kap7-iris.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "colab": {
      "name": "kap7-iris.ipynb",
      "provenance": [],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/DJCordhose/buch-machine-learning-notebooks/blob/master/kap7-iris.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v4eRosoGpl8_",
        "colab_type": "text"
      },
      "source": [
        "# Kapitel 7: Neuronale Netzwerke - Grundlagen"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KYbHMTMjpl8_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "83Ttu0Pnpl9D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "85822c65-8551-41f3-9bf1-ea0dfc9291f6"
      },
      "source": [
        "%matplotlib inline\n",
        "%pylab inline"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Populating the interactive namespace from numpy and matplotlib\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rJ1DvBlApl9G",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pylab as plt\n",
        "import numpy as np"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g0rvxIp4pl9J",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "colors = 'bwr'#['b','y','r']\n",
        "CMAP = colors#plt.cm.rainbow"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n5AGdJTFpl9N",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0b1b66a4-ee5b-4b89-c789-5ca2e27f31f4"
      },
      "source": [
        "import sklearn\n",
        "print(sklearn.__version__)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.22.2.post1\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f5bwPx-ipl9P",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "49d8eea1-8c56-439a-f5bc-a82f32e6b6aa"
      },
      "source": [
        "import tensorflow as tf\n",
        "print(tf.__version__)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.3.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ssBE6lbypl9S",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "92738183-eed5-40d3-ccf6-f0d3e4e810af"
      },
      "source": [
        "import keras\n",
        "print(keras.__version__)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.4.3\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_9-yW3Skpl9V",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "19437069-f13b-4c88-f89b-e4635850f0e2"
      },
      "source": [
        "import pandas as pd\n",
        "print(pd.__version__)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1.0.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lKGPNR5Vpl9Y",
        "colab_type": "text"
      },
      "source": [
        "## Iris mit Neuronalen Netzwerken"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5h7LeQWFpl9Y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.datasets import load_iris\n",
        "iris = load_iris()"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FtNqwr6Gpl9a",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0caf111b-fe2f-4113-987b-ef3bdaf81934"
      },
      "source": [
        "print(iris.DESCR)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            ".. _iris_dataset:\n",
            "\n",
            "Iris plants dataset\n",
            "--------------------\n",
            "\n",
            "**Data Set Characteristics:**\n",
            "\n",
            "    :Number of Instances: 150 (50 in each of three classes)\n",
            "    :Number of Attributes: 4 numeric, predictive attributes and the class\n",
            "    :Attribute Information:\n",
            "        - sepal length in cm\n",
            "        - sepal width in cm\n",
            "        - petal length in cm\n",
            "        - petal width in cm\n",
            "        - class:\n",
            "                - Iris-Setosa\n",
            "                - Iris-Versicolour\n",
            "                - Iris-Virginica\n",
            "                \n",
            "    :Summary Statistics:\n",
            "\n",
            "    ============== ==== ==== ======= ===== ====================\n",
            "                    Min  Max   Mean    SD   Class Correlation\n",
            "    ============== ==== ==== ======= ===== ====================\n",
            "    sepal length:   4.3  7.9   5.84   0.83    0.7826\n",
            "    sepal width:    2.0  4.4   3.05   0.43   -0.4194\n",
            "    petal length:   1.0  6.9   3.76   1.76    0.9490  (high!)\n",
            "    petal width:    0.1  2.5   1.20   0.76    0.9565  (high!)\n",
            "    ============== ==== ==== ======= ===== ====================\n",
            "\n",
            "    :Missing Attribute Values: None\n",
            "    :Class Distribution: 33.3% for each of 3 classes.\n",
            "    :Creator: R.A. Fisher\n",
            "    :Donor: Michael Marshall (MARSHALL%PLU@io.arc.nasa.gov)\n",
            "    :Date: July, 1988\n",
            "\n",
            "The famous Iris database, first used by Sir R.A. Fisher. The dataset is taken\n",
            "from Fisher's paper. Note that it's the same as in R, but not as in the UCI\n",
            "Machine Learning Repository, which has two wrong data points.\n",
            "\n",
            "This is perhaps the best known database to be found in the\n",
            "pattern recognition literature.  Fisher's paper is a classic in the field and\n",
            "is referenced frequently to this day.  (See Duda & Hart, for example.)  The\n",
            "data set contains 3 classes of 50 instances each, where each class refers to a\n",
            "type of iris plant.  One class is linearly separable from the other 2; the\n",
            "latter are NOT linearly separable from each other.\n",
            "\n",
            ".. topic:: References\n",
            "\n",
            "   - Fisher, R.A. \"The use of multiple measurements in taxonomic problems\"\n",
            "     Annual Eugenics, 7, Part II, 179-188 (1936); also in \"Contributions to\n",
            "     Mathematical Statistics\" (John Wiley, NY, 1950).\n",
            "   - Duda, R.O., & Hart, P.E. (1973) Pattern Classification and Scene Analysis.\n",
            "     (Q327.D83) John Wiley & Sons.  ISBN 0-471-22361-1.  See page 218.\n",
            "   - Dasarathy, B.V. (1980) \"Nosing Around the Neighborhood: A New System\n",
            "     Structure and Classification Rule for Recognition in Partially Exposed\n",
            "     Environments\".  IEEE Transactions on Pattern Analysis and Machine\n",
            "     Intelligence, Vol. PAMI-2, No. 1, 67-71.\n",
            "   - Gates, G.W. (1972) \"The Reduced Nearest Neighbor Rule\".  IEEE Transactions\n",
            "     on Information Theory, May 1972, 431-433.\n",
            "   - See also: 1988 MLC Proceedings, 54-64.  Cheeseman et al\"s AUTOCLASS II\n",
            "     conceptual clustering system finds 3 classes in the data.\n",
            "   - Many, many more ...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LEmm7Weipl9d",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "9fe728e2-2ea0-486c-e520-a3e958b7b3d4"
      },
      "source": [
        "iris_df = pd.DataFrame(iris.data, columns=iris.feature_names)\n",
        "pd.plotting.scatter_matrix(iris_df, c=iris.target, cmap=CMAP, edgecolor='black', figsize=(20, 20))"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[<matplotlib.axes._subplots.AxesSubplot object at 0x7f4580aec588>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f4580aca828>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f4580a85a90>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f4580a38cf8>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f45809edf60>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45809aa208>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f458095f470>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45809916a0>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f4580991710>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45808f9ba8>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45808ade10>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f458086c0b8>],\n",
              "       [<matplotlib.axes._subplots.AxesSubplot object at 0x7f458081e320>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45807d2588>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45808057f0>,\n",
              "        <matplotlib.axes._subplots.AxesSubplot object at 0x7f45807b8a58>]],\n",
              "      dtype=object)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oa3iYStkpl9f",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# plt.savefig('ML_0701.png', bbox_inches='tight')"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "c-exG806pl9g",
        "colab_type": "text"
      },
      "source": [
        "## Das künstliche Neuron"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PrYfSutGpl9h",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "w0 = 3\n",
        "w1 = -4\n",
        "w2 = 2\n",
        "\n",
        "def neuron_no_activation(x1, x2):\n",
        "    sum = w0 + x1 * w1 + x2 * w2\n",
        "    return sum"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yhw0DMUWpl9j",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "dcc8b20c-be1b-43b6-c60c-2c130cfd3c5b"
      },
      "source": [
        "iris.data[0]"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([5.1, 3.5, 1.4, 0.2])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rVqDyYWKpl9l",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "4208fc66-5604-4799-e353-65e52fea9e98"
      },
      "source": [
        "neuron_no_activation(5.1, 3.5)"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "-10.399999999999999"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TpcxmzDspl9o",
        "colab_type": "text"
      },
      "source": [
        "### Activation Functions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wP80A7hkpl9o",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def centerAxis(uses_negative=False):\n",
        "    # http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.plot\n",
        "    ax = plt.gca()\n",
        "    ax.spines['left'].set_position('center')\n",
        "    if uses_negative:\n",
        "        ax.spines['bottom'].set_position('center')\n",
        "    ax.spines['right'].set_color('none')\n",
        "    ax.spines['top'].set_color('none')\n",
        "    ax.xaxis.set_ticks_position('bottom')\n",
        "    ax.yaxis.set_ticks_position('left')"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yNq7ZiU0pl9r",
        "colab_type": "text"
      },
      "source": [
        "#### Step Function: abrupter, nicht stetig differenzierbarer Übergang zwischen 0 und 1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZqUtjXQ1pl9r",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def np_step(X):\n",
        "    return 0.5 * (np.sign(X) + 1)"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BIBsOI4epl9t",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "f6676a2a-0eda-4b1f-b4ee-a9aab2298c2d"
      },
      "source": [
        "x = np.arange(-10,10,0.01)\n",
        "y = np_step(x)\n",
        "\n",
        "centerAxis()\n",
        "plt.plot(x, y, lw=3)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f4581ffe1d0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "trtAvxTmpl9w",
        "colab_type": "text"
      },
      "source": [
        "#### Sigmoid Function: Fließender Übergang zwischen 0 und 1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BN8Y7fXPpl9x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def np_sigmoid(X):\n",
        "    return 1 / (1 + np.exp(X * -1))"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HyC7OoVnpl9z",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "bc6f637c-c76a-4afd-89bc-2f27b6622948"
      },
      "source": [
        "x = np.arange(-10,10,0.01)\n",
        "y = np_sigmoid(x)\n",
        "\n",
        "centerAxis()\n",
