GANS_IN_ACTION_Chapter_4_DCGAN(Edited_Layer4).ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/zergswim/GAN/blob/main/GANS_IN_ACTION_Chapter_4_DCGAN(Edited_Layer4).ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "98kdUTdQZSYS"
      },
      "source": [
        "# 4장: 심층 합성곱 GAN (DCGAN)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LxvhXefnZSYW"
      },
      "source": [
        "<table align=\"left\">\n",
        "  <td>\n",
        "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/rickiepark/gans-in-action/blob/master/chapter-4/Chapter_4_DCGAN.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />구글 코랩에서 실행하기</a>\n",
        "  </td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "XiVqM0RRZSYX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2eadd60f-ad12-4dd7-8b0e-b28fed87ae12"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[name: \"/device:CPU:0\"\n",
            "device_type: \"CPU\"\n",
            "memory_limit: 268435456\n",
            "locality {\n",
            "}\n",
            "incarnation: 12599676818369461874\n",
            "xla_global_id: -1\n",
            ", name: \"/device:GPU:0\"\n",
            "device_type: \"GPU\"\n",
            "memory_limit: 14465892352\n",
            "locality {\n",
            "  bus_id: 1\n",
            "  links {\n",
            "  }\n",
            "}\n",
            "incarnation: 3292122390270365714\n",
            "physical_device_desc: \"device: 0, name: Tesla T4, pci bus id: 0000:00:04.0, compute capability: 7.5\"\n",
            "xla_global_id: 416903419\n",
            "]\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "#print(\"TensorFlow version: \", tf.__version__)\n",
        "#tf.debugging.set_log_device_placement(True)\n",
        "from tensorflow.python.client import device_lib\n",
        "print(device_lib.list_local_devices())\n",
        "#import os|\n",
        "#os.environ[\"CUDA_VISIBLE_DEVICES\"]='1'\n",
        "#os.environ[\"CUDA_VISIBLE_DEVICES\"]='0'\n",
        "\n",
        "%matplotlib inline\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "from tensorflow.keras.datasets import mnist, fashion_mnist\n",
        "from tensorflow.keras.layers import Activation, BatchNormalization, Dense, Dropout, Flatten, Reshape\n",
        "from tensorflow.keras.layers import LeakyReLU\n",
        "from tensorflow.keras.layers import Conv2D, Conv2DTranspose\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.optimizers import Adam"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "Da95dC9tZSYZ"
      },
      "outputs": [],
      "source": [
        "img_rows = 28\n",
        "img_cols = 28\n",
        "channels = 1\n",
        "\n",
        "# 입력 이미지 차원\n",
        "img_shape = (img_rows, img_cols, channels)\n",
        "\n",
        "# 생성자 입력으로 사용될 잡음 벡터의 크기\n",
        "z_dim = 100"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6oufNd2OZSYZ"
      },
      "source": [
        "## 생성자"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "gNs6dv45ZSYa",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "615a4294-f313-4218-acba-ad08c566feb5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " dense (Dense)               (None, 25088)             2533888   \n",
            "                                                                 \n",
            " reshape (Reshape)           (None, 7, 7, 512)         0         \n",
            "                                                                 \n",
            " conv2d_transpose (Conv2DTra  (None, 7, 7, 256)        1179904   \n",
