Traffic_Sign_Classifier.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "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": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "## Step 0: Load The Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "# Load pickled data\n",
    "import pickle\n",
    "\n",
    "# TODO: Fill this in based on where you saved the training and testing data\n",
    "\n",
    "training_file = \"traffic-signs-data/train.p\"\n",
    "validation_file = \"traffic-signs-data/valid.p\"\n",
    "testing_file = \"traffic-signs-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']\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAACRCAYAAADTnUPWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwVAS2IJuMBgMJcH/B8gcM2iIoyZ1AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd4ec4fbe0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(104397, 32, 32, 3)\n",
      "(104397,)\n"
     ]
    }
   ],
   "source": [
    "# create more fake data\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "from PIL import ImageFilter\n",
    "\n",
    "X_train_blured = []\n",
    "for sign in X_train:\n",
    "    im = Image.fromarray(sign).filter(ImageFilter.SMOOTH)\n",
    "    X_train_blured.append(np.asarray(im))\n",
    "    \n",
    "X_train_sharpen = []\n",
    "for sign in X_train:\n",
    "    im = Image.fromarray(sign).filter(ImageFilter.SHARPEN)\n",
    "    X_train_sharpen.append(np.asarray(im))\n",
    "\n",
    "fig = plt.figure()\n",
    "a=fig.add_subplot(1,3,1)\n",
    "imgplot = plt.imshow(X_train[4].squeeze())\n",
    "a.set_title('Before')\n",
    "a=fig.add_subplot(1,3,2)\n",
    "imgplot = plt.imshow(X_train_blured[4].squeeze())\n",
    "a.set_title('After blured')\n",
    "a=fig.add_subplot(1,3,3)\n",
    "imgplot = plt.imshow(X_train_sharpen[4].squeeze())\n",
    "a.set_title('After sharpen')\n",
    "\n",
    "plt.show()\n",
    "\n",
    "X_train = np.concatenate((X_train, X_train_blured))\n",
    "X_train = np.concatenate((X_train, X_train_sharpen))\n",
    "y_train = np.concatenate((y_train, y_train, y_train))\n",
    "print(X_train.shape)\n",
    "print(y_train.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "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": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Provide a Basic Summary of the Data Set Using Python, Numpy and/or Pandas"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of training examples = 104397\n",
      "Number of testing examples = 12630\n",
      "Number of validation examples = 4410\n",
      "Image data shape = (32, 32, 1)\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",
    "import pandas as pd\n",
    "from collections import Counter\n",
    "\n",
    "# read signnames\n",
    "import csv\n",
    "reader = csv.reader(open('signnames.csv'))\n",
    "# skip header\n",
    "next(reader, None)\n",
    "signnames = {}\n",
    "for row in reader:    \n",
    "    signnames[row[0]] = row[1]\n",
    "\n",
    "# TODO: Number of training examples\n",
    "n_train = len(y_train)\n",
    "\n",
    "# TODO: Number of validation examples\n",
    "n_validation = len(y_valid)\n",
    "\n",
    "# TODO: Number of testing examples.\n",
    "n_test = len(y_test)\n",
    "\n",
    "# TODO: What's the shape of an traffic sign image?\n",
    "#image_shape = X_train[0].shape\n",
    "image_shape = (32,32,1)\n",
    "\n",
    "# TODO: How many unique classes/labels there are in the dataset.\n",
    "n_classes = len(signnames)\n",
    "\n",
    "print(\"Number of training examples =\", n_train)\n",
    "print(\"Number of testing examples =\", n_test)\n",
    "print(\"Number of validation examples =\", n_validation)\n",
    "print(\"Image data shape =\", image_shape)\n",
    "print(\"Number of classes =\", n_classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd409ab8d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd40aa5ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    " \n",
    "# data to plot\n",
    "n_groups = 43\n",
    "means_frank = list(Counter(y_train).values())\n",
    "means_guido = list(Counter(y_test).values())\n",
    "\n",
    "#n_groups = 4\n",
    "#means_frank = [1,2,5,6]\n",
    "#means_guido = [1,2,7,3]\n",
    " \n",
    "# create plot\n",
    "fig, ax = plt.subplots()\n",
    "index = np.arange(n_groups)\n",
    "bar_width = 0.35\n",
    "opacity = 0.8\n",
    " \n",
    "rects1 = plt.bar(index, means_frank, bar_width,\n",
    "                 alpha=opacity,\n",
    "                 color='b',\n",
    "                 label='y_train')\n",
    " \n",
    "rects2 = plt.bar(index + bar_width, means_guido, bar_width,\n",
    "                 alpha=opacity,\n",
    "                 color='g',\n",
    "                 label='y_test')\n",
    " \n",
    "plt.xlabel('sign')\n",
    "plt.ylabel('exaples')\n",
    "plt.title('Exeples of signs')\n",
    "plt.xticks(index + bar_width, list(Counter(y_test).keys()))\n",
    "plt.legend()\n",
    " \n",
    "plt.tight_layout()\n",
    "plt.figure(figsize=(25,25))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Include an exploratory visualization of the dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "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": 6,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Speed limit (120km/h)\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAADHCAYAAAAJSqg8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd770819e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Data exploration visualization code goes here.\n",
    "### Feel free to use as many code cells as needed.\n",
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.image as mpimg\n",
    "%matplotlib inline\n",
    "\n",
    "# grayscale and normalize\n",
    "def normalizeImg(img):    \n",
    "    gray = np.dot(img.squeeze()[...,:3], [0.299, 0.587, 0.114])\n",
    "    normalize = (gray - 128) / 128   \n",
    "    \n",
    "    return normalize.reshape(32,32,1)\n",
    "\n",
    "# make all gray\n",
    "X_train_gray = []\n",
    "for img in X_train:\n",
    "    X_train_gray.append(normalizeImg(img))\n",
    "    \n",
    "X_valid_gray = []\n",
    "for img in X_valid:\n",
    "    X_valid_gray.append(normalizeImg(img))\n",
    "    \n",
    "X_test_gray = []\n",
    "for img in X_test:\n",
    "    X_test_gray.append(normalizeImg(img))\n",
    "      \n",
    "\n",
    "# show result\n",
    "index = random.randint(0, len(X_train))\n",
    "\n",
    "fig = plt.figure()\n",
    "a=fig.add_subplot(1,2,1)\n",
    "imgplot = plt.imshow(X_train[index].squeeze())\n",
    "a.set_title('Before')\n",
    "a=fig.add_subplot(1,2,2)\n",
    "imgplot = plt.imshow(X_train_gray[index].squeeze(), cmap = plt.get_cmap('gray'))\n",
    "a.set_title('After')\n",
    "\n",
    "\n",
    "print(signnames[str(y_train[index])])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "X_train = X_train_gray\n",
    "X_valid = X_valid_gray\n",
    "X_test = X_test_gray"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "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": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Pre-process the Data Set (normalization, grayscale, etc.)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "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": 8,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": 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",
    "### Feel free to use as many code cells as needed.\n",
    "from sklearn.utils import shuffle\n",
    "\n",
    "X_train, y_train = shuffle(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Model Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "### Define your architecture here.\n",
    "### Feel free to use as many code cells as needed.\n",
    "import tensorflow as tf\n",
    "\n",
    "from tensorflow.contrib.layers import flatten\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",
    "    # SOLUTION: Layer 1: Convolutional. Input = 32x32x3. Output = 28x28x6.\n",
    "    conv1_W = tf.Variable(tf.truncated_normal(shape=(5, 5, image_shape[2], 6), mean = mu, stddev = sigma))\n",
    "    conv1_b = tf.Variable(tf.zeros(6))\n",
    "    conv1   = tf.nn.conv2d(x, conv1_W, strides=[1, 1, 1, 1], padding='VALID') + conv1_b\n",
    "\n",
    "    # SOLUTION: Activation.\n",
    "    conv1 = tf.nn.relu(conv1)\n",
    "    \n",
    "    return1 = conv1\n",
    "\n",
    "    # SOLUTION: Pooling. Input = 28x28x6. Output = 14x14x6.\n",
    "    conv1 = tf.nn.max_pool(conv1, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n",
    "\n",
    "    # SOLUTION: Layer 2: Convolutional. Output = 10x10x16.\n",
    "    conv2_W = tf.Variable(tf.truncated_normal(shape=(5, 5, 6, 16), mean = mu, stddev = sigma))\n",
    "    conv2_b = tf.Variable(tf.zeros(16))\n",
    "    conv2   = tf.nn.conv2d(conv1, conv2_W, strides=[1, 1, 1, 1], padding='VALID') + conv2_b\n",
    "    \n",
    "    # SOLUTION: Activation.\n",
    "    conv2 = tf.nn.relu(conv2)\n",
    "    return2 = conv2\n",
    "\n",
    "    # SOLUTION: Pooling. Input = 10x10x16. Output = 5x5x16.\n",
    "    conv2 = tf.nn.max_pool(conv2, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='VALID')\n",
    "\n",
    "    # SOLUTION: Flatten. Input = 5x5x16. Output = 400.\n",
    "    fc0   = flatten(conv2)\n",
    "    \n",
    "    # SOLUTION: Layer 3: Fully Connected. Input = 400. Output = 120.\n",
    "    fc1_W = tf.Variable(tf.truncated_normal(shape=(400, 120), mean = mu, stddev = sigma))\n",
    "    fc1_b = tf.Variable(tf.zeros(120))\n",
    "    fc1   = tf.matmul(fc0, fc1_W) + fc1_b\n",
    "    \n",
    "    # SOLUTION: Activation.\n",
    "    fc1    = tf.nn.relu(fc1)\n",
    "    fc1 = tf.nn.dropout(fc1, keep_prob_l3)\n",
    "\n",
    "    # SOLUTION: Layer 4: Fully Connected. Input = 120. Output = 84.    \n",
    "    fc2_W  = tf.Variable(tf.truncated_normal(shape=(120, 84), mean = mu, stddev = sigma))\n",
    "    fc2_b  = tf.Variable(tf.zeros(84))\n",
    "    fc2    = tf.matmul(fc1, fc2_W) + fc2_b\n",
    "    \n",
    "    # SOLUTION: Activation.\n",
    "    fc2    = tf.nn.relu(fc2)\n",
    "    fc2 = tf.nn.dropout(fc2, keep_prob_l4)\n",
    "\n",
    "    # SOLUTION: Layer 5: Fully Connected. Input = 84. Output = 43.\n",
    "    fc3_W  = tf.Variable(tf.truncated_normal(shape=(84, n_classes), 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, return1, return2\n",
    "\n",
    "\n",
    "# x is a placeholder for a batch of input images. y is a placeholder for a batch of output labels.\n",
    "x = tf.placeholder(tf.float32, (None, image_shape[0],image_shape[1],image_shape[2]))\n",
    "y = tf.placeholder(tf.int32, (None))\n",
    "one_hot_y = tf.one_hot(y, n_classes)\n",
    "\n",
    "keep_prob_l3 = tf.placeholder(tf.float32)\n",
    "keep_prob_l4 = tf.placeholder(tf.float32)\n",
    "\n",
    "# Training Pipeline\n",
    "rate = 0.001\n",
    "\n",
    "logits, layer1, layer2 = 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)\n",
    "\n",
    "# Model Evaluation\n",
    "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, keep_prob_l3:1, keep_prob_l4:1})\n",
    "        total_accuracy += (accuracy * len(batch_x))\n",
    "    return total_accuracy / num_examples"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Train, Validate and Test the Model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "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": 15,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training...\n",
      "\n",
      "EPOCH 1 ...\n",
      "Validation Accuracy = 0.812\n",
      "\n",
      "EPOCH 2 ...\n",
      "Validation Accuracy = 0.905\n",
      "\n",
      "EPOCH 3 ...\n",
      "Validation Accuracy = 0.924\n",
      "\n",
      "EPOCH 4 ...\n",
      "Validation Accuracy = 0.939\n",
      "\n",
      "EPOCH 5 ...\n",
      "Validation Accuracy = 0.944\n",
      "\n",
      "EPOCH 6 ...\n",
      "Validation Accuracy = 0.950\n",
      "\n",
      "EPOCH 7 ...\n",
      "Validation Accuracy = 0.952\n",
      "\n",
      "EPOCH 8 ...\n",
      "Validation Accuracy = 0.956\n",
      "\n",
      "EPOCH 9 ...\n",
      "Validation Accuracy = 0.958\n",
      "\n",
      "EPOCH 10 ...\n",
      "Validation Accuracy = 0.963\n",
      "\n",
      "Model saved\n",
      "Test Accuracy = 0.935\n"
     ]
    }
