you don’t specify anything, no activation is applied
(ie. “linear” activation: <cite>a(x) = x</cite>).</p></li>
<li><p><strong>weight_initializer</strong> (str or <cite>Initializer</cite>, (Default value = None)) – Initializer for the <cite>weight</cite> weights matrix for the convolution layer
for generating the output features.</p></li>
<li><p><strong>bias_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros)) – Initializer for the bias vector for the convolution layer
for generating the output features.</p></li>
<li><p><strong>offset_weight_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros)) – Initializer for the <cite>weight</cite> weights matrix for the convolution layer
for generating the offset.</p></li>
<li><p><strong>offset_bias_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros),) – Initializer for the bias vector for the convolution layer
for generating the offset.</p></li>
<li><p><strong>offset_use_bias</strong> (<em>bool</em><em>, </em><em>(</em><em>Default value = True</em><em>)</em>) – Whether the layer for generating the offset uses a bias vector.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: 4D input tensor with shape
<cite>(batch_size, in_channels, height, width)</cite> when <cite>layout</cite> is <cite>NCHW</cite>.
For other layouts shape is permuted accordingly.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: 4D output tensor with shape
<cite>(batch_size, channels, out_height, out_width)</cite> when <cite>layout</cite> is <cite>NCHW</cite>.
out_height and out_width are calculated as:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">out_height</span> <span class="o">=</span> <span class="n">floor</span><span class="p">((</span><span class="n">height</span><span class="o">+</span><span class="mi">2</span><span class="o">*</span><span class="n">padding</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="n">dilation</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">*</span><span class="p">(</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="n">stride</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span><span class="o">+</span><span class="mi">1</span>
<span class="n">out_width</span> <span class="o">=</span> <span class="n">floor</span><span class="p">((</span><span class="n">width</span><span class="o">+</span><span class="mi">2</span><span class="o">*</span><span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">-</span><span class="n">dilation</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">*</span><span class="p">(</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="n">stride</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span><span class="o">+</span><span class="mi">1</span>
</pre></div>
</div>
</li>
</ul>
</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.cnn.DeformableConvolution.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em>, <em class="sig-param">offset_weight</em>, <em class="sig-param">deformable_conv_weight</em>, <em class="sig-param">offset_bias=None</em>, <em class="sig-param">deformable_conv_bias=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/cnn/conv_layers.html#DeformableConvolution.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.cnn.DeformableConvolution.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.cnn.ModulatedDeformableConvolution">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.cnn.</code><code class="sig-name descname">ModulatedDeformableConvolution</code><span class="sig-paren">(</span><em class="sig-param">channels</em>, <em class="sig-param">kernel_size=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">strides=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">padding=(0</em>, <em class="sig-param">0)</em>, <em class="sig-param">dilation=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">groups=1</em>, <em class="sig-param">num_deformable_group=1</em>, <em class="sig-param">layout='NCHW'</em>, <em class="sig-param">use_bias=True</em>, <em class="sig-param">in_channels=0</em>, <em class="sig-param">activation=None</em>, <em class="sig-param">weight_initializer=None</em>, <em class="sig-param">bias_initializer='zeros'</em>, <em class="sig-param">offset_weight_initializer='zeros'</em>, <em class="sig-param">offset_bias_initializer='zeros'</em>, <em class="sig-param">offset_use_bias=True</em>, <em class="sig-param">op_name='ModulatedDeformableConvolution'</em>, <em class="sig-param">adj=None</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/cnn/conv_layers.html#ModulatedDeformableConvolution"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.cnn.ModulatedDeformableConvolution" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.block.HybridBlock</span></code></p>
<p>2-D Deformable Convolution v2 (Dai, 2018).</p>
<p>The modulated deformable convolution operation is described in <a class="reference external" href="https://arxiv.org/abs/1811.11168">https://arxiv.org/abs/1811.11168</a></p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>channels</strong> (<em>int</em><em>,</em>) – The dimensionality of the output space
i.e. the number of output channels in the convolution.</p></li>
<li><p><strong>kernel_size</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the dimensions of the convolution window.</p></li>
<li><p><strong>strides</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the strides of the convolution.</p></li>
<li><p><strong>padding</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>0</em><em>,</em><em>0</em><em>)</em><em>)</em>) – If padding is non-zero, then the input is implicitly zero-padded
on both sides for padding number of points.</p></li>
<li><p><strong>dilation</strong> (<em>int</em><em> or </em><em>tuple/list of 2 ints</em><em>, </em><em>(</em><em>Default value =</em><em> (</em><em>1</em><em>,</em><em>1</em><em>)</em><em>)</em>) – Specifies the dilation rate to use for dilated convolution.</p></li>
<li><p><strong>groups</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 1</em><em>)</em>) – Controls the connections between inputs and outputs.
At groups=1, all inputs are convolved to all outputs.
At groups=2, the operation becomes equivalent to having two convolution
layers side by side, each seeing half the input channels, and producing
half the output channels, and both subsequently concatenated.</p></li>
<li><p><strong>num_deformable_group</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 1</em><em>)</em>) – Number of deformable group partitions.</p></li>
<li><p><strong>layout</strong> (<em>str</em><em>, </em><em>(</em><em>Default value = NCHW</em><em>)</em>) – Dimension ordering of data and weight. Can be ‘NCW’, ‘NWC’, ‘NCHW’,
‘NHWC’, ‘NCDHW’, ‘NDHWC’, etc. ‘N’, ‘C’, ‘H’, ‘W’, ‘D’ stands for
batch, channel, height, width and depth dimensions respectively.
Convolution is performed over ‘D’, ‘H’, and ‘W’ dimensions.</p></li>
<li><p><strong>use_bias</strong> (<em>bool</em><em>, </em><em>(</em><em>Default value = True</em><em>)</em>) – Whether the layer for generating the output features uses a bias vector.</p></li>
<li><p><strong>in_channels</strong> (<em>int</em><em>, </em><em>(</em><em>Default value = 0</em><em>)</em>) – The number of input channels to this layer. If not specified,
initialization will be deferred to the first time <cite>forward</cite> is called
and input channels will be inferred from the shape of input data.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em>, </em><em>(</em><em>Default value = None</em><em>)</em>) – Activation function to use. See <a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a>.
If you don’t specify anything, no activation is applied
(ie. “linear” activation: <cite>a(x) = x</cite>).</p></li>
<li><p><strong>weight_initializer</strong> (str or <cite>Initializer</cite>, (Default value = None)) – Initializer for the <cite>weight</cite> weights matrix for the convolution layer
for generating the output features.</p></li>
<li><p><strong>bias_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros)) – Initializer for the bias vector for the convolution layer
for generating the output features.</p></li>
<li><p><strong>offset_weight_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros)) – Initializer for the <cite>weight</cite> weights matrix for the convolution layer
for generating the offset.</p></li>
<li><p><strong>offset_bias_initializer</strong> (str or <cite>Initializer</cite>, (Default value = zeros),) – Initializer for the bias vector for the convolution layer
for generating the offset.</p></li>
<li><p><strong>offset_use_bias</strong> (<em>bool</em><em>, </em><em>(</em><em>Default value = True</em><em>)</em>) – Whether the layer for generating the offset uses a bias vector.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: 4D input tensor with shape
<cite>(batch_size, in_channels, height, width)</cite> when <cite>layout</cite> is <cite>NCHW</cite>.
