Mid_ML_Keras_training_practice.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Mid ML Keras_training_practice.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/oObelix/DataScience/blob/master/Mid_ML_Keras_training_practice.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6kuM0Ad00gEi"
      },
      "source": [
        "# Практика по обучению моделей\n",
        "\n",
        "В этом практическом задании мы закрепим на практике как обучать модели в tensorflow на синтетических данных. В задачах ниже необходимо получить точность выше 90%. Советуем не сдаваться на этом -- везде реально получить 100%. А в последнем задании можно получить бонусные баллы за модель с минимальным количеством параметров."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q4nbSbaodI3M",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "868f72e3-e4eb-4759-ba67-2cf810b1a420"
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "# tf.enable_eager_execution()\n",
        "print(tf.__version__)"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.5.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZHDLtJa_PyWk",
        "cellView": "form"
      },
      "source": [
        "#@title\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.cm as cm\n",
        "from matplotlib.colors import ListedColormap, LinearSegmentedColormap, Normalize\n",
        "from sklearn.metrics import accuracy_score\n",
        "\n",
        "\n",
        "%matplotlib inline\n",
        "np.random.seed(10)\n",
        "\n",
        "colors = ['red', \"blue\"]\n",
        "labels_cmap = ListedColormap(colors, 2)\n",
        "colors = [(1, 0, 0), (1, 1, 1), (0, 0, 1)]  # R -> W -> B\n",
        "main_cmap = LinearSegmentedColormap.from_list(\"main_scheme\", colors, N=300)\n",
        "\n",
        "def show_data(X, y):\n",
        "    plt.figure(figsize=(5, 5))\n",
        "    plt.scatter(X[:, 0], X[:, 1], s=120, color=labels_cmap(y))\n",
        "    \n",
        "\n",
        "def show_descision_boundary(clf, limits, binary=False, X=None, y=None, n_lines=10, show_lines=False,\n",
        "                           figsize=(5, 5), ax=None):\n",
        "    if limits is None:\n",
        "        if X is not None:\n",
        "            xs = [X[:, 0].min() - .3, X[:, 0].max() + .3]\n",
        "            ys = [X[:, 1].min() - .3, X[:, 1].max() + .3]\n",
        "        else:\n",
        "            xs = [-1, 1]\n",
        "            ys = [-1, 1]\n",
        "    else:\n",
        "        xs, ys = limits\n",
        "    \n",
        "    x_min, x_max = xs\n",
        "    y_min, y_max = ys\n",
        "    \n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.01),\n",
        "                         np.arange(y_min, y_max, 0.01))\n",
        "    \n",
        "    if ax is None:\n",
        "        fig = plt.figure(figsize=figsize)\n",
        "        ax = fig.add_subplot(1, 1, 1)\n",
        "    \n",
        "    if binary:\n",
        "        Z = clf.predict(np.c_[xx.ravel(), xx.ravel()])\n",
        "        Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "        norm = Normalize(vmin=0.,vmax=1.)\n",
        "    else:\n",
        "        Z = clf(np.c_[xx.ravel(), xx.ravel()])\n",
        "        Z = clf(np.c_[xx.ravel(), yy.ravel()])\n",
        "        #if clf.prob_output:\n",
        "        #    norm = Normalize(vmin=0.,vmax=1.)\n",
        "        #else:\n",
        "        norm = Normalize(vmin=-10.,vmax=10., clip=True)\n",
        "        Z = Z.numpy() \n",
        "    \n",
        "    Z = Z.reshape(xx.shape)\n",
        "    Z = Z.astype(np.float32)\n",
        "    \n",
        "    ax.contourf(xx, yy, Z, n_lines, alpha=0.4, cmap=main_cmap, norm=norm)\n",
        "    if show_lines:\n",
        "        cp = ax.contour(xx, yy, Z, n_lines)\n",
        "        ax.clabel(cp, inline=True, \n",
        "              fontsize=10, colors=\"green\")\n",
        "    \n",
        "    if y is not None:\n",
        "        X = np.array(X)\n",
        "        y = np.array(y)\n",
        "        ax.scatter(X[:, 0], X[:, 1], s=120, color=labels_cmap(y),\n",
        "                   zorder=4)\n",
        "        \n",
        "\n",
        "def eval_model(model, X, y):\n",
        "    accuracy = model.evaluate(X, y)[1]\n",
        "    if accuracy == 1.0:\n",
        "        print(\"Perfect!\")\n",
        "    elif  accuracy > 0.9:\n",
        "        print(\"Well done! Can you make 100%?\")\n",
        "    else:\n",
        "        print(\"Don't give up!\")\n",
        "    return accuracy"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4Io7wFhHgdXU"
      },
      "source": [
        "**1a**. Начнем с классической задачи -- XOR."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1MUslBwSPy1y",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "4a54d07d-647f-4491-c8af-0e59ad65121e"
      },
      "source": [
        "X = np.random.rand(200, 2) - 0.5\n",
        "y = ((X[:, 0] < 0) ^ (X[:, 1] < 0))\n",
        "X = X*2\n",
        "show_data(X, y)"
      ],
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sEs6b7mlQg5d",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "251ffcf7-3c54-4c5d-a266-dd1b51357400"
      },
      "source": [
        "model = tf.keras.Sequential()\n",
        "# <YOUR CODE STARTS HERE>\n",
        "# 1. добавьте скрытые слои в модель (в первом из них не забудьте про input_shape)\n",
        "#    ~3 скрытых слоев достаточно\n",
        "# 2. добавьте выходной слой он должен быть без активации\n",
        "# 3. определите оптимизатор \n",
        "# 4. определите лосс (не забудьте указать from_logits=True!!) \n",
        "\n",
        "model = tf.keras.Sequential()\n",
        "model.add(tf.keras.layers.Dense(16, input_shape=(2,), activation=\"relu\"))\n",
        "model.add(tf.keras.layers.Dense(4, activation=\"sigmoid\"))\n",
        "model.add(tf.keras.layers.Dense(1))\n",
        "\n",
        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.01)\n",
        "loss = tf.keras.losses.BinaryCrossentropy(from_logits=True)\n",
        "\n",
        "# <YOUR CODE ENDS HERE>\n",
        "\n",
        "model.compile(optimizer=optimizer, loss=loss, metrics=[\"accuracy\"]) \n",
        "history = model.fit(X, y, epochs=200, verbose=1, batch_size=64, shuffle=True)"
      ],
      "execution_count": 95,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.7029 - accuracy: 0.5850\n",
            "Epoch 2/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.6790 - accuracy: 0.5550\n",
            "Epoch 3/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.6672 - accuracy: 0.4750\n",
            "Epoch 4/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.6557 - accuracy: 0.4750\n",
            "Epoch 5/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.6443 - accuracy: 0.4750\n",
            "Epoch 6/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.6310 - accuracy: 0.4950\n",
            "Epoch 7/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.6179 - accuracy: 0.5350\n",
            "Epoch 8/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.6034 - accuracy: 0.5900\n",
            "Epoch 9/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.5873 - accuracy: 0.6250\n",
            "Epoch 10/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.5696 - accuracy: 0.6550\n",
            "Epoch 11/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.5524 - accuracy: 0.6800\n",
            "Epoch 12/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.5345 - accuracy: 0.7100\n",
            "Epoch 13/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.5151 - accuracy: 0.7150\n",
            "Epoch 14/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.4963 - accuracy: 0.7250\n",
            "Epoch 15/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.4794 - accuracy: 0.7550\n",
            "Epoch 16/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.4597 - accuracy: 0.8000\n",
            "Epoch 17/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.4391 - accuracy: 0.8400\n",
            "Epoch 18/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.4199 - accuracy: 0.8850\n",
            "Epoch 19/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.4017 - accuracy: 0.9100\n",
            "Epoch 20/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.3833 - accuracy: 0.9150\n",
            "Epoch 21/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.3660 - accuracy: 0.9150\n",
            "Epoch 22/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.3498 - accuracy: 0.9200\n",
            "Epoch 23/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.3339 - accuracy: 0.9450\n",
            "Epoch 24/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.3178 - accuracy: 0.9300\n",
            "Epoch 25/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.3036 - accuracy: 0.9200\n",
            "Epoch 26/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.2908 - accuracy: 0.9250\n",
            "Epoch 27/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.2785 - accuracy: 0.9400\n",
            "Epoch 28/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.2642 - accuracy: 0.9400\n",
            "Epoch 29/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.2537 - accuracy: 0.9600\n",
            "Epoch 30/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.2445 - accuracy: 0.9650\n",
            "Epoch 31/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.2361 - accuracy: 0.9550\n",
            "Epoch 32/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.2263 - accuracy: 0.9650\n",
            "Epoch 33/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.2158 - accuracy: 0.9550\n",
            "Epoch 34/200\n",
            "4/4 [==============================] - 0s 7ms/step - loss: 0.2058 - accuracy: 0.9550\n",
            "Epoch 35/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1987 - accuracy: 0.9500\n",
            "Epoch 36/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.1934 - accuracy: 0.9550\n",
            "Epoch 37/200\n",
            "4/4 [==============================] - 0s 9ms/step - loss: 0.1869 - accuracy: 0.9450\n",
            "Epoch 38/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1815 - accuracy: 0.9500\n",
            "Epoch 39/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1741 - accuracy: 0.9550\n",
            "Epoch 40/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1673 - accuracy: 0.9600\n",
            "Epoch 41/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1616 - accuracy: 0.9650\n",
            "Epoch 42/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1601 - accuracy: 0.9550\n",
            "Epoch 43/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1554 - accuracy: 0.9550\n",
            "Epoch 44/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1468 - accuracy: 0.9600\n",
            "Epoch 45/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1438 - accuracy: 0.9650\n",
            "Epoch 46/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1422 - accuracy: 0.9850\n",
            "Epoch 47/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1383 - accuracy: 0.9850\n",
            "Epoch 48/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1336 - accuracy: 0.9850\n",
            "Epoch 49/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.1289 - accuracy: 0.9600\n",
            "Epoch 50/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1267 - accuracy: 0.9600\n",
            "Epoch 51/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1236 - accuracy: 0.9650\n",
            "Epoch 52/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1212 - accuracy: 0.9750\n",
            "Epoch 53/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1189 - accuracy: 0.9800\n",
            "Epoch 54/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1160 - accuracy: 0.9700\n",
            "Epoch 55/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1119 - accuracy: 0.9650\n",
            "Epoch 56/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1097 - accuracy: 0.9750\n",
            "Epoch 57/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1086 - accuracy: 0.9750\n",
            "Epoch 58/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.1069 - accuracy: 0.9700\n",
            "Epoch 59/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1028 - accuracy: 0.9750\n",
            "Epoch 60/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.1029 - accuracy: 0.9800\n",
            "Epoch 61/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.1009 - accuracy: 0.9850\n",
            "Epoch 62/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0982 - accuracy: 0.9700\n",
            "Epoch 63/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0971 - accuracy: 0.9700\n",
            "Epoch 64/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0950 - accuracy: 0.9650\n",
            "Epoch 65/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0919 - accuracy: 0.9650\n",
            "Epoch 66/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0890 - accuracy: 0.9800\n",
            "Epoch 67/200\n",
            "4/4 [==============================] - 0s 9ms/step - loss: 0.0909 - accuracy: 0.9850\n",
            "Epoch 68/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0882 - accuracy: 0.9800\n",
            "Epoch 69/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0855 - accuracy: 0.9850\n",
            "Epoch 70/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0844 - accuracy: 0.9800\n",
            "Epoch 71/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0837 - accuracy: 0.9800\n",
            "Epoch 72/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0819 - accuracy: 0.9750\n",
            "Epoch 73/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0811 - accuracy: 0.9850\n",
            "Epoch 74/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0803 - accuracy: 0.9850\n",
            "Epoch 75/200\n",
            "4/4 [==============================] - 0s 7ms/step - loss: 0.0780 - accuracy: 0.9800\n",
