Week1.RNN_exam.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Tony_RNN.exam.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jinwoo95/Seq2Seq-Modelling-Study/blob/main/Week1.RNN_exam.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1eXgVtsdX6JT"
      },
      "source": [
        "참고자료 : https://www.kaggle.com/kanncaa1/recurrent-neural-network-with-pytorch\n",
        "\n",
        "# RNN 실습하기\n",
        "이론으로 학습한 rnn을 실전적 예제에 적용해봄으로써 rnn에 대한 심층적 이해와 실전적 활용능력을 배양하고자 한다.\n",
        "\n",
        "## Introduction\n",
        "- 많은 고전 모델들(eg.ARIMA)가 TS data를 다루는 데에 사용되어 왔다.\n",
        "- 하지만, 최근 몇 년간 LSTM이 TS 분야에서 좋은 성능을 내고 있다.\n",
        "- RNN은 Sequence data를 다루는 데에 있어서 좋은 성능을 보이며, RNN에서는 Sequence 초기에 있는 데이터의 정보를 Sequence 뒷 단계에서도 추출해 낼 수 있다는 장점이 있다.\n",
        "\n",
        "### Step :\n",
        "  1. Import Libraries\n",
        "  2. Prepare Dataset\n",
        "  3. Create RNN Model\n",
        "    - hidden layer dimension is 100\n",
        "    - number of hidden layer is 1\n",
        "  4. Instantiate Model\n",
        "  5. Instantiate Loss\n",
        "    - Cross entropy loss\n",
        "    - It also has softmax(logistic function) in it.\n",
        "  6. Instantiate Optimizer\n",
        "    - SGD Optimizer\n",
        "  7. Traning the Model\n",
        "  8. Prediction"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9myt0QqN8KAa",
        "outputId": "ef1d2915-609e-44a7-e43a-040bf5f8fd24"
      },
      "source": [
        "#공유 드라이브 연결\n",
        "from google.colab import drive\n",
        "import os\n",
        "drive.mount('/content/gdrive/', force_remount=True)\n",
        "os.chdir('gdrive/MyDrive/DoVie/Week1 RNN/data/Tony_RNN.exam')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/gdrive/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ecZZuUNUZnyI"
      },
      "source": [
        "\n",
        "## Import Libraries\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uamcLszwX0pV"
      },
      "source": [
        "import torch\n",
        "import numpy as np # linear algebra\n",
        "import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "\n",
        "import torch.nn as nn\n",
        "from torch.autograd import Variable\n",
        "from sklearn.model_selection import train_test_split\n",
        "from torch.utils.data import DataLoader, TensorDataset"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j8TYIY3ia6ZI"
      },
      "source": [
        "### Prepare Dataset\n",
        "- Data는 MNIST data이다."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "id": "dp3qg-gpYC6O",
        "outputId": "fc7c9c29-4bfe-461c-811e-7cf308bd8107"
      },
      "source": [
        "train = pd.read_csv(r\"train.csv\",dtype = np.float32)\n",
        "\n",
        "train.head()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>pixel0</th>\n",
              "      <th>pixel1</th>\n",
              "      <th>pixel2</th>\n",
              "      <th>pixel3</th>\n",
              "      <th>pixel4</th>\n",
              "      <th>pixel5</th>\n",
              "      <th>pixel6</th>\n",
              "      <th>pixel7</th>\n",
              "      <th>pixel8</th>\n",
              "      <th>pixel9</th>\n",
              "      <th>pixel10</th>\n",
              "      <th>pixel11</th>\n",
              "      <th>pixel12</th>\n",
              "      <th>pixel13</th>\n",
              "      <th>pixel14</th>\n",
              "      <th>pixel15</th>\n",
              "      <th>pixel16</th>\n",
              "      <th>pixel17</th>\n",
              "      <th>pixel18</th>\n",
              "      <th>pixel19</th>\n",
              "      <th>pixel20</th>\n",
              "      <th>pixel21</th>\n",
              "      <th>pixel22</th>\n",
              "      <th>pixel23</th>\n",