For other layouts shape is permuted accordingly.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: 4D output tensor with shape
<cite>(batch_size, channels, out_height, out_width)</cite> when <cite>layout</cite> is <cite>NCHW</cite>.
out_height and out_width are calculated as:</p>
<div class="highlight-default notranslate"><div class="highlight"><pre><span></span><span class="n">out_height</span> <span class="o">=</span> <span class="n">floor</span><span class="p">((</span><span class="n">height</span><span class="o">+</span><span class="mi">2</span><span class="o">*</span><span class="n">padding</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="n">dilation</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">*</span><span class="p">(</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">0</span><span class="p">]</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="n">stride</span><span class="p">[</span><span class="mi">0</span><span class="p">])</span><span class="o">+</span><span class="mi">1</span>
<span class="n">out_width</span> <span class="o">=</span> <span class="n">floor</span><span class="p">((</span><span class="n">width</span><span class="o">+</span><span class="mi">2</span><span class="o">*</span><span class="n">padding</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">-</span><span class="n">dilation</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">*</span><span class="p">(</span><span class="n">kernel_size</span><span class="p">[</span><span class="mi">1</span><span class="p">]</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">-</span><span class="mi">1</span><span class="p">)</span><span class="o">/</span><span class="n">stride</span><span class="p">[</span><span class="mi">1</span><span class="p">])</span><span class="o">+</span><span class="mi">1</span>
</pre></div>
</div>
</li>
</ul>
</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.cnn.ModulatedDeformableConvolution.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">x</em>, <em class="sig-param">offset_weight</em>, <em class="sig-param">deformable_conv_weight</em>, <em class="sig-param">offset_bias=None</em>, <em class="sig-param">deformable_conv_bias=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/cnn/conv_layers.html#ModulatedDeformableConvolution.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.cnn.ModulatedDeformableConvolution.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</dd></dl>
<span class="target" id="module-mxnet.gluon.contrib.rnn"></span><p>Contrib recurrent neural network module.</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv1DRNNCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv1DRNNCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv1DRNNCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv1DRNNCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvRNNCell</span></code></p>
<p>1D Convolutional RNN cell.</p>
<div class="math notranslate nohighlight">
\[h_t = tanh(W_i \ast x_t + R_i \ast h_{t-1} + b_i)\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCW’ the shape should be (C, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>,</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCW’ and ‘NWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_rnn_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv2DRNNCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv2DRNNCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv2DRNNCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv2DRNNCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvRNNCell</span></code></p>
<p>2D Convolutional RNN cell.</p>
<div class="math notranslate nohighlight">
\[h_t = tanh(W_i \ast x_t + R_i \ast h_{t-1} + b_i)\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCHW’ the shape should be (C, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCHW’ and ‘NHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_rnn_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv3DRNNCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv3DRNNCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCDHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv3DRNNCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv3DRNNCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvRNNCell</span></code></p>
<p>3D Convolutional RNN cells</p>
<div class="math notranslate nohighlight">
\[h_t = tanh(W_i \ast x_t + R_i \ast h_{t-1} + b_i)\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCDHW’ the shape should be (C, D, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCDHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCDHW’ and ‘NDHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_rnn_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv1DLSTMCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv1DLSTMCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv1DLSTMCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv1DLSTMCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvLSTMCell</span></code></p>
<p>1D Convolutional LSTM network cell.</p>
<p><a class="reference external" href="https://arxiv.org/abs/1506.04214">“Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting”</a> paper. Xingjian et al. NIPS2015</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
i_t = \sigma(W_i \ast x_t + R_i \ast h_{t-1} + b_i) \\
f_t = \sigma(W_f \ast x_t + R_f \ast h_{t-1} + b_f) \\
o_t = \sigma(W_o \ast x_t + R_o \ast h_{t-1} + b_o) \\
c^\prime_t = tanh(W_c \ast x_t + R_c \ast h_{t-1} + b_c) \\
c_t = f_t \circ c_{t-1} + i_t \circ c^\prime_t \\
h_t = o_t \circ tanh(c_t) \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCW’ the shape should be (C, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>,</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCW’ and ‘NWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in c^prime_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_lstm_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv2DLSTMCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv2DLSTMCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv2DLSTMCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv2DLSTMCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvLSTMCell</span></code></p>
<p>2D Convolutional LSTM network cell.</p>
<p><a class="reference external" href="https://arxiv.org/abs/1506.04214">“Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting”</a> paper. Xingjian et al. NIPS2015</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
i_t = \sigma(W_i \ast x_t + R_i \ast h_{t-1} + b_i) \\
f_t = \sigma(W_f \ast x_t + R_f \ast h_{t-1} + b_f) \\
o_t = \sigma(W_o \ast x_t + R_o \ast h_{t-1} + b_o) \\
c^\prime_t = tanh(W_c \ast x_t + R_c \ast h_{t-1} + b_c) \\
c_t = f_t \circ c_{t-1} + i_t \circ c^\prime_t \\
h_t = o_t \circ tanh(c_t) \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCHW’ the shape should be (C, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCHW’ and ‘NHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in c^prime_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_lstm_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv3DLSTMCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv3DLSTMCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCDHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv3DLSTMCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv3DLSTMCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvLSTMCell</span></code></p>
<p>3D Convolutional LSTM network cell.</p>
<p><a class="reference external" href="https://arxiv.org/abs/1506.04214">“Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting”</a> paper. Xingjian et al. NIPS2015</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
i_t = \sigma(W_i \ast x_t + R_i \ast h_{t-1} + b_i) \\
f_t = \sigma(W_f \ast x_t + R_f \ast h_{t-1} + b_f) \\
o_t = \sigma(W_o \ast x_t + R_o \ast h_{t-1} + b_o) \\
c^\prime_t = tanh(W_c \ast x_t + R_c \ast h_{t-1} + b_c) \\
c_t = f_t \circ c_{t-1} + i_t \circ c^\prime_t \\
h_t = o_t \circ tanh(c_t) \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCDHW’ the shape should be (C, D, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCDHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCDHW’ and ‘NDHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in c^prime_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_lstm_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv1DGRUCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv1DGRUCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv1DGRUCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv1DGRUCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvGRUCell</span></code></p>
<p>1D Convolutional Gated Rectified Unit (GRU) network cell.</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
r_t = \sigma(W_r \ast x_t + R_r \ast h_{t-1} + b_r) \\
z_t = \sigma(W_z \ast x_t + R_z \ast h_{t-1} + b_z) \\
n_t = tanh(W_i \ast x_t + b_i + r_t \circ (R_n \ast h_{t-1} + b_n)) \\
h^\prime_t = (1 - z_t) \circ n_t + z_t \circ h \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCW’ the shape should be (C, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>,</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>,</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCW’ and ‘NWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in n_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_gru_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv2DGRUCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv2DGRUCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv2DGRUCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv2DGRUCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvGRUCell</span></code></p>
<p>2D Convolutional Gated Rectified Unit (GRU) network cell.</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
r_t = \sigma(W_r \ast x_t + R_r \ast h_{t-1} + b_r) \\
z_t = \sigma(W_z \ast x_t + R_z \ast h_{t-1} + b_z) \\
n_t = tanh(W_i \ast x_t + b_i + r_t \circ (R_n \ast h_{t-1} + b_n)) \\
h^\prime_t = (1 - z_t) \circ n_t + z_t \circ h \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCHW’ the shape should be (C, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCHW’ and ‘NHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in n_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_gru_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.Conv3DGRUCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">Conv3DGRUCell</code><span class="sig-paren">(</span><em class="sig-param">input_shape</em>, <em class="sig-param">hidden_channels</em>, <em class="sig-param">i2h_kernel</em>, <em class="sig-param">h2h_kernel</em>, <em class="sig-param">i2h_pad=(0</em>, <em class="sig-param">0</em>, <em class="sig-param">0)</em>, <em class="sig-param">i2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">h2h_dilate=(1</em>, <em class="sig-param">1</em>, <em class="sig-param">1)</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">conv_layout='NCDHW'</em>, <em class="sig-param">activation='tanh'</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/conv_rnn_cell.html#Conv3DGRUCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.Conv3DGRUCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.rnn.conv_rnn_cell._ConvGRUCell</span></code></p>
<p>3D Convolutional Gated Rectified Unit (GRU) network cell.</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
r_t = \sigma(W_r \ast x_t + R_r \ast h_{t-1} + b_r) \\
z_t = \sigma(W_z \ast x_t + R_z \ast h_{t-1} + b_z) \\
n_t = tanh(W_i \ast x_t + b_i + r_t \circ (R_n \ast h_{t-1} + b_n)) \\
h^\prime_t = (1 - z_t) \circ n_t + z_t \circ h \\
\end{array}\end{split}\]</div>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>input_shape</strong> (<em>tuple of int</em>) – Input tensor shape at each time step for each sample, excluding dimension of the batch size
and sequence length. Must be consistent with <cite>conv_layout</cite>.
