pythorch.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "pythorch.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPtFki/25aAq4beAkSqTPs1",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Junseokee/Study-Python/blob/main/pythorch.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#데이콘 수화데이터 분류"
      ],
      "metadata": {
        "id": "5B9O8Pt3YDtm"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "##환경 설정 및 데이터 로드"
      ],
      "metadata": {
        "id": "k9rAsXMkYKd0"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HTiJuuG1Tr52",
        "outputId": "9512d3b4-9ec1-42ab-b8cc-1612e9afc202"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "%cd \"/content/drive/MyDrive/Colab Notebooks/dacon\"\n",
        "!unzip \"user_data.zip\""
      ],
      "metadata": {
        "id": "ZbUqPDbyUQJi"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "import torch\n",
        "device = torch.device(\"cuda\") if torch.cuda.is_available() else torch.device('cpu') #GPU 할당"
      ],
      "metadata": {
        "id": "dk2wYu0yUQR6"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#하이퍼 파라미터 튜닝\n",
        "CFG = {\n",
        "  'IMG_SIZE':128, #이미지 사이즈\n",
        "  'EPOCHS':50, #에포크\n",
        "  'LEARNING_RATE':2e-2, #학습률\n",
        "  'BATCH_SIZE':12, #배치사이즈\n",
        "  'SEED':41 #시드\n",
        "}"
      ],
      "metadata": {
        "id": "2R4oizOyUQWB"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import random\n",
        "import numpy as np"
      ],
      "metadata": {
        "id": "DBiaVaLEUQZ5"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# random seed 고정\n",
        "def seed_everything(seed):\n",
        "  random.seed(seed)\n",
        "  os.environ['PYTHONHASHSEED'] = str(seed)\n",
        "  np.random.seed(seed)\n",
        "  torch.manual_seed(seed)\n",
        "  torch.cuda.manual_seed(seed)\n",
        "  torch.backends.cudnn.deterministic = True\n",
        "  torch.backends.cudnn.benchmark = True\n",
        "\n",
        "seed_everything(CFG['SEED'])"
      ],
      "metadata": {
        "id": "kXdpeV5hUQdh"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "label_df = pd.read_csv('/content/drive/MyDrive/Colab Notebooks/dacon/train.csv')\n",
        "label_df.head() #라벨 확인"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "NT3sB-2gUQhJ",
        "outputId": "9628b6a0-a820-4584-b0bd-7cc6277ae644"
      },
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  file_name label\n",
              "0   001.png  10-2\n",
              "1   002.png  10-1\n",
              "2   003.png     3\n",
              "3   004.png     8\n",
              "4   005.png     9"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-fa219182-4692-4217-a4a1-bfa70d2f5ec7\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>file_name</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>001.png</td>\n",
              "      <td>10-2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>002.png</td>\n",
              "      <td>10-1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>003.png</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>004.png</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>005.png</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-fa219182-4692-4217-a4a1-bfa70d2f5ec7')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-fa219182-4692-4217-a4a1-bfa70d2f5ec7 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-fa219182-4692-4217-a4a1-bfa70d2f5ec7');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "label_df.info()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tsc71kSSUQk4",
        "outputId": "54621502-d0f7-4e5a-aa79-8bb6592b36ff"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 858 entries, 0 to 857\n",
            "Data columns (total 2 columns):\n",
            " #   Column     Non-Null Count  Dtype \n",
            "---  ------     --------------  ----- \n",
            " 0   file_name  858 non-null    object\n",
            " 1   label      858 non-null    object\n",
            "dtypes: object(2)\n",
            "memory usage: 13.5+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##데이터 전처리"
      ],
      "metadata": {
        "id": "tbD9IEUCXoyd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#10-1 10-2라벨을 10 과 0으로 변경\n",
