Nomades2_0_Rashid_05(withoutTranningAndBackward).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Pt-home/Notebooks/blob/main/Nomades2_0_Rashid_05(withoutTranningAndBackward).ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy\n",
        "# библиотека scipy.special содержит сигмоиду expit()\n",
        "import scipy.special\n",
        "import matplotlib.pyplot\n",
        "%matplotlib inline"
      ],
      "metadata": {
        "id": "uo9X1Kg96cc_"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "XGT7rXEVzWGg"
      },
      "outputs": [],
      "source": [
        "# определение класса нейронной сети class neuralNetwork:\n",
        "class neuralNetwork:\n",
        "\n",
        "  def __init__(self, inputnodes, hiddennodes, outputnodes, learningrate) :\n",
        "    # задать количество узлов во входном, скрытом и выходном слое\n",
        "    self.inodes = inputnodes\n",
        "    self.hnodes = hiddennodes\n",
        "    self.onodes = outputnodes\n",
        "\n",
        "    # коэффициент обучения\n",
        "    self.lr = learningrate\n",
        "\n",
        "    #матрицы весов\n",
        "    self.wih = numpy.random.rand(self.hnodes, self.inodes) - 0.5\n",
        "    self.who = numpy.random.rand(self.onodes, self.hnodes) - 0.5\n",
        "\n",
        "    # использование сигмоиды в качестве функции активации\n",
        "    self.activation_function = lambda x: scipy.special.expit(x)\n",
        "    #обратная активация\n",
        "    self.inverse_activation_function = lambda x: scipy.special.logit(x)\n",
        "\n",
        "  def train(self, inputs_list, targets_list):\n",
        "\n",
        "    # Прямой проход\n",
        "\n",
        "    # преобразовать список входных значений в двухмерный массив\n",
        "    inputs =  numpy.array(inputs_list, ndmin = 2).T\n",
        "    targets = numpy.array(targets_list, ndmin = 2).T\n",
        "\n",
        "    # рассчитать входящие сигналы для скрытого слоя\n",
        "    hidden_inputs = numpy.dot(self.wih, inputs)\n",
        "    # рассчитать исходящие сигналы для скрытого слоя\n",
        "    hidden_outputs = self.activation_function(hidden_inputs)\n",
        "\n",
        "    # рассчитать входящие сигналы для выходного слоя\n",
        "    final_inputs = numpy.dot(self.who, hidden_outputs)\n",
        "    # рассчитать исходящие сигналы для выходного слоя\n",
        "    final_outputs = self.activation_function(final_inputs)\n",
        "\n",
        "    # Анализ и коррекция весов\n",
        "\n",
        "    # ошибка = целевое значение - фактическое значение\n",
        "    output_errors = targets - final_outputs\n",
        "\n",
        "    # ошибки скрытого слоя - это ошибки output_errors, распределенные пропорционально весовым коэффициентам связей и рекомбинированные на скрытых узлах\n",
        "    hidden_errors = numpy.dot(self.who.T, output_errors)\n",
        "\n",
        "    # обновить весовые коэффициенты связей между скрытым и выходным слоями\n",
        "    self.who += self.lr * numpy.dot((output_errors * final_outputs * (1.0 - final_outputs)), numpy.transpose(hidden_outputs))\n",
        "    # обновить весовые коэффициенты связей между входным и скрытым слоями\n",
        "    self.wih += self.lr * numpy.dot((hidden_errors * hidden_outputs * (1.0 - hidden_outputs)), numpy.transpose(inputs))\n",
        "\n",
        "  def query(self, inputs_list):\n",
        "    # преобразовать список входных значений в двухмерный массив\n",
        "    inputs = numpy.array(inputs_list, ndmin = 2).T\n",
        "\n",
        "    # рассчитать входящие сигналы для скрытого слоя\n",
        "    hidden_inputs = numpy.dot(self.wih, inputs)\n",
        "    # рассчитать исходящие сигналы для скрытого слоя\n",
        "    hidden_outputs = self.activation_function(hidden_inputs)\n",
        "\n",
        "    # рассчитать входящие сигналы для выходного слоя\n",
        "    final_inputs = numpy.dot(self.who, hidden_outputs)\n",
        "    # рассчитать исходящие сигналы для выходного слоя\n",
        "    final_outputs = self.activation_function(final_inputs)\n",
        "\n",
        "    return final_outputs\n",
        "\n",
        "  def backquery(self, targets_list):\n",
        "     # transpose the targets list to a vertical array\n",
        "     final_outputs = numpy.array(targets_list, ndmin=2).T\n",
        "\n",
        "     # calculate the signal into the final output layer\n",
        "     final_inputs =  self.inverse_activation_function(final_outputs)\n",
        "\n",
        "     # calculate the signal out of the hidden layer\n",
        "     hidden_outputs = numpy.dot(self.who.T, final_inputs)\n",
        "\n",
        "     # scale them back to 0.01 to .99\n",
        "     hidden_outputs -= numpy.min(hidden_outputs)\n",
        "     hidden_outputs /= numpy.max(hidden_outputs)\n",
        "     hidden_outputs *= 0.98\n",