For example, for layout ‘NCDHW’ the shape should be (C, D, H, W).</p></li>
<li><p><strong>hidden_channels</strong> (<em>int</em>) – Number of output channels.</p></li>
<li><p><strong>i2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Input convolution kernel sizes.</p></li>
<li><p><strong>h2h_kernel</strong> (<em>int</em><em> or </em><em>tuple of int</em>) – Recurrent convolution kernel sizes. Only odd-numbered sizes are supported.</p></li>
<li><p><strong>i2h_pad</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>0</em><em>, </em><em>0</em><em>, </em><em>0</em><em>)</em>) – Pad for input convolution.</p></li>
<li><p><strong>i2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Input convolution dilate.</p></li>
<li><p><strong>h2h_dilate</strong> (<em>int</em><em> or </em><em>tuple of int</em><em>, </em><em>default</em><em> (</em><em>1</em><em>, </em><em>1</em><em>, </em><em>1</em><em>)</em>) – Recurrent convolution dilate.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the input convolutions.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the input convolutions.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the input convolution bias vectors.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default zeros</em>) – Initializer for the recurrent convolution bias vectors.</p></li>
<li><p><strong>conv_layout</strong> (<em>str</em><em>, </em><em>default 'NCDHW'</em>) – Layout for all convolution inputs, outputs and weights. Options are ‘NCDHW’ and ‘NDHWC’.</p></li>
<li><p><strong>activation</strong> (<em>str</em><em> or </em><a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a><em>, </em><em>default 'tanh'</em>) – Type of activation function used in n_t.
If argument type is string, it’s equivalent to nn.Activation(act_type=str). See
<a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.Activation" title="mxnet.ndarray.Activation"><code class="xref py py-func docutils literal notranslate"><span class="pre">Activation()</span></code></a> for available choices.
Alternatively, other activation blocks such as nn.LeakyReLU can be used.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'conv_gru_</span></code>’) – Prefix for name of layers (and name of weight if params is None).</p></li>
<li><p><strong>params</strong> (<em>RNNParams</em><em>, </em><em>default None</em>) – Container for weight sharing between cells. Created if None.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.VariationalDropoutCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">VariationalDropoutCell</code><span class="sig-paren">(</span><em class="sig-param">base_cell</em>, <em class="sig-param">drop_inputs=0.0</em>, <em class="sig-param">drop_states=0.0</em>, <em class="sig-param">drop_outputs=0.0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#VariationalDropoutCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.VariationalDropoutCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.rnn.rnn_cell.ModifierCell</span></code></p>
<p>Applies Variational Dropout on base cell.
<a class="reference external" href="https://arxiv.org/pdf/1512.05287.pdf">https://arxiv.org/pdf/1512.05287.pdf</a></p>
<p>Variational dropout uses the same dropout mask across time-steps. It can be applied to RNN
inputs, outputs, and states. The masks for them are not shared.</p>
<p>The dropout mask is initialized when stepping forward for the first time and will remain
the same until .reset() is called. Thus, if using the cell and stepping manually without calling
.unroll(), the .reset() should be called after each sequence.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>base_cell</strong> (<a class="reference internal" href="../rnn/index.html#mxnet.gluon.rnn.RecurrentCell" title="mxnet.gluon.rnn.RecurrentCell"><em>RecurrentCell</em></a>) – The cell on which to perform variational dropout.</p></li>
<li><p><strong>drop_inputs</strong> (<em>float</em><em>, </em><em>default 0.</em>) – The dropout rate for inputs. Won’t apply dropout if it equals 0.</p></li>
<li><p><strong>drop_states</strong> (<em>float</em><em>, </em><em>default 0.</em>) – The dropout rate for state inputs on the first state channel.
Won’t apply dropout if it equals 0.</p></li>
<li><p><strong>drop_outputs</strong> (<em>float</em><em>, </em><em>default 0.</em>) – The dropout rate for outputs. Won’t apply dropout if it equals 0.</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.rnn.VariationalDropoutCell.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">inputs</em>, <em class="sig-param">states</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#VariationalDropoutCell.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.VariationalDropoutCell.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.rnn.VariationalDropoutCell.reset">
<code class="sig-name descname">reset</code><span class="sig-paren">(</span><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#VariationalDropoutCell.reset"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.VariationalDropoutCell.reset" title="Permalink to this definition"></a></dt>
<dd><p>Reset before re-using the cell for another graph.</p>
</dd></dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.rnn.VariationalDropoutCell.unroll">
<code class="sig-name descname">unroll</code><span class="sig-paren">(</span><em class="sig-param">length</em>, <em class="sig-param">inputs</em>, <em class="sig-param">begin_state=None</em>, <em class="sig-param">layout='NTC'</em>, <em class="sig-param">merge_outputs=None</em>, <em class="sig-param">valid_length=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#VariationalDropoutCell.unroll"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.VariationalDropoutCell.unroll" title="Permalink to this definition"></a></dt>
<dd><p>Unrolls an RNN cell across time steps.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>length</strong> (<em>int</em>) – Number of steps to unroll.</p></li>
<li><p><strong>inputs</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em>, </em><em>list of Symbol</em><em>, or </em><em>None</em>) – <p>If <cite>inputs</cite> is a single Symbol (usually the output
of Embedding symbol), it should have shape
(batch_size, length, …) if <cite>layout</cite> is ‘NTC’,
or (length, batch_size, …) if <cite>layout</cite> is ‘TNC’.</p>
<p>If <cite>inputs</cite> is a list of symbols (usually output of
previous unroll), they should all have shape
(batch_size, …).</p>
</p></li>
<li><p><strong>begin_state</strong> (<em>nested list of Symbol</em><em>, </em><em>optional</em>) – Input states created by <cite>begin_state()</cite>
or output state of another cell.
Created from <cite>begin_state()</cite> if <cite>None</cite>.</p></li>
<li><p><strong>layout</strong> (<em>str</em><em>, </em><em>optional</em>) – <cite>layout</cite> of input symbol. Only used if inputs
is a single Symbol.</p></li>
<li><p><strong>merge_outputs</strong> (<em>bool</em><em>, </em><em>optional</em>) – If <cite>False</cite>, returns outputs as a list of Symbols.
If <cite>True</cite>, concatenates output across time steps
and returns a single symbol with shape
(batch_size, length, …) if layout is ‘NTC’,
or (length, batch_size, …) if layout is ‘TNC’.
If <cite>None</cite>, output whatever is faster.</p></li>
<li><p><strong>valid_length</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em>, </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a><em> or </em><em>None</em>) – <cite>valid_length</cite> specifies the length of the sequences in the batch without padding.
This option is especially useful for building sequence-to-sequence models where
the input and output sequences would potentially be padded.
If <cite>valid_length</cite> is None, all sequences are assumed to have the same length.
If <cite>valid_length</cite> is a Symbol or NDArray, it should have shape (batch_size,).