            "Epoch 76/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0762 - accuracy: 0.9750\n",
            "Epoch 77/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0764 - accuracy: 0.9700\n",
            "Epoch 78/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0734 - accuracy: 0.9800\n",
            "Epoch 79/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0725 - accuracy: 0.9850\n",
            "Epoch 80/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0719 - accuracy: 0.9750\n",
            "Epoch 81/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0770 - accuracy: 0.9600\n",
            "Epoch 82/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0753 - accuracy: 0.9650\n",
            "Epoch 83/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0685 - accuracy: 0.9850\n",
            "Epoch 84/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0681 - accuracy: 0.9950\n",
            "Epoch 85/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0680 - accuracy: 0.9950\n",
            "Epoch 86/200\n",
            "4/4 [==============================] - 0s 8ms/step - loss: 0.0677 - accuracy: 0.9900\n",
            "Epoch 87/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0651 - accuracy: 0.9950\n",
            "Epoch 88/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0626 - accuracy: 0.9850\n",
            "Epoch 89/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0657 - accuracy: 0.9800\n",
            "Epoch 90/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0655 - accuracy: 0.9800\n",
            "Epoch 91/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0625 - accuracy: 0.9800\n",
            "Epoch 92/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0600 - accuracy: 0.9850\n",
            "Epoch 93/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0617 - accuracy: 0.9800\n",
            "Epoch 94/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0619 - accuracy: 0.9900\n",
            "Epoch 95/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0609 - accuracy: 0.9900\n",
            "Epoch 96/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0578 - accuracy: 0.9900\n",
            "Epoch 97/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0580 - accuracy: 0.9750\n",
            "Epoch 98/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0577 - accuracy: 0.9850\n",
            "Epoch 99/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0565 - accuracy: 0.9950\n",
            "Epoch 100/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0555 - accuracy: 0.9900\n",
            "Epoch 101/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0566 - accuracy: 0.9900\n",
            "Epoch 102/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0556 - accuracy: 1.0000\n",
            "Epoch 103/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0593 - accuracy: 0.9950\n",
            "Epoch 104/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0629 - accuracy: 0.9950\n",
            "Epoch 105/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0554 - accuracy: 0.9950\n",
            "Epoch 106/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0529 - accuracy: 0.9900\n",
            "Epoch 107/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0514 - accuracy: 0.9900\n",
            "Epoch 108/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0496 - accuracy: 0.9900\n",
            "Epoch 109/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0531 - accuracy: 0.9900\n",
            "Epoch 110/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0499 - accuracy: 0.9950\n",
            "Epoch 111/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0502 - accuracy: 0.9850\n",
            "Epoch 112/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0507 - accuracy: 0.9750\n",
            "Epoch 113/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0514 - accuracy: 0.9800\n",
            "Epoch 114/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0510 - accuracy: 0.9850\n",
            "Epoch 115/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0511 - accuracy: 0.9900\n",
            "Epoch 116/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0490 - accuracy: 0.9950\n",
            "Epoch 117/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0486 - accuracy: 1.0000\n",
            "Epoch 118/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0492 - accuracy: 1.0000\n",
            "Epoch 119/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0472 - accuracy: 0.9950\n",
            "Epoch 120/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0448 - accuracy: 0.9900\n",
            "Epoch 121/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0441 - accuracy: 0.9950\n",
            "Epoch 122/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0442 - accuracy: 0.9900\n",
            "Epoch 123/200\n",
            "4/4 [==============================] - 0s 7ms/step - loss: 0.0451 - accuracy: 0.9900\n",
            "Epoch 124/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0443 - accuracy: 0.9900\n",
            "Epoch 125/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0427 - accuracy: 0.9950\n",
            "Epoch 126/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0416 - accuracy: 1.0000\n",
            "Epoch 127/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0410 - accuracy: 0.9950\n",
            "Epoch 128/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0411 - accuracy: 0.9950\n",
            "Epoch 129/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0431 - accuracy: 0.9900\n",
            "Epoch 130/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0464 - accuracy: 0.9850\n",
            "Epoch 131/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0434 - accuracy: 0.9900\n",
            "Epoch 132/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0414 - accuracy: 0.9900\n",
            "Epoch 133/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0397 - accuracy: 0.9900\n",
            "Epoch 134/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0385 - accuracy: 0.9900\n",
            "Epoch 135/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0377 - accuracy: 1.0000\n",
            "Epoch 136/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0421 - accuracy: 0.9950\n",
            "Epoch 137/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0431 - accuracy: 0.9900\n",
            "Epoch 138/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0389 - accuracy: 0.9900\n",
            "Epoch 139/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0387 - accuracy: 0.9850\n",
            "Epoch 140/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0379 - accuracy: 0.9900\n",
            "Epoch 141/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0366 - accuracy: 0.9950\n",
            "Epoch 142/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0361 - accuracy: 0.9950\n",
            "Epoch 143/200\n",
            "4/4 [==============================] - 0s 8ms/step - loss: 0.0354 - accuracy: 0.9950\n",
            "Epoch 144/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0358 - accuracy: 0.9950\n",
            "Epoch 145/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0355 - accuracy: 0.9850\n",
            "Epoch 146/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0358 - accuracy: 1.0000\n",
            "Epoch 147/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0359 - accuracy: 1.0000\n",
            "Epoch 148/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0358 - accuracy: 1.0000\n",
            "Epoch 149/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0345 - accuracy: 0.9950\n",
            "Epoch 150/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0375 - accuracy: 0.9850\n",
            "Epoch 151/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0422 - accuracy: 0.9900\n",
            "Epoch 152/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0486 - accuracy: 0.9950\n",
            "Epoch 153/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0405 - accuracy: 0.9950\n",
            "Epoch 154/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0360 - accuracy: 0.9950\n",
            "Epoch 155/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0350 - accuracy: 0.9850\n",
            "Epoch 156/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0346 - accuracy: 0.9900\n",
            "Epoch 157/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0339 - accuracy: 0.9950\n",
            "Epoch 158/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0345 - accuracy: 0.9950\n",
            "Epoch 159/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0334 - accuracy: 1.0000\n",
            "Epoch 160/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0299 - accuracy: 0.9950\n",
            "Epoch 161/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0421 - accuracy: 0.9750\n",
            "Epoch 162/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0442 - accuracy: 0.9750\n",
            "Epoch 163/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0340 - accuracy: 0.9900\n",
            "Epoch 164/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0324 - accuracy: 1.0000\n",
            "Epoch 165/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0361 - accuracy: 0.9900\n",
            "Epoch 166/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0374 - accuracy: 0.9900\n",
            "Epoch 167/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0344 - accuracy: 0.9850\n",
            "Epoch 168/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0318 - accuracy: 0.9900\n",
            "Epoch 169/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0294 - accuracy: 0.9900\n",
            "Epoch 170/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0284 - accuracy: 1.0000\n",
            "Epoch 171/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0283 - accuracy: 1.0000\n",
            "Epoch 172/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0280 - accuracy: 1.0000\n",
            "Epoch 173/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0278 - accuracy: 1.0000\n",
            "Epoch 174/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0276 - accuracy: 1.0000\n",
            "Epoch 175/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0280 - accuracy: 1.0000\n",
            "Epoch 176/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0279 - accuracy: 1.0000\n",
            "Epoch 177/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0279 - accuracy: 1.0000\n",
            "Epoch 178/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0275 - accuracy: 1.0000\n",
            "Epoch 179/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0274 - accuracy: 0.9950\n",
            "Epoch 180/200\n",
            "4/4 [==============================] - 0s 9ms/step - loss: 0.0315 - accuracy: 0.9900\n",
            "Epoch 181/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0285 - accuracy: 0.9950\n",
            "Epoch 182/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0260 - accuracy: 0.9950\n",
            "Epoch 183/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0281 - accuracy: 1.0000\n",
            "Epoch 184/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0296 - accuracy: 1.0000\n",
            "Epoch 185/200\n",
            "4/4 [==============================] - 0s 8ms/step - loss: 0.0282 - accuracy: 1.0000\n",
            "Epoch 186/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0259 - accuracy: 1.0000\n",
            "Epoch 187/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0257 - accuracy: 0.9950\n",
            "Epoch 188/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0257 - accuracy: 0.9950\n",
            "Epoch 189/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0254 - accuracy: 0.9950\n",
            "Epoch 190/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0248 - accuracy: 0.9950\n",
            "Epoch 191/200\n",
            "4/4 [==============================] - 0s 6ms/step - loss: 0.0245 - accuracy: 1.0000\n",
            "Epoch 192/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0263 - accuracy: 1.0000\n",
            "Epoch 193/200\n",
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0262 - accuracy: 1.0000\n",
            "Epoch 194/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0255 - accuracy: 1.0000\n",
            "Epoch 195/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0248 - accuracy: 0.9950\n",
            "Epoch 196/200\n",
            "4/4 [==============================] - 0s 5ms/step - loss: 0.0305 - accuracy: 0.9850\n",
            "Epoch 197/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0286 - accuracy: 0.9950\n",
            "Epoch 198/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0282 - accuracy: 0.9900\n",
            "Epoch 199/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0273 - accuracy: 0.9900\n",
            "Epoch 200/200\n",
            "4/4 [==============================] - 0s 4ms/step - loss: 0.0231 - accuracy: 0.9950\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P8VXug0TRKl1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 354
        },
        "outputId": "a7ef5214-4d2e-45af-df21-add2bb0098bf"
      },
      "source": [
        "limits = [[-0.6, 0.6], [-0.6, 0.6]]\n",
        "eval_model(model, X, y)\n",
        "show_descision_boundary(limits=None, clf=model, binary=False,\n",
        "                                X=X,\n",
        "                                y=y, \n",
        "                                n_lines=50,\n",
        "                                show_lines=False)"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7/7 [==============================] - 0s 3ms/step - loss: 0.0256 - accuracy: 1.0000\n",
            "Perfect!\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WgU9ZW5vo8NS"
      },
      "source": [
