mnist-inception.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Digit Recognition with Inception Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "keras version: 2.2.4\n",
      "tensorflow version: 1.12.0\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import time\n",
    "\n",
    "import keras\n",
    "from keras import backend as K\n",
    "from keras import layers\n",
    "from keras import optimizers\n",
    "from keras.models import Model\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.utils import to_categorical\n",
    "\n",
    "%matplotlib inline\n",
    "pd.set_option('display.max_colwidth', -1)\n",
    "pd.set_option(\"display.max_rows\", None)\n",
    "\n",
    "keras_version = keras.__version__\n",
    "tf_version = K.tensorflow_backend.tf.VERSION\n",
    "\n",
    "print(\"keras version:\", keras_version)\n",
    "print(K.backend(), \"version:\", tf_version)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Prep"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load data\n",
    "rawdata = np.loadtxt('train.csv', dtype=int, delimiter=',', skiprows=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Raw data shape: (42000, 785)\n",
      "Labels shape: (42000,)\n",
      "Pixels shape: (42000, 784)\n",
      "Pixels reshaped shape: (42000, 28, 28)\n"
     ]
    }
   ],
   "source": [
    "# inspect data\n",
    "print(\"Raw data shape:\", rawdata.shape)\n",
    "\n",
    "# split labels and pixel values\n",
    "y = rawdata[:, 0]\n",
    "X = rawdata[:, 1:]\n",
    "print(\"Labels shape:\", y.shape)\n",
    "print(\"Pixels shape:\", X.shape)\n",
    "\n",
    "# convert pixel values to 2d arrays\n",
    "X = np.reshape(X, (42000, 28, 28))\n",
    "print(\"Pixels reshaped shape:\", X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 70 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# display random sample of images\n",
    "plt.rcParams['figure.figsize'] = (10.0, 8.0)\n",
    "plt.rcParams['image.interpolation'] = 'nearest'\n",
    "plt.rcParams['image.cmap'] = 'gray'\n",
    "\n",
    "num_classes = 10\n",
    "samples_per_class = 7\n",
    "for cls in range(num_classes):\n",
    "    idxs = np.flatnonzero(y == cls)\n",
    "    idxs = np.random.choice(idxs, samples_per_class, replace=False)\n",
    "    for i, idx in enumerate(idxs):\n",
    "        plt_idx = i * num_classes + cls + 1\n",
    "        plt.subplot(samples_per_class, num_classes, plt_idx)\n",
    "        plt.imshow(X[idx])\n",
    "        plt.axis('off')\n",
    "        if i == 0:\n",
    "            plt.title(cls)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(37800, 28, 28, 1) (4200, 28, 28, 1)\n",
      "(37800, 10) (4200, 10)\n",
      "[420 420 420 420 420 420 420 420 420 420]\n"
     ]
    }
   ],
   "source": [
    "# prep the data for training\n",
    "\n",
    "# one hot encode labels\n",
    "y_oh = to_categorical(y, num_classes)\n",
    "\n",
    "# convert pixel values to be between -1 and 1\n",
    "X_norm = X / 127.5 - 1\n",
    "X_norm = np.expand_dims(X_norm, -1) # channels last\n",
    "\n",
    "# split data into train and validation sets\n",
    "num_val = int(y.shape[0] * 0.1)\n",
    "validation_mask = np.zeros(y.shape[0], np.bool)\n",
    "np.random.seed(1)\n",
    "for c in range(num_classes):\n",
    "    idxs = np.random.choice(np.flatnonzero(y == c), num_val // 10, replace=False)\n",
    "    validation_mask[idxs] = 1\n",
    "np.random.seed(None)  \n",
    "    \n",
    "X_train = X_norm[~validation_mask]\n",
    "X_val = X_norm[validation_mask]\n",
    "print(X_train.shape, X_val.shape)\n",
    "\n",
    "y_train = y_oh[~validation_mask]\n",
    "y_val = y_oh[validation_mask]\n",
    "print(y_train.shape, y_val.shape)\n",
    "\n",
    "# confirm validation set is balanced across classes\n",
    "print(np.bincount(y[validation_mask]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Useful Classes / Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "class LinearDecay(keras.callbacks.Callback):\n",
    "    '''\n",
    "    Rate of change = (lr0 - end_lr) / M\n",
    "    '''\n",
    "    def __init__(self, M=None, end_lr=0, min_lr=0.0001, rate_of_change=None):\n",
    "        self.i = 1\n",
    "        self.M = M\n",
    "        self.lr0 = None\n",
    "        self.end_lr = end_lr\n",
    "        self.min_lr = min_lr  \n",
    "        self.rate_of_change = rate_of_change\n",
    "        \n",
    "    def on_train_begin(self, logs=None):    \n",
    "        lr = K.get_value(self.model.optimizer.lr)                                 \n",
    "        if self.lr0 is None:\n",
    "            self.lr0 = lr\n",
    "        else:\n",
    "            print('Learning rate is %s.' % (lr))\n",
    "        if self.rate_of_change is None:\n",
    "            self.rate_of_change = (self.lr0 - self.end_lr) / self.M\n",
    "        \n",
    "    def on_batch_end(self, batch, logs=None):\n",
    "        lr = self.lr0 - self.rate_of_change * self.i  \n",
    "        self.i += 1\n",
    "        if lr > self.min_lr:\n",
    "            K.set_value(self.model.optimizer.lr, lr)\n",
    "        else:\n",
    "            K.set_value(self.model.optimizer.lr, self.min_lr)\n",
    "            \n",
    "    def on_epoch_end(self, epoch, logs=None):\n",
    "        logs = logs or {}\n",
    "        logs['lr'] = K.get_value(self.model.optimizer.lr)\n",
    "        \n",
    "        \n",
    "def plot_training_log(log, acc=True, loss=True, lr=True, figsize=(10.0, 2.5)):\n",
    "    \"\"\"Plot training history.\n",
    "    \n",
    "    # Arguments\n",
    "        log: Dictionary of training history, same as ` History.history`  that  \n",
    "            is returned when fitting a Keras model. Should have records for 'acc', \n",
    "            'val_acc', 'loss', 'val_loss', and optionally 'lr'.\n",
    "        acc: if true, plot both 'acc' and 'val_acc' on one plot.\n",
    "        loss: if true, plot both 'loss' and 'val_loss' on one plot.\n",
    "        lr: if true, and if 'lr' is in log, plot both 'lr' and 'val_acc' on one plot.\n",
    "        figsize: size of each plot.\n",
    "    \"\"\"\n",
    "    \n",
    "    plt.rcParams['figure.figsize'] = figsize\n",
    "    max_val_acc_epoch = log['val_acc'].idxmax() + 1\n",
    "    epochs = range(1, len(log['epoch']) + 1)\n",
    "    \n",
    "    def plot(ytype, ylabel, max_val_acc_epoch):\n",
    "        plt.axvline(x=max_val_acc_epoch, color='0.5', linestyle='--')\n",
    "        plt.plot(epochs, log[ytype], label='Train')\n",
    "        plt.plot(epochs, log['val_' + ytype], label='Validation')\n",
    "        plt.minorticks_on()\n",
    "        plt.grid(b=True, axis='x', which='both', color='0.8', linestyle='-')\n",
    "        plt.xlabel('Epoch')\n",
    "        plt.ylabel(ylabel)\n",
    "        plt.xlim(0, epochs[-1] + 1)\n",
    "        plt.legend()\n",
    "        plt.tight_layout()\n",
    "        plt.show()\n",
    "        \n",
    "    if acc:\n",
    "        plot('acc', 'Accuracy', max_val_acc_epoch)\n",
    "        \n",
    "    if loss:\n",
    "        plot('loss', 'Loss', max_val_acc_epoch)\n",
    "    \n",
    "    if lr and 'lr' in log.columns:\n",
    "        fig, ax1 = plt.subplots()\n",
    "        plt.axvline(x=max_val_acc_epoch, color='0.5', linestyle='--')\n",
    "\n",
    "        ln1 = ax1.plot(epochs, log['lr'], 'C4o-', label='Learning Rate')   \n",
    "        ax1.set_xticks(epochs, minor=True)\n",
    "        ax1.grid(b=True, axis='x', which='both', color='0.8', linestyle='-')\n",
    "        ax1.set_xlabel('Epoch')\n",
    "        ax1.set_ylabel('Learning Rate')\n",
    "        ax1.set_xlim(0, epochs[-1] + 1)\n",
    "\n",
    "        ax2 = ax1.twinx()\n",
    "        ln2 = ax2.plot(epochs, log['val_acc'], 'C1o-', \n",
    "                       label='Validation Accuracy [0.98 - 1.0]')        \n",
    "        #ax2.set_ylabel('Val Accuracy')\n",
    "        ax2.set_yticks([])\n",
    "        ax2.set_ylim(0.98, 1.0)        \n",
    "\n",
    "        lns = ln1 + ln2\n",
    "        plt.legend(lns, [l.get_label() for l in lns])\n",
    "        plt.tight_layout()\n",
    "        plt.show()\n",
    "        \n",
    "        \n",
    "def evaluate_model(model, X, y, log=None, pre_train_weights=None):\n",
    "    \"\"\"Display accuracy, display misclassified digits, and visualize weights in \n",
    "    first convolution layer for the given model with X and y as prediction inputs.\n",
    "    Optionally, accuracy and loss is plotted with the log argument and additional\n",
    "    weights can be visualized with the pre_train_weights argument.\n",
    "    \n",
    "    # Arguments\n",
    "        model: Keras model with first convolution layer named 'conv_1'.\n",
    "        X: Numpy array of MNIST digits, in channels last format.\n",
    "        y: Labels for X.\n",
    "        log: (optional) Dictionary of training history, same as ` History.history`  that  \n",
    "            is returned when fitting a Keras model. Should have records for 'acc', \n",
    "            'val_acc', 'loss', 'val_loss', and optionally 'lr'.\n",
    "        pre_train_weights: (optional) Numpy array of weights in first convolution layer \n",
    "            of the model.\n",
    "    \"\"\"\n",
    " \n",
    "    if log is not None:\n",
    "        print(\"Max Validation Accuracy:\", max(log['val_acc']))\n",
    "        plot_training_log(log)\n",
    "\n",
    "    scores = model.predict(X)\n",
    "    predictions = np.argmax(scores, axis=1)\n",
    "    y_digits = np.nonzero(y)[1]       \n",
    "    print(\"Accuracy:\",  np.mean(predictions == y_digits))\n",
    "    \n",
    "    post_train_weights = model.get_layer('conv_1').get_weights()[0]\n",
    "    post_train_weights = post_train_weights.transpose(3, 2, 0, 1)\n",
    "    num_weights = len(post_train_weights)\n",
    "    if pre_train_weights is not None:\n",
    "        plt.rcParams['figure.figsize'] = (10.0, 3.0)\n",
    "        for i in range(num_weights):\n",
    "            plt.subplot(4, num_weights // 4, i + 1)\n",
    "            ker = pre_train_weights[i, 0]\n",
    "            low, high = np.amin(ker), np.max(ker)\n",
    "            plt.imshow(255 * (ker - low) / (high - low))\n",
    "            plt.axis('off')\n",
    "        plt.suptitle('Pre-Training Weights from First Convolutional Layer')\n",
    "        plt.show()\n",
    "    plt.rcParams['figure.figsize'] = (10.0, 3.0)\n",
    "    for i in range(num_weights):\n",
    "        plt.subplot(4, num_weights // 4, i + 1)\n",
    "        ker = post_train_weights[i, 0]\n",
    "        low, high = np.amin(ker), np.max(ker)\n",
    "        plt.imshow(255 * (ker - low) / (high - low))\n",
    "        plt.axis('off')\n",
    "    plt.suptitle('Post-Training Weights from First Convolutional Layer')\n",
    "    plt.show()\n",
    "\n",
    "    misclassified_mask = (predictions != y_digits)\n",
    "    samples_per_class = 7\n",
    "    num_classes = 10\n",
    "    plt.rcParams['figure.figsize'] = (10.0, 7.0)    \n",
    "    counts = [];\n",
    "    for cls in range(10):\n",
    "        idxs = np.flatnonzero(y_digits[misclassified_mask] == cls)\n",
    "        counts.append(len(idxs))\n",
    "        if len(idxs) > samples_per_class:\n",
    "            idxs = np.random.choice(idxs, samples_per_class, replace=False)       \n",
    "        for i, idx in enumerate(idxs):\n",
    "            plt_idx = i * num_classes + cls + 1\n",
    "            plt.subplot(samples_per_class, num_classes, plt_idx)\n",
    "            plt.imshow(X[misclassified_mask][idx,:,:,0])\n",
    "            plt.axis('off')\n",
    "            plt.text(14, 27, 'Pred:' + str(predictions[misclassified_mask][idx]), \n",
    "                     horizontalalignment='center', verticalalignment='top')        \n",
    "            if i == 0:\n",
    "                plt.title(cls)            \n",
    "    plt.suptitle('Misclassified Digits | Norm of Counts = ' + str(np.linalg.norm(counts)))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inception Modules"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv2D(x, filters, kernel_size, strides, padding='valid', kernel_initializer='glorot_uniform', batch_norm=True):\n",
    "    if batch_norm:\n",
    "        x = layers.Conv2D(filters=filters, \n",
    "                          kernel_size=kernel_size, \n",
    "                          strides=strides, \n",
    "                          padding=padding, \n",
    "                          kernel_initializer=kernel_initializer, \n",
    "                          #kernel_constraint=keras.constraints.MaxNorm(4., axis=(0,1,2)),\n",
    "                          use_bias=False)(x)\n",
    "        x = layers.BatchNormalization(scale=False)(x)\n",
    "    else:\n",
    "        x = layers.Conv2D(filters=filters, \n",
    "                          kernel_size=kernel_size, \n",
    "                          strides=strides, \n",
    "                          padding=padding, \n",
    "                          kernel_initializer=kernel_initializer, \n",
    "                          use_bias=True)(x)\n",
    "\n",
    "    return layers.Activation('relu')(x)\n",
    "    #return layers.PReLU(alpha_initializer=keras.initializers.Constant(value=0.25), alpha_regularizer=None, alpha_constraint=None, shared_axes=[1, 2])(x)\n",
    "\n",
    "\n",
    "def inception_module_A(x, filters=None, dropout=None, kernel_initializer='glorot_uniform', batch_norm=True):\n",
    "    if filters is None:\n",
    "        filters = int(x.shape[-1])\n",
    "    branch_filters = filters // 4\n",
    "        \n",
    "    b1 = conv2D(x, \n",
    "                filters=(branch_filters // 3) * 2, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    if dropout is not None:\n",
    "        #b1 = layers.Dropout(dropout, noise_shape=(1, 1, 1, (branch_filters // 3) * 2))(b1)\n",
    "        b1 = layers.SpatialDropout2D(dropout)(b1)\n",
    "    b1 = conv2D(b1, \n",
    "                filters=branch_filters, \n",
    "                kernel_size=3, \n",
    "                strides=1, \n",
    "                padding='same', \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    \n",
    "    b2 = conv2D(x, \n",
    "                filters=(branch_filters // 3) * 2, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    if dropout is not None:\n",
    "        #b2 = layers.Dropout(dropout, noise_shape=(1, 1, 1, (branch_filters // 3) * 2))(b2)\n",
    "        b2 = layers.SpatialDropout2D(dropout)(b2)        \n",
    "    b2 = conv2D(b2, \n",
    "                filters=branch_filters, \n",
    "                kernel_size=3, \n",
    "                strides=1, \n",
    "                padding='same', \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    b2 = conv2D(b2, \n",
    "                filters=branch_filters, \n",
    "                kernel_size=3, \n",
    "                strides=1, \n",
    "                padding='same', \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "        \n",
    "    b3 = conv2D(x, \n",
    "                filters=branch_filters, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    \n",
    "    pool = layers.AveragePooling2D(pool_size=(3, 3), strides=1, padding='same')(x)\n",
    "    pool = conv2D(pool, \n",
    "                  filters=branch_filters, \n",
    "                  kernel_size=1, \n",
    "                  strides=1, \n",
    "                  kernel_initializer=kernel_initializer, \n",
    "                  batch_norm=batch_norm)\n",
    "\n",
    "    x = layers.concatenate([b1, b2, b3, pool])\n",
    "    if dropout is not None:\n",
    "        #x = layers.Dropout(dropout, noise_shape=(1, 1, 1, branch_filters * 4))(x)\n",
    "        x = layers.SpatialDropout2D(dropout)(x)\n",
    "    \n",
    "    return x\n",
    "\n",
    "\n",
    "def inception_module_B(x, filters=None, dropout=None, kernel_initializer='glorot_uniform', batch_norm=True):\n",
    "    if filters is None:\n",
    "        filters = int(x.shape[-1])\n",
    "    branch_filters = filters // 6\n",
    "        \n",
    "    b1 = conv2D(x, \n",
    "                filters=(branch_filters // 2) * 3, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    if dropout is not None:\n",
    "        #b1 = layers.Dropout(dropout, noise_shape=(1, 1, 1, (branch_filters // 2) * 3))(b1)\n",
    "        b1 = layers.SpatialDropout2D(dropout)(b1)\n",
    "        \n",
    "    b1a = conv2D(b1, \n",
    "                 filters=branch_filters, \n",
    "                 kernel_size=(1, 3), \n",
    "                 strides=1, \n",
    "                 padding='same', \n",
    "                 kernel_initializer=kernel_initializer, \n",
    "                 batch_norm=batch_norm)\n",
    "    \n",
    "    b1b = conv2D(b1, \n",
    "                 filters=branch_filters, \n",
