dac18_keras_train_n_inference.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Reference: \n",
    "#   Cunxi Yu, Houping Xiao and Giovanni De Micheli. \n",
    "#   \"Developing Synthesis Flows without Human Knowledge.\" ACM/IEEE Design Automation Conference (DAC'18)\n",
    "#   Contact: Cunxi Yu.  Email: cunxi.yu@epfl.ch\n",
    "#\n",
    "#\n",
    "#Demo: Input design/datasets: 64-bit GF-Mont with delay labels.\n",
    "#        Number of classes: 7 (0-6)\n",
    "#Demo \n",
    "#    Training Results:\n",
    "#        1) Angel flows classifcation (class-0): 93%\n",
    "#        2) Devil flows classifcation (class-0): 89%\n",
    "#        3) overall accuracy of classifer 74.2%         \n",
    "#\n",
    "#    Testing Results (20,000 * 3 test points): \n",
    "#        1) Angel flows classifcation (class-0): 93%,92%,93%\n",
    "#        2) Devil flows classifcation (class-0): 90%,90%,90%\n",
    "#        3) overall accuracy of classifer 73.6%, 73.4%, 73.6%\n",
    "#\n",
    "#     Saved model: dac18_GF_delay_model_save_test.h5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/Users/cunxiyu/miniconda2/lib/python2.7/site-packages/matplotlib/__init__.py:1405: UserWarning: \n",
      "This call to matplotlib.use() has no effect because the backend has already\n",
      "been chosen; matplotlib.use() must be called *before* pylab, matplotlib.pyplot,\n",
      "or matplotlib.backends is imported for the first time.\n",
      "\n",
      "  warnings.warn(_use_error_msg)\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout\n",
    "from keras.layers import Conv2D,Flatten,MaxPooling2D\n",
    "from keras.utils import to_categorical\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.utils import shuffle\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "matplotlib.use('Agg')\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import itertools\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_csv_data(csv_filename, header=None, num_funcs=6, shuffle_data=True):\n",
    "    \"\"\"helper function to read data from csv file\n",
    "    return X and Y\n",
    "    X.shape = (num_samples, length_seq, num_funcs)\n",
    "    Y.shape = (num_samples, num_classes)\n",
    "    \"\"\"\n",
    "    # read data\n",
    "    data = pd.read_csv(csv_filename, header=header)\n",
    "    data = data.as_matrix()\n",
    "    if (shuffle_data):\n",
    "        data = shuffle(data)\n",
    "\n",
    "    # extract X and Y\n",
    "    X = data[:, :-1]#.reshape((-1, int((data.shape[1]-1)/num_funcs), num_funcs))\n",
    "    X = data[:, :-1].reshape(X.shape[0], 1, 24, 6)\n",
    "    Y = data[:, -1]\n",
    "\n",
    "    # convert Y to one-hot vectors\n",
    "    Y = to_categorical(Y, num_classes=Y.max()+1)\n",
    "\n",
    "    return X, Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.callbacks import Callback\n",
    "from IPython.display import clear_output\n",
    "class PlotLearning(Callback):\n",
    "    def on_train_begin(self, logs={}):\n",
    "        self.i = 0\n",
    "        self.x = []\n",
    "        self.losses = []\n",
    "        self.val_losses = []\n",
    "        self.acc = []\n",
    "        self.val_acc = []\n",
    "        self.fig = plt.figure()\n",
    "        self.logs = []\n",
    "\n",
    "    def on_epoch_end(self, epoch, logs={}):\n",
    "        self.logs.append(logs)\n",
    "        self.x.append(self.i)\n",
    "        self.losses.append(logs.get('loss'))\n",
    "        self.val_losses.append(logs.get('val_loss'))\n",
    "        self.acc.append(logs.get('acc'))\n",
    "        self.val_acc.append(logs.get('val_acc'))\n",
    "        self.i += 1\n",
    "\n",
    "        #self.plot_history()\n",
    "\n",
    "    def plot_history(self):\n",
    "        #f, ax1 = plt.subplots()\n",
    "        f, (ax1, ax2) = plt.subplots(1, 2, sharex=True)\n",
    "        clear_output(wait=True)\n",
    "        ax1.set_yscale('log')\n",
    "        ax1.plot(self.x, self.losses, label=\"loss\")\n",
    "        ax1.plot(self.x, self.val_losses, label=\"val_loss\")\n",
    "        ax1.legend()\n",
    "        ax2.plot(self.x, self.acc, label=\"accuracy\")\n",
    "        ax2.plot(self.x, self.val_acc, label=\"validation accuracy\")\n",
    "        ax2.legend()\n",
    "        plt.show()\n",
    "\n",
    "plot_learning = PlotLearning()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "('X.shape = ', (20000, 1, 24, 6))\n",
      "('Y.shape = ', (20000, 7))\n"
     ]
    }
   ],
   "source": [
    "##training input data: test.csv\n",
    "folder_path = 'datasets/64bitGF/delay/'\n",
    "csv_filename = 'test.csv'\n",
    "X, Y = read_csv_data(folder_path + csv_filename, header=None, num_funcs=6)\n",
    "print(\"X.shape = \", X.shape)\n",
    "print(\"Y.shape = \", Y.shape)\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 1, 23, 32)         416       \n",
      "_________________________________________________________________\n",
      "conv2 (Conv2D)               (None, 1, 22, 64)         4160      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 1, 22, 64)         0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1, 22, 64)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1408)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 64)                90176     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 7)                 455       \n",
      "=================================================================\n",
      "Total params: 95,207\n",
      "Trainable params: 95,207\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 14000 samples, validate on 6000 samples\n",
      "Epoch 1/100\n",
      " - 2s - loss: 1.2511 - acc: 0.5182 - val_loss: 0.9710 - val_acc: 0.6165\n",
      "Epoch 2/100\n",
      " - 2s - loss: 1.0143 - acc: 0.5904 - val_loss: 0.9191 - val_acc: 0.6357\n",
      "Epoch 3/100\n",
      " - 2s - loss: 0.9730 - acc: 0.6100 - val_loss: 0.8879 - val_acc: 0.6420\n",
      "Epoch 4/100\n",
      " - 2s - loss: 0.9365 - acc: 0.6220 - val_loss: 0.8797 - val_acc: 0.6488\n",
      "Epoch 5/100\n",
      " - 2s - loss: 0.9176 - acc: 0.6267 - val_loss: 0.8593 - val_acc: 0.6487\n",