The ith element will be the length of the ith sequence in the batch.
The last valid state will be return and the padded outputs will be masked with 0.
Note that <cite>valid_length</cite> must be smaller or equal to <cite>length</cite>.</p></li>
</ul>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p><ul class="simple">
<li><p><strong>outputs</strong> (<em>list of Symbol or Symbol</em>) – Symbol (if <cite>merge_outputs</cite> is True) or list of Symbols
(if <cite>merge_outputs</cite> is False) corresponding to the output from
the RNN from this unrolling.</p></li>
<li><p><strong>states</strong> (<em>list of Symbol</em>) – The new state of this RNN after this unrolling.
The type of this symbol is same as the output of <cite>begin_state()</cite>.</p></li>
</ul>
</p>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.rnn.LSTMPCell">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.rnn.</code><code class="sig-name descname">LSTMPCell</code><span class="sig-paren">(</span><em class="sig-param">hidden_size</em>, <em class="sig-param">projection_size</em>, <em class="sig-param">i2h_weight_initializer=None</em>, <em class="sig-param">h2h_weight_initializer=None</em>, <em class="sig-param">h2r_weight_initializer=None</em>, <em class="sig-param">i2h_bias_initializer='zeros'</em>, <em class="sig-param">h2h_bias_initializer='zeros'</em>, <em class="sig-param">input_size=0</em>, <em class="sig-param">prefix=None</em>, <em class="sig-param">params=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#LSTMPCell"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.LSTMPCell" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.rnn.rnn_cell.HybridRecurrentCell</span></code></p>
<p>Long-Short Term Memory Projected (LSTMP) network cell.
(<a class="reference external" href="https://arxiv.org/abs/1402.1128">https://arxiv.org/abs/1402.1128</a>)</p>
<p>Each call computes the following function:</p>
<div class="math notranslate nohighlight">
\[\begin{split}\begin{array}{ll}
i_t = sigmoid(W_{ii} x_t + b_{ii} + W_{ri} r_{(t-1)} + b_{ri}) \\
f_t = sigmoid(W_{if} x_t + b_{if} + W_{rf} r_{(t-1)} + b_{rf}) \\
g_t = \tanh(W_{ig} x_t + b_{ig} + W_{rc} r_{(t-1)} + b_{rg}) \\
o_t = sigmoid(W_{io} x_t + b_{io} + W_{ro} r_{(t-1)} + b_{ro}) \\
c_t = f_t * c_{(t-1)} + i_t * g_t \\
h_t = o_t * \tanh(c_t) \\
r_t = W_{hr} h_t
\end{array}\end{split}\]</div>
<p>where <span class="math notranslate nohighlight">\(r_t\)</span> is the projected recurrent activation at time <cite>t</cite>,
<span class="math notranslate nohighlight">\(h_t\)</span> is the hidden state at time <cite>t</cite>, <span class="math notranslate nohighlight">\(c_t\)</span> is the
cell state at time <cite>t</cite>, <span class="math notranslate nohighlight">\(x_t\)</span> is the input at time <cite>t</cite>, and <span class="math notranslate nohighlight">\(i_t\)</span>,
<span class="math notranslate nohighlight">\(f_t\)</span>, <span class="math notranslate nohighlight">\(g_t\)</span>, <span class="math notranslate nohighlight">\(o_t\)</span> are the input, forget, cell, and
out gates, respectively.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>hidden_size</strong> (<em>int</em>) – Number of units in cell state symbol.</p></li>
<li><p><strong>projection_size</strong> (<em>int</em>) – Number of units in output symbol.</p></li>
<li><p><strong>i2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the input weights matrix, used for the linear
transformation of the inputs.</p></li>
<li><p><strong>h2h_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the recurrent weights matrix, used for the linear
transformation of the hidden state.</p></li>
<li><p><strong>h2r_weight_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the projection weights matrix, used for the linear
transformation of the recurrent state.</p></li>
<li><p><strong>i2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a><em>, </em><em>default 'lstmbias'</em>) – Initializer for the bias vector. By default, bias for the forget
gate is initialized to 1 while all other biases are initialized
to zero.</p></li>
<li><p><strong>h2h_bias_initializer</strong> (<em>str</em><em> or </em><a class="reference internal" href="../../initializer/index.html#mxnet.initializer.Initializer" title="mxnet.initializer.Initializer"><em>Initializer</em></a>) – Initializer for the bias vector.</p></li>
<li><p><strong>prefix</strong> (str, default <code class="docutils literal notranslate"><span class="pre">'lstmp_</span></code>’) – Prefix for name of <cite>Block`s
(and name of weight if params is `None</cite>).</p></li>
<li><p><strong>params</strong> (<a class="reference internal" href="../parameter.html#mxnet.gluon.Parameter" title="mxnet.gluon.Parameter"><em>Parameter</em></a><em> or </em><em>None</em>) – Container for weight sharing between cells.
Created if <cite>None</cite>.</p></li>
<li><p><strong>Inputs</strong> – <ul>
<li><p><strong>data</strong>: input tensor with shape <cite>(batch_size, input_size)</cite>.</p></li>
<li><p><strong>states</strong>: a list of two initial recurrent state tensors, with shape
<cite>(batch_size, projection_size)</cite> and <cite>(batch_size, hidden_size)</cite> respectively.</p></li>
</ul>
</p></li>
<li><p><strong>Outputs</strong> – <ul>
<li><p><strong>out</strong>: output tensor with shape <cite>(batch_size, num_hidden)</cite>.</p></li>
<li><p><strong>next_states</strong>: a list of two output recurrent state tensors. Each has
the same shape as <cite>states</cite>.</p></li>
</ul>
</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.rnn.LSTMPCell.hybrid_forward">
<code class="sig-name descname">hybrid_forward</code><span class="sig-paren">(</span><em class="sig-param">F</em>, <em class="sig-param">inputs</em>, <em class="sig-param">states</em>, <em class="sig-param">i2h_weight</em>, <em class="sig-param">h2h_weight</em>, <em class="sig-param">h2r_weight</em>, <em class="sig-param">i2h_bias</em>, <em class="sig-param">h2h_bias</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#LSTMPCell.hybrid_forward"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.LSTMPCell.hybrid_forward" title="Permalink to this definition"></a></dt>
<dd><p>Overrides to construct symbolic graph for this <cite>Block</cite>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>x</strong> (<a class="reference internal" href="../../symbol/symbol.html#mxnet.symbol.Symbol" title="mxnet.symbol.Symbol"><em>Symbol</em></a><em> or </em><a class="reference internal" href="../../ndarray/ndarray.html#mxnet.ndarray.NDArray" title="mxnet.ndarray.NDArray"><em>NDArray</em></a>) – The first input tensor.</p></li>
<li><p><strong>*args</strong> (<em>list of Symbol</em><em> or </em><em>list of NDArray</em>) – Additional input tensors.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.rnn.LSTMPCell.state_info">
<code class="sig-name descname">state_info</code><span class="sig-paren">(</span><em class="sig-param">batch_size=0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/rnn/rnn_cell.html#LSTMPCell.state_info"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.rnn.LSTMPCell.state_info" title="Permalink to this definition"></a></dt>
<dd><p>shape and layout information of states</p>
</dd></dl>
</dd></dl>
<span class="target" id="module-mxnet.gluon.contrib.data.sampler"></span><p>Dataset sampler.</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.data.sampler.IntervalSampler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.data.sampler.</code><code class="sig-name descname">IntervalSampler</code><span class="sig-paren">(</span><em class="sig-param">length</em>, <em class="sig-param">interval</em>, <em class="sig-param">rollover=True</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/data/sampler.html#IntervalSampler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.data.sampler.IntervalSampler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.data.sampler.Sampler</span></code></p>
<p>Samples elements from [0, length) at fixed intervals.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>length</strong> (<em>int</em>) – Length of the sequence.</p></li>
<li><p><strong>interval</strong> (<em>int</em>) – The number of items to skip between two samples.</p></li>
<li><p><strong>rollover</strong> (<em>bool</em><em>, </em><em>default True</em>) – Whether to start again from the first skipped item after reaching the end.