    "                 kernel_size=(3, 1), \n",
    "                 strides=1, \n",
    "                 padding='same', \n",
    "                 kernel_initializer=kernel_initializer, \n",
    "                 batch_norm=batch_norm)\n",
    "    \n",
    "    b2 = conv2D(x, \n",
    "                filters=(branch_filters // 2) * 3, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    if dropout is not None:\n",
    "        #b2 = layers.Dropout(dropout, noise_shape=(1, 1, 1, (branch_filters // 2) * 3))(b2)    \n",
    "        b2 = layers.SpatialDropout2D(dropout)(b2)  \n",
    "    b2 = conv2D(b2, \n",
    "                filters=(branch_filters // 4) * 7, \n",
    "                kernel_size=(1, 3), \n",
    "                strides=1, \n",
    "                padding='same', \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    b2 = conv2D(b2, \n",
    "                filters=branch_filters * 2, \n",
    "                kernel_size=(3, 1), \n",
    "                strides=1, \n",
    "                padding='same', \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "\n",
    "    b2a = conv2D(b2, \n",
    "                 filters=branch_filters, \n",
    "                 kernel_size=(1, 3), \n",
    "                 strides=1, \n",
    "                 padding='same', \n",
    "                 kernel_initializer=kernel_initializer, \n",
    "                 batch_norm=batch_norm)\n",
    "    \n",
    "    b2b = conv2D(b2, \n",
    "                 branch_filters, \n",
    "                 kernel_size=(3, 1), \n",
    "                 strides=1, \n",
    "                 padding='same', \n",
    "                 kernel_initializer=kernel_initializer, \n",
    "                 batch_norm=batch_norm)\n",
    "        \n",
    "    b3 = conv2D(x, \n",
    "                filters=branch_filters, \n",
    "                kernel_size=1, \n",
    "                strides=1, \n",
    "                kernel_initializer=kernel_initializer, \n",
    "                batch_norm=batch_norm)\n",
    "    \n",
    "    pool = layers.AveragePooling2D(pool_size=(3, 3), strides=1, padding='same')(x)\n",
    "    pool = conv2D(pool, \n",
    "                  filters=branch_filters, \n",
    "                  kernel_size=1, \n",
    "                  strides=1, \n",
    "                  kernel_initializer=kernel_initializer, \n",
    "                  batch_norm=batch_norm)\n",
    "    \n",
    "    x = layers.concatenate([b1a, b1b, b2a, b2b, b3, pool])\n",
    "    if dropout is not None:\n",
    "        #x = layers.Dropout(dropout, noise_shape=(1, 1, 1, branch_filters * 6))(x)\n",
    "        x = layers.SpatialDropout2D(dropout)(x)\n",
    "    \n",
    "    return x\n",
    "\n",
    "\n",
    "def reduction_module(x, filters, padding='same', batch_norm=True):\n",
    "    branch_filters = (filters - int(x.shape[-1])) // 2\n",
    "        \n",
    "    b1 = conv2D(x, (branch_filters // 3) * 2, 1, strides=1, batch_norm=batch_norm)\n",
    "    b1 = conv2D(b1, branch_filters, 3, strides=2, padding=padding, batch_norm=batch_norm)\n",
    "    \n",
    "    b2 = conv2D(x, (branch_filters // 3) * 2, 1, strides=1, batch_norm=batch_norm)\n",
    "    b2 = conv2D(b2, branch_filters, 3, strides=1, padding='same', batch_norm=batch_norm)\n",
    "    b2 = conv2D(b2, branch_filters, 3, strides=2, padding=padding, batch_norm=batch_norm)\n",
    "    \n",
    "    pool = layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding=padding)(x)\n",
    "    #pool = conv2D(pool, branch_filters, 1, strides=1, batch_norm=batch_norm)\n",
    "\n",
    "    return layers.concatenate([b1, b2, pool])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Label Probabilities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[1.000e-04 9.991e-01 1.000e-04 ... 1.000e-04 1.000e-04 1.000e-04]\n",
      " [9.991e-01 1.000e-04 1.000e-04 ... 1.000e-04 1.000e-04 1.000e-04]\n",
      " [1.000e-04 9.991e-01 1.000e-04 ... 1.000e-04 1.000e-04 1.000e-04]\n",
      " ...\n",
      " [1.000e-04 1.000e-04 1.000e-04 ... 9.991e-01 1.000e-04 1.000e-04]\n",
      " [1.000e-04 1.000e-04 1.000e-04 ... 1.000e-04 1.000e-04 1.000e-04]\n",
      " [1.000e-04 1.000e-04 1.000e-04 ... 1.000e-04 1.000e-04 9.991e-01]]\n"
     ]
    }
   ],
   "source": [
    "# label smoothing\n",
    "epsilon = 0.001\n",
    "y_train_smooth = y_train * (1 - epsilon) + epsilon / 10\n",
    "print(y_train_smooth)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Inception Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 28, 28, 1)    0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv_1 (Conv2D)                 (None, 12, 12, 64)   1600        input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_1 (BatchNor (None, 12, 12, 64)   192         conv_1[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "activation_1 (Activation)       (None, 12, 12, 64)   0           batch_normalization_1[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_3 (Conv2D)               (None, 12, 12, 16)   1024        activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_4 (BatchNor (None, 12, 12, 16)   48          conv2d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_4 (Activation)       (None, 12, 12, 16)   0           batch_normalization_4[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_1 (Conv2D)               (None, 12, 12, 16)   1024        activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_4 (Conv2D)               (None, 12, 12, 24)   3456        activation_4[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_2 (BatchNor (None, 12, 12, 16)   48          conv2d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_5 (BatchNor (None, 12, 12, 24)   72          conv2d_4[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_2 (Activation)       (None, 12, 12, 16)   0           batch_normalization_2[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_5 (Activation)       (None, 12, 12, 24)   0           batch_normalization_5[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_1 (AveragePoo (None, 12, 12, 64)   0           activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_2 (Conv2D)               (None, 12, 12, 24)   3456        activation_2[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_5 (Conv2D)               (None, 12, 12, 24)   5184        activation_5[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_6 (Conv2D)               (None, 12, 12, 24)   1536        activation_1[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_7 (Conv2D)               (None, 12, 12, 24)   1536        average_pooling2d_1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_3 (BatchNor (None, 12, 12, 24)   72          conv2d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_6 (BatchNor (None, 12, 12, 24)   72          conv2d_5[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_7 (BatchNor (None, 12, 12, 24)   72          conv2d_6[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_8 (BatchNor (None, 12, 12, 24)   72          conv2d_7[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "activation_3 (Activation)       (None, 12, 12, 24)   0           batch_normalization_3[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_6 (Activation)       (None, 12, 12, 24)   0           batch_normalization_6[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_7 (Activation)       (None, 12, 12, 24)   0           batch_normalization_7[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_8 (Activation)       (None, 12, 12, 24)   0           batch_normalization_8[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_1 (Concatenate)     (None, 12, 12, 96)   0           activation_3[0][0]               \n",
      "                                                                 activation_6[0][0]               \n",
      "                                                                 activation_7[0][0]               \n",
      "                                                                 activation_8[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_10 (Conv2D)              (None, 12, 12, 16)   1536        concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_11 (BatchNo (None, 12, 12, 16)   48          conv2d_10[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_11 (Activation)      (None, 12, 12, 16)   0           batch_normalization_11[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_8 (Conv2D)               (None, 12, 12, 16)   1536        concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_11 (Conv2D)              (None, 12, 12, 24)   3456        activation_11[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_9 (BatchNor (None, 12, 12, 16)   48          conv2d_8[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_12 (BatchNo (None, 12, 12, 24)   72          conv2d_11[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_9 (Activation)       (None, 12, 12, 16)   0           batch_normalization_9[0][0]      \n",
      "__________________________________________________________________________________________________\n",
      "activation_12 (Activation)      (None, 12, 12, 24)   0           batch_normalization_12[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_2 (AveragePoo (None, 12, 12, 96)   0           concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_9 (Conv2D)               (None, 12, 12, 24)   3456        activation_9[0][0]               \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_12 (Conv2D)              (None, 12, 12, 24)   5184        activation_12[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_13 (Conv2D)              (None, 12, 12, 24)   2304        concatenate_1[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_14 (Conv2D)              (None, 12, 12, 24)   2304        average_pooling2d_2[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_10 (BatchNo (None, 12, 12, 24)   72          conv2d_9[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_13 (BatchNo (None, 12, 12, 24)   72          conv2d_12[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_14 (BatchNo (None, 12, 12, 24)   72          conv2d_13[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_15 (BatchNo (None, 12, 12, 24)   72          conv2d_14[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_10 (Activation)      (None, 12, 12, 24)   0           batch_normalization_10[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_13 (Activation)      (None, 12, 12, 24)   0           batch_normalization_13[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_14 (Activation)      (None, 12, 12, 24)   0           batch_normalization_14[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_15 (Activation)      (None, 12, 12, 24)   0           batch_normalization_15[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_2 (Concatenate)     (None, 12, 12, 96)   0           activation_10[0][0]              \n",
      "                                                                 activation_13[0][0]              \n",
      "                                                                 activation_14[0][0]              \n",
      "                                                                 activation_15[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "spatial_dropout2d_1 (SpatialDro (None, 12, 12, 96)   0           concatenate_2[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2D)  (None, 6, 6, 96)     0           spatial_dropout2d_1[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_18 (Conv2D)              (None, 6, 6, 36)     3456        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_19 (BatchNo (None, 6, 6, 36)     108         conv2d_18[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_19 (Activation)      (None, 6, 6, 36)     0           batch_normalization_19[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_19 (Conv2D)              (None, 6, 6, 42)     4536        activation_19[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_20 (BatchNo (None, 6, 6, 42)     126         conv2d_19[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_20 (Activation)      (None, 6, 6, 42)     0           batch_normalization_20[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_15 (Conv2D)              (None, 6, 6, 36)     3456        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_20 (Conv2D)              (None, 6, 6, 48)     6048        activation_20[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_16 (BatchNo (None, 6, 6, 36)     108         conv2d_15[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_21 (BatchNo (None, 6, 6, 48)     144         conv2d_20[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_16 (Activation)      (None, 6, 6, 36)     0           batch_normalization_16[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_21 (Activation)      (None, 6, 6, 48)     0           batch_normalization_21[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "average_pooling2d_3 (AveragePoo (None, 6, 6, 96)     0           max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_16 (Conv2D)              (None, 6, 6, 24)     2592        activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_17 (Conv2D)              (None, 6, 6, 24)     2592        activation_16[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_21 (Conv2D)              (None, 6, 6, 24)     3456        activation_21[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_22 (Conv2D)              (None, 6, 6, 24)     3456        activation_21[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_23 (Conv2D)              (None, 6, 6, 24)     2304        max_pooling2d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv2d_24 (Conv2D)              (None, 6, 6, 24)     2304        average_pooling2d_3[0][0]        \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_17 (BatchNo (None, 6, 6, 24)     72          conv2d_16[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_18 (BatchNo (None, 6, 6, 24)     72          conv2d_17[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_22 (BatchNo (None, 6, 6, 24)     72          conv2d_21[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_23 (BatchNo (None, 6, 6, 24)     72          conv2d_22[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_24 (BatchNo (None, 6, 6, 24)     72          conv2d_23[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "batch_normalization_25 (BatchNo (None, 6, 6, 24)     72          conv2d_24[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "activation_17 (Activation)      (None, 6, 6, 24)     0           batch_normalization_17[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_18 (Activation)      (None, 6, 6, 24)     0           batch_normalization_18[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_22 (Activation)      (None, 6, 6, 24)     0           batch_normalization_22[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_23 (Activation)      (None, 6, 6, 24)     0           batch_normalization_23[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_24 (Activation)      (None, 6, 6, 24)     0           batch_normalization_24[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "activation_25 (Activation)      (None, 6, 6, 24)     0           batch_normalization_25[0][0]     \n",
      "__________________________________________________________________________________________________\n",
      "concatenate_3 (Concatenate)     (None, 6, 6, 144)    0           activation_17[0][0]              \n",
      "                                                                 activation_18[0][0]              \n",
      "                                                                 activation_22[0][0]              \n",
      "                                                                 activation_23[0][0]              \n",
      "                                                                 activation_24[0][0]              \n",
      "                                                                 activation_25[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "global_average_pooling2d_1 (Glo (None, 144)          0           concatenate_3[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "dropout_1 (Dropout)             (None, 144)          0           global_average_pooling2d_1[0][0] \n",
      "__________________________________________________________________________________________________\n",
      "logits (Dense)                  (None, 10)           1450        dropout_1[0][0]                  \n",
      "__________________________________________________________________________________________________\n",
      "softmax (Activation)            (None, 10)           0           logits[0][0]                     \n",
      "==================================================================================================\n",
      "Total params: 76,264\n",
      "Trainable params: 74,916\n",
      "Non-trainable params: 1,348\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "K.clear_session()\n",
    "\n",
    "stem_width = 64\n",
    "kernel_initializer = 'glorot_uniform'\n",
    "\n",
    "inputs = layers.Input(shape=X_norm.shape[1:])\n",
    "\n",
    "#x = layers.ZeroPadding2D(padding=(2,2))(inputs)\n",
    "\n",
    "x = layers.Conv2D(filters=stem_width, \n",
    "                  kernel_size=5, \n",
    "                  strides=2, \n",
    "                  padding='valid', \n",
    "                  use_bias=False, \n",
    "                  kernel_initializer=kernel_initializer, \n",
    "                  #kernel_constraint=keras.constraints.MaxNorm(4., axis=(0,1,2)),\n",
    "                  name='conv_1')(inputs)\n",
    "x = layers.BatchNormalization(scale=False)(x)\n",
    "x = layers.Activation('relu')(x)\n",
    "#x = layers.SpatialDropout2D(0.2)(x)\n",
    "\n",
    "x = inception_module_A(x, filters=int(1.5*stem_width), kernel_initializer=kernel_initializer, batch_norm=True)\n",
    "#x = layers.SpatialDropout2D(0.3)(x)\n",
    "\n",
    "x = inception_module_A(x, filters=int(1.5*stem_width), kernel_initializer=kernel_initializer, batch_norm=True)\n",
    "x = layers.SpatialDropout2D(0.3)(x)\n",
    "\n",
    "x = layers.MaxPooling2D(pool_size=(3, 3), strides=2, padding='same')(x)\n",
    "\n",
    "x = inception_module_B(x, filters=int(2.25*stem_width), kernel_initializer=kernel_initializer, batch_norm=True)\n",
    "\n",
    "x = layers.GlobalAveragePooling2D()(x)\n",
    "x = layers.Dropout(0.5)(x)\n",
    "\n",
    "x = layers.Dense(num_classes, name='logits')(x)\n",
    "x = layers.Activation('softmax', name='softmax')(x)\n",
    "\n",
    "model = Model(inputs=inputs, outputs=x)\n",
    "model.compile(loss='categorical_crossentropy', \n",
    "              #optimizer=optimizers.SGD(lr=0.04, momentum=0.5, nesterov=False),\n",
    "              #optimizer=optimizers.RMSprop(lr=0.045, rho=0.9, epsilon=1.0, clipvalue=2.0), # incep v3 paper setting max out at .9940, weight vis still noisy after 100 epochs\n",
    "              #optimizer=optimizers.Adam(lr=0.003, beta_1=0.9, beta_2=0.999),\n",
    "              optimizer=optimizers.Adamax(lr=0.01, beta_1=0.49, beta_2=0.999),\n",
    "              metrics=['accuracy'])\n",
    "model.summary()\n",
    "\n",
    "# save weights for visualization\n",
    "pre_train_weights = model.get_layer('conv_1').get_weights()[0]\n",
    "pre_train_weights = pre_train_weights.transpose(3, 2, 0, 1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# used when saving results to file, see \"Save Results\" section\n",
    "optimizer_args = \"(lr=0.01, beta_1=0.49, beta_2=0.999)\"\n",
    "dropout_type = \"spatial\"\n",
    "dropout_rates = \"(0,0,.3,.5)\"\n",
    "annealer_args = \"(max_lr=0.01, min_lr=0.0, T=10, T_mul=1)\"\n",
    "notes = \"stem_width=64, A_width=96, B_width=144, only elastic_transform(alpha_range=[8, 11], sigma=3),\\\n",
    " width_shift_range=2, height_shift_range=2\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.ndimage.interpolation import map_coordinates\n",
    "from scipy.ndimage.filters import gaussian_filter\n",
    "\n",
    "def elastic_transform(image, alpha_range, sigma, random_state=None):\n",
    "    \"\"\"Elastic deformation of images as described in [Simard2003]_.\n",
    "    .. [Simard2003] Simard, Steinkraus and Platt, \"Best Practices for\n",
    "       Convolutional Neural Networks applied to Visual Document Analysis\", in\n",
    "       Proc. of the International Conference on Document Analysis and\n",
    "       Recognition, 2003.\n",
    "    \"\"\"\n",
    "    if random_state is None:\n",
    "        random_state = np.random.RandomState(None)\n",
    "        \n",
    "    alpha = np.random.uniform(low=alpha_range[0], high=alpha_range[1])\n",
    "\n",
    "    shape = image.shape\n",
    "    dx = gaussian_filter((random_state.rand(*shape) * 2 - 1), sigma) * alpha\n",
    "    dy = gaussian_filter((random_state.rand(*shape) * 2 - 1), sigma) * alpha\n",
    "\n",
    "    x, y, z = np.meshgrid(np.arange(shape[0]), np.arange(shape[1]), np.arange(shape[2]), indexing='ij')\n",
    "    indices = np.reshape(x+dx, (-1, 1)), np.reshape(y+dy, (-1, 1)), np.reshape(z, (-1, 1))\n",
    "\n",
    "    return map_coordinates(image, indices, order=1, mode='reflect').reshape(shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CosineAnneal(keras.callbacks.Callback):\n",
    "    def __init__(self, max_lr, min_lr, T, T_mul=1):\n",
    "        self.max_lr = max_lr\n",
    "        self.min_lr = min_lr\n",
    "        self.T = T\n",
    "        self.T_cur = 0\n",
    "        self.T_mul = T_mul\n",
    "        self.step = 0\n",
    "        \n",
    "    def on_batch_begin(self, batch, logs=None):\n",
    "        if self.T <= self.T_cur:\n",
    "            self.T *= self.T_mul\n",
    "            self.T_cur = 0\n",
    "            self.step = 0\n",
    "        lr = self.min_lr + 0.5 * (self.max_lr - self.min_lr) * (1 + np.cos(self.T_cur * np.pi / self.T))        \n",
    "        K.set_value(self.model.optimizer.lr, lr)\n",
    "        # use self.step to avoid floating point arithmetic errors at warm restarts\n",
    "        self.step += 1\n",
    "        self.T_cur = self.step / self.params['steps']\n",
    "        #print(\"\\n\", batch, \":\", K.get_value(self.model.optimizer.lr))\n",
    "            \n",
    "    def on_epoch_end(self, epoch, logs=None):\n",
    "        logs = logs or {}\n",
    "        logs['lr'] = K.get_value(self.model.optimizer.lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/60\n",
      "591/591 [==============================] - 328s 556ms/step - loss: 0.4029 - acc: 0.8809 - val_loss: 0.0796 - val_acc: 0.9776\n",
      "Epoch 2/60\n",
      "591/591 [==============================] - 354s 599ms/step - loss: 0.1436 - acc: 0.9627 - val_loss: 0.1945 - val_acc: 0.9419\n",
      "Epoch 3/60\n",
      "591/591 [==============================] - 332s 563ms/step - loss: 0.1074 - acc: 0.9723 - val_loss: 0.0654 - val_acc: 0.9805\n",
      "Epoch 4/60\n",
      "591/591 [==============================] - 2957s 5s/step - loss: 0.0908 - acc: 0.9765 - val_loss: 0.0568 - val_acc: 0.9840\n",
      "Epoch 5/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0771 - acc: 0.9806 - val_loss: 0.0334 - val_acc: 0.9910\n",
      "Epoch 6/60\n",
      "591/591 [==============================] - 338s 573ms/step - loss: 0.0701 - acc: 0.9837 - val_loss: 0.0287 - val_acc: 0.9905\n",
      "Epoch 7/60\n",
      "591/591 [==============================] - 9445s 16s/step - loss: 0.0614 - acc: 0.9858 - val_loss: 0.0276 - val_acc: 0.9917\n",
      "Epoch 8/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0525 - acc: 0.9881 - val_loss: 0.0258 - val_acc: 0.9919\n",
      "Epoch 9/60\n",
      "591/591 [==============================] - 320s 541ms/step - loss: 0.0542 - acc: 0.9881 - val_loss: 0.0231 - val_acc: 0.9936\n",
      "Epoch 10/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0493 - acc: 0.9889 - val_loss: 0.0233 - val_acc: 0.9929\n",
      "Epoch 11/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0783 - acc: 0.9805 - val_loss: 0.0434 - val_acc: 0.9881\n",
      "Epoch 12/60\n",
      "591/591 [==============================] - 336s 568ms/step - loss: 0.0728 - acc: 0.9825 - val_loss: 0.0351 - val_acc: 0.9914\n",
      "Epoch 13/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0695 - acc: 0.9838 - val_loss: 0.0372 - val_acc: 0.9893\n",
      "Epoch 14/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0614 - acc: 0.9857 - val_loss: 0.0326 - val_acc: 0.9914\n",
      "Epoch 15/60\n",
      "591/591 [==============================] - 321s 544ms/step - loss: 0.0559 - acc: 0.9874 - val_loss: 0.0244 - val_acc: 0.9936\n",
      "Epoch 16/60\n",
      "591/591 [==============================] - 322s 544ms/step - loss: 0.0513 - acc: 0.9885 - val_loss: 0.0235 - val_acc: 0.9936\n",
      "Epoch 17/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0466 - acc: 0.9901 - val_loss: 0.0249 - val_acc: 0.9936\n",
      "Epoch 18/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0417 - acc: 0.9910 - val_loss: 0.0221 - val_acc: 0.9945\n",
      "Epoch 19/60\n",
      "591/591 [==============================] - 323s 546ms/step - loss: 0.0401 - acc: 0.9916 - val_loss: 0.0216 - val_acc: 0.9943\n",
      "Epoch 20/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0403 - acc: 0.9922 - val_loss: 0.0218 - val_acc: 0.9945\n",
      "Epoch 21/60\n",
      "591/591 [==============================] - 325s 549ms/step - loss: 0.0599 - acc: 0.9857 - val_loss: 0.0428 - val_acc: 0.9874\n",
      "Epoch 22/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0564 - acc: 0.9873 - val_loss: 0.0326 - val_acc: 0.9917\n",
      "Epoch 23/60\n",
      "591/591 [==============================] - 330s 558ms/step - loss: 0.0561 - acc: 0.9872 - val_loss: 0.0226 - val_acc: 0.9940\n",
      "Epoch 24/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0530 - acc: 0.9879 - val_loss: 0.0299 - val_acc: 0.9917\n",
      "Epoch 25/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0458 - acc: 0.9907 - val_loss: 0.0236 - val_acc: 0.9943\n",
      "Epoch 26/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0440 - acc: 0.9905 - val_loss: 0.0233 - val_acc: 0.9938\n",
      "Epoch 27/60\n",
      "591/591 [==============================] - 322s 546ms/step - loss: 0.0391 - acc: 0.9922 - val_loss: 0.0237 - val_acc: 0.9948\n",
      "Epoch 28/60\n",
      "591/591 [==============================] - 325s 549ms/step - loss: 0.0378 - acc: 0.9925 - val_loss: 0.0221 - val_acc: 0.9943\n",
      "Epoch 29/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0373 - acc: 0.9927 - val_loss: 0.0212 - val_acc: 0.9948\n",
      "Epoch 30/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0363 - acc: 0.9931 - val_loss: 0.0210 - val_acc: 0.9948\n",
      "Epoch 31/60\n",
      "591/591 [==============================] - 327s 554ms/step - loss: 0.0521 - acc: 0.9887 - val_loss: 0.0276 - val_acc: 0.9931\n",
      "Epoch 32/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0526 - acc: 0.9884 - val_loss: 0.0266 - val_acc: 0.9921\n",
      "Epoch 33/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0473 - acc: 0.9896 - val_loss: 0.0306 - val_acc: 0.9914\n",
      "Epoch 34/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0477 - acc: 0.9898 - val_loss: 0.0265 - val_acc: 0.9933\n",
      "Epoch 35/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0440 - acc: 0.9909 - val_loss: 0.0260 - val_acc: 0.9945\n",
      "Epoch 36/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0382 - acc: 0.9923 - val_loss: 0.0223 - val_acc: 0.9943\n",
      "Epoch 37/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0381 - acc: 0.9928 - val_loss: 0.0212 - val_acc: 0.9955\n",
      "Epoch 38/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0336 - acc: 0.9939 - val_loss: 0.0213 - val_acc: 0.9948\n",
      "Epoch 39/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0344 - acc: 0.9935 - val_loss: 0.0217 - val_acc: 0.9950\n",
      "Epoch 40/60\n",
      "591/591 [==============================] - 352s 596ms/step - loss: 0.0329 - acc: 0.9935 - val_loss: 0.0216 - val_acc: 0.9950\n",
      "Epoch 41/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0470 - acc: 0.9896 - val_loss: 0.0294 - val_acc: 0.9929\n",
      "Epoch 42/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0463 - acc: 0.9900 - val_loss: 0.0251 - val_acc: 0.9940\n",
      "Epoch 43/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0440 - acc: 0.9911 - val_loss: 0.0264 - val_acc: 0.9931\n",
      "Epoch 44/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0434 - acc: 0.9909 - val_loss: 0.0251 - val_acc: 0.9936\n",
      "Epoch 45/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0401 - acc: 0.9924 - val_loss: 0.0236 - val_acc: 0.9950\n",
      "Epoch 46/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0361 - acc: 0.9933 - val_loss: 0.0249 - val_acc: 0.9936\n",
      "Epoch 47/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0341 - acc: 0.9933 - val_loss: 0.0223 - val_acc: 0.9952\n",
      "Epoch 48/60\n",
      "591/591 [==============================] - 325s 551ms/step - loss: 0.0325 - acc: 0.9944 - val_loss: 0.0210 - val_acc: 0.9950\n",
      "Epoch 49/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0316 - acc: 0.9944 - val_loss: 0.0206 - val_acc: 0.9952\n",
      "Epoch 50/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0297 - acc: 0.9953 - val_loss: 0.0206 - val_acc: 0.9952\n",
      "Epoch 51/60\n",
      "591/591 [==============================] - 323s 546ms/step - loss: 0.0434 - acc: 0.9909 - val_loss: 0.0318 - val_acc: 0.9919\n",
      "Epoch 52/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0418 - acc: 0.9917 - val_loss: 0.0276 - val_acc: 0.9931\n",
      "Epoch 53/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0414 - acc: 0.9913 - val_loss: 0.0269 - val_acc: 0.9938\n",
      "Epoch 54/60\n",
      "591/591 [==============================] - 327s 554ms/step - loss: 0.0383 - acc: 0.9922 - val_loss: 0.0238 - val_acc: 0.9952\n",
      "Epoch 55/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0354 - acc: 0.9934 - val_loss: 0.0227 - val_acc: 0.9952\n",
      "Epoch 56/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0339 - acc: 0.9940 - val_loss: 0.0249 - val_acc: 0.9945\n",
      "Epoch 57/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0337 - acc: 0.9939 - val_loss: 0.0211 - val_acc: 0.9955\n",
      "Epoch 58/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0318 - acc: 0.9943 - val_loss: 0.0219 - val_acc: 0.9955\n",
      "Epoch 59/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0303 - acc: 0.9944 - val_loss: 0.0216 - val_acc: 0.9952\n",
      "Epoch 60/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0291 - acc: 0.9949 - val_loss: 0.0212 - val_acc: 0.9955\n"
     ]
    }
   ],
   "source": [
    "batch_size = 64\n",
    "epochs = 60\n",
    "examples = y_train.shape[0]\n",
    "time_id = time.strftime(\"%Y-%m-%d_%H-%M-%S\")\n",
    "log_filename = time_id + '.csv'\n",
    "\n",
    "# setup callbacks\n",
    "def exp_decay(epoch, lr, rate=0.94, min_lr=0.001):\n",
    "    \"\"\"Exponential decay every 2 epochs\"\"\"\n",
    "    if epoch % 2 or epoch == 0:\n",
    "        return lr\n",
    "    return max(lr * rate, min_lr)\n",
    "\n",
    "def step_decay(epoch, lr, step_size=0.001, min_lr=0.001):\n",
    "    \"\"\"Decay by step_size every epoch\"\"\"\n",
    "    if epoch == 0:\n",
    "        return lr\n",
    "    return max(lr - step_size, min_lr)\n",
    "\n",
    "#annealer = None\n",
    "#annealer = LinearDecay(rate_of_change=3e-06, min_lr=1e-04)\n",
    "#annealer = LinearDecay(M=(examples // batch_size + 1) * epochs, end_lr=0.035, min_lr=0.001)\n",
    "#annealer = keras.callbacks.ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=4, min_lr=0.0001, verbose=1)\n",
    "#annealer = keras.callbacks.LearningRateScheduler(step_decay, verbose=0)\n",
    "annealer = CosineAnneal(max_lr=0.01, min_lr=0.0, T=10, T_mul=1)\n",
    "chkpt = keras.callbacks.ModelCheckpoint('./mnist-inception-chkpt-{epoch:02d}-{val_acc:.4f}_' + time_id + '.hdf5',\n",
    "                                        save_best_only=True, monitor='val_acc', verbose=False)\n",
    "chkpt.best = 0.995 # don't save models with val_acc under this threshold\n",
    "csv_logger = keras.callbacks.CSVLogger('logs/' + log_filename, append=True)\n",
    "\n",
    "preprocessing_function = lambda x: elastic_transform(x, alpha_range=[8, 10], sigma=3)\n",
    "\n",
    "# define data augmentations\n",
    "datagen = ImageDataGenerator(\n",
    "    #rotation_range=15,\n",
    "    width_shift_range=2,\n",
    "    height_shift_range=2,\n",
    "    #shear_range=15,\n",
    "    #zoom_range=0.1,\n",
    "    preprocessing_function=preprocessing_function\n",
    ")\n",
    "\n",
    "history = model.fit_generator(\n",
    "    datagen.flow(X_train, y_train_smooth, batch_size=batch_size, shuffle=True),\n",
    "    epochs=epochs,\n",
    "    steps_per_epoch=591,\n",
    "    validation_data=(X_val, y_val),\n",
    "    callbacks=[annealer, chkpt, csv_logger]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0403 - acc: 0.9918 - val_loss: 0.0282 - val_acc: 0.9938\n",
      "Epoch 2/60\n",
      "591/591 [==============================] - 319s 541ms/step - loss: 0.0402 - acc: 0.9921 - val_loss: 0.0233 - val_acc: 0.9945\n",
      "Epoch 3/60\n",
      "591/591 [==============================] - 319s 539ms/step - loss: 0.0381 - acc: 0.9927 - val_loss: 0.0249 - val_acc: 0.9948\n",
      "Epoch 4/60\n",
      "591/591 [==============================] - 317s 536ms/step - loss: 0.0382 - acc: 0.9924 - val_loss: 0.0249 - val_acc: 0.9940\n",
      "Epoch 5/60\n",
      "591/591 [==============================] - 320s 541ms/step - loss: 0.0351 - acc: 0.9939 - val_loss: 0.0259 - val_acc: 0.9952\n",
      "Epoch 6/60\n",
