material/deep_learning/classification_hidden_mnist_in_lettern.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "classification_hidden_mnist_in_lettern.ipynb",
      "provenance": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "upvzpuGa9hD6"
      },
      "source": [
        "# 개요\n",
        "\n",
        "Dacon의 대회 https://dacon.io/competitions/official/235626/data 에 대한 풀이.\n",
        "\n",
        "중앙대학교 정영석, 김민환, 김학개, 진소연의 결과를 가져다 정리함."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "htP1c1HGRY75"
      },
      "source": [
        "# 데이터"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X4B2LS9s7NYb"
      },
      "source": [
        "## 다운로드"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JGJYeMDHRVvI",
        "outputId": "762b3853-ed88-4590-a991-b397cfe7ce4f"
      },
      "source": [
        "!wget https://github.com/dhrim/2022_WISET/raw/main/material/deep_learning/data/hiden_mnist_in_letter.csv"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2021-09-16 01:58:06--  https://github.com/dhrim/2022_WISET/raw/main/material/deep_learning/data/hiden_mnist_in_letter.csv\n",
            "Resolving github.com (github.com)... 140.82.114.3\n",
            "Connecting to github.com (github.com)|140.82.114.3|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://raw.githubusercontent.com/dhrim/2022_WISET/main/material/deep_learning/data/hiden_mnist_in_letter.csv [following]\n",
            "--2021-09-16 01:58:08--  https://raw.githubusercontent.com/dhrim/2022_WISET/main/material/deep_learning/data/hiden_mnist_in_letter.csv\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4004157 (3.8M) [text/plain]\n",
            "Saving to: ‘hiden_mnist_in_letter.csv’\n",
            "\n",
            "hiden_mnist_in_lett 100%[===================>]   3.82M  --.-KB/s    in 0.04s   \n",
            "\n",
            "2021-09-16 01:58:08 (93.1 MB/s) - ‘hiden_mnist_in_letter.csv’ saved [4004157/4004157]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ys3yJvn47QMC"
      },
      "source": [
        "## 데이터 로딩"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Mewb_UpPR7iu"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "id": "FdYBh4ibSND2",
        "outputId": "e7d031ee-327c-46af-c2eb-acf2ef2118f6"
      },
      "source": [
        "df = pd.read_csv('hiden_mnist_in_letter.csv')\n",
        "df.head()"
      ],
      "execution_count": 2,
      "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>id</th>\n",
              "      <th>digit</th>\n",
              "      <th>letter</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>10</th>\n",
              "      <th>11</th>\n",
              "      <th>12</th>\n",
              "      <th>13</th>\n",
              "      <th>14</th>\n",
              "      <th>15</th>\n",
              "      <th>16</th>\n",
              "      <th>17</th>\n",
              "      <th>18</th>\n",
              "      <th>19</th>\n",
              "      <th>20</th>\n",
              "      <th>21</th>\n",
              "      <th>22</th>\n",
              "      <th>23</th>\n",
              "      <th>24</th>\n",
              "      <th>25</th>\n",
              "      <th>26</th>\n",
              "      <th>27</th>\n",
              "      <th>28</th>\n",
              "      <th>29</th>\n",
              "      <th>30</th>\n",
              "      <th>31</th>\n",
              "      <th>32</th>\n",
              "      <th>33</th>\n",
              "      <th>34</th>\n",
              "      <th>35</th>\n",
              "      <th>36</th>\n",
              "      <th>...</th>\n",
              "      <th>744</th>\n",
              "      <th>745</th>\n",
              "      <th>746</th>\n",
              "      <th>747</th>\n",
              "      <th>748</th>\n",
              "      <th>749</th>\n",
              "      <th>750</th>\n",
              "      <th>751</th>\n",
              "      <th>752</th>\n",
              "      <th>753</th>\n",
              "      <th>754</th>\n",
              "      <th>755</th>\n",
              "      <th>756</th>\n",
              "      <th>757</th>\n",
              "      <th>758</th>\n",
              "      <th>759</th>\n",
              "      <th>760</th>\n",
              "      <th>761</th>\n",
              "      <th>762</th>\n",
              "      <th>763</th>\n",
              "      <th>764</th>\n",
              "      <th>765</th>\n",
              "      <th>766</th>\n",
              "      <th>767</th>\n",
              "      <th>768</th>\n",
              "      <th>769</th>\n",
              "      <th>770</th>\n",
              "      <th>771</th>\n",
              "      <th>772</th>\n",
              "      <th>773</th>\n",
              "      <th>774</th>\n",
              "      <th>775</th>\n",
              "      <th>776</th>\n",
              "      <th>777</th>\n",
              "      <th>778</th>\n",
              "      <th>779</th>\n",
              "      <th>780</th>\n",
              "      <th>781</th>\n",
              "      <th>782</th>\n",
              "      <th>783</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>5</td>\n",
              "      <td>L</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>...</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>B</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>...</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>L</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>9</td>\n",
              "      <td>D</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>6</td>\n",
              "      <td>A</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>...</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 787 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   id  digit letter  0  1  2  3  4  ...  776  777  778  779  780  781  782  783\n",
              "0   1      5      L  1  1  1  4  3  ...    0    1    2    4    4    4    3    4\n",
              "1   2      0      B  0  4  0  0  4  ...    0    1    4    1    4    2    1    2\n",
              "2   3      4      L  1  1  2  2  1  ...    3    0    2    0    3    0    2    2\n",
              "3   4      9      D  1  2  0  2  0  ...    2    0    1    4    0    0    1    1\n",
              "4   5      6      A  3  0  2  4  0  ...    3    2    1    3    4    3    1    2\n",
              "\n",
              "[5 rows x 787 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "id": "3EHnFG2iSUro",
