8.5-introduction-to-gans.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'2.2.2'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import keras\n",
    "keras.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 적대적 생성 신경망 소개\n",
    "\n",
    "이 노트북은 [케라스 창시자에게 배우는 딥러닝](https://tensorflow.blog/%EC%BC%80%EB%9D%BC%EC%8A%A4-%EB%94%A5%EB%9F%AC%EB%8B%9D/) 책의 8장 5절의 코드 예제입니다. 책에는 더 많은 내용과 그림이 있습니다. 이 노트북에는 소스 코드에 관련된 설명만 포함합니다.\n",
    "\n",
    "---\n",
    "[...]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## GAN 구현 방법\n",
    "\n",
    "이 절에서는 케라스에서 가장 기본적인 형태의 GAN을 구현하는 방법을 설명하겠습니다. GAN은 수준 높은 기술이기 때문에 기술적인 내용을 깊이 설명하는 것은 이 책의 범위를 벗어납니다. 구체적인 구현은 심층 합성곱 GAN(DCGAN)입니다. 생성자와 판별자가 심층 컨브넷입니다. 특히 생성자에서 이미지 업샘플링을 위해 `Conv2DTranspose` 층을 사용합니다.\n",
    "\n",
    "CIFAR10 데이터셋의 이미지로 GAN을 훈련하겠습니다. 이 데이터셋은 32 × 32 크기의 RGB 이미지 50,000개로 이루어져 있고 10개의 클래스를 가집니다(클래스마다 5,000개의 이미지가 있습니다). 문제를 간단하게 만들기 위해 “frog” 클래스의 이미지만 사용하겠습니다.\n",
    "\n",
    "GAN 구조는 다음과 같습니다:\n",
    "\n",
    "1.\t`generator` 네트워크는 `(latent_dim,)` 크기의 벡터를 `(32, 32, 3)` 크기의 이미지로 매핑합니다.\n",
    "2.\t`discriminator` 네트워크는 (32, 32, 3) 크기의 이미지가 진짜일 확률을 추정하여 이진 값으로 매핑합니다.\n",
    "3.\t생성자와 판별자를 연결하는 `gan` 네트워크를 만듭니다. `gan(x) = discriminator(generator(x))` 입니다. 이 `gan` 네트워크는 잠재 공간의 벡터를 판별자의 평가로 매핑합니다. 판별자는 생성자가 잠재 공간의 벡터를 디코딩한 것이 얼마나 현실적인지를 평가합니다.\n",
    "4.\t“진짜\"/”가짜\" 레이블과 함께 진짜 이미지와 가짜 이미지 샘플을 사용해 판별자를 훈련합니다. 일반적인 이미지 분류 모델을 훈련하는 것과 동일합니다.\n",
    "5. 생성자를 훈련하려면 `gan` 모델의 손실에 대한 생성자 가중치의 그래디언트를 사용합니다. 이 말은 매 단계마다 생성자에 의해 디코딩된 이미지를 판별자가 “진짜\"로 분류하도록 만드는 방향으로 생성자의 가중치를 이동한다는 뜻입니다. 다른 말로하면 판별자를 속이도록 생성자를 훈련합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 훈련 방법\n",
    "\n",
    "GAN을 훈련하고 튜닝하는 과정은 어렵기로 유명합니다. 알아두어야 할 몇 가지 유용한 기법이 있습니다. 딥러닝의 대부분이 그렇듯이 이는 과학보다는 연금술에 가깝습니다. 이런 기법들은 이론에 바탕을 둔 지침이 아니고 경험을 통해 발견된 것입니다. 실제 일어난 현상을 직관적으로 이해하는 수준에서 검증되었습니다. 모든 문제에 반드시 적용해야 것은 아니지만 경험상 잘 작동한다고 알려져 있습니다.\n",
    "\n",
    "다음은 이 절에서 GAN 생성자와 판별자를 구현하는 데 사용할 몇 가지 기법입니다. 이 목록이 GAN에 관련된 전체 팁이 아닙니다. GAN 논문들에서 더 많은 방법을 볼 수 있습니다.\n",
    "\n",
    "* 생성자의 마지막 활성화로 다른 종류의 모델에서 널리 사용하는 `sigmoid` 대신 `tanh` 함수를 사용합니다. \n",
    "* 균등 분포가 아니고 정규 분포(가우시안 분포)를 사용하여 잠재 공간에서 포인트를 샘플링합니다.\n",
    "* 무작위성은 모델을 견고하게 만듭니다. GAN 훈련은 동적 평형을 만들기 때문에 여러 방식으로 갇힐 가능성이 높습니다. 훈련하는 동안 무작위성을 주입하면 이를 방지하는 데 도움이 됩니다. 무작위성은 두 가지 방법으로 주입합니다. 판별자에 드롭아웃을 사용하거나 판별자를 위해 레이블에 랜덤 노이즈를 추가합니다.\n",
    "* 희소한 그래디언트는 GAN 훈련을 방해할 수 있습니다. 딥러닝에서 희소는 종종 바람직한 현상이지만 GAN에서는 그렇지 않습니다. 그래디언트를 희소하게 만들 수 있는 것은 최대 풀링 연산과 ReLU 활성화 두 가지입니다. 최대 풀링 대신 스트라이드 합성곱을 사용해 다운샘플링을 하는 것이 좋습니다. 또 ReLU 활성화 대신 `LeakyReLU` 층을 사용하세요. ReLU와 비슷하지만 음수의 활성화 값을 조금 허용하기 때문에 희소가 조금 완화됩니다.\n",
    "* 생성자에서 픽셀 공간을 균일하게 다루지 못해 생성된 이미지에서 체스판 모양이 종종 나타납니다(그림 8-17 참조). 이를 해결하기 위해 생성자와 판별자에서 스트라이드 `Conv2DTranpose`나 `Conv2D`를 사용할 때 스트라이드 크기로 나누어질 수 있는 커널 크기를 사용합니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 생성자\n",
    "\n",
    "먼저 벡터(훈련하는 동안 잠재 공간에서 무작위로 샘플링됩니다)를 후보 이미지로 변환하는 `generator` 모델을 만들어 보죠. GAN에서 발생하는 많은 문제 중 하나는 생성자가 노이즈 같은 이미지를 생성하는 데서 멈추는 것입니다. 판별자와 생성자 양쪽에 모두 드롭아웃을 사용하는 것이 해결 방법이 될 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 32)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 32768)             1081344   \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_1 (LeakyReLU)    (None, 32768)             0         \n",
      "_________________________________________________________________\n",
      "reshape_1 (Reshape)          (None, 16, 16, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 16, 16, 256)       819456    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_2 (LeakyReLU)    (None, 16, 16, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_transpose_1 (Conv2DTr (None, 32, 32, 256)       1048832   \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_3 (LeakyReLU)    (None, 32, 32, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 32, 32, 256)       1638656   \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_4 (LeakyReLU)    (None, 32, 32, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 32, 32, 256)       1638656   \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_5 (LeakyReLU)    (None, 32, 32, 256)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 32, 32, 3)         37635     \n",
      "=================================================================\n",
      "Total params: 6,264,579\n",
      "Trainable params: 6,264,579\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras import layers\n",
    "import numpy as np\n",
    "\n",
    "latent_dim = 32\n",
    "height = 32\n",
    "width = 32\n",
    "channels = 3\n",
    "\n",
    "generator_input = keras.Input(shape=(latent_dim,))\n",
    "\n",
    "# 입력을 16 × 16 크기의 128개 채널을 가진 특성 맵으로 변환합니다\n",
    "x = layers.Dense(128 * 16 * 16)(generator_input)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Reshape((16, 16, 128))(x)\n",
    "\n",
    "# 합성곱 층을 추가합니다\n",
    "x = layers.Conv2D(256, 5, padding='same')(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "\n",
    "# 32 × 32 크기로 업샘플링합니다\n",
    "x = layers.Conv2DTranspose(256, 4, strides=2, padding='same')(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "\n",
    "# 합성곱 층을 더 추가합니다\n",
    "x = layers.Conv2D(256, 5, padding='same')(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Conv2D(256, 5, padding='same')(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "\n",
    "# 32 × 32 크기의 1개 채널을 가진 특성 맵을 생성합니다\n",
    "x = layers.Conv2D(channels, 7, activation='tanh', padding='same')(x)\n",
    "generator = keras.models.Model(generator_input, x)\n",
    "generator.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 판별자\n",
    "\n",
    "다음은 후보 이미지(진짜 혹은 가짜)를 입력으로 받고 두 개의 클래스로 분류하는 `discriminator` 모델을 만들겠습니다. 이 클래스는 '생성된 이미지' 또는 '훈련 세트에서 온 진짜 이미지'입니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_2 (InputLayer)         (None, 32, 32, 3)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 30, 30, 128)       3584      \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_6 (LeakyReLU)    (None, 30, 30, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_6 (Conv2D)            (None, 14, 14, 128)       262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_7 (LeakyReLU)    (None, 14, 14, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_7 (Conv2D)            (None, 6, 6, 128)         262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_8 (LeakyReLU)    (None, 6, 6, 128)         0         \n",
      "_________________________________________________________________\n",
      "conv2d_8 (Conv2D)            (None, 2, 2, 128)         262272    \n",
      "_________________________________________________________________\n",
      "leaky_re_lu_9 (LeakyReLU)    (None, 2, 2, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 513       \n",
      "=================================================================\n",
      "Total params: 790,913\n",
      "Trainable params: 790,913\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "discriminator_input = layers.Input(shape=(height, width, channels))\n",