        "**1b**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EDC9kIa6b0KN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "00f4fe16-cd0b-44cb-e3bf-7cb849f76c74"
      },
      "source": [
        "from sklearn.datasets import make_moons\n",
        "\n",
        "X, y = make_moons(noise=0.04)\n",
        "show_data(X, y)"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OvRCmDeqS87n",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a5d1c9fa-adf6-4af0-ba6d-d6716e6e9a74"
      },
      "source": [
        "model = tf.keras.Sequential()\n",
        "# <YOUR CODE STARTS HERE>\n",
        "# 1. добавьте скрытые слои в модель (в первом из них не забудьте про input_shape)\n",
        "#    ~3 скрытых слоев достаточно\n",
        "# 2. добавьте выходной слой он должен быть без активации\n",
        "# 3. определите оптимизатор \n",
        "# 4. определите лосс (не забудьте указать from_logits=True!!) \n",
        "\n",
        "model = tf.keras.Sequential()\n",
        "model.add(tf.keras.layers.Dense(16, input_shape=(2,), activation=\"relu\"))\n",
        "model.add(tf.keras.layers.Dense(4, activation=\"relu\"))\n",
        "model.add(tf.keras.layers.Dense(1))\n",
        "\n",
        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.01)\n",
        "loss = tf.keras.losses.BinaryCrossentropy(from_logits=True)\n",
        "\n",
        "# <YOUR CODE ENDS HERE>\n",
        "\n",
        "model.compile(optimizer=optimizer, loss=loss, metrics=[\"accuracy\"]) \n",
        "history = model.fit(X, y, epochs=200, verbose=1, batch_size=64, shuffle=True)"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/200\n",
            "2/2 [==============================] - 1s 6ms/step - loss: 0.7932 - accuracy: 0.5000\n",
            "Epoch 2/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.7411 - accuracy: 0.5000\n",
            "Epoch 3/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.7037 - accuracy: 0.5000\n",
            "Epoch 4/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.6805 - accuracy: 0.5000\n",
            "Epoch 5/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.6573 - accuracy: 0.5000\n",
            "Epoch 6/200\n",
            "2/2 [==============================] - 0s 3ms/step - loss: 0.6339 - accuracy: 0.5200\n",
            "Epoch 7/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6067 - accuracy: 0.6900\n",
            "Epoch 8/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5763 - accuracy: 0.8300\n",
            "Epoch 9/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5443 - accuracy: 0.8400\n",
            "Epoch 10/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.5137 - accuracy: 0.8500\n",
            "Epoch 11/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4824 - accuracy: 0.8600\n",
            "Epoch 12/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.4538 - accuracy: 0.8700\n",
            "Epoch 13/200\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.4236 - accuracy: 0.8700\n",
            "Epoch 14/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.3972 - accuracy: 0.8700\n",
            "Epoch 15/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3728 - accuracy: 0.8700\n",
            "Epoch 16/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3521 - accuracy: 0.8700\n",
            "Epoch 17/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3321 - accuracy: 0.8700\n",
            "Epoch 18/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3163 - accuracy: 0.8800\n",
            "Epoch 19/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.3056 - accuracy: 0.8800\n",
            "Epoch 20/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2956 - accuracy: 0.8800\n",
            "Epoch 21/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2861 - accuracy: 0.8800\n",
            "Epoch 22/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.2786 - accuracy: 0.8800\n",
            "Epoch 23/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2733 - accuracy: 0.8800\n",
            "Epoch 24/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2681 - accuracy: 0.8700\n",
            "Epoch 25/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.2647 - accuracy: 0.8700\n",
            "Epoch 26/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2604 - accuracy: 0.8700\n",
            "Epoch 27/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.2564 - accuracy: 0.8700\n",
            "Epoch 28/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2525 - accuracy: 0.8800\n",
            "Epoch 29/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2493 - accuracy: 0.8700\n",
            "Epoch 30/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2447 - accuracy: 0.8800\n",
            "Epoch 31/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.2431 - accuracy: 0.8900\n",
            "Epoch 32/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2389 - accuracy: 0.8900\n",
            "Epoch 33/200\n",
            "2/2 [==============================] - 0s 21ms/step - loss: 0.2352 - accuracy: 0.8900\n",
            "Epoch 34/200\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 0.2294 - accuracy: 0.8800\n",
            "Epoch 35/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2253 - accuracy: 0.8600\n",
            "Epoch 36/200\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.2209 - accuracy: 0.8600\n",
            "Epoch 37/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2193 - accuracy: 0.8600\n",
            "Epoch 38/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.2144 - accuracy: 0.8700\n",
            "Epoch 39/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2088 - accuracy: 0.8700\n",
            "Epoch 40/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2091 - accuracy: 0.8800\n",
            "Epoch 41/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2028 - accuracy: 0.8800\n",
            "Epoch 42/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1979 - accuracy: 0.8900\n",
            "Epoch 43/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1928 - accuracy: 0.9100\n",
            "Epoch 44/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1892 - accuracy: 0.9300\n",
            "Epoch 45/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.1854 - accuracy: 0.9300\n",
            "Epoch 46/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1803 - accuracy: 0.9200\n",
            "Epoch 47/200\n",
            "2/2 [==============================] - 0s 24ms/step - loss: 0.1753 - accuracy: 0.9100\n",
            "Epoch 48/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.1708 - accuracy: 0.9100\n",
            "Epoch 49/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1663 - accuracy: 0.9100\n",
            "Epoch 50/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.1619 - accuracy: 0.9100\n",
            "Epoch 51/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1583 - accuracy: 0.9100\n",
            "Epoch 52/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1546 - accuracy: 0.9100\n",
            "Epoch 53/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1488 - accuracy: 0.9100\n",
            "Epoch 54/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1439 - accuracy: 0.9100\n",
            "Epoch 55/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1393 - accuracy: 0.9300\n",
            "Epoch 56/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1348 - accuracy: 0.9300\n",
            "Epoch 57/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1300 - accuracy: 0.9100\n",
            "Epoch 58/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1248 - accuracy: 0.9100\n",
            "Epoch 59/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1204 - accuracy: 0.9300\n",
            "Epoch 60/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1156 - accuracy: 0.9500\n",
            "Epoch 61/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1116 - accuracy: 0.9600\n",
            "Epoch 62/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1067 - accuracy: 0.9700\n",
            "Epoch 63/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1022 - accuracy: 0.9700\n",
            "Epoch 64/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0977 - accuracy: 0.9700\n",
            "Epoch 65/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0941 - accuracy: 0.9700\n",
            "Epoch 66/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0893 - accuracy: 0.9800\n",
            "Epoch 67/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0853 - accuracy: 0.9800\n",
            "Epoch 68/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0826 - accuracy: 0.9800\n",
            "Epoch 69/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0781 - accuracy: 0.9800\n",
            "Epoch 70/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0743 - accuracy: 0.9800\n",
            "Epoch 71/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0708 - accuracy: 0.9800\n",
            "Epoch 72/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0677 - accuracy: 0.9800\n",
            "Epoch 73/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0643 - accuracy: 0.9800\n",
            "Epoch 74/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0606 - accuracy: 0.9900\n",
            "Epoch 75/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0577 - accuracy: 1.0000\n",
            "Epoch 76/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0549 - accuracy: 1.0000\n",
            "Epoch 77/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0524 - accuracy: 1.0000\n",
            "Epoch 78/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0490 - accuracy: 1.0000\n",
            "Epoch 79/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0463 - accuracy: 1.0000\n",
            "Epoch 80/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0441 - accuracy: 1.0000\n",
            "Epoch 81/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0420 - accuracy: 1.0000\n",
            "Epoch 82/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0396 - accuracy: 1.0000\n",
            "Epoch 83/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0375 - accuracy: 1.0000\n",
            "Epoch 84/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0355 - accuracy: 1.0000\n",
            "Epoch 85/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0339 - accuracy: 1.0000\n",
            "Epoch 86/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0321 - accuracy: 1.0000\n",
            "Epoch 87/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0306 - accuracy: 1.0000\n",
            "Epoch 88/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0290 - accuracy: 1.0000\n",
            "Epoch 89/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0279 - accuracy: 1.0000\n",
            "Epoch 90/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0264 - accuracy: 1.0000\n",
            "Epoch 91/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0253 - accuracy: 1.0000\n",
            "Epoch 92/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0241 - accuracy: 1.0000\n",
            "Epoch 93/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0229 - accuracy: 1.0000\n",
            "Epoch 94/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0221 - accuracy: 1.0000\n",
            "Epoch 95/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0212 - accuracy: 1.0000\n",
            "Epoch 96/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0203 - accuracy: 1.0000\n",
            "Epoch 97/200\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.0195 - accuracy: 1.0000\n",
            "Epoch 98/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0187 - accuracy: 1.0000\n",
            "Epoch 99/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0179 - accuracy: 1.0000\n",
            "Epoch 100/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0173 - accuracy: 1.0000\n",
            "Epoch 101/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0166 - accuracy: 1.0000\n",
            "Epoch 102/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0160 - accuracy: 1.0000\n",
            "Epoch 103/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0155 - accuracy: 1.0000\n",
            "Epoch 104/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0150 - accuracy: 1.0000\n",
            "Epoch 105/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0144 - accuracy: 1.0000\n",
            "Epoch 106/200\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.0138 - accuracy: 1.0000\n",
            "Epoch 107/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0134 - accuracy: 1.0000\n",
            "Epoch 108/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0130 - accuracy: 1.0000\n",
            "Epoch 109/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0125 - accuracy: 1.0000\n",
            "Epoch 110/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0122 - accuracy: 1.0000\n",
            "Epoch 111/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0117 - accuracy: 1.0000\n",
            "Epoch 112/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0114 - accuracy: 1.0000\n",
            "Epoch 113/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0110 - accuracy: 1.0000\n",
            "Epoch 114/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0107 - accuracy: 1.0000\n",
            "Epoch 115/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0104 - accuracy: 1.0000\n",
            "Epoch 116/200\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0101 - accuracy: 1.0000\n",
            "Epoch 117/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0100 - accuracy: 1.0000\n",
            "Epoch 118/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0096 - accuracy: 1.0000\n",
            "Epoch 119/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0093 - accuracy: 1.0000\n",
            "Epoch 120/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0090 - accuracy: 1.0000\n",
            "Epoch 121/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0088 - accuracy: 1.0000\n",
