material/deep_learning/denoising_autoencoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "denoising_autoencoder.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "7u-zHeaOyjn2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "outputId": "6b864f77-0928-403e-e8aa-7a71f5196f8a"
      },
      "source": [
        "%tensorflow_version 1.x"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 1.x selected.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jLKCaOxfGQg3",
        "colab_type": "text"
      },
      "source": [
        "copy from https://blog.naver.com/ckdgus1433/221443838135"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3gUCJI7zK6V0",
        "colab_type": "text"
      },
      "source": [
        "# 모듈 임포팅"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y3ayalJEK4Me",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "from tensorflow.keras.datasets import mnist\n",
        "from tensorflow.keras.layers import Dense\n",
        "from tensorflow.keras.models import Sequential"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-8b8fVWQK9JD",
        "colab_type": "text"
      },
      "source": [
        "# 데이터 로딩"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D0QMMukcK56L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# MNIST 로딩 (라벨은 필요없기 때문에 버림)\n",
        "(x_train, _), (x_test, _) = mnist.load_data()"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OOAbPf_ZK_6b",
        "colab_type": "text"
      },
      "source": [
        "# Normalization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WWimdhYPLBTj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# 데이터 정규화\n",
        "x_train = x_train.astype('float32') / 255.\n",
        "x_test = x_test.astype('float32') / 255."
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "COdPgRsoLBjD",
        "colab_type": "text"
      },
      "source": [
        "# Flattening"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wjDc3mmbLB-D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "x_train = np.reshape(x_train, (len(x_train), 784))\n",
        "x_test = np.reshape(x_test, (len(x_test), 784))"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zNfjHZQSLN-s",
        "colab_type": "text"
      },
      "source": [
        "# 노이즈 추가"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dTSIWBn7NeuA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "noise = np.random.normal(loc=0.0, scale=1.0, size=10000) "
      ],
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zZplNkNxN5dl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "outputId": "04a3e20e-ea29-4c13-8681-4015f75d7b1b"
      },
      "source": [
        "plt.hist(noise, bins=1000, range=(-2., 2.))\n",
        "plt.show()"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tvOfQ4hMLLo7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# 원본데이터에 Noise 추가\n",
        "noise_factor = 0.5\n",
        "x_train_noisy = x_train + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_train.shape) \n",
        "x_test_noisy = x_test + noise_factor * np.random.normal(loc=0.0, scale=1.0, size=x_test.shape) \n",
        "\n",
        "# 최대가 1, 최소가 0이 되도록 \n",
        "x_train_noisy = np.clip(x_train_noisy, 0., 1.)\n",
        "x_test_noisy = np.clip(x_test_noisy, 0., 1.)"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6na4p0mnLj8j",
        "colab_type": "text"
      },
      "source": [
        "# 노이즈 추가된 영상 보기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y--Ep7aULL17",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 111
        },
        "outputId": "7081071f-4258-48d2-e908-6ff7e9f22330"
      },
      "source": [
        "# Noise가 추가된 데이터 확인\n",
        "n = 10\n",
        "plt.figure(figsize=(20, 2))\n",
        "for i in range(1,n):\n",
        "    ax = plt.subplot(1, n, i)\n",
        "    plt.imshow(x_test_noisy[i].reshape(28, 28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x144 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i01JyXUqL7yF",
        "colab_type": "text"
      },
      "source": [
        "# 모델 정의와 학습 실행"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bvY_c15CLL48",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "41c905e7-ad90-419e-9ec7-a524b252f5d3"
      },
      "source": [
        "# 모형 구성\n",
        "model = Sequential()\n",
        "model.add(Dense(128, activation='relu', input_dim=784))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dense(32, activation='relu'))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dense(128, activation='relu'))\n",
        "model.add(Dense(784, activation='sigmoid'))\n",
        "model.compile(optimizer='adam', loss='binary_crossentropy')\n",
        "\n",
        "\n",
        "# 모형 학습\n",
        "model.fit(x_train_noisy, x_train, \n",
        "          epochs=100,\n",
        "          batch_size=256,\n",
        "          shuffle=True,\n",
        "          validation_data=(x_test_noisy, x_test))\n"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 1s 22us/sample - loss: 0.2679 - val_loss: 0.2134\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1875 - val_loss: 0.1733\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1680 - val_loss: 0.1618\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1579 - val_loss: 0.1533\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1514 - val_loss: 0.1474\n",
            "Epoch 6/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1463 - val_loss: 0.1434\n",
            "Epoch 7/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1426 - val_loss: 0.1404\n",
