breast_cancer_detection_Model04.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[name: \"/device:CPU:0\"\n",
      "device_type: \"CPU\"\n",
      "memory_limit: 268435456\n",
      "locality {\n",
      "}\n",
      "incarnation: 927931377932929411\n",
      ", name: \"/device:GPU:0\"\n",
      "device_type: \"GPU\"\n",
      "memory_limit: 4950209331\n",
      "locality {\n",
      "  bus_id: 1\n",
      "  links {\n",
      "  }\n",
      "}\n",
      "incarnation: 1486559521091559177\n",
      "physical_device_desc: \"device: 0, name: GeForce GTX 1060, pci bus id: 0000:01:00.0, compute capability: 6.1\"\n",
      "]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from tensorflow.python.client import device_lib\n",
    "print(device_lib.list_local_devices())\n",
    "import matplotlib.pylab as plt\n",
    "import seaborn as sns\n",
    "from glob import glob\n",
    "import fnmatch\n",
    "import cv2\n",
    "import argparse\n",
    "import pickle\n",
    "import os\n",
    "import sklearn\n",
    "from sklearn import model_selection\n",
    "from sklearn.model_selection import train_test_split, KFold, cross_val_score, StratifiedKFold, learning_curve, GridSearchCV\n",
    "from sklearn.metrics import confusion_matrix, make_scorer, accuracy_score, classification_report\n",
    "import keras\n",
    "from keras import backend as K\n",
    "K.tensorflow_backend._get_available_gpus()\n",
    "from keras.callbacks import Callback, EarlyStopping, ReduceLROnPlateau, ModelCheckpoint\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.utils import plot_model\n",
    "from keras.utils.np_utils import to_categorical\n",
    "from keras.utils.vis_utils import model_to_dot\n",
    "from keras.models import Sequential, model_from_json\n",
    "from keras.optimizers import SGD, RMSprop, Adam, Adagrad, Adadelta\n",
    "from keras.layers import Dense, Dropout, Embedding, SpatialDropout1D, Activation, Flatten, BatchNormalization, Conv2D, MaxPool2D, MaxPooling2D\n",
    "%matplotlib inline\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1001_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1051_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1101_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1151_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1201_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1251_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1301_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1351_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1501_class0.png\n",
      "IDC_regular_ps50_idx5\\10253\\0\\10253_idx5_x1001_y1551_class0.png\n"
     ]
    }
   ],
   "source": [
    "# source: https://www.kaggle.com/paultimothymooney/predicting-idc-in-breast-cancer-histology-images/notebook\n",
    "imagePatches = glob('IDC_regular_ps50_idx5/**/*.png', recursive=True) # search pathname/folder to find .png files recursively then parse files\n",
    "for filename in imagePatches[0:10]:\n",
    "    print(filename)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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"
    }
   ],
   "source": [
    "image_name = \"IDC_regular_ps50_idx5/9135/1/9135_idx5_x1701_y1851_class1.png\" #Image to be used as query\n",
    "def plotImage(image_location):\n",
    "    image = cv2.imread(image_name)\n",
    "    image = cv2.resize(image, (50,50))\n",
    "    plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB)); plt.axis('off')\n",
    "    return\n",
    "plotImage(image_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "IDC(-)\n",
      "\n",
      " ['IDC_regular_ps50_idx5\\\\10253\\\\0\\\\10253_idx5_x1001_y1001_class0.png', 'IDC_regular_ps50_idx5\\\\10253\\\\0\\\\10253_idx5_x1001_y1051_class0.png', 'IDC_regular_ps50_idx5\\\\10253\\\\0\\\\10253_idx5_x1001_y1101_class0.png', 'IDC_regular_ps50_idx5\\\\10253\\\\0\\\\10253_idx5_x1001_y1151_class0.png', 'IDC_regular_ps50_idx5\\\\10253\\\\0\\\\10253_idx5_x1001_y1201_class0.png'] \n",
      "\n",
      "IDC(+)\n",
      "\n",
      " ['IDC_regular_ps50_idx5\\\\10253\\\\1\\\\10253_idx5_x501_y351_class1.png', 'IDC_regular_ps50_idx5\\\\10253\\\\1\\\\10253_idx5_x501_y401_class1.png', 'IDC_regular_ps50_idx5\\\\10253\\\\1\\\\10253_idx5_x551_y301_class1.png', 'IDC_regular_ps50_idx5\\\\10253\\\\1\\\\10253_idx5_x551_y351_class1.png', 'IDC_regular_ps50_idx5\\\\10253\\\\1\\\\10253_idx5_x551_y401_class1.png']\n"
     ]
    }
   ],
   "source": [
    "patternZero = '*class0.png'\n",
    "patternOne = '*class1.png'\n",
    "classZero = fnmatch.filter(imagePatches, patternZero) #filename pattern matching returns subset of list of names\n",
    "classOne = fnmatch.filter(imagePatches, patternOne)\n",
    "print(\"IDC(-)\\n\\n\",classZero[0:5],'\\n')\n",
    "print(\"IDC(+)\\n\\n\",classOne[0:5])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def proc_images(lowerIndex,upperIndex):\n",
    "    \"\"\"\n",
    "    Returns two arrays: \n",
    "        x is an array of resized images\n",
    "        y is an array of labels\n",
    "    \"\"\" \n",
    "    x = []\n",
    "    y = []\n",
    "    WIDTH = 50\n",
    "    HEIGHT = 50\n",
    "    for img in imagePatches[lowerIndex:upperIndex]:\n",
    "        full_size_image = cv2.imread(img)\n",
    "        x.append(cv2.resize(full_size_image, (WIDTH,HEIGHT), interpolation=cv2.INTER_CUBIC))\n",
    "        if img in classZero:\n",
    "            y.append(0)\n",
    "        elif img in classOne:\n",
    "            y.append(1)\n",
    "        else:\n",
    "            return\n",
    "    return x,y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Wall time: 0 ns\n"
     ]
    }
   ],
   "source": [
    "%time\n",
    "X,Y = proc_images(0,100000)\n",
    "df = pd.DataFrame()\n",
    "df[\"images\"]=X\n",
    "df[\"labels\"]=Y\n",
    "X2=df[\"images\"]\n",
    "Y2=df[\"labels\"]\n",
    "X2=np.array(X2)\n",
    "imgs0=[]\n",
    "imgs1=[]\n",
    "imgs0 = X2[Y2==0] # (0 = no IDC, 1 = IDC)\n",
    "imgs1 = X2[Y2==1] "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of images: 100000\n",
      "Number of IDC(-) Images: 70901\n",
      "Number of IDC(+) Images: 29099\n",
      "Percentage of positive images: 29.10%\n",
      "Image shape (Width, Height, Channels): (50, 50, 3)\n"
     ]
    }
   ],
   "source": [
    "def describeData(a,b):\n",
    "    print('Total number of images: {}'.format(len(a)))\n",
    "    print('Number of IDC(-) Images: {}'.format(np.sum(b==0)))\n",
    "    print('Number of IDC(+) Images: {}'.format(np.sum(b==1)))\n",
    "    print('Percentage of positive images: {:.2f}%'.format(100*np.mean(b)))\n",
    "    print('Image shape (Width, Height, Channels): {}'.format(a[0].shape))\n",
    "describeData(X2,Y2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                              images  labels\n",
      "0  [[[206, 164, 226], [196, 154, 224], [211, 175,...       0\n",
      "1  [[[197, 150, 219], [201, 158, 217], [205, 173,...       0\n",
      "2  [[[249, 245, 248], [248, 246, 248], [253, 246,...       0\n",
      "3  [[[249, 247, 249], [249, 247, 249], [249, 247,...       0\n",
      "4  [[[237, 231, 241], [245, 242, 246], [235, 222,...       0\n",
      "5  [[[248, 246, 248], [248, 246, 248], [248, 246,...       0\n",
      "6  [[[175, 120, 212], [145, 88, 175], [132, 72, 1...       0\n",
      "7  [[[87, 38, 142], [120, 94, 165], [195, 180, 21...       0\n",
      "8  [[[235, 228, 240], [242, 240, 246], [252, 244,...       0\n",
      "9  [[[238, 223, 242], [239, 235, 241], [250, 245,...       0\n",
      "\n",
      "{0: 'IDC(-)', 1: 'IDC(+)'}\n"
     ]
    }
   ],
   "source": [
    "dict_characters = {0: 'IDC(-)', 1: 'IDC(+)'}\n",
    "print(df.head(10))\n",
    "print(\"\")\n",
    "print(dict_characters)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plotOne(a,b):\n",
    "    \"\"\"\n",
    "    Plot one numpy array\n",
    "    \"\"\"\n",
    "    plt.subplot(1,2,1)\n",
    "    plt.title('IDC (-)')\n",
    "    plt.imshow(a[100])\n",
    "    plt.subplot(1,2,2)\n",
    "    plt.title('IDC (+)')\n",
    "    plt.imshow(b[100])\n",
    "plotOne(imgs0, imgs1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X=np.array(X)\n",
    "X=X/255.0\n",
    "\n",
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.20, random_state=42)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] training network...\n",
      "Train on 80000 samples, validate on 20000 samples\n",
      "Epoch 1/75\n",
      "80000/80000 [==============================] - 16s 195us/step - loss: 0.5639 - acc: 0.7273 - val_loss: 0.5104 - val_acc: 0.7517\n",
      "Epoch 2/75\n",
      "80000/80000 [==============================] - 8s 96us/step - loss: 0.5002 - acc: 0.7624 - val_loss: 0.4905 - val_acc: 0.7704\n",
      "Epoch 3/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4898 - acc: 0.7687 - val_loss: 0.4831 - val_acc: 0.7710\n",
      "Epoch 4/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4882 - acc: 0.7697 - val_loss: 0.4864 - val_acc: 0.7726\n",
      "Epoch 5/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4846 - acc: 0.7717 - val_loss: 0.4823 - val_acc: 0.7711\n",
      "Epoch 6/75\n",
      "80000/80000 [==============================] - 7s 94us/step - loss: 0.4841 - acc: 0.7716 - val_loss: 0.4817 - val_acc: 0.7708\n",
      "Epoch 7/75\n",
      "80000/80000 [==============================] - 8s 94us/step - loss: 0.4836 - acc: 0.7721 - val_loss: 0.4820 - val_acc: 0.7718\n",
      "Epoch 8/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4817 - acc: 0.7720 - val_loss: 0.4791 - val_acc: 0.7721\n",
      "Epoch 9/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4800 - acc: 0.7726 - val_loss: 0.4794 - val_acc: 0.7714\n",
      "Epoch 10/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4796 - acc: 0.7716 - val_loss: 0.4803 - val_acc: 0.7715\n",
      "Epoch 11/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4785 - acc: 0.7734 - val_loss: 0.4769 - val_acc: 0.7731\n",
      "Epoch 12/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4770 - acc: 0.7738 - val_loss: 0.4757 - val_acc: 0.7733\n",
      "Epoch 13/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4745 - acc: 0.7752 - val_loss: 0.4758 - val_acc: 0.7725\n",
      "Epoch 14/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4736 - acc: 0.7759 - val_loss: 0.4730 - val_acc: 0.7746\n",
      "Epoch 15/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4718 - acc: 0.7761 - val_loss: 0.4715 - val_acc: 0.7748\n",
      "Epoch 16/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4697 - acc: 0.7765 - val_loss: 0.4784 - val_acc: 0.7715\n",
      "Epoch 17/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4686 - acc: 0.7780 - val_loss: 0.4749 - val_acc: 0.7728\n",
      "Epoch 18/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4668 - acc: 0.7783 - val_loss: 0.4696 - val_acc: 0.7751\n",
      "Epoch 19/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4661 - acc: 0.7780 - val_loss: 0.4678 - val_acc: 0.7748\n",
      "Epoch 20/75\n",
      "80000/80000 [==============================] - 8s 95us/step - loss: 0.4643 - acc: 0.7810 - val_loss: 0.4667 - val_acc: 0.7771\n",
      "Epoch 21/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4643 - acc: 0.7795 - val_loss: 0.4700 - val_acc: 0.7803\n",
      "Epoch 22/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4627 - acc: 0.7818 - val_loss: 0.4707 - val_acc: 0.7808\n",
      "Epoch 23/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4628 - acc: 0.7821 - val_loss: 0.4710 - val_acc: 0.7736\n",
      "Epoch 24/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4619 - acc: 0.7812 - val_loss: 0.4681 - val_acc: 0.7750\n",
      "Epoch 25/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4607 - acc: 0.7828 - val_loss: 0.4657 - val_acc: 0.7804\n",
      "Epoch 26/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4603 - acc: 0.7835 - val_loss: 0.4632 - val_acc: 0.7781\n",
      "Epoch 27/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4593 - acc: 0.7839 - val_loss: 0.4811 - val_acc: 0.7826\n",
      "Epoch 28/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4587 - acc: 0.7836 - val_loss: 0.4713 - val_acc: 0.7839\n",
      "Epoch 29/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4579 - acc: 0.7844 - val_loss: 0.4675 - val_acc: 0.7832\n",
      "Epoch 30/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4579 - acc: 0.7842 - val_loss: 0.4672 - val_acc: 0.7817\n",
      "Epoch 31/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4568 - acc: 0.7847 - val_loss: 0.5347 - val_acc: 0.7481\n",
      "Epoch 32/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4562 - acc: 0.7857 - val_loss: 0.4615 - val_acc: 0.7787\n",
      "Epoch 33/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4557 - acc: 0.7855 - val_loss: 0.4678 - val_acc: 0.7756\n",
      "Epoch 34/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4560 - acc: 0.7857 - val_loss: 0.4970 - val_acc: 0.7741\n",
      "Epoch 35/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4546 - acc: 0.7875 - val_loss: 0.4631 - val_acc: 0.7769\n",
      "Epoch 36/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4544 - acc: 0.7880 - val_loss: 0.4636 - val_acc: 0.7775\n",
      "Epoch 37/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4539 - acc: 0.7875 - val_loss: 0.4583 - val_acc: 0.7843\n",
      "Epoch 38/75\n",
      "80000/80000 [==============================] - 8s 95us/step - loss: 0.4526 - acc: 0.7887 - val_loss: 0.4580 - val_acc: 0.7854\n",
      "Epoch 39/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4524 - acc: 0.7893 - val_loss: 0.4578 - val_acc: 0.7822\n",
      "Epoch 40/75\n",
      "80000/80000 [==============================] - 7s 94us/step - loss: 0.4515 - acc: 0.7894 - val_loss: 0.4630 - val_acc: 0.7879\n",
      "Epoch 41/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4511 - acc: 0.7906 - val_loss: 0.4600 - val_acc: 0.7795\n",
      "Epoch 42/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4500 - acc: 0.7915 - val_loss: 0.4576 - val_acc: 0.7880\n",
      "Epoch 43/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4504 - acc: 0.7907 - val_loss: 0.4709 - val_acc: 0.7891\n",
      "Epoch 44/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4494 - acc: 0.7918 - val_loss: 0.4562 - val_acc: 0.7846\n",
      "Epoch 45/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4495 - acc: 0.7906 - val_loss: 0.4581 - val_acc: 0.7898\n",
      "Epoch 46/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4492 - acc: 0.7925 - val_loss: 0.4554 - val_acc: 0.7845\n",
      "Epoch 47/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4480 - acc: 0.7922 - val_loss: 0.4550 - val_acc: 0.7885\n",
      "Epoch 48/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4487 - acc: 0.7925 - val_loss: 0.4566 - val_acc: 0.7849\n",
      "Epoch 49/75\n",
      "80000/80000 [==============================] - 7s 94us/step - loss: 0.4477 - acc: 0.7929 - val_loss: 0.4684 - val_acc: 0.7915\n",
      "Epoch 50/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4466 - acc: 0.7944 - val_loss: 0.4631 - val_acc: 0.7909\n",
      "Epoch 51/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4463 - acc: 0.7934 - val_loss: 0.4983 - val_acc: 0.7736\n",
      "Epoch 52/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4461 - acc: 0.7937 - val_loss: 0.4544 - val_acc: 0.7908\n",
      "Epoch 53/75\n",
      "80000/80000 [==============================] - 8s 95us/step - loss: 0.4459 - acc: 0.7941 - val_loss: 0.4635 - val_acc: 0.7796\n",
      "Epoch 54/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4454 - acc: 0.7942 - val_loss: 0.4550 - val_acc: 0.7900\n",
      "Epoch 55/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4453 - acc: 0.7960 - val_loss: 0.4558 - val_acc: 0.7930\n",
      "Epoch 56/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4450 - acc: 0.7948 - val_loss: 0.4572 - val_acc: 0.7851\n",
      "Epoch 57/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4440 - acc: 0.7962 - val_loss: 0.4639 - val_acc: 0.7806\n",
      "Epoch 58/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4439 - acc: 0.7948 - val_loss: 0.4566 - val_acc: 0.7915\n",
      "Epoch 59/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4434 - acc: 0.7969 - val_loss: 0.4565 - val_acc: 0.7847\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 60/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4429 - acc: 0.7968 - val_loss: 0.4533 - val_acc: 0.7906\n",
      "Epoch 61/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4436 - acc: 0.7962 - val_loss: 0.4841 - val_acc: 0.7729\n",
      "Epoch 62/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4435 - acc: 0.7969 - val_loss: 0.4608 - val_acc: 0.7939\n",
      "Epoch 63/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4415 - acc: 0.7980 - val_loss: 0.4626 - val_acc: 0.7938\n",
      "Epoch 64/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4422 - acc: 0.7968 - val_loss: 0.4536 - val_acc: 0.7869\n",
      "Epoch 65/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4418 - acc: 0.7973 - val_loss: 0.4545 - val_acc: 0.7895\n",
      "Epoch 66/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4421 - acc: 0.7973 - val_loss: 0.4670 - val_acc: 0.7804\n",