            "Epoch 122/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0086 - accuracy: 1.0000\n",
            "Epoch 123/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0084 - accuracy: 1.0000\n",
            "Epoch 124/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0082 - accuracy: 1.0000\n",
            "Epoch 125/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0080 - accuracy: 1.0000\n",
            "Epoch 126/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0078 - accuracy: 1.0000\n",
            "Epoch 127/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0076 - accuracy: 1.0000\n",
            "Epoch 128/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0074 - accuracy: 1.0000\n",
            "Epoch 129/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0073 - accuracy: 1.0000\n",
            "Epoch 130/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0071 - accuracy: 1.0000\n",
            "Epoch 131/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0070 - accuracy: 1.0000\n",
            "Epoch 132/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0068 - accuracy: 1.0000\n",
            "Epoch 133/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0066 - accuracy: 1.0000\n",
            "Epoch 134/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0065 - accuracy: 1.0000\n",
            "Epoch 135/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0064 - accuracy: 1.0000\n",
            "Epoch 136/200\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.0062 - accuracy: 1.0000\n",
            "Epoch 137/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.0061 - accuracy: 1.0000\n",
            "Epoch 138/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0059 - accuracy: 1.0000\n",
            "Epoch 139/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0058 - accuracy: 1.0000\n",
            "Epoch 140/200\n",
            "2/2 [==============================] - 0s 18ms/step - loss: 0.0057 - accuracy: 1.0000\n",
            "Epoch 141/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0056 - accuracy: 1.0000\n",
            "Epoch 142/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0055 - accuracy: 1.0000\n",
            "Epoch 143/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0054 - accuracy: 1.0000\n",
            "Epoch 144/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0053 - accuracy: 1.0000\n",
            "Epoch 145/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0052 - accuracy: 1.0000\n",
            "Epoch 146/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0051 - accuracy: 1.0000\n",
            "Epoch 147/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0050 - accuracy: 1.0000\n",
            "Epoch 148/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0049 - accuracy: 1.0000\n",
            "Epoch 149/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.0048 - accuracy: 1.0000\n",
            "Epoch 150/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0047 - accuracy: 1.0000\n",
            "Epoch 151/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0046 - accuracy: 1.0000\n",
            "Epoch 152/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0045 - accuracy: 1.0000\n",
            "Epoch 153/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0045 - accuracy: 1.0000\n",
            "Epoch 154/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0044 - accuracy: 1.0000\n",
            "Epoch 155/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0043 - accuracy: 1.0000\n",
            "Epoch 156/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0042 - accuracy: 1.0000\n",
            "Epoch 157/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0041 - accuracy: 1.0000\n",
            "Epoch 158/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0041 - accuracy: 1.0000\n",
            "Epoch 159/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0040 - accuracy: 1.0000\n",
            "Epoch 160/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0039 - accuracy: 1.0000\n",
            "Epoch 161/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0039 - accuracy: 1.0000\n",
            "Epoch 162/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0038 - accuracy: 1.0000\n",
            "Epoch 163/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0037 - accuracy: 1.0000\n",
            "Epoch 164/200\n",
            "2/2 [==============================] - 0s 23ms/step - loss: 0.0037 - accuracy: 1.0000\n",
            "Epoch 165/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0036 - accuracy: 1.0000\n",
            "Epoch 166/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0036 - accuracy: 1.0000\n",
            "Epoch 167/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0035 - accuracy: 1.0000\n",
            "Epoch 168/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.0035 - accuracy: 1.0000\n",
            "Epoch 169/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0034 - accuracy: 1.0000\n",
            "Epoch 170/200\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.0034 - accuracy: 1.0000\n",
            "Epoch 171/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0033 - accuracy: 1.0000\n",
            "Epoch 172/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0033 - accuracy: 1.0000\n",
            "Epoch 173/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0032 - accuracy: 1.0000\n",
            "Epoch 174/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0032 - accuracy: 1.0000\n",
            "Epoch 175/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0031 - accuracy: 1.0000\n",
            "Epoch 176/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0031 - accuracy: 1.0000\n",
            "Epoch 177/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0030 - accuracy: 1.0000\n",
            "Epoch 178/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0030 - accuracy: 1.0000\n",
            "Epoch 179/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0030 - accuracy: 1.0000\n",
            "Epoch 180/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0029 - accuracy: 1.0000\n",
            "Epoch 181/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0029 - accuracy: 1.0000\n",
            "Epoch 182/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0028 - accuracy: 1.0000\n",
            "Epoch 183/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.0028 - accuracy: 1.0000\n",
            "Epoch 184/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0028 - accuracy: 1.0000\n",
            "Epoch 185/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 186/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0027 - accuracy: 1.0000\n",
            "Epoch 187/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0026 - accuracy: 1.0000\n",
            "Epoch 188/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0026 - accuracy: 1.0000\n",
            "Epoch 189/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0026 - accuracy: 1.0000\n",
            "Epoch 190/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 191/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 192/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 193/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.0025 - accuracy: 1.0000\n",
            "Epoch 194/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.0024 - accuracy: 1.0000\n",
            "Epoch 195/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.0024 - accuracy: 1.0000\n",
            "Epoch 196/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.0024 - accuracy: 1.0000\n",
            "Epoch 197/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0023 - accuracy: 1.0000\n",
            "Epoch 198/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.0023 - accuracy: 1.0000\n",
            "Epoch 199/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.0023 - accuracy: 1.0000\n",
            "Epoch 200/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.0023 - accuracy: 1.0000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a4IYwbm-jIbI",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 354
        },
        "outputId": "bf3ec328-d3ba-40a5-9a7f-a28acd4a4b74"
      },
      "source": [
        "eval_model(model, X, y)\n",
        "show_descision_boundary(limits=None, clf=model, binary=False,\n",
        "                                X=X,\n",
        "                                y=y, \n",
        "                                n_lines=50,\n",
        "                                show_lines=False)"
      ],
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0022 - accuracy: 1.0000\n",
            "Perfect!\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aT83SJ8lpFzw"
      },
      "source": [
        "**1c**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GQujffe6jXRo",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 320
        },
        "outputId": "eb7913b6-c8ec-48d7-f283-3e7e4ac825df"
      },
      "source": [
        "from sklearn.datasets import make_circles\n",
        "\n",
        "X, y = make_circles(factor=0.5, noise=0.05)\n",
        "show_data(X, y)"
      ],
      "execution_count": 81,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ThMesjHJkIkY",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "63825038-a809-4d2d-b5ae-0a6df0fffea5"
      },
      "source": [
        "model = tf.keras.Sequential()\n",
        "# <YOUR CODE STARTS HERE>\n",
        "# 1. добавьте скрытые слои в модель (в первом из них не забудьте про input_shape)\n",
        "#    Постарайтесь сделать количество слоев и нейронов минимальным.\n",
        "# 2. добавьте выходной слой он должен быть без активации\n",
        "# 3. определите оптимизатор \n",
        "# 4. определите лосс (не забудьте указать from_logits=True!!) \n",
        "\n",
        "model = tf.keras.Sequential()\n",
        "model.add(tf.keras.layers.Dense(4, input_shape=(2,), activation=\"relu\"))\n",
        "model.add(tf.keras.layers.Dense(1))\n",
        "\n",
        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.01)\n",
        "loss = tf.keras.losses.BinaryCrossentropy(from_logits=True)\n",
        "\n",
        "# <YOUR CODE ENDS HERE>\n",
        "\n",
        "model.compile(optimizer=optimizer, loss=loss, metrics=[\"accuracy\"]) \n",
        "history = model.fit(X, y, epochs=200, verbose=1, batch_size=64, shuffle=True)"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.6730 - accuracy: 0.5000\n",
            "Epoch 2/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6690 - accuracy: 0.5000\n",
            "Epoch 3/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6660 - accuracy: 0.5000\n",
            "Epoch 4/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6636 - accuracy: 0.5000\n",
            "Epoch 5/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6600 - accuracy: 0.5000\n",
            "Epoch 6/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.6566 - accuracy: 0.5000\n",
            "Epoch 7/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6532 - accuracy: 0.5000\n",
            "Epoch 8/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6495 - accuracy: 0.5000\n",
            "Epoch 9/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6456 - accuracy: 0.5000\n",
            "Epoch 10/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6413 - accuracy: 0.5000\n",
            "Epoch 11/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6373 - accuracy: 0.5000\n",
            "Epoch 12/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6324 - accuracy: 0.5000\n",
            "Epoch 13/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6280 - accuracy: 0.5000\n",
            "Epoch 14/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.6242 - accuracy: 0.5000\n",
            "Epoch 15/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6209 - accuracy: 0.5000\n",
            "Epoch 16/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.6183 - accuracy: 0.5000\n",
            "Epoch 17/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.6154 - accuracy: 0.5000\n",
            "Epoch 18/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6127 - accuracy: 0.5000\n",
            "Epoch 19/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6104 - accuracy: 0.5000\n",
            "Epoch 20/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.6075 - accuracy: 0.5000\n",
            "Epoch 21/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.6049 - accuracy: 0.5000\n",
            "Epoch 22/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.6020 - accuracy: 0.5000\n",
            "Epoch 23/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.5988 - accuracy: 0.5000\n",
            "Epoch 24/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5959 - accuracy: 0.5000\n",
            "Epoch 25/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.5929 - accuracy: 0.5000\n",
            "Epoch 26/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.5897 - accuracy: 0.5000\n",
            "Epoch 27/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.5865 - accuracy: 0.5000\n",
            "Epoch 28/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5831 - accuracy: 0.5000\n",
            "Epoch 29/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5797 - accuracy: 0.5000\n",
            "Epoch 30/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5761 - accuracy: 0.5200\n",
            "Epoch 31/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5730 - accuracy: 0.5300\n",
            "Epoch 32/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5692 - accuracy: 0.5400\n",
            "Epoch 33/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5659 - accuracy: 0.5400\n",
            "Epoch 34/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5626 - accuracy: 0.5400\n",
            "Epoch 35/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5590 - accuracy: 0.5700\n",
            "Epoch 36/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5555 - accuracy: 0.5800\n",
