Traffic_Sign_Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Self-Driving Car Engineer Nanodegree\n",
    "\n",
    "## Deep Learning\n",
    "\n",
    "## Project: Build a Traffic Sign Recognition Classifier\n",
    "\n",
    "In this notebook, a template is provided for you to implement your functionality in stages, which is required to successfully complete this project. If additional code is required that cannot be included in the notebook, be sure that the Python code is successfully imported and included in your submission if necessary. \n",
    "\n",
    "> **Note**: Once you have completed all of the code implementations, you need to finalize your work by exporting the iPython Notebook as an HTML document. Before exporting the notebook to html, all of the code cells need to have been run so that reviewers can see the final implementation and output. You can then export the notebook by using the menu above and navigating to  \\n\",\n",
    "    \"**File -> Download as -> HTML (.html)**. Include the finished document along with this notebook as your submission. \n",
    "\n",
    "In addition to implementing code, there is a writeup to complete. The writeup should be completed in a separate file, which can be either a markdown file or a pdf document. There is a [write up template](https://github.com/udacity/CarND-Traffic-Sign-Classifier-Project/blob/master/writeup_template.md) that can be used to guide the writing process. Completing the code template and writeup template will cover all of the [rubric points](https://review.udacity.com/#!/rubrics/481/view) for this project.\n",
    "\n",
    "The [rubric](https://review.udacity.com/#!/rubrics/481/view) contains \"Stand Out Suggestions\" for enhancing the project beyond the minimum requirements. The stand out suggestions are optional. If you decide to pursue the \"stand out suggestions\", you can include the code in this Ipython notebook and also discuss the results in the writeup file.\n",
    "\n",
    "\n",
    ">**Note:** Code and Markdown cells can be executed using the **Shift + Enter** keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step 0: Load The Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import cv2\n",
    "import pickle\n",
    "import random\n",
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "from tensorflow.contrib.layers import flatten\n",
    "import tensorflow as tf\n",
    "from sklearn.utils import shuffle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load training, validation and test data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load pickled data\n",
    "# Fill this in based on where you saved the training and testing data\n",
    "\n",
    "training_file = '/home/jzhang/udacity/self_drive_nd/lesson8_cnn/p2/data/train.p'\n",
    "validation_file='/home/jzhang/udacity/self_drive_nd/lesson8_cnn/p2/data/valid.p'\n",
    "testing_file = '/home/jzhang/udacity/self_drive_nd/lesson8_cnn/p2/data/test.p'\n",
    "\n",
    "with open(training_file, mode='rb') as f:\n",
    "    train = pickle.load(f)\n",
    "with open(validation_file, mode='rb') as f:\n",
    "    valid = pickle.load(f)\n",
    "with open(testing_file, mode='rb') as f:\n",
    "    test = pickle.load(f)\n",
    "    \n",
    "X_train, y_train = train['features'], train['labels']\n",
    "X_valid, y_valid = valid['features'], valid['labels']\n",
    "X_test, y_test = test['features'], test['labels']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## Step 1: Dataset Summary & Exploration\n",
    "\n",
    "The pickled data is a dictionary with 4 key/value pairs:\n",
    "\n",
    "- `'features'` is a 4D array containing raw pixel data of the traffic sign images, (num examples, width, height, channels).\n",
    "- `'labels'` is a 1D array containing the label/class id of the traffic sign. The file `signnames.csv` contains id -> name mappings for each id.\n",
    "- `'sizes'` is a list containing tuples, (width, height) representing the original width and height the image.\n",
    "- `'coords'` is a list containing tuples, (x1, y1, x2, y2) representing coordinates of a bounding box around the sign in the image. **THESE COORDINATES ASSUME THE ORIGINAL IMAGE. THE PICKLED DATA CONTAINS RESIZED VERSIONS (32 by 32) OF THESE IMAGES**\n",
    "\n",
    "Complete the basic data summary below. Use python, numpy and/or pandas methods to calculate the data summary rather than hard coding the results. For example, the [pandas shape method](http://pandas.pydata.org/pandas-docs/stable/generated/pandas.DataFrame.shape.html) might be useful for calculating some of the summary results. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Provide a Basic Summary of the Data Set Using Python, Numpy and/or Pandas\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples = 34799\n",
      "Number of validation examples = 4410\n",
      "Number of testing examples = 12630\n",
      "Image data shape = (32, 32, 3)\n",
      "Number of classes = 43\n"
     ]
    }
   ],
   "source": [
    "### Replace each question mark with the appropriate value. \n",
    "### Use python, pandas or numpy methods rather than hard coding the results\n",
    "\n",
    "# Number of training examples\n",
    "n_train = X_train.shape[0]\n",
    "\n",
    "# Number of validation examples\n",
    "n_validation = X_valid.shape[0]\n",
    "\n",
    "# Number of testing examples.\n",
    "n_test = X_test.shape[0]\n",
    "\n",
    "# What's the shape of an traffic sign image?\n",
    "image_shape = X_train.shape[1:] \n",
    "\n",
    "# How many unique classes/labels there are in the dataset.\n",
    "n_classes = len(set(y_train))\n",
    "\n",
    "print(\"Number of training examples =\", n_train)\n",
    "print(\"Number of validation examples =\", n_validation)\n",
    "print(\"Number of testing examples =\", n_test)\n",
    "print(\"Image data shape =\", image_shape)\n",
    "print(\"Number of classes =\", n_classes)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Include an exploratory visualization of the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Visualize the German Traffic Signs Dataset using the pickled file(s). This is open ended, suggestions include: plotting traffic sign images, plotting the count of each sign, etc. \n",
    "\n",
    "The [Matplotlib](http://matplotlib.org/) [examples](http://matplotlib.org/examples/index.html) and [gallery](http://matplotlib.org/gallery.html) pages are a great resource for doing visualizations in Python.\n",
    "\n",
    "**NOTE:** It's recommended you start with something simple first. If you wish to do more, come back to it after you've completed the rest of the sections. It can be interesting to look at the distribution of classes in the training, validation and test set. Is the distribution the same? Are there more examples of some classes than others?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fb892d57908>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAFsAAABZCAYAAABR/liSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nA0PbN2rdGfnXAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb8930ce470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAFsAAABZCAYAAABR/liSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892e36518>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAFsAAABZCAYAAABR/liSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892dce2e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Data exploration visualization code goes here.\n",
    "### Feel free to use as many code cells as needed.\n",
    "\n",
    "# Visualizations will be shown in the notebook.\n",
    "%matplotlib inline\n",
    "index = random.randint(0, len(X_train))\n",
    "image = X_train[index].squeeze()\n",
    "\n",
    "plt.figure(figsize=(1,1))\n",
    "plt.imshow(image)\n",
    "\n",
    "index = random.randint(0, len(X_train))\n",
    "image = X_train[index].squeeze()\n",
    "\n",
    "plt.figure(figsize=(1,1))\n",
    "plt.imshow(image)\n",
    "\n",
    "index = random.randint(0, len(X_train))\n",
    "image = X_train[index].squeeze()\n",
    "\n",
    "plt.figure(figsize=(1,1))\n",
    "plt.imshow(image)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Collect histogram for training, validataion and test data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAs8AAAE/CAYAAAC5CC4zAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/x8ug0/wCZtO9AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb89f9ba080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAskAAAE/CAYAAAC0Fl50AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892d78da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAskAAAE/CAYAAAC0Fl50AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892b140f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# historgram collection\n",
    "labels_train, counts_train = np.unique(y_train, return_counts=True)\n",
    "labels_valid, counts_valid = np.unique(y_valid, return_counts=True)\n",
    "labels_test, counts_test = np.unique(y_test, return_counts=True)\n",
    "# Plot histogram \n",
    "plt.rcParams[\"figure.figsize\"] = [12, 5]\n",
    "axes = plt.gca()\n",
    "axes.set_xlim([-0.5,42.5])\n",
    "\n",
    "plt.bar(labels_train, counts_train, tick_label=labels_train, width=0.6, align='center')\n",
    "plt.title('Training data histogram')\n",
    "plt.show()\n",
    "\n",
    "plt.bar(labels_valid, counts_valid, tick_label=labels_valid, width=0.6, align='center')\n",
    "plt.title('Validation data histogram')\n",
    "plt.show()\n",
    "\n",
    "plt.bar(labels_test, counts_test, tick_label=labels_test, width=0.6, align='center')\n",
    "plt.title('Test data histogram')\n",
    "plt.show()\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "\n",
    "## Step 2: Design and Test a Model Architecture\n",
    "\n",
    "Design and implement a deep learning model that learns to recognize traffic signs. Train and test your model on the [German Traffic Sign Dataset](http://benchmark.ini.rub.de/?section=gtsrb&subsection=dataset).\n",
    "\n",
    "The LeNet-5 implementation shown in the [classroom](https://classroom.udacity.com/nanodegrees/nd013/parts/fbf77062-5703-404e-b60c-95b78b2f3f9e/modules/6df7ae49-c61c-4bb2-a23e-6527e69209ec/lessons/601ae704-1035-4287-8b11-e2c2716217ad/concepts/d4aca031-508f-4e0b-b493-e7b706120f81) at the end of the CNN lesson is a solid starting point. You'll have to change the number of classes and possibly the preprocessing, but aside from that it's plug and play! \n",
    "\n",
    "With the LeNet-5 solution from the lecture, you should expect a validation set accuracy of about 0.89. To meet specifications, the validation set accuracy will need to be at least 0.93. It is possible to get an even higher accuracy, but 0.93 is the minimum for a successful project submission. \n",
    "\n",
    "There are various aspects to consider when thinking about this problem:\n",
    "\n",
    "- Neural network architecture (is the network over or underfitting?)\n",
    "- Play around preprocessing techniques (normalization, rgb to grayscale, etc)\n",
    "- Number of examples per label (some have more than others).\n",
    "- Generate fake data.\n",
    "\n",
    "Here is an example of a [published baseline model on this problem](http://yann.lecun.com/exdb/publis/pdf/sermanet-ijcnn-11.pdf). It's not required to be familiar with the approach used in the paper but, it's good practice to try to read papers like these."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocess Data\n",
    "\n",
    "Minimally, the image data should be normalized so that the data has mean zero and equal variance. For image data, `(pixel - 128)/ 128` is a quick way to approximately normalize the data and can be used in this project. \n",
    "\n",
    "Other pre-processing steps are optional. You can try different techniques to see if it improves performance. \n",
    "\n",
    "Use the code cell (or multiple code cells, if necessary) to implement the first step of your project."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "### Preprocess the data here. It is required to normalize the data. Other preprocessing steps could include \n",
    "### converting to grayscale, etc.\n",
    "\n",
    "def graylize_images(X_data):\n",
    "    \"\"\"\n",
    "    Conver the input image data from RGB to Gray8\n",
    "    \"\"\"\n",
    "    images_gray = np.average(X_data, axis=3, weights=[0.2989, 0.5870, 0.1140])\n",
    "    images_gray = np.expand_dims(images_gray, axis=3)\n",
    "    return images_gray\n",
    "\n",
    "def normalize_images(X_data):\n",
    "    \"\"\"\n",
    "    Normalize the input image data\n",
    "    \"\"\"\n",
    "    X_data = X_data.astype('float32')\n",
    "    X_data = (X_data - 128.) / 128.\n",
    "    return X_data\n",
    "\n",
    "X_train = normalize_images(X_train)\n",
    "X_test = normalize_images(X_test)\n",
    "X_valid = normalize_images(X_valid)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Architecture\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "\n",
    "EPOCHS = 30\n",
    "BATCH_SIZE = 128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "### Define your architecture here.\n",
    "\n",
    "def LeNet(x):    \n",
    "    # Arguments used for tf.truncated_normal, randomly defines variables for the weights and biases for each layer\n",
    "    mu = 0\n",
    "    sigma = 0.1\n",
    "    \n",
    "    # Layer 1: Convolutional. Input = 32x32x3. Output = 28x28x12.\n",
    "    conv1_W = tf.Variable(tf.truncated_normal(shape=(5, 5, image_shape[2], 12), mean = mu, stddev = sigma))\n",
    "    conv1_b = tf.Variable(tf.zeros(12))\n",
    "    conv1   = tf.nn.conv2d(x, conv1_W, strides=[1, 1, 1, 1], padding='VALID') + conv1_b\n",
    "\n",
    "    # Activation.\n",
    "    conv1 = tf.nn.relu(conv1)\n",
    "\n",
    "    # Pooling. Input = 28x38x12. Output = 14x14x12.\n",
    "    conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n",
    "    \n",
    "    # Layer 2: Convolutional. INput = 14x14x12, Output = 10x10x12.\n",
    "    conv2_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 12, 12), mean = mu, stddev = sigma))\n",
    "    conv2_b = tf.Variable(tf.zeros(12))\n",
    "    conv2   = tf.nn.conv2d(conv1, conv2_W, strides=[1, 1, 1, 1], padding='VALID') + conv2_b\n",
    "    \n",
    "    # Activation.\n",
    "    conv2 = tf.nn.relu(conv2)\n",
    "\n",
    "    # Pooling. Input = 10x10x12. Output = 5x5x12.\n",
    "    conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n",
    "\n",
    "    # Flatten. Input = 5x5x12. Output = 300.\n",
    "    fc0   = flatten(conv2)\n",
    "    \n",
    "    # Layer 3: Fully Connected. Input = 300. Output = 240.\n",
    "    fc1_W = tf.Variable(tf.truncated_normal(shape=(300, 240), mean = mu, stddev = sigma))\n",
    "    fc1_b = tf.Variable(tf.zeros(240))\n",
    "    fc1   = tf.matmul(fc0, fc1_W) + fc1_b\n",
    "        \n",
    "    # Activation.\n",
    "    fc1    = tf.nn.relu(fc1)\n",
    "\n",
    "    # Fully Connected. Input = 240. Output = 84.\n",
    "    fc2_W  = tf.Variable(tf.truncated_normal(shape=(240, 84), mean = mu, stddev = sigma))\n",
    "    fc2_b  = tf.Variable(tf.zeros(84))\n",
    "    fc2    = tf.matmul(fc1, fc2_W) + fc2_b\n",
    "\n",
    "        \n",
    "    # Activation, with .3 chance of dropout\n",
    "    fc2    = tf.nn.relu(fc2)\n",
    "    fc2    = tf.nn.dropout(fc2, 0.7)\n",
    "    \n",
    "    # Fully Connected. Input = 84. Output = 43.\n",
    "    fc3_W  = tf.Variable(tf.truncated_normal(shape=(84, 43), mean = mu, stddev = sigma))\n",
    "    fc3_b  = tf.Variable(tf.zeros(n_classes))\n",
    "    logits = tf.matmul(fc2, fc3_W) + fc3_b\n",
    "    \n",
    "    return logits"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Train, Validate and Test the Model\n",
    "A validation set can be used to assess how well the model is performing. A low accuracy on the training and validation\n",
    "sets imply underfitting. A high accuracy on the training set but low accuracy on the validation set implies overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "x = tf.placeholder(tf.float32, (None, 32, 32, 3))\n",
    "y = tf.placeholder(tf.int32, (None))\n",
    "one_hot_y = tf.one_hot(y, 43)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rate = 0.001\n",
    "\n",
    "logits = LeNet(x)\n",
    "cross_entropy = tf.nn.softmax_cross_entropy_with_logits(labels=one_hot_y, logits=logits)\n",
    "loss_operation = tf.reduce_mean(cross_entropy)\n",
    "optimizer = tf.train.AdamOptimizer(learning_rate = rate)\n",
    "training_operation = optimizer.minimize(loss_operation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model Evaluation\n",
    "Evaluate how well the loss and accuracy of the model for a given dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "correct_prediction = tf.equal(tf.argmax(logits, 1), tf.argmax(one_hot_y, 1))\n",
    "accuracy_operation = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
    "saver = tf.train.Saver()\n",
    "\n",
    "def evaluate(X_data, y_data):\n",
    "    num_examples = len(X_data)\n",
    "    total_accuracy = 0\n",
    "    sess = tf.get_default_session()\n",
    "    for offset in range(0, num_examples, BATCH_SIZE):\n",
    "        batch_x, batch_y = X_data[offset:offset+BATCH_SIZE], y_data[offset:offset+BATCH_SIZE]\n",
    "        accuracy = sess.run(accuracy_operation, feed_dict={x: batch_x, y: batch_y})\n",
    "        total_accuracy += (accuracy * len(batch_x))\n",
    "    return total_accuracy / num_examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Train the Model\n",
    "Run the training data through the training pipeline to train the model.\n",
    "\n",
    "Before each epoch, shuffle the training set.\n",
    "\n",
    "After each epoch, measure the loss and accuracy of the validation set.\n",
    "\n",
    "Save the model after training.\n",
    "\n",
    "You do not need to modify this section."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training...\n",
      "\n",
      "EPOCH 1 ...\n",
      "Validation Accuracy = 0.747\n",
      "\n",
      "EPOCH 2 ...\n",
      "Validation Accuracy = 0.834\n",
      "\n",
      "EPOCH 3 ...\n",
      "Validation Accuracy = 0.873\n",
      "\n",
      "EPOCH 4 ...\n",
      "Validation Accuracy = 0.882\n",
      "\n",
      "EPOCH 5 ...\n",
      "Validation Accuracy = 0.878\n",
      "\n",
      "EPOCH 6 ...\n",
      "Validation Accuracy = 0.907\n",
      "\n",
      "EPOCH 7 ...\n",
      "Validation Accuracy = 0.915\n",
      "\n",
      "EPOCH 8 ...\n",
      "Validation Accuracy = 0.914\n",
      "\n",
      "EPOCH 9 ...\n",
      "Validation Accuracy = 0.909\n",
      "\n",
      "EPOCH 10 ...\n",
      "Validation Accuracy = 0.918\n",
      "\n",
      "EPOCH 11 ...\n",
      "Validation Accuracy = 0.909\n",
      "\n",
      "EPOCH 12 ...\n",
      "Validation Accuracy = 0.924\n",
      "\n",
      "EPOCH 13 ...\n",
      "Validation Accuracy = 0.934\n",
      "\n",
      "EPOCH 14 ...\n",
      "Validation Accuracy = 0.932\n",
      "\n",
      "EPOCH 15 ...\n",
      "Validation Accuracy = 0.929\n",
      "\n",
      "EPOCH 16 ...\n",
      "Validation Accuracy = 0.933\n",
      "\n",
      "EPOCH 17 ...\n",
      "Validation Accuracy = 0.929\n",
      "\n",
      "EPOCH 18 ...\n",
      "Validation Accuracy = 0.938\n",
      "\n",
      "EPOCH 19 ...\n",
      "Validation Accuracy = 0.931\n",
      "\n",
      "EPOCH 20 ...\n",
      "Validation Accuracy = 0.936\n",
      "\n",
      "EPOCH 21 ...\n",
      "Validation Accuracy = 0.938\n",
      "\n",
      "EPOCH 22 ...\n",
      "Validation Accuracy = 0.936\n",
      "\n",
      "EPOCH 23 ...\n",
      "Validation Accuracy = 0.946\n",
      "\n",
      "EPOCH 24 ...\n",
      "Validation Accuracy = 0.929\n",
      "\n",
      "EPOCH 25 ...\n",
      "Validation Accuracy = 0.928\n",
      "\n",
      "EPOCH 26 ...\n",
      "Validation Accuracy = 0.946\n",
      "\n",
      "EPOCH 27 ...\n",
      "Validation Accuracy = 0.952\n",
      "\n",
      "EPOCH 28 ...\n",
      "Validation Accuracy = 0.941\n",
      "\n",
      "EPOCH 29 ...\n",
      "Validation Accuracy = 0.939\n",
      "\n",
      "EPOCH 30 ...\n",
      "Validation Accuracy = 0.947\n",
      "\n",
      "Model saved\n"
     ]
    }
   ],
   "source": [
    "\n",
    "accuracy = np.zeros(EPOCHS)\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    num_examples = len(X_train)\n",
    "    \n",
    "    print(\"Training...\")\n",
    "    print()\n",
    "    for i in range(EPOCHS):\n",
    "        X_train, y_train = shuffle(X_train, y_train)\n",
    "        for offset in range(0, num_examples, BATCH_SIZE):\n",
    "            end = offset + BATCH_SIZE\n",
    "            batch_x, batch_y = X_train[offset:end], y_train[offset:end]\n",
    "            sess.run(training_operation, feed_dict={x: batch_x, y: batch_y})\n",
    "            \n",
    "        validation_accuracy = evaluate(X_valid, y_valid)\n",
    "        accuracy[i] = validation_accuracy\n",
    "        print(\"EPOCH {} ...\".format(i+1))\n",
    "        print(\"Validation Accuracy = {:.3f}\".format(validation_accuracy))\n",
    "        print()\n",
    "        \n",
    "    saver.save(sess, './lenet')\n",
    "    print(\"Model saved\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Discussion on training results\n",
    "In the beginning, I did not normalize, and I found that my model accuray was stuck around ~0.7. Once I normalize the images, the accuray immediately goes up to 0.94.\n",
    "\n",
    "Also, converting images into grayscale does not help improve the my model. I am seeing accuray drops down to 0.5 for some reason.   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fb88ca0af28>]"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtoAAAFNCAYAAAA+ZchVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3Xd40+e9/vH3x9t4socHe68QDGSSHbLTpD0t2Tvtr21O\n956nK+k5Pd1pG7JHaZLTJmnaZjZ7MUwAM8Iw07IBG4z3lp7fH/qaCMeAbCzL435dly5L+q6PZCXc\nfvQMc84hIiIiIiJdKybaBYiIiIiI9EUK2iIiIiIiEaCgLSIiIiISAQraIiIiIiIRoKAtIiIiIhIB\nCtoiIiIiIhGgoC0iUWdmN5jZ89GuQ6LLzG41s9ejeP3Pm1mpmdWYWUa06gipJ6rvh4gcPwVtEfkI\nL2i03gJmVh/y+Jquvp5z7mHn3IVdfV45Pmb2mJk5Mzsx5LkpZtYSzboiwcySgF8AZznnUp1zldGu\nSUR6PwVtEfkIL2ikOudSgd3ApSHP/bnt/mYW1/1V9nxmFmNmvf3/s+XAT6JdREd14jM5Akh0zm2I\nRD0i0j/19n8ARCQKzOwnZvaEmf3FzKqBa83sZDNbZmYVZrbHzH5rZvHe/nFey+inzazQzA6a2W9D\nznfoK/Iw9o01s1+b2QEz225md5jZEZe4NbPvevtVm9kGM7uszfZPm9kmb/t6M5vtPT/azJ4xszIz\n229mvwl57Q+FHD8h9Ppm9raZ/djM3gNqgVzv9X3gXWObmd3apoYrzWyNmVV5r/l8M7vKzJa32e/r\nZva3dl7jNWa2rM1zXzOzp7z7l4Rc32dmXzrS+9WOB4E8Mzu1vY3e+c4MeXzo/Wl9b8zsRm+/cjO7\nzcwWmNk677PymzanjDGzP5hZpVfzWSHnzjSzB73Pl8/MftT6h4z3Hr/pfe7Kge+2U2uSt32PmRWb\n2S/NLMHMpgIbvH1qzOylI7zWU0M+42vMbGHItrfN7Kdmlu/V/rSZDQzZfoX3+asws1fNbHLItnY/\nax9utl95x203s/PDfD8mee9HpXfOpe29JhGJLAVtEemsK4ClQAbwBNACfAEYApwKXAB8us0xFwFz\ngTkEw/m5Rzn/kfb9f8C5wCwgD7jyGHVu8erJAH4KLDWz4QBmdhXBQHYNkO6dq9yCraH/AgqBMUAO\n8OQxrhPqOuBm75w+YB9wsff4NuB3ZjbLq+EU4AHgK0AmcBawC3gGmGxmE9uc95F2rvd3YKaZjQt5\n7mqCvx8IhuVbnHNpBN+3NzrwWmqAuwi+d52VB4wHrgV+C3wTOBuYQfB3GxriTwE2Efwc/Rh4yswy\nvW2PAvXeueYSfE9vanPsB8BQ4Oft1PF9r5ZZBD9XpwLfcs59AMyGQ9/mnN/2QDPLAZ4FfgAM8l7D\nU2Y2OGS3673bKMCAX3nHTvVqv8Or7d/As2YWH8Zn7RRgHTDYO9/9IduO9n781DvvQCAbuLud90NE\nIkxBW0Q6623n3D+ccwHnXL1zbqVzbrlzrsU5tx1YApzR5pg7nXOVzrmdwOvACUc5/5H2/STwK+dc\nsXOunPYD1SHOuSedc3u8OpcCOwmGLYBbgbucc6tc0BbnXBFwMsGg9w3nXK33+t4J610JesA594Fz\nrtl7P/7hnNvuXeNV4BXgdG/fW4B7nXOveDUWOec2O+fqgf8jGE4xsxOAkcBz7bzGGuAfwGJv3ynA\nOOCf3i7NwDQzS3POlTvn3u/AawH4AzDRzM7r4HGtfuyca3TOPQc0AY8558qccz7gbYKht9Ue4Hfe\ne7cU2AFcaGZZBP/A+pJzrs45tw/4Nd5r9ux2zv3ROef33r+2rgF+6F27FPgRwT9ewnE98Kxz7kXv\n9/QCsJbgH5StHnbObXTO1RIM9YvNzLwan3XOveqcayb4h0sGsIBjf9a2OececM75gYeBbDMbEsb7\n0UwwuI90zjV08PMrIl1EQVtEOqso9IEFB8n9y8z2mlkVwRAzpM0xe0Pu1wGpRzn/kfYd1ebah9XR\nltdtYa331XsFMCWkrhxgWzuH5QA7vXDTGW3fm0vMbLnXdaICOD+MGiAYrFoHn14LPOEFtfYsBa7y\n7l8DPOWca/AeXwFcBuw2s9fNbEFHXox3np8QbGHuMC8Etqon2MIf+jj0c+BzzoV2BdpF8Hc+GkgE\n9oX8Lu8Ghofse9TPgneeXW3OnRXWiwhe/6rWa3vXP8k7Z3vX3+XVO6jtdZ1zAYLfdGRx7M9a2/8O\nIPh+Hev9+AoQD+R73XRuCPN1ikgXUtAWkc5q2y/6HmA9MME5l06wRc8icN09BL8Kb5VzpB29rhR/\nJNjdZLBzLpNgt4TWuooIfu3eVhEw2sxi29lWCwwIeTyinX1C+2wnA38F7gSGezW8FEYNOOfe9s5x\nKsGuII+2t5/nBSDLzGYSDNyH+uR63zRcBgwj2Mr9+FHOcyT3ecdf1ub5cN6Pjshu8zgXKCH4PtUB\ng5xzmd4t3Tk3K2TfI/bV95QQDKih5y4Os64i4MGQa2c651Kcc/8Tsk/oZzEXaCQ4mPSw63r9qLO9\nax/ts3aseo74fnjf4tzqnBsJfA5YYmZjO3gNETlOCtoi0lXSgEqg1uuT2rZ/dld5EviimY3yBpt9\n7Sj7phIMX2UEB5XdRrBFu9V9wNfNbI4FTfT64r4HHAB+ZmYDzCw5pB/xGuAMM8vx+g5/8xj1JgIJ\nXg1+M7sEOCdk+/3ArWZ2lgVnKckOHShHMFz/Eahxzh024DGUc64J+BvwSyAFeJXgi042s6vNLN1r\nDa8GAseoub3zNwP/BXyjzaY1BLtIxJnZfI7dZ/5YRlpwPus4M1tM8I+QF7wuPW8AvzCzdO+9mhA6\nIDEMfwG+73W9GAp8D3gszGMfBa4ws/MsOCA3yfudhbZoX+99s5NC8L160mudfxK4zMzOtOAA4a8R\n/D0s5+iftSM61vthZp/0upcAVBD876Cz39CISCcpaItIV/kKcAPBAHEPwQGSkfBHgn221wGrCA74\nampvR+dcAfA7YAXBlvDJBMNN6/a/EOzj/QRQBTwFDHTOtQCXAFMJthzuBj7hHfYC8LR3/RUEB8gd\nkXOuAviSd0y5d55/hmx/l+AAyd8S/EPlNQ5vGX2E4KDBo7Vmt1pKsN/uk226ItwA7PK69NzCh/2+\nx1lwlo1RHz1Vux4DSts89x2Cf7xUEAyuxzu7xbvAdILv1Q+BjzvnDnrbriX4R8RG4CDBPuwdaUH/\nL4L9qtcDBQQ/C3eGc6A3VuAKgq+xjOBn4isc/u/oowTfoz1ALPBF79gNBH8Hf/SOvQC4rLUPP0f+\nrB3L0d6PBcBKM6sl+Ln+nHNud5jnFZEuYod3hRMR6V3M7FLg1865drtf9HZe62gpMMM5tyPa9Uj7\nzOxt4D7n3EPRrkVEeg61aItIr2JmKWZ2gde1IJtgX/Cno11XBH0OeEchW0Sk99FqbiLS2xjBOYL/\nSnAg3j8Jdgnoc8zMR3CatsujXYuIiHScuo6IiIiIiESAuo6IiIiIiESAgraIiIiISAT0mT7aQ4YM\ncWPGjIl2GSIiIiLSx61atWq/c27osfbrM0F7zJgx5OfnR7sMEREREenjzGxXOPup64iIiIiISAQo\naIuIiIiIRICCtoiIiIhIBChoi4iIiIhEgIK2iIiIiEgEKGiLiIiIiESAgraIiIiISAQoaIuIiIiI\nRICCtoiIiIhIBEQ0aJvZBWa22cwKzeyb7WwfbWavmFmBmb1uZtkh2/xmtsa7PRvJOkVERET6kvLa\nJl7csJfy2qZol9KvRWwJdjOLBe4GzgN8wEoze9Y5tzFkt18AjzjnHjazs4E7geu8bfXOuRMiVZ+I\niIhIX1JW3chLG/fy3Lo9LNtejj/gyEiO5+sXTOaqebnExFi0S+x3Iha0gflAoXNuO4CZPQ5cDoQG\n7WnAl737rwHPRLAeERERkT6ltKqBFzYEw/WKHeUEHIwdksJnzhhH3uhB/OmNbXzn6fU8me/jJ5fP\nYGZ2RrRL7lciGbSzgKKQxz5gQZt91gJXAr8BrgDSzGywc+4AkGRm+UALcJdzTiFcRERE+r09lfW8\nsH4vz6/by8pd5TgH44em8PmzJnDRrJFMHp6GWbD1+szJQ/n7mhJ+8q8PuOzut7l2wWi+ev5kMgbE\nR/lV9A+RDNrh+CrwezO7EXgTKAb83rbRzrliMxsHvGpm65xz20IPNrPbgdsBcnNzu69qERERkW7k\nO1jHC+uDLdfv764AYMqINL54ziQumjmCicPT2j3OzPjYnCzOmjKMX728hUfe28lz6/bw7YumcuWJ\nWYcCuUSGOecic2Kzk4EfOucWeY+/BeCcu/MI+6cCm5xz2e1sewj4p3Pur0e6Xl5ensvPz++K0kVE\nRESibveBOp5fv4fn1u1hra8SgGkj07l41kgumDGC8UNTO3zO9cWVfPeZ9awpqmD+mEH8+GMzmDyi\n/ZAuR2Zmq5xzecfcL4JBOw7YApxDsKV6JXC1c25DyD5DgHLnXMDMfgr4nXPfN7OBQJ1zrtHb5z3g\n8jYDKQ+joC0iIiK93Y79tTy3bg/Pr9/D+uIqAGZlZ3DhjJFcOGMEY4akHPc1AgHHk/lF3PXCJqob\nWrj51DF84dxJpCZGu6ND1yqvbeKp933ceMoY4mK7dqK9cIN2xN5R51yLmX0eeBGIBR5wzm0wsx8B\n+c65Z4EzgTvNzBHsOvI57/CpwD1mFiA4BeFdRwvZIiIiIr3Vnsp6/prv41/r9rBpbzUAc3Iz+c5F\nU7lgxghyBg3o0uvFxBiL5+dy/vQR/PcLm7j3rR08u7aE710yjYtnjuz13Um27qvmgXd28NT7xTS2\nBJiRlcFJ4wZHpZaItWh3N7Voi4iI9C4NzX4efW8Xm/dV8/1Lp5Ge1P8G6G0rq2HxkmWUVTeSN3og\nF84MtlyPykzuthre332Q7z2zng0lVZw+cQj/ddl0xnWiW0o0Oed4a+t+7nt7B29uKSMxLoYrT8zm\nltPGMGFY13eNiXrXke6moC0iItI7BAKOZ9YU878vbaG4oh4zmDEqg0duns/AlIRol9dttnshO+Ac\nj926gCkj0qNWiz/geGzZLn7x4mYaWwLcvnAcnztrAskJsVGrKRwNzX6eWV3MA+/sYMu+GoamJXL9\nSaO55qTRDIrgZ0lBW0RERHoU5xxvbt3PXc9v4oM9VczMyuBbF06hocXPZx57nzGDB/DYLQsYlp4U\n7VIjbsf+WhYveY8Wv+Mvt5/EpCPMGtLdSqsbuPO5TTy9upjsgcn88NLpnDtteLTL+oiy6kYeXbaL\nPy/bxYHaJqaNTOeW08ZyyeyRJMZF/o8DBW0RERHpMdYXV3Ln8x/wTuEBcgYl89XzJ3PprFGHVit8\nd9t+bn04n6Fpifz51gVkD+zafsk9ya4DtXzqnmU0+QMsvS26LdlH8t62A3z/7+vZWlrDuVOH8YNL\np3d5X/HO2LS3ivvf2sHf15TQHAhwzpRh3HLaOE4aN6hb+5YraIuIiEjUFZXX8T8vbubZtSUMHBDP\nHWdP5JqTctttdXx/90FufGAFKYlx/PnWBb2un3A4dh+oY/GS96hr9rP01pOYNqrnhexWzf4AD7y9\ng9+8shV/wHHH2RO4beG4bmkxDhUIOF7fUsr9b+/gncIDJMfH8h952dx06ljGdsEsLJ2hoC0iIiJR\nU17bxO9fLeTRZTuJjTFuOW0snz5j/DEHPG4sqeK6+5djBo/esoCpI3tuEO2oovI6Fi9ZRk1jC0tv\nW8D0Ub1jOfSSinp+/M+NPL9+L+OGpPDVRZOZNDyNUZlJDEiI3JSA9U1+/va+jwfe2cH2slpGpCdx\nwyljuGp+DpkDotuXX0FbREREul19k58H3tnBn17fRm1TC5/My+GL505iREb4/a63ldVw7X3LqW1s\n4eGb5zMnd2AEK+4evoPBkF1V38zS205iRlbvCNmh3thSxg/+vp6dB+oOPZeRHM+ozGRGZSQxMjOJ\nkRnJjGr9mZHMiIwkEuI6Nof1vqoGHn53J0tX7KairplZ2RncctpYLpo5kvgung+7sxS0RUREpNv4\nA46/ririVy9vZW9VA+dOHcY3LphyxKXBj6WovI5r7lvOgZpG7rthHiePj848yF2hpKKeTy15j4q6\nZpbeehIzs3tfyG7V2OJnze4KSirrKaloYE9lPXsqGiipbKCkop7K+uaPHDMkNZEsL3yPzExilPez\nNZQPS0siNsZYX1zJ/W/v4J8FJfgDjvOnjeCW08eSN3pgj5vbW0FbREREIs45x6ubSrnr+U1sLa3h\nhJxMvn3RVOaPHXTc595X1cC19y1nd3kdf7p2LmdNGdYFFXevPZX1LF6yjPKaJh67dQGzczKjXVJE\n1TW1tAngIT+9MF7X5D/smNgYY1BKAmXVjaQkxPLJeTncdMpYcgdHf/DlkShoi4iI9BNNLQHKahrZ\nV9VAaVUjpdUNOAcjM5IYlZnMyIwkBqUkdHmr4OrdB7nz+U2s2FHO2CEpfH3RZC6YMaJLr1Ne28T1\nDyxn895qfv2pOVw8a2SXnTvS9lY2cNW9wcVoHr2lb3SBOV7OOarqW7zgHdIqXtnAtJHpfHJeTq9Y\nuCjqS7CLiIh0twJfBUve3M6WfdVcPHMUn5yXzciM7lthr6u1+APsr2liX1VD8FbdSFlVA/uqGtlX\nHfxZWtXAgdqmY54rMS6GkRkffn2flZn8ka/yww04O/bX8j8vbuK5dXsZkprAjz82g8XzciLSf3ZQ\nSgJLbzuJmx9cyR1/eZ+6pln8R15Ol1+nq5VWNXD1vcsorWrgEYXsQ8yMjAHxZAyI71MDXY9ELdoi\nItKrBQKO1zaXsuTN7SzfUU5aYhxTRqaxcudBYgzOmDSUT83L5Zypw3rMQKpW1Q3N5O88yJ7KYJAu\nrQ4G59YQvb+mkbb/TMdYsM/r8PQkhqcnMiw9ieFpwfvD05MYmhb8acahr+xLKj782n5PZQN7KurZ\nV92IP3D4yVMT44JhPDP5wz61Ia3iifGx3PPGNpYu301CXAy3nT6O2xaOIzUx8u12dU0tfPrRVby1\ndT//ddl0bjhlTMSv2Vml1Q0sXrKMvZUNPHzzfOaNOf5uNNKzqOuIiIj0aY0tfv6+uoQlb22nsLSG\nURlJ3HzaWD41L4e0pHiKyut4Mr+IJ/OL2FfVyJDURD4xN5tPzcuJ2ty7EFzR7t8f7OPFDXt5t/AA\nTf4AAGYwOCWBYSGheZgXpoNBOnh/cGoisTHH3zWjxR/sblJS8eHX96Ff45dUNLC/pvEjx8XGGFfN\nz+E/z5nIsLTuXcGxscXPHUtX89LGfXxt0WQ+d9aEbr1+OMqqG7nq3mWUVNTz0E3zu6SvuvQ8Ctoi\nItInVdY189jyXTz07k7KqhuZOjKdTy8cx8Wz2p/6q8Uf4I0tZTy+sohXN5XiDzhOGjeIxfNyuWDG\nCJLiI7/4RlF5HS9u2MuLG/aSv+sgzkHOoGQWTRvB2VOHMXZICkNSE3tci3tji5+9XujeU1nPgZom\nzp46jPFRXEimxR/ga38t4OnVxfy/M8fz9UWTe8yMFPtrGrn63mUUldfz4E3zOGlc750pRY5OQVtE\npItUNzSzsaSKYelJjMxI6pZgJh9VVF7HA+/s4ImVRdQ1+Tl94hA+vXA8p04YHHbQ2lfVwF9X+Xhi\nZRG7y+vISI7nijlZfGpeTpf2F3XOsWlvtReu9/HBnioApoxIY9H0ESyaPoKpI9N6TEDsbQIBx/f+\nvp4/L9/NDSeP5geXTj+0lHu0lNc2cfW9y9h5oJYHbpzHKeOHRLUeiSwFbRGR49TY4uexZbv5/atb\nOVj34dywg1ISDg0qG5X5Yf/V1p/D05N6XMvk8SipqGflznIamwPMyMpg4vDUbn1963yV3PPmNp5b\nt4cYMy6bPYrbFo47rmAcCDiWbT/A4yuLeGH9Xpr8AWZnZ7B4fi6Xzh7VqT7H/oDj/d0HeXH9Xl7a\nuI/d5XWYQd7ogSyaPoLzp43o0dOV9TbOOe58fhNL3tzOJ+Zmc9eVM4mL0n93B2ubuOreZezYHwzZ\np05QyO7rFLRFRDopEHD8o6CEX7y0maLyek6bMITrTx5NdUNLsB+rN5isxBtoVt3QctjxZjAsLfGw\nFdJGZnizPHgrqA1JTYx6C1x7nHNs31/Lih3lrNxRzvId5RRX1B+2T2JcDFNHpjMrO4OZWRnMys5k\n/NCULg05zjle31zGkje38972A6QmxnH1glxuOnVMl88icrC2iadXF/P4yt1s2VfDgIRYLpk1ksXz\nc5mTk3nUVufGFj/vbjvASxv28vLGfeyvaSIhNoZTJgxm0fQRnDt1OEPTEru0XvmQc47fvVrIL1/e\nwkUzR/DrT83p8CqEx6uiromr711OYVkN912fx8JJQ7v1+hIdCtoiIp3wTuF+7nz+A9YXVzF1ZDrf\nunDKMf/hrGlsCQbv1gB+aHaHDxdqaGgOHHZMfKwxPD2JMYNTmDg8lYnD0ryfqWQOSIjkSzyMP+D4\nYE9VMFjvDN721wSnihucksD8sYOYN2YQ88cOYkBCLOuKK1nnq6SguJINxZXUegtPJMXHMH1Ua/AO\n3sYOSe3woL3GFj9/X1PCfW9tZ8u+GkakJ3HzaWNYPD834nPrOudYXVTBEyuK+EdBCXVNfiYNT+VT\n83K5ck4WA1OCv5eaxhZe21TKixv28vrmMmoaW0hJiOXMKcNYNH0EZ00eSlovmAe4L7n/7R38+J8b\nOXPyUP507dxu695VWdfM1fctY+u+GpZcP5czJ/e+BXWkcxS0RUQ6YGNJFXe9sIk3t5SRlZnMVxdN\n4vLZWV3S6uyco6Ku+bAV0koqgmF854Fatu6rob75w5XShqQmMnFY6qHgPcEL4YO7YMGRxhY/Bb5K\nVuwoZ8WOct7fdZDqxmCLfFZmMgvGDmLe2GCwHjck5ajXCwSCrd/riitY56tiXXEF64urDr2WlITY\nYPj2gvfMrAzGDE5p9z2trG/mz8t38dA7OymtbmTKiDRuXziOS2aN6vYWSgiG6X+sLeHxlUWsLaog\nITaG86YPp66xhXe8mUIGpyRw7tThLJoxnFPGD1Hf/Sh7fMVuvvX0OuaPGcT9N86L+JSDlfXNXHtf\ncCGde67rnatWSucpaIuIhMF3sI5fvrSFp9cUk54Uz+fPmsB1J4/u1tAUCDhKKuvZWlpD4b4atpZW\nH7rfGoIBBg6IZ+KwNCZ4Aby1FXxYWuIRA3FNYwvv7zrIyp3BbiBriipoagm2rk8clhoM1WOC4Tor\n8/i7ZPgDjm1lNazzVbKuuJICXwUbSqpo9K6ZlhjH9Kx0ZmVnMjMrg9GDB/D3NSU8vmI3td4Ax9tO\nH8fpE4f0mIGCH+yp4omVRTy9upi0pDivv/Vw8sYM6pJp9qTrPLu2hC8/sYbpWRk8fNO8iH07VNXQ\nzHX3LWfjnir+dO1czpk6PCLXkZ5LQVtE5Cgq6pq4+7VCHn5vFwA3nTqGz54xgYwBPecrf+cc+6oa\n2VpazZZ9NRSWVrN1Xw1bS2uorP9wcGZ6UhwTh6d5rd+pDE1LZJ2vkhU7y9lQUoU/4IgxmJGVcagb\nyLwxgxiU0j1dVFr8AQrLaijwVR4K4Bv3VB0K/HExxqWzR3Hr6WOZPiqjW2rqjNZ/L3vKHwDSvn9v\n3Mdnl77PuCEpfO+SaQxIiCU5IZbk+OAt0fsZH2ud+l1WNzRz/QMrWF9cyR+umct50xSy+yMFbRGR\ndjQ0+3n43Z3c/Voh1Y0tfPzEbL583iRGdUFrbndxzlFW0+i1fnst4PtqKCytObQUd0JcDCfkZDLf\nC9Ynjh7YLav3havZH2DLvmoKS2uYN2ZQr3r/ped7p3A/tz6cf1iXrLZiY4ykuBiSE2JJ8sL3oZ8J\nsYe2tT7fuu21zaWsL67k7mtOZNH0Ed34qqQnUdAWEQnhDzieXl3ML1/aTEllA2dNHso3LpzClBFd\nN3dyT3CgppG9VQ2MH5qqPsPSr+2ramB7WS0NzX7qm/2HftY3Be83NAeCj5tbHwe31Xvb2tu/yR8g\nKT6GX33yBC6cOTLaL1GiKNyg3XOaN0REIsA5x+tbyvj585vYtLeaWdkZ/OKTs/vsYhKDU4NLdIv0\nd8El67t2iXh/wBFwrk/Nky+RpaAtIn1Wga+CO5/bxHvbD5A7aAC/u2oOF88c2SPnrxaRni82xohF\n//+Q8Cloi0ifs/tAHf/z0mb+sbaEQSkJ/PDSaVy9YHRUpokTEZH+S0FbRPqMhmY///3CZh5dtpPY\nGOOOsydw+8JxWjxERESiQkFbRPqE8tombnskn1W7DrJ4Xg5fOm9Sl/fPFBER6QgFbRHp9baX1XDT\nQyvZU9nA3VefyMWzNBuAiIhEn4K2iPRqy7cf4PZHVxEbY/zltpOYO3pgtEsSEREBFLRFpBd76n0f\n3/hbAbmDBvDgjfPJHTwg2iWJiIgcoqAtIr2Oc45f/3srv3llKyePG8yfrp3bo5ZOFxERAQVtEell\nGlv8fPNv63h6dTGfmJvNz66YqWn7RESkR1LQFpGjKq1q4L3tBzhryjDSozxNXkVdE7c/uooVO8r5\n6vmT+NxZEzDT4hEiItIzKWiLSLv2VNZzzxvbWbpiN00tATIHxPPZM8dz/cljSIqP7fZ6du6v5eaH\nVuI7WM9vFp/A5SdkdXsNIiIiHaGgLSKHKa6o54+vF/LkSh8B57jyxCwunDmSh97Zyc+e28QDb+/k\ni+dO5BNzs4mL7Z4uG/k7y7ntkXwA/nzbAuaNGdQt1xURETkeCtoiAkBReR1/eH0bf11VBMAn5ubw\n2TPHkzMoOJPHWZOH8d62A/z3i5v45lPrWPLmdr66aDIXzhgR0e4bz64t4atPriVrYDIP3DiPsUNS\nInYtERGRrmTOuWjX0CXy8vJcfn5+tMuQPqi+yU9TS4CYGIiNseDNgj/7Qv/gXQdqufu1Qp56v5gY\nMz41L4fPnDmerMzkdvd3zvHyxn38z4ub2Vpaw8ysDL5xwRROmzikS+tyznH3a4X84qUtzB8ziHuu\nm8vAlIQuvYaIiEhnmNkq51zeMfdT0Bb5qEDA8c62/SxdvpuXN+6jJdD+fydmEBdjxFhIAI8N/owJ\nCeSttxgcFDbjAAAgAElEQVRrDesxZCbHc8r4wSycNJSZWRnExHRvaN9eVsPdr23jmTXFxMYYV8/P\n5TNnjGdERnjLlvsDjqdXF/Orl7dQXFHPKeMH8/ULpnBCTuZx19bUEuDbT6/jr6t8fOyEUfz8E7NI\njOv+fuEiIiLtUdAW6YQDNY38dZWPpSt2s+tAHZkD4rlyTjZZA5MJBBwtAUfAOfyBkJtzBELutz7f\nul9LwNvu8M4RwB8IDjbcUFIFwMAB8Zw+cSgLJw1l4cQhDEsPL+x2RmFpNb9/tZBn15aQEBfDNQtG\n8+mF4zp9zcYWP0uX7+b3rxZyoLaJC6aP4KuLJjFhWFqnzldZ18xnHlvFe9sP8IVzJvLFcyf2iW8O\nRESk71DQFgmTc47lO8pZunw3L6zfS5M/wLwxA7lmwWgumDEiojNs7K9p5O2t+3lzSxlvbt3P/ppG\nAKaMSOOMScHgnTdmYJe05m7eW83vXt3Kv9btISkulutPHs2tp49jaFricZ8boKaxhfvf2sG9b22n\nrqmFT8zN5ovnTmLUEbqgtGf3gTpuemgFu8vr+PnHZ3HlidldUpuIiEhXUtAWOYaKuib+9n4xS5fv\nYltZLWlJcXz8xGyuXpDLpOGda409HoGA44O9VbyxpYw3t5SxatdBmv2O5PhYTh4/mIUTh7Bw0lDG\nDknpUAvvxpIqfv/aVp5bt5eUhFiuP2UMt542lsGpXROw2zpQ08gfXt/Go+/tAoPrTxrNZ8+awKBj\n9K9+f/dBbns4n5aA457r5nLSuMERqU9EROR4KWiLtMM5x/u7K1i6fDf/LCihsSXACTmZXL0gl0tn\njSI5oef0A65pbGHZtgO8ubWMN7aUsetAHQA5g5JZ6HUzOWX8YNKOsIjM+uJKfvvKVl7auI+0xDhu\nPHUMN586ttsGFBZX1PPrl7fwt/d9DEiI4/aF47jltLGkJH50sqPn1u3hS0+sYURGEg/cOI/xQ1O7\npUYREZHOUNAWCVHd0Mwzq4v58/LdbNpbTUpCLB+bk8XVC3KZPioj2uWFZdeBWt7cUsYbW/bz3rb9\n1Db5iYsxTswdyMJJwdbuGaMyWOcF7Fc2lZKeFMfNp43lplPGkjEgOqs6FpZW84sXt/DChr0MTkng\njrMncNWCXBLjYnHOcc+b27nr+U3MHT2Qe6/PO2bLt4iISLQpaIsA63yVLF2xi7+vKaGuyc/0Uelc\ns2A0l50witR2WlZ7i6aWAKt2HeTNrcFuJq2DKtOT4qhqaCEjOZ5bTxvLDaeOifqy6a1W7z7If7+w\nmfe2HyB7YDJfOncS+bvK+cuKIi6ZNZJf/MfsqKw4KSIi0lEK2tJv1TW18OyaEpau2E2Br5Kk+Bgu\nmz2KqxeMZnZ2Rp+cwaKsupG3C8t4t/AA44amct3Jo3vkHxLOOd4u3M/PX9jE+uLgHwefP2sCXz5v\nUrdPbygiItJZCtrS7xSV13HvW9t5+v1iqhtbmDQ8lavn53LFidlkJPeMVl0JCgQcL23chxksmj4i\n2uWIiIh0SLhBu+c1eYl0QkOzn+vuX05JZQMXzxzJ1QtyyRs9sE+2XvcFMTHGBTMUsEVEpG9T0JY+\n4e7XCtl5oI4/37qAUyd07VLgIiIiIp0RE+0CRI5XYWk1f3pjG1fMyVLIFhERkR5DQVt6Necc3356\nPQMS4vjOxVOjXY6IiIjIIQra0qv93yofK3aU860LpzAkQisdioiIiHRGRIO2mV1gZpvNrNDMvtnO\n9tFm9oqZFZjZ62aWHbLtBjPb6t1uiGSd0jsdqGnkZ899wLwxA/lkXk60yxERERE5TMSCtpnFAncD\nFwLTgKvMbFqb3X4BPOKcmwX8CLjTO3YQ8ANgATAf+IGZDYxUrdI7/ey5TdQ0tPDTK2ZqDmYRERHp\ncSLZoj0fKHTObXfONQGPA5e32Wca8Kp3/7WQ7YuAl51z5c65g8DLwAURrFV6mXe37edv7/v49Bnj\nmDQ8LdrliIiIiHxEJIN2FlAU8tjnPRdqLXCld/8KIM3MBod5rPRTjS1+vvv0enIHDeCOsydGuxwR\nERGRdkV7MORXgTPMbDVwBlAM+MM92MxuN7N8M8svKyuLVI3Sw/zx9W1s31/Ljz82g6T42GiXIyIi\nItKuSAbtYiB0hFq299whzrkS59yVzrk5wHe85yrCOdbbd4lzLs85lzd06NCurl96oO1lNfzhtW1c\nOnsUZ0zS71xERER6rkgG7ZXARDMba2YJwGLg2dAdzGyImbXW8C3gAe/+i8D5ZjbQGwR5vvec9GPO\nOb77zHoS42P43iWaM1tERER6togFbedcC/B5ggH5A+BJ59wGM/uRmV3m7XYmsNnMtgDDgZ96x5YD\nPyYY1lcCP/Kek37s6dXFvLvtAN+4YArD0pKiXY6IiIjIUZlzLto1dIm8vDyXn58f7TIkQg7WNnHO\nL99g9OAB/O0zp2g6PxEREYkaM1vlnMs71n7RHgwpEpa7nt9EZX0zP9Oc2SIiItJLKGhLj7diRzlP\n5Bdx6+ljmToyPdrliIiIiIRFQVt6tKaWAN9+eh1Zmcl84RzNmS0iIiK9R1y0CxA5mnvf2k5haQ0P\n3jiPAQn6uIqIiEjvoRZt6bF2Hajlt69s5aKZIzhryrBolyMiIiLSIQra0iO1zpkdHxvDDy6dHu1y\nRERERDpMQVt6pGfXlvDW1v18bdFkhqdrzmwRERHpfRS0pceprGvmx//8gNnZGVx70uholyMiIiLS\nKRpdJj3Oz1/cRHltIw/dNI9YzZktIiIivZRatKVHWbXrIEuX7+bmU8cyIysj2uWIiIiIdJqCtvQY\nzf4A335qHaMykvjSeZOiXY6IiIjIcVHXEekx7n97B5v3VXPv9XmkJOqjKSIiIr2bWrSlRygqr+PX\n/97C+dOGc9604dEuR0REROS4KWhL1Dnn+P7f1xNrxg8v05zZIiIi0jcoaEvUPb9+L69tLuPL509m\nVGZytMsRERER6RLqCCtHtXlvNf6AY9LwVOJiu/7vsqqGZn747AZmZKVzw8maM1tERET6DgVtadeG\nkkr+96UtvLqpFIABCbHMyMpgTk4mJ+RkckJuJiMzjr/1+X9f3Mz+mkbuuyEvIkFeREREJFoUtOUw\nhaXV/Orlrfxr3R7Sk+L42qLJZGUms6aogtVFFTz4zk6a/AEAhqcnBkN3zkBOyMlkVnZGh2YLWVNU\nwSPLdnHDyWOYlZ0ZqZckIiIiEhUK2gLA7gN1/PqVLTyzupjk+Fj+8+wJ3HL6ODKS4wH42JwsABpb\n/GwsqWJNUcWh24sb9gEQYzBpeJoXvoOt3hOHpbW7umOLN2f2sLREvnK+5swWERGRvkdBu5/bU1nP\n714t5MmVRcTGGLecNpbPnDGewamJ7e6fGBfLnNyBzMkdeOi58tom1not3muKKnh+/V4eX1kEQEpC\nLDOzMw61es/JzWR4ehIPvbuTjXuq+NO1J5KWFN8tr1VERESkOylo91P7axr54+vbeHTZLpxzXDU/\nl8+fPYHh6UkdPteglATOmjKMs6YMA4LT9e3YX3tYq/f9b2+n2e8AGJWRRHldE+dMGcai6SO69HWJ\niIiI9BQK2v1MZV0zS97axoPv7KSh2c/HT8zmP8+ZSM6gAV12DTNj3NBUxg1N5coTswFoaPazcU8V\na3YHg3dJRT3/dfl0zD7arURERESkL1DQ7idqGlt48O0dLHlrO9UNLVwyayRfOm8S44emdsv1k+Jj\nOTF3ICeGdDkRERER6csUtPu4hmY/jy3bxR9e30Z5bRPnTh3OV86fxNSR6dEuTURERKRPU9Duo5pa\nAjyZX8TvXt3KvqpGTp84hK+cP5kTcjSNnoiIiEh3UNDuY/wBx9Ori/nNK1soKq8nb/RAfrN4DieN\nGxzt0kRERET6FQXtPiIQcDy3fg+/enkL28pqmZmVwY9vmsEZk4ZqwKGIiIhIFCho9xF3v1bI/768\nhUnDU/nTtXNZNH24AraIiIhIFClo9wF1TS3c/84OzpkyjCXX57W7EqOIiIiIdK+YaBcgx+//8n1U\n1DXz2bPGK2SLiIiI9BAK2r2cP+C47+3tnJibydzRg6JdjoiIiIh4FLR7uRfW76WovJ7bF46Pdiki\nIiIiEkJBuxdzzrHkzW2MGTyA86YNj3Y5IiIiIhJCQbsXW7GjnLW+Sm49fZz6ZouIiIj0MAravdi9\nb21nUEoCHz8xO9qliIiIiEgbCtq9VGFpDf/+oJTrThpNckJstMsRERERkTbCCtpm9pSZXWxmCuY9\nxH1vbScxLobrTx4d7VJEREREpB3hBuc/AFcDW83sLjObHMGa5BhKqxt46v1iPjE3m8GpidEuR0RE\nRETaEVbQds792zl3DXAisBP4t5m9a2Y3mVl8JAuUj3r0vV00BwLcevq4aJciIiIiIkcQdlcQMxsM\n3AjcCqwGfkMweL8ckcqkXXVNLTy6bBfnTxvO2CEp0S5HRERERI4gLpydzOxpYDLwKHCpc26Pt+kJ\nM8uPVHHyUa3Lrd++UK3ZIiIiIj1ZWEEb+K1z7rX2Njjn8rqwHjmKFn+A+97eztzRA7XcuoiIiEgP\nF27XkWlmltn6wMwGmtlnI1STHMGLG/ZRVF7PbeqbLSIiItLjhRu0b3POVbQ+cM4dBG6LTEnSntbl\n1scOSdFy6yIiIiK9QLhBO9bMDq3xbWaxQEJkSpL2tC63fstpY7XcuoiIiEgvEG4f7RcIDny8x3v8\nae856Saty61/Yq6WWxcRERHpDcIN2t8gGK7/n/f4ZeC+iFQkH9G63PoXz51IUryWWxcRERHpDcIK\n2s65APBH7ybdrHW59etO0nLrIiIiIr1FuPNoTwTuBKYBSa3PO+c0/UWEtS63/h95Wm5dREREpDcJ\ndzDkgwRbs1uAs4BHgMciVZR86JF3tdy6iIiISG8UbtBOds69Aphzbpdz7ofAxZErS0DLrYuIiIj0\nZuEOhmw0sxhgq5l9HigGUiNXlkBwufXK+mZuXzg+2qWIiIiISAeF26L9BWAA8J/AXOBa4IZjHWRm\nF5jZZjMrNLNvtrM918xeM7PVZlZgZhd5z48xs3ozW+Pd/hT+S+obDl9ufWC0yxERERGRDjpmi7a3\nOM2nnHNfBWqAm8I5sXfc3cB5gA9YaWbPOuc2huz2XeBJ59wfzWwa8Bwwxtu2zTl3QtivpI9pXW79\nOxdNi3YpIiIiItIJx2zRds75gdM6ce75QKFzbrtzrgl4HLi87emBdO9+BlDSiev0OVpuXURERKT3\nC7eP9mozexb4P6C29Unn3FNHOSYLKAp57AMWtNnnh8BLZnYHkAKcG7JtrJmtBqqA7zrn3mp7ATO7\nHbgdIDc3N8yX0vO1Lrf+k4/N0HLrIiIiIr1UuH20k4ADwNnApd7tki64/lXAQ865bOAi4FFv0OUe\nINc5Nwf4MrDUzNLbHuycW+Kcy3PO5Q0dOrQLyukZtNy6iIiISO8X7sqQYfXLbqMYyAl5nO09F+oW\n4ALvGu+ZWRIwxDlXCjR6z68ys23AJCC/E3X0KoWl1VpuXURERKQPCHdlyAcJ9qc+jHPu5qMcthKY\naGZjCQbsxcDVbfbZDZwDPGRmUwm2nJeZ2VCg3DnnN7NxwERgezi19nb3vbVDy62LiIiI9AHh9tH+\nZ8j9JOAKjjFw0TnX4s25/SIQCzzgnNtgZj8C8p1zzwJfAe41sy8RDPI3OuecmS0EfmRmzUAA+Ixz\nrrxDr6wXal1u/ZPztNy6iIiISG8XbteRv4U+NrO/AG+HcdxzBKfsC33u+yH3NwKnHuF6f2v7fF/X\nutz6LadpuXURERGR3i7cwZBtTQSGdWUh/V3rcuuLpo3QcusiIiIifUC4fbSrObyP9l7gGxGpqJ96\ncmURlfXN3LZQrdkiIiIifUG4XUfSIl1If9biD3D/Ozu03LqIiIhIHxJW1xEzu8LMMkIeZ5rZxyJX\nVv/Sutz67WrNFhEREekzwu2j/QPnXGXrA+dcBfCDyJTUv4Qut37uVC23LiIiItJXhBu029sv3KkB\n5Shal1u/5bSxWm5dREREpA8JN2jnm9kvzWy8d/slsCqShfUXWm5dREREpG8KN2jfATQBTwCPAw3A\n5yJVVH/Rutz69SeP1nLrIiIiIn1MuLOO1ALfjHAt/U7rcuvXnzwm2qWIiIiISBcLd9aRl80sM+Tx\nQDN7MXJl9X2ty63/R142g1ISol2OiIiIiHSxcLuODPFmGgHAOXcQrQx5XLTcuoiIiEjfFm7QDphZ\nbusDMxvD4StFSgdouXURERGRvi/cKfq+A7xtZm8ABpwO3B6xqvo4LbcuIiIi0veFOxjyBTPLIxiu\nVwPPAPWRLKyv8gcc97+zgzwtty4iIiLSp4UVtM3sVuALQDawBjgJeA84O3Kl9U0f7KmiqLyeL507\nKdqliIiIiEgEhdtH+wvAPGCXc+4sYA5QcfRDpD1rfcG3Ta3ZIiIiIn1buEG7wTnXAGBmic65TcDk\nyJXVdxUUVZI5IJ7cQQOiXYqIiIiIRFC4gyF93jzazwAvm9lBYFfkyuq71voqmJmVgZlFuxQRERER\niaBwB0Ne4d39oZm9BmQAL0Ssqj6qvsnP1tIazp06PNqliIiIiEiEhduifYhz7o1IFNIfbCipxB9w\nzMrOiHYpIiIiIhJh4fbRli6w1lcJwOyczGPsKSIiIiK9nYJ2NyrwVTA8PZHh6UnRLkVEREREIkxB\nuxsV+CqZla3WbBEREZH+QEG7m1TWN7Njfy2z1T9bREREpF9Q0O4m67z+2WrRFhEREekfFLS7SeuK\nkJpxRERERKR/UNDuJgW+CkYPHkDmgIRolyIiIiIi3UBBu5toIKSIiIhI/6Kg3Q1KqxvYU9mggZAi\nIiIi/YiCdjcoKNJASBEREZH+RkG7GxT4KogxmJGVHu1SRERERKSbKGh3g7W+SiYOS2NAQly0SxER\nERGRbqKgHWHOOQp8FczOUf9sERERkf5EQTvCfAfrOVjXrP7ZIiIiIv2MgnaEtS5UM1tBW0RERKRf\nUdCOsAJfJQmxMUwekRbtUkRERESkGyloR9jaogqmjkonIU5vtYiIiEh/ovQXQf6AY31xpRaqERER\nEemHFLQjaHtZDbVNfg2EFBEREemHFLQjaK0vuCKkWrRFRERE+h8F7Qgq8FWQkhDLuKGp0S5FRERE\nRLqZgnYErfVVMiMrg9gYi3YpIiIiItLNFLQjpKklwAclVczOUf9sERERkf5IQTtCNu+tpskfYJb6\nZ4uIiIj0SwraEaIVIUVERET6NwXtCCnwVTBwQDzZA5OjXYqIiIiIRIGCdoQU+CqZlZ2JmQZCioiI\niPRHCtoRUNfUwpZ91Zo/W0RERKQfU9COgA0lVQQcWhFSREREpB9T0I6AtUXBgZCzctSiLSIiItJf\nKWhHQIGvkpEZSQxLS4p2KSIiIiISJREN2mZ2gZltNrNCM/tmO9tzzew1M1ttZgVmdlHItm95x202\ns0WRrLOrFfgqNH+2iIiISD8XsaBtZrHA3cCFwDTgKjOb1ma37wJPOufmAIuBP3jHTvMeTwcuAP7g\nna/Hq6xrZueBOvXPFhEREennItmiPR8odM5td841AY8Dl7fZxwHp3v0MoMS7fznwuHOu0Tm3Ayj0\nztfjFRRroRoRERERiWzQzgKKQh77vOdC/RC41sx8wHPAHR04tkcq8FUCMFNdR0RERET6tWgPhrwK\neMg5lw1cBDxqZmHXZGa3m1m+meWXlZVFrMiOWFtUwdghKWQkx0e7FBERERGJokgG7WIgJ+Rxtvdc\nqFuAJwGcc+8BScCQMI/FObfEOZfnnMsbOnRoF5beecEVIdWaLSIiItLfRTJorwQmmtlYM0sgOLjx\n2Tb77AbOATCzqQSDdpm332IzSzSzscBEYEUEa+0SpVUN7K1q0EBIERERESEuUid2zrWY2eeBF4FY\n4AHn3AYz+xGQ75x7FvgKcK+ZfYngwMgbnXMO2GBmTwIbgRbgc845f6Rq7Sprvf7ZWnpdRERERCIW\ntAGcc88RHOQY+tz3Q+5vBE49wrE/BX4ayfq6WoGvgtgYY/ooBW0RERGR/i7agyH7lLW+SiYOSyU5\noVdM+S0iIiIiEaSg3UWccxT4KjR/toiIiIgACtpdpqi8noq6ZmblqNuIiIiIiChod5m1Pq0IKSIi\nIiIfUtDuIgW+ChLiYpg8Ii3apYiIiIhID6Cg3UXW+iqZNjKd+Fi9pSIiIiKioN0l/AHH+uJKzZ8t\nIiIiIocoaHeBbWU11DX5tSKkiIiIiByioN0F1hZ5AyE144iIiIiIeBS0u0CBr5LUxDjGDUmNdiki\nIiIi0kMoaHeBAl8FM7LSiYmxaJciIiIiIj2EgvZxamoJ8MGeas2fLSIiIiKHUdA+Tpv2VtHkD2gg\npIiIiIgcRkH7OK31VQIwS1P7iYiIiEgIBe3jVFBUwaCUBLIHJke7FBERERHpQRS0j1OBr5JZ2RmY\naSCkiIiIiHxIQfs41Da2sLW0Wv2zRUREROQjFLSPw/riSgIOLb0uIiIiIh+hoH0cCg4NhFSLtoiI\niIgcTkH7OKz1VTAqI4mhaYnRLkVEREREehgF7eMQHAip1mwRERER+SgF7U46WNvE7vI6ZuWof7aI\niIiIfJSCdicVFAf7Z2vpdRERERFpj4J2JxUUVQAwI0st2iIiIiLyUQranbTWV8m4ISlkJMdHuxQR\nERER6YEUtDupwFfBLM2fLSIiIiJHoKDdCXsrGyitbtSMIyIiIiJyRAranbDWF+yfPVszjoiIiIjI\nEShod0KBr4LYGGPaSAVtEREREWmfgnYnFPgqmTQ8jeSE2GiXIiIiIiI9lIJ2BznnKPBVMlsDIUVE\nRETkKBS0O2jXgToq65s1EFJEREREjkpBu4NaB0Jqaj8RERERORoF7Q4q8FWSGBfD5BFp0S5FRERE\nRHowBe0OKvBVMH1UOvGxeutERERE5MiUFjugxR9gfXGV+meLiIiIyDEpaHdAYVkN9c1+LVQjIiIi\nIsekoN0BBUWVAGrRFhEREZFjUtDugLW+CtIS4xg7OCXapYiIiIhID6eg3QEFvkpmZmcQE2PRLkVE\nREREejgF7TA1tvjZtFcDIUVEREQkPAraYfpgTzXNfqel10VEREQkLAraYSpoXREyRy3aIiIiInJs\nCtphWltUyZDUBEZlJEW7FBERERHpBRS0w1Tgq2BWdiZmGggpIiIiIsemoB2GmsYWCstqmKX+2SIi\nIiISJgXtMKwvrsQ5mK0ZR0REREQkTAraYTg0EFIt2iIiIiISJgXtMKz1VZKVmczg1MRolyIiIiIi\nvYSCdhgKfBXMzlFrtoiIiIiET0H7GMprmygqr9eKkCIiIiLSIREN2mZ2gZltNrNCM/tmO9t/ZWZr\nvNsWM6sI2eYP2fZsJOs8GvXPFhEREZHOiIvUic0sFrgbOA/wASvN7Fnn3MbWfZxzXwrZ/w5gTsgp\n6p1zJ0SqvnAV+Coxg5lZCtoiIiIiEr5ItmjPBwqdc9udc03A48DlR9n/KuAvEaynUwp8FYwbkkJa\nUny0SxERERGRXiSSQTsLKAp57