Digit Reconnition TF.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 211,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import scipy.io as sio\n",
    "import tensorflow as tf\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = sio.loadmat('digits.mat')\n",
    "training_inputs = data['data']['training'][0][0][0]['inputs'][0]\n",
    "training_targets = data['data']['training'][0][0][0]['targets'][0]\n",
    "validation_inputs = data['data']['validation'][0][0][0]['inputs'][0]\n",
    "validation_targets = data['data']['validation'][0][0][0]['targets'][0]\n",
    "test_inputs = data['data']['test'][0][0][0]['inputs'][0]\n",
    "test_targets = data['data']['test'][0][0][0]['targets'][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "def theta_to_model(theta):\n",
    "    n_hid = int(len(theta) / (n_input+n_classes))\n",
    "    input_to_hid = theta[:n_input*n_hid].reshape(n_input,n_hid).T\n",
    "    hid_to_class = theta[n_input*n_hid:].reshape(n_hid, n_classes).T\n",
    "    return input_to_hid, hid_to_class\n",
    "    \n",
    "def initial_model(n_hid, random = False):\n",
    "    # n_hid: number of hidden logistic units\n",
    "    n_params = (n_input+n_classes) * n_hid\n",
    "    # No initialized ramdomly to get always the same result\n",
    "    if random:\n",
    "        as_row_vector = np.random.normal(0,0.01, n_params)\n",
    "    else:\n",
    "        as_row_vector = np.cos(range(0,(n_params)))*0.1\n",
    "    return theta_to_model(as_row_vector)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# Create 2 hidden layers model\n",
    "def multilayer_perceptron_2_layers(relu = True):\n",
    "    # tf Graph input\n",
    "    x = tf.placeholder(\"float\", [None, n_input])\n",
    "    y = tf.placeholder(\"int32\", [None, n_classes])\n",
    "    # Store layers weight & bias\n",
    "    mean = 0.0\n",
    "    stddev = 0.1\n",
    "    weights = {\n",
    "        'h1': tf.Variable(tf.random_normal([n_input, n_hidden_1], mean=mean,stddev=stddev)),\n",
    "        'h2': tf.Variable(tf.random_normal([n_hidden_1, n_hidden_2], mean=mean,stddev=stddev)),\n",
    "        'out': tf.Variable(tf.random_normal([n_hidden_2, n_classes], mean=mean,stddev=stddev)),\n",
    "    }\n",
    "    biases = {\n",
    "        'b1': tf.Variable(tf.random_normal([n_hidden_1], mean=mean,stddev=stddev)),\n",
    "        'b2': tf.Variable(tf.random_normal([n_hidden_2], mean=mean,stddev=stddev)),\n",
    "        'out': tf.Variable(tf.random_normal([n_classes], mean=mean,stddev=stddev)),\n",
    "    }\n",
    "    # Hidden layer with RELU activation\n",
    "    layer_1 = tf.add(tf.matmul(x, weights['h1']), biases['b1'])\n",
    "    if relu:\n",
    "        layer_1 = tf.nn.relu(layer_1)\n",
    "    else:\n",
    "        layer_1 = tf.nn.sigmoid(layer_1)\n",
    "    # Hidden layer with RELU activation\n",
    "    layer_2 = tf.add(tf.matmul(layer_1, weights['h2']), biases['b2'])\n",
    "    if relu:\n",
    "        layer_2 = tf.nn.relu(layer_2)\n",
    "    else:\n",
    "        layer_2 = tf.nn.sigmoid(layer_2)\n",
    "    # Output layer with linear activation\n",
    "    out_layer = tf.matmul(layer_2, weights['out']) + biases['out']\n",
    "    return out_layer,x, y, weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Create 1 hidden layer model\n",
    "# tf Graph input\n",
    "\n",
    "def multilayer_perceptron(random = False):\n",
    "    mean = 0.0\n",
    "    stddev = 0.1\n",
    "    # Hidden layer with sigmoid activation\n",
    "    biases = {\n",
    "        'b1': tf.cast(tf.Variable(tf.random_normal([n_hidden_1], mean=mean,stddev=stddev)), dtype=\"float64\"),\n",
    "        'out': tf.cast(tf.Variable(tf.random_normal([n_classes], mean=mean,stddev=stddev)), dtype=\"float64\"),\n",
    "    }\n",
    "    if random:\n",
    "        x = tf.placeholder(\"float\", [None, n_input])\n",
    "        y = tf.placeholder(\"int32\", [None, n_classes])\n",
    "        weights = {\n",
    "            'h1': tf.Variable(tf.random_normal([n_input, n_hidden_1])),\n",
    "            'out': tf.Variable(tf.random_normal([n_hidden_1, n_classes]))\n",
    "        }\n",
    "    else:\n",
    "        x = tf.placeholder(\"float64\", [None, n_input])\n",
    "        y = tf.placeholder(\"int32\", [None, n_classes])\n",
    "        w_hi_initial,w_out_initial = initial_model(n_hidden_1)\n",
    "        weights = {\n",
    "            'h1': tf.Variable(w_hi_initial.T),\n",
    "            'out': tf.Variable(w_out_initial.T)\n",
    "        }\n",
    "    layer_1 = tf.add(tf.matmul(x, weights['h1']), biases['b1'])\n",
    "    #layer_1 = tf.matmul(x, weights['h1'])\n",
    "    layer_1 = tf.nn.sigmoid(layer_1)\n",
    "    # Output layer with linear activation\n",
    "    #out_layer = tf.matmul(layer_1, weights['out'])\n",
    "    out_layer = tf.matmul(layer_1, weights['out']) + biases['out']\n",
    "    return out_layer, x, y, weights\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "def run_model(pred, x, y, weights, beta = 0):\n",
    "    # Define loss and optimizer\n",
    "    loss = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=pred, labels=y))\n",
    "    \n",
    "    regularizer_out = tf.nn.l2_loss(weights['out'])\n",
    "    regularizer_h1 = tf.nn.l2_loss(weights['h1'])\n",
    "    \n",
    "    regularizers = regularizer_out + regularizer_h1\n",
    "    \n",
    "    if 'h2' in weights:\n",
    "        regularizers = regularizers + tf.nn.l2_loss(weights['h2'])\n",
    "    \n",
    "    cost = loss + beta*regularizers\n",
    "    optimizer = tf.train.MomentumOptimizer(learning_rate=learning_rate, momentum = momentum).minimize(cost)\n",
    "    #optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate).minimize(cost)\n",
    "\n",
    "    # Initializing the variables\n",
    "    init = tf.global_variables_initializer()\n",
    "\n",
    "    # Launch the graph\n",
    "    n_training_cases = training_inputs.shape[1]\n",
    "    training_data_losses = []\n",
    "    validation_data_losses = []\n",
    "    with tf.Session() as sess:\n",
    "        sess.run(init)\n",
    "\n",
    "        # Training cycle\n",
    "        for optimization_iteration_i in range(iterations):\n",
    "            training_batch_start = (optimization_iteration_i*mini_batch_size)%n_training_cases\n",
    "            training_batch_inputs = training_inputs[:,training_batch_start:(training_batch_start+mini_batch_size)]\n",
    "            training_batch_targets = training_targets[:,training_batch_start:(training_batch_start+mini_batch_size)]\n",
    "\n",
    "            _, c = sess.run([optimizer, cost], feed_dict={x: training_batch_inputs.T,y: training_batch_targets.T})\n",
    "            if optimization_iteration_i % display_step == 0:\n",
    "                _, c = sess.run([optimizer, cost], feed_dict={x: training_inputs.T,y: training_targets.T})\n",
    "                c_cv = sess.run(cost, feed_dict={x: validation_inputs.T,y: validation_targets.T})\n",
    "                training_data_losses.append(c)\n",
    "                validation_data_losses.append(c_cv)\n",
    "        print(\"Optimization Finished!\")\n",
    "\n",
    "        # Test model\n",
    "        correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))\n",
    "        # Calculate accuracy\n",
    "        accuracy = tf.reduce_mean(tf.cast(correct_prediction, \"float\"))\n",
    "        print(\"Accuracy:\", accuracy.eval({x: training_inputs.T, y: training_targets.T}))\n",
    "        print(\"Accuracy:\", accuracy.eval({x: validation_inputs.T, y: validation_targets.T}))\n",
    "        print(\"Accuracy:\", accuracy.eval({x: test_inputs.T, y: test_targets.T}))\n",
    "        print(optimization_iteration_i, training_data_losses[-1], validation_data_losses[-1])\n",
    "    plt.figure(figsize = (20,10))\n",
    "    plt.plot(training_data_losses, label = 'Training')\n",
    "    plt.plot(validation_data_losses, label = 'Cross Validation')\n",
