Wavelet_Transform.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import librosa.display\n",
    "import pylab\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score\n",
    "import librosa    \n",
    "import glob\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.utils import to_categorical\n",
    "from keras import optimizers\n",
    "import keras\n",
    "import pywt\n",
    "\n",
    "%matplotlib inline\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "audio_data = []\n",
    "labels = []\n",
    "sampling_rate = []\n",
    "file_names = []\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(220000,)\n",
      "44000\n"
     ]
    }
   ],
   "source": [
    "#samplerate, data = wavfile.read(\"dataset/dog/1-30344-A.wav\")\n",
    "data, samplerate = librosa.load(\"dataset/dog/1-30344-A.wav\", sr=44000) # Downsample 44.1kHz to 8kHz\n",
    "\n",
    "print(data.shape)\n",
    "print(samplerate)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5.0\n",
      "2.2727272727272726e-05\n",
      "176000.0\n"
     ]
    }
   ],
   "source": [
    "time_sec = (len(data)/samplerate)\n",
    "step = time_sec/len(data)\n",
    "print(time_sec)\n",
    "print(step)\n",
    "i=0\n",
    "time_divion=[]\n",
    "while i<=time_sec-step:\n",
    "    \n",
    "    time_divion.append(i)\n",
    "    i=i+step\n",
    "# the fourth second step\n",
    "four_sec_step_number = (4*len(time_divion))/time_sec\n",
    "print(four_sec_step_number)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['004 - Baby cry', '005 - Clock tick', '006 - Person sneeze', '007 - Helicopter', '008 - Chainsaw', '009 - Rooster', '010 - Fire crackling', 'dog', 'rain', 'sea']\n",
      "the class = 004 - Baby cry, the label = 0\n",
      "the class = 005 - Clock tick, the label = 1\n",
      "the class = 006 - Person sneeze, the label = 2\n",
      "the class = 007 - Helicopter, the label = 3\n",
      "the class = 008 - Chainsaw, the label = 4\n",
      "the class = 009 - Rooster, the label = 5\n",
      "the class = 010 - Fire crackling, the label = 6\n",
      "the class = dog, the label = 7\n",
      "the class = rain, the label = 8\n",
      "the class = sea, the label = 9\n",
      "0\n"
     ]
    }
   ],
   "source": [
    "classes = []\n",
    "label_number=0\n",
    "audio_data = []\n",
    "labels = []\n",
    "sampling_rate = []\n",
    "file_names = []\n",
    "data = []\n",
    "noisy_removed=[]\n",
    "noise=[]\n",
    "for filepath in glob.iglob('dataset/*'):\n",
    "    \n",
    "    #print(filepath[9:])\n",
    "    \n",
    "    #print(filepath)\n",
    "    classes.append(filepath[8:])\n",
    "\n",
    "print(classes)\n",
    "\n",
    "for i in classes:\n",
    "    print(\"the class = \"+i+\", the label = \"+str(label_number))\n",
    "    for j in glob.iglob('dataset/'+i+'/*'):\n",
    "        #samplerate, data = wavfile.read(j)\n",
    "        y, s = librosa.load(j, sr=44000) # Downsample 44.1kHz to 8kHz\n",
    "        #reduced_noise = nr.reduce_noise(audio_clip=y, noise_clip=y, verbose=False)\n",
    "        #print(s)\n",
    "        #print(j)\n",
    "        data.append([y,label_number])\n",
    "        #noise.append(y)\n",
    "        #labels.append(label_number)\n",
    "        \n",
    "    label_number = label_number + 1\n",
    "print(len(labels))\n",
    "#print(labels)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "\n",
    "random.shuffle(data)\n",
    "audio_data=[]\n",
    "labels=[]\n",
    "for i,j in data:\n",
    "    audio_data.append(i)\n",
    "    labels.append(j)\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9\n"
     ]
    }
   ],
   "source": [
    "print(labels[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9\n"
     ]
    }
   ],
   "source": [
    "example = random.randint(0,400)\n",
    "print(labels[example])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_path='sea_before.jpg'\n",
    "plt.plot(time_divion[0:192000],audio_data[example][0:192000])\n",
    "plt.show()\n",
    "plt.title('signal in real time')\n",
    "#pylab.savefig(save_path, bbox_inches=None, pad_inches=0)\n",
    "#pylab.close()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 192000)\n"
     ]
    }
   ],
   "source": [
    "scales = np.arange(1, 101)\n",
    "coeff1, freqs1 = pywt.cwt(audio_data[example][0:192000], scales, 'morl')\n",
    "print(coeff1.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "38"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_path = 'sea.jpg'\n",
    "\n",
    "plt.figure(1, figsize=(20,10))\n",
    "plt.subplot(121)\n",
    "plt.imshow(coeff1, cmap='coolwarm', aspect='auto')  \n",
    "plt.show() \n",
    "plt.title('Wavlet Transform')\n",
    "#pylab.savefig(save_path, bbox_inches=None, pad_inches=0)\n",
    "#pylab.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "1\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "2\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "3\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "4\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "5\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "6\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "7\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "8\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "9\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "10\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "11\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "12\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "13\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "14\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "15\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "16\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "17\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "18\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "19\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "20\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "21\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "22\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "23\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "24\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "25\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "26\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "27\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "28\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "29\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "30\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "31\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "32\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "33\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "34\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "35\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "36\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "37\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "38\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "39\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "40\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "41\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "42\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "43\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "44\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "45\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "46\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "47\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "48\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "49\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "50\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "51\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "52\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "53\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "54\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "55\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "56\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "57\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "58\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "59\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "60\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "61\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "62\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "63\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "64\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "65\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "66\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "67\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "68\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "69\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "70\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "71\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "72\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "73\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "74\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "75\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "76\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "77\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "78\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "79\