dlnd_face_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Face Generation\n",
    "\n",
    "In this project, you'll define and train a DCGAN on a dataset of faces. Your goal is to get a generator network to generate *new* images of faces that look as realistic as possible!\n",
    "\n",
    "The project will be broken down into a series of tasks from **loading in data to defining and training adversarial networks**. At the end of the notebook, you'll be able to visualize the results of your trained Generator to see how it performs; your generated samples should look like fairly realistic faces with small amounts of noise.\n",
    "\n",
    "### Get the Data\n",
    "\n",
    "You'll be using the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train your adversarial networks.\n",
    "\n",
    "This dataset is more complex than the number datasets (like MNIST or SVHN) you've been working with, and so, you should prepare to define deeper networks and train them for a longer time to get good results. It is suggested that you utilize a GPU for training.\n",
    "\n",
    "### Pre-processed Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. Some sample data is show below.\n",
    "\n",
    "<img src='assets/processed_face_data.png' width=60% />\n",
    "\n",
    "> If you are working locally, you can download this data [by clicking here](https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip)\n",
    "\n",
    "This is a zip file that you'll need to extract in the home directory of this notebook for further loading and processing. After extracting the data, you should be left with a directory of data `processed_celeba_small/`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Archive:  processed_celeba_small.zip\n",
      "replace processed_celeba_small/.DS_Store? [y]es, [n]o, [A]ll, [N]one, [r]ename: ^C\n"
     ]
    }
   ],
   "source": [
    "# can comment out after executing\n",
    "!unzip processed_celeba_small.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = 'processed_celeba_small/'\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "import pickle as pkl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import problem_unittests as tests\n",
    "#import helper\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize the CelebA Data\n",
    "\n",
    "The [CelebA](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations, you'll only need the images. Note that these are color images with [3 color channels (RGB)](https://en.wikipedia.org/wiki/Channel_(digital_image)#RGB_Images) each.\n",
    "\n",
    "### Pre-process and Load the Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. This *pre-processed* dataset is a smaller subset of the very large CelebA data.\n",
    "\n",
    "> There are a few other steps that you'll need to **transform** this data and create a **DataLoader**.\n",
    "\n",
    "#### Exercise: Complete the following `get_dataloader` function, such that it satisfies these requirements:\n",
    "\n",
    "* Your images should be square, Tensor images of size `image_size x image_size` in the x and y dimension.\n",
    "* Your function should return a DataLoader that shuffles and batches these Tensor images.\n",
    "\n",
    "#### ImageFolder\n",
    "\n",
    "To create a dataset given a directory of images, it's recommended that you use PyTorch's [ImageFolder](https://pytorch.org/docs/0.4.0/torchvision/datasets.html#imagefolder) wrapper, with a root directory `processed_celeba_small/` and data transformation passed in."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# necessary imports\n",
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision import transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dataloader(batch_size, image_size, data_dir='processed_celeba_small/'):\n",
    "    \"\"\"\n",
    "    Batch the neural network data using DataLoader\n",
    "    :param batch_size: The size of each batch; the number of images in a batch\n",
    "    :param img_size: The square size of the image data (x, y)\n",
    "    :param data_dir: Directory where image data is located\n",
    "    :return: DataLoader with batched data\n",
    "    \"\"\"\n",
    "    \n",
    "    # TODO: Implement function and return a dataloader\n",
    "    \n",
    "    # resize and normalize the images\n",
    "    transform = transforms.Compose([transforms.Resize(image_size), # resize to 128x128\n",
    "                                    transforms.ToTensor()])\n",
    "    \n",
    "    \n",
    "    # define datasets using ImageFolder\n",
    "    dataset = datasets.ImageFolder(data_dir, transform)\n",
    "    \n",
    "    # create and return DataLoaders\n",
    "    data_loader = torch.utils.data.DataLoader(dataset=dataset, batch_size=batch_size, shuffle=True)\n",
    "    \n",
    "    \n",
    "    \n",
    "    return data_loader\n",
    "    \n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a DataLoader\n",
    "\n",
    "#### Exercise: Create a DataLoader `celeba_train_loader` with appropriate hyperparameters.\n",
    "\n",
    "Call the above function and create a dataloader to view images. \n",
    "* You can decide on any reasonable `batch_size` parameter\n",
    "* Your `image_size` **must be** `32`. Resizing the data to a smaller size will make for faster training, while still creating convincing images of faces!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define function hyperparameters\n",
    "batch_size = 32\n",
    "img_size = 32\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# Call your function and get a dataloader\n",
    "celeba_train_loader = get_dataloader(batch_size, img_size)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, you can view some images! You should seen square images of somewhat-centered faces.\n",
    "\n",
    "Note: You'll need to convert the Tensor images into a NumPy type and transpose the dimensions to correctly display an image, suggested `imshow` code is below, but it may not be perfect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABHEAAADqCAYAAAAlBtnSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsvcmzJNl55Xd9jDnize/lUJlZA6oKA0EQbLAJzqORZt2iZOrWZJKZpP4jZKaFFpL+A5lJe2206NaqZVIb1QaRokiABAEUUACqUHNWzplvfjFH+KBF+PPzc1aEUFkV2c2UfWdTtyL9uV+/w3evu5/zHS/Pc2cwGAwGg8FgMBgMBoPBYPj7Df/fdgUMBoPBYDAYDAaDwWAwGAw/H/YSx2AwGAwGg8FgMBgMBoPhOYC9xDEYDAaDwWAwGAwGg8FgeA5gL3EMBoPBYDAYDAaDwWAwGJ4D2Escg8FgMBgMBoPBYDAYDIbnAPYSx2AwGAwGg8FgMBgMBoPhOYC9xDEYDAaDwWAwGAwGg8FgeA5gL3EMBoPBYDAYDAaDwWAwGJ4D2Escg8FgMBgMBoPBYDAYDIbnAOHTHNzpbea7+1edc841G3H5exTqXZCXL/9bb9VJ8Q9Zrj9O06wsz+dz/J6qnCT62wwX9vxPlDPc6mQyK8uDkcrzROfI8H4r91h7/F65kXxp2XMrGuQpkE2OXT7rr2zCp0EtDPNmbdF3ORo/CNU+HtpvOkP7ZOqTTwfdexRGulYQlOXZbFqW2f/L4KEf8p9z7OL4FbX6/F3ymTCdTo/yPN9dx7k6jVq+3W0555ybTcbl7xnmB3srR1+wnTlv8Kcu5TFoL5Y5dyP0aehjjuQZyvkn/o7/s6r/K7PPW/67t+J/PF/16mxsluWtnZ2yzLHPwZFX5vECtz++646OTtYyF5txmPca0WVFy99rrVZZjusN/AXGP+MK68w2zNEOrHFeHRnLzlMFG33ZrS+/PscWz5whjrCcphiLyRzH4PIYAAH7Gdf1V8Ri1v3+ycXa5mK7Gefbvebi2hhvPvrU97kmfXJcOVedKwTXvAzrYhApptbqTfyBjqnGhk+ul+xOH3M4qml9Z5tzLa6MluoAU+nTxNoVB62O8Yvfj87Grj+crmUuhoGfR2Hwid9XxR6uYTX2Q7y8zONXbQlW3W/190/O12zV/K9AlU8zjacEY2I+UznhHM2Xz6F6o16We92ufq/rd46pVTf+wx+/s7a5GEdhXo8XYzdzXAc0/zY3N8pyr9PRH69YWz4NVg7z5dOievySP85XHLyqe6vnWzGOuC/GIcPhoCyPBir7K2JqZYwV7dofjdx4up65WIvDvNGoOeecm4y1P6wMQxzvY2L6XBMq60O29Hc2RLDqb93y3xmtkyJGJyv2TWyz6lYF6xnGaLWGXFOX7Kfc32kb9DNjt1fZPyw/z3iWrm0uBkGQh2VsXH7//J+fE+6Lw5dP0nzV+Kz0+/L1uHJ4cZ7q3m9FH3FNx3Mh1/3KOOIWYMUxKeLxcDAsyxli9qdBnudrmYs7Ozv5zZs3nXPOzfAsOB6rbudnJ/p9NGIdUF4+bypzwTHeCEGg/+NzZF7ZGOIP8svrLB8T1TUSbc9nlRXjtTp3VS+u7108Z2xsbpXlEHX3Vj2cAt///g8+1Vx8qpc4u/tX3X/3P/xz55xzv/y1a+XvVza1aEcZNrG5Bp63oin8QI01SdSIZxdaTB4/flyWT441YC5OTsvyaIzNZaD6ePHigWjq66HtJ2/dK8t/88ZHZfneic4x9vQgNfNrOrenzW2eqxMzpwHue7jvXOf0vAzHLA8iyx6SJ9/979260KzF7vdef90551zCgbetsRJFar/3Pr5blg8xcRMufpUXWxj8mBT7u3tleaPXK8u379wuy9O5FmxOqDBYDFO2GR9sOJ54zQCjmxNujpd/q1/KrR/vvffex+s613a35f6b//QPnHPO3Xn7p+Xvo4vzsjxGuJxg7E3nEx0/0O/nfZ3/fKZj8M7TjeYY8wiu+9vaxO82NF9SvKSbFgsU9/Up+m4yx8MqAnSM49v4nwALW4i6uFhzNOxow/77/84/Kcv/0X/5z8pyd3u7LOd5gjLj12J0/MNf/SO3LvQakftnv/aic865LNZD+Ku//M2y/MLrXynLGV4QpGifhPMGLz/8jC/WdN0sneAYxE2+LOBk8DGRLudiJZ6rzbjQT6f4HXNuMtOLhcFQdT8/0++nTw7L8niqetWbik0t3FQNLxcansqBh7nuaxH9r/6XP13fXOw13X/9X/yWc865ZkvzoFbTC7hmU+UAax7XijnmHDvg4vyiLM+GisHtPa3BL77+df1tX8c8eOcnZXlwflyWp8PFtQK2Ya9dlq+8dKMsR5jPDx8/Kstpprb18cIiRSRNsaZ7bvmmrjLusuXlZQ8d/+3/9OduXYjCwL14ZbFH8H1uJlWHJmLPNh7+X7p2VeXrKt+8fqUsb25ozcvQJg4vLrNk+f0mmNMObX65Bo4x/xOco/Lwg4e5/lCB/hB7qPsPtc86vdAYGmBuefh49/oXXyvLf/QHf1iWv/TFL5bl1qbmQ4p4ygfm3o1fX9tcrMex+9WvvOqcc26U46VkU/31T//9PynLf/T7v606YWHiPoIvMLLKk5jD7yseAHA8Yyq7KSvGgFd5WF/+QZPzKVs+jCov6Sov/gOsy3PF9O/+9V+V5Te+8xdluT3XWGtkiusJxmBeX+yT//mffcutC41Gzf3ON7/snHPurZ9+UP4+n+gmY7RVq6n7quHlcz3SvYep6h8EeCjHPGviIauBcoiPPH6k80/RAU9Oz5xzzp1gDzOZ40M04yA2/I1Y5+60sEb4fLDHOadaI/kxOk35TIU2wIvWAC+pZwk/nmp+//DD47XNxTCK3NVbixcAWaK6Rmh/voBL+NEn5eRSMajMP8wRhw/92NOkc103xPhvN9s4XueZFzFqnKp96p7GQqetvwvxsjoMtH9r+Hp2jGrYr7ZU9zhUXfhi4uxU8fhvvvPtsjzsY3PuuKZ+Eut8jrl586b76+/8jXPOubt39az85o++W5b/j3/5L8ryT978flmez9QP00T3PkN4ivCQ5mGONrEGb3bU5gd4jpyOtV/inM6K+DTGvJmhLlxnM7yIG2D/NcecCNCisc8+1Dxr9fSy5o//vf+gLP+7/+Q/Kcvbuwc6p88XOk7wuDeMP9VcNDmVwWAwGAwGg8FgMBgMBsNzgKdi4rSadferX1985WjVQc3GG3t+EUjw5WU61ZtNz+HTOr5g9Id6y8wvfnfviA3y+OHDsnz85InODyaOj6/aYX3xFs+ri+J0dVdfyH7l6y/pHN/XW//756qLj6+GaYVaCdZRRpYNv6ik+N0JOKZKC1tODV0X0ix3/aIvujtiIGzuiYnz+BG+gs9IZ10uw8hXaOgi0MynE73lPMRY4FcGUhyzlF9tF2V+AapKS/gzaeP8B36WJMV7OYvn7zuSZOqOHi/ejHtO7dkCUyHG295sjC84mK+jif52CMZDgq9IGWU5Ab6m4it1NtWX23mIuYA35HHxFcvD1ypKIiP2Bb+WefwSiS86NX3N8DHWJpnG0QT08Ld+pK8Ejz/+nbLcaevLCRkFeUW+93R01k8Dz/NKSUytpS/F29v62sCYkeJrKL8wOdxvHOhrnh+R4gemYOVLsQ4hAbdBFgnYTLX2glHgNxUv+JV/hq8jw3O1/cWpYsr5mWJ4mog5ljU0Fs4yjcsQFOM2vjvUMHZizN2IjMeK/u7ZcO08z3NB8TWNX20pK6kcjzg3gpSBbKnZFF++EYNrdY3V/QMxcbYgZ0kQOxsvv1KWpxOte+OCrTODhKbeg4yvia+JYF/sidDqzs711XCGr2hkmedUEeGLZ0VmXPm4ukLSUGGuXsoy17c+es5ztYKJkWGuBBgzNXzJbiBORGABepQhIc6yprvb+mrXrDVxzPKxOkfMm05EVx8XY2c4UuydYtwEYO7V6urDnS2Nlc2e4k6I9oyDo7IcDXXNPlg/h4+0R7t3Wx8Nr+3ri2NUxxfnGtajCht2fYjDwF3dWcSoYarrfeWX/2FZ/q1f+03VCV/nK2wvltEXAfYOFXk3WQEVSQ2ZyapnQDZ68dU5zZa3SUWKGXAkcQ/JkwdLjqjeE8mVXwSj6v5H75Xli4/Vp3PuAzneLxkF3vq+Bwee7zrF+tPAuu5jzLQj3UCrpnK7BbYFWBBt9HMdfdJra/7tQvqwiXjKth1ivzTC1/3z7uI54wzPMDNsQEdgpU6xds8wtzOwMsn4mpIJEGtdjmLF6MlUdUkSsvgwpjB0QjJg1r+1WSDPXV7EI6bc4KicZ5wrOIR1BXPVA8sGP7tatJyl7EI8C4L94mMMh4hLl+f3U9QXREiydrycz3Z4BgYTOMLixrWtquHCGoM9bQP7hyoTB3+K8vqfFhcXuJRhjwdiBE/B5rx3V8/n56DzU+7tY18Uo2l7YLb1IAnf62mcbzRxDPbqaYrYnVF+v+iXKYZBVZKvuiTonydnZ2X5+Fz7UipEqmoQnfT4idbL/+1f/K9leXCutfM//M/+87J85foLZbmypnwGFZwxcQwGg8FgMBgMBoPBYDAYngPYSxyDwWAwGAwGg8FgMBgMhucATyWnCnznthoLqlJOav8ECVSRBG+O5Fmk/E4gfepfiH71+FCJ9e7cuVOW799Dct1HouKfQE41Hoi25Hm6raCgHwagoXe3RPfdfeHLZfmVa6JQno+UQDkB5TxHkismkPNArarIFVZkSl+ZnGqJQ9cq15LPgjTP3GmRHK1/okSXj0DXm4zVt5MVdEzS4b1V0iZK63CefEXyyqwi8/ikE5VfSW7JuqxwEmPyzJXpvlY49awkKj4T0uJTw8tzFxaSk14HzjSZxv7RmebErI/M8kNkwMdcJP0wYSJoSCKYfC6kUxzkcjMkJYtAe750KmALhnQwA8U1AOWRGfvJA58gca+H+07Q7wHqePJAMeVH3/6/yzITJ+++cLMsN7pKqOY/g9fduXMuLaiUPUiWmByXUz+FZIUy1CiGhC6ENCwABZ7JVJGc2mOSOVC1211Ry9ub0tBErYVUIe5JTtXuqL4x6MvsNsqpPnrvR2X5vbffVLWQcLvVFVU2mWocBxn6NkA8gMxghEZLZkjoPUXi4DXC8/xS2heCQhxg7DOGzyAnnUDmSPo0f6dErXsVEqpdlZukk4OvnCERY4bEma63kBdPIZvMIRcKIcukVCSrQdpYw30gceEUUrgp4itzVXrZ8ljr5cvXlYrstYjx3hpjsec5FxX3GUF6xNjQRQLVLcgwOui3Gt1LEHs8zL9GpPNsdCVnqlVcnZgQF1IMJDYdFWs2qfYcWzESZka4EZ5vZ0PX19HONTGO60eSzd19IglV/0j7hw/fe78sXwdVPMd5WmizWvhUW89PjcD33UZjEQM3m4qpv/6Nf1CWey04ryEpOxNyV03+KLlm8ukV43PFNqLi1rfUCWX5H1aMZioSQybBXD4XaBDhciax1+81yEmu3bpVlvsPHqjuY5gI1CATbSzWA7pAfl6kSeIuCgOTNhLz1iFZ6SDGdWoq720oxvXqqtN+Z7k8Yx8uldeuXy/L3bbmBWUTp2eSWZwh4fxl32Z0voW0qg/XnhGT/2P/ezrS8Yd4njnq6/cTJHemFD6vbs5LzCjRqySTh9RzjVK4CjzngqhwB2aiZj7jrNiXx5DqR5CcdTFWr8KQogtZKNua5QR7B6YUyHw49BXSmc2G9lF1SPQYO+fYPOU5nodz1TGh0Q/WyNxf7tAZIS62sXYfH0muU+nripvqM4AndeYUxjbf+Ys/L8uDUz0r+5AGhUiwvdGDeyHu8VZXaU72ulofNhhiAqRc8HmXlDbReayog6/rp+SrBEyZoGNe3FVdHp1IWnUPbX9YkVlhDmEfPYHM/C+/9X+W5Rdual38k3/6H6s6kFR/Fl6NMXEMBoPBYDAYDAaDwWAwGJ4D2Escg8FgMBgMBoPBYDAYDIbnAE/Fac3TxM1HCwptBm70ZCw62Pm5KFfDqZw3SEMbXOiYxw8koTo5F4Xp6LGkUpfUSuecG4PC6JgxH64NKajLs3RxznmuW318W7KK996WI9WVV3+tLB90JRcYHoJyyzTXFaWUqKqkX+ak64IWWBHrVNhxdAUq5FTZ+uRUmfPcpKB8pxPRNHNSMytUYtK6V8idKpIk/T5PmTEfUhDcfbZCqbRMzFSRUPmkG4PLloPOTopzRf7FE6EuK+n5q4iK//akVWEQuJ3Wgv53caF5NkA/noP6OwK1dzSC9AnWBLnPLOmiGWZIzx/6On6HNOy2+I/boKs3QWmuFxToMPqkY9XinlRO4GA1GMLxaKbfz8aQhUG6mcL9jjKoFBKVH33n/ynLJ4/ul+WrX5BTxy9+U/Fg//oN5xyccdYBzyvlYZ2eqP/M5D9FO+Ron0ZL7d2Cs1UNVPc8gUQScjq6jcUx+gr03agp+vkcMSktYm44VRz20fdBjHM0dI4AfX5+IQp7+46oyrO+Omt3S+0Ros99UNFn0P8dwgnrbLjcrSeZPpv5muV5OY82G6LGziEtZowfgzZPdygvh/MJxnyOby1bBzfKcqMNR6oJ3CsmagtKO2jUFYSL/4ki9V2GyUKXSbowNkBF7uOgOqQ7lBPP02WykWrZq1xrebm6OKy/Hz3n3KX6rVUHpR3OXD3Mid3NnsobKm90NP47LUi4W/q9Udd5YrjAxfjdD3GPiGF11KFVnL/X1fXpIkQHGh/yyyTRuOxAYkiJDR0zHOTpfTiUHA805h49lMzqozuS4dS25LSXwR2mtqGxu054nu9CbzGmX7ghaewL1+Cmlys+0Ek053YYVHy6k9JBqLr+s70+KQV3ruo+RSn+5Wly7F3yihScl+S4WO74U93esI7cu/L6+oP9a5IUPdySXGU6Ur/HcJFxrYXsdp1yqizL3LBIs9DDvElixfJ6qnF7fVvH7EI2dX1Xa8gLe5JKtHHM9rbWohYcqUJIHonOttbaq3CnSopYn2CfNYTzYB+uPQM4Zs6wvl9greqe6PjYSZ7h4ETq4Jozg1TPb6juY6xBCVMXYK6H4fr67u/icvhllA/ieYZy+jrG9s0txbSXdlTeb6vvNjuKryHkxFNIxfqQIo+xj5hhjowxFy6KZ6Io0Lr48lWtuRn2mfeO9Yz63mP10QTnm+NeA6SUSDF3Q8TaAG1AOVUQUjq0fA+61r3p5Tmz1I3Hi7n/9k/fKH//4J23dBDWk0aouFmHzH8Ha9sB3BFv9tS3DaxRwVwxuo53B/WYrn6qgkcZe/EPWUW2hz2s475M5+vGiv+bB4oL+5ua83cgrbrzUHLiIeSP2Aq5CE5l/WO960iRIsBBTpV9BsdNY+IYDAaDwWAwGAwGg8FgMDwHsJc4BoPBYDAYDAaDwWAwGAzPAZ5KTpUkc/f40YISlM5EAbu4EIXwoi8JxxAUTIds4+RqB6C+bYOKfGVL9MfJC6J43rtzuyw/uPNhWT4EJdSjRCRZ/N4f6bfRVBSnyUgSricf/Fjn2Hq9LDc80b+mKem3oL+mdF/S+SuUr4qkR6jS4EApKyRD2RrZ457vu6Cg/M8hL6FMzPOXS5JcvqL+K+RRFaeR5cnzK8z41WTAxUFzn8N1uRtY1aWBtOaKNQquj/OsbOi/f+5ULnMuL5x35qDknvY15y5GkNMkpHiiXeDqFEA6V0O29wakazevilr4lWtyMLq1CRlBExnqI7pPFf8F9TUk9zCDdA/hYpboHP2J5tkTxJ27h4o7t09FeR2jbVIwkU8fi9o4H+hv796Wy8rhoeiy3/jt33fOVV2DPj885wWL9m/BpYaOIl4syvDGliQB+1dVbrbVh4OhKJ7jU/Uz29yHVI6Ob3OMkQQyixAZ/BuF1CTLRI+dJnDB8fU7M/8nOSQ+iJWksGagDGc11TeBQ8kAFPbjvur44UPRmcc4P79ShNmzccTxfd/VL2VUFSkFrg3K+giOGSFiLan1Q0get3clSdk80FoYBHC/ooMX11rGTKyRQSHdCODqlUN+Mwc9n65oDUj9eqB7n40Vd5pN/T7GupumlL2iWg7lipxqeVs+C/i+5J4xpH+UVm31RHve35FU48qu4uAWJBkbkDn14M7RgMyQbmYeZQ1wNvPRR57TnAoKKWoUan4kkFhUHGjgTFTjOXDN7W2s45Cfj6b6/eGRYuJgpjE6gRTk7kM50e2/pvgSZ6pnegHXnDXDK+LIwVW5kNbqaGfQ6R2o+hyHPiRkPqRCma/z0GAtSdl2kBMm2dIy50Jeuq3hHlbsY4jKnFghMfS5AV1u7FJxkNzbV6y59coXyvIHR1ovQ8ip4sJJ0V+jnCrPnbtU+/iBztuFVHi3JsnQjT3NxR76+fWX5QazDQfPOqWNkKR6kHbnHt3cuL/EeI65d1ocz/Umo1QPG5oAKRZmcLMKMc+5p5uO6UIlNOBwdIrzjClnRfsxpiefYuysA5euvT7uP8T+poX489VXbpXl3/v6V8vyPtzH3Eh7tRxulg57VI9z1FO/53Q8hSw4hXPfvJh3c+wVrm1KVpiOtY+5v4W9Tq5z/OyJ9oh0nqKMNcJaS+cuxuMunofpWpUxdrhni8l07N5//23nnHP/17f+tPx91Ndzc51y39py+fEVODseQKJMCVWEcREhnjRwzmaDjosqh5BxXW4i8owpFhh7Ke3Tn2Vw6pyxLk21fQP77q2W1vcPH94ry4/h+jyFhP/9t/V+4eiBUji80NH+Yf4ZpqIxcQwGg8FgMBgMBoPBYDAYngPYSxyDwWAwGAwGg8FgMBgMhucAT8Uvn88S9+j+Ijvz4SPRp8egh4+nolnRkSgDlWwK95AcvL4UFFIP1L8xsrlfnMmpaorf26A9x7loVnlBSwxBrWqCYkjq/dnoYVkeQdpR775UlsNM8oa5E0XMc8vpl8zEHgQrqK3kdNGqo6CArZN4nOe5m17KyXLStCFloKvTCnp7lZq7wiVhqcfUagnV8qOFMKMkRL9TPuBX3MN0PJ2XUsiGUv/ZUUmfJXKXuXmRKf9iKNkU5VQjZMMfw/liynZEv8cYt61Qf/vV166W5V/9BZWv90RhbEGu4yDzmKOtL8cPJTyUEla6Dr8noD+2EC9acJnYgNRh41Dx5e4TxaPjgWJA/1Tym3ymedyAHOH+uz8ryy/denlRF7i/fF54nueCQk7RaCuuBKCK71yTw8qNl+Wc1WrrmPO+qJmzc7oagXaLxk3RhgHcqeKG6lADzTys6ffSzQiU5RkkTqcnyt7PODIfqb2PjyW3GEGeNoPcIEFcTnAfEzjeTXAM/zZl9EA8TdcaSYUwDN3W9oJyPYEMqg76f4pxRSlFLdYYntFJC5LHg+svluUGpEoeHFpyjEsffePgsuEHaouwWK8CSkUqRjaIqYgjdKrrwBVtCucdB+p3qwE5CdxBkgxOjRXZLeMFHXcoY14/POe5uKCyx1jOQtC9m5H+YQPucNs99ckW3DZ6HbVPC64q9ZrahxIKD/sSD23oYd9Q2R6UEg6dg6ZWGSSxHju3IvfWNSO632FcdiGb28T9nQ+11pxBon50ot8PEY82r6s9crc++Q3h+Z6LCjnv/tX98ne/4jbEDQg6uyI9hFQe+555irg3p9tUpRL6W8jSfMhLodYqx3+a0c0z/cS//3+WKw6my903uU2LECNiSJ4DrON7B2q/DxHLfEigm8VY9j+Dq8pqeC4tGmiAZ4tmV334woFkDTf3NCZ3ehpjB5CG1SB9CuEG46Ps4b4op6rsbrHmZHB+uty8+JDpRHRjw/jLUPa45uGZp4kY0YWEJIMUsoZ4lMPZ8eJY+8ExJCIJ98AYLk307TqR57mbF+uIj/WhDhnYL7z2cln+R7/3W2X5FpzF0r72DlNf8SSj4yYkvz7mRYiYGmIM+JBQeXWM7aJc68qdKMR8TsaYw+jr32xrrJ381Q/L8iNI/h3aP8EejA55ntM569gHRpDdTqfr24P+PCTzuXtSuLcePpEDYYZY1YUb2ib2BLtYKw4oJ8aaV8OC1W7gmBh7csipYsglGYcof73c02S5YkcO2SzfRTimQcG+bArX0BHmeR375Q3skTtd1eWt2+rPs776/8kDuWLf//ijsnz9C19W3T4DrcaYOAaDwWAwGAwGg8FgMBgMzwHsJY7BYDAYDAaDwWAwGAwGw3OAp5JTDQcj991vv7EogyZ7dvaYR5Wl7R1R0uag3N8DlciDYcVkJurRBDQkvmlqIst1rysqZKsrym9IZ43xgmYXgG46HosqVUMWeJeB2j8WdWwICmO9eUv3RNkWKHd0k0jpZgWKbNX1ib+jPpflNVpz5Hnu0oLCGbrlMqgM8q6Kaqoiv3I4nr+Dho9jVrlxVd4irqLkFv0Z5aJMRr6uE2EQBThHinpNUJ6DvpjheNLSvZVNvkoA9m8WSZK4o0K68vBQ2dDPMbbHoNMPUU58Urx1fMdX+Ve+ImeP3/2GXCqub4JOCmlFgmz+80zHhKAl5/mlU4HooNkK6jcHDKVzUaC+bkFmECAu1HbhLgPa7O1D0YwfHcvl4OwMVFzQzCfIMv/Gn3/LOefcqK9zfG54nvOLrPpNUDO39tT2t155pSw3OpByzhRnhxeSlTJuzgeik84g85mc6x6mdLUAnTSpyFpARS/aJ4STRw1tH9fh2IBOnIxVl+EAkljUa4o4O4GDVuZDCgF6dB3uBHXIInPIHHB4RaqzblyGNMpT5pAhDSH9DeNPOgw559zoXMdENbXjzt6VslwHhTi9kHTNp/NTJEpzUEMZq73vLc4TgBIeQDrkEvWLX3E8wlyhUxXm8cRxjuqYIdzEUq55lJ9AaulVXASXOziuC57LXVBQ/1ux9hXbmHPbXZV7TVHwm6Dp1+GwFlLK4pLlZUgLaPDjR+wsjFvuJ7LLunPvQRnIirZE23MPwHKMuN0CrX8L7XECGv35UGP37EwS1jv35WZ15SXVfXdbji/rRBAGrre1oPHvX5EciC6IeUV5BMcaSLzSnG4wcPtBOaXLY0XmRF0w90aQcFC+XuxNKzJ8xz1k8olji5PgOipSisq4m0JSUIfUJ25QRqZiAJe2HHIObtqacSGX8de+l1PHAAAgAElEQVS3F8ryvHSQnYy0znV8ODU25dL3wp6eM3a3NK7qdR3vQSaWI+ZS+ulRcsd9C+pWcftylP9e/jukjXThRPvFkIpk0DX5lecDXbUGaVUHa2GOMT2mDGeg8xzDITip9O2zkTMSucvdrHDFC2ENug3n4a9/WRLxvU2tVfW67jNsSKp0irk1xfNXNtU48X3GV0ir0F4R+jrEWnspHe9saBzN0ObhVa0NCeRF/kj39/pN7d9OfyxJfpYgpjDWVNY2PMugi2rYMwy95XHnMl1BvsbnxTRJ3EXxnDGHYyljXB0Spy1IHvd6WDdacJLCel9H27dbG/gdEn7I/cI6XagQQyGLvexnL9BexYP8K8V9cC32MJ9rkFDVxti3YL7O8OwU1yUdy/JrZfmDe3o3copnh/fffbcs//JvauwEAWW/nw7GxDEYDAaDwWAwGAwGg8FgeA5gL3EMBoPBYDAYDAaDwWAwGJ4DPJ2cajhy3/vuQk6VzkgZFZVtc1O0KboaDC5E6zw+Ep1+NkL2bjhW1EDzrtdFYWuClsis1TVQ6wNmnM4K2jiohCHpxLmoTHQbaDpRqPoT0aDyEDKvQPeX5MvdBMh49UizrVB6+T9L5FRrlu2UUi6c1l/27865vGIltVzu5FW1UvqdsilQ32L0G6VbsxnobzhlraCwXoNs48Vroli2QedL4JJxguz1Hz2SNOYE0r4E9HdSqElnJpWfv1db49+stGqepO7wdEGVnUxVkyHK51ONpSGG1RzyiA3Ik77xC7fK8h/9mlwDrvVEZwwyzYthSgkF7x9yFtBovVIOx5aDO44j/OVlur2BQlnxV4C9TA7XKj/UUR5kU7cfivJ/ciaZVbOl+e3GRQzIOD8/JzyvdERptCSn2r8i+UwAB4rxWNTok0M56d376HZZppsOs+2ThV+DI0enLWpz0ACtGLKoGNTjuJDHRI7U7+VuYwmcMTK4hmSB6MakhCdwb0ngIDAHnXUOevjgVLKND977uCy/8+57Zfnew3tluT/UedaJNE3dxcVifDTRhiO4MA4gN6EkeI77HMKJ5cpVyQW6HY2NGuj6wwnWTsxFR7eREOMf8yIIF/XkPMgp84GcIID8Mp9DrgP9Ty3jGMD6i7FTq6kNZqA0U07MxcRbUdaasd6YGxTn2+yqvW9clRvf1R1RpjcgK2pA+lSDxDNEX/mIGzn2ORmkp6R8X+5bijOpWDHsWrQDpRfVsuPBuI5+pkQ6QJytYV624d6y1e0sLR9DKneBdfz0WHP0yaHKUay4s06EQei2NxfymkZTfZTljIVwu5vANS7Q3KVTFeXiOaTAHqRweUKXE7hWcV5SWpVTTlX8Thcy3lRFbggpeMo9Z8UeqyxFkGH4KeSt6KPzvmSZOcbgFMcEkDGkXAPno8s/dOtEXrQJ5Qtskxj/08I90tHHo34Usikf7Ux5VECpFKUanLsr5trl7iWFBCbjQwfW8aAGVx3sxbIVktEYkvAZZFOMKa0OUktATpWfKuZy7xpX2oyxZn3Is9xNizWqDYetPTgV5XBWHT6Ra6V3rn6kq28Y6T63bmAfgXk5upDMaXqhZ80Aax1dIeng6PxFu9BVbop/D7e1N9vvIV3Ie2+V5Wvb2je2IUnsw/GIfZ1VHHKXe/TWKSNauex5P++Ap8ZsNnX3PlqkPxlhf0LH2BlcqMMQ8jXKjTDEmGIh5BrPUMnnAo97RMgffcZLSFWXOKjmqCMvRBlWhv0PJW4eXBtjxPzQ55qq8h5Su5x38d7jXOP77j05VVFq36trTf20MCaOwWAwGAwGg8FgMBgMBsNzAHuJYzAYDAaDwWAwGAwGg8HwHOCp5FRJMnePjhbZlrM5qJag1k+mosGdnYlim4HOlExVTnGeCDSwAI5DEWhZIX5fJQEijTKoL45iZvIEsqlLpybnnGvBnaoLOtU5Mp8PJ6LzO2S/Dxwo8syaTu7yDNm9k+VOPLlHSdHib9eZbdw5r6SsJ97yzNx+voKmhoqG6PNmqN83Osow3gOtrNkQJfHqgaRQHmj4D0Axy1PRenc3F/TLV/d3dQ6Um6DQzqa6p/uPRaWMMtEdT2aiXtY3RJG/C8nV8YXGwhhuBnSzCjM6jkA6QinW002xT4155tyDweL6pxdqq1O4U52Cbj2HzIoUxtdeVRb+P/zazbJ8fVN9x74Yw5EjgeOVXxmjpB8Ti7HkV1xbUC/QIzEtXYqzVMSGmDcJFVegMdci/e1BCy5LO6L0JhPNywcnklMdHYtmvlWMkzxb31wMwtB1dxbtf+mo4pxzo6FkB6d9lVM46Y37qmeE2BqA7klpShtypopUFdLGGE4BPiikAWJAUEgCfMhQU4yzGeo4R+xL0KHjMeQkAdYI0IojuFM0N9RXbluxY+uqznn9lS+W5V/65q+U5bsfK6a8+96HZflbP7nt1oXcyWVtPILzF5xhxnDhiiDVmOOYGWRje/uibdPNilKc+VjzMkghF6BMFjISSqRcIUvOIYkiJdjHGkqHjxzji2unjwnoQUYXR4r79ZbG+HCs8ZvPIa0iXZrSATjhycFqhZvhZ0Dg+67VXFD1N9uaKzsbkv1sb26V5W4LrhqQHjYhZQuw56g4cEFWk6LNM7rTYI/kBVynhUvZjof5R5eygLEKwTLPUJcMLjx0pIM1SsPX+NtAH/Z6KtchkR9OtXbmA+2Xjh6dluVRqr9dJ3zfd+14sS+jDPpirDg6uQDNv6HY4secQ5DxJIp1Y8Tjo8eSbZw8gpvpGdIFjOGaQ0l/qPaNC7lau6M2afU07uo9jbuojXZDHE85MtC/HmQhAR1UMS997FEePfgAdVdsCrBOjI7Vp+PzRZkuS58Xvu+5duFmM53B3Q5ylDHaNWQMg7uLV4lb2CCEHOdok1VpBFat+YinpYNjJWUC4in2TT4d4egOSNdQSDQhpHEh9lkR5OEHsfbdaaDnkrsnGq9jxOV6A5Lq+rORUxGUZM/g/PPwUPOmGah++3h2iDFX2puaF3s35Zqa4Vnm7EJS8wSxaI7dY4K9TmVNK5rXO5P0xW1ALhRI5hxGmK+QqG911J67W+qLwUPtDThfEjjtch9Np8FGE7JoBJKMcf0ZZHSYjMfuZz9+0znn3HSoeOBD4pRTtoS9d4Lno2SGWILkBzNfY4HOezO0Tw17mJqPdsC12CauWAMDzOcZrs+6OEi/qTNOkQYix143xxhKITel7DOkDA46shzyx9NTPVsMsUZ2tyUR/LQwJo7BYDAYDAaDwWAwGAwGw3MAe4ljMBgMBoPBYDAYDAaDwfAc4OnkVFnqTi8KyQlcNSJQmaZT0YdCX/QxELkdEnZXqJwR6FHMSE6aWALKcYpyElBogfMXdfNB1avVIatwdCQQrXuCa7ZAQ+ynoL/PRemcQvKRgFJJmh/dJyqJ6CmnolPHJW18rTy5XJRPMkYrFlOgmMJdoQYq2a0D0Rp/5auvluWXkTF+b0c04AbkVDU4jGVwdRgOXi/LCSRPtYLm3ABtlmOFrl8J5ByNOmiYna/p+AaonaAYf+d7b5fl7/1EDjdznH9eofsvp/NXTLyeBcfROTdLUnfveDG/+iNdYwB24JBSJdTjoK17+OUXlWH/5gYHhOZuivsnhTHGvM8rgxiOX5ijWVGHZa4OizpSEwXKKCn//B3SyqwyHuD4EmjOhbjubgduG1clKZtjjh71lTX+wf37zjnnZpB/fl7U6w33+pd+0Tnn3AAufYcfaOyRepyjLenAFyFmRKAGN0GTbjQlEelg/LdCOFKFjIWgiiMkXf6a00kKsTemlA4xZQ63vwiOHHRtOjySS9j9I/3uN0Rbvvria2V5a1+xhtI6H246+7ckEewdyGnI/Y//s1sXPOdcWKxXJ090Dylku6MhXA0xF1PEuS4kb62Oyh76ZTrW8ZQlZ1jJZ6muFaeg68MJ6dKViLEzqzhMOYBuLpDZgX4cQcYwhfsfJQeNpijyQQiJaqI1NaOkx1seDzxvedz9PPA8z0XhpdwTsj5Q2uuU+IHeXoPcNqAjTWWRX16mVDqnEw8o9sEKl67yNFijM0g4KP2kC1VeceRA33rLnceCCH0IGUavLSlSCzKy85HG/fRC8/j4sSjkg1wyg7XC81xW7C/OB1rDBpDfbHUlVYwwJhM63DyRq917b/5tWX73zR+V5VPMdQ+xuYG9K6UgdASKIE8KCunMEZ39EHN9OPJs7kvOceUlyUnaiG15S22bhiukc5T0Yax9dPt+WZ6OJHlMUsWai75k54/8xe/zNa6Lvu+5RmsxnmoD7LcRV84hT51jTuSUptDti05x2MP5dMHh3oLzkvsSlLlV94u2pYSRMuPq9gj1hWzKg/QiRHwMmHoB/UaHIH6NDyEL7MEhOIYcrYlY1mpRsLU+BL7vNot9RzuGAxTq3cccPTzWfTZyjadeE7EF+7P65m0dsyE5WQwnsDRAP1ZSekjCEgeaX2HhRJVSZjPHGLzQ2J9lWpcr7kT7+2X5lde1Fn988qbqheeUDHs5ji+OmWaTUnccwzWjmAfrfOTI0tT1i33wyrQelEFNIT1FGw7oyIj2jjHOK4++eC9QdanEngNzh85+88sT4Z+zOVO7QO6LVBEpxlYy0zFZqr9luoIpnl0HmFsjyMxHE607c7TBBGMxW77N+dQwJo7BYDAYDAaDwWAwGAwGw3MAe4ljMBgMBoPBYDAYDAaDwfAc4KnkVFmWunFBDwrJGiY9MaUMAlIcZF5npvgIVGRSpUh/zEh5xHunAA4qdMGgvKtWlFNQzJtNULnpWoUs1A1Q4WFq49qg+WWQViUenTQgRcBrsgq9k5Rq2ol4n5SlJGumj19emxKYqpyKTkbiuO12ROn77a/LDeabv6Ty9pbkOTW4RgXoc9JMSSHfaIDanJBimP2dWjlHn6IMVF8v0u89ZKBvbopu6VCX/ljj4os3REn++PbdsjwYQMKwoi9WMhi9ZyOnynLnBoXMb0oJAujblCc1IHm5ua12uYVyhozsgxmdLzjnQLPHdSmRo1tECunMJWs0pZwEZg5z0DIHE9X3YiCa42iCbPbgXwYISBEkH3kIWmaoMVVDpvg9SKuGB5IAzkC1HA4W1PIsXS7b/EzwfOeK+Hd2DucCjm3QLqd9UYBT0JD3tzS2m21RVVugFbdB1Sa1OYJUJ4ckhnPEq1CLL+tGpzrQ9B2A2MH574EGm83hoIX6uonm5Xe+KwnDxZ/9VVn+jd//g7L8hddeVh3gijCnQ9sancWIPM9dMl1QbicjrQmUPg2HKlNCSvnbjWuKPxH6i8Yq84HkKTW6UNGcAf2VoS3yCm17UeZala6QENB3xq9EYf1L4ODgAVepGWIH3c8op0rpmoWYEmLPkOaU6RWyzDUui3mel+4R7J8cMSkMKI2h9AExhp1FCccKJ42KzKMi/6DMabmDTnbZR95yCn6G2UjHRLp3RnDvSyEbz2P8LaVjoKJH2H+F2EoGoKVnkPxV3I4mmg/rhOd7LmxeOhvpfnZ2JEOKQ93PFLT5D376g7L8wz/7Vlk+vP9+Waa8vIk42mqp3MCellI7vzJOIF0r4mHGfSDda+Bad/HBz8ry6Z2PynJ9H659L90qyxtXFVPqDe3NfDioHp/KNazfF/1/gN8DzPXRSOvQcLI4ZpasT07VbNTd1760kOh/9+KN8vcZZHqVtSris4XakO5UlEhWYhib3K2Sk+J37AmWxQm/4gamOkLJ5oIa60K3KUg3sZ9JcP0Y46YZ6/wNODiOnPqQypsanpEonY0ghVwnAs9z7SJGNBDjO7hejArS8ZiuUimOqUEqPe5Lnhlhn839X9LU3ogui66BeQlnqTgunkHgPOxhb+thDW21dUx0oPhyNFWcOzrTej3HnpNpM6rLWCUXg6oL+RtlRDPIeJ4Fwihyu4U87KM7H5e/c+xPRpAMjdVXU+iEpvBYi2KmUME6ivWVZU7GvNKGTNJC98XFf2eI/zwHnZidxzmsNYluVtxPThHnxkh1MEJ7zNFv07H6P4Gb1QTPnUwR81mk4sbEMRgMBoPBYDAYDAaDwWB4DmAvcQwGg8FgMBgMBoPBYDAYngM8lZzKOefSgoYXBMupofkKx54K5bdCZ6SkB8dA5pSFpGjpkCmyttfwOzOv1wpnhXQOShYorlMnGpQf0Z1FdCeHYxwkVKGD44iPTOVo1pxcdKbfrrhS0P4FkrLykPXKAC6vUPHLQD9Q1hYiu/sXbl4ry18EZXd7Q+4xNTgjhKDKeXhfmOWk3ureQ2855fvSBYrnSEFr9Ty6EcGdAFcJyI+tNKeu88KBXIquX9kty3fel7SqonarakfcMnjPyJ3K9z3XqC3aNwAlHuolh59dE1RSOjMx4/zHp/rjE1AISUvMIDMkhfXmNWXkv34g2nYIGu2lXHE2Ux/N4HJ3OhAt8707j8vykyPMOdCouy2V6zBYCDKNuzxWv9SbcGuqOOHpPnbamsdDyJTuPD68PKNbF7Isc6NCNpRXKJiKKw8+vl2WD++Kznp9V+5wV+BqRMewCDTgOKqhDDeGGE4moGenYKr6AWJkIZtJs+U0Xj8gtR3UckgvfJS9FPKZhsZZbwO0ZTg1vfHGT8ry+exfleU/if9xWT7YlSSOjl55wmC8PuRZ5mZFn80gfZiOQL2F+xilT/WO6OHdnijxlFN5aOtkqDlSgyTCIQbmTv1IqTNj1OWYj8hIhlTE0WUJkbQiuqX5CqULlAqjzetwSgljOG6NMGZwfrr80S0JK6RbFzznnF/Ik0jxrqOeddLuIU0IMeZ9n44ZkFVDSkPnN94D74ZH+CvWGcmsQCXHHiqAPCOjPB3OKHTKCSlzBD283kScTdAGcKoKEDuCUDHd5wYITmz5mPur9SFJEnd8ulg7ao8hMdp9oSxPIet683t/UZa/96//dx1zclaWaw21Yw/7my4c/1p1jZMG5SJoF+6BKR2+dPGjTKLydbUJSTLWX7oTjQcnZfn4DclMTt9RfVs7lGtqLt7pw0UW/UUp2GxEp1kdc/YMZMaBy92WW9zzQVdj7CHkNo2m7qvGcYu5kEBugW21y7Dn8CHh9SBhqmwXcXye080G/VLIbBhtuf6GkATNZjpqjr0V97E59zCO0lo880CWsrmh/cDWgcb6mx9oz/Dhx9pTxdgXhdfk1rZe5KW0MuLzBaRStUz926zsC7CnRSyKIWH0sUm5e+dBWR7DIYkxtQmXq66nuRtF2APVFuVOR880QUN9N0Jajv5MUpn3P1Y7f/e73yvLjx8+VF3mdG1cvo/0KaOlFA5Oo3U4/Q6H2m88G3jOFWsaJXhezucsRivENVgt5VgjHfY2DmtqgHhab2s8BzwGMuwQ5+HaHBT1TfAw5Du0PcbQfI40BlgvIzR+hj0k5fL5HFK5CRzMIPlyiNchRuNGW88WLezB/c9gT2VMHIPBYDAYDAaDwWAwGAyG5wD2EsdgMBgMBoPBYDAYDAaD4TnA08mpcs/lRZr1DFRhMinpwhBWKNDItg8a6BROGszMPIekx0EKlYBSeXwqymsHFLP9bVGxdjYXtLgeqHRkmI9wvsenolY9PFIG/tMp3GsgrZrBfcmviQoWou50fWJm7QxyIDpvuGyJnGqtkhzPZYXEhQ5ZpMhS9lWHlO1gWzKFZkMUsDwglV/3S+ehAO1AV6OMUjVQDH2U84Li66OOIWilOSixKem0bhkFv0qPiyP9S7MlSnSnq/sjt52OH5RFkB1JCZWfr4/yT/jOuVZB/8tnkJbRBQhyww200UYDzlPso6b694UNyckohRmda14cPrxXlhNQ9CkvyDJk0i9kdJSZDEFD7MNxaQpZSqupfjnYFyV8e0uU8FqoG5lN1OYj0PbHcN8aVXWOZSnCSNmB09PZcDHXA1/1/fzIXVq0yRAOaD/82x+W5fsffliWb+xL7tdpSXpDx4ZGW7/HTbVPUFfZQSLiIAtxoH8HmPch5silzJGOcKTS0+0IKjU3J5UYx8AcyzVm+oNNyKn290T3rtXUHvfviUL91k/fKsv1r8otj7TYPF1nHBXSNHUXZ4t5kWIezKa65wS02hRtUa+Jtu0hLs6mOv7JQ9Hg3/+BxkaIdTHqqX83djV3dzZVhprJ1Ysx79W1buWUTSFucU33KvOGsRsSLh+yGbh50CGGMYVybFK2Waa7WXAZeNdoT+X7vmsW7lkbHVGduyjX4K7lwY3SQULlQz7jQe4QQLfmh5TBcXGB5InSP3/5ulgWuceAPJn7rAoVfqWDVba0PMN4nVEGFMNhDHKxGf62gSAQYs8TzZ6NO1WSJu709Ng559zWPijuQ61bDz58tyyfP5DDUwdrZNxQX9faipHbXY2HXgsyVkiPmpBW1RuUmSFFgPukXITuRCHlb3A1SSHtSTFGEoydKeLcCO5go5PDsnyRPCnLA845xKAG6h5COpBAujwq8ht4a3ThbEah+/r1hTx7cq71+xxS0gh9Ral8gjVkTCcpSJ8oE2MGhXoL6x/3UWiTFA6HdFa97M95NF/6dwGcejzuUYear0M8i5zBNa0P+dEF3Bz9Gtdr3QjryHWcoaaJ8gFcltYJ3/Nco76obwuOXFe2tEe5tqFFqQfHqDbmUNzSMRM8vD08Pi/LP3rvnbL86FzPhTOskXXsRXfamrtbW1qDN3cXdbuJdAo7u/r3FGvVzz7627L85lsflGXK/P/4t36tLP/pn3+7LN8/lvMbXQmptK1mF1Fft+AWeOLpXitL85rg+b6rFc96ucdnex2z0aM0DTJADDgPEv4s1vEzyDo9pCTJ4GbVhNwthuQqhCyqhr1uo1inJ2PFCx8uVAlkcMmI7qxwjHJ6FpkHehZJPaYIwfxLMC+num4b+/HamWLDxqb28k3U/bNIxI2JYzAYDAaDwWAwGAwGg8HwHMBe4hgMBoPBYDAYDAaDwWAwPAd4aneqyz9h5ukUDKAQ1Os0X049DHBMDCepOmQBnU3RkJod0ammE1GbZsgUnQ5VfvQIVNGLBY326o6cCqKaaF4PD491LOjvMSiGTWQGH89JrYQUDFnLPTjo5HBfSeeqezJDVnFKx0A/9oo2ztfIk8td7vKC+lpxvcjd0v+pOEbhmBkkSU8ga7v4+L7OAjlPpyHKWBeyJTBCHUwGXAR69iUtNacdygoaeAoO4hRjbsbyBHIeODNc9DGeKk0OSYZbBf6Lt6K8TuQuTxb3kecYYzmo/ZDFHUAe9pVXbpTlr3311bJMZ7FpRoqyxnm7p77bbMitrF5Df+Qa23NKmApXkHPIpk7PRYkdwsGnBSpmDXNxpwc3gybcMxJck5nzMR7qkA7NMPinM9Ef6RQXY8B3W4u/DYL19WeSpu70bNEWP/jej8rff/ampEGvXL9elne35frVBMU4rqPfMOcujtXORxkong1JC7obkEj24DKHWMz4XkpZapDDgNafIobOIPOjI8hson6eYCzMh6K5BqD472+JetoD9fT0Qsefw03m/ELnpINHnj0DvrFzLksz1y/qksB5rd9X/ehUlCNedbqgB0Ni5DCPHRxMGpAFD851n+++89OyPH1Dfd1rah29deNmWb5xczGudg8UCxqgmAcB/ZE0pmaQwZyfixJ+eqZ1dNjX+EphC3P1i5K5hYzvdOTAVSmv9n1Khj75758XgR+4je6irXYgG97AnIghp3KQjFLKwjUvhbQpwDFBQnkiY5XWJbpxBWjDEP1y2Saev3xc03lwBklOgvJooDE6OFO/DRCXR1PF1gtIq84piQKNfoy9TTbUedpjzO/2s3FV8Z3n4mL/RVehBx++XZbn53JvunFFroo9jMOzx3J7ch6kOJCiRpA11CAtCyGzD+CgUsNeN4CEIrvsJ7opoS7JlI53cPUKNXY4hyi/91M4ZaWQOWMf7SOOcocVo461WHWbQ4vTLBzyOD8/L6IgcFc6i7n4iy/JaemDB9pbOqQmoJT75FR7/4fYl07hWMTYSvljBy6rOyjvbSiG1rH++9h3RcXPPqRPlLvNIOU5gkTsnbuSuL0LefB9rGd9SMJzzLmtfY3dvK7nGwcnni6c5W7sKa51sUc7+eBn7lkgCHy3Wch8b+5qDX/pQFKlzTr2Z2jbOmJtivLDY60zHz1W+fxMa9F0pHt7cq49/dmZ1ssI+9sWpK5bm4v2evmG9l1fgBNvF2vko7u65sm9R2U52ECqiWvaI29jLXl4prjoh8HSsoc5RQcopn3gGpit0Tn1EmEYuZ0ihQHlVA2MK6ae8OhO5Sv2XEx1X2d4/prkirPTqeYCHXE3ekqP8uorX9DvXczXLUrSirohfqV4Dh9NdR+Hp3q2uf9E17/36I7+NtcY8nwdX8Prk2aN6QfgmtWBDA4pAhp4p4Ew8pmeFo2JYzAYDAaDwWAwGAwGg8HwHMBe4hgMBoPBYDAYDAaDwWAwPAd4SjmVV9K3MtDRwBR2+NklyJJfAw2YVP0aqKcRqKcx6Pq9TdGpSBUdnYiSNvVB570Atb6gVDZ8na/R0TVnyGC/syeaX3ND13x8IXpefgx3mpFufALZSJbqmAyyENKlI7jpUCZEd4lL1451igA855xXXMOryITI6dKwSOkkBvnaBG4PDvIxZgEfwxnh9uOHZXl4oX67zF7vnHN7cN+5efOWjimkWGTP55T7gE57fKF6vX9H9NvDU1DFIbcIMV4DyHAOQafOacLyKahvNKTKnpmcynN50U80MmG5g9n95Rui3v7Sl18py62uqJkjSGEoRWP7TpHxPYIUgH0zh/yK5aSQ4A0hszk+hyQGcra4Bjkj6IkOMowEc34AyvkAEoF0rHlJ6mkNUqw6nC4mE7q7qO5h0Y/eGvtzPJ66N3/8vnPOuR/84Cfl701Qavd2RIHutDXPKPGk8wwUBK6GmJfiXvqnkhPMBpqLnVO1eRdZ9ZugEPuFU5jf1rhJQG2fg7ZOScZFX+WzU82t8Qh9BXeOJKUji8otrB0eYgBd4KYTxi1gwH4AACAASURBVGJIvTKKBdaHLMvcqHBcyxO18xS04UZH0otOV5KwzQ253QQhJGqg/N95IKr2ByiPMY9O4G7mjzWPfEhn7mWaI2Gy6IMATlJ7gVzAcsiF6PD24KGuf+e+4usIMo/NLUj0WrrvGc4TYB77FUsOFSmFpFrkWYjigjBwm4VsYqNL1yG4LqEtx2jjOfp5jnk2q7DbIetzK6QvuMlWU3OxB9p4BzTsRuGgFEWk1AsJFoNjuKHcvydXQcoQHPZrDcwzH3KDyroLCWsDbjIN7OPoKOlSuHxAwrpW5LnzC2e3MaSax3Nd76BF5xPdZxPOI1tNtfl4DJo/XObyHLJsjO3JRP3YhkQ5CjTvI1wrLuZCBgejypo7goQRkpAxpDVhHXIu9F0AmZcf6Joh5ll7qj6aTHFdyKy2sB4PIN8NCqk/HeY+L/wgdM2NhXT4l7D3f//+XR2D651BPnP37sdlOYZ72AEcDvtDyG2OtBbeu6fzQx3hXrmmv/3CLclPO3DluUx5MBkoJg94HfTbm+/LYfEjOA8OIeVOIAVpxHCzSpEK4oHmsQsUZ7/2C6+V5R2sO3GGZ7CR5v3xQ8m41gnf81yjGHPbPa1zO1gfOjHi35JnH+ecO8eec4ByB3sQl2m+Jona7iLGc1akPQjdF5s4T1zkdxiONc/v3tcY6bQgFXWaZzdvvVyWTx9qDP7oDblJnqH9vYDuhnAQq5Qhp8JzJGNKiL1iVqT3WKeZcRzH7uq1Fz5RNz6rUxqazNXeT87VV1mmedHE83dIiWkNcs8EMjg44t6+rbmzCxdqN8feIl/ERR9pSlI851+c6f3Ax3fkTvjxQz2jHmO/mmBc1hFnY/yeTVXHdgOuwJAR0nk64mNEgmeUbOaeFsbEMRgMBoPBYDAYDAaDwWB4DmAvcQwGg8FgMBgMBoPBYDAYngM8pZwqLx0PmIyeMgVSuXJwo+nsEEZwlQFtNAGtKKPECEzNDMdPLkTX2myINrgDF5dHhRMK60gKWgOyhBy6mUFflKtkAj0NnBfqyL7tse7IRJ+BNhs4nYeyI4d7Je1LN75OAnleuhaRUph7lFYtp09P6B4D6RHMvlwIyiAdWSYYCz7O6YHCTbpZClnCZZm0vRTU9gGoqu/+9N2y/J033izLR2eqe68rqciNK1fLcheOBDU4xYRoJzp6+XAvIoPx2XjgVJHluZsUc4H0UY7hg03NuVevisIagLJ3Dup36un4C9B2J5BEDQYa/wHkGbtdjfONFt3ZILsshkwewCXKqZ0vZnCnauiYDLqEs4eS4tTheEZJ0aivuBDi9wBymmwuuma7pWtFcOFIBqrPpesQ58PnxXA4ct/928UYPYOLws41OU30IHfb2lCMuwLp4Rbcm2bot/5A5xxOVA7gZpPBbSYeqZ+HFTkT+sVfjKO4DnkIzjGAlPXwUDTkAeQGMHByXk192O7AnSVEHOwotu7uiULbeSwqehtuXbMRXLEgW03SZzMzs9y5aSHhmMN5K8XYo2x4D7LdCHKHIcbbJBGd91/9xV+X5fc+fK8szxFrdzc0Bjad5iLlWmyLew8W0oEWHDP24crk1xTzTo80595554OyfPehnGDqDVG8W03VZXcHThpO8BExg5ALPI8RKNcsZcBrdafyXa+QKtH9g5WeYu3vQxI8gWSzgXEYQRI1nVHut1yWRefNENT1HuRdV69I2rF3sIgTdMMkhkONpydPKBuRfGII6SndrHoYF82e7ol7ty7u9TyGCxUcf+YVh1Ldd55jfq8RcRC6F4ox34eD5gDU+qQBCj/q1wZVHiPAtZqQi8Kqb4Z1cQh31DGkhReQC6eQKm1DJnTZvinW5bMLzbkHjzXPGDsD7Hm9uvrCr2FfSh0i/thHbOqEihdDjNMsV3tsQEJ1F45jSVw4h65RTuV832XF9bYijZMv3pJj0CPESjol3rip/VynrTGceWqrVhcOlJRVH2pNi+E8i0cE52HP43uQvxZ73TOkHPjoruROj08lX5pnihcvvwiHT8jguBeeQip+jrXtyUASjlmiPU//TPI/PuBNsS8K8JzR6m7gqMduXfCc58LiWSKC9L2G/qo1qZNVnSaQZc8RN7xKOg21aX+sNorQ73lfv2/XNAaud/QMcPO6HNB6O4vfa4gFHchZA9T9L7//47Kc4pmPe7bzPhwc4WzsrXBnrJQpJ4ZzVxuSWqYduUxdsF6PKq+U+dKxsiJN91h/3dcGnE8bHe1p80DrWR9x82IAiS322U1ImhPsRRM8uyRTPHPMFv0com1mkDtdnEsGPp5orjgf6U6wX/Mxi5jmYQfpDVqR9nSToeZQgGfHFDL/KfbJd96WG+3VmyanMhgMBoPBYDAYDAaDwWD4/yWeiomTO8+lBTskqFJxUFQ5wFtFsl/4hXKA5HMJkgPHeMs86usYD+8ZQ1xrC1/P4vonWT9TMGXiGd6y4Y3fBRNxTvlVRG8IR3O8FcTXaA/3xETPKb4AVJL84UuUD1YL2QJJ6Um/znerXpk0Mw1UHyYJDXEDbfRbG8m4Bkha9fjJcVlO8BX47FT99viBkkYF+MpxZV9fpPZ2kayVX839y+R5Gk8zfPG+97ESiZ08Vl2u7+rLzXyqr48pvoQePtJb2dhpjPSQ/LCGxLcDMs2YwRifbJn81l/va3Fc27nLD3tk4jiMpZ2u3hqnSJp394G+yo7q6t+LqcoPnuAr7pm+BKZgsFxDwtsv3dDX915LXy7CSG2azIrxAGpdD4n3zjHPjk9UPj/WOfqpxsjFWH3nPMwnX/e6i/NvY77ut3QeD8yUCAldffRjfpnQcp1Z41xexoEQX69bSFy31VX9D3YPynI91pfXQyRnfANMiYeH+gp8H1/it5Bs7Zu/9LWyPK/r3rb5VSzVV4O8yPhI5mMCNsHhob4+Hp0phj8+Ufmtd5VMbjwFc+xAX2tiX+Nv/4q+cmyBXbazrS+I9brqg9x6boKkl0wKuk5kWebGBTNwgi+CNST/63XAqNrCl0J8ueqfIWn+UF/wvFw3tLOhNprhy/rv/savl+UwRwL6E82RAIloL44W9T0/xVd1rHl5rK/Cbz1UTP3XP71Tlk9x/W5N/fje+4r1v/PrXynLL+8q1ucev2iXRefh6zaT/DuUg+LL3zpTxvu+7xrFHOS+hfM9wn6jzsTMIMKETAh9oX64/0TMipNTzScyTjfBBN0GEyYEA3UywV6kMBeIG2Iekyk4BvttjqT1YyTKPTzSOJsiae9D/M7r37gupgPnXwgW5RTsniESPWex7jtsadytE40gdF/pLdrjFOvQxzPMrSlHDsYV1k4/Ijtax5+d6j4fHsMsYarzR9jbdfCVuo49YryjtrtMpjrC/vekrzY8O9dc8ZDgNMDm4v4IyeKxN9oB66gLZpiH9caHQUHqVPdrTa0xNeyfR2DdTItkzZlbHxMnzTJ3WjBHN5qKQ52OyimCRntDsbWJR5qLJ4r3dx4obs0a2rfcwdp5/6PbZflLSGC8qWZzGeZ9CoZAXDBNmNj/w4+0zvXn6sNrt14qy40Nrek/fU+JlS+QcDwEa59GLBH22r2Gfk9hVhDkapvJSON1MtR9b+4ofqwTucvdrKj7BGsPAzfXSGgTXJaTcobkvdi7u4nuuVYXsyZD/w4GSoR7vaVjXruuZ4OvfPELZTkqkhxnYF/tdMB0RuW/H79Tlo9PNNZufeXLZfn2fdX93iOxPjzcU4hywP0nxjiTHJMp2kJS5tElq2WN/g2T4cC999ffds45NwdTNG8wYbvKG2CwbKLdxonu8f1H2iPOoHh4cqwYtoX178YLYqvlUx0znGmcD4cwIJgs9oI1sEYnYNycXuh5xothIHCgMXEPc5H7kAbcYh48lsnDqy+ovntXtEebYDOae7rv++9LJfKX/1JrxFe+8bvuaWFMHIPBYDAYDAaDwWAwGAyG5wD2EsdgMBgMBoPBYDAYDAaD4TnAUyY2ds4VCeKoJCEFzEOy1wjJmFiOQUueeZVMsWWxiSRfmx3ROhtIVhSDztgGbXR0LtpuUFBYPchJEtCMY1yzw4RwoHlNR6Dw5aBtgXYdJaL2BUiQRHlRClq8QxK7kHIpJl8tyvlaJRzOXb6743mr1HQknSTVLwKNFjTwEWh2TSSJZcLYi6b+dhOJwvwAiW9B2w7qokRGzQUNn4mYR3PRF+8/FjU0QWLf/QMlgaTEjQm0Ww31D+nh02PR1n3IqdwciRrZL5UGXPX7+jDPPXc4Kdp0prY9qKmu+5C8nJ2Ievg+EiV+2Nc4PJ3oPCPkGDufidq93UVS8GvqX3+qOdUEnbwW6pjjfFG3DHT7BpqK0qf+WP07nWhORF0lS+sjadzRqSqc4fqPjlWvbSRrnR2or/0dtVm3sVwC6rv1Sxvr9Zp7/dUFtfqjDyV72d5U3Q6QyHRnD8nhEEou+qD1g0odYT6tijc3QCuOkUDRRxvOIJXoF/KbdkvzczZSf/bPRXdlEsIMNPALxOcEczpBDJ2NdB+co13Q61++qYSEEfpqhqTnj0B5fR9Ss3Uiz3M3LZLWzSFJ2u6JTryzhYTMTY3zMEJi/Q4SjA50ni7kga0N0X/bHc2Fq9tI4ol44DVERXZDjZPT+eL82RwJp8fql0Zb7RxzLd4S/b8bawy8+IIkbz/79p+V5cmcCf8h20CiWcqxU9DocxzPoey8Z6BRzXOXF5OKSdITSIyayHC6gTXs7ELj/85dzeOP70lWdoqkojOsnV0k2b9+TW1bw7V6PbUz19dLyUElpywNHLB27+8rdvj4gytI7J9Troxk5aenoLlvacx1K3Jd3dMEcWRGU4CJYvpo9GzkVF6aurBIJrzb05ybdVVOxqrr1Nf4HCLZ5PBEa+TRoeo6gnTglN4UEWRL0Ne1sQHoUvIfIeF+sV8MKmkJ1EctxNoQv8c9SEggoUghSRzBCGT0GP0IGbsX6fwdJEjexFwMIXvNR6rDpXxonQn/c+e5rBg3U0z81qbutx4oPrZ93W8Qa3/e6KpOOzBHiJAE3tWRPBax8MWXbpblbcTCFtscUth5cf9ZrvNNkVibcrMeYsfOnqRM77yr9akBmXuMZ6fsTHGkCblID+tiE6YQza39stzHPuHh+5LIRpDOrhO+51yjqHsDzw6tOp/hFOci3I+HvavfUJsHkJFfuwETB+wRz4YaG81XJJXaRhu9eF3tsrOvPvCKZ8o5E39Dql1H//+jP/7DsnznnpLZXkWMHGBM+YfYG62QU/kY7/6KhMf8vYNn4yNIdtcFP3eumSzG9g4S2ddrEcqaB12kAmjCJGSKZO+bcMLZQLLyDqWTkMS99vLLZflU2zk366vNZ0i2/uRBIeHbUBzu4999pF7Y3tG6WN/UdY4HSI+CZ8rXXlF9H777A1UmlZy2WcN4xb60CYlejmeqdKx52T/Ts+ynhTFxDAaDwWAwGAwGg8FgMBieA9hLHIPBYDAYDAaDwWAwGAyG5wBPKafKnVdmaqYbD2Q54PkHnk4fggJdB7WuuyvqWQQqJCmBnaZoqKRxxaChzeEKcgG5VLOg6zWQBb2GzPxhCMcJ0P99OG/UkL19lqqOx1O4NYE62c9F55uhiTNQTvNMNF7KDuhydUnvXq8jjnpuldKHV6MEzAf1sbcjOu4OaOARjk9BRd/qiSo5B926jn4+ALW71dH5o9qCWpehxiGofbtXdf3OWDQ1ZnR/patjwE517YbGVo7xen+gTOJJxZFqVV+gb1HP7FlQ/4t6TAuXrxrolaT+7m+IBtrs6t4GD0Tfy0DtfwE00Hlb9R7MdP6Drtrray+JkvryjsZGH05Ic0gxhhcLSuO9B6I29nrqx05L9b15RW24kWq+DiCL3IjgbAXqNCUcDvdXR5b5OhypPEo40F8Bxrtf0CLpvvd54eW5CwqqZgsyuJduSSZ0FXKnZkdU1WwGF5GrmjchpDrDvubZ1774almOQtjpQH4VQp5aA0U2hbQuKdwhxpDm0AXHQVYRg/p/FVLFP/iNX9Hh+ksXw/mjHuq++anh4wfi0x4Fmuvzieisdx/pmPc+kDtFH7KXdSLLcjcpHH84n9qQT3Tb6jsoSF3MWNTTMXkbNNxfVN9lGdwC4VTV2UC8DFWu+arPyYPbZTksXOa8QHPlHDK3zobm9mtXJMmqf0NSgOFMc+jkXHP+lRfl7HL9hv6WdHm6RTKmpnAudFgvK9OukCys1Scuz9y0cPWhYxQv24DrZRc09hZcT3Ks960WJJuQEITYi8TYi2xvwr0L7bAJCn+IvVNSzDXK1AK4ITUgMW+D5k4J13isfphChgixm5tDWtXAfYzhdkOJVr2u66YJ3Eo9nd/P4Ba1Rni+c3Ehm/Uyxb/dmuLiDDLA/qlkUzncNx9cQPoFiUNtS/f/QghJKdx3zuEgNoQTVu5Bjhdp/EhyDylvU300wd4yhP44aeqaCdxfaAQ0Huv6kzEcHOHsswtXtA42R7VUa0OWIkXARHFi0D8v/l3n+9zInfMKR7AcMq49ONZMIF+IIFXysYfrQE5Rh1MZ9+HtrtaZl/cks4oxjzqYRzEkVHRzu3QKhNGXa0aQG2Ju1aHt76LdfvNLr5Tl43NJxftYo2eQFjcgLUoijAvsYVrYX21T5gNn0dozklPFQeBe2Fpc52ADKTEgs4lCtXOO578YTmq9Xc3dnX08HyE6jyBXHcFx8Qt7uucZHYqbmscZ9jpRIWOtT/Hcg4UmwoLw2ku3yvKN61rzUqSXuHMoSa0PaQ3TO3CRqbg7U2aF+OphjG/0sMb494p/Z/T+fAg9z20UqUVubsPFrKlrtOD62sQ+th2r3NvX3NpP9LczSIx2byp1QKum8elj/LexVxjjWTmdQNJfpGLw5thvwFW3jjlczyB3G0ti+A9e1P6H4yZIdJ2Xr2n/tdFAjI4UFwZYX7fbmgNxqL3EHuYlZfefFsbEMRgMBoPBYDAYDAaDwWB4DmAvcQwGg8FgMBgMBoPBYDAYngM8lZzKc57zC4p8SEsEsLdy/E/FUAL0YA/03w1Qe3c2RXmkO1UImj2zdweQkQxBraqDHpgU0iafFCrQfRsNUa7oGpE50ay2kCmbTdYc6voD0LWOB6BogRJHTnhOR4kU0g60X1hIPubPSJLzaQRVJOaRBk5px8a26GBeSmcPUO6QJT+Bw1MLtMn2hqiBXg3SkcLhyAfFtdkWtW9nX1TGKSjeIaiaccws+MgMD8nVoC9qa4q+Yqb6pzWeekbmVM5zkq5Fvuq9A/eaXZgwbG3r/tvbohB7oGE3A82LCVw7HBw59mLR/bZ6Ov7oruQUP3r7QVn+xa9IGnR0uKCZvv2Bsspfh3nOyy/CwScS/XijAZot3Ol80KiTBNRJ0KWTlPRbgc52HnopxXjIUb506KFLwOdFmqbuonB+2QPt+SbkUW24qtThTBNUYoburAnKeZLCkQPxr8q2hZQFlPnZXJTgKbLzN+MFDXQ6hdMQ5rOf0d0LdFo42VzZUR1rmJd0XcjBN74Yiir96JHGDnUDJ2dwX3v3ff3tQHF5jQYqFeRZ7ibTxRwMPMhwIf2NKk5nAmXAzZbGMCnK27dEM/ZBww3qipcB5CIIk24+VrtMIIWpF9T16UR9fn4mqvDWVbXb1oFibX1L/TgfQ0J8Cnr6S3KqoqQhCeE+grjDbuH6zr1EDpluKTleo8w4z/PSHW0Kp71khvbG8TVILWto+yYkVGNIQeZwGPPh5plmXO91hRrWqAblSdhH+cmirXxK29n5K6SfLTqIYP2dzZZLYug8NINUdQS3mwbcZA721P8XkCXMUc/YYX1ZI3LfuayoShOBbtyXPCpsYj1pq97HcAx5+85HZfk80HkaY/RRxSkV+8iR2qv/WG16O0N6AbTXK43FmBlkquNRpr9761DraR7C2bGFfve1Nhw90t8ePdA5L/qSQV3ta07/7rbWmAASSco/JnB0mc0Py/J0vIgTdJL7vPB8z4XFmh/HXON1TAqZKJYcF9GpyuNzA/54qnuJEH0o56ZjUMQ5NcP8gwzbK1zAArQD19wEErvTh5Lw7WCeb0YcH0hdEOoZKYLcIk1VryfY955hHk9G6v8GXYSQRiB+RpvUOA7dS1cXsaADF7DJSOMwDNTmUahYVGP781kQWuQ5ZIY+1s5GoL3UFNKmiyHmMZxQW221hbvc63jq53Sqvhtf0LkXa1Wu/uJ+cmsHEr0a9teQ4nEF9CoOdXD281nWdSmnqhXnT5I1zkXnuVoxj/Yg2c7rcIfDGG6iXMf+r0a5Ldx+55xPdOCCk2EGZ9U5XA1zzCkPMv68eEfhpXBWxlypSP7nOkc8lYSx4TSeHPZ0lI1mTPnS4LM9ZMmow5UdpfQIPY2/ra6ejQ8O9Lz0aWFMHIPBYDAYDAaDwWAwGAyG5wD2EsdgMBgMBoPBYDAYDAaD4TnAU8mpcudcfsldpB6IEhNQ8yh3oiNQDTS0JuiSm5A2dUBzjUBDC2uUSumcc0h3RqCqNQvaaoRzOzhCHFwVVf1iIEoiCWld1JHOIgFu6nAsOt29kajo41z0qzkoZbm3hB7u/m5m8fVlGf+s8NGhUbRcksayDzprXiGg00EAVMmADkAq56CiZsXxAWRlMcZBuwOHM5wvgANRBApfGJEGCUrmUOcfTkHhyygR5Lj/+XT+Zymniou6NOGwtr8haug2xvwuqN9XOnBbwBgLIFmYoS9IG+3ghjJPsoN3H4tineQ6Hixz120s6vvSVdEyPVAVQzRngsndinWSdkw5lfoIxh9uRokQncLYd5W5qGMoxbqkSDsnhzzSWj8vsixz4+EiVlwDVTWgnBFtGUQa8y3QrdH9rk13ODh5ZYiJXsrzKz6NQW2lgrDiclPE0wbmGRRRroU65ujbHqRgEWRGFUesCLIR0F9JD25iQNXQnQEj9gpqf75WP6PqeS8dT2JUKogZw1g9yFOGWiuaGJ+kgVNaXIf8pQYKuYd5keNa+ZwyLriwFXK52Qmo4hj7477kkV243DUaun4nUqyBasrNU8UF9sQEfRcm+oOgIjlG3fE/OdqM0qp1wfM9VyucSUYDUbnHaNd0JhlgChlaTBkcnDTjOiS5HHs+Y5J+DyHbiCAhR/h1E7iwlfTzyqn9pWU68lCeFUHiltcRF0A5TzH/hrgY5yXlJ/sHoodvY0wP4YjkBU+19fzUyPPcjQsJfR2yGQdHFAfXuFZTbXR6zrio/q01OD51z6NUc3QOJ6wcjlDTIWIwXAEZiS6lLRGsb/wjUfu5n02xp0kRr7HMumSsulPyNoQraIr9TcuhXxA7PUgJuaaHmWKGPyvmSr4+CYfzcucVNk859mqMoWmodsjQbl7AdRGtAs1Vgn079zwBbIi4zueQLjuMee5HL2XWc0g/EsiQU/T4CHM4malPWjXJgHY2kVqADk4Y08eQv1KKmaaSKlZcxSATbXC9SJ6NtNFzeSn1TyFbGWN8RjHkSTXKz9RGIZ4pYvYpxvAEMXs01pi/lPs551wd8roGHPfcOdxai/3wfKa/C2I4S85xDswJOr8liBF1yCb5/DSmEyAdiSvuxMv3K5RZtZHWolHEkTEk2p8XfuC7ZmeRTmQj15icJCeqQ0Qp2/JnZT5/8RmtjjlHZ08fNm90apxDhluZfxEd1hbtXGtKyhYkimUTSMzZlo0anwu5x+ACC4exyjMS6ou1Da8u3Nam6tNAWorupqStbThVfVoYE8dgMBgMBoPBYDAYDAaD4TmAvcQxGAwGg8FgMBgMBoPBYHgO8NTuVIG/oBCRnks5QhCA7ga6GSm5robj8bd1OD702qIWxk1R0hJQhE/7ohM+OjouywPQRmsF/bs/ESVrACpbvSW+03wG5wXQ2bugjTchBZij7hs4TyPW3/oz0FNBj6tS+0ndSpf8vmYZwCUNjXQ0OlyQPca/gzzDR5mDKAatzQNtNV/iuuWcc54HWiEouZkD3b7I/O6BekfafYhxQ4obrxODzudD/jGHbGcCquqTM2XQZx75qkBqBd2R/5M/G0GV7zxXLyRn26DsXd8RvXKzq3nTgwyiKgnS3VGS0oRcMkeDxaGOn8Kp5PpVzZHrL2nubvYgxyv6NHNyWxjDSWE+gZyRri1w6nHIfp/SlgLj0QMtM69Ih6j1DJeWAzrEQNLQKWixwRrdqTzPc3Eh7dzZgANeAjo8pJltSEx9yAlJpWbtoEh0Gdw8ctDzKdeiLIvzOwPNuVZQm2tohxTXJ017gvERoC1bPUjHQtBgUZdkOFj6e+VeEajoWlFxRWTIWKOb0d/FZTyP6XzYhAwYNOPJCHT3kfpiOFXMcZBihR3IQmPIddqaOyE6j+Oczo5kCJ+eLK41PBctOvElFxqfi3q/dwVzzls+XgI42czQA0PsE3zElAD6xxDzj3LsrCKtEpJ1Sjcu6+MHrtNql+VLpIhPM8zFGXjSHNtRQMciSAipKw0pM8YaTPu0fDmdnNLW7LKdMa79yl4Mzj44R4UqTqkwxo2HPvTorMX+oWsVJCeUv++CHj4EzX+0Rso/Mc8y92S0GLs9xK02nIdGiIUBxvNmV+vWi9fkEHh/KDchmsrAoKUioRhh/LQOFOte6u6V5Zd3Ve4VY6aZqd2uQwrQw5g6TbRGJnDN8rAL62zIHWxnDIvKw3tlcX9Lx9Sxv82xH3OQscwgRakjxse1xRhYYYT2meBl/y97b/Jjy5Ze9+1oT5/97ZvXVk9WqYpilTrABNQCAjQ0PNFEFqChAQ88NGD/AZ5o5JE9MAzbMmyBEm1JZJEsqSgXTRapKlbz+nf7JvvM00frwYmM9Qu+c1w3650LIYlvTd5+ceNE7NjNt7+IXGuv0nmVXDGCG1iBXI2SYB95YVByPuGijIMx1hnImWh/VWBc0P2R86uge16VO1GG2Fyf8DusrVySQkrBuLUArpPwvQs/plqQ/eZDwvNiX86OPt6BIkiz14qydHmVy6TM+eGCWjacd5FzoF+SbZrJ7AAAIABJREFUQvIX6uYTyLLmE+UL6UxrV4j4FlFmhvfRBG0xq45HkNa08G7HYd7mugXHu3HG2LY85yhWjN8CcTRn3KVzE9qyhVjb7y9y/FM4dX5eeJ7vwiqnCSd0e+LeB1ojfac4y7p5lCfyvZC5PZb1IGCbQGY6g+Nwrnvh1d2VxaKedKSOELPmKddQ3BTt2m4zJtIxE3OXAcZjv+F4Tmkd3qMQ00PI5f0uhbGvBmPiGAwGg8FgMBgMBoPBYDBcAdhHHIPBYDAYDAaDwWAwGAyGK4DLWQR4Zb07PxhGDTFQQUcJUA/z5Qwjl4MeSMVVTikWmFsj7Fp+BMnLBBS6HHThs8pxim43ORxZMrgQ5KBTptj5PMKO/WVfNCsPu8l3O6B5YTd5PyF1mfIiPCxokdypvKikLqt2Kf9lcUH951V5j1UShHwuzlo2VXtTvhLGlDuAll5SlEQHHZZRH1AM62uCHsfdy0PQ1ueldqkPHGV7lN7oeDbjbv+iIb48kcwgQRsEK/ymvMZxb2lx3Qiq57i3Idri27c0PvtQ6EQxaPPolxIhwCsoYVN/ZYHKCfquSNSmG6BFNqjl56Axl4sKDeBIUFA6hLHf6ug5fJxfOkqBcByuEF6+3ImJ84+05wLjKgdtPG84nVWSvjV2aBD4rle5C1zb1c7/IdqP7goJKMZFofZpzAvKvSDbCCDzKQvKIjX+C1BVM0hHCsgDXHtR36KhvYIcDf2WpJCVUv4VU2aiQVpCHpdC2kiXOypOPAQqOvjQNZCuiOuOo6qH58KqApsDrQl0WKQMbHymtnUTPCckLxEo3DPIXOKW2rfoo327cE0BXb1A+eCl1sujg8W6SHei45Hu/+ljyZO3b0F+vEu5rB4jxvOFOWOK1gwffzOiy1+I8UN6fVOigbWzuPj39c1Fz/NcXMlyI4/yRKx/dBShuwljCanlaFs6zHi0c6O0gsLdhjkGYi7dg6rjlHtQhkWaPudKw0mobCRdKJMqjnNwnRTj9XQkOYMHKeA25K87PUmLYpy/TqR56Z5VUsA391SPmDI0Smzh3rMJ2f63v/6X6vJ7Dz+uywcnco2aQ2o3x3U6yC9uQ9r09bvv1OVbW5I5RdWcppPiTleS46/sykH16RkcpjBeSsgY4oHuuQ0Zwdfv6/hbb8kRxUGOXsaqQ29L5zx/rnwoCuBWV9XT9/XvnxeFc25WBfpZqvqcnyhu5tAPeQ15EtoETkZ+wHiD/G+K3yK/9UpI9BuKH5zfWAMX1w8hk6CrVYZrJ8iV5nzpQb7BuFwgt0kQ8+iamiFmDXY0zxJI4YtGnqz6bO7ddK8DZVm6ZL7oD8qJmZdkWJPniKlpqfcLP1ku/6Ykiu9TvJdHx1W4JtLaOEKcjNqL9Xs2xVp4/LQutyCJGWP+t/uQiA/ghIT30gSOhgXGb9Z4F+SYgoyoIaeinFjnbGwucqngOeW6nw9FkbtR5fzlNy6L+iOHTOB0VjjlQi3IAxv9T4cn5N4R15kZtsWAU9hkCjdN5MbdKo63kSvR6XIy4fouRGjviO+LmJd0uPUiuM9hLqZYC9gengd5GeZrzHedtsmpDAaDwWAwGAwGg8FgMBj+QsI+4hgMBoPBYDAYDAaDwWAwXAFcTk7lJGPxSTdylDtQDgQ0dnhWeQ650ek5aOaBpC3hRJS00Vzl0zOdE2EX9t09Udsudnzv9yRXyEBJHk9EMU9RYR872A9n3H2bbjugtuunLiT1Df/iNVtEAF+TDk3lqvM/J+q7UY7SMH6CHIGbkM9EDZuNxzguKlsYqX8iH+0DmiAY5w3ZRAGKWasLCcFFc6JdQ7iEtXuioI1PIfkClS2CCw4pmVM8x9MX2r3/GMcLULFhbNGg+5Nx22Cur9O2gfdwznnVXHzzlmjPd7c0zvtwRGkFkEGAPpp7av+Gm0MJ2aCTRK2cg044xLg9Ub/PMUfzQJTK6XxRX8ryQsSFdgzJjw/XOIwROkn5lCdmGF/FcvkNSagNgxb8TwY6dopreq9BieP5getULnybm5IVDQaiq8/nGodzSEmzTO1K9zDOkQKV9mNyYSFzhWvfFNYrE8S8TdDqWxuL2BpGlCeqZXuglR5PNW4mqPsglwuSF+r8gG4CkF8FoPvP0+UU1jDgOEJsCJc7jqwTnue5VhWP9na19vTgCEeXtNGJHGY80LZJS+4W+m2bbkk9uHZMNDbCtsZPDMo3587z5wd1+exsMQc7LdX32cHzujw7U71ubUKe8cadunztvhx86D6SY6g1HI/Q/j6o03SDajoUCkVD3vMaYmpZ1uMpgASFDk8NTz84r9F5M1jh/EQJI+WelEfRHTPFfHGQLTF2X/w2h3wngnyJTlZTrIUz1jGAwxkdqfC0hc+cTuUJKOz7J5Lf0RFn65pkQzcGGtN9X/FrncjL0p1Pq37sag0L2poTeUY5vX7LZPidW/fr8q1NyZmOjyWnOj3TPJ4MlYumQ/Xd3rakKjcgmc3hLFVEi7Zm29JZ8M7Ojbq8iefwsC1AtyX5VRQqFoSx1pJ2pLk7CUT5P0J+zTw6wppByUnsaTz2KweawF/fnJwXuft4upDbfXIsZ7AQ+eebGxhLkNskLUrM4EI2R1yBhDsbao3ykJe0+rp+Q0KdYk1lXljFCUqwU8S+Ntx2x5BlPjnSvBnsaa6EdJhiDo61dorr5Ji7XK9fwgWuD0eyBJK4mXd5CcerY9EuBeqXNXIytdE8pTsV5NxoR4RLF620MoQsG7F2DllahDUng+Tpw48WDm4/f+9RfayNd407d9W20aef6Phbb9fle1/+Ql0+PNA7xRTzLKM7VWNfC5Q9rp3L51cIme729mZ1bH1yKuec86o1ucS2GTlylVlGia3ytpydhRdMrjOtiHIq9Bv6JJvpt0PISUcjSrf0zN3BIs4OdpCfpJLvBmONrTnWsGzOOgox5MF8Scqcfosh1xivKcZ0GzntANLibl+xm++mrwpj4hgMBoPBYDAYDAaDwWAwXAHYRxyDwWAwGAwGg8FgMBgMhiuAS8upLhw+CvC+GjRpSDi4nXVOpyrQjRLQ48dT0pNAOR+rmjNQCEcjygtAnaTspqI5tkD/72Pnf1Iep4l22PcddhUHhfF4rPt0sNv0PFOZ8gKfNGq0U5O6jLaki0JFBy3WzB6/UFmQ6p6hf+he1INUw0NfTaeislGGMQDFNyi4SzwokZBq5KDChm3uzo+2qnZjL0r1tweKeYC6x6h7mqjtk5C7qWsMnQ0lp3sAucGE2hu3fDfz0i3nQfIcCIHWCs9zrlXRsHd3RJ/ut0WTjWDN46M9SX0O0S+5xx3WQZ3MNP8KyDNm56IoZiNc04PbBSmYlSOJBxeCEJKfBPM/TdVHztMzFS3SUCGnKSjDwFij1JPyQUgQGsebfFYVL6QR63Q48nyXhwvK/3mpNr4OWVkKWnGK2DefwhkB9PaAbnKIN3THSSknwDUpp+vRQSpE/1+0M/rYD9Enkdqs24ODFnRtMzhf+SXp+5CfeHByQJ+MIP+aZZSRqM2++IZotAViE12rfvhQc/3zwvOcC6rn3toSTTaClCHHWjU5kiQDqgo3R0zbgGtDHy4IyUTt4o+1RsVdSKsQgHyn89/5kijf09lCRjAZInZDitkb7OAaWtNjxPSgIUlUuWRMQawpIXPzc0iWfDrbqT4+pAsNKdNrkFMlWeae7C+kG+/cu6c60FUKE4SSjBzy0QKOGB6kfGVG+ZV+O5sr5s0gS6Zci49LqWB9T0qs0FIe8q8EbpvzFI53oG93ce0YeVyGPpxg7p5hTB+eSpZCd8KdW5Jw9LbgatRWfFknyrJ0WZWDZHAP6jL3gmNaTochruGILTHS5M2WZDZ+G3ke8iF/W7GoBdfSw4Nndbk90XXmG4vxQyPTw4MXdfn8XDkKY3EPY60Pqn4XWwH4kEsnUFnkDU07JPyQLRapYooHZx8aIAYXLkprnJNl6bm0WFx3hPyggxuPsPfBZoI4EVKEqfFJOWM2wTxDDuND2uHDyclDm5RwRfXpCnnxW0hvwjbkfJCEnI/1nvHJY42JvT3F3B5lbXgBmCBHoiNvAvnVaCRp3xxtk0EmmnT1TDM22Trhec6v5hdjdgLJegcuYDnG4Zx5IyS2HuZljlyjIXulzB77H7Th0Bkjvzk40Bg4OF7EsZ3byiE6XdQROcTuzvW6vHNdksc56vtsX3LACZxGM8QUGncVWF+b22kgv23YpWlOXNTTW6O0MfB9t7mxeL9IjxTjzmZ6rpGn2N/Fu8J8DndO5ItdPHCOdWk6VLyhEyTXyAnWxRLzNfR1Ly9cSBdTjK0Zcokig9sUxtx4hC0hMCk8X/0WtZEPdOBOOF/hGoj3SEpeY6x/2DnCOeT7rwpj4hgMBoPBYDAYDAaDwWAwXAHYRxyDwWAwGAwGg8FgMBgMhiuAS8mpPM+54ILOCUoX5UvcyZ0OVnR1oiIhSbnbOlwbfNGpSlBCp6Acn41Eg6OXRbsl2mBUUYQ97DDdg0vCGNc7OoZTCGiIEehplCuUoDbO8D2MbhV8WJqjlHCuaJxDG5zXIeFwruamlyuUI+zDuAWuF/qZu3qzPC0hg3HkCaqYghpIFwbSCjO4cxQVFZ00XkpLpnApK9H2PujRWUaHM42nJ4eitj58qZ3ks4J9ovs2nadWlFVsSOVeF3LI90pST1EpH7I4zldKH1LIOdJCbQT2q8unkLZM4FQCiuIcnZpD3tauXD7KIWRYGDsdUJ6DM1DeC9AZQckH89QlK6zCOMabkg/QxjE2KfUsIAu4cK5ZZ2+WZemyitb+wccP6uM9xKpBX5TsLGV8An0T0ocCscor0A9zyAkTuO1h7HQow4lBW55L2ppVUoQ8BX2VDjugpPbBE42w1HhzXXuWq59dl5IvHT8fia57dCppAaW47Y7utbsl9yvakPnrNW3QdX3fdbqLNWUAOVW7rXXmaF8SqiShzAZzF/Nv2NLzt1P1dYC4GCS6/gwuZh7iawmZx5s34TI1+M7iPiPF0Y0Nyak6kGX2OhqDMeS+MaRtnGczSB1KyFVCuj5BUknZT7mC2u83HB8vfrg+2niSZu7Jy4VTzO0bcvHxNiAPRyyjfHs6h8x3rnEYon4l5KkJf4t5HGDex6D+jyFhGp6ony9koG1QszcauZXaeAS3xQ7jM8bK8BzyPEirMvTPyZnm4jmkKHSKo8xjNNJ9x3BnWbeDygVKV7qsmkdzuBmFkE+0c8QZSORy5JzTnHI5rIuQFeUZ2hSuYAO4jXho6xkkTAnm+pP9haRmirk4HsLxE5K3XqQ5322xH0Ht9/ncWJchC0rgLtOQ16LNziBHms6w1UCAhKAurm9lLIrSzSrJU3cgqTjvMC4hfaC0Hv1GKUOKNS+DbGOG+dfdVNvmkKTlQ7jGQULYcKet2vMMzoMvDiTZPRxjvUZ7b3c0bg72lX/2IBXqwWWNc2sKN7sJxvE57rsBlzU3o+QD4z5/PXqqoizcuFq7+nTN4/sOAj4l/Anl9KhrwFdWxNcwjJce95D3drpa08pA52+EWrO/snXXOedcBHeyHJIszqdrdO3E2nYyVlzcP9B4oLStRG5JE6cCcq2ysRUAtg6A/SPX3bCSYK9ysvplUJalS6t1Pg4ogVaZsvYpcoKzBO9lkDOVqcZwym0NCrxfYsuNEcY8LuMmM42RFNsLpOPFNecTzNtUdRkPFcs2B3gvyumsDIfTWDFoinw5PYVzLNaIEnMx6mBbAsTNPIcrdnJYl4enl5cZGxPHYDAYDAaDwWAwGAwGg+EKwD7iGAwGg8FgMBgMBoPBYDBcAVzOncrz6l3YS1DZKZsK6IhSUmYlFHSvIP2XNkygWWXgm50M4fIB6ifp9DF28HcVLbnEv4ON5nqgUO7MtNv4iPTHue5TgNoOdlyD2kgHrRKShqa1EXa/brjggHZWvB4pTlndj9ItVq2g9gn0wbArqhfpfTmdN7B7eAZ9FB2vAlDs6djlIOGZTETbzmtHHJ2aYDfw0VDnkqoZgiqeo3/Ox+qf9x4/r8v7Y0i46NaFbvAaWilQAanm4fmvSU5VOueSavIcn4oqOBlLErEJSn5YcLd/XadAZTNQ4ulUxIF+diQK4fMj0Om35ayw+9Wv1uWoLypiVMWD2QvRByfPnuo2x0eqL26fTkVz9GKNoxBSo2IFhbTgrv6NjkEMQh9RFgCWqMsqynG5TmljWTpXUYifPpPc5uxYdM8vvvtGXb5186Z+Sro/5l8KOVWEwZrTwQNjIcM8KlDmtKRTTVBJa/IVzdCFs9YI7kl0ZsgR1+Ke6jvHvMywS//puSjJB6dqG8pSOjmcQrDuFGiDdbv81fcLAtevaP+dbr8+HsIxazqh25bqN0NcCjAOzzDm4wQym1zX79Ct7FwxcOZrXs55fXRaXAWsXczP9qauHaDPKf1tSC5zNihkmeiXErKEiDJnSAqCGe4FuZiHdYhrVVD/+/qQ5rl7cbwYZ6djtdn9a6qbB8ewAg4UM7hTZRqqLp6zz+E4hxjiQxLT2VRfhIjdGZ50BHnc+dniZuWxYsc5xkHDsRONtbUhWQHXyyFkAMORHqTAOD4+1fEZ5OStWLGYzjJ0sRwN4XDkvR45VeGcG1X6358+flgfn2zLPcaD5Lu/p7ZIIdtPGw6jkN4iLhXIdTb6knD2eppHPmQeLV/XHGWIh5UVJvPoNuSMBRLWENK2AHOCfT1CP3ZD5bcZ+rqgnBhrZJqI5p9ndH/V2uxBrlW7PK5xXSyK3I2rrRIyxMrBABJers3QTWXIz7mrgQfrrxxSnZAuqJCKOkh7ogBrC1zARnC5OjxdzMGjF5JEOcTeyNf9O309xwBS1dlE93zxTO5kg0043vl0xNUDniPW+D2NuV1sh9BBTEkg0WQusU7kReFGlVxt0NU47ECWw3xqStkM5h9jv9d4d4AjFfKehrMY5IwJ5g5jcIo16iJ1SDDnE5wbYeC9OFfc7cApifLXk1PlumlK5yGsK6iXBxkdl1euhY2tAJAzBJUEaJ3vjUWZuWlROVEFio/dNqRvM8RKvAjRsWnGfkj0DjHo6X2lgMyVztOUZ3uQxQ4Qc9u7WjuDqm3LXPfv492135F8zoN7Z5IobqaQhUW+5l+7A/etdLnkttVV34a+2qwoNBYCD1LVCZwg9Sr3yjAmjsFgMBgMBoPBYDAYDAbDFYB9xDEYDAaDwWAwGAwGg8FguAK4lJwqCALX31hQmOZwHfDoCESabEO2wV23IZXCOaS4kWzLXfUnkHGFHdESd2/fqst37spd4tbthQRhe0+0rTAQ9a0ExfzmS9HjPvrww7r84umzukyZQYYHnIFSRpp/gR23C1DI8wblbTmNs6axvzaDI9AUcTQHXW9GJylI0jxQT+m8UaB/8gA0TVCyPcjdSsqW0A6tQPS3C+nU+Zn6p0xJ4QM1EtTQ0QTUYDzr8yPRNt97JDnVmM5jdEPJl7cTUTbK/L/XtfO/c9OKb3lwCscQUHyv74iml0OSyP5tOryAqoq+uKDtO+fcGPeKS/VjvydqY5/uJ9hBPq+o9f5EsaON8TKm5AsU2gmoilNIdHqgSNK9hpq3As9EOQ0pp6SkFiso53X7rZE27nnOxZXOYYouoZzj5QlkheXycZUm0xXnwMUAR1O41szGpFWrnWPQxl3D2WwxpmKMpwQU9jkkFkkGt5WpqKoJ5Aw7XcQC9EkKevgQ9ORjOIu0QT32QH8Gs7ox+17PTFy4vfV7C7p4BFp1npJyDAkCfkv5XgSHi5zyJI5JzNcZ2iiZw31hwj7AeokGaFVzt9MRDZlsb7oHtaDF8VpYxzfoioY6Yhw5XzR6H2sGnZgCSPDclM55cMijvrNcp5BqgSzP3UEl23t5Isne7I5yCecr32gqoDUX2CcTOGYkCWQbkDX0Y1G7S8Rcz0NsJW28rf4a9Bf1PdqX9GIMd7gWpNA7O7pPCJkzZRWMfXM6xWBeJsj1xhjTOdbICBKiKRxCErhCpli/14nCOTepHu8Hn3xUH//B+c/q8hvXd+vyf/I3vlWX25S4Y+xxjUw9tWm8qZzSh5ysgBNPFKkf23hkH2tnJ1iMqznc+bi1QIIcsuHgA7kO8xWvRH4LTTDXBo6BAFKfzGmt90rNg8BTfAkhC/Nqzcc610XfhZWccwBpWob1pGgj9kA24XxqqNAmmAsB5h8e3QUNB13kDYhVDrGqnCBGVjnKm196uz5GF9yXkJCfJYgXuM8581jIb2dYufZ2FIMo86F1awvvRXTzHZ6rD+l41Qkvt6vGq6IoSzeutqLIs+VS/bzhmAbpe8F3SsgGMf5hGuVcqPNjyKNO4Ww5hENfgHkZhZCuVs6RrRT/jrHjZ5Rf6vZDSGumiJFDOPg1nCghx6VrZo7jdJnykMsxL83Qfn419stVFo+/BMJW6Xbequo0Vfs5yKmmDzXGolztQzcrriHMLTLklgnWkNDXvO+jr+iyW0A23wm1Lm4MFmtdiO05skJyULpa5QVyjwz1gpZtNkM8R1oc4PlabcQXjMVWV3Vo9/DNJMTWAZkuGmTaluJVYUwcg8FgMBgMBoPBYDAYDIYrAPuIYzAYDAaDwWAwGAwGg8FwBXApHp3vB64/WFB7A1L1QSH2KJuiVCP47A7gzjUpnjloyTlkPNx9PgXNrtPeqMvb1+Q+0NsQzXU0XdQtP5EUpz8QtTgKRKeK2qI1HR8d1OXT4/26PNjSb1M8yBAOFXPSAintoIQDNPoSz+Ry0HirNihfn55qKSh3O4Dz0yHK926o7Us8bwrJmJdBVpMu70NSGUPSydFWWUUJDSDVyKE/GZ2JJno8Em11Dpqt3xYl/ScfPanLL89EcUvBa/QdnbWow1GRpH464jQsrNbpZgSUzrmkmmvDmdp8hLlIwjrle6x3wF39G7RVUj/V1lvbojnu9tW/2VztmD+QFJEyLu/imnRGwJjyQJVNES88UqcDUstBW2R8YXeh7yit4jM1JFToYDonNObomlCWpcsqaZMP55sQ1PxGjIE0we0sd+/z4Jjig+6ZQdpzcqRYSBmaH1J+peN0NZlX9U1BDU4gd0s5V0DZHo00R4uhylvXFE9jOI9NxrrnKSR0p5B/XdsQXb0psEG/Nf7l9dhTeZ5zcdV2NDhJJnA7gISQbk8BZRCoXtiCbAMxkk4WJVxT9l+ILvz8g0d12Z+QWs05teibwe5efYwOIi2MBbpJxBu65/V3bqu+WxizGnbOZ3xB44RwmWhD0uWfSwJbeHB5pPT3NcipirJ003Qxzo5O1JZnQ8lLbm9rzWO/UapLinUJuj8dH89AsR9B2jjBGtnfhIQK8ogW2m17sKBetzAmpnCBiyPUEU4hT59LfkWXzI1NPd/WQH2yf6YYPc9V91PIbJMETljR8nV/grnb7oCXvkYsJKqL5x7i+AvI/xNIHn8tVd/1YtL2l8ty4i4c55CPvDhEm0IW0I3hqITUrqHWqqRTM9Rlgjxqhvx3kiJfQX5DOv/2rij5AdznPDi+tDEvA1TGK9Vqoae5GMORKkR7RNX2Cd4aY6vney6scvENjMMccmuX0oFPh2F26KBIdF6LMleVD57LTcrDOOca3I0VFweQ0HUD9W1/c1Hf6VhzIuuqvYcon6LuwzOVA0hfe3vIs3blmutjW4IjONEN4YLb3YMUHmttAsfDCNtedCB/XCcW+c1i7M6Ro7KcBZRlr3CkwprKdJrnj+EsND3VHPn0kbbCCPDMOz2Nqw5y/VaVb212FAv7gfqi5anPX2C7hucnel88x1wYQnKaNXJwPCvWZa4TWQI5ceMdke8amBPV8xVrdKfyw9J1t6sctQvZIpKVp08hA0Se2e1DJtTSHGr31IZRV/nfIWTML/c1j0KM7Y2++qUFyf/8EP1f5TyMU0kOpz08X16o78/gnsic+gZi6wDx1C8UeNpYa8epvhd0Y/RzC+/8kOUGEfJBrE2vCmPiGAwGg8FgMBgMBoPBYDBcAdhHHIPBYDAYDAaDwWAwGAyGK4BLyanCMHLX9haypRFovTNQVXM4EDjSH2laxU9HPN6g0HFnblKpVT7aF20phxTjIWRRp6cLipYP2lkL9O3ZDI4sqFcxE/0qpPMNnHewybwb0nmBdO+G+w9dgZY7BJE6ePHYr0cE0LxwQwFEOiAofQ9fqL3fuSFa6d6OqN8Z6OGk2M7RPmenoniSDRhyYIAqmVW05QAuONOxrnc6BpXOF8UtBq34/SeiyP/ok6c6v4Q8Af1QNtoDkj+U6cTmoT8plWs07JrhVTTrkBR+0OObzjzc+R2yCjIzcT4MNlwfUglPrEg3P9McoctJ03EOlNALImNPNQt7ulGv0HPMQS0PVkhiGsFrxXM3PlO/gsFN4zDNH17HLCzLWuYUQBpDycz5CJT9I9F3v3BHUpYWOov9HwaQx6WMxXqWGPTQHuZLZ1vlErRVv9qeP88Yk0EBhoNBCgnVcA4pWAbJF+jhlLXNINd6ti+q8jEkHNuQazXc/jAwGrOveF1yKq+WroShBtz8XPTcZAaqbrJcJtRwQ+Oah/EQwsGKkpvnzyQLOJupD/wEcibQ5gfVdXqQB2zsitrsIOEYPtK4K8/VdwGGVEzbhhiOOJSoYu0MIV1pg7regvxkBglQWSx37VgfylrOPUVfnZxKeni+rfbe3ITrVqR2jVoak5twkgoiHZ9AhnN0qvE8PFa/TUeKmwGuH6P/40q2VMCFI0107QmklSeHkhlnOH73/t26vIH5nzM3mGq8HmH+nUMeFYIG7iO+kF5/BtlzGL0eR5zAObdRLWoTjDcf4+oUOd8QLkC3OuqjyIMUFWtIGzJDnhPAYaRhpAbpgxfq+iGkoGXlVBUloNVDnu/PJJsp4OwXg4bf39TzbW5r3NFNrDH/GhZ+Gj++0/VDrN2UujPviar2WOecLErnksp1hy5aCK0uS/QsY3RQ19PztrmVQUn5CiTW6KseJH7nRxqr01K6ppA+AAAgAElEQVT55WD3muowUJvPqzx5jH4bYd5O4HC0P1oucbq3ofhy/4YkVHt7mpefPNc8PkUcmWLCts4Ur4OuYmsHsWkCmc8EUvh1wne+a3mL9X0CKfjxUO1ZYLIkSMriFtyhumo7Dwkr14dDSMueHej5h4ijg5baYvpE62UbueY7Nyo3wh1IUSGtGU3Ud6O55D9jHD+Hjm+K96cUOUrAXATSzRLbaXiINXnBd9BVWwQsftt4//ic8JzvosqVsfQU+328LHtYkzCs3Ca2CAjgCN0OIaVuQSrc17NH2C5gsCk3wdu35BZ5B47UjfeCat5l0K/2MIYSuFONIWd8/OjjunwMeezudUnOWy09axpACon3pQSOcEEb7qsht2TQdTy8v/lBx10WxsQxGAwGg8FgMBgMBoPBYLgCsI84BoPBYDAYDAaDwWAwGAxXAJfitAZB4DYrd6oeHBNOQklVZhPRA9O56LZFDgo5dxunhKlh6oPd80HJvn9LMoLbN0WnuntPNKt+T1TV8WhBs5tORLeLY1H43//gg7r885/9tC5vgmYegPKVgrY4gSvLGO4MKdhspCWXq6yNVuzE/rqcjVSf5dWhnAob/7sD0KFfQuJwA3KqTh87j4MO2IU7AKUaKajaPiUaIZwaKreHDJVsQ37Sj3VupyVJwEPsNv5HH35Slw9BpyswBSjV8xx3EofbTcPVQfVpuLK55XTHdcJ3petVO/tvd/UMvZYGUOBT1oUfUzbWKEMq1gK1lTumUysF15wSMquo4fAE2uDFBA+Xyy3SueiGU9Dc6biVgR7O8ZuvlDAux6qp5Tdo47ikWx9FlVe9mPAeAiEliedwZnr/4wd1+evvvl2Xey265nDcKia1scP+5obo9sOp+K/FOWQ4oGG323CQ8xdU2BYGfw5HnAxubyeP5AKXnYiKuzHQYCkQQ+dw5zhErPn5hx/V5ckUzhYZZFygEBcl5x/o+K+jC51zrixrCnOAmDcCbTyHrGSeUmIIVzWMySEouf1UY2AXtN023LzuXIMU5q6o+MlE94ozuHNUkuNtuGR0AowdaBc27mmdDUEP7oNmPOtDcgmplIdnpWQwYB9BUtfraS2ZJ3Cqorz6Yq6sMbR6znNRuKhfCNnJBFLxfci3o/bNutyBew3rGcJVr9PWutjr6xlbLc3FM8hTY0gYW5B5ZLS8qqQIOSRzrQ5Sujb6BMkVHcN24epJF8AzSDKeQ4r16RPMacynCPVNIXvOIR0cDiH1xDxZJzzfuVY1nNqQmbZRP0pLz8eo66bOibFGcc2n9CHs6Pobofo0hPTBg1zbwWWlKCnXWPw3aWOMJ3Bk5DrrQ8YXqL4bG3LQibCmDsfIe0NIuCCpy7gFAurlITfyIRHxEBsunFXWKacqi7KWPMwhD2whriSoT+Grfdpw0vMbMjE4WaI9d69LHrWzqRh6c1sx9PxIbXh+onnRgUQqrCRmrS2tm4NjrWHbc7XZuz3dJ9tQvbbg2rOxpbhA2eKDF8q7x7nmE2XJGXKVbTjqnQ7Vludzbnvwev6W73u+60eLZ5pCTv30QO+LLaxhcZsSVfQvAj0dSSfIOyYY531IXjbbeo8rZ7rODtx1d1uaO/lk0UZPjiWnacGpzENcyGLFtl28cz5/onjptHQ7D05IRUPbz60AILlBTuOtcF9d9n5RrvW9MXDOq8aiL/mYhzzTg0x7MlSbTLEOeHCBThONhU6u8t6uxnwJ+dXhqfr2dKg69CaaLzduKUfZvb6Yu4WvMV4ghxo90Bw+PtTWGuMTrW3XBoqte3RShF6sxFo8w3yiy3aIsdVw2PTUHj5ktn50eddGY+IYDAaDwWAwGAwGg8FgMFwB2Eccg8FgMBgMBoPBYDAYDIYrgEvJqcqidPOK5jgAJT7bEK2ILgXTMSijc1HLPQdKMGhllEQkkCeRVkb63TYobNdQ3oGLRFjt7F6CTlzAHeV6R1Sw6aHo0omv50hDOGZAHzWFFKFBkffoFqPHoOPWKmkVpTiv0djoM2h6aIHqDn0XnzGBDGICZ6Ix3GnCFim4kGR0QEMF9XEG6ci0gCyvvWjPDLS9BG4lMVxqhona/ucPP6zLz89FbU3Rt36D+gypTqNPBFLRS0qLcpTJFA/WyPkHwsBze/3FjW5ug6oPORXdHBpjrCGtwkkBZWNqixwSJlI//Q6kZaA/5qSBckP2Si6VQG9YIhZk6NMiAiUR988wj5OGa51u4zekbQTpjJhnpKqioXzQlV+PH07psmrc0GEqgaNBgjn3s/cl/Xz89a/W5Z2u5lanpevMUzwL2i0G/XyAZ59BCjWZiVqatERhbbUWcZZym2wCV50jxdD0WNTvDqRSIebZGE5Ao0Tljx4+qMtPnj13Ah3klruW5Qi6jTj7muJp6eCshXE1gmteijHMcgl3jhyOFcVclR0hps5QPjyBOyPmy+CW1r9ipDUqhIvLhfxrWKhvR4eiLYdw6umA5t+CdLbAWh/CccKDdCUpVd8ArngB/n6UImCSuk43HQbVC5mHt8ZZuXAYW8wLurhEWLfOh2iffdV/x5esLIBTB51XyoZUV+jCNWkeUYaq67dZHzq1+IvzPY4tOLYMz+WsFULKFmPNdZDezJBzPX4mOcHP3n+/Lj97KVeXIFLdiTzn2qFrJpDLDtGWa4XvubK/aLtgjvUJWsoMcX0CmQedtDpol0aH4dnoGufQviViKrcC4G/TieSl84qin8xRLzhVzaBb8yEpotucFy13HsKjsqvdZKh5yfUmKPEcTmOtoBQEa4ni1/qCa1mWLqvqNIXTXpZj7GOLgwlkGxPEzS6cfgJIK0o460TI/7NIHR3h+pvIL2fIZyL0V3K4kGgEcBLbOFe9Crj5lluSWBzN4PiDvHiM96KHB1pHX8IdroTkqA9nKw9x9pROVe0+yqpPnnPThPUh8AO3WcmyJ1M9w+lI62ILsspt5CUNmyPkoii6DLElhj3QjT1J5DZ7cjY6eKI6zOEK+GhfsW42Xoy3AuNiBMllC/n15o7OuXZHcp4Z3iNjyBa5zqUe8xU6MQvMbzzKxelo6X02H1qnnKr0PFdc5PZ4VQghK+tgzRtmes8aI8bx/c9hjgYTzMVIcqpdtL8/0bwYPnxYlz95+qwuH+6onzer/g8RH/NMdTnZl4QqhavYRlv1uo4tByKnGDSeaT4lcHUbItfrbmluMS+iw6aP4z7fR4PLy4yNiWMwGAwGg8FgMBgMBoPBcAVgH3EMBoPBYDAYDAaDwWAwGK4AvMtQrzzPO3DOPfyFJxrWjTfKsrz2i0/7xbA+/I8K68erD+vDvxiwfrz6sD78iwHrx6sP68O/GLB+vPqwPvyLgVfqx0t9xDEYDAaDwWAwGAwGg8FgMPzHgcmpDAaDwWAwGAwGg8FgMBiuAOwjjsFgMBgMBoPBYDAYDAbDFYB9xDEYDAaDwWAwGAwGg8FguAKwjzgGg8FgMBgMBoPBYDAYDFcA9hHHYDAYDAaDwWAwGAwGg+EKwD7iGAwGg8FgMBgMBoPBYDBcAdhHHIPBYDAYDAaDwWAwGAyGKwD7iGMwGAwGg8FgMBgMBoPBcAVgH3EMBoPBYDAYDAaDwWAwGK4A7COOwWAwGAwGg8FgMBgMBsMVgH3EMRgMBoPBYDAYDAaDwWC4ArCPOAaDwWAwGAwGg8FgMBgMVwD2EcdgMBgMBoPBYDAYDAaD4QrAPuIYDAaDwWAwGAwGg8FgMFwB2Eccg8FgMBgMBoPBYDAYDIYrAPuIYzAYDAaDwWAwGAwGg8FwBWAfcQwGg8FgMBgMBoPBYDAYrgDsI47BYDAYDAaDwWAwGAwGwxWAfcQxGAwGg8FgMBgMBoPBYLgCsI84BoPBYDAYDAaDwWAwGAxXAPYRx2AwGAwGg8FgMBgMBoPhCsA+4hgMBoPBYDAYDAaDwWAwXAGElznZ87zS8y6++5SrzsH/6ZwgCOpyr9upy/yKlBc6vyiKupxl+dLjRalyWeq3rJnv+585WK6oO6/B5/CWn+4cHrX53EKzvqzj8vN51K/aOi9yVxTF8h9cEns7m+Ubd25U183q41mmcp6jvXPVv25L55znr6qOjheFrpOnaV1Ok2T5OcXye/nV2PF8Dlfdp91u1eW4FeM59ExpovtzMHi4DtsgjKK6HKHsh/HycqC6se4Off7DP/nxYVmW19wasDnolTeu7VT/t7wvmnMCz4x2LtH+K8cz5wJOan4BxjVxL9YhmS/6IE3V/7xPHKk9GS+8wMdxf+k5zlv1PRp1XzFkVz1rA9VjPHlx6I5Ph2uZi3EUlu1qvHLMMNxwTJaIjx7Oj0KNz16vV5cHg35dbrXVtl6jf9RvjHMZ5uvwfFiXp5OJc645bxuxEnGhMYeXzOdFBVRsxnMd57iMMEZardbScwif9UE93/vo0drmYhj4ZSuqYhTusWo9ay4ny8dno00b82/ZL/8ccH4jBjQaddWituTqK/piVR1d4546zLDIccJxzTUyjNW/Ozt7Ol7F+/0XL9z56dla5uLW9nZ5685t55xzMWJ5gHKKdTFDOcCDhXgurnNnZ6d1eTjUfCKa8YzjCGskcqGL9Y3zNoxYd9UrxXxuDAN0aImW9BFPvUYZv1wxLpjbcN5vbGwsLf/0pz9d21zc2uiVt64v1kXGKD5/hDbiuhFFGm+sN69TIGdiQ67KnzxPfRq32nWZYyaZzRb3RP836oL4Xq6K3Y1Z8EpR4nJY0dcXePz0hTs+OV3LzXzPKy/mUdkMOEvrUCK2NrH8XQThxgU4hXM3CtUXYbC8zHlxEQ8SjLPGewsnHYqN9RLX5vhjrGEMWrlIAo0zVoyRP7dGrG0u7uzslHfv3vlM7Rqr0Ip44rnl/8B1dD4d1WXG1PlsWpezVPPSa+Tlq+7lfaa+BPMJrnlcw1atv8Sqdbkxc1estXGseNDC+85F1Q5OR244nq1lLu7u7ZX333jTOedckWtsn54e1+XRUGtbifEZor3boWJu1MgLdS8+o/9KOSXmC9fOJWCsbLzf8v288a6u3+YYpHgddvMMc73xnqxn3djcrMsDrHk8x6341vDDP/nhK83FS37E8V2rXoj4wYPnLJ+uO1t6ofjON79Rl1uIouOpJt9wOKvLBwcndXmOxGiajHU80/EcDdFud51zzvn4BpLgXA8tkOBFPwrw4o7feo0PUxhQCPocMBMElDmCelbwpRMvpihfvOCdIgH8vHjjzg337//5P3XOOXc+UrsenRzU5eHpWV0enauNL9rSOedaXSUkJR8FC9t8pCB7/PxJXT54+kjXR/AdjnV+3NXLaH9rp7rnTn3MYdJ+4UtfqMv337hbl8/OFGhePNP9PcxQJlNHR0d1+dqNGyrfvqe67Knc2bmDOur8Fl4+HAJf0Lrz0K0JN67tuH/63/6Xi/9BMONikpf4SFdozEfzeV1Ox+c6X+uBayHo8mNViAnTbnwZxcLp9MzJXHV4+unTxX+fPtX1sCDdv3WrLm9tK/hFGxp3g4EC4eaGzvHi5R/UuECvSpjCcPkHuBKxrKja9R/8o//GrQvtVuz+8te/7JxzrtvBfEK77h8c1uU52rLdUpvcvKmx951vf7su/8Zv/PW6/O679+tyECApnGl+Rxg7+0+f1+Xvf/f36/Kf/Yc/dc45N57odz5iOONCBx+UOni+QX9Ql5nATKeK+QkWywLj+/Ydze+33n1bzxQyudKPmfB026rDX/v7/2Rtc7EVBe6r97adc85FmPvDqebcaKxnS/FsHG9hiHUA60kLzxBglQ/LFckNxvYcH0xTxKKLNsq5WDYCOT+0LU+uWrHmDevu+PKc6WE7LZ3f6uiZLl5inXNuggRr7/abdfk/+4f/eV3e/sIi3v9X//ifuHXh1p3b7n/83/5n55xz93Y1nzb6W3V5Hx8zD88UN7c6movbHfX/s6caYv/6t36zLn/v935vaR26G5oXDh8axlPNtdNjrWnD48X67ZVqsxs3rqvuG5p/z19oPqfo88Ix3umecUt/aGuhHKH/A+Q5HBc5Ply0OqrD3/67f68u/62/9Tfr8te++pW1zcVb13fc//Df/RfOOecmyC3295/V5Rs3lBf7kWLCzVvv1OUO8o/JVH09m6r9S8yt/Zcv6/IpcqaorfXq7beVp2yiTR988L5zzrn+xnZ97Podxbb+jsZjijbPCq25vs95jA8NTvNs1R8aX+mPQCv+wHmBv/ef/uMV1748Qt9z16t1ZM44xDiIcDPDWuQw9gKUPcSkCcp95ATXepq7N3Y1F6/vql+2e+rPOFIfHlWx4eFL5TaHfG+Z848lumeIGNrb1D3biAVHyP+fn2j8FYitPtuJf5jEH9f4EcPD+TmuM0tma5uLd+/ecb/5L/5P55xzMV+KsVakqHaE71N8DyoxtpO53hHe/9Ef1OXv/9536/KH7/28Lh/u672Gcaws+c6lcRUGizyS48tzqli7vfwPjSkTFoy75h9v8Bx4j80S5eNIY1wUqRJca+/fVZ787jvK66Jgcf3/+r//l25duP/Gm+53/+D/dc45N52oLf/FP/+f6vL3vqu1LRvr3fE68q0vX9O6dAMkjk6L+cTynKfTUZszLre7miPdgd4FLt6H+K6e4D2cH50mY8WOMfLPcaLfnmOQnuH4R4fKzU/OJnU57qguf+fv/4O6/Bt/8+/oOfo6J/DwrQGxO4y8V5qLl/qIs/gos+qr92fBr2wc/GWJvwrhcgEmS5ZqYM/5dRuT2ws1GAJ+zSp0flzNCiYZfoSXXhznX1NYsRhMjBZmd4iXlxgvkRMuKhEXVCFFEGFAYfMW1V/aVjGHfhkUZeEm1cev83N8rEFSysS7PUBiiXr4Tm0VYFEsG18wFajwTuK6eLljv7APz89Ut/0qWd2+pom4jcTmB9/9d3X5+btKxL71nW/V5Zs39fL38pkWWrZ3Hwv00aEWSwaLrb3bqsNAybvv48VlgqQPH6bWj+qvVY0/t+AviPyjDb4U8y8VeaLAFfAvlI5fzvFRBAlIhpfCiKsP3lLzmebxRc3IbIoaf+nWmPL514au2jla8bEmWMHiWcWs8Vd83Q9WfNEvq/NXJ8KXR5bl9YfDI34IR90CfEzexFf9t99Skv+Xf+2bdfk73/41nfO2xryHuTg5V3Lp4y/Ljx5rXvzx//OHdfmj996ry1GVUN2+rsR2c0v16uOvDT0wgdhvZDfwWef4uHg+1hh98lwvSI8efFKXQ8TWN99+Q9dEH00nGt/zqa6/Tvi+77rVi/wcf/kji7H5Vyb9li9Hjb8QlfwLEe6FgMV4GfFDZOMlQYlU7Gu9zPPFeLhgxznnXIq6u2LFXzDJtGSsxx8lOEUirJEh/+jBtQQNUiB2HODF+IP336/Lv/7WW8651X/l/GXge77rVR8cQrAgeAd+bArRPxGe9wAvcb/z2/93Xf7DP9J84gAIEK/5MpoWHEeau61U8yJuL67TbWk9vX9D83JnR+W9jsbHEDH5ZKh7Hp2qPDnDR8dY9+wgFnNgNj7ugXGSI4/70Z/+sep5T38AWS9K51fryOmRkuw2+jSdqz3bgerKMZ/MlJSfHO3XZR+56xgvA+cnylf8QLHu7m3lCztb+CD4RB+V/GrudjbxgaCP2NkFoxJr5xwvtMlMH6warIAGA+VVcvflLIIG62DZtFvfVHSlcy6t4lzJv7aT1YfYyhzCw3rJF2jG1j4u9KXbYvj96tt6IX73vvLLPX5cwb0c/rg7ThZj58Wp1tz9Q730Pnuh/n65rz8WDseaHwny8RLvMG3E0C18iBjn+C3aI2vQH1UsmwlhXWzG9/WhdGJS5WhzZlAh2pAfaxrPMFcsevr+j+vyj3+gvP/TDz+qy48eab5y7dwcgKXcAnsZuWOnvWhfspvZVlyryNafYR2dIL6MEWub+S3GbIwP4HjX5TXnc7XHpw/1fDna6Y37iw8l+Upm2uVRloUr8kX7//CPflAf/97v/k5dTkaKPbcQw97Ex88+8pMQC2bUxnsG2iSGuqKFP4wEyPkK5LTTsXLaixZhHpJn6gfPaTy1YrVxg3GJKZHij4JztO3dLY2nEAHwGfLr3/++Pi5u448Hv/5rf1W/jS75GebPwfbEMRgMBoPBYDAYDAaDwWC4ArCPOAaDwWAwGAwGg8FgMBgMVwCX5vFc6MzKX7TTmWtq89ugr2Wp6EyRL3rS9S3RryYjUXgpt08S0ABBz+4EoFxBk7hV6cJJrZxBj8h9avy2qIpepueLcf8eZCN9aPX6oBkP8Uzn2OeHlKvcV31nuFeW4vnK5n/XgSxP3fHpguZ5fCy6Z5aI6kfpwxwSs8kY2txDUd3TkWig3OwtpPTFV/+3oZPvgh66geMD0BnPThb0tMnLx/Wx2YHq3hns1uWPfiS6ZYI9k779N75Tl2/clD7z5XPtFRBjPwtuvnt6pHu125/qmfANNGyJcptAH5tn3FB5vbgYFmRPZo2NL5fLE9Op6I9BQYkA9z6APIKbNnJ8xpz3oEhCzlFQflG1aQuUZNLcuUFhAdpqB5KQNvafcNxwOqKumZsPul8IntPYW2QZhXx9aipXutIlFYXfb0hj1CY97B/zta98pS7/Fex986tf014Lb9y7WZeLBHteHbzQjRF/jylZ+Ym05C8+fVCXN3uax7feXuwJdeemqKED7E0UYqM9D/2QrdjsNsXYopyq31OfU07y6KmoxJ98+IHui0Xi3S++W5cpcz05UXusFaU2rpxjXzVuQsuBQ0p22ZBNQT/PfUvQRvyrC/c7SBh3QV3mpu/sm4vZ0tgThc/U2EMU0qoVsrCAkihcaAN7JA16mrtZSekW5LgTlZNMkpbnT7SnWVrJQdcqp/J916v2fGvQ5JkrQIYa4dYzyCD+zb/6v+ry73z3X9dlyvpamBcJ5Kwx9kBgu+1si6K+fV9r12ZFM9+APHlzS+dyH4XpbeVWU8TwA+zf8uCZ5taHn2qtPYUkuIDM1YMMYQYZQA4pUgdShMN9xaDv/f733OtAnmXuvMprZqD5byBGzdEXA+xDE0DOcXYCyUKi5xlhD4UTyLViyHhu39W+ebuQUI0gEZ9DmtfrLWJ8gLnqYdP2HLlTEGFfwoA5NU0qNE4pEeMmspeeOeVyueTnuOLqWznnsmXvGZCjeNDn97FPRoncYjiSrJ37mrx5XX3y137li3X5G5BT3dzG3hseNmtdsYGtHy/68Ct3JJ9zoa49wia8Lw4kp/r0sebEx4+Uiz56oZzzHPup7WILid6GYsTBWGM0T7iJOSS9bnndGzq114QcC1eEeoTIXRPUI4dU//CB1vk/+d6/qcuPP5TE9vil2ivCHiMb2Fvo7rbm+u0dSRR7WBcvzDXCxp5wOf6dsiDdZwzp0zniyykksrOUhi66PNeAMTdlxhJ5do5NhYfYBgOx9riS8Uywt8vnRTKfuUdV+/8f//v/Uh8/OZA88O62YtK9Hc2tLUifIshQY0rJAsjsQm6Vghw+XC6/b2yKjHzCq/KiJGFMhJQeMqyywHYPnurId3XuPch9QDchs0q7qu95ovOfPpP8/7d/+7fq8t3bkl2+eU9xwnmXl8IZE8dgMBgMBoPBYDAYDAaD4QrAPuIYDAaDwWAwGAwGg8FgMFwBXH5b5HKZS8tyjYEHujM2WHchqI3f/MbX6vJXv/TVuvzvfyAng/FQ5XAoKlSwgjYeQ1q1UTkI5aBBwojJTWgB3dXvaN8XwJ65h03Lt+H4MIBVWg+WdR3Y+XZgoTaEyianXKXhILBG7UZ9r9ydXsiTQMHsoC3TsWjIz56KVv38mcoeJBl9SOU6fVHrupCndXuiM8egBNOdo4DVXh8OF52KWnw+FH15fCZ66uxEVPvNbUmrnn8sGuaP0D/f+rbcfPZuiP538FJysQ7GwhhysfMjUV7pY7S1I9u/sCWqJqV960RZOpdeOBKA7sed6VPaGMKmNp2r3wPoAgLQRkvIwKagkFOi1d1UP5aUUMFaeXaqsdSunCaCntqHI5xflP0pLdE1P+JNjdOyDWp/uNxhqimJWm6ZynNYH9rPX5y+zhnp+Z5rV7KJAvGpA8nYO+/Ibe2v/hVJAr/yxbfq8v27kjZ5hebF4fOP6/LwRBTu6anG+bNP5WJ48lzU7huQcNy7L3nA3t7CzaPboksYHQP0fIxkdBOhDKiktIjOA7jQ3pbGC2WfT0ChfvLogeoOqdctuMNwvVgn8iKvZRYJndkabiAq0vmCckMPsSJf4ZRISnywwlWNrl0Rjdqw1l24Q0H560K60YD+HELPG6PuIaTFfkPCqOO3tuHsB+eKkyGkDminM46lXNd59vhRXT56uYjBGejpnxe+81yrcuGjHHMMlzTmBA5ytz/8Y7l2/PZ35dpxDDvgmNoz/PYu2ueNO3LEuXVdY/jarta0zZ7W1H61LjJu0oWjUe5r7GeI1duQSm60IFHBVPngiebZ6Qzyatgmz7HWtBGLG3XAnP4QUsh1Ik3m7smjBYWd9tplrhgSY20fwIZ9NJTccgb5ywS2uUdHip0bkLre3JPMLYYU6nhfctU51rEQvRZdjDcML8op2cNFoT6KIW0LkVOlkGo5yAyKS6pmGo5Ur7B9wtpQlq6o8owC7kWIiM73NEBbHeWKJeuGNWcD7wdfg2zqq+9oHb2xoTEygLSxgzjXiJF0matyXcZwOmVto463t+SI9eX7clU8+hXFxAdPJOf7yXsf1uUPHiv/HMM1NOpIKnRwrrE7hSMjJUrFL3IbWwM855xXy0Nok7VC14eKnB0oL/mT7/+ruvz4Y+U0+4fYFgDXfHMPucueYuc9OGruYN63MDaiSt7folQOlWxh/WPVZ1gbRnCnOhmr34dT5d0znDOdw9kPL8owKHPTGVwkZ7rzeKrj+weLHITuwp8XJycn7p/9s//VOefchx/8rD6+3cHcamkcbkEG1YL8MYRuriGhQm7TyMMxth22Hgmw5QXlVIH77DXzAsDDaVQAACAASURBVHJTtEnZeMdmToStFOhIRhdGxJQQz9dB8L6OrQDoVvrgY7m8/uAPv1+Xb1xXjtpu0w361WBMHIPBYDAYDAaDwWAwGAyGKwD7iGMwGAwGg8FgMBgMBoPBcAXwOfjlq+RUpMqJYhSBeri3KcrQN776pbr8zhckHei0QCGdiJb88Qcf6fLgV9KVgbTdC6rcFLt+x218u2qLVgcmvEsz0slBF+uqvN0XbWoT0qGiFLX1HLS5o7GojU9PVZ8EVC/uWp7m69dwFHleS5ECXPgEDkyfvC/a3MunkiqFcAPb2tmpyx1IpbbhxrC1KdroZl+0xiCA8wwemE5OEeQa+cV9Qdvr9UU9HYO2PhpJctWFa9X+A9HxH0Aq8hZkKXNIeI7hPNFuY+fxMx1vgdoXBarv1i4of+Fr/E5aUUgbLlSUYYDul8813kq4/XigDYfoi/m5qKqnh6KZ9/twgkC7nBzr/OmR+iMALbxb7fwfQRKQguI7xS79s+eiE59OVF8fO9t3BhpHVCs0JFoNSRQIsKRFk8a5yqnKLZORfl54zqvu0YIk8e69O3X5m9/6Rl1+6y3Jmt56U/K9yFdcefZYu+HvP3tQl6fnkgGc7WsMn6OvrmNO37//dl3e3NacDisKv0/XMcQsjw5LxfI2Jm3WBw06grTPd5pPRRdSLMhP5pnGxQiODE8fi4q9u6tnGmDsrhNl6dw8vaDwLn9+Sp9CUKYDX3OIkoUU4zOFLIQUfjpyRZAARaA0kwpcoL3KKmZE+PcYC2oMnRVdGLttlUs6W6FPBxsaL9c2tBZuban9Mzj+zGd0fMTaMIF0B/Kr0dliLBc53b8+HzzPubhqN8aJyUT1TOCe9uHP5YzyL39LjlT7B5LPlLnOb0V6xq+8qXn8rS9r/XnrtuRUW1tYX7sa8xHd/Kr8Kptr7KdzOhPp3IY7YbZcKtfBWjBo001S938f7nAPj9QndHN0kMEVyBmKUnWbZ69HklMUuZtMFvXa3pLEKULusLmjfMUP6Uilda7A2jlE7NzGb/cg3abj1fGh4itdo0LIa3ro0/JCogGHGw/yNJ8Olz6c6uAO1oIL3wRxpKR09ZK6GUqkvRXym4vj5RrdqZyTSUuBmJQj9s0x9XM4AIWQM4WID/d3JU9857YcHPewJnQhT6OzW8ORqKHh0L2CKl+l804JRy9/TrdBOt9AtrWl612D5PwG3jN6bV3/R59onSvgyDTHtg35VPea5VybIHn0X89cdK50Xu3shXZpvC7qf9Kx4smP/p1cqB787Id1+eBQMWQ8U/9ewzvlF65JNnVvT/N1c6C1iJLHEP3brqyIKZtqbAXC/kfftRFHuDVFD31xwliDednx1L8t5FVBwyJSzzqaa66PxooTk2ElQczXNxdPz87cb/7WwlWpHWgNCbtqnwFy1zYbC/HMR/x1jfwHZWytEUdqkyBUOYILcIC1hQ6XF5dshQgSkGflcO/jsuVjTgQhtlkJKG2HCyMkVyHm1gDj4gbymQfYWuLffv/36vJXfkU5/q9+WeVXhTFxDAaDwWAwGAwGg8FgMBiuAOwjjsFgMBgMBoPBYDAYDAbDFcDl5VQVVam5V/1yOh5p8y1KcfpwLQKFdzYR3Wh7ILrb1794ty7vQjpQgpYYwRFoDjqZf0GbAw0qA1WVlPcJaMmnQzkSgJ3udrGr+Q6cmHqxaFMOtLBJous/OdQ10wISIMh4pgkoYHXd1ktVvaDYD89EE/7xD0VZfPn407rcB013C44ZO5uiEu/siJ66vSX6Yr8j2RIdw5q706ucgBJXgLYYdRdtjg3RXU6Kq0d6ntr4bKRd+oNSnfjxT0SF394WbX0PUrB0JorueKLr0B1mAteqEJS/uK2x2B2I2rlOlKVzabpoL8qp6IhTwEGsgHzBwW3Nd5qLJTjKh89Em6dTzjXIFmcnev5TyM+4g3sMl7FWRT+OEAuiEJRwMlUnmuenz+XIcJqoX96M1e/dm5IXeZCoUMbCecQ28/3lVPEGXgPlOPB91+8tYsj2tsbJN78pSuXbkFDdvSt5QOCpT45eSvL4HHKq8yNJO1I40U1GakPKNt66/25d3tjUcZ+ODP5nZWVlClcz0IQ9j/IoWq+A1o3ncHRVwPVjBOABpHh7mcaiDweRMeQPT59IRnnvDUlX1onSOZdVsTpPl0t8Akr28PyUbXCeUSpEqVQrgJwQ6wznDt0f6UQXQprQqWj/3QhU6FAxehMOC4M2ZTm6xhz6Yw9j5Po1yGu3uzhHdZ9BZjWGZCnmfIXUawaHvPH5gna/TjmV7/uuXcWqMaTX47Fi/+mxHN7+4A/+bV1+8UKujQEsRfrIbX7lTa2Rf/0bX67L797TnL6GtajbVzzw4DzkENsuZFEJ10KUKZstC80b5j9BqDWiFdMRUucHEWSIlP9hTO+f676TGXI05GJJqrZJMrXrOhGEgdvZWazjm1jPBxui4feRw80h4fUgIZyPNCb3ttQXlNg+f6TYMoeLmYecsiHFwVro2E/VHPQom8La7Wdwm6L5kgeHQMx55ig5XVsaS1i5tNg8A7KsRmhesl6u2+GollNBhoQynapCSBWZz3chN72/p7l1G3L6PuQubZRDyqIQfx3d+SBnc1WeXKJfKa3lWhjmlI0gtsJVLA40P25fUz75ra9pjQ5Q3588UI40x1qf4Q2voNKest/89f0tv+4ZDjcORMTLhz//cV3+4E/lSHwOmePJmWLXoKMYdXdXc/0W+nezqznXjinRUd9RAtSpyoyj/gopOJ3IQsrCmf9yTPl0edT16YBIF0Pf49jHXB8pvmZz9d3R2aLfy8va0P3/IM9zd1ptrbDTpUuXxm07YuLOGK/xHGL9CTGHGB/bbcimMLdC3CuIlp/DbTkuWtALsf5hCmcZ/sdnW0Hyjv6MI2wLg1jjQTbsJTiHDr6UiKE/H+Md+/e/97t1+Y27l89RjYljMBgMBoPBYDAYDAaDwXAFYB9xDAaDwWAwGAwGg8FgMBiuAC4tpxJdfrm8oHGUO0aDZtWLITc6E0U5B7WcdNIYVNG7NyXjoUTHj0U59kEFjrqL45REeXDEKRJR06YzURjPIaEpQHNvgXoHM5WG4wN3zW5jN/nxVPWNPdGvghWU17xq63UyVcuicOl08Ww//5Hoi0+w030Hrkt7NyRTuX1fsratHUmrunCviH24dEHuNoUMpqDkB+UEDhp00wgqeiJ3jJ/loCbHoke3emrXbiq5z2gk+ZoDre3Zx3ruza+J5r63LWetBPRwH9TBOZxAJthZ3wslRQriyysWXwVlWdZSjBz0zQwUesqpykzP4EHaEoAqfPxCbXR6IHnUXfQ7Ha/GR5q7PmSAMSR43HH+okRqMaVVsY/fgWYZTOCihnod/Nl7dfnGQHTaaAexQDVv0sAblHDOPzjuYOb5VT3X6U4VhoHbrRyh3nlHblBf+qLKb9y/XZc7DZc0ucm9fCFa/3go6rHDuJjP4EgD6d8d0Dc3tiXtCFtqwxJtElQU9YaENlguAyrBwc/B5c7z5Tv8U4pFxysGQBoS9Lugojtdc5Ko/PKZpGYx6e9rxgWDmeMqgvSJEgfKqbySsj5drw2ZUwtztAcniAHcVELEUQcXqrCjOgy66tPruwvJ005nBa0ccSsExXuKWDiFdCzsiPLf39IasLGp2DzLEOshRT7rqbx/qnViPpMsIEcO8PL5C+ecc2lKUcXnhee8qr+OjyXHO0S8+eRjOWM+/PTDupzMtc5swhHjy7c1n379y5rTb93W2rm9qbjVg0S51VM5Bxc8g+OVX0kRGG/pnpKlHO9wSotSHJ0uLTu43dxEmkjaeoLciTrUEfKvGcZICjkVY8M64Tm4k2B+DDY0PjnP5pDO+WijHTisjSE7f/xANPiCa16w3MEtAl3fgyNXzrW5rNqrpAwS82CqmO6liHkJpT1wpENuWayQUF1acrFq2au1C+vLUsvSORmzQuKAOjBHjeE8k8ENtg+H0xtwFdugxAbxNELZx9gJWKacCvIcv4rX+Qp33iDG/MN8poQ9gAtujnKrq/l0+wZkXmibDiT88UcP6vIHTyWzmiEVpeo3X6Ms9TMoL+TXOIRxmE2VN3/44x/U5dN95aIvD7UOMD9787okcjcgnaRrbIB3RDr7hZAz0YmsfVGG3Ipyex8JiI/13YfMNIcsPMRaXCC/yREvOX7jYPlWAFM4I44DOt7hfafnVXVxa0Pp5PZcYP1uN/J2SGYzrNllgvJy6RNeEV1BVRafofE81HWqGCJfuhhfBeX5OfoT3wLoPJX5cNNCmf3QwvtfzM8VPt5pmdOFXJshc8XWLe//XO8xL15ovr4qjIljMBgMBoPBYDAYDAaDwXAFYB9xDAaDwWAwGAwGg8FgMBiuAC6p9Sid85bJqZpeVfXFSX9sUWIhutEMMpT5HJR7uGrMJ6K8hqDTcXd4Uh7bcB/wK1kGXXBKUJlIfQp9OOlgZ/uG+8+Mjj9wvPIoHYBcAPITH+4gbchFuthlPZxh13jQj9eF+XzqPvro5845554+lpSo31abbe+KpnjrDdHAd65LYtQFZTok9w2UuwxlyimKcoW0IhX9zqHNg5p+B0kIOHHzALTWls5p9zCGEsnjhqdy7Xn0kcbNLlwobt6VjGwwEBX36FgyFkpKRkNRrn3QOc/ORMdfJ8qydGnVvnS1KT26DoC+DlpnC7IcD+4hhy/0bByfEeiEx4c6x82W98eqL8NeWXOkdS6oqhFopbxIC64hW6CVnj2UK8zLW5Iu3AO1lvobUnHpFpSVn93Z/s/jwvWrXGlfdXm02233pS990Tnn3BuQrN25rbG3tSWp4mwq977zU9GNJ2Mdj0ErnU/hFADq/7UdtVUXsgEHF4DSB7UctOGoiuOU5HmgnuaY83ReakilsPO/16D4Y+w21BZ0hNBRH+OCziI0vzsfSpbw/OlT97rgVzRoUm8jzAlKqAI6lUDnG0DO1APF+3ZX51yHq1MfC2yM65PqTMnZAOVuZ9G/HC+c85T45g1HKsXaDPM1gpwxaGttaMGR52ystXOKuNOFKwXjAZnTBdaPRw8XkpaGnOdzoihLN6+e8+RMOcnjh1oj/+SP/rAun55IMuvB7e82nLn+0jv36/Lda1pbupQ/wW2jCDtLy40BTTlV1UeUePqB+tPzKL3ANTzQxqG4ihraG42hAcbW7R3Fo3kix60R3ARnR4pHfUjlOL/DcFWk/Xzw/cANBgt3mo0NudS00OYp1sU28sYCErKDx8wRJKPzIaHudSDjDrAuQFpfItdlmpTCFWtWXTNC387hJlkMIcWBhMfD/Gu14QLHmFou16U2JcQ45VX+rvt6uq5x/cy7kIZQyoB4g7W8mCMPwWW6cNXbhKsmXcIKxOgCc8fHb0O65kBOFeD8oIqjjO1MFeiYSblbGSt2zBprIaQ6uaRS7ULjbw9STFeqLt2e5mjYUd3/7JGkxRlicTZb/3vGAp7kcHQBwpA82dea/PKhJKrHJ8rXj88V+9+4o2e+DidcugPT8StGLk5ZMre88DGPvOpdgtsMBFijfa7X0Wf737mmzDefQjbVwvrL92SMDW4vwPk9nOmaJyP1HZ2hsu5rkFOVZf08HgyGo0Z+oPokXGY4zhGHfMQYH+2W4ZrcniSK8O6IPCBBLPbRXxflxvzzlkvbi5JSYTo7Ip9EruLhol1sCzOHa1U6g4QOcYrvt8kU7rtwvRzDWe5VYUwcg8FgMBgMBoPBYDAYDIYrAPuIYzAYDAaDwWAwGAwGg8FwBXBp65wLtiAZvj6cIyLsvN4BVTgGTbpN6hGpp3PRcJOx6HQl6GkR6Fc+qaKQRZHSf7GDvA/3Dson6G/R2KSdVKxixT+A/h44yBjgCjOFdGsOiRglMBlcC0pQmsPPqk8+N2bzufvo44+dc87FcCbpY3f73WtyxuiDMsy+9UExo2wi90jfxXM5UuuWU04dnYEglah3cs8hVQDtkC4U8xw0VMjt6CRQjkSXf/Hpx3X5pxhbG6CN9/qitg8ncMgiJRU7oU/HcispGu2xXuQVPT3HKPZByabDmwc3ihacNNIJaN0T0OCvyUFlOpIkJRnp+pQ2+gF5lCqWoD/W064Ilp3qPPBAY8pSUF8fDjsjOBs8ev+ndXnr/p26PLgldyePEiHMMzrn+XTuQpC72EV/nU5xURS52zcWkoSbN27Ux3cgoUpn6pPZuWiXc4yxiHRfUMUnuWIopWo7G5JLxpAQ0H2gdKJ+lohnZdXPYPE25rajfGCFS0qjDfk/iCl5wz0FMRGuEk5D1/nQKuiJmmP09HDfvS741cIYYC6GHsY5pGoxHCVCxkW0XR/yjDf76tMv3pEsYLuj63chNehAChBEkLOwZcKKIk1pD5q8BPW3xFoV4vwItH0XK0ZmaPM0V98VmPfHI10/g6MFnSu8jFJI1eHZ84WMMqX89nOiKHI3quJfkaohjp7LLeLBJ+/X5TFcjTZB2X/7Ftwcr4v63+2oHwK4w5WgczPPyZELNZz0ICl2FxI+pkGOzhg8LtD5jbKBGHPLyzVWEoyhGLY2O3D5uYucgXT/EWRDLcTuCHnFOhHGkdutYv5gQ3KLDiQWAcbVaKTnefyxcoFncCILMM56kKf0Q+YryD+Ru2TIWYI510g4KsWLsTSdKl7TnTHLNdaiFJJHSO4Y30tIfph9+A15wXKJ6qqtEfj3Xq+5ai859/OhdJ7Lvc9KQwoki3PkW7xzj+6xXKOwADHPyNBCfgx5KpyqGm5HKPvIJy6uGQSUEzucy8mIdRHjxqNcrAUHJ1wnwUtKq614sdXD9gOU2ATv6rfImdMHimvzVPmdy7Cofk54rnRh9ayZg2wNMezlE7m9ne5r64GzM8hK8N6xhzWnBQl1BHc2yqYou6NU12s4TuFW1dzlGCHTgdKnEDJz51HOA/kNrhMij21DihNiPMwwHluUHEMKvYFy0nj3WlzTX6OeynOaX9wSha573CpjVuActFyAWdrCOtOG9LohC+ZrNuNWtvx9ilueXLhcUY7GvD7Be1GOLUsK5Ln8hsD4Qpl5iISpoBMr5muA+d3Bup9hzh0caYuK418iRzUmjsFgMBgMBoPBYDAYDAbDFYB9xDEYDAaDwWAwGAwGg8FguAK4lJzKc54Lql3QG0YGJeVGohjm2DF9DncU0tC4Y3MOmuscuzcH4P+Cke1y0M18UL4LuFaUFf27AD+LO1w3nKeK5ZIf3MblIFcVLHPzdfxPSroW+ZVosyJbfk5dhTVqODwn2trGpqj5vZboze0OHRBA5W5IIiiJQt82duQHPY6uI6AkZmgr9lEK2lxZfWskZY0U4DZorQHofG4GSjIoln2U53NRmD967726fOctuQW9+eUv1OVeV7KzGRw5uGv56ExOVVFwacXiK8NfRmXmGCYPFwhQp5NzuRxllDJAojZd4QKzbEf4P399uvVc0Bv9FRTXICJVFdIPUM8pqRt0FTuODk7q8sHHcpTpbEk6FPRAhSadvOQYp+uSbnshx/TWKqiSIvMNuKGFPtyAJuqf2VRyqrIExR5NNZ+iTQ4ludroiIZMp76CTnqQc2SQlJDTnuWL34a4hg+qdIOmT+ox3cAaMRcuI6BNl/j7QsNxjUEI4z6BU2CKgN2G/Gc+X5/8hvCcc141dyhta4M2XDIW8regYUOd4TqQZd0dqIPv9BDH2voBJVQtuPK4tiQlHmQxQbRoo4Z3DeZ/g0JOVzdQlLkuzhHHGQv5TEFPz9E7xZo+UwzmWlhiPShBG5+enlXnLo9vvwzyPHfDypVqPtXcKhoaa7QPnmt7oLXz5p5cqProkwhyM8qpKPEskY75HkWBAqUsZSXhKwOulaSBI5ZxOuMaAfMiLN50LQzRn+x/Orz02qrv9gCSV8RlulVGr8niKAwCt7OzGPMx5NQJ+vTkSHV69InkHC8ePqrLMXLEHtalHrYCiCALKRADfY/9iLbj2oLr1xIWyAodXOByOKVkyJFbfbgJUmaK+ep5y+Nl46j3Cn1BR0GvEcF+8W8vjVLthrji0+1vlTkuXXMgjzg5Vp9PpnB/HPC5kMM0nDJxX+YlzO2q+M61jfUNw+XzmXKYAGvq3KM8RHXMKa3EWhN1EP9x/kZPsek+XC+fn2isTedqp7P5+uRUzrl6ePB9KsM738FLucANIdtP8T436G+jjO0dAuaLy+WBZUPmgr6hO1Vja5CL/1IGjDWgWC65KRo5Dd6HmlasqhfKdLYqGg5+XG8gR6IEOsT7c7WVwhoNVJ3nnOtUt96AJLiNsUo5bwG5rQt/8btPq7EvS7D0HLpWUaPIXDxCzL3IV/idofnujTXPsb6oC+J8gLjQCWDRhXFMKRbX6BTjglL0GNsecKwPsY3Mq8KYOAaDwWAwGAwGg8FgMBgMVwD2EcdgMBgMBoPBYDAYDAaD4QrgUlqPKO66W/e+7pxzrgB9je4lRUaZjWhCpOkNp9htHC4ILhc9qbFTNWhRpNlzt3pSpDLIP9JZdT53+wdlO89RbkiZuPO0buMaThEEOXyg1uH6Meha/TalBqD54aK1BGGNjNUgCN3OduWaAU1clsGNZIVsiZzsHFTtkDYAjVbBWKAxw4oyT/Ihy6vlLqDxhuDpFxg3LiaFnfQ1lTvY3b09Fa30/FxuRx/8RG5He7eu65q4DimZdBCZDEWJm74CpfCXQunqgUnqJ2nzKam6pHJibJ+d6plL0BkpCcwgVYkbVHzdi44AdOWh1MZVTixeSSorHKNwDShrGrvtBwhZvUL0znGqNj9+IFr8tTffqMsRdsIPQL9krCk4vykjcOtHp91yX/3yO8455zYgNUkmkuMlU9HAE0j/6MxXwN3u5TNJrk6P5XCzfV+ysnlCGigtNOggQVmZjl+46vk+aNfoTxoG0h2uwNxNIDFNEkhBoJUt2T/kO+cMGIjpqa4zmapuIerQAYV1nfB9zw0qOjslJnRVyArE14bpi+oUQm6xAYebvZ6es4sFokP5VRvOF5DJhptyDfIijX91KR3JhIaDYAKHOciiizld3eCOQgklHjaE+8sGpGAtOnTRbYNiL6ypwYXUeo288SLP3XkluRidS4a4fyC6P9fFFiQ2XTiNDNAPMd1m4I7EcsDjXLswbvmYDQnZhfMLXeAoZ86XrwsRpY1wz8h89aeHNT1GMJ4iFlA222tBWoV+ppxoNJakKStfz98PPc9zrWoOTuBCeQDnm4OnL+ry0eMndbmNPKIXY87BqYgSmRLrHNdOH050IZ2NIJ1jrL34bUhnMbhM5pAQZ3DfDLuYo4g7pPP7+S/OzVZjuYTKa6yR63en8j3PxVV+F69wA6L0KUV+wsdi3sA14eRE6+LetuSmrrE9AvN/akIh16c72YWmpLHPBNsJRxsuYXqmFL+lPDVlfsL3LrrjYox6kGdEMV22sEYgNrWi15SjOs8VF/2EZ0jh7Hd0KGeehHJLzJvtTTjFdrSGUUIVUs4PmRnv61O6jPMbbmtVuy/d1sI5lyF38ujmiRwlw2+5BQSlO5yVPqR2CeRIjK+NLQjwW8r+skoCu045VRj4bntj0ebcfiNus43RDg19NmSAmEMpvhdkyAXpsMenzDM+kH4bNRz8cM2qXyKsQ5QWc47SpZZuWsxX+d6ZUdqYLpfTUfpGORVdq1uQP272Nb739vSu+aowJo7BYDAYDAaDwWAwGAwGwxWAfcQxGAwGg8FgMBgMBoPBYLgCuBSPrt3dcF/8tb/rnHOu1ZOUoYXdtela4nnYgf/8w7qYh5ILcFf/mKw2MhhBZ6KcpeGctGyL8eoOzjmXg6bd3AEclDVHCQko5DjuNbjNpHvzlqSX6TptUJRvbKr9bu/pp6cztY0r088+zueE73muXVEVJ9gNPuddwIkjHZoPSSlNRDogWWukxnMn74w7qqOtQE/0G04Bi7IXUNSCvgJNjVS9HPS8AvRM11W9wplotu2hqH3PIck5eCb69c0379flOJYUYjyj843qdnpy5F4LPLAVqSrM2OagKmK39QTU4uGpqMXoLjfD+KcNDd1MSlIU/eUczpz8ymrutEDxDhtl0GNBgw0wR4MOyqAK5xinQ7iPnD97Wpc723Dt6aHsk3ZJ+dcS6eQaqapRGLgbuwtpYzbDXMwkm0omOt6g8kJ6NDpTPz99IDer+UjHDw4lLZhOdLwHJ7ouqMqkqlLm5gWLOhSOTlbq4/Nz1fdsJNr0+URjLkMMpcSpjzVl0NXxhiMFKOFBwPGt32Znuu8Eu/1vbsnlYp3wPa+mpJcYhz6dVUClLhhrQfkv0L87W2p/Ok/5kN/EXT1zPNB4jnqi53otxL0l0s6SFGYqASADdpTc4fmakmdI5xinIQsrE/22D9eU3S05/oVPtf5xqnUiSgEW1w/Wp+BwSTJ3jx994pxz7j/86Z/Vxz/55APVoa/2PjmVPCcodZyOKVEEqRyo8aTyxzEd/uh8SVes5e55F/HUXxErPYYy/CzH3F3lOOk1yrw/YwGd0iA/guIkSzTvsey4Ntww14k8Td1J5XhzNkQsOpVE7vil1nM/k2wphGSvg3yBzPqGaxHapeHW15AtYv7B2cijRLhqRzp15siRKEn04VhTou5liW0B6PJXLM/fVi1kXtPaB2c39hRYUV4PPM9znSrOx5g30QrZScKxSiknqsbtAk7hHjqbak0oij7OR44IqSCdHf+/9t7sR5I0u/L7bHMz3z32XGvt6i42ydFoBkOMqAdhIMyT/uABRq8iJGgEiODW7O7prjUzIyJj8fDddj24hZ+fsSPRlUxPEUHc81JfeVrY8i33u+Z+zj0FYttOfsN3AsrRWsc+7KDFXInvMy3ZFPdltEu8ZxR0vqUMlU6gOGeeP+w++qGoHfY6uv2m2J8XyD8pW8PgUU6V4P2vw1yAcY/yMMyZ4B3vqR3kkfdyaPa5hzFiMKDEqW65cnJ8NS4sb0A3pfodtSb4btpP5uaroQAAIABJREFU4GiI+U4nxbzJ2es96qk8r3adePv8YQy5PfqSLpKufFjaS0VUgWdMKRlDfAx9lLzoaH/luIV09XIP5Oo+JWi6gX4PUmGWCsjhDsiXIaxLOpzmdOJq1QWBtAv9scHxlAKOkN+eHJ+494UxcQwGg8FgMBgMBoPBYDAYHgHsSxyDwWAwGAwGg8FgMBgMhkeA95JTeWHkOkdnzjnnuqAA9UGJj+EAk6CC9eZCdP4IcgFXicpHmQ1dM3xSqEB981GFnNXkPVYnb6iqlATQYYByIdL86VJD+VUFWVCJqvikc7HidqtCd6XPk0DnPBmq/w7hUhO57bUu5vujrFZV6VbzLYUxW6vv+11RSWPQ+DqgrEV0OALd9PZOY3s3g2wKFM8KbgshbFtCOHb5oOWFcP8IG0qkD3cHSrXWkJzkGB9S2ZZrfT4DzTqDbIjV5pc3eqZXv5e0anICpyrS+TCPAlDkF7jWPlHVtVs3zxqgOnyAOVbXpPupuV6I176c6v4qSCJ+fH25a8N8xR0Ote6Dd1A/yYqtW44P2/FIIYPxS1BicziyFDrJErTxdQ7HrQ77X3HHC0SXnl/IXWb87Ew3llCWA9kDbbHcA3RZb59rsXKb9TYWRnSJ2Sg+Lmd6Fl6ajlTn5+dow00n1TFTSIy4ziZjjecnz9U/46H6sxNDwtY4cmSQUM0XoqGeX0jOdQUXkBliDSnJ/b7iDp0nYtCgz0615p4+e6LjB6LCp9BxJomuNZvJ/aKqtKb3Cs/b0bkLxLyUElLsZxFiW8C4ATfHCfolBo05gQwpAc086kGeAqo295/2zG1o4470cEqlEBfxVzXlXxEdPkCvbjk7UN4JWUKi5zs6knNaHMAtCKcZQRbW723nSehrfn0olsuF+3/+218555z777/9/e5zD7E1TRGHVlpP3UhzshOC9h6KBl57kJshxlCpViNvKFvOQA+P4b1jkM8xhMSAFH/KT1YFnNwQrNepjvEgoW3lMC0XTrU7iP/cI4aQAt4uFNdarnh7RFVWbnO33dMKSA3mV5LY1szbsEbpGEK9OOd8zXwFDlYljl9Col35zAvUpiNQ1NDvM6oqkHPS+ZEyq9W1jvnhQnGu01NcfPHis12bFP7WmqaLaP2wjI5oHePtT7qhc3ouauYT74GuimvI1ykZC7BXJZh7FWJryXcOxOUS+VLRKr8A+SFKJYSYF/fLpW7JqSh9ZN8/LGtqlYfAmFBmGbTkPJBEYRxy5OZ0gozwXhRh7lJytW/cxyDKWdKFpI2LO63LDXI7Stz7cJNl3GDesfS11umOOh5on+kgvxh0md/ANa4BJUslJayQUy4xHzO4NhYtJySUeoAUp4CklfLalhETxpfuk6M+ZO+Bcpr799F9yqkCV7tR2cSzjfo7ClUDxPf5vJD1UU6O0hMF3r/v0Ic5JFQR8oa0QsmLSP087PM9nuvLb+6dQBxB/2TcI7G2Uki7ikr3njm4BiLPan9f8LD0NGe5EOy7Pe4Fnff6SmZ7rvf+C4PBYDAYDAaDwWAwGAwGw//vsC9xDAaDwWAwGAwGg8FgMBgeAd5PTuV7Lmwoii3HmJZUCbKKVsVuUsh1ztUcFHfIX1qV1OHq5FNmwwrZuM/6AZlHB/R8Uh5JTyULrQbdqaooiaKDBJ8Vx4MundPhBuekBKaHUehCahR1t8eHe/yqra4lDyM9fDwY79rDAarBww2F5cbXK1HrZjeiRM6nkn8sMLYrSp7Q5xGojEcnogH3h6JBRs2cY8VyOi8tQc9ckUI90/VndzomBPU0wmTs4Zqbuajf53A4+uT2y117eDzZtbkeclCry4cZyfvB7jp0Q6sebHuQoqWp+i6GS0gVi/6/BFX17Y0cttKJ5ox3pOePIO/phYwHoKI3H6ctNy0dO78VhXIJZyO6FvRHcuE5PJYMYzLUvVel5vISUr8ZZEfBSM8dDiHRbJEwGb/8P/jsQ1HXlSsbqRjlK7fo7yzTuolAFd9sNIZXN691TKLOPTmVzKNAhLy+leRpulD/JFdw9qjVh4eHigf3cW6xFMX1LaQK0ztJXChnPIAz1OQQ4zbWdRJIQXJIO3Ks9dtb3e/ZmWRWo5HmIvuGbllvr+QotE/Ude2yhsK+Wut6pMxGoPaT1h6UioV97HMd7Kl5RmcvOCWgXWRal5HT2okptQHV2TWuEBGo4h4dGSBhzGErlEJ+usRCXmP9F1hDPvZoD5aTzAcmGLsx3EfGga4bQSZa1dtj6j1axZVl4e6m23U3vdU8oQPfBnGTLkW9rmIJTf3Op4qhARTkUazPu3CHG2L/GQyHDx5DCe99O4M8ivvyFM4v11iXdG2iVDxAAkSn0MDpuYsCbTpLIi6OkT98Bonm7fJhqfU+UZaFu2viG/eQCnLqbKnPe8g/Qsy9qiWb5VrUMUu4/13PFANv7uDIBY1UBSeWBEndpL+NDWdPtbe9eHq0a9O1b7lRv/3q23/Ytc/neqa//F/+867N0gEt56mfsnaQ03qtVwVK/fb/O3BVVW7VyG8pWapKur5A+tfK9+GY1te6OTnW/jOGSyXHeYn17QWQUy2151Dy1Ie0tWr2KL+mrIlyG80JOrxRcnIHueEMuWsG9yif7yKQZ+QIPFRz0NEr8HUMyyQEH+unfM+TyxYkXtdvJJldTLVuMjxPSOcpSBIz7HMVy2Ig/67grLiAjDuAW2CvR8fLP3Sn8hFTc0il7lLFzlu806yQo9CsKcZY85y+e/hd10WU6OhjGj5OkLs+wTvTdLO9z+j7/Tnijvs997/9xb9zzjn32zeSbMaQEnmQOxF03aJr8RT7T1FqX5rP9Z41RZkHhqrnZ3Jv+uLFi137EG6zSdh8R4H+y5DnnM81587vlAtfXKuEBOV0PThjco+MMNB05CxROoQxK4XMdrPSnBrDiTXp/KG074/BmDgGg8FgMBgMBoPBYDAYDI8A9iWOwWAwGAwGg8FgMBgMBsMjwPvJqTxvV1WfdO8QlMqQzhj4jiiMRBX+8VvRvV7ffLtrB3C2OjoUxfoI5+wkouX1cN0AnEAfdM+6cWWg2QIrnJdsg79W1LrmGk45K5is0KGlxnkK0ObAIHYs5D8CRSuBa0EXlPrhZNsH0fn+HI6qqnLLxlljjWrpB5A4jCBxCEBPJR1seivZ1OWlHHGmM9HjNqDhF6CqlqBWpleSOc3mooI/hZPQ5HBLGQzB8U43og9Pp7rmxaXm1h3OF4A+OkhEX2MF+DICxRHz7A5SrCu4HQ3GkHyBzrkmLR3ObfuE54lq7GHeei05FejugeQclP4tMKGnc7UxNVyG+T+/0LivQVUtIW06QfX8Lmjj98zSKiU9VRdarDRfMn6/jMr2OWQMlLCxwn+nr9iRrySNmF+KOhmNNHa9jujqDo56pLP6O47yPt04KldW27U9h0vDFHKqpKs4UZao6j+H89MCLnMD0UoHlIzBjSHs6Xlv36p/bm4hw4F08fREsqy4uZ/rWzicreDYB6OLGP3a66G/IT8JQImuEDdbrh2I/z5lthhzulwNhrq3g0PRjSm13CfKsnLzRrZFR7ge+jyA81sEKW1ca+xO4epDuvcdXMYuvxXlt/zxza7dxzx5OpHM7MUTuUj0x1pHfmc7vj7dWbAWp9da5z+e65qXM1H+c1DFhwOtuYOxYkHSirVwq0Bc7GPff4nxirC/OtDrv13cf75fp7j7+VEUdKnRvkXZUgSZMSWjN5BELEilxx4fYV9MFtrbh0v97RhSqNNjjeForPVdNnvXCvKg2ymkkpCSLjeaZwX6NewgVmO+0k2mKiBtpbTD57rU+HQTjcuTQ+USJ2PFuG/fak/dJ7b5zbZPc8gqM8iDY4wXpUo+nKc4B7i/bjKtixnGi0rEPnI7Gv8sUo37Yqb72Vw2n680tkd9OpZorv32G635tzc633/4n/5y1/7y85/t2ow7LXvDn7J0PEqxIJ9oSaj8Bz77MNR17TaN006EGE9XJ+bzAebqEPv6M+SQf/6nX+/ana5ywRvI5u8gSTvB+8fhoc7pIC1MF5oLYTP/Pcpga8XTOaQ3V8id1yldAJErIVZSepgjj67QCZTNOcxvull5dPFCm+9y+0W9c0oqEDfefCf3vxTy4xy5aw+OVFQbDSEz7SC/aLmgYj7kiHtT7F2dQNKgI8SowWAbAyNIoZfIc2eI15uCblp0/tK9h5CzddEuUzhrQdJOd922a5YCSRdulce496PmZTPc43h2wsC9ON1e4xXf7ei0zHInkA8xZiyxR3WQEzBvC+nm6BS3mBtfvFXpgGylz58eSGZ1MjluzoH3jDXKQ6DMxwXcX7kuKUlM6I6NtdXtIndFGRSWEfBZTgX5wwHcvZ8eKecJ6Gz2E2FMHIPBYDAYDAaDwWAwGAyGRwD7EsdgMBgMBoPBYDAYDAaD4RHgveVUUUOd6nRA2cNXQaSkepDidIaifdWg1n/7K0kHqlA0qzdXojydTESzenEG2caBqHUjVIonzf6+cj2rypegO+WgzS5Bc55OQbOcLR88PgX9fLPS8V4FmQmqjU+GoisfxKLcJbE6kNS+l0+3FLH416L+fSiqunbLRhY1BjX75EySiS6o/LTsukPF/O9+UIX5168lMbpeg06KvqpSVthXvw1ADYxALWV/1uNtv5WglaYZXDhmomaf30BmskZ1dIy5qzXOMeQ+HadrRoHuPcKcvryQw9GzT57rPBhnn/LC4L2W2E9HDXosaKh0eKvA8W65t+Ge7laiOX4D558ZXKM6HVC7UeV9DslFRj75mZw1nh9pPvebKux0waHbxi0kXN/fakyvN4oRIVx+jg907l5Hle1DSCCOOnqOzx2oy4moqodHkp8EkBSUmPv+/WPX+5NTlWXhZovt2r6+lGQlh6tKEuveilzjcHOtzsozSFYK3d/5hcYz6KhPHOihcSKpRllhzUFOSgezJNn+LeUTPpzPuC/QqWd6q3V5C/ljiLgcozJ/gvHpQ/5VgsY7neqYCdyvjo60R1A2QOnsXlHXrsi3DxtFpP/T3QUUZ2gsEsjGxpB21L7G9HypY2ZwwShCdfBoBeekQjTjARwUQk/j2G0kswUkE/OF4uLlVPH6rtCau4AkhG4u3dUcn+u5n8PZ7PBQz+SBKd6F/OQzOI59eqL1vUBfXv9+G3eCQNKhD0VdSy41xly6uVLsoUtaghxjNNIzHhwr9rmePl9hLdCxMMV87mR0f1R7Odda7PfgptSMLeVOLodTCJxUclDhV5AZ+ZAk5F3MP8S5wUCfHxzq+YpMn8/utBf4qe5n1NM8Gvbh/nT5cdyp6rp2RbO/zCGDZv4xgWyxD2p/CPc/uhTWGLBlxtwO0uU15OIYOxeCZg+HrE6kzz9t8rwnB8qLE+QTGWL6r7+RnMBFml8/vFKO+PwLzYHuAPMxfJfDlPeOzyGT8CmneiDG7VNO5TxXNfdB9yYHN76KLjF4rslEOW2vq7i1hGzneq74S/dCHzFmBRnOcq15VEOeOoQUcZhsr5vmOscGee4t5M+UcNHZkf06gVPdIdy0RohNdBC9hRQkCug8BXcvxK8Szxd/xBz1fswWyNFvr5TrcNbkeJ/qjFFuAjnfDHLVV6++3bU3ueLJ6TO9y5yivwLkPXOMQcUY3Nn+bYVjF62yELoO1/M1JDpz7JEd5C5jSMFO4abaQ/mKDf62gLSqxPM5lEkIkT90G3c9vut+KCqvdmnj1NbBe/gS7pUxpH+TntZEJ0BeileUJdbcvWzSOed6XcjasBclfJyaDk/q88sbvYN2713p8G73DeYK5cRdxLWfnT7TvQw1b4ZDrb8Vynj88FbvwOuN7qWLuJDjWsxvnz/RHD2BhJwS/J8KY+IYDAaDwWAwGAwGg8FgMDwCvB8Tx3kuaooVsYBXiV8fQSpp/Wrf6+LXqidPdQO/+OWunaAg4nfffK/2K7EfFvgm9u5Y35A9Q/E/frPsNz+hk4nB4lRvr/QN9hQFQvHDiluArdDFt/sxf7nB+QcoYDXBN2sHfRSWxDeo/OJ0gG/rRk1xL36b/qEIgsCNR9uibU/wq+fkQN8GxrHueYVvmN+igPGrC/1CeXGrb4mvUQQ3xbeQAX7hGeBbbg+/kHXjh4uZRk0RzNpHEU4USpuxgONSA3e7BHMBv1SXuJcO+rta6fynfbTHKFx5rV9/5/h1ZYjC0AGL+VYf53vS2jlX3f9KVejZPHxD3v4cX4Xjl6sNaHRLfNu/BHdnU6mPCvRjjPVd+fpV6GCsXzo/72qOHT3drlEW1fvxH9SHv0PR5O/nus4cv4quUZyuv9YYsSDYAHPjl2cqUHiCeJFMNWdYuHqCX5p9/DJW3a/vPdY1LvLcXb7ZMojm13quGMwWD3FiikKaN1P1d4bxYVHk23OtV/5CGOMXj6enipsvnihGDw/wiwDi6T35goUlByP0K+b7+VuN5zmKSrP4eIVflcZjXf9krOu/eKoxQf1et15ofAYojjscaS36LLT//jXjfhJq51zd7DM1KCYBfqHij+AZmXpgobCI/F9/pwKm2Rq/5kzURwF+ohpgnpeh9qgS97NYaH0nw+21PKx/snLCgeZIttBzZGDlXM30S2GBIp5/9903u/Yvnmpu/M//7s927ZNEMSIOtB8fjrRHjhPdw6+vNPeD+0mwx18cfd/fFWFeIjZk+LU1xV41HIIR2CqereN/949/u2tXWHMsglmCIfAJisN/8lSFGrs9nbMma6JZOykL5iI/WiJvmeG+1oin332rPGuNAuVkR/WR1H391ee79ukJiqiDOeBjTfdQrHLUR5xIFAP2iSIv3M3rLSuFbIMD/CI+RBDpYGHWtfolQ45wg33+x1eKadd3LNCve7hcakxTFEidgnXx2bHm+Sdn274LUcw4RMHPvIPc+UzFej/7WuvpxSdf6JgTxfQARUdbxYlbRakRnMio8Zmj4vMWE6eJI3tk4jgnwmuFIvAtxhDaEdi2h4eaYxEKw755I/aHh8Kz15jzFcYqhlqgizzjFrGhe0Zm4fZvqwKsLTAVQeBwOXLkBdjHl7cyGfAvxKw6A7vo559/qntEsfcAbBWQL12Ed5Qa+1EJtnLw0Qobq5Z9jdzSp0IDsaWLsU4wFiUYWD+cq19+853GdAmW9VcDjd3RmdgVa9wD+2UJQ42ySWx8T8eWiGcF1gQL5P79b363a79eKp8cYG84SnRfT5Ez/fxLMfqPUJh7jfeXJeJ3XYNhX2IC3U+yao9Jau05V23v44AGI+wz5FU9zNUuzAvevNY7/OWlxnAK1i7qR7thomdPfD3vEfLSJ0+0R4Y13uOamMT9bAE2ZQJG1BK5/8WF4vziR+XOLA5/AvVKgfe8o+NPdu0I7/9z5Lq9BAYkke6nZQTR03z5qTAmjsFgMBgMBoPBYDAYDAbDI4B9iWMwGAwGg8FgMBgMBoPB8AjwnnIq5/x7KjD8z1tFG+HN3qL+goI5HKHY089Ed7uZiW705aeimJVPRKfv+KIBLqaiZd1cia5/Cgqb1xSWm6Eo75R+95RKgeL98z/5atdmQV8woV2/J5rVHShiJxMUzkMxYw/3nqNwb1mI9uVnoqkFjSzEq/enAwh83w0buuHxkfqVxQtD0CtraBBIQUx6osolXRRBhGyjrvQsLGDaxXm6oPMPQcWLe6Bqj7eSnAR0vnKNAo+e+p7zowu62yR5mIrrIANIUcA4wjGkp66Weqa7udr9iY4PfFFb90f4b6Oua5c30qkAhdk8tEsURnNod0EhP+yDnoiCj70cfYEiY6ix6gJ0jBeoXVEKl+hvJyfb+VZDi5MMRYmtnCQkJ5D3PYc8ZgZJEeNIuoG0DTHoFMXPHenEkEbM3orGPHqGYtUoul02f1vvcUCLonDTyy09P12gwNuhnov09ulUVP5NhmKmDtRoUE/9CLIsHJ2VKHhaa011IaGZHGouFIhP6/m2XYPeezRQH3dR5NXHGuomipWDHmn9KOzLArE9raGzE82FMeZo0kUhUMipOogTYYQC8qCc7xWec/W9JKElNUBMIOcY8/MAhWq/gFTqLYofT1GUOoGMp4v5fwo5Z11qbueQDqwxHpumcHUE2roXUFoCmUGpOHd2pD5nPUzKlcsD9f/n2McHKHjuseDyXPfbwedlR2P9GjKh7mRLXfYDybY+FFVZulUjm+E6W6OQYQEJwuFYVO5np2qnkNJE2EcPn0oGs8F6urvSs0eQaieQ/3Ad1whAeUOfL1F4k1LVfqxzFCgUOsHnlzeik9PwYQRpIwsVU1aToIBjTWkbDQoS3U8n0ryoqo+kbawqV6239zuGrKKHuFFDZlNASpRjX5pDLvD6QjKo37zW3JgvIYvsKl9xofqF0v7jA82Hpyfq37iRP/ooJu87xMu+4vLnn3+2a3/99de79tkzydwiFEUOseYr5FoeJfo+JVeUTSEH4jGtfN9vPtsfPKe9o3XdVt6mdohj+lhDA+SonUx9MoCRQYQ8gHvFs0PJU4ulxn+BouA1jneNUUmd6V1hM1NeUbA+Aww/Tp+93LUvC8qdUDQ70dxaoCBv/0RxZwQ5Y7lhDghzBow/pVid6CP9lu9p/Dox8ww9Twd7eBcxrwNJSr5hfFO/nCBfGOA95WyssXuKQvPprcaxi9IWJd657l+1MhTc5bSLEYs5jkPGaxTc7cJw5wxlRH781a937STQfhC81H5ZQ5qeZTrG8/WsBe6hqPZv3JBnqTv/frvPepA3RxiHDO/Qgx5zV8jmYJbTG2t/CGu+a0LCC4mkS1U6wFV8h9bcvoTMbuJt59cm11q5xb8foagwtnQX9/R8q0qDfr3U9W/eooAyDAWeFTpnD2sxgMxv0tXcGeL5higG/c8pnWJMHIPBYDAYDAaDwWAwGAyGRwD7EsdgMBgMBoPBYDAYDAaD4RHgveRUzvN2Ve0DSiY8uqmI4lSTdemBvkfKo9Pfrq5FP9xMJY/ogYrXBw3uaKSK0HRoGMNxpZtsv6eK+7rmi2eitaWZaGHTuWhTMVxQhk9FVXeooO5QGfz0U1E0+z26g+j4FWnGuXhcmw3oYqBaDxu6pr/Hyv+e57mkoSr2IUHgNTzITnro76+//sWuffZMVfLPL0XJPgdt7RaVv1NQImFm5o4GuocTUFj7I9DvGmnVZKh/T1fqs5NTyS2Cjp7jGFS5EvQ4H89HJw2/hAyjoNxN1yrw+XIuCV9ewtUI1feL8uPQxj1X75zXXAnXLjhSlZRWZXABgyPXESSBNWSAUzjibHKN3eEA7juB2gnO2U9Ec/XgzlY39GYYLrmzzyQzeHENx7Mrrf96pbjwNNF1oPhxk+eSZR7gmv0atFhfayuERG5zC0kB5BAB6P/5zolrf5X/67J2WSNPKkGljiF3yDJ9vlyAVlzr8yM4+nx2qpjo/gTSRsTfGjF3jEr9h4iRCeQRFVzmimZedAMdG4JM3wMd9OefKM7eQQ7z6XO4R2EyJAi6B3ByGOEeu6CeeiHcQeCCE3TUpmyD598n6tq5e/M3D/OD0pcy1xoNIVXqIjycddV3nx+LVu3g5tcdioreg8NFAKp8ALlWRPe5hdZUlW3vrYBk0MGF4fBI49vpQgajo50fRg9+TjnuaKC/jUDbr2FFEaPPDo4Uy39IddZvsb8Pf/nn29v9q//L7QtlWbpV43K4gmNUAVkbnQyfHUvK8OJEe3/yUsc8f/nZrl3DtcbD/hD8TK5CCbTa3Y7yK8rWfLgMlul2HhWQfDEODyEr6ENWsUas+cv/4d/u2inyGeYAQ8gQx3BNSiCLzjZqV7HynDimQ6kCtre/MNpCGPjuaLidcwniWYh7LTE/6chWh5QBay08eQbp+LH2mflCc2O21Bhk2HNiyjyRJw9CHdPpN32Edb7BOB8dKZ/9y5//ya7dP9TemRVci5DiIAbRtMbH/uHeKdfX3/I8rcFr/nbfw3kvYaGEijIxr35YGpRgDA8oZTlFTgKZ5inWWRfScgpvM8jsNxvFBg+OOEWTX87gSEWp3gGksv2ezreGhC6Ew5iPHh3iuTu+4mYPDnIhzrPcwFkJuSElb8fItSuUHXDfXrj9wdvJ9lLKfnysp7HmthdJWt9NdMyLM0lVPoWE7BaOhTD/au0h3krjdQDnp4CvVHALu9/IKzicduEW2o81M46wFy/hmpq81LtRfwKJOvLlA8T90UAxiOtyQ5kX9vSyxPhif9qzQZxzzrkoDN2TRs6eehoTyowjlI9IIc/tQSo3wFrpwY2pwPOWOCaAE3EVqJ1leJ+eSx43QLzuRtsxpLS4y5gId+oY1+c1B3DT/hz76AbOY4s7OW6FkFZ5qGIRY3x62I8i7BE13q/4/v9TYUwcg8FgMBgMBoPBYDAYDIZHAPsSx2AwGAwGg8FgMBgMBoPhEeC95VR+o4UJQ1bAF5XJDyBfaLldiDKWt8iKoFL3QW2sRHn0QXnlHZfgB4agVMZ9Uri350+6oMr6OrbXQwVt0O0rcNOKCo4vcM8g/TJp8YNF3VpDDpGhQv1iifZK1D0f1El3LyPz91f736s9F5aNxAzVwH04MzhP909qvAfXl6MDPTtdvQ66Ov7qRvS4LFMfRqBv9ijDgVSii2rmSUMbDfF3XciADo9F2+v0dO+kDxe4fgVpUUQZma9rlms9x2yuc8JMxN0tJLPagH4ZgTbZKn++ZwTN2vEwPymhKnJ9voJMqI/xGkPOVoIiOeqD2o1K/T7WsQf6fwf01yFcIXw4OBSN/KKTwKXjpejp/yNo/t9/+xr3BeemLiQ3kH/RnaHD/sdYb1L1QQfU9QruLjmkVfUIlONg/99313Xt0ubZYkjAYtDAV2uM4QJzDHP70yeSc5yiTxwom2mJeFYiFkIW1Y8ejgcF3cQaSYwPcjYd+zo4R4yx6oKGvMGYlJi7XP8HXfV9CNmOgyShxHNUdFLBwqe86WM5xTnnuaqRG9AfghTvCnRyOv6RPt2B68TRSH03hlNMB06uAAAgAElEQVRQjHnuEDtr/B7jQy63huy0oGtVvb27utK41KDkRx311mQMqZr/sLwhwCSo0f8+Nmk/1THVBvOhwtyH09m3V3KrKw9Eo588/2x7zWh/8riqrHbuVFnK+akx6UAe8exEMoAJ5ZtweTw4hjy3hIQH5/Qg1QiRQ6zhwsc1UteUSmzXN2n0QUB5iO5lCMckD+MctP4YUkDKxhn66EJW4L4go/QxF7h2mV8xf9wnAt+5YX97/SimM5OeLcf6o0uTD3lmh/R8qI3GPvsfMZXOQhjHEHM0QyzPV5I2JtG2ryNIr+gwt17q2OPPJWn3IzjBYCwcJOIFnKRq3BfjRStG+g/Hy7p+OL919cf5HfjeJYv3Rqeq6B1SbrpxDbDnHI/hfNrX3uJFfIlgXNbHd3C2SnO802Cep02fo2KCi+HCFNElEfNmglj9/EwxzmEvqArNhRp7Zw2pVAZ5Kt2c0o3kGRxD5n15/fFy1PtrlnRqhNw+hCMQ43mM+NBHewKn1KdDyaaylFJQSLQxN3zUcZjNlOd1uBaadVxhDdUBJccal8mxxpfzqMK49GP18xDtsCeJWII4RSfQG7goL5nL4Z0yx7qvGulOvUdxYxx33GefbR3U3twql/j+lfbmGP26RKyqmIdBkjafwu0ZLk1QSrku3hHwauFilFZhhOoHPXy6nf80XTuD/HmN/beL0ifMW7JC91ij5EofsW+Akh8eZJksIxPgPSrE+zbnZYFyKjXytZ8KY+IYDAaDwWAwGAwGg8FgMDwC2Jc4BoPBYDAYDAaDwWAwGAyPAO/FafWcc34jK/AhL2g5G/FzUKh80C7na32eXsENJheVKAAXipXoSUWn40cXzgc9UHi98v6cqE7t+HeiYVGSVdak3EIeQtkUK7/DFSiDm0sOmuNqJprjaqXzLFJQ6iH7+s3F1v0ozUnS/zDUVeWyxoUrrNCv+cP9U8BFK8d9+KFoaBO4x5CmefZU7jQF+qcmnRzVu32PbgigzjbXCiH9iGLJCuIuXFowhp0I1c4xR6tclMUCNEw6la3RByEorAXGbToTlTEFjzagLG9/Q9eC55zzG2q7B6mGV5J6j0ruoOmVMFvro2L/EhTTEPTzBNRlVpD30L8hXG7iWH3dAeXYy7dtDzTnISq2f/ZCc38AeuTdtVzAIvCcKa0agHPJcS+6uN+p+qZFFadT3O10185fSOrl7eQI+xPl1M65vKG+DiBBIzXz9goOb6DjPz+SG9qTiQb0bKLx9MDrTBHPFpCLLJeaFz5oy5Q/ZqB/B40Uz0N8putQF7GgA5otDI7cuH64D/kp3TYcrl9WkKXgNwhSYYsKTgGQhlIKu2/cy7lggudSxDwaNVJCFcLlsQepWB9/0cGeSpedAv1LtYMPaUeF+Eaad9n8AV0mW33lOL6QTsNNLKQjFqQOFaUl5cMyK1L4M8T9N0vdw5tC63v4Qq5rk/5pc03I7D4QdVW59Wr9B/dJuU2IWNaD9CXkeILKHiUP71cF1mKJfsjhpFJCNsX15XuUC2+PiSCh8jtwFYMLFh18YriJ0NWshgCwoGwLe2GFe8/QrrnvoA8qj30JSUXIFbE/eL7nuo2kG9uAq3BPFOG17o+3FFE2wzZcxrB2HNcRjvGxjrMY+wzGrLp3v4RszqWSJWyu5NpTLBTbBl3IbxxlFRi7Vk7FNp6b7Yrx0n/w+JYXVb3/mOr5not3ORSdthDjkfNQquoolWD+h7kaOY0DnRXxuK5yjHNqd7jW+CrQyGBiD9JzyHAGzIkwA70KrpGQeBc4eYXr0310A4l8BvnlBns6JfVUTsaQsUwGlKh8HPSQFzz7VM5Mi5lyuwEkbwnySea0MUp3sESGH/NVFnsR4tsafUeXvwkkvPeupUzbK6xL7tdHY71zDrgvw6mO1+nDObeL96QAufZsAQdV5BIB1rFj/tQyltv/Wqzq2q0byVZF+SDGJEDsORriPRzvQT76bXiq3LXAC1KJryP4DpGgjEYBzVWKdoT3kujeQRtjf5BA+gbXr3FP+zJdpTfIkddrrSduWx3mPFijK8QalvHoQ3Y9Qj+FzAfw/vFTYUwcg8FgMBgMBoPBYDAYDIZHAPsSx2AwGAwGg8FgMBgMBoPhEeA93anaVFF9Dsoj/pkOJgEoqUFXlblvxBp1Cao0Dzpwfmq5Rul40pu7cDZp0UAbajSlJVUFKivcFtpyHlTyL+lIgKrloHEWcPBZQ36zmInmmC4pTdK9r1AcflHrb//+1a+3n8E950MR+IE7GG4rwrP/8hRyJ5AJS9ABowCUbDop4KvAHNX2Y8qfWn0Ilwb0Fen+EZwigua6dEMaDFTVvttV++ZWdEQ6piSgsoWx6JNZDLupuaidJarNU56TwsFptdS1MlD0SGPP0vevNv5TcT9fPVSob8mpWhI/9S0r2newbhwdmOiwgUrqpAI70IU5ZxI6roDimW+2fV2C5h7R4QrXPBxq/ONa45Wn6mfeboDrkPJIynsMmeUG0k0P193c3e7aKeQNEeLR3uB5Lmiq7YfQGxWIcZeQkq0x30YvRQ/to2I/6d45XKVKxKorUDbfnF/t2t2OJGzzmeb5bKE+H4y3jhAnZ3KG6IKnXc9U1f94onXZi3XuDtwBPWgYqoo0VM1dxiAc4kpwifMMFHI46/h03PlY2kbnnLfb60B9J1UY6y/GXA0K9V2co53CBSpTH/loUzoaUIwGh4sCm4uvj92O9Q+jSM+nPETNAHLKllHHO34CqmrKG3DRQvNoWerzv72WZPBvZrqHRQIHqBNJG/17B8Q9ujbWTsZLdBSKmLdQeoS9qkLM9TDmPiZrjN/LwLx2eQ6JCF0AkZdELVo6JXSNhA/n6EB61YcjToigSCl8S0NClxYkWj6XDaU3uEd+Tsk76fIpno8uWvuE55wLmv0naFHf/8lBDXJIJYoKeVb1jonuqU/pZuWwt1StaUmpHSRycG1brxpZNNZNiHHxCsXi9fTtrj05lly9hOSgpHjTp1ySTpyQq3gPS9vo8tMWu76rvR8EfuD6jUNjlmtMmHuzxAKfK6XTJHLasgOZSk/tnk93Pt3DHHIKD3KHERxpY06wZs5TBsTwlGBNdDGfSu6dWKMJYw3eD+at9YdcL4MLFSRUdNfjWuTQ9uP9yVLbqN39/Of72WdffrVrV5nejy6//Yddu4f3I0aKCJ3ahZw4Qp5HV9oVcl3Kz2L0ew9x0mtyJjpcUk6XQPFIh9wM+2iBMeK7aMC5A8l8gblWYeyqlisyZZwPx+x7SbO3xyVZVZVbNnn7GiU3KDdMMfcqvHP0EON7iJV872QplgA5dpw8vF9e3Sg/z0qNJ0NYd7jNNXsopXCHMZlfKUb0sObOhijLMaEjtp6jhGtcCRe42ULzeA3nwRy5K10L6QSZLZT/3P3337n3hTFxDAaDwWAwGAwGg8FgMBgeAexLHIPBYDAYDAaDwWAwGAyGR4D35rTeSzhqSo+84uFjQQ+lY0HYFRU/7J3t2tnN9w+eJ+iQBgqqGlxuYshlYLq0c1PwSBUGq4mUyxCUWNJZi1LHFBu1K1SeXkEKNk9Bf4Ukg65Va1CgM0gdap+02O3w1HukrAaB70bjpiI3GfhwNGj1FWi6nY4ocRElRqD7snp4ToebjNW+1SeLpahno6HkF/0DVQ0PG1peDqpeAZpkF2Mf3em+lnc6Nx2IenACqsHVqyARq2LIAGKNGx1nvA2elZR6j9K9/Unh2qhd3dAtfdAuPXBvq1LXLuCelUGSFEWQmYF6XYFOSup71FoXvK7uLMF5unD5SBqJWoi4UEKeloKGWGPNJVBwhZybLSopJjOlIPicbjoe1rHzQT/O4fIwF80xHjYxa49uHJ4XuDDc0sZ9SFZIT51OJU+iSoWy1gI08xXGdg2655sbPcv/+7e/2bV//82bXXs8lGvAeq17uLgUhf8+tn76yenus5fP5ZJyMJJsKjs7xeeTXXsylptWJ9JapNHCKle8SNeQYbRo0xqLtMK8h7yXNFvGoH3C93zXbejCdBWqsS7rltuPwDiWZZTbqh1tGHPQhhNRDWp5inPewrWkoFNGM8cYO+imwvVP+nbLiRIjRho13f8C7Af1Wu1Fpvl+EUo25T+T1PrzY7W7x9obOo2zhx/sWcrR0Lw7yAP6fUk5KaWdz9WvHDdK9iiJCvE5Y1LLJYjyR8q8Qfd3cOEr0+0xpVPuQfl2Eiu2jw8gN4czTU06O+KpD01nhnlTv0Oi2XI+g4RjtdGau0OfrTfYX/eO7XMwUnOuUAXjUXuBPSdday/yWq6cyD+71AVDWoVYzhjgeeqLlmS9WYsBZMg0JO1ChpXPXu/a6zutm2DyfNeufDmf+NiL/XaP4L7Ydmi/SzbF33737zLmB74bjbbrbr3R+ddrrb8y51qBuw/LF0Ba5Y0klagDSBzodgsdXIauWiPPmQy1v0VITPJGNspzsMxAivUcdx7Or4MEeRYklOsZ9i3sI475F9osC5HllNTrmAHufTRQ3+wTnpNak3FmiLE4OUWOcAh3qjmeB+dkyMdSbMmscuYFyCNy7EujgfKRDnLUtDkGU6qVZ/r+O6SHuLEI5QdCyKbCMXJtGtGtKH97OI7SLblEzOJ9Bs07mbdPPZVzzm/uaQ05GgelwE0skQes8SyjvtbN6aHyTOa9fkCZvc4zvb7etUtcGK/Tbr7Wuo+72+efruDyOlcsWKAswaCrz/0jOJb1lW9EfcXTHN8/5CtdkzEoQg5eO0jNMC4+1r2Xql+nv1Vu/lNhTByDwWAwGAwGg8FgMBgMhkcA+xLHYDAYDAaDwWAwGAwGg+ER4L3kVJ7znO9v6U90OCDDjJ+zqH9FyQ3pm5GoSmkJxybQxrux2jEo5F3QnBzovEFNCmnjVAA6aA16cEb3jA4qgIM21aIKk/IOKdYGtOEUTkWkxOVoZ6y0DyrgcCAZ0eXt9jz7JMdFUeTOnjYURtB3K1B2yxpV31G9u4SeY7GUzGO9EfU4W2mcM8hCVqD1z+c63sf54zFco3D+Rbq97nQt2vgGkrUCDgYe7NFY6X02hfNUqXkTxpordBzhwojpUIJ/ySCnKkDvraDnK6o9U/7vUUtO1bIvgOuLo1QCVOoUfdeBa1ACSni6gpsDHE9a6wK01QxrgRKgApTm9HZbWf4ac+HkRFKcGn24XuoeA9AQKaehW1oIWqaHyu+UeTFOBaRRo89KzKUcUr/qfk3XFP18KDzn6qg5LWLfSrFkuVA/dH2NTwYHovlGdMzNRg98fan7v4a0kHrTyQhrDusrg5zq7Fj0195gu3Yi8IFnN6K7Hg7g2ID7enUnR6y7qdoDULlraJ82lE6mpJAjjlPyQ1o61nQAZwHSyfcJ3/Ndv3H2Wmd65oCSCcy9kk4Jhcb6NtdY9wtQsiEp8NeITHRCgSxymeseZqnGvZ8o7g0ap4wAziRFRdkMHQrhqoP1R0lIAap6jbnpMp1zQ/mxB0r9l5/s2k/O1O7AzYPS6drbroPOnh2O7kM15VQxnKrWWB9vMeeXkC2OC8XTAnM4xdgyMSrQVylo2Ms51jTyleFYMoD7GHkHaVKEOH8FGnqBtTI4kJy914dsGfscHSSrAhTy9GFHzgLSKspB6dpx8Vb3Q2n5fuG5uolvLSc7/I8Hh5EA+UfAYyB/ocQ2hctR4EPy2DJ20xxw6JdyqfmzvlRfhM366o01FtSBlpDkz1/9dtemEuz4TxWjg67mSNhKP9Ah2FN/mvwCx/O333uHzH1mqbVkhnTlodtoHVAyA/norWTDNXJBSmYotMsQk7jOFpjDnDuUKvF+0kZ/5WFvpZzqFlKU5UIxf4KxmiCGd+kyGTzseMVchM5ddB9dQQLKsgxHh5I0D5KP405VO6nP6QDacmzCs7HMQTlTjpBhb8lz7QMVZHEV+mKDmMrYzOv2UYIh6eqc93/LnKPlYMxSHBXzCf1DH2M3SfiuofY8U45M6TTHkeUKNpTFcZ+mA/SuL/e3Fj3nXNTM+RhryG9JgtXeIHfdJCgl0UMcRG6ZjNX3zIu4Lu/uNIdfn2t9/+N3P+7aKfao6Wr7ThFC6vnmUq5WUzjQZpg3w4HGp4RTaBd73mCCfBUBuMYcqVpuqninwf5elMorOrAIzaeX7n1hTByDwWAwGAwGg8FgMBgMhkcA+xLHYDAYDAaDwWAwGAwGg+ER4D35yN5OzkCplBdS7oB/AN0wB610eiu3k7s3opX1PdFQhwGcDGpRtGIHqnPI6vC6bBjCzaHaUqS6Aapmg6pVgIoZwQUih5NJVZHKBpclSC/WcDagQ0lVgaIMKl4Kim4Iyn8Hlc3v6WX788NxrvacqxoqHCtnh6Cn+rXuodpoPNNSY7XKRGtbgEKXw6VrAyrnElTi1VLH93qSc1y8vtq1f/xBLgz3lMEDyDpYrZ30uLrUuTt4JkpmKtANO6Volc4nbZIuA5hcCaV4cJ9pKZroWvVxXDg8V7u4mXN1BQejQv3sFXRBgFTFwakLTj5JV+tvs8B5Klbnh1uZp3mSYf5f32o8Lt/I/ahXb8cvhZxmBkZqCHewGjK0HPcbdrTmqabwKaGihqqiFxDpt+TFqknHsWohCnS+2fZxvUc5VV07lzfzLMd8S0EBTjPKLiFTmKuPYx/zDVTV62utpx6kKV9+IUeIT4uHHalWaw1MAhrwqHGW8OnCxPnU1fjEGKslZJY3P/6wa/ch54g6jOek1lLa8/AaJRW719OzxmiTFr9XeLWrG0c6SmMj7H8RNqgQ972EPOwCMXKSaS0OUshlEFODAvTzDFRnyF4TWHh0sdbvd7oQMucZYlUILYAHmWONc2T02aJDE+j8a+ydS9L/E+21R8+e7drVQPuBhwWeBezL7fm9Pe6Mnue5uHHPC/CMdJ1cQ2JxeaP1dw158OEI+wmcSehq5yF5KtE/NVzy1rmue3spivVgrGvdpy6UgA6wDgZ97OOg5i+u4DYHxx/G/xz2SNzHNxncBHHdLH/YFW8KSeolJM21t8+spg2/OTdp/nQP8elwijlMd64O7i9F39Ub5nbML5D3RspBKIlYTpH33p3v2ofD7ZxJOpo7KebdEhKhoqd1k0FOPPlMOUDck1SG+g+vfle6X7+j/S+D2tUubWQtLec3zMmKjjhIvu4QKymxqSrN2yqFfAax8g7S7yvK7yHV8AvWjtB5lo08Jltr7OljxBSyQK6yXGhOLFGS4fBAso2C5RnwXpIj51ph7s6xzy3pMgrpfIxcawGXnX3jfntLsJ/QPTKI+L4DJzJIgpdYi8sNJPTYXwO8yywxZ+bI4bg3+4jNHu7hXrbEucM8Y4XYljjFWkqicpYUgctYd46YMlNcLFKt9QrvL3Q6LLB/VMjlWTqg37yH+3tUNnq1c53mlvpw4+N7QIG4ktIpDvviXaQYFsDNmq6dDlJtD5LmpKccsXLYC5HTJrifH5v9jTK80NNaDTHedFDeoITLPNCY/HChd5ijU8gQEYs3WKMp5M0e5lGB7w5a5V8gu3ZLvRv9VBgTx2AwGAwGg8FgMBgMBoPhEcC+xDEYDAaDwWAwGAwGg8FgeAR4P3cqr3a+lzZtULo8VjdHdXZIq0JQdf2N6FH946c65pwSGtHgMlZbBy2/9kXR6g3HuzZNSO4ZTKu5PpzDMYEykx6qU+egzeaodk76VQ4aVFnofuuKlFtU34aMJwMtme4/QSiK2JOX/8Y559yvLv5Pty8UReGubrZU0c5AYxKiWnsIChgdZtYb9fdyrfZqxf6BIweorXTZoTNCBFrjfEmqOKizjbRmTbcgUAbp6kGWdgi6cQWK6XoNii4o+50Ex9esAK85kiSUX+E70JoSPVAiIQXcJ7y6dl7jrFZhfpagENLhIoSrG6VEpHt36FIAKqcHGrNPyiOev4t+KUAnPZ+J5t0bbu/h2fMnuEdqotT/iQ/XnFJzxyNHWYc7H7IR0mZJu27RTGntgblRgRZbkLrczLF6j5Kc2jmXN/IgUjkd5FF0cckhsbi6Ee2yAIU7xTrjvDg+EA10DAljpyNKcNJV/PMhVSWVt2jWxQYuTDkclio68kC+FnR1nfWd5sTVa0mrKCUNAsxXLna0o5Y7EZzlIkhvKtCvw/26Gd2jritXNPE/4JoDJZhTL8Da8mNIq0DJXaeQ7fbhtgVZQFLBvQnTpwvXjqMR9h/IYm5vt84OpCqX6J8wgcMOXbYwH0s8R035MdYQXcZSSE4qxIuoq/lYBZSL0IkM7Xsq8j6VH57bBRcucbrAbeDAR7nfLajx8wkchjqMp2q2HDyx//XgTPn8OeRxgSQXlFf2mzU1hGMG5z6dF3NIGHLsywvIdgaxxoSy+DX2CFLUNxjzVcq25vEN1vp6zZxH8WCf8Dzn/GbekPLPxKBuyWwc2g/HdlLiO5Cw1C0HK+QLERwOW8pe7KnITe4at6I0k1Tgjm6CcN77/OWXu/aTX/77XTuZSBbbjjWQfLU2z3ctnvddVPU/+e+Ho65qlzVzNIXEiHKqgmsU+fYCe+Ec7wobzE8PwTIvGFvVjkPNzzd3klZdz7/ZtQ+nyoHDJuYOEUMHeJ/oD7W2A+xbdGq9vJLErkh17g4SF76XUGZMB6WW5BvDcofc4K9/+92u/f0bvXftE7VT2QOmWy1nNMhkKXPhM1CquVrjb7Fe6WzkY54HiN+ztXKWH240vtFUTnHzt9u8KsaaS/Be0MM4jo/l8sfYdnslGXv29mGpa4X3JLoe0gkygIyL7m90k6OjZxhvz8+8/ENRpKW7+nbbJ5c+5LOQDDnk3n4Hkji8O06nlP9jbLFXBLH6pz9WvtrpKyf48hc/27VffoV3OsTl128vnHPO3cHh+GCg8/l4h+/hXkJI3zL0fYgc+Yfv5Yh1eKh9t8I5l5DH1RiLJNB5fOSGKWJWhPP8VBgTx2AwGAwGg8FgMBgMBoPhEcC+xDEYDAaDwWAwGAwGg8FgeAR4L3557TmXNw4OIWhcPuQmpHJ6qBofeqzaLirWeCTK4eoK1b5RBR5sSZeloPHfgp4bg64F6cjscktF9kCn7A1Emxoeih7nryDnAfU1y0hPBLWRdFpUsKcUIAX1bYPK4wVofsO+qOsHE1HLv/qP/9k559x/+83fu31hvdm4v/mHf3TOORcNRAf8kz/9xa79yTEq4xfqk8Va7eVSkrTNEnIq0PpXoC9moF73+hrzMNJYdEinPNH4zxpJzh1cjUpPF/JA53OgZ5JqWGEcyhYNGjIoH1IhSAWyAjK4VMf3DyGtopyK1cZpfbRH1K5yddnQ9jCXKrr3oG/9dzg/UR7kRZAhQcqwgYNKVKFP0e0lHaxAFSdtcDrbrtdBX3TD8VhzMIzU/xHosTVcAGgDhtt1Hdxv6ZOuC0knJAIRHWIwvnRx4RotGslQXe+PNl5VlVs1tNSoCyp3D/MQ8zmFe9QMDm8FpFj5SvTULuRxG8Sw8gZuCJXanUjr1YcsA6YROwlFSglf+bBjSwhJFinkXoiYn4nyOkd86YR6vjDQc3TgIBhQH8e5AIkKpap7tW0A6rrauZ9EkfYTD/ONs8Z/xzrLQb2dgZ59BAe9kg4tkPd4GCQoZ1wMxxuHvotG2z2HVOEYEpcc41XgOQrQvWu4WZW12nXLzRFz09O4uwGoyKBRl+wzOssxB2io9nt1bay9XQzjNElBey+wb9Dhku5Udys9lxdrTArcLV3YIsQeupZ1EENfnkp+yvEvm7W2wXgzb2FcoDwhgLQoxMOuIZGkm1aJ3/pyjH8BZx86OC3g4nX+VlIw7rtB5+P9fhg0sb0tp+IRcHqhwySek3tb7Sjb0Od0La1rSqspj9ea68WaG52J5vz0divXuV1oHiVDyaOefvnVrv3VX/wnHXP2ue4lQrmC1t5JUIaBj7mnea1/eLBdP/D5PtdiVVc7uQbnG3MVyqkKSO8XiJtXkHIvNiq3gFTdFZTQIf/sjjRWX0A2cQfpZIj5cnK0dXycHJzsPmvlJKnua41SCpSfzJfauzdvlV8PIXP0ajop0cFRc9HnaGDdv7rQWvzhQpKfu/VHylHr2qXNCwGdGikH8ujgiLXCfIFzcgNpFXN9yo3oLhhA3hYhv3x7Lukayy5MBhPnnHMHcLVkyYfBWO9nNZzfQuRa8VBrcT7XO0uK96QIOQ1jJ8s4hJSaUVaP4Y1QdqDXlCwJfMomPxC151y+7Z/78g3OOVetcZ89lJuI6R6rdTl3mvOdFBKjCmuOudBCa7fTxTs6JMo9PHuMUgzd3vPt9aNP9XfIHzZYwxu886/poA133gjSuhz5ycVbfefQxTEZ8iLfV5wPKH+8RfmVOZyq8vfPUY2JYzAYDAaDwWAwGAwGg8HwCGBf4hgMBoPBYDAYDAaDwWAwPAK8nzuV81zUSKdCuMqEkFPxhD553aCYdhJRG0PQDG9X4jzR1WKdiVrot1wwIHMBzdiDXMRveJcHcEdJIcm4XonuxurUdNyig0+/BwcdUILDjihfNSnKkB3Nl6JQBei/CZy74kjXOj/fVsLP6Tj0gcjywv3YUCnnP4pKNjl5sWs/g6SrgPPMBmNF96galcE3qDa+WomS5oHuRoqZj8/pTtMPQHluJEkeaL8ZpGkl3FgqfC+5gRtKmtJhSX0cY+7m0IL5NR05dN0NJC0HHY0bHclySK587+NIOFxdubrY9m/dmh+gckNCFVKy0HLnAC0S0hMP8zyFy0MJGY2HObCcQtqY0vEKNMeGZvp2DonFkf7u4HSya/cn6NsIVG5+7Qy6fBWQOv+wU1XLFQaU0xJjlCHuJKDRFw1FktT6D0VZVu52ul2DYQynLzjteaBpVniWDGuhRxlgqGNSjPM35xe79t0dHGn6oop/8snLXbuDPkkgiZvPtvHs9lpxc7FQXLuFG00Jzn6caP51u6JN+6Dc9nsa8w5o0JQI0nkhRLzgPItacirIgiD/2Cc8pynXCR+OJ6S1x5A1TSFzW3h6hgqpv7cAABR0SURBVBno8XT+GXThoADnFq/g/Ncz15Bi+VwL3na/qkrdS+hpjGo6eXmUn0A2hbjvMV57XDeQC0BqVgwlO8hrOKFBlkn3nwI6lnvniP2txC3FP2jkFB7lM2s4DSHeXVzJ0eT8Ru41X7483bVjuIHRSQ0sfJdvHpbEQZ3mCsjmCsyLulmjFdZ/hj5jLONF4wTyAMT5PtZlniPm0/mGLk9YiwXixe1c8eDqVnGCkoeWHmaP8DztuS3ZBmjwFeVULRdK4B2SXB8xrW49Au3HIOeIuc/ArQ7rq9vcm5dqrR4caC8cHcrBNe6f4b60T9TIvCuuV+QDdO7kWLSMg9y78hUe9XHGjpe6l0i1pO+MB9C7URK1gGvON2+0531+gv6EXLlGLuhBzhFBWtXDWD2NJf9O+ur/uHF8rNbIP+E2WMAFbg23zTWlypgfd1NI61B+IkL8b+XdkJ1R/pihnxaIRyvEnc1qf+8XRFXXbt3kggFKHlACVCH/xiO08vWjsfaNELlgB25P8xkk4nDf62JPHXXggoi8ykfuMB5v58l4oGNDHBsg76J0nXOQrpQxchqWo+hQz8i8NKQUCms3paQTsl7IrM6v3zjn2hLgD4XnnAub9+mBh/fUUu5pmJ4ugtvbAO7HdJhlKYwlZf4pSwTgHQJzgZK7kKUS6CbWSBQpVWautIJs9XIO5zm859xONZ96iAufnknm2kPZlrzU8dS7eZA3B0s993ijuXOQYW38M1JUY+IYDAaDwWAwGAwGg8FgMDwC2Jc4BoPBYDAYDAaDwWAwGAyPAO8lp3JObgYh6LkB5Cw+nYLgwlGD1pbEoiHOblUx/eZO0qNuAXocpC0+aFGRx+rUah8eivJ42EgTyg1crUB9ykrS40BJBGUthdtCCYpkdCyaXwUnliXobDegFq+WomtNerrHcVc02tVMNO2b+e+318z1zB8K3/dd3N/e9w/fS/owW8CJB5QukmhLUKBr0DSpGKqc+pbneftKY3v+Ss8ze67+7A81hiXOv1411NoHqOTOOVdBBgC1k/vuu1c6ptQ4vHyp/oaSzYU+3axAWQSdryU/wlegpNpXoLH7wXsvsZ+GunJ1sfmDe/IwDzuQUIVwRMnpcEMbDtAGE1BMC9AMPUhE1jP16d1bzaUIcg46XwTJln6Y3ul8d3PNl9evXu/akyOtraPnx7v2CJIrvwuqOk2IWk4amLN04QBdPsdaX2Gt9eCOcy9p3Kc7VV4UO1lGMtRYPYGr2BBuCJs1nRn0wAdHiiWxp+OnU43J775/s2sXWC//4T/9x137z/78T3WeDlxz0J8XP/7onHPu9q3i1MWFaOvffPvdrv32StX7L99e7toVqvd/9dUnu/aXn0vOlUFCcod9gTHIB/WYsqmWUQwlceRr7xGe57tOI630sYbSpeZ2AheoCnvh9+ijI0jOJpDXXa3hVDKhg6PWUQZKfJHr/HEIGRtkXLuQhvutsKdnfbjzVVj/pe6ll3A9YR8FxbvItY7vDiQ1Crtaxx6cKD3sx3SRySrSyZvGHi1xPM+5qHFk26TcC+HIgTxkjZu7mIlanmO+UbZDtzcyps+vQOe+1nm6PckcRxNQuI8lQ4ua+ULph4f8yAfdf3aL61xo7fa7+tvPPnu2a4fYGCtK2EH9plvQcq1jvnulfXdGCTnm4nCoebFP1LXnqob+TykzTYhK7klhqwCAzlNxD4F7jI/R8yhhfHjPL+H+V1Aux99P/e3f0uH07VuNV9VRvjy5VUw/Gmm8uM8yN6pbLrKUQVEiiXnKR/pISvA/hrquXdnkUwX2YOaEJeIEyydskIf9gNj65lYSil5yuGtnkDu8efPjrr2609qJ0W097DN9OBh1GwedGrlfhf7OEDuWkPXOEV8Yi7/4THshz7kqkANgKtIRNMc/3GHtLuHsu0ZphGCPOQ1R185lTcmLlPMTkihKf/m+MJvr/p4eKL+JIAU9GqLsw1T98uvffLtrV0gGxhOtoxGkWCcjSKey7TxZz1hOQWNEZzG6EFFOlbHkANRRh4fa82LILH1Ku5DXbTjffY3jmvnYWvHlfHrb3Nf+ZOOFc+6mud41JEA3eA9O4LBWYg5T7hZANhUj/1nCWvEG7xOXiH8Z5UaYLwO834yxn8T394N4m0NmeTdTXLhZoY3n2yAWewNdZ/Cl1uUYrlmvLzQOlH3SnfPMKf86wQtLjPWde++/Fo2JYzAYDAaDwWAwGAwGg8HwCGBf4hgMBoPBYDAYDAaDwWAwPAK8pzuVc37DD2O1f1bvJyWQn9NVxKEieQpK+GYhapNfijba7Yvae3YsSna5Bs01E4WsB2lVt7lfH1XC6xqynRJOSaA8D8eivq3guDRFNWsPFcYXoIW9uhJ18/ZGNK4ENLLjsa7bJy0d1N1esaVX+25/9Dg/CFy/kWgUpZ7l/I3kFtmfPtcf0IUsgGOXD/ow3WByUcaSRP05Gol6lm1QYT5QPw97kKeBcpo0JbvpfBW0KMuYWwvR2khZHI1ESR+PQT9vuUpoPOmsM4dGyw8o4VCb8o/AZ6V0jfNeUTvnNbS9GmsxgCTDB92QlPsgIFUc6xV9kdDNrSOaY4Wq8QH41hGc3Qp8NzyHg9v9sjvEvcSQuORvtf5ncGfwsG76AzjbQbrpRZRNYb2AWkr6P+mvc7ioZe9Ya9U913ePzOOqknvDcgm3NdBoY7hRBKCtjiaSWyQ45uWxpGcvTxUrN5CKXi0gG4U86Xfffb9rrzeSzVSgBy8a96kNJHbrtdp+X2v4Cdbz5FAU9jpTf//y51/q+GOt0esbudpsQIWtIdHtYV/g2s0g+QPT3i1wn/uE5/subNZ5Crp1QVklXSqwht7MdEz0Ws98NpZs5iiDyx4cF3ys3RVi1GoJNwfswX1IAe4p97WnmEq3hWqtY6MQEsNY916UkKtAQjXFXnyFWLBgPIL7kSsktfNAf6+xdnM4St7LYapqf/K4uq5d0czzHJLKFHuO13J90fO+udY+ush1TxPEigKU+RiugQHk1hFi7gHW7vNPPtu1BwdaI64Zz5xUfjjFrW8kz8oga3JYl+ODvk7XV7tCHKwgJ6CpSg558/m5xvANnPBCSEMPIAWbjCSR2C9qyYuxL/rYkwNoHIKW3Bn7P/YlyoqYx3KuUrZRwSlug7wgw7xaI6bljbaxN9TYrtDn35+r5MBXyD/ogsVxaTlotewckevwmeg49s797Y9tfPvbGOu6dnnT/wXWOF23/olmdgcYw7hL7OtXkPt9PdD66yPmXV7rneP7y/NdezBSzvGLr77YtbsHWq+9/rbdYY6MzWeJ/Te91HUGKLFwBEfOl2faL9NCa3e+gEMs3Hn5PnYJRzi+uwzGWt+HyKmORrrWt3/1f7t9oaqcSxtXpbDmeyFKEkDmH0Luy+ecLfQMQ7hphojHLydybXvlS5b/3Z36ov9cLm95rHG6RX5zr0Dt5H8opXfOuRplFkrMzTWcv1LIco6OFecGQ917jHIFOcaI7xQr5FiU/OfID2/ulCffb7v7VMfNssz9lx9+cM45d41339EEjm1wGF5XWmd1IJkaSzu4Cu+IPToWqj8PDjQnZ1Ploh7y1S7mzriva/WