CNN VAE in Edward

CNN_VAE_Edward.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CNNを使った変分オートエンコーダ(VAE)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Variable AutoEncoder(VAE)は\n",
    "edwardでは単一のクラスとして提供してはおらずencoder部分、decoder部分をそれぞれ定義することで記述することができる。\n",
    "\n",
    "TensorflowベースのライブラリTensorFlow Slimで定義したConvolutional Neural Net(CNN)を組み込む。\n",
    "- [exampleのコード](https://github.com/blei-lab/edward/blob/master/examples/vae_convolutional.py)\n",
    "- [１層全結合NNでのVAE](VAE_Edward.ipynb)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import absolute_import\n",
    "from __future__ import division\n",
    "from __future__ import print_function\n",
    "\n",
    "import edward as ed\n",
    "import numpy as np\n",
    "import os\n",
    "import tensorflow as tf\n",
    "import tensorflow.contrib.slim as slim\n",
    "\n",
    "from edward.models import Bernoulli, Normal\n",
    "from edward.util import Progbar\n",
    "\n",
    "from observations import mnist\n",
    "\n",
    "ed.set_seed(42)\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 実行"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def generator(array, batch_size):\n",
    "  \"\"\"Generate batch with respect to array's first axis.\"\"\"\n",
    "  start = 0  # pointer to where we are in iteration\n",
    "  while True:\n",
    "    stop = start + batch_size\n",
    "    diff = stop - array.shape[0]\n",
    "    if diff <= 0:\n",
    "      batch = array[start:stop]\n",
    "      start += batch_size\n",
    "    else:\n",
    "      batch = np.concatenate((array[start:], array[:diff]))\n",
    "      start = diff\n",
    "    batch = batch.astype(np.float32) / 255.0  # normalize pixel intensities\n",
    "    batch = np.random.binomial(1, batch)  # binarize images\n",
    "    yield batch\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "訓練用データをミニバッチごとに分割する関数generatorを定義する。\n",
    "メモリ節約のためdata x_train,x_test をpython文法のgeneratorとして定義する。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data_dir = \"/tmp/data\"\n",
    "out_dir = \"/tmp/out\"\n",
    "if not os.path.exists(out_dir):\n",
    "    os.makedirs(out_dir)\n",
    "\n",
    "M = 100  # batch size during training\n",
    "d = 2  # latent dimension"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from observations.util import maybe_download_and_extract\n",
    "\n",
    "def mnist(path,url='http://yann.lecun.com/exdb/mnist/'):\n",
    "#  url = 'https://storage.googleapis.com/cvdf-datasets/mnist/'\n",
    "    path = os.path.expanduser(path)\n",
    "    train_images = 'train-images-idx3-ubyte'\n",
    "    train_labels = 'train-labels-idx1-ubyte'\n",
    "    test_images = 't10k-images-idx3-ubyte'\n",
    "    test_labels = 't10k-labels-idx1-ubyte'\n",
    "\n",
    "    if not os.path.exists(os.path.join(path, train_images)):\n",
    "        maybe_download_and_extract(path, url + train_images + '.gz')\n",
    "    if not os.path.exists(os.path.join(path, train_labels)):\n",
    "        maybe_download_and_extract(path, url + train_labels + '.gz')\n",
    "    if not os.path.exists(os.path.join(path, test_images)):\n",
    "        maybe_download_and_extract(path, url + test_images + '.gz')\n",
    "    if not os.path.exists(os.path.join(path, test_labels)):\n",
    "        maybe_download_and_extract(path, url + test_labels + '.gz')\n",
    "\n",
    "    with open(os.path.join(path, train_images)) as f:\n",
    "        loaded = np.fromfile(file=f, dtype='uint8')\n",
    "    x_train = loaded[16:].reshape((60000, 28 * 28)).astype(float)\n",
