Image_Classifier/skin_cancer_classification_2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Image Classifier using Convolutions\n",
    "\n",
    "This notebook is an example notebook of using image data and classifying them into 2 or more classes.\n",
    "\n",
    "Classifying images into labels is a long-lasting problem in almost all vector of industry. \n",
    "\n",
    "Here is a notebook to do the same using neural network model to classify the images with cyclic learning rate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "('2.2.2',\n",
       " '1.10.0',\n",
       " '3.6.6 |Anaconda, Inc.| (default, Jun 28 2018, 11:27:44) [MSC v.1900 64 bit (AMD64)]')"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import keras, tensorflow, sys\n",
    "keras.__version__, tensorflow.__version__, sys.version"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Packages for processing/ loading image files\n",
    "from glob import glob \n",
    "import cv2\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import time\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.metrics import f1_score, precision_score, recall_score, roc_auc_score, accuracy_score\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.utils import class_weight\n",
    "\n",
    "# Packages for the model\n",
    "from keras.layers import LeakyReLU, Concatenate, concatenate, Lambda, UpSampling2D, Add, Input, Dense\n",
    "from keras.layers import Conv2D, MaxPooling2D, Activation, Dropout, Flatten, BatchNormalization\n",
    "from keras.models import Model,load_model\n",
    "from keras.optimizers import Adam\n",
    "import tensorflow as tf\n",
    "import keras.backend as K\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.utils.np_utils import to_categorical\n",
    "\n",
    "# Import the snapshot builder callback file locally which is used for dynamic LR.\n",
    "from snapshot import SnapshotCallbackBuilder\n",
    "\n",
    "# Used for training on 1st GPU in a multi-GPU environment.\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\" \n",
    "\n",
    "# Pacakge for viewing the images\n",
    "from matplotlib import pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Function to load and process Image\n",
    "\n",
    "Load image using opencv function imread(path, 1 for read as colour image).\n",
    "\n",
    "Cv2 loads the image with B channel first. We converte this to R channel first.\n",
    "\n",
    "Gaussian Blur is applied to image to have smoother transitions in image. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocess image for of top view training/ detection\n",
    "def load_and_process_image(img_file):\n",
    "    im = cv2.imread(img_file, 1)\n",
    "    im = cv2.cvtColor(im, cv2.COLOR_BGR2RGB)\n",
    "    im = cv2.GaussianBlur(im, (5,5), 0)\n",
    "    return im"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Dataset\n",
    "\n",
    "The dataset for this notebook is taken from -> https://www.kaggle.com/kmader/skin-cancer-mnist-ham10000 .\n",
    "\n",
    "The dataset contains images of 7 different types of skin problems. There is total of 10000 jpg images, accompanied with a csv file conating the labels for the images."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "imageid_path_dict = {os.path.splitext(os.path.basename(x))[0]: x\n",
    "                     for x in glob(\"./skin-cancer-mnist-ham10000/*part*/*.jpg\")}\n",
    "\n",
    "lesion_type_dict = {\n",
    "    'nv': 'Melanocytic nevi',\n",
    "    'mel': 'dermatofibroma',\n",
    "    'bkl': 'Benign keratosis-like lesions ',\n",
    "    'bcc': 'Basal cell carcinoma',\n",
    "    'akiec': 'Actinic keratoses',\n",
    "    'vasc': 'Vascular lesions',\n",
    "    'df': 'Dermatofibroma'\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>lesion_id</th>\n",
       "      <th>image_id</th>\n",
       "      <th>dx</th>\n",
       "      <th>dx_type</th>\n",
       "      <th>age</th>\n",
       "      <th>sex</th>\n",
       "      <th>localization</th>\n",
       "      <th>path</th>\n",
       "      <th>cell_type</th>\n",
       "      <th>cell_type_idx</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5831</th>\n",
       "      <td>HAM_0001092</td>\n",
       "      <td>ISIC_0031962</td>\n",
       "      <td>nv</td>\n",
       "      <td>follow_up</td>\n",
       "      <td>50.0</td>\n",
       "      <td>male</td>\n",
       "      <td>lower extremity</td>\n",
       "      <td>./skin-cancer-mnist-ham10000\\HAM10000_images_p...</td>\n",
       "      <td>Melanocytic nevi</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>473</th>\n",
