Assignment_2_M2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Assignment 2 M2.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "TPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "wHg6i-MrwyWB",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## MA 598 Programming Assignment 2:  Due Monday, February 25, 8pm.\n",
        "\n",
        "### Model 2\n",
        "Team member names:  \n",
        "\n",
        "In this assignment we use a GRU together with\n",
        "\n",
        "* *Recurrent dropout*, a specific, built-in way to use dropout to fight overfitting in recurrent layers.\n",
        "\n",
        "**Model 1: **  Modify the model described in Chollet, Section 6.3.6 (GRU plus reccurent dropout) as follows: \n",
        "\n",
        "- use a skip connection that adds the baseline prediction of the same temperature 24 hours ago to the output of the GRU, \n",
        "- use only 16 nodes in the GRU instead of 32, and \n",
        "- use a lookback of 432 (3 days).\n",
        "\n",
        "**Model 2:** Extend Model 1 by encoding the date using 2 floating point numbers of the form (cos 2 π t/365, sin 2 π t/365), where t is the number of days from January 1. Likewise, encode the time of day using a similar format.  Include these values as input to the GRU + skip connection from Model 1. \n",
        "\n",
        "\n",
        "### To do:  \n",
        "1. Specify the model number above and indicate the team member names.\n",
        "- Specify your model structure below. \n",
        "- Restart kernel and run all.  \n",
        "- Answer the questions in the next text box - in this assignment for Model 2 only.\n",
        "- Print to pdf.\n",
        "- Combine the pdfs for the 3 models into a single pdf.\n",
        "- Submit on gradescope (indicate the start of each model and your team members on gradescope).  "
      ]
    },
    {
      "metadata": {
        "id": "UUqHpRkBS5nb",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Questions (Model 2 only):  \n",
        "Give brief (one or two sentence) answers to the questions below.\n",
        "\n",
        "**Q1:** What other data preprocessing could you do that might improve accuracy? \n",
        "\n",
        "**Answer: **\n",
        "1. This is typical time series system. Therefore, a preprocessing like Fourier Transform may help to improve a lot.\n",
        "2. We may also change the temperature into several classes, like scaled resolution. The model then becomes a classification than regression.\n",
        "\n",
        "**Q2: ** Why do you think that dropout didn't really improve the best test performance in the comparison in the original code from Section 6.3.6?  \n",
        "\n",
        "**Answer: **\n",
        "1. The model is not deep and complex enough to have overfitting problem.\n",
        "2. The inputs have been normalized before trainning, so there is not much irregular inputs.\n",
        "3. The number of training epoches are small. If increase it, dropout may effect more.\n"
      ]
    },
    {
      "metadata": {
        "id": "p8r1E-ZWgoPm",
        "colab_type": "code",
        "outputId": "b617147b-a0c9-4697-e347-f2b8844108ce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "import keras\n",
        "keras.__version__"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.2.4'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "metadata": {
        "id": "tVNXhrapgoP0",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Recurrent neural networks for weather prediction\n",
        "\n",
        "This notebook adapts the code samples found in Chapter 6, Section 3 of [Deep Learning with Python](https://www.manning.com/books/deep-learning-with-python?a_aid=keras&a_bid=76564dff). Note that the original text features far more content, in particular further explanations and figures: in this notebook, you will only find source code and related comments.\n",
        "\n",
        "---\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "2lpEeW_bgoP2",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## A temperature forecasting problem\n",
        "\n",
        "This model uses weather \n",
        "timeseries dataset recorded at the Weather Station at the Max-Planck-Institute for Biogeochemistry in Jena, Germany: http://www.bgc-jena.mpg.de/wetter/.\n",
        "\n",
        "In this dataset, fourteen different quantities (such air temperature, atmospheric pressure, humidity, wind direction, etc.) are recorded \n",
        "every ten minutes, over several years. The original data goes back to 2003, but we limit ourselves to data from 2009-2016. This dataset is \n",
        "perfect for learning to work with numerical timeseries. We will use it to build a model that takes as input some data from the recent past (a \n",
        "few days worth of data points) and predicts the air temperature 24 hours in the future."
      ]
    },
    {
      "metadata": {
        "id": "7A7JSD2tgoP4",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "First get access to the data:"
      ]
    },
    {
      "metadata": {
        "id": "g2dXPhm7sMxz",
        "colab_type": "code",
        "outputId": "66ed8863-b390-432d-93e1-6e129d908b0d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "CRGW3KyAR63r",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Look at the header and a few lines to understand the data better.  "
      ]
    },
    {
      "metadata": {
        "id": "R6b8z4uwgoP5",
        "colab_type": "code",
        "outputId": "2cd0664c-aa99-4a35-86dc-b8c73cb7735d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "cell_type": "code",
      "source": [
        "import os\n",
        "\n",
        "data_dir = '/content/gdrive/My Drive/Datasets'\n",
        "fname = os.path.join(data_dir, 'jena_climate_2009_2016.csv')\n",
        "\n",
        "f = open(fname)\n",
        "data = f.read()\n",
        "f.close()\n",
        "\n",
        "lines = data.split('\\n')\n",
        "header = lines[0].split(',')\n",
        "lines = lines[1:]\n",
        "\n",
        "print(header)\n",
        "print(len(lines))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['\"Date Time\"', '\"p (mbar)\"', '\"T (degC)\"', '\"Tpot (K)\"', '\"Tdew (degC)\"', '\"rh (%)\"', '\"VPmax (mbar)\"', '\"VPact (mbar)\"', '\"VPdef (mbar)\"', '\"sh (g/kg)\"', '\"H2OC (mmol/mol)\"', '\"rho (g/m**3)\"', '\"wv (m/s)\"', '\"max. wv (m/s)\"', '\"wd (deg)\"']\n",
            "420551\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "gd8Mo15wFo1P",
        "colab_type": "code",
        "outputId": "83f662cd-edd6-49fe-ab62-9d2d33dd464e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 55
        }
      },
      "cell_type": "code",
      "source": [
        "print(lines[1000:1003])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['07.01.2009 22:50:00,997.79,-10.79,262.54,-11.84,91.90,2.68,2.46,0.22,1.53,2.47,1323.57,0.24,0.63,139.40', '07.01.2009 23:00:00,997.88,-10.68,262.64,-11.74,91.80,2.70,2.48,0.22,1.55,2.48,1323.13,0.10,0.50,184.20', '07.01.2009 23:10:00,998.05,-10.99,262.32,-12.14,91.10,2.63,2.40,0.23,1.50,2.40,1324.97,0.20,0.63,146.60']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "L5jLNcUGgoP_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Convert all of these 420,551 lines of data into a Numpy array:"
      ]
    },
    {
      "metadata": {
        "id": "hjGJdAw7goQB",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import datetime as dt\n",
        "import time\n",
        "\n",
        "#float_data = np.zeros((len(lines), len(header) - 1))\n",
        "# Add extra element\n",
        "float_data = np.zeros((len(lines), len(header)+3))\n",
        "values = np.zeros(len(header)+3)\n",
        "for i, line in enumerate(lines):\n",
        "    a = dt.datetime.strptime(line.split(',')[0],'%d.%m.%Y %H:%M:%S')\n",
        "    d = float(a.timetuple().tm_yday)\n",
        "    values[0] = np.cos(2*np.pi*d/365)\n",
        "    values[1] = np.sin(2*np.pi*d/365)\n",
        "    t = float(a.timetuple().tm_hour*60+a.timetuple().tm_min)\n",
        "    values[2] = np.cos(2*np.pi*t/24/60)\n",
        "    values[3] = np.sin(2*np.pi*t/24/60)\n",
        "    values[4:] = [float(x) for x in line.split(',')[1:]]\n",
        "    # Change date information into required format\n",
        "    float_data[i, :] = values\n",
        "    "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "b81QVldCF0jz",
        "colab_type": "code",
        "outputId": "24862579-8736-42d7-d8f9-33118bca5fe0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 277
        }
      },
      "cell_type": "code",
      "source": [
        "print(float_data[1000:1003])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[ 9.92748722e-01  1.20208045e-01  9.53716951e-01 -3.00705800e-01\n",
            "   9.97790000e+02 -1.07900000e+01  2.62540000e+02 -1.18400000e+01\n",
            "   9.19000000e+01  2.68000000e+00  2.46000000e+00  2.20000000e-01\n",
            "   1.53000000e+00  2.47000000e+00  1.32357000e+03  2.40000000e-01\n",
            "   6.30000000e-01  1.39400000e+02]\n",
            " [ 9.92748722e-01  1.20208045e-01  9.65925826e-01 -2.58819045e-01\n",
            "   9.97880000e+02 -1.06800000e+01  2.62640000e+02 -1.17400000e+01\n",
            "   9.18000000e+01  2.70000000e+00  2.48000000e+00  2.20000000e-01\n",
            "   1.55000000e+00  2.48000000e+00  1.32313000e+03  1.00000000e-01\n",
            "   5.00000000e-01  1.84200000e+02]\n",
            " [ 9.92748722e-01  1.20208045e-01  9.76296007e-01 -2.16439614e-01\n",
            "   9.98050000e+02 -1.09900000e+01  2.62320000e+02 -1.21400000e+01\n",
            "   9.11000000e+01  2.63000000e+00  2.40000000e+00  2.30000000e-01\n",
            "   1.50000000e+00  2.40000000e+00  1.32497000e+03  2.00000000e-01\n",
            "   6.30000000e-01  1.46600000e+02]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "97qHRCELgoQV",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Preparing the data\n",
        "\n",
        "\n",
        "The exact formulation of our problem will be the following: given data going as far back as `lookback` timesteps (a timestep is 10 minutes) \n",
        "and sampled every `steps` timesteps, can we predict the temperature in `delay` timesteps?\n",
        "\n",
        "We preprocess the data by subtracting the mean of each timeseries and dividing by the standard deviation. We plan on using the first \n",
        "200,000 timesteps as training data, so we compute the mean and standard deviation only on this fraction of the data:"
      ]
    },
    {
      "metadata": {
        "id": "_k5_M3cOgoQW",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "mean = float_data[:200000].mean(axis=0)\n",
        "float_data -= mean\n",
        "std = float_data[:200000].std(axis=0)\n",
        "float_data /= std"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "U3r_D2SjgoQa",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "\n",
        "Now here is the data generator that we will use. It yields a tuple `(samples, targets)` where `samples` is one batch of input data and \n",
        "`targets` is the corresponding array of target temperatures. It takes the following arguments:\n",
        "\n",
        "* `data`: The original array of floating point data, which we just normalized in the code snippet above.\n",
        "* `lookback`: How many timesteps back should our input data go.\n",
        "* `delay`: How many timesteps in the future should our target be.\n",
        "* `min_index` and `max_index`: Indices in the `data` array that delimit which timesteps to draw from. This is useful for keeping a segment \n",
        "of the data for validation and another one for testing.\n",
        "* `shuffle`: Whether to shuffle our samples or draw them in chronological order.\n",
        "* `batch_size`: The number of samples per batch.\n",
        "* `step`: The period, in timesteps, at which we sample data. We will set it 6 in order to draw one data point every hour."