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "80\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "81\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "82\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "83\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "84\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "85\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "86\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "87\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "88\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "89\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "90\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "91\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "92\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "93\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "94\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "95\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "96\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "97\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "98\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "99\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "100\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "101\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "102\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "103\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "104\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "105\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "106\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "107\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "108\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "109\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "110\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "111\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "112\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "113\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "114\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "115\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "116\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "117\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "118\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "119\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "120\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "121\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "122\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "123\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "124\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "125\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "126\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "127\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "128\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "129\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "130\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "131\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "132\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "133\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "134\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "135\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "136\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "137\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "138\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "139\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "140\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "141\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "142\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "143\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "144\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "145\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "146\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "147\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "148\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "149\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "150\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "151\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "152\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "153\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "154\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "155\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "156\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "157\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "158\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "159\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "160\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "161\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "162\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "163\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "164\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "165\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "166\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "167\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "168\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "169\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "170\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "171\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "172\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "173\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "174\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "175\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "176\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "177\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "178\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "179\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "180\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "181\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "182\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "183\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "184\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "185\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "186\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "187\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "188\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "189\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "190\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "191\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "192\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "193\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "194\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "195\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "196\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "197\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "198\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "199\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "200\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "201\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "202\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "203\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "204\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "205\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "206\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "207\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "208\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "209\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "210\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "211\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "212\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "213\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "214\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "215\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "216\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "217\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "218\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "219\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "220\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "221\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "222\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "223\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "224\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "225\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "226\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "227\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "228\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "229\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "230\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "231\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "232\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "233\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "234\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "235\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "236\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "237\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "238\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "239\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "240\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "241\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "242\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "243\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "244\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "245\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "246\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "247\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "248\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "249\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "250\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "251\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "252\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "253\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "254\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "255\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "256\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "257\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "258\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "259\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "260\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "261\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "262\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "263\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "264\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "265\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "266\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "267\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "268\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "269\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "270\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "271\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "272\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "273\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "274\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "275\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "276\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "277\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "278\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "279\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "280\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "281\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "282\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "283\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "284\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "285\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "286\n",
      "(100, 192000)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(100, 100)\n",
      "287\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "288\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "289\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "290\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "291\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "292\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "293\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "294\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "295\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "296\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "297\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "298\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "299\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "300\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "301\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "302\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "303\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "304\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "305\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "306\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "307\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "308\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "309\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "310\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "311\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "312\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "313\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "314\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "315\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "316\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "317\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "318\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "319\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "320\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "321\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "322\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "323\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "324\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "325\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "326\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "327\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "328\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "329\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "330\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "331\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "332\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "333\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "334\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "335\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "336\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "337\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "338\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "339\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "340\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "341\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "342\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "343\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "344\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "345\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "346\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "347\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "348\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "349\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "350\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "351\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "352\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "353\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "354\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "355\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "356\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "357\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "358\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "359\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "360\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "361\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "362\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "363\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "364\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "365\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "366\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "367\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "368\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "369\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "370\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "371\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "372\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "373\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "374\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "375\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "376\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "377\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "378\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "379\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "380\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "381\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "382\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "383\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "384\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "385\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "386\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "387\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "388\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "389\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "390\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "391\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "392\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "393\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "394\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "395\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "396\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "397\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "398\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "399\n",