    "x = layers.Conv2D(128, 3)(discriminator_input)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Conv2D(128, 4, strides=2)(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Conv2D(128, 4, strides=2)(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Conv2D(128, 4, strides=2)(x)\n",
    "x = layers.LeakyReLU()(x)\n",
    "x = layers.Flatten()(x)\n",
    "\n",
    "# 드롭아웃 층을 넣는 것이 아주 중요합니다!\n",
    "x = layers.Dropout(0.4)(x)\n",
    "\n",
    "# 분류 층\n",
    "x = layers.Dense(1, activation='sigmoid')(x)\n",
    "\n",
    "discriminator = keras.models.Model(discriminator_input, x)\n",
    "discriminator.summary()\n",
    "\n",
    "# 옵티마이저에서 (값을 지정하여) 그래디언트 클리핑을 사용합니다\n",
    "# 안정된 훈련을 위해서 학습률 감쇠를 사용합니다\n",
    "discriminator_optimizer = keras.optimizers.RMSprop(lr=0.0008, clipvalue=1.0, decay=1e-8)\n",
    "discriminator.compile(optimizer=discriminator_optimizer, loss='binary_crossentropy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 적대적 네트워크\n",
    "\n",
    "마지막으로 생성자와 판별자를 연결하여 GAN을 설정합니다. 훈련할 때 생성자가 판별자를 속이는 능력이 커지도록 학습합니다. 이 모델은 잠재 공간의 포인트를 “진짜\" 또는 “가짜\"의 분류 결정으로 변환합니다. 훈련에 사용되는 타깃 레이블은 항상 '진짜 이미지'입니다. `gan`을 훈련하는 것은 `discriminator`가 가짜 이미지를 보았을 때 진짜라고 예측하도록 만들기 위해 `generator`의 가중치를 업데이트하는 것입니다. 훈련하는 동안 판별자를 동결(학습되지 않도록)하는 것이 아주 중요합니다. `gan`을 훈련할 때 가중치가 업데이트되지 않습니다. 판별자의 가중치가 훈련하는 동안 업데이트되면 판별자는 항상 “진짜\"를 예측하도록 훈련됩니다. 이것이 우리가 원하는 바는 아니죠!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 판별자의 가중치가 훈련되지 않도록 설정합니다(gan 모델에만 적용됩니다)\n",
    "discriminator.trainable = False\n",
    "\n",
    "gan_input = keras.Input(shape=(latent_dim,))\n",
    "gan_output = discriminator(generator(gan_input))\n",
    "gan = keras.models.Model(gan_input, gan_output)\n",
    "\n",
    "gan_optimizer = keras.optimizers.RMSprop(lr=0.0004, clipvalue=1.0, decay=1e-8)\n",
    "gan.compile(optimizer=gan_optimizer, loss='binary_crossentropy')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## DCGAN 훈련 방법\n",
    "\n",
    "이제 훈련을 시작합니다.  훈련 반복의 내용을 요약 정리해 보겠습니다.\n",
    "\n",
    "```\n",
    "매 반복마다 다음을 수행합니다:\n",
    "\n",
    "    1.잠재 공간에서 무작위로 포인트를 뽑습니다(랜덤 노이즈).\n",
    "    2.이 랜덤 노이즈를 사용해 `generator`에서 이미지를 생성합니다.\n",
    "    3.생성된 이미지와 진짜 이미지를 섞습니다.\n",
    "    4.진짜와 가짜가 섞인 이미지와 이에 대응하는 타깃을 사용해 `discriminator`를 훈련합니다. 타깃은 “진짜\"(실제 이미지일 경우) 또는 “가짜\"(생성된 이미지일 경우)입니다.\n",
    "    5.잠재 공간에서 무작위로 새로운 포인트를 뽑습니다.\n",
    "    6.이 랜덤 벡터를 사용해 `gan`을 훈련합니다. 모든 타깃은 “진짜\"로 설정합니다. 판별자가 생성된 이미지를 모두 “진짜 이미지\"라고 예측하도록 생성자의 가중치를 업데이트합니다(`gan` 안에서 판별자는 동결되기 때문에 생성자만 업데이트합니다). 결국 생성자는 판별자를 속이도록 훈련합니다.\n",
    "```\n",
    "\n",
    "실제로 만들어 보죠:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/haesun/anaconda3/envs/deep-learning-with-python/lib/python3.6/site-packages/keras/engine/training.py:479: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
      "  'Discrepancy between trainable weights and collected trainable'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "스텝 0에서 판별자 손실: 0.6949256\n",
      "스텝 0에서 적대적 손실: 0.65713036\n",
      "스텝 100에서 판별자 손실: 0.6902039\n",
      "스텝 100에서 적대적 손실: 0.8107964\n",
      "스텝 200에서 판별자 손실: 0.69638944\n",
      "스텝 200에서 적대적 손실: 0.7792972\n",
      "스텝 300에서 판별자 손실: 0.6901457\n",
      "스텝 300에서 적대적 손실: 0.77436054\n",
      "스텝 400에서 판별자 손실: 0.67664975\n",
      "스텝 400에서 적대적 손실: 0.73274004\n",
      "스텝 500에서 판별자 손실: 0.6887272\n",
      "스텝 500에서 적대적 손실: 0.7381878\n",
      "스텝 600에서 판별자 손실: 0.6828704\n",
      "스텝 600에서 적대적 손실: 0.76748234\n",
      "스텝 700에서 판별자 손실: 0.6951365\n",
      "스텝 700에서 적대적 손실: 0.7289097\n",
      "스텝 800에서 판별자 손실: 0.7190336\n",
      "스텝 800에서 적대적 손실: 0.74828637\n",
      "스텝 900에서 판별자 손실: 0.7041694\n",
      "스텝 900에서 적대적 손실: 0.76092464\n",
      "스텝 1000에서 판별자 손실: 0.6961935\n",
      "스텝 1000에서 적대적 손실: 0.7705952\n",
      "스텝 1100에서 판별자 손실: 0.6952446\n",
      "스텝 1100에서 적대적 손실: 0.7563175\n",
      "스텝 1200에서 판별자 손실: 0.6940303\n",
      "스텝 1200에서 적대적 손실: 0.86164886\n",
      "스텝 1300에서 판별자 손실: 0.6923666\n",
      "스텝 1300에서 적대적 손실: 0.75390154\n",
      "스텝 1400에서 판별자 손실: 0.69698876\n",
      "스텝 1400에서 적대적 손실: 0.74170864\n",
      "스텝 1500에서 판별자 손실: 0.6879484\n",
      "스텝 1500에서 적대적 손실: 0.73414576\n",
      "스텝 1600에서 판별자 손실: 0.70734346\n",
      "스텝 1600에서 적대적 손실: 0.73252755\n",
      "스텝 1700에서 판별자 손실: 0.7116456\n",
      "스텝 1700에서 적대적 손실: 0.7651757\n",
      "스텝 1800에서 판별자 손실: 0.72208524\n",
      "스텝 1800에서 적대적 손실: 0.75683695\n",
      "스텝 1900에서 판별자 손실: 0.7225605\n",
      "스텝 1900에서 적대적 손실: 0.74279666\n",
      "스텝 2000에서 판별자 손실: 0.681931\n",
      "스텝 2000에서 적대적 손실: 0.93539745\n",
      "스텝 2100에서 판별자 손실: 0.6859837\n",
      "스텝 2100에서 적대적 손실: 0.77806115\n",
      "스텝 2200에서 판별자 손실: 0.69303304\n",
      "스텝 2200에서 적대적 손실: 0.7447194\n",
      "스텝 2300에서 판별자 손실: 0.6980282\n",
      "스텝 2300에서 적대적 손실: 0.74495566\n",
      "스텝 2400에서 판별자 손실: 0.6896796\n",
      "스텝 2400에서 적대적 손실: 0.77484715\n",
      "스텝 2500에서 판별자 손실: 0.78387755\n",
      "스텝 2500에서 적대적 손실: 0.7450498\n",
      "스텝 2600에서 판별자 손실: 0.7133597\n",
      "스텝 2600에서 적대적 손실: 0.8398307\n",
      "스텝 2700에서 판별자 손실: 0.71523935\n",
      "스텝 2700에서 적대적 손실: 0.7921444\n",
      "스텝 2800에서 판별자 손실: 0.69421625\n",
      "스텝 2800에서 적대적 손실: 0.7998878\n",
      "스텝 2900에서 판별자 손실: 0.6895941\n",
      "스텝 2900에서 적대적 손실: 0.7869561\n",
      "스텝 3000에서 판별자 손실: 0.83328277\n",
      "스텝 3000에서 적대적 손실: 0.79146093\n",
      "스텝 3100에서 판별자 손실: 0.70107824\n",
      "스텝 3100에서 적대적 손실: 0.73397076\n",
      "스텝 3200에서 판별자 손실: 0.70538676\n",
      "스텝 3200에서 적대적 손실: 0.69698036\n",
      "스텝 3300에서 판별자 손실: 0.6913239\n",
      "스텝 3300에서 적대적 손실: 0.7309754\n",
      "스텝 3400에서 판별자 손실: 0.7014526\n",
      "스텝 3400에서 적대적 손실: 0.7882652\n",
      "스텝 3500에서 판별자 손실: 0.69528425\n",
      "스텝 3500에서 적대적 손실: 0.74922466\n",
      "스텝 3600에서 판별자 손실: 0.6883661\n",
      "스텝 3600에서 적대적 손실: 0.73456806\n",
      "스텝 3700에서 판별자 손실: 0.6836599\n",
      "스텝 3700에서 적대적 손실: 0.72061574\n",
      "스텝 3800에서 판별자 손실: 0.74770033\n",
      "스텝 3800에서 적대적 손실: 0.7646421\n",
      "스텝 3900에서 판별자 손실: 0.6970914\n",
      "스텝 3900에서 적대적 손실: 0.75626713\n",
      "스텝 4000에서 판별자 손실: 0.71431667\n",
      "스텝 4000에서 적대적 손실: 0.753148\n",
      "스텝 4100에서 판별자 손실: 0.72747624\n",
      "스텝 4100에서 적대적 손실: 0.77221537\n",
      "스텝 4200에서 판별자 손실: 0.6636469\n",
      "스텝 4200에서 적대적 손실: 0.88101995\n",
      "스텝 4300에서 판별자 손실: 0.68881196\n",
      "스텝 4300에서 적대적 손실: 0.7100999\n",
      "스텝 4400에서 판별자 손실: 0.6997429\n",
      "스텝 4400에서 적대적 손실: 0.7232765\n",
      "스텝 4500에서 판별자 손실: 0.69985485\n",
      "스텝 4500에서 적대적 손실: 0.7265221\n",
      "스텝 4600에서 판별자 손실: 0.7126813\n",
      "스텝 4600에서 적대적 손실: 0.73262024\n",
      "스텝 4700에서 판별자 손실: 0.691203\n",
      "스텝 4700에서 적대적 손실: 0.7504519\n",
      "스텝 4800에서 판별자 손실: 0.6840957\n",
      "스텝 4800에서 적대적 손실: 0.76147616\n",
      "스텝 4900에서 판별자 손실: 0.69118893\n",
      "스텝 4900에서 적대적 손실: 0.7457582\n",
      "스텝 5000에서 판별자 손실: 0.6922146\n",
      "스텝 5000에서 적대적 손실: 0.77010024\n",
