deep_learning_tutorial.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Deep learning, c'est quoi?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Pour faire fonctionner ce notebook sur sa machine :\n",
    "* Télécharger et installer anacodna https://www.anaconda.com/download/\n",
    "* Aller dans le repertoire anaconda3/bin\n",
    "* créer un environnement tensorflow `conda create -n tensorflow python=3.5` cf https://www.tensorflow.org/versions/r0.12/get_started/os_setup#anaconda_installation \n",
    "* activer l'environnement : `source activate tensorflow`, installer tensorflow `conda install -c conda-forge tensorflow`\n",
    "* installer Keras `pip install keras` et matplotlib `pip install matplotlib`\n",
    "* Dans l'interface d'anaconda <img src=\"anaconda.png\"> installer et lancer jupyter notebook\n",
    "* Télécharger et ouvrir ce notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Apprentissage :  60000\n",
      "Test :  10000\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers.core import Dense, Activation\n",
    "from keras.optimizers import SGD\n",
    "from keras.utils import np_utils\n",
    "\n",
    "nb_classes = 10 #On doit reconnaitre 10 chiffres différents\n",
    "\n",
    "(training_images, training_labels), (test_images, test_labels) = mnist.load_data()\n",
    "print ('Apprentissage : ', training_images.shape[0])\n",
    "print ('Test : ', test_images.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11cf4e4a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(6):\n",
    "    plt.subplot(2,3,i+1)\n",
    "    plt.imshow(training_images[i], cmap='gray', interpolation='none')\n",
    "    plt.title(\"Chiffre : {}\".format(training_labels[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "training_images = training_images.reshape(60000, 784)\n",
    "test_images = test_images.reshape(10000, 784)\n",
    "training_images = training_images.astype('float32')\n",
    "test_images = test_images.astype('float32')\n",
    "training_images /= 255\n",
    "test_images /= 255\n",
    "\n",
    "training_labels_categories = np_utils.to_categorical(training_labels, nb_classes)\n",
    "test_labels_categories = np_utils.to_categorical(test_labels, nb_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/1\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 0.4688 - acc: 0.8612 - val_loss: 0.2356 - val_acc: 0.9312\n",
