Convolutional Neural Networks in Python with Keras - Deep Learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from keras.utils import to_categorical\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.model_selection import train_test_split\n",
    "%matplotlib inline\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras.datasets import fashion_mnist\n",
    "(train_X,train_Y), (test_X,test_Y) = fashion_mnist.load_data()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training data shape :  (60000, 28, 28) (60000,)\n",
      "Testing data shape :  (10000, 28, 28) (10000,)\n",
      "Total number of outputs :  10\n",
      "Output classes :  [0 1 2 3 4 5 6 7 8 9]\n"
     ]
    }
   ],
   "source": [
    "print('Training data shape : ', train_X.shape, train_Y.shape)\n",
    "\n",
    "print('Testing data shape : ', test_X.shape, test_Y.shape)\n",
    "\n",
    "# Find the unique numbers from the train labels - there are 10 [0..9]\n",
    "classes = np.unique(train_Y)\n",
    "nClasses = len(classes)\n",
    "print('Total number of outputs : ', nClasses)\n",
    "print('Output classes : ', classes)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Ground Truth : 9')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATwAAACuCAYAAACr3LH6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e602b72518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=[5,5])\n",
    "\n",
    "# Display the first image in training data\n",
    "plt.subplot(121)\n",
    "plt.imshow(train_X[0,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(train_Y[0]))\n",
    "\n",
    "# Display the first image in testing data\n",
    "plt.subplot(122)\n",
    "plt.imshow(test_X[0,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(test_Y[0]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "As you could see in the above plot, the images are grayscale images have pixel values that range from 0 to 255. \n",
    "Also, these images have a dimension of 28 x 28. As a result, you'll need to preprocess the data before you feed it into the model. \n",
    "\n",
    "As a first step, convert each 28 x 28 image of the train and test set into a matrix of size 28 x 28 x 1 which is fed into \n",
    "the network.\n",
    "\"\"\"\n",
    "train_X = train_X.reshape(-1, 28,28, 1)\n",
    "test_X = test_X.reshape(-1, 28,28, 1)\n",
    "train_X.shape, test_X.shape\n",
    "\"\"\"\n",
    "The data right now is in an int8 format, so before you feed it into the network you need to convert its type to float32, \n",
    "and you also have to rescale the pixel values in range 0 - 1 inclusive. So let's do that!\n",
    "\"\"\"\n",
    "train_X = train_X.astype('float32')\n",
    "test_X = test_X.astype('float32')\n",
    "train_X = train_X / 255.\n",
    "test_X = test_X / 255.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original label: 9\n",
      "After conversion to one-hot: [0. 0. 0. 0. 0. 0. 0. 0. 0. 1.]\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Now you need to convert the class labels into a one-hot encoding vector.\n",
    "In one-hot encoding, you convert the categorical data into a vector of numbers. The reason why you convert the categorical \n",
    "data in one hot encoding is that machine learning algorithms cannot work with categorical data directly. You generate \n",
    "one boolean column for each category or class. Only one of these columns could take on the value 1 for each sample. \n",
    "Hence, the term one-hot encoding.\n",
    "\n",
    "For your problem statement, the one hot encoding will be a row vector, and for each image, it will have a dimension of 1 x 10. \n",
    "The important thing to note here is that the vector consists of all zeros except for the class that it represents, \n",
    "and for that, it is 1. For example, the ankle boot image that you plotted above has a label of 9, so for all the ankle \n",
    "boot images, the one hot encoding vector would be [0 0 0 0 0 0 0 0 1 0].\n",
    "\n",
    "So let's convert the training and testing labels into one-hot encoding vectors:\n",
    "\"\"\"\n",
    "# Change the labels from categorical to one-hot encoding\n",
    "train_Y_one_hot = to_categorical(train_Y)\n",
    "test_Y_one_hot = to_categorical(test_Y)\n",
    "\n",
    "# Display the change for category label using one-hot encoding\n",
    "print('Original label:', train_Y[0])\n",
    "print('After conversion to one-hot:', train_Y_one_hot[0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((48000, 28, 28, 1), (12000, 28, 28, 1), (48000, 10), (12000, 10))"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "This last step is a crucial one. In machine learning or any data specific task, you should partition the data correctly. \n",
    "For the model to generalize well, you split the training data into two parts, one designed for training and another \n",
    "one for validation. In this case, you will train the model on 80\\% of the training data and validate it on 20\\% of \n",
    "the remaining training data. This will also help to reduce overfitting since you will be validating the model on the \n",
    "data it would not have seen in training phase, which will help in boosting the test performance.\n",
    "\"\"\"\n",
    "\n",
    "train_X,valid_X,train_label,valid_label = train_test_split(train_X, train_Y_one_hot, test_size=0.2, random_state=13)\n",
    "train_X.shape,valid_X.shape,train_label.shape,valid_label.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "The images are of size 28 x 28. You convert the image matrix to an array, rescale it between 0 and 1, reshape it so that \n",
