[E-10]Learning_GAN_Fashion_MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "published-growing",
   "metadata": {},
   "source": [
    "## 학습 목표\n",
    "\n",
    "- 생성 모델링 개념을 이해하며 판별 모델링과의 차이 알기\n",
    "- Pix2Pix, CycleGAN 등의 이미지 관련 다양한 생성 모델링의 응용을 접하며 흥미 가지기\n",
    "- Fashion MNIST 데이터셋의 의미를 알기\n",
    "- 생성적 적대 신경망(GAN)의 구조와 원리를 이해하기\n",
    "- 텐서플로우로 짠 DCGAN 학습 코드를 익히며 응용하기"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "detailed-vegetable",
   "metadata": {},
   "source": [
    "## 여러 가지 생성 모델링 기법과 친해지기 (1) Pix2Pix\n",
    "- 이미지 관련 모델\n",
    "- paired data 필요\n",
    "- 단순화된 input을 넣으면 ground truth를 반환하도록 학습"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "intense-carrier",
   "metadata": {},
   "source": [
    "## 여러 가지 생성 모델링 기법과 친해지기 (2) CycleGAN\n",
    "\n",
    "- 그림 <-> 실사 양방향으로 변환 가능(Pix2Pix는 한 방향)\n",
    "- 스타일(?)을 학습해서 다른 것에 입히는 방식이라 labelling 필요 없음"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "attached-quarter",
   "metadata": {},
   "source": [
    "## 여러 가지 생성 모델링 기법과 친해지기 (3) Neural Style Transfer\n",
    "- Base image와 Style image 두 장을 놓고 새로운 이미지를 만들어 냄"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "close-duration",
   "metadata": {},
   "source": [
    "## 패션을 디자인하려면? 먼저 패션을 배워야지! (2) 코드로 살펴보기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "returning-trust",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensorflow 2.4.1\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import glob\n",
    "import time\n",
    "\n",
    "import PIL\n",
    "import imageio\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "from IPython import display\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "print(\"tensorflow\", tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "tracked-republic",
   "metadata": {},
   "source": [
    "> fashion_mnist는 tensorflow 안에 모듈로 있다.  \n",
    "> 분류 문제가 아니기 때문에 y(라벨) 데이터는 가져오지 않는다"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "adjacent-shape",
   "metadata": {},
   "outputs": [],
   "source": [
    "fashion_mnist = tf.keras.datasets.fashion_mnist\n",
    "\n",
    "(original_train_x, _), (original_test_x, _) = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "african-frank",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max pixel: 255\n",
      "min pixel: 0\n"
     ]
    }
   ],
   "source": [
    "print(\"max pixel:\", original_train_x.max())\n",
    "print(\"min pixel:\", original_train_x.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "attached-machinery",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "max pixel: 1.0\n",
      "min pixel: -1.0\n"
     ]
    }
   ],
   "source": [
    "train_x = (original_train_x - 127.5) / 127.5 # 이미지를 [-1, 1]로 정규화합니다.\n",
    "\n",
    "print(\"max pixel:\", train_x.max())\n",
    "print(\"min pixel:\", train_x.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "pursuant-county",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_x.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bacterial-withdrawal",
   "metadata": {},
   "source": [
    "> 여기서 주의할 점은 딥러닝에서 이미지를 다룰 때는 `채널 수`도 필요함.  \n",
    "> 그래서 4차원 tensor로 만들어줘야 함\n",
    "> 그런데 RGB는 채널 3개이지만, 흑백은 채널이 1개이니 그렇게 변환"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "demanding-prototype",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28, 1)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_x = train_x.reshape(train_x.shape[0], 28, 28, 1).astype('float32')\n",
    "train_x.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afraid-marker",
   "metadata": {},
   "source": [
    "----------------------\n",
    "----------------------\n",
    "----------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "infrared-expansion",
   "metadata": {},
   "source": [
    "> #### 실험\n",
    "> - 한 번 3차원(가로, 세로, 채널)을 `plt.imshow()`에 넣어도 되는지 확인해보고 싶다.\n",
    "> - 정규화된 사이즈가 달라도 이미지가 제대로 나오는지 확인해보고 싶다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "wooden-wiring",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAATEAAAD4CAYAAACE9dGgAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAAc7ElEQVR4nO3de3Bc5Znn8e8jWfJFlm/YCANODMQkcZLFsA4QoDIkzIRLpcawyVBQs8SZocbsLuyEKf6AYWcrbE2xRWUDbGYyYccENqYKwjIBFoZxhYtDQkiGizEOvi2xARNjfDfYxrZsqfvZP/ootCyd5xypW+o+5vehTql1nn77vD6SHs7lOe9r7o6ISFG1NLoDIiK1UBITkUJTEhORQlMSE5FCUxITkUIbM5oba7exPo6O0dykyEdKN/s57Iesls+48Esdvmt3Kdd7X3nt0JPuflEt26tVTUnMzC4Cvge0Aj9099ui94+jg7Psglo2KSKBF31ZzZ+xa3eJl578WK73ts5cP73mDdZo2KeTZtYK/ANwMTAXuNLM5tarYyLSGA6Uc/6XxcxmmdmzZrbWzNaY2beS9beY2WYzW5ksl1S1+Wsz22Bmr5vZhVnbqOVI7Exgg7u/mWz4QWABsLaGzxSRBnOcHs93OplDL3CDu68ws07gFTN7Oond6e7frX5zciB0BfAZ4HjgGTM71T29Q7Vc2D8B2FT1/TvJun7MbJGZLTez5T0cqmFzIjJa6nUk5u5b3H1F8nofsI5B8kSVBcCD7n7I3d8CNlA5YEo14ncn3X2xu8939/ltjB3pzYlIjRyn5PkWYHrfQUqyLEr7XDObDZwOvJisus7MXjOze81sarIu18FRtVqS2GZgVtX3JybrRKTgyniuBdjZd5CSLIsH+zwzmwg8DFzv7nuBu4BTgHnAFuD24fa1liT2MjDHzE4ys3Yq57GP1/B5ItIEHCjhuZY8zKyNSgK7390fAXD3be5ecvcycDcfnjIO+eBo2EnM3XuB64AnqZznPuTua4b7eSLSPIZwJBYyMwPuAda5+x1V62dWve0yYHXy+nHgCjMba2YnAXOAl6Jt1FQn5u5LgaW1fIaINBcHeuo3RNe5wFXAKjNbmay7mUpJ1rxkcxuBawDcfY2ZPUSlyqEXuDa6MwmjXLEvIs3Ph3CqmPlZ7s8Dgz1BkHrw4+63Arfm3YaSmIj051Aq0FipSmIi0k+lYr84lMRE5AhGadAzwOakJCYi/VQu7CuJiUhBVerElMREpMDKOhITkaLSkZiIFJpjlAo0cr2SmIgMoNNJESksxzjsrY3uRm5KYiLST6XYVaeTIlJgurAvzcMyfhlrHK2g9ZhpYfy9C09NjU164IWatp31b7Mxbakx7zlc27ZrlfVzidRvhImUjzdKriMxESmwso7ERKSoKhf2i5MaitNTERkVurAvIoVXUp2YiBSVKvZFpPDKujspIkVVeQBcSUyahLXGj494b28Yb5k3N4yvu2Zi3P5geqxtfzg7PWMOxoMktz21PIzXVAuWVYOWsV+xOAnU0jcbE/zZxj/OXByjR48diUhRuaNiVxEpMlOxq4gUl6MjMREpOF3YF5HCckyDIopIcVWmbCtOaihOT0VklGjyXGkiYU0R2XVimy6cEsb/9Au/DOO/2nFyauztsceFbX18GGbMH34hjJ/6g82psd6Nv4s/PGPMrqz9lqV16tT0YKkUti3t3ZserMNQY85HqGLfzDYC+4AS0Ovu8+vRKRFprI/akdiX3H1nHT5HRJqAu310jsRE5OhTubD/0XnsyIGnzMyBf3T3xUe+wcwWAYsAxjGhxs2JyMgr1hj7tfb0PHc/A7gYuNbMvnjkG9x9sbvPd/f5bYytcXMiMtIqF/Yt15LFzGaZ2bNmttbM1pjZt5L108zsaTNbn3ydmqw3M/s7M9tgZq+Z2RlZ26gpibn75uTrduBRIB6WQEQKoURLriWHXuAGd58LnE3lYGcucBOwzN3nAMuS76FyQDQnWRYBd2VtYNhJzMw6zKyz7zXwFWD1cD9PRJpDX8V+PY7E3H2Lu69IXu8D1gEnAAuAJcnblgCXJq8XAPd5xQvAFDObGW2jlmtiXcCjVhl3aQzwgLv/tIbPkxFQ7u6uqf3h0z8I41+fHI/pNa6lJzX2i5Z4vLDNP5sVxkv/Ju7b23d0psbKr54Ttj1mdVyrNenVLWF85xdPCOM7/m16QVdXxnScU595IzVmu+tzr24IE4VMN7PqX4LFg10bBzCz2cDpwItAl7v37cStVPIJVBLcpqpm7yTrUnf4sP/F7v4mcNpw24tIc3KHnnLuJLYzT32omU0EHgaud/e9VjXopLt7cnNwWFRiISL9VE4n63d30szaqCSw+939kWT1NjOb6e5bktPF7cn6zUD1IfiJybpUxbmPKiKjppQ8P5m1ZLHKIdc9wDp3v6Mq9DiwMHm9EHisav03kruUZwN7qk47B6UjMRHpp6/Eok7OBa4CVpnZymTdzcBtwENmdjXwNnB5ElsKXAJsAA4Af5a1ASUxETlC/U4n3f15SD1ku2CQ9ztw7VC2oSQmIgNojH0ZXdH0YhlDynxw+dlh/Btzfx7G3+iZEcZPbN+dGvuT418J2/Lv4/j3X/+DML7/zcmpsZaOeL9sPTs+Etm8IP53e088VM/UFel/ei0Lt4Vt9x5OH96otKz2p2Iqdyc/Os9OishRRsNTi0jh6XRSRAqrzncnR5ySmIgMoEERRaSw3I1eJTERKTKdTopIYemamAxdVOc1ws6+8aUw/qWJa2v6/BOCOcT2e3vY9v1SRxj/9tx/CeM7Tk0fiidrctgfro+H6vkgqEEDaO2Nf6Zn//mrqbGvTXs5bPudhz+XGmvx/WHbvJTERKSwVCcmIoWnOjERKSx36M0/KGLDKYmJyAA6nRSRwtI1MREpPFcSE5Ei04V9GZqMMb9G0voPjg3juyZNDONbe6eE8WNa06dV62w5GLad3bYzjO8opdeBAbS2pU8Jd9jj8bL+22f+OYx3f7otjLdZPOXbOePeTY39ydpvhG07eDOM18pd18REpNCMku5OikiR6ZqYiBSWnp0UkWLzhl6mHTIlMREZQHcnRaSwXBf2RaTodDophTFjbHodF8A46wnj7RbPr/huz9TU2PqDnwzb/nZvXMN2UdeaMN4T1IK1BuOcQXad1/Ft74Xxbo/ryKK9em5XXAe2MozWR5HuTmYeM5rZvWa23cxWV62bZmZPm9n65Gv6b6qIFIp7JYnlWZpBnhPfHwEXHbHuJmCZu88BliXfi8hRouyWa2kGmUnM3Z8DjpyLfgGwJHm9BLi0vt0SkUZyz7c0g+FeE+ty9y3J661AV9obzWwRsAhgHBOGuTkRGS2OUS7Q3cmae+ruDulXSd19sbvPd/f5bYytdXMiMgo859IMhpvEtpnZTIDk6/b6dUlEGuoovLA/mMeBhcnrhcBj9emOiDSFAh2KZV4TM7MfA+cD083sHeDbwG3AQ2Z2NfA2cPlIdvKolzHvpLXGY195b3qtVuvUuPrlD6asCuM7SpPC+Pul+DrnlNYDqbF9vePCtrsPxp/9qbFbwviKA7NTYzPa4zqvqN8AGw9PD+Nzxm4N49/ZdkFqbNa4I++j9dd7wRdTY/7iv4Zt82qWo6w8MpOYu1+ZEkr/KYhIYTlQLtcniZnZvcBXge3u/tlk3S3AXwA7krfd7O5Lk9hfA1cDJeAv3f3JrG0U5xaEiIwOB9zyLdl+xMA6U4A73X1esvQlsLnAFcBnkjY/MLP4NAQlMREZRL3qxFLqTNMsAB5090Pu/hawATgzq5GSmIgMlP/C/nQzW161LMq5hevM7LXksca+C7cnAJuq3vNOsi6kB8BF5AhDKp/Y6e7zh7iBu4C/pZIG/xa4HfjzIX7G7+lITEQGGsESC3ff5u4ldy8Dd/PhKeNmYFbVW09M1oV0JNYMMi4u2Jj4xxSVWGy6+tNh2y9PiKcm+3V3fDQ/Y8y+MB4NhzNz7J6wbWdXdxjPKu+YNiZ9mKF9pfFh2wkth8J41r/7jPZ4urm/euaM1FjnZ3eFbSe1Bcce9bip6OB1ujs5GDObWfXY4mVA3wg5jwMPmNkdwPHAHOClrM9TEhORQdStxGKwOtPzzWwelWO5jcA1AO6+xsweAtYCvcC17h4P7IaSmIgMpk7V+Cl1pvcE778VuHUo21ASE5GBmuSRojyUxESkv75i14JQEhORAZplwMM8lMREZKARvDtZb0piIjKA6UhMhsLa2sN4uTuul4pMX3U4jO8sxVOLTWmJh6Rpz5ja7HBQJ3bOtLfCtjsyarlWHDwpjHe2HkyNzWiJ67xmtcW1Wqu6Z4Xxpfs/Ecav/uozqbEfL/6jsG37T3+dGjOPf165NNFYYXkoiYnIEXKPUNEUlMREZCAdiYlIoZUb3YH8lMREpD/ViYlI0enupIgUW4GSmMYTE5FCK9aRWDC1mY2J652sNSNft8TxcncwvlQ5c7SQkPfEtVy1+N4/fj+Mb+qdEsa39sTxrKnNSsGQLi8cnBy2HdfSE8ZnjNkbxveW4zqzyL5yPJ1cNE4aZPf9xmPWp8Ye2fOHYdvRoNNJESkuR48diUjB6UhMRIpMp5MiUmxKYiJSaEpiIlJU5jqdFJGi093J4allfsWsWiuPy3Ya6uCCM8P4pkvjOrQ/PT19ar6tvZ1h21cPzA7jk4MxuQA6MuZn7Pb0+r13D09NjUF2rVU0ryTAsUEdWcnjusDNPXHfsmTVz73TG8yJ+cfxWGdT7htWl4akSEdimRX7ZnavmW03s9VV624xs81mtjJZLhnZborIqBrBGcDrLc9jRz8CLhpk/Z3uPi9Zlta3WyLSMP7hdbGspRlkJjF3fw7YPQp9EZFmcZQdiaW5zsxeS043Uy8gmNkiM1tuZst7iK+fiEhzsHK+pRkMN4ndBZwCzAO2ALenvdHdF7v7fHef38bYYW5ORGRww0pi7r7N3UvuXgbuBuLbayJSLEf76aSZzaz69jJgddp7RaRgCnZhP7NOzMx+DJwPTDezd4BvA+eb2TwquXgjcE09OhPVgdVqzMzjwnjPSV1hfPenJ6TGDhwXFwbOu2RdGP9m1/8O4ztKk8J4m6Xvt009x4RtT5+wMYz/bM/cML5zzMQwHtWZndORPqYWwPvl9H0OcPyY98L4jRu+nhrrmhDXYv3w4/EN9x6PLwi93hNfOtlTTh+P7C/nPhu2fZQZYbwumiRB5ZGZxNz9ykFW3zMCfRGRZnE0JTER+WgxmufOYx5KYiLSXxNd78pDE4WIyEB1ujuZ8tjiNDN72szWJ1+nJuvNzP7OzDYkNahn5OmqkpiIDFS/EosfMfCxxZuAZe4+B1iWfA9wMTAnWRZRqUfNpCQmIgPUq8Qi5bHFBcCS5PUS4NKq9fd5xQvAlCPKuQbVVNfEDl38+TB+7H95MzU2b9I7Ydu5458P493leMq3aFiYtQdPCNseKLeH8fWH4/KPPb1xqUFrcBV2++F4KJ7b34qnB1t25v8K43/z7mBjA3yoZXz6b/quUlye8bWJ8ZRsEP/MrvnYc6mxk9u3h22f2B//7bybMVRPV9ueMD67bUdq7N91/jZsexSUWHS5+5bk9Vagr77pBGBT1fveSdZtIdBUSUxEmoAP6e7kdDNbXvX9YndfnHtT7m5W220EJTERGSh/Wtnp7vOH+OnbzGymu29JThf7Dos3A7Oq3ndisi6ka2IiMsAIP3b0OLAweb0QeKxq/TeSu5RnA3uqTjtT6UhMRAaq0zWxlMcWbwMeMrOrgbeBy5O3LwUuATYAB4A/y7MNJTER6a+OI1SkPLYIcMEg73Xg2qFuQ0lMRPoxilWxryQmIgMoiaWxeFq2s/77y2HzCzrXpMYOeDz0SVYdWFbdT2TymHh6rkM98W7e3hMPtZPl1LFbU2OXTVoZtn3u+2eF8fO6/3MYf+PL8TBCyw6mDzmzozf+d1/x1pfD+IrfzQrjZ89+KzX2uc74pldWbV5na3cYj4ZHAthfTv99faE7rp8bFUpiIlJoSmIiUlgFG8VCSUxEBlISE5Ei06CIIlJoOp0UkeJqounY8lASE5GBlMQG13NsB+9elT7P7i2T/z5s/8Dus1Njs8YdOe5afx9v3xnGTxv/dhiPdLbENUOfnBTXDD2x/8Qw/vP3PxXGZ7a9nxr75YFTwrYP3vI/wvg3/+qGMP6Fpf8hjO+dnT7GQG9H/Jcy6bRdYfxvTv+XMN5updTY+6W4Dmza2P1hfEprXBuYJapr7GxJn+YOoPWTn0iN2cZ43Lw8VLEvIoVn5eJkMSUxEelP18REpOh0OikixaYkJiJFpiMxESk2JTERKayhzXbUcKOaxFp6YMK29L3zxN55YfuTx6fP1bezJ55f8ckPPhfGTxz/Xhif3Jpeu/OJYDwvgJXdU8L4T3d8JowfPz6ef3Fbz+TU2K6ejrDtgWBcK4B77rwjjN++LZ638rJpK1Jjp7XHdWDvl+N5bNZmzNe5rzwuNdbt8fhyezLqyDqD3weAHo//tFo9/e9gSktcg7b3c8ekxkrbav+TLlqdWOZsR2Y2y8yeNbO1ZrbGzL6VrJ9mZk+b2frk6/BHFRSR5uKeb2kCeaZs6wVucPe5wNnAtWY2F7gJWObuc4BlyfcichQY4Snb6iozibn7FndfkbzeB6yjMrX4AmBJ8rYlwKUj1EcRGU0+hKUJDOkE2sxmA6cDLwJdVRNbbgW6UtosAhYBtHfojFOkCIp0YT/3DOBmNhF4GLje3ftdaU7mixs0L7v7Ynef7+7zx4yNLzKLSHOwcr6lGeRKYmbWRiWB3e/ujySrt5nZzCQ+E9g+Ml0UkVHlFOrCfubppJkZcA+wzt2r77c/DiykMiX5QuCxrM9qPVymc9Oh1HjZLWz/s53pQ9J0jdsXtp3XuSmMv34gvl2/6uDxqbEVYz4Wth3f2hPGJ7fHQ/l0jEnfZwDT29L/7SeNjf/fEg1XA/Byd/xv+48zfh7Gf9ebfgnhn/efGrZdeyB9nwNMzZgqb9Xe9PYHetvDtodK8Z9Gd29csjN5bPwz/fy09KGfXmdm2HbHacHwRr8Km+bWLBft88hzTexc4CpglZmtTNbdTCV5PWRmVwNvA5ePSA9FZPQdTUnM3Z+nUv82mAvq2x0RabSiFbvqsSMR6c9dgyKKSMEVJ4cpiYnIQDqdFJHickCnkyJSaMXJYaOcxD44SMsvXk0N/9NT54bN/+uCf0qN/SJjWrMntsZ1PXsPx0PSzJiQPoXXpKBOC2BaWzz91+SMeqdxFk/59l5v+pMQh1riIWdKqTeeK7YeSh/mB+BX5TlhvKfcmho7FMQgu75u9+HpYfz48XtSY/t604fpAdi4b1oY37lnYhjvnhD/aT1fSp9K76Lj1oRtx29P/5m1xL8quel0UkQKrZ53J81sI7APKAG97j7fzKYB/weYDWwELnf3eFC/FLmfnRSRj4iRGcXiS+4+z93nJ9/XbSgvJTER6adS7Oq5lhrUbSgvJTERGaicc4HpZra8alk0yKc58JSZvVIVzzWUVx66JiYiAwzhKGtn1SlimvPcfbOZHQs8bWb/rzro7m42/FsJOhITkf7qfE3M3TcnX7cDjwJnUsehvJTEROQIlWcn8yxZzKzDzDr7XgNfAVbz4VBekHMorzRNdTp58o3/GsZ/8NrX09v+p9fDthcftzqMr9gbj5v1u6Bu6DfBWGMAbS3xEJgT2g6H8XEZ9VLtreljgrVk/O+ynFEn1tEa9y1rrLNpY9Nr5Dpb4zG3WmocOrQ1+Le/tGd22LZrQlz794lJO8N4r8fHB1+Y/EZq7N63zgnbdv39r1NjGz2uScytfgMedgGPVoYlZAzwgLv/1Mxepk5DeTVVEhORJlDHyXPd/U3gtEHW76JOQ3kpiYnIQE0y9HQeSmIiMlBxcpiSmIgMZOUmmcooByUxEenP6StkLQQlMRHpx6j5kaJRpSQmIgMpiQVagjGkyvEciJPvfyE1tuv+eLM/+dqFYfysm18O41+d/ZvU2Kfat4Vt2zKOzcdl3M/uaIlrubqDX7isaubnD84K46WMT/jZe58O4+/3jE+NbTswKWzbFtS/5RHNY3qwNx5nbc/BeLyx1pb4j7z75/FYZ2+tTR//bvLS+HdxVCiJiUhh6ZqYiBSd7k6KSIG5TidFpMAcJTERKbjinE0qiYnIQKoTE5FiO5qSmJnNAu6jMi6QA4vd/XtmdgvwF8CO5K03u/vSzC1m1IKNlI6HXwzjqx+O26/mpNSYff6Pw7YHj0uvlQIYuysek2vfx+P2k95IH0Oq5VA8EWH5N+vCeLYPami7N4zGo6jVpj0jPqPmLfy25k9oGHcoFed8Ms+RWC9wg7uvSEZofMXMnk5id7r7d0eueyLSEEfTkVgyI8mW5PU+M1sHnDDSHRORBipQEhvSGPtmNhs4Heg7N7vOzF4zs3vNbGpKm0V90zn1EJ82iUgTcKDs+ZYmkDuJmdlE4GHgenffC9wFnALMo3Kkdvtg7dx9sbvPd/f5bYytvcciMsIcvJxvaQK57k6aWRuVBHa/uz8C4O7bquJ3A0+MSA9FZHQ5hbqwn3kkZpVpSu4B1rn7HVXrZ1a97TIq0zCJyNHAPd/SBPIciZ0LXAWsMrOVybqbgSvNbB6VvL0RuGYE+lcI/vKqMB4P6pJtUvoMXZmK8/9TaSpNkqDyyHN38nkYdHLC7JowESmg5jnKykMV+yLSnwMaikdECk1HYiJSXEffY0ci8lHi4E1SA5aHkpiIDNQk1fh5KImJyEC6JiYiheWuu5MiUnA6EhOR4nK81JjBS4dDSUxE+usbiqcglMREZKAClVgMaVBEETn6OeBlz7XkYWYXmdnrZrbBzG6qd3+VxESkP6/foIhm1gr8A3AxMJfK6Ddz69ldnU6KyAB1vLB/JrDB3d8EMLMHgQXA2nptYFST2D7e2/mM/+TtqlXTgZ2j2YchaNa+NWu/QH0brnr27eO1fsA+3nvyGf/J9JxvH2dmy6u+X+zui6u+PwHYVPX9O8BZtfax2qgmMXfvN52fmS139/mj2Ye8mrVvzdovUN+Gq9n65u4XNboPQ6FrYiIykjYDs6q+PzFZVzdKYiIykl4G5pjZSWbWDlwBPF7PDTT6wv7i7Lc0TLP2rVn7BerbcDVz32ri7r1mdh3wJNAK3Ovua+q5DfMCPSMlInIknU6KSKEpiYlIoTUkiY30Ywi1MLONZrbKzFYeUf/SiL7ca2bbzWx11bppZva0ma1Pvk5tor7dYmabk3230swuaVDfZpnZs2a21szWmNm3kvUN3XdBv5pivxXVqF8TSx5D+C3wR1QK314GrnT3ulXw1sLMNgLz3b3hhZFm9kXgA+A+d/9ssu47wG53vy35H8BUd7+xSfp2C/CBu393tPtzRN9mAjPdfYWZdQKvAJcC36SB+y7o1+U0wX4rqkYcif3+MQR3Pwz0PYYgR3D354DdR6xeACxJXi+h8kcw6lL61hTcfYu7r0he7wPWUakcb+i+C/olNWhEEhvsMYRm+kE68JSZvWJmixrdmUF0ufuW5PVWoKuRnRnEdWb2WnK62ZBT3WpmNhs4HXiRJtp3R/QLmmy/FYku7A90nrufQeWp+2uT06am5JVrAc1UI3MXcAowD9gC3N7IzpjZROBh4Hp331sda+S+G6RfTbXfiqYRSWzEH0OohbtvTr5uBx6lcvrbTLYl11b6rrFsb3B/fs/dt7l7ySuTFt5NA/edmbVRSRT3u/sjyeqG77vB+tVM+62IGpHERvwxhOEys47kgitm1gF8BVgdtxp1jwMLk9cLgcca2Jd++hJE4jIatO/MzIB7gHXufkdVqKH7Lq1fzbLfiqohFfvJLeT/yYePIdw66p0YhJmdTOXoCyqPZD3QyL6Z2Y+B86kM1bIN+Dbwf4GHgI8BbwOXu/uoX2BP6dv5VE6JHNgIXFN1DWo0+3Ye8EtgFdA3ct/NVK4/NWzfBf26kibYb0Wlx45EpNB0YV9ECk1JTEQKTUlMRApNSUxECk1JTEQKTUlMRApNSUxECu3/A4BqExKmbJN8AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(original_train_x[0])\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "mechanical-emission",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "original_train_x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "statewide-young",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('uint8')"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "original_train_x.dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "running-recycling",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28, 1)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t = original_train_x.reshape(60000, 28, 28, 1)\n",
    "t.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "australian-praise",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('uint8')"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t.dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "educated-airport",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('uint8')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t.astype('float32')\n",
    "t.dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "fitted-primary",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dtype('float32')"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t = t.astype('float32')\n",
    "t.dtype"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "suburban-satellite",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd33eb1cb50>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(original_train_x[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "peaceful-supply",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd33eaa8890>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(t[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "blank-christmas",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28, 1)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "t.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "vulnerable-cowboy",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd33e8a1e10>"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(t[0], cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "reserved-princeton",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(t[0], cmap='gray')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "nominated-making",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0\n",
      "0.0\n"
     ]
    }
   ],
   "source": [
    "tt = t / 127.5\n",
    "print(tt.max())\n",
    "print(tt.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "promotional-roommate",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(tt[0])\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "yellow-objective",
   "metadata": {},
   "source": [
    "> #### 결과\n",
    "> - 3차원으로 `plt.imshow()`에 넣어도 오류나지 않는다.\n",
    "> - 값 스케일이 바뀌어도 `plt.imshow()`로서 이미지가 잘 나타난다."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "passing-theorem",
   "metadata": {},
   "source": [
    "----------------------\n",
    "----------------------\n",
    "----------------------"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "standard-research",
   "metadata": {},
   "source": [
    "> 데이터 하나만 꺼내서 확인하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "apparent-quantity",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(train_x[0].reshape(28, 28), cmap='gray')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dental-master",
   "metadata": {},
   "source": [
    "> 위의 컬러바를 통해 실제 값과 컬러의 분포를 확인할 수 있음  \n",
    "> -1 ~ 1까지"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "optional-boulder",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 10 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 10개 확인\n",
    "plt.figure(figsize=(10, 5))\n",
    "\n",
    "for i in range(10):\n",
    "    plt.subplot(2, 5, i+1)\n",
    "    plt.imshow(train_x[i].reshape(28, 28), cmap='gray')\n",
    "    plt.title(f'index: {i}')\n",
    "    plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "illegal-relief",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x864 with 25 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 25개 랜덤으로 확인\n",
    "\n",
    "plt.figure(figsize=(10, 12))\n",
    "\n",
    "for i in range(25):\n",
    "    plt.subplot(5, 5, i+1)\n",
    "    random_index = np.random.randint(1, 60000)\n",
    "    plt.imshow(train_x[random_index].reshape(28, 28), cmap='gray')\n",
    "    plt.title(f'index: {random_index}')\n",
    "    plt.axis('off')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "accessible-intersection",
   "metadata": {},
   "source": [
    "> tf.data.Dataset 화 시켜서 매번 우리가 직접 섞지 않아도 되게끔 해주기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "recognized-firmware",
   "metadata": {},
   "outputs": [],
   "source": [
    "BUFFER_SIZE = 60000 # 전체 다 섞어주기\n",
    "BATCH_SIZE = 256 # 미니배치 학습"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "japanese-woman",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = tf.data.Dataset.from_tensor_slices(train_x).shuffle(BUFFER_SIZE).batch(BATCH_SIZE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "upset-issue",
   "metadata": {},
   "source": [
    "## 그림을 만들어내는 화가 생성자, 그리고 평가하는 비평가 판별자 (1) GAN 이해하기"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "suburban-school",
   "metadata": {},
   "source": [
    "- 이후 소개하는 코드는 Tensorflow 2.0 Tutorial의 DCGAN 구현을 기반으로 상세한 설명을 추가함"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "irish-australian",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_generator_model():\n",
    "\n",
    "    # Start\n",
    "    model = tf.keras.Sequential()\n",
    "\n",
    "    # First: Dense layer\n",
    "    model.add(layers.Dense(7*7*256, use_bias=False, input_shape=(100,)))\n",
    "    model.add(layers.BatchNormalization())\n",
    "    model.add(layers.LeakyReLU())\n",
    "\n",
    "    # Second: Reshape layer\n",
    "    model.add(layers.Reshape((7, 7, 256)))\n",
    "\n",
    "    # Third: Conv2DTranspose layer\n",
    "    model.add(layers.Conv2DTranspose(128, kernel_size=(5, 5), strides=(1, 1), padding='same', use_bias=False))\n",
    "    model.add(layers.BatchNormalization())\n",
    "    model.add(layers.LeakyReLU())\n",
    "\n",
    "    # Fourth: Conv2DTranspose layer\n",
    "    model.add(layers.Conv2DTranspose(64, kernel_size=(5, 5), strides=(2, 2), padding='same', use_bias=False))\n",
    "    model.add(layers.BatchNormalization())\n",
    "    model.add(layers.LeakyReLU())\n",
    "\n",
    "    # Fifth: Conv2DTranspose layer\n",
    "    model.add(layers.Conv2DTranspose(1, kernel_size=(5, 5), strides=(2, 2), padding='same', use_bias=False, \\\n",
    "                                     activation='tanh'))\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "headed-reproduction",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "\n",
    "`make_generator_model`이라는 함수를 만들어서 언제든 생성자를 생성할 수 있도록 했습니다.\n",
    "\n",
    "함수 내부에서는 먼저 `tf.keras.Sequential()`로 모델을 시작한 후 레이어를 차곡차곡 쌓아줍니다.\n",
    "\n",
    "여기에서 가장 중요한 레이어는 바로 `Conv2DTranspose` 레이어입니다. `Conv2DTranspose` 층은 일반적인 Conv2D와 반대로 이미지 사이즈를 넓혀주는 층입니다. 이 모델에서는 세 번의 `Conv2DTranspose` 층을 이용해 `(7, 7, 256)` → `(14, 14, 64)` → `(28, 28, 1)` 순으로 이미지를 키워나갑니다. 여기서 최종 사이즈인 `(28, 28, 1)`은 우리가 준비했던 데이터셋과 형상이 동일합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "lesser-soundtrack",
   "metadata": {},
   "source": [
    "레이어의 사이사이에 특정 층들이 반복되는 것을 확인할 수 있는데, `BatchNormalization` 레이어는 신경망의 가중치가 폭발하지 않도록 가중치 값을 정규화시켜줍니다. 또한 중간층들의 활성화 함수는 모두 `LeakyReLU`를 사용하였습니다. 다만 마지막 층에는 활성화 함수로 `tanh`를 사용하는데, 이는 우리가 `-1 ~ 1` 이내의 값으로 픽셀 값을 정규화시켰던 데이터셋과 동일하게 하기 위함입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "worst-detroit",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 12544)             1254400   \n",