      "Score sur le dataset de test: 93.12 %\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(500, input_shape=(784,)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(Dense(10))\n",
    "model.add(Activation('softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer=SGD(0.5), metrics=['accuracy'])\n",
    "\n",
    "history = model.fit(training_images, training_labels_categories, batch_size=500, epochs=1,\n",
    "                    verbose=1, validation_data=(test_images, test_labels_categories))\n",
    "score = model.evaluate(test_images, test_labels_categories, verbose=0)\n",
    "print('Score sur le dataset de test:', score[1]*100, \"%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134c392e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAACRCAYAAADTnUPWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAFkZJREFUeJzt3Xu4VVW5x/HvKyJeQEUUQyAwQEs9B0HzUbMOKl4yRczLY3o8ZN6OYYKYhqapeMljKj5RRwMvWEmllUceTAMx8GSIiYKJKBdTwDYgoaCCHtBx/piTyZiTvdaee+91Hfv3eZ71MMYac675rvUuxh5rzJs55xARkfq3TbUDEBGR0lCHLiISCHXoIiKBUIcuIhIIdegiIoFQhy4iEoi67tDNrLeZOTPbNq4/YWbDqh0XbB1bI+1Xm9m9Xv0UM1tmZh+Y2QAz29fM5prZ+2Z2aeUirz7lNUw1ntdBZra8SPs9ZnatV7/YzFbGee1iZl8ys0VxfWhlom6Ec66sD+BNYAPwAbASmAh0LNFr9wYcsG0jbd8E/pzjNQYDLwIfAsuBM5qx/X2AR4DVwFrgZWAU0K5YbAVeawlwsle/DxhbwjwMAj6N87D5MUx5re+8ZrZ1fxxb31DzGi+7G/BO3uW99Q4B/gC8B6wBngfOjdsGActzvk77+DPq7z03HRhRwlx2AMYC/wDeBf4baN/UepUaoZ/knOsIDAQOBq7JLmCRiv5iMLP9gEnA94FdgP7AnJzr9gFmA8uAf3HO7QKcTvT+OrUgnF7A/CL17PbbtWAb/3DOdfQeD7bgNXzKa9MqkVfM7AigT0vWbURN5tXzX8CC5qxgZocBTwMzgb5AF+Bi4Kst2P6ewPbkzGsLP6vRRJ/9AUQDjIE0koetlGOk0Mhf/MFe/UfAlLg8A7gZeJboL15fov+A9wENwNvATUC7ePl2wO1EI6c3gOF4f/Hj1zsf+ALwEfAJ0UjjvQKxTQJubOH7+iXweI7RyDBgaRzz97326+PX6BDH6IhGk0uIvnifxO/hgzihE4G7iUYYHxKNQDvEn8dSotHUPcAOBeIZRM4RiPJaP3mNt7kt8BLwr5RmhF6TeY3XORyYBZxLM0bowJ+BnxZpH0T0S+5yYFX8fs712ifG722fOE8ujvXpOLefsuWXTYfmflaNxPMCcLpXPwtY1tT7rPTIqSdwAtGXb7NzgAuJRj9vEX1wm4g+gAHAsURJB7gAODF+/mDgtMa245xbAPwnMMtFI9FdC4R0aBzX38yswcx+aWa75Xw7g4Hf5ljuCGBf4GjgB2b2hUysH7toNATRT7g+zrmjgP8FLonjXxi3n0X0JelE9AW9legLdiDR59Ud+EGRWLrG835/N7OxZrZTrnfaBOW16nm9DHjGOfdynjeYV63lNf718hPgEqIONe/72BE4jKbz+hmiTrc7cB7wUzPrnIl1IbB/XN3VOXeUc64P0R/fk+L4P47bm/NZNRp6ptzDzHYp9gYq1aH/j5m9R/RlnQnc4rVNdM7Nd85tIpobOwEY6Zz70Dm3imge6cx42TOAu5xzy5xza4AftjKuHkQf+qlAP2AHYFzOdbsQ/aVtyg3OuQ3OuXnAPKKf/y31mHPuWefcp8DHRF+Wy5xza5xz7xN9rmcWWPc1og6iG3AUcBBwZytiAeW16nmNO92LKN7hN1et5vVSYLZzLtf0maczUV/XVF43AmOccxudc38gGm3v2/wwE835rLKeBEaY2R5m9hmi9w6wY7ENNrqnvgyGOueeKtC2zCv3Itrh0GCW/HHaxltmr8zyb7Uyrg3AA5tHSmZ2C1Aozqx/EnWOTVnhldcDHQstmIP/3vcgSu4c77Myop+5W3HOrfBi+buZXQlMIeoMWkp5jVQtr8BdRJ3Q2lZsP6vm8mpmexF1age1YPV3iaZEuhENbAr5Z9z5blbKvDb1WWXdDOwKzCX6Iz+BaFS/stgGK9WhF+P/dFpGFPzumQ92swagp1f/bM7XLeTlzHLNufTkU0QjwAeasU5r+fGtJuq49nfOvd3C1yrnLzTlNb/W5PVo4Agzu817bpaZjXDOTSplkLFq5fUQog751bhD3AHYwcxWAN2dc58UfGHn1pvZLKK8/qmJ7ZRScz6r9IrObSCaWroEwMwuBObEv+IKqqnj0J1zDcBU4A4z29nMtjGzPmb2b/EiDwOXmlmPeG5rdJGXW0k057RdkWUeAM41s8/F82yjiUatAJjZm2b2zQLrXgccbmY/in8SYWZ94/naQnO7JRMndgIw1sy6xtvvbmbHNba8mR1pZr3iPe49ieZpHyt3nHGsymtOzc0r0Vx7f6LptAPj504CHq1ArJXM6xNEO6Q3v88fEM3tH7i5M7foGPdBBda/EvimmV1hZl3i5fub2a/zvt/WyPFZpcQ53yv+/3oocC3Rd7OomurQY/8BbAe8SvRT6bds+Qk8Afgj0Zzli8Dvi7zO00SHEa0ws9WNLeCcux/4OdFham8R/QW9FCD+YnUBniuw7hKiHS29gflmthb4HdHe6fdzvdPW+x6wGHjOzNYRjS4LzfkNAP5CtIf+L8Df2DIvVwnKa3658+qcW+WcW7H5ET+9Oh7hVUJF8hrvZPbf51pg4+b3HA9S3if6Xm/FOfcXon1HRwFvmNkaYDzR0UWVUuyzyurDlv+vDwKjnXNTm9qAxYfESIZFx/UOd859o9qxSOkor2Eys38nmqa6qtqxVJM6dBGRQNTilIuIiLSAOnQRkUC0qkM3s+PN7HUzW2xmxfZgSx1RXsOl3AauqWsDFHoQneiwBPgc0Z7becB+Tazj9KiNh/Ia5qOU/2er/V70SD3eydMvt2aEfgiw2Dn3hnPu/4BfAye34vWkNiiv4VJu61eus2xb06F3J33a6vL4uRQzu9DMXjCzF1qxLakc5TVcTeZWea1vZT/13zk3nugAfszMlXt7UhnKa5iU1/rWmhH626Sv09Ajfk7qm/IaLuU2cK3p0P8K9DOzvePTqc8EJpcmLKki5TVcym3gWjzl4pzbZGaXEF2roR1wv3Ou4K21pD4or+FSbsNX0VP/NSdXO5xz1vRS+SivtUN5DdYc59zBTS2kM0VFRAKhDl1EJBDq0EVEAqEOXUQkEOrQRUQCoQ5dRCQQZT/1vy3r3Llzqv7Zzxa76fkWb72Vvg7PZZddlpRfeeWVVNvChQuT8rx585obYnB23TV9H+d169Yl5U8/LXrDdJG6pxG6iEgg1KGLiARCHbqISCA0h95KX/va11L1IUOGJOVBgwal2vr27ZvrNf15cYBevXol5Q4dOhRcr127drleP2SPPPJIqv7hhx8m5XvvvTfVNmXKlIrElEfXrl1T9TVr1iTlTZs2VTocqVMaoYuIBEIduohIIDTl0og+ffqk6sOHD0/KF1xwQapthx12SNXNWn+xu3322afVr9FWvfjii6n6FVdckZRnzpxZ6XByGzlyZKrevn37pOy/h7bqzjvvTNX9zyub8zfffDMpd++evnvis88+m6q/9NJLSXnGjBmptoaGhqRcL4e8aoQuIhIIdegiIoFQhy4iEgjNoTeiR48eqfqIESPKvs3XXnstKc+fr7uCtdSyZcuqHUJuxxxzTFIeNWpUqm277bZLym11Dn3w4MFJeejQoam2U089NSl/9NFHqTZ/2U6dOqXazj333IL17LLPPfdcUj777LNTbbX6PdMIXUQkEOrQRUQCEfSUy+67756q+1Mn2cOXnnzyyaT88ccfp9rWrl2blP0zDwF22mmnVH3q1KlJOXtlxNmzZydl/3ApgA0bNhTchuR38cUXVzuE3Pwzif0pFtj6ULy26Fvf+lZSfvvtt1Ntjz76aMH1nnjiiRZt7/LLL0/V33jjjaTsn7lbyzRCFxEJhDp0EZFAqEMXEQlEcHPo/py2P58N0L9//6R8yimnFHwN/3AlgIEDByZl/7Ri2PouRMuXL0/K9XK6cL074IADknL2VO9a5h+2mDVmzJgKRlKb/P+vzz//fNm3d8cdd5R9G+WmEbqISCCa7NDN7H4zW2