Chapter_09_텍스트를_위한_인공_신경망.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPGZY+Fd6l4EyrtWxSRazfp",
      "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/Sangyeonglee353/ai-hongong/blob/main/Chapter_09_%ED%85%8D%EC%8A%A4%ED%8A%B8%EB%A5%BC_%EC%9C%84%ED%95%9C_%EC%9D%B8%EA%B3%B5_%EC%8B%A0%EA%B2%BD%EB%A7%9D.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Ch09-1 | 순차 데이터와 순환 신경망"
      ],
      "metadata": {
        "id": "5rzl1ubOie92"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Ch09-2 | 순환 신경망으로 IMDB 리뷰 분류하기"
      ],
      "metadata": {
        "id": "jMLvXiIXizRB"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## IMDB 리뷰 데이터셋"
      ],
      "metadata": {
        "id": "0L3yuPPXYFta"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. IMDB 리뷰 데이터셋 가져오기\n",
        "# [참고] imdb.load_data: 전체 어휘 사전에 있는 단어를 등장 순서대로 나열한 다음 가장 많이 등장한 {num_words} 단어 선택\n",
        "from tensorflow.keras.datasets import imdb\n",
        "(train_input, train_target), (test_input, test_target) = imdb.load_data(num_words=300)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rOBmHDaaYKLp",
        "outputId": "5e691c7a-0203-442a-da94-1438eaf83bbb"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz\n",
            "17464789/17464789 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 훈련 세트와 테스트 세트 크기 확인\n",
        "## 리뷰 텍스트 길이가 제각각 -> 별도의 파이썬 리스트로 넘파이 배열로 보관 -> 메모리 효율적 사용\n",
        "print(train_input.shape, test_input.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4La-NRjtYZXo",
        "outputId": "ecc79fef-e16c-4d0f-e9a0-e8a9c2688ff6"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(25000,) (25000,)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3-1. 첫 번째 리뷰 길이 확인\n",
        "## [참고] 넘파이 배열: 정수나 실수 외에도 파이썬 객체로 담을 수 있음.\n",
        "print(len(train_input[0]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6oB0tuK2Yydg",
        "outputId": "eccb7b9f-68a1-4716-c159-cd5c087265cb"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "218\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3-2. 두 번째 리뷰 길이 확인\n",
        "print(len(train_input[1]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UEj-wITzY_Xy",
        "outputId": "ac856608-86af-4fbc-c66c-6a64ebb57ffe"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "189\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3-3. 첫 번째 리뷰에 담긴 내용 출력\n",
        "print(train_input[0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8SPf3f0cZKiF",
        "outputId": "7b142d18-1f21-4dc2-a0ad-1f3a2aa2535a"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[1, 14, 22, 16, 43, 2, 2, 2, 2, 65, 2, 2, 66, 2, 4, 173, 36, 256, 5, 25, 100, 43, 2, 112, 50, 2, 2, 9, 35, 2, 284, 5, 150, 4, 172, 112, 167, 2, 2, 2, 39, 4, 172, 2, 2, 17, 2, 38, 13, 2, 4, 192, 50, 16, 6, 147, 2, 19, 14, 22, 4, 2, 2, 2, 4, 22, 71, 87, 12, 16, 43, 2, 38, 76, 15, 13, 2, 4, 22, 17, 2, 17, 12, 16, 2, 18, 2, 5, 62, 2, 12, 8, 2, 8, 106, 5, 4, 2, 2, 16, 2, 66, 2, 33, 4, 130, 12, 16, 38, 2, 5, 25, 124, 51, 36, 135, 48, 25, 2, 33, 6, 22, 12, 215, 28, 77, 52, 5, 14, 2, 16, 82, 2, 8, 4, 107, 117, 2, 15, 256, 4, 2, 7, 2, 5, 2, 36, 71, 43, 2, 2, 26, 2, 2, 46, 7, 4, 2, 2, 13, 104, 88, 4, 2, 15, 297, 98, 32, 2, 56, 26, 141, 6, 194, 2, 18, 4, 226, 22, 21, 134, 2, 26, 2, 5, 144, 30, 2, 18, 51, 36, 28, 224, 92, 25, 104, 4, 226, 65, 16, 38, 2, 88, 12, 16, 283, 5, 16, 2, 113, 103, 32, 15, 16, 2, 19, 178, 32]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. 타깃 데이터 출력\n",
        "print(train_target[:20])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1JhRAcBfZxFo",
        "outputId": "4965c249-dfa6-48c0-baa3-1fe16d41f56d"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[1 0 0 1 0 0 1 0 1 0 1 0 0 0 0 0 1 1 0 1]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 긍정(0), 부정(1) 판단 -> 이진분류 문제\n",
        "# 5. 훈련 세트와 검증 세트 분리(8:2)\n",
        "from sklearn.model_selection import train_test_split\n",
        "train_input, val_input, train_target, val_target = train_test_split(train_input, train_target, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "mIUhhyJFZ2y1"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 6. 훈련 세트 조사\n",
        "# 6.1 리뷰의 길이 계산 for 평균 리뷰 길이, 최소 리뷰 길이, 최대 리뷰 길이 확인\n",
        "import numpy as np\n",
        "lengths = np.array([len(x) for x in train_input])\n",
        "\n",
        "# 6.2 리뷰 길이의 평균과 중간값 구하기\n",
        "print(np.mean(lengths), np.median(lengths))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gAGgANA2aqph",
        "outputId": "18a5585f-e397-4edc-afb4-4412bf1d8e1d"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "239.00925 178.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 6.3 lengths 배열 > 히스토그램으로 표현\n",
        "## [결과 해석] 대부분 리뷰 길이 300미만, 대부분 짧으므로 100개의 단어만 사용\n",
        "## [결과 해석] 단, 100보다 작은 리뷰는 패딩(0)으로 길이를 맞춤.\n",
        "import matplotlib.pyplot as plt\n",
        "plt.hist(lengths)\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('frequency')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "9gqYsXNIbfgm",
        "outputId": "dab9af16-9cdc-4333-b16f-0f12c995555c"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 7. 길이 100으로 조정\n",
        "# 7.1 길이 조정\n",
        "# [참고] pad_sequence(): 길이가 짧으면 앞부분을 0으로 패딩\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "train_seq = pad_sequences(train_input, maxlen=100)\n",
        "\n",
        "# 7.2 train_seq 크기 확인\n",
        "print(train_seq.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lw3z2ER_cu-o",
        "outputId": "b95acd3a-0854-4902-85b2-c5fb9f77d942"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(20000, 100)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 7.3 첫 번째 샘플 출력\n",
        "print(train_seq[0])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tE-K-627dRNu",
        "outputId": "b474b431-5a35-4f60-9872-6d851871547f"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[ 10   4  20   9   2   2   2   5  45   6   2   2  33 269   8   2 142   2\n",
            "   5   2  17  73  17 204   5   2  19  55   2   2  92  66 104  14  20  93\n",
            "  76   2 151  33   4  58  12 188   2 151  12 215  69 224 142  73 237   6\n",
            "   2   7   2   2 188   2 103  14  31  10  10   2   7   2   5   2  80  91\n",
            "   2  30   2  34  14  20 151  50  26 131  49   2  84  46  50  37  80  79\n",
            "   6   2  46   7  14  20  10  10   2 158]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 7.4 원본 샘플 끝 확인(잘린 부분 체크)\n",
        "## [결과 해석] 앞 부분이 짤림. 뒷 부분과 일치하므로\n",
        "print(train_input[0][-10:])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fhCVW4rHdWUw",
        "outputId": "d195933f-9846-4a25-c7f5-214b1be3628d"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[6, 2, 46, 7, 14, 20, 10, 10, 2, 158]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 7.5 여섯 번째 샘플 출력\n",
        "print(train_seq[5])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T_KsDNKxdot1",
        "outputId": "af15131c-beec-4aa9-d236-c4ca8641aa5b"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[  0   0   0   0   1   2 195  19  49   2   2 190   4   2   2   2 183  10\n",
            "  10  13  82  79   4   2  36  71 269   8   2  25  19  49   7   4   2   2\n",