      "_________________________________________________________________\n",
      "batch_normalization (BatchNo (None, 12544)             50176     \n",
      "_________________________________________________________________\n",
      "leaky_re_lu (LeakyReLU)      (None, 12544)             0         \n",
      "_________________________________________________________________\n",
      "reshape (Reshape)            (None, 7, 7, 256)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_transpose (Conv2DTran (None, 7, 7, 128)         819200    \n",
      "_________________________________________________________________\n",
      "batch_normalization_1 (Batch (None, 7, 7, 128)         512       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_transpose_1 (Conv2DTr (None, 14, 14, 64)        204800    \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 14, 14, 64)        256       \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_2 (LeakyReLU)    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_transpose_2 (Conv2DTr (None, 28, 28, 1)         1600      \n",
      "=================================================================\n",
      "Total params: 2,330,944\n",
      "Trainable params: 2,305,472\n",
      "Non-trainable params: 25,472\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# generator 모델 생성\n",
    "generator = make_generator_model()\n",
    "\n",
    "generator.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "immune-bandwidth",
   "metadata": {},
   "source": [
    "> (1, 100) 형태를 갖는 열벡터를 만들어서 인풋으로 넣어줄 예정  \n",
    "> 아직 훈련되지 않아서 별 의미없는 output이 나올 것임"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "generous-software",
   "metadata": {},
   "outputs": [],
   "source": [
    "noise = tf.random.normal([1, 100])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "polyphonic-output",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "텐서플로우 2.0 이후 버전에서는 레이어와 모델에 `call` 메소드를 구현해 놓기 때문에, 방금 만들어진 생성자 모델에 입력값으로 노이즈를 넣고 바로 모델을 호출하면 간단히 결과 이미지가 생성됩니다 (내부적으로는 생성자의 `call` 함수가 호출됩니다).\n",
    "\n",
    "단, 지금은 학습하는 중이 아니니 `training=False`를 설정해 주어야 합니다! `Batch Normalization` 레이어는 훈련 시기와 추론(infernce) 시기의 행동이 다르기 때문에 `training=False`을 주어야 올바른 결과를 얻을 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "subtle-consistency",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 100), dtype=float32, numpy=\n",
       "array([[ 3.96854639e-01, -9.03411031e-01, -1.03277966e-01,\n",
       "        -1.33564128e-02,  1.02616894e+00, -1.47861993e+00,\n",
       "         3.29661757e-01, -1.76281497e-01, -9.49293792e-01,\n",
       "         4.59632784e-01,  7.27090836e-01, -3.15581769e-01,\n",
       "         1.91859746e+00,  5.37751615e-01,  2.14032784e-01,\n",
       "         1.17724679e-01,  3.75476182e-01,  4.61798199e-02,\n",
       "         1.34195030e+00, -2.33857036e-01,  1.39860439e+00,\n",
       "        -2.01601014e-01,  7.66258061e-01, -3.12837660e-01,\n",
       "         1.42178297e+00,  2.79477268e-01,  1.56311941e+00,\n",
       "         1.09847806e-01,  1.99261144e-01,  1.65953999e-03,\n",
       "        -1.13957787e+00,  1.91445518e+00, -1.36811960e+00,\n",
       "        -4.50607866e-01,  4.37245041e-01,  7.96103597e-01,\n",
       "        -6.59848750e-01, -9.80238736e-01, -4.69153263e-02,\n",
       "        -7.39774466e-01, -1.93865955e-01,  4.70025033e-01,\n",
       "         1.86829254e-01,  3.37503016e-01,  9.14283752e-01,\n",
       "         1.20782495e-01, -2.56878942e-01, -1.57627368e+00,\n",
       "        -4.82582510e-01, -2.38460749e-01,  5.33107400e-01,\n",
       "         9.21795785e-01, -1.02051401e+00,  1.24347154e-02,\n",
       "        -1.09277022e+00,  7.22795844e-01, -1.54712105e+00,\n",
       "        -5.34086704e-01,  1.01934874e+00,  1.00953653e-01,\n",
       "         3.74880340e-03,  1.35820127e+00,  5.07589519e-01,\n",
       "        -1.09337103e+00, -1.72534752e+00,  1.25932908e+00,\n",
       "         1.98038924e+00, -5.66969104e-02,  2.27832031e+00,\n",
       "        -1.36630133e-01,  4.08770777e-02,  6.16267443e-01,\n",
       "         5.22248000e-02,  5.05880177e-01, -1.55623519e+00,\n",
       "        -4.75246429e-01,  3.07505548e-01,  1.11799061e-01,\n",
       "        -1.05705678e+00, -5.70644200e-01,  1.17127144e+00,\n",
       "        -5.92848957e-01, -4.14751261e-01, -4.64318812e-01,\n",
       "         3.75705391e-01, -6.53238297e-02, -3.00637241e-02,\n",
       "         1.31952727e+00,  1.17409670e+00, -8.27786982e-01,\n",
       "        -2.95131302e+00, -1.26418102e+00, -3.75436127e-01,\n",
       "         7.58018136e-01,  9.52141881e-01,  1.22645962e+00,\n",
       "         8.70974600e-01,  8.43034983e-01,  1.29235470e+00,\n",
       "         4.28438652e-03]], dtype=float32)>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "promising-immigration",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 28, 28, 1])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generated_image = generator(noise, training=False)\n",
    "generated_image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "engaged-burning",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(generated_image[0, :, :, 0], cmap='gray')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "roman-imaging",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([28, 28, 1])"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "generated_image[0, :, :, :].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "insured-davis",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(generated_image[0, :, :, :], cmap='gray')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fewer-placement",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "역시 -1과 1 사이의 값에서 적당히 잘 생성된 것을 확인할 수 있습니다.\n",
    "\n",
    "아직은 모델이 전혀 학습하지 않은 상태이기 때문에 아무런 의미가 없는 노이즈 같은 이미지가 나타나..."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "animated-direction",
   "metadata": {},
   "source": [
    "## 그림을 만들어내는 화가 생성자, 그리고 평가하는 비평가 판별자 (3) 판별자 구현하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "alike-positive",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_discriminator_model():\n",
    "\n",
    "    # Start\n",
    "    model = tf.keras.Sequential()\n",
    "\n",
    "    # First: Conv2D Layer\n",
    "    model.add(layers.Conv2D(64, (5, 5), strides=(2, 2), padding='same', input_shape=[28, 28, 1]))\n",
    "    model.add(layers.LeakyReLU())\n",
    "    model.add(layers.Dropout(0.3))\n",
    "\n",
    "    # Second: Conv2D Layer\n",
    "    model.add(layers.Conv2D(128, (5, 5), strides=(2, 2), padding='same'))\n",
    "    model.add(layers.LeakyReLU())\n",
    "    model.add(layers.Dropout(0.3))\n",
    "\n",
    "    # Third: Flatten Layer\n",
    "    model.add(layers.Flatten())\n",
    "\n",
    "    # Fourth: Dense Layer\n",
    "    model.add(layers.Dense(1))\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cross-admission",
   "metadata": {},
   "source": [
    "판별자 또한 `make_discriminator_model` 함수로 구현하였습니다.\n",
    "\n",
    "`Conv2DTranspose` 층을 사용해서 이미지를 키워나갔던 생성자와 반대로, 판별자는 Conv2D 층으로 이미지의 크기를 점점 줄여나갑니다.\n",