      ]
    },
    {
      "metadata": {
        "id": "yxUhBRARgoQc",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def generator(data, lookback, delay, min_index, max_index,\n",
        "              shuffle=False, batch_size=128, step=6):\n",
        "    if max_index is None:\n",
        "        max_index = len(data) - delay - 1\n",
        "    i = min_index + lookback\n",
        "    while 1:\n",
        "        if shuffle:\n",
        "            rows = np.random.randint(\n",
        "                min_index + lookback, max_index, size=batch_size)\n",
        "        else:\n",
        "            if i + batch_size >= max_index:\n",
        "                i = min_index + lookback\n",
        "            rows = np.arange(i, min(i + batch_size, max_index))\n",
        "            i += len(rows)\n",
        "\n",
        "        samples = np.zeros((len(rows),\n",
        "                           lookback // step,\n",
        "                           data.shape[-1]))\n",
        "        targets = np.zeros((len(rows),))\n",
        "        for j, row in enumerate(rows):\n",
        "            indices = range(rows[j] - lookback, rows[j], step)\n",
        "            samples[j] = data[indices]\n",
        "            targets[j] = data[rows[j] + delay][5] # change to 5 as temperature\n",
        "        yield samples, targets"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Eu5fxIqlgoQm",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "\n",
        "Now let's use our abstract generator function to instantiate three generators, one for training, one for validation and one for testing. \n",
        "Each will look at different temporal segments of the original data: the training generator looks at the first 200,000 timesteps, the \n",
        "validation generator looks at the following 100,000, and the test generator looks at the remainder."
      ]
    },
    {
      "metadata": {
        "id": "5kKT16HrgoQm",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# lookback = 720\n",
        "# Modified by Model 1 c)\n",
        "lookback = 432\n",
        "step = 6\n",
        "delay = 144 # This is a 24 hours delay\n",
        "batch_size = 128\n",
        "\n",
        "train_gen = generator(float_data,\n",
        "                      lookback=lookback,\n",
        "                      delay=delay,\n",
        "                      min_index=0,\n",
        "                      max_index=200000,\n",
        "                      shuffle=True,\n",
        "                      step=step, \n",
        "                      batch_size=batch_size)\n",
        "val_gen = generator(float_data,\n",
        "                    lookback=lookback,\n",
        "                    delay=delay,\n",
        "                    min_index=200001,\n",
        "                    max_index=300000,\n",
        "                    step=step,\n",
        "                    batch_size=batch_size)\n",
        "test_gen = generator(float_data,\n",
        "                     lookback=lookback,\n",
        "                     delay=delay,\n",
        "                     min_index=300001,\n",
        "                     max_index=None,\n",
        "                     step=step,\n",
        "                     batch_size=batch_size)\n",
        "\n",
        "# This is how many steps to draw from `val_gen`\n",
        "# in order to see the whole validation set:\n",
        "val_steps = (300000 - 200001 - lookback) // batch_size\n",
        "\n",
        "# This is how many steps to draw from `test_gen`\n",
        "# in order to see the whole test set:\n",
        "test_steps = (len(float_data) - 300001 - lookback) // batch_size"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "APgzTXL-goRK",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## A recurrent model\n",
        "\n",
        "\n",
        "Instead of the `LSTM` layer introduced in the previous section, we will use the `GRU` layer, developed by Cho et al. in 2014. `GRU` layers \n",
        "(which stands for \"gated recurrent unit\") work by leveraging the same principle as LSTM, but they are somewhat streamlined and thus cheaper \n",
        "to run, albeit they may not have quite as much representational power as LSTM. This trade-off between computational expensiveness and \n",
        "representational power is seen everywhere in machine learning.\n",
        "\n",
        "The model below uses recurrent dropout to help prevent overfitting.  "
      ]
    },
    {
      "metadata": {
        "id": "0VKAkSyggoRZ",
        "colab_type": "code",
        "outputId": "f37fc14e-7b6a-4b11-b160-19e04ed842f1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 735
        }
      },
      "cell_type": "code",
      "source": [
        "from keras.models import Sequential\n",
        "from keras import layers, Input, Model\n",
        "from keras.optimizers import RMSprop\n",
        "\n",
        "baseline = Sequential()\n",
        "baseline.add(layers.Flatten(input_shape=(lookback // step, float_data.shape[-1])))\n",
        "baseline.add(layers.Dense(32, activation='relu'))\n",
        "baseline.add(layers.Dense(1))\n",
        "\n",
        "model = Sequential()\n",
        "#model.add(layers.GRU(32,\n",
        "#                     dropout=0.2,\n",
        "#                     recurrent_dropout=0.2,\n",