       "      <td>HAM_0001057</td>\n",
       "      <td>ISIC_0032045</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>75.0</td>\n",
       "      <td>female</td>\n",
       "      <td>back</td>\n",
       "      <td>./skin-cancer-mnist-ham10000\\HAM10000_images_p...</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>HAM_0004373</td>\n",
       "      <td>ISIC_0029014</td>\n",
       "      <td>bkl</td>\n",
       "      <td>histo</td>\n",
       "      <td>70.0</td>\n",
       "      <td>female</td>\n",
       "      <td>face</td>\n",
       "      <td>./skin-cancer-mnist-ham10000\\HAM10000_images_p...</td>\n",
       "      <td>Benign keratosis-like lesions</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        lesion_id      image_id   dx    dx_type   age     sex  \\\n",
       "5831  HAM_0001092  ISIC_0031962   nv  follow_up  50.0    male   \n",
       "473   HAM_0001057  ISIC_0032045  bkl      histo  75.0  female   \n",
       "266   HAM_0004373  ISIC_0029014  bkl      histo  70.0  female   \n",
       "\n",
       "         localization                                               path  \\\n",
       "5831  lower extremity  ./skin-cancer-mnist-ham10000\\HAM10000_images_p...   \n",
       "473              back  ./skin-cancer-mnist-ham10000\\HAM10000_images_p...   \n",
       "266              face  ./skin-cancer-mnist-ham10000\\HAM10000_images_p...   \n",
       "\n",
       "                           cell_type  cell_type_idx  \n",
       "5831                Melanocytic nevi              4  \n",
       "473   Benign keratosis-like lesions               2  \n",
       "266   Benign keratosis-like lesions               2  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tile_df = pd.read_csv(\"./skin-cancer-mnist-ham10000/HAM10000_metadata.csv\")\n",
    "tile_df['path'] = tile_df['image_id'].map(imageid_path_dict.get)\n",
    "tile_df['cell_type'] = tile_df['dx'].map(lesion_type_dict.get) \n",
    "tile_df['cell_type_idx'] = pd.Categorical(tile_df['cell_type']).codes\n",
    "tile_df.sample(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10015, 10)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tile_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((10015, 450, 600, 3), (10015,))"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.array([load_and_process_image(x) for x in tile_df.path])\n",
    "y = np.array(tile_df['cell_type_idx'])\n",
    "X.shape, y.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This shows the distribution of data in each of the class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4    6705\n",
       "6    1113\n",
       "2    1099\n",
       "1     514\n",
       "0     327\n",
       "5     142\n",
       "3     115\n",
       "dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(y).value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Sample Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1ec9d89bc88>"
      ]
     },
     "execution_count": 9,
     "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": [
    "plt.imshow(X[49])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((9013, 450, 600, 3), (1002, 450, 600, 3))"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train, x_test, y_train, y_test = train_test_split(X, y, test_size=0.1, random_state=3)\n",
    "x_train.shape, x_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((8111, 450, 600, 3), (902, 450, 600, 3))"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train, x_val, y_train, y_val = train_test_split(x_train, y_train, test_size=0.1, random_state=3)\n",
    "x_train.shape, x_val.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Computing class weights to weight the loss function according to the number of instances in each class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_weights = class_weight.compute_class_weight('balanced',\n",
    "                                                 np.unique(y_train),\n",
    "                                                 y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Converting y_train to one hot encoded vectors, which is required for the training of neural network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train = to_categorical(y_train, len(lesion_type_dict))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model\n",
    "\n",
    "A Convolution based model is used here to classify the images.\n",
    "\n",
    "There are 6 convolution lyers with kernel size of (1,1) and number of kernels increases from 16 to 512.\n",
    "\n",
    "Each Convolution layer is followed by batch normaliztion layer and Leaky ReLU activation.\n",