    "첫 번째 Conv2D 층에서 입력된 `[28, 28, 1]` 사이즈의 이미지는 다음 층을 거치며 `(28, 28, 1) → (14, 14, 64) → (7, 7, 128)`까지 줄어들게 됩니다.\n",
    "\n",
    "마지막에는 `Flatten` 층을 사용해 3차원 이미지를 1차원으로 쭉 펴서 `7x7x128=627`2, 즉 `(1, 6272)` 형상의 벡터로 변환합니다. 이는 생성자의 `Reshape` 층에서 1차원 벡터를 3차원으로 변환했던 것과 정확히 반대 역할을 합니다. 1차원 벡터로 변환한 후에는 마지막 `Dense Layer`를 거쳐 단 하나의 값을 출력하게 됩니다.\n",
    "\n",
    "그럼 판별 모델을 `discriminator`라는 변수 이름으로 생성하고, 모델 세부 내용인 `summary`를 출력해 봅시다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "registered-entity",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d (Conv2D)              (None, 14, 14, 64)        1664      \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_3 (LeakyReLU)    (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 14, 14, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 7, 7, 128)         204928    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_4 (LeakyReLU)    (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 7, 7, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 6272)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 6273      \n",
      "=================================================================\n",
      "Total params: 212,865\n",
      "Trainable params: 212,865\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "discriminator = make_discriminator_model()\n",
    "\n",
    "discriminator.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "moderate-ebony",
   "metadata": {},
   "source": [
    "> - 아까 생성한 이미지를 넣으면 어떤 값이 나오는지 확인  \n",
    "> - 물론 아직은 의미 없는 값이 나올 것임"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "straight-harris",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 1), dtype=float32, numpy=array([[-0.00516815]], dtype=float32)>"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "decision = discriminator(generated_image, training=False)\n",
    "decision"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "congressional-fabric",
   "metadata": {},
   "source": [
    "----------------\n",
    "----------------\n",
    "----------------"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "illegal-break",
   "metadata": {},
   "source": [
    "#### 실험\n",
    "> - `padding='same'` 의 원리를 알고 싶다.\n",
    "> - `strides=1`일 때는 input과 동일한 shape이 나오는 걸 알겠는데,\n",
    "> - `strides=2`이상 되면 `same`이 무슨 의미일까?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "victorian-professor",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 14, 14, 32])"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=5, strides=2, padding='same')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "czech-conversion",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 12, 12, 32])"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=5, strides=2, padding='valid')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "divine-browse",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 26, 26, 32])"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=3, strides=1, padding='valid')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "thick-refund",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 13, 13, 32])"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=3, strides=2, padding='valid')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "connected-european",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 14, 14, 32])"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=3, strides=2, padding='same')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "corresponding-circle",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 10, 10, 32])"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=3, strides=3, padding='same')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "loved-pricing",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([1, 9, 9, 32])"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ten = tf.zeros([1, 28, 28, 1])\n",
    "\n",
    "con1 = tf.keras.layers.Conv2D(32, kernel_size=3, strides=3, padding='valid')\n",
    "out = con1(ten)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "unexpected-insured",
   "metadata": {},
   "source": [
    "#### feedback\n",
    "> - 결론적으로 `same`을 하면 strides가 2이상일 때 9, 11 같은 output shape을 10, 12 같이 even하게 만들어주는 것 같다."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "banner-hungary",
   "metadata": {},
   "source": [
    "## 생성 모델이 똑똑해지기 위한 기나긴 여정 (1) 손실함수와 최적화 함수"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "romance-exercise",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "GAN은 손실함수로 교차 엔트로피(Cross Entropy)를 사용합니다.\n",
    "교차 엔트로피는 앞서 분류 모델을 설계할 때 많이 사용했던 손실함수이죠. 점점 가까워지기 원하는 두 값이 얼마나 큰 차이가 나는지를 정량적으로 계산할 때 많이 쓰입니다. 특히 판별자는 한 개의 이미지가 가짜인지 진짜인지 나타내는 2개 클래스 간 분류 문제를 풀어야 하므로, 이진 교차 엔트로피`(binary cross entropy)`를 사용할 것입니다.\n",
    "\n",
    "자, 이쯤에서 다시 생성자와 판별자의 목적이 무엇이었는지 한번 정리해 볼까요? 생성자는 자기가 생성한 `Fake Image`를 판별자가 진짜라고 판단하도록 속이고 싶어 합니다. 판별자는 생성자가 생성한 가짜 이미지를 가짜로, 진짜 이미지를 진짜로 잘 판별하는 것이 목표입니다.\n",
    "\n",
    "좋습니다!  \n",
    "그렇다면 `Real Image`에 대한 라벨을 `1`, `Fake Image`에 대한 라벨을 `0`으로 두었을 때, 각각의 손실함수를 이용해 정량적으로 달성해야 하는 목표하는 결과는 다음과 같겠죠.\n",
    "\n",
    "`생성자` : `판별자`가 Fake Image에 대해 판별한 값, 즉 `D(fake_image)` 값이 `1`에 가까워지는 것  \n",
    "`판별자` : Real Image 판별값, 즉 `D(real_image)`는 `1`에, Fake Image 판별값, 즉 `D(fake_image)`는 `0`에 가까워지는 것  \n",
    "결국 생성자든 구분자든, 결국 손실함수에 들어가는 값은 모두 판별자의 판별값이 되겠네요! 이러한 기준을 가지고 생성자, 판별자 각각에 대한 손실함수를 설계해 봅시다.\n",
    "\n",
    "손실함수에 사용할 교차 엔트로피 함수는 `tf.keras.losses` 라이브러리 안에 있습니다.\n",
    "\n",
    "다만, 우리가 교차 엔트로피를 계산하기 위해 입력할 값은 판별자가 판별한 값인데, 판별자 모델의 맨 마지막 Layer에는 값을 정규화시키는 sigmoid나 tanh 함수와 같은 활성화 함수가 없었습니다. 즉, 구분자가 출력하는 값은 범위가 정해지지 않아 모든 실숫값을 가질 수 있죠.\n",
    "\n",
    "그런데 tf.keras.losses의 `BinaryCrossEntropy` 클래스는 기본적으로 본인에게 들어오는 인풋값이 0-1 사이에 분포하는 확률값이라고 가정합니다. 따라서 `from_logits`를 `True`로 설정해 주어야 BinaryCrossEntropy에 입력된 값을 함수 내부에서 sigmoid 함수를 사용해 0~1 사이의 값으로 정규화한 후 알맞게 계산할 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "stylish-disposition",
   "metadata": {},
   "outputs": [],
   "source": [
    "cross_entropy = tf.keras.losses.BinaryCrossentropy(from_logits=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "disturbed-operation",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "\n",
    "그럼, 이제 cross_entropy를 활용해 계산할 loss들은 `fake_output`와 `real_output`, 두 가지를 활용합니다.\n",
    "\n",
    "`fake_output` : 생성자가 생성한 Fake Image를 구분자에 입력시켜서 판별된 값, 즉 `D(fake_image)`\n",
    "`real_output` : 기존에 있던 Real Image를 구분자에 입력시켜서 판별된 값, 즉 `D(real_image)`\n",
    "그럼 이제, fake_output과 real_output을 각각 1 또는 0에 비교를 해야 하는데, 어떻게 비교할 수 있을까요?\n",
    "\n",
    "바로 `tf.ones_like()`와 `tf.zeros_like()` 함수를 활용하면 됩니다.\n",
    "이 함수들은 특정 벡터와 동일한 크기이면서 값은 1 또는 0으로 가득 채워진 벡터를 만들고 싶을 때 사용합니다.\n",
    "어떤 식으로 활용되는지, 한번 실험해 볼까요?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "regulated-flood",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(2, 3), dtype=int32, numpy=\n",
       "array([[1, 1, 1],\n",
       "       [1, 1, 1]], dtype=int32)>"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vector = [[1, 2, 3],\n",
    "          [4, 5, 6]]\n",
    "\n",
    "tf.ones_like(vector)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fluid-mustang",
   "metadata": {},
   "source": [
    "### generator_loss"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "prompt-union",