              "      <th>pixel24</th>\n",
              "      <th>pixel25</th>\n",
              "      <th>pixel26</th>\n",
              "      <th>pixel27</th>\n",
              "      <th>pixel28</th>\n",
              "      <th>pixel29</th>\n",
              "      <th>pixel30</th>\n",
              "      <th>pixel31</th>\n",
              "      <th>pixel32</th>\n",
              "      <th>pixel33</th>\n",
              "      <th>pixel34</th>\n",
              "      <th>pixel35</th>\n",
              "      <th>pixel36</th>\n",
              "      <th>pixel37</th>\n",
              "      <th>pixel38</th>\n",
              "      <th>...</th>\n",
              "      <th>pixel744</th>\n",
              "      <th>pixel745</th>\n",
              "      <th>pixel746</th>\n",
              "      <th>pixel747</th>\n",
              "      <th>pixel748</th>\n",
              "      <th>pixel749</th>\n",
              "      <th>pixel750</th>\n",
              "      <th>pixel751</th>\n",
              "      <th>pixel752</th>\n",
              "      <th>pixel753</th>\n",
              "      <th>pixel754</th>\n",
              "      <th>pixel755</th>\n",
              "      <th>pixel756</th>\n",
              "      <th>pixel757</th>\n",
              "      <th>pixel758</th>\n",
              "      <th>pixel759</th>\n",
              "      <th>pixel760</th>\n",
              "      <th>pixel761</th>\n",
              "      <th>pixel762</th>\n",
              "      <th>pixel763</th>\n",
              "      <th>pixel764</th>\n",
              "      <th>pixel765</th>\n",
              "      <th>pixel766</th>\n",
              "      <th>pixel767</th>\n",
              "      <th>pixel768</th>\n",
              "      <th>pixel769</th>\n",
              "      <th>pixel770</th>\n",
              "      <th>pixel771</th>\n",
              "      <th>pixel772</th>\n",
              "      <th>pixel773</th>\n",
              "      <th>pixel774</th>\n",
              "      <th>pixel775</th>\n",
              "      <th>pixel776</th>\n",
              "      <th>pixel777</th>\n",
              "      <th>pixel778</th>\n",
              "      <th>pixel779</th>\n",
              "      <th>pixel780</th>\n",
              "      <th>pixel781</th>\n",
              "      <th>pixel782</th>\n",
              "      <th>pixel783</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 785 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   label  pixel0  pixel1  pixel2  ...  pixel780  pixel781  pixel782  pixel783\n",
              "0    1.0     0.0     0.0     0.0  ...       0.0       0.0       0.0       0.0\n",
              "1    0.0     0.0     0.0     0.0  ...       0.0       0.0       0.0       0.0\n",
              "2    1.0     0.0     0.0     0.0  ...       0.0       0.0       0.0       0.0\n",
              "3    4.0     0.0     0.0     0.0  ...       0.0       0.0       0.0       0.0\n",
              "4    0.0     0.0     0.0     0.0  ...       0.0       0.0       0.0       0.0\n",
              "\n",
              "[5 rows x 785 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "qLd6qNLUbl27",
        "outputId": "8c6a4781-700e-4dd3-ae7e-f79548df09c0"
      },
      "source": [
        "# feature 와 label 분리\n",
        "targets_numpy = train.label.values\n",
        "features_numpy = train.loc[:,train.columns != \"label\"].values/255 # normalization\n",
        "\n",
        "# train set과 test set의 분리.\n",
        "features_train, features_test, targets_train, targets_test = train_test_split(features_numpy,\n",
        "                                                                             targets_numpy,\n",
        "                                                                             test_size = 0.2,\n",
        "                                                                             random_state = 42) \n",
        "\n",
        "# create feature and targets tensor for train set. As you remember we need variable to accumulate gradients. Therefore first we create tensor, then we will create variable\n",
        "featuresTrain = torch.from_numpy(features_train) # numpy 배열을 Torch Tensor로 변환하기.\n",
        "targetsTrain = torch.from_numpy(targets_train).type(torch.LongTensor) # data type is long\n",
        "\n",
        "# create feature and targets tensor for test set.\n",
        "featuresTest = torch.from_numpy(features_test)\n",
        "targetsTest = torch.from_numpy(targets_test).type(torch.LongTensor) # data type is long\n",
        "\n",
        "# batch_size, epoch and iteration\n",
        "batch_size = 100\n",
        "n_iters = 10000\n",
        "num_epochs = n_iters / (len(features_train) / batch_size)\n",
        "num_epochs = int(num_epochs)\n",
        "\n",
        "# Pytorch train and test sets\n",
        "train = TensorDataset(featuresTrain,targetsTrain) # Dataset wrapping tensors. 즉, 학습데이터 X와 레이블 Y를 묶어 놓는 컨테이너 역할\n",
        "test = TensorDataset(featuresTest,targetsTest)\n",
        "\n",
        "# data loader\n",
        "## 순서로 섞어서 batch_size 개수씩 데이터를 변환하는 DataLoader 작성\n",
        "train_loader = DataLoader(train, batch_size = batch_size, shuffle = False)\n",
        "test_loader = DataLoader(test, batch_size = batch_size, shuffle = False)\n",
        "\n",
        "# visualize one of the images in data set\n",
        "plt.imshow(features_numpy[10].reshape(28,28))\n",
        "plt.axis(\"off\")\n",
        "plt.title(str(targets_numpy[10]))\n",
        "plt.savefig('graph.png')\n",
        "plt.show()"
      ],
      "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": "EDrL3Ilnp4wi"
      },
      "source": [
        "### RNN 모델 class 생성"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Jt-FqE0DWHZ"
      },
      "source": [
        "##RNNMOdel은 nn.Module을 상속받는다. \n",
        "##super().__init__은 부모 클래스의 init을 상속받는 것이다. \n",
        "class RNNModel(nn.Module):\n",
        "  #구조를 쌓기 위한 변수 선언.\n",
        "    def __init__(self, input_dim, hidden_dim, layer_dim, output_dim):\n",
        "        super(RNNModel, self).__init__()\n",
        "        \n",
        "        # Number of hidden dimensions(Hidden의 차원 수)\n",
        "        self.hidden_dim = hidden_dim\n",
        "        \n",
        "        # Number of hidden layers(Hidden layer의 개수)\n",
        "        self.layer_dim = layer_dim\n",
        "        \n",
        "        # RNN\n",
        "        self.rnn = nn.RNN(input_dim, hidden_dim, layer_dim, batch_first=True, nonlinearity='relu')\n",
        "        \n",
        "        # Readout layer (fully-connected-layer)\n",
        "        self.fc = nn.Linear(hidden_dim, output_dim)\n",
        "    \n",
        "    #feedforward를 위한 각 구조 연결.\n",
        "    def forward(self, x):\n",
        "        \n",
        "        # Initialize hidden state with zeros, layer_dim은 Hidden layer의 개수.즉, cell의 개수.\n",
        "        h0 = Variable(torch.zeros(self.layer_dim, x.size(0), self.hidden_dim)) #x.size는 전체 원소의 개수 반환\n",
        "        \n",
        "        # One time step\n",
        "        out, hn = self.rnn(x, h0)\n",
        "        out = self.fc(out[:, -1, :])\n",
        "        return out\n",
        "\n",
        "# batch_size, epoch and iteration\n",
        "batch_size = 100\n",
        "n_iters = 8000\n",
        "num_epochs = n_iters / (len(features_train) / batch_size)\n",
        "num_epochs = int(num_epochs)\n",
        "\n",
        "# Pytorch train and test sets\n",
        "train = TensorDataset(featuresTrain,targetsTrain)\n",