            " nspose)                                                         \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 7, 7, 256)        1024      \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " leaky_re_lu (LeakyReLU)     (None, 7, 7, 256)         0         \n",
            "                                                                 \n",
            " conv2d_transpose_1 (Conv2DT  (None, 14, 14, 128)      295040    \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 14, 14, 128)      512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " leaky_re_lu_1 (LeakyReLU)   (None, 14, 14, 128)       0         \n",
            "                                                                 \n",
            " conv2d_transpose_2 (Conv2DT  (None, 14, 14, 64)       73792     \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_2 (Batc  (None, 14, 14, 64)       256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " leaky_re_lu_2 (LeakyReLU)   (None, 14, 14, 64)        0         \n",
            "                                                                 \n",
            " conv2d_transpose_3 (Conv2DT  (None, 28, 28, 1)        577       \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " activation (Activation)     (None, 28, 28, 1)         0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 4,084,993\n",
            "Trainable params: 4,084,097\n",
            "Non-trainable params: 896\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "def build_generator(z_dim):\n",
        "    model = Sequential()\n",
        "    model.add(Dense(256*2 * 7 * 7, input_dim=z_dim))\n",
        "    model.add(Reshape((7, 7, 256*2)))\n",
        "    model.add(Conv2DTranspose(256, kernel_size=3, strides=1, padding='same'))\n",
        "    model.add(BatchNormalization())\n",
        "    model.add(LeakyReLU(alpha=0.2))\n",
        "    model.add(Conv2DTranspose(128, kernel_size=3, strides=2, padding='same'))\n",
        "    model.add(BatchNormalization())\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "    model.add(Conv2DTranspose(64, kernel_size=3, strides=1, padding='same'))\n",
        "    model.add(BatchNormalization())\n",
        "    model.add(LeakyReLU(alpha=0.1))\n",
        "    model.add(Conv2DTranspose(1, kernel_size=3, strides=2, padding='same'))\n",
        "    model.add(Activation('tanh'))\n",
        "\n",
        "    return model\n",
        "\n",
        "generator = build_generator(z_dim)\n",
        "generator.summary()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5nvRR-bSZSYb"
      },
      "source": [
        "## 판별자"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "XsjOs-TWZSYb"
      },
      "outputs": [],
      "source": [
        "def build_discriminator(img_shape):\n",
        "\n",
        "    model = Sequential()\n",
        "\n",
        "    # 28x28x1에서 14x14x32 텐서로 바꾸는 합성곱 층\n",
        "    model.add(\n",
        "        Conv2D(32,\n",
        "               kernel_size=3,\n",
        "               strides=2,\n",
        "               input_shape=img_shape,\n",
        "               padding='same'))\n",
        "\n",
        "    # LeakyReLU 활성화 함수\n",
        "    model.add(LeakyReLU(alpha=0.01))\n",
        "\n",
        "    # 14x14x32에서 7x7x64 텐서로 바꾸는 합성곱 층\n",
        "    model.add(\n",
        "        Conv2D(64,\n",
        "               kernel_size=3,\n",
        "               strides=2,\n",
        "               padding='same'))\n",
        "\n",
        "    # LeakyReLU 활성화 함수\n",
        "    model.add(LeakyReLU(alpha=0.01))\n",