    "    plt.legend(loc='best')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Equivalente a la de Hinton pero con tensorflow"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization Finished!\n",
      "Accuracy: 1.0\n",
      "Accuracy: 0.919\n",
      "Accuracy: 0.920556\n",
      "999 0.00321078741991 0.319957267094\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIQAAAJCCAYAAACxsxylAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ntjf4CWzyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x144febe10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Parameters\n",
    "learning_rate = 0.35\n",
    "iterations = 1000\n",
    "mini_batch_size = 100\n",
    "display_step = 10\n",
    "momentum = 0.9\n",
    "\n",
    "# Network Parameters\n",
    "n_hidden_1 = 37 # 1st layer number of features\n",
    "n_input = 256 # MNIST data input (img shape: 28*28)\n",
    "n_classes = 10 # MNIST total classes (0-9 digits)\n",
    "run_model(*multilayer_perceptron(random = False))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# La de Hinton con regularizacion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization Finished!\n",
      "Accuracy: 1.0\n",
      "Accuracy: 0.923\n",
      "Accuracy: 0.924889\n",
      "999 0.248217297442 0.435532659804\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIQAAAJCCAYAAACxsxylAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nK8/X3CnzAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x143436f98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "run_model(*multilayer_perceptron(random = False), beta = 0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Se agrega un \"hidden layer\" con relu's"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization Finished!\n",
      "Accuracy: 1.0\n",
      "Accuracy: 0.941\n",
      "Accuracy: 0.934222\n",
      "49999 0.0683524 0.288334\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHsAAAJCCAYAAABQ9MuxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1448849b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Parameters\n",
    "learning_rate = 0.005\n",
    "iterations = 50000\n",
    "mini_batch_size = 200\n",
    "display_step = 100\n",
    "momentum = 0.9\n",
    "\n",
    "# Network Parameters\n",
    "n_hidden_1 = 500 \n",
    "n_hidden_2 = 300 \n",
    "n_input = 256 \n",
    "n_classes = 10 \n",
    "run_model(*multilayer_perceptron_2_layers(), beta = 0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Igual a anterior pero con sigmoidea"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Optimization Finished!\n",
      "Accuracy: 0.999\n",
      "Accuracy: 0.916\n",
      "Accuracy: 0.912333\n",
      "49999 0.300516 0.529006\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABIQAAAJCCAYAAACxsxylAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x149c46940>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Parameters\n",
    "learning_rate = 0.005\n",
    "iterations = 50000\n",
    "mini_batch_size = 200\n",
    "display_step = 100\n",
    "momentum = 0.9\n",
    "\n",
    "# Network Parameters\n",
    "n_hidden_1 = 500 \n",
    "n_hidden_2 = 300 \n",
    "n_input = 256 \n",
    "n_classes = 10 \n",
    "run_model(*multilayer_perceptron_2_layers(relu = False), beta = 0.001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Parameters\n",
    "learning_rate = 0.01\n",
    "iterations = 50000\n",
    "mini_batch_size = 200\n",
    "display_step = 500\n",
    "momentum = 0.9\n",
    "\n",
    "# Network Parameters\n",
    "n_hidden_1 = 256 \n",
    "n_hidden_2 = 128 \n",
    "n_input = 256 \n",
    "n_classes = 10 \n",
    "run_model(*multilayer_perceptron_2_layers(), beta = 0.001)\n",
    "\n",
    "Optimization Finished!\n",
    "Accuracy: 1.0\n",
    "Accuracy: 0.933\n",
    "Accuracy: 0.933444\n",
    "49999 0.0567606 0.285725"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda env:aind]",
   "language": "python",
   "name": "conda-env-aind-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}