            "   2   2   2  10  10  48  25  40   2  11   2   2  40   2   2   5   4   2\n",
            "   2  95  14 238  56 129   2  10  10  21   2  94   2   2   2   2  11 190\n",
            "  24   2   2   7  94 205   2  10  10  87   2  34  49   2   7   2   2   2\n",
            "   2   2 290   2  46  48  64  18   4   2]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 7.6 검증 세트의 길이도 100으로 조정\n",
        "val_seq = pad_sequences(val_input, maxlen=100)"
      ],
      "metadata": {
        "id": "hvGwZkVEdvsf"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 순환 신경망 만들기"
      ],
      "metadata": {
        "id": "T8sXGHVWd8i8"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. SimpleRNN 클래스 사용\n",
        "from tensorflow import keras\n",
        "model = keras.Sequential()\n",
        "model.add(keras.layers.SimpleRNN(8, input_shape=(100, 300)))\n",
        "model.add(keras.layers.Dense(1, activation='sigmoid')) # 이진 분류이므로, 활성화 함수: sigmoid"
      ],
      "metadata": {
        "id": "HEoEdGnud_MG"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 원-핫 인코딩\n",
        "# 2.1 인코딩 진행\n",
        "train_oh = keras.utils.to_categorical(train_seq)\n",
        "\n",
        "# 2.2 배열 크기 확인\n",
        "print(train_oh.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gFEeR2A9f5RQ",
        "outputId": "22c4d5ea-b74c-4875-92ab-8ce633e33815"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(20000, 100, 300)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2.3 인코딩 결과 확인\n",
        "# 첫 번째 토큰의 10의 인코딩 확인\n",
        "print(train_oh[0][0][:12])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s4_h44RSicmB",
        "outputId": "0655a281-37dd-49e2-dabb-293cb6436817"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2.4 나머지가 모두 0인지 확인\n",
        "print(np.sum(train_oh[0][0]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ugo3kHVNisqg",
        "outputId": "ab960c1d-6b03-4b4a-ca33-0e24daa52344"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1.0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2.5 검증 세트의 원-핫 인코딩\n",
        "val_oh = keras.utils.to_categorical(val_seq)"
      ],
      "metadata": {
        "id": "XCu74pkoi0Pc"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2.6 모델 구조 출력\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tV4QfLuZjHgQ",
        "outputId": "914ba97a-973f-49f9-96d8-746da2f79d28"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " simple_rnn_1 (SimpleRNN)    (None, 8)                 2472      \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 9         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2481 (9.69 KB)\n",
            "Trainable params: 2481 (9.69 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 순환 신경망 훈련하기"
      ],
      "metadata": {
        "id": "HZL71LbmkHcM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. 모델 훈련\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-simplernn-model.h5', save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model.fit(train_oh, train_target, epochs=100, batch_size=64,\n",
        "                    validation_data=(val_oh, val_target),\n",
        "                    callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T6q4xdNIkKBo",
        "outputId": "ff9ef038-27ef-466f-950c-461a51b07f23"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 13s 38ms/step - loss: 0.7033 - accuracy: 0.5033 - val_loss: 0.6985 - val_accuracy: 0.5094\n",
            "Epoch 2/100\n",
            "  7/313 [..............................] - ETA: 6s - loss: 0.7053 - accuracy: 0.4911"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 10s 32ms/step - loss: 0.6961 - accuracy: 0.5103 - val_loss: 0.6955 - val_accuracy: 0.5176\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.6927 - accuracy: 0.5196 - val_loss: 0.6929 - val_accuracy: 0.5220\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 10s 31ms/step - loss: 0.6898 - accuracy: 0.5319 - val_loss: 0.6904 - val_accuracy: 0.5312\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.6867 - accuracy: 0.5440 - val_loss: 0.6870 - val_accuracy: 0.5448\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 8s 27ms/step - loss: 0.6824 - accuracy: 0.5604 - val_loss: 0.6815 - val_accuracy: 0.5634\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 8s 27ms/step - loss: 0.6743 - accuracy: 0.5848 - val_loss: 0.6696 - val_accuracy: 0.6074\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 9s 28ms/step - loss: 0.6591 - accuracy: 0.6217 - val_loss: 0.6541 - val_accuracy: 0.6306\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.6470 - accuracy: 0.6379 - val_loss: 0.6431 - val_accuracy: 0.6456\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.6377 - accuracy: 0.6513 - val_loss: 0.6359 - val_accuracy: 0.6498\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.6307 - accuracy: 0.6582 - val_loss: 0.6278 - val_accuracy: 0.6592\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.6226 - accuracy: 0.6661 - val_loss: 0.6220 - val_accuracy: 0.6658\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 8s 24ms/step - loss: 0.6157 - accuracy: 0.6730 - val_loss: 0.6201 - val_accuracy: 0.6644\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.6081 - accuracy: 0.6816 - val_loss: 0.6069 - val_accuracy: 0.6786\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.6007 - accuracy: 0.6898 - val_loss: 0.6014 - val_accuracy: 0.6844\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5945 - accuracy: 0.6963 - val_loss: 0.5949 - val_accuracy: 0.6916\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 10s 31ms/step - loss: 0.5894 - accuracy: 0.7020 - val_loss: 0.5911 - val_accuracy: 0.6942\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 10s 30ms/step - loss: 0.5846 - accuracy: 0.7058 - val_loss: 0.5901 - val_accuracy: 0.6934\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 11s 37ms/step - loss: 0.5811 - accuracy: 0.7090 - val_loss: 0.5858 - val_accuracy: 0.6966\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 15s 47ms/step - loss: 0.5760 - accuracy: 0.7135 - val_loss: 0.5811 - val_accuracy: 0.6996\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 17s 55ms/step - loss: 0.5727 - accuracy: 0.7161 - val_loss: 0.5781 - val_accuracy: 0.7022\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 16s 53ms/step - loss: 0.5698 - accuracy: 0.7183 - val_loss: 0.5743 - val_accuracy: 0.7064\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.5665 - accuracy: 0.7207 - val_loss: 0.5725 - val_accuracy: 0.7070\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 10s 31ms/step - loss: 0.5628 - accuracy: 0.7232 - val_loss: 0.5851 - val_accuracy: 0.6974\n",
            "Epoch 25/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5590 - accuracy: 0.7261 - val_loss: 0.5673 - val_accuracy: 0.7134\n",
            "Epoch 26/100\n",
            "313/313 [==============================] - 8s 27ms/step - loss: 0.5556 - accuracy: 0.7286 - val_loss: 0.5626 - val_accuracy: 0.7176\n",
            "Epoch 27/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.5515 - accuracy: 0.7323 - val_loss: 0.5599 - val_accuracy: 0.7186\n",
            "Epoch 28/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.5481 - accuracy: 0.7358 - val_loss: 0.5588 - val_accuracy: 0.7198\n",
