miniproject1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Miniproject 1: Image Classification\n",
    "\n",
    "## Introduction\n",
    "\n",
    "### Important dates:\n",
    "\n",
    "- Project release: Friday, 15th March 2019\n",
    "- **Submission deadline**: Monday, 29th April 2019, 11:59 pm\n",
    "\n",
    "### Description\n",
    "\n",
    "One of the deepest traditions in learning about deep learning is to first [tackle the exciting problem of MNIST classification](http://yann.lecun.com/exdb/mnist/). [The MNIST database](https://en.wikipedia.org/wiki/MNIST_database) (Modified National Institute of Standards and Technology database) is a large database of handwritten digits that is commonly used as a first test for new classification algorithms. \n",
    "We follow this tradition to investigate the performance of artificial neural networks of different complexity on MNIST. However, since MNIST is too easy for accessing the full power of modern machine learning algorithms (see e.g. [this post](https://twitter.com/goodfellow_ian/status/852591106655043584)) we will extend our analysis to the recently introduced, harder [Fashion-MNIST dataset](https://github.com/zalandoresearch/fashion-mnist).\n",
    "\n",
    "\n",
    "### Prerequisites\n",
    "\n",
    "- You should have a running installation of [tensorflow](https://www.tensorflow.org/install/) and [keras](https://keras.io/). Feel free to gain inspiration from the [Keras example directory](https://github.com/keras-team/keras/tree/master/examples) for your implementations.\n",
    "- You should know the concepts \"multilayer perceptron\", \"stochastic gradient descent with minibatches\", \"convolutional neural network\", \"training and validation data\", \"overfitting\" and \"early stopping\".\n",
    "\n",
    "### What you will learn\n",
    "\n",
    "- You will learn how to define feedforward neural networks in keras and fit them to data.\n",
    "- You will be guided through a prototyping procedure for the application of deep learning to a specific domain.\n",
    "- You will get in contact with concepts discussed later in the lecture, like \"regularization\", \"batch normalization\" and \"convolutional networks\".\n",
    "- You will gain some experience on the influence of network architecture, optimizer and regularization choices on the goodness of fit.\n",
    "- You will learn to be more patient :) Some fits may take your computer quite a bit of time; run them over night (or on an external server).\n",
    "\n",
    "### Evaluation criteria\n",
    "\n",
    "The evaluation is (mostly) based on the figures you submit and your answer sentences. Provide clear and concise answers respecting the indicated maximum length (answers to the questions should be below the line that says \"Answer to question ...\").\n",
    "\n",
    "**The submitted notebook must be run by you!** We will only do random tests of your code and not re-run the full notebook. There will be fraud detection sessions at the end of the semester.\n",
    "\n",
    "### Your names\n",
    "\n",
    "**Before you start**: please enter your full name(s) in the field below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-03-09T09:08:24.514461Z",
     "start_time": "2018-03-09T09:08:24.506410Z"
    }
   },
   "outputs": [],
   "source": [
    "student1 = \"Fredrik Bixo\"\n",
    "student2 = \"Armen\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-22T21:52:59.697375Z",
     "start_time": "2018-02-22T21:52:59.689443Z"
    }
   },
   "source": [
    "## Some helper functions\n",
    "\n",
    "For your convenience we provide here some functions to preprocess the data and plot the results later. Simply run the following cells with `Shift-Enter`.\n",
    "\n",
    "### Dependencies and constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T14:27:09.352019Z",
     "start_time": "2018-02-23T14:27:08.476310Z"
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import numpy as np\n",
    "import time\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy.io\n",
    "\n",
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Conv2D, MaxPooling2D, Dropout, Flatten\n",
    "from keras.optimizers import SGD, Adam"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T15:11:52.252208Z",
     "start_time": "2018-02-23T15:11:52.121360Z"
    }
   },
   "outputs": [],
   "source": [
    "def plot_some_samples(x, y = [], yhat = [], select_from = [], \n",
    "                      ncols = 6, nrows = 4, xdim = 28, ydim = 28,\n",
    "                      label_mapping = range(10)):\n",
    "    \"\"\"plot some input vectors as grayscale images (optionally together with their assigned or predicted labels).\n",
    "    \n",
    "    x is an NxD - dimensional array, where D is the length of an input vector and N is the number of samples.\n",
    "    Out of the N samples, ncols x nrows indices are randomly selected from the list select_from (if it is empty, select_from becomes range(N)).\n",
    "    \n",
    "    Keyword arguments:\n",
    "    y             -- corresponding labels to plot in green below each image.\n",
    "    yhat          -- corresponding predicted labels to plot in red below each image.\n",
    "    select_from   -- list of indices from which to select the images.\n",
    "    ncols, nrows  -- number of columns and rows to plot.\n",
    "    xdim, ydim    -- number of pixels of the images in x- and y-direction.\n",
    "    label_mapping -- map labels to digits.\n",
    "    \n",
    "    \"\"\"\n",
    "    fig, ax = plt.subplots(nrows, ncols)\n",
    "    if len(select_from) == 0:\n",
    "        select_from = range(x.shape[0])\n",
    "    indices = np.random.choice(select_from, size = min(ncols * nrows, len(select_from)), replace = False)\n",
    "    for i, ind in enumerate(indices):\n",
    "        thisax = ax[i//ncols,i%ncols]\n",
    "        thisax.matshow(x[ind].reshape(xdim, ydim), cmap='gray')\n",
    "        thisax.set_axis_off()\n",
    "        if len(y) != 0:\n",
    "            j = y[ind] if type(y[ind]) != np.ndarray else y[ind].argmax()\n",
    "            thisax.text(0, 0, (label_mapping[j]+1)%10, color='green', \n",
    "                                                       verticalalignment='top',\n",
    "                                                       transform=thisax.transAxes)\n",
    "        if len(yhat) != 0:\n",
    "            k = yhat[ind] if type(yhat[ind]) != np.ndarray else yhat[ind].argmax()\n",
    "            thisax.text(1, 0, (label_mapping[k]+1)%10, color='red',\n",
    "                                             verticalalignment='top',\n",
    "                                             horizontalalignment='right',\n",
    "                                             transform=thisax.transAxes)\n",
    "    return fig\n",
    "\n",
    "def prepare_standardplot(title, xlabel):\n",
    "    fig, (ax1, ax2) = plt.subplots(1, 2)\n",
    "    fig.suptitle(title)\n",
    "    ax1.set_ylabel('categorical cross entropy')\n",
    "    ax1.set_xlabel(xlabel)\n",
    "    ax1.set_yscale('log')\n",
    "    ax2.set_ylabel('accuracy [% correct]')\n",
    "    ax2.set_xlabel(xlabel)\n",
    "    return fig, ax1, ax2\n",
    "\n",
    "def finalize_standardplot(fig, ax1, ax2):\n",
    "    ax1handles, ax1labels = ax1.get_legend_handles_labels()\n",
    "    if len(ax1labels) > 0:\n",
    "        ax1.legend(ax1handles, ax1labels)\n",
    "    ax2handles, ax2labels = ax2.get_legend_handles_labels()\n",
    "    if len(ax2labels) > 0:\n",
    "        ax2.legend(ax2handles, ax2labels)\n",
    "    fig.tight_layout()\n",
    "    plt.subplots_adjust(top=0.9)\n",
    "\n",
    "def plot_history(history, title):\n",
    "    fig, ax1, ax2 = prepare_standardplot(title, 'epoch')\n",
    "    ax1.plot(history.history['loss'], label = \"training\")\n",
    "    ax1.plot(history.history['val_loss'], label = \"validation\")\n",
    "    ax2.plot(history.history['acc'], label = \"training\")\n",
    "    ax2.plot(history.history['val_acc'], label = \"validation\")\n",
    "    finalize_standardplot(fig, ax1, ax2)\n",
    "    return fig\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 1: Data import and visualization (4 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "### Loading the data\n",
    "\n",
    "The datasets we use in this project (MNIST, Fashion-MNIST) consists of grayscale images with 28x28 pixels. Keras comes with a convenient in-built [data importer](https://keras.io/datasets/) for common datasets.\n",
    "\n",
    "1. As a warm-up exercise, use this importer to (down-)load the MNIST and Fashion-MNIST dataset. Assign useful variables to test & train images and labels for both datasets respectively. (2 pts)\n",
    "2. Use the corresponding plotting function defined above to plot some samples of the two datasets. What do the green digits at the bottom left of each image indicate? (1 sentence max.) (2 pts)\n",
    "\n",
    "The low resolution (and grayscale) of the images certainly misses some information that could be helpful for classifying the images. However, since the data has lower dimensionality due to the low resolution, the fitting procedures converge faster. This is an advantage in situations like here (or generally when prototyping), were we want to try many different things without having to wait too long for computations to finish.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T14:27:44.442862Z",
     "start_time": "2018-02-23T14:27:09.505547Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28)\n",
      "(60000,)\n",
      "(10000, 28, 28)\n",
      "(10000,)\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.datasets import fashion_mnist\n",
    "\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "(x_fashion_train, y_fashion_train), (x_fashion_test, y_fashion_test) = fashion_mnist.load_data()\n",
    "\n",
    "print(x_train.shape)\n",
    "print(y_train.shape)\n",
    "print(x_test.shape)\n",
    "print(y_test.shape)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 2:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Data pre-processing**: To prepare for fitting we transform the labels to one hot coding, i.e. for 5 classes, label 2 becomes the vector [0, 0, 1, 0, 0] (python uses 0-indexing). Furthermore we reshape (flatten) the input images to input vectors and rescale the data into the range [0,1]."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[5 0 4 ... 5 6 8]\n",
      "[[0. 0. 0. ... 0. 0. 0.]\n",
      " [1. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " ...\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 0. 0.]\n",
      " [0. 0. 0. ... 0. 1. 0.]]\n",
      "(10000,)\n"
     ]
    }
   ],
   "source": [
    "print(y_train)\n",
    "y_train_c = keras.utils.to_categorical(y_train)\n",
    "y_test_c = keras.utils.to_categorical(y_test)\n",
    "print(y_train_c)\n",
    "\n",
    "y_fashion_train = keras.utils.to_categorical(y_fashion_train)\n",
    "y_fashion_test = keras.utils.to_categorical(y_fashion_test)\n",
    "\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "ename": "IndexError",
     "evalue": "tuple index out of range",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-25-d10a2f9ca9b3>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mx_train\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mx_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mx_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mx_fashion_train\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx_fashion_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_fashion_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_fashion_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mx_fashion_train\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_fashion_train\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mx_fashion_test\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mx_fashion_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreshape\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_fashion_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_fashion_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mx_fashion_test\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_fashion_test\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mIndexError\u001b[0m: tuple index out of range"
     ]
    }
   ],
   "source": [
    "x_train = x_train.reshape(x_train.shape[0], x_train.shape[1]*x_train.shape[2])/np.max(x_train)\n",
    "x_test = x_test.reshape(x_test.shape[0], x_test.shape[1]*x_test.shape[2])/np.max(x_test)\n",
    "\n",
    "x_fashion_train = x_fashion_train.reshape(x_fashion_train.shape[0], x_fashion_train.shape[1]*x_fashion_train.shape[2])/np.max(x_fashion_train)\n",
    "x_fashion_test = x_fashion_test.reshape(x_fashion_test.shape[0], x_fashion_test.shape[1]*x_fashion_test.shape[2])/np.max(x_fashion_test)\n",
    "\n",
    "print(y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot MNIST-Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 24 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_some_samples(x_train, (y_train+9)%10);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plot Fashion-MNIST-Train"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 24 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_some_samples(x_fashion_train, y_fashion_train);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 2: No hidden layer (10 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "Define and fit a model without a hidden layer (since we will use multi-layer models later in this project, you can define a general constructor function for models with an arbitrary number of hidden layers already at this point). (1 pt for each step)\n",
    "\n",
    "1. Use the softmax activation for the output layer. D\n",
    "2. Use the categorical_crossentropy loss. D\n",
    "3. Add the accuracy metric to the metrics. D\n",
    "4. Choose stochastic gradient descent for the optimizer. D\n",
    "5. Choose a minibatch size of 128. D\n",
    "6. Fit for as many epochs as needed to see no further decrease in the validation loss.\n",
    "7. Plot the output of the fitting procedure (a history object) using the function plot_history defined above.\n",
    "8. Determine the indices of all test images that are misclassified by the fitted model and plot some of them using the function \n",
    "   `plot_some_samples(x_test, y_test, yhat_test, error_indices)`. Explain the green and red digits at the bottom of each image.\n",
    "9. Repeat the above steps for fitting the network to the Fashion-MNIST dataset.\n",
    "\n",
    "\n",
    "Hints:\n",
    "* Read the keras docs, in particular [Getting started with the Keras Sequential model](https://keras.io/getting-started/sequential-model-guide/).\n",
    "* Have a look at the keras [examples](https://github.com/keras-team/keras/tree/master/examples), e.g. [mnist_mlp](https://github.com/keras-team/keras/blob/master/examples/mnist_mlp.py)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_3 (Dense)              (None, 10)                7850      \n",
      "=================================================================\n",
      "Total params: 7,850\n",
      "Trainable params: 7,850\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 1.2896 - acc: 0.6914 - val_loss: 0.8122 - val_acc: 0.8356\n",
      "Epoch 2/10\n",
      "60000/60000 [==============================] - 1s 13us/step - loss: 0.7166 - acc: 0.8413 - val_loss: 0.6055 - val_acc: 0.8628\n",
      "Epoch 3/10\n",
      "60000/60000 [==============================] - 1s 13us/step - loss: 0.5870 - acc: 0.8606 - val_loss: 0.5235 - val_acc: 0.8746\n",
      "Epoch 4/10\n",
      "60000/60000 [==============================] - 1s 13us/step - loss: 0.5251 - acc: 0.8697 - val_loss: 0.4777 - val_acc: 0.8807\n",
      "Epoch 5/10\n",
      "22528/60000 [==========>...................] - ETA: 0s - loss: 0.4988 - acc: 0.8755"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-9-b132c84dd5b3>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     26\u001b[0m                     \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     27\u001b[0m                     \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 28\u001b[0;31m                     validation_data=(x_test, y_test_c))\n\u001b[0m\u001b[1;32m     29\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     30\u001b[0m \u001b[0mplot_history\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Hist\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    202\u001b[0m                     \u001b[0mbatch_logs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mo\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    203\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 204\u001b[0;31m                 \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_batch_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch_index\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_logs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    205\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mcallback_model\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstop_training\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    206\u001b[0m                     \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/callbacks.py\u001b[0m in \u001b[0;36mon_batch_end\u001b[0;34m(self, batch, logs)\u001b[0m\n\u001b[1;32m    113\u001b[0m         \u001b[0mt_before_callbacks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    114\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mcallback\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 115\u001b[0;31m             \u001b[0mcallback\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_batch_end\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlogs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    116\u001b[0m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_delta_ts_batch_end\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mt_before_callbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    117\u001b[0m         \u001b[0mdelta_t_median\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmedian\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_delta_ts_batch_end\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/callbacks.py\u001b[0m in \u001b[0;36mon_batch_end\u001b[0;34m(self, batch, logs)\u001b[0m\n\u001b[1;32m    234\u001b[0m             \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    235\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtotals\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 236\u001b[0;31m                     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtotals\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mv\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    237\u001b[0m                 \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    238\u001b[0m                     \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtotals\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mk\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mv\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 10\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax', input_dim=784))\n",
    "\n",
    "model.summary()\n",
    "\n",
    "# Optimizer:\n",
    "sgd = optimizers.SGD(lr=0.01)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "# Fitting:\n",
    "history = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "\n",
    "plot_history(history, \"Hist\")\n",
    "\n",
    "score = model.evaluate(x_test, y_test_c, verbose=0)\n",
    "# test\n",
    "yhat_test = model.predict(x_test)\n",
    "\n",
    "#targetValues = y_test.argmax(axis=1)\n",
    "predictedValues = yhat_test.argmax(axis=1)\n",
    "\n",
    "missmatches = []\n",
    "\n",
    "for i in range(0, len(predictedValues)):\n",
    "    if predictedValues[i] != y_test[i]:\n",
    "        missmatches.append(i) \n",
    "        \n",
    "plot_some_samples(x_test, (y_test+9)%10, (yhat_test+9)%10,missmatches);\n",
    "    \n",
    "print('Target values:', y_test)\n",
    "print('Predicted values:', predictedValues)\n",
    "\n",
    "print('Test loss:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 10:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 3: One hidden layer, different optizimizers & overfitting (10 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "Train a network with one hidden layer and compare different optimizers.\n",
    "\n",
    "1. Use one hidden layer with 128 units and the 'relu' activation. Use the [summary method](https://keras.io/models/about-keras-models/) to display your model in a compact way. (1 pt)\n",
    "2. Fit the model for 50 epochs with different learning rates of stochastic gradient descent (SGD). (1pt)\n",
    "3. Replace the stochastic gradient descent optimizer with the [Adam optimizer](https://keras.io/optimizers/#adam). (1pt)\n",
    "4. Plot the learning curves of SGD with a reasonable learning rate (i.e. in the range [0.01,0.1]) together with the learning curves of Adam in the same figure. Take care of a reasonable labeling of the curves in the plot. (2pts)\n",
    "5. Answer the questions below. (4pts)\n",
    "6. Run the network (using the Adam optimizer) on the Fashion-MNIST dataset and plot the learning curves using the plot_history function defined above. (1pt)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# This plotting routine might help you ...\n",
    "def comparison_plot(history_sgd, history_adam, label1, label2, title):\n",
    "    fig, ax1, ax2 = prepare_standardplot(title, \"epochs\")\n",
    "    ax1.plot(history_sgd.history['loss'], label=label1 + ' training')\n",
    "    ax1.plot(history_sgd.history['val_loss'], label=label1 + ' validation')\n",
    "    ax1.plot(history_adam.history['loss'], label=label2 + ' training')\n",
    "    ax1.plot(history_adam.history['val_loss'], label=label2 + ' validation')\n",
    "    ax2.plot(history_sgd.history['acc'], label=label1 + ' training')\n",
    "    ax2.plot(history_sgd.history['val_acc'], label=label1 + ' validation')\n",
    "    ax2.plot(history_adam.history['acc'], label=label2 + ' training')\n",
    "    ax2.plot(history_adam.history['val_acc'], label=label2 + ' validation')\n",
    "    finalize_standardplot(fig, ax1, ax2)\n",
    "    return fig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_31 (Dense)             (None, 128)               100480    \n",
      "_________________________________________________________________\n",
      "dense_32 (Dense)             (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 101,770\n",
      "Trainable params: 101,770\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.3650 - acc: 0.9015 - val_loss: 0.1960 - val_acc: 0.9450\n",
      "Epoch 2/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1681 - acc: 0.9518 - val_loss: 0.1412 - val_acc: 0.9599\n",
      "Epoch 3/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1215 - acc: 0.9652 - val_loss: 0.1143 - val_acc: 0.9664\n",
      "Epoch 4/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0936 - acc: 0.9726 - val_loss: 0.0981 - val_acc: 0.9711\n",
      "Epoch 5/50\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.0753 - acc: 0.9784 - val_loss: 0.0888 - val_acc: 0.9727\n",
      "Epoch 6/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0623 - acc: 0.9821 - val_loss: 0.0844 - val_acc: 0.9739\n",
      "Epoch 7/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0523 - acc: 0.9845 - val_loss: 0.0825 - val_acc: 0.9750\n",
      "Epoch 8/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0439 - acc: 0.9869 - val_loss: 0.0760 - val_acc: 0.9772\n",
      "Epoch 9/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0367 - acc: 0.9896 - val_loss: 0.0765 - val_acc: 0.9778\n",
      "Epoch 10/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0317 - acc: 0.9911 - val_loss: 0.0768 - val_acc: 0.9782\n",
      "Epoch 11/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.0259 - acc: 0.9932 - val_loss: 0.0762 - val_acc: 0.9772\n",
      "Epoch 12/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0231 - acc: 0.9942 - val_loss: 0.0728 - val_acc: 0.9771\n",
      "Epoch 13/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0190 - acc: 0.9957 - val_loss: 0.0774 - val_acc: 0.9774\n",
      "Epoch 14/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0158 - acc: 0.9966 - val_loss: 0.0745 - val_acc: 0.9781\n",
      "Epoch 15/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0150 - acc: 0.9963 - val_loss: 0.0769 - val_acc: 0.9769\n",