    "\n",
    "Max pooling of features after each convolution block is performed to reduce the size of features by 2.\n",
    "\n",
    "Finally dense layers are used to classify the features found from the convolution layer.\n",
    "\n",
    "Categorical Cross entropy is used as the loss function and adam is the optimizer with cosine annealed learning rate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 450, 600, 3)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 450, 600, 16)      64        \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 450, 600, 16)      64        \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 450, 600, 16)      0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 225, 300, 16)      0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 225, 300, 32)      544       \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 225, 300, 32)      128       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_2 (LeakyReLU)    (None, 225, 300, 32)      0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 112, 150, 32)      0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 112, 150, 64)      2112      \n",
      "_________________________________________________________________\n",
      "batch_normalization_3 (Batch (None, 112, 150, 64)      256       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_3 (LeakyReLU)    (None, 112, 150, 64)      0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 56, 75, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 56, 75, 128)       8320      \n",
      "_________________________________________________________________\n",
      "batch_normalization_4 (Batch (None, 56, 75, 128)       512       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_4 (LeakyReLU)    (None, 56, 75, 128)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 28, 37, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 28, 37, 256)       33024     \n",
      "_________________________________________________________________\n",
      "batch_normalization_5 (Batch (None, 28, 37, 256)       1024      \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_5 (LeakyReLU)    (None, 28, 37, 256)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 14, 18, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_6 (Conv2D)            (None, 14, 18, 512)       131584    \n",
      "_________________________________________________________________\n",
      "batch_normalization_6 (Batch (None, 14, 18, 512)       2048      \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_6 (LeakyReLU)    (None, 14, 18, 512)       0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_6 (MaxPooling2 (None, 7, 9, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 32256)             0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 64)                2064448   \n",
      "_________________________________________________________________\n",
      "batch_normalization_7 (Batch (None, 64)                256       \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 16)                1040      \n",
      "_________________________________________________________________\n",
      "batch_normalization_8 (Batch (None, 16)                64        \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 16)                0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 7)                 119       \n",
      "=================================================================\n",
      "Total params: 2,245,607\n",
      "Trainable params: 2,243,431\n",
      "Non-trainable params: 2,176\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Initalize the Inception based model\n",
    "\n",
    "inp = Input(X[0].shape)\n",
    "x = Lambda(lambda x: x/255)(inp)\n",
    "\n",
    "x = Conv2D(16,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(inp)\n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2), strides=2)(x)  #225,300\n",
    "\n",
    "\n",
    "x = Conv2D(32,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(x)   \n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2), strides=2)(x) #112, 150\n",
    "\n",
    "\n",
    "x = Conv2D(64,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(x)    \n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2), strides=2)(x) #56, 75\n",
    "\n",
    "\n",
    "x = Conv2D(128,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(x)  \n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2))(x) #28, 37\n",
    "\n",
    "\n",
    "x = Conv2D(256,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(x)    \n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2))(x) #14, 18\n",
