keras_barebones.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "keras.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyO5u6vhyQjFkGKEtPUMyAGy",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/foxtrotmike/CS909/blob/master/keras_barebones.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OKyS2u1M6Wf9",
        "colab_type": "text"
      },
      "source": [
        "# Simple Classification Example with Keras\n",
        "\n",
        "## by Fayyaz Minhas\n",
        "\n",
        "This is a \"barebones\" example with Keras in which we train a simple neural network and visualize its decision boundary.\n",
        "\n",
        "Let's load some visualization code first (You don't need to understand this part!).\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1C2x72Q928yC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from numpy.random import randn #importing randn\n",
        "import numpy as np #importing numpy\n",
        "import matplotlib.pyplot as plt #importing plotting module\n",
        "import itertools\n",
        "import warnings\n",
        "\n",
        "def plotit(X,Y=None,clf=None,  conts = None, ccolors = ('b','k','r'), colors = ('c','y'), markers = ('s','o'), hold = False, transform = None,extent = None,**kwargs):\n",
        "    \"\"\"\n",
        "    A function for showing data scatter plot and classification boundary\n",
        "    of a classifier for 2D data\n",
        "        X: nxd  matrix of data points\n",
        "        Y: (optional) n vector of class labels\n",
        "        clf: (optional) classification/discriminant function handle\n",
        "        conts: (optional) contours (if None, contours are drawn for each class boundary)\n",
        "        ccolors: (optional) colors for contours   \n",
        "        colors: (optional) colors for each class (sorted wrt class id)\n",
        "            can be 'scaled' or 'random' or a list/tuple of color ids\n",
        "        markers: (optional) markers for each class (sorted wrt class id)\n",
        "        hold: Whether to hold the plot or not for overlay (default: False).\n",
        "        transform: (optional) a function handle for transforming data before passing to clf\n",
        "        kwargs: any keyword arguments to be passed to clf (if any)        \n",
        "    \"\"\"\n",
        "    if clf is not None and X.shape[1]!=2:\n",
        "        warnings.warn(\"Data Dimensionality is not 2. Unable to plot.\")\n",
        "        return\n",
        "    if markers is None:\n",
        "        markers = ('.',)\n",
        "    eps=1e-6\n",
        "    d0,d1 = (0,1)\n",
        "    if extent is None:\n",
        "        minx, maxx = np.min(X[:,d0])-eps, np.max(X[:,d0])+eps\n",
        "        miny, maxy = np.min(X[:,d1])-eps, np.max(X[:,d1])+eps\n",
        "        extent = [minx,maxx,miny,maxy]\n",
        "    else:\n",
        "        [minx,maxx,miny,maxy] = extent\n",
        "    if Y is not None:\n",
        "        classes = sorted(set(Y))\n",
        "        if conts is None or len(conts)<2:\n",
        "            #conts = list(classes)\n",
        "            vmin,vmax = classes[0]-eps,classes[-1]+eps\n",
        "        else:            \n",
        "            vmin,vmax= np.min(conts)-eps,np.max(conts)+eps\n",
        "        \n",
        "    else:\n",
        "        vmin,vmax=-2-eps,2+eps\n",
        "        if conts is None or len(conts)<2:            \n",
        "            conts = sorted([-1+eps,0,1-eps])\n",
        "        else:\n",
        "            vmin,vmax= np.min(conts)-eps,np.max(conts)+eps\n",
        "        \n",
        "    if clf is not None:\n",
        "        npts = 150\n",
        "        x = np.linspace(minx,maxx,npts)\n",
        "        y = np.linspace(miny,maxy,npts)\n",
        "        t = np.array(list(itertools.product(x,y)))\n",
        "        if transform is not None:\n",
        "            t = transform(t)\n",
        "        z = clf(t,**kwargs)\n",
        "        \n",
        "        z = np.reshape(z,(npts,npts)).T        \n",
        "        \n",
        "        \n",
        "        plt.contour(x,y,z,conts,linewidths = [2],colors=ccolors,extent=extent, label='f(x)=0')\n",
        "        #plt.imshow(np.flipud(z), extent = extent, cmap=plt.cm.Purples, vmin = -2, vmax = +2); plt.colorbar()\n",
        "        plt.pcolormesh(x, y, z,cmap=plt.cm.Purples,vmin=vmin,vmax=vmax);plt.colorbar()\n",
        "        plt.axis(extent)\n",
        "    \n",
        "    if Y is not None:        \n",
        "        for i,y in enumerate(classes):\n",
        "            \n",
        "            if colors is None or colors=='scaled':\n",
        "                cc = np.array([[i,i,i]])/float(len(classes))\n",
        "            elif colors =='random':\n",
        "                cc = np.array([[np.random.rand(),np.random.rand(),np.random.rand()]])\n",
        "            else:\n",
        "                cc = colors[i%len(colors)]\n",
        "            \n",
        "            mm = markers[i%len(markers)]\n",
        "            plt.scatter(X[Y==y,d0],X[Y==y,d1], marker = mm,c = cc, s = 50)     \n",
        "         \n",
        "    else:\n",
        "        plt.scatter(X[:,d0],X[:,d1],marker = markers[0], c = 'k', s = 5)\n",
        "    plt.xlabel('$x_1$')\n",
        "    plt.ylabel('$x_2$')   \n",
        "    if not hold:\n",
        "        plt.grid()        \n",
        "        plt.show()\n",
        "    return extent"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ow0gOvoi6_fK",
        "colab_type": "text"
      },
      "source": [
        "Let's create some data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lUZKbmp03a6L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = np.array([[-1,0],[0,1],[4,4],[2,3]])\n",
        "y = np.array([-1,-1,+1,+1])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "z7-8Mive7C3w",
        "colab_type": "text"
      },
      "source": [
        "Let's make a simple neural network with 2 neurons in the hidden layer. Each neuron takes 2 inputs (the two features) and the output of the two neurons is fed to a single output neuron. \n",
        "\n",
        "We use MSE as the loss function with a gradient descent based optimizer. We run the optimizer for 1000 epochs with a batch size of 2 (i.e., two examples are picked, their error computed and weights updated at once)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zmtxywBp3sZO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "b61d4331-e488-429d-a84a-1855eb625b12"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "\n",
        "model = Sequential()\n",
        "model.add(Dense(2,input_dim=2,init='uniform',activation='relu'))\n",
        "model.add(Dense(1,init='uniform',activation='linear'))\n",
        "model.compile(loss = 'mse',optimizer='adam',metrics=['accuracy'])\n",
        "model.fit(X,y,nb_epoch = 1000, batch_size = 4)\n"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:5: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(2, input_dim=2, activation=\"relu\", kernel_initializer=\"uniform\")`\n",
            "  \"\"\"\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:6: UserWarning: Update your `Dense` call to the Keras 2 API: `Dense(1, activation=\"linear\", kernel_initializer=\"uniform\")`\n",
            "  \n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:8: UserWarning: The `nb_epoch` argument in `fit` has been renamed `epochs`.\n",
            "  \n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/1000\n",
            "4/4 [==============================] - 0s 92ms/step - loss: 1.0017 - acc: 0.0000e+00\n",
            "Epoch 2/1000\n",
            "4/4 [==============================] - 0s 730us/step - loss: 1.0014 - acc: 0.0000e+00\n",
            "Epoch 3/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 1.0012 - acc: 0.0000e+00\n",
            "Epoch 4/1000\n",
            "4/4 [==============================] - 0s 403us/step - loss: 1.0009 - acc: 0.0000e+00\n",
            "Epoch 5/1000\n",
            "4/4 [==============================] - 0s 377us/step - loss: 1.0007 - acc: 0.0000e+00\n",
            "Epoch 6/1000\n",
            "4/4 [==============================] - 0s 560us/step - loss: 1.0005 - acc: 0.0000e+00\n",
            "Epoch 7/1000\n",
            "4/4 [==============================] - 0s 381us/step - loss: 1.0003 - acc: 0.0000e+00\n",
            "Epoch 8/1000\n",
            "4/4 [==============================] - 0s 355us/step - loss: 1.0001 - acc: 0.0000e+00\n",
            "Epoch 9/1000\n",
            "4/4 [==============================] - 0s 400us/step - loss: 0.9999 - acc: 0.0000e+00\n",
            "Epoch 10/1000\n",
            "4/4 [==============================] - 0s 409us/step - loss: 0.9998 - acc: 0.0000e+00\n",
            "Epoch 11/1000\n",
            "4/4 [==============================] - 0s 426us/step - loss: 0.9998 - acc: 0.0000e+00\n",
            "Epoch 12/1000\n",
            "4/4 [==============================] - 0s 511us/step - loss: 0.9997 - acc: 0.0000e+00\n",
            "Epoch 13/1000\n",
            "4/4 [==============================] - 0s 602us/step - loss: 0.9997 - acc: 0.0000e+00\n",
            "Epoch 14/1000\n",
            "4/4 [==============================] - 0s 393us/step - loss: 0.9996 - acc: 0.0000e+00\n",
            "Epoch 15/1000\n",
            "4/4 [==============================] - 0s 384us/step - loss: 0.9996 - acc: 0.0000e+00\n",
            "Epoch 16/1000\n",
            "4/4 [==============================] - 0s 326us/step - loss: 0.9996 - acc: 0.0000e+00\n",
            "Epoch 17/1000\n",
            "4/4 [==============================] - 0s 394us/step - loss: 0.9996 - acc: 0.0000e+00\n",
            "Epoch 18/1000\n",
            "4/4 [==============================] - 0s 434us/step - loss: 0.9995 - acc: 0.0000e+00\n",
            "Epoch 19/1000\n",
            "4/4 [==============================] - 0s 442us/step - loss: 0.9995 - acc: 0.0000e+00\n",
            "Epoch 20/1000\n",
            "4/4 [==============================] - 0s 444us/step - loss: 0.9995 - acc: 0.0000e+00\n",
            "Epoch 21/1000\n",
            "4/4 [==============================] - 0s 456us/step - loss: 0.9994 - acc: 0.0000e+00\n",
            "Epoch 22/1000\n",
            "4/4 [==============================] - 0s 467us/step - loss: 0.9994 - acc: 0.0000e+00\n",
            "Epoch 23/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.9993 - acc: 0.0000e+00\n",
            "Epoch 24/1000\n",
            "4/4 [==============================] - 0s 459us/step - loss: 0.9993 - acc: 0.0000e+00\n",
            "Epoch 25/1000\n",
            "4/4 [==============================] - 0s 494us/step - loss: 0.9992 - acc: 0.0000e+00\n",
            "Epoch 26/1000\n",
            "4/4 [==============================] - 0s 462us/step - loss: 0.9992 - acc: 0.0000e+00\n",
            "Epoch 27/1000\n",
            "4/4 [==============================] - 0s 516us/step - loss: 0.9991 - acc: 0.0000e+00\n",
            "Epoch 28/1000\n",
            "4/4 [==============================] - 0s 525us/step - loss: 0.9991 - acc: 0.0000e+00\n",
            "Epoch 29/1000\n",
            "4/4 [==============================] - 0s 464us/step - loss: 0.9990 - acc: 0.0000e+00\n",
            "Epoch 30/1000\n",
            "4/4 [==============================] - 0s 508us/step - loss: 0.9989 - acc: 0.0000e+00\n",
            "Epoch 31/1000\n",
            "4/4 [==============================] - 0s 447us/step - loss: 0.9988 - acc: 0.0000e+00\n",
            "Epoch 32/1000\n",
            "4/4 [==============================] - 0s 503us/step - loss: 0.9987 - acc: 0.0000e+00\n",
            "Epoch 33/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.9987 - acc: 0.0000e+00\n",
            "Epoch 34/1000\n",
            "4/4 [==============================] - 0s 466us/step - loss: 0.9986 - acc: 0.0000e+00\n",
            "Epoch 35/1000\n",
            "4/4 [==============================] - 0s 477us/step - loss: 0.9985 - acc: 0.0000e+00\n",
            "Epoch 36/1000\n",
            "4/4 [==============================] - 0s 502us/step - loss: 0.9984 - acc: 0.0000e+00\n",
            "Epoch 37/1000\n",
            "4/4 [==============================] - 0s 438us/step - loss: 0.9983 - acc: 0.0000e+00\n",
            "Epoch 38/1000\n",
            "4/4 [==============================] - 0s 426us/step - loss: 0.9981 - acc: 0.0000e+00\n",
            "Epoch 39/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.9980 - acc: 0.0000e+00\n",
            "Epoch 40/1000\n",
            "4/4 [==============================] - 0s 555us/step - loss: 0.9979 - acc: 0.0000e+00\n",
            "Epoch 41/1000\n",
            "4/4 [==============================] - 0s 440us/step - loss: 0.9978 - acc: 0.0000e+00\n",
            "Epoch 42/1000\n",
            "4/4 [==============================] - 0s 559us/step - loss: 0.9977 - acc: 0.0000e+00\n",
            "Epoch 43/1000\n",
            "4/4 [==============================] - 0s 466us/step - loss: 0.9975 - acc: 0.0000e+00\n",
            "Epoch 44/1000\n",
            "4/4 [==============================] - 0s 503us/step - loss: 0.9974 - acc: 0.0000e+00\n",
            "Epoch 45/1000\n",
            "4/4 [==============================] - 0s 462us/step - loss: 0.9972 - acc: 0.0000e+00\n",
            "Epoch 46/1000\n",
            "4/4 [==============================] - 0s 458us/step - loss: 0.9971 - acc: 0.0000e+00\n",
            "Epoch 47/1000\n",
            "4/4 [==============================] - 0s 448us/step - loss: 0.9969 - acc: 0.0000e+00\n",
            "Epoch 48/1000\n",
            "4/4 [==============================] - 0s 438us/step - loss: 0.9967 - acc: 0.0000e+00\n",
            "Epoch 49/1000\n",
            "4/4 [==============================] - 0s 444us/step - loss: 0.9966 - acc: 0.0000e+00\n",
            "Epoch 50/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.9964 - acc: 0.0000e+00\n",
            "Epoch 51/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.9962 - acc: 0.0000e+00\n",
            "Epoch 52/1000\n",
            "4/4 [==============================] - 0s 719us/step - loss: 0.9960 - acc: 0.0000e+00\n",
            "Epoch 53/1000\n",
            "4/4 [==============================] - 0s 812us/step - loss: 0.9958 - acc: 0.0000e+00\n",
            "Epoch 54/1000\n",
            "4/4 [==============================] - 0s 552us/step - loss: 0.9956 - acc: 0.0000e+00\n",
            "Epoch 55/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.9954 - acc: 0.0000e+00\n",
            "Epoch 56/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.9952 - acc: 0.0000e+00\n",
            "Epoch 57/1000\n",
            "4/4 [==============================] - 0s 561us/step - loss: 0.9949 - acc: 0.0000e+00\n",
            "Epoch 58/1000\n",
            "4/4 [==============================] - 0s 776us/step - loss: 0.9947 - acc: 0.0000e+00\n",
            "Epoch 59/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.9945 - acc: 0.0000e+00\n",
            "Epoch 60/1000\n",
            "4/4 [==============================] - 0s 542us/step - loss: 0.9943 - acc: 0.0000e+00\n",
            "Epoch 61/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.9940 - acc: 0.0000e+00\n",
            "Epoch 62/1000\n",
            "4/4 [==============================] - 0s 800us/step - loss: 0.9937 - acc: 0.0000e+00\n",
            "Epoch 63/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.9935 - acc: 0.0000e+00\n",
            "Epoch 64/1000\n",
            "4/4 [==============================] - 0s 698us/step - loss: 0.9932 - acc: 0.0000e+00\n",
            "Epoch 65/1000\n",
            "4/4 [==============================] - 0s 570us/step - loss: 0.9929 - acc: 0.0000e+00\n",
            "Epoch 66/1000\n",
            "4/4 [==============================] - 0s 557us/step - loss: 0.9926 - acc: 0.0000e+00\n",
            "Epoch 67/1000\n",
            "4/4 [==============================] - 0s 546us/step - loss: 0.9923 - acc: 0.0000e+00\n",
            "Epoch 68/1000\n",
            "4/4 [==============================] - 0s 772us/step - loss: 0.9920 - acc: 0.0000e+00\n",
            "Epoch 69/1000\n",
            "4/4 [==============================] - 0s 738us/step - loss: 0.9917 - acc: 0.0000e+00\n",
            "Epoch 70/1000\n",
            "4/4 [==============================] - 0s 598us/step - loss: 0.9914 - acc: 0.0000e+00\n",
            "Epoch 71/1000\n",
            "4/4 [==============================] - 0s 805us/step - loss: 0.9910 - acc: 0.0000e+00\n",
            "Epoch 72/1000\n",
            "4/4 [==============================] - 0s 664us/step - loss: 0.9907 - acc: 0.0000e+00\n",