    "it's of size 28 x 28 x 1, and feed this as an input to the network.\n",
    "\n",
    "You'll use three convolutional layers:\n",
    "\n",
    "The first layer will have 32-3 x 3 filters,\n",
    "The second layer will have 64-3 x 3 filters and\n",
    "The third layer will have 128-3 x 3 filters.\n",
    "\"\"\"\n",
    "#import modules to train the model\n",
    "import keras\n",
    "from keras.models import Sequential,Input,Model\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.advanced_activations import LeakyReLU\n",
    "\n",
    "\"\"\"\n",
    "You will use a batch size of 64 using a higher batch size of 128 or 256 is also preferable it all depends on the memory. \n",
    "It contributes massively to determining the learning parameters and affects the prediction accuracy. You will train \n",
    "the network for 20 epochs.\n",
    "\"\"\"\n",
    "\n",
    "batch_size = 64\n",
    "epochs = 20\n",
    "num_classes = 10\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fashion_model = Sequential()\n",
    "fashion_model.add(Conv2D(32, kernel_size=(3, 3),activation='linear',input_shape=(28,28,1),padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D((2, 2),padding='same'))\n",
    "fashion_model.add(Conv2D(64, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Conv2D(128, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))                  \n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Flatten())\n",
    "fashion_model.add(Dense(128, activation='linear'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))                  \n",
    "fashion_model.add(Dense(num_classes, activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 28, 28, 32)        320       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 28, 28, 32)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 14, 14, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 14, 14, 64)        18496     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_2 (LeakyReLU)    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 7, 7, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 7, 7, 128)         73856     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_3 (LeakyReLU)    (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 128)               262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_4 (LeakyReLU)    (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 356,234\n",
      "Trainable params: 356,234\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "After the model is created, you compile it using the Adam optimizer, one of the most popular optimization algorithms. \n",
    "You can read more about this optimizer here. Additionally, you specify the loss type which is categorical cross entropy \n",
    "which is used for multi-class classification, you can also use binary cross-entropy as the loss function. \n",
    "Lastly, you specify the metrics as accuracy which you want to analyze while the model is training.\n",
    "\"\"\"\n",
    "fashion_model.compile(loss=keras.losses.categorical_crossentropy, optimizer=keras.optimizers.Adam(),metrics=['accuracy'])\n",
    "\n",
    "\"\"\"\n",
    "Let's visualize the layers that you created in the above step by using the summary function. This will show some \n",
    "parameters (weights and biases) in each layer and also the total parameters in your model.\n",
    "\"\"\"\n",
    "fashion_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 48000 samples, validate on 12000 samples\n",
      "Epoch 1/20\n",
      "48000/48000 [==============================] - 211s 4ms/step - loss: 0.4585 - acc: 0.8330 - val_loss: 0.3137 - val_acc: 0.8861\n",
      "Epoch 2/20\n",
      "48000/48000 [==============================] - 227s 5ms/step - loss: 0.2792 - acc: 0.8977 - val_loss: 0.2956 - val_acc: 0.8924\n",
      "Epoch 3/20\n",
      "48000/48000 [==============================] - 208s 4ms/step - loss: 0.2343 - acc: 0.9135 - val_loss: 0.2577 - val_acc: 0.9059\n",
      "Epoch 4/20\n",
      "48000/48000 [==============================] - 211s 4ms/step - loss: 0.2045 - acc: 0.9242 - val_loss: 0.2386 - val_acc: 0.9170\n",
      "Epoch 5/20\n",
      "48000/48000 [==============================] - 196s 4ms/step - loss: 0.1794 - acc: 0.9337 - val_loss: 0.2194 - val_acc: 0.9225\n",
      "Epoch 6/20\n",
      "48000/48000 [==============================] - 194s 4ms/step - loss: 0.1578 - acc: 0.9411 - val_loss: 0.2237 - val_acc: 0.9213\n",
      "Epoch 7/20\n",
      "48000/48000 [==============================] - 180s 4ms/step - loss: 0.1358 - acc: 0.9493 - val_loss: 0.2406 - val_acc: 0.9178\n",
      "Epoch 8/20\n",
      "48000/48000 [==============================] - 182s 4ms/step - loss: 0.1183 - acc: 0.9564 - val_loss: 0.2737 - val_acc: 0.9143\n",
      "Epoch 9/20\n",
      "48000/48000 [==============================] - 218s 5ms/step - loss: 0.0996 - acc: 0.9620 - val_loss: 0.2506 - val_acc: 0.9200\n",
      "Epoch 10/20\n",
      "48000/48000 [==============================] - 221s 5ms/step - loss: 0.0872 - acc: 0.9668 - val_loss: 0.2678 - val_acc: 0.9217\n",
      "Epoch 11/20\n",
      "48000/48000 [==============================] - 191s 4ms/step - loss: 0.0705 - acc: 0.9736 - val_loss: 0.3146 - val_acc: 0.9177\n",
      "Epoch 12/20\n",
      "48000/48000 [==============================] - 181s 4ms/step - loss: 0.0630 - acc: 0.9761 - val_loss: 0.3046 - val_acc: 0.9180\n",
      "Epoch 13/20\n",
      "48000/48000 [==============================] - 189s 4ms/step - loss: 0.0531 - acc: 0.9802 - val_loss: 0.3446 - val_acc: 0.9215\n",
      "Epoch 14/20\n",