      "Epoch 6/100\n",
      " - 2s - loss: 0.9034 - acc: 0.6336 - val_loss: 0.8340 - val_acc: 0.6598\n",
      "Epoch 7/100\n",
      " - 2s - loss: 0.8873 - acc: 0.6404 - val_loss: 0.8554 - val_acc: 0.6527\n",
      "Epoch 8/100\n",
      " - 2s - loss: 0.8739 - acc: 0.6454 - val_loss: 0.8028 - val_acc: 0.6730\n",
      "Epoch 9/100\n",
      " - 2s - loss: 0.8598 - acc: 0.6499 - val_loss: 0.7972 - val_acc: 0.6758\n",
      "Epoch 10/100\n",
      " - 2s - loss: 0.8446 - acc: 0.6504 - val_loss: 0.7882 - val_acc: 0.6723\n",
      "Epoch 11/100\n",
      " - 2s - loss: 0.8330 - acc: 0.6546 - val_loss: 0.7639 - val_acc: 0.6827\n",
      "Epoch 12/100\n",
      " - 2s - loss: 0.8180 - acc: 0.6614 - val_loss: 0.7343 - val_acc: 0.6898\n",
      "Epoch 13/100\n",
      " - 2s - loss: 0.8010 - acc: 0.6689 - val_loss: 0.7343 - val_acc: 0.6948\n",
      "Epoch 14/100\n",
      " - 2s - loss: 0.8008 - acc: 0.6699 - val_loss: 0.7143 - val_acc: 0.6992\n",
      "Epoch 15/100\n",
      " - 2s - loss: 0.7879 - acc: 0.6752 - val_loss: 0.7040 - val_acc: 0.6997\n",
      "Epoch 16/100\n",
      " - 2s - loss: 0.7818 - acc: 0.6691 - val_loss: 0.6915 - val_acc: 0.7045\n",
      "Epoch 17/100\n",
      " - 2s - loss: 0.7667 - acc: 0.6836 - val_loss: 0.6919 - val_acc: 0.7063\n",
      "Epoch 18/100\n",
      " - 2s - loss: 0.7622 - acc: 0.6811 - val_loss: 0.6772 - val_acc: 0.7103\n",
      "Epoch 19/100\n",
      " - 2s - loss: 0.7506 - acc: 0.6824 - val_loss: 0.6670 - val_acc: 0.7132\n",
      "Epoch 20/100\n",
      " - 2s - loss: 0.7408 - acc: 0.6861 - val_loss: 0.6773 - val_acc: 0.7145\n",
      "Epoch 21/100\n",
      " - 2s - loss: 0.7331 - acc: 0.6886 - val_loss: 0.6671 - val_acc: 0.7140\n",
      "Epoch 22/100\n",
      " - 2s - loss: 0.7255 - acc: 0.6954 - val_loss: 0.6720 - val_acc: 0.7140\n",
      "Epoch 23/100\n",
      " - 2s - loss: 0.7126 - acc: 0.6965 - val_loss: 0.6700 - val_acc: 0.7072\n",
      "Epoch 24/100\n",
      " - 2s - loss: 0.7115 - acc: 0.6998 - val_loss: 0.6532 - val_acc: 0.7190\n",
      "Epoch 25/100\n",
      " - 2s - loss: 0.7046 - acc: 0.7036 - val_loss: 0.6397 - val_acc: 0.7207\n",
      "Epoch 26/100\n",
      " - 2s - loss: 0.6991 - acc: 0.7031 - val_loss: 0.6282 - val_acc: 0.7243\n",
      "Epoch 27/100\n",
      " - 2s - loss: 0.6918 - acc: 0.7088 - val_loss: 0.6380 - val_acc: 0.7153\n",
      "Epoch 28/100\n",
      " - 2s - loss: 0.6847 - acc: 0.7106 - val_loss: 0.6257 - val_acc: 0.7245\n",
      "Epoch 29/100\n",
      " - 2s - loss: 0.6937 - acc: 0.7028 - val_loss: 0.6347 - val_acc: 0.7222\n",
      "Epoch 30/100\n",
      " - 2s - loss: 0.6796 - acc: 0.7147 - val_loss: 0.6363 - val_acc: 0.7225\n",
      "Epoch 31/100\n",
      " - 2s - loss: 0.6810 - acc: 0.7126 - val_loss: 0.6228 - val_acc: 0.7245\n",
      "Epoch 32/100\n",
      " - 2s - loss: 0.6768 - acc: 0.7095 - val_loss: 0.6219 - val_acc: 0.7230\n",
      "Epoch 33/100\n",
      " - 2s - loss: 0.6791 - acc: 0.7076 - val_loss: 0.6175 - val_acc: 0.7308\n",
      "Epoch 34/100\n",
      " - 2s - loss: 0.6639 - acc: 0.7124 - val_loss: 0.6147 - val_acc: 0.7272\n",
