Adam.sc

README.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Adam\n",
    "\n",
    "This plugin provides the Adam optimizer for SGD. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Dependencies for use in Ammonite REPL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import $ivy.`com.thoughtworks.deeplearning::plugins-builtins:2.0.0`\n",
    "\n",
    "import scala.concurrent.ExecutionContext.Implicits.global\n",
    "import org.nd4j.linalg.factory.Nd4j\n",
    "import org.nd4j.linalg.api.ndarray.INDArray\n",
    "import com.thoughtworks.feature.Factory\n",
    "import com.thoughtworks.deeplearning.plugins.INDArrayWeights\n",
    "import com.thoughtworks.deeplearning.plugins.Builtins"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "interp.load(scala.io.Source.fromURL(new java.net.URL(\"https://gist.github.com/Rabenda/0c2fc6ba4cfa536e4788112a94200b50/raw/233cbc83932dad659519c80717d145a3983f57e1/Adam.sc\")).mkString)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "interp.load(\"\"\"\n",
    "    val hyperparameters = Factory[Builtins with Adam].newInstance()\n",
    "\"\"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### For sbt projects"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "// build.sbt\n",
    "libraryDependencies += \"com.thoughtworks.deeplearning\" %% \"plugins-builtins\" % \"latest.release\"\n",
    "\n",
    "addCompilerPlugin(\"com.thoughtworks.import\" %% \"import\" % \"latest.release\")\n",
    "\n",
    "// XXX.scala\n",
    "import $exec.`https://gist.github.com/Rabenda/0c2fc6ba4cfa536e4788112a94200b50/raw/233cbc83932dad659519c80717d145a3983f57e1/Adam.sc`\n",
    "\n",
    "import scala.concurrent.ExecutionContext.Implicits.global\n",
    "import org.nd4j.linalg.factory.Nd4j\n",
    "import org.nd4j.linalg.api.ndarray.INDArray\n",
    "import org.nd4j.linalg.ops.transforms.Transforms\n",
    "import com.thoughtworks.feature.Factory\n",
    "import com.thoughtworks.deeplearning.plugins.INDArrayWeights\n",
    "import com.thoughtworks.deeplearning.plugins.Builtins\n",
    "             \n",
    "val hyperparameters = Factory[Builtins with Adam].newInstance()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Scala",
   "language": "scala",
   "name": "scala"
  },
  "language_info": {
   "codemirror_mode": "text/x-scala",
   "file_extension": ".scala",
   "mimetype": "text/x-scala",
   "name": "scala211",
   "nbconvert_exporter": "script",
   "pygments_lexer": "scala",
   "version": "2.11.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}

Adam.sc

trait Adam extends INDArrayWeights {
  import org.nd4j.linalg.ops.transforms.Transforms

  trait INDArrayWeightApi extends super.INDArrayWeightApi {
    this: INDArrayWeight =>
    private var m0: Option[INDArray] = None
    def m = m0.getOrElse(Nd4j.zeros(data.shape: _*))
    def m_=(value: INDArray) = m0 = Some(value)
    
    private var v0: Option[INDArray] = None
    def v = v0.getOrElse(Nd4j.zeros(data.shape: _*))
    def v_=(value: INDArray) = v0 = Some(value)
  }

  override type INDArrayWeight <: INDArrayWeightApi with Weight

  trait INDArrayOptimizerApi extends super.INDArrayOptimizerApi {
    this: INDArrayOptimizer =>
    def beta1: scala.Double = 0.9

    def beta2: scala.Double = 0.999

    def eps: scala.Double = 1e-8

    abstract override def delta = delta0
    
    private lazy val delta0: INDArray = {
      import org.nd4s.Implicits._
      import weight._
      m = m * beta1 + super.delta * (1.0 - beta1)
      v = v * beta2 + (super.delta * super.delta) * (1.0 - beta2)
      m / (Transforms.sqrt(v) + eps)
    }
  }

  override type INDArrayOptimizer <: INDArrayOptimizerApi with Optimizer
}

LICENSE

MIT License

Copyright (c) 2017 ThoughtWorks Inc.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.