      "48000/48000 [==============================] - 217s 5ms/step - loss: 0.0493 - acc: 0.9815 - val_loss: 0.3711 - val_acc: 0.9173\n",
      "Epoch 15/20\n",
      "48000/48000 [==============================] - 213s 4ms/step - loss: 0.0400 - acc: 0.9848 - val_loss: 0.3745 - val_acc: 0.9193\n",
      "Epoch 16/20\n",
      "48000/48000 [==============================] - 191s 4ms/step - loss: 0.0396 - acc: 0.9848 - val_loss: 0.4027 - val_acc: 0.9220\n",
      "Epoch 17/20\n",
      "48000/48000 [==============================] - 191s 4ms/step - loss: 0.0359 - acc: 0.9864 - val_loss: 0.4042 - val_acc: 0.9176\n",
      "Epoch 18/20\n",
      "48000/48000 [==============================] - 217s 5ms/step - loss: 0.0341 - acc: 0.9878 - val_loss: 0.4207 - val_acc: 0.9201\n",
      "Epoch 19/20\n",
      "48000/48000 [==============================] - 222s 5ms/step - loss: 0.0320 - acc: 0.9881 - val_loss: 0.4474 - val_acc: 0.9182\n",
      "Epoch 20/20\n",
      "48000/48000 [==============================] - 194s 4ms/step - loss: 0.0318 - acc: 0.9887 - val_loss: 0.4547 - val_acc: 0.9198\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "It's finally time to train the model with Keras' fit() function! The model trains for 20 epochs. \n",
    "The fit() function will return a history object; By storying the result of this function in fashion_train, \n",
    "you can use it later to plot the accuracy and loss function plots between training and validation which \n",
    "will help you to analyze your model's performance visually.\n",
    "\"\"\"\n",
    "fashion_train = fashion_model.fit(train_X, train_label, batch_size=batch_size,epochs=epochs,verbose=1,validation_data=(valid_X, valid_label))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.4860444114053622\n",
      "Test accuracy: 0.9136\n"
     ]
    }
   ],
   "source": [
    "test_eval = fashion_model.evaluate(test_X, test_Y_one_hot, verbose=0)\n",
    "print('Test loss:', test_eval[0])\n",
    "print('Test accuracy:', test_eval[1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGJNALOkbY0wCsaRvjDEJ5P8DFlHYIDKN830AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e601426a20>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e60170f4a8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "\"\"\"\n",
    "The test accuracy looks impressive. It turns out that your classifier does better than the benchmark that was reported here, \n",
    "which is an SVM classifier with mean accuracy of 0.897. Also, the model does well compared to some of the deep learning \n",
    "models mentioned on the GitHub profile of the creators of fashion-MNIST dataset.\n",
    "However, you saw that the model looked like it was overfitting. Are these results really all that good?\n",
    "Let's put your model evaluation into perspective and plot the accuracy and loss plots between training and validation data:\n",
    "\"\"\"\n",
    "accuracy = fashion_train.history['acc']\n",
    "val_accuracy = fashion_train.history['val_acc']\n",
    "loss = fashion_train.history['loss']\n",
    "val_loss = fashion_train.history['val_loss']\n",
    "epochs = range(len(accuracy))\n",
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(list, 20, 0.8330416666666667)"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(accuracy),len(accuracy),accuracy[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "You can add a dropout layer to overcome the problem of overfitting to some extent. Dropout randomly turns off a fraction of \n",
    "neurons during the training process, reducing the dependency on the training set by some amount. How many fractions of neurons \n",
    "you want to turn off is decided by a hyperparameter, which can be tuned accordingly. This way, turning off some neurons will \n",
    "not allow the network to memorize the training data since not all the neurons will be active at the same time and the \n",
    "inactive neurons will not be able to learn anything.\n",
    "\n",
    "So let's create, compile and train the network again but this time with dropout. And run it for 20 epochs with a batch \n",
    "size of 64.\n",
    "\"\"\"\n",
    "batch_size = 64\n",
    "epochs = 20\n",
    "num_classes = 10\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_4 (Conv2D)            (None, 28, 28, 32)        320       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_5 (LeakyReLU)    (None, 28, 28, 32)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 14, 14, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 14, 14, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 14, 14, 64)        18496     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_6 (LeakyReLU)    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 7, 7, 64)          0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 7, 7, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_6 (Conv2D)            (None, 7, 7, 128)         73856     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_7 (LeakyReLU)    (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "max_pooling2d_6 (MaxPooling2 (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 4, 4, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 2048)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 128)               262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_8 (LeakyReLU)    (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 356,234\n",
      "Trainable params: 356,234\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "fashion_model = Sequential()\n",
    "fashion_model.add(Conv2D(32, kernel_size=(3, 3),activation='linear',padding='same',input_shape=(28,28,1)))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D((2, 2),padding='same'))\n",
    "fashion_model.add(Dropout(0.25))\n",
    "fashion_model.add(Conv2D(64, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))\n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Dropout(0.25))\n",
    "fashion_model.add(Conv2D(128, (3, 3), activation='linear',padding='same'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))                  \n",
    "fashion_model.add(MaxPooling2D(pool_size=(2, 2),padding='same'))\n",
    "fashion_model.add(Dropout(0.4))\n",
    "fashion_model.add(Flatten())\n",
    "fashion_model.add(Dense(128, activation='linear'))\n",
    "fashion_model.add(LeakyReLU(alpha=0.1))           \n",
    "fashion_model.add(Dropout(0.3))\n",
    "fashion_model.add(Dense(num_classes, activation='softmax'))\n",
    "fashion_model.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 48000 samples, validate on 12000 samples\n",
      "Epoch 1/20\n",
      "48000/48000 [==============================] - 208s 4ms/step - loss: 0.6090 - acc: 0.7730 - val_loss: 0.3766 - val_acc: 0.8592\n",
      "Epoch 2/20\n",
      "48000/48000 [==============================] - 201s 4ms/step - loss: 0.3779 - acc: 0.8589 - val_loss: 0.3124 - val_acc: 0.8822\n",
      "Epoch 3/20\n",
      "48000/48000 [==============================] - 200s 4ms/step - loss: 0.3281 - acc: 0.8786 - val_loss: 0.2781 - val_acc: 0.8985\n",
      "Epoch 4/20\n",
      "48000/48000 [==============================] - 200s 4ms/step - loss: 0.2971 - acc: 0.8892 - val_loss: 0.2581 - val_acc: 0.9048\n",
      "Epoch 5/20\n",
      "48000/48000 [==============================] - 205s 4ms/step - loss: 0.2789 - acc: 0.8956 - val_loss: 0.2561 - val_acc: 0.9088\n",
      "Epoch 6/20\n",
      "48000/48000 [==============================] - 236s 5ms/step - loss: 0.2673 - acc: 0.9010 - val_loss: 0.2409 - val_acc: 0.9112\n",
      "Epoch 7/20\n",
      "48000/48000 [==============================] - 232s 5ms/step - loss: 0.2562 - acc: 0.9051 - val_loss: 0.2336 - val_acc: 0.9160\n",
      "Epoch 8/20\n",
      "48000/48000 [==============================] - 217s 5ms/step - loss: 0.2483 - acc: 0.9071 - val_loss: 0.2420 - val_acc: 0.9097\n",
      "Epoch 9/20\n",
      "48000/48000 [==============================] - 228s 5ms/step - loss: 0.2430 - acc: 0.9086 - val_loss: 0.2238 - val_acc: 0.9187\n",
      "Epoch 10/20\n",
      "48000/48000 [==============================] - 214s 4ms/step - loss: 0.2327 - acc: 0.9137 - val_loss: 0.2381 - val_acc: 0.9149\n",
      "Epoch 11/20\n",
      "48000/48000 [==============================] - 207s 4ms/step - loss: 0.2294 - acc: 0.9139 - val_loss: 0.2213 - val_acc: 0.9168\n",
      "Epoch 12/20\n",
      "48000/48000 [==============================] - 218s 5ms/step - loss: 0.2219 - acc: 0.9156 - val_loss: 0.2167 - val_acc: 0.9202\n",
      "Epoch 13/20\n",
      "48000/48000 [==============================] - 220s 5ms/step - loss: 0.2185 - acc: 0.9173 - val_loss: 0.2144 - val_acc: 0.9213\n",
      "Epoch 14/20\n",
      "48000/48000 [==============================] - 207s 4ms/step - loss: 0.2167 - acc: 0.9185 - val_loss: 0.2135 - val_acc: 0.9248\n",
      "Epoch 15/20\n",
      "48000/48000 [==============================] - 206s 4ms/step - loss: 0.2130 - acc: 0.9191 - val_loss: 0.2098 - val_acc: 0.9245\n",
      "Epoch 16/20\n",
      "48000/48000 [==============================] - 207s 4ms/step - loss: 0.2088 - acc: 0.9210 - val_loss: 0.2161 - val_acc: 0.9217\n",
      "Epoch 17/20\n",
      "48000/48000 [==============================] - 206s 4ms/step - loss: 0.2067 - acc: 0.9219 - val_loss: 0.2111 - val_acc: 0.9253\n",
      "Epoch 18/20\n",
      "48000/48000 [==============================] - 202s 4ms/step - loss: 0.2008 - acc: 0.9227 - val_loss: 0.2141 - val_acc: 0.9258\n",
      "Epoch 19/20\n",
      "48000/48000 [==============================] - 198s 4ms/step - loss: 0.2011 - acc: 0.9237 - val_loss: 0.2065 - val_acc: 0.9263\n",
      "Epoch 20/20\n",
      "48000/48000 [==============================] - 210s 4ms/step - loss: 0.1970 - acc: 0.9254 - val_loss: 0.2051 - val_acc: 0.9262\n"
     ]
    }
   ],
   "source": [
    "fashion_model.compile(loss=keras.losses.categorical_crossentropy, optimizer=keras.optimizers.Adam(),metrics=['accuracy'])\n",
    "fashion_train_dropout = fashion_model.fit(train_X, train_label, batch_size=batch_size,epochs=epochs,verbose=1,validation_data=(valid_X, valid_label))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 14s 1ms/step\n",
      "Test loss: 0.2218779006898403\n",
      "Test accuracy: 0.9218\n"
     ]
    }
   ],
   "source": [
    "#\n",
    "# Save the model so we don't have to retrain it each time\n",
    "#\n",
    "fashion_model.save(\"fashion_model_dropout.h5py\")\n",
    "\n",
    "\"\"\"\n",
    "To reload the model use the following command\n",
    "myModel = keras.models.load_model(filepath)\n",
    "\"\"\"\n",
    "\n",
    "\"\"\"\n",
    "Model Evaluation on the Test Set\n",
    "Finally, let's also evaluate your new model and see how it performs!\n",
    "\"\"\"\n",
    "\n",
    "test_eval = fashion_model.evaluate(test_X, test_Y_one_hot, verbose=1)\n",
    "print('Test loss:', test_eval[0])\n",
