CHAP_09.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "CHAP-09.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "authorship_tag": "ABX9TyNhcj80/26fjQblA1lmggcs",
      "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/metamath1/noviceml/blob/master/CHAP_09.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WtqACm8bGFyr"
      },
      "source": [
        "# 9장 그림 및 예제 코드"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-9x7EUKIGBy8"
      },
      "source": [
        "# 기본적인 import들\n",
        "# 이후 그래프 그리는 코드에는 중복으로 적지 않음.\n",
        "# 다른 곳으로 그래프 그리는 코드를 복사-붙이기 할 때는 \n",
        "# 이 import 코드와 함께 복사-붙이기 해야함\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib as mpl\n",
        "from mpl_toolkits import mplot3d\n",
        "import matplotlib.font_manager as mfm\n",
        "import sympy\n",
        "\n",
        "# numpy 출력 형식 지정\n",
        "np.set_printoptions(precision=4, linewidth=150)\n",
        "\n",
        "# matplotlib 스타일 지정\n",
        "mpl.style.use('bmh')\n",
        "mpl.style.use('seaborn-whitegrid')\n",
        "style = plt.style.library['bmh']\n",
        "# 스타일 컬러를 쉽게 쓸 수 있도록 리스트 저장\n",
        "style_colors = [ c['color'] for c in style['axes.prop_cycle'] ]\n",
        "\n",
        "# 그림을 로컬 폴더에 저장하고 싶으면 True로 수정 \n",
        "file_print = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KjKWC7TxGKCs"
      },
      "source": [
        "def plot_boundary(X, model, ax):\n",
        "    # meshgrid를 짜서 Model에 포워드 시키기\n",
        "    # 뉴럴 네트웍 모델 자체가 함수이므로 (x,y)->z를 하는 것이다.\n",
        "    x = np.linspace(min(X[:,0]), max(X[:,0]))\n",
        "    y = np.linspace(min(X[:,1]), max(X[:,1]))\n",
        "    \n",
        "    xx, yy = np.meshgrid(x, y)\n",
        "    grid_tensor = torch.Tensor(np.c_[xx.flatten(), yy.flatten()])\n",
        "    zz = model.forward(grid_tensor).detach().numpy().reshape(xx.shape)\n",
        "    ax.contour(xx, yy, zz, cmap='gray', levels=[0.5])\n",
        "\n",
        "def logistic(x):\n",
        "    return 1 / (1+np.exp(-x))\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o-Jdsj4bGQrQ"
      },
      "source": [
        "## 샘플데이터 만들기"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nHiAcsNJGS4n"
      },
      "source": [
        "### 결정 경계"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o1gl2aENGPgr",
        "outputId": "b21ada09-e8bc-4306-c7b5-0404a6bb0e8f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 429
        }
      },
      "source": [
        "dcs_bnd_1 = lambda x: -(3/5)*x + 3\n",
        "dcs_bnd_1_imp = lambda x, y: (3)*x + (5)*y + (-15)\n",
        "\n",
        "dcs_bnd_2 = lambda x: -2*x + 6\n",
        "dcs_bnd_2_imp = lambda x, y: (-6)*x + (-3)*y + (18)\n",
        "\n",
        "x = np.linspace(0, 5, 50)\n",
        "\n",
        "fig = plt.figure(figsize=(7, 7))\n",
        "ax = fig.add_subplot(1, 1, 1)\n",
        "\n",
        "ax.plot(x, dcs_bnd_1(x), color='k')\n",
        "ax.plot(x, dcs_bnd_2(x), '--', color='k')\n",
        "\n",
        "ax.set_xlim(0, 5)\n",
        "ax.set_ylim(0, 5)\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TSfGct4oGWfj"
      },
      "source": [
        "### 무작위점 500개 생성"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vtszym3YGUNb"
      },
      "source": [
        "np.random.seed(3)\n",
        "\n",
        "# 0~5에서 무작위 점 500개 추출\n",
        "N = 500\n",
        "samples = (np.random.rand(N*2)*5).reshape(N,2)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m9m6BzVNGXnb",
        "outputId": "485f755e-9b1d-4c1c-cc49-e76fcdbee129",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "samples.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(500, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iOZKBdIQGZ11"
      },
      "source": [
        "### 생성된 점 500개 그리기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YbldnKTcGYoD",
        "outputId": "538ecee8-1b4a-4c5d-ec7b-8c169a08c1d2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 429
        }
      },
      "source": [
        "fig = plt.figure(figsize=(7, 7))\n",
        "ax = fig.add_subplot(1, 1, 1)\n",
        "\n",
        "ax.plot(samples[:,0], samples[:,1], 'o', color='k')\n",
        "ax.set_ylim(0, 5)\n",
        "ax.set_xlim(0, 5)\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TAabodEBGdNF"
      },
      "source": [
        "## 샘플 데이터에 대한 레이블(타겟) 만들기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SgTh0qekGbSL"
      },
      "source": [
        "y1_bin = dcs_bnd_1_imp(samples[:,0], samples[:,1]) > 0\n",
        "y2_bin = dcs_bnd_2_imp(samples[:,0], samples[:,1]) < 0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lE7-EslNGegj",
        "outputId": "d30527d0-bcd8-4989-c71b-fb20efe91708",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "positive_where = np.where((y1_bin | y2_bin))[0]\n",
        "target = np.zeros(N)\n",
        "target[positive_where] = 1\n",
        "\n",
        "print(target.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(500,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d2FSmcNTGgx-"
      },
      "source": [
        "- 이제 다음처럼 데이터가 준비되었음\n",
        "\n",
        "- `samples`: 점 500개게 (500, 2) 형태로 저장\n",
        "\n",
        "- `target`: 각점에 대한 레이블 0, 1이 (500,) 형태로 저장"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MO8aiIlvGmmW"
      },
      "source": [
        "## 그림 9-1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nu1JbkPjGfXz",
        "outputId": "57eb0e7e-063c-46af-c48c-bd0dbfb50b4d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        }
      },
      "source": [
        "fig = plt.figure(figsize=(7, 7))\n",
        "ax = fig.add_subplot(1, 1, 1)\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=18)\n",
        "ax.yaxis.set_tick_params(labelsize=18)\n",
        "ax.set_xlabel('$x$', fontsize=25)\n",
        "ax.set_ylabel('$y$', fontsize=25)\n",
        "\n",
        "ax.plot(samples[np.where(target==1)[0],0], samples[np.where(target==1)[0],1], 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "ax.plot(samples[np.where(target==0)[0],0], samples[np.where(target==0)[0],1], '^', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "x1 = np.linspace(0, 15/7, 50)\n",
        "x2 = np.linspace(15/7, 3, 50)\n",
        "\n",
        "ax.plot(x1, dcs_bnd_1(x1), '--', color='k')\n",
        "ax.plot(x2, dcs_bnd_2(x2), '--', color='k')\n",
        "\n",
        "ax.set_ylim(0,5)\n",
        "ax.set_xlim(0,5)\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap9/fig9-1.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap9/fig9-1.pdf\", format='pdf', bbox_inches='tight')\n",
        "    \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "74-1jev5GtyK"
      },
      "source": [
        "## 그림 9-3: $ 3x + 5y - 15 = 0 $ 로 분류"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-DUjaluLGn27",
        "outputId": "f0e625e3-0017-47ee-a585-c22475299aed",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        }
      },
      "source": [
        "fig = plt.figure(figsize=(7, 7))\n",
        "ax = fig.add_subplot(1, 1, 1)\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=18)\n",
        "ax.yaxis.set_tick_params(labelsize=18)\n",
        "ax.set_xlabel('$x$', fontsize=25)\n",
        "ax.set_ylabel('$y$', fontsize=25)\n",
        "\n",
        "ax.plot(x, dcs_bnd_1(x), color='k')\n",
        "\n",
        "ax.plot(samples[y1_bin, 0], samples[y1_bin,1], 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "not_y1_bin = np.invert(y1_bin)\n",
        "not_y1_and_target = np.logical_and(not_y1_bin,target)\n",
        "\n",
        "ax.plot(samples[not_y1_and_target, 0], samples[not_y1_and_target, 1], 'o',\n",
        "        markerfacecolor='white', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.plot(samples[np.where(target==0)[0], 0], samples[np.where(target==0)[0], 1], '^', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.set_ylim(0,5)\n",
        "ax.set_xlim(0,5)\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap9/fig9-3.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap9/fig9-3.pdf\", format='pdf', bbox_inches='tight')\n",
        "    \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tNNRVn76GxiG"
      },
      "source": [
        "## 그림 9-5: $ -6x - 3y + 18 = 0 $로 분류"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9tLHuXJgGvb7",
        "outputId": "5e258a43-739d-4111-928f-2268439e28f0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        }
      },
      "source": [
        "fig = plt.figure(figsize=(7,7))\n",
        "ax = fig.add_subplot(1,1,1)\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=18)\n",
        "ax.yaxis.set_tick_params(labelsize=18)\n",
        "ax.set_xlabel('$x$', fontsize=25)\n",
        "ax.set_ylabel('$y$', fontsize=25)\n",
        "\n",
        "ax.plot(x, dcs_bnd_2(x), color='k')\n",
        "\n",
        "ax.plot(samples[y2_bin, 0], samples[y2_bin, 1], 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "not_y2_bin = np.invert(y2_bin)\n",
        "not_y2_and_target = np.logical_and(not_y2_bin,target)\n",
        "\n",
        "ax.plot(samples[not_y2_and_target, 0], samples[not_y2_and_target, 1], 'o',\n",
        "        markerfacecolor='white', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.plot(samples[np.where(target==0)[0], 0], samples[np.where(target==0)[0], 1], '^', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.set_ylim(0,5)\n",
        "ax.set_xlim(0,5)\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap9/fig9-5.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap9/fig9-5.pdf\", format='pdf', bbox_inches='tight')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BbEMiR8uG0ij"
      },
      "source": [
        "## 그림 9-8: 위 두 경우를 합성하여 분류하는 경우"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yns4xK30Gyu8",
        "outputId": "83a856bc-2bc2-4eb1-8a42-ab6fab0d52c8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        }
      },
      "source": [
        "xx, yy = np.meshgrid(x, x)\n",
        "\n",
        "a, b, c = 10, -9, 4\n",
        "o = lambda x, y: logistic( a*logistic(dcs_bnd_1_imp(x, y)) + b*logistic(dcs_bnd_2_imp(x, y)) + c )\n",
        "\n",
        "fig = plt.figure(figsize=(7,7))\n",
        "ax = fig.add_subplot()\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=18)\n",
        "ax.yaxis.set_tick_params(labelsize=18)\n",
        "ax.set_xlabel('$x$', fontsize=25)\n",
        "ax.set_ylabel('$y$', fontsize=25)\n",
        "\n",
        "pred = o(samples[:,0], samples[:,1])\n",
        "pred_pos = pred >= 0.5\n",
        "pred_neg = pred < 0.5\n",
        "\n",
        "ax.contour(xx, yy, o(xx,yy), cmap='gray', levels=[0.5])\n",
        "\n",
        "# for positive samples\n",
        "TP = np.logical_and(target==1, pred_pos)\n",
        "FN = np.logical_and(target==1, pred_neg)\n",
        "\n",
        "# for negative samples\n",
        "TN = np.logical_and(target==0, pred_neg)\n",
        "FP = np.logical_and(target==0, pred_pos)\n",
        "\n",
        "ax.plot(samples[TP,0], samples[TP,1], 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "ax.plot(samples[TN,0], samples[TN,1], '^', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.plot(samples[FN,0], samples[FN,1], 'o', \n",
        "        markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15)\n",