      "(100, 192000)\n",
      "(100, 100)\n",
      "(400, 100, 100)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.decomposition import PCA\n",
    "\n",
    "pca_2 = PCA(n_components=100)\n",
    "\n",
    "wavelet=[]\n",
    "for x in range(0,len(audio_data)):\n",
    "    \n",
    "    scales = np.arange(1, 101)\n",
    "    coeff1, freqs1 = pywt.cwt(audio_data[x][0:192000], scales, 'morl')\n",
    "    print(x)\n",
    "    print(coeff1.shape)\n",
    "    \n",
    "    print(pca_2.fit_transform(coeff1).shape)\n",
    "    wavelet.append(pca_2.fit_transform(coeff1))\n",
    "\n",
    "wavelet=np.array(wavelet) \n",
    "print(wavelet.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(400, 10000)"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "wavelet=wavelet.reshape(400,-1)\n",
    "wavelet.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "646.8539956483597\n",
      "1.0000000000000002\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import MinMaxScaler\n",
    "scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "scaler.fit(wavelet)\n",
    "\n",
    "\n",
    "print(np.amax(wavelet))\n",
    "\n",
    "normalized_features = scaler.transform(wavelet)\n",
    "print(np.amax(normalized_features))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeClassifier(class_weight=None, criterion='gini', max_depth=None,\n",
       "            max_features=None, max_leaf_nodes=None,\n",
       "            min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "            min_samples_leaf=1, min_samples_split=2,\n",
       "            min_weight_fraction_leaf=0.0, presort=False, random_state=None,\n",
       "            splitter='best')"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import tree\n",
    "X_train, X_test, y_train, y_test = train_test_split(normalized_features, labels, test_size=0.20, random_state=1150)\n",
    "clf = tree.DecisionTreeClassifier()\n",
    "clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.35\n"
     ]
    }
   ],
   "source": [
    "y_pred = clf.predict(X_test)\n",
    "print(accuracy_score(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "features_convolution = np.reshape(normalized_features,(400,100, -1,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "y=keras.utils.to_categorical(labels, num_classes=10, dtype='float32')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Conv2D(16, (3, 3), input_shape=features_convolution.shape[1:]))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "#'''\n",
    "#model.add(Dropout(0.2))\n",
    "\n",
    "model.add(Conv2D(32, (3, 3)))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "\n",
    "#model.add(Dropout(0.2))\n",
    "\n",
    "#'''\n",
    "#'''\n",
    "model.add(Conv2D(64, (3, 3),padding='same'))\n",
    "model.add(Activation('relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "#'''\n",
    "\n",
    "\n",
    "model.add(Flatten())  # this converts our 3D feature maps to 1D feature vectors\n",
    "\n",
    "#model.add(Dense(1000))#input_shape=features.shape[1:]\n",
    "model.add(Dense(64))#input_shape=features.shape[1:]\n",
    "\n",
    "model.add(Dense(10))\n",
    "model.add(Activation('softmax'))\n",
    "sgd = optimizers.SGD(lr=0.0000001, decay=1e-6, momentum=0.9, nesterov=True)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 320 samples, validate on 80 samples\n",
      "Epoch 1/40\n",
      "320/320 [==============================] - 10s 31ms/step - loss: 2.3322 - acc: 0.0750 - val_loss: 2.3450 - val_acc: 0.0750\n",
      "Epoch 2/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 2.2789 - acc: 0.1219 - val_loss: 2.3938 - val_acc: 0.0750\n",
      "Epoch 3/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 2.2477 - acc: 0.1406 - val_loss: 2.2926 - val_acc: 0.0875\n",
      "Epoch 4/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 2.1666 - acc: 0.1781 - val_loss: 2.2968 - val_acc: 0.0875\n",
      "Epoch 5/40\n",
      "320/320 [==============================] - 4s 12ms/step - loss: 2.0797 - acc: 0.2375 - val_loss: 2.0228 - val_acc: 0.2750\n",
      "Epoch 6/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 1.9743 - acc: 0.2844 - val_loss: 1.9715 - val_acc: 0.2875\n",
      "Epoch 7/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.7807 - acc: 0.3594 - val_loss: 1.8099 - val_acc: 0.3875\n",
      "Epoch 8/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.6260 - acc: 0.4156 - val_loss: 1.6222 - val_acc: 0.4000\n",
      "Epoch 9/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.7029 - acc: 0.4156 - val_loss: 1.6537 - val_acc: 0.3375\n",
      "Epoch 10/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.4137 - acc: 0.4906 - val_loss: 1.4820 - val_acc: 0.4375\n",
      "Epoch 11/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.3227 - acc: 0.5625 - val_loss: 1.3682 - val_acc: 0.5875\n",
      "Epoch 12/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 1.2160 - acc: 0.5687 - val_loss: 1.4040 - val_acc: 0.5500\n",
      "Epoch 13/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 1.1859 - acc: 0.5844 - val_loss: 1.4116 - val_acc: 0.5125\n",
      "Epoch 14/40\n",
      "320/320 [==============================] - 5s 15ms/step - loss: 1.1135 - acc: 0.6406 - val_loss: 1.3254 - val_acc: 0.5750\n",
      "Epoch 15/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 1.0270 - acc: 0.6500 - val_loss: 1.4400 - val_acc: 0.5500\n",
      "Epoch 16/40\n",
      "320/320 [==============================] - 5s 14ms/step - loss: 0.9267 - acc: 0.7344 - val_loss: 1.1948 - val_acc: 0.6125\n",
      "Epoch 17/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 0.8745 - acc: 0.7094 - val_loss: 1.3423 - val_acc: 0.5500\n",
      "Epoch 18/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 0.8289 - acc: 0.7219 - val_loss: 1.2897 - val_acc: 0.6125\n",
      "Epoch 19/40\n",
      "320/320 [==============================] - 4s 12ms/step - loss: 0.6934 - acc: 0.7812 - val_loss: 1.3981 - val_acc: 0.5625\n",
      "Epoch 20/40\n",
      "320/320 [==============================] - 4s 11ms/step - loss: 0.6115 - acc: 0.8188 - val_loss: 1.4254 - val_acc: 0.5750\n",
      "Epoch 21/40\n",
      "320/320 [==============================] - 4s 11ms/step - loss: 0.5756 - acc: 0.8219 - val_loss: 1.4646 - val_acc: 0.5750\n",
      "Epoch 22/40\n",
      "320/320 [==============================] - 4s 12ms/step - loss: 0.5223 - acc: 0.8438 - val_loss: 1.4736 - val_acc: 0.5875\n",
      "Epoch 23/40\n",
      "320/320 [==============================] - 4s 12ms/step - loss: 0.5059 - acc: 0.8531 - val_loss: 1.5830 - val_acc: 0.5375\n",
      "Epoch 24/40\n",
      "320/320 [==============================] - 4s 12ms/step - loss: 0.4762 - acc: 0.8406 - val_loss: 1.6124 - val_acc: 0.5125\n",
      "Epoch 25/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 0.4941 - acc: 0.8281 - val_loss: 1.4844 - val_acc: 0.6000\n",
      "Epoch 26/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.3757 - acc: 0.9000 - val_loss: 1.4899 - val_acc: 0.5125\n",
      "Epoch 27/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.3635 - acc: 0.8625 - val_loss: 1.6169 - val_acc: 0.5625\n",
      "Epoch 28/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 0.5224 - acc: 0.8375 - val_loss: 1.6762 - val_acc: 0.5500\n",
      "Epoch 29/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.4931 - acc: 0.8562 - val_loss: 1.6639 - val_acc: 0.5500\n",
      "Epoch 30/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.3210 - acc: 0.9125 - val_loss: 1.5203 - val_acc: 0.5625\n",
      "Epoch 31/40\n",
      "320/320 [==============================] - 4s 13ms/step - loss: 0.2594 - acc: 0.9250 - val_loss: 1.5598 - val_acc: 0.6000\n",
      "Epoch 32/40\n",
      "320/320 [==============================] - 5s 15ms/step - loss: 0.2418 - acc: 0.9563 - val_loss: 1.8347 - val_acc: 0.5250\n",
      "Epoch 33/40\n",
      "320/320 [==============================] - 5s 14ms/step - loss: 0.2157 - acc: 0.9281 - val_loss: 1.7423 - val_acc: 0.5500\n",
      "Epoch 34/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.2237 - acc: 0.9469 - val_loss: 1.7895 - val_acc: 0.5500\n",
      "Epoch 35/40\n",
      "320/320 [==============================] - 5s 15ms/step - loss: 0.1862 - acc: 0.9500 - val_loss: 1.6385 - val_acc: 0.5875\n",
      "Epoch 36/40\n",
      "320/320 [==============================] - 5s 15ms/step - loss: 0.2028 - acc: 0.9438 - val_loss: 1.8627 - val_acc: 0.5500\n",
      "Epoch 37/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.2028 - acc: 0.9406 - val_loss: 1.7613 - val_acc: 0.5750\n",
      "Epoch 38/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.1681 - acc: 0.9500 - val_loss: 1.7873 - val_acc: 0.5625\n",
      "Epoch 39/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.1375 - acc: 0.9625 - val_loss: 1.7646 - val_acc: 0.5625\n",
      "Epoch 40/40\n",
      "320/320 [==============================] - 4s 14ms/step - loss: 0.1467 - acc: 0.9625 - val_loss: 1.9131 - val_acc: 0.5625\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(features_convolution, y,batch_size=16, epochs=40,validation_split=0.2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend(loc=0)\n",
    "plt.figure()\n",
    "\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}