    "print('Test accuracy:', test_eval[1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEICAYAAACzliQjAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xl4FFXW+PHvYTMg+yLKHpVRFhMI\nIYgiICAvMAqjgoDgiIioM7j9dJQZmJEXZV73ddRXxhFHjSDqq6AjIiKKCiphSRCRRWSJLIYtgAgY\nOL8/bnVomu6kk+6kk/T5PE89XcutqlOVzunbt6pviapijDEmPlSKdQDGGGNKjyV9Y4yJI5b0jTEm\njljSN8aYOGJJ3xhj4oglfWOMiSOW9OOQiFQWkQMi0iKaZWNJRM4WkajffywifURko9/0GhG5KJyy\nxdjXCyLyl+Kub0w4qsQ6AFM4ETngN1kDOAwc9aZvVNX0omxPVY8CNaNdNh6o6jnR2I6IjAFGqmpP\nv22Pica2jSmIJf1yQFXzk65Xkxyjqh+FKi8iVVQ1rzRiM6Yw9n4sW6x5pwIQkftF5HURmS4i+4GR\nItJVRL4Ukb0isk1EnhKRql75KiKiItLKm37VWz5HRPaLyGIRSSxqWW95fxFZKyK5IvK0iHwhIqNC\nxB1OjDeKyHoR2SMiT/mtW1lEHheRXSLyPdCvgPMzUURmBMx7RkQe88bHiMhq73i+92rhobaVLSI9\nvfEaIvKKF9sqoFOQ/W7wtrtKRAZ6888D/gFc5DWd7fQ7t5P81r/JO/ZdIvKOiJwRzrkpynn2xSMi\nH4nIbhHZLiJ3++3nr9452SciGSLSJFhTmoh87vs7e+dzobef3cBEEWktIgu8Y9npnbc6fuu39I4x\nx1v+pIgkeDG38St3hogcFJEGoY7XFEJVbShHA7AR6BMw737gCHAZ7oO8OtAZ6IL7NncmsBYY55Wv\nAijQypt+FdgJpAJVgdeBV4tR9jRgPzDIW/b/gF+BUSGOJZwYZwF1gFbAbt+xA+OAVUAzoAGw0L2d\ng+7nTOAAcKrftn8CUr3py7wyAvQCfgGSvGV9gI1+28oGenrjjwCfAPWAlsC3AWWvAs7w/iZXezE0\n9paNAT4JiPNVYJI33teLsQOQADwLfBzOuSniea4D7ABuA04BagNp3rI/A5lAa+8YOgD1gbMDzzXw\nue/v7B1bHnAzUBn3fvwN0Buo5r1PvgAe8Tueb7zzeapX/kJv2VRgit9+7gTejvX/YXkeYh6ADUX8\ng4VO+h8Xst5dwBveeLBE/r9+ZQcC3xSj7GjgM79lAmwjRNIPM8bz/Zb/H3CXN74Q18zlWzYgMBEF\nbPtL4GpvvD+wtoCy7wF/9MYLSvqb/f8WwB/8ywbZ7jfAb73xwpL+v4G/+y2rjbuO06ywc1PE83wN\nkBGi3Pe+eAPmh5P0NxQSw2BgiTd+EbAdqByk3IXAD4B40yuAK6L9fxVPgzXvVBxb/CdE5FwR+Y/3\ndX0fMBloWMD62/3GD1LwxdtQZZv4x6HuvzQ71EbCjDGsfQGbCogX4DVguDd+NZB/8VtELhWRr7zm\njb24WnZB58rnjIJiEJFRIpLpNVHsBc4Nc7vgji9/e6q6D9gDNPUrE9bfrJDz3BxYHyKG5rjEXxyB\n78fTRWSmiPzoxfBSQAwb1d00cAJV/QL3raGbiLQHWgD/KWZMBmvTr0gCb1d8HlezPFtVawN/w9W8\nS9I2XE0UABERTkxSgSKJcRsuWfgUdkvp60AfEWmGa356zYuxOvAm8D+4ppe6wIdhxrE9VAwicibw\nHK6Jo4G33e/8tlvY7aVbcU1Gvu3VwjUj/RhGXIEKOs9bgLNCrBdq2c9eTDX85p0eUCbw+B7E3XV2\nnhfDqIAYWopI5RBxvAyMxH0rmamqh0OUM2GwpF9x1QJygZ+9C2E3lsI+3wNSROQyEamCayduVEIx\nzgRuF5Gm3kW9ewoqrKo7cE0Q04A1qrrOW3QKrp05BzgqIpfi2p7DjeEvIlJX3O8Yxvktq4lLfDm4\nz78xuJq+zw6gmf8F1QDTgetFJElETsF9KH2mqiG/ORWgoPM8G2ghIuNEpJqI1BaRNG/ZC8D9InKW\nOB1EpD7uw2477oaByiIyFr8PqAJi+BnIFZHmuCYmn8XALuDv4i6OVxeRC/2Wv4JrDroa9wFgImBJ\nv+K6E7gWd2H1eVxNt0R5iXUo8Bjun/gsYDmuhhftGJ8D5gMrgSW42nphXsO10b/mF/Ne4A7gbdzF\n0MG4D69w3Iv7xrERmINfQlLVLOAp4GuvzLnAV37rzgPWATtExL+Zxrf+B7hmmLe99VsAI8KMK1DI\n86yqucAlwJW4C8drgR7e4oeBd3DneR/uomqC12x3A/AX3EX9swOOLZh7gTTch89s4C2/GPKAS4E2\nuFr/Ztzfwbd8I+7vfERVFxXx2E0A38URY6LO+7q+FRisqp/FOh5TfonIy7iLw5NiHUt5Zz/OMlEl\nIv1wX9cP4W75y8PVdo0pFu/6yCDgvFjHUhFY846Jtm7ABtzX/n7A7+zCmykuEfkf3G8F/q6qm2Md\nT0VgzTvGGBNHrKZvjDFxpMy16Tds2FBbtWoV6zCMMaZcWbp06U5VLegWaaAMJv1WrVqRkZER6zCM\nMaZcEZHCfpUOWPOOMcbEFUv6xhgTRyzpG2NMHLGkb4wxccSSvjHGxBFL+sYYE0cs6RtjTIylp0Or\nVlCpkntNTy9sjeIrc/fpG2NMrKjCr7/CL7/AoUMnvoYzLy/PJe5KlUDk+HhB00uXwmuvuf0CbNoE\nY8e68RHF7Uy7AJb0jTEVQnY2bNgABw6cOOzff/K8UMPBg3DsWKyPxMUxYYIlfWNMCdi3Dz7+GD78\nEHJy4JRTjg/Vqp04XdiyVq3grLNcLbak/forLFoE77/vhm++CV22ShWoVQtq1jxxaN78xOkaNaB6\ndTckJJz4Gmo8IQFmzYLJk2HLFrfN+++H4cPdN4djx44PwaYbhnhq8uYS6lPUkr4xcUYVsrJgzhz4\n4AP44gvXLOFLgocPnzwcORL+9hs1ggsugAsvdK+dOrnEGA1bt7qY338f5s1zH1hVq0L37jBqFHTo\nEDy5V6tW8HbT013NevNmaNECpkwJv5adng633upq5+C2cdNNrukmnG20bOmadAK1KOypz8VU5rpW\nTk1NVet7x5jo2r3bJckPPnDDdu8BjcnJ0K8f9O8PXbuGTo6qLvEfORL8Q+HwYdemvWaN+xBZtAjW\nr3frVqvmEr//B0HjxuHFnZcHX33lkvycObB8uZvftCkMGOCG3r1h9uzIkvbYsceTNrga/9Sp4W2j\nVavgSbtlS9i4seT37yMiS1U1tdBylvSNqXiOHnUXCH21+a+/ds0J9epB374u0fftC02alFwMP/3k\nkv+iRe6DICPj+DeGM888/gFwwQWQmQl//atL2k2bwqWXwt69MHcu7NkDlSu7cr5Ef955x5uQYp20\nK1VyH4qBRMK/PhDJN43j+4ti0vcegfckUBl4QVUfCFjeEngRaIR7uPRIVc0WkQ64B1jXBo4CU1S1\nwIdfW9I3pugOHnRNH4sWuST/4Yewa5dLPGlpLsn36wedO7sEGguHD8OyZce/CXzxhftgCKVOHbj8\ncpfkL7kE6tYNXi7WSTvS/UdL1JK+93DrtcAlQDawBBiuqt/6lXkDeE9V/y0ivYDrVPUaEfkNoKq6\nTkSaAEuBNqq6N9T+LOkb4+TluaS4fXvhw/79x9dr3Ph4ku/TJ/SFwmgqTk1V1d1t06WL+4AK1KJF\n8GQaKNZJO1rNM5EKN+mHcyE3DVivqhu8Dc/APaT4W78ybYE7vPEFwDsAqrrWV0BVt4rIT7hvAyGT\nvjHxZs8ed/fM/Pnw/ffHE3lOTvBkVqcOnH66G1JSjo83buza6Dt0cImwKCK9kOmf9MK9z1zE3emz\ne3fw5Vu2hLf/UB8O4V4InTIleNKeMiW89X3HGGnzTGkJJ+k3BfxPfzbQJaBMJnAlrgnocqCWiDRQ\n1fzPbxFJA6oB3wfuQETGAmMBWpTUJWtjyojDh2HxYndh9aOPXFv3sWPuLpO2bSEx0V1U9SVz/6Fx\nY3ebYDQVN2n7TJhwYsKEot1nXhGS9ogRZTfJn0RVCxyAIbh2fN/0NcDTAWWaAP8HLMcl/mygjt/y\nM4A1wPmF7a9Tp05qTFFt3Kh65ZWqQ4aovvqq6u7dsY7ouGPHVLOyVB99VLV/f9UaNVRBtXJl1Qsu\nUL33XtXPPlM9cqT4+3j1VdWWLVVF3Ourr4a/bsuWLp7AoWXL8NYXCb6+SPix+86Jb6hRo2jHEMnx\nVxRAhhaSX9Wd3kKTfldgrt/0n4E/F1C+JpDtN10bWAYMCScgS/qmqGbOVK1TR7VWLdXGjd27ukoV\n1V69VJ980n0glLbsbNWXXlIdOVL19NOPJ7NzzlEdN071nXdU9+6Nzr4iTZqRJu1IPzR8xxDvSTtS\n0Uz6VYANQCKueSYTaBdQpiFQyRufAkz2xqsB84HbwwlGLembIjhwQPX66927+PzzVTdsUD16VHXx\nYtXx41XbtDmegJKTVf/2N9Vly1zNO5oOHVJdtUr1zTdVb71VtW3b4/tt1Eh1+HDVf/1LddOm0NuI\nZU090vWjUVM3kYta0nfbYgDuDp7vgQnevMnAQG98MLDOK/MCcIo3fyTwK7DCb+hQ0L4s6ZdPx46p\nZmaq7tpVOvtbvtzVmkVU//KX0E0ja9eqPvywardux2u0LVq42va8eeE3qRw7pvrTT6oLF6pOnap6\n552qv/2t6tlnq1aqdDzZJSSo9u2r+tBDLsajRwvfdqxr6ta8UjFENemX5mBJv3zZvt0luHPOce+m\nWrVUJ0woueR/7JjqE0+oVqum2qSJ6vz54a/700+qL76oOmiQavXqLt46dVxNfNw41ebN3bwmTVRv\nv131gQdUR41S7dpVtV69E5NiQoJqUpK7hvDXv6refPPxZpzyVlNXtaRdEVjSNyUmL0/1P/9RveIK\n13YOrib93HMuCfqS/8SJ0b2gumOH6oABbvuXXaaak1P8bf38s+qsWaqjR6vWrh08aYLqGWeoXnyx\n6k03uQ+bOXNUf/jhxBp8Raipm/LPkr6Jug0bXCJv1kzz26vvukt19eoTy2VlqQ4e7MrUru3a0vfs\niWzfH37oatKnnKL69NPRbZdv0SJ40m3ePLz1raZuygJL+iYqfvlFdfp01T593LulUiVX237rLdXD\nhwteNzPTfRvwNaNMmlT0O1YOH1a9+263jTZt3DajLdKattXUTVlgSd9EJCvL3YlSv757l7RqpTp5\nsurmzUXf1vLlqr/7ndtO3bpuO7m5ha+3bp1qaqpb78YbXZNMSbCauqkILOmbIsvNVX3+edW0NPfO\nqFZNdehQd5dLOHehFGbZMtWBA92269VTve++0Mn/5ZdVa9Z0HxJvvRX5vgsSaU3bauqmLLCkbwr0\n00+unfyhh1SvvtrdW+679bB9e3fRcufOktl3Roa7EAvum8SUKar79rllubnuB02getFFBd/bHk2R\n1rStpm5iLdykb/3pV3DHjrmeDFesOHH48cfjZZo3d510JSe7fszT0krncXcZGTBpEvznP1C/Ptx8\nM0yf7no2/NvfXF8oVcJ8tls0+iM3pjyLZi+bppw4dMg9J9Q/uWdmugc+g+tHvW1b6NXLJXlfom/Q\nIDbxpqbCe++5B3xMmuQSdYsW8Omn0K1b+NuJtMMwY+KJ1fTLuX37XE35jTfcU5IOHXLza9U63s2u\nb2jXLnrPKo229HT4059g2zbXj3lRaupl5SEWxsSS1fQrsL174d134c033ePkDh+GM86A66+Hiy92\nCT4xseh9qsdKpDX1zZuLNt+YeGY1/XJi926YNcsl+nnz4NdfoVkzGDzYDV27lp8kHyjSmrrV9I0J\nv6ZfTtNEfNi5E154wT32rnFjGD0aVq2CW291D+HYtAkef9w9YDrWCT893SXfSpXca3p6+OtGWlOf\nMsU9NMNfUR6iYUw8seadMmbHDnj7bVej/+QTOHoUzjwT7rzT1eg7dSqdO2uKItLmmUifnFTeHldn\nTCxZ804ZoAoLFsD//I97VuqxY9C6NQwZ4hJ9hw5lL9H7qygPljamPLMLueXEJ5/AvffCwoXQpImr\nrQ4ZAu3bl+1E7y/S5hmrqRtTeizpx8jChS7Zf/KJu/PmySddbbes3lJZkEibZ6CcPVjamHLMLuSW\nss8+g969oUcP+O47eOIJ+P57d3E2lgk/kguxdiHVmPIjrKQvIv1EZI2IrBeR8UGWtxSR+SKSJSKf\niEgzv2XXisg6b7g2msGXJ198AX36QPfu7g6cxx5z3SPcdhtUrx7b2Hxt6ps2uesLvgux4Sb+ESNc\n+3vLlq5JqmVLa483pqwq9EKuiFTGPfv2EiAbWAIMV9Vv/cq8Abynqv8WkV7Adap6jYjUBzKAVECB\npUAnVd0Tan8V7ULu4sWuGWfePDjtNLjnHrjpppNrxrFk97kbU/5F8z79NGC9qm5Q1SPADGBQQJm2\nwHxvfIHf8v8C5qnqbi/RzwP6hXMA5d1XX7n76y+4wPWB8/DDrmb///5f2Ur4YL9oNSaehJP0mwJb\n/KazvXn+MoErvfHLgVoi0iDMdRGRsSKSISIZOTk54cZeJn39NQwYAOefD0uXwoMPwg8/wF13wamn\nxjq64EJdcC3KhVhjTPkQTtIPduNgYJvQXUAPEVkO9AB+BPLCXBdVnaqqqaqa2qhRozBCKnsOHYIr\nroAuXVzif+ABl+zvvrvsJnsfuxBrTPwI55bNbKC533QzYKt/AVXdClwBICI1gStVNVdEsoGeAet+\nEkG8ZdZtt7lf0k6eDLff7nq5LC/sPnlj4kc4Nf0lQGsRSRSRasAwYLZ/ARFpKCK+bf0ZeNEbnwv0\nFZF6IlIP6OvNq1BeecXdrTJ+PPz1r7FJ+JHccgkuwW/c6H4NvHGjJXxjKqpCk76q5gHjcMl6NTBT\nVVeJyGQRGegV6wmsEZG1QGNgirfubuA+3AfHEmCyN6/C+OYbdzdOjx5w332xiSHSWy6NMfHD+t6J\nwP790Lmz699++XL3y9pYsFsujTHW904JU4UbboB162D+/NglfLBbLo0x4bNuGIrp2Wfh9dfh/vuh\nZ8/YxmK3XBpjwmVJvxi+/hruuAN++1v3C9tYs1sujTHhsqRfRLt2ua6PmzSBl1+O/ROrwPq+McaE\nz9r0i+DYMbjmGti+HT7/HOrXj3VEx1nXxMaYcJSBemr58cADMGeOey5t587R3Xak99kbY0w4rKYf\npo8/dj+8Gj4cbr45utuO9BmzxhgTLrtPPwxbt0LHjq45Z8kSqFkzutu3++yNMZGy+/SjJC8Phg2D\nAwdcbT/aCR/sPntjTOmxNv1CTJjgHnE4dSq0a1cy+7D77I0xpcWSfgFmz4aHHoIbbyzZtnW7z94Y\nU1os6YewYQNcey2kpLiHl5cku8/eGFNarE0/iEOH3A+wAN58ExISSn6fdp+9MaY0WNIP4vbbYdky\nmDULEhNjHY0xxkSPNe8EePVVeP5595jDgQMLL2+MMeWJJX0/q1a5i7bdu9tFVGNMxWRJ33P4MAwe\n7B51OGMGVLGGL2NMBRRW0heRfiKyRkTWi8j4IMtbiMgCEVkuIlkiMsCbX1VE/i0iK0VktYj8OdoH\nEC3PPgvffQfTpsX2gSjGGFOSCk36IlIZeAboD7QFhotI24BiE3HPzu2Ie3D6s978IcApqnoe0Am4\nUURaRSf06MnNdc05l1wC/fvHOhpjjCk54dT004D1qrpBVY8AM4BBAWUUqO2N1wG2+s0/VUSqANWB\nI8C+iKOOsocfdv3kP/BA8bdhvWQaY8qDcJJ+U2CL33S2N8/fJGCkiGQD7wO3ePPfBH4GtgGbgUdU\ndXfgDkRkrIhkiEhGTk5O0Y4gQtu2wWOPud4zU1KKtw1fL5mbNrln5/p6ybTEb4wpa8JJ+hJkXmDX\nnMOBl1S1GTAAeEVEKuG+JRwFmgCJwJ0icuZJG1OdqqqpqpraqFGjIh1ApP77v+HXX+G++4q/jQkT\njneL7HPwoJtvjDFlSThJPxto7jfdjOPNNz7XAzMBVHUxkAA0BK4GPlDVX1X1J+ALoNCuP0vL2rXw\nwgtw001w1lnF3471kmmMKS/CSfpLgNYikigi1XAXamcHlNkM9AYQkTa4pJ/jze8lzqnA+cB30Qo+\nUhMmQPXq7uEokbBeMo0x5UWhSV9V84BxwFxgNe4unVUiMllEfL9ZvRO4QUQygenAKHVPZ3kGqAl8\ng/vwmKaqWSVwHEX21VeuX5277oLTTotsW9ZLpjGmvIjLJ2epQq9e7he433/vfpAVqfR0981h82ZX\nw58yxTpQM8aUHntyVgHmzoVPPoGnn45OwgfrJdMYUz7EXTcMx47BPffAmWcef/i4McbEi7ir6b/2\nGmRlwfTpUK1arKMxxpjSFVc1/cOHYeJE6NgRrroq1tEYY0zpi6ua/v/+r/u17D//6bpLMMaYeBM3\nqS831/3qtk8f17GaMcbEo7hJ+o88EnmnasYYU97FRdL3dao2bBh06hTraIwxJnbiIunfdx8cORJZ\np2rGGFMRVPikv3YtTJ3qnn179tmxjsYYY2Krwif9iRMhISHyTtWMMaYiqNBJ/+uv4Y03XKdqjRvH\nOhpjjIm9Cpv0VWH8eGjUCO68M9bRGGNM2VBhf5z14YewYAE89VT0OlUzxpjyrkLW9H2dqiUmugu4\nxhhjnApZ058+HTIzXedq1qmaMcYcV+Fq+v6dqg0dGutojDGmbAkr6YtIPxFZIyLrRWR8kOUtRGSB\niCwXkSwRGeC3LElEFovIKhFZKSIJ0TyAQM8/Dxs3uu4WrFM1Y4w5UaFpUUQq45512x9oCwwXkbYB\nxSbinp3bEffg9Ge9dasArwI3qWo7oCfwa9SiD7Bvn/vVbe/eRe9ULT0dWrVyHxStWrlpY4ypaMJp\n008D1qvqBgARmQEMAr71K6NAbW+8DrDVG+8LZKlqJoCq7opG0KE88gjs3Olq+SLhr5ee7p6idfCg\nm9606fhTtewRiMaYiiScBpCmwBa/6Wxvnr9JwEgRyQbeB27x5v8GUBGZKyLLROTuYDsQkbEikiEi\nGTk5OUU6AJ/t2+HRR93DUVILfTTwiSZMOJ7wfQ4edPONMaYiCSfpB6sza8D0cOAlVW0GDABeEZFK\nuG8S3YAR3uvlItL7pI2pTlXVVFVNbdSoUZEOwKdGDbjtNrj//qKvu3lz0eYbY0x5FU7Szwaa+003\n43jzjc/1wEwAVV0MJAANvXU/VdWdqnoQ9y0gJdKgg6ldG/7+d2jduujrtmhRtPnGGFNehZP0lwCt\nRSRRRKrhLtTODiizGegNICJtcEk/B5gLJIlIDe+ibg9OvBZQJkyZ4r4p+KtRw803xpiKpNCkr6p5\nwDhcAl+Nu0tnlYhMFpGBXrE7gRtEJBOYDoxSZw/wGO6DYwWwTFX/UxIHEokRI1z3yy1bugvALVu6\nabuIa4ypaEQ1sHk+tlJTUzUjIyPWYRhjTLkiIktVtdDbWOznS8YYE0cs6RtjTByxpG+MMXHEkr4x\nxsQRS/rGGBNHLOkbY0wcsaRvjDFxxJK+McbEEUv6xhgTRyzpG2NMHLGkb4wxccSSvjHGxBFL+sYY\nE0cs6RtjTByxpG+MMXHEkr4xxsSRsJK+iPQTkTUisl5ExgdZ3kJEFojIchHJEpEBQZYfEJG7ohW4\nMcaYois06YtIZeAZoD/QFhguIm0Dik3EPUaxI+4Zus8GLH8cmBN5uMYYYyIRTk0/DVivqhtU9Qgw\nAxgUUEaB2t54HWCrb4GI/A7YAKyKPFxjjDGRCCfpNwW2+E1ne/P8TQJGikg28D5wC4CInArcA/x3\nQTsQkbEikiEiGTk5OWGGbowxpqjCSfoSZF7g09SHAy+pajNgAPCKiFTCJfvHVfVAQTtQ1amqmqqq\nqY0aNQonbmOMMcVQJYwy2UBzv+lm+DXfeK4H+gGo6mIRSQAaAl2AwSLyEFAXOCYih1T1HxFHbowx\npsjCSfpLgNYikgj8iLtQe3VAmc1Ab+AlEWkDJAA5qnqRr4CITAIOWMI3xpjYKbR5R1XzgHHAXGA1\n7i6dVSIyWUQGesXuBG4QkUxgOjBKVQObgIwxxsSYlLXcnJqaqhkZGbEOwxhjyhURWaqqqYWVs1/k\nGmNMHLGkb4wxccSSvjHGxBFL+sYYE0cs6RtjTByxpG+MMXHEkr4xxsQRS/rGGBNHLOkbY0wcsaRv\njDFxxJK+McbEEUv6xhgTRyzpG2NMHLGkb4wxccSSvjHGxBFL+sYYE0cs6RtjTBwJK+mLSD8RWSMi\n60VkfJDlLURkgYgsF5EsERngzb9ERJaKyErvtVe0D8AYY0z4Cn0wuohUBp4BLgGygSUiMltVv/Ur\nNhH37NznRKQt8D7QCtgJXKaqW0WkPe45u02jfAzGGGPCFE5NPw1Yr6obVPUIMAMYFFBGgdreeB1g\nK4CqLlfVrd78VUCCiJwSedjGGGOKI5yk3xTY4jedzcm19UnASBHJxtXybwmynSuB5ap6OHCBiIwV\nkQwRycjJyQkrcGOMMUUXTtKXIPM0YHo48JKqNgMGAK+ISP62RaQd8CBwY7AdqOpUVU1V1dRGjRqF\nF7kxxpgiCyfpZwPN/aab4TXf+LkemAmgqouBBKAhgIg0A94Gfq+q30casDHGmOILJ+kvAVqLSKKI\nVAOGAbMDymwGegOISBtc0s8RkbrAf4A/q+oX0QvbGGNMcRSa9FU1DxiHu/NmNe4unVUiMllEBnrF\n7gRuEJFMYDowSlXVW+9s4K8issIbTiuRIzHGGFMocbm57EhNTdWMjIxYh2GMMeWKiCxV1dTCytkv\nco0xJo5Y0jfGmDhiSd8YY+KIJX1jjIkjlvSNMSaOWNI3xpg4YknfGGPiiCV9Y4yJI5b0jTEmjljS\nN8aYOGJJ3xhj4oglfWOMiSOW9I0xJo5Y0jfGmDhiSd8YY+KIJX1jjIkjlvSNMSaOhJX0RaSfiKwR\nkfUiMj7I8hYiskBElotIlogM8Fv2Z2+9NSLyX9EM3hhjTNFUKayAiFQGngEuAbKBJSIyW1W/9Ss2\nEffs3OdEpC3wPtDKGx8GtAMjzDUNAAAUT0lEQVSaAB+JyG9U9Wi0D8QYY0zhwqnppwHrVXWDqh4B\nZgCDAsooUNsbrwNs9cYHATNU9bCq/gCs97ZnjDEmBsJJ+k2BLX7T2d48f5OAkSKSjavl31KEdRGR\nsSKSISIZOTk5YYZujDGmqMJJ+hJkngZMDwdeUtVmwADgFRGpFOa6qOpUVU1V1dRGjRqFEZIxxpji\nKLRNH1c7b+433YzjzTc+1wP9AFR1sYgkAA3DXNcYY0wpCaemvwRoLSKJIlINd2F2