        "\n",
        "    # 7x7x64에서 3x3x128 텐서로 바꾸는 합성곱 층\n",
        "    model.add(\n",
        "        Conv2D(128,\n",
        "               kernel_size=3,\n",
        "               strides=2,\n",
        "               padding='same'))\n",
        "\n",
        "    # LeakyReLU 활성화 함수\n",
        "    model.add(LeakyReLU(alpha=0.01))\n",
        "    \n",
        "    # sigmoid 활성화 함수를 사용한 출력층\n",
        "    model.add(Flatten())\n",
        "    model.add(Dense(1, activation='sigmoid'))\n",
        "\n",
        "    return model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YBLXvJrKZSYc"
      },
      "source": [
        "## 모델 만들기"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "mNj-eGB7ZSYd"
      },
      "outputs": [],
      "source": [
        "def build_gan(generator, discriminator):\n",
        "\n",
        "    model = Sequential()\n",
        "\n",
        "    # 생성자 -> 판별자로 연결된 모델\n",
        "    model.add(generator)\n",
        "    model.add(discriminator)\n",
        "\n",
        "    return model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "GGrnRiBaZSYd"
      },
      "outputs": [],
      "source": [
        "generator = build_generator(z_dim)\n",
        "#generator.summary()\n",
        "optimizer_g = tf.keras.optimizers.Adam()\n",
        "train_loss = tf.keras.metrics.Mean()\n",
        "\n",
        "@tf.function\n",
        "def train_step(z, real):\n",
        "    with tf.GradientTape() as tape:\n",
        "        gen_imgs = generator(z)\n",
        "        pred = discriminator(gen_imgs)\n",
        "        loss = tf.keras.losses.binary_crossentropy(real, pred) #? real or fake?\n",
        "\n",
        "    gradients = tape.gradient(loss, generator.trainable_variables)\n",
        "    optimizer_g.apply_gradients(zip(gradients, generator.trainable_variables))\n",
        "\n",
        "    train_loss(loss)\n",
        "    g_loss = train_loss.result()\n",
        "    \n",
        "    return gen_imgs, g_loss\n",
        "\n",
        "# 판별자 모델을 만들고 컴파일하기\n",
        "discriminator = build_discriminator(img_shape)\n",
        "discriminator.compile(loss='binary_crossentropy',\n",
        "                      optimizer=Adam(),\n",
        "                      metrics=['accuracy'])\n",
        "\n",
        "# # 생성자 모델 만들기\n",
        "# generator = build_generator(z_dim)\n",
        "\n",
        "# # 생성자를 훈련하는 동안 판별자의 파라미터를 유지\n",
        "# discriminator.trainable = False\n",
        "\n",
        "# # 생성자를 훈련하기 위해 동결된 판별자로 GAN 모델을 만들고 컴파일\n",
        "# gan = build_gan(generator, discriminator)\n",
        "# gan.compile(loss='binary_crossentropy', optimizer=Adam())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z5Ui9V_uZSYe"
      },
      "source": [
        "## 훈련"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "ue_mENCRZSYe"
      },
      "outputs": [],
      "source": [
        "losses = []\n",
        "accuracies = []\n",
        "iteration_checkpoints = []\n",
        "\n",
        "\n",
        "def train(iterations, batch_size, sample_interval):\n",
        "\n",
        "    # MNIST 데이터셋 로드\n",
        "    (X_train, _), (_, _) = mnist.load_data()\n",
        "    #(X_train, _), (_, _) = fashion_mnist.load_data()\n",
        "\n",
        "    # [0, 255] 흑백 픽셀 값을 [-1, 1] 사이로 스케일 조정\n",
        "    X_train = X_train / 127.5 - 1.0\n",
        "    X_train = np.expand_dims(X_train, axis=3)\n",
        "\n",
        "    # 진짜 이미지 레이블: 모두 1\n",
        "    real = np.ones((batch_size, 1))\n",
        "\n",
        "    # 가짜 이미지 레이블: 모두 0\n",