      "591/591 [==============================] - 320s 541ms/step - loss: 0.0325 - acc: 0.9944 - val_loss: 0.0267 - val_acc: 0.9943\n",
      "Epoch 7/60\n",
      "591/591 [==============================] - 340s 575ms/step - loss: 0.0297 - acc: 0.9947 - val_loss: 0.0235 - val_acc: 0.9938\n",
      "Epoch 8/60\n",
      "591/591 [==============================] - 348s 589ms/step - loss: 0.0294 - acc: 0.9951 - val_loss: 0.0234 - val_acc: 0.9952\n",
      "Epoch 9/60\n",
      "591/591 [==============================] - 388s 657ms/step - loss: 0.0282 - acc: 0.9954 - val_loss: 0.0236 - val_acc: 0.9955\n",
      "Epoch 10/60\n",
      "591/591 [==============================] - 399s 675ms/step - loss: 0.0275 - acc: 0.9957 - val_loss: 0.0233 - val_acc: 0.9957\n",
      "Epoch 11/60\n",
      "591/591 [==============================] - 343s 581ms/step - loss: 0.0368 - acc: 0.9927 - val_loss: 0.0258 - val_acc: 0.9952\n",
      "Epoch 12/60\n",
      "591/591 [==============================] - 335s 566ms/step - loss: 0.0379 - acc: 0.9928 - val_loss: 0.0298 - val_acc: 0.9938\n",
      "Epoch 13/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0379 - acc: 0.9924 - val_loss: 0.0237 - val_acc: 0.9936\n",
      "Epoch 14/60\n",
      "591/591 [==============================] - 320s 542ms/step - loss: 0.0346 - acc: 0.9935 - val_loss: 0.0235 - val_acc: 0.9931\n",
      "Epoch 15/60\n",
      "591/591 [==============================] - 321s 544ms/step - loss: 0.0338 - acc: 0.9937 - val_loss: 0.0264 - val_acc: 0.9936\n",
      "Epoch 16/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0320 - acc: 0.9946 - val_loss: 0.0250 - val_acc: 0.9938\n",
      "Epoch 17/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0294 - acc: 0.9949 - val_loss: 0.0252 - val_acc: 0.9950\n",
      "Epoch 18/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0288 - acc: 0.9948 - val_loss: 0.0230 - val_acc: 0.9950\n",
      "Epoch 19/60\n",
      "591/591 [==============================] - 320s 542ms/step - loss: 0.0272 - acc: 0.9957 - val_loss: 0.0235 - val_acc: 0.9952\n",
      "Epoch 20/60\n",
      "591/591 [==============================] - 321s 544ms/step - loss: 0.0267 - acc: 0.9958 - val_loss: 0.0237 - val_acc: 0.9952\n",
      "Epoch 21/60\n",
      "591/591 [==============================] - 323s 547ms/step - loss: 0.0361 - acc: 0.9927 - val_loss: 0.0338 - val_acc: 0.9926\n",
      "Epoch 22/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0376 - acc: 0.9927 - val_loss: 0.0222 - val_acc: 0.9945\n",
      "Epoch 23/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0385 - acc: 0.9924 - val_loss: 0.0305 - val_acc: 0.9929\n",
      "Epoch 24/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0323 - acc: 0.9945 - val_loss: 0.0242 - val_acc: 0.9950\n",
      "Epoch 25/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0332 - acc: 0.9941 - val_loss: 0.0239 - val_acc: 0.9950\n",
      "Epoch 26/60\n",
      "591/591 [==============================] - 324s 547ms/step - loss: 0.0297 - acc: 0.9948 - val_loss: 0.0240 - val_acc: 0.9943\n",
      "Epoch 27/60\n",
      "591/591 [==============================] - 322s 546ms/step - loss: 0.0275 - acc: 0.9957 - val_loss: 0.0228 - val_acc: 0.9955\n",
      "Epoch 28/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0276 - acc: 0.9957 - val_loss: 0.0224 - val_acc: 0.9957\n",
      "Epoch 29/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0262 - acc: 0.9961 - val_loss: 0.0226 - val_acc: 0.9962\n",
      "Epoch 30/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0264 - acc: 0.9962 - val_loss: 0.0228 - val_acc: 0.9962\n",
      "Epoch 31/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0355 - acc: 0.9930 - val_loss: 0.0319 - val_acc: 0.9905\n",
      "Epoch 32/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0354 - acc: 0.9934 - val_loss: 0.0220 - val_acc: 0.9940\n",
      "Epoch 33/60\n",
      "591/591 [==============================] - 322s 545ms/step - loss: 0.0339 - acc: 0.9935 - val_loss: 0.0245 - val_acc: 0.9943\n",
      "Epoch 34/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0322 - acc: 0.9941 - val_loss: 0.0236 - val_acc: 0.9948\n",
      "Epoch 35/60\n",
      "591/591 [==============================] - 324s 549ms/step - loss: 0.0291 - acc: 0.9953 - val_loss: 0.0218 - val_acc: 0.9950\n",
      "Epoch 36/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0294 - acc: 0.9953 - val_loss: 0.0267 - val_acc: 0.9943\n",
      "Epoch 37/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0276 - acc: 0.9959 - val_loss: 0.0223 - val_acc: 0.9960\n",
      "Epoch 38/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0254 - acc: 0.9963 - val_loss: 0.0223 - val_acc: 0.9952\n",
      "Epoch 39/60\n",
      "591/591 [==============================] - 325s 551ms/step - loss: 0.0258 - acc: 0.9961 - val_loss: 0.0223 - val_acc: 0.9955\n",
      "Epoch 40/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0263 - acc: 0.9957 - val_loss: 0.0221 - val_acc: 0.9957\n",
      "Epoch 41/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0343 - acc: 0.9932 - val_loss: 0.0276 - val_acc: 0.9931\n",
      "Epoch 42/60\n",
      "591/591 [==============================] - 325s 549ms/step - loss: 0.0352 - acc: 0.9933 - val_loss: 0.0211 - val_acc: 0.9943\n",
      "Epoch 43/60\n",
      "591/591 [==============================] - 324s 548ms/step - loss: 0.0322 - acc: 0.9942 - val_loss: 0.0222 - val_acc: 0.9957\n",
      "Epoch 44/60\n",
      "591/591 [==============================] - 329s 556ms/step - loss: 0.0320 - acc: 0.9940 - val_loss: 0.0225 - val_acc: 0.9943\n",
      "Epoch 45/60\n",
      "591/591 [==============================] - 329s 556ms/step - loss: 0.0296 - acc: 0.9950 - val_loss: 0.0270 - val_acc: 0.9936\n",
      "Epoch 46/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0284 - acc: 0.9951 - val_loss: 0.0213 - val_acc: 0.9955\n",
      "Epoch 47/60\n",
      "591/591 [==============================] - 326s 551ms/step - loss: 0.0283 - acc: 0.9951 - val_loss: 0.0221 - val_acc: 0.9952\n",
      "Epoch 48/60\n",
      "591/591 [==============================] - 325s 551ms/step - loss: 0.0255 - acc: 0.9961 - val_loss: 0.0223 - val_acc: 0.9950\n",
      "Epoch 49/60\n",
      "591/591 [==============================] - 325s 550ms/step - loss: 0.0252 - acc: 0.9961 - val_loss: 0.0217 - val_acc: 0.9960\n",
      "Epoch 50/60\n",
      "591/591 [==============================] - 327s 553ms/step - loss: 0.0246 - acc: 0.9966 - val_loss: 0.0217 - val_acc: 0.9962\n",
      "Epoch 51/60\n",
      "591/591 [==============================] - 327s 552ms/step - loss: 0.0322 - acc: 0.9943 - val_loss: 0.0256 - val_acc: 0.9940\n",
      "Epoch 52/60\n",
      "591/591 [==============================] - 326s 552ms/step - loss: 0.0350 - acc: 0.9935 - val_loss: 0.0226 - val_acc: 0.9952\n",
      "Epoch 53/60\n",
      "591/591 [==============================] - 318s 538ms/step - loss: 0.0326 - acc: 0.9943 - val_loss: 0.0256 - val_acc: 0.9940\n",
      "Epoch 54/60\n",
      "591/591 [==============================] - 318s 538ms/step - loss: 0.0311 - acc: 0.9947 - val_loss: 0.0288 - val_acc: 0.9938\n",
      "Epoch 55/60\n",
      "591/591 [==============================] - 318s 537ms/step - loss: 0.0290 - acc: 0.9950 - val_loss: 0.0274 - val_acc: 0.9940\n",
      "Epoch 56/60\n",
      "591/591 [==============================] - 322s 544ms/step - loss: 0.0289 - acc: 0.9954 - val_loss: 0.0284 - val_acc: 0.9931\n",
      "Epoch 57/60\n",
      "591/591 [==============================] - 319s 539ms/step - loss: 0.0261 - acc: 0.9961 - val_loss: 0.0260 - val_acc: 0.9933\n",
      "Epoch 58/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0262 - acc: 0.9961 - val_loss: 0.0260 - val_acc: 0.9936\n",
      "Epoch 59/60\n",
      "591/591 [==============================] - 320s 542ms/step - loss: 0.0237 - acc: 0.9969 - val_loss: 0.0248 - val_acc: 0.9940\n",
      "Epoch 60/60\n",
      "591/591 [==============================] - 321s 543ms/step - loss: 0.0246 - acc: 0.9965 - val_loss: 0.0246 - val_acc: 0.9936\n"
     ]
    }
   ],
   "source": [
    "history = model.fit_generator(\n",
    "    datagen.flow(X_train, y_train_smooth, batch_size=batch_size, shuffle=True),\n",
    "    epochs=epochs,\n",
    "    steps_per_epoch=591,\n",
    "    validation_data=(X_val, y_val),\n",
    "    callbacks=[annealer, chkpt, csv_logger]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Max Validation Accuracy: 0.9961904798235212\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAscAAACsCAYAAACXWgGeAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzs3Xd4XFeZ+PHvmaqZ0aj3YltyL3GPE6c4lSUmjQTSCBASIJSEhAV+lGVpAXZhlxIChCWQAmzABEjHaes4cZrjHse2LPciy1bvM5p6fn+cK1mS1SxrJEt5P88zz7Q797x37tw77z333HOU1hohhBBCCCEE2EY7ACGEEEIIIU4XkhwLIYQQQghhkeRYCCGEEEIIiyTHQgghhBBCWCQ5FkIIIYQQwiLJsRBCCCGEEBZJjoUQQgghhLBIciyEEEIIIYRFkmMhhBBCCCEsjtEO4GRlZGTo0tLSESkrFArhdrulrDFUnpQlZZ0OZY10eVKWlDWSZdXV1QGQmZmZ8LL6M9a/Rylr5MvauHFjrdY6e8AZaK3HxA24EnhgwoQJeqTs2LFDyhpj5UlZUtbpUNZIlydlSVkjWdbDDz+sH3744REpqz9j/XuUska+LGCDHkTOOWaaVWitn9Fa3+73+0c7FCGEEEIIMU4lNDlWSl2mlCpXSu1RSn29n+k+rJTSSqnFiYxHCCGEEEKI/iQsOVZK2YFfA8uBWcBNSqlZvUznB+4C3h5KOS3tETYebKClPXIq4QohhBBiEPLy8sjLyxvtMIRImERekLcE2KO13geglFoBXA3s6DHd94H/Ar4ylELerWjiI79/m798+myWTs4c+ANCCCGEGLLLLrtstEMYNyKRCBUVFbS3t3d7raysbMTKH49lRaNRIpEITqdzSJ9PZHJcCBzu8rwCOKvrBEqpBUCx1vpZpVSfybFS6nbgdoD8/Pxu76V5XQA0BcPDErQQQgghxEioqKjA7/czadIklFIABINBPB7PiJQ/HsvSWnP06FEqKiooKSkZ0jwS2eZY9fKa7nxTKRvwc+DLA81Ia/2A1nqx1npxRkZGt/fSvOaooDEgzSqEEEKIRHv88cd5/PHHRzuMcaG9vZ3MzMzOxFicOqUUaWlp3WrjT1Yia44rgOIuz4uAyi7P/cAc4BXrR5EHPK2UukprvWGwhXQmx0FJjoUQQohEa25uHu0QxhVJjIffqX6niaw5Xg9MVUqVKKVcwI3A0x1vaq2btNZZWutJWutJwFrgpBJjAI/Tjstuk5pjIYQQQghxyhKWHGuto8CdwAtAGfCY1nq7UuoepdRVw1WOUopUr5PGgLQ5FkIIIYQYrLq6OubPn8/8+fPJy8ujsLCw83k4PLi86tZbb6W8vDzBkY6shA4frbVeCazs8dq3+5j2wqGWk+51Ss2xEEIIIcRJyMzMZMuWLQB897vfJTk5ma98pXv/CJ2jxtl6r099+OGHEx7nSEtocjyclFJXAldOmDDhhPfSPC4apbcKIYQQIuGKiopGO4Rx6XvPbGdHZTPxeLzPRPRkzSpI4TtXzj7pz+3Zs4cPfvCDnHfeebz99ts8++yzfO9732PTpk0Eg0FuuOEGvv1tU9d53nnn8atf/Yo5c+aQlZXFZz/7WZ577jm8Xi9PPfUUOTk5w7IsI2lcDB+dKjXHQgghxIi49NJLufTSS0c7DJFgO3bs4JOf/CSbN2+msLCQH/3oR2zYsIF33nmHl156iR07eg5bAU1NTVxwwQW88847LF26lIceemgUIj91Y6bmuD9pHifbpLcKIYQQQoxRHTW8I9n3cH8mT57MmWee2fn8L3/5Cw8++CDRaJTKykp27NjBrFndBz72eDwsX74cgEWLFvHaa6+NaMzDZXwkx1JzLIQQQoyIxx57DIDrr79+lCMRieTz+Tof7969m1/84hesW7eOtLQ0PvrRj/baj7DL5ep8bLfbiUajIxLrcBszzSr6k+Z1EYzEaI/ERjsUIYQQYlwLBAIEAoHRDkOMoObmZvx+PykpKRw9epQXXnhhtENKqHFTcwzQFIyQ5LSPcjRCCCGEEOPHwoULmTVrFnPmzKG0tJRzzz13tENKqDGTHA/UWwWYIaRzU5JGODIhhBBCiLHtu9/9bufjKVOmdHbxBmZMiT/96U+9fu7111/vfNzY2Nj5+MYbb+TGG28c/kBHwJhpVtFfbxWdQ0jLQCBCCCGEEOIUjJma4/6keqzkWHqsEEIIIRKqpKRktEMQIqHGRXIsNcdCCCHEyLjgggtGOwQhEmrMNKvoT7r3eJtjIYQQQgghhmpcJMdelx2nXUmzCiGEECLBHn30UR599NHRDkOIhBkzybFS6kql1AMtLS29vUeqxyU1x0IIIUSCRSIRIhH5vxXj15hJjvvrrQJMu+OmoLQ5FkIIIYQYjPe///0nDOhx77338vnPf77PzyQnJwNQWVnJhz/84V6nufDCC9mwYUO/Zd97773dBpP5wAc+0K0ruNE0ZpLjgaR5ZAhpIYQQQojBuu6661ixYkW311asWMFNN9004GcLCgr4+9//PuSyeybHK1euJC0tbcjzG07jorcKMENIH2kMjnYYQgghhBAn77mvw7F3ccVjYBum0X7zzoDlP+rz7WuuuYZ77rmHUCiE2+3mwIEDVFZWMn/+fC655BIaGhqIRCL84Ac/4Oqrr+722QMHDnDFFVewbds2gsEgt956Kzt27GDmzJkEg8fzsc997nOsX7+eQCDAddddx/e+9z3uu+8+Kisrueiii8jKymL16tVMmjSJDRs2kJWVxc9+9jMeeughAD71qU/xxS9+kQMHDrB8+XLOO+883nzzTQoLC3nqqafweDzD8111MX5qjr1OmqQrNyGEECKhpk2bxrRp00Y7DDEMMjMzWbJkCc8//zxgao1vuOEGPB4PTzzxBJs2bWL16tV8+ctfRmvd53x+85vf4PV62bp1K9/85jfZuHFj53s//OEP2bBhA+vWrePVV19l69at3HXXXRQUFLB69WpWr17dbV4bN27k4Ycf5u2332bt2rX87ne/Y/PmzQDs3r2bO+64g+3bt5OWlsY//vGPBHwr46nm2OOU3iqEEEKIBDvnnHNGO4TxyarhDQeDCakN7ctNN93EihUruPrqq1mxYgUPPfQQWmv+7d/+jTVr1mCz2Thy5AhVVVXk5eX1Oo81a9Zw1113ATB37lzmzp3b+d5jjz3GAw88QCQS4dixY+zYsaPb+z29/vrrXHPNNfh8PgCuvfZaXnvtNa666ipKSkqYP38+AIsWLeLAgQPD9C10N2ZqjvvrrQJMzXEgHCMUjY1wZEIIIYQQY9MHP/hBVq1axaZNmwgGgyxcuJBHH32UmpoaNm7cyJYtW8jNzaW9vb3f+SilTnht//79/OQnP2HVqlWsW7eOyy+/fMD59FdD7Xa7Ox/b7Xai0egASzc0YyY5Hqi3ilRrIJAmqT0WQgghEuaRRx7hkUceGe0wxDBJTk7mwgsv5Lbbbuu8EK+pqYmcnBycTierV6/m4MGD/c5j2bJlnX1fb9u2ja1btwLQ3NyMz+cjNTWVqqoqnnvuuc7P+P1+eqvwXLZsGU8++SSBQIC2tjaeeOIJzj///OFa3EEZN80q0juHkI6Q408a5WiEEEIIIcaGm266iWuvvbaz54qbb76ZK6+8ksWLFzN//nxmzJjR7+c/97nPceuttzJ37lzmz5/PkiVLAJg3bx4LFixg9uzZTJw4kXPPPbfzM7fffjvLly8nPz+/W7vjhQsX8olPfKJzHp/61KdYsGBBwppQ9CahybFS6jLgF4Ad+L3W+kc93v8scAcQA1qB27XWO4ZSVppHhpAWQgghhDhZ11xzTbfmDFlZWbz11lu9Ttva2grApEmT2LZtGwAej+eELuE6dJxlCPZoS/2FL3yBL3zhC53Puya/X/rSl/jSl77UbT5dywP4yle+MoglG5qENatQStmBXwPLgVnATUqpWT0m+7PW+gyt9Xzgv4CfDbW8tM6aY+mxQgghhBBCDE0i2xwvAfZorfdprcPACqBbJ3la6+YuT31A362wB5DqsZJjaXMshBBCCCGGKJHNKgqBw12eVwBn9ZxIKXUH8CXABVzc24yUUrcDtwPk5+f3WlhHzXGTNKsQQgghEmb27NmjHcK4orXutacHMXT99XgxGImsOe5tTZ8Qrdb611rrycDXgH/vbUZa6we01ou11oszMjJ6LSzZ7cBhUzRIswohhBAiYc4880zOPPPM0Q5jXEhKSqKuru6UkzlxnNaaxsZGkpKG3jlDImuOK4DiLs+LgMp+pl8B/GaohSmlSPPKQCBCCCFEIkUi5n/W6XSOciRjX1FRERUVFdTU1HS+FolERuy7Ha9lRaPRAXvY6E8ik+P1wFSlVAlwBLgR+EjXCZRSU7XWu62nlwO7OQWpHqc0qxBCCCESqKM/20984hOjG8g44HQ6KSkp6fZaWVkZM2fOHJHyx3NZp5KIJyw51lpHlVJ3Ai9gunJ7SGu9XSl1D7BBa/00cKdS6lIgAjQAt5xKmWleF41BaVYhhBBCCCGGJqH9HGutVwIre7z27S6P7x7O8tI8To419z8soRBCCCGEEH0ZM8NHK6WuVEo90NtQgx1SvU4ZBEQIIYQQQgzZmEmOtdbPaK1v9/v9fU6T7nXJICBCCCGEEGLIEtqsYqSleZy0hWOEo3FcjjGT9wshhBBjxvz580c7BCESanwlxx0DgQQjZPvdoxyNEEIIMf5IcizGu3FVvZrqdQHQJD1WCCGEEAkRCAQIBAKjHYYQCTOukuM0j6k5lovyhBBCiMR47LHHeOyxx0Y7DCESZswkx4PprSLdqjlukORYCCGEEEIMwZhJjgfTW0VHm2PpsUIIIYQQQgzFmEmOByO1ywV5QgghhBBCnKxxlRz73Q7sNiVtjoUQQgghxJCMq67clFKkepw0Sm8VQgghREIsXrx4tEMQIqHGVXIMpt2x1BwLIYQQiTFnzpzRDkGIhBpUswql1GSllNt6fKFS6i6lVFpiQzshhgF7qwDTnZskx0IIIURiNDU10dTUNNphCJEwg21z/A8gppSaAjwIlAB/TlhUvRhMbxUAaV6XNKsQQgghEuSJJ57giSeeGO0whEiYwSbHca11FLgGuFdr/a9AfuLCGjqpORZCCCGEEEM12OQ4opS6CbgFeNZ6zZmYkE5NqtdJkyTHQgghhBBiCAabHN8KLAV+qLXer5QqAf43cWENXZrHRUsoSiQWH+1QhBBCCCHEGDOo3iq01juAuwCUUumAX2v9o0QG1pNS6krgygkTJvQ7Xbrv+EAgWcnuEYhMCCGEEEKMF4PtreIVpVSKUioDeAd4WCn1s8SG1t1gL8hL9XQMIS1NK4QQQojhtnTpUpYuXTraYQiRMIPt5zhVa92slPoU8LDW+jtKqa2JDGyo0rwuAJqkxwohhBBi2E2fPn20QxAioQbb5tihlMoHruf4BXkDUkpdppQqV0rtUUp9vZf3v6SU2qGU2qqUWqWUmjjYefclTWqOhRBCiISpra2ltrZ2tMMQImEGmxzfA7wA7NVar1dKlQK7+/uAUsoO/BpYDswCblJKzeox2WZgsdZ6LvB34L9OJvjepHklORZCCCES5dlnn+XZZwddTybEmDOo5Fhr/Tet9Vyt9ees5/u01h8a4GNLgD3WtGFgBXB1j/mu1loHrKdrgaKTC/9EHc0qGgLSrEIIIYQQQpycwV6QV6SUekIpVa2UqlJK/UMpNVAiWwgc7vK8wnqtL58Enuuj/NuVUruVUjVHjx7tt1C/24FNmd4qhBBCCCGEOBmDbVbxMPA0UIBJcJ+xXuuP6uU13euESn0UWAz8d2/va60f0FpP1Vpn5+f3PzCfzaZI9TiZv+8B2PSnAUIUQgghhBDiuMEmx9la64e11lHr9giQPcBnKoDiLs+LgMqeEymlLgW+CVyltQ4NMp5+pXmcnF39GGw+LccpEUIIIYQQp6nBduVWa9Xu/sV6fhNQN8Bn1gNTrdH0jgA3Ah/pOoFSagHwW+AyrXX1oKMewISkAL62ZqjfO1yzFEIIIQSwbNmy0Q5BiIQabHJ8G/Ar4OeYphFvYoaU7pPWOqqUuhPTy4UdeEhrvV0pdQ+wQWv9NKYZRTLwN6UUwCGt9VVDWpIuZjqsCuq2GmhvhqSUU52lEEIIIYDS0tLRDkGIhBrs8NGHgG5Jq1Lqi8C9A3xuJbCyx2vf7vL40kFHehImqy6tN+r3QcH8RBQjhBBCvOccO3YMgLy8vFGORIjEGGyb4958adiiGASl1JVKqQdaWloGnLY41qWTjPp9CYxKCCGEeG95/vnnef7550c7DCES5lSS4956o0gYrfUzWuvb/X7/gNPmRw6xJ15gnki7YyGEEEIIMUinkhz32i3b6SAreICtupR4ch7U7x/tcIQQQgghxBjRb5tjpVQLvSfBCvAkJKJTFWrBF6piT/wCQikxPHVScyyEEEIIIQan3+RYaz1wG4bTTe0uAPbqQtp8Ck/l6lEOSAghhBBCjBWD7cpt7KgxyfEeXUCj105WW7V05yaEEEIMk0suuWS0QxAioU6lzfGIGnRvFbXlaJuDgzqXWmehea1B2h0LIYQQw6G4uJji4uKBJxRijBozyfGge6uoKSeeXkoUB0ftVnIs7Y6FEEKIYXH48GEOHz488IRCjFFjJjketJpybNnTUQoOK6uDcunrWAghhBgWq1atYtWqVaMdhhAJM76S42gIGvajsqeT6nFSG7KDP1+SYyGEEEIIMSjjKzmu2ws6DtnTSfM4qW0NQUapJMdCCCGEEGJQxldyXFtu7rOmsaQkg1Vl1QT9E6XNsRBCCCGEGJQxkxwPqrcKqxs3sqZy50VTicU1b9SnQFs1hAbo5UIIIYQQQrznjZnkeFC9VdSWQ+oEcPmYkOnlusVFPHnIGshPmlYIIYQQp+yyyy7jsssuG+0whEiYMZMcD0rNLsie3vn0joumcFDnmieSHAshhBCnLC8vj7y8vNEOQ4iEGT/JcTwGdbu7JcdF6V4WL1wEQFPFztGKTAghhBg39u3bx759UuEkxq/xkxw3HoJoO2RN6/by7ZeeQZVOZ/fOraMUmBBCCDF+rFmzhjVr1ox2GEIkzPhJjmuti/G61BwD5Kd6CPonouv2crCubRQCE0IIIYQQY8WYSY4H7K2i5ng3bj3lTpzJRFXFL1/ek8AIhRBCCCHEWDdmkuMBe6uoLQdfNngzTnjLkzeNHNXIC5t2s6+mNcGRCiGEEEKIsWrMJMcDqtkFWdN7fy+jFIApjlruW7V7BIMSQgghhBBjSUKTY6XUZUqpcqXUHqXU13t5f5lSapNSKqqU+vCQC9La1Bxnn9ikAoDMyQB8fHqUp96pZMW6Q0MuSgghhHgvu+KKK7jiiitGOwwhEiZhybFSyg78GlgOzAJuUkrN6jHZIeATwJ9PqbDWamhvGrDm+PLCAMumZvP1x9/ll6t2o7U+pWKFEEKI95qsrCyysrJGOwwhEiaRNcdLgD1a631a6zCwAri66wRa6wNa661A/JRKqrH6MO6r5tjlg+Q8XE0H+f0ti7l2YSE/fWkX33pqG7G4JMhCCCHEYJWXl1NeXj7aYQiRMI4EzrsQONzleQVw1lBmpJS6HbgdID8//8QJOrpx66vmGEzTivq9OO02fnrdPHL8SfzPq3upbQlz743zSXLahxKaEEII8Z7y1ltvATB9ej//uUKMYYmsOVa9vDakalqt9QNa68Va68UZGSf2RkFNObj8kFLQ90wySjqHkFZK8fXlM/j2FbN4YccxPv7gOpoCkaGEJoQQQgghxpFEJscVQHGX50VAZUJKqi2HrKmgesvHLRmTobUKQsf7Sb7tvBLuu3EBWw43csWvXuOdw40JCU8IIYQQQowNiUyO1wNTlVIlSikXcCPwdEJKqtl1wsh4J7AuyqN+f7eXr5xXwIrPnE08Dh/+nzd58PX9cqGeEEIIIcR7VMKSY611FLgTeAEoAx7TWm9XSt2jlLoKQCl1plKqArgO+K1SavtJF9TeBK3Heh0ZrxurOzfq957w1sIJ6ay863wump7D95/dwaf/uJHGQPikQxFCCCGEEGNbIi/IQ2u9EljZ47Vvd3m8HtPcYuhqrIvxBqo5Ti8x91a7455SvU5++7FFPPLmAf5jZRkf+MVrfPHsdGbOPKXohkxrzYOv78ftsHHJzFwK0jyjE4gQQgjRxTXXXDPaIQiRUAlNjoeTUupK4MoJEyZ0f8ObAefeDfnz+p+BOxmS86Cu9+TYKoNbzy1h0cR07vzzZr76wlH+tO11bjizmKvmF5CS5Dz1BRmkZ7ce5Qf/LAPgW09tZ2Z+CpfOzOGSmbnMLUzFZuunfbUQQgiRIKmpqaMdghAJNWaSY631M8Azc+bM+XS3NzInw/vuGdxMMkr7rDnuam5RGv+86zx+vXIjrxwK8+9PbuMH/9zBB+bkc8OZxSwpyUD1d/HfKWoKRPjeM9uZV5TKT66bx+ryav6vrJpfr97DL1/eQ5rXycy8FGbmpzAz38/M/BSm5ibjdkh3dEIIIRJr27ZtAMyZM2eUIxEiMcZMcjwsMkth90uDmtSf5OTqmal87ZoZvHukiRXrD/PMlkoe33yEWfkpfOaCUi4/Ix+Hffibbf/HyjIaAhH+eNtZTM31MzXXz+3LJtPQFuaVXdW8va+esmMt/HndQdojZvwUp13xL7PzuO3cEhZOSEto8i6EEOK9a8OGDYAkx2L8em8lxxmlpju37U/AzKvANnBNq1KKuUVpzC1K41uXz+KZdyr57Zq93L1iCz95sZxPn1/KdYuK8biGp9b2rb11/HXDYT57wWRmFaR0ey/d5+KaBUVcs8A0047FNQfq2ig72szGgw38fWMF/9x6lHnFadx27iSWz8nH5UhkhyRCCCGEEOPLeytzmn2tSZD/9gn41Zmw4WGItA/64x6XnevPLOalf72ABz62iOxkN99+ajvn/vhl7lu1m6bgqQ0k0h6J8c0n3mVChpe7L5k64PR2m2JydjJXzC3gO1fOZu03LuH7V8+mJRjh7hVbOP+/XuanL5bzzuFG4qfRMNnRWJza1tBohyGEEEIIcYL3WM1xCdy5AcqegTfuhWe/CK/8J5z1WTjrM+DyDWo2NptpwvC+WblsONjAb17Zy89e2sXv1uzjlnMmcdt5JWT4XCcd3v2r97Cvto0/fXLJkGqifW4HH1s6iZvPmsiru2p46I39ne2Us/1uLp6ew8Uzc8iOxU963sOlrjXEjQ+sZXd1K9l+N3MKUphdkMrsghSm5fnxOO3YbQqbUthtCrtS+JMccgGiEEIIIUbEmEmO++yt4mTZ7DD7gzDrati/xiTJq74H+16Bjz0xqKYWXWLizEkZnPmJDLZXNvHr1Xv49St7ePD1/Xz07Al86vxSXHYbta0halpD1LWGqWsNkeV3c87krG4J9MGGML959QjXLijk/KnZp7aINsVFM3K4aEYO9W1hXt1VzaqyalZuO8pfNxzGYYOFbzZz9uRMzi7NYOGEdJKcib+Yr7k9wi0Pr+NQfYAvXjqVQ/UBdlQ2s2Z3LbF+araTnDam5viZmpvM9Fw/03L9zC5IISclKeExnyqtNbG4TkjbdCGEEEIMvzGTHPfZW8VQKQWlF5jbpj/C01+Al78Pl353SLObXZDK/TcvYndVC/e/spcHX9/P717b3+9nZuWncO6UTM6ZksV9b9WQ7HbwzcuHt2PljC7tlCOxOBsONPCPN8vY3RTjVy/v5r5V4LLbmD8hjbNLM1lamsmCCWnDniy3R2J86g8b2Hm0hd99fDEXzcjp9l75sRb21rQSicWJxSEWjxOLa6JxzdGmdnZVtfD67loe33QEMKvvvClZ3HBmMe+blXta9NTx57cP8bvX9hEIRwlF44QicULRGHENU3OSuXhGDhdOz2HxpHSckiwLIcao66+/frRDECKhxkxynFALPw5HNsHrP4eChTDrqiHPamqun5/fMJ+7L5nKP989itdlJzPZTVayi+xkNxk+FwfrA7y5p5Y39tTxhzcPdibRP71uHpnJ7uFaqhM47TaWTs4kLZzJzJkzaW6PsOFAPWv31fPW3jorWd6N22Fj0cR0lpZmcs6UTOYWpZ1SMheJaT73vxtZf6CeX9y4oFtiDJDktDOvOI15xWkDzqsxEGZXVStv7q3lbxsquPPPm0n3OrlmQRE3LikecoynIhbX/OfKMn7/+n4WTkjjrJIM3A4bbqcdl92GTcHGQw089MZ+frtmH/4kB8umZnNGUSqxuCYSixOJxYnGTC1zaXYyc4tSmZbrPy0uqIzFNf9XVkU4GsfjtON12fFYtwyf+V1L7yhCvHd4vd7RDkGIhJLkuMPyH8Oxd+HJz0H2DMgeYDjqAUzK8nHHRVN6fS8z2c3CCencefFUguEYGw7W8075fq5dWNh9wqrtkJwLvqxTiqUvKUlOLp6Ry8UzcgFoCkZYt98kym/tq+OnL+3ipy9BstvB2aWZnD81i/OmZlGa5Rt0MhSLa376ejWvHmjjP645g6vmFUA0DGhwnPyBQJrXxZKSDJaUZPCFi6fyxp5a/rr+MH9ae4CH3tiPx6nIT60i2+8mJyWJHL+bSVk+lpZmMjl78HEPViAc5e4VW3hpRxWfOGcS37piFvY+2ke3hqK8vruW1TurWV1ezT/fPdr5nt2mcNoVWkMoatqEuxw2ZuanMLcwlQJXkMKSyIgORANQ0xLi7hWbeXNvXZ/TuBw2itI8FKZ7KEr3UpCahMdlx+20m4MEhw23w05BWhJTc/zD1rPLcNhT3UJta5hsv5usZDcpSQ5J9IUYwJYtWwCYP3/+KEciRGJIctzB4Ybr/wi/XQZ/vRk+/fKIFOtx2Tl/ajZZ0druf8plz8JjHwO7C+bdBEvvgKyBe7A4FakeJ++blcv7ZplkuaEtzNp9dby2p5bXd9fyf2VVABSkJjE5J5lMn4vMZDeZyS6yfG6SkxxoDTGt0VoT15rXdtXy6oE2vr58Bh85awKEWuCRK6DpMFz4DVj0CbAPLeGz2xTLpmWzbFo2da0h/vnuUTbuOkzU6aO6uZ2tFY1UNbd39gWdl5LEuVOyOG9qJktLs8g/0Sh+AAAgAElEQVRNObUaz7pAlK/+di3bK5v4zpWzuPXckn6nT3Y7uGxOHpfNyUNrTWsoitNuw2m3dSbUWmsO1QfYWtHEu0ea2FrRyBObj9AaivKzN17i7NJMczHozFzyUhPb5nrDgXru+PMmGgMRfnTtGSyamE4gHCMYiREMxwiEY9S3hahoCFq3AC9WHqOuLdznPJWCkkwfM/L9TM9NwR9rI6soRLY/cWdMehOLa3718h5+sWoXXZu7uxw2spPdpHmdOO02HDZlHbjYcNgVxelepuUmMzXXz/RcP+lDuPB2JOypbmH9gQZsCmzKXOBqs4HdZpavIC2J3JSkEbnWQIw/khyL8W7MJMfDdkFef1IL4bqH4Y8fhCc/D2d8I3Fl9efA6/D326BgAeTOgS1/ho2PwPTlcM4XYMJSiEeh5Sg0VUDTEdN/c84MKD7bDJU9DNJ9Lpafkc/yM/IBOFQX4LU9Nby5t47KxiAH6wLUtYZoC8f6nc8NZ6Tx2QsmQywCj91iaugLFsDKr8Db/2NGOJz+AZM5DVFmspuPL53EmWlBZs483m67I9l8Y08db+ypZdXOKv6xqQIwA6ekelyke52ke12kep343Q6TsDpMQuSym5pPn9tBcpIDf5KZJhbXfP2fRwhE4XcfX8wlM3NPKl6lFP5eaoGVUkzM9DEx08eV8woAk8g9+doWdgWSeHF7Fd96chvfenIbM/L8+NwOFOarUyhQMDHDy9mlmZw9OZPCNM9Jf5daa37/2j5+9NxOitI9PPz5JSf0ud2fcDROezTW2eY6FI0TDMc4XB9g57EWyo+1sKOymee2HUNruGd1FUXpHhZMSGfhhDTmF6dRmpVMiicxtbg1LSG++NfNvLGnjg/OL+DDi4qpbQ2ZC2dbzK0xGCEa10RjcaJx3dmOfOOBBlpC0c55ZSW7mZabzOTsZKbkHL/PTRnZZL9DayjKL/5vFw+9caDfi1w7ZPpc5KUmkex2oAE0aDRam2ZYJdk+puUkM80ajCgreXQPBhoDYVaXV9PQFqGlPUpLu3UfihCNmeXt+MkoFF6XnWl5fqbn+ZmR5ycvJem0PTNQ2RjkSGOQutYw9W1hGgJh6lrDNLdHCISjtIVinfdxrTlnchZXzMtnQbEM+iTEcBszyfGwX5DXl5Jl8L7vwYv/TrbKhJLvgmfgtrDD5uhW+MtNkD4Jbv47eDPg4n+H9b+Hdb+D8pXgzYRgA+heumRTdpN4TjoXJp1vEulhSpYnZHq5OXMiN581sdvrwXCMurYQraGoVUt1vLYqyWmnoXI/aG26ztu7Cq68z7Tz3vU8vPgtWPERmHgu/MsPoHDhsMTaoWuy+ZGzJhCPa3YcbWbd/npqWkM0BsI0tEVoCIQ5VBegLRy12gBrItE44Zi56V7yjEyPncc+s5Q5hanDGnNPdptiVk4SH5o5k69fNoO9Na28sL2KdfvricZNbNpKauJxeKmsir9tNAcAxRkezi7JZHZBCuFYnLaQqfltC0UJhmN43XYyfG4yfS4yfC4yfS7uf7Wa1w+28f7Zufz3dfNOuimHy2EzbaV7VGzPKUztPNAC0yTlubfepcGWyuZDjWw4UM8z71R2vp/ktJGXkkReahJ5KaamMyvZTZbfZe6tmz/JgdthG1SC8OaeWu7+6xaagxF+/KEzuH5x8UklFlprjjW3U36shd1VreyqamF3dStPbj7SLWn2uuykuhV56fVkeM13m5HswudydOmq0GwnboeNbL+bbH8SuSlusv3uk77AVGvNynePcs8zOzjW3M5NS4r5zLLJOB024nGT7Ma1JhyLU9MSorIxyLGmdiqb2jnaFCQYjpmDLBsobChlLpT959aj/LlL/+3pXif5Phsz3w1TkuVlYqaPkiwfk7J8JLsT93fSFIjw4Ov7eOiNA7T2+J791kGro0tzpo7ttSkY4fHNRzpfT/U4mZabTIbPrAuv247X5cDrshNsamJvpJLsZNMkK9vvTugyddhV1cL9L2yBnStZH5/OEY73VuRz2Un1OPG5HXjdDnwuOwVpLsKxOP+79iAPvbGfwjQPl8/N5/Iz8inN9tEWitEWjtIWitIainLwaHBUzs5EY3Ge3FLJ/a/s4XB9ANXjv8FpV6Rb20a6z2W2k2SzD8r2u8lOdpNl3ad6nKdVd57tkRhPbTnCuv0NtIWind93IByjPRKjIM3TeaDccdB8qmcqR1IwHGN/bRsH6trYX9vGvhrz2GW3MSHDS3GGh+IML8UZXorSPKT7XOPuIvMxkxyPqKV3QuVmsrY9DGV/gPx5UHKBSZwnnD3o/pBPWv0++N8PgTsFPva4SYwBknPgon+Dc78I7/wZKjeDvwBSi0xtd2qxSZiPbYUDb8DBN+Ct++GNX4DLD4tuMf04pyWm1t3jslPk6vsCjYZK4NX/gs3/C8u+auIBUxM+5VLY9AdY/Z/wu4tgzofMwUBGad8FttVCdRkULznpdss2m2JOYepJJbRaawLhGC3tUVpDEes+iqu16uQT41jEtCU/stE0MSlabC4C7ef760opxZQcP1Ny/NxxUe/TxOOa8qoW1u6rY+2+um7JMtB5UV2S005bOEpjoPvgNTYF3/zATD51fsnxnXlTBfhywDF8NYdel4NZOUnMnHl8XR9tCvLO4UYqGkzydqy5narmdjYcbKC6JUQ42nsf3TZFZ7LjczlI8zrJSzUJdb51/9b2ev66bR+lWT7+9MklzIjugj983pyFKT7LHEgWLzm+3fVCKUV+qof8VA8XTj9+YanWmpqWEHtqWtlb08b+mjb2VdYQczg42tTOjqPN1LWF+4y/p3Svk2y/uYA3M9lNlnWf7nORkuQgxeM090lOQtE43/m/Y2ys3M+s/BTu/+hCFk5I73Pe03L9g4qh63Lt6jwQaGH7oRre2FPLPzZ1H0Ap2++mJMtHaZZJmCdm+vAnOXDYFA67Dadd4bDZ8LjspFgJ7UAXnLaGY/z8pV089MZ+WtqjLJ+Tx2cvmMyEDC/JSeYsD+EAVG0z+8GU/BPm0RSIsPNYM+VVLew81sKeqlYO1JoD4WDYJJIdTa9Y371tvc9lZ0Kmj5IsL5MyzUFAaZaPrGQ3bqdpS+922Eiy+mc/GftqWrnvpTL82x/lW45/kOlqJmb3UH3m/4Ozbic92dt7s5f6/bD7JVovXsSLdTn8891jPPzGfh5Ys6/Psr7x4lGy/W5m5acwqyCFWfkplGT5KEzzkOZ1DmvSFo3FeWLzEX718m6mN67hl94XcOVksqboM1R7phCLa+IawrEYDYEIDW1hDtcHeOdwI/VtYaK9nPHwOO1MzTVnMKbn+pmW52dqTjKNwRjHmtqJxs0FzdG47jzYTESTocrGIH9ae5C/rDtEYyBCjt9NutfVud/JSnbjctioaAiecMAMZj/l6NJcy+2wk5+aRFG6h8I0j7lP99Ja205Sdhup1nY+El2BxuOaN/bW8se3DrKqrKpbc7PcFDeTMn2EojFeLq+mpuXEQbxSkhxmH+V1kuFzk5viJj81ydpfmkqO5vYYTcGIVUFA53gGp2NirXRvVWKnsTlz5uht27YlvqB4jAOv/ZVJ+hDsexUq1kM8YmpmvZmQlGKS2KQUSEoFT4a5eC45B/x51uNcSCkYVN/Juza9zrTX7oD2ZrjthVO+IJBwAA6/DVsehW2Pm9dmXQVL76Ssxdet6UGiVa78bwrW/QDmfQQ+eH/vzSfam00yv/Z+iIVh8W0mkU62alHicdi32iTSO1eadeFOhZlXmJEPSy8Au5OysrIRW7ZuZQUbYMdT5nfi9JoDKFcyuP0mga/dDRUb4Og7EA12n5HNYZrPFJ9lkmVPumlr7nCb9th2N7uOtTBt/tIhxRmPa+oDYTxOOx6n/YQamGgsTkMgQn1b2JwBqDnCv5w9D6Ih89tZ9wBUbgJHEhQushLJs00i6Uk3o0wGG47fwm1mu/BkmEQzKQ3svR+Hn8z60lrTEopSazV9qG018XY93RwImwOXhkCYo03tHGtqJ9Cl2c+1Cwr5/vsL8b32Q9NUyZ9nDjIrt5jfFJgLcvPmmnXo9Jj14PCAMwnSJkLOTHPwNkBb+Z7LZtrhYyUH5o88FteEojFqWkJUt4Sobm6nqjlEVXO76Re9zeofvS3c7wicXqfiq5fN5KNnFeM4uAY2Pwpt1TBpGUy+CPLn97kOBiUaMgdz7U3sK99GaVEe7cFmauoaqGtopLotyrZwHutbMtldF6a2te925115nHZSPCZR7jhg6/idOh02VpcdozUc5/2zc7n7kmnMyvVCw36zLR3ZYLa3Y9tAW+u4+CzTh/3MqyBtkL3XtDcTqy6nvGwb3sKZVMW8VIZ8VAXhWFM7B+vaOFAX4HB9oNfErYPLbiPd5+x2JibD58LntuN22HFZF6e6HDZeeWc/9oOv8DXHCkrUUSLF5+A8/4uw/kHY/YJZX1fdZypmOhxeD2/90gxi1XHWMHMKzPkQLZOv4sWaNOrbwvjcDnxWsuZ12zlw4CBBdwY7KpvZcbSZPdUtRGLHl8PnslNoJWeZye7Os1DmjJSZzp/ktGp3zf2GN17BqTTvv2w5ceusRCyuWb9jH0+XtzKjaQ1f8zxFaWw/Or0EFaw3+/h5N5qKnj4qarTWNAUjVhOnMDWtIWpbQhxuCLC7qpXyqpZeEzMAN2EW2nbjIsrm+BRISjUXZvuTyElxk5/qoTAticJ0DwVp5uZx2q2k2iTXEes+FDVNwtojcUIRk9D9cc1O3jocQOs4t01u45asMorcQdSsq01zRpvthGXpPGCubqWuLWz1TqSJxc3ZyVA0RmVjOxUNAY40BiiKHuZMWzkN2s+a+FwC1um3ZGud2pSymtEplJVc5qUkMTnHx+Ts5M5bfloSDpsa1EHP+ne2sbXFx6NrD7Kvto1Mn4vrFuZxRnEWk6yDQl+PMyjBcIyKhgCH6tpoqDqEvbYcb8NOUlv3kB3cR0HkEFFsNMR9NOCnUSfTQDIh7cKngiTT3nmfRIhGewbNSYVEU4pxZE7CnzcVb1o2mo6zoub7VDYbuVk5ZGZlowaoUOrrv0UptVFrvXig70WS4350+3LDbXBoLRx6C9pqzIYeaob2JvM4UGdu9Pg+7W7zR5o52bqZjRYdh3jMWvNx2lf/N0ltlXDLM1C0aHgXpKnCJDgbHoFQE8H0mXgKZpjEzeU39+5k8wcYqLeWpdbcR0OmiUfmFHPLmmruPekmqRto49uzCv3odaiS8+Ejfxu45rHlGLz6Y9j4B5OYLL3TlLP5j9B4yCRc826CCWdB+fOw81mzHjwZMOsqqqLJ5OYVmsTFZgeb08QYi1i3sEmC4lGTXPuyzM2bBb5sKzEdXBKx893NzLAdgK1/g90vmvl6s8y8w63mvoPdbf7oihabBLNwkTm4OrLBHMQcXmdqkyOBvgtMLTbzyJ8PBfNNQu3NHNbaXIDdG19lauMakzwGaiFrGsy/2bRrP7TWnKHoWDZHEkQHMQS7O9Ukyr4sE7M3E7wZ1DQFyE52dk+u25vM95hZaradjFLImGySHXfqCX9C/elIqKua2tm1ezcf8O1AvfRtCDbC2Z+DC79ufv+RoOnO8dBbZn1U7zQHMdGQeS/eIzG1Oc33kjPDbB9JqebWccDs9FFRvomiZA3NR8wFqE0V5oCpYL45W1CwwHx2kLV24WicxkCY5vYozVZb2+ZghGA4Rmn7DhZHN5rrE5oOmwOS1GKoevf4919yPkw8x2wHbTXmDExbjblFOg7Yuuy/4lEItZqkODbI4d5tDsiaRjhzBvXeybTbPOaAQEMsfrxpRzgSIRSOEI5ECUfMfXtMEYorgnE7oZgiGFMUO1u4rDhGRrgSGg6YZev47bn8phlW0ZnmO63ZaQ5Qj1nLXLDQHDQ7PNa+wGFuAI0HoaYcaneZazd64/Qdb1IXj6F1jHgsSjwWI+xIps2TT7M7nyZXHg3OPBpUGoFQiPZggFB7kHAoSCQURMeiKB3DThw7Mewqzvm2bSyy7SKaMQ3H+78P095vfgdaw/Yn4Lmvmf3v0s+b5Xvrfji81vzGFt1qksxDa2H747D/NbPecmab/YPNbu3/zPLWNTSSmZFh/nO0JhaP0RQI0xhxUBfzUB1J4mi7k8MBJ9UhJxHlJKqcRJWDqHIR1k7aQmHa2sMmfuLdbqrL4yJVy796nqUkdgCdOQW17P/BnA9DuMV0lbr2f0ysZ37aXGQeajEHOw0HrNtBa936jlcwuHxmuZNzwZ9Lgy2D3QEfuxvixPe/xly1j7z6t8mq34w9HrJ+xYqqpMnsdM9mMzN4K1TCoRZNtMflMX1lPzFsxLBb9zYUmmWuXXwmbxfzAm/haK0ElDlwjrabbe2MD8MZ10PurF5mGLH2I9Hjt1jE7OsOrYWDr6MPvIEK1B7/iM1FZcZZlKct4x3vUqriKZ2JYty0k4J4mOrGNg7UttIYCGFDo9BoFCHMOrTZTFLtsCncTjtJ1pkOt9NOktPG/soqpsT2c3lmJe9LPUJhWxmq8QCkFEH29OO9d2VNN/9PNeVQU2bd7zTL0CE511QeZE03yxCoJ9xcS6ytDoINxCNB4k4/YYeXiM1L2O4jrJw4AlWktleSEa/D1uda6S6Mk6DdT9iZQtzmIqYVUa2IaYhqRSQaZ9rXXsXh7n7djSTHw+CkayFjEfOH03oMWquhudI0lajbC3V7zE4g1nuNStzmwvaRFTDlkmGKvhehVtjyKIF1f8JLu9kxhVvNrYMr2SQxXiuJsbvMMtTv7SV2a+dgd5sEzeawEv6YdR+HcBvtKSUkfW61SRoGq3YPvHyP+bMD06Rl4S0w88ruTSki7aYd87bHofw5iLQN+evp5Eiyan2TzR+wy3f8QEDZzJ+P1sQOvY09GoDkPGvHeJ35c+pIdqIh851H2sw0AyWxsSjU7TafiYXM52MRiIWoKl9PbqzS1HDW7z0x3o6kzO0366xzu7buddysv44DhI57AKzlsm66+QhKx02zlyWfhtKLuidw4TaTSB5ea3aMnvTuN6cPQk0QaIBgvTng6rjvOIgM1JvEO9puEjdP2vHPJ6VAa41Zztaq7suqbNZ0Vq10Ry273WkSVrvD3Pdytiaw7228de+a2sXLfwZ5cwb1cwDM7zncZraFmp2mWU/NTqjeYS6I1f1clOr0HW8C1d5kkreO796TYf5M4lFztifSZt0HzDrsmtR1PLY7zTJ3PI5FzAELCiZfDAtuhumXm5rutlrY/6oZAXTvK9B0yPrNeMxZGV+22dY7mop1rmfrN+G2Dp47zpS5/VRUN1BUMs06Q+I1yxdtN99J9XZzX7XjeFmnypNhDiIySqz7UnNwmTWt97NydXuh7Gmz76jcQq8pkCvZfD57euf9gaomJuX4u/9Ggw3mu7DZzFlDm918L8FGk6g3HoaWyt6v/+hHOCkb1/v+HeZ/tPeD8WADvPQdc6YMTE3r2XfAgo+eeA1JS5VZ1u1PmAOwePR4BUA8RjwaxmZ3mOVQ6vg6DgdOPOgbBjpzKuqCr5omcj3XT1MFvPKf5iCu53fmSjZnZmw2s6113lrplbIdn0fuHNP0sfQCsz88/DYcfNOcWejr80Ph9JptbPpymPp+U4FTvhK2PgZ7Xzb7gewZZrpQi1WB1nzi2cKeUoth0nnm2puJ53Bw21omBrebyp/Gg4Ayv9N41OwbOr6b/vY7lph1kBNTTuJKdf41dCTa/ngzNqzvMaXIHHBmTTW/7Zqd5qxnz/g79lvZM8w2lDPLPB+g29mBciodaafx2H6qD5UTajW97dBRWw7EYzFamuoINNURbq1HBxuxhRpxEsVpA4dN4bRpHDZQ8RjTv/wCHs84T4679Fbx6YMHD45ImcN+ij4eMzvUcFuXZMQOSlFeUcv0eWcNX1n9OGG54nGzA7E7zcbeX+y1e0zSEmo2yVvHLRYyG27nH4j1R+7ysjv9IqYuPH9owdaUm0Qgo/9u0gCIRdm5/R1mTC01iWbHn4TWVgLlMn9EdpeJr73xeO1ZoNY8DjaaWo6QddDQkdzG41bNS9zskHScRlcBaed/2uzUTmLY8aHots7am8yFm9Vl5nGoyTrlbZ3N6Kyx7rhs39q1dDbTcB1PrpTqslzmyr7asJOsf/mSSUQSrGz7NmbO7idJDbWag8q6vaYGtmuiHaw366sj2Y9HrPUe6TVZCSs3rkv+zdSCn0Tt84C0Ntt01zNJ4Vb2VbdSOv98k8B3PbiIhk0SeWSTuX6gbo9ZHy5f94RT2brXNMVj1jJGjt/HTBJU7ZtOzvvuNkl4f3G21RyvlRuiQe8Xw21m36A1xw/SrPuOJLNjX6FsZruKRehas1Z+8BjT55055FgBs+12/R51zNSs96ixH/L+PhYxFSFttWb/4kiymkUlmW2u46Cmy76xrLx8cGVVbDTrbMqlQ24W0+dyaW0Oatqt322o2TpLELb26eHjj5Wt2wHann37QdmYMnV6t/V4oLKaSeffMPD+sLoMdr9kmh2mTzI3b2bvZ1HicRNba7WpeGqpMgfNgTqOxNIpPPem483veopFzdmTo+90P5PXsfx96VrJE4+CjnMoks6EZR/p+z+ytcYcoOx63iyHdTBpDixTzee6ntG0OcxrhYsgvftF7p3rTGtzAL7zn+ZAz5lk7SOSrf2E12re1bWCw9qnd6zDrv/RPStNgJoAZM/7F3Omxd9Lr0vxuDnQrd1t4s2eMeSxFxLR7LEjh+3ZhORUm1WMmQvyRqy3ikSy2ftMOOLVg2ujlxA228C1uh2xp08CLj2p2UfLyoYamTkqHSy7A+30mmRkMJJzzG2IjpaVkVY6cm23OyVZp8hLhnjAMYCasjKyRiAxBgb+E3UnQ94Z5naK9iaqPbpSJk53svmzt4QiZb1f3OdwmSYVBQuGLYS6sjJy+kuMO+I8hd/7STvFJBwgfnQYav1sNrC5gAR1Q2d3muSmR4IzLIa7iV1XSplkx+npPSnqx+sbHwFgyswrur0ejJYNrqIgZ6a5DYbNZp1ZSjvhWpzmsjIK+0qMwRxQDNO21lZW1ndiDCZBP+t2cxsuSkHubHNLkNqyMrKn97MubLYu//2nn0T1AHL6XSIohBBCCCHEKElocqyUukwpVa6U2qOU+nov77uVUn+13n9bKTUpkfEIIYQQQgjRn4Qlx0opO/BrYDkwC7hJKdXzMs5PAg1a6ynAz4EfJyoeIYQQQgghBpLImuMlwB6t9T6tdRhYAVzdY5qrAetyXP4OXKLGyhAyQgghhBBi3EnkBXmFwOEuzyuAnt0xdE6jtY4qpZqATKC260RKqduBjlbuIaXUyPTlBqlA04BTSVkDyaLHOk2g8fo9jmRZsr7GXnmyzsZWWeNifd16660jVlYfZJ1JWX3pa30N7spZrXVCbsB1wO+7PP8Y8Mse02wHiro83wtkDjDfDYmKuZeyHpCyhqU8WWdjqyxZX2OsPFlnY64sWV9jrDxZZ2OurFNaX4lsVlEBdB2/swio7GsapZQDc1RRn8CYTtYzUtaYM16/x/G6zsbzdyjrTMo6HYzn71DWmZSVEAkbBMRKdncBlwBHgPXAR7TW27tMcwdwhtb6s0qpG4FrtdbXDzDfDXoQHTiL04ess7FF1tfYI+tsbJH1NfbIOhtbTnV9JazNsTZtiO8EXgDswENa6+1KqXsw1d1PAw8Cf1JK7cHUGN84iFk/kKiYRcLIOhtbZH2NPbLOxhZZX2OPrLOx5ZTW15gZPloIIYQQQohEkxHyhBBCCCGEsEhyLIQQQgghhGVMJccDDUctRpdSqlgptVopVaaU2q6Uutt6PUMp9ZJSard1nz7asYrjlFJ2pdRmpdSz1vMSazj33dbw7q7RjlEcp5RKU0r9XSm109rWlso2dnpTSv2rtU/cppT6i1IqSbaz04dS6iGlVHXXMRT62qaUcZ+Vh2xVSi0cvcjfu/pYZ/9t7Re3KqWeUEqldXnvG9Y6K1dKvX+g+Y+Z5HiQw1GL0RUFvqy1ngmcDdxhraOvA6u01lOBVdZzcfq4Gyjr8vzHwM+t9dWAGeZdnD5+ATyvtZ4BzMOsO9nGTlNKqULgLmCx1noO5gL1G5Ht7HTyCHBZj9f62qaWA1Ot2+3Ab0YoRtHdI5y4zl4C5mit52J6S/sGgJWH3AjMtj5zv5VT9mnMJMcMbjhqMYq01ke11pusxy2YP+1Cug8T/gfgg6MToehJKVUEXA783nqugIsxw7mDrK/TilIqBViG6ekHrXVYa92IbGOnOwfgsbo49QJHke3stKG1XsOJYyz0tU1dDfxRG2uBNKVU/shEKjr0ts601i9qraPW07WY8TXArLMVWuuQ1no/sAeTU/ZpLCXHvQ1HXThKsYgBKKUmAQuAt4FcrfVRMAk0kDN6kYke7gW+CsSt55lAY5cdjGxnp5dSoAZ42GoK83ullA/Zxk5bWusjwE+AQ5ikuAnYiGxnp7u+tinJRcaG24DnrMcnvc7GUnKsenlN+qE7DSmlkoF/AF/UWjePdjyid0qpK4BqrfXGri/3MqlsZ6cPB7AQ+I3WegHQhjShOK1ZbVWvBkqAAsCHOTXfk2xnY4PsI09zSqlvYpp5PtrxUi+T9bvOxlJyPJjhqMUoU0o5MYnxo1rrx62XqzpOO1n31aMVn+jmXOAqpdQBTDOlizE1yWnW6V+Q7ex0UwFUaK3ftp7/HZMsyzZ2+roU2K+1rtFaR4DHgXOQ7ex019c2JbnIaUwpdQtwBXCzPj6Qx0mvs7GUHK8HplpX+LowjaufHuWYRBdWe9UHgTKt9c+6vPU0cIv1+BbgqZGOTZxIa/0NrXWR1noSZnt6WWt9M7Aa+LA1mayv04jW+hhwWCk13XrpEmAHso2dzg4BZyulvNY+smOdyXZ2eutrm3oa+LjVa8XZQFNH8wsxupRSlwFfA67SWge6vPU0cKNSyq2UKsFcTLmu33mNpRHylFIfwNRsdfVDC0AAAALHSURBVAxH/cNRDkl0oZQ6D3gNeJfjbVj/DdPu+DFgAuaP4jqtdc+LH8QoUkpdCHxFa32FUqoUU5OcAWwGPqq1Do1mfOI4pdR8zAWULmAfcCumokO2sdOUUup7wA2YU72bgU9h2jzKdnYaUEr9BbgQyAKqgO8AT9LLNmUd4PwK0+tBALhVa71hNOJ+L+tjnX0DcAN11mRrtdaftab/JqYdchTT5PO5nvPsNv+xlBwLIYQQQgiRSGOpWYUQQgghhBAJJcmxEEIIIYQQFkmOhRBCCCGEsEhyLIQQQgghhEWSYyGEEEIIISySHAshxGlGKRVTSm3pchu2UfCUUpOUUtuGa35CCDHeOAaeRAghxAgLaq3nj3YQQgjxXiQ1x0IIMUYopQ4opX6slFpn3aZYr09USq1SSm217idYr+cqpZ5QSr1j3c6xZmVXSv1OKbVdKfWiUsozagslhBCnGUmOhRDi9OPp0azihi7vNWutl2BG6brXeu1XwB+11nOBR4H7rNfvA17VWv//9u5YlcMwiuP495ckJQuLMlhMBiVX4BYMyCSThUluwC1YDK5DySayWayyUQwGi6Rj8NB/YPzzf/X9LO95Tu/wPNvpdN73mQcWgOuWnwUOqmoOeAKW+3weSeoMb8iTpAGT5Lmqxr7J3wJLVXWTZBi4r6qJJI/AVFW9tvxdVU0meQCme68lTjIDnFTVbFvvAcNVtd//k0nS4LNzLEndUj/EP73znZee+A2/P5GkLxbHktQtKz3PixafA6stXgfOWnwKbAEkGUoy/lublKSuslsgSYNnNMlVz/q4qj5/5zaS5JKP5sZay20DR0l2gQdgo+V3gMMkm3x0iLeAu77vXpI6zJljSeqINnO8WFWPf70XSfqvHKuQJEmSGjvHkiRJUmPnWJIkSWosjiVJkqTG4liSJElqLI4lSZKkxuJYkiRJat4Bo7sjHk3p95AAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 720x180 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x180 