      "Epoch 35/100\n",
      " - 2s - loss: 0.6630 - acc: 0.7170 - val_loss: 0.6120 - val_acc: 0.7330\n",
      "Epoch 36/100\n",
      " - 2s - loss: 0.6630 - acc: 0.7156 - val_loss: 0.6306 - val_acc: 0.7262\n",
      "Epoch 37/100\n",
      " - 2s - loss: 0.6520 - acc: 0.7228 - val_loss: 0.6232 - val_acc: 0.7260\n",
      "Epoch 38/100\n",
      " - 2s - loss: 0.6439 - acc: 0.7245 - val_loss: 0.6048 - val_acc: 0.7358\n",
      "Epoch 39/100\n",
      " - 2s - loss: 0.6478 - acc: 0.7294 - val_loss: 0.6154 - val_acc: 0.7270\n",
      "Epoch 40/100\n",
      " - 2s - loss: 0.6426 - acc: 0.7301 - val_loss: 0.6021 - val_acc: 0.7385\n",
      "Epoch 41/100\n",
      " - 2s - loss: 0.6386 - acc: 0.7298 - val_loss: 0.6066 - val_acc: 0.7330\n",
      "Epoch 42/100\n",
      " - 2s - loss: 0.6447 - acc: 0.7306 - val_loss: 0.5925 - val_acc: 0.7423\n",
      "Epoch 43/100\n",
      " - 2s - loss: 0.6336 - acc: 0.7297 - val_loss: 0.5940 - val_acc: 0.7363\n",
      "Epoch 44/100\n",
      " - 2s - loss: 0.6345 - acc: 0.7300 - val_loss: 0.6089 - val_acc: 0.7395\n",
      "Epoch 45/100\n",
      " - 2s - loss: 0.6287 - acc: 0.7296 - val_loss: 0.5969 - val_acc: 0.7395\n",
      "Epoch 46/100\n",
      " - 2s - loss: 0.6278 - acc: 0.7320 - val_loss: 0.6077 - val_acc: 0.7312\n",
      "Epoch 47/100\n",
      " - 2s - loss: 0.6271 - acc: 0.7337 - val_loss: 0.5923 - val_acc: 0.7408\n",
      "Epoch 48/100\n",
      " - 2s - loss: 0.6191 - acc: 0.7388 - val_loss: 0.6125 - val_acc: 0.7292\n",
      "Epoch 49/100\n",
      " - 2s - loss: 0.6185 - acc: 0.7376 - val_loss: 0.5961 - val_acc: 0.7358\n",
      "Epoch 50/100\n",
      " - 2s - loss: 0.6172 - acc: 0.7386 - val_loss: 0.5972 - val_acc: 0.7357\n",
      "Epoch 51/100\n",
      " - 2s - loss: 0.6137 - acc: 0.7391 - val_loss: 0.6049 - val_acc: 0.7320\n",
      "Epoch 52/100\n",
      " - 2s - loss: 0.6152 - acc: 0.7388 - val_loss: 0.5894 - val_acc: 0.7373\n",
      "Epoch 53/100\n",
      " - 2s - loss: 0.6098 - acc: 0.7446 - val_loss: 0.6098 - val_acc: 0.7350\n",
      "Epoch 54/100\n",
      " - 2s - loss: 0.6097 - acc: 0.7412 - val_loss: 0.5992 - val_acc: 0.7347\n",
      "Epoch 55/100\n",
      " - 2s - loss: 0.5995 - acc: 0.7452 - val_loss: 0.5890 - val_acc: 0.7385\n",
      "Epoch 56/100\n",
      " - 2s - loss: 0.6056 - acc: 0.7424 - val_loss: 0.5817 - val_acc: 0.7380\n",
      "Epoch 57/100\n",
      " - 2s - loss: 0.5957 - acc: 0.7481 - val_loss: 0.5846 - val_acc: 0.7403\n",
      "Epoch 58/100\n",
      " - 2s - loss: 0.5979 - acc: 0.7486 - val_loss: 0.5857 - val_acc: 0.7392\n",
      "Epoch 59/100\n",
      " - 2s - loss: 0.5932 - acc: 0.7490 - val_loss: 0.5835 - val_acc: 0.7420\n",
      "Epoch 60/100\n",
      " - 2s - loss: 0.5892 - acc: 0.7530 - val_loss: 0.5930 - val_acc: 0.7377\n",
      "Epoch 61/100\n",
      " - 2s - loss: 0.5954 - acc: 0.7501 - val_loss: 0.5829 - val_acc: 0.7405\n",
      "Epoch 62/100\n",
      " - 2s - loss: 0.5800 - acc: 0.7524 - val_loss: 0.6047 - val_acc: 0.7353\n",
      "Epoch 63/100\n",
      " - 2s - loss: 0.5855 - acc: 0.7559 - val_loss: 0.5875 - val_acc: 0.7408\n",
      "Epoch 64/100\n",
      " - 2s - loss: 0.5820 - acc: 0.7539 - val_loss: 0.5886 - val_acc: 0.7422\n",
      "Epoch 65/100\n",
      " - 2s - loss: 0.5747 - acc: 0.7581 - val_loss: 0.6000 - val_acc: 0.7397\n",
      "Epoch 66/100\n",
      " - 2s - loss: 0.5795 - acc: 0.7594 - val_loss: 0.5918 - val_acc: 0.7362\n",
      "Epoch 67/100\n",
      " - 2s - loss: 0.5718 - acc: 0.7596 - val_loss: 0.5934 - val_acc: 0.7412\n",
      "Epoch 68/100\n",
      " - 2s - loss: 0.5774 - acc: 0.7573 - val_loss: 0.5834 - val_acc: 0.7452\n",
      "Epoch 69/100\n",
      " - 2s - loss: 0.5703 - acc: 0.7600 - val_loss: 0.5908 - val_acc: 0.7440\n",
      "Epoch 70/100\n",
      " - 2s - loss: 0.5726 - acc: 0.7560 - val_loss: 0.5976 - val_acc: 0.7437\n",
      "Epoch 71/100\n",
      " - 2s - loss: 0.5626 - acc: 0.7611 - val_loss: 0.5923 - val_acc: 0.7363\n",
      "Epoch 72/100\n",
      " - 2s - loss: 0.5639 - acc: 0.7618 - val_loss: 0.5893 - val_acc: 0.7428\n",
      "Epoch 73/100\n",
      " - 2s - loss: 0.5714 - acc: 0.7598 - val_loss: 0.5992 - val_acc: 0.7430\n",
      "Epoch 74/100\n",
      " - 2s - loss: 0.5634 - acc: 0.7624 - val_loss: 0.5844 - val_acc: 0.7468\n",
      "Epoch 75/100\n",
      " - 2s - loss: 0.5536 - acc: 0.7639 - val_loss: 0.5911 - val_acc: 0.7407\n",
      "Epoch 76/100\n",
      " - 2s - loss: 0.5579 - acc: 0.7645 - val_loss: 0.5873 - val_acc: 0.7453\n",
      "Epoch 77/100\n",
      " - 2s - loss: 0.5580 - acc: 0.7676 - val_loss: 0.5997 - val_acc: 0.7412\n",
      "Epoch 78/100\n",
      " - 2s - loss: 0.5566 - acc: 0.7664 - val_loss: 0.5897 - val_acc: 0.7472\n",
      "Epoch 79/100\n",
      " - 2s - loss: 0.5540 - acc: 0.7706 - val_loss: 0.5811 - val_acc: 0.7470\n",
      "Epoch 80/100\n",
      " - 2s - loss: 0.5502 - acc: 0.7696 - val_loss: 0.5892 - val_acc: 0.7457\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 81/100\n",
      " - 2s - loss: 0.5492 - acc: 0.7660 - val_loss: 0.5875 - val_acc: 0.7418\n",
      "Epoch 82/100\n",
      " - 2s - loss: 0.5486 - acc: 0.7668 - val_loss: 0.5845 - val_acc: 0.7463\n",
      "Epoch 83/100\n",
      " - 2s - loss: 0.5491 - acc: 0.7676 - val_loss: 0.6038 - val_acc: 0.7423\n",
      "Epoch 84/100\n",
      " - 2s - loss: 0.5412 - acc: 0.7723 - val_loss: 0.5929 - val_acc: 0.7468\n",
      "Epoch 85/100\n",
      " - 2s - loss: 0.5436 - acc: 0.7674 - val_loss: 0.5886 - val_acc: 0.7392\n",
      "Epoch 86/100\n",
      " - 2s - loss: 0.5405 - acc: 0.7726 - val_loss: 0.5958 - val_acc: 0.7438\n",
      "Epoch 87/100\n",
      " - 2s - loss: 0.5450 - acc: 0.7704 - val_loss: 0.5964 - val_acc: 0.7393\n",
      "Epoch 88/100\n",
      " - 2s - loss: 0.5318 - acc: 0.7756 - val_loss: 0.5921 - val_acc: 0.7455\n",
      "Epoch 89/100\n",
      " - 2s - loss: 0.5340 - acc: 0.7744 - val_loss: 0.5887 - val_acc: 0.7440\n",
      "Epoch 90/100\n",
      " - 2s - loss: 0.5341 - acc: 0.7751 - val_loss: 0.6081 - val_acc: 0.7390\n",
      "Epoch 91/100\n",
      " - 2s - loss: 0.5281 - acc: 0.7817 - val_loss: 0.6046 - val_acc: 0.7385\n",
      "Epoch 92/100\n",