            "Epoch 29/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.5450 - accuracy: 0.7386 - val_loss: 0.5543 - val_accuracy: 0.7230\n",
            "Epoch 30/100\n",
            "313/313 [==============================] - 11s 35ms/step - loss: 0.5411 - accuracy: 0.7422 - val_loss: 0.5512 - val_accuracy: 0.7280\n",
            "Epoch 31/100\n",
            "313/313 [==============================] - 16s 50ms/step - loss: 0.5379 - accuracy: 0.7423 - val_loss: 0.5485 - val_accuracy: 0.7300\n",
            "Epoch 32/100\n",
            "313/313 [==============================] - 16s 51ms/step - loss: 0.5353 - accuracy: 0.7455 - val_loss: 0.5462 - val_accuracy: 0.7288\n",
            "Epoch 33/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5330 - accuracy: 0.7456 - val_loss: 0.5448 - val_accuracy: 0.7278\n",
            "Epoch 34/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5299 - accuracy: 0.7493 - val_loss: 0.5432 - val_accuracy: 0.7320\n",
            "Epoch 35/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5276 - accuracy: 0.7502 - val_loss: 0.5418 - val_accuracy: 0.7300\n",
            "Epoch 36/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5246 - accuracy: 0.7539 - val_loss: 0.5400 - val_accuracy: 0.7322\n",
            "Epoch 37/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5226 - accuracy: 0.7539 - val_loss: 0.5374 - val_accuracy: 0.7342\n",
            "Epoch 38/100\n",
            "313/313 [==============================] - 8s 27ms/step - loss: 0.5208 - accuracy: 0.7548 - val_loss: 0.5386 - val_accuracy: 0.7310\n",
            "Epoch 39/100\n",
            "313/313 [==============================] - 9s 27ms/step - loss: 0.5188 - accuracy: 0.7545 - val_loss: 0.5373 - val_accuracy: 0.7322\n",
            "Epoch 40/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5166 - accuracy: 0.7576 - val_loss: 0.5344 - val_accuracy: 0.7348\n",
            "Epoch 41/100\n",
            "313/313 [==============================] - 10s 31ms/step - loss: 0.5148 - accuracy: 0.7584 - val_loss: 0.5347 - val_accuracy: 0.7352\n",
            "Epoch 42/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5131 - accuracy: 0.7581 - val_loss: 0.5350 - val_accuracy: 0.7372\n",
            "Epoch 43/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.5108 - accuracy: 0.7589 - val_loss: 0.5326 - val_accuracy: 0.7372\n",
            "Epoch 44/100\n",
            "313/313 [==============================] - 10s 30ms/step - loss: 0.5101 - accuracy: 0.7602 - val_loss: 0.5357 - val_accuracy: 0.7344\n",
            "Epoch 45/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5081 - accuracy: 0.7619 - val_loss: 0.5347 - val_accuracy: 0.7360\n",
            "Epoch 46/100\n",
            "313/313 [==============================] - 11s 34ms/step - loss: 0.5067 - accuracy: 0.7634 - val_loss: 0.5324 - val_accuracy: 0.7380\n",
            "Epoch 47/100\n",
            "313/313 [==============================] - 11s 35ms/step - loss: 0.5062 - accuracy: 0.7618 - val_loss: 0.5297 - val_accuracy: 0.7412\n",
            "Epoch 48/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.5043 - accuracy: 0.7634 - val_loss: 0.5301 - val_accuracy: 0.7382\n",
            "Epoch 49/100\n",
            "313/313 [==============================] - 10s 31ms/step - loss: 0.5029 - accuracy: 0.7650 - val_loss: 0.5287 - val_accuracy: 0.7412\n",
            "Epoch 50/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5021 - accuracy: 0.7653 - val_loss: 0.5299 - val_accuracy: 0.7386\n",
            "Epoch 51/100\n",
            "313/313 [==============================] - 8s 26ms/step - loss: 0.5008 - accuracy: 0.7650 - val_loss: 0.5300 - val_accuracy: 0.7422\n",
            "Epoch 52/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5001 - accuracy: 0.7660 - val_loss: 0.5269 - val_accuracy: 0.7430\n",
            "Epoch 53/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.4988 - accuracy: 0.7660 - val_loss: 0.5340 - val_accuracy: 0.7368\n",
            "Epoch 54/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.4985 - accuracy: 0.7673 - val_loss: 0.5255 - val_accuracy: 0.7448\n",
            "Epoch 55/100\n",
            "313/313 [==============================] - 9s 30ms/step - loss: 0.4973 - accuracy: 0.7660 - val_loss: 0.5271 - val_accuracy: 0.7446\n",
            "Epoch 56/100\n",
            "313/313 [==============================] - 10s 30ms/step - loss: 0.4954 - accuracy: 0.7680 - val_loss: 0.5261 - val_accuracy: 0.7456\n",
            "Epoch 57/100\n",
            "313/313 [==============================] - 8s 27ms/step - loss: 0.4951 - accuracy: 0.7685 - val_loss: 0.5256 - val_accuracy: 0.7446\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 훈련 손실과 검증 손실 그래프 출력\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "zmZOso0YmGwj",
        "outputId": "52b26d21-f1ca-4194-b212-0d54a9be6c58"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. train_seq 배열과 train_oh 배열의 nbytes 속성으로 크기 확인\n",
        "# [참고] 원-핫 인코딩의 단점: 입력 데이터가 엄청 커짐\n",
        "print(train_seq.nbytes, train_oh.nbytes)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vFOv-v7vmeW2",
        "outputId": "f9ccb1d6-c985-4d67-dd72-dd176849c87c"
      },
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "8000000 2400000000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 단어 임베딩을 사용하기"
      ],
      "metadata": {
        "id": "dyBlCLWlnSOj"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. 두 번째 순환 신경망 만들기\n",
        "model2 = keras.Sequential()\n",
        "model2.add(keras.layers.Embedding(300, 16, input_length=100))\n",
        "model2.add(keras.layers.SimpleRNN(8))\n",
        "model2.add(keras.layers.Dense(1, activation='sigmoid'))"
      ],
      "metadata": {
        "id": "kb-2cUa5nVH9"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 모델 구조 출력\n",
        "model2.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "G4CSt3i1oLog",
        "outputId": "69685599-28db-45a4-a1df-5886f14a61e2"
      },
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_1 (Embedding)     (None, 100, 16)           4800      \n",
            "                                                                 \n",
            " simple_rnn_3 (SimpleRNN)    (None, 8)                 200       \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 1)                 9         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 5009 (19.57 KB)\n",
            "Trainable params: 5009 (19.57 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. 모델 훈련하기\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model2.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-embedding-model.h5',\n",
        "                                                save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model2.fit(train_seq, train_target, epochs=100, batch_size=64,\n",
        "                     validation_data=(val_seq, val_target),\n",
        "                     callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ypwFOPg5qnii",
        "outputId": "6d3d32db-20b0-4423-96d5-fc99ec699bae"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 10s 27ms/step - loss: 0.6960 - accuracy: 0.5078 - val_loss: 0.6942 - val_accuracy: 0.5088\n",
            "Epoch 2/100\n",
            "  7/313 [..............................] - ETA: 5s - loss: 0.6936 - accuracy: 0.5335"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 13s 41ms/step - loss: 0.6920 - accuracy: 0.5200 - val_loss: 0.6930 - val_accuracy: 0.5066\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 15s 49ms/step - loss: 0.6893 - accuracy: 0.5389 - val_loss: 0.6916 - val_accuracy: 0.5162\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 15s 47ms/step - loss: 0.6865 - accuracy: 0.5566 - val_loss: 0.6899 - val_accuracy: 0.5276\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 7s 21ms/step - loss: 0.6825 - accuracy: 0.5724 - val_loss: 0.6858 - val_accuracy: 0.5560\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.6752 - accuracy: 0.5993 - val_loss: 0.6759 - val_accuracy: 0.5958\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 7s 24ms/step - loss: 0.6623 - accuracy: 0.6340 - val_loss: 0.6630 - val_accuracy: 0.6352\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 7s 22ms/step - loss: 0.6475 - accuracy: 0.6635 - val_loss: 0.6473 - val_accuracy: 0.6654\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.6333 - accuracy: 0.6862 - val_loss: 0.6333 - val_accuracy: 0.6820\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 7s 21ms/step - loss: 0.6205 - accuracy: 0.7034 - val_loss: 0.6214 - val_accuracy: 0.6970\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.6089 - accuracy: 0.7129 - val_loss: 0.6131 - val_accuracy: 0.7048\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 7s 21ms/step - loss: 0.5980 - accuracy: 0.7229 - val_loss: 0.6022 - val_accuracy: 0.7138\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5885 - accuracy: 0.7275 - val_loss: 0.5968 - val_accuracy: 0.7142\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 7s 24ms/step - loss: 0.5817 - accuracy: 0.7313 - val_loss: 0.5921 - val_accuracy: 0.7168\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 7s 22ms/step - loss: 0.5755 - accuracy: 0.7332 - val_loss: 0.5850 - val_accuracy: 0.7182\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5685 - accuracy: 0.7381 - val_loss: 0.5811 - val_accuracy: 0.7138\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 7s 23ms/step - loss: 0.5637 - accuracy: 0.7369 - val_loss: 0.5765 - val_accuracy: 0.7240\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 9s 27ms/step - loss: 0.5610 - accuracy: 0.7375 - val_loss: 0.5756 - val_accuracy: 0.7226\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 9s 29ms/step - loss: 0.5576 - accuracy: 0.7370 - val_loss: 0.5732 - val_accuracy: 0.7180\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 7s 22ms/step - loss: 0.5532 - accuracy: 0.7418 - val_loss: 0.5731 - val_accuracy: 0.7156\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5519 - accuracy: 0.7397 - val_loss: 0.5717 - val_accuracy: 0.7164\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 7s 21ms/step - loss: 0.5494 - accuracy: 0.7409 - val_loss: 0.5729 - val_accuracy: 0.7110\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 8s 25ms/step - loss: 0.5472 - accuracy: 0.7423 - val_loss: 0.5755 - val_accuracy: 0.7048\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 7s 21ms/step - loss: 0.5446 - accuracy: 0.7422 - val_loss: 0.5787 - val_accuracy: 0.7202\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. 훈련 손실과 검증 손실 그래프 출력\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "24bpBsRjrdCR",
        "outputId": "46641209-99de-47dc-8ca6-85e0bb14d40f"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Ch09-3 | LSTM과 GRU 셀"
      ],
      "metadata": {
        "id": "fvQoDRGBi3ZS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## LSTM 신경망 훈련하기"
      ],
      "metadata": {
        "id": "5hL5Q1wHj4qZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. IMDB 리뷰 데이터 로드 및 훈련 세트와 검증 세트로 나누기\n",
        "from tensorflow.keras.datasets import imdb\n",
        "from sklearn.model_selection import train_test_split\n",
        "(train_input, train_target), (test_input, test_target) = imdb.load_data(num_words=500)\n",
        "train_input, val_input, train_target, val_target = train_test_split(\n",
        "    train_input, train_target, test_size=0.2, random_state=42)"
      ],
      "metadata": {
        "id": "kT8KmV-iKver"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 샘플 길이: 100으로 조정 및 패딩 추가\n",
        "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
        "train_seq = pad_sequences(train_input, maxlen=100)\n",
        "val_seq = pad_sequences(val_input, maxlen=100)"
      ],
      "metadata": {
        "id": "MfvdhupnLOV-"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. LSTM 셀을 사용한 순환층 만들기\n",
        "from tensorflow import keras\n",
        "model = keras.Sequential()\n",
        "model.add(keras.layers.Embedding(500, 16, input_length=100))\n",
        "model.add(keras.layers.LSTM(8))\n",
        "model.add(keras.layers.Dense(1, activation='sigmoid'))"
      ],
      "metadata": {
        "id": "mT9-f37rLh5S"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. 모델 구조 확인\n",
        "model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nsCxnjK7Lzov",
        "outputId": "d34dc146-61b9-43bf-d1c2-7b1778e4426b"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_4\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_2 (Embedding)     (None, 100, 16)           8000      \n",
            "                                                                 \n",
            " lstm (LSTM)                 (None, 8)                 800       \n",
            "                                                                 \n",
            " dense_4 (Dense)             (None, 1)                 9         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 8809 (34.41 KB)\n",
            "Trainable params: 8809 (34.41 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 5. 모델 컴파일 & 훈련\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-lstm-model.h5', save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model.fit(train_seq, train_target, epochs=100, batch_size=64,\n",
        "                    validation_data=(val_seq, val_target),\n",
        "                    callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KlJ8LXUnL4F_",
        "outputId": "01faf6c2-dd64-4175-c74a-2b8880d70d2e"
      },
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 18s 49ms/step - loss: 0.6920 - accuracy: 0.5454 - val_loss: 0.6906 - val_accuracy: 0.5886\n",
            "Epoch 2/100\n",
