Demo20201203.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Demo20201203.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2_hOMgRZTIqp",
        "outputId": "4ef78bde-e651-4537-911d-ab7590bc1793"
      },
      "source": [
        "! nvidia-smi"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Thu Dec  3 01:52:06 2020       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 455.38       Driver Version: 418.67       CUDA Version: 10.1     |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|                               |                      |               MIG M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla P100-PCIE...  Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   41C    P0    28W / 250W |      0MiB / 16280MiB |      0%      Default |\n",
            "|                               |                      |                 ERR! |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                                  |\n",
            "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\n",
            "|        ID   ID                                                   Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6MGqlRPdUC4t"
      },
      "source": [
        "## Python Basic"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IyjtIOuYR2nK",
        "outputId": "e6e74c3d-64b2-4d1d-f097-ed03069af0a6"
      },
      "source": [
        "a = 3\n",
        "b = 2\n",
        "a + b"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I3gbcg2Ik38K"
      },
      "source": [
        "## Decision Tree"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tnjMQ-GEk3iK"
      },
      "source": [
        "from sklearn.datasets import load_iris\n",
        "iris = load_iris()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X04tY2uKk98e",
        "outputId": "55bd5602-256c-4c4e-a5da-d679faea5508"
      },
      "source": [
        "iris.data[0:3]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[5.1, 3.5, 1.4, 0.2],\n",
              "       [4.9, 3. , 1.4, 0.2],\n",
              "       [4.7, 3.2, 1.3, 0.2]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pTdS-BM3lPE6",
        "outputId": "e88f5de7-06e6-41f7-82ef-c219ed9cf561"
      },
      "source": [
        "iris.target"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8B5AcuZIlcew",
        "outputId": "6bc57e63-c051-4c8f-9379-ac6c19145649"
      },
      "source": [
        "from sklearn import tree\n",
        "clf = tree.DecisionTreeClassifier()\n",
        "clf.fit(iris.data, iris.target)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini',\n",
              "                       max_depth=None, max_features=None, max_leaf_nodes=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=1, min_samples_split=2,\n",
              "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
              "                       random_state=None, splitter='best')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SZUAZCvTlq8e"
      },
      "source": [
        "tree.export_graphviz(clf,'tree.dot')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sGAxd_eAlw3n",
        "outputId": "900e0338-91eb-42e1-a626-89b76c691926"
      },
      "source": [
        "! cat tree.dot"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "digraph Tree {\n",
            "node [shape=box] ;\n",
            "0 [label=\"X[2] <= 2.45\\ngini = 0.667\\nsamples = 150\\nvalue = [50, 50, 50]\"] ;\n",
            "1 [label=\"gini = 0.0\\nsamples = 50\\nvalue = [50, 0, 0]\"] ;\n",
            "0 -> 1 [labeldistance=2.5, labelangle=45, headlabel=\"True\"] ;\n",
            "2 [label=\"X[3] <= 1.75\\ngini = 0.5\\nsamples = 100\\nvalue = [0, 50, 50]\"] ;\n",
            "0 -> 2 [labeldistance=2.5, labelangle=-45, headlabel=\"False\"] ;\n",
            "3 [label=\"X[2] <= 4.95\\ngini = 0.168\\nsamples = 54\\nvalue = [0, 49, 5]\"] ;\n",
            "2 -> 3 ;\n",
            "4 [label=\"X[3] <= 1.65\\ngini = 0.041\\nsamples = 48\\nvalue = [0, 47, 1]\"] ;\n",
            "3 -> 4 ;\n",
            "5 [label=\"gini = 0.0\\nsamples = 47\\nvalue = [0, 47, 0]\"] ;\n",
            "4 -> 5 ;\n",
            "6 [label=\"gini = 0.0\\nsamples = 1\\nvalue = [0, 0, 1]\"] ;\n",
            "4 -> 6 ;\n",
            "7 [label=\"X[3] <= 1.55\\ngini = 0.444\\nsamples = 6\\nvalue = [0, 2, 4]\"] ;\n",
            "3 -> 7 ;\n",
            "8 [label=\"gini = 0.0\\nsamples = 3\\nvalue = [0, 0, 3]\"] ;\n",
            "7 -> 8 ;\n",
            "9 [label=\"X[0] <= 6.95\\ngini = 0.444\\nsamples = 3\\nvalue = [0, 2, 1]\"] ;\n",
            "7 -> 9 ;\n",
            "10 [label=\"gini = 0.0\\nsamples = 2\\nvalue = [0, 2, 0]\"] ;\n",
            "9 -> 10 ;\n",
            "11 [label=\"gini = 0.0\\nsamples = 1\\nvalue = [0, 0, 1]\"] ;\n",
            "9 -> 11 ;\n",
            "12 [label=\"X[2] <= 4.85\\ngini = 0.043\\nsamples = 46\\nvalue = [0, 1, 45]\"] ;\n",
            "2 -> 12 ;\n",
            "13 [label=\"X[1] <= 3.1\\ngini = 0.444\\nsamples = 3\\nvalue = [0, 1, 2]\"] ;\n",
            "12 -> 13 ;\n",
            "14 [label=\"gini = 0.0\\nsamples = 2\\nvalue = [0, 0, 2]\"] ;\n",
            "13 -> 14 ;\n",
            "15 [label=\"gini = 0.0\\nsamples = 1\\nvalue = [0, 1, 0]\"] ;\n",
            "13 -> 15 ;\n",
            "16 [label=\"gini = 0.0\\nsamples = 43\\nvalue = [0, 0, 43]\"] ;\n",
            "12 -> 16 ;\n",
            "}"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "1YFPMPR0mFkC",
        "outputId": "5caa09fa-37fb-4b2b-8cd1-2d3dfa8f735a"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "plt.scatter(iris.data[:,2], iris.data[:,3], c = iris.target)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f7b5b7e02e8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qo67go9WohCh"
      },
      "source": [
        "## Decision Tree Boundary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "27etcAI9ogep"
      },
      "source": [
        "X = iris.data[:, [2,3]]\n",
        "y = iris.target"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xCRlcUkym78V",
        "outputId": "c8a8342f-bbc0-4fac-ed64-8a9e64932d66"
      },
      "source": [
        "clf = tree.DecisionTreeClassifier(max_depth= 2)\n",
        "clf.fit(X,y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "DecisionTreeClassifier(ccp_alpha=0.0, class_weight=None, criterion='gini',\n",
              "                       max_depth=2, max_features=None, max_leaf_nodes=None,\n",
              "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
              "                       min_samples_leaf=1, min_samples_split=2,\n",
              "                       min_weight_fraction_leaf=0.0, presort='deprecated',\n",
              "                       random_state=None, splitter='best')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "yU2EIDDbo0fX",
        "outputId": "25f5cb10-eb85-4203-c5db-4e51bf2f4acf"
      },
      "source": [
        "import numpy as np\n",
        "x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
        "y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
        "xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1),\n",
        "                     np.arange(y_min, y_max, 0.1))\n",
        "\n",
        "plt.plot()\n",
        "Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "Z = Z.reshape(xx.shape)\n",
        "\n",
        "plt.contourf(xx, yy, Z, alpha=0.4, cmap = plt.cm.rainbow)\n",
        "plt.scatter(X[:, 0], X[:, 1], c=y, alpha=1, cmap = plt.cm.RdYlBu)\n",
        "plt.title('Decision Tree')\n",
        "plt.xlabel('Petal.Length')\n",
        "plt.ylabel('Petal.Width')\n",
        "plt.show()\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3tmgUwZ5sdus"
      },
      "source": [
        "## Logisitc Regression"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eEV9AwbDpJqc",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e66173e2-b143-4cb8-edcc-abd3089f9b05"
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "clf = LogisticRegression(max_iter = 300 )\n",
        "clf.fit(iris.data, iris.target)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
              "                   intercept_scaling=1, l1_ratio=None, max_iter=300,\n",
              "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
              "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                   warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5Px1dlRcsmEd"
      },
      "source": [
        "X = iris.data[:, [2,3]]\n",
        "y = iris.target"
      ],
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QWSYNKmgs0wa",
        "outputId": "31bf6c92-124a-4d67-e5f7-be52ec4af95c"
      },
      "source": [
        "clf.fit(X, y)"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
              "                   intercept_scaling=1, l1_ratio=None, max_iter=300,\n",
              "                   multi_class='auto', n_jobs=None, penalty='l2',\n",
              "                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n",
              "                   warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "_d8l6Ts5s3dB",
        "outputId": "eba28fa3-32a1-4a77-dec2-c1482035fb51"
      },
      "source": [
        "import numpy as np\n",
        "x_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
        "y_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
        "xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.1),\n",
        "                     np.arange(y_min, y_max, 0.1))\n",
        "\n",
        "plt.plot()\n",
        "Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "Z = Z.reshape(xx.shape)\n",
        "\n",
        "plt.contourf(xx, yy, Z, alpha=0.4, cmap = plt.cm.rainbow)\n",
        "plt.scatter(X[:, 0], X[:, 1], c=y, alpha=1, cmap = plt.cm.RdYlBu)\n",
        "plt.title('Logistic Regression')\n",
        "plt.xlabel('Petal.Length')\n",
        "plt.ylabel('Petal.Width')\n",
        "plt.show()\n",
        "\n"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WG45RZTWITRF"
      },
      "source": [
        "## Activation Function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NuOki-ceISTk",
        "outputId": "121f531a-8c37-4126-d99e-e84988403216"
      },
      "source": [
        "def threshold_function(x):\n",
        "  y = x > 0\n",
        "  return y.astype(int)\n",
        "\n",
        "import numpy as np\n",
        "threshold_function(np.array([-1,1,2]))"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 1, 1])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 29
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "sqTBMAVLs5kx",
        "outputId": "c3bed540-e791-4718-9168-b83b40b348b0"
      },
      "source": [