        "test = TensorDataset(featuresTest,targetsTest)\n",
        "\n",
        "# data loader\n",
        "train_loader = DataLoader(train, batch_size = batch_size, shuffle = False)\n",
        "test_loader = DataLoader(test, batch_size = batch_size, shuffle = False)\n",
        "\n",
        "# Create RNN\n",
        "input_dim = 28    # input dimension \n",
        "hidden_dim = 100  # hidden layer dimension #Hidden layer에 포함된 유닛의 개수 \n",
        "layer_dim = 1     # number of hidden layers #Hidden layer의 개수 \n",
        "output_dim = 10   # output dimension, 출력 값이 10개로 나온다는 뜻. \n",
        "\n",
        "#모델 구축\n",
        "model = RNNModel(input_dim, hidden_dim, layer_dim, output_dim)\n",
        "\n",
        "# Cross Entropy Loss \n",
        "error = nn.CrossEntropyLoss()\n",
        "\n",
        "# SGD Optimizer\n",
        "learning_rate = 0.05\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 172
        },
        "id": "MYBDCSbb66oW",
        "outputId": "8ba7e6b6-127a-41f7-e811-b11cb1e8eaff"
      },
      "source": [
        "len(train.tensors[0][1]) / 28"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-a3666f65ce61>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtensors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0;36m28\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m: name 'train' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BFafGVjZsaiN"
      },
      "source": [
        "### 학습하기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bV9pTLkjDXBu",
        "outputId": "1fb98555-adf9-4005-dbe6-7ea614d2e94e"
      },
      "source": [
        "seq_dim = 28  #sequence의 길이 784 = 28(input_dimension) * 28 (sequence_dimension)\n",
        "#필요한 list 정의\n",
        "loss_list = []\n",
        "iteration_list = []\n",
        "accuracy_list = []\n",
        "count = 0\n",
        "\n",
        "for epoch in range(num_epochs): #epoch는 전체 데이터를 한 번 보는 것.\n",
        "    for i, (images, labels) in enumerate(train_loader): #i는 데이터의 순서, image는 feature들\n",
        "\n",
        "        #train_loader에는 batch_size 만큼의 데이터가 한 단위로 들어가 있다. \n",
        "        train  = Variable(images.view(-1,  seq_dim, input_dim)) #tensor의 크기(size)나 모양을 변경하고 싶을 때 torch.view사용. numpy의 reshape 같은 역할. 앞의 -1은 batch_size 단위로 학습시키기 위해 붙여주는 듯.\n",
        "        labels = Variable(labels)\n",
        "        \n",
        "        # 역전파 단계 전에, Optimizer 객체를 사용하여 (모델의 학습 가능한 가중치인)\n",
        "        # 갱신할 변수들에 대한 모든 변화도를 0으로 만듭니다. 이렇게 하는 이유는\n",
        "        # 기본적으로 .backward()를 호출할 때마다 변화도가 버퍼(buffer)에 (덮어쓰지 않고)\n",
        "        # 누적되기 때문입니다. 더 자세한 내용은 torch.autograd.backward에 대한 문서를\n",
        "        # 참조하세요.\n",
        "        # Clear gradients\n",
        "        optimizer.zero_grad()\n",
        "        \n",
        "        # Forward propagation\n",
        "        outputs = model(train)\n",
        "        \n",
        "        # Calculate softmax and ross entropy loss\n",
        "        loss = error(outputs, labels)\n",
        "        \n",
        "        # Calculating gradients, 역전파 단계\n",
        "        loss.backward()\n",
        "        \n",
        "        # Update parameters , 파라미터 update\n",
        "        optimizer.step()\n",
        "        \n",
        "        count += 1\n",
        "        #250번마다 출력.\n",
        "        if count % 250 == 0:\n",
        "            # Calculate Accuracy         \n",
        "            correct = 0\n",
        "            total = 0\n",
        "            # Iterate through test dataset\n",
        "            for images, labels in test_loader:\n",
        "                images = Variable(images.view(-1, seq_dim, input_dim))\n",
        "                \n",
        "                # Forward propagation\n",