            "  2/313 [..............................] - ETA: 20s - loss: 0.6915 - accuracy: 0.5547"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 20s 64ms/step - loss: 0.6892 - accuracy: 0.6010 - val_loss: 0.6864 - val_accuracy: 0.6326\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 19s 59ms/step - loss: 0.6786 - accuracy: 0.6557 - val_loss: 0.6635 - val_accuracy: 0.6796\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.6252 - accuracy: 0.6956 - val_loss: 0.6003 - val_accuracy: 0.6994\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 22s 69ms/step - loss: 0.5871 - accuracy: 0.7134 - val_loss: 0.5769 - val_accuracy: 0.7212\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 10s 33ms/step - loss: 0.5635 - accuracy: 0.7344 - val_loss: 0.5572 - val_accuracy: 0.7308\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.5429 - accuracy: 0.7501 - val_loss: 0.5379 - val_accuracy: 0.7458\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.5241 - accuracy: 0.7631 - val_loss: 0.5194 - val_accuracy: 0.7634\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.5069 - accuracy: 0.7735 - val_loss: 0.5046 - val_accuracy: 0.7680\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4913 - accuracy: 0.7799 - val_loss: 0.4910 - val_accuracy: 0.7790\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4775 - accuracy: 0.7884 - val_loss: 0.4829 - val_accuracy: 0.7776\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4657 - accuracy: 0.7947 - val_loss: 0.4702 - val_accuracy: 0.7868\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4565 - accuracy: 0.7970 - val_loss: 0.4623 - val_accuracy: 0.7922\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.4487 - accuracy: 0.8004 - val_loss: 0.4574 - val_accuracy: 0.7914\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4419 - accuracy: 0.8047 - val_loss: 0.4516 - val_accuracy: 0.7954\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 11s 35ms/step - loss: 0.4370 - accuracy: 0.8065 - val_loss: 0.4484 - val_accuracy: 0.7966\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 10s 33ms/step - loss: 0.4328 - accuracy: 0.8063 - val_loss: 0.4457 - val_accuracy: 0.7962\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 11s 37ms/step - loss: 0.4292 - accuracy: 0.8090 - val_loss: 0.4439 - val_accuracy: 0.7978\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.4260 - accuracy: 0.8080 - val_loss: 0.4436 - val_accuracy: 0.7976\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.4238 - accuracy: 0.8100 - val_loss: 0.4431 - val_accuracy: 0.7948\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4217 - accuracy: 0.8109 - val_loss: 0.4390 - val_accuracy: 0.7966\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 14s 43ms/step - loss: 0.4196 - accuracy: 0.8113 - val_loss: 0.4392 - val_accuracy: 0.7970\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.4180 - accuracy: 0.8110 - val_loss: 0.4393 - val_accuracy: 0.8012\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 11s 37ms/step - loss: 0.4164 - accuracy: 0.8115 - val_loss: 0.4387 - val_accuracy: 0.8020\n",
            "Epoch 25/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.4151 - accuracy: 0.8117 - val_loss: 0.4364 - val_accuracy: 0.7966\n",
            "Epoch 26/100\n",
            "313/313 [==============================] - 14s 45ms/step - loss: 0.4140 - accuracy: 0.8109 - val_loss: 0.4380 - val_accuracy: 0.7938\n",
            "Epoch 27/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.4127 - accuracy: 0.8114 - val_loss: 0.4344 - val_accuracy: 0.8002\n",
            "Epoch 28/100\n",
            "313/313 [==============================] - 12s 37ms/step - loss: 0.4125 - accuracy: 0.8122 - val_loss: 0.4337 - val_accuracy: 0.7996\n",
            "Epoch 29/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4117 - accuracy: 0.8121 - val_loss: 0.4338 - val_accuracy: 0.7978\n",
            "Epoch 30/100\n",
            "313/313 [==============================] - 11s 34ms/step - loss: 0.4108 - accuracy: 0.8128 - val_loss: 0.4330 - val_accuracy: 0.7986\n",
            "Epoch 31/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4103 - accuracy: 0.8128 - val_loss: 0.4326 - val_accuracy: 0.7998\n",
            "Epoch 32/100\n",
            "313/313 [==============================] - 11s 36ms/step - loss: 0.4094 - accuracy: 0.8130 - val_loss: 0.4326 - val_accuracy: 0.8006\n",
            "Epoch 33/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.4089 - accuracy: 0.8131 - val_loss: 0.4336 - val_accuracy: 0.8000\n",
            "Epoch 34/100\n",
            "313/313 [==============================] - 12s 38ms/step - loss: 0.4085 - accuracy: 0.8138 - val_loss: 0.4333 - val_accuracy: 0.8012\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 6. 훈련 손실과 검증 손실 그래프 출력\n",
        "# [결과 해석] 기본 순환층 보다 LSTM이 과대적합을 억제하면서 훈련을 잘 수행\n",
        "import matplotlib.pyplot as plt\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "cfb_45lxMwHD",
        "outputId": "b0072fb0-c508-4e12-c0d9-395048d5345d"
      },
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 순환층에 드롭아웃 적용하기"
      ],
      "metadata": {
        "id": "bA7sYUDSO4XD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. 드롭 아웃 적용 모델 생성\n",
        "# [참고] 순환층은 자체적으로 드롭아웃 기능 제공\n",
        "# [참고] dropout 매개변수: 셀의 입력에 드롭 아웃 / recurrent_dropout: 순환되는 은닉 상태에 드롭 아웃\n",
        "model2 = keras.Sequential()\n",
        "model2.add(keras.layers.Embedding(500, 16, input_length=100))\n",
        "model2.add(keras.layers.LSTM(8, dropout=0.3))\n",
        "model2.add(keras.layers.Dense(1, activation='sigmoid'))"
      ],
      "metadata": {
        "id": "kki9sft1Pz-o"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 모델 컴파일 & 훈련\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model2.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-dropout-model.h5', save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model2.fit(train_seq, train_target, epochs=100, batch_size=64,\n",
        "                     validation_data=(val_seq, val_target),\n",
        "                     callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8aHLQKGEQu27",
        "outputId": "2ff1eda6-4f30-4392-d4d4-701d409e665e"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 17s 42ms/step - loss: 0.6926 - accuracy: 0.5415 - val_loss: 0.6921 - val_accuracy: 0.5464\n",
            "Epoch 2/100\n",
            "  3/313 [..............................] - ETA: 10s - loss: 0.6922 - accuracy: 0.5521"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 13s 41ms/step - loss: 0.6912 - accuracy: 0.5630 - val_loss: 0.6903 - val_accuracy: 0.6114\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.6882 - accuracy: 0.6198 - val_loss: 0.6856 - val_accuracy: 0.6568\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.6775 - accuracy: 0.6547 - val_loss: 0.6624 - val_accuracy: 0.6590\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 16s 50ms/step - loss: 0.6300 - accuracy: 0.6958 - val_loss: 0.6131 - val_accuracy: 0.7162\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 24s 78ms/step - loss: 0.6022 - accuracy: 0.7178 - val_loss: 0.5937 - val_accuracy: 0.7256\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 14s 44ms/step - loss: 0.5852 - accuracy: 0.7298 - val_loss: 0.5770 - val_accuracy: 0.7404\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.5685 - accuracy: 0.7402 - val_loss: 0.5621 - val_accuracy: 0.7482\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 20s 65ms/step - loss: 0.5533 - accuracy: 0.7473 - val_loss: 0.5468 - val_accuracy: 0.7558\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 14s 44ms/step - loss: 0.5387 - accuracy: 0.7580 - val_loss: 0.5324 - val_accuracy: 0.7580\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 19s 60ms/step - loss: 0.5237 - accuracy: 0.7637 - val_loss: 0.5180 - val_accuracy: 0.7668\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.5100 - accuracy: 0.7693 - val_loss: 0.5045 - val_accuracy: 0.7720\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 14s 45ms/step - loss: 0.4955 - accuracy: 0.7773 - val_loss: 0.4909 - val_accuracy: 0.7806\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 13s 43ms/step - loss: 0.4824 - accuracy: 0.7826 - val_loss: 0.4789 - val_accuracy: 0.7846\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4707 - accuracy: 0.7887 - val_loss: 0.4696 - val_accuracy: 0.7858\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4627 - accuracy: 0.7926 - val_loss: 0.4614 - val_accuracy: 0.7920\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4543 - accuracy: 0.7948 - val_loss: 0.4551 - val_accuracy: 0.7928\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4486 - accuracy: 0.7981 - val_loss: 0.4508 - val_accuracy: 0.7944\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4439 - accuracy: 0.7986 - val_loss: 0.4473 - val_accuracy: 0.7950\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4406 - accuracy: 0.8000 - val_loss: 0.4444 - val_accuracy: 0.7968\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4358 - accuracy: 0.8018 - val_loss: 0.4420 - val_accuracy: 0.7962\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4345 - accuracy: 0.8043 - val_loss: 0.4408 - val_accuracy: 0.8002\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4317 - accuracy: 0.8048 - val_loss: 0.4398 - val_accuracy: 0.7938\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4299 - accuracy: 0.8050 - val_loss: 0.4383 - val_accuracy: 0.8000\n",
            "Epoch 25/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4279 - accuracy: 0.8051 - val_loss: 0.4371 - val_accuracy: 0.8024\n",
            "Epoch 26/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4263 - accuracy: 0.8080 - val_loss: 0.4368 - val_accuracy: 0.8036\n",
            "Epoch 27/100\n",
            "313/313 [==============================] - 14s 44ms/step - loss: 0.4254 - accuracy: 0.8040 - val_loss: 0.4351 - val_accuracy: 0.7976\n",
            "Epoch 28/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.4233 - accuracy: 0.8059 - val_loss: 0.4358 - val_accuracy: 0.7966\n",
            "Epoch 29/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4240 - accuracy: 0.8044 - val_loss: 0.4342 - val_accuracy: 0.8008\n",
            "Epoch 30/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4216 - accuracy: 0.8089 - val_loss: 0.4346 - val_accuracy: 0.7968\n",
            "Epoch 31/100\n",
            "313/313 [==============================] - 13s 43ms/step - loss: 0.4224 - accuracy: 0.8069 - val_loss: 0.4329 - val_accuracy: 0.7968\n",
            "Epoch 32/100\n",
            "313/313 [==============================] - 17s 56ms/step - loss: 0.4202 - accuracy: 0.8084 - val_loss: 0.4327 - val_accuracy: 0.7978\n",
            "Epoch 33/100\n",
            "313/313 [==============================] - 15s 49ms/step - loss: 0.4206 - accuracy: 0.8073 - val_loss: 0.4327 - val_accuracy: 0.7976\n",
            "Epoch 34/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4194 - accuracy: 0.8082 - val_loss: 0.4321 - val_accuracy: 0.8008\n",
            "Epoch 35/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4210 - accuracy: 0.8061 - val_loss: 0.4318 - val_accuracy: 0.7994\n",
            "Epoch 36/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.4189 - accuracy: 0.8087 - val_loss: 0.4326 - val_accuracy: 0.7968\n",
            "Epoch 37/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4180 - accuracy: 0.8085 - val_loss: 0.4322 - val_accuracy: 0.8020\n",
            "Epoch 38/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.4170 - accuracy: 0.8094 - val_loss: 0.4320 - val_accuracy: 0.8008\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. 훈련 손실과 검증 손실 그래프 출력\n",
        "# [결과 해석] 드롭아웃 효과 발휘\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "ovz3y_Z3Rjd5",
        "outputId": "e8b586a4-5596-47ab-98d9-ac642a3fa67c"
      },
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 2개의 층을 연결하기"
      ],
      "metadata": {
        "id": "rNNb-ossUK2q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. 2개의 층 쌓기\n",
        "# [주의] 순환층에서 모든 타임스텝의 은닉 상태를 출력해야 함\n",
        "#        > 마지막을 제외한 다른 모든 순환층에서 return_sequence 매개변수 = True 설정\n",
        "model3 = keras.Sequential()\n",
        "model3.add(keras.layers.Embedding(500, 16, input_length=100))\n",
        "model3.add(keras.layers.LSTM(8, dropout=0.3, return_sequences=True))\n",
        "model3.add(keras.layers.LSTM(8, dropout=0.3))\n",
        "model3.add(keras.layers.Dense(1, activation='sigmoid'))"
      ],
      "metadata": {
        "id": "7mSgNJaSUPXa"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 모델 구조 확인\n",
        "model3.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xDD2v3y9VOwt",
        "outputId": "226a8501-8402-448a-acb7-9850e657e5e6"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_6\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_4 (Embedding)     (None, 100, 16)           8000      \n",
            "                                                                 \n",
            " lstm_2 (LSTM)               (None, 100, 8)            800       \n",
            "                                                                 \n",
            " lstm_3 (LSTM)               (None, 8)                 544       \n",
            "                                                                 \n",
            " dense_6 (Dense)             (None, 1)                 9         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 9353 (36.54 KB)\n",
            "Trainable params: 9353 (36.54 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. 모델 컴파일 & 훈련\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model3.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-2rnn-model.h5', save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model3.fit(train_seq, train_target, epochs=100, batch_size=64,\n",
        "                     validation_data=(val_seq, val_target),\n",
        "                     callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AveKrn8KVXv-",
        "outputId": "85c30a6f-e2d5-4db5-e232-ae026bdbacbd"
      },
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 31s 85ms/step - loss: 0.6930 - accuracy: 0.5166 - val_loss: 0.6927 - val_accuracy: 0.5434\n",
            "Epoch 2/100\n",