        "plt.plot(x,y,lw=3)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f45806bf160>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WAsTI6Chpl91",
        "colab_type": "text"
      },
      "source": [
        "#### Tangens Hyperbolicus Function: Fließender Übergang zwischen -1 und 1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_WDP-00Lpl92",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "b9014b7e-a164-4e50-bec4-76d1a0071941"
      },
      "source": [
        "x = np.arange(-10,10,0.01)\n",
        "y = np.tanh(x)\n",
        "\n",
        "centerAxis()\n",
        "plt.plot(x,y,lw=3)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f457e7eef98>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JWZW6pkrpl95",
        "colab_type": "text"
      },
      "source": [
        "#### Relu: Einfach zu berechnen, setzt kompletten negativen Wertebereich auf 0"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6LTGgI99pl96",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def np_relu(x):\n",
        "    return np.maximum(0, x)"
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mqtXbFQvpl98",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "outputId": "b560795b-7df1-412f-a8b3-726a99006aff"
      },
      "source": [
        "x = np.arange(-10,10,0.01)\n",
        "y = np_relu(x)\n",
        "\n",
        "centerAxis()\n",
        "plt.plot(x,y,lw=3)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f457e79f630>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pDcTe_YHpl9-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# https://docs.python.org/3/library/math.html\n",
        "import math as math\n",
        "def sigmoid(x):\n",
        "    return 1 / (1 + math.exp(x * -1))\n",
        "\n",
        "w0 = 3\n",
        "w1 = -4\n",
        "w2 = 2\n",
        "\n",
        "def neuron(x1, x2):\n",
        "    sum = w0 + x1 * w1 + x2 * w2\n",
        "    return sigmoid(sum)"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8jQTVmi6pl-A",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e7ce10f8-c0a0-43c5-971c-6dd870ee5649"
      },
      "source": [
        "neuron(5.1, 3.5)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3.043155690056538e-05"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-Ci-Qwh3pl-C",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Version that takes as many values as you like\n",
        "weights_with_bias = np.array([3, -4, 2])\n",
        "def np_neuron(X):\n",
        "    inputs_with_1_for_bias = np.concatenate((np.array([1]), X))\n",
        "    return np_sigmoid(np.sum(inputs_with_1_for_bias*weights_with_bias));"
      ],
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_kIgFmAhpl-F",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "11c1ab58-7f3b-4ce0-818c-b4a5b069ffc8"
      },
      "source": [
        "np_neuron(np.array([5.1, 3.5]))"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "3.043155690056538e-05"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fcynLtuepl-H",
        "colab_type": "text"
      },
      "source": [
        "## Unser erste Neuronales Netz mit Keras"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uv0NsPfhpl-H",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.layers import Input\n",
        "inputs = Input(shape=(4, ))"
      ],
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8id5zq2Bpl-K",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.layers import Dense\n",
        "fc = Dense(3)(inputs)"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RiFhyYi3pl-L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.models import Model\n",
        "model = Model(inputs=inputs, outputs=fc)"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wKyUqPAupl-N",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        },
        "outputId": "29c0ce17-5d53-4174-cd6b-863383e59ae7"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"functional_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         [(None, 4)]               0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 3)                 15        \n",
            "=================================================================\n",
            "Total params: 15\n",
            "Trainable params: 15\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lhrAh8GApl-P",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss='categorical_crossentropy',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-WcVU5Rgpl-R",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "42dd925d-cc8c-4594-cffb-ed5843f9f511"
      },
      "source": [
        "model.predict(np.array([[ 5.1,  3.5,  1.4,  0.2]]))"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[-7.2692757,  6.4111967,  5.368948 ]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i1OwT7TWpl-T",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "inputs = Input(shape=(4, ))\n",
        "fc = Dense(3)(inputs)\n",
        "predictions = Dense(3, activation='softmax')(fc)\n",
        "model = Model(inputs=inputs, outputs=predictions)"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KTxUA8jOpl-X",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "outputId": "f404a16b-f8f0-417b-f768-87548a4328d1"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"functional_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_2 (InputLayer)         [(None, 4)]               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 3)                 15        \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 3)                 12        \n",
            "=================================================================\n",
            "Total params: 27\n",
            "Trainable params: 27\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mGszqEMDpl-a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss='categorical_crossentropy',\n",
        "              metrics=['accuracy'])"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MI_BdRKbpl-c",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "a1b7dcd4-05ac-46bc-b813-bc3ea1f2f1d4"
      },
      "source": [
        "model.predict(np.array([[ 5.1,  3.5,  1.4,  0.2]]))"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.00201356, 0.21839522, 0.7795912 ]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "collapsed": true,
        "id": "sdoL_Z7bpl-e",
        "colab_type": "text"
      },
      "source": [
        "# Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fQ1oRcd-pl-e",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e2597cc0-5451-48da-f469-42d0fe6884e3"
      },
      "source": [
        "X = np.array(iris.data)\n",
        "y = np.array(iris.target)\n",
        "X.shape, y.shape"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((150, 4), (150,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c4LEAqRlpl-h",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "db3f7986-2b35-45d7-b081-30475080d3a8"
      },
      "source": [
        "y[100]"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-2gMaXFGpl-m",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.utils.np_utils import to_categorical\n",
        "\n",
        "num_categories = 3\n",
        "\n",
        "y = to_categorical(y, num_categories)"
      ],
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_GdCZmogpl-o",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9b966b62-bff3-4274-bb14-12b4dab702e7"
      },
      "source": [
        "y[100]"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0., 0., 1.], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TJL9v7tQpl-r",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=42, stratify=y)"
      ],
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2W10jFK_pl-t",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "cf0f8104-5b4c-410d-fe9b-bfafde834311"
      },
      "source": [
        "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((120, 4), (30, 4), (120, 3), (30, 3))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JNNf0b4kpl-w",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# !rm -r tf_log\n",
        "\n",
        "# https://keras.io/callbacks/#tensorboard\n",
        "# tb_callback = keras.callbacks.TensorBoard(log_dir='./tf_log')\n",
        "# To start tensorboard\n",
        "# tensorboard --logdir=/mnt/c/Users/olive/Development/ml/tf_log\n",
        "# open http://localhost:6006"
      ],
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Px5unIxQpl-0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "aeaeb4c7-e800-4fa2-b98b-3bcaa0881dd8"
      },
      "source": [
        "# %time model.fit(X_train, y_train, epochs=500, validation_split=0.3, callbacks=[tb_callback])\n",
        "%time model.fit(X_train, y_train, epochs=500, validation_split=0.3)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/500\n",
            "3/3 [==============================] - 0s 59ms/step - loss: 3.2124 - accuracy: 0.3214 - val_loss: 2.9324 - val_accuracy: 0.3611\n",
            "Epoch 2/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 3.1541 - accuracy: 0.3214 - val_loss: 2.8819 - val_accuracy: 0.3611\n",
            "Epoch 3/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 3.1031 - accuracy: 0.3214 - val_loss: 2.8316 - val_accuracy: 0.3611\n",
            "Epoch 4/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 3.0458 - accuracy: 0.3214 - val_loss: 2.7831 - val_accuracy: 0.3611\n",
            "Epoch 5/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.9976 - accuracy: 0.3214 - val_loss: 2.7348 - val_accuracy: 0.3611\n",
            "Epoch 6/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.9466 - accuracy: 0.3214 - val_loss: 2.6886 - val_accuracy: 0.3611\n",
            "Epoch 7/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.8950 - accuracy: 0.3214 - val_loss: 2.6438 - val_accuracy: 0.3611\n",
            "Epoch 8/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.8478 - accuracy: 0.3214 - val_loss: 2.5997 - val_accuracy: 0.3611\n",
            "Epoch 9/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 2.7974 - accuracy: 0.3214 - val_loss: 2.5573 - val_accuracy: 0.3611\n",
            "Epoch 10/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.7565 - accuracy: 0.3214 - val_loss: 2.5152 - val_accuracy: 0.3611\n",
            "Epoch 11/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.7085 - accuracy: 0.3214 - val_loss: 2.4752 - val_accuracy: 0.3611\n",
            "Epoch 12/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.6647 - accuracy: 0.3214 - val_loss: 2.4362 - val_accuracy: 0.3611\n",
            "Epoch 13/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.6232 - accuracy: 0.3214 - val_loss: 2.3980 - val_accuracy: 0.3611\n",
            "Epoch 14/500\n",
            "3/3 [==============================] - 0s 15ms/step - loss: 2.5801 - accuracy: 0.3214 - val_loss: 2.3611 - val_accuracy: 0.3611\n",
            "Epoch 15/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.5408 - accuracy: 0.3214 - val_loss: 2.3247 - val_accuracy: 0.3611\n",
            "Epoch 16/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 2.5005 - accuracy: 0.3214 - val_loss: 2.2893 - val_accuracy: 0.3611\n",
            "Epoch 17/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 2.4608 - accuracy: 0.3214 - val_loss: 2.2545 - val_accuracy: 0.3611\n",
            "Epoch 18/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.4225 - accuracy: 0.3214 - val_loss: 2.2201 - val_accuracy: 0.3611\n",
            "Epoch 19/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.3822 - accuracy: 0.3214 - val_loss: 2.1865 - val_accuracy: 0.3333\n",
            "Epoch 20/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.3459 - accuracy: 0.3214 - val_loss: 2.1528 - val_accuracy: 0.3333\n",
            "Epoch 21/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.3068 - accuracy: 0.3214 - val_loss: 2.1196 - val_accuracy: 0.3333\n",
            "Epoch 22/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 2.2709 - accuracy: 0.3214 - val_loss: 2.0862 - val_accuracy: 0.3333\n",
            "Epoch 23/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.2320 - accuracy: 0.3214 - val_loss: 2.0534 - val_accuracy: 0.3333\n",
            "Epoch 24/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.1956 - accuracy: 0.3214 - val_loss: 2.0208 - val_accuracy: 0.3333\n",
            "Epoch 25/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.1580 - accuracy: 0.3214 - val_loss: 1.9885 - val_accuracy: 0.3333\n",
            "Epoch 26/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.1225 - accuracy: 0.3095 - val_loss: 1.9562 - val_accuracy: 0.3333\n",
            "Epoch 27/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.0861 - accuracy: 0.3095 - val_loss: 1.9243 - val_accuracy: 0.3333\n",
            "Epoch 28/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.0523 - accuracy: 0.3095 - val_loss: 1.8924 - val_accuracy: 0.3333\n",
            "Epoch 29/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 2.0137 - accuracy: 0.3095 - val_loss: 1.8616 - val_accuracy: 0.3333\n",
            "Epoch 30/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.9787 - accuracy: 0.3095 - val_loss: 1.8307 - val_accuracy: 0.3333\n",
            "Epoch 31/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.9422 - accuracy: 0.3095 - val_loss: 1.8002 - val_accuracy: 0.3333\n",
            "Epoch 32/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.9073 - accuracy: 0.3095 - val_loss: 1.7696 - val_accuracy: 0.3333\n",
            "Epoch 33/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.8721 - accuracy: 0.3095 - val_loss: 1.7391 - val_accuracy: 0.3333\n",
            "Epoch 34/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 1.8421 - accuracy: 0.3095 - val_loss: 1.7079 - val_accuracy: 0.3333\n",
            "Epoch 35/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.8037 - accuracy: 0.3095 - val_loss: 1.6780 - val_accuracy: 0.3333\n",
            "Epoch 36/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.7703 - accuracy: 0.3095 - val_loss: 1.6484 - val_accuracy: 0.3333\n",
            "Epoch 37/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.7355 - accuracy: 0.3095 - val_loss: 1.6195 - val_accuracy: 0.3333\n",
            "Epoch 38/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.7042 - accuracy: 0.3095 - val_loss: 1.5904 - val_accuracy: 0.3333\n",
            "Epoch 39/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.6696 - accuracy: 0.3095 - val_loss: 1.5623 - val_accuracy: 0.3333\n",
            "Epoch 40/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 1.6364 - accuracy: 0.3095 - val_loss: 1.5346 - val_accuracy: 0.3333\n",
            "Epoch 41/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.6066 - accuracy: 0.3095 - val_loss: 1.5068 - val_accuracy: 0.3333\n",
            "Epoch 42/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.5738 - accuracy: 0.3095 - val_loss: 1.4802 - val_accuracy: 0.3333\n",
            "Epoch 43/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.5420 - accuracy: 0.3095 - val_loss: 1.4542 - val_accuracy: 0.3333\n",
            "Epoch 44/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.5137 - accuracy: 0.2976 - val_loss: 1.4283 - val_accuracy: 0.3333\n",
            "Epoch 45/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.4825 - accuracy: 0.2976 - val_loss: 1.4035 - val_accuracy: 0.3333\n",
            "Epoch 46/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.4573 - accuracy: 0.2976 - val_loss: 1.3786 - val_accuracy: 0.3333\n",
            "Epoch 47/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.4266 - accuracy: 0.2976 - val_loss: 1.3554 - val_accuracy: 0.3333\n",
            "Epoch 48/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.3984 - accuracy: 0.2976 - val_loss: 1.3334 - val_accuracy: 0.3333\n",
            "Epoch 49/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 1.3732 - accuracy: 0.2976 - val_loss: 1.3118 - val_accuracy: 0.3333\n",
            "Epoch 50/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 1.3482 - accuracy: 0.2976 - val_loss: 1.2911 - val_accuracy: 0.3333\n",
            "Epoch 51/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.3253 - accuracy: 0.2976 - val_loss: 1.2714 - val_accuracy: 0.3333\n",
            "Epoch 52/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.3017 - accuracy: 0.2976 - val_loss: 1.2530 - val_accuracy: 0.3333\n",
            "Epoch 53/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 1.2792 - accuracy: 0.2976 - val_loss: 1.2358 - val_accuracy: 0.3333\n",
            "Epoch 54/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.2591 - accuracy: 0.2976 - val_loss: 1.2196 - val_accuracy: 0.3333\n",
            "Epoch 55/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.2384 - accuracy: 0.2857 - val_loss: 1.2044 - val_accuracy: 0.3333\n",
            "Epoch 56/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.2193 - accuracy: 0.2857 - val_loss: 1.1900 - val_accuracy: 0.3333\n",
            "Epoch 57/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.2026 - accuracy: 0.2857 - val_loss: 1.1762 - val_accuracy: 0.3333\n",
            "Epoch 58/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.1849 - accuracy: 0.2857 - val_loss: 1.1634 - val_accuracy: 0.2778\n",
            "Epoch 59/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.1696 - accuracy: 0.2857 - val_loss: 1.1513 - val_accuracy: 0.2778\n",
            "Epoch 60/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.1539 - accuracy: 0.2738 - val_loss: 1.1401 - val_accuracy: 0.2778\n",
            "Epoch 61/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.1395 - accuracy: 0.2738 - val_loss: 1.1295 - val_accuracy: 0.2778\n",
            "Epoch 62/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.1264 - accuracy: 0.2738 - val_loss: 1.1194 - val_accuracy: 0.2500\n",
            "Epoch 63/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.1134 - accuracy: 0.2738 - val_loss: 1.1099 - val_accuracy: 0.2500\n",
            "Epoch 64/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.1013 - accuracy: 0.2738 - val_loss: 1.1007 - val_accuracy: 0.2222\n",
            "Epoch 65/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 1.0911 - accuracy: 0.2619 - val_loss: 1.0920 - val_accuracy: 0.2222\n",
            "Epoch 66/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 1.0796 - accuracy: 0.2619 - val_loss: 1.0839 - val_accuracy: 0.2222\n",
            "Epoch 67/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0689 - accuracy: 0.2500 - val_loss: 1.0761 - val_accuracy: 0.2222\n",
            "Epoch 68/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 1.0582 - accuracy: 0.2500 - val_loss: 1.0686 - val_accuracy: 0.1944\n",
            "Epoch 69/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0484 - accuracy: 0.2500 - val_loss: 1.0613 - val_accuracy: 0.1389\n",
            "Epoch 70/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0405 - accuracy: 0.2381 - val_loss: 1.0541 - val_accuracy: 0.1389\n",
            "Epoch 71/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 1.0311 - accuracy: 0.2262 - val_loss: 1.0471 - val_accuracy: 0.1389\n",
            "Epoch 72/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0223 - accuracy: 0.2262 - val_loss: 1.0404 - val_accuracy: 0.1111\n",
            "Epoch 73/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0144 - accuracy: 0.2262 - val_loss: 1.0337 - val_accuracy: 0.1111\n",
            "Epoch 74/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 1.0056 - accuracy: 0.2143 - val_loss: 1.0272 - val_accuracy: 0.1111\n",
            "Epoch 75/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.9983 - accuracy: 0.1905 - val_loss: 1.0209 - val_accuracy: 0.1111\n",
            "Epoch 76/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9902 - accuracy: 0.1905 - val_loss: 1.0146 - val_accuracy: 0.1111\n",
            "Epoch 77/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.9830 - accuracy: 0.1786 - val_loss: 1.0084 - val_accuracy: 0.1111\n",
            "Epoch 78/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.9756 - accuracy: 0.1548 - val_loss: 1.0023 - val_accuracy: 0.0833\n",
            "Epoch 79/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.9681 - accuracy: 0.1310 - val_loss: 0.9963 - val_accuracy: 0.1111\n",
            "Epoch 80/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.9619 - accuracy: 0.1310 - val_loss: 0.9904 - val_accuracy: 0.1389\n",
            "Epoch 81/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9545 - accuracy: 0.1429 - val_loss: 0.9845 - val_accuracy: 0.1667\n",
            "Epoch 82/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9479 - accuracy: 0.1548 - val_loss: 0.9787 - val_accuracy: 0.1667\n",
            "Epoch 83/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9413 - accuracy: 0.1905 - val_loss: 0.9729 - val_accuracy: 0.1944\n",
            "Epoch 84/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9350 - accuracy: 0.2262 - val_loss: 0.9674 - val_accuracy: 0.1944\n",
            "Epoch 85/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9285 - accuracy: 0.2500 - val_loss: 0.9619 - val_accuracy: 0.2500\n",
            "Epoch 86/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9220 - accuracy: 0.2619 - val_loss: 0.9564 - val_accuracy: 0.2500\n",
            "Epoch 87/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9160 - accuracy: 0.2857 - val_loss: 0.9509 - val_accuracy: 0.3333\n",
            "Epoch 88/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.9104 - accuracy: 0.3095 - val_loss: 0.9456 - val_accuracy: 0.3333\n",
            "Epoch 89/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.9040 - accuracy: 0.3333 - val_loss: 0.9403 - val_accuracy: 0.3333\n",
            "Epoch 90/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8981 - accuracy: 0.3690 - val_loss: 0.9351 - val_accuracy: 0.3333\n",
            "Epoch 91/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8924 - accuracy: 0.3810 - val_loss: 0.9298 - val_accuracy: 0.3889\n",
            "Epoch 92/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8868 - accuracy: 0.3929 - val_loss: 0.9247 - val_accuracy: 0.3889\n",
            "Epoch 93/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8811 - accuracy: 0.4048 - val_loss: 0.9197 - val_accuracy: 0.3889\n",
            "Epoch 94/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8755 - accuracy: 0.4167 - val_loss: 0.9148 - val_accuracy: 0.3889\n",
            "Epoch 95/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.8701 - accuracy: 0.4286 - val_loss: 0.9099 - val_accuracy: 0.4167\n",
            "Epoch 96/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8646 - accuracy: 0.4524 - val_loss: 0.9051 - val_accuracy: 0.4167\n",
            "Epoch 97/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8595 - accuracy: 0.4643 - val_loss: 0.9004 - val_accuracy: 0.4167\n",
            "Epoch 98/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8541 - accuracy: 0.4643 - val_loss: 0.8957 - val_accuracy: 0.4444\n",
            "Epoch 99/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8491 - accuracy: 0.4643 - val_loss: 0.8911 - val_accuracy: 0.4444\n",
            "Epoch 100/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8440 - accuracy: 0.4643 - val_loss: 0.8866 - val_accuracy: 0.4444\n",
            "Epoch 101/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8390 - accuracy: 0.4762 - val_loss: 0.8822 - val_accuracy: 0.4444\n",
            "Epoch 102/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8341 - accuracy: 0.4762 - val_loss: 0.8778 - val_accuracy: 0.4444\n",
            "Epoch 103/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8291 - accuracy: 0.4762 - val_loss: 0.8734 - val_accuracy: 0.4444\n",
            "Epoch 104/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8243 - accuracy: 0.4881 - val_loss: 0.8692 - val_accuracy: 0.4444\n",
            "Epoch 105/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8199 - accuracy: 0.4762 - val_loss: 0.8651 - val_accuracy: 0.4722\n",
            "Epoch 106/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.8150 - accuracy: 0.5000 - val_loss: 0.8609 - val_accuracy: 0.4722\n",
            "Epoch 107/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.8104 - accuracy: 0.5119 - val_loss: 0.8568 - val_accuracy: 0.4722\n",
            "Epoch 108/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.8059 - accuracy: 0.5119 - val_loss: 0.8526 - val_accuracy: 0.4722\n",
            "Epoch 109/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.8014 - accuracy: 0.5119 - val_loss: 0.8485 - val_accuracy: 0.4722\n",
            "Epoch 110/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.7973 - accuracy: 0.5238 - val_loss: 0.8444 - val_accuracy: 0.4722\n",
            "Epoch 111/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7931 - accuracy: 0.5238 - val_loss: 0.8405 - val_accuracy: 0.4722\n",
            "Epoch 112/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.7886 - accuracy: 0.5238 - val_loss: 0.8365 - val_accuracy: 0.4722\n",
            "Epoch 113/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7850 - accuracy: 0.5238 - val_loss: 0.8326 - val_accuracy: 0.4722\n",
            "Epoch 114/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.7806 - accuracy: 0.5238 - val_loss: 0.8289 - val_accuracy: 0.4722\n",
            "Epoch 115/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7764 - accuracy: 0.5357 - val_loss: 0.8249 - val_accuracy: 0.4722\n",
            "Epoch 116/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7725 - accuracy: 0.5357 - val_loss: 0.8212 - val_accuracy: 0.4722\n",
            "Epoch 117/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7685 - accuracy: 0.5476 - val_loss: 0.8175 - val_accuracy: 0.4722\n",
            "Epoch 118/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7651 - accuracy: 0.5595 - val_loss: 0.8140 - val_accuracy: 0.4444\n",
            "Epoch 119/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7611 - accuracy: 0.5595 - val_loss: 0.8103 - val_accuracy: 0.4722\n",
            "Epoch 120/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7576 - accuracy: 0.5595 - val_loss: 0.8065 - val_accuracy: 0.4722\n",
            "Epoch 121/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7536 - accuracy: 0.5595 - val_loss: 0.8030 - val_accuracy: 0.4722\n",
            "Epoch 122/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7499 - accuracy: 0.5595 - val_loss: 0.7996 - val_accuracy: 0.4444\n",
            "Epoch 123/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7463 - accuracy: 0.5714 - val_loss: 0.7962 - val_accuracy: 0.4444\n",
            "Epoch 124/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7428 - accuracy: 0.5714 - val_loss: 0.7928 - val_accuracy: 0.4444\n",
            "Epoch 125/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7393 - accuracy: 0.5714 - val_loss: 0.7894 - val_accuracy: 0.4444\n",
            "Epoch 126/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7358 - accuracy: 0.5833 - val_loss: 0.7861 - val_accuracy: 0.4444\n",
            "Epoch 127/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7325 - accuracy: 0.5833 - val_loss: 0.7826 - val_accuracy: 0.4722\n",
            "Epoch 128/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7292 - accuracy: 0.5833 - val_loss: 0.7794 - val_accuracy: 0.4722\n",
            "Epoch 129/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7257 - accuracy: 0.5952 - val_loss: 0.7762 - val_accuracy: 0.4722\n",
            "Epoch 130/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7227 - accuracy: 0.6071 - val_loss: 0.7732 - val_accuracy: 0.4444\n",
            "Epoch 131/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7192 - accuracy: 0.6071 - val_loss: 0.7702 - val_accuracy: 0.4722\n",
            "Epoch 132/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7161 - accuracy: 0.6071 - val_loss: 0.7672 - val_accuracy: 0.4722\n",
            "Epoch 133/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.7130 - accuracy: 0.5952 - val_loss: 0.7642 - val_accuracy: 0.4722\n",
            "Epoch 134/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7098 - accuracy: 0.6071 - val_loss: 0.7611 - val_accuracy: 0.4722\n",
            "Epoch 135/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7067 - accuracy: 0.6071 - val_loss: 0.7581 - val_accuracy: 0.4722\n",
            "Epoch 136/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.7037 - accuracy: 0.6071 - val_loss: 0.7550 - val_accuracy: 0.4722\n",
            "Epoch 137/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.7007 - accuracy: 0.6071 - val_loss: 0.7521 - val_accuracy: 0.4722\n",
            "Epoch 138/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6977 - accuracy: 0.6071 - val_loss: 0.7491 - val_accuracy: 0.4722\n",
            "Epoch 139/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6948 - accuracy: 0.6071 - val_loss: 0.7462 - val_accuracy: 0.4722\n",
            "Epoch 140/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6920 - accuracy: 0.6071 - val_loss: 0.7434 - val_accuracy: 0.4722\n",
            "Epoch 141/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6893 - accuracy: 0.6071 - val_loss: 0.7409 - val_accuracy: 0.4722\n",
            "Epoch 142/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.6866 - accuracy: 0.6190 - val_loss: 0.7383 - val_accuracy: 0.4722\n",
            "Epoch 143/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6835 - accuracy: 0.6190 - val_loss: 0.7355 - val_accuracy: 0.4722\n",
            "Epoch 144/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6807 - accuracy: 0.6190 - val_loss: 0.7328 - val_accuracy: 0.4722\n",
            "Epoch 145/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6779 - accuracy: 0.6310 - val_loss: 0.7301 - val_accuracy: 0.4722\n",
            "Epoch 146/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6752 - accuracy: 0.6548 - val_loss: 0.7276 - val_accuracy: 0.5000\n",
            "Epoch 147/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6726 - accuracy: 0.6667 - val_loss: 0.7250 - val_accuracy: 0.5278\n",
            "Epoch 148/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6701 - accuracy: 0.6667 - val_loss: 0.7224 - val_accuracy: 0.5278\n",
            "Epoch 149/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6675 - accuracy: 0.6667 - val_loss: 0.7200 - val_accuracy: 0.5556\n",
            "Epoch 150/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6650 - accuracy: 0.6786 - val_loss: 0.7175 - val_accuracy: 0.5833\n",
            "Epoch 151/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.6622 - accuracy: 0.6786 - val_loss: 0.7148 - val_accuracy: 0.5833\n",
            "Epoch 152/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6597 - accuracy: 0.6786 - val_loss: 0.7122 - val_accuracy: 0.5833\n",
            "Epoch 153/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6572 - accuracy: 0.6786 - val_loss: 0.7096 - val_accuracy: 0.5556\n",
            "Epoch 154/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6549 - accuracy: 0.6667 - val_loss: 0.7070 - val_accuracy: 0.5278\n",
            "Epoch 155/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6524 - accuracy: 0.6667 - val_loss: 0.7046 - val_accuracy: 0.5278\n",
            "Epoch 156/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6502 - accuracy: 0.6786 - val_loss: 0.7023 - val_accuracy: 0.6111\n",
            "Epoch 157/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6476 - accuracy: 0.6786 - val_loss: 0.6998 - val_accuracy: 0.6111\n",
            "Epoch 158/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6453 - accuracy: 0.6786 - val_loss: 0.6974 - val_accuracy: 0.6111\n",
            "Epoch 159/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6429 - accuracy: 0.6786 - val_loss: 0.6950 - val_accuracy: 0.6111\n",
            "Epoch 160/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6408 - accuracy: 0.6786 - val_loss: 0.6927 - val_accuracy: 0.6111\n",
            "Epoch 161/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6384 - accuracy: 0.6786 - val_loss: 0.6906 - val_accuracy: 0.6111\n",
            "Epoch 162/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6360 - accuracy: 0.6786 - val_loss: 0.6884 - val_accuracy: 0.6111\n",
            "Epoch 163/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6340 - accuracy: 0.6786 - val_loss: 0.6864 - val_accuracy: 0.6389\n",
            "Epoch 164/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6318 - accuracy: 0.6786 - val_loss: 0.6841 - val_accuracy: 0.6389\n",
            "Epoch 165/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6298 - accuracy: 0.6667 - val_loss: 0.6821 - val_accuracy: 0.6389\n",
            "Epoch 166/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6274 - accuracy: 0.6667 - val_loss: 0.6798 - val_accuracy: 0.6389\n",
            "Epoch 167/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.6253 - accuracy: 0.6667 - val_loss: 0.6776 - val_accuracy: 0.6389\n",
            "Epoch 168/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6233 - accuracy: 0.6667 - val_loss: 0.6753 - val_accuracy: 0.6389\n",
            "Epoch 169/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6211 - accuracy: 0.6667 - val_loss: 0.6733 - val_accuracy: 0.6389\n",
            "Epoch 170/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6191 - accuracy: 0.6667 - val_loss: 0.6711 - val_accuracy: 0.6389\n",
            "Epoch 171/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6172 - accuracy: 0.6667 - val_loss: 0.6691 - val_accuracy: 0.6389\n",
            "Epoch 172/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6151 - accuracy: 0.6786 - val_loss: 0.6670 - val_accuracy: 0.6389\n",
            "Epoch 173/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6131 - accuracy: 0.6786 - val_loss: 0.6649 - val_accuracy: 0.6389\n",
            "Epoch 174/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6110 - accuracy: 0.6786 - val_loss: 0.6625 - val_accuracy: 0.6389\n",
            "Epoch 175/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6090 - accuracy: 0.6667 - val_loss: 0.6603 - val_accuracy: 0.6389\n",
            "Epoch 176/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6071 - accuracy: 0.6667 - val_loss: 0.6581 - val_accuracy: 0.6389\n",
            "Epoch 177/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6056 - accuracy: 0.6667 - val_loss: 0.6558 - val_accuracy: 0.6389\n",
            "Epoch 178/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.6035 - accuracy: 0.6667 - val_loss: 0.6538 - val_accuracy: 0.6389\n",
            "Epoch 179/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.6016 - accuracy: 0.6667 - val_loss: 0.6518 - val_accuracy: 0.6389\n",
            "Epoch 180/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5997 - accuracy: 0.6667 - val_loss: 0.6498 - val_accuracy: 0.6389\n",
            "Epoch 181/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5979 - accuracy: 0.6667 - val_loss: 0.6478 - val_accuracy: 0.6389\n",
            "Epoch 182/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5962 - accuracy: 0.6667 - val_loss: 0.6458 - val_accuracy: 0.6389\n",
            "Epoch 183/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5942 - accuracy: 0.6786 - val_loss: 0.6439 - val_accuracy: 0.6389\n",
            "Epoch 184/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5925 - accuracy: 0.6786 - val_loss: 0.6421 - val_accuracy: 0.6389\n",
            "Epoch 185/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5907 - accuracy: 0.6786 - val_loss: 0.6403 - val_accuracy: 0.6389\n",
            "Epoch 186/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5889 - accuracy: 0.6786 - val_loss: 0.6385 - val_accuracy: 0.6389\n",