      "스텝 5100에서 판별자 손실: 0.68561333\n",
      "스텝 5100에서 적대적 손실: 0.73494273\n",
      "스텝 5200에서 판별자 손실: 0.68722475\n",
      "스텝 5200에서 적대적 손실: 0.7371446\n",
      "스텝 5300에서 판별자 손실: 0.75649905\n",
      "스텝 5300에서 적대적 손실: 0.7871337\n",
      "스텝 5400에서 판별자 손실: 1.1356353\n",
      "스텝 5400에서 적대적 손실: 0.83297\n",
      "스텝 5500에서 판별자 손실: 0.69317013\n",
      "스텝 5500에서 적대적 손실: 0.7214963\n",
      "스텝 5600에서 판별자 손실: 0.6956296\n",
      "스텝 5600에서 적대적 손실: 0.7718533\n",
      "스텝 5700에서 판별자 손실: 0.68193114\n",
      "스텝 5700에서 적대적 손실: 0.7388134\n",
      "스텝 5800에서 판별자 손실: 0.69178975\n",
      "스텝 5800에서 적대적 손실: 0.7142013\n",
      "스텝 5900에서 판별자 손실: 0.7055496\n",
      "스텝 5900에서 적대적 손실: 0.6911681\n",
      "스텝 6000에서 판별자 손실: 0.68773377\n",
      "스텝 6000에서 적대적 손실: 0.79507333\n",
      "스텝 6100에서 판별자 손실: 0.69964826\n",
      "스텝 6100에서 적대적 손실: 0.7587262\n",
      "스텝 6200에서 판별자 손실: 0.6865734\n",
      "스텝 6200에서 적대적 손실: 0.7368025\n",
      "스텝 6300에서 판별자 손실: 0.7116674\n",
      "스텝 6300에서 적대적 손실: 0.70878893\n",
      "스텝 6400에서 판별자 손실: 0.7232849\n",
      "스텝 6400에서 적대적 손실: 0.69788635\n",
      "스텝 6500에서 판별자 손실: 0.6915632\n",
      "스텝 6500에서 적대적 손실: 0.7307524\n",
      "스텝 6600에서 판별자 손실: 0.6972132\n",
      "스텝 6600에서 적대적 손실: 0.73100436\n",
      "스텝 6700에서 판별자 손실: 0.6878956\n",
      "스텝 6700에서 적대적 손실: 0.83057404\n",
      "스텝 6800에서 판별자 손실: 0.46024114\n",
      "스텝 6800에서 적대적 손실: 8.443697\n",
      "스텝 6900에서 판별자 손실: 0.71324813\n",
      "스텝 6900에서 적대적 손실: 1.5412463\n",
      "스텝 7000에서 판별자 손실: 0.7004356\n",
      "스텝 7000에서 적대적 손실: 0.868543\n",
      "스텝 7100에서 판별자 손실: 0.69475687\n",
      "스텝 7100에서 적대적 손실: 0.81978834\n",
      "스텝 7200에서 판별자 손실: 0.68780637\n",
      "스텝 7200에서 적대적 손실: 0.85392743\n",
      "스텝 7300에서 판별자 손실: 0.679514\n",
      "스텝 7300에서 적대적 손실: 0.78175056\n",
      "스텝 7400에서 판별자 손실: 0.6796296\n",
      "스텝 7400에서 적대적 손실: 0.71190035\n",
      "스텝 7500에서 판별자 손실: 0.7143494\n",
      "스텝 7500에서 적대적 손실: 0.82242143\n",
      "스텝 7600에서 판별자 손실: 0.69852495\n",
      "스텝 7600에서 적대적 손실: 0.7110367\n",
      "스텝 7700에서 판별자 손실: 0.6860732\n",
      "스텝 7700에서 적대적 손실: 0.7677852\n",
      "스텝 7800에서 판별자 손실: 0.6929771\n",
      "스텝 7800에서 적대적 손실: 0.783628\n",
      "스텝 7900에서 판별자 손실: 0.67653257\n",
      "스텝 7900에서 적대적 손실: 0.7711078\n",
      "스텝 8000에서 판별자 손실: 0.6838045\n",
      "스텝 8000에서 적대적 손실: 0.86231613\n",
      "스텝 8100에서 판별자 손실: 0.6750197\n",
      "스텝 8100에서 적대적 손실: 0.74030095\n",
      "스텝 8200에서 판별자 손실: 0.7556801\n",
      "스텝 8200에서 적대적 손실: 0.97776\n",
      "스텝 8300에서 판별자 손실: 0.66886103\n",
      "스텝 8300에서 적대적 손실: 0.8162745\n",
      "스텝 8400에서 판별자 손실: 1.0085771\n",
      "스텝 8400에서 적대적 손실: 0.8776614\n",
      "스텝 8500에서 판별자 손실: 0.66519976\n",
      "스텝 8500에서 적대적 손실: 0.796661\n",
      "스텝 8600에서 판별자 손실: 0.7255496\n",
      "스텝 8600에서 적대적 손실: 0.66313\n",
      "스텝 8700에서 판별자 손실: 0.71208185\n",
      "스텝 8700에서 적대적 손실: 0.80712444\n",
      "스텝 8800에서 판별자 손실: 0.68681073\n",
      "스텝 8800에서 적대적 손실: 0.82072306\n",
      "스텝 8900에서 판별자 손실: 0.68857586\n",
      "스텝 8900에서 적대적 손실: 0.75071067\n",
      "스텝 9000에서 판별자 손실: 0.67631257\n",
      "스텝 9000에서 적대적 손실: 0.7848218\n",
      "스텝 9100에서 판별자 손실: 0.67863333\n",
      "스텝 9100에서 적대적 손실: 0.79472554\n",
      "스텝 9200에서 판별자 손실: 0.7040019\n",
      "스텝 9200에서 적대적 손실: 0.80204326\n",
      "스텝 9300에서 판별자 손실: 0.66886413\n",
      "스텝 9300에서 적대적 손실: 0.8013617\n",
      "스텝 9400에서 판별자 손실: 0.63577974\n",
      "스텝 9400에서 적대적 손실: 0.6830619\n",
      "스텝 9500에서 판별자 손실: 0.69877046\n",
      "스텝 9500에서 적대적 손실: 0.79317737\n",
      "스텝 9600에서 판별자 손실: 0.7716285\n",
      "스텝 9600에서 적대적 손실: 1.103823\n",
      "스텝 9700에서 판별자 손실: 0.87332267\n",
      "스텝 9700에서 적대적 손실: 0.7923678\n",
      "스텝 9800에서 판별자 손실: 0.7400392\n",
      "스텝 9800에서 적대적 손실: 0.8203319\n",
      "스텝 9900에서 판별자 손실: 0.67820084\n",
      "스텝 9900에서 적대적 손실: 0.9951621\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "from keras.preprocessing import image\n",
    "\n",
    "# CIFAR10 데이터를 로드합니다\n",
    "(x_train, y_train), (_, _) = keras.datasets.cifar10.load_data()\n",
    "\n",
    "# 개구리 이미지를 선택합니다(클래스 6)\n",
    "x_train = x_train[y_train.flatten() == 6]\n",
    "\n",
    "# 데이터를 정규화합니다\n",
    "x_train = x_train.reshape(\n",
    "    (x_train.shape[0],) + (height, width, channels)).astype('float32') / 255.\n",
    "\n",
    "iterations = 10000\n",
    "batch_size = 20\n",
    "save_dir = './datasets/gan_images/'\n",
    "if not os.path.exists(save_dir):\n",
    "    os.mkdir(save_dir)\n",
    "\n",
    "# 훈련 반복 시작\n",
    "start = 0\n",
    "for step in range(iterations):\n",
    "    # 잠재 공간에서 무작위로 포인트를 샘플링합니다\n",
    "    random_latent_vectors = np.random.normal(size=(batch_size, latent_dim))\n",
    "\n",
    "    # 가짜 이미지를 디코딩합니다\n",
    "    generated_images = generator.predict(random_latent_vectors)\n",
    "\n",
    "    # 진짜 이미지와 연결합니다\n",
    "    stop = start + batch_size\n",
    "    real_images = x_train[start: stop]\n",
    "    combined_images = np.concatenate([generated_images, real_images])\n",
    "\n",
    "    # 진짜와 가짜 이미지를 구분하여 레이블을 합칩니다\n",
    "    labels = np.concatenate([np.ones((batch_size, 1)),\n",
    "                             np.zeros((batch_size, 1))])\n",
    "    # 레이블에 랜덤 노이즈를 추가합니다. 아주 중요합니다!\n",
    "    labels += 0.05 * np.random.random(labels.shape)\n",
    "\n",
    "    # discriminator를 훈련합니다\n",
    "    d_loss = discriminator.train_on_batch(combined_images, labels)\n",
    "\n",
    "    # 잠재 공간에서 무작위로 포인트를 샘플링합니다\n",
    "    random_latent_vectors = np.random.normal(size=(batch_size, latent_dim))\n",
    "\n",
    "    # 모두 “진짜 이미지\"라고 레이블을 만듭니다\n",
    "    misleading_targets = np.zeros((batch_size, 1))\n",
    "\n",
    "    # generator를 훈련합니다(gan 모델에서 discriminator의 가중치는 동결됩니다)\n",
    "    a_loss = gan.train_on_batch(random_latent_vectors, misleading_targets)\n",
    "    \n",
    "    start += batch_size\n",
    "    if start > len(x_train) - batch_size:\n",
    "      start = 0\n",
    "\n",
    "    # 중간 중간 저장하고 그래프를 그립니다\n",
    "    if step % 100 == 0:\n",
    "        # 모델 가중치를 저장합니다\n",
    "        gan.save_weights('gan.h5')\n",
    "\n",
    "        # 측정 지표를 출력합니다\n",
    "        print('스텝 %s에서 판별자 손실: %s' % (step, d_loss))\n",
    "        print('스텝 %s에서 적대적 손실: %s' % (step, a_loss))\n",
    "\n",
    "        # 생성된 이미지 하나를 저장합니다\n",
    "        img = image.array_to_img(generated_images[0] * 255., scale=False)\n",
    "        img.save(os.path.join(save_dir, 'generated_frog' + str(step) + '.png'))\n",
    "\n",
    "        # 비교를 위해 진짜 이미지 하나를 저장합니다\n",
    "        img = image.array_to_img(real_images[0] * 255., scale=False)\n",