If true, this sampler would start again from the first skipped item until all items
are visited.
Otherwise, iteration stops when end is reached and skipped items are ignored.</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="n">sampler</span> <span class="o">=</span> <span class="n">contrib</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">IntervalSampler</span><span class="p">(</span><span class="mi">13</span><span class="p">,</span> <span class="n">interval</span><span class="o">=</span><span class="mi">3</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">list</span><span class="p">(</span><span class="n">sampler</span><span class="p">)</span>
<span class="go">[0, 3, 6, 9, 12, 1, 4, 7, 10, 2, 5, 8, 11]</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">sampler</span> <span class="o">=</span> <span class="n">contrib</span><span class="o">.</span><span class="n">data</span><span class="o">.</span><span class="n">IntervalSampler</span><span class="p">(</span><span class="mi">13</span><span class="p">,</span> <span class="n">interval</span><span class="o">=</span><span class="mi">3</span><span class="p">,</span> <span class="n">rollover</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="nb">list</span><span class="p">(</span><span class="n">sampler</span><span class="p">)</span>
<span class="go">[0, 3, 6, 9, 12]</span>
</pre></div>
</div>
</dd></dl>
<span class="target" id="module-mxnet.gluon.contrib.data.text"></span><p>Text datasets.</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.data.text.WikiText2">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.data.text.</code><code class="sig-name descname">WikiText2</code><span class="sig-paren">(</span><em class="sig-param">root='/home/jenkins_slave/.mxnet/datasets/wikitext-2'</em>, <em class="sig-param">segment='train'</em>, <em class="sig-param">vocab=None</em>, <em class="sig-param">seq_len=35</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/data/text.html#WikiText2"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.data.text.WikiText2" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.data.text._WikiText</span></code></p>
<p>WikiText-2 word-level dataset for language modeling, from Salesforce research.</p>
<p>From
<a class="reference external" href="https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/">https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/</a></p>
<p>License: Creative Commons Attribution-ShareAlike</p>
<p>Each sample is a vector of length equal to the specified sequence length.
At the end of each sentence, an end-of-sentence token ‘&lt;eos&gt;’ is added.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>root</strong> (<em>str</em><em>, </em><em>default $MXNET_HOME/datasets/wikitext-2</em>) – Path to temp folder for storing data.</p></li>
<li><p><strong>segment</strong> (<em>str</em><em>, </em><em>default 'train'</em>) – Dataset segment. Options are ‘train’, ‘validation’, ‘test’.</p></li>
<li><p><strong>vocab</strong> (<code class="xref py py-class docutils literal notranslate"><span class="pre">Vocabulary</span></code>, default None) – The vocabulary to use for indexing the text dataset.
If None, a default vocabulary is created.</p></li>
<li><p><strong>seq_len</strong> (<em>int</em><em>, </em><em>default 35</em>) – The sequence length of each sample, regardless of the sentence boundary.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.data.text.WikiText103">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.data.text.</code><code class="sig-name descname">WikiText103</code><span class="sig-paren">(</span><em class="sig-param">root='/home/jenkins_slave/.mxnet/datasets/wikitext-103'</em>, <em class="sig-param">segment='train'</em>, <em class="sig-param">vocab=None</em>, <em class="sig-param">seq_len=35</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/data/text.html#WikiText103"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.data.text.WikiText103" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.data.text._WikiText</span></code></p>
<p>WikiText-103 word-level dataset for language modeling, from Salesforce research.</p>
<p>From
<a class="reference external" href="https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/">https://blog.einstein.ai/the-wikitext-long-term-dependency-language-modeling-dataset/</a></p>
<p>License: Creative Commons Attribution-ShareAlike</p>
<p>Each sample is a vector of length equal to the specified sequence length.
At the end of each sentence, an end-of-sentence token ‘&lt;eos&gt;’ is added.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>root</strong> (<em>str</em><em>, </em><em>default $MXNET_HOME/datasets/wikitext-103</em>) – Path to temp folder for storing data.</p></li>
<li><p><strong>segment</strong> (<em>str</em><em>, </em><em>default 'train'</em>) – Dataset segment. Options are ‘train’, ‘validation’, ‘test’.</p></li>
<li><p><strong>vocab</strong> (<code class="xref py py-class docutils literal notranslate"><span class="pre">Vocabulary</span></code>, default None) – The vocabulary to use for indexing the text dataset.
If None, a default vocabulary is created.</p></li>
<li><p><strong>seq_len</strong> (<em>int</em><em>, </em><em>default 35</em>) – The sequence length of each sample, regardless of the sentence boundary.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<span class="target" id="module-mxnet.gluon.contrib.estimator"></span><p>Gluon Estimator Module</p>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.BatchProcessor">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">BatchProcessor</code><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/batch_processor.html#BatchProcessor"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.BatchProcessor" 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>BatchProcessor Class for plug and play fit_batch &amp; evaluate_batch</p>
<p>During training or validation, data are divided into minibatches for processing. This
class aims at providing hooks of training or validating on a minibatch of data. Users
may provide customized fit_batch() and evaluate_batch() methods by inheriting from
this class and overriding class methods.</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.BatchProcessor" title="mxnet.gluon.contrib.estimator.BatchProcessor"><code class="xref py py-class docutils literal notranslate"><span class="pre">BatchProcessor</span></code></a> can be used to replace fit_batch() and evaluate_batch()
in the base estimator class</p>
<dl class="method">
<dt id="mxnet.gluon.contrib.estimator.BatchProcessor.evaluate_batch">
<code class="sig-name descname">evaluate_batch</code><span class="sig-paren">(</span><em class="sig-param">estimator</em>, <em class="sig-param">val_batch</em>, <em class="sig-param">batch_axis=0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/batch_processor.html#BatchProcessor.evaluate_batch"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.BatchProcessor.evaluate_batch" title="Permalink to this definition"></a></dt>
<dd><p>Evaluate the estimator model on a batch of validation data.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>estimator</strong> (<a class="reference internal" href="#mxnet.gluon.contrib.estimator.Estimator" title="mxnet.gluon.contrib.estimator.Estimator"><em>Estimator</em></a>) – Reference to the estimator</p></li>
<li><p><strong>val_batch</strong> (<em>tuple</em>) – Data and label of a batch from the validation data loader.</p></li>
<li><p><strong>batch_axis</strong> (<em>int</em><em>, </em><em>default 0</em>) – Batch axis to split the validation data into devices.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.estimator.BatchProcessor.fit_batch">
<code class="sig-name descname">fit_batch</code><span class="sig-paren">(</span><em class="sig-param">estimator</em>, <em class="sig-param">train_batch</em>, <em class="sig-param">batch_axis=0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/batch_processor.html#BatchProcessor.fit_batch"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.BatchProcessor.fit_batch" title="Permalink to this definition"></a></dt>
<dd><p>Trains the estimator model on a batch of training data.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>estimator</strong> (<a class="reference internal" href="#mxnet.gluon.contrib.estimator.Estimator" title="mxnet.gluon.contrib.estimator.Estimator"><em>Estimator</em></a>) – Reference to the estimator</p></li>
<li><p><strong>train_batch</strong> (<em>tuple</em>) – Data and label of a batch from the training data loader.</p></li>
<li><p><strong>batch_axis</strong> (<em>int</em><em>, </em><em>default 0</em>) – Batch axis to split the training data into devices.</p></li>
</ul>
</dd>
<dt class="field-even">Returns</dt>
<dd class="field-even"><p><ul class="simple">
<li><p><strong>data</strong> (<em>List of NDArray</em>) – Sharded data from the batch. Data is sharded with
<cite>gluon.split_and_load</cite>.</p></li>
<li><p><strong>label</strong> (<em>List of NDArray</em>) – Sharded label from the batch. Labels are sharded with
<cite>gluon.split_and_load</cite>.</p></li>
<li><p><strong>pred</strong> (<em>List of NDArray</em>) – Prediction on each of the sharded inputs.</p></li>
<li><p><strong>loss</strong> (<em>List of NDArray</em>) – Loss on each of the sharded inputs.</p></li>
</ul>
</p>
</dd>