        "x = np.arange(-3,3,0.1)\n",
        "plt.plot(x, threshold_function(x))"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5b87ec88>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAR8klEQVR4nO3de6xlZ1nH8e9vzplTVMAaO1xsp0wTRsNoVPRYjRol4eKUmI7GWxuJN2KjsYoBL1VM0RL+QBI0aL2MgXgJUiuomcQhRbDGSyx2UKx0anVSxU4ldKiAEsLs2+Mfe53pZrOnZ/fMPrPn3Xw/ySR7r71mn2elv/P0nXe9a61UFZKk9u1ZdgGSpMWwoUvSirChS9KKsKFL0oqwoUvSilhf1g++4oor6sCBA8v68ZLUpPe9730fqap9sz5bWkM/cOAAJ06cWNaPl6QmJfng+T5zykWSVoQNXZJWhA1dklaEDV2SVoQNXZJWxLYNPclbkjya5APn+TxJ3pTkVJL7knzV4suUJG1nnhH67wKHn+Dz64CD3Z+bgN+88LIkSU/WtuvQq+qvkxx4gl2OAL9f4/vw3pPk8iTPrqoPLahGaSnuO/0x3n3yw8suQyvohc97Jl+x//KFf+8iLiy6Enh44v3pbttnNPQkNzEexXP11Vcv4EdLu+dN7/l33v3AoyTLrkSr5hlPf8ol29DnVlVHgaMAm5ubPllDl7RP9Ud89XO+gHf86NcvuxRpLotY5fIIsH/i/VXdNqlpvcGIvWsOz9WORTT0Y8D3datdvg74uPPnWgVnhyM21teWXYY0t22nXJK8DXgBcEWS08BrgL0AVfVbwHHgpcAp4JPAD+5WsdLF1B+M2FjzUg21Y55VLjdu83kBP7awiqRLRG84YmPdKRe1w+GHdB49R+hqjGmVzqM/HLGx7q+I2mFapfMYr3LxV0TtMK3SefQGjtDVFtMqncdZp1zUGNMqzVBV4zl0p1zUENMqzTAYFVXY0NUU0yrN0BuMAJxyUVNMqzRDfzhu6K5yUUtMqzSDI3S1yLRKM5y1oatBplWaYWvKxZOiaolplWboDR2hqz2mVZrh3By6I3Q1xLRKM/QdoatBplWaYeukqMsW1RLTKs3gskW1yLRKM/SHBcBlNnQ1xLRKM/ScclGDTKs0Q284BJxyUVtMqzRDfzCecrGhqyWmVZrh7Lmbc2XJlUjzs6FLM2zNoV+2trbkSqT52dClGbywSC0yrdIMj69yccpF7bChSzP0BiP2BNZdtqiGmFZpht5w5HSLmmNipRl6g5EXFak5JlaaoTccedm/mmNipRl6g5H3QldzTKw0Q384Yq8jdDVmrsQmOZzkwSSnktwy4/Ork9yd5J+S3JfkpYsvVbp4HKGrRdsmNskacDtwHXAIuDHJoandfgG4s6qeD9wA/MaiC5Uupt7AVS5qzzyJvRY4VVUPVVUPuAM4MrVPAU/vXn8+8N+LK1G6+HpDV7moPfMk9krg4Yn3p7ttk34ReFmS08Bx4MdnfVGSm5KcSHLizJkzOyhXujgcoatFi0rsjcDvVtVVwEuBP0jyGd9dVUerarOqNvft27egHy0tnssW1aJ5EvsIsH/i/VXdtkkvB+4EqKq/B54CXLGIAqVl6A89Kar2zJPYe4GDSa5JssH4pOexqX3+C3ghQJLnMW7ozqmoWV4pqhZtm9iqGgA3A3cBDzBezXJ/ktuSXN/t9irgh5P8M/A24AeqqnaraGm3OYeuFq3Ps1NVHWd8snNy260Tr08C37DY0qTl6Q/Lhq7mmFhphrNOuahBJlaaoTcYuspFzTGx0gxOuahFJlaaYXylqI+fU1ts6NKU4agYjoqNtbVllyI9KTZ0aUp/OH5AtFMuao2JlaacHYwbulMuao0NXZrS6xq6q1zUGhMrTek55aJGmVhpSv/clIu/HmqLiZWmOEJXq0ysNGVrDt3b56o1JlaasjVC3+sIXY0xsdKUc6tcHKGrMSZWmnJuysURuhpjYqUpW1eKuspFrTGx0hRH6GqViZWmuGxRrTKx0hSXLapVJlaa4ghdrTKx0hRH6GqViZWmeD90tcrESlN63pxLjTKx0pSeD7hQo2zo0pTesNhY30NiQ1dbbOjSlN5g5AlRNcnUSlN6w6EnRNUkUytN6Q/KEbqaZGqlKb3hiL3rzp+rPTZ0aYpz6GrVXKlNcjjJg0lOJbnlPPt8d5KTSe5P8oeLLVO6eM4ORmysry27DOlJW99uhyRrwO3Ai4HTwL1JjlXVyYl9DgI/B3xDVX00yTN2q2Bpt/WHIzZcg64GzTNCvxY4VVUPVVUPuAM4MrXPDwO3V9VHAarq0cWWKV08vcHIVS5q0jypvRJ4eOL96W7bpC8GvjjJ3yW5J8nhWV+U5KYkJ5KcOHPmzM4qlnZZb2hDV5sWldp14CDwAuBG4HeSXD69U1UdrarNqtrct2/fgn60tFj94cj7uKhJ86T2EWD/xPurum2TTgPHqqpfVf8B/BvjBi81x1UuatU8qb0XOJjkmiQbwA3Asal9/ozx6JwkVzCegnlogXVKF41z6GrVtqmtqgFwM3AX8ABwZ1Xdn+S2JNd3u90FPJbkJHA38NNV9dhuFS3tpt7QEbratO2yRYCqOg4cn9p268TrAl7Z/ZGa5ghdrTK10hRXuahVplaa0vekqBplaqUp45tz+auh9phaacJoVPSH3j5XbTK10oT+aPw8UefQ1SJTK03YekC0I3S1yNRKE841dEfoapCplSb0hwXY0NUmUytN2Bqhe3MutcjUShN6wyHgCF1tMrXShN6gm3JxhK4GmVppQm+4dVLUR9CpPTZ0acLjyxZ9SLTaY0OXJrhsUS0ztdKE/nBrlYtTLmqPDV2acNYRuhpmaqUJWydFL7Ohq0GmVprQ98IiNczUShMeX7bor4baY2qlCd5tUS0ztdKEc6tcHKGrQaZWmnDWEboaZmqlCU65qGWmVprQH47Yuxb27PHCIrXHhi5N6A1GLllUs0yuNKE3HLlkUc0yudKE/nDk/LmaZXKlCWedclHDTK40oTcYeR8XNcvkShP6zqGrYSZXmuAqF7VsruQmOZzkwSSnktzyBPt9R5JKsrm4EqWLx1Uuatm2yU2yBtwOXAccAm5McmjGfk8DXgG8d9FFShdLf1CuclGz5knutcCpqnqoqnrAHcCRGfu9Fng98KkF1iddVGeHI2/MpWbNk9wrgYcn3p/utp2T5KuA/VX150/0RUluSnIiyYkzZ8486WKl3dYbuA5d7brg5CbZA7wReNV2+1bV0ararKrNffv2XeiPlhauNxi6bFHNmie5jwD7J95f1W3b8jTgy4C/SvKfwNcBxzwxqhb1h8XeNW/MpTbN09DvBQ4muSbJBnADcGzrw6r6eFVdUVUHquoAcA9wfVWd2JWKpV3UG7jKRe3aNrlVNQBuBu4CHgDurKr7k9yW5PrdLlC6mFy2qJatz7NTVR0Hjk9tu/U8+77gwsuSlqPvhUVqmMmVJpx1hK6GmVypU1Xjm3M5QlejTK7UGYwKwCkXNcvkSp1zD4h2ykWNMrlSx4au1plcqdMf2tDVNpMrdc52I3Tn0NUqkyt1et0I3Xu5qFUmV+qcm3JxhK5GmVyp03PKRY0zuVLHVS5qncmVOj1XuahxJlfqOOWi1plcqbPV0F3lolaZXKnjlItaZ3KlztayRadc1CqTK3Vc5aLWmVypc66hO0JXo0yu1OkNx/dDt6GrVSZX6jjlotaZXKljQ1frTK7U6Q9H7Ams7cmyS5F2xIYudXrDkaNzNc30Sp3eYOQJUTXN9EodR+hqnemVOo7Q1TrTK3V6A0foapvplTp9p1zUONMrdXqDkTfmUtNMr9TxpKhaN1d6kxxO8mCSU0lumfH5K5OcTHJfkvckec7iS5V2lydF1bpt05tkDbgduA44BNyY5NDUbv8EbFbVlwNvB3550YVKu80Rulo3T3qvBU5V1UNV1QPuAI5M7lBVd1fVJ7u39wBXLbZMafc5Qlfr5knvlcDDE+9Pd9vO5+XAO2d9kOSmJCeSnDhz5sz8VUoXgatc1LqFpjfJy4BN4A2zPq+qo1W1WVWb+/btW+SPli6Yq1zUuvU59nkE2D/x/qpu26dJ8iLg1cA3V9XZxZQnXTxeWKTWzZPee4GDSa5JsgHcAByb3CHJ84HfBq6vqkcXX6a0+zwpqtZtm96qGgA3A3cBDwB3VtX9SW5Lcn232xuApwJ/nOT9SY6d5+ukS5YnRdW6eaZcqKrjwPGpbbdOvH7RguuSLjpH6Gqd6ZU6jtDVOtMrAcNRMSpc5aKmmV4JHxCt1WB6JWzoWg2mV2J8QhRgYy1LrkTaORu6xERDd4SuhpleCadctBpMr8T4xlwAG2trS65E2jkbusTjI/S9zqGrYTZ0CTjrlItWgOmVmJhysaGrYaZXYuKkqFeKqmGmV8JVLloNplfCKRetBtMr8fiFRd6cSy0zvRITq1xs6GqY6ZV4fMrlMqdc1DDTKzF5YZG/EmqX6ZVwlYtWg+mVsKFrNZheicfn0Nf3eC8XtcuGLgFnhyM21veQ2NDVLhu6xHjK5TJPiKpxJlhiPOWy1/lzNc4ES4xH6F5UpNaZYImuoTtCV+NMsAT0h+XTitQ8G7rE+F4uG+s+T1Rts6FLjO+26JSLWmeCJaDvskWtABMsMR6h7113Dl1tm6uhJzmc5MEkp5LcMuPzy5L8Uff5e5McWHSh0m5y2aJWwbYJTrIG3A5cBxwCbkxyaGq3lwMfrarnAr8CvH7RhUq7qe8culbA+hz7XAucqqqHAJLcARwBTk7scwT4xe7124FfT5KqqgXWCsCd9z7M7/zNQ4v+Wn2W++Bjn+S5z3jqssuQLsg8Df1K4OGJ96eBrz3fPlU1SPJx4AuBj0zulOQm4CaAq6++ekcFX/65ezn4TH/xtFgHn/lUvudr9i+7DOmCzNPQF6aqjgJHATY3N3c0en/Jlz6Ll3zpsxZalyStgnkmDR8BJocuV3XbZu6TZB34fOCxRRQoSZrPPA39XuBgkmuSbAA3AMem9jkGfH/3+juBv9yN+XNJ0vltO+XSzYnfDNwFrAFvqar7k9wGnKiqY8CbgT9Icgr4H8ZNX5J0Ec01h15Vx4HjU9tunXj9KeC7FluaJOnJcOGtJK0IG7okrQgbuiStCBu6JK2ILGt1YZIzwAd3+NevYOoq1Mat0vGs0rGAx3MpW6VjgfmP5zlVtW/WB0tr6BciyYmq2lx2HYuySsezSscCHs+lbJWOBRZzPE65SNKKsKFL0opotaEfXXYBC7ZKx7NKxwIez6VslY4FFnA8Tc6hS5I+U6sjdEnSFBu6JK2IZht6ktcmuS/J+5O8K8kXLbumC5HkDUn+tTumP01y+bJr2qkk35Xk/iSjJM0uK9vu4egtSfKWJI8m+cCya7lQSfYnuTvJyS5nr1h2TTuV5ClJ/iHJP3fH8ksX9H2tzqEneXpV/W/3+ieAQ1X1I0sua8eSvITxfeQHSV4PUFU/u+SydiTJ84AR8NvAT1XViSWX9KR1D0f/N+DFjB+7eC9wY1WdfMK/eIlK8k3AJ4Dfr6ovW3Y9FyLJs4FnV9U/Jnka8D7g21r8b5MkwOdV1SeS7AX+FnhFVd2zk+9rdoS+1cw7nwe0+X+mTlW9q6oG3dt7GD8ZqklV9UBVPbjsOi7QuYejV1UP2Ho4epOq6q8ZP6ugeVX1oar6x+71/wEPMH6ucXNq7BPd273dnx33smYbOkCS1yV5GPhe4Nbt9m/IDwHvXHYRn+VmPRy9yaaxypIcAJ4PvHe5lexckrUk7wceBf6iqnZ8LJd0Q0/y7iQfmPHnCEBVvbqq9gNvBW5ebrXb2+54un1eDQwYH9Mla55jkXZTkqcC7wB+cupf7E2pqmFVfSXjf5Vfm2THU2JzPbFoWarqRXPu+lbGT1R6zS6Wc8G2O54kPwB8K/DCS/2ZrE/iv02r5nk4upakm29+B/DWqvqTZdezCFX1sSR3A4eBHZ28vqRH6E8kycGJt0eAf11WLYuQ5DDwM8D1VfXJZdejuR6OriXoTiS+GXigqt647HouRJJ9WyvaknwO45PwO+5lLa9yeQfwJYxXU3wQ+JGqanYE1T1g+zLgsW7TPa2u2kny7cCvAfuAjwHvr6pvWW5VT16SlwK/yuMPR3/dkkvasSRvA17A+BatHwZeU1VvXmpRO5TkG4G/Af6F8e8/wM93zz5uSpIvB36Pccb2AHdW1W07/r5WG7ok6dM1O+UiSfp0NnRJWhE2dElaETZ0SVoRNnRJWhE2dElaETZ0SVoR/w+Eoc0V/GGovQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "br02ZEOIIPuh"
      },
      "source": [
        "def sigmoid_function(x):\n",
        "    return 1/ (1 + np.exp(-x))\n"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-7zcpE-EJDIw",
        "outputId": "079b4a35-001b-445c-d346-8c0dbf41e622"
      },
      "source": [
        "x = np.array([-1,1,2])\n",
        "sigmoid_function(x)"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.26894142, 0.73105858, 0.88079708])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "esgPn0BgJ0hX",
        "outputId": "f7c0fa34-22fe-4e3f-9906-f97f95a22d55"
      },
      "source": [
        "x = np.array([-1,1,2,5,6])\n",
        "sigmoid_function(x)"
      ],
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.26894142, 0.73105858, 0.88079708, 0.99330715, 0.99752738])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 35
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "SiUo1r9uJJKW",
        "outputId": "e30026ed-5178-4bd0-ca14-7e994e79683b"
      },
      "source": [
        "s = np.arange(-9,9,0.1)\n",
        "plt.plot(s, sigmoid_function(s))"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5b0d2c88>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zOfMmEUdJnGx",
        "outputId": "06f6e167-12dc-4f2d-c55e-9c622d728039"
      },
      "source": [
        "def tangent_function(x):\n",
        "    return (1 - np.exp(-2*x)) / (1 + np.exp(-2*x))\n",
        "\n",
        "x = np.array([-1,1,2])\n",
        "tangent_function(x)\n"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.76159416,  0.76159416,  0.96402758])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4aNxHAlUKIvA",
        "outputId": "e9c8b09d-a72a-4253-ff9a-df9983922b0f"
      },
      "source": [
        "np.tanh(x)"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-0.76159416,  0.76159416,  0.96402758])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 37
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "RUSuDlORKNhe",
        "outputId": "0df2ee9b-6031-4678-be84-6c36178620cc"
      },
      "source": [
        "s = np.arange(-9,9,0.1)\n",
        "plt.plot(s, np.tanh(s))"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5b0ae4e0>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rG1Tf8yWKSxe",
        "outputId": "76dcfbe3-ede6-4a0c-ea3d-a4355e989791"
      },
      "source": [
        "def relu_function(x):\n",
        "    return np.maximum(0,x)\n",
        "\n",
        "x = np.array([-1,1,2])\n",
        "relu_function(x)\n"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 1, 2])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "GYRtfXrUKm6j",
        "outputId": "0a06bd16-62ba-4f86-f510-569be1fd26ea"
      },
      "source": [
        "s = np.arange(-9,9,0.1)\n",
        "plt.plot(s, relu_function(s))"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5b00fe10>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 40
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wUwJM6v7LnXk"
      },
      "source": [
        "## Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4qbtmdfsLmwB"
      },
      "source": [
        "import numpy as np\n",
        "X = np.array([1,2])\n",
        "W = np.array([[1,3,5],\n",
        "              [2,4,6]])"
      ],
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VezCQG9JKr8B",
        "outputId": "e4bb96d1-8c0f-4fe5-c783-bcc4be9d8e11"
      },
      "source": [
        "X.shape"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cqp_Pb_vL6St",
        "outputId": "03c0818a-7120-4836-e594-6ecbf7d4ca5f"
      },
      "source": [
        "W.shape"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(2, 3)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 44
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1dW-EyA2L8_M",
        "outputId": "10cc358e-61ac-46de-8b6e-ffdc7440c5a2"
      },
      "source": [
        "a = np.dot(X,W)\n",
        "a"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 5, 11, 17])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "znCHHIB5L_hV"
      },
      "source": [
        "a1 = sigmoid_function(a)"
      ],
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hFXZ3OBHMYgX",
        "outputId": "8d1361ab-6867-4090-a5bc-149cedfdb424"
      },
      "source": [
        "a1"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.99330715, 0.9999833 , 0.99999996])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_FMxOFieMY7Z"