    "        img.save(os.path.join(save_dir, 'real_frog' + str(step) + '.png'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "가짜 이미지 몇개를 출력해 보죠:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARsAAAEYCAYAAABsuVKPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3XuMXPd1H/DvuXfes7PvB5cPkRQlO1ZtR1ZZwYDaNI86kNUAspGX/YehPwwwKGIgBtI/hBRoXKB/OEVtI38ULuhKiFq4fjS2YSEx2hiCCyFFoXgtUxIpOiEtUxTJ5ZLLfc3Oe+ae/rEjg6V5zh1yl79Zrb4fgODunL0zv7lz5+zd+Z17fqKqICK616JhD4CI3h2YbIgoCCYbIgqCyYaIgmCyIaIgmGyIKAgmGyIKgsmGiIJgsiGiIDIhH6xQFK1U7PwWS2zGcll/qNrrufFeLzFj3a5fRa0pu0nFfk6JPyx0Ox3/ByB+1Au7QaDXS3neaj+vbmLvTwBIku1UpvvjVqTctxNOq5hPG7U4Y0vZ3UjbJWmPHac9gHvn/usVRfZ9ZzP2+xIA6s3WsqrOpA1hW8lGRB4H8OcAYgD/RVU/7/18pRLht3+3YMbHcuNm7ND8lDuWzuq6G1/bqJuxlRX/ZW5hwo9HZTNW3/Dve+3KkhtHnHXDGS8c5dxt16t+omt17e1Xq/b+BIBaw79vdU6qNfIPbk1Nkna82eze9bYAEEX2Ds9k/GTQSvnN0+3525fz9ts1SklVSbfp33cub8b2zVTcbV8+e+5N9wf67vrPKBGJAfwnAB8F8BCAT4rIQ3d7f0S0t23nM5tHAZxX1TdUtQ3g6wCe3JlhEdFes51kcwDAWzd9f6l/2/9HRE6IyIKILDQbvMKc6N1qO8nmdn9g/kI2UdWTqnpcVY8Xitv4gIuI3tG2k2wuATh00/cHAVzZ3nCIaK/aTrL5IYAHReSoiOQAfALA8zszLCLaa+566ltVuyLyGQD/C1tT38+q6hlvm3wuh6MHj5jx5bcumbGlSy13PBPTk258OhozY52WP41byNpT8gCwstE2Y8UoZSp1ctqNZyWlUAf2FHOt5U8hw65CAABUG04s9g+dUkpdVKNjv57t1Nojn8CuKcnF/uvRTSuMcn49J7G/vyUlnnHqzACg49T4dLv++yMD/7573nGaMq5BbavORlW/B+B7OzISItrTeLkCEQXBZENEQTDZEFEQTDZEFASTDREFEbTFhESCuGxPo407lzBnsyPufR+bvc+Nq9pX+0rK1N7iVf9K4WxSM2Nt+FdeT1b8+efa5qYbz8T2/bey/hRyvuc/9kavasaixL/vtMnSjNOWI5PS0qDV818PtOy497gA0GunXK2u9vRzPmNfOQ0ASHnsTtqV2x17Wt5roQIATrcQAEC9YZdvLK/7x+CgeGZDREEw2RBREEw2RBQEkw0RBcFkQ0RBMNkQURBMNkQURNA6m0SBdtfOb2Ozh81YKbZbRABAsunXRzTrdnf5Rs1fwSAes1dPAICRvN0mopBSehGl1NGs1p0+DwCK++z6o7KU3G1vbNh1NADQWLXbFmSK/n03mv64O5FdF5Kk1PCkLGKAxGlvkUlZmSGf0gYiX7JfUE2pLmqL/9g5pzUGAOTz9mPXUlZmQEptksKOt7opdU0D4pkNEQXBZENEQTDZEFEQTDZEFASTDREFwWRDREEw2RBREEHrbJq9BGeqdl3J+9tFM9Yp+vUoufq6G8+OjZqxfNevs7m+6D92wymP2Fz1a3Q6Db8+IsnY4waAXN6ONxO/rqPT83vtlKbtJWxk065bAoB2x7/vbMb+Pdds+Pu73fbrUQpOP5w48mthKhm/fijp2du3U0pdMinjjvL+7/52x96+001Z/ial303JeQtkUvbZoHhmQ0RBMNkQURBMNkQUBJMNEQXBZENEQTDZEFEQYZdyyWSRn5w34611ezidlGnDqdif+ks27alYsTspAAAKZX86tOa0qEi6dXfbTM5eQgMAkpL/2K2uvbxHK239jrw/V1t0Whp0Gv7+nhjzl4lJnLYFVfG3rXf99hWR2PtExZ+ST2unEOXslh652N/fWfGP4aTpH4jZxC43GC/609M9Z8oeAOKufRwW45Q5/QFtK9mIyAUAVQA9AF1VPb4TgyKivWcnzmx+TVWXd+B+iGgP42c2RBTEdpONAvgbEfmRiJy43Q+IyAkRWRCRhWZK604i2ru2+2fUY6p6RURmAXxfRH6iqi/e/AOqehLASQCYvq/iX6xDRHvWts5sVPVK//9rAL4D4NGdGBQR7T13nWxEpCwilbe/BvCbAE7v1MCIaG/Zzp9RcwC+IyJv389/V9X/6W2QqKLRsesYug27XqWY+PUqpZJd/wAAI6N2fCWlPUXc8VtQVDfsz6KqKZf+l512CACgKUuPrG3Y+6WR89eRyaT8Udto2HUfvYbfYqKY92tl1hrOMjItvx5ltlJx4+rU2dSb/pNO6bqBesuuR9HIf85x7B9HuZJ/LDR69mutSUqNTkqbiF7PqddKa18xoLtONqr6BoBf3pFRENGex6lvIgqCyYaIgmCyIaIgmGyIKAgmGyIKgsmGiIII2s8m6SXYrNq9SNpq1ymMj+1z77tVs+sEAABrdn1FZ8Pv19FN/LqPiZxdnNEt+ONqrzr1JgAQ+XUh45P2UjHrbX9JFGn5tRnTTp1OedbfZ4WUFijdurMsifo1PFPTfs1I1LMP68rIpLvtsWOH3Pj/+fEbZuynl1fdbXOFlLdbSs+ZXNE+N4h7/nmDiH8cZUr22OLszqQJntkQURBMNkQUBJMNEQXBZENEQTDZEFEQTDZEFETQqe9sFGG/0wqi2Lanicvwlzwp+t0UUMraS6K0W5fdbTeuL7nxq02ndYAznQ8AnY4/zQvxlxapOi0TWlV/Wr3gtGIAgH/0XnvZnc1V/9C5suy3BBkbHTVjvba/z9Dyl3KprW+YsXHncQFgZsTvMXFgesyMrW36r2Wtl7LcUM8vsYgy9nsg9mfNkU95q7fb9tg6bf8YHBTPbIgoCCYbIgqCyYaIgmCyIaIgmGyIKAgmGyIKgsmGiIIIWmeTiTOYGh03451l+xL9et2vUWinXL6vNfu+u12/FQMiv85AWnZ9RS5nt4AAgPKkX/ehkf+8Gk4NRK7r93k4eNBv2zGbt1+rwqw/rgtX7FoXAIh7dnuLct6vden2/PqhVqZmxqprfk3Vmdf9eq5M1q4Te+yR+91tlzb8JYOQ+L/7Ly2tmLG164vutt2OX4gzMlY0Y5pNKWLzy9B+jmc2RBQEkw0RBcFkQ0RBMNkQURBMNkQUBJMNEQXBZENEQaTW2YjIswB+C8A1VX1//7ZJAN8AcATABQC/p6r+OhYAekkPa5t2/cVkZOe+diel1kX8vDkyYvchKTjLWABA0V/xBD2x63SiXEqPkzW/p0xm1F+Co+P0ABoZ9etVjs5Pu3Fx2sp0V/y6p1Lk96TRyO5J01qx60kAIJ/z66IOjk/YwdivPaqt+TU87/3AQTOWwO9HMzLu11TluhU3Xlu390sVTk8lACNTdj8nAKjE9vZrjXD9bP4CwOO33PY0gBdU9UEAL/S/JyIypSYbVX0RwK0p9UkAz/W/fg7Ax3Z4XES0x9ztZzZzqroIAP3/Z3duSES0F93zD4hF5ISILIjIQqPm/51PRHvX3SabJRGZB4D+/9esH1TVk6p6XFWPF8v+h1hEtHfdbbJ5HsBT/a+fAvDdnRkOEe1VqclGRL4G4P8CeK+IXBKRTwP4PICPiMg5AB/pf09EZEqts1HVTxqh37ibB1TYdQ4qdp1CIn7NyErH/zyou2HXKNRqfg8TSfw//8rjds+ascKku21b/DqcxC+PQClv10B0UvqjNBt+vcrfb9j3vVJdc7dtFPzXa7xo77PVyO+F0+n6dR+FnL1mVdTz+7pMzdo9fABgcszu7XJjyd8n+8am3HgvpU6nWLaf99Gj/nEWR/4xXNu03z9xN2VRqgGxgpiIgmCyIaIgmGyIKAgmGyIKgsmGiIJgsiGiIIIu5dJNgNWm3TJhrmxfgp+ktGLIFfzpuWLOXoIjqvhtBbJ5fzfNJPYyGHn1566vbPjT7rrmT+lX5uwpzwMVfwmOJOvvs4uX3jJjhcgfV1H91hhjZXtqPG7b7UAAoFH3p+zjpj2FXMz4+6So/rT61TfeMGObTf84mpnxywHqHf842zdhX4KoTbtlBwBsrvtxbyWkaNRvjTEontkQURBMNkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFEbTORhNFp2bXMZSd+fzimH+JfKtt14QAwJWlJTN2Y8NuSQAAhYy/m9pdu84GkV9Hs379hhtPin69ynTLXjKldMhvO9BNactx8ardMiEPvx4lm9IBdnXEroXJi7+/NfJrZWTTbtshUyntEqr+MjKSsY/DTNlfvubSdf8YbdT9FhOFvL1Ejeb946Q94u/TkaK9X/KxXaN2J3hmQ0RBMNkQURBMNkQUBJMNEQXBZENEQTDZEFEQTDZEFETQOhuBIAt7Pl8yds+aOLaXgNmK+z1QWpHdA0Xzfp8RgV8/sbFu9zHJ5Pz7zhz0x91N6XEyMn/AjJULfn3EpZpfX9Qq2fUsvbZf6zLrLG8DAGMj9us5PZXSP6XoF/FcvnjdjPUy/nG00fX7Jo2N22MrF/xasI76NVeNXsuNj+ScvkkZv86m1vT3WWfT7nejsd8LZ1A8syGiIJhsiCgIJhsiCoLJhoiCYLIhoiCYbIgoiKBT3wCgTvuATteObTb8vNhzpu4AII7sqcEkZRq31vTbEmys29OpuYLd7gAAqht+q4ZGy5+KzR46bMaKOX8ZmaTpT8WW4nEztrZ82d02U6q48amcHe91/Ne62kpZ/kbtcoKxfdPutld+ds6Nry+um7F83i+RmKn4U/pxzm8xUb+xbMdif2o7TvwSjPyYHZfIaaFyB1LPbETkWRG5JiKnb7rtcyJyWURO9f89sSOjIaI9a5A/o/4CwOO3uf1Lqvpw/9/3dnZYRLTXpCYbVX0RgN++jIgoxXY+IP6MiLza/zPL7FcoIidEZEFEFlr1lF6RRLRn3W2y+TKAYwAeBrAI4AvWD6rqSVU9rqrH8yX/2hEi2rvuKtmo6pKq9lQ1AfAVAI/u7LCIaK+5q2QjIvM3fftxAKetnyUiAgaosxGRrwH4VQDTInIJwJ8C+FUReRiAArgA4A8GebBOp4srl+0lVQ5P2/Uupaz/J1gJfh3BgblZM9aK7CUyAKC96V/6P96y21esd/wanSs1e38AQDnrb58dt+tVMpG/z5ZW/GVkmh27nqUy69eMrDltNwDgzZxd97Sx7Le+qDrjAoCxQ/vM2GTWb32RnX+PG887v55bLf85NyO/5irT8+ui1jr29tG6X6MzOenXAOXVWaIm5y8JNKjUZKOqn7zNzc/syKMT0bsGL1cgoiCYbIgoCCYbIgqCyYaIgmCyIaIgmGyIKIig/WwiiVDJ2MuLNDY2zFh92a8jeM/77CVNAKAjdr3Kxav20h8AkO34y3/Um/ZurG36NSHZpt/PJi/+74OVxUUzVtn0r58tdPx6ltHIfj2yKcvbXN+0xwUA9apdp9NylvQBgInReTeOpl1z9ZMzF91NKxV/aZ35+2bM2JVlv6fSG03/ed035h8LUrDrcPZN+bVirZTjrNe0j4WxrH/8D4pnNkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFEXTqO4ZiLLangmO1pw6TyJ9WHM/t9x87saf+8ikrVWjXbocAANXONTO2Xq2525bL/jIyua6//T/8+HUz1kiZSk2KfpuI/UX7eXcbq+62ubLf3uLqut3eolb128d2O/7zqmTtliBd8ceVL/stPRLYJRj75ubcbZd+5reguHLNb0FRmbHLDfaX/TYQvZTlb0pl+1jIpLTlGBTPbIgoCCYbIgqCyYaIgmCyIaIgmGyIKAgmGyIKgsmGiIIIWmeTaA+1pl1r0FC7jiCT9+tR3nzLX7qqMnHUjJUy9nIoABDN+LUZ5bZ9CX6zm7KL6369irb81hqFih2/serXJhVafsuDXnPNjNXrfv3P5JhfmxGP2LUy+yvT7rbXluxxAcDGRfu+44LfQuL8uv+8GvcdNmP//DF/rcZq068f6qzb9VoAcHDGXqKmF/ntQjbhv9aFgl1fVBf/GBwUz2yIKAgmGyIKgsmGiIJgsiGiIJhsiCgIJhsiCoLJhoiCSK2zEZFDAP4rgH0AEgAnVfXPRWQSwDcAHAFwAcDvqapbNKKaoKN2HUPSs+fzo55fJ7Be9+s6umr3T7m46u+G++f9XjkPvf+IGbu86C+ncv6S3XsFAHJXU5Z6ie0an9Kk35tlPGsveQIAa84KHiORv7xHpef3ADpUsZceKc/Nutu+UvNrYU6v2/v0xsY5d9uLm/6xMNq0e9Yc+6UH3G0LJX+ffeQD/8yNHzhg13v98MevuNte3PRreNpte4klHbFjd2KQM5sugD9W1fcB+DCAPxSRhwA8DeAFVX0QwAv974mIbis12ajqoqq+3P+6CuAsgAMAngTwXP/HngPwsXs1SCJ657ujz2xE5AiADwF4CcCcqi4CWwkJwG3PfUXkhIgsiMhCO6X0noj2roGTjYiMAPgWgM+qqr0u6y1U9aSqHlfV47k8P48merca6N0vIllsJZqvquq3+zcvich8Pz4PwP8Eioje1VKTjYgIgGcAnFXVL94Ueh7AU/2vnwLw3Z0fHhHtFYO0mHgMwKcAvCYip/q3/QmAzwP4poh8GsBFAL+bdkexRBjNlsx4Re2pwUbdvzy/XPKXwait27GVq37LgtHIHjMAjDTs6eks/NYYmdif0q+U7LYbAFB2lsZpbvrtKzbL/rT81Jw9dX6g7LdqKBXH3XhR7dczHvWXPKnU/Sn7w85U7emL/jTub03Pu/EP9ewlT85fWHC3Xa/6bSB+/9ceceP3OUvFnDl93t02if1SBIj93uvtUIuJ1GSjqn8LwBrJb+zIKIhoz+MntkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFEXQpl572UG3bNS0tZziaMtc/kvWXWylN2PUu1xt+Lcu51/zL9zdhP6d9h/+xu+14yX8Jilm/LcH0uLP8Tc6vKcmN+