    "    with open(os.path.join(path, train_labels)) as f:\n",
    "        loaded = np.fromfile(file=f, dtype='uint8')\n",
    "    y_train = loaded[8:].reshape((60000,))\n",
    "    with open(os.path.join(path, test_images)) as f:\n",
    "        loaded = np.fromfile(file=f, dtype='uint8')\n",
    "    x_test = loaded[16:].reshape((10000, 28 * 28)).astype(float)\n",
    "    with open(os.path.join(path, test_labels)) as f:\n",
    "        loaded = np.fromfile(file=f, dtype='uint8')\n",
    "    y_test = loaded[8:].reshape((10000,))\n",
    "\n",
    "    return (x_train, y_train), (x_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generative network"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "データを生成するネットワーク、\n",
    "入力は隠れ変数z,出力は尤度"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def generative_network(z):\n",
    "  \"\"\"Generative network to parameterize generative model. It takes\n",
    "  latent variables as input and outputs the likelihood parameters.\n",
    "  logits = neural_network(z)\n",
    "  \"\"\"    \n",
    "  with slim.arg_scope([slim.conv2d_transpose],\n",
    "                      activation_fn=tf.nn.elu,\n",
    "                      normalizer_fn=slim.batch_norm,\n",
    "                      normalizer_params={'scale': True}):\n",
    "        net = tf.reshape(z, [M, 1, 1, d])\n",
    "        net = slim.conv2d_transpose(net, 128, 3, padding='VALID')\n",
    "        net = slim.conv2d_transpose(net, 64, 5, padding='VALID')\n",
    "        net = slim.conv2d_transpose(net, 32, 5, stride=2)\n",
    "        net = slim.conv2d_transpose(net, 1, 5, stride=2, activation_fn=None)\n",
    "        net = slim.flatten(net)\n",
    "  return net\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def inference_network(x):\n",
    "  \"\"\"Inference network to parameterize variational model. It takes\n",
    "  data as input and outputs the variational parameters.\n",
    "\n",
    "  loc, scale = neural_network(x)\n",
    "  \"\"\"\n",
    "  with slim.arg_scope([slim.conv2d, slim.fully_connected],\n",
    "                      activation_fn=tf.nn.elu,\n",
    "                      normalizer_fn=slim.batch_norm,\n",
    "                      normalizer_params={'scale': True}):\n",
    "    net = tf.reshape(x, [M, 28, 28, 1])\n",
    "    net = slim.conv2d(net, 32, 5, stride=2)\n",
    "    net = slim.conv2d(net, 64, 5, stride=2)\n",
    "    net = slim.conv2d(net, 128, 5, padding='VALID')\n",
    "    net = slim.dropout(net, 0.9)\n",
    "    net = slim.flatten(net)\n",
    "    params = slim.fully_connected(net, d * 2, activation_fn=None)\n",
    "\n",
    "  loc = params[:, :d]\n",
    "  scale = tf.nn.softplus(params[:, d:])\n",
    "  return loc, scale"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ">> Downloading /tmp/data/train-images-idx3-ubyte.gz.part \n",
      ">> [9.5 MB/9.5 MB] 105% @765.9 KB/s,[0s remaining, 13s elapsed]        \n",
      "URL http://yann.lecun.com/exdb/mnist/train-images-idx3-ubyte.gz downloaded to /tmp/data/train-images-idx3-ubyte.gz \n",
      ">> Downloading /tmp/data/train-labels-idx1-ubyte.gz.part \n",
      ">> [28.2 KB/28.2 KB] 3630% @868.6 KB/s,[0s remaining, 1s elapsed]        \n",
      "URL http://yann.lecun.com/exdb/mnist/train-labels-idx1-ubyte.gz downloaded to /tmp/data/train-labels-idx1-ubyte.gz \n",
      ">> Downloading /tmp/data/t10k-images-idx3-ubyte.gz.part \n",
      ">> [1.6 MB/1.6 MB] 127% @768.7 KB/s,[0s remaining, 2s elapsed]        \n",
      "URL http://yann.lecun.com/exdb/mnist/t10k-images-idx3-ubyte.gz downloaded to /tmp/data/t10k-images-idx3-ubyte.gz \n",