   "metadata": {},
   "source": [
    "그러면 `generator_loss`는 다음과 같이 구현할 수 있겠군요.\n",
    "\n",
    "generator_loss는 fake_output가 1에 가까워지기를 바라므로, 다음과 같이 `tf.ones_like`와의 교차 엔트로피값을 계산하면 됩니다.\n",
    "즉, `cross_entropy(tf.ones_like(fake_output), fake_output)` 값은 fake_output이 (Real Image를 의미하는) 1에 가까울수록 `작은 값`을 가집니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "excited-handbook",
   "metadata": {},
   "outputs": [],
   "source": [
    "def generator_loss(fake_output):\n",
    "    return cross_entropy(tf.ones_like(fake_output), fake_output)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "closed-myrtle",
   "metadata": {},
   "source": [
    "### discriminator_loss"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "personalized-gazette",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "\n",
    "반면, `discriminator_loss`는 `real_output` 값은 `1`에 가까워지기를, `fake_output` 값은 `0`에 가까워지기를 바라므로, 두 가지 loss값을 모두 계산합니다. real_output은 1로 채워진 벡터와, fake_output은 0으로 채워진 벡터와 비교하면 되겠죠!\n",
    "\n",
    "최종 `discriminator_loss 값은 이 둘을 더한 값`입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "imperial-model",
   "metadata": {},
   "outputs": [],
   "source": [
    "def discriminator_loss(real_output, fake_output):\n",
    "    real_loss = cross_entropy(tf.ones_like(real_output), real_output)\n",
    "    fake_loss = cross_entropy(tf.zeros_like(fake_output), fake_output)\n",
    "    total_loss = real_loss + fake_loss\n",
    "    return total_loss"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "stupid-flavor",
   "metadata": {},
   "source": [
    "### discriminator accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "everyday-bryan",
   "metadata": {},
   "source": [
    "#### 아이펠 설명\n",
    "\n",
    "한편, 판별자가 real output, fake output을 얼마나 정확히 판별하는지의 accuracy를 계산해 보는 것도 GAN의 학습에서 매우 중요합니다. 특히 두 `accuracy`를 따로 계산해서 비교해 보는 것이 매우 유용합니다.\n",
    "\n",
    "만약 판별자가 real output과 fake output을 accuracy가 `1.0`에 가까울 정도로 정확하게 판별해 낸다면 좋은 것일까요? 그렇지 않습니다. 이 경우 생성자가 만들어내는 fake output이 real output과 차이가 많이 나기 때문에 판별자가 매우 `쉽게 판별`해 내고 있다는 뜻입니다. 그래서, real accuracy와 fake accuracy는 초반에는 1.0에 가깝게 나오다가, 서서히 낮아져서 둘 다 `0.5`에 가까워지는 것이 이상적입니다. fake accuracy가 1.0에 더 가깝다면 아직은 생성자가 판별자를 충분히 잘 속이지 못하고 있다는 뜻입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "annual-choice",
   "metadata": {},
   "outputs": [],
   "source": [
    "def discriminator_accuracy(real_output, fake_output):\n",
    "    real_accuracy = tf.reduce_mean(tf.cast(tf.math.greater_equal(real_output, tf.constant([0.5])), tf.float32))\n",
    "    fake_accuracy = tf.reduce_mean(tf.cast(tf.math.less(fake_output, tf.constant([0.5])), tf.float32))\n",
    "    return real_accuracy, fake_accuracy"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "downtown-bacteria",
   "metadata": {},
   "source": [
    "위 함수 안에 사용된 tensorflow 함수들의 역할을 순차적으로 예를 들어 정리하면 다음과 같습니다.\n",
    "\n",
    "ex.) `real_output = tf.Tensor([0.2, 0.4, 0.7, 0.9])` 라면,\n",
    "\n",
    "- (1) `tf.math.greater_equal(real_output, tf.constant([0.5])` : real_output의 각 원소가 0.5 이상인지 True, False로 판별 - >> `tf.Tensor([False, False, True, True])`\n",
    "- (2) `tf.cast( (1), tf.float32)` : (1)의 결과가 `True이면 1.0`, `False이면 0.0으로 변환` - >> `tf.Tensor([0.0, 0.0, 1.0, 1.0])`\n",
    "- (3) `tf.reduce_mean( (2))` : (2)의 결과를 평균내어 이번 배치의 정확도(accuracy)를 계산 - >> `0.5`"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "portuguese-choice",
   "metadata": {},
   "source": [
    "### 최적화 함수(optimizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "creative-profession",
   "metadata": {},
   "source": [
    "중요한 점 한 가지는 생성자와 구분자는 따로따로 학습을 진행하는 개별 네트워크이기 때문에 optimizer를 따로 만들어주어야 한다는 점입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "centered-acting",
   "metadata": {},
   "outputs": [],
   "source": [
    "generator_optimizer = tf.keras.optimizers.Adam(1e-4)\n",
    "discriminator_optimizer = tf.keras.optimizers.Adam(1e-4)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "hungry-judge",
   "metadata": {},
   "source": [
    "또한, 매번 학습이 어떻게 진행되어가고 있는지를 확인하기 위해 생성자가 생성한 샘플을 확인할 것입니다.\n",
    "\n",
    "샘플은 한 번에 16장을 생성하도록 하겠습니다.  \n",
    "생성할 샘플은 매번 같은 노이즈로 생성해야 그에 대한 진전 과정을 확인할 수 있으므로, 고정된 `seed` 노이즈를 만들어두어야 합니다.\n",
    "즉, 100차원의 노이즈를 총 16개, `(16, 100)` 형상의 벡터를 만들어 두도록 하겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "muslim-homeless",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([16, 100])"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "noise_dim = 100\n",
    "num_examples_to_generate = 16\n",
    "\n",
    "seed = tf.random.normal([num_examples_to_generate, noise_dim])\n",
    "seed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "unknown-poultry",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(1, 100), dtype=float32, numpy=\n",
       "array([[ 1.2627834 ,  0.38518155, -0.9515627 ,  0.0875157 , -0.32265577,\n",
       "         0.25068024, -0.11028544, -0.24491279,  0.6183259 ,  0.83429825,\n",
       "         0.7682573 , -1.4271178 , -2.0610085 ,  1.0442251 ,  2.1625097 ,\n",
       "        -0.26752424,  1.4999356 , -0.22690766,  0.9004858 ,  0.6464795 ,\n",
       "         1.3431352 , -0.8213151 , -0.23106466, -1.1284685 ,  0.6144815 ,\n",
       "        -0.06441844, -0.03768314,  0.5690817 ,  0.04751236, -0.16303144,\n",
       "        -1.355594  , -0.5115844 , -0.31575105, -2.0118873 ,  1.3598229 ,\n",
       "         1.1545432 , -0.70045036,  0.1300467 ,  0.74650246,  0.5513549 ,\n",
       "        -0.02698994, -0.77925885,  0.8454625 ,  2.1379788 ,  1.5662848 ,\n",
       "         0.64801633, -1.4340749 , -0.03694333,  0.25454614, -0.3965547 ,\n",
       "         0.23265162,  1.7766    , -0.74997395,  1.1660575 , -0.831553  ,\n",
       "         1.2857306 ,  0.680116  ,  0.35116115,  1.0427655 , -0.1990515 ,\n",
       "        -0.49894363,  0.6006104 , -0.21008171,  1.0238844 , -0.30149165,\n",
       "         0.76349276, -1.9857934 , -0.4069542 , -1.2116657 ,  0.38255537,\n",
       "         0.8625307 ,  1.0955262 ,  2.1060526 ,  0.36560544,  0.21076408,\n",
       "        -0.3212393 ,  1.2680686 , -0.22836526, -0.63706243,  0.53120774,\n",
       "         0.3869012 ,  1.4587908 ,  0.07715912,  2.3357635 ,  0.08761414,\n",
       "        -1.0183656 ,  0.55281466, -0.78365904, -1.4399743 ,  0.7417362 ,\n",
       "        -1.561811  ,  0.09942132,  1.2588651 , -0.80192184,  0.94536936,\n",
       "         0.58854973, -1.2719394 ,  0.2650859 , -1.4812465 ,  0.26207373]],\n",
       "      dtype=float32)>"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seed[:1, :]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "streaming-rehabilitation",
   "metadata": {},
   "source": [
    "##  생성 모델이 똑똑해지기 위한 기나긴 여정 (2) 훈련과정 설계"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "italian-century",
   "metadata": {},
   "source": [
    "그렇다면 이제 진짜 훈련을 위한 코드를 만들어 보도록 합시다.\n",
    "\n",
    "하나의 미니 배치 당 진행할 `train_step` 함수를 먼저 만들어야 합니다.\n",
    "텐서플로우 2.0 이후부터는 1.x의 이전 버전과 다르게, session을 사용하지 않습니다. 대신, 학습시킬 훈련 함수 위에 `@tf.function` 이라는 데코레이터를 붙여서 사용합니다. 이 데코레이터는 우리가 직접 session을 열어서 학습했다가, 학습이 완료되면 다시 닫아주는 등의 번거로운 과정을 내부적으로 처리해서 더욱 편리하게 학습시킬 수 있게 해줍니다.\n",
    "\n",