            "Epoch 187/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5872 - accuracy: 0.6667 - val_loss: 0.6367 - val_accuracy: 0.6389\n",
            "Epoch 188/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5855 - accuracy: 0.6667 - val_loss: 0.6350 - val_accuracy: 0.6389\n",
            "Epoch 189/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5838 - accuracy: 0.6667 - val_loss: 0.6334 - val_accuracy: 0.6667\n",
            "Epoch 190/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5820 - accuracy: 0.6786 - val_loss: 0.6317 - val_accuracy: 0.6667\n",
            "Epoch 191/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5808 - accuracy: 0.6786 - val_loss: 0.6303 - val_accuracy: 0.6667\n",
            "Epoch 192/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5788 - accuracy: 0.6786 - val_loss: 0.6285 - val_accuracy: 0.6667\n",
            "Epoch 193/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5771 - accuracy: 0.6786 - val_loss: 0.6267 - val_accuracy: 0.6944\n",
            "Epoch 194/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5757 - accuracy: 0.6786 - val_loss: 0.6248 - val_accuracy: 0.6944\n",
            "Epoch 195/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5739 - accuracy: 0.6786 - val_loss: 0.6232 - val_accuracy: 0.6944\n",
            "Epoch 196/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5722 - accuracy: 0.6667 - val_loss: 0.6216 - val_accuracy: 0.6944\n",
            "Epoch 197/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5709 - accuracy: 0.6667 - val_loss: 0.6201 - val_accuracy: 0.6944\n",
            "Epoch 198/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5693 - accuracy: 0.6667 - val_loss: 0.6183 - val_accuracy: 0.6944\n",
            "Epoch 199/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5677 - accuracy: 0.6667 - val_loss: 0.6167 - val_accuracy: 0.6944\n",
            "Epoch 200/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5663 - accuracy: 0.6667 - val_loss: 0.6150 - val_accuracy: 0.6944\n",
            "Epoch 201/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5646 - accuracy: 0.6667 - val_loss: 0.6133 - val_accuracy: 0.6944\n",
            "Epoch 202/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5631 - accuracy: 0.6667 - val_loss: 0.6115 - val_accuracy: 0.6944\n",
            "Epoch 203/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5616 - accuracy: 0.6667 - val_loss: 0.6098 - val_accuracy: 0.6944\n",
            "Epoch 204/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5601 - accuracy: 0.6667 - val_loss: 0.6080 - val_accuracy: 0.6944\n",
            "Epoch 205/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5585 - accuracy: 0.6667 - val_loss: 0.6062 - val_accuracy: 0.6944\n",
            "Epoch 206/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5572 - accuracy: 0.6786 - val_loss: 0.6043 - val_accuracy: 0.6944\n",
            "Epoch 207/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5557 - accuracy: 0.6905 - val_loss: 0.6025 - val_accuracy: 0.6944\n",
            "Epoch 208/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5544 - accuracy: 0.7381 - val_loss: 0.6008 - val_accuracy: 0.7222\n",
            "Epoch 209/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5530 - accuracy: 0.7381 - val_loss: 0.5991 - val_accuracy: 0.6944\n",
            "Epoch 210/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5517 - accuracy: 0.7262 - val_loss: 0.5975 - val_accuracy: 0.6944\n",
            "Epoch 211/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5504 - accuracy: 0.7262 - val_loss: 0.5961 - val_accuracy: 0.7222\n",
            "Epoch 212/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5487 - accuracy: 0.7381 - val_loss: 0.5946 - val_accuracy: 0.7222\n",
            "Epoch 213/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5475 - accuracy: 0.7381 - val_loss: 0.5932 - val_accuracy: 0.7222\n",
            "Epoch 214/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5460 - accuracy: 0.7381 - val_loss: 0.5917 - val_accuracy: 0.7222\n",
            "Epoch 215/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5446 - accuracy: 0.7381 - val_loss: 0.5902 - val_accuracy: 0.7222\n",
            "Epoch 216/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5433 - accuracy: 0.7262 - val_loss: 0.5889 - val_accuracy: 0.6944\n",
            "Epoch 217/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5419 - accuracy: 0.7024 - val_loss: 0.5875 - val_accuracy: 0.6944\n",
            "Epoch 218/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5406 - accuracy: 0.7024 - val_loss: 0.5861 - val_accuracy: 0.6944\n",
            "Epoch 219/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5392 - accuracy: 0.7024 - val_loss: 0.5848 - val_accuracy: 0.6944\n",
            "Epoch 220/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.5379 - accuracy: 0.7143 - val_loss: 0.5834 - val_accuracy: 0.6944\n",
            "Epoch 221/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5368 - accuracy: 0.7143 - val_loss: 0.5820 - val_accuracy: 0.6944\n",
            "Epoch 222/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5353 - accuracy: 0.7143 - val_loss: 0.5806 - val_accuracy: 0.6944\n",
            "Epoch 223/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5342 - accuracy: 0.7143 - val_loss: 0.5790 - val_accuracy: 0.6944\n",
            "Epoch 224/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5328 - accuracy: 0.7143 - val_loss: 0.5776 - val_accuracy: 0.6944\n",
            "Epoch 225/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5316 - accuracy: 0.7143 - val_loss: 0.5761 - val_accuracy: 0.6944\n",
            "Epoch 226/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5303 - accuracy: 0.7143 - val_loss: 0.5748 - val_accuracy: 0.6944\n",
            "Epoch 227/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5291 - accuracy: 0.7143 - val_loss: 0.5734 - val_accuracy: 0.6944\n",
            "Epoch 228/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5279 - accuracy: 0.7143 - val_loss: 0.5721 - val_accuracy: 0.6944\n",
            "Epoch 229/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5267 - accuracy: 0.7143 - val_loss: 0.5707 - val_accuracy: 0.6944\n",
            "Epoch 230/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5254 - accuracy: 0.7143 - val_loss: 0.5693 - val_accuracy: 0.6944\n",
            "Epoch 231/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5242 - accuracy: 0.7143 - val_loss: 0.5679 - val_accuracy: 0.6944\n",
            "Epoch 232/500\n",
            "3/3 [==============================] - 0s 12ms/step - loss: 0.5230 - accuracy: 0.7262 - val_loss: 0.5666 - val_accuracy: 0.6944\n",
            "Epoch 233/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5218 - accuracy: 0.7143 - val_loss: 0.5653 - val_accuracy: 0.6944\n",
            "Epoch 234/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5207 - accuracy: 0.7143 - val_loss: 0.5640 - val_accuracy: 0.6944\n",
            "Epoch 235/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5196 - accuracy: 0.7262 - val_loss: 0.5628 - val_accuracy: 0.6944\n",
            "Epoch 236/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5183 - accuracy: 0.7262 - val_loss: 0.5614 - val_accuracy: 0.6944\n",
            "Epoch 237/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.5173 - accuracy: 0.7262 - val_loss: 0.5603 - val_accuracy: 0.6944\n",
            "Epoch 238/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5160 - accuracy: 0.7262 - val_loss: 0.5589 - val_accuracy: 0.6944\n",
            "Epoch 239/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5149 - accuracy: 0.7262 - val_loss: 0.5575 - val_accuracy: 0.7222\n",
            "Epoch 240/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5139 - accuracy: 0.7262 - val_loss: 0.5562 - val_accuracy: 0.7222\n",
            "Epoch 241/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5126 - accuracy: 0.7262 - val_loss: 0.5548 - val_accuracy: 0.7222\n",
            "Epoch 242/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5115 - accuracy: 0.7500 - val_loss: 0.5534 - val_accuracy: 0.7222\n",
            "Epoch 243/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5107 - accuracy: 0.7500 - val_loss: 0.5520 - val_accuracy: 0.7222\n",
            "Epoch 244/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5093 - accuracy: 0.7619 - val_loss: 0.5507 - val_accuracy: 0.7222\n",
            "Epoch 245/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5082 - accuracy: 0.7619 - val_loss: 0.5495 - val_accuracy: 0.7222\n",
            "Epoch 246/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.5071 - accuracy: 0.7619 - val_loss: 0.5483 - val_accuracy: 0.7222\n",
            "Epoch 247/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5061 - accuracy: 0.7619 - val_loss: 0.5470 - val_accuracy: 0.7222\n",
            "Epoch 248/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5050 - accuracy: 0.7619 - val_loss: 0.5457 - val_accuracy: 0.7222\n",
            "Epoch 249/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5038 - accuracy: 0.7619 - val_loss: 0.5445 - val_accuracy: 0.7222\n",
            "Epoch 250/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5028 - accuracy: 0.7619 - val_loss: 0.5434 - val_accuracy: 0.7222\n",
            "Epoch 251/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.5018 - accuracy: 0.7619 - val_loss: 0.5421 - val_accuracy: 0.7222\n",
            "Epoch 252/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.5008 - accuracy: 0.7619 - val_loss: 0.5410 - val_accuracy: 0.7222\n",
            "Epoch 253/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4998 - accuracy: 0.7619 - val_loss: 0.5399 - val_accuracy: 0.7222\n",
            "Epoch 254/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4986 - accuracy: 0.7619 - val_loss: 0.5385 - val_accuracy: 0.7222\n",
            "Epoch 255/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4976 - accuracy: 0.7619 - val_loss: 0.5372 - val_accuracy: 0.7500\n",
            "Epoch 256/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4967 - accuracy: 0.7619 - val_loss: 0.5359 - val_accuracy: 0.7500\n",
            "Epoch 257/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4956 - accuracy: 0.7738 - val_loss: 0.5348 - val_accuracy: 0.7500\n",
            "Epoch 258/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4946 - accuracy: 0.7738 - val_loss: 0.5335 - val_accuracy: 0.7500\n",
            "Epoch 259/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4936 - accuracy: 0.7857 - val_loss: 0.5323 - val_accuracy: 0.7500\n",
            "Epoch 260/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4927 - accuracy: 0.7976 - val_loss: 0.5311 - val_accuracy: 0.7500\n",
            "Epoch 261/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4916 - accuracy: 0.7976 - val_loss: 0.5300 - val_accuracy: 0.7500\n",
            "Epoch 262/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4907 - accuracy: 0.7857 - val_loss: 0.5289 - val_accuracy: 0.7500\n",
            "Epoch 263/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4897 - accuracy: 0.7857 - val_loss: 0.5278 - val_accuracy: 0.7500\n",
            "Epoch 264/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4887 - accuracy: 0.7857 - val_loss: 0.5267 - val_accuracy: 0.7500\n",
            "Epoch 265/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4877 - accuracy: 0.7738 - val_loss: 0.5257 - val_accuracy: 0.7500\n",
            "Epoch 266/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4866 - accuracy: 0.7738 - val_loss: 0.5247 - val_accuracy: 0.7500\n",
            "Epoch 267/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4857 - accuracy: 0.7857 - val_loss: 0.5237 - val_accuracy: 0.7500\n",
            "Epoch 268/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4848 - accuracy: 0.7857 - val_loss: 0.5226 - val_accuracy: 0.7500\n",