</dl>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.CheckpointHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">CheckpointHandler</code><span class="sig-paren">(</span><em class="sig-param">model_dir</em>, <em class="sig-param">model_prefix='model'</em>, <em class="sig-param">monitor=None</em>, <em class="sig-param">verbose=0</em>, <em class="sig-param">save_best=False</em>, <em class="sig-param">mode='auto'</em>, <em class="sig-param">epoch_period=1</em>, <em class="sig-param">batch_period=None</em>, <em class="sig-param">max_checkpoints=5</em>, <em class="sig-param">resume_from_checkpoint=False</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#CheckpointHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochEnd</span></code></p>
<p>Save the model after user define period</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="mxnet.gluon.contrib.estimator.CheckpointHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">CheckpointHandler</span></code></a> saves the network architecture after first batch if the model
can be fully hybridized, saves model parameters and trainer states after user defined period,
default saves every epoch.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>model_dir</strong> (<em>str</em>) – File directory to save all the model related files including model architecture,
model parameters, and trainer states.</p></li>
<li><p><strong>model_prefix</strong> (<em>str default 'model'</em>) – Prefix to add for all checkpoint file names.</p></li>
<li><p><strong>monitor</strong> (<a class="reference internal" href="../../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a><em>, </em><em>default None</em>) – The metrics to monitor and determine if model has improved</p></li>
<li><p><strong>verbose</strong> (<em>int</em><em>, </em><em>default 0</em>) – Verbosity mode, 1 means inform user every time a checkpoint is saved</p></li>
<li><p><strong>save_best</strong> (<em>bool</em><em>, </em><em>default False</em>) – If True, monitor must not be None, <a class="reference internal" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="mxnet.gluon.contrib.estimator.CheckpointHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">CheckpointHandler</span></code></a> will save the
model parameters and trainer states with the best monitored value.</p></li>
<li><p><strong>mode</strong> (<em>str</em><em>, </em><em>default 'auto'</em>) – One of {auto, min, max}, if <cite>save_best=True</cite>, the comparison to make
and determine if the monitored value has improved. if ‘auto’ mode,
<a class="reference internal" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="mxnet.gluon.contrib.estimator.CheckpointHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">CheckpointHandler</span></code></a> will try to use min or max based on
the monitored metric name.</p></li>
<li><p><strong>epoch_period</strong> (<em>int</em><em>, </em><em>default 1</em>) – Epoch intervals between saving the network. By default, checkpoints are
saved every epoch.</p></li>
<li><p><strong>batch_period</strong> (<em>int</em><em>, </em><em>default None</em>) – Batch intervals between saving the network.
By default, checkpoints are not saved based on the number of batches.</p></li>
<li><p><strong>max_checkpoints</strong> (<em>int</em><em>, </em><em>default 5</em>) – Maximum number of checkpoint files to keep in the model_dir, older checkpoints
will be removed. Best checkpoint file is not counted.</p></li>
<li><p><strong>resume_from_checkpoint</strong> (<em>bool</em><em>, </em><em>default False</em>) – Whether to resume training from checkpoint in model_dir. If True and checkpoints
found, <a class="reference internal" href="#mxnet.gluon.contrib.estimator.CheckpointHandler" title="mxnet.gluon.contrib.estimator.CheckpointHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">CheckpointHandler</span></code></a> will load net parameters and trainer states,
and train the remaining of epochs and batches.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.EarlyStoppingHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">EarlyStoppingHandler</code><span class="sig-paren">(</span><em class="sig-param">monitor</em>, <em class="sig-param">min_delta=0</em>, <em class="sig-param">patience=0</em>, <em class="sig-param">mode='auto'</em>, <em class="sig-param">baseline=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#EarlyStoppingHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.EarlyStoppingHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainEnd</span></code></p>
<p>Early stop training if monitored value is not improving</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>monitor</strong> (<a class="reference internal" href="../../metric/index.html#mxnet.metric.EvalMetric" title="mxnet.metric.EvalMetric"><em>EvalMetric</em></a>) – The metric to monitor, and stop training if this metric does not improve.</p></li>
<li><p><strong>min_delta</strong> (<em>float</em><em>, </em><em>default 0</em>) – Minimal change in monitored value to be considered as an improvement.</p></li>
<li><p><strong>patience</strong> (<em>int</em><em>, </em><em>default 0</em>) – Number of epochs to wait for improvement before terminate training.</p></li>
<li><p><strong>mode</strong> (<em>str</em><em>, </em><em>default 'auto'</em>) – One of {auto, min, max}, if <cite>save_best_only=True</cite>, the comparison to make
and determine if the monitored value has improved. if ‘auto’ mode, checkpoint
handler will try to use min or max based on the monitored metric name.</p></li>
<li><p><strong>baseline</strong> (<em>float</em>) – Baseline value to compare the monitored value with.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.Estimator">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">Estimator</code><span class="sig-paren">(</span><em class="sig-param">net</em>, <em class="sig-param">loss</em>, <em class="sig-param">train_metrics=None</em>, <em class="sig-param">val_metrics=None</em>, <em class="sig-param">initializer=None</em>, <em class="sig-param">trainer=None</em>, <em class="sig-param">context=None</em>, <em class="sig-param">val_net=None</em>, <em class="sig-param">val_loss=None</em>, <em class="sig-param">batch_processor=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/estimator.html#Estimator"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.Estimator" 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>Estimator Class for easy model training</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.Estimator" title="mxnet.gluon.contrib.estimator.Estimator"><code class="xref py py-class docutils literal notranslate"><span class="pre">Estimator</span></code></a> can be used to facilitate the training &amp; validation process</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>net</strong> (<a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a>) – The model used for training.</p></li>
<li><p><strong>loss</strong> (<a class="reference internal" href="../loss/index.html#mxnet.gluon.loss.Loss" title="mxnet.gluon.loss.Loss"><em>gluon.loss.Loss</em></a>) – Loss (objective) function to calculate during training.</p></li>
<li><p><strong>train_metrics</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 EvalMetric</em>) – Training metrics for evaluating models on training dataset.</p></li>
<li><p><strong>val_metrics</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 EvalMetric</em>) – Validation metrics for evaluating models on validation dataset.</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>) – Initializer to initialize the network.</p></li>
<li><p><strong>trainer</strong> (<a class="reference internal" href="../trainer.html#mxnet.gluon.Trainer" title="mxnet.gluon.Trainer"><em>Trainer</em></a>) – Trainer to apply optimizer on network parameters.</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>) – Device(s) to run the training on.</p></li>
<li><p><strong>val_net</strong> (<a class="reference internal" href="../block.html#mxnet.gluon.Block" title="mxnet.gluon.Block"><em>gluon.Block</em></a>) – <p>The model used for validation. The validation model does not necessarily belong to
the same model class as the training model. But the two models typically share the
same architecture. Therefore the validation model can reuse parameters of the
training model.</p>
<p>The code example of consruction of val_net sharing the same network parameters as
the training net is given below:</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="n">net</span> <span class="o">=</span> <span class="n">_get_train_network</span><span class="p">()</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">val_net</span> <span class="o">=</span> <span class="n">_get_test_network</span><span class="p">(</span><span class="n">params</span><span class="o">=</span><span class="n">net</span><span class="o">.</span><span class="n">collect_params</span><span class="p">())</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">net</span><span class="o">.</span><span class="n">initialize</span><span class="p">(</span><span class="n">ctx</span><span class="o">=</span><span class="n">ctx</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">est</span> <span class="o">=</span> <span class="n">Estimator</span><span class="p">(</span><span class="n">net</span><span class="p">,</span> <span class="n">loss</span><span class="p">,</span> <span class="n">val_net</span><span class="o">=</span><span class="n">val_net</span><span class="p">)</span>
</pre></div>
</div>