            "Epoch 73/1000\n",
            "4/4 [==============================] - 0s 494us/step - loss: 0.9903 - acc: 0.0000e+00\n",
            "Epoch 74/1000\n",
            "4/4 [==============================] - 0s 545us/step - loss: 0.9900 - acc: 0.0000e+00\n",
            "Epoch 75/1000\n",
            "4/4 [==============================] - 0s 478us/step - loss: 0.9896 - acc: 0.0000e+00\n",
            "Epoch 76/1000\n",
            "4/4 [==============================] - 0s 429us/step - loss: 0.9892 - acc: 0.0000e+00\n",
            "Epoch 77/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.9888 - acc: 0.0000e+00\n",
            "Epoch 78/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.9885 - acc: 0.0000e+00\n",
            "Epoch 79/1000\n",
            "4/4 [==============================] - 0s 560us/step - loss: 0.9881 - acc: 0.0000e+00\n",
            "Epoch 80/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.9876 - acc: 0.0000e+00\n",
            "Epoch 81/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.9872 - acc: 0.0000e+00\n",
            "Epoch 82/1000\n",
            "4/4 [==============================] - 0s 627us/step - loss: 0.9868 - acc: 0.0000e+00\n",
            "Epoch 83/1000\n",
            "4/4 [==============================] - 0s 477us/step - loss: 0.9864 - acc: 0.0000e+00\n",
            "Epoch 84/1000\n",
            "4/4 [==============================] - 0s 487us/step - loss: 0.9859 - acc: 0.0000e+00\n",
            "Epoch 85/1000\n",
            "4/4 [==============================] - 0s 434us/step - loss: 0.9855 - acc: 0.0000e+00\n",
            "Epoch 86/1000\n",
            "4/4 [==============================] - 0s 444us/step - loss: 0.9850 - acc: 0.0000e+00\n",
            "Epoch 87/1000\n",
            "4/4 [==============================] - 0s 444us/step - loss: 0.9846 - acc: 0.0000e+00\n",
            "Epoch 88/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.9841 - acc: 0.0000e+00\n",
            "Epoch 89/1000\n",
            "4/4 [==============================] - 0s 429us/step - loss: 0.9836 - acc: 0.0000e+00\n",
            "Epoch 90/1000\n",
            "4/4 [==============================] - 0s 419us/step - loss: 0.9831 - acc: 0.0000e+00\n",
            "Epoch 91/1000\n",
            "4/4 [==============================] - 0s 458us/step - loss: 0.9826 - acc: 0.0000e+00\n",
            "Epoch 92/1000\n",
            "4/4 [==============================] - 0s 609us/step - loss: 0.9821 - acc: 0.0000e+00\n",
            "Epoch 93/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.9816 - acc: 0.0000e+00\n",
            "Epoch 94/1000\n",
            "4/4 [==============================] - 0s 470us/step - loss: 0.9810 - acc: 0.0000e+00\n",
            "Epoch 95/1000\n",
            "4/4 [==============================] - 0s 679us/step - loss: 0.9805 - acc: 0.0000e+00\n",
            "Epoch 96/1000\n",
            "4/4 [==============================] - 0s 817us/step - loss: 0.9799 - acc: 0.0000e+00\n",
            "Epoch 97/1000\n",
            "4/4 [==============================] - 0s 689us/step - loss: 0.9794 - acc: 0.0000e+00\n",
            "Epoch 98/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.9788 - acc: 0.0000e+00\n",
            "Epoch 99/1000\n",
            "4/4 [==============================] - 0s 655us/step - loss: 0.9783 - acc: 0.0000e+00\n",
            "Epoch 100/1000\n",
            "4/4 [==============================] - 0s 445us/step - loss: 0.9777 - acc: 0.0000e+00\n",
            "Epoch 101/1000\n",
            "4/4 [==============================] - 0s 489us/step - loss: 0.9771 - acc: 0.0000e+00\n",
            "Epoch 102/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.9765 - acc: 0.0000e+00\n",
            "Epoch 103/1000\n",
            "4/4 [==============================] - 0s 484us/step - loss: 0.9759 - acc: 0.0000e+00\n",
            "Epoch 104/1000\n",
            "4/4 [==============================] - 0s 688us/step - loss: 0.9753 - acc: 0.0000e+00\n",
            "Epoch 105/1000\n",
            "4/4 [==============================] - 0s 665us/step - loss: 0.9747 - acc: 0.0000e+00\n",
            "Epoch 106/1000\n",
            "4/4 [==============================] - 0s 463us/step - loss: 0.9740 - acc: 0.0000e+00\n",
            "Epoch 107/1000\n",
            "4/4 [==============================] - 0s 703us/step - loss: 0.9734 - acc: 0.0000e+00\n",
            "Epoch 108/1000\n",
            "4/4 [==============================] - 0s 662us/step - loss: 0.9727 - acc: 0.0000e+00\n",
            "Epoch 109/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.9721 - acc: 0.0000e+00\n",
            "Epoch 110/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.9714 - acc: 0.0000e+00\n",
            "Epoch 111/1000\n",
            "4/4 [==============================] - 0s 809us/step - loss: 0.9707 - acc: 0.0000e+00\n",
            "Epoch 112/1000\n",
            "4/4 [==============================] - 0s 740us/step - loss: 0.9700 - acc: 0.0000e+00\n",
            "Epoch 113/1000\n",
            "4/4 [==============================] - 0s 782us/step - loss: 0.9693 - acc: 0.0000e+00\n",
            "Epoch 114/1000\n",
            "4/4 [==============================] - 0s 731us/step - loss: 0.9686 - acc: 0.0000e+00\n",
            "Epoch 115/1000\n",
            "4/4 [==============================] - 0s 650us/step - loss: 0.9679 - acc: 0.0000e+00\n",
            "Epoch 116/1000\n",
            "4/4 [==============================] - 0s 640us/step - loss: 0.9672 - acc: 0.0000e+00\n",
            "Epoch 117/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.9664 - acc: 0.0000e+00\n",
            "Epoch 118/1000\n",
            "4/4 [==============================] - 0s 729us/step - loss: 0.9657 - acc: 0.0000e+00\n",
            "Epoch 119/1000\n",
            "4/4 [==============================] - 0s 667us/step - loss: 0.9650 - acc: 0.0000e+00\n",
            "Epoch 120/1000\n",
            "4/4 [==============================] - 0s 717us/step - loss: 0.9642 - acc: 0.0000e+00\n",
            "Epoch 121/1000\n",
            "4/4 [==============================] - 0s 678us/step - loss: 0.9635 - acc: 0.0000e+00\n",
            "Epoch 122/1000\n",
            "4/4 [==============================] - 0s 743us/step - loss: 0.9627 - acc: 0.0000e+00\n",
            "Epoch 123/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.9619 - acc: 0.0000e+00\n",
            "Epoch 124/1000\n",
            "4/4 [==============================] - 0s 508us/step - loss: 0.9611 - acc: 0.0000e+00\n",
            "Epoch 125/1000\n",
            "4/4 [==============================] - 0s 463us/step - loss: 0.9603 - acc: 0.0000e+00\n",
            "Epoch 126/1000\n",
            "4/4 [==============================] - 0s 628us/step - loss: 0.9595 - acc: 0.0000e+00\n",
            "Epoch 127/1000\n",
            "4/4 [==============================] - 0s 805us/step - loss: 0.9587 - acc: 0.0000e+00\n",
            "Epoch 128/1000\n",
            "4/4 [==============================] - 0s 555us/step - loss: 0.9578 - acc: 0.0000e+00\n",
            "Epoch 129/1000\n",
            "4/4 [==============================] - 0s 554us/step - loss: 0.9570 - acc: 0.0000e+00\n",
            "Epoch 130/1000\n",
            "4/4 [==============================] - 0s 977us/step - loss: 0.9561 - acc: 0.0000e+00\n",
            "Epoch 131/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.9553 - acc: 0.0000e+00\n",
            "Epoch 132/1000\n",
            "4/4 [==============================] - 0s 549us/step - loss: 0.9544 - acc: 0.0000e+00\n",
            "Epoch 133/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.9535 - acc: 0.0000e+00\n",
            "Epoch 134/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.9526 - acc: 0.0000e+00\n",
            "Epoch 135/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.9518 - acc: 0.0000e+00\n",
            "Epoch 136/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.9509 - acc: 0.0000e+00\n",
            "Epoch 137/1000\n",
            "4/4 [==============================] - 0s 422us/step - loss: 0.9500 - acc: 0.0000e+00\n",
            "Epoch 138/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.9491 - acc: 0.0000e+00\n",
            "Epoch 139/1000\n",
            "4/4 [==============================] - 0s 534us/step - loss: 0.9482 - acc: 0.0000e+00\n",
            "Epoch 140/1000\n",
            "4/4 [==============================] - 0s 455us/step - loss: 0.9472 - acc: 0.0000e+00\n",
            "Epoch 141/1000\n",
            "4/4 [==============================] - 0s 500us/step - loss: 0.9463 - acc: 0.0000e+00\n",
            "Epoch 142/1000\n",
            "4/4 [==============================] - 0s 459us/step - loss: 0.9454 - acc: 0.0000e+00\n",
            "Epoch 143/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.9444 - acc: 0.0000e+00\n",
            "Epoch 144/1000\n",
            "4/4 [==============================] - 0s 485us/step - loss: 0.9434 - acc: 0.0000e+00\n",
            "Epoch 145/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.9425 - acc: 0.0000e+00\n",
            "Epoch 146/1000\n",
            "4/4 [==============================] - 0s 493us/step - loss: 0.9415 - acc: 0.0000e+00\n",
            "Epoch 147/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.9405 - acc: 0.0000e+00\n",
            "Epoch 148/1000\n",
            "4/4 [==============================] - 0s 559us/step - loss: 0.9395 - acc: 0.0000e+00\n",
            "Epoch 149/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.9385 - acc: 0.0000e+00\n",
            "Epoch 150/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.9375 - acc: 0.0000e+00\n",
            "Epoch 151/1000\n",
            "4/4 [==============================] - 0s 477us/step - loss: 0.9365 - acc: 0.0000e+00\n",
            "Epoch 152/1000\n",
            "4/4 [==============================] - 0s 472us/step - loss: 0.9354 - acc: 0.0000e+00\n",
            "Epoch 153/1000\n",
            "4/4 [==============================] - 0s 449us/step - loss: 0.9344 - acc: 0.0000e+00\n",
            "Epoch 154/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.9334 - acc: 0.0000e+00\n",
            "Epoch 155/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.9324 - acc: 0.0000e+00\n",
            "Epoch 156/1000\n",
            "4/4 [==============================] - 0s 507us/step - loss: 0.9313 - acc: 0.0000e+00\n",
            "Epoch 157/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.9302 - acc: 0.0000e+00\n",
            "Epoch 158/1000\n",
            "4/4 [==============================] - 0s 526us/step - loss: 0.9292 - acc: 0.0000e+00\n",
            "Epoch 159/1000\n",
            "4/4 [==============================] - 0s 423us/step - loss: 0.9281 - acc: 0.0000e+00\n",
            "Epoch 160/1000\n",
            "4/4 [==============================] - 0s 503us/step - loss: 0.9270 - acc: 0.0000e+00\n",
            "Epoch 161/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.9259 - acc: 0.0000e+00\n",
            "Epoch 162/1000\n",
            "4/4 [==============================] - 0s 2ms/step - loss: 0.9248 - acc: 0.0000e+00\n",
            "Epoch 163/1000\n",
            "4/4 [==============================] - 0s 796us/step - loss: 0.9237 - acc: 0.0000e+00\n",
            "Epoch 164/1000\n",
            "4/4 [==============================] - 0s 680us/step - loss: 0.9226 - acc: 0.0000e+00\n",
            "Epoch 165/1000\n",
            "4/4 [==============================] - 0s 646us/step - loss: 0.9215 - acc: 0.0000e+00\n",
            "Epoch 166/1000\n",
            "4/4 [==============================] - 0s 646us/step - loss: 0.9203 - acc: 0.0000e+00\n",
            "Epoch 167/1000\n",
            "4/4 [==============================] - 0s 605us/step - loss: 0.9192 - acc: 0.0000e+00\n",
            "Epoch 168/1000\n",
            "4/4 [==============================] - 0s 2ms/step - loss: 0.9181 - acc: 0.0000e+00\n",
            "Epoch 169/1000\n",
            "4/4 [==============================] - 0s 805us/step - loss: 0.9169 - acc: 0.0000e+00\n",
            "Epoch 170/1000\n",
            "4/4 [==============================] - 0s 956us/step - loss: 0.9157 - acc: 0.0000e+00\n",
            "Epoch 171/1000\n",
            "4/4 [==============================] - 0s 889us/step - loss: 0.9146 - acc: 0.0000e+00\n",
            "Epoch 172/1000\n",
            "4/4 [==============================] - 0s 937us/step - loss: 0.9134 - acc: 0.0000e+00\n",
            "Epoch 173/1000\n",
            "4/4 [==============================] - 0s 565us/step - loss: 0.9122 - acc: 0.0000e+00\n",
            "Epoch 174/1000\n",
            "4/4 [==============================] - 0s 644us/step - loss: 0.9110 - acc: 0.0000e+00\n",
            "Epoch 175/1000\n",
            "4/4 [==============================] - 0s 488us/step - loss: 0.9098 - acc: 0.0000e+00\n",
            "Epoch 176/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.9086 - acc: 0.0000e+00\n",
            "Epoch 177/1000\n",
            "4/4 [==============================] - 0s 749us/step - loss: 0.9074 - acc: 0.0000e+00\n",
            "Epoch 178/1000\n",
            "4/4 [==============================] - 0s 745us/step - loss: 0.9062 - acc: 0.0000e+00\n",
            "Epoch 179/1000\n",
            "4/4 [==============================] - 0s 717us/step - loss: 0.9049 - acc: 0.0000e+00\n",
            "Epoch 180/1000\n",
            "4/4 [==============================] - 0s 740us/step - loss: 0.9037 - acc: 0.0000e+00\n",
            "Epoch 181/1000\n",
            "4/4 [==============================] - 0s 813us/step - loss: 0.9025 - acc: 0.0000e+00\n",
            "Epoch 182/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.9013 - acc: 0.0000e+00\n",
            "Epoch 183/1000\n",
            "4/4 [==============================] - 0s 759us/step - loss: 0.9000 - acc: 0.0000e+00\n",
            "Epoch 184/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.8987 - acc: 0.0000e+00\n",
            "Epoch 185/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.8975 - acc: 0.0000e+00\n",
            "Epoch 186/1000\n",
            "4/4 [==============================] - 0s 857us/step - loss: 0.8962 - acc: 0.0000e+00\n",
            "Epoch 187/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.8949 - acc: 0.0000e+00\n",
            "Epoch 188/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.8937 - acc: 0.0000e+00\n",
            "Epoch 189/1000\n",
            "4/4 [==============================] - 0s 748us/step - loss: 0.8924 - acc: 0.0000e+00\n",
            "Epoch 190/1000\n",
            "4/4 [==============================] - 0s 802us/step - loss: 0.8911 - acc: 0.0000e+00\n",
            "Epoch 191/1000\n",
            "4/4 [==============================] - 0s 473us/step - loss: 0.8898 - acc: 0.0000e+00\n",
            "Epoch 192/1000\n",
            "4/4 [==============================] - 0s 814us/step - loss: 0.8884 - acc: 0.0000e+00\n",
            "Epoch 193/1000\n",
            "4/4 [==============================] - 0s 623us/step - loss: 0.8871 - acc: 0.0000e+00\n",
            "Epoch 194/1000\n",
            "4/4 [==============================] - 0s 725us/step - loss: 0.8858 - acc: 0.0000e+00\n",
            "Epoch 195/1000\n",
            "4/4 [==============================] - 0s 677us/step - loss: 0.8844 - acc: 0.0000e+00\n",
            "Epoch 196/1000\n",
            "4/4 [==============================] - 0s 825us/step - loss: 0.8831 - acc: 0.0000e+00\n",
            "Epoch 197/1000\n",
            "4/4 [==============================] - 0s 954us/step - loss: 0.8818 - acc: 0.0000e+00\n",
            "Epoch 198/1000\n",
            "4/4 [==============================] - 0s 700us/step - loss: 0.8804 - acc: 0.0000e+00\n",
            "Epoch 199/1000\n",
            "4/4 [==============================] - 0s 600us/step - loss: 0.8791 - acc: 0.0000e+00\n",
            "Epoch 200/1000\n",
            "4/4 [==============================] - 0s 763us/step - loss: 0.8777 - acc: 0.0000e+00\n",
            "Epoch 201/1000\n",
            "4/4 [==============================] - 0s 875us/step - loss: 0.8763 - acc: 0.0000e+00\n",
            "Epoch 202/1000\n",
            "4/4 [==============================] - 0s 702us/step - loss: 0.8749 - acc: 0.0000e+00\n",
            "Epoch 203/1000\n",
            "4/4 [==============================] - 0s 862us/step - loss: 0.8735 - acc: 0.0000e+00\n",
            "Epoch 204/1000\n",
            "4/4 [==============================] - 0s 580us/step - loss: 0.8721 - acc: 0.0000e+00\n",
            "Epoch 205/1000\n",
            "4/4 [==============================] - 0s 715us/step - loss: 0.8708 - acc: 0.0000e+00\n",
            "Epoch 206/1000\n",
            "4/4 [==============================] - 0s 876us/step - loss: 0.8693 - acc: 0.0000e+00\n",
            "Epoch 207/1000\n",
            "4/4 [==============================] - 0s 683us/step - loss: 0.8679 - acc: 0.0000e+00\n",
            "Epoch 208/1000\n",
            "4/4 [==============================] - 0s 732us/step - loss: 0.8665 - acc: 0.0000e+00\n",
            "Epoch 209/1000\n",
            "4/4 [==============================] - 0s 739us/step - loss: 0.8651 - acc: 0.0000e+00\n",
            "Epoch 210/1000\n",
            "4/4 [==============================] - 0s 908us/step - loss: 0.8637 - acc: 0.0000e+00\n",
            "Epoch 211/1000\n",
            "4/4 [==============================] - 0s 887us/step - loss: 0.8623 - acc: 0.0000e+00\n",
            "Epoch 212/1000\n",
            "4/4 [==============================] - 0s 557us/step - loss: 0.8609 - acc: 0.0000e+00\n",
            "Epoch 213/1000\n",
            "4/4 [==============================] - 0s 662us/step - loss: 0.8595 - acc: 0.0000e+00\n",
            "Epoch 214/1000\n",
            "4/4 [==============================] - 0s 575us/step - loss: 0.8580 - acc: 0.0000e+00\n",
            "Epoch 215/1000\n",
            "4/4 [==============================] - 0s 646us/step - loss: 0.8566 - acc: 0.0000e+00\n",
            "Epoch 216/1000\n",
            "4/4 [==============================] - 0s 594us/step - loss: 0.8551 - acc: 0.0000e+00\n",
            "Epoch 217/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.8536 - acc: 0.0000e+00\n",
            "Epoch 218/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.8521 - acc: 0.0000e+00\n",
            "Epoch 219/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.8506 - acc: 0.0000e+00\n",
            "Epoch 220/1000\n",
            "4/4 [==============================] - 0s 453us/step - loss: 0.8492 - acc: 0.0000e+00\n",
            "Epoch 221/1000\n",
            "4/4 [==============================] - 0s 476us/step - loss: 0.8477 - acc: 0.0000e+00\n",
            "Epoch 222/1000\n",
            "4/4 [==============================] - 0s 484us/step - loss: 0.8461 - acc: 0.0000e+00\n",
            "Epoch 223/1000\n",
            "4/4 [==============================] - 0s 496us/step - loss: 0.8446 - acc: 0.0000e+00\n",
            "Epoch 224/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.8432 - acc: 0.0000e+00\n",
            "Epoch 225/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.8416 - acc: 0.0000e+00\n",
            "Epoch 226/1000\n",
            "4/4 [==============================] - 0s 604us/step - loss: 0.8401 - acc: 0.0000e+00\n",
            "Epoch 227/1000\n",
            "4/4 [==============================] - 0s 475us/step - loss: 0.8386 - acc: 0.0000e+00\n",
            "Epoch 228/1000\n",
            "4/4 [==============================] - 0s 551us/step - loss: 0.8370 - acc: 0.0000e+00\n",
            "Epoch 229/1000\n",
            "4/4 [==============================] - 0s 670us/step - loss: 0.8356 - acc: 0.0000e+00\n",
            "Epoch 230/1000\n",
            "4/4 [==============================] - 0s 566us/step - loss: 0.8341 - acc: 0.0000e+00\n",
            "Epoch 231/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.8326 - acc: 0.0000e+00\n",
            "Epoch 232/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.8310 - acc: 0.0000e+00\n",
            "Epoch 233/1000\n",
            "4/4 [==============================] - 0s 438us/step - loss: 0.8295 - acc: 0.0000e+00\n",
            "Epoch 234/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.8279 - acc: 0.0000e+00\n",
            "Epoch 235/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.8263 - acc: 0.0000e+00\n",
            "Epoch 236/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.8247 - acc: 0.0000e+00\n",
            "Epoch 237/1000\n",
            "4/4 [==============================] - 0s 474us/step - loss: 0.8232 - acc: 0.0000e+00\n",
            "Epoch 238/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.8217 - acc: 0.0000e+00\n",
            "Epoch 239/1000\n",
            "4/4 [==============================] - 0s 521us/step - loss: 0.8199 - acc: 0.0000e+00\n",
            "Epoch 240/1000\n",
            "4/4 [==============================] - 0s 494us/step - loss: 0.8184 - acc: 0.0000e+00\n",
            "Epoch 241/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.8169 - acc: 0.0000e+00\n",
            "Epoch 242/1000\n",
            "4/4 [==============================] - 0s 424us/step - loss: 0.8153 - acc: 0.0000e+00\n",
            "Epoch 243/1000\n",
            "4/4 [==============================] - 0s 458us/step - loss: 0.8138 - acc: 0.0000e+00\n",
            "Epoch 244/1000\n",
            "4/4 [==============================] - 0s 650us/step - loss: 0.8122 - acc: 0.0000e+00\n",
            "Epoch 245/1000\n",
            "4/4 [==============================] - 0s 813us/step - loss: 0.8106 - acc: 0.0000e+00\n",
            "Epoch 246/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.8089 - acc: 0.0000e+00\n",
            "Epoch 247/1000\n",
            "4/4 [==============================] - 0s 484us/step - loss: 0.8073 - acc: 0.0000e+00\n",
            "Epoch 248/1000\n",
            "4/4 [==============================] - 0s 451us/step - loss: 0.8056 - acc: 0.0000e+00\n",
            "Epoch 249/1000\n",
            "4/4 [==============================] - 0s 812us/step - loss: 0.8039 - acc: 0.0000e+00\n",
            "Epoch 250/1000\n",
            "4/4 [==============================] - 0s 726us/step - loss: 0.8024 - acc: 0.0000e+00\n",
            "Epoch 251/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.8008 - acc: 0.0000e+00\n",
            "Epoch 252/1000\n",
            "4/4 [==============================] - 0s 718us/step - loss: 0.7990 - acc: 0.0000e+00\n",
            "Epoch 253/1000\n",
            "4/4 [==============================] - 0s 525us/step - loss: 0.7974 - acc: 0.0000e+00\n",
            "Epoch 254/1000\n",
            "4/4 [==============================] - 0s 483us/step - loss: 0.7958 - acc: 0.0000e+00\n",
            "Epoch 255/1000\n",
            "4/4 [==============================] - 0s 445us/step - loss: 0.7942 - acc: 0.0000e+00\n",
            "Epoch 256/1000\n",
            "4/4 [==============================] - 0s 456us/step - loss: 0.7925 - acc: 0.0000e+00\n",
            "Epoch 257/1000\n",
            "4/4 [==============================] - 0s 492us/step - loss: 0.7909 - acc: 0.0000e+00\n",
            "Epoch 258/1000\n",
            "4/4 [==============================] - 0s 555us/step - loss: 0.7892 - acc: 0.0000e+00\n",
            "Epoch 259/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.7875 - acc: 0.0000e+00\n",
            "Epoch 260/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.7858 - acc: 0.0000e+00\n",
            "Epoch 261/1000\n",
            "4/4 [==============================] - 0s 411us/step - loss: 0.7840 - acc: 0.0000e+00\n",
            "Epoch 262/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.7825 - acc: 0.0000e+00\n",
            "Epoch 263/1000\n",
            "4/4 [==============================] - 0s 464us/step - loss: 0.7808 - acc: 0.0000e+00\n",
            "Epoch 264/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.7790 - acc: 0.0000e+00\n",
            "Epoch 265/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.7774 - acc: 0.0000e+00\n",
            "Epoch 266/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.7757 - acc: 0.0000e+00\n",
            "Epoch 267/1000\n",
            "4/4 [==============================] - 0s 580us/step - loss: 0.7741 - acc: 0.0000e+00\n",
            "Epoch 268/1000\n",
            "4/4 [==============================] - 0s 456us/step - loss: 0.7724 - acc: 0.0000e+00\n",
            "Epoch 269/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.7707 - acc: 0.0000e+00\n",
            "Epoch 270/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.7690 - acc: 0.0000e+00\n",
            "Epoch 271/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.7672 - acc: 0.0000e+00\n",
            "Epoch 272/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.7655 - acc: 0.0000e+00\n",
            "Epoch 273/1000\n",
            "4/4 [==============================] - 0s 494us/step - loss: 0.7637 - acc: 0.0000e+00\n",
            "Epoch 274/1000\n",
            "4/4 [==============================] - 0s 567us/step - loss: 0.7619 - acc: 0.0000e+00\n",
            "Epoch 275/1000\n",
            "4/4 [==============================] - 0s 447us/step - loss: 0.7602 - acc: 0.0000e+00\n",
            "Epoch 276/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.7585 - acc: 0.0000e+00\n",
            "Epoch 277/1000\n",
            "4/4 [==============================] - 0s 667us/step - loss: 0.7568 - acc: 0.0000e+00\n",
            "Epoch 278/1000\n",
            "4/4 [==============================] - 0s 681us/step - loss: 0.7550 - acc: 0.0000e+00\n",
            "Epoch 279/1000\n",
            "4/4 [==============================] - 0s 338us/step - loss: 0.7533 - acc: 0.0000e+00\n",
            "Epoch 280/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.7515 - acc: 0.0000e+00\n",
            "Epoch 281/1000\n",
            "4/4 [==============================] - 0s 706us/step - loss: 0.7498 - acc: 0.0000e+00\n",
            "Epoch 282/1000\n",
            "4/4 [==============================] - 0s 751us/step - loss: 0.7480 - acc: 0.0000e+00\n",
            "Epoch 283/1000\n",
            "4/4 [==============================] - 0s 406us/step - loss: 0.7463 - acc: 0.0000e+00\n",
            "Epoch 284/1000\n",
            "4/4 [==============================] - 0s 856us/step - loss: 0.7446 - acc: 0.0000e+00\n",
            "Epoch 285/1000\n",
            "4/4 [==============================] - 0s 609us/step - loss: 0.7428 - acc: 0.0000e+00\n",
            "Epoch 286/1000\n",
            "4/4 [==============================] - 0s 670us/step - loss: 0.7411 - acc: 0.0000e+00\n",
            "Epoch 287/1000\n",
            "4/4 [==============================] - 0s 581us/step - loss: 0.7393 - acc: 0.0000e+00\n",
            "Epoch 288/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.7375 - acc: 0.0000e+00\n",
            "Epoch 289/1000\n",
            "4/4 [==============================] - 0s 720us/step - loss: 0.7357 - acc: 0.0000e+00\n",
            "Epoch 290/1000\n",
            "4/4 [==============================] - 0s 786us/step - loss: 0.7340 - acc: 0.0000e+00\n",
            "Epoch 291/1000\n",
            "4/4 [==============================] - 0s 545us/step - loss: 0.7322 - acc: 0.0000e+00\n",
            "Epoch 292/1000\n",
            "4/4 [==============================] - 0s 610us/step - loss: 0.7303 - acc: 0.0000e+00\n",
            "Epoch 293/1000\n",
            "4/4 [==============================] - 0s 489us/step - loss: 0.7286 - acc: 0.0000e+00\n",
            "Epoch 294/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.7268 - acc: 0.0000e+00\n",
            "Epoch 295/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.7250 - acc: 0.0000e+00\n",
            "Epoch 296/1000\n",
            "4/4 [==============================] - 0s 598us/step - loss: 0.7232 - acc: 0.0000e+00\n",
            "Epoch 297/1000\n",
            "4/4 [==============================] - 0s 586us/step - loss: 0.7215 - acc: 0.0000e+00\n",
            "Epoch 298/1000\n",
            "4/4 [==============================] - 0s 682us/step - loss: 0.7196 - acc: 0.0000e+00\n",
            "Epoch 299/1000\n",
            "4/4 [==============================] - 0s 861us/step - loss: 0.7179 - acc: 0.0000e+00\n",
            "Epoch 300/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.7161 - acc: 0.0000e+00\n",
            "Epoch 301/1000\n",
            "4/4 [==============================] - 0s 691us/step - loss: 0.7143 - acc: 0.0000e+00\n",
            "Epoch 302/1000\n",
            "4/4 [==============================] - 0s 570us/step - loss: 0.7125 - acc: 0.0000e+00\n",
            "Epoch 303/1000\n",
            "4/4 [==============================] - 0s 614us/step - loss: 0.7107 - acc: 0.0000e+00\n",
            "Epoch 304/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.7088 - acc: 0.0000e+00\n",
            "Epoch 305/1000\n",
            "4/4 [==============================] - 0s 885us/step - loss: 0.7070 - acc: 0.0000e+00\n",
            "Epoch 306/1000\n",
            "4/4 [==============================] - 0s 589us/step - loss: 0.7053 - acc: 0.0000e+00\n",
            "Epoch 307/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.7034 - acc: 0.0000e+00\n",
            "Epoch 308/1000\n",
            "4/4 [==============================] - 0s 591us/step - loss: 0.7015 - acc: 0.0000e+00\n",
            "Epoch 309/1000\n",
            "4/4 [==============================] - 0s 581us/step - loss: 0.6997 - acc: 0.0000e+00\n",
            "Epoch 310/1000\n",
            "4/4 [==============================] - 0s 576us/step - loss: 0.6979 - acc: 0.0000e+00\n",
            "Epoch 311/1000\n",
            "4/4 [==============================] - 0s 453us/step - loss: 0.6961 - acc: 0.0000e+00\n",
            "Epoch 312/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.6942 - acc: 0.0000e+00\n",
            "Epoch 313/1000\n",
            "4/4 [==============================] - 0s 562us/step - loss: 0.6925 - acc: 0.0000e+00\n",
            "Epoch 314/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.6906 - acc: 0.0000e+00\n",
            "Epoch 315/1000\n",
            "4/4 [==============================] - 0s 841us/step - loss: 0.6888 - acc: 0.0000e+00\n",
            "Epoch 316/1000\n",
            "4/4 [==============================] - 0s 611us/step - loss: 0.6870 - acc: 0.0000e+00\n",
            "Epoch 317/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.6852 - acc: 0.0000e+00\n",
            "Epoch 318/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.6834 - acc: 0.0000e+00\n",
            "Epoch 319/1000\n",
            "4/4 [==============================] - 0s 550us/step - loss: 0.6815 - acc: 0.0000e+00\n",
            "Epoch 320/1000\n",
            "4/4 [==============================] - 0s 631us/step - loss: 0.6796 - acc: 0.0000e+00\n",
            "Epoch 321/1000\n",
            "4/4 [==============================] - 0s 840us/step - loss: 0.6777 - acc: 0.0000e+00\n",
            "Epoch 322/1000\n",
            "4/4 [==============================] - 0s 647us/step - loss: 0.6760 - acc: 0.0000e+00\n",
            "Epoch 323/1000\n",
            "4/4 [==============================] - 0s 554us/step - loss: 0.6742 - acc: 0.0000e+00\n",
            "Epoch 324/1000\n",
            "4/4 [==============================] - 0s 573us/step - loss: 0.6722 - acc: 0.0000e+00\n",
            "Epoch 325/1000\n",
            "4/4 [==============================] - 0s 543us/step - loss: 0.6704 - acc: 0.0000e+00\n",
            "Epoch 326/1000\n",
            "4/4 [==============================] - 0s 483us/step - loss: 0.6685 - acc: 0.0000e+00\n",
            "Epoch 327/1000\n",
            "4/4 [==============================] - 0s 466us/step - loss: 0.6667 - acc: 0.0000e+00\n",
            "Epoch 328/1000\n",
            "4/4 [==============================] - 0s 511us/step - loss: 0.6648 - acc: 0.0000e+00\n",
            "Epoch 329/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.6630 - acc: 0.0000e+00\n",
            "Epoch 330/1000\n",
            "4/4 [==============================] - 0s 455us/step - loss: 0.6611 - acc: 0.0000e+00\n",
            "Epoch 331/1000\n",
            "4/4 [==============================] - 0s 462us/step - loss: 0.6593 - acc: 0.0000e+00\n",
            "Epoch 332/1000\n",
            "4/4 [==============================] - 0s 455us/step - loss: 0.6574 - acc: 0.0000e+00\n",
            "Epoch 333/1000\n",
            "4/4 [==============================] - 0s 441us/step - loss: 0.6555 - acc: 0.0000e+00\n",
            "Epoch 334/1000\n",
            "4/4 [==============================] - 0s 502us/step - loss: 0.6537 - acc: 0.0000e+00\n",
            "Epoch 335/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.6518 - acc: 0.0000e+00\n",
            "Epoch 336/1000\n",
            "4/4 [==============================] - 0s 443us/step - loss: 0.6500 - acc: 0.0000e+00\n",
            "Epoch 337/1000\n",
            "4/4 [==============================] - 0s 485us/step - loss: 0.6481 - acc: 0.0000e+00\n",
            "Epoch 338/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.6463 - acc: 0.0000e+00\n",
            "Epoch 339/1000\n",