        "                outputs = model(images)\n",
        "                \n",
        "                # Get predictions from the maximum value , max인 값 하나 출력.\n",
        "                predicted = torch.max(outputs.data, 1)[1]\n",
        "                \n",
        "                # Total number of labels\n",
        "                total += labels.size(0)\n",
        "                \n",
        "                correct += (predicted == labels).sum()\n",
        "            \n",
        "            accuracy = 100 * correct / float(total)\n",
        "            \n",
        "            # store loss and iteration\n",
        "            loss_list.append(loss.data)\n",
        "            iteration_list.append(count)\n",
        "            accuracy_list.append(accuracy)\n",
        "            if count % 500 == 0:\n",
        "                # Print Loss\n",
        "                print('Iteration: {}  Loss: {}  Accuracy: {} %'.format(count, loss.data, accuracy))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Iteration: 500  Loss: 1.8662906885147095  Accuracy: 25.904762268066406 %\n",
            "Iteration: 1000  Loss: 1.2908999919891357  Accuracy: 57.66666793823242 %\n",
            "Iteration: 1500  Loss: 0.5768565535545349  Accuracy: 82.27381134033203 %\n",
            "Iteration: 2000  Loss: 0.5902754068374634  Accuracy: 84.75 %\n",
            "Iteration: 2500  Loss: 0.236233189702034  Accuracy: 89.85713958740234 %\n",
            "Iteration: 3000  Loss: 0.10372806340456009  Accuracy: 92.71428680419922 %\n",
            "Iteration: 3500  Loss: 0.3212709128856659  Accuracy: 90.32142639160156 %\n",
            "Iteration: 4000  Loss: 0.09733286499977112  Accuracy: 94.39286041259766 %\n",
            "Iteration: 4500  Loss: 0.5112283825874329  Accuracy: 84.89286041259766 %\n",
            "Iteration: 5000  Loss: 0.09765458852052689  Accuracy: 94.8452377319336 %\n",
            "Iteration: 5500  Loss: 0.22413517534732819  Accuracy: 95.03571319580078 %\n",
            "Iteration: 6000  Loss: 0.23084676265716553  Accuracy: 95.69047546386719 %\n",
            "Iteration: 6500  Loss: 0.1013358011841774  Accuracy: 95.80952453613281 %\n",
            "Iteration: 7000  Loss: 0.15087687969207764  Accuracy: 96.10713958740234 %\n",
            "Iteration: 7500  Loss: 0.09949839860200882  Accuracy: 96.23809814453125 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        },
        "id": "7efT5Jsupv7y",
        "outputId": "cc810d0a-08ce-44fc-a199-4c2c1f7564a9"
      },
      "source": [
        "# visualization loss \n",
        "plt.plot(iteration_list,loss_list)\n",
        "plt.xlabel(\"Number of iteration\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.title(\"RNN: Loss vs Number of iteration\")\n",
        "plt.show()\n",
        "\n",
        "# visualization accuracy \n",
        "plt.plot(iteration_list,accuracy_list,color = \"red\")\n",
        "plt.xlabel(\"Number of iteration\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.title(\"RNN: Accuracy vs Number of iteration\")\n",
        "plt.savefig('graph.png')\n",
        "plt.show()"
      ],
      "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"
          }
        },
        {
          "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": "0EGxCKHNu5zK",
        "outputId": "0ff8c73a-7eaa-429c-cc91-2b088be10281"
      },
      "source": [
        "labels.size()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "torch.Size([100])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "D4ohsRqevqnP",
        "outputId": "c0131110-7f9f-467c-eba4-db7d0fe639fd"
      },
      "source": [
        "iteration_list"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pcJFQOCqylr9"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}