      ">> Downloading /tmp/data/t10k-labels-idx1-ubyte.gz.part \n",
      ">> [4.4 KB/4.4 KB] 23086% @1.6 MB/s,[0s remaining, 0s elapsed]        \n",
      "URL http://yann.lecun.com/exdb/mnist/t10k-labels-idx1-ubyte.gz downloaded to /tmp/data/t10k-labels-idx1-ubyte.gz \n"
     ]
    }
   ],
   "source": [
    "ed.set_seed(42)\n",
    "\n",
    "data_dir = \"/tmp/data\"\n",
    "out_dir = \"/tmp/out\"\n",
    "if not os.path.exists(out_dir):\n",
    "  os.makedirs(out_dir)\n",
    "M = 128  # batch size during training\n",
    "d = 10  # latent dimension\n",
    "\n",
    "# DATA. MNIST batches are fed at training time.\n",
    "(x_train, _), (x_test, _) = mnist(data_dir)\n",
    "x_train_generator = generator(x_train, M)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Encoder,Decoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# MODEL(Decoder)\n",
    "z = Normal(loc=tf.zeros([M, d]), scale=tf.ones([M, d]))\n",
    "logits = generative_network(z)\n",
    "x = Bernoulli(logits=logits)\n",
    "\n",
    "# INFERENCE (Encoder)\n",
    "x_ph = tf.placeholder(tf.int32, [M, 28 * 28])\n",
    "loc, scale = inference_network(tf.cast(x_ph, tf.float32))\n",
    "qz = Normal(loc=loc, scale=scale)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "### 推測処理"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "p(x, z)とq(z | x) の合体"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "data = {x: x_ph}\n",
    "inference = ed.KLqp({z: qz}, data)\n",
    "optimizer = tf.train.AdamOptimizer(0.01, epsilon=1.0)\n",
    "inference.initialize(optimizer=optimizer)\n",
    "\n",
    "hidden_rep = tf.sigmoid(logits)\n",
    "tf.global_variables_initializer().run()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## lossの計算と出力"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "n_epoch = 100\n",
    "n_iter_per_epoch = x_train.shape[0] // M\n",
    "for epoch in range(1, n_epoch + 1):\n",
    "  print(\"Epoch: {0}\".format(epoch))\n",
    "  avg_loss = 0.0\n",
    "\n",
    "  pbar = Progbar(n_iter_per_epoch)\n",
    "  for t in range(1, n_iter_per_epoch + 1):\n",
    "    pbar.update(t)\n",
    "    x_batch = next(x_train_generator)\n",
    "    info_dict = inference.update(feed_dict={x_ph: x_batch})\n",
    "    avg_loss += info_dict['loss']\n",
    "\n",
    "    #1つの画像に対する平均周辺化尤度の下限を表示\n",
    "  avg_loss = avg_loss / n_iter_per_epoch\n",
    "  avg_loss = avg_loss / M\n",
    "  print(\"-log p(x) <= {:0.3f}\".format(avg_loss))\n",
    "\n",
    "  # 事前予測のチェック\n",
    "    images = x.eval()\n",
    "    for m in range(M):\n",
    "        im= Image.fromarray(images[m].reshape(28, 28))\n",
    "        im.save(os.path.join(out_dir, '_%d.png') % m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "import numpy as np\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d0245e90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d017d3d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nfJEG+n+54qmsecouOwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d00b0cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87cffe9490>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAGiFJREFUeJztnV2MJFd5hp9vdma6Z9Zry0KxLeHgSRQpd8hKFN8YCSMQ\nQhGSIy4cYoT4iRAXcYIEFwZuLKJcABeWCBIXcQyyEYg/idjcEIOidWQkgiE4McHGSMmYP+/aish6\nf6Z7ZnZOLrpOb3V1VZ3qrur66XofqeTenj7Vp8v99qlz3u/7jjnnEEL0i42mOyCEqB8JX4geIuEL\n0UMkfCF6iIQvRA+R8IXoIaWEb2ZvM7PnzewFM7u/qk4JIVaLLevjm9kG8ALwZuA3wNPAO51zzyde\np0ABIRrCOWdpz2+WOOcdwM+dcy8CmNlXgLuB5+df+kDs8VngrhJvu2rOov6V4Szt7d9Z2ts3qL5/\nn8j8S5lb/dcCv4z9+1fRc0KIlqPFPSF6SJlb/V8Dr4v9+9bouRTOxh4PS7xlHew13YEAe013IMBe\n0x3IYa/pDgTYK9l+PzrClFncOwX8jMni3kvAD4C/cM49l3idm53jCyHq4RPVL+45566a2X3AE0ym\nDA8nRS+EaCdlbvVxzn0b+MOK+iKEqAkt7gnRQyR8IXqIhC9ED5HwheghEr4QPUTCF6KHSPhC9BAJ\nX4geIuEL0UMkfCF6iIQvRA+R8IXoIRK+ED1Ewheih0j4QvQQCV+IHiLhC9FDJHwheoiEL0QPkfCF\n6CESvhA9RMIXoodI+EL0EAlfiB4i4QvRQyR8IXqIhC9EDym1d57oIg7LORyW+yqgZPvUzVtFzUj4\nPcNwnOJq6rHBCcD0mRM25l4FlGrvJPxWIOH3DC/8LY5mjk2O2eIIIPGXLY6jr8lJNDMs09418JnF\nPBJ+z4gLf8B4emxzyIAxQOzZwXQUP2FjKuCy7UXzlPo/YWb7wAXgBDhyzt1RRafE6kgKf4cDdjhg\nyIgdDgA4YIdNjmdEe5VTWDRel20vmqfsT/AJcJdz7rdVdEasnjThn+Yyu1zhNJcB5kR7zCaHbGcK\nf9H2onnKCt+QJdgpsoR/HZc4w0UcliraU1zNFf4i7UXzlBW+A75jZleBf3DOPVRBn8QKSRP+Llc4\nw0Vu4MJ01T0u2hFDTnG1kvaiHZQV/p3OuZfM7HeY/AA855x7av5lZ2OP96JDrIZ8n37IaO6Iz9MB\nDtnmkO2ZRTp/pJ1jkfahXsrnL8N+dIQpJXzn3EvRf18xs28CdwApwr+rzNuIBQj59DsccIaLnOYy\nOxwwYMwWR2xwknkrf8zm1MozXKn2A8by+VfGHrOD6pOZr1xa+Ga2C2w45y6Z2WngrcAnlj2fqIaQ\nT+9vzXe5wpAR2xyyyXHmHD4uWi/uMu0HjOXzt4AyI/7NwDfNzEXn+ZJz7olquiWWJeTTx2/N4yN2\nlnDjot3kGKBU+wFj+fwtYOkr7Zz7H+D2CvsiKiDk0w8Zzc28s4SbFO02hwCl2g8Yy+dvAfqJXTNC\nPv2Q0dw0IEu4MCtav3hXpv2AsXz+FiDhrxkhn37IKHPxLy5cmBVtkSSdIu23OZTP3wIk/DUj5LMP\nGeXafcDUAdjkeOG03FB7P12Qz98sEn7nKOfTD6NEmjwmAl5+9A21D8UJ5KF8/2qQ8DtGFT59G/qf\n5fMfsp3bPu79Kw5geST8jlGFT9+W/qf5/Eds5ba/yinFAVSAhN8xqvDp29T/pOUX8vKP2VQcQAXo\nSnWMKnz6tvQ/zecvInzFAZRHwu8YVfj0bek/zPv83jLM4ogtxQFUgITfMarw6dvQf0j3+f1dQBZp\nt/eKA1gcCb9jVOHTt6H/WT5/aFV+xBBQHEBZJPw1ICnsIgLKO1fynFnnTztC7Y1ycQKGU75/BUj4\nHcNhU0trzGBuoctHxi3LBiczcQHJ6QKQMZG4dpue177sXUcoDkD5/sWQ8DtGUvjJhS6/aLYsWTEC\nwMyGGWleuvfg89tXK3zl+y+HhN8xksJPLnR5kS2Lz5n3R1o+fvz9/eFvvYFg+zKE4gCU718MXYmO\nERd+2up2FcJP3j77W3Z/m5ycahyww4ghB+wABNuXIRQHoHz/Ykj4HSMufKh+dXuLoznRelFlCf8y\np7nCLpc5DRBsX4ZQHIDy/Ysh4XcML3yY97H9gloZ0m6fvc+eJ/xLXMdFzmC4YPsyhOIAlO9fDAm/\nY3jh+y911V69r7KTFVGX5ipcYZeLnOECN0zff9GIvKKE4gCU718MCb9zrDYUx4vHz5eTh8PmMv7j\n83wgt33yUywWvhMm7+poxL+GhC9mCMUJOIyLnOEypzlghzEDjtiKrLL0qUC8/ZhBKZ/f9y8rjuCA\nnWD/hIQvEoTiBBwWZfvvMmLIIdscs5m5BpBsP2ZQyueP9y8tjsBPPfL6JyR8kSAUJwDEsv2vjahZ\nwk+2HzMo5fMn+5eMI4hPPbL6JyR8kSAUJwDMzdyzhJ/WfsyglM+fNhWJxxFMqgrm909I+CJBKE4A\nSMn2Txd+Wvsxg1I+f5rw43EEI4bB/gkJXyQIxQlAdpJOXPhZ7Q/ZLuXzZwnfxxGMGAb7JyR8kSAU\nJ+Bfk2fG5bX304Vlff5QHMGIYbB/QsIXS1AmkiBPgEVHfB97f8zmzNrBZCoxXKpffUPCFzOE6vZX\nERLc5rr/fUHCFzOE6vaXDX3d5rDVdf/7goQvZgjV7a8q7betdf/7QlD4ZvYw8HbgvHPu9dFzNwJf\nBW4D9oF7nHMXVthPUROhuv1VF/qQ8JuhyIj/BeCzwKOx5z4KfNc592kzux/4WPSc6Dihuv2rKO0l\n4ddPUPjOuafM7LbE03cDb4wePwKcRcJfC0J1+6ss5tnGuv99Ydk5/k3OufMAzrlzZnZThX0SDRKq\n2x9PrV32/G2u+98XqlrcC/wfOxt7vBcdYjnyheO98GXryhtuOipvRk55fEPOYUnhhz9d/if0CTk+\nI89n3oV24OkH+9ERZlnhnzezm51z583sFuDl/JffteTbiCQhnx2uhdR2sa688u3LsMfsoPpk5iuL\nCj85VDwOvBf4FPAe4LEFeidKEPLZYT6Jpkt15ZVvXw9F7LwvMxmyX2NmvwAeAD4JfN3M3g+8CNyz\nyk6Ka4R8dphNm+1aXXnl29dDkVX9ezP+9JaK+yIKEPLZgU7XlVe+fT20fwgQM4R8dqDTdeWVb18P\nEn7HCPnsDut0XXnl29eDhN8xQj67//J3ta688u3rQcJvHfk+/XCuqv1oZp7v6977lX6f+XYt7TX/\n/H6hMK19NZ8uvwfzn262br/y7atBwm8ZIZ9+h4PcfPbQqFf2/GWRT98OJPyWEfLp/a39svnsqz5/\nCPn07UDCbxkhnz5+a79MPvuqzx9CPn07kPBbRsin91tBL5vPvurzh5BP3w4k/JYR8umHjErls6/6\n/CHk07cDCb9lhHz6IaNS+eyrPn8I+fTtQMJvGSGffsgo145r+vwh5NO3Awm/dsr59KF8+LR8+i2O\npot3ae+xyPmr+PSqi988En7NrNpHTxvR43vVGU517YWEXzd1+vQDxnMbVBpOde2FhF83dfv0afvP\nq669kPBrpk6fPm1XWkB17YWEXzd1+vQwvystoLr2QsKvm7p8epjff95vRa269kLCr5m6fHov+jQf\nfJU+vegGEn7lLJbvHrfyqhDfpBzyJOu9jYTiDMruCyCKIeFXTMinHzBmlytsc8gGJ5ywwSHbXGEX\nmFTIjb9+3W7HQ3EGA8Yz+fkK210NEn7FhHz6bQ6nR1z4AMdsTu8Gkm2BqArOegk/GWcwYNzpfQG6\ngoRfMSGf3kfJ+cML/5jNacmseLs0H77LhOIMBow7vS9AV9CVrJiQT3+Kq9N4dX/4Ue+EDTY5nrv9\n9bf863CbG4ozGDDu9L4AXUHCr5iQT7/ByXTTRz/SH8YmAKe4Oid6L4p1Ez7MxxkMGHd6X4CuIOFX\nTMinN9y0ppwXuF/cu8Iup7iaKvp1WdgKxRn4tQ/o5r4AXUHCr5iQT+/xc/q48F/l+umXPjkS+uCb\nrhOKM/BhxV3dF6ArSPgVk+ZTxxNwHDYTIpssTOHz5dP2fy8y4ofq1jeNH7XjvUrGMPhpT7L23pjB\n9Dzy+csh4a8Zobr1fhrRFPHYhKw4Bfn8q0fCXzNCdeubnjJkxTjAtXoB8vlXj4S/ZoTq1jfthXuv\nPitOQT5/PQSvlJk9DLwdOO+ce3303APAB4CXo5d93Dn37ZX1UhQmVLe+aXGE4hTk89dDkW/BF4DP\nAo8mnn/QOfdg9V0SZQjVrT9iq9H+bXGUG6fgfwTk86+WoPCdc0+Z2W0pf9IqSgsJ1a33eQFNkXb7\nHo9TkM9fD2Xu++4zs3cDPwQ+4py7EGogVk+obn3cEmsCXwUoK05BPn89LCv8zwF/65xzZvZ3wIPA\nX2a//Gzs8V50rCdpdeN9eO6YAQ6bhuwmffoi7UM0v798fj2CuAWX/Gx+M4088moWaMTfj44wSwnf\nOfdK7J8PAd/Kb3HXMm/TSdJG3Pic1GHB/d/z2vsRL4sxg0b3ly9bj8DP7dv6+drNHrOD6pOZrywq\n/JmQKDO7xTl3LvrnO4CfLNS/NSYp/ORClMOC+7/ntQ8JIx7338T+8mXrEYRu2Zv+fOtCETvvy0yG\n7NeY2S+AB4A3mdntwAmTe4sPrrCPnSIp/DSfObT/e177UE7+MZuN7i9fth5B6Ha96c+3LhRZ1b83\n5ekvrKAva0Fc+Gmrz0Bw//e89kWE3+T+8mXrEYRo+vOtCwp1qpi48GF+9RlIqWqfLvy09qFb4bRw\n3aaFv0g9gpD4m/5864KEXzFe+DDvM3vRhvZ/z2vv7wKySEtcqTOBpWw9glBkYdOfb12Q8CvGC9+L\nNq1ufShtNq99aA6cb6Y1I/xF6hGEAoya/nzrgoRfOeWr45evrt8u4tL0j/zIHffx/ZEnbYm7GiR8\nUSll4xhCcQBwbaqkfPzlkfBFpZSNYwjFAcD84qjy8RdHwheVUjaOIRQHALN2qPLxl0NXSlRK2TiG\nUBwAoHz8CpDwRaWUjWOI5+OnxQEAysevAAlfVErZOIZQHID/cUg7v4RfHAlfVErZOIZQHIBftVc+\nfjkkfFE7eZEO89UERjPzfIdNNxf1+xZ4q08jfnEkfFEpIR8+FHI8YMwZLnKay+xwMJPRpwCe6pDw\nRaWEfPjQLfk2h5HLf4UhI7Y5nI7qGtGrQ8IXlRLy4UNpxb78tj/8iC/hV4uELyol5MMXEX5yxzwJ\nv3okfFEpoXz8UOmwtGmChF89Er6olJAPHyoWmraRZnxDTVENEr6olJAP7+Pt89rnHaIaJPzKyf/i\nejsqu5DE+tlV8U/o7TyV0mgWCb9i+p5PHsrHHzOYuR66nW8GCb9i+p5PHsrHHzNIvTZA9LquX4Fu\nIOFXTN/zyUP5+GMGM9cluYGmqIfuf9NaRt/zyUP5+GMG7HAwt1W2r7cv6kHCr5iQjw3rnU8eyscf\nM5gTvd81V8KvDwm/YvqeTx7Kx49vmhEX/TosbHYJCb9i+p5PHsrH9+W34qIfMpr+WIh6kPBrIBmA\nkjWyd2PELxdgk+fW+zPkvcOYwXQLruMoI9//0IjiSPgVU7aufNupI98+qzTXCRscsNPp69cWJPyK\nKVtXvu2sOt8+fv2SsQ5HbHHATqevX1uQ8CumbF35trPqfPvk9fOHtwJHDDt9/dpCUPhmdivwKHAz\ncAI85Jz7ezO7EfgqcBuwD9zjnLuQeaKeULaufNtZdb592lTpgJ2p4L0l2NXr1xaKjPjHwIedc8+Y\n2XXAj8zsCeB9wHedc582s/uBjwEfXWFfO0HZuvJtZ9X59vHNNL3wL3OaK+xymdOMGHb6+rWFoPCd\nc+eAc9HjS2b2HHArcDfwxuhljwBnkfCDPjbk15VvO6vOt08b8S9zmktcx0XOMGLY6evXFhaa45vZ\nHnA78H3gZufceZj8OJjZTZX3roOUrSvfdladb58m/CvscpEzXOAGRgw7ff3aQmHhR7f53wA+FI38\nSdM2x8Q9G3u8Fx39Jc/zXn2+frl6AT6n/hRXp3Xt/cLeMJqBl+2dv933Pn18YW8cTZdEGvvREaaQ\n8M1sk4nov+iceyx6+ryZ3eycO29mtwAvZ5/hrkKdWQfK+tzxHP1V5OuXrRcg2swes4Pqk5mvLDri\nfx74qXPuM7HnHgfeC3wKeA/wWEq73lHW587ysKGafP2y9QLEelDEzrsTeBfwrJn9mMkt/ceZCP5r\nZvZ+4EXgnlV2tCuU9bl9zvqq8vX7Xi9ATCiyqv89yLzHe0u13ek+ZX3uYzZXmq/f93oBYoJ+wium\nrM99xNZK8/X7Xi9ATJDwK6asz512e11lvn7f6wWICRJ+xZT1uX1036ry9fteL0BMkPAXJt8H9wtl\nfqU8bpUVyVmPR7b5lfa4Tx5qX9aHBzhke+qdJw/DzXy++P70xa6e8u3bgIS/ICEffMCYXa6wzSEb\nnHDCBodsc4VdgEJz/CvsTktUbXAyTWX1f88jFAdQ9PPF7wjiNfIMl5tPH0L59u1Awl+QkA++PR0v\nZ4UPkxX7Ij6+P0Nc+DBZdAuJNxQHsMjnGzCeK4xpuFL71yvfvh1I+AsS8sH96OcPL/xjNhkxLHSr\nfhI7gxf+JscMGQW//GXjAJKfL63ufZn965Vv3w4k/AUJ+eC+PnxcvH7ULDri+nm4P7y9VmQeXTYO\nIP750qrhAqX2r1e+fTuQ8Bck5INvcDJdnPIj/WFsAhAS/ymuThfP/EifnD7kUTYOIP75YL4aLpCb\nTx8iK+1W+fb1IuEvSMgHN9x0jupHer+455/LY4uj6RzaCzhepy4U+Vc2DsB/Ppive+/XF/Ly6UMo\n374dSPgLEvLBPX5OHxf+q1w/XejLwvv8fk4fF/71vBoMACobB+A/nxd9Wr57np0ZQvn27UDCr4Bk\nPns8p9zf8ie98DzhxL32uI8/8drDhS7K+PCGr++fVS9g3of3VlsRYY6irP34EZ/nK9++HiT8BUkb\nsRapmx+KA9jhoJRP3rQPH1rDGDOQT98CJPwFSQp/0br5oTgAP3VY1idv2ocPxRnE1zvk0zeHhL8g\nSeEvWjc/FAfgbcFlffKmffjQ4qW3G+XTN4uEvyDxL/4ydfNDcQBDRqV88qZ9+CLCl0/fPBL+gsS/\n+LB43fz4ol1aHMCQUSmfvGkfPpRLkDZNkPDrR8JfEP/Fh+Xq5ofiAIaMSvnkTfvwIbsyLXFIPn39\nSPgL4r/4y9bND8UB+NTbZX3ypn34MYNge/n0zSPhL0xRiSx+xuQPR9Zrk20WEW7o/CHkw68HEn7N\nhOIAFtmCKm07KsieahRNFMpDPvx6IOHXTCgOYJlNJ71NF98QY1mfPYR8+PVAwq+ZUBzAottMp/n0\nZXz2EPLh1wMJv2ZCcQBFhJ8Mw/W3/F54ZXz2EPLh1wMJv2ZCcQChLLotjuZE7y27LOEv4rOHkA+/\nHkj4NROKAwgV2ki7vfc+fZ7wi/rsIeTDrwcSfs2E4gBClpyPvktG5Pkfk7I+e5H+y4fvPhJ+7ZSL\nA3DYtAzXfLb9GIfJZxdBJPyOUbYegBAg4XeOsvUAhADCYVxmdquZ/YuZ/ZeZPWtmfx09/4CZ/crM\n/j063rb67oqsJBk/h3+V61NHfAlfxCky4h8DH3bOPWNm1wE/MrPvRH970Dn34Oq6J5KUrQcgBBQQ\nvnPuHHAuenzJzJ4DXhv9Wd+kmilbD0AIWHCOb2Z7wO3AvwFvAO4zs3cDPwQ+4py7kN1aVEHZegBC\nQIE5vie6zf8G8CHn3CXgc8DvO+duZ3JHoFv+GgjN8S9wg+b4IkihEd/MNpmI/ovOuccAnHOvxF7y\nEPCt7DOcjT3eiw6xHNXXAxDrwn50hCl6q/954KfOuc/4J8zslmj+D/AO4CfZze8q+DZCiOXZY3ZQ\nfTLzlUHhm9mdwLuAZ83sx0w2U/k4cK+Z3Q6cMPmZ+eCy3RVC1EuRVf3vQWr1hm9X3x0hRB2Uq8Mk\nhOgkEr4QPUTCF6KHSPhC9BAJX4geIuEL0UMkfCF6iIQvRA+R8IXoIRK+ED1Ewheih0j4QvSQBoS/\nX/9bLsR+0x0IsN90BwLsN92BHPab7kCA/dreScKfY7/pDgTYb7oDAfab7kAO+013IMB+be+kW30h\neoiEL0QPMedWW73NzFQeToiGcM6lVlhdufCFEO1Dt/pC9BAJX4geUpvwzextZva8mb1gZvfX9b5F\nMbN9M/sPM/uxmf2gBf152MzOm9l/xp670cyeMLOfmdk/m9kNLetfazZSTdns9W+i51txDZvejLaW\nOb6ZbQAvAG8GfgM8DbzTOff8yt+8IGb238AfO+d+23RfAMzsDcAl4FHn3Ouj5z4F/K9z7tPRj+eN\nzrmPtqh/DwAX27CRqpndAtwS3+wVuBt4Hy24hjn9+3NquIZ1jfh3AD93zr3onDsCvsLkQ7YJo0VT\nH+fcU0DyR+hu4JHo8SPAn9XaqRgZ/YOWbKTqnDvnnHsmenwJeA64lZZcw4z+1bYZbV1f9NcCv4z9\n+1dc+5BtwQHfMbOnzewDTXcmg5ucc+dhuovxTQ33J437zOwZM/vHJqcicWKbvX4fuLlt1zCxGS3U\ncA1bM8K1gDudc38E/CnwV9GtbNtpmxfbuo1UUzZ7TV6zRq9hU5vR1iX8XwOvi/371ui51uCceyn6\n7yvAN5lMT9rGeTO7GaZzxJcb7s8MzrlX3LVFo4eAP2myP2mbvdKia5i1GW0d17Au4T8N/IGZ3WZm\n28A7gcdreu8gZrYb/fJiZqeBt5K7CWhtGLPzvceB90aP3wM8lmxQMzP9i4TkCWykWgtzm73SrmuY\nuhlt7O8ru4a1Re5FtsRnmPzYPOyc+2Qtb1wAM/s9JqO8Y7Kf4Jea7p+ZfZnJNsOvAc4DDwD/BHwd\n+F3gReAe59z/tah/b2IyV51upOrn0w30707gX4Fnmfx/9Zu9/gD4Gg1fw5z+3UsN11Ahu0L0EC3u\nCdFDJHwheoiEL0QPkfCF6CESvhA9RMIXoodI+EL0EAlfiB7y/1peyiZEeGNcAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d002d250>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAGChJREFUeJztnUusJFd5x3/fffa948GyUDyWmOCbKFJ2yEoUb4zEIBCy\nokiOWDjECPESYhEnSLAwsBmBWAALSwSJRRyDbATiJRGbDTEoGkdGIpiHgwk2RkquweAZWxEZ5nHf\n98ui+7Srq6vrVN+qrq7q+v+kkvt29+k6rul/n1Pn/33fMXdHCNEtlubdASFE/Uj4QnQQCV+IDiLh\nC9FBJHwhOoiEL0QHKSV8M7vTzJ41s+fM7L6qOiWEmC12Uh/fzJaA54A3Ab8FngTe5u7Ppt6nQAEh\n5oS7W9bzKyU+83bgl+7+PICZfQW4C3h2/K3nE48vAOdKnHbWXED9K8MFmtu/CzS3b1B9/z428ZUy\nU/3XAL9O/P3C4DkhRMPR4p4QHaTMVP83wGsTf58dPJfBhcTjXolT1sHWvDsQYWveHYiwNe8O5LA1\n7w5E2CrZfntwxCmzuLcM/IL+4t6LwA+Av3X3Z1Lv89F7fCFEPXys+sU9dz8ys3uBx+jfMjyYFr0Q\nopmUmerj7t8G/rSivgghakKLe0J0EAlfiA4i4QvRQSR8ITqIhC9EB5HwheggEr4QHUTCF6KDSPhC\ndBAJX4gOIuEL0UEkfCE6iIQvRAeR8IXoIBK+EB1Ewheig0j4QnQQCV+IDiLhC9FBJHwhOoiEL0QH\nkfCF6CASvhAdRMIXooNI+EJ0EAlfiA4i4QvRQSR8ITqIhC9EB5HwheggEr4QHWSlTGMz2wYuA8fA\ngbvfXkWnhBCzpZTw6Qv+nLv/rorOCCHqoexU3yr4DCFEzZQVrQPfMbMnzex9VXRICDF7yk7173D3\nF83sD+j/ADzj7k+Mv+1C4vHW4GgrjuUcRVrnfQqD14WYnu3BEaeU8N39xcF/XzazbwK3AxnCP1fm\nNI3CcJY5yjyWOGaJ49z2xywNWyQfh8MlfHFithgdVB+f+M4TC9/MNoEld79qZqeAtwAfO+nntYUg\n/FUORo4VDlnlgGWOctsfsZxqucrh4J/hmKUCcwYhylNmxD8DfNPMfPA5X3L3x6rpVnNJCn+dveGx\nxj7r7LHCYW77Q1YSrdaHM4RjloY/AELMmhN/09z9f4DbKuxLK0gLf4MdNtihxy4b7BQS/g4brHA4\nIvojlgutEQhRBRpipiRL+Ke4xibXOcU1VjnIbX/A6pjoD1lhnzUJX9SGhD8lk4R/A1c5zRXW2M9t\nnzW932eNZY4kfFEbEv6UZAl/k+uc5go3cpl19nLb79IDRkW/Sy+6KChElUj4Y+T79GEhL6zkJ608\nw4ejed5n7LPGPmsji3zhiI36igMQVSDhp4j59Ovsscl11thniWOOWWKfNa6zCfSn8sn3pz8ja8Zw\nyArHgyDKfdZy+6c4AFEFEn6KmE+/NhyvR4UP/RX7MBtItwWGs4Kk8JOiX+KYA1Zz+6c4AFEFEn6K\nmE+/ysEwQi8p/ENW2KU31i4p6hUOxz4//XrMy1ccgKgCfVNSxHz6ZY5wLCH9peGofcwSKxyOTd/D\nlN+xkc9Pi36N/ULCVxyAKIuEnyLm0y9xzD5rHLA6HOn3EzcAyxyNiT6I2rHhj0Dw+5Ov99jliOXc\n/ikOQFSBhJ8i5tMbznU2uc7mUOBhce86myxzlCn6sPAWPj/r9bBgl4fiAEQVSPgpYj59INzTJ4X/\ne141FOWkkTx8fng925KbjOIARBVI+AVIyjI8CvfVYcofFtuA4cR/3KXfy/zMsNofl32fsIi4wuEw\nliB8hoiRH6fRlTgJCT+FY0NB77E+tpDmGFc4zTVOscMGe6xzwOrASut/KfLaF/H5xeyIxWkAnYiT\nkPBTpIWfXkhzbHg/v0tvuMCX/FLktd9jPdfnR8KfKbE4DaATcRISfoq08LN88p2hyffKiD9J+On2\ne6zn+vxitsTiNIBOxEkszv9JRSSFn7V6DozduU8Sflb7PdZzfX4xW2JxGkAn4iQk/BRJ4cP46jmM\nTwUnCT+r/R7ruT6/mC2xOA2gE3ESEn6KIHwY98mDZZa9NDQq/Ent91nL9fnFbInFaYQgK1jsOAkJ\nP0UQfvhHzyqfnWcIAbntw+3CtBF7ohpicRrh33DR4yQk/DGKVsifRDyfftLIXmTED77/MkdDHz9k\nDcaKgFRB3OfOr0VQtn1ZYvUUwvmz0IgvJhLziTfY4TRXOMU1NtgZyfgr8sUqm89fllg9AKCUTx5r\nH9u3IMYqB7n1FI5ZisZpLAISfsXEfOIwtdzkOj12WWN/OPKcRPjT5vOXJVYPACjlk8fal51yL3OU\nW0/hiOVonMYiIOFXTMwnDrZROMKIf1LhT5vPX5Yi9QDK+uRl9i2IEW6VJtVTOGI5GqexCEj4FRPz\niXvsjsXvn1T4J8nnL0uRegBlffIy+xbESOZaZNVTSP+wSfiiEDGfuMduRhTAyYQP9bsDReoBlPXJ\n89rH9i2IccRybj2FrBgNCV9EifnEPXYnLl5NI3w4WT5/WYrUAyjrk+e1j+1bEOOA1dx6CvusReM0\nFgEJv2JiPnGP3VJ2Vdl8/rLE6gGU9clj7ctaliF1elI9hT3Wo3Eai4CEXwNVetFG8JPn4ymHIKS8\nfQF67I4cyft0oFT7XknhG84e60MvP6uWQd6/06Lk60v4FRPL5297Pn4sjsDwaJxC2faz7H9X9jWQ\n8Csmls/f9nz8WByB4dE4hbLtZ9n/ruxrIOFXTCyfv+35+LE4AiAap1C2/Sz735V9DaI9NbMHgb8C\nLrn76wbP3QR8FbgV2AbudvfLEz+kQ8Ty+duejx+LIwCicQpl28+y/13Z16DIT9QXgM8CDyee+zDw\nXXf/tJndB3xk8FznieXztz0fPxZHAETjFMq2n2X/u7KvQVT47v6Emd2aevou4A2Dxw8BF5DwgXg+\nf9vz8WNxBDA5ySYp/DLtZ9n/ruxrcNKbkpvd/RKAu180s5sr7FOrieXztz0fPxZHEN6TF6dQtv0s\n+x/78V2UfQ2qWo2I/ItcSDzeGhyLSvwrmld3P/bFC588rxCTKuII5hmHECP245L3rzv/EX97cMQ5\nqfAvmdkZd79kZrcAL+W//dwJT7N4xHz+WEhq0vtvYxzAvAnXP+sIiTp57LHe4Hz9LUYH1ccnvrOo\n8NMhSY8C7wI+BbwTeGSK3nWamM8fS0KZlKsO7YgDmDfJ65/24kMyTh7JuP425+sXsfO+TH/IfrWZ\n/Qo4D3wS+LqZvQd4Hrh7lp1cJGI+f8zLX+Gw1XEA8yZ9/cMRrNaidl7b8/WLrOrfM+GlN1fcl04Q\n8/mLCL/NcQDzJutWa4cNdumxw8bUATwLK3xRLTGfP7Y6vMpBq+MA5k2W8K9xiutsco1TJwrZlfBF\nlJjPHysmmTW9b1McwLyZJPyr3MAVTk+VpNPWBB2Q8Gsn5vPHVuVDdFtb4wDmTZbwr7PJFU5zmRuH\n+fp57RchX1/Cr51yoSiOlYoDmDfzjkMI9faOWB7sSvBKDkU/pLo34x40Awm/ZZSNA5g3ikNoBhJ+\nyygbBzBvFIfQDCT8llE2DmDeKA6hGUj4LaNsHMC8URxCM5DwW0bZOIB5oziEZiDht4yycQDzRnEI\nzUDCbxll4wDmjeIQmoGE3zraIO98euwOM+JCZlsz0lq7w2z3WxJCNBIJX4gOIuEL0UEkfCE6iIQv\nRAeR8IXoIBK+EB1EPr5YKPoV+ydn+4fCmllxBF1CwhcLRaxu/g4bDa6LXx8SvlgoYnXzQ6mtttfF\nL4uELxaKWN38UEa77XXxyyLhi4UiVjc/bFXe9rr4ZZHwxUIRq5u/S28h6uKXRcIXC0Wsbv4uvYWo\ni18WCV8sFLG6+bv0FqIuflkkfFEx+VtOrLPHGvvD6rorHE5VOSjm0+/SGzuS9/ldqZsfQ8IXlWL4\nhIl0X9wb7HCaK5ziGhvssM4eqxywxHGh8iLy6atBwheVEoSfVTt/lQM22Bm46Nfpscsa+8NRfxrh\ny6cvh4QvKiUp/KRpFrb86rGbcNFfGfFPInz59CcnGqBsZg+a2SUz+2niufNm9oKZ/Xhw3Dnbboq2\nkBb+Bjuc4hqnucKNXOZGLmdO9U8q/LBqHxbvLnNj5lRfwh+lyIj/BeCzwMOp5+939/ur75JoM5OE\nv8l1TnGNHrsZLnp54cunn46o8N39CTO7NeMlXUUxxiTh38BVTnOFHrsTF//KCF8+/XSUuce/18ze\nAfwQ+JC7X66oT6LFZAl/k+vDqX6P3Vy7L4Z8+mo4qfA/B3zc3d3MPgHcD7x38tsvJB5vDQ7RTPJ9\n+BjjLvori3m9gZOef3b59Cdne3DEOZHw3f3lxJ8PAN/Kb3HuJKcRcyDmw8cCbdbZk08/N7YYHVQf\nn/jOosI3Evf0ZnaLu18c/PlW4GdT9U80lpgPH9uUc419+fQtICp8M/sy/SH71Wb2K+A88EYzuw04\npj+3eP8M+yhqJObDx7bhDttgy6dvNkVW9e/JePoLM+iLaABZi3Ph/nyDnULCH812L+/TK5++ehS5\nJ0aI+fCrHOS2z7pNkE/fPCR8MULMh19jP7f9Esfy6VuAhC9GiPnw6xE7Ls8KlE/fHCR8MYLhw1E7\n5MsnE2xiPnwM+fTNQMIXtSKfvhlI+KJW5NM3Awlf1Ip8+mYg4YtakU/fDCR8USvy6ZuBhC9qRT59\nM5DwRa3Ip28GEn7nmG3d+yJnP2aJI5aH+9MnF/bk09eDhN8xZl33XrQDCb9jzLruvWgHEn7HmHXd\ne9EOJPyOEcu377FbKp9etAMJv2PMuu69aAcSfseYdd170Q4k/I4x67r3oh1I+AtHvk9ftu59/Oz5\nPQgJOSEjL2TeHce3cRQVIuEvGPP26ZVv3w4k/AVj3j698u3bgYS/YMzbp1e+fTuQ8BeMefv0yrdv\nBxL+gjFvn1759u1Awl8w5u3TK9++HUj4C8a8fXrl27cDCb91zNanj/nwMVQXvx1I+C1j1j59zIeP\nBdrssS6fvgVI+C1j1j59zIc/Yjm3/T5r8ulbgITfMmbt08d8+MPIV+aQFfn0LSAqfDM7CzwMnAGO\ngQfc/R/N7Cbgq8CtwDZwt7tfnmFfBbP36WM+fBHhy6dvPkVG/EPgg+7+lJndAPzIzB4D3g18190/\nbWb3AR8BPjzDvgpm79PHfPgDVnPbZ90mSPjNIyp8d78IXBw8vmpmzwBngbuANwze9hBwAQl/5sza\np4/58Pus5bYPFXTl0zebqe7xzWwLuA34PnDG3S9B/8fBzG6uvHdijFn79DEffo/1aHv59M2nsPAH\n0/xvAB8YjPzpb1HOt+pC4vHW4BDZTFf3PmnlzapYRrIHwc6bLGyJe35sD444hYRvZiv0Rf9Fd39k\n8PQlMzvj7pfM7BbgpcmfcK5QZ0Tcp19nj02us8Y+SxxzzNLQQoO+j558/7TT/awZxSErQ8Gvszfi\n62s63yS2GB1UH5/4zqIj/ueBn7v7ZxLPPQq8C/gU8E7gkYx2YkpiPv0a+8MjKXzor6iH2UC6LTDY\nDWc64SdFH354sjx+YBCkI9pAETvvDuDtwNNm9hP635yP0hf818zsPcDzwN2z7GhXiPn0IQovHEH4\nh6ywS2+sXVK04QdgmvOn26+zN2LVha21jlmKWn2iORRZ1f8eTAzXenO13RExn36Zo+G9djjCqHzM\nEiscjk3Pw5S/yDQ8ef606MOPzw4brHA4IvojlmeyviBmg36iG0bMp1/ieFisMoz0+4kbgGWOxkQf\nRDut8NPtQ3BQWvShDxJ+e5DwG0bMpzd8GAsfBJ6Mj1/mKFP0RRfewvmz2h+xPFxbgFHRq+5+u5Dw\nG0bMpw+Ee/qk8H/Pq4aiTI/UseSa9PlD+7Rdt8Y+MCr6XXrDHwvRDiT82imXTw99yy6s3ic9/OSn\nTz5zuXz7vFgBjfjtQcKvmSry6fN8dsNz2yvfXoCEXztV5NPn+eyG57ZXvr0ACb92qsinz/PZgdz2\nyrcXIOHXThX59Hk+O5DbXvn2AiT82qkinz7PZwdy24dgG+XbdxsJv2aqyKfP89nD65PaK99egIRf\nO1Xk0+f57OEckw7l2wuQ8GfAbOveQ/DLsz3zLJ8+WG2Oqe69ACT8yqnCpy9DzKfX/vQCJPzKqcKn\nL0PMp9f+9AIk/MqpwqcvQ8yn1/70AiT8yqnCpy9DzKfX/vQCJPzKqcKnL0OsLr72pxcg4VdOFT59\nGWI+vfanFyDhV04VPn0ZYj699qcXIOFXjuHDGncrHI7Uqivq0+cRy6eXTy+KIOG3DPn0ogok/JYh\nn15UgYTfMuTTiyqQ8FuGfHpRBRJ+y5BPL6pAwm8Z8ulFFUj4LUM+vagCCX+M/Hz6GOn960PmXdjo\nIubDh3PkFdIIVXAOB2dILuzJpxdFkPBTxPLpg4Ansc5eqbr2wMj50u8Sogok/BSxfPqY+NbYL1XX\nHsaLZY7uby9EeaLCN7OzwMPAGeAY+Cd3/6yZnQfeB7w0eOtH3f3bM+tpTcTy6WN7zIdtqk9a1x5G\ny2NPu7+9EEUoMuIfAh9096fM7AbgR2b2ncFr97v7/bPrXv3E8umLCL9MXXug1P72QhQhKnx3vwhc\nHDy+ambPAK8ZvLxw38RYPn2oZz+JrNuEaeraA6X2txeiCFPd45vZFnAb8B/A64F7zewdwA+BD7n7\n5cmt20Esnz7sVjOJrAW5aerahx+H8FnT7m8vRBEKC38wzf8G8IHByP854OPu7mb2CeB+4L0z6mdt\nxPLp1yNptflV5+N17YOdd9L97YUoQiHhm9kKfdF/0d0fAXD3lxNveQD41uRPuJB4vDU4mknZfPqy\nde2h7wyEc6YPx9hnbegEhMy72PbWogtsD444RUf8zwM/d/fPhCfM7JbB/T/AW4GfTW5+ruBp2k8V\n+fLpGcEKh4kAIFO+vZjAFqOD6uMT31nEzrsDeDvwtJn9hP6g9lHgHjO7jb7Ftw28/6TdXSSqyJdP\nCj8p+iBu5duLshRZ1f8eZN5gtt6znwVV5Msn2ydFH7awVr69KIsi9yqminz50D4t+rCTrfLtRVkk\n/IqpIl8+tIdR0YfFP+Xbi7JI+BVTRb58sO6Sok8m6SjfXpRFwq+YKvLlgwNwyEpmWrDy7UVZJPwU\nWfnuB6yOJNHkUUVd+2q21hBiMhJ+iqwRO2mpxUJ291iXzy4aj4SfIi38tI8eS9LZZ00+u2g8En6K\ntPDTllosLfeQFfnsovFI+CmSws/y0YsIXz67aDoSfoqk8GHcR4+V3soK15XwRdOQ8FME4UO2jx6r\nexccAfnsoslI+CmC8Cf56DGjLT8bX8IXzWAOwt+myfn48DzOVoN99G2aff22aW7/tmlu36DO/s2h\nesN2/aeciu15dyDC9rw7EGF73h3IYXveHYiwXduZVLZFiA4i4QvRQcx9tnezZtbc22UhFhx3z1xR\nnrnwhRDNQ1N9ITqIhC9EB6lN+GZ2p5k9a2bPmdl9dZ23KGa2bWb/aWY/MbMfNKA/D5rZJTP7aeK5\nm8zsMTP7hZn9q5nd2LD+nTezF8zsx4Pjzjn276yZ/ZuZ/ZeZPW1m/zB4vhHXMKN/fz94vpZrWMs9\nvpktAc8BbwJ+CzwJvM3dn535yQtiZv8N/Lm7/27efQEws9cDV4GH3f11g+c+Bfyvu3968ON5k7t/\nuEH9Ow9cacJGqmZ2C3BLcrNX4C7g3TTgGub072+o4RrWNeLfDvzS3Z939wPgK/T/J5uE0aBbH3d/\nAkj/CN0FPDR4/BDw17V2KsGE/kFDNlJ194vu/tTg8VXgGeAsDbmGE/pX22a0dX3RXwP8OvH3C7zy\nP9kUHPiOmT1pZu+bd2cmcLO7X4LhLsY3z7k/WdxrZk+Z2T/P81YkSWKz1+8DZ5p2DVOb0UIN17Ax\nI1wDuMPd/wz4S+DvBlPZptM0L/ZzwB+7+230t1ZvwpR/ZLNXxq/ZXK9hRv9quYZ1Cf83wGsTf58d\nPNcY3P3FwX9fBr5J//akaVwyszMwvEd8ac79GcHdX/ZXFo0eAP5inv3J2uyVBl3DSZvR1nEN6xL+\nk8CfmNmtZrYGvA14tKZzRzGzzcEvL2Z2CngLuZuA1oYxer/3KPCuweN3Ao+kG9TMSP8GQgpENlKt\nhbHNXmnWNczcjDbx+syuYW2RewNb4jP0f2wedPdP1nLiApjZH9Ef5Z1+qvKX5t0/M/sy/W2GXw1c\nAs4D/wJ8HfhD4Hngbnf/vwb1743071WHG6mG++k59O8O4N+Bpwm7l/c3e/0B8DXmfA1z+ncPNVxD\nhewK0UG0uCdEB5HwheggEr4QHUTCF6KDSPhCdBAJX4gOIuEL0UEkfCE6yP8D2hCMVUhCw7QAAAAA\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d002d490>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d01c4cd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nHrsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d03b2c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87d02ea150>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f87cff55610>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for m in range(10):\n",
    "    im = Image.open('tmp/out/CNN_%d.png'%m, \"r\")\n",
    "    plt.imshow(np.array(im))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}