            "  2/313 [..............................] - ETA: 19s - loss: 0.6925 - accuracy: 0.5234"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 23s 73ms/step - loss: 0.6922 - accuracy: 0.5562 - val_loss: 0.6914 - val_accuracy: 0.6034\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 30s 97ms/step - loss: 0.6890 - accuracy: 0.6053 - val_loss: 0.6844 - val_accuracy: 0.6340\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 31s 99ms/step - loss: 0.6703 - accuracy: 0.6470 - val_loss: 0.6434 - val_accuracy: 0.6792\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 24s 78ms/step - loss: 0.6057 - accuracy: 0.6934 - val_loss: 0.5746 - val_accuracy: 0.7138\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.5661 - accuracy: 0.7172 - val_loss: 0.5538 - val_accuracy: 0.7296\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 23s 74ms/step - loss: 0.5435 - accuracy: 0.7367 - val_loss: 0.5242 - val_accuracy: 0.7522\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 22s 72ms/step - loss: 0.5243 - accuracy: 0.7461 - val_loss: 0.5095 - val_accuracy: 0.7584\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 23s 74ms/step - loss: 0.5098 - accuracy: 0.7584 - val_loss: 0.4961 - val_accuracy: 0.7682\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4978 - accuracy: 0.7655 - val_loss: 0.4865 - val_accuracy: 0.7774\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 23s 72ms/step - loss: 0.4912 - accuracy: 0.7702 - val_loss: 0.4801 - val_accuracy: 0.7806\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.4828 - accuracy: 0.7772 - val_loss: 0.4754 - val_accuracy: 0.7824\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.4771 - accuracy: 0.7775 - val_loss: 0.4690 - val_accuracy: 0.7844\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 24s 76ms/step - loss: 0.4690 - accuracy: 0.7818 - val_loss: 0.4657 - val_accuracy: 0.7830\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4661 - accuracy: 0.7847 - val_loss: 0.4629 - val_accuracy: 0.7856\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 23s 72ms/step - loss: 0.4604 - accuracy: 0.7875 - val_loss: 0.4596 - val_accuracy: 0.7864\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 23s 72ms/step - loss: 0.4565 - accuracy: 0.7904 - val_loss: 0.4621 - val_accuracy: 0.7858\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.4552 - accuracy: 0.7921 - val_loss: 0.4535 - val_accuracy: 0.7918\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4527 - accuracy: 0.7911 - val_loss: 0.4526 - val_accuracy: 0.7906\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4473 - accuracy: 0.7958 - val_loss: 0.4509 - val_accuracy: 0.7882\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.4455 - accuracy: 0.7944 - val_loss: 0.4497 - val_accuracy: 0.7898\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 23s 72ms/step - loss: 0.4457 - accuracy: 0.7936 - val_loss: 0.4460 - val_accuracy: 0.7922\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4411 - accuracy: 0.7996 - val_loss: 0.4519 - val_accuracy: 0.7904\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4412 - accuracy: 0.8001 - val_loss: 0.4512 - val_accuracy: 0.7914\n",
            "Epoch 25/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4382 - accuracy: 0.7987 - val_loss: 0.4432 - val_accuracy: 0.7942\n",
            "Epoch 26/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4360 - accuracy: 0.8005 - val_loss: 0.4482 - val_accuracy: 0.7942\n",
            "Epoch 27/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4353 - accuracy: 0.8000 - val_loss: 0.4442 - val_accuracy: 0.7946\n",
            "Epoch 28/100\n",
            "313/313 [==============================] - 22s 70ms/step - loss: 0.4344 - accuracy: 0.8022 - val_loss: 0.4417 - val_accuracy: 0.7932\n",
            "Epoch 29/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4332 - accuracy: 0.8020 - val_loss: 0.4460 - val_accuracy: 0.7954\n",
            "Epoch 30/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4295 - accuracy: 0.8044 - val_loss: 0.4399 - val_accuracy: 0.7956\n",
            "Epoch 31/100\n",
            "313/313 [==============================] - 21s 68ms/step - loss: 0.4321 - accuracy: 0.8055 - val_loss: 0.4394 - val_accuracy: 0.7966\n",
            "Epoch 32/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4310 - accuracy: 0.8045 - val_loss: 0.4413 - val_accuracy: 0.7884\n",
            "Epoch 33/100\n",
            "313/313 [==============================] - 23s 73ms/step - loss: 0.4281 - accuracy: 0.8033 - val_loss: 0.4451 - val_accuracy: 0.7980\n",
            "Epoch 34/100\n",
            "313/313 [==============================] - 22s 72ms/step - loss: 0.4297 - accuracy: 0.8051 - val_loss: 0.4384 - val_accuracy: 0.7956\n",
            "Epoch 35/100\n",
            "313/313 [==============================] - 22s 69ms/step - loss: 0.4272 - accuracy: 0.8033 - val_loss: 0.4388 - val_accuracy: 0.7972\n",
            "Epoch 36/100\n",
            "313/313 [==============================] - 23s 72ms/step - loss: 0.4284 - accuracy: 0.8035 - val_loss: 0.4381 - val_accuracy: 0.7984\n",
            "Epoch 37/100\n",
            "313/313 [==============================] - 23s 75ms/step - loss: 0.4254 - accuracy: 0.8073 - val_loss: 0.4375 - val_accuracy: 0.7964\n",
            "Epoch 38/100\n",
            "313/313 [==============================] - 23s 74ms/step - loss: 0.4271 - accuracy: 0.8046 - val_loss: 0.4371 - val_accuracy: 0.7980\n",
            "Epoch 39/100\n",
            "313/313 [==============================] - 23s 74ms/step - loss: 0.4249 - accuracy: 0.8031 - val_loss: 0.4385 - val_accuracy: 0.7976\n",
            "Epoch 40/100\n",
            "313/313 [==============================] - 24s 75ms/step - loss: 0.4245 - accuracy: 0.8057 - val_loss: 0.4398 - val_accuracy: 0.7976\n",
            "Epoch 41/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.4258 - accuracy: 0.8051 - val_loss: 0.4383 - val_accuracy: 0.8002\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. 훈련 손실과 검증 손실 그래픽 출력\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "uZni-_mVWVr0",
        "outputId": "a17a4a81-24d8-4f1a-edb5-c39bfe353fab"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## GRU 신경망 훈련하기"
      ],
      "metadata": {
        "id": "UsDyRc6eaUwD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. GRU로 모델 생성\n",
        "model4 = keras.Sequential()\n",
        "model4.add(keras.layers.Embedding(500, 16, input_length=100))\n",
        "model4.add(keras.layers.GRU(8))\n",
        "model4.add(keras.layers.Dense(1, activation='sigmoid'))"
      ],
      "metadata": {
        "id": "xHzhQHcDau_K"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# 2. 모델 구조 확인\n",
        "model4.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_Pfdbj9RbDDO",
        "outputId": "2aa1e42d-5e08-4902-cd0e-3b51043c6400"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_7\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_5 (Embedding)     (None, 100, 16)           8000      \n",
            "                                                                 \n",
            " gru (GRU)                   (None, 8)                 624       \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 1)                 9         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 8633 (33.72 KB)\n",
            "Trainable params: 8633 (33.72 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 3. 모델 컴파일 & 훈련\n",
        "rmsprop = keras.optimizers.RMSprop(learning_rate=1e-4)\n",
        "model4.compile(optimizer=rmsprop, loss='binary_crossentropy', metrics=['accuracy'])\n",
        "checkpoint_cb = keras.callbacks.ModelCheckpoint('best-gru-model.h5', save_best_only=True)\n",
        "early_stopping_cb = keras.callbacks.EarlyStopping(patience=3, restore_best_weights=True)\n",
        "history = model4.fit(train_seq, train_target, epochs=100, batch_size=64,\n",
        "                     validation_data=(val_seq, val_target),\n",
