{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "!THEANO_FLAGS=device=gpu python DeepLearningTutorials/code/logistic_sgd.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "!THEANO_FLAGS=device=gpu python DeepLearningTutorials/code/convolutional_mlp.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "!THEANO_FLAGS=device=cpu python DeepLearningTutorials/code/logistic_sgd.py"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "%%time\n",
    "!THEANO_FLAGS=device=cpu timeout 1200 python DeepLearningTutorials/code/convolutional_mlp.py"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import theano.tensor as T\n",
    "from theano.tensor.shared_randomstreams import RandomStreams\n",
    "from theano import function, shared\n",
    "from theano.compile.sharedvalue import SharedVariable\n",
    "import numpy as np\n",
    "\n",
    "srng = RandomStreams(seed=1234)\n",
    "\n",
    "x = T.fcol()\n",
    "a = srng.uniform((10, 1))\n",
    "a_shared = SharedVariable(\n",
    "    value=np.zeros((10, 1)),\n",
    "    type=a.type,\n",
    "    name='a',\n",
    "    strict=True\n",
    ")\n",
    "\n",
    "b = srng.uniform((10, 1))\n",
    "b_shared = SharedVariable(\n",
    "    value=np.zeros((10, 1)),\n",
    "    type=b.type,\n",
    "    name='b',\n",
    "    strict=True\n",
    ")\n",
    "\n",
    "z = (x + a).T.dot(b)\n",
    "f = function([x], z, updates=[(a_shared, a), (b_shared, b)], allow_input_downcast=True)\n",
    "\n",
    "f(np.ones((10, 1)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "a_shared.get_value()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "b_shared.get_value()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "(np.ones((10, 1)) + a_shared.get_value()).T.dot(b_shared.get_value())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Problem 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def fib(n):\n",
    "    if n <= 1:\n",
    "        return 0\n",
    "    elif n == 2:\n",
    "        return 1\n",
    "    else:\n",
    "        a = shared(0)\n",
    "        b = shared(1)\n",
    "        f = function([], a + b, updates=[(b, a + b), (a, b)])\n",
    "        for i in range(1, n):\n",
    "            f()\n",
    "            \n",
    "    return b.get_value()\n",
    "        \n",
    "print('fib(10): {}'.format(fib(10)))\n",
    "print('fib(20): {}'.format(fib(20)))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}