            "4/4 [==============================] - 0s 527us/step - loss: 0.6445 - acc: 0.0000e+00\n",
            "Epoch 340/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.6427 - acc: 0.0000e+00\n",
            "Epoch 341/1000\n",
            "4/4 [==============================] - 0s 470us/step - loss: 0.6408 - acc: 0.0000e+00\n",
            "Epoch 342/1000\n",
            "4/4 [==============================] - 0s 461us/step - loss: 0.6389 - acc: 0.0000e+00\n",
            "Epoch 343/1000\n",
            "4/4 [==============================] - 0s 427us/step - loss: 0.6370 - acc: 0.0000e+00\n",
            "Epoch 344/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.6351 - acc: 0.0000e+00\n",
            "Epoch 345/1000\n",
            "4/4 [==============================] - 0s 465us/step - loss: 0.6332 - acc: 0.0000e+00\n",
            "Epoch 346/1000\n",
            "4/4 [==============================] - 0s 417us/step - loss: 0.6315 - acc: 0.0000e+00\n",
            "Epoch 347/1000\n",
            "4/4 [==============================] - 0s 570us/step - loss: 0.6297 - acc: 0.0000e+00\n",
            "Epoch 348/1000\n",
            "4/4 [==============================] - 0s 741us/step - loss: 0.6277 - acc: 0.0000e+00\n",
            "Epoch 349/1000\n",
            "4/4 [==============================] - 0s 691us/step - loss: 0.6258 - acc: 0.0000e+00\n",
            "Epoch 350/1000\n",
            "4/4 [==============================] - 0s 487us/step - loss: 0.6240 - acc: 0.0000e+00\n",
            "Epoch 351/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.6221 - acc: 0.0000e+00\n",
            "Epoch 352/1000\n",
            "4/4 [==============================] - 0s 458us/step - loss: 0.6203 - acc: 0.0000e+00\n",
            "Epoch 353/1000\n",
            "4/4 [==============================] - 0s 663us/step - loss: 0.6184 - acc: 0.0000e+00\n",
            "Epoch 354/1000\n",
            "4/4 [==============================] - 0s 692us/step - loss: 0.6165 - acc: 0.0000e+00\n",
            "Epoch 355/1000\n",
            "4/4 [==============================] - 0s 430us/step - loss: 0.6146 - acc: 0.0000e+00\n",
            "Epoch 356/1000\n",
            "4/4 [==============================] - 0s 475us/step - loss: 0.6127 - acc: 0.0000e+00\n",
            "Epoch 357/1000\n",
            "4/4 [==============================] - 0s 576us/step - loss: 0.6109 - acc: 0.0000e+00\n",
            "Epoch 358/1000\n",
            "4/4 [==============================] - 0s 681us/step - loss: 0.6090 - acc: 0.0000e+00\n",
            "Epoch 359/1000\n",
            "4/4 [==============================] - 0s 926us/step - loss: 0.6070 - acc: 0.0000e+00\n",
            "Epoch 360/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.6052 - acc: 0.0000e+00\n",
            "Epoch 361/1000\n",
            "4/4 [==============================] - 0s 541us/step - loss: 0.6033 - acc: 0.0000e+00\n",
            "Epoch 362/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.6014 - acc: 0.0000e+00\n",
            "Epoch 363/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.5996 - acc: 0.0000e+00\n",
            "Epoch 364/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.5977 - acc: 0.0000e+00\n",
            "Epoch 365/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.5959 - acc: 0.0000e+00\n",
            "Epoch 366/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.5941 - acc: 0.0000e+00\n",
            "Epoch 367/1000\n",
            "4/4 [==============================] - 0s 467us/step - loss: 0.5922 - acc: 0.0000e+00\n",
            "Epoch 368/1000\n",
            "4/4 [==============================] - 0s 561us/step - loss: 0.5904 - acc: 0.0000e+00\n",
            "Epoch 369/1000\n",
            "4/4 [==============================] - 0s 521us/step - loss: 0.5885 - acc: 0.0000e+00\n",
            "Epoch 370/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.5866 - acc: 0.0000e+00\n",
            "Epoch 371/1000\n",
            "4/4 [==============================] - 0s 568us/step - loss: 0.5846 - acc: 0.0000e+00\n",
            "Epoch 372/1000\n",
            "4/4 [==============================] - 0s 648us/step - loss: 0.5828 - acc: 0.0000e+00\n",
            "Epoch 373/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.5810 - acc: 0.0000e+00\n",
            "Epoch 374/1000\n",
            "4/4 [==============================] - 0s 464us/step - loss: 0.5790 - acc: 0.0000e+00\n",
            "Epoch 375/1000\n",
            "4/4 [==============================] - 0s 560us/step - loss: 0.5772 - acc: 0.0000e+00\n",
            "Epoch 376/1000\n",
            "4/4 [==============================] - 0s 735us/step - loss: 0.5753 - acc: 0.0000e+00\n",
            "Epoch 377/1000\n",
            "4/4 [==============================] - 0s 797us/step - loss: 0.5734 - acc: 0.0000e+00\n",
            "Epoch 378/1000\n",
            "4/4 [==============================] - 0s 740us/step - loss: 0.5716 - acc: 0.0000e+00\n",
            "Epoch 379/1000\n",
            "4/4 [==============================] - 0s 767us/step - loss: 0.5697 - acc: 0.0000e+00\n",
            "Epoch 380/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.5679 - acc: 0.0000e+00\n",
            "Epoch 381/1000\n",
            "4/4 [==============================] - 0s 483us/step - loss: 0.5661 - acc: 0.0000e+00\n",
            "Epoch 382/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.5642 - acc: 0.0000e+00\n",
            "Epoch 383/1000\n",
            "4/4 [==============================] - 0s 435us/step - loss: 0.5624 - acc: 0.0000e+00\n",
            "Epoch 384/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.5605 - acc: 0.0000e+00\n",
            "Epoch 385/1000\n",
            "4/4 [==============================] - 0s 494us/step - loss: 0.5586 - acc: 0.0000e+00\n",
            "Epoch 386/1000\n",
            "4/4 [==============================] - 0s 525us/step - loss: 0.5567 - acc: 0.0000e+00\n",
            "Epoch 387/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.5549 - acc: 0.0000e+00\n",
            "Epoch 388/1000\n",
            "4/4 [==============================] - 0s 810us/step - loss: 0.5531 - acc: 0.0000e+00\n",
            "Epoch 389/1000\n",
            "4/4 [==============================] - 0s 689us/step - loss: 0.5511 - acc: 0.0000e+00\n",
            "Epoch 390/1000\n",
            "4/4 [==============================] - 0s 428us/step - loss: 0.5493 - acc: 0.0000e+00\n",
            "Epoch 391/1000\n",
            "4/4 [==============================] - 0s 723us/step - loss: 0.5475 - acc: 0.0000e+00\n",
            "Epoch 392/1000\n",
            "4/4 [==============================] - 0s 436us/step - loss: 0.5457 - acc: 0.0000e+00\n",
            "Epoch 393/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.5439 - acc: 0.0000e+00\n",
            "Epoch 394/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.5420 - acc: 0.0000e+00\n",
            "Epoch 395/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.5402 - acc: 0.0000e+00\n",
            "Epoch 396/1000\n",
            "4/4 [==============================] - 0s 644us/step - loss: 0.5383 - acc: 0.0000e+00\n",
            "Epoch 397/1000\n",
            "4/4 [==============================] - 0s 487us/step - loss: 0.5364 - acc: 0.0000e+00\n",
            "Epoch 398/1000\n",
            "4/4 [==============================] - 0s 492us/step - loss: 0.5345 - acc: 0.0000e+00\n",
            "Epoch 399/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.5327 - acc: 0.0000e+00\n",
            "Epoch 400/1000\n",
            "4/4 [==============================] - 0s 460us/step - loss: 0.5308 - acc: 0.0000e+00\n",
            "Epoch 401/1000\n",
            "4/4 [==============================] - 0s 445us/step - loss: 0.5290 - acc: 0.0000e+00\n",
            "Epoch 402/1000\n",
            "4/4 [==============================] - 0s 500us/step - loss: 0.5271 - acc: 0.0000e+00\n",
            "Epoch 403/1000\n",
            "4/4 [==============================] - 0s 625us/step - loss: 0.5252 - acc: 0.0000e+00\n",
            "Epoch 404/1000\n",
            "4/4 [==============================] - 0s 617us/step - loss: 0.5235 - acc: 0.0000e+00\n",
            "Epoch 405/1000\n",
            "4/4 [==============================] - 0s 642us/step - loss: 0.5216 - acc: 0.0000e+00\n",
            "Epoch 406/1000\n",
            "4/4 [==============================] - 0s 740us/step - loss: 0.5198 - acc: 0.0000e+00\n",
            "Epoch 407/1000\n",
            "4/4 [==============================] - 0s 891us/step - loss: 0.5180 - acc: 0.0000e+00\n",
            "Epoch 408/1000\n",
            "4/4 [==============================] - 0s 850us/step - loss: 0.5162 - acc: 0.0000e+00\n",
            "Epoch 409/1000\n",
            "4/4 [==============================] - 0s 775us/step - loss: 0.5144 - acc: 0.0000e+00\n",
            "Epoch 410/1000\n",
            "4/4 [==============================] - 0s 715us/step - loss: 0.5125 - acc: 0.0000e+00\n",
            "Epoch 411/1000\n",
            "4/4 [==============================] - 0s 755us/step - loss: 0.5106 - acc: 0.0000e+00\n",
            "Epoch 412/1000\n",
            "4/4 [==============================] - 0s 458us/step - loss: 0.5089 - acc: 0.0000e+00\n",
            "Epoch 413/1000\n",
            "4/4 [==============================] - 0s 574us/step - loss: 0.5071 - acc: 0.0000e+00\n",
            "Epoch 414/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.5051 - acc: 0.0000e+00\n",
            "Epoch 415/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.5034 - acc: 0.0000e+00\n",
            "Epoch 416/1000\n",
            "4/4 [==============================] - 0s 492us/step - loss: 0.5015 - acc: 0.0000e+00\n",
            "Epoch 417/1000\n",
            "4/4 [==============================] - 0s 427us/step - loss: 0.4997 - acc: 0.0000e+00\n",
            "Epoch 418/1000\n",
            "4/4 [==============================] - 0s 466us/step - loss: 0.4980 - acc: 0.0000e+00\n",
            "Epoch 419/1000\n",
            "4/4 [==============================] - 0s 561us/step - loss: 0.4961 - acc: 0.0000e+00\n",
            "Epoch 420/1000\n",
            "4/4 [==============================] - 0s 483us/step - loss: 0.4943 - acc: 0.0000e+00\n",
            "Epoch 421/1000\n",
            "4/4 [==============================] - 0s 512us/step - loss: 0.4926 - acc: 0.0000e+00\n",
            "Epoch 422/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.4908 - acc: 0.0000e+00\n",
            "Epoch 423/1000\n",
            "4/4 [==============================] - 0s 456us/step - loss: 0.4890 - acc: 0.0000e+00\n",
            "Epoch 424/1000\n",
            "4/4 [==============================] - 0s 402us/step - loss: 0.4872 - acc: 0.0000e+00\n",
            "Epoch 425/1000\n",
            "4/4 [==============================] - 0s 472us/step - loss: 0.4853 - acc: 0.0000e+00\n",
            "Epoch 426/1000\n",
            "4/4 [==============================] - 0s 457us/step - loss: 0.4835 - acc: 0.0000e+00\n",
            "Epoch 427/1000\n",
            "4/4 [==============================] - 0s 527us/step - loss: 0.4817 - acc: 0.0000e+00\n",
            "Epoch 428/1000\n",
            "4/4 [==============================] - 0s 460us/step - loss: 0.4799 - acc: 0.0000e+00\n",
            "Epoch 429/1000\n",
            "4/4 [==============================] - 0s 479us/step - loss: 0.4782 - acc: 0.0000e+00\n",
            "Epoch 430/1000\n",
            "4/4 [==============================] - 0s 539us/step - loss: 0.4764 - acc: 0.0000e+00\n",
            "Epoch 431/1000\n",
            "4/4 [==============================] - 0s 499us/step - loss: 0.4746 - acc: 0.0000e+00\n",
            "Epoch 432/1000\n",
            "4/4 [==============================] - 0s 429us/step - loss: 0.4728 - acc: 0.0000e+00\n",
            "Epoch 433/1000\n",
            "4/4 [==============================] - 0s 729us/step - loss: 0.4710 - acc: 0.0000e+00\n",
            "Epoch 434/1000\n",
            "4/4 [==============================] - 0s 553us/step - loss: 0.4692 - acc: 0.0000e+00\n",
            "Epoch 435/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.4675 - acc: 0.0000e+00\n",
            "Epoch 436/1000\n",
            "4/4 [==============================] - 0s 496us/step - loss: 0.4657 - acc: 0.0000e+00\n",
            "Epoch 437/1000\n",
            "4/4 [==============================] - 0s 555us/step - loss: 0.4639 - acc: 0.0000e+00\n",
            "Epoch 438/1000\n",
            "4/4 [==============================] - 0s 489us/step - loss: 0.4622 - acc: 0.0000e+00\n",
            "Epoch 439/1000\n",
            "4/4 [==============================] - 0s 470us/step - loss: 0.4604 - acc: 0.0000e+00\n",
            "Epoch 440/1000\n",
            "4/4 [==============================] - 0s 479us/step - loss: 0.4587 - acc: 0.0000e+00\n",
            "Epoch 441/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.4569 - acc: 0.0000e+00\n",
            "Epoch 442/1000\n",
            "4/4 [==============================] - 0s 470us/step - loss: 0.4551 - acc: 0.0000e+00\n",
            "Epoch 443/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.4534 - acc: 0.0000e+00\n",
            "Epoch 444/1000\n",
            "4/4 [==============================] - 0s 633us/step - loss: 0.4516 - acc: 0.0000e+00\n",
            "Epoch 445/1000\n",
            "4/4 [==============================] - 0s 2ms/step - loss: 0.4499 - acc: 0.0000e+00\n",
            "Epoch 446/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.4482 - acc: 0.0000e+00\n",
            "Epoch 447/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.4465 - acc: 0.0000e+00\n",
            "Epoch 448/1000\n",
            "4/4 [==============================] - 0s 533us/step - loss: 0.4447 - acc: 0.0000e+00\n",
            "Epoch 449/1000\n",
            "4/4 [==============================] - 0s 462us/step - loss: 0.4430 - acc: 0.0000e+00\n",
            "Epoch 450/1000\n",
            "4/4 [==============================] - 0s 449us/step - loss: 0.4413 - acc: 0.0000e+00\n",
            "Epoch 451/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.4396 - acc: 0.0000e+00\n",
            "Epoch 452/1000\n",
            "4/4 [==============================] - 0s 550us/step - loss: 0.4378 - acc: 0.0000e+00\n",
            "Epoch 453/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.4360 - acc: 0.0000e+00\n",
            "Epoch 454/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.4343 - acc: 0.0000e+00\n",
            "Epoch 455/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.4326 - acc: 0.0000e+00\n",
            "Epoch 456/1000\n",
            "4/4 [==============================] - 0s 496us/step - loss: 0.4309 - acc: 0.0000e+00\n",
            "Epoch 457/1000\n",
            "4/4 [==============================] - 0s 465us/step - loss: 0.4292 - acc: 0.0000e+00\n",
            "Epoch 458/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.4275 - acc: 0.0000e+00\n",
            "Epoch 459/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.4258 - acc: 0.0000e+00\n",
            "Epoch 460/1000\n",
            "4/4 [==============================] - 0s 534us/step - loss: 0.4241 - acc: 0.0000e+00\n",
            "Epoch 461/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.4224 - acc: 0.2500\n",
            "Epoch 462/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.4207 - acc: 0.2500\n",
            "Epoch 463/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.4190 - acc: 0.2500\n",
            "Epoch 464/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.4173 - acc: 0.2500\n",
            "Epoch 465/1000\n",
            "4/4 [==============================] - 0s 741us/step - loss: 0.4156 - acc: 0.2500\n",
            "Epoch 466/1000\n",
            "4/4 [==============================] - 0s 700us/step - loss: 0.4140 - acc: 0.2500\n",
            "Epoch 467/1000\n",
            "4/4 [==============================] - 0s 806us/step - loss: 0.4123 - acc: 0.2500\n",
            "Epoch 468/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.4106 - acc: 0.2500\n",
            "Epoch 469/1000\n",
            "4/4 [==============================] - 0s 581us/step - loss: 0.4090 - acc: 0.2500\n",
            "Epoch 470/1000\n",
            "4/4 [==============================] - 0s 629us/step - loss: 0.4074 - acc: 0.2500\n",
            "Epoch 471/1000\n",
            "4/4 [==============================] - 0s 438us/step - loss: 0.4057 - acc: 0.2500\n",
            "Epoch 472/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.4040 - acc: 0.2500\n",
            "Epoch 473/1000\n",