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9935714285714285\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAADVCAYAAACVHr5AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3Xm8VfP+P/DXR0XzQCo5TURpToaoKJfQpNKkNAopDWigmTTcopIhUmmiUYgbkiipKOqmNE+a50FJoc/vj7W6trNf7+zD4nx/976ej4eHep11Pue991p77U9rr/f5OO89REREROTPOy+1CxARERH5b6GJlYiIiEhENLESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhERBMr+Z/inLvSOXck6m1Tk3PuPufcuwlu28Y5N/cvquNS59wi59z3zrn+f8XP+KOcc5ucczekdh1/J+fcHudcxT/4vX/Lsf9XHo8iqUUTK/nTnHNbnXMnnXPHnXN7nXOvOecyRzDuZeGYZ//zzrkTMX9P8Rul93699z571NumhHOuinPugHPOxWQTjWx4AnWO8d7XjKi2Jc65e//gt7cFsNV7n8V73yOKelLCOTfFOXcq2TFTGwC895d77xf/gTGLOud+TmC7Ys65mc65g865I865Fc65DrH78/+y5JOwv+rYT2FNCT33Iv/XaGIlUanpvc8M4GoA1wLomXwDF0j4mPPeb/beZw7HPXuSL342Y2+Uzrk0f7D+v9MSAFkAlIjJKgE4kCy7CcCCv7GuP6sAgG+tLzrn0v4NNfSLOT4ye+/f/r1v+LN1OeeKAlgMYB2C4zM7gMYI9ukFf2ZsSV1/0zEr/2U0sZJIee93Angf4QTBOfepc66/c+5zAD8AuMw5l805N8Y5t9s5t9M59/QfnRCFVylGOOfmOOdOALjBOVfHOfdv59wx59w251z3mO1/86/g8ApNn/D/x5xzs51zOVK6bfj11s6575xz+51zXa2PYrz3JwEsQzBxgnMuH4CTAN5JluUH8Fn49wudcxPCMbeHdZwXfu03H6c456o75zaEV06Gk6tQLnzOjoQfkd0ahs8imBSPDq/2POucS+OceyF8TEfD57UI2Q+TATQE0Cv83krOuUHOuTecc1Odc98DaOScy+CcezHc9zucc0Occ+nCMe5wzm10zvUMr97tdM5Vc87dFdZ50Dn3WCLHBanvP/vCqKuCc255uF/3OOcGht+6AEAa9+sVsLJk+KcBzPXeP+G93wMA3vtvvff1vfc/hj/zbufct+FzPtc5d0Wy2h5xzq0Kn+PXnXPnh1/bfHb/hH+/INym2O+Nm+zxT3HO9Yz5+x3OuY3hn6cDyAVgTvgYO5BjP394vB9yzq13zjWP+dqgsObJLvgYeKVzrkzM13s757aEX1vlnKue0E47h3B/fRE+F7ucc8NcOAlywbmlf7LtP3LOtQn/nM859054jG0+m8c8lt8cG3+2Vvnfo4mVRMoFE4JqAJbHxE0BPIDgKs02AOMB/AygMICyAKoCaP0nfuy9AHqF4y8FcAzBFYPsAOoA6Oycu+Mc398YQBMAl4Tf0zGl24ZvJEMBNACQFP6X8xzjLEA4iQJwM4IJ1MJk2Rrv/f7w768DOArgMgDXAaiN4Hn9DedcHgBTATwC4GIAuwCUS7bZTQgmdhcBeAHAaADw3j+G4PlrHV7teQxAjfD7LweQI3z8h5P/XO/9PQDexK9XjD4Lv3Q3gv2dLfz6kwBKASgZjlsZQNeYoQoA+AlAHgCDAIwFUC/8nlsB9HfOXZr85/8Byet6AcAA731WAFcAOHul6yYAv8RcAVtOxroVwAzrBznnSgAYh+Cj0lwA5gOY5X57NaQegH8geE1cj+B5BoApAO6J2a4GgC3e+28THPd3ee/rA9gHoGr4GEeQzaYjuCJ3SVjbMOdchZiv10Gwr7ID+BhA7EfY6wDciOC5/ieAKc65c702EvETgIcBXIjgymBN/HoOGQ+gsXPBx7DOubwAKgCY5oJ/wM0GsAhAXgB3AOjunLs5Zuzkx4ZIimhiJVF52wU3uy5EcIIfEPO1cd771d77nxGcCO8E0Ml7f8J7vw/AMPy5fxnO8N5/4b0/470/5b3/OPx5Z7z3XwOYhmCiYnnVe7/Je38CwRtkmT+wbQMAb3rvl3jvTyH4KPRcr6/5CN4QEP7/7MSqYkw2HwCccwUQvME/6r3/wXu/G8AI8OesFoCl3vv3vPc/AXgG8ROhdd77Cd77XxC8gRRwzln30/wEICuAogB8+LzuO8fjinuc3vvZ4b44iWBS2sd7f8B7vxfB1Z7YCeIPAIaEx8oUALkBPBMeK8sBbEIwKbP0CK/eHHHO7UhBXT8BuNI5d5H3/nvv/ReJPLjwjTobgN3n2OweAG957z/13p9G8NrICeCamG2Gee/3hhPp2fj1uHodQN2zV7AQTGreSMG4f1p4Faw0gO7h62sZguMmdr/N895/FB5TE2Pqh/d+qvd+d/hcTwSwE/GT/RTx3n/pvV/qvf/Fe78JwT8Ozr7GPwPg8etrqTGAD7z3h8Isvff+n97709779QBew29fS8mPDZEU0cRKolLbe5/de1/Ae9822Qlpe8yfCwBIB2D32TdAAK8g+Bc3nHOrYz52qYTExI5/9mOC+S78+ApAC5z76tGemD//AOBcN95b2+aNrcN7fwzBFSbL5wByhW9aNwH4zHt/GMDhmOzs/VUFAKQHsD/mOXsOwaQjueR1nEHwRnauxwDYj/l9AGMQ7KO9zrmXXMoaE/5TS3gFIQ+Cq5ZnbQMQewVqf1gzEHw8CgB7Y75+8hy1AkD/8DjM7r1PSqSuUHMEV8XWhx8x3X6O7/2PcCJxFMGVHEtexDzm8Ht24rePmx5X3vvVYa13OueyIvhHyeQUjBuFvAj2S+xrOvl+M19DLuhaXRlz7BbGuV+Pv8sFzQLvu6BZ5hiA3mfH9N57ABMQXMlG+P+J4Z8LACgYM/k+AuBRBMflWcmPDZEU0cRK/g4+5s/bAZwCkDPmDTCr9744AHjvY29O/4yOdu7xgeAK1VQA+bz32RB8XPJXd2ftRvDxHwAgfBPMZm3svf8ewAoAdQFk8t5vDb/0WZgVwa8Tq+0AjgPIkew5uzqBOs5Dyt5of/Nc+sBQ731ZBBOP0jj3R6XmeOEb3h4Eb25n5Uf8xO/vkPxxrvHeN0QwwR8BYGZ4lSj5scXMRfDxkWUXYh5zeJXrUiT+uCcjuDp1N4KrkWff+FMy7gkAGWP+nifZ18/1OHcBuNg5lyEmS2i/OeeuBPA8glsBLvTBjf0b8edfj68C+BrA5eHHt08lG3MCgHrOuXIA8gH4V5hvB7A25nWU3QddrHVivjeRfS5i0sRK/lbhx1hzADzrnMvqnDvPOXd5snsc/rDwqkhmAAe99z86524EUD+KsX/HNAB3O+euDd+QnwJw5ne+ZwGCfy0vjMkWhtkWHzQCwHu/BUEn4WDnXJbwObvC8d9RNAvA9S646TttOFYOsp1lL4L7uAAAzrnyzrlrwrFOADgN4JcUjJfcZAB9nHMXOedyAegBYNKfGC8Szrlm4ceAZ69AeQT7bx+Cm9fzn+PbewG4zQVNGrnD8YqEN0CnRzDJr+Ocu8kFN+o/DuAggvvcEjEZwb1VrfHrx4BI4bgrANRwzmUP71Frn+zrv9nvyWwEsBLA0y64ef5qBFf4Xk+g9swInsf9AM4LbxQvnMD3/YdzLn2y/xyC+ymPeu+PO+eKA7g/9nu895sRdKi+BmBq+FEpEL7WnHOdwrHSOudKhY9JJBKaWElqaAbgfAQnvsMI7lU610cpCQuvirQB8EzY1dMVwY23f6nw/p8uAN5C8C/53QjeoE+d49vmI7hCEjux+izMkv+ahXsQ3Bi8FsAhBG+qcR8FhhPXexBcdTmA4OrVN79TR6xhAJo55w475waHP3McgCMANiP4CIjd3Jyo3gj2+2oEb/afAxj8J8aLSg0A68JjZiCABt77n8OPZwcD+Cr86Cju/jvv/RoEN2cXA7Am/HhpCoJ9ecp7vxLAfQg+Tt2P4Cb1u8L7yH5XeDXz3wg6NqfH5CkZdyyCCdJ3AN7Drx8nntUfQWPAEefcw8l+vkdwD2ExBFccpwLoksgVZR/c4/gygsnebgCFkPiEEgDSIPj4N/a/CgiaM1o7544DeDGsKbnxCO7HO/sxIML7Dqsh2F/bEDxvI3Huj5dFUsQFrxkRiZILfg3DIQB5w8lOatWRFsGbYU3/B35Bpsj/r5xzVQG85L1P0RUykT9LV6xEIuKcq+WC39OUGcGvXvgiNSZVzrk7XfC7wtID6IPgZuKv/u46RFJL+HF8BwCjUrsW+d+jiZVIdOojuDq0A8FNxE1SqY6bAGxBcH/QPwDUibnHROS/Wvhx7WEE92G9mMrlyP8gfRQoIiIiEhFdsRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiEREEysRERGRiGhiJSIiIhIRTaxEREREIqKJlYiIiEhENLESERERiYgmViIiIiIR0cRKREREJCJpU+sHX3XVVXSRwjx58tDte/fuTfPdu3fTfObMmTSfMWOGS56NHj2a1lK7dm06Rs6cOWn+xBNP0HzMmDE037dvX1wtAJAvXz5aT8mSJek4VapUofn5559P806dOtHcex9XzyOPPEJreeCBB+gYO3fupHm7du1o3q1bN5q3atUqrpZTp07RWrJkyULHyJ8/P81XrlxJ84kTJ9L8wQcfjKtl2bJltBbreNywYQPNc+XKRfNp06bRfNasWfSYWbVqFa1n165ddJzhw4fTfNCgQTTPmjUrzQsWLBhXT82aNWktbdu2pWPceeedNH///fdp3qtXL5ovW7Ysrpb169fTWi6++GI6xosv8jV7x40bR/O3336b5iVKlIirJUOGDLSWpk2b0jFq1KhB86SkJJrffffdNN+yZQs9Zs4//3xaz3XXXUfHeffdd2luPWdXX301zatVqxZXT5cuXWgtx48fp2NcddVVNL///vtpniFDBpoDiKtl6tSptJYbb7yRDjBv3jyaL126NEX5F198EVdLpkyZaC3ZsmWjY0yePJnmadKkoXnp0qVpniVLFnrMVKtWjdZjvVY//fRTmp93Hr+mU6hQIZonJSXF1dO0aVNai7X+8aRJk2j+4IMP0rxmzZo0r1GjBn1uYumKlYiIiEhENLESERERiYgmViIiIiIR0cRKREREJCKpdvP6kCFDaN6nTx+ad+jQgeYbN26k+bFjxxKuxbqB2rop/KuvvqK5dePvN998k3AtANC9e3eat2nThuYnT56k+aZNm2i+bNmyhGuxxi5WrBjNrRs8S5UqRfMcOXIkXMuZM2dofu+999K8ePHiNM+YMSPN33vvPZqzmxtPnz5Nt7VuIF69ejXNP/jgA5pbNy5bypUrR3OraWDNmjU0/+STT2hu3by9fPnyuKxevXp025YtW9L85ZdfpnnDhg1pfuTIEZoz77zzDs2//fZbms+ZM4fmVsOD9ZhmzZoVl1nHQLNmzWi+fv16mls3dFs321qsG50PHjxIc6sJp0KFCinKmfLly9Pc2n8jR46k+YkTJ2j+6KOP0vyCCy6Iy7777ju67VtvvUXzKVOm0Hzbtm00v/zyy2nOWA0ArVu3prl1vraOmYoVKyZcCwDUqlWL5tZ71q233krzoUOH0vzUqVM0Zw0bVqOU9X775JNP0txq1ujYsSPNEzk364qViIiISEQ0sRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiETEWb/+/a/Wr18/+oN79OhBt+/ZsyfNre40azmXBg0axP06+p9//pnWUrduXTqGtTyJ1VW2YsUKmt999930V+PXrVuX1rNgwQI6jrWki9XtULRoUZo/+eSTcfU88MADtJZRo0bRMZ599lmaW11lXbp0oXmzZs3iailYsCCtpUSJEnSMN998k+Z79+6l+TXXXENzY+khWovVPWZ1eM2dO5fmH3/8Mc03bNhAj5nx48fTeqylLKzutz179qRoe5AlQSpVqkRrueuuu+gA+/bto7nVBVq5cmWas6UmGjZsSGv58MMP6RjWUimZMmWiubWf2PJQa9eupbVce+21dIxGjRrR3OpwK1OmDM1PnTplLcFB63nqqafoxq+99hrNrde81aWYLl26uHoOHz5Ma7G6by+99FKaW0vXWMtV1a5dO66Wiy66iNZidaNar7HNmzfTvFKlSjRnSw9Zz0vhwoXpGNbyLDfffDPNrWNp9erV9Jg5ffo0rcfq8rM6IJs0aUJz6z13ypQpcfU89dRTtBbrebfesx5//HGaW0twFS5cWEvaiIiIiPxdNLESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhEJNXWCrQ62awOC2tNOWtdsgsvvJDmDRo0iMusrpann36a5la33fXXX0/zX375hebWunJffPEFzQ8cOEDzf/7znzR//vnnaW51ZzG33HILzT/99FOa33bbbTR/7LHHaP7MM88kXEv27Nlp3rdvX5pPnz6d5p06daK51YXFzJs3j+bW8WitU5UrVy6aW12tlhYtWtDcWt/N6qS01mazun5YV9GkSZPotlOnTqV5//79I9mereFlnQcWL15Mc6sr8KOPPqJ5/fr1ac40bdqU5iNGjKC5tQ6htW7qgAEDEq4FAAYPHkzzL7/8kubW+W3p0qU0t47hlOynqlWrpuhnWmvTtWrViua1a9eOy6y1G4cPH07zOnXq0Pziiy+mudWhyFid4OnTp6c5W7sTABo3bkzzF154IeFaAOCGG26g+ejRo2letmxZmrP3YSBla/xa6wpb54fq1avT3OqMfPXVV2luvd/G0hUrERERkYhoYiUiIiISEU2sRERERCKiiZWIiIhIRDSxEhEREYlIqnUF/vzzzzR/5ZVXaD527FiaW+v/WGsOMdOmTaP5/fffT/M1a9bQfOvWrTQvVapUwrUAwPbt22ludVLkzp2b5taahta6fYy1FmPp0qVp/sYbb9D8H//4B82tdQs7d+4cl7Vv355u+8MPP9D83XffpbnVCXPPPfckvH1KH8+JEydoni1bNpoXKVKE5pajR4/SfMmSJTS3jm2rQ3bt2rUJ11KwYEGaW+tCWms6WmvfWWutMdYaaf369aO51ZFrdT1ZncNszTark8vqRHSOL0l2++2309xaz9DSrVs3mpcvX57mVnez1SE7Y8YMmrOuQKsbvHv37jS31oOz1mW1OveY8ePH09xaLzMpKYnm1vnnq6++onnFihXjMta1CNjnQqsr2VqD11r/sEqVKjS3zjMzZ86kudXZar1nfffddzRnrOfXeq1acwVrHOs9LhG6YiUiIiISEU2sRERERCKiiZWIiIhIRDSxEhEREYmIJlYiIiIiEXHe+9SuQUREROS/gq5YiYiIiEREEysRERGRiGhiJSIiIhIRTaxEREREIqKJlYiIiEhENLESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISEU2sRERERCKSNrV+8Ny5c+nqzydPnqTb16pVi+aPPfYYzQsXLkzzNm3auOTZyJEjaS233XYbHcPKP/nkE5o//vjjNJ8yZUpcLQDQqFEjWs8dd9xBx3nmmWdoPnz4cJpfcMEFNK9UqVJcPSdPnqS1zJo1i47xzTff0HzSpEk0f/TRR2neoUOHuFr69etHa7n//vvpGAcOHKB5nTp1aD548GBr+7haMmbMSGtZv349HcM6ZqpWrUrzxo0b0/z666+nx0xSUhKtp0CBAnSczz//nOYzZ86keZUqVWieI0eOuHpKlixJaylfvjwdY9WqVTTfunUrzW+99VaaT5w4Ma6WRYsW0VqsMb788kualyxZkuabNm2i+WWXXRZXS9WqVWktU6dOpWM0b96c5mXLlqX5yJEjab5v3z56zHTv3p3W06pVKzpO+/btaW6dC6xjZuHChXH1VKtWjdayf/9+Osa+fftoXr16dZpbz433Pq6WI0eO0Fo2b95Mxxg9ejTNrefLOmZq1KgRV0uzZs1oLeeffz4dY9SoUTS33oMWLFhA8yVLltBjZsOGDbSeK6+8ko6TNWtWmufNm5fmhw8fpvmePXvi6jl+/DitxZor5MqVi+bjx4+nec+ePWk+ZMgQ+tzE0hUrERERkYhoYiUiIiISEU2sRERERCKiiZWIiIhIRDSxEhEREYlIqnUFPvLIIzR/9913ad60aVOa7927l+Zp0qRJuJb77ruP5k8//TTNrdpHjBhB84MHDyZcCwDkyJGD5la3Q7169WjevXt3mt97770J15IpUyaaT5kyhebZsmWjeb9+/Whu7VfG6szKnDkzzVevXk3zMWPG0NzqMGVdhBMnTqTbdu3aleaDBg2i+RVXXEHzYsWK0dzSrFkzmlsdk/Pnz6f5E088QfO+ffvS/J577onLKlasSLfduHEjzV9++WWat2vXjuZW5zBTqFAhmlvdfyVKlKD5ypUraW51ZzHPP/88zQ8dOkRz61xYqlQpmv/0008J1wIA3bp1o/nNN99M8zVr1tDc6qpKyX6aPXs2za3uMatrdvHixTTPkydPwrVY598zZ87Q/LrrrqP50KFDaZ6S82/BggVpbu2LN998k+YDBgyg+YoVKxKuBQB27dpF8yZNmtDc6pS3OtyLFi2acC0PPfQQzbNkyUJzqyvZ6pK/6667Eq4lOV2xEhEREYmIJlYiIiIiEdHESkRERCQimliJiIiIREQTKxEREZGIpFpX4IwZM2hudaRYnUPWWlLWelrMSy+9RHOrw6tcuXI0X7duHc2XLl2acC2Ava6c1V1odRtZ+aJFi2i+bNmyuOyNN96g21pdNtaaUdZ6cJ06daI5W+fQWjuud+/eNLe6KK31tKz9x6RLl47m1vPlHF9eyuqirFy5Ms3ffvttmltrcllrZlrjWF1uVvciY70+smfPTnNrfTBr7UZrzTbGeg0cOXKE5hMmTKD58uXLab5jxw6aJyUlxWXWmp5WV6/1OrXWoLPWNrNY+8M6/zRs2JDmVuel1V3HWB2Ebdu2pbn1WIsXL07zb7/9NuFa5s2bR/OWLVvSfMuWLTS3nl9rv7L3PqtL1zrurPNvixYtaP7aa6/R3GKtuWh1GW/fvp3m1lql1vs/Y3UcTp8+neZr166l+SuvvEJz6zFZXc+xdMVKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISEU2sRERERCKSal2Bbdq0ofno0aNpXr16dZpba5ilZM0ha61Aaw26XLly0dx6TNaaXzt37qT5nj17aG6tybVp0yaae+9pni9fPpoz1ppc1jp81ppfX3zxBc2tdbaYChUq0PzAgQM0r1KlCs2tTpixY8cmXIvVhWetX2Xta6tD6L333ku4FgCYOnUqza31sdgafwCwYcMGmqekY7JAgQI0t7pvevToQXOrY9Q6R6Rk25S+Vq0OxWrVqiVcyzvvvEPzW265hebjxo2jufWaSclrCbDXQrPOb1bnnrUu65AhQxKu5YYbbqB5hgwZaM46mAFg/fr1NLfOnewYs45Ta31Uqzvtl19+oXmNGjVozjoAf/zxR7qt1ZX92Wef0fyiiy6ied26dWk+c+ZMmlvdtNZrfsmSJTS3jm1rrUO2Dq+1/uqnn35K84EDB9LcWjvWWnMwEbpiJSIiIhIRTaxEREREIqKJlYiIiEhENLESERERiYgmViIiIiIRSbWuwI8//pjmVhfB6dOnaW51DllrFLHOma+//ppue/ToUZovXLiQ5ta6Till1T5r1iyaW52R1vpuixcvTriWbt260fzDDz+keefOnWl+5swZmludi4zViWitFZgzZ06aWx2m1jp5zLRp02jeq1cvmvfp04fm1rpsGTNmTLgWwO5aa9SoEc1PnDhBc2udLWvdRaZq1ao037t3L82tbk9rf1hrZrLOp40bN9JtrcdvvYatzsVjx47RnDnvPP7vWKtD0+qktY7rf/3rXzSvVKkSza11+Kz9Z3XuWc+B1aXI1nG999576bZWt6tVu3VumzNnDs3Z+0dKOxGtteOs7sKUrOmYPn16mlvraw4dOpTmw4YNo3nevHkTrgUAChYsSHNrHUVrfOtcYHVPM9YavKyDEAD69+9Pc2suYq0zmghdsRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiEREEysRERGRiLiUdGWJiIiIiE1XrEREREQioomViIiISEQ0sRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiEREEysRERGRiGhiJSIiIhIRTaxEREREIqKJlYiIiEhENLESERERiYgmViIiIiIRSZtaP/iWW26hqz+XL1+ebv/JJ5/QvGjRojS/+uqrad6+fXuXPHv00UdpLbNnz6ZjVKpUieavvvoqzbt3707zAQMGxNUCAK1bt6b1TJgwgY4zb948mlt1btu2jeb58+ePq+eaa66htSxbtoyOMWTIEJpfdtllNK9evTrN06dPH1fLwIEDaS3W8/vkk0/S/LvvvqP5mDFjaO69j6vl9OnTtJY5c+bQMSwvvPACzU+ePEnz+fPn02Mmc+bMtJ5y5crRcW666SaaN2/enOYlS5ak+cmTJ+PqKVeuHK1lzZo1dIzcuXPT3Nq+TJkyNF+7dm1cLZ07d6a1WMfjrFmzaF6iRAmat2rViubFihWLqyVdunS0lg0bNtAxrHNbxYoVaf7111/T/NChQ/SY6dixI62nRYsWdJxatWrRfMeOHTRPyXnGOUdryZ49Ox2jSJEiNP/ll19o3qVLF5o3aNAgrpYKFSrQWvLly0fH+Oyzz2i+c+dOmh89epTm2bJli6tl27ZttJbt27fTMZo2bUrzFStW0Nx6Hvfs2UOPmb59+9J6rPfb06dP09w6hidNmkTzzp07x9Wza9cuWsvll19Ox6hatSrNy5YtS/OkpCSat27dmj43sXTFSkRERCQimliJiIiIREQTKxEREZGIaGIlIiIiEpFUu3n99ddfp/ngwYNp3qtXL5pbN7sPHTo04VqsG3NbtmxJ8yeeeILm9evXp3mWLFkSrgUAKleuTPOaNWvSvF+/fjS3npsjR47JTKRiAAAPf0lEQVTQPH/+/HHZggUL6Lbr16+n+QUXXEDznj170rxt27Y037t3b1z21Vdf0W1/+OEHmn/00Uc0L126dIq2Z84//3yaWzdrWjfVWjfvWzehWrZs2ULzO+64g+Z16tSh+erVq2nesGHDhGuZP38+zWvXrk3znDlz0ty6SX3dunUJ12IdX5kyZaK5dVO/dfP60qVLE65l4sSJNB81ahTNM2fOTPNrr72W5vv370+4FgB47rnnaG41Dbzxxhs0nzt3Ls2t19N9990Xl3lP70M2zxtbt26leevWrWluHe8NGjSIy6zn/bzz+HUIq2Hn4MGDNLdeS+x5bNeuHd3WOj9Yz8vmzZtpbp3HLFZD0OjRo2nO9jUADBw4kObWOZ6xmgCsGq3mMmsucubMmYRrSU5XrEREREQioomViIiISEQ0sRIRERGJiCZWIiIiIhHRxEpEREQkIqnWFbh8+XKaDxs2jObO8d8iX6xYMZpbS3MwVvfcAw88QPPPP/+c5mPHjv3TtQDAnXfeSfPdu3fT/KKLLqL5pZdeSvNSpUolXEuGDBloniNHDppbnZG33norza2lFphChQrR3Hq+rM7FJk2a0PzLL79MuJb27dvTnHUzAsD06dNpbnV6Wh2pllWrVtHc2k9XXnklzXPlykVzq6Nv3LhxcZnVZWM9Z9a5wOqqWrt2Lc0Z63Faz6/VETd58mSaW0vjMIcPH6b5okWLaM6eW8BeesnqOrR069aN5tY5tXjx4jS3upKtOpmVK1fS3Oqabdy4Mc2tbrmnnnoq4VqsZamszsI8efLQvE2bNjS3jiXG6jK2zm3We1OHDh1oXq1atYRrAYDChQvT3DpvWMeA1WFqvf4Yazkeq2vW2k/Tpk2j+bPPPktzq3M6lq5YiYiIiEREEysRERGRiGhiJSIiIhIRTaxEREREIqKJlYiIiEhEUq0rcOPGjTS31vbq27cvzSdNmkTzDz74gOZdu3aNy6y1yr799luaZ82aleZLliyhudXxYq1naHXzWetmWV0m+fLlo/mgQYNo/vjjjydci9XtYXWZHDp0iOZWpyOTLl06mlvrSFrr8y1evJjmuXPnTriWZs2a0dxaj+rGG2+kubWu3ksvvURzdvwCdjeQdezddtttNE+blp8SpkyZQnPG6th6+OGHaW51CFnHdffu3Wk+YMCAuKx37950W6vbtVatWjTv2LEjzevWrUvze++9Ny6zOmCtY2D27Nk0r1GjBs2tNdKsLuB69erR3Fp70+qss7odrbX1GKvLrVGjRjS33iesjsaRI0fSnHWqWp262bJlo7n1Wv3+++9pPmPGDJo/9NBDcdmmTZvottY6j9999x3NrY7n8ePH09xirR373nvv0dzqSrZe2ydOnEi4Fuv98Pjx4zS31si1arz44osTriU5XbESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhERBMrERERkYikWlegtV7bsWPHaF60aNEUbd+jR4+Ea8mcOTPNk5KSaM46fgC7Y2vChAkJ1wLYHV5WV5W1Dp21zpbVMclYa6F16tSJ5o888gjNGzRoQPNRo0YlXAvrWgSALVu20Nx6nG+//TbNre4ptkaYtb6bte7UmTNnaG51Pe3fv5/mlpSu71a2bFmaW52qKWGt5Wd18G7dupXmS5cupbnVBcrcfvvtNG/Xrh3NT506RXNrLcZvvvmG5uwc8corr9BtrTUUrQ6sdevW0dxaI80ya9Ysmltdbhs2bKC51e24bNkymj/22GNx2ddff023tbo6rW5M69iw1vBkKlSoQPOPPvqI5lYnrbV2o7UeLnPzzTenKN+3bx/N9+zZQ3Or09pirc938uRJmlsdltY4Vvf49ddfH5f16dOHbmu9bqxuzGuuuYbmVodlInTFSkRERCQimliJiIiIREQTKxEREZGIaGIlIiIiEhFNrEREREQi4rz3qV2DiIiIyH8FXbESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISEU2sRERERCKiiZWIiIhIRDSxEhEREYmIJlYiIiIiEdHESkRERCQimliJiIiIRCRtav3gw4cP09Wfv/zyS7r94sWLab57926aP/TQQzQvU6aMS57lyZOH1mItUP3WW2/RvF27djRfsWIFzb33cbUAwODBg+kPTpMmDR0nR44cNM+XLx/NJ06cSPMJEybE1bNhwwZaS5cuXegY1v6oWbMmzWvVqkXzUqVKxdUyevRoWsuDDz5Ix9iyZQvNM2bMSPNnn32W5gMHDoyr5dprr6W19OjRg45RpkwZmrdp04bmt9xyC827du1Kj5mePXvSet599106zpIlS2g+d+5cmjdp0oTmx44di6unYsWKtJYJEybQMXLmzEnzVatW0fzGG2+kOYC4WpxztJY5c+bQAXbu3Enz/fv30zx37tw0b9asWcK1nD59mo5xySWX0PyTTz6heVJSEs1z5MhBj5nKlSvTer7//ns6ToUKFWj+6KOP0vy5556j+bBhw+LqadGiBa2lXr16dAzLE088QfN+/frRvHbt2nG1jBgxgtby2muv0TGyZ89O888//5zmmzdvpnlSUlJcLc899xyt5fDhw3QM6/Vbo0YNmrds2ZLmuXPnpsfMe++9R+uxzvvPPPMMzVu3bk3z0qVL07xq1apx9fTp04fW8vLLL9MxWrVqRfOBAwfS/KWXXqJ527Zt6XMTS1esRERERCKiiZWIiIhIRDSxEhEREYmIJlYiIiIiEdHESkRERCQiqdYVWLRoUZr36tWL5n379qX52LFjaW7d6T916tS4rFixYnRbqytgypQpNK9SpQrNmzdvTnPLhRdeSPONGzfS/ODBgzRv3LgxzYsXL55wLVdeeSXNrc6LAgUK0Lxp06YJ/0wAKFWqVFz24osv0m2tTi7reZ8+fTrN//3vfydYHbB06VKad+zYkeZ58+al+W233Ubzzp0707xr1640t7rWrOfd6qQcNWoUzT/88EOaM/Pnz6e59VrNnz8/zdkxAAD79u2jea5cueIyq2tv+fLlNLe6KK3OtwULFtC8WbNmcdn1119Pt7U6pBYtWkRza99Zz8vq1atpPmnSJJr/+OOPNF+4cCHNH374YZpb52DG6jpduXIlzU+dOkXzV155hebr169PuJbZs2fT3OrstboFL730UpqfOXMm4VrKlStHc6sz1nqfbNGiBc3z5MlDc6sj3npM1vnN2n/p06en+YABA2hetWrVuOypp56i21pdmh06dKB5pkyZaJ45c2aat23bluaxdMVKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISEU2sRERERCKSal2BVlfO7bffTvMNGzbQ3OoiLFiwYMK1WGuVderUieZWd9odd9xBc2vdN0v//v1pbq11dM0119Dc6l601odinSB79uyh21rdVtdeey3Ns2XLRvO9e/fSnLGOmQwZMtDc6uq0uka2bduWcC0//PADzS+//HKaWx1hVhfT/fffn3AtgL0W2ptvvklzq8vvwIEDNLfW1rvhhhvisu3bt9Ntrf1hrSeWLl06mhcqVIjmbI07qyPOem1b63Fa67tZr1XGWu/U2tfjxo2jubXO42WXXZZwLYC9tqBzfCk0qxPW6lCzOr+Yu+++m+ZWB5a1pmz37t1pXr58+YRrsc6bVlfroUOHaN6gQQOaHzt2LOFarK7eadOm0dzqkLbWTWXddufyzTff0NzqjBw+fDjNixQpQvNq1aolXIu1ZmZK15e0uvnr16+fonFi6YqViIiISEQ0sRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiETEWWsC/dUqV65Mf7C17p3V3WR1F1hrF+3cuZO1vNBasmbNSsewupXmzZtHc2udqvr169P2mxtuuIHWY627tHXrVppffPHFNB80aBDNH3/88bh6nHO0FquT4vXXX6f5Bx98QPOrrrqK5uXLl4+rpVOnTrQWaz3Ddu3a0fzEiRM0L1myJM03bdoUV8sFF1xAa+nSpQsdw1rn0VoX8qabbqJ5o0aN6DFzxRVX0HqGDh1Kx3n++edpztbbA+yuwH379iV8zKRNy5uQ33rrLZqvWrWK5lZX1YABA+JqKVu2LK3FWhPQ6t7MmDEjzRs2bEjzp59+Oq6WTZs20VpeeOEFOkbNmjVpbp1/hg0bRvOZM2fSYyZNmjS0HqvLbdeuXTQ/ffo0zWfNmkXzXr16xdWzbNkyWsuyZcvoGFZXsrW2Xr58+WjuSAtk48aNaS2TJ0+mY2TJkoXm1mvMWiuwZcuWcbUUKlSI1vL+++/TMawO9CeffJLm1r7u1q0bPWYmTpxI61m3bh0dZ8eOHTS3Ol6tDuHOnTvH1ZMtWzZay0MPPUTHuO6662jep08fmv/88880X7NmDW+bjaErViIiIiIR0cRKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISkVTrClyyZAn9wfv27aPbW50w1lpEI0aMoHmRIkXi7ujPmjVriroLTp48SXOri8daI6x37960uyBLliy0HmutqunTp9PcWt+sePHiNB8/fnxcPQ0bNqS1tG/fno5RpkwZmp93Hp/DW51GhQsXjqulbt26KerWsTqBrPXwrG6zzJkzx9WSNm1aWsuECRPoGFaHYu3atWludc147+kxU758eVrPwoUL6ThjxoyheZs2bayfS3MAcfVUqlSJbtyxY0c6gNU9ZR1js2fPpvnYsWMT7vitVKkSHWPkyJE0t7paFy1aRHPWiXf8+HFai/XasDpG16xZQ3NrDUWQfQTYna3Wz7X23yWXXEJz6zzZpk2buHpuv/12Wkvz5s3pGE2aNKF5xYoVaW6tlbdjx464WvLly0drsc7jp06dorn12nv55Zdpvnjx4rhaPvroI1qLtSagdWwUKFCA5lb3cYkSJegxU6RIkRR1klrr7TVr1ozm1tqIGzZsiKunePHiKeqyPXr0KM2tLs2yZcvSvFChQuoKFBEREfm7aGIlIiIiEhFNrEREREQioomViIiISEQ0sRIRERGJSKp1BYqIiIj8t9EVKxEREZGIaGIlIiIiEhFNrEREREQioomViIiISEQ0sRIRERGJiCZWIiIiIhHRxEpEREQkIppYiYiIiEREEysRERGRiGhiJSIiIhIRTaxEREREIqKJlYiIiEhENLESERERiYgmViIiIiIR0cRKREREJCKaWImIiIhERBMrERERkYhoYiUiIiISEU2sRERERCKiiZWIiIhIRDSxEhEREYmIJlYiIiIiEdHESkRERCQi/w/oP4GZcqUmfQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 27 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "log = pd.read_csv('logs/' + log_filename)\n",
    "log['epoch'] += 1\n",
    "evaluate_model(model, X_val, y_val, log=log, pre_train_weights=pre_train_weights)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Best Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9961904761904762\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model2 = keras.models.load_model('./mnist-inception-chkpt-29-0.9962_2018-12-02_13-58-36.hdf5')\n",
    "evaluate_model(model2, X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Some other decent models for comparison."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9959523809523809\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 17 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# best model 1st 60 epochs\n",
    "model2 = keras.models.load_model('./mnist-inception-chkpt-37-0.9960_2018-11-25_01-29-15.hdf5')\n",
    "evaluate_model(model2, X_val, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9961904761904762\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# best model 2nd 60 epochs\n",
    "model2 = keras.models.load_model('./mnist-inception-chkpt-27-0.9962_2018-11-25_01-29-15.hdf5')\n",