dkCZzUBvABFp\nAyQAOV65YSJyiogkAq2Br6MVvDHGmKIptKavqnkiMg6YC1QGXlTVVSIyGchQ1dnAncA/ReQOXPPN\nKFVVYJWIzAS+BfKAP9qdO8YYEzvicnPZkZqaqhkZGbEOwxhjyhURWaqqqYWVs1/kGmNMHAnnQq4x\nJkZ+/fVXsrOzOXToUKxDMWVEQkICzZo1o2rVqsVa35K+MWVYdnY2tWrVolWrVogEuwPaxBNVZdeu\nXWRnZ5OYmFisbVjzjjFl2KFDh2jQoIElfAOAiNCgQYOIvvlZ0jemjLOEb/xF+n6wpG+MMXHEkr4x\nFUh6OrRqBZUqudf09Mi2t2vXLjp06ECHDh04/fTTadq0af70kSNHwtrGddddx5o1awos88wzz5Ae\nabAmLHYh15gKIj0dxo6Fgwfd9KZNbhpgxIjibbNBgwasWLECgEmTJlGzZk3uuuuuE8qoKqpKpUrB\n65DTpk0rdD9//OMfixdgDOXl5VGlSvlLoVbTN6aCmDDheML3OXjQzY+29evX0759e2666SZSUlLY\ntm0bY8eOJTU1lXbt2jF58uT8st26dWPFihXk5eVRt25dxo8fT3JyMl27duWnn34CYOLEiTzxxBP5\n5cePH09aWhrnnHMOixYtAuDnn3/myiuvJDk5meHDh5Oampr/geTv3nvvpXPnzvnx+X6AunbtWnr1\n6kVycjIpKSls3LgRgL///e+cd955JCcnM8E7Wb6YAbZv387ZZ58NwAsvvMCwYcO49NJL6d+/P/v2\n7aNXr16kpKSQlJTEe++9lx/HtGnTSEpKIjk5meuuu469e/dy5plnkpeXB8DevXtJTEzk6NHS7aTA\nkr4xFcTmzUWbH6lvv/2W66+/nuXLl9O0aVMeeOABMjIyyMzMZN68eXz77bcnrZObm0uPHj3IzMyk\na9euvPjii0G3rap8/fXXPPzww/kfIE8//TSnn346mZmZjB8/nuXLlwdd97bbbmPJkiWsXLmS3Nxc\nPvjgAwCGDx/OHXfcQWZmJosWLeK0007j3XffZc6cOXz99ddkZmZy5513Fnrcixcv5pVXXmHevHlU\nr16dWbNmsWzZMj766CPuuOMOADIzM3nwwQf55JNPyMzM5NFHH6Vu3bpceOGF+fG89tprXHXVVVSu\nXLnwkx1FlvSNqSBatCja/EidddZZdO7cOX96+vTppKSkkJKSwurVq4Mm/erVq9O/f38AOnXqlF/b\nDnTFFVecVObzzz9n2LBhACQnJ9OuXbug686fP5+0tDSSk5P59NNPWbVqFXv27GHnzp1cdtllgPuB\nU40aNfjoo48YPXo01atXB6B+/fqFHnffvn2pV68e4D6c7rnnHpKSkujbty9btmxh586dfPzxxwwd\nOjR/e77XMWPG5Dd3TZs2jeuuu67Q/UWbJX1jKogpU6BGjRPn1ajh5peEU089NX983bp1PPnkk3z8\n8cdkZWXRr1+/oPeSV6tWLX+8cuXK+U0dgU455ZSTyoTTT9jBgwcZN24cb7/9NllZWYwePTo/jmC3\nOqpq0PlVqlTh2LFjACcdh/9xv/zyy+Tm5rJs2TJWrFhBw4YNOXToUMjt9ujRg7Vr17JgwQKqVq3K\nueeeW+gxRZslfWMqiBEjYOpUaNkSRNzr1KnFv4hbFPv27aNWrVrUrl2bbdu2MXfu3Kjvo1u3bsyc\nOROAlStXBv0m8csvv1CpUiUaNmzI/v37eeuttwCoV68eDRs25N133wVcIj948CB9+/blX//6F7/8\n8gsAu3fvBqBVq1YsXboUgDfffDNkTLm5uZx22mlUqVKFefPm8eOPPwLQp08fZsyYkb893yvAyJEj\nGTFiRExq+WBJ35gKZcQI2LgRjh1zr6WR8AFSUlJo27Yt7du354YbbuDCCy+M+j5uueUWfvzxR5KS\nknj00Udp3749derUOaFMgwYNuPbaa2nfvj2XX345Xbp0yV+Wnp7Oo48+SlJSEt26dSMnJ4dLL72U\nfv36kZqaSocOHXj88ccB+NOf/sSTTz7JBRdcwJ49e0LGdM0117Bo0SJSU1N54403aN26NQBJSUnc\nfffddO/enQ4dOvCnP/0pf50RI0aQm5vL0KFDo3l6wmZdKxtThq1evZo2bdrEOowyIS8vj7y8PBIS\nEli3bh19+/Zl3bp15e62yRkzZjB37tywbmUNJdj7ItyulcvX2TLGxK0DBw7Qu3dv8vLyUFWef/75\ncpfwb775Zj766KP8O3hioXydMWNM3Kpbt25+O3t59dxzz8U6hPDa9EWkn4isEZH1IjI+yPLHRWSF\nN6wVkb1+yx4SkVUislpEnhLrPcoYY2Km0Jq+iFQGngEuAbKBJSIyW1XzL52r6h1+5W8BOnrjFwAX\nAkne4s+BHsAnUYrfGGNMEYRT008D1qvqBlU9AswABhVQfjgw3RtXIAGoBpwCVAV2FD9cY4wxkQgn\n6TcFtvhNZ3vzTiIiLYFE4GMAVV0MLAC2ecNcVV0dZL2xIpIhIhk5OTlFOwJjjDFhCyfpB2uDD3Wf\n5zDgTVU9CiAiZwNtgGa4D4peItL9pI2pTlXVVFVNbdSoUXiRG2NKXM+ePU/6odUTTzzBH/7whwLX\nq1mzJgBbt25l8ODBIbdd2O3ZTzzxBAf9epEbMGAAe/fuLWANU5hwkn420NxvuhmwNUTZYRxv2gG4\nHPhSVQ+o6gFgDnB+cQI1xpS+4cOHM2PGjBPmzZgxg+HDh4e1fpMmTQr8RWthApP++++/T926dYu9\nvdKmqvndOZQV4ST9JUBrEUkUkWq4xD47sJCInAPUAxb7zd4M9BCRKiJSFXcR96TmHWNM4W6/HXr2\njO5w++0F73Pw4MG89957HD58GICNGzeydetWunXrln/ffEpKCueddx6zZs06af2NGzfSvn17wHWR\nMGzYMJKSkhg6dGh+1wfg7l/3dct87733AvDUU0+xdetWLr74Yi6++GLAdY+wc+dOAB577DHat29P\n+/bt87tl3rhxI23atOGGG26gXbt29O3b94T9+Lz77rt06dKFjh070qdPH3bscJcaDxw4wHXXXcd5\n551HUlJSfjcOH3zwASkpKSQnJ9O7d2/APV/gkUceyd9m+/bt2bhxY34Mf/jDH0hJSWHLli1Bjw9g\nyZIlXHDBBSQnJ5OWlsb+/fu56KKLTugy+sILLyQrK6vgP1QRFHr3jqrmicg4YC5QGXhRVVeJyGQg\nQ1V9HwDDgRl64k983wR6AStxTUIfqOq7UYveGFOiGjRoQFpaGh988AGDBg1ixowZDB06FBEhISGB\nt99+m9q1a7Nz507OP/98Bg4cGPIZrs899xw1atQgKyuLrKwsUlJS8pdNmTKF+vXrc/ToUXr37k1W\nVha33norjz32GAsWLKBhw4YnbGvp0qVMmzaNr776ClWlS5cu9OjRg3r16rFu3TqmT5/OP//5T666\n6ireeustRo4cecL63bp148svv0REeOGFF3jooYd49NFHue+++6hTpw4rV64EYM+ePeTk5HDDDTew\ncOFCEhMTT+hHJ5Q1a9Ywbdo0nn322ZDHd+655zJ06FBef/11OnfuzL59+6hevTpjxozhpZde4okn\nnmDt2rUcPnyYpKSkQvYYvrB+nKWq7wPvB8z7W8D0pCDrHQVujCA+Y4zHq8yWOl8Tjy/p+/rAV1X+\n8pe/sHDhQipVqsSPP/7Ijh07OP3004NuZ+HChdx6662A65vGP5HNnDmTqVOnkpeXx7Zt2/j2228L\nTHSff/45l19+eX6Pl1dccQWfffYZAwcOJDExkQ4dOgChu2/Ozs5m6NChbNu2jSNHjpCYmAjARx99\ndEJzVr169Xj33Xfp3r17fplwul9u2bIl559/vCU72PGJCGeccUZ+99S1a9cGYMiQIdx33308/PDD\nvPjii4waNarQ/RVFhelwLdrPBjXGOL/73e+YP38+y5Yt45dffsmvoaenp5OTk8PSpUtZsWIFjRs3\nDtqdsr9g3wJ++OEHHnnkEebPn09WVha//e1vC91OQX2G+bplhtDdN99yyy2MGzeOlStX8vzzz+fv\nL1iXyOF0vwwndsHs3/1yqOMLtd0aNWpwySWXMGvWLGbOnMnVV18d8liLo0Ikfd+zQTdtAtXjzwa1\nxG9M5GrWrEnPnj0ZPXr0CRdwfd0KV61alQULFrBp06YCt9O9e/f8h59/8803+e3U+/bt49RTT6VO\nnTrs2LGDOXPm5K9Tq1Yt9u/fH3Rb77zzDgcPHuTnn3/m7bff5qKLLgr7mHJzc2na1N15/u9//zt/\nft++ffnHP/6RP71nzx66du3Kp59+yg8//ACc2P3ysmXLAFi2bFn+8kChju/cc89l69atLFmyBID9\n+/fnf0CNGTOGW2+9lc6dO4f1zaIoKkTSL81ngxoTj4YPH05mZmb+k6vAdRGckZFBamoq6enphT4Q\n5Oabb+bAgQMkJSXx0EMPkZaWBrinYHXs2JF27doxevToE7plHjt2LP3798+/kOuTkpLCqFGjSEtL\no0uXLowZM4aOHTuGfTyTJk1iyJAhXHTRRSdcL5g4cSJ79uyhffv2JCcns2DBAho1asTUqVO54oor\nSE5Ozu8S+corr2T37t106NCB5557jt/85jdB9xXq+KpVq8brr7/OLbfcQnJyMpdcckn+t4VOnTpR\nu3btEulzv0J0rVypkqvhBxJx/YobU15Z18rxaevWrfTs2ZPvvvuOSpVOrptH0rVyhajpl/azQY0x\npqS8/PLLdOnShSlTpgRN+JGqEEm/tJ8NaowxJeX3v/89W7ZsYciQISWy/QqR9GP5bFBjSlpZa4I1\nsRXp+6HCPERlxAhL8qbiSUhIYNeuXTRo0CDkj55M/FBVdu3aRUJCQrG3UWGSvjEVUbNmzcjOzsZ6\nnzU+CQkJNGvWrNjrW9I3pgyrWrVq/i9BjYmGCtGmb4wxJjyW9I0xJo5Y0jfGmDhS5n6RKyI5QMGd\neBSsIbAzSuGUBIsvMhZfZCy+yJTl+FqqaqGPHixzST9SIpIRzk+RY8Xii4zFFxmLLzJlPb5wWPOO\nMcbEEUv6xhgTRypi0p8a6wAKYfFFxuKLjMUXmbIeX6EqXJu+McaY0CpiTd8YY0wIlvSNMSaOlMuk\nLyL9RGSNiKwXkfFBlp8iIq97y78SkValGFtzEVkgIqtFZJWI3BakTE8RyRWRFd7wt9KKzy+GjSKy\n0tv/SY8qE+cp7xxmiUhKKcZ2jt+5WSEi+0Tk9oAypXoOReRFEflJRL7xm1dfROaJyDrvtV6Ida/1\nyqwTkWtLMb6HReQ77+/3tojUDbFuge+FEoxvkoj86Pc3HBBi3QL/30swvtf9YtsoIitCrFvi5y+q\nVLVcDUBl4HvgTKAakAm0DSjzB+B/vfFhwOulGN8ZQIo3XgtYGyS+nsB7MT6PG4GGBSwfAMwBBDgf\n+CqGf+/tuB+exOwcAt2BFOAbv3kPAeO98fHAg0HWqw9s8F7reeP1Sim+vkAVb/zBYPGF814owfgm\nAXeF