        "#                     input_shape=(None, float_data.shape[-1])))\n",
        "#model.add(layers.Dense(1))\n",
        "# Modified by Model 1 b)\n",
        "\n",
        "Data_input = Input(shape = [None, float_data.shape[-1]])\n",
        "GRU_layer = layers.GRU(16,dropout=0.2,recurrent_dropout=0.2,\n",
        "                       input_shape=Data_input.shape)(Data_input)\n",
        "# Temperature from 24 hour ago\n",
        "T24_input = layers.Lambda(lambda x: x[:,-1,5])(Data_input)\n",
        "\n",
        "Dense_layer = layers.Dense(1)(GRU_layer)\n",
        "Merge_layer = layers.add([Dense_layer, T24_input])\n",
        "\n",
        "model = Model(inputs = Data_input, outputs = Merge_layer)\n",
        "\n",
        "model.summary()\n",
        "\n",
        "from keras.utils.vis_utils import plot_model\n",
        "# Plot model graph\n",
        "plot_model(model, show_shapes=True, show_layer_names=True, to_file='model.png')\n",
        "from IPython.display import Image\n",
        "Image(filename='model.png')\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_2 (InputLayer)            (None, None, 18)     0                                            \n",
            "__________________________________________________________________________________________________\n",
            "gru_2 (GRU)                     (None, 16)           1680        input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_8 (Dense)                 (None, 1)            17          gru_2[0][0]                      \n",
            "__________________________________________________________________________________________________\n",
            "lambda_2 (Lambda)               (None,)              0           input_2[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "add_2 (Add)                     (None, 1)            0           dense_8[0][0]                    \n",
            "                                                                 lambda_2[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 1,697\n",
            "Trainable params: 1,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwAAAAGVCAIAAADG1wEAAAAABmJLR0QA/wD/AP+gvaeTAAAgAElE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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "metadata": {
        "id": "BblYefbtkrBJ",
        "colab_type": "code",
        "outputId": "619249ba-0128-4050-af1f-774453777a85",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1404
        }
      },
      "cell_type": "code",
      "source": [
        "\n",
        "model.compile(optimizer=RMSprop(), loss='mae')\n",
        "history = model.fit_generator(train_gen,\n",
        "                              steps_per_epoch=500,\n",
        "                              epochs=40,\n",
        "                              validation_data=val_gen,\n",
        "                              validation_steps=val_steps)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/40\n",
            "500/500 [==============================] - 31s 62ms/step - loss: 0.3105 - val_loss: 0.2642\n",
            "Epoch 2/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2910 - val_loss: 0.2621\n",
            "Epoch 3/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2873 - val_loss: 0.2584\n",
            "Epoch 4/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2844 - val_loss: 0.2581\n",
            "Epoch 5/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2823 - val_loss: 0.2575\n",
            "Epoch 6/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2791 - val_loss: 0.2590\n",
            "Epoch 7/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2781 - val_loss: 0.2579\n",
            "Epoch 8/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2787 - val_loss: 0.2563\n",
            "Epoch 9/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2780 - val_loss: 0.2586\n",
            "Epoch 10/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2760 - val_loss: 0.2566\n",
            "Epoch 11/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2758 - val_loss: 0.2560\n",
            "Epoch 12/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2747 - val_loss: 0.2556\n",
            "Epoch 13/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2735 - val_loss: 0.2588\n",
            "Epoch 14/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2720 - val_loss: 0.2563\n",
            "Epoch 15/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2722 - val_loss: 0.2566\n",
            "Epoch 16/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2729 - val_loss: 0.2551\n",
            "Epoch 17/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2714 - val_loss: 0.2553\n",
            "Epoch 18/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2700 - val_loss: 0.2549\n",
            "Epoch 19/40\n",
            "500/500 [==============================] - 31s 61ms/step - loss: 0.2716 - val_loss: 0.2566\n",
            "Epoch 20/40\n",
            "500/500 [==============================] - 32s 64ms/step - loss: 0.2705 - val_loss: 0.2571\n",
            "Epoch 21/40\n",
            "500/500 [==============================] - 31s 63ms/step - loss: 0.2689 - val_loss: 0.2551\n",
            "Epoch 22/40\n",
            "500/500 [==============================] - 32s 64ms/step - loss: 0.2697 - val_loss: 0.2552\n",
            "Epoch 23/40\n",
            "500/500 [==============================] - 31s 62ms/step - loss: 0.2703 - val_loss: 0.2554\n",
            "Epoch 24/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2689 - val_loss: 0.2571\n",
            "Epoch 25/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2690 - val_loss: 0.2549\n",
            "Epoch 26/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2687 - val_loss: 0.2554\n",
            "Epoch 27/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2685 - val_loss: 0.2561\n",
            "Epoch 28/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2679 - val_loss: 0.2558\n",
            "Epoch 29/40\n",
            "500/500 [==============================] - 31s 61ms/step - loss: 0.2678 - val_loss: 0.2559\n",
            "Epoch 30/40\n",
            "500/500 [==============================] - 31s 62ms/step - loss: 0.2665 - val_loss: 0.2539\n",
            "Epoch 31/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2669 - val_loss: 0.2528\n",
            "Epoch 32/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2652 - val_loss: 0.2537\n",
            "Epoch 33/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2677 - val_loss: 0.2524\n",
            "Epoch 34/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2669 - val_loss: 0.2531\n",
            "Epoch 35/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2651 - val_loss: 0.2534\n",
            "Epoch 36/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2658 - val_loss: 0.2538\n",
            "Epoch 37/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2657 - val_loss: 0.2529\n",
            "Epoch 38/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2652 - val_loss: 0.2556\n",
            "Epoch 39/40\n",
            "500/500 [==============================] - 30s 60ms/step - loss: 0.2651 - val_loss: 0.2539\n",
            "Epoch 40/40\n",
            "500/500 [==============================] - 30s 61ms/step - loss: 0.2652 - val_loss: 0.2535\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "ocgxmgoKgoRh",
        "colab_type": "code",
        "outputId": "35bedba7-c993-401e-c745-9cf166131b34",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        }
      },
      "cell_type": "code",
      "source": [
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(loss))\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAecAAAFZCAYAAACizedRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzs3XlcVNX/x/HXLKyCBQYuZaWWe1ba\nZpqmYuJSZr9SstLSNDNLyxYjFc1c2kytvmmLlVmJGVp9S80ly8yyMssFc6nI3ABFBVln+f0xX0eJ\nbYCBucD7+XjwgLkz997zuXeYz5xzzz3H5HQ6nYiIiIhhmH1dABEREclPyVlERMRglJxFREQMRslZ\nRETEYJScRUREDEbJWURExGCUnKVai4uLIzo6mujoaFq1akWXLl3cjzMyMkq1rejoaFJTU4t9zYsv\nvsiHH35YniJ73d13301CQoJXttWsWTMOHTrEqlWrePLJJ8u1v8WLF7v/9uTYemrcuHH85z//8cq2\nRHzF6usCiFSkyZMnu//u2rUrzz33HFdccUWZtrVixYoSXzN27Ngybbuq6d69O927dy/z+ikpKbz5\n5pv0798f8OzYitQkqjlLjXbXXXfx0ksv0bNnTzZv3kxqaipDhw4lOjqarl278vbbb7tfe6rW+MMP\nPzBgwABefPFFevbsSdeuXdm0aROQv9bWtWtXFi1axK233krHjh2ZMWOGe1tz586lffv2/N///R/v\nv/8+Xbt2LbR8H330ET179uSGG27gjjvuYP/+/QAkJCTw0EMPERsbS48ePejVqxe7d+8GYN++fdx2\n221ERUUxduxY7HZ7ge1+/fXX3HjjjfmW9e3bl2+++abYY3BKQkICd999d4n7W7NmDTfeeCM9evTg\nlltuITExEYCYmBgOHDhAdHQ0ubm57mMLsGDBAnr16kV0dDT3338/R48edR/bOXPmcM8999ClSxfu\nuecesrKyijq1AOzcuZOYmBiio6Pp27cv69evB+DkyZM88MAD9OzZk27dujF+/Hjy8vKKXC5S2ZSc\npcbbtm0bn3/+OW3btuW1117jvPPOY8WKFbz77ru8+OKLHDx4sMA6O3bs4NJLL2X58uUMHDiQ1157\nrdBt//jjj8THx/Pxxx+zcOFCDh06xO7du3nzzTf55JNP+OCDD4qsNR45coSnn36at99+my+//JLz\nzz8/X3PtN998w8CBA1m5ciVXX3017777LgAvvPAC7du3Z/Xq1QwePJjNmzcX2Hb79u05dOgQ+/bt\nA1wJ9tChQ1x77bUeH4NTitqfzWZj3LhxTJkyhZUrV9K1a1eeffZZAKZNm0b9+vVZsWIF/v7+7m1t\n2bKFt956i/fee48VK1bQoEEDXnzxRffzK1as4KWXXmLVqlUcPXqUVatWFVkuh8PBI488wp133smK\nFSt45plnGDt2LBkZGSxbtozatWuzfPlyVq5cicViYc+ePUUuF6lsSs5S43Xu3Bmz2fWvMH78eCZM\nmABAw4YNiYiI4J9//imwTq1atYiKigKgVatWHDhwoNBt33jjjVgsFurWrUudOnU4ePAgP/74I1dd\ndRWRkZEEBATwf//3f4WuW6dOHX7++Wfq1asHwBVXXOFOpgBNmjShdevWALRs2dKdQH/66Sd69eoF\nQJs2bWjcuHGBbfv7+9OlSxfWrl0LwOrVq4mKisJqtXp8DE4pan9Wq5XvvvuOyy67rNDyF2bdunX0\n6NGDOnXqAHDbbbexYcMG9/OdO3fm7LPPxmq10rRp02K/NPzzzz+kpqbSu3dvAC655BIaNGjA1q1b\nCQ8P55dffuHbb7/F4XAwefJkWrRoUeRykcqma85S45111lnuv7du3equKZrNZlJSUnA4HAXWCQ0N\ndf9tNpsLfQ1ASEiI+2+LxYLdbufEiRP59lm3bt1C17Xb7cyZM4e1a9dit9s5efIkjRo1KrQMp7YN\ncPz48Xz7rV27dqHb79GjBwsWLGDw4MGsXr2akSNHluoYnFLc/t577z2WLl1Kbm4uubm5mEymIrcD\ncPToUSIjI/Nt68iRIyXGXNS2QkND8+2zdu3aHD16lN69e3P8+HFmz57NH3/8wU033cSTTz5Jz549\nC11+Zu1epDKo5ixyhscee4wePXqwcuVKVqxYQVhYmNf3ERISQmZmpvtxcnJyoa/74osvWLt2LQsX\nLmTlypU89NBDHm2/du3a+Xqin7pm+2/XXXcdO3fu5K+//uKvv/7immuuAUp/DIra3+bNm3njjTd4\n7bXXWLlyJc8880yJZT/nnHM4duyY+/GxY8c455xzSlyvMHXq1OH48eOcObfPsWPH3LXymJgYPvro\nI7744gu2b9/OsmXLil0uUpmUnEXOcOTIEVq3bo3JZGLp0qVkZWXlS6Te0KZNG3744QeOHj1Kbm5u\nkR/+R44c4dxzzyU8PJy0tDSWL1/OyZMnS9z+ZZdd5r4Wu3nzZv7+++9CX+fv70/Hjh15/vnn6dat\nGxaLxb3f0hyDovZ39OhR6tSpQ4MGDcjKymLp0qVkZmbidDqxWq1kZmZis9nybev6669n1apVpKWl\nAbBo0SI6d+5cYsyFOe+886hXrx5ffPGFu2ypqam0adOGV199lSVLlgCulovzzjsPk8lU5HKRyqbk\nLHKG0aNH88ADD3DjjTeSmZnJgAEDmDBhQpEJrizatGlDv3796NevH4MGDaJLly6Fvq5Pnz4cO3aM\n7t27M3bsWMaMGcOhQ4fy9fouzGOPPcZXX31FVFQU77//Ptdee22Rr+3RowerV6+mZ8+e7mWlPQZF\n7e+6664jMjKSqKgohgwZwuDBgwkNDeWhhx6iWbNmnHXWWXTo0CHf9fo2bdowfPhw7rjjDqKjo0lP\nT+fhhx8uNt6imEwmZs6cycKFC+nZsyfPPPMMs2fPJjg4mL59+/LJJ5/Qo0cPoqOj8fPzo2/fvkUu\nF6lsJs3nLFL5nE6nu0a2bt06Zs2apeZTEXFTzVmkkh09epRrrrmG/fv343Q6Wb58ubtHs4gIqOYs\n4hMffvgh8+fPx2Qy0bhxY6ZOneruqCQiouQsIiJiMGrWFhERMRglZxEREYMxzAhhKSnpXt1eWFgw\naWnevT/VlxSPsSkeY1M8xlZT44mICC3yuWpbc7ZaLb4uglcpHmNTPMameIxN8RRUbZOziIhIVaXk\nLCIiYjBKziIiIgaj5CwiImIwSs4iIiIGo+QsIiJiMErOIiIiBmOYQUhERKT6ePnll/j990SOHj1C\ndnY2DRqcS+3aZzFt2vMlrvvFF59Rq1YInTsXPtf57NkvctttMTRocG6ZyjZq1HAeeeRxGje+qEzr\nV4Zql5yXLrUya5Y/u3ZB06bBjBmTS79+Nl8XS0TE0E5/dppp2tRR7s/OBx98GHAl2j/+2MuoUWM8\nXrdXrxuLfX706LFlLldVUa2S89KlVu67L8j9ODHR8r/HWUrQIiJFqMzPzs2bf2LRooVkZmYyatTD\n/PLLz2zYsI6cnDzat+/AkCHDeeuteZx99tk0atSEhITFmExmkpL+5PrruzFkyHB3zferr9Zw8mQG\nf/+dxP79//DQQ2Np374DCxe+w+rVX9KgwbnYbDZiYu6gbdsrCpQlIyODqVMnkZGRjs1mY8yYx2jW\nrDmzZj3Pzp2J2O12+vW7lV69bix0WUXy6JrztGnTGDBgADExMfz222/5nlu8eDH9+/cnJiaGSZMm\ncWoGyl27dhEVFcXChQu9X+oizJrlX+jy2bMLXy4iIpX/2bl37x5mznyF5s1bAPDBBx/w+uvvsHz5\nfzl5MiPfa3fs2M5TT01i7ty3+fjj+ALbSk4+zAsvzGH06Ef59NMETpw4TkLCR8ybN59HHx3Hli2b\niyzHRx99SKtWrXn55XmMHj2Wl1+eyYkTx/nuu2+ZO3c+r732FjabrdBlFa3EmvOmTZtISkoiPj6e\nvXv3EhsbS3y86wBlZWXx+eef8/777+Pn58egQYP45ZdfaN68OVOmTKF9+/YVHsCZdu0q/LtGUctF\nRKTyPzsvuuhi/P1diT8wMJA777wThwOOHTvGiRMn8r22WbPmBAYGFrmtNm0uAyAyMpKMjAz++Wcf\njRs3ISAgkICAQFq0aFXkujt37mDQoKEANG/ekn/+2Uft2mfRsOEFjBv3CF26RBEd3Rt/f/8Cyypa\niUd+48aNREVFAdCkSROOHz9ORobrm01QUBDvvvsufn5+ZGVlkZGRQUREBP7+/rzxxhtERkZWbOn/\npWlTR6mWi4hI5X92+vn5AXDo0EHi49/nzTff5JVXXqdevXoFXmuxFD+JxJnPO51OnE4wm0+nNpOp\n6HVNJpO7tRfA4XDF++KLc7jnnuHs3r2LJ554uMhlFanE5JyamkpYWJj7cXh4OCkpKfle8/rrr9O9\ne3eio6Np2LAhVqu12G86FWXMmNxCl48eXfhyERHx3WfnsWPHCAsLo1atWvz++04OHTpEXl5eubZZ\nv359/vhjLzabjbS0NHbuTCzytc2bt+SXX34CYNu2rTRq1ISDBw/w0UeLaNasOaNGjeH48eOFLqto\npe4Qdua3jFOGDx/OoEGDGDZsGO3ataNdu3alLkhYWHC5p9kaPhxq14bp02HHDmjZEp58EmJigkpe\nuQoobu7PqkjxGJviMTZvxlORn52hoYEEB/u7y3v22cEEBPgRERFKeHg73nmnNjExMbRr147bb4/h\n5ZdfoF27doSEBOZ7LbhquhERofj7WwkLq0WtWgGEhAQSERFKWlot/P2tNGt2IX373sT9999DkyZN\nuOyyS6lTJzTf8Tq1/v33DyM2NpaxYx/A6XQyefJELrjgAt5+ey4PPjgMPz8/YmL607x5owLLSjr+\n5T0/Jmdh2fYML7/8MhEREcTExADQrVs3PvnkE0JCQjh27Bi7d+/myiuvBOCNN94AYNiwYe51w8LC\nuPPOO0ssSEpKerkC+beIiFCvb9OXFI+xKR5jUzzG5u14vvjiM7p3j8ZisTBoUAwzZ75MZGRdr22/\nJJ7GU1wCL7FZu0OHDqxcuRKA7du3ExkZSUhICAA2m41x48Zx8uRJALZu3UqjRo08KryIiEhFOHLk\nCMOHD2bEiCHccEN0pSZmbymxWbtt27a0atWKmJgYTCYTcXFxJCQkEBoaSvfu3XnggQcYNGgQVquV\nZs2a0a1bN7Zt28azzz7L/v37sVqtrFy5kpdffpmzzz67MmISEZEa7K677uauu+72dTHKpcRm7cqi\nZu3iKR5jUzzGpniMrabGU65mbREREalcSs4iIiIGo+QsIiJiMErOIiLidffdd0+BAUDmzn2FDz8s\nfL6FzZt/Yvz4xwEYN+6RAs9//HE8b701r8j97dmzm7//TgIgLu5JcnKyy1p0br31RjIzM8u8vjco\nOYuIiNd1796DtWtX5Vu2bt1aoqJuKHHdGTNmlnp/X3+9ln37/gZg8uTpBARU/iiV3lStpowUERFj\n6NbtBu6/fygjRz4EwM6diURERBAREcmPP/7Am2/Oxc/Pj9DQUP7zn1fyrdu7dzc+/3wNP/20iTlz\nXiQ8vA516pzjngJy6tRJpKQkk5WVxZAhw6lXrz6ffJLA11+vJSwsjIkTn2TBgngyMtKZPv1p8vLy\nMJvNjBs3AZPJxNSpk2jQ4Fz27NlN06bNGDduQqExJCcfLrB+ZGRdnn56AkeOpJKbm8vQofdxxRVX\n5Vv2yCNjaNHi8nIdPyVnEZFqbtKkAD77zLsf9zfeaGPSpJwinw8LC6dBg3PZsWMbLVu2Zu3aVXTv\nHg1Aeno6cXHP0KDBuUyZMpFvv/0WKDhDxbx5rzBhwhQuvrgpjz76EA0anEt6+gmuuuoaevbsw/79\n/zBhwjjmz1/I1Ve35/rru9GyZWv3+m++OZc+ffrSrdsNfPXVaubPf52hQ+/j998TmTx5GmFh4fTr\n14v09HRCQwve1lTY+rfddjvHjx/j1VffID09nY0bN7B37558y7Zv/7ncx1fN2iIiUiG6d49mzRpX\n0/aGDd9w/fXdADj77LN59tlnGDVqOL/88jPHjh0rdP2DBw9y8cVNAbjssrYAhIbWJjFxO/ffP4Sp\nUydx4kTRk1D8/nsil1/umuuhbdsr2L37dwDOPbchdeqcg9ls5pxzIgrMIV3c+hdccCGZmSeZMmUC\nmzf/SFTUDQWW9e5d/iklVXMWEanmJk3KKbaWW1E6d+7CggXz6d69Bw0bnk/t2rUBmD59Cs8/P4sL\nL2zEzJnPFrn+mVM/nhova9WqFZw4cYJXX32TEydOcO+9dxVTgtNTQubl2TCZXNv79zSURY/FVXD9\nwMBA5s17h61bf2P58s/YsGE9sbFx+Zb9/PP3PPJIbHGHpkSqOYuISIUIDq5FkyYXs2DB2+4mbYCT\nJzOoW7ce6enpbN78c5HTRJ5zTgR///0XTqeTX35xNRUfO3aM+vUbYDab+frrte51TSYTdrs93/ot\nWrRk82bXlJBbtvxM8+YtSlX+wtb//fedrFq1gksvvYxHH32Sv/76s8CyvXv3lmo/hVHNWUREKkz3\n7tE880wccXFT3MtuueU27r9/KA0bns8ddwxi3rx53Hvv/QXWHT58JOPHP0G9evXdk1dcf31Xxo17\nhB07ttG7901ERkby9ttvcOmllzNr1vMEBwe717/33hFMnz6Fzz5bhtXqx5NPTsBms3lc9sLWDwgI\nZN68V/nkkwTMZjMDB95F/foN8i0bOnRoOY6Yi8bWriIUj7EpHmNTPMZWU+PR2NoiIiJViJKziIiI\nwSg5i4iIGIySs4iIiMEoOYuIiBiMkrOIiIjBKDmLiIgYjJKziIiIwSg5i4iIGIySs4iIiMEoOYuI\niBiMkrOIiIjBKDmLiIgYjJKziIiIwSg5i4iIGIySs4iIiMEoOYuIiBiMkrOIiIjBKDmLiIgYjJKz\niIiIwSg5i4iIGIySs4iIiMEoOYuIiBiMkrOIiIjBKDmLiIgYjJKziIiIwVg9edG0adP49ddfMZlM\nxMbG0qZNG/dzixcvZsmSJZjNZpo3b05cXBwmk6nYdURERKRoJSbnTZs2kZSURHx8PHv37iU2Npb4\n+HgAsrKy+Pzzz3n//ffx8/Nj0KBB/PLLL9hstiLXERERkeKV2Ky9ceNGoqKiAGjSpAnHjx8nIyMD\ngKCgIN599138/PzIysoiIyODiIiIYtcRERGR4pVYc05NTaVVq1bux+Hh4aSkpBASEuJe9vrrr7Ng\nwQIGDRpEw4YNPVrn38LCgrFaLWWNo1AREaFe3Z6vKR5jUzzGpniMTfHk59E15zM5nc4Cy4YPH86g\nQYMYNmwY7dq182idf0tLyyxtUYoVERFKSkq6V7fpS4rH2BSPsSkeY6up8RSXwEts1o6MjCQ1NdX9\nODk5mYiICACOHTvGjz/+CEBgYCCdOnVi8+bNxa4jIiIixSsxOXfo0IGVK1cCsH37diIjI93N0zab\njXHjxnHy5EkAtm7dSqNGjYpdR0RERIpXYrN227ZtadWqFTExMZhMJuLi4khISCA0NJTu3bvzwAMP\nMGjQIKxWK82aNaNbt26YTKYC64iIiIhnTE5PLghXAm9fb6ip1zCqCsVjbIrH2BSPsVXKNWcRERGp\nXErOIiIiBqPkLCIiYjBKziIiIgaj5CwiImIwSs4iIiIGo+QsIiJiMErOIiIiBqPkLCIiYjBKziIi\nIgaj5CwiImIwSs4iIiIGo+QsIiJiMErOIiIiBqPkLCIiYjBKziIiIgaj5CwiImIwSs4iIiIGo+Qs\nIiJiMErOIiIiBqPkLCIiYjBKziIiIgaj5CwiImIwSs4iIiIGo+QsIiJiMErOIiIiBqPkLCIiYjBK\nziIiIgaj5CwiImIwSs4iIiIGo+QsIiJiMErOIiIiBqPkLCIiYjBKziIiIgZTo5Pz0qVWOncOpn79\nEDp3DmbpUquviyQiIkKNzUZLl1q5774g9+PERMv/HmfRr5/NdwUTEZEar8bWnGfN8i90+ezZhS8X\nERGpLDU2Oe/aVXjoRS0XERGpLB41a0+bNo1ff/0Vk8lEbGwsbdq0cT/3/fffM3PmTMxmM40aNWLq\n1KkAxMXFsXv3bvz8/Jg0aRJNmjSpmAjKqGlTB4mJlkKXi4iI+FKJ1cRNmzaRlJREfHw8U6dOdSff\nUyZOnMicOXNYtGgRJ0+eZP369axZs4b09HQWLVrE1KlTee655yosgLIaMya30OWjRxe+XEREpLKU\nWHPeuHEjUVFRADRp0oTjx4+TkZFBSEgIAAkJCe6/w8PDSUtLIyUlxV27Pv/88zlw4AB2ux2LpWBN\n1Vdcnb6ymD3bn127zDRt6mD06Fx1BhMREZ8rMTmnpqbSqlUr9+Pw8HBSUlLcCfnU7+TkZDZs2MDo\n0aP57bffePfddxk8eDBJSUns27ePtLQ0zjnnnCL3ExYWjNXq3eQdERFa7PPDh7t+XCxAUDGv9r2S\n4qlqFI+xKR5jUzzGVt54Sn0rldPpLLDsyJEjjBgxgri4OMLCwujcuTObN2/mjjvuoFmzZjRu3LjQ\n9c6UlpZZ2qIUKyIilJSUdK9u05cUj7EpHmNTPMZWU+MpLoGXmJwjIyNJTU11P05OTiYiIsL9OCMj\ng2HDhjFmzBg6duzoXv7www+7/46KiqJOnTolFlREREQ86BDWoUMHVq5cCcD27duJjIx0N2UDzJgx\ng8GDB9OpUyf3sp07d/Lkk08C8M0339CyZUvMZt2iJCIi4okSa85t27alVatWxMTEYDKZiIuLIyEh\ngdDQUDp27MiyZctISkpiyZIlAPTp04fbbrsNp9PJrbfeSkBAAC+88EKFByIiIlJdeHTN+dFHH833\nuHnz5u6/t23bVug6M2bMKEexREREai61NYuIiBiMkrOIiIjBKDmLiIgYjJKziIiIwSg5e2DpUiud\nOwdTv34InTsHs3RpjZ0GW0REKoGyTAmWLrVy332nh/VMTLT873GWxuEWEZEKoZpzCWbN8i90+ezZ\nhS8XEREpLyXnEuzaVfghKmq5iIhIeSnDlKBpU0eplouIiJSXknMJxozJLXT56NGFLxcRESkvJecS\n9OtnY968LFq2tGO1OmnZ0s68eeoMJiIiFUe9tT3Qr59NyVhERCqNas4iIiIGo+QsIiJiMErOIiIi\nBqPkLCIiYjBKziIiIgaj5OxFmiBDRES8QdnDSzRBhoiIeItqzl6iCTJERMRblJy9RBNkiIiItyhz\neIkmyBAREW9RcvYSTZAhIiLeouTsJZogQ0REvEW9tb1IE2SIiIg3qOYsIiJiMErOIiIiBqPk7CMa\nTUxERIqijOADGk1MRESKo5qzD2g0MRERKY6Ssw9oNDERESmOsoEPaDQxEREpjpKzD2g0MRERKY6S\nsw9oNDERESmOemv7iEYTExGRoqjmLCIiYjBKziIiIgbjUbP2tGnT+PXXXzGZTMTGxtKmTRv3c99/\n/z0zZ87EbDbTqFEjpk6dSlZWFk888QTHjx8nLy+PBx54gOuuu67CghAREalOSkzOmzZtIikpifj4\nePbu3UtsbCzx8fHu5ydOnMiCBQuoV68eDz30EOvXr2ffvn00atSIsWPHcvjwYQYPHsyKFSsqNBAR\nEZHqosRm7Y0bNxIVFQVAkyZNOH78OBkZGe7nExISqFevHgDh4eGkpaURFhbGsWPHADhx4gRhYWEV\nUXYREZFqqcSac2pqKq1atXI/Dg8PJyUlhZCQEAD37+TkZDZs2MDo0aMJCwsjISGB7t27c+LECebN\nm1dBxRcREal+Sn0rldPpLLDsyJEjjBgxgri4OMLCwvjkk09o0KABb731Fjt37iQ2NpaEhIRitxsW\nFozVailtcYoVERHq1e35muIxNsVjbIrH2BRPfiUm58jISFJTU92Pk5OTiYiIcD/OyMhg2LBhjBkz\nho4dOwKwefNm99/NmzcnOTkZu92OxVJ08k1LyyxzEIWJiAglJSXdq9v0haVLrcya5c+uXRaaNrUz\nZkxutbg/urqcn1MUj7EpHmOrqfEUl8BLvObcoUMHVq5cCcD27duJjIx0N2UDzJgxg8GDB9OpUyf3\nsgsuuIBff/0VgP3791OrVq1iE7MU7tTUkomJFuz201NLau5nEZHqrcRP+bZt29KqVStiYmIwmUzE\nxcWRkJBAaGgoHTt2ZNmyZSQlJbFkyRIA+vTpw4ABA4iNjeXOO+/EZrMxadKkio6jWipuasny1J5P\n18bNNG3qqDa1cRGR6sLkLOwisg94u0mjOjST1K8fgt1uKrDcanVy4EBGIWuU7FRt/N8qe2zv6nB+\nzqR4jE3xGFtNjadczdriOxUxtWRxtXERETEGJWcDK+3UkkuXWuncOZj69UPo3Dm40GvTu3YVfsoL\nW+7J9kRExPv0aWtgrmbmLGbPPt1be/Towq8P/7u5+lTnMcjfXN20qYPExIKd8/5dG/d0eyIi4n2q\nORtcv3421q3LJC8P1q3LLDIxetpc7WltXM3fIiK+o+RcTXjaXN2vn41587Jo2dKO1eqkZUt7oZ3B\nStP8LSIi3qVm7WrC0+ZqcCXokpqmS7M9ERHxLlWDqonSdh6r7O2JiIjnlJyrCU+bq321PRER8Zya\ntasRT5qrfbk9ERHxjGrOIiIiBqPkLCIiYjBKzuIVGk1MRMR7lJyl3PJPbWkqdmrLU0ncakVJXESk\nCErOUm6ejiam+alFRDyj5Czl5uloYhoSVETEM0rOUm6eTm2pIUFFRDyjT0UpN09HE6uI+alFRKoj\nJWcpN09HE6uI+alFRKojfdqJV3gymlhFzE8tIlIdKTlLpTqVxCMiQklJySzydcV1HlNyFpHqTs3a\nYkjqPCYiNZk+6cSQ1HlMRGoyJWcxpNJ0HlPHMRGpbvQpJoaUv/OYmaZNHYV2HlPHMRGpjpScxbA8\n6QGujmMiUh2pWVuqNF93HNNEHiJSEZScpUrzZccxTeQhIhVFyVmqtNKOOuZNmshDRCqKkrNUaZ4O\nHVoRfN2kLiLVlz5FpMrr18/GunWZHDiQwbp1mUUm5tLccuXJa3UvtohUFCVnqRHyXx82FXt92NPX\n+rJJXUSqNyVnqRFKc33Y09fmb1KnUpvURaR6U3KWGqE014dL89pTTep5eRTbpA7eH8lMI6OJVF9K\nzlIjlOb6cEVcSy5Ns7ovtid3GzQ/AAAgAElEQVQixqLkLDVCaa4PV8S1ZG/fdqXbuESqNyVnqRFK\nc8tVRdye5e3brnQbl0j1pjYwqTE8Gau7LK/1RNOmDhITLYUuN8L2RMRY9DVbpBJ4u6m8tNtT5zGR\nqsWj/9Bp06bx66+/YjKZiI2NpU2bNu7nvv/+e2bOnInZbKZRo0ZMnTqVjz/+mE8//dT9mm3btvHL\nL794v/QiVYSnU2BWxPZ8Pa3m0qVWZs06Xc4xY8oet0hNUWJy3rRpE0lJScTHx7N3715iY2OJj493\nPz9x4kQWLFhAvXr1eOihh1i/fj233XYbt912m3v95cuXV1wEIlWEp03lniYzT7fny2k1ff3FQKSq\nKrFZe+PGjURFRQHQpEkTjh8/TkZGhvv5hIQE6tWrB0B4eDhpaWn51n/11VcZOXKkN8ssUm1VxC1S\nFdV5zJOmcvUqFymbEv87U1NTCQsLcz8ODw8nJSXF/TgkJASA5ORkNmzYQOfOnd3P/fbbb9SvX5+I\niAhvllmk2qqIZObL+7bVq1ykbEr9ddzpdBZYduTIEUaMGEFcXFy+RL5kyRL69evn0XbDwoKxWgv2\nPi2PiIhQr27P1xSPsXkjnl27ilpuKfP2J06E228vuHzChOK3Wdxzr7xS+PJXXw1i+PDTj1u2hK1b\nC76uZUtTpZ9/vd+MTfHkV2JyjoyMJDU11f04OTk5X004IyODYcOGMWbMGDp27Jhv3R9++IHx48d7\nVJC0tExPy+yRiIhQUlLSvbpNX1I8xuateJo2DS7iFik7KSll+x/p1g3mzbMW6DzWrZuNMxrB8ikp\nnh07QgBTIcudpKScvuw1alT+a86nPPBAFikplXfNWe83Y6up8RSXwEtsW+rQoQMrV64EYPv27URG\nRrqbsgFmzJjB4MGD6dSpU771Dh8+TK1atfD317UlEU9V1ExXpZ1W02ql2FuuPG0qL82ALt6e0lOk\nKivxHd22bVtatWpFTEwMJpOJuLg4EhISCA0NpWPHjixbtoykpCSWLFkCQJ8+fRgwYAApKSmEh4dX\neAAi1Ym3b7kqjdL0rB4zJrfQGnFhXyI86VVemn2X9rWunu+uVgndxiVVhclZ2EVkH/B2k0ZNbSap\nKhSP8XTuXHiTesuWdtatK9ikvnRpwabysia+0uzb09f+O4mfUlzNvarcj10d3m9nqqnxFNesrbYg\nEQFK37Pam0OcVsSUnqW5v1v3Y4vR6H4GEQEq5pariti3p68tTcLX/dhiNErOIgJUXGc0b+/b09eW\nJuGXJpF72hlNndakPJScRQT4d89qvDJVZtn27Z0pPUuT8D1N5J4OvlIRI71VFH2JMCZ1CKsiFI+x\nKR5jOt1pzULTpnaPJwc55d9J39POaKXtXFda3jo/pe00V1Gqy/vtFHUIExEpxqlOa64Py6KToqe3\nsHna/F1Vhi315aQoUjxjvVNERHzEk4FaPG3+rqjOdZ4OEuOp0n6JUBN45VFyFhHxkKfXsSuic13+\n69h45Tp2ab5EVKXr6NWBkrOIiIc87YxWmg5u4LvpN0vzJUK3m1UufeURESkFTwdf8fR1ng6AUhHX\nsUszXGxVuY5eXSg5i4j4kKedspo2dRQxY1n5rmN7+iWiovYvhdNXHhERH/K0RurLQWKMsP+aRslZ\nRMSHyjb9ZuUOElNw/yVfR/dEaXqf17Se4tU7OhERgyvL9Jsl3bddUbw52UlFTRNaXajmLCLiQxVR\nIwXj1zRL0/vb1z3FfXEslZxFRHzMkwFQSsPX9yR7kswqYppQT/ddmtf66lgqOYuIVDO+rGl6mswq\nYprQ0iRST1/rq2Op5CwiUs348p5kT5NZRUwTWhFN5b46lkrOIiLVTEWN7e0JT5NZaXqfe3pdviKa\nyn11LJWcRUSqGV/ek1yaZHbqWnteHiVea/fmxCSlea2vjqWSs4hINVNRPcA94csvBhXRVO6rY2ms\nvvVe8sEHVn77DSZMgFq1fF0aEZHK5817kku7X0/H6/blvkv72so+ltUyOScmWpg/H7ZtC+KDD7Ko\nXdvXJRIRqTl89cWgtPv2ZTlLUi2btSdOzCEmBjZtsnLrrcGkpfm6RCIiIp6rlsnZzw8WLoTbb89j\nyxYLN98cTHKyydfFEhER8Ui1TM4AFgu89FI2Q4bkkpho4eabgzhwQAlaRESMr9omZwCzGaZPz+GB\nB3LZs8fCTTcFk5SkBC0iIsZWrZMzgMnkugb92GM5/P23mb59g9m7VwlaRESMq9onZ3Al6Mcey2Xi\nxGwOHDBz003BJCbWiNBFRKQKqlEZatSoPKZPzyYlxczNNwfz6681KnwREakialx2Gjo0j1mzsjh2\nDG65JZhNm2rcIRAREYOrkZlp4EAbr72WTWYm9O+vGrSIiBhLjc1Kt9xi4403ssnMNPHoo4HY7b4u\nkYiIiEuNTc4AffrYuPXWPH791cJ77/n5ujgiIiJADU/OAHFxOYSEOJk2LYAjR3SLlYiI+F6NT851\n6zp5/PEcjh0zMW2av6+LIyIiouQMrh7czZvbWbjQj19+0SERERHfUibCNVHGjBk5OJ0mxo0LxOHw\ndYlERKQm8yg5T5s2jQEDBhATE8Nvv/2W77nvv/+e/v37ExMTw5NPPonjf5nt008/5aabbuKWW25h\n3bp1Xi+4t117rZ1bbsnjl18svP++OoeJiIjvlJicN23aRFJSEvHx8UydOpWpU6fme37ixInMmTOH\nRYsWcfLkSdavX09aWhqvvvoqH3zwAXPnzmXNmjUVFoA3TZqUQ61aTp55JoCjR31dGhERqalKTM4b\nN24kKioKgCZNmnD8+HEyMjLczyckJFCvXj0AwsPDSUtLY+PGjbRv356QkBAiIyOZMmVKBRXfu+rV\nc/LYYzmkpZmYNi3A18UREZEaqsTknJqaSlhYmPtxeHg4KSkp7schISEAJCcns2HDBjp37sw///xD\ndnY2I0aMYODAgWzcuLECil4xhg3Lo1kzO++958eWLbokLyIilc9a2hWcTmeBZUeOHGHEiBHExcW5\nE/mxY8d45ZVXOHDgAIMGDeKrr77CZCr6PuKwsGCsVktpi1OsiIjQMq332mvQtStMmFCL775zzQtt\nBGWNx6gUj7EpHmNTPMZW3nhKTM6RkZGkpqa6HycnJxMREeF+nJGRwbBhwxgzZgwdO3YEoE6dOlx+\n+eVYrVbOP/98atWqxdGjR6lTp06R+0lLyyxPHAVERISSkpJepnVbt4abbw5k2TI/5szJ5o478rxa\ntrIoTzxGpHiMTfEYm+IxNk/jKS6Bl1gn7NChAytXrgRg+/btREZGupuyAWbMmMHgwYPp1KmTe1nH\njh35/vvvcTgcpKWlkZmZma9pvCqYPDmH4GAnzzzjT1qar0sjIiI1SYk157Zt29KqVStiYmIwmUzE\nxcWRkJBAaGgoHTt2ZNmyZSQlJbFkyRIA+vTpw4ABA+jRowf9+/cHYPz48ZiN0jbsofr1nTz6aA5P\nPx3I9OkBPPdcjs/KsmaNhYMH4eab4YzvRSIiUk2ZnIVdRPYBbzdpeKOZJDcXunQJZs8eM6tWZdKm\nTeWOTpKRAePHB/DBB65hRcPDHTz4YC733JNHcHClFsXrvN2MtWGDBYsFrrnGN9OL1dRmuapC8Rhb\nTY2nXM3aNZm/P0yf7ho57IknKnfksO+/t3D99bX44AN/Wre2M3482GwmJk8O5Oqra/HWW37kVEJl\n/vffzTz7rD/Hj1f8vsrqt9/M3HZbEP36BbF2rXc7FYqI+IKScwk6dbJz0015/PyzhenT/fnuOwv7\n9pmw2Spmfzk5MGWKP337BvHPPyZGj85hxYpMpkyBH3/MYMyYHNLTTTz5ZCDXXluLDz6wVlhZdu82\n069fEC++GMBNNwVz8KDxZu3KzoYHHgjEZjNhscDQoUFs3aq3tYhUbWrW9sCBAyY6dKjFyZOnk5PF\n4qRBAyfnneegYUMnDRs6/vfjpFEjB+edV/rDmphoZuTIQLZvt3DBBQ5eeSWbq6+2F4gnJcXEnDn+\nvPOOHzk5Jho3dvD44zncfLPNa7d9/f23iRtvDObgQTPXXWdj/Xor557rID4+i6ZNy9+E4K3zExcX\nwGuv+TNkSC4dOti5995A6tZ1snx5JueeW3lv7ZraLFdVKB5jq6nxFNesreTsoT//NPHtt1b27TOx\nb5/Z/fvQIRNOZ8Ea5fnnO+jc2cb119vp2NFGcZ3VHQ6YO9ePadMCyM01cddduUyenJOv81dh8Rw8\naGLmTH/ef98Pm81EixZ2Hn88l169bBRzS3mJDh92Jea//jITF5fNyJF5zJnjz9SpAYSFOVm4MJMr\nryxfgvbG+fnuOwv9+gXRqJGTNWtOUquW6zhOnBhIixZ2Pvssk9q1y7ULj9XUD5eqQvEYW02NR8m5\nAuXmwv79Jv75x5Ww//7bTGKimQ0brJw44cqQZrOTSy91cP31Njp3tnPFFXb8/zd19L59Jh56KJAN\nG6ycc46Dl17KpkePgp2aiosnKcnECy8E8NFHVhwOE9HRecyalU14eOnjOXoUbr45mJ07LTzySA7j\nxuW6n1u0yMrDDwfi7w+vv55VaDk9Vd7zk54O119fi/37TXz+eSbt2rm+LDid8NRTAbz5pj/XXWfj\nww+z3Me6ItXUD5eqQvEYW02NR8nZB2w22LLFzLp1Vr7+2sLPP1uw2VzJOjjYybXX2mnVys78+f6k\np5vo2TOPF1/M4ZxzCj8dnsSze7eZceMCWL/eSv36Dl57LZtrr/U8gaanw//9XzBbtlgYPjyXKVNy\nCtTAV6+2cO+9QeTkwAsv5JR5gJbynp8xY1y92P/9BQLAbod77glkxQo/+vfP4+WXs8vVkuAJX7/f\nvE3xGJviMTb11jYwqxWuuMLBo4/m8tlnWezalcF772Vy7725nHuug9WrrcyeHYDDAbNnZ/HOO9lF\nJmZPXXyxg8WLs3jqqRySk0306xfEjBn+HnUYy8yEO+8MYssWCwMH5vL00wUTM0BUlJ2PP87krLOc\nPPxwIDNn+lPZX++WL7fywQf+XHKJnUceyS3wvMUCc+dm07atncWL/XjuuUqoOouIeJFqzj6yf7+J\nzZstXH653aPOY6WN58cfzYwYEcS+fWauusrG3LnZRe4nNxcGDQpi7VorN92Ux7x52VhKuCNpzx4T\nAwYEs2+fmbvvzmX69JwS1ylPPKekpJjo3DmY9HQTq1Zl0rx50de+U1JM9OoVTFKSmVmzshg4sIK6\ntWPs91tGBvzyi4Vt28x06GD36H59X8WTkeG6Z71rVzt+XpxW3cjnpywUj7Gp5lyFnXuukxtvtJWp\nV7cnrrzSwdq1J+nbN49Nm6x06VKL//634IBwNhuMGBHI2rVWunWz8Z//lJyYAS66yMnnn2fSqpWd\nd97x5957A8nOroBAzuB0wmOPBZCaaiY2NqfYxAwQEeFk0aJMwsKcjB0bWCPugXY4YOdOM++/78cj\njwTQuXMwTZqE8H//F0xcXCC9egWzdGmp57upFImJZrp3r8VddwUzZEhQpdzHL2JUSs7V2Flnweuv\nZ/PSS9nk5sKQIUE89lgAWVmu5x0OeOSRQP77Xz+uvdbG/Pml6zxVr56TTz7JpEMHG59/7seAAUEc\nPlxxF3cXL7byxReust53n2fXups0cbJgQRZWq+se6G3bPH/L5/l+vpMS2WyufgAzZvhz661BXHxx\nCJ061eLhhwNZuNCfpCQz11xjZ9SoHKZPzyYgAO67L4hXXvGr9MsRxfnoIys9ewazd6+Zxo0drFxp\nZfDgIPd7VaSmUbN2FVHeeHbtMjN8eCA7dlho3tzOvHnZLFzoxxtv+HP55XaWLMkktIwznOXkuAYC\n+fRTVztkmzZ2unSx0aVL/p7p5Ynnn39MdO5cC6cT1q07yfnnl+5t+9lnVu69N5B69Vz3QDdo4Fo/\nPR3+/NPMX3+Z+fNPM3/+afrfbzOHDpm55hobs2Zl07hx8fvzxfvt0CETw4YF8sMPp2vCF11kp107\nB+3a2WnXzk6LFg6sZ1SUd+wwM3BgEAcOmLnnnlymTSv8ckRlxZOT4xqi9t13/QkNdTJnTjbdutkY\nMiSI1autXHedjQULsqhVq3z70eeBsdXUeNRbuxrwRjzZ2TB5cgBvveWPxeLEbnfdG710aWaZbrs6\nk8MB8+f78cUXVn74wUJenqsGXauWk44d7Vx/vY0uXWzuJFeaeBwOuPXWIL791srs2VncfnvZrh2/\n9pofcXGBnH++g8hIJ3/9ZSI1tWBN2mRy0rChk7POcrJ1q4WgICfjx+cwdGhekYO8VPb77bvvLAwb\nFkhKipnevfO466482ra1c/bZJa974ICJgQOD2LHDQnR0HnPnZhcYq70y4vn7bxP33uvqhNiypZ35\n87Pc74/cXBg+PJAvvvDj6qttfPBBVpm/PII+D4yupsaj5FwNeDOe5cutjBkTSFiYq1m6bl3vvgUy\nMlzJY906K199ZWXv3tMZ7YILHHTpYqN3b38uuCCDCy5wlnib0+uv+zF+fCDR0Xm8+27Zb4tyOmHi\nxADmzfPHanVy/vmu0dwuvNBBo0anfxo2dBIQ4Frnk0+sPPFEAEePmmnf3sbs2dlceGHB41VZ7zen\nE1591Y+pUwMwmWDSpByGDcsr9TE5cQLuuSeI9euttGtn5733svLdLVDR8axebWHkyCCOHTMRE5PH\njBkFvyDk5blaZJYt86NdOzuLFmVy1lll219x8TgcsHmzmeXLrXzzjZXQUNf7onFjB40bn36PBAaW\nbd8VQZ9vxqbkXIyaerI9lZkJZjOV8oHz998mvvrKyrp1FtavPz04C0BoqJPWre1ccomD1q3ttG7t\noGlTh7spfNcuM1FRwYSEOPn660wiIsr/dj16FGrXJl9zb3GSk0089lgAy5f7ERzsZOLEHO6+O38t\nujLebydOwIMPBrJ8uR/16jl4443Tw7uWRW6uq8/B4sV+XHihg0WLMsvUslEadjs8/7w/M2cGEBDg\nZPp0173yRX25sNth9GhXGdu0sbN4cdlaef4dT24ufPutheXLraxYYeXwYdfJ9PNzult9zmQyOTn3\nXKf7C1zjxg7OOst1To4fN5GebuL4cRMnTsCJEyb3z/HjJiwWJ08/nUP//t67W0Cfb8am5FyMmnqy\njc5mg59/tpCYGMz33+exdauZPXvM+YZA9fd30qyZg9atHWzZYiYx0cLbb2fRu3fF3QpVEqcTEhKs\nPPlkIMeOmbjuOhsvvZTtvvZd0edn+3YzQ4YE8eefZjp2tDFvXrZXvqg4nfDss65kWaeOgwULsrjy\nSkeFxJOaauL++wP5+msr55/vYP78LI9u63I44NFHA1i40J+WLe189FFWqWOPiAjljz/SWbPGyvLl\nVlavtpKe7nrPhYc7uOEGO7165dGpk+vLzl9/mfnjj9P9EE79ffCgZx0KQ0Jcl0Vq13ayb5+ZjAwT\nkya5hsL1huryeXCKt+P54QcLf/5pon9/7803UBpKzsXQm9fYzozn5EnXbTTbtlnYutXM9u0Wduww\nk53t+vAcMMA1ypcRHD5sYuzYQL780kqtWk4mT87hrrvyiIz07Pzk5VHq+3fj4608/nggWVmuWcqe\neCLX41q/pxYu9OOxxwLw84PXXsvm7ruDyvR+s9lcnezS0101x4wMV20yNdXEc88FcOCAmRtusPHK\nK1keXR8/xeGA2NgA5s/3p2lTOx9/nFXi5ZjsbNd0oj/9ZOGHHwJZs8ZJbq7rPdWwoYNevWz07Gnj\nqqvsHh/PkydPJ+6TJ113RNSu7XT/nHWWk9BQ8nWyS0w0M2BAEIcOmRk1KocJE3LLPWJddf48KK+t\nW8306RNMVpaJ7t1tvPpq6d5r3qDkXAy9eY2tpHhsNti711Vruf56u6Gu9zmdrtu6nnoqkBMnTHTu\nbGPOHCuHD58kOdlESoqZ5GTT//42/e9vMykprubPunUdtGhx6sdOy5aupvx/x5id7erJvGCBP7Vr\nO3nllSyio8vejF2SNWssDB3qun1p3DgToaHZZGaayMyErCwTWVn5f59anp7uaspNTzeRmVl01jGb\nnTz5ZC4PPphbptqM0+mahWzuXH8aNXKQkHB65jGn0zVO/c8/W/jpJ9dwuVu3mvM1UbdqZadnTxu9\netlo1cpR4UO6nmnfPhP9+7tuFbv99jxefDG7XF+watrngaeSk0306BHMgQMmLr3UwZYtllK10niL\nknMx9OY1tuoQz8GDJh55JJA1a4r/lDWZnJxzjpPISCdnn+0kKcnMP//kz05ms5PGjU8n7YsvdvDq\nq/5s2WL53xjsWTRqVPH/qr/+6rrVKiXFs+wZEOAkNNRVWwwNddUeQ0Kc1K595mNX7fLyyz0bnaw4\nTidMn+7PrFkBnH++g0GD8vj5ZzM//2whOfl0ma1WJ5dc4uCKK1y3lPXsGURQkG/fb6mprl7yW7ZY\n6NHDxrx5WQU6wXmqOvz/nMkb8eTkwC23BPPjjxaeeiqHUaNy8/VvmDGj7HMBlJaSczH05jW26hLP\nqVr0Dz8EUatWLhERTiIjXbdqRUY6iYhwUqeOs0At6cQJSEy0kJhodv/s2GHJ11kO4PbbXT2Zg4Iq\nL6bDh0389lsI2dlZBAc7CQrC/TsoKP/v0gzZ6k0vvujPs88GuB/Xr386Ebdr56BNG3u+Y2aU91tG\nBtx9dxDffGPl6qttvPde2ZpcjRKPt5Q3HqcTxowJ5MMP/ejXz3V74KmWkVWrLDzwgOvOgIEDXUMN\nV/T/k5JzMfTmNTbFU5DT6boH2ZWsLTRp4rou6gtV4fx8842FY8dMXHGF3T2oTFGMFE9uLowa5bpF\nrEULO/HxWdSrV/oObkaJxxvKG8+p2y0vvdTOJ59kFmiRSEoyMXRoEL/9ZqF1a1dLVGG3RHqLN5Kz\nMQfZFamBTCbXmOvnnmsnKqriri1XF6d6Vlc1/v6uWdPq1HHy1lv+9O4dzOLFmTRpYoh6UpXz1VcW\nJk4MICLCwbvvFn6p4IILnPz3v5nExrp6/XfvXotXX83ihhuM+x7S2NoiIpXMbIZp03J44okc9u1z\n9S7eskUfx6W1d6+J4cODsFrhnXeyim1BCQyEmTNzmDUri5wcuPPOYKZP98du0Pysd4OIiA+YTDB2\nbC7PP5/N0aMm+vUL5r//tRpqQpLKUpaYT5yAu+4K4vhxEy+8kM2VV3rW2XDgQBuff57JBRc4eOml\nAPr3D2LpUivbtpkrfGa90lCztoiIDw0enEd4uJP77w9kyJAgLrvMztixOdxwg71cQ9Vu2mThiy+s\nXH216xayyrx1rDB2u2u0wL17zeze7Rp86NRPVhbcdFPA/8aIL/k2N7vdNbvanj0W7r8/l5iY0vXN\nuOQSB6tWnWTUqCC+/NLK+vWuVGg2u4b1bdbMwcUX22na1OH+CQkpa+Rlow5hVYTiMTbFY2xVIZ7E\nRDPPP+/Pf/97ena3sWNziY4umFiLiicjAz7+2I+33/Zjx47TXek7drQxZUoOrVpV/L2+aWmuMQpO\n/eze7fr9xx9m9yAwp5yaZAbM/P23a1nLlnbuuiuPW2/NK3Is9cmTA3j1VX+6drXx/vtZZb5rwOFw\nzQOwY4eZXbtO/xw9WrBRuUEDByNH5jJ8eMm3Y6m3djGqwj9jaSgeY1M8xlaV4klMNDNzpj+ffmrF\n6TTRurUrSffseXooyn/H8/vvZt55x4/4eD8yMkxYrU569bJx8802PvzQj1WrrJjNTu68M49x43Lz\nTXJSFtnZrqlWz0zCrgRs4siRgoktJMTJxRc7aNLEdQ//RRe5/m7c2DX4Tp06oXz8cSbvvefH8uVW\nbDYTQUFObr7Zxl135dKu3ena9OLFVkaNCqJJEwcrVpws82QoxUlNNbF7d/6EvWePmb59bUyalFPi\n+krOxahK/4yeUDzGpniMrSrGs3OnmZde8mfZMleSbtnSlaR797ZRt24oBw6ks2KFlfnz/diwwdUs\nW6+ea2CWO+/My3d71tq1rh7Nu3ZZqF3bydixrilQC5trvTDZ2bB+vYUVK6x8/bWVfftM+cbDB7BY\nnFxwgdOddE8l4Isvdt33X1xT9Znn5/BhE/Hxfrz3nh9JSa5E36KFnUGD8mjc2MGgQUEEBMCKFSe5\n6CJDpK8ClJyLURX/GYujeIxN8RhbVY5n1y5XTXrZMisOh2sO9l69LCxc6HDPpnXddTbuuSePHj1s\nRY7dnpcH77zjx/PPB3DsmIkmTRxMnpxN9+6FX9s+csTEqlWuhLxundU9NOvZZztp2dJOkyaOfD8X\nXOAs9bjxpxR2fhwO1xeC995zzRNvs7n2bzY7+eCDLLp2NWg3a5Sci1WV/xkLo3iMTfEYW3WIZ88e\nEy+9FMDHH7uSdGiok5iYPO6+O4+LL/b8WvLRo/DccwG8+64fdruJLl1sPP10Ds2aOfjjDxMrVrim\n0dy0yYLD4UqIjRs7iI62ER1t48or7V4fGa6k85Oc7KpNf/KJlbvuymPw4MoZhrOslJyLUR3+Gc+k\neIxN8RhbdYrnjz9MHDwYwmWXpVOrVtm3s3OnmQkTAvj6a6u7SfqPP1w1cZPJyRVXOOjRwzVz10UX\nVexEIdXp/IBGCBMRqXEaN3Zy9dWQklK+7TRv7mDx4iy+/NJCXFwgBw6YiI7Oo0cPO92724iMNES9\nrcZSchYRqaFMJujRw84NN5zEbsfr84RL2WmEMBGRGs5kUmI2GiVnERERg1FyFhERMRglZxEREYNR\nchYRETEYJWcRERGD8ah/3rRp0/j1118xmUzExsbSpk0b93Pff/89M2fOxGw206hRI6ZOncqPP/7I\n6NGjufjiiwFo2rQpEyZMqJgIREREqpkSk/OmTZtISkoiPj6evXv3EhsbS3x8vPv5iRMnsmDBAurV\nq8dDDz3E+vXrCQwM5KqrrmLOnDkVWngREZHqqMRm7Y0bNxIVFQVAkyZNOH78OBkZGe7nExISqFev\nHgDh4eGkpaVVUFFFRNbO32IAAAYdSURBVERqhhKTc2pqKmFhYe7H4eHhpJwxblxISAgAycnJbNiw\ngc6dOwOwZ88eRowYwe23386GDRu8XW4REZFqq9RjwhQ2T8aRI0cYMWIEcXFxhIWFceGFFzJq1Ch6\n9uzJvn37GDRoEF9++SX+xUweGhYWjNXq3alOihtUvCpSPMameIxN8Rib4smvxOQcGRlJamqq+3Fy\ncjIRERHuxxkZGQwbNowxY8bQsWNHAOrWrUuvXr0AOP/88znnnHM4fPgwDRs2LHI/aWmZZQ6iMDV1\nlpOqQvEYm+IxNsVjbJUyK1WHDh14+eWXiYmJYfv27URGRrqbsgFmzJjB4MGD6dSpk3vZp59+SkpK\nCkOHDiUlJYUjR45Qt27dMheyrPRNzNgUj7EpHmNTPMZW3ng8ms/5hRde4KeffsJkMhEXF8eOHTsI\nDQ2lY8eOXHnllVx++eXu1/bp04fevXvz6KOPcuLECfLy8hg1apT7WrSIiIgUz6PkLCIiIpVHI4SJ\niIgYjJKziIiIwSg5i4iIGIySs4iIiMGUehCSqqC4iTqqmh9++KFaTCKya9cuRo4cyd13382dd97J\nwYMHefzxx7Hb7URERPD8888XO0iN0fw7nnHjxrF9+3bOPvtsAIYOHcr111/v20KWwnPPPcfPP/+M\nzWbjvvvu45JLLqnS5+ff8axdu7bKnp+srCzGjRvHkSNHyMnJYeTIkTRv3rzKnp/C4lm5cmWVPT+n\nZGdn06dPH0aOHEn79u3LfX6qXXIuaaKOqqiqTyKSmZnJlClTaN++vXvZnDlzGDhwID179mTmzJks\nWbKEgQMH+rCUnissHoBHHnmELl26+KhUZff999+ze/du4uPjSUtLo1+/frRv377Knp/C4rnmmmuq\n7Pn56quvaN26NcOGDWP//v0MGTKEtm3bVtnzU1g8l19+eZU9P6e89tprnHXWWYB3Pt+qXbN2SRN1\nSOXz9/fnjTfeIDIy0r3shx9+oFu3bgB06dKFjRs3+qp4pVZYPFXZlVdeyezZswGoXbs2WVlZVfr8\nFBaP3W73canKrlevXgwbNgyAgwcPUrdu3Sp9fgqLp6rbu3cve/bscdf2vXF+ql1yLmmijqqoqk8i\nYrVaCQwMzLcsKyvL3cxTp06dKnWOCosHYOHChQwaNIiHH36Yo0eP+qBkZWOxWAgODgZgyZIldOrU\nqUqfn8LisVgsVfb8nBITE8Ojjz5KbGxslT4/p5wZD1Td/x+AZ599lnHjxrkfe+P8VLtm7X+r6mOs\nlGUSkaqmqp8jgL59+3L22WfTokULXn/9dV555RUmTpzo62KVyurVq1myZAnz58/nhhtucC+vqufn\nzHi2bdtW5c/PokWLSExM5LHHHst3Tqrq+TkzntjY2Cp7fpYtW8Zll11W5NwRZT0/1a7mXNJEHVXN\nqUlETCZTvklEqrrg4GCys7MBOHz4cJVvIm7fvj0tWrQAoGvXruzatcvHJSqd9evXM3fuXN544w1C\nQ0Or/Pn5dzxV+fxs27aNgwcPAtCiRQvsdju1atWqsuensHiaNm1aZc/PunXrWLNmDf379+ejjz7i\nP//5j1f+f6pdcu7QoQMrV64EKHSijqrm008/5a233gLweBKRquDaa691n6cvv/yS6667zsclKp8H\nH3yQffv2Aa7rTad611cF6enpPPfcc8ybN8/dW7Yqn5/C4qnK5+enn35i/vz5gOuyXWZmZpU+P4XF\nM3HixCp7fmbNmsXHH3/M4sWLue222xg5cqRXzk+1HFv73xN1NG/e3NdFKrOMjIwqP4nItm3bePbZ\nZ9m/fz9Wq5W6devywgsvMG7cOHJycmjQoAHTp0/Hz8/P10X1SGHx3Hnnnbz++usEBQURHBzM9OnT\nqVOnjq+L6pH4+HhefvllGjVq5F42Y8YMxo8fXyXPT2Hx3HLLLSxcuLBKnp/s7GyeeuopDh48SHZ2\nNqNGjaJ169Y88cQTVfL8FBZPcHAwz/9/O3ZsA0AMw0Ase3gVb5gyC3/7fRo5IFdQcYD2HrnP3zln\nVdXq7ut9nowzAEz23K0NANOJMwCEEWcACCPOABBGnAEgjDgDQBhxBoAw4gwAYT4k88x6nNM7qgAA\nAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}