/s0n7NbOZT2+7VHszm/vqgyWTRjmZTXutXwW0x0rtnx3Af8+p/C1AE3fjmyfz/X1+2aJwC4VvZf68biFTe+2t40Y7ru1yZNFvzliiqwj6Oe//YYGM9siCgIJhsiCoLJhoiCYLIhoiCYbIgoCCYbIgqCyYaIgghaZwNVJB2794uqPaE/MeHXR+Riv9+N1u1+Hrm2XxNS7/h1H0nFrmFQpNSyqN/jJFv0l3pp1Ozn/UsPTrnbjpftWhcA0IpdN1LJ+fv7wH3+8jdZZ2keSfzeRN3pJTc+94Dz2O2D7rYbFX/5m+mqPbZm7L+d3ko5jjrRshu/cNHuxtvK+fddKPvHeLlr101Jye9dNCie2RBREEw2RBQEkw0RBcFkQ0RBMNkQURBMNkQURNCp70QETecS/SS2p9+qTX/Jk9k5e4kNAJibsKc8Vxv+kibtjZTWABl7SjIPZw0ZAKNZfwp5tuC3icjk7dYaRx445G47PuO3iRidP2rGVP02DyM5//UoO0+revav3W0PH/Cf1+gHf9uMFebs5wQAK1fPuvGfnXvdiU672x6b8vdJu2q3kACAiz+5asayRb+FxJFx/zhK1uwSiw781hmD4pkNEQXBZENEQTDZEFEQTDZEFASTDREFwWRDREEw2RBREEHrbKIoRrkyYf9Ay162pJeU3fsuj0668a7TqSHf9OsIxib8x447dq1MGf6l/bK66Ma7iV+nc/i9h83Y9KjdzgMAZvf59SqT7/knZiwSv/VF1PWX1oH8zAwlo/5hmRlNaafg7PJMzl8aZ/bAcTfeq9vPu7Xu33d786Ibx4ZfZ9OqvmXG8uK/lp1cw42P5u06nHZuZ85JUu9FRA6JyA9E5KyInBGRP+rf/jkRuSwip/r/ntiRERHRnjTImU0XwB+r6ssiUgHwIxH5fj/2JVX9j/dueES0V6QmG1VdBLDY/7oqImcB+MsGEhHd4o7+GBORIwA+BOCl/k2fEZFXReRZEbnthzEickJEFkRkoZ1yfRMR7V0DJxsRGQHwLQCfVdUNAF8GcAzAw9g68/nC7bZT1ZOqelxVj+cKfn9XItq7Bko2IpLFVqL5qqp+GwBUdUlVe6qaAPgKgEfv3TCJ6J1ukNkoAfAMgLOq+sWbbp+/6cc+DuD0zg+PiPaKQWajHgPwKQCvicip/m1/AuCTIvIwAAVwAcAfpN1RAqDulEi0svafWTm54d53lPh/omVh91/JVfy6jUre7lcDAPUr182Y5u3laQAgiv14rr7qxtcXz5ix5n5/SZT8MbtGBwCS+jkzlq084G4rWHPj3eWfmLHCuL/ETGH0X7rxXmS/nknjmrttVPR70swc/qAZayz5fZE2bviPXe+cd+PlCbuO59qKv7xNo51ynOWdz1Pzfp3ZoAaZjfpbALd7lt/bkREQ0bsCL1cgoiCYbIgoCCYbIgqCyYaIgmCyIaIgmGyIKIig/WwkihGN2P1sSok9nHI85d53N+v3pFm7aNfCdGv+NVsZ8Wt4JmftOoRM5Pd1SZK6Gx9PWe+n2bZ7oCyv+XUdMyt2TxkAaOTtsU2V/HWK4sy4G2+LXfeRmfmou61U/DqcKLKPsaTn11Q1Nv1amW7DrtfqwF8DDFn/tYzG/J5MYxm7zqaT8c8bWhv+Md5J7JqsbOz36RkUz2yIKAgmGyIKgsmGiIJgsiGiIJhsiCgIJhsiCiLo1Hc3Aa5X7anH2Z491bra9VtMTKYsHbJ/fsaMXVu1WykAQN2fnUYta+fs2Wm/XfP6DX/KvjXhP6+WjJqx1cifsry65C8TMzXzoBnLrfqtL6Kc/7zyxfeYsV7KtHoHKR0fM/Zr3en6S+dsOFPbANCuVc2YqL+/OyX/7Zad8ae+G027bUci/rI9mvhja6g9Nd7e8FuVDIpnNkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFwWRDREEErbPJRFnMje4z43PZvBnrdv2WBb2c3WoBAOKJOTNWecCvQeht+LUXcZQzY9Mz/tIgY5P+Ehuzeb9+ohXbNSfjY/b+BIDCgUfcuIrdEmGl6e+zSsryH5mRg2YsyfrPOYrt/Q0ASWy3oKjFdg0OAHRH/d+/naxdj5L0Gu62Tb+sCTdW7TYoANCo2MdwTv1xF8WvldnctIvJtOUvZTQontkQURBMNkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFkVpnIyIFAC8CyPd//i9V9U9F5CiArwOYBPAygE+pqtvERESQEfshm4ndKySX8XuYZFLqOqrry2asi5a77chMyY0nRbvuo5nSHyU76993teWPDaP2ftko+EuHFIv2/gaAOuzntX/arpMBAOT8Pjy12K6L6qUst7Jv1H/sjlPQstx80912xHnOABAX7fqi2rLfc2kl8QttWh2/FubgYbs30uVTV91t31xdcuMry3aN0Oio/1oOapAzmxaAX1fVXwbwMIDHReTDAP4MwJdU9UEAqwA+vSMjIqI9KTXZ6Ja3fw1l+/8UwK8D+Mv+7c8B+Ng9GSER7QkDfWYjIrGInAJwDcD3AfwUwJqqvl1XfgnAbc/xROSEiCyIyEK77reKJKK9a6Bko6o9VX0YwEEAjwJ43+1+zNj2pKoeV9XjuZL/9zAR7V13NBulqmsA/jeADwMYF/n5p70HAVzZ2aER0V6SmmxEZEZExvtfFwH8CwBnAfwAwO/0f+wpAN+9V4Mkone+QVpMzAN4TkRibCWnb6rqX4nI6wC+LiL/HsCPATyTdkea9NBq2perr8ButzDidzTAct2fLl1M7KVHeikrVVTFnyIu5Owp5mrX3/bQlN/yoB3506XjWfuxyzm/HCC36bdEyMb2a1Wsddxtq+v+uEXsF7QX+b8DSwV/eZzqpj2tvla3209sxf1Sg0LWLlV4/S17qRUAqKV8Zrna8JdymVyx367nlv02EI2UpY66sMfWqPstPwaVmmxU9VUAH7rN7W9g6/MbIqJUrCAmoiCYbIgoCCYbIgqCyYaIgmCyIaIgmGyIKAhRve1VBvfmwUSuA7j5Gv9pAHbvh+HZreMCdu/YOK47t1vHdqfjOqyqfsEYAiebX3hwkQVVPT60ARh267iA3Ts2juvO7dax3atx8c8oIgqCyYaIghh2sjk55Me37NZxAbt3bBzXndutY7sn4xrqZzZE9O4x7DMbInqXYLIhoiCGkmxE5HER+XsROS8iTw9jDBYRuSAir4nIKRFZGOI4nhWRayJy+qbbJkXk+yJyrv//xC4a2+dE5HJ/v50SkSeGMK5DIvIDETkrImdE5I/6tw91vznj2g37rCAifycir/TH9u/6tx8VkZf6++wbIrL9nr6qGvQfgBhbDdPvB5AD8AqAh0KPwxnfBQDTu2AcvwLgEQCnb7rtPwB4uv/10wD+bBeN7XMA/vWQ99k8gEf6X1cA/AOAh4a935xx7YZ9JgBG+l9nAbyErba/3wTwif7t/xnAv9ruYw3jzOZRAOdV9Q3dWtTu6wCeHMI4djVVfRHAyi03P4mtZXOAIS6fY4xt6FR1UVVf7n9dxVb72gMY8n5zxjV0uiXIUk3DSDYHALx10/fmMjBDogD+RkR+JCInhj2YW8yp6iKwdQADmB3yeG71GRF5tf9n1lD+xHubiBzBVofJl7CL9tst4wJ2wT7bzlJNd2IYyeZ2zWd30/z7Y6r6CICPAvhDEfmVYQ/oHeLLAI5ha9XURQBfGNZARGQEwLcAfFZV/ea8Ad1mXLtin+k2lmq6E8NINpcAHLrp+121DIyqXun/fw3Ad7C7+iwvicg8APT/vzbk8fycqi71D9oEwFcwpP0mIllsvaG/qqrf7t889P12u3Htln32Nr3HSzUNI9n8EMCD/U+7cwA+AeD5IYzjF4hIWUQqb38N4DcBnPa3Cup5bC2bA+yy5XPefjP3fRxD2G+ytWTDMwDOquoXbwoNdb9Z49ol+yzcUk1D+gT8CWyIaPznAAAAmUlEQVR9Iv9TAP9mmJ/G3zKu+7E1O/YKgDPDHBuAr2Hr1LqDrbPBTwOYAvACgHP9/yd30dj+G4DXALyKrTf3/BDG9U+xdbr/KoBT/X9PDHu/OePaDfvsg9haiulVbCW7f9u//X4AfwfgPID/ASC/3cfi5QpEFAQriIkoCCYbIgqCyYaIgmCyIaIgmGyIKAgmGyIKgsmGiIL4f4zhwFbeRpzXAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARsAAAEYCAYAAABsuVKPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3XuM3WeZH/Dvc65zztwvHnt8i2PHuQIJwZsN4tIFym6gKwWk7Rb+QFGFFFSBukjbPyiVurTqH+yqgPaPFVUo0Wa3lEsXWKIV3UJZKLCFgBOMncSbxHZ8H8+Mx577nPvTP+a48gY/3zO2J+8Mk+9HsjxznnnPeed3fvPMb877nOc1d4eIyKsts94TEJHXBiUbEUlCyUZEklCyEZEklGxEJAklGxFJQslGRJJQshGRJJRsRCSJXMoHK5QzXhrIhvFyV1cYK+bicQCA1jINmzVIjN+3dThM1UY1jOUzRTq2k0bT+BcYiRuft3e4a/dWfNeIjycAZNi8ALRa9TCWz/Dnw5u86r1ej+fWqPOx2UyexsvFcvy4LX7fbvH3DABdHX4a85n42iCXK9Gx9WqFxpfn9sXB3HN07MmJ2kV330K/CDeZbMzsIQB/CiAL4L+6+6fZ15cGsnjw0aEwfv+dd4ax24f66Fway/yA5PNTYayYG6Rjc5kRGj958XgYGy3vpWOd/8xiZq7DU1Qk8fwoHVrJ8h+OZi1O4BlM07F9RX7RvLQ0Hsa2lvnz0Zip0fiFC5fC2MUJ/kupr7yTxu+77d4wNrnIk0klM0njdw/HyR0Axvq6w9jw0Ovo2Asvv0Djv/zO18OYb3k9Hfsv/+TEKfoFbTf8Z5StXA78GYD3ALgbwAfN7O4bvT8R2dxu5jWbBwAcc/cT7l4D8BUAD6/NtERks7mZZLMDwJmrPj/bvu0fMbNHzeygmR2sLfHLRBHZvG4m2VzrFcBfeRHA3R9z9wPufqBQ1uKXyGvVzfz0nwWw66rPdwI4f3PTEZHN6maSzc8B7DezW82sAOADAJ5cm2mJyGZzw0vf7t4ws48B+F9YWfp+3N3p+nMhV8DuwT1hvHYxHjudKdD5DNgtNF6obg1jC5fPhDEA6I5X61dMkRKDoWE6tD7Dl5CLHZaBMR/XT2QH+NPbWF6i8a5sXCPUqMQ1UQBgzus+qhcvh7FKTy8dOz8VlxoAQGM2roXJVfvp2MXFHho/a/HztVTiNTrnjddc3TbGz+Fi761hLE9qogDAz/NzYai5O4zN1+OfnetxU3U27v5tAN9ek5mIyKamV2xFJAklGxFJQslGRJJQshGRJJRsRCSJpC0mcpkctnbHS8GDC+RdxtN8iXjvGH9ndhMTYaxrkS8btnKLNH5rrhnG6pfO0bF9Vf7YdczQeIu0DmiSpWsAyFf4u6eRid/FXCXv2gaAnmXeJqK4EN93X42/G71xirdLyOXiDgGe4cf71DR/V3iuEc9tljwuAEyM8fPohVOHaHy0ED9231C8dA0ArT5+XTF4V7xs39PHz8HV0pWNiCShZCMiSSjZiEgSSjYikoSSjYgkoWQjIkko2YhIEknrbFoOLMYlKdi3dVcYy1T4NgSFLH97f2s5bluQ38E70zezvGt+vhx3vV+e7ZDPm7xmZL7CW2v098fHZY5PG10l3sqhi2yf46S2CAC8xrdy8Z64xqdU4e0pagN8B4TFmfiYTy51aKvR4s/X8lL8IzNrfPeE3Ai/76bN0fji+R+FsfzQO+nYHXtvp/GpmZfD2IXptemJpysbEUlCyUZEklCyEZEklGxEJAklGxFJQslGRJJQshGRJJLW2Sw16vjFVNzf5S7EW3D0DPItTUod6j6aubjuo57hNTy5Bq/hWV6Mx9eafMuTrMXfMwBkcrz/SqEU16T05MkWMwCWqrM07oi3W8mUeB0NcnxLlMZ8vH3OUiV+XAAw5/1unGxrkm3w49lf5t/Xucn5MDZf5zU885m4HgsAZoz3jZnqJ3sdVR6hY3tG76XxpYH4HO5r/IyOBfZ1iK/QlY2IJKFkIyJJKNmISBJKNiKShJKNiCShZCMiSSRd+s5mcujr2RrGL2UGwli9xpeI5zts/9Fqxi0NCi1+30vgy+pLtfi+L1f5snl/iS+HVltVGp+vxk9hp61cmuDxhWq8HNrTs52OXazybWJa8/GyfMEX6NhSzxCNT42fDmONef5cV6vx0jYAjJHz1yp8bFcf38qli5QaAEB/IV62z+TjrYoAIJvlj10uxefp0Og9dOxq3VSyMbOTAOYBNAE03P3AWkxKRDaftbiyeYe7k2ojERG9ZiMiidxssnEA3zGzp83s0Wt9gZk9amYHzexgbbFDn0oR2bRu9s+ot7j7eTMbBfBdM/sHd//h1V/g7o8BeAwABnb18VdxRWTTuqkrG3c/3/5/EsA3ATywFpMSkc3nhpONmXWbWe+VjwH8NoBn12piIrK53MyfUVsBfNPMrtzPf3f3v2UDMpZH2eL6inPn4y0j7t91C51Mf5lvS1Isxo97qc5fS+pu8O1Wms24RmGwwOts0Ii3SwGArjpvO7BscQuLTq0acuA1PpnmWBibzfDvqzLOFygbmbhlSF+F3/eFJf7X+NJS/Dt0Kct/vy5MdWhfQVp+9PZ32PLHec3UrR3apPRkzoaxyotfpWMzez5E4/mB+GdvaJC3CwH4z8cVN5xs3P0EAN4kQ0SkTUvfIpKEko2IJKFkIyJJKNmISBJKNiKShJKNiCSRtJ9NBk2UcnHPj5LFPTeGh/gWHO7LND5++vkwdulih14gpR38sfvjmpFCcZSOPXn2FI33dffTuC1eCmOTU7zWZcnibWAAoNgV15zkZnm/mnx3hz4+pD6o3uLP9ezEHI33DPaFscFpPq/5ef585PrjfjjFfn48+zr8tI318u2Kto0cD2OZeV5P2zz6BI0X9t8exrL9O+lYYHVNH3RlIyJJKNmISBJKNiKShJKNiCShZCMiSSjZiEgSSZe+6/Uazp87E8b/yT37w9juDi0kGlN8+RrTcbw+zZdSZ3K8DUTzbNwGYmQ3z+czx0/S+HKH7VjKuXgptqvJx5by/PuqL86GsWaJ3/eOgb003szEZQ6FRd7mwcp8W5K6xcvbx8n5BwCzMx22kfH4PBqqxkvuADBSv0DjXXPHaBxdR8JQocHbiVgtLpEAAKs/HcZy5Xfzea2SrmxEJAklGxFJQslGRJJQshGRJJRsRCQJJRsRSULJRkSSSFpnk4WhPxM/ZKEe10eMX+A1CLXpSRrv2XZbGGtVT9Kx9ZkCjS814vYWtYm4VgUAZuZ4jc9Qb4PGe0rx2//LPXwLjt1v2E3jlUr82LUGr0fpKvEWFJPPHSZRft/bh/jzcfZsvCVKZYlvjVPcNkTjw4W4NqnV4PedqfDWGYU+3mKiZ1vc6iS3vETH+ileZ5Op3hGPbfEtaFZLVzYikoSSjYgkoWQjIkko2YhIEko2IpKEko2IJKFkIyJJdKyzMbPHAfwugEl3f137tiEAXwWwB8BJAL/v7ryhBoBWxlArxTUSy/XpMNZb4XUb5SzvdzNUjvu+LGT4fddbfJuYFuK6jswy77OzcO5lGh8Yo2HMVabC2Niee/h9l7fSeK073nbn0pm/p2Mzl+KxAFCqx6dLocCfy2KV/45czMS9dm7ZymuPCh2OSbYenwtdw3ybmPJ2GoY7r7mqXjofP3aNn6PZEd5fyLvI8+V8exuA1z1dsZormz8H8NArbvsEgO+5+34A32t/LiIS6phs3P2HAF5ZfvgwgCu7Xj0B4H1rPC8R2WRu9DWbre4+DgDt//m2jyLymveqv0BsZo+a2UEzO1hbrL/aDyciG9SNJpsJMxsDgPb/4bsg3f0xdz/g7gcKHfZ/FpHN60aTzZMAHml//AiAb63NdERks+qYbMzsywB+AuAOMztrZh8G8GkA7zazlwC8u/25iEioY52Nu38wCL3reh/M3JCrxQ85NRX3+xjZzXt9VOd5jcKJZ58JYzM1Pra0rZvGL5yJ64P6jO/NtP/uuEcJAOzexse3zo+HsUaHPj1LU3zvp8JAXO/Sxw8JMnl+amUtjne3eC1MrZvHM8MDYWzLzi469uQE7wszPx6fK+VuXm/StXWExnND/BzPluJCncwyrxXzDrUwlidrPNlhOna1VEEsIkko2YhIEko2IpKEko2IJKFkIyJJKNmISBJJt3KxTBb5Un8Yzy3HS99z5/mS5GCeb5OR64qXFYfyTsf2lfn2Htnt8RJxbZkvtZaaZ2i8MnWOxgfK8TLwQD+v2B4c4Evfl87+Mozls3xJvmcw3mIGAHJT8TYxvTu30LH1Rf7Y44ficoDWctwOBAByuV00PjASn79bdtxKxw6P8jYQ5Z49NG61p8KY5/m2PNnSPn7fg/HPlxX5+Q/w7Yqu0JWNiCShZCMiSSjZiEgSSjYikoSSjYgkoWQjIkko2YhIEknrbJrNJi4vxFtGXK7FuS9fL9H77t/Lay+8Gdd1NFsd2pX28vsuF/vC2PYuXnsxeTreigUAcrUObQv64++rPBq3WgCA4qjxeC3e1iTTWqBjkeG1Mq2h42GsusRrixYv83Nh6A1x7dHEWX68p47xxx7csieM9e7mdUvF/gqNZ3v5FjZd3W+Lxy6/RMfiMq/38np8rjQX+XZEq6UrGxFJQslGRJJQshGRJJRsRCQJJRsRSULJRkSSULIRkSSS1tlUG3W8PBH3Grlny21hbKGH9yGpNeJ6EwDo7YnrPrKFuEcJAFiL19nMLV0OY4NlXsOTaXbI9znea2egFNfhdJV5jU4mt43G64W4T4kt8l45F078hMYXF58NY83zl+jY4X0P0vjO/e8LYz7I+wf17+Dn0Z2vf08Y693G65qW6y/SuM/8NY03WzNkcFwTBQCZft7Pplk7HMZa86foWCDeYuYfzWFVXyUicpOUbEQkCSUbEUlCyUZEklCyEZEklGxEJImkS9/uWTTqcTuGiWWS+wpx2wAAePEcX9Lc1RM/bhEdWi1k+TYxxXK87Fgaih8XAJoXeNuBDPgSM8bi5euW8e09rLyDxgfIima9Gi/3A0B++y00vlB9OIxVTzxPxw7t/x0az43eHsa667yUwBZ564y58fj7zvbw7YZm6nwrl9ISX3Yf7orPM8+cpWMb48dovMXaqJTX5pqk472Y2eNmNmlmz15126fM7JyZHWr/e++azEZENq3VpKw/B/DQNW7/nLvf1/737bWdlohsNh2Tjbv/EAAv6RQR6eBm/hj7mJkdbv+ZFe5ta2aPmtlBMzvYqvC/SUVk87rRZPN5APsA3AdgHMBnoi9098fc/YC7H8h0JX09WkQ2kBtKNu4+4e5Nd28B+AKAB9Z2WiKy2dxQsjGzsas+fT+A+C28IiJYRZ2NmX0ZwG8BGDGzswD+CMBvmdl9ABzASQAfWc2DedNQX4jX83MjcU1Jb7ab3vfCMm8TUcjGRSO1LK91aTX56+NDfbvCWF8fbzuwc/dv0Pj4kZ/TeLa1J4zVynQolmY7bCNTjudeLt9Jx7Za/JjOnZoMY339b6Jj97/5P9J4pjAcxqbnv0TH/ujJr9L4jtH4dcedC/wcHL2LhjFU5jVV1dmDYazcy+uabOteGq/X/z6MtTL8PFmtjsnG3T94jZu/uCaPLiKvGXq7gogkoWQjIkko2YhIEko2IpKEko2IJKFkIyJJJH3/QH9PCe958A1h/O2jca1APc/72Uyc4Vu9TFyOtyW51HiBjt2a4fftjfkwVqyV6NjlCu9DUugt0nhpNO5nk+kK37IGAGjk42MCANVq/H1ZlteE5Lt4vDwQP5/Fnp10bLYr3pYHAJDpCkM/+wXfTuVff+bvaPzffvRfhLHC/g6FTXxXH+wf5oU4xdqbw1iLt8oBzvEeQa2e+Pnw8oUOd746urIRkSSUbEQkCSUbEUlCyUZEklCyEZEklGxEJImkS9+lfAFv2BlvH5JHvH43fuYcve9G9TyNL1YWw9jC5bjdAQBsGRqi8WYr3kbm/DG+bJ6v821ktmzhrTUK1XiJ2fr52HqG/67xfNwOxIxvb5MBbwFrpCVIrpeXC3iDr/M2KvF2K1/8O96y46FP3k3jt3/8Wk0QVrz9dXvo2GJvXEoAAKjxZfdMJj6HrUa2YgFgw7zFRK4rLgmoZebo2NXSlY2IJKFkIyJJKNmISBJKNiKShJKNiCShZCMiSSjZiEgSSetslqs1HD5+OoxvycW1AmdO8lYMvc63DlmejbdjGZ/gW7WM5qs0niElJcuVcTrWq3zeua57afz23fGWKtbLW0xMzZ/gj52Nfxdli7z1xdSp/0PjPz/0f8PYmw68n46dmz5G47PTE2Fs907evuL33jZK4/9sdiaM1Su8HqV7gNcP2XKBxy0e7wtxrRcANM7x/ha+M64H827+XK+WrmxEJAklGxFJQslGRJJQshGRJJRsRCQJJRsRSULJRkSS6FhnY2a7APwFgG0AWgAec/c/NbMhAF8FsAfASQC/7+5xIxEAi/OL+MkPngnjb7snroEoWod+HQ2eN7f3xn10Bgu8jiBnvEZhai7utTOcjfu2AMCy8e0/XveOj9D4yB2/G8Yazbj/CQBcNr49jlfimpLpS7yu4/RZp/Hs6FvDWKXO+/D89H/+GY3n+m4PY/ffy+uWhrMv0fhs82QYm5/lz2X/8AiNZ/K85qoxHR/zfKZDjU6pRuNej7f1se64bul6rObKpgHgD939LgAPAviomd0N4BMAvufu+wF8r/25iMg1dUw27j7u7s+0P54HcBTADgAPA3ii/WVPAHjfqzVJEfn1d12v2ZjZHgBvBPAUgK3uPg6sJCQA16zzNrNHzeygmR1s1HmrSBHZvFadbMysB8DXAXzc3VfdlNTdH3P3A+5+IJdP+lYsEdlAVpVszCyPlUTzJXf/RvvmCTMba8fHAPCu4SLymtYx2ZiZAfgigKPu/tmrQk8CeKT98SMAvrX20xORzWI1f9e8BcCHABwxs0Pt2z4J4NMAvmZmHwZwGsA/X80Dsi1AqjXyNvdFnheXOix9bxuIv9VF3kECk5N8m5gXT8fxvjJfctw1xJdDe7bHy7gr4u97em6Kjpxe4tvMjHQNhLHuEb5sfs+bb6Xx4pkXwtjkad764tiRaRpf6DkexrzAt+Xxi3yLGhuJSxksw0skLi6d4o89e5LGszNxKUJxkZc5tCa30Hi9Fm8zkynyrXNWq2OycfcfA4g2N3rXmsxCRDY9VRCLSBJKNiKShJKNiCShZCMiSSjZiEgSSjYikkTS9w9kDChZtIoOTFyItz0pGn8L/f49cQsJADg/ezGMnThNO2Mg283bJUyT2o3THteTAMCt991H47Um3/7j7KlDYexvXniOjsVAnoZ/55bfjIcObaNjK8vx8QaAv/6Lx8PYS0fi7wkAHnz7O2h85kJcc/L8S7+gY3PzvL3FHUPxj8zCxT4+r1rcigQAyksXaLy3Eo9vTvM6myypMwOA/Fj8fVfzz9Kxq6UrGxFJQslGRJJQshGRJJRsRCQJJRsRSULJRkSSULIRkSTS9ulstZCrxL0xFpfi7VqKwzwvFptx/c7Kfcd9ZeZneA+T7X27aLw2tBTGdm59Nx37wOs+QOPTUws0/o3v/O8w9t9O8v4p+7bz2qTmG+Purw/eej8d22FXH4yM3hPGym+6m44tk75HAHDkXFwX0rrM+/AMZfm5MHMxfuzjL/2Uju0e7aLxPbl4OxUAyA+Q7VbAt6BpHH+expuX4xoe27s21yS6shGRJJRsRCQJJRsRSULJRkSSULIRkSSUbEQkiaRL3616E/MT8fJdtUq2W5nj22TsGuTx199zbzx25y107MwsX8YtI357/7t+8zfo2H17D9D4iefjbUkA4FtHXg5j52p8KbXRw9t2tP42fuyjW/hS6p17+BY0t+/fG8b6y2N07FM/+C6Nd7Xic2FwDx2K7DneTmSsFO/7U65V6NitfFUd3Vm+zUxu7mdhLN/Hl9UbdV4u0PJ46dyzvH3FaunKRkSSULIRkSSUbEQkCSUbEUlCyUZEklCyEZEklGxEJIm0dTYOVCqk2KAV1zBUO9UJNOLWFQCQI+MztfhxAaAyH7enAIAabZvBiyvm5ngNw5mZeRpfKIyGsUHwbWB2NflWLsPZuHajvMx/T/UZfz7GmnEtzHBvLx1b2ce3kRmbjo/ZiYsv0rGZygyNDyzHx+T2kRE6Nj/Iz+HuYoPGS4N3hLFsh21iWscmaDzTtzuM5XOTdOxqdbyyMbNdZvZ9MztqZs+Z2R+0b/+UmZ0zs0Ptf+9dkxmJyKa0miubBoA/dPdnzKwXwNNmdqWE83Pu/p9fvemJyGbRMdm4+ziA8fbH82Z2FABv8SYi8grX9QKxme0B8EYAT7Vv+piZHTazx81sMBjzqJkdNLODzWaHN4eIyKa16mRjZj0Avg7g4+4+B+DzAPYBuA8rVz6fudY4d3/M3Q+4+4FsVotfIq9Vq/rpN7M8VhLNl9z9GwDg7hPu3nT3FoAvAHjg1ZumiPy6W81qlAH4IoCj7v7Zq26/ug/A+wHELe1F5DVvNatRbwHwIQBHzOxQ+7ZPAvigmd0HwAGcBPCRznflgMc1K8VcHMvxNiM49vwRGr944Wh83/lrvtz0/5U79ArpLsXbgxw/zus6Ls/xHijNFt+iZsdgHOcjgV0damH2jcT30LrMt4npvszrhwo9w2HsUv1HdGxr9jSN9/XFx/z2JX7K58d47dFQd7zdUHnhGB07sIvXBxWz8dY5ANA6R86l6jgdm93DtyOyAXKO7+T1Wqu1mtWoH+Pa5+2312QGIvKaoFdsRSQJJRsRSULJRkSSULIRkSSUbEQkCSUbEUkiaT+bjGVQ6orX7AvNuDaj1YjrGwBgPs97ztTn4p41W0u8jsCLfE+qYil+7OU5ft9Hx/n+S9vHdtL4u3bH9SrIxb1uAODS8XivIAAAqZWxy0t0aNc+/j64bCvuv1LlLXww3GGPsK5cXxiby56hY7uNVydlinG/m3I3n1d3P68PIj8aAIBshvTaWeT1WrbMe+nYAPm+jffZWS1d2YhIEko2IpKEko2IJKFkIyJJKNmISBJKNiKSRNKlb4ehavFb+Ivd5TBWAF/athZfdswU4se9uHSBjq3V+NYiaMXLwMO7O2wDM8OXkKs9/CkaLcfxiUXeQmLy7HEan2rGx2W4wOdVaWyl8fEzL4exsx36ieTHD9P4jm1xucByjZ8ntV6+9N2Yjpevd+zibRyWaidoPL/8NI0XKz8NY9lqgd/3CL+u8O59cdDiUoLroSsbEUlCyUZEklCyEZEklGxEJAklGxFJQslGRJJQshGRJMy9wx4pa6i3t+z33X9HGC9avJVFbz/fYiPrcQsJAKjm4vqJriyvhcnV4q1aAKCZvRzGysb7BrD6HwDo6e2m8XJ33GIiP8y3oCk4bzvQU45bCxQ8rokCgO07+bz782R8h5YflcWzNN4zGLcjqS3y77l3gLcy6SrHc+vp5b0xshafJwDQmvsSjQ/1jISxTOUf6Fhr3EnjjWLcTqSw5a10bM/2F5529wP0i6ArGxFJRMlGRJJQshGRJJRsRCQJJRsRSULJRkSSULIRkSQ69rMxsy4APwRQbH/9X7n7H5nZrQC+AmAIwDMAPuTutGClXC7ijW+6NYxPnIp7nJy/PEnnWe7vp/FiOc6rlQav6yhkeC1SoxbXjFya5Fts7N7H+5AsOv99UFuKazv6Omxv09vPv+9aJX7s/kF+TJac9+mp1C6GMZ/hPWesj9cPLU2Oh7FWH6+ZunSB9