        "label_df['label'][label_df['label'] == '10-1'] = 10 ##label : 10-1 -> 10\n",
        "label_df['label'][label_df['label'] == '10-2'] = 0\n",
        "label_df['label'] = label_df['label'].apply(lambda x : int(x)) ##Dtype : object -> int"
      ],
      "metadata": {
        "id": "XGL7XTWnYRGV"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "label_df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "faiO1pVyYRjW",
        "outputId": "f8d9f95a-f6be-4f6c-b391-54f1839b6913"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "  file_name  label\n",
              "0   001.png      0\n",
              "1   002.png     10\n",
              "2   003.png      3\n",
              "3   004.png      8\n",
              "4   005.png      9"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-85072265-f260-4a76-b171-baf2a2b3c4d4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>file_name</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>001.png</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>002.png</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>003.png</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>004.png</td>\n",
              "      <td>8</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>005.png</td>\n",
              "      <td>9</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-85072265-f260-4a76-b171-baf2a2b3c4d4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-85072265-f260-4a76-b171-baf2a2b3c4d4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-85072265-f260-4a76-b171-baf2a2b3c4d4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "label_df.info() #label이 object -> int64로 변경"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V7vOMk0eYRp9",
        "outputId": "3d0b1b88-ebcf-4266-d498-454bcdf95434"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 858 entries, 0 to 857\n",
            "Data columns (total 2 columns):\n",
            " #   Column     Non-Null Count  Dtype \n",
            "---  ------     --------------  ----- \n",
            " 0   file_name  858 non-null    object\n",
            " 1   label      858 non-null    int64 \n",
            "dtypes: int64(1), object(1)\n",
            "memory usage: 13.5+ KB\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#데이터 이미지의 local adress와 label 값을 list에 저장\n",
        "from glob import glob\n",
        "\n",
        "def get_train_data(data_dir):\n",
        "    img_path_list = []\n",
        "    label_list = []\n",
        "    \n",
        "    # get image path\n",
        "    img_path_list.extend(glob(os.path.join(data_dir, '*.png')))\n",
        "    img_path_list.sort(key=lambda x:int(x.split('/')[-1].split('.')[0]))\n",
        "    #print('wow', img_path_list)\n",
        "        \n",
        "    # get label\n",
        "    #label_df = pd.read_csv(data_dir+'/train.csv')\n",
        "    label_list.extend(label_df['label'])\n",
        "    #print('wow2', label_list)\n",
        "                \n",
        "    return img_path_list, label_list\n",
        "\n",
        "def get_test_data(data_dir):\n",
        "    img_path_list = []\n",
        "    \n",
        "    # get image path\n",
        "    img_path_list.extend(glob(os.path.join(data_dir, '*.png')))\n",
        "    img_path_list.sort(key=lambda x:int(x.split('/')[-1].split('.')[0]))\n",
        "    #print(img_path_list)\n",
        "    \n",
        "    return img_path_list"
      ],
      "metadata": {
        "id": "9MB644IDYRsd"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "all_img_path, all_label = get_train_data('/content/drive/MyDrive/Colab Notebooks/dacon/train')\n",
        "test_img_path = get_test_data('/content/drive/MyDrive/Colab Notebooks/dacon/test')"
      ],
      "metadata": {
        "id": "S6c0ORltYRvP"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "all_label[0:7]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VVNjRPy4YRzI",
        "outputId": "c52c43c6-d4bf-4f19-999c-5904a71fad13"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0, 10, 3, 8, 9, 9, 8]"
            ]
          },
          "metadata": {},
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##DataSet 구성"
      ],
      "metadata": {
        "id": "xxNGaqgxYRzn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import torchvision.datasets as datasets # 이미지 데이터셋 집합체\n",