POe+wgzGw2MBV4NeTrJzPLNbJmZfSxyZR6Zc461vkrNny0iIiIi\nHRaxriMdtBj4q3POH/LcaOdcsZmNA141s3XOuW2hB5nZ7cDtALm5uV1e1N6qBsqqG9U/W0REREQ6\nLJIt2sVATsjjbO+59iymTbcR51yx93M78DqH999u3WeJcy7POZc3dOjQrqj5MGuLKgGYlaMWbRER\nERHpmEgG7ZXARDMba2YJBMP0R2YPMbMpwEDgvZDnBppZond/CHAqsLHtsZFW4KsgLsaYNjK9uy8t\nIiIiIr1cxLqOOOdazOzzwItALPCAc26Dmf0IyHfOtYbuxcDjzjkXcvhU4B4zCxD8Y+Cu0NlKukuB\nr5LJI9JIio/t7kuLiIiISC9nh+fb3isvL8/l5+d36TlLqxo4UNvEVLVoi4iIiIjHzFY55/KOtV9P\nGQzZIw1LT2JYelK0yxARERGRXkhLsIuIiIiIRICCtoiIiIhIBChoi4iIiIhEgIK2iIiIiEgEKGiL\niIiIiESAgraIiIiISAQoaIuIiIiI/P/27j/kzrqM4/j70zbLNtGJS0SXpgWVZluKsDSRougHocGy\nLMX6pwIjpX9mv8iEIKIfYoRaNJi4mr82k/5JE1n5R7pp89fWzMRoYznCVi5Ic179cb5jj2vP4xTP\nc9/34f2Ch+c+3/N97nOd5+J6zvXc53ufewxstCVJkqQxsNGWJEmSxsBGW5IkSRqDVFXXMbwqkjwD\nbO06Dh20o4C/dx2EDpr5GhbzNSzma1jM17CMK1/HV9Wil5o0dwwP3JWtVXV610Ho4CTZaL6Gw3wN\ni/kaFvM1LOZrWLrOl0tHJEmSpDGw0ZYkSZLGYJIa7Z90HYBeFvM1LOZrWMzXsJivYTFfw9Jpvibm\nZEhJkiSpTybpiLYkSZLUGxPRaCf5YJKtSR5PcnnX8WhmSZ5M8nCSTUk2dh2PXizJyiQ7kzwyZezI\nJHcm+VP7vrDLGLXPNPm6Isn2VmObkny4yxi1T5LFSe5OsjnJo0kubePWWA/NkC9rrIeSvC7JfUke\nbPn6Vht/U5J7W594Y5JDZi2moS8dSTIHeAx4P7AN2ABcUFWbOw1M00ryJHB6Vfk5pD2U5GxgN3B9\nVZ3Sxr4LPF1V32n/zC6sqhVdxqmRafJ1BbC7qr7XZWz6f0mOAY6pqgeSHAbcD5wHfAZrrHdmyNf5\nWGO9kyTA/KranWQecA9wKfBlYG1VrUlyLfBgVV0zGzFNwhHtM4DHq+qJqnoOWAOc23FM0mBV1W+B\np/cbPhdY1bZXMXqhUQ9Mky/1VFXtqKoH2vYzwBbgWKyxXpohX+qhGtndbs5rXwW8F7iljc9qfU1C\no30s8Ncpt7dhEfRdAXckuT/J57oORgfl6Kra0bb/BhzdZTA6KF9M8lBbWuIyhB5KcgKwFLgXa6z3\n9hhokCQAAAQjSURBVMsXWGO9lGROkk3ATuBO4M/Arqp6vk2Z1T5xEhptDc9ZVfUu4EPAJe2tbw1E\njdabDXvN2eS7BjgJWALsAL7fbTjaX5IFwK3AZVX1r6n3WWP9c4B8WWM9VVV7qmoJcByjVQ9v7TKe\nSWi0twOLp9w+ro2pp6pqe/u+E1jHqBDUb0+1tYp71yzu7DgezaCqnmovNi8AP8Ua65W2dvRWYHVV\nrW3D1lhPHShf1lj/VdUu4G5gGXBEkrntrlntEyeh0d4AvKWdUXoI8Eng9o5j0jSSzG8nlJBkPvAB\n4JGZf0o9cDtwcdu+GPhlh7HoJext2JqPYY31RjtZ62fAlqr6wZS7rLEemi5f1lg/JVmU5Ii2fSij\nD8rYwqjhXt6mzWp9Df5TRwDax+pcBcwBVlbVtzsOSdNIciKjo9gAc4Gfm69+SfIL4BzgKOAp4JvA\nbcBNwBuBvwDnV5Un4PXANPk6h9Fb2gU8CXx+yvpfdSjJWcDvgIeBF9rwVxmt+7XGemaGfF2ANdY7\nSU5ldLLjHEYHk2+qqitb77EGOBL4A3BhVT07KzFNQqMtSZIk9c0kLB2RJEmSesdGW5IkSRoDG21J\nkiRpDGy0JUmSpDGw0ZYkSZLGwEZbkgRAknOS/KrrOCRpUthoS5IkSWNgoy1JA5PkwiT3JdmU5Lok\nc5LsTvLDJI8muSvJojZ3SZLfJ3koybokC9v4m5P8JsmDSR5IclLb/YIktyT5Y5LV7cp4JDktyfok\n9yf59ZTLhX8pyea2/zWd/EIkqadstCVpQJK8DfgEcGZVLQH2AJ8G5gMbq+pkYD2jK0QCXA+sqKpT\nGV3dbu/4auDHVfVO4N3A3qvaLQUuA94OnAicmWQe8CNgeVWdBqwE9l7R9XJgadv/F8bzrCVpmOZ2\nHYAk6WV5H3AasKEdbD4U2Mno8tA3tjk3AGuTHA4cUVXr2/gq4OYkhwHHVtU6gKr6D0Db331Vta3d\n3gScAOwCTgHubHPmsK8xfwhYneQ24LbxPGVJGiYbbUkalgCrquorLxpMvrHfvHqF+392yvYeRq8T\nAR6tqmUHmP8R4Gzgo8DXkryjqp5/hY8tSRPFpSOSNCx3AcuTvAEgyZFJjmf093x5m/Mp4J6q+ifw\njyTvaeMXAeur6hlgW5Lz2j5em+T1MzzmVmBRkmVt/rwkJyd5DbC4qu4GVgCHAwte1WcrSQPmEW1J\nGpCq2pzk68AdrdH9L3AJ8G/gjHbfTkbruAEuBq5tjfQTwGfb+EXAdUmubPv4+AyP+VyS5cDVbTnK\nXOAq4DHghjYW4Oqq2vXqPmNJGq5UvdJ3FyVJfZFkd1V5NFmSesSlI5IkSdIYeERbkiRJGgOPaEuS\nJEljYKMtSZIkjYGNtiRJkjQGNtqSJEnSGNhoS5IkSWNgoy1JkiSNwf8ANGQGg/E2fvoAAAAASUVO\nRK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88325cb00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams[\"figure.figsize\"] = [12, 5]\n",
    "ax = plt.gca()\n",
    "ax.set_xlim([0, EPOCHS+1])\n",
    "ax.set_xlabel('epoches')\n",
    "ax.set_ylabel('accuracy')\n",
    "\n",
    "plt.title('Traning accuracy vs. Number of epoches')\n",
    "x = range(1, EPOCHS + 1)\n",
    "plt.plot(x, accuracy)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluate the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from ./lenet\n",
      "Test Accuracy = 0.935\n"
     ]
    }
   ],
   "source": [
    "# Check model accuray by using test data set\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "\n",
    "    test_accuracy = evaluate(X_test, y_test)\n",
    "    print(\"Test Accuracy = {:.3f}\".format(test_accuracy))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "\n",
    "## Step 3: Test a Model on New Images\n",
    "\n",
    "To give yourself more insight into how your model is working, download at least five pictures of German traffic signs from the web and use your model to predict the traffic sign type.\n",
    "\n",
    "You may find `signnames.csv` useful as it contains mappings from the class id (integer) to the actual sign name."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load and Output the Images\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAM0AAABlCAYAAADu82fqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb89276af98>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOYAAABlCAYAAABZTp/NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nnizEK6bHk4V4xfR4shCvmB5PFvL/5gQzVvt1xi4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88f47c6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAANUAAABlCAYAAADaiqb2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892a148d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAMQAAABlCAYAAAAS5kwgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBIEIBFIEgQgEUgSBCARSBIEIBFL8PyLGyw/O88FiAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892a399b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAJEAAABlCAYAAABN2wEsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\npyu22yAmAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892a4db38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Load the images and plot them here.\n",
    "### Feel free to use as many code cells as needed.\n",
    "from PIL import Image\n",
    "\n",
    "imgs  = ['new_images/' + img for img in os.listdir('new_images')]\n",
    "processed_imgs = []\n",
    "\n",
    "for img in imgs:\n",
    "    image = Image.open(img)\n",
    "    # Covert format from whatever into RGB \n",
    "    image = image.convert('RGB')\n",
    "    image = image.resize((32, 32))    \n",
    "    image = np.array(list(image.getdata()), dtype='uint8')\n",
    "    # Convert array from 1024x3 to 32x32x3\n",
    "    image = np.reshape(image, (32, 32, 3))\n",
    "    processed_imgs.append(image)\n",
    "\n",
    "images = np.array(processed_imgs, dtype='uint8')\n",
    "\n",
    "for i, img in enumerate(images):\n",
    "    plt.figure(figsize=(1,1))\n",
    "    plt.imshow(img)\n",
    "    plt.title(imgs[i])\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### Predict the Sign Type for Each Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from ./lenet\n"
     ]
    }
   ],
   "source": [
    "### Run the predictions here and use the model to output the prediction for each image.\n",
    "### Make sure to pre-process the images with the same pre-processing pipeline used earlier.\n",
    "### Feel free to use as many code cells as needed.\n",
    "\n",
    "# Step 1: pre-process the image\n",
    "X_images = normalize_images(images)\n",
    "y_correct = [4, 6,7,21,14]\n",
    "y_correct = np.array(y_correct, dtype='uint8')\n",
    "\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))\n",
    "    y_pred = sess.run(logits, feed_dict={x: X_images})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Analyze Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "with tf.Session() as sess:\n",
    "    logits = tf.placeholder('float', [None, 43])\n",
    "    predict_labels = tf.argmax(logits, 1)\n",