    "evaluate_model(model2, X_val, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9952380952380953\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlIAAAFeCAYAAABO9fGVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzs3Xm8TeX+wPHPN0SoyFRkKGMloRJJKgmXVJerNFFXGq7mueh30xzN5VZyb0lp0EySJqRUUiklDYSQqTKEwvf3x7Oetfc5Z59jn332eM73/Xqdl2ONz3PW2ms/6/tMoqoYY4wxxpji2ynTCTDGGGOMyVVWkDLGGGOMSZAVpIwxxhhjEmQFKWOMMcaYBFlByhhjjDEmQVaQMsYYY4xJkBWkTKFE5GERGVaC/QeKyPvJTFO+408WkQFR/79ZRFaLyAoRaSAiG0SkXALHbSQiKiLlS5C2uP92Jf075zvWIhFplIxjGWOM2TErSJVBwZftnyJSM9/yz4MCRCMAVT1PVW/KRBrjoao9VPUJABGpD1wO7K+qe6rqYlWtqqrbkn3e4O+3SUTWi8hvIvKBiJwnIuHnqTh/u+htReQoEVma7DQbJ7h2x6bxfL5AH/2jInJ5IdtPzrftnyLyZYztOgfHuTlq2UAR2ZZv/6Oi1h8uIh8H9+1cETki3zFricjTwT39q4g8lW/9sSIyR0Q2isgSEekXta5c8CKzLDj+ZyJSLVhXUUTuCdb9KiKjRKRC1LoxIvJT1H49Cvnb/F+Q52OjlvULPn9/iMh7RVyHAcG+g6KWVQxeYn4RkbUi8pqI1ItaP05ElovIOhFZEL3vjtJlyhYrSJVdC4H+/j8iciCwS+aSU2INgTWqujJN5zteVXcNzns7cDUwJk3nzriSROvKkqgCfVVVrQocCGwHXihk+x75tv8AeD56m6AQch/wUYxDfBi9v6q+F+yzB/AqMAKoBtwJvCYi1aP2fRFYgbunawMjo865P/A0cD2wO9Aa+DRq3xuBw4EOwG7AGcDmYN01wCFAS6AZ0BYYGqwrDywBOgfHHQY8lz+qKiKNgb7A8nz5XQvci/sMxhTk8VpgXr5VFwfpbQXUBX4DHohafxvQSFV3A3oDN4vIwXGmy5QhVpAqu54Ezoz6/wBgbPQGIvK4f+MVkZoiMjF4W10rIjN8BEZE6ovIiyKySkTWiMiDsU4oIvcFb7LrRORTEekUta6diMwO1v0iIncHyysFb4ZrgnN/IiJ1gnXvicig4E1wKlA3eAt/XPJVz4nI7sGb73IR+Tl4ey4XrCsnIiPFVQv+CPSM94+oqr+r6qvAycAAEWmZ/28X/P+q4NzLgjSriDSJ3lZEqgCTo/KxQUTqFva3KS4JqlqDvP4qIguj3/6Dc70aXN/vReScqHX/FpEJwbVYBwwMlj0fLFsvIl+KSDMRuVZEVgbX+rgi0hPzvhGRnURkqLgoxUoRGSsiuwfrCkTsJCrKFKTpuWCf9SIyT0QOCdY9CTTAFSA2BNek0PsrRc4Epqvqoh1tGBQmOuE+q9EuB94E5hfjvIcDv6jq86q6TVXHAauAvwfnOg6oD1wZ3NN/qepnUfsPBR5R1cmqulVV16jqD8G+1YFLgHNU9Sd1vlJVX5A6HrhfVdeq6irgfuBsAFXdqKr/VtVFqrpdVSfiXvLyFFiAB3EvK39GL1TVt1T1OWBZEXm/LTjn6nzL9wGmqOovQVqfAQ6IOvY8Vd3i/xv8NI4nXaZssYJU2TUL2E1E9gsKFCcD44rY/nJgKVALqANcB2iw70TgJ6ARUA/3QIrlE9yb7B64t9vnRaRSsO4+4L7g7a8x8FywfADuTbU+UAM4D9gUfVBVfQvoASwL3sIHxjj3E8BWoAnQBjgO8KH6c4BewfJDcG+YxaKqH+P+Pp3yrxOR7sBlwLHB+TsXcoyN+fJRVVWXUfjfJhGHAd8CNXFRiTEiIsG68UEe6uL+BreKSJeofU8AJuAiGr7a53jcF3114DNgCu65Ug8YDjwSKxE7uG8GBj9HA/sCVXFfWPHqHRyrGi4K8yCAqp4BLMZFE6uq6p3EcX9Fpdm/SMT6mRhn2s7E3YvxbjtDVRdGpaEhrhAyvJB92gQvBAtEZJhEIocS/OTJEi5KBNAed188ERQqPxGR6Pu0fXD+L4MXgnHiolzgomxbgb7i2icuEJF/5TuP5Pv/3r5wnCdBrhDbjKjokYj8A/hTVV8vJM+FEpF2uM/0wzFWjwE6Bi8QlYHTcC8y0fuPEpE/cIXW5cDrUesSTpcpXawgVbb5qFRX3IPi5yK2/QvYC2gYvK3OUDdRYzvcF++VwdvlZlWN2cBcVccFb7JbVfUuoCLQPOr4TUSkpqpuUNVZUctrAE2CN+lPVXVdcTIZPJx7AJcEaVwJ3AOcEmzSD7hXVZeo6lrcG2wiluEKifn1A/4XvOH+gasGKY7C/jaJ+ElVRwdtx57AXdM64tqYHQFcHVzDz4HHcFU03oeq+nIQOfCFjRmqOkVVt+KqoGoBt6vqX7jCTCMJ2srkU9R9cxpwt6r+qKobcNUyp0j81Ynvq+rrQR6fBA4qYtu47y9V7aWq1Qr56bWjRImLwNbBFUbjcSbweL5l9wPDgr9LftNxBaPaQB9c1f2VwboPcJHO/iJSQVwnjcZA5WD93riXi3eBPYG7gFck0o5yb9y90AdoimsG8EDUut1xBaB9cIXwf4tI12D9ZOBicW2w9gQuCpb7cwNhleVTwBOqOj9YVhW4FRfxKpagsD4KuFBVt8fYZAGuYP0zsA7Yj3wFVFW9ANgV94L0IrClpOkypY8VpMq2J4FTcW//Y4velBHA98CbIvKjiFwTLK+P+3LeuqOTicjlIvKNiPwuIr/hHr7+Qf1P3IN4fvA27L+YnsRFOZ4RVy12Z/DALY6GQAVguY8g4CIltYP1dXHtNLyfinl8rx6uzUZ++Y+/JMY2RSnsb5OIFf6XoFAHLuJTF1irquujtv0JlycvVrp/ifp9E7BaIw38fWGraoz9irpv6pL3GvyEa0sTb5Xbiqjf/wAqFVEIS8b9Fa8BwAuFFILyENcQfE+iCl0icjywq6o+G2ufoOC5MCjofokrFPQN1q3BRRQvw12z7sBbuAgkuGu1SFXHBC9Kz+Cud8eo9f9T1QVB+m8F/ha1DmC4qm5S1bm4QrRffwsuWvk5rkD3Mq4AG7ZnFNdM4ElcFdmQqGzdCDwZHZUrhguAuar6YSHr/wNUwhWkq+AKSpPzbxQUsN/HFRjPT0K6TCljBakyTFV/wrVH+BvuIVLUtutV9XJV3RdXnXNZUO2zBGiwo2hB8DZ+NS46U11VqwG/E4T8VfU7Ve2PK9zcAUwQkSrBQ/1GVd0f186jF3nbdsVjCe5NsmZUBGE3VfXtIZbjvti9BsU8PiJyKK7QESsatxz3EPbqx9jG0wILCvnbFDeNO7AM2ENEdo1a1oC8UcoCaSuBou6bZbjCb3Q6tuIKABuJimQEUYdaxThvnjwU5/6Sgj3qon8KfAHn23cX4B/EX603AHgxX6GrC3BIUH22Alcdf4mIvFLIMZSoKjVVnaaqh6rqHrjoUnPg42D1XIq+vkWtnxt1voKJcIWrIapaL3h+rAE+9QXuoGp5DK6g3CeIZkbn+aKoPNfHNUa/uoi0Ru97UtS+hwN3SaQN50HA4+rabm3BRdjaSb7ezFHKE2kjVZJ0mVLGClLmn8AxQfucQolILxFpEjz01gHbgp+PcQWF20WkirjGux1jHGJX3JfhKqC8iNyA693jj3+6iNQKQvC/BYu3icjRInJg8IW5DvcmW6whDVR1Oa5x7l0ispu4xsyNo9qAPId7KO4truHsNYUeLJ/geL1wb+DjgkhAfs8BZ4lrj1YZuKGIQ/4C1IhuP1LY3ybeNMZDVZfgogW3BdewFe7eeKroPRNW1H0zHrhURPaJqkJ5NoheLcBFmHoGkaOhuCrieP2Ca3cFQHHuL83Xoy7fT8wu+1FOwl27d3eUwKhC1+P5Vg3DRSZbBz+vAqOBs4L9ekikI0aLYPuwkCUibYJqvd1wPfKWquqUYPVLQHVxwwSUE5G+uBeDmcH6/+Hu4X2De/hqXBs31DU6nwFcL25Igf1whbyJwXnriWuHJCLSPkjX/0Xl6z+4arXjo6qMvS646kqf52XAucBDwbHLiWtnWR7YKbiPfERxYHBcv+9sXCTp+mD9J8CZ4jqiVMBFsJap6moRqS0ip4hI1eAc3XBVpe/Eky5TtlhBqoxT1R9UdXYcmzbFVQVsAD4ERqnqe8Fb5fG4RtSLcVUFJ8fYfwoubL4AV1WzmbxVRd2BeSKyAde4+hR1PWl89cY64BtgGkU3ii/MmcDOwNfAr8Ex9wrWjQ7S9wUwhx1E5wKvicj6IA/XA3cTfKHlp6qTcW1b3sVVj/qqhi0xtp2PK0j8GFRD1qXwv02y9cc1/F6G+2L9P1WdmoLzsIP75r+4ap7puIjpZuDCYL/fcV94j+GiZRuJVE/F4zZgaPC3vYLk3V87MgAYq6p5ojYi0im4rtFOxEVr8xS6gqjwCv+Dq1LbqK5dH7gv97kishHXKPpFXCHUuwrXc20J7t4/KerYa3GN9K8Izn0NcIKqrg7W/xdX/f8R7vO7hUhbJ3D3TkNctGkSrh3X28G6xrhC+kZcRO4aVX0zyH9DXAGkNbAiKsJ3WnDeNfnyvA34NSpSd0bwd/gPrh3TJtznGVX9Ld++fwLrgnuIIK+bge9wL3h/i/qbKK4abynueTES18bylTjTZcoQyfe5NsakWPDG/hVQMZ62ZcU89iLgKI2je70xxpiSs4iUMWkgIieJyM5B1eEdwGvJLkQZY4xJPytIGZMe5+KqD37AVQOcX/TmCbuXSDsqY4wxKWZVe8YYY4wxCbKIlDHGGGNMgqwgZYwxxhiTICtIGWOMMcYkyApSxhhjjDEJsoKUMcYYY0yCrCBljDHGGJMgK0gZY4wxxiTIClLGGGOMMQmygpQxxhhjTIKsIGWMMcYYkyArSBljjDHGJMgKUsYYY4wxCbKClDHGGGNMgqwgZYwxxhiTICtIGWOMMcYkyApSxhhjjDEJsoKUMcYYY0yCrCBljDHGGJMgK0gZY4wxxiTIClLGGGOMMQmygpQxxhhjTIKsIGWMMcYYk6CcL0iJyB4i8pKIbBSRn0Tk1EynKZlEZIiIzBaRLSLyeKbTYxIjIuNEZLmIrBORBSIyKNNpShURaSoim0VkXKbTYhIjIqeIyDfBc/UHEemU6TQlg4hsyPezTUQeyHS6kk1EGonI6yLyq4isEJEHRaR8ptOVTCKyn4i8IyK/i8j3InJSptKS8wUp4CHgT6AOcBrwHxE5ILNJSqplwM3AfzOdEFMitwGNVHU3oDdws4gcnOE0pcpDwCeZToRJjIh0Be4AzgJ2BY4EfsxoopJEVav6H9x3xibg+QwnKxVGASuBvYDWQGfggoymKImCQuErwERgD2AwME5EmmUiPTldkBKRKkAfYJiqblDV94FXgTMym7LkUdUXVfVlYE2m02ISp6rzVHWL/2/w0ziDSUoJETkF+A14O9NpMQm7ERiuqrNUdbuq/qyqP2c6USnQF1fYmJHphKTAPsBzqrpZVVcAbwClKcDQAqgL3KOq21T1HWAmGfruz+mCFNAM2KaqC6KWfUHpumFMKSEio0TkD2A+sBx4PcNJSioR2Q0YDlye6bSYxIhIOeAQoFZQXbI0qBbaJdNpS4EBwFhV1UwnJAXuA04RkcoiUg/ogStMlRZSyLKW6U4I5H5Bqirwe75lv+PC0cZkFVW9AHdvdgJeBLYUvUfOuQkYo6pLMp0Qk7A6QAVctKYTrlqoDTA0k4lKNhFpgKvueiLTaUmRabiAwjpgKTAbeDmjKUqu+bho4pUiUkFEjsNdz8qZSEyuF6Q2ALvlW7YbsD4DaTFmh4Iw9PvA3sD5mU5PsohIa+BY4J5Mp8WUyKbg3wdUdbmqrgbuBv6WwTSlwpnA+6q6MNMJSTYR2QmYgntZqwLUBKrj2r2VCqr6F3Ai0BNYgYuCP4crNKZdrhekFgDlRaRp1LKDgHkZSo8x8SpP6WojdRTQCFgsIiuAK4A+IjInk4kyxaOqv+K+jEpjdVe0Mym90ag9gPrAg6q6RVXXAP+jlBWGVXWuqnZW1Rqq2g3YF/g4E2nJ6YKUqm7ElbqHi0gVEekInAA8mdmUJY+IlBeRSkA5oJyIVCpt3VhLOxGpHXQnryoi5USkG9AfeCfTaUuiR3EFw9bBz8PAJKBbJhNlEvI/4MLgvq0OXILrHVUqiMjhQD1KZ289gijiQuD84PujGq492BeZTVlyiUir4Puwsohcgeuh+Hgm0pLTBanABcAuuPrS8cD5qlqaIlJDceH2a4DTg99LVXuFMkBx1XhLgV+BkcAlqvpKRlOVRKr6h6qu8D+4avfNqroq02kzxXYTbviKBcA3wGfALRlNUXINAF5U1dLcBOTvQHdgFfA9sBW4NKMpSr4zcJ12VgJdgK5RPaPTSkpnhwVjjDHGmNQrDREpY4wxxpiMsIKUMcYYY0yCrCBljDHGGJMgK0gZY4wxxiQord3oRSRnW7araqwh6QsoC3ks7crCNSwLeSztysI1tDxmN8ujYxEpY4wxxpgE2cCOxhhjQpdeeikdO3YEoG/fvhlOjTHZzyJSxhhjjDEJsoiUMcaYUPPmzWnevHmmk2FMzrCIlDHGGGNMgnI6IrXrrrsCsGjRIvbYYw8A5s+fD8AxxxwDwPLlyzOTOGOMyVHvv/9+ppNgTM7IyYLU7rvvDsAzzzwDQLVq1ZgzZw4Ar7/+ep5tSlNBqkaNGgA88MADAPTv3x+A9evX06FDBwDmzStN8zXnnh3NXSlSdnru33///QBceOGFAIwZM4ZBgwZlMkmmDKlZsyYAb775Jvvttx8As2bNCpfdddddAPz555+ZSWCSVa5cGYBy5coBcNFFF1GlShWAsPPAG2+8AcCDDz7I5s2bAfjrr7/SndRSx6r2jDHGGGMSlHMRKRFh+PDhABx33HEAbNiwgcGDBwPw6aefZixtqXbVVVcBcPLJJwOwfft2AKpUqUL9+vUBi0hlwo6iUEVtWxojVHvvvTcAZ511FhC5T7t160alSpUAwrdhY1LFR6QOOuigcNmRRx4JQOfOncPmIFdeeWX6E5dk7dq1Y+LEiUAk39H8c6ZTp04A3HLLLTzxxBMAXHHFFQCsWbMmHUlNOR+FO+ecc+jTpw8AjRo1AmDYsGE8/vjjST+nRaSMMcYYYxKUcxGpnj17MmTIkDzLmjRpwqpVqzKUovQYMmQIl19+eaaTYSheBKosuuSSS4BImw1v5syZpaY9Sml20kknceutt2Y6GSX2ww8/AHDAAQdw9dVX51k3YMCAMFLl79M//vgjvQlMgvbt2wPw8ssvx4xEFWXAgAGA+/uAi9Jt2rQpuQlMo27dugFw3XXXAXDEEUeE63xEbvDgwSmJSOVcQWro0KHh79999x1ATl/8HWnWrBkAV199NT/99BMA27ZtA6Bx48YZS5dJDl8oKy1VfJUrV6Zr164x123YsCGs5jPZ56STTgLc8/Spp57KcGpKzjeinj9/fljN7A0YMIAuXboAkYbYU6dOTW8Ck+D//u//AKhdu3aR240fPx6Ad999F4BHHnkkXHfIIYcAcPPNN+fcy3q9evU488wzAbjxxhuBSGP7WD744IOUpMOq9owxxhhjEpQzEakePXoAcPDBB4fLmjZtCsC3337LlClTAFeqBli4cCGQ+9Uw/k2iVq1aYUNJ37DeIlIm2xx44IG0bNky5rqXX345zakxxeGrRKZPn87q1asznJrUeuihh7jgggsAOPXUU4HcjEjF4vOx5557cuCBBwKE//prPHXq1AKR4xYtWqQxlSXTrl07wA2p4qsmY33Xr1+/HoDbbrsNIBzyItksImWMMcYYk6CciUg1aNAAgJ122ilsFLhy5UrAdW30Def8v//6178AePjhh9Od1KTwXXN9A8KJEyfy8ccfAzBjxow8227evLnUv0Ga7Obbmfg3v2iffPIJQBg1NtnFdxf3ja7Lwqjmd911FwMHDgQiw+j4IWSWLFmSqWQV2/fffw9EGlpH++CDD8JIlI8Sv/LKKwD88ssvBbb/8ccfU5XMpNltt90AGDduHFB0rcyWLVvCWpy5c+emNF05U5Dq169f+LvvUfLYY48BcNppp3HRRRcB0LBhQwDuvfdewBVEHnroIQB+/fXXtKW3JCpVqhTe8CNHjgTgf//7X9jLZKed8gYS33jjDWbPnp3eRJqkUtWcbnDuP5++QBXN99SzEZSzk6/S8RMVv/jii5lMTlosWrSIsWPHAnDeeecBBXuZ5oJnn30WiAQOgEI7ewAFqvpiHSubTZgwAXA99b3CnpuVKlWib9++QOoLUla1Z4wxxhiToJyJSEVbsWIFQDh21L333st///tfAG666SYgUkK/8cYb+e233wDCyFS2N0Dfd999w6q9V199FXBdx3031fwRKR+9MpkhIkm5p+I9RjZFrvzbvP/X5B7fANlXy0Y3E/DVfn5ohClTppSaMfu++eabTCchK/hqPv+9ms18jVOsZ2X+ZVu3bg3HEks1i0gZY4wxxiQo5yJS27dvj/lGtG7dOgAuvvjiPMuHDBnCfffdB8DSpUuB7O2G7ecDevbZZznjjDMAwkE4ITLHXn4WkUqvTEeEMjWIp2+j17t3bwC6dOlChw4dAChf3j1K/GCxEBkYz68rX748W7duTVt6zY61aNGCE088EXDzr8VaD4SjQV9wwQU8+uijaUufKZrvgHT11VeHcwbGGuHcD0S58847A3mHEfIDW/uG69nMDxh6xx13AK7xuW+AXrVq1Tzbjho1KpxPMNVyriC1efPmcHLGogwbNgxwI4P7Xhl+VPRsLUjtt99+gJtEcs6cOXnWtWvXjtatW+dZtnbtWiDvl5fJDF+oyfZq45K4++67ATjqqKMKrPP34JVXXkn//v0BOPTQQ4HIFBT77rsvCxYsSENKTbyOPPLIIgvk/pnkt3n44YfD5+n9998PuHGnTGb4jhwjRozgwQcfBKBNmzaAm07NF6B8J5BrrrmmwDFmzpyZjqQmhf/u952rtm/fHi7zhUPflMd3OEsHq9ozxhhjjElQzkWk4uWr+s4991w+/PBDgDCi06tXr7iiWunm3x42b95cYN3gwYOpUaNGnmVvvvkm4BqiG5Nq3377LZA3IuWrnkeMGAHAf/7znzAi5S1btgzAxjrLQi1atAgbXc+fPx9wUaprr70WiIyxFB1p9VWB/vmTqxEpP4yOj7Zlusq+pPycs/575IMPPgir8nxtTHQely9fDkSGEcolvmH80KFDw05Y/h597rnngLzNYlLNIlLGGGOMMQnKuYjUTjvtRJ06dYDYo7Pmt3jx4nC7PffcE4iMYJttfv/99wLLfIPdww47rMC60aNHpzxNxni+I4dvnFqlShWeeeYZoOiGqs2aNQPc7AS+XZ/JDp06dQrbQfko1PXXXx++3a9ZswYgjIYvWbIkjADkeoTR5zH/v6WJjxq2b98eyJtHPyBpuoYISCbfsPzcc88No2zz5s0DIkMgpZNFpIwxxhhjEpRzEalKlSrxt7/9DXDTpsTjiy++ACLdt/fff//UJC4FrrrqKiBvmn3d72effZaRNJnCpaP3XqbacvgpXu65554it3vjjTeASK89b6+99uLzzz9PTeJMwvy96iNSqhoOheDbz0yaNAkoPfN6tmzZkgoVKgCRiIxvV5vr/HAAPXv25PDDD4+5zV9//cV7772XxlQll+9BXLdu3fD+feSRR4BIm8x0yrmCVCJ8Acp75513MpSS4jv99NMLLPPpj1UVaLJDskY7z0WFVRX06dOHyZMnpzk1piirV68OZ0r4+uuvAdeA96WXXgIiXcr9i9ySJUvCcYpyuUD1zjvvULFiRSDS/d83vs4lfpyv4cOHh8v22WcfIO9YUZ4fx23YsGHhC08u8Y3no1/SfFMD38g8E6xqzxhjjDEmQTkZkerRowcQX9VeixYtwhHDfUPXjz76KGVpS5aWLVsCeRvGL168GICbb745I2kymZcLXbT9YLJ+GI9KlSoBULly5TD6sX379swkzuRxxhln0KBBAyAy/MEff/wRrvdDI/gIVcOGDXM6EuXVrFmzVESMTzjhBAD69u0b1/b+2t15550pS1Mq9ezZE4BWrVqFy/wcuitXrsxImsAiUsYYY4wxCcuZiJQfPK1du3a0a9cOiAxnEGvW6l122QVww8TvvvvuQGTQwEw0RiuOSpUqhUMb+NnXAd566y0AFi1alIlkGRMX3w3Zt5XyU8ScfPLJDBo0CMgb9TCZs3r16iIjTP46TZ06FXBTxLRt2xagwDRWuWrMmDGZTkLCiju9i28X1rlzZ6ZNm5aKJKXUeeedV2CZH6Ijk3KmIPX2228Dbq6dfv365VnWp0+f8I/pG9j5ufbat28fjvydKwWQ3r17h4VFb/369bzwwgsZSpHJlFyoyiuMn9PSF6QgUmXtJ1s1ucEXjl944YWwCjCXRXcGadq0KQDvv/9+JpOUFtWrVwdcz1o/+bgvJGe7Sy+9lK5du+ZZNnPmTJ566qkMpSjCqvaMMcYYYxIk6WxwJyIlPlmLFi3CEnTdunUB16XTNzTzy7zffvutQAQrEaoaV2igJHn0jeLfeecdGjZsmGfdFVdcscPxe0oq3jyWdsm4T6MV9zNWkihUOu7TeB177LEAvPjii4CrpvbV6r7b9saNG4t9XLtPnXRcw1TJ9H06efLkcB5B36En2U0+0pFH34TlwQcf5KyzzirqHD5N4bJ3330XgC5duiR6+pTm0T8jzj//fADOPPPMcIws3/Gqa9euRc6qkAzx5NEiUsYYY4wxCcq5iBRE2lmcc845AAwaNCjsYu35oQ7atm3LkiVLSnzOdLxd+GEdJk6cGC7zb0lHH310VpS8ywJ7008u39X68ssvD5f57tq+W31x2H3q2H2auE2bNoWDO+ZyRMqrWLEil112GRD5TJ155pl07twZIBzhPPr73g/O2a1bNyASoSqOVOWxTp06TJ8+HYi0YVPVsC20z+u4ceOKc9iExJMQtVapAAAgAElEQVTHnCxIZUI6PhS+Ue4HH3wQTlfQvXt3INLYM5XsC8qx+zS5GjduDLgeX8cccwxgBalksPs0cddee204Hl9pKEgVxhcWfSGpQ4cOgJvy58YbbwQIp23yBaviSFUeX3jhBU488US/L+CmtenYsSMAs2fPLlY6S8Kq9owxxhhjUsgiUnHKhreLVLM3facsXMOykMfSrixcQ8tjdktVHo877rhwfC8/bNHQoUMzMvaVRaSMMcYYY1LIIlJxsreLsqMsXMOykMfSrixcQ8tjdrM8OhaRMsYYY4xJkBWkjDHGGGMSlNaClKrKjn6A7cAXwDxgAlAlnv0KOdbRwKRC1jUEpgLzgW+AfXZwrKTkMR35A44JzuF/tgAnxXE8Q9bdpyOAr3H36QPATum4T9OcxynA74Wtt/s0tmy6hsH63YFlwEPJuobZlEdSdJ9mWR7vDM4xDzglXXnM1msYbx6zMSK1SVVbq2pL4E8gz3TP4iQj3WOBEaq6H9AOWJmEY8Yj5flT1XeDc7TGFar+AN4syTFNASm/jiJyONARaAW0BA4FOpfkmMWUrs/iCOCMJBzHFJSuawhwE5D+blVl4z5Nx/OmJ9AWaA0cBlwpIruV5JjFkNPXMBsLUtFmAE1EpJGIfCMio4A5QH0ROU5EPhSROSLyvIhUBRCR7iIyX0TeB/4e66Aisj9QXlWnAqjqBlX9I015ipaS/OXTF5icofyVFam6jgpUAnYGKgIVgF9Sn52YUnavqurbwPq05KJsS9k1FJGDgTpk/oWtLNynqcrj/sA0Vd2qqhtxEaLu6chQPjl3DbO2ICUi5YEewJfBoubAWFVtA2wEhgLHqmpbYDZwmYhUAkYDxwOdgD2jjneIiDwW/LcZ8JuIvCgin4nICBEpl5aMRdKTyvxFOwUYn7KMlHGpvI6q+iHwLrA8+Jmiqt+kJWNR0nivmhRJ5TUMIgV3AVemKTsxlYX7NMV5/ALoISKVRaQmroqsfhqyFcrVa1g+1SdIwC4i8nnw+wxgDFAX+ElVZwXL2+NKzzPFDR+/M/Ah0AJYqKrfAYjIOGAwgKrOBgYF+5fH/cHbAIuBZ4GBwblSLR35I1i/F3Agrl7YJFfKr6OINAH2A/YOjjdVRI5U1ekpzpuXtnvVpEw6ruEFwOuquiTYP93Kwn2a8jyq6psicijwAbAq2Lf488YkJqevYTYWpDYFbXtCwR9tY/QiYKqq9s+3XWtcdciOLAU+U9Ufg/1exl2kdBSk0pE/rx/wkqr+lWBaTeHScR1PAmap6oZgv8m4+zRdBal03qsmNdJxDTsAnUTkAqAqsLOIbFDVa0qU8viVhfs0LXlU1VuAW4L9nga+K0GaiyOnr2HWVu3twCygY/DGThCKbIbr2bSPiDQOtutfyP6fANVFpFbw/2NwPaOyRUnzR9R6q9bLnJJex8VAZxEpLyIVcA3N0161twPJuldN5pToGqrqaaraQFUbAVfgqmLSVYiKV1m4T0uURxEpJyI1gt9b4Tq5ZLrNW7SsvYY5WZBS1VW4qrjxIjIX9wduoaqbcSG9SeIanf3k95G8bU+24T7wb4vIl7iS7uj05qJwJc1f8P9GuPrtTPSiMSTlOk4AfsC1F/gC+EJVX0tjFnYoSffqDOB5oIuILBWRbunMQ1mXjGuY7crCfZqEPFYAZojI18CjwOmqmq6qvR3K5msomsYpYowxxhhjSpO0tpGSUj7fjikdysJ9WhbyWNqVhWtoecxulkcnJ6v2jDHGGGOygRWkjDHGGGMSZAUpY4wxxpgEWUHKGGOMMSZBVpAyxhhjjEmQFaSMMcYYYxKUjVPEGGOMyVIXX3wxAPfee2+4bOnSpQDUr5/WOW6NyQo5V5A64IADqF69OgCzZs0Kl+266655tvvss88A2LhxI8YYY0rutddeo0ePHgBs3749w6mJ2GWXXQAYP348J5xwAgDff/89AG+//Tb/+te/ANi2bVtmEphknTt3BuC9994DYOHChXTs2BGA5cuXZypZZZZV7RljjDHGJCjrI1KVKlUC4JBDDgHcG0fdunUBmDBhAgBHHXUUNWvWBMIZo5k0aRIAt912Gx988EFa02xMYerUqQNA+/btATjyyCMBaNu2LevXrwdgr732AuDQQw/NQArj1759e9q0aQPAyy+/DCT2Nvyf//wHgN122y1cdtpppyUhhaakfFXda6+5KR7333//8BkbrUqVKgCce+65ADzyyCNpSqFTvrz7KuvZsyd+2rPGjRuH/7Zr1w6A4cOHAzB58mS2bNmS1jQmU9u2bYFIVHDdunVs2LAhk0kq0ywiZYwxxhiToLROWlzc+XYqVarESy+9BMBxxx0X7zkAwreS008/nfHjxxfntDHZ/F5lRyLzQlWsWBEgjNA0b96cv/3tbwW2O/roowHYY489/LmAyP0azb9lF0cq577yUeFLL70UgGOOOYZatWoB8OOPPwLwxx9/MGTIECC+doqXXHIJd9xxBwDlypUDXNuWFi1aFLqPfRadVM5f1qBBAwAmTpwIuHao8Xj99dcBOP7444vcLlX36Zlnnsnjjz8ORKI1GzduZKedXMzAR87Wr1/Pq6++CsB9990HwNy5cwH4888/i3PKQqXys+g/gyNHjvTn4uyzzwZg7NixxT1cwrJxrj3/3GzevDktW7YEIpH/7du3889//hOI1HZ99dVXtGrVqtDjxZPHrCxInXXWWQDceOON1KtXr8D6+fPnA/DDDz8AcPfdd4e/+z/YySefDED16tXp1KlTCVNuD++yJN77tFmzZuG///73vwFo3bp1rOMBsQtL0et8z6dRo0YBcOeddxYz5cl7sPmHka8yHzFiRNiIt3Llyjs6NhD5Ul2xYgUAF154IZs3bwYIv9hGjBgR9gLz1S5PP/102FA4FvssOqn6cmrYsGFYyPBfRLH8/vvvALzwwgthtfSwYcOAHXfySdUXcIUKFXjmmWeASPV5vXr1wnv2nHPOAaBfv3506NAhz76+89KQIUOYM2dOcU4bUyoLGb6x+TvvvOPPxRdffAG4pi5AeE3idfDBB/Ppp58Wa59sKEj169cPIHw++ZeAww8/PPr8QOxnMERe4mKxSYuNMcYYY1IoqyJSZ5xxBgAPPPAAQIEhDcBFqZ566ikgEpEqSsuWLfnqq6+Kndb87C247NjRfeo7O0ydOhWAJk2ahBGcHUWdClu3bNkyrr/+eqBkoflkvSH66NB1111X5HHuuecegDDSFL2Pz++mTZsA6NOnD2+++SYQeWv0VYIQecueOXNmkee0z6KT7Lf8hg0bAvDqq68WGYnyvv32W8A1QC+uVEYyfI2Gj47tu+++BbapUqVKWBXWt29fIBLBWL9+fRgN9v8mMtRDOqI1vXr1Atzn1Ue0e/bsCUC1atXiikr5KPpVV13FqaeeWqzzpyOPtWvXDn/3HXD8s7Jx48Zh84L8z9e//vqL3377DYAZM2YArlNMkyZNgEhzjHvuuYdVq1YVen6LSBljjDHGpFBWRaR8qT9Wmny7qN69excZifL14r179y50m3Xr1nH77bfvOMFR7C247ChuG6k77riDE088EYj95uqjTqtXrw4jVz6S47ssT5o0iQULFpQw5cl5Q7z99tvDdksVKlQIl7/99tuAi54B3HDDDfz8889A3nznH93ar/PbQiTqduqpp4bLu3XrBkQ+64Wxz6JTkrd8H3GKjtb7dn4+klMYH4nynSkWLVpU7POnMpKx8847A5GoZ1Ht7aL1798fgEcffTRslD569GgAzj///GJHpdLZfqhGjRo8/PDDQGRohIULF3LssccWuo9/fvl2yLNmzQqjx/FKVR6PO+64sFH4P/7xD3+uAtstW7aMjz/+GKBAp7JVq1Yxbdq04pw2ppxrbB6rIDVv3jyAcDTd6IexV61atfDB3LVrVyDyYSrsPP4L7PnnnwfgscceCy9ILPbwLjuK+6Fv1qwZ33zzDVB09d0ll1zClClTAJJSaIolGQ+2bdu2FcjH888/H1aFFPdhG833/PONmWvXrs3TTz8NuB5X8bDPopPIF7AfW8lXS7/33nu88cYbADzxxBNA7Gfn2rVrARg8eDBffvklEH8BJZZsaKRcmN69e4fjonl9+vQJe5DHK1N5fPLJJwHYc889w+/DWJ599lkg0pO4ZcuWrFy5sljnSlYefXWb7yjQpEmTsGmPr557+umnw2eP75n566+/pnwkd6vaM8YYY4xJoawe2XzDhg3heCSxIlEDBw4EXLdq32BuR90cwXW99qMo+yrAnj17hvMx5X8bMaYoCxYsCEfSjzV2lHfTTTdxyimn5Fk2ffp0wN2vftwlz0dLs8H3339fokiU57ufRzcgNenjn6e+YfmAAQOKjOL7SJSPFk6ePDkdycyoNWvWFFjWqlWrYkekMiV66Ab/efPL/PAGnTt3pmnTpgBhFVqsfKfDKaecwogRI4BINeN3330XNih/6KGHMpKu4rCIlDHGGGNMgrKqjZRvK+EH0/zzzz/DN/hXXnmlwPa//vorkHeOLt/N0de1xlK9enVuvvnmAst9RMo32otm7TLKjkTaLFSrVg2IdMuNdtlllwGxo6T+nov11rVgwYKwEXu8baqS0WZh1qxZYVsmb8uWLeH8d4lGbA844IBwAMEaNWqEy/1wEvG2z7DPopPIfer/1r7tafSzM5Z3330XoMhGy4nIxjZSu+yyC+Da3/gGzn5g0QMOOIDFixcX63iZzuOuu+4aRh794Kk++nTxxReHzyP/XXnBBRcU+xzJyOO0adM44ogjgMgwKnfffXc40nymI/M519j8/PPPB/J+qfjedbHGs5k9ezbgQpK+x0889thjD2666SYABg0aBLjeSb5aMNaknPbwLjtS9WDr1atXGE73fO+9MWPGhFXVUeko8v6PJRkPtn79+jFmzBgg8uUCkWoePwn40KFDwwf0kiVLCj2XHzV4xIgRXHTRRXnWLV++nP322w8g7klX7bPolOQ+9ePm+KmKCuMb9fov4GTJdCEjFl/F6TuEQORFx0+sXRzZkEf/PXfttdfmWf7ll1+Gz55Mj1vXoEGDMLDRvXt3IO+Llq9yLKysUlRzHt+Tz+c/keYJ1tjcGGOMMSaFsioi5Uuh/o23SZMm4ZuTbzQ3c+bM8M3Yh6n79+/PXXfdlVCaoqsHfXfe5s2bF9jO3oLLjkx0uW7UqFFYLb3XXnv5dITV3BMmTIjrOMl6C/bzV/nJUf2Iwvl99913QGTOL4D3338fgE8++QSAq6++GoiMOB3tuuuuK/acgvZZdEpyn/rn5SWXXBJzvY/2+wiBf04mSzZEazw/QbYfBqJBgwa89dZbQKTzyNatW4t93EznsU6dOuFQFfkjj/fee284TIKfoy8Ryc6jn0i4V69eYbOeWFFT/93vx7SL1qhRIyDSoQIiVdR9+/YNh1OIl0WkjDHGGGNSKKuGP/B1ob6x98iRI8N5dHwD11NOOYXXXnsNiJRGE4lG+ZJvdHsoX2dsTLp17tw5jERFa9OmDRB/RCpZnnvuuTz/duvWjSFDhgB5h3jwoyP7fyHSaDWekaBjtUc0qbejwTS3bNkCuA4/pVWsSBTA559/Hg73kEgkKls89thjhbaBa9u2bRi5ufDCCwFSPrBlPHxj8wkTJhT5zPP58rVT0XyZ4aqrrgo7+vjnaM2aNYsdkYqHRaSMMcYYYxKUVREpzw98dvbZZxeYWfyZZ54Je1X4yJWfbiJeu+++exjh8sPQG5NJsQby3LBhQxgRyrQpU6aEbRcPOuggAG6++ebw7S+6XWFRc2Z6vh3Vf//735Sk15RMx44dAbj11lsBwrkXS4s6deoUiET5XqN9+/ZlxYoVGUtbsvTs2TP8DPp2UJ07dwZg/fr19OrVK2NpK6lYkSjPt6t+/fXXufzyywHCuRP9MDXJlpUFKT8JZo8ePXjzzTeBvA9qP9RBp06dgLwjsu60kwuy+Uawd999d7juwAMPBFyXVt9YzZswYQLjxo1LZjaM2SE/we9BBx1UoJrrmGOOKVFD0GRbv349ECkEHXXUUWH1gJ8sHCLVdX7YA9+F/NBDDw1Hbz/11FOByEPPZCc/npIv0M+cOTOTySkxP5zHqFGjwgLU0qVLgcg9+eOPP2YmcUniC0gPPvhgWKXlvwf9Zxhg4sSJ6U9cGvXu3TssSPrygO9EkWxWtWeMMcYYk6CsjEh5S5cuDeeG6tOnDwDDhg2jcuXKAAX+hcjbsJ+z56ijjiryHH7k2hEjRsQ9IKAxJeUjolOnTgXcUB/+7ck3+vTzYmUzHz32/0bzoxVHD53gO4jE6rZssk+dOnWASGeHrl278tVXX2UySSUydOhQAE466aTweT948GAgEmnNVV26dAHc/IkAt9xyC6tXrwYiUbeyoEePHgCcd9554bJRo0al9JwWkTLGGGOMSVBWR6QAfvjhB4Bw0L6FCxey7777ApFGr37QwkScffbZQOrqTo2J5t8aR48eDUQGjVPVcFBLPwhmrvPTU0TL5WhGWVa7dm3ARfpz8Rr6uSOvvPLKcJmPUvhG57nuww8/BODbb78F3HdmPFP7HHzwwUBuRMDjMWzYMAB23nnn8F5NZIqf4sj6glR+0RMYVqhQAcgbwvP85LHRjcp9WDd6lN7oxnfGpFKvXr3CXlC+oas3evTosGdUro/dc/rppwPQunXrAutiTQhu0u+vv/4CXAG+tI/ltcsuuzB8+HAAypd3X3nTpk2LOcF4Ltt7772ByHfaBRdcENe17du3L+BmJ9h9991Tl8A0ady4cfh7ugr9VrVnjDHGGJOgnItIRfNvVf7faH5+L2MyzY+gPHLkSJo2bQpExqzxY6Ddc889OR+JAjjxxBN54okngILjSL333nth42XfQcRXQ5j0euyxxwA47LDDwjkQS2tkqlWrVuGcgd4BBxwQdpBYsmQJAAsWLADcXK+PPvpoWtOYDD79fo66atWqhY3N586dC0SeO2vWrGHQoEF59svFPENklHPfXMJXQy9YsCActT3VLCJljDHGGJOgnI5IGZOtatSoEXaQGDhwIOAiNP4N8e9//zuQ+wMc5te1a9dC13Xo0CFsUO+HHRkwYEA46K5Jv3POOSfsAOE7PsSyePFiYMdz9GWjww8/vMCymjVr8vvvvwPw1ltvAZG5VqPb0OaiZ555BnDPIN+A3EeAfURq7dq1bNq0CYg0wPefzVzTtm1bAI4++mggMrPCtGnTihwBPZmsIGVMgvwX0PDhw8OqEh8u32OPPcJqPO/DDz8MG7iWtgJUPCpWrBiORzRy5EjAestmA98LurQaPXo0vXv3BiKFpH//+9/MmzcPgG3btmUsbam0Zs0abrjhBiAyq4fvode0adNwonFfoMpFLVu2DJtHVKxYEXBNCADuvffetKXDqvaMMcYYYxIkRU0smvSTiaTvZEmmqqWzJaYpIN771I9jNm3atHAuJz+3FcCcOXMAN7kvuFHMU/32F+99mqrP4hFHHMG0adN8WoDIG+8VV1zBU089BVCiWQTss+iUheep5TG7ZTqPZ511