        "     hidden_outputs += 0.01\n",
        "\n",
        "     # calculate the signal into the hidden layer\n",
        "     hidden_inputs = self.inverse_activation_function(hidden_outputs)\n",
        "\n",
        "     # calculate the signal out of the input layer\n",
        "     inputs = numpy.dot(self.wih.T, hidden_inputs)\n",
        "\n",
        "     # scale them back to 0.01 to .99\n",
        "     inputs -= numpy.min(inputs)\n",
        "     inputs /= numpy.max(inputs)\n",
        "     inputs *= 0.98\n",
        "     inputs += 0.01\n",
        "\n",
        "     return inputs"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# количество входных, скрытых и выходных узлов\n",
        "input_nodes = 784\n",
        "hidden_nodes = 200\n",
        "output_nodes = 10\n",
        "# коэффициент обучения равен 0,1\n",
        "learning_rate = 0.1\n",
        "\n",
        "epochs = 5"
      ],
      "metadata": {
        "id": "q-rExmIw23tP"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# создать экземпляр нейронной сети\n",
        "n = neuralNetwork(input_nodes, hidden_nodes, output_nodes, learning_rate)"
      ],
      "metadata": {
        "id": "GPcbv6Hs4R8n"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# загрузить cеть\n",
        "f = open('wih_' + str(hidden_nodes) + '_' + str(learning_rate) + '_' + str(epochs) + '.dat','r')\n",
        "record = f.readlines()[0].split(',')\n",
        "n.wih = numpy.asfarray(record).reshape(n.hnodes, n.inodes)\n",
        "n.wih.shape"
      ],
      "metadata": {
        "id": "XQhMlR4COwKv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "928b4a6f-d76d-4e7f-accb-e68fc6133204"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(200, 784)"
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "f = open('who_' + str(hidden_nodes) + '_' + str(learning_rate) + '_' + str(epochs) + '.dat','r')\n",
        "record = f.readlines()[0].split(',')\n",
        "n.who = numpy.asfarray(record).reshape(n.onodes, n.hnodes)\n",
        "n.who.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rUyfJ7baDb8K",
        "outputId": "4403fcd9-0f36-4b9e-b90b-f2b50ac6817e"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(10, 200)"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# загрузить в список тестовый набор данных CSV-файла набора MNIST\n",
        "test_data_file = open(\"mnist_test.csv\", 'r')\n",
        "test_data_list = test_data_file.readlines()\n",
        "test_data_file.close()"
      ],
      "metadata": {
        "id": "fdLlE80uFVDX"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# тестирование нейронной сети\n",
        "\n",
        "# журнал оценок работы сети, первоначально пустой\n",
        "scorecard = []\n",
        "# перебрать все записи в тестовом наборе данных\n",
        "for record in test_data_list:\n",
        "\n",
        "  # получить список значений, используя символы запятой (1,1) в качестве разделителей\n",
        "  all_values = record.split(\",\")\n",
        "  # масштабировать и сместить входные значения\n",
        "  inputs = (numpy.asfarray(all_values[1:]) / 255.0 * 0.99) + 0.01\n",
        "\n",
        "  outputs = n.query(inputs)\n",
        "\n",
        "  # индекс наибольшего значения является маркерным значением\n",
        "  label = numpy.argmax(outputs)\n",
        "  correct_label = int(all_values[0])\n",
        "\n",
        "  if correct_label==label:\n",
        "    scorecard.append(1)\n",
        "  else:\n",
        "    scorecard.append(0)\n",
        "\n",
        "print(f\"Эффективность: {sum(scorecard) / len(scorecard)}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9VnVKfy_HKAR",
        "outputId": "857fb0cb-297a-449b-faef-4ad433edfac0"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Эффективность: 0.9713\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 0\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "OYV5Bqx1K8Ar",
        "outputId": "21934b62-cf08-4a54-ceea-8613238239ff"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c793b5e0>"
            ]
          },
          "metadata": {},
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 1\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "x2urxsXVL_cE",
        "outputId": "a5076c0f-6084-4851-eb41-a03bc250cacb"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c61e3ca0>"
            ]
          },
          "metadata": {},
          "execution_count": 20