    "    predict_labels = sess.run([predict_labels], feed_dict={logits: y_pred})\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "4"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Calculate the accuracy for these 5 new images. \n",
    "### For example, if the model predicted 1 out of 5 signs correctly, it's 20% accurate on these new images.\n",
    "predict_labels[0][0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The correct labels are y_correct = [4, 6,7,21,14]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 4,  6,  7, 21, 14], dtype=uint8)"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_correct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuray is 0.600\n"
     ]
    }
   ],
   "source": [
    "#Calculate accuray \n",
    "num_diff = 0\n",
    "for i in range(5):\n",
    "    if y_correct[i] != predict_labels[0][i]:\n",
    "        num_diff += 1\n",
    "        \n",
    "print(\"Accuray is %.3f\" % ((5-num_diff)/5))\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Output Top 5 Softmax Probabilities For Each Image Found on the Web"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For each of the new images, print out the model's softmax probabilities to show the **certainty** of the model's predictions (limit the output to the top 5 probabilities for each image). [`tf.nn.top_k`](https://www.tensorflow.org/versions/r0.12/api_docs/python/nn.html#top_k) could prove helpful here. \n",
    "\n",
    "The example below demonstrates how tf.nn.top_k can be used to find the top k predictions for each image.\n",
    "\n",
    "`tf.nn.top_k` will return the values and indices (class ids) of the top k predictions. So if k=3, for each sign, it'll return the 3 largest probabilities (out of a possible 43) and the correspoding class ids.\n",
    "\n",
    "Take this numpy array as an example. The values in the array represent predictions. The array contains softmax probabilities for five candidate images with six possible classes. `tk.nn.top_k` is used to choose the three classes with the highest probability:\n",
    "\n",
    "```\n",
    "# (5, 6) array\n",
    "a = np.array([[ 0.24879643,  0.07032244,  0.12641572,  0.34763842,  0.07893497,\n",
    "         0.12789202],\n",
    "       [ 0.28086119,  0.27569815,  0.08594638,  0.0178669 ,  0.18063401,\n",
    "         0.15899337],\n",
    "       [ 0.26076848,  0.23664738,  0.08020603,  0.07001922,  0.1134371 ,\n",
    "         0.23892179],\n",
    "       [ 0.11943333,  0.29198961,  0.02605103,  0.26234032,  0.1351348 ,\n",
    "         0.16505091],\n",
    "       [ 0.09561176,  0.34396535,  0.0643941 ,  0.16240774,  0.24206137,\n",
    "         0.09155967]])\n",
    "```\n",
    "\n",
    "Running it through `sess.run(tf.nn.top_k(tf.constant(a), k=3))` produces:\n",
    "\n",
    "```\n",
    "TopKV2(values=array([[ 0.34763842,  0.24879643,  0.12789202],\n",
    "       [ 0.28086119,  0.27569815,  0.18063401],\n",
    "       [ 0.26076848,  0.23892179,  0.23664738],\n",
    "       [ 0.29198961,  0.26234032,  0.16505091],\n",
    "       [ 0.34396535,  0.24206137,  0.16240774]]), indices=array([[3, 0, 5],\n",
    "       [0, 1, 4],\n",
    "       [0, 5, 1],\n",
    "       [1, 3, 5],\n",
    "       [1, 4, 3]], dtype=int32))\n",
    "```\n",
    "\n",
    "Looking just at the first row we get `[ 0.34763842,  0.24879643,  0.12789202]`, you can confirm these are the 3 largest probabilities in `a`. You'll also notice `[3, 0, 5]` are the corresponding indices."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [],
   "source": [
    "with tf.Session() as sess:\n",
    "    logits = tf.placeholder('float', [None, 43])\n",
    "    y_softmax = tf.nn.softmax(logits)\n",
    "    top_5_prob, top_5_label = tf.nn.top_k(y_softmax, k=len(y_correct))\n",
    "    top_5_probs, top_5_labels = sess.run([top_5_prob, top_5_label], feed_dict={logits: y_pred})\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Load signnames.csv to map label number to sign string\n",
    "label_map = {}\n",
    "with open('signnames.csv', 'r') as f:\n",
    "    line0 = True\n",
    "    for line in f:\n",
    "        # Ignore first line\n",
    "        if line0:\n",
    "            line0 = False\n",
    "            continue\n",
    "        # Parse CSV file into a map\n",
    "        label_idx, label_name = line.split(',')\n",
    "        label_idx = int(label_int)\n",
    "        label_map[label_idx] = label_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def show_pred_results(image, top_k_val, top_k_idx):\n",
    "    top_k_pred = [label_map[idx] for idx in top_k_idx]\n",
    "    plt.imshow(image)\n",
    "    plt.show()\n",
    "    \n",
    "    for i in range(5):\n",
    "        print('%s: %.2f%%' % (top_k_pred[i].replace('\\n', ''), top_k_val[i] * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATUAAAEyCAYAAACbGke8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88c9f8ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Speed limit (70km/h): 99.56%\n",
      "General caution: 0.44%\n",
      "Keep left: 0.00%\n",
      "Go straight or left: 0.00%\n",
      "Wild animals crossing: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATUAAAEyCAYAAACbGke8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88c9dda90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "End of no passing: 56.41%\n",
      "End of all speed and passing limits: 43.58%\n",
      "End of speed limit (80km/h): 0.00%\n",
      "Speed limit (60km/h): 0.00%\n",
      "End of no passing by vehicles over 3.5 metric tons: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATUAAAEyCAYAAACbGke8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb892b31b38>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Speed limit (100km/h): 100.00%\n",
      "Speed limit (120km/h): 0.00%\n",
      "Speed limit (50km/h): 0.00%\n",
      "Speed limit (30km/h): 0.00%\n",
      "Vehicles over 3.5 metric tons prohibited: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATUAAAEyCAYAAACbGke8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDKWqecgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88c8b7b00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dangerous curve to the right: 99.45%\n",
      "Children crossing: 0.47%\n",
      "Beware of ice/snow: 0.07%\n",
      "Right-of-way at the next intersection: 0.00%\n",
      "Slippery road: 0.00%\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATUAAAEyCAYAAACbGke8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nFIWiBqAo/wdctnVnYSCJEAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb88ca0ab00>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Stop: 100.00%\n",
      "Bicycles crossing: 0.00%\n",
      "Speed limit (30km/h): 0.00%\n",
      "Priority road: 0.00%\n",
      "Speed limit (80km/h): 0.00%\n"
     ]
    }
   ],
   "source": [
    "for i in range(5):\n",
    "    show_pred_results(images[i], top_k_vals[i], top_k_idxs[i])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.2"
  },
  "widgets": {
   "state": {},
   "version": "1.1.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}