Neural_Network_Back_Propagation.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Neural Network - Back Propagation.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyOsZ+2wHvFWCDo9JoGEBg8Y",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/tecnocrata/Bromine/blob/master/Neural_Network_Back_Propagation.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "84Ahuz6rM7JU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import scipy as sc\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "#Esto es para crar un dataset con una distribucion circular\n",
        "from sklearn.datasets import make_circles"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hw2TIMfdOXBj",
        "colab_type": "code",
        "outputId": "c5b253bd-dc21-4ad0-f06c-b9100a671e40",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        }
      },
      "source": [
        "#n es el numero de registros en nuestros datos\n",
        "n=500\n",
        "#p la cantidad de propiedades/caracteristicas de cada uno de los registros \n",
        "p=2\n",
        "\n",
        "#Crear el dataset\n",
        "# Donde X es el array que contiene las dos propiedades\n",
        "# Donde Y es el array que contiene el resultado de aplicar la funcion, \n",
        "# Y es 0 si pertenece a un circulo o 1 si pertenece al otro\n",
        "\n",
        "#Factor es la distancia entre ambos circulos\n",
        "#Noise es para modificar la uniformidad, modificando este valor los valores no son uniformes, pruebalo\n",
        "X, Y = make_circles (n_samples=n, factor=0.5, noise=0.04)\n",
        "\n",
        "#Corrigiendo Y, porque la funcion devuelve una serie de numeros y no un array, que es lo que queremos\n",
        "Y = Y[:, np.newaxis]\n",
        "\n",
        "print (len(X))\n",
        "print (len(Y))\n",
        "# Plotear solo los puntos que pertenecen al circulo 0\n",
        "plt.scatter(X[Y[:,0]==0,0],X[Y[:,0]==0,1], c=\"skyblue\")\n",
        "\n",
        "# Plotear solo los puntos que pertenecen al circulo 1\n",
        "plt.scatter(X[Y[:,0]==1,0],X[Y[:,0]==1,1], c=\"salmon\")\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "500\n",
            "500\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": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ErgtPdOYcPZR",
        "colab_type": "text"
      },
      "source": [
        "El objetivo es que la red neuronal, separe en dos (clases) toda esta nube de puntos"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X7y50oZ6cMv5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#Una forma de pensar es ver cada capa es un modulo\n",
        "#Dentro de una misma capa se realizaran las mismas operaciones\n",
        "#El MODULO PRINCIPAL en una RN es es \"la capa\"\n",
        "\n",
        "class NeuralLayer ():\n",
        "  def __init__(self, connections, neurons, activationFunction, deriv_activationFunction):\n",
        "    self.activationFunction = activationFunction\n",
        "    self.deriv_activationFunction = deriv_activationFunction\n",
        "    # bias = sesgo, es un array, del tamano de la cantidad de neuronas\n",
        "    # de valores entre -1 y 1 (normalizado), por eso multiplicamos el valor * 2 y le quitamos 1\n",
        "    self.bias = np.random.rand(1, neurons) * 2 -1\n",
        "\n",
        "    #Este es el la matriz de pesos, de  connections x neurons\n",
        "    self.w =  np.random.rand(connections, neurons) * 2 -1\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Oz-uKnQxfm1m",
        "colab_type": "code",
        "outputId": "4ddcdba2-232d-40b1-fddb-3f931e03f81e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "# Estas son las diferentes funciones de activacion que usaremos\n",
        "sigm =  lambda x: 1/ (1 + np.e ** (-x))\n",
        "# generamos un vector de 100 elementos de valores entre -5 y 5\n",
        "_x = np.linspace (-5,5, 100)\n",
        "#generamos la grafica\n",
        "plt.plot(_x,sigm(_x), 'r')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f0ec3dd0b70>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UXDCG1fHijaB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X6Y0SK2Igtfw",
        "colab_type": "code",
        "outputId": "5eaf2cc8-b6ee-4446-9581-9ab13674dd7e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "# Necesitaremos la derivada de la funcion de activacion, para el algoritmo de backpropagation\n",
        "# en este caso dde la funcion sigmoide\n",
        "deriv_sigm = lambda x: x* (1-x)\n",
        "plt.plot(_x,deriv_sigm(_x), 'r')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f0ec39cc1d0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JHm6qAERhxB6",
        "colab_type": "code",
        "outputId": "735ba86e-2244-4612-f089-27c623a37190",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "# otra funcion de activacion\n",
        "relu = lambda x: np.maximum (0, x)\n",
        "plt.plot(_x,relu(_x))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f0ec3de35c0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_IZh46Bhii30",
        "colab_type": "code",
        "outputId": "f8b5b75b-4674-4227-e8d4-9f09e65212a6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 221
        }
      },
      "source": [
        "# layer 0\n",
        "l0 = NeuralLayer (p, 4, sigm, deriv_sigm)\n",
        "# layer 1\n",
        "l1 = NeuralLayer (4, 8, sigm, deriv_sigm)\n",
        "\n",
        "# Definimos una variable que nos permitira modificar la topologia de la RNN de manera dinamica\n",
        "# Contiene la cantidad de neuronas que tendra cada capa\n",
        "# La capa final tiene una sola neurona, porque el resultado que queremos obtener es binario\n",
        "# Es decir 0 = pertenede al primer conjunto de datos o 1 = pertenece al segundo conjunto\n",
        "topology = [p, 4, 8, 16, 8, 4, 1]\n",
        "\n",
        "# Tambien podriamos tener otro vector que diga que funciones de activacion usaremos en cada capa\n",
        "# functions = []\n",
        "# Que learning rate usaremos por cada capa\n",
        "# learning_rate = []\n",
        "# Estos se llaman los hiper-parametros\n",
        "# Por simplicidad mantendremos topology simplemente en base al numero de neuronas por capa\n",
        "# Tambien por simplicidad definimos que todas las capas tendran la misma funcion de activacion, esto puede cambiarse\n",
        "def createNeuralNetwork (topology, activation_function, deriv_activation_function):\n",
        "  nn = []\n",
        "  #esta forma es sin manipular el array\n",
        "  # for i, n_neurons in enumerate(topology):\n",
        "  #   if (i+1 < len(topology)):\n",
        "  #     nn.append (NeuralLayer(connections=n_neurons, neurons=topology[i+1]), activation_function)\n",
        "  for i, n_neurons in enumerate(topology [:-1]): #descartando el ultimo elemento para evitar el if interno\n",
        "    print (n_neurons, topology[i])\n",
        "    nn.append (NeuralLayer(connections=topology[i], neurons=topology[i+1], activationFunction= activation_function, deriv_activationFunction=deriv_activation_function))\n",
        "  return nn\n",
        "\n",
        "createNeuralNetwork (topology, sigm, deriv_sigm)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2 2\n",
            "4 4\n",
            "8 8\n",
            "16 16\n",
            "8 8\n",
            "4 4\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<__main__.NeuralLayer at 0x7f0ec3a348d0>,\n",
              " <__main__.NeuralLayer at 0x7f0ec3a34fd0>,\n",
              " <__main__.NeuralLayer at 0x7f0ec3a34dd8>,\n",
              " <__main__.NeuralLayer at 0x7f0ec3a344a8>,\n",
              " <__main__.NeuralLayer at 0x7f0ec3a347b8>,\n",
              " <__main__.NeuralLayer at 0x7f0ec3a34a58>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hMci5rsy9CrA",
        "colab_type": "code",
        "outputId": "19daaf2a-dc55-4b63-910f-1ebdf3a1ab32",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        }
      },
      "source": [
        "#Creando la logica de entrenamiento\n",
        "# Este tiene proceso tiene tres pasos/elementos\n",
        "# 1) Un procesamiento hacia adelante\n",
        "# Mostrarle a la RNN un tipo de entrada y la salida que quiero obtener, \n",
        "# la RNN procesara estos dos elementos por todas sus capas y todas sus neuronas, \n",
        "# ejecutando sumas ponderadas y funciones de activacion y retornara un valor\n",
        "# Al principio cuando la red no esta entrenada, el valor resultante sera aleatorio\n",
        "# Sin embargo con el paso del tiempo/entrenamiento ese valor se ira pareciendo al del vector Y\n",
        "# 2) Una propagacion hacia atras, Una vez que se tiene el resultado anterior, se compara este resultado con el valor esperado\n",
        "# Esta comparacion se la hara usando la funcion de coste, que nos dice como se diferencian ambos resultados (el error)\n",
        "# El resultado de esa comparacion sera un error, que usaremos para hacer una propagacion hacia atras\n",
        "# 3) Aplicar el gradient descent, mediante aplicar el algoritmo del descenso del gradiente\n",
        "# Consiste en calcular la derivada parciales de la funcion de coste\n",
        "# este calculo permitira optimizar la funcion de coste y por ende entrenar la red\n",
        "\n",
        "topology = [p, 4, 8, 8, 16, 1]\n",
        "nn= createNeuralNetwork (topology, sigm, deriv_sigm)\n",
        "\n",
        "# Definiendo la funcion de coste\n",
        "# Donde Ypredicted = es la Y que la RNN predecira o sacara como resultado\n",
        "# Donde Yreal = es la Y real esperada\n",
        "# Esta formula es la del error cuadratico medio\n",
        "# Carlos la llama l2_cost\n",
        "cost= lambda Ypredicted, Yreal: np.mean((Ypredicted-Yreal)**2)\n",
        "deriv_cost = lambda Ypredicted, Yreal: Ypredicted - Yreal\n",
        "\n",
        "# El parametro train=True, le indica al metodo que sera usado para entrenamiento\n",