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaAAAAGdCAYAAABU0qcqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAoc0lEQVR4nO3de3CV9Z3H8U8IySFACITcIYSEq1yLKJFy2QhsILYuKNuRSmexo7i4wVllu+3QabXubieVznRtHRansxXsrHhhWmCxiqtQQrVAhwiL3DIkyyWQCxDMhQBJSJ79gyFtBCXfnwm/JL5fM2cGTp4Pzy/PeU4+nJxzvicsCIJAAADcZj18LwAA8OVEAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwoqfvBXxac3OzSktLFR0drbCwMN/LAQAYBUGg2tpapaSkqEePz36c0+kKqLS0VKmpqb6XAQD4gkpKSjR48ODP/HqnK6Do6GhJ0tKlSxUZGdnm3IULF8z7GjlypDkjSZMmTTJn3nrrLXPG5RHgoEGDbst+JOnAgQPmzOzZs82ZDRs2mDOZmZnmjCSNGDHCnNmxY4c543LM77jjDnOmsrLSnJGkY8eOmTMu5151dbU543K/LS4uNmekP/88sjh37pw5Y/lZd11TU5M547ov63FoaGjQL3/5y1vmOqyAVq9erZ/+9KcqLy/XxIkT9eKLL2rKlCm3zF2/Y0ZGRioUCrV5fy4HtVevXuaMJPXp08eccVmfyw8pl+/JtYAiIiLMmaioKHOmZ0/7aep62/bu3duc6cy3reU+9JdcbluXfd2u+63LfiS378nl2HX2AnI9j251nnfIixDeeOMNrVixQs8++6w++ugjTZw4UXPnztXZs2c7YncAgC6oQwroZz/7mZYuXapvf/vbGjNmjF566SX17t1bL7/8ckfsDgDQBbV7ATU0NKigoEBz5sz580569NCcOXO0a9euG7avr69XTU1NqwsAoPtr9wI6f/68mpqalJiY2Or6xMRElZeX37B9Xl6eYmJiWi68Ag4Avhy8vxF15cqVqq6ubrmUlJT4XhIA4DZo91fBxcXFKTw8XBUVFa2ur6ioUFJS0g3bh0Ih51dYAAC6rnZ/BBQZGanJkydr27ZtLdc1Nzdr27Ztmjp1anvvDgDQRXXI+4BWrFihJUuW6K677tKUKVP0wgsvqK6uTt/+9rc7YncAgC6oQwrooYce0rlz5/TMM8+ovLxcX/nKV7R169YbXpgAAPjy6rBJCMuXL9fy5cud84cPHza9A/6TTz4x7+OrX/2qOSNJBw8eNGcSEhLMmb1795ozQRCYMy6jQyR97pDBz+LyMvvs7GxzxuXYSdeew7Ry+dVybW2tObNnzx5zxmWygyTddddd5symTZvMmZycHHPG5XsaMGCAOSNJV65cMWdcJhRs377dnHG93/7qV78yZ6wvDmvrcfP+KjgAwJcTBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALzosGGkX1R8fLwiIiLavP19991n3ofLAFPJbUDh4MGDzZn09HRzprCw0Jxx/UDAqKgoc8blODQ2NpozdXV15owknTx50py58847zRmXoaxnz541Z2bNmmXOSNKMGTPMmVGjRpkzH374oTlzsw+2vBXLz5K/1NzcbM5MmjTJnLl8+bI5k5GRYc5I0u9+9ztzxvpzpa0DWXkEBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC867TTsuLg405Tm8+fPm/dRVFRkzkhS3759zZkgCMyZ+vp6cyYtLc2cOXTokDkjSVevXjVn9u/f77Qvq+zsbKecy5TqLVu2mDOTJ082ZxYtWmTOuEzQlqSqqipzxmWi87Bhw8yZ2tpac8Zlorrkdr91uT89+OCD5sy7775rzkhu99usrCzT9vX19SooKLjldjwCAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvOu0w0iAITIMAXQZ3RkdHmzOSNH36dHPm8OHD5kxDQ4M5c/ToUXOmvLzcnJGkO+64w5y5cuWKOeMyGPM3v/mNOSNJMTEx5kxGRoY543I7uRy7/v37mzOS1KOH/f+mYWFh5szHH39szqSnp5szLj8fJKmystKciYiIMGd+8pOfmDPLli0zZyTpk08+MWd69rRVRVvPVR4BAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXnXYY6enTp01D/ZKSksz7OHjwoDkjSVlZWeZMZGSkOZOQkGDOlJaWmjNxcXHmjCTTsNjrJk6caM6cOnXKnOnXr585I0nJycnmzKFDh8wZ63BHSRozZow5U1ZWZs5I0v79+80Zl2MeHh5uzpw4ccKcSUtLM2ckt/tTfHy8OXPnnXeaM1evXjVnJLdhpNafX20d/sojIACAFxQQAMCLdi+gH/3oRwoLC2t1GT16dHvvBgDQxXXIc0Bjx47V+++//+edOPy+GwDQvXVIM/Ts2dPpRQEAgC+PDnkO6NixY0pJSVFGRoYWL178ua9iqq+vV01NTasLAKD7a/cCyszM1Lp167R161atWbNGx48f14wZM1RbW3vT7fPy8hQTE9NySU1Nbe8lAQA6oXYvoJycHH3jG9/QhAkTNHfuXL399tuqqqrSm2++edPtV65cqerq6pZLSUlJey8JANAJdfirA/r376+RI0eqqKjopl8PhUIKhUIdvQwAQCfT4e8DunjxooqLi53eYQ4A6L7avYC+853vKD8/XydOnNAf//hHPfDAAwoPD9c3v/nN9t4VAKALa/dfwZ0+fVrf/OY3VVlZqfj4eE2fPl27d+92mo8EAOi+wgKXiZIdqKamRjExMXr77bfVp0+fNuc2btxo3teIESPMGUnavn27OTNq1Chz5pVXXjFnHnvsMXMmJyfHnJHcBkm6fE/Nzc3mjKuvfOUr5ozLsFSXV3u63E4ua5Pk9HaI//zP/zRnhg4das64qKqqcspNmzbNnHnnnXfMGZdBrgMHDjRnJLdhrhkZGabt6+vr9cILL6i6uvpzvzdmwQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFx3+gXSuTp48qaioqDZv7zJ0cfjw4eaM5DbEdMCAAebMwoULzZnjx4+bM6+99po5I0mVlZXmzOXLl82Z+vp6c+bcuXPmjCTt2bPHnFm1apU5s3btWnMmKSnJnHH9iPsXX3zRnDl//rw5M2jQIHPG5RxfsGCBOSNJR48eNWdSUlLMmSlTppgzLueqJMXFxZkzMTExpu2vXLnSpu14BAQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvOu007MbGRvXs2fblpaWlmffx3nvvmTOS22Rdl6m64eHh5syMGTPMmcjISHNGkgoKCswZl6m/7777rjnTv39/c0aSkpOTzZmXXnrJnJk1a5Y547K2hoYGc0ayTz+WpKtXr5ozLvdbl+NQVFRkzkhSWVmZOfOtb33LnHH5nlyOtyTFx8ebM4WFhU77uhUeAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAF512GOnHH39sGpIZBIF5H3fffbc5I0k7d+40Z5qamswZl+/pgw8+MGdOnTplzrjm3n//fXPm7/7u78yZkydPmjOS1NzcbM4sXLjQnElKSjJnLly4YM7U1NSYM5LbsNTDhw+bMy5DY8+cOWPOjBkzxpyRpH79+pkzLsN9jx07Zs643rYnTpwwZ1JTU03bt/UY8AgIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALzotMNIz549q4iIiDZv7zLU8NKlS+aMJA0cONCccRlGOn/+fHPm4MGD5kx1dbU5I0nf+MY3zJlz586ZMy6DGhctWmTOSFJUVJQ5c/XqVXNmwoQJ5syAAQPMma1bt5ozkvT222+bMy73i759+5ozLrfR2LFjzRlJqqioMGcGDx5szoSHh5szLkNFJbfhuWvWrDFt39DQ0KbteAQEAPCCAgIAeGEuoJ07d+r+++9XSkqKwsLCtGnTplZfD4JAzzzzjJKTkxUVFaU5c+Y4/QoFANC9mQuorq5OEydO1OrVq2/69VWrVukXv/iFXnrpJe3Zs0d9+vTR3LlzdeXKlS+8WABA92F+EUJOTo5ycnJu+rUgCPTCCy/oBz/4QcsT6L/+9a+VmJioTZs2OT8xDADoftr1OaDjx4+rvLxcc+bMabkuJiZGmZmZ2rVr100z9fX1qqmpaXUBAHR/7VpA5eXlkqTExMRW1ycmJrZ87dPy8vIUExPTcrF+9jgAoGvy/iq4lStXqrq6uuVSUlLie0kAgNugXQsoKSlJ0o1v3qqoqGj52qeFQiH169ev1QUA0P21awGlp6crKSlJ27Zta7mupqZGe/bs0dSpU9tzVwCALs78KriLFy+qqKio5e/Hjx/X/v37FRsbqyFDhuipp57Sv/3bv2nEiBFKT0/XD3/4Q6WkpGjBggXtuW4AQBdnLqC9e/fq3nvvbfn7ihUrJElLlizRunXr9N3vfld1dXV6/PHHVVVVpenTp2vr1q3q1atX+60aANDlhQVBEPhexF+qqalRTEyMZsyYoZ49296P999/v3lfpaWl5ox07XkrK5cBhVlZWebM7XzDr8tQyH379pkzU6ZMMWcuXLhgzkhyeg7S5a0DLudQQkKCOfPxxx+bM5L0wQcfmDNlZWXmTFxcnDnjct4lJyebM5IUHx9vzlRVVZkzLj+GXe5LktSjh/2Zl5kzZ5q2v3TpkhYvXqzq6urPvU95fxUcAODLiQICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC/MH8dwu8ycOdP0EQ7V1dXmfbhOjo6OjjZnXCYFV1ZWmjN9+/Y1Zw4dOmTOSNLYsWPNmZEjRzrty+r06dNOOZdJwUOGDDFnXM49l+8pJibGnJHcJk7HxsaaM6mpqebMH/7wB3OmoaHBnJGkiIgIc8blfuFy216+fNmckaTx48ebM7W1tabt27o2HgEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBeddhjpvn37nAYBWrgMCJWkv/mbvzFnQqGQORMEgTkTHx9vzkyaNMmckdwGd7oMkrzrrrvMmZqaGnNGks6ePWvOHD161JzJysoyZ0aNGmXO/PznPzdnXF24cMGc2bNnjznTu3dvc+bw4cPmjCSlpaWZM9u3bzdnEhISzJnMzExzRnI7FtbjwDBSAECnRgEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvOu0w0iNHjpiGXY4dO9a8j4ULF5ozknTmzBlzxmXYoMuw1L59+5ozr732mjkjuQ1LLS4uNmf+93//15y59957zRlJeuedd8yZr3/96+bMK6+8Ys64DJp1GeQqSYcOHTJnysvLzZmePe0/gr72ta+ZM7/85S/NGcltsOjf//3fmzPf//73zZnp06ebM5J06tQpc6atw0Wva2hoaNN2PAICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC867TDSv/3bv1UoFGrz9qWlpeZ9DB482JyRpIiICHMmKirKnNm6das5c/ToUXPGZeip5DZI0kV2drY5U1JS4rQvl+/pww8/NGeSk5PNmVmzZpkzGzZsMGckt2Gu586dM2fOnz9vzqxcudKcsQ7TvG7MmDHmzKuvvmrOTJo0yZy5cOGCOSNJ48aNM2esA4EbGxvbtB2PgAAAXlBAAAAvzAW0c+dO3X///UpJSVFYWJg2bdrU6uuPPPKIwsLCWl3mzZvXXusFAHQT5gKqq6vTxIkTtXr16s/cZt68eSorK2u5uH7gGQCg+zI/45qTk6OcnJzP3SYUCikpKcl5UQCA7q9DngPasWOHEhISNGrUKD3xxBOqrKz8zG3r6+tVU1PT6gIA6P7avYDmzZunX//619q2bZuef/555efnKycnR01NTTfdPi8vTzExMS2X1NTU9l4SAKATavc3cixatKjlz+PHj9eECRM0bNgw7dixQ7Nnz75h+5UrV2rFihUtf6+pqaGEAOBLoMNfhp2RkaG4uDgVFRXd9OuhUEj9+vVrdQEAdH8dXkCnT59WZWWl0zu/AQDdl/lXcBcvXmz1aOb48ePav3+/YmNjFRsbq+eee04LFy5UUlKSiouL9d3vflfDhw/X3Llz23XhAICuzVxAe/fubTUn6vrzN0uWLNGaNWt04MABvfLKK6qqqlJKSoqys7P1r//6r6a5bgCA7s9cQFlZWQqC4DO//u67736hBV13+fJlNTc3t3l7l4Ga1gF71w0aNMiciY+PN2dcng/76KOPzJnhw4ebM5I+85WNnyctLc2ciY6ONmeOHDlizkjSY489Zs64DLp0+Z5c7lvjx483Z1z3tWbNGnNm8eLF5szDDz9szrz++uvmjCR98skn5kxdXZ05k5iYaM5cunTJnJGkhoYGc+aee+4xbX/lyhVt2bLlltsxCw4A4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABetPtHcreXnTt3Kjw8vM3bZ2Vlmffx3//93+aMJD366KPmjMuUausEWkkaNmyYOXP+/HlzRpKioqLMmbvuusucqampMWdcP3+qsbHRnHGZjv673/3OnNmwYYM5M3ToUHNGcjuPRo8ebc64fEzLmTNnzJns7GxzRnKbbJ2enm7OfNYnRn+e+fPnmzOS2jSl+tMGDBhg2r6tE7d5BAQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXnTaYaT33XefaVBhfX29eR9//dd/bc5I0jvvvGPOTJ061ZzZt2+fOZORkWHOlJWVmTOSVFFRYc6EhYU57cvqySefdMqVlpaaM8eOHTNndu/ebc706dPHnBk4cKA5I0lnz541Z3r16mXOXLp0yZwZPny4OeN63sXHx5szLoNFo6OjzZk333zTnJGkp59+2px5+eWXTdu3dagvj4AAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwItOO4w0MjLSNIz0D3/4g3kfd955pzkjuQ1D7Nu3rzlz9OhRc8ZlIKTlOP+lcePGmTMff/yxOTN//nxz5r/+67/MGUlKSEgwZwoLC80Zl8GiLgMrXYb0StK3vvUtc8bltk1NTTVn9u/fb8643tfPnDljzrgMEXYZllpZWWnOSFJ+fr45k5ycbNq+oaGhTdvxCAgA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvAgLgiDwvYi/VFNTo5iYGD3//POKiopqc85lMF9NTY05I0lJSUnmzOnTp80Z6wBASSoqKjJnhg4das5I0tmzZ82Z2NhYc+by5cvmTL9+/cwZSaqrqzNnXIbG9uhh/7+fy+DOiooKc0a6NgzYateuXebMggULzBmXwb6XLl0yZyS389XlPnjhwgVzZtSoUeaM5Hb8fvOb35i2b2pq0v79+1VdXf2590UeAQEAvKCAAABemAooLy9Pd999t6Kjo5WQkKAFCxbc8FkoV65cUW5urgYOHKi+fftq4cKFzr8GAAB0X6YCys/PV25urnbv3q333ntPjY2Nys7ObvV786efflpbtmzRhg0blJ+fr9LSUj344IPtvnAAQNdm+kTUrVu3tvr7unXrlJCQoIKCAs2cOVPV1dX61a9+pfXr12vWrFmSpLVr1+qOO+7Q7t27dc8997TfygEAXdoXeg6ourpa0p9fKVJQUKDGxkbNmTOnZZvRo0dryJAhn/kKmfr6etXU1LS6AAC6P+cCam5u1lNPPaVp06Zp3LhxkqTy8nJFRkaqf//+rbZNTExUeXn5Tf+dvLw8xcTEtFxcXmoKAOh6nAsoNzdXBw8e1Ouvv/6FFrBy5UpVV1e3XEpKSr7QvwcA6BpMzwFdt3z5cr311lvauXOnBg8e3HJ9UlKSGhoaVFVV1epRUEVFxWe+eTMUCikUCrksAwDQhZkeAQVBoOXLl2vjxo3avn270tPTW3198uTJioiI0LZt21quKyws1KlTpzR16tT2WTEAoFswPQLKzc3V+vXrtXnzZkVHR7c8rxMTE6OoqCjFxMTo0Ucf1YoVKxQbG6t+/frpySef1NSpU3kFHACgFVMBrVmzRpKUlZXV6vq1a9fqkUcekST9+7//u3r06KGFCxeqvr5ec+fO1X/8x3+0y2IBAN1Hpx1Geu+996pnz7b34/Dhw837qqqqMmcktwGPLq/uswxjva6hocGciYuLM2ckKSIiwpxJSEgwZ5qamsyZ9evXmzOSWl7RabF9+3ZzZtq0aeaMy3Olhw8fNmck6etf/7o5c+TIEXOmoKDAnBkwYIA5M3LkSHNGkmbPnm3OuAzp3b17tzlTWlpqzkhq9bx9W40ZM8a0fX19vZ5//nmGkQIAOicKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8cPpE1NthxIgRioyMbPP2sbGx5n3U1dWZM5K0bNkyc+bnP/+5OZOWlmbOuAw3T05ONmcktXwelMX//M//mDNhYWHmjMu0aUnKzs42Z+rr680Zl+9p1qxZ5szkyZPNGUl6/fXXzZkePez/n3X5ng4dOmTOuJ7jBw4cMGdqa2vNGZdp+S7HTnI7fn369DFtHx4e3qbteAQEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF502mGkUVFRCoVCbd5+//795n2MHj3anJGkH//4x+bMpUuXzJmsrCxzprGx0ZzZs2ePOSNJ8fHx5syZM2fMmdzcXHPm1KlT5ozkNnxy0qRJ5kxCQoI54zL89fLly+aM5DbM1TI8+Lpjx46ZM1/72tfMmRMnTpgzkjR9+nRzpqSkxJxxOXbbt283ZyQpJyfHnLEOWG1qamrTdjwCAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvOu0w0h49eqhHj7b3o8vgzkOHDpkzkvTYY4+ZMxUVFebMkCFDzJkBAwaYM7169TJnJKm4uNicWbx4sTkzfvx4c+bIkSPmjCQtXLjQnCkrKzNnPvnkE3NmwYIF5ozLEFxJ2rJliznjMlDTZdjnhAkTzBnrMM3rtm3bZs6MGTPGnHEZlup6vz137pw5k5ycbNq+rUNweQQEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF502mGk//d//6eIiIg2b+8yzO+rX/2qOSNJKSkp5ozL4NP09HRz5o033jBnBg0aZM5I0oULF8yZ+vp6c+bMmTPmTFNTkzkjSX/84x/NmZ497Xcjl0GuLsM+T506Zc5IUnh4uDlzxx13mDMNDQ3mzMsvv2zOuBw7SUpISDBnSkpKzJnq6mpzZtGiReaMJP3pT38yZ3bs2GHavq23K4+AAABeUEAAAC9MBZSXl6e7775b0dHRSkhI0IIFC1RYWNhqm6ysLIWFhbW6LFu2rF0XDQDo+kwFlJ+fr9zcXO3evVvvvfeeGhsblZ2drbq6ulbbLV26VGVlZS2XVatWteuiAQBdn+nZ061bt7b6+7p165SQkKCCggLNnDmz5frevXsrKSmpfVYIAOiWvtBzQNdfuREbG9vq+ldffVVxcXEaN26cVq5c+bkfC1xfX6+amppWFwBA9+f8Muzm5mY99dRTmjZtmsaNG9dy/cMPP6y0tDSlpKTowIED+t73vqfCwkL99re/vem/k5eXp+eee851GQCALsq5gHJzc3Xw4EF98MEHra5//PHHW/48fvx4JScna/bs2SouLtawYcNu+HdWrlypFStWtPy9pqZGqamprssCAHQRTgW0fPlyvfXWW9q5c6cGDx78udtmZmZKkoqKim5aQKFQSKFQyGUZAIAuzFRAQRDoySef1MaNG7Vjx442vVN///79kqTk5GSnBQIAuidTAeXm5mr9+vXavHmzoqOjVV5eLkmKiYlRVFSUiouLtX79et13330aOHCgDhw4oKefflozZ87UhAkTOuQbAAB0TaYCWrNmjaRrbzb9S2vXrtUjjzyiyMhIvf/++3rhhRdUV1en1NRULVy4UD/4wQ/abcEAgO7