      },
      "source": [
        "W2 = np.array([[1,2],[3,4],[5,6]])"
      ],
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PmfhnAfoMegI",
        "outputId": "c6e4596e-2313-4cbf-dbe9-bf70dab0b6ff"
      },
      "source": [
        "np.dot(a1, W2)"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 8.99325684, 11.98654724])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3gWTO3vxMgG9"
      },
      "source": [
        "network = {}\n",
        "network['w1'] = np.array([[0.1,0.3,0.5],[0.2,0.4,0.6]])\n",
        "network['b1'] = np.array([0.1,0.2,0.3])\n",
        "network['w2'] = np.array([[0.1,0.4],[0.2,0.5],[0.3,0.6]])\n",
        "network['b2'] = np.array([0.1,0.2])\n"
      ],
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "a2idxhHbNaIf"
      },
      "source": [
        "x = np.array([1,0.5])\n"
      ],
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TMneJZLbNax3"
      },
      "source": [
        "a = np.dot(x, network['w1']) + network['b1']"
      ],
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k5bqGCadNesv",
        "outputId": "2ed3e364-b603-4f0f-af25-970a190ff64e"
      },
      "source": [
        "a"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.3, 0.7, 1.1])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 54
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VQ6o9tAyNfEP",
        "outputId": "7a86d6f5-4f1a-45f6-8f88-bca914084eb5"
      },
      "source": [
        "z = sigmoid_function(a)\n",
        "z"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.57444252, 0.66818777, 0.75026011])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 55
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_27bwdOnNk_n"
      },
      "source": [
        "y = np.dot(z, network['w2']) + network['b2']"
      ],
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PvGEAqYwNqUN",
        "outputId": "ea8381f2-9cce-48ee-a38a-ecb6794a3667"
      },
      "source": [
        "y"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.51615984, 1.21402696])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "c4PHAvJBNrqE",
        "outputId": "c1dd1388-1364-4a4b-e85e-a4b3bb04da6b"
      },
      "source": [
        "\n",
        "def softmax_function(x):\n",
        "    return np.exp(x) / np.sum(np.exp(x))\n",
        "\n",
        "y = np.array([-1,1,2])\n",
        "softmax_function(y)\n"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.03511903, 0.25949646, 0.70538451])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-vY7zHcEOoTc"
      },
      "source": [
        "def mean_squared_err(y_hat, y):\n",
        "    return 0.5 * np.sum((y_hat - y) ** 2)\n"
      ],
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sFQtGfC9P8g3"
      },
      "source": [
        "def cross_entropy_err(y_hat, y):\n",
        "    delta = 1e-8\n",
        "    return -np.sum(y*np.log(y_hat + delta))\n"
      ],
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "IGYyHUVbQNp9",
        "outputId": "40f9a6e6-d806-4744-93c3-7f0450ad5076"
      },
      "source": [
        "def func(x):\n",
        "    return x ** 2 \n",
        "s = np.arange(-9,9,0.1)\n",
        "plt.plot(s, func(s))\n"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5aca85f8>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAD4CAYAAAD1jb0+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3xW9dnH8c+VTQYJGcwEwgh7E5Y4GYoTByJOrAPrqnZZa5fV52mrtVpbWysVFRHRiotaRRxUZBP2JoEkJGEkJBAIIft6/shNH4rB3JDcOfdJrvfrlVfucZLzfZ2EK4ff+Z3fJaqKMcYY9wlwOoAxxpizYwXcGGNcygq4Mca4lBVwY4xxKSvgxhjjUkFNubP4+HhNTk5uyl0aY4zrrVmz5qCqJpz6epMW8OTkZNLS0ppyl8YY43oikl3X6zaEYowxLmUF3BhjXMoKuDHGuJQVcGOMcSkr4MYY41JWwI0xxqWsgBtjjEu5ooD/a+M+3lhR5zRIY4xpsbwq4CLyfRHZIiKbRWSuiISJSFcRWSkiGSLytoiE+Crkx5v28czCHZRVVvtqF8YY4zr1FnAR6QR8D0hV1f5AIDAVeAp4TlV7AIeAO30VcuqIJA6XVvLplv2+2oUxxriOt0MoQUArEQkCwoF9wFhgnuf9WcDVjR+v1pju8STFtuKtVTm+2oUxxrhOvQVcVfOAZ4A91BbuYmANcFhVqzyb5QKd6vp6EZkuImkiklZQUHB2IQOEqcM7s3x3IZkHj53V9zDGGCeUVVaTfuCoT763N0MobYBJQFegIxABTPR2B6o6Q1VTVTU1IeEbi2l57fphiQQGCG+t3nPW38MYY5raJ5v3MeG5xWzIOdzo39ubIZTxQKaqFqhqJfAeMAaI8QypACQCeY2e7iRtW4cxtndb3l2TS0VVjS93ZYwxjWbuyhyS48IZ0Cm60b+3NwV8DzBKRMJFRIBxwFZgETDZs8004MNGT3eKm0Z05mBJBZ9vO+DrXRljTINl5JewKquIG4Z3JiBAGv37ezMGvpLai5VrgU2er5kB/AT4gYhkAHHAzEZPd4rzeybQMTqMuatsGMUY4//eXr2HoABh8rBEn3x/rxo6qOqvgF+d8vJuYESjJ/oWgQHC9alJ/OnLdHKKSkmKDW/K3RtjjNfKq6qZtyaXCX3bkRAV6pN9uOJOzJNNGZ6EAG+vtimFxhj/tXDLAQ6VVjJ1RGef7cN1BbxTTCsu6JnAO2tyqKq2i5nGGP/01uo9dIppxXk94n22D9cVcICpIzpz4Eg5i3ac3bxyY4zxpezCYyzNKGTq8CSfXLw8wZUFfGzvtrSNCrWLmcYYv/TW6hwCBK5PTfLpflxZwIMDA7g+NZF/78hnX/Fxp+MYY8x/VFTV8E5aDmN7t6V9dJhP9+XKAg5wQ2pnahT+sTrX6SjGGPMfC7fu52BJBTeP6uLzfbm2gHeOC+e8lHjeXr2H6hp1Oo4xxgDwxopsEtu04vyUs186xFuuLeAAU4d3Zm9xGYvT7WKmMcZ5GfklrNhdxI0jOhPow4uXJ7i6gE/o2464iBDmrLCLmcYY5725cg/BgcIUH1+8PMHVBTwkKIAbhifx5fYD5B22i5nGGOeUVVYzb00Ol/Rr77M7L0/l6gIOcNPIzigwd6WdhRtjnPPPDXs5UlbFzSN9f/HyBNcX8MQ24Yzr3Za3Vu+xZWaNMY6Zs3IP3RMiGNUttsn26foCDnDLqC4cLKlggfXMNMY4YHNeMetzDnPzyC7UrrrdNJpFAT8/JYEuceHMXp7ldBRjTAs0Z+UewoIDuG6ob5aNPZ1mUcADAoRbRnZhddYhtu074nQcY0wLcrSskg/X53HlwI5Ehwc36b696YnZS0TWn/RxREQeFpFYEflMRNI9n9s0ReDTmTwskdCgAN5Yke1kDGNMC/PB+r2UVlQ3yZ2Xp/KmI88OVR2sqoOBYUAp8D7wKPCFqqYAX3ieO6ZNRAhXDurI++vyOFpW6WQUY0wLoarMWZFNv46tGZTY+D0v63OmQyjjgF2qmk1tp/pZntdnAVc3ZrCzceuoLpRWVPP+Op/2VzbGGABWZRaxff9Rbh3VtBcvTzjTAj4VmOt53E5V93ke7wfa1fUFIjJdRNJEJK2gwLe3vA9KimFQYjSzl2ejauujGGN86/Xl2US3CmbS4E6O7N/rAi4iIcBVwDunvqe11bLOiqmqM1Q1VVVTExJ8v7jLLaO6kJ5fwsrMIp/vyxjTcu0vLmPBlv3cMDyJViGBjmQ4kzPwS4G1qnrA8/yAiHQA8HzOb+xwZ+PKQR2JbhXM7OV2MdMY4ztvrsymRpVbmvDOy1OdSQG/kf8fPgGYD0zzPJ4GfNhYoRoiLDiQKamJfLplPweOlDkdxxjTDJVXVfPmqj2M7dWWznHhjuXwqoCLSAQwAXjvpJd/B0wQkXRgvOe5X7h5ZBeqapS3VlnnemNM41uwubZpw23nJDuaw6sCrqrHVDVOVYtPeq1QVcepaoqqjldVvxl0To6P4IKeCby5KptK61xvjGlkry3Lomt8hE87znujWdyJWZdbR3XhwJFyPt96oP6NjTHGSxtzD7Nuz2FuHdXFpx3nvdFsC/hFvduS2KYVry7LcjqKMaYZeX15NuEhgUxObdp1T+rSbAt4YIAwbXQyqzKL2LK3uP4vMMaYehQdq2D+hr1cO7QTrcOadt2TujTbAg4wZXgS4SGBvLo0y+koxphm4O3VOVRU1XDb6GSnowDNvIBHtwrmuqGJzF+/l4Ml5U7HMca4WFV1DW+syGZ0tzh6totyOg7QzAs4wLRzkqmorrGWa8aYBvlsa23v3WkOTx08WbMv4D3aRnJ+zwRmr8i2lmvGmLP2ytJMkmJbMaFvncs+OaLZF3CA74xJJv9oOZ9s3lf/xsYYc4qNuYdZnXWI28/pSqDDUwdP1iIK+AUpCXSLj+AVu5hpjDkLry7NIjI0iCl+MHXwZC2igAcECLePSWZDzmHW7jnkdBxjjIscOFLGPzfs5frURKL8YOrgyVpEAQe4bmgiUaFBNqXQGHNGZi/PplqV2/3o4uUJLaaAR4QGMWV4Ep9s2sf+Ylul0BhTv7LKauaszGZCn3Z0iYtwOs43tJgCDjBtdDLVqsxekeV0FGOMC3ywLo9DpZXccW5Xp6PUqUUV8M5x4Vzctx1vrNhDaUWV03GMMX5MVXllaSZ9O7RmZNdYp+PUqUUVcIDp53ej+Hgl89bkOh3FGOPHlmYUsvNACXec29WRhsXe8LahQ4yIzBOR7SKyTURGi0isiHwmIumez218HbYxDOsSy5DOMcxckkl1jTU+NsbU7ZWlmcRHhnDloA5ORzktb8/AnwcWqGpvYBCwDXgU+EJVU4AvPM9d4e7zupFdWMpntla4MaYOGfklfLk9n1tGdSE0yJmGxd6ot4CLSDRwPjATQFUrVPUwMAmY5dlsFnC1r0I2tkv6tScpthV//3q301GMMX7o5a93ExoUwK2jnGtY7A1vzsC7AgXAqyKyTkRe9vTIbKeqJ+5N3w/UuUCAiEwXkTQRSSsoKGic1A0UGCDcMaYra7IP2Y09xpj/kn+0jPfW5jF5WCJxkaFOx/lW3hTwIGAo8KKqDgGOccpwiaoqUOeAsqrOUNVUVU1NSEhoaN5GMyU1idZhQbxsZ+HGmJO8viybypoa7jqvm9NR6uVNAc8FclV1pef5PGoL+gER6QDg+Zzvm4i+EREaxE0ju7Bg835yikqdjmOM8QPHyquYvSKbi/u2o2u8/924c6p6C7iq7gdyRKSX56VxwFZgPjDN89o04EOfJPSh289JJkCEmUsynY5ijPED76TlUHy8kunnd3c6ile8nYXyIDBHRDYCg4HfAL8DJohIOjDe89xV2keHcdWgjvwjLYfi0kqn4xhjHFRVXcPLSzJJ7dKGYV1cMSvauwKuqus949gDVfVqVT2kqoWqOk5VU1R1vKoW+TqsL9x1XjdKK6qZsyrb6SjGGAct2LKf3EPHuft8/x/7PqHF3Yl5qr4dW3NeSjyvLs2irLLa6TjGGAeoKjMW76ZrfAQT+vhPx536tPgCDnDvBd0pOFrOe2vznI5ijHHAit1FbMwt5q7zuhLgRx136mMFHBjdPY6BidHMWLzLbq83pgX6+9e7iYsI4bqh/tVxpz5WwAER4d4LupNVWMqCzfudjmOMaUI7Dxzly+353DY6mbBg/71tvi5WwD0u7teebvERvPhVBrX3JRljWoIX/72L8JBAbhvt37fN18UKuEdggDD9/G5szjvC0oxCp+MYY5pATlEp8zfs5aYRnWkTEeJ0nDNmBfwk1wztRNuoUF78KsPpKMaYJvDS4l0ECK64bb4uVsBPEhoUyF3ndWVpRiEbcw87HccY40P5R8v4R1ouk4cl0j46zOk4Z8UK+CluHNGZqLAg/vbVLqejGGN8aOaSTKqqa7jHJbfN18UK+CmiwoK5bXQXPtm8n90FJU7HMcb4QPHxSuas2MNlAzqQ7IJFq07HCngdbj+nKyGBAbz0lS01a0xzNHt5FiXlVdx7oXvPvsEKeJ0SokK5YXgS763LJe/wcafjGGMa0fGKal5ZmsVFvRLo1zHa6TgNYgX8NO65oPYv80s2Fm5Ms/LW6j0UHavgvot6OB2lwayAn0anmFZcNzSRt1bnkH+kzOk4xphGUFFVw4zFuxme3IbhybFOx2kwK+Df4r4Le1Bdo7y02MbCjWkO3l+Xy77iMu670P1n32AF/Ft1jgtn0qCOzFmZTWFJudNxjDENUFldwwuLMhiYGM2FvfynP29DeFXARSRLRDaJyHoRSfO8Fisin4lIuuezO1pYnKH7LupBeVVtpw5jjHt9sC6PnKLjfG9sCiLuWTL225zJGfhFqjpYVVM9zx8FvlDVFOALTulU31z0aBvJ5QM68PqyLA6XVjgdxxhzFqqqa/jLogz6dWzNuD5tnY7TaBoyhDIJmOV5PAu4uuFx/NMDY3twzDP1yBjjPvM37CWrsJTvjWs+Z9/gfQFXYKGIrBGR6Z7X2qnqPs/j/UCdfYhEZLqIpIlIWkFBQQPjOqN3+9Zc3Lcdry3N5EiZNT82xk2qa5QXvsygd/soV7VL84a3BfxcVR0KXArcLyLnn/ym1i6gXeci2qo6w9MQOTUhwb0XDh4cm8KRsipmL7fmx8a4yUcb97L74DEeGpfiqnZp3vC2K32e53M+8D4wAjggIh0APJ/zfRXSHwxIjOaiXgn8/evdHLWzcGNcobpG+fOXGfRqF8Ul/do7HafR1VvARSRCRKJOPAYuBjYD84Fpns2mAR/6KqS/eHh8Tw6XVjJrWZbTUYwxXvh40z4y8kt4cFyPZnf2Dd6dgbcDlojIBmAV8C9VXQD8DpggIunAeM/zZm1QUgzj+7RlxuLdFB+3s3Bj/FlNjfLnL9Pp0TaSS/t3cDqOT9RbwFV1t6oO8nz0U9X/9bxeqKrjVDVFVcerapHv4zrv4fE9OVJWxSs2L9wYv/bJ5v3sPFDCg2N7ENgMz77B7sQ8Y/07RTOxX3teWZJp88KN8VPVNcpzn++kR9tIrhjY0ek4PmMF/Cw8PCGFo+VVvPy1nYUb44/mb8gjI7+EH0zo2WzPvsEK+Fnp3b41lw/swKtLMyk6ZmfhxviTyuoa/vh5On07tGZiM5x5cjIr4Gfp4XEplFZW89JiWy/cGH/y7ppcsgtL+eHFPZvlzJOTWQE/Syntopg0qCOvL8um4KitVGiMPyivquZPX6QzOCmGsb2bz5onp2MFvAG+Ny6F8qpq69pjjJ+Yu3IPe4vL+PElvZrVmienYwW8AbolRHLNkERmr8hmX7H1zjTGSccrqnlh0S5GdYvlnO5xTsdpElbAG+jh8SnUqPKnL9KdjmJMi/b68iwOlpTzw4tbxtk3WAFvsKTYcG4e2YV/pOWyq6DE6TjGtEhHyyr521e7uKBnQrPodektK+CN4IGxPQgNCuDZhTudjmJMizRzSSaHSiv54cU9nY7SpKyAN4L4yFDuOq8b/9q0j425h52OY0yLUnC0nL8v3s1lA9ozMDHG6ThNygp4I7n7vK7ERoTw9IIdTkcxpkV54ct0yqpq+NHFvZyO0uSsgDeSqLBg7r+oB0syDrIk/aDTcYxpEbILjzFn5R6mDk+iW0Kk03GanBXwRnTzyM50imnF059up7ZJkTHGl55ZuJPgwAAeGp/idBRHWAFvRGHBgTw8PoWNucUs2Lzf6TjGNGsbcw/zzw17ufu8rrSNCnM6jiO8LuAiEigi60TkI8/zriKyUkQyRORtEQnxXUz3uHZoIiltI/n9wh1UVtc4HceYZklV+d0n24mNCOHu87s5HccxZ3IG/hCw7aTnTwHPqWoP4BBwZ2MGc6vAAOGRib3ZXXCMt1btcTqOMc3S1+kHWbarkAfH9iAqLNjpOI7xqoCLSCJwOfCy57kAY4F5nk1mAVf7IqAbje/TlpFdY3nu83SOWANkYxpVTU3t2XdSbCtuGtnZ6TiO8vYM/I/AI8CJMYE44LCqVnme5wKd6vpCEZkuImkiklZQUNCgsG4hIvz88r4UHavgxX/bQlfGNKb5G/aydd8RfnRxL0KDAp2O4yhvutJfAeSr6pqz2YGqzlDVVFVNTUhIOJtv4UoDEqO5dkgnZi7JJPdQqdNxjGkWjldU8/SC7fTv1Jorm3GrNG95cwY+BrhKRLKAt6gdOnkeiBGRIM82iUCeTxK62I8u6YUAz3xqN/cY0xhe/no3e4vL+MXlfZt9swZveNOV/qeqmqiqycBU4EtVvRlYBEz2bDYN+NBnKV2qY0wr7jqvKx+s38uGHLvF3piGOHCkjBe/2sXEfu0Z2a1lLBdbn4bMA/8J8AMRyaB2THxm40RqXu69sAfxkSH877+22c09xjTAM5/uoKpa+ellvZ2O4jfOqICr6r9V9QrP492qOkJVe6jq9apqfcXqEBkaxMPje7Iqq4iFWw84HccYV9qcV8y8tbncPiaZLnERTsfxG3YnZhOYOjyJlLaR/O6T7VRU2c09xpwJVeXJj7bSJjyEB8b2cDqOX7EC3gSCAgN47PI+ZB48xqxlWU7HMcZVPt1ygJWZRXx/Qk9at+CbdupiBbyJXNSrLWN7t+X5L9LJP1rmdBxjXKG8qprffrKNnu0iuXF4ktNx/I4V8Cb0iyv6Ul5VbWuGG+Ol15dlk11Yys8u70tQoJWrU9kRaUJd4yO489xuzFuTy7o9h5yOY4xfyz9SxvNfpHNRrwQu6NlybgI8E1bAm9gDY3vQNiqUx+dvoabGphUaczq/9Vz0/9WV/ZyO4resgDexyNAgfnpZbzbk1k6LMsZ806rMIt5fl8f087uRHG/TBk/HCrgDrh7ciaGdY3h6wXZbrdCYU1RV1/DLDzfTMTqM+y7q7nQcv2YF3AEiwq+v6k/hsQr+9Hm603GM8StzVu5h+/6j/PyKvoSHBNX/BS2YFXCHDEiM5obUJF5blkX6gaNOxzHGLxwsKecPC3dwbo94Lu3f3uk4fs