        "    fake = np.zeros((batch_size, 1))\n",
        "\n",
        "    import time\n",
        "    start = time.time()  # 시작 시간 저장        \n",
        "\n",
        "    for iteration in range(iterations):\n",
        "\n",
        "        # -------------------------\n",
        "        #  판별자 훈련\n",
        "        # -------------------------\n",
        "\n",
        "        # 진짜 이미지에서 랜덤 배치 가져오기\n",
        "        idx = np.random.randint(0, X_train.shape[0], batch_size)\n",
        "        imgs = X_train[idx]\n",
        "\n",
        "        # 가짜 이미지 배치 생성\n",
        "        z = np.random.normal(0, 1, (batch_size, 100))\n",
        "        # gen_imgs = generator.predict(z)\n",
        "\n",
        "        # 생성자 훈련?\n",
        "        gen_imgs, g_loss = train_step(z, real)\n",
        "\n",
        "        # 판별자 훈련\n",
        "        d_loss_real = discriminator.train_on_batch(imgs, real)\n",
        "        d_loss_fake = discriminator.train_on_batch(gen_imgs, fake)\n",
        "        d_loss, accuracy = 0.5 * np.add(d_loss_real, d_loss_fake)\n",
        "\n",
        "        # ---------------------\n",
        "        #  생성자 훈련\n",
        "        # ---------------------\n",
        "\n",
        "        # 가짜 이미지 배치 생성\n",
        "        # z = np.random.normal(0, 1, (batch_size, 100))\n",
        "        # gen_imgs = generator.predict(z)\n",
        "\n",
        "        # 생성자 훈련\n",
        "        # g_loss = gan.train_on_batch(z, real)\n",
        "\n",
        "        if (iteration + 1) % sample_interval == 0:\n",
        "\n",
        "            # 훈련이 끝난 후 그래프를 그리기 위해 손실과 정확도 저장\n",
        "            losses.append((d_loss, g_loss))\n",
        "            accuracies.append(100.0 * accuracy)\n",
        "            iteration_checkpoints.append(iteration + 1)\n",
        "\n",
        "            # 훈련 과정 출력\n",
        "            print(\"%d [D 손실: %f, 정확도: %.2f%%] [G 손실: %f]\" %\n",
        "                  (iteration + 1, d_loss, 100.0 * accuracy, g_loss))\n",
        "            print(\"time :\", time.time() - start)  # 현재시각 - 시작시간 = 실행 시간\n",
        "            # 생성된 이미지 샘플 출력\n",
        "            sample_images(generator)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "pLeOBhl6ZSYf"
      },
      "outputs": [],
      "source": [
        "def sample_images(generator, image_grid_rows=4, image_grid_columns=4):\n",
        "\n",
        "    # 랜덤한 잡음 샘플링\n",
        "    z = np.random.normal(0, 1, (image_grid_rows * image_grid_columns, z_dim))\n",
        "\n",
        "    # 랜덤한 잡음에서 이미지 생성하기\n",
        "    gen_imgs = generator.predict(z)\n",
        "\n",
        "    # 이미지 픽셀 값을 [0, 1] 사이로 스케일 조정\n",
        "    gen_imgs = 0.5 * gen_imgs + 0.5\n",
        "\n",
        "    # 이미지 그리드 설정\n",
        "    fig, axs = plt.subplots(image_grid_rows,\n",
        "                            image_grid_columns,\n",
        "                            figsize=(4, 4),\n",
        "                            sharey=True,\n",
        "                            sharex=True)\n",
        "\n",
        "    cnt = 0\n",
        "    for i in range(image_grid_rows):\n",
        "        for j in range(image_grid_columns):\n",
        "            # 이미지 그리드 출력\n",
        "            axs[i, j].imshow(gen_imgs[cnt, :, :, 0], cmap='gray')\n",
        "            axs[i, j].axis('off')\n",
        "            cnt += 1\n",
        "    plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kp1YgmIGZSYg"
      },
      "source": [
