samples/jupyter-notebooks/Autoencoder.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Autoencoder\n",
    "This notebook demonstrates the invocation of the SystemML autoencoder script, and alternative ways of passing in/out data."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### This notebook is supported with SystemML 0.14.0 and above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "!pip show systemml"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from systemml import MLContext, dml\n",
    "ml = MLContext(sc)\n",
    "print(ml.info())\n",
    "sc.version"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## SystemML Read/Write data from local file system"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "FsPath = \"/tmp/data/\"\n",
    "inp  = FsPath + \"Input/\"\n",
    "outp = FsPath + \"Output/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Generate Data and write out to file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "X_pd = pd.DataFrame(np.arange(1,2001, dtype=np.float)).values.reshape(100,20)\n",
    "# X_pd = pd.DataFrame(range(1, 2001,1),dtype=float).values.reshape(100,20)\n",
    "script =\"\"\"\n",
    "    write(X, $Xfile)\n",
    "\"\"\"\n",
    "prog = dml(script).input(X=X_pd).input(**{\"$Xfile\":inp+\"X.csv\"})\n",
    "ml.execute(prog)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "!ls -l /tmp/data/Input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "autoencoderURL = \"https://raw.githubusercontent.com/apache/systemml/master/scripts/staging/autoencoder-2layer.dml\"\n",
    "rets = (\"iter\", \"num_iters_per_epoch\", \"beg\", \"end\", \"o\")\n",
    "\n",
    "prog = dml(autoencoderURL).input(**{\"$X\":inp+\"X.csv\"}) \\\n",
    "                          .input(**{\"$H1\":500, \"$H2\":2, \"$BATCH\":36, \"$EPOCH\":5 \\\n",
    "                                    , \"$W1_out\":outp+\"W1_out\", \"$b1_out\":outp+\"b1_out\" \\\n",
    "                                    , \"$W2_out\":outp+\"W2_out\", \"$b2_out\":outp+\"b2_out\" \\\n",
    "                                    , \"$W3_out\":outp+\"W3_out\", \"$b3_out\":outp+\"b3_out\" \\\n",
    "                                    , \"$W4_out\":outp+\"W4_out\", \"$b4_out\":outp+\"b4_out\" \\\n",
    "                                   }).output(*rets)\n",
    "iter, num_iters_per_epoch, beg, end, o = ml.execute(prog).get(*rets)\n",
    "print (iter, num_iters_per_epoch, beg, end, o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "!ls -l /tmp/data/Output"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Alternatively to passing in/out file names, use Python variables."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "autoencoderURL = \"https://raw.githubusercontent.com/apache/systemml/master/scripts/staging/autoencoder-2layer.dml\"\n",
    "rets = (\"iter\", \"num_iters_per_epoch\", \"beg\", \"end\", \"o\")\n",
    "rets2 = (\"W1\", \"b1\", \"W2\", \"b2\", \"W3\", \"b3\", \"W4\", \"b4\")\n",
    "\n",
    "prog = dml(autoencoderURL).input(X=X_pd) \\\n",
    "                          .input(**{ \"$H1\":500, \"$H2\":2, \"$BATCH\":36, \"$EPOCH\":5}) \\\n",
    "                          .output(*rets) \\\n",
    "                          .output(*rets2)\n",
    "result = ml.execute(prog)\n",
    "iter, num_iters_per_epoch, beg, end, o = result.get(*rets)\n",
    "W1, b1, W2, b2, W3, b3, W4, b4 = result.get(*rets2)\n",
    "\n",
    "print (iter, num_iters_per_epoch, beg, end, o)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}