        "ax.plot(samples[FP,0], samples[FP,1], '^', \n",
        "         markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15)\n",
        "\n",
        "ax.set_ylim(0,5)\n",
        "ax.set_xlim(0,5)\n",
        "\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap9/fig9-8.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap9/fig9-8.pdf\", format='pdf', bbox_inches='tight')\n",
        "    \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "czkFSla8G3qe"
      },
      "source": [
        "## 그림 9-9"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KvMRbpx2G1sr",
        "outputId": "e5a01440-e1dd-4abd-825b-ecd86de2a6d4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 575
        }
      },
      "source": [
        "fig = plt.figure(figsize=(10,10))\n",
        "ax = fig.add_subplot(projection='3d')\n",
        "\n",
        "plt.rc('xtick',labelsize=15)\n",
        "plt.rc('ytick',labelsize=15)\n",
        "\n",
        "x1 = np.linspace(0, 5, 101)\n",
        "x2 = np.linspace(0, 5, 101)\n",
        "\n",
        "X1, X2 = np.meshgrid(x1, x2)\n",
        "\n",
        "ax.contour(X1, X2, o(X1, X2), linewidths=5, cmap='gray', levels=[0.5])\n",
        "ax.plot_surface(X1, X2, o(X1, X2), color='w', edgecolor='k', alpha=0.5)\n",
        "\n",
        "ax.plot3D(samples[TP,0], samples[TP,1], 1.0, 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='k', markersize=8, zorder=10)\n",
        "\n",
        "ax.plot3D(samples[TN,0], samples[TN,1], 0.0, '^', \n",
        "        markerfacecolor='gray', markeredgecolor='k', markersize=8, zorder=10)\n",
        "\n",
        "ax.plot3D(samples[FN,0], samples[FN,1], 0.0, 'o', \n",
        "        markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15, zorder=10)\n",
        "ax.plot3D(samples[FP,0], samples[FP,1], 1.0, '^', \n",
        "        markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15, zorder=10)\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=15) \n",
        "ax.yaxis.set_tick_params(labelsize=15)\n",
        "ax.zaxis.set_tick_params(labelsize=15)\n",
        "ax.set_xlabel(r'$x$', fontsize=20)\n",
        "ax.set_ylabel(r'$y$', fontsize=20)\n",
        "ax.set_zlabel(r'$a$', fontsize=20)\n",
        "ax.view_init(30, -80)\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap10/fig10-9.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap10/fig10-9.pdf\", format='pdf', bbox_inches='tight')\n",
        "    \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5u9OHxrtG66t"
      },
      "source": [
        "## network 함수"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_JonZN7QG5Cr"
      },
      "source": [
        "def network(X, W):\n",
        "    \"\"\"\n",
        "    X : (N, D)\n",
        "    W : (3, 3)\n",
        "        [b^(1)_1, w^(1)_11, w^(1)_12]\n",
        "        [b^(1)_2, w^(1)_21, w^(1)_22]\n",
        "        [b^(2)_1, w^(2)_11, w^(2)_12]\n",
        "    \n",
        "    ret : (N,)\n",
        "    \n",
        "    D, H, A = 2, 2, 1\n",
        "    \"\"\"\n",
        "    X = np.hstack( (np.ones(X.shape[0]).reshape(-1,1), X) ) # (N,D)->(N,D+1)\n",
        "    Z1 = np.dot(W[:2,:], X.T) # (H,N)=(H,D+1)*(D+1,N)\n",
        "    A1 = logistic(Z1)           # (H,N)\n",
        "    A1 = np.vstack((np.ones(A1.shape[1]), A1)) # (H,N)->(H+1,N)\n",
        "    Z = np.dot(W[-1,:], A1) # (H+1,)*(H+1,N)\n",
        "    A2 = logistic(Z)         # (N,)\n",
        "    \n",
        "    return A2"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7djzBCXHG9bH"
      },
      "source": [
        "## network 함수 검증"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bQ_ZH0oRG8MC",
        "outputId": "3033824f-cf48-4bf7-edf0-20668dfdc802",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "W = np.array([ [-15, 3, 5], [18, -6, -3], [4, 10, -9] ])\n",
        "pred = network(samples, W)\n",
        "\n",
        "pred[pred>=0.5] = 1\n",
        "pred[pred<0.5] = 0\n",
        "\n",
        "result = pred==target\n",
        "\n",
        "np.size(result) - np.count_nonzero(result)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yw9I7rg5G-lq",
        "outputId": "3d94e303-1b17-43cc-8d42-39c9e7d456dd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "np.random.seed(17)\n",
        "W = np.random.randn(9)\n",
        "\n",
        "W.reshape(3,3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.2763, -1.8546,  0.6239],\n",
              "       [ 1.1453,  1.0372,  1.8866],\n",
              "       [-0.1117, -0.3621,  0.1487]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OUMoDdgPG_vL",
        "outputId": "7b7a66eb-e579-4dce-bcf4-fd469b549396",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "pred = network(samples, W.reshape(3,3))\n",
        "\n",
        "pred[pred>=0.5] = 1\n",
        "pred[pred<0.5] = 0\n",
        "\n",
        "result = pred==target\n",
        "\n",
        "np.size(result) - np.count_nonzero(result)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "163"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vWoMzvcRHAvz"
      },
      "source": [
        "def J(W, X, T):\n",
        "    \"\"\"\n",
        "    W: 함숫값을 결정하는 변수, 가중치 (9,)\n",
        "    X: 주어진 점 데이터 X, X: (N,D)\n",
        "    T: 데이터에 대한 클래스 T, 0 또는 1, T: (N,)\n",
        "    \"\"\"\n",
        "    N = X.shape[0]\n",
        "    W = W.reshape(3,3)\n",
        "    \n",
        "    Y = network(X, W)\n",
        "    return (1/(2*N)) * ((T-Y)**2).sum()\n",
        "    \n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fT1l4LVCHBw8",
        "outputId": "ddb8c3a4-1739-4cc1-ed9b-15dceca691f0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# 초기 상태에서 목적함숫값\n",
        "J(W, samples, target)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.1263148192185165"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aaNF_M9LHI_T",
        "outputId": "55fa8752-21dd-4ec1-b4af-5b658e6fe819",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "from scipy import optimize\n",
        "\n",
        "# https://docs.scipy.org/doc/scipy/reference/generated/scipy.optimize.fmin_cg.html\n",
        "W_star = optimize.fmin_cg(J, W, args=(samples, target),  gtol=1e-06)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Optimization terminated successfully.\n",
            "         Current function value: 0.000003\n",
            "         Iterations: 313\n",
            "         Function evaluations: 12694\n",
            "         Gradient evaluations: 1154\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Re56OeN1HLTj",
        "outputId": "d6c4eb99-a674-4fcc-980c-4b290241d922",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "W_star = W_star.reshape(3,3)\n",
        "W_star"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 42.9811,  -9.4741, -13.7029],\n",
              "       [-27.8283,   9.6497,   3.2045],\n",
              "       [ 19.2321, -40.3183,  39.3389]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2GFH4JWVHPxu"
      },
      "source": [
        "## 그림 9-12"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oNaLZq_qHNpL",
        "outputId": "82f2e770-db45-4f8d-fdef-8bb3233c60c5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        }
      },
      "source": [
        "x = np.linspace(0, 5, 200)\n",
        "\n",
        "xx, yy = np.meshgrid(x, x)\n",
        "\n",
        "dcs_bnd_1_imp_ = lambda x, y: W_star[0, 1]*x + W_star[0, 2]*y + W_star[0, 0] \n",
        "dcs_bnd_2_imp_ = lambda x, y: W_star[1, 1]*x + W_star[1, 2]*y + W_star[1, 0] \n",
        "\n",
        "a, b, c = W_star[2, 1], W_star[2, 2], W_star[2, 0] \n",
        "o = lambda x, y: logistic( a*logistic(dcs_bnd_1_imp_(x, y)) + b*logistic(dcs_bnd_2_imp_(x, y)) + c )\n",
        "\n",
        "fig = plt.figure(figsize=(7,7))\n",
        "ax = fig.add_subplot()\n",
        "\n",
        "ax.xaxis.set_tick_params(labelsize=18)\n",
        "ax.yaxis.set_tick_params(labelsize=18)\n",
        "ax.set_xlabel('$x$', fontsize=25)\n",
        "ax.set_ylabel('$y$', fontsize=25)\n",
        "\n",
        "pred = o(samples[:,0], samples[:,1])\n",
        "pred_pos = pred >= 0.5\n",
        "pred_neg = pred < 0.5\n",
        "\n",
        "ax.contour(xx, yy, o(xx,yy), cmap='gray', levels=[0.5])\n",
        "\n",
        "# for positive samples\n",
        "TP = np.logical_and(target==1, pred_pos)\n",
        "FN = np.logical_and(target==1, pred_neg)\n",
        "\n",
        "# for negative samples\n",
        "TN = np.logical_and(target==0, pred_neg)\n",
        "FP = np.logical_and(target==0, pred_pos)\n",
        "\n",
        "ax.plot(samples[TP,0], samples[TP,1], 'o', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "ax.plot(samples[TN,0], samples[TN,1], '^', \n",
        "        markerfacecolor='gray', markeredgecolor='gray', markersize=8)\n",
        "\n",
        "ax.plot(samples[FN,0], samples[FN,1], 'o', \n",
        "        markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15)\n",
        "ax.plot(samples[FP,0], samples[FP,1], '^', \n",
        "        markerfacecolor='k', markeredgecolor='w', markeredgewidth=2, markersize=15)\n",
        "\n",
        "ax.set_ylim(0,5)\n",
        "ax.set_xlim(0,5)\n",
        "\n",
        "if file_print == True :\n",
        "    fig.savefig(\"imgs/chap9/fig9-12.png\", dpi=300, bbox_inches='tight')\n",
        "    fig.savefig(\"imgs/chap9/fig9-12.pdf\", format='pdf', bbox_inches='tight')\n",
        "   \n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcMAAAHBCAYAAAAVTtDVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOyde3wU1d3/P5vdbJI1SPDBqlCBVGVaqyAkIZC1iSjWS/WBpxVtvaGtVH3VFipG1FYkT4tW01bRWrSlihf0Z0HbVEOLRlcXNxYSiEJNGLC2oEl8DFLKhgQme/n9ESbdbOZyZubMzJnZeb9evtTs7M7Zs2e+3/O9Hl86nYaHh4eHh0cuk2f3ADw8PDw8POzGU4YeHh4eHjmPpww9PDw8PHIeTxl6eHh4eOQ8njL08PDw8Mh5PGXo4eHh4ZHzBOwegBQcx70JoEbm5RU8z//YwuF4eHh4eLgcJpXhUbYBuFHi711WD8TDw8PDw92wrAzjPM+32j0IDw8PDw/348UMPTw8PDxyHk8Zenh4eHjkPCy7ScdyHLcGwLkATgDwAYBf8zz/qK2j8vDw8PBwHSwrw1IALwL4FoAxAG4C8CuO44p4nv955oVbt271uo17eHh4eAyjrKzMR3otq8rw6wASPM8fFP/AcVwjgHcA/C/HcY/zPB/PfENZWZnFQ3Q+XV1dGDdunK1jEAQBq1evxr/+9S8kEomhvwcCAYwZMwY33HADgsGg5eOqr69HX1+f7OuhUAi1tbWmjyMSiaC5uXnY3IgEAgFUVVVh9uzZ1N9rBiysNxqorQ1gcH2Ul5cjHA4bXr9umTc9kMy1HJdccomm65mMGfI8vz9TER79WxpAA4AiAF+2ZWAe1InFYiMUIQAkEgn861//QiwWs2Vc5eXlCASk94p+vx/l5eXD/iYIAiKRCOrr61FXV4f6+npEIhEIgmBoHK2trZLKDBico9ZW+YRrI+/1kIdEOPf19aG5uRmrV682vAZyGaXnkDZMKkOO4/I4jpOagaKj/z5s5Xg8zINFgS0IApLJJJLJ5IjXAoEAjj32WITD4WHXr169Gs3NzUOCkpYwVBO8/f39przXQ55QKER0nd0bOjcQDocxZsyYEQrRDAXJnDLkOO4UAP0A7sv6ux/APACfAXjfhqF5mABrAltUbJs3b4bUwdcnnngiLr744mGuLzOtWzXBW1RUJPuakfd6yKPFWvEscGMEg0HccMMNqKqqQigUgs/nQygUQlVVFfX1y1zMkOf5v3Mc9wcAP+Q4LgHgNQDFAL4H4EwAN/A8P2DnGD3oEQqFFBWi1QJbTrGJdHZ2YsOGDbjpppuGFCKJdas3NldeXq4Y98t219J6r4c84XAYHR0diuskE88CN0YwGMTs2bMlnyG59a0H5pThURZgsB3bQgBLABwB0Abgv3mef9nOgXnQhVRgC4KAWCyG1tZW9PX1UU1QyERJsQFAOp3GwYMHEYvFhh5OM61bOcErJhhlumtpvtdDHtFayVyPSiht6Kxa11Zg9XdR2pT4fMRJpEMwqQx5nj8C4IGj/3i4GBKBLZVxKsbkOjo6qGackiRHJJPJYdaemdZttuDt7+9HUVERkZAx8l6P4cgJ+kWLFiEWi+mywK1c12Zjx3dRW987duzQ9HlMKkOP3IFEYEciEdWYHK0SATXFJpJp7ZntjlRyE5n5Xo9B1AT9tddeq8sCJ4k1O+V3s+u70FzfnjL0sB21BW1mTC4bJcWWSaa157kjnYcWl56aoG9padFlgdNa1yy4Wq18Rs3CU4YezGNlxqmo2D777DOkUinJa7LrDD13pLPQ6tIjFfRaLRQa61rLdzFTabKWFa4HTxnqhIXdWK5gZcapqNii0Siam5tHlFcEAgGMGjVqhLXnuSOdg1aXnlmCnsa6Jv0uZsf0aD+jdshX5uoMnYCZRdYeI1Gq6zKjRCAYDGLOnDm44447UF1dPaK+ae7cud6Gx8FobfRgVr0mjXVN+l3M7vRE8xm1S756ylAHrLYQcytKXSjMjMmJ1l5tbS2WLVuG2tpazJ49G/n5+abcz8MatFp6Zm3GaKxr0u9idqcnms+oXfLVc5PqwCnBYre4cr2YnAdNtLr0zEqQorGuSb+L2TE9ms+oXfLVU4Y6cEKwmCRG4CRIY3Ju2QB4mIfWUhgzN2NGY82k38WKuDutuLld8tVThjpgrYWYFCSuBo7jbBqdObipiNnDPPRYemYmSBnZwJF+Fye15rNLvnoxQx1YndChBxZPgzAbL5brQYJS82erN0xGk0VIv4tdcXc92CVfPctQB04osnaCK5c2ahuAlpYWJmK5HvbDSikMDQ8OyXdxUtzdLvnqKUMdOGFhOcGVSxuSDYAgCEz8PizjtLir08abCYkHh1Y4g5UNgBp2yVdPGeqE9YXlpBgBQEegkfQVJemR6GThahSnxV2dNt5sctGDQ4Id8tWLGboUJ8UIaBXZkhYpWzEWp+K0uKvTxpuNdwAzO3jK0KWwlCSgBi2BRqLg1XbaTheuRnFa4pXTxpuNE5LxcgXPTepiWE0Hz4ZWkW0wGERRUZGiwlPbaVsZw2ERp7ntnDbebEiSRfbt22fjCHMHTxl6aIZ2nIamQKuoqDAUK3W6cDWK1sQrM+KrWj7T6YliTkjGyxU8ZcggrCdw0D7Ik6ZAM5qWzapwtWpNaEm8MiN5RetnOi1RTArWk/H0wrocy8a/fPlyu8dgmO7u7uXjxo2zexhUEIXB7t27ceTIEQDAwMAAOjs70dHRgalTp8Lv9+v+7Gg0ivXr16OpqQnt7e0YGBjA+PHjNX3m+vXrh8aWTSqVQk9Pj6YEHUEQ0NnZKXl+YCAQQGVlJUpLS4k+y+/3Y+rUqQCAnp4eJBIJhEIhVFZWYt68eaoPIclYxo4di1GjRhGNhwZmrolsxo8fj46ODgiCMGwOxM3EvHnzhu4VjUaxe/fuEYoolUoNJRpl/m7xeFx13rR+ppbxOhWSeWMNK9esHN3d3Rg3blwd6fVeAg1jmJXAIZUlefjwYV1ZkrRdibQzX+VOmyDZjbKYhWtlUo+WxCszkle0fqaTEsVyCScmonluUsYwo2O7IAhYu3Ytenp6JD9Tq2uTtiuRpbgJS2MRsbqLP6nbzoz4qp7PdKub0ck45WSfTDxlyBi0BYxoEUopQhGti9OMOA1LAo2lsQDsJvWYEV9lIYHHwzisrlklPGXIGLQFjOiuUEPL4nRCb1Y3wWpSjxmbIrsTeDI/21Oy+mF1zSrhxQwZg3YRrpK7IhMti9OL01gLq4XZZsRXlT6zpKQEyWQS9fX1qKurQ319Pfbt22dJfD2XuhDRgNU1q4SXTcoYctlxeXl5SKVS2LNnD1paWiAIAlEWaFNTk+o9tWZsAoNZm6WlpaioqIDP50NXVxc++OADbNu2jXhsTsXq7D5WMya1Zu6SzJvcZ5aXl+Ozzz7D3//+96HsRKmMXxE9Wc0iWjNazcaJ2aQsrFmt2aSem5QxshM4+vr64PP5kE6nkU6nAWhzBZE0r9a7i3d6k2SnwGJST+bYaMdXpT4zEongwIEDRF4OEb1xKScmf0hhp6uX5TUrh6cMGSRTGEQiEckYCmkWqFIMBgCOPfZY9Pb24r777tP8sBgtvvfiMuSwltQjYtVvSOruz0RvXMqJyR/ZsLBRZXXNyuEpQ8YxuktVSnZJpVI4dOgQkskkgP88LO3t7eA4Dm1tbYoCzsjYWHhYPYxh5W+opqCyMRKXcmLyRza0u0QB7t+8egk0jGN0lyqX7CLGWEVFKJJIJLBv3z6i5AEjY3NiUa7HcKz8DdWOOspEKYFHEAREIpFhSTiRSGTYunZi8kc2tBsi5EJSkWcZasCOnRGNXaqUu6K+vl4xAUGMT4pI7SiNjM0tcZlcxsrfUMndn5eXh7y8PCSTScW4FKkl64bSIdquXjMsTdbwlCEhdrn1zGpErNXtBIwUcEbG5oa4TK5j5W+opqBInj9Sga6U/FFRUeEIVyFtV28ubF5dpQzNtNzs2hmZtUslyTKVIlPAGRmbG+IyuY6VvyGN7EQtAl3Km+KkODftTXQubF5dowzNXqh27YzMSlEuLy9HLBYbETNUI1PAGRmbG47e0YrbEhCs/g2NZicaFehOchXS3kTnwubVNcrQ7IVq587IjBTlcDiMHTt2IB6PD5szsbhfCikBp3dsbojLaMFJVgUpTvsNjQp0J7kKaW+ic2Hz6hplSGuhyu3ei4qKFBWe03ZGwWAQc+fOxYcffjjsYZk2bRp4nh9R4ExbwDmxKNcITrIqSHHab2hUoDvNVUhzE+20jY8eXKMMaSxUpd17MBiE3++XdCs6dWeUn58v+bBUV1dbIuCcVpRrBCdZFVpQ+w1Zcg0bFei54CqUw2kbHz24RhnSWKhKu3cAKCgogCAIrt0ZieSSkrIKp1kVNGDNNWxUoOeCq1AJt8sF1yhDGgtVbfceDAZRVVXl2p2Rh3lYZVWwZImx6Bo2ItBzwVWYy7hGGWpdqFJCg2T37uadkYd5WGFVmGGJGVGubnMN54KrMJdxjTLUslDlhIYabo4JeJgLTatCTkElk0mqlphR5epG17DbXYW5jGuUIUC+UOXcN0rkQkzAwzxoWRVKCiqZTI5ooyeixxIz6ubM5YSTXIYlV70WXKUMSdF6HIwTYgJOXYC5hNxmTWweTfLbqSV5KaHVEjPq5iR1DXtr1z2wljSlhZxUhiQu0VAo5JiYgJMXYK4j99tFo1FEo1EUFRWhoqJiaP3pOddPRKslZtTNSeIa9tauu2AxaYqUnFSGau6bUCiE2tpaC0eknezdtBROWIC5jprLvr+/f5hi0NNPFtDn5jfq5sx2DUtZfZFIhDnh6Vmq+nFy0lROKkOn1wtJ7ablYH0BmoGThBmJpZepGNQUlM/ng9/vp5L6T+M5yXQNd3V1DZ2jKcKa8PQsVWM4OWkqJ5Wh0+uFtCYAsbwAaeM0YUZq6YmKQU1BVVZWwu/3U0n9t+I5YU14OtnNJ2LnZtDJSVM5qQydXi+kNW7E8gKkjZIw6+npwdq1a3HVVVcx8xtrOUqrv79fVUFVV1cPWWNGseI5YU14smapasWqzaCcwp02bRo2b97sSK9bTipDwNn1QlriRqwvQNqobRT27t2L1atXM2MhKll62RQVFQ0pqGg0OkLoTJ48mfr4zH5OWAtZsGapasUKy1ZJ4ZaUlKCkpMT0Rv9mkLPK0MmQWhOkC9BJMTY1SOaFJXeXnKWXTbZi2LVr17DXE4kENm/ejF27djGj6EmwI2QhrveWlpZhyq2oqAiBQEDxd2Ddy2KFZaukcA8cOEDVVW8lnjLUgd3Kg8SaIB2T02JsapBsFFhyd2W6IrOFs0i2YnBDXEvE6pCFuN73798/4gSa/v5++Hw+2fc6wctihWWrpnDb2tpQW1uL2bNnD5OV0WiU6Y22pww1woLyUNtNaxmDmwQrQO52ZMndlemKzBQecoqB9u7f7s2dlSELcb1LHcUGYKiDT/Yh105w8wHWxGBJFS4LslILnjLUCAvKg+Zu2q6EAbMEsLhR6OnpUbyOVXcXiWKguft3msAyCmnyWV5eHgoLCx3l5gOsicGSKlwWZKUWPGWoEVayzWjtptUEq94ibyXMFMDiRmHt2rXYu3ev5DVOcHcpQXP37zSBZRTS9ZxMJplvvCGFFTFYUoXLiqwkJc/uATgNFrPNxN6W9fX1qKurQ319PSKRCARBUH1vKBRSfN3n8xF9jhZIBLARgsEgrrrqKhx//PEIBIbv95zi7lKivLx8xPcS0aroSQSWm1Bb7yKseg7UEDeDVVVVCIVC8Pl8CIVCqKqqomblh8NhjBkzRvXZYlFWKuFZhhphrS7KqJVVXl6Ot99+e1h8JBva1oEVO0an15IqQXP37zSBZRSSmLLTPQdmx2BJny3WZKUanjIkgKQPKED3ISKNqRl1c4XDYWzatEn29XQ6Td2dYZUAdnItqRI0Fb3TBJZRxI2EVDYp4EzPgR0JUCTPFms1pGp4ylAF0j6gNB8iLdaeUSsrGAzKnoEnQts6yDUBbAa0FL3TBJZRlEpZWE77l4PlBCgtHgy7M5oBTxmqQtIHlPYPp8Xao2FlWa2cck0As4zT+/TqwU0eA5YToEg9GKwodE8ZqqCWim3GcU9arD0aikxJOYkJNHV1ddTLH3JJALOKm2OruYBZ8XdalhrJxoMVhe4pQxXsSDDQck8aVpZSS7B0Om1a+YMngIdjl6vITZZSrmGGfLLaUmOlBMNThirYEd/Sck8aVpaUcvL7/UilUiOyTGnt1jwBPBxWXEUezsIM+WS1pcZKRrNXZ6gCzZouM+5Jq65IVE61tbVYtmwZgsGgbLmFG+vP7Mbs2ksPd2KGfLK69lSt9tOqhDpHWIYcx1UDeBNAlOf5c5Supe1qsiO+pfWeZlhZrOzWcgVWXEUezsIM+WT1s89KQh3zypDjuAIAvwEg307+KHKupmg0img0iqKiIlRUVGhSjHbEt1iIqXnlD9bibT489GCGrLD62WcloY55ZQjgxwDGAFC1zdXKIPr7+3XFYOyIb9kdU6O5W2Ohhoh1vM2Hh15oywqrLTUWNv8A48qQ47gzACwFcCOABWrXk3SkZ6H+xgnQ2q15iSFksOIq8vCoqKiQjVGn02lUVFRQv6fdm3+AYWXIcVweBt2jzTzPP8lxnKoyJO1IbyQG09vbi3Xr1g07EWHChAmYP38+iouLNX8eq9DarbFSQ8Q6rLiK3ITnkdBHS0uL6utufGaZVYYAbgYwHcBU0jeQnHIuoicG09vbi4ceemhET8O9e/fioYcewuLFi12nEI3u1rzEEDJYcRW5AUEQEI1G0dzcPKzVoBkeCTcq3NbWVtnDj5PJpC3PrBXzzKQy5Dju8wDuA/Aznud5kvd0dXXhi1/8It577z3ZHzKTgoICdHV1aRrXyy+/rLhI1q5di0svvVTTZ9pJPB7XPAdaITkv0ewx0MbMeeM4DhzHDfvbvn37TLmX1Zi93gYGBtDW1obt27fL9ttNJBLYv38/Nm7ciLKyMsP3a2howMGDB4fkQl9fH2KxGHbs2IG5c+ciPz/f0D0Aa57TTFh7Zq2aZyaVIYBHAXRhUCESMW7cOIwdOxYfffQRUVPtGTNmYNy4cZoG9cknn6i+rvUz7aSrq8v08apZ66FQyFFzBmibNzdaDnoxc72Jsel9+/apNp5PJpPYuXOn4Y1rJBJBPB4fsUFOJpOIx+P48MMPqVhQVjynmdB8Zmmsf73z3N3dTfT5IswpQ47jvgHgUgCXAMjnOE5U+f6jrxcDEHieH3HirFJHehE3xmBoClzawjuXE0O85CHrEGPTaopQhEapiltDALSeWVrr36p5Zk4ZYlAR+gA0yrweB1AHYLnUi5lxrkzB7tYYDE2Ba4bwzuXEEKclDznZiiXJJM+ERqmKW2tDaT2ztNa/VfPMojJcAWC1xN8fOfrv7wPYK/H6CGin606YMGFYFqnU61YTjUYlXUN6BK4ZwjuXE0OcZDk43YolTZwD6Hkk3FobSuuZpbX+rZpn5pQhz/O7AezO/jvHcf8++vrblg/qKPPnz5fMJgUAv9+P+fPnWzoeQRBGZMxlolXgmiW8WaghsgMnWQ5Os2KzIc0k9/l81DwSbg4B0Hhmaa1/q+aZOWXIMsXFxVi8eLGhOkOarqhYLEb1lHonCW8n4CTLwUlWrBRKAlPE5/OhqqoK1dXVVKzcXA4BAOqyjNb6t2qeHXNqBc/z56g16baCYDCISZMmDXVaD4VCmDRpEtHDJbqimpubhxaJ6IpavXo1BGFETpAiJN3jtQhcVrrHuwU7TjzRi9M3QuFwGGPGjJGcb5/Ph3A4jDvuuANz5syh5u6ldWKMEyGRZUrrX7y+vr4ekUhEUfZZNc+eZagBo3EV2q4oEreQFoHrZrePHTjJcnCSFSuFXbFpK0IAAwMDiEQiTCU2kcgypUPDRUjlpxXz7BjLkAWMnjlH+5wwNUtO3BGTIre7ZlF4OwEnWQ5OsmLlyD6Ts7a2FrNnz2ZqnrUiCAIaGhqoeZNoQSLLste/HKyc2elZhhowGleh7YpSsuTy8vIwa9YsTYIglzM/zcIpyUNOsmJziVgsNqzziohebxKtnAVSWZa5/uvr62Xfx0Jc2lOGGjCizARBgM/nU0x40eqKUhNg1dXVmj4PcI7w9qCLtxFiE6U+oVoVCM3yGT1uddbj0p4y1ICRuIqaCyAvL0+zK8oTYB408TZC7EFTgdDMWdCTX8B6XNpThhowkmDS2tqqaBWm02ldrihPgHl4uBeaCoRm+YwetzrrCXpeAo0GjCSYqO3w0um0Z8l5eHgMo7y8HH6/X/I1rQqEppWpJzmM9QQ9zzLUgBG3JEkneA8PD49MwuEwduzYgXg8bjixibabUqtXSqv8tLpXbk4qQyOTrNctybqLwMMjl2G1SXkwGMTcuXPx4YcfGs4LYEEGkcpPO3rl5pwytKshsZe67uHBJqw3Kc/Pz9e1Ac9W8EVFRUPfg3UZZEev3JxQhtmLQgqzGxJ7mZ9ksLpDdyvefDu/SbkUUgq+v78ffr8fBQUFCAaDTMsgO3rlul4ZSi0KOcwu/PQyP5VhfYfuNrz5HsSpTcqVNjJyCj6ZTEIQBFRVVTH5nUTsqEl0fTap3KKQw+7Cz1zGaLs7D2148z0I68XgUqg1ym5paaHa+tFq7Dg0wPWWodYTsP1+PwRByIkdMWs4dYfuVLz5HoT1YnAp1DYyajLPagWv1R1vR7KP65WhlhOwASCVSmH16tU54yJiKWbkxB26kzHaXpCVdWMUFrIstaK2kVHDSgWvxx1vR8Kh65Uh6QnYIqlUyrFBc60MDAwwFTNy4g7dyeidb7fFGo0KXjs2BiQyLRAIMKHg9SQo2ZFw6HplSHICdjZucxH19vZi3bp12Lt379DfJkyYgGOPPZapLDqjO3Q3WStWoHe+3ZZ9aUTw2rUxIGniccwxxzBRyqXXHW91wqHrlSHJAZNSuMUl19vbi4ceemhE5/tMxSiFHRsCIzt0Vq0VlhW03vl2Y6xRr+C1a2NAspERs0rtLuVySvjD9cpQbtcnCIKicnSLS27dunWyR8CoYfUiNbJDZ9FaYVVBi+idb6cINyuwa2NAspFhpZTLKeEP1ytDQHrXF4lEHBc014OaBaiEHYtU7wPMorVCoqAzd+92WI565tspws0K7NoYsNjEQ84LMm3aNGzevJl5WZsTylAKrz2aMlYuUhquRBatFTUF3dLSMmINsmQ5yuHE7EuzsHNjwIrlByh7QUpKSlBSUoIDBw4wLWtzVhmyuLOyi+ysMysXKS1XIovWComCHhgYYMq1S4K3kfwPubIxUNuwKnlBDhw4gMrKSvj9fqZlbc4qQ4CtnZUWtFhSEyZMUHSVnnzyySgtLbVtkdKK9bEolEjKelhz7ZLgbST/Qy5sDEg2rGodb9ra2lBbW8vkehbJaWXoRLRaUvPnz5fMJgWAvLw8XH755SguLrZtkdKK9bEolNQUtFp2c19fny3dkEg2W07dSNImFzYGahvWaDSqGoZwQlKVpwwdhlZLqri4GIsXL5asM/zKV76C4uJiy8YuBa1YH4tCSU1B9/b2qn4/2t2Q1BQd6xmwLOL2jYHahnXz5s2qn+GEpCpPGTKMlOBSKgmRs6SKi4tx/fXXj7i+q6vLlHFrgWasjzWhpKagY7GYakMImrFDEkXHYomKh73odfVn4oTYqacMGUVOcKnhBHdEJizG+miipKBFy7Gnp0f2/TRjhySKjsUSFQ970drSUgonxE49ZcgoWo+eEnGCOyITFmN9ViFajvfdd5/idbQ2OCSKzsoSFUEQEI1Gh9WgBQIBVFZWorq62nPHMoLahhVQtg5DoZAjfktPGTKK1qOnAGdaUnpifSy3ONNKMBi0rCyERNFpHYve30IQBPz2t7/Fvn37hv09kUggFouB53ksXLjQUb+nm9ZlJmob1smTJzuiqF4NTxkyila3hJMsKTmhsWjRIlWh4cYED6tcxSSKTstYjPwWsVgM+/fvlx3LZ5995qj4pBvXpYjahhUAdu3a5XjvjqcMVbBrt6cmuAKBAILBIBNZk1owKjTcmOBhlauYtLkz6ViM/Batra1IpVKyY02n06bEJ816nq1el1bLJbXkNNYyufXgKUMF7NztqQmuqqoqxwl9wLjQcGOCh1VlIaTNnUnHYuS3sCMZzMzn2cp1yaIVylomtx48ZaiAnVaIWxNLjAoNGgkeLMZ2rBAmpIqOdCxGfguSDEXayWBmPs9WJh650TvCAp4yVECv4KYhbFksIqeBUaFhNNmExV21ldBUukZ+i/Lycrz99tuyrlKfz0c98cJM683K3rhu9I6wgCuVIa2dvx7BTVPYOtX1oDT/RoWG0WQTb1dNDyO/RTgcRnt7+4hsUpH/+q//ou79UHue+/r6EIlEdMkNK+tlWTyhxQ3k2T0A2ojKqLm5eWjRiMpo9erVEASB+LNCoZDi61KCm0TYuhm1+Z82bdpQbVI2JEIjHA5jzJgxIz6D1H1Msqv2GIkgCIhEIqivr0ddXR3q6+uRTCZRUlKi67cIBoNYuHAhwuEw8vPzh703HA6bUlah9jz7fD7dcsPoutSCHrnkoY7rLEOaO389u71cd2GozT8AjBkzRncs1Kj7ONd31Xq8JnLejs2bN6OkpASVlZVoa2vT/FsEg0HMmTMHc+bMof49pVB6nn0+H3w+n265YWVYQ49cYjFOrhWzv4PrlCFNZaQniSXXha3a/Le1tWHRokWGhIYR9zGL5x5ahV4XvtpZdX6/H7W1tZZ8ByMoPc/JZFI2fkkqN6wKa2iVS26Ik1vxHVynDGkqIz27PVaErV07QZL5tzMWqrSrBv7jDtQzT6zvvvV6Tdzi7VB6nqPRqOJ77TpOSwqtcskNcXIrvoPrlCFtZaRVcNMKpBsRrHbuBFnZDMght6sWSSQSuubJCbtvvUrNTd4OueeZpC8r7eO0jKBFLrlhM2PFd3CdMrT7FAQa9YFO7tJi9/yrkbmrfueddzAwMDDiGj3z5ITdt16lxvoGRytSG82xY8eis7NT8hBsEVZ+R624YTNDkglcX18/zHA49thjNSp8AogAACAASURBVN3DdcpQaeefTCaRTCZNdXfQCKQ7uUuLE5oFiLvq1tZWSWUIaJ8nJ+y+9So11jc4WpDbaJKc7ZlIJNDS0gIAxB4bFlznbtjMkDRpyM4CvvDCCzXdw3XKUFRG0WgUzc3NSKfTQ6+l02ls3rwZu3btMtXdYTQmxkKXFr04qVkAzXlywu5br1JzwgaHFKWNplzJTyb9/f3D5lDJY2Ol61xJ6bphM6MW68+G9LpMXKcMgUGB7Pf74ff7mXZbyWF3lxZSlB5AJzQLoDlPTth961VqTtrgqKG20SSBVKZY5TpXU7rXXnut5O/u8/mQTCYRjUbR2trK9O+pFuungX/58uWmfLCVdHd3L1+7di0EQcD48ePh9/uxfv16HDlyRPL6VCqFnp4eZne0LS0tsu47YFDwKo1dEAR0dnZKpoqLh6eWlpYiHo9j1KhRusYoPoC7d+8emueBgQF0dnaio6MDU6dOhd/vH7o2Go1i/fr1aGpqQktLy7Dfyi5I5ykbqXnT+1lW4vf7MXXqVABAT08PEokEQqEQKisrMW/ePEUh6Pf7UVpainA4jJqaGoTDYZSWlmr6/YysN1o0NTWpXhMIBBRP1JBCSqbQkkFq8xaNRrF79+4RSiKVSkEQBPj9fsybNw/A4O8+MDAwVFcpes7knl1WkFu7SnJy8uTJGDduXB3pPVxjGWa7H5zgtpLDqFvDCrcW6a6X5SxLmvPkFFeimgufhRiXmRQVFSk++0VFRSguLpb8HdUskuzPtUoGkYZVxH8ikYikfGHdaya1dsWkGRq4qh1b5o/p5JZFRls7iW6tqqoqhEIh+Hw+hEIhVFVVUVM+pG3NWG5PR3OerJhzs6HZypBFBEGAz+dTvGZgYACTJ09GZWXliN9RTWZkv26VDNKqdN3UkrC8vJwo1kuCayxDEfHHdHLQmEaMxuzCdtIHUE8ykJXWCc15cmpjdREnlIcYIRaLybotRRKJBDZv3owxY8Zg0aJFI9abFplilQzSGq92stcsGyWPjFZcpwyBwR/TKW4rOaQEq6gkWlpahi3YoqIiVFRUECsL8XO2bNmCw4cP61I2pA+g1gePZbeq23FCeYgRWltbFesIReSUv1aZYpUM0qp0nZDsRYqS4aAVVyrDoqIiV2XAAdJKQqS/vx+xWIxIWdBSNqQPoNYHz+3WiQiLsTk3WQxSaIktSSl/rTLFKhmkVek62WsmhZxHZuvWrZo+x3XKMPPHdLrbKhM5JSGSTCaJlAUtZUP6AGp98NxunQDsWr9ushikICnczkRK+WuVKVbIIK1K1+leM7NwlTJ084+ppCRESJQFLWVD+gBqffDcbp0A7Fq/brMYstFauO0k5a9F6brNa0YL1yhDFtxMZkK6o1VTFrRP9VB7ALU+eG63TgB2rV+3WwxaCreVlD+LLm6tuMlrRgvXKEMnnKdmBFIXj5qysEPZaHnw7LJOrBRwrFq/brcYsr9fX1/fUPF5ZpG9kvJn1cXtYRzXKEO3Q+LiIVEWrLvC7LBOrBZwLFu/brcYsr9f5iaIRPmz6uL2MI6nDB0CiYuH5FQO1l1hwWAQ1157LdatW4e9e/cO/X3cuHGYP3++KbtuqwUc6xuSXEKr8mfVxe1hHNd0oMkUnG4ku8OJFOl0Gu+88w5++9vfynYLyfycwsJC5jqlCIKAp59+esSROl1dXXj66ad1d0ERT7Cvr69HXV0d6uvrEYlEIAiC5R05jHYY8rAPVl3cHsZxjWV4xhln4JFHHsGCBQvsHoppZO5im5qaJFuZpVIp7Nu3D9FoFHPmzFH8HI7jMG7cOLOHrQkzrDQ1NygtAUcad7QzNueG5A87YdnF7WEM1yjDeDyO6667DpFIBKtWrXL9oty8ebPi61u2bJFVhixjhhtKTcGqNWEmWUta4452xObsSv6g0fGIFTwX9yBu3FS5xk365JNPoqioCE899RTOOuss/PnPf7Z7SKailhqudLQJy5jhhiI5w06ulyGpgGO5IbmIHWPMbP59+PBhAM5u/u25uN3b0N01luF1112H6dOn44orrsDOnTtx8cUX4xvf+AYeeeQRnHTSSXYPz4MQM9xQago2kUjg+OOPN5RU5ITECqNj1GMN6HF7s2x1uL38hAS3ZtS6xjIEgClTpuC9997DL37xCxQXF+PFF1/El770JTz77LN2D406al3ZaR1rYjVKR7LodUOpHaUTCoUMH7/khMQKI2PUaw1oTU5ygtUhurhra2uxbNky1NbWYvbs2TmhCAF3HQGVCbMSk+O4SwHcDuAMAEEA7wH4Oc/zLym9LxgM4tZbb8U3vvENfO9730NjYyOuueYaxGIxPPTQQygoKLBg9OZTWVmp6NaqrKy0cDT0MKP0gyTOYzSG54TECiNj1GsNaFXAVlgdLFueTsAJGz89MGkZchx3NYA/AfgngMsBXAFgAMCLHMddQfIZEydOxMsvv4zf/OY3CAaDeOyxxzB9+nQ0NzebNm4rqa6uxtixYyUPKy0qKsLMmTNtGJVxzDgk14o4jxkWLW2MjFGvNaD1gFuzrQ4nWJ6s4+SD05Vg1TL8KYBNPM9fI/6B47gogI8A3AjgBZIP8fl8WLhwIaZPn45vfetbaG9vx9lnn43vfe97qKurw3HHHWfO6C0gGAxiwYIFWLVq1Yidmlirx0LdoBJKO3SamZZWxHlYb2YAGBujXmtAa/Yl6X30WndujXdZiVszapmzDDmOKwRQD+DuzL/zPH8QwE4AE7V+ZllZGbZv34677roLPp8Pv/rVr3Daaafh97//PZ1B20RLS4vkTjbzOCdWsXqHbnacxwyLljZGxqjXGtBqlZPcx8jacWu8y0rcmlHLnGXI8/xhAI9m/53juHwAEwC06fncwsJCrFixAldccQUWL16MSCSCK664As3Nzbj//vsdGUt0QgajHG7coTuhr6feMeq1BjKt8i1btuDIkSOKVjnJfYysHbfGu6zErRm1zCnDbDiO8wM4BcB9AAoBLDPyeVOmTMHrr7+ORx55BEuWLMHKlSvxxhtv4KmnnsK0adNoDNkynPxgO1mR5yJGXKxaOh6R3GflypWqayccDku6UYuKihSfC6fGu6zGCRs/rTCtDDmOuw7Ak0f/910Ac3ie3yp1bXYvSzUuu+wynHLKKbjllluwY8cOzJw5E9///vfx3e9+F8XFxYbGbRWFhYVDhcxSFBQUKM5LPB7XPG+0UFPkfX19to1NDTvnzU6+9rWvYfv27Whvb8fhw4dRWFiI008/HVOmTMG+fftU3086b2r3IVk7jz32GA4ePIhkMjn0t1gshvz8fPj9/qG/Z+L3+/HFL36Rud82V9eb1TCtDDGYUVoG4CQAVwOIcRx3E8/za7Iv1NNjc9y4cZg9ezZuu+02rFq1Cr/4xS/wzDPP4IknnsDXvvY1o2M3nRkzZii6lGbMmKE4L11dXbb1JlVL8w+FQsz1TRWxc97sZuLEibj00kt1vVfLvCndR23t5OfnIx6Pj1B44v8XFBRAEARJy/OCCy5gxs3npjZ2SmhJhtJybXd3t6Zx+NLptOEvYxUcxz0L4BsAxvE8/y/x71u3bk2XlZUZ+uxoNIqlS5fir3/9KwDg1ltvxb333st0LFGq1yTwnwdbLSnCTqEeiUQUFXlVVRWzLhiWlKGTauZozZva2gGU2xUWFRWhoqKC6XiX0Web1hjMXltavqfWOdm6dSvKyspG1p7JwGI26Ukcx93AcdzpEi9vw2DccDLt+1ZXV6O5uRkrVqyA3+/HL3/5S1RWVqK9vZ32rajhhAxGOdyakWYluVozp7Z21Pr2Hj58mPkOMnb3urVqbWn5nmbPCXPKEEABgN8CuFPitVlH/23K4YU+nw933XUXYrEYTjnlFLz33nuorKzEL37xC8XYnJ04tTWUkxU5K9gtMO1Cbe24oSjc7hIQq9aWlu9p9pwwFzPkef6fHMc9A+AajuMOAvjj0Ze+DuAyAGt4ntfmDNZIZWUl2tra8N3vfhf/7//9P9x22214+OGHsXbtWpx99tlm3jqncGNGmpW4MSNXy5mQcmvHDUXhdmeKW7W2tHxPs+eEOWV4lG9jsBfpgqP/fQTAhwCWAnjQigGMGjUKzz33HK699lrccccd2L59O2pqarBkyRL85Cc/YTqWSBunxKXEcba0tAx7MMQYEWvjNYrdApM2tM5bpNENyO41T9JHNhKJmDY+q9aWln65Zvf/ZVIZ8jyfAPCLo//Yhs/nw0UXXYTzzz8fy5Ytw/3334/6+nq8/vrrWLduHb7whS/YOTxLsOtAWK2I49y/f/+ILML+/n7mxksDNeFQWFhoqsCkDa1GDEaLwllY82rWLYBhr9Men1WN57VY8WZb/P7ly5cb+gAW6O7uXm5mdl9eXh7OO+88XHDBBXjjjTfQ0dGBZ555BqNHj8a0adOQl8di6FWdeDyOUaNGKV4TjUaxe/fuEQswlUoNBdFLS0uH/i4IAqLRKNavX4+mpqahlnHjx4+H3++n/yWyxilVP6Y0Xj2QzJsVCIKAzs5OpFIpydcTiQT27NkzdNDzwMAAOjs70dHRgalTp5r6e0ihNm/r16/HkSNHJF9LpVLo6ekhTqzy+/0oLS1FOBxGTU0NwuEwSktLib6z1jVvBuPHj0dHRwcEQRj2+wYCAQSDQRw5cmTEWqc5PqW1FQgEUFlZSWUOlL7nmDFjMG/evKHfTMu1wGBpxbhx4+pIx+IpQw18/vOfx4IFC9De3o733nsPjY2NaGhowMyZM3HiiSeafn/akAh1LQJK3FHv3r176D1WCWClcWaOd8+ePYYVNCvKUE44KGGlQM9Gbd6ampoU359IJFBTU0N7WCOgqZT14vf7MXXqVADA//3f/yGZTCIUCqGyshKdnZ2y2Zzi+CoqKgxtSrUqHhrfs6enB4lEYuh7zps3b5iFq+VaQLsyZNJNyjIlJSVoaGjAiy++iNtuuw3vvfceysvLh2KJLLqfjKAldmBnv1G1cWZf6wa3abY7kHQOWE2uYeVMSK3xMrPii3Jt7KLRqOr4jbp5rew/qiWRzsykO08Z6sDn8+Gyyy7DhRdeiB//+Md4+OGH8cADD+Ctt97C+vXr8fnPf37oWrsD8UbRIqDszG5UG6fUeJzaEDyTTOFQX19PPAcsJtewkgWqZc3bEV9UG18gEKAWe9WjeJwq85wZ7NJIPB7HmjVr0NvbS/Vzi4uL8dBDDyEWi+Hkk0/G5s2bcdZZZ+HZZ59FOp12RVG0lgNh7cxuVBqnHG47skfLZoDFWjvSRgyCICASiaC+vh51dXWor69HJBKh9jxpWfN21HqqjU+8vxRG1jzJvDtZ5uWEMoxGo9i7dy/eeustUz5/1qxZ2LZtGy644AJ89tlnuOaaa3DRRRfh5ZdfdnxRtJZOMXYWO4vj1BrHYNFC0ova/IuwWmtH0ojBCmGrZc3bURxvtAOPnjVPOu9ObgTh+gSaeDyOhoYGpFIpfPrpp5g+fbqiqR6Px/H888/j1FNP1WTSh0IhXHXVVZg0aRI2bdqEv/3tb5g0aZKscKYV6DYCSSKIlqC1VRloauP89NNPVQWCSCgU0pwMwUoCTTZq2aUA/QQILZCuN6UsUCsyPbWseSuSfrLnTW1827ZtG8oglkLPmieddxaSj0S8BJosotEoxGbk6XQab731luKJFJlWpNaTK3w+H6677jpceOGFuPPOO1UtIRqBbisgjR3QKHamPU61ps4sWkh6kZt/EafEbpSwKi5NuubtOh/R6g48pPNOcryWIAhMrj9Xu0nj8TjefffdoXqcZDKJd999VzZ2KF6fTqcVr1PjxBNPxJNPPqn6g+fn5zvWpSAFi/1Gc6khuNz8V1dX484773RM31olWOu6c/zxxyu+PnbsWItG8h/MWPOk807iql+5ciX1/A0auNoyzLQKRZSsQ61WpBqzZs1CLBaTLAQPBAJIp9Ou6y3JWr9RK1PEWYC1+acNK+UXIj09PYZeNwO5NT927Fj09PTgvvvu0+wlIJ13JatUpK+vD6tWrcKiRYuYev5cqwyzrUIR0TqsqakZdqK9nBWZfZ0W5NxWSv78TNyU3GEnblcQucS0adPQ3Nw8YpMLSLsAzU7zV3tG7TrtJnPNi8kvXV1dusMxpK5XUeapbQL6+vqYK2tyrZtUyioUEa0+teulrtNCttsKGFSEsVgM9957r6pSZDH93cPDLgRBAM/zss91SUnJMBegFZmnTjguikaGJ6nrVZR5JLBW1uRay/Djjz+W7VOZTCbx8ccfD/2/VitSC1JWycaNG3HVVVchFovh7LPPlqwZCgQCmDZtmqMaLXt4mEksFsOBAwckX8vLywPHccOeCys6IrHSKEAJGklHWsINwWCQqAkGa54v1yrDG2+8kfhaEivSSOwwmwsuuADbt2/Ht7/9bXz22Wc47rjjkJ+fP/R6IBBASUkJeJ7HgQMHVF0bTuv44LTxmgFrc8DaeKRQEuqpVAptbW2YM2cO0fW0YvJ2Z1BLkf1bqkGqlLSEG8rLy1XbxrFgNWfiWjepFrRYkXqQ6oAzbtw4bNiwAaeccgqam5tx6NAhpNNp+Hw+VFVVgeO4YYpQJNu14bSOD04brxmwNgesjUcOrZaGFZYJaxnUUr+lGmYopXA4rOhCZsVqzsS1lqEWtFiRepCrXczLy8OSJUtw7rnn4vbbbx8q4P3a176Gr3zlK0S7WjubY+vBaeM1A9bmgLXxyKE1k9SqzFOWErTkfks5zArHBINB3HzzzVi1atWI34DVsibPMjQZktrFadOm4bXXXsMLL7yA0aNHo7GxUXXXKr5uRzsoIzhtvGbA2hywNh45tPQMVbve5/Nh2rRp1Mcohdm9VDNR+i2zyQzHmOEVKC4uxqJFi1BdXc2E1ayGZxmajJbaxcsvvxxVVVW4/vrr0dfXh2OOOUb2c8VdrdOC1E4brxmwNgesjUcOrfG5cDiM9vZ27Nu3b8RnpdNp8DyP6upqU4Wy1adakLhGfT7fUPJLMpnE5s2bTfMKsGQ1q+FZhiaitQMOMHiA8MaNG1FcXCxbepG5C3ZCancmThuvGbA2B6yNRw6t8blgMAiO45CXJy3mDhw4YHqXJ6sbV6v9lqFQCMuWLRvqRtTW1qboFWhpaaE6PpbxLEMT0doBRyQvLw933HEHHn74YRw8eHBYQ2W/3z9sF+yE1O5MtIzXCRmOemDtN2NtPEpotTTa2tpkG5db0eXJ6jM+tf6WJF4BVnuJ0sazDE1CrXZRrTdfMBjED37wA5xzzjnw+XxIp9M4dOgQ3nnnHUyePHlocTqt96aWM+uckOGoB9Z+M9bGQxO7XcBW31/rb0nSS9RpPZL14vojnOyiqakJ3d3dkvWLPp8PR44cweTJkxU/QzzOpqamBqeccgpWrVqFV199FWvXrkU6ncasWbNQUFBAfNxMNnYcRUR6PI4VR/Xoxei8aTkiyAqsGo8d662lpYX6cUZW31/LvGn9LQVBwJ49exQ/c8+ePXjrrbcsPV6OBlqPcPLJFZs7ia1bt6bLysrsHsYwHn/8cXzyySeyr5944omaSzqSySTuvvtu3HfffQAAjuPw+OOP6z4vraurC3ZuIpTcoCtXrlTcVYdCIdTW1lo42v9g97w5FTvmTe0Ir6qqKlPdpDTub+a8CYIwJE9IEC1MFrNBs9m6dSvKysp8pNfndMwwHo/jxRdfxGWXXaa73ZocZtQu+v1+3Hvvvbjgggtw0003YefOnTjnnHNw5ZVX4pe//CVOOOEE6vc0C7UsO7vdW27BTXFXPd/F7g4xdt9fjWAwqHomYyas1Z7SJKfdpE1NTdi5cycEQVB1WbLEpEmTsHDhQgQCAfz1r39FW1sbnnvuOcyaNQsnn3wy8efYeWK7mhs0Ly9P8cR2UveWIAiIRqNYv349mpqaqLh6jM6bGWOSu8/q1auxe/fuodPHBwYG0NnZiY6ODkydOtVSd5eRedP7Xex2SdO4v9nPqSAI6OzsVHzeMrH6xHq9mH7SPcdxEwB8DsA2nufJZo9BsovhjTTjtoOCggLcc889WLBgAa655hq8/fbbqKmpwT333IOlS5cO63XKImpZdoFAAIFAwFCGo9U1XiRYNSZBELB27VrJo3ScuLs30iXH7lo3u++vhpz1qoRdnhkzPR16skm3ANgM4F8cxzVyHHcbx3FlHMcR+2ZZQKoY3olMmjQJb7zxBpYsWYJEIoG7774bM2fORHt7u91DU0TNDZpIJAxnOFpd40WCFWMSFe7evXtlr2GpswwJTumSoxUru9PIIVW/qYYdtadmZ5jrUYbXAWgAIAC4CMADGFSQ+zmO+xPHcT/kOO4sQ6MyGT3F8CyTn5+Pn//853j99dcxadIkbNu2DWeeeSYWLlyIzz77zO7hSUJSHGy0ATKLAtSKMYkKVw0nxV3dGENmqXxItF5ra2uxbNkyVFdXa2p9ZwVmbyQ1u0l5nv8LgL8AAMdxXwZQA+AcANUALgHwtaOvHQDwJwAP8zzfZmiUlNFbDM865557Ltra2vCjH/0Ijz/+OFavXo3XX38dL7zwAioqKiTfY1eCBUlxsFb3kllH19DECqFO2p+Slc4yJFjVdNtKWG6QzmLij9kNDAwV3fM8/z7P87/mef5ynudPBHA6gBUAkgAKASwAsIXjuP81ch+aGC2GZ52SkhI8+uijeP/991FWVoZ//OMfqKqqwr333jviOw8MDNi2M6Vd6M3K0TVqWNH6jOT727W71+sW1Nqk2wmw6LkQYe1oKsD8jSTVDjQ8z+/keX4ZgO8BqAPwLQBvAPgxx3E/p3kvvZAc5KsXqXML7YLjOMRiMSxevBiJRAI/+tGPcN555+Gf//zn0DXbt2+3LaZG+2HTenQNMPhwWR2jsUKok3QVsWN3b2Tz5cYuOay7frNdp2I/U7tKcszeSOpShhzHFXIcJzsjPM//FsBFPM+/wPP8BQDmAfgOx3Hn6BsmPcw8yDfz3EIWKCgowIMPPohXX30VJ554It566y1MnjwZixcvRm9vL9rb223dmdJ82LQcXZOJ1TEaK4S6ksIFgAkTJtiyuzey+WLRUjEKDeHOQgKOVZi9kdRTWrEAwONH//svAJ4AsIHn+WxJNFQYw/P8nziO+yaAWwG8qXu0FDDrIF+WSzXOP/98bNu2DbfffjvWrl2LlStX4tVXX8UVV1yh+D67d6ZaIHWNSmFljEYU6mJss7+/f+g4HVpxWrV4z1VXXWWL8iDZfCnNP+slClox2iCdxdKh7PHRzEcwO46ppwPNjwA8jMH44EUALgVwgOO4VwG0AjgI4FwAw2oQeZ7fyHHcPYZGyzBazi20g5NOOgnPPPMMlixZgm9961vo6OjAoUOHiM5MdAJqCRbiLlzuGitOMBAxW6hboXD1cPjwYcXXrd582d2dx6hwZzkBh6aizv6dMmuQaf5mepThfp7nbwcAjuOOx2CpxVUArjj6DwD0YzCzFBzHLQLQxfP8OgDa/VgOQK5UgyXrUOSss87C1q1bceedd6KlpQXhcFiyQF9pZ2q3EJGCZJcdjUYVP8NJlrAaLFpRhYWFigoxnU6jrq5OcT3RWntahbUZa97opsXq46G0QEtRS/1OYlOO448/nqr1q0cZruE47lcAlvA83wOgHkA9x3ETMWgtHgNgE8/z/3f0+h8DKOE4bjyAf1IYM3M4rVQjFAph5cqVeOqpp9Da2orRo0cPU4hKO1NWXTMku2y1kgsnWcJO5PTTT8f27dtVY7ty64nm2tMirOXuG41GEY1GUVRUhIqKCl2K0cimheUEHFqK2krrV3MCDc/zj2Gw6L6B47ifZfx9D8/zf+Z5fn2GIgSAawD8GcBEDGaZugonl2qcf/75+MEPfoD9+/fj0KFDSKVSOHLkCL7whS/IChYWu7oAZAkWbkzPdxJTpkyRTB6SQmo90Vx7WsoaotEoPvvsM9nr+/v7bSmUt6JMRy+0FLWV5Se6Tq3gef41AK9xHPclgmuHivTdCEmpBovWochpp52GX//614hEIvj2t7+Nf/7znwgEAvjkk0+wePHiEa2ZWHbNqO2yWSwkziXy8/NHuAWVjpDLXk801x6psBYEAc3NzYrjFO9vdZyOJDRgV0iDVpMEK61fo0X3HbQG4lTMLNWwktmzZ2PHjh34/ve/j0QigVtvvRWXX3459u3bN+w6ll0zatBKzx8YGMiZdHbaZJfTqJG5nmiuPVKrKhaLqSpCEasL5dXKdCoqKmxrqkHLC2Ol9ZvT5xnSwKxSDTsoLi7Gww8/jNmzZ2PBggVYv3493njjDdTX1+O6665DXl6e49tiGU0sEQQBDQ0NiMfjTMVMnYqW9URz7ZGWNWhVblZuBtUScOzMNqXlhTFafqIFTxl6jOB//ud/MGXKFNx0001oamrCd77zHTzxxBP49a9/benitBoSl1IsFsPBgwdHeANYSGd3IlrWE821RyqstdavWr0ZVNrc2RnSoFXeY2Vow1OGHpKccsopePXVV/Hcc89hyZIliMViKC8vx4MPPogxY8ZQWZwslWiQZiq2trbKusXtjpk6ES3CjuRa0jVFKqzVrNFMWNsM2h3SoFHeY2XNbE6ddB+Px/H888/j1FNP9VxZUD9B2+fzYcqUKbjhhhvw73//G1u2bMGGDRswevRoXHjhhUNCSc/J4aydwh6NRrF79+4RO+lUKjUUWyktLUVTU5Pi5yQSCdTU1Jg2Ticjtd60nASvdi0ATWvK7/ejtLQU4XAYNTU1CIfDKC0tHXYN6SnwokKeN28e9XWr96T7lpYWDAwMyL4eCoUckTRG8jtJYfpJ904ms3coyxmerFFSUoJVq1ahuroaCxcuxPPPP4833ngDq1evxiWXXEL0GVJdJFKp1AghI7obo9Eo/H6/JqvRiKVJ6lJSsxT8fj/q6+tNsXRZsqRposWCULo2EolQj5HJWaNilnU6nWb2sa6ZnAAAIABJREFUd3BzSMMMcsYyjMfjaGhoQCqVwqefforp06cztXDtQOuO88wzz8T8+fPx7rvvor29Hc8//zy2bt2KadOm4fjjj5d9n5QVmEqlZLP0UqkUPv74Y3R2dhJbjUYtTVKLTxAEfPzxx4oZhqIlSdPSZc2S1oNeC4eU9evXD81NNqlUCj09PZotITlrdNasWbjyyitx3nnnEVsqetE7b+PHj0dHRwcEQRi26fT5fEin09izZw9aWlogCALGjx/P/PrRilbLkOoRTiwj1TvUQzunnXYa3nzzTdTX16O4uBivvPIKpk+fjjVr1sgqCD3HK6XTaU3F1UYLsklTuMPhMI499tgRaeOipSBn6RptRsBqswOWMCtGxtpRRqRklxIBg+tUVIaA9ae2sExOKEO53qEsd4dhmby8PNx222344IMPsGDBAhw+fBjXX389rr76ahw4cGDE9XqPV5JCrpbLaKcK0rqoYDCIuXPnjqhVVNpV06g/Y/kgWDuQOrpIrbMN62U/ZpCpyKurq+H3+03bsDmdnIgZOq13qFM44YQT8OSTT6Kmpga33HILnnvuOUSjUbzwwguoqqoaus7I8UpS9PX1IRKJDIudGbUKtGQ15ufnj4hb1dUpe2OMZu7ZnRnIEnKZv3l5ynv7QCCABx54AP39/czG+cxEa6mFW2PUcrjeMnRy71An4PP5cP3116OtrQ0zZszAxx9/jHA4jCuvvBL/+Mc/AJCdvC4SCARGtICTumd2Vw011KwCo91pzO6UwXIfSquRcxmLFo/c+jl48ODQpiEX3YNaNlTihsOO7jV24XplSNI7NJN4PI41a9Z4SlIjkydPxttvv42lS5ciGAzi+eefx1lnnYXf//73ii5In8+H/Pz8YcqnqqpK9npx96/F7UqaOWckNmR2E3Cvyfh/UHO7q22mRHLNPahlQ5WLMWrXK0OtvUMzyy/cgJXKPT8/Hz/72c+wa9cuzJ07FwcPHsQVV1yBJ554AqNHj5bsoTh27Fjcdtttw5RPdXW1bM/FdDpN3CtSfI8VTbjV+kQavb/Zn+8k1CwctZrATHIp3qplQ8VqjFoqVkyrL7DrY4ZaeoeKLtV0Os3s4bxasaO2cuLEifjDH/6A3/zmN7j11lvx7LPPorW1Fffccw8++ugj1S4SSl0n1A7oBQZ3wFaf7m52pwxWT6+3Ay1dYUjIlXirlri4HTFqtRil2Wepul4ZakGq/MLJCTZ2Knefz4cbb7wR55xzDr75zW/i3XffxdVXX42FCxdi6dKlePPNN1FRUSG7eOWKq9UO6A2FQqitraX6XUgx+3R5q0+vZzWBQq2YHNDmRs+VeKuWDZXVDflJFJ3Zjcdd7yYlxY3lFyzUVnIch1gshu99b/Bc58ceewy1tbXYs2ePrvF4sTNrYDmBQs1lXFlZSXSAsPieXFozpHFxq58zEkVntuvWtcpQa6xMqfzCibCk3EOhEH71q19h+/btOO+888BxHABgy5Yt+PTTTzV9lhc7swaWEyjUMn/lYs7ZeGtGHqufMxJFZ7br1rXKUEsijBvLL1hU7qeffjoWL16M/Px8AIOJDkuWLMHOnTuJP4PWAb0eyrCaQAGMdN9mu/qk1khRUREmTJjgrRlCrH7OSBSd2eVFrowZao2VkZRfOCl2qKbc7UoMisfjeO+994b+PxAIYOLEiaiqqsIPfvAD3HnnnSgoKFD9HKtjZyxhVRyP1SJ/0iSKXF4jtLByDklilGY3HnelMtSaCKO1/IJ1WFXuUuMKBAKYNWsW6urq8Nxzz+HRRx/F+eefP/Q6q0kcdqCkCFpbW5FOp6l1V7E6gYIUmkkU3tpiBxJFZ/ZBv65ThnKxMjlrKB6Po6CgAEuWLLGtjCIej+PFF1/EZZddRmUMdit3qe8jZ636fD7MnDkT3d3daGtrw1e/+lXccMMNeOSRR5CXl2dqKrXTUFIEmX+jMUesHv9D6/R2s9P0PcgRBAHJZFJSZmUqOrPLi1ynDLX2IWXhjEPaY9BSW2kGUt9HyVr1+Xyoq6tDe3s7li9fjtWrV2Pbtm246667TE2ldhpaGp4bnSOzd+F6oeW+VdpY7N+/H48++igOHjw49PcJEyZg/vz5jq87Zo3MTUm2fPD5fKisrER1dfWQojPTdeuqBBqtiTDZsUU7EmVYGANNMr9PW1sbfve736G3t1fVWu3q6sLSpUsRi8VQWlqKbdu2YcuWLcwmcdiB1kJzI3PEaqISrSQKpY1FMpkcpggBYO/evXjooYcc/3yyhtLxbn6/H36/37K15irLUGusjIUiexbGQJPM7yO6ZN966y1ia3X69OnYtm0bbrnlFlXBliudQ0T0dF4xMkcsJqHQct/q6WCTTCaxbt06XH/99Zrf6yENLbc3DVxlGWqJlbFQh8fCGIySWc8pZ5lr/U4lJSV49tlnvfPpslAqhJbDbXNEq/5Ny0kqmezdu1fX+zykYSlrmVnLkOO4OQCWA5gO4DCA9wHcx/P8Brn3aImVsXDGIQtjMEp2PaeUZZ5MJnV9p3A4jFgsJhtYz6XOIYB8HE8ON84RrSQKJQuTJQRBwNatW7F27VpXZryylLXMpDLkOO5SAH8C8GcAX8egBftDAI0cx13O8/w6I5/PQh0eC2MwSna8M51OSyouvb1R5YT/wMAAgsFgznUOkVMEAHDkyJFhc293ootWtJQ50HDfat1Y2IGYXLJ///6h39ZtGa8sZS2z6ia9F8AuAHN5nv/LUWtwLoD9AH5g9MO1nnFoBiyMwSjZ8UE5F7X4utbvlJ3EAQzGEWKxGJYvX44rr7wSu3bt0v8FHIhUb8lFixYhHA4zleiiBTv6oMolCB177LGK75swYQL1scghJpdkP1cstMSjBUvtFZmzDDmO8wH4CYAenucHxL/zPN/HcdxuACcbvYfddXisjMEI2Zat2hmDeq3DbCsgnU5j1apV2Lx5M1588UU0NDTg/vvvx+LFi4cO/s21YmoWE120YPZpBHJIzVtvby8eeughyWfT7/dj/vz51MchB0vJJWY9UywdTebTclCqnXAclw+gG8C7PM/PyXxt69at6bKyMnsG5mC6urowbtw4Xe9tbGxEW1vbMKHh9/tx3HHHDXPrZJKXl4dQKIQbb7yRWCHKNSTo7OzEsmXL8MQTTwAAZs+ejaeffhqf+9znRhRTA//ZadKwlozMWy4jN2/19fWqcaPbb7/dzKENo7e3F+vWrRuWLGNHnWFdXZ3i6z6fD8uWLTN9HFINCgC6z5QZbN26FWVlZT7S65mzDBWoA/BfAH4t9WJXV5e1o3EB8Xhc17z19fWNUITAoEXb09Mj+75UKoXe3l5s2LABZ599NtG93n77bezZs2fEe3w+H37yk5/g7LPPxu23345IJIIpU6agrq4O//73vyVdS/v378fGjRthdOOkd95yHbl5I8ko3LNnz1CD92wGBgawfft2tLe34/DhwygsLMTpp5+OKVOmyL5HjQsuuGDE3w4ePDii/tBMCgsLcfjwYdnXCwoKLFmHW7duldzg0nymWMARypDjuBsB3AFgDc/zL0ld4+3UtaPXwmlsbJR9ze/3Y9q0aSMyR+PxOB5++GEkEgns3r0bF198seouOx6PY/fu3QAg+57rr78el1xyCa655hps3LgRH330EY455hjJz0smk9i5cycuvfRSkq8pi2cZ6kNu3kjqJz/88ENJl6CU1XL48GFs374dH330kS1WCy2X4owZMxSTS2bMmGHJOly7dq1iSIfGM2UG3d3dmq5nNYFmCI7jlgF4DMBzABbaPBwP6It36jlomOQ98XgcjY2N+P3vf49ly5ap1o/lWqG+EyDJGJTrpMPauYs0k4HE5BK/3z/s71Ynl7BUC2gmTFuGHMetAnATgAcA3MHzvDMCnC5Ha+9Tkubp2bFB0obrYp3jpk2bUFdXh/vuu09R4FhRt5RrCTxGCYfDiEajitfICVyWkkwAuslAYnLJxo0bsXPnTtuSS1iqBTQTZi1DjuNWALgRwCKe55farQgzO614aIPkoOHs4n2S90j1dZ05c6ZslxYr6pbsKBNwOsFgUFWgyr3OmtVC+1DkYDCIsrKyYaU0s2fPtnRTpdT5yE2NHZhUhhzHzQVwFwatwYftHg8wUlh7kEHSPD1bqX3yySdEDdel3KhydUsDAwNIJpOYOXOmid+WPbedU6ioqNAlcM0+/VwrrClnGrBUCwgMbjgjkQjq6+tRV1eH+vp6RCIRwxtN5pQhx3EBAL8E8A8AEY7jyiX+sdTX5LaTJayEpLlAtlJ76aWXVN8j50YVBGFEoX4ymUQsFsOKFStw6aWX4u9//7sJ33QQ2paB2xEFW0tLi2yiiJLAZc1qYU0504ClE0zM9LywGDP8PIAvHP3vLTLXlAL4pyWjgftOlrAStWSbPXv2DOuyoVaeISboqPV1zS6mfuWVV9De3o7XX38dX/rSl/DDH/4QP/3pT3Wl3g8MDCASiUjGBNUsg76+PtTX17smhmgkPipXvyZC8lmsnbvIUnsxmrDS2MHMBg2OKbpXwsyi+8ySAJFAIIBFixYx3ztUDRZKBOSK96XKM0SkfhMRpd+mq6sLd911F5566ikAg4J07dq1mDhxIvF4BUHAY489hng8Lil8e3t7iVxhrBcsk6C1GDt7vUUiEUXFUVVVRXxqPQsdTMSx0C5QZ+E5tZPsDZcSoVAItbW1ALQX3fuXL19uaKAs0N3dvdysxdLU1ITu7u5hVojP58ORI0cwefJkU+5pFfF4HKNGjbL1/g0NDSMsx3Q6jU8//RTTp0+XFBxSv4mI0m8zatQozJs3D1/96lfxl7/8Be3t7XjyySdx+umng+M4ojFHo1F88MEHI8acSqUgCAJOPPFEHDp0CKlUSvFzxOsBoLS0lOjerBGNRrF79+4Rykzuu2Wvt/Xr1+PIkSOSn51KpbBnzx689dZbaGlpgSAIGD9+/IgyA2Bw81RaWopwOIyamhqEw2GUlpZKXms2fr8fU6dOBQD09PQgkUggFAqhsrIS8+bN06Wc7X5O7UTcXOzevVt2rWSSSCRQU1MDYLDOcNy4ccptfDJg0U3KDG44WYJltB7GLGK0r2tVVRXee+89fPvb38bLL7+MuXPnYtmyZbj77rtVzwtsbW2VvXcikcC+ffswZswYotMQtKb+s1ayYbSsgfSAXaed1MCKS1ErrK0vQN4tKoeRmKynDBXQK6w9yNCr1LTWOUoxduxYNDQ04IEHHsCdd96J//3f/8Urr7yC3/3udzjrrLNk30eSLbho0SJitw5pdqGU+81uJWH0u5F0nhExu2l3rsPi+gKUN1zZGI3JespQAaefLME6NJSaEXw+H5YuXYqysjJ85zvfwbZt21BRUYGlS5fixz/+MQoLC0e8h6QAOdMyIGlCTYJdJzsooTYXgUAAgiDIClGtB+zaUUQvwqLVRBMW1xdA7j2gkTDlKUMF7BbWHtYwZ84c/O1vf8Ndd92FRx99FCtWrEBDQwP+9Kc/jYjnlZeXIxaLSW6SpHamtLILWeu0Aqgrs0QigdWrV8taFXoO2NVTp2dUkbFqNdGExfUFkHkPaG1MmKsz9PCwg1GjRuGRRx7Bpk2bMHnyZPztb39DeXk5/vKXvwy7LhwO49hjjyUuQKZVsKzFJWlWUXI24ncTz5HMJp1OKzYakKpfU0NrTIhGXVouNFJgtVmAWh1pdXU1ta48njL08MggHA5jy5YtuPjii7F//35cfPHF+MlPfjIUOw4Gg5g7dy5xATKtgmXSYm4r28GJ301OGQLqjQZEl7LYbqy6uppqET0NRZYLjRRYbRZgZfcbz03q4UjkDv2lwejRo/Hyyy9jxYoVuOeee7Bs2TLs2LEDjz/+OMaMGYP8/HxN2YI0sgtJ3a1Wx36CwaCqi1OLVUG7iJ6G+89qqynbrVtYWIgZM2aYGp9ktVmAuOGyoo7Usww9HInZvWLz8vJw991345VXXkFxcTHWrVuHM888E2+88YYp91ODdIdshxVD06qg3fqLhiKz0mqSsuwPHz5seqN31vqPZpLtPTCrWbmnDD0ch5W9Yi+++GJs27YNM2fORGdnJ8477zzcfffdOHTokGn3lIJUSdgR+6HdH5Sm8KOhyKzsf2pXfJKl/qN24blJPRyH1b1iTzvtNGzatAkrVqzAT3/6UzzxxBPYvHkzXnnlFUyaNMm0+2ZD4m614+w51vqDZkLD/Wfl97Mzq1NufYkJWW4tKxHxLEMPRyF3WoXZJ4kEAgHcc8892LJlC0499VS8//77KCsrw2uvvWbqfbVixykOLFsVJO4/texbK78fa1mduXQ+p2cZejgKtdMqzGbatGloaGjA7bffjsbGRlx00UV4/PHH8Z3vfMf0e5Ngl5XGagsytQQMAEQ1hFZ9P9ZOlWe1GN8MvEbdOYzTGgDrbexNm4GBASxcuBCCIGDTpk3405/+hMOHD+Pcc88lqpUzEzMaRdPCrvWm1Mhba7NxsxEEAZ2dnZKN3gOBACorKy0dj1oz9Z6eHlvd4Ep4jboZwsz0fxr09fVhzZo1zI4vG5Z6xebl5eG+++7DF77wBdx88824//778cEHH+DZZ58d0cbN6lZerFppLMJa5xXW4q+suW3NxFOGJpKZ/s9iQ+9t27YxPb5sWOwVu3DhQpSWluKyyy7Diy++iO7ubrz00ks44YQTALijlZeb+3KyJuyl3LoFBQWm1xnKwZrb1kxcrQzttMyy0/9ZO+4pHo9j165dzI5PClZ7xc6ZMwdvv/02LrroIjQ3N6OiogIbNmzAGWec4fiYixuUuRIsCvtsy97Ow31ZLcY3A1dnk5pdmK127+z0f5ZgfXxGiMfjWLNmjekZppmcccYZaG1txaxZs/DRRx+hqqoKL7/8suNbebm9L6cd2bdOguVifNq4VhlaWZgtd2+r0/9JEccnBulZG59R7NoEnXDCCXj99dcxf/58xONxzJ07lzk3nFacrszVoCXsrWqObjUsl83QxrXK0E7LRyn9nwVYGh9tK87OTRAw6FZ74YUXsGLFCqTTadVONazHXJyuzNWgIezdXotnVTs0u3FlzFDOMrMiLpZ9bxErx6AEa+OjnWRkdXcaKXw+H+666y588YtfxGOPPYYZM2YgPz9/xHVOcMOxGFOjjdHsW9bjwm5OgKKJKy1DOy0fkvR/O2FpfLStONbc01//+tdx7733Ih6PY2BgYNhrTom5eDE1dVh2JbvdaqWJ65ShmuVjtmBkMf0/E5bGR9uVzZL7V6S8vBzf//73wfM8Dh06hHQ6jYKCAsfEXHIpgUIvLLuS3Z4ARRPXuUntLsxmNf1fJHN8dqZs03Zls+b+zeTUU0/F448/josuughbtmzBSSedhKamJuYVIWDteXJORBAEBAIBxTMd7XQls9ZUgGVcpwxZsnw85KHdY9TuTZAaxx13HJqamvDf//3fePPNN3HOOecgEongy1/+sm1jIsXraCON6IKUkzeA/a5klq1W1nCdMmTdMvMwx4pzwiZo1KhR2LBhA+bNm4dXX30V5557Lt544w1HKESPkYguSLlNWF5enu2u5FxIgKKF65ShB/uYYcU5ZRNUVFSEhoYGzJs3Dxs3bsR5552Ht99+G6eeeqrdQ/PQiJILEhhUhnbHhXOpg4xRPGXoYTksWXF2tOwrLCzEH//4R1x66aVoamrCnDlzsGnTJpx88slD1xhJh7frvbmGmgsymUzaPmdmN/5203rxye3QncTWrVvTZWVlttyb9ZMplLAqgYZkjv7whz9g+/btmDp1KubNm2f6mEQaGxuxdetWlJWVEVujtOatt7cX559/Pv7617/iy1/+MmKxGEaPHi3ZDxT4jwBTsjbsei8JdiZsmUF9fb2qQqShHIzOW6bCopkAZfZ6McrR55r4TDXXlVZYjd7WX3b0z7QLtTmKx+PYsWMHAGD79u2WzUl2neMnn3xi6W9SXFyMDRs24Etf+hLef/99zJ8/f0hw6U2Ht+u9uYhSDaYICzV9ch1kABhqIee29eIpQwMYKRq3qn+mXUpXvO8nn3yiOkdNTU3D6g2bmposGWN2neNLL71keU/TMWPGoLGxEZ/73Ofw2muv4aabbjJUxG3Xe3MRuRrMbBKJBPbt24doNGrRyNShUYyvtl7eeecdRxX1e8rQAHqLxq3sn2lX02rxvi+99JLiHGVahSJWWIdSdY49PT229DQtLS1FY2MjioqK8OSTTxpKh7frvYB7m1XLkd3XVIl0Oo1YLIYHHniAiTmhYdWprZeBgQFHdbnxlKFOjLT+sqqJuF1NqzPv29PTozhHmVahiBXWIUtt6YBBl9vTTz8NAIaae6sJZbPem6ttvzJdkCT09/cjFovZPic0vABq6wWAo9ylnjLUid7WX1b2z7Tr5A5SRSNlFYqYaR3K1TmK2NXT9LLLLsODDz6IlpaWEb1MRdTS4Y30EjXyXiviR6xbniTKARhcX/v379c8JzS/P41ifJKYqZPc654y1IGR/qdW9c+0q2m1FkUjZRWKmGkdKinrzPvb0dN08eLFOPPMM7F//35dzb2N9BJVioElEgm0tLTICl+z441OsDy11Owlk0lNc0L7+xvxAoiI60UNUbGyvpnxlKEO9LrYrGwiblfTai2KZvfu3YrX7dq1i+bQhlCqcxSx88SLBx98ED09PYjFYkOChPSMPSPn86nFwPr7+2WFr9ltv5yQuRgOh+HzEWfya5oT2t+fxmkk4npRsw6LioqIlbmdCtMruteB3qJxq/pn2tW0Ws0qzBzHxx9/jNGjRysKhNGjR9MeIoDh3WoaGxvR1tYmOWa7epr6/X4888wzOP/883H//fejvLwc0WiUuHWWkV6i4nsBSHYukTujz0jbL5LCbSc0nA4Gg6iqqsI777yDVCqler2WVmi0vz+tYnzxO6t1uSFR5uFweETdoqgwOzo6TK9b9JShDvS2/rKq84pdTauV7uv3+zFt2jRbm2VLwVI3nEwKCwvxhz/8ARUVFWhtbcWVV16Jl156SZPlYQStwldv2y+pwm0pAeiUhtPV1dXYtWsXenp6VK/V4lal/f1pnkZColhXrlxJ5Ea385BkTxlaCKkSNdrVxi4Bz6piUYLlnqZjx45FY2MjwuEw/vjHP+Kpp57CddddZ8m9tQpfEoG4b9++EZ9Dekq80YbTVrUNE5XM2rVrsXfvXtnrQqGQplZoZjTcpnUaCYliJVlPdlv/njJkkMzaQD2WlF0CXuq+Tm5XxwKnn346Hn74YVx77bX44Q9/iPPOO29YD1Oz0Cp89VoapALQSMNpEusTADVlGQwGcdVVV0m2KgMG5/bmm2/W9LmsN9xWU6wk68lu699LoGEMvbWBrLZ3s6vo301cffXVuPTSS3HgwAEsXLhQNUGJBnoSLOTafikJfTUB2NfXh/r6eiSTSZSUlOjKklWzPqPRKPVMVblEpurqaixatEjzxtBIljALkKwnGhmuRvAsQ8aQqg0ksQ6NWpNmkK3Y7Txt3sn4fD785je/wZe//GVs3LgRjz32GG6++WZT72n2aQciahYDMKiYNm/ejJKSElRWVqKtrU1TjEvN+tyyZQvS6bSqq1arq5Xmocg0Y3x2QLqe7LR+/cuXLzf1BlbQ3d293A3d8OPxOBoaGobibul0Gp9++immT5+uuNjF96VSKaLrM983atQoauPPpqmpCd3d3Uin0/D5fDhy5AgmT55s2v2swux5k6K4uBiTJk3C+vXr8dprr+GrX/0qPv/5z5t2P7/fj6lTpwIAenp6kEgkEAqFUFlZiXnz5ukSvlLzJggCOjs7VbMvU6kUBEHAxIkTMX/+fPh8PnR1deGDDz7Atm3bIAgCxo8fD7/fP+K9avWqqVRK9v6pVAo9PT2oqKjA6tWrsXv3bhw5cgTAYLuxzs5OdHR0YOrUqZL3poE4b36/H6WlpQiHw6ipqUE4HEZpaalp96UJyXoaP348Ojo6IAjCsN9DVJjz5s3T9F27u7sxbty4OtLrPcuQIZRqA5WsPb3WpJnIFf171qF+rrjiCmzatAmPPvooLr/8cnR0dBB3PdEDTctGDjmLQQqx8D/7erX0exLrUwmxhZqdmY5uQG092W39epYhI2RbhSJq1qFea1J8r1kWTqZVKOIW69AOy1Bkzpw5aGxsBM/zKCwsRE1NjS3j0IPUvGVbDHJt6EQSiQQEQRihlETLERhsfJ6JkvUZCASQl5enaJmGQiF0dXUNWYTZiNajWXE7O9dbNoIgIBqNYv369WhqakJLS4uiVa4VmtavVsvQNQk0LCaPaEFvVxu7Os0oYWWnHaeiN+EpGAziwQcfBAD87Gc/w4cffgiA/VZXSmQm3pBYulrbvqkln1RWVqomd9id6cgCTmiJZwTXKEOnZyzqqdFTUjptbW22KR3WToRgESNZtjU1NfjmN7+Jvr4+LF682FVCSi3rUA0ppaTWoq66ulo1U9PuTEerUNpUOaElnhFcEzM0K2PRqjo5PbWBSkonmUzaFjt0YvG9ldDIsv3lL3+JxsZGvPzyy1izZo1r4llqWYe9vb2KVpicUlKLV6nFqliv86OBWj1mb28v8y3xjOAaZQiYkzxidsmCEWWr1nB6z549pt5fDpa7urAAjYSnk046CXV1dbj11lvxj3/8A4WFhZLXOU1IqSVRxGIxU5SSmrK0qtTETtQsP7UEJ6e7il2lDGlnLFpRJ2dE2WYrncym036/HxMnTjT1/pl4nWbIoJlle8stt+DJJ59EQUGB4nVOE1JKiskupUSS6WhVyzezUKvHVMMsV7FV8+oqZQjQtQ7NLlmgqWz1CNm+vj5q92ex6J9F9JbPSJGfn4+VK1fiz3/+M4455hjZ69wSzwL0p9/TEKhKSpq04TjLkJSfBAIBS13FVs6raxJoRGhlLFpxOC7Nk+j1ZJVu27aNyv31tpDLNczIsj3nnHPQ2dn+cJWYAAAgAElEQVQpW5LglnhWJlrbvlmRYKQluYTVzF+1JKFQKGR5Szgrk3ZcpwwBOhmLZpcs0FS2eoRsPB7Hrl27qNyfplJ3M2Zk2fp8Plx11VXYv3//CIXopniWEawQqCQNxwG2yxNI+ofqPThaL6TzSgNXKkOjGYtW1MnRVLZ6hCyt+1thQbsFs7Js/3975x8lVXXt+U91N4iIxifoIGOMjuhx9CmroQFF7YgBFoqAQYmyCMJEHZ38eEYZoswKtAHzMjPMm7hc854QGYOjJqgo8fFb25A0gQQVkI4wHCExMUzrCj8fjaTTQPf8UVX9muLWvefeuuf+qNqftVgNVXVv7bpd3O/Z++wft912Gx9//DEbN26ko6MjkptUmojihmpah5jk8gSTZuBBmrGXgknf2rAomz3DhoaG0M5lezhuKZPonRJV/N5kP/nkE8cbQJBkjjD3wKIirmQfm1m2DQ0NDBs2jObmZvbu3VtW+4SlEkXBvOnIq7hn9rkRdzs0J4rtUebp0aNHeO8V2pnKCNt1cqWIrVOiit+b7PLly4s+50fIShF1E2yJVjkm+wwdOpS6ujree+89lixZYn2qRZqwMRi3ENM6xKR3sgm7H63tTNAwx5mJGDpgu04uqNiGkX3a2trKvn37ij7vR+xte9A2RCvNY6W8Fgff+c53+MpXvsKCBQu4//77Q101p5koCuZNSz6iEOakEDQTtLuAmjRvDwsRwxgIKrZhlHo0NTVRXV3NyZMnqaqqYvDgwYwbNy6QF2bTg7YlWmGXy0QZcvVaHEyaNIkrrriCDz/8kNdee4177rnHqj1pIYraRNMQYyV0sskTZNKHk4C6EebUFhHDlBBGsXbhOTo6OrrOEcQLs+lB26jxtDFWKqqQq8nioLq6mkcffZSHHnqIp556ylEM014YHoSo9sJMQoyV0MkmT5D90WIC6kTYiwcRw5QQRqJKsXM0NjayY8eOxIQObc1CDDvZJ8qQq+niYNq0aTz++ONs3ryZnTt3ctVVV3U9Vw6F4UGxMZsxyMIiiUkqtgiyP2oSGgU7i4eyLK1IG17jfMIo9XA7R3Nzc9c8tyTUCdqo8bRRLhNVfaWf8pXevXtz1113AfDCCy+c8lyS0/rTRin1glGXJ8RFkEkfJqUStsqGRAwTgNc4nzCKtb3OkRfDuOsESxWtYguLsAveo6yv9Ls4mDFjBgALFy7kyJEjXY9HWcBc7sjCwhuTIv5CTLrg2Fo8JFoMlVKDlFK7lFKdSqkr47bHBiatzMJIVPGacNGdOL3DUkWr2MIi7GSfqIYqB1kc3HDDDdTX13P48GEWLVrU9XjS0/rThCwsvDEp4i8kiICGRWL3DJVSXwf+J3AwbltsYrIXFEaiitM59uzZw9KlS63VCQahFNFy28MLM9kn3+A8iusWtHxl5syZNDU18ZOf/IRZs2YBlZXWb5P29nZZWBgQZH80zgSjRIqhUuqLwD8AXwcuBsJrL1MiYabS20oUMWXr1q1d4dFC4uoiU4po2Z4ykqd7g/NCwn7foIuDMWPGcPbZZ/P++++zZ88eBg4cWFFp/bbI7xV6IQuLLH4Tl+JMMEqkGAIHgBFa621KqSfiNqY7YabSx93K7NNPPy16U497Ir3fRUeUC4s///nPVjsUdSfo4qBXr17ccccdvPDCCzz//PPMnz/f16o7ySUYcdqW3yt0QxYWpWEj89eERIqh1vqDuG1wwub8wTxReof9+/fnyJEjXcOAa2trE9OezO+iI8qFxaRJkxgwYECo57TB1772NV544QWee+455s6da7zqTnIJRty2maT+l1u9YKWQSDEMQktLi/X3+NWvftUVVuzo6GD16tXceOONJZ+rkFLPbcKxY8dOG+G0bds2lFKhdnUIatu2bdvo7Ow0sin/eqeFhY3P1NraGsn3rVQuv/xyLrvsMn73u9/x/PPPc9tttwGglEIpdcpr9+/f3/X3LVu2cPDgwdOu54kTJzh48CDr1q1jyJAhvu0J47rZss0Uk9T/cePGnXI9SyUt37e0UzZiaHul3trayu7du08Rw927d3PbbbcF8uAOHTrkKoaHDh2y+plWrVrlGCLVWsfuHa5ateqUf3vZVPj6QsL+TC0tLanwDAG+9a1v8e1vf5vXX3+d+++/3+iYl156yTUMvGvXLsaPH+/bljCumy3bTPFKQurduzdf+MIXQn3PNH3fksQnn3zi6/VlI4a2CTsMZ7sZuBv5EG2hGMeZRVpom5+9P9tTRpKGn/3UKVOm8Mgjj/D2229z9OhRo99rkjMl47ZNkpD8keS950JEDA1Iwv5emNieNlEKQRYdcS4s4sDPfuoFF1zAsGHD2Lx5M2vWrGHy5Mme509yCUbctlVSb9FSiXt/1y+JLrpPCl7i8dZbb7m2U0saSfWkbLRMKzdMmjQUMmXKFACjkgCIt/DZi7htyychjRgxgt69e5PJZKy1B0s7aevSI56hB62trWzfvt1VPHbv3k1bW1tqhsXmPaliexH5lmZRT4JPsseaFILUUuabd7/55pt89NFHXHrppa6vT7L3kwTb4kr9TxtBplbESSLFUCl1CdAv98/83fpqpVT+ztystS7eCTdEmpqaOHHiBHV1dY43ndbWVp5++unETHwIg7gmwSfVY00KQWspzzvvPO68805eeuklXnzxRebMmeO5l5PUyQpJti0JJGmPLuj+blyfIZFiCDwBTC94bFm3v18K/MG2ESZ1hVF1PYmKOCfBB9n7i3K4btyUksQ1bdo0XnrpJZYuXcpjjz1mtJeTVO8nybbFSdL26ILs78b5GRK5Z6i1nqG1zrj8+UMUdniN6IlyckFURDWWKCy8Jn6UC6Xup44cOZKzzz6bnTt3snr16