            "Epoch 37/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.5518 - accuracy: 0.6100\n",
            "Epoch 38/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5480 - accuracy: 0.6300\n",
            "Epoch 39/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5445 - accuracy: 0.6400\n",
            "Epoch 40/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5406 - accuracy: 0.6700\n",
            "Epoch 41/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5365 - accuracy: 0.6700\n",
            "Epoch 42/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.5328 - accuracy: 0.6800\n",
            "Epoch 43/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.5289 - accuracy: 0.6900\n",
            "Epoch 44/200\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.5251 - accuracy: 0.7000\n",
            "Epoch 45/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5213 - accuracy: 0.7000\n",
            "Epoch 46/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.5173 - accuracy: 0.7000\n",
            "Epoch 47/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.5135 - accuracy: 0.7100\n",
            "Epoch 48/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.5096 - accuracy: 0.7300\n",
            "Epoch 49/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.5056 - accuracy: 0.7400\n",
            "Epoch 50/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.5014 - accuracy: 0.7500\n",
            "Epoch 51/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.4975 - accuracy: 0.7600\n",
            "Epoch 52/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.4935 - accuracy: 0.8000\n",
            "Epoch 53/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.4890 - accuracy: 0.8100\n",
            "Epoch 54/200\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.4855 - accuracy: 0.8100\n",
            "Epoch 55/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4811 - accuracy: 0.8200\n",
            "Epoch 56/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.4773 - accuracy: 0.8200\n",
            "Epoch 57/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.4731 - accuracy: 0.8200\n",
            "Epoch 58/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.4690 - accuracy: 0.8300\n",
            "Epoch 59/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4648 - accuracy: 0.8300\n",
            "Epoch 60/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.4607 - accuracy: 0.8300\n",
            "Epoch 61/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.4566 - accuracy: 0.8300\n",
            "Epoch 62/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.4522 - accuracy: 0.8300\n",
            "Epoch 63/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.4481 - accuracy: 0.8400\n",
            "Epoch 64/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.4436 - accuracy: 0.8600\n",
            "Epoch 65/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4394 - accuracy: 0.8600\n",
            "Epoch 66/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4352 - accuracy: 0.8700\n",
            "Epoch 67/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.4308 - accuracy: 0.8800\n",
            "Epoch 68/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.4265 - accuracy: 0.8800\n",
            "Epoch 69/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.4222 - accuracy: 0.8800\n",
            "Epoch 70/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4181 - accuracy: 0.8900\n",
            "Epoch 71/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.4141 - accuracy: 0.8900\n",
            "Epoch 72/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.4096 - accuracy: 0.8900\n",
            "Epoch 73/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.4053 - accuracy: 0.8900\n",
            "Epoch 74/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.4008 - accuracy: 0.8900\n",
            "Epoch 75/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.3964 - accuracy: 0.8900\n",
            "Epoch 76/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.3919 - accuracy: 0.9100\n",
            "Epoch 77/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.3877 - accuracy: 0.9200\n",
            "Epoch 78/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.3834 - accuracy: 0.9300\n",
            "Epoch 79/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3789 - accuracy: 0.9400\n",
            "Epoch 80/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.3746 - accuracy: 0.9400\n",
            "Epoch 81/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.3709 - accuracy: 0.9400\n",
            "Epoch 82/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3661 - accuracy: 0.9400\n",
            "Epoch 83/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.3615 - accuracy: 0.9400\n",
            "Epoch 84/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.3576 - accuracy: 0.9400\n",
            "Epoch 85/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3532 - accuracy: 0.9400\n",
            "Epoch 86/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3493 - accuracy: 0.9400\n",
            "Epoch 87/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.3452 - accuracy: 0.9400\n",
            "Epoch 88/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.3415 - accuracy: 0.9400\n",
            "Epoch 89/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3377 - accuracy: 0.9400\n",
            "Epoch 90/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3334 - accuracy: 0.9400\n",
            "Epoch 91/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.3296 - accuracy: 0.9400\n",
            "Epoch 92/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3256 - accuracy: 0.9400\n",
            "Epoch 93/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.3218 - accuracy: 0.9400\n",
            "Epoch 94/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3178 - accuracy: 0.9600\n",
            "Epoch 95/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.3141 - accuracy: 0.9600\n",
            "Epoch 96/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.3103 - accuracy: 0.9700\n",
            "Epoch 97/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.3063 - accuracy: 0.9800\n",
            "Epoch 98/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.3026 - accuracy: 0.9800\n",
            "Epoch 99/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2991 - accuracy: 0.9900\n",
            "Epoch 100/200\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.2954 - accuracy: 0.9900\n",
            "Epoch 101/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2918 - accuracy: 0.9900\n",
            "Epoch 102/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2881 - accuracy: 0.9900\n",
            "Epoch 103/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.2847 - accuracy: 0.9900\n",
            "Epoch 104/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.2816 - accuracy: 0.9900\n",
            "Epoch 105/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.2781 - accuracy: 0.9900\n",
            "Epoch 106/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2748 - accuracy: 0.9900\n",
            "Epoch 107/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2715 - accuracy: 0.9900\n",
            "Epoch 108/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2683 - accuracy: 0.9900\n",
            "Epoch 109/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2652 - accuracy: 1.0000\n",
            "Epoch 110/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.2616 - accuracy: 1.0000\n",
            "Epoch 111/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2587 - accuracy: 1.0000\n",
            "Epoch 112/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2556 - accuracy: 1.0000\n",
            "Epoch 113/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2525 - accuracy: 1.0000\n",
            "Epoch 114/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2496 - accuracy: 1.0000\n",
            "Epoch 115/200\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 0.2468 - accuracy: 1.0000\n",
            "Epoch 116/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.2436 - accuracy: 1.0000\n",
            "Epoch 117/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.2407 - accuracy: 1.0000\n",
            "Epoch 118/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.2379 - accuracy: 1.0000\n",
            "Epoch 119/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2353 - accuracy: 1.0000\n",
            "Epoch 120/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2324 - accuracy: 1.0000\n",
            "Epoch 121/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2296 - accuracy: 1.0000\n",
            "Epoch 122/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2270 - accuracy: 1.0000\n",
            "Epoch 123/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2242 - accuracy: 1.0000\n",
            "Epoch 124/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2214 - accuracy: 1.0000\n",
            "Epoch 125/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2190 - accuracy: 1.0000\n",
            "Epoch 126/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.2162 - accuracy: 1.0000\n",
            "Epoch 127/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2134 - accuracy: 1.0000\n",
            "Epoch 128/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.2107 - accuracy: 1.0000\n",
            "Epoch 129/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.2082 - accuracy: 1.0000\n",
            "Epoch 130/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.2057 - accuracy: 1.0000\n",
            "Epoch 131/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.2036 - accuracy: 1.0000\n",
            "Epoch 132/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.2007 - accuracy: 1.0000\n",
            "Epoch 133/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1984 - accuracy: 1.0000\n",
            "Epoch 134/200\n",
            "2/2 [==============================] - 0s 11ms/step - loss: 0.1961 - accuracy: 1.0000\n",
            "Epoch 135/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1938 - accuracy: 1.0000\n",
            "Epoch 136/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1916 - accuracy: 1.0000\n",
            "Epoch 137/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1890 - accuracy: 1.0000\n",
            "Epoch 138/200\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.1867 - accuracy: 1.0000\n",
            "Epoch 139/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1845 - accuracy: 1.0000\n",
            "Epoch 140/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1824 - accuracy: 1.0000\n",
            "Epoch 141/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1803 - accuracy: 1.0000\n",
            "Epoch 142/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1783 - accuracy: 1.0000\n",
            "Epoch 143/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1765 - accuracy: 1.0000\n",
            "Epoch 144/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1745 - accuracy: 1.0000\n",
            "Epoch 145/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.1726 - accuracy: 1.0000\n",
            "Epoch 146/200\n",
            "2/2 [==============================] - 0s 12ms/step - loss: 0.1707 - accuracy: 1.0000\n",
            "Epoch 147/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1689 - accuracy: 1.0000\n",
            "Epoch 148/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1669 - accuracy: 1.0000\n",
            "Epoch 149/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1652 - accuracy: 1.0000\n",
            "Epoch 150/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1631 - accuracy: 1.0000\n",
            "Epoch 151/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1613 - accuracy: 1.0000\n",
            "Epoch 152/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1598 - accuracy: 1.0000\n",
            "Epoch 153/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1580 - accuracy: 1.0000\n",
            "Epoch 154/200\n",
            "2/2 [==============================] - 0s 19ms/step - loss: 0.1562 - accuracy: 1.0000\n",
            "Epoch 155/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1545 - accuracy: 1.0000\n",
            "Epoch 156/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1532 - accuracy: 1.0000\n",
            "Epoch 157/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1514 - accuracy: 1.0000\n",
            "Epoch 158/200\n",
            "2/2 [==============================] - 0s 15ms/step - loss: 0.1497 - accuracy: 1.0000\n",
            "Epoch 159/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1485 - accuracy: 1.0000\n",
            "Epoch 160/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1467 - accuracy: 1.0000\n",
            "Epoch 161/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1452 - accuracy: 1.0000\n",
            "Epoch 162/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1439 - accuracy: 1.0000\n",
            "Epoch 163/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1422 - accuracy: 1.0000\n",
            "Epoch 164/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1407 - accuracy: 1.0000\n",
            "Epoch 165/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1394 - accuracy: 1.0000\n",
            "Epoch 166/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1379 - accuracy: 1.0000\n",
            "Epoch 167/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1365 - accuracy: 1.0000\n",
            "Epoch 168/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1351 - accuracy: 1.0000\n",
            "Epoch 169/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1339 - accuracy: 1.0000\n",
            "Epoch 170/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1325 - accuracy: 1.0000\n",