Vmr3jP7WZm08xsUfxv52KvIbVHeQ2Xctt25ZlymQj8BPi599xoYLpz7lYzGx3Xv1f68JqWPdxr0qRJSdn/yQZwyy23JOWnnnoq9zay0yy+pUuX5n6dGjORGs5rcxx22GFJeeeddy64XLUPB83enMS/oqJ/2CrAM88805pNTaQOc5v9fPx69hBgaVyTI3Tn3DNA9iDMk4EH4/KDwFCkriiv4VJu266W7hTd0zm3+WLBK4A9Cy1oZhcCF7ZwO1JZymu4cuVWea1vrT7KxTnnzMwVaR8PjAcotpzUFuU1XMVyq7zWt5Z26CvNrJtzrsHMugGrShlUUzp27JiUr7rqqlTbiSeemJRXr16darv99tuT8vr168sUXV2ral7z8vMPW5+y7fNPER8/fnzZYsrj5JNPTtX9fTwTJkxItb333nul3nzN5zZ7yOlee+2VlHfZZZdKh1OXWnrY4mRgWFweBjxWmnCkypTXcCm3bUCewxZ/BcwC9jWz5WZ2HnArcIyZLQIGx3WpI8pruJTbtqvJKRfn3DcKNB1d4lhy8y9gP3r06FSbfxjhl7/85VSbf9XEtq4W85rX2LFjU/ViN9WupTMu/asHllO95ta/uiHA5MmTk/J3v/vdVNu4ceOS8jvvvNPibd52221Jedq0aam2bL0e6ExREZFAqEMXEQmEOnQRkUDU5dUWDz/88IJt/p2A/CsfSn0bMmRIUj799NMLLpe9TMPrr79erpBy8Q+369q1axUjqT+LFi1Kyttvv32qzf8+3HfffS3exkEHHZSUzSzVpjl0ERGpGnXoIiKBqMspl9NOO61g2/HHH5+Ur7vuulTbY49tOZdi7ty5pQ9MSiZ71cRrr722YJvv61//eqr+0UcflTawZtp7772T8oEHHlhwudZMG4Sq2I2ys1MweWWn6z7/+c8n5XvvvbdFr1lLNEIXEQmEOnQRkUCoQxcRCYQ5V7krZJbqcpx+zM25EbO/7D333JNq828Mnb3x8+LFi5Py/Pnzi25j//33T8qzZs1KtdXSYZTOOWt6qXzKcZnVvn37puoLFy4suKx/RcUzzjgj1fbJJ5+UNrBm8ufNi80J9+vXL1VfsmRJi7ZX63ltjm222TLenDlzZqqtd+/eSdmfB4fid6bKXnHz/PPPT8rZw6GzN4uvsjnOuYObWkgjdBGRQKhDFxEJhDp0EZFA1OVx6P6dh0aNGpV7PX9O7tvf/naqLVsvhexlPWfMmJGUzzzzzJJvr94dd9xxSfnmm28uuJy/TwPgkksuScrZOXP/dO6ddtopdywbN25M1du3b19wWX/Otjn7pPy5/+ylYyW9z+uRRx5Jtd11111J+corr0y1XX/99Uk5e7z6F7/4xVQ9e7p/vdMIXUQkEOrQRUQCUZdTLv5din7zm9+k2iZNmpSUt902/fZ69uyZlP3pl3LZY489UnX/kgXXXHNNqu2mm24qezy1zr+J8sCBAwsu16FDh1T96quvLris/x246KKLCi6X/emdPcRwwIABBdc9++yzk/KUKVNSbccee2zB9fwbQVfy8OF69OMf/zhVP/XUU5Oyf1kIgEMOOSQpZ6fKslfj9G/UHQKN0EVEAqEOXUQkEOrQRUQCUZdz6P6haS+88EKqrdgd4I8+estNz7Nza/6hTtlDm0rFn6f175QiEX9etBh/XwjA8OHDW73t7Bx2p06dUvVly5Yl5XXr1qXaHnrooaQ8Z86cVFt2P4rv7rvvbnacEhk6dGhSPu+881JtvXr1Sso33nhjqi277yo0GqGLiARCHbqISCDqcsqlpaZPn16wzb8qXnbKZdOmTUn5gQceSLVNmDAhVR85cmRSPuuss1oUZ1s1ZsyYpDxu3LiCyy1dujRVX7lyZVL+4IMPCq739NNPp+rZ6Trf888/n6rvuOOOSXn16tWptq985StJ+Tvf+U6qzZ9amzdvXqrNvwmyNM+7776blP0zx5uSzUFoNEIXEQlEkx26mfU0sz+Z2atmNt/MRsTP72Zm08xsUfxv5/KHK6WivIZJeW3b8ozQNwGXO+f2Aw4FhpvZfsBoYLpzrh8wPa5L/VBew6S8tmFNzqE75xqAhrj8vpktALoDJwOD4sUeBGYA3ytLlBUwderUpJy90p9/+vgFF1yQasveWWfQoEG5tlftuxfVYl5/9rOfJeXZs2cXXG7FihWp+po1a5Ly+vXrSx8Y6TnbrGnTpiXlIUOGFFwuewertWvXtj6wjFrMq1ROs3aKmllvYAAwG9gz/vIArAD2LLDOhcCFLQ9Ryk15DZPy2vbk3ilqZh2B3wEjnXOpMytcdFZGo1cXcs6Nd84dnOd+eFJ5ymuYlNe2KdcI3czaE305HnLO/T5+eqWZdXPONZhZN