        "outputId": "4bddf6fa-8da5-478a-8d9f-d17f9585eb84"
      },
      "source": [
        "df['letter'].value_counts().plot(kind='bar')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fe4a6002190>"
            ]
          },
          "metadata": {},
          "execution_count": 3
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 280
        },
        "id": "hqJZJvk-T_Fi",
        "outputId": "fba23011-e86f-49a0-8c3f-a4ffee2a8252"
      },
      "source": [
        "df['digit'].value_counts().plot(kind='bar')"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fe4a5a68e90>"
            ]
          },
          "metadata": {},
          "execution_count": 4
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        },
        "id": "_oReYWWHXQ8l",
        "outputId": "02b5eb24-cb2d-4f6c-dfd1-b7479ac50cc4"
      },
      "source": [
        "df.head()"
      ],
      "execution_count": 5,
      "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>id</th>\n",
              "      <th>digit</th>\n",
              "      <th>letter</th>\n",
              "      <th>0</th>\n",
              "      <th>1</th>\n",
              "      <th>2</th>\n",
              "      <th>3</th>\n",
              "      <th>4</th>\n",
              "      <th>5</th>\n",
              "      <th>6</th>\n",
              "      <th>7</th>\n",
              "      <th>8</th>\n",
              "      <th>9</th>\n",
              "      <th>10</th>\n",
              "      <th>11</th>\n",
              "      <th>12</th>\n",
              "      <th>13</th>\n",
              "      <th>14</th>\n",
              "      <th>15</th>\n",
              "      <th>16</th>\n",
              "      <th>17</th>\n",
              "      <th>18</th>\n",
              "      <th>19</th>\n",
              "      <th>20</th>\n",
              "      <th>21</th>\n",
              "      <th>22</th>\n",
              "      <th>23</th>\n",
              "      <th>24</th>\n",
              "      <th>25</th>\n",
              "      <th>26</th>\n",
              "      <th>27</th>\n",
              "      <th>28</th>\n",
              "      <th>29</th>\n",
              "      <th>30</th>\n",
              "      <th>31</th>\n",
              "      <th>32</th>\n",
              "      <th>33</th>\n",
              "      <th>34</th>\n",
              "      <th>35</th>\n",
              "      <th>36</th>\n",
              "      <th>...</th>\n",
              "      <th>744</th>\n",
              "      <th>745</th>\n",
              "      <th>746</th>\n",
              "      <th>747</th>\n",
              "      <th>748</th>\n",
              "      <th>749</th>\n",
              "      <th>750</th>\n",
              "      <th>751</th>\n",
              "      <th>752</th>\n",
              "      <th>753</th>\n",
              "      <th>754</th>\n",
              "      <th>755</th>\n",
              "      <th>756</th>\n",
              "      <th>757</th>\n",
              "      <th>758</th>\n",
              "      <th>759</th>\n",
              "      <th>760</th>\n",
              "      <th>761</th>\n",
              "      <th>762</th>\n",
              "      <th>763</th>\n",
              "      <th>764</th>\n",
              "      <th>765</th>\n",
              "      <th>766</th>\n",
              "      <th>767</th>\n",
              "      <th>768</th>\n",
              "      <th>769</th>\n",
              "      <th>770</th>\n",
              "      <th>771</th>\n",
              "      <th>772</th>\n",
              "      <th>773</th>\n",
              "      <th>774</th>\n",
              "      <th>775</th>\n",
              "      <th>776</th>\n",
              "      <th>777</th>\n",
              "      <th>778</th>\n",
              "      <th>779</th>\n",
              "      <th>780</th>\n",
              "      <th>781</th>\n",
              "      <th>782</th>\n",
              "      <th>783</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>1</td>\n",
              "      <td>5</td>\n",
              "      <td>L</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>...</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>B</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>...</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>L</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>4</td>\n",
              "      <td>9</td>\n",
              "      <td>D</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>...</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>5</td>\n",
              "      <td>6</td>\n",
              "      <td>A</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>0</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>...</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>0</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>4</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>2</td>\n",
              "      <td>1</td>\n",
              "      <td>3</td>\n",
              "      <td>4</td>\n",
              "      <td>3</td>\n",
              "      <td>1</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>5 rows × 787 columns</p>\n",
              "</div>"
            ],
            "text/plain": [
              "   id  digit letter  0  1  2  3  4  ...  776  777  778  779  780  781  782  783\n",
              "0   1      5      L  1  1  1  4  3  ...    0    1    2    4    4    4    3    4\n",
              "1   2      0      B  0  4  0  0  4  ...    0    1    4    1    4    2    1    2\n",
              "2   3      4      L  1  1  2  2  1  ...    3    0    2    0    3    0    2    2\n",
              "3   4      9      D  1  2  0  2  0  ...    2    0    1    4    0    0    1    1\n",
              "4   5      6      A  3  0  2  4  0  ...    3    2    1    3    4    3    1    2\n",
              "\n",
              "[5 rows x 787 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RSmt-VaSTAnO"
      },
      "source": [
        "raw_x = df.iloc[:,3:].to_numpy()\n",
        "digit = df.digit.to_numpy() \n",
        "letter = df.letter.to_numpy()"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U7jgibm1WcT8",
        "outputId": "c6ee4c0c-3218-4904-b53b-5ee9a5e9287d"
      },
      "source": [
        "print(digit)\n",
        "print(letter)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[5 0 4 ... 9 0 5]\n",
            "['L' 'B' 'L' ... 'A' 'Z' 'Z']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OiwR8MWc7T8n"