    "데코레이터의 개념은 어렵지 않으시더라도 `@tf.function` 이 정확히 어떤 역할을 해주는 것인지 명확히 이해가 안 되실 수도 있을 것 같습니다. Tensorflow Tutorial에는 이 데코레이터의 역할을 다음과 같이 정의하고 있습니다.\n",
    "\n",
    "> tf.function : Compiles a function into a callable TensorFlow graph\n",
    "\n",
    "아래 두 코드를 실행해 보면 위 정의의 의미가 명확해질 것입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "finite-mount",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'numpy.ndarray'>\n",
      "<class 'numpy.ndarray'>\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([7, 7])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "\n",
    "def f(x, y):\n",
    "    print(type(x))\n",
    "    print(type(y))\n",
    "    return x ** 2 + y\n",
    "\n",
    "x = np.array([2, 3])\n",
    "y = np.array([3, -2])\n",
    "f(x, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "incident-discharge",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'tensorflow.python.framework.ops.Tensor'>\n",
      "<class 'tensorflow.python.framework.ops.Tensor'>\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tf.Tensor: shape=(2,), dtype=int64, numpy=array([7, 7])>"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "\n",
    "@tf.function    # 위와 동일한 함수이지만 @tf.function 데코레이터가 적용되었습니다.\n",
    "def f(x, y):\n",
    "    print(type(x))\n",
    "    print(type(y))\n",
    "    return x ** 2 + y\n",
    "\n",
    "x = np.array([2, 3])\n",
    "y = np.array([3, -2])\n",
    "f(x, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "combined-decision",
   "metadata": {},
   "source": [
    "`Numpy array` 입력 x, y를 동일하게 사용했지만 f(x,y)의 결과 타입은 다릅니다. `@tf.function` 데코레이터가 사용된 함수에 입력된 입력은 Tensorflow의 `graph 노드`가 될 수 있는 타입으로 자동 변환됩니다. 알아두면 향후 Tensorflow 모델 구현에 매우 편리하게 사용할 수 있을 것 같습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "guilty-universe",
   "metadata": {},
   "outputs": [],
   "source": [
    "@tf.function\n",
    "def train_step(images):  #(1) 입력데이터\n",
    "    noise = tf.random.normal([BATCH_SIZE, noise_dim])  #(2) 생성자 입력 노이즈\n",
    "\n",
    "    with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:  #(3) tf.GradientTape() 오픈\n",
    "        generated_images = generator(noise, training=True)  #(4) generated_images 생성\n",
    "\n",
    "        #(5) discriminator 판별\n",
    "        real_output = discriminator(images, training=True)\n",
    "        fake_output = discriminator(generated_images, training=True)\n",
    "\n",
    "        #(6) loss 계산\n",
    "        gen_loss = generator_loss(fake_output)\n",
    "        disc_loss = discriminator_loss(real_output, fake_output)\n",
    "\n",
    "        #(7) accuracy 계산\n",
    "        real_accuracy, fake_accuracy = discriminator_accuracy(real_output, fake_output) \n",
    "    \n",
    "    #(8) gradient 계산\n",
    "    gradients_of_generator = gen_tape.gradient(gen_loss, generator.trainable_variables)\n",
    "    gradients_of_discriminator = disc_tape.gradient(disc_loss, discriminator.trainable_variables)\n",
    "\n",
    "    #(9) 모델 학습\n",
    "    generator_optimizer.apply_gradients(zip(gradients_of_generator, generator.trainable_variables))\n",
    "    discriminator_optimizer.apply_gradients(zip(gradients_of_discriminator, discriminator.trainable_variables))\n",
    "\n",
    "    return gen_loss, disc_loss, real_accuracy, fake_accuracy  #(10) 리턴값"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "billion-astronomy",
   "metadata": {},
   "source": [
    "train_step 함수를 하나하나 뜯어보면 다음과 같습니다.\n",
    "\n",
    "- (1) 입력 데이터: Real Image 역할을 할 images 한 세트를 입력으로 받음\n",
    "- (2) 생성자 입력 노이즈 : generator가 FAKE IMAGE를 생성하기 위한 noise를 images 한 세트와 같은 크기인 BATCH_SIZE 만큼 생성함\n",
    "- (3) `tf.GradientTape()`는 가중치 갱신을 위한 Gradient를 자동 미분으로 계산하기 위해 `with` 구문 열기\n",
    "- (4) `generated_images` 생성 : generator가 noise를 입력받은 후 generated_images 생성\n",
    "- (5) `discriminator` 판별 : discriminator가 Real Image인 images와 Fake Image인 generated_images를 각각 입력받은 후 real_output, fake_output 출력\n",
    "- (6) loss 계산 : fake_output, real_output으로 generator와 discriminator 각각의 loss 계산\n",
    "- (7) accuracy 계산 : fake_output, real_output으로 discriminator의 정확도 계산\n",
    "- (8) gradient 계산 : `gen_tape`와 `disc_tape`를 활용해 gradient를 자동으로 계산\n",
    "- (9) 모델 학습 : 계산된 gradient를 optimizer에 입력해 가중치 갱신\n",
    "- (10) 리턴값 : 이번 스텝에 계산된 loss와 accuracy를 리턴"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cooked-conditions",
   "metadata": {},
   "source": [
    "그러면, 이렇게 한 단계씩 학습할 train_step과 함께 일정 간격으로 학습 현황을 볼 수 있는 샘플을 생성하는 함수를 만들어 보죠.\n",
    "아까 만들어 놓았던 고정된 `seed`를 이용해서 결과물을 만들어내므로, 고정된 seed에 대한 결과물이 얼마나 나아지고 있는지를 확인할 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "accessible-overhead",
   "metadata": {},
   "outputs": [],
   "source": [
    "def generate_and_save_images(model, epoch, it, sample_seeds):\n",
    "\n",
    "    predictions = model(sample_seeds, training=False)\n",
    "\n",
    "    fig = plt.figure(figsize=(4, 4))\n",
    "    for i in range(predictions.shape[0]):\n",
    "        plt.subplot(4, 4, i+1)\n",
    "        plt.imshow(predictions[i, :, :, 0], cmap='gray')\n",
    "        plt.axis('off')\n",
    "   \n",
    "    plt.savefig('{}/aiffel/dcgan_newimage/fashion/generated_samples/sample_epoch_{:04d}_iter_{:03d}.png'\n",
    "                    .format(os.getenv('HOME'), epoch, it))\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "entitled-civilian",
   "metadata": {},
   "source": [
    "model이 16개의 seed가 들어있는 sample_seeds를 입력받아서 만들어낸 prediction을 `matplotlib`으로 시각화해주는 과정이 담겨 있는데, 어딘가 익숙하지 않나요?\n",
    "\n",
    "바로 우리가 위에서 해봤던, subplot을 이용해서 총 16개의 sample을 시각화하는 과정입니다. 16개이므로 4행 4열로 나누어 `subplot(4, 4, i+1)`로 시각화하는 것이죠.\n",
    "\n",
    "또한, plt에 저장되어 보이는 이미지를 `plt.savefig`로 간단히 파일로 저장합니다.\n",
    "\n",
    "\n",
    "학습 과정을 체크하기 위해 시각화해 보아야 할 중요한 것으로 `loss`와 `accuracy` 그래프를 빼놓을 수 없겠습니다. GAN의 학습 과정은 지도학습 모델보다 까다로운데, 이것은 두 모델이 서로의 학습 과정에 영향을 주고받기 때문입니다. `train_step()` 함수가 리턴하는 `gen_loss`, `disc_loss`, `real_accuracy`, `fake_accuracy` 이상 4가지 값을 `history`라는 `dict` 구조에 리스트로 저장하고 있다가 매 `epoch`마다 시각화하는 함수를 만들어 봅시다. 예를 들어 생성자의 loss의 history는 `history['gen_loss']`로 접근할 수 있는 list로 관리할 예정입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "alternate-michigan",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib.pylab import rcParams\n",
    "rcParams['figure.figsize'] = 15, 6    # matlab 차트의 기본 크기를 15,6으로 지정해 줍니다.\n",
    "\n",
    "def draw_train_history(history, epoch):\n",
    "    # summarize history for loss  \n",
    "    plt.subplot(211)  \n",
    "    plt.plot(history['gen_loss'])  \n",
    "    plt.plot(history['disc_loss'])  \n",
    "    plt.title('model loss')  \n",
    "    plt.ylabel('loss')  \n",
    "    plt.xlabel('batch iters')  \n",
    "    plt.legend(['gen_loss', 'disc_loss'], loc='upper left')  \n",
    "\n",
    "    # summarize history for accuracy  \n",
    "    plt.subplot(212)  \n",
    "    plt.plot(history['fake_accuracy'])  \n",
    "    plt.plot(history['real_accuracy'])  \n",
    "    plt.title('discriminator accuracy')  \n",
    "    plt.ylabel('accuracy')  \n",
    "    plt.xlabel('batch iters')  \n",
    "    plt.legend(['fake_accuracy', 'real_accuracy'], loc='upper left')  \n",
    "    \n",
    "    # training_history 디렉토리에 epoch별로 그래프를 이미지 파일로 저장합니다.\n",