        "## GAN 훈련과 출력 점검"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "l1mUj8CYZSYg",
        "outputId": "8da3e504-7507-4d1b-c207-bc24448a8859",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1000 [D 손실: 0.092842, 정확도: 97.27%] [G 손실: 7.068509]\n",
            "time : 41.981138706207275\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "2000 [D 손실: 0.003808, 정확도: 100.00%] [G 손실: 8.792349]\n",
            "time : 76.43957543373108\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3000 [D 손실: 0.000206, 정확도: 100.00%] [G 손실: 9.656578]\n",
            "time : 114.45717573165894\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4000 [D 손실: 0.000005, 정확도: 100.00%] [G 손실: 11.767351]\n",
            "time : 149.69792366027832\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "5000 [D 손실: 0.000001, 정확도: 100.00%] [G 손실: 13.849445]\n",
            "time : 184.0895709991455\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "6000 [D 손실: 0.000004, 정확도: 100.00%] [G 손실: 14.710831]\n",
            "time : 218.82390117645264\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "7000 [D 손실: 0.000003, 정확도: 100.00%] [G 손실: 15.433132]\n",
            "time : 253.3684265613556\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "8000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 15.875152]\n",
            "time : 287.9946777820587\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "9000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 15.986144]\n",
            "time : 322.8369083404541\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 16.208960]\n",
            "time : 357.51191329956055\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "11000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 16.720135]\n",
            "time : 392.0103394985199\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "12000 [D 손실: 0.000011, 정확도: 100.00%] [G 손실: 17.077097]\n",
            "time : 426.4084782600403\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "13000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 17.970600]\n",
            "time : 461.0630724430084\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "14000 [D 손실: 0.000002, 정확도: 100.00%] [G 손실: 18.577988]\n",
            "time : 495.45064306259155\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "15000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 19.261410]\n",
            "time : 529.9501070976257\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "16000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 20.077206]\n",
            "time : 564.5400912761688\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAOwAAADnCAYAAAAdFLrXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAZjklEQVR4nO3df1AU5xkH8O9753E/aINBZFChjuCvpBJBrGGmgZoYp81EyBnRdPKfiXGmU0mgJb+m6dj8kUyxChQ74ww46EysnSGmIJhOJsYxQUwIF5WidWKrGCFFovgDe3fccdy+/UPvAsgdd3DLu+/d85nZmeNub+952Htud999913GOQchRA460QEQQkJHBUuIRKhgCZEIFSwhEqGCJUQmnPOAEwAu28QY401NTfyeoPnJnOfIKdQ8Zcy1pKSEjxGVeYa6TmkLS4hEqGAJkciMybwpLS0NJSUlMBgM4JyjtrYW586di3Rsk8I5R01NDY4ePYrq6mrR4ZAwGQwGlJaWIjU1FQCQlZUlOCKNCfc4wGQy8ccee4wPDg5yzjlXFIUXFBQI3+cfb4r14x0Zc7VYLPz06dNjj1vpGPbeFNYW1mQyoa6uDitWrIDRaAznrYSQCAirYHU6HZYuXYolS5aoFQ8hJAhqdCJEImEVrMfjwcGDB/G3v/0Nw8PDasVEYhh9xyYwmQP3nJwc7nA4uKIo3Ov1UqOTJHnKkitjjK9cuZI7nU5qdJpKo5PPxYsXUVRUhBkz7r69vb19Mosh5D5GoxHl5eXUsBkAu/drNP6LjAV+UQKccxbKfLGSJ6D9XC0WC1pbW5GdnR1olphep9ToRIhEJrVLTKLPzJkzkZeXh8uXL+PcuXPIzc1FSkrKffN5vV60tLRgYGBAQJTRKyMjA5mZmWhra0NfX1/gGWPxwD1W8wylIbGyspIzxvjhw4fHa/Dh//vf/3hWVpZqucRqT6dXXnllVANuRBudSHRijOHxxx9HdXU1MjMzx51Hp6OjKDUwxqDT6bB161asXbs24HxUsDHMYDDAYDAAAMxmMxhjWL58OZYvXy44su8NDw9jaGjI/7fFYhEYjXo8Hg+cTieeeuop6PX6gPNRwcaw0tJS/PKXvwRwtxC0eBqlvr4ef/rTn/x/nzlzRmA06qmvr8cXX3yByspK5OfnB5yPCjaGpaamBjt9IoSiKPjXv/4FRVEAAF999RU6OjoER6W+69evo7+/H6dOnUJ8fDxycnLGnY8KlmiKy+XCli1b/MfKsdQ9kXOON954A3q9Hk6nc9x5qGBjTElJif+xFi8OnzFjBp599lkAwPvvvx8zBZuVlYXVq1dPPGM0NY2PnaL1FMBk87w3TYnD4VD9tM6pU6e4zWbjFoslqtcpY8w/lZaWjv1X02kdon1utxulpaXIyMjAgQMH8I9//AN79+4VHVbE+fpMp6enAwAWLFgQ0vuoYMmEXC4Xbt68CUVR4HK5VN1N9fWkun37Nn73u9+hs7NTtc8SyTcYxI9//GMAwA9/+MOQ3ked/xE7efpmD3f5H330EV588UX/bll/f7/qx5YGgwGzZs2C0+nEnTt3/M9HyzpljGHWrFn+8+AvvfQS3n777VGzjPc+2sKSCblcLvT29k7rZ3o8nuB9aiXn++HzOXPmDBobGwP24fahfmaEaEBzczM2bNgAm80WdD5hW9iEhASUlZUhISEBAHDo0CG0tLSICmfaFBUVITc3F1VVVfj222+n/fNffvnlUbEE6lVTU1PjH2v64sWL0xJbLFu9ejU2bNiAZcuWBZ9RVNP4vHnzeG9vr78Nu7i4WNjpDjXz9E16vZ5bLBa+Z88e7nQ6eW5uLrdYLJOaDAbDpPIcm+vOnTu50+nkiqKMOp9AY01P//TrX/+aOxwOPjw8TKd1tCAvLw+7du1CWloaTCYT9u/fH7A3y0QOHjyInTt3TjmmqqoqHDp0CPv376ehawWjvsQapCgKuru70d3dPer55ORkpKWlhbyczz77LCLxfPvtt+jv74fNZoPdbvc/zznHrVu3IvIZJDS+y+sYC94ITgU7TVpaWgL+cm7btg07duyY5ojuGtt318ftdguJJ1Zt2rQJO3bsmPCKKWEFa7fbUVtbi+zsbBQUFIgKY9ooioLBwcFxX/N4PJNaZmFhob+nzFRQcYo3Y8YMmM3mCecTdlpnYGAAf/jDH7Bv3z5RIUjLt/v04osvoqKiQnQ4ZBoJ28ImJiaioqICDz/8sKgQpFVQUIAFCxbg6NGj2Lt3L5qamkSHRKaJsC1sXFwcHnnkEWRkZIgKQTiDwYC5c+f6z0WHKj09HQUFBbDb7Th9+rRK0ZHpZLfb0dvbC5fLFXxGUeeydDodT05O5ps3b46J87DjTY888gjv6uriAwMDYy+tCsnNmzd957IndR5Wxila84yPj+dz587lH3/8sTbPwyqKgmvXruH8+fNoaGiIyd40DocDNpsNcXFxADBhP9KxLly4gL6+PlitVrVCJNPE4XDA4XDg+PHjcDgcgddpNP1KjZ1k+jXW6XS8qakp5K2roii8sLAwrDy1kmusrFM18qTzsBqhKApqamrQ3t6OV199FQ888EDQ+RljeOmll/Dkk09OU4Ryqa6unnAezjlqa2v9faalEIu/UlrN02g08oULF/KrV6+GvKW9h7aw908TkrHPNG1hNYIxhvLycqxZswZJSUmiwyEaRdfDaoiiKFAUxbclJOQ+VLAawfndMWmfffZZ3LhxQ3Q4RKOoYDVkaGgI/f39qKmpQXNzs+hwiAZRwWrMwMAAtm/fjrq6Ov/tKgjxoYLVqLa2NlitVhw5ckR0KERDqGA1qq+vD83Nzbh8+bLoUIiGBB2XmBCiLbSFJUQiVLCESIQKlhCJUMESIhEqWEIkQgVLiESoYAmRCBUsITIJ9SJgNQZJU3uazMXOUZ4nrFYr903p6enCY1crV9FxqpUnXcAeYxoaGvyPi4uL8Ze//EVgNCRcIRdsUVER5s+fL+y+pmoaec/USN1oSqtG5hoL9+ONNhP1JR714uDgIJ544gl0dHRMPOCxBnDOg98K7B7GWNB/gtaFmicQO7lGa55hFSznHP/+979hs9mwZcsWzd9EKdZX7nhiJddozTOsY1jGGJYsWQK73X7f7QkJIeqjqiNEItRKHGNKSkr8j48fP45//vOfAqMh4Zp0wTLG/Ld3p4vg5VFZWel//Morr6Czs5PWn0TCanTyGRgYwMmTJ+H1esE5xzvvvIP29nZ1IpyCWG+gCDS778G5c+fw9ddf4/XXX0dXV5cKkUVezK/TUHsABSLj7Q5ipVdMKOvU6XTyn//853z27NnC86B1OnGe1OgU40wmEw4cOID33nsPJpNJdDhkAlSwMY4xhqSkJCQlJfnbJIh2UcESIhE6rUMAAGlpadi1axeGh4fBuYT3TdWwhIQElJWV4cqVK9i7d++UlhW0YJ1O54QLUBQFw8PDUwpiKkwmk7/XlcfjgcfjERaLzJKTk/GrX/0KAMA5xyeffEIFGyHx8fF4/vnnYbPZ1C3Yn/70pyEtRNQpAZPJhLq6OixduhQAcPDgQezcuVNILIQEcv36dTzzzDOw2+1TXlbQgu3o6JjyB6hJp9Nh6dKlyM7OBhD9l8ZNh2vXrqG7uxu3bt2CwWDAQw89BIfDgUuXLokOTVoejydieyvU6ERGqa+vR35+Pj7//HPMnj0bjY2NePfdd0WHRe6RutHJ4/Hg4MGD+Prrr7Fx40bR4Ujt2rVrqK+vx7FjxzA4OAgAcDgceO+99/DNN9+IDY58Lxp6i+Tk5HCHw8ErKir8zzHGYr5XTCg9nRRF4YqicJvNxs1m86Q+nzHmn7SS61RymK5cJpOn1FtYn4sXL6KoqAjd3d0AgLfeegurVq0SHJX21dTU+O8/e+vWrUkNSFBYWIgtW7YAAG7cuIHf/OY3uHXrVkTjVFtiYiIqKiqQmJjof87tdmuzj7XMWx7GGJ89e7a/H2x8fDxPTU3lH3/8sW8DQlvYMVNPTw/3TZs3b57yZ5eUlPi31j09PXzOnDmayDXU5SUmJvKcnBz+3XffjdrzcDgcPCsrS3PrVOotrMlkwl//+lcoigKr1Yrnn38eb7/99qhfSjLayD2PgYEBgZGIxxhDRUUFnnrqKSQlJYkOJyRSF6yiKDh79iw451AUBfHx8ZgzZ47osDTt6tWrEV3epUuX0NjYiNzcXP9zc+bMwaOPPorOzk7t7VKOwDnHhQsXkJKSgtWrV8NoNIoOaWLRtKs4cveMdomnL1edTsebmpr8u8RWq5UrisK3bdsmLNdwljlv3jze29tLu8QkNiiKgpqaGixatAivvfYaFi9ePGpEEhI5VLAkIo4cOYJ58+bBZrP5D0vi4uJgsVgA3C1ql8uFGTNmIC4uDm63G16vV2TIfpxzDA4Owu12a363mHo6EdWUlJSgtbUVra2tqK2thdFoxMaNG3HixImQ+6lPB19f3zfffFN0KBOSegur0+mwePFiAMCFCxcER0PGSk1NRWpqKoC7LforV67EypUrkZ2djYSEBMHRfc/X13fhwoWiQ5mQ1AVrNBqxb98+DA8PY+3ataLDIUEsXboUn3zyCfR6vehQpCZ1wQ4PD+ODDz6A1+vVzPEQGR9jbNSYUYWFhcjIyABw99RQc3OzqND8Ll68iKqqKjz++ONYtGiR6HDGJ/vpjpF9P0tLS+m0jsBcxzs9EoqGhoaI5zqV79Pu3bvptI4ajEYjysvLkZ6eDgBYsGCB4IiIzHJzc/Hmm28iMzNTdCgBSV2wer0e+fn5/gvYiVherxd9fX0wGo1Sdg9NSUlBYWEhgNCGRxKBTuuQiLl27Rp+8YtfoKysTHQoUUvqLazX68Xx48dx5cqVcV+3Wq3THFFsUxQF165dw/nz59HQ0BByT6cvv/xS5ciiCDXGxE6esZTrZJZttVr9DWFabXSiXWJC7uns7ERxcTFaWlpEhxKQdAWr1+thsVhgsVhgNpvpTvBkSoxGI8xmMxhjuHLlCurq6nD+/HnRYQUk3bc9Ly8PJ06cQGtrK44dO6bdE9xE8xhjKC8vR0NDA2bNmuX/bm3YsEF0aAFJ0+jkGyM3JycHK1asAHC36d1sNguOjMgsPT0dK1asQE5ODpYtWzbqu6VF0hSsb4zcefPmiQ6FRBnfd0uGwytpCtY3Ru4DDzzgf25oaAjXr18XGBWJFrLcG1eagh0YGMD27dtFh0GijO90iSyjY2h/H4AQlXDO8c477+CFF16QZixlabawhKihvb0d//3vf+FyuUSHEhJ2r1cIIUQCtEtMiESoYAmRCBUsIRKhgiVEIlSwhEiECpYQiVDBEiIRKlhCJEIFS4hMgo2L09DQwN944w0OgFdXV/vHu/nqq6+42WwWPr7P2KmgoIA3NDTwq1ev+kKdlvF/xqqsrNTE+D+ixnTy/W+Ki4uF5Sr6u6hWnkH7Eq9fvz7Yy5rT3NyMDz/8EI2NjSgoKBAdDiERF3Ln/0OHDuE///kPgLvjzw4NDakW1FT4bi589OhRVFdXq/Y5vgG7Nm7ciPz8fNU+RyYJCQkoKyvzj9pAVBCLuxWRzHPkoUKs7xKPvbcO7RJHPk9qdCJEIlSwU9TT04MzZ85gcHAQTqcTZ86cQXd3t+iwSJSiC9inqKqqCnV1dTh27BgURcGaNWtgt9tFh0WiFBXsFHk8Hty5cwcHDhyA1+uF3W6Hx+MRHZYQdrsdtbW1yM7OplZ6tcTigXus5jmd52G9Xi/ftm0brdMI5xl0iBjGWOAXJcA5D2kovFjJE5ieXFNSUvCTn/wEZ8+exTfffBPRZcf6OqWCRezkCcROrtGaJ7USEyIRKlhCJEIFS4hEqGAJkQgVLCESoYIlRCJUsIRIhAqWEIlQwRIikaCd/0fe5DZYjyhCyPQIWrCHDx/2P66pqcGRI0dUD4gQEljQgh15idTRo0dVD4YQEhwdwxIiESpYQiRCBUuIREIu2KKiIuzcuROpqalqxkMICSLoBey4O1yF3+DgIJ544gl0dHTA5XKpG1kEyHyxs16vh9FoBHD3lJrb7YaiKOPOSxew3y9a8wxrl9hkMmH//v2ora31f5mIOvLy8nDixAm0trbi2LFjWLRokeiQiAaENWoiYwxLliyB3W6HTkeHv2owGAx46KGHkJOT47/lhcvlwsqVK/GDH/wAwN1bpfT09IgMk4gywehz49Lq3evGTjKOsDd37lze1dXF3W73qP+5y+XiTqeTO51OvmPHjknlqbVcY2WdRjJPGpdYYxhjMJlMiIuLG/X8yEMQg8Ew3WHFhIyMDKxbtw6MMXg8HtTX1+P69euiwxolrIL1NVBRv2L1jOy/TaZXZmYmqqqqAABOpxMnT56Ut2BH9iW+desW3G63akHFqrfeegv5+flITEwUHQrRqJAL9ty5c2hublYzlqik0+mQnJwMvV4PALh9+zYcDsd98zHGsGrVKqxdu3a6QyQSoaZelSUnJ+Ojjz6CzWaDzWbDpk2bRIdEJBZ0C9vY2Oh/fPHiRdWDiUZDQ0M4ffo0HnzwQQBAX1+f4IiIzIIW7Pr166crjqh18+ZNvPDCC6LDIFGCdok1gnOOmpoa/P73v8edO3dCes/WrVtRXV2tcmSxo7OzE8XFxWhpaREdSmCxePJZy3nOmzeP9/b2BuqzwjnnfPfu3Tw+Pp5/+OGHvqeo40QE8tTr9dxisfA9e/Zwh8PBs7KyNJcndZyQ0HPPPYfHHnsM6enpokOJKnl5edi1axfS0tJEhxIQFayEZs+ejdmzZ4sOI+rMnDnT33/b6XQKjmZ8dAxLiESoYDXGbrejtraWOqmQcVHBaszAwAC2b9+Ourq6gBesA6D+3CpRFAWc86D/e5GoYDWqra0NVqt13LGgm5qaYLVa0d7eLiCy6OX7nz/zzDPYtGkTurq6RId0PzoFoO08q6ur7zutU1lZyRljvKmpiU7rxFieE43pRAjRENolJkQiVLCESIQKlhCJUMESIhEqWEIkQgVLiESoYAmRCBUsIRKhgiVEIkGvh43WO4CNFUqejDEcPnwYBQUFEy6vubkZdXV1AO7e8e/TTz9VdRznUPMEaJ3KImCesdgfczJ5jum7G7Kenh4+Z84cTeRJ61SeKVBetEtMiESoYMPgdrsxODhI16ISYahgQ8Q5x+uvv47169fjxo0bosMhMYoKNgw6nY5uZE2EolETQ8QYwx//+Ec8/fTTMJlMosMhMYo2F2GIi4ujYiVCUcESIhEqWEIkQgVLiESoYAmRSFS1EmdkZCAzM1N0GISoJqoKdt26daiqqhIdBiGqiaqCZSzki1bCsnr1amzYsAHLli1TZfmEhCqqCjbSdDodjEYjVq1ahW3btsHtdod9G0KXy0V9j0nEUMEGsXjxYuzbtw+pqan+vsSfffZZWMsYHh5Gf3+/ShGSWBNVBfvdd9/h1KlTWLhwIRISEqa8vPj4eCxfvhxmsxmcc3R1daGjoyMCkRIyOVF1Wuf9999Hfn4+Tp48KToUQlQRFVvY5ORkbNq0CTNmzABjDD/60Y9Eh0SIKqKiYNPS0rBjxw6YzeaILZMxplqrMyGTFRUFG2lGoxHl5eVYsWIFjEaj6HAI8aOCHYder0d+fj6ys7NFh0LIKFHV6ERItIuKgr19+zaam5tx9uxZ0aEQoqqoKNhLly7hueee8w/eTUi0kvoY1mAwoLS0FKmpqQCArKysKS+zqKgIa9as8S/ThzGGrVu3Yu3atQCAzs5O7N27d8qfR0hYZB493Ww2888//5wPDg4GGng/7FHiq6urQxrRv6GhQXj+volG/td2nowxbjabudFonHKeUu8Su91ubN68GVu2bFH13jWETEVSUhIOHz6Md999d8rLknqXWFEUXLhwAYwx2Gw2zJ8/H2lpaWEvJycnx/84OTk5kiFGhE6nw+LFiwEAFy5coKt/JKQoSmTWmyy7FcEmxhg3mUz81VdfndQusdPp5L5peHhYc7vEZrOZf/HFF/zEiRPcZDJNejdRpnUaaJI1T5PJxOPi4qacp9S7xD6cc7hcLrS3t+PPf/4zurq6wnq/2Wz2T3q9XqUoJ2f16tUoLi5Gamoq5s+fj+LiYvzsZz8THRYJk8vlwtDQ0NQXJNOvVCiTTqcbeVvIULc8YZuOLSxjjO/evfu+z66srJzUVkfWdTqZXEXHqVaeUbGF9SksLERjYyNWrVolOpQpy83NRWNjI55