      },
      "source": [
        "## normalization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vCIujQEuVMCD"
      },
      "source": [
        "x = raw_x / 255.0\n",
        "y = digit"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VZ7cu4LeTw3v",
        "outputId": "a7fed57e-f7bd-4d93-c997-b9561211c356"
      },
      "source": [
        "print(x.shape)\n",
        "print(y.shape)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(2048, 784)\n",
            "(2048,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GjVM9ijn7WTA"
      },
      "source": [
        "## 3d이미지로 reshape"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j0RKPSrwc74_",
        "outputId": "b89ce294-35d7-40c6-f275-b2b802c7d8f2"
      },
      "source": [
        "x = x.reshape((-1,28,28,1))\n",
        "print(x.shape)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(2048, 28, 28, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 608
        },
        "id": "8GFJa8aAWEvR",
        "outputId": "3c7caa66-634a-4911-9fa9-d1f13ae36ea5"
      },
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5, i + 1)\n",
        "    plt.title('%s, %s' % (digit[i], letter[i]))\n",
        "    plt.imshow(x[i].squeeze())"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a1qjZS7V7dSu"
      },
      "source": [
        "## 데이터 분리"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "29uOo5lDjI97"
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "train_x, test_x, train_y, test_y = train_test_split(x, y, test_size=0.2, shuffle=True)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "13KD7WBc7f2U"
      },
      "source": [
        "# 학습"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MyaxzoJLYVHZ"
      },
      "source": [
        "import tensorflow as tf\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator,load_img,img_to_array \n",
        "\n",
        "import sklearn as skl\n",
        "from sklearn.preprocessing import OneHotEncoder\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.model_selection import StratifiedKFold\n",
        "\n",
        "from tensorflow.keras import Sequential, Model\n",
        "from tensorflow.keras.layers import Dense, Dropout, Activation, Input, BatchNormalization, GlobalAveragePooling2D, Conv2D, MaxPooling2D, Flatten\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.callbacks import EarlyStopping\n",
        "from tensorflow.keras.experimental import CosineDecay\n",
        "from tensorflow.keras.applications import EfficientNetB1\n",
        "from tensorflow.keras.layers.experimental.preprocessing import CenterCrop, RandomRotation, RandomZoom, RandomWidth, RandomHeight\n",
        "from tensorflow.keras.optimizers import Adam"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UbwG04pr7hde"
      },
      "source": [
        "## 모델 정의"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dK4-QBlA5y4D",
        "outputId": "ac17ad02-846d-4e7f-9fd5-d55792816b26"
      },
      "source": [
        "model = keras.Sequential()\n",
        "model.add(Input((28,28,1)))\n",
        "model.add(Conv2D(32, (3, 3),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(32, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(32, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(32, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D((2, 2)))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Conv2D(64, (3, 3),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(64, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(64, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D((2, 2)))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Conv2D(128, (3, 3),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Conv2D(128, (5, 5),activation='relu',padding='same'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(MaxPooling2D((2, 2)))\n",
        "model.add(Dropout(0.5))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Dense(256, activation='relu'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(BatchNormalization())\n",
        "model.add(Dropout(0.5))\n",
        "model.add(Dense(10, activation='softmax'))\n",
        "\n",
        "model.summary()\n",
        "\n",
        "model.compile(optimizer=\"Adam\", loss=\"sparse_categorical_crossentropy\", metrics=[\"accuracy\"])"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 28, 28, 32)        320       \n",
            "_________________________________________________________________\n",
            "batch_normalization (BatchNo (None, 28, 28, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 28, 28, 32)        25632     \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 28, 28, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 28, 28, 32)        25632     \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 28, 28, 32)        128       \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 28, 28, 32)        25632     \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 28, 28, 32)        128       \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 14, 14, 32)        0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 14, 14, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_4 (Conv2D)            (None, 14, 14, 64)        18496     \n",
            "_________________________________________________________________\n",
            "batch_normalization_4 (Batch (None, 14, 14, 64)        256       \n",
            "_________________________________________________________________\n",
            "conv2d_5 (Conv2D)            (None, 14, 14, 64)        102464    \n",
            "_________________________________________________________________\n",
            "batch_normalization_5 (Batch (None, 14, 14, 64)        256       \n",
            "_________________________________________________________________\n",
            "conv2d_6 (Conv2D)            (None, 14, 14, 64)        102464    \n",
            "_________________________________________________________________\n",
            "batch_normalization_6 (Batch (None, 14, 14, 64)        256       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 7, 7, 64)          0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 7, 7, 64)          0         \n",
            "_________________________________________________________________\n",
            "conv2d_7 (Conv2D)            (None, 7, 7, 128)         73856     \n",
            "_________________________________________________________________\n",
            "batch_normalization_7 (Batch (None, 7, 7, 128)         512       \n",
            "_________________________________________________________________\n",
            "conv2d_8 (Conv2D)            (None, 7, 7, 128)         409728    \n",
            "_________________________________________________________________\n",
            "batch_normalization_8 (Batch (None, 7, 7, 128)         512       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 3, 3, 128)         0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 3, 3, 128)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 1152)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 256)               295168    \n",
            "_________________________________________________________________\n",
            "batch_normalization_9 (Batch (None, 256)               1024      \n",
            "_________________________________________________________________\n",
            "dropout_3 (Dropout)          (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 64)                16448     \n",
            "_________________________________________________________________\n",
            "batch_normalization_10 (Batc (None, 64)                256       \n",
            "_________________________________________________________________\n",
            "dropout_4 (Dropout)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                650       \n",
            "=================================================================\n",
            "Total params: 1,100,074\n",
            "Trainable params: 1,098,282\n",
            "Non-trainable params: 1,792\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xd4ujV4O7ji0"
      },
      "source": [
        "## 학습 실행"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dZq03R6FkU71",
        "outputId": "ca3840aa-d88a-48b3-bf79-87892efb20e9"
      },
      "source": [
        "history = model.fit(train_x,train_y, verbose=1, batch_size=128, epochs=200, validation_split=0.1)"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "12/12 [==============================] - 5s 90ms/step - loss: 3.4691 - accuracy: 0.1004 - val_loss: 2.3192 - val_accuracy: 0.0976\n",
            "Epoch 2/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 3.2058 - accuracy: 0.1187 - val_loss: 2.3581 - val_accuracy: 0.0671\n",
            "Epoch 3/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.9340 - accuracy: 0.1533 - val_loss: 2.4351 - val_accuracy: 0.0610\n",
            "Epoch 4/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.8120 - accuracy: 0.1547 - val_loss: 2.4794 - val_accuracy: 0.0610\n",
            "Epoch 5/200\n",
            "12/12 [==============================] - 0s 29ms/step - loss: 2.6120 - accuracy: 0.1859 - val_loss: 2.5195 - val_accuracy: 0.0610\n",
            "Epoch 6/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.5495 - accuracy: 0.2035 - val_loss: 2.5989 - val_accuracy: 0.0610\n",
            "Epoch 7/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.3760 - accuracy: 0.2266 - val_loss: 2.7129 - val_accuracy: 0.0610\n",
            "Epoch 8/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.2897 - accuracy: 0.2483 - val_loss: 2.7765 - val_accuracy: 0.0610\n",
            "Epoch 9/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.2018 - accuracy: 0.2483 - val_loss: 2.9514 - val_accuracy: 0.0610\n",
            "Epoch 10/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 2.1673 - accuracy: 0.2856 - val_loss: 3.1839 - val_accuracy: 0.0610\n",
            "Epoch 11/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 2.0231 - accuracy: 0.3060 - val_loss: 3.5460 - val_accuracy: 0.0610\n",
            "Epoch 12/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.9277 - accuracy: 0.3168 - val_loss: 4.5392 - val_accuracy: 0.0610\n",
            "Epoch 13/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.8617 - accuracy: 0.3657 - val_loss: 5.6301 - val_accuracy: 0.0610\n",
            "Epoch 14/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.7583 - accuracy: 0.3833 - val_loss: 6.3210 - val_accuracy: 0.0610\n",
            "Epoch 15/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 1.6988 - accuracy: 0.4152 - val_loss: 7.1557 - val_accuracy: 0.0610\n",
            "Epoch 16/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 1.5391 - accuracy: 0.4851 - val_loss: 7.5892 - val_accuracy: 0.0610\n",
            "Epoch 17/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.4813 - accuracy: 0.4871 - val_loss: 7.4445 - val_accuracy: 0.0610\n",
            "Epoch 18/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.3841 - accuracy: 0.5285 - val_loss: 7.7804 - val_accuracy: 0.0610\n",
            "Epoch 19/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 1.3060 - accuracy: 0.5488 - val_loss: 8.6432 - val_accuracy: 0.0610\n",
            "Epoch 20/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.2238 - accuracy: 0.5712 - val_loss: 9.3731 - val_accuracy: 0.0610\n",
            "Epoch 21/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 1.2325 - accuracy: 0.5665 - val_loss: 9.5442 - val_accuracy: 0.0610\n",
            "Epoch 22/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.1299 - accuracy: 0.6242 - val_loss: 9.2959 - val_accuracy: 0.0610\n",
            "Epoch 23/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.0913 - accuracy: 0.6445 - val_loss: 8.9363 - val_accuracy: 0.0610\n",
            "Epoch 24/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 1.0310 - accuracy: 0.6411 - val_loss: 9.9130 - val_accuracy: 0.0610\n",
            "Epoch 25/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 1.0016 - accuracy: 0.6737 - val_loss: 9.7828 - val_accuracy: 0.0610\n",
            "Epoch 26/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.9374 - accuracy: 0.6866 - val_loss: 9.9287 - val_accuracy: 0.0610\n",
            "Epoch 27/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.8762 - accuracy: 0.7157 - val_loss: 9.6408 - val_accuracy: 0.0610\n",
            "Epoch 28/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.8778 - accuracy: 0.7157 - val_loss: 9.7734 - val_accuracy: 0.0610\n",
            "Epoch 29/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.8232 - accuracy: 0.7334 - val_loss: 10.2614 - val_accuracy: 0.0610\n",
            "Epoch 30/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.8216 - accuracy: 0.7402 - val_loss: 9.6516 - val_accuracy: 0.0610\n",
            "Epoch 31/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.8067 - accuracy: 0.7544 - val_loss: 9.6250 - val_accuracy: 0.0610\n",
            "Epoch 32/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.7430 - accuracy: 0.7510 - val_loss: 10.3193 - val_accuracy: 0.0610\n",
            "Epoch 33/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.7098 - accuracy: 0.7666 - val_loss: 10.1043 - val_accuracy: 0.0610\n",
            "Epoch 34/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.6630 - accuracy: 0.7782 - val_loss: 9.4516 - val_accuracy: 0.0610\n",
            "Epoch 35/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.6623 - accuracy: 0.7938 - val_loss: 9.6104 - val_accuracy: 0.0610\n",
            "Epoch 36/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.6665 - accuracy: 0.7836 - val_loss: 10.4375 - val_accuracy: 0.0610\n",
            "Epoch 37/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.6105 - accuracy: 0.8060 - val_loss: 10.0494 - val_accuracy: 0.0610\n",
            "Epoch 38/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.6308 - accuracy: 0.8060 - val_loss: 6.7094 - val_accuracy: 0.0793\n",
            "Epoch 39/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.5709 - accuracy: 0.8202 - val_loss: 8.9226 - val_accuracy: 0.0610\n",
            "Epoch 40/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.5558 - accuracy: 0.8263 - val_loss: 9.1610 - val_accuracy: 0.0610\n",
            "Epoch 41/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.5365 - accuracy: 0.8392 - val_loss: 7.9350 - val_accuracy: 0.0610\n",
            "Epoch 42/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.5556 - accuracy: 0.8121 - val_loss: 7.4778 - val_accuracy: 0.0610\n",
            "Epoch 43/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.4786 - accuracy: 0.8501 - val_loss: 5.4668 - val_accuracy: 0.1159\n",
            "Epoch 44/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.4590 - accuracy: 0.8616 - val_loss: 6.0714 - val_accuracy: 0.0854\n",
            "Epoch 45/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.4757 - accuracy: 0.8494 - val_loss: 5.5681 - val_accuracy: 0.1098\n",
            "Epoch 46/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.4324 - accuracy: 0.8684 - val_loss: 4.9406 - val_accuracy: 0.2012\n",
            "Epoch 47/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.4016 - accuracy: 0.8745 - val_loss: 5.2425 - val_accuracy: 0.1646\n",
            "Epoch 48/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.4131 - accuracy: 0.8758 - val_loss: 4.5909 - val_accuracy: 0.2012\n",
            "Epoch 49/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.4144 - accuracy: 0.8718 - val_loss: 3.3095 - val_accuracy: 0.3537\n",
            "Epoch 50/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.3722 - accuracy: 0.8894 - val_loss: 3.7063 - val_accuracy: 0.2988\n",
            "Epoch 51/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.3666 - accuracy: 0.8915 - val_loss: 3.3091 - val_accuracy: 0.3659\n",
            "Epoch 52/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.3648 - accuracy: 0.8833 - val_loss: 2.0962 - val_accuracy: 0.5305\n",
            "Epoch 53/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.3409 - accuracy: 0.9016 - val_loss: 2.8906 - val_accuracy: 0.4390\n",
            "Epoch 54/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.3384 - accuracy: 0.8948 - val_loss: 2.5483 - val_accuracy: 0.5061\n",
            "Epoch 55/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.3086 - accuracy: 0.9071 - val_loss: 3.7755 - val_accuracy: 0.3780\n",
            "Epoch 56/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.3005 - accuracy: 0.9043 - val_loss: 1.9334 - val_accuracy: 0.6159\n",
            "Epoch 57/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.3237 - accuracy: 0.8976 - val_loss: 1.8872 - val_accuracy: 0.6402\n",
            "Epoch 58/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2883 - accuracy: 0.9111 - val_loss: 1.7377 - val_accuracy: 0.5793\n",
            "Epoch 59/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2908 - accuracy: 0.9084 - val_loss: 1.1596 - val_accuracy: 0.7012\n",
            "Epoch 60/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2639 - accuracy: 0.9206 - val_loss: 1.3166 - val_accuracy: 0.6707\n",
            "Epoch 61/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2711 - accuracy: 0.9206 - val_loss: 1.2753 - val_accuracy: 0.6890\n",
            "Epoch 62/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2186 - accuracy: 0.9349 - val_loss: 1.3957 - val_accuracy: 0.6768\n",
            "Epoch 63/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.2333 - accuracy: 0.9328 - val_loss: 1.0896 - val_accuracy: 0.6829\n",
            "Epoch 64/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1887 - accuracy: 0.9498 - val_loss: 0.8055 - val_accuracy: 0.7683\n",
            "Epoch 65/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.1902 - accuracy: 0.9423 - val_loss: 1.1592 - val_accuracy: 0.7317\n",
            "Epoch 66/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.2027 - accuracy: 0.9362 - val_loss: 0.9562 - val_accuracy: 0.7317\n",
            "Epoch 67/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.2035 - accuracy: 0.9423 - val_loss: 0.7920 - val_accuracy: 0.7927\n",
            "Epoch 68/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1730 - accuracy: 0.9518 - val_loss: 0.8897 - val_accuracy: 0.7744\n",
            "Epoch 69/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1701 - accuracy: 0.9552 - val_loss: 0.7041 - val_accuracy: 0.7988\n",
            "Epoch 70/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.1709 - accuracy: 0.9484 - val_loss: 0.4949 - val_accuracy: 0.8476\n",
            "Epoch 71/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.1710 - accuracy: 0.9512 - val_loss: 0.4792 - val_accuracy: 0.8720\n",
            "Epoch 72/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1526 - accuracy: 0.9559 - val_loss: 0.6402 - val_accuracy: 0.8354\n",
            "Epoch 73/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1622 - accuracy: 0.9545 - val_loss: 0.5952 - val_accuracy: 0.8293\n",
            "Epoch 74/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1598 - accuracy: 0.9613 - val_loss: 0.5814 - val_accuracy: 0.8415\n",
            "Epoch 75/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1530 - accuracy: 0.9573 - val_loss: 0.6429 - val_accuracy: 0.8293\n",
            "Epoch 76/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1651 - accuracy: 0.9478 - val_loss: 0.6774 - val_accuracy: 0.8293\n",
            "Epoch 77/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1417 - accuracy: 0.9539 - val_loss: 0.5984 - val_accuracy: 0.8354\n",
            "Epoch 78/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1490 - accuracy: 0.9532 - val_loss: 0.6822 - val_accuracy: 0.8293\n",
            "Epoch 79/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1251 - accuracy: 0.9647 - val_loss: 0.4530 - val_accuracy: 0.8841\n",
            "Epoch 80/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1256 - accuracy: 0.9640 - val_loss: 0.5282 - val_accuracy: 0.8720\n",
            "Epoch 81/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.1199 - accuracy: 0.9654 - val_loss: 0.5751 - val_accuracy: 0.8537\n",
            "Epoch 82/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0950 - accuracy: 0.9749 - val_loss: 0.4857 - val_accuracy: 0.8598\n",
            "Epoch 83/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1152 - accuracy: 0.9695 - val_loss: 0.5368 - val_accuracy: 0.8598\n",
            "Epoch 84/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0896 - accuracy: 0.9776 - val_loss: 0.5332 - val_accuracy: 0.8598\n",
            "Epoch 85/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0956 - accuracy: 0.9735 - val_loss: 0.4874 - val_accuracy: 0.8415\n",
            "Epoch 86/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0830 - accuracy: 0.9796 - val_loss: 0.4483 - val_accuracy: 0.8841\n",
            "Epoch 87/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0854 - accuracy: 0.9756 - val_loss: 0.5358 - val_accuracy: 0.8780\n",
            "Epoch 88/200\n",
            "12/12 [==============================] - 0s 29ms/step - loss: 0.0989 - accuracy: 0.9729 - val_loss: 0.4341 - val_accuracy: 0.8780\n",
            "Epoch 89/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0967 - accuracy: 0.9722 - val_loss: 0.4447 - val_accuracy: 0.9085\n",
            "Epoch 90/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0963 - accuracy: 0.9722 - val_loss: 0.5300 - val_accuracy: 0.8659\n",
            "Epoch 91/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.1016 - accuracy: 0.9674 - val_loss: 0.6189 - val_accuracy: 0.8232\n",
            "Epoch 92/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0976 - accuracy: 0.9695 - val_loss: 0.4175 - val_accuracy: 0.8841\n",
            "Epoch 93/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0873 - accuracy: 0.9742 - val_loss: 0.5388 - val_accuracy: 0.8537\n",
            "Epoch 94/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0837 - accuracy: 0.9810 - val_loss: 0.4735 - val_accuracy: 0.8780\n",
            "Epoch 95/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0804 - accuracy: 0.9796 - val_loss: 0.5648 - val_accuracy: 0.8841\n",
            "Epoch 96/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0770 - accuracy: 0.9810 - val_loss: 0.6027 - val_accuracy: 0.8659\n",
            "Epoch 97/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0804 - accuracy: 0.9790 - val_loss: 0.5358 - val_accuracy: 0.8598\n",
            "Epoch 98/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0822 - accuracy: 0.9763 - val_loss: 0.4272 - val_accuracy: 0.8720\n",
            "Epoch 99/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0791 - accuracy: 0.9824 - val_loss: 0.5051 - val_accuracy: 0.8720\n",
            "Epoch 100/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0672 - accuracy: 0.9844 - val_loss: 0.5751 - val_accuracy: 0.8415\n",
            "Epoch 101/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0740 - accuracy: 0.9803 - val_loss: 0.7043 - val_accuracy: 0.8293\n",
            "Epoch 102/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.0580 - accuracy: 0.9864 - val_loss: 0.6150 - val_accuracy: 0.8598\n",
            "Epoch 103/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0652 - accuracy: 0.9864 - val_loss: 0.5054 - val_accuracy: 0.8598\n",
            "Epoch 104/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0628 - accuracy: 0.9858 - val_loss: 0.5547 - val_accuracy: 0.8476\n",
            "Epoch 105/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0598 - accuracy: 0.9878 - val_loss: 0.4767 - val_accuracy: 0.8841\n",
            "Epoch 106/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0569 - accuracy: 0.9851 - val_loss: 0.5587 - val_accuracy: 0.8598\n",
            "Epoch 107/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0659 - accuracy: 0.9810 - val_loss: 0.5305 - val_accuracy: 0.8720\n",
            "Epoch 108/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0547 - accuracy: 0.9858 - val_loss: 0.5077 - val_accuracy: 0.8720\n",
            "Epoch 109/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0496 - accuracy: 0.9885 - val_loss: 0.5036 - val_accuracy: 0.8841\n",
            "Epoch 110/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.0632 - accuracy: 0.9783 - val_loss: 0.5483 - val_accuracy: 0.8963\n",
            "Epoch 111/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0652 - accuracy: 0.9824 - val_loss: 0.4376 - val_accuracy: 0.8841\n",
            "Epoch 112/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0611 - accuracy: 0.9824 - val_loss: 0.5727 - val_accuracy: 0.8720\n",
            "Epoch 113/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0681 - accuracy: 0.9830 - val_loss: 0.5098 - val_accuracy: 0.8720\n",
            "Epoch 114/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0599 - accuracy: 0.9830 - val_loss: 0.5612 - val_accuracy: 0.8841\n",
            "Epoch 115/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0606 - accuracy: 0.9830 - val_loss: 0.6480 - val_accuracy: 0.8720\n",
            "Epoch 116/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0651 - accuracy: 0.9830 - val_loss: 0.5326 - val_accuracy: 0.8720\n",
            "Epoch 117/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0711 - accuracy: 0.9830 - val_loss: 0.8314 - val_accuracy: 0.8354\n",
            "Epoch 118/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0807 - accuracy: 0.9783 - val_loss: 0.5537 - val_accuracy: 0.8780\n",
            "Epoch 119/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0566 - accuracy: 0.9858 - val_loss: 0.5303 - val_accuracy: 0.8780\n",
            "Epoch 120/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0559 - accuracy: 0.9864 - val_loss: 0.5309 - val_accuracy: 0.8720\n",
            "Epoch 121/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0508 - accuracy: 0.9858 - val_loss: 0.5230 - val_accuracy: 0.8963\n",
            "Epoch 122/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0434 - accuracy: 0.9864 - val_loss: 0.5437 - val_accuracy: 0.8902\n",
            "Epoch 123/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0491 - accuracy: 0.9864 - val_loss: 0.6149 - val_accuracy: 0.8598\n",
            "Epoch 124/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0520 - accuracy: 0.9844 - val_loss: 0.4724 - val_accuracy: 0.9024\n",
            "Epoch 125/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0687 - accuracy: 0.9810 - val_loss: 0.5804 - val_accuracy: 0.8659\n",
            "Epoch 126/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0851 - accuracy: 0.9756 - val_loss: 0.6786 - val_accuracy: 0.8415\n",
            "Epoch 127/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0913 - accuracy: 0.9688 - val_loss: 0.4275 - val_accuracy: 0.9268\n",
            "Epoch 128/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0687 - accuracy: 0.9763 - val_loss: 0.5419 - val_accuracy: 0.8720\n",
            "Epoch 129/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0829 - accuracy: 0.9729 - val_loss: 1.3471 - val_accuracy: 0.7012\n",
            "Epoch 130/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0861 - accuracy: 0.9763 - val_loss: 0.4726 - val_accuracy: 0.8963\n",
            "Epoch 131/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0864 - accuracy: 0.9749 - val_loss: 0.6107 - val_accuracy: 0.8780\n",
            "Epoch 132/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0547 - accuracy: 0.9817 - val_loss: 0.5929 - val_accuracy: 0.8780\n",
            "Epoch 133/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0517 - accuracy: 0.9837 - val_loss: 0.5316 - val_accuracy: 0.8902\n",
            "Epoch 134/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0536 - accuracy: 0.9796 - val_loss: 0.6077 - val_accuracy: 0.8659\n",
            "Epoch 135/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0445 - accuracy: 0.9864 - val_loss: 0.5660 - val_accuracy: 0.8902\n",
            "Epoch 136/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0457 - accuracy: 0.9871 - val_loss: 0.6752 - val_accuracy: 0.8780\n",
            "Epoch 137/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0476 - accuracy: 0.9878 - val_loss: 0.7018 - val_accuracy: 0.8720\n",
            "Epoch 138/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0503 - accuracy: 0.9878 - val_loss: 0.5078 - val_accuracy: 0.8902\n",
            "Epoch 139/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0503 - accuracy: 0.9844 - val_loss: 0.5594 - val_accuracy: 0.8598\n",
            "Epoch 140/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0347 - accuracy: 0.9939 - val_loss: 0.5605 - val_accuracy: 0.8780\n",
            "Epoch 141/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0351 - accuracy: 0.9898 - val_loss: 0.4952 - val_accuracy: 0.9024\n",
            "Epoch 142/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0391 - accuracy: 0.9919 - val_loss: 0.6209 - val_accuracy: 0.9024\n",
            "Epoch 143/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0280 - accuracy: 0.9939 - val_loss: 0.6419 - val_accuracy: 0.8841\n",
            "Epoch 144/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0329 - accuracy: 0.9932 - val_loss: 0.5240 - val_accuracy: 0.8902\n",
            "Epoch 145/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0302 - accuracy: 0.9946 - val_loss: 0.5208 - val_accuracy: 0.8902\n",
            "Epoch 146/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0287 - accuracy: 0.9919 - val_loss: 0.5866 - val_accuracy: 0.8720\n",
            "Epoch 147/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0312 - accuracy: 0.9905 - val_loss: 0.6275 - val_accuracy: 0.8841\n",
            "Epoch 148/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0341 - accuracy: 0.9898 - val_loss: 0.4614 - val_accuracy: 0.8902\n",
            "Epoch 149/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0398 - accuracy: 0.9905 - val_loss: 0.7104 - val_accuracy: 0.8476\n",
            "Epoch 150/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0429 - accuracy: 0.9858 - val_loss: 0.5771 - val_accuracy: 0.8780\n",
            "Epoch 151/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0295 - accuracy: 0.9925 - val_loss: 0.6989 - val_accuracy: 0.8537\n",
            "Epoch 152/200\n",
            "12/12 [==============================] - 0s 26ms/step - loss: 0.0389 - accuracy: 0.9912 - val_loss: 0.5410 - val_accuracy: 0.9024\n",
            "Epoch 153/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0229 - accuracy: 0.9946 - val_loss: 0.5456 - val_accuracy: 0.8963\n",
            "Epoch 154/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0477 - accuracy: 0.9864 - val_loss: 0.6977 - val_accuracy: 0.8598\n",
            "Epoch 155/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0378 - accuracy: 0.9898 - val_loss: 0.4679 - val_accuracy: 0.9024\n",
            "Epoch 156/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0405 - accuracy: 0.9858 - val_loss: 0.5789 - val_accuracy: 0.8537\n",
            "Epoch 157/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0303 - accuracy: 0.9919 - val_loss: 0.5495 - val_accuracy: 0.8841\n",
            "Epoch 158/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0295 - accuracy: 0.9932 - val_loss: 0.6032 - val_accuracy: 0.8780\n",
            "Epoch 159/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0431 - accuracy: 0.9891 - val_loss: 0.6204 - val_accuracy: 0.8598\n",
            "Epoch 160/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0395 - accuracy: 0.9858 - val_loss: 0.5563 - val_accuracy: 0.8780\n",
            "Epoch 161/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0472 - accuracy: 0.9844 - val_loss: 0.4975 - val_accuracy: 0.9146\n",
            "Epoch 162/200\n",
            "12/12 [==============================] - 0s 29ms/step - loss: 0.0530 - accuracy: 0.9837 - val_loss: 0.5506 - val_accuracy: 0.8902\n",
            "Epoch 163/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0434 - accuracy: 0.9878 - val_loss: 0.5363 - val_accuracy: 0.8902\n",