            "4/4 [==============================] - 0s 577us/step - loss: 0.4023 - acc: 0.2500\n",
            "Epoch 474/1000\n",
            "4/4 [==============================] - 0s 635us/step - loss: 0.4006 - acc: 0.2500\n",
            "Epoch 475/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.3990 - acc: 0.2500\n",
            "Epoch 476/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.3974 - acc: 0.2500\n",
            "Epoch 477/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.3958 - acc: 0.2500\n",
            "Epoch 478/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.3941 - acc: 0.2500\n",
            "Epoch 479/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.3924 - acc: 0.2500\n",
            "Epoch 480/1000\n",
            "4/4 [==============================] - 0s 867us/step - loss: 0.3909 - acc: 0.2500\n",
            "Epoch 481/1000\n",
            "4/4 [==============================] - 0s 736us/step - loss: 0.3893 - acc: 0.2500\n",
            "Epoch 482/1000\n",
            "4/4 [==============================] - 0s 641us/step - loss: 0.3875 - acc: 0.2500\n",
            "Epoch 483/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.3860 - acc: 0.2500\n",
            "Epoch 484/1000\n",
            "4/4 [==============================] - 0s 539us/step - loss: 0.3845 - acc: 0.2500\n",
            "Epoch 485/1000\n",
            "4/4 [==============================] - 0s 632us/step - loss: 0.3829 - acc: 0.2500\n",
            "Epoch 486/1000\n",
            "4/4 [==============================] - 0s 621us/step - loss: 0.3813 - acc: 0.2500\n",
            "Epoch 487/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.3797 - acc: 0.2500\n",
            "Epoch 488/1000\n",
            "4/4 [==============================] - 0s 856us/step - loss: 0.3780 - acc: 0.2500\n",
            "Epoch 489/1000\n",
            "4/4 [==============================] - 0s 508us/step - loss: 0.3764 - acc: 0.2500\n",
            "Epoch 490/1000\n",
            "4/4 [==============================] - 0s 483us/step - loss: 0.3747 - acc: 0.2500\n",
            "Epoch 491/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.3731 - acc: 0.2500\n",
            "Epoch 492/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.3715 - acc: 0.2500\n",
            "Epoch 493/1000\n",
            "4/4 [==============================] - 0s 579us/step - loss: 0.3700 - acc: 0.2500\n",
            "Epoch 494/1000\n",
            "4/4 [==============================] - 0s 557us/step - loss: 0.3684 - acc: 0.2500\n",
            "Epoch 495/1000\n",
            "4/4 [==============================] - 0s 468us/step - loss: 0.3669 - acc: 0.2500\n",
            "Epoch 496/1000\n",
            "4/4 [==============================] - 0s 577us/step - loss: 0.3652 - acc: 0.2500\n",
            "Epoch 497/1000\n",
            "4/4 [==============================] - 0s 549us/step - loss: 0.3637 - acc: 0.2500\n",
            "Epoch 498/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.3622 - acc: 0.2500\n",
            "Epoch 499/1000\n",
            "4/4 [==============================] - 0s 700us/step - loss: 0.3606 - acc: 0.2500\n",
            "Epoch 500/1000\n",
            "4/4 [==============================] - 0s 514us/step - loss: 0.3591 - acc: 0.2500\n",
            "Epoch 501/1000\n",
            "4/4 [==============================] - 0s 485us/step - loss: 0.3575 - acc: 0.2500\n",
            "Epoch 502/1000\n",
            "4/4 [==============================] - 0s 516us/step - loss: 0.3559 - acc: 0.2500\n",
            "Epoch 503/1000\n",
            "4/4 [==============================] - 0s 600us/step - loss: 0.3544 - acc: 0.2500\n",
            "Epoch 504/1000\n",
            "4/4 [==============================] - 0s 574us/step - loss: 0.3528 - acc: 0.2500\n",
            "Epoch 505/1000\n",
            "4/4 [==============================] - 0s 598us/step - loss: 0.3513 - acc: 0.2500\n",
            "Epoch 506/1000\n",
            "4/4 [==============================] - 0s 580us/step - loss: 0.3498 - acc: 0.2500\n",
            "Epoch 507/1000\n",
            "4/4 [==============================] - 0s 607us/step - loss: 0.3482 - acc: 0.2500\n",
            "Epoch 508/1000\n",
            "4/4 [==============================] - 0s 603us/step - loss: 0.3468 - acc: 0.2500\n",
            "Epoch 509/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.3453 - acc: 0.2500\n",
            "Epoch 510/1000\n",
            "4/4 [==============================] - 0s 658us/step - loss: 0.3437 - acc: 0.2500\n",
            "Epoch 511/1000\n",
            "4/4 [==============================] - 0s 534us/step - loss: 0.3422 - acc: 0.2500\n",
            "Epoch 512/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.3407 - acc: 0.2500\n",
            "Epoch 513/1000\n",
            "4/4 [==============================] - 0s 600us/step - loss: 0.3392 - acc: 0.2500\n",
            "Epoch 514/1000\n",
            "4/4 [==============================] - 0s 601us/step - loss: 0.3378 - acc: 0.2500\n",
            "Epoch 515/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.3363 - acc: 0.2500\n",
            "Epoch 516/1000\n",
            "4/4 [==============================] - 0s 641us/step - loss: 0.3347 - acc: 0.2500\n",
            "Epoch 517/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.3333 - acc: 0.2500\n",
            "Epoch 518/1000\n",
            "4/4 [==============================] - 0s 724us/step - loss: 0.3318 - acc: 0.2500\n",
            "Epoch 519/1000\n",
            "4/4 [==============================] - 0s 972us/step - loss: 0.3303 - acc: 0.2500\n",
            "Epoch 520/1000\n",
            "4/4 [==============================] - 0s 650us/step - loss: 0.3289 - acc: 0.2500\n",
            "Epoch 521/1000\n",
            "4/4 [==============================] - 0s 621us/step - loss: 0.3274 - acc: 0.2500\n",
            "Epoch 522/1000\n",
            "4/4 [==============================] - 0s 721us/step - loss: 0.3259 - acc: 0.2500\n",
            "Epoch 523/1000\n",
            "4/4 [==============================] - 0s 614us/step - loss: 0.3245 - acc: 0.2500\n",
            "Epoch 524/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.3231 - acc: 0.2500\n",
            "Epoch 525/1000\n",
            "4/4 [==============================] - 0s 660us/step - loss: 0.3216 - acc: 0.2500\n",
            "Epoch 526/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.3202 - acc: 0.2500\n",
            "Epoch 527/1000\n",
            "4/4 [==============================] - 0s 461us/step - loss: 0.3187 - acc: 0.2500\n",
            "Epoch 528/1000\n",
            "4/4 [==============================] - 0s 513us/step - loss: 0.3173 - acc: 0.2500\n",
            "Epoch 529/1000\n",
            "4/4 [==============================] - 0s 716us/step - loss: 0.3159 - acc: 0.2500\n",
            "Epoch 530/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.3145 - acc: 0.2500\n",
            "Epoch 531/1000\n",
            "4/4 [==============================] - 0s 644us/step - loss: 0.3130 - acc: 0.2500\n",
            "Epoch 532/1000\n",
            "4/4 [==============================] - 0s 620us/step - loss: 0.3116 - acc: 0.2500\n",
            "Epoch 533/1000\n",
            "4/4 [==============================] - 0s 391us/step - loss: 0.3102 - acc: 0.2500\n",
            "Epoch 534/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.3088 - acc: 0.2500\n",
            "Epoch 535/1000\n",
            "4/4 [==============================] - 0s 469us/step - loss: 0.3074 - acc: 0.2500\n",
            "Epoch 536/1000\n",
            "4/4 [==============================] - 0s 501us/step - loss: 0.3060 - acc: 0.2500\n",
            "Epoch 537/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.3046 - acc: 0.2500\n",
            "Epoch 538/1000\n",
            "4/4 [==============================] - 0s 463us/step - loss: 0.3033 - acc: 0.2500\n",
            "Epoch 539/1000\n",
            "4/4 [==============================] - 0s 545us/step - loss: 0.3019 - acc: 0.2500\n",
            "Epoch 540/1000\n",
            "4/4 [==============================] - 0s 586us/step - loss: 0.3004 - acc: 0.2500\n",
            "Epoch 541/1000\n",
            "4/4 [==============================] - 0s 597us/step - loss: 0.2991 - acc: 0.2500\n",
            "Epoch 542/1000\n",
            "4/4 [==============================] - 0s 604us/step - loss: 0.2977 - acc: 0.2500\n",
            "Epoch 543/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.2963 - acc: 0.2500\n",
            "Epoch 544/1000\n",
            "4/4 [==============================] - 0s 534us/step - loss: 0.2950 - acc: 0.2500\n",
            "Epoch 545/1000\n",
            "4/4 [==============================] - 0s 595us/step - loss: 0.2936 - acc: 0.2500\n",
            "Epoch 546/1000\n",
            "4/4 [==============================] - 0s 579us/step - loss: 0.2923 - acc: 0.2500\n",
            "Epoch 547/1000\n",
            "4/4 [==============================] - 0s 561us/step - loss: 0.2910 - acc: 0.2500\n",
            "Epoch 548/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.2896 - acc: 0.2500\n",
            "Epoch 549/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.2883 - acc: 0.2500\n",
            "Epoch 550/1000\n",
            "4/4 [==============================] - 0s 521us/step - loss: 0.2869 - acc: 0.2500\n",
            "Epoch 551/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.2856 - acc: 0.2500\n",
            "Epoch 552/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.2843 - acc: 0.2500\n",
            "Epoch 553/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.2830 - acc: 0.2500\n",
            "Epoch 554/1000\n",
            "4/4 [==============================] - 0s 466us/step - loss: 0.2817 - acc: 0.2500\n",
            "Epoch 555/1000\n",
            "4/4 [==============================] - 0s 417us/step - loss: 0.2804 - acc: 0.2500\n",
            "Epoch 556/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.2791 - acc: 0.2500\n",
            "Epoch 557/1000\n",
            "4/4 [==============================] - 0s 460us/step - loss: 0.2778 - acc: 0.2500\n",
            "Epoch 558/1000\n",
            "4/4 [==============================] - 0s 493us/step - loss: 0.2765 - acc: 0.2500\n",
            "Epoch 559/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.2752 - acc: 0.2500\n",
            "Epoch 560/1000\n",
            "4/4 [==============================] - 0s 526us/step - loss: 0.2739 - acc: 0.2500\n",
            "Epoch 561/1000\n",
            "4/4 [==============================] - 0s 620us/step - loss: 0.2726 - acc: 0.2500\n",
            "Epoch 562/1000\n",
            "4/4 [==============================] - 0s 569us/step - loss: 0.2714 - acc: 0.2500\n",
            "Epoch 563/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.2702 - acc: 0.2500\n",
            "Epoch 564/1000\n",
            "4/4 [==============================] - 0s 549us/step - loss: 0.2689 - acc: 0.7500\n",
            "Epoch 565/1000\n",
            "4/4 [==============================] - 0s 580us/step - loss: 0.2676 - acc: 0.7500\n",
            "Epoch 566/1000\n",
            "4/4 [==============================] - 0s 489us/step - loss: 0.2663 - acc: 0.7500\n",
            "Epoch 567/1000\n",
            "4/4 [==============================] - 0s 579us/step - loss: 0.2652 - acc: 0.7500\n",
            "Epoch 568/1000\n",
            "4/4 [==============================] - 0s 525us/step - loss: 0.2640 - acc: 0.7500\n",
            "Epoch 569/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.2627 - acc: 0.7500\n",
            "Epoch 570/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.2614 - acc: 0.7500\n",
            "Epoch 571/1000\n",
            "4/4 [==============================] - 0s 593us/step - loss: 0.2602 - acc: 0.7500\n",
            "Epoch 572/1000\n",
            "4/4 [==============================] - 0s 523us/step - loss: 0.2590 - acc: 0.7500\n",
            "Epoch 573/1000\n",
            "4/4 [==============================] - 0s 417us/step - loss: 0.2578 - acc: 0.7500\n",
            "Epoch 574/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.2565 - acc: 0.7500\n",
            "Epoch 575/1000\n",
            "4/4 [==============================] - 0s 499us/step - loss: 0.2554 - acc: 0.7500\n",
            "Epoch 576/1000\n",
            "4/4 [==============================] - 0s 433us/step - loss: 0.2542 - acc: 0.7500\n",
            "Epoch 577/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.2529 - acc: 0.7500\n",
            "Epoch 578/1000\n",
            "4/4 [==============================] - 0s 489us/step - loss: 0.2517 - acc: 0.7500\n",
            "Epoch 579/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.2506 - acc: 0.7500\n",
            "Epoch 580/1000\n",
            "4/4 [==============================] - 0s 596us/step - loss: 0.2494 - acc: 0.7500\n",
            "Epoch 581/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.2482 - acc: 0.7500\n",
            "Epoch 582/1000\n",
            "4/4 [==============================] - 0s 541us/step - loss: 0.2470 - acc: 0.7500\n",
            "Epoch 583/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.2459 - acc: 0.7500\n",
            "Epoch 584/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.2447 - acc: 0.7500\n",
            "Epoch 585/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.2435 - acc: 0.7500\n",
            "Epoch 586/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.2424 - acc: 0.7500\n",
            "Epoch 587/1000\n",
            "4/4 [==============================] - 0s 566us/step - loss: 0.2413 - acc: 0.7500\n",
            "Epoch 588/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.2401 - acc: 0.7500\n",
            "Epoch 589/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.2390 - acc: 0.7500\n",
            "Epoch 590/1000\n",
            "4/4 [==============================] - 0s 460us/step - loss: 0.2378 - acc: 0.7500\n",
            "Epoch 591/1000\n",
            "4/4 [==============================] - 0s 559us/step - loss: 0.2367 - acc: 0.7500\n",
            "Epoch 592/1000\n",
            "4/4 [==============================] - 0s 460us/step - loss: 0.2356 - acc: 0.7500\n",
            "Epoch 593/1000\n",
            "4/4 [==============================] - 0s 546us/step - loss: 0.2345 - acc: 0.7500\n",
            "Epoch 594/1000\n",
            "4/4 [==============================] - 0s 425us/step - loss: 0.2333 - acc: 0.7500\n",
            "Epoch 595/1000\n",
            "4/4 [==============================] - 0s 463us/step - loss: 0.2323 - acc: 0.7500\n",
            "Epoch 596/1000\n",
            "4/4 [==============================] - 0s 486us/step - loss: 0.2312 - acc: 0.7500\n",
            "Epoch 597/1000\n",
            "4/4 [==============================] - 0s 468us/step - loss: 0.2300 - acc: 0.7500\n",
            "Epoch 598/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.2289 - acc: 0.7500\n",
            "Epoch 599/1000\n",
            "4/4 [==============================] - 0s 561us/step - loss: 0.2278 - acc: 0.7500\n",
            "Epoch 600/1000\n",
            "4/4 [==============================] - 0s 512us/step - loss: 0.2268 - acc: 0.7500\n",
            "Epoch 601/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.2257 - acc: 0.7500\n",
            "Epoch 602/1000\n",
            "4/4 [==============================] - 0s 502us/step - loss: 0.2246 - acc: 0.7500\n",
            "Epoch 603/1000\n",
            "4/4 [==============================] - 0s 538us/step - loss: 0.2235 - acc: 0.7500\n",
            "Epoch 604/1000\n",
            "4/4 [==============================] - 0s 508us/step - loss: 0.2225 - acc: 0.7500\n",
            "Epoch 605/1000\n",
            "4/4 [==============================] - 0s 455us/step - loss: 0.2214 - acc: 0.7500\n",
            "Epoch 606/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.2204 - acc: 0.7500\n",
            "Epoch 607/1000\n",
            "4/4 [==============================] - 0s 551us/step - loss: 0.2193 - acc: 0.7500\n",
            "Epoch 608/1000\n",
            "4/4 [==============================] - 0s 612us/step - loss: 0.2183 - acc: 0.7500\n",
            "Epoch 609/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.2172 - acc: 0.7500\n",
            "Epoch 610/1000\n",
            "4/4 [==============================] - 0s 579us/step - loss: 0.2162 - acc: 0.7500\n",
            "Epoch 611/1000\n",
            "4/4 [==============================] - 0s 446us/step - loss: 0.2152 - acc: 0.7500\n",
            "Epoch 612/1000\n",