      " - 2s - loss: 0.5346 - acc: 0.7769 - val_loss: 0.5938 - val_acc: 0.7405\n",
      "Epoch 93/100\n",
      " - 2s - loss: 0.5331 - acc: 0.7783 - val_loss: 0.5959 - val_acc: 0.7437\n",
      "Epoch 94/100\n",
      " - 2s - loss: 0.5217 - acc: 0.7787 - val_loss: 0.6102 - val_acc: 0.7338\n",
      "Epoch 95/100\n",
      " - 2s - loss: 0.5264 - acc: 0.7749 - val_loss: 0.6013 - val_acc: 0.7388\n",
      "Epoch 96/100\n",
      " - 2s - loss: 0.5322 - acc: 0.7779 - val_loss: 0.5880 - val_acc: 0.7445\n",
      "Epoch 97/100\n",
      " - 2s - loss: 0.5257 - acc: 0.7799 - val_loss: 0.5953 - val_acc: 0.7397\n",
      "Epoch 98/100\n",
      " - 2s - loss: 0.5255 - acc: 0.7794 - val_loss: 0.5851 - val_acc: 0.7448\n",
      "Epoch 99/100\n",
      " - 2s - loss: 0.5218 - acc: 0.7788 - val_loss: 0.5962 - val_acc: 0.7455\n",
      "Epoch 100/100\n",
      " - 2s - loss: 0.5223 - acc: 0.7800 - val_loss: 0.5967 - val_acc: 0.7418\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x126b9f9d0>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x126ce77d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "input_shape = (1, 24, 6)\n",
    "model = Sequential()\n",
    "\n",
    "num_classes=7\n",
    "model.add(Conv2D(32, kernel_size=(1, 2),\n",
    "                 activation='elu',\n",
    "                 input_shape=input_shape))\n",
    "model.add(Conv2D(64, (1, 2), activation='elu', name = \"conv2\"))\n",
    "model.add(MaxPooling2D(pool_size=(1, 1)))\n",
    "model.add(Dropout(0.3))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(64, activation='selu'))\n",
    "model.add(Dropout(0.4))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "model.summary()\n",
    "model.compile(optimizer='rmsprop', loss='categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "model.fit(X_train, Y_train,\n",
    "          batch_size=64,\n",
    "          epochs=100,\n",
    "          verbose=2,\n",
    "          validation_data=(X_test, Y_test), callbacks=[plot_learning])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import h5py\n",
    "model.save('dac18_GF_delay_model_save_test.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_confusion_matrix(cm, classes,\n",
    "                          normalize=False,\n",
    "                          print_raw_cm=False,\n",
    "                          title='Confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        if (print_raw_cm):\n",
    "            print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        if (print_raw_cm):\n",
    "            print('Confusion matrix, without normalization')\n",
    "\n",
    "    if (print_raw_cm):\n",
    "        print(cm)\n",
    "\n",
    "    plt.imshow(cm, interpolation='nearest', cmap=plt.cm.Greys)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, format(cm[i, j], fmt),\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    #plt.tight_layout()\n",
    "    plt.ylabel('True label')\n",
    "    plt.xlabel('Predicted label')\n",
    "def evaluate(model, X_test, Y_test, classes,NAME):\n",