            "Epoch 164/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0376 - accuracy: 0.9891 - val_loss: 0.5548 - val_accuracy: 0.8720\n",
            "Epoch 165/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0399 - accuracy: 0.9919 - val_loss: 0.5280 - val_accuracy: 0.8780\n",
            "Epoch 166/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0332 - accuracy: 0.9898 - val_loss: 0.4240 - val_accuracy: 0.8963\n",
            "Epoch 167/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0366 - accuracy: 0.9905 - val_loss: 0.5648 - val_accuracy: 0.8659\n",
            "Epoch 168/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0414 - accuracy: 0.9898 - val_loss: 0.6443 - val_accuracy: 0.8598\n",
            "Epoch 169/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0192 - accuracy: 0.9966 - val_loss: 0.5175 - val_accuracy: 0.8780\n",
            "Epoch 170/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0199 - accuracy: 0.9966 - val_loss: 0.4674 - val_accuracy: 0.8902\n",
            "Epoch 171/200\n",
            "12/12 [==============================] - 0s 29ms/step - loss: 0.0140 - accuracy: 0.9980 - val_loss: 0.5658 - val_accuracy: 0.8780\n",
            "Epoch 172/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0152 - accuracy: 0.9980 - val_loss: 0.5981 - val_accuracy: 0.8780\n",
            "Epoch 173/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0150 - accuracy: 0.9966 - val_loss: 0.5362 - val_accuracy: 0.8902\n",
            "Epoch 174/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0163 - accuracy: 0.9959 - val_loss: 0.5303 - val_accuracy: 0.8902\n",
            "Epoch 175/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0165 - accuracy: 0.9959 - val_loss: 0.6989 - val_accuracy: 0.8659\n",
            "Epoch 176/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0202 - accuracy: 0.9946 - val_loss: 0.6342 - val_accuracy: 0.8902\n",
            "Epoch 177/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0180 - accuracy: 0.9966 - val_loss: 0.6057 - val_accuracy: 0.8902\n",
            "Epoch 178/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0203 - accuracy: 0.9932 - val_loss: 0.5557 - val_accuracy: 0.8902\n",
            "Epoch 179/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0266 - accuracy: 0.9919 - val_loss: 0.4585 - val_accuracy: 0.9146\n",
            "Epoch 180/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0235 - accuracy: 0.9932 - val_loss: 0.5321 - val_accuracy: 0.8902\n",
            "Epoch 181/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0180 - accuracy: 0.9959 - val_loss: 0.6372 - val_accuracy: 0.8902\n",
            "Epoch 182/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0311 - accuracy: 0.9919 - val_loss: 0.5857 - val_accuracy: 0.8659\n",
            "Epoch 183/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0256 - accuracy: 0.9925 - val_loss: 0.4884 - val_accuracy: 0.9085\n",
            "Epoch 184/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0330 - accuracy: 0.9891 - val_loss: 0.5417 - val_accuracy: 0.8841\n",
            "Epoch 185/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0324 - accuracy: 0.9912 - val_loss: 0.5736 - val_accuracy: 0.9085\n",
            "Epoch 186/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0242 - accuracy: 0.9919 - val_loss: 0.5880 - val_accuracy: 0.8902\n",
            "Epoch 187/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0279 - accuracy: 0.9946 - val_loss: 0.7220 - val_accuracy: 0.8415\n",
            "Epoch 188/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0233 - accuracy: 0.9932 - val_loss: 0.4928 - val_accuracy: 0.8963\n",
            "Epoch 189/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0302 - accuracy: 0.9891 - val_loss: 0.4666 - val_accuracy: 0.9085\n",
            "Epoch 190/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0215 - accuracy: 0.9932 - val_loss: 0.6593 - val_accuracy: 0.8780\n",
            "Epoch 191/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0263 - accuracy: 0.9925 - val_loss: 0.9331 - val_accuracy: 0.8171\n",
            "Epoch 192/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0332 - accuracy: 0.9905 - val_loss: 0.6651 - val_accuracy: 0.8659\n",
            "Epoch 193/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0427 - accuracy: 0.9851 - val_loss: 0.5620 - val_accuracy: 0.8963\n",
            "Epoch 194/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0209 - accuracy: 0.9932 - val_loss: 0.4683 - val_accuracy: 0.8780\n",
            "Epoch 195/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0365 - accuracy: 0.9891 - val_loss: 0.4869 - val_accuracy: 0.8963\n",
            "Epoch 196/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0229 - accuracy: 0.9946 - val_loss: 0.5549 - val_accuracy: 0.8659\n",
            "Epoch 197/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0298 - accuracy: 0.9925 - val_loss: 0.4760 - val_accuracy: 0.8902\n",
            "Epoch 198/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0403 - accuracy: 0.9898 - val_loss: 0.7122 - val_accuracy: 0.8720\n",
            "Epoch 199/200\n",
            "12/12 [==============================] - 0s 28ms/step - loss: 0.0300 - accuracy: 0.9919 - val_loss: 0.6932 - val_accuracy: 0.8598\n",
            "Epoch 200/200\n",
            "12/12 [==============================] - 0s 27ms/step - loss: 0.0338 - accuracy: 0.9891 - val_loss: 0.6070 - val_accuracy: 0.8720\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZCNQ-GJB7ll2"
      },
      "source": [
        "## 로스 그래프"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "mntqs26g6afW",
        "outputId": "a7ac2dee-be85-4996-b812-279c83c74945"
      },
      "source": [
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.show()"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JYcVirDQ7nhb"
      },
      "source": [
        "## 성능 보기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XwCFA_mz2PTz",
        "outputId": "bd153d47-6680-44e0-9d3d-e3eb0e747426"
      },
      "source": [
        "loss, acc = model.evaluate(test_x, test_y)\n",
        "print(loss, acc)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "13/13 [==============================] - 0s 13ms/step - loss: 0.6614 - accuracy: 0.8659\n",
            "0.6613637804985046 0.8658536672592163\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yNQkQfWO7KrX"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}