8KuIN+fEkvIkKD+NkHm22dFGOApxdsp7Simsev6ktt3xjzbayAOyguMpRHL+3Nqswi3l1rq/Galm3dnkP8Iy2XO8/tSo+2UU7HcQUr4A67ITWJoZ1j+M3H2zhcWuF0HGMcUVVdwy8+3Ey71qE8OC7F6TiuYQXcYQEBwv9eM4Di45U8tWC703GMccRry7LYnHeEX1zRl8hQu3DpLSvgfqBPh9bceW5X5q7KYU12kdNxjGlSeYeP8+xnO7moVwKXD+jgdBxXsQLuJx4al0LH6DB+9v5mKqttyVnTMqgqv/pwM6rwxKT+duHyDHnT1DhMRFaJyAYR2SIiv/a83lVEVopIhoi8LSIhvo/bfEWEBvGrq/qxff9RXl2a6XQcY5rEgs37+XxbPj+Y0JOk2HCn47iON2fg5cBYVR0EDAYmisgo4CngOVXtARwC7vRdzJbh4r7tGN+nLc99lk7e4eNOxzHGp46UVfKr+Vvo26E13xmT7HQcV/KmqbGqaonnabDnQ6ntTj/P8/os4GqfJGxBRITHr+qHCDz23iabG26atWc+3cHBknJ+e+0AWyr2LHl11EQkUETWA/nAZ8Au4LCqVnk2yQU6neZrp4tImoikFRQUNEbmZi2xTTiPXNKLr3YW8P46mxtumqe1ew4xe0U2t41OZlBSjNNxXMurAq6q1ao6GEgERgBet4VW1RmqmqqqqQkJtqavN24bncywLm144qOtFBy1XtGmeamoquGx9zbRvnUYP7qkl9NxXO1Mu9IfBhYBo4EYETkxYTMRsNPFRhIQIDx13UBKK6r51fzNTscxplH9ZVEG2/cf5clJ/W3OdwN5MwslQURiPI9bAROAbdQW8smezaYBH/oqZEvUo20kD41L4eNN+1mweZ/TcYxpFFv2FvOXRRlcO6QT4/u2czqO63lzBt4BWCQiG4HVwGeq+hHwE+AHIpIBxAEzfRezZZp+fjf6dWzNzz/YQnGprRtu3K2yuoYfvbORNhEh/PLKvk7HaRa8mYWyUVWHqOpAVe2vqk94Xt+tqiNUtYeqXq+qNljbyIIDA3h68kAOlVbw5L+2Oh3HmAb566JdbNt3hN9cM4CYcLttpDHY3B0/169jNN+9oLYR8r935Dsdx5izsnXvEf78ZTqTBndkgg2dNBor4C7w4NgUUtpG8pN3N9qKhcZ1Kqtr+PG8DcSEB/O4NShuVFbAXSAsOJDnbhhMYUkFv/hwi9NxjDkjL321iy17j/A/Vw+gTYQNnTQmK+Au0b9TNA+PT+GfG/Yyf8Nep+MY45XNecU8/0U6VwzswERrkdborIC7yHcv6M6QzjH84oPN7C8uczqOMd+qrLKah99eT2xECE9O6u90nGbJCriLBAUG8OyUwVRU1fDIuxttrRTj13778TYy8kt45vpBNnTiI1bAXaZrfASPXd6HxTsLeGPlHqfjGFOnRTvymbU8mzvGdOW8FFtCw1esgLvQLSM7c37PBH7zr21kHjzmdBxj/kthSTmPzNtIr3ZRPDLR1jrxJSvgLiQi/H7yQEKCAvje3HVUVFkHH+MfVJWfvreJ4tJK/jh1MGHBgU5HatasgLtUu9ZhPD15IJvyinnamiEbP/GPtBwWbj3AIxN70adDa6fjNHtWwF3skn7tuW10F15eksmi7XaXpnHWroISfv3PrZzTPY47xnR1Ok6LYAXc5R67rA+920fxw3c2cOCITS00ziirrOb+OWsJCw7k2SmDCQiw5sRNwQq4y4UFB/LCTUM5XlHNw2+tp7rGphaapvfrf25h+/6jPDtlEO2jw5yO02JYAW8GerSN5IlJ/Vi+u5C/LspwOo5pYT5Yl8fcVTncd2F3LuzV1uk4LYoV8GZi8rBEJg3uyHOf72RVZpHTcUwLsaughMfe38Tw5Db8YEJPp+O0ON505EkSkUUislVEtojIQ57XY0XkMxFJ93xu4/u45nREhP+5uj+dY8N5cO5a8o/aeLjxrZPHvf904xDrLO8Ab454FfBDVe0LjALuF5G+wKPAF6qaAnzheW4cFBUWzIu3DKP4eCUPvLmOymqbH25859f/3Pqfce8O0a2cjtMiedORZ5+qrvU8PkptP8xOwCRglmezWcDVvgppvNenQ2t+e+0AVmUW8dQnNj/c+MY7aTnMXbWHe23c21Fn9H8eEUkGhgArgXaqeqLb7n6gzjYbIjJdRNJEJK2goKABUY23rhmSyDTP/PCPNtrSs6Zxbcg5zM8+2MyYHnH80Ma9HeV1AReRSOBd4GFVPXLye1q7LF6d89dUdYaqpqpqakKCLWrTVH52eV+GdWnDI/M2svPAUafjmGai4Gg598xeQ0JkKH++caiNezvMq6MvIsHUFu85qvqe5+UDItLB834HwG4F9CMhQQH89eahhIcE8d3ZazhaZl3tTcNUVNVw/5y1HD5ewYzbhhFrS8Q6zptZKALMBLap6rMnvTUfmOZ5PA34sPHjmYZo1zqMv9w0hOyiUr7/9gZq7CYf0wD/86+trMoq4qnrBtKvY7TTcQzenYGPAW4FxorIes/HZcDvgAkikg6M9zw3fmZktzh+cXkfPt92gKc/3eF0HONS/0jL4fXl2dx9XlcmDe7kdBzjEVTfBqq6BDjdwgbjGjeO8YVp5ySTnl/C377aRfeECK5PTXI6knGRNdmH+Pn7tRctfzKxt9NxzEnsCkQLICI8flU/xvSI47H3N9mdmsZrewpLufv1NDrEhPGCXbT0O/bTaCGCAwP4603DSGoTzj2z09hTWOp0JOPniksruf21VVTXKK/ePtz6WvohK+AtSHR4MDNvH06Nwp2zVnPEZqaY06ioquGeN9LILTrOjFuH0S0h0ulIpg5WwFuYrvERvHjLUDIPHuP+OWvtdnvzDarKo+9tZMXuIp6ePJCR3eKcjmROwwp4C3RO93h+c80Avk4/yCPzNtr0QvNf/vxlBu+tzeP743ty9RCbceLP6p2FYpqnKcOTKCgp5/ef7iAhKpTHLuvjdCTjB+atyeXZz3Zy7ZBOfG9cD6fjmHpYAW/B7ruwO/lHypixeDfxkSFMP7+705GMgxZu2c9P3t3ImB5x/Pa6AdTew2f8mRXwFkxE+OWV/Th4rILffLyd+MhQrh2a6HQs44Bluw7ywNx19O8UzYxbUwkNCnQ6kvGCFfAWLjBAeHbKIA4dq+CReRtpExHCRbY8aIuyMfcwd89Ko0tsOK/dPpyIUCsLbmEXMQ2hQYG8dOsweraL4t431rBid6HTkUwTycg/yrRXVtEmIoTZd460ud4uYwXcALXdfF6/cwRJbcK547XVrM6yuzWbu9xDpdw6cxWBAQG8cedI6ybvQlbAzX/ER4Yy5+7af8i3v7KKNdmHnI5kfCSnqJSpM1ZQUl7F63eMIDk+wulI5ixYATf/pW1UGHPvHkVCVCi3v7KK9TmHnY5kGtmJ4n3keCVz7hpJ346tnY5kzpIVcPMN7VqH8ebdo4iJCOa2mSvZnFfsdCTTSPYUlnLDS8spKa/izbtHMTAxxulIpgGsgJs6dYxpxdy7RxEVFszNL69k3R4bTnG7rIPHuGHGckorq5lz10j6d7KmDG7nTUeeV0QkX0Q2n/RarIh8JiLpnlI0m0wAAAyiSURBVM9tfBvTOCGxTThvTR9FTHhtEV+SftDpSOYsZR48xtQZKyirrObNu0ZZ8W4mvDkDfw2YeMprjwJfqGoK8IXnuWmGkmLDeeee0XSOrZ2dsmDzfqcjmTO0KbeYyS8uo6K6hjfvHmVj3s1IvQVcVRcDp84pmwTM8jyeBVzdyLmMH2nbOoy3po+iX6fW3DdnDe+k5TgdyXjp6/QCps5YTlhwIPO+O5o+Hax4NydnOwbeTlX3eR7vB9o1Uh7jp2LCQ5hz10jG9Ijnx/M2MnNJptORTD0+XJ/HHa+tJik2nPfuO8fW9G6GGnwRU1UVOO16pCIyXUTSRCStoKCgobszDgoPCeLlaalc2r89T360lSf+uZVqW4rWL81ckslDb61nSOc2vH3PaNq1tpt0mqOzLeAHRKQDgOdz/uk2VNUZqpqqqqkJCQlnuTvjL0KDAnnhpqHcMaYrryzN5O7X0ygpr3I6lvGorlF+8/E2nvxoKxP7tef1O0YQ3SrY6VjGR862gM8HpnkeTwM+bJw4xg0CA4RfXtmXJ6/uz1c7C5j84jLyDh93OlaLV1JexfTX05ixeDe3je7CX24eSliwrSrYnHkzjXAusBzoJSK5InIn8DtggoikA+M9z00Lc+uoLrx6+3DyDh1n0gtL7a5NB+UUlXLdX5fx750FPDmpH09M6k9ggK3n3dxJ7RB200hNTdW0tLQm259pGukHjnLHrNXkHynnN9cM4LphtqZ4U1qacZAH566jqrqGv948jHNT4p2OZBqZiKxR1dRTX7c7MU2DpbSL4v37xjCkcww/fGcDP3t/E+VV1U7HavZqapS/LMrg1pkriYsI4YP7x1jxbmGsgJtGER8Zyht3juSeC7oxZ+UepvxtObmHSp2O1WwVl1YyfXYav/90B1cM7MgH94+xaYItkBVw02iCAgP46aV9+Nstw9hdcIzLnv+ajzbudTpWs7NydyGXPr+Yf+8o4NdX9eP5qYOti04LZQXcNLqJ/dvz0ffOpVtCJA+8uY5H5m3gmE01bLDK6hr+sHAHN/59BSFBAbx77zlMOyfZmg+3YPZn2/hEl7gI3vnuaJ7/PJ2//DuDVZlF/P76QQxPjnU6miulHzjKj+dtZH3OYSYPS+Txq/oRaWfdLZ6dgRufCQ4M4EeX9GLu3aOoqlGmvLScx+dvobTCzsa9VVVdw18WZXD5n5aQVXiMP984hGeuH2TF2wB2Bm6awKhucXz68Pk8vWA7ry3L4ovtB/jfqwdwfk+7M/fbbM4r5qfvbWJTXjGX9m/PE5P6kxAV6nQs40dsHrhpUit3F/Loe5vIPHiMywa05+eX96VjTCunY/mV4tJKnlm4gzdWZhMXEcKvr+rP5QM7OB3LOOh088CtgJsmV15Vzd8X7+aFRRkIwgNje3DnuV1b/G3f1TXKO2k5/P7THRwqreC20cl8f0JPW8vEWAE3/ienqJQnP9rKwq0H6BAdxg8m9OTaoYkt7hZwVWXRjnx+98l2dh4oYWjnGJ6Y1N+65pj/sAJu/NaK3YX89uNtbMgtpnf7KB4al8Il/doT0MwLuaqyYncRf/x8Jyszi0iOC+cnE3szsX97mxpo/osVcOPXVJV/bdrHHxbuJPPgMXq2i+T+i3pwxcCOze6MXFX5Ov0gf/4yndVZh0iICuXBsT24cURnggNtYpj5JivgxhWqa5SPNu7lhS8zSM8vISm2FdNGJ3N9apLrx4LLKquZv34vry7LYtu+I3SIDuPeC7szJTWpxY//m29nBdy4Sk2NsnDrfl5ZksWqrCJaBQdyzdBOTElNYlBitKuGGDLyS5i3Jpe3V+/hUGklvdtH8Z0xyVw9pBOhQVa4Tf2sgBvX2rK3mNeWZjF/w17Kq2ronhDB5GFJXDGwA0mx4U7Hq1PB0XIWbN7HvLV5bMg5TGCAMK53W74zpiujusW66g+QcZ4VcON6R8oq+XjjPuatySUt+xAAfTq05pJ+7Rjfpx19O7R27MKnqrL74DG+3JbPp1v2s2bPIVShV7sork9N5KrBHWkbZX0pzdnxSQEXkYnA80Ag8LKqfmtnHivgprFkFx5j4ZYD/1UsY8KDGd0tjtHd4xiUGEPvDlE+G6Koqq4ho6CEjTnFrNhdyLJdhew/Ugb8/x+Vi/u2p0+HKDvbNg3W6AVcRAKBncAEIBdYDdyoqltP9zVWwI0vFBwtZ0lGAcsyagvpif6cQQFCr/ZR9GoXRZe4CJLjw+kcG05CVCjxkaH1Xjgsr6qm6FgFB49WkHuolMzCY2QdPEZ6fgnb9h2hrLIGgLiIEEZ3j+Oc7vGclxLvt8M6xr18UcBHA4+r6iWe5z8FUNXfnu5rrIAbX1NVcg8dZ3NeMRvzitmcV8yu/BL2Fpd9Y9vwkEDCQwIJCQwgNDgQAcqraiivqqGsspqSOpbAjY8MpVtCBP07RjMwMZr+naLpnhBhZ9nGp05XwBuymFUnIOek57nAyDp2PB2YDtC5c+cG7M6Y+okISbHhJMWGc+mA/18/pKyymuzCUnIPlVJ4rIKDJeUUlVRwvLKa8qoaKqpqqFElNCiQ0OAAQoMCiIsIIS4ylNiIEDrFtKJLXDhRYe6eymiaF5+vRqiqM4AZUHsG7uv9GVOXsODA2uGU9lFORzGm0TTktq88IOmk54me14wxxjSBhhTw1UCKiHQVkRBgKjC/cWIZY4ypz1kPoahqlYg8AHxK7TTCV1R1S6MlM8YY860aNAauqh8DHzdSFmOMMWfAlj4zxhiXsgJujDEuZQXcGGNcygq4Mca4VJOuRigiBUB2k+3w7MQDB50O4QW35AT3ZHVLTnBPVrfkBP/O2kVVE059sUkLuBuISFpdaw74G7fkBPdkdUtOcE9Wt+QEd2U9wYZQjDHGpayAG2OMS1kB/6YZTgfwkltygnuyuiUnuCerW3KCu7ICNgZujDGuZWfgxhjjUlbAjTHGpVp0AReRt0VkvecjS0TWn2a7LBHZ5NnOkZ5wIvK4iOSdlPey02w3UUR2iEiGiDzqQM7fi8h2EdkoIu+LSMxptnPsmNZ3jEQk1PO7kSEiK0UkuSnzeTIkicgiEdkqIltE5KE6trlQRIpP+p34ZVPnPCnLt/48pdafPMd0o4gMdSBjr5OO1XoROSIiD5+yjd8cU6+oqn3UXgf4A/DL07yXBcQ7nO9x4Ef1bBMI7AK6ASHABqBvE+e8GAjyPH4KeMqfjqk3xwi4D/ib5/FU4G0HcnYAhnoeR1HbQPzUnBcCHzV1trP5eQKXAZ8AAowCVjqcNxDYT+0NMn55TL35aNFn4CdIbUfaKcBcp7M00AggQ1V3q2oF8BYwqSkDqOpCVT3RDXgFtZ2a/Ik3x2gSMMvzeB4wTpq4a7Gq7lPVtZ7HR4Ft1PahdatJwOtaawUQIyId6vsiHxoH7FJVf78z/FtZAa91HnBAVdNP874CC0VkjadJs1Me8Pz38xURaVPH+3U1mnbyH/0d1J511cWpY+rNMfrPNp4/RsVAXJOkq4NnCGcIsLKOt0eLyAYR+URE+jVpsP9W38/T3343p3L6EzZ/Oab18nlTY6eJyOdA+zre+pmqfuh5fCPffvZ9rqrmiUhb4DMR2a6qi5syK/Ai8CS1/1CepHbI547GzuANb46piPwMqALmnObbNMkxdTsRiQTeBR5W1SOnvL2W2iGAEs81kQ+AlKbO6OGan6enBeRVwE/reNufjmm9mn0BV9Xx3/a+iAQB1wLDvuV75Hk+54vI+9T+N7zRfznry3qCiPwd+KiOt5qk0bQXx/R24ApgnHoGFuv4Hk1yTOvgzTE6sU2u5/cjGihsgmz/RUSCqS3ec1T1vVPfP7mgq+rHIvJXEYlX1SZfkMmLn6c/NUG/FFirqgdOfcOfjqk3bAgFxgPbVTW3rjdFJEJEok48pvYi3eYmzHcix8njhdecJoPjjaZFZCLwCHCVqpaeZhsnj6k3x2g+MM3zeDLw5en+EPmKZ8x9JrBNVZ89zTbtT4zNi8gIav89O/GHxpuf53zgNs9slFFAsarua+KoJ5z2f9z+cky91ezPwL3wjbEwEekIvKyqlwHtgPc9P9Mg4E1VXdDkKeFpERlM7RBKFnDPqVnVPxpNvwCEUvvfaIAVqvpdfzmmpztGIvIEkKaq86ktnLNFJAMoovZ3pKmNAW4FNsn/T299DOgMoKp/o/aPy70iUgUcB6Y29R8ajzp/niLy3ZOyfkztTJQMoBT4jgM5T/yBmYDn34/ntZNz+ssx9YrdSm+MMS5lQyjGGONSVsCNMcalrIAbY4xLWQE3xhiXsgJujDEuZQXcGGNcygq4Mca41P8BiXxz4+WbQRgAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nr-tjGXUYchz"
      },
      "source": [
        "def dfunc(f, x):\n",
        "    h = 1e-4 \n",
        "    return (f(x+h) - f(x)) / (h)"
      ],
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZcSwl9vDYeYi",
        "outputId": "298bb697-07b4-49f7-abed-ce74b3e5ca50"
      },
      "source": [
        "func(3)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6QuCjCIiYgec",
        "outputId": "0b5aeed1-51db-4a00-e0f5-52479e55ba00"
      },
      "source": [
        "dfunc(func, 3)"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6.000100000012054"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "NVsRvaRvYiUs",
        "outputId": "bf9fbedc-3d37-44ba-e357-e59231802473"
      },
      "source": [
        "# 切線函數\n",
        "def tfunc(f, x, t):\n",
        "    d = dfunc(f, x)\n",
        "    y = f(x) - d*x\n",
        "    return d*t + y\n",
        "\n",
        "# 繪製 x ^ 2 \n",
        "x = np.arange(-6, 6, 0.01)\n",
        "y = func(x)\n",
        "plt.plot(x, y)\n",
        "\n",
        "# 繪製x = 3 時的切線\n",
        "x2 = np.arange(0, 5, 0.01)\n",
        "y2 = tfunc(func, 3, x2)\n",
        "plt.plot(x2, y2)\n"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f7b5afca748>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FPn-woY-YpFj"
      },
      "source": [
        "def dfunc(f, x):\n",
        "    h = 1e-4 \n",
        "    return (f(x+h) - f(x-h)) / (2*h)\n"
      ],
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VoTNIrhYY6MP",
        "outputId": "59dbe8b8-7d78-4812-a760-6396139e6dae"
      },
      "source": [
        "dfunc(func, 3)"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "6.000000000012662"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_pisSmGTY7of"
      },
      "source": [
        "# intial: 3 , gradient: 6, learning rate: 0.8\n",
        "# step : 3 - 0.8 * 6 = -1.8, gradient: -3.6\n",
        "# step : -1.8 - (-3.6 * 0.8)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "laRo9dTXZvFr"
      },
      "source": [
        "def gradient_descent(func, init_x, lr = 0.3, epochs = 100):    \n",
        "    x = init_x       \n",
        "    res = [x]\n",
        "    for i in range(epochs):         \n",
        "        grad = dfunc(func,x)        \n",
        "        x = x - grad * lr\n",
        "        res.append(x)\n",
        "    return np.array(res)"
      ],
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XSTW1v8sZ0Tk",
        "outputId": "ab47702e-69a6-4c3e-acc9-037cca548192"
      },
      "source": [
        "x = gradient_descent(func, 3, lr=0.8)\n",
        "x"
      ],
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 3.00000000e+00, -1.80000000e+00,  1.08000000e+00, -6.48000000e-01,\n",