++mnRoRANkbrRaaz09PSQ7pAug4SEBKxYsQIzZ84UHYoqEhMT/VdXDQ4O4ubNm4IjkkNUbWGjyaeffopVq1ahvr5edCgRxxhDRUUFbDYbbDYbdu7cKTokaUTVFvbSpUtobGz0/221WkN638j3hOrLL78M+z3hSExMxKOPPoqUlJT7XsvIyIDVakVbWxv6+vpUjSNSZs6ciby8POj1ejDG8PDDD2POnDkAgAcffFBwdBKJxQN3GfK0Wq1cUZSAjV5er5cXFBSElafIXHNycrjD4ZhyA57W84zUFBONTtGks7MTL7/8MlpaWsZ9XafTYevWraiurp7myCanp6cHZWVl+Pvf/y46FLnF4q+UDHnq9XpusVj4nj17Am5lR9D8Flan03Gz2cxfe+01TW1hjUYjt1gs3GKxhNXXV9R3N6qOYaNJXl4edu3aNamullo0coxnrWCMoby8HPn5+QCA48eP47e//a3gqIKjgtWwiPU/1YCRYzxrSXp6un8ooCtXrgiOZmJ0DKtRLS0tyM/PxwcffCA6FKIhtIXVqAULFmDdunV0Ay4yChWsRmVmZqKiogI6He0Eke/Rt0Gj2traYLVaceTIEdGhEA2hgtWovr4+NDc34/Lly6JDIRrCoqUVkpBYQFtYQiRCBUuIRKhgCZEIFSwhEqGCJUQiVLCESOT/8rGgkTtbU/wAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "17000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 20.645252]\n",
            "time : 599.25426197052\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "18000 [D 손실: 0.000001, 정확도: 100.00%] [G 손실: 20.919775]\n",
            "time : 633.7753727436066\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "19000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 21.107119]\n",
            "time : 668.429276227951\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "20000 [D 손실: 0.000000, 정확도: 100.00%] [G 손실: 21.248672]\n",
            "time : 703.4061691761017\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# 하이퍼파라미터 설정\n",
        "iterations = 20000\n",
        "batch_size = 128\n",
        "sample_interval = 1000\n",
        "\n",
        "# 지정된 반복 횟수만큼 DCGAN 훈련\n",
        "train(iterations, batch_size, sample_interval)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "87MsILUKZSYh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "outputId": "b4d43b95-917f-45ac-ee86-e91ff5180761"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fe4206fb450>"
            ]
          },
          "metadata": {},
          "execution_count": 10
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "losses = np.array(losses)\n",
        "\n",
        "# 판별자와 생성자의 훈련 손실 그래프\n",
        "plt.figure(figsize=(15, 5))\n",
        "plt.plot(iteration_checkpoints, losses.T[0], label=\"Discriminator loss\")\n",
        "plt.plot(iteration_checkpoints, losses.T[1], label=\"Generator loss\")\n",
        "\n",
        "plt.xticks(iteration_checkpoints, rotation=90)\n",
        "\n",
        "plt.title(\"Training Loss\")\n",
        "plt.xlabel(\"Iteration\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "D9ZzVg41ZSYh",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 378
        },
        "outputId": "ae91b58c-fff0-428b-b56a-bbbb97943ba9"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fe3b2238b10>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x360 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "accuracies = np.array(accuracies)\n",
        "\n",
        "# 판별자의 정확도 그래프\n",
        "plt.figure(figsize=(15, 5))\n",
        "plt.plot(iteration_checkpoints, accuracies, label=\"Discriminator accuracy\")\n",
        "\n",
        "plt.xticks(iteration_checkpoints, rotation=90)\n",
        "plt.yticks(range(0, 100, 5))\n",
        "\n",
        "plt.title(\"Discriminator Accuracy\")\n",
        "plt.xlabel(\"Iteration\")\n",
        "plt.ylabel(\"Accuracy (%)\")\n",
        "plt.legend()"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "TensorFlow 2.3 on Python 3.6 (CUDA 10.1)",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.7"
    },
    "colab": {
      "name": "GANS_IN_ACTION_Chapter_4_DCGAN(Edited_Layer4).ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}