      "Epoch 16/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0122 - acc: 0.9971 - val_loss: 0.0814 - val_acc: 0.9759\n",
      "Epoch 17/50\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.0111 - acc: 0.9975 - val_loss: 0.0818 - val_acc: 0.9778\n",
      "Epoch 18/50\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.0091 - acc: 0.9982 - val_loss: 0.0838 - val_acc: 0.9786\n",
      "Epoch 19/50\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.0078 - acc: 0.9985 - val_loss: 0.0792 - val_acc: 0.9786\n",
      "Epoch 20/50\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.0070 - acc: 0.9987 - val_loss: 0.0872 - val_acc: 0.9778\n",
      "Epoch 21/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0056 - acc: 0.9992 - val_loss: 0.0821 - val_acc: 0.9802\n",
      "Epoch 22/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0049 - acc: 0.9993 - val_loss: 0.1023 - val_acc: 0.9746\n",
      "Epoch 23/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0057 - acc: 0.9989 - val_loss: 0.0889 - val_acc: 0.9797\n",
      "Epoch 24/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.0034 - acc: 0.9998 - val_loss: 0.0864 - val_acc: 0.9798\n",
      "Epoch 25/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0038 - acc: 0.9994 - val_loss: 0.0997 - val_acc: 0.9756\n",
      "Epoch 26/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.0027 - acc: 0.9998 - val_loss: 0.0941 - val_acc: 0.9792\n",
      "Epoch 27/50\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.0017 - acc: 0.9999 - val_loss: 0.0895 - val_acc: 0.9798\n",
      "Epoch 28/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0017 - acc: 0.9998 - val_loss: 0.1108 - val_acc: 0.9762\n",
      "Epoch 29/50\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.0103 - acc: 0.9968 - val_loss: 0.1025 - val_acc: 0.9784\n",
      "Epoch 30/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0032 - acc: 0.9994 - val_loss: 0.1002 - val_acc: 0.9801\n",
      "Epoch 31/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0013 - acc: 0.9999 - val_loss: 0.0978 - val_acc: 0.9789\n",
      "Epoch 32/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 7.9730e-04 - acc: 1.0000 - val_loss: 0.0964 - val_acc: 0.9797\n",
      "Epoch 33/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 6.2066e-04 - acc: 1.0000 - val_loss: 0.0960 - val_acc: 0.9803\n",
      "Epoch 34/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0093 - acc: 0.9972 - val_loss: 0.1039 - val_acc: 0.9770\n",
      "Epoch 35/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0025 - acc: 0.9995 - val_loss: 0.1035 - val_acc: 0.9786\n",
      "Epoch 36/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 7.8003e-04 - acc: 1.0000 - val_loss: 0.0982 - val_acc: 0.9798\n",
      "Epoch 37/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 4.5281e-04 - acc: 1.0000 - val_loss: 0.0987 - val_acc: 0.9802\n",
      "Epoch 38/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 3.6843e-04 - acc: 1.0000 - val_loss: 0.1004 - val_acc: 0.9807\n",
      "Epoch 39/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 3.3059e-04 - acc: 1.0000 - val_loss: 0.1006 - val_acc: 0.9804\n",
      "Epoch 40/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 2.9238e-04 - acc: 1.0000 - val_loss: 0.1010 - val_acc: 0.9804\n",
      "Epoch 41/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 2.7260e-04 - acc: 1.0000 - val_loss: 0.1025 - val_acc: 0.9800\n",
      "Epoch 42/50\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 2.4915e-04 - acc: 1.0000 - val_loss: 0.1046 - val_acc: 0.9806\n",
      "Epoch 43/50\n",
      "58752/60000 [============================>.] - ETA: 0s - loss: 2.3508e-04 - acc: 1.0000"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-27-c365dca844eb>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     52\u001b[0m                     \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     53\u001b[0m                     \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 54\u001b[0;31m                     validation_data=(x_test, y_test_c))\n\u001b[0m\u001b[1;32m     55\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     56\u001b[0m \u001b[0mplot_history\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhistorySGD\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Hist\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    210\u001b[0m                         val_outs = test_loop(model, val_f, val_ins,\n\u001b[1;32m    211\u001b[0m                                              \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mbatch_size\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 212\u001b[0;31m                                              verbose=0)\n\u001b[0m\u001b[1;32m    213\u001b[0m                         \u001b[0mval_outs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval_outs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    214\u001b[0m                         \u001b[0;31m# Same labels assumed.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mtest_loop\u001b[0;34m(model, f, ins, batch_size, verbose, steps)\u001b[0m\n\u001b[1;32m    390\u001b[0m                 \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    391\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 392\u001b[0;31m             \u001b[0mbatch_outs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    393\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbatch_outs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    394\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0mbatch_index\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2713\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2714\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2715\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2716\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2717\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2673\u001b[0m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_metadata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2675\u001b[0;31m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2676\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2677\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 50\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(128, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "model.summary()\n",
    "\n",
    "learningRates = [0.01,0.05,0.1]\n",
    "\n",
    "# SDG optimzer\n",
    "#for i in learningRates:\n",
    "\n",
    "    # Optimizer:\n",
    " #   sgd = optimizers.SGD(lr=0.01)\n",
    "\n",
    "  ##  model.compile(loss='categorical_crossentropy',\n",
    "  #            optimizer=sgd,\n",
    "   #           metrics=['accuracy'])\n",
    "    # Fitting:\n",
    " #   history = model.fit(x_train, y_train_c,\n",
    " #                   batch_size=batch_size,\n",
    "  #                  epochs=epochs,\n",
    " #                   verbose=1,\n",
    "  #                  validation_data=(x_test, y_test_c))\n",
    "\n",
    "  #  plot_history(history, \"Learning rate: \" + str(i))\n",
    "    \n",
    "\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "    \n",
    "    # Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "    # Optimizer: Adam\n",
    "  \n",
    "    \n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T15:42:45.497806Z",
     "start_time": "2018-02-23T15:42:44.961166Z"
    }
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Question**: What happens if the learning rate of SGD is A) very large B) very small? Please answer A) and B) with one full sentence each (double click this markdown cell to edit).\n",
    "\n",
    "**Answer**:\n",
    "\n",
    "A) If the learning rate is very small, then the algorithm overfits more since it converges faster to a maximum on the traning data.\n",
    "\n",
    "B) If the learning rate is very low, then the algorithm overfits less since it converges slower to a maximum on the traning data.\n",
    "\n",
    "**Question**: At which epoch (approximately) does the Adam optimizer start to overfit (on MNIST)? Please answer with one full sentence.\n",
    "\n",
    "**Answer**: The adam optimizer starts overfitting after around 17 epochs, when the time when the training error decreases and the test error increases.\n",
    "\n",
    "**Question**: Explain the qualitative difference between the loss curves and the accuracy curves with respect to signs of overfitting. Please answer with at most 3 full sentences.\n",
    "\n",
    "**Answer**:\n",
    "\n",
    "The accuracy graph is less useful for determining when the graph is overfitting. This is because the loss curves measure the loss in terms of probabilities, while the accuracy only counts the number of correctly predicted labels. Therefore, the accuracy graph remains flat while the loss curve starts slowly increasing because of overfitting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 4: Model performance as a function of number of hidden neurons (8 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "Investigate how the best validation loss and accuracy depends on the number of hidden neurons in a single layer.\n",
    "\n",
    "1. Fit a reasonable number of models (e.g. 5) with different hidden layer sizes (between 10 and 1000 hidden neurons) to the MNIST dataset. You may use the Adam optimizer and a meaningful number of epochs (overfitting!). (3 pts)\n",
    "2. Plot the best validation loss and accuracy versus the number of hidden neurons. Is the observed trend in accordance with the [general approximation theorem](https://en.wikipedia.org/wiki/Universal_approximation_theorem)? If not, what might be practical reasons for the deviation? (2 sentences max.) (3 pts)\n",
    "3. Repeat steps 1. & 2. for the Fashion-MNIST dataset. (2 pts)\n",
    "\n",
    "In this exercise we fit each model only for one initialization and random seed. In practice one would collect some statistics (e.g. 25-, 50-, 75-percentiles) for each layer size by fitting each model several times with different initializations and the random seeds. You may also want to do this here. It is a good exercise, but not mandatory as it takes quite a bit of computation time.\n",
    "\n",
    "### Solution\n",
    "\n",
    "We can se that the validation loss decreseas as a function of the number of neurons in the hidden layer while the accuracy incrases. Since the validation loss decreases and the accuracy increases, the neural network becomes a better approximator for the real function that we try to model. This is in accordance with the general approximation theorem, which states that any continious function could be approximated with a finite number of neurons in the hidden layer, however the number of neurons needed might be exponential."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEdCAYAAABZtfMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xl8VPXV+PHPyUYSlpCERUiABFQUEAgkuAPuS11xqVZ9BBW07rZ9qnaR1mrtr1W0Prb6oAJu1Sou1daKyINiq5aEILsiQhJC2LOxJJDl/P64N3EIWSbL5E5mzvv1mldm7r0zc2bkeuZ+tyOqijHGGBNsIrwOwBhjjGmMJShjjDFByRKUMcaYoGQJyhhjTFCyBGWMMSYoWYIyxhgTlCxBGdNKIjJYRPaKSGQzx6iIHNmZcRkTaixBGeMHEckTkTMBVLVAVXuoao2772MRucnbCI0JPZagjDHGBCVLUMa0QEReAgYD77lNez91m/CiRORh4FTgKXffU408v5uIPCoiBSKyXUSeEZG4zv4cxnQ1lqCMaYGqXgcUABeqag/gdZ99Pwc+BW53m/1ub+Ql/h9wNDAWOBJIAR4IeODGdHGWoIwJIBERYDpwj6oWq+oe4LfAVd5GZkzwi/I6AGNCXF8gHljm5CoABGhyBKAxxmEJyhj/NLfsf3P7dgEVwEhV3dKxIRkT2qyJzxj/bAeGtnafqtYCzwKPi0g/ABFJEZFzAhKlMSHEEpQx/nkE+IWIlAKXN9j3R+ByESkRkScbee69wAbgCxEpBz4Chgc0WmNCgFjBQmOMMcHIrqCMMcYEJUtQxhhjgpIlKGOMMUHJEpQxxpigZAnKGGNMULIEZYwxJihZgjLGGBOULEEZY4wJSpagjDHGBCVLUMYYY4KSJShjjDFBycptNKFPnz6alpbmdRimC1q2bNkuVe3rdRyBZueIaSt/zxFLUE1IS0sjJyfH6zBMFyQi+V7H0BnsHDFt5e85Yk18xhhjgpIlKGOMMUHJEpQxxpigZH1QYaSqqorCwkIqKyu9DiUkxMbGkpqaSnR0tNehGBOSLEGFkcLCQnr27ElaWhoi4nU4XZqqsnv3bgoLC0lPT/c6HGNCkjXxhZHKykqSk5MtOXUAESE5OdmuRo0JIEtQYcaSU8ex79KYwLIE1QrL8ov50+INXodhjDFBS1X5Zvse/mfRN2wprWjXa1mCaoUvNhbzhwVfU7LvoNehhIUePXoAUFRUxOWXX97oMZMnT25xsugTTzzB/v376x+ff/75lJaWdlygxoS52lpleUEJv/vnV5zx2Cec9fgSHlu4npy84na9rg2SaIUJ6UkA5OSXcNaI/h5HEz4GDhzI/Pnz2/z8J554gmuvvZb4+HgA3n///Y4KzZiwVVVTy382FrNgzTY+XLuN7eUHiIoQThyWzLRT0jl7RH/694pt13tYgmqF41ISiImMICev2BJUG9x7770MGTKEW2+9FYBf/epXiAhLliyhpKSEqqoqHnroIS6++OJDnpeXl8cFF1zA6tWrqaioYNq0aaxdu5Zjjz2WiorvmhB++MMfkp2dTUVFBZdffjm//vWvefLJJykqKuK0006jT58+LF68uH6Jnj59+jBr1izmzJkDwE033cTdd99NXl4e5513HqeccgqfffYZKSkp/O1vfyMuLq7zvixjglDFwRqWfLOTBWu2sWjdDsoqqoiNjmDS0X05Z+QRnHFMfxLiO27ahSWoVoiNjmR0agJL23nZGgx+/d4a1haVd+hrjhjYi5kXjmxy/1VXXcXdd99dn6Bef/11PvjgA+655x569erFrl27OOGEE7jooouaHIDw9NNPEx8fz8qVK1m5ciXjxo2r3/fwww+TlJRETU0NZ5xxBitXruTOO+9k1qxZLF68mD59+hzyWsuWLWPu3Ln85z//QVU5/vjjmTRpEomJiXzzzTe8+uqrPPvss1x55ZW8+eabXHvttR3wLRnTtZTtr2LRV9tZsGYbn6zfSWVVLb1iozhzRH/OGXkEE4/qS1xMZEDe2xJUK2WlJ/HcpxupOFgTsP8ooSojI4MdO3ZQVFTEzp07SUxMZMCAAdxzzz0sWbKEiIgItmzZwvbt2zniiCMafY0lS5Zw5513AjB69GhGjx5dv+/1119n9uzZVFdXs3XrVtauXXvI/ob+9a9/cemll9K9e3cApkyZwqeffspFF11Eeno6Y8eOBWD8+PHk5eV10LdgTPDbUV7JgrXb+XDNNj7/djfVtUr/Xt24Yvwgzhl5BMcPTSI6MvBDGCxBtVJWWiJPf6x8ubmUE4clex1OmzV3pRNIl19+OfPnz2fbtm1cddVVvPLKK+zcuZNly5YRHR1NWlpai3OLGru62rRpE48++ijZ2dkkJiYyderUFl9HVZvc161bt/r7kZGRhzQlGhOKNu3ax4I121iwZhvLC5xBROl9unPTqUM5Z2R/xqT2JiKic6dWWIJqpfGDkxCBnLziLp2gvHLVVVcxffp0du3axSeffMLrr79Ov379iI6OZvHixeTnN78K/8SJE3nllVc47bTTWL16NStXrgSgvLyc7t27k5CQwPbt2/nnP//J5MmTAejZsyd79uw5rIlv4sSJTJ06lfvuuw9V5e233+all14KyOc2JtioKmuKyvlwzTYWrNnO19v3ADAqpRc/Putozhl1BEf16+HpfL+wSlAiMhT4OZCgqo2PW25BQnw0w/v3JDu/pGODCxMjR45kz549pKSkMGDAAK655houvPBCMjMzGTt2LMccc0yzz//hD3/ItGnTGD16NGPHjmXChAkAjBkzhoyMDEaOHMnQoUM5+eST658zY8YMzjvvPAYMGMDixYvrt48bN46pU6fWv8ZNN91ERkaGNeeZkFVTqyzLL6m/UiosqSBCIDMtiQcuGMHZI/uTmhjvdZj1pLlmjna9sMgg4EXgCKAWmK2qf2zja80BLgB2qOqoBvvOBf4IRALPqerv/Hi9+S0lqMzMTG1qfs0v31nN28u3sGLm2UR28iVve6xbt45jjz3W6zBCSmPfqYgsU9VMj0LqNM2dIyZ4HKiu4bMNu1mwZhsfrdvOrr0HiYmM4JSj+nDOyP6ceWx/knt0a/mFOpC/50ggr6CqgR+raq6I9ASWichCVV3rE2Q/oEJV9/hsO1JVGy7XMA94Cifh4XNsJPAn4CygEMgWkXdxktUjDV7jBlXd0REfLDMtkZe+yGfd1nJGpSR0xEsaY0yHUFVK9lfx2be7WLBmO4u/2sHeA9X06BbF5OF9OXfUEUwe3o8e3YK/AS1gEarqVmCre3+PiKwDUoC1PodNAn4oIueraqWITAcuBc5v8FpLRCStkbeZAGxQ1Y0AIvIacLGqPoJzxdVqInIhcOGRRx7Z5DFZac6E3ey8YktQxphOp6rs3HOAvN37ydu9jwL3b/7u/eTv3kd5ZTUAyd1juGD0AM4ZeQQnHZlMt6iuNfK4U1Kom1wygP/4blfVN0QkHXhNRN4AbsC5GvJXCrDZ53EhcHwzcSQDDwMZInK/m8gOoarvAe9lZmZOb+p1BvaOI6V3HDl5JUw7uWuVWlBVW+S0gwSqedwYcPqLtpZVkN9oEtpPRVVN/bGREUJqYhxDkruTMbg3g5PiOS4lgcy0pC7VDdFQwBOUiPQA3gTuVtXDZoaq6u/dK5+ngWGqurc1L9/Itib/r6Gqu4FbWvH6TcpKS+Szb3d3qf/hx8bGsnv3biu50QHq6kHFxrZvKRcT3qpqaiksqXASz6595Bfvr09IhcUVHKyprT82JjKCwcnxpCXHc9KwPqT1iWdIcnfSkuMZ2DuuU+YldbaAJigRicZJTq+o6ltNHHMqMAp4G5gJ3N6KtygEBvk8TgWK2hZt62SlJ/HOl0UUFO9nSHL3znjLdktNTaWwsJCdO3d6HUpIqKuoa0xzKqtqKCjeT96ufc5f9yoob/c+ikorqan97jd1fEwkQ5K7M7x/T84a0Z+05O4MSXYS0RG9Yrv01VBbBCxBifMT/XlgnarOauKYDOBZ4HvAJuBlEXlIVX/h59tkA0e5zYRbgKuAH7Q7eD981w9V0mUSVHR0tFV/NSbAikoreGrxBr7dsZf83fvZVn7ohPFesVGk9+nO2EGJXDL2u6ugwcnx9O3RzVo3fATyCupk4DpglYh86W77mar6LiUdD1yhqt8CiMj1wNSGLyQirwKTgT4iUgjMVNXnVbVaRG4HFuCM3JujqmsC9YF8Hdm3Bwlx0WRvKuby8fYr2hgDJfsOcu3z/6GotIKRAxM46cjkQ66C0pLj6R0f43WYXUYgR/H9i8b7iHyP+XeDx1U4V1QNj7u6mdd4H+j0+gkREUJWWiLZ+V1/4VhjTPtVHKzhxheyKSyp4OUbj68vz2PaLvR61TpRZloSG3fuY9feA16HYozxUHVNLXe8upzlm0t58qqxlpw6iCWodqjrh8rJs2WPjAlXqsov/7aaj9Zt58GLRnLuqAFehxQyLEG1w6iUXnSLimh3WWNjTNf1x0Xf8OrSzdx22jCuOzHN63BCiiWodugWFcmYQb3JtgRlTFh6dWkBT3z0DZePT+UnZw/3OpyQYwmqnSakJbG6qJz9B6u9DsUY04kWrt3Oz99exeThfXlkynE2PDwALEG1U2ZaIjW1Wl/gyxgT+pbll3DHq7kcl5LAn68ZF5KrOAQD+1bbafyQRCIEa+YzJkxs2LGXG1/I5ohescyZmkV8TPCvCt5VWYJqp56x0RxzRC8byWdMGNheXsn1c5YSFSG8eMPxnV5HKdxYguoAE9KTyC0oodpnYUdjTGgpr6zi+jlLKd1/kHnTJjA4OXgqz4YqS1AdIDMtkf0Ha1i79bDF2o0xIeBAdQ03v7iMDTv28sx1460OXCexBNUB6ibsLt1k/VDGOyJyroh8LSIbROS+RvYPEZFFIrJSRD4WkVSffb8XkTUisk5EnhQbklavtlb50esr+Hzjbv5wxWhOPaqv1yGFDUtQHaB/r1gGJ8VbP5TxjIhEAn8CzgNGAFeLyIgGhz0KvKiqo4EHgUfc556Es7jzaJzSN1k41a7Dnqry0D/W8Y+VW7n/vGO4NMMWhu5MlqA6SGZaItl5xVZl1XhlArBBVTeq6kHgNeDiBseMABa59xf77FcgFogBugHRwPaAR9wFPPvpRub8exPTTk5jxsShXocTdixBdZAJaUns3neQTbv2eR2KCU8pwGafx4XuNl8rgMvc+5cCPUUkWVU/x0lYW93bAlVd19ibiMgMEckRkZxQL3z59vJCfvv+V3xv9AB++b0RNhHXA5agOkimLRxrvNXY/z0bXs7/BJgkIstxmvC2ANUiciRwLE5F6hTgdBGZ2NibqOpsVc1U1cy+fUO3L+bTb3by32+s5MShycy6cgwRYVbJNlhYguogw/p2J6l7DEttwq7xRiEwyOdxKlDke4CqFqnqFFXNAH7ubivDuZr6QlX3qupe4J/ACZ0TdvBZvaWMW15axpH9evC//zWeblGRXocUtixBdRARIXNIoq1sbrySDRwlIukiEgNcBbzre4CI9BGRunP+fmCOe78A58oqSkSica6uGm3iC3UFu/czde5SesfH8MINE+gVG+11SGHNElQHykpLIm/3fnbsqfQ6FBNmVLUauB1YgJNcXlfVNSLyoIhc5B42GfhaRNYD/YGH3e3zgW+BVTj9VCtU9b3OjD8Y7N57gP+a8x+qa5UXbphA/16xXocU9mwRqQ6Ulf5dP9T5x1nRMtO5VPV94P0G2x7wuT8fJxk1fF4NcHPAAwxi+w5Uc8O8bLaWVfKX6cdzZL8eXodksCuoDjVyYC9ioyNs4VhjupCqmlpu+0suq7aU8dQPxjF+iJVrDxYtJigRsf9afoqOjCBjUKIlKGO6CFXl/rdW8fHXO3nokuM4a0R/r0MyPvy5gvqPiLwhIufb8icty0pPYm1ROXsPWAFDY4LdYx+uZ/6yQu464yh+cPxgr8MxDfiToI4GZgPXARtE5LcicnRgw+q6stISqVXIzbf5UMYEs5c+z+OpxRu4esIg7j7zKK/DMY1oMUGpY6GqXg3cBFwPLBWRT0TkxIBH2MVkDE4kMkJsuLkxQeyD1Vt54N01nHlsP35z8ShbJSJItTiKT0SSgWtxrqC2A3fgzK8YC7wBpAcywK6mR7coRgzoRbatKGFMUFq6qZg7X/uSjEG9+Z+rxxFl5dqDlj//ZT4HegGXqOr3VPUtVa1W1RzgmcCG1zVlpiWyfHMJB6utgKExwWT99j3c9EI2qYlxPH99FnExtkpEMPMnQQ1X1d8A5SLS03eHqv6/wITVtU1IS6KyqpY1RWVeh2KMcRWVVnD9nKXERkfywrQJJHaP8Tok0wJ/EtR4EVkFrARWi8gKERkf4Li6tLqFY224uTHBoWx/FVPnLmVvZTXzpk1gUJKVa+8K/ElQc4BbVTVNVYcAtwFzAxtW19a3ZzfS+3S3fihjgkBlVQ3TX8xh0659/O914xkxsJfXIRk/+ZOg9qjqp3UPVPVfwJ7AhRQa6haOra21AobGeKWmVrn7tS9ZmlfMY1eO5aQj+3gdkmkFfxLUUhH5XxGZLCKTROTPwMciMk5ExgU6wK4qKz2Jkv1VbNy11+tQjAlLqsqv31vDB2u28csLRnDRmIFeh2RayZ/FYse6f2c22H4STkG00zs0ohCRVd8PVcKR/Xq2cLQxpqP9+eNvefHzfGZMHMqNp9hsmK6oxQSlqqd1RiChJi05nj49YsjeVMzVE2wJFdM8P9e8rFXV0oAHEwLeyNnMHxZ8zSVjB3Lfucd4HY5pI38m6ibgXD3VlYD+BHjQrcRpmiAiZKUlkZ1vI/mMX4rcW3NLGkQC9munBYu/2sF9b63ilCP78PvLrVx7V+bvKL49wJXurRwbxeeXzLQkNhdXsK3MChiaFq1T1aGqmt7UDdjtdZDBrqZWuef1LxnevyfPXDeemChbJaIr8+e/3jBVnamqG93br4GhgQ4sFEyw+VDGf/6sa2lrX7Zg/fY9lO6vYvrEdHp0s3qsXZ0/CapCRE6peyAiJwMVgQspdBw7oCfxMZG2cKxpkapWAojISw331W2rO8Y0LbfAmXs4bnCix5GYjuDPT4xbgBfdviiAEpwVzU0LoiIjGDc4kaU2Ydf4b6TvAxGJBGzlFj/l5peS3D2GwbZSREho9gpKRCJw1uIbA4wGRqtqhqqu7JToQkBWWhJfbSunvLLK61BMEBOR+0VkDzBaRMrd2x5gB/A3j8PrMpYXlJAxONHKZ4SIZhOUqtYCt7v3y1W1vFOiCiFZaYmowjIrYGiaoaqPqGpP4A+q2su99VTVZFW93+v4uoLifQfZuGsf44b09joU00H86YNaKCI/EZFBIpJUdwt4ZCFi7ODeRFkBQ+O/pT7N6YhIbxG5xMuAuorl1v8Ucvzpg7rB/XubzzbFRvL5JT4mipEpCWRvsiso45eZqvp23QNVLRWRmcA7HsbUJeQWlBAZIYxOTWj5YNMl+JOgjm04ekhEYgMUT0jKGpLIi1/kc6C6hm5RViDNNKuxVg0bL+2H3PxSd+SsfV2hwp8mvs/83GaakJWexMHqWlZvscU3TItyRGSWiAwTkaEi8jiwzOuggl11TS0rCkuteS/ENJmgROQItzBhnIhk1K1eLiKTARvD2QqZQ5yTZqk185mW3QEcBP4KvI4z5/C2Zp9h+Hr7HvYfrLEEFWKauxY+B5gKpAKzfLbvAX4WwJhCTnKPbgzr290dKDHM63BMEFPVfcB9ItJDVa1Wi59yC5w1dC1BhZYmE5SqvgC8ICKXqeqbnRhTSMpKS+Kfq7dRW6u2eKVpkoicBDwH9AAGi8gY4GZVvdXbyILb8vwS+vSIYVBSnNehmA7kT2/i30XkB0Ca7/Gq+mCgggpFmWlJvJa9mW927GX4EVYfyjTpcZzWi3cBVHWFiExs/ikm1ybohiR/Bkn8DbgYqAb2+dxMK9jCscZfqrq5waYaTwLpInbvPUDe7v3WvBeC/LmCSlXVcwMeSYgblBRHv57dyM4r5toThngdjglem91mPhWRGOBOYJ3HMQW15fX9T7aCRKjxa5i5iBwX8EhCnIiQlZ5Eji0ca5p3C86ovRSgEBiLjeJrVm5BCVERwuhUS1Chxp8EdQqwTES+FpGVIrJKRGyx2DbIGpLIltIKtpRatRJzOHfl8utU9RpV7a+q/VT1WlX1q1ChiJzrnqcbROS+RvYPEZFF7nn8sYikuttPE5EvfW6VXWl5pdyCEo4d0Iu4GJsEH2r8aeI7L+BRhImsdKcfKievmJSxKR5HY4KNqtaIyMU4AyVaxU1ufwLOwrnyyhaRd1V1rc9hjwIvquoLInI68AhOQlyMc6WGu87mBuDD9n2azlFdU8uKzWVcmZnqdSgmAFq8glLVfGAQcLp7f78/zzOHO+aIXvToFsXSTTZQwjTp3yLylIic6jM5fpwfz5sAbHCrXh8EXsMZ3ORrBLDIvb+4kf0AlwP/VNX9bf0AnemrbXuoqKph3BAbIBGKWryCcheqzASGA3OBaOBl4OTAhhZ6IiOEcUMSrR/KNOck96/vNA4FTm/heSmA7+i/QuD4BsesAC4D/ghcCvQUkeQGTYhXcejE/EOIyAxgBsDgwYNbCCnwbAXz0ObPldClwEW4Q8tVtQiwiTxtNCEtka+376FsvxUwNIdyC4Q+raqnNbi1lJwAGpsApA0e/wSYJCLLgUnAFpzpI3XvPwA4DljQ1Juo6mxVzVTVzL59+/oRVmDlFpTSp0c3UhNtgm4o8idBHVRVxf3HLiLdAxtSaMt050Pl5FsznzmUb4HQNijEaYqvkwoUNXj9IlWdoqoZwM/dbb4rGF8JvK2qXebXU25BCeMG97YJuiHKnwT1uoj8L9BbRKYDHwHPBjas0DV2UG+iI4Vsa+YzjWtrgdBs4CgRSXfnT12FuxpFHRHp416lAdwPzGnwGlcDr7b3A3SWXXsPkL97v/U/hbAW+6BU9VEROQsox+mHekBVFwY8shAVGx3JcSkJtqKEaUqbCoSqarWI3I7TPBcJzFHVNSLyIJCjqu8Ck4FHRESBJb7vISJpOFdgn3TMxwi85bZAbMjzq7KXm5AsKXWQrLQk5vx7E5VVNcRG29wN8x1VTW/Hc98H3m+w7QGf+/OB+U08Nw9noEWX8d0EXaugG6psuLgHstKSqKpRVhZaAUNzKBGJFpE7RWS+e7tdRKK9jisY5eaXMGJgL/uRF8IsQXlgvNtmbs18phFPA+OBP7u38e4246O6ppaVhWXWvBfi/GriqyMiicAgVbWljtohsXsMR/fvYQnKNCZLVcf4PP4/EVnhWTRBqm6CboYtEBvSWryCctfs6uWOJFoBzBWRJifyGf9kpiWxLK+EmtqGU1VMmKsRkfqyyyIyFCu3cZhcm6AbFvxp4ktQ1XJgCjBXVccDZwY2rNCXlZbIngPVfL1tj9ehmODy38Bi94fhJ8D/AT/2OKagk5tfQt+eNkE31PnTxBflzjC/Endyn2m/LJ8ChiMG9vI4GhMsVHWRiByFM6VDgK9U9YDHYQWd3IJSm6AbBvy5gnoQZ27FBlXNdpscvglsWKEvpXccAxJirR/KHEJEbgPiVHWlqq4A4kXkVq/jCia79h6goNgq6IYDf1Yzf0NVR6vqre7jjap6WeBDC20iQlZaEtl5xTgrSRkDwHRVLa17oKolwHQP4wk6uflu/5OtIBHy/Bkk8Xt3kES0W+xsl4hc2xnBhbqstES2lx+gsMQKGJp6EeLTbuXWeYrxMJ6gk1tQSlSEcFyKTdANdf408Z3tDpK4AGdByqNxOnJNO9UVMLRmPuNjAc76l2e4RQVfBT7wOKagkltQwkiboBsW/ElQdbPYzwdeVVX7v2kHObpfT3rGRlmCMr7uxSkq+EOctfIWAT/1NKIgUlVTy8rCUjKs/yks+DOK7z0R+QqoAG4Vkb5AZWDDCg8REULmkERb2dzUc0tuPOPeTANfbd1DZVWt9T+FCX8GSdwHnAhkunVi9tF4qWjTBlnpSWzYsZfifQe9DsWYoPfdBF1bQSIc+DNIIhq4DviriMwHbgR2N/8s46+6+VA51sxnTItyC0ro17MbKb1tgm448KcPquHileOwxSs7zOjUBGKiIsjJt2Y+cygRiRURm8Xtw6mgm2gTdMOEP31QtnhlAHWLimRMagJLN9kVlPmOiNyE03IRISKfqurPvI7Jazv3HGBzcQXXnTDE61BMJ/HnCsoWrwywzLQkVm8po+Kgfa3hSkQubLDpTFWdpKqnAt/zIqZgYwvEhh9/EpQtXhlgE9KSqK5Vlm+2Zr4wNkZE/iYida0VK0XkFRF5GVjjZWDBIreghOhIYZRN0A0bzTbxiUgEzvByW7wygMYNSUQEcvJKOGlYH6/DMR5Q1YdE5AjgQbd/5QGgBxBv9dccy/NLGTEwwSbohpFmr6DcORmPqeqBusUrLTl1vIS4aIb372kTds0+4G7gT8Bs4GpgvacRBYmqmlpWbim14eVhxp8mvg9F5DKxYTMBlZWWRG5+CdU1tV6HYjwgIg8B/8BZOeI0Vb0Ip0DoP0TkOk+DCwLrtpY7E3St/yms+JOgfgS8ARwQkXIR2SMi5QGOK+xkpiWy72ANX1kBw3B1gapOBE4C/gtAVd8FzgGSvAwsGNgK5uGpxWHmqtqzMwIJdxPchWOXbiq2TuDwtFpEXgLigE/qNqpqNfBHz6IKErkFpfTv1Y2BCbFeh2I6kT8rSVwqIgk+j3uLyCWBDSv8DEiII6V3HDn51g8VjlT1WuD3wC9U9R6v4wk2NkE3PPnTxDdTVcvqHrjF1GYGLqTwNSE9iey8EitgGIZEZJyqrlLVr5o7pjNjChY79lRSWFJh/U9hyJ8E1dgx/qxAYVopMy2RnXsOkL97v9ehmM43V0QSRSSpqRvwvNdBeiE33ykwPG6IjeALN/4kmhwRmYUz9FWBO4BlAY0qTNUtHJudV0xan+4eR2M6WQLOedVcG9bOToolqCx3J+iOHGh9s+HGnwR1B/BL4K/u4w+BXwQsojB2ZN8e9I6PJjuvmCsyB3kdjulEqprmdQzByqmgaxN0w5E/o/j2Afd1Qixhr66AYY4VMDQGgINWlKrsAAAgAElEQVTVtawsLOOa422B2HDkTx+U6URZaUls3LWPnXtswQ5j1m0t50B1rfU/hSlLUEEm0+2HWmbDzY2xFczDnCWoIHNcSgLdoiLItma+sCQib4rI99yFmsNebkEpR/SKZaBV0A1LTfZBicj/4Izaa5Sq3hmQiMJcTFQEYwf1toVjw9fTwDTgSRF5A5jX3NyoUJebX2LNe2GsuV9pOTjDXpu6mQDJSktiTVE5+w5Uex2K6WSq+pGqXgOMA/KAhSLymYhME5Ho5p4rIueKyNciskFEDhvYJCJDRGSRiKx067ul+uwbLCIfisg6EVkrImkd+8lab0d5JVtKbYJuOGvyCkpVX+jMQMx3stKTeGrxBpYXlHLKUVYfKtyISDJwLU7J9+XAK8ApwPXA5CaeE4kzV/EsoBDIFpF3VXWtz2GPAi+q6gsicjrwiPseAC8CD6vqQhHpAXi+rH5d/1OGJaiw1eIwcxHpC9wLjADqV2pU1dMDGFdYGze4NxHiTNi1BBVeROQt4BjgJeBCVd3q7vqriOQ089QJwAZV3ei+zmvAxYBvghoB1K3ztxh4xz12BBClqgsBVHVvB32cdsktKCUmMoJRKb28DsV4xJ+O2FeAdUA68GucZofsAMYU9nrGRnPsgF62cGx4ekpVR6jqIz7JCQBVzWzmeSnAZp/Hhe42XyuAy9z7lwI93au1o4FSEXlLRJaLyB/cK7LDiMgMEckRkZydOwO7sEVufgkjU3rRLcom6IYrfxJUsqo+D1Sp6ieqegNwQoDjCntOAcNSqqyAYbg5VkTqRwW46/Pd6sfzGlsiqeEgp58Ak0RkOTAJ2AJU47SknOruzwKGAlMbexNVna2qmaqa2bdvXz/CapuD1bWs3FJm/U9hzp8EVeX+3eoOf80AUpt7gmm/zLREKqpqWFtktSHDzHS3YgAAqloCTPfjeYWA7/pYqUCR7wGqWqSqU1Q1A/i5u63Mfe5yVd3o1p96B2eQhmfWbi3nYLVV0A13/iSoh9x6UD/G+YX1HN+1Y3cJIjJURJ4Xkflex+Iv34VjTViJEJ+iR25TW4wfz8sGjhKRdBGJAa4C3vU9QET6+Myvuh+Y4/PcRLe/GeB0Du276nTfVdC1IebhrMUEpap/V9UyVV2tqqep6ni3FHWnEJE5IrJDRFY32N7skFpf7i/DGwMbacfq3yuWwUnxlqDCzwLgdRE5wx1p9yrwQUtPcq98bnefvw54XVXXiMiDInKRe9hk4GsRWQ/0Bx52n1uD8+NzkYiswmkufLZjP1br5BaUMCAhlgEJNkE3nPkziu8F4K66ZgcRSQQec/uiOsM84CmcYbB1MTU6pBaIxBk66+sGVd3ROaF2rKy0JD7+egeqapVEw8e9wM3AD3ESxYc4rRYtUtX3gfcbbHvA5/58oNFWBHcE3+i2hdzxlheUWvOe8avcxuiGbeJuP1SnUNUljUwabHRIrao+AlzQ1vcSkRnADIDBgwe39WU6TFZaIm/mFrJx1z6G9e3hdTimE6hqLc5qEk97HYtXtrsTdKednOZ1KMZjflXUda+aAHAre3pdUdefIbX1RCRZRJ4BMkTk/qaO66wRSv6qWzg2x5r5woaIHCUi893VHDbW3byOqzN91/9kV1Dhzp9E8xjwmc8Agytw26495M+Q2u92qO4GbglcOIExrG93krrHsHRTCd/P8v6KznSKucBM4HHgNJx1+cKqfTe3oISYyAhGDrQJuuHOn0ESL+JM7tsO7ACmqOpLgQ6sBS0OqQ0FIm4BQ5uwG07iVHURIKqar6q/whlVFzZyC0oZZRN0Dc0kKBHp5f5NArYBf8FZVWKbu81LLQ6pDRUT0pPI372fHeWVXodiOkelOxT8GxG5XUQuBfp5HVRnOVhdy6otZYy35j1D81dQf3H/LsNZ2bzuVve4U4jIq8DnwHARKRSRG5saUttZMXWmzPr5UFYfKkzcDcQDdwLjcRaNvd7TiDrRmqIym6Br6jW3mvkF7oTBSapa0IkxNYzj6ia2HzakNhSNHNiLuOhIsvOK+d7oAV6HYwLInT5xpar+N7AXp/8prOQWOAOGbYCEgRb6oFRVgbc7KRbTiOjICDIGWwHDcOBOmB3vu5JEuMktKCGldxz9e8W2fLAJef4MM/9CRLICHolpUmZaEuu2lrOnsqrlg01Xtxz4m4hcJyJT6m5eB9VZlueXkDHYljcyDn+GmZ8G3Cwi+cA+nCGvqqpBM+s81E1IS6JWneaPSUd7Pz/LBFQSsJtDR+4p8JY34XSebWWVFJVVcpP1PxmXPwnqvIBHYZqVMbg3kRFCTl6xJagQp6ph1+9Up66CrvU/mTotJihVzReRMTj1YgA+VdUVgQ3L+OreLYqRA3uxdJP1Q4U6EZlLI5POO3HtS8/k5pfQLSqCEQNsgq5xtNgHJSJ34cx/6ufeXhaROwIdmDlU5pAkvtxcysFqK2AY4v4O/MO9LQJ64YzoC3m5BSUcl5JATJQ/XeMmHPjzL+FG4HhVfcBdGfkE/CugZjpQVloiB6prWV1U5nUoJoBU9U2f2yvAlcAor+MKtAPVNazeUm7Ne+YQ/iQoAWp8HtcQZmuDBYP6CbvWzBdujgJCfiHGNUXlHKypZZyN4DM+/BkkMRf4j4jUzYe6BHg+cCGZxvTt2Y30Pt3Jzivh5kleR2MCRUT2cGgf1DacGlEhrX4FcxvBZ3z4M0hiloh8DJyCc+U0TVWXBzowc7istEQ+XLud2lolIsIuYkORqvb0OgYvLC8oJaV3HP1sgq7x4c8giSQgD3gZeAnIF5HoAMflGRG5UERml5UFX19PZloSpfur+HZnWPSZhyURuVREEnwe9xaRS7yMqTPkFpRY/5M5jD99ULnATmA98I17f5OI5IrI+EAG5wVVfU9VZyQkJLR8cCfLsoVjw8FMVa3/deRWs57pYTwBt7Wsgq1lldb/ZA7jT4L6ADhfVfuoajLOxN3XgVuBPwcyOHOotOR4UnrH8dynGyndf9DrcExgNHZOel3BOqBy890FYq3/yTTgT4LKVNUFdQ9U9UNgoqp+AXQLWGTmMCLCE1eNpbCkghkvLeNAdU3LTzJdTY6IzBKRYSIyVEQexylxE7JyC5wJusfaBF3TgD8JqlhE7hWRIe7tp0CJWxrAZo12sqy0JP5wxWiWbirm/jdX4Sw4b0LIHcBB4K84LRUVwG2eRhRguQUljE61CbrmcP40HfwApw38Hffxv9xtkTiTCE0nu3hsCvm79zNr4XrS+nTnzjOO8jok00FUdR9wn9dxdJYD1TWs2VLOtJPTvA7FBCF/hpnvAu4QkR6q2nD42IbAhGVacsfpR5K3ex+zFq5ncFI8l2SkeB2S6QAishC4wh0cgYgkAq+p6jneRhYYq7c4E3QzrP/JNMKfYeYnichaYK37eIyI2OAIj4kIv5symhOGJvHT+SttIdnQ0acuOQGoagnOGpghaXn9CuY2gs8czp9G38eBc3Bq1OCuZD4xkEEZ/8RERfDMteNJTYzj5pdy2LRrn9chmfarFZH6pY1EZAiNrG4eKnILSkhNjKNfT5ugaw7nV6+kqm5usMmGjwWJ3vExzJ2WhYhww7xsSvbZ8PMu7ufAv0TkJRF5CVgC3O9xTAGTm19qw8tNk/xJUJtF5CRARSRGRH4CrAtwXKYVhiR3Z/Z149lSWsHNNvy8S1PVD4BxfDeKb7zvNI9QUlRawbZym6BrmuZPgroFZ5hrClAIjMWZpGuCSGZaEo9eMYalecXcZ8PPu7oaYAdQBowQkZBsUrcKuqYl/gwzH66q1/huEJGTgX8HJiTTVheNGUjB7n08+uF6hiTHc/eZR3sdkmklEbkJuAtIBb7Eqb/2OXC6l3EFQm5+KbHRNkHXNM2fK6j/8XObCQK3nXYkl41L5YmPvuHt5YVeh2Na7y4gC8hX1dOADJz1L0NObkEJo1N6Ex1pE3RN45q8ghKRE4GTgL4i8iOfXb1wJumaICQiPDLlOLaU7ufe+asYmBDH8UOTvQ7L+K9SVStFBBHppqpfichwr4PqaJVVNawpKuOGU9K9DsUEseZ+usQAPXCSWE+fWzlweeBDM20VExXB/16bSWpSHDe/vIyNVp6jKykUkd44K7csFJG/AUX+PFFEzhWRr0Vkg4gcthqFu1TZIhFZKSIfi0iqz74aEfnSvb3bYZ+mCWuKyqiqURvBZ5rV5BWUqn4CfCIi81Q1vxNjMh0gIT6aeVMncMmf/80N87J569aTSeoe43VYpgWqeql791cishhIwKko0Cx3bcw/AWfhDGbKFpF3VXWtz2GPAi+q6gsicjrwCHCdu69CVcd21Odoia1gbvzhT+PvfhH5g4i8LyL/V3cLeGSm3QYnx/Psf42nqKySm1/KseHnXYyqfqKq76qqP5PbJgAbVHWje/xrwMUNjhkBLHLvL25kf6fJLShhUFIcfXtaQQTTNH8S1CvAV0A68Guc6rrZAYzJU8FcUbctxg9J4rErxpCdV8JP56+04eehKwXwnVBf6G7ztQK4zL1/KdBTROo6KGNFJEdEvmiugq+IzHCPy9m5s21jN1TVqaBrV0+mBf4kqGRVfR6ocn/R3YAz9DUkBXNF3ba6cMxA/vuc4fztyyIe/+gbr8MxgSGNbGv4a+QnwCQRWQ5MArYA1e6+waqaiVOp4AkRGdbYm6jqbFXNVNXMvn37tinQorJKtpcfsARlWuTPPKgq9+9WEfkeTodtajPHmyB06+Rh5O3ax5OLvmFIUjyXjbf/hCGmEBjk8ziVBoMrVLUImAIgIj2Ay+rKy7v7UNWNIvIxzvD2bwMRaG6+O0HXEpRpgT9XUA+JSALwY5xfYM8B9wQ0KtPhRISHLz2Ok4Ylc99bK/li426vQzIdKxs4SkTSRSQGuAo4ZDSeiPQRkbpz/n5gjrs9UUS61R0DnIxbvSAQcgtKiI2O4JgBPQP1FiZEtJigVPXvqlqmqqtV9TRVHa+qAR+GajpeTFQET18znsFJ8dz80jK+teHnIUNVq4HbgQU4a2W+rqprRORBEbnIPWwy8LWIrAf6Aw+724/FKTW/AmfwxO8ajP7rULkFpYxOtQm6pmX+1IN6wZ2XUfc4UUTmBDYsEygJ8dHMnTqBqAhn9fNiW/08ZKjq+6p6tKoOU9WH3W0P1P2gVNX5qnqUe8xNqnrA3f6Zqh6nqmPcv88HKsbKqhrWFpVZ857xiz8/YUY3UkAtI3AhmUAbnBzPs9dnsq2skhkv5lBZZcPPTedYvaVugq6tYG5a5k+CinDLTgMgIkn4N7jCBLFxgxOZdeVYcvJt+LnpPLaCuWkNfxLNY8BnIjIfZ9jqlXzXdm26sO+NHkB+8XB+/8HXpCXH86OzQ27JNxNkcvNLGZwUT58eNkHXtKzFBKWqL4pIDs5y/wJMCWQHqulcP5w0jPxd+3ny/zYwOLk7l9vwcxMgdRN0Txpmixcb//jVVOcmJEtKIUhEeOjSURSW7uf+t1aS0juOE+1/ICYAtpRWsGPPAWveM36zcZ6G6MgI/nzNeIYkd+fml3LYsMOGn5uOl1tgC8Sa1rEEZQBIiItm7tQsoiMjuGFeNrv3HvA6JBNicvNLiIuO5JgjbIKu8Y8lKFNvUJIz/Hx7eSUzXlpmw89Nh1peUMLo1ASibIKu8ZP9SzGHGDc4kce/P5Zl+SX85I0V1Nba8HPTfk4F3XLrfzKtYgnKHOb84wZw77nH8PeVW5m1cL3X4ZgQsGpLGdW1VkHXtI5NuDWNumXSUPJ37+OpxRsYnBzPlZmDWn6SMU2oW8E8w1aQMK1gCco0SkT4zSWjKCyp4GdvrSK1dxwnHdnH67BMF5VbUMKQZJuga1rHmvhMk6IjI/jzteNI79OdW15exoYde7wOyXRBzgTdUmveM61mCco0q1dsNHOmZhETFcG0ednssuHnppUKSyrYueeALRBrWs0SlGnRoKR4nrs+ix3lB2z1c9NqdQvEZtgVlGklS1ANiMiFIjK7rKzM61CCythBvXni+2PJLSjlx6+voKqm1uuQTBexvKCU+BiboGtazxJUA6r6nqrOSEhI8DqUoHPecQO4/7xj+MeqrVz4P/9iWX6x1yGZLiDXJuiaNrJ/MaZVbp40jNnXjaesoorLnv6c+99aRdn+Kq/DMkHKqaBbbgMkTJtYgjKtdvbII1j4o0nceEo6f80u4IxZH/PO8i1W9NAcZmWhTdA1bWcJyrRJj25R/PKCEbx7+ymk9I7j7r9+yXXPL2XTrn1eh2aCyDKboGvawRKUaZdRKQm8devJ/ObikazYXMo5Tyzhjx99w4FqG+lnnP6ntOR4km2CrmkDS1Cm3SIjhOtOTGPRjydx9oj+PP7Res7746d8/u1ur0MzHlJVlheUWPOeaTNLUKbD9OsVy1M/GMe8aVlU1dRy9bNf8KPXv7TaUmFqc3EFu/YeJMNWMDdtZAnKdLjJw/vx4d2TuHXyMN79sogzZn3CX7MLrHRHmKmboGsrSJi2sgRlAiIuJpKfnnsM7991Kkf368m9b67i+7M/Z/12W88vXOQWlBAfE8nw/jZB17SNrWZuAuro/j15bcYJzF9WyG//uY7z//gpMyYO5Y7TjyIuJtLr8EwA5RaUMCa1d5eeoFtVVUVhYSGVlZVeh9IlxcbGkpqaSnR0dJuebwnKBFxEhHBl1iDOOLYfv33/K/788be8t7KI31w8isnD+3kdngmA/QerWbd1D7dMGup1KO1SWFhIz549SUtLQ0S8DqdLUVV2795NYWEh6enpbXqNrvvTxnQ5yT268diVY3h1+glER0YwdW42t/0llx3l9us01KwsLKMmBCboVlZWkpycbMmpDUSE5OTkdl19WoIyne7EYcn8865T+dFZR7Nw7XbOeOwTXvw8jxobRBEyQmkFc0tObdfe784SlPFEt6hI7jzjKBbcPZExg3rzwN/WMOXpz1hTZKvIh4Lc/FLS+3QnqXuM16GYLswSlPFUep/uvHTjBP541Vi2lOznoqf+zUN/X8u+A9Veh2baqG6Cri1vZNrLEpTxnIhw8dgUFv1oMt/PGsRz/9rEmbM+YcGabV6HZtqgoHg/u/cd7PL9T+Gmujr4fhTaKD4TNBLio/ntpcdx2bhUfv72Km5+aRlnHtufX188kpTecV6HF/RE5Fzgj0Ak8Jyq/q7B/iHAHKAvUAxcq6qFPvt7AeuAt1X19rbG8d0E3dBKUL9+bw1ri8o79DVHDOzFzAtHtnjcJZdcwubNm6msrOSuu+5ixowZfPDBB/zsZz+jpqaGPn36sGjRIvbu3csdd9xBTk4OIsLMmTO57LLL6NGjB3v37gVg/vz5/P3vf2fevHlMnTqVpKQkli9fzrhx4/j+97/P3XffTUVFBXFxccydO5fhw4dTU1PDvffey4IFCxARpk+fzogRI3jqqad4++23AVi4cCFPP/00b731Vod9P5agTNAZPySR9+44hTn/2sQTH33DWbM+4Z4zj2bayWldek5NIIlIJPAn4CygEMgWkXdVda3PYY8CL6rqCyJyOvAIcJ3P/t8An7Q3ltz8UrrHRDLcKuh2mDlz5pCUlERFRQVZWVlcfPHFTJ8+nSVLlpCenk5xsVM89De/+Q0JCQmsWrUKgJKSkhZfe/369Xz00UdERkZSXl7OkiVLiIqK4qOPPuJnP/sZb775JrNnz2bTpk0sX76cqKgoiouLSUxM5LbbbmPnzp307duXuXPnMm3atA793JagTFCKjozg5knDOP+4Afzq3TU8/P463lq+hd9eOiokRoYFwARgg6puBBCR14CLAd8ENQK4x72/GHinboeIjAf6Ax8Ame0JJLeghDGDehMZEVqj3/y50gmUJ598sv5KZfPmzcyePZuJEyfWzy9KSkoC4KOPPuK1116rf15iYsvnyhVXXEFkpDNpvqysjOuvv55vvvkGEaGqqqr+dW+55RaioqIOeb/rrruOl19+mWnTpvH555/z4osvdtAndtjPURPUBiXF89z1mTxz7ThK9h1kytOf8Yt3VrG5eD9l+6s4UF1jhRIdKcBmn8eF7jZfK4DL3PuXAj1FJFlEIoDHgP9u6U1EZIaI5IhIzs6dOw/bv/9gNV9t2xNyzXte+vjjj/noo4/4/PPPWbFiBRkZGYwZM6bRIdyq2uh2320N5yV17969/v4vf/lLTjvtNFavXs17771Xf2xTrztt2jRefvllXn31Va644or6BNZRwuoKSkQuAb4H9AP+pKofehyS8YOIcO6oAZxyVF8e+/BrXvgsj5e/KKjfHyEQFx1JXEwUcTERzv3oSGKjI4mLiXT3RdZvj4tx97n34xs89n1uvPu3W1REsM+HaSy4hpn7J8BTIjIVWAJsAaqBW4H3VXVzS59RVWcDswEyMzMP+2WwYrM7QXeIjeDrKGVlZSQmJhIfH89XX33FF198wYEDB/jkk0/YtGlTfRNfUlISZ599Nk899RRPPPEE4DTxJSYm0r9/f9atW8fw4cN5++236dmz8ebXsrIyUlKc3zXz5s2r33722WfzzDPPMHny5PomvqSkJAYOHMjAgQN56KGHWLhwYYd/9oAmKBHpDTwHjMI5WW5Q1c/b8DpzgAuAHao6qsG+ZjuGfanqO8A7IpKI0x5vCaoL6dEtipkXjuT7WYNYXlBKxcEaKqpqqKyqoeJgDfuraqh0t1W420r2HWRLI8e15aLr0AQWwU/PPYZzRh7R8R+0bQqBQT6PU4Ei3wNUtQiYAiAiPYDLVLVMRE4EThWRW4EeQIyI7FXV+1obRP0E3UF2BdVRzj33XJ555hlGjx7N8OHDOeGEE+jbty+zZ89mypQp1NbW0q9fPxYuXMgvfvELbrvtNkaNGkVkZCQzZ85kypQp/O53v+OCCy5g0KBBjBo1qn7AREM//elPuf7665k1axann356/fabbrqJ9evXM3r0aKKjo5k+fTq33+6Mo7nmmmvYuXMnI0aM6PDPLoFsHhGRF4BPVfU5EYkB4lW11Gd/P6BCVff4bDtSVTc0eJ2JwF6cDt5RPtsjgfX4dAwDV+Mkq0cahHODqu5wn/cY8Iqq5jYVe2Zmpubk5LTlY5sgp6ocrKmtT3CHJrpa9h+sPiShVVTVusdVu39rqayq4boTh3DC0OTDXl9Elqlqu/pxWktEonDOhTNwroyygR+o6hqfY/oAxapaKyIPAzWq+kCD15kKZPoziq+xc+St3EI+/WYXj39/bHs/UlBYt24dxx57rNdhBLXbb7+djIwMbrzxxkb3N/Yd+nuOBOwKyh2yOhGYCqCqB4GDDQ6bBPxQRM5X1UoRmY7TNn6+70GqukRE0hp5m0Y7hlX1EZwrroYxCfA74J/NJScT2kSEblGRdIuKJFQaolS1WkRuBxbg/ECbo6prRORBIEdV3wUmA4+IiOI08d3W0XFMGZfKlHGpHf2yJkiNHz+e7t2789hjjwXk9QPZxDcU2AnMFZExwDLgLlXdV3eAqr4hIunAayLyBnADztWQvxrrGD6+mePvAM4EEtwrtWcaHiAiFwIXHnnkka0Iwxjvqer7wPsNtj3gc38+ML+F15gHzAtAeCYELVu2LKCvH8hRfFHAOOBpVc0A9gGHtWmr6u+BSuBp4CJVbbxxtHH+dAz7vteTqjpeVW9pLDm5x7ynqjMSEhJaEYYxJlTZKNG2a+93F8gEVQgUqup/3MfzcRLWIUTkVJxBFG8DM9vwHs12DBtjTFvFxsaye/duS1JtUFcPKjY2ts2vEbAmPlXdJiKbRWS4qn6N03nrO2kQEckAnsUZ+r0JeFlEHlLVX/j5NtnAUW4z4RbgKuAHHfYhjDFhLTU1lcLCQhqb82VaVldRt60CPQ/qDuAVdwTfRqDhOhjxwBWq+i2AiFyPO6jCl4i8itPB20dECoGZqvp8Ux3DgfowxpjwEh0d3eZqsKb9ApqgVPVLmlk2RVX/3eBxFc4VVcPjrm7mNQ7rGDbGGNP12VJHxhhjgpIlKGOMMUEpoCtJdGUishPIb2RXH2BXJ4cTbOw7cDT1PQxR1b6dHUxns3OkWfYdONp1jliCaiURyensZWyCjX0HDvseGmffi30Hddr7PVgTnzHGmKBkCcoYY0xQsgTVerO9DiAI2HfgsO+hcfa92HdQp13fg/VBGWOMCUp2BWWMMSYoWYIyxhgTlCxBtYKInCsiX4vIBhFpdTnsrk5EBonIYhFZJyJrROQur2PyiohEishyEfm717EEEztH7Byp0xHniCUoP7nl5f8EnAeMAK4WkRHeRtXpqoEfq+qxwAnAbWH4HdS5C1jndRDBxM4RwM4RX+0+RyxB+a++vLxbvv414GKPY+pUqrpVVXPd+3tw/vGleBtV5xORVJwSMc95HUuQsXPEzhGg484RS1D+a6y8fNj9w6sjImlABvCf5o8MSU8APwVqvQ4kyNg54sPOkfafI5ag/Neq8vKhTER6AG8Cd6tqudfxdCYRuQDYoarLvI4lCNk54rJzpGPOEUtQ/rPy8oCIROOceK+o6ltex+OBk4GLRCQPpwnrdBF52duQgoadI9g5QgeeIzZR108iEgWsxyldvwWn3PwPwqmCr4gI8AJQrKp3ex2P10RkMvATVb3A61iCgZ0jdo401N5zxK6g/KSq1UBdefl1wOvhdOK5Tgauw/lF9KV7O9/roExwsHMEsHOkQ9kVlDHGmKBkV1DGGGOCkiUoY4wxQckSlDHGmKBkCcoYY0xQsgRljDEmKFmCMsYYE5QsQZk2c1evNsY0wc6R9rEEFSZEJM2tUfOsW6fmQxGJE5FhIvKBiCwTkU9F5Bj3+HkicrnP8/e6fye79W7+Aqxyt/1IRFa7t7ubez93350islZEVorIa53+ZRjTCDtHgpCq2i0MbkAaTq2ase7j14FrgUXAUe6244H/c+/PAy73ef5e9+9kYB+Q7j4ej3MSdgd6AGtwVnBu9P3c+0VAN/d+b6+/G7vZTYDsaAcAAAFDSURBVNXOkWC8RbUmmZkub5OqfuneX4ZzgpwEvOEsIQZANz9eZ6mqbnLvnwK8rar7AETkLeBU4N0m3g9gJfCKiLwDvNPmT2NMx7NzJIhYggovB3zu1wD9gVJVHdvIsdW4TcDuApgxPvv2+dxvrMRCU+8X597/HjARuAj4pYiMVGcdN2O8ZudIELE+qPBWDmwSkSvAOclEZIy7Lw+naQKcqqjRTbzGEuASEYkXke7ApcCnTb2hiEQAg1R1MU5Bs944zR7GBCM7RzxkCcpcA9woIitw2sbrSnQ/C0wSkaU47e77GnuyOuWt5wFLcSqHPqeqy5t5v0jgZRFZBSwHHlfV0o74IMYEiJ0jHrHVzI0xxgQlu4IyxhgTlCxBGWOMCUqWoIwxxgQlS1DGGGOCkiUoY4wxQckSlDHGmKBkCcoYY0xQ+v/r/zBORSCifwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Plotting:\n",
    "\n",
    "\n",
    "\n",
    "valHist = [0.2131, 0.0886,0.0696,0.0612,0.0607]\n",
    "accHist = [0.9394, 0.9748,0.9836,0.9847,0.9872]\n",
    "\n",
    "fig, ax1, ax2 = prepare_standardplot(\"title\", 'neurons')\n",
    "ax1.plot(valHist, label = \"validation\")\n",
    "ax2.plot(accHist, label = \"accuracy\")\n",
    "finalize_standardplot(fig, ax1, ax2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/100\n",
      "60000/60000 [==============================] - 2s 41us/step - loss: 0.8434 - acc: 0.7495 - val_loss: 0.4160 - val_acc: 0.8805\n",
      "Epoch 2/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.3719 - acc: 0.8940 - val_loss: 0.3302 - val_acc: 0.9054\n",
      "Epoch 3/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.3216 - acc: 0.9079 - val_loss: 0.3082 - val_acc: 0.9115\n",
      "Epoch 4/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2994 - acc: 0.9150 - val_loss: 0.2908 - val_acc: 0.9170\n",
      "Epoch 5/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.2855 - acc: 0.9191 - val_loss: 0.2830 - val_acc: 0.9193\n",
      "Epoch 6/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.2758 - acc: 0.9227 - val_loss: 0.2716 - val_acc: 0.9225\n",
      "Epoch 7/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.2677 - acc: 0.9246 - val_loss: 0.2689 - val_acc: 0.9242\n",
      "Epoch 8/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.2606 - acc: 0.9258 - val_loss: 0.2595 - val_acc: 0.9266\n",
      "Epoch 9/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.2472 - acc: 0.9298 - val_loss: 0.2487 - val_acc: 0.9288\n",
      "Epoch 10/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.2368 - acc: 0.9331 - val_loss: 0.2467 - val_acc: 0.9301\n",
      "Epoch 11/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.2302 - acc: 0.9348 - val_loss: 0.2382 - val_acc: 0.9327\n",
      "Epoch 12/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.2241 - acc: 0.9373 - val_loss: 0.2387 - val_acc: 0.9335\n",
      "Epoch 13/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2197 - acc: 0.9378 - val_loss: 0.2331 - val_acc: 0.9336\n",
      "Epoch 14/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2164 - acc: 0.9384 - val_loss: 0.2334 - val_acc: 0.9329\n",
      "Epoch 15/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.2127 - acc: 0.9402 - val_loss: 0.2273 - val_acc: 0.9346\n",
      "Epoch 16/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.2092 - acc: 0.9405 - val_loss: 0.2253 - val_acc: 0.9355\n",
      "Epoch 17/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2066 - acc: 0.9418 - val_loss: 0.2287 - val_acc: 0.9335\n",
      "Epoch 18/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2033 - acc: 0.9422 - val_loss: 0.2239 - val_acc: 0.9330\n",
      "Epoch 19/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.2011 - acc: 0.9430 - val_loss: 0.2240 - val_acc: 0.9376\n",
      "Epoch 20/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1993 - acc: 0.9433 - val_loss: 0.2187 - val_acc: 0.9370\n",
      "Epoch 21/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1966 - acc: 0.9445 - val_loss: 0.2241 - val_acc: 0.9353\n",
      "Epoch 22/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1951 - acc: 0.9449 - val_loss: 0.2166 - val_acc: 0.9368\n",
      "Epoch 23/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1929 - acc: 0.9447 - val_loss: 0.2201 - val_acc: 0.9367\n",
      "Epoch 24/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1911 - acc: 0.9457 - val_loss: 0.2171 - val_acc: 0.9380\n",
      "Epoch 25/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1885 - acc: 0.9464 - val_loss: 0.2155 - val_acc: 0.9384\n",
      "Epoch 26/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1873 - acc: 0.9473 - val_loss: 0.2163 - val_acc: 0.9386\n",
      "Epoch 27/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1853 - acc: 0.9478 - val_loss: 0.2142 - val_acc: 0.9387\n",
      "Epoch 28/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1843 - acc: 0.9471 - val_loss: 0.2157 - val_acc: 0.9378\n",
      "Epoch 29/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1825 - acc: 0.9483 - val_loss: 0.2154 - val_acc: 0.9396\n",
      "Epoch 30/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1814 - acc: 0.9482 - val_loss: 0.2190 - val_acc: 0.9377\n",
      "Epoch 31/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1796 - acc: 0.9489 - val_loss: 0.2171 - val_acc: 0.9380\n",
      "Epoch 32/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1780 - acc: 0.9495 - val_loss: 0.2144 - val_acc: 0.9380\n",
      "Epoch 33/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.1770 - acc: 0.9495 - val_loss: 0.2152 - val_acc: 0.9384\n",
      "Epoch 34/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1760 - acc: 0.9498 - val_loss: 0.2166 - val_acc: 0.9390\n",
      "Epoch 35/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1741 - acc: 0.9505 - val_loss: 0.2223 - val_acc: 0.9365\n",
      "Epoch 36/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1737 - acc: 0.9502 - val_loss: 0.2140 - val_acc: 0.9381\n",
      "Epoch 37/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1727 - acc: 0.9504 - val_loss: 0.2148 - val_acc: 0.9384\n",
      "Epoch 38/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1717 - acc: 0.9508 - val_loss: 0.2143 - val_acc: 0.9388\n",
      "Epoch 39/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1704 - acc: 0.9513 - val_loss: 0.2164 - val_acc: 0.9383\n",
      "Epoch 40/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1699 - acc: 0.9513 - val_loss: 0.2154 - val_acc: 0.9392\n",
      "Epoch 41/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1690 - acc: 0.9509 - val_loss: 0.2166 - val_acc: 0.9380\n",
      "Epoch 42/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1685 - acc: 0.9516 - val_loss: 0.2131 - val_acc: 0.9391\n",
      "Epoch 43/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1676 - acc: 0.9514 - val_loss: 0.2157 - val_acc: 0.9394\n",
      "Epoch 44/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1667 - acc: 0.9523 - val_loss: 0.2165 - val_acc: 0.9383\n",
      "Epoch 45/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1652 - acc: 0.9528 - val_loss: 0.2148 - val_acc: 0.9397\n",
      "Epoch 46/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1652 - acc: 0.9529 - val_loss: 0.2161 - val_acc: 0.9389\n",
      "Epoch 47/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1642 - acc: 0.9528 - val_loss: 0.2177 - val_acc: 0.9375\n",
      "Epoch 48/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.1640 - acc: 0.9529 - val_loss: 0.2152 - val_acc: 0.9388\n",
      "Epoch 49/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1636 - acc: 0.9532 - val_loss: 0.2172 - val_acc: 0.9390\n",
      "Epoch 50/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1626 - acc: 0.9534 - val_loss: 0.2175 - val_acc: 0.9378\n",
      "Epoch 51/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1616 - acc: 0.9536 - val_loss: 0.2149 - val_acc: 0.9392\n",
      "Epoch 52/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1610 - acc: 0.9536 - val_loss: 0.2155 - val_acc: 0.9395\n",
      "Epoch 53/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1604 - acc: 0.9540 - val_loss: 0.2147 - val_acc: 0.9389\n",
      "Epoch 54/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1600 - acc: 0.9540 - val_loss: 0.2189 - val_acc: 0.9377\n",
      "Epoch 55/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1593 - acc: 0.9538 - val_loss: 0.2183 - val_acc: 0.9397\n",
      "Epoch 56/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1588 - acc: 0.9546 - val_loss: 0.2154 - val_acc: 0.9396\n",
      "Epoch 57/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1584 - acc: 0.9542 - val_loss: 0.2214 - val_acc: 0.9381\n",
      "Epoch 58/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1580 - acc: 0.9550 - val_loss: 0.2226 - val_acc: 0.9376\n",
      "Epoch 59/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1576 - acc: 0.9545 - val_loss: 0.2193 - val_acc: 0.9390\n",
      "Epoch 60/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1567 - acc: 0.9545 - val_loss: 0.2215 - val_acc: 0.9383\n",
      "Epoch 61/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1565 - acc: 0.9550 - val_loss: 0.2188 - val_acc: 0.9402\n",
      "Epoch 62/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1555 - acc: 0.9555 - val_loss: 0.2257 - val_acc: 0.9374\n",
      "Epoch 63/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1556 - acc: 0.9554 - val_loss: 0.2175 - val_acc: 0.9388\n",
      "Epoch 64/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1550 - acc: 0.9552 - val_loss: 0.2189 - val_acc: 0.9392\n",
      "Epoch 65/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1551 - acc: 0.9551 - val_loss: 0.2235 - val_acc: 0.9383\n",
      "Epoch 66/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1546 - acc: 0.9552 - val_loss: 0.2233 - val_acc: 0.9375\n",
      "Epoch 67/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1540 - acc: 0.9561 - val_loss: 0.2202 - val_acc: 0.9395\n",
      "Epoch 68/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1534 - acc: 0.9564 - val_loss: 0.2214 - val_acc: 0.9386\n",
      "Epoch 69/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1535 - acc: 0.9564 - val_loss: 0.2205 - val_acc: 0.9393\n",
      "Epoch 70/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1529 - acc: 0.9561 - val_loss: 0.2220 - val_acc: 0.9392\n",
      "Epoch 71/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1526 - acc: 0.9558 - val_loss: 0.2219 - val_acc: 0.9401\n",
      "Epoch 72/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1518 - acc: 0.9564 - val_loss: 0.2225 - val_acc: 0.9370\n",
      "Epoch 73/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1519 - acc: 0.9563 - val_loss: 0.2241 - val_acc: 0.9378\n",
      "Epoch 74/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1519 - acc: 0.9566 - val_loss: 0.2221 - val_acc: 0.9391\n",
      "Epoch 75/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1515 - acc: 0.9568 - val_loss: 0.2256 - val_acc: 0.9387\n",
      "Epoch 76/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1510 - acc: 0.9566 - val_loss: 0.2233 - val_acc: 0.9388\n",
      "Epoch 77/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1506 - acc: 0.9566 - val_loss: 0.2223 - val_acc: 0.9384\n",
      "Epoch 78/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1504 - acc: 0.9567 - val_loss: 0.2265 - val_acc: 0.9379\n",
      "Epoch 79/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1499 - acc: 0.9571 - val_loss: 0.2280 - val_acc: 0.9377\n",
      "Epoch 80/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1500 - acc: 0.9567 - val_loss: 0.2234 - val_acc: 0.9381\n",
      "Epoch 81/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1492 - acc: 0.9570 - val_loss: 0.2230 - val_acc: 0.9389\n",
      "Epoch 82/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1495 - acc: 0.9564 - val_loss: 0.2280 - val_acc: 0.9381\n",
      "Epoch 83/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1489 - acc: 0.9575 - val_loss: 0.2262 - val_acc: 0.9384\n",
      "Epoch 84/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1483 - acc: 0.9579 - val_loss: 0.2272 - val_acc: 0.9372\n",
      "Epoch 85/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1484 - acc: 0.9573 - val_loss: 0.2273 - val_acc: 0.9379\n",
      "Epoch 86/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1481 - acc: 0.9581 - val_loss: 0.2266 - val_acc: 0.9371\n",
      "Epoch 87/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1478 - acc: 0.9574 - val_loss: 0.2272 - val_acc: 0.9371\n",
      "Epoch 88/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1481 - acc: 0.9580 - val_loss: 0.2284 - val_acc: 0.9362\n",
      "Epoch 89/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1477 - acc: 0.9578 - val_loss: 0.2261 - val_acc: 0.9372\n",
      "Epoch 90/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1471 - acc: 0.9580 - val_loss: 0.2283 - val_acc: 0.9381\n",
      "Epoch 91/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1468 - acc: 0.9578 - val_loss: 0.2275 - val_acc: 0.9368\n",
      "Epoch 92/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1467 - acc: 0.9579 - val_loss: 0.2319 - val_acc: 0.9376\n",
      "Epoch 93/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1465 - acc: 0.9573 - val_loss: 0.2288 - val_acc: 0.9378\n",
      "Epoch 94/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1459 - acc: 0.9586 - val_loss: 0.2294 - val_acc: 0.9382\n",
      "Epoch 95/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1458 - acc: 0.9581 - val_loss: 0.2313 - val_acc: 0.9378\n",
      "Epoch 96/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1457 - acc: 0.9587 - val_loss: 0.2306 - val_acc: 0.9382\n",
      "Epoch 97/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1455 - acc: 0.9585 - val_loss: 0.2290 - val_acc: 0.9373\n",
      "Epoch 98/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1452 - acc: 0.9583 - val_loss: 0.2364 - val_acc: 0.9364\n",
      "Epoch 99/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1451 - acc: 0.9587 - val_loss: 0.2325 - val_acc: 0.9370\n",
      "Epoch 100/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1451 - acc: 0.9584 - val_loss: 0.2291 - val_acc: 0.9378\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Model 1 - 10 neurons \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 100\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(10, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "    \n",
    "    # Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "    # Optimizer: Adam\n",
    "  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T14:58:15.181352Z",
     "start_time": "2018-02-23T14:31:52.623267Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/50\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.4341 - acc: 0.8824 - val_loss: 0.2462 - val_acc: 0.9303\n",
      "Epoch 2/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.2154 - acc: 0.9390 - val_loss: 0.1846 - val_acc: 0.9465\n",
      "Epoch 3/50\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1689 - acc: 0.9517 - val_loss: 0.1540 - val_acc: 0.9549\n",
      "Epoch 4/50\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1414 - acc: 0.9590 - val_loss: 0.1309 - val_acc: 0.9612\n",
      "Epoch 5/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1210 - acc: 0.9652 - val_loss: 0.1240 - val_acc: 0.9635\n",
      "Epoch 6/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1061 - acc: 0.9697 - val_loss: 0.1099 - val_acc: 0.9674\n",
      "Epoch 7/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0952 - acc: 0.9724 - val_loss: 0.1053 - val_acc: 0.9690\n",
      "Epoch 8/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0857 - acc: 0.9753 - val_loss: 0.1012 - val_acc: 0.9697\n",
      "Epoch 9/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.0770 - acc: 0.9775 - val_loss: 0.1001 - val_acc: 0.9690\n",
      "Epoch 10/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0699 - acc: 0.9797 - val_loss: 0.0921 - val_acc: 0.9724\n",
      "Epoch 11/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0638 - acc: 0.9816 - val_loss: 0.0921 - val_acc: 0.9725\n",
      "Epoch 12/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0581 - acc: 0.9834 - val_loss: 0.0883 - val_acc: 0.9721\n",
      "Epoch 13/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0541 - acc: 0.9843 - val_loss: 0.0904 - val_acc: 0.9711\n",
      "Epoch 14/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0496 - acc: 0.9856 - val_loss: 0.0920 - val_acc: 0.9712\n",
      "Epoch 15/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0454 - acc: 0.9869 - val_loss: 0.0877 - val_acc: 0.9736\n",
      "Epoch 16/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0420 - acc: 0.9882 - val_loss: 0.0906 - val_acc: 0.9731\n",
      "Epoch 17/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0387 - acc: 0.9894 - val_loss: 0.0905 - val_acc: 0.9719\n",
      "Epoch 18/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0356 - acc: 0.9900 - val_loss: 0.0886 - val_acc: 0.9720\n",
      "Epoch 19/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0332 - acc: 0.9907 - val_loss: 0.0922 - val_acc: 0.9718\n",
      "Epoch 20/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0308 - acc: 0.9916 - val_loss: 0.0932 - val_acc: 0.9726\n",
      "Epoch 21/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0279 - acc: 0.9926 - val_loss: 0.0904 - val_acc: 0.9726\n",
      "Epoch 22/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0261 - acc: 0.9932 - val_loss: 0.0931 - val_acc: 0.9730\n",
      "Epoch 23/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0242 - acc: 0.9941 - val_loss: 0.0944 - val_acc: 0.9729\n",
      "Epoch 24/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0223 - acc: 0.9944 - val_loss: 0.0943 - val_acc: 0.9726\n",
      "Epoch 25/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0214 - acc: 0.9948 - val_loss: 0.0934 - val_acc: 0.9720\n",
      "Epoch 26/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0192 - acc: 0.9954 - val_loss: 0.0962 - val_acc: 0.9712\n",
      "Epoch 27/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0176 - acc: 0.9959 - val_loss: 0.0995 - val_acc: 0.9711\n",
      "Epoch 28/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0174 - acc: 0.9960 - val_loss: 0.0987 - val_acc: 0.9732\n",
      "Epoch 29/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0155 - acc: 0.9965 - val_loss: 0.1023 - val_acc: 0.9718\n",
      "Epoch 30/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0141 - acc: 0.9970 - val_loss: 0.1016 - val_acc: 0.9723\n",
      "Epoch 31/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0135 - acc: 0.9972 - val_loss: 0.1001 - val_acc: 0.9724\n",
      "Epoch 32/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0123 - acc: 0.9975 - val_loss: 0.1076 - val_acc: 0.9719\n",
      "Epoch 33/50\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0116 - acc: 0.9977 - val_loss: 0.1124 - val_acc: 0.9718\n",
      "Epoch 34/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0101 - acc: 0.9980 - val_loss: 0.1028 - val_acc: 0.9718\n",
      "Epoch 35/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0104 - acc: 0.9978 - val_loss: 0.1143 - val_acc: 0.9716\n",
      "Epoch 36/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0095 - acc: 0.9981 - val_loss: 0.1060 - val_acc: 0.9748\n",
      "Epoch 37/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0085 - acc: 0.9984 - val_loss: 0.1174 - val_acc: 0.9707\n",
      "Epoch 38/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0082 - acc: 0.9985 - val_loss: 0.1198 - val_acc: 0.9701\n",
      "Epoch 39/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0079 - acc: 0.9986 - val_loss: 0.1150 - val_acc: 0.9721\n",
      "Epoch 40/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0063 - acc: 0.9991 - val_loss: 0.1163 - val_acc: 0.9718\n",
      "Epoch 41/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0068 - acc: 0.9988 - val_loss: 0.1223 - val_acc: 0.9709\n",
      "Epoch 42/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0057 - acc: 0.9991 - val_loss: 0.1226 - val_acc: 0.9719\n",
      "Epoch 43/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0048 - acc: 0.9994 - val_loss: 0.1202 - val_acc: 0.9726\n",
      "Epoch 44/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0050 - acc: 0.9993 - val_loss: 0.1226 - val_acc: 0.9715\n",
      "Epoch 45/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0066 - acc: 0.9985 - val_loss: 0.1284 - val_acc: 0.9705\n",
      "Epoch 46/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0043 - acc: 0.9996 - val_loss: 0.1252 - val_acc: 0.9716\n",
      "Epoch 47/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0033 - acc: 0.9998 - val_loss: 0.1247 - val_acc: 0.9736\n",
      "Epoch 48/50\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0058 - acc: 0.9988 - val_loss: 0.1351 - val_acc: 0.9695\n",
      "Epoch 49/50\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0045 - acc: 0.9991 - val_loss: 0.1265 - val_acc: 0.9723\n",
      "Epoch 50/50\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.0033 - acc: 0.9996 - val_loss: 0.1380 - val_acc: 0.9700\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Model 2 - 100 neurons \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 50\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(50, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "    \n",
    "    # Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "    # Optimizer: Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.3247 - acc: 0.9107 - val_loss: 0.1781 - val_acc: 0.9494\n",
      "Epoch 2/50\n",
      "60000/60000 [==============================] - 2s 40us/step - loss: 0.1443 - acc: 0.9578 - val_loss: 0.1294 - val_acc: 0.9607\n",
      "Epoch 3/50\n",
      "60000/60000 [==============================] - 3s 43us/step - loss: 0.0996 - acc: 0.9708 - val_loss: 0.0965 - val_acc: 0.9705\n",
      "Epoch 4/50\n",
      "60000/60000 [==============================] - 2s 37us/step - loss: 0.0754 - acc: 0.9781 - val_loss: 0.0801 - val_acc: 0.9750\n",
      "Epoch 5/50\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.0594 - acc: 0.9828 - val_loss: 0.0759 - val_acc: 0.9762\n",
      "Epoch 6/50\n",
      "60000/60000 [==============================] - 2s 37us/step - loss: 0.0473 - acc: 0.9864 - val_loss: 0.0789 - val_acc: 0.9753\n",
      "Epoch 7/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0386 - acc: 0.9893 - val_loss: 0.0718 - val_acc: 0.9780\n",
      "Epoch 8/50\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.0317 - acc: 0.9911 - val_loss: 0.0698 - val_acc: 0.9792\n",
      "Epoch 9/50\n",
      "60000/60000 [==============================] - 2s 38us/step - loss: 0.0247 - acc: 0.9937 - val_loss: 0.0669 - val_acc: 0.9779\n",
      "Epoch 10/50\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.0208 - acc: 0.9946 - val_loss: 0.0658 - val_acc: 0.9797\n",
      "Epoch 11/50\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.0171 - acc: 0.9957 - val_loss: 0.0682 - val_acc: 0.9790\n",
      "Epoch 12/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0136 - acc: 0.9969 - val_loss: 0.0710 - val_acc: 0.9799\n",
      "Epoch 13/50\n",
      "60000/60000 [==============================] - 2s 37us/step - loss: 0.0114 - acc: 0.9976 - val_loss: 0.0722 - val_acc: 0.9798\n",
      "Epoch 14/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.0099 - acc: 0.9978 - val_loss: 0.0742 - val_acc: 0.9793\n",
      "Epoch 15/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0069 - acc: 0.9989 - val_loss: 0.0732 - val_acc: 0.9796\n",
      "Epoch 16/50\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.0069 - acc: 0.9988 - val_loss: 0.0696 - val_acc: 0.9814\n",
      "Epoch 17/50\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.0055 - acc: 0.9991 - val_loss: 0.0780 - val_acc: 0.9794\n",
      "Epoch 18/50\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.0055 - acc: 0.9989 - val_loss: 0.0818 - val_acc: 0.9795\n",
      "Epoch 19/50\n",
      "60000/60000 [==============================] - 2s 38us/step - loss: 0.0051 - acc: 0.9991 - val_loss: 0.0767 - val_acc: 0.9806\n",
      "Epoch 20/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0057 - acc: 0.9986 - val_loss: 0.0792 - val_acc: 0.9806\n",
      "Epoch 21/50\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.0041 - acc: 0.9992 - val_loss: 0.0887 - val_acc: 0.9786\n",
      "Epoch 22/50\n",
      "60000/60000 [==============================] - 2s 38us/step - loss: 0.0025 - acc: 0.9997 - val_loss: 0.0773 - val_acc: 0.9812\n",
      "Epoch 23/50\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.0012 - acc: 0.9999 - val_loss: 0.0856 - val_acc: 0.9809\n",
      "Epoch 24/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0021 - acc: 0.9996 - val_loss: 0.0916 - val_acc: 0.9793\n",
      "Epoch 25/50\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.0094 - acc: 0.9970 - val_loss: 0.0849 - val_acc: 0.9805\n",
      "Epoch 26/50\n",
      "60000/60000 [==============================] - 2s 30us/step - loss: 0.0019 - acc: 0.9998 - val_loss: 0.0811 - val_acc: 0.9813\n",
      "Epoch 27/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 6.9606e-04 - acc: 1.0000 - val_loss: 0.0806 - val_acc: 0.9816\n",
      "Epoch 28/50\n",
      "60000/60000 [==============================] - 2s 30us/step - loss: 4.4268e-04 - acc: 1.0000 - val_loss: 0.0809 - val_acc: 0.9823\n",
      "Epoch 29/50\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 3.5583e-04 - acc: 1.0000 - val_loss: 0.0817 - val_acc: 0.9820\n",
      "Epoch 30/50\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 8.8624e-04 - acc: 0.9999 - val_loss: 0.1047 - val_acc: 0.9773\n",
      "Epoch 31/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.0155 - acc: 0.9947 - val_loss: 0.0876 - val_acc: 0.9817\n",
      "Epoch 32/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.0016 - acc: 0.9998 - val_loss: 0.0875 - val_acc: 0.9820\n",
      "Epoch 33/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 4.5504e-04 - acc: 1.0000 - val_loss: 0.0839 - val_acc: 0.9833\n",
      "Epoch 34/50\n",
      "60000/60000 [==============================] - 2s 30us/step - loss: 2.8884e-04 - acc: 1.0000 - val_loss: 0.0843 - val_acc: 0.9835\n",
      "Epoch 35/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 2.3999e-04 - acc: 1.0000 - val_loss: 0.0845 - val_acc: 0.9836\n",
      "Epoch 36/50\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 2.0866e-04 - acc: 1.0000 - val_loss: 0.0853 - val_acc: 0.9834\n",
      "Epoch 37/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 1.8248e-04 - acc: 1.0000 - val_loss: 0.0862 - val_acc: 0.9829\n",
      "Epoch 38/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 1.6290e-04 - acc: 1.0000 - val_loss: 0.0861 - val_acc: 0.9834\n",
      "Epoch 39/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 1.4625e-04 - acc: 1.0000 - val_loss: 0.0868 - val_acc: 0.9829\n",
      "Epoch 40/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 1.3288e-04 - acc: 1.0000 - val_loss: 0.0877 - val_acc: 0.9831\n",
      "Epoch 41/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 1.1391e-04 - acc: 1.0000 - val_loss: 0.0878 - val_acc: 0.9830\n",
      "Epoch 42/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.0101 - acc: 0.9969 - val_loss: 0.1139 - val_acc: 0.9744\n",
      "Epoch 43/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.0084 - acc: 0.9970 - val_loss: 0.0995 - val_acc: 0.9804\n",
      "Epoch 44/50\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 6.9188e-04 - acc: 0.9999 - val_loss: 0.0952 - val_acc: 0.9813\n",
      "Epoch 45/50\n",
      "60000/60000 [==============================] - 2s 37us/step - loss: 2.2105e-04 - acc: 1.0000 - val_loss: 0.0950 - val_acc: 0.9817\n",
      "Epoch 46/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 3.9783e-04 - acc: 0.9999 - val_loss: 0.0927 - val_acc: 0.9818\n",
      "Epoch 47/50\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 1.3171e-04 - acc: 1.0000 - val_loss: 0.0930 - val_acc: 0.9816\n",
      "Epoch 48/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 1.0770e-04 - acc: 1.0000 - val_loss: 0.0937 - val_acc: 0.9817\n",
      "Epoch 49/50\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 9.4018e-05 - acc: 1.0000 - val_loss: 0.0944 - val_acc: 0.9814\n",
      "Epoch 50/50\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 8.2268e-05 - acc: 1.0000 - val_loss: 0.0943 - val_acc: 0.9817\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Model 3 - 100 neurons \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 50\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(200, activation='relu', input_dim=784))\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "# Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "# Optimizer: Adam"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 2:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/50\n",
      "60000/60000 [==============================] - 5s 89us/step - loss: 0.2697 - acc: 0.9230 - val_loss: 0.1452 - val_acc: 0.9588\n",
      "Epoch 2/50\n",
      "60000/60000 [==============================] - 4s 71us/step - loss: 0.1092 - acc: 0.9681 - val_loss: 0.0915 - val_acc: 0.9732\n",
      "Epoch 3/50\n",
      "60000/60000 [==============================] - 4s 74us/step - loss: 0.0709 - acc: 0.9785 - val_loss: 0.0780 - val_acc: 0.9757\n",
      "Epoch 4/50\n",
      "60000/60000 [==============================] - 4s 75us/step - loss: 0.0512 - acc: 0.9850 - val_loss: 0.0734 - val_acc: 0.9767\n",
      "Epoch 5/50\n",
      "60000/60000 [==============================] - 4s 67us/step - loss: 0.0376 - acc: 0.9888 - val_loss: 0.0679 - val_acc: 0.9781\n",
      "Epoch 6/50\n",
      "60000/60000 [==============================] - 4s 61us/step - loss: 0.0286 - acc: 0.9918 - val_loss: 0.0612 - val_acc: 0.9815\n",
      "Epoch 7/50\n",
      "60000/60000 [==============================] - 4s 59us/step - loss: 0.0203 - acc: 0.9945 - val_loss: 0.0652 - val_acc: 0.9808\n",
      "Epoch 8/50\n",
      "60000/60000 [==============================] - 4s 59us/step - loss: 0.0152 - acc: 0.9963 - val_loss: 0.0621 - val_acc: 0.9810\n",
      "Epoch 9/50\n",
      "60000/60000 [==============================] - 3s 58us/step - loss: 0.0125 - acc: 0.9970 - val_loss: 0.0657 - val_acc: 0.9804\n",
      "Epoch 10/50\n",
      "60000/60000 [==============================] - 3s 58us/step - loss: 0.0114 - acc: 0.9971 - val_loss: 0.0682 - val_acc: 0.9803\n",
      "Epoch 11/50\n",
      "60000/60000 [==============================] - 4s 61us/step - loss: 0.0094 - acc: 0.9975 - val_loss: 0.0777 - val_acc: 0.9784\n",
      "Epoch 12/50\n",
      "60000/60000 [==============================] - 4s 62us/step - loss: 0.0067 - acc: 0.9985 - val_loss: 0.0684 - val_acc: 0.9804\n",
      "Epoch 13/50\n",
      "60000/60000 [==============================] - 4s 63us/step - loss: 0.0061 - acc: 0.9987 - val_loss: 0.0704 - val_acc: 0.9820\n",
      "Epoch 14/50\n",
      "60000/60000 [==============================] - 4s 59us/step - loss: 0.0039 - acc: 0.9993 - val_loss: 0.0742 - val_acc: 0.9809\n",
      "Epoch 15/50\n",
      "60000/60000 [==============================] - 4s 61us/step - loss: 0.0033 - acc: 0.9994 - val_loss: 0.0800 - val_acc: 0.9798\n",
      "Epoch 16/50\n",
      "60000/60000 [==============================] - 4s 69us/step - loss: 0.0077 - acc: 0.9979 - val_loss: 0.0836 - val_acc: 0.9788\n",
      "Epoch 17/50\n",
      "60000/60000 [==============================] - 3s 57us/step - loss: 0.0073 - acc: 0.9980 - val_loss: 0.0802 - val_acc: 0.9801\n",
      "Epoch 18/50\n",
      "60000/60000 [==============================] - 3s 53us/step - loss: 0.0046 - acc: 0.9989 - val_loss: 0.0785 - val_acc: 0.9803\n",
      "Epoch 19/50\n",
      "60000/60000 [==============================] - 3s 52us/step - loss: 0.0016 - acc: 0.9997 - val_loss: 0.0716 - val_acc: 0.9834\n",
      "Epoch 20/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 5.3941e-04 - acc: 1.0000 - val_loss: 0.0712 - val_acc: 0.9833\n",
      "Epoch 21/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0086 - acc: 0.9972 - val_loss: 0.0904 - val_acc: 0.9796\n",
      "Epoch 22/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0054 - acc: 0.9984 - val_loss: 0.0787 - val_acc: 0.9819\n",
      "Epoch 23/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0025 - acc: 0.9995 - val_loss: 0.0793 - val_acc: 0.9826\n",
      "Epoch 24/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0034 - acc: 0.9991 - val_loss: 0.0862 - val_acc: 0.9798\n",
      "Epoch 25/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0030 - acc: 0.9993 - val_loss: 0.0934 - val_acc: 0.9803\n",
      "Epoch 26/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 0.0052 - acc: 0.9983 - val_loss: 0.0848 - val_acc: 0.9826\n",
      "Epoch 27/50\n",
      "60000/60000 [==============================] - 3s 52us/step - loss: 0.0019 - acc: 0.9995 - val_loss: 0.0769 - val_acc: 0.9823\n",
      "Epoch 28/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 9.7123e-04 - acc: 0.9998 - val_loss: 0.0760 - val_acc: 0.9847\n",
      "Epoch 29/50\n",
      "60000/60000 [==============================] - 3s 53us/step - loss: 1.5558e-04 - acc: 1.0000 - val_loss: 0.0759 - val_acc: 0.9844\n",
      "Epoch 30/50\n",
      "60000/60000 [==============================] - 3s 57us/step - loss: 9.1040e-05 - acc: 1.0000 - val_loss: 0.0761 - val_acc: 0.9855\n",
      "Epoch 31/50\n",
      "60000/60000 [==============================] - 3s 52us/step - loss: 7.1278e-05 - acc: 1.0000 - val_loss: 0.0772 - val_acc: 0.9851\n",
      "Epoch 32/50\n",
      "60000/60000 [==============================] - 3s 55us/step - loss: 6.0907e-05 - acc: 1.0000 - val_loss: 0.0779 - val_acc: 0.9854\n",
      "Epoch 33/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 5.1603e-05 - acc: 1.0000 - val_loss: 0.0782 - val_acc: 0.9849\n",
      "Epoch 34/50\n",
      "60000/60000 [==============================] - 3s 57us/step - loss: 4.4798e-05 - acc: 1.0000 - val_loss: 0.0794 - val_acc: 0.9851\n",
      "Epoch 35/50\n",
      "60000/60000 [==============================] - 3s 54us/step - loss: 3.8175e-05 - acc: 1.0000 - val_loss: 0.0799 - val_acc: 0.9849\n",
      "Epoch 36/50\n",
      "60000/60000 [==============================] - 4s 61us/step - loss: 3.3537e-05 - acc: 1.0000 - val_loss: 0.0801 - val_acc: 0.9850\n",
      "Epoch 37/50\n",
      "60000/60000 [==============================] - 4s 67us/step - loss: 2.8808e-05 - acc: 1.0000 - val_loss: 0.0806 - val_acc: 0.9849\n",
      "Epoch 38/50\n",
      "60000/60000 [==============================] - 3s 55us/step - loss: 2.5062e-05 - acc: 1.0000 - val_loss: 0.0820 - val_acc: 0.9848\n",
      "Epoch 39/50\n",
      "60000/60000 [==============================] - 3s 57us/step - loss: 2.3634e-05 - acc: 1.0000 - val_loss: 0.0823 - val_acc: 0.9850\n",
      "Epoch 40/50\n",
      "60000/60000 [==============================] - 3s 58us/step - loss: 0.0195 - acc: 0.9949 - val_loss: 0.1034 - val_acc: 0.9778\n",
      "Epoch 41/50\n",
      "60000/60000 [==============================] - 3s 53us/step - loss: 0.0050 - acc: 0.9984 - val_loss: 0.0931 - val_acc: 0.9807\n",
      "Epoch 42/50\n",
      "60000/60000 [==============================] - 3s 48us/step - loss: 0.0015 - acc: 0.9996 - val_loss: 0.0867 - val_acc: 0.9820\n",
      "Epoch 43/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 3.3663e-04 - acc: 1.0000 - val_loss: 0.0863 - val_acc: 0.9836\n",
      "Epoch 44/50\n",
      "60000/60000 [==============================] - 3s 51us/step - loss: 1.0575e-04 - acc: 1.0000 - val_loss: 0.0851 - val_acc: 0.9834\n",
      "Epoch 45/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 7.2370e-05 - acc: 1.0000 - val_loss: 0.0856 - val_acc: 0.9837\n",
      "Epoch 46/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 5.7780e-05 - acc: 1.0000 - val_loss: 0.0860 - val_acc: 0.9840\n",
      "Epoch 47/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 4.8393e-05 - acc: 1.0000 - val_loss: 0.0866 - val_acc: 0.9838\n",
      "Epoch 48/50\n",
      "60000/60000 [==============================] - 3s 48us/step - loss: 4.0893e-05 - acc: 1.0000 - val_loss: 0.0872 - val_acc: 0.9839\n",
      "Epoch 49/50\n",
      "60000/60000 [==============================] - 3s 48us/step - loss: 3.4956e-05 - acc: 1.0000 - val_loss: 0.0872 - val_acc: 0.9839\n",
      "Epoch 50/50\n",
      "60000/60000 [==============================] - 3s 50us/step - loss: 2.9575e-05 - acc: 1.0000 - val_loss: 0.0879 - val_acc: 0.9840\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEdCAYAAABZtfMGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXd4HNXVuN+z6l2yJblI7g0bbIwt040xoXfjQOgtkAqEdEh+HyQEQvKFkC8BQoDQCTHEJGBawBTbdGzj3rstF1ku6nV3z++POyutZJVRWe2ufN/nmcfambl3zsqjOXPOPUVUFYvFYrFYIg1PuAWwWCwWi6UlrIKyWCwWS0RiFZTFYrFYIhKroCwWi8USkVgFZbFYLJaIxCooi8VisUQkVkFZLN2AiKwSkVPDLYfF0puwCspicYGIbBWR05vtu15EPgZQ1SNVdV47cwwVERWR2BCKarH0GqyCslgsFktEYhWUxdINBFtYInKsiCwSkTIRKRKRB53TFjj/lohIhYicEB5pLZbowLoaLJbu58/An1X1eRFJBY5y9p8CbAEyVdUbNukslijBKiiLxT2vikiwYokHvmrhvHpgpIhkq+o+4PMekc5i6WVYF5/F4p6LVTUzsAHfa+W8bwKjgbUislBEzu85ES2W3oO1oCyWbkZVNwBXiIgHuASYLSJ9Ads6wGLpANaCsli6GRG5WkRyVNUPlDi7fUAx4AeGh004iyWKsArKYul+zgZWiUgFJmDiclWtUdUq4D7gExEpEZHjwyqlxRLhiG1YaLFYLJZIxFpQFovFYolIrIKyWCwWS0RiFZTFYrFYIhKroCwWi8USkVgFZbFYLJaIxCooi8VisUQkVkFZLBaLJSKxCspisVgsEYlVUBaLxWKJSKyCslgsFktEYhWUxWKxWCKSXtluIzs7W4cOHRpuMSy9lMWLF+9T1ZxwyxFK7N+QJZS4/RvqlQpq6NChLFq0KNxiWHopIrIt3DKEGvs3ZAklbv+GepWLT0QuEJHHS0tLwy2KxWKxWLpIr1JQqvq6qn4rIyMj3KJYLBaLpYv0KgVlsVgslt6DVVAWi8ViiUisgrJYLBZLRGIVlMUSBYjIUyKyV0RWtnJcROQvIrJRRJaLyKSgY9eJyAZnu67npLZYukavUlA2is/Si3kGOLuN4+cAo5ztW8CjACLSB7gbOA44FrhbRLJCKqnF0k30qjwoVX0deL2goODm5sf8fuWZT7cyul8aJ4/KDoN0FkvnUdUFIjK0jVMuAp5TVQU+F5FMERkAnArMVdUDACIyF6Po/hlaiaOHPaU1rN1TxrHD+pAc3/2PxLKaejYUVbChqJwDVXXdPn9bjMhJ5awj+7d73v6KWr7YcoCy6noq63zUen1dvnZ8jIebpg7v0hy9SkG1hccjPPzhRs46sp9VUJbeSB6wI+hzobOvtf2HICLfwlhfDB48ODRSRgiVtV7uf3sNc1cXUVRWC8C3pw3nznPGtjpGVRGRdudWVT7asI/564v5aEMx64squk3ujhLjEVb86swWFa+qMmvhDl5bupMvtxzAr9177ZT4GKugOsKw7BQ2F1eGWwyLJRS09OTUNvYfulP1ceBxgIKCgm5+XIWG3aXVPDZ/Mxv3VnCgsg6v389d5x/Z5kvoxr0VfPeFxWwqruDc8QOYPCSLuauLmL2okB+fMYb42KYrH9V1Pm6btYTPN+1n+hG5nDt+ABlJcSzdUcK6PWWcN2EgZ4zrB5iH/q9fX80zn24lPtbDccP6cNHEPMb0S2N0vzRy0xNC+vsIZv76Yr79/GJWFJZy3PC+hxx/cO56HvpgI6NyU/n+9JGcdkQu/dITSYmPJSEuMlZ/DjsF9dGG4nCLYbGEgkJgUNDnfGCXs//UZvvn9ZhUIeJAZR1//2gzT368BVU4Mi+dARmJrN9bzm2zlvDWbVPpn5F4yLi3Vuzmp/9aRkJcDM/deFyDIhvaN4UbnlnIe2uKOHf8gIbzS6vqufHZhSzZfpCzj+rPRxuKmbNsV8PxtMRY5izbxf2XjOcbUwbz2ILNPPPpVm44aSg/P/sIEuNiQv/LaIWCIWapcVlhySEK6oXPt/HQBxu5fMog7r9kvCvLMBwcdgpq9uJCKmu9pCQcVl/d0vuZA9wiIrMwARGlqrpbRN4BfhsUGHEmcGe4hOwKc1cX8eynW1lfVM7ecuOWm3FMHj86YzSD+iQDxjq64KGP+cGsJbx48/HEeMyD1+vz84d31vHYgs1MHJTJX6+axMDMpIa5Txmdw8CMRP755fYGBVVcXss1T37B5uJKHrlyEueMH4DX5+eLLQeo8/k5Oj+TpLgYvv3CYn7+ygo+2bifOct2ccHRA/mf88bh8YT3od83NYFBfZJYuqOkyf53V+3hrtdWctoRudx78VERq5zgMFRQAFv2VXJUni2HZIkeROSfGEsoW0QKMZF5cQCq+jfgLeBcYCNQBdzgHDsgIr8BFjpT3RMImIgmXlq4nTv/vYLBfZKZOiqH0f1SmToqh3ED05ucNzI3ld9cfBQ/+dcy/vjuOs6bMIB9FXU8vmATn2zcz9XHD+Z/zh9HQmxTyybGI1xaMIi/fLCBHQeq6Jsaz43PLGTb/iqevL6AqaNM4e3YGA8njWzqPnzi2sncPmspc5bt4oThfXng0glhV04Bjs7PZMn2RgVV7/Pzs1eWc1ReBg9feQyxMZHhymsNq6AslihAVa9o57gC32/l2FPAU6GQqyf4+0ebuffNNZwyOofHrp5MUnzbbrOvT87n0037+Ou8Tfx13iYA4mM9/O/XJ3BZwaBWx102xSiof365nfVF5azaVcoT1zYqp9ZIiI3hoSuO4d3VRZwyOucQ5RdOJg7K5I3lu9lbXkNuWiJfbjlASVU9/ztzZEgiFrubyJewGxna1yiorftsoITFEqn4/crmfZXMXV3Ef1ftYdmOEs45qj//d/lE1w//384Yz7TROcTHeMhOS2Bo3xRy0toOUMjLTGLa6Bwenb8JVfjNRUfytbH9XF0vNsbTZO0qUpg4KBOAZTtKOWNcIu+u2kNinKddpRsp9CoFJSIXABeMHDmyxeNJ8TEMyEhki1VQFktEUXiwiscXbGbZjhI27K2gqs7k4UzIz+CX547lhpOGdsgdlRgXw0UTW4ymb5OrjxvCvHXF3Dx1GNecMLTD4yONo/IyiPEIS3cc5PSxuby7uoipo3LatUIjhV6loNpK1A0wLDuFzVZBWSwRQWlVPY/M28gzn2wFgSlDs7isYBBj+qdxyugc8oICGXqC08f1Y+4PT2FETmqPXjdUJMbFcET/NJbuKGHlzjJ2l9bwozNGh1ss1/QqBeWGYdkpvLlid7jFsFgOe2rqfcz46yds2V/JzEn5/OiM0U0i68LFqH5p4RahW5k4KJM5S3fxzqo9eATXbstIILJDOELAsOwUSqrqOVjZsyVHLBZLU55YsJnN+yp56ropPHDp0RGhnHojEwdlUl7r5YUvtjFlaB/6pMSHWyTXHJYKCrBuPosljBQerOKReRs5d3x/ph+RG25xejWBQImSqnrOdFGXL5I4fBSUrx5Wz+EI33rARvJZLOHkt2+tAeCX540LsyS9nxE5qaQ6hQnOHBc97j04nBSUeOD12xiw/gViPGIj+SyWMPHxhn28tWIPt0wf2eNBEIcjHo9w7LA+TMjPaKi4ES0cPkESnhgY8TU8m95ncOZlVkFZLGHij3PXMahPUpcrXVvc86dvTMTf3eXKe4DDx4ICGHUmVBZzavouuwZlsYSBVbtKWbK9hOtPHBbWQqqHGxlJcWRFUXBEgMNLQY38GiBMk6Vs3VeJqQ5jsVh6in98sZ3EOA9fn5QfblEsUcDhpaBSsiFvEkdWfUF1va+hUZnFYgk95TX1vLpkJxdMGEhGcly4xbFEARGvoERkuIg8KSKzu2XCUWeSXbqSLMrYvC98nS4tlsONV5fspKrOx1XHDwm3KJYoIaQKSkSeEpG9IrKy2f6zRWSdiGwUkTvamkNVN6vqN7tNqFFnICineJazamdZt01rCSEVxeD3h1sKSxdQVf7xxXaOykvn6HzbScDijlBbUM8AZwfvEJEY4BHgHGAccIWIjBOR8SLyRrOt+zP4BhwDydlclLKK99YUdfv0lm5m93J4cCy8MAOqoq6NkcVh8baDrN1TztXHDYnoBnmWyCKkCkpVFwDNnyrHAhsdy6gOmAVcpKorVPX8Zttet9cSkW+JyCIRWVRc3EZbd48HRp7O8bqUxVv3Hd4ljyI9SEQV3v45xCXDtk/hsWmwe1m4pbJ0gpcW7iA1IZYLJw4MtyiWKCIca1B5wI6gz4XOvhYRkb4i8jfgGBFptVW1qj6uqgWqWpCT006vk1FnkOwtZSLr+WCtax0Y+VQdgL1r2z+vch+8dgvcPwi+eCxyFdWqf8P2T+GMX8ON/wX1wZNnwnu/huqS9sdbIoLqOh9vr9zDOUf1j4omeZbIIRx3S0v2fatPSFXdD3zH1cTt9INqYOTpaFIWf+BJ/ryigJmTe0HIqyrMuhIKF8LlL8Losw49x++Dxc/A+/dAXQX0OxLe/hls+gAuesREOUYKdVXw7l3QfwJMutYkWn9rPrxzJ3z8ICx6EgpuhNT+EJcEiemQNsBsMXFmfH0VeGugvtr8W1sOtWVQWwF1leZ34K0F9QNqfq4pNeeMPgdO+F64fwtNEJGzgT8DMcDfVfV3zY4PwXTOzcF4Lq5W1ULn2O+B85xTf6OqL/WU3O+u3kNFrZdLbGi5pYOEQ0EVAsF9l/OBXd0xsZt+UAAkZSKXPc/gZy9i5pb/oaZ2LokJbXfb7BFUYediWPUf81AdfTYMOwXiEtsfu2YObP8MkrPhpWvgypdgxPTG43tWwOu3w85FMHQqnPsA5IwxFtTc/4FHT4JLHofh00L3/VpDFSqK4MAWqD5g6iZueh/KCmHmE0Y5AaTmwMy/w0k/MEr24z917bpxyRCbYMpgIRCbaBRdQjpE2DpJ0NrtGZi/oYUiMkdVVwed9gDwnKo+KyKnAfcD14jIecAkYCKQAMwXkbdVtUeihP6zZCd5mUkcN6xPT1zO0osIh4JaCIwSkWHATuBy4Moel2LYVDYeew9Tv/wlO2b/mEFXPtT9D6X6GjNnbDvKz1sHC5+Az/8GpdshJt5si56EuBTIL4DcsWYbdxEkZTUbXwtz74LccXDtHHjuImNNTf2xediX7oBls8y4GY/DhMsav+vx34GhJ8HsG824U34C0+6AGJe3Rn21sTpqSh2LpA58tcaNWLYTynYZd1xdRdBx55z6arNVHzTWTnOOvhKGnHjo/v7j4ap/mbF1VeB15igvgvLd4PdCfIqxrOKSjOKJTYKENLPFp5jNE1WVDBrWbgFEZBZwERCsoMYBP3R+/hB4NWj/fFX1Al4RWYYJXno51EIXl9fy0YZ9fPuU4Xg8kaX0LZFPSBWUiPwTOBXIFpFC4G5VfVJEbgHewbgqnlLVVd10PXcuPoehZ36XZ7/4lOs2PA+PfgUn3gpHzYTYbigJcnAbPHuBUTTXvgYZLSyz+f2w9nWYezcc3GKspem/gDHnmAfrlo9g3Vuwawl89Zx5iM/7nXHHjfxa4zxfPgEHt8LVrxgr49pXzbU/+I05npAOx1wNp/8Kklt4i+0/Hr41z7j7FvwBFj0N6QMhtZ+RQzxGofl95uHvrYGy3Ub51Ja2/XuIT4PkLIhPNUohNhHi0iEmwVEgyZCYAX2GQdYw42aMiTO/t77t/D8GFBBARr75Hr2XltZuj2t2zjJgJsYNOANIE5G+zv67ReRBIBmYTlPFBphAI+BbAIMHD+4Woecs24XPr1wyqePt1y0W6Y3lfgoKCnTRokWuzv3+CwvJ2fIqd/f9ANm7GtLz4cx74MhLOm9RHdgMz15o1jJUjeVy3RzIGmqOl+2GZS/CV88bxZQzFs66F0ae3vqcfr9xz712C+xbZ9ZlBh1nHvhv/gjyCuCafwed7zNWTWJGxyyF1XNgw7tQsde43bw15juoHzyxZouJM2s9GXmQ1h8SM8114lOMYolNgKQ+5nhi78t5EZHFqlrQw9e8FDhLVW9yPl8DHKuqtwadMxB4GBgGLMAoqyNVtVREfglcChQDe4EvVfXPrV2vI39DbXHeXz4ixiPMueXkLs9l6T24/Rs67ENqzjhyILevPJGzr/whx/uXmLWN2TfCwifh3D+YQAK3+Lwm6uzf3zYP9uteNxbH85fA0+fC8OlQ+CXsMz2pGHKysZiOvKR9l5rHA4OOhW/Phw/uhc8eMVYVgMTAmb9pdn5My9ZSe4y70GyWSKPdtVtV3QVcAiAiqcBMVS11jt0H3OccexHYEGqB1xeVs2pXGXdfYHs+WTpHr1JQHXXxAZx5ZD/SEmN58csdHH/FGTDiNFjyvFFUj0+H8x4w1gqYNY91b0Of4TDg6Ea31/p3YNk/YfN84/JKyTHKqf9RZtz1b8KLlxl33aBjYcI34MgZ0HdEx79kXBKcdR+c8lOoKTEyxSU1WmeW3kq7a7cikg0cUFU/cCcmoi8QYJGpqvtFZAIwAXg31ALPXW0S4c+bMCDUl7L0UnqVgnIdxRdEcnwsMyfl848vtrGvYhzZqQkw+Xo44gJ45UaYcysULjJK6bNHoNLJm8odZyyitW9AyTazXnPkRcZNN3y6iQYL0P8ouH2lUWjdFYiRlGk2y2GBqnpbWrsVkXuARao6B7Pee7+IKMbF931neBzwkVPBoQwTfu4Ntczz1u3lqLx0ctNcRKFaLC3QqxRUZ7n6+ME88+lWXl60g++d6lhfKX3h6n8bd9rHD5p9I06DE24xCmnZLPj8ERhyEpxxDxxxfttuOk/E1+W1RDiq+hbwVrN9dwX9PBs4pKiyqtZgIvl6jNKqer7aXsL3Tu2El8BicehVCqozLj6AkblpHD+8Dy9+sZ1vnzKCmEA4rCcGTr/bWEXxyTDwmMZBBTc2utcsFksTPtpYjM+vnDqmnaouFksb9KrXelV9XVW/lZHR8cixq48fQuHBahasb6GO39CTmiqnAFY5WSwtMm9dMRlJcUwclNX+yRZLK/QqBdUVzhzXn+zUBF74fFu4RbFYohq/X5m/vpipo7IbvREWSyewCsohPtbD5VMG8cG6vew40EJVA4vF4orVu8soLq9l+pju75ZjObzoVQpKRC4QkcdLS9upbtAKVx0/GI8Iz1srymLpNPPWmUjXU0bb9adWObDZJOxb2qTdIAkR6aOqUdEprjNh5sEMyEji7KP6M+vL7dx++ijbGsBi6QTz1hUzPi+DnLQIKMAciWyeBy9ebpL5h08zifplu0xXgaJVMPZ8U3at/3iTZ3lgsylzlj3GXeHoXoSbJ/AXIrIUeBp4W3tjbaQgbjhxKG8u382/v9rJ1ccPCbc4FktUUVpdz1fbD3LL9I5F0vZavLVQWmjyKEUalVOfYSY1ZflL8Pptpt7lwEmmGPSaOWZ/9mgzNlBIWWIge5QJzqoohspi83NqLqTkmn9Tc035s+J1ULQSkvuaOpsRVp3fLW4U1GjgdOBG4CEReQl4RlXXh1SyMDF5SBZH5aXzzKdbueq4wbY9tcXSAdbsLsOvMGmIjd4D4N3/gS8fM4n8Q0+GtW8Z5XTd66Yw8vRfGKspI6+xS8HZvzV927Z+YlJc+h1l0lyKVpnNW2vqd6b0NR0TKvcahbVnuamhWVdpFFlSFuxeajoLpEanu7VdBeVYTHOBuSIyHXgB+J5Tsv8OVf0sxDL2KCLC9ScO4yf/WsYnG/dz8qgIauJnsUQ464vKARjTPy3MkkQAlftNvcxh00z5s80fmtqeV77U2BxUpLEkWoCkLDj5h2YL5sgZ7q6raubd8B78Yybs39B1BVW5H7YuMHP3O8op0yamf1vVfuOKBNNKp3KfUZr1VTDlpi5d1s0aVF/gauAaoAi4FZiDaX72L0zl5Iigs4m6zTl/wgDuf2sNT3+yxSooi6UDrC8qJy0xlv7ph9daSYssfML0Kjv3D6Y5aEBxhJrANbKd5+C+DS33VQvGW2dci+l5RonGJprC1ps+MNuupTRpfB6TYAphq6/1OWMSoOCbXfrOblx8nwHPAxcH2kc7LBKRv3X6yiGgq0ESARLjYrjmhCH833sbWLO7jLED0tsfZLFYWL+ngjH90qLDNf7KzcYFltrPbNN+Djmju2fuuirTrXr0OUY5Qc+vA2UMMkpin4vVmA/vhU+Cuq/EJJimohID+VOMK3L4dNNKp2gl7F1tWuuk5Bpr0OOoEk+ssRZTc8yxLn5nNwpqjKqqiKSLSJqqlgcOqOrvu3T1COaGE4fx94+28PAHG3nkqknhFsdiiXhUlXVF5Zw7Pgqql+/fBCteNhVi/F5jPSSkwQX/1z3zL/2HcX+ddFv3zNcZPDHGFbd/Y9vnbfsMPvkLTLzKBGnsWQFVB0wFnaEnH9rTbcCE0MncDDcKarKIPA2kASIiJcCNqro4tKKFl4zkOK49YQiPzt/Exr3ljMy1PnWLpS2Ky2spra5nTL/UcIvSPiv+BQhc/qLpHv3S1aZtTktuuNpy08m6rrJ1BVa5H/51HWQOgRHTTeeDvAIYfELIv0qb9B1pAitao7YCXv0OZA6Gc35vlPTos3pOvnZwk6j7FPA9VR2qqkMwJfyfDq1YkcFNU4eTFBfDwx+08wZisVhY5wRIjO6JAImy3VBe1LmxqrD8ZWMdpA80+0afDeW7jPUQzPp34JHj4bOHYfHTsHt5y3N+/lfY+rFpv/PKN02n7JNuC394d/YoOLjVrDG1xLv/Dw5ugxl/M8opwnCjoMpV9aPAB1X9GChv4/xeQ5+UeK4+fghzlu1iy77KcItjsUQ06/Y4EXz9QvygqzoAT0yHR0+AotUdH7/rKziwCSZc1rhv1Jnm3/XvNO5b+qJpNJqQBle+bNZlljx/6Hw1pfDlE6YT9c82w80fwMwnTU+5cNN3lAlkOLj10GPLXjJK98Rb2g+iCBNuFNSXIvKYiJwqItNE5K/APBGZJCIRtTjT1VJHLXHT1GHExXh4fMGmbpvTYumNbCiqoG9KPH1TW6ggsfSf8Mex5qHf1Vz/t35qklQ9sfDs+R1XUitmmwX+sRc27kvNhbzJsP6/5rPPa9x6AyfBtxcYt9fYC0wCbX110/kWPmk6aZ/8I7PukzcZxn89MnrAZY8y/+7f0HT/ts9gzi0wdCqcdteh4yIEN7/BiZhk3buBXwFjgROBPwIPhEyyTtCVdhutkZuWyEUTB/La0l1U1Ia8CanF0ioicraIrBORjSJyRwvHh4jI+yKyXETmiUh+0LH/FZFVIrJGRP4iIQizW1dUzujWrKdFT0H5bnj1u/D8xbDy3zDv9/DydfDh/VBT5u4iK1+BlbNh2h1ww9tG0XRESfl9Zo5RZx7akXr02bBzsUl2XfVv05j0lJ9CbLw5Pvk6Yy2tntM4pq7KrDeNPB0GTnQnQ0/SNyjUPMCBLfDSVWbd6bLnGr9fBNKuglLV6W1sp/WEkOHmG1MGU1Xn483lu8ItiiVKEZE+LrbMNsbHAI8A52C6414hIs275D4APKeqE4B7gPudsScCJwETgKOAKcC07vx+qsqGovKWE3TLdpmcmum/gHMfgMLFMPsGmHc/7PwK5v8O/nKMcZMd3Goe+i1Rvgfe/LGxUE7+oYlQu/5NR0ld4E5JbZkPFUVN3XsBRp8FqHHzffQg5I4zSivAkJMha5hJvg2w5Hmo2gdTf9z+tcNBUqYJ+w5YUKrw8rVGUV/5MiT3Ca987eAmUTcDYz2d4uyaD9yjqt3nR4twJg3OZFRuKrMW7uAbUwaHWxxLdLLL2dqyXGKA1m6wY4GNqroZQERmARcBwU/lcUCg/MCHwKvOzwokAvHO9eMwSffdxs6SairrfIxqKYJvzRuOdBeZnKBxF0PJdvNzQqpRWHPvgrd+0jgmtR9cMQvyglYR/nuHca/NeAxinEdXQEk9c55RUte9Dv2a6e3yIljynAmA2P4FJKTDqBYi1fpPgLQB8OF9xtq75O9N3XQeD0y6Bt6/B3YvM26y+b+HwSdG7BoOYNah9jmBXju+NCWRLnzIqQYR2biN4isHLnO2Mg6TKL4AIsI3pgxiyfaShoVgi6WDrFHV4ao6rLUN2N/G+DxgR9DnQmdfMMuAmc7PM4A0EenrlCP7ENjtbO+o6ppu+VYODSWOWnLxrZljKnEHElZTcyB/slFOYH6+/g248R248GH42l2AGGvJ7zfnFC6CVf+Bk25vXFcJ0HcEXPcGxMQZJbVnZeOxmlJ47iL44F4TgTdoCsz8e8tVwUWMFVW+G7KGtlxaaOJVJnn1sWnw35+bgq7nP9ih31WPkz2y0YJa+gLEpZgK6lGAGwU1QlXvVtXNzvZrYHioBYs0LpmUT1yM8NLCHe2fbLEcipuEmLbOacnyah5t8BNgmogswbjwdgJeERmJWTvOxyi100TklGZjEZFvicgiEVlUXFzsQtxG1u2pAGBUcwVVuQ+2fWIi3NpCBAYfbyyUqT+GM+4x0XbLZxm31Lv/z1QmOPHWlsdnj3SUVDw8fS5s+9QEOsy+0Tycr30NfrAUvvFC23k+Y841/550e6OVFkxafzj5djjyYvjmXPjmO5A7tu3vFm76jjL18kp3wsr/GNkToiBXDXcKqlpETg58EJGTgOo2zu+V9EmJ58xx/fnPkkJqvW3Un7JYWkBVawBE5JA45cC+wDmtUAgMCvqcj3EZBl9jl6peoqrHAL909pVirKnPVbVCVSuAt4HjW5DxcVUtUNWCnJyOFRddX1RO//REMpLimh5Y+yao30TAdYTxl5oSO+/9yuQsbf8Mpt/Z9oM1eyR8810Tkff8DJh1BWx8D877Iww/1d11R50J186BSde1fs7X7oJLn4FBx7r/PuEkYHF+/CDUlcPEK8MrTwdwo6C+AzwiIltFZCvwMPDtkEoVoXxjyiAOVtXz7qpudd9bDi+ODP7gBD9MdjFuITBKRIaJSDxwOaZoc/Bc2SIS+Ju+E+OeB9iOsaxiRSQOY111u4uvxQTdNXNMdYX+HSyP4/HA2b83AQ2vfse40o65tv1xmYOMq7DfkbDhXTjhFph8vfvripgmgpEQIt5d9HUU1OKbaQE+AAAgAElEQVRnjOtyyEnhlKZDtPm/4NzsY1T1aEwE0ARVPUZVW0mnDi+hyIMK5uSR2Qzpm8yTH2+hl/dtjBhqvT4ufuQTvtwSFU2dW0VE7hSRcmCCiJQ5WzmwF3itvfGq6gVuAd7BKJeXVXWViNwjIgH/2anAOhFZD/QD7nP2zwY2ASsw61TLnMLK3YKqsrm4khE5KU0PVJfA5vnGvdeZqPb8yXD0FcYCO/3XLbvcWiKlrwmWuGKWcRUe7mQNMTljfq+zhhYFhXwd2lRQqurH/FGgqmWq6jJZITyEIg8qGI9HuOnkYSzdUcLibQdDcg1LUw5W1rN0RwlfbG4rfiDyUdX7VTUN+IOqpjtbmqr2VdU7Xc7xlqqOVtURqnqfs+8uVZ3j/DxbVUc559ykqrXOfp+qfltVx6rqOFX9UXd+t7JqL9X1PgZmJDU9sOAP4K+HcS77GLXEuQ/A1a/AmHM6Ni4+xYzxxHT+2r2FmDgTHo8YhR9FuLFj54rIT0RkUHDORsgli1C+PnkQWclxPLZgc7hFOSyo95korr3ltWGWpNv40kndAEBEMkXk4nAK1FX2lJmls34ZQZFxq/5j6tdNuclYQp0lIdUkwUbRW39EMuYcOPpy4wKNItzYzDc6/34/aJ9yGEbyASTFx3DN8UN46MONbCquYEROdETDRCtev3Gl7i1vK34gqrhbVf8T+KCqJSJyN405S1FHkaOgGpoU7l0Lr37fBDmcdX8YJbM0cOZvwi1Bp3BjQY1tIV+jeQb7YcU1JwwlLsbDkx9vCbcovR5v77OgWvqbc7m4EpnsCVZQFcWmdUV8csSX0bFEPm4U1Kcu9x025KQlMHNSHrMXF1Lcex6cEUm9z7GgynrN73mRiDwoIiNEZLiI/AmI6t5qRaWOi8+/B546E0oL4dJnG1tZWCydpFUFJSL9RWQykCQixwSql4vIqUByj0kYodw8dTiqyv+8utJG9IUQr1NJoLi8trf8nm8F6oCXgJcxOYXfb3NEhLOnrIbjknaS8OzZphXGdXNMN1aLpYu05Vo4C7gekxAYXMujHPhFCGWKCobnpPLjM8fwu7fXMntxIZcWRNfiY7QQsKDqfH5Kq+vJTI5ul5GqVgJ3iEiqkzQb9RwoKeVR7gNPCtw4B3KPCLdIll5CqxaUqj6rqtOB65tVML9QVf/dgzJGLDdPHc6xw/rw69dXs+NAKxWYLV0isAYFUNQL3HwicqKIrMYp8ioiRzs91qKWcfvepo+WwCWPW+Vk6VbcrEG9ISJXisgvROSuwBZyyaKAGI/w4GVHI8DtLy2lqs72i+puAlF80Gsi+f6E8U7sB1DVZTR2Cog+VDmv8lV2Jo6KqgoFlujAjYJ6DVPW3wtUBm0WID8rmXtnHMVX2w9ywUMfs3ZPROcyRx31QRZUbwmUUNXmFYejtrijd8P7DKeQ5YOutrlKlm7HTXhrvqqe3f5phy8XTcwjOzWB219aykUPf8K9Fx9l16S6Ca8v2ILqFQpqh9NAUJ2aerfRzXXxehLvJw9zQDMpHd7BYrAWiwtchZmLyPiQS9IGInKxiDwhIq+JyJnhlKU1ThqZzVu3TWXykCx+/spyFm6N7tpxkUIgig96jYvvO5iovTxMhfKJRGsUX/E6Erd9yHPeM8jNsgnrlu7HjYI6GVgsIutEZLmIrBAR18ViReQpEdkrIiub7T/bmXOjiNzR1hyq+qqq3oyJKvyG22v3NDlpCTx+bQF5WUn86OWlVNTaNamuEojiE4l+C8qpXH6Nql6lqv1UNVdVr1bV6Cw0+MVj+DzxvOj7Gv3SW2gAaLF0ETcK6hxgFHAmcAFwvvOvW54BmrgInT/UR5y5xwFXiMg4ERkvIm8023KDhv4/Z1zEkpoQy58um8jOg9Xc+8bq9gdY2iSwBtUvLZHiKF+DUlUfZj23d7DrK/ZkTeYA6Y1ljiyWbqRdBaWq2zCN0k5zfq5yMy5o/AKgub/rWGCj06G3DpgFXKSqK1T1/GbbXjH8HnhbVb9q6Tpd6Qba3RQM7cN3po1g1sIdzF1te0d1hcAa1MDMxN7i4vtERB4WkalBye+Twi1Up6jaz0EyiIsR+qREd36aJTJpN0jCKWRZAIwBngbigBeArsSU5gHBkUyFwHFtnH8rcDqQISIjVfVvzU9Q1ceBxwEKCgrCXnLg9tNH88Havfzu7TWcPjYXsRFOnaLeWYMamJnE2j3lYZamWzjR+Te4UZECp4VBlq5RdZD9qWnkpiXa+9sSEtxYQjOAC3FCy1V1F9BC68wO0dLd3KpSUdW/qOpkVf1OS8qpYdIQNyzsCPGxHm48eRibiitZ1EO9oxZvO8A/vtjWI9fqKQIWVF5mElV1vqhe13MagD7aLPF9uqpGn3Ly1kFdOUX1yfTPsO49S2hwo6Dq1BRBUwARSWnnfDcUYtyGAfKBXV2dNNQNCzvK+RMGkJoQy6wvm6e9hIZ/fLGdu15bRVlNfY9crycIrEENzDTN8PaWRa+bL7gBaNRTbbz2O+uS7PqTJWS4UVAvi8hjQKaI3Ay8BzzRxesuBEaJyDAnF+RyYE4X54w4kuNjuXDiQN5csYvSavdKY295Des64c6qqvXh8yufbNjX4bGRSqCSRIOCivJIPrrQALS9yFcRGSIi7zvRtvNEJN/ZP11ElgZtNV1uklhlAg+3VyfaCD5LyHATJPEAMBt4BbMOdZeqPuT2AiLyT+AzYIyIFIrIN1XVi3mTfAeTpPiyqq7qzBdodq2IcfEFuHzKIGrq/cxZ5t5A/OV/VnLW/y1g5qOf8sbyXU3q0bVFpVNqad668AaJdCfeBgvKPAR7gYK6EZP3tADTZmMxsKi9Qa1FvjY77QHgOVWdgFnjuh9AVT9U1YmqOhGz1lUFvNulb1FlLKg93hT6ZyR0aSqLpTVcNUpT1bnA3M5cQFWvaGX/W8BbnZmzjWu9DrxeUFBwc3fO2xXG52UwbkA6Ly3czjXHD3E1Zu2eMkbmplJcXsstLy7hxpOGcdcF7feIrHTWZ+avL0ZVe8XCdSAPamBG9Lv4AJyGn52hIfIVQERmYULWg3MZxgE/dH7+kJa79H4dEw3bterGjgV1QNOsBWUJGa7DxS2dQ0S4/NhBrNxZxsqd7Vt2NfU+Cg9Wc+74AXz4k1M5Y1w/Xl++C7+//cDEqjofHjH9eXpJxBtev58Yj5CZHEd8rCfqG0SKSJyI3CYis53tFhGJczG0pcjXvGbnLANmOj/PANJEpG+zcy4H/tmKbO5TNZw1qIOaZtegLCGjVymoSHTxganVlxQXwy0vfsXibW2XQNq2vwpVGJGTQoxHOPvI/hSX17J6d/tFaCvrvBw3zDyPeoubz+tTYj2CiJCTmtAbXHyPApOBvzrbZGdfe7iJfP0JME1ElgDTgJ2YIs9mApEBwHiMa/3QyVQfV9UCVS3IyclpWxrHgioh1UbxWUJGhxSUiGSJyIRQCdNVIi2KL0BGUhzP3DAFr1+59G+fcf9ba5pU6Q5mU7HpYTcix9Q2mzYmBxH4YO3edq9TWetjeE4KR/RPY/769s+PBup9SlyMuU37pSf0hmTdKap6nap+4Gw3AFNcjGs38lVVd6nqJap6DPBLZ1/w29plwH9UtethnlUHqYtJpo446+KzhIx2FZQTDZTuRBotA54WkQfbG2dpynHD+/Lf20/hG1MG8diCzby8qOXQ882OghqWbaL5s1MTmJCf6VJBeUlNiOXUMbks2nqQ8l4Qbu71+4mLMcZDblpib2i54ROREYEPIjIcd+022o18FZFsJ9cK4E7gqWZzXEEr7r0OU7WfypgMMpLiSIyL6ZYpLZbmuLGgMlS1DLgEeFpVJ2OqOkQckeriC5CaEMtvZ4xnSN9k3mulBNKm4koGZCSSktAYv3LamFyWFZawv6L1h7PX56fW6yc5PpZTx+Tg9SufbIzOGqTB1Pv8xDoWVG56r3Dx/RT40Hnxmw98APy4vUGtRb6KyD0icqFz2qnAOhFZD/QD7guMF5GhGAtsfrd8i+oDlGLXnyyhxY2CinV815cBb4RYni4RqS6+YESE6WNy+XTTfqrrDn1x3lxcwfCcprnQpx2Ri6qJzmuNSmeulIQYJg/JIi0htle4+ep9SpwnYEElUFpdT0191Pb3Q1XfxxRfvs3Zxqjqhy7HvqWqo1V1hKre5+y7S1XnOD/PVtVRzjk3qWpt0NitqprnJAt3nar9lJBma/BZQoobBXUP5q1to6oudFwSG0IrVu/ma2NzqfX6+XRT04RaVWVzcWXD+lOAIwemk52a0KabL9BuPiUhlrgYD5OGZLFshztL8tlPt3LqHz7EFAyJLLzBFlSaeVuP5kg+Efk+kKSqy51278ki8r1wy9VhqvZTKmnEx/aqOCtLhOEmUfdfqjpBVb/nfN6sqjPbG2dpnWOH9SElPuYQhVNcXkt5rZfh2U0tKI9HmD4mhwXri1tN2q2sNVZFcrxZDxjTP42NxRWuknzfWbWHrfur2FdR15mvE1Lq/UqsswaVk24SQqM8UOJmVS0JfFDVg0DE5O25puogJaQ1BLBYLKHATZDE/zpBEnFOGZV9InJ1TwjXUSJ9DSpAQmwMJ4/K5oO1e5tYLZuKKwEYnnNod9LpR+RSVuPlq+0lhxyDxiTdlHizdjW6Xxp1Xj/bDrSdj1nv87PEmXPHwa7lboYCr89PnMfcpoH1jj2l0WtBAR4JyqB2KkREl5/MVw+1pRwkjQRrQVlCiJu760wnSOJ8TKjraMxCb8QRDWtQAb52RD92l9awZndjQm1DiHnuoQrqpJHZAK22kq8McvEBjOlnCs6vbydhd9WuMqqdNZ0d7SizcOD1NVpQA5x8m92l1eEUqau8g6lv+TUROQ0TVfffMMvUMapNdf6DmtoQYWmxhAI3CiqQ5X4u8E9VbTvT1OKKU48wiZAfrmt0820uriQxzsOAFiKjMpLi6JsST+HBlh/OVbWNQRIAI3NTEYH1RRVtyrEoSOG1Nnc4MS4+c5uakGYPRdFd7ujnwPvAdzE1+d4HfhZWiTpKoMyRP9W6+Cwhxc3d9bqIrMU0LXxfRHKAqH5CRAK5aYlMyM/g/TWN4eab91UwLDsVj6flt9L8rCQKW3HDBSyoZMfFlxQfw+A+yawvatuC+nLLAYb0TSY7NT5CLSh/QxSfiNA/PZHdpdF7+6mqX1X/pqpfV9WZqvqY0wo+enAKxe7zpxFnXXyWEOImSOIO4ASgwMlAr8QUqYw4omUNKsBpR+SyZEdJg8tqU3EFI3Jab7eV3ye5VSVS2cyCArMOta4NBaWqLNp2kIIhfcjPSo7QNahGFx9A/4zEaLegoh/HgtrvTyHeWlCWEOImSCIOuAZ4SURmA98EIjIDNJrWoABmHJNHYmwMP3xpKVV1XgoPVrcYIBEgPyuJnSXVLRaOrWq2BgVmHWrLvkpqvS2/oG8qruRAZR3HDstiUJ9kdhyIRBefv4kbaUBGUlRbUL0Cp1DsXl+qDTO3hBQ3d1fz4paTcFfc0tIOQ/qm8JuLj+LzzQf4yb+WNRSJbY1BWcnU+5SiFsKsG8LMg8rOjO6fhs+vbNlX2bAvOOw8EHAxZWgfBmUlsaukGp+Lquk9SaBYbIB+6caCclPdPdIRkUQRSQ+3HB3GsaCKfSk2SMISUtwoqM4Wt7S44OuT8/n65HzeWrEH4JAk3WDys0xPpJaCGSrrvCTEehoCCgBG9zNzBbrzbiqu4JjfzOWh902e9cKtB8hOjWdYdgqD+iTj9WvERcjV+5pbUInU+5QDVZGXs9URROQmTETfmyLy23DL0yGqDqBxyVT642yQhCWkuLm7Olvc0uKSey46klFO1N2w7DYsqD7JQMvh4IFCscEMz04l1iMNgRLPf7aN8hovf5y7nicWbGbh1gMUDOmDiDAoKzB3ZCkor1+bPAQDrR32RJmbT0QuaLbrdFWdpqpTgfPCIVOnqToASVkAVkFZQoqbjrqB4pabMT1phgA3hFSqw4zk+FieufFYlu8oabKG1Jy8zNYtqKo6H8kJTatKx8d6GJadwro9FVTWenllcSHnTxiAAve9tQaA6080DV4H9TFz7zhYxQk073EXPkyx2KAgifRALlQNR+VFx1qjw9GO1XSXU+JouYj8A9PTaVV4Resg1QfwJ/UBsIm6lpDSpoJySvdXY4pbjsEoqLXBRSgjCect9YKRI0eGW5QOk5eZ1KCAWiMxLobctIRWLahAFYlgRvdPY+XOUl5buovyWi83nDSUCfmZ1Nb7eW9NEScMN8poYGYSHoHCCAs1N2tQTV18YLoGRxOqeq+I9AfucQpJ3AWkAsmqujyswnWUqv34E42CshaUJZS0eXc5lY//qKq1geKWkaqcIPqi+DqDyYVqeQ2qJetrTL80th+o4smPNzN2QDqTBmcRF+Phr1dN4q3bpjJuoFmjj4vxMCAjiR0Rlqxr1qAaLai+qQnEeoQ9EbZW5pJK4HbgEeBxTH+m9WGVqDNUHcCXaF18ltDj5u56V0RmBtcPs4SPQX1azleqrPU1FIoNZnS/VFRNSPk1xw8h8N8YH+tpUE4B8rOSIi5Z1+tvmgcV4xFy0xKiLtRcRO4F3sRUjpiuqhdiGoC+KSLXhFW4jlK1n/qEgIKyjwVL6HCjoH4E/AuoFZEyESkXkbIQy2VphfwskwfUvEp5VV0rLj6nJl9aQiwXTRzY5tytKb9wUu/zN3HxQdQm656vqqcAJwLXAjh9nM4C+oRTsA7h80JNSYOCsnlQllDSbpCEqqb1hCAWdwzKSsbnV3aX1jRE9YFjQSUcakEN6ZtCVnIcMyfltxmAEZi7qKyWmnpfxLTx9vr0kLf0/hmJrG2nCG4EslJEngeSCOpq63TK/XPYpOooNabyfX18JoCtJGEJKW4qScwQkYygz5kicnFoxbK0Rr4TDt58Haqy7tAwczAusfd+NI2fn3NEu3MHIvl2lkTO+o7X72+S2wXQPz2JPaU1EdlgsTVU9Wrgf4H/p6o/7MwcInK2iKwTkY0ickcLx4c4LXGWOy3l84OODRaRd0VkjYisdlrAdxwnSbc23q5BWUKPm7vrblVtKG7nNFu7O3QiWdoiOBw8mKpaX0Oh2Ob0TU1w9SBpK88qHKhqk5bvAQZkJFJV56Pc6YEVDYjIJFVdoapr2zqnjWMxmOCKc4BxwBUiMq7ZaQ8Az6nqBEwn7PuDjj0H/EFVxwLHAq23Z24Lp1BsTZx5Z7XFYi2hxM3d1dI5bvKnLCFgQEYSIk0tqDqvnzqfn5QWgiQ6QkOyboRE8gXKLjW3oPpFZ7Lu0yKSJSJ9WtuAJ9sYfyyw0eloXQfM4tCizeMwQRgAHwaOO4osVlXnAqhqhap27i3EsaBq4qyLzxJ63CiaRSLyIObtTYFbgcUhlcrSKvGxpl9UcL5SS4ViO0NuWgLxsZ6IyYXyNiioQy0oMMm6gSCQKCAD83fTVthbcRvH8oAdQZ8LgeOanbMMmIlZ05oBpIlIX0yT0RIR+TcwDHgPuKN5mw8R+RbwLYDBgwe3LIVTKLY6NhOoID7WRvFZQoeb159bgTrgJeBlTOLu90MpVGeJtnYbnaV5a4zKukNbbXQGj0fIz0yKmEi+eidSMa55FJ9TTaIoiiwoVR2qqsNVdVgb27FtTNGSJmi+CPcTYJqILAGmATsBL+ZFdKpzfAowHLi+BRkfV9UCVS3IyclpWQrHgqqKbcyfs1hChZt+UJWqekfgxlXVX6hqZXvjwsHhkKgLkN+nabJuVW3TZoVdYXDfZLbsiwwF5fWZ52/zKL5+QeWODiMKgUFBn/OBXcEnqOouVb1EVY8BfunsK3XGLnHcg17gVUxXgo5TdQBiE6klAbAKyhJa7N0VheRnJbOnrKahz1NFbcDF1/XQ8HED0tlQVE5NffjrAdf7jQXVfA0qPtZDdmp81JU76iILgVEiMkxE4oHLgTnBJ4hItlOeDOBO4KmgsVlON2yA04DVnZIiawiMOYc6Jw3PKihLKLF3VxQyKCsJVdhdYh7QVQEXXzdYUBPyM/D6lTW7w5+LXd+KBQUmFypKyx11CsfyuQXTomMN8LKqrhKRe0TkQue0U4F1IrIe6Afc54z1Ydx774vICoy78IlOCTLlJrj0Geq8RkPZYrGWUGKj8aKQ/IZouyqGZqdQWds9QRIAE/JNdNaKnaUcMziry/N1hUC1jOaVJMCsQ7VUkzDSEZFXMJbN206tS9eo6lvAW8323RX082xgditj5wITOixwKzSsD1oLyhJCWn2iichDHLoI24Cq3hYSiSztEmhcuNN5QFfWBdaguu7iG5CRSHZqPMsLwx9oErCgmkfxgbGgFm872NMidQePYtrV/EVE/gU801ZuVKTSqKBsFJ8ldLT1yr2ox6SwdIj+GYl4pLHiQ6Dde0uVJDqKiDA+L4PlhSVdnqureP2tv6X3T0/kYFV9RJVlcoOqvge851RnuQKYKyI7MC63F1S1PqwCuiTg4rOJupZQ0uoTTVWf7UlBLO6Ji/HQLz2xwYIK5EEld4OCAuPmm7++2PSY6qY5O0Mgii/Wc+hbem6aieQrLq9tUpMwGnByk64GrgGWAP8ATgauw6wjRTx1jgVlE3UtoaTdp48T+fNzTJZ6YmC/qp4WQrks7ZCXmURhMwsquZssiQn5GfgVVu8uY8rQ8BXabmudIzstHoB9FdGloJxk2SOA54ELVHW3c+glEYkar0W9NxDAYhWUJXS4ubv+gYkaGgb8GtiKCVu1hJG8rKTGNahaL8nxMXhasDQ6w/h8k0e2bEd43XytVZIAyE41eTj7Kup6VKZu4GFVHaeq9wcpJwBUtSBcQnWUep+fGI8Q0033nMXSEm4UVF9VfRKoV9X5qnojcHyI5bK0Q15mEnvKTF+oyrrWC8V2hty0RAZkJLJiZ3gDJerbiOILKKji8oht8NwaY0UkM/DBqc/3vXAK1Bmadzq2WEKBGwUVWLTdLSLnicgxmCz2HkFExorI30Rktoh8t6euG+nkZSXh8ytF5bWmWWE3JOkGMz4vgxVhjuRrrZIEQN/URhdflHGz0xEAAFU9CNwcRnk6RZ3Pb917lpDj5g6714k4+jEm2e/vgKt+NiLylIjsFZGVzfa32dcmGFVdo6rfAS4DosYFEmryMhtDzY2Lr3uDGSbkZ7B5XyWl1eELKvO2UkkCICE2hoykuGhUUB4RadC4ThuN+DDK0ynqvH6bpGsJOW466r7h/FgKTO/g/M8AD2N60QBN+tqcgakRtlBE5gAxNO1fA3Cjqu51MuXvcOay0JgLtaukmspaH6ndbUE5CburdpZy4sjsbp3bLfVtRPEBZKfGR6OCegd4WUT+hskz/A7w3/CK1HHqrQVl6QHcdNR9tgWf+VNtjQmgqguAA812t9jXxmnmdn6zba8zzxxVPRG4yu0X6+0MzGzsfltVFwILKs8ESiwP4zpUwMUX38qbenZqAvvKoy5I4ufAB8B3MV0B3gd+FlaJOkG9T62CsoQcN0+1Cc195s46VGdx09emARE5FbgESKBZmZdm57Xfy6YXkRwfS5+UeAoPVlNR6yXPsai6i6yUePqnJ7KhqKJb5+0IjUESrVhQaQms3hX+moEdwSlv9KizRS11NkjC0gO4UVAeEclyFnNxOn925XXdTV+bxgOq84B57U2qqo8DjwMUFBS0Ol9vIi8zybGgfN1SKLY5fVLiKakKn4XSXr23nNQE9kVZFJ+IjMK4spvnFQ4Pm1CdoM7rJz42eip4WKITN0+1PwKfikigCOWlOFWSO0m7fW06i4hcAFwwcuTI7pgu4snLTGLD3vKQVXzISonjYBgVVFt5UAA5aQmU13qjrdzR08DdwJ8wa7o30HaX3Yik3ucn3lpQlhDjpmHhc5g20kXAXuASVX2+C9dst69NZzlcGhYGyMtqtKC6o1BsczKT4ykJZxRfG3lQYIIkIOpCzZNU9X1AVHWbqv4K058pqrBBEpaeoNU7TETSnX/7AHuAFzFVJfY4+9pFRP4JfAaMEZFCEflma31tuvY1Gq53WLR8D5CXmURNvR+vX0NiQWUmxVFSFT4F1VY/KIjaZN0ap6ngBhG5RURmALnhFqqj1HttkIQl9LT1VHsROB9YTNM1InE+t+szV9UrWtl/SF+b7kBVXwdeLygoiLrEx84QHBiREgILKivZrEH5/dptZZQ6Qlt5UBC15Y5uB5KB24DfYNx814VVok5Q6/OTER8XbjEsvZy2qpmf7yQUTlPV7T0ok8UlgWRd6L5K5sFkJsfhVyiv9ZKR1PMPo/byoHLSAgoqOiwoJwfwMlX9KVCBWX/qyPizgT9jcgb/rqq/a3Z8CKYZYg4mveNqVS10jvmAFc6p21X1QrpAvddvK5lbQk6bd5iqKvCfHpKlyxxuLr78JhZUKBSUWeMJVyRfY6mjlm/ThnJHUeLic1qvTw6uJOGWoAT3czARgFeIyLhmpz0APKeqE4B7aJr4Xq2qE52tS8oJnCCJWBskYQktbl6BPheRKSGXpBs43IIkMpLiGlx73V2LDyAr2VhN4VqH8vr9iNBqxeyE2BjSE2OjxoJyWAK8JiLXiMglgc3FuBYT3JudMw6T+AvwYQvHuw0bJGHpCdzcYdOBz0Rkk4gsF5EVIrI81IJZ2kdEGtahQhIk4SiocIWa1/uUuFYi+AJkpyVQHF0Kqg+wHxO5d4Gzne9iXEsJ7nnNzlmGibgFmAGkOc0RARJFZJGIfC4iF3dW+AC2koSlJ3DzVDsn5FJ0E4dbHhSYdaj1RRUhCzOHMFpQPn+rOVABcqKs3JGqdmjdKQg3Ce4/AR4WkeuBBcBOwOscG6yqu0RkOPCBiKxQ1U1NLtCBaiy1Xn+rJagslu7CTbHYbSJyNDDV2fWRqi4LrVid43CL4oPGSL7UUCTqhnsNyt/+W3p2WgJroqjckYg8TQuVU5w+a23RboK7qu7ClAVDRFKBmapaGnQMVd0sIl1GfWsAAB/vSURBVPOAY4BNzca7rsZiEnWtgrKEFjfFYn+AyX/KdbYXROTWUAtmccegLNPuPBQKKj3RzHkwTBaUm6Z4OalR5+J7A3jT2d4H0jERfe3RboK7iGQ7OVYAd2Ii+gIFnhMC5wAnAau78iVsw0JLT+DmqfZN4DhVrQQQkd9jkm8fCqVgFndcPmUweVlJ9HVygrqT2BgP6YmxYesJVe/zt1pFIkB2ajzlNdFT7khVXwn+7CSzv+dinFdEAgnuMcBTqrpKRO4BFqnqHOBU4H4RUYyL7/vO8LHAYyLix7yU/k5Vu0FBWQvKElrcKCgBfEGffURo7bDDcQ0qIzmO8ycMDNn8mcnxYQuS8Pq03TWoxmTdWvIdazLKGAW4Kr/fUoK7qt4V9PNsYHYL4z4FxndNzEb8fqXep3YNyhJy3Ciop4EvRCSQD3Ux8GToROo8h+MaVKjJSo4Ln4vPxRpUY7JuXVQoKBEpp+ka1B5Mj6iood7fdpV5i6W7cBMk8aCzqHoyxnK6QVWXhFowS2SQmRy+lhten7/VKhIBGiyo6EnWTQu3DF0lUOHDBklYQk27CsopDLvV2QL74lQ1fFVELT1GZnIcW/ZVhuXa9T5ttQ5fgOzoK3c0A/ggEF3ndKs+VVVfDa9k7qn3BiyoiPT0W3oRbl6BvgKKgfXABufnLSLylYhMDqVwlvCTFU4Lyt9+pFjfFBMKH0UVze8OKCcAp1v13WGUp8PUBRpJ2jUoS4hxc4f9FzhXVbNVtS8mcfdl4HvAX0MpXEc53Grx9QQZSXGU1XgbejP1JF6ftuviS4yLIS26yh219DfX/TkCIaTOsaCsi88SatzcYQWq+k7gg6q+C5yiqp8D3R/b3AUOt1p8PUGgHl84Qs3rff52XXxgAiWiqOXGIhF5UERGiMhwEfkTpqVN1FDvvKzYKD5LqHFzhx0QkZ+LyBBn+xlw0Kmu3POv1ZYeJctxoYWjs66pJNH+Okd2dCXr3grUAS9hPBHVNOYrRQX17VSZt1i6CzeuhSsxPvLAIu7Hzr4Y4LIQyWWJEAJ9oMKxDuX1+Yl1USEjJzWBNbujo9yRk/B+R7jl6AoBC8oqKEuoafcOU9V9qnorMFVVj1HVW1W1WFXrVHVjD8hoCSOBenwHK8Ph4nNnQfXPSGR3aQ2mfVlkIyJznci9wOcsEXmnrTGRRq3XuvgsPYObWnwnishqnNpdInK0iERUcIQldARaboTHxeeunM6AjESq631hK8nUQbKdyD0AVPUgpsZl1NBoQdkwc0tocfMK9CfgLEwPG5xK5qeEUqjOYqP4up9wdtX1usiDAtNyBGBnSXWoReoO/CLSUNrIadMe+aZfEA1BEtbFZwkxru4wVd3RbJevxRPDjI3i637SE2OJ8UhYekLV+fzEtRNmDjDAUVC7S2pCLVJ38EvgYxF5XkSexxR1vTPMMnUIuwZl6SncBEnsEJETAXXK/N8GrAmtWJZIQUTISIoLS8FYN8ViAQZmJAKwqzTyLShV/a+ITAKOx5QO+6Gq7guzWB2izmsVlKVncHOHfQcTBpuHaZo2EZOkazlMyEyOC4sF5fW7y4PKTk0gLkbYFR0WFBgPxF6gFBgnIhHpMm+NukAtPhskYQkxbiyoMap6VfAOETkJ+CQ0IlkijcykOEqqe96CqvepKxefxyNOJF/kW1AichPwA0xH3KUYS+oz4LRwytUR6m0lCUsP4eYOa6kxoW1WeBiRlRwfljBzr8tKEgADM5LYFR1BEj8ApgDbVHU6pvV6cXhF6hgNa1CxNorPElpataBE5ATgRCBHRH4UdCgdk6RrOUzITI5n7Z7yHr9uvd/dGhTAwMwkvtxyIMQSdQs1qlojIohIgqquFZEx4RaqI9TZIAlLD9HWHRYPpGKUWFrQVgZ8PfSiWSKFzORwBUn4iWun5XuAgZmJ7CmrweeP+IjtQidR91Vgroi8BuxyM1BEzhaRdSKyUUQOqUbhlCJ7X0SWi8g8EclvdjxdRHaKyMNd+QJ1NlHX0kO0akGp6nxgvog8o6rbelAmS4SRlRxHVZ2PWq+PhNieMZ79fsWvuLagBmQk4fMrxeW19Hei+iIRVZ3h/PgrEfkQyMB0DGgTp/blI8AZmGClhSL/v707j6+qPBM4/nuykABhCQlgWGQRlEUhIQGZgrjUTsG6gRmL2wjqMKId1FqXjhUrldFWbB3s6AgVxcqMIsqihdaCKcgoSMISIVBBRIlBSBHCIpR7k2f+uCfhEkJyc3O3nPt8Px8+3Hvuuee8J/e873vOe973eWWJqpb4rTYDeFVV54rIZcCTwC1+n/8CWNnUY7AJC02kBHKGfSsiT4vIUhF5v/pf2FMWBBuoGx7Vg3UrItiTr7HTindp76uUmslgXcB3EaiqS1Q1kNvTYcAOVd3prP86cE2tdQYAK5zXBf6fO3O3dQbea2q6bRyUiZRAzrB5wDagF/A4vpl114UxTUGzgbrhUR3u6EAEKyivc5Xe0HxQ1bpUD9ZtBj35gtQV8B8wX+os87cJuM55PRZoIyIZIpIAPAM8UN8ORGSSiBSKSGF5+Zn7bZzwVpEgkBjgb2NMsAKpoDJU9SXA41zx3Yava6yJEzUBYyP4HMrbyCkdsto1q2gSwairNqj9wO0nwMUisgG4GPgK8OIbt7i0jogwp25MdZaq5qlqXseOHc+4nqcysBiJxjRVIOOgqi+b94jID/A90O1Wz/rGZU5OuRGNJr7ArtLbpiaRlpLUrJr4GqkU6O73vhu1OleoahkwDkBE0oDrVLXC6ZF7kYjcha/jUwsROaKqQU37caKyyjpImIgIpIJ6QkTaAffjG//UFrgvrKkyMaVm0sIo3EEFOg5KRMhqJoN1g7QO6CsivfDdGY3HNy9bDRHJBL5R1Sp88f3mAPgPtBeRCfhmyQ56TipPZZV1kDAR0WAFparvOi8rgEvDmxwTizqmpZCSlMBf90ZuLFT1g/hAn0GB7zlUMwp31Ciq6hWRHwF/wjcOcY6qbhGRaUChqi4BLgGeFBHFF4Q2LDP1erxqTXwmIhqsoERkLnBP9Rw2IpIOPOM8izJxoEVSArk90lm7M3IDYYPpKdalfSpbytzbg1NVlwJLay2b6vd6AbCggW28ArzSlHScqKyyKBImIgLJ/YPqmGAtJ3xJMrFoeO8Mtn59KGLNfN6q6ia+wAvCrHYt+duRExz3xORsMK5xwpr4TIQEcpYlOHdNAIhIBwJ7dmVcZHjvDFSJWDihk018jbmD8vXk+7rCnc18scLjtV58JjICOcueAT4UkV847d0fAr8Kb7JMrBncvR0pSQmsjVAFdbKbeSOeQTWjeaGaM4/14jMREkgniVdFpBDfdAACjKsVXsXEgZSkRIacnc6anfsjsj+v08080F58cPIOyq0dJWLFCRsHZSIk0CnfS1T1t6r6XDQqJxFpLSJFInJlpPdtThreO4OSPYeoOBb+8VDV8d4CmQ+qWnUMvj3uHQsVEzxetWdQJiLCepaJyBwR2Scim2strzcqcx0eAuaHJ5UmUBf27oAqrItAM19jx0EBpCYnkpnWwpr4wszXi88qKBN+4T7LXgFG+y/wi8o8Bl9wyxtEZICIXCAi79b610lELgdKgL1hTqtpQHb39rRISohIM5+npomvcd2Zz+mYxvovDja8ogmab6CudTM34RfW3niqukpEetZaXBOVGUBEXgeuUdUngdOa8ETkUqA1vsrsmIgsdUbK115vEjAJ4Oyzzw7lYRhHanIiQ85uH5GOEjWdJBrRiw/gB4OymLp4C9u+PkS/s9qGI2lx74T14jMREo2zLJCozDVU9RFVvRf4H2B2XZWTs15AgS5N01zYK4MtZRUcOh7e51DeIKcVv+KCLBIThCUbA5oD0ATBgsWaSInGWRZIVObTV1B9xS/skomSkX0zqVJYvOGrsO7HUz1Qt5F3UJlpKYzsk8nijWWoxvzsus2Sp1Ktm7mJiGgMuG0wKnOwROQq4Ko+ffqc9pnH46G0tJTjx60LclOkAXPHdWXXgX1UHO1Mu9Ytw7Ifj7dx0cz9XZPdhR/P38T6Lw+Q26NDqJMW99zczdzKidBKTU2lW7duJCcnB/X9aFRQDUZlDpaqvgO8k5eX9y+1PystLaVNmzb07NkTEXvA2xTdj3tI+6KMwi07+O6wC8Kyj2DGQVX7x4FnkZL0CUs2llkFFQZu7iRh5UToqCr79++ntLSUXr16BbWNcHcz/1/gI+A8ESkVkdtV1QtUR2XeCsxX1S0h2t8Zp3w/fvw4GRkZdtKFQFpqMhkZGVR5Pew7HJ4rzWDGQVVLS0ni8v6debd4T82zLBM6bu4kYeVE6IgIGRkZTbobDetZpqo3qGqWqiarajdnZl5Udamqnquq56jq9BDur94p3+2kC52sdi0RYOaK7WHZfnXFEswdFMDV2V3Yf/QEq3f8LZTJMrg/1JGVE6HT1L+le88yE1YpyYm0Tknkfz/ezf+FoRIIJpq5v0vO60hmWgt++/4O6ywRQqqKp9LmgzKR4aqzrL4mvmg7ePAgzz//fKO/d8UVV3DwYP0DT6dOncry5cuDTVrQ2rZM5pyOrbnztSJ27DsS0m17ghwHVS0lKZEHv9+Pwi8OsGhjeHscxpPq38XNd1DR5MZyoilcdZY11MQXTWc68Sor65+7aOnSpbRv377edaZNm8bll1/epPQFI0GEl24dSkpSArfPXcc3R0M3V9TJJr7gmwjyc7sxuHt7/mPpNg6HedxWvDhRGXzvStMwN5YTTRGX8zo9/s4WSsoOhXSbA7q05bGrBp7x84cffpjPPvuM7OxskpOTSUtLIysri40bN1JSUsK1117L7t27OX78OPfccw+TJk0CoGfPnhQWFnLkyBHGjBnDyJEj+fDDD+natSuLFy+mZcuWTJgwgSuvvJL8/Hx69uzJrbfeyjvvvIPH4+HNN9+kX79+lJeXc+ONN7J//36GDh3KH//4R4qKisjMzGzScXfv0IoXb8njhtlruPO1Il67/cKQXF2fHAcVfEGYkCBMu3og1z7/fzz3/g7+/Yr+TU5XNInIaOA/8U35/jtVfarW5z2AOUBH4BvgZlUtdZa/7XwvGXhOVf87mDSc7P7vqmvbOlk5EbpyIliuOstiuYnvqaee4pxzzmHjxo08/fTTfPzxx0yfPp2SEl9w+Dlz5lBUVERhYSEzZ85k//7T491t376du+++my1bttC+fXveeuutOveVmZnJ+vXrmTx5MjNmzADg8ccf57LLLmP9+vWMHTuWL7/8MmTHltsjnafzB/Hx59/wxB9CE+zeW1lFUoI0+SHr4O7tuT63O3NWf85fvz4ckrRFw5liWNZabQbwqqoOAqYBTzrL9wDfUdVs4ELgYRHpEkw6qieStCa+8HBzOREMV91B1TcOyl99VzCRMmzYsFPGBsycOZOFCxcCsHv3brZv305GRsYp3+nVqxfZ2dkA5ObmsmvXrjq3PW7cuJp13n77bQBWr15ds/3Ro0eTnp5e53eDdU12V7aUHWLWqp2c36Ud1w/t3vCX6uGt0iY17/l7cPR5rNi2lx/9z3oW/2gErVo0y9O+zhiW+AIpVxsA3Oe8LgAWAaiqf9trCk24MD3ZxOf+CsrKidCXE43l/rMsRrVu3brm9V/+8heWL1/ORx99xKZNm8jJyalz7EBKSkrN68TERLxeb53brl7Pf51I9GR78PvnMbJPJj9btJn1Xx5o0rY8lVVBd5CoLSMthd/8MJsd5UeYujgkQ+6iIZAYlpuA65zXY4E2IpIBICLdRaTY2cYvVfW06C0iMklECkWksLy8vM5EnHCa+Gw+qMhwYznRGHaWRUibNm04fLjuJqaKigrS09Np1aoV27ZtY82aNSHf/8iRI5k/3zel1nvvvceBA02rQOqSlJjAczfk0LldCnfMLWxSzz5vpYZ0zqGL+nbk3y7tw4KiUt4s3N3wF2JPIDEsfwJcLCIbgIvxRWrxAqjqbqfprw9wq4h0Pm1jAQRcruldaRVUWMRDOdEYrjrLYvkZVEZGBiNGjOD888/ngQceOOWz0aNH4/V6GTRoEI8++ijDhw8P+f4fe+wx3nvvPYYMGcKyZcvIysqiTZs2Id9PeusWvHrbhSSIcPPv1rL7m2+D2o63qqpJHSTqcs/l5zK8dwceXbyZrXtC+/A7AhqMYamqZao6TlVzgEecZRW11wG2ABcFkwh7BhVe8VJOBExVXfcvNzdXayspKTltWTw5fvy4ejweVVX98MMPdfDgwU3eZn1/05KyCr3gsT/qRb98X/dWHGv0tn/8xkb9zpMrmpK8Ou09dEyHPvFnHfWr9/Xg0RNBbQMo1Aif0/ieF+8EegEt8DXnDay1TiaQ4LyeDkxzXncDWjqv04FPgQvq219deUhVteiLb7THQ+9qwba9Qf3tYp2VE5EpJwLNQ3YZFCe+/PJLhg4dyuDBg5kyZQqzZ88O6/76Z7Vl7m3D+NuRv3P73EKOnah/HEdt3qqqkHWS8NepTSov3JxL2cFj3PvGBqqqYqvN/Uz0DDEsRWSaiFztrHYJ8FcR+RTojK+SAugPrBWRTcBKYIaqfhJMOjz2DMrVIl1ONKRZdmcyjde3b182bNgQ0X3mnJ3OzPE5/MvvC/nJgk389oacgLuNeys15E181XJ7pDP1qoE8umgzzy7/lB//43lh2U+oqepSYGmtZVP9Xi8AFtTxvT8Dg0KRhppefNbE50rRKCfqY2eZCavLB3TmodH9+EPxHmau2BHw98I9a+vNF57NP+V2Y+b7O3hnk82+G6iaZ1B2B2UiwFVnWSx3kohn/zqqN+OGdOU3yz9l/rrAetCFchxUXUSEJ8aez9Ce6dz/5qYmd4uPFye81ovPRI6rzjKN4Vh88UxEeHLcBYw6tyMPvlXMa2u+aPA7nsqqRk/33lgpSYm8eEseWe1SmfRqYdA9DuPJyV58FovPhJ+rKigTu1KSEpl1Sy7f7deJny3azEurP693fW+lRiQgaYfWLXjp1qGc8FZxx9xCjv697kGNxudEHMXiM9FnZ1mMSktLA6CsrIz8/Pw617nkkksoLCysdzvPPvss33578s4gkLD84ZKanMgLN+cyeuBZ/OLdknqb+3zjoCJzevbplMZvbxzC9n2HefCt4pgbTR9LPHEU6qg5cGM54c/OshjXpUsXFiw4rWNWwGqfeIGE5Q+nFkkJPHdjDhf1zeSnCz9h5ad1h9TxVIb3GVRto87tyAPf93XmmP3Bzojtt7mxgbqxyW3lRLX47Ga+7GH4OqhhIGd21gUw5qkzfvzQQw/Ro0cP7rrrLgB+/vOfIyKsWrWKAwcO4PF4eOKJJ7jmmmtO+d6uXbu48sor2bx5M8eOHWPixImUlJTQv39/jh07VrPe5MmTWbduHceOHSM/P5/HH3+cmTNnUlZWxqWXXkpmZiYFBQU1YfkzMzP59a9/zZw5cwC44447uPfee9m1a9cZw/WHSnJiAs/fNITrX1zDXa8VMf/Of2Bgl1OfG3qrwtuLry53XtybT746yFPLtjEgqx0j+0ZnioFYdiKeQh1ZORHVcgJcdgcVy734xo8fzxtvvFHzfv78+UycOJGFCxeyfv16CgoKuP/+++ttXnrhhRdo1aoVxcXFPPLIIxQVFdV8Nn36dAoLCykuLmblypUUFxczZcoUunTpQkFBAQUFBadsq6ioiJdffpm1a9eyZs0aZs+eXTP+IdBw/U3RJjWZlycMpW3LZCa+vI4v9h895fNIPYPyJyI8nT+YPp3S+GBH3Xd28c6CxYaXlROnctUdlAY43UZ9VzDhkpOTw759+ygrK6O8vJz09HSysrK47777WLVqFQkJCXz11Vfs3buXs846q85trFq1iilTpgAwaNAgBg06OfZy/vz5zJo1C6/Xy549eygpKTnl89pWr17N2LFja6Iljxs3jg8++ICrr7464HD9TXVWu1Tm3jaMH774ETfOXsubd/4DXdr7rsA8lVUkRaEQbJ2SxFuTv0Ob1OSI77s58MTTjLpWTkS9nHBVBRXr8vPzWbBgAV9//TXjx49n3rx5lJeXU1RURHJyMj179qwzfL6/uiIxfP7558yYMYN169aRnp7OhAkTGtxOfVdgtcP1+zcRhNq5ndvw+9sv5IZZa7jpd2t541+H06lNKt4qJTlMkSQaYpXTmXkqq0gQonLxEC+snDjJzrIIGj9+PK+//joLFiwgPz+fiooKOnXqRHJyMgUFBXzxRf3jg0aNGsW8efMA2Lx5M8XFxQAcOnSI1q1b065dO/bu3cuyZctqvnOm8P2jRo1i0aJFfPvttxw9epSFCxdy0UVBBbhusvO7tuOV24ay99Bxxjz7Ab/586cc/bvXCsEYdCLMET6MlRP+7A4qggYOHMjhw4fp2rUrWVlZ3HTTTVx11VXk5eWRnZ1Nv3796v3+5MmTmThxIoMGDSI7O5thw4YBMHjwYHJychg4cCC9e/dmxIgRNd+ZNGkSY8aMISsr65T25SFDhjBhwoSabdxxxx3k5OSErTmvIbk9OvD6pOE8u3w7/7liOxAnzUjNjMer9vwpzKycOEncOOYjLy9Pa/f737p1K/37949SitwpXH/Tz8qP8GZhKd8b0IncHh1Cvv2mEpEiVc2LdjrCqa48BDC/cDdrd37DM9cPjkKqws/KidCr628aaB6yOygTc87pmMbDY+q/SjTRcX1ed67P697wisaEgN2rG2OMiUmuqqAaGgflxubMaLG/pXErO7dDp6l/S1dVUPVFM09NTWX//v128oWAqrJ//35SU1OjnZS4IiKjReSvIrJDRB6u4/MeIrJCRIpF5C8i0s1Zni0iH4nIFuezH0Y+9c2DlROhE4pyIm6eQXXr1o3S0lLKyy1CQCikpqbSrVu3aCcjbohIIvBfwPeAUmCdiCxR1RK/1WYAr6rqXBG5DHgSuAX4FvhnVd0uIl2AIhH5k6pGPxpojLFyIrSaWk7ETQWVnJxMr169op0MY4I1DNihqjsBROR14BrAv4IaANznvC4AFgGo6qfVK6hqmYjsAzoCVkHVYuVEbHFVE58xLtYV8J+fpNRZ5m8TcJ3zeizQRkQy/FcQkWFAC+Cz2jsQkUkiUigihXYHYWKBVVDGNA91jVqu/aDkJ8DFIrIBuBj4CqiZgVFEsoDfAxNVteq0janOUtU8Vc3r2LFj6FJuTJDiponPmGauFPAfgNQNKPNfQVXLgHEAIpIGXKeqFc77tsAfgJ+p6pqIpNiYJnJlJAkRKQfOFLAqE/hbBJMTaXZ84ddDVSN6iyEiScCnwHfx3RmtA25U1S1+62QC36hqlYhMBypVdaqItACWAe+o6rMB7i+e8xC4/xijfXwB5SFX3kHVd+AiUujmMDV2fO6kql4R+RHwJyARmKOqW0RkGlCoqkuAS4AnRUSBVcDdztevB0YBGSIywVk2QVU31rO/uM1D4P5jbC7H58oKyhg3UtWlwNJay6b6vV4AnDbvt6q+BrwW9gQaE2LWScIYY0xMiscKala0ExBmdnwm3OLhN3D7MTaL43NlJwljjDHNXzzeQRljjGkGrIIyxhgTk+KmgmooEnRzJCLdRaRARLY6karvcZZ3EJE/i8h25//0aKe1KUQkUUQ2iMi7zvteIrLWOb43nHE+JgLclo8sD8V2HoqLCsovEvQYfAE1bxCRAdFNVUh4gftVtT8wHLjbOa6HgRWq2hdY4bxvzu4Btvq9/yXwG+f4DgC3RyVVccal+cjyUAznobiooPCLBK2qJ4DqSNDNmqruUdX1zuvD+E7ArviOba6z2lzg2uiksOmcOY1+APzOeS/AZZwc79Osj6+ZcV0+sjwExPDxxUsFFUgk6GZNRHoCOcBaoLOq7gFfBgQ6RS9lTfYs8CBQHdw0AzioqtVBUF33W8YwV+cjy0OxJ14qqEAiQTdbTmDQt4B7VfVQtNMTKiJyJbBPVYv8F9exqmt+yxjn2r+95aHY/B3jJdRRg5GgmysRScaXseap6tvO4r0ikqWqe5wpFvZFL4VNMgK4WkSuAFKBtviuBtuLSJJzBeia37IZcGU+sjwUu79jvNxBrQP6Oj1XWgDjgSVRTlOTOW3JLwFbVfXXfh8tAW51Xt8KLI502kJBVX+qqt1UtSe+3+x9Vb0J32yx+c5qzfb4miHX5SPLQ0AMH19cVFDOVUJ1JOitwHz/aQqasRHALcBlIrLR+XcF8BTwPRHZDnzPee8mDwE/FpEd+NrTX4pyeuKCS/OR5aEYzkMW6sgYY0xMios7KGOMMc2PVVDGGGNiklVQxhhjYpJVUMYYY2KSVVDGGGNiklVQpkEickl1FGRjTHAsHzWeVVDGGGNiklVQLiIiN4vIx85gwxedOWCOiMgzIrJeRFaISEdn3WwRWSMixSKysHq+GxHpIyLLRWST851znM2nicgCEdkmIvOcEfjGuI7lo9hhFZRLiEh/4IfACFXNBiqBm4DWwHpVHQKsBB5zvvIq8JCqDgI+8Vs+D/gvVR0MfAfY4yzPAe7FNw9Qb3wj8I1xFctHsSVegsXGg+8CucA656KsJb4Al1XAG846rwFvi0g7oL2qrnSWzwXeFJE2QFdVXQigqscBnO19rKqlzvuNQE9gdfgPy5iIsnwUQ6yCcg8B5qrqT09ZKPJorfXqi21VX3PD3/1eV2LnjnEny0cxxJr43GMFkC8inQBEpIOI9MD3G1dHLb4RWK2qFcABEbnIWX4LsNKZB6dURK51tpEiIq0iehTGRJfloxhitbdLqGqJiPwMeE9EEgAPcDdwFBgoIkVABb72dfCF2P9vJ+PsBCY6y28BXhSRac42/imCh2FMVFk+ii0WzdzlROSIqqZFOx3GNGeWj6LDmviMMcbEJLuDMsYYE5PsDsoYY0xMsgrKGGNMTLIKyhhjTEyyCsoYY0xMsgrKGGNMTPp/bsbv+ksRA5wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Model 4 - 500 neurons \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 50\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(500, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "    \n",
    "    # Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "    # Optimizer: Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/50\n",
      "45184/60000 [=====================>........] - ETA: 2s - loss: 0.2730 - acc: 0.9229"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-78-3c6c9e5a4ae1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     30\u001b[0m                     \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     31\u001b[0m                     \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 32\u001b[0;31m                     validation_data=(x_test, y_test_c))\n\u001b[0m\u001b[1;32m     33\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     34\u001b[0m \u001b[0mplot_history\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhistorySGD\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Hist\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    197\u001b[0m                     \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 199\u001b[0;31m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    200\u001b[0m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    201\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2713\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2714\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2715\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2716\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2717\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2673\u001b[0m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_metadata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2675\u001b[0;31m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2676\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2677\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "#Model 5 - 1000 neurons \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 50\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(1000, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer='Adam',\n",
    "              metrics=['accuracy'])\n",
    "    \n",
    "    # Fitting:\n",
    "historySGD = model.fit(x_train, y_train_c,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "    # Optimizer: Adam"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 5: Going deeper: tricks and regularization (8 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "Adding hidden layers to a deep network does not necessarily lead to a straight-forward improvement of performance. Overfitting can be counteracted with regularization and dropout. Batch normalization is supposed to mainly speed up convergence. Since the MNIST dataset is almost perfectly solved already by a one-hidden-layer network we use the Fashion-MNIST dataset in this exercise.\n",
    "\n",
    "1. Add one or two hidden layers with 50 hidden neurons (each) and train the network for a sufficiently long time (at least 100 epochs). Since deep models are very expressive you will most probably encounter overfitting. Try to improve the best validation scores of the model (even if it is only a minor improvement) by experimenting with batch_normalization layers, dropout layers and l1- and l2-regularization on weights (kernels) and biases. (4 pts)\n",
    "2. After you have found good settings, plot the learning curves for both models, naive (=no tricks/regularization) and tuned (=tricks + regularized), preferably together in a comparison plot. Discuss your results; refer to the model performance with only 1 hidden layer. (2 sentences max.) (2pts)\n",
    "3. Fit your best performing (probably regularized deep) model also to MNIST for having a reference for the next exercise. Plot the resulting learning curves. (2 pts)\n",
    "\n",
    "### Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T16:26:50.480763Z",
     "start_time": "2018-02-23T16:06:32.938435Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_71 (Dense)             (None, 50)                39250     \n",
      "_________________________________________________________________\n",
      "dense_72 (Dense)             (None, 10)                510       \n",
      "=================================================================\n",
      "Total params: 39,760\n",
      "Trainable params: 39,760\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/100\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.6703 - acc: 0.7701 - val_loss: 0.5467 - val_acc: 0.8016\n",
      "Epoch 2/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.4741 - acc: 0.8330 - val_loss: 0.5301 - val_acc: 0.8067\n",
      "Epoch 3/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.4314 - acc: 0.8467 - val_loss: 0.4311 - val_acc: 0.8496\n",
      "Epoch 4/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.4042 - acc: 0.8571 - val_loss: 0.4319 - val_acc: 0.8450\n",
      "Epoch 5/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.3870 - acc: 0.8624 - val_loss: 0.4168 - val_acc: 0.8534\n",
      "Epoch 6/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.3731 - acc: 0.8671 - val_loss: 0.4043 - val_acc: 0.8585\n",
      "Epoch 7/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.3616 - acc: 0.8711 - val_loss: 0.4023 - val_acc: 0.8569\n",
      "Epoch 8/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.3504 - acc: 0.8740 - val_loss: 0.3974 - val_acc: 0.8581\n",
      "Epoch 9/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.3425 - acc: 0.8769 - val_loss: 0.4008 - val_acc: 0.8585\n",
      "Epoch 10/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.3353 - acc: 0.8791 - val_loss: 0.3787 - val_acc: 0.8660\n",
      "Epoch 11/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.3281 - acc: 0.8827 - val_loss: 0.4034 - val_acc: 0.8559\n",
      "Epoch 12/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.3217 - acc: 0.8847 - val_loss: 0.3779 - val_acc: 0.8632\n",
      "Epoch 13/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.3153 - acc: 0.8863 - val_loss: 0.3724 - val_acc: 0.8647\n",
      "Epoch 14/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.3109 - acc: 0.8871 - val_loss: 0.3643 - val_acc: 0.8703\n",
      "Epoch 15/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.3048 - acc: 0.8897 - val_loss: 0.3603 - val_acc: 0.8667\n",
      "Epoch 16/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2994 - acc: 0.8903 - val_loss: 0.3969 - val_acc: 0.8552\n",
      "Epoch 17/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2972 - acc: 0.8922 - val_loss: 0.3699 - val_acc: 0.8677\n",
      "Epoch 18/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2921 - acc: 0.8952 - val_loss: 0.3593 - val_acc: 0.8726\n",
      "Epoch 19/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2876 - acc: 0.8960 - val_loss: 0.3511 - val_acc: 0.8741\n",
      "Epoch 20/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2851 - acc: 0.8967 - val_loss: 0.3552 - val_acc: 0.8731\n",
      "Epoch 21/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2803 - acc: 0.8990 - val_loss: 0.3727 - val_acc: 0.8612\n",
      "Epoch 22/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2775 - acc: 0.8984 - val_loss: 0.3502 - val_acc: 0.8755\n",
      "Epoch 23/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.2744 - acc: 0.9003 - val_loss: 0.3780 - val_acc: 0.8593\n",
      "Epoch 24/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2714 - acc: 0.9009 - val_loss: 0.3510 - val_acc: 0.8747\n",
      "Epoch 25/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2677 - acc: 0.9032 - val_loss: 0.3640 - val_acc: 0.8693\n",
      "Epoch 26/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2657 - acc: 0.9032 - val_loss: 0.3444 - val_acc: 0.8761\n",
      "Epoch 27/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2634 - acc: 0.9041 - val_loss: 0.3396 - val_acc: 0.8805\n",
      "Epoch 28/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2602 - acc: 0.9055 - val_loss: 0.3749 - val_acc: 0.8608\n",
      "Epoch 29/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2566 - acc: 0.9068 - val_loss: 0.3446 - val_acc: 0.8775\n",
      "Epoch 30/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2543 - acc: 0.9078 - val_loss: 0.3637 - val_acc: 0.8677\n",
      "Epoch 31/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2524 - acc: 0.9081 - val_loss: 0.3435 - val_acc: 0.8772\n",
      "Epoch 32/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2505 - acc: 0.9095 - val_loss: 0.3560 - val_acc: 0.8683\n",
      "Epoch 33/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2481 - acc: 0.9094 - val_loss: 0.3878 - val_acc: 0.8584\n",
      "Epoch 34/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2444 - acc: 0.9116 - val_loss: 0.3432 - val_acc: 0.8768\n",
      "Epoch 35/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2418 - acc: 0.9133 - val_loss: 0.3380 - val_acc: 0.8812\n",
      "Epoch 36/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2404 - acc: 0.9127 - val_loss: 0.3345 - val_acc: 0.8809\n",
      "Epoch 37/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2392 - acc: 0.9126 - val_loss: 0.3364 - val_acc: 0.8809\n",
      "Epoch 38/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2361 - acc: 0.9133 - val_loss: 0.3650 - val_acc: 0.8703\n",
      "Epoch 39/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.2346 - acc: 0.9149 - val_loss: 0.3658 - val_acc: 0.8728\n",
      "Epoch 40/100\n",
      "60000/60000 [==============================] - 1s 16us/step - loss: 0.2325 - acc: 0.9159 - val_loss: 0.3342 - val_acc: 0.8836\n",
      "Epoch 41/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2305 - acc: 0.9168 - val_loss: 0.3533 - val_acc: 0.8747\n",
      "Epoch 42/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2302 - acc: 0.9167 - val_loss: 0.3563 - val_acc: 0.8732\n",
      "Epoch 43/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2267 - acc: 0.9179 - val_loss: 0.3534 - val_acc: 0.8762\n",
      "Epoch 44/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2262 - acc: 0.9179 - val_loss: 0.3650 - val_acc: 0.8717\n",
      "Epoch 45/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2240 - acc: 0.9188 - val_loss: 0.3524 - val_acc: 0.8752\n",
      "Epoch 46/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2223 - acc: 0.9185 - val_loss: 0.3425 - val_acc: 0.8755\n",
      "Epoch 47/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2201 - acc: 0.9200 - val_loss: 0.3530 - val_acc: 0.8758\n",
      "Epoch 48/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2178 - acc: 0.9209 - val_loss: 0.3356 - val_acc: 0.8806\n",
      "Epoch 49/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2174 - acc: 0.9212 - val_loss: 0.3792 - val_acc: 0.8667\n",
      "Epoch 50/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2160 - acc: 0.9216 - val_loss: 0.3444 - val_acc: 0.8814\n",
      "Epoch 51/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2130 - acc: 0.9224 - val_loss: 0.3585 - val_acc: 0.8746\n",
      "Epoch 52/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2123 - acc: 0.9227 - val_loss: 0.3409 - val_acc: 0.8820\n",
      "Epoch 53/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2114 - acc: 0.9235 - val_loss: 0.3465 - val_acc: 0.8774\n",
      "Epoch 54/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2085 - acc: 0.9247 - val_loss: 0.3460 - val_acc: 0.8801\n",
      "Epoch 55/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2087 - acc: 0.9243 - val_loss: 0.3575 - val_acc: 0.8723\n",
      "Epoch 56/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2065 - acc: 0.9253 - val_loss: 0.3820 - val_acc: 0.8715\n",
      "Epoch 57/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2064 - acc: 0.9244 - val_loss: 0.3392 - val_acc: 0.8819\n",
      "Epoch 58/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2028 - acc: 0.9263 - val_loss: 0.3461 - val_acc: 0.8824\n",
      "Epoch 59/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2031 - acc: 0.9264 - val_loss: 0.3771 - val_acc: 0.8754\n",
      "Epoch 60/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2011 - acc: 0.9263 - val_loss: 0.3453 - val_acc: 0.8819\n",
      "Epoch 61/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2013 - acc: 0.9260 - val_loss: 0.3515 - val_acc: 0.8752\n",
      "Epoch 62/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1990 - acc: 0.9283 - val_loss: 0.3826 - val_acc: 0.8714\n",
      "Epoch 63/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1982 - acc: 0.9281 - val_loss: 0.3607 - val_acc: 0.8785\n",
      "Epoch 64/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1978 - acc: 0.9288 - val_loss: 0.3479 - val_acc: 0.8823\n",
      "Epoch 65/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1929 - acc: 0.9302 - val_loss: 0.3680 - val_acc: 0.8771\n",
      "Epoch 66/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1938 - acc: 0.9295 - val_loss: 0.3530 - val_acc: 0.8814\n",
      "Epoch 67/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1929 - acc: 0.9295 - val_loss: 0.3808 - val_acc: 0.8725\n",
      "Epoch 68/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1925 - acc: 0.9302 - val_loss: 0.3514 - val_acc: 0.8823\n",
      "Epoch 69/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1909 - acc: 0.9313 - val_loss: 0.3674 - val_acc: 0.8768\n",
      "Epoch 70/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1894 - acc: 0.9316 - val_loss: 0.4136 - val_acc: 0.8642\n",
      "Epoch 71/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1877 - acc: 0.9316 - val_loss: 0.3776 - val_acc: 0.8761\n",
      "Epoch 72/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1871 - acc: 0.9326 - val_loss: 0.3643 - val_acc: 0.8737\n",
      "Epoch 73/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1869 - acc: 0.9323 - val_loss: 0.3623 - val_acc: 0.8791\n",
      "Epoch 74/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1853 - acc: 0.9324 - val_loss: 0.3603 - val_acc: 0.8796\n",
      "Epoch 75/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1835 - acc: 0.9339 - val_loss: 0.3672 - val_acc: 0.8795\n",
      "Epoch 76/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1837 - acc: 0.9331 - val_loss: 0.3784 - val_acc: 0.8770\n",
      "Epoch 77/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1799 - acc: 0.9347 - val_loss: 0.3768 - val_acc: 0.8731\n",
      "Epoch 78/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1803 - acc: 0.9351 - val_loss: 0.3761 - val_acc: 0.8754\n",
      "Epoch 79/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1789 - acc: 0.9350 - val_loss: 0.3594 - val_acc: 0.8816\n",
      "Epoch 80/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1780 - acc: 0.9356 - val_loss: 0.3591 - val_acc: 0.8823\n",
      "Epoch 81/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1781 - acc: 0.9361 - val_loss: 0.3879 - val_acc: 0.8742\n",
      "Epoch 82/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1762 - acc: 0.9369 - val_loss: 0.3721 - val_acc: 0.8812\n",
      "Epoch 83/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1753 - acc: 0.9369 - val_loss: 0.3693 - val_acc: 0.8805\n",
      "Epoch 84/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1743 - acc: 0.9379 - val_loss: 0.3668 - val_acc: 0.8791\n",
      "Epoch 85/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1728 - acc: 0.9375 - val_loss: 0.3788 - val_acc: 0.8770\n",
      "Epoch 86/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1719 - acc: 0.9385 - val_loss: 0.3663 - val_acc: 0.8789\n",
      "Epoch 87/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1714 - acc: 0.9380 - val_loss: 0.4196 - val_acc: 0.8653\n",
      "Epoch 88/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1705 - acc: 0.9386 - val_loss: 0.3881 - val_acc: 0.8755\n",
      "Epoch 89/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1693 - acc: 0.9383 - val_loss: 0.3672 - val_acc: 0.8826\n",
      "Epoch 90/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1685 - acc: 0.9398 - val_loss: 0.3616 - val_acc: 0.8827\n",
      "Epoch 91/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1678 - acc: 0.9390 - val_loss: 0.3658 - val_acc: 0.8812\n",
      "Epoch 92/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1676 - acc: 0.9398 - val_loss: 0.3800 - val_acc: 0.8760\n",
      "Epoch 93/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1674 - acc: 0.9390 - val_loss: 0.4264 - val_acc: 0.8613\n",
      "Epoch 94/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1661 - acc: 0.9394 - val_loss: 0.3720 - val_acc: 0.8799\n",
      "Epoch 95/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1653 - acc: 0.9414 - val_loss: 0.4441 - val_acc: 0.8534\n",
      "Epoch 96/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1620 - acc: 0.9421 - val_loss: 0.3749 - val_acc: 0.8784\n",
      "Epoch 97/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1630 - acc: 0.9412 - val_loss: 0.3811 - val_acc: 0.8780\n",
      "Epoch 98/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1613 - acc: 0.9418 - val_loss: 0.3864 - val_acc: 0.8767\n",
      "Epoch 99/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1612 - acc: 0.9417 - val_loss: 0.3939 - val_acc: 0.8748\n",
      "Epoch 100/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1608 - acc: 0.9416 - val_loss: 0.3969 - val_acc: 0.8754\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Model 1 - no regulrization\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 100\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(50, activation='relu', input_dim=784))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = optimizers.SGD(lr=0.1)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "# Fitting:\n",
    "history = model.fit(x_fashion_train, y_fashion_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_fashion_test, y_fashion_test))\n",
    "    \n",
    "plot_history(history, \"Hist\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_63 (Dense)             (None, 50)                39250     \n",
      "_________________________________________________________________\n",
      "dense_64 (Dense)             (None, 50)                2550      \n",
      "_________________________________________________________________\n",
      "dense_65 (Dense)             (None, 10)                510       \n",
      "=================================================================\n",
      "Total params: 42,310\n",
      "Trainable params: 42,310\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/100\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.9097 - acc: 0.7371 - val_loss: 0.7048 - val_acc: 0.8027\n",
      "Epoch 2/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.6752 - acc: 0.8131 - val_loss: 0.7381 - val_acc: 0.7776\n",
      "Epoch 3/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.6172 - acc: 0.8306 - val_loss: 0.6920 - val_acc: 0.8122\n",
      "Epoch 4/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.5827 - acc: 0.8419 - val_loss: 0.6069 - val_acc: 0.8341\n",
      "Epoch 5/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.5536 - acc: 0.8499 - val_loss: 0.5778 - val_acc: 0.8400\n",
      "Epoch 6/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.5307 - acc: 0.8552 - val_loss: 0.5545 - val_acc: 0.8489\n",
      "Epoch 7/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.5185 - acc: 0.8573 - val_loss: 0.5327 - val_acc: 0.8518\n",
      "Epoch 8/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.5008 - acc: 0.8623 - val_loss: 0.5166 - val_acc: 0.8544\n",
      "Epoch 9/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.4872 - acc: 0.8644 - val_loss: 0.5145 - val_acc: 0.8522\n",
      "Epoch 10/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.4755 - acc: 0.8672 - val_loss: 0.5286 - val_acc: 0.8502\n",
      "Epoch 11/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.4663 - acc: 0.8680 - val_loss: 0.4998 - val_acc: 0.8543\n",
      "Epoch 12/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.4555 - acc: 0.8709 - val_loss: 0.5030 - val_acc: 0.8482\n",
      "Epoch 13/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.4481 - acc: 0.8723 - val_loss: 0.4768 - val_acc: 0.8616\n",
      "Epoch 14/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.4416 - acc: 0.8742 - val_loss: 0.4804 - val_acc: 0.8590\n",
      "Epoch 15/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.4337 - acc: 0.8748 - val_loss: 0.5265 - val_acc: 0.8469\n",
      "Epoch 16/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.4265 - acc: 0.8772 - val_loss: 0.4671 - val_acc: 0.8653\n",
      "Epoch 17/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.4235 - acc: 0.8768 - val_loss: 0.4773 - val_acc: 0.8564\n",
      "Epoch 18/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.4189 - acc: 0.8778 - val_loss: 0.4618 - val_acc: 0.8632\n",
      "Epoch 19/100\n",
      "37248/60000 [=================>............] - ETA: 0s - loss: 0.4134 - acc: 0.8816"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-39-aa197fa162c8>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     33\u001b[0m                     \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     34\u001b[0m                     \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 35\u001b[0;31m                     validation_data=(x_fashion_test, y_fashion_test))\n\u001b[0m\u001b[1;32m     36\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m \u001b[0mplot_history\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhistoryL2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"Hist\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    197\u001b[0m                     \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 199\u001b[0;31m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    200\u001b[0m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    201\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2713\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2714\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2715\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2716\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2717\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2673\u001b[0m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_metadata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2675\u001b[0;31m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2676\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2677\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "#Model 2 - l2 regulrized. \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "from keras import regularizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 100\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(50, activation='relu', kernel_regularizer=regularizers.l2(0.001), input_dim=784))\n",
    "model.add(Dense(50, activation='relu', kernel_regularizer=regularizers.l2(0.001)))\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = optimizers.SGD(lr=0.05)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "    # Fitting:\n",
    "historyL2 = model.fit(x_fashion_train, y_fashion_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_fashion_test, y_fashion_test))\n",
    "    \n",
    "plot_history(historyL2, \"Hist\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Model 2.1 - L1 regularized\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "from keras import regularizers\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 100\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "#Input layer: 784 neurons\n",
    "\n",
    "#Hidden layer: 128 neurons\n",
    "model.add(Dense(50, activation='relu', kernel_regularizer=regularizers.l1(0.01), input_dim=784))\n",
    "model.add(Dense(50, activation='relu', kernel_regularizer=regularizers.l1(0.01)))\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = optimizers.SGD(lr=0.05)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "    # Fitting:\n",
    "historyL1 = model.fit(x_fashion_train, y_fashion_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_fashion_test, y_fashion_test))\n",
    "    \n",
    "plot_history(historyL1, \"Hist\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_66 (Dense)             (None, 50)                39250     \n",
      "_________________________________________________________________\n",
      "batch_normalization_8 (Batch (None, 50)                200       \n",
      "_________________________________________________________________\n",
      "dense_67 (Dense)             (None, 50)                2550      \n",
      "_________________________________________________________________\n",
      "batch_normalization_9 (Batch (None, 50)                200       \n",
      "_________________________________________________________________\n",
      "dense_68 (Dense)             (None, 10)                510       \n",
      "=================================================================\n",
      "Total params: 42,710\n",
      "Trainable params: 42,510\n",
      "Non-trainable params: 200\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/100\n",
      "60000/60000 [==============================] - 3s 56us/step - loss: 0.5444 - acc: 0.8134 - val_loss: 0.4655 - val_acc: 0.8373\n",
      "Epoch 2/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.4072 - acc: 0.8561 - val_loss: 0.4508 - val_acc: 0.8423\n",
      "Epoch 3/100\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.3732 - acc: 0.8663 - val_loss: 0.4113 - val_acc: 0.8556\n",
      "Epoch 4/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.3523 - acc: 0.8741 - val_loss: 0.4048 - val_acc: 0.8590\n",
      "Epoch 5/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.3375 - acc: 0.8790 - val_loss: 0.4081 - val_acc: 0.8544\n",
      "Epoch 6/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.3249 - acc: 0.8837 - val_loss: 0.3900 - val_acc: 0.8605\n",
      "Epoch 7/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.3128 - acc: 0.8879 - val_loss: 0.3797 - val_acc: 0.8644\n",
      "Epoch 8/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.3044 - acc: 0.8909 - val_loss: 0.3772 - val_acc: 0.8700\n",
      "Epoch 9/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.2946 - acc: 0.8945 - val_loss: 0.3782 - val_acc: 0.8659\n",
      "Epoch 10/100\n",
      "60000/60000 [==============================] - 2s 29us/step - loss: 0.2885 - acc: 0.8965 - val_loss: 0.3700 - val_acc: 0.8707\n",
      "Epoch 11/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2830 - acc: 0.8964 - val_loss: 0.3669 - val_acc: 0.8748\n",
      "Epoch 12/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2763 - acc: 0.8999 - val_loss: 0.3705 - val_acc: 0.8678\n",
      "Epoch 13/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2707 - acc: 0.9020 - val_loss: 0.4047 - val_acc: 0.8591\n",
      "Epoch 14/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2649 - acc: 0.9031 - val_loss: 0.3760 - val_acc: 0.8697\n",
      "Epoch 15/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2624 - acc: 0.9041 - val_loss: 0.3863 - val_acc: 0.8647\n",
      "Epoch 16/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2559 - acc: 0.9075 - val_loss: 0.3712 - val_acc: 0.8734\n",
      "Epoch 17/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2517 - acc: 0.9072 - val_loss: 0.3683 - val_acc: 0.8711\n",
      "Epoch 18/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.2478 - acc: 0.9099 - val_loss: 0.3797 - val_acc: 0.8655\n",
      "Epoch 19/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2447 - acc: 0.9112 - val_loss: 0.3811 - val_acc: 0.8680\n",
      "Epoch 20/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2408 - acc: 0.9133 - val_loss: 0.3643 - val_acc: 0.8772\n",
      "Epoch 21/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2367 - acc: 0.9129 - val_loss: 0.3797 - val_acc: 0.8705\n",
      "Epoch 22/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2327 - acc: 0.9145 - val_loss: 0.4082 - val_acc: 0.8592\n",
      "Epoch 23/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2280 - acc: 0.9171 - val_loss: 0.4087 - val_acc: 0.8608\n",
      "Epoch 24/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.2262 - acc: 0.9178 - val_loss: 0.3753 - val_acc: 0.8715\n",
      "Epoch 25/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2218 - acc: 0.9185 - val_loss: 0.3829 - val_acc: 0.8730\n",
      "Epoch 26/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2191 - acc: 0.9196 - val_loss: 0.3779 - val_acc: 0.8727\n",
      "Epoch 27/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2177 - acc: 0.9206 - val_loss: 0.3858 - val_acc: 0.8702\n",
      "Epoch 28/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2127 - acc: 0.9227 - val_loss: 0.3677 - val_acc: 0.8748\n",
      "Epoch 29/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2120 - acc: 0.9224 - val_loss: 0.4041 - val_acc: 0.8658\n",
      "Epoch 30/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2088 - acc: 0.9232 - val_loss: 0.3976 - val_acc: 0.8701\n",
      "Epoch 31/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.2071 - acc: 0.9245 - val_loss: 0.3957 - val_acc: 0.8679\n",
      "Epoch 32/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.2026 - acc: 0.9263 - val_loss: 0.4085 - val_acc: 0.8588\n",
      "Epoch 33/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.2020 - acc: 0.9256 - val_loss: 0.4362 - val_acc: 0.8553\n",
      "Epoch 34/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1991 - acc: 0.9271 - val_loss: 0.4102 - val_acc: 0.8656\n",
      "Epoch 35/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1967 - acc: 0.9286 - val_loss: 0.3853 - val_acc: 0.8708\n",
      "Epoch 36/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1932 - acc: 0.9289 - val_loss: 0.4277 - val_acc: 0.8588\n",
      "Epoch 37/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1912 - acc: 0.9289 - val_loss: 0.4052 - val_acc: 0.8710\n",
      "Epoch 38/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1900 - acc: 0.9300 - val_loss: 0.3863 - val_acc: 0.8741\n",
      "Epoch 39/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1879 - acc: 0.9308 - val_loss: 0.4180 - val_acc: 0.8673\n",
      "Epoch 40/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1843 - acc: 0.9329 - val_loss: 0.4419 - val_acc: 0.8590\n",
      "Epoch 41/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1843 - acc: 0.9321 - val_loss: 0.4472 - val_acc: 0.8594\n",
      "Epoch 42/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1807 - acc: 0.9334 - val_loss: 0.4179 - val_acc: 0.8683\n",
      "Epoch 43/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1803 - acc: 0.9330 - val_loss: 0.4227 - val_acc: 0.8652\n",
      "Epoch 44/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1786 - acc: 0.9345 - val_loss: 0.4097 - val_acc: 0.8687\n",
      "Epoch 45/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1761 - acc: 0.9361 - val_loss: 0.3983 - val_acc: 0.8740\n",
      "Epoch 46/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1759 - acc: 0.9345 - val_loss: 0.4463 - val_acc: 0.8605\n",
      "Epoch 47/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1720 - acc: 0.9374 - val_loss: 0.4548 - val_acc: 0.8529\n",
      "Epoch 48/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1709 - acc: 0.9369 - val_loss: 0.4258 - val_acc: 0.8662\n",
      "Epoch 49/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.1688 - acc: 0.9380 - val_loss: 0.4297 - val_acc: 0.8654\n",
      "Epoch 50/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.1666 - acc: 0.9383 - val_loss: 0.4473 - val_acc: 0.8629\n",
      "Epoch 51/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1657 - acc: 0.9396 - val_loss: 0.4509 - val_acc: 0.8595\n",
      "Epoch 52/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.1668 - acc: 0.9389 - val_loss: 0.4182 - val_acc: 0.8723\n",
      "Epoch 53/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1614 - acc: 0.9404 - val_loss: 0.4230 - val_acc: 0.8692\n",
      "Epoch 54/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1599 - acc: 0.9408 - val_loss: 0.4242 - val_acc: 0.8706\n",
      "Epoch 55/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1600 - acc: 0.9414 - val_loss: 0.4546 - val_acc: 0.8638\n",
      "Epoch 56/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1586 - acc: 0.9426 - val_loss: 0.4687 - val_acc: 0.8590\n",
      "Epoch 57/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1563 - acc: 0.9426 - val_loss: 0.4344 - val_acc: 0.8708\n",
      "Epoch 58/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1571 - acc: 0.9424 - val_loss: 0.4448 - val_acc: 0.8686\n",
      "Epoch 59/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1529 - acc: 0.9451 - val_loss: 0.4500 - val_acc: 0.8655\n",
      "Epoch 60/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1524 - acc: 0.9441 - val_loss: 0.4283 - val_acc: 0.8663\n",
      "Epoch 61/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1526 - acc: 0.9425 - val_loss: 0.4475 - val_acc: 0.8699\n",
      "Epoch 62/100\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.1498 - acc: 0.9444 - val_loss: 0.4504 - val_acc: 0.8644\n",
      "Epoch 63/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.1481 - acc: 0.9452 - val_loss: 0.4717 - val_acc: 0.8664\n",
      "Epoch 64/100\n",
      "60000/60000 [==============================] - 2s 38us/step - loss: 0.1462 - acc: 0.9455 - val_loss: 0.4436 - val_acc: 0.8650\n",
      "Epoch 65/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1458 - acc: 0.9470 - val_loss: 0.4745 - val_acc: 0.8572\n",
      "Epoch 66/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.1448 - acc: 0.9463 - val_loss: 0.4457 - val_acc: 0.8677\n",
      "Epoch 67/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.1436 - acc: 0.9472 - val_loss: 0.4567 - val_acc: 0.8696\n",
      "Epoch 68/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1400 - acc: 0.9487 - val_loss: 0.4738 - val_acc: 0.8653\n",
      "Epoch 69/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1391 - acc: 0.9498 - val_loss: 0.4560 - val_acc: 0.8663\n",
      "Epoch 70/100\n",
      "60000/60000 [==============================] - 2s 30us/step - loss: 0.1413 - acc: 0.9487 - val_loss: 0.4845 - val_acc: 0.8534\n",
      "Epoch 71/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1392 - acc: 0.9493 - val_loss: 0.5006 - val_acc: 0.8577\n",
      "Epoch 72/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.1366 - acc: 0.9499 - val_loss: 0.5078 - val_acc: 0.8689\n",
      "Epoch 73/100\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.1382 - acc: 0.9491 - val_loss: 0.4734 - val_acc: 0.8628\n",
      "Epoch 74/100\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.1358 - acc: 0.9498 - val_loss: 0.5079 - val_acc: 0.8590\n",
      "Epoch 75/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1331 - acc: 0.9512 - val_loss: 0.5207 - val_acc: 0.8542\n",
      "Epoch 76/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1328 - acc: 0.9509 - val_loss: 0.4719 - val_acc: 0.8557\n",
      "Epoch 77/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1316 - acc: 0.9525 - val_loss: 0.4827 - val_acc: 0.8658\n",
      "Epoch 78/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1320 - acc: 0.9520 - val_loss: 0.5039 - val_acc: 0.8617\n",
      "Epoch 79/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.1314 - acc: 0.9518 - val_loss: 0.5273 - val_acc: 0.8692\n",
      "Epoch 80/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1295 - acc: 0.9523 - val_loss: 0.5220 - val_acc: 0.8616\n",
      "Epoch 81/100\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.1280 - acc: 0.9527 - val_loss: 0.4891 - val_acc: 0.8694\n",
      "Epoch 82/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1285 - acc: 0.9527 - val_loss: 0.5392 - val_acc: 0.8582\n",
      "Epoch 83/100\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.1255 - acc: 0.9539 - val_loss: 0.5225 - val_acc: 0.8589\n",
      "Epoch 84/100\n",
      "60000/60000 [==============================] - 2s 34us/step - loss: 0.1281 - acc: 0.9523 - val_loss: 0.5049 - val_acc: 0.8659\n",
      "Epoch 85/100\n",
      "60000/60000 [==============================] - 2s 39us/step - loss: 0.1251 - acc: 0.9542 - val_loss: 0.4945 - val_acc: 0.8652\n",
      "Epoch 86/100\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.1224 - acc: 0.9553 - val_loss: 0.5447 - val_acc: 0.8613\n",
      "Epoch 87/100\n",
      "60000/60000 [==============================] - 2s 37us/step - loss: 0.1246 - acc: 0.9541 - val_loss: 0.5058 - val_acc: 0.8664\n",
      "Epoch 88/100\n",
      "60000/60000 [==============================] - 2s 36us/step - loss: 0.1223 - acc: 0.9550 - val_loss: 0.5093 - val_acc: 0.8640\n",
      "Epoch 89/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1203 - acc: 0.9560 - val_loss: 0.5454 - val_acc: 0.8514\n",
      "Epoch 90/100\n",
      "60000/60000 [==============================] - 2s 35us/step - loss: 0.1219 - acc: 0.9546 - val_loss: 0.5043 - val_acc: 0.8606\n",
      "Epoch 91/100\n",
      "60000/60000 [==============================] - 2s 31us/step - loss: 0.1187 - acc: 0.9566 - val_loss: 0.5416 - val_acc: 0.8595\n",
      "Epoch 92/100\n",
      "60000/60000 [==============================] - 2s 30us/step - loss: 0.1196 - acc: 0.9555 - val_loss: 0.5241 - val_acc: 0.8614\n",
      "Epoch 93/100\n",
      "60000/60000 [==============================] - 2s 33us/step - loss: 0.1187 - acc: 0.9564 - val_loss: 0.5309 - val_acc: 0.8608\n",
      "Epoch 94/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1177 - acc: 0.9567 - val_loss: 0.5178 - val_acc: 0.8598\n",
      "Epoch 95/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1169 - acc: 0.9574 - val_loss: 0.5068 - val_acc: 0.8584\n",
      "Epoch 96/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1152 - acc: 0.9579 - val_loss: 0.5689 - val_acc: 0.8493\n",
      "Epoch 97/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1135 - acc: 0.9577 - val_loss: 0.5752 - val_acc: 0.8611\n",
      "Epoch 98/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1147 - acc: 0.9578 - val_loss: 0.5892 - val_acc: 0.8467\n",
      "Epoch 99/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1128 - acc: 0.9579 - val_loss: 0.5292 - val_acc: 0.8652\n",
      "Epoch 100/100\n",
      "60000/60000 [==============================] - 2s 32us/step - loss: 0.1116 - acc: 0.9587 - val_loss: 0.5412 - val_acc: 0.8634\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Model 3 - Batch normalized \n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Activation\n",
    "from keras import optimizers\n",
    "from keras import regularizers\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "\n",
    "batch_size = 128\n",
    "epochs = 100\n",
    "\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Dense(50, input_dim=784))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Dense(50, activation='relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = optimizers.SGD(lr=0.05)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "# Fitting:\n",
    "historyBN = model.fit(x_fashion_train, y_fashion_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_fashion_test, y_fashion_test))\n",
    "    \n",
    "plot_history(historyBN, \"Hist\")\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'optimizers' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-26-a106bd2bde22>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     11\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mDense\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactivation\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'softmax'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     12\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 13\u001b[0;31m \u001b[0msgd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0moptimizers\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mSGD\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     14\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     15\u001b[0m model.compile(loss='categorical_crossentropy',\n",
      "\u001b[0;31mNameError\u001b[0m: name 'optimizers' is not defined"
     ]
    }
   ],
   "source": [
    "# Dropout\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(50, input_dim=784, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = optimizers.SGD(lr=0.1)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "# Fitting:\n",
    "historyDropout = model.fit(x_fashion_train, y_fashion_train,\n",
    "                    batch_size=batch_size,\n",
    "                    epochs=epochs,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_fashion_test, y_fashion_test))\n",
    "    \n",
    "plot_history(historySGD, \"Hist\")\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 2 (comments):"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "comparison_plot(history,historyDropout, \"Normal\", \"Dropout\", \"Title\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The dropout regularized network seems to converge slower on the trining data to the global minimum of the cross entropy loss function. However, the validation error seems to converge while the validation error for the normal network increases over the number of epochs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_84 (Dense)             (None, 50)                39250     \n",
      "_________________________________________________________________\n",
      "dropout_56 (Dropout)         (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_85 (Dense)             (None, 50)                2550      \n",
      "_________________________________________________________________\n",
      "dropout_57 (Dropout)         (None, 50)                0         \n",
      "_________________________________________________________________\n",
      "dense_86 (Dense)             (None, 10)                510       \n",
      "=================================================================\n",
      "Total params: 42,310\n",
      "Trainable params: 42,310\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/100\n",
      "60000/60000 [==============================] - 2s 29us/step - loss: 0.7639 - acc: 0.7591 - val_loss: 0.2960 - val_acc: 0.9132\n",
      "Epoch 2/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.4025 - acc: 0.8810 - val_loss: 0.2199 - val_acc: 0.9329\n",
      "Epoch 3/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.3351 - acc: 0.9026 - val_loss: 0.1911 - val_acc: 0.9428\n",
      "Epoch 4/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2923 - acc: 0.9162 - val_loss: 0.1749 - val_acc: 0.9466\n",
      "Epoch 5/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.2668 - acc: 0.9223 - val_loss: 0.1562 - val_acc: 0.9527\n",
      "Epoch 6/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.2480 - acc: 0.9289 - val_loss: 0.1433 - val_acc: 0.9566\n",
      "Epoch 7/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.2288 - acc: 0.9343 - val_loss: 0.1445 - val_acc: 0.9557\n",
      "Epoch 8/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.2172 - acc: 0.9371 - val_loss: 0.1347 - val_acc: 0.9587\n",
      "Epoch 9/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.2081 - acc: 0.9390 - val_loss: 0.1288 - val_acc: 0.9622\n",
      "Epoch 10/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.2007 - acc: 0.9417 - val_loss: 0.1246 - val_acc: 0.9634\n",
      "Epoch 11/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1923 - acc: 0.9438 - val_loss: 0.1211 - val_acc: 0.9625\n",
      "Epoch 12/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1872 - acc: 0.9451 - val_loss: 0.1205 - val_acc: 0.9633\n",
      "Epoch 13/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1809 - acc: 0.9476 - val_loss: 0.1175 - val_acc: 0.9650\n",
      "Epoch 14/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1727 - acc: 0.9490 - val_loss: 0.1132 - val_acc: 0.9667\n",
      "Epoch 15/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1677 - acc: 0.9511 - val_loss: 0.1129 - val_acc: 0.9662\n",
      "Epoch 16/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1652 - acc: 0.9520 - val_loss: 0.1074 - val_acc: 0.9677\n",
      "Epoch 17/100\n",
      "60000/60000 [==============================] - 2s 27us/step - loss: 0.1599 - acc: 0.9527 - val_loss: 0.1099 - val_acc: 0.9657\n",
      "Epoch 18/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.1581 - acc: 0.9535 - val_loss: 0.1070 - val_acc: 0.9677\n",
      "Epoch 19/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1548 - acc: 0.9547 - val_loss: 0.1064 - val_acc: 0.9695\n",
      "Epoch 20/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1504 - acc: 0.9552 - val_loss: 0.1022 - val_acc: 0.9703\n",
      "Epoch 21/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1484 - acc: 0.9563 - val_loss: 0.1014 - val_acc: 0.9707\n",
      "Epoch 22/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.1447 - acc: 0.9567 - val_loss: 0.1068 - val_acc: 0.9685\n",
      "Epoch 23/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.1436 - acc: 0.9569 - val_loss: 0.1017 - val_acc: 0.9698\n",
      "Epoch 24/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1405 - acc: 0.9572 - val_loss: 0.1005 - val_acc: 0.9714\n",
      "Epoch 25/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1371 - acc: 0.9598 - val_loss: 0.1010 - val_acc: 0.9695\n",
      "Epoch 26/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1395 - acc: 0.9576 - val_loss: 0.1027 - val_acc: 0.9703\n",
      "Epoch 27/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1366 - acc: 0.9590 - val_loss: 0.0997 - val_acc: 0.9729\n",
      "Epoch 28/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1311 - acc: 0.9602 - val_loss: 0.1005 - val_acc: 0.9710\n",
      "Epoch 29/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1311 - acc: 0.9615 - val_loss: 0.1000 - val_acc: 0.9699\n",
      "Epoch 30/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1315 - acc: 0.9604 - val_loss: 0.0979 - val_acc: 0.9704\n",