    "\n",
    "\n",
    "x = Conv2D(512,(1,1),padding='same',kernel_initializer=\"glorot_normal\")(x)    \n",
    "x = BatchNormalization(axis = 3)(x)\n",
    "x = LeakyReLU()(x)\n",
    "x = MaxPooling2D((2, 2))(x) #7, 9\n",
    "\n",
    "x = Flatten()(x)\n",
    "\n",
    "x = Dense(64, activation=\"relu\", kernel_initializer=\"glorot_normal\")(x)\n",
    "x = BatchNormalization()(x)\n",
    "x = Dropout(0.2)(x)\n",
    "\n",
    "x = Dense(16, activation=\"relu\", kernel_initializer=\"glorot_normal\")(x)\n",
    "x = BatchNormalization()(x)\n",
    "x = Dropout(0.2)(x)\n",
    "\n",
    "pred = Dense(len(lesion_type_dict), activation=\"softmax\", kernel_initializer=\"glorot_normal\")(x)\n",
    "\n",
    "model = Model(inp, pred)\n",
    "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
    "model.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Initializing callback for tensorboard. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.callbacks import TensorBoard,LearningRateScheduler\n",
    "tb = TensorBoard(log_dir='./logs/skin_cancer_snapshot', histogram_freq=0, batch_size=16, write_graph=True, write_grads=True, write_images=True,\n",
    "                 embeddings_freq=0, embeddings_layer_names=None, embeddings_metadata=None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_val = to_categorical(y_val, len(lesion_type_dict))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Snapshot callback\n",
    "\n",
    "In neural networks, the model weights can easily get stuck in a local minima if the model is trained with low learning rate throughout. In Adam the momentum does help to get out of a local minima but the model can still be stuck in a local minima.\n",
    "\n",
    "To mitigate this problem, learning rate is increased to a high value (alpha_zero) after certain number of epochs (nb_epoch/M). After each cycle the learning rate is increased so that the model weights now come out of the local minima and go towards another, perhaps, better minima.\n",
    "\n",
    "Generally because of the stochasticity in the training, the same model trained multiple times produce different results on the same test set. To mitigate this different snapshots of the model taken during training by above process is used to evaluate the results. In this notebook only the last model found in the training is used for the results, but according to the paper of [snapshot ensemble](https://arxiv.org/pdf/1704.00109.pdf) and my other experiments averaging results from mutiple snapshots will improve the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Model parameters\n",
    "# Number of model snapshots\n",
    "M = 6\n",
    "# Number of epochs\n",
    "nb_epoch = T = 78\n",
    "# Learning rate to be set at the start of a snapshot cycle\n",
    "alpha_zero = 1e-3\n",
    "\n",
    "snapshot = SnapshotCallbackBuilder(T, M, alpha_zero)\n",
    "timestr = time.strftime(\"%Y-%m-%d_%H-%M-%S\")\n",
    "model_prefix = './conv_snpsht{}'.format(timestr)\n",
    "\n",
    "# Train model\n",
    "callbacks = [tb] + snapshot.get_callbacks(model_prefix = model_prefix)\n",
    "model.fit(x_train, y_train, batch_size=16, epochs=nb_epoch, callbacks=callbacks, verbose=0,\n",
    "          validation_data=[x_val, y_val], class_weight=class_weights)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred = model.predict(x_test)\n",
    "pred = np.argmax(pred, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion Matrix:\n",
      " [[  7  12   5   1   6   0   5]\n",
      " [  2  31   9   0  11   0   7]\n",
      " [  2   3  53   0  41   0  12]\n",
      " [  1   2   1   4   2   0   0]\n",
      " [  1   3  18   0 620   0  12]\n",
      " [  0   0   0   0   0  13   1]\n",
      " [  2   0  14   1  54   0  46]]\n",
      "f1_score: 0.5931103930878011 \n",
      " precision_score: 0.667154283794211 \n",
      " recall_score: 0.5511906632483304 \n",
      " accuracy_score: 0.7724550898203593\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix,f1_score, precision_score, recall_score, roc_auc_score, accuracy_score\n",
    "print(\"Confusion Matrix:\\n\", confusion_matrix(y_test, pred))\n",
    "print(\"f1_score:\",f1_score(y_test, pred, average=\"macro\"), \"\\n\",\n",
    "      \"precision_score:\",precision_score(y_test, pred, average=\"macro\"), \"\\n\",\n",
    "      \"recall_score:\",recall_score(y_test, pred, average=\"macro\"), \"\\n\",\n",
    "      \"accuracy_score:\",accuracy_score(y_test, pred))"
   ]
  },
  {
   "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
}