8fcv8P+9pOILWP4o8LdYnb9oDuWxpp8GrFfVDap6BJgBDAooMwjwPeL+TaC3lFJn5Kq6TVWX\neeP7gdVA09LYd5QNAl5W50ugroicEYM4egPfq2okv9KOmKouBHYHzPZ/n/0b+F2QVf8LmKequ1V1\nDzAP6Fca8anqh6qa501+CRS/P94IhTh/4Qjn/z1iBcXn5Y6rgOnR3m8slMek3xTY4jedzclJNb+M\n96bPBRqUSnR+vGaljsBXQRZ3FZFMEZkjIu1KNTBHgQ9FZKmIjA2yPJzzXBqGEfqfLdbnsLGqbgP3\nYQ+cFqRMWTmPo3Hf3IIp7L1QksZ5zU8vhmgeKwvn7yJgh6quC7E8luevyMpj0g9WYw+87zScMiVK\nRGoCbwG3q+q+gMXLcM0VycDTwDulGZvnQlVNAfoDfxSR7gHLy8I5rAYMBN4IsrgsnMNwlIXzOAHI\nA9JDFCnsvVBSngPOAjoA23BNKIFifv6A4RRcy4/V+SuW8pj0s4HmftPNgK2hyohIFaAOxftqWSwi\nUhWX8NNV9f8Cl6vqPlU94I2/D1QVkYalFZ+3363e60/A27iv0f7COc8lrT+wTFV3BC4oC+cQ2OFr\n8vJefwpSJqbn0btwfCkwQr0G6EBhvBdKhKruUNWjqnoM+GeI/cb6/FUBrgBeD1UmVuevuMpj0l8C\ntBaRRK8mOAyYHVBmNuC7S2Iw8HGoN3y0ee1//wJWq+pjIcqc7rvGICJpuL/DrtKIz9vnqSJSyzeO\nu+D3TUCx2cDvvbt4zgdyfU0ZpShkDSvW59Dj/z67FpgVpMxcoK+I1POaL/p680qciPQD7gEGqurB\nEGXCeS+UVHz+14guD7HfcP7fS1If4DtVzQ62MJbnr9hifSW5OAPuzpK1uKv6E7x5k3FvboAEXJPA\neuBr4MxSjK0b7utnFrDCGwYANwE3eWXGAatwdyJ8CVxQyufvTG/fmV4cvnPoH6MAz3jneCWQWsox\n1sAl8Tp+82J2DnEfPtuAX3G1z+tx14nmA+u81/pe2VTgBb91R3vvxfXAdaUY33pce7jvfei7o60J\n8H5B74VSiu8V772VhUvkZwTG502f9P9eGvF581/yvef8ypb6+YvmYN0wGGNMHCmPzTvGGGOKyZK+\nMcbEEUv6xhgTRyzpG2NMHLGkb4wxccSSvjHGxBFL+sYYE0f+P1eBa4sj9HQAAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e6030307f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e60304d0b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#\n",
    "# Now replot the accuracy after adding the dropout\n",
    "#\n",
    "accuracy = fashion_train_dropout.history['acc']\n",
    "val_accuracy = fashion_train_dropout.history['val_acc']\n",
    "loss = fashion_train_dropout.history['loss']\n",
    "val_loss = fashion_train_dropout.history['val_loss']\n",
    "epochs = range(len(accuracy))\n",
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((10000,), (10000,))"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"\n",
    "Since the predictions you get are floating point values, it will not be feasible to compare the predicted labels with \n",
    "true test labels. So, you will round off the output which will convert the float values into an integer. \n",
    "Further, you will use np.argmax() to select the index number which has a higher value in a row.\n",
    "\n",
    "For example, let's assume a prediction for one test image to be 0 1 0 0 0 0 0 0 0 0, the output for this should be \n",
    "a class label 1.\n",
    "\"\"\"\n",
    "predicted_classes = fashion_model.predict(test_X)\n",
    "predicted_classes = np.argmax(np.round(predicted_classes),axis=1)\n",
    "predicted_classes.shape, test_Y.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 9176 correct labels\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEYCAYAAACUdWs9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS0WSNBhLRZI0GEtFkjSYW3KgtYFj4xRsAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e602e70ba8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#\n",
    "# Show the correctly predicted labels\n",
    "#\n",
    "correct = np.where(predicted_classes==test_Y)[0]\n",
    "print(\"Found %d correct labels\" % len(correct))\n",
    "for i, correct in enumerate(correct[:9]):\n",
    "    plt.subplot(3,3,i+1)\n",
    "    plt.imshow(test_X[correct].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predicted {}, Class {}\".format(predicted_classes[correct], test_Y[correct]))\n",
    "    plt.tight_layout()\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 824 incorrect labels\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEYCAYAAACUdWs9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNSoiIlKa/wDUG1AoVYemqwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1e602f62208>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#\n",
    "# Now show the incorrectly predicted labels\n",
    "#\n",
    "incorrect = np.where(predicted_classes!=test_Y)[0]\n",
    "print(\"Found %d incorrect labels\" % len(incorrect))\n",
    "for i, incorrect in enumerate(incorrect[:9]):\n",
    "    plt.subplot(3,3,i+1)\n",
    "    plt.imshow(test_X[incorrect].reshape(28,28), cmap='gray', interpolation='none')\n",
    "    plt.title(\"Predicted {}, Class {}\".format(predicted_classes[incorrect], test_Y[incorrect]))\n",
    "    plt.tight_layout()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "             precision    recall  f1-score   support\n",
      "\n",
      "    Class 0       0.79      0.86      0.82      1000\n",
      "    Class 1       0.99      0.98      0.99      1000\n",
      "    Class 2       0.91      0.85      0.88      1000\n",
      "    Class 3       0.92      0.93      0.92      1000\n",
      "    Class 4       0.88      0.88      0.88      1000\n",
      "    Class 5       0.99      0.98      0.99      1000\n",
      "    Class 6       0.77      0.77      0.77      1000\n",
      "    Class 7       0.97      0.98      0.97      1000\n",
      "    Class 8       0.99      0.98      0.99      1000\n",
      "    Class 9       0.98      0.97      0.98      1000\n",
      "\n",
      "avg / total       0.92      0.92      0.92     10000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "Classification report will help us in identifying the misclassified classes in more detail. You will be able to observe \n",
    "for which class the model performed bad out of the given ten classes.\n",
    "\"\"\"\n",
    "\n",
    "from sklearn.metrics import classification_report\n",
    "target_names = [\"Class {}\".format(i) for i in range(num_classes)]\n",
    "print(classification_report(test_Y, predicted_classes, target_names=target_names))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}