              "        3.88800000e-01, -2.33280000e-01,  1.39968000e-01, -8.39808000e-02,\n",
              "        5.03884800e-02, -3.02330880e-02,  1.81398528e-02, -1.08839117e-02,\n",
              "        6.53034701e-03, -3.91820820e-03,  2.35092492e-03, -1.41055495e-03,\n",
              "        8.46332972e-04, -5.07799783e-04,  3.04679870e-04, -1.82807922e-04,\n",
              "        1.09684753e-04, -6.58108519e-05,  3.94865112e-05, -2.36919067e-05,\n",
              "        1.42151440e-05, -8.52908641e-06,  5.11745185e-06, -3.07047111e-06,\n",
              "        1.84228266e-06, -1.10536960e-06,  6.63221759e-07, -3.97933055e-07,\n",
              "        2.38759833e-07, -1.43255900e-07,  8.59535400e-08, -5.15721240e-08,\n",
              "        3.09432744e-08, -1.85659646e-08,  1.11395788e-08, -6.68374727e-09,\n",
              "        4.01024836e-09, -2.40614902e-09,  1.44368941e-09, -8.66213646e-10,\n",
              "        5.19728188e-10, -3.11836913e-10,  1.87102148e-10, -1.12261289e-10,\n",
              "        6.73567731e-11, -4.04140639e-11,  2.42484383e-11, -1.45490630e-11,\n",
              "        8.72943780e-12, -5.23766268e-12,  3.14259761e-12, -1.88555858e-12,\n",
              "        1.13133514e-12, -6.78801083e-13,  4.07280642e-13, -2.44368387e-13,\n",
              "        1.46621023e-13, -8.79726202e-14,  5.27835761e-14, -3.16701351e-14,\n",
              "        1.90020956e-14, -1.14012587e-14,  6.84074063e-15, -4.10445368e-15,\n",
              "        2.46266555e-15, -1.47760599e-15,  8.86562228e-16, -5.31946644e-16,\n",
              "        3.19169267e-16, -1.91512191e-16,  1.14908596e-16, -6.89504936e-17,\n",
              "        4.13821652e-17, -2.48187536e-17,  1.48792984e-17, -8.93026836e-18,\n",
              "        5.35679518e-18, -3.21279597e-18,  1.92895872e-18, -1.14815316e-18,\n",
              "        6.98113967e-19, -4.33469111e-19,  2.61362604e-19, -1.55536425e-19,\n",
              "        8.26915915e-20, -4.30398616e-20,  1.65171425e-20, -3.33519223e-21,\n",
              "       -3.33519223e-21, -3.33519223e-21, -3.33519223e-21, -3.33519223e-21,\n",
              "       -3.33519223e-21, -3.33519223e-21, -3.33519223e-21, -3.33519223e-21,\n",
              "       -3.33519223e-21])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "Tf2mQHc_aAE6",
        "outputId": "5186026a-9d83-473b-c8e4-4091b0c87116"
      },
      "source": [
        "\n",
        "t = np.arange(-6.0, 6.0, 0.01)\n",
        "plt.plot(t, func(t), c='b')\n",
        "plt.plot(x, func(x), c='r')    \n",
        "plt.scatter(x, func(x), c='r')    \n"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f7b5ab9ac18>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 72
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "NtBJJHqxaBLP",
        "outputId": "cbe6fc0e-9b47-4c73-e3e6-dd042439da2e"
      },
      "source": [
        "x = gradient_descent(func, 3, lr=1)\n",
        "t = np.arange(-6.0, 6.0, 0.01)\n",
        "plt.plot(t, func(t), c='b')\n",
        "plt.plot(x, func(x), c='r')    \n",
        "plt.scatter(x, func(x), c='r')  "
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f7b5ab06ba8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 73
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "1fj87wrwaSbf",
        "outputId": "ad1d3291-2df7-474a-eff5-247da646b3c5"
      },
      "source": [
        "x = gradient_descent(func, 3, lr=0.001)\n",
        "t = np.arange(-6.0, 6.0, 0.01)\n",
        "plt.plot(t, func(t), c='b')\n",
        "plt.plot(x, func(x), c='r')    \n",
        "plt.scatter(x, func(x), c='r')  "
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f7b5aaf31d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wEnQ5Kc2aU9H",
        "outputId": "86e38391-6d43-48a5-9ece-37ab625f6a59"
      },
      "source": [
        "# 初始網路\n",
        "x  = np.array([0.6, 0.9])\n",
        "\n",
        "np.random.seed(42)\n",
        "weight   = np.random.randn(2,3)\n",
        "z        = np.dot(x, weight)\n",
        "\n",
        "# 取得預測值 y_hat\n",
        "y_hat    = softmax_function(z)\n",
        "y_hat"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.73222998, 0.10288817, 0.16488185])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 75
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4rqa1iFCbAjh"
      },
      "source": [
        "# 計算代價(損失)\n",
        "y        = np.array([0, 0, 1])\n"
      ],
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8aWXbQQ8bGrS",
        "outputId": "aae8cbca-81a3-4efe-cf72-b8955fad50e4"
      },
      "source": [
        "def cross_entropy_err(y_hat, y):\n",
        "    delta = 1e-8\n",
        "    return -np.sum(y*np.log(y_hat + delta))\n",
        "\n",
        "cross_entropy_err(y_hat, y)"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1.8025260716416405"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y6WrqU3CbKi_"
      },
      "source": [
        "def dfunc(f, x):\n",
        "    h = 1e-4 \n",
        "    grad = np.zeros_like(x)\n",
        "    it = np.nditer(x, flags=['multi_index'])\n",
        "    while not it.finished:     \n",
        "        idx = it.multi_index\n",
        "        tmp_val = x[idx]\n",
        "        x[idx] = float(tmp_val) + h\n",
        "        fxh1 = f(x) # f(x+h)\n",
        "        x[idx] = tmp_val - h \n",
        "        fxh2 = f(x) # f(x-h)\n",
        "        grad[idx] = (fxh1 - fxh2) / (2*h)\n",
        "        x[idx] = tmp_val\n",
        "        it.iternext()   \n",
        "    return grad\n"
      ],
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MC2kVs_6bZM2",
        "outputId": "fbb04f4b-754e-4392-908b-0f9ec6da1176"
      },
      "source": [
        "def predict(x):\n",
        "    return np.dot(x, weight)\n",
        "\n",
        "def loss(x, y):\n",
        "    z = predict(x)\n",
        "    y_hat = softmax_function(z)\n",
        "    loss = cross_entropy_err(y_hat, y)\n",
        "    return loss\n",
        "\n",
        "func = lambda w: loss(x, y)\n",
        "dfunc(func, weight)\n"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.43933796,  0.0617329 , -0.50107086],\n",
              "       [ 0.65900694,  0.09259935, -0.75160629]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "06BTzt0Wdag7"
      },
      "source": [
        "## Training Neural Network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GOKICOW3bk1z"
      },
      "source": [
        "def softmax_function(x):\n",
        "    return np.exp(x) / np.sum(np.exp(x))\n",
        "\n",
        "def sigmoid_function(x):\n",
        "    return 1/ (1 + np.exp(-x))\n",
        "\n",
        "def dfunc(f, x):\n",
        "    h = 1e-4 \n",
        "    grad = np.zeros_like(x)\n",
        "    it = np.nditer(x, flags=['multi_index'])\n",
        "    while not it.finished:     \n",
        "        idx = it.multi_index\n",
        "        tmp_val = x[idx]\n",
        "        x[idx] = float(tmp_val) + h\n",
        "        fxh1 = f(x) # f(x+h)\n",
        "        x[idx] = tmp_val - h \n",
        "        fxh2 = f(x) # f(x-h)\n",
        "        grad[idx] = (fxh1 - fxh2) / (2*h)\n",
        "        x[idx] = tmp_val\n",
        "        it.iternext()   \n",
        "    return grad\n",
        "\n",
        "def cross_entropy_err(y_hat, y):\n",
        "    delta = 1e-8\n",
        "    return -np.sum(y*np.log(y_hat + delta))"
      ],
      "execution_count": 109,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3YTh_vL0eI19"
      },
      "source": [
        "class ANN:\n",
        "    def __init__(self, input_size, hidden_size, output_size, weight_init_std=0.01):\n",
        "        self.params = {}\n",
        "        self.params['W1'] = weight_init_std * np.random.randn(input_size, hidden_size)\n",
        "        self.params['b1'] = np.zeros(hidden_size)\n",
        "        self.params['W2'] = weight_init_std * np.random.randn(hidden_size, output_size)\n",
        "        self.params['b2'] = np.zeros(output_size)\n",
        " \n",
        "    def predict(self, x):\n",
        "        W1, W2 = self.params['W1'], self.params['W2']\n",
        "        b1, b2 = self.params['b1'], self.params['b2']\n",
        "        a1 = np.dot(x, W1) + b1\n",
        "        z1 = sigmoid_function(a1)\n",
        "        a2 = np.dot(z1, W2) + b2\n",
        "        y = softmax_function(a2)\n",
        "        return y\n",
        "        \n",
        "    def loss(self, x, y):\n",
        "        y_hat = self.predict(x)\n",
        "        return cross_entropy_err(y_hat, y)\n",
        "        \n",
        "    def numerical_gradient(self, x, y):\n",
        "        loss_W = lambda W: self.loss(x, y)\n",
        "        grads = {}\n",
        "        grads['W1'] = dfunc(loss_W, self.params['W1'])\n",
        "        grads['b1'] = dfunc(loss_W, self.params['b1'])\n",
        "        grads['W2'] = dfunc(loss_W, self.params['W2'])\n",
        "        grads['b2'] = dfunc(loss_W, self.params['b2'])\n",
        "        return grads"
      ],
      "execution_count": 110,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nAEQvONeedqJ"
      },
      "source": [
        "net = ANN(input_size=4, hidden_size=5, output_size=3)"
      ],
      "execution_count": 111,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NZcObiNIeqkP"
      },
      "source": [
        "from sklearn.datasets import load_iris\n",
        "iris = load_iris()\n",
        "x= iris.data"
      ],
      "execution_count": 112,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1_XggWemexHu",
        "outputId": "cef51cd9-a9ba-4954-e583-73e76f512a93"
      },
      "source": [
        "iris.target"
      ],
      "execution_count": 100,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
              "       0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
              "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
              "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 100
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PpvJdspve602"
      },
      "source": [
        "y = np.zeros((len(iris.target), 3))\n",
        "for idx, val in enumerate(iris.target):\n",
        "    y[idx, val] = 1"
      ],
      "execution_count": 113,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fMT5nLghfBbN",
        "outputId": "1bf74c30-c076-4a9c-ced0-648f8f56341b"
      },
      "source": [
        "y[0:3]"
      ],
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1., 0., 0.],\n",
              "       [1., 0., 0.],\n",
              "       [1., 0., 0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 102
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YmtFEw_MfCNF",