            "Epoch 269/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4839 - accuracy: 0.7857 - val_loss: 0.5216 - val_accuracy: 0.7500\n",
            "Epoch 270/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4831 - accuracy: 0.7976 - val_loss: 0.5203 - val_accuracy: 0.7500\n",
            "Epoch 271/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4819 - accuracy: 0.7976 - val_loss: 0.5193 - val_accuracy: 0.7500\n",
            "Epoch 272/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4810 - accuracy: 0.7976 - val_loss: 0.5182 - val_accuracy: 0.7500\n",
            "Epoch 273/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4800 - accuracy: 0.7857 - val_loss: 0.5173 - val_accuracy: 0.7500\n",
            "Epoch 274/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4792 - accuracy: 0.7857 - val_loss: 0.5164 - val_accuracy: 0.7500\n",
            "Epoch 275/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4783 - accuracy: 0.7857 - val_loss: 0.5153 - val_accuracy: 0.7500\n",
            "Epoch 276/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4774 - accuracy: 0.7857 - val_loss: 0.5144 - val_accuracy: 0.7500\n",
            "Epoch 277/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4765 - accuracy: 0.7857 - val_loss: 0.5134 - val_accuracy: 0.7500\n",
            "Epoch 278/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4757 - accuracy: 0.7857 - val_loss: 0.5121 - val_accuracy: 0.7500\n",
            "Epoch 279/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4746 - accuracy: 0.7857 - val_loss: 0.5110 - val_accuracy: 0.7500\n",
            "Epoch 280/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4737 - accuracy: 0.7857 - val_loss: 0.5099 - val_accuracy: 0.7500\n",
            "Epoch 281/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4729 - accuracy: 0.7857 - val_loss: 0.5087 - val_accuracy: 0.7778\n",
            "Epoch 282/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4719 - accuracy: 0.8214 - val_loss: 0.5075 - val_accuracy: 0.7778\n",
            "Epoch 283/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4711 - accuracy: 0.8333 - val_loss: 0.5064 - val_accuracy: 0.8056\n",
            "Epoch 284/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4702 - accuracy: 0.8333 - val_loss: 0.5053 - val_accuracy: 0.8056\n",
            "Epoch 285/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4694 - accuracy: 0.8333 - val_loss: 0.5042 - val_accuracy: 0.8333\n",
            "Epoch 286/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4685 - accuracy: 0.8333 - val_loss: 0.5031 - val_accuracy: 0.8333\n",
            "Epoch 287/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4676 - accuracy: 0.8333 - val_loss: 0.5019 - val_accuracy: 0.8333\n",
            "Epoch 288/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4669 - accuracy: 0.8571 - val_loss: 0.5008 - val_accuracy: 0.8333\n",
            "Epoch 289/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4660 - accuracy: 0.8571 - val_loss: 0.4998 - val_accuracy: 0.8333\n",
            "Epoch 290/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.4650 - accuracy: 0.8571 - val_loss: 0.4988 - val_accuracy: 0.8333\n",
            "Epoch 291/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4644 - accuracy: 0.8690 - val_loss: 0.4977 - val_accuracy: 0.8611\n",
            "Epoch 292/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4634 - accuracy: 0.8690 - val_loss: 0.4968 - val_accuracy: 0.8333\n",
            "Epoch 293/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4627 - accuracy: 0.8333 - val_loss: 0.4960 - val_accuracy: 0.8333\n",
            "Epoch 294/500\n",
            "3/3 [==============================] - 0s 10ms/step - loss: 0.4617 - accuracy: 0.8214 - val_loss: 0.4950 - val_accuracy: 0.8333\n",
            "Epoch 295/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4608 - accuracy: 0.8333 - val_loss: 0.4939 - val_accuracy: 0.8333\n",
            "Epoch 296/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4599 - accuracy: 0.8333 - val_loss: 0.4929 - val_accuracy: 0.8333\n",
            "Epoch 297/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4591 - accuracy: 0.8690 - val_loss: 0.4919 - val_accuracy: 0.8333\n",
            "Epoch 298/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4583 - accuracy: 0.8810 - val_loss: 0.4909 - val_accuracy: 0.8611\n",
            "Epoch 299/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4575 - accuracy: 0.8810 - val_loss: 0.4899 - val_accuracy: 0.8611\n",
            "Epoch 300/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4567 - accuracy: 0.8810 - val_loss: 0.4890 - val_accuracy: 0.8333\n",
            "Epoch 301/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4558 - accuracy: 0.8690 - val_loss: 0.4881 - val_accuracy: 0.8333\n",
            "Epoch 302/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4551 - accuracy: 0.8571 - val_loss: 0.4873 - val_accuracy: 0.8333\n",
            "Epoch 303/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4543 - accuracy: 0.8333 - val_loss: 0.4864 - val_accuracy: 0.8333\n",
            "Epoch 304/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4535 - accuracy: 0.8333 - val_loss: 0.4855 - val_accuracy: 0.8333\n",
            "Epoch 305/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4528 - accuracy: 0.8333 - val_loss: 0.4846 - val_accuracy: 0.8333\n",
            "Epoch 306/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4521 - accuracy: 0.8333 - val_loss: 0.4840 - val_accuracy: 0.8333\n",
            "Epoch 307/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4512 - accuracy: 0.8333 - val_loss: 0.4830 - val_accuracy: 0.8333\n",
            "Epoch 308/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4503 - accuracy: 0.8333 - val_loss: 0.4821 - val_accuracy: 0.8333\n",
            "Epoch 309/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4496 - accuracy: 0.8333 - val_loss: 0.4812 - val_accuracy: 0.8333\n",
            "Epoch 310/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4488 - accuracy: 0.8333 - val_loss: 0.4803 - val_accuracy: 0.8333\n",
            "Epoch 311/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4480 - accuracy: 0.8333 - val_loss: 0.4792 - val_accuracy: 0.8611\n",
            "Epoch 312/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4473 - accuracy: 0.8571 - val_loss: 0.4781 - val_accuracy: 0.8611\n",
            "Epoch 313/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4465 - accuracy: 0.8690 - val_loss: 0.4772 - val_accuracy: 0.8611\n",
            "Epoch 314/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4458 - accuracy: 0.8690 - val_loss: 0.4762 - val_accuracy: 0.8611\n",
            "Epoch 315/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4449 - accuracy: 0.8690 - val_loss: 0.4752 - val_accuracy: 0.8889\n",
            "Epoch 316/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4442 - accuracy: 0.8690 - val_loss: 0.4741 - val_accuracy: 0.9167\n",
            "Epoch 317/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4434 - accuracy: 0.8810 - val_loss: 0.4731 - val_accuracy: 0.9167\n",
            "Epoch 318/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4426 - accuracy: 0.8810 - val_loss: 0.4722 - val_accuracy: 0.9167\n",
            "Epoch 319/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4419 - accuracy: 0.8929 - val_loss: 0.4713 - val_accuracy: 0.9167\n",
            "Epoch 320/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4411 - accuracy: 0.8929 - val_loss: 0.4703 - val_accuracy: 0.9167\n",
            "Epoch 321/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4403 - accuracy: 0.8929 - val_loss: 0.4693 - val_accuracy: 0.9167\n",
            "Epoch 322/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4396 - accuracy: 0.8929 - val_loss: 0.4684 - val_accuracy: 0.9167\n",
            "Epoch 323/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4389 - accuracy: 0.8929 - val_loss: 0.4674 - val_accuracy: 0.9167\n",
            "Epoch 324/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4381 - accuracy: 0.8929 - val_loss: 0.4664 - val_accuracy: 0.9167\n",
            "Epoch 325/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4374 - accuracy: 0.8929 - val_loss: 0.4655 - val_accuracy: 0.9167\n",
            "Epoch 326/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4366 - accuracy: 0.8929 - val_loss: 0.4645 - val_accuracy: 0.9167\n",
            "Epoch 327/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4359 - accuracy: 0.9048 - val_loss: 0.4636 - val_accuracy: 0.9444\n",
            "Epoch 328/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4352 - accuracy: 0.9048 - val_loss: 0.4627 - val_accuracy: 0.9167\n",
            "Epoch 329/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4344 - accuracy: 0.9048 - val_loss: 0.4618 - val_accuracy: 0.9167\n",
            "Epoch 330/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4337 - accuracy: 0.8929 - val_loss: 0.4610 - val_accuracy: 0.9167\n",
            "Epoch 331/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4330 - accuracy: 0.8929 - val_loss: 0.4602 - val_accuracy: 0.9167\n",
            "Epoch 332/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4323 - accuracy: 0.8929 - val_loss: 0.4594 - val_accuracy: 0.9167\n",
            "Epoch 333/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4315 - accuracy: 0.8810 - val_loss: 0.4586 - val_accuracy: 0.9167\n",
            "Epoch 334/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4307 - accuracy: 0.8810 - val_loss: 0.4579 - val_accuracy: 0.9167\n",
            "Epoch 335/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4300 - accuracy: 0.8810 - val_loss: 0.4571 - val_accuracy: 0.9167\n",
            "Epoch 336/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4294 - accuracy: 0.8810 - val_loss: 0.4563 - val_accuracy: 0.9167\n",
            "Epoch 337/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4287 - accuracy: 0.8690 - val_loss: 0.4555 - val_accuracy: 0.9167\n",
            "Epoch 338/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4286 - accuracy: 0.8690 - val_loss: 0.4551 - val_accuracy: 0.9167\n",
            "Epoch 339/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4274 - accuracy: 0.8690 - val_loss: 0.4542 - val_accuracy: 0.9167\n",
            "Epoch 340/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4267 - accuracy: 0.8690 - val_loss: 0.4532 - val_accuracy: 0.9167\n",
            "Epoch 341/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4260 - accuracy: 0.8690 - val_loss: 0.4522 - val_accuracy: 0.9167\n",
            "Epoch 342/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4253 - accuracy: 0.8690 - val_loss: 0.4512 - val_accuracy: 0.9167\n",
            "Epoch 343/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4245 - accuracy: 0.8690 - val_loss: 0.4503 - val_accuracy: 0.9167\n",
            "Epoch 344/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4241 - accuracy: 0.8810 - val_loss: 0.4493 - val_accuracy: 0.9167\n",
            "Epoch 345/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4231 - accuracy: 0.8810 - val_loss: 0.4486 - val_accuracy: 0.9167\n",
            "Epoch 346/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4224 - accuracy: 0.8810 - val_loss: 0.4477 - val_accuracy: 0.9167\n",
            "Epoch 347/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.4217 - accuracy: 0.8810 - val_loss: 0.4469 - val_accuracy: 0.9167\n",
            "Epoch 348/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4212 - accuracy: 0.8810 - val_loss: 0.4459 - val_accuracy: 0.9167\n",
            "Epoch 349/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4206 - accuracy: 0.8810 - val_loss: 0.4452 - val_accuracy: 0.9167\n",
            "Epoch 350/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4198 - accuracy: 0.8810 - val_loss: 0.4443 - val_accuracy: 0.9167\n",