    "    \"\"\"evaluate dataset and plot confusion matrix\"\"\"\n",
    "    # predict from X_test\n",
    "    Y_pred = model.predict(X_test)\n",
    "    Y_pred_label = np.argmax(Y_pred, axis=-1)\n",
    "    Y_test_label = np.argmax(Y_test, axis=-1)\n",
    "\n",
    "\n",
    "    print(\"Prediction accuracy = %f\" %(float(sum(Y_pred_label == Y_test_label))/Y_test_label.shape[0]))\n",
    "    cnf_matrix = confusion_matrix(Y_test_label, Y_pred_label)\n",
    "\n",
    "\n",
    "    np.set_printoptions(precision=2)\n",
    "    plt.figure()\n",
    "    plot_confusion_matrix(cnf_matrix, classes=classes, normalize=True,\n",
    "                          title='Normalized confusion matrix\\n Angel-class:class-0\\n Devil-class:class-6')\n",
    "    \n",
    "    plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction accuracy = 0.741833\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x128bb4790>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "##evaluate training results\n",
    "evaluate(model, X_test, Y_test, classes=range(0,Y.shape[1]),NAME = \"origin_delay_\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x128d8da50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_learning.plot_history()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "((20000, 1, 24, 6), (20000, 7))\n",
      "Prediction accuracy = 0.736200\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x12838cb10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## testing (inference)\n",
    "# 1st testing input datasets: 20000 untested flows, test2.csv\n",
    "\n",
    "folder_path = 'datasets/64bitGF/delay/'\n",
    "csv_filename = 'test2.csv'\n",
    "X, Y = read_csv_data(folder_path + csv_filename, header=None, num_funcs=6)\n",
    "print(X.shape, Y.shape)\n",
    "evaluate(model, X, Y, classes=range(0,Y.shape[1]),NAME = \"origin_delay_\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction accuracy = 0.733800\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x128da7750>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## testing (inference)\n",
    "# 2nd testing input datasets: 20000 untested flows, test3.csv\n",
    "\n",
    "folder_path = 'datasets/64bitGF/delay/'\n",
    "csv_filename = 'test3.csv'\n",
    "X, Y = read_csv_data(folder_path + csv_filename, header=None, num_funcs=6)\n",
    "#print(X.shape, Y.shape)\n",
    "evaluate(model, X, Y, classes=range(0,Y.shape[1]),NAME = \"origin_delay_\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction accuracy = 0.736150\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x128412dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "## testing (inference)\n",
    "# 3rd testing input datasets: 20000 untested flows, test4.csv\n",
    "\n",
    "folder_path = 'datasets/64bitGF/delay/'\n",
    "csv_filename = 'test4.csv'\n",
    "X, Y = read_csv_data(folder_path + csv_filename, header=None, num_funcs=6)\n",
    "#print(X.shape, Y.shape)\n",
    "evaluate(model, X, Y, classes=range(0,Y.shape[1]),NAME = \"origin_delay_\")"
   ]
  }
 ],
 "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": 2
}