            "4/4 [==============================] - 0s 447us/step - loss: 0.2141 - acc: 0.7500\n",
            "Epoch 613/1000\n",
            "4/4 [==============================] - 0s 490us/step - loss: 0.2131 - acc: 0.7500\n",
            "Epoch 614/1000\n",
            "4/4 [==============================] - 0s 679us/step - loss: 0.2121 - acc: 0.7500\n",
            "Epoch 615/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.2111 - acc: 0.7500\n",
            "Epoch 616/1000\n",
            "4/4 [==============================] - 0s 490us/step - loss: 0.2101 - acc: 0.7500\n",
            "Epoch 617/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.2091 - acc: 0.7500\n",
            "Epoch 618/1000\n",
            "4/4 [==============================] - 0s 514us/step - loss: 0.2081 - acc: 0.7500\n",
            "Epoch 619/1000\n",
            "4/4 [==============================] - 0s 482us/step - loss: 0.2071 - acc: 0.7500\n",
            "Epoch 620/1000\n",
            "4/4 [==============================] - 0s 478us/step - loss: 0.2061 - acc: 0.7500\n",
            "Epoch 621/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.2052 - acc: 0.7500\n",
            "Epoch 622/1000\n",
            "4/4 [==============================] - 0s 566us/step - loss: 0.2043 - acc: 0.7500\n",
            "Epoch 623/1000\n",
            "4/4 [==============================] - 0s 583us/step - loss: 0.2033 - acc: 0.7500\n",
            "Epoch 624/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.2023 - acc: 0.7500\n",
            "Epoch 625/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.2013 - acc: 0.7500\n",
            "Epoch 626/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.2003 - acc: 0.7500\n",
            "Epoch 627/1000\n",
            "4/4 [==============================] - 0s 488us/step - loss: 0.1994 - acc: 0.7500\n",
            "Epoch 628/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.1984 - acc: 0.7500\n",
            "Epoch 629/1000\n",
            "4/4 [==============================] - 0s 552us/step - loss: 0.1975 - acc: 0.7500\n",
            "Epoch 630/1000\n",
            "4/4 [==============================] - 0s 515us/step - loss: 0.1966 - acc: 0.7500\n",
            "Epoch 631/1000\n",
            "4/4 [==============================] - 0s 587us/step - loss: 0.1956 - acc: 0.7500\n",
            "Epoch 632/1000\n",
            "4/4 [==============================] - 0s 480us/step - loss: 0.1947 - acc: 0.7500\n",
            "Epoch 633/1000\n",
            "4/4 [==============================] - 0s 545us/step - loss: 0.1938 - acc: 0.7500\n",
            "Epoch 634/1000\n",
            "4/4 [==============================] - 0s 577us/step - loss: 0.1929 - acc: 0.7500\n",
            "Epoch 635/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.1920 - acc: 0.7500\n",
            "Epoch 636/1000\n",
            "4/4 [==============================] - 0s 526us/step - loss: 0.1910 - acc: 0.7500\n",
            "Epoch 637/1000\n",
            "4/4 [==============================] - 0s 760us/step - loss: 0.1902 - acc: 0.7500\n",
            "Epoch 638/1000\n",
            "4/4 [==============================] - 0s 929us/step - loss: 0.1893 - acc: 0.7500\n",
            "Epoch 639/1000\n",
            "4/4 [==============================] - 0s 602us/step - loss: 0.1883 - acc: 0.7500\n",
            "Epoch 640/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.1875 - acc: 0.7500\n",
            "Epoch 641/1000\n",
            "4/4 [==============================] - 0s 479us/step - loss: 0.1866 - acc: 0.7500\n",
            "Epoch 642/1000\n",
            "4/4 [==============================] - 0s 478us/step - loss: 0.1857 - acc: 0.7500\n",
            "Epoch 643/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.1848 - acc: 0.7500\n",
            "Epoch 644/1000\n",
            "4/4 [==============================] - 0s 559us/step - loss: 0.1840 - acc: 0.7500\n",
            "Epoch 645/1000\n",
            "4/4 [==============================] - 0s 452us/step - loss: 0.1831 - acc: 0.7500\n",
            "Epoch 646/1000\n",
            "4/4 [==============================] - 0s 499us/step - loss: 0.1822 - acc: 0.7500\n",
            "Epoch 647/1000\n",
            "4/4 [==============================] - 0s 648us/step - loss: 0.1814 - acc: 0.7500\n",
            "Epoch 648/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.1806 - acc: 0.7500\n",
            "Epoch 649/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.1797 - acc: 0.7500\n",
            "Epoch 650/1000\n",
            "4/4 [==============================] - 0s 565us/step - loss: 0.1789 - acc: 0.7500\n",
            "Epoch 651/1000\n",
            "4/4 [==============================] - 0s 481us/step - loss: 0.1781 - acc: 0.7500\n",
            "Epoch 652/1000\n",
            "4/4 [==============================] - 0s 407us/step - loss: 0.1772 - acc: 0.7500\n",
            "Epoch 653/1000\n",
            "4/4 [==============================] - 0s 582us/step - loss: 0.1764 - acc: 0.7500\n",
            "Epoch 654/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.1756 - acc: 0.7500\n",
            "Epoch 655/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.1747 - acc: 0.7500\n",
            "Epoch 656/1000\n",
            "4/4 [==============================] - 0s 463us/step - loss: 0.1739 - acc: 0.7500\n",
            "Epoch 657/1000\n",
            "4/4 [==============================] - 0s 525us/step - loss: 0.1731 - acc: 0.7500\n",
            "Epoch 658/1000\n",
            "4/4 [==============================] - 0s 516us/step - loss: 0.1723 - acc: 0.7500\n",
            "Epoch 659/1000\n",
            "4/4 [==============================] - 0s 504us/step - loss: 0.1715 - acc: 0.7500\n",
            "Epoch 660/1000\n",
            "4/4 [==============================] - 0s 628us/step - loss: 0.1708 - acc: 0.7500\n",
            "Epoch 661/1000\n",
            "4/4 [==============================] - 0s 510us/step - loss: 0.1700 - acc: 0.7500\n",
            "Epoch 662/1000\n",
            "4/4 [==============================] - 0s 547us/step - loss: 0.1692 - acc: 0.7500\n",
            "Epoch 663/1000\n",
            "4/4 [==============================] - 0s 573us/step - loss: 0.1684 - acc: 0.7500\n",
            "Epoch 664/1000\n",
            "4/4 [==============================] - 0s 479us/step - loss: 0.1676 - acc: 0.7500\n",
            "Epoch 665/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.1669 - acc: 0.7500\n",
            "Epoch 666/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.1661 - acc: 0.7500\n",
            "Epoch 667/1000\n",
            "4/4 [==============================] - 0s 500us/step - loss: 0.1654 - acc: 0.7500\n",
            "Epoch 668/1000\n",
            "4/4 [==============================] - 0s 550us/step - loss: 0.1646 - acc: 0.7500\n",
            "Epoch 669/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.1639 - acc: 0.7500\n",
            "Epoch 670/1000\n",
            "4/4 [==============================] - 0s 514us/step - loss: 0.1631 - acc: 0.7500\n",
            "Epoch 671/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.1624 - acc: 0.7500\n",
            "Epoch 672/1000\n",
            "4/4 [==============================] - 0s 426us/step - loss: 0.1616 - acc: 0.7500\n",
            "Epoch 673/1000\n",
            "4/4 [==============================] - 0s 585us/step - loss: 0.1609 - acc: 0.7500\n",
            "Epoch 674/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.1602 - acc: 0.7500\n",
            "Epoch 675/1000\n",
            "4/4 [==============================] - 0s 480us/step - loss: 0.1595 - acc: 0.7500\n",
            "Epoch 676/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.1588 - acc: 0.7500\n",
            "Epoch 677/1000\n",
            "4/4 [==============================] - 0s 524us/step - loss: 0.1580 - acc: 0.7500\n",
            "Epoch 678/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.1573 - acc: 0.7500\n",
            "Epoch 679/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.1567 - acc: 0.7500\n",
            "Epoch 680/1000\n",
            "4/4 [==============================] - 0s 658us/step - loss: 0.1560 - acc: 0.7500\n",
            "Epoch 681/1000\n",
            "4/4 [==============================] - 0s 530us/step - loss: 0.1552 - acc: 0.7500\n",
            "Epoch 682/1000\n",
            "4/4 [==============================] - 0s 484us/step - loss: 0.1545 - acc: 0.7500\n",
            "Epoch 683/1000\n",
            "4/4 [==============================] - 0s 497us/step - loss: 0.1538 - acc: 0.7500\n",
            "Epoch 684/1000\n",
            "4/4 [==============================] - 0s 368us/step - loss: 0.1532 - acc: 0.7500\n",
            "Epoch 685/1000\n",
            "4/4 [==============================] - 0s 397us/step - loss: 0.1525 - acc: 0.7500\n",
            "Epoch 686/1000\n",
            "4/4 [==============================] - 0s 442us/step - loss: 0.1518 - acc: 0.7500\n",
            "Epoch 687/1000\n",
            "4/4 [==============================] - 0s 475us/step - loss: 0.1511 - acc: 0.7500\n",
            "Epoch 688/1000\n",
            "4/4 [==============================] - 0s 468us/step - loss: 0.1504 - acc: 0.7500\n",
            "Epoch 689/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.1498 - acc: 0.7500\n",
            "Epoch 690/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.1491 - acc: 0.7500\n",
            "Epoch 691/1000\n",
            "4/4 [==============================] - 0s 565us/step - loss: 0.1484 - acc: 0.7500\n",
            "Epoch 692/1000\n",
            "4/4 [==============================] - 0s 595us/step - loss: 0.1478 - acc: 0.7500\n",
            "Epoch 693/1000\n",
            "4/4 [==============================] - 0s 575us/step - loss: 0.1472 - acc: 0.7500\n",
            "Epoch 694/1000\n",
            "4/4 [==============================] - 0s 491us/step - loss: 0.1465 - acc: 0.7500\n",
            "Epoch 695/1000\n",
            "4/4 [==============================] - 0s 558us/step - loss: 0.1459 - acc: 0.7500\n",
            "Epoch 696/1000\n",
            "4/4 [==============================] - 0s 475us/step - loss: 0.1453 - acc: 0.7500\n",
            "Epoch 697/1000\n",
            "4/4 [==============================] - 0s 542us/step - loss: 0.1447 - acc: 0.7500\n",
            "Epoch 698/1000\n",
            "4/4 [==============================] - 0s 707us/step - loss: 0.1441 - acc: 0.7500\n",
            "Epoch 699/1000\n",
            "4/4 [==============================] - 0s 851us/step - loss: 0.1434 - acc: 0.7500\n",
            "Epoch 700/1000\n",
            "4/4 [==============================] - 0s 514us/step - loss: 0.1427 - acc: 0.7500\n",
            "Epoch 701/1000\n",
            "4/4 [==============================] - 0s 419us/step - loss: 0.1421 - acc: 0.7500\n",
            "Epoch 702/1000\n",
            "4/4 [==============================] - 0s 532us/step - loss: 0.1415 - acc: 0.7500\n",
            "Epoch 703/1000\n",
            "4/4 [==============================] - 0s 630us/step - loss: 0.1409 - acc: 0.7500\n",
            "Epoch 704/1000\n",
            "4/4 [==============================] - 0s 577us/step - loss: 0.1403 - acc: 0.7500\n",
            "Epoch 705/1000\n",
            "4/4 [==============================] - 0s 557us/step - loss: 0.1397 - acc: 0.7500\n",
            "Epoch 706/1000\n",
            "4/4 [==============================] - 0s 427us/step - loss: 0.1392 - acc: 0.7500\n",
            "Epoch 707/1000\n",
            "4/4 [==============================] - 0s 593us/step - loss: 0.1386 - acc: 0.7500\n",
            "Epoch 708/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.1380 - acc: 0.7500\n",
            "Epoch 709/1000\n",
            "4/4 [==============================] - 0s 482us/step - loss: 0.1374 - acc: 0.7500\n",
            "Epoch 710/1000\n",
            "4/4 [==============================] - 0s 490us/step - loss: 0.1368 - acc: 0.7500\n",
            "Epoch 711/1000\n",
            "4/4 [==============================] - 0s 865us/step - loss: 0.1363 - acc: 0.7500\n",
            "Epoch 712/1000\n",
            "4/4 [==============================] - 0s 567us/step - loss: 0.1357 - acc: 0.7500\n",
            "Epoch 713/1000\n",
            "4/4 [==============================] - 0s 571us/step - loss: 0.1351 - acc: 0.7500\n",
            "Epoch 714/1000\n",
            "4/4 [==============================] - 0s 656us/step - loss: 0.1345 - acc: 0.7500\n",
            "Epoch 715/1000\n",
            "4/4 [==============================] - 0s 492us/step - loss: 0.1340 - acc: 0.7500\n",
            "Epoch 716/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.1335 - acc: 0.7500\n",
            "Epoch 717/1000\n",
            "4/4 [==============================] - 0s 556us/step - loss: 0.1329 - acc: 0.7500\n",
            "Epoch 718/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.1324 - acc: 0.7500\n",
            "Epoch 719/1000\n",
            "4/4 [==============================] - 0s 451us/step - loss: 0.1318 - acc: 0.7500\n",
            "Epoch 720/1000\n",
            "4/4 [==============================] - 0s 513us/step - loss: 0.1313 - acc: 0.7500\n",
            "Epoch 721/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.1307 - acc: 0.7500\n",
            "Epoch 722/1000\n",
            "4/4 [==============================] - 0s 503us/step - loss: 0.1302 - acc: 0.7500\n",
            "Epoch 723/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.1297 - acc: 0.7500\n",
            "Epoch 724/1000\n",
            "4/4 [==============================] - 0s 476us/step - loss: 0.1292 - acc: 0.7500\n",
            "Epoch 725/1000\n",
            "4/4 [==============================] - 0s 505us/step - loss: 0.1287 - acc: 0.7500\n",
            "Epoch 726/1000\n",
            "4/4 [==============================] - 0s 585us/step - loss: 0.1282 - acc: 0.7500\n",
            "Epoch 727/1000\n",
            "4/4 [==============================] - 0s 417us/step - loss: 0.1276 - acc: 0.7500\n",
            "Epoch 728/1000\n",
            "4/4 [==============================] - 0s 553us/step - loss: 0.1272 - acc: 0.7500\n",
            "Epoch 729/1000\n",
            "4/4 [==============================] - 0s 596us/step - loss: 0.1267 - acc: 0.7500\n",
            "Epoch 730/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.1262 - acc: 0.7500\n",
            "Epoch 731/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.1257 - acc: 0.7500\n",
            "Epoch 732/1000\n",
            "4/4 [==============================] - 0s 605us/step - loss: 0.1252 - acc: 0.7500\n",
            "Epoch 733/1000\n",
            "4/4 [==============================] - 0s 496us/step - loss: 0.1247 - acc: 0.7500\n",
            "Epoch 734/1000\n",
            "4/4 [==============================] - 0s 422us/step - loss: 0.1243 - acc: 0.7500\n",
            "Epoch 735/1000\n",
            "4/4 [==============================] - 0s 560us/step - loss: 0.1238 - acc: 0.7500\n",
            "Epoch 736/1000\n",
            "4/4 [==============================] - 0s 717us/step - loss: 0.1233 - acc: 0.7500\n",
            "Epoch 737/1000\n",
            "4/4 [==============================] - 0s 508us/step - loss: 0.1228 - acc: 0.7500\n",
            "Epoch 738/1000\n",
            "4/4 [==============================] - 0s 447us/step - loss: 0.1223 - acc: 0.7500\n",
            "Epoch 739/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.1219 - acc: 0.7500\n",
            "Epoch 740/1000\n",
            "4/4 [==============================] - 0s 553us/step - loss: 0.1214 - acc: 0.7500\n",
            "Epoch 741/1000\n",
            "4/4 [==============================] - 0s 556us/step - loss: 0.1209 - acc: 0.7500\n",
            "Epoch 742/1000\n",
            "4/4 [==============================] - 0s 507us/step - loss: 0.1205 - acc: 0.7500\n",
            "Epoch 743/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.1201 - acc: 0.7500\n",
            "Epoch 744/1000\n",
            "4/4 [==============================] - 0s 487us/step - loss: 0.1196 - acc: 0.7500\n",
            "Epoch 745/1000\n",
            "4/4 [==============================] - 0s 509us/step - loss: 0.1191 - acc: 0.7500\n",
            "Epoch 746/1000\n",
            "4/4 [==============================] - 0s 552us/step - loss: 0.1187 - acc: 0.7500\n",
            "Epoch 747/1000\n",
            "4/4 [==============================] - 0s 516us/step - loss: 0.1183 - acc: 0.7500\n",
            "Epoch 748/1000\n",
            "4/4 [==============================] - 0s 966us/step - loss: 0.1179 - acc: 0.7500\n",
            "Epoch 749/1000\n",
            "4/4 [==============================] - 0s 448us/step - loss: 0.1174 - acc: 0.7500\n",
            "Epoch 750/1000\n",
            "4/4 [==============================] - 0s 730us/step - loss: 0.1170 - acc: 0.7500\n",
            "Epoch 751/1000\n",
            "4/4 [==============================] - 0s 541us/step - loss: 0.1165 - acc: 0.7500\n",
            "Epoch 752/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.1161 - acc: 0.7500\n",
            "Epoch 753/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.1157 - acc: 0.7500\n",
            "Epoch 754/1000\n",
            "4/4 [==============================] - 0s 481us/step - loss: 0.1153 - acc: 0.7500\n",
            "Epoch 755/1000\n",
            "4/4 [==============================] - 0s 487us/step - loss: 0.1149 - acc: 0.7500\n",
            "Epoch 756/1000\n",
            "4/4 [==============================] - 0s 502us/step - loss: 0.1145 - acc: 0.7500\n",
            "Epoch 757/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.1141 - acc: 0.7500\n",
            "Epoch 758/1000\n",
            "4/4 [==============================] - 0s 546us/step - loss: 0.1137 - acc: 0.7500\n",
            "Epoch 759/1000\n",
            "4/4 [==============================] - 0s 498us/step - loss: 0.1133 - acc: 0.7500\n",
            "Epoch 760/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.1129 - acc: 0.7500\n",
            "Epoch 761/1000\n",
            "4/4 [==============================] - 0s 471us/step - loss: 0.1125 - acc: 0.7500\n",
            "Epoch 762/1000\n",
            "4/4 [==============================] - 0s 687us/step - loss: 0.1121 - acc: 0.7500\n",
            "Epoch 763/1000\n",
            "4/4 [==============================] - 0s 598us/step - loss: 0.1117 - acc: 0.7500\n",
            "Epoch 764/1000\n",
            "4/4 [==============================] - 0s 575us/step - loss: 0.1113 - acc: 0.7500\n",
            "Epoch 765/1000\n",
            "4/4 [==============================] - 0s 793us/step - loss: 0.1109 - acc: 0.7500\n",
            "Epoch 766/1000\n",
            "4/4 [==============================] - 0s 777us/step - loss: 0.1106 - acc: 0.7500\n",
            "Epoch 767/1000\n",
            "4/4 [==============================] - 0s 554us/step - loss: 0.1102 - acc: 0.7500\n",
            "Epoch 768/1000\n",
            "4/4 [==============================] - 0s 766us/step - loss: 0.1098 - acc: 0.7500\n",
            "Epoch 769/1000\n",
            "4/4 [==============================] - 0s 734us/step - loss: 0.1095 - acc: 0.7500\n",
            "Epoch 770/1000\n",
            "4/4 [==============================] - 0s 583us/step - loss: 0.1091 - acc: 0.7500\n",
            "Epoch 771/1000\n",
            "4/4 [==============================] - 0s 928us/step - loss: 0.1087 - acc: 0.7500\n",
            "Epoch 772/1000\n",
            "4/4 [==============================] - 0s 991us/step - loss: 0.1084 - acc: 0.7500\n",
            "Epoch 773/1000\n",
            "4/4 [==============================] - 0s 2ms/step - loss: 0.1080 - acc: 0.7500\n",
            "Epoch 774/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1077 - acc: 0.7500\n",
            "Epoch 775/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1073 - acc: 0.7500\n",
            "Epoch 776/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1070 - acc: 0.7500\n",
            "Epoch 777/1000\n",
            "4/4 [==============================] - 0s 753us/step - loss: 0.1067 - acc: 0.7500\n",
            "Epoch 778/1000\n",
            "4/4 [==============================] - 0s 917us/step - loss: 0.1063 - acc: 0.7500\n",
            "Epoch 779/1000\n",
            "4/4 [==============================] - 0s 997us/step - loss: 0.1060 - acc: 0.7500\n",
            "Epoch 780/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1056 - acc: 0.7500\n",
            "Epoch 781/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1053 - acc: 0.7500\n",
            "Epoch 782/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1050 - acc: 0.7500\n",
            "Epoch 783/1000\n",
            "4/4 [==============================] - 0s 722us/step - loss: 0.1047 - acc: 0.7500\n",
            "Epoch 784/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1043 - acc: 0.7500\n",
            "Epoch 785/1000\n",
            "4/4 [==============================] - 0s 746us/step - loss: 0.1040 - acc: 0.7500\n",
            "Epoch 786/1000\n",
            "4/4 [==============================] - 0s 674us/step - loss: 0.1037 - acc: 0.7500\n",
            "Epoch 787/1000\n",
            "4/4 [==============================] - 0s 654us/step - loss: 0.1034 - acc: 0.7500\n",
            "Epoch 788/1000\n",
            "4/4 [==============================] - 0s 714us/step - loss: 0.1031 - acc: 0.7500\n",
            "Epoch 789/1000\n",
            "4/4 [==============================] - 0s 692us/step - loss: 0.1028 - acc: 0.7500\n",
            "Epoch 790/1000\n",
            "4/4 [==============================] - 0s 604us/step - loss: 0.1025 - acc: 0.7500\n",
            "Epoch 791/1000\n",
            "4/4 [==============================] - 0s 775us/step - loss: 0.1021 - acc: 0.7500\n",
            "Epoch 792/1000\n",
            "4/4 [==============================] - 0s 919us/step - loss: 0.1018 - acc: 0.7500\n",
            "Epoch 793/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1015 - acc: 0.7500\n",
            "Epoch 794/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1013 - acc: 0.7500\n",
            "Epoch 795/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1010 - acc: 0.7500\n",
            "Epoch 796/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1007 - acc: 0.7500\n",
            "Epoch 797/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1004 - acc: 0.7500\n",
            "Epoch 798/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.1001 - acc: 0.7500\n",
            "Epoch 799/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0998 - acc: 0.7500\n",
            "Epoch 800/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0996 - acc: 0.7500\n",
            "Epoch 801/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0993 - acc: 0.7500\n",
            "Epoch 802/1000\n",
            "4/4 [==============================] - 0s 677us/step - loss: 0.0990 - acc: 0.7500\n",
            "Epoch 803/1000\n",
            "4/4 [==============================] - 0s 642us/step - loss: 0.0988 - acc: 0.7500\n",
            "Epoch 804/1000\n",
            "4/4 [==============================] - 0s 694us/step - loss: 0.0985 - acc: 0.7500\n",
            "Epoch 805/1000\n",
            "4/4 [==============================] - 0s 611us/step - loss: 0.0982 - acc: 0.7500\n",
            "Epoch 806/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.0980 - acc: 0.7500\n",
            "Epoch 807/1000\n",
            "4/4 [==============================] - 0s 604us/step - loss: 0.0977 - acc: 0.7500\n",
            "Epoch 808/1000\n",
            "4/4 [==============================] - 0s 665us/step - loss: 0.0974 - acc: 0.7500\n",
            "Epoch 809/1000\n",
            "4/4 [==============================] - 0s 594us/step - loss: 0.0971 - acc: 0.7500\n",
            "Epoch 810/1000\n",
            "4/4 [==============================] - 0s 590us/step - loss: 0.0969 - acc: 0.7500\n",
            "Epoch 811/1000\n",
            "4/4 [==============================] - 0s 693us/step - loss: 0.0966 - acc: 0.7500\n",
            "Epoch 812/1000\n",
            "4/4 [==============================] - 0s 600us/step - loss: 0.0964 - acc: 0.7500\n",
            "Epoch 813/1000\n",
            "4/4 [==============================] - 0s 666us/step - loss: 0.0961 - acc: 0.7500\n",
            "Epoch 814/1000\n",
            "4/4 [==============================] - 0s 763us/step - loss: 0.0959 - acc: 0.7500\n",
            "Epoch 815/1000\n",
            "4/4 [==============================] - 0s 686us/step - loss: 0.0957 - acc: 0.7500\n",
            "Epoch 816/1000\n",
            "4/4 [==============================] - 0s 566us/step - loss: 0.0954 - acc: 0.7500\n",
            "Epoch 817/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.0952 - acc: 0.7500\n",
            "Epoch 818/1000\n",
            "4/4 [==============================] - 0s 836us/step - loss: 0.0949 - acc: 0.7500\n",
            "Epoch 819/1000\n",
            "4/4 [==============================] - 0s 904us/step - loss: 0.0947 - acc: 0.7500\n",
            "Epoch 820/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0945 - acc: 0.7500\n",
            "Epoch 821/1000\n",
            "4/4 [==============================] - 0s 720us/step - loss: 0.0942 - acc: 0.7500\n",
            "Epoch 822/1000\n",
            "4/4 [==============================] - 0s 688us/step - loss: 0.0940 - acc: 0.7500\n",
            "Epoch 823/1000\n",
            "4/4 [==============================] - 0s 693us/step - loss: 0.0938 - acc: 0.7500\n",
            "Epoch 824/1000\n",
            "4/4 [==============================] - 0s 689us/step - loss: 0.0936 - acc: 0.7500\n",
            "Epoch 825/1000\n",
            "4/4 [==============================] - 0s 583us/step - loss: 0.0934 - acc: 0.7500\n",
            "Epoch 826/1000\n",
            "4/4 [==============================] - 0s 630us/step - loss: 0.0931 - acc: 0.7500\n",
            "Epoch 827/1000\n",
            "4/4 [==============================] - 0s 726us/step - loss: 0.0929 - acc: 0.7500\n",
            "Epoch 828/1000\n",
            "4/4 [==============================] - 0s 593us/step - loss: 0.0927 - acc: 0.7500\n",
            "Epoch 829/1000\n",
            "4/4 [==============================] - 0s 679us/step - loss: 0.0925 - acc: 0.7500\n",
            "Epoch 830/1000\n",
            "4/4 [==============================] - 0s 604us/step - loss: 0.0923 - acc: 0.7500\n",
            "Epoch 831/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.0921 - acc: 0.7500\n",
            "Epoch 832/1000\n",
            "4/4 [==============================] - 0s 671us/step - loss: 0.0918 - acc: 0.7500\n",
            "Epoch 833/1000\n",
            "4/4 [==============================] - 0s 638us/step - loss: 0.0916 - acc: 0.7500\n",
            "Epoch 834/1000\n",
            "4/4 [==============================] - 0s 586us/step - loss: 0.0915 - acc: 0.7500\n",
            "Epoch 835/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.0913 - acc: 0.7500\n",
            "Epoch 836/1000\n",
            "4/4 [==============================] - 0s 633us/step - loss: 0.0910 - acc: 0.7500\n",
            "Epoch 837/1000\n",
            "4/4 [==============================] - 0s 570us/step - loss: 0.0908 - acc: 0.7500\n",
            "Epoch 838/1000\n",
            "4/4 [==============================] - 0s 700us/step - loss: 0.0907 - acc: 0.7500\n",
            "Epoch 839/1000\n",
            "4/4 [==============================] - 0s 836us/step - loss: 0.0905 - acc: 0.7500\n",
            "Epoch 840/1000\n",
            "4/4 [==============================] - 0s 698us/step - loss: 0.0903 - acc: 0.7500\n",
            "Epoch 841/1000\n",
            "4/4 [==============================] - 0s 648us/step - loss: 0.0901 - acc: 0.7500\n",
            "Epoch 842/1000\n",
            "4/4 [==============================] - 0s 540us/step - loss: 0.0899 - acc: 0.7500\n",
            "Epoch 843/1000\n",
            "4/4 [==============================] - 0s 757us/step - loss: 0.0897 - acc: 0.7500\n",
            "Epoch 844/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0895 - acc: 0.7500\n",
            "Epoch 845/1000\n",
            "4/4 [==============================] - 0s 814us/step - loss: 0.0893 - acc: 0.7500\n",
            "Epoch 846/1000\n",
            "4/4 [==============================] - 0s 624us/step - loss: 0.0892 - acc: 0.7500\n",
            "Epoch 847/1000\n",
            "4/4 [==============================] - 0s 932us/step - loss: 0.0890 - acc: 0.7500\n",
            "Epoch 848/1000\n",
            "4/4 [==============================] - 0s 426us/step - loss: 0.0888 - acc: 0.7500\n",
            "Epoch 849/1000\n",
            "4/4 [==============================] - 0s 369us/step - loss: 0.0886 - acc: 0.7500\n",
            "Epoch 850/1000\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0884 - acc: 0.7500\n",
            "Epoch 851/1000\n",
            "4/4 [==============================] - 0s 729us/step - loss: 0.0883 - acc: 0.7500\n",
            "Epoch 852/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.0881 - acc: 0.7500\n",
            "Epoch 853/1000\n",
            "4/4 [==============================] - 0s 565us/step - loss: 0.0879 - acc: 0.7500\n",
            "Epoch 854/1000\n",
            "4/4 [==============================] - 0s 527us/step - loss: 0.0878 - acc: 0.7500\n",
            "Epoch 855/1000\n",
            "4/4 [==============================] - 0s 631us/step - loss: 0.0876 - acc: 0.7500\n",
            "Epoch 856/1000\n",
            "4/4 [==============================] - 0s 516us/step - loss: 0.0875 - acc: 0.7500\n",
            "Epoch 857/1000\n",
            "4/4 [==============================] - 0s 537us/step - loss: 0.0873 - acc: 0.7500\n",
            "Epoch 858/1000\n",
            "4/4 [==============================] - 0s 462us/step - loss: 0.0871 - acc: 0.7500\n",
            "Epoch 859/1000\n",
            "4/4 [==============================] - 0s 448us/step - loss: 0.0870 - acc: 0.7500\n",
            "Epoch 860/1000\n",
            "4/4 [==============================] - 0s 506us/step - loss: 0.0868 - acc: 0.7500\n",
            "Epoch 861/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.0867 - acc: 0.7500\n",
            "Epoch 862/1000\n",
            "4/4 [==============================] - 0s 664us/step - loss: 0.0865 - acc: 0.7500\n",
            "Epoch 863/1000\n",
            "4/4 [==============================] - 0s 519us/step - loss: 0.0864 - acc: 0.7500\n",
            "Epoch 864/1000\n",
            "4/4 [==============================] - 0s 548us/step - loss: 0.0862 - acc: 0.7500\n",
            "Epoch 865/1000\n",
            "4/4 [==============================] - 0s 482us/step - loss: 0.0861 - acc: 0.7500\n",
            "Epoch 866/1000\n",
            "4/4 [==============================] - 0s 742us/step - loss: 0.0859 - acc: 0.7500\n",
            "Epoch 867/1000\n",
            "4/4 [==============================] - 0s 517us/step - loss: 0.0858 - acc: 0.7500\n",
            "Epoch 868/1000\n",
            "4/4 [==============================] - 0s 553us/step - loss: 0.0856 - acc: 0.7500\n",
            "Epoch 869/1000\n",
            "4/4 [==============================] - 0s 520us/step - loss: 0.0855 - acc: 0.7500\n",
            "Epoch 870/1000\n",
            "4/4 [==============================] - 0s 690us/step - loss: 0.0853 - acc: 0.7500\n",
            "Epoch 871/1000\n",
            "4/4 [==============================] - 0s 700us/step - loss: 0.0852 - acc: 0.7500\n",
            "Epoch 872/1000\n",
            "4/4 [==============================] - 0s 692us/step - loss: 0.0851 - acc: 0.7500\n",
            "Epoch 873/1000\n",
            "4/4 [==============================] - 0s 647us/step - loss: 0.0849 - acc: 0.7500\n",
            "Epoch 874/1000\n",
            "4/4 [==============================] - 0s 707us/step - loss: 0.0848 - acc: 0.7500\n",
            "Epoch 875/1000\n",
            "4/4 [==============================] - 0s 578us/step - loss: 0.0847 - acc: 0.7500\n",
            "Epoch 876/1000\n",
            "4/4 [==============================] - 0s 694us/step - loss: 0.0845 - acc: 0.7500\n",
            "Epoch 877/1000\n",
            "4/4 [==============================] - 0s 680us/step - loss: 0.0844 - acc: 0.7500\n",
            "Epoch 878/1000\n",
            "4/4 [==============================] - 0s 794us/step - loss: 0.0843 - acc: 0.7500\n",
            "Epoch 879/1000\n",
            "4/4 [==============================] - 0s 805us/step - loss: 0.0842 - acc: 0.7500\n",
            "Epoch 880/1000\n",
            "4/4 [==============================] - 0s 872us/step - loss: 0.0840 - acc: 0.7500\n",
            "Epoch 881/1000\n",
            "4/4 [==============================] - 0s 642us/step - loss: 0.0839 - acc: 0.7500\n",
            "Epoch 882/1000\n",
            "4/4 [==============================] - 0s 677us/step - loss: 0.0838 - acc: 0.7500\n",
            "Epoch 883/1000\n",
            "4/4 [==============================] - 0s 780us/step - loss: 0.0837 - acc: 0.7500\n",
            "Epoch 884/1000\n",
            "4/4 [==============================] - 0s 724us/step - loss: 0.0835 - acc: 0.7500\n",