lTt5QhmJG2PLsj+bpyfIZFimARMhC6qyQVRkTZxD5JMklZKnebRs2dPxowZA8D27dtl4kIZkrRJGkGmVsT5GUQMi+AldOWY+Wgi7l6DiKMkbV6sCcWubxj7qdOnZ3ceiolqHpm4kE7irsEsJEjmbZyfIal7hrFiUldYbpmPprWUcSXXFBL3xA9bFLu+YdRS3nrrrVx00UUcO3aMs846q+jrZOJCOom7BtMJv/u7cX4GEUMH3ISuo6ODX/7yl2WX+Wgi7vX19bEl1xQS98QPG9hOXqqpqWHq1Kls3ryZ+vp6qqpODwxFXUeYpOxHvyTN9qi649j83HF2+BExdMBN6Do6Oti7d2/ZdT0xEfekZM6muSOQW/arzeubf9+JEyfywx/+kGuvvZYLLrgg1jrCpGU/+iEu292EKIoaTNufe+jQoUWTZDo7Oxk6dGjgc3shYuhAodDlawlPnDjRtbouN7zEPX8NkhCWTHOIulgY1HbYN/++559/Pv369WPhwoUsWbKEvXv3xlarZ5I5mNTyiThsNxGiUmswvbw+25/73Xff9Xze1ndCEmgMKMdEDb8kKXM2rSFqt+xXm9e38H0nTpxIe3s7W7ZsYdasWcydO5dZs2YxcuTISD2xpGU/+iEO202EKL9HF+T3mhfbTZs2de3b5cV28eLFtLe3W//c7733nuv/bZvfCfEMPSjXRA0/JC0smdYQdbEwqO3rW/i+Q4cO5ZlnnmHJkiU8+eST9OrVK9B5S907Slr2ox9s2e50Ta+88kr69etnvb2Zidja/p3F+Z0Qz9CDKDyiJJUrOFFqjZvgXsNp8/o6vW9LSwvDhw/nwIEDrFmzJtB5TbwIL7wGLic5q9WG7cWu6fbt21m8eLF1oTARW9u/szi/EyKGLkRVS5j0LippDUsmCbdFVZDra7qAKva+t99+OwCvvPKKn4/RRRidQtw6lGQyGWpra2lvb2f9+vUsWLCA733veyxYsID169cbia1NbEzPKHZNT548yaFDh4q+X55ShcJEbG1PDYlzKomESV2IIlEjzl6gpqQ1LJkUvBZVDz/8sO/fuUm9p9v7VldX06dPH9auXduVGOaHMEJ2N9xwAzt37mT//v2nPdfZ2cmuXbvQWnP48OHYs00Lw5dnnnkmPXv2pLOz85TrW0rmptc1rampoaamxlrZgUmNn+2M1Tinkohn6EIUHpEk55Q/YYdBTdvQub0vwIQJEzh8+HCgfo9hhOx69uyJUsqx3hHg0KFDHDhwIPZem07hy7/85S+0t7dzxhlncOaZZ4Yy19Drmp44ccJ3ezM/mHhlYc5zdPL6N27cyL333hvLvEjxDF2w7RFJck5lEPaiyrQe0et9Bw4cCMCPf/xjvvjFL/qywcSLKPSmevXqxbBhw05JsNm2bRsdHR2O5yj2OEQ7D88tJAwwYsSIUOzwuqa9e/e2Or7K1CsLY2qISZlI1GU1IoYxUo5dVITT6b6oWrVqFVu2bGHIkCGBfsd+FlBei7k9e/Ywb948Xn75ZZ566inOPfdcYzu8OoXU1taedrNra2s7LcTp5Q25EVW2aVRDak26r9gcXxXlrMgk1phKmDQmyrHRt+BOGFM2wsxuHjhwICNHjqStrY033njD17FeEwlOnjzJ/v37PUOcXtmDbkSVbRpVun+xa1pdXR1ZZ6BS6hT9kMQaUxHDmJByhcqj1P1hGwuou+++G4Bly5Z5vPJU3PaO7r33XjZv3lz0+939Zue2T1VVVUUmk3F8LsoeqlGl+xe7poMGDUp0a7ogBF1g2MwurtgwadwT0qVcobIIY3/YRnbzhAkTeOihh/j5z39OW1ubrwL8YiG79evXG4+Jctunyodtu2eT5p+LsodqlM2jna5pS0tLWQkhBJtOYbsvasWKYT41vbGxkcOHD0cuilKuUFmEsT9sYwF14YUXMmjQILZv385bb73F+PHjfZ+jEJMQV/5m57VPBUSyh+VGnOn+5UqQBYbtfcaKFMPuezfNzc2StCJYJax2a7YWUFOmTOnqchKGGJokxXS/2XklhdhKGDElysSSSiHIAsN2IlNFimHh3g0gJQ2CNZI+ZWP69OnMmTOHlStX8umnn9K/f/+SzucVAstkMrF6U0F6qtrM4qxEgiwwbCcyVZwYFlul54f2incohE3S94f79+/P2LFjWbFiBcuWLeOb3/xmSedzC4FVVVVx/fXXx+ZNJWWGYtIGA8eB3wVGkH1GP1ScGBZbpXd0dIh3KFghDfvDd999NytWrGD58uUli2GxEFh1dTXnnXce9fX1pZobmCTUtyVFkNOG7USmiiqtKOYV5sl7h6W+R5InUAj+qJTf59ixY6mqqmLDhg20traWdK4klwjYqG/zm+4fRpPzSsSrtrXU0HtFiaFXr8a8d1jKjS/pEygEfzQ1NfHHP/6RRYsWlbUg9u3bl+uuu47jx4/7LsB3wql4e8iQIbF7PGHvOwUZZZXEgvM0EGZfVCcqSgzd9m7ylFLwHkaHESE4YXtx+d8nwNGjR2lsbEyEXba49957AfjRj34UsyX2CLuAPoiXl+ahxnFjs0NORYnhgw8+SENDg2u2XCkJDbYnUKTlphoXYXvlhZGE5ubmwC3U0hAtmDJlCmeddRYbNmzg97//fdzmWCHseXlBvLw0DzUuZyougQbsJDREMYHCZIZdpRL2XEin/eXOzk4aGxsZNmxYbHbZ5JxzzmHChAn89Kc/5eWXX2b27NlxmxQ6YRfQB/HyouxoU07YzsCtKM/QJmE2UHZCQrDuhO2VF9tfbm5u9jVpIU3zKltbWxk8eDB9+vRh5cqVcZtjhbD3nYJ4ebYTQcqRIHuzfqlIzzBswuow4obpDLtKJCyvPN+vduzYsUWzjjs7O3nnnXe6ZgFGYZdfgvbdbWpq4rPPPuPmm29m9erVHDhwgL59+1qzMy7CLKAP4uVF2dHG1JtKet2j297sgQMHaGpqYtSoUSW9h3iGIWB7AkWxm6p4h1nC8srzYejXX3/dNet49+7dRtfedrTA7X397lF2Txaqra2ld+/eLF261JaJsRH21IOgXl4Uo5JMvakovK5Scdub7ejoYNOmTSXbKWIYArY7jMR1U00DYY016h6G3rdvn2fWsde1j2tepVM43STxqvt3rLq6mvr6ehYtWuQ5fSJN2Ljp2073b29vZ8uWLYHE2zTTNQ11j15bE52dnSXbKWHSELDZYSSKEGyaCavvZ6EY9OzZ0zXF3WuBY6sfqVcItNgepVvildN3rLa2lqamJj744AOuueaarseTHk5zw1b3GVt9S/PiffDgwa7fjZ9ONaaNrW03wA4Dr1ZsAL/+9a9L+h6KGCacpDd5jpswvHKnMPTx48eZOXOmo+C0tLQwYMAA63Y54ZZRXCyc3tnZ6ZrN6vQdq6mpob6+nldeeaVLDNPeRiwNN/3u5MW78HtkKt6mma5pqHusq6ujqanJ9TXHjx9n8eLFgb+HIoYJJ+lNnuMmDK88jFmDNuwqxKtMw+lznDx50jXxqljkIZPJUFtby9tvv82cOXPYuHFj0USRKPt6lkIabvrdKVW8TRtb226AXQr5SMS7775r9PpSvocihj4ImqVXCmlo8pxm0hSGdssoLvY5uouj02dyizxkMhn69u3LwoULaW1tLXpjhmR6VoUk+abvRKnibZrpmtS6R6dIhBelfA8lgcYHaekkIphjOxM4LLwyir367uYp/ExukYeamhqUUsY3o6R5VoWE3X3GNn5qGJ2yZE+ePMm5557rmema1LrHYnu8XgT9HooYGmKz6F3arMVHWsLQXhnFJn134XQRzbcorKurI5PJUFdXR0NDAw0NDQwfPpyqKvNbRNI8q0KSetMvhql4F8uS3bx5MwDDhw93zXS1nREbFLcwsRtBv4cSJjXEZtG7tFmLjzSEoU1CuYWfY9WqVWzbtq1o4wCnEGvhXuSYMWP4zW9+Y2RjEj2rQqIsdg+DfOu47tmkcLp4u2XJHj58mOrqambNmuX6XrYyYkvBT6enPKV8D0UMDbDZSSRNvSuFeAiSUezH4y220KuqqqKzs5NMJuNqXzHPqrAMo1evXgwbNixW4UniTb8YefFet24du3btKireacuSNcWknKI7pXr4IoYG2Mg2dDq3lEqEQxyJTjYJEso19Xi9FnqXXnope/bsoUePHo7H9+jRg+uvv96xvVdh8kNbW1tqyjCSQs+ePRkyZAjjx48v+pq0Zcma4pXYM2DAAPbv3x+ahy9i6IHNbMO4eleWO+UWds4Lmw2R91ro3XXXXcyePZvPfe5zpwhifhVeTNRsFbjbIs3NBNKWJWuK14SRqVOnhvq7kQQaD2xmG1ZymzVbSUPlPN0j7Gxmk5Zxffr04dixY2zcuLErZGqSXJGmae5p6M3pRtqyZE2JOrFHPMMcxVbdtrIN01TfZgNb3lv3BUZHRweLFi3iwQcfTP21tLG3bLoXmb8h7dixg5aWlqI33u6kKXSXNi+2kLBnNCaJKPd4y84zDOpxFFt151PPH3300a6bQE1NDTNnzqShoSFwNmJa6ttsYMt7K1xgdHR0cPToURobG0M5f5zYmItoutC77rrruPzyy9m3bx/r1683OndSprmbTKlIkxfrRFJLI9JG2XmGQTwOk1V32Ikuaalvs0FjY2PXzSfMpCG3gbyjRo1KrXdoa2/ZdCGXyWS45557mD9/PsuXL2f06NGex0TR1cRrn8+0l2qavNhipClLNqmUlRgGDSV5CZ2Nm1GU9W1Jyq5sbW3lt7/9bde/w7qxFws7Q/Z32tjYyB133BH4/Cbvb+sa28xmNmXcuHHMnz+ftWvXGpVb2A7dmQidafizXBNQKo3CxdHtt9/u6/iyCpMGCSWZDM5Ne6JLktrINTY2WrmWXu3ImpubA4djTecB2rjGQeYi2khOqqur4/zzz+ejjz5iy5Ytnq93Ct316tUrtNCdidCZhj/LNQGlknBKgvJL2Yhh0GnwXkIX15DWsEhSdmWhV5gnjGvp1Y6sFMH1Ejqb1zjI3rINYa6urmbq1KkAPPvss0bHFE5znzZtWmjT3E2EzjT8mbY2bcLpBO1j2p2yEcMg3puJ0KU90cVG4kVQnLzCPKXa5pToVEgQoTIROpvX2O/esk1hfuCBBwBYunQpbW1toZ03CCZCZ5rEIwko6SdoH9PulM2eYZAyBROhS3OiS9KK+nfv3l30ubCuZdjDkJ2Erra2tuv5INfYz/6i371lmx2NrrrqKgYPHszWrVtZu3at1T1YL0z2+fwk8UgCSroJGhrtTtmIYZAboInQpaGRczGSkHiRp7W1lePHjzs+V1NTw8MPPxyKQIe5eCkmdEqprtcEuca2aiyjWPxMnjyZrVu3smLFiljF0EToyrn+TjgVv31MnSgbMQxyA0yz0HlhUtQfJWF7bMUI83daTOi2bt3KwIEDAzVOsNmYPYrFz6233srs2bNZs2YNHR0dvkY8hYmJ0KVtSoUQHLfFkSllI4YzZ86MvWwgSZiIT/dwn23SFm52E7oPP/zQeD+5UIRshTGj6mh07bXXcvHFF/Pxxx+zfv16vvSlL5V8ziCYCp2EPyuDYosjP5SNGJZLU+awMBGfKMUwbV64jf1km2HMqDzvTCbDjBkzmDdvHs8++2xsYggidMK/4rQ48kvZiGEl9PP0g4n4tLS0RGBJOnETuo6Ojq795NbWVp5++mlOnDjhufdpM4wZped93333MW/ePN544w2OHDnCOeecE9q5BSEohYsjk3rY7pSNGMosQCFM3BYTLS0tDBgwADAPe9oOY7rZm89ezU+h8INT5uvFF1/MTTfdxIYNG1ixYkVX/aEgpJmyqTNMSxG8UD74afQQZ71qKUX4xY698847AVizZk0oNgrBMGlEngTSYGfZiCHEX1QuVBZ+Gj3ElUBUShG+27Fjx44FYO3atUVLZsqduG/waZnDmBY7yyZMCvEXlcdNkhpylzt+w55xJRCVkr3qduwVV1zB1VdfzY4dO1i5ciVf/vKX7XyAhGI6EcMmaZnDmBY7y8ozhMr2DpPUkLvcSUObvqD9ek2OzWQy3HfffQAsXrzY0idILiY3eNukZQ5jWuwsOzFMYs1aFCSpIXclkIa6yVKmrZgcO23aNGpqali3bh379u0Lx+iUkIQbfFrmMKbFzrIJkzY0NMRtQqzY7EkpnI7NsGcY4e5SsldNj+3Xrx+jR49mzZo1/OxnP+tq5F0JxHGDL5zX50VS5jCmZV5kYj1DpdQgpdQqpdS/KKWOKaV+qZSKtodYjPiZSVdKOExIHmGEu0sJ4/o5dsKECQCsW7cusK1pxHQiRlj4ndeXpDmMtuZFhp3AlEgxVEpdBjQB/YCpwO3AvwBvKqWGx2lbVPi5IaZ9+LDwr4QV7i4ljOvn2HxW6bp16/jss88C2ZpGoh4I7GdeX9IakduYF2kjQzWpYdI5ZG0bp7XeD6CU2gjsBr4PjIrRNuv4aeYcVU9KIRrCCneXEsb1c+wll1zCiBEj2LRpE6+++iozZswI/L5pIuqJGCbz+jKZTCIbkdtomG6SwOS3M1LixFAplQHuAN7KCyGA1vqvSqnXgL9TSp2rtT4cm5GW8XNDjKonpWCfpM2fNOW+++5j06ZNPPfccxUjhlFPxPAKjWYyGebOnRvqe4ZJ2H1kTRKYbrnlFl/nTJwYAhcDnwM+cHhuB9nQ7jXAhiiNigq/N8Q0ZDUKZiRp/qQfJk+ezDe+8Q02bNjAn/70Jz7/+c/HbVIkRNkoPC1JKFFhI4EpU8yriAul1FDgHeARrfVTBc/dASwH7tJav5Z/fMuWLcn6EIIgCELsDBkyJGP62iR6hr1yP//q8Fx+V/SUZZCfDywIgiAIhSQxmzTv3zoF3c/I/fTOLRYEQRAEQ5Iohp/mfp7v8Ny/yf38JCJbBEEQhAogcWKotd4L7AeudXj6WuA48NtIjRIEQRDKmiTuGQIsA2YopfprrT8FUEqdBdwJrNZaH809Ngj4e+BGoAfwLjBXay3V5i7krtvLgAL+vdZ6V8wmpQKl1CjgCWAw0EY2u/kHWuvVcdqVdJRS44HvAH9LdvtjO/A/tNavx2pYilBK1QO/AJq01jfHa01yUUr9AijWqez7WuvvFjs2cZ5hjieBz4AVSqlxSqnRwBvAWcB/AelSExSl1NeBzYC/itQKJ3dDfws4AkwCvkpWEFcppSbHaVuSUUp9Ffhn4A/AV4C7yUZ3XlNK3R2jaalBKXUG8CNAEgXN2AoMdfjzT24HJdIz1Fr/P6XUTcB/B35KVrR/Ddystd6Ze1lFd6kJQq636z8AXydbz1nZ3c398ffAh8BErfVx6FqF/gn4O+DV+ExLNE8CG7TW0/IPKKWayF63B8lGKAR3vgv8DZCMWUfJp1Vr7ftaJVIMAbTW/xcY7/ScdKkJzAFghNZ6m1LqibiNSQu579t8YF9eCAG01seUUruByqgy94lSqhewgIIGGlrrI0qpXcAXYjEsRSil/hZ4jOzCYXrM5pQ1iRVDDyq6S01QtNZO10vwQGvdCbxS+LhSqgcwEHg/cqNSgNa6DfjHwsdz1+1iYFvkRqUIpVQV2fDoJq31j5VSIoYWSasYXpD7ud/huf0FrxEEW3wP6IvHXoSQRSlVDVwG/IBsc43kNtNMBv+JbLLWoLgNSRn9lFJLgFvIluPtAf5Ja33awqw7SU2g8cJ3lxpBCBOl1IPA48ASyYr0Rik1AzgBaODfAaOC7OtUCkqpi8guGv6r1lrHbU/KuBT4IzCFbAXCR8D/Ukr9Z7eD0iqG0qVGiA2l1FxgIfAToHLGu5fGPwNDyGZ97wI25gRScOYfgRaygiiYMwm4UGvdoLXeqLVeSTb3ZDMwTyl1drED0xomlS41QiwopZ4BHiKb6fx4bj9R8EBrfRA4mPvnKqXUi8AzSqk3tNaHYjQtcSil7iR7A78d6JHbYwWozj3fB2jXWgcb6V7G5L5nhY91KqXeAIYDVwO/cTo2lZ6hdKkR4kAp9X2yWX0Pa60fEyF0Ryl1oVLqfqXUVQ5PbyW73XFFxGalgfFkawpXAa3d/tyY+9NKrt5aOBWlVKb3xrsAAANeSURBVJVSysnJy2+btRU7NpVimGMZMFop1T//gFOXGkEIA6XURLI3oMe11k/HbU9KOAN4Fpjt8Nz1uZ8fR2dOavg+cJPDn/dzf24CnovNuoSSa8TyFwpCy7nErTvIlpbtKHZ84uYZmqKU+rdk2zp9RLZFVjvZepzrgGHdivOFHEqpS8h27AH4j2T3u+4iu9kM0Cyhl9PJrTQ12dX63YDTfxq5dg4opf4PMI1sxu3Pcg9PIhtqXqK1/g9x2ZY2ck0ekHZsxVFKLSV7T1tAtmNUH+AbwBjgfq31/y52bFr3DE271Ain8gSnF+4u6/b3S8m2zRJO5SKyGZCQHTzthFw7Z75GdtE6Pff3vwK/J7tw/WGMdgnlyXSyIfgHgJlkv2/bgAla6xVuB6bWMxQEQRCEsEjznqEgCIIghIKIoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFY+IoSAIglDxiBgKgiAIFU9qh/sKgpBFKTUO+DZwBtn/048A7wH/DRiWe9kSrfVz8VgoCMlHPENBSDFKqenARGCi1roeeBVYCzwDvAOMBo4Ci5VS58VmqCAkHBFDQUgpSqkBwJ3Ag1rrY7mHdwDnAhdorV8BrgRuBX4HHInFUEFIASKGgpBevgos0Fp3dnvs6tzPpQBa6+3ANUCd1vpExPYJQmqQPUNBSC+FQggwEugEfp5/QGv9QaRWCUIKyXR2Fv5fEgQhjSilegAHgD9ora+N2x5BSBMSJhWE8uF64Gzg7cInlFKZ6M0RhPQgYigIKUUpdYFSamC3h0bnfv6i4HUPAJOisksQ0ojsGQpCClFK/Q3wAdBXKdUPOAxMzj29o9vrzgamAmMiN1IQUoR4hoKQTi4BzgdeJCuE3wXeBE6QzR4lJ5IvA9/VWrfHY6YgpANJoBGElKKUmgvcQjbCs0pr/QOl1HjgCeCvQBswV2v9q/isFIR0IGIoCIIgVDwSJhUEQRAqHhFDQRAEoeIRMRQEQRAqHhFDQRAEoeIRMRQEQRAqHhFDQRAEoeIRMRQEQRAqHhFDQRAEoeIRMRQEQRAqHhFDQRAEoeL5/843Md+F1l1IAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8n_JXPgGHT3W"
      },
      "source": [
        "## 신경망 미분하기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1E4tqFvXHRgL"
      },
      "source": [
        "# pytorch를 임포트 한다.\n",
        "import torch"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5It-5ppAHXEo"
      },
      "source": [
        "### 순전파 함수(넘파이)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JPDltI-GHVf9"
      },
      "source": [
        "def forward(X, W, T, retopt='all'):\n",
        "    \"\"\"\n",
        "    네트워크를 피드포워드 시킨다. numpy 버전\n",
        "    X : 네트워크의 입력벡터 shape:(N,2)\n",
        "    retopt : 네트워크가 순전파되면서 각 레이어에서 계산된 결과 값을 \n",
        "    되돌릴 방법을 설정한다.\n",
        "        - 'all'  : 모든 층에서 계산된 결과를 튜플 형태로 되돌린다.\n",
        "        - 'fval' : 함수의 최종 출력값만 되돌린다.\n",
        "    \"\"\"\n",
        "    N = X.shape[0]\n",
        "    \n",
        "    H1 = np.dot(W[:2,1:], X.T)\n",
        "    Z1 = H1 + W[:2,0].reshape(-1,1)\n",
        "    A1 = logistic(Z1)\n",
        "    \n",
        "    H2 = np.dot(W[2,1:], A1)\n",
        "    Z2 = H2 + W[2,0]\n",
        "    A2 = logistic(Z2)\n",
        "\n",
        "    C = (1/(2*N)) * ((T-A2)**2).sum()\n",
        "    \n",
        "    if retopt == 'all':\n",
        "        return (H1, Z1, A1, H2, Z2, A2, C)\n",
        "    elif retopt == 'fval':\n",
        "        return C"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EA6Ph-I8HZ1p"
      },
      "source": [
        "### 순전파 함수(파이토치)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "losYdj2pHYjT"
      },
      "source": [
        "def forward_torch(X, W, T, retopt='all'):\n",
        "    \"\"\"\n",
        "    네트워크를 피드포워드 시킨다. pytorch 버전\n",
        "    X : 네트워크의 입력벡터 size:(N,2)\n",
        "    retopt : 네트워크가 순전파되면서 각 레이어에서 계산된 결과 값을 \n",
        "    되돌릴 방법을 설정한다.\n",
        "        - 'all'  : 모든 층에서 계산된 결과를 튜플 형태로 되돌린다.\n",
        "        - 'fval' : 함수의 최종 출력값만 되돌린다.\n",
        "    \"\"\"\n",
        "    N = X.size()[0]\n",
        "    T = torch.tensor(T, dtype=torch.double)\n",
        "    \n",
        "    # 계산 결과 검증을 위해 pytorch를 사용하므로 numpy 어레이 뿐 아니라\n",
