experiments/synthetic/Localized_CNNs_on_synthetic_data.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Localized CNNs on synthetic data",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "qSSwdfIKTIPG",
        "colab_type": "code",
        "outputId": "9b7b3028-9ade-493b-b79c-e713eb1b74b0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/gdrive')\n",
        "%cd /gdrive"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /gdrive\n",
            "/gdrive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7rMB8CGvgvtk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import sys\n",
        "\n",
        "sys.path.append(\"/gdrive/My Drive\")\n",
        "sys.path.append('..')\n",
        "sys.path.append('../..')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TJ7LD0RLU-sE",
        "colab_type": "text"
      },
      "source": [
        "# Configuration variables: task to run, paths to datasets, ..."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8GZfiYEJUBcF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "ball_train_data_path = \"/gdrive/My Drive/bouncing_balls_training_data.mat\"\n",
        "ball_test_data_path = \"/gdrive/My Drive/bouncing_balls_testing_data.mat\"\n",
        "\n",
        "save_results_path = \"/gdrive/My Drive\""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eAW3Mw4PUCOf",
        "colab_type": "text"
      },
      "source": [
        "# Utility and processing functions\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iBFs8_LXS_JS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# imports\n",
        "import glob\n",
        "import pandas as pd\n",
        "from shapely.geometry import Point\n",
        "from shapely.geometry.polygon import Polygon\n",
        "import numpy as np\n",
        "\n",
        "import copy\n",
        "import os\n",
        "import pickle\n",
        "import time\n",
        "from random import shuffle\n",
        "import copy\n",
        "import pickle\n",
        "import random\n",
        "import numpy as np\n",
        "\n",
        "\n",
        "\n",
        "def get_bouncing_split_sets(future=5):\n",
        "    data_train = pickle.load(open(ball_train_data_path, \"rb\"))[\"Data\"]\n",
        "    data_test = pickle.load(open(ball_test_data_path, \"rb\"))[\"Data\"]\n",
        "\n",
        "\n",
        "    horizon = 30 - future\n",
        "\n",
        "    train_len = int(0.75 * len(data_train))\n",
        "\n",
        "    def shift_array(data):\n",
        "      return np.swapaxes(np.swapaxes(data, 1, 2), 3, 2)\n",
        "\n",
        "    def form_data(in_data):\n",
        "        data_out_x = copy.deepcopy(in_data)\n",
        "        data_out_y = []\n",
        "\n",
        "        for i in range(len(in_data)):\n",
        "            data_out_x[i] = in_data[i].reshape((30, 30, 30))\n",
        "\n",
        "            data_out_y.append(data_out_x[i][-1, :, :])\n",
        "            data_out_x[i] = data_out_x[i][:-future, :, :]\n",
        "        return data_out_x, data_out_y\n",
        "\n",
        "    pre_train_x, pre_train_y = form_data(data_train)\n",
        "    test_x, test_y = form_data(data_test)\n",
        "\n",
        "    train_x, train_y = pre_train_x[:train_len], pre_train_y[:train_len]\n",
        "    valid_x, valid_y = pre_train_x[train_len:], pre_train_y[train_len:]\n",
        "\n",
        "    train_x, train_y = np.stack(train_x, axis=0), np.expand_dims(np.stack(train_y, axis=0), 1)\n",
        "    valid_x, valid_y = np.stack(valid_x, axis=0), np.expand_dims(np.stack(valid_y, axis=0), 1)\n",
        "    test_x, test_y = np.stack(test_x, axis=0), np.expand_dims(np.stack(test_y, axis=0), 1)\n",
        "\n",
        "    return shift_array(train_x), shift_array(train_y), shift_array(valid_x), shift_array(valid_y), shift_array(test_x), shift_array(test_y)\n",
        "\n",
        "def get_errors(predicted, test_y, horizon, min_val, max_val):\n",
        "    total_nodes = predicted.shape[1] * predicted.shape[2]\n",
        "\n",
        "    errs = np.zeros((predicted.shape[1], predicted.shape[2], horizon))\n",
        "    mape = np.zeros((predicted.shape[1], predicted.shape[2], horizon))\n",
        "    mae = np.zeros((predicted.shape[1], predicted.shape[2], horizon))\n",
        "\n",
        "    for i in range(test_y.shape[0]):\n",
        "        denorm_truth = test_y[i, :, :, :] * (max_val - min_val) + min_val\n",
        "        denorm_pred = predicted[i, :, :, :] * (max_val - min_val) + min_val\n",
        "\n",
        "        mape += np.abs((denorm_truth - denorm_pred) / denorm_truth)\n",
        "        errs += (denorm_truth - denorm_pred) ** 2\n",
        "        mae += np.abs(denorm_pred - denorm_truth)\n",
        "    errs = np.sum(errs, axis=2) / test_y.shape[0]\n",
        "\n",
        "    rmse = np.sqrt(np.sum(errs) / total_nodes)\n",
        "    mape = np.sum(np.sum(mape, axis=2) / test_y.shape[0]) / total_nodes * 100\n",
        "    mae = np.sum((np.sum(mae, axis=2) / test_y.shape[0])) / total_nodes\n",
        "\n",
        "    return rmse, mape, mae"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NzikFTzmVqgM",
        "colab_type": "text"
      },
      "source": [
        "# Define models classes"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jIuc1NhAVp3N",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "3e6495fa-6f4a-41d7-cdcb-2c721895e7de"
      },
      "source": [
        "import keras\n",
        "from keras import Input, Model, metrics\n",
        "from keras.layers import Conv2D\n",
        "\n",
        "from keras import Input, Model, metrics\n",
        "from keras.layers import Conv2D, Dense, Activation, Reshape, Flatten, LocallyConnected2D\n",
        "from keras.layers import MaxPooling2D, LSTM, Concatenate\n",
        "\n",
        "class CNN():\n",
        "    def __init__(self, train_x, train_y, valid_x, valid_y, test_x, test_y, layers=None):\n",
        "        self.input_dim, self.look_back, self.horizon =  (train_x[0].shape[1], train_x[0].shape[2]), train_x[0].shape[-1], 1\n",
        "        self.train_x, self.train_y, self.valid_x, self.valid_y, self.test_x, self.test_y = train_x, train_y, valid_x, valid_y, test_x, test_y\n",
        "        self.layers = layers\n",
        "\n",
        "    def get_model(self):\n",
        "        I1 = Input(shape=(self.input_dim[0], self.input_dim[1], self.look_back))\n",
        "\n",
        "        if self.layers is None:\n",
        "            Conved = Conv2D(30, kernel_size=5, activation='relu', padding=\"same\")(I1)\n",
        "            Conved = Conv2D(30, kernel_size=4, activation='relu', padding=\"same\")(Conved)\n",
        "            Conved = Conv2D(30, kernel_size=3, activation='relu', padding=\"same\")(Conved)\n",
        "        else:\n",
        "            for idx, layer in enumerate(self.layers):\n",
        "                if idx == 0:\n",
        "                    Conved = Conv2D(layer[1], kernel_size=layer[0], activation='relu', padding=\"same\")(I1)\n",
        "                else:\n",
        "                    Conved = Conv2D(layer[1], kernel_size=layer[0], activation='relu', padding=\"same\")(Conved)\n",
        "\n",
        "        Sing = Conv2D(self.horizon, kernel_size=1, padding=\"same\")(Conved)\n",
        "        model = Model(inputs=[I1], outputs=[Sing])\n",
        "        model.compile(optimizer='adam', loss='mean_squared_error', metrics=[metrics.mse])\n",
        "\n",
        "        self.model = model\n",
        "        return self.model\n",
        "\n",
        "    def train(self, epochs=1):\n",
        "        hist = self.model.fit(self.train_x, self.train_y, validation_data=(self.valid_x, self.valid_y), epochs=epochs,\n",
        "                              batch_size=200, verbose=2)\n",
        "        return hist\n",
        "\n",
        "    def predict(self, data):\n",
        "        predicted = self.model.predict(data)\n",
        "        return predicted\n",
        "\n",
        "############ =================================================================================== ######################\n",
        "class LICNN():\n",
        "    def __init__(self, train_x, train_y, valid_x, valid_y, test_x, test_y, learnable_inputs_count, layers=None):\n",
        "        self.input_dim, self.look_back, self.horizon = (train_x[0].shape[1], train_x[0].shape[2]), train_x[0].shape[-1], 1\n",
        "        self.train_x, self.train_y, self.valid_x, self.valid_y, self.test_x, self.test_y = train_x, train_y, valid_x, valid_y, test_x, test_y\n",
        "        self.learnable_inputs_count = learnable_inputs_count\n",
        "        self.layers = layers\n",
        "\n",
        "    def get_model(self):\n",
        "        I1 = Input(shape=(self.input_dim[0], self.input_dim[1], self.look_back))\n",
        "\n",
        "        # To implement learnable inputs, we use auxillary input...\n",
        "        I2 = Input(shape=(self.input_dim[0],self.input_dim[1], 1))\n",
        "        # ...that receives a matrix of 1s and implements locally connected layer without biases with a 1x1 kernel.\n",
        "        C1 = LocallyConnected2D(self.learnable_inputs_count, kernel_size=(1, 1), input_shape=(self.input_dim[0], self.input_dim[1], 1),\n",
        "                                bias_initializer='zeros', use_bias=False)(I2)\n",
        "        \n",
        "        # Finally, we concatinate real input with the learnable input. \n",
        "        Merged1 = Concatenate(-1)([I1, C1])\n",
        "\n",
        "        if self.layers is None:\n",
        "            Conved = Conv2D(30, kernel_size=5, activation='relu', padding=\"same\")(Merged1)\n",
        "            Conved = Conv2D(30, kernel_size=4, activation='relu', padding=\"same\")(Conved)\n",
        "            Conved = Conv2D(30, kernel_size=3, activation='relu', padding=\"same\")(Conved)\n",
        "        else:\n",
        "            for idx, layer in enumerate(self.layers):\n",
        "                if idx == 0:\n",
        "                    Conved = Conv2D(layer[1], kernel_size=layer[0], activation='relu', padding=\"same\")(Merged1)\n",
        "                else:\n",
        "                    Conved = Conv2D(layer[1], kernel_size=layer[0], activation='relu', padding=\"same\")(Conved)\n",
        "\n",
        "        Sing = Conv2D(self.horizon, kernel_size=1, padding=\"same\")(Conved)\n",
        "\n",
        "        model = Model(inputs=[I1, I2], outputs=Sing)\n",
        "        model.compile(optimizer='adam', loss='mean_squared_error', metrics=[metrics.mse])\n",
        "        self.model = model\n",
        "        return self.model\n",
        "\n",
        "    def train(self, epochs = 1):\n",
        "        hist = self.model.fit(self.train_x, self.train_y, validation_data=(self.valid_x, self.valid_y), epochs=epochs, batch_size=200, verbose=2)\n",
        "        return hist\n",
        "    def predict(self, data):\n",
        "        predicted = self.model.predict(data)\n",
        "        return predicted\n",
        "# ================================================================================================================"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qN4gsXAIV0yM",
        "colab_type": "text"
      },
      "source": [
        "# Experiment driver"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ulDkMXHLf3su",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def run_experiment():\n",
        "  train_x, train_y, valid_x, valid_y, test_x, test_y = get_bouncing_split_sets(future = 5)\n",
        "  print(train_x.shape)\n",
        "  # spatial dimension\n",
        "  input_dim = (train_x.shape[1], train_x.shape[2])\n",
        "  look_back = train_x.shape[3]\n",
        "\n",
        "  train_ones = np.ones((train_x.shape[0], train_x.shape[1], train_x.shape[2], 1))\n",
        "  valid_ones = np.ones((valid_x.shape[0], valid_x.shape[1], valid_x.shape[2], 1))\n",
        "  test_ones = np.ones((test_x.shape[0], test_x.shape[1], test_x.shape[2], 1))\n",
        "\n",
        "  rand_noise_in = np.expand_dims(np.random.random((train_x.shape[1], train_x.shape[2], 2))*2 - 1, 0)\n",
        "\n",
        "  train_rand = np.repeat(rand_noise_in, train_x.shape[0],axis=0)\n",
        "  valid_rand = np.repeat(rand_noise_in, valid_x.shape[0],axis=0)\n",
        "  test_rand = np.repeat(rand_noise_in, test_x.shape[0],axis=0)\n",
        "\n",
        "  coords= np.zeros((1,30,30, 2))\n",
        "  for i in range(30):\n",
        "      for j in range(30):\n",
        "          coords[0, i, j, 1] = j / train_x.shape[2]\n",
        "          coords[0, i, j, 0] = i / train_x.shape[1]\n",
        "\n",
        "  train_coords = np.repeat(coords, train_x.shape[0],axis=0)\n",
        "  valid_coords = np.repeat(coords, valid_x.shape[0],axis=0)\n",
        "  test_coords =  np.repeat(coords, test_x.shape[0],axis=0)\n",
        "\n",
        " \n",
        "  models = [\n",
        "      {\n",
        "          \"name\": \"CNN\",\n",
        "          \"skip\": True,\n",
        "          \"model\": CNN([train_x], train_y, [valid_x], valid_y, [test_x], test_y, layers=[(20, 22), (10, 10)])\n",
        "      },\n",
        "      {\n",
        "        \"name\": \"LI CNN\",\n",
        "        \"skip\": True,\n",
        "        # 2 learnable inputs are used\n",
        "        \"model\": LICNN([train_x, train_ones], train_y, [valid_x, valid_ones], valid_y, [test_x, test_ones], test_y, 2,layers=[(20, 20), (10,10)])\n",
        "      },\n",
        "      {\n",
        "        \"name\": \"CoordConv\",\n",
        "        \"skip\": False,\n",
        "        # coordinates are concatinated with the input\n",
        "        \"model\": CNN([np.concatenate( (train_x, train_coords), axis=-1)], train_y, [np.concatenate( (valid_x, valid_coords), axis=-1)], valid_y,\n",
        "                                     [np.concatenate( (test_x, test_coords), axis=-1)],  test_y, layers=[(20, 21), (10, 10)])\n",
        "      },\n",
        "      { \n",
        "        \"name\": \"RandomConv\",\n",
        "        \"skip\": True,\n",
        "        # random noise is concatinated with the input\n",
        "        \"model\": CNN([np.concatenate( (train_x, train_rand), axis=-1)], train_y, [np.concatenate( (valid_x, valid_rand), axis=-1)], valid_y,\n",
        "                                     [np.concatenate( (test_x, test_rand), axis=-1)],  test_y, layers=[(20, 21), (10, 10)])\n",
        "      },\n",
        "  ]\n",
        "\n",
        "  epochs_count = 100\n",
        "  extra_text = \"\"\n",
        "\n",
        "  # for clearier code the random matrix passed to RandomConv is fixed here, but was regenerated in each experiment  \n",
        "  repeat_exp_count = 5\n",
        "\n",
        "  for exp_iter in range(repeat_exp_count):\n",
        "      print(\"Starting {} / {} experiments batch\".format(exp_iter, repeat_exp_count))\n",
        "      for model in models:\n",
        "          if \"skip\" in model and model[\"skip\"] is True:\n",
        "              print(\"Skipping {} \".format(model[\"name\"]))\n",
        "              continue\n",
        "          # initialize\n",
        "          model[\"model\"].get_model()\n",
        "          model[\"model\"].model.summary()\n",
        "          print(\"Training {} model, with {} learnable parameters\".format(model[\"name\"], model[\"model\"].model.count_params()))\n",
        "\n",
        "          train_error, valid_error, test_error = [], [], []\n",
        "\n",
        "          for epoch in range(epochs_count):\n",
        "              hist = model[\"model\"].train()\n",
        "\n",
        "              predicted_valid = model[\"model\"].predict(model[\"model\"].valid_x) # this corresponds to the first testing set in the paper\n",
        "              predicted_test = model[\"model\"].predict(model[\"model\"].test_x)         \n",
        "\n",
        "              valid_rmse, valid_mape, valid_mae = get_errors(predicted_valid, model[\"model\"].valid_y, 1, 0, 1)\n",
        "              test_rmse, test_mape, test_mae = get_errors(predicted_test, model[\"model\"].test_y, 1, 0, 1)\n",
        "\n",
        "              print(\"Model - {} , epoch {} / {}.\".format(model[\"name\"], epoch + 1, epochs_count))\n",
        "              print(\"Test RMSE - {}, MAPE - {}, MAE - {}\".format(test_rmse, test_mape, test_mae))\n",
        "              \n",
        "              test_error.append((test_rmse, test_mape, test_mae))\n",
        "              valid_error.append((valid_rmse, valid_mape, valid_mae))\n",
        "              train_error.append((hist.history[\"loss\"]))\n",
        "\n",
        "          folder = \"{}/results/synthetic/bouncing_ball/\".format(save_results_path)\n",
        "          if not os.path.isdir(folder):\n",
        "              os.makedirs(folder)\n",
        "\n",
        "          file_name = \"{}/{}_{}_{}_{}\".format(folder,\n",
        "                                                    model[\"name\"],\n",
        "                                                    model[\"model\"].model.count_params(),\n",
        "                                                    epochs_count,\n",
        "                                                    str(time.ctime()).replace(\" \", \"_\").replace(\":\", \"-\"))\n",
        "          pickle.dump({\"name\": model[\"name\"],\n",
        "                      \"learnable_parameters\": model[\"model\"].model.count_params(),\n",
        "                      \"horizon\": 1,\n",
        "                      \"predict_horizon\": 1,\n",
        "                      \"look_back\": look_back,\n",
        "                      \"summary\": model[\"model\"].model.to_json(),\n",
        "                      \"model_descriptor\": model[\"model\"].model.to_json(),\n",
        "                      \"test_error\": test_error,\n",
        "                      \"time\": time.ctime(),\n",
        "                      \"valid_error\": valid_error,\n",
        "                      \"train_error\": train_error},\n",
        "                      open(file_name, \"wb\"))\n",
        "      # write the results to the file"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PrsAxh_aUZ2d",
        "colab_type": "code",
        "outputId": "0a3946df-c7e3-4ac6-925d-7904efbfce47",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "run_experiment()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1500, 30, 30, 25)\n",
            "Starting 0 / 5 experiments batch\n",
            "Skipping CNN \n",
            "Skipping LI CNN \n",
            "Model: \"model_10\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_10 (InputLayer)        (None, 30, 30, 27)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_28 (Conv2D)           (None, 30, 30, 21)        226821    \n",
            "_________________________________________________________________\n",
            "conv2d_29 (Conv2D)           (None, 30, 30, 10)        21010     \n",
            "_________________________________________________________________\n",
            "conv2d_30 (Conv2D)           (None, 30, 30, 1)         11        \n",
            "=================================================================\n",
            "Total params: 247,842\n",
            "Trainable params: 247,842\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Training CoordConv model, with 247842 learnable parameters\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0434 - mean_squared_error: 0.0434 - val_loss: 0.0334 - val_mean_squared_error: 0.0334\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:66: RuntimeWarning: divide by zero encountered in true_divide\n",
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:66: RuntimeWarning: overflow encountered in true_divide\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Model - CoordConv , epoch 1 / 100.\n",
            "Test RMSE - 0.18355873470111575, MAPE - inf, MAE - 0.06472045912436208\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0334 - mean_squared_error: 0.0334 - val_loss: 0.0329 - val_mean_squared_error: 0.0329\n",
            "Model - CoordConv , epoch 2 / 100.\n",
            "Test RMSE - 0.18240068853224375, MAPE - inf, MAE - 0.07339491872682209\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0326 - mean_squared_error: 0.0326 - val_loss: 0.0319 - val_mean_squared_error: 0.0319\n",
            "Model - CoordConv , epoch 3 / 100.\n",
            "Test RMSE - 0.17902388374427966, MAPE - inf, MAE - 0.07035664462629763\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0315 - mean_squared_error: 0.0315 - val_loss: 0.0309 - val_mean_squared_error: 0.0309\n",
            "Model - CoordConv , epoch 4 / 100.\n",
            "Test RMSE - 0.17627206018644573, MAPE - inf, MAE - 0.08296699113179848\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0304 - mean_squared_error: 0.0304 - val_loss: 0.0298 - val_mean_squared_error: 0.0298\n",
            "Model - CoordConv , epoch 5 / 100.\n",
            "Test RMSE - 0.17181712313684494, MAPE - inf, MAE - 0.07557023400869045\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0293 - mean_squared_error: 0.0293 - val_loss: 0.0288 - val_mean_squared_error: 0.0288\n",
            "Model - CoordConv , epoch 6 / 100.\n",
            "Test RMSE - 0.16834918607141536, MAPE - inf, MAE - 0.07597275572244928\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0283 - mean_squared_error: 0.0283 - val_loss: 0.0278 - val_mean_squared_error: 0.0278\n",
            "Model - CoordConv , epoch 7 / 100.\n",
            "Test RMSE - 0.1654627259346488, MAPE - inf, MAE - 0.07108760704791849\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0273 - mean_squared_error: 0.0273 - val_loss: 0.0269 - val_mean_squared_error: 0.0269\n",
            "Model - CoordConv , epoch 8 / 100.\n",
            "Test RMSE - 0.16336849357940225, MAPE - inf, MAE - 0.06505569866434535\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0264 - mean_squared_error: 0.0264 - val_loss: 0.0260 - val_mean_squared_error: 0.0260\n",
            "Model - CoordConv , epoch 9 / 100.\n",
            "Test RMSE - 0.16219311804335773, MAPE - inf, MAE - 0.0635955908386831\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0255 - mean_squared_error: 0.0255 - val_loss: 0.0252 - val_mean_squared_error: 0.0252\n",
            "Model - CoordConv , epoch 10 / 100.\n",
            "Test RMSE - 0.1602347158468875, MAPE - inf, MAE - 0.059205594734987446\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0247 - mean_squared_error: 0.0247 - val_loss: 0.0244 - val_mean_squared_error: 0.0244\n",
            "Model - CoordConv , epoch 11 / 100.\n",
            "Test RMSE - 0.15843611055203696, MAPE - inf, MAE - 0.05848615056059415\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0239 - mean_squared_error: 0.0239 - val_loss: 0.0237 - val_mean_squared_error: 0.0237\n",
            "Model - CoordConv , epoch 12 / 100.\n",
            "Test RMSE - 0.15688183312624088, MAPE - inf, MAE - 0.05685620304355552\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0231 - mean_squared_error: 0.0231 - val_loss: 0.0230 - val_mean_squared_error: 0.0230\n",
            "Model - CoordConv , epoch 13 / 100.\n",
            "Test RMSE - 0.1547677036285703, MAPE - inf, MAE - 0.054994189414594816\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0224 - mean_squared_error: 0.0224 - val_loss: 0.0223 - val_mean_squared_error: 0.0223\n",
            "Model - CoordConv , epoch 14 / 100.\n",
            "Test RMSE - 0.15210431267058125, MAPE - inf, MAE - 0.05369446354514496\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0216 - mean_squared_error: 0.0216 - val_loss: 0.0217 - val_mean_squared_error: 0.0217\n",
            "Model - CoordConv , epoch 15 / 100.\n",
            "Test RMSE - 0.15084803904460203, MAPE - inf, MAE - 0.05180440272609054\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0209 - mean_squared_error: 0.0209 - val_loss: 0.0211 - val_mean_squared_error: 0.0211\n",
            "Model - CoordConv , epoch 16 / 100.\n",
            "Test RMSE - 0.1483907907352693, MAPE - inf, MAE - 0.05181414255161142\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0201 - mean_squared_error: 0.0201 - val_loss: 0.0205 - val_mean_squared_error: 0.0205\n",
            "Model - CoordConv , epoch 17 / 100.\n",
            "Test RMSE - 0.14629343161959368, MAPE - inf, MAE - 0.05104322461876608\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0195 - mean_squared_error: 0.0195 - val_loss: 0.0199 - val_mean_squared_error: 0.0199\n",
            "Model - CoordConv , epoch 18 / 100.\n",
            "Test RMSE - 0.1447540984387481, MAPE - inf, MAE - 0.048787545997694304\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0189 - mean_squared_error: 0.0189 - val_loss: 0.0195 - val_mean_squared_error: 0.0195\n",
            "Model - CoordConv , epoch 19 / 100.\n",
            "Test RMSE - 0.1430983448437156, MAPE - inf, MAE - 0.04763363759968781\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0184 - mean_squared_error: 0.0184 - val_loss: 0.0191 - val_mean_squared_error: 0.0191\n",
            "Model - CoordConv , epoch 20 / 100.\n",
            "Test RMSE - 0.14153275177918104, MAPE - inf, MAE - 0.04664664361996821\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0181 - mean_squared_error: 0.0181 - val_loss: 0.0184 - val_mean_squared_error: 0.0184\n",
            "Model - CoordConv , epoch 21 / 100.\n",
            "Test RMSE - 0.1402909769835961, MAPE - inf, MAE - 0.04519933082872257\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0174 - mean_squared_error: 0.0174 - val_loss: 0.0182 - val_mean_squared_error: 0.0182\n",
            "Model - CoordConv , epoch 22 / 100.\n",
            "Test RMSE - 0.14014458917790656, MAPE - inf, MAE - 0.0441006388634549\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0170 - mean_squared_error: 0.0170 - val_loss: 0.0179 - val_mean_squared_error: 0.0179\n",
            "Model - CoordConv , epoch 23 / 100.\n",
            "Test RMSE - 0.13717037198329385, MAPE - inf, MAE - 0.04520402570282232\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0167 - mean_squared_error: 0.0167 - val_loss: 0.0175 - val_mean_squared_error: 0.0175\n",
            "Model - CoordConv , epoch 24 / 100.\n",
            "Test RMSE - 0.13823155929972203, MAPE - inf, MAE - 0.04326656456463007\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0161 - mean_squared_error: 0.0161 - val_loss: 0.0172 - val_mean_squared_error: 0.0172\n",
            "Model - CoordConv , epoch 25 / 100.\n",