   ],
   "source": [
    "### Train your model here.\n",
    "### Calculate and report the accuracy on the training and validation set.\n",
    "### Once a final model architecture is selected, \n",
    "### the accuracy on the test set should be calculated and reported as well.\n",
    "### Feel free to use as many code cells as needed.\n",
    "\n",
    "#Train the Model\n",
    "EPOCHS = 10\n",
    "BATCH_SIZE = 128\n",
    "\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, keep_prob_l3:0.5, keep_prob_l4:0.5})\n",
    "            \n",
    "        validation_accuracy = evaluate(X_valid, y_valid)\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\")\n",
    "    \n",
    "    \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": {
    "deletable": true,
    "editable": true
   },
   "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": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Load and Output the Images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd27b64320>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nBDkAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd27a43fd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd27a74550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQiSKgl+IRFHwC5EoCn4hEkXBL0Si/F/n/TzjBZuUggAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd770df7b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nNwxPseA3DE+x4DcMT/l/E6WguHOj6t8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fbd37220080>"
      ]
     },
     "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",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from PIL import Image\n",
    "\n",
    "sign1 = np.asarray(Image.open('test/sign1.jpg').resize((32, 32)))\n",
    "sign2 = np.asarray(Image.open('test/sign2.jpg').resize((32, 32)))\n",
    "sign3 = np.asarray(Image.open('test/sign3.jpg').resize((32, 32)))\n",
    "sign4 = np.asarray(Image.open('test/sign4.jpg').resize((32, 32)))\n",
    "sign5 = np.asarray(Image.open('test/sign5.jpg').resize((32, 32)))\n",
    "\n",
    "plt.imshow(sign1)\n",
    "plt.show()\n",
    "plt.imshow(sign2)\n",
    "plt.show()\n",
    "plt.imshow(sign3)\n",
    "plt.show()\n",
    "plt.imshow(sign4)\n",
    "plt.show()\n",
    "plt.imshow(sign5)\n",
    "plt.show()\n",
    "\n",
    "sign1 = normalizeImg(sign1)\n",
    "sign2 = normalizeImg(sign2)\n",
    "sign3 = normalizeImg(sign3)\n",
    "sign4 = normalizeImg(sign4)\n",
    "sign5 = normalizeImg(sign5)\n",
    "\n",
    "test_images = [sign1,sign2,sign3,sign4,sign5]\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Predict the Sign Type for Each Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Priority road\n",
      "Yield\n",
      "Bumpy road\n",
      "Priority road\n",
      "Yield\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",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))    \n",
    "    \n",
    "    sess = tf.get_default_session()\n",
    "    accuracy = sess.run(tf.argmax(logits, 1), feed_dict={x: test_images, keep_prob_l3:1, keep_prob_l4:1})      \n",
    "    \n",
    "    print(signnames[str(accuracy[0])])\n",
    "    print(signnames[str(accuracy[1])])\n",
    "    print(signnames[str(accuracy[2])])\n",
    "    print(signnames[str(accuracy[3])])\n",
    "    print(signnames[str(accuracy[4])])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Analyze Performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80.0\n"
     ]
    }
   ],
   "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",
    "\n",
    "accuracy = 4 / len(test_images) * 100\n",
    "print(accuracy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Output Top 5 Softmax Probabilities For Each Image Found on the Web"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "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": 14,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.900278 Priority road\n",
      "0.0789745 Yield\n",
      "0.0103441 No passing\n",
      "0.00420427 Ahead only\n",
      "0.00329284 Turn left ahead\n",
      "---\n",
      "1.0 Yield\n",
      "0.0 Speed limit (20km/h)\n",
      "0.0 Speed limit (30km/h)\n",
      "0.0 Speed limit (50km/h)\n",
      "0.0 Speed limit (60km/h)\n",
      "---\n",
      "0.999871 Bumpy road\n",
      "0.000128809 Bicycles crossing\n",
      "1.89637e-08 Dangerous curve to the left\n",
      "6.70351e-09 Go straight or right\n",
      "5.05163e-09 Traffic signals\n",
      "---\n",
      "0.994981 Priority road\n",
      "0.00273771 Roundabout mandatory\n",
      "0.00179186 Speed limit (50km/h)\n",
      "0.000359271 Speed limit (30km/h)\n",
      "4.77334e-05 Keep right\n",
      "---\n",
      "0.979558 Yield\n",
      "0.00772195 Go straight or right\n",
      "0.00725152 Road work\n",
      "0.00421104 Ahead only\n",
      "0.000352258 Turn left ahead\n",
      "---\n"
     ]
    }
   ],
   "source": [
    "### Print out the top five softmax probabilities for the predictions on the German traffic sign images found on the web. \n",
    "### Feel free to use as many code cells as needed.\n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))    \n",
    "    \n",
    "    sess = tf.get_default_session()\n",
    "    accuracy = sess.run(tf.nn.softmax(logits), feed_dict={x: test_images, keep_prob_l3:1, keep_prob_l4:1})  \n",
    "    top_5 = sess.run(tf.nn.top_k(tf.constant(accuracy), k=5))\n",
    "\n",
    "    for result in [0,1,2,3,4]:\n",
    "        for prediction in [0,1,2,3,4]:\n",
    "            print(top_5.values[result][prediction], signnames[str(top_5.indices[result][prediction])])\n",
    "        print(\"---\")\n",
    "        "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "### Project Writeup\n",
    "\n",
    "Once you have completed the code implementation, document your results in a project writeup using this [template](https://github.com/udacity/CarND-Traffic-Sign-Classifier-Project/blob/master/writeup_template.md) as a guide. The writeup can be in a markdown or pdf file. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "> **Note**: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this 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."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "---\n",
    "\n",
    "## Step 4 (Optional): Visualize the Neural Network's State with Test Images\n",
    "\n",
    " This Section is not required to complete but acts as an additional excersise for understaning the output of a neural network's weights. While neural networks can be a great learning device they are often referred to as a black box. We can understand what the weights of a neural network look like better by plotting their feature maps. After successfully training your neural network you can see what it's feature maps look like by plotting the output of the network's weight layers in response to a test stimuli image. From these plotted feature maps, it's possible to see what characteristics of an image the network finds interesting. For a sign, maybe the inner network feature maps react with high activation to the sign's boundary outline or to the contrast in the sign's painted symbol.\n",
    "\n",
    " Provided for you below is the function code that allows you to get the visualization output of any tensorflow weight layer you want. The inputs to the function should be a stimuli image, one used during training or a new one you provided, and then the tensorflow variable name that represents the layer's state during the training process, for instance if you wanted to see what the [LeNet lab's](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) feature maps looked like for it's second convolutional layer you could enter conv2 as the tf_activation variable.\n",
    "\n",
    "For an example of what feature map outputs look like, check out NVIDIA's results in their paper [End-to-End Deep Learning for Self-Driving Cars](https://devblogs.nvidia.com/parallelforall/deep-learning-self-driving-cars/) in the section Visualization of internal CNN State. NVIDIA was able to show that their network's inner weights had high activations to road boundary lines by comparing feature maps from an image with a clear path to one without. Try experimenting with a similar test to show that your trained network's weights are looking for interesting features, whether it's looking at differences in feature maps from images with or without a sign, or even what feature maps look like in a trained network vs a completely untrained one on the same sign image.\n",