            "Epoch 885/1000\n",
            "4/4 [==============================] - 0s 673us/step - loss: 0.0834 - acc: 0.7500\n",
            "Epoch 886/1000\n",
            "4/4 [==============================] - 0s 687us/step - loss: 0.0833 - acc: 1.0000\n",
            "Epoch 887/1000\n",
            "4/4 [==============================] - 0s 535us/step - loss: 0.0832 - acc: 1.0000\n",
            "Epoch 888/1000\n",
            "4/4 [==============================] - 0s 694us/step - loss: 0.0831 - acc: 1.0000\n",
            "Epoch 889/1000\n",
            "4/4 [==============================] - 0s 775us/step - loss: 0.0830 - acc: 1.0000\n",
            "Epoch 890/1000\n",
            "4/4 [==============================] - 0s 705us/step - loss: 0.0828 - acc: 1.0000\n",
            "Epoch 891/1000\n",
            "4/4 [==============================] - 0s 801us/step - loss: 0.0827 - acc: 1.0000\n",
            "Epoch 892/1000\n",
            "4/4 [==============================] - 0s 564us/step - loss: 0.0826 - acc: 1.0000\n",
            "Epoch 893/1000\n",
            "4/4 [==============================] - 0s 899us/step - loss: 0.0825 - acc: 1.0000\n",
            "Epoch 894/1000\n",
            "4/4 [==============================] - 0s 470us/step - loss: 0.0824 - acc: 1.0000\n",
            "Epoch 895/1000\n",
            "4/4 [==============================] - 0s 428us/step - loss: 0.0823 - acc: 1.0000\n",
            "Epoch 896/1000\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0822 - acc: 1.0000\n",
            "Epoch 897/1000\n",
            "4/4 [==============================] - 0s 672us/step - loss: 0.0821 - acc: 1.0000\n",
            "Epoch 898/1000\n",
            "4/4 [==============================] - 0s 620us/step - loss: 0.0820 - acc: 1.0000\n",
            "Epoch 899/1000\n",
            "4/4 [==============================] - 0s 658us/step - loss: 0.0819 - acc: 1.0000\n",
            "Epoch 900/1000\n",
            "4/4 [==============================] - 0s 811us/step - loss: 0.0818 - acc: 1.0000\n",
            "Epoch 901/1000\n",
            "4/4 [==============================] - 0s 500us/step - loss: 0.0817 - acc: 1.0000\n",
            "Epoch 902/1000\n",
            "4/4 [==============================] - 0s 869us/step - loss: 0.0816 - acc: 1.0000\n",
            "Epoch 903/1000\n",
            "4/4 [==============================] - 0s 582us/step - loss: 0.0815 - acc: 1.0000\n",
            "Epoch 904/1000\n",
            "4/4 [==============================] - 0s 634us/step - loss: 0.0814 - acc: 1.0000\n",
            "Epoch 905/1000\n",
            "4/4 [==============================] - 0s 581us/step - loss: 0.0813 - acc: 1.0000\n",
            "Epoch 906/1000\n",
            "4/4 [==============================] - 0s 646us/step - loss: 0.0812 - acc: 1.0000\n",
            "Epoch 907/1000\n",
            "4/4 [==============================] - 0s 669us/step - loss: 0.0811 - acc: 1.0000\n",
            "Epoch 908/1000\n",
            "4/4 [==============================] - 0s 646us/step - loss: 0.0810 - acc: 1.0000\n",
            "Epoch 909/1000\n",
            "4/4 [==============================] - 0s 531us/step - loss: 0.0809 - acc: 1.0000\n",
            "Epoch 910/1000\n",
            "4/4 [==============================] - 0s 601us/step - loss: 0.0808 - acc: 1.0000\n",
            "Epoch 911/1000\n",
            "4/4 [==============================] - 0s 588us/step - loss: 0.0808 - acc: 1.0000\n",
            "Epoch 912/1000\n",
            "4/4 [==============================] - 0s 802us/step - loss: 0.0807 - acc: 1.0000\n",
            "Epoch 913/1000\n",
            "4/4 [==============================] - 0s 641us/step - loss: 0.0806 - acc: 1.0000\n",
            "Epoch 914/1000\n",
            "4/4 [==============================] - 0s 621us/step - loss: 0.0805 - acc: 1.0000\n",
            "Epoch 915/1000\n",
            "4/4 [==============================] - 0s 533us/step - loss: 0.0804 - acc: 1.0000\n",
            "Epoch 916/1000\n",
            "4/4 [==============================] - 0s 722us/step - loss: 0.0803 - acc: 1.0000\n",
            "Epoch 917/1000\n",
            "4/4 [==============================] - 0s 749us/step - loss: 0.0802 - acc: 1.0000\n",
            "Epoch 918/1000\n",
            "4/4 [==============================] - 0s 637us/step - loss: 0.0802 - acc: 1.0000\n",
            "Epoch 919/1000\n",
            "4/4 [==============================] - 0s 634us/step - loss: 0.0801 - acc: 1.0000\n",
            "Epoch 920/1000\n",
            "4/4 [==============================] - 0s 926us/step - loss: 0.0800 - acc: 1.0000\n",
            "Epoch 921/1000\n",
            "4/4 [==============================] - 0s 717us/step - loss: 0.0799 - acc: 1.0000\n",
            "Epoch 922/1000\n",
            "4/4 [==============================] - 0s 786us/step - loss: 0.0798 - acc: 1.0000\n",
            "Epoch 923/1000\n",
            "4/4 [==============================] - 0s 947us/step - loss: 0.0798 - acc: 1.0000\n",
            "Epoch 924/1000\n",
            "4/4 [==============================] - 0s 605us/step - loss: 0.0797 - acc: 1.0000\n",
            "Epoch 925/1000\n",
            "4/4 [==============================] - 0s 550us/step - loss: 0.0796 - acc: 1.0000\n",
            "Epoch 926/1000\n",
            "4/4 [==============================] - 0s 626us/step - loss: 0.0795 - acc: 1.0000\n",
            "Epoch 927/1000\n",
            "4/4 [==============================] - 0s 991us/step - loss: 0.0795 - acc: 1.0000\n",
            "Epoch 928/1000\n",
            "4/4 [==============================] - 0s 668us/step - loss: 0.0794 - acc: 1.0000\n",
            "Epoch 929/1000\n",
            "4/4 [==============================] - 0s 889us/step - loss: 0.0793 - acc: 1.0000\n",
            "Epoch 930/1000\n",
            "4/4 [==============================] - 0s 825us/step - loss: 0.0792 - acc: 1.0000\n",
            "Epoch 931/1000\n",
            "4/4 [==============================] - 0s 522us/step - loss: 0.0792 - acc: 1.0000\n",
            "Epoch 932/1000\n",
            "4/4 [==============================] - 0s 545us/step - loss: 0.0791 - acc: 1.0000\n",
            "Epoch 933/1000\n",
            "4/4 [==============================] - 0s 581us/step - loss: 0.0790 - acc: 1.0000\n",
            "Epoch 934/1000\n",
            "4/4 [==============================] - 0s 610us/step - loss: 0.0790 - acc: 1.0000\n",
            "Epoch 935/1000\n",
            "4/4 [==============================] - 0s 612us/step - loss: 0.0789 - acc: 1.0000\n",
            "Epoch 936/1000\n",
            "4/4 [==============================] - 0s 571us/step - loss: 0.0788 - acc: 1.0000\n",
            "Epoch 937/1000\n",
            "4/4 [==============================] - 0s 662us/step - loss: 0.0788 - acc: 1.0000\n",
            "Epoch 938/1000\n",
            "4/4 [==============================] - 0s 610us/step - loss: 0.0787 - acc: 1.0000\n",
            "Epoch 939/1000\n",
            "4/4 [==============================] - 0s 829us/step - loss: 0.0786 - acc: 1.0000\n",
            "Epoch 940/1000\n",
            "4/4 [==============================] - 0s 511us/step - loss: 0.0786 - acc: 1.0000\n",
            "Epoch 941/1000\n",
            "4/4 [==============================] - 0s 925us/step - loss: 0.0785 - acc: 1.0000\n",
            "Epoch 942/1000\n",
            "4/4 [==============================] - 0s 559us/step - loss: 0.0784 - acc: 1.0000\n",
            "Epoch 943/1000\n",
            "4/4 [==============================] - 0s 518us/step - loss: 0.0784 - acc: 1.0000\n",
            "Epoch 944/1000\n",
            "4/4 [==============================] - 0s 599us/step - loss: 0.0783 - acc: 1.0000\n",
            "Epoch 945/1000\n",
            "4/4 [==============================] - 0s 567us/step - loss: 0.0782 - acc: 1.0000\n",
            "Epoch 946/1000\n",
            "4/4 [==============================] - 0s 609us/step - loss: 0.0782 - acc: 1.0000\n",
            "Epoch 947/1000\n",
            "4/4 [==============================] - 0s 603us/step - loss: 0.0781 - acc: 1.0000\n",
            "Epoch 948/1000\n",
            "4/4 [==============================] - 0s 783us/step - loss: 0.0781 - acc: 1.0000\n",
            "Epoch 949/1000\n",
            "4/4 [==============================] - 0s 816us/step - loss: 0.0780 - acc: 1.0000\n",
            "Epoch 950/1000\n",
            "4/4 [==============================] - 0s 612us/step - loss: 0.0779 - acc: 1.0000\n",
            "Epoch 951/1000\n",
            "4/4 [==============================] - 0s 802us/step - loss: 0.0779 - acc: 1.0000\n",
            "Epoch 952/1000\n",
            "4/4 [==============================] - 0s 842us/step - loss: 0.0778 - acc: 1.0000\n",
            "Epoch 953/1000\n",
            "4/4 [==============================] - 0s 565us/step - loss: 0.0778 - acc: 1.0000\n",
            "Epoch 954/1000\n",
            "4/4 [==============================] - 0s 613us/step - loss: 0.0777 - acc: 1.0000\n",
            "Epoch 955/1000\n",
            "4/4 [==============================] - 0s 528us/step - loss: 0.0776 - acc: 1.0000\n",
            "Epoch 956/1000\n",
            "4/4 [==============================] - 0s 606us/step - loss: 0.0776 - acc: 1.0000\n",
            "Epoch 957/1000\n",
            "4/4 [==============================] - 0s 553us/step - loss: 0.0775 - acc: 1.0000\n",
            "Epoch 958/1000\n",
            "4/4 [==============================] - 0s 495us/step - loss: 0.0775 - acc: 1.0000\n",
            "Epoch 959/1000\n",
            "4/4 [==============================] - 0s 529us/step - loss: 0.0774 - acc: 1.0000\n",
            "Epoch 960/1000\n",
            "4/4 [==============================] - 0s 625us/step - loss: 0.0774 - acc: 1.0000\n",
            "Epoch 961/1000\n",
            "4/4 [==============================] - 0s 544us/step - loss: 0.0773 - acc: 1.0000\n",
            "Epoch 962/1000\n",
            "4/4 [==============================] - 0s 757us/step - loss: 0.0773 - acc: 1.0000\n",
            "Epoch 963/1000\n",
            "4/4 [==============================] - 0s 607us/step - loss: 0.0772 - acc: 1.0000\n",
            "Epoch 964/1000\n",
            "4/4 [==============================] - 0s 767us/step - loss: 0.0772 - acc: 1.0000\n",
            "Epoch 965/1000\n",
            "4/4 [==============================] - 0s 631us/step - loss: 0.0771 - acc: 1.0000\n",
            "Epoch 966/1000\n",
            "4/4 [==============================] - 0s 571us/step - loss: 0.0771 - acc: 1.0000\n",
            "Epoch 967/1000\n",
            "4/4 [==============================] - 0s 567us/step - loss: 0.0770 - acc: 1.0000\n",
            "Epoch 968/1000\n",
            "4/4 [==============================] - 0s 690us/step - loss: 0.0769 - acc: 1.0000\n",
            "Epoch 969/1000\n",
            "4/4 [==============================] - 0s 536us/step - loss: 0.0769 - acc: 1.0000\n",
            "Epoch 970/1000\n",
            "4/4 [==============================] - 0s 563us/step - loss: 0.0768 - acc: 1.0000\n",
            "Epoch 971/1000\n",
            "4/4 [==============================] - 0s 732us/step - loss: 0.0768 - acc: 1.0000\n",
            "Epoch 972/1000\n",
            "4/4 [==============================] - 0s 867us/step - loss: 0.0767 - acc: 1.0000\n",
            "Epoch 973/1000\n",
            "4/4 [==============================] - 0s 802us/step - loss: 0.0767 - acc: 1.0000\n",
            "Epoch 974/1000\n",
            "4/4 [==============================] - 0s 583us/step - loss: 0.0766 - acc: 1.0000\n",
            "Epoch 975/1000\n",
            "4/4 [==============================] - 0s 638us/step - loss: 0.0766 - acc: 1.0000\n",
            "Epoch 976/1000\n",
            "4/4 [==============================] - 0s 529us/step - loss: 0.0766 - acc: 1.0000\n",
            "Epoch 977/1000\n",
            "4/4 [==============================] - 0s 607us/step - loss: 0.0765 - acc: 1.0000\n",
            "Epoch 978/1000\n",
            "4/4 [==============================] - 0s 997us/step - loss: 0.0765 - acc: 1.0000\n",
            "Epoch 979/1000\n",
            "4/4 [==============================] - 0s 847us/step - loss: 0.0764 - acc: 1.0000\n",
            "Epoch 980/1000\n",
            "4/4 [==============================] - 0s 884us/step - loss: 0.0764 - acc: 1.0000\n",
            "Epoch 981/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0763 - acc: 1.0000\n",
            "Epoch 982/1000\n",
            "4/4 [==============================] - 0s 813us/step - loss: 0.0763 - acc: 1.0000\n",
            "Epoch 983/1000\n",
            "4/4 [==============================] - 0s 755us/step - loss: 0.0762 - acc: 1.0000\n",
            "Epoch 984/1000\n",
            "4/4 [==============================] - 0s 799us/step - loss: 0.0762 - acc: 1.0000\n",
            "Epoch 985/1000\n",
            "4/4 [==============================] - 0s 754us/step - loss: 0.0761 - acc: 1.0000\n",
            "Epoch 986/1000\n",
            "4/4 [==============================] - 0s 404us/step - loss: 0.0761 - acc: 1.0000\n",
            "Epoch 987/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0760 - acc: 1.0000\n",
            "Epoch 988/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0760 - acc: 1.0000\n",
            "Epoch 989/1000\n",
            "4/4 [==============================] - 0s 815us/step - loss: 0.0760 - acc: 1.0000\n",
            "Epoch 990/1000\n",
            "4/4 [==============================] - 0s 813us/step - loss: 0.0759 - acc: 1.0000\n",
            "Epoch 991/1000\n",
            "4/4 [==============================] - 0s 982us/step - loss: 0.0759 - acc: 1.0000\n",
            "Epoch 992/1000\n",
            "4/4 [==============================] - 0s 742us/step - loss: 0.0758 - acc: 1.0000\n",
            "Epoch 993/1000\n",
            "4/4 [==============================] - 0s 789us/step - loss: 0.0758 - acc: 1.0000\n",
            "Epoch 994/1000\n",
            "4/4 [==============================] - 0s 793us/step - loss: 0.0758 - acc: 1.0000\n",
            "Epoch 995/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0757 - acc: 1.0000\n",
            "Epoch 996/1000\n",
            "4/4 [==============================] - 0s 759us/step - loss: 0.0757 - acc: 1.0000\n",
            "Epoch 997/1000\n",
            "4/4 [==============================] - 0s 781us/step - loss: 0.0756 - acc: 1.0000\n",
            "Epoch 998/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0756 - acc: 1.0000\n",
            "Epoch 999/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0755 - acc: 1.0000\n",
            "Epoch 1000/1000\n",
            "4/4 [==============================] - 0s 1ms/step - loss: 0.0755 - acc: 1.0000\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7fc6b21b8dd8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qP2w4Ms97l6B",
        "colab_type": "text"
      },
      "source": [
        "We then determine the output over training examples."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kKI3bDrR4qFv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "df49f209-f6c4-45dd-90ea-ad247eb0de1f"
      },
      "source": [
        "predictions = model.predict(X)\n",
        "print(predictions)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[-1.2274415 ]\n",
            " [-0.5307533 ]\n",
            " [ 0.8947075 ]\n",
            " [ 0.86262316]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "upKbiOAt7uDg",
        "colab_type": "text"
      },
      "source": [
        "We now visualize the output of the neural network."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1jjNolWL5rQ1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 319
        },
        "outputId": "9bed53d6-262f-4b6e-aee6-513496177bcd"
      },
      "source": [
        "plotit(X,y,clf=model.predict,conts=[0],extent = [-5,+5,-5,+5])"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:63: UserWarning: The following kwargs were not used by contour: 'label'\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAEMCAYAAAAidwoiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhmEYXYwZAcMwjC7GjIBhGEYX8/8DIKODN3arrVcAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[-5, 5, -5, 5]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Xfeb0NeH7zQc",
        "colab_type": "text"
      },
      "source": [
        "What you can do:\n",
        "Change the structure of the neural network (number of neurons, layers, activation functions, epochs, batch size, etc.) and see how it affects the final output."
      ]
    }
  ]
}