zbauvfuMDYzeYSOHRjm/W7mzl6i8dKde8OYz/yYjl2YfJbf9974vvuzcWzFCx3iK1ZzZVMF8E53vxfAfQAeMrMHAfwxgM+5+34AlwF8eFWPKCKvSR2Tja9YaH+ab/9zAO8E8Fft258A8L5XZYYisims6jUbM8ua2SEAkwC+C+A4gBl3v1LPfhbAjmDso2Z20MwOLi/ztxSIyOa1qmTj7k13vw/ATgAPALjrWl8WjH3M3Q+4+4FSqXjjMxWRX2vXtRrl7jMAfgDgQQADZnblBeadAM6v7dREZDPpmGzMbIuZDbQ/LgH4pwCOAvg+gN9rf9kjAL71ak1SRH79rWYrlzEAT5hZFivJ6Wvu/jdm9jyAr5jZfwLwCwBf7HRHmWwW3f0DYfyOu+NlxdYp3rIA4Eut1iLBPF/GrVX5EnLD4j8PpxbjthkAMDDLJgbMXYyXiAFgcDDeZuPE4hQde0uNb9FRL8dL+jP5GTq2F3xJvzsftzQ4U+Hf81bn9z1l8blw11b+XM9n+fOxPR//fn5xcZaOHVvgLSbO5Mdo/Lbu28LY0tCb6diZC3ybmBPnH4wft7U2r7V2TDbufhjAG69x+wmsvH4jItKRKohFJAklGxFJQslGRJJQshGRJJRsRCQJJRsRSSLpVi5mNgXg1FU3jQDgRRXrY6POC9i4c9O8rt9Gndv1zusWd9/S6YuSJptfeXCzg6vZbya1jTovYOPOTfO6fht1bq/WvPRnlIgkoWQjIkmsd7J5bJ0fP7JR5wVs3LlpXtdvo87tVZnXur5mIyKvHet9ZSMirxFKNiKSxLokGzN7yMxeMLNjZvaJ9ZhDxMxOmtkRMztkZgfXcR6Pm9mkmT171W1DZvZdM3up/f/gBprbp8zsXPu4HTKz967DvHaZ2ffN7KiZPWdmf9C+fV2PG5nXRjhmXWb2MzP7ZXtu/6F9+61m9lT7mH3VzHgTodVw96T/AGSx0jB9L4ACgF8CuDv1PMj8TgIY2QDzeDuA+wE8e9VtfwLgE+2PPwHgjzfQ3D4F4N+s8zEbA3B/++NeAC8CuHu9jxuZ10Y4Zgagp/1xHsBTWGn7+zUAH2jf/l8A/Kubfaz1uLJ5AMAxdz/hK5vafQXAw+swjw3N3X8I4JW7lj2MlW1zgHXcPieY27pz93F3f6b98TxW2tfuwDofNzKvdecrkmzVtB7JZgeAM1d9Hm4Ds04cwHfM7Gkze3S9J/MKW919HFg5gQGMrvN8XuljZna4/WfWuvyJd4WZ7cFKh8mnsIGO2yvmBWyAY3YzWzVdj/VINnaN2zbS+vtb3P1+AO8B8FEze/t6T+jXxOeAdXC3AAABXklEQVQB7MPKrqnjAD6zXhMxsx4AXwfwcXfnTaATusa8NsQx85vYqul6rEeyOQvg6k2PN9Q2MO5+vv3/JIBvYmP1WZ4wszEAaP/PN0BPyN0n2idtC8AXsE7HzczyWPmB/pK7f6N987oft2vNa6Mcsyv8Vd6qaT2Szc8B7G+/2l0A8AEAT67DPH6FmXWbWe+VjwH8NgC+I3taT2Jl2xxgg22fc+WHue39WIfjZmaGlV0+jrr7Z68Kretxi+a1QY5Zuq2a1ukV8Pdi5RX54wD+3Xq+Gv+Kee3FyurYLwE8t55zA/BlrFxa17FyNfhhAMMAvgfgpfb/Qxtobn8J4AiAw1j54R5bh3m9FSuX+4cBHGr/e+96Hzcyr41wzN6Ala2YDmMl2f379u17AfwMwDEA/wNA8WYfS29XEJEkVEEsIkko2YhIEko2IpKEko2IJKFkIyJJKNmISBJKNiKSxP8DYO226LlnazQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAARsAAAEYCAYAAABsuVKPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3WuMnOd1H/D/mevOzl5nd7lc7vIqyYok25EURlGiNo2dxJCFALKR1rA/GPpggGkRAzGQfhBcIHGBonCC2kY+pC7oWo3aOr40tmOhUVvbqhPBjiyZkiiJMimRonhZci/c5V5m5z7znn7Yoc3KPOcdcpfPrFb/H0BwOWffd555553Dd+Y5cx5RVRAR3WyJbg+AiN4ZmGyIKAgmGyIKgsmGiIJgsiGiIJhsiCgIJhsiCoLJhoiCYLIhoiBSIe+sP9+jY4V+My7Ots163d95TCV0OpszY6mUfxhq9ZobL5dL9v1m/H0P9NvHAwAk4W+/VlwzY9VG0902nc648UzGjkeI3G1LVf+YVSoVMyZJ/zHnnHEBQLXWMGPlStndthX551HGOWbe+QsAzab/fESRf0zFuYdkIua6Qf19+y8f/5g0W7qgqmP+ADaYbETkQQB/ASAJ4D+r6me93x8r9OPff+r37f05j3jp/Hl3LFHLT0YTB+6yxzU64m57+txZN37kxR+bsakpf9/v/2fvd+OZvlE3/szT/2DGTkwvudtO7Nztxif3TZmxWuQnk+dOvOHGXzlxzIxl+vzz9q49/rhPnJoxYy8ef9HddnXNTlQAMLVrlxlLtvx0s7jkPx/lop8Ik2K/XAv5HnfbqOE/X41ay4yJ+Ilqbrniv0DabvhtlIgkAfwlgA8CuBPAx0TkzhvdHxFtbxv5zOY+AKdU9bSq1gF8DcDDmzMsItpuNpJsJgFc/d5mun3b/0dEDonIERE5UixVN3B3RPR2tpFkc603qL/woYuqHlbVg6p6sD/mfSURbV8bSTbTAK7+pG4KwMWNDYeItquNJJufALhNRPaLSAbARwE8sTnDIqLt5oanvlW1KSKfBPB/sD71/ZiqvuptEzUUlXl7Cq5StGsvFk+uuuOpo9eN9ybtacWhyJ9enhzwp1pXR+3HNL5zp7vtcOF2Ny4te0oSAHQxaY/rdNHdFmuX3fB77vgVM1Zurbjbthbt+h8AWJyza06SS/4UcKbkl0Gcc8okUg3/rXwhP+jGy87YWpFfR1Np+vUqvbm8v33N3v9qxf88VNQ+TwAgofa0fRr+OdipDdXZqOqTAJ7clJEQ0bbGrysQURBMNkQUBJMNEQXBZENEQTDZEFEQQVtM1Op1nDwzbcZPX3zTjCXq/rRha8Cfsly+aD/Uhay7KcbH+9z47Ig9NXi5vOxue0vD/5bx2uV5N/4Pi6fM2EvL9refAWBqdMCNv7tif0u5XPGn1Z+/MOvG55r2VG2y6n+DXyp+i4mly/b0dLbf/2Z2vWSXXwBA5HQm6BV/35WY57rlf7kaSe+L2zGz08l0TAmFc98N8b8x3ile2RBREEw2RBQEkw0RBcFkQ0RBMNkQURBMNkQUBJMNEQURtM5GW4q6070+WbNrGHJ5v/5hturHL6zZtQK1Vb/QZqHgF0BczNqPqVb061F2zb7uxs+8YdceAUBzasiM7Rn261H64T/u50+eNmNV9VsaFJN+S4MdBXvViWjNr6nKpf3TtjdnP1/JRkxr2phlexI5ZykXiWnjENNiIlGNGZs69x3T3qIWU6fmXXVsToMJXtkQUSBMNkQUBJMNEQXBZENEQTDZEFEQTDZEFASTDREFEbTOJmpGKM07S3wk7FoASfiz/bWK3zemuOgsE7Pk90/JXyq48UuRXdfRn/KXifnHtaNuvLriL7eS7bX7+PQ1/ad39bK/HMtri3Y/m1XnuQKAIfHjy2WnNmnJr3Vpxiys2ttj/x+aULsuCQAi9ZeRyabtWKVacreF+ueZt28AqFXs46Ixz4c6NWwAkOi171w3qdCGVzZEFASTDREFwWRDREEw2RBREEw2RBQEkw0RBRF26lsj1Fr29GDUsKeQ6/436FEQf7mV3K4JM9bM+tsurvnToa2iPSWZ3OUvl3J53m+NkWz5846jlUUztrpkT10DQP+AP6XfHLZPj2bJX7akXPfbWzRaq2Ys3e/PAdfcNU2ARN0+j1Ixy62slvyWIDt67LG1/NMEPZF/342qf5Jr5Lxcm/62vWm/nYhEdimCwm+x4h+xn9tQshGRM+37agFoqurBjeyPiLavzbiyeZ+qLmzCfohoG+NnNkQUxEaTjQL4rog8LyKHrvULInJIRI6IyJFqzPtKItq+Nvo26gFVvSgiOwB8T0ROqOrTV/+Cqh4GcBgARntz/hc0iGjb2tCVjapebP89D+DbAO7bjEER0fZzw8lGRPIi0n/lZwAfAHBsswZGRNvLRt5GjQP4tqzXLaQA/LWq/m9vgyiKsFayl6uors2bsTsf2O8OpumXMACJvBmaXfCX0IirUZjY47QtKPoTdZecmhAAGBvwW2dMFXabsdKg//S2an6vhkTJrgEqOPUmAHC27tcHFRt2XUc+8v8PbLX8Gp4osis/0kl/3/mYVg1FtR93vek/5nrWP49y/kowSDTtX6iW/PYWych/XNmUfR5qwj9HO3XDyUZVTwP45U0ZBRFte5z6JqIgmGyIKAgmGyIKgsmGiIJgsiGiIJhsiCiIoP1sNBmh1WfXA2jCruu4+9fvcfe9uOIsEQPgmR+9asaWYupsRkZ3uXFN231jUgO97rYDTbsfDQAMxtRHSMWuTYoW/BqfTEz90Nq8vYzMYNJeQgYAemr+9+BGxF7WJJGJqaNZ82tKCjvs+qFm1S/ISg/7j2uuaD+upZhir1bkL+WSrPl9ejIZ+/kacGIAkEvF9dKx656STg3Ous6uWXhlQ0RBMNkQURBMNkQUBJMNEQXBZENEQTDZEFEQQae+M+k09u62l1TZN3qrGZua2uPue2nlNTd+9tKKGbu05E85tnr9Vgw7yvZyLemY6c7GjD9lP3aLvxTMHXsPmLFLKX/c5077y8j0OtOhUdo+ngCws9dvQdHTmzNjy4v2/QJANfLXTKmu2G05+nqddiAAkpE/7mrVeb76/H2n4T/X2R5/yj/nLEOTF/+Yqd/9Aom6c57GtN3oFK9siCgIJhsiCoLJhoiCYLIhoiCYbIgoCCYbIgqCyYaIgghaZyMqSDfsOobBwogZm33jrLvvHz/3ihtvwq7rKMcs77HstBUAgDrsmpO4JU9qCX/fjZglUTJ5u4VFfpdfo9NT8ZeJiSp224J05I+r5ZcuIQd7uRWk/I3zO/19p8RZLqh83t12pH/cjS827XqUwoD/cqrW/DYQw+LXRd061m/Gigt+nc3stN0uBAASzksg0xezxozf8ePn99HZrxERbQyTDREFwWRDREEw2RBREEw2RBQEkw0RBcFkQ0RBxNbZiMhjAH4PwLyqvrt9WwHA1wHsA3AGwEdU1V7PpK0RtTBTtmtSGsfsWoChHj8vnrs068bTg/ZyLMMpv/4BGb9e5cTFC/b9it8zZmLSri0CgLma3w9netZ+3BdiapPKWf/p7xuxa2mG0v62hYJ/zGbnp81YddZf3mY0t8ONS9KuOVls+TUj2vLPs+bOYTNWafhLnlxs+C+R0ZFRN/7ee+0aoMa0X691PJF342cv2nVP/X3upsBcTLytkyubvwLw4FtuexTAU6p6G4Cn2v8mIjLFJhtVfRrAWy85HgbwePvnxwF8aJPHRUTbzI1+ZjOuqjMA0P7bv64lone8m/7dKBE5BOAQAORjPiMgou3rRq9s5kRkAgDaf5sLTqvqYVU9qKoHs+mYL3QR0bZ1o8nmCQCPtH9+BMB3Nmc4RLRdxSYbEfkqgGcA3C4i0yLyCQCfBfC7InISwO+2/01EZIr9EEVVP2aEfvt676zZbGJh3q6hWEzZdQi3Tvo1CImYtZ2ySbvHSS3l92aZ3OM3UGnssOsfknV/jaOVZb/24vKqvz7TWsqu06kM+n1h6vBrM4adAouBjF9TMtTnH7Nmy64PWi36z3V9yT9mWafnzJ4xf1zVmH42K0v281F0+ugAQAL++ku37LZ7LgFAUuzGMZH49Vgjo/6+F0r2dUciZk0qwO/Z9LP9dPRbREQbxGRDREEw2RBREEw2RBQEkw0RBcFkQ0RBhF3KRRJIZO0paqnaX4PPpu2v9gPA5IA/tbdWtqcdz8VML1++4E/zRv32Y8rGLGmSSTlLmgAYHPSnYt997z12rH6nu+0bx/3lbxYur5qx/QX/63Da9I9pqsfevrfgT6WenPGn7Av99vI2+/barUYAYL7il0GUj9n9FHZMjLnbPnD/XW48Hfn/98+ds1uZRHNrMfv2q/dHc/Zrb1X9Kf1O8cqGiIJgsiGiIJhsiCgIJhsiCoLJhoiCYLIhoiCYbIgoiKB1NqmUYGTIrodZW7XbMTQa/tfzG5G/ZErF6fRQa/k1OhdKfrHM4qpdh3Bb1m8xcc9de934r+zy61kGU3aNT67fX04lueuAGy9feMOMVct2uwMAKJ/3H3dF7HqWTMt/rsdG/ZqrsX0T9r7H/fYVsz961Y3nnVPlznv3uNvu2ecv2zP901NuPGrax7wBv8VEK4ppf+HUv9UaGXfbTvHKhoiCYLIhoiCYbIgoCCYbIgqCyYaIgmCyIaIgmGyIKIigdTaJVAp9O+xag+aq3T9lZXnZ3XdF/aVcdgzYtRfpnF/XcfT0JTdeyto1DkXx8/n8xRk3fmzuvBu/dOIFM/beX/fraIYyfr1Ksm4vmVKe8Zf3KPdk3Xg9smuX5hYuu9sOjfjLsURlu5/NmycW3G33vOc2N/6e/Q+asfHJ33C3nZ95040vxJzj9ZJdZ5No+j2XevwWQNCqfQ73JP2aKcCvU7uCVzZEFASTDREFwWRDREEw2RBREEw2RBQEkw0RBRF46lvQO2x/XX06bU+n1ur+3N2q+Mt/DPbYU4O1pj+NW1qzp+QBIJEdtINpcbc9v+RPd2ay/uPKjNrT9ifPnHO3/SW/uwXye+xygZ6GPx2qvf7SIakle6q1OuKXIgwVCm58xGmdsW+wz9125y2/6sZXp+2lXF47fsLd9sQZu2UHAPz0hL/8zagzw5xb81/KkfrPh5cJRkf9UoNOxV7ZiMhjIjIvIseuuu0zInJBRI62/zy0KaMhom2rk7dRfwXgWpVMX1DVu9t/ntzcYRHRdhObbFT1aQB+SScRUYyNfED8SRF5uf02y6x7F5FDInJERI6UKzE100S0bd1osvkigFsA3A1gBsDnrF9U1cOqelBVD/bmgn4eTURbyA0lG1WdU9WWqkYAvgTgvs0dFhFtNzeUbETk6jnRDwM4Zv0uERHQQZ2NiHwVwG8BGBWRaQB/CuC3RORuAArgDIA/6OTOJJFAKm/X2ZTqdn2F1v26jqlJf4mOyQl7WZPiWf/z753D/W583vkoKpPwD3EqY7dDWN/erwHqb9j7byz6NTxR75AbH3TKXbJ9Y+62FfWfr9Ueu13C3snb3W3v+tUPuvEDt99vxho1f1me8qpf6/LC8b83Y3/3t//X3bY+4P/fnkv6S6akG3atWM+g39IjSvj1WqNOPVg04p+jgH/MrohNNqr6sWvc/OWO9k5E1MavKxBREEw2RBQEkw0RBcFkQ0RBMNkQURBMNkQURNDvD6TSKYxP2fUwI7vsepZ05NfR5HP+Ui6lkl330YT/na2hmDqDtRW7NwtW7HoSAMhPONsC6M362/eP2PVDe8f8hjWDBX8Jjqho33ej4S9vk0v5fWP2Fuz4wfd9xN22sNuuowGAZMo+j146+r/cbf/6v/+dG399YdGMJUZ2uNvmY5YM2r/Dfz56qxUzJst+XVPlsh8v9dqpoDrjL3/TKV7ZEFEQTDZEFASTDREFwWRDREEw2RBREEw2RBRE0KnvnmwWt+23l9mo3WEvmbK0WHT3rc01N96bsr9inxR/uZVk5MczzoxlbaXqbltfbbnxWp/dVgAAkrN2PDXlL9/Rl8m7ca8lQjZvdoIFAOzfuceNj4zfa8YGp+wYANQa/v+R33/m+2bsQ4/+ubvtb8z7+/79D77HjH3gnv3utr2pmOnpi/5SL8sn7VYOpfKMu23RP42QydotQxbKMRt3iFc2RBQEkw0RBcFkQ0RBMNkQURBMNkQUBJMNEQXBZENEQYRtMZFKYmR0xIwPDdl1H4sLdg0OALRqfi3MbMWucZhfm3e3XS75rQEaKXvc6XG/PUW5x24bAACn52fd+M683WIifclfbiXX5y9rkhm9xYxlh/2WH4MTd7rxnqFJM9ZS///AE6f9epS//P4zZuzTd/vLxDywd8qN/86g/Vwf2O0f77UFv1XD2fN2+woAmLlob99a8+u1Er1++4pi1V4yqAl/iRnAryX72Rg6+i0iog1isiGiIJhsiCgIJhsiCoLJhoiCYLIhoiCYbIgoiNg6GxHZDeC/AtgJIAJwWFX/QkQKAL4OYB+AMwA+oqpL3r4iBWoNuzfGmlNy0opZqiXnlxGgtWb3dkn2+X1dhqt+nU1p1a5XafjlP5BBu04GAEYnfsmNv+8jv2fGdu3+l+62lYVzbrwE+3Gv1YfcbQvw+900nJ40Z197yd32T/7jf3Hjc0t2zcnBd4272x7I++eCFO2+SeU5u1YFAKolf/mbesKvV9nRZx/zniG/Pmi16i+t06rZz3UyFfPigl8Dd0UnVzZNAH+sqncAuB/AH4rInQAeBfCUqt4G4Kn2v4mIrik22ajqjKq+0P65COA4gEkADwN4vP1rjwP40M0aJBG9/V3XZzYisg/APQCeBTCuqjPAekICcM3lAEXkkIgcEZEjKzGrQxLR9tVxshGRPgDfBPApVe3sTRoAVT2sqgdV9eCg870SItreOko2IpLGeqL5iqp+q33znIhMtOMTAPxvMxLRO1psshERAfBlAMdV9fNXhZ4A8Ej750cAfGfzh0dE20UnLSYeAPBxAK+IyNH2bZ8G8FkA3xCRTwA4B+BfxO0o0hZKtWUznhmyv8q+a8BfJuPU66fd+NKKPS2Z7h10t9W83waiunzZjPWk/WUwBhP+tPr4pN+iYmRwwowNDfttHrRhL28DAGfetKfGa4v+0jmprFsFgdVlu3XGE0/ZS7EAwMsnz7rxodxOM/aTHx9xtz015E/z/tPb7H339fvT/cvL/sV/puE/12sV+7XTEv+lvLBSd+NFsV97rWH/POlUbLJR1R8CsKpFfntTRkFE2x4riIkoCCYbIgqCyYaIgmCyIaIgmGyIKAgmGyIKIuhSLgrAWzxk58ReM9Y7kHX3XW/6y2DsGLVbOZxfOuNuO19vuvH8Lnt5mv6sX2cj/TE1PjGtNWbX7CNavXjB3fb0nL+0yIWy/a2U4d5rfhXuZ378zHNu/Ls/fNaMTV/2WzXs3lVw45UVuwZoesavN+nL+7Uu+WH7ua76u8bMm/4vtJb92qVM2j5XmpHfniKT818/+bwdL/mrxHSMVzZEFASTDREFwWRDREEw2RBREEw2RBQEkw0RBcFkQ0RBBK2ziaII1YpdDzBV2GXG9uy/x933xLBdowMAL/z0DTP26pPT7rYXL/o5udy045mCXydTuxRTwzPh90iZc2pS/vZH33O3fXPO7wntlbPsmLKXxgGAlVm79woA1PvsvjG3Toz62y75x2wxbT+u9Ki/pEkh778kjh2z+yal3CoyoLK84sZrC358cpddKyYxawatNuN60tjHpbTWcRdgF69siCgIJhsiCoLJhoiCYLIhoiCYbIgoCCYbIgoi6NS3SBKZlD19Nzayz4zl834rhlpt3I1nU5fMmKbtZSwAIMr5U63FRXuad7Duj3ts/IAb35WzywEA4Mz5GTP2jy+ccLcdHvTHJj32tHszpj1Fn/qn1q8duN2MnVv1p+Srq/5SLqm6PS0/kPX/f12LaUFRiopm7NZ3+a0v3vVr+9z4hTftcxQAWhX7uEjK7wNRqvnLEdWc/hgLRX9Kv1O8siGiIJhsiCgIJhsiCoLJhoiCYLIhoiCYbIgoCCYbIgoi7FIukaJSc5YeKds1DKdOvO7u+42ZN934qdklMxb5385Hoc+vR1lesuts6mV/5+Ulv67j1KxdRwMAF4v28h+T437tUa4n5ukXpy1Bw18apCfZ78Zr1bIZS/hlNsj1+8ut5BL2ObZQ8ds4pLN+vUoiadfwpGJqi1Lwl/UZzg658YUV+/WxVvL3XVy1jzcA1JzHhVrMGjUdir2yEZHdIvIDETkuIq+KyB+1b/+MiFwQkaPtPw9tyoiIaFvq5MqmCeCPVfUFEekH8LyIXOnK9AVV/Q83b3hEtF3EJhtVnQEw0/65KCLHAUze7IER0fZyXR8Qi8g+APcAuLJ26idF5GUReUxErvlFGhE5JCJHROTIasz7RiLavjpONiLSB+CbAD6lqqsAvgjgFgB3Y/3K53PX2k5VD6vqQVU9ODDgf7BHRNtXR8lGRNJYTzRfUdVvAYCqzqlqS1UjAF8CcN/NGyYRvd11MhslAL4M4Liqfv6q2yeu+rUPAzi2+cMjou2ik9moBwB8HMArInK0fdunAXxMRO4GoADOAPiDuB2V6zW8csZeUuX8i3afkumiX3yxXLOXNAGAfVO7zdhYwa8Jaaz6S1kM9NhvD/NpfwmNjPqfY60u+r1ESmt2nU1fxt82V7d7CwFApseuOSku+ce7NuzXsyQr9nEZGfKXv8mn8m68qnZdiI762/Zl/fqhUss5plX/PJmb8f9vL87azyUArJXtY7666tcHrZT8fjbNHnspl2zSPyZATGFUWyezUT8EcK3KtCc7ugciIvDrCkQUCJMNEQXBZENEQTDZEFEQTDZEFASTDREFEbSfTVRroXjGrkU4N2PXnFSTfl6UjN/P49yZeTO2tuTXo0jarkEAgITY415e8OtR+vxloVCu+l/xKDs9Thp1f72r7Lgf73XW6iqr/7iWL/g9UIZzVTOWj+kLU1/x16yqLNm9iyTh9G0B0GzFxEt2PculmLWbCpmYWpi6f47XIjVjixW/jqZc8vedTtrPZyumD0+neGVDREEw2RBREEw2RBQEkw0RBcFkQ0RBMNkQURBBp74RRdCSPSXaD3v6radlT/sBQGnGnkoFgHLanuZdi2lPkYxpaTDa67SR6PPbOPTm/CnL+YVpN35pdtGMFYb9Vg3zS35Lg5mq/VyNDfvHpJz1/x9bqdjPV27Nn9pevmyXMQBARu37Xqra0+IAIJEfL9bs8yjtnQcAzi37j6tPMm58ecU+Tyt1v3wj3ec/HyW1H5fElAN0ilc2RBQEkw0RBcFkQ0RBMNkQURBMNkQUBJMNEQXBZENEQQSts2lFEUpFu7ajkLPrDJrit1oorfjLSUjLrhnJOTEAKEZ+a4ByZNc/jPf5tROS9Z+ClvhjGxyxl6EZ3u3X+LSq11o04+eaLbsWJl2w208AQK/4+y5dtuumllb9Wpdmzl9apKn2MjLZwVF3W8S0auiP7OcznfTPk7WaXyvWavjx3JB9bZBFzDGJ/BqgqGGPvRpzTDrFKxsiCoLJhoiCYLIhoiCYbIgoCCYbIgqCyYaIgmCyIaIgYutsRKQHwNMAsu3f/xtV/VMR2Q/gawAKAF4A8HFVdYtCFIIa7N4YRWdJld6cv5xEvidm+Y+6HV9Wv44gGbOUxaUVe/vLi/YyLwAwMub3nIlbWqSctOtZTp33+76Mjo258Wzaril5/eIld9vKiv+4Gxm7f8qA01sFACpNP15dtuts+gb857Kn6dejSMI+j1YvXXa3TfX6tTBLZb8nTSZtj61U8euxWs64AaDWtK87Eim/ZqpTnVzZ1AC8X1V/GcDdAB4UkfsB/BmAL6jqbQCWAHxiU0ZERNtSbLLRdVfKftPtPwrg/QD+pn374wA+dFNGSETbQkef2YhIUkSOApgH8D0AbwBYVv3Z9e40gElj20MickREjtRi2m8S0fbVUbJR1Zaq3g1gCsB9AO641q8Z2x5W1YOqejCb9d8PE9H2dV2zUaq6DODvAdwPYEhErnzqNAXg4uYOjYi2k9hkIyJjIjLU/jkH4HcAHAfwAwD/vP1rjwD4zs0aJBG9/XXSYmICwOMiksR6cvqGqv5PEfkpgK+JyL8D8CKAL8fuKRJI3X4rpWJP85bL/tf3V5b96esosqcG8yl/+rlc9fetS3a8Kn4+jwZzbnyob8iN9ziPa6HoT4cOZPz2F4msPVVbj5lqnRN/mZjhHrs1RstZ0gcA+mOmcZOwp2oTTgwAFiv2tDkADObs7S+3/HN0Z8x9N9VvMdGXsuOLNX8po2TCP8eTsOM96c1ZyiU22ajqywDuucbtp7H++Q0RUSxWEBNREEw2RBQEkw0RBcFkQ0RBMNkQURBMNkQUhGjM3P6m3pnIJQBnr7ppFMBCsAF0bquOC9i6Y+O4rt9WHdv1jmuvqvr9ShA42fzCnYscUdWDXRuAYauOC9i6Y+O4rt9WHdvNGhffRhFREEw2RBREt5PN4S7fv2WrjgvYumPjuK7fVh3bTRlXVz+zIaJ3jm5f2RDROwSTDREF0ZVkIyIPishrInJKRB7txhgsInJGRF4RkaMicqSL43hMROZF5NhVtxVE5HsicrL99/AWGttnRORC+7gdFZGHujCu3SLyAxE5LiKvisgftW/v6nFzxrUVjlmPiDwnIi+1x/Zv27fvF5Fn28fs6yLiNz/qhKoG/QMgifWG6QcAZAC8BODO0ONwxncGwOgWGMdvArgXwLGrbvtzAI+2f34UwJ9tobF9BsC/7vIxmwBwb/vnfgCvA7iz28fNGddWOGYCoK/9cxrAs1hv+/sNAB9t3/6fAPyrjd5XN65s7gNwSlVP6/qidl8D8HAXxrGlqerTAN666tnDWF82B+ji8jnG2LpOVWdU9YX2z0Wst6+dRJePmzOurtN1QZZq6kaymQRw/qp/m8vAdIkC+K6IPC8ih7o9mLcYV9UZYP0EBrCjy+N5q0+KyMvtt1ldeYt3hYjsw3qHyWexhY7bW8YFbIFjtpGlmq5HN5LNtRqxbqX59wdU9V4AHwTwhyLym90e0NvEFwHcgvVVU2cAfK5bAxGRPgDfBPCKrFYpAAABSklEQVQpVV3t1jje6hrj2hLHTDewVNP16EaymQaw+6p/b6llYFT1YvvveQDfxtbqszwnIhMA0P7bX8w7IFWda5+0EYAvoUvHTUTSWH9Bf0VVv9W+uevH7Vrj2irH7Aq9yUs1dSPZ/ATAbe1PuzMAPgrgiS6M4xeISF5E+q/8DOADAI75WwX1BNaXzQG22PI5V17MbR9GF46biAjWV/k4rqqfvyrU1eNmjWuLHLNwSzV16RPwh7D+ifwbAP5NNz+Nf8u4DmB9duwlAK92c2wAvor1S+sG1q8GPwFgBMBTAE62/y5sobH9NwCvAHgZ6y/uiS6M659g/XL/ZQBH238e6vZxc8a1FY7Ze7G+FNPLWE92f9K+/QCA5wCcAvA/AGQ3el/8ugIRBcEKYiIKgsmGiIJgsiGiIJhsiCgIJhsiCoLJhoiCYLIhoiD+HwQ5XvnboNsDAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 잠재 공간에서 랜덤한 포인트를 샘플링합니다\n",
    "random_latent_vectors = np.random.normal(size=(10, latent_dim))\n",
    "\n",
    "# 가짜 이미지로 디코딩합니다\n",
    "generated_images = generator.predict(random_latent_vectors)\n",
    "\n",
    "for i in range(generated_images.shape[0]):\n",
    "    img = image.array_to_img(generated_images[i] * 255., scale=False)\n",
    "    plt.figure()\n",
    "    plt.imshow(img)\n",
    "    \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Froggy with some pixellated artifacts."
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}