            "Epoch 351/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4190 - accuracy: 0.8810 - val_loss: 0.4436 - val_accuracy: 0.9167\n",
            "Epoch 352/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4184 - accuracy: 0.8810 - val_loss: 0.4427 - val_accuracy: 0.9167\n",
            "Epoch 353/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4177 - accuracy: 0.8810 - val_loss: 0.4419 - val_accuracy: 0.9167\n",
            "Epoch 354/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4170 - accuracy: 0.8810 - val_loss: 0.4411 - val_accuracy: 0.9167\n",
            "Epoch 355/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4164 - accuracy: 0.8810 - val_loss: 0.4402 - val_accuracy: 0.9167\n",
            "Epoch 356/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4158 - accuracy: 0.8810 - val_loss: 0.4395 - val_accuracy: 0.9167\n",
            "Epoch 357/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4150 - accuracy: 0.8810 - val_loss: 0.4386 - val_accuracy: 0.9167\n",
            "Epoch 358/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4144 - accuracy: 0.8810 - val_loss: 0.4378 - val_accuracy: 0.9167\n",
            "Epoch 359/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4137 - accuracy: 0.8810 - val_loss: 0.4368 - val_accuracy: 0.9722\n",
            "Epoch 360/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4130 - accuracy: 0.8810 - val_loss: 0.4359 - val_accuracy: 0.9722\n",
            "Epoch 361/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4123 - accuracy: 0.9048 - val_loss: 0.4350 - val_accuracy: 0.9722\n",
            "Epoch 362/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4118 - accuracy: 0.9048 - val_loss: 0.4340 - val_accuracy: 0.9722\n",
            "Epoch 363/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4110 - accuracy: 0.9167 - val_loss: 0.4331 - val_accuracy: 0.9722\n",
            "Epoch 364/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4103 - accuracy: 0.9167 - val_loss: 0.4322 - val_accuracy: 1.0000\n",
            "Epoch 365/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4098 - accuracy: 0.9286 - val_loss: 0.4313 - val_accuracy: 1.0000\n",
            "Epoch 366/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4091 - accuracy: 0.9286 - val_loss: 0.4305 - val_accuracy: 1.0000\n",
            "Epoch 367/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4084 - accuracy: 0.9286 - val_loss: 0.4296 - val_accuracy: 1.0000\n",
            "Epoch 368/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4079 - accuracy: 0.9286 - val_loss: 0.4288 - val_accuracy: 1.0000\n",
            "Epoch 369/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4071 - accuracy: 0.9286 - val_loss: 0.4281 - val_accuracy: 1.0000\n",
            "Epoch 370/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.4064 - accuracy: 0.9286 - val_loss: 0.4273 - val_accuracy: 1.0000\n",
            "Epoch 371/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4057 - accuracy: 0.9286 - val_loss: 0.4266 - val_accuracy: 1.0000\n",
            "Epoch 372/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4053 - accuracy: 0.9167 - val_loss: 0.4260 - val_accuracy: 1.0000\n",
            "Epoch 373/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4048 - accuracy: 0.9167 - val_loss: 0.4254 - val_accuracy: 0.9722\n",
            "Epoch 374/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4039 - accuracy: 0.9167 - val_loss: 0.4246 - val_accuracy: 0.9722\n",
            "Epoch 375/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4034 - accuracy: 0.9167 - val_loss: 0.4238 - val_accuracy: 0.9722\n",
            "Epoch 376/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4027 - accuracy: 0.9167 - val_loss: 0.4230 - val_accuracy: 1.0000\n",
            "Epoch 377/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4021 - accuracy: 0.9167 - val_loss: 0.4220 - val_accuracy: 1.0000\n",
            "Epoch 378/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4013 - accuracy: 0.9167 - val_loss: 0.4211 - val_accuracy: 1.0000\n",
            "Epoch 379/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4008 - accuracy: 0.9286 - val_loss: 0.4202 - val_accuracy: 1.0000\n",
            "Epoch 380/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.4001 - accuracy: 0.9286 - val_loss: 0.4194 - val_accuracy: 1.0000\n",
            "Epoch 381/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3995 - accuracy: 0.9286 - val_loss: 0.4186 - val_accuracy: 1.0000\n",
            "Epoch 382/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3988 - accuracy: 0.9286 - val_loss: 0.4178 - val_accuracy: 1.0000\n",
            "Epoch 383/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3982 - accuracy: 0.9286 - val_loss: 0.4170 - val_accuracy: 1.0000\n",
            "Epoch 384/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3976 - accuracy: 0.9286 - val_loss: 0.4162 - val_accuracy: 1.0000\n",
            "Epoch 385/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3971 - accuracy: 0.9286 - val_loss: 0.4155 - val_accuracy: 1.0000\n",
            "Epoch 386/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.3963 - accuracy: 0.9286 - val_loss: 0.4147 - val_accuracy: 1.0000\n",
            "Epoch 387/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3957 - accuracy: 0.9286 - val_loss: 0.4139 - val_accuracy: 1.0000\n",
            "Epoch 388/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3951 - accuracy: 0.9286 - val_loss: 0.4131 - val_accuracy: 1.0000\n",
            "Epoch 389/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3947 - accuracy: 0.9286 - val_loss: 0.4122 - val_accuracy: 1.0000\n",
            "Epoch 390/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3939 - accuracy: 0.9405 - val_loss: 0.4114 - val_accuracy: 1.0000\n",
            "Epoch 391/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3932 - accuracy: 0.9286 - val_loss: 0.4108 - val_accuracy: 1.0000\n",
            "Epoch 392/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3926 - accuracy: 0.9286 - val_loss: 0.4101 - val_accuracy: 1.0000\n",
            "Epoch 393/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3920 - accuracy: 0.9286 - val_loss: 0.4094 - val_accuracy: 1.0000\n",
            "Epoch 394/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3914 - accuracy: 0.9167 - val_loss: 0.4087 - val_accuracy: 1.0000\n",
            "Epoch 395/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3909 - accuracy: 0.9167 - val_loss: 0.4079 - val_accuracy: 1.0000\n",
            "Epoch 396/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3903 - accuracy: 0.9167 - val_loss: 0.4072 - val_accuracy: 1.0000\n",
            "Epoch 397/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3896 - accuracy: 0.9167 - val_loss: 0.4064 - val_accuracy: 1.0000\n",
            "Epoch 398/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3890 - accuracy: 0.9167 - val_loss: 0.4057 - val_accuracy: 1.0000\n",
            "Epoch 399/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3884 - accuracy: 0.9167 - val_loss: 0.4050 - val_accuracy: 1.0000\n",
            "Epoch 400/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3878 - accuracy: 0.9167 - val_loss: 0.4042 - val_accuracy: 1.0000\n",
            "Epoch 401/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3873 - accuracy: 0.9167 - val_loss: 0.4033 - val_accuracy: 1.0000\n",
            "Epoch 402/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3865 - accuracy: 0.9167 - val_loss: 0.4025 - val_accuracy: 1.0000\n",
            "Epoch 403/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3860 - accuracy: 0.9167 - val_loss: 0.4018 - val_accuracy: 1.0000\n",
            "Epoch 404/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3855 - accuracy: 0.9167 - val_loss: 0.4011 - val_accuracy: 1.0000\n",
            "Epoch 405/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3847 - accuracy: 0.9167 - val_loss: 0.4001 - val_accuracy: 1.0000\n",
            "Epoch 406/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.3841 - accuracy: 0.9405 - val_loss: 0.3992 - val_accuracy: 1.0000\n",
            "Epoch 407/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3834 - accuracy: 0.9405 - val_loss: 0.3983 - val_accuracy: 1.0000\n",
            "Epoch 408/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3829 - accuracy: 0.9405 - val_loss: 0.3975 - val_accuracy: 1.0000\n",
            "Epoch 409/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3823 - accuracy: 0.9405 - val_loss: 0.3967 - val_accuracy: 1.0000\n",
            "Epoch 410/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3818 - accuracy: 0.9405 - val_loss: 0.3959 - val_accuracy: 1.0000\n",
            "Epoch 411/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3812 - accuracy: 0.9524 - val_loss: 0.3952 - val_accuracy: 1.0000\n",
            "Epoch 412/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3805 - accuracy: 0.9524 - val_loss: 0.3944 - val_accuracy: 1.0000\n",
            "Epoch 413/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3800 - accuracy: 0.9405 - val_loss: 0.3936 - val_accuracy: 1.0000\n",
            "Epoch 414/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3793 - accuracy: 0.9405 - val_loss: 0.3929 - val_accuracy: 1.0000\n",
            "Epoch 415/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3788 - accuracy: 0.9524 - val_loss: 0.3921 - val_accuracy: 1.0000\n",
            "Epoch 416/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3780 - accuracy: 0.9405 - val_loss: 0.3915 - val_accuracy: 1.0000\n",
            "Epoch 417/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3774 - accuracy: 0.9405 - val_loss: 0.3908 - val_accuracy: 1.0000\n",
            "Epoch 418/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3768 - accuracy: 0.9286 - val_loss: 0.3903 - val_accuracy: 1.0000\n",
            "Epoch 419/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3765 - accuracy: 0.9286 - val_loss: 0.3898 - val_accuracy: 1.0000\n",
            "Epoch 420/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3758 - accuracy: 0.9167 - val_loss: 0.3890 - val_accuracy: 1.0000\n",
            "Epoch 421/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3752 - accuracy: 0.9167 - val_loss: 0.3884 - val_accuracy: 1.0000\n",
            "Epoch 422/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3747 - accuracy: 0.9167 - val_loss: 0.3877 - val_accuracy: 1.0000\n",
            "Epoch 423/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3742 - accuracy: 0.9167 - val_loss: 0.3870 - val_accuracy: 1.0000\n",
            "Epoch 424/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3735 - accuracy: 0.9167 - val_loss: 0.3860 - val_accuracy: 1.0000\n",
            "Epoch 425/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3728 - accuracy: 0.9286 - val_loss: 0.3851 - val_accuracy: 1.0000\n",
            "Epoch 426/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3721 - accuracy: 0.9286 - val_loss: 0.3841 - val_accuracy: 1.0000\n",
            "Epoch 427/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3715 - accuracy: 0.9524 - val_loss: 0.3833 - val_accuracy: 1.0000\n",
            "Epoch 428/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3712 - accuracy: 0.9524 - val_loss: 0.3824 - val_accuracy: 1.0000\n",
            "Epoch 429/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3706 - accuracy: 0.9524 - val_loss: 0.3816 - val_accuracy: 1.0000\n",
            "Epoch 430/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3703 - accuracy: 0.9524 - val_loss: 0.3809 - val_accuracy: 1.0000\n",
            "Epoch 431/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3694 - accuracy: 0.9524 - val_loss: 0.3802 - val_accuracy: 1.0000\n",
            "Epoch 432/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3688 - accuracy: 0.9524 - val_loss: 0.3795 - val_accuracy: 1.0000\n",
            "Epoch 433/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3682 - accuracy: 0.9524 - val_loss: 0.3789 - val_accuracy: 1.0000\n",