        "import torchvision.transforms as transforms # 이미지 변환 툴\n",
        "\n",
        "from torch.utils.data import DataLoader # 학습 및 배치로 모델에 넣어주기 위한 툴\n",
        "from torch.utils.data import DataLoader, Dataset\n",
        "\n",
        "class CustomDataset(Dataset):\n",
        "    def __init__(self, img_path_list, label_list, train_mode=True, transforms=None): #필요한 변수들을 선언\n",
        "        self.transforms = transforms\n",
        "        self.train_mode = train_mode\n",
        "        self.img_path_list = img_path_list\n",
        "        self.label_list = label_list\n",
        "\n",
        "    def __getitem__(self, index): #index번째 data를 return\n",
        "        img_path = self.img_path_list[index]\n",
        "        # Get image data\n",
        "        image = cv2.imread(img_path)\n",
        "        if self.transforms is not None:\n",
        "            image = self.transforms(image)\n",
        "\n",
        "        if self.train_mode:\n",
        "            label = self.label_list[index]\n",
        "            return image, label\n",
        "        else:\n",
        "            return image\n",
        "    \n",
        "    def __len__(self): #길이 return\n",
        "        return len(self.img_path_list)"
      ],
      "metadata": {
        "id": "MCT3JcQMYR0J"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Train : Validation = 0.75 : 0.25 Split\n",
        "train_len = int(len(all_img_path)*0.75)\n",
        "Vali_len = int(len(all_img_path)*0.25)\n",
        "\n",
        "train_img_path = all_img_path[:train_len]\n",
        "train_label = all_label[:train_len]\n",
        "\n",
        "vali_img_path = all_img_path[train_len:]\n",
        "vali_label = all_label[train_len:]"
      ],
      "metadata": {
        "id": "QQS6-YoaYR0l"
      },
      "execution_count": 81,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print('train set 길이 : ', train_len)\n",
        "print('vaildation set 길이 : ', Vali_len)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Rv2qTyBYYR1V",
        "outputId": "5ba71494-36b8-475f-85be-990221c2a84a"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "train set 길이 :  643\n",
            "vaildation set 길이 :  214\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_transform = transforms.Compose([\n",
        "                    transforms.ToPILImage(), #Numpy배열에서 PIL이미지로\n",
        "                    transforms.Resize([CFG['IMG_SIZE'], CFG['IMG_SIZE']]), #이미지 사이즈 변형\n",
        "                    transforms.ToTensor(), #이미지 데이터를 tensor\n",
        "                    transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5)) #이미지 정규화\n",
        "                    \n",
        "                    ])\n",
        "\n",
        "test_transform = transforms.Compose([\n",
        "                    transforms.ToPILImage(),\n",
        "                    transforms.Resize([CFG['IMG_SIZE'], CFG['IMG_SIZE']]),\n",
        "                    transforms.ToTensor(),\n",
        "                    transforms.Normalize(mean=(0.5, 0.5, 0.5), std=(0.5, 0.5, 0.5))\n",
        "                    ])"
      ],
      "metadata": {
        "id": "iA_SJ1KOknjp"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Get Dataloader\n",
        "\n",
        "#CustomDataset class를 통하여 train dataset생성\n",
        "train_dataset = CustomDataset(train_img_path, train_label, train_mode=True, transforms=train_transform) \n",
        "#만든 train dataset를 DataLoader에 넣어 batch 만들기\n",
        "train_loader = DataLoader(train_dataset, batch_size = CFG['BATCH_SIZE'], shuffle=True, num_workers=0) #BATCH_SIZE : 24\n",
        "\n",
        "#vaildation 에서도 적용\n",
        "vali_dataset = CustomDataset(vali_img_path, vali_label, train_mode=True, transforms=test_transform)\n",
        "vali_loader = DataLoader(vali_dataset, batch_size = CFG['BATCH_SIZE'], shuffle=False, num_workers=0)\n",
        "\n",
        "train_batches = len(train_loader)\n",
        "vali_batches = len(vali_loader)\n",
        "\n",
        "print('total train imgs :',train_len,'/ total train batches :', train_batches)\n",
        "print('total valid imgs :',Vali_len, '/ total valid batches :', vali_batches)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xzYlfJ7Gltlh",