            "Epoch 8/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1397 - val_loss: 0.1378\n",
            "Epoch 9/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1373 - val_loss: 0.1370\n",
            "Epoch 10/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1351 - val_loss: 0.1336\n",
            "Epoch 11/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1334 - val_loss: 0.1324\n",
            "Epoch 12/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1319 - val_loss: 0.1306\n",
            "Epoch 13/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1303 - val_loss: 0.1292\n",
            "Epoch 14/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1289 - val_loss: 0.1281\n",
            "Epoch 15/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1276 - val_loss: 0.1268\n",
            "Epoch 16/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1265 - val_loss: 0.1261\n",
            "Epoch 17/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1255 - val_loss: 0.1256\n",
            "Epoch 18/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1246 - val_loss: 0.1250\n",
            "Epoch 19/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1240 - val_loss: 0.1236\n",
            "Epoch 20/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1232 - val_loss: 0.1235\n",
            "Epoch 21/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1226 - val_loss: 0.1229\n",
            "Epoch 22/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1220 - val_loss: 0.1228\n",
            "Epoch 23/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1216 - val_loss: 0.1219\n",
            "Epoch 24/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1209 - val_loss: 0.1214\n",
            "Epoch 25/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1206 - val_loss: 0.1211\n",
            "Epoch 26/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1201 - val_loss: 0.1207\n",
            "Epoch 27/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1196 - val_loss: 0.1205\n",
            "Epoch 28/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1193 - val_loss: 0.1201\n",
            "Epoch 29/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1190 - val_loss: 0.1199\n",
            "Epoch 30/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1186 - val_loss: 0.1199\n",
            "Epoch 31/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1182 - val_loss: 0.1193\n",
            "Epoch 32/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1179 - val_loss: 0.1186\n",
            "Epoch 33/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1176 - val_loss: 0.1189\n",
            "Epoch 34/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1173 - val_loss: 0.1187\n",
            "Epoch 35/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1171 - val_loss: 0.1178\n",
            "Epoch 36/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1167 - val_loss: 0.1180\n",
            "Epoch 37/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1165 - val_loss: 0.1176\n",
            "Epoch 38/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1161 - val_loss: 0.1173\n",
            "Epoch 39/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1159 - val_loss: 0.1178\n",
            "Epoch 40/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1157 - val_loss: 0.1168\n",
            "Epoch 41/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1154 - val_loss: 0.1165\n",
            "Epoch 42/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1152 - val_loss: 0.1164\n",
            "Epoch 43/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1150 - val_loss: 0.1163\n",
            "Epoch 44/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1149 - val_loss: 0.1165\n",
            "Epoch 45/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1146 - val_loss: 0.1162\n",
            "Epoch 46/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1145 - val_loss: 0.1159\n",
            "Epoch 47/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1144 - val_loss: 0.1158\n",
            "Epoch 48/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1142 - val_loss: 0.1162\n",
            "Epoch 49/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1140 - val_loss: 0.1159\n",
            "Epoch 50/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1138 - val_loss: 0.1157\n",
            "Epoch 51/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1137 - val_loss: 0.1159\n",
            "Epoch 52/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1136 - val_loss: 0.1156\n",
            "Epoch 53/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1134 - val_loss: 0.1152\n",
            "Epoch 54/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1133 - val_loss: 0.1149\n",
            "Epoch 55/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1132 - val_loss: 0.1148\n",
            "Epoch 56/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1131 - val_loss: 0.1149\n",
            "Epoch 57/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1130 - val_loss: 0.1153\n",
            "Epoch 58/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1130 - val_loss: 0.1157\n",
            "Epoch 59/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1128 - val_loss: 0.1146\n",
            "Epoch 60/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1127 - val_loss: 0.1146\n",
            "Epoch 61/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1126 - val_loss: 0.1144\n",