        "outputId": "48b30272-043a-4183-f58c-abcd5b8b49db"
      },
      "source": [
        "net.predict(x)[0:3]"
      ],
      "execution_count": 103,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.00225455, 0.00220925, 0.0022026 ],\n",
              "       [0.00225454, 0.0022093 , 0.00220259],\n",
              "       [0.00225456, 0.0022093 , 0.0022026 ]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 103
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "plPkspYlfNYE",
        "outputId": "d691aaf1-53c6-4c25-939c-9281f6cc0441"
      },
      "source": [
        "import time\n",
        "a = time.clock()\n",
        "epochs = 3000\n",
        "lr     = 0.01\n",
        "\n",
        "train_loss = []\n",
        "for i in range(epochs):         \n",
        "    grad = net.numerical_gradient(x,y)\n",
        "    for key in ('W1', 'b1', 'W2', 'b2'):\n",
        "        net.params[key] = net.params[key] - lr * grad[key]\n",
        "    loss = net.loss(x, y)\n",
        "    train_loss.append(loss)\n",
        "time.clock() -a"
      ],
      "execution_count": 114,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "23.955813000000006"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 114
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Vdxj_h6Wf6EK",
        "outputId": "bf6ff661-f85f-4d33-82b0-fc87e8d1dead"
      },
      "source": [
        "net.predict(x)[0:3]"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[6.71113882e-03, 1.02078880e-05, 4.82566274e-08],\n",
              "       [6.63676429e-03, 1.03189021e-05, 4.87173747e-08],\n",
              "       [6.71130948e-03, 1.02715294e-05, 4.80402166e-08]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 115
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "s1amZJ6agNzk",
        "outputId": "02c3768c-6753-4501-8b97-aac1c6788bf3"
      },
      "source": [
        "%pylab inline\n",
        "import matplotlib.pyplot as plt\n",
        "plt.scatter(range(0,3000),train_loss)\n",
        "plt.xlabel('Iteration')\n",
        "plt.ylabel('Loss')\n",
        "plt.title('Training Loss')\n",
        "plt.show()"
      ],
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Populating the interactive namespace from numpy and matplotlib\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x5C3qVpdgpiA"
      },
      "source": [
        "predicted = np.argmax(net.predict(x), axis = 1)"
      ],
      "execution_count": 119,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JmPdgwp2htlm",
        "outputId": "e06de1c4-ee4f-4da8-9d7e-fa496359c3f9"
      },
      "source": [
        "from sklearn.metrics import accuracy_score, confusion_matrix\n",
        "accuracy_score(iris.target, predicted)"
      ],
      "execution_count": 120,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9733333333333334"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 120
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "90kconQph9CH",
        "outputId": "ac1a189a-bcd6-4785-b0ad-703e0b0e64b6"
      },
      "source": [
        "confusion_matrix(iris.target, predicted)"
      ],
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[50,  0,  0],\n",
              "       [ 0, 46,  4],\n",
              "       [ 0,  0, 50]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lYi36_QomgTU"
      },
      "source": [
        "## Stochastic Gradient Descent"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6vG0Iu1Mh-Jk"
      },
      "source": [
        "def cross_entropy_err(y_hat, y):\n",
        "    y      = y.reshape(1, y.size)\n",
        "    y_hat  = y_hat.reshape(1, y_hat.size)\n",
        "    batch_size = y_hat.shape[0]\n",
        "    return -np.sum(y * np.log(y_hat)) / batch_size"
      ],
      "execution_count": 122,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u4n-QOrzmnm2",
        "outputId": "6e127728-3fae-4cdc-a4c9-b357901f5e20"
      },
      "source": [
        "net = ANN(input_size=4, hidden_size=5, output_size=3)\n",
        "\n",
        "epochs     = 3000\n",
        "lr         = 0.01\n",
        "batch_size = 30\n",
        "\n",
        "a = time.clock()\n",
        "train_loss = []\n",
        "for i in range(epochs): \n",
        "    idx = np.random.choice(iris.data.shape[0], batch_size)\n",
        "    x_batch   = iris.data[idx]\n",
        "    y_batch   = y[idx]\n",
        "    grad = net.numerical_gradient(x_batch,y_batch)\n",
        "    for key in ('W1', 'b1', 'W2', 'b2'):\n",
        "        net.params[key] = net.params[key] - lr * grad[key]\n",
        "    loss = net.loss(x_batch, y_batch)\n",
        "    train_loss.append(loss)\n",
        "time.clock() - a"
      ],
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "12.731513000000007"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 123
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        },
        "id": "IUWo1Tdum5xj",
        "outputId": "23c37701-fc4c-4a1f-b0a2-7f566017a043"
      },
      "source": [
        "plt.scatter(range(0,3000),train_loss)\n",
        "plt.xlabel('Iteration')\n",
        "plt.ylabel('Loss')\n",
        "plt.title('Training Loss')"
      ],
      "execution_count": 124,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Training Loss')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 124
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QstCMvxInDRK"
      },
      "source": [
        "predicted = np.argmax(net.predict(x), axis = 1)"
      ],
      "execution_count": 125,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-6NwAGGWnhRd",
        "outputId": "dd497f64-01d6-4c37-8d36-e9573862930b"
      },
      "source": [
        "accuracy_score(iris.target, predicted)"
      ],
      "execution_count": 126,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.98"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 126
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "narw6pxOnk3f",
        "outputId": "072c900e-169c-4067-dea2-8b13e8f909bd"
      },
      "source": [
        "confusion_matrix(iris.target, predicted)"
      ],
      "execution_count": 127,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[50,  0,  0],\n",
              "       [ 0, 48,  2],\n",
              "       [ 0,  1, 49]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 127
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "s7H9n9DeqR_T"
      },
      "source": [
        "## Backward Propogation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XTvsM0q6qRRm"
      },
      "source": [
        "class Relu:\n",
        "    def __init__(self):\n",
        "        self.cache = None\n",
        "    \n",
        "    def forward(self, x):\n",
        "        self.cache = (x <=0)\n",
        "        out = np.maximum(0,x)\n",
        "        return out\n",
        "\n",
        "    def backward(self, dout):\n",
        "        dout[self.cache] = 0\n",
        "        dx = dout\n",
        "        return dx"
      ],
      "execution_count": 128,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m8n6KU5cqXBW"
      },
      "source": [
        "import numpy as np\n",
        "x= np.array([1.0 , -0.5])\n",
        "layer = Relu()"
      ],
      "execution_count": 129,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "imP5DXusqc4W",
        "outputId": "29f46c3d-f0fd-41d7-87a3-01e1cee9dc97"
      },
      "source": [
        "layer.forward(x)"
      ],
      "execution_count": 130,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1., 0.])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 130
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KdP6yT-6qepm",
        "outputId": "cf3fd32a-e7d4-46a6-d791-b70bbd1de914"
      },
      "source": [
        "layer.cache"
      ],
      "execution_count": 131,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([False,  True])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 131
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3SKfT7qfqjBN",
        "outputId": "9d43f99f-037d-4ebc-f72b-54e0d8e06395"
      },
      "source": [
        "layer.backward(np.array([1,1]))"
      ],
      "execution_count": 133,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([1, 0])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 133
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yz8f1N8bqrZj"
      },
      "source": [
        "class Sigmoid:\n",
        "    def __init__(self):\n",
        "        self.out = None\n",
        "\n",
        "    def forward(self, x):\n",
        "        out = 1/ (1 + np.exp(-x))\n",
        "        self.out = out\n",
        "        return out\n",
        "\n",
        "    def backward(self, dout):\n",
        "        y  = self.out\n",
        "        dx = dout * y * (1-y)\n",
        "        return dx"
      ],
      "execution_count": 134,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fgRJjE77sQ2J",
        "outputId": "8e3b24b6-3268-44df-efe3-e193f9a0d570"
      },
      "source": [
        "x"
      ],
      "execution_count": 138,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 1. , -0.5])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 138
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N4eA9Jgdr4bb",
        "outputId": "f3730185-ac9d-4421-b919-5f922360d231"
      },
      "source": [
        "layer = Sigmoid()\n",
        "layer.forward(x)"
      ],
      "execution_count": 135,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.73105858, 0.37754067])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 135
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yNPISDuIsDsA",
        "outputId": "97a60298-b2d7-4f33-975b-03ee957f3485"
      },
      "source": [
        "0.37754067 * (1- 0.37754067)"
      ],
      "execution_count": 137,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.23500371249595112"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 137
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-OutPYXYsX_M",
        "outputId": "bea98888-4344-43aa-9127-c5a1bd195a71"
      },
      "source": [
        "0.73105858 * (1- 0.73105858)"
      ],
      "execution_count": 139,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.19661193260838358"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 139
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cSGAkG8or8eY",
        "outputId": "0cb7f2ed-3b15-4d66-c6fc-d2b534f98c08"
      },
      "source": [
        "layer.backward(x)"
      ],
      "execution_count": 136,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([ 0.19661193, -0.11750186])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 136
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DhtxvL0_r_L_"
      },
      "source": [