        "outputId": "bf7205b9-028b-4671-a55b-22e1fd295208"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "total train imgs : 643 / total train batches : 54\n",
            "total valid imgs : 214 / total valid batches : 18\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_batches = len(train_loader)\n",
        "vali_batches = len(vali_loader)\n",
        "\n",
        "print('total train imgs :',train_len,'/ total train batches :', train_batches)\n",
        "print('total valid imgs :',Vali_len, '/ total valid batches :', vali_batches)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3QTxjWxglvV4",
        "outputId": "1329e2f6-78cc-45ac-f885-07ac7ed0ccf6"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "total train imgs : 643 / total train batches : 54\n",
            "total valid imgs : 214 / total valid batches : 18\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "train_features, train_labels = next(iter(train_loader)) #iter는 반복 가능한 객체에서 이터레이터를 반환하고, \n",
        "                                                        #next는 이터레이터에서 값을 차례대로 꺼냅니다. \n",
        "img = train_features[0]\n",
        "label = train_labels[0]\n",
        "plt.imshow(img[0], cmap=\"gray\")\n",
        "plt.show()\n",
        "print(f\"Label: {label}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 286
        },
        "id": "19TPFJvnnxtI",
        "outputId": "2411e95f-30f6-4c95-e58a-0b8a60a6d036"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Label: 10\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"Feature batch shape: {train_features.size()}\")\n",
        "print(f\"Batch Labels shape: {train_labels}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dlPQxGWTn6gH",
        "outputId": "e90afe8e-1206-4441-de8c-9b212024d92d"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Feature batch shape: torch.Size([12, 3, 128, 128])\n",
            "Batch Labels shape: tensor([10,  2,  9,  7,  6,  2,  4,  7,  1,  8,  0,  9])\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tqdm.auto import tqdm\n",
        "import torch.nn as nn # 신경망들이 포함됨\n",
        "#import torch.nn.init as init # 텐서에 초기값을 줌\n",
        "\n",
        "class CNNclassification(torch.nn.Module):\n",
        "    def __init__(self):\n",
        "        super(CNNclassification, self).__init__()\n",
        "        self.layer1 = torch.nn.Sequential(\n",
        "            nn.Conv2d(3, 8, kernel_size=3, stride=1, padding=1), #cnn layer\n",
        "            nn.ReLU(), #activation function\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)) #pooling layer\n",
        "        \n",
        "        self.layer2 = torch.nn.Sequential(\n",
        "            nn.Conv2d(8, 16, kernel_size=3, stride=1, padding=1), #cnn layer\n",
        "            nn.ReLU(), #activation function\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)) #pooling layer\n",
        "        \n",
        "        self.layer3 = torch.nn.Sequential(\n",
        "            nn.Conv2d(16, 32, kernel_size=3, stride=1, padding=1), #cnn layer\n",
        "            nn.ReLU(), #activation function\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)) #pooling layer\n",
        "        \n",
        "        self.layer4 = torch.nn.Sequential(\n",
        "            nn.Conv2d(32, 64, kernel_size=4, stride=1, padding=1), #cnn layer\n",
        "            nn.ReLU(), #activation function\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)) #pooling layer\n",
        "        \n",
        "        self.fc_layer = nn.Sequential( \n",
        "            nn.Linear(3136, 11) #fully connected layer(ouput layer)\n",
        "        )    \n",
        "        \n",
        "    def forward(self, x):\n",
        "        \n",
        "        x = self.layer1(x) #1층\n",
        "        \n",
        "        x = self.layer2(x) #2층\n",
        "         \n",
        "        x = self.layer3(x) #3층\n",
        "        \n",
        "        x = self.layer4(x) #4층\n",
        "        \n",
        "        x = torch.flatten(x, start_dim=1) # N차원 배열 -> 1차원 배열\n",
        "        \n",
        "        out = self.fc_layer(x)\n",
        "        return out"
      ],
      "metadata": {
        "id": "dxHDcu0ioXfg"
      },
      "execution_count": 88,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "tev8dSSCootQ"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}