mirror of
https://github.com/tensorflow/haskell.git
synced 2024-07-01 17:08:34 +02:00
5 lines
10 KiB
HTML
5 lines
10 KiB
HTML
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>TensorFlow.Minimize</title><link href="ocean.css" rel="stylesheet" type="text/css" title="Ocean" /><script src="haddock-util.js" type="text/javascript"></script><script src="https://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML" type="text/javascript"></script><script type="text/javascript">//<![CDATA[
|
|
window.onload = function () {pageLoad();setSynopsis("mini_TensorFlow-Minimize.html");};
|
|
//]]>
|
|
</script></head><body><div id="package-header"><ul class="links" id="page-menu"><li><a href="src/TensorFlow.Minimize.html">Source</a></li><li><a href="index.html">Contents</a></li><li><a href="doc-index.html">Index</a></li></ul><p class="caption">tensorflow-ops-0.1.0.0: Friendly layer around TensorFlow bindings.</p></div><div id="content"><div id="module-header"><table class="info"><tr><th>Safe Haskell</th><td>None</td></tr><tr><th>Language</th><td>Haskell2010</td></tr></table><p class="caption">TensorFlow.Minimize</p></div><div id="synopsis"><p id="control.syn" class="caption expander" onclick="toggleSection('syn')">Synopsis</p><ul id="section.syn" class="hide" onclick="toggleSection('syn')"><li class="src short"><span class="keyword">type</span> <a href="#t:Minimizer">Minimizer</a> a = <span class="keyword">forall</span> m. <a href="../tensorflow-0.1.0.2/TensorFlow-Build.html#t:MonadBuild">MonadBuild</a> m => [<a href="TensorFlow-Variable.html#t:Variable">Variable</a> a] -> [<a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Tensor">Tensor</a> <a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Value">Value</a> a] -> m <a href="../tensorflow-0.1.0.2/TensorFlow-Output.html#t:ControlNode">ControlNode</a></li><li class="src short"><a href="#v:minimizeWith">minimizeWith</a> :: (<a href="../tensorflow-0.1.0.2/TensorFlow-Build.html#t:MonadBuild">MonadBuild</a> m, <a href="TensorFlow-Gradient.html#t:GradientCompatible">GradientCompatible</a> a) => <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> a -> <a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Tensor">Tensor</a> v a -> [<a href="TensorFlow-Variable.html#t:Variable">Variable</a> a] -> m <a href="../tensorflow-0.1.0.2/TensorFlow-Output.html#t:ControlNode">ControlNode</a></li><li class="src short"><a href="#v:gradientDescent">gradientDescent</a> :: <a href="TensorFlow-Gradient.html#t:GradientCompatible">GradientCompatible</a> a => a -> <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> a</li><li class="src short"><span class="keyword">data</span> <a href="#t:AdamConfig">AdamConfig</a> = <a href="#v:AdamConfig">AdamConfig</a> {<ul class="subs"><li><a href="#v:adamLearningRate">adamLearningRate</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li><li><a href="#v:adamBeta1">adamBeta1</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li><li><a href="#v:adamBeta2">adamBeta2</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li><li><a href="#v:adamEpsilon">adamEpsilon</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li></ul>}</li><li class="src short"><a href="#v:adam">adam</a> :: <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li><li class="src short"><a href="#v:adam-39-">adam'</a> :: <a href="TensorFlow-Minimize.html#t:AdamConfig">AdamConfig</a> -> <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></li></ul></div><div id="interface"><h1>Documentation</h1><div class="top"><p class="src"><span class="keyword">type</span> <a id="t:Minimizer" class="def">Minimizer</a> a = <span class="keyword">forall</span> m. <a href="../tensorflow-0.1.0.2/TensorFlow-Build.html#t:MonadBuild">MonadBuild</a> m => [<a href="TensorFlow-Variable.html#t:Variable">Variable</a> a] -> [<a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Tensor">Tensor</a> <a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Value">Value</a> a] -> m <a href="../tensorflow-0.1.0.2/TensorFlow-Output.html#t:ControlNode">ControlNode</a> <a href="src/TensorFlow.Minimize.html#Minimizer" class="link">Source</a> <a href="#t:Minimizer" class="selflink">#</a></p><div class="doc"><p>Functions that minimize a loss w.r.t. a set of <code><a href="TensorFlow-Variable.html#t:Variable">Variable</a></code>s.</p><p>Generally only performs one step of an iterative algorithm.</p><p><code><a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a></code>s are defined as a function of the gradients instead of
|
|
the loss so that users can apply transformations to the gradients.</p></div></div><div class="top"><p class="src"><a id="v:minimizeWith" class="def">minimizeWith</a> <a href="src/TensorFlow.Minimize.html#minimizeWith" class="link">Source</a> <a href="#v:minimizeWith" class="selflink">#</a></p><div class="subs arguments"><p class="caption">Arguments</p><table><tr><td class="src">:: (<a href="../tensorflow-0.1.0.2/TensorFlow-Build.html#t:MonadBuild">MonadBuild</a> m, <a href="TensorFlow-Gradient.html#t:GradientCompatible">GradientCompatible</a> a)</td><td class="doc empty"> </td></tr><tr><td class="src">=> <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> a</td><td class="doc empty"> </td></tr><tr><td class="src">-> <a href="../tensorflow-0.1.0.2/TensorFlow-Tensor.html#t:Tensor">Tensor</a> v a</td><td class="doc"><p>Loss.</p></td></tr><tr><td class="src">-> [<a href="TensorFlow-Variable.html#t:Variable">Variable</a> a]</td><td class="doc"><p>Parameters of the loss function.</p></td></tr><tr><td class="src">-> m <a href="../tensorflow-0.1.0.2/TensorFlow-Output.html#t:ControlNode">ControlNode</a></td><td class="doc empty"> </td></tr></table></div><div class="doc"><p>Convenience wrapper around <code><a href="TensorFlow-Gradient.html#v:gradients">gradients</a></code> and a <code><a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a></code>.</p></div></div><div class="top"><p class="src"><a id="v:gradientDescent" class="def">gradientDescent</a> <a href="src/TensorFlow.Minimize.html#gradientDescent" class="link">Source</a> <a href="#v:gradientDescent" class="selflink">#</a></p><div class="subs arguments"><p class="caption">Arguments</p><table><tr><td class="src">:: <a href="TensorFlow-Gradient.html#t:GradientCompatible">GradientCompatible</a> a</td><td class="doc empty"> </td></tr><tr><td class="src">=> a</td><td class="doc"><p>Learning rate.</p></td></tr><tr><td class="src">-> <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> a</td><td class="doc empty"> </td></tr></table></div><div class="doc"><p>Perform one step of the gradient descent algorithm.</p></div></div><div class="top"><p class="src"><span class="keyword">data</span> <a id="t:AdamConfig" class="def">AdamConfig</a> <a href="src/TensorFlow.Minimize.html#AdamConfig" class="link">Source</a> <a href="#t:AdamConfig" class="selflink">#</a></p><div class="subs constructors"><p class="caption">Constructors</p><table><tr><td class="src"><a id="v:AdamConfig" class="def">AdamConfig</a></td><td class="doc empty"> </td></tr><tr><td colspan="2"><div class="subs fields"><p class="caption">Fields</p><ul><li><dfn class="src"><a id="v:adamLearningRate" class="def">adamLearningRate</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></dfn><div class="doc empty"> </div></li><li><dfn class="src"><a id="v:adamBeta1" class="def">adamBeta1</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></dfn><div class="doc empty"> </div></li><li><dfn class="src"><a id="v:adamBeta2" class="def">adamBeta2</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></dfn><div class="doc empty"> </div></li><li><dfn class="src"><a id="v:adamEpsilon" class="def">adamEpsilon</a> :: <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a></dfn><div class="doc empty"> </div></li></ul></div></td></tr></table></div><div class="subs instances"><p id="control.i:AdamConfig" class="caption collapser" onclick="toggleSection('i:AdamConfig')">Instances</p><div id="section.i:AdamConfig" class="show"><table><tr><td class="src clearfix"><span class="inst-left"><span id="control.i:id:AdamConfig:Default:1" class="instance expander" onclick="toggleSection('i:id:AdamConfig:Default:1')"></span> Default <a href="TensorFlow-Minimize.html#t:AdamConfig">AdamConfig</a></span> <a href="src/TensorFlow.Minimize.html#line-77" class="link">Source</a> <a href="#t:AdamConfig" class="selflink">#</a></td><td class="doc empty"> </td></tr><tr><td colspan="2"><div id="section.i:id:AdamConfig:Default:1" class="inst-details hide"><div class="subs methods"><p class="caption">Methods</p><p class="src"><a href="#v:def">def</a> :: <a href="TensorFlow-Minimize.html#t:AdamConfig">AdamConfig</a></p></div></div></td></tr></table></div></div></div><div class="top"><p class="src"><a id="v:adam" class="def">adam</a> :: <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a> <a href="src/TensorFlow.Minimize.html#adam" class="link">Source</a> <a href="#v:adam" class="selflink">#</a></p><div class="doc"><p>Perform one step of the adam algorithm.</p><p>See <a href="https://arxiv.org/abs/1412.6980">https://arxiv.org/abs/1412.6980</a>.</p><p>NOTE: Currently requires all <code><a href="TensorFlow-Variable.html#t:Variable">Variable</a></code>s to have an <code><a href="TensorFlow-Variable.html#v:initializedValue">initializedValue</a></code>.</p></div></div><div class="top"><p class="src"><a id="v:adam-39-" class="def">adam'</a> :: <a href="TensorFlow-Minimize.html#t:AdamConfig">AdamConfig</a> -> <a href="TensorFlow-Minimize.html#t:Minimizer">Minimizer</a> <a href="../base-4.9.1.0/Prelude.html#t:Float">Float</a> <a href="src/TensorFlow.Minimize.html#adam%27" class="link">Source</a> <a href="#v:adam-39-" class="selflink">#</a></p></div></div></div><div id="footer"><p>Produced by <a href="http://www.haskell.org/haddock/">Haddock</a> version 2.17.3</p></div></body></html> |