      "Epoch 67/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4411 - acc: 0.7977 - val_loss: 0.4961 - val_acc: 0.7725\n",
      "Epoch 68/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4409 - acc: 0.7990 - val_loss: 0.4516 - val_acc: 0.7915\n",
      "Epoch 69/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4405 - acc: 0.8002 - val_loss: 0.4523 - val_acc: 0.7947\n",
      "Epoch 70/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4398 - acc: 0.7998 - val_loss: 0.4555 - val_acc: 0.7852\n",
      "Epoch 71/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4416 - acc: 0.7981 - val_loss: 0.5816 - val_acc: 0.7123\n",
      "Epoch 72/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4397 - acc: 0.7998 - val_loss: 0.4609 - val_acc: 0.7817\n",
      "Epoch 73/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4402 - acc: 0.7998 - val_loss: 0.4596 - val_acc: 0.7826\n",
      "Epoch 74/75\n",
      "80000/80000 [==============================] - 7s 92us/step - loss: 0.4391 - acc: 0.7999 - val_loss: 0.4607 - val_acc: 0.7934\n",
      "Epoch 75/75\n",
      "80000/80000 [==============================] - 7s 93us/step - loss: 0.4395 - acc: 0.7995 - val_loss: 0.4694 - val_acc: 0.7882\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten_1 (Flatten)          (None, 7500)              0         \n",
      "_________________________________________________________________\n",
      "dense_28 (Dense)             (None, 64)                480064    \n",
      "_________________________________________________________________\n",
      "dense_29 (Dense)             (None, 10)                650       \n",
      "=================================================================\n",
      "Total params: 480,714\n",
      "Trainable params: 480,714\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#Source: (fashion MNIST) https://www.tensorflow.org/tutorials/keras/basic_classification\n",
    "#Source: https://www.pyimagesearch.com/2018/09/10/keras-tutorial-how-to-get-started-with-keras-deep-learning-and-python/?fbclid=lwAR3WMQfZB513JosNLudYcv6DaaVwEv5pD6rQEMB4l_fDRxlwWfNG8gObmJU\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout\n",
    "from keras.layers import Conv2D\n",
    "from keras.layers import Flatten, Activation\n",
    "import keras.backend as K\n",
    "\n",
    "model=Sequential()\n",
    "\n",
    "model.add(Flatten(input_shape=(50, 50, 3)))\n",
    "model.add(Dense(64, activation='sigmoid'))\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "INIT_LR = 0.01\n",
    "EPOCHS = 75\n",
    "\n",
    "print(\"[INFO] training network...\")\n",
    "opt = SGD(lr=INIT_LR)\n",
    "\n",
    "#Source: (fashion MNIST) https://www.tensorflow.org/tutorials/keras/basic_classification\n",
    "model.compile(loss='sparse_categorical_crossentropy',\n",
    "              optimizer=opt,\n",
    "                 metrics=['accuracy'])\n",
    "# train the network\n",
    "H = model.fit(X_train, Y_train, validation_data = (X_test, Y_test), epochs=EPOCHS, batch_size=100)\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] evaluating network...\n",
      "20000/20000 [==============================] - 1s 70us/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x1e980710da0>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Source: https://www.pyimagesearch.com/2018/09/10/keras-tutorial-how-to-get-started-with-keras-deep-learning-and-python/?fbclid=lwAR3WMQfZB513JosNLudYcv6DaaVwEv5pD6rQEMB4l_fDRxlwWfNG8gObmJU\n",
    "\n",
    "# evaluate the network\n",
    "print(\"[INFO] evaluating network...\")\n",
    "score = model.evaluate(X_test, Y_test, batch_size=100)\n",
    "\n",
    "\n",
    "# plot the training loss and accuracy\n",
    "N = np.arange(0, EPOCHS)\n",
    "plt.style.use(\"ggplot\")\n",
    "plt.figure()\n",
    "plt.plot(N, H.history[\"loss\"], label=\"train_loss\")\n",
    "plt.plot(N, H.history[\"val_loss\"], label=\"val_loss\")\n",
    "plt.plot(N, H.history[\"acc\"], label=\"train_acc\")\n",
    "plt.plot(N, H.history[\"val_acc\"], label=\"val_acc\")\n",
    "plt.title(\"Training Loss and Accuracy\")\n",
    "plt.xlabel(\"Epoch #\")\n",
    "plt.ylabel(\"Loss/Accuracy\")\n",
    "plt.legend()\n"
   ]
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}