<p>Proper namespace match is required for weight sharing between two networks. Most networks
inheriting <code class="xref py py-class docutils literal notranslate"><span class="pre">Block</span></code> can share their parameters correctly. An exception is
Sequential networks that Block scope must be specified for correct weight sharing. For
the naming in mxnet Gluon API, please refer to the site
(<a class="reference external" href="https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/blocks/naming.html">https://mxnet.apache.org/api/python/docs/tutorials/packages/gluon/blocks/naming.html</a>)
for future information.</p>
</p></li>
<li><p><strong>val_loss</strong> (<em>gluon.loss.loss</em>) – Loss (objective) function to calculate during validation. If set val_loss
None, it will use the same loss function as self.loss</p></li>
<li><p><strong>batch_processor</strong> (<a class="reference internal" href="#mxnet.gluon.contrib.estimator.BatchProcessor" title="mxnet.gluon.contrib.estimator.BatchProcessor"><em>BatchProcessor</em></a>) – BatchProcessor provides customized fit_batch() and evaluate_batch() methods</p></li>
</ul>
</dd>
</dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.estimator.Estimator.evaluate">
<code class="sig-name descname">evaluate</code><span class="sig-paren">(</span><em class="sig-param">val_data</em>, <em class="sig-param">batch_axis=0</em>, <em class="sig-param">event_handlers=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/estimator.html#Estimator.evaluate"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.Estimator.evaluate" title="Permalink to this definition"></a></dt>
<dd><p>Evaluate model on validation data.</p>
<p>This function calls <code class="xref py py-func docutils literal notranslate"><span class="pre">evaluate_batch()</span></code> on each of the batches from the
validation data loader. Thus, for custom use cases, it’s possible to inherit the
estimator class and override <code class="xref py py-func docutils literal notranslate"><span class="pre">evaluate_batch()</span></code>.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>val_data</strong> (<a class="reference internal" href="../data/index.html#mxnet.gluon.data.DataLoader" title="mxnet.gluon.data.DataLoader"><em>DataLoader</em></a>) – Validation data loader with data and labels.</p></li>
<li><p><strong>batch_axis</strong> (<em>int</em><em>, </em><em>default 0</em>) – Batch axis to split the validation data into devices.</p></li>
<li><p><strong>event_handlers</strong> (<em>EventHandler</em><em> or </em><em>list of EventHandler</em>) – List of <code class="xref py py-class docutils literal notranslate"><span class="pre">EventHandlers</span></code> to apply during validation. Besides
event handlers specified here, a default MetricHandler and a LoggingHandler
will be added if not specified explicitly.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="method">
<dt id="mxnet.gluon.contrib.estimator.Estimator.fit">
<code class="sig-name descname">fit</code><span class="sig-paren">(</span><em class="sig-param">train_data</em>, <em class="sig-param">val_data=None</em>, <em class="sig-param">epochs=None</em>, <em class="sig-param">event_handlers=None</em>, <em class="sig-param">batches=None</em>, <em class="sig-param">batch_axis=0</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/estimator.html#Estimator.fit"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.Estimator.fit" title="Permalink to this definition"></a></dt>
<dd><p>Trains the model with a given <code class="xref py py-class docutils literal notranslate"><span class="pre">DataLoader</span></code> for a specified
number of epochs or batches. The batch size is inferred from the
data loader’s batch_size.</p>
<p>This function calls <code class="xref py py-func docutils literal notranslate"><span class="pre">fit_batch()</span></code> on each of the batches from the
training data loader. Thus, for custom use cases, it’s possible to inherit the
estimator class and override <code class="xref py py-func docutils literal notranslate"><span class="pre">fit_batch()</span></code>.</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="../data/index.html#mxnet.gluon.data.DataLoader" title="mxnet.gluon.data.DataLoader"><em>DataLoader</em></a>) – Training data loader with data and labels.</p></li>
<li><p><strong>val_data</strong> (<a class="reference internal" href="../data/index.html#mxnet.gluon.data.DataLoader" title="mxnet.gluon.data.DataLoader"><em>DataLoader</em></a><em>, </em><em>default None</em>) – Validation data loader with data and labels.</p></li>
<li><p><strong>epochs</strong> (<em>int</em><em>, </em><em>default None</em>) – Number of epochs to iterate on the training data.
You can only specify one and only one type of iteration(epochs or batches).</p></li>
<li><p><strong>event_handlers</strong> (<em>EventHandler</em><em> or </em><em>list of EventHandler</em>) – List of <code class="xref py py-class docutils literal notranslate"><span class="pre">EventHandlers</span></code> to apply during training. Besides
the event handlers specified here, a StoppingHandler,
LoggingHandler and MetricHandler will be added by default if not
yet specified manually. If validation data is provided, a
ValidationHandler is also added if not already specified.</p></li>
<li><p><strong>batches</strong> (<em>int</em><em>, </em><em>default None</em>) – Number of batches to iterate on the training data.
You can only specify one and only one type of iteration(epochs or batches).</p></li>
<li><p><strong>batch_axis</strong> (<em>int</em><em>, </em><em>default 0</em>) – Batch axis to split the training data into devices.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="attribute">
<dt id="mxnet.gluon.contrib.estimator.Estimator.logger">
<code class="sig-name descname">logger</code><em class="property"> = None</em><a class="headerlink" href="#mxnet.gluon.contrib.estimator.Estimator.logger" title="Permalink to this definition"></a></dt>
<dd><p>logging.Logger object associated with the Estimator.</p>
<p>The logger is used for all logs generated by this estimator and its
handlers. A new logging.Logger is created during Estimator construction and
configured to write all logs with level logging.INFO or higher to
sys.stdout.</p>
<p>You can modify the logging settings using the standard Python methods. For
example, to save logs to a file in addition to printing them to stdout
output, you can attach a logging.FileHandler to the logger.</p>
<div class="doctest highlight-default notranslate"><div class="highlight"><pre><span></span><span class="gp">&gt;&gt;&gt; </span><span class="n">est</span> <span class="o">=</span> <span class="n">Estimator</span><span class="p">(</span><span class="n">net</span><span class="p">,</span> <span class="n">loss</span><span class="p">)</span>
<span class="gp">&gt;&gt;&gt; </span><span class="kn">import</span> <span class="nn">logging</span>
<span class="gp">&gt;&gt;&gt; </span><span class="n">est</span><span class="o">.</span><span class="n">logger</span><span class="o">.</span><span class="n">addHandler</span><span class="p">(</span><span class="n">logging</span><span class="o">.</span><span class="n">FileHandler</span><span class="p">(</span><span class="n">filename</span><span class="p">))</span>
</pre></div>
</div>
</dd></dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.GradientUpdateHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">GradientUpdateHandler</code><span class="sig-paren">(</span><em class="sig-param">priority=-2000</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#GradientUpdateHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.GradientUpdateHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code></p>
<p>Gradient Update Handler that apply gradients on network weights</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.GradientUpdateHandler" title="mxnet.gluon.contrib.estimator.GradientUpdateHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">GradientUpdateHandler</span></code></a> takes the priority level. It updates weight parameters
at the end of each batch</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>priority</strong> (<em>scalar</em><em>, </em><em>default -2000</em>) – priority level of the gradient update handler. Priority level is sorted in ascending
order. The lower the number is, the higher priority level the handler is.</p></li>
<li><p><strong>----------</strong> – </p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.LoggingHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">LoggingHandler</code><span class="sig-paren">(</span><em class="sig-param">log_interval='epoch'</em>, <em class="sig-param">metrics=None</em>, <em class="sig-param">priority=inf</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#LoggingHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.LoggingHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code></p>
<p>Basic Logging Handler that applies to every Gluon estimator by default.</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.LoggingHandler" title="mxnet.gluon.contrib.estimator.LoggingHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">LoggingHandler</span></code></a> logs hyper-parameters, training statistics,
and other useful information during training</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>log_interval</strong> (<em>int</em><em> or </em><em>str</em><em>, </em><em>default 'epoch'</em>) – Logging interval during training.
log_interval=’epoch’: display metrics every epoch
log_interval=integer k: display metrics every interval of k batches</p></li>
<li><p><strong>metrics</strong> (<em>list of EvalMetrics</em>) – Metrics to be logged, logged at batch end, epoch end, train end.</p></li>
<li><p><strong>priority</strong> (<em>scalar</em><em>, </em><em>default np.Inf</em>) – Priority level of the LoggingHandler. Priority level is sorted in
ascending order. The lower the number is, the higher priority level the
handler is.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.MetricHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">MetricHandler</code><span class="sig-paren">(</span><em class="sig-param">metrics</em>, <em class="sig-param">priority=-1000</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#MetricHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.MetricHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code></p>
<p>Metric Handler that update metric values at batch end</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.MetricHandler" title="mxnet.gluon.contrib.estimator.MetricHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">MetricHandler</span></code></a> takes model predictions and true labels
and update the metrics, it also update metric wrapper for loss with loss values.
Validation loss and metrics will be handled by <a class="reference internal" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="mxnet.gluon.contrib.estimator.ValidationHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">ValidationHandler</span></code></a></p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>metrics</strong> (<em>List of EvalMetrics</em>) – Metrics to be updated at batch end.</p></li>
<li><p><strong>priority</strong> (<em>scalar</em>) – Priority level of the MetricHandler. Priority level is sorted in ascending
order. The lower the number is, the higher priority level the handler is.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.StoppingHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">StoppingHandler</code><span class="sig-paren">(</span><em class="sig-param">max_epoch=None</em>, <em class="sig-param">max_batch=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#StoppingHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.StoppingHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochEnd</span></code></p>
<p>Stop conditions to stop training
Stop training if maximum number of batches or epochs
reached.</p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>max_epoch</strong> (<em>int</em><em>, </em><em>default None</em>) – Number of maximum epochs to train.</p></li>
<li><p><strong>max_batch</strong> (<em>int</em><em>, </em><em>default None</em>) – Number of maximum batches to train.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
<dl class="class">
<dt id="mxnet.gluon.contrib.estimator.ValidationHandler">
<em class="property">class </em><code class="sig-prename descclassname">mxnet.gluon.contrib.estimator.</code><code class="sig-name descname">ValidationHandler</code><span class="sig-paren">(</span><em class="sig-param">val_data</em>, <em class="sig-param">eval_fn</em>, <em class="sig-param">epoch_period=1</em>, <em class="sig-param">batch_period=None</em>, <em class="sig-param">priority=-1000</em>, <em class="sig-param">event_handlers=None</em><span class="sig-paren">)</span><a class="reference internal" href="../../../_modules/mxnet/gluon/contrib/estimator/event_handler.html#ValidationHandler"><span class="viewcode-link">[source]</span></a><a class="headerlink" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="Permalink to this definition"></a></dt>
<dd><p>Bases: <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.TrainBegin</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.BatchEnd</span></code>, <code class="xref py py-class docutils literal notranslate"><span class="pre">mxnet.gluon.contrib.estimator.event_handler.EpochEnd</span></code></p>
<p>Validation Handler that evaluate model on validation dataset</p>
<p><a class="reference internal" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="mxnet.gluon.contrib.estimator.ValidationHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">ValidationHandler</span></code></a> takes validation dataset, an evaluation function,
metrics to be evaluated, and how often to run the validation. You can provide custom
evaluation function or use the one provided my <a class="reference internal" href="#mxnet.gluon.contrib.estimator.Estimator" title="mxnet.gluon.contrib.estimator.Estimator"><code class="xref py py-class docutils literal notranslate"><span class="pre">Estimator</span></code></a></p>
<dl class="field-list simple">
<dt class="field-odd">Parameters</dt>
<dd class="field-odd"><ul class="simple">
<li><p><strong>val_data</strong> (<a class="reference internal" href="../data/index.html#mxnet.gluon.data.DataLoader" title="mxnet.gluon.data.DataLoader"><em>DataLoader</em></a>) – Validation data set to run evaluation.</p></li>
<li><p><strong>eval_fn</strong> (<em>function</em>) – A function defines how to run evaluation and
calculate loss and metrics.</p></li>
<li><p><strong>epoch_period</strong> (<em>int</em><em>, </em><em>default 1</em>) – How often to run validation at epoch end, by default
<a class="reference internal" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="mxnet.gluon.contrib.estimator.ValidationHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">ValidationHandler</span></code></a> validate every epoch.</p></li>
<li><p><strong>batch_period</strong> (<em>int</em><em>, </em><em>default None</em>) – How often to run validation at batch end, by default
<a class="reference internal" href="#mxnet.gluon.contrib.estimator.ValidationHandler" title="mxnet.gluon.contrib.estimator.ValidationHandler"><code class="xref py py-class docutils literal notranslate"><span class="pre">ValidationHandler</span></code></a> does not validate at batch end.</p></li>
<li><p><strong>priority</strong> (<em>scalar</em><em>, </em><em>default -1000</em>) – Priority level of the ValidationHandler. Priority level is sorted in
ascending order. The lower the number is, the higher priority level the
handler is.</p></li>
<li><p><strong>event_handlers</strong> (<em>EventHandler</em><em> or </em><em>list of EventHandlers</em>) – List of <code class="xref py py-class docutils literal notranslate"><span class="pre">EventHandler</span></code> to apply during validaiton. This argument
is used by self.eval_fn function in order to process customized event
handlers.</p></li>
</ul>
</dd>
</dl>
</dd></dl>
</div>
</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 class="localtoc">
<p class="caption">
<span class="caption-text">Table Of Contents</span>
</p>
<ul>
<li><a class="reference internal" href="#">gluon.contrib</a><ul>
<li><a class="reference internal" href="#neural-network">Neural Network</a></li>
<li><a class="reference internal" href="#convolutional-neural-network">Convolutional Neural Network</a></li>
<li><a class="reference internal" href="#recurrent-neural-network">Recurrent Neural Network</a></li>
<li><a class="reference internal" href="#data">Data</a></li>
<li><a class="reference internal" href="#text-dataset">Text Dataset</a></li>
<li><a class="reference internal" href="#estimator">Estimator</a></li>
<li><a class="reference internal" href="#event-handler">Event Handler</a></li>
<li><a class="reference internal" href="#module-mxnet.gluon.contrib">API Reference</a></li>
</ul>
</li>
</ul>
</div>
</div>
</div>
<div class="clearer"></div>
</div><div class="pagenation">
<a id="button-prev" href="../trainer.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>gluon.Trainer</div>
</div>
</a>
<a id="button-next" href="../data/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>gluon.data</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>