        "    # pytorch tensor형태에 대해서도 동일한 연산을 한다.\n",
        "    H1 = torch.mm(W[:2,1:], torch.t(X))\n",
        "    Z1 = H1 + W[:2,0].view(-1,1)\n",
        "    A1 = torch.sigmoid(Z1) #(2,1)\n",
        "    \n",
        "    H2 = torch.mm(W[2:,1:], A1)\n",
        "    Z2 = H2 + W[2,0]\n",
        "    A2 = torch.sigmoid(Z2)\n",
        "\n",
        "    C = (1/(2*N)) * ((T-A2)**2).sum()\n",
        "    \n",
        "    if retopt == 'all':\n",
        "        return (H1, Z1, A1, H2, Z2, A2, C)\n",
        "    elif retopt == 'fval':\n",
        "        return C"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kxg77ow5HcXM"
      },
      "source": [
        "### 넘파이 어레이와 텐서 프린트 보조함수"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DTeXp1ipHa-j"
      },
      "source": [
        "np.set_printoptions(precision=4, linewidth =150)\n",
        "\n",
        "def print_tensor(t):\n",
        "    \"\"\"\n",
        "    텐서형 자료를 보기 좋게 프린트하기 위한 보조 함수\n",
        "    \"\"\"\n",
        "    def namestr(obj, namespace):\n",
        "        return [name for name in namespace if namespace[name] is obj]\n",
        "    \n",
        "    var_name = namestr(t, globals())[0]\n",
        "    \n",
        "    print(\"{}:{},{}\".format(var_name, t.shape, t.dtype))\n",
        "    print(t)\n",
        "    print(\"-------------------------------------------\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Rq6o7ilTHe9c"
      },
      "source": [
        "### 가중치 무작위 초기화"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JlJCwpeZHdrz",
        "outputId": "d9be6666-9d7d-4619-b71a-ae740e0e3665",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 187
        }
      },
      "source": [
        "np.random.seed(17)\n",
        "\n",
        "W = np.random.randn(9).reshape(3,3)\n",
        "W_torch = torch.tensor(W, dtype=torch.double); W_torch.requires_grad=True\n",
        "\n",
        "print_tensor(W)\n",
        "print_tensor(W_torch)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "W:(3, 3),float64\n",
            "[[ 0.2763 -1.8546  0.6239]\n",
            " [ 1.1453  1.0372  1.8866]\n",
            " [-0.1117 -0.3621  0.1487]]\n",
            "-------------------------------------------\n",
            "W_torch:torch.Size([3, 3]),torch.float64\n",
            "tensor([[ 0.2763, -1.8546,  0.6239],\n",
            "        [ 1.1453,  1.0372,  1.8866],\n",
            "        [-0.1117, -0.3621,  0.1487]], dtype=torch.float64, requires_grad=True)\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0miWkfH4HiTx"
      },
      "source": [
        "### 초기 상태에서 목적함숫값"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5UBYQGC2HgUc",
        "outputId": "f8af7638-3ee9-4562-f361-dec8665372be",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# 초기 상태에서 목적함숫값\n",
        "J(W, samples, target)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.1263148192185165"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cKTCxmTGHjkE",
        "outputId": "21afbb70-ee37-4abb-f576-f2cb4909cc5d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "H1, Z1, A1, H2, Z2, A2, C = forward(samples, W, target)\n",
        "\n",
        "C"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.1263148192185165"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EycQWx8-Hmz6"
      },
      "source": [
        "### 데이터 하나로 순전파 시키기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0KHDxSXBHkfj",
        "outputId": "73eef01a-d0c6-4583-fa13-d69a3094bcb8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        }
      },
      "source": [
        "N = 1\n",
        "x = samples[[0]]\n",
        "x_torch = torch.tensor(x, dtype=torch.double); x_torch.requires_grad=True\n",
        "t = target[[0]]\n",
        "\n",
        "print_tensor(x)\n",
        "print_tensor(x_torch)\n",
        "print_tensor(t)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "x:(1, 2),float64\n",
            "[[2.754  3.5407]]\n",
            "-------------------------------------------\n",
            "x_torch:torch.Size([1, 2]),torch.float64\n",
            "tensor([[2.7540, 3.5407]], dtype=torch.float64, requires_grad=True)\n",
            "-------------------------------------------\n",
            "t:(1,),float64\n",
            "[1.]\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "daU7I4_7HoKz",
        "outputId": "a869ef58-e198-4c80-f9f5-01151df654de",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        }
      },
      "source": [
        "H1, Z1, A1, H2, Z2, A2, C = forward(x, W, t)\n",
        "\n",
        "print_tensor(H1)\n",
        "print_tensor(Z1)\n",
        "print_tensor(A1)\n",
        "print_tensor(H2)\n",
        "print_tensor(Z2)\n",
        "print_tensor(A2)\n",
        "print_tensor(C)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "H1:(2, 1),float64\n",
            "[[-2.8986]\n",
            " [ 9.5365]]\n",
            "-------------------------------------------\n",
            "Z1:(2, 1),float64\n",
            "[[-2.6223]\n",
            " [10.6818]]\n",
            "-------------------------------------------\n",
            "A1:(2, 1),float64\n",
            "[[0.0677]\n",
            " [1.    ]]\n",
            "-------------------------------------------\n",
            "H2:(1,),float64\n",
            "[0.1242]\n",
            "-------------------------------------------\n",
            "Z2:(1,),float64\n",
            "[0.0125]\n",
            "-------------------------------------------\n",
            "A2:(1,),float64\n",
            "[0.5031]\n",
            "-------------------------------------------\n",
            "C:(),float64\n",
            "0.12344827837124746\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xTjK0X91HpzR",
        "outputId": "1421e235-8571-48d2-9c70-ba4f08f04f72",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        }
      },
      "source": [
        "H1_torch, Z1_torch, A1_torch, H2_torch, Z2_torch, A2_torch, C_torch = forward_torch(x_torch, W_torch, t)\n",
        "\n",
        "print_tensor(H1_torch)\n",
        "print_tensor(Z1_torch)\n",
        "print_tensor(A1_torch)\n",
        "print_tensor(H2_torch)\n",
        "print_tensor(Z2_torch)\n",
        "print_tensor(A2_torch)\n",
        "print_tensor(C_torch)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "H1_torch:torch.Size([2, 1]),torch.float64\n",
            "tensor([[-2.8986],\n",
            "        [ 9.5365]], dtype=torch.float64, grad_fn=<MmBackward>)\n",
            "-------------------------------------------\n",
            "Z1_torch:torch.Size([2, 1]),torch.float64\n",
            "tensor([[-2.6223],\n",
            "        [10.6818]], dtype=torch.float64, grad_fn=<AddBackward0>)\n",
            "-------------------------------------------\n",
            "A1_torch:torch.Size([2, 1]),torch.float64\n",
            "tensor([[0.0677],\n",
            "        [1.0000]], dtype=torch.float64, grad_fn=<SigmoidBackward>)\n",
            "-------------------------------------------\n",
            "H2_torch:torch.Size([1, 1]),torch.float64\n",
            "tensor([[0.1242]], dtype=torch.float64, grad_fn=<MmBackward>)\n",
            "-------------------------------------------\n",
            "Z2_torch:torch.Size([1, 1]),torch.float64\n",
            "tensor([[0.0125]], dtype=torch.float64, grad_fn=<AddBackward0>)\n",
            "-------------------------------------------\n",
            "A2_torch:torch.Size([1, 1]),torch.float64\n",
            "tensor([[0.5031]], dtype=torch.float64, grad_fn=<SigmoidBackward>)\n",
            "-------------------------------------------\n",
            "C_torch:torch.Size([]),torch.float64\n",
            "tensor(0.1234, dtype=torch.float64, grad_fn=<MulBackward0>)\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Rma1f-n1HteM"
      },
      "source": [
        "### 단계별 계산"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T_9W4uUPHvey"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{a}^{(2)}}$: (1,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AmhH32FMHr50",
        "outputId": "317903bf-5b66-40b5-d668-85114ac6a261",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dA2 = -(t-A2)/N\n",
        "dA2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.4969])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "941UXpSIHyqm",
        "outputId": "792f29ee-054e-4e04-c648-28a674d41a59",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dA2_torch = torch.autograd.grad(C_torch, A2_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dA2_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[-0.4969]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wqa2H9mLH1mo"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{z}^{(2)}}$: (1,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Mk0iumwHz58",
        "outputId": "568cba60-d4b9-4e55-9b6e-700976b871ad",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dA2_dZ2 = logistic(Z2)*(1-logistic(Z2))\n",
        "dZ2 = dA2_dZ2 * dA2\n",
        "dZ2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.1242])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c3y0MsLvH2zs",
        "outputId": "fb959464-bebb-47b0-967a-d0962e5e937a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dZ2_torch = torch.autograd.grad(C_torch, Z2_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dZ2_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[-0.1242]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o3963sheH5aQ"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{b}^{(2)}}$: (1,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l1IUyGLpH33E",
        "outputId": "a2895a57-d286-4410-a5ce-945881c1e752",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "db2 = dZ2\n",
        "db2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.1242])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vnNIgUPOH6lc",
        "outputId": "0833a75a-0f26-4e53-97f0-40af91f9c1b1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dW_torch = torch.autograd.grad(C_torch, W_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dW_torch[2,0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor(-0.1242, dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9FoZpRU7H9Th"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{h}^{(2)}}$: (1,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JyFWXbr7H7vU"
      },
      "source": [
        "dH2 = dZ2\n",
        "dH2_torch = dZ2_torch"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "le3kHspFH_d-"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{W}^{(2)}}$: (1,2)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eN6BMfnFH-e0",
        "outputId": "aaf4d56c-632d-49fc-85f9-78039cd6f462",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2.reshape(-1,1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[-0.1242]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3ypEuShnIBEE",
        "outputId": "c4623e7f-96ca-4b5d-dc67-75a78b431a7e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2_dW2 = A1.T # dH2/dW2는 야코비안이 아니라 일반화 야코비안으로 (1, (1,2))가 되야 함\n",
        "\n",
        "dW2 = np.dot(dH2_dW2.T, dH2.reshape(-1,1)) # 야코비안 전치 곱하기 그레디언트\n",
        "dW2.T #여기서 결과를 다시 전치시켜야 (1,2)가 되는 이유는 일반화 야코비안과 관련있고 뒤에 다시 설명해야함."