    "\n",
    "<figure>\n",
    " <img src=\"visualize_cnn.png\" width=\"380\" alt=\"Combined Image\" />\n",
    " <figcaption>\n",
    " <p></p> \n",
    " <p style=\"text-align: center;\"> Your output should look something like this (above)</p> \n",
    " </figcaption>\n",
    "</figure>\n",
    " <p></p> \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4CckUBj8hAQKg5+QQGHwExIo/w/kT15AINKxtwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f53e558b710>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Stop\n",
      "No passing\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAApQAAACICAYAAABZRHv1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f53ce7ad6d8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "### Visualize your network's feature maps here.\n",
    "### Feel free to use as many code cells as needed.\n",
    "\n",
    "# image_input: the test image being fed into the network to produce the feature maps\n",
    "# tf_activation: should be a tf variable name used during your training procedure that represents the calculated state of a specific weight layer\n",
    "# activation_min/max: can be used to view the activation contrast in more detail, by default matplot sets min and max to the actual min and max values of the output\n",
    "# plt_num: used to plot out multiple different weight feature map sets on the same block, just extend the plt number for each new feature map entry\n",
    "\n",
    "def outputFeatureMap(image_input, tf_activation, activation_min=-1, activation_max=-1 ,plt_num=1):\n",
    "    # Here make sure to preprocess your image_input in a way your network expects\n",
    "    # with size, normalization, ect if needed\n",
    "    # image_input =\n",
    "    # Note: x should be the same name as your network's tensorflow data placeholder variable\n",
    "    # If you get an error tf_activation is not defined it may be having trouble accessing the variable from inside a function\n",
    "    activation = tf_activation.eval(session=sess,feed_dict={x : image_input})\n",
    "    featuremaps = activation.shape[3]\n",
    "    plt.figure(plt_num, figsize=(15,15))\n",
    "    for featuremap in range(featuremaps):\n",
    "        plt.subplot(6,8, featuremap+1) # sets the number of feature maps to show on each row and column\n",
    "        plt.title('FeatureMap ' + str(featuremap)) # displays the feature map number\n",
    "        if activation_min != -1 & activation_max != -1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmin =activation_min, vmax=activation_max, cmap=\"gray\")\n",
    "        elif activation_max != -1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmax=activation_max, cmap=\"gray\")\n",
    "        elif activation_min !=-1:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", vmin=activation_min, cmap=\"gray\")\n",
    "        else:\n",
    "            plt.imshow(activation[0,:,:, featuremap], interpolation=\"nearest\", cmap=\"gray\")\n",
    "            \n",
    "# show result\n",
    "index = random.randint(0, len(X_test))\n",
    "image = test_images[0] #X_test[index]\n",
    "\n",
    "plt.imshow(image.squeeze(), cmap = plt.get_cmap('gray'))\n",
    "plt.show()\n",
    "#print(signnames[str(y_test[index])])\n",
    "print(\"Stop\")\n",
    "            \n",
    "with tf.Session() as sess:\n",
    "    saver.restore(sess, tf.train.latest_checkpoint('.'))    \n",
    "    \n",
    "    sess = tf.get_default_session()\n",
    "    accuracy = sess.run(tf.argmax(logits, 1), feed_dict={x: [image], keep_prob_l3:1, keep_prob_l4:1})      \n",
    "    \n",
    "    print(signnames[str(accuracy[0])])\n",
    "    \n",
    "    outputFeatureMap([image], layer1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": 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
}