    "    plt.savefig('{}/aiffel/dcgan_newimage/fashion/training_history/train_history_{:04d}.png'\n",
    "                    .format(os.getenv('HOME'), epoch))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "wrapped-falls",
   "metadata": {},
   "source": [
    "다음은 정기적으로 모델을 저장하기 위한 checkpoint를 만들어줍니다.\n",
    "\n",
    "`tf.train.Checkpoint`를 활용하면 매번 모델을 직접 저장해 주지 않아도, 코드 한 줄로 빠르고 편하게 버전 관리를 할 수 있습니다.\n",
    "모델이 복잡해지고 학습 속도가 오래 걸릴수록, 모델에 대한 저장 및 버전 관리는 필수적입니다. 아주 오랜 시간 동안 학습을 시켰는데 네트워크가 끊긴다거나, 예상치 못한 문제로 인해 모델을 잃어버린다면 그보다 끔찍한 일은 없을 테니까요..!\n",
    "\n",
    "`checkpoint`에는 optimizer와 `생성자, 판별자`를 모두 넣어 저장합니다. 정확히는 생성자와 판별자가 학습한 모델 `가중치`를 저장하는 것이죠. 게임의 체크포인트처럼 우리의 학습 환경과 학습한 내용을 저장해 둡니다.\n",
    "\n",
    "checkpoint 모델을 저장하기 위해 작업환경 내에 `training_checkpoints` 라는 디렉토리를 사용하겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "intelligent-sydney",
   "metadata": {},
   "outputs": [],
   "source": [
    "checkpoint_dir = os.getenv('HOME')+'/aiffel/dcgan_newimage/fashion/training_checkpoints'\n",
    "\n",
    "checkpoint_prefix = os.path.join(checkpoint_dir, \"ckpt\")\n",
    "checkpoint = tf.train.Checkpoint(generator_optimizer=generator_optimizer,\n",
    "                                 discriminator_optimizer=discriminator_optimizer,\n",
    "                                 generator=generator,\n",
    "                                 discriminator=discriminator)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "pleasant-beaver",
   "metadata": {},
   "source": [
    "## 생성 모델이 똑똑해지기 위한 기나긴 여정 (3) 학습 시키기"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "herbal-marathon",
   "metadata": {},
   "source": [
    "자.. 이제 대망의 모델 학습입니다.\n",
    "\n",
    "지금까지 한 단계를 학습하는 `train_step`, 샘플 이미지를 생성하고 저장하기 위한 `generate_and_save_images()`, 학습 과정을 시각화하는 `draw_train_history()`, 그리고 모델까지 저장하기 위한 `checkpoint`까지 모두 준비가 되었으니 이것들을 한곳에 합치기만 하면 됩니다.\n",
    "\n",
    "한번 확인해 볼까요!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "presidential-rocket",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(dataset, epochs, save_every):\n",
    "    start = time.time()\n",
    "    history = {'gen_loss':[], 'disc_loss':[], 'real_accuracy':[], 'fake_accuracy':[]}\n",
    "\n",
    "    for epoch in range(epochs):\n",
    "        epoch_start = time.time()\n",
    "        for it, image_batch in enumerate(dataset):\n",
    "            gen_loss, disc_loss, real_accuracy, fake_accuracy = train_step(image_batch)\n",
    "            history['gen_loss'].append(gen_loss)\n",
    "            history['disc_loss'].append(disc_loss)\n",
    "            history['real_accuracy'].append(real_accuracy)\n",
    "            history['fake_accuracy'].append(fake_accuracy)\n",
    "\n",
    "            if it % 50 == 0:\n",
    "                display.clear_output(wait=True)\n",
    "                generate_and_save_images(generator, epoch+1, it+1, seed)\n",
    "                print('Epoch {} | iter {}'.format(epoch+1, it+1))\n",
    "                print('Time for epoch {} : {} sec'.format(epoch+1, int(time.time()-epoch_start)))\n",
    "\n",
    "        if (epoch + 1) % save_every == 0:\n",
    "            checkpoint.save(file_prefix=checkpoint_prefix)\n",
    "\n",
    "        display.clear_output(wait=True)\n",
    "        generate_and_save_images(generator, epochs, it, seed)\n",
    "        print('Time for training : {} sec'.format(int(time.time()-start)))\n",
    "\n",
    "        draw_train_history(history, epoch)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "august-thursday",
   "metadata": {},
   "source": [
    "출력하고자 하는 print문과 몇 번마다 출력하는지에 대한 구문을 원하는 대로 수정해서 마음에 드는 형태로 확인해 보세요!\n",
    "\n",
    "그러면 모델을 저장하는 간격을 담당하는 `save_every`와 전체 학습 에포크를 결정하는 `EPOCHS` 파라미터도 설정해 준 후 훈련시켜 보도록 하겠습니다!\n",
    "\n",
    "\n",
    "\n",
    "본 학습은 이미지를 활용한 것이기 때문에 CPU 환경에서는 매우 느릴 수 있습니다.\n",
    "RTX2070 또는 구글 Colab의 T4 정도 사양의 경우 1 Epoch 당 10초가 걸리는 반면, 일반적인 CPU 환경에서는 최소 3분에서 10분 이상이 걸릴 수 있으므로 최대한 GPU 환경에서 실습해 보는 것을 추천합니다.\n",
    "\n",
    "이제 다음과 같이 train() 함수를 실행시키면 드디어 모델이 학습해 나가면서 만들어내는 결과물을 실시간으로 확인할 수 있습니다!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "ahead-pierce",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "save_every = 5\n",
    "EPOCHS = 50\n",
    "\n",
    "# 사용가능한 GPU 디바이스 확인\n",
    "tf.config.list_physical_devices(\"GPU\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "simple-alexandria",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 288x288 with 16 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time for training : 1495 sec\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 7min 7s, sys: 46.8 s, total: 7min 54s\n",
      "Wall time: 24min 56s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "train(train_dataset, EPOCHS, save_every)\n",
    "\n",
    "# 학습과정의 loss, accuracy 그래프 이미지 파일이 ~/aiffel/dcgan_newimage/fashion/training_history 경로에 생성되고 있으니\n",
    "# 진행 과정을 수시로 확인해 보시길 권합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "postal-seven",
   "metadata": {},
   "source": [
    "학습이 성공적으로 진행되었나요?\n",
    "\n",
    "만약 생성자가 만든 Fake Image에 대한 판별자의 Accuracy(fake_accuracy)가 계속 1에 가깝게 유지되고 있다면, 생성자가 만든 이미지가 아직 판별자를 성공적으로 속이지 못하고 있다는 뜻입니다.\n",
    "\n",
    "보다 나은 결과물을 만들기 위해서는 더욱 많은 epoch으로 학습을 시켜야 합니다. 결과물이 아직 흐릿하다면 더욱 많은 학습을 시켜보세요!\n",
    "과연 생성 모델이 이전에는 없던 새로운 디자인의 패션을 만들어내는 것을 확인할 수 있을 거예요..! (물론 시간은 그와 비례하게 늘어나겠죠..?)\n",
    "\n",
    "어쩌면 단순히 학습 수행 시간만 늘리는 것보다, 모델 구조나 학습 방법을 바꾸어 봐야 할 수도 있습니다. 그래프를 유심히 보면서 학습 결과가 더 이상 개선되고 있지 않은지 여부를 지속적으로 살펴봅시다."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "musical-valuable",
   "metadata": {},
   "source": [
    "## 학습과정 시각화하기"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "sweet-hudson",
   "metadata": {},
   "source": [
    "학습이 끝난 후 우리가 생성했던 샘플 이미지들을 합쳐 GIF 파일, 일명 움짤로 만들어 보겠습니다. 더욱 생생한 학습 진행 과정을 확인해 볼 수 있을 테니까요!\n",
    "\n",
    "GIF 파일은 우리가 오래전에 import 해놓았던 `imageio` 라이브러리를 활용해 만들 수 있습니다.\n",
    "`imageio.get_writer`를 활용해서 파일을 열고, 거기에 `append_data`로 이미지를 하나씩 붙여나가는 방식입니다.\n",
    "아래 코드가 무사히 실행되면, `fasion_mnist_dcgan.gif` 파일이 저장되어 있는 것을 확인할 수 있을 겁니다!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "welsh-motorcycle",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-rw-r--r-- 1 root root 1317797 Feb  8 07:14 /aiffel/aiffel/dcgan_newimage/fashion/fashion_mnist_dcgan.gif\r\n"
     ]
    }
   ],
   "source": [
    "anim_file = os.getenv('HOME')+'/aiffel/dcgan_newimage/fashion/fashion_mnist_dcgan.gif'\n",
    "\n",
    "with imageio.get_writer(anim_file, mode='I') as writer:\n",
    "    filenames = glob.glob('{}/aiffel/dcgan_newimage/fashion/generated_samples/sample*.png'.format(os.getenv('HOME')))\n",
    "    filenames = sorted(filenames)\n",
    "    last = -1\n",
    "    for i, filename in enumerate(filenames):\n",
    "        frame = 2*(i**0.5)\n",
    "        if round(frame) > round(last):\n",
    "            last = frame\n",
    "        else:\n",
    "            continue\n",
    "        image = imageio.imread(filename)\n",
    "        writer.append_data(image)\n",
    "    image = imageio.imread(filename)\n",
    "    writer.append_data(image)\n",
    "\n",
    "!ls -l ~/aiffel/dcgan_newimage/fashion/fashion_mnist_dcgan.gif"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}