FmPGjAFg5cqVQKSjme/0UlLx5NEiUsYYY4wxCbKIVJzsLbjsKAv3aVnIY2lXFq6h5TG7ZSqPfjiLMWPGsNdeewEwduxYAM4//3wgeW2/4smjFaTiZA/vsqMs3KdlIY+lXVm4hpbH7GZ5dKxqzxhjjDEmQWktSKmq7OgH2A58AcwDJgBV4tmvkGMdDUwqZF1DYCowH/gG2GcHxzJlRJbdp/48XwCvxXGsXMzjncE55gGnJCuPpV2WXcMRwNe45+kDwE6l6T4F2gCzgjx+SRLv0yzKYyNgTnCer4Hz05XHdOQv33niep7Gm8dsjEhtUtXWqtoS+BPIMyOxOMlI91hghKruB7QDVibhmKbsSNd96s/TWlV7J+F4iZw7ZXkUkZ5AW6A1cBhwpYjsVpJjmjzScQ0PBzoCrYCWwKFA55Ics5jS8Vn8AzhTVQ8AugP3iki1Eh6zONKRx+XA4arqP4vXiEjdEh4zXjn9PM3GglS0GUATEWkkIt+IyChcibm+iBwnIh+KyBwReV5EqgKISHcRmS8i7wN/j3VQEdkfKK+qUwFUdYOq/pGmPJnSJyX3aZZJVR73B6ap6lZV3Yh7U+xeyLamZFJ1DRWoBOwMVAQqAL+kPjsxpSSPqrpAVb8Lfl+Ge/GulZ4sFZCqPP6pqluC/1Ykc+WDnHueZm1BSkTKAz1wYVSA5sBYVW0DbASGAseqaltgNnCZiFQCRgPHA52APaOOd4iIPBb8txnwm4i8KCKficgIESmXloyZUiXF9ylAJRGZLSKzROTE1OeooBTn8Qugh4hUFpGauLB8/TRkq0xJ5TVU1Q+Bd3ERjeXAFFX9Ji0Zi5KGz6Jf3g5XaPwhhdmJKdV5FJH6IjIXWALcERQa0yZnn6eqmlU/wDbg8+DnAdwN2whYGLVNL2B11HZfA2Nw1QPTo7brDUyMcY6+wO/Avrhpcl4A/pnpvNtP7vyk4z4N1tUN/t0XWAQ0LoV5vD7YdyrwFHBxpq9vaflJ0/O0CTAJqBr8fAgcWZryGLV+L+BboH1pu475zlcX+BioU5ryR4qep9k4194mdXW0IREBVxoNFwFTVbV/vu1a48LMO7IU+ExVfwz2exloj7soxsQjHfcpGrwRquqPIvIertFrut6E05XHW4Bbgv2eBr4rQZpNXum4hicBs1R1Q7DfZNzzdHoJ0l0cablPxbXdmwQMVdVZJUpx8aUlj56qLhORebgIz4SEUlw8Of08zdqqvR2YBXQUkSYAQbVAM1yPkX1EpHGwXf9C9v8EqC4ivo77GFzp1phkKtF9KiLVRaRi8HtNXIPebLtPS5rHciJSI/i9Fa7B8pupT7aJUtLn6WKgs4iUF5EKuIbmaa/a24GS3qc7Ay/hqpmeT0eCE1DSPO4tIrsEv1fHPW++TX2y45a1z9OcLEip6ipgIDA+qM+dBbRQ1c3AYGBS0OjsJ79Pvjr9bcAVwNsi8iWupDs6vbkwpV1J71NgP2C2iHyBa4Nyu6pmVUEqCXmsAMwQka+BR4HTVXVrOvNQ1iXhGk7AvdV/SdC1XFVfS2MWdigJeewHHAkMFJHPg588EZRMS9Lz5qPgeTMNGKmqX5Ilsvl5KkF9oTHGGGOMKaa0tpGSUj5MvCkd7D41xpjkKe3P1Jys2jPGGGOMyQZWkDLGGGOMSZAVpIwxxhhjEmQFKWOMMcaYBFlByhhjjDEmQdk4snlCmjRpAsBFF10EwCGHHALAtm3bGD3aDRE1duzYzCTOlGn+3rzuuuvo0qULAMcccwwAP/yQ9um6jDHGJFFax5FKVRfItm3bMnnyZABq165dYP22bdsAOOGEEwCYNGlSsc9h3crLjmTfpx999BEA7dq18/M9hcs6dOiQzFPZfWqMyTrJeKbWrl2b6667DoB//etfANx+++0MGzas0H2qVKkCQLly5QBYt25dsc9rwx8YY4wxxqRQTlftVa1aFYAJEyZQrVo1AAYNGgTAG2+8AcApp5zCiBEjAGjatGkGUmnKOh99OvTQQwssM8YYU7g6deoAMHv2bOrVqwfA559/DsCdd95Z5L6+HOBrAo488siUpNEiUsYYY4wxCcrJiJRvvOvbRdWrV4/u3bsD8O677+bZ9q677uLmm29ObwKNibJmzZrw9xUrVgDw6KOPZio5xhiT9URc06RZs2YB7nt+wYIFAHTt2hWA9evXF7p/rVq1OOyww/JsV6NGjTzP42TJyYLUAw88AEQKVOecc06BApTXvXt3KlasCMDmzZvTk0BjgIYNGwIwcODAcNnzzz8PwNdfJ2XScWOMKZXuueceIPIcXbduHf369QOIqzBUv379sJF59erVAdf4PBUFKavaM8YYY4xJUM5FpPbdd1+OO+44AF577TUAxo0bx047uTKhjz4dddRRgItW+YZmL7zwQppTa8qy1q1bA5E3KoANGzYAcMYZZwCRquilS5emOXXGGJOdqlevTv/+/fMse+yxx5g7d27cx6hRo0ZYPeh16NCBxYsXJyWN0SwiZYwxxhiToJwZkNNHmmbMmBF2Iz/ooIPC9UOHDgXgH//4R4F9V61aBUSGP/j999+LfX4b6LDsSMbgcfXr1+fll18GoE2bNoVuN2fOHACGDRsWdp4oCbtPjTHZprjP1Iceeojzzz8fgK1btwLQqFEjli1bFvcxunbtypQpU/Is+/zzz2nXrl2e4+5IPM/UnKnaO/zwwwE3Fs/jjz8ORBqPf/rpp+GYUrHUqlULgGuvvRaAa665JoUpNQbOPvvsIgtQXtu2bQF46aWX6NOnD5DYyPvGGJPrfBOdo48+Olw2c+ZMgGIVogpTr169MNgyfvz4Eh/Ps6o9Y4wxxpgEZX1EarfddgNgzJgx4TJfjffJJ58AboRz31j3qquuAiIjnL///vvhBLEHHnhgehJtyiw/8q6//6KtXbs2DFcvXLgQcOOcAXTq1Ckcnd8YY8oiP4deixYtwmXTp09P6Fh+RPRotWrVCssDFpEyxhhjjMkCWR+R2mWXXQDYZ599AHj66adp1aoVAHvttRcAU6ZM4bzzzgNg0aJFALzzzjuA625ev359APbff3/ANVzfsmVLejJgyhR/j+2xxx6sXr0aiLxlff3118ybNy/P9n4eyE6dOnHOOecA8NRTT6UrucYYkzWiO5B5o0aNSuhYv/zyS8zlyWhrlV/WF6Q6d+4MRCYdHDt2LC1btsyzzdVXXx0WoLxYf8QGDRoArirQClImFaZOnQrAMcccw8aNGwH46quvCt1+yZIl4e/+/jTGGOP4QEhhBaPCxKraW7RoEQ8++GBS0hXNqvaMMcYYYxKU9RGpXXfdFSCMIE2ZMoXTTjstzza//fZbXMd6++23gfjm6TGmJD766KO4tqtbt26KU2KMMbnr4YcfBmDIkAFXL7sAABELSURBVCHhBMaxlC/vijN+eKMrrrgi5rH8uJLJZBEpY4wxxpgEZXVEqm7dutx9991AZE6y6N/9fGX//Oc/ufHGGwHYtm1bgeP4effuv//+VCbXmGLznSTAzRlpjDFl1ejRowE49thjwzajftDimTNn8scffxTYZ+XKlQDUrl0bgCpVqhR6fN9uNdksImWMMcYYk6CsjkhVrlw5HJDTD74JhHOSvfXWW4Cbp6xJkyYAnHrqqXmO0adPHxo2bAi4wTmNyQb+fj3kkEPCZd9++22mkmOMMRnn25Z26dKFAQMGAJGp3cqVKxcz2uSHRsqkrC5IRfPz6gGsWLECgG7dugHQrl07vv/++zzb+xGmR44cyXPPPQe4OfmMyQZ+zCg/Ftq8efPCOaWMMaYs++GHH7jhhhsAeP311wE4+eSTOeCAAwps58fr843IP/jgA8CN3+cLY57vcJZsVrVnjDHGGJOgnIlIHXbYYQWWbd++HSBPl0jfBXLw4MEA7L777tx2221AZFBPYzLl4IMPBiJV0Fu3bgVc+Dr/oLLGGFPW+e/3ooY+iMVHqqJ17NiR+fPnJyVd0SwiZYwxxhiToKyOSK1du5alS5cC0L17d8C1ffr5558LbFuhQgUgMsSB71Y+fPhw5s6dm47kGlOk5s2b8/LLLwORNnz+LWvixIkZS5cxxpQ2X3zxRfi7bz/1wgsvpORcWV+QmjRpEgDnnnsu4OYtGz9+PAALFy4EoHXr1rRu3RqITBo7e/ZsIHV/OGOi7bzzzgD07NkTgDfeeINNmzYBhPfmsGHDwgKU7xxx+umnpzupxhhT6kU35fEd1OKdBaW4rGrPGGOMMSZBks4G2CJS7JP5caR89ZwfEyo/nw9fVdKrVy/ARbWSQVUlKQcyWS+R+/Sss84CYMyYMYAbE+q1114DIiPw16lTJ2xc7u/PN998s+QJjmL3qTEm2yTyTC2p++67jwsvvBCAdevWAdC0adNiz7UXzzPVIlLGGGOMMQnK+oiU16hRIwBuvfXWcFmfPn0AN9L5E088AcBLL71UghQWzt70y46S3Ke+3V6syOno0aPDoThSNdSB3afGmGyT6YiUiHssHnrooWH76XjF80zNmYJUptkXVNlh96kxxiRPJp6p7du358UXXwRg+vTpgOvc45tXxMuq9owxxhhjUsgiUnGyN/2yw+5TY4xJntL+TLWIlDHGGGNMgqwgZYwxxhiTICtIGWOMMcYkyApSxhhjjDEJSmtjc2OMMcaY0iTrIlIisk1EPheRr0TkeRGpXIJjHSUiEwtZd6eIzBORb0TkfvEjdhkThzTep/48n4vIq4mn2Bhjsle6nqnB+t1E5GcReTDRc0TLuoIUsElVW6tqS+BP4LzoleKUKN0icjjQEWgFtAQOBTqX5JimzEn5fZrvPK1VtXcSjmeMMdkoXc9UgJuAaUk6VlYWpKLNAJqISKMgcjQKmAPUF5HjRORDEZkTlF6rAohIdxGZLyLvA38v5LgKVAJ2BioCFYBfUp8dU0ql6j41xpiyKGXPVBE5GKgDJG3G+KwtSIlIeaAH8GWwqDkwVlXbABuBocCxqtoWmA1cJiKVgNHA8UAnYM+o4x0iIo8BqOqHwLvA8uBniqp+k5aMmVIllfdpoJKIzBaRWSJyYupzZIwxmZPKZ2oQ0boLuDKZaS6fzIMlyS4i8nnw+wxgDFAX+ElVZwXL2wP7AzODpk07Ax8CLYCFqvodgIiMAwYDqOpsYFCwvAmwH7B3cLypInKkqk5Pcd5M6ZHy+zTQQFWXici+wDsi8qWq/pDarBljTNql45l6AfC6qi5JZrPobCxIbVLV1tELggxvjF4ETFXV/vm2a42rttuRk4BZqroh2G8y7gJZQcrEKx33Kaq6LPj3RxF5D2gDWEHKGFPapOOZ2gHoJCIXAFWBnUVkg6peU5KEZ23V3g7MAjoGkSVEpLKINAPmA/uISONgu/6F7L8Y6Cwi5UWkAq6huVXtmWQr0X0qItVFpGLwe01cB4mvU59sY4zJSiV6pqrqaaraQFUbAVfgqgxLVIiCHC1IqeoqYCAwXkTm4v64LVR1My6cNylocPaT3ydf25MJuLf6L4Ev4P/bu9/QrOs1juPvS5f/nTXEdJlpjbT0hElknLC/IzKomIgVQRlEegqfCEGET7InJziPeiBEkyKKESmZUXYadAiTDCy06amTA/Nok5AmmDrXSb/nwX3fc7q5zd/uf6v3C27cfvfvzyWIfLi+F98fe1JKH5bxr6A/gSL8O70B2BURe8jN9P09pWSQkvSnVIT/U0vCDTklSZIyqsYZKUmS9AcRESO2Y5NSGnQqfUQu7UmSJFUDg5QkSVJGBilJkqSMDFKSJEkZGaQkSZIyMkhJkiRlZJCSJEnKyCAlSZKUkUFKkiQpI4OUJElSRgYpSZI0YjQ1NdHU1ERbWxtnz54977Nt2za2bdvGjBkzylaPLy2WJEklU6x37S1fvhyAN954A4CJEydysQzT1NTE1q1bh/1M37UnSZJUQnakJElSyRSjIzV16lQOHToEwJgxYwD4+OOPWbduHQDd3d0A7Nu3D4Bjx44xderU4T7WjpQkSVIp1VS6AEmSpIEsXbq0pxNVsHHjRvbs2QPAtGnTzvuupaWlbLUZpCRJUlVqaGgA4OWXX+45VljO27JlS0VqupBLe5IkSRnZkZIkSVVp0aJFAMyaNYv9+/cD8Nprr/U5r66uDoCIQWfDi86OlCRJUkZ2pCRJUlV6+OGHATh9+jQvvvgiAJ2dnX3Oe/TRRwF6Nuj84IMPylShQUqSJFWRp556ildeeQWA2tpaALZu3crmzZsHvKa3jo6O0hV4AZf2JEmSMrIjJUmSqsb27dvZu3cvAO3t7QCsWbNmwGvGjRsHQGtrKwCHDx8uYYXnsyMlSZKUkR0pSZJUNdrb27nnnnsyXdvc3AzA8ePHi1nSgOxISZIkZWRHSpIkjUgrVqwAzr1r74477gBg06ZNZavBICVJkkakwr5RZ8+eBWDBggVlr8GlPUmSpIwMUpIkSRkZpCRJkjJyRkqSJP0h/PTTT2V/pkFKkiT9IZTzZcUFLu1JkiRlZEdKkiSNSM899xwAJ0+eBGD37t1lr8GOlCRJUkZ2pCRJ0ohU2JDzzJkzAHR3d5e9BoOUJEkacerq6qivrwegtrYWgMWLFwNw6NChstXh0p4kSVJGBilJkjTidHZ20tHRQUdHR0XrMEhJkiRl5IyUJEkakdra2gC49dZbAWhvby97DVGYeJckSSq2iBixQSOlFIOd49KeJElSRgYpSZJUMimlGOgDnAX2APuATcDEwa4Z4F53Ax/1c/ye/DMKn26gaQj3G5RBSpIkVVJXSmlhSmkB8BuwuveXkTOsvJJS+lf+GQvJhapTwKfDuWeBQUqSJFWL7UBDRMyOiO8iYgPwDXB1RNwXEV9GxDcR8V5ETAKIiPsj4vuI+AJYNoRnLAe2pZROFaNgg5QkSaq4iKgBlgJt+UNzgbdSSjcDJ4F1QGNKaRGwC1gbEeOA14EHgSXA9F73uyUimvt51KNAS7HqdvsDSZJUSeMjYnf+5+3ARqAeOJhS2pk/fhtwI7AjIgDGAF8C84ADKaX9ABHxNvAMQEppF/B07wdFxAzgL8A/i1W8QUqSJFVSV352qUc+LJ3sfQhoTSk9dsF5C4FL2V5hBfB+Sul/GWvtw6U9SZJU7XYCt0dEA0BETIiI64HvgTkRcV3+vMcudoNe3xdtWQ8MUpIkqcqllI4CK4GWiPiWXLCal1I6TW4p76P8sPnBwjUXzkhFxGzgauDzYtbmzuaSJEkZOSMlSZJKxlfESJIkqV8GKUmSpIwMUpIkSRkZpCRJkjIySEmSJGVkkJIkScrIICVJkpSRQUqSJCkjg5QkSVJGBilJkqSMDFKSJEkZ+a49SZI04owaNYpRo3L9oGXLlgHw7rvv9nz/6quvAvDSSy8BcPz4cX7//fei1xEpjdh3CUqSpCpXjJcWjx49mgkTJgDw+OOPA3DXXXfR2toKwPr16wGYPn36Re/x9ddfs2TJEgC6u7uH9FxfWixJklRCdqQkSVLJFKMj9cgjj7Bq1SoArrjiCgAaGhrYsGEDAB0dHQCsWbMGgDlz5vR7nylTpgBw4sSJIT3XjpQkSVIJOWwuSZKq2meffcZDDz0EwBNPPAHk5qF27dp13nk///wzAO+8807ZajNISZKkqnLZZZcBcO+99wLwySef9AyZFxw+fLjPdYWlu3JyaU+SJCkjO1KSJKmqXHXVVQBEDDrrDcA111wDwNq1a0tW08XYkZIkScrIjpQkSaoqP/7443l/9mfSpEk9u5avXLkSgMsvv7zPeZ2dnQBs2bKF06dPF7VOcB8pSZJUQoPtI1VTk+vpFPaHOnr06ID3mzFjBgCNjY28+eabFz1v7969ADz77LMA7NixY2gF9+I+UpIkSSXk0p4kSaqY2bNnA1BXVwfkXi5cW1sLnOtSzZ8/n4kTJwJw5513AtDU1NTnXr/88guQGzrfvHkzAF1dXaUrHjtSkiRJmdmRkiRJFVFTU8MDDzwAwOrVqwF4//33e4bGly1bBsC0adN6rjlw4ABwrlsFfeegCvNR5WBHSpIkKSM7UpIkqSJSStTX1wMwd+5cAF544YU+5/3www+sX78eOLfVwbFjx2hubgbOvVvvxIkTpS65D4OUJEmqiJqaGq699tqLfv/pp58C8Pzzz/cs17W0tAC59+qdOXMGqEyAKnBpT5IkKSM35JQkSSUz2IacM2fOBM4Nlk+ZMoWDBw8C8OGHHwK5ZbybbroJgPHjxwPw1VdflabgXtyQU5IkqYTsSEmSpJIZrCM1FPPnz+/Z9mDUqFwPqBxzUUPpSDlsLkmSqtqRI0c4depUpcvol0t7kiRJGbm0J0mSSuZSl/YmT57MlVdeCUB7e3tJahoqh80lSZJKyI6UJEkqmUvtSI0ePZoxY8YA0NXVVZKahsphc0mSVNUK+0KtWrUKgF9//ZWNGzdWsqRL4tKeJElSRnakJElSxTQ2NgIwadIkAMaOHVvJci6ZHSlJkqSMHDaXJEklM9Cw+eTJk+ns7ARym24CzJo1qzyFDcFQhs0NUpIkqWSK8YqYSnEfKUmSpBIySEmSpJJJKcVgH+AssAfYB2wCJg7luovc627go36Ozwa+yT/n38DfhnCvQRmkJElSpXWllBamlBYAvwGre38ZOcPNLEeAv6aUFgKLgRcion6Y9zRISZKkqrIdaIiI2RHxXURsINdJujoi7ouILyPim4h4LyImAUTE/RHxfUR8ASzr76Yppd9SSt35X8dSpAxkkJIkSVUhImqApUBb/tBc4K2U0s3ASWAd0JhSWgTsAtZGxDjgdeBBYAkwvdf9bomI5l6/Xx0R3wKHgFdSSh3DrdkgJUmSKm18ROwmF47+CxTeEXMwpbQz//NtwI3Ajvy5TwLXAPOAAyml/Sm3FcHbhZumlHallJ7u9fuhlNJNQAPwZERcOdzC3dlckiRVWld+dqlHRECuC9VzCGhNKT12wXkLgUvaYiGl1BER+8h1sDZlqjjPjpQkSRoJdgK3R0QDQERMiIjrge+BORFxXf68x/q7OCJmRsT4/M9XALcD/xluUQYpSZJU9VJKR4GVQEt+zmknMC+ldBp4BvgoP2x+sHDNBTNSNwBfRcQe4HPgHymlNobJnc0lSZIysiMlSZKUkUFKkiQpI4OUJElSRgYpSZKkjAxSkiRJGRmkJEmSMjJISZIkZWSQkiRJysggJUmSlJFBSpIkKSODlCRJUkYGKUmSpIwMUpIkSRkZpCRJkjIySEmSJGX0f4EtFDsnNRJRAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x504 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = keras.models.load_model('./mnist-inception-chkpt-12-0.9952_2018-11-15_20-02-30.hdf5')\n",
    "evaluate_model(model, X_val, y_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.9952380952380953\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x216 with 64 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x504 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model = keras.models.load_model('./mnist-inception-chkpt-48-0.9952_2018-10-26_22-46-37.hdf5')\n",
    "evaluate_model(model, X_val, y_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Save Results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'val_acc': 0.9962,\n",
       " 'val_loss': 0.0226,\n",
       " 'lr': 0.0002455394,\n",
       " 'epoch': 89,\n",
       " 'optimizer': 'Adamax',\n",
       " 'optimizer args': '(lr=0.01, beta_1=0.49, beta_2=0.999)',\n",
       " 'annealer': 'CosineAnneal',\n",
       " 'annealer args': '(max_lr=0.01, min_lr=0.0, T=10, T_mul=1)',\n",
       " 'dropout type': 'spatial',\n",
       " 'dropout rates': '(0,0,.3,.5)',\n",
       " 'notes': 'stem_width=64, A_width=96, B_width=144, only elastic_transform(alpha_range=[8, 10], sigma=3), width_shift_range=2, height_shift_range=2',\n",
       " 'batch size': 64,\n",
       " 'epochs': 120,\n",
       " 'log': '2018-12-02_13-58-36.csv',\n",
       " 'keras': '2.2.4',\n",
       " 'tf': '1.12.0'}"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_result = {}\n",
    "\n",
    "i = log['val_acc'].idxmax()\n",
    "new_result['val_acc'] = round(log['val_acc'][i], 4)\n",
    "new_result['val_loss'] = round(log['val_loss'][i], 4)\n",
    "if 'lr' in log.keys():\n",
    "    new_result['lr'] = log['lr'][i]\n",
    "new_result['epoch'] = i + 1\n",
    "\n",
    "new_result['optimizer'] = model.optimizer.__class__.__name__\n",
    "new_result['optimizer args'] = optimizer_args\n",
    "\n",
    "if annealer is None:\n",
    "    new_result['annealer'] = ''\n",
    "    new_result['lr'] = K.get_value(model.optimizer.lr)\n",
    "else:\n",
    "    new_result['annealer'] = annealer.__class__.__name__\n",
    "    if new_result['annealer'] == 'LearningRateScheduler':\n",
    "        new_result['annealer'] = annealer.schedule.__doc__    \n",
    "new_result['annealer args'] = annealer_args\n",
    "\n",
    "new_result['dropout type'] = dropout_type\n",
    "new_result['dropout rates'] = dropout_rates\n",
    "\n",
    "new_result['notes'] = notes\n",
    "\n",
    "new_result['batch size'] = batch_size\n",
    "new_result['epochs'] = len(log['acc'])\n",
    "\n",
    "new_result['log'] = log_filename\n",
    "\n",
    "new_result['keras'] = keras_version\n",
    "new_result['tf'] = tf_version\n",
    "\n",
    "new_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "scrolled": true
   },
   "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>val_acc</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>lr</th>\n",
       "      <th>epoch</th>\n",
       "      <th>optimizer</th>\n",
       "      <th>optimizer args</th>\n",
       "      <th>annealer</th>\n",
       "      <th>annealer args</th>\n",
       "      <th>dropout type</th>\n",
       "      <th>dropout rates</th>\n",
       "      <th>notes</th>\n",
       "      <th>batch size</th>\n",
       "      <th>epochs</th>\n",
       "      <th>log</th>\n",
       "      <th>keras</th>\n",
       "      <th>tf</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>153</th>\n",
       "      <td>0.9962</td>\n",
       "      <td>0.0226</td>\n",
       "      <td>0.000246</td>\n",
       "      <td>89</td>\n",
       "      <td>Adamax</td>\n",
       "      <td>(lr=0.01, beta_1=0.49, beta_2=0.999)</td>\n",
       "      <td>CosineAnneal</td>\n",
       "      <td>(max_lr=0.01, min_lr=0.0, T=10, T_mul=1)</td>\n",
       "      <td>spatial</td>\n",
       "      <td>(0,0,.3,.5)</td>\n",
       "      <td>stem_width=64, A_width=96, B_width=144, only elastic_transform(alpha_range=[8, 10], sigma=3), width_shift_range=2, height_shift_range=2</td>\n",
       "      <td>64</td>\n",
       "      <td>120</td>\n",
       "      <td>2018-12-02_13-58-36.csv</td>\n",
       "      <td>2.2.4</td>\n",
       "      <td>1.12.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     val_acc  val_loss        lr  epoch optimizer  \\\n",
       "153  0.9962   0.0226    0.000246  89     Adamax     \n",
       "\n",
       "                           optimizer args      annealer  \\\n",
       "153  (lr=0.01, beta_1=0.49, beta_2=0.999)  CosineAnneal   \n",
       "\n",
       "                                annealer args dropout type dropout rates  \\\n",
       "153  (max_lr=0.01, min_lr=0.0, T=10, T_mul=1)  spatial      (0,0,.3,.5)    \n",
       "\n",
       "                                                                                                                                       notes  \\\n",
       "153  stem_width=64, A_width=96, B_width=144, only elastic_transform(alpha_range=[8, 10], sigma=3), width_shift_range=2, height_shift_range=2   \n",
       "\n",
       "     batch size  epochs                      log  keras      tf  \n",
       "153  64          120     2018-12-02_13-58-36.csv  2.2.4  1.12.0  "
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = pd.read_csv('results.csv')\n",
    "new_results = results.append(new_result, ignore_index=True)\n",
    "new_results.iloc[[-1]]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Overwrite results.csv?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Type y to save updated results to results.csv: y\n",
      "results.csv saved\n"
     ]
    }
   ],
   "source": [
    "results_filename = 'results.csv'\n",
    "prompt = \"Type y to save updated results to \" + results_filename + \": \"\n",
    "overwrite = input(prompt)\n",
    "if overwrite == 'y':\n",
    "    new_results.to_csv(results_filename, index=False, float_format='%.4f')\n",
    "    print(results_filename, \"saved\")\n",
    "else:\n",
    "    print(results_filename, \"not saved\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predict on Test Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load test set data\n",
    "test_data = np.loadtxt('test.csv', dtype=int, delimiter=',', skiprows=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Raw data shape: (28000, 784)\n",
      "Pixels reshaped shape: (28000, 28, 28, 1)\n"
     ]
    }
   ],
   "source": [
    "# inspect data\n",
    "print(\"Raw data shape:\", test_data.shape)\n",
    "\n",
    "# convert pixel values to 2d arrays\n",
    "test_images = np.reshape(test_data, (28000, 28, 28))\n",
    "test_images = test_images / 127.5 - 1\n",
    "test_images = np.expand_dims(test_images, -1)\n",
    "print(\"Pixels reshaped shape:\", test_images.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 70 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# display random sample of images\n",
    "plt.rcParams['figure.figsize'] = (10.0, 8.0)\n",
    "for i in range(70):\n",
    "    plt.subplot(7, 10, i + 1)\n",
    "    plt.imshow(test_images[i,:,:,0])\n",
    "    plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_scores = model.predict(test_images)\n",
    "test_predictions = np.argmax(test_scores, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(28000, 2)\n"
     ]
    }
   ],
   "source": [
    "header = 'ImageId,Label'\n",
    "submission = np.stack((range(1, 28001), test_predictions), axis=1)\n",
    "print(submission.shape)\n",
    "np.savetxt('submission.csv', submission, fmt='%i', delimiter=',', header=header, comments='')"
   ]
  }
 ],
 "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"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}