            "Test RMSE - 0.13665524035067494, MAPE - inf, MAE - 0.04320376933979869\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0157 - mean_squared_error: 0.0157 - val_loss: 0.0169 - val_mean_squared_error: 0.0169\n",
            "Model - CoordConv , epoch 26 / 100.\n",
            "Test RMSE - 0.13391501434608458, MAPE - inf, MAE - 0.04377549558728464\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0153 - mean_squared_error: 0.0153 - val_loss: 0.0167 - val_mean_squared_error: 0.0167\n",
            "Model - CoordConv , epoch 27 / 100.\n",
            "Test RMSE - 0.1337224150299722, MAPE - inf, MAE - 0.0438571053787628\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0154 - mean_squared_error: 0.0154 - val_loss: 0.0164 - val_mean_squared_error: 0.0164\n",
            "Model - CoordConv , epoch 28 / 100.\n",
            "Test RMSE - 0.13287550231614328, MAPE - inf, MAE - 0.041914370878809976\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0152 - mean_squared_error: 0.0152 - val_loss: 0.0168 - val_mean_squared_error: 0.0168\n",
            "Model - CoordConv , epoch 29 / 100.\n",
            "Test RMSE - 0.13657174889651974, MAPE - inf, MAE - 0.039841015738611874\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0153 - mean_squared_error: 0.0153 - val_loss: 0.0163 - val_mean_squared_error: 0.0163\n",
            "Model - CoordConv , epoch 30 / 100.\n",
            "Test RMSE - 0.1310951669217724, MAPE - inf, MAE - 0.043137327936948945\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0147 - mean_squared_error: 0.0147 - val_loss: 0.0162 - val_mean_squared_error: 0.0162\n",
            "Model - CoordConv , epoch 31 / 100.\n",
            "Test RMSE - 0.13420672678998177, MAPE - inf, MAE - 0.03943468039624793\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0143 - mean_squared_error: 0.0143 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 32 / 100.\n",
            "Test RMSE - 0.13047403029001087, MAPE - inf, MAE - 0.03999711998883964\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 33 / 100.\n",
            "Test RMSE - 0.1286494881457685, MAPE - inf, MAE - 0.039957853662606437\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0146 - mean_squared_error: 0.0146 - val_loss: 0.0199 - val_mean_squared_error: 0.0199\n",
            "Model - CoordConv , epoch 34 / 100.\n",
            "Test RMSE - 0.14848050417423375, MAPE - inf, MAE - 0.038690251513083924\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0187 - mean_squared_error: 0.0187 - val_loss: 0.0177 - val_mean_squared_error: 0.0177\n",
            "Model - CoordConv , epoch 35 / 100.\n",
            "Test RMSE - 0.13519831407570188, MAPE - inf, MAE - 0.045706323916028295\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0164 - mean_squared_error: 0.0164 - val_loss: 0.0167 - val_mean_squared_error: 0.0167\n",
            "Model - CoordConv , epoch 36 / 100.\n",
            "Test RMSE - 0.13795415744562803, MAPE - inf, MAE - 0.043414958427303287\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0149 - mean_squared_error: 0.0149 - val_loss: 0.0158 - val_mean_squared_error: 0.0158\n",
            "Model - CoordConv , epoch 37 / 100.\n",
            "Test RMSE - 0.13201472843576095, MAPE - inf, MAE - 0.04098044794899242\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0153 - val_mean_squared_error: 0.0153\n",
            "Model - CoordConv , epoch 38 / 100.\n",
            "Test RMSE - 0.1285272702950853, MAPE - inf, MAE - 0.04151676324859419\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
            "Model - CoordConv , epoch 39 / 100.\n",
            "Test RMSE - 0.12782274490917644, MAPE - inf, MAE - 0.04005224120448988\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0133 - mean_squared_error: 0.0133 - val_loss: 0.0147 - val_mean_squared_error: 0.0147\n",
            "Model - CoordConv , epoch 40 / 100.\n",
            "Test RMSE - 0.12749220939018682, MAPE - inf, MAE - 0.03897237223786018\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0129 - mean_squared_error: 0.0129 - val_loss: 0.0146 - val_mean_squared_error: 0.0146\n",
            "Model - CoordConv , epoch 41 / 100.\n",
            "Test RMSE - 0.12767155054988802, MAPE - inf, MAE - 0.038614305938542565\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0128 - mean_squared_error: 0.0128 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n",
            "Model - CoordConv , epoch 42 / 100.\n",
            "Test RMSE - 0.1260253107109309, MAPE - inf, MAE - 0.03844279301714123\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
            "Model - CoordConv , epoch 43 / 100.\n",
            "Test RMSE - 0.12708095801760935, MAPE - inf, MAE - 0.037448374009836596\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
            "Model - CoordConv , epoch 44 / 100.\n",
            "Test RMSE - 0.12546839863339257, MAPE - inf, MAE - 0.03740992120941386\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0123 - mean_squared_error: 0.0123 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
            "Model - CoordConv , epoch 45 / 100.\n",
            "Test RMSE - 0.12580836331294382, MAPE - inf, MAE - 0.037795110279373226\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0123 - mean_squared_error: 0.0123 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
            "Model - CoordConv , epoch 46 / 100.\n",
            "Test RMSE - 0.12545768786762207, MAPE - inf, MAE - 0.03654726047332869\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0121 - mean_squared_error: 0.0121 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
            "Model - CoordConv , epoch 47 / 100.\n",
            "Test RMSE - 0.12504453665116028, MAPE - inf, MAE - 0.03652652796638012\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0119 - mean_squared_error: 0.0119 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
            "Model - CoordConv , epoch 48 / 100.\n",
            "Test RMSE - 0.12513246119712082, MAPE - inf, MAE - 0.03605092476050239\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0118 - mean_squared_error: 0.0118 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
            "Model - CoordConv , epoch 49 / 100.\n",
            "Test RMSE - 0.12408466553780426, MAPE - inf, MAE - 0.03640508724001609\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0122 - mean_squared_error: 0.0122 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
            "Model - CoordConv , epoch 50 / 100.\n",
            "Test RMSE - 0.12790381643435442, MAPE - inf, MAE - 0.035422560224420814\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0131 - mean_squared_error: 0.0131 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
            "Model - CoordConv , epoch 51 / 100.\n",
            "Test RMSE - 0.12595257696332865, MAPE - inf, MAE - 0.036317671460182895\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0121 - mean_squared_error: 0.0121 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
            "Model - CoordConv , epoch 52 / 100.\n",
            "Test RMSE - 0.1296639466659074, MAPE - inf, MAE - 0.03524870068363032\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
            "Model - CoordConv , epoch 53 / 100.\n",
            "Test RMSE - 0.1240472065860151, MAPE - inf, MAE - 0.03604288875452655\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0116 - mean_squared_error: 0.0116 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
            "Model - CoordConv , epoch 54 / 100.\n",
            "Test RMSE - 0.12339482095271821, MAPE - inf, MAE - 0.037042445801819986\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0115 - mean_squared_error: 0.0115 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
            "Model - CoordConv , epoch 55 / 100.\n",
            "Test RMSE - 0.1224763035742497, MAPE - inf, MAE - 0.03692787131511482\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0117 - mean_squared_error: 0.0117 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
            "Model - CoordConv , epoch 56 / 100.\n",
            "Test RMSE - 0.12475665837405686, MAPE - inf, MAE - 0.03543143212981074\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0113 - mean_squared_error: 0.0113 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
            "Model - CoordConv , epoch 57 / 100.\n",
            "Test RMSE - 0.12253544208543711, MAPE - inf, MAE - 0.03599231808165198\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0112 - mean_squared_error: 0.0112 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
            "Model - CoordConv , epoch 58 / 100.\n",
            "Test RMSE - 0.12250408740024783, MAPE - inf, MAE - 0.03470573176488707\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0110 - mean_squared_error: 0.0110 - val_loss: 0.0128 - val_mean_squared_error: 0.0128\n",
            "Model - CoordConv , epoch 59 / 100.\n",
            "Test RMSE - 0.12260566614961405, MAPE - inf, MAE - 0.03495140240875249\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0109 - mean_squared_error: 0.0109 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
            "Model - CoordConv , epoch 60 / 100.\n",
            "Test RMSE - 0.12046531582036293, MAPE - inf, MAE - 0.035929344435715244\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0111 - mean_squared_error: 0.0111 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 61 / 100.\n",
            "Test RMSE - 0.12258738914738383, MAPE - inf, MAE - 0.0340444344104727\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0110 - mean_squared_error: 0.0110 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
            "Model - CoordConv , epoch 62 / 100.\n",
            "Test RMSE - 0.12203691348621341, MAPE - inf, MAE - 0.03531170837882914\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0111 - mean_squared_error: 0.0111 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 63 / 100.\n",
            "Test RMSE - 0.11958933634272387, MAPE - inf, MAE - 0.034670645730631736\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0106 - mean_squared_error: 0.0106 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 64 / 100.\n",
            "Test RMSE - 0.12239619670437019, MAPE - inf, MAE - 0.03373014022708574\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0105 - mean_squared_error: 0.0105 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
            "Model - CoordConv , epoch 65 / 100.\n",
            "Test RMSE - 0.12158650227929647, MAPE - inf, MAE - 0.03369803560634555\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
            "Model - CoordConv , epoch 66 / 100.\n",
            "Test RMSE - 0.11937531134387087, MAPE - inf, MAE - 0.03386963722026203\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 67 / 100.\n",
            "Test RMSE - 0.11881806984758582, MAPE - inf, MAE - 0.03385096128830422\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0107 - mean_squared_error: 0.0107 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
            "Model - CoordConv , epoch 68 / 100.\n",
            "Test RMSE - 0.1384737584745359, MAPE - inf, MAE - 0.034611087070618086\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
            "Model - CoordConv , epoch 69 / 100.\n",
            "Test RMSE - 0.12436905616812807, MAPE - inf, MAE - 0.03449353138199618\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0111 - mean_squared_error: 0.0111 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 70 / 100.\n",
            "Test RMSE - 0.1188340645895205, MAPE - inf, MAE - 0.03487277484678963\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0105 - mean_squared_error: 0.0105 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 71 / 100.\n",
            "Test RMSE - 0.12049052812855193, MAPE - inf, MAE - 0.033404696730608904\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0102 - mean_squared_error: 0.0102 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
            "Model - CoordConv , epoch 72 / 100.\n",
            "Test RMSE - 0.1180502915313277, MAPE - inf, MAE - 0.03397001269882215\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
            "Model - CoordConv , epoch 73 / 100.\n",
            "Test RMSE - 0.11869083279396281, MAPE - inf, MAE - 0.03384583481809845\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0105 - mean_squared_error: 0.0105 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
            "Model - CoordConv , epoch 74 / 100.\n",
            "Test RMSE - 0.12642852566865007, MAPE - inf, MAE - 0.03295026777909912\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0106 - mean_squared_error: 0.0106 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 75 / 100.\n",
            "Test RMSE - 0.1190408068454188, MAPE - inf, MAE - 0.03369751286578402\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0105 - mean_squared_error: 0.0105 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
            "Model - CoordConv , epoch 76 / 100.\n",
            "Test RMSE - 0.11956530968214196, MAPE - inf, MAE - 0.03275027387677624\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
            "Model - CoordConv , epoch 77 / 100.\n",
            "Test RMSE - 0.12214312295946902, MAPE - inf, MAE - 0.032559147865281676\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
            "Model - CoordConv , epoch 78 / 100.\n",
            "Test RMSE - 0.12084485303226812, MAPE - inf, MAE - 0.03618535621353699\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0107 - mean_squared_error: 0.0107 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 79 / 100.\n",
            "Test RMSE - 0.11660483975533957, MAPE - inf, MAE - 0.0340709326291835\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0101 - mean_squared_error: 0.0101 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 80 / 100.\n",
            "Test RMSE - 0.12502455187452388, MAPE - inf, MAE - 0.032742168455347116\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0100 - mean_squared_error: 0.0100 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 81 / 100.\n",
            "Test RMSE - 0.1173102078240897, MAPE - inf, MAE - 0.032504950752593506\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0097 - mean_squared_error: 0.0097 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
            "Model - CoordConv , epoch 82 / 100.\n",
            "Test RMSE - 0.11631305930805926, MAPE - inf, MAE - 0.03268212145723158\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0094 - mean_squared_error: 0.0094 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
            "Model - CoordConv , epoch 83 / 100.\n",
            "Test RMSE - 0.11757655989100287, MAPE - inf, MAE - 0.03226460140333576\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0094 - mean_squared_error: 0.0094 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 84 / 100.\n",
            "Test RMSE - 0.11712679745832509, MAPE - inf, MAE - 0.03402209664821596\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0094 - mean_squared_error: 0.0094 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 85 / 100.\n",
            "Test RMSE - 0.11711960955026905, MAPE - inf, MAE - 0.03192097172350637\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0093 - mean_squared_error: 0.0093 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 86 / 100.\n",
            "Test RMSE - 0.11526931995475026, MAPE - inf, MAE - 0.03260752204153767\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
            "Model - CoordConv , epoch 87 / 100.\n",
            "Test RMSE - 0.11721010529903257, MAPE - inf, MAE - 0.0313846742456563\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
            "Model - CoordConv , epoch 88 / 100.\n",
            "Test RMSE - 0.1170308133775191, MAPE - inf, MAE - 0.03144469810378147\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
            "Model - CoordConv , epoch 89 / 100.\n",
            "Test RMSE - 0.11776643702808554, MAPE - inf, MAE - 0.03159385930986194\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
            "Model - CoordConv , epoch 90 / 100.\n",
            "Test RMSE - 0.11764161203537947, MAPE - inf, MAE - 0.031284408971787435\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 91 / 100.\n",
            "Test RMSE - 0.11647862294060272, MAPE - inf, MAE - 0.033407398072281\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
            "Model - CoordConv , epoch 92 / 100.\n",
            "Test RMSE - 0.12411608551888882, MAPE - inf, MAE - 0.031550825794886404\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0100 - mean_squared_error: 0.0100 - val_loss: 0.0158 - val_mean_squared_error: 0.0158\n",
            "Model - CoordConv , epoch 93 / 100.\n",
            "Test RMSE - 0.12742854187910024, MAPE - inf, MAE - 0.039729382508211425\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0120 - mean_squared_error: 0.0120 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
            "Model - CoordConv , epoch 94 / 100.\n",
            "Test RMSE - 0.13101847276427842, MAPE - inf, MAE - 0.03318571185128099\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0114 - mean_squared_error: 0.0114 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 95 / 100.\n",
            "Test RMSE - 0.11909714661415766, MAPE - inf, MAE - 0.031869504904776746\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0097 - mean_squared_error: 0.0097 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
            "Model - CoordConv , epoch 96 / 100.\n",
            "Test RMSE - 0.11693550630380911, MAPE - inf, MAE - 0.03169724316985784\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
            "Model - CoordConv , epoch 97 / 100.\n",
            "Test RMSE - 0.11516704089554052, MAPE - inf, MAE - 0.03205123472716952\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
            "Model - CoordConv , epoch 98 / 100.\n",
            "Test RMSE - 0.11569267557251574, MAPE - inf, MAE - 0.031999650599403955\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
            "Model - CoordConv , epoch 99 / 100.\n",
            "Test RMSE - 0.11821298700733857, MAPE - inf, MAE - 0.03128709954870673\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0087 - mean_squared_error: 0.0087 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
            "Model - CoordConv , epoch 100 / 100.\n",
            "Test RMSE - 0.11462975082439483, MAPE - inf, MAE - 0.030921854834474446\n",
            "Skipping RandomConv \n",
            "Starting 1 / 5 experiments batch\n",
            "Skipping CNN \n",
            "Skipping LI CNN \n",
            "Model: \"model_11\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_11 (InputLayer)        (None, 30, 30, 27)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_31 (Conv2D)           (None, 30, 30, 21)        226821    \n",
            "_________________________________________________________________\n",
            "conv2d_32 (Conv2D)           (None, 30, 30, 10)        21010     \n",
            "_________________________________________________________________\n",
            "conv2d_33 (Conv2D)           (None, 30, 30, 1)         11        \n",
            "=================================================================\n",
            "Total params: 247,842\n",
            "Trainable params: 247,842\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Training CoordConv model, with 247842 learnable parameters\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0395 - mean_squared_error: 0.0395 - val_loss: 0.0338 - val_mean_squared_error: 0.0338\n",
            "Model - CoordConv , epoch 1 / 100.\n",
            "Test RMSE - 0.18409890304397739, MAPE - inf, MAE - 0.047595556491429804\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0332 - mean_squared_error: 0.0332 - val_loss: 0.0321 - val_mean_squared_error: 0.0321\n",
            "Model - CoordConv , epoch 2 / 100.\n",
            "Test RMSE - 0.17996432021952155, MAPE - inf, MAE - 0.07034238044492414\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0316 - mean_squared_error: 0.0316 - val_loss: 0.0309 - val_mean_squared_error: 0.0309\n",
            "Model - CoordConv , epoch 3 / 100.\n",
            "Test RMSE - 0.17533738651745515, MAPE - inf, MAE - 0.07959552834727734\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0303 - mean_squared_error: 0.0303 - val_loss: 0.0300 - val_mean_squared_error: 0.0300\n",
            "Model - CoordConv , epoch 4 / 100.\n",
            "Test RMSE - 0.17165604993859376, MAPE - inf, MAE - 0.07617502909033612\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0290 - mean_squared_error: 0.0290 - val_loss: 0.0285 - val_mean_squared_error: 0.0285\n",
            "Model - CoordConv , epoch 5 / 100.\n",
            "Test RMSE - 0.16773395741949604, MAPE - inf, MAE - 0.07202619868450631\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0277 - mean_squared_error: 0.0277 - val_loss: 0.0273 - val_mean_squared_error: 0.0273\n",
            "Model - CoordConv , epoch 6 / 100.\n",
            "Test RMSE - 0.16510846155421588, MAPE - inf, MAE - 0.06714994661650174\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0265 - mean_squared_error: 0.0265 - val_loss: 0.0264 - val_mean_squared_error: 0.0264\n",
            "Model - CoordConv , epoch 7 / 100.\n",
            "Test RMSE - 0.16323170384813968, MAPE - inf, MAE - 0.06418900040908912\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0254 - mean_squared_error: 0.0254 - val_loss: 0.0254 - val_mean_squared_error: 0.0254\n",
            "Model - CoordConv , epoch 8 / 100.\n",
            "Test RMSE - 0.16109819480430945, MAPE - inf, MAE - 0.058926329962656135\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0244 - mean_squared_error: 0.0244 - val_loss: 0.0246 - val_mean_squared_error: 0.0246\n",
            "Model - CoordConv , epoch 9 / 100.\n",
            "Test RMSE - 0.15878083308898236, MAPE - inf, MAE - 0.062057996360959274\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0234 - mean_squared_error: 0.0234 - val_loss: 0.0237 - val_mean_squared_error: 0.0237\n",
            "Model - CoordConv , epoch 10 / 100.\n",
            "Test RMSE - 0.15730601575826372, MAPE - inf, MAE - 0.056204159053854535\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0225 - mean_squared_error: 0.0225 - val_loss: 0.0229 - val_mean_squared_error: 0.0229\n",
            "Model - CoordConv , epoch 11 / 100.\n",
            "Test RMSE - 0.15476352542167188, MAPE - inf, MAE - 0.05287947615137924\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0216 - mean_squared_error: 0.0216 - val_loss: 0.0221 - val_mean_squared_error: 0.0221\n",
            "Model - CoordConv , epoch 12 / 100.\n",
            "Test RMSE - 0.15252222180369573, MAPE - inf, MAE - 0.055315443924526385\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0209 - mean_squared_error: 0.0209 - val_loss: 0.0214 - val_mean_squared_error: 0.0214\n",
            "Model - CoordConv , epoch 13 / 100.\n",
            "Test RMSE - 0.14932089837707235, MAPE - inf, MAE - 0.0534020408903734\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0200 - mean_squared_error: 0.0200 - val_loss: 0.0206 - val_mean_squared_error: 0.0206\n",
            "Model - CoordConv , epoch 14 / 100.\n",
            "Test RMSE - 0.1476354228966016, MAPE - inf, MAE - 0.052656432551984364\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0190 - mean_squared_error: 0.0190 - val_loss: 0.0198 - val_mean_squared_error: 0.0198\n",
            "Model - CoordConv , epoch 15 / 100.\n",
            "Test RMSE - 0.1449853996409466, MAPE - inf, MAE - 0.04830830789472328\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0183 - mean_squared_error: 0.0183 - val_loss: 0.0191 - val_mean_squared_error: 0.0191\n",
            "Model - CoordConv , epoch 16 / 100.\n",
            "Test RMSE - 0.14390586130851235, MAPE - inf, MAE - 0.04689929953125023\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0174 - mean_squared_error: 0.0174 - val_loss: 0.0184 - val_mean_squared_error: 0.0184\n",
            "Model - CoordConv , epoch 17 / 100.\n",
            "Test RMSE - 0.13961860884616684, MAPE - inf, MAE - 0.04750692656769382\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0167 - mean_squared_error: 0.0167 - val_loss: 0.0177 - val_mean_squared_error: 0.0177\n",
            "Model - CoordConv , epoch 18 / 100.\n",
            "Test RMSE - 0.13939550746348336, MAPE - inf, MAE - 0.04481610687966555\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0160 - mean_squared_error: 0.0160 - val_loss: 0.0173 - val_mean_squared_error: 0.0173\n",
            "Model - CoordConv , epoch 19 / 100.\n",
            "Test RMSE - 0.13564185210790283, MAPE - inf, MAE - 0.04391553996728042\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0156 - mean_squared_error: 0.0156 - val_loss: 0.0167 - val_mean_squared_error: 0.0167\n",
            "Model - CoordConv , epoch 20 / 100.\n",
            "Test RMSE - 0.13576461159855296, MAPE - inf, MAE - 0.04246879293607348\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0149 - mean_squared_error: 0.0149 - val_loss: 0.0167 - val_mean_squared_error: 0.0167\n",
            "Model - CoordConv , epoch 21 / 100.\n",
            "Test RMSE - 0.13579380606596136, MAPE - inf, MAE - 0.04107161769986105\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0146 - mean_squared_error: 0.0146 - val_loss: 0.0161 - val_mean_squared_error: 0.0161\n",
            "Model - CoordConv , epoch 22 / 100.\n",
            "Test RMSE - 0.13216636779567814, MAPE - inf, MAE - 0.04071263682090389\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0141 - mean_squared_error: 0.0141 - val_loss: 0.0161 - val_mean_squared_error: 0.0161\n",
            "Model - CoordConv , epoch 23 / 100.\n",
            "Test RMSE - 0.12961217483424867, MAPE - inf, MAE - 0.04134631878473846\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 24 / 100.\n",
            "Test RMSE - 0.12986068105317547, MAPE - inf, MAE - 0.03920885001130566\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0132 - mean_squared_error: 0.0132 - val_loss: 0.0155 - val_mean_squared_error: 0.0155\n",
            "Model - CoordConv , epoch 25 / 100.\n",
            "Test RMSE - 0.12768822055019688, MAPE - inf, MAE - 0.038778878991598896\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0128 - mean_squared_error: 0.0128 - val_loss: 0.0147 - val_mean_squared_error: 0.0147\n",
            "Model - CoordConv , epoch 26 / 100.\n",
            "Test RMSE - 0.1278168088528956, MAPE - inf, MAE - 0.03797266015329955\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n",
            "Model - CoordConv , epoch 27 / 100.\n",
            "Test RMSE - 0.12486120183189857, MAPE - inf, MAE - 0.037190330764935456\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0122 - mean_squared_error: 0.0122 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
            "Model - CoordConv , epoch 28 / 100.\n",
            "Test RMSE - 0.13206811382828584, MAPE - inf, MAE - 0.03632688048647967\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0126 - mean_squared_error: 0.0126 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
            "Model - CoordConv , epoch 29 / 100.\n",
            "Test RMSE - 0.12806315183646075, MAPE - inf, MAE - 0.036662259699075245\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0117 - mean_squared_error: 0.0117 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
            "Model - CoordConv , epoch 30 / 100.\n",
            "Test RMSE - 0.12276746253367354, MAPE - inf, MAE - 0.03573183995397883\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0110 - mean_squared_error: 0.0110 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
            "Model - CoordConv , epoch 31 / 100.\n",
            "Test RMSE - 0.12030794178301944, MAPE - inf, MAE - 0.03567690658648338\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0108 - mean_squared_error: 0.0108 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
            "Model - CoordConv , epoch 32 / 100.\n",
            "Test RMSE - 0.11940777679271937, MAPE - inf, MAE - 0.034896829906337586\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
            "Model - CoordConv , epoch 33 / 100.\n",
            "Test RMSE - 0.11763719959287988, MAPE - inf, MAE - 0.033254924346679725\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
            "Model - CoordConv , epoch 34 / 100.\n",
            "Test RMSE - 0.11723614247161535, MAPE - inf, MAE - 0.03447753129033192\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0100 - mean_squared_error: 0.0100 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 35 / 100.\n",
            "Test RMSE - 0.11578807955505203, MAPE - inf, MAE - 0.03241254231152538\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0098 - mean_squared_error: 0.0098 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 36 / 100.\n",
            "Test RMSE - 0.11596708956728692, MAPE - inf, MAE - 0.03370128469806889\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0095 - mean_squared_error: 0.0095 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
            "Model - CoordConv , epoch 37 / 100.\n",
            "Test RMSE - 0.11467293679084026, MAPE - inf, MAE - 0.03137802715960403\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0093 - mean_squared_error: 0.0093 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
            "Model - CoordConv , epoch 38 / 100.\n",
            "Test RMSE - 0.11554218179614414, MAPE - inf, MAE - 0.03232831272837131\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 39 / 100.\n",
            "Test RMSE - 0.11554420233427394, MAPE - inf, MAE - 0.03165176031903606\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
            "Model - CoordConv , epoch 40 / 100.\n",
            "Test RMSE - 0.11163971871661067, MAPE - inf, MAE - 0.03233715337599406\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0086 - mean_squared_error: 0.0086 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
            "Model - CoordConv , epoch 41 / 100.\n",
            "Test RMSE - 0.11397047010275864, MAPE - inf, MAE - 0.03037786213793888\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 42 / 100.\n",
            "Test RMSE - 0.10919430010290651, MAPE - inf, MAE - 0.0310348004774237\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0086 - mean_squared_error: 0.0086 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
            "Model - CoordConv , epoch 43 / 100.\n",
            "Test RMSE - 0.10801423327926442, MAPE - inf, MAE - 0.030326615673389082\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0081 - mean_squared_error: 0.0081 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
            "Model - CoordConv , epoch 44 / 100.\n",
            "Test RMSE - 0.10835973228409448, MAPE - inf, MAE - 0.030089137599130976\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0081 - mean_squared_error: 0.0081 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
            "Model - CoordConv , epoch 45 / 100.\n",
            "Test RMSE - 0.10694896144843635, MAPE - inf, MAE - 0.030805110022015383\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0078 - mean_squared_error: 0.0078 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
            "Model - CoordConv , epoch 46 / 100.\n",
            "Test RMSE - 0.10733471176585371, MAPE - inf, MAE - 0.029104003006336802\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0078 - mean_squared_error: 0.0078 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
            "Model - CoordConv , epoch 47 / 100.\n",
            "Test RMSE - 0.10529408644779568, MAPE - inf, MAE - 0.02990992808735033\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0076 - mean_squared_error: 0.0076 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
            "Model - CoordConv , epoch 48 / 100.\n",
            "Test RMSE - 0.10464873259198272, MAPE - inf, MAE - 0.028461311786152316\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0073 - mean_squared_error: 0.0073 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
            "Model - CoordConv , epoch 49 / 100.\n",
            "Test RMSE - 0.10614768586163745, MAPE - inf, MAE - 0.028527906873003384\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0071 - mean_squared_error: 0.0071 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
            "Model - CoordConv , epoch 50 / 100.\n",
            "Test RMSE - 0.1045151923151749, MAPE - inf, MAE - 0.027919419390775902\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0070 - mean_squared_error: 0.0070 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
            "Model - CoordConv , epoch 51 / 100.\n",
            "Test RMSE - 0.10641284380500529, MAPE - inf, MAE - 0.027853645944701796\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0079 - mean_squared_error: 0.0079 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 52 / 100.\n",
            "Test RMSE - 0.10630580043930807, MAPE - inf, MAE - 0.03072035302551459\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0078 - mean_squared_error: 0.0078 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
            "Model - CoordConv , epoch 53 / 100.\n",
            "Test RMSE - 0.10446580572396547, MAPE - inf, MAE - 0.02979296090604046\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0073 - mean_squared_error: 0.0073 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
            "Model - CoordConv , epoch 54 / 100.\n",
            "Test RMSE - 0.10405886376335874, MAPE - inf, MAE - 0.02979212707422771\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0069 - mean_squared_error: 0.0069 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
            "Model - CoordConv , epoch 55 / 100.\n",
            "Test RMSE - 0.10097240498522009, MAPE - inf, MAE - 0.028060765192354092\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0066 - mean_squared_error: 0.0066 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
            "Model - CoordConv , epoch 56 / 100.\n",
            "Test RMSE - 0.10088970932967613, MAPE - inf, MAE - 0.02722035808451468\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0093 - val_mean_squared_error: 0.0093\n",
            "Model - CoordConv , epoch 57 / 100.\n",
            "Test RMSE - 0.10109363208392311, MAPE - inf, MAE - 0.02709089140581112\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
            "Model - CoordConv , epoch 58 / 100.\n",
            "Test RMSE - 0.10098694516705611, MAPE - inf, MAE - 0.02605493212715215\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0061 - mean_squared_error: 0.0061 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 59 / 100.\n",
            "Test RMSE - 0.10045856709767347, MAPE - inf, MAE - 0.027248766735763704\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
            "Model - CoordConv , epoch 60 / 100.\n",
            "Test RMSE - 0.09931959363397005, MAPE - inf, MAE - 0.026453616674735544\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0061 - mean_squared_error: 0.0061 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 61 / 100.\n",
            "Test RMSE - 0.10065277347424427, MAPE - inf, MAE - 0.027231966912328476\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 62 / 100.\n",
            "Test RMSE - 0.10009566763996229, MAPE - inf, MAE - 0.025904594282061287\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0063 - mean_squared_error: 0.0063 - val_loss: 0.0095 - val_mean_squared_error: 0.0095\n",
            "Model - CoordConv , epoch 63 / 100.\n",
            "Test RMSE - 0.09997572573049084, MAPE - inf, MAE - 0.02774107473458214\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0063 - mean_squared_error: 0.0063 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 64 / 100.\n",
            "Test RMSE - 0.09952976801256641, MAPE - inf, MAE - 0.027643676140452357\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 65 / 100.\n",
            "Test RMSE - 0.09759712464408221, MAPE - inf, MAE - 0.02619199871543517\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0060 - mean_squared_error: 0.0060 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
            "Model - CoordConv , epoch 66 / 100.\n",
            "Test RMSE - 0.09927820761329333, MAPE - inf, MAE - 0.025763931183441616\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
            "Model - CoordConv , epoch 67 / 100.\n",
            "Test RMSE - 0.09917757315122751, MAPE - inf, MAE - 0.02543854773432024\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0055 - mean_squared_error: 0.0055 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 68 / 100.\n",
            "Test RMSE - 0.09676643260427946, MAPE - inf, MAE - 0.025406650348883505\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0055 - mean_squared_error: 0.0055 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
            "Model - CoordConv , epoch 69 / 100.\n",
            "Test RMSE - 0.09974840990842458, MAPE - inf, MAE - 0.025039877143699825\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0055 - mean_squared_error: 0.0055 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 70 / 100.\n",
            "Test RMSE - 0.09756022794021657, MAPE - inf, MAE - 0.026210887111294753\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 71 / 100.\n",
            "Test RMSE - 0.09736754207822701, MAPE - inf, MAE - 0.025692552939207094\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 72 / 100.\n",
            "Test RMSE - 0.10206216325230766, MAPE - inf, MAE - 0.025733210425066013\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0057 - mean_squared_error: 0.0057 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
            "Model - CoordConv , epoch 73 / 100.\n",
            "Test RMSE - 0.09797541557225926, MAPE - inf, MAE - 0.025203109624888954\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0087 - val_mean_squared_error: 0.0087\n",
            "Model - CoordConv , epoch 74 / 100.\n",
            "Test RMSE - 0.09716749696667742, MAPE - inf, MAE - 0.02537494650921232\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0087 - val_mean_squared_error: 0.0087\n",
            "Model - CoordConv , epoch 75 / 100.\n",
            "Test RMSE - 0.0964136897925122, MAPE - inf, MAE - 0.02564646574430082\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0086 - val_mean_squared_error: 0.0086\n",
            "Model - CoordConv , epoch 76 / 100.\n",
            "Test RMSE - 0.09698445156394335, MAPE - inf, MAE - 0.024930649844652618\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0086 - val_mean_squared_error: 0.0086\n",
            "Model - CoordConv , epoch 77 / 100.\n",
            "Test RMSE - 0.09554905883326432, MAPE - inf, MAE - 0.02496700489393479\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0052 - mean_squared_error: 0.0052 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 78 / 100.\n",
            "Test RMSE - 0.09639883933098697, MAPE - inf, MAE - 0.025863861386301798\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
            "Model - CoordConv , epoch 79 / 100.\n",
            "Test RMSE - 0.10214366648541187, MAPE - inf, MAE - 0.024958066486806842\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 80 / 100.\n",
            "Test RMSE - 0.09895572413077568, MAPE - inf, MAE - 0.025371014830884383\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 81 / 100.\n",
            "Test RMSE - 0.09667166680456103, MAPE - inf, MAE - 0.02572146064011822\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 82 / 100.\n",
            "Test RMSE - 0.09734521867968722, MAPE - inf, MAE - 0.026363509323200397\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0083 - val_mean_squared_error: 0.0083\n",
            "Model - CoordConv , epoch 83 / 100.\n",
            "Test RMSE - 0.0950382814864704, MAPE - inf, MAE - 0.024284764727474394\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0049 - mean_squared_error: 0.0049 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 84 / 100.\n",
            "Test RMSE - 0.09742142262708328, MAPE - inf, MAE - 0.024399584183142066\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0049 - mean_squared_error: 0.0049 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 85 / 100.\n",
            "Test RMSE - 0.09502845653013285, MAPE - inf, MAE - 0.024559672280319748\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 86 / 100.\n",
            "Test RMSE - 0.093073031407308, MAPE - inf, MAE - 0.024337486999032817\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 87 / 100.\n",
            "Test RMSE - 0.09421913271498879, MAPE - inf, MAE - 0.02424820023264818\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 88 / 100.\n",
            "Test RMSE - 0.09469926998715145, MAPE - inf, MAE - 0.02438370932724208\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 89 / 100.\n",
            "Test RMSE - 0.09525757161001434, MAPE - inf, MAE - 0.02408294831981797\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0047 - mean_squared_error: 0.0047 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 90 / 100.\n",
            "Test RMSE - 0.09338421819690038, MAPE - inf, MAE - 0.023716955542383115\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 91 / 100.\n",
            "Test RMSE - 0.09688887847245921, MAPE - inf, MAE - 0.02423423676246711\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0047 - mean_squared_error: 0.0047 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 92 / 100.\n",
            "Test RMSE - 0.09555293717505596, MAPE - inf, MAE - 0.023984655910016212\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0047 - mean_squared_error: 0.0047 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 93 / 100.\n",
            "Test RMSE - 0.09442923968964823, MAPE - inf, MAE - 0.024629700040064526\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0047 - mean_squared_error: 0.0047 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 94 / 100.\n",
            "Test RMSE - 0.09296605194170299, MAPE - inf, MAE - 0.023474655162424255\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0080 - val_mean_squared_error: 0.0080\n",
            "Model - CoordConv , epoch 95 / 100.\n",
            "Test RMSE - 0.09343356845032123, MAPE - inf, MAE - 0.024003150510567665\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0045 - mean_squared_error: 0.0045 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 96 / 100.\n",
            "Test RMSE - 0.09314169500004789, MAPE - inf, MAE - 0.024182605649350222\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 97 / 100.\n",
            "Test RMSE - 0.09493897351634362, MAPE - inf, MAE - 0.023399809639693396\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0080 - val_mean_squared_error: 0.0080\n",
            "Model - CoordConv , epoch 98 / 100.\n",
            "Test RMSE - 0.09255607832455029, MAPE - inf, MAE - 0.02395012153570651\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0079 - val_mean_squared_error: 0.0079\n",
            "Model - CoordConv , epoch 99 / 100.\n",
            "Test RMSE - 0.09268596937548024, MAPE - inf, MAE - 0.02372114746927664\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0044 - mean_squared_error: 0.0044 - val_loss: 0.0079 - val_mean_squared_error: 0.0079\n",
            "Model - CoordConv , epoch 100 / 100.\n",
            "Test RMSE - 0.09424112075310301, MAPE - inf, MAE - 0.02358122152293954\n",
            "Skipping RandomConv \n",
            "Starting 2 / 5 experiments batch\n",
            "Skipping CNN \n",
            "Skipping LI CNN \n",
            "Model: \"model_12\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_12 (InputLayer)        (None, 30, 30, 27)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_34 (Conv2D)           (None, 30, 30, 21)        226821    \n",
            "_________________________________________________________________\n",
            "conv2d_35 (Conv2D)           (None, 30, 30, 10)        21010     \n",
            "_________________________________________________________________\n",
            "conv2d_36 (Conv2D)           (None, 30, 30, 1)         11        \n",
            "=================================================================\n",
            "Total params: 247,842\n",
            "Trainable params: 247,842\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Training CoordConv model, with 247842 learnable parameters\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0406 - mean_squared_error: 0.0406 - val_loss: 0.0348 - val_mean_squared_error: 0.0348\n",
            "Model - CoordConv , epoch 1 / 100.\n",
            "Test RMSE - 0.18651281895544858, MAPE - inf, MAE - 0.04550299381092488\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0347 - mean_squared_error: 0.0347 - val_loss: 0.0345 - val_mean_squared_error: 0.0345\n",
            "Model - CoordConv , epoch 2 / 100.\n",
            "Test RMSE - 0.18564911034938728, MAPE - inf, MAE - 0.04109378997132505\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0343 - mean_squared_error: 0.0343 - val_loss: 0.0341 - val_mean_squared_error: 0.0341\n",
            "Model - CoordConv , epoch 3 / 100.\n",
            "Test RMSE - 0.18461788468380577, MAPE - inf, MAE - 0.04523142636722835\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0339 - mean_squared_error: 0.0339 - val_loss: 0.0337 - val_mean_squared_error: 0.0337\n",
            "Model - CoordConv , epoch 4 / 100.\n",
            "Test RMSE - 0.18358629557791264, MAPE - inf, MAE - 0.05001109670044449\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0336 - mean_squared_error: 0.0336 - val_loss: 0.0334 - val_mean_squared_error: 0.0334\n",
            "Model - CoordConv , epoch 5 / 100.\n",
            "Test RMSE - 0.1825991258956348, MAPE - inf, MAE - 0.054926561573301204\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0332 - mean_squared_error: 0.0332 - val_loss: 0.0330 - val_mean_squared_error: 0.0330\n",
            "Model - CoordConv , epoch 6 / 100.\n",
            "Test RMSE - 0.1815197471942697, MAPE - inf, MAE - 0.06033196630875548\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0329 - mean_squared_error: 0.0329 - val_loss: 0.0327 - val_mean_squared_error: 0.0327\n",
            "Model - CoordConv , epoch 7 / 100.\n",
            "Test RMSE - 0.1802240212263561, MAPE - inf, MAE - 0.06453136895833064\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0324 - mean_squared_error: 0.0324 - val_loss: 0.0323 - val_mean_squared_error: 0.0323\n",
            "Model - CoordConv , epoch 8 / 100.\n",
            "Test RMSE - 0.17823417292951638, MAPE - inf, MAE - 0.06778137815918386\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0318 - mean_squared_error: 0.0318 - val_loss: 0.0314 - val_mean_squared_error: 0.0314\n",
            "Model - CoordConv , epoch 9 / 100.\n",
            "Test RMSE - 0.17535500503587995, MAPE - inf, MAE - 0.06978651636619132\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0309 - mean_squared_error: 0.0309 - val_loss: 0.0305 - val_mean_squared_error: 0.0305\n",
            "Model - CoordConv , epoch 10 / 100.\n",
            "Test RMSE - 0.17257613084699366, MAPE - inf, MAE - 0.07183365668780806\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0296 - mean_squared_error: 0.0296 - val_loss: 0.0292 - val_mean_squared_error: 0.0292\n",
            "Model - CoordConv , epoch 11 / 100.\n",
            "Test RMSE - 0.16821765699214516, MAPE - inf, MAE - 0.08046275228193193\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0279 - mean_squared_error: 0.0279 - val_loss: 0.0271 - val_mean_squared_error: 0.0271\n",
            "Model - CoordConv , epoch 12 / 100.\n",
            "Test RMSE - 0.16409137322189446, MAPE - inf, MAE - 0.07560251610789828\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0258 - mean_squared_error: 0.0258 - val_loss: 0.0255 - val_mean_squared_error: 0.0255\n",
            "Model - CoordConv , epoch 13 / 100.\n",
            "Test RMSE - 0.15902444285095285, MAPE - inf, MAE - 0.07529747778491615\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0238 - mean_squared_error: 0.0238 - val_loss: 0.0236 - val_mean_squared_error: 0.0236\n",
            "Model - CoordConv , epoch 14 / 100.\n",
            "Test RMSE - 0.1545083123062083, MAPE - inf, MAE - 0.06933087370792412\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0221 - mean_squared_error: 0.0221 - val_loss: 0.0222 - val_mean_squared_error: 0.0222\n",
            "Model - CoordConv , epoch 15 / 100.\n",
            "Test RMSE - 0.15162123839730687, MAPE - inf, MAE - 0.06398418083735624\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0205 - mean_squared_error: 0.0205 - val_loss: 0.0205 - val_mean_squared_error: 0.0205\n",
            "Model - CoordConv , epoch 16 / 100.\n",
            "Test RMSE - 0.14514828520984327, MAPE - inf, MAE - 0.06076724206823196\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0190 - mean_squared_error: 0.0190 - val_loss: 0.0190 - val_mean_squared_error: 0.0190\n",
            "Model - CoordConv , epoch 17 / 100.\n",
            "Test RMSE - 0.14042769351008244, MAPE - inf, MAE - 0.054290993084991725\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0176 - mean_squared_error: 0.0176 - val_loss: 0.0183 - val_mean_squared_error: 0.0183\n",
            "Model - CoordConv , epoch 18 / 100.\n",
            "Test RMSE - 0.13805486647015297, MAPE - inf, MAE - 0.05057823845159044\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0173 - mean_squared_error: 0.0173 - val_loss: 0.0176 - val_mean_squared_error: 0.0176\n",
            "Model - CoordConv , epoch 19 / 100.\n",
            "Test RMSE - 0.1347965756191018, MAPE - inf, MAE - 0.049694924587168426\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0162 - mean_squared_error: 0.0162 - val_loss: 0.0170 - val_mean_squared_error: 0.0170\n",
            "Model - CoordConv , epoch 20 / 100.\n",
            "Test RMSE - 0.13559026771603988, MAPE - inf, MAE - 0.0445643521041441\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0152 - mean_squared_error: 0.0152 - val_loss: 0.0166 - val_mean_squared_error: 0.0166\n",
            "Model - CoordConv , epoch 21 / 100.\n",
            "Test RMSE - 0.13304864962955176, MAPE - inf, MAE - 0.042872302587398464\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0146 - mean_squared_error: 0.0146 - val_loss: 0.0160 - val_mean_squared_error: 0.0160\n",
            "Model - CoordConv , epoch 22 / 100.\n",
            "Test RMSE - 0.1323776420160521, MAPE - inf, MAE - 0.04049782023936725\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 23 / 100.\n",
            "Test RMSE - 0.13077085654921644, MAPE - inf, MAE - 0.03993283546043379\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0157 - val_mean_squared_error: 0.0157\n",
            "Model - CoordConv , epoch 24 / 100.\n",
            "Test RMSE - 0.1278250932195929, MAPE - inf, MAE - 0.04249012242250049\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0133 - mean_squared_error: 0.0133 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
            "Model - CoordConv , epoch 25 / 100.\n",
            "Test RMSE - 0.1268640969956684, MAPE - inf, MAE - 0.03717604362746117\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0127 - mean_squared_error: 0.0127 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
            "Model - CoordConv , epoch 26 / 100.\n",
            "Test RMSE - 0.13189242834960463, MAPE - inf, MAE - 0.03614889925522553\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0126 - mean_squared_error: 0.0126 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
            "Model - CoordConv , epoch 27 / 100.\n",
            "Test RMSE - 0.12444209433757364, MAPE - inf, MAE - 0.03638310473969742\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0119 - mean_squared_error: 0.0119 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
            "Model - CoordConv , epoch 28 / 100.\n",
            "Test RMSE - 0.12293910819301689, MAPE - inf, MAE - 0.035976630178405064\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0115 - mean_squared_error: 0.0115 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
            "Model - CoordConv , epoch 29 / 100.\n",
            "Test RMSE - 0.12435636742516336, MAPE - inf, MAE - 0.0347330518900471\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0114 - mean_squared_error: 0.0114 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
            "Model - CoordConv , epoch 30 / 100.\n",
            "Test RMSE - 0.12138284865218452, MAPE - inf, MAE - 0.03531466372356882\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0110 - mean_squared_error: 0.0110 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
            "Model - CoordConv , epoch 31 / 100.\n",
            "Test RMSE - 0.12007370424709332, MAPE - inf, MAE - 0.03563560053912283\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0113 - mean_squared_error: 0.0113 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
            "Model - CoordConv , epoch 32 / 100.\n",
            "Test RMSE - 0.12826653912479022, MAPE - inf, MAE - 0.0333755192508268\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0113 - mean_squared_error: 0.0113 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
            "Model - CoordConv , epoch 33 / 100.\n",
            "Test RMSE - 0.1203767058915897, MAPE - inf, MAE - 0.0336956136617074\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0112 - mean_squared_error: 0.0112 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
            "Model - CoordConv , epoch 34 / 100.\n",
            "Test RMSE - 0.11803509569023225, MAPE - inf, MAE - 0.03574168492901313\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0106 - mean_squared_error: 0.0106 - val_loss: 0.0128 - val_mean_squared_error: 0.0128\n",
            "Model - CoordConv , epoch 35 / 100.\n",
            "Test RMSE - 0.11702512460184059, MAPE - inf, MAE - 0.03323986555332138\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0101 - mean_squared_error: 0.0101 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
            "Model - CoordConv , epoch 36 / 100.\n",
            "Test RMSE - 0.11580059452099058, MAPE - inf, MAE - 0.03302008035330127\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0097 - mean_squared_error: 0.0097 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
            "Model - CoordConv , epoch 37 / 100.\n",
            "Test RMSE - 0.11531850691734931, MAPE - inf, MAE - 0.03298985455746997\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0096 - mean_squared_error: 0.0096 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 38 / 100.\n",
            "Test RMSE - 0.1180999094104981, MAPE - inf, MAE - 0.03127688819069711\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0094 - mean_squared_error: 0.0094 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
            "Model - CoordConv , epoch 39 / 100.\n",
            "Test RMSE - 0.11587850312240261, MAPE - inf, MAE - 0.030960771985210876\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
            "Model - CoordConv , epoch 40 / 100.\n",
            "Test RMSE - 0.11482081296346172, MAPE - inf, MAE - 0.03369091119152675\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0092 - mean_squared_error: 0.0092 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
            "Model - CoordConv , epoch 41 / 100.\n",
            "Test RMSE - 0.11382434086463072, MAPE - inf, MAE - 0.030346209218275005\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
            "Model - CoordConv , epoch 42 / 100.\n",
            "Test RMSE - 0.11056791287924951, MAPE - inf, MAE - 0.031056299311279817\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
            "Model - CoordConv , epoch 43 / 100.\n",
            "Test RMSE - 0.11518098722510445, MAPE - inf, MAE - 0.029702693533855396\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 44 / 100.\n",
            "Test RMSE - 0.11001274620423986, MAPE - inf, MAE - 0.030585789985884394\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0083 - mean_squared_error: 0.0083 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
            "Model - CoordConv , epoch 45 / 100.\n",
            "Test RMSE - 0.11080163024569907, MAPE - inf, MAE - 0.029279058623439134\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0082 - mean_squared_error: 0.0082 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
            "Model - CoordConv , epoch 46 / 100.\n",
            "Test RMSE - 0.11041485708343766, MAPE - inf, MAE - 0.029073072179547493\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0084 - mean_squared_error: 0.0084 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
            "Model - CoordConv , epoch 47 / 100.\n",
            "Test RMSE - 0.10764533982271277, MAPE - inf, MAE - 0.029372239666124098\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0079 - mean_squared_error: 0.0079 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
            "Model - CoordConv , epoch 48 / 100.\n",
            "Test RMSE - 0.11215340942326095, MAPE - inf, MAE - 0.028917326205742593\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0077 - mean_squared_error: 0.0077 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 49 / 100.\n",
            "Test RMSE - 0.10952628129761605, MAPE - inf, MAE - 0.031048594378116814\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0083 - mean_squared_error: 0.0083 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
            "Model - CoordConv , epoch 50 / 100.\n",
            "Test RMSE - 0.11019616780480353, MAPE - inf, MAE - 0.02840986634943587\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0077 - mean_squared_error: 0.0077 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
            "Model - CoordConv , epoch 51 / 100.\n",
            "Test RMSE - 0.11173794814880458, MAPE - inf, MAE - 0.02857324108878912\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0083 - mean_squared_error: 0.0083 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
            "Model - CoordConv , epoch 52 / 100.\n",
            "Test RMSE - 0.11007956688031043, MAPE - inf, MAE - 0.030979019659131486\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0080 - mean_squared_error: 0.0080 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
            "Model - CoordConv , epoch 53 / 100.\n",
            "Test RMSE - 0.10751996143648287, MAPE - inf, MAE - 0.028245721598398516\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0075 - mean_squared_error: 0.0075 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
            "Model - CoordConv , epoch 54 / 100.\n",
            "Test RMSE - 0.1099795163573315, MAPE - inf, MAE - 0.027823036771483307\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0075 - mean_squared_error: 0.0075 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
            "Model - CoordConv , epoch 55 / 100.\n",
            "Test RMSE - 0.1069775392184725, MAPE - inf, MAE - 0.028344654908242013\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0073 - mean_squared_error: 0.0073 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
            "Model - CoordConv , epoch 56 / 100.\n",
            "Test RMSE - 0.10645234047376295, MAPE - inf, MAE - 0.029226951678308414\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0072 - mean_squared_error: 0.0072 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
            "Model - CoordConv , epoch 57 / 100.\n",
            "Test RMSE - 0.10489595750235484, MAPE - inf, MAE - 0.027886623458556904\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0069 - mean_squared_error: 0.0069 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
            "Model - CoordConv , epoch 58 / 100.\n",
            "Test RMSE - 0.10682659498416242, MAPE - inf, MAE - 0.027059137462436784\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0069 - mean_squared_error: 0.0069 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
            "Model - CoordConv , epoch 59 / 100.\n",
            "Test RMSE - 0.10412710813487838, MAPE - inf, MAE - 0.027488859607898247\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0073 - mean_squared_error: 0.0073 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
            "Model - CoordConv , epoch 60 / 100.\n",
            "Test RMSE - 0.10363657908789844, MAPE - inf, MAE - 0.027602552044684026\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0070 - mean_squared_error: 0.0070 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
            "Model - CoordConv , epoch 61 / 100.\n",
            "Test RMSE - 0.1066859910429854, MAPE - inf, MAE - 0.027245727008011134\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0067 - mean_squared_error: 0.0067 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
            "Model - CoordConv , epoch 62 / 100.\n",
            "Test RMSE - 0.1073226597202834, MAPE - inf, MAE - 0.026676273869073064\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0067 - mean_squared_error: 0.0067 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 63 / 100.\n",
            "Test RMSE - 0.10347065975712516, MAPE - inf, MAE - 0.026715022816432704\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0093 - val_mean_squared_error: 0.0093\n",
            "Model - CoordConv , epoch 64 / 100.\n",
            "Test RMSE - 0.10375987987748889, MAPE - inf, MAE - 0.026653978626668733\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0095 - val_mean_squared_error: 0.0095\n",
            "Model - CoordConv , epoch 65 / 100.\n",
            "Test RMSE - 0.10667090271265679, MAPE - inf, MAE - 0.026599851725940062\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 66 / 100.\n",
            "Test RMSE - 0.10355049630398595, MAPE - inf, MAE - 0.026471203007302\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 67 / 100.\n",
            "Test RMSE - 0.1032469482169571, MAPE - inf, MAE - 0.026874601356386196\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0061 - mean_squared_error: 0.0061 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 68 / 100.\n",
            "Test RMSE - 0.10526777534471174, MAPE - inf, MAE - 0.026238838504612054\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 69 / 100.\n",
            "Test RMSE - 0.10202497853316772, MAPE - inf, MAE - 0.026851868479930655\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0064 - mean_squared_error: 0.0064 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 70 / 100.\n",
            "Test RMSE - 0.1042445415076798, MAPE - inf, MAE - 0.025682850526741383\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0062 - mean_squared_error: 0.0062 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
            "Model - CoordConv , epoch 71 / 100.\n",
            "Test RMSE - 0.10291114240876297, MAPE - inf, MAE - 0.02583438948535475\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0059 - mean_squared_error: 0.0059 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 72 / 100.\n",
            "Test RMSE - 0.1022555442907339, MAPE - inf, MAE - 0.02608977062168375\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0058 - mean_squared_error: 0.0058 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 73 / 100.\n",
            "Test RMSE - 0.10279036994977754, MAPE - inf, MAE - 0.025408660621506436\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0058 - mean_squared_error: 0.0058 - val_loss: 0.0094 - val_mean_squared_error: 0.0094\n",
            "Model - CoordConv , epoch 74 / 100.\n",
            "Test RMSE - 0.1054228625328088, MAPE - inf, MAE - 0.02542332241347255\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0059 - mean_squared_error: 0.0059 - val_loss: 0.0093 - val_mean_squared_error: 0.0093\n",
            "Model - CoordConv , epoch 75 / 100.\n",
            "Test RMSE - 0.10097560383051624, MAPE - inf, MAE - 0.026719226692251286\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0057 - mean_squared_error: 0.0057 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 76 / 100.\n",
            "Test RMSE - 0.10352655628760589, MAPE - inf, MAE - 0.02526558368845845\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 77 / 100.\n",
            "Test RMSE - 0.10120537455434434, MAPE - inf, MAE - 0.02603465332636083\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0086 - val_mean_squared_error: 0.0086\n",
            "Model - CoordConv , epoch 78 / 100.\n",
            "Test RMSE - 0.10107002424471673, MAPE - inf, MAE - 0.025242051841805037\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0088 - val_mean_squared_error: 0.0088\n",
            "Model - CoordConv , epoch 79 / 100.\n",
            "Test RMSE - 0.10157315254838993, MAPE - inf, MAE - 0.024662520198647408\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0057 - mean_squared_error: 0.0057 - val_loss: 0.0090 - val_mean_squared_error: 0.0090\n",
            "Model - CoordConv , epoch 80 / 100.\n",
            "Test RMSE - 0.10006754897084102, MAPE - inf, MAE - 0.026045979157578406\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0060 - mean_squared_error: 0.0060 - val_loss: 0.0091 - val_mean_squared_error: 0.0091\n",
            "Model - CoordConv , epoch 81 / 100.\n",
            "Test RMSE - 0.10342630536593028, MAPE - inf, MAE - 0.025056375186793094\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0059 - mean_squared_error: 0.0059 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 82 / 100.\n",
            "Test RMSE - 0.09953105083545562, MAPE - inf, MAE - 0.024680302234840957\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0085 - val_mean_squared_error: 0.0085\n",
            "Model - CoordConv , epoch 83 / 100.\n",
            "Test RMSE - 0.10112576102291886, MAPE - inf, MAE - 0.024989333005262602\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0053 - mean_squared_error: 0.0053 - val_loss: 0.0085 - val_mean_squared_error: 0.0085\n",
            "Model - CoordConv , epoch 84 / 100.\n",
            "Test RMSE - 0.10069572136446812, MAPE - inf, MAE - 0.02491488880582495\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0051 - mean_squared_error: 0.0051 - val_loss: 0.0085 - val_mean_squared_error: 0.0085\n",
            "Model - CoordConv , epoch 85 / 100.\n",
            "Test RMSE - 0.09976631249237447, MAPE - inf, MAE - 0.02477585716339568\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0050 - mean_squared_error: 0.0050 - val_loss: 0.0083 - val_mean_squared_error: 0.0083\n",
            "Model - CoordConv , epoch 86 / 100.\n",
            "Test RMSE - 0.09873227177580488, MAPE - inf, MAE - 0.024452618086405167\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0050 - mean_squared_error: 0.0050 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 87 / 100.\n",
            "Test RMSE - 0.09896650906391034, MAPE - inf, MAE - 0.02444772028638313\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0049 - mean_squared_error: 0.0049 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 88 / 100.\n",
            "Test RMSE - 0.09927704280328051, MAPE - inf, MAE - 0.02436276582379939\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0092 - val_mean_squared_error: 0.0092\n",
            "Model - CoordConv , epoch 89 / 100.\n",
            "Test RMSE - 0.10617142786200535, MAPE - inf, MAE - 0.024972157373610138\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0086 - val_mean_squared_error: 0.0086\n",
            "Model - CoordConv , epoch 90 / 100.\n",
            "Test RMSE - 0.10104442574507427, MAPE - inf, MAE - 0.024935087539133108\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0056 - mean_squared_error: 0.0056 - val_loss: 0.0089 - val_mean_squared_error: 0.0089\n",
            "Model - CoordConv , epoch 91 / 100.\n",
            "Test RMSE - 0.10267003775660083, MAPE - inf, MAE - 0.024592958964288838\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0054 - mean_squared_error: 0.0054 - val_loss: 0.0085 - val_mean_squared_error: 0.0085\n",
            "Model - CoordConv , epoch 92 / 100.\n",
            "Test RMSE - 0.10001064629640213, MAPE - inf, MAE - 0.02438673702844824\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0050 - mean_squared_error: 0.0050 - val_loss: 0.0083 - val_mean_squared_error: 0.0083\n",
            "Model - CoordConv , epoch 93 / 100.\n",
            "Test RMSE - 0.09795843997221411, MAPE - inf, MAE - 0.024226012823141376\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.0083 - val_mean_squared_error: 0.0083\n",
            "Model - CoordConv , epoch 94 / 100.\n",
            "Test RMSE - 0.09890419513770493, MAPE - inf, MAE - 0.024104357044683583\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.0084 - val_mean_squared_error: 0.0084\n",
            "Model - CoordConv , epoch 95 / 100.\n",
            "Test RMSE - 0.09673331008299801, MAPE - inf, MAE - 0.024157919690309428\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0049 - mean_squared_error: 0.0049 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 96 / 100.\n",
            "Test RMSE - 0.09755287435771487, MAPE - inf, MAE - 0.02394914013891182\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0048 - mean_squared_error: 0.0048 - val_loss: 0.0083 - val_mean_squared_error: 0.0083\n",
            "Model - CoordConv , epoch 97 / 100.\n",
            "Test RMSE - 0.09887902717947693, MAPE - inf, MAE - 0.023954527087717652\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0046 - mean_squared_error: 0.0046 - val_loss: 0.0081 - val_mean_squared_error: 0.0081\n",
            "Model - CoordConv , epoch 98 / 100.\n",
            "Test RMSE - 0.09761699300227879, MAPE - inf, MAE - 0.023700933221263078\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0045 - mean_squared_error: 0.0045 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 99 / 100.\n",
            "Test RMSE - 0.09711077157273834, MAPE - inf, MAE - 0.023714039204414217\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0045 - mean_squared_error: 0.0045 - val_loss: 0.0082 - val_mean_squared_error: 0.0082\n",
            "Model - CoordConv , epoch 100 / 100.\n",
            "Test RMSE - 0.09729468388088482, MAPE - inf, MAE - 0.024161716999943204\n",
            "Skipping RandomConv \n",
            "Starting 3 / 5 experiments batch\n",
            "Skipping CNN \n",
            "Skipping LI CNN \n",
            "Model: \"model_13\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_13 (InputLayer)        (None, 30, 30, 27)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_37 (Conv2D)           (None, 30, 30, 21)        226821    \n",
            "_________________________________________________________________\n",
            "conv2d_38 (Conv2D)           (None, 30, 30, 10)        21010     \n",
            "_________________________________________________________________\n",
            "conv2d_39 (Conv2D)           (None, 30, 30, 1)         11        \n",
            "=================================================================\n",
            "Total params: 247,842\n",
            "Trainable params: 247,842\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Training CoordConv model, with 247842 learnable parameters\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0704 - mean_squared_error: 0.0704 - val_loss: 0.0339 - val_mean_squared_error: 0.0339\n",
            "Model - CoordConv , epoch 1 / 100.\n",
            "Test RMSE - 0.18268908982234552, MAPE - inf, MAE - 0.05608235904414857\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0327 - mean_squared_error: 0.0327 - val_loss: 0.0316 - val_mean_squared_error: 0.0316\n",
            "Model - CoordConv , epoch 2 / 100.\n",
            "Test RMSE - 0.17421198125653956, MAPE - inf, MAE - 0.0738393038927235\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0310 - mean_squared_error: 0.0310 - val_loss: 0.0302 - val_mean_squared_error: 0.0302\n",
            "Model - CoordConv , epoch 3 / 100.\n",
            "Test RMSE - 0.1708474978996395, MAPE - inf, MAE - 0.07760414787349106\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0296 - mean_squared_error: 0.0296 - val_loss: 0.0288 - val_mean_squared_error: 0.0288\n",
            "Model - CoordConv , epoch 4 / 100.\n",
            "Test RMSE - 0.16523680335429264, MAPE - inf, MAE - 0.06859670326908497\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0281 - mean_squared_error: 0.0281 - val_loss: 0.0276 - val_mean_squared_error: 0.0276\n",
            "Model - CoordConv , epoch 5 / 100.\n",
            "Test RMSE - 0.1612866532831082, MAPE - inf, MAE - 0.0678320119153511\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0266 - mean_squared_error: 0.0266 - val_loss: 0.0262 - val_mean_squared_error: 0.0262\n",
            "Model - CoordConv , epoch 6 / 100.\n",
            "Test RMSE - 0.15811701360932928, MAPE - inf, MAE - 0.06189151048458222\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0250 - mean_squared_error: 0.0250 - val_loss: 0.0245 - val_mean_squared_error: 0.0245\n",
            "Model - CoordConv , epoch 7 / 100.\n",
            "Test RMSE - 0.15485289796752147, MAPE - inf, MAE - 0.06198497016663516\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0236 - mean_squared_error: 0.0236 - val_loss: 0.0274 - val_mean_squared_error: 0.0274\n",
            "Model - CoordConv , epoch 8 / 100.\n",
            "Test RMSE - 0.1673335019703634, MAPE - inf, MAE - 0.049330275151273396\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0256 - mean_squared_error: 0.0256 - val_loss: 0.0245 - val_mean_squared_error: 0.0245\n",
            "Model - CoordConv , epoch 9 / 100.\n",
            "Test RMSE - 0.15638244269021132, MAPE - inf, MAE - 0.05283076518998138\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0226 - mean_squared_error: 0.0226 - val_loss: 0.0228 - val_mean_squared_error: 0.0228\n",
            "Model - CoordConv , epoch 10 / 100.\n",
            "Test RMSE - 0.14971682062906433, MAPE - inf, MAE - 0.05848467962450044\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0211 - mean_squared_error: 0.0211 - val_loss: 0.0217 - val_mean_squared_error: 0.0217\n",
            "Model - CoordConv , epoch 11 / 100.\n",
            "Test RMSE - 0.14810833773374907, MAPE - inf, MAE - 0.04951075629391947\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0198 - mean_squared_error: 0.0198 - val_loss: 0.0203 - val_mean_squared_error: 0.0203\n",
            "Model - CoordConv , epoch 12 / 100.\n",
            "Test RMSE - 0.14408224921057358, MAPE - inf, MAE - 0.051759490045498305\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0187 - mean_squared_error: 0.0187 - val_loss: 0.0196 - val_mean_squared_error: 0.0196\n",
            "Model - CoordConv , epoch 13 / 100.\n",
            "Test RMSE - 0.14228169670577334, MAPE - inf, MAE - 0.049334172580419784\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0179 - mean_squared_error: 0.0179 - val_loss: 0.0192 - val_mean_squared_error: 0.0192\n",
            "Model - CoordConv , epoch 14 / 100.\n",
            "Test RMSE - 0.1405833573261484, MAPE - inf, MAE - 0.04942972944363764\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0176 - mean_squared_error: 0.0176 - val_loss: 0.0204 - val_mean_squared_error: 0.0204\n",
            "Model - CoordConv , epoch 15 / 100.\n",
            "Test RMSE - 0.14576634911232791, MAPE - inf, MAE - 0.044571226010893854\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0181 - mean_squared_error: 0.0181 - val_loss: 0.0184 - val_mean_squared_error: 0.0184\n",
            "Model - CoordConv , epoch 16 / 100.\n",
            "Test RMSE - 0.1387660576026745, MAPE - inf, MAE - 0.04551348373747994\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0165 - mean_squared_error: 0.0165 - val_loss: 0.0181 - val_mean_squared_error: 0.0181\n",
            "Model - CoordConv , epoch 17 / 100.\n",
            "Test RMSE - 0.1368813029855124, MAPE - inf, MAE - 0.043489777486561444\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0162 - mean_squared_error: 0.0162 - val_loss: 0.0204 - val_mean_squared_error: 0.0204\n",
            "Model - CoordConv , epoch 18 / 100.\n",
            "Test RMSE - 0.1418220284538039, MAPE - inf, MAE - 0.0525187004451612\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0206 - mean_squared_error: 0.0206 - val_loss: 0.0202 - val_mean_squared_error: 0.0202\n",
            "Model - CoordConv , epoch 19 / 100.\n",
            "Test RMSE - 0.14641319759755875, MAPE - inf, MAE - 0.04449938561315622\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0184 - mean_squared_error: 0.0184 - val_loss: 0.0188 - val_mean_squared_error: 0.0188\n",
            "Model - CoordConv , epoch 20 / 100.\n",
            "Test RMSE - 0.13884918869340854, MAPE - inf, MAE - 0.051949258893988576\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0168 - mean_squared_error: 0.0168 - val_loss: 0.0179 - val_mean_squared_error: 0.0179\n",
            "Model - CoordConv , epoch 21 / 100.\n",
            "Test RMSE - 0.13854360139587013, MAPE - inf, MAE - 0.04147420272082657\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0158 - mean_squared_error: 0.0158 - val_loss: 0.0171 - val_mean_squared_error: 0.0171\n",
            "Model - CoordConv , epoch 22 / 100.\n",
            "Test RMSE - 0.13392597027165765, MAPE - inf, MAE - 0.04300024332487567\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0154 - mean_squared_error: 0.0154 - val_loss: 0.0170 - val_mean_squared_error: 0.0170\n",
            "Model - CoordConv , epoch 23 / 100.\n",
            "Test RMSE - 0.13250312950254217, MAPE - inf, MAE - 0.04383180469844444\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0149 - mean_squared_error: 0.0149 - val_loss: 0.0166 - val_mean_squared_error: 0.0166\n",
            "Model - CoordConv , epoch 24 / 100.\n",
            "Test RMSE - 0.13204292005311588, MAPE - inf, MAE - 0.04239693910808194\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0145 - mean_squared_error: 0.0145 - val_loss: 0.0163 - val_mean_squared_error: 0.0163\n",
            "Model - CoordConv , epoch 25 / 100.\n",
            "Test RMSE - 0.13114997271822143, MAPE - inf, MAE - 0.04281944422110968\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0142 - mean_squared_error: 0.0142 - val_loss: 0.0163 - val_mean_squared_error: 0.0163\n",
            "Model - CoordConv , epoch 26 / 100.\n",
            "Test RMSE - 0.1307921678521399, MAPE - inf, MAE - 0.04203676223644611\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0141 - mean_squared_error: 0.0141 - val_loss: 0.0159 - val_mean_squared_error: 0.0159\n",
            "Model - CoordConv , epoch 27 / 100.\n",
            "Test RMSE - 0.13011627957093447, MAPE - inf, MAE - 0.04020847253654807\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0158 - val_mean_squared_error: 0.0158\n",
            "Model - CoordConv , epoch 28 / 100.\n",
            "Test RMSE - 0.13082224152310998, MAPE - inf, MAE - 0.04063453515461852\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0137 - mean_squared_error: 0.0137 - val_loss: 0.0155 - val_mean_squared_error: 0.0155\n",
            "Model - CoordConv , epoch 29 / 100.\n",
            "Test RMSE - 0.12868679787518675, MAPE - inf, MAE - 0.04007430713841586\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0159 - val_mean_squared_error: 0.0159\n",
            "Model - CoordConv , epoch 30 / 100.\n",
            "Test RMSE - 0.13169769494050726, MAPE - inf, MAE - 0.03739768850792914\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0142 - mean_squared_error: 0.0142 - val_loss: 0.0157 - val_mean_squared_error: 0.0157\n",
            "Model - CoordConv , epoch 31 / 100.\n",
            "Test RMSE - 0.13026256634284394, MAPE - inf, MAE - 0.04144339814054166\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0136 - mean_squared_error: 0.0136 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 32 / 100.\n",
            "Test RMSE - 0.1276456473207794, MAPE - inf, MAE - 0.040261946182406944\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0151 - val_mean_squared_error: 0.0151\n",
            "Model - CoordConv , epoch 33 / 100.\n",
            "Test RMSE - 0.12918213885609606, MAPE - inf, MAE - 0.03791441468326224\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0130 - mean_squared_error: 0.0130 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
            "Model - CoordConv , epoch 34 / 100.\n",
            "Test RMSE - 0.12582692571776577, MAPE - inf, MAE - 0.03856870943585568\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0128 - mean_squared_error: 0.0128 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
            "Model - CoordConv , epoch 35 / 100.\n",
            "Test RMSE - 0.12640492470535727, MAPE - inf, MAE - 0.03741676051296305\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0133 - mean_squared_error: 0.0133 - val_loss: 0.0172 - val_mean_squared_error: 0.0172\n",
            "Model - CoordConv , epoch 36 / 100.\n",
            "Test RMSE - 0.13675428375211315, MAPE - inf, MAE - 0.03666856935886705\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0144 - mean_squared_error: 0.0144 - val_loss: 0.0151 - val_mean_squared_error: 0.0151\n",
            "Model - CoordConv , epoch 37 / 100.\n",
            "Test RMSE - 0.12753191359090357, MAPE - inf, MAE - 0.0378419660670725\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0129 - mean_squared_error: 0.0129 - val_loss: 0.0146 - val_mean_squared_error: 0.0146\n",
            "Model - CoordConv , epoch 38 / 100.\n",
            "Test RMSE - 0.12546345603489836, MAPE - inf, MAE - 0.038251806880558174\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
            "Model - CoordConv , epoch 39 / 100.\n",
            "Test RMSE - 0.12456274796829403, MAPE - inf, MAE - 0.037381890901250624\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0123 - mean_squared_error: 0.0123 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
            "Model - CoordConv , epoch 40 / 100.\n",
            "Test RMSE - 0.12637153880280338, MAPE - inf, MAE - 0.03746851868831065\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0121 - mean_squared_error: 0.0121 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
            "Model - CoordConv , epoch 41 / 100.\n",
            "Test RMSE - 0.12397539778196075, MAPE - inf, MAE - 0.036287977913958144\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0119 - mean_squared_error: 0.0119 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
            "Model - CoordConv , epoch 42 / 100.\n",
            "Test RMSE - 0.12349773581926698, MAPE - inf, MAE - 0.037518640880108096\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
            "Model - CoordConv , epoch 43 / 100.\n",
            "Test RMSE - 0.12731584940969243, MAPE - inf, MAE - 0.03535773213679858\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0143 - mean_squared_error: 0.0143 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
            "Model - CoordConv , epoch 44 / 100.\n",
            "Test RMSE - 0.1275069829155726, MAPE - inf, MAE - 0.04020814716192662\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
            "Model - CoordConv , epoch 45 / 100.\n",
            "Test RMSE - 0.1266684899220271, MAPE - inf, MAE - 0.04048130287660179\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0129 - mean_squared_error: 0.0129 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
            "Model - CoordConv , epoch 46 / 100.\n",
            "Test RMSE - 0.12416760751725549, MAPE - inf, MAE - 0.036029419292116886\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0120 - mean_squared_error: 0.0120 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
            "Model - CoordConv , epoch 47 / 100.\n",
            "Test RMSE - 0.12813638049064086, MAPE - inf, MAE - 0.035462699108454225\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
            "Model - CoordConv , epoch 48 / 100.\n",
            "Test RMSE - 0.12448341880926275, MAPE - inf, MAE - 0.03540803603674946\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0122 - mean_squared_error: 0.0122 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
            "Model - CoordConv , epoch 49 / 100.\n",
            "Test RMSE - 0.12303821705649264, MAPE - inf, MAE - 0.03694184896472885\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0120 - mean_squared_error: 0.0120 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
            "Model - CoordConv , epoch 50 / 100.\n",
            "Test RMSE - 0.12265889144145366, MAPE - inf, MAE - 0.03691341516210826\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0115 - mean_squared_error: 0.0115 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
            "Model - CoordConv , epoch 51 / 100.\n",
            "Test RMSE - 0.12071737144041053, MAPE - inf, MAE - 0.03622959021965325\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0111 - mean_squared_error: 0.0111 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
            "Model - CoordConv , epoch 52 / 100.\n",
            "Test RMSE - 0.12218967250049584, MAPE - inf, MAE - 0.0352918077573971\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0112 - mean_squared_error: 0.0112 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
            "Model - CoordConv , epoch 53 / 100.\n",
            "Test RMSE - 0.12337281181197382, MAPE - inf, MAE - 0.03460092069020917\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0109 - mean_squared_error: 0.0109 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
            "Model - CoordConv , epoch 54 / 100.\n",
            "Test RMSE - 0.122863768757488, MAPE - inf, MAE - 0.035137597297495826\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0109 - mean_squared_error: 0.0109 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
            "Model - CoordConv , epoch 55 / 100.\n",
            "Test RMSE - 0.12294457442850668, MAPE - inf, MAE - 0.0342520299458163\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0108 - mean_squared_error: 0.0108 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
            "Model - CoordConv , epoch 56 / 100.\n",
            "Test RMSE - 0.11999169877986816, MAPE - inf, MAE - 0.03517340445380945\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0109 - mean_squared_error: 0.0109 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
            "Model - CoordConv , epoch 57 / 100.\n",
            "Test RMSE - 0.12711504109030133, MAPE - inf, MAE - 0.03387937088354277\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0144 - mean_squared_error: 0.0144 - val_loss: 0.0186 - val_mean_squared_error: 0.0186\n",
            "Model - CoordConv , epoch 58 / 100.\n",
            "Test RMSE - 0.14206431032086103, MAPE - inf, MAE - 0.036316049248389035\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0147 - mean_squared_error: 0.0147 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
            "Model - CoordConv , epoch 59 / 100.\n",
            "Test RMSE - 0.1300085473366217, MAPE - inf, MAE - 0.03871036327512299\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
            "Model - CoordConv , epoch 60 / 100.\n",
            "Test RMSE - 0.1269759954700756, MAPE - inf, MAE - 0.03588559203749574\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0115 - mean_squared_error: 0.0115 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
            "Model - CoordConv , epoch 61 / 100.\n",
            "Test RMSE - 0.12280722917526479, MAPE - inf, MAE - 0.03522901126686141\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0110 - mean_squared_error: 0.0110 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
            "Model - CoordConv , epoch 62 / 100.\n",
            "Test RMSE - 0.12069632459089862, MAPE - inf, MAE - 0.034757076052158005\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0106 - mean_squared_error: 0.0106 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
            "Model - CoordConv , epoch 63 / 100.\n",
            "Test RMSE - 0.12209586925192072, MAPE - inf, MAE - 0.034198257414871154\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
            "Model - CoordConv , epoch 64 / 100.\n",
            "Test RMSE - 0.12024524802226663, MAPE - inf, MAE - 0.03386595501513328\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
            "Model - CoordConv , epoch 65 / 100.\n",
            "Test RMSE - 0.12043334884105614, MAPE - inf, MAE - 0.03513331508693014\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
            "Model - CoordConv , epoch 66 / 100.\n",
            "Test RMSE - 0.1212261555939002, MAPE - inf, MAE - 0.03305487559158275\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
            "Model - CoordConv , epoch 67 / 100.\n",
            "Test RMSE - 0.12073898435395164, MAPE - inf, MAE - 0.03646011876847688\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0104 - mean_squared_error: 0.0104 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 68 / 100.\n",
            "Test RMSE - 0.11983744942536305, MAPE - inf, MAE - 0.03395888136150548\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0100 - mean_squared_error: 0.0100 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 69 / 100.\n",
            "Test RMSE - 0.11899244883918579, MAPE - inf, MAE - 0.033972095914547695\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0099 - mean_squared_error: 0.0099 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
            "Model - CoordConv , epoch 70 / 100.\n",
            "Test RMSE - 0.11764193908796738, MAPE - inf, MAE - 0.03402039268986417\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0100 - mean_squared_error: 0.0100 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 71 / 100.\n",
            "Test RMSE - 0.11849820325366614, MAPE - inf, MAE - 0.03276933890682024\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0099 - mean_squared_error: 0.0099 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
            "Model - CoordConv , epoch 72 / 100.\n",
            "Test RMSE - 0.12001985925636527, MAPE - inf, MAE - 0.03231919631315516\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0103 - mean_squared_error: 0.0103 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
            "Model - CoordConv , epoch 73 / 100.\n",
            "Test RMSE - 0.1210968784851676, MAPE - inf, MAE - 0.03587647050753499\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0101 - mean_squared_error: 0.0101 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
            "Model - CoordConv , epoch 74 / 100.\n",
            "Test RMSE - 0.11948803318516074, MAPE - inf, MAE - 0.032639571658416196\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0097 - mean_squared_error: 0.0097 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 75 / 100.\n",
            "Test RMSE - 0.11643742736187289, MAPE - inf, MAE - 0.03289552125331838\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0095 - mean_squared_error: 0.0095 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
            "Model - CoordConv , epoch 76 / 100.\n",
            "Test RMSE - 0.11647232034008281, MAPE - inf, MAE - 0.033427278761606956\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0098 - mean_squared_error: 0.0098 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 77 / 100.\n",
            "Test RMSE - 0.11651824745204244, MAPE - inf, MAE - 0.0326294835986858\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0121 - mean_squared_error: 0.0121 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
            "Model - CoordConv , epoch 78 / 100.\n",
            "Test RMSE - 0.12105649150419233, MAPE - inf, MAE - 0.03279540373712322\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0118 - mean_squared_error: 0.0118 - val_loss: 0.0146 - val_mean_squared_error: 0.0146\n",
            "Model - CoordConv , epoch 79 / 100.\n",
            "Test RMSE - 0.13103438729659334, MAPE - inf, MAE - 0.03408250910745955\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0132 - mean_squared_error: 0.0132 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
            "Model - CoordConv , epoch 80 / 100.\n",
            "Test RMSE - 0.11977732366834164, MAPE - inf, MAE - 0.03792842352204988\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0112 - mean_squared_error: 0.0112 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
            "Model - CoordConv , epoch 81 / 100.\n",
            "Test RMSE - 0.1176548580346459, MAPE - inf, MAE - 0.03558878770156602\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0101 - mean_squared_error: 0.0101 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
            "Model - CoordConv , epoch 82 / 100.\n",
            "Test RMSE - 0.11686812351530065, MAPE - inf, MAE - 0.03285109200422682\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0097 - mean_squared_error: 0.0097 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 83 / 100.\n",
            "Test RMSE - 0.11650656763372068, MAPE - inf, MAE - 0.03261627652527367\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0096 - mean_squared_error: 0.0096 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
            "Model - CoordConv , epoch 84 / 100.\n",
            "Test RMSE - 0.11911122609012842, MAPE - inf, MAE - 0.03186451810888495\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0095 - mean_squared_error: 0.0095 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
            "Model - CoordConv , epoch 85 / 100.\n",
            "Test RMSE - 0.12000968459180913, MAPE - inf, MAE - 0.03162004546896852\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0094 - mean_squared_error: 0.0094 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
            "Model - CoordConv , epoch 86 / 100.\n",
            "Test RMSE - 0.11565381554097216, MAPE - inf, MAE - 0.03253681716240485\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
            "Model - CoordConv , epoch 87 / 100.\n",
            "Test RMSE - 0.11471356575952768, MAPE - inf, MAE - 0.03140528513627863\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 88 / 100.\n",
            "Test RMSE - 0.1154989759410878, MAPE - inf, MAE - 0.031740243466196284\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 89 / 100.\n",
            "Test RMSE - 0.11550795850613652, MAPE - inf, MAE - 0.031817042453816224\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 90 / 100.\n",
            "Test RMSE - 0.11494314914902619, MAPE - inf, MAE - 0.03133533675453152\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0089 - mean_squared_error: 0.0089 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
            "Model - CoordConv , epoch 91 / 100.\n",
            "Test RMSE - 0.12225643739254895, MAPE - inf, MAE - 0.031146241674684348\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0111 - mean_squared_error: 0.0111 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
            "Model - CoordConv , epoch 92 / 100.\n",
            "Test RMSE - 0.12399950551852287, MAPE - inf, MAE - 0.038214932993828984\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0108 - mean_squared_error: 0.0108 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 93 / 100.\n",
            "Test RMSE - 0.11521843179120493, MAPE - inf, MAE - 0.03308865240398945\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0093 - mean_squared_error: 0.0093 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
            "Model - CoordConv , epoch 94 / 100.\n",
            "Test RMSE - 0.11578920346723398, MAPE - inf, MAE - 0.032056361908720396\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 95 / 100.\n",
            "Test RMSE - 0.1152661077992984, MAPE - inf, MAE - 0.031205455648581966\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0090 - mean_squared_error: 0.0090 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
            "Model - CoordConv , epoch 96 / 100.\n",
            "Test RMSE - 0.11580479598081282, MAPE - inf, MAE - 0.031571945064068636\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
            "Model - CoordConv , epoch 97 / 100.\n",
            "Test RMSE - 0.11600648830810908, MAPE - inf, MAE - 0.03147514750791216\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
            "Model - CoordConv , epoch 98 / 100.\n",
            "Test RMSE - 0.11408014795535089, MAPE - inf, MAE - 0.03152006819820608\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0086 - mean_squared_error: 0.0086 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 99 / 100.\n",
            "Test RMSE - 0.11740018465393531, MAPE - inf, MAE - 0.030604486170238122\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0088 - mean_squared_error: 0.0088 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
            "Model - CoordConv , epoch 100 / 100.\n",
            "Test RMSE - 0.11331533147027471, MAPE - inf, MAE - 0.03176514759399792\n",
            "Skipping RandomConv \n",
            "Starting 4 / 5 experiments batch\n",
            "Skipping CNN \n",
            "Skipping LI CNN \n",
            "Model: \"model_14\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_14 (InputLayer)        (None, 30, 30, 27)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_40 (Conv2D)           (None, 30, 30, 21)        226821    \n",
            "_________________________________________________________________\n",
            "conv2d_41 (Conv2D)           (None, 30, 30, 10)        21010     \n",
            "_________________________________________________________________\n",
            "conv2d_42 (Conv2D)           (None, 30, 30, 1)         11        \n",
            "=================================================================\n",
            "Total params: 247,842\n",
            "Trainable params: 247,842\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Training CoordConv model, with 247842 learnable parameters\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.1591 - mean_squared_error: 0.1591 - val_loss: 0.0346 - val_mean_squared_error: 0.0346\n",
            "Model - CoordConv , epoch 1 / 100.\n",
            "Test RMSE - 0.1860423938810153, MAPE - inf, MAE - 0.04269307915796142\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0346 - mean_squared_error: 0.0346 - val_loss: 0.0346 - val_mean_squared_error: 0.0346\n",
            "Model - CoordConv , epoch 2 / 100.\n",
            "Test RMSE - 0.18605114910238738, MAPE - inf, MAE - 0.04268736690393601\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0346 - mean_squared_error: 0.0346 - val_loss: 0.0345 - val_mean_squared_error: 0.0345\n",
            "Model - CoordConv , epoch 3 / 100.\n",
            "Test RMSE - 0.18580061359115896, MAPE - inf, MAE - 0.04157997196683871\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0345 - mean_squared_error: 0.0345 - val_loss: 0.0344 - val_mean_squared_error: 0.0344\n",
            "Model - CoordConv , epoch 4 / 100.\n",
            "Test RMSE - 0.18542383850591437, MAPE - inf, MAE - 0.041952375526293925\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0343 - mean_squared_error: 0.0343 - val_loss: 0.0342 - val_mean_squared_error: 0.0342\n",
            "Model - CoordConv , epoch 5 / 100.\n",
            "Test RMSE - 0.18499160393421854, MAPE - inf, MAE - 0.04372161056947504\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0342 - mean_squared_error: 0.0342 - val_loss: 0.0341 - val_mean_squared_error: 0.0341\n",
            "Model - CoordConv , epoch 6 / 100.\n",
            "Test RMSE - 0.18454453589558212, MAPE - inf, MAE - 0.04566444707367727\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0340 - mean_squared_error: 0.0340 - val_loss: 0.0339 - val_mean_squared_error: 0.0339\n",
            "Model - CoordConv , epoch 7 / 100.\n",
            "Test RMSE - 0.18410683234271136, MAPE - inf, MAE - 0.04769538071439629\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0338 - mean_squared_error: 0.0338 - val_loss: 0.0337 - val_mean_squared_error: 0.0337\n",
            "Model - CoordConv , epoch 8 / 100.\n",
            "Test RMSE - 0.18369288784584642, MAPE - inf, MAE - 0.04975785537710231\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0337 - mean_squared_error: 0.0337 - val_loss: 0.0336 - val_mean_squared_error: 0.0336\n",
            "Model - CoordConv , epoch 9 / 100.\n",
            "Test RMSE - 0.18331104858037603, MAPE - inf, MAE - 0.05181214513238374\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0336 - mean_squared_error: 0.0336 - val_loss: 0.0335 - val_mean_squared_error: 0.0335\n",
            "Model - CoordConv , epoch 10 / 100.\n",
            "Test RMSE - 0.18296569275977898, MAPE - inf, MAE - 0.053828488870474596\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0334 - mean_squared_error: 0.0334 - val_loss: 0.0334 - val_mean_squared_error: 0.0334\n",
            "Model - CoordConv , epoch 11 / 100.\n",
            "Test RMSE - 0.18265841365380417, MAPE - inf, MAE - 0.05578517265256137\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0333 - mean_squared_error: 0.0333 - val_loss: 0.0333 - val_mean_squared_error: 0.0333\n",
            "Model - CoordConv , epoch 12 / 100.\n",
            "Test RMSE - 0.1823889295577339, MAPE - inf, MAE - 0.05766578794232831\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0332 - mean_squared_error: 0.0332 - val_loss: 0.0332 - val_mean_squared_error: 0.0332\n",
            "Model - CoordConv , epoch 13 / 100.\n",
            "Test RMSE - 0.18215569427320616, MAPE - inf, MAE - 0.059457946439343046\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0332 - mean_squared_error: 0.0332 - val_loss: 0.0331 - val_mean_squared_error: 0.0331\n",
            "Model - CoordConv , epoch 14 / 100.\n",
            "Test RMSE - 0.18195629965861326, MAPE - inf, MAE - 0.06115243493606187\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0331 - mean_squared_error: 0.0331 - val_loss: 0.0331 - val_mean_squared_error: 0.0331\n",
            "Model - CoordConv , epoch 15 / 100.\n",
            "Test RMSE - 0.1817877441271676, MAPE - inf, MAE - 0.06274242468555352\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0330 - mean_squared_error: 0.0330 - val_loss: 0.0330 - val_mean_squared_error: 0.0330\n",
            "Model - CoordConv , epoch 16 / 100.\n",
            "Test RMSE - 0.18164642102307635, MAPE - inf, MAE - 0.06422102586343889\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0330 - mean_squared_error: 0.0330 - val_loss: 0.0330 - val_mean_squared_error: 0.0330\n",
            "Model - CoordConv , epoch 17 / 100.\n",
            "Test RMSE - 0.18152629694450184, MAPE - inf, MAE - 0.06555983386559995\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0329 - mean_squared_error: 0.0329 - val_loss: 0.0329 - val_mean_squared_error: 0.0329\n",
            "Model - CoordConv , epoch 18 / 100.\n",
            "Test RMSE - 0.18142746172387766, MAPE - inf, MAE - 0.06678968626403575\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0329 - mean_squared_error: 0.0329 - val_loss: 0.0329 - val_mean_squared_error: 0.0329\n",
            "Model - CoordConv , epoch 19 / 100.\n",
            "Test RMSE - 0.1813441075644639, MAPE - inf, MAE - 0.06786598623290697\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0329 - mean_squared_error: 0.0329 - val_loss: 0.0329 - val_mean_squared_error: 0.0329\n",
            "Model - CoordConv , epoch 20 / 100.\n",
            "Test RMSE - 0.18126583596582124, MAPE - inf, MAE - 0.06877375316198198\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0329 - val_mean_squared_error: 0.0329\n",
            "Model - CoordConv , epoch 21 / 100.\n",
            "Test RMSE - 0.18124159079272767, MAPE - inf, MAE - 0.07011196023950737\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 22 / 100.\n",
            "Test RMSE - 0.18120464737420822, MAPE - inf, MAE - 0.07102451702283105\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 23 / 100.\n",
            "Test RMSE - 0.1811750496729225, MAPE - inf, MAE - 0.07180381771465104\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 24 / 100.\n",
            "Test RMSE - 0.18115270026864977, MAPE - inf, MAE - 0.07248629849329566\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 25 / 100.\n",
            "Test RMSE - 0.18113583610312328, MAPE - inf, MAE - 0.07308514714374433\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 26 / 100.\n",
            "Test RMSE - 0.18112315860208383, MAPE - inf, MAE - 0.07361034694334632\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 27 / 100.\n",
            "Test RMSE - 0.18111368376989534, MAPE - inf, MAE - 0.07406995336078268\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 28 / 100.\n",
            "Test RMSE - 0.18110666159592942, MAPE - inf, MAE - 0.07447030797322447\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 29 / 100.\n",
            "Test RMSE - 0.18110148931124923, MAPE - inf, MAE - 0.07481828185578762\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 30 / 100.\n",
            "Test RMSE - 0.18109772102509544, MAPE - inf, MAE - 0.0751186256886303\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 31 / 100.\n",
            "Test RMSE - 0.18109499690035968, MAPE - inf, MAE - 0.07537682684680555\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 32 / 100.\n",
            "Test RMSE - 0.18109304188012987, MAPE - inf, MAE - 0.07559805322365834\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 33 / 100.\n",
            "Test RMSE - 0.1810916536640066, MAPE - inf, MAE - 0.07578630979265752\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 34 / 100.\n",
            "Test RMSE - 0.18109067542235813, MAPE - inf, MAE - 0.07594584833937612\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 35 / 100.\n",
            "Test RMSE - 0.1810899886861849, MAPE - inf, MAE - 0.07608106195159883\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 36 / 100.\n",
            "Test RMSE - 0.18108951464898962, MAPE - inf, MAE - 0.07619413921767056\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 37 / 100.\n",
            "Test RMSE - 0.18108918725684656, MAPE - inf, MAE - 0.07628905125973873\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 38 / 100.\n",
            "Test RMSE - 0.18108896427098986, MAPE - inf, MAE - 0.07636786935584708\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 39 / 100.\n",
            "Test RMSE - 0.18108881361750326, MAPE - inf, MAE - 0.07643293517073112\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 40 / 100.\n",
            "Test RMSE - 0.18108871218749384, MAPE - inf, MAE - 0.07648655520611028\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 41 / 100.\n",
            "Test RMSE - 0.1810886442742566, MAPE - inf, MAE - 0.07653056689452979\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 42 / 100.\n",
            "Test RMSE - 0.1810885994235468, MAPE - inf, MAE - 0.07656619875681912\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 43 / 100.\n",
            "Test RMSE - 0.1810885696988141, MAPE - inf, MAE - 0.07659510491218738\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 44 / 100.\n",
            "Test RMSE - 0.18108854987243458, MAPE - inf, MAE - 0.07661868089804709\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 45 / 100.\n",
            "Test RMSE - 0.18108853690484558, MAPE - inf, MAE - 0.07663748374747768\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 46 / 100.\n",
            "Test RMSE - 0.18108852841601722, MAPE - inf, MAE - 0.07665239442284974\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 47 / 100.\n",
            "Test RMSE - 0.18108852266743714, MAPE - inf, MAE - 0.07666455630079977\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 48 / 100.\n",
            "Test RMSE - 0.18108851886496863, MAPE - inf, MAE - 0.076674185564418\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 49 / 100.\n",
            "Test RMSE - 0.18108851638113527, MAPE - inf, MAE - 0.07668162298516522\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 50 / 100.\n",
            "Test RMSE - 0.1810885147175313, MAPE - inf, MAE - 0.0766874254915061\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 51 / 100.\n",
            "Test RMSE - 0.1810885135794187, MAPE - inf, MAE - 0.07669198296543052\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 52 / 100.\n",
            "Test RMSE - 0.18108851282057511, MAPE - inf, MAE - 0.07669541336191638\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 53 / 100.\n",
            "Test RMSE - 0.18108851237715387, MAPE - inf, MAE - 0.07669761839099669\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 54 / 100.\n",
            "Test RMSE - 0.1810885119853386, MAPE - inf, MAE - 0.07669972840396506\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 55 / 100.\n",
            "Test RMSE - 0.18108851174472507, MAPE - inf, MAE - 0.07670111760504984\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 56 / 100.\n",
            "Test RMSE - 0.18108851158069003, MAPE - inf, MAE - 0.07670211363601628\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 57 / 100.\n",
            "Test RMSE - 0.18108851139709925, MAPE - inf, MAE - 0.07670328331711833\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 58 / 100.\n",
            "Test RMSE - 0.1810885112844748, MAPE - inf, MAE - 0.0767040336167608\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 59 / 100.\n",
            "Test RMSE - 0.18108851128206074, MAPE - inf, MAE - 0.07670404999884907\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 60 / 100.\n",
            "Test RMSE - 0.1810885112029971, MAPE - inf, MAE - 0.07670459388417944\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 61 / 100.\n",
            "Test RMSE - 0.1810885111726129, MAPE - inf, MAE - 0.07670480685132686\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 62 / 100.\n",
            "Test RMSE - 0.18108851110327978, MAPE - inf, MAE - 0.07670530159039243\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 63 / 100.\n",
            "Test RMSE - 0.18108851108250495, MAPE - inf, MAE - 0.07670545230560445\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 64 / 100.\n",
            "Test RMSE - 0.18108851107846027, MAPE - inf, MAE - 0.07670548179336331\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 65 / 100.\n",
            "Test RMSE - 0.18108851113382748, MAPE - inf, MAE - 0.0767050820704097\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 66 / 100.\n",
            "Test RMSE - 0.1810885111352035, MAPE - inf, MAE - 0.07670507224115672\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 67 / 100.\n",
            "Test RMSE - 0.18108851105698862, MAPE - inf, MAE - 0.07670563906141066\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 68 / 100.\n",
            "Test RMSE - 0.18108851102291376, MAPE - inf, MAE - 0.0767058913455699\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 69 / 100.\n",
            "Test RMSE - 0.18108851103436868, MAPE - inf, MAE - 0.07670580615871093\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 70 / 100.\n",
            "Test RMSE - 0.18108851104055873, MAPE - inf, MAE - 0.0767057602888638\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 71 / 100.\n",
            "Test RMSE - 0.18108851107935875, MAPE - inf, MAE - 0.07670547524052801\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 72 / 100.\n",
            "Test RMSE - 0.18108851109558322, MAPE - inf, MAE - 0.07670535728949252\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 73 / 100.\n",
            "Test RMSE - 0.18108851111236263, MAPE - inf, MAE - 0.07670523606203937\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 74 / 100.\n",
            "Test RMSE - 0.18108851111145227, MAPE - inf, MAE - 0.07670524261487467\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 75 / 100.\n",
            "Test RMSE - 0.18108851111008942, MAPE - inf, MAE - 0.07670525244412763\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 76 / 100.\n",
            "Test RMSE - 0.18108851108250495, MAPE - inf, MAE - 0.07670545230560445\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 77 / 100.\n",
            "Test RMSE - 0.18108851106859708, MAPE - inf, MAE - 0.07670555387455169\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 78 / 100.\n",
            "Test RMSE - 0.18108851101150994, MAPE - inf, MAE - 0.07670597653242889\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 79 / 100.\n",
            "Test RMSE - 0.18108851101545248, MAPE - inf, MAE - 0.07670594704467003\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 80 / 100.\n",
            "Test RMSE - 0.18108851100320988, MAPE - inf, MAE - 0.0767060387843643\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 81 / 100.\n",
            "Test RMSE - 0.18108851106010934, MAPE - inf, MAE - 0.0767056161264871\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 82 / 100.\n",
            "Test RMSE - 0.18108851111828325, MAPE - inf, MAE - 0.07670519346860988\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 83 / 100.\n",
            "Test RMSE - 0.18108851123512884, MAPE - inf, MAE - 0.07670437108777903\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 84 / 100.\n",
            "Test RMSE - 0.18108851119877006, MAPE - inf, MAE - 0.0767046233719383\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 85 / 100.\n",
            "Test RMSE - 0.18108851116379282, MAPE - inf, MAE - 0.07670486910326227\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 86 / 100.\n",
            "Test RMSE - 0.18108851116889618, MAPE - inf, MAE - 0.07670483306266808\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 87 / 100.\n",
            "Test RMSE - 0.18108851116843136, MAPE - inf, MAE - 0.07670483633908573\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 88 / 100.\n",
            "Test RMSE - 0.18108851117214844, MAPE - inf, MAE - 0.07670481012774452\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 89 / 100.\n",
            "Test RMSE - 0.18108851114393412, MAPE - inf, MAE - 0.07670500998922133\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 90 / 100.\n",
            "Test RMSE - 0.18108851116611208, MAPE - inf, MAE - 0.07670485272117399\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 91 / 100.\n",
            "Test RMSE - 0.18108851118612931, MAPE - inf, MAE - 0.07670471183521493\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 92 / 100.\n",
            "Test RMSE - 0.18108851116843136, MAPE - inf, MAE - 0.07670483633908573\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 93 / 100.\n",
            "Test RMSE - 0.18108851112695942, MAPE - inf, MAE - 0.07670513121667448\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 94 / 100.\n",
            "Test RMSE - 0.18108851107307608, MAPE - inf, MAE - 0.07670552111037517\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 95 / 100.\n",
            "Test RMSE - 0.18108851103127985, MAPE - inf, MAE - 0.07670582909363452\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 96 / 100.\n",
            "Test RMSE - 0.18108851101939935, MAPE - inf, MAE - 0.07670591755691114\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 97 / 100.\n",
            "Test RMSE - 0.18108851100845022, MAPE - inf, MAE - 0.07670599946735246\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 98 / 100.\n",
            "Test RMSE - 0.1810885110290754, MAPE - inf, MAE - 0.07670584547572279\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 99 / 100.\n",
            "Test RMSE - 0.18108851099972334, MAPE - inf, MAE - 0.07670606499570552\n",
            "Train on 1500 samples, validate on 500 samples\n",
            "Epoch 1/1\n",
            " - 1s - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
            "Model - CoordConv , epoch 100 / 100.\n",
            "Test RMSE - 0.18108851099146256, MAPE - inf, MAE - 0.07670612724764092\n",
            "Skipping RandomConv \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cdnLLnCznQmU",
        "colab_type": "text"
      },
      "source": [
        "# Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "360Rvf2dzcca",
        "colab_type": "code",
        "outputId": "f6af53dc-0201-4b67-eaa9-9e2416789cda",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 609
        }
      },
      "source": [
        "# import importlib\n",
        "# importlib.reload(matplotlib)\n",
        "import glob\n",
        "import os\n",
        "import pickle\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "import constants\n",
        "import seaborn as sns\n",
        "from itertools import cycle\n",
        "sns.set_palette(\"cubehelix\")\n",
        "\n",
        "cycol = cycle(sns.color_palette())\n",
        "\n",
        "def read_folder(folder_path):\n",
        "    os.chdir(folder_path)\n",
        "\n",
        "    results = {}\n",
        "    for file in reversed(sorted(glob.glob(\"*\"))):\n",
        "        res = pickle.load(open(file, \"rb\"))\n",
        "        if \"learnable_maps_count\" in res:\n",
        "            res[\"name\"] = \"{} #learn maps-{}\".format(res[\"name\"], res[\"learnable_maps_count\"])\n",
        "        if res[\"name\"] not in results:\n",
        "\n",
        "            results[res[\"name\"]] = []\n",
        "\n",
        "        results[res[\"name\"]].append(res)\n",
        "\n",
        "    return results\n",
        "folder = \"{}/results/synthetic/bouncing_ball/\".format(save_results_path)\n",
        "results = read_folder(folder)\n",
        "colors = [next(cycol) for _ in range(0, 10)]\n",
        "colors[0] = \"black\"\n",
        "colors[1] = \"green\"\n",
        "iter = 0\n",
        "\n",
        "for key,val in results.items():\n",
        "    # aggregate\n",
        "    train_errs,valid_errs,test_mape_errs, test_mse_errs = None,None,None,None\n",
        "    learnable_params = []\n",
        "\n",
        "    abstract_name = None\n",
        "    for run in val:\n",
        "        if len(run[\"train_error\"]) != 100:\n",
        "            continue\n",
        "        abstract_name = run[\"name\"]\n",
        "        learnable_params.append(run[\"learnable_parameters\"])\n",
        "        valid_errs = np.array(run[\"valid_error\"])[:,0][None].T if valid_errs is None else np.hstack((valid_errs,np.array(run[\"valid_error\"])[:,0][None].T ))\n",
        "        test_mse_errs = np.array(run[\"test_error\"])[:,0][None].T if test_mse_errs is None else np.hstack((test_mse_errs,np.array(run[\"test_error\"])[:,0][None].T))\n",
        "    learnable_params = np.mean(learnable_params).astype(int)\n",
        "    print(\" {} has {} learnable params.\".format(key, learnable_params))\n",
        "    def get_batch_stats(batch):\n",
        "        mean_batch = np.mean(batch, axis=-1)\n",
        "\n",
        "        min_batch = np.min(batch, axis=-1)\n",
        "        max_batch = np.max(batch,axis=-1)\n",
        "\n",
        "        return mean_batch, min_batch, max_batch\n",
        "\n",
        "    plt.figure(1)\n",
        "    mean_test, min_test, max_test = get_batch_stats(test_mse_errs)\n",
        "\n",
        "    mean_valid, min_valid, max_valid = get_batch_stats(valid_errs)\n",
        "\n",
        "    # test error\n",
        "    plt.fill_between(np.linspace(1,mean_test.shape[0],mean_test.shape[0],endpoint=True),min_test,max_test,color=colors[iter%len(colors)],alpha=0.1)\n",
        "    plt.semilogy(np.linspace(1,mean_test.shape[0],mean_test.shape[0],endpoint=True),mean_test,  color=colors[iter%len(colors)], label=abstract_name,basey = 2)\n",
        "    # validation error\n",
        "    plt.figure(2)\n",
        "    plt.fill_between(np.linspace(1, mean_valid.shape[0], mean_valid.shape[0], endpoint=True), min_valid, max_valid,\n",
        "                     color=colors[iter % len(colors)], alpha=0.1)\n",
        "    plt.semilogy(np.linspace(1, mean_valid.shape[0], mean_valid.shape[0], endpoint=True), mean_valid, color=colors[iter % len(colors)],label=abstract_name,basey = 2)\n",
        "    iter+=1\n",
        "plt.figure(1)\n",
        "plt.grid(linestyle=\":\")\n",
        "\n",
        "ax = plt.gca()\n",
        "handles, labels = ax.get_legend_handles_labels()\n",
        "labels, handles = zip(*sorted(zip(labels, handles), key=lambda t: len(t[0])))\n",
        "ax.legend(handles, labels)\n",
        "\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"RMSE\")\n",
        "\n",
        "plt.gca().spines[\"top\"].set_visible(False)\n",
        "plt.gca().spines[\"right\"].set_visible(False)\n",
        "\n",
        "plt.figure(2)\n",
        "plt.grid(linestyle=\":\")\n",
        "plt.legend()\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"RMSE\")\n",
        "\n",
        "\n",
        "plt.gca().spines[\"top\"].set_visible(False)\n",
        "plt.gca().spines[\"right\"].set_visible(False)\n",
        "\n",
        "display(plt.figure(2))\n",
        "display(plt.figure(1))\n",
        "\n",
        "plt.close()\n",
        "plt.figure(2)\n",
        "plt.close()"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            " RandomConv has 247842 learnable params.\n",
            " LI CNN has 237841 learnable params.\n",
            " CoordConv has 247842 learnable params.\n",
            " CNN has 242043 learnable params.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEGCAYAAABlxeIAAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOydd5xcZb3/36fPzPaW3fSEFNILgSTUhNA74dK9ChdB0QuCShHQi1fx4hVUEPGnghQLRRC4KlJEkVAChIQ0SCebns2W2emnn98fszPZzc6WJDu7m5z9vF55ZXbPmec8nz0z5/t8y/P5Cp7nMYABDGAAAxiA2NcTGMAABjCAAfQPDBiEAQxgAAMYADBgEAYwgAEMYAAtGDAIAxjAAAYwAGDAIAxgAAMYwABaIPf1BA4UZ555pvfqq6/29TQGMIABDOBQg9DRgUPWQ2hoaNjv92zatCkPM+nf8CNn8CdvP3IGf/LOF+dD1iAcCMrLy/t6Cr0OP3IGf/L2I2fwJ+98cfaVQUgmk309hV6HHzmDP3n7kTP4k3e+OPvKIIiir+gC/uQM/uTtR87gT9754nzIJpUPBIqi9PUUeh1+5Az+5N2XnC3LYvv27ei63uvXdhyH5ubmXr9uX6I7nAOBAMOGDduvz4WvDEI8HqeysrKvp9Gr8CNn8CfvvuS8fft2ioqKGDVqFILQYRFLXmAYBpqm9eo1+xpdcfY8j8bGRrZv387o0aO7Pa6vfC2/PSDAn5zBn7z7krOu61RUVPS6MQCQZV+ta4GuOQuCQEVFxX57bL4yCNu3b+/rKfQ6/MgZ/Mm7rzn3hTEAME2zT67bl+gO5wO5H74yCGPHju3rKfQ6/MgZ/Mnbj5whHSv3G/LF2VcG4ZNPPgHAcSziiWYs+/BfWWQ4+w1+5O1HzgCpVCr7evfu3Vx++eWMGTOGWbNmcfbZZ7N+/XoEQeChhx7KnnfDDTfwxBNPAHD11VczdOhQDMMA0pteR40a1ZsU9hutOfckfGUQpk+fDoDt2ERj9TSGd9IY3klKj+O6bh/PLj/IcPYb/Mjbj5wBQqEQkE6kLly4kPnz57Np0yaWLl3KvffeS11dHYMGDeLBBx/sMNQiSRKPPfZYb077oJDh3NPwlUFYunQpnuex5ONXqF2/FklIJ2YisQYawjtIJCO4rtPHs+xZLF26tK+n0CfwI28/cgZIJBIAvPnmmyiKwvXXX589Nn36dIYPH05VVRWnnHIKTz75ZM4xbr75Zn76059i23avzPlgkeHc0/BVen7WrFnsadjC03/9AcMrJyOpBYwaM46AFsJ1HeLJMPFkMwXBEoKBQiTp0P/zzJo1q6+n0CfwI+/+wnnLy38jsWtXj45ZMHgwI885O/exggIAVq9e3enf4Pbbb+ess87immuuaXdsxIgRnHDCCfzud7/jvPPO65lJ5xEZzj0N33kIgypHcvTUs9lav5rNu1ayYeNaYrEYruuhqSEURSOpR6lv2k401ohtW3097YOCX1eNfuTtR87Q/dXyEUccwZw5c3jqqadyHr/jjju47777Donw8YCH0APIrB7mz72S5Z/+g082/YPikmrWbTWRRQVFVqgoq6CmsgbP89CNBCk9hqqGKAgVoyqHXjVDf1k19jb8yLu/cO5oJZ8vZFbLkydP5vnnn+/03DvvvJOLL76YefPmtTs2btw4ZsyYwR//+Me8zLMnMeAh9ABWrVoFQDBQzLgRc2ls3k79no3g2liyRdyNs2b7GraHd+B6LqoaQNNC2I5JuHk3Tc27Ma3e35p/MMhw9hv8yNuPnGGv0NuCBQswDINf//rX2WMrV65k27Zt2Z8nTJjApEmT+Mtf/pJzrLvuuov7778/vxPuAQyI2/UAxo8fTzga5uHf/i+JpExF6XBWbHgdwTAIiCohNUShVkh9eA/bo9tJ2enSLkVW0bQQrmtnDcOhUrI6fvz4vp5Cn8CPvP3IGfbW5AuCwIsvvsgbb7zBmDFjmDx5MnfccQc1NTVtzr/rrrs63MQ3efJkjjrqqLzP+WCRr30Igud5eRk43zj66KO9jz76aL/es2HDBuKJOsznX+O51CqOO+UM3njvEcYOn8UxMy8CTcXz0jW+NYOrsbAo1UopD5a3URe0bBPXsQkFiykIlSCKUk/T6zFs2LCBcePG9fU0eh1+5N2XnNesWcPEiRP75Nq6rvtuc1p3OXdwXw6/jmkHgurqambOOIFP3V3MkmsoK6hmwhHHs3HbR+zc8Sm4LoIAoiigJ3WK1CLiZpxt0W0krb0bQRRZRVWDpPQYDeGdWJbRh6w6R3V1dV9PoU/gR95+5AwDyrY9CV8ZhIxc7NpgnCOUKjZ8vJppR55KSdEgPlz9fxiJKACaFqA5EsF2bIJKEFmS2RnbgWHvffALgoCqBpFEicbmXZhmfnYOHiz8JgucgR95+5EzcMjsHehJ5IuzrwxCxsUaf9xJmNhEP9uEKErMnXEJhpnko+UvgechCOkHfiKeLu2SBRlNDlAXr2tXkiZJMoqiEY7UoRv9r3OT31zpDPzI24+cYaBBTo+Om5dR+znmTTiVt4X1TA1WYuopykuGMGX8ArbuXk3tZ+m8REDTaI5EyORYVFHB8RyaUk3txpNECUXRiET3HHJVSAMYwAAGkIGvDEJGG7w0UMr74jZKlRCb310OwMQxJ1JROoylq/6CrscQRAHPc9ton4SUEM16MwmzvScgihKyrNIcrcdx+o8L2xcdrPoD/Mjbj5yBQ2IjWU8jX5x9ZRBKS0uzrwNDB9NIHGFnA5B+oM+edhG2bfDxx//X8juxnapgUA2yJ1GHk0PzSJJk8Dyi8Sb6S/VWa85+gh95+5EzDDTI6Un4yiDU1dVlX8+smM5r4momlFaQCKeTcaXF1UwaexK125ezc+caVFUlnoi3GUMWZARBoDHZmPMaqhrAMBMkU9H8EdkPtObsJ/iRtx85Q7qfM0BhYWG7Y9/97nc73Gj229/+lilTpjB16lRmzpyZPa8zOeza2tpOpbR7CxnOPQ1fGYQRI0YAIAoCk4sn8qq4GkWU2Pr2suw5k8bOp7igiiXL/oTjWti2g7nPHz8oB4ka0ZyhIwBVDRKLN/WLctQMZ7/Bj7z9yBlAVdX9fs8rr7zCAw88wOuvv86qVat4//33KSkpyR7vTA67Kynt3sCBcO4OfGUQ1q9fD4AiKQSlAAWV1axxdlEVN/FaYnKSJDN72kKSqQgrV72CIAgYRvvYbFANUp/YkzN0JAoisqLSHKvvczntDGe/wY+8/cgZDix3cu+993L//fczZMgQADRN47rrrsse70wOuysp7d5AvvJFvgq+TZ06FWh5YIsKR1cexbP1H/Jd6QJWvr2ECfPmAFBVMZKxw49mw8Z3GTn8KDRVpaiwqM1YsiBjYRFOhaksaN/cXJYUTEsnFm+iuKiyz/rNZjj7DX7k3V843/zqzSzfvbxHx5xRM4MHznwg57EDaRbTlVR2V3LYnUlp9wYGGuT0AFrLA2uSyoyKafxD+JSwmcTbuKPNudMnno6mFrD04xdI6SnsHJVDQSVIs9GMbue21qoSaFFMjec83hvwqySyH3n7kTPkTwq6MznsrqS0840B+eseQOsVgSppTCw5kgKtiDfsdfzboBmsXrqSCbOmpY+rIWZOPJPFK55n85YlVA+qRg61/3NpcoD6RAPDioeRywlQ1ACxeCOKoqHI+Yn7dYb+Ionc2/Aj7/7CuaOVfL5wIFLQkydPZunSpSxYsKDDc7qSw+5MSjvfGJC/7gG0XkEpoowsytww6cs8Kb+F67k0fdi2SfnIYTOoqRjDmnV/Z0/Djn2HA9Ib1kzHIGHm9gJEQUQUJZLJSM8R2Q/4ddXoR95+5AwHtlq+4447uPXWW9m9ezcApmny6KOPtjuvMznsrqS084l8eQi+MgitV1CymBaHOn/kOQwuH8kicQNHVVWzad3G7DmCIHD05PNwXYcPljyXM2wE6dBRQ7Khw80isqyiG4k+6b7WX1aNvQ0/8vYjZ9i7Wk4mkwwbNiz77yc/+UmH7zn77LO54YYbOPXUU7OS19Fo+1LxruSwO5PSzify5SH4Sv56xYoVTJ8+nUQ0ypa164gpKcBjY/Qz7vvnd/mV8wX+unUTp3/139u8b+2GRXy87jVmz1rIsXMW5hw7aSUp0UooD5XnPG6aOgGtgOKiiv2a88Eiw9lv8CPvvuTcl/LXyWQyb0nW/oruch6Qv+4EkydP3vuDAAE5gO3aTCgdz4Qxs1nHbiYUFREJh9u878gjjmdY9WSWLH2Jbds/zTl2UA4S1sMdehHpXs2xXpe1aMPZR/Ajbz9yBggGg309hV5Hvjj7yiBs3JgOBymKiud5BKUArpcO83xl0rW8In/K2LJBfPD3f7V5nyBJzJ16AQUFFfzt9Z8Ti7XfpSwIAqIgkrByx/bSxwVSen5ifx0hw9lv8CNvP3IGf2o45YuzrwzCsGHDAFA0FUVR8FzwSIfMCpVCFhx7GXF0NK8Jx2m7oUzRQsyddBGOY/PXV3+WcxeyJmtE9AgdReFkRSORau7VzWoZzn6DH3n7kTPkb9duf8bATuUeQENDWshOEEUKS0vBdBAQskJ0R9XMYpXYxLyC8Tz67m/avlkQqCgbwtGTL2RPfS2vvfFLvBbvwvNcPvzoz/zjn49iOiaGk9t6i4IIeL3qJWQ4+w1+5O1HzjDQIKcn4SuD0Fr8qqisDMe2USUNx9u7Yp91wnmoyIixelY3r2nzflFVqCkezXHHXMymzUt5Z/Gz6HqCP7/8UxZ/+Dxr1r3Lrh1riemxDuegyBkvoXcke3MJfvkBfuTtR84w0CCnR8fNy6j9FK0VArVQKJ1YltKJ5QxC1VVs1eNc6Mzkzo++x87k7jZjCIrM2OqZTJ20gGXLX+G3T9/O1u2rmX/i5yksKGPVyjeImbEOH/iiKOF5HkYHwng9jXypIvZ3+JG3HzkD/UZqvjeRL86+MgitH9KqqqKEQqiulE0sZxCaMI6hYhnTzMHctuxuoubeFb+qacRTSY6bdC6jR0xHEmUuvvAupk89jRnTzmD7zjU0NGwhaXX8wFdklVgi3Ctegh+bh4A/efuR877YvXs3l19+OWPGjGHWrFmcffbZPS76d/XVV/P8888DaSP8rW99i3HjxnHUUUdx7LHH8sorr/To9XoTvjII+9btFpaU4FkuHl4bi1s9ayoRXefzDcexK7GbOz/+PoaTlroVRAFBkmiKRDhtzn/whQv+m5qqIwCYMvlkVDXI6tX/JKx33PBcFCVcz8EwUx2e01PwW312Bn7k7UfOsDd84nkeCxcuZP78+WzatImlS5dy7733HlSfiK5i9d/5znfYtWsXq1evZtmyZbz00kvEYh2HjHsKAyGjHkBTU9t+yIUlJQiCQKlWRsrZ+3AWRJG68kKmlg/lW1zJyuZP+PGnP88eV1UVV4CmSDOO4+HGk3iOi6YGmTJpPps2fURD8w4Mu+N+CIqsEU+G8+7u7svZL/Ajbz9yhr0P7TfffBNFUbj++uuzx6ZPn84JJ5zArbfemm2G8+yzzwJpA5Lr9//617848cQTOf/885k0aRKe53HDDTdw5JFHcuqpp7Jnzx4gvTnskUce4aGHHkLTNACqq6u59NJLAXj66aeZOnUqU6ZM4fbbb8/OqbCwkLvuuovp06czd+5c6urqiEQijBw5MuvlJRIJhg8f3mEYMF9JZV+J22W0zzMIBAJ4skyRKBMnhuO6SC2Wd/zp89j46DMcX1jFNVOv4LHapzmp+jhOGDQXSBsFy7RoCDdSVlxCABALg8yYdgbLV77O2k/eorJ0KEOLhyLmUL2TRAnL0jEtHU3N38aafTn7BX7k3V84P/fyfWzfta5Hxxw2+EguOefWnMcyJZgdSVq/8MILLF++nBUrVtDQ0MAxxxzDSSedxHvvvZfz9wDLli1j9erVjB49mhdeeIF169bx6aefUldXx6RJk7jmmmvYuHEjI0aMoLi4uN01d+7cye23387SpUspKyvj9NNP56WXXuLCCy8kkUgwd+5cfvCDH3DbbbfxyCOP8O1vf5sZM2bw1ltvcfLJJ/PXv/6VM844A0VROuXc0/CVh7B58+Y2PyuKghIIgOtRqpai23u9BFlR2BCUKQ0EubBxMmMKR3P/Jz9vk09QVAVJlGgIN1HfUI/e1ExhsJQjx81lzdq3aY7tIWp0LGonS0reW23uy9kv8CNvP3IGsq0uO8I777zDFVdcgSRJVFdXM2/ePJYsWdLh7wFmz57N6NGjAVi0aFH2vCFDhnSqkJrBkiVLmD9/PlVVVciyzOc+9zkWLVoEpB/m5557LpDWn6qtrQXgsssuy3opzzzzDJdddtkBcz5Q+MpDmDBhQrvfFZaUEA+HCTrpHgm2YyNL6T/LiRecxSsPPMLpkyZz9+wb+Y+Vt/DQuke4a+o3su+XZImgHMQ2LfbU11MtSsw++kLWb/yQpR/+H8H5nyckh1BzSF/LsophJHEcC0nKvRLIB2c/wI+8+wvnjlby+UIgEADS0h2ZZO/BojvicWPHjmXr1q1Eo9GcXkJHUBQl2zBLkqRs+Of888/nzjvvpKmpqUtp7gznnoavPITly9t3cSoqKsKVJOSiIkrlYnQzhdMSxystK2W9DLplMXKTwb+PvpTXdv6D9/Z82G4cWVWQVIVELEqRVMysGWezfuP7NNRtpi6xB7eDXIEgCOhG/kpQc3H2A/zI24+cIR3LB1iwYAGGYfDrX/86e2zlypWUlpby7LPP4jgO9fX1LFq0iNmzZ3PiiSfm/P2+OOmkk7Ln7dq1izfffBNIJ/G/+MUvctNNN2X7K9fX1/Pcc88xe/Zs3nrrLRoaGnAch6effrrLvgmFhYUcc8wx3HTTTZx77rlIktQl556GrwxCLhnbTDJIkGUKy6qoClSQspLYLfIS515yIb95exFaOMF/FJ/FEYWj+PGnPydht78hiqISNwxsPcWsyadTXFTJO+88jWGnOgwdybJKUo/lLbncmXTv4Qw/8vYjZ9i7mhcEgRdffJE33niDMWPGMHnyZO644w6uvPJKpk2bxvTp01mwYAE/+tGPqKmpYeHChTl/vy8WLlzIuHHjmDRpEl/4whc49thjs8fuueceqqqqmDRpElOmTOHcc8+luLiYwYMH88Mf/pCTTz6Z6dOnM2vWLC644IIuuVx22WX8/ve/7zRc1JpzT8NX8tdLly7NmXQKh8Ps2bOHoqIi7HiClJlgj1GPKqjIksyXr/lP7jr+JGoGV7NyepCvLLmFhSPO4eaJX2kzjmmmsCyHsrIKCrUAtfVrePmNhznhuCsYP/EEhpcMR5O1dtfXjQTlpYNRlZ53AzvifLjDj7z7knNfyl8nEom8PSD7K7rLeUD+uhN09GUpLS0lGAyi6zqiphIQFAYXDMH0TBzX5aJLF/LT115DTOhMjpSwcMQ5vLj1ZT5tXpsdw3UdJFFGUSQS8TjIEqOqJjByyGQ+WPIiRipGXbwu5+YhSZTRO5G7yAfnwx1+5O1HzpC/1XJ/xkALzR7AsmXLcv5eEARqamqwbRtPFEEQUEWFqkAVup1iwSnz+LSxgfVNjcjrt/GlkZ+jUivnvk9+npW9cF0HVQ0iiTKWbWKaJqKqMm/25QjAG3//FaZt0JRqXysuyyopI5EXFdSOOB/u8CNvP3KG/LWT7M8YaKHZA5gxY0aHx1RVpaqqimQyiaAo4DgUqAUUaoU4gsPCfzuf7/3pBbAcCjc1cNPE69kU38yztS+mB/A8JFlDUYOIAkRj6XLSkpJBLDj289TVb2b5+3+mKd5IYh8do3TFgZCXncudcT6c4UfefuQM/tyhnS/OvjIIa9eu7fR4aWkpsiwjKApeS2inXCtHQOD8i85lc2MDK6Nh5C11zBMncXzVHH772TM0G80IgogkSihyAFmWMXQjW3kwbvTRTBl/Ess+fZ362k/ZHd7errNavvYkdMX5cIUfefuRMww0yOlJ+MogZDaadARBECgtLSVlmiCk+yRIokRlqIqSiiLmn3wi33nqaZyCAPLHG7h+5OdIOTrPbXkJRQkgCAKSJCHLKoIIzZHmbPXQicdcQkXpEP75/h8wwo3siuxsU4oqSTK2beZsvJNPzocr/Mjbj5xhb6Wgn5Avzr4yCDt37uzynMLCQjzPQ9Q0vJYNI0E5QLFazAUXn8uexibeScXAdhiz0WBe1XG8sPVlUp6ZHUNVQsiSiGEYpPR0GEiWVc446ToMK8mSVa9gRCOEE21bcYqiRErvWS+hO5wPR/iRtx85A1lP3E/IF2dfGYTy8vIuz9E0DU3TcGirOV6ilTBj1jSOGDOanz/xB5JjhiDWR/i6dh4JJ8nztf+XPVeSZQRBRFUUwuEwTkuyuKJsCLOmnMG6zR/Q2LCVxqadxI29yaH0noQEjtNzwlXd4Xw4wo+8/cgZQJbTygKSJDFjxgymTJnCeeedR3Nzx4rD+4MnnniCG264oUfGisfjfPnLX87Kc8+fP58PPvhgv8fJcO5p+MogdHd3X3l5OYZtI4hiNpcgizKVgUpuuvWr7Ny1i/969EmsQSVU1yb4XPHpPPPZcyRaeiAIgoiiBAEPz4N4dG9J6dFTz6KkqIq3PnoGxRXZXb8F026R1hYEBIEe3bmcrx2N/R1+5O1HzrC3D0QwGGT58uWsXr2a8vJyHn744T6eWXtce+21lJeXs2HDBpYuXcrjjz9+QK1P89X7wlcGobsa4pkMvqhpeM7eUtACtYC5c+bwjVtv5N133ufxZUvxgirXx+Yimy7PbX4he64sK3ieh6ZpRGMxEskEnuchyyrz515Jc3QPK9a9iWJ77Gzcmr3BiqyR1KM9tnPZj+0FwZ+8/ci5Ixx77LHs2LEDgA8//JBjjz2WmTNnctxxx7FuXVqJ9YknnuCiiy7izDPPZNy4cdx2223Z9z/++OOMHz+e2bNn8+6772Z/X1tby4IFC5g2bRqnnHIKW7duBdJNc77yla8wd+5cjjjiCP71r39xzTXXMHHiRK6++moANm3axAcffMA999yTvVejR4/mnHPOAeAnP/kJU6ZMYcqUKTzwwAPZ602cOJHrrruOyZMnc/rpp5NKpVi7dm0bmY3a2lqmTp160H83X4nbdSQluy9kWaaoqIhUIoHU6sEsCALlgXLOvegsPttYy2NPPsOR/3UL8wjxkHQVN2x4igtHnUepWoIoygiAIKTDUE1NTcSkKKVlZQwfPJHxo2fz0apXGTfqaIJuIbul7QyuGIEoSliWgWXpqD0gi91dzocb/Mi7v3C++eabe1xXacaMGdmH5L4Q9pGXdxyHf/zjH3zxi18E0qJ/b7/9NrIs88Ybb3DnnXfypz/9CUjrP3388cdomsaRRx7JjTfeiCzL3H333SxdupSSkhJOPvlkZs6cCcCNN97IVVddxVVXXcVjjz3G1772NV566SUgrXiwePFi/vznP3P++efz7rvv8uijj3LMMcewfPlytm7dyowZM3JqFGW8hQ8++ADP85gzZw7z5s2jrKyMDRs28PTTT/PII49w6aWX8qc//YnLLrsM0zTZvHkzo0eP5tlnn+1S7qI78NWSIh6Pd/vckpISbNdFkOVs2AjSiqglWik3fvN6jpo1g7t/9HPCo4ZwhFnGVfox/Gz1L4D0ak2QZRzXRpREgsEggihSX19PfUM9c6ZfgCwpvL3kj2jBAhLRMM2RPbiWjSiIxBI90zxnfzgfTvAjbz9yhr3hk1QqxYwZM6ipqaGuro7TTjsNgEgkwiWXXMKUKVP4+te/zieffJJ97ymnnEJJSQmBQIBJkyaxZcsWPvjgg6x0taqqbR60ixcv5sorrwTg85//PO+880722HnnnYcgCEydOpXq6mqmTp2KKIpMnjw5K3HdEd555x0WLlxIQUEBhYWFXHTRRbz99ttA2ovI7DHJyGW7rsull16alcvuKYPgKw+hsrKy2+cGg8G0JRcEPMNAaOWOF6tFxMwo3/uf73DFxVdx+4MP8+ubbuCyTfDN2mf4aPgyjq46ClUOYOjxrNnNSGVbpkVMN5g+4TSWrPozn21bzqjh02iO1BO0FQQRDMEhFSgiFCzqNc6HE/zIu79w7mglny9kEqyZHEIymeSMM87g4Ycf5mtf+xrf+c53OPnkk3nxxRepra1l/vz52fe2Lt9sLUV9IMiMJYpim3FFUcS2bWbOnMmKFStwHKdTJdOOxs3MMZVKIcsyl112GZdccgkXXXQRgiAwbty4A557dq4HPcIhhO3bt3f7XFEUqaiowLBt2GelLosypVopxeVF3Patb7ByxSqeWvExXkkh/+VewC8+/jmGYyBJCh7tV/mKqhAIBhlcOYWy4sG8veSPeI6DI4IlOUiKguQINDftwnFyt9DLB+fDCX7k7UfO0L4EMxQK8bOf/Ywf//jH2LZNJBJh6NChQDpv0BXmzJnDW2+9RWNjI5Zl8dxzz2WPHXfccTzzzDMA/OEPf+DEE0/s9jzHjBnD0Ucfzd133531/mtra3n55Zc58cQTeemll0gmkyQSCV588cVOxzZNkzFjxiBJEt///vd7xDsAnxmEsWPH7tf5RUVF6ce5orRJLgMUqUVIgsSpZy5gwanz+X+/+A07B1dQKAS5Nno0T6z7PaLY4mF47SsCBAFCoQKmjDuNeCLMklV/Q5FVmo10qZysKHgpg0hkbwWC53ltwlf54Hy4wI+8/cgZcjeLmTlzJtOmTePpp5/mtttu44477mDmzJnd8gAGDx7Md7/7XY499liOP/74NmqhDz30EI8//jjTpk3jd7/7HQ8++OB+zfXRRx+lrq6OsWPHMmXKFK6++moGDRrEUUcdxdVXX83s2bOZM2cO1157bTZv0RnnjFx2po/zwcJX8tcrVqxg+vTp+/Weuro6YuEwiuMg7tPHNGEmqE/WY8QNLrv4C1QNquLJe+6Gj9fwsPRPzjn1eoaqldi22WFHNMMwWL72r9TuWMEV5/8XWrCIwUWD0SQV17YxbYPyIaPT3dwS6UolrbS0TQirpzkfDvAj777k3Jfy18lk0nd6Rt3lPCB/3QkO5MtSWlqKDVkpi9YIKenWmEWlRdxy+82sW7ueV1Z/gjW4jOud+cYA3zEAACAASURBVPz9vd8jCBKe27HRVVWN8SPnocgab7zzBCICcSOdHBRlGVGQaNq1Bb25GY+0l2Dth9Kh3x6KGfiRtx85w4C4XU/CVwZh6dKl+/0eTdMIhUJYrgv7hGsEQaBYKcZ0TE47fQGTJk/kl//vUYQZU4iWSFwXm83SJS93Or4gQDBQxOxpF1DXsJlP1r5FzIzhtISZFFUDz8XyLERJQlIUbF3H6ebW9QPhfDjAj7z9yBkG5K97Er4yCAfaQKSyshLL89rlEQCCSjDrgN309a9SV7eHZ557kcpTTuOTUANzdhUTX/Npp70ONE2jvPgIjhgxkw9W/IXmyG4SZvqGC4KAGgiRSDXjtFxfUhSsWKxb+QS/Nk3xI+++5txX4eeBBjm5cSD3w1cG4UBXUMFgEFlVc9QLgSRKFCqFmI7JrKNncsKJx/HEb35PJBZjxIKzeU3+lEFbU3ibd3R4gwQhXU42a9K5aGqIRYufoilRn72eiIAgiFl5bEEU8TwPMx7HMQxc2+7QOPh11ehH3n3JORAI0NjY2CdGYcBDaA/P82hsbMyZcO8MvkoqHwx27txJsr4eNRBol9DVbYPdiV2ElBAbN37GlZdezeVXXsI3brmRN7b/E+39T5nNGKxjJiBUlSII7e2w54Gup0iau3n97UcRBJGaylGMGjyJk466iFCwGMPQKS8dnK27di0r/QX0QBAF1OJixH6yW3UA/oJlWWzfvt2XvQn6KwKBAMOGDcu1g73DpLKvNqatWrXqgPU+CgoKiDY0oLhuO4OgSSqKqGC7DmPHHsG555/FH5/5E5dcupBTRyzg7qGLGLm9kkErNmIeOwmxIIS4j1EQBJAlmbLCkVx4+tep3bGa+oYtfLD6FXQjwcIFNyBJEolUmOLCSgRBaPPw91wXo7kZpagIudWq4GA49zai8SYKQyXpct2DxKHEu6fQl5wVRemzfgwD97rn4KuQ0fjx4w/4vZqmgSS1SyxDS3JZLcZ0081tvvqf16GqCj/58UMA3DjzRr4XeBnXMAms2orbQUJYURUM06CidAQnzPo3zjrlq8ycsIBVG98lqcdQZBXD1EkkI+3nIIqIioIZi2ElEliJBEY4zIjKSqxEImf+o7/BsoysVPjB4mDu9aEKP3IGf/LOF2dfGYSMMuGBQFVVhE62mweVtBCd53lUVlXyxeuu5u233uW9d9+nMlDBhdOv5H+kvyA2RlDX7ci5WQ3ShiccDuN6Lq7nMm3CfGzHZPm6f6WPqwGSeoyU3l63RhBFJFXFTqZwdB0EgZ0NDdipFHo4jJVM5ozxuo7TZwnBDDzPw3EtvB4yCAdzrw9V+JEz+JN3vjj7yiBUV1cf8HtFUSQQDOJ0sFtYFmUKlUIMJ+0lXPG5Sxg+fBg/vu9nWJbFOcPPJDq4kGelJchb9yBs3tXhdTwPotEoiqQQKCxleM2RLPnkNVzPRUBAVQPE4mEsq328VhAEXAlERUEQRQZVVCCpKqKiYCUSmJEobitvwdZ1jKYmzEh0v3dB9yRc18FxnHa9pg8UB3OvD1X4kTP4k3e+OPvKIBxsB6XCwkIsQegw/FKkFmO3rHBVVeWbt36NLbVbefaZPyEIAnfNuI3H1PdYpexGWbMVGtqHfgACAY14PE4ilkA3dY6aeBqNkV1s3rEKSFcdKapGOFrfxih4nkciGaE5uifbdS0SSzfnEQQBWdPwHBsjHMYxDIxoFCsWQ1QUXMfGaG7GPQhxr4OB53m4nnPQ2k0Z9FS3rEMJfuQM/uSdL86+Mgj7W4K1L4LBIIhiO7G7DDRJpUAJYTrpHMEJJx3HccfP5ZFfPkZt7Vaqg4O4Y+atfN37Hc2yibJ0PUIkd/lYMBgkHosTi8YYMXQKoUARS1a/lj0uCSKKohGO1KMbiawxSKSigIfZYij2bcYtKgqiLGNGY7iWhaRp6VBTS4LajERwDKPXQ0ie5yJAj3kIB3uvD0X4kTP4k3e+OPvKIBwsVFVNG4QcMhYZFGsl2O7eh9rtd34DVVW56T+/SUNDI6cOXcDJI0/hy95vMLGR312FuLUu51jBUBAjmaK2bjOTxp3A2tolROON2eOSIKJqAaLxJiKxBpJGDE0NIMsaeo4cQwaCKCJpatYIZCDKMoIkYUajWS/Ccxw81827gfC8dPWW4/ivYfoABtBf4CuDcLA10rIsp2t6JQk6CBsF5QCqpGVXukOHDuGBh+6jsTHMzTfeSiKR5NZpN+MVBrlWfQKzNIi88jOkFRtzjhlQNVzdZdCgI3E9lw8/ea3NcREBTQ1gOyYBJYiAgCSIWI6FZVsYhrFfHNPGQkMQBMxYDL2pKf2voQEjHMbW9bxULLmeiyiIOI7dI8bHj/XwfuQM/uSdL86+MgilpaUHPUYoFMIWhE4TsGVaKaa7d6U7ecpEfnjf99mwfhPfuvXbqKjcc/TdbLHr+ApPkjqiGmlbPfI7qyCeyr4vs2ouKCxCcgKMGDqVxSv/SiLVNvcgIKDKbUNDgiBgmjolRQfWYEeQJCRVRdK07P8AVjyO3tSEGYu1MwwHIs+dgePYCKS9r85kPrqLnrjXhxr8yBn8yTtfnH1lEOrqcodm9gcFBQXYntdp2CggB1BEBatVgvSEE4/ljm/fyuL3PuSXv3iUyWUT+Z+j/5u1sQ3cnngc/eixCLqF8vZKxB31QPohqahBNDWELEscOfJEbMfknx89l/O6raHIKroZZ09jY5fndhetjYRjGOjhMHYqhWvbWMkkRlMTRjh8QIlp13XSvXE9D7eDktz9QU/c60MNfuQM/uSdL86+MggjRow46DHUlp4IgqLk3KQG6dV5ZagK13OzCWaACxeey8J/O58nHvs977y9mPlDTuKOGbfwYdMyfrD7cfQTJuMVFyB/vDEdQrJsFDmAJCnIokxxYRUjBk9n6ad/JxxryHntDERBxHVsaqry01ZRUlVEWc5ugLOTyfQ+DVE8oGqljEHwWl4fLHriXh9q8CNn8CfvfHH2lUFYv379QY+hqiqSJKFbFlYn8XlNUqkprEEURHR7b7zvm7fexPjxY7n7299n9+46Fo46n+snXsvf697iz/X/wJo7EWfsUMRt9QQWr0UMRxFFEUEUUBSZiWPm4Xkury7+bZdzFQSRTVtqD5pzh+O3bITLhJUEUUSUJARJShuFjNZSN+C4NoIgIghCtmT2YNAT97o/wvM8dCN3Zdrhyrkr+JF3vjgPiNsdAAzDIJFI0LRtG4ZtU1BQ0GHTbNu12ZOsx/UcVCntXWzdso1/v+Iaxo4dw69+8xCyLPOld26gNraF3859mOJACV59mMDqLZDUoawYN6jiBBW8kYP5aNPrbNzyAWed8GUKgyEi8UZqKkcyZljbBikuHrZlUFZSgyT1rmyV5zg4tpNWctW0rNHoCPVN25EkGce2CGgFFBWW9+JsDx04jk0kWk9ZaU06xDaAAew/BjqmQc/JA2uaRnl5OcNHjaayvLzTjL8sypQFytqUoo4YOZxv3/0tVq5czZ23341t29w85QbCZjN/2vkytm3ilRUhnnYcwoTRCKqCGEkgf7YL+cM1TD7iJERJ5m/v/JI//v0nvLb4SZ78y/d4+tUf0Ryrz15HROCzrbtpiuzO7kvoLQiShKyld0i7loUZiWDF4x16DK7rIAoigihi2x2XnnZ3AXO4yl+7novtWjnDaocr567gR9754jzgIRwEHNNED4fZvns3gUAAsYM+x57nsTO+E0EQkVspeT7z9PPc/78PcPKCk7j3f7/Hd1f8gDd3LeLZk35DpVZFILC3CYbruiRrPyOwdCPOqBr2DFOIJ5oRJI2jxs1mzWfv8tbS5wCBs47/D46edNreeXoulmUQChShqQEQJERB6FWvwfM8XMtCVFXUwsI0J9vGdRxERaGheTuaGsL1XBzHpqp8WM5xjGgUXBeloMCXUt+mmaKxeReV5cNQ5I49rgEMoBMMeAjQ81ZVlGVESaK4uLjTev+MGqrltl35Xn7Fxdxy+828+c9F3HH7f/GlcdfgeR6PbvxDG2MAaY0jBpVjj6pGqt1NpVnAyKFTqCodxpbG7cyaciY3XP4gI2om8Je3fsW6LWmu6z/bhiSIaGoA3UwQjtXTHNlNU/PunFpI+YIgCEiqittiRPWmJsxoFDuRwGhuzvadziTDc4rwWVZ6s1yL1LfR3IxjmjnPPVxXjWmDOeAhtIYfeQ94CPugP3gIAFYigR6LsWP3bkKhUIdxXcd12BbfRlAKtjvnmaee4/4fPcjlV1yMeq7K7zY8zZPzHmFS2YQ25xlmAjMZR3t/LYJpY500HTSFeDJOsDBAaVEZqiDz3Cs/ojm2hy9f9EMqy4Z2OB/Pcykr6f1YtLdPTwkjlSQSq6egohIEAT0Ro6ywCrWgELGlGZDneZjNzXiA2JKvcR0Hz7YRJCntMajqYR9XjycihKO7qSgdQih4YHtMBuB7DHgIACtWrOjxMeVgEDUQpCAYRO/ES2jdanNfXH7lJVxx5SU88/TzjNo2kopAOXcvu6dNdVJ6DBVPFHFmjkuXpC5agbh+GwUomHGLaCRGwkox7/jPI4kyT736Q9Zs3Jh9v9vK+EuihO3aGGayB/4K+4d9GwyJSvoBb0aimOEIdiKBpacwI5Fs+apjGOnwUqvkvShJe3dVR6PpzXItHPNxr/sDXNdClmRsu/1n7XDl3BX8yDtfnH1lECZPntzjYwqiiFpUSHFxMXYHjW8yKFSKcLzcNfZf+/pXmTZtCvd/7wG+MuQ6NsdqefCTX7Q5R5JkZEXFDirYcybhlRQgr9+O+ubHFHy2C0s3CDc0ExALmXfcv9MUqeO91U8QjtaRTKbSbUCTe3dCq7JGPNmcDT+kSxqTWHbPKI52F57nIcgSoiwjqUp6pS9LIAhpsT3TxE4kst7CvhBaDINrGNjJtIHLx73uD3AcG0lScpbm5puzZRk5mzP1NQ7Xe90Z8sXZVwZhY6vVck9CkCSKKisJahqpTiqONElFldSsRHZrKIrCvfd9D1XT+MN3n+HSkf/Gc5tf4O3d7+69jiAQDBQhSDJ2SRB79kSM+dOxh1Yg19YR/HgjAVEiGoujCKXMnrmQHXs28dAzN/P3xU/heh576vdgWukHviiIeJ5LSk/gODbReCPReCPN0TqsTip9ehqe54LntQn3OI6NKMt4QOPOzVim3s6z2BeiqmIlk9iGkbd73dew3bRByHV/8s3ZcR0sq/+JDx6u97oz5IuzrwzCsGG5K1d6AqKiUDF4MHYq1eE5giBQrBRjOrlDS9XVg/jBvXezefMWtj++g7FFY/jesntp0BtbjSESCpYAYFsGriYjzZqCO/NIxKYY6nurCekWISVARfEY5s26jqGVR7J0/d/4y7s/RRAF6uv34DjpXdaqEiCRitAU2YVtmwTUIJIktxiF/RPGO1A4joPQqse0KErZ0lPLtUiZSWyh61yXIAhIioIVjTJ40CBc284qtXqOk61qyoX91WDqaJx8wvM8XMdGEiU8XNx95pzPzzdkmhj1P4OQb979Efni7CuD0NDQudzDwaKwtBRJUbCtjkMuQSUIQsf19HPmHsPtd36T9xa9z5DFNSTtJF9bfAsRM5o9RxRFQqEyFCVAqKAcVQ2gjBmFOXt8Orfw7mrU15dQ/nEtlTtTnHDk+Zw+54vUNdXy0bq/4HlQ31CP63kICCiKhqIEsmWMsigjSyrhyJ798hRc1z2gcJPjWG1W/6Ig4roOrusQTzajBkKYVseGtjUEUUSUZfbs3InR3JxVatXDYYxwc1praZ/7Y6dS6E1N2N1UkHRMc7/O7ymk264KLa/bS3zk+/Ptuk6b/TT9Bfnm3R+RL86+MgiFLfXv+YIoSZTX1GAkO354dZZczuDiSy7ky1/5Iouef5cTdhzH5lgtN7z3dWJmbO84kkQwWJzdIS0IAmrNYFLHTcSeORZ31GBQZLQd9ciLVjB4TYLTKk5n94bl7Kpfh67rxONxnJQOuom4T+GBJErIskI8Ee72ZjDDTBJLNO63fLXruW3CRYIoYrt2um+056JICrZt4XRzVS5IEkXFxWlZjTaKrem+2EarZLWVTGLG44iShBWLpRPTrpsW7GspibV1PcvJMU3MSARRkrATiW57Fp7r4ma8FMvaL4/EcxzMWAxbN/BIz0MQBNx98lH5/ny7ro3ruj2iNdWTyDfv/oh8ce5dPYM+htXJyr2nUFJRQf327e1KK1ujUCkibnbcwAbg2i9dTVNTmOd+9QIX3XkBf/Ze5obF3+Dh435KoZL7wyArGkJAww4FEYdW4boOZlQnGI4gbNvDkEaVIRyHu7gWR9tJvWlT73kIksSwS85HLWsrqSuLMrqVQjcSBAOdfwA9zyOpx7BtE9uxUeTubxpzvX1CRghYjk0yFUVVWzpDCeC4VocSIfvC6kBcT5QkXMBobkYOBLBTqbQOkyAgiiKOYWDreto8CgKCJGHF49jJJFIggB1PIKrpftWZZLfSgcS457ppCQ9dxzYM0s9yDxAQZQm1pKTLvAiAnUxi6zqWbWInYtjFEoheuwdzvj/fjmNnQ1Wi2L370Bvoje91f0O+OPvKQ9g35poPqKpKUWUlerLjcs7OkssZCILArbffzIknHceff/QyX63+Emub1/OzT/5fp+8JaIU4toVtGuB5SIEQ1rBKrOOnkFgwHXPGEaxmE7vEJsQxIwnOnIogS9QvWpxzZa/KGrF4E5Ha2k4b49i2gevYiKKEae5fKMV13Xb7B1zPQ5RkhBbPRRREDKN7YSPoPMafEeCzUqk2excym+daexVii+S3IAjYyWTWGEA6b2TpejYE1dqrSDU0oDc1ZTfPiYqS7lKnaUiaius4WIm2InW5+km4to2t6+m5KDKI4CQSeI6LvU94Lt+fb9d1EGjvmfQ1euN73d+QL86+MgihUKhXrlNRXY3tOB0+QLtKLmcgiiLf/f63qaqq5Ln/foEzak7l5W2v0mx2XPonKxqqFiIUKqWgoJxAMJRezQngqBLS8CEUz5zB66m3WBfciVFTSckxM9F37ia2rn3lgiiIRD5czmd/fI5tf3+jw3BQIhVDkuRsH4buho0yidJ9Q1ZBLYQi7fUyJFnBtLtvaILBYKfHRUlCbtnDsC9y/i5jGFqt6AVBSEuAx+MY0ShGUzhbVCAqStawiIrSbkxJVdOeQ0sewrUszJbd162lw61EEkHcqwIriBKiJOGlDEyr7ecn359vp0WivL89gHvre92fkC/OvjIITU1NvXKdYDCIVlKCZVkdGoWukssZlJQU88P7vk99fQO7X9qN4Ri8VPvnDs/PlKbKitrSNc0hECjCsx00SSVuxBk7fA5jhs7k/dUv0RTbjlVdSaBmEI2Ll+DskyhtXvkp8dXrkEuLaVy5krolS3Acp828LdvCtHRkSckmhO1WdfK2bRONNeTk2jpR2hlEBFzHwe5mn4XmaLTrk3oAoiThui6eZSEF9sqAd2fHtKiqWLEYRjSK3rILG8huyHMtC8c0sppNjmOnBQAlCRwPO9XWC239+c7kQRzTzMp9HAw8z0t38Gu5v/0JvfW97k/IF2dfGYQhQ4b0ynUEQaCquhpTkrKibvsik1w2uvASIN2C8+u33MiyV1Yw1BjCHz97odvVHsGQiiwrqFoI17ZQZZWwGWbeUVdSGCrnn8ueJNzcQMExM3FNk4Z3P0w/QDyPxOatNL77AaFRIxh68fkERgxl11uL2PnJMsItCqqe56Hr8XbCfhmdJM/zSKTCpPR4zt21+5WAFjzsbvy9AKor89MYKBckRTkgob2Mh+FZFrKmIUrpzXmZDXlWS7I7A9dzsh6KrKqY8Vib0Njgmhoc08SIRNPhqnAzZjSKEYng7Gdv7X2RMQLpvtf9K2bfW9/r/oR8cfaVQdi8eXOvXauoqIhQQQGOmq5s8XKsbIvU4hZNoa4fipdedhHnnHsmO57fyR69nn/s/Fe35pGIpR/MqhoCQYCWL3bSNTljzrUkUhHeWf00m5u3UTBpAvH1m6h97ClqH/sDu19/E21QJdWnnoSsKAw+/WTU8lKa/vkejW8tpva1V9n8xmvEdm9vo7ypyCq6kU6aG2YKw9RRFI2kHms3v7ScRveMgiTK3c5PbN2xo1vn9TUESWpnTDJGwXWcNruz3RYPAdKVWB5gJhLYuo4RibJxzRrMaBTPsbNVVZmciJ1MHpSX4LW0jRVEsY331x/Qm9/r/oJ8ce5U3E4QhAWe5/2z5fVoz/M2tzp2ked5L+RlVt3AgYjbpasjes8G6rpObW0thcEgdjyes0FMk95E3IwTkANdjmdbNt/4+rd4b+b7DK8axovnP9Ple7xWO4Bd18EwElimju6alBdUULt1GW99/AzgUVY4mLnVx3Nk1WSceBJch7JjZiK3ildasTh73ngLKxbHs21cM71aLJkygbJjZiJpWpq7maKsuIpIvDEbSjLMFGUlg5FbPeQsSyccqyegdB7zB/DwME2dirKhXd7H3r7XvYGGpu0oipZNtKeMJCXBchQlHabyBKHDKizHNJGDQZSCgpzHu4JlGTRFdqHIGp7rUlE+tNNKut7E4Xivu8JBcu4wntmVQVjmed5R+77O9XNv40AMwrJlyzjqqN6dcl1dHbFYDMW2EVpWWK1huzbb49sJSIHs6q8z6CmdS3/0eXZO38WNBV/hqtM+1+n5TfUxyqvalkXajk0qGSFlp6gqGITnmGza/jHrt33E7sZNHDF0GleceRua2vVD2tF1mj78mOin6xA1lYJRI1CKi6AgQMGIEaCKqHLaSJiWQTBQSEGoJPt+w0wRjTegdcMgQIuhKRmE0jJmR1i5Zg3TJk7s1piHAlzXobF5Z5u/k2HplBRVoirpxURnnDOhy0BZWToHkeO45zjp3INlZfWhMosJw0zRHN2DqgQwjSSlgQpwnG6XzuYTffG97mscJOcDNggfe543c9/XuX7ubfQX+euuYNs2mzdvRhEEBMNEVNvHmiNGhGajmaDcvYfi7sY6zv/HJbg7XU7ePo9v3vI1amqq92terusQjTVgYVMdqiGQeahsWMSiFU9TUzGSz5/zbSzb4JPP3mfrrrWccezVVJTmvo7R0EjTh8sw6htxWjbmySVFDL3gbOSCtIfh4mFZBhWlg7N17Ck9TjzZjKZ07SEBWLaJqmgUFVbsF99DHY5j09S8O93gqAWGpVMQLOm2DHZrLyGTcHYtG8+x08YAAdjrUYqShFJYiKgo6EaCaKwBWVRINjVSVlKNKIhImoaaYx9GJjzV2kPta8MxgCwOWP7a6+B1rp/7PfqikYYsy1RVVWE4TodVRUVqESIiTjdjvDUV1dxx9C0wChZ573DxhZ/jicd/n3OzSlN9+7g9pPWCQsFCJETqUw3ZndPTxp3EmXOuoz68gwee+k9++oev8vri37Kudgl/XfRohzddq6xg8NmnMeqqyxl97b9Tc/apOIkUO//vFexEuhomU1raup1nrj0InUGWFXQjmVPtszVWfPppt8c8FOB6Luyj5ySKUpsEb1ecRUXBTibTch7h5nRewbERxPSDXdZU5MyublVNFwy0yH+kGuqxmqNYkSgIIEhyu9JZz/Owk8nsHgy9sRGj5X+9sRG9qalLDajO9KYyMM1Um89Q5nttJ5OYsdyf9/4AyzLQjZ6Rm++TBjmCIDQDi0hblBNbXtPy8wme55XlZVbdwKHiIUDaS/jss88IiCJeyyalfZEwE9Ql9xBSgt0KHXmex93L7uHVbX9n6vLJrHhpFaNGj+T2O77BMbNndWtenucSTzThuA6SJFMVGpS99tZd61i9+V8MqTyC6tLx1NYt56O1f+ML5/w3Y0dM6db4qV117Hr5deRQiCHnn4lcWIDjuXiuQ0nRICRJIhZvxLTNNnsOuoJpGwS1tqGnwx2mpROJNbTxpBw3vcgoK+6+d5hZue/Pat1zXeLJCIaZRJFVTEuntCVs53kermmiFhdjJ5M4tp3dyLcv3JawaUdhJtswsFrKhUVVQwkFc35XYvEmPM+luGhvJZmt65jRKIIgoBQVIQe653H2JsKROiRRajPvPsIBewgXAD8G7m/1OvPzhT01u97CsmXL+uS6sixTUlKC2aK8mQsFagFVwUqSVqpbnoIgCHxr+jcZVjiUHXN3cs+Dd2PbNl/50k3c/78PZM/ryENIjyESDJQgegKW6xBrJaA3YvCRLDjqKo4cfgI1VSOYNfEMgloRry9+Mju/XQ2befiP3+AXz93C8288wKJlL5BI7R0jOLiaIeeejp1Mse3Zl4isXovopb2C5tgeLNtKJ8e6YQBbQ5ZVknqs03r4lWvW7NeY+cbBdiZ0HDutaNcKoiji2FZ27O5wFkRxv0M3gpiWSRcz+yuEvTtlBUFAkGXMSAQPOtzsB+nqKddxMOPtNy46pokVjbbs6NbwbCvdJjUabbeXx7ZNdCOZHWPpRx9hxWItu7kVrHi8T9RoO4NtW+hGokvPtrvI17Nsv1poCoKgAFOAHZ7n7cnLjLqJQ6HKqDV0XWfLli0EBQFcN2diD/Z6Cqqk4HouruciiRKalDuJuj6ygavf+jIjC4dz59TbePnXr/D8cy9xz713c+ZZp7WJ4XaElB7DMlKYOFQXVKN1kLBdvWkRby77A+efdCPlpRU8/cr/oqkhBpWPoD68jUi8gbHDZ/KFc7/d5n1mOELD24tJ7diFVlVJaMRQ9PoGzMYwUihI5YlzCVYP6sZfcS8MS6cwVNqhxlJ/qzxJJKNIkkRAO7Aqn0QyQsqIZxP0GehWiuKCCgJaKK+cw9E6QEASxC7/9l3BMQzkUChbkeY5DmaLMdjXWDmWhQAoRUXZKr09DVvxPJfy4hpEBPRoFLnVe9NJcRm1pLjftFSNxhpJGXEEBAZVjjjo8fJVZdTpiIIg/FIQhMktr0uAFcBvgY8FQbjiQGfTV1i7dm2fXTsQCKBpGo4odrp6KVALIgAQHgAAIABJREFUGFw4GE3SKNFKGBQaBB4deg3jS8Zx35wf0GSEufa9r1J0YRFTZkzi3h/cz84du4g2dx2zDGgFiLKChEBjqjEdisiBSaOPp6yohjc+eJLf/fUeSgorue6ie/nCud/mm5//FWccexUbt33M+i1tVy9qWQmDzzuDQafOw04kCC9biR2LExg8CDuRYOeLf6PhvSW4VvdXT0qLl9DRgmZDP6pNd12XlB49qHalTqs9CK2hyBqJVATP8/LKuXWuJ7057cBXumLLvgijuTm7eS6XMYD0pr+MQm2qoYFkQz1mJJ3LSDTUY8ZibNq2ra18uqLgWGa2e15fw3UdUkYMVQnk7GNxIMjXs6yrHMInnudlDMLNwHzP8y4UBKEGeOVQqzJKpVJdatzkE/F4nB07dhAURLyWjmDdQcyI0ag3ElI61i+JmFF+uuoh/rrtFWq0appeDDPBHM8vfvkggUDnJZqQ/sInEk2YroWmBKgMVuZ8AH22cwUvv/sLBpWN5uy511NaUokkiSSTSVJ6khcX/QhFUfnPS3+CJLXn5zkunuumhdpIrxYbF39EbM16EEUkTUVUVeTCAkIjhhEaNRy1NHeuQLdSBNUCgoHiNnsbAFK6TrCH48jd8bZyQTeSROONCNCtPRS50HqF3m58K0VxQTmeJ/U4Z0jzbgjvyOYvbNdGEmVK/j975x1nV1nn//fpt8/MnZ6Z9AQS0ggJLfSiIIKg2NDVtbviLrpiQVwX9OeKq+JaWBUXGxawgIpIs1ETWkIaaZNJMpPp5fZ26vP748zcZDIlk0mGoPh5vfLKzNznnHuee895vs+3fT4vcSxcCIFtW6TSPciKiqEHiEaqx/yuhRC4loVRUTFm/8/B46bbiygUs2TzCQw9iGkWqYnPQDmCnNkwPOfAmnGUa9mUcwgHk/a/CvgtgBCiZ6pXcjzR1dV1XN8/FAqhKAqSoYMQk44rh/UwiqRMmFuo0GPcvOozfOvMW4kZUazLLDaftZVPfvszpNOH5/XxRXcq0SWf8nqwmPArW4bgug6u5zBvxgredOENXH3B9QQDEQqFPJlMBiEgGomx+sQrGEh18vy2R8Z8H0mRy8YAQDEM6s4/i8pLzkY7oZnw3NkYtdW4xSKD655j/1330vbTX9H9wJ8YXPcc2Z27cXI+S6ihBTCtIsl0N5nsILZdKmsm9A4JiHiT7AQ/HDzPI5PtP+zO2HbMUWOKpazfyT1E4T0VuI6DPM7C5XsJGXr6pieK6/NNHfgMZUnGOw7dyn7+QoAso2o6puVTp/SOIRZTVs/LZsf1yD3HwUwmcQ+jhX408Klb0qjDnfySmJKH4DmO34U+dOx0rWWH26KmJEm6HOgEzgLeCyBJkgocv632FBGPx4/r+8uyTDwep7+/n5Bh4JVKSBPsXsrHSTKVRqXvJcgTsxyeWX86p9edyl+6HuMLT3yJtVVPc+k1V3LB8nNZs+YMXxfXspk5q5nTzzh1xLGKohIMV0EhRd7MIEkQN6pwHAtZVvBcF0VXaaieWz7m0J35gpmnsGP/U/zluV+wbOE5hAKTq5G/b8tP6R7Yw8ffeQe65ns0diZLoW0/xa5erGSKwv5OGHogtKpKgo31SJqGJEFWkQmeONfvqhYAJgOJDjwh0FSdimjNCA5/X93NHBLe8RfpYDA6goLjYBRLeYqlPLKsEo2MfR8JIcjkEsiSTEW0FlmWsR0be0ia1HVlLMs6bFPdofA8zyeWG2djp0gyjmcRDh3ZeScLIQSIgwSMJOm4Kac5ju1zQCHhCQ/XdamMxcYcK8kywnWxs1n0WGxEWEl4HvaQGJKZTqOFQqih0GG9hSPtzrbtEp7rHHRfSbiey5H6B65p4lkW3lAV13StZYczCB8Evgk0AB89yDO4CPjDtFzRNKJQKFBVddwqZQGoqqoin89jlkposjzpGyykhUiaSTxx+KocWZK5uOkCVl61gqsfuQbtHTrrvvsMjzz05xHjvvXtWzlzzekj/qYqKpFIHM0skikOUrKKNFQ0URGsIptP4LrOmKGgYei6xuoTr+C+p27le/fewIqF57Fs4dmEgzH6k50MpjppqJ5DY+288jE9A/vY27UVgBdb17Jy0QUAaLEog3GXDb2becOb/xUJCSuRotjRSaGji9zuvQjPRXgCPI/slh3Ez1iFaG4gk8lRVRHzj3FMMrkEsUi1X5njumTzCSy7hCzLyJLs19ynCwSNMKFgbMQcXdchX0wRCEaGxIJGh6jggCaEJ0G+kCIaiY8g/lNUDcsuEGZyRrL8/p5z2KYfVdXpH0xSW113zEMgh/JNSUhlCc+XWihneHMCPjWX61kUi8VxjYKsaX4FUz6PFg6XnzWnWBza4Ojl/gnPddEjkTGfRyEEbrGINXQe7RD6ac91x2QiyBVSI+4lCemIyQGFELilkj+XIW2M6VrLJjQIQohdwKVj/P1h4OFjeSGSJC0GPgLUAH8WQoyvBDNFvByqTiRJorGxkfb2dhzPQ/U8PMdFQiCp6qgbSjgOwhMoukaFUUHSTBJSJ8eFXh2I88ml1/Ofmz/P+7/9bi6NvApd05EVmX+79no+d9MXuftXd1J5SIxekmQCgTCaFsDxbJJ2hpJdIqKEsdzihAYBoLFmLhevfg8tnU/z6PO/5K/P/2LE60EjynVv/RbhkL8wrtvyBzTVIBKqZP32P5UNgic8Hnji+/Ql93PKoguZ17wMoyaOUROn8uRlI85pJdP0P76WgcfXIVVV4MViJLv7kFQVWdfxFEExFKWieTZFSgjhERii5vAsGzuTxcnmGMhkUGNhak9cSsDwP+dCMY0sychISJJMsZQZ00solvLIioqmaBTNPIria0MMU0sMV+i4rjtp5TchBLl8EvUwn7kiyUhC4LhH7oEcHp6/+h4MCTxPMNVHyrQK6FrwiI2X6zllgyDLKqZZHMEIOxZkTfN32LZd7qr2xY78XbskSSiGgWv7pa56NFrufxBCIBynXMqqDJW1SrJc7nVwLctPjCsjVfBs29esGL6PwPdajtQgDEuuyrpepjKfrrVswrtMkqRvTvS6EOK6CY6diV+RVI+/vfieEOIbE5xrO/Avkq+leCdwzA2CNgWK4umAqqo0NTXR1taGGgr5noLr4hYKMCRpCQzRCfhxd+G6RLQISTN5RO910YwLeHLwKX64+yecf965zK70S94+/1+f5V3/9AG+9F9f5ZYvf37MB1NRFBRFwQDMYp5Bu0RQNVAPszPUNJWZtUs4edG5FIpptux+ipJZpCLsN6M99Mz3eODJH3H1q/6VQjHD5l2Pc8rii6iK1fPIujvpS7RTF5/F9j3P0JfcD8DmlieY17xs3PfUqyqY8bpLSW3bSeK5FyDTSbJt/6hxPYBWVUGoeQae7VDq7cdOpkaNSz2zkZrVq6g+aTFFs1A2Hpqmj+kluK5LySqUqSV0PUC+kPQZQg8O9UhgmwWU0OS8hGIpVw45HQ6qplIq5dAix9Yg+DHvQ3wUIRC4TEWF1/Nc8oU0SkQf09MaD0IIXNcuG1hVUYd0OCbeIA0r4Xmui5lK+d+Jqo4WLdI0f0wy5dOteB6uZfle/JBAEvjiRnY262thuC52Jous+8da6bQfnlIUCsUMyiHPyVQqtJxiCUlRytcrXHfa1rLDfRv/AmwFfgl0MRklkwNwgOuFEBskSYoC6yVJ+iOgALccMvY9Qog+SZJeB3wI+MkRvM+kkcvlqHkJefIngmEYNDc309nZiSPLBINBJEXBGW69lySE56FGoyBEufEmoASwXXvSnb227fLJFR9j/cBGPrv+c3z/nO8Q1aMsWnQCH/zQe/nfb93OeQ/8kde89tUTnkdXdFzPJe3mcD2baKBywvGGYTAwMIAsyyyefTbaQaphnQPnsrX1MZbPP5eu5E5cz+GMZZcRDET58zM/Z/32P3PJmn/m0ed/SU1lEzNq57NtzzouP/f9qBPMW5IklFlNaLEw2XyJhfPnoCLh2RaeaeOWShR7eil1dJPZtgtJUwnU1xJZMBe9qhI1GkYNhyns7yS1cSs9f/wLA888R3zNagJz5/jvgR8WONRLsOwifs/WUGkmks9MetCiIzyPzPOb6di0jdmXXUZ8yUkTfoaO45ArpA7oSh8Glu1SsoqEj/EO0m8AHMtDmFr5pOvaWFYJ17NRj8CgeJ43wixJSAgE2VyO6kmET2RFQQyFacfzKobHOMViuenu0GS+JMvIqoo5JGqkDMmqKrLsK99lMsjhICUzj36IIZdkGduefBJbuC6eZZZ7NqQh3e/pWssOd9c0At8DLgHeAWjA74QQPxZC/HiiA4UQ3UKIDUM/Z4HtQJMQYosQ4vJD/vUNjbtPCPEaYEwKT0mSPiBJ0vOSJD3f3d3NwMAA3d3ddHZ2kkwmaW1tpVgssm3bNjzPK3fzDfN+9PX14Xke27Zto1gs0traSjKZpLOzk+Hz7du3j1wux44dO3Ach02bNo04x/D/W7ZswTRNWlpayGQytLe309fXR19fH+3t7WQyGVpaWjBNky1btox5jpaWFpqbmxkcHGRwcJD+wUGKQCqbJZFMUpIk+gYGsFyXnmQS17bJJUxszy53ICf6/Vr8dDKP47hkM0Us06aQNykWTCQJVNPgppM/Q3uug4+s+wRdfX5VxuWvvYLlK5Zxy399lS2bt5JK5HBdj2y6gG055HMlSkWLUtEiX7DxhIxXUBkoJOju9xWbuvuTI/7vHUzjeR7pXBHdCGC54HiCfNEkmy9SNC0WzX4Vhh7mwad/xDNbHqSxehGupdGyu5v5zStZv/0vbG15it5EO2etuIqZdSdTsgo8/+ITDKYypLN5evoTFEsm7V29uK5La5uvf7CnvQtd1ylZNq7jsq+7D6Hp9JUsvIoKnJkzCZ67hso3XUno8kuJnX8OhaYmjFnNdJVs1HCIHj3AzLdchbtyBUgSfQ/9la4//JF9G7eS2Lmb3he20bV9B/u79tPV20synWZPexuuJ7Ovo2eoJ2A/siTTsqfD/6537aH7gT+R2fgikq7R/sgjDHR1093bS+/AAIlUivauLvKFAi1792LbNpt3vIgiK+weOseuPfvLc7Rth47ufvKFEn0DSZLpLDISA4kMqUySPW1tmJbF9pYW4ADP0fD/W3fuxHEcWvbuJV8o0N7VRSKVondggO7eXlKZDPs6OiiWSrTu89+3Ze/+g65D4sVdLXiex87WVoqlEvs6OkhlMuPOyXEctu7ciWWb7Ovox3Gs8vVsb2nBtCz2tLWRzeXo6O6mP5GgP5Ggo7ubbC7HnvY2bMdhT3vXQdch6B8cmPSc2jo7MW2bna2teJ5X7u4evo7N27cjhGD3/v2YrktbZ+eYcyqYJns6OxGyzIsHfcayprFp+3YyPZ3saWnHKpm07+sgk0jRsbedgc4eEgMJ2traJrVGbNy4EcfzynPq6Oujr7cXz/Mmve5t2LBhxLo3ESbdqSxJUjPwVuBjwKeEEJPexUuSNAefB2mpEGLMGkhJks4H3gAYwGYhxP9OdM6p9CHs2LGDRYsWHdExLwWEECQSCQYGBohGo36oSIgRfQrCdXEyWVwFOnIdhLXJdbxmUgVilb5L/afOv3LjczdxRt2p3HrGl9BkjZ6eXv7l/dcxOJDga9/40pg8SMLPIIKqgW2Tt3IUrSxNsZmo8kHXKPymm8PlGAC27X2KPz9/JwCXnv4hZjcsQVUVWjs28dAz30VTDaLhOB9+i0/D8bU738+cpiW85dUfH/eclm3T1dVFOBymdyDJzBn1VIyTbJwshOuS2ryN5PMbR4kcVZ1/BlWLFxMwQqSzg9h7O8ls20l43mwiC+ejhoI4hQL5PW2kNr2Ik8tTe84ZyLVx+n77ENFZs5h39RvGDNcVilnyhdSkKMiH0d7Vy4z6OIqiUhGtPap5H4xsLoHlmCO8UtuxMPQQkfDEnuJYGEx1A75qYGVs8h3qY1Gl265NT2+SJSdO7G29lPA8l4FkJ5o0/BxIIEBW/OIFs1SgtnEORjg6LmMB+M+dmUyOohtxTZN9/f0sPmnKcx430jMpf02SpFOAa/B7ER4EJk21J0lSBLgHv0pp3IJ4IcSjwKOTPe9UsGDBguk8/ZQhSRLxeJxsNotlWei6PuobkxQFOWCAaWIoARzPGbEYj4dI7MDDc3HTBeTsPF/Y+CVuWv8F/mv1zTQ01HPHD/6XD3/oY3zkXz/BF//7c5x/wTkjT+IJUFQkVUFYFqFAFNMuMJjvpy7SgCRJCOFhD4m++5w3Ezufi+ecybZ9a7HtEgtmrigvigtmriC6pZpsYZAls8+nq9PfDZ44+zQ2734c0yqOu0gWi8VyqKQiEqZYLB61QZAUhaqVy4ieuAA7mULWdWRdo+/Rp0g+/gyKYVCa0UB+6w7Sz21CCQYZXPscg08/j15VhTXoe1JaVSVNV76GQEMdpl2k+oxTGXhyHYmtW6leNjI3YtulIwoVDaOpvgZFUTCtI0tcHw6e546qbJNlBXuScqYHw3EcPNdB1wNYQ30Ek00s+yWnI69DkRXqal5eJIfFUn6oD2J0CbMESLqKXSggTLvM3STr+qgwlpPP++GtQ/MsksSc5uZpufbDUVd8XpKk9fhewWPAaiHEe4UQk+IWHuI+ugf42fFUVxvGiy++eLwvYVxIkkRtbS2l0vgSkbJhgCwTkYLYk2xwSifzI36/as7l/Mui9/FI559Z1/cMADW1NXzv+7ex8IT5fPzfP8273/lB7v75rxgYGPQPEh4oMsMlJbIsUxWtp+iVyBb9EJFtm4SClQSDUWzn8NcmSTJXnfsRrr7gEyMWBEmSOe2k1zKzfjHLF55DKBQiEAjQWLUYx7XYvveZMc8ngFwuiz6U+EvnCliWdcw42tVQkGBTI0ZtNVpFjMbXXIQejzP45ydJP/4M6ec2EVk4j9nveBMz3/J6KpcvQdY0qlafzMy3XMWst76eQIO/G9a0AIETZ6M31NLx5z+T6eosN885jsNg736yz28h9exGil09fsFBsURu9176n1hHZvuuMZvt9nUMVYVLAtuZnNToZOB57ugErKxg29aEBINjwXFNnz4bCQHlRsLJHWuNKmaQJZm2zp4jvo7pgmWXyBdS5cT3WJAkCVRfgEgIgZ3PYyaTZRpxADufxykWx2R7lVWV7dNE3ng46goP2AsMk4IMD5YAIYRYPsGxEvBjICGE+OixudwD+Fuiv54shBC0t7fjeR6GMXaliPA8Stk0nck2QqGpkXfZns3Vf3obUS3KT86/A8kVIEsUSyV++Yt7efjBP9KyqxVN0/jGbV/h1FNORgoFQVUQ2RwMVTw4jk0y30eNEaciXDtEsOaSyvQPJZGPXWLTcRx+8tBnqYvP5F2v+89Rr5uWRXd3N+GDJCLz+TzNzc2ox2inPOqaCkW6fvsAdjpDxYolVJ956hF9H1YmQ+c99+OZFpFFC6g94zTSLbtJPvsCwrH9p034nbnDDXnDP0cWzKX2vLPGFFwa5qKqqjgy0aTxMJDo9GU6D/FbS3aRqmgt2iTFjcAneXNcC1XRRim+HQ6DqW4URS3ragzDtIpUVtSP21T4UsFxHJKZblTVGJNiZBiWbRIKxkYIGw0r1cmGgayqOPk88jg04uCHjQI1NVPtOZmyYtrsic4qhGib4NizgSeALcBwOcKNQogHJrzUSWIqBmH9+vWsWjU5rYDjhWKxSFtbG7EJQh1CCDr694Bl+/F6RR73xhjsSxOvHW04/tD+EDdt+AJfXHUzr2rw6/4l7cDCuad1Lzd84rMkkyl++tP/o37ebCRFwcsPlcYOeQsluwhCMCPWXA5hFUsZimah3HF8rPDUpnvZsOsR5jefzJL5Z3LS/NMx1BC5fI50Oo2iyOi6/57d/UliIZ36+vpp4fcZhlMoYA0kCc1qmtLxbskk8dwLZF4cIisTgsCMBmrPPRMlFKTU3UuxqxfF0Ak2z8CoqSa1aSuJZzegVcSoOet09Jo4SjBAy94OTpg306dLGBxAJHO4iTSyLNNwxhkjtLH9Ek7Xb8yboCJJCMFAomPMMN2R6lJ4nsdgshNdD5QbBoNGlHDo8GE9z/MYTHWOKbW6ffdeTl6yBEOfXH/OdMDzXNLZATzhHbYC0HZtdNUYs5fFtW0Y6jmYaLHfuGULp65Z89IahHEP8rd+1wghfjaVqzkW+Hv0EIaxf/9+SqUSkiQNJWmVUURWKTNFItdPQPiJXp+MXhlxgwjb8XeUQiCpI3fJrnC55i/vwnFtfnnJXSimg3TIbnPvnn288+3vZ+HC+XzvZ3egGRqeaYFpIh0U1yw5JVRZpS5UhyqrB3kJE9/URwrLLvH01t/R2vkCuWISRVZZufASls+/kGAwNGphKxaLVFZWEhtD4vHlBnMwSWbrdgKN9UQWzjs8ZXlnN71/fAx3qGpENnRkw8AzTf87GoaigPB8zqhzzyI0bw6ecLBtE4GfC4iEKjH0sZvEXNclke4eU+LUEx6e5xKvbJzUHG3HIpnuLfdUuEOhqPESyyUzjyIrfoOk45DM9Ix5Hb6samBcSpGXAvlCmmIpOylv53DzngyOl4cQAz4MNAH3AX8E/hW4HtgkhLhyKldzLPD36iEAmKZJIpFA13V0Xaenp4fQITwrpmuyP7ufsOqHSIRlQ8n0vQVZ9o2BppJM5KmKGqOMBcCjnY/x8ec+w2dO/QxX1b96TE/j4Yf+xGduuJm3v+vtfOzGjyEcF5EvjvAmYLRRKJQylKbBSwB/19qXbGP9jodo7XyBuqpZXHTqu6ipOLBL7+5PUl0RRtM06mqPXcXNywmuaWH29WMl09jJFKlkmsrqKhRDRwmHCNTVosUrKQ4OMPjY09gDCfSaOAiBWyzhlkoHhaUktHAELRbFqKgkNm8eFfPnIxR51EJsDgyS3rKdQEMd+rxmqquaJpXAzheyFErp8rnEsMZ2VdOo+851Hb8aSQjCoQo0zSCV7Scwhoewc087C2bPoKqy4eg+0ClCCMFgqhtV1UaFs8aCh8CxLWriU/Mq4Th5CJIk/Q5IAuvw+Yvqhk72ESHExqlcybHC37OHcCj6+vrIZrOjvITB4iApM1UuQRWuv1jjeaBpSKGA72WYFpRMJO3gElGBsB3e+8y/sS/TxsnVy6nVq1lYuYA3zL0SRRrqlvY8vvrlb/CLu+/h7e9+O9d+9EPoljPiXMMwHRNFVmgIN4AQpHP9SJKEqkxfbHd3x3oe3XAXplVg5Ymv4tTFl5WpGzzPw7IsZk5TRcbfEoTnkd78IrnWNpSggRIKoQSMcujPdR2cXB43X8BOZfCGumMjc2Yhx2OE6upRwkFSG7eS29VazmXojXXMvvRSwjV1mIkE+a5u7GwG1zRxTYvIzGbiS5b4vSnZfpBGUniXrCLxisZRHcvZ3CCmVUIbYjVVFRXXdUaFroTr+noJdpHqyqlRix8tbMcmle45ohLho73e4+UhbBFCLBv6WQG6gVlCiGNXwjBFTMUgbNmyhWXLxqc/eLmiVCqxb9++UXkFIQS9hV4KTqHMbyQ8gbBtJN3vDE72JKmsr0Rk8yAfIN8Stgu6yh6zg9s23kZvoYe+Qh8pK831y67jmvlv9sc5LrYEX7v1W/z6579m/sL5fO6mGzjhpBMpFkuYpkk8fqBLtOAUiGpRqoPVuK5NNp9AIKaBX+cAimaOJzf9mh1t64iG4py94k2EA3NorImTz+eZ2TwTRTn+PFbTjT3tXcybNeOoz+MKj0JnF4XWNortnbi5A5VqkqJQsWwxlacsJ9+6j4G1z4HwkFWfeG0Y8pCwjVsqUb1iOdHTVuAIZ1TIp2QVqYzVjgi12I5FItmN5ko+hQR+3F2WlRHGpNDeSfeDf4KmGcy44HTiNTPR1JeenqZQzJIvpscMZ40H0yoRr2yYVM/OWHhxxw5OPvXIihgOwpQNwgYhxCnj/X48MRWDYJrmuNU7L3fs27fPFxA/pAzNEx5duS5szyaojt6huI6Loip4lg2Fkl9oLABVRQoGkOQhfhRP4GWyfGz9Z3h+YAO/uPBOmsIzELaDFAoiaSprH1/L5278HIP9g+X8BsAVV7yGz9z0qfIuL2/nqQ3UEjEieJ5LLp/E8XwK4OkUI+nsb+GxF+5iMN2JJMnEwjVUhOu47Jz3MKN2wvqIvwvYtoM2hud2NPAQuCUTJ5HCTmUIzpyBFj0gnWnlciSee4GAHiI8YwZqbSUl1UPTdHQ1ROKZ5xhc/wJGQx0Nl1yIGhp5jx6aWBZC0Ld3BwOPrsUaSBBbuojq01eVieiG4eQL7P/l75BkCbdQRImGmfmaS6iaM788RngeqZYWElu2El+yhKrFo5tSjwVjayLVjXSIsTocTLtIZWzqlVHFXI5YY+NLbhBcYHh7IOFrIBQ4UHZ6dF0/R4GpGISWlhYWLlw4TVc0vchkMvT09BCJjNaxdTyH/dn96LI+ikwrM5AhVhPzQ0TFEkiy7z2MsWP28gV6Cj289a//zJKqk/jfNf8DjosUCZU7KtOpNHf/6Oc4RZNILEJPfz+//Pmvueji8/nCLTehab4WdNEp0hieQUA1hiQEc5hmHiR5Wg2D67ns6dxIR98eSlaStp6tNNbM471X/b8D5GBC8NiGe5hVf+KEhHl/a+jo7qe58aXPl5hWiaqKBoRwSKb70Y2Az7/lWAgE5p5OBh9b65OyVVUSqKsh0FhPeM4sMDRfPyJWi53P07VuLYkXNqEEAgRnziC3qxUlHKL2nDMIzZnlN0F6Ht33P0Kpt5/mN15BT0cP3sbNuLk8RjxOoDqOFomS3r0bO5st37sLr3kr4Rm+B+W6LvlCCtu1iFc0TPl+dByHZLr7iMJF4BuEWKTmiI8bxu49e1i8YsVLaxBezpiKQchkMhOWc76c4boura2thMPhMW+CrJmlr9g3itLCKlnogcntQvxcg8U9HffxpU23ctPKG7l8xquRouFDxEWGqCyGktA9p8cdAAAgAElEQVQ/veNO/ufL32DNWWfw5a9+gUAwgOM6lDwTXdLQVZ2gEiSoBLCcIqVSDlXVp5VL37RsDF1jw44/8dSWX3HNpTeweK4vCPTci4/w+8dvR1V0/vmKm5jd+PKjM5kK8oUS4dD0ldiOB9MuEg5Wki+mxq3BtxJJcq37MPsHMPsGcIslkCQCMxpQomHsvkGshM+HFVl8AjVnrva7wHv76X/0KaxEEj1eScWyk3ByBZLrN1J7/lnEFp9AvlBCVyCzaRsik8dMJDFTKcLNTdSdsopQ0wx2/eSneI7Done+E1eXyOaTSEMiO1UVdUcc0rSyWVI7duDKYKse4epatIrJry1j9SIcCVKJBLWzZ//DIAxjKgahvb2dWbNmTdMVTT96e3vJ5XJjaqkKIWjPtKPIyghKi3wqT7hycrxHwnURuQJClfngk//G7kwrPzrrdubMmHjBFEJw7513c8sXb2XlKSv4n29+mfBQ7NcTHq7nYrkWMSNGPBDHdixy+QG0KfDhTxbpXIGKSAjXc/nZQzchywrXXfMN0vkBvv2L62msnUeukCJfTPO+13+Buvjf7n0xjL6BJHU1L70AlO3auI6DqmmTolMRQmANJsi17iPfug+nUECvryXY1Ehk9kyM6pHlo8L1yO7aTXrrdqwBnwoksmAedRefiyRJ5XmbVrGsWX0oJUaxr49dP/s5ejxO/NJzMYIhZEnGckwCeviIOJnMVIqWu3+BnRnJxFN38XlEF84b56iRcDwHCb/0dCrPQPv+/cw/6aR/GIRhTMUg9PX1UVc39drf443h5HIkEhmzOiFn5ejJ9xDRD4SVSrkSgcjkdo1CCD/5rMjsz3fwnsf/BVmSue3C/+XEqhMnPNazLB76zR+4+eZbOGnJIr5x21dGeWMFu0BQDVIbqsW08hSLuSPm65ks8kWTcNDf9e3p3Mgf1n6Hi097Jy37n6NnsI0Pv/lrCCG44zc3IkkS11z6KRpq5o4Kub3c8OBTPyQeq+f0ZZeNei2ZzlJVcXx6LlzhHVEM/WBMls9ICEGpp4/i/i4qT15SzisMz9vvWG4YN7Hcu2UTXQ8+gqSpqOEwSig4RAcjEQhH0cJhAvE4Rjzuax7kcr4YjqoSmzMXxdAx02l233U3rmUx5/VXkXfzKLbH4DPrKfX00viaV026SbFkFYmEKiftJRysVtjb20vzggX/MAjDeCUaBIB0Ol2maFAUBSEExWIR27YJhUL0lHpAgKZoOJ5DPpsnFps8xYVXLIFpgSKzL9vGvz7zCXJOjm+c9w1W1q0c97hhY/Lo409y46duZu68OXzztq9QUzuSs73oFNFkjYASwLFKCM8hbBz7RexggyCE4N5Hb6V7sBUhPF533rWcfOL5yLJMX6KNH/z2s5SsApqqU189hxNnr+aMZZchSSq5XI7KyspxBe5fSuQKKb7y4/cRCkT5+Dv/b1SFyvE0CMcTZYMwQVy+ZBbI5AaxO3oxu3pw8kWcfB7PsvEcG8nzu84ZZz2UVJXY3DkUenvxTIu5b7waraaSbD6JoQVwTYuu3z2Inc4w43WXEqg/fC7HQ2BbpQmN2DCKmRTZQpLq2lkoivIPg3AoXokho2Hkcjk6OztRVRXbtqmqqiIcDtPR0YFsyHQXupElGU3WKGVK6FGdgDp5LwHH9RvdbIdeKcWHH/8IPYUebjj1Bq6Ye8W4N6FnWVAose7p5/jEJ/4DzxNc/OoLefNbXs/SZUvKxzmeg+u52J5DsZAirlehK5ovP3mM+GiGQ0bD6Eu28Ys/3cKsupO4aNV7fDZKRaGmpgbbLbC3cytd/a109LWwv2cnAT3CkjnnceLMM6mva6Sq8shpnocxkOrixda1nLn88knz9oyFZ7c+xP1P/B8Ab7/sRk6cPbLJ8niFjI43huftuDaeEMQicTTVFydyXZeSmSNfyKDrxph65JZtEgxECBoRrFSK0uAgrmWjRSPo0RhOPs/g9u2kd+3Ec11qX3M+ek01CIGiHgiTOYUCnb95ALdYwqitRgkGUCNhKpYuRouNbaht10aRFSqitWM+V8Lz6Fu/nu4nngRZou6sM2lcfRr7Ozr+ETI6GK+0pPKhKBaLJBIJ4vF4OaeQzWbp7OzE0zwCWgBDMUjn0gy6g5PWTzgYwhMgQdJMcsOTN7C+bz0XzbyIz5z2GSqM0fw1w8YEoG1fO7/4yd384fcPks8XuPhVF/CFL96EekhznGOWQJaoNqowzQKO62AEj56TZjipfDAGUh1UROrKRsdxHEqlEvF4nFgshuM4vnBS+xae2/4AHf3bkSWZeKyJuU1LqK5swLSLWFaRaDjOaUsvnVDBzXZMHt9wL0++8Ftcz+HM5ZfzmrPePeU5/fB3N5HJD1I088xrWsqbX339iNePV1L5eOPgebtDNOy6ZqCpBoVSBmlYvW6cdXA4zzVcbWQ7NulMH0J4yIqKIitYtoksK6jIo+moD4KdyZJ4doPf5Fco4uTyIEnET19FxdJFIElYg0mKHV0I1wFJxhEulXMXEJ81Z8QCX+juYf8f/0ihpwejuRHJE5S6egg3zaDy9NNpWrnyHwZhGK+0stPJIpPJ0NXVRTQaRZIkuru7cSK+sIsmT71px/VcfrLjJ3x707epDlRzy9m3cHLtyRMeI4QgN5Dgrjvv5vbbf8AFF57LLf/9+bJRELYDQYOCZDEjPIOApJHubccWLiUsNEknNEXCskQ6R7xidInuWNdYKBRQVRXHcZBlGU3TUFWVnsG97OnaSPdAK33JNhzX5wjSVB3bsairmsmVF1zLzPoT6OzbzdpN99HasYWAESIcqCBbSJDK9rN84TlIksTmlif54NVfYkbt/MNc1WjkCmm+cuf7OPeUN1Ay86zf/ic+8c/fJ2gcMPTHq+z0eGOseTuujeu5aJoxKTqJ4fyDLEkkM33IsuzzciEQnjfl3JKdzTHw+FoK7Z3o1VW4JRM3XxhzbGBGA/WrV+MWigxu2Uqxtxc1FKLyzFMIz/fzW8lt28k8uwmhaaz6+PXjSoEeBv8wCPC33Zh2JOju7qZQKBAMBrFtG1OY9BZ6p+QlHIrtie18+slP013o5lOrP8UbFrxhwvHD/Q93/+QX3Hrrt7jgovO44cbr2bxxCxvWb0QLBbnyLVfS2NTIzNhMnGKBtq5tFEQRWdIIqAEqjUr0IwwlOa57RLTXhxOUyWYzRGNh4lU1KLLCrrb13PfY7WQLSeris+gd3Iehh1g85zQczyZf9CtQzj3lDcxvXk7RzPOtu68jGorzgau/dMQLzHCp7LVvuhXXc7j9nk9x5XkfYtVJF5fHTEdj2t8CjsW8TbtEwIjgOOakGEuPBEIIci17SG7YjF4ZIzRnFqFZTSgBv1/DNS1yu3aT2rINN+cbi0BtLdXLlhFaOIeCUygTAlqOierIiLxF3bKl//AQhvFKoq44UpRKJdra2ohGo7S3t9M8s5m2TBuGMnYM9UiRsTLc+NSNrOtexxsXvpGPnfIxDGV8QyuEQBRK3PWTu/na124r/90wDBzXwXM91py/hms/fC0Lly6kkBpAcvx4sOmaeAhqgtWE9eiIB8Bx7XK3tBACVVHLida+RIa6+LELD3qeR6FQIBaNEY6EMQyDXD7Nw2vvpLOvhZWLL+S0pa8mMIFHs7V1Lb985FYuXfNu1qy4HMe1KZZyRMOHj/v/6L6bSecGuO6abwHwrbuvIxys5L1X/b/ymGNFXTEesvkkkVDltHabTwXHYt6e8IY4k5RppVmZCMLzyO/vROgKek0cCfA8MSL3IRBYtklfX+6lp654OeOVRG43FRxKdZEsJUmZqTK9hWmaaJo2dXItz+W2Tbdx5/Y7qTAquGreVVy98GqaImOX3A17Cn95+M/sa9vPKatXsvS0U0glUvz6rl9zz933kM/l+foPvs7pq1YjcgWQfS3akl1EkxQCio6mGQgBjmOhIKFrQRAgJCgV0mihyKiH5EhkGieCEALLssqhJc/zynQdgUCAmurqwx7/0we+yL6urVREakike/CEx/mr3sSFp7113OPyxTRf/vH7OGflVVx8+tsBeGz9Pfz52Z/z72//DlVHQaM8WaRzg3z9Z9fymrPezWlLL5329zseOJrS2emAQPja6odcU8kqEjFiVDQ2H3OD8PKZ/UuA9esnLQX9N4/q6mpM06S1tRWAiB7BFa6/MAuBbdsTynUeDoqs8JGVH+H2i25nVd0qfrrjp1x535XctfOuMcdLkoQUDHDhpRfznne/nZVnnoqu69Q11HHtv1/LvQ/fS2NTIzf+243s7+iCYAB0HSkUxKioxJUEASOGbZs4jk1QixCN1BCuaSBSP4NIbSMYBsL25Tu7+5P+XC0bYTtjSk4eKSRJwjAMwuEwgUCAcNj3FHRdJ5fL4breYY+/4tz3M6N2PnVVMzl75es5ae7pPLr+V+zY+9y4x23f+wxCeCyZv6b8txUn+LrXm3Y9Vv7brj37j3KG42NX2/O4nsMLO/86be8xVRyreb+cjAH4MqNjefSqqrG1pWVa3vPl9QlMM/4WtBCOFcLhMLIsM3fuXMBPKFfoFRTdIpZlEQwF8cTEC9hksLp+NV855yvcf+X9rJmxhq+/8HV2JneOOVaSJKSAgRQOl/llhlFRWcE37/gmkiRx3fuvI53PIwcMJE1F0XRcQwU8YtFaKiLVGHoANRIpU2pIkkQwVoUrCYTj0FBT6RPzBQ3kgAGOc9RzPXQuB/8syzL5Qn6CI3xURut471Vf4K2XfpKLT38bV1/8UWbUzueev3yTgVTXqPGu67Bx5+PEYw00VM8ZcZ75M1fw1Kb7GEh2AnDCvJlHP7FxsLPN30x19u1mcIzrPJ6Yznm/HKHKKnMa4sdkk3MoXlEG4ZXkIciyTDweZ/fu3eW/VQer0WSNvJmnqrKKUDCEZVkTnGXyqAvV8fkzP0+lUcln134W0zXHHCdJ0ij1tmHMnDWTr33na/R29/KBf/oAd3z7Djau34g9pDVblBxkV4AnkEOhUUbFMEIIQwdJorsviRwKgqbiqoCqIg4yCkL4hkO4LsLzJvVwTTRG13Wy2QxH+ohqqs5bL/kEiqxy90NfJp0bLL/W1b+H2+/5FO092zl16SWjwgNXnvchFFnl5w/9NyWrMG0egu2Y7O3cwklzT0dCYlPLE9PyPlOBcF12to6r5Pt3i71dA9NyXuXmm2+elhNPN773ve/d/IEPfOCIjpkxY/oSbi9HDJdS6kP6rJIkYSgGA/kB6mvr0TWdVDqFfhC1sOu5SEhTik0G1ADzK+bzs50/o+gUWTNjDZ7w6Mh1oMoq+iSEchoaGzhh8QlseHYDD9//MPfdcx93/+RuBvsGidZVMqeuGUU3/AqNQyBLCqbti5PHKqLImopl+6pgqhHwm+08r9wrIWnaUDRVKsuNjjdv4Xn+GIkRRH/l95ZlSiUTQ9VQFWXMMeN+bkaYxtp5rNt8P2s33cfGnY+yt3Mrj6y7E4Hg6ouu45RFF4153Iy6+Ty9+Q/0Jzs4d9UltHZs5oGnfsAzWx9E1wLUVjYhHWUopHX/JjbufJRLz3o3mXyC9p7tnL70suOWXHaGtBEAPNeluqoC4XpjMvj+vSJqKESr66f6HXxuvBdeUQZh06ZNNDQcH5m94wFFUWhpaUHX9XJy2bEcYqEYtmYTMkJkMhkUxZfXdD2XglvA8ZxJLd5jYWZ0JqlSirt33c3Gvo18bcPX+PH2H7N1cCuvnfvaSd3As+fO5o3XvJG3/NNbWLZiGa7r8tB9D3HvXffy2Nq1NDQ1M3fu3FHnkiSfvbJg5envSxEMahhGGCE8JElG1jQ/SRcMIAUMZE3z/+kaIMB2xlxUfG/CRQ4F/RzFOIZDCIFtmoSCIaSJjMuQV3Kw0YjH6lky70ziFQ24nkP3wB4WzzmNt1/2aRprxidMq4rVo2shnt7yB57Z+jDrt/+RklVACMHz2x5hy+4nCRlR6qsnrwcx7AkNX/+6zX+gL7mfy8/9AEiwceejnDDrFGKRiZPo04HWjk1851cfpz4+i5qqJoTrsa93gKpo2NcOH8MQC89DOO7flcFobe9i5uzRz8Ak8Q+DAFBTU3NcJPaOJ2pra0mn06iqiiRJlEolmhubUVSFrJVFldRyY1bBKVAfqqfklpCRp1yiuqp+Fc/3Pk/GznBGwxksii/iL/v/wuLqxcyOTX5hCgQDzF0wl4suuYg3ve1NVNZU8vy657nzzjt59NFHaWpqYtasWSMfCklif3IvkgFhLUQsXA0IbLuEovoGYMzFWpYRljX2Yu84SIaOEjD8xXIcw6FIEnnLJFpZieS4YzYNCSHwHMd/D0FZoAggHKxgZsOJLFtwFqtPeBUzaxcTDkYO62001y8kX0zjujYXn/52rrrgWs5Yfhn18dl09LXw7IsP4Xoec5sOX7fuOY6/gLqe39krwR+e/D9mNZzIySeeR1WsnnWbfo8iKcxvXDo0bmyvaTx0D+yld7CdeMWRb85+9+h3GUx30dqxmeULz8XQDKob6lEMA7dUKnvCI+ZkO/7nfMjn/beMoCpRUTNlHYdxDcIrqotl9+7dLFr098F9P1ns27ePxsbGEeI6gUCAkOzT//ZYPTiOQ8EpUBWoImbE8PBIlBKE5Kl1CQfVID989Q/Lv9uezQt9L/CNF77BmsY1k6JIPhQVlRW8413v4Kq3XsUzDz7DN7/5Td72trfR2NjIa1/7Ws4//3za2tp4+vmn2blrB+99/7u44rLXIcsKmhagaOYmPL8ky0gBA1Gy4GB6DdcFWfYT04Bs6Hi2PWqHD/5ONBCK0JtMUBeJIrvyqDyHcF1M10NWFDTXQZa10WWynkc2lyNrm4SDIZRxvA3X9bBsi2AgwBXnfoD2rl5mzagvv75k/pnMbz6FXz38dR7f8Gssu8ClZ727bOgd16Zo5iiZeTzPpTrSgKwb6OEgwhM42Rx9g/tJZfs59xS/ATFohDlxzmq2tq7lsos/hAy4po1nWSj64b3Kopnjp3/4Lyy7xCff9YMj4q7qGdjH3s4tnLLoIrbsfoLf/PU2/uny/6Cjo5s5c2eiRSPYmSyyqpa/G8+2UXQN2TCws1mUadT3finRm8gwHan0V5RBaH4Fiq03NzcTDodJJBLk83kikUi5I7c2XIsqq/T19xHwAlQF/AapqBYlUUrgCe+YNLJpssZ1K6/j+sev57etv+WNC984pfNIkoSiKlz8+ot5/RtfzyMPPsL999/PnXfeyR133AFAvDqO53n89xduZcWqVSybFy/zDR2uH0HSNETJKo8TrusnsCOh8nGSLCMHg7iFPJJ8YHHxj5HRQwFs26E3naI2HCFwkKCRJwSpVIqUZaGoCk119WMupJZpkjFLSJqKLUnIjuPnO4ZQKpXI5wtkc1nA/45VRaE2PpqAr1AocOZJbyQSjvH0lgfoHWxHIBhMd5PNJ0aMndO0lEvPfx+zYouRFNAqouzashGAhbMOVOgtnbuGbXueZk/nZk6cdyqKYeDkfOZQ+TAdww88+QOyBV8IZ/f+jSyee9qE4w/G2s2/R1MNLlnzTprq5vP7x7/Hs9seZtXyywFfx1mNRHByOWRNAyEQSCjhEEgSkqoiXHeUkZ4qjlV/y1QQjx0968BYeEUZhIGBgTElKP+eMTznhoYG9uzZQ0XFSFK6qmAVK+etpLerF8f2E9CKrFBlVJE0k4TUoyeaAziv6TxOqTuF27fcziWzL2F7YjuPtD1CTI/xgWUfmDQba1gLU3SKFNwCF772Qi6/8nLSmTQbN24kXB9mZvNM9rXs5V1vfjefuO4T3PPrewgHwuhqAMezJySj870EHVE08SSf8lgOGaMWEElTkTQd4ThIw0Rnrotk+Ml7Xfd3/T3JJBXFIkYohKbrpAcGyToWsXgcy7IYSKepr6jEtUw/lCVJCNfviFYMHVVVKboOhmKU8w6WbdHT3Yuia4RCIUql0lAHdZR0Nk8wcKDL1nU9stkskWiEUxddQbyyjudefIiKSC3zm5cTjzUQCkQxlCB5J8sTG37Dd3/2UU5aeBYnn3QB82evZHfnJhpq5hLRfRlWPI8T5qwmGqnm/j99m9n/fBsBI4wSDOBa1oSL5I59z7Fp12Oce8obeH7bH9m6+6lJG4RsPsmWlidZfdKrCBoRVp/0anbuW88fn/oxVbH5LD3Jl3pXDB0I+6RygBo5oPanhkLYmeyEFCVHAs/2Q38HG8G+xH6e2fogF512DaHA9FGRZwtT7yGaCK8og/BKMwZwYM7BYJCmpqYx1dbi0TjhOWHa29sBv4QypsdImMfOS5AkiY+u/CjvfPidXPKbSzBdk6AapOgUear7Kf777P9mTmzOpM4VVP0eisHiIIMMIskS81fOR5EVNEVj4aITuOE/b+DmG2/m5s/dzFdu+QqGEcLKD8JhOGokTUPYDorh90CMBzlg4OXd8o5TCDGCBVPTVOTKCgrFIpm+PoTrIqsqsdqacoNbLpcj51hEgwHcYglZUbAtm4xlEh5i5c3lc8Qrq/DyOeSAQdGyUMIhDE+UO9Gz2QyxaJTAIeyuw30RsiRh2w5nnXwV5626esQY4boIJLSKKKeveh1PPn8vTz1/L9tanvKrk4TgvDPeghIK4haGeHYqKnjrFZ/m+3d/knsevJW3XflZJEVBCRh4JQtJU2nv2cHD6+7Etk2WLjiLhbNWct9j36W+ejbnr34zhVKWzbuewHbMSVFFPPviQ3ieyxnLX+t/T5LE6875IP97z/U8sf6nLFl8gPlTMfxudmHbZREdAFlTkTQVz3EmZCydDIQQ+B+PKBvB/mQHP7rvZnLF1JDuxr8c1XtMhEO/62OFV1RSOZ1OE42+sgREDp5zIBAYd/emqirhcJhk0u/w1YYqcnJODl0+NnHXulAdrucS0SJ8cNkH+Y/T/4OVtSu5f+/9/Lrl1zSFm1hQuWBS55IkCV3R0RUdTdbQlJHyjbNmzqJklfj5j3+OETA4dfVpmGYe5XAGQZKQdf2wFSnD/RTCtAGBJEvIh5TCyrKMpuvo4RC6rqGFQiMSzZqmkU6nCUVj6KEQnu2QL+QxEeVSYcuyfO6kaAQhy/QPDBAMh8BzwfNQVBXTtAjoOsWiSUAbip9LEv39/WiSjCzLOKZFQDfQjJFhLs910KJRZEVGVTXmzVrO2adezcK5q4hFqtH1IOef8VYi0SokRUbSNBRdp6qiHk0LsHb9bwkYIWY1nYSkKCQHu3jgqR/w4FM/ACEIGhE27nqU57c9guPavOOyG6mI1KCqOht2/JkZtfOorZo4lGs7Jr/+09eZ37yc05e+xr92z0NXA4Qrq3l20/3UVs+koXbOgc9eVVGGPLYR35ui+MnnCarAJgPhusiajqzrCMsmke3hh/fdhEBw4pzVbNr5GAtnrZy2SqxsJkvt1Kkr/pFUBspEaK8kHMmcA4EAs2bNoru7m1wuR0WwgqJTpOSUJh3SORyuXXHtiN/XzFjDXa+5i08/9WluXHsjO1M7+fDyDx+1lKXwBNd94jq6u7u55Yu30LavjU9+6jo8zy3XsB8tJEVBDgdxc3m/CW68cZKENEbfhCRJBAJBuru7kGWZSCRK1rZGeHGqqpLL+jramUwGWfYXeBEI4ObyflWUEBQLBeRgECUcws0XKJVMnJJJoKoSKWCgyjJFy8KwbCTpgDCYEg4hH9IoqCgqc5qXMqd56ci/H8IUfPapV9PetY2HHr2D3fs20NO/j0xuAEVWOXvFVZy/5hoCwQh9Ha1sa3uGqmgdjbV+Ce3sxpMIB2JsbV3LSfPOGPNz82yHQinDb5/4LoVSljUrrii/JjwPWdM4dfmlrFv/ex786/dYNO80DGPiEKesqiihEG6hMKkk+HgQnudvHFSFRH8HP/z9zbiey7tf9zkqojW0dW3j949/jw9e/SVkWWF/7y6efOE3vPqMd1BdefT9UN40cdC9ojwE13XHDJn8PeNI56yqKhUVFaiqSjqdJqAGKLgFBAJFmh694YgW4bI5l5E0k9y18y52JHZwdtPZ5O08e9J7KDpFKo0jUyzzhIdu6Fx4yYWU7BI/+9HP2Lx5K2efs4ZQ6Ngl5CRZRlJV/98UdmuKIvt5G0WhVCqV9RgOvK6Qz+eJRqP09fWVvTxpSEgey0JWVExFIhqrIBQJIal+oYAaDqGFgr7Xo8hYnktVdTWypqEEAqjBwFEtipIkccLc1exue4FCMcusGYtZueRiLjvn/axYfhF6IIgkywSDUZor51I/tIP3bAfJEyRz/by4Zx1nLr8cRVHp7t/DCzv/imkVCWhBehLt/OTBL9A72MZlZ72HpQsOcDkJ10UOGCiaRmW0kWc334/nuSyce3h6GklVQIBrmlPVE0B4LmrI7zd58Inv09G7i/de9QUaamajKhqxSDXPbH2AYCBCS/sL/PYvt9Gf7EAIwQmzj55CxyqVqK5v+oeHcDRIJBJUVb2yJAanMmdJkqisrCQYDLJv3z7qQ/V05juRkadNhF5TNG487UYWVC7gq+u/yoW/vhBXDHUUI/Gdi77DqfWnTvp8VsHCCBqoisp1H7uO5tnN3HrzV/mnt7+H797+LebMmXPMrl06yng0+OGlwDhehBCCRCLh5ykOKnWVDd1nXA0YmMUiyUSSSDhEd18fQtcIHLQRUJT/z955x8lV1vv/fc6ZXrb3lt30Hgg9NAG5gCKISLkWLngRRPHCFUWuPwt6EQuilKsgV1CK2PAKigrSpYY0QkJ6sr3P7uzUM6c+vz9mZ5LN7iabTTYJmfm8XnnB7syceT5zdp7v822fr4KqqghZRhlHOmQy8Lj9fP7T9+zxOYo7nRS3kunpeIrfiyzLzG88kZUbn2Nz60rC0V5eXPFbbNvahbtMUUE511x2B5WBWmzDzCZwhSCb7A94azh20bm8vupPLJ73AWqrxh+CNRTtw+3y4fH5EbY94pq7Qth2ujJplw03lgzTN/fLh9EAACAASURBVNjO9JqFSHK6Gz0aH2Dt5pc5etYZVJbsHM+7cMYyVm18nr+/ni6/XjL7dFJagnXbXuPck6/cY3HDRJBQx5aG2V/klPy1qqo55yHsL+dQKMTg4CC4oFftRUbGo4yfizgQSeg1fWt4ueNlKn2VVPuruXftvSSNJL857zfZ0ti9wTTMEeM6Tdtk+fLlfPOGbyKEzd333MmJJ52wX+s8WEilUui6gd/vG7dCJpVKoSgOhBBYljnmPU8kElRVVePbQ3hrqiCGB8EoLmfWu9EiQ9z56LVoehLTMlgw4yTOXXYloYEOOsMtGELntOMvwesJImwbIxJDkobzAIaBq7houNlSw7RVfvrwF9ANjc9c+r0xjcLKd5/lqefuRZEdnHDUh1l2zEfxSd50uaxDya7LNgwkWUEIO93MJoGQJX7x5Nfp7NvKFR/6OjOnH4PD5+PZfz7EP9/6PTdccR+FzkJkxZHNPw0MdfHky/dxwsJzWTjzZLa0ruaxv32Xfz33ZuY17d/fXjQ8SOP8pZNttB3XrcipkNHWrVspL8+tEYP7y9ntdhONRnE73BR50mGbUCxEUksizLSMNgyHNswEuqXvcVjORFDtr+ak6pNYVLaIpsImjio/it9t/h1bhrZw7rRz9+omCyGIhWJ4AjtP3LIkU1Nbw7KzlvHWy2/x6CO/zk6WKyoqQNNVtm7ZwdsrVhKPxamunnQX6AGHoijDoaTxPRFFUejq7MXn84x7ALBtG1mWD4lBkCQp3TCW6eeQJGTFQWwoRPdAMxec/jnOOv4TuB0eigoqmTH7OGY2Ls1WIEmShOR0YqVSwxIVCg5v+v62tXZSWVnF/FnLWLfpFd5e+zeaGhZTGEz/3ZuWwV+e/xnPv/Yw0xsWU1s5ixXvPsOba/6ML1hEQ918LM0AKy2Trvj9OAI+FK8nLWsiYPnav7J684t4XD62d7zLsUvOwxYWv3v6B8yZcTwnHXshstOJrWlpqRRZxucJsnTumVQMew5FBRWs2vAcKS3Owpkn7/UzE5aFbVpjdl93dPdRu9sM5n3AuCGjnPIQMl+IXMKB4JxIJGhvbycQCBCPx3F5XahOFQCn5CQcDpNIJfD7d8bm92d+81j43ebf8cNVP+TGo2/kU3M/Na6W0P9t+z/ufedebj/5dpbVLBv1HMM2GBga4P7v3ceLz79EIjG2ZPXChQv49FWf4JQzl1HsKz4gpbdTjb01SlmWhWVa1DccPnLRqaEwtmnicnvTshlC4CwIjlsWausGRjSK4vPi8KUTyMIWWUmKoWgfD/72q8QSYRrrFmCYOpFYP4ND3Zx2/CWcfdpVKLLC4FA3T/7jHlra1/Glzz5Eob8US093Ne/edzIU7eeuBz9LfeVsPnDspTz45Nc5fsmHKC+t468v/pzrPnU39TVpBQRhWRjxRFo7abgSLRP2wxb8/c1fseK9Z/nKv/0i26eQlgsRWe8HSUrnWZR0h7w93HOQuRaAGotQO3vJAfcQcsogrF69mqVLl07Rig5PHCjOnZ2dxONxqqqqKCgowLANWodacTvcCFvQ0tVCiVSCw+MglArhIp2s3DVBuj8QQnDTqzfxSscrFLuLmVU0izklc1hWvYyjK47GFjY/WPkDntr+FLIkMzMwk8fPf3zMDdKwDXRbp8RRwup3VvPqa68iyRL1DbU01Faxef12Hnn413S0ddI4cxo/uOu7LJm9+IDwmEp0dfZRU7vn6WmJRIJpDdMOaB5hf2AbJkY0PX9adrlw+Lx77SS2NA1JVrKx/00btzF33s5y5WhsgP979ickkxEcDjcul4djF53DwjmnjrjOULSPH//vZ1g09zQu+fDN477fr5/8Dpu3r+C6j/2QsuIannnrEd5892m8ngCVZY1c84k7Rzw/I4ZoGzq2ZmTFFVFkunq28MCTX+Mjp13DcQvOSecxLBtn0I9tWtipVNpL8XnTvTCShLBtrFQqXS5LuuN689Zmlp1+Zt4gZJAfoXlwYZomtm2PkMpWDZXWSCtCCKoD1Rhxg95QL13JLooDxQhbkNJS+Ly+A+KZJY0kf9nxFzYPbWZbeBtbh7ai2zpBZ5BiTzFtsTb+fcG/U+mr5PYVt3PvB+4d00uAYaNg6SiSgktxIUsylm0RUwdJJqMoksLbr63hztvuRgjBd+/4Nuecefao69jCJpwKE3QV4DqAg9mnCqqqosgKFZUVuN2HZnbw7jATSSSHki7jPMhhumdfeYhXlv+OL1zxP6PyDuFID2s3vMQ/Xv0V55z2GU5e9BHMRBLhc3Lfb75EKNzBFR/7NnNnjl02C6OVYy1N556Hr8Pj9nP1Rd/FNoy0R+TcKa8yrmqrZWFpGlZKQ9dVamcuzhuEDCZjEFatWpVTU9Ng6jlHU1ESRoKqQDrmblkW7dF2ZFlGkRSi0SihUAiHw3HANyDVVHmr+y1e6XyFreGtXL3was6oPwPDMjj/yfOpLajlwQ8+uM+bjGnqyLIDWZZpa23jpuv+k+btrXzms//G1f/+GYqL07kUIQSDqUHCWphST+mohLdt28OnQ2n439SGnSbiIQDouoGmpSgpKaGoqBj5fa4AunHDVubNH7+yaE9IaQl+9MBVVJZN4+rLf0hCjfDmqqd4d9MrDAxPoptWt4CrL/shsiRjxBI4/F76hzrYtG05p55wyT6HE196/XGee/1hrjjvGzRNPxq3349lmTS3v8uW5pXMnn4cM6cdnX2+ZZn86dm78LoDfOjMa0EI3lu/iZNPPv2Az1TOKYOQx8FBJBWhL9GH35XOKei6Tnd397DOz+i6d8u20FIalmXh841fSbMvyOQcHjjrAY6p3D+DOBQe4Hu33s7zf38Zj9fDRRddwEUXXYi32AtegdflxhI2DYXTsl9QIQS6oeJyerGFhW1bCGHjch4eVW5CCJLJJE6ni6qqygMW2ns/4s3Vf+Yvz/+U+bNOZmvzSkxTZ2bTUuZMP56ZjUspL6kfcV/314uJxPq56xefRTNUnA43VeVN9A+2k9LS+SxFcfKpi77FnOnHYQubJ/56B+9seBGAs0/9N8446RNoWpKKsml5g5BBPocwMRwKzrql0xxuJuhOJ81M2ySSjDDUO4TH68n2Mli2haqqOB3OrOheaCBEwL//mlNdbV1cufJKZhTO4L6z7tuvawkhiCcG2bZ5K79//E+89MzLGMbOcZxNMxu59c7/YvHMo/AOb/i6oeJxB/F7C4evYROJ9SMQ+12DPh4m6iHsCk3TME2TyopK/IGpUdCcauyeQ9hXWJbJPb/8HIND3Ry14CxOO/4SykunNvEeiYVoaV9He/cmunq3UVpcy7yZJ1JbNYtH/3grvQOtfOqib7FlxwreXP0U/3LqlfQNtPPOhhe4/CP/hWJVseykU/MGIYN8ldHEcKg4N4ebUeS06FtGyC4cDTMUGsLv96PrenojqqzEPywRbds2bW1tOF3O/ZeuEILHNj3GXWvu4op5V3Bq7aksKl2Ec5KbsWUZROP9OBQXA6Ewa9asJT4UIzIY4bePPkF9Yx0/vv97zKyajWWZIASFBRUjwkSmZTAU7cHl9ExJ+Giyp1fLskgkkrjdLjweD263G7/fj+MANNwdDOxaZTRZJNUolm0R9B/6xtWkGuWh3/0X3f07EMLmlOMu5rwPfBbLMnjw97fQ2b2FT3/sOxy7ePSc7QkibxAANmzYwPz586doRYcnDhXnsBqmO9aNU3FSX1iPIik0DzWjxTRCoRB+v5+KiopRIaRoNEpff99+ewlDvUO4Slzc/NrNLO9Zji1sfA4fNx59IxfPGqn4qZoqLtm1VyOk6ylSqSi2SOshCVsQCJTy+itvcNPnb+LE047nnrt/hEtxUhiswDHG8Bc1FSOWDON2eQ94KWtf7yAVlSUAGLaJcx8HEWXKUk3LxOFwUltbOyK/kG6Q0ykYVmI9XLBjeyvTZ0x8Et/7AUk1yq+f/G8qyxv5yFmfz278iWSE+x67Acu0ue0rT6MokzLaeYMA+U7lgwnd0umKdlEdrMY93Fw0qA7SH+9HtmS8Xu+YnsuB8hJ27VSO6TFW9K7gia1PsLxnObccdwuXzLoEgNe7Xucbb36D+kA9d51+1147odO5gRSaFsfnLcxu+o8//Dh3fvdOPnbJhVzxqU8zf95iXF5Xuq5+eCqaEIKIGmHNxuVUVVdQHaxBmcT0uPFgGCZOpwPLtuhOdFPuLc9+9vuKZDJJQUEhpaXDBsYw6OjowLIsGhoaxswFHSqkUhoez+FRMXUw0D/QTiIRZ8mC0/IeQgaTMQjbt29nxowZU7SiwxOHkvPuIQxb2DSHm7NS1aZtYlhGNu6eQTQapb+/P9voFtNjuBU3rn0YfxgbiBEsHSl1blgGN792M//s/CdfPfarhNQQD773II0FjXQnuqnwVnDvGfdSH9z3+LEQgu9/5/s88esnsr8rrypn+qzpzJwzk9rqWlavWs2KN1YQGYrwpW/eyIUXnkeRqxDF4TwgeYXBgQglpYVEtCh9ai9lnrIJS32MxScej1NbW4vb5aazsxNBeq9wOBxUV9cwkb3INM0pDz11dHRTV1c9pe9xuKGlpZUli47NG4QMJmMQwuFwzonbHW6cY1qMjmgHsiTjlNNegCWsEfLaGS/Btm0kRcKWbJDS09ImCk3VcHvdWJaFYRhZ4Tjd0rn51Zt5tetVAD4646N85ZivsCW8hRv/eSMyMnedfhcLyxbu6fJjQgjB+k3radvRRkdLB+3N7WzbvI3mHc2YhklZeRknnnIizduaaW1p5VdP/ooZdU0ohgmKQkxPj8Qs9U5OQ19VU7jcTjrinbhkJ7plUF9QN+nQlGGYWFZ6ip6maVlPMxaLUVVZRSC457BeJBIhFApRW1s7pnDfgUI0GqOgIIiu6xiGid9/YKb8Hc4YCIWY3jQvbxAymIxB6OzspLa2dopWdHjicOMshCCqpbWR3IobzdJoHWrNViRloGkaqVSKgegATtvJoDpIwJmeB+1yufaaKE9GkvgKfcQTcSSkESEq3dK59517mVcyjw81fSj7mpZoC1986YsMpAa49cRb+Zdp/7LP/CzbQpZGDl8xdINQKETVsD5SW0sbl3/kcpadtoxb77oVU02QMlM4HS5sYVPmKSPgHnuz7e8P4fG4xxz0FI3EER57ePSpl6ShUuWv3K9ZFqlUCsuyRsiSWJaFruvU19VjWiaJRAJN0yguLs5u/OFwmFBoAI/Hg2WZ6bnPU+Qp9PWGqKgso6urC8uyqa8/8mend3Z2Mn/uUQfcIORUyU2uVRjB4cdZkiQKPYV4HGnFVI/Dg9fpRbf0Ec9zu90UFhZSUVHBUXOPYk7THErKS/D7/RiGQSKRIBwNk1DH1iKSZCntGbg9FBcXk0rtnEHrUlzcdMxNI4wBQGNBIw+f8zDzSubxX6//Fz9b+zNsMXrAkBACaxeZ5l2hyMqoL6nT5aS6pjr7+4bGBq754jW89NxLvPXCW7jcfjyKC4/Dg0fx0BXr4e2VKwmHh0Zc55ln/sE5Z3+YD533UZYvXzF6XQgi2hBeJb0pO+SdXsdk4fF4RhgDINsn0tbeRkdHB9FoNJtj6OzsJBQaYGBggGAwsHO+dE8vtp0+fAoBtnXghlXJikwqlUJVVQxDxzLHvjd7g23Z6LqOqqokk+oBW99UYKq+1znlIYRCIcrKyqZoRYcn3g+cE3qCjmjHKC/Bsi0My2B6yfRRzW6maTIQHyASiSCbMg4l3Qmd2XRTiRSmMKmtrcXpdNLa2orP55vQiUq3dL6/4vs8teMpFpQs4JTaUzih6gRkSeb5tud5of0FhBA88MEHqA1MzvsyTZMrPn4Fof4QV117FaoaxdBN1r3zHmtWvEMqpeHzebnsXy/h4k9cxK8f/Q2/+dXvWbhoAfF4gtaWVq77/DV84frPZTfo0NAAcSlOLBRFVVM0NDaQslPUB+qmZI7FWCXNum5gWeaoca3JZDL7+SeTSSzLoqqy6oD0PgwNRUkkYtiWjWEak5L4ti2b9o52LMvKcmtsbDxsS2/7+/qYOWNBPmSUwWQMQktLywEdjPJ+wPuBsxCC5qFmHLJjxFzkhJ6g1FdKibcE3dJpGWoh4EqHUlRDxev0IiERjocxVINIJILX50WRFYb6hwiWBLPhssHBwfTjE6y4yiinPrn9STYObtyZUJUdnFB1AutC6wi6gjz4wQcp942WF98+tJ3HNz/OdYuvo8w7tkHe+N5Grv3UtSMUV2vrazl+2bHMWzKHt19dwYvPvpId5P7hj53LtTddg7AsHvzxI/z5yb9QWVnB/AXzmD1nFn0D/axesYbW5jYAGqdP4/RzTuXSj32cGdOmT4j3gYZhalkJ61QqhSzJOJzpuQ2qqu41v5ApgzVNC4/bPaYoX1trB7qRIhgMkkql8Pv8lJXv2yEoEU/Q09tDIJD++4rH41RX1xwSqfCJoK21nUULl+YNQgaTMQjxeDx7w3MF7xfO0VSUnnjPiNh5TIvRWNSYLZ3cMbgDp5JOREe1KI1FjciSTMtQC36nn0QiQW9vLw6ng0Q8QVNTU9YAmKZJS2sLfp9/n79EES3Cyt6V6LbOKTWnEHQFWRdax+df/DxV/ioeOOuBEdU8MT3Gp5/5NO3xdqYVTOP+M++nwjd2B7GaVIfDWQI1FaawKF3mKQSYtkFPRy9/eOyPzFs0j/MuSOc0UmYKt+Jh9StreO65F9i8aQvNzS0oisySpYs58ZTjcbqcvPDMS7y7Zj0er4fHfv0rFi48uP0olm1iGClczrHFDQ3DxDB0amtrR5Wx2ragt7eHZDKJJEmYpkVJSQllZaXYwk6rfg7fx+bmFhyKgmt4gpymaUyb1jihKqgMOjs7sW2By5Wu9kqlUgQCwWzZ7eGGoaEw0+pn5w1CBpMxCJs2bWLu3LlTtKLDE+8XzrawaQm3IBB4nd4R4aIMBpIDhFNhFElBkRUaCtODR0KJEOFUGL/LTyQeobO7EzNusvjokZLVAwPpEJPPd2CqUFb2ruQ/Xv4PpgWn8YNTfkBDQQNCCG5+7WZe6XiFG46+gfvfvZ8STwn3n3U/1f49l0bGEgMIYY/Zm7B7CW9ST1LmTSef41qczqFOpKSDosqRYbeOtk5uvPYrmLrJY7/9JU3DQ1VsYWMJG4c0OudxoKAbKRTZgW1bOJ1j9wnomo4tbKqqqrLih7Yt6OvrJZlMZvMXGe2laQ3TMCwNJPC4fKiqysYNW6itq8peM5FIUFdXN+FeCcMwaGtrG3FwMk0T27L3e3ZEIp72/g60LMj2bTtYevQJeYOQwWQMwsGoiT7c8H7irFs6/Yn+bCK00l9JsXfnyVs1VNoi6XBIfWE9Pmd6Y7dsi+Zwc7Z8tcxTRmu4lQJ/wYiSS9M0h0+C9ojQkS1sVEvF79j3L+0bXW/wtde/hmEb3HD0DeiWzk/W/IQbj76RT8/7NOtC67j+pevxO/1cPPNijqs8jvml80eExjIwjBTx5CDSqFoPgSTJ7PyuCoQAQ7Kp8lXSq/bhkd3pU/MYEg7N21u4/qobKSwu5J6HfkxBYRBTpIfRFHmKJt2rsDfoRgqfp4CEGsHtGj/0klFfLSoqpqiokMHBQeLx+KhkdjKZpLi4GK/XlZ6HICl0dXUBEm63a8TzysrKxqzEGguRSITBwcFRB4V4PE5TYxOyMrkErhDQ3t6GoigHvNIvmYxRXTk9bxAymIxBWLt2LUuWLJmiFR2eeD9yjmtx+pP91BbUjmhGE0KwbXAbiqzQVNQ0cgC6FkO3dEq8JUiSxPJVyyltKs0ajQxM06S3r5eUmsomOVUzXVGSmY2wr+hL9vGd5d/hze43AfhA3Qf40ak/yq5v4+BGblt+G5vCmwAIOoN844RvcFbDWaOuZVlmOl8hbIQAWU6Pz8wYBCEEtrDQ9SQxNYwlbDxOH26Hm8RACn+pJ336N3SQJBTFgSwrrF29ji9dezP10+o45QPLaJwxjWlN0yirK2VaScN+laaOh6QWw3Q6cJsM6zeNv3kJIUilUth2uvpod2MA6URvKqVSXV1JQbCMjs4dxOMqsWiSquqdeRxdN5BlaUKbcGbTdjgco1R2M415k+2hSCSS9PR0A9A4rfGADiXauGELJ55wct4gZJCXvz6yMZ5QW2+8F6/DS4Fnz3o6mqnRMtQyqnIJ0htLOBwmPBQm4A+QMBIUe4rTYad9aH7bfb1/3PZH3up+i2+d+C2CrtHvG06FWdm3ksc2PsZ7A+9xy3G38PFZH5/U+6V5WMTVMLZp4HAMT9cSNqah4fWlT/1aKoZtWyBLvPHKch649yE627qyG6/iUKifVsdRi5fw2c9+hhkzd4boTNvEtNOqrors2Kfwkmka9KX6wemiWPbjlBUcEzC2exNjjESHqK6qpaykirXr3qawsAh5jAqqRCJB47TGvZ7uU6kUnZ2dY+bZkskkJSUlWSXefYEQ0NHRgSRJ6LpGRXnFXhv5MrBtsdcZFXn5692QH5AzMRxpnCeq6Llq1SpKppegyMqY4RmA7u7u9KnUYdNY0EhXogs48POgd4dqqtzy2i281vUa1y66lqsXXj3pbuL0yTpOSouhR8BVJOFyeXG5/ThlZ7pnwtKxLAvbNrFti2QyQXtbB63NbbRsa2Pr1m2sX7MBLaVx+Scv4YprPonsljGFCUikC6zSYSu3w4VLcaEIBYcs4x/D4AIMxvtJCJ2ArwhhmQQk14hZELplAAJZSg9SmujGpqpx3K4gfl8BsXgYG51QX3SU7Hc8Hqempga324NpGkiSNObMh1BogEQiPqYXoOsGiiJTU1MzobWNXKdKZ2cnwWBwn66j6zqDg4NUVVXt8XnvrdvIsmV5LaMs8h5CHnvDqEE9lo5lWztnFug6W1u2UhIsoTpQTcJI0BPvyT5/f7En42XYBrctv42nm58m4AywqGwRi0oXUeGrwO/0E3AGWFqxdEKhHCEEz+14hqdbnqYz2U1HohOfw8eV86/k0tmX4nWMjN/bto1l6aS0BLZt4nC46Orr4eGfPcozT/0Df8BHZVUFLrcLt8dNRWUFNXXV1NTVMGfhLKrrqjBMDUmSqCuoxzVcBSaEYOXKVUiKTLDOS0VJ+rGoFqEAd9Z4GLZJV7xrOCcikJAp85ZO6HPX9SSy5EeRHQQCfsKRbpzO9GeU0NMJ3IA7QGp4NnHGEwKor68fYRQMw6C9vX3c/pRMIrupafo+VSzBcNWSZeMazm3E4/EJhY0GBwcJhQaYPr1pj7m/vIewG/IewsSQi5whzXvxUYvZEd5B0B1EM7WsrIRD2dnv0NbThsfwUFFUgRCCLf1bwAaXw7XfKrEJY7jCZJwwlBCC59qeY2XvSt4Nvcu2oW3ZfgeAJeVL+OkZPx2xoXcnuvEonhGJ4OfanuNrr3+NMlcZC8oXMK1gGlvDW3m9+3XKveVct/g6Lph+wajNw7ZtYvF+ZMWBLMkIAZs3bOaJx/8PNZmWuk4mVXq7eunvC2WT2iVlJSxZupCFRy9k3vxZnHLMqbzy8qvcf///8t76DQCUlpdy1jlnceknL6VyWiUOw6LAXYAiO+hN9KFZqayxs22blJWi2FNCoXv8UKAQAsPUKC6oQhqWB0kkh2hpbcNZIqOZGgJBTaAGtyOtY5XJC+haOqeSkfS2TIuuri4EYo+jXSdasRQKDZBMJvD5/LhcTvr6+kYktROJBJWVVXvUWRICWlqas017ewox5T2E3ZD3EPKYCNoibaiGikN2UF9YT8pIjeh3iKgRpCEJn8dHKpXCkA0sr4WlWsRiMTwez4iY9kQlA4QQJM0kDsmxx7DVrlBNlaienlG9tn8tt6+4nZOqTuLO0+7EITv40/Y/ccfKO3AqTq5fcj0Xz7yYN7vf5Ev//BILyxaOMh5r+tZw7zv3sja0lstmX8ZNS28a1bGs6yrJ5BBOV3pzNgwNt8uP1xvM8ojHB9B0nZ7OXta9s57VK97h3dXrCPUPZD8T27apb6jj8isvAUXw1ssrePuNlfgDfh7+48P4C92UOguxJWgdaEO2ZIqKd8bmhQDVTBJ0BSnxlIy50Vm2iYxMQXBnAjmiDrGtbz1eVwC3w405nJSvDlRnw3CWbWHaJpZuEQwEKSktpaenG13X92r0E4nEXqfJqapKR0cnfr8PwzCwLAun05XtaYC0QXI4HVRXj196rKoqXV1duFxuHA5lj8/Newi7YTIGYd26dSxatGiKVnR4Ihc5w07ecT1Of6KfuoI6nIoz2+/gUBw7T7xKCe3t7ZSXl1NSWkJbpA3TNpEtmVAohGma2U3Ptm3EcOzb6/WO+4XULR2n4iToDNKb7J1UsvpP2/7EbW/fxtkNZ6NICs+0PsMJVScAsLxnOXOK59ASbaGpoIn7z7ofc9CkuGpkCaktbO5acxe/3vRrzqg7g9uW3YYsyWwd2krCSHBc5XEkkoPpxDMgSQoBf/GIiW6GoZFIDI4wGn5fcXpy3MpVrFmzimMWH8Oxpy8FBRQBXk8hXe29XHHJFUyfOZ07H7yDQoeHLS07+Np/fJPBgTDf+t7XOOm0E7PvkzYKKkFXgDeee4vHHn2c//7urcyePSv7vl5PEK9n52jW9lg7se4h3IUSjuHQVdJUKRg2LKqhElJDCAS1gVqSiSRerxdN08asZNodmqbh9XgprxjdjQ47JS8cDsceQzyZ8FPjtEZsYROJRAFGNL719/WTVJN4PJ69JsU3b9rK8cctyxuEDCZjEDRN26N7eCQiFznDTt5CiOwGnkE0FaU73o0syVT4KyhwF5BKpbKnRcu2CCVDhNUwPpdvxOneNE0Mw0BVVQYHB7NSGbsjYSao8FQQcAXojHdiC3tSJa0Pb3iYe965B1mS+dyiz3HVgquQkPhH6z+4c/WdFLmL+PlZP6fYU4xlWuPGqH+z+TfcuepOij3FRLQIlkgbgJ+f9XOOLltCLB4CJIKBslFTuIQQxOL9WSMhbJtgw9xcywAAIABJREFUsDy7GYUinZimgdvlxaW4MA0Nf6AUh+LkxWdf5Ctf/AoXXHIBZ553Ct+66TaELSivLGfH1mau+Y9/5xNXXjZiY3v9tTf5+o23YpkWBQVBfnbfPRx73DHoeoqCYFlWCiOkhojpMRy2TEId3KXSCpJGAp/Lj6oncTvcaLZOsbuYoDMwQsp7b7DttOBdTU3NcGI+3c2c2fxDoQFiseiEmh0zvRWJRAJFUTBNk+qqavwBP7Zl09rWmj1k7C6dYdkW3ZEeyoKleBwe4vEotdUzDrhBUG699dbJXPCQ44EHHrj1mmuu2afXNDc3U1o6Oa359ytykTPs5C1J0qgvjUtxEdWiGLZBRaACh+wYkWyUJZmAK4DX6SWqRdEsDd3UMWwj3dAlCySHhNfjJRqJIsvyaJE3U6fEU4IsybgdbsJaGIfkSMt6a+l6+4k0DB5VfhS1/lo+Ne9TfKjpQ9jCJplMMq98HpfPuZyPzfxYVt8pPhjH7Rvb+C8qW8Sc4jlEtAin153O5XMuZ23/WrYNbePCmR9FkK5OGqujOP0Zyuh6EoTA4wmOGA/qcnqxDB2ZdFgHJLyeIJIk0TSzCUM3+N0jv+P5v75EcUkRP/rZ7Vx2xSV0d/bwh1//H9u37KC+sY7SslKat7Vwy/Vfp6qmkp/+79289drbPPbob5gxYzqN0+vxeQqRJJmUmaI32YvP4SMxlMThdrB21RpKy8tQFAUFmZSp4nP503kjyUHcSBB0BlBkacxS1bEgSRKqqhKNRonFYkRiESJDkWzSemBgIDsTfG+QZRnDMPB6vTidaaMSjUbTGkyaRjwez+YqbNsGIfD7/Vi2xbbebbR2tqCaKpZsMdgboaa6brIG4dvj8s0lDyEajR5282CnGrnIGfbOO6bFGFAHaCxq3ON1hBBpQzA83U2RFZyyM70hxXtRbIWenh6EELg9bhRZwbRMYskY9YF6HA4HpmnSp/aRslKUF5Xj8/kIhULIsjxmKeSekEgk8Hg8Y8a/9ZSOyzNxL+S3m3/LHavu4Odn/ZxjK4/d6+cQi/VjC5uCYPmoDdW2bYSwsG0LIQSuXTqTLcvilhtuYSg8xA/u+QGBoBdVjSIQ/OGxJ3n4gUdR1RSLly6it6sX0zS575F7KSgPIqdkvvLFr7H2nXeZv2AeH7/4Ej7ykY9gBYffR3GxbvU67vz+nax7Zx1nf+hMbvn2TUhICGwcDjcb12+irqEWh99JUPESdKYN2nhGQQiBbum4FNeIDVe3dLrjPbgUJwElgGRJuN0enMNifZPZnFOpFE5nWiLcNEycLieGbaBIClpKo6Ghgfahdto72wl4Aui6TnF5CYamc9L80ycrg50PGQG0tbXR0NAwRSs6PJGLnGHvvIUQmLaJc5KjKzPjQF2KC2Gnx01GIhEM00AzNRoqG5hRPSPrBehmWq3V60yHmFKpFB2dHfi8Ywu/jQXDMBBCUF1dTVtb26i51ImhBP6ifZgqZ2lc8OcLqPfXc9fJd+FQHHvsytV0FcvU8PmKJvwe48G2TVQ1hmGkSKoaf3vyGf702z8TiUT58f23s2DxomxOQeg2zz35Av/4+wusX/8eABVVFcydPxevz8tzf3uOwqJClh53FC88+xLXf/k6/u3qq9C0BI/98jHuv+tBGqdP4+5f3InDBzNL55HS4sgOJ+awkXfIDiQkUmaKsBZGt3S8Dh9l3lIUWUG3DHoS3TgkBwKBYRkUegopdBciSzK6riJgjxId4yEejyOEIBgMEtEiDKqDSLKMpZpUVFUwEB5EETJutxvbsklpGoqkcObx5+QNQgaTMQh9fX1UVIytOnmkIhc5w8HhHVbD9Cf6sxVLQgg0TUO1VaaXTB81KzqmxeiMdVIwXFoZjUbp6+ubsBptRkrB6/UyNDSUDVdkkIqn8AQmLrOQSqX4/dbfc/e6u3n0I48yyzMLy7Jwu928N/Aez7Y+y6q+VXyw4YN8cu4ncSvuSZ+Ex4Ouq6ipKA7FjdPpQdMMFMXCMFI4hsNXKUNFM1T8vhK6O3t58+U32bZxG1s2baGrs4vzLzyf6798PT6/j1tu+CovP/8K//Pg/7Dirbf55c9/xTEnLGXdmnU0zWzk+/d/j8aqGbR1buWvTz3D4MAQhUUFBAuD1NXVMH3udIoKCnHIDjRLQxISpd5SQmooPfZ1+AAhBKhGkmJvOi/R1t5GTVUdyBYu575JXaSbBy10W0+HwZx+JAkSyQQOpwPLsEbkKAzdIBKNcOEZl+QNQgZ5gzAx5CJnODi8Ldti++B2/K6dMWQhBEkjyYySGWN2H3fHukkaSbxOL0II+vr6iMViKIqSrVQZa8PNJL0rKyuBdIimvb09HS9XlLQhiqkEi4MTyk2kUikURaG4vJjTHz6d6mA1H575Yd5ufZv1A+vpSnbhlJ1ML5zO5vBmavw13Hj0jZxZf+YBV0fd3cgIYROPDyKwkSQZ09LxuIO4XL5RCXwhBFpCyxrCZCLJlZddScuOFizT4oKLz+eLN1/Dijff4Rs3fZslSxdz9AlL+cOjfyAaieJyOdF1I3s9SZJonD6N45Ydy2Wf+jjFZUVoloZTcY4qChBCkNAT3Hf7A/zlqb/hcrmYMWM68xfM4eyzz+bkU04iaSdJmSlKvKU491B6rFs6XYlu3LIry9G2bZKJtOLr7qKFfX19fPTMS0fpL00QeYMAuRk+yUXOcPB4hxIhhrShrIieZmo4ZSd1hWPP9TVtk5ZwCy6HC4fsGBZsS434Zws7HcZwOIYrW2xsYTOtYdqInEMymaSjowNZkSkqLEKLa2i2htPp3GNuIlMpVVdXh9Pp5Ffv/Ir/9+L/A6DKX8XMgpksq1zGaTWnEXQGWTu4lrvfvZvtke0sLF3ItYuu5aTqk7KbuKZpWJa1xzLcfYVtm8TiA0iAz1c8Iom9O3YPlbW3tfOFK7/AOeefw+duuIZYPIQkSbzy/Jt857++ixCC0848jcuvuIh5i+ejazpDgxHaWtrYuH4z7727kZXLVyFLMh++6Dwu+eTHqGuoHbOx7/vf/hHPPPUPrrji0wQCQdatW8c776whFotTWFjIKR9cxlXXfZpAUZAydylepxfDNlBNFd02cMgKCgpxPZ6W15hgCLOvp5+PfvCSvEHIIJ9UnhhykTMcPN6GZbAjvAOPw4MlLFJGirqCuj2K7yWNJB2RDmRZHqXGmgk7JZNJNE3D4UhXQLnd7jFLJROJBG63G0VR6BvoI+AP0N/bjyRJY5YbW7aFqqrU1dahS+nEqUtxsbp7NY1FjZT5yjAMA13Xsxtgd3c3TreTp1ue5sH1D9KT7GFx2WI+3vRxTqo8iZKCEpxOJ+FwGK/Xm92kuhPduBU3JZ7JDZmxLBNJGl3BtTvGSqbv6nUkEkPE4mHKy+pZs/IdCgoKmD1vNpqWRE1Fx6ys6urs5tcP/Za/P/UspmlSUBhk1tyZzJjdRNOM9L+/PfUMf37iaS7/zMf572/ejm+410RVE/ztmb/w1NNP888XXqWmvoY7fvo9CsqCIElIQkKS0uq6AoEtbGRk2ra3s3L5ao4+dglz5s/eI+fe/l4uOuOyvEHIYDIGYevWrcyaNWuKVnR4Ihc5w8HlHUqEUM10N7RLcVHkKdrrDGPTNulP9BPRIrgV96iKln2FaqgMdAxQUF2AV/HS3d2NYaRDIQKBtMseUFlZieJOJ1JTZmpMRdhdEQ6HGRwcxO/3o1s6f97xZx5a/xC9ai+F7kIunHsh58w4h/lF84kPxulVe3l066M83fw0QVeQH57yQ46pHFs+JZVKYZrpGcyTndsRDUUpKBvfAEdjESzLIOAvHOE57S7dMRZ6u3t5459vsXnDFrZt3k7z9tbs5wrwiasu4+rrr8LnL6IusLMMNK5F2dr7HlvX7eD//eetBAsC/Pj+HxIsCPLPF17ljVeXY5kWPp8Hp8vFu6vX0dPdC4DL7eKb3/sap515CpAecvTj2+8hGoly8ukncdqZpxAMFPDRf/l43iBkkG9MmxhykTO8f3irhsqgOkjCSOw81Q5/JSVJmrDQXkyLUeWpYsgcwhY2boc7m6yM6TFciis9g3p4WlrKTDGtaBqtQ634nGOLu2VgWRZtbW243On4tqalJ5btsHbwxIYneGbbM6Ss9Lzk2SWz2RbehoTERTMu4u3et2mPtfPlY77MRTMuYmXfSl7peIXOWCeFzkIqA5XUF9dTRhkLqxdS5N73CqZMQ56u69nQVQamaWJZFuXl5XR1dY3qGdB0FVUdygrkZZD+7AyELVCcTrzuAIriwrIstm3ZyPZt23G53Bx74hIC/lJSto7b4abIXYTX6aUr1oUQNroaZcum7Xz1i99A13W0VArLsqmtryEYDJBUU6RUlZlzZnLK6Scxd8Es7vjvu9m8YQs3fPULKIrCT++8H8XhYFpTAxvWbUQIwZwFs3lv7Ya8QcggL10xMeQiZ3j/8baFjWZqmLaJIivIkkxPrAck9trhnJlbkOhIMGveLFqHWgm6g2kdIj1OsaeYpJFEt3T8Lj8xPUZtsJagO0hPrIeEkRhVEbU7otEoff19+Lw+kmqShvqGbBNVQk+wqnsVyzuWs7J7JdOLpvPJmZ/EY3qwHTbffOubvNr1Kh7FQ8pK4VE81AfqiZnpXhDD3nnirgvUceX8Kzl/+vkTliEP94RpppnbVtzGZbMu48LGC7PlsxkZbJ/PR39/P/F4fITByGg1CUlkR5fatoVlGbhdflwuL8pucX3LMonFQzgcTizTxB8oIaVqmMLE4XFkjbrf6ceyTOLxEJ2dvfzvvQ9RP62OM//lA8ycM7rL2DR1nE4Ptinx1Rtu5q3XVgBw7AlLueU7X6GispyB0CCvv/w6kYEY9937QN4gZJAXt8vjSEfSSNIWacuWqWZgWMaI5GNci1MZqKTQkxaL6451kzASmJZJqa+UMl8ZAkF/op9QMkSJt4TqYHX2Pdoj7XsNG2WqmnRdp6qqakLjKSORCP2hfmRZ5onmJ2iPt3N82fGcVHMSTfVNKIqCEIKeeA8b+jawfNtyXup+iQ2DG6j113LpjEuZWzyXWWWzKHTvFMLLdAl7POkpbC+0vcDX3/g6kiShWRpXzLmC6xZdB4CiKNTU1KQbv0yT9vb2rKeTgWnqxOMDOF2erDEI+Ev3mMhOaXGSyQhOhxdJdlFTXUMikSAajY6S09Z1lYQ6hHNYWmPMz1fY2JZJMFCOLMuEh/p45H8fo7yijI9c/OFsHkUIgWFoqKbFh044P192mkFe/npiyEXOcGTwFkLQOtQ6wkuIa+lqFFmSs6WrCT3B9JLprF2zlmOOOQbDMmgON1PkLaLcVz5iE0roCTwOz87SRmGzbXAbfufe5ReSySTxeJzy8vIJ5zs0TSMSiRCNRRG2wOf3UVVZNebJNhaL0dPTw/L+5Ty06SE2D23OPlbsLuaM+jM4q+osTmw4EVmWaetv48XuF7nn3XtYXL6Yhz/2MD98/Yc8vv5xzq0/l68s+QqNDY0jmu3i8Tg9PT2jej+SySF0Iz1KNeAvyQrljQchbKKxEMJWqK1tIBgMYts2XV1d2ZzIrlDVGJqewOl0Y9s7O7qR0tPoLNPA5yvKdnjvLiiYgWlouFw+BntifPiDH84bhAzyHkIeuYBdT/CqoeJUnFQHqumKdWXHW/qdfqqCIydsaaY24UR1T6yHpJmckrnKGei6TiqVIhAIjLuJCSHo7u7G7XZTWFhIS7SFHYM72NizkZWdK3m953U0S6PcV07CSJA0kgCcXnc6v/joL/C5fAgh+PGbP+bHb/2Yk6pP4sGLHsx6Tpn36OruwtCNEZu2bZvE44P4fIV7NQYZRGMRyssqKC7eqTBrmiYdHR0oijIigS2ETTwRxrR0FEnB7Q6gKA4sS0fVVWzTpqiwfIQnkBEUzEhsWJaJBAQCpYSiA5y99Oy8Qcgg7yFMDLnIGY4c3kII2qPtpMwUDslBQ1EDDtmBZVt0RDtIGAmaiprwOr2T5jxe2EgIgWZpyJI8KaXWAwlN04gbcZ5vfZ7XWl+jxFdCfUE9TUVN1CXqmLNozojn/3b9b/nq81+lqaiJhz/6MNOKpmUf03WdtvY2/L6RXpEQ9gjZ7z3BNE2EENTVjRaYy8iSeD3eEZ5QxjNQFGf2NZaVLgMuKCggGoumxRDdbgzDIJmKk1IjeL3pEFTGGFiWYLB7kPPOOi8vf51B3kPII1eQNJJ0RbuoL6zHvcvp1bIthlJDlHjHHigzUdjCZvvg9mxPhG7p6JaOhETAHSCmxfaaYzgc8Xrb61zzl2uQZZkvL/syZzWdRV1BHYZl8MKmF/jHtn9Q5Cvi3Gnn0lCws4lRCIEQYo+n73giTk11zbiy18lkkq6uLny+nVpVmSqoTA7DNE00TaOmpgav14tpmkSj0Wzi2+v1EokOEI3GCAaCSJKCJMmoqkpxRTELqhfkDUIGkzEIa9euZcmSJVO0osMTucgZjjzelm3ttbdhfzj3xHoY0oZQJAW/y0+BuwCvIy3E1xXrQjXUvVYiHSq0bG6hcU7jmI9tD2/n2r9cy8bQRgCmF09nIDlARIvgkl0YtoFAMLdoLidUnEB9oJ6GYANNBU24ZTeynD6x72ocMqWsDQ0Ne9yQX9j8Ane8eQduxU2xu5jqYDUfaPgAs/2zsaz0ONfq6uo9zmYQQtDb20symcTn8xGPx6msrKSnvYfTjs+P0MxiMgbBNM1JN7+8X5GLnCE3ee8PZ8MysISFWxldCZMyU9lS1j0hE2IybRNZkpElGUVSJq0oO1HsaTBQZl3bw9t5qeUlXm19lRJvCefMOIfjK49nS+sWXu55mRe7X2RjaGN2pnWBu4ArF1/JZTMvQ6QEuq2zKbaJlkgLhmFQWlRKXUkdZzaeOaa+0iPvPsI3X/omFb4KSt2lDOqD9CX6MGyD6kA1F8y+gM8e/VmqC0eOyXxhxwsMqAPUBGuoCdbQUNiAJCS6urpQVZXCwkIqKyuJqlHmlM/JG4QMJmMQNm3axNy5c6doRYcncpEz5CbvqeTcEenAsI0RIStIb34pM4VlW0iSRMAVwO9KD3XRLR3VUEmZKTxOz5TlITqbO6ltqp3UazVNw+VKJ981U6Mt0saO8A7+sOEP/H3b3wm6giyqWMSq7lVoljbq9TOKZ/CfJ/4nF8y5AN3S2R7ezsPvPMzj6x/nzKYz+Z/z/ieb1E4aSZ7b/hx/2vQnXmp5iQp/Bb+88JcsrFiIYRl88+Vv8sjaR0Zc3+f0cXTV0RxTdQynVpzKCTNPQFEUmrc1s+yo/AjNLCZjEOLx+ISlho8U5CJnyE3eU8lZNVRah1qzGk2WbaEaKgJBobuQQk8hbod7TAmIpJGkP9FPykwhSRIO2ZH9tyss20pLTiONe62xkEqm8PgOfIXUe/3vcfdbd7MjvINl9ctYWrqUKqqorKrE5XWxpnsNP3nzJ2wa2ESRp4ioFsUWNgDXH389Ny+7edww3/q+9Vz11FUMqoPcdsZt/HHjH3mz402uO/Y6/nXRv9Id66Yz1sm63nWs6FrBhv4NuGQX3/vg97h0waUMRgZZUJvPIWQxGYPQ0tJCY2Pj1CzoMEUucobc5D3VnFuHWrNd0Q7ZQZmvDL/LP2pjHwsZT0K3dFJmCtVQ0S09vTWJtN6SU3ZS5C3CsAxieiztdQzvXZIkZcM5kBaGy+Q0+jr7qKg9OBLvu3oUkE7I/3XrX3l+x/M0FDQwu2w2C8sX0lTctNdr9Sf6ufovV7OyayVuxc0dZ9/BxfMvHvO5fYk+vvC3L/BG+xtcseQKPjvjs5y88OS8QchgMgYhFApRVlY2RSs6PJGLnCE3eU8158xJv9RXOqFGtr3BFja6pWNYRlZ1dde5EpnRpbawMS0zO9lMlmQ6oh04FScO2UEsHCNY/P6rgoJ0v8gvVv+CUxpOYUnVngsCTNvk+699n/tW3sfRFUfz9jVv41AmlTMa98blVNZtV5XCXEEucobc5D3VnH1O34h6/v2FLMl4HJ4xG+IkScoaibFQHazOynqYpjnisYwulCRJU9psdyDgdrj5wvFfmNBzHbKDr5/2dY6qOormnua9Vp1NBjllEGzbPtRLOOjIRc6Qm7xzibPP6aPYU0xMi2V5q4aarXAq9BSS0BNopjYqEf5+x/mzz6fL3zUl184pgzBeE8mRjFzkDLnJO9c4l/nK0gbBYRPTYhS4CyjyFOFxpEXvdI9OS7gFh+zYp9O0LWziWpyAOzDhxPbBhtMzNaW8hyfbKcLg4OChXsJBRy5yhtzknWucFVmhpqAGSZWYXjyd6mB1duYDpAUBa4I12VkTE4EtbOJ6nCJPEXE9fkDWadomqqES1+PE9Xg6mb6fUGPqAVjZaOSUh1BTU3Ool3DQkYucITd55yJnn9PHrMZZ4za/BdwByswy+pP9uBRXtpxVM7XsxpzJZUiSRFyPUx2optBTiCRJRLXouEOKLDvdbTxecl0IQUyL4XF40p3fTi+2bTOkDRHTYjhkx6S7v4OlU5NEzykPobm5+VAv4aAjFzlDbvLORc6wd96lvlIaChsIuAJopkZMi+GUndQW1NJU3JSeI20ZxLWdxgCg3F+OU3aimSMb0jJehGZp43oRGWNQ6itlWtE0yv3lBFwBCjwFNBQ20FTclG2GmwzC3eFJvW5vyKmyU9u2JysX+75FLnKG3OSdi5xh33mPpQuVKXPdvapJt3Rah1rTgne7zKYu95dT4C6gN95L0kjic43M38S0GMXe4lHzKHaFZmq0RlqzmlH7gpgWY3bp7APeh5BTfz3vvPPOoV7CQUcucobc5J2LnGHfeY+1+WbKXHeHS3ExrWgadYV11BbUUh2sZnrJdIq9xSiyQmWgEofsIGWmgLTuUzQVpcgzejjR7nA73FT5qyaVq+je2r3Pr5kIcspDyCOPPPI40NAtnbahNixhEXAFKPGWZHMSE0FvvJewGsbr9GZzIRmhQMu2Rky4yyCux5lVMivvIewPVq1adaiXcNCRi5whN3nnImc49Lxdiov6wnqaipuoLagdUek0EZT7y6kOViNJEjEtRkyLkTSSeB1eSn2laGY6V2FYBoZlkDJTdG7unBIueQ8hjzzyyOMwgWVbmLY5QsYj0xcxqA6iyEq2WqrIUzTZt8l7CACrV68+1Es46MhFzpCbvHORMxxZvBVZwe0YOZNClmQKPAU0FjdSX1hPZaCSHRt2TMn755SHkItVGLnIGXKTdy5yhtzkvZ+c8x4CpAeI5BpykTPkJu9c5Ay5yXuqOOeUQWhq2rtG+ZGGXOQMuck7FzlDbvKeKs45ZRC6uqZGIfBwRi5yhtzknYucITd5TxXnnDIIJSUlh3oJBx25yBlyk3cucobc5D1VnHPKICSTyUO9hIOOXOQMuck7FzlDbvKeKs45ZRByrRIBcpMz5CbvXOQMucl7qjjn1CfpdE7NUInDGbnIGXKTdy5yhtzkPVWc37d9CJIk9QOt+/iyMiA0Bcs5nJGLnCE3eeciZ8hN3vvDOSSEOHesB963BmEykCRppRDi2EO9joOJXOQMuck7FzlDbvKeKs45FTLKI4888shjfOQNQh555JFHHkDuGYQHDvUCDgFykTPkJu9c5Ay5yXtKOOdUDiGPPPLII4/xkWseQh555JFHHuMgbxDyyCOPPPIAcsQgSJJ0riRJmyVJ2iZJ0i2Hej1TBUmS6iVJekmSpA2SJL0nSdINw78vkSTpOUmStg7/t/hQr/VAQ5IkRZKkNZIkPT38c5MkScuH7/nvJEkaPUH9fQxJkookSXpCkqRNkiRtlCTppBy5z/85/Le9XpKk30iS5DkS77UkSQ9JktQnSdL6XX435v2V0rhnmP+7kiQtnez7HvEGQZIkBfgpcB4wH/hXSZLmH9pVTRlM4CYhxHzgROALw1xvAV4QQswCXhj++UjDDcDGXX7+AfATIcRMIAz8+yFZ1dThbuAZIcRcYAlp7kf0fZYkqRb4D+BYIcRCQAEu58i8178Cdm8eG+/+ngfMGv53DXDfZN/0iDcIwPHANiHEDiGEDvwWuPAQr2lKIIToFkKsHv7/GOlNopY034eHn/Yw8NFDs8KpgSRJdcCHgV8M/ywBZwJPDD/liOIsSVIhcBrwIIAQQhdCDHGE3+dhOACvJEkOwAd0cwTeayHEP4HB3X493v29EHhEpPEWUCRJUvVk3jcXDEIt0L7Lzx3DvzuiIUlSI3A0sByoFEJ0Dz/UA1QeomVNFe4Cbgbs4Z9LgSEhhDn885F2z5uAfuCXw2GyX0iS5OcIv89C/P/27iZEjiIM4/j/wcSwGtCoIMoqixg8iBrFQzAeJHoKogeFIAGD5JSD6EUkeBI8iXiIiqB4CBI8qFEXD6JuRATFaGBN/EKNLjHBfB1cUSQs4fFQtdJsnDUZZrZDz/ODZqprhpkq3mXeqeraah8GngEOUhLBLLCXbse6qVd8B/YdNwoJYeRIWgm8CTxq+4/mcy7rjDuz1ljS3cAx23vbbssSWgbcArxo+2bgLxZMD3UtzgB1zvxeSkK8EriQ06dVRsKw4jsKCeEwcFXjfLzWdZKk5ZRksNP2rlp9dH4IWR+PtdW+IVgH3CNphjIduJ4yv35xnVaA7sX8EHDI9uf1/A1KguhynAHuAn6xfdz2HLCLEv8ux7qpV3wH9h03CgnhC2B1XYlwPuUi1GTLbRqKOnf+CvCd7WcbT00Cm2t5M/DOUrdtWGxvsz1ue4IS2922NwEfAffXl3Wtz0eAXyVdV6vuBL6lw3GuDgJrJV1Q/9bn+93ZWC/QK76TwIN1tdFaYLYxtXR2bHf+ADYAPwAHgCfabs8Q+3k7ZRi5D5iuxwbKnPoU8CPwIXBJ220dUv/vAN6t5WuAPcBPwOvAirbbN+C+rgG+rLF+G1g1CnEGngS+B74GXgVWdDHWwGuU6yRzlBHhll7xBURZSXkA2E9ZhdXX52brioiIAEa18E4DAAABk0lEQVRjyigiIs5AEkJERABJCBERUSUhREQEkIQQERFVEkLEIiSdkjTdOAa2YZykieZulhFtW/b/L4kYaX/bXtN2IyKWQkYIEX2QNCPpaUn7Je2RdG2tn5C0u+5LPyXp6lp/uaS3JH1Vj9vqW50n6eW6x//7ksZa61SMvCSEiMWNLZgy2th4btb2DcDzlB1XAZ4Ddti+EdgJbK/124GPbd9E2Xfom1q/GnjB9vXA78B9Q+5PRE/5T+WIRUj60/bK/6ifAdbb/rluKHjE9qWSTgBX2J6r9b/ZvkzScWDc9snGe0wAH7jc8ARJjwPLbT81/J5FnC4jhIj+uUf5bJxslE+R63rRoiSEiP5tbDx+VsufUnZdBdgEfFLLU8BW+Pf+zxctVSMjzlR+jUQsbkzSdOP8PdvzS09XSdpH+ZX/QK17mHIns8codzV7qNY/ArwkaQtlJLCVsptlxDkj1xAi+lCvIdxq+0TbbYkYlEwZRUQEkBFCRERUGSFERASQhBAREVUSQkREAEkIERFRJSFERAQA/wBmCviIUvPYtwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ScYAePIrqfkY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}