B/Cu4z5Oamqr8/PwvtCAAwJdDp52GXVZWZpoyPHv2bPM+CgoKzBlJOnnypDnTv39/c2bPnj3mTHx8vDkzfvx4c0aSzp8/b864TBeura01Z1x/5evyPrTJkyebM2PHjjVnDhw4YM7ExcWZM5K0efNmc6Zfv37mTGlpqTnjMjn63nvvNWekG9983xZDhw41Zx566CFzxmVtktt9sLGx0bT91atX27Qdw0gBAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwItOO4x00KBBioiIaPP2LgP2XIYaSm7DMR944AFz5vDhw+ZMZGTkbdmPJM2fP9+cuXTpkjlz/UMNLU6cOGHOSLrlJ/zeTHFxsTmTkZFhztxqGv3NuH4QpMtte/DgQXNm1KhR5szFixfNGZcBwpI0ZswYc+aee+4xZ1wGD1sHhF63e/duc8ZlyHFb8AgIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB40elmwV2fd2Wdc3T58mXzvurr680ZSWpoaDBn6urqzJkrV66YM01NTeaM63Fwmet2u24n1zlZLretC5fb1mUWnMttJN2+2+l27cflfiFJYWFh5ozLMXf5nlzPVZf7hnVf1/dxq3M2LHA5qzvQ6dOnlZqa6nsZAIAvqKSk5HMH/Ha6AmpublZpaamio6Nv+N9HTU2NUlNTVVJSon79+nlaoX8ch2s4DtdwHK7hOFzTGY5DEASqra1VSkqKevT47Gd6Ot2v4Hr06HHLkfj9+vX7Up9g13EcruE4XMNxuIbjcI3v4xATE3PLbXgRAgDACwoIAOBFlyqgUCikZ599VqFQyPdSvOI4XMNxuIbjcA3H4ZqudBw63YsQAABfDl3qERAAoPuggAAAXlBAAAAvKCAAgBddpoBWr16toUOHqlevXsrMzNSf/vQn30u67X70ox8pLCys1WX06NG+l9Xhdu7cqfvvv18pKSkKCwvTpk2bWn09CAI988wzSk5OVlRUlObMmaNjx475WWwHutVxeOSRR244P+bNm+dnsR0kLy9Pd999t6Kjo5WQkKAFCxaosLCw1TZXrlxRbm6uBg4cqL59+2rhwoWqqKjwtOKO0ZbjkJWVdcP5sGzZMk8rvrkuUUBvvPGGVqxYoWeffVYfffSRJk6cqLlz5+rs2bO+l3bbjR07VmVlZS2XDz74wPeSOlxdXZ0mTpyo1atX3/Trq1at0i9+8Qu99NJL2rNnj/r06aO5c+c6DfzszG51HCRp3rx5rc6P11577TausOPl5+crNzdXu3fv1nvvvafGxkZlZ2e3Gvb79NNPa8uWLdqwYYPy8/NVWlqqBx980OOq219bjoMkLV26tNX5sGrVKk8r/gxBFzBlypQgNze35e9NTU1BSkpKkJeX53FVt9+zzz4bTJw40fcyvJIUbNy4seXvzc3NQVJSUvDTn/605bqqqqogFAoFr732mocV3h6fPg5BEARLliwJ5s+f72U9vpw9ezaQFOTn5wdBcO22j4iICDZs2NCyzZEjRwJJwa5du3wts8N9+jgEQRD81V/9VfCP//iP/hbVBp3+EVBDQ4MKCgo0Z86clut69OihOXPmaNeuXR5X5sexY8eUkpKijIwMLV68WKdOnfK9JK+OHz+u8vLyVudHTEyMMjMzv5Tnx44dO5SQkKBRo0bpiSeeUGVlpe8ldajq6mpJUmxsrCSpoKBAjY2Nrc6H0aNHa8iQId36fPj0cbju1VdfVVxcnMaNG6eVK1c6fzxHR+l0w0g/7fz582pqalJiYmKr6xMTE3X06FFPq/IjMzNT69at06hRo1RWVqbnnntOM2bM0MGDBxUdHe17eV6Ul5dL0k3Pj+tf+7KYN2+eHnzwQaWnp6u4uFjf//73lZOTo127dik8PNz38tpdc3OznnrqKU2bNk3jxo2TdO18iIyMVP/+/Vtt253Ph5sdB0l6+OGHlZaWppSUFB04cEDf+973VFhYqN/+9rceV9tapy8g/FlOTk7LnydMmKDMzEylpaXpzTff1KOPPupxZegMFi1a1PLn8ePHa8KECRo2bJh27Nih2bNne1xZx8jNzdXBgwe/FM+Dfp7POg6PP/54y5/Hjx+v5ORkzZ49W8XFxRo2bNjtXuZNdfpfwcXFxSk8PPyGV7FUVFQoKSnJ06o6h/79+2vkyJEqKiryvRRvrp8DnB83ysjIUFxcXLc8P5YvX6633npLv//971t9fEtSUpIaGhpUVVXVavvuej581nG4mczMTEnqVOdDpy+gyMhITZ48Wdu2bWu5rrm5Wdu2bdPUqVM9rsy/ixcvqri4WMnJyb6X4k16erqSkpJanR81NTXas2fPl/78OH36tCorK7vV+REEgZYvX66NGzdq+/btSk9Pb/X1yZMnKyIiotX5UFhYqFOnTnWr8+FWx+Fm9u/fL0md63zw/SqItnj99deDUCgUrFu3Ljh8+HDw+OOPB/379w/Ky8t9L+22+qd/+qdgx44dwfHjx4MPP/wwmDNnThAXFxecPXvW99I6VG1tbbBv375g3759gaTgZz/7WbBv377g5MmTQRAEwU9+8pOgf//+webNm4MDBw4E8+fPD9LT04PLly97Xnn7+rzjUFtbG3znO98Jdu3aFRw/fjx4//33gzvvvDMYMWJEcOXKFd9LbzdPPPFEEBMTE+zYsSMoKytruVy6dKllm2XLlgVDhgwJtm/fHuzduzeYOnVqMHXqVI+rbn+3Og5FRUXBv/zLvwR79+4Njh8/HmzevDnIyMgIZs6c6XnlrXWJAgqCIHjxxReDIUOGBJGRkcGUKVOC3bt3+17SbffQQw8FycnJQWRkZDBo0KDgoYceCoqKinwvq8P9/ve/DyTdcFmyZEkQBNdeiv3DH/4wSExMDEKhUDB79uygsLDQ76I7wOcdh0uXLgXZ2dlBfHx8EBEREaSlpQVLly7tdv9Ju9n3LylYu3ZtyzaXL18O/uEf/iEYMGBA0Lt37+CBBx4IysrK/C26A9zqOJw6dSqYOXNmEBsbG4RCoWD48OHBP//zPwfV1dV+F/4pfBwDAMCLTv8cEACge6KAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAF/8PvEn2EBqpq5EAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 2\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "1ES2KCNcMRI6",
        "outputId": "5babac62-7984-4eeb-9883-d2ce6aa3790f"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c6288df0>"
            ]
          },
          "metadata": {},
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaAAAAGdCAYAAABU0qcqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAoZ0lEQVR4nO3dfXCV5Z3G8SsEckggnBhC3jABgrwKxBEBEUQsEUhbB5Xdgu3ugNOR0QZnke3WYaeV2t2Z7Oq069RldTrTlTojSt2psDoOqyCEaoEKgohAmsQoYEiAYHLynpA8+wdDahQwv9skdxK+n5kzA8l98dx5eHIuDufkd6KCIAgEAEAPG+B7AwCAaxMFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMCLgb438GVtbW0qKytTfHy8oqKifG8HAGAUBIFqamqUnp6uAQOu/Din1xVQWVmZMjIyfG8DAPANnTx5Utdff/0VP9/rCig+Pl6S9MgjjygUCnU6V19fbz5WYmKiOSNJ48aNM2e2bNliztx1113mTEtLizlz8803mzOStG3bNnPG5VHtxx9/bM4MGjTInJGkefPmmTNvvvmmOTNjxgxzJikpyZx54403zBnp4v9EWE2fPt2cudq/jq/E5XqIjY01ZyTp1ltvNWdqamrMmR07dpgzrv9QHzjQfrefmppqWt/Y2KjHHnus/f78insx76STNmzYoKeeekrl5eXKzs7WM888o5kzZ35t7tIdVCgUMhVQa2ureY+DBw82ZyQpLi7OnHG5Q3T5pnG5uIYOHWrOSG7nz6WAYmJizBnXAnL5u3XZn8vfbU9dd5JbAblcD9HR0eaMy/m23Jd8kcs5d7kv6smvqafui6Sv/37vlhchbN68WWvXrtX69ev1/vvvKzs7W4sWLdKZM2e643AAgD6oWwroV7/6lR588EE98MADmjx5sp577jnFxcXpv//7v7vjcACAPqjLC6i5uVkHDhxQTk7OXw8yYIBycnK0Z8+er6xvampSJBLpcAMA9H9dXkDnzp1Ta2urUlJSOnw8JSVF5eXlX1mfn5+vcDjcfuMVcABwbfD+g6jr1q1TdXV1++3kyZO+twQA6AFd/iq4pKQkRUdHq6KiosPHKyoqLvtSPuur3QAA/UOXPwKKiYnR9OnTO7yuva2tTTt27NDs2bO7+nAAgD6qW34OaO3atVqxYoVuueUWzZw5U08//bTq6ur0wAMPdMfhAAB9ULcU0LJly3T27Fk9/vjjKi8v10033aRt27Z95YUJAIBrV1QQBIHvTXxRJBJROBzWfffdZ/qJXZefqE5PTzdnJCkhIcGcaWxsNGeam5vNGReTJ092yh05csScSU5ONmfOnj1rzowYMcKckaSRI0eaMy5/ty7X0NGjR82ZzMxMc0aSjh07Zs4UFRWZM3feeac543KXVVdXZ85I9hE0knTq1Clz5oMPPjBnXMY5SXJ6IHC5VzBfTWNjo9avX6/q6moNGzbsiuu8vwoOAHBtooAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAX3TINuytEIhHTMNJvfetb5mO4vv23yxvo1dfXmzPnz583Z1zeUfajjz4yZyS3gZ/Dhw83Z0aPHm3ObNmyxZyRpLKyMnNmzJgx5szp06fNGZfhuWlpaeaMJA0dOtScWblypTlTWFhozrgoKCjokeNI0o033mjOuNw/fPlNPzurqanJnImLizOtHziwc9XCIyAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB40WunYd90002mqdMDBti79E9/+pM548plgnZ2drY54zLp9tixY+aMJNO08kuKiorMmcGDB5szLpOjJam8vNycee+998yZ3Nxcc+bUqVPmTENDgzkjSQkJCebM8ePHzZnm5mZzxuXv9o477jBnJOkvf/mLOfP++++bM6mpqebMhx9+aM5IUk1NjTlz4cIF0/rO/r3yCAgA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvOi1w0iHDx9uGkLpMuSyra3NnJHcBhtWVlaaMy77O3DggDkzY8YMc0aSPvvsM3MmNjbWnHEZCFldXW3OSNKJEyfMmW9/+9vmTFVVlTnjMnDX5bqTemZgpSRlZWWZMy7nbsKECeaMJA0ZMsScmTlzpjmzefNmc+bee+81ZyTp6NGj5sznn39uWt/Zocg8AgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAAL3rtMNIzZ84oFAp1en1jY6P5GJFIxJyRpE8//dScGTZsmDkTHR1tziQnJ5szu3fvNmckt6GsZ8+eNWcqKip6JCNJc+fONWcyMjLMmZ4aGvvb3/7WnJGk5cuXmzNBEJgzLsNfXb7Xa2trzRlJuu6668yZ48ePmzO33nqrOXP69GlzRpLi4+PNGetw2tbW1k6t4xEQAMALCggA4EWXF9DPf/5zRUVFdbhNnDixqw8DAOjjuuU5oBtvvFHbt2//60EG9tqnmgAAnnRLMwwcOFCpqand8UcDAPqJbnkOqKioSOnp6crKytIPfvCDq77SpampSZFIpMMNAND/dXkBzZo1Sxs3btS2bdv07LPPqrS0VLfffvsVX8aXn5+vcDjcfnN5SSsAoO/p8gLKzc3V3/7t32ratGlatGiR3njjDVVVVen3v//9ZdevW7dO1dXV7beTJ0929ZYAAL1Qt786ICEhQePHj1dxcfFlPx8KhUw/cAoA6B+6/eeAamtrVVJSorS0tO4+FACgD+nyAvrxj3+sgoICffLJJ/rTn/6ke++9V9HR0br//vu7+lAAgD6sy/8L7tSpU7r//vtVWVmpESNGaO7cudq7d69GjBjR1YcCAPRhUYHLBMFuFIlEFA6H9cQTT2jw4MGdzrkMn2xpaTFnJJn2dYnLAECXAYouexs7dqw5I7kNZXUZfOryg8wpKSnmjOR2TbgcKy4uzpzp7IDHL6qvrzdnJGno0KHmTGJiojnT0NBgzgwZMsSc2bRpkzkjScuWLTNnXP5uy8vLzRnX+y+Xwaf79+83rW9qatJTTz2l6urqqw5iZhYcAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHjR7W9I5yocDis2NrbT613eSXX+/PnmjCTTvi55++23e+Q4jY2N5syxY8fMGUkqKyszZ86dO2fOTJw40Zw5cuSIOSNJN954ozlTWVlpziQnJ5szZ8+eNWcmTJhgzkhug3A/+OADc2b8+PHmjMuA1VtuucWckS6+oaaVy/Bcl/dLq6mpMWcktyHC3YVHQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCi107DHj58uOLi4jq9PhwOm4/x+uuvmzOStGTJEnMmKirKnLnrrrvMmddee82ccdmb5DYx2eVr2r9/vznT0NBgzkhSTEyMOeMyXbitrc2cmTFjhjlTXV1tzkjSZ599Zs4kJiaaMz01vX3x4sXmjCQNGGD/N7rL98WiRYvMmXfffdeckdwmb1v3V1tbq6eeeupr1/EICADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC86LXDSD/66COFQqFOr09ISDAfIxKJmDOStGvXLnMmKyvLnNm9e7c5M2TIEHPmyJEj5owk07DYS86fP2/OXHfddebMyJEjzRlJunDhgjmzYsUKc6alpcWccRmUOnnyZHNGchvmWlBQYM5MmzbNnJkzZ445U1JSYs5I0sSJE80Zl2GkxcXF5ozLtep6rIMHD5rWd3bILI+AAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMCLXjuMdMSIEYqNje30+mPHjpmPMW7cOHNGkioqKsyZQYMGmTOdHej3RS6DEBMTE80ZSWptbTVnemqg5owZM8wZye38uQyazczMNGeGDh1qzvzyl780ZyS3AbBNTU3mTBAE5syJEyfMmQkTJpgzkrRjxw5z5nvf+545k5aWZs4MGOD2+CE6Otqc2bx5s2l9c3Nzp9bxCAgA4AUFBADwwlxAu3fv1t1336309HRFRUVpy5YtHT4fBIEef/xxpaWlKTY2Vjk5OSoqKuqq/QIA+glzAdXV1Sk7O1sbNmy47OeffPJJ/frXv9Zzzz2nffv2aciQIVq0aJHT8xkAgP7L/CKE3Nxc5ebmXvZzQRDo6aef1k9/+lMtWbJEkvTCCy8oJSVFW7Zs0fLly7/ZbgEA/UaXPgdUWlqq8vJy5eTktH8sHA5r1qxZ2rNnz2UzTU1NikQiHW4AgP6vSwuovLxckpSSktLh4ykpKe2f+7L8/HyFw+H2W0ZGRlduCQDQS3l/Fdy6detUXV3dfjt58qTvLQEAekCXFlBqaqqkr/6gZkVFRfvnviwUCmnYsGEdbgCA/q9LC2jMmDFKTU3t8NPDkUhE+/bt0+zZs7vyUACAPs78Krja2loVFxe3/760tFSHDh1SYmKiMjMztWbNGv3rv/6rxo0bpzFjxuhnP/uZ0tPTdc8993TlvgEAfZy5gPbv368777yz/fdr166VJK1YsUIbN27UT37yE9XV1WnVqlWqqqrS3LlztW3bNg0ePLjrdg0A6POiApdpgN0oEokoHA7r7/7u70yDK6dPn24+lmspFhYWmjMuwx1HjBhhztTU1JgznR0c+GXDhw83Z1wut/Hjx5szycnJ5ozkNjTWZbCoy3FcvPzyyz1yHMntOtq/f785c/bsWXMmKSnJnJHc7lfq6urMmZtvvtmc2b59uzkjSdXV1eaMdSBwQ0ODVq9ererq6qs+r+/9VXAAgGsTBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXpjfjqGnzJ07V7GxsZ1eHx0dbT5GUVGROSNJubm55kxVVZU5M3r0aHNmyJAh5kxDQ4M5I0kffPCBORMXF2fOREVFmTOuX9Mnn3xizrhMgXaZhp2SkmLOhMNhc0Zym1p+9OhRc8blHZBDoZA5Ex8fb85IF9/vzCojI8Oc2bp1qznj+kYGLtPlrd9PnV3PIyAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8KLXDiP9z//8T9OA0YkTJ5qPccMNN5gzktugywED7F3vMix1+vTp5syFCxfMGUmaNGmSOfPRRx+ZM2fOnDFnRowYYc5I0ocffmjOvPjii+bMsmXLzJna2lpzpqCgwJyR3L43Dh06ZM7ExMSYMy7X0PDhw80ZyW2wqMv3bX19vTnzne98x5yR3IYIDx061LS+qampU+t4BAQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXvTaYaR///d/r9jY2E6vb2xsNB/DZcilJDU3N5szLsMGlyxZYs4cPnzYnHnmmWfMGUm6+eabzRmXwafl5eXmjMv1IEk1NTXmzNSpU82Z4uJicyYSiZgzCxYsMGckaceOHeZMWVmZOTNq1ChzZs2aNebMH//4R3NGklpaWswZl6/p4MGD5swnn3xizkjS0aNHzZnbbrvNtL61tbVT63gEBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABe9NphpFFRUYqKiur0+uTkZPMxxo0bZ864qq6uNmd++ctfmjNJSUnmzJAhQ8wZSfrNb35jzgwYYP83z+zZs82ZpqYmc0aSxo4da85UVFQ4HcvK5dy98sorTseaMmWKOePyPVhQUGDOtLW1mTMug1Il6f777zdnXK4Hl+/B9957z5yR3AbUnj171rS+s99/PAICAHhBAQEAvDAX0O7du3X33XcrPT1dUVFR2rJlS4fPr1y5sv2/zy7dFi9e3FX7BQD0E+YCqqurU3Z2tjZs2HDFNYsXL9bp06fbby+99NI32iQAoP8xvwghNzdXubm5V10TCoWUmprqvCkAQP/XLc8B7dq1S8nJyZowYYIefvhhVVZWXnFtU1OTIpFIhxsAoP/r8gJavHixXnjhBe3YsUP//u//roKCAuXm5l7xPcLz8/MVDofbbxkZGV29JQBAL9TlPwe0fPny9l9PnTpV06ZN09ixY7Vr167Lvv583bp1Wrt2bfvvI5EIJQQA14Bufxl2VlaWkpKSVFxcfNnPh0IhDRs2rMMNAND/dXsBnTp1SpWVlUpLS+vuQwEA+hDzf8HV1tZ2eDRTWlqqQ4cOKTExUYmJiXriiSe0dOlSpaamqqSkRD/5yU90ww03aNGiRV26cQBA32YuoP379+vOO+9s//2l529WrFihZ599VocPH9bvfvc7VVVVKT09XQsXLtS//Mu/KBQKdd2uAQB9nrmA5s+fryAIrvj5//u///tGG7qkrKzMVFrHjh0zH2PGjBnmjCRNnz7dnLlw4YI54zKgcNeuXT1yHEm69dZbzZm6ujpzZvTo0eaM63Xo8gIYl39czZ8/35zZuXOnOeMyBFeS/ud//seccRmWmp2dbc64/B298MIL5owkxcXFmTPl5eXmjGXw8iXHjx83ZyQpPT3dnLEOmr1aR3wRs+AAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgRZe/JXdXKSws1KBBgzq93mW68DvvvGPOSG4TnV3e6dUlk5WVZc5cd9115owkvf/+++bM1KlTzZlz586ZM4WFheaMJN1yyy3mjMs07KqqKnPmtddeM2cyMzPNGUmaMGGCOVNfX2/O3HTTTebM/v37zRmX+wfJbRJ7S0uLOeMytfyBBx4wZyS3ad3hcNi0vrGxsVPreAQEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF702mGkS5cuVVxcXKfXuwzYcxmEKEmRSMSccRmo6fI1jRw50pzZt2+fOSNJZWVl5sy4cePMGZeBmr/5zW/MGUnauXOnOTNwoP3bqLPDGr8oOzvbnHEZcilJn376qTmTlpZmzlRWVpozLl/TbbfdZs5I0vbt282Zjz/+2Jyx3Nddkpuba85IUmJiojlTXFxsWt/c3NypdTwCAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAveu0w0n379ikmJqbT61tbW83HGD16tDkjuQ0OrKmpMWc+++wzc6a2ttac+fzzz80ZSZo+fbo54zLI9eTJk+bMm2++ac5I0qxZs8yZo0ePmjOdHdb4RR988IE54/J9IUnf/e53e+RY6enp5ozLeXAZBiy5DZq94447zBmX+5Tjx4+bM5Lb0NiUlBTT+qampk6t4xEQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHjRa4eRLl68WEOGDOn0epfhk21tbeaMJJ0/f96cOXv2rDmTmZlpzgRBYM64DOCUpPr6enNm3Lhx5kxJSYk54zp8Mj4+3pyJjo42Z+rq6syZcDhszgwdOtSckaTY2FhzprS01JzZv3+/OZObm2vOFBUVmTOS2+BOl+vh9OnT5syoUaPMGck+WFSSjhw5Ylrf2WG7PAICAHhBAQEAvDAVUH5+vmbMmKH4+HglJyfrnnvuUWFhYYc1jY2NysvL0/DhwzV06FAtXbpUFRUVXbppAEDfZyqggoIC5eXlae/evXrrrbfU0tKihQsXdvj/7EcffVSvvfaaXnnlFRUUFKisrEz33Xdfl28cANC3mV6EsG3btg6/37hxo5KTk3XgwAHNmzdP1dXV+u1vf6tNmzbpW9/6liTp+eef16RJk7R3717deuutXbdzAECf9o2eA6qurpYkJSYmSpIOHDiglpYW5eTktK+ZOHGiMjMztWfPnsv+GU1NTYpEIh1uAID+z7mA2tratGbNGs2ZM0dTpkyRJJWXlysmJkYJCQkd1qakpKi8vPyyf05+fr7C4XD7LSMjw3VLAIA+xLmA8vLydOTIEb388svfaAPr1q1TdXV1++3kyZPf6M8DAPQNTj+Iunr1ar3++uvavXu3rr/++vaPp6amqrm5WVVVVR0eBVVUVCg1NfWyf1YoFFIoFHLZBgCgDzM9AgqCQKtXr9arr76qt99+W2PGjOnw+enTp2vQoEHasWNH+8cKCwt14sQJzZ49u2t2DADoF0yPgPLy8rRp0yZt3bpV8fHx7c/rhMNhxcbGKhwO64c//KHWrl2rxMREDRs2TI888ohmz57NK+AAAB2YCujZZ5+VJM2fP7/Dx59//nmtXLlSkvQf//EfGjBggJYuXaqmpiYtWrRI//Vf/9UlmwUA9B+mAurMoMvBgwdrw4YN2rBhg/OmpIvDGi2DNV0G83344YfmjOQ2FNJloKbLANM77rjDnLnSKxS/zoULF8wZl68pKSnJnHEZ9ilJv/jFL8wZl+cw77rrLnNm+fLl5swzzzxjzkh//dEKC5eBmpWVleaMy9DT2267zZyR3O4jXAbhTpgwwZypqakxZ6SL02qsGhoaTOtbWlo6tY5ZcAAALyggAIAXFBAAwAsKCADgBQUEAPCCAgIAeEEBAQC8oIAAAF5QQAAALyggAIAXFBAAwAsKCADgBQUEAPAiKrCMnO4BkUhE4XBY+fn5Gjx4cKdzLpN4Bw50ekNYRUdHmzMuE2jnzZtnzrzxxhvmjMveXB07dsycudK76V5NTEyMOSNJU6ZMMWcOHz5szowcOdKccZn4fv78eXNGkj799FNzZtCgQeZMbGysOTNt2jRz5tChQ+aM5DZ5e9KkSeZMa2urOeMy+V6Sjhw5Ys5kZmaa1jc0NOhHP/qRqqurNWzYsCuu4xEQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwggICAHjhNo2zB4wfP15Dhgzp9HqXYaQJCQnmjCSVlZWZM1cbyHclH330kTkzd+5cc8ZlmKYklZeXmzPf+c53zJm6ujpzxmXoqeQ26HLBggXmzNGjR82Z+Ph4c+bUqVPmjCStWrXKnHn55ZfNmffee8+cqa6uNmeam5vNGUmmgciXZGVlmTMuM6H/93//15yRpMmTJ5sztbW1pvUNDQ2dWscjIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwotcOI/3jH/+oUCjU6fWWwaWXVFVVmTOS2zDECRMmmDPp6enmzPnz580ZlwGmkvTWW2+ZM3/5y1/Mmb/5m78xZyorK80ZSfr444/NmXfeececmTRpkjlz8uRJc8ZlCK4kvfDCC+ZMa2urObN48WJzZuzYseaMy9BTSRo5cqQ589lnn5kzLt+3LnuT3O6Lzp0753Ssr8MjIACAFxQQAMALCggA4AUFBADwggICAHhBAQEAvKCAAABeUEAAAC8oIACAFxQQAMALCggA4AUFBADwotcOI83MzFRsbGyn13/44YfmY9x0003mjCTNmzfPnHHZ3/XXX2/OFBUVmTOuQ1lTU1PNmdtvv92ccRnu6DK4U7p43fVE5s033zRnLN8PlwwePNickaRwOGzOpKSkmDMXLlwwZ1y4DDCVpIED7XeRiYmJ5ozLeZgyZYo5I7ndF7W0tJjWNzU1dWodj4AAAF5QQAAAL0wFlJ+frxkzZig+Pl7Jycm65557VFhY2GHN/PnzFRUV1eH20EMPdemmAQB9n6mACgoKlJeXp7179+qtt95SS0uLFi5cqLq6ug7rHnzwQZ0+fbr99uSTT3bppgEAfZ/pGbZt27Z1+P3GjRuVnJysAwcOdHhiPi4uzukJagDAteMbPQd06a2pv/yqjxdffFFJSUmaMmWK1q1bp/r6+iv+GU1NTYpEIh1uAID+z/ll2G1tbVqzZo3mzJnT4eWA3//+9zVq1Cilp6fr8OHDeuyxx1RYWKg//OEPl/1z8vPz9cQTT7huAwDQRzkXUF5eno4cOaJ33nmnw8dXrVrV/uupU6cqLS1NCxYsUElJyWVfi79u3TqtXbu2/feRSEQZGRmu2wIA9BFOBbR69Wq9/vrr2r1799f+sOSsWbMkScXFxZctoFAopFAo5LINAEAfZiqgIAj0yCOP6NVXX9WuXbs0ZsyYr80cOnRIkpSWlua0QQBA/2QqoLy8PG3atElbt25VfHy8ysvLJV0c2xEbG6uSkhJt2rRJ3/72tzV8+HAdPnxYjz76qObNm6dp06Z1yxcAAOibTAX07LPPSrr4w6Zf9Pzzz2vlypWKiYnR9u3b9fTTT6uurk4ZGRlaunSpfvrTn3bZhgEA/YP5v+CuJiMjQwUFBd9oQwCAa0OvnYZ99uxZ0yTfhIQE8zHOnTtnzkhu02THjx9vzhw+fNicKSsrM2e+973vmTOSdOzYMXNm586d5kxDQ4M5c9ttt5kzkttk65KSEnNm0qRJ5szkyZPNGVelpaXmjMv3k8vE9y+P/+qMSy+Gsjp69KhTzmrOnDnmjOv9l8v3k3W6fHNzc6fWMYwUAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALzotcNIS0tLFRMT0+n1U6dONR/j888/N2ckafjw4eZMfHy8OVNXV2fOTJw40ZzZvHmzOSNJycnJ5kxWVpY5c+lNDS1c39b90ntcWdTW1poz48aNM2f+/Oc/mzMjRowwZyRp4ED7XYPLgNVTp06ZM42NjebMu+++a85IbgM/c3JyzJmioiJzxuV6kKTly5ebM21tbab1TU1NnVrHIyAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOBFr5sFFwSBJKmlpcWUc5kP1dl5RV9mnYskSQ0NDT2SuXDhgjnjeh5cznlra6s509zcbM64nDup564jl/25HMfl65HcZsG5fE09db5drjvJ7dqrr683Z1zOg/U+8hKX/VnP+aX1l+7PryQq+LoVPezUqVPOgyQBAL3HyZMndf3111/x872ugNra2lRWVqb4+HhFRUV1+FwkElFGRoZOnjypYcOGedqhf5yHizgPF3EeLuI8XNQbzkMQBKqpqVF6eroGDLjyMz297r/gBgwYcNXGlKRhw4Zd0xfYJZyHizgPF3EeLuI8XOT7PITD4a9dw4sQAABeUEAAAC/6VAGFQiGtX79eoVDI91a84jxcxHm4iPNwEefhor50HnrdixAAANeGPvUICADQf1BAAAAvKCAAgBcUEADAiz5TQBs2bNDo0aM1ePBgzZo1S3/+8599b6nH/fznP1dUVFSH28SJE31vq9vt3r1bd999t9LT0xUVFaUtW7Z0+HwQBHr88ceVlpam2NhY5eTkqKioyM9mu9HXnYeVK1d+5fpYvHixn812k/z8fM2YMUPx8fFKTk7WPffco8LCwg5rGhsblZeXp+HDh2vo0KFaunSpKioqPO24e3TmPMyfP/8r18NDDz3kaceX1ycKaPPmzVq7dq3Wr1+v999/X9nZ2Vq0aJHOnDnje2s97sYbb9Tp06fbb++8847vLXW7uro6ZWdna8OGDZf9/JNPPqlf//rXeu6557Rv3z4NGTJEixYtch7E2Vt93XmQpMWLF3e4Pl566aUe3GH3KygoUF5envbu3au33npLLS0tWrhwoerq6trXPProo3rttdf0yiuvqKCgQGVlZbrvvvs87rrrdeY8SNKDDz7Y4Xp48sknPe34CoI+YObMmUFeXl7771tbW4P09PQgPz/f46563vr164Ps7Gzf2/BKUvDqq6+2/76trS1ITU0NnnrqqfaPVVVVBaFQKHjppZc87LBnfPk8BEEQrFixIliyZImX/fhy5syZQFJQUFAQBMHFv/tBgwYFr7zySvuaY8eOBZKCPXv2+Npmt/vyeQiCILjjjjuCf/iHf/C3qU7o9Y+AmpubdeDAAeXk5LR/bMCAAcrJydGePXs87syPoqIipaenKysrSz/4wQ904sQJ31vyqrS0VOXl5R2uj3A4rFmzZl2T18euXbuUnJysCRMm6OGHH1ZlZaXvLXWr6upqSVJiYqIk6cCBA2ppaelwPUycOFGZmZn9+nr48nm45MUXX1RSUpKmTJmidevWOb0VQ3fqdcNIv+zcuXNqbW1VSkpKh4+npKTo+PHjnnblx6xZs7Rx40ZNmDBBp0+f1hNPPKHbb79dR44cUXx8vO/teVFeXi5Jl70+Ln3uWrF48WLdd999GjNmjEpKSvTP//zPys3N1Z49exQdHe17e12ura1Na9as0Zw5czRlyhRJF6+HmJgYJSQkdFjbn6+Hy50HSfr+97+vUaNGKT09XYcPH9Zjjz2mwsJC/eEPf/C42456fQHhr3Jzc9t/PW3aNM2aNUujRo3S73//e/3whz/0uDP0BsuXL2//9dSpUzVt2jSNHTtWu3bt0oIFCzzurHvk5eXpyJEj18TzoFdzpfOwatWq9l9PnTpVaWlpWrBggUpKSjR27Nie3uZl9fr/gktKSlJ0dPRXXsVSUVGh1NRUT7vqHRISEjR+/HgVFxf73oo3l64Bro+vysrKUlJSUr+8PlavXq3XX39dO3fu7PD2LampqWpublZVVVWH9f31erjSebicWbNmSVKvuh56fQHFxMRo+vTp2rFjR/vH2tratGPHDs2ePdvjzvyrra1VSUmJ0tLSfG/FmzFjxig1NbXD9RGJRLRv375r/vo4deqUKisr+9X1EQSBVq9erVdffVVvv/22xowZ0+Hz06dP16BBgzpcD4WFhTpx4kS/uh6+7jxczqFDhySpd10Pvl8F0Rkvv/xyEAqFgo0bNwZHjx4NVq1aFSQkJATl5eW+t9aj/vEf/zHYtWtXUFpaGrz77rtBTk5OkJSUFJw5c8b31rpVTU1NcPDgweDgwYOBpOBXv/pVcPDgweDTTz8NgiAI/u3f/i1ISEgItm7dGhw+fDhYsmRJMGbMmKChocHzzrvW1c5DTU1N8OMf/zjYs2dPUFpaGmzfvj24+eabg3HjxgWNjY2+t95lHn744SAcDge7du0KTp8+3X6rr69vX/PQQw8FmZmZwdtvvx3s378/mD17djB79myPu+56X3ceiouLg1/84hfB/v37g9LS0mDr1q1BVlZWMG/ePM8776hPFFAQBMEzzzwTZGZmBjExMcHMmTODvXv3+t5Sj1u2bFmQlpYWxMTEBCNHjgyWLVsWFBcX+95Wt9u5c2cg6Su3FStWBEFw8aXYP/vZz4KUlJQgFAoFCxYsCAoLC/1uuhtc7TzU19cHCxcuDEaMGBEMGjQoGDVqVPDggw/2u3+kXe7rlxQ8//zz7WsaGhqCH/3oR8F1110XxMXFBffee29w+vRpf5vuBl93Hk6cOBHMmzcvSExMDEKhUHDDDTcE//RP/xRUV1f73fiX8HYMAAAvev1zQACA/okCAgB4QQEBALyggAAAXlBAAAAvKCAAgBcUEADACwoIAOAFBQQA8IICAgB4QQEBALyggAAAXvw/5gAibpVRBtYAAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 3\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "Y41nWrxWMaVD",
        "outputId": "88a4a1d0-3d10-4977-b8fc-6196873438ff"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c74f9a20>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 4\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "9ZDqP3wBMjJ-",
        "outputId": "357d97db-1605-4d32-c33c-ac67a474dbba"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c7562da0>"
            ]
          },
          "metadata": {},
          "execution_count": 23
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 5\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "drJztyoGMpiC",
        "outputId": "b2161eb2-435f-4692-d208-d509d79c0f3a"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c73e3340>"
            ]
          },
          "metadata": {},
          "execution_count": 24
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 6\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "z_qaTDsgMwxi",
        "outputId": "3a8142cf-84d8-4a04-deca-bfcc57a6f990"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c74f8760>"
            ]
          },
          "metadata": {},
          "execution_count": 25
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 7\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "yNOqlNOEM2lS",
        "outputId": "63038dba-7450-4d64-b133-135a42eac996"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c74664a0>"
            ]
          },
          "metadata": {},
          "execution_count": 26
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 8\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "rGk26x9UM9si",
        "outputId": "f50f72de-5a4f-4537-ee4f-e58e68267955"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c732f100>"
            ]
          },
          "metadata": {},
          "execution_count": 27
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# label to test\n",
        "label = 9\n",
        "\n",
        "# create the output signals for this label\n",
        "targets = numpy.zeros(output_nodes) + 0.01\n",
        "targets[label] = 0.99\n",
        "\n",
        "# get image data\n",
        "image_data = n.backquery(targets)\n",
        "\n",
        "# plot image data\n",
        "matplotlib.pyplot.imshow(image_data.reshape(28,28), cmap='Greys', interpolation='None')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "8pyH_Ia8ND5M",
        "outputId": "8e7e3ff3-ec7d-46ea-ca26-81acfc0d0f2a"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7dd5c73b84c0>"
            ]
          },
          "metadata": {},
          "execution_count": 28
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}