            "Epoch 171/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1312 - accuracy: 1.0000\n",
            "Epoch 172/200\n",
            "2/2 [==============================] - 0s 20ms/step - loss: 0.1300 - accuracy: 1.0000\n",
            "Epoch 173/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1286 - accuracy: 1.0000\n",
            "Epoch 174/200\n",
            "2/2 [==============================] - 0s 28ms/step - loss: 0.1274 - accuracy: 1.0000\n",
            "Epoch 175/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1262 - accuracy: 1.0000\n",
            "Epoch 176/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1250 - accuracy: 1.0000\n",
            "Epoch 177/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1238 - accuracy: 1.0000\n",
            "Epoch 178/200\n",
            "2/2 [==============================] - 0s 14ms/step - loss: 0.1226 - accuracy: 1.0000\n",
            "Epoch 179/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1215 - accuracy: 1.0000\n",
            "Epoch 180/200\n",
            "2/2 [==============================] - 0s 16ms/step - loss: 0.1203 - accuracy: 1.0000\n",
            "Epoch 181/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1193 - accuracy: 1.0000\n",
            "Epoch 182/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1182 - accuracy: 1.0000\n",
            "Epoch 183/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1171 - accuracy: 1.0000\n",
            "Epoch 184/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1161 - accuracy: 1.0000\n",
            "Epoch 185/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1149 - accuracy: 1.0000\n",
            "Epoch 186/200\n",
            "2/2 [==============================] - 0s 9ms/step - loss: 0.1138 - accuracy: 1.0000\n",
            "Epoch 187/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1128 - accuracy: 1.0000\n",
            "Epoch 188/200\n",
            "2/2 [==============================] - 0s 8ms/step - loss: 0.1117 - accuracy: 1.0000\n",
            "Epoch 189/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1109 - accuracy: 1.0000\n",
            "Epoch 190/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1099 - accuracy: 1.0000\n",
            "Epoch 191/200\n",
            "2/2 [==============================] - 0s 6ms/step - loss: 0.1089 - accuracy: 1.0000\n",
            "Epoch 192/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.1080 - accuracy: 1.0000\n",
            "Epoch 193/200\n",
            "2/2 [==============================] - 0s 7ms/step - loss: 0.1071 - accuracy: 1.0000\n",
            "Epoch 194/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1061 - accuracy: 1.0000\n",
            "Epoch 195/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1051 - accuracy: 1.0000\n",
            "Epoch 196/200\n",
            "2/2 [==============================] - 0s 17ms/step - loss: 0.1042 - accuracy: 1.0000\n",
            "Epoch 197/200\n",
            "2/2 [==============================] - 0s 4ms/step - loss: 0.1033 - accuracy: 1.0000\n",
            "Epoch 198/200\n",
            "2/2 [==============================] - 0s 5ms/step - loss: 0.1024 - accuracy: 1.0000\n",
            "Epoch 199/200\n",
            "2/2 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 1.0000\n",
            "Epoch 200/200\n",
            "2/2 [==============================] - 0s 13ms/step - loss: 0.1006 - accuracy: 1.0000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YOXDdx84kMkJ",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 371
        },
        "outputId": "8dffdee4-4b98-4d1b-d1e8-2ad3628730b6"
      },
      "source": [
        "accuracy = eval_model(model, X, y)\n",
        "if model.count_params() < 18 and accuracy == 1.0:\n",
        "    print(\"Bonus points!\")\n",
        "else:\n",
        "    print(f\"Model has {model.count_params()} parameters\")\n",
        "\n",
        "show_descision_boundary(limits=None, clf=model, binary=False,\n",
        "                                X=X,\n",
        "                                y=y, \n",
        "                                n_lines=50,\n",
        "                                show_lines=False)"
      ],
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "4/4 [==============================] - 0s 3ms/step - loss: 0.0999 - accuracy: 1.0000\n",
            "Perfect!\n",
            "Bonus points!\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUQAAAEvCAYAAAA92bhfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOxdd5wTZf5+JtnsJruUA6SDgCKggpW7s+spqGBBz4p6dhFFPfXUs9zv7B3Ps9dTsYsNsXc9T7EAigiKooCA0uuW7G428/vj2deZTN6ZeSeZJJNsns8nn+xOJslkyjPf7/Ntmq7rKKOMMsooAwgVegPKKKOMMoKCMiGWUUYZZbSiTIhllFFGGa0oE2IZZZRRRivKhFhGGWWU0YoyIZZRRhlltKKi0Btgh03atdP7d+liLKitBdq1S10p18tqa/kchGV1dXxu3z59mXm9+no+19QYyxoaUv8Xy6qrU5fF40AsBt/R0GD/WjTq//fJIPaLDLn4zXYQx8zPbTDvX+txdlvfeg7IEI8bf6tsZ2Oj8bfK8W1qynz9qir39Zubjb8rKzFjzpxVuq53la0aWELs36ULpl92Gf/573/5vMcexgrZLjP/n49lqtv20Ud83n13Y9lnn6UvmzGDzzvvbCybNYvPf/iDseybb/i8ww7Gsu++4/M22xjL5s/n81ZbIW+YO9f+tYED87cdX38tXz5kSP62YeZM+9eGDrV/TXZ8nSA79k7wel4sWMDnwYPV1l+8mM+qx/uXX/jcv7/a+itW8LlPn98WaVtuuchu9cAS4m/IBRnafUemy1QI0rodTsvMxOe0LJdkKE5sMwYMSF+WDZwuMjuyzAVR2pGDHVEC/pOlHaHNnGkcRy/vs6JMhEoIPiEC7kSSK9LMN7l+9FE68X32WfqyGTNSyVDATIYCsgtGdlHIyNB6cs+bl/4+wH+itG6PGfm0Kp2IUhCMFfkiSlWUidATgk2IubC+MlmWD3IVrrIZwlU2Q7jKZgjr0Ixvvkm/mL77Lv3CmD9ffvLLTnC7k96OKIH8WZVz5xoXsxWlSJRuMG+HChma950KGZo9CBUyFEQIqB0PQYSAGhkKIgQyJkMgyIQoAgtm5MO1zTXh5lM3NEN2ocoIROYqu8GJKO0+ry0RJZA/97tMhM5Ys8bx5eASIhBc1zYX5Bok3VDV/XFDkIkSKC33u0yEznAhQoHgEqI1JQYIhmvb1nTDxYuBvn3T35sNnC6osvtNqBJlmQidYSbCHj1cVw8uIZoRFNe2GHRDK7LRDcXJbT7JrcgXWQbBqgwCUQLp7nc5YCKHIEMFIhQIPiEGzbUNum4oc5XNUNUNVU7u+fPtybItESUQDPfbDmUiVEawCTFIrm0uchCDqhuqnuB2rweBKIG2535bUXaPPSPYhAgEw7X1m3DtXOUg6IZe7/YyuFmVdmhLVmUuibJMhBkjuIRYWxsM1zZXuqHMVTajULohkFvLJchWZRCIEsjc/S4ToTM2bHBdJbiEaI0yB628rpR1QzPMJ7UZvXrJl2eKIBMlEHz32+29ZrRlItxkE8dVg0uIZgRNS8y3biiQb93QTvxeuLDwRAmU3e9M0NYCJoIMXYhQoDgIEWjbuuGsWf7phgKZkqHdMoGFC+1fa0tWZT6JUgVlIlRC8AmxrBumL8tGN1ywQF03VD25Vd4TBKsyCEQJ5Nf9LhOhJwSbEItNN5QtK1bdMBMydEKQiRIoPfe7rBNmhOASomjuUEy6oR3hFqNuaIb5ZO7WLf31bFB2vw34QZRlInSGXcfyVgSXEIFguLttXTe0ujhLlqR/rkC+yDIIVmUQiBKQu99lIkyHCxEKBJcQZWk3xaQl5ks3tCIXuqH5hLY7uZcsSb0IzGhLRAnYu99+EyXgvTNRWybCTp1cVw8uIZpRbLXLhZ6LYkY2+YYrVqiL30EmSsDe/fabKAFnq9IOuSBLK9pawESQoQIRCgSfEMu6IZ/znW9oR2Je4XRx2LnffhMl4GxV2qFU3O8yESoj+IQIlIZuaEUx6Ya5gpNVaYey+02oEGWpECGQE/dYhmATYinphjJX2Ywg64aAc8fhzp3tX3NDSwvw00+cFdy1K9CzZ9n9NiMT91vl/Wa0MZ3QCcElxGxnqpR1w1RkoxuuWWN/Ii9bZk+WTkTZ0AA88gjw2GMcbB4KcaD4wIHAmWcCe++d/p5CuN/LlwOTJwNffAEkk7zJHH00sMUW6esGxf32ahECbYMIGxtdVwkuIbZrV1w5iIB/uqFAEHRDt1kUTif6smXy5ZEIcMwx3C7rSTpnDnDBBcBxxwHnn+/83Wb47X4nk8DNNwPPPcf/m5r4PHcu8MorPDa33AJUVxvvCbr7LdCWibB9e8dVfSFETdMeAnAggBW6rg+VvK4BuA3AaAD1AE7UdX2m8hcEvXbZb91QNkIUyEw3FMhEN8yi87Dj+8aOBRYtokUoQ0MDLccttwRGjcrsuwUydb+vuw54+WWDCAVaWvj44gvgtNNo5UYiztsQFPc7H6M9gxYwEWToQoQCflmIjwC4E8CjNq+PArBF6+OPAO5pfVZHWTdMhapuuHhx5rohkHXDzTQsXMjfbUeGAvE4cNttwB9bT5NsdEoZnC7wjz4Cpk51drGam3mszjkHuPrqzMvF8hX9DmrAJCBEKOALIeq6/l9N0/o7rDIGwKO6rusAPtU07XeapvXUdf1X1w8Psm4o0FZ0Q2uDzQ4d0j/LDrW1wOuvA08/7U6G5u1av54kL9zvhQuBt97iax06AAcdxP2laerb4oY33gASCbV1P/kE+POfgccfBzbd1L9t8Mv9LhOhJ+RLQ+wNwHx1Lmld5kyIQdcNZSNEgdLUDWUn8qpV6esBqUTZ0gLceCPwxBMMnNTXy98jQ0UFt3nwYL53wgRqjC0tJCxNoyXXqRNw7bXAoEGp78/Uqvz4Y36HKtavB04+mW5+OJz6WhDShMo6oTICFVTRNG0cgHEAsKk4mcu6YfqyfOuGdnd1uxNbEKWuAxdeSIKJx+XruqGqCli9Ghgzhs9motJ16o0NDSTLZ54x9k2m0W9A3YI1o76edcX77mssC0qakBPaGhG63OjyRYhLAZgzSfu0LkuBruv3A7gfAIZ36aIHNgexkLqhFbnUDTduBF58kdbdhg1ATQ1wyCGMADtd0OLEnzoVmDYtczJsbgaGDQP++U9eiE4nc3099by336blmE2aUM+e7tF1K+rqaCGaCTGTNCEgN5U6VpSJUIp8EeJUAGdpmvY0GExZ76ofqs5UKVTtcqF0Q5mrbIYX3dAKs2748cfA+PGGFSZw113AnXdyn2y+OXWzgw+Wn8R33eXNRTYjHCa5JJPAm2+qaXrLlnEfb7ed/ToqaUKHHWYki3uBncsqQ6GSz9syEcZirqv7lXbzFIC9AGyiadoSAJcDiACAruv3AngNTLmZD6bdnOTpC8q6IZ/91g1lrjJAUhk3Tm7ZiTSUd9/lIxZjlPXQQ4FrrqGLCwBr1wI//JD+fhWEQsDvfkfL8L33qCUqJNUiHgfeeceZEJ0gLvijjyaZx+O8IajCLf1GBbkkyrYWMBFkqECEAn5Fmce6vK4DmJDRhxerbmhFMemGV16p7uYKK2rKFODHH6njRSJ0tyMRb3qcpvHk7dGDUdvu3WkpqpKSrtsHerygqgp48kngyCMZMFFBRQVThMS+zGeakJv7XSZCZQQqqGKLYtQNZa6yGfnUDevqWF3x5JPAypW84HfdFTjxRINoxcm/YIHzzA87xOP8Tf/5D/CXvwCVld7IsKYG2Gcf6pM77WSk0WyyCS1GFUQi/B1283e9pAkNHAi89hqrZWQ3PSvCYeCMM4z9aFelA+SPLIVVWSZCZQSbEIOWg1iMuuGcOdQCm5sNa66uDnj1VQYg9t2XJWjm7VUlICsaGoCHHwbOPpsn/Y47Ap9+6v6+mhrgqqtokVmx667q2xMKsSTQLfotg4wse/Tg7/nzn3ks7HTMWIxEbs5D9Lv2OxN4JUKgTeiETgguIXqdqWKG31piseqGFRUsL5O1T08madW99RbL1P7xDy6vrfWmm1mxdi0DEptvDpx7LvPz3AIrmkbr0LqdNTW0+k4+Gbj3Xmc3PhKhZelEAk5E6WRVPv00byqzZlFDFRdgVRX31YknAn/7m/33mhEEohQoE2EagkuIQDByEItNNxTo3x+49FL3SGk8zrSa00/nBd6vX+YWIkAS3riRf//pT8AJJ7De1247YjF+f8+eqcvXrjUuqFNPBT7/HJg5U06KkQjff8cdmW2zClHecw/w/fckx/nzgWiU1uvYsVlNeUNLC/D++8D999OaTyb5W045hSlONTX+u99lIrRFcAnRmnYDlHVDFd1QjBCtrWXENZlM/ywrNA146ilaOnvvzQiv4lCeNDQ3p16kl18ObLYZO8fU1xvbk0wCW28NXH89sO226Z9jvZCefZaEd889huuqabw4xoyhhZZlL7w0yBLRd9mFf5vd70xLGmtruc/nzUu1ohct4n657TYel803l7/fq1VZJkJXBJcQzQiKbihbFkTdEGApVySilq7S2Eit74IL+P8ll7DCxGseHkAdrXt343urqoDjj6fG9sknxnYNH25/octQUQGcdx71yU8/ZdVKu3Z0k2tquM7atfbvF+v4hUzdbwFdp5wxZ056Rx2A+z4eB446irKGjOAycb/bcMBEBcEnxCDphrIRokBmuqFArvINk0lvWqDZkhwzhik099zDE1f1c6qrSaTmk1y4zwDw+9/zkQ0qKoDddpO/Zncxmt1vKwpFlF9+CcyeLSdDAV3ndj/xBG8EqsimQ1EbJUKBYBNiKeuG332XG91QIBZzvtjMiERSt6+ujpriPvuwKuWdd7jcqVqkupqa12GHpS63O/E3brS3XkVyt18IIlE+84xarmdTEwNKZ56Z3jjCT7RB91iGYBMikP8cxGLWDc3o2JGW1AcfuFt4oRDdWjM6daI1N2kSyauujhfN//0f65MjEVqVoRAJ7PzzmYen2oYryEQJ5N79nj1b3fKOx5kaNX68scxLTqUTykSYguASot1MlXzULherbiggKhPOPtu9uUI0CowYYeTQ1dWln+iVlUCXLvz7xRe5HR9/TOtl0025b/yyXpwuGrP7bUUpWpVmPPIIm1d06KCuUzqhTIRSBJcQ7WaqmFHWDfnsVJr3738zHzCRSHd5YzFagf/+N/+XXegya61XL+CII/h+EXiRdRPx28ULslXplSiHDgV+/VXdStQ0znc5+WRnAnPrUVkmQkcElxDNKAbd0IpC6IaymtVhw1gX/NprnB4n0ju23poW5L77krhk4rhMALeK3HYncUODfcultkSUgNz9Pv54noeqNePxOPC//5EQnaBSpdNGAyYqKA5CBIKvG8pcZTO86IZWqOqGgLxSY/vt+bjsMrq5YgayFV7J0AlBJkrA3v32mygBOaGMGMFh8146AqkGyWTwkjzeBolQIPiE2BZ1Q5mrbIaX/oZWxOPpZCjTDQH5CZ7tSez0fru8x1xEV52sSjv4SZaaxkYYe+6p5jZHIqk3xdpaI4exTx9gwIDst6kNE6FAsAkxF7qhFaWoGwLOowDMUNUNvcwYyRROVqUditn9HjyYGu7nn7uvGw6zR+VPPwG3387mHBUVJNbmZlYDXXghLU+vaGM6oROCS4hemztk4yqbUWy6oUAmc1Ey1Q3zjVJ2vy+9lPXQTqQfjbKkEmDnnfXr0wNkc+cy7enMM0mcZthFv8tEmIbgEiJQ1g1VdEO7fneyagWZjuTmKjuRYWVl+rJsdC6vKAX3e+edgRtuAC6+mIRjrT2vrmYA7I472Cxj7Vr7+vR4nNVFO+9snOtO0W+gTIQWBJcQZTNV2ppuqOvAV1+xquHHH6kj7bMPU1422cTbCFErGcp0w8ZGdd1QRoZOy4FgkGUQ3e9DD+UI1bvv5vyYUIgW4IABwFlnsfrn449Zv+3WrKOhAbj1VuN8z6bhRRsiQoHgEqIZpaAbCqiSYadOwLHHsuuxebbHokXAQw9x7sdxxzFSKRB03dCOLINClIV0v3faiZZgYyOwbh23U+QOJpNsVKvagWj6dKbZZNOWrAQDJioIPiGWim44f766btijB8lw+fL0i1SQxzPP8GQ96yz+n0/d0MkKzARloiTat+dDRmSybAM7VFbSe8iEENsoEQoEnxAB/7TEYtENH3yQd3gnC62xEXjgAbaHEmk0BdANE0mjdrkilEWnbRmCTJRAft1vWX9QO7S02H//hg1M0J8yhTfHXr3Yk3HXXY33tBH3WIZgE2IuchAzGQWQT92wpYVdmVUuek2jpXjMMemv2bnKPuqGiaQGhFovomRLCjmakTeiBIJBlrmwKseMYUMIlTnXHTqwxtwqgzz/PHtdappB5nPnst69QweS5KBBzp9dokQokEWv+BzDrrlDpstUdcMZMwqnGw4cyPpW1Yu6sZFziwF1V9n6fisUdcMUMgT4t80jkdRsH76jslL+yCdiMfkD4P6VPdxwzDFq3c+jUabedOxouODt27PJ7KWXUo+2WrZ1dWwqu99+wM8/yz/XvJ3m31NiCLaFGGTd0AqvumEyyYTcKVMYOPnd75hUO3SoegstgOSZZ93QM5GFbKyiIFiVQbAoAXf3u2NH4MYbmZ5jt25lJRO0TzstdXlTE7uNO32HrvPcufJKVtAIlLhFaEVwCTGTmSr51A1lrrIZTrrh4sU8QdevTz1Jv/uORKka3dU0YIst0pfnUDf8jcDsSM4LgkyUQDDI0ux+H3ssj/nFF/NZ3OgiEZLmLrswGm39rJdfVisPTCbZBGTtWp4vJRYwUUFwCdGMUtINO3WiiL1xY/pJKjqfhEI84d1O4miUU+0E8qAbcvty2LnZ5fMTSfubRZuwKk85he7zk0+yk3k8zpvi8cfbz6h59125BCVDZSXTdvbeu00RoUDwCTFb3dCKQuuGV1/tPvtYRSsKhymcixklhdAN4d69KhpV+mh1BNmqzBdRxmIkxlNOkb9udY1VW4wBPC81rU2SIRB0QszVCFEz8qkb1tZyBq8K4UUifE4k0smzspKpNpMm8eQtkG5oNmhlSCbtr8U2RZRAYa3KLbdkMwiVbWhpMbqnt0EEN8osUCq6oXiPIDo3NDcD223H+tXKShboi+qFM87gCW5u45WJbiiQgW7oRobiNdlDvF/28B3l6DfdadVAXZ8+DOy1UQTXQqytLR3dUKTYNDd7Gw1aVQXcdRctwBUrmIPWo0dqM4dsdMOWlqx0QycydILT+9zGv/iKIFuVflqUffsypeatt5x3cFUV8M9/+ve9RYjgWoiy5g5WBFk3FLCW5jmN8jQjHDY+r0MHDirv00dOhgXQDTMlQzcE2qoECm9RZmpVPvgga6WdtMGWFjaYWL06s+8oAQSXEM0oRt1QNkK0Tx+gd+/0z5ShpYW/xzwLIwD5hjICqq1Nf/iNQBNlMbjfNTXA22+z4sXOfU4kGGH+05+cW5eVMIqDEAF3VzmouqGAqEYZP17d9/vhB+DII4GFC/2rUxbwSTcU5Ge9LmUkmU+iDIXsibLNWZUCiQQwdaqzbNPUxEIBMYmxjSH4hOglBzET3VAgH3XKAwcyv+v449Vcn0SCFqK5cgDw1tIrD7qh9a1O3p4dUWZClsuXA19+yXJcmRIRaKsSBSDK559XC640NgL33kvNu40h2ISYD91w1qzc6oayuSgHHEDXRUWIa25m9YC4Wr26ymZ40Q0tkOmGtbXeJS0nI0aVKD/6CBg1itkko0fTw+vXD7jqKhb/uCHQROngfmeNF15Q76mYSMg9ohJHsKPMQPHrhoB8LkpdnVo+IsCrdfZs47PzoRtKXGUz/HZ/nYKt5u968EHgmmuM3GMz7//732wU9P77QPfu3rchk+i375FvwDb6nXWVjkqnnN+2IeRc+1yiCLaFWCq6oRliap7qnRpIbdcUMN1QYN06PvyG2Yr8/PNUMrSisRFYupRzmLxkN6mgJHTK/v3V8xEbGznm9OWX5ed7iSK4FqLKTBWgOHRDAfMI0d69mVqj4sYmEvLJaQHSDQHOQ7Ijxd/9Tr7cC264wd1oSSSA77/n6bLjjumve+mzqoJAV+kg1aqsGDfOaAzrhqYmo3diUxNTdq69Vn5tlRCCbSEKeNENrQiKbmidp3zooZyrq4JNNuFnBVQ3XLeOZAjwWfYQFqTsoYLly5kRooJ4nFWNTjrlsmUcZLfrrsCwYdQhJ03y5lU6IZPod05gsiKx8848/1Sa1Oo6U282bODGzZjBc/b553O0ocFAcC1EAT9K88zwohtaka1uaB4hOmgQe9f98INzsnYsxv525jK9AOmGboS2ZAnw6KPA11+TDHbemc1aunTh6/X1alblL79wk+1mxZuRTHKeuxXiJ7/+OnDqqcb3A1x/zhzgootIjLvtxuV+W5SAvVWZ6yqdhB5CxSuvsE3YqlXeo8gNDcDpp9NwGDAg+w0KIIrDQiyUbihzlc3wqhuasWIFZ6J062Zf3xyLAQceyHZhVgRENwQM69CMpiY2bt5lF44Kfv99dqG65RZg++2pBeq6vUVptSqbmtRjUACLLZYsSV/+4YdsElNfn24N1tXxd44dy9PH7zQhN+Q0+i2Ob58+/HGnnUYZpkMHNp+tqFDTFxMJHtAShab7rT77hOH9+unTd989XTfMZhSA1VXOt24oIOYp9+nDPJE776QrIq6ARIIW4WmnMUVHxTrMUDdULc1zc5XNSCY5IfXjj+0v2liMluJ118lft6K5mbtfloIpQzjMa3y33Xj9du1KAt5uO/su+WaEQrQkZcoK4FxqnAur0g7mm4SrFZlsSY1G19dz7ndjI/CXv1CXUEHHjvKbfJFAq6mZoev6cNlrwXWZ7RLQrAiKbijghQwBnlxnnAGcfDKwciVP0u7dWZmiaUWhGwqIEth33+XcIicLpqEBeOIJEqds11oRidBQvu8+NbdZjAD58EPeK//3P6oTqmW6ySRTRd96ixqjFZWVPFQzZ/Kw9OjB00bT7K3HXLvf8bg7KSaSmkGK1dU04QH1Ow3Am7jom1hi8IUQNU3bH8BtAMIAHtR1/QbL6ycCuBnA0tZFd+q6/qDjh7ZrV1y6oRghaoWZDAXMDRoE+vXjQ2DDhsDmG8o0P0E07dsDjzyiFpxobCTB3Xab83qrVwMvvsjR05EIrUVV9zmR4PvPOINRZxUyFYjHqSm+/nrq8nXrGHB98kkjLtbSwkNz4YU0tmRc4eRm58WqDIUBu1zGaFQ977CqqiTJEPCBEDVNCwO4C8BIAEsAfKFp2lRd1+daVn1G1/WzsvqyUtANrWS4Zk16nbKVDAUCqhuayTCZlN+TZNB1ko2145QIuDQ2kpBefJHXn0oPRhkSCd5PBw9WH1cj8NVXPOQihrBqFasvly1Lj0nU1TFT5bPPqIJYOUM1+dyMvLnfBx5IhnfbQZoG7LNPfrapAPAjqPIHAPN1Xf9J1/UmAE8DGOPD5xqQ6YZAZvmGAoXSDc1YsyZ9mZc6ZSvymG8oYCZDwHvgsr4+dVpm+/b8zGXLmGD94ov8qWbrVViHXoyUcJizk2SapxNCIQaERHBn7Fg5GZp/z5QpjKyrwks5oxtUAi3SMsBzz1Wrw4zFOCCtROEHIfYGYDaRlrQus+IwTdO+1jTtOU3T+so+SNO0cZqmTdc0bfpKcfT91g2/+abwuiFgkKHfLb1c4KWll2oQxbwZXr2p5mb2rjDXILdvz6Ysc+Y4u7he4oFNTamKhBdoGn/34sU8fdxIv74emDgx+2oZr20QlSxnuyTubbYBJkxwvmNUVwNHHSU3OkoE+Uq7