        "class Affine:\n",
        "    def __init__(self, W, b):\n",
        "        self.W = W\n",
        "        self.b = b\n",
        "        self.x = None\n",
        "        self.dW = None\n",
        "        self.db = None\n",
        "\n",
        "    def forward(self, x):\n",
        "        self.x = x\n",
        "        out = np.dot(self.x, self.W) + self.b\n",
        "        return out\n",
        "\n",
        "    def backward(self, dout):\n",
        "        dx = np.dot(dout, self.W.T)\n",
        "        self.dW = np.dot(self.x.T, dout)\n",
        "        self.db = np.sum(dout)\n",
        "        return dx"
      ],
      "execution_count": 140,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2tVXD_X2sop6"
      },
      "source": [
        "class SoftmaxWithLoss:\n",
        "    def __init__(self):\n",
        "        self.loss = None\n",
        "        self.y_hat = None \n",
        "        self.y = None \n",
        "\n",
        "    def forward(self, x, y):\n",
        "        self.y = y\n",
        "        self.y_hat = softmax_function(x)\n",
        "        self.loss = cross_entropy_err(self.y_hat, self.y)\n",
        "        return self.loss\n",
        "\n",
        "    def backward(self, dout=1):\n",
        "        batch_size = self.y.shape[0]\n",
        "        dx = (self.y_hat - self.y) / batch_size\n",
        "        return dx"
      ],
      "execution_count": 141,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bx8tFP-Jsxbw"
      },
      "source": [
        "\n",
        "def softmax_function(x):\n",
        "    if x.ndim == 2:\n",
        "        x = x.T\n",
        "        x = x - np.max(x, axis=0)\n",
        "        y = np.exp(x) / np.sum(np.exp(x), axis=0)\n",
        "        return y.T \n",
        "\n",
        "    x = x - np.max(x)\n",
        "    return np.exp(x) / np.sum(np.exp(x))"
      ],
      "execution_count": 142,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3UXZi3fptBm2"
      },
      "source": [
        "\n",
        "def cross_entropy_err(y_hat, y):\n",
        "    y      = y.reshape(1, y.size)\n",
        "    y_hat  = y_hat.reshape(1, y_hat.size)\n",
        "    batch_size = y_hat.shape[0]\n",
        "    return -np.sum(y * np.log(y_hat)) / batch_size"
      ],
      "execution_count": 143,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1CVui-_GtDIG"
      },
      "source": [
        "import numpy as np\n",
        "from collections import OrderedDict\n",
        "\n",
        "class ANN:\n",
        "\n",
        "    def __init__(self, input_size, hidden_size, output_size, weight_init_std = 0.01):\n",
        "        self.params = {}\n",
        "        self.params['W1'] = weight_init_std * np.random.randn(input_size, hidden_size)\n",
        "        self.params['b1'] = np.zeros(hidden_size)\n",
        "        self.params['W2'] = weight_init_std * np.random.randn(hidden_size, output_size) \n",
        "        self.params['b2'] = np.zeros(output_size)\n",
        "\n",
        "        self.layers = OrderedDict()\n",
        "        self.layers['Affine1'] = Affine(self.params['W1'], self.params['b1'])\n",
        "        self.layers['Sigmoid1'] = Sigmoid()\n",
        "        self.layers['Affine2'] = Affine(self.params['W2'], self.params['b2'])\n",
        "\n",
        "        self.lastLayer = SoftmaxWithLoss()\n",
        "        \n",
        "    def predict(self, x):\n",
        "        for layer in self.layers.values():\n",
        "            x = layer.forward(x)\n",
        "        return x\n",
        "        \n",
        "    def loss(self, x, y):\n",
        "        y_hat = self.predict(x)\n",
        "        return self.lastLayer.forward(y_hat, y)\n",
        "  \n",
        "    def gradient(self, x, y):\n",
        "        # forward\n",
        "        self.loss(x, y)\n",
        "\n",
        "        # backward\n",
        "        dout = 1\n",
        "        dout = self.lastLayer.backward(dout)\n",
        "        \n",
        "        layers = list(self.layers.values())\n",
        "        layers.reverse()\n",
        "        for layer in layers:\n",
        "            dout = layer.backward(dout)\n",
        "\n",
        "        grads = {}\n",
        "        grads['W1'], grads['b1'] = self.layers['Affine1'].dW, self.layers['Affine1'].db\n",
        "        grads['W2'], grads['b2'] = self.layers['Affine2'].dW, self.layers['Affine2'].db\n",
        "\n",
        "        return grads"
      ],
      "execution_count": 144,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JvgHRXaXtYPL"
      },
      "source": [
        "\n",
        "from sklearn.datasets import load_iris\n",
        "iris = load_iris()\n",
        "\n",
        "y = np.zeros((len(iris.target), 3))\n",
        "for idx, val in enumerate(iris.target):\n",
        "    y[idx, val] = 1"
      ],
      "execution_count": 145,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "557MukjitagS",
        "outputId": "17b28d77-c7c2-4760-9765-f7867458082a"
      },
      "source": [
        "import time\n",
        "a = time.clock()\n",
        "epochs     = 3000\n",
        "lr         = 0.01\n",
        "batch_size = 30\n",
        "\n",
        "net = ANN(input_size=4, hidden_size=5, output_size=3)\n",
        "\n",
        "train_loss = []\n",
        "for i in range(epochs): \n",
        "    idx = np.random.choice(iris.data.shape[0], batch_size)\n",
        "    x_batch   = iris.data[idx]\n",
        "    y_batch   = y[idx]\n",
        "    grad = net.gradient(x_batch,y_batch)\n",
        "    for key in ('W1', 'b1', 'W2', 'b2'):\n",
        "        net.params[key] -= lr * grad[key]\n",
        "    loss = net.loss(x_batch, y_batch)\n",
        "    train_loss.append(loss)\n",
        "    \n",
        "time.clock() - a"
      ],
      "execution_count": 146,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7028819999999882"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 146
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t3AwfEVfwkul"
      },
      "source": [
        "## Sklearn On IRIS"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u29i4mIotmdp"
      },
      "source": [
        "from sklearn.datasets import load_iris\n",
        "iris = load_iris()\n",
        "X = iris.data\n",
        "y = iris.target"
      ],
      "execution_count": 147,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QkfPC5HYwvnG",
        "outputId": "7ca3e107-d149-4995-a916-2019e5d88034"
      },
      "source": [
        "from sklearn.neural_network import MLPClassifier\n",
        "#?MLPClassifier((5,))\n",
        "clf = MLPClassifier((5,), max_iter  = 3000, activation = 'logistic')\n",
        "clf.fit(X, y)"
      ],
      "execution_count": 159,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "MLPClassifier(activation='logistic', alpha=0.0001, batch_size='auto',\n",
              "              beta_1=0.9, beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
              "              hidden_layer_sizes=(5,), learning_rate='constant',\n",
              "              learning_rate_init=0.001, max_fun=15000, max_iter=3000,\n",
              "              momentum=0.9, n_iter_no_change=10, nesterovs_momentum=True,\n",
              "              power_t=0.5, random_state=None, shuffle=True, solver='adam',\n",
              "              tol=0.0001, validation_fraction=0.1, verbose=False,\n",
              "              warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 159
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BYBG0gBiw4T1"
      },
      "source": [
        "predicted = clf.predict(X)"
      ],
      "execution_count": 160,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DurR49e8xDED",
        "outputId": "c48bee9e-732c-4549-8d08-97687caecc35"
      },
      "source": [
        "accuracy_score(y, predicted)"
      ],
      "execution_count": 161,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.98"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 161
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r8rMzSxwxdep",
        "outputId": "257017a3-6d02-4646-d24e-02ee85a56edb"
      },
      "source": [
        "confusion_matrix(y, predicted)"
      ],
      "execution_count": 162,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[50,  0,  0],\n",
              "       [ 0, 47,  3],\n",
              "       [ 0,  0, 50]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 162
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LnTPpJMkx-U1"
      },
      "source": [
        "## MNIST"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bCxXMSOcxqRf"
      },
      "source": [
        "from sklearn.datasets import load_digits\n",
        "digits = load_digits()"
      ],
      "execution_count": 163,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xPQpULb_yDdD",
        "outputId": "f45449de-0859-4152-fb1b-40d023ff5a82"
      },
      "source": [
        "digits.data"
      ],
      "execution_count": 164,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.,  0.,  5., ...,  0.,  0.,  0.],\n",
              "       [ 0.,  0.,  0., ..., 10.,  0.,  0.],\n",
              "       [ 0.,  0.,  0., ..., 16.,  9.,  0.],\n",
              "       ...,\n",
              "       [ 0.,  0.,  1., ...,  6.,  0.,  0.],\n",
              "       [ 0.,  0.,  2., ..., 12.,  0.,  0.],\n",
              "       [ 0.,  0., 10., ..., 12.,  1.,  0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 164
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "31Qhur-OyF5z",
        "outputId": "d231328a-7cf0-406d-c109-ba3de5d2055f"
      },
      "source": [
        "digits.data.shape"
      ],
      "execution_count": 165,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1797, 64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 165
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JOfBjvJ5yIDT",
        "outputId": "adcf3b2b-772b-41cd-cded-bd3420f46329"
      },
      "source": [
        "digits.data[0].reshape(8,8)"
      ],
      "execution_count": 167,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[ 0.,  0.,  5., 13.,  9.,  1.,  0.,  0.],\n",
              "       [ 0.,  0., 13., 15., 10., 15.,  5.,  0.],\n",
              "       [ 0.,  3., 15.,  2.,  0., 11.,  8.,  0.],\n",
              "       [ 0.,  4., 12.,  0.,  0.,  8.,  8.,  0.],\n",
              "       [ 0.,  5.,  8.,  0.,  0.,  9.,  8.,  0.],\n",
              "       [ 0.,  4., 11.,  0.,  1., 12.,  7.,  0.],\n",
              "       [ 0.,  2., 14.,  5., 10., 12.,  0.,  0.],\n",
              "       [ 0.,  0.,  6., 13., 10.,  0.,  0.,  0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 167