            "Epoch 434/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3680 - accuracy: 0.9524 - val_loss: 0.3784 - val_accuracy: 1.0000\n",
            "Epoch 435/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3672 - accuracy: 0.9524 - val_loss: 0.3776 - val_accuracy: 1.0000\n",
            "Epoch 436/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3666 - accuracy: 0.9524 - val_loss: 0.3769 - val_accuracy: 1.0000\n",
            "Epoch 437/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3659 - accuracy: 0.9524 - val_loss: 0.3763 - val_accuracy: 1.0000\n",
            "Epoch 438/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3653 - accuracy: 0.9524 - val_loss: 0.3756 - val_accuracy: 1.0000\n",
            "Epoch 439/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3648 - accuracy: 0.9524 - val_loss: 0.3749 - val_accuracy: 1.0000\n",
            "Epoch 440/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3642 - accuracy: 0.9524 - val_loss: 0.3743 - val_accuracy: 1.0000\n",
            "Epoch 441/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3637 - accuracy: 0.9405 - val_loss: 0.3736 - val_accuracy: 1.0000\n",
            "Epoch 442/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3632 - accuracy: 0.9405 - val_loss: 0.3729 - val_accuracy: 1.0000\n",
            "Epoch 443/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3626 - accuracy: 0.9405 - val_loss: 0.3722 - val_accuracy: 1.0000\n",
            "Epoch 444/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.3620 - accuracy: 0.9405 - val_loss: 0.3714 - val_accuracy: 1.0000\n",
            "Epoch 445/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3614 - accuracy: 0.9405 - val_loss: 0.3707 - val_accuracy: 1.0000\n",
            "Epoch 446/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3609 - accuracy: 0.9524 - val_loss: 0.3699 - val_accuracy: 1.0000\n",
            "Epoch 447/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3602 - accuracy: 0.9524 - val_loss: 0.3690 - val_accuracy: 1.0000\n",
            "Epoch 448/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3597 - accuracy: 0.9524 - val_loss: 0.3682 - val_accuracy: 1.0000\n",
            "Epoch 449/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3591 - accuracy: 0.9524 - val_loss: 0.3674 - val_accuracy: 1.0000\n",
            "Epoch 450/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3588 - accuracy: 0.9643 - val_loss: 0.3666 - val_accuracy: 1.0000\n",
            "Epoch 451/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3582 - accuracy: 0.9643 - val_loss: 0.3659 - val_accuracy: 1.0000\n",
            "Epoch 452/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3577 - accuracy: 0.9643 - val_loss: 0.3652 - val_accuracy: 1.0000\n",
            "Epoch 453/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3570 - accuracy: 0.9643 - val_loss: 0.3644 - val_accuracy: 1.0000\n",
            "Epoch 454/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3565 - accuracy: 0.9643 - val_loss: 0.3637 - val_accuracy: 1.0000\n",
            "Epoch 455/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3560 - accuracy: 0.9643 - val_loss: 0.3630 - val_accuracy: 1.0000\n",
            "Epoch 456/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3553 - accuracy: 0.9643 - val_loss: 0.3623 - val_accuracy: 1.0000\n",
            "Epoch 457/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3547 - accuracy: 0.9643 - val_loss: 0.3616 - val_accuracy: 1.0000\n",
            "Epoch 458/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3545 - accuracy: 0.9524 - val_loss: 0.3609 - val_accuracy: 1.0000\n",
            "Epoch 459/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3536 - accuracy: 0.9524 - val_loss: 0.3602 - val_accuracy: 1.0000\n",
            "Epoch 460/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3529 - accuracy: 0.9643 - val_loss: 0.3595 - val_accuracy: 1.0000\n",
            "Epoch 461/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3524 - accuracy: 0.9643 - val_loss: 0.3588 - val_accuracy: 1.0000\n",
            "Epoch 462/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3519 - accuracy: 0.9643 - val_loss: 0.3581 - val_accuracy: 1.0000\n",
            "Epoch 463/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3514 - accuracy: 0.9643 - val_loss: 0.3573 - val_accuracy: 1.0000\n",
            "Epoch 464/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3509 - accuracy: 0.9643 - val_loss: 0.3566 - val_accuracy: 1.0000\n",
            "Epoch 465/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3502 - accuracy: 0.9643 - val_loss: 0.3559 - val_accuracy: 1.0000\n",
            "Epoch 466/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3496 - accuracy: 0.9643 - val_loss: 0.3553 - val_accuracy: 1.0000\n",
            "Epoch 467/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3491 - accuracy: 0.9643 - val_loss: 0.3546 - val_accuracy: 1.0000\n",
            "Epoch 468/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3489 - accuracy: 0.9524 - val_loss: 0.3541 - val_accuracy: 1.0000\n",
            "Epoch 469/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3481 - accuracy: 0.9524 - val_loss: 0.3533 - val_accuracy: 1.0000\n",
            "Epoch 470/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3474 - accuracy: 0.9524 - val_loss: 0.3526 - val_accuracy: 1.0000\n",
            "Epoch 471/500\n",
            "3/3 [==============================] - 0s 6ms/step - loss: 0.3469 - accuracy: 0.9524 - val_loss: 0.3519 - val_accuracy: 1.0000\n",
            "Epoch 472/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3465 - accuracy: 0.9524 - val_loss: 0.3512 - val_accuracy: 1.0000\n",
            "Epoch 473/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3458 - accuracy: 0.9524 - val_loss: 0.3506 - val_accuracy: 1.0000\n",
            "Epoch 474/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3452 - accuracy: 0.9524 - val_loss: 0.3500 - val_accuracy: 1.0000\n",
            "Epoch 475/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3450 - accuracy: 0.9405 - val_loss: 0.3495 - val_accuracy: 1.0000\n",
            "Epoch 476/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3443 - accuracy: 0.9405 - val_loss: 0.3488 - val_accuracy: 1.0000\n",
            "Epoch 477/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3437 - accuracy: 0.9405 - val_loss: 0.3481 - val_accuracy: 1.0000\n",
            "Epoch 478/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3432 - accuracy: 0.9405 - val_loss: 0.3474 - val_accuracy: 1.0000\n",
            "Epoch 479/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3427 - accuracy: 0.9524 - val_loss: 0.3466 - val_accuracy: 1.0000\n",
            "Epoch 480/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3421 - accuracy: 0.9524 - val_loss: 0.3459 - val_accuracy: 1.0000\n",
            "Epoch 481/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3415 - accuracy: 0.9643 - val_loss: 0.3451 - val_accuracy: 1.0000\n",
            "Epoch 482/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3410 - accuracy: 0.9643 - val_loss: 0.3443 - val_accuracy: 1.0000\n",
            "Epoch 483/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3404 - accuracy: 0.9643 - val_loss: 0.3436 - val_accuracy: 1.0000\n",
            "Epoch 484/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3401 - accuracy: 0.9643 - val_loss: 0.3429 - val_accuracy: 1.0000\n",
            "Epoch 485/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3394 - accuracy: 0.9643 - val_loss: 0.3422 - val_accuracy: 1.0000\n",
            "Epoch 486/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3388 - accuracy: 0.9643 - val_loss: 0.3415 - val_accuracy: 1.0000\n",
            "Epoch 487/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3385 - accuracy: 0.9643 - val_loss: 0.3409 - val_accuracy: 1.0000\n",
            "Epoch 488/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3378 - accuracy: 0.9643 - val_loss: 0.3403 - val_accuracy: 1.0000\n",
            "Epoch 489/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3373 - accuracy: 0.9643 - val_loss: 0.3395 - val_accuracy: 1.0000\n",
            "Epoch 490/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3368 - accuracy: 0.9643 - val_loss: 0.3389 - val_accuracy: 1.0000\n",
            "Epoch 491/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3361 - accuracy: 0.9643 - val_loss: 0.3382 - val_accuracy: 1.0000\n",
            "Epoch 492/500\n",
            "3/3 [==============================] - 0s 9ms/step - loss: 0.3357 - accuracy: 0.9643 - val_loss: 0.3375 - val_accuracy: 1.0000\n",
            "Epoch 493/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3351 - accuracy: 0.9643 - val_loss: 0.3368 - val_accuracy: 1.0000\n",
            "Epoch 494/500\n",
            "3/3 [==============================] - 0s 8ms/step - loss: 0.3346 - accuracy: 0.9643 - val_loss: 0.3361 - val_accuracy: 1.0000\n",
            "Epoch 495/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3340 - accuracy: 0.9643 - val_loss: 0.3355 - val_accuracy: 1.0000\n",
            "Epoch 496/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3336 - accuracy: 0.9643 - val_loss: 0.3348 - val_accuracy: 1.0000\n",
            "Epoch 497/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3329 - accuracy: 0.9643 - val_loss: 0.3341 - val_accuracy: 1.0000\n",
            "Epoch 498/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3327 - accuracy: 0.9643 - val_loss: 0.3334 - val_accuracy: 1.0000\n",
            "Epoch 499/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3319 - accuracy: 0.9643 - val_loss: 0.3327 - val_accuracy: 1.0000\n",
            "Epoch 500/500\n",
            "3/3 [==============================] - 0s 7ms/step - loss: 0.3315 - accuracy: 0.9643 - val_loss: 0.3321 - val_accuracy: 1.0000\n",
            "CPU times: user 13 s, sys: 866 ms, total: 13.8 s\n",
            "Wall time: 14.2 s\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f45752fbe10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pmsuPaqYpl-2",
        "colab_type": "text"
      },
      "source": [
        "# Bewertung"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iCCbIyrjpl-3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8279ee31-3dfc-4575-eb27-b0a83a909bb4"
      },
      "source": [
        "model.predict(np.array([[ 5.1,  3.5,  1.4,  0.2]]))"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.9625648 , 0.0360607 , 0.00137442]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jQaEAtyVpl-4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "4bdab368-1515-4a61-e77b-399c5741bc2e"
      },
      "source": [
        "X[0], y[0]"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([5.1, 3.5, 1.4, 0.2]), array([1., 0., 0.], dtype=float32))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wrzGbtchpl-7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "5c1af142-3b53-4bd4-8bd0-7467ef8bf6d8"
      },
      "source": [
        "train_loss, train_accuracy = model.evaluate(X_train, y_train)\n",
        "train_loss, train_accuracy"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4/4 [==============================] - 0s 2ms/step - loss: 0.3313 - accuracy: 0.9750\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.331294983625412, 0.9750000238418579)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hnXEggalpl-8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "50a0eed0-1d06-4a5b-9894-23c55f073178"
      },
      "source": [
        "test_loss, test_accuracy = model.evaluate(X_test, y_test)\n",
        "test_loss, test_accuracy"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "1/1 [==============================] - 0s 1ms/step - loss: 0.3250 - accuracy: 0.9333\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.3249962031841278, 0.9333333373069763)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    }
  ]
}