eRlAf13XtwHwNoBJspV0Xb9f1/Xhuq4P79quXenrhm4tvfKoG8pcZTOcdEMzli/37tLecAPvW+++y/91nc1W/CylraxkhtO4cd4aUoRCBpE+9ZR6f981axjIyVVJY05w3XUcZl9dnVoZVV1NjXHCBGazl6h+CPgTVFkKwGzx9YERPAEA6LpuvnQeBHCT8qeXgm5oRYnphgIia6N9e28kILZj/Hi6mj17ytWEbJBI8HQYOZKnwttvq73PHIhdvNhbw/P16/l+1eRzN4i8SDddUSXaLIWmAX//O3DWWdQTP/uM+sQOOwBHH82cRTvoOjcuEslRx4v8wA8L8QsAW2iaNkDTtEoARwOYal5B07Sepn8PBvCt66fW1paObihzlc0oMt1QQEaGnToBBx2kVh1mRTxOA6WuLrP326GyEth3Xw6T69iRLvruu7sbOrEYcMEFxj7yGuAQh8Uu8RzIrIzRCapus2M7sZoazvu++26a6uPG2ZPhL78wKtarF8/zbt04pOo//ynKbjlZE6Ku6wkAZwF4EyS6ybquz9E07SpN0w5uXe0cTdPmaJo2C8A5AE50/WDr2eelpVdZN0yB37rh6tVyI1Vs7gknqJdpW7FkCTBvnrcZS9GofbFPZSUtzhtMiWCaRkt0n33sR4zEYsDhhzMvfvVqPkaPVifFlhZ5cwkz7AgyEFDRFT77jGVHd97J8ziR4A9fsAC4+GIWiptHYBQBfNEQdV1/Tdf1Qbqub67r+rWty/6p6/rU1r8v0XV9a13Xt9V1/U+6rnvrPOnVVTbDi25oRVk3TIOdbmjGgAHA+edn5jnpOr9XtotlqKkBbr4Z2H9/BnRqavjzq6v5/6hRnNjasaPxnpUreRj/9S/g+utZVi4mvUYiwPDhrG657joaRiL6vf32alNkIxFOfvAa0VZZX7jNbshZswiBxYuBgw/mxsg8l/p6FogfdJD/vdhyiOBWqgjkUzeUucpmFJlu+BvyoBtaN/nUU0lC11/Pa0M1HUfXue6ECXRX3byuaJSHfOedgb/+lYGMjRtJZIceamxXSwtTeO6+G1i0iKTV1ESS+7//46GvraWR8+qrwGWXAWefTZI8+GBWyfTtS6I88UR7wqmo4CEdN477S+RU5hOhkFriekrVilfccYd7lntzM0nxww+BvfbK7HvyjOATIlAcuuHPPwPPPAO89x6vlu7dWax7wAHyqzoPuqGsThnIjW4osNQUTttlF5LQ/fcDzz2nFpCIRLg799qLtdAvvijffZpGsnr0UeOe0kk/DYAAACAASURBVKmTkUC9di0Jb/lyfu/557PMThCZcMk//5zkddJJwLbbklR13djtGzey8mbSJJLkSScBDz9Mwm5sNA5ZOMxt33pr9uzo2pXvtVrUKgS5bp0//SMd4VS1IlBZKdcuEgnuFJW7XG0tybNICDHYzR1GjOA/Ktahim6YzSgAmav800/87jvukI9trK7mlXLLLanZ/fkcIapgHWaqGy5fnm4ZLl2a2osC4DW1665qrl6nTnQKwmES05NPcvfW1xvls83NwE47Uct360LV0MDuNp9/7uy5VVbSqnIi7WgUuPFGJoy3tPDe9/rrPPR9+zJFTyY3m7Fxozsp1tc7E6LgKCc9U1iIjrKFtdmD05eZsXAhu2Somv39+nESWEBQnM0dBEEEVTf85hv6WsuW2V9FoqD33HNp6pg/w81VLjLdEEi1Dq2bNm4c3U0nFzgaZUBm2TJj2bHHsjrkiy/4+ZWVDFb07Gn/OQJvv03LUKWuWiWIE48DV1zBpg/hMFN4Ro50f5/fsDPczMiJ29zYyHP+oYe8lST5mTKQYwSXEIHg6oaLF9N/Wr5cLdLb0ADcdBN9R5luWFdXVLrhjBlUB8Ru3X57prX06pVuHQqceioJbepUOSnGYkx1GzcuNR1GkGyfPkaMKpnk8t6yeqhWfPAB70N+BxcaG3lvE6dmJsOsCqUtpsGL29zYyAjVrFnedmo4XFSVLcElRBkZZKIbCvipG37xBdlDtVOArrPqZsGC9CuhQLphUxODC83NJBZBZDLdcO1aTrT7/HP+hI0bjbaNAAcCPv44sN9+bPwq42dNAy6/nMrBAw9Qz6uo4M/6/e/Z+lF2eO0IduVKe4u0Z082hchFpLWhgR7jwQdzv8gsZSeSbN9erq6YIZrjuumIKknavuG669j23OtOraxkoneRILiEaEY2+YYy3RDILN9Q4I03vA/fqKzktowx9b0oQL7hmjXMyXv0UUOTa2piqsnf/w788Y/GusuWkeCmTqVVZvaSzO6YIMa33+ZuueOOdMN60035Xbvvzkc8zgu6fXumx3iFE1G++KJ/s1Gs0HXjcMiM+rVrc/O9Vqi6zZ57JMrQ1MQ+bV4TrWMx3iVl119AEXxCDJJu+MsvDIEuXOi0xfaQTc7Lo274yy9sMbl6dfrF9PHHDApcey11u1WraGV9+KF6D8F4HPjkE06u3H9/Y/myZfZdqmVSajbo2pX6fa6KJGpqnHMkO3Wi1ejmSheV2zxtmvf65epqYMQI6o1FhGATYpB0Q3O+YSbF7YkEh0oJ5Fk31HVWWqxYYW9wNjRQGu3Zk9bgRx9595AaGoBnn00lRFkuuoAdUW66qbfvtW5DLrHvvtm9v+jc5tWrvSVX9+sHPPaYe6lOABFsQgSCoxsCRjXKsGGcBujlJBk40CBEL7qhDBnohh9+yJ/h5n3H4yxfrazMnFjmzFFf144sneaeuJFl//50w710xxa71MkNjUbZDTsTFz8XUHGbAR/c5o4d1Y2AUAg47LCiJEMgyIQoI41C6oZ1dRTZxOQkL2QYjbK2U3wOkJ2rbIJqnfJdd6nNJ9d1ekiVlZlXXKl2hHGCHVE6ud+CKA86CLj1VvXv6tCBgZ4ZM4Dbb0+f8KdpPIRHHMEUILcId5DcZqX0Gze3ec891Q9qNMoSoSJFcAmxXbvc64ZWyHTDJUs4DemVV7zPwgR4glx+OU8qgQLkG3qRPcXckkyRy4tc1f3efXe6/E4WVCjEa/ef/+Ru3WYbBpUeeojDpVpaSIZ77MEZYMNbU3lXrsz+d6i4zUBA3ObqauD441mi42R2axrvSjJvrUgQXEI0I1e6ocxVNuOXX6iFvPZaZjkc++7LdutDh/L/HOuG69fzYn7uOV4kvXoxQLLvvt4aLSSTJLVMJk1WVbFiUTYo3gyznOoXzGR5/fXMafzuO/mhi0a5X66+OtUb3Goro9t1czNvDjJv0c1K9AOil6IT/HSbHXHddezgu2iR/As1jaz8xBNZfEnhURyECBRGN9ywwbAM3RCJkNj69aPZMXJkarebHOuGTz0FnH46z0txEc2dC3z6qdGgYO5cNV6vrmZe4I03er8PRCKs6gLkzX4EnAjTD7KsqgIefJABnkmTaGmFwyS5zTdnovjQofL7oEgRsuuA1bWrmpWo4jbnA6puc8KplK99ewrR48cDb75pTP2qqOBBHzyYd2O32sWAI/iEaOcq50M3nDFDXUzWNBa2iqvI7zplE6y64XPPkQxlQZDaWn79U0+peftVVSSLP/2JydiqhBgK0QK55x4aCk5kCNi/vmSJPVl6JcpIhNbq2LEkvvp6HgI3kvr55+yi3AC/77HHjBrwAw5gSp65fZgYqOUmMeSl2YMKunblibRsGZM9ly/njxg1Sm24dhEg2ITo5Cqb4bduKPINZ81SD1VWVXES+tZbZ9bSK0PdsLmZM4edIsK6TjIYOJA/127dSIRu9ujRtConTeIg+YYGe8NVRFz3359WZXOzOxk6IROr0o0ohbTlBZmS4q+/cj8sXMhTR9yE3n+fMbk77+TNRhWqbnNORwtY0aMHT7oSRHAJUZwFhdANBbyGWevq1OqUGxt90w2nTlULACaTvEjPPpsXZShkEGMoRGIbPJj6WU0NN7lTJ/YGvPlmVqGILtiJBONde+1F/u/Thxeum26YLezI0m/3u0eP1AYTMnTtmq4jrlpFtWTNmvQbiHASxo2jK2+OseUDvrjNbQDBJUQgM91QwI98w6FDWXqhoiE2NqZ3zc5DvuGbb6q11QJIetttR8P3iSdYgdjSQk3thBPYD9Da0kvTWLFy5ZV0ZwFeWNXVqUEMQUpm0vr+e/l2DBqktr2qyIf7rYKJE1m655TrGY8D55zDWu5wuAjdZtUoTpEiuIRoLXPLpr8hoK4bmkntz39mDacKdtgh9XvzlG/oNejR3MxNGjuW1ooZsv6GgFE3LIjs55/l6S8yYrLOQl60KDhE6UaKXtzmjRtZsqhirTc2speiauswFbcZyLPbXKIILiGakU/d0IyWFk4amjLFvZHfP/6RvjxHuiFgEOGQIfx6FWJsaeEFnk1/Q1lS9E8/pZPP99/LB8PbDYvPN1EKa9cOXt3mr76iBqtyHOrqmOfoZy/FgvVILDEEnxDzrRsKiLkol13GBL/3308nxcpKnmU33UR/UyDHuuFvi0LMM7zuuvTVZeje3WisqjoKwNxVRpChzFU2w47YnGBHlG6f5zdZZgqvlrpVTQmM25zNaAErEgmeIE1NjNZ17uzPNuYQwSdEIL+6IZA+QnTiREYVnngC+PJLHugOHYBDDmFXU5F4DWSvG1og0w3jcaMSpWdPlqq98orzRRmLMcqpac4jRAHvZCjTDZ0IzityZVX66TZ3767eHrOiIvU3qTZ78MttzjnWrWMN5H33UaMJhUiKe+4JXHKJPDYQEASbEAuhGwqYr/K1a1m3deCB/F/XjTM4H3NRTJCR3sSJJIc5c+QXTSzGfNo99lCbiwLI+w1mqht++y2ft9wyfd1s4EaUTqTol9sM8AYybBitNxXS0jTexPxGXnsk2uHXX5l+sHJluiHw9tusp7z1VnbJCCB8mcucE8g0u2x1QytkuuGKFe7zlEWydh7zDQH7OuVYjOkxV1/NYUfRqDGfeKedOCDttNPSvy4fuuG33zKKLf6WPfyG2IZMXHevEDcOTWNjaJfYGADqeGPGsKrTLDnItN2CQCLPpEEm6+g6S6KWLZN7RbrO6/q889h1PoAItoWYjxGiZtjNU7bCboRonnRDu7eefDLz4H75hedely58qM5T9ls3NJOdIEUZnEgxU6uyXz9aim5QcZtVceSRTGmymxsj0NJizIl+7TXesHbeuch6JMrwv/9xp7uF2uNx1oU+91x+tssDgkuIsu7S+dQNAYMMzWzgNELUjBzqhgKyEaKaBmyxRfpnuumGAn7phoLknIhQwG6dH3+0J0s/3G9Vt1lFRxTR5muv5YyYO+7gTcdJ121poYt90kmM2flBYgV1m++7T00z0HXmHa1dKz8JC4jgusxmFEI39EqGedYNvYwQVdENlyzh/eSJJ4zYka5nrhsCamTohM03lz+A3LnbmcB8E9E0WunvvsvqNpVmss3NLJMEisxttmL+fPXqrspKNYE2zwiuhSgQFN1QIBPdUMDnecqqI0TNkOmGjz1mNJAVvK1pbJR86aXUyAF7V9lJN8wFxGf/+KPzev36uQdXAH/dZgFNo66rUgrf1MSphRdfXGRus9Uc9ZLx3dISnNbjJgSbEEtFN1QcIQrAs24IqM9TlumGt91GT0d24cbjTNW55BLjELiV5smsttmz05cJDBtm/1o+kAu3WcCLAbRhg38VcXkbLWDFfvvxYKskZcZi/uZm+YTgu8yZtPRqY7ohIJ+n7KYbfv+9PRmat+Haa9ONaFXdUJDh4MHpD/G67KGKILrNAmEPXqeuM/Cq6/lxm2U32fSVPLrNJ5+s5jJXVVFP8LKD8oTgEmJDg5puKBshCmSmGwqUgG5ohVk3TCTYFOK009RcumSSM7WscNMNzWQog4wkrUTpBBW3XLjNbshFp56dd1YknlacfDK7eH/+ufu6suNeEJjvzt2706WQ3Z0FIhFWE0yYkPttywDBJUTrTs2HbrhmTcnohjJXGQDeeQfYfnvmzLm5ir9tY4K9bwW86IaZNFA2E2M+oNq/0WkSoIA5j/PUU73JZPX1fP/48WxObQcnvhEQPRLdkMlkDEdcdBFw4YX84WZ/XNOYObLVVowwd+jg8xf7g+ASogxB0Q2tKBLd8OWXebHV13uflSV+th+6oReofE4+3Gan4VYCVrd5m22Affbx3l0mHgdOOcXbGNVMoOo2yzwYW4h+cd99R2vxj39kz7nDD+cJ+PHHwZirYIPiIkQr3FzlXOmGMlfZDC+6oQW50g3jceD88zMfLRqLZaYbCnz5ZepDBSpWomo0Ox9VK1ZoGnDLLXIZ3A0tLYxSOyGQbrNAt24kxvfeIwk+8gjJUXUkR4FQ3ISogqDrhhJX2Qy/dMM77/RuFQpUVBitqlTzDa1kCLC79tZbG8tUiTFbqAQz+/RR0xFV3GYzpkwBvv7a23sAWvFOhRyBdptlaGnhSbp+feZ35Tyg9AnRDV50w0WLgDPP5FCMESOAq66Su9oB0w0BYPLk9E1SRTjMmmgV3XD2bLllJ4hQ/C0eKqTol/udLVTdZqEjrl3Lmc+ZEo5MsfEbqkEfT26zGUuWMMGyVy96b/368QS56y71Vu95RJkQVbBsGRsPHnooC1DnzuVVOmkS3YDzzjMKVHOoGwpkkm+4fr3kdykgGgWOOCKdDFR1w2ytQFW3WUVHzLfb/Oyz2XmIKr0P8+I2Z9rsYdo0YMcdmdtVW8vkyOZm3jEuv5x32YBVq5QJ0Q2NjcDxx/Nqt0Yjmpv5+uTJHJShgGx0w3XrMss3BLylfwAMEnbqxDrbUaMy1w2BVOuwUCiE2/zyy5lbhzU1wGGHOa+j4jYDBXKbFy9mv1BBhFY0NHCd0aMz13JygDIhumHKFI6rc0r9r6+n5SirJSugbiiwdKm3ZgiaBtxwA1NtdtstM91QwI0MS9FtBrjPVfocyCCa+O62W2bvN0NmuFmRE7f5zjvdy2USCe6od95R/9wco3gJccAAYN683H/P3Xc793ISaGpiixM3uLjKfuqGgKFnnXOOuvu2zz68GH/8MXPd8Msv/bMMi81tFuk33bpl9v6aGtaXh0JF1uxBnLQtLYwqq9QP1taqXTd5QvESogp69aJ1lykaGtTf39LCHk4+wC/dUJDhN98A//63WnCvspLpY/nON8wHVEtn/XKbjzlG3a0V6NaNPRKHDJFLIVaIZg9uyGv8Ys0aykmqKEROlA1KmxBVYSfsNjd7q7f0chJkCdX+hg8/DPztb2rEVV3NrisiVTMT3TBTBMVtVqlaUXWbhw1jypIXdOlC58dPqLrNKjqiktsciXjTBQNU0+wLIWqatr+mafM0TZuvadrFkterNE17pvX1zzRN6+/H9/oCp0lg7drx4KoiQN07li5lA+OnnnL3+KuqSJqvvmpcjNnohl6h4lqrfk+Qmj1EIhzI6AUybig6t7ljR3namgzhMLDLLtltl4/ImhA1TQsDuAvAKABbARiraZq1Hc0pANbquj4QwK0Absz2e3OCRIKDcO69lyUGr73GkiMVUmzXjgXCAYGuU5pRuevrOhOv27Wzd5VV8w0LiSA2e9hzT3W3ORJhFpcZRek2axpFa5UBM5WVwNln536bFOFHP8Q/AJiv6/pPAKBp2tMAxgCYa1pnDIArWv9+DsCdmqZpuu6gamUaossUzz7L0GpzM79b1+VjDGQIhXjm+jl5PEt8+6167qGmkfvFIDSZq5wPfPklG08UGn72SFy5klripEnuikooBJx4ovp2Ll4MfPYZM7+6dQP239/e+8x7j8STTqJhsXSp/YyVWIwbbZ5pXmD44TL3BmBuMbOkdZl0HV3XEwDWA3AZyy3B3Lnu62SCp54CrrySt9m6OiP6UFfHh6ZRDJKdbZWVdLtffllNrHEbAu4Tfv1VParc2Og8kCmX3a8FVN3moDR7UIGINp92Gq1vp+MRizHvX1ZpCqS6zfPmcd1dd2W58D/+AZx+OgMx996beWWcrz0S27fnHIXNNktv3R0O06zdf3/goYc8b2cuEaigiqZp4zRNm65p2vSV1q4ysv6GMsjahDlhwQLggQechbamJp4te+5Jwa2mhge0fXu2j/nkEyXWyGjObYaorPRWJZFJN3e3rKc5c7x/ZrYIotu8ySZ0QLp2TXc6wmFaZKNHA9dfL3+/2W3++msmyk+bxhtZXR0dmro6YNUqjqJ1auKRF7dZGAY9ewLTp9M83nNPZn307cvSp7ffZgRPxYjII/xwmZcC6Gv6v0/rMtk6SzRNqwDQEUCaVKzr+v0A7geA4X36eGePvn3l/RCd8Pjjat1pEgmemd9+y6h0IkGTxecD2q6dfXcbL9hqK/WW9NXVwB/+IH9tyy3lGuKwYbTW5s2Ta4nbb08XeM4cb/mIxeQ2A+qjBQYMAD74AHjjDRpFS5dSM9x5Z/LDnnu6f1ciAYwdK+9AJ1Bfz8Kpvfbi7Gcz/HSblREO0xLcf3+fPjC38MNC/ALAFpqmDdA0rRLA0QCmWtaZCuCE1r8PB/Ceo36YS3Trlnqmv/66+xxZgGkEb77JW/zmm9O3KdDdrUsX92FEXboweKfiBlVWOldFbLklk7SthTjDhvExb57cWhTE5qelGCS32WuPxKoqktRLL9FwmjYN+Ne/6OrKkuqtePNNtRqB+np+bibISY/EIkLWhNiqCZ4F4E0A3wKYrOv6HE3TrtI07eDW1f4DoIumafMBnA8gLTWnYFA5wwTi8UC3LjKjd2/qSm4RzqoqalBmeVSmJ4rSP1l1ohgUlQ0pitf9sA797JGYi9ECVqiMJm7fnpalk3Voxrx58t7IeUPAXGFV+KIh6rr+mq7rg3Rd31zX9Wtbl/1T1/WprX/HdV0/Qtf1gbqu/0FEpAMBlbwGgerqVGHOzUwrMHr35oxlmXYVi5EMzzkn1V0TIzvzSYpeyNCvak3VZg8q8NojMVN4yUesrJSvH6geiQFEsMeQ5gNjxqjlRITD7N4hUFWV+x7vPmDQIFYUfvghy0s3bCA5jh4NHHCAvEhn0CBaT4sWUR+bOZP/6zq1sA4dSIpedEWZppgJEaqMLXWb1+wnevRw72AleiT2tuZetGLDBiY6vPsu5ew+fdhgadddU++/Xu7diUTmY0tCIYVCk1AYiWRLXgOF+UBwCTEWY0gtk/7rXnDMMaykd0MkwoiyVzQ1Fdx9CIeBvfdmoF6mey1Zkm4NDRrEeNNZZ9FaEPeLykruCiHYeyVFILVML1dk6KW7TyHxzDPM+AqFDPVm9mxOOezWjRZ+39aQ5ZFHsi5dJUW3e3f3YE/OoRrFCRAClXbjGX50vOnZk90MnMJqsRhwwQU5Kc2IRvPbDs5qzWy2GZ+tEdUpU5jTtnYtL9REgo/6eiZ8T57MpOBMgy3i4QTz+93I0LwdbmS4aBEfQh5wQy5GCwBMxbnqKjoaVim7ro7bePDBxjE7+OD0z5AhFnPublR2m+1R3ISoApWON/vtB9xzD9khFuMjGqVm2LcvcNttwHHH5WVz7eBH4Y6wGGSkuNlmJMUlS5jUPXGi+wD7xx4zLjqvuqIbzESoQoYAiVCFDAE1MhQ3CXHTsIPX0QIACfDKK51JJ5lkbuHvf8+khuOPZwGIU0VcLAbssINRdZQpcj5aIKAIrsucbwwaxGTRr79mzkZ9Pduf77gjr3rV8F4O8LvfqdWqrl3rHrHcdFNaM8uW8UJesIC9bX/+mfw/ZAj1QhWrtbkZePpp4IoruMu86ooymMlTlQgBdSIE3MnQbC27kaG4uXh1T198UXk4IwCu++mnPD232IIlgRs2GKdlLMZjdsQRwGWX0QJUGUGQFUJh98or4TY3NfFH33orT5ZkksbG2WdzPEdA5jSXCRFg6Z0YFrXNNrwSP/yQsyDmzCEhDhtGP2S77Qq7rTbo3l2ey7ZyZfq84E03pRYlxucmEsbF+f776tZoMkmtCzASuLMhxVxphbmwCgFvZLhypfH3N9/QVc6kW1x9PW9YBx4IHH00cxkbGmhBjhljkKDKMaytTa+qywlWr2Zi9qJFqYbFokWcwnXTTcBbb5HpC4wyIVqxYQOnhH/7bWre4ZIlnDG7yy7Af/6j1smjwOjdm26alRTr6iiJyuruvbrmTU0k03A4O1LMBRnm2ioE1Jo7CPTuTT3ymGOya50Zj7N0/pJLgN13z+wzVOIdokdiVs0ekkmy9w8/yH90fT0ZfeRIRttUkjJziNLXEL2guZla4Tff8CCZk7CTSS77+GPghBPkCdrhsLdE7zxApHqYL8wHHjCqD7NFJJKa1C10PNVgi/hbVSv88UfvWqEqGQot1Qlmq1CVDHv3No7Dtdf6owfrOrPFCg63Zg9vvcW7gNMdQNdprj7yiK+blgmKjxBz1fEGYKHpggXOt854HPjiC3ZfzRQ57Hizdm36MjMpNjcDTz7pTwqlplHwzzaJu9Rc5JUrU8lQYMUK9gHxo9ipqQn46ivndQLRI/GWW9S+pKEhELNVistlHjjQezcbL7jvPrXbd309B23bdURwQEVI9y0yt3FjarJu9+58ls1XERfme+95E/OdoOtG8H3RovTqDzcXWgXFRISAnAgFZs3yNzXPLhIsTmG3oEpe3GYvheyrVvncWcI7is9CdEPfvpmRpq57yw+ZMcP7dyiiXTv3E7VLazdJWfWgIEaZtdihQ3bD080IhxmUybTczw2lRIaAvyN3qqponZtRX69OhiLw60tQRbVHoht03b+TM0OUHiGqwNrxBuDB8OLLBGC4diak2LWrfxZiSwutHsBfUswkydq8DXYQeZaqWuGyZepa4cqVqVqhDH37+rfvAeYlCpiJUIUMATUyzPo095KV0bNnZo05fURxucy5RCjETp7m6IMTVCv/fUR9fXr3mi5dmNVgdZ8BIxXHnJ/Yrx8ffjVJMN9DBCGJLjJmF1oQm2x6nxnFbhW2tLDv4SuvMGGhRw+O5dluO67TpYtan0UnxGKUKsTcZ1WrEMiMDF09WCdN/G9/Y26Qm45YXQ389a/uG5VjtE0L0Q4nnKCmX1RVGXVUeep445Rf5tVSPPtsf7KGIhH5fSETazHXViGgbhUCmQVOpk8HdtoJOO88YOpUEuPkyawaGTmSZHTppe77XtSLW29+oZAxauCqqwwXWdUqFC6yGxkmk3xEo+pkaJt286c/sZuHk+UXCvEHZFte4wPKhGjG0UerNWJoamJEbMQI34bTq8CNFEXjWCsxdu/Ox9q1fIwaxVy4bElR03ihyyweL6SYaeldkNJpZszg/XTNmtTcY5GttWgR0/GGDwfOOMN+38dinLz35ZfMM+zfn92Jfvc73oNFsYco+SuYVQi4kyFAsnv1VUbRZBtQU8Py2nfeCUS1SrAJccgQ1irlC506sUC3fXvn0aO6zrP8++/ZhdUtISyDsKIIrFjfKqwBs4Buhqq1eOmlwHXX8YKLxfiTa2p4I99nH3cpJxplZcQOO/D/bEnRTxc5n1YhwNPhvPPc65I3bAAuv5zrPvIIu5RHItz/kQh7ldxwA8shu3YFxo1jA42ffqLEcd99TGrPJHDiu1XY2vrLtf1XUxN/4HvvsY33pptyYFtFBd2LiRMpRAdkprlWqE7+bhi+xRb69H/9i2FMcwuw+fM5MMSMBQtSyx8WL04dSvXLL+njzFasSPX31qwxqvSXL+cZ+/jjamk4sRgtxcGDKSLJbv8myzOR1NIic/G4PI1CSC8yw1Xkmcm6YovmoLIeeua0HF0nry9YwEqSlhbeqDt1YhJxS0s6KcdirJC4/nqe14DREUbmQss0RS8IolZoxuefs7hJ5VSpqGBNsrg5/fILb1C9exuaoB0KrhUCalahdSPicZrFL73E/0USbLt2TFW4/XYKrXmCVlMzQ9f14bLX2k5QZeFC+xmPVnTvzpZgoRDw8MPuWczNzcxLvP32rDYxmUwnRTF0StZWUTR9yCbYEo2S+196iRerGDCo68DQobyvvPEGrZtIhJ7PEUfQ7TZnSGy2GUnRrrcikBkxBp0MAWqFqtUniQRvJhMnUl5r3979tDR/tmoEGVCzCgW8BE6UrEKB5mbgoIPYZdh6HYm7/fjxfO3YY102IvdoG4TYq5fagAnRAkbgpZfUSjoSCeC557IixGiUN9J8kuKiRTwHFy7kz7T+1K++okH+wgvchWYCFP3/zLtLEJGdtWjuxO1GisVAhAJeqz0aGhh8ve8+92F0RWsVCjz6KE8kJz2hoYGNUw44IA8tepwRbA0xn+jcOX3Z+vXq74/HvRUHS1IVxMkpy/0SJ7lMjsw02DJpEHmoFQAAIABJREFUkkGG0k1M8mKfMCE9X9autyJg33QWcNYVBYqJDAHuX69N0eNxdiN3kpe9aoWAulYIqGuFgKJVaP0xuk7dUMV81jS6Kg0NBe0HUCZEJ3gpIQqHDUHNBb+dXDakKLpoW4lRiON+BFvicbWa5mSSpPnZZ+mv9e3LmNddd/Fc/u9/jcTjTEgx1+k0fiZZA0bU/i9/yazAIplkvqIVXtNpAPXACeA9ncaTi2zGr7+6D5wRaGgAnn+e15FokuL0yBHKhOiEESPUSE7TOLQEsO94YzlpfjvRTHdhM3JlLQIkxWnT1LsiNzayRt/cV+O99/iTL7+c4zHvuAO4+GIue/ZZruOFFPORTuMGL1ahyOvs3p2ywZ/+5L3Ioq6OGSlmFJWLLLszm1Fbq2wkAEg9aQUxyh5AzsiyODXEuXPTI825wDHHUEd0c4XFEAs3SETA35o9JFvSIs+51BWXLfNWljVnDjB2LGWeHXcErrkmXRYSu2niRAbxJ0zwFmwJSs9CJ5gT3EWkGGBu4KGHMo3IS82y0B+LigjNG+GETTbxlnKmMosBSO03J0MWpFh8FqI5nSbX6NcPOPdc9yEWRx3FmZEqkJwgbi404L+l2L69NzcvmeT3vP46mxw7aeTxOLV0kUJqndliRdCSrO1gtgrNZAjwWEydylNBFeEwt68kyRCgLm/tQGGHdu2Yu+QH3KxLBxQfIaog0443Mpx1FvCPf/CAmae9V1eTDMePpz9phdNdKgCkOGpUZoX7wmJ1Q2Njer66kwvthHz2LJRBaIVAOhGaEYsx4V3F5QZo2R9xBP/OReAkJ0nWqmQo3NcLLpAnylpRXe0ecvcDLqRYnC6zHxAdb1SaNBx/PG/9r7/OgtVIhPWZY8ZQI7GaWuEwowsNDfbWZQbuM2BYZmYXWlwgsiRucaFZk7hjMWMktR/NYq3QdeaqWzs6ObnQVgQhgqxChAIrVrAKZcUK93UrKtjgYrfd3NctKqsQMIyBcBjYay8Wz99+u9xICIVoaDz+eOrrsoqCPKDtEqIKOnRg08pNNqFifsghLN61zn2QJfqpkiKQwmDi5EwISzGHuuKll7Lx94IFuSPFxYvTSUqFFIuJDDdu5KEeP54ygZuGWFFB2UIEn+wQuCRrN5iJ0IyLL2YV1zXX8GCJ15uaGIW74or07HS3pik5IswyIfqBqirjAJoPlDkiBvhqLfpBijU1wGuvARdeyKFFfnVyFgiFaExb890B+yTuYiJCwDjsM2cy1qdyY9l3X0blZamvAkVjFeo6S3VuuYU1jIkED/b48YzCdezI9Q49lAbFN9+wB1wkwo7z1pGQAk4h+8ZGZ8LMgizLhJgJZAOQq6qMA5WptVggUrz7bt68t9/eOVjiBaEQm0SIihRZZQuQai2al7khSGTYpQtw553qo7vjcX/IsOBWYWMjpaRPPkkfL3r11Rwv+tJLxryI2lpagv37Z9cI1u29WbTkC35QJd8db9wgAiuy/vziQMkOiNValKEAwRYRcOncGTj5ZO8VF3aorAROOsn4X6Wyxfq3DJl0p/GSZC2LIFshqn5EFRDgPvDJjM8/ly/32p0GyGOStQwnngh89JH8TlBfzx164IG824lroqoq912xxXfIHi4IPiGqYMAA/1pAqyBgpJhNZQtgkOIpp6jl0VZVOafsRCJsLbnttqnLu3Zlo4gzzgDOPBO4+Wbql0Bu02nckKlVaIaX0QCytNaicZEBnr+zZnHEqJtLUV9vZGEUeDyAyja0LZfZS8cbM374gaUZ335LxthpJ2YoJ5PZuc+A3IUuULClTx829znxRG6+lWBFFxxNcx4/o+tsJfbTTwbJPf8824lpmkHAn37KwMJ22/GasZN+8p1kLYP5/mYlQ4BJC6pVaj17Gn8XFRECxnn7wANqtfuJBNuGX3ml83pW46JAA+tLw0JUQa9eauuZz+rmZpox++/P5+efB555hnmJ223H/2tqUhPVBISJLuuqYC1BskMBXOi99mKgZfRobn779nxvVZWxyW5GQSJBlWPCBGqHkydTo2xoSM2JbGkh8c6cSRKW7YpCJFlbYbYKZWQI0LpWIarqalrIQBGTYTjM2eSqzUw0zbnblDgYNTXGQ1xTdo8coe0QogqE0i3q2s46i72vmppSw4d1dbyyr7kGuPfewLnQQHakuNVWwH/+Q1J74QXq4nvsQQJT7Sfc3Mza/pdeomXoRKLNzUzPeeqp1OW5CJyoJlkDqfcyOyIU+POfnZusC1RVcd1cdqfJKMm6sZER4C+/ZKqZGSLJWrHaQxniYAgSNMNMjtaH+b0+k2XbcplV0Lkzu2dPnszomVsft4kT2QCzd28SZZFHoAEjiVvojGvXspON18qWxkZ2wVFpItHYyNzcE080EpsLHUH+5RfOMPnoIxpDgwcDp57KWm6zhlpfz9NmyhRmlsTj6bmIYmjU888bVrcb8mIVrlhBveKRR/ijNI0HY489gMsuY5dgIJ0Id9iBEpKKeJpMpuoEQKpV6BVu78mCFMsWoqyOrHNn4Ikn1Pq46bpRoxYwSzGTYAuQHmyZMSPz6HM8rp7KU1tLAwUoLBmuX8/xCAccQKns66+ZY/jSSyy1Gz2a90wgtQ55wACS+uGHGylN4vmkk4APP2SBU2B6Fi5axIlW99/Pnb9xI1ujNzYCb7/N+s7nnpNbhaefrmYSV1QARx5p3OydrEK/4GZdOqB4CdHciypTiCEWVlJsbOQsFxU0NaX2cMqGFM194Oy+qwBtxHJRxSJDOEzXtNDpNLfeSh5obEzdN8kk98fs2XQKVq/mfpoxgyQ5eDDLIV94gQbRhRfSsv72W6blbbFFgNJpWlq40atW2euIDQ2sRZSlvW21FQVnlWlkZ53Fv7OxCv2Cy3cXJyH62fFGRor19d76uFkJzCoMm+EUbAECqSt26ZJZIwivSCRYJemETKzCLl2YIXLrrWzg/Oqr6b9XHIr16+kcOO3+5maeLq+8wqqz449nf8nmZqomjY1sA3bTTextoGkB7Fn49tskQ7cD29jIHWfFxo2sTx4+XE4y0ShLX599lhqBFzK0aph5ROlqiKLjjZk8e/WSp95YSbFHD/XoAUDmEDXPZtTUlISuOGwYn1UrMczYdFPKVCrqw3bb2VdwZJpO8+GHtM4SCWP7q6u5iy+4wBj2JoImEyeqyWLxON3qeNyePOvr2Wn88stZqmeHgkSQ77lHbRhMMsm0g7o6g8zE3aNjR5rS777LUp2ZM7nzunal2HrssYbY6oUIRQ8BJ7jdOTNEcVqI2UCQ4sKF6a8JYly2jE0cVKIBsRgnlAPygxgwXRHwbil26sRGEF4H28ditKA6d3bfldGokY5iRabpNFOmMENq3Tpe+7rOR10dpbLrrwduuy01gvzBB+oNXteude9F2tDAUQ1Cc7SiYOk0IiNeBRUVxlxbc8UJwAM7ciQF1sWLGYmaNQs47jh1Mly1KpUMxbPdw/we2SMLlD4hyvoiipxEJ1I86CC1SEJzM7UU84GyopCkaCHGTIMtY8fSNVQlxViMBLfPPkzd7NjRPmOjqorR5Z12Sl2eTc/Cn36itee06xoa2Mh21ixjmZdu16rQNCYtmFHwnoVe5gWJ4dwbN6qVwFnzCp1gJkJxDbnBiSyFdZkhYZY2Ifbty+dMSHHgQGDcOPeDr+tkiw0b1EhxyRLenZcuNRgpl8EWwBdrUdOAG25g2uWWW/L97doZUqtI4o5G2f369tupxwN0m995h+0jo1GuJ9YdOpRDqo45JtUtzjbJ+sEH1fKGGxvpPQpkUsjkhoYGY0wCEJAk6xEj1EvpOnfmNaGyfiZaoSoRqsKNMB2g6V60MuubNa0zgGcA9AewEMCRuq6nmUGaprUAmN3678+6rh/s9tnDt9hCn/6vf/Gf774DttkmdYX589PnqixYwFCfFYsX81kWjBEZ9LIrYcUKdjwQ77dDJMIi9okT+f+GDXw26xyzZ1NIevddskgySVY4/XR+hziBREhXVscmxC0nM01i1SaSra5LKN1EkzWcFZA1nBX45huez507894iWl/17m3IpbKfsHAhyaGqiuuaI8pLl6aum006zbbb2rupVkSjxpCrN9+kZZuJXuqEv/6Vefzm73RErnsWLlvGhEq3FIJolLrDaac5rxcEIlSEtvnmM3RdHy57LVsL8WIA7+q6vgWAd1v/l6FB1/XtWh+uZJgGLx1vZE0eMrUU6+vV2h83N1N4FkRoNd1ffJHlCW++yZO3vp5MtHIlSXTECEOjKRJdcehQBkE23ZQ/daedgD335D3HOvLUjP792Q9wyJB0whTWoNUqTCbZyPa442idDhzIFJbtt7cf+6sSxBGIx40Y2i67GC38/EK7dsbInYJ2pzGX3vXuTZZ2au9fWcn+bcce6/y5RUSGbsiWEMcAEJMzJgE4JMvPyw4DBvDZjhRF5NlKjL16yYMt772n3sYkHGZJgxkdOnBc3UUX2Q8jicdpgR5zjHFVFgkpOs2CFnW/dtlFwqJzKypobKQBffrpTG1pauJuSiZ5/3nySd5Prr46NTFAJc1FoKbGaDoRDgMHe79lOyIa5U0g54Ph7WCWW8xC7sUXkxSrqlK9jnCY/2+/PSdn2ZGmlyRrYRyoaIWix5vTI0fIlhC767r+a+vfywDYpb5GNU2brmnap5qm+UOaAwfST7MmaA8YYLQDy9ZaXL9evYC9pUVe9fLQQ+5uSXMz/ciZM41lOSJFuyRuPytbzHCyFlVI8a9/ZeqKU7VLUxODIzffbCw7+GC1mFhFBeNn5sHwfnaSi8WA++5TMJ7yYRVao1qaxpv1N98wB+mPf2RJ3lFHMcly6lT76Ve5sgoF2XXubP8Q6+WAMF3zEDVNeweAbGDqZeZ/dF3XNU2zO5L9dF1fqmnaZgDe0zRttq7rP0q+axyAcQCwqV1rcTMGDiSxyeY0DxhATXHevHRdsW9fWmXWPEWApPjLLyTFzp2pD6qEHisqeMDN/fLXr6evp6LTxuPsqLDjjsYyt1xFwHMbMaBws6CtP6O2lka10BRHjKBLLDBvHl9Xuf7jcRLPKadw2445htajGyoqaIGqjACVoaaGhyKZTE3rq67mYX/wQQWLM1c9CwXcGjJUVVEjPO009QgyoB5BBtSsQgGnduIqr69ZkzEpulqIuq6P0HV9qOTxEoDlmqb1BIDWZ6ngpuv60tbnnwB8AGB7m/Xu13V9uK7rw7vKhByZjigITVbK5+ZCA86W4pAh6sOLW1qMMYriYCxbpl4EnEwyeFQClS1ulmJzM2c7b7899fp772Ve7+GHs3x2dmv47ZFHvBlDmsZ+jgCTySdOdL++jz46fXTw8OHqAdhEgqV7d90F7L4777077sgOP4sXs4zXFrkcAQqodafx0sk6V+k0VqswWzhZly6fn63LPBVAa1YyTgDwknUFTdM6aZpW1fr3JgB2BeC9EHnIED7nkxS7d+f3up1UlZXMKampMXb6smU0mbzUvIkTsoiTuN1IMZGgRv/UU7Tq6upoSSUS/P/775nF9OWXJBovu0/0VhTke/jhJDune9ozz7DYwowTT1T7Pk2jNtirF7/rtdc4pfbTT1m+6xicyUfPQjdYk6ydoOoiy5Ks7WB2b/0gQh+QLSHeAGCkpmk/ABjR+j80TRuuadqDretsCWC6pmmzALwP4AZd190JUZb3kA0p2umKbsGWG290ziqORBixu+ii1OWdO3O5SrULwJNyr72KqrLFbTyBLNjyyiucP+LWVe3UU71VTwqYJd833yRBOn1OQwPHBosgP8DDfsQR7knosRjwf/+Xx56FTvDSs1B4G6pWoZfACVAYq9AnZEWIuq6v1nV9H13Xt2h1rde0Lp+u6/qprX9/ouv6MF3Xt219/o/yF5iDDAJDhhhpOFZiHDjQPtgCZGYtbrIJzZnf/95oaldZyeeqKuYf3nabPOmtWzf1zqEA80qAgpBiJsEWwJu1qOtMxXQrdwN4P/Q6FrWiIlVKvuMOtfQbTQMee8z4v6mJ98FRo4wItBnRKJdPnmxonl7TaVxRTFYhkHngxA1LlsgfOUJWidm5xPDNN9en33QTI2AAo19WiBZd1qRtwCA2a7AFMOo4vSZxz51Lkq6qYvbxHnsYjCDyFa2T19eto6q+Zo19Co8o5D3//NTlwkqWzZcwR64DnsQtGs7+/DNJxkuOoBdEo8x7HziQu27QIPUkgcGD2apLoF07Evh//wv8+998TiR4DZ92Gvsbiih5wZOsAW9aoRuCQoSAUUprhVt+sPU6NEHbckvbxOzgEyIQLFJ0q2wB0g/GsmVsALFyZepJXFXFq27cOIY6NU3exUOFGIuAFKdPZxRY5K/7icpK3p+eeIL/r1jB4IhqL8devahbBqaTtRtK3SoE7MnQDS5kqY0cWeSECBQPKQJyYkwmqbY/9BBr1KqrqRkee6yRpiMr+RMoAVKcPp3GsorL7AXRKDtRP/uscR2vWcNIs6qFuPXW9vOSzfCkFbai4FYhYE+GySSjV/Pn85zcYQf3TrqZpNPkgwgVoW27bQkQIpAbUgTyay0Cht7YQ5LeWQKkCMiJUddZY+xWGq6KUIiy8F//aki15pb+hx3GYiM3VFczTWbcOPt1Ss4qFEmSt99unFehEHOiDj6Y/d5kk7WCahV6QHESYo8e+vS77kp/wYkUgWBZiyVAioC/1uLDDzP3MFMdMRxmN+qGBl7rZs/OTIYAc+KPPNLdIm3XjqeG3VzooiJDFatQ14Hx4xmGl+2cigqS4auvGtZirpKszQGSPJAh4EyIwW3/VV1NH8bqxwwdysfMmfZRaCD/SdxOvRXtBlkB8oz6bHorus2C9thGDPA3X/GEE9gEwmuzWYFQiJ/XubOxK8zpPeaKk912Y6zKqX9BdTUrWmRkWPCehYD3dBrAPZ3mkUeAN96wPw8SCeYhnXoq/89VOo3ZKnQjQ9FnwK65s08IroW42Wb69GuuMbp3/uEP6SsVi66YqaUIBMpa9MtSbGkBrruOw95EB2tVVFSwVZfowygjQisee4wucW0tCU7T+DxkCMeFyE6torIKAfXAidAtli93/8xYjOQ5aFDhrEIz+QnjA3AefA84NrYsTpdZECJQPKQIlLwL7af7HI9z+NOjjzKtxa1kXNNYHfnII6kut1sdclMTCe6TT3h4wmGmlcoOvdnSLXjgJBcR5NmzWYqjchcKhWjS/+MfzuvlSisUZGgmQhW4kKU2ZkyREyKQf1IEgmUtlgApAnJiXLuWKoibrhiLMZK89db8X6Uhg+ChcjpNq8QybRpwzjnyxH4Z9tsPuPtu+9eDRoYK0H7/+xIgRMCZFIFgWYsqpKhpTJwTRbtbbskI32abBYsUgZxbi2+8wVRMO1krFmOJnRhE5ScZBsoqBPwJnAhY8wqnTWMyqCohbrIJZ0fsuWfqwSxCIhQoTkLs1k2fftRRwM47p79YLC60HSkmEiyCfecdajrCVwyHmTuy/fbseCoCPWZkSopAoKxFGSl+/DFLwn/6ySiXC4W4uRdcwDSaslWYZZJ1QwNbnXsJ81dX8z0PPMAdVcRkCBQrIQ4YoE8/8ED+U2qkeOmlrDGz63BQVcXvueMO+xKkEnCh7ZK4Z8+m3tfSQqN5u+1IkCVHhrm2CgF5xcnllwOPP+6NtKuqWPojcoO9RJDd4IUIxTVqp+EroHgJ8fLL6U4CxU2KgEGMdXX0DZ3avQBGtrD4bUFyofNcA11yRAgUtvRuwwYWlf/6q7e5q1VV7GgxaJDzerkmQ9Hg2QkOhFnchAgEixSB7KzFiRPpG6o0+ttqK+D559tssEUVZTJshZeehevWAX//O68f1YLvcBg45JDU8YFm+JFOI4M5/1fkBDvBhSy1sWOLnBCBzEkRCJa1OGIERwuoQNO47aFQMEkRyJu1KIOZg3IROCk5IgTSS+/+9z8GWby0BZoyJX15PqxCn6DttlsRVqpYseOOfEybxocZ227Lh6yyBWBOBxCMyhYvLaA1zThRg1jZAuStE7fd14pGtU4wW4VlMkRqknWfPt7McmtzSHN/QtVqE0DNKswBGbohuIS4cWP6WE/AGMJkJUWApAgEmxS91GvW1KSerObxBFZitM6Ctn5OTY3R/diKgHXitvnI35Z7Safx2skayNMIUBly2cnabgRoz57qrcnD4VRPymvpHWCM/HWCmQjzSIZAkAlRHORSI8WxY9UKeSMRYPRo9zro5mYO8zjiCE6LHzWKlQUff5z+viIaTyD7yHxYhXkbAWpFoXoWVlVxIIyohXRCRQVw/PH8u4hcZC8IrobYr58+/bLLOJQX4EgzK/IdbAGy1xX79OEgYBFCtUN1NfDii4bYJku/mT2b2cq1tamlWCKfcZ99OM7OerIHMNgCOOuKAr4GToo1yRrwNgIU4DHv2FHe6WLWLOAvf3Eu56uqYgdeMTuokESoMjhbdg22wklDDD4hArkhRaBwwRaAzfcaGtLF7FCIJ99tt3FwOCAv91u7lsS6Zo29yxONsvTqzjvTXyuiYIsqyhHkVqxaxZ3x0UccVv3DD7wptrQAu+zCch8xe9U8Mvfkk/lbrZ5AdTVzEP/+dx7XIJChrGhBQFyDNtBOOaXICREwSBEIhrWYLSm2bw9MmsQxdEKoTiS4/eedl37AraR4223svu12sVZVcdamkBnMKKLKFjcUFRnm2ipMJICLL2Z2u0zmiMU40EzMWxUSzMaNjCA/+ijPt3CYVVNjxvBZ07xphW7wmwgVoe29dwkQokCQXGg/0nLicSbItrRQ4BYpOf378zNmzOBF2asXE2IrKph6s+uuasNJIhFOYxcDrIo8X9GKoiJCID9a4RVXsAuGU/J/NMrE/9GjnT83CFqhj2QIFCshdumiTx89Wk58pUaKVsybB1x5JRv/hUK86iMR/n3SSXR3Tj1VPaF28GAGXkokiVugTIatMJPh+vU891XOjb592TXbmkoDBCPJ2qwV+kSGgDMhBjfKLNRzWZRZaGtBiUDbzYIeMICPefPShWCRUmDOtxLfcdpp1H2amsgOTU0krI0bWWD/8MPp2+EEQVrZ5CsCeY1AO82CDlQ6TaEHw1vTaaZMkROc3fuFIWBGENJpzFahj2TohuASImBYgPkkRbvxBEOG8PH117lLzWlupn7oRC4NDfxddjOerdC01NrTQpKihWx+S3MxkZYZstScwKXTuCHf6TTz57vXyQuEQhyWbUYuXGQvSdZm48GNCIURIntkCIXkowJj991Jeh99lO4iW0nR/LqVFM0utJUUrS700KG8c86cme5CDxlCF/rrr9Nd6IEDeeDnzk13oQcMoAs9b166Cy2K1Z9+Wq2EqqGBF0JdnTsxVlXRzTajQwe6z6tWpbvPVlI0u9BWUjS70FZStHOhm5rSXOiKkE4XOtmS5kJHo7y+21TPQhl+/pmeweTJzEWKRoG992bZnfk8jEScv8sK8TuKTSsUpCeTnMQ1mAGCqyHKgioy4hMoBV1x/Hi5CyNDJMKT2ckaiES4DXfdRUsxSHXQHnVFZZSiVvjii8z/a2lJ7U4TCnE/HnsscMklPMZvvglceKHaiIDKSmrLggNKgQwVoO2/fxFqiDIUwoUG8lfZourqADyZL7qIpCOrMojF+F333WfM1/W7DhrIm66ohFz1LPRaegeoa4WAMxn+738kuHg8vVVXMsnlTz7J4wxwvoJqiH6rrUiGXqbeqWqF8+era4Xz5qlphcIdFrp9DhBcQqyt5eQhK4qJFL0GW1SGfws0NzPJ9oUXmFZTXU1iDIWA3r3ZhHbyZFYnALlrDgEUnhRzOQIU8HcEKJBKhG564eWXu0eMGxroBYhxhFdc4a4rRKPAmWe2GatQFcF2mQX57bFH+gq5dJ+BwrjQr78O/Otf7pPVAZLe3XcbnbiTSbpJkYjhBhfLICsgcxe6mFxkr0nW8+ezpE7Fc4jFmGt63HH8//nnOYYCSCXUWIy/6dprgb32cv7MXKfTFIgInVzmYBPiZZcZVqJXUgSKT1fs359Dptz0n2iU7rKwevM13Q8Ijq5YTEQIZBZB/uADNupQHVx93HGAWXdfs4YJ2q++ys9o357lnkcc4W5BFptVaPXS7HoRoNgJUSBTYiw2UtywgWVXdm5SNMptuvlmuseZjjwFMifG9euBDz/kXI6ffqKbttNOrJEVKT65JMVSIMOvvwbefpv7smdPElWfPqnpNC+/TILLlBAFii3J2iwxuZGhmQiFhCWuQRtoF11UAoQItB1SjMeB66/nxdHcTHc4GiWhjBxJq8EaSMnXLOhff6WFsXJl6uS2cJhkNnIkcOutdN1zQYqtKNq8wu++AyZMYJpVYyOPrfi9O+7IwWLC+v7hB7bsV3GZq6s5yfGQQ1KXl1I6jRWCDAURKkI7+OASIUSgNEgRcCdGXWdO4ldf8blHD87GFRZEprOgsyHFDRtIeMuX2+c/xmLssHP77fzfR1JURpDJ8JBDeCORXXeRCInr5ZeNYNiBBwLffuv+fdXV7IFpdoULTYb5tAo9oDgJsUsXffqoUXLiyzcpAsHKV8xkFjSQPSk+8wytP7eoZyzG/LbNN+f/MlIE/CdGv5KsW1qYz3fnnbyodZ3HaPx4zsRxSmuxC5zoOhOpf/zRuUN1JELSvOEG/j9tGuvWnazEWAw46yy27wIKT4RA4KxCM4qTEN2izG0p2BIEUly3jtahWMcJ4TDd6ttu4/85CLakwS+rcNUqBrZ+/jldu6up4bF45hlqfAsXMvBRVwd07860r3bt5IGTL78EjjpKLYOgqoopYKKe/4knGBXW9dTfKRoBH3ccI8yaViZDBRQvIQqX2Yn8SsGFzjcpAt6JcfVqYLfd1IlnwADg00+N/3NJin6RYWMjz6MFC+znFVdWcn+3b085QwwCq6ykHnjEEcwdtLrV118P3HOP2vySdu2AW24xcmABIwH7hReYo1uOGi6aAAAOoklEQVRVxY7oJ5wAbLllZoGTUiRCkS8suwZbUfyECJRJEbBvi+6ntbh+PUvFnnqKf7drx555BxzABHDVapq+fYHp01OX+U2KfmuFTz8N/O1vqcEiO2ianNyiUVaLTJ6cur1//zsJTQU1NdyO/fYzEuqdULYKCVnxhATaFVeUACECpUWKQDCsRTMpfvwxcM45RkmYgLiwdd3ecrJijz2YA2eFX6SYi8DJLruoBTDcIKpAzjvPWHbttWzdptKlqLqaw+BHjXJer1jTaXJFhGaL2gHa4YcXISF26aJPv/769BeCRIpAsKzFbEjx22+Z7O1kAYbDJEW32dI1NUwf2WUX7wncgHp1ixu8RJA3bOAF7WVuthM6dqRLLYaJrV5NK1ulcWu7drw5OckEZauQ8EiGgDMhBreWGZDXMguy8/qauQbaWuv8xz/yIXttxx35mDYt+M0h7GZBA8bFYLYUBLp1o77l5g63tLgTRkUFsOmm1LfsZkGLel9zMwQzVOqg3ZBJOo2fxkFLC/Dee/y7Qwcet222cR/3GY0ySJJvMsx1z0JALZ1GlQzNPUs9kKEbgmshukWZgWBZi8VsKS5axGCAivUiggXJZLr7HI3ywnv++dRk7lzXQQN0oSsqUn+D156FO+4on4OdCURgZMQIY5mIYK9eLe97GY3y4n7oIXlfw7JVSGRJhDmzEDVNO0LTtDmapiU1TZN+Qet6+2uaNk/TtPmapl2s/AV77MHHf/+bO2vRiqB1zAEysxRl4wkAo32TaOckPl9lUDlAwtl3Xw4sr6kx+jL26kXN7PHH3ZvOmpFpx5xff2UQZOxYEkSXLnRTDzgAeOst9zb6su40Z5yRvZtuhkiuFthkEyZdjx7N76yu5veJrjd/+UtpkGGRWYVmZGUhapq2JYAkgPsAXKDr+nTJOmEA3wMYCWAJgC8AjNV13bGlrW9R5mKxFIHCWYtz5lDEV4muArQmzz2XF+fGjSSumhqSUK475nz5JdNaPvzQvrt4TQ1ThB59NN31dOpkvXEjj+mqVdm7zx068OYoI7dly7ifZsxgknfnzvxemZtcTEQIBMMqlBklJmg33ZQbC1HX9W91XZeYKCn4A4D5uq7/pOt6E4CnAYzx/GWZWoOFsBS33ZYHxXpgnGa2AIXTFauq1AkgFjNOyhUraAW1a2dYZLnqrRgOc4DSyJHAu+86j1qoq+PxPvfc1OVuPQvbtwdeeonWpqwbTDTK7XCzPquqaEHbkSFAyeKww4BDD2VJZpkM7aFKhuZrTlyHsocDfNEQNU37APYW4uEA9td1/dTW//8C4I+6rp/l9Jm+R5mdXgOCZS3m21JcupTu4q+/pr9mRTTKLs41Nc6VLYDcWly3jpUejzxC0qys5H494wxg+HCSjcxa/Pln7mNVKxYgMc2cSZLw0rNwwwbgueeASZPYxELX2X/y9NM5sP3oo0m6dnmIgwczj9P8XebGvG65hblOpyk1IgRSidAF2rHHZm4hapr2jqZp30ge3q089+8ap2nadE3Tpq+MxzOzBu00Rzc9si3oioDcUuzdGzjxRPcSuaoq6lzC0uvcmY9ly9Q6cc+YwVScW27hRR+Pk3xefx048kjW47a0yK3FBx5QG8BlxX33eRsBCrAd19lnM7H8hx9YfzxtGvfRttsCr7wCDBtmfGY4TD2wqgoYM8aeDMX+ckI+RoC6oYTJ0A35sBB3BnCFruv7tf5/CQDoui4x/wwoNYgtdV0xU0sRyMxafOopNjRIJNLTa2Ix6nIXXMC0GivcmkMsWgQcc4yzhReLsd73xhv5v9lS7N+flTNesffe/F1uyKSB6+rVPJcaGmgF77dfugZqJkMnFGuSNaDenSYgROhkIeaDECvAoMo+AJaCQZVjdF2f4/SZvtQylwIpAvl1ob/7Drj/fsPiTSZpBZ90Eq27lSu53GtziAkTGP11y2MUjQ3EZ4iegXY12W7Yd1/gscfsX/djFrIMqkQIlLVCgTxZhTkjRE3TDgVwB4CuANYB+ErX9f00TesF4EFd10e3rjcawL8BhAE8pOv6tW6fnZcocy5JEQiGtZiprihc2kGD0lNyvDaG2LiRsoNqnuOECaz9FWhsBDbbzLvLHItxYt2ECfLXy2QYjJ6FZtlIJXAi4EaGMhkLgHbnnUVYuudnLXM52MLnQrUR+//2ziVEjiKM47+/hiTkoiYa8QFqUIzeFOMTPKgH9WDUKHjyQdSICF4DuelBFE+iICIBRVHRg67g2yg5JbqHmPhAjYJgiElUUHLxlfJQ1Unv7Mx09+x2V1XP94Ohe6cru1Wd7d9+3d9XNbt3+2ePhw/PbzuMyy6DmZm5723Y4JM5TVi2zE+fGxRT0w97gv6JEPodFRYyLJ7xl9A992Q4dW/YxVO3vKZJQsWSLX47yXS/1av9hT04u2Mw2TLJH93B0pyHHvKJi7qsWOEXdB0lwzofAWoy9LQhwyZF1oPlNOMoT7EdIsMq0o4QJ10gNsVb6BQiRYgTLf75p8++1lmhZulS2LgRHn7Yf10UcTvn39+2bfwcZ+lYNvzRR4/VDE4SFUK1DNsup+mbCCFKVFgmzwgR2inGTq2IGyaLFNeunTvtqeDcc/2rPH2qTFvRIgyfB7xypZfKFVeMX3q/QDo2LRCOiUzyyZ5bb51fQlNIr8hSv/een3kzKMOmUWFdGbZZTlOFydCzgKiwTNoRYqwscxuRIqQVLXYdKc7Own33jV9RZ/lyP8+3+OgBGD7d78ABnznetcvXAK5b50t6hkmpbuIkpagQ4pfT9FiE4yLEPIQIaUkR+nEL3bUUP/jAf+b0kSPzM84rVviZKk8+6QujBxk3D3oYlkH2WFToKd29aevWHggRFiawlKLFPkkRmonx4EG/osubb/pni0uW+F/o++/357iYndJ0cYgyOckwpagQ4pfTtCXC4rEWoE2bMhTiqLnMkFa02AcpQpxki3PDo8FJV8zJSYRgUWFBhzKEXIWYWpY5RSlCWtFiG58FXVeKJkNPkzULC2JHhdCsyLpKhuXk5oAMIWchxpjLbMkWv81FimXaSJz0TYQwlVFhmbyFCGmJr1ySk1K02FcpQrUYLSr02xxkGCkqLJO/EKG9bHFK0WIfpAjdR4vjyLWcpm8ihKhRYbmtXnqpB0KEOHKbZilCWtFiUylaVOgxGc4tu8lSiKtWudkbbkhDbrlIEdKKFmNKsQ0Z5lpkDXmsWdhiVAgcvdbyvWXOJctsyRa/TUGKFhV6LCqc27Z0jeUrxJyyzFXH+3ALHUuKUE+MfZVhrkXWUD9x0oEIC/IWIqQnt1xuoVOSIrQbLUL/RAgWFRYskgyhD0KEvBIqfZIipBUtjirNqYPJ0NNkzcKCDMpp5rQd9RiKvgixICW55SJFSCta7FqKtmahZ4qjwjJ5CrEPWWZLtvhtTClaVOiJLcPIUSFw9HrUyy9nKMS+ZJmrjud2C52CFKFajCkVWYOtWQhRo8LyNZhnhJhilrmt7zvNUoTFjxYtKvSYDIdOtc1biJCewKZZipBWtDgoRVuzcLJymr6JEEZeb/kLsSAlgfVBipBWtLgY9YoWFfrtNEeFMPY6648QIS2BdS1FSCtaTEmKdclJhrkWWUP9xEmHIizIU4h9zzJXHc/lFjoXKeYkQrCocLDtIskQchVialnmGD9zWqQI7YrRZOjJaM3CScppgEoZsn17xmU3bWWZF/JvTYp5JFvA1iwssKhwTtu8hQjVfwFSElgfpAj530JbVOjJSYYtRoXltvneMk9bltmSLX67ECnmWmQNtmYhtBYVltv2R4iQlqQs2eK3KUkRLCo0GY5tm6cQc/lc5uJ4G8dG/cy+SxG8GMfNbBkmxbqYDD1NZZi5CAuyFKKkQ8BPsfsBnAz8WtkqL2xM6dO38UA6YzrLOXfKsAPJCjEVJM2O+muSKzam9OnbeCCPMR0XuwOGYRipYEI0DMMImBCreS52B1rAxpQ+fRsPZDAme4ZoGIYRsAjRMAwjYEIcQNLtkr6SdETSyIyYpOslfStpr6TNXfaxKZJWSvpQ0vdhe9KIdv9J2hVeM133s4qqcy5pmaTXwvGdks7uvpfNqDGmuyUdKv2/3Bujn3WRtFXSQUlfjjguSU+F8e6WNKS4NSLOOXuVXsAFwPnAp8AlI9ocD/wArAGWAl8AF8bu+5gxPQFsDvubgcdHtDscu69jxlB5zoEHgWfD/h3Aa7H7vQhjuht4OnZfG4zpauBi4MsRx28E3gUEXA7sjN3n8ssixAGcc984576taHYpsNc596Nz7m/gVWB9+72bmPXAC2H/BeDmiH2ZlDrnvDzON4BrJanDPjYlt9+jSpxz24HfxzRZD7zoPDuAEyWd1k3vqjEhTsYZQGkxP34O76XKqc65/WH/F+DUEe2WS5qVtENSatKsc86PtnHO/Qv8AazqpHeTUff3aEO4vXxDUo05hEmT9LWzJHYHYiDpI2DY51ducc691XV/FoNxYyp/4ZxzkkaVFpzlnNsnaQ2wTdIe59wPi91XoxFvA6845/6StAkfAV8TuU+9ZSqF6Jy7boHfYh9Q/kt9ZngvGuPGJOmApNOcc/vD7cnBYe2cc/vC9kdJnwIX4Z9xpUCdc160+VnSEuAE4LduujcRlWNyzpX7/zz+eXDOJHftlLFb5sn4HDhP0jmSluIf4CeXlS0xA9wV9u8C5kXBkk6StCzsnwxcBXw92C4idc55eZy3AdtceJKfKJVjGni+dhPwTYf9a4MZ4M6Qbb4c+KP0OCc+sbM6qb2AW/DPNf4CDgDvh/dPB94ptbsR+A4fQW2J3e+KMa0CPga+Bz4CVob3LwGeD/tXAnvwmc49wMbY/R4yjnnnHHgEuCnsLwdeB/YCnwFrYvd5Ecb0GPBV+H/5BFgbu88V43kF2A/8E66jjcADwAPhuIBnwnj3MKKSI9bLZqoYhmEE7JbZMAwjYEI0DMMImBANwzACJkTDMIyACdEwDCNgQjQMwwiYEA3DMAImRMMwjMD/gMwN2T/I4zsAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 360x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5avEVO-SBAkq"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}