2BVuYKshAULFiTlhx9+ONWWvfGw78gjjyzYlr3ZwuOPP56U/StGVkut5dU/PLTYIYW1zD/0EtI3YmhoaMguXha1ltdaMnjw4GqHUFZ5jnIx4D5ggXPuTq9pMjAsLg8DHit9eFIuymuYlNe2Lc8I/UvAOcDfzGxu/NzVwK3Aw2Z2HvAWUHgYK7VIeQ2T8tqG5TnK5c9AoRvvHV3gealxymuYlNe2rU2d+l/Mhg0bkrJ/+j5Ax44dk/LBB6f3FXXt2jVV9++I8otf/CLV5l/RUcKUvTF4ti7Vlb1j0apVW3YlzJ07l3qnU/9FRAKhDl1EJBCacmmEf/U+gJNOOikpn3POOam2Qw89NFW/4YYbkrL/c05Eqm/x4sWpepcuXZLyAQcckGqrx0NnNUIXEQmEOnQRkUCoQxcRCYTm0Jspeyhiti4i9WPjxo1JudjdruqFRugiIoFQhy4iEghNuYhIm9G9e/dU3b9xyWuvvVbpcEpOI3QRkUCoQxcRCYQ6dBGRQFh0N6oKbcyschuTopxzhS6x2mzKa+1QXoM1J89tATVCFxEJhDp0EZFAqEMXEQmEOnQRkUCoQxcRCYQ6dBGRQFT61P/VwFvA7nG5FrTFWHqV+PWU1+KU19Jpq7Hkym1Fj0NPNmr2Qp5jKitBsZROLcWvWEqnluJXLMVpykVEJBDq0EVEAlGtDn18lbbbGMVSOrUUv2IpnVqKX7EUUZU5dBERKT1NuYiIBEIduohIICraoZvZ8Wb2upktNrPRldx2vP37zWyVmb3iPbebmU0zs0Xxv50rEEdPM/uTmb1qZvPNbES1YikF5TUVSzC5VV5TsdRFXivWoZtZO+CnwFeB/YBvmNl+ldp+bCJwfOa50cB051w/YHpcL7dNwOXOuf2AQ4Hh8WdRjVhaRXndShC5VV63Uh95dc5V5AEcBvzRq18FXFWp7Xvb7Q284tVfB7rF5W7A61WI6THgmFqIRXlVbpXX+s1rJadcugPLvPry+Llq29M51xCXVwB7VnLjZtYbGADMrnYsLaS8FlDnuVVeC6jlvGqnqMdFf2YrdhynmXUEfgeMdM6tq2YsIavGZ6nclp/yurVKduhvAz29eo/4uWpbaWbdAOJ/V1Vio2bWnuiL8ZBz7vfVjKWVlNeMQHKrvGbUQ14r2aH/FehnZnub2XbAmcDkCm6/kMnAsLg8jGhurKzMzID7gAXOuTurGUsJKK+egHKrvHrqJq8V3pFwArAQWAJ8vwo7Mn4FNAAbieYEzwO6EO2dXgQ8BexWgTiOIPpp9jIwN36cUI1YlFflVnkNJ6869V9EJBDaKSoiEgh16CIigVCHLiISCHXoIiKBUIcuIhIIdegiIoFQhy4iEoj/B+8xPkllod7cAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x134c39978>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "classes_predites = model.predict_classes(test_images)\n",
    "correct_indices = np.nonzero(classes_predites == test_labels)[0]\n",
    "incorrect_indices = np.nonzero(classes_predites != test_labels)[0]\n",
    "\n",
    "plt.figure()\n",
    "for i, correct in enumerate(correct_indices[:3]):\n",
    "    plt.subplot(1,3,i+1)\n",
    "    plt.imshow(test_images[correct].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predit {}, Chiffre {}\".format(classes_predites[correct], test_labels[correct]))\n",
    "\n",
    "plt.figure()\n",
    "for i, incorrect in enumerate(incorrect_indices[:3]):\n",
    "    plt.subplot(1,3,i+1)\n",
    "    plt.imshow(test_images[incorrect].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predit {}, Chiffre {}\".format(classes_predites[incorrect], test_labels[incorrect]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}