            "Epoch 62/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1124 - val_loss: 0.1144\n",
            "Epoch 63/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1124 - val_loss: 0.1144\n",
            "Epoch 64/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1123 - val_loss: 0.1142\n",
            "Epoch 65/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1122 - val_loss: 0.1143\n",
            "Epoch 66/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1121 - val_loss: 0.1143\n",
            "Epoch 67/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1121 - val_loss: 0.1144\n",
            "Epoch 68/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1120 - val_loss: 0.1141\n",
            "Epoch 69/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1119 - val_loss: 0.1139\n",
            "Epoch 70/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1118 - val_loss: 0.1141\n",
            "Epoch 71/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1118 - val_loss: 0.1141\n",
            "Epoch 72/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1117 - val_loss: 0.1141\n",
            "Epoch 73/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1116 - val_loss: 0.1137\n",
            "Epoch 74/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1116 - val_loss: 0.1142\n",
            "Epoch 75/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1115 - val_loss: 0.1141\n",
            "Epoch 76/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1114 - val_loss: 0.1136\n",
            "Epoch 77/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1114 - val_loss: 0.1137\n",
            "Epoch 78/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1113 - val_loss: 0.1136\n",
            "Epoch 79/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1112 - val_loss: 0.1138\n",
            "Epoch 80/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1112 - val_loss: 0.1139\n",
            "Epoch 81/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1112 - val_loss: 0.1138\n",
            "Epoch 82/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1111 - val_loss: 0.1137\n",
            "Epoch 83/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1111 - val_loss: 0.1135\n",
            "Epoch 84/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1110 - val_loss: 0.1137\n",
            "Epoch 85/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1109 - val_loss: 0.1144\n",
            "Epoch 86/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1110 - val_loss: 0.1138\n",
            "Epoch 87/100\n",
            "60000/60000 [==============================] - 1s 18us/sample - loss: 0.1109 - val_loss: 0.1133\n",
            "Epoch 88/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1108 - val_loss: 0.1133\n",
            "Epoch 89/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1108 - val_loss: 0.1136\n",
            "Epoch 90/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1107 - val_loss: 0.1134\n",
            "Epoch 91/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1106 - val_loss: 0.1136\n",
            "Epoch 92/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1106 - val_loss: 0.1135\n",
            "Epoch 93/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1106 - val_loss: 0.1135\n",
            "Epoch 94/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1106 - val_loss: 0.1135\n",
            "Epoch 95/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1106 - val_loss: 0.1133\n",
            "Epoch 96/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1104 - val_loss: 0.1136\n",
            "Epoch 97/100\n",
            "60000/60000 [==============================] - 1s 17us/sample - loss: 0.1105 - val_loss: 0.1134\n",
            "Epoch 98/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1103 - val_loss: 0.1132\n",
            "Epoch 99/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1103 - val_loss: 0.1133\n",
            "Epoch 100/100\n",
            "60000/60000 [==============================] - 1s 16us/sample - loss: 0.1102 - val_loss: 0.1135\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f5ff47ee898>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eP3_P6rDMCFa",
        "colab_type": "text"
      },
      "source": [
        "# 결과 확인"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2y4hQ-VfFxF4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 303
        },
        "outputId": "8ff774c8-2bf6-4ded-90ed-b1b013f7c62b"
      },
      "source": [
        "decoded_imgs = model.predict(x_test_noisy)\n",
        "\n",
        "n = 10\n",
        "plt.figure(figsize=(20, 6))\n",
        "for i in range(1,n):\n",
        "    # 원본 표시\n",
        "    ax = plt.subplot(3, n, i)\n",
        "    plt.imshow(x_test[i].reshape(28, 28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "\n",
        "    # 노이즈 첨가 된 것\n",
        "    ax = plt.subplot(3, n, i + n)\n",
        "    plt.imshow(x_test_noisy[i].reshape(28, 28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "\n",
        "    # 노이즈 제거 된 것\n",
        "    ax = plt.subplot(3, n, i + 2*n)\n",
        "    plt.imshow(decoded_imgs[i].reshape(28, 28))\n",
        "    plt.gray()\n",
        "    ax.get_xaxis().set_visible(False)\n",
        "    ax.get_yaxis().set_visible(False)\n",
        "plt.show()"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x432 with 27 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fTqVK4F6F1S2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}