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "M-myf4j2yQQR",
        "outputId": "d5f3ff88-2c87-482a-a316-29d8fc65f3b9"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "plt.imshow(digits.data[0].reshape(8,8))"
      ],
      "execution_count": 168,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7f7b5a612588>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 168
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAPUAAAD4CAYAAAA0L6C7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAALGUlEQVR4nO3d/6uW9R3H8ddrR81Vplu2Co8sGSXEYlnOIUYwpWErKthYCjUWA2FQFMmiRmPbPxDuhxGI1YJc0qwgWl8Wq2iBM7/kKr8Nk4ZHKo2+C6kn3/vh3ILFsXPd97muz3Wf954PkM6Xm/vzvrGn132uc9/XxxEhAHl8re0BANSLqIFkiBpIhqiBZIgaSGZSE3c6xafEVJ3WxF23anhm2cd0zjnvF1tr/6EZxdaaOnS02FpxdLjYWiV9pkM6Eoc92vcaiXqqTtMPvKSJu27Vez9ZWHS9X69cV2yt3265tthaF9z+drG1ht95t9haJW2Mf5z0ezz9BpIhaiAZogaSIWogGaIGkiFqIBmiBpIhaiAZogaSqRS17aW2d9veY/vOpocC0Lsxo7Y9IOlPkq6UdKGk5bYvbHowAL2pcqReIGlPROyNiCOS1kkq90JhAF2pEvUsSftO+Hyo87UvsL3C9mbbm4/qcF3zAehSbSfKImJ1RMyPiPmTdUpddwugS1Wi3i9p9gmfD3a+BqAPVYl6k6Tzbc+xPUXSMklPNDsWgF6NeZGEiBi2fbOkZyUNSLo/IrY3PhmAnlS68klEPCXpqYZnAVADXlEGJEPUQDJEDSRD1EAyRA0kQ9RAMkQNJNPIDh1ZldwxQ5KWTfug2FqrZnxabK2/bX222FqX/v5XxdaSpJmrNxRdbzQcqYFkiBpIhqiBZIgaSIaogWSIGkiGqIFkiBpIhqiBZIgaSKbKDh332z5g+40SAwEYnypH6j9LWtrwHABqMmbUEfGSpPcLzAKgBrW9S8v2CkkrJGmqTq3rbgF0iW13gGQ4+w0kQ9RAMlV+pfWwpA2S5toesv3L5scC0Ksqe2ktLzEIgHrw9BtIhqiBZIgaSIaogWSIGkiGqIFkiBpIZsJvuzO8+NJiay2btq3YWpJ05dJlxdaa/tquYmv97OUlxdZ6f97nxdaSpJlFVxsdR2ogGaIGkiFqIBmiBpIhaiAZogaSIWogGaIGkiFqIBmiBpKpco2y2bZfsL3D9nbbt5YYDEBvqrz2e1jSyojYanuapC22n4uIHQ3PBqAHVbbdeTsitnY+/kTSTkmzmh4MQG+6epeW7fMkzZO0cZTvse0O0AcqnyizfbqkRyXdFhEff/n7bLsD9IdKUduerJGg10bEY82OBGA8qpz9tqT7JO2MiHuaHwnAeFQ5Ui+SdKOkxba3df78uOG5APSoyrY7L0tygVkA1IBXlAHJEDWQDFEDyRA1kAxRA8kQNZAMUQPJEDWQzITfS+uzM8s9hLsPXFRsLUk6VnB/q5I2vf6dtkdIjSM1kAxRA8kQNZAMUQPJEDWQDFEDyRA1kAxRA8kQNZBMlQsPTrX9iu1/d7bd+UOJwQD0psprLA9LWhwRn3YuFfyy7acj4l8NzwagB1UuPBiSPu18OrnzJ5ocCkDvql7Mf8D2NkkHJD0XEaNuu2N7s+3NR3W47jkBVFQp6oj4PCIuljQoaYHt745yG7bdAfpAV2e/I+JDSS9IWtrMOADGq8rZ77Nsz+h8/HVJV0jK+UZfIIEqZ7/PlfSg7QGN/CPwSEQ82exYAHpV5ez3axrZkxrABMAryoBkiBpIhqiBZIgaSIaogWSIGkiGqIFkiBpIZuJvu/ONcv8urd2wsNhaknSBXim6XimTph8pttbwR1OKrdUvOFIDyRA1kAxRA8kQNZAMUQPJEDWQDFEDyRA1kAxRA8kQNZBM5ag7F/R/1TYXHQT6WDdH6lsl7WxqEAD1qLrtzqCkqyStaXYcAONV9Ui9StIdko6d7AbspQX0hyo7dFwt6UBEbPmq27GXFtAfqhypF0m6xvZbktZJWmz7oUanAtCzMaOOiLsiYjAizpO0TNLzEXFD45MB6Am/pwaS6epyRhHxoqQXG5kEQC04UgPJEDWQDFEDyRA1kAxRA8kQNZAMUQPJTPhtd6Z+cNL3mNTu+xe9WWwtSfqo4FqTzjm72FrXX/iVbyOo1SNPX1ZsrX7BkRpIhqiBZIgaSIaogWSIGkiGqIFkiBpIhqiBZIgaSIaogWQqvUy0cyXRTyR9Lmk4IuY3ORSA3nXz2u8fRsR7jU0CoBY8/QaSqRp1SPq77S22V4x2A7bdAfpD1affl0XEftvfkvSc7V0R8dKJN4iI1ZJWS9IZ/mbUPCeAiiodqSNif+e/ByQ9LmlBk0MB6F2VDfJOsz3t+MeSfiTpjaYHA9CbKk+/z5b0uO3jt/9LRDzT6FQAejZm1BGxV9L3CswCoAb8SgtIhqiBZIgaSIaogWSIGkiGqIFkiBpIZsJvu3PG7nKb0/xu8Mlia0nSz1fcXmytydcdLLZWSXPu2tD2CMVxpAaSIWogGaIGkiFqIBmiBpIhaiAZogaSIWogGaIGkiFqIJlKUdueYXu97V22d9pe2PRgAHpT9bXff5T0TET81PYUSac2OBOAcRgzatvTJV0u6ReSFBFHJB1pdiwAvary9HuOpIOSHrD9qu01net/fwHb7gD9oUrUkyRdIuneiJgn6ZCkO798o4hYHRHzI2L+ZJ1S85gAqqoS9ZCkoYjY2Pl8vUYiB9CHxow6It6RtM/23M6Xlkja0ehUAHpW9ez3LZLWds5875V0U3MjARiPSlFHxDZJ8xueBUANeEUZkAxRA8kQNZAMUQPJEDWQDFEDyRA1kAxRA8lM+L20jr22q9ha19+7sthaknT3yoeLrbXqzSXF1tp08UCxtf4fcaQGkiFqIBmiBpIhaiAZogaSIWogGaIGkiFqIBmiBpIZM2rbc21vO+HPx7ZvKzEcgO6N+TLRiNgt6WJJsj0gab+kxxueC0CPun36vUTSmxHx3yaGATB+3b6hY5mkUd9lYHuFpBWSNJX984DWVD5Sd675fY2kv472fbbdAfpDN0+/r5S0NSLebWoYAOPXTdTLdZKn3gD6R6WoO1vXXiHpsWbHATBeVbfdOSTpzIZnAVADXlEGJEPUQDJEDSRD1EAyRA0kQ9RAMkQNJEPUQDKOiPrv1D4oqdu3Z86U9F7tw/SHrI+Nx9Web0fEWaN9o5Goe2F7c0TMb3uOJmR9bDyu/sTTbyAZogaS6aeoV7c9QIOyPjYeVx/qm5+pAdSjn47UAGpA1EAyfRG17aW2d9veY/vOtuepg+3Ztl+wvcP2dtu3tj1TnWwP2H7V9pNtz1In2zNsr7e9y/ZO2wvbnqlbrf9M3dkg4D8auVzSkKRNkpZHxI5WBxsn2+dKOjcittqeJmmLpOsm+uM6zvbtkuZLOiMirm57nrrYflDSPyNiTecKuqdGxIdtz9WNfjhSL5C0JyL2RsQRSeskXdvyTOMWEW9HxNbOx59I2ilpVrtT1cP2oKSrJK1pe5Y62Z4u6XJJ90lSRByZaEFL/RH1LEn7Tvh8SEn+5z/O9nmS5kna2O4ktVkl6Q5Jx9oepGZzJB2U9EDnR4s1nYtuTij9EHVqtk+X9Kik2yLi47bnGS/bV0s6EBFb2p6lAZMkXSLp3oiYJ+mQpAl3jqcfot4vafYJnw92vjbh2Z6skaDXRkSWyysvknSN7bc08qPSYtsPtTtSbYYkDUXE8WdU6zUS+YTSD1FvknS+7TmdExPLJD3R8kzjZtsa+dlsZ0Tc0/Y8dYmIuyJiMCLO08jf1fMRcUPLY9UiIt6RtM/23M6XlkiacCc2u90gr3YRMWz7ZknPShqQdH9EbG95rDosknSjpNdtb+t87TcR8VSLM2Fst0ha2znA7JV0U8vzdK31X2kBqFc/PP0GUCOiBpIhaiAZogaSIWogGaIGkiFqIJn/ASA9oV0xPR7gAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-jA0PMlxyYfZ",
        "outputId": "edd21490-56a8-4f4e-8f09-a66aa66c00d5"
      },
      "source": [
        "digits.target"
      ],
      "execution_count": 169,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 1, 2, ..., 8, 9, 8])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 169
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UKb7I344yfWv",
        "outputId": "e50b30b8-8a6d-4b44-ad9e-72aeec66b0fa"
      },
      "source": [
        "digits.target[0]"
      ],
      "execution_count": 170,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 170
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 610
        },
        "id": "AEgbS_j4yhLw",
        "outputId": "5669c799-bdfc-48c0-c858-a7e4ab3d3e06"
      },
      "source": [
        "fig = plt.figure(figsize = (8,8))\n",
        "\n",
        "fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)\n",
        "\n",
        "for i in range(36):\n",
        "    ax = fig.add_subplot(6, 6, i+1, xticks=[], yticks=[])\n",
        "    ax.imshow(digits.images[i],cmap=plt.cm.binary,interpolation='nearest')\n",
        "    ax.text(0, 7, str(digits.target[i]), color=\"red\", fontsize = 20)\n",
        "\n"
      ],
      "execution_count": 171,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 36 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gKhyfEGby4XV",
        "outputId": "e20f6971-74a6-47bd-b56b-6786af90a5a2"
      },
      "source": [
        "x = digits.data\n",
        "x.shape"
      ],
      "execution_count": 172,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1797, 64)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 172
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-C5THEFPzA4T",
        "outputId": "ebb532ea-ed81-478e-c7ec-ad98ae5cda2f"
      },
      "source": [
        "y = digits.target\n",
        "y.shape"
      ],
      "execution_count": 173,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1797,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 173
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mh1a4gUgz3dj"
      },
      "source": [
        "x = x / 15.0"
      ],
      "execution_count": 178,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-tdjwZehz2rC",
        "outputId": "6230e30e-f7d9-4094-e2ad-5f360dc319e6"
      },
      "source": [
        "from sklearn.neural_network import MLPClassifier\n",
        "clf = MLPClassifier((32,32,32,), max_iter = 100 )\n",
        "clf.fit(x,y)"
      ],
      "execution_count": 184,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/neural_network/_multilayer_perceptron.py:571: ConvergenceWarning: Stochastic Optimizer: Maximum iterations (100) reached and the optimization hasn't converged yet.\n",
            "  % self.max_iter, ConvergenceWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,\n",
              "              beta_2=0.999, early_stopping=False, epsilon=1e-08,\n",
              "              hidden_layer_sizes=(32,), learning_rate='constant',\n",
              "              learning_rate_init=0.001, max_fun=15000, max_iter=100,\n",
              "              momentum=0.9, n_iter_no_change=10, nesterovs_momentum=True,\n",
              "              power_t=0.5, random_state=None, shuffle=True, solver='adam',\n",
              "              tol=0.0001, validation_fraction=0.1, verbose=False,\n",
              "              warm_start=False)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 184
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 610
        },
        "id": "0aHQq-cv0OPu",
        "outputId": "3edca404-af76-4a89-a2c0-e21ffbc7ee89"
      },
      "source": [
        "predicted = clf.predict(x)\n",
        "fig = plt.figure(figsize = (8,8))\n",
        "fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)\n",
        "for i in range(36):\n",
        "    ax = fig.add_subplot(6, 6, i+1, xticks=[], yticks=[])\n",
        "    ax.imshow(digits.images[i],cmap=plt.cm.binary,interpolation='nearest')\n",
        "    ax.text(0, 7, str('{}-{}'.format(digits.target[i],predicted[i])), color=\"red\", fontsize = 20)\n"
      ],
      "execution_count": 185,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 36 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3oILwtmH0bZ3",
        "outputId": "859415b4-202e-4098-a77e-8c6f7144220b"
      },
      "source": [
        "accuracy_score(y, predicted)"
      ],
      "execution_count": 186,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9838619922092376"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 186
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wrbLIRVx0sV4",
        "outputId": "9547df73-d2c7-44ad-865b-5d902ce470af"
      },
      "source": [
        "clf.coefs_[0].shape"
      ],
      "execution_count": 190,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(64, 32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 190
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Iw-PIwdTzRBu"
      },
      "source": [
        "### Standardized "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YVQP_gxfzDhS",
        "outputId": "e8a91608-9d7d-4821-a362-07b44c2f41ad"
      },
      "source": [
        "a = np.array([50,60,70,80,90])\n",
        "a.mean(), a.std()"
      ],
      "execution_count": 175,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(70.0, 14.142135623730951)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 175
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZcTB1RozzXOr",
        "outputId": "87cdee15-e48c-4296-baea-daa1cd84870e"
      },
      "source": [
        "(a - a.mean()) / a.std()"
      ],
      "execution_count": 176,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([-1.41421356, -0.70710678,  0.        ,  0.70710678,  1.41421356])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 176
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "z90BGLjRzcR_",
        "outputId": "3f787302-30a8-47b9-ce25-76dc580b8707"
      },
      "source": [
        "a = np.array([5, 13,15])\n",
        "sigmoid_function(a)"
      ],
      "execution_count": 177,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.99330715, 0.99999774, 0.99999969])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 177
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8iP-iwC4zogd"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}