      "Epoch 31/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1300 - acc: 0.9614 - val_loss: 0.0963 - val_acc: 0.9718\n",
      "Epoch 32/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1262 - acc: 0.9619 - val_loss: 0.1013 - val_acc: 0.9702\n",
      "Epoch 33/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1251 - acc: 0.9626 - val_loss: 0.0989 - val_acc: 0.9712\n",
      "Epoch 34/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1230 - acc: 0.9624 - val_loss: 0.1004 - val_acc: 0.9715\n",
      "Epoch 35/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1197 - acc: 0.9642 - val_loss: 0.0997 - val_acc: 0.9716\n",
      "Epoch 36/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1203 - acc: 0.9636 - val_loss: 0.0980 - val_acc: 0.9720\n",
      "Epoch 37/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1197 - acc: 0.9634 - val_loss: 0.0967 - val_acc: 0.9714\n",
      "Epoch 38/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1206 - acc: 0.9630 - val_loss: 0.0966 - val_acc: 0.9713\n",
      "Epoch 39/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1175 - acc: 0.9637 - val_loss: 0.0964 - val_acc: 0.9733\n",
      "Epoch 40/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1195 - acc: 0.9639 - val_loss: 0.0951 - val_acc: 0.9725\n",
      "Epoch 41/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1171 - acc: 0.9645 - val_loss: 0.0960 - val_acc: 0.9722\n",
      "Epoch 42/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1129 - acc: 0.9663 - val_loss: 0.0997 - val_acc: 0.9717\n",
      "Epoch 43/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1148 - acc: 0.9652 - val_loss: 0.1008 - val_acc: 0.9713\n",
      "Epoch 44/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1123 - acc: 0.9659 - val_loss: 0.1000 - val_acc: 0.9727\n",
      "Epoch 45/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.1096 - acc: 0.9667 - val_loss: 0.0944 - val_acc: 0.9734\n",
      "Epoch 46/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.1120 - acc: 0.9660 - val_loss: 0.0973 - val_acc: 0.9724\n",
      "Epoch 47/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1106 - acc: 0.9664 - val_loss: 0.0993 - val_acc: 0.9722\n",
      "Epoch 48/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1068 - acc: 0.9672 - val_loss: 0.0978 - val_acc: 0.9713\n",
      "Epoch 49/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1083 - acc: 0.9669 - val_loss: 0.0977 - val_acc: 0.9726\n",
      "Epoch 50/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1040 - acc: 0.9682 - val_loss: 0.0973 - val_acc: 0.9734\n",
      "Epoch 51/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.1067 - acc: 0.9675 - val_loss: 0.0991 - val_acc: 0.9714\n",
      "Epoch 52/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.1091 - acc: 0.9663 - val_loss: 0.1015 - val_acc: 0.9710\n",
      "Epoch 53/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.1073 - acc: 0.9666 - val_loss: 0.1013 - val_acc: 0.9722\n",
      "Epoch 54/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1082 - acc: 0.9675 - val_loss: 0.1012 - val_acc: 0.9723\n",
      "Epoch 55/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1034 - acc: 0.9682 - val_loss: 0.0982 - val_acc: 0.9734\n",
      "Epoch 56/100\n",
      "60000/60000 [==============================] - 1s 22us/step - loss: 0.1030 - acc: 0.9680 - val_loss: 0.0980 - val_acc: 0.9731\n",
      "Epoch 57/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.0987 - acc: 0.9685 - val_loss: 0.1003 - val_acc: 0.9720\n",
      "Epoch 58/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1030 - acc: 0.9679 - val_loss: 0.0993 - val_acc: 0.9725\n",
      "Epoch 59/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.1006 - acc: 0.9689 - val_loss: 0.1020 - val_acc: 0.9726\n",
      "Epoch 60/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.1020 - acc: 0.9681 - val_loss: 0.1013 - val_acc: 0.9736\n",
      "Epoch 61/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1003 - acc: 0.9692 - val_loss: 0.1025 - val_acc: 0.9719\n",
      "Epoch 62/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1023 - acc: 0.9687 - val_loss: 0.1040 - val_acc: 0.9717\n",
      "Epoch 63/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.1001 - acc: 0.9689 - val_loss: 0.0979 - val_acc: 0.9725\n",
      "Epoch 64/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.0980 - acc: 0.9698 - val_loss: 0.0979 - val_acc: 0.9728\n",
      "Epoch 65/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0977 - acc: 0.9699 - val_loss: 0.0995 - val_acc: 0.9729\n",
      "Epoch 66/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0974 - acc: 0.9699 - val_loss: 0.0988 - val_acc: 0.9738\n",
      "Epoch 67/100\n",
      "60000/60000 [==============================] - 2s 25us/step - loss: 0.0947 - acc: 0.9702 - val_loss: 0.1032 - val_acc: 0.9728\n",
      "Epoch 68/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0949 - acc: 0.9710 - val_loss: 0.1005 - val_acc: 0.9733\n",
      "Epoch 69/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0951 - acc: 0.9703 - val_loss: 0.0973 - val_acc: 0.9747\n",
      "Epoch 70/100\n",
      "60000/60000 [==============================] - 1s 20us/step - loss: 0.0963 - acc: 0.9700 - val_loss: 0.1003 - val_acc: 0.9718\n",
      "Epoch 71/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.0919 - acc: 0.9716 - val_loss: 0.0969 - val_acc: 0.9739\n",
      "Epoch 72/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.0972 - acc: 0.9694 - val_loss: 0.0968 - val_acc: 0.9731\n",
      "Epoch 73/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0935 - acc: 0.9705 - val_loss: 0.0989 - val_acc: 0.9718\n",
      "Epoch 74/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.0932 - acc: 0.9711 - val_loss: 0.0994 - val_acc: 0.9730\n",
      "Epoch 75/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0901 - acc: 0.9722 - val_loss: 0.0962 - val_acc: 0.9736\n",
      "Epoch 76/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0939 - acc: 0.9707 - val_loss: 0.1022 - val_acc: 0.9731\n",
      "Epoch 77/100\n",
      "60000/60000 [==============================] - 1s 23us/step - loss: 0.0933 - acc: 0.9707 - val_loss: 0.1017 - val_acc: 0.9727\n",
      "Epoch 78/100\n",
      "60000/60000 [==============================] - 1s 15us/step - loss: 0.0924 - acc: 0.9709 - val_loss: 0.0998 - val_acc: 0.9730\n",
      "Epoch 79/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0935 - acc: 0.9705 - val_loss: 0.0995 - val_acc: 0.9740\n",
      "Epoch 80/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0908 - acc: 0.9711 - val_loss: 0.1008 - val_acc: 0.9735\n",
      "Epoch 81/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0891 - acc: 0.9719 - val_loss: 0.1014 - val_acc: 0.9728\n",
      "Epoch 82/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0908 - acc: 0.9715 - val_loss: 0.0975 - val_acc: 0.9731\n",
      "Epoch 83/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.0876 - acc: 0.9727 - val_loss: 0.1006 - val_acc: 0.9735\n",
      "Epoch 84/100\n",
      "60000/60000 [==============================] - 1s 24us/step - loss: 0.0876 - acc: 0.9721 - val_loss: 0.1017 - val_acc: 0.9737\n",
      "Epoch 85/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0902 - acc: 0.9723 - val_loss: 0.0982 - val_acc: 0.9734\n",
      "Epoch 86/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0877 - acc: 0.9726 - val_loss: 0.1044 - val_acc: 0.9724\n",
      "Epoch 87/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0913 - acc: 0.9712 - val_loss: 0.0989 - val_acc: 0.9726\n",
      "Epoch 88/100\n",
      "60000/60000 [==============================] - 1s 19us/step - loss: 0.0887 - acc: 0.9725 - val_loss: 0.0970 - val_acc: 0.9741\n",
      "Epoch 89/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0870 - acc: 0.9727 - val_loss: 0.1017 - val_acc: 0.9725\n",
      "Epoch 90/100\n",
      "60000/60000 [==============================] - 1s 17us/step - loss: 0.0878 - acc: 0.9726 - val_loss: 0.1019 - val_acc: 0.9733\n",
      "Epoch 91/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0870 - acc: 0.9722 - val_loss: 0.1010 - val_acc: 0.9739\n",
      "Epoch 92/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0879 - acc: 0.9717 - val_loss: 0.0974 - val_acc: 0.9739\n",
      "Epoch 93/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0861 - acc: 0.9733 - val_loss: 0.1052 - val_acc: 0.9722\n",
      "Epoch 94/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0852 - acc: 0.9728 - val_loss: 0.1016 - val_acc: 0.9739\n",
      "Epoch 95/100\n",
      "60000/60000 [==============================] - 1s 18us/step - loss: 0.0848 - acc: 0.9734 - val_loss: 0.1066 - val_acc: 0.9732\n",
      "Epoch 96/100\n",
      "60000/60000 [==============================] - 1s 21us/step - loss: 0.0866 - acc: 0.9731 - val_loss: 0.1011 - val_acc: 0.9727\n",
      "Epoch 97/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.0847 - acc: 0.9733 - val_loss: 0.1015 - val_acc: 0.9725\n",
      "Epoch 98/100\n",
      "60000/60000 [==============================] - 2s 26us/step - loss: 0.0878 - acc: 0.9724 - val_loss: 0.0971 - val_acc: 0.9732\n",
      "Epoch 99/100\n",
      "60000/60000 [==============================] - 1s 25us/step - loss: 0.0850 - acc: 0.9732 - val_loss: 0.0971 - val_acc: 0.9731\n",
      "Epoch 100/100\n",
      "60000/60000 [==============================] - 2s 28us/step - loss: 0.0843 - acc: 0.9736 - val_loss: 0.1015 - val_acc: 0.9728\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Dropout regularized deep\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "model.add(Dense(50, input_dim=784, activation='relu'))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(50, activation='relu'))\n",
    "model.add(Dropout(0.3))\n",
    "\n",
    "# Output layer: 10 Neurons\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "sgd = SGD(lr=0.1)\n",
    "\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=sgd,\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.summary()\n",
    "    \n",
    "# Fitting:\n",
    "historyDropout = model.fit(x_train, y_train_c,\n",
    "                    batch_size=128,\n",
    "                    epochs=100,\n",
    "                    verbose=1,\n",
    "                    validation_data=(x_test, y_test_c))\n",
    "    \n",
    "plot_history(historyDropout, \"Hist\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 6: Convolutional neural networks (CNNs) (10 points)\n",
    "\n",
    "### Description\n",
    "\n",
    "Convolutional neural networks have an inductive bias that is well adapted to image classification.\n",
    "\n",
    "1. Design a convolutional neural network, play with different architectures and parameters. Hint: You may get valuable inspiration from the keras [examples](https://github.com/keras-team/keras/tree/master/examples). (4 pts)\n",
    "2. Plot the learning curves of the convolutional neural network for MNIST and Fashion-MNIST. (4 pts)\n",
    "3. How does the CNN performance compare to the so far best performing (deep) neural network model for the two data sets? (2 sentences max.) (2 pts)\n",
    "\n",
    "### Solution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-02-23T16:05:21.840299Z",
     "start_time": "2018-02-23T15:51:11.993053Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train shape: (60000, 28, 28, 1)\n",
      "60000 train samples\n",
      "10000 test samples\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/12\n",
      "60000/60000 [==============================] - 112s 2ms/step - loss: 0.2710 - acc: 0.9167 - val_loss: 0.0587 - val_acc: 0.9814\n",
      "Epoch 2/12\n",
      "16768/60000 [=======>......................] - ETA: 1:20 - loss: 0.1026 - acc: 0.9698"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-48-737fedbc7ffc>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     68\u001b[0m           \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mepochs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     69\u001b[0m           \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 70\u001b[0;31m           validation_data=(x_test, y_test))\n\u001b[0m\u001b[1;32m     71\u001b[0m \u001b[0mscore\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mverbose\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     72\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Test loss:'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscore\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, **kwargs)\u001b[0m\n\u001b[1;32m   1037\u001b[0m                                         \u001b[0minitial_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0minitial_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1038\u001b[0m                                         \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1039\u001b[0;31m                                         validation_steps=validation_steps)\n\u001b[0m\u001b[1;32m   1040\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1041\u001b[0m     def evaluate(self, x=None, y=None,\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/engine/training_arrays.py\u001b[0m in \u001b[0;36mfit_loop\u001b[0;34m(model, f, ins, out_labels, batch_size, epochs, verbose, callbacks, val_f, val_ins, shuffle, callback_metrics, initial_epoch, steps_per_epoch, validation_steps)\u001b[0m\n\u001b[1;32m    197\u001b[0m                     \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mins_batch\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtoarray\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    198\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 199\u001b[0;31m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mins_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    200\u001b[0m                 \u001b[0mouts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_list\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    201\u001b[0m                 \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout_labels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mouts\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2713\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_legacy_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2714\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2715\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2716\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2717\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mpy_any\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mis_tensor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py\u001b[0m in \u001b[0;36m_call\u001b[0;34m(self, inputs)\u001b[0m\n\u001b[1;32m   2673\u001b[0m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun_metadata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2674\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 2675\u001b[0;31m             \u001b[0mfetched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_callable_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0marray_vals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   2676\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mfetched\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   2677\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/anaconda3/envs/cs456env/lib/python3.6/site-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1380\u001b[0m           ret = tf_session.TF_SessionRunCallable(\n\u001b[1;32m   1381\u001b[0m               \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_handle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstatus\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1382\u001b[0;31m               run_metadata_ptr)\n\u001b[0m\u001b[1;32m   1383\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1384\u001b[0m           \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "'''Trains a simple convnet on the MNIST dataset.\n",
    "Gets to 99.25% test accuracy after 12 epochs\n",
    "(there is still a lot of margin for parameter tuning).\n",
    "16 seconds per epoch on a GRID K520 GPU.\n",
    "'''\n",
    "\n",
    "from __future__ import print_function\n",
    "import keras\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras import backend as K\n",
    "\n",
    "batch_size = 128\n",
    "num_classes = 10\n",
    "epochs = 12\n",
    "\n",
    "# input image dimensions\n",
    "img_rows, img_cols = 28, 28\n",
    "\n",
    "# the data, split between train and test sets\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape data depending on image format.\n",
    "if K.image_data_format() == 'channels_first':\n",
    "    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)\n",
    "    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)\n",
    "    input_shape = (1, img_rows, img_cols)\n",
    "else:\n",
    "    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)\n",
    "    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)\n",
    "    input_shape = (img_rows, img_cols, 1)\n",
    "\n",
    "# Normalize to values between 0-1\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "x_train /= 255\n",
    "x_test /= 255\n",
    "print('x_train shape:', x_train.shape)\n",
    "print(x_train.shape[0], 'train samples')\n",
    "print(x_test.shape[0], 'test samples')\n",
    "\n",
    "# convert class vectors to binary class matrices\n",
    "y_train = keras.utils.to_categorical(y_train, num_classes)\n",
    "y_test = keras.utils.to_categorical(y_test, num_classes)\n",
    "\n",
    "# Convolutional neural network.\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(3, 3),\n",
    "                 activation='relu',\n",
    "                 input_shape=input_shape))\n",
    "model.add(Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.25))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128, activation='relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss=keras.losses.categorical_crossentropy,\n",
    "              optimizer=keras.optimizers.Adadelta(),\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.fit(x_train, y_train,\n",
    "          batch_size=batch_size,\n",
    "          epochs=epochs,\n",
    "          verbose=1,\n",
    "          validation_data=(x_test, y_test))\n",
    "score = model.evaluate(x_test, y_test, verbose=0)\n",
    "print('Test loss:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 3:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The test accuracy and the training accuracy is significanly high. Likewise, the test and traning loss significanly lower."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exercise 7: Sigmoidal activation function and batch-normalization (6 points)\n",
    "\n",
    "### Description:\n",
    "\n",
    "In the original publication of batch normalization [Ioffe and Szegedy, 2014](https://arxiv.org/pdf/1502.03167.pdf), the authors mention a particularly beneficial effect of their method on networks with sigmoidal activation functions. This is because such networks usually suffer from saturating activations/vanishing gradients. Here we want to reproduce this behaviour (Chose either MNIST or Fashion-MNIST for this exercise).\n",
    "\n",
    "1. Implement the same convolutional network as in the previous exercise, but using the sigmoid activation function instead of the standard choice ReLU. Train the network for a reasonable amount of time. What do you observe? (1 sentence max.) (3 pts)\n",
    "2. Add batch-normalization layers to all convolutional and fully-connected layers (i.e. before each layer with learnable parameters). How does the performance change? Can the network reach the ReLU-CNN performance of the previous exercise? (1 sentence max.) (3 pts)\n",
    "3. **BONUS (optional, not graded**): Investigate our initial guess that saturating activity/vanishing gradients might be the cause of this behaviour. For that, create histograms of the hidden activitions for different hidden layers for the sigmoid-CNN and the sigmoid-CNN with batch-normalization (counting over both, samples and neurons per layer). You may only chose layers with learnable parameters. What do you observe?\n",
    "Hint: You can use the [keract](https://github.com/philipperemy/keract) package to access neural activation values for all layers of your network model.\n",
    "\n",
    "\n",
    "\n",
    "### Solution:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x_train shape: (60000, 28, 28, 1)\n",
      "60000 train samples\n",
      "10000 test samples\n",
      "Train on 60000 samples, validate on 10000 samples\n",
      "Epoch 1/12\n",
      "60000/60000 [==============================] - 99s 2ms/step - loss: 2.3173 - acc: 0.1098 - val_loss: 2.3008 - val_acc: 0.1135\n",
      "Epoch 2/12\n",
      "60000/60000 [==============================] - 116s 2ms/step - loss: 2.1662 - acc: 0.1758 - val_loss: 0.6203 - val_acc: 0.8148\n",
      "Epoch 3/12\n",
      "60000/60000 [==============================] - 117s 2ms/step - loss: 0.4377 - acc: 0.8686 - val_loss: 0.2661 - val_acc: 0.9229\n",
      "Epoch 4/12\n",
      "60000/60000 [==============================] - 109s 2ms/step - loss: 0.2991 - acc: 0.9121 - val_loss: 0.1940 - val_acc: 0.9433\n",
      "Epoch 5/12\n",
      "60000/60000 [==============================] - 99s 2ms/step - loss: 0.2512 - acc: 0.9259 - val_loss: 0.1624 - val_acc: 0.9521\n",
      "Epoch 6/12\n",
      "60000/60000 [==============================] - 101s 2ms/step - loss: 0.2260 - acc: 0.9342 - val_loss: 0.1464 - val_acc: 0.9568\n",
      "Epoch 7/12\n",
      "55680/60000 [==========================>...] - ETA: 6s - loss: 0.2089 - acc: 0.9389"
     ]
    }
   ],
   "source": [
    "'''Trains a simple convnet on the MNIST dataset.\n",
    "Gets to 99.25% test accuracy after 12 epochs\n",
    "(there is still a lot of margin for parameter tuning).\n",
    "16 seconds per epoch on a GRID K520 GPU.\n",
    "'''\n",
    "\n",
    "from __future__ import print_function\n",
    "import keras\n",
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras import backend as K\n",
    "\n",
    "batch_size = 128\n",
    "num_classes = 10\n",
    "epochs = 12\n",
    "\n",
    "# input image dimensions\n",
    "img_rows, img_cols = 28, 28\n",
    "\n",
    "# the data, split between train and test sets\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "# Reshape data depending on image format.\n",
    "if K.image_data_format() == 'channels_first':\n",
    "    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)\n",
    "    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)\n",
    "    input_shape = (1, img_rows, img_cols)\n",
    "else:\n",
    "    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)\n",
    "    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)\n",
    "    input_shape = (img_rows, img_cols, 1)\n",
    "\n",
    "# Normalize to values between 0-1\n",
    "x_train = x_train.astype('float32')\n",
    "x_test = x_test.astype('float32')\n",
    "x_train /= 255\n",
    "x_test /= 255\n",
    "print('x_train shape:', x_train.shape)\n",
    "print(x_train.shape[0], 'train samples')\n",
    "print(x_test.shape[0], 'test samples')\n",
    "\n",
    "# convert class vectors to binary class matrices\n",
    "y_train = keras.utils.to_categorical(y_train, num_classes)\n",
    "y_test = keras.utils.to_categorical(y_test, num_classes)\n",
    "\n",
    "# Convolutional neural network.\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(3, 3),\n",
    "                 activation='sigmoid',\n",
    "                 input_shape=input_shape))\n",
    "model.add(Conv2D(64, (3, 3), activation='sigmoid'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.25))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128, activation='sigmoid'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "\n",
    "model.compile(loss=keras.losses.categorical_crossentropy,\n",
    "              optimizer=keras.optimizers.Adadelta(),\n",
    "              metrics=['accuracy'])\n",
    "\n",
    "model.fit(x_train, y_train,\n",
    "          batch_size=batch_size,\n",
    "          epochs=epochs,\n",
    "          verbose=1,\n",
    "          validation_data=(x_test, y_test))\n",
    "score = model.evaluate(x_test, y_test, verbose=0)\n",
    "print('Test loss:', score[0])\n",
    "print('Test accuracy:', score[1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 1:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Answer to question 2:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "cs456env",
   "language": "python",
   "name": "cs456env"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}