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[-0.0084, -0.1242]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jufRLC3YICds",
        "outputId": "0c07db21-6275-4f1c-81de-49b79920cee3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dW_torch[2,1:]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([-0.0084, -0.1242], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L83KeiwlIFRp"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{a}^{(1)}}$: (2,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2BF3spM2IDpC",
        "outputId": "84771415-1881-4303-8378-45591faa768d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2_dA1 = W[2:,1:] # dH2/dA1은 (1,2)인 야코비안이 맞고\n",
        "dH2_dA1.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1, 2)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y2fl4DQiIGb8",
        "outputId": "6f19a774-1520-413c-a529-94fa64e025cf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dA1 = np.dot(dH2_dA1.T, dH2.reshape(-1,1)) # 따라서 야코비안 전치 그래디언트가 성립, W.T * delta\n",
        "\n",
        "dA1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.045 ],\n",
              "       [-0.0185]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZYK-xbirIHfd",
        "outputId": "e9468364-9267-4349-bdb5-e9478c834a43",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dA1_torch = torch.autograd.grad(C_torch, A1_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dA1_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 0.0450],\n",
              "        [-0.0185]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8aH_iRkyIMaQ"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{z}^{(1)}}$: (2,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4F8t4HgDIKPc",
        "outputId": "7920717d-cfca-409a-a5a0-ea4954e4b76a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dA1_dZ1 =  np.zeros((A1.shape[0], Z1.shape[0]))\n",
        "dA1_dZ1[np.diag_indices(Z1.shape[0])] = (logistic(Z1)*(1-logistic(Z1))).reshape(-1)\n",
        "\n",
        "dA1_dZ1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[6.3132e-02, 0.0000e+00],\n",
              "       [0.0000e+00, 2.2958e-05]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CEa7_RnWINxU",
        "outputId": "0be81433-4b91-4968-ba29-0899a539ab35",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dZ1 = np.dot(dA1_dZ1.T, dA1)\n",
        "dZ1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 2.8396e-03],\n",
              "       [-4.2398e-07]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 47
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oFHZDWanIOiM",
        "outputId": "8ad03839-01fb-4fd1-f245-61c51b2e366a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dZ1_torch = torch.autograd.grad(C_torch, Z1_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dZ1_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 2.8396e-03],\n",
              "        [-4.2398e-07]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7Jh6ViFNIQ-I"
      },
      "source": [
        "- 굳이 이렇게 어렵게 하지말고 그냥 엘리먼트 와이즈(아다마르 곱)로 처리하는 것이 더 간편하다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MQiZ3hhdIPac",
        "outputId": "b997ca57-0ee9-4303-ced3-1c05651923e3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dA1_dZ1 = logistic(Z1)*(1-logistic(Z1))\n",
        "\n",
        "dA1_dZ1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[6.3132e-02],\n",
              "       [2.2958e-05]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l0zHYc2TISP0",
        "outputId": "a6f0f335-ed04-4869-86ad-f01537903192",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dZ1_ = dA1_dZ1 * dA1\n",
        "dZ1_"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 2.8396e-03],\n",
              "       [-4.2398e-07]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X_F95ZyTIVAE"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{b}^{(1)}}$: (2,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jax_e84KITEc",
        "outputId": "d8818d59-d1a5-4662-8ff7-33ffa9a78279",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "db1 = dZ1\n",
        "db1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 2.8396e-03],\n",
              "       [-4.2398e-07]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wd6Z_wAUIWGM",
        "outputId": "e51158a4-a982-4b30-bde5-a3edbf1b67b1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dW_torch[:2,0].view(-1,1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 2.8396e-03],\n",
              "        [-4.2398e-07]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r-xj8sqkIZ--"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{h}^{(1)}}$: (2,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hkcCP2BVIXCU",
        "outputId": "a71662c0-0b1b-49b3-d57a-f1012f8cc63f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dH1 = dZ1\n",
        "dH1_torch = dZ1_torch\n",
        "\n",
        "dH1_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 2.8396e-03],\n",
              "        [-4.2398e-07]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 53
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qzTATAXLIfpJ",
        "outputId": "c32ba038-d0fd-4fc3-e4c8-192658070b2e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dH1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 2.8396e-03],\n",
              "       [-4.2398e-07]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6AiXQS7FIdnm"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{W}^{(1)}}$: (2,2)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cSuisLUhIbmE",
        "outputId": "d8a2a5de-b161-4af2-818b-e7ad34f95058",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 136
        }
      },
      "source": [
        "dH1_dW1 = np.zeros((2,2,2))\n",
        "dH1_dW1[0,0,:] = x\n",
        "dH1_dW1[1,1,:] = x\n",
        "\n",
        "print_tensor(dH1_dW1)\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dH1_dW1:(2, 2, 2),float64\n",
            "[[[2.754  3.5407]\n",
            "  [0.     0.    ]]\n",
            "\n",
            " [[0.     0.    ]\n",
            "  [2.754  3.5407]]]\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CjGVJln0IiyE",
        "outputId": "3c22ae35-b9c2-4abd-f36a-2157f9e66f9c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dW1 = (dH1_dW1*dH1.reshape(dH1.shape[0], 1, 1)).sum(axis=0)\n",
        "dW1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 7.8203e-03,  1.0054e-02],\n",
              "       [-1.1676e-06, -1.5012e-06]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LwCu6MJfIkq1",
        "outputId": "3f15874d-47ac-4ca2-b139-a257b8fa5d13",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dW_torch[:2,1:]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 7.8203e-03,  1.0054e-02],\n",
              "        [-1.1676e-06, -1.5012e-06]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "r2ol_leTIl3d",
        "outputId": "1f06ec79-3dcb-425b-d817-ca4df0e803c0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        }
      },
      "source": [
        "print(dH1.shape)\n",
        "print(x.shape)\n",
        "np.dot(dH1, x)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(2, 1)\n",
            "(1, 2)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 7.8203e-03,  1.0054e-02],\n",
              "       [-1.1676e-06, -1.5012e-06]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LNoPD5WGIozx"
      },
      "source": [
        "### 텐서 평활화"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JpAaBdGwInU8",
        "outputId": "5047ffbb-d298-4f84-eea8-f2c53e6c2c98",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "T = np.arange(8).reshape(2,2,2)\n",
        "T"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[[0, 1],\n",
              "        [2, 3]],\n",
              "\n",
              "       [[4, 5],\n",
              "        [6, 7]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NgVddqLaIqLV",
        "outputId": "2934f8b3-9db5-454c-e108-0ed2e82da18c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        }
      },
      "source": [
        "T.reshape(-1,2)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0, 1],\n",
              "       [2, 3],\n",
              "       [4, 5],\n",
              "       [6, 7]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xkW4hG51Irdl",
        "outputId": "731ed905-84ce-4788-e99b-9de691f5e5be",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 102
        }
      },
      "source": [
        "T.reshape(-1,2).reshape(2,2,2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[[0, 1],\n",
              "        [2, 3]],\n",
              "\n",
              "       [[4, 5],\n",
              "        [6, 7]]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mPiSkwzOIuWE"
      },
      "source": [
        "### 텐서를 평활화해서 경사도벡터와 곱하는 한줄 코드"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OjJhRABEIsnc",
        "outputId": "cc36817c-e1ed-42a5-93d3-f88528ab33ee",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "np.dot(dH1_dW1.transpose(1,2,0).reshape(-1,2), dH1).reshape(2, 2, 1).squeeze()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 7.8203e-03,  1.0054e-02],\n",
              "       [-1.1676e-06, -1.5012e-06]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "27h4Ed7oIxff"
      },
      "source": [
        "## 데이터가 여러개인 경우"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Bxh3t_SIvjW",
        "outputId": "b532ef55-14a0-47a9-e89d-45dccd0760a3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        }
      },
      "source": [
        "N=3\n",
        "x = samples[[0,5,10]]\n",
        "x_torch = torch.tensor(x, dtype=torch.double); x_torch.requires_grad=True\n",
        "t = target[[0,5,10]]\n",
        "\n",
        "print_tensor(x)\n",
        "print_tensor(x_torch)\n",
        "print_tensor(t)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "x:(3, 2),float64\n",
            "[[2.754  3.5407]\n",
            " [0.1494 2.2842]\n",
            " [1.4176 3.4657]]\n",
            "-------------------------------------------\n",
            "x_torch:torch.Size([3, 2]),torch.float64\n",
            "tensor([[2.7540, 3.5407],\n",
            "        [0.1494, 2.2842],\n",
            "        [1.4176, 3.4657]], dtype=torch.float64, requires_grad=True)\n",
            "-------------------------------------------\n",
            "t:(3,),float64\n",
            "[1. 0. 1.]\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D95H3iS8I2X0"
      },
      "source": [
        "### 순전파와 역전파를 동시에 하는 함수"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4O3f9Fo1Iz2z"
      },
      "source": [
        "def forward_backward(X, W, T):\n",
        "    \"\"\"\n",
        "    네트워크를 포워드, 백워드 시킨다. numpy 버전\n",
        "    X : 네트워크의 입력벡터 shape:(N,2)\n",
        "    반환 : 목적함수, (W1에 대한 미분계수, b1에 대한 미분계수,\n",
        "                     W2에 대한 미분계수, b2에 대한 미분계수)\n",
        "    \"\"\"\n",
        "    # forward\n",
        "    N = X.shape[0]\n",
        "    \n",
        "    H1 = np.dot(W[:2,1:], X.T)\n",
        "    Z1 = H1 + W[:2,0].reshape(-1,1)\n",
        "    A1 = logistic(Z1)\n",
        "    \n",
        "    H2 = np.dot(W[2,1:], A1)\n",
        "    Z2 = H2 + W[2,0]\n",
        "    A2 = logistic(Z2)\n",
        "    \n",
        "    C = (1/(2*N)) * ((T-A2)**2).sum()\n",
        "    \n",
        "    # backward, dA->dZ->db->dW\n",
        "    dA2 = -(T-A2)/N\n",
        "    dA2_dZ2 = logistic(Z2)*(1-logistic(Z2))\n",
        "    dZ2 = dA2_dZ2 * dA2\n",
        "    \n",
        "    db2 = dZ2\n",
        "    dW2 = np.dot(dZ2, A1.T)\n",
        "\n",
        "    dH2_dA1 = W[2:,1:]\n",
        "    dA1 = np.dot(dH2_dA1.T, dZ2.reshape(-1,1))\n",
        "    dA1_dZ1 = logistic(Z1)*(1-logistic(Z1))\n",
        "    dZ1 = dA1_dZ1 * dA1\n",
        "\n",
        "    db1 = dZ1\n",
        "    dW1 = np.dot(dZ1, X)\n",
        "    \n",
        "    return C, (dW1, db1, dW2, db2)\n",
        "    "
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eyaRIsQ3I51R"
      },
      "source": [
        "### 모든 데이터에 대해 각각 역전파한 미분계수를 평균하기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0zuVbc3gI4Gt",
        "outputId": "bc0301b7-1d53-40fc-9a70-965cb78d0e3f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        }
      },
      "source": [
        "f, dW1, db1, dW2, db2 = 0, 0, 0, 0, 0\n",
        "\n",
        "for x_, t_ in zip(x, t):\n",
        "    x_ = np.array([x_])\n",
        "    t_ = np.array([t_])\n",
        "\n",
        "    fv, derivs = forward_backward(x_, W, t_)\n",
        "    \n",
        "    f += fv\n",
        "    dW1 += derivs[0]\n",
        "    db1 += derivs[1]\n",
        "    dW2 += derivs[2]\n",
        "    db2 += derivs[3]\n",
        "\n",
        "f   /= N\n",
        "dW1 /= N\n",
        "db1 /= N\n",
        "dW2 /= N\n",
        "db2 /= N\n",
        "\n",
        "print('C:', f)\n",
        "print_tensor(dW1)\n",
        "print_tensor(db1)\n",
        "print_tensor(dW2)\n",
        "print_tensor(db2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "C: 0.12001678911062634\n",
            "dW1:(2, 2),float64\n",
            "[[7.9160e-03 1.2443e-02]\n",
            " [1.5168e-06 4.1280e-05]]\n",
            "-------------------------------------------\n",
            "db1:(2, 1),float64\n",
            "[[2.8791e-03]\n",
            " [1.8509e-05]]\n",
            "-------------------------------------------\n",
            "dW2:(2,),float64\n",
            "[ 0.0061 -0.0499]\n",
            "-------------------------------------------\n",
            "db2:(1,),float64\n",
            "[-0.0497]\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3ArgDuGBI93o"
      },
      "source": [
        "### 역전파 한번으로 미분계수 구하기 "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OuC_ebzLI7EN",
        "outputId": "e3925501-8b63-49fc-f2ec-f18bd4ad65b5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "x"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[2.754 , 3.5407],\n",
              "       [0.1494, 2.2842],\n",
              "       [1.4176, 3.4657]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "01UE_SFQI_SF",
        "outputId": "aeb96106-4827-47b8-8589-335c2809968a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        }
      },
      "source": [
        "H1, Z1, A1, H2, Z2, A2, C = forward(x, W.reshape(3,3), t)\n",
        "\n",
        "print_tensor(H1)\n",
        "print_tensor(Z1)\n",
        "print_tensor(A1)\n",
        "print_tensor(H2)\n",
        "print_tensor(Z2)\n",
        "print_tensor(A2)\n",
        "print_tensor(C)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "H1:(2, 3),float64\n",
            "[[-2.8986  1.148  -0.4669]\n",
            " [ 9.5365  4.4643  8.0089]]\n",
            "-------------------------------------------\n",
            "Z1:(2, 3),float64\n",
            "[[-2.6223  1.4243 -0.1907]\n",
            " [10.6818  5.6096  9.1542]]\n",
            "-------------------------------------------\n",
            "A1:(2, 3),float64\n",
            "[[0.0677 0.806  0.4525]\n",
            " [1.     0.9964 0.9999]]\n",
            "-------------------------------------------\n",
            "H2:(3,),float64\n",
            "[ 0.1242 -0.1437 -0.0152]\n",
            "-------------------------------------------\n",
            "Z2:(3,),float64\n",
            "[ 0.0125 -0.2554 -0.1269]\n",
            "-------------------------------------------\n",
            "A2:(3,),float64\n",
            "[0.5031 0.4365 0.4683]\n",
            "-------------------------------------------\n",
            "C:(),float64\n",
            "0.12001678911062633\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FYNrIxzmJAj9",
        "outputId": "4d2922dc-a373-4884-9168-1b003dec05e5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        }
      },
      "source": [
        "H1_torch, Z1_torch, A1_torch, H2_torch, Z2_torch, A2_torch, C_torch = forward_torch(x_torch, W_torch, t)\n",
        "\n",
        "print_tensor(H1_torch)\n",
        "print_tensor(Z1_torch)\n",
        "print_tensor(A1_torch)\n",
        "print_tensor(H2_torch)\n",
        "print_tensor(Z2_torch)\n",
        "print_tensor(A2_torch)\n",
        "print_tensor(C_torch)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "H1_torch:torch.Size([2, 3]),torch.float64\n",
            "tensor([[-2.8986,  1.1480, -0.4669],\n",
            "        [ 9.5365,  4.4643,  8.0089]], dtype=torch.float64,\n",
            "       grad_fn=<MmBackward>)\n",
            "-------------------------------------------\n",
            "Z1_torch:torch.Size([2, 3]),torch.float64\n",
            "tensor([[-2.6223,  1.4243, -0.1907],\n",
            "        [10.6818,  5.6096,  9.1542]], dtype=torch.float64,\n",
            "       grad_fn=<AddBackward0>)\n",
            "-------------------------------------------\n",
            "A1_torch:torch.Size([2, 3]),torch.float64\n",
            "tensor([[0.0677, 0.8060, 0.4525],\n",
            "        [1.0000, 0.9964, 0.9999]], dtype=torch.float64,\n",
            "       grad_fn=<SigmoidBackward>)\n",
            "-------------------------------------------\n",
            "H2_torch:torch.Size([1, 3]),torch.float64\n",
            "tensor([[ 0.1242, -0.1437, -0.0152]], dtype=torch.float64,\n",
            "       grad_fn=<MmBackward>)\n",
            "-------------------------------------------\n",
            "Z2_torch:torch.Size([1, 3]),torch.float64\n",
            "tensor([[ 0.0125, -0.2554, -0.1269]], dtype=torch.float64,\n",
            "       grad_fn=<AddBackward0>)\n",
            "-------------------------------------------\n",
            "A2_torch:torch.Size([1, 3]),torch.float64\n",
            "tensor([[0.5031, 0.4365, 0.4683]], dtype=torch.float64,\n",
            "       grad_fn=<SigmoidBackward>)\n",
            "-------------------------------------------\n",
            "C_torch:torch.Size([]),torch.float64\n",
            "tensor(0.1200, dtype=torch.float64, grad_fn=<MulBackward0>)\n",
            "-------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ri0X5LQhJEdA"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{a}^{(2)}}$: (1,N)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PyFfFBsyJB4c",
        "outputId": "e5daa199-b132-4d9e-9b5f-d7221a440ba2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dA2 = -(t-A2)/N\n",
        "\n",
        "dA2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.1656,  0.1455, -0.1772])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YmJD44pyJHty"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{z}^{(2)}}$: (1,N)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ks9njX9xJFf1",
        "outputId": "3bc7f34b-624f-41a6-be78-9bc8a68067b5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dA2_dZ2 = logistic(Z2)*(1-logistic(Z2))\n",
        "dZ2 = dA2_dZ2 * dA2\n",
        "\n",
        "dZ2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.0414,  0.0358, -0.0441])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y5M4DQmlJLJd"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{b}^{(2)}}$: (1,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zGBRJv72JJJd",
        "outputId": "67838ad6-cbd3-40b9-83ac-949512823fb6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "db2 = dZ2.sum(axis=0, keepdims=True)\n",
        "dH2 = dZ2\n",
        "\n",
        "db2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.0497])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mr0IQojEJN-8"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{W}^{(2)}}$: (1,2)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NVcKaHhwJMct",
        "outputId": "2daa3d14-2889-4767-e805-f49bd293b815",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2_dW2 = A1.T\n",
        "dW2 = np.dot(dH2, dH2_dW2) # 정리된 결과로 야코비안 전치 평활화 따위 하지 않고 바로 delta * a.T\n",
        "\n",
        "dW2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0.0061, -0.0499])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 72
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qB4EoAVWJPjU",
        "outputId": "c740abb2-de79-49ee-ccb0-c7c962d3d62e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 73
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "22OGgqDNJRBs",
        "outputId": "37d4a767-df43-45f1-ac8b-5d8692e434b0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2.reshape(1,-1).shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lazWTXGsJSHU",
        "outputId": "b415a15e-a31f-4813-e8b1-e75fa4366a0d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "dH2.reshape(-1,1).shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(3, 1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TLG645dYJUt2"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{a}^{(1)}}$: (2,N)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SicbogOnJTBM",
        "outputId": "a0b2aa20-00b5-4e22-c126-06c1d3f03c55",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dH2_dA1 = W[2:,1:] # 분자레이아웃 아코비안 W2 (1,2)\n",
        "dA1 = np.dot(dH2_dA1.T, dH2.reshape(1,-1)) # W.T * delta\n",
        "\n",
        "dA1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.015 , -0.013 ,  0.016 ],\n",
              "       [-0.0062,  0.0053, -0.0066]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 76
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3Xr8a9CSJXlr"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{z}^{(1)}}$: (2,N)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hkEwLFdKJWRc",
        "outputId": "f5f073ba-e78a-492b-ce73-f687f217ce7e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dA1_dZ1 =  logistic(Z1)*(1-logistic(Z1))\n",
        "    \n",
        "dA1_dZ1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[6.3132e-02, 1.5636e-01, 2.4774e-01],\n",
              "       [2.2958e-05, 3.6357e-03, 1.0576e-04]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uOpfeQMBJY9k",
        "outputId": "7a49d46e-8461-4619-90c1-fef8704a8b8b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dZ1 = dA1_dZ1 * dA1 # 엘리먼트와이즈 곱\n",
        "\n",
        "dZ1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 9.4654e-04, -2.0261e-03,  3.9587e-03],\n",
              "       [-1.4133e-07,  1.9344e-05, -6.9385e-07]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qOtx7fKXJbyU"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{b}^{(1)}}$: (2,1)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cq-djEpdJaI8",
        "outputId": "960d3bf2-9d20-49e9-94a1-c1050d660e5e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "db1 = dZ1.sum(axis=1, keepdims=True)\n",
        "db1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[2.8791e-03],\n",
              "       [1.8509e-05]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XXHp0f7YJefR"
      },
      "source": [
        "### $\\frac{\\partial C}{\\partial \\mathbf{W}^{(1)}}$: (2,2)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PcOuMzfqJc_c",
        "outputId": "72898f36-558f-4e6a-95e9-06218b7b6e31",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "dW1 = np.dot(dZ1, x) # delta * a.T\n",
        "dW1"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[7.9160e-03, 1.2443e-02],\n",
              "       [1.5168e-06, 4.1280e-05]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w4GcUqs7JhQ-"
      },
      "source": [
        "### 파이토치로 미분하기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-yzegpyOJftc",
        "outputId": "9d0ebc93-c2b3-42b3-9741-e26a82049036",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "#                              종속변수, 독립변수, 종속변수와 곱해지는 상위 그래디언트 \n",
        "dW_torch = torch.autograd.grad(C_torch, W_torch, torch.tensor(1, dtype=torch.double), retain_graph=True)[0]\n",
        "dW_torch"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[ 2.8791e-03,  7.9160e-03,  1.2443e-02],\n",
              "        [ 1.8509e-05,  1.5168e-06,  4.1280e-05],\n",
              "        [-4.9747e-02,  6.0737e-03, -4.9872e-02]], dtype=torch.float64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 81
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jLREAmKqJi2c"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}