        "                     callbacks=[checkpoint_cb, early_stopping_cb])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qulsBydEbNek",
        "outputId": "86da9fb5-34c4-43d3-e4f2-858595c4d42d"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "313/313 [==============================] - 23s 61ms/step - loss: 0.6922 - accuracy: 0.5240 - val_loss: 0.6917 - val_accuracy: 0.5344\n",
            "Epoch 2/100\n",
            "  3/313 [..............................] - ETA: 10s - loss: 0.6916 - accuracy: 0.5677"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/engine/training.py:3103: UserWarning: You are saving your model as an HDF5 file via `model.save()`. This file format is considered legacy. We recommend using instead the native Keras format, e.g. `model.save('my_model.keras')`.\n",
            "  saving_api.save_model(\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 21s 66ms/step - loss: 0.6898 - accuracy: 0.5706 - val_loss: 0.6889 - val_accuracy: 0.5818\n",
            "Epoch 3/100\n",
            "313/313 [==============================] - 22s 71ms/step - loss: 0.6862 - accuracy: 0.5982 - val_loss: 0.6850 - val_accuracy: 0.5856\n",
            "Epoch 4/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.6805 - accuracy: 0.6146 - val_loss: 0.6786 - val_accuracy: 0.5994\n",
            "Epoch 5/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.6715 - accuracy: 0.6284 - val_loss: 0.6679 - val_accuracy: 0.6192\n",
            "Epoch 6/100\n",
            "313/313 [==============================] - 18s 58ms/step - loss: 0.6567 - accuracy: 0.6431 - val_loss: 0.6503 - val_accuracy: 0.6434\n",
            "Epoch 7/100\n",
            "313/313 [==============================] - 19s 60ms/step - loss: 0.6317 - accuracy: 0.6668 - val_loss: 0.6185 - val_accuracy: 0.6760\n",
            "Epoch 8/100\n",
            "313/313 [==============================] - 19s 61ms/step - loss: 0.5820 - accuracy: 0.7017 - val_loss: 0.5548 - val_accuracy: 0.7220\n",
            "Epoch 9/100\n",
            "313/313 [==============================] - 13s 43ms/step - loss: 0.5222 - accuracy: 0.7429 - val_loss: 0.5150 - val_accuracy: 0.7484\n",
            "Epoch 10/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.4965 - accuracy: 0.7619 - val_loss: 0.4964 - val_accuracy: 0.7610\n",
            "Epoch 11/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4789 - accuracy: 0.7718 - val_loss: 0.4796 - val_accuracy: 0.7742\n",
            "Epoch 12/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4651 - accuracy: 0.7825 - val_loss: 0.4695 - val_accuracy: 0.7802\n",
            "Epoch 13/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4550 - accuracy: 0.7915 - val_loss: 0.4643 - val_accuracy: 0.7862\n",
            "Epoch 14/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4467 - accuracy: 0.7965 - val_loss: 0.4602 - val_accuracy: 0.7932\n",
            "Epoch 15/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.4409 - accuracy: 0.7996 - val_loss: 0.4536 - val_accuracy: 0.7926\n",
            "Epoch 16/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4358 - accuracy: 0.8020 - val_loss: 0.4493 - val_accuracy: 0.7946\n",
            "Epoch 17/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4317 - accuracy: 0.8045 - val_loss: 0.4472 - val_accuracy: 0.7934\n",
            "Epoch 18/100\n",
            "313/313 [==============================] - 13s 42ms/step - loss: 0.4289 - accuracy: 0.8080 - val_loss: 0.4465 - val_accuracy: 0.7906\n",
            "Epoch 19/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4264 - accuracy: 0.8083 - val_loss: 0.4471 - val_accuracy: 0.7968\n",
            "Epoch 20/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4243 - accuracy: 0.8106 - val_loss: 0.4434 - val_accuracy: 0.7920\n",
            "Epoch 21/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4224 - accuracy: 0.8109 - val_loss: 0.4436 - val_accuracy: 0.7920\n",
            "Epoch 22/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.4211 - accuracy: 0.8113 - val_loss: 0.4420 - val_accuracy: 0.7934\n",
            "Epoch 23/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4201 - accuracy: 0.8120 - val_loss: 0.4414 - val_accuracy: 0.7948\n",
            "Epoch 24/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4184 - accuracy: 0.8132 - val_loss: 0.4437 - val_accuracy: 0.7906\n",
            "Epoch 25/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4182 - accuracy: 0.8140 - val_loss: 0.4408 - val_accuracy: 0.7964\n",
            "Epoch 26/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4177 - accuracy: 0.8137 - val_loss: 0.4404 - val_accuracy: 0.7952\n",
            "Epoch 27/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4168 - accuracy: 0.8152 - val_loss: 0.4449 - val_accuracy: 0.7940\n",
            "Epoch 28/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4159 - accuracy: 0.8137 - val_loss: 0.4402 - val_accuracy: 0.7962\n",
            "Epoch 29/100\n",
            "313/313 [==============================] - 13s 41ms/step - loss: 0.4155 - accuracy: 0.8145 - val_loss: 0.4409 - val_accuracy: 0.7950\n",
            "Epoch 30/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4148 - accuracy: 0.8148 - val_loss: 0.4405 - val_accuracy: 0.7960\n",
            "Epoch 31/100\n",
            "313/313 [==============================] - 12s 39ms/step - loss: 0.4143 - accuracy: 0.8164 - val_loss: 0.4390 - val_accuracy: 0.7974\n",
            "Epoch 32/100\n",
            "313/313 [==============================] - 12s 40ms/step - loss: 0.4143 - accuracy: 0.8150 - val_loss: 0.4451 - val_accuracy: 0.7878\n",
            "Epoch 33/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4140 - accuracy: 0.8156 - val_loss: 0.4391 - val_accuracy: 0.7966\n",
            "Epoch 34/100\n",
            "313/313 [==============================] - 13s 40ms/step - loss: 0.4134 - accuracy: 0.8156 - val_loss: 0.4451 - val_accuracy: 0.7888\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 4. 훈련 손실과 검증 손실 그래프 출력\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.xlabel('epoch')\n",
        "plt.ylabel('loss')\n",
        "plt.legend(['train', 'val'])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 449
        },
        "id": "nn64m0Smc4Ix",
        "outputId": "c31807f2-b49c-4ec0-95ec-1e2ff40eaa53"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## LSTM과 GRU 셀로 훈련 과정"
      ],
      "metadata": {
        "id": "yxv3nZr-dN0C"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# 1. 테스트 세트의 성능 계산\n",
        "test_seq = pad_sequences(test_input, maxlen=100)\n",
        "rnn_model = keras.models.load_model('best-2rnn-model.h5')\n",
        "rnn_model.evaluate(test_seq, test_target)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x_LLQXDpdQks",
        "outputId": "3bb80a96-71ec-4f9a-b7f5-d95b07049d5d"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 13s 15ms/step - loss: 0.4306 - accuracy: 0.7988\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.43060389161109924, 0.7987599968910217]"
            ]
          },
          "metadata": {},
          "execution_count": 52
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "vGqeu-e4gaBy"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}