        "# El parametro train=False, le indica al metodo que sera usado para predecir o dar un resultado unicamente (NO ME AGRADA!)\n",
        "def train(nn, X, Y, cost_function, deriv_cost_function, learning_rate=0.5, train= True):\n",
        "  # (1) Fordward pass\n",
        "  # Este bloque es lo que hariamos sin iteracion\n",
        "  # z = X @ nn[0].w + nn[0].bias\n",
        "  # activation = nn[0].activationFunction (z)\n",
        "\n",
        "  # Esto es requerido porque al principio el previous output es los valores de entrada\n",
        "  previous_output = X\n",
        "  \n",
        "  # Carlos quiere crear un vector out donde\n",
        "  # se guarde los siguientes pares out = [(z0,a0), (z1,a1), ....(zn, an)]\n",
        "  # Este par de resultados se guardara para el algoritmo de back-propagation\n",
        "  out = [(None, X)]\n",
        "\n",
        "  for i, layer in enumerate(nn):\n",
        "    z = previous_output @ nn[i].w + nn[i].bias\n",
        "    # Este es el resultado de la capa i, llamado tambien la activacion\n",
        "    a = nn[i].activationFunction (z)\n",
        "    previous_output = a\n",
        "    out.append ((z,a))\n",
        "  #print (previous_output)\n",
        "\n",
        "  if train:\n",
        "    # (2) Backward pass\n",
        "    deltas = []\n",
        "    # Recorremos la RNN de atras hacia adelante\n",
        "    # Las derivadas se calculan de diferente manera para la ultima capa que para las capas intermedias\n",
        "    # print (\"rango--->\",reversed(range(0, len(nn))))\n",
        "    for i in reversed(range(0, len(nn))):\n",
        "        # i+1 porque quiero acceder a la capa de adelante\n",
        "        z = out [i+1][0]\n",
        "        a = out [i+1][1]\n",
        "        #print (a.shape)\n",
        "        \n",
        "        if (i == len(nn)-1):\n",
        "          # Calcular delta de la ultima capa\n",
        "          # El delta se calcula en funcion al coste\n",
        "          deltas.insert(0, deriv_cost_function(a,Y) * nn[i].deriv_activationFunction(a))\n",
        "        else:\n",
        "          # Calcular el delta de la capa intermedia, en funcion a la capa previa\n",
        "          #print (deltas[0].shape, nn[i+1].w.shape)\n",
        "          # deltas.insert(0, deltas[0] @ nn[i+1].w.T * nn[i].deriv_activationFunction(a))\n",
        "          deltas.insert(0, deltas[0] @ _W.T * nn[i].deriv_activationFunction(a))\n",
        "        _W = nn[i].w\n",
        "        #(3) Gradiente descendiente\n",
        "        nn[i].bias = nn[i].bias - np.mean(deltas[0], axis=0, keepdims=True) * learning_rate\n",
        "        nn[i].w = nn[i].w - out[i][1].T @ deltas[0] * learning_rate\n",
        "    #(3) Gradiente descendiente YO sugiero hacer otro bucle para evitar la confusion\n",
        "    #for i in range(0, len(nn)):\n",
        "\n",
        "  #return previous_output\n",
        "  return out[-1][1]\n",
        "\n",
        "\n",
        "train (nn, X, Y, cost, deriv_cost, 0.05)\n",
        "print (\"\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2 2\n",
            "4 4\n",
            "8 8\n",
            "8 8\n",
            "16 16\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wLgF1T-nYltF",
        "colab_type": "code",
        "outputId": "5331db30-92ed-4c52-ed2a-7e70100dbe32",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 517
        }
      },
      "source": [
        "import time\n",
        "from IPython.display import clear_output\n",
        "\n",
        "topology = [p, 4, 8, 16, 1]\n",
        "nn= createNeuralNetwork (topology, sigm, deriv_sigm)\n",
        "\n",
        "# Se guardara los costes, para poder graficar como evoluciona\n",
        "loss = []\n",
        "for i in range(2500):\n",
        "  Ypredicted = train (nn, X, Y, cost, deriv_cost, learning_rate= 0.01)\n",
        "  if i%25==0:\n",
        "    loss.append(cost(Ypredicted, Y))\n",
        "    \n",
        "    #cantidad de puntos, resolucion\n",
        "    res = 50\n",
        "    _x0=np.linspace(-1.5, 1.5, num=res)\n",
        "    _x1=np.linspace(-1.5, 1.5, num=res)\n",
        "\n",
        "    _Y = np.zeros((res,res))\n",
        "\n",
        "    for i0, x0 in enumerate(_x0):\n",
        "      for i1, x1 in enumerate(_x1):\n",
        "        _Y[i0, i1] = train (nn, np.array([[x0, x1]]), Y, cost, deriv_cost, 0.05, train=False)[0][0]\n",
        "\n",
        "    plt.pcolormesh(_x0, _x1, _Y, cmap=\"coolwarm\")\n",
        "    plt.axis(\"equal\")\n",
        "\n",
        "    # Plotear solo los puntos que pertenecen al circulo 0\n",
        "    plt.scatter(X[Y[:,0]==0,0],X[Y[:,0]==0,1], c=\"skyblue\")\n",
        "    # Plotear solo los puntos que pertenecen al circulo 1\n",
        "    plt.scatter(X[Y[:,0]==1,0],X[Y[:,0]==1,1], c=\"salmon\")\n",
        "\n",
        "    #Limpiar el contenido para mostrar de forma animada los resultados\n",
        "    clear_output(wait=True)\n",
        "    plt.show()\n",
        "    plt.plot (range(len(loss)), loss)\n",
        "    plt.show()\n",
        "    time.sleep(0.5)\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1jvaxC0j3CVD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l5VYVvmf2pDt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JHa_8kRg2Wnw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Vefb1zw3xyjx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DVllBsCHxrAU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d5fUhA1Xxfhs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CqttGeapxSse",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "no2ciJVjr-6L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ib06xv-2rUdE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QneaUSO8rPn8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C34sKyZHrHBe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l-LJCg12ziSF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cn9fci0xzKrX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9UfM25xuyzzH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "30FTqsMPyErr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xmbdXhwqx-j6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zwfpSQMkxZ8v",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U9XTykTuwaHE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CO3FonACu6dE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gxq8StZ5umcM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7E2ZWSnjuYEv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1lU1a4Y7uDwP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dyDyV3XwtLcc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wFvMsCXLs6xZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xXYMM8DmsXwV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZemKXE_NsUcr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v4a0tX8GYUIG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rbzoFII2YDJu",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H_tEmFguXiPv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EBLjny1HXCMX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e-RHhgBZW-kh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wrwWmgZzR7u1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e4qIBMSCPbso",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-8HAnHJgMhCK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "My49uYLzMFQc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sHrzvBkeL_H_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "e-KXSB_WHxjp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-LZJEzw7Hv4z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M64BtTkHHNSO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "21tdPkwZHH3Q",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZSH2Iz4FHEHF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zsqz4nwPG5kI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BQ1GcWiLGs-5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yXymND8gGoj7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6a4habiBBTdx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z-bog3fL80VB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qcPhIviEr59s",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "38g949S1rvV1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gIZJx-CQpqWq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gpHdNxArpfje",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_iG_jZouigsE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M1B-Orwbid9N",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6AO225YSiast",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YsyqcIBCiSy0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z4oUOTWagrbe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OM2B8Ih8TOad",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yRjQKG7KTKYG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "k5Lb-DPfTFNP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U0bVl3NLSuMG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yRO95uSSSsep",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CQf8SMCBSrKo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LvoVCuUpQxdx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YAvkn2rbQuMB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dHCqCX7mQh-B",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j08COj1GQSHa",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ovHrsCFzQPtp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SLqBpJnEQLWi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IlJiS9t7QFbz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xvq7F_gkP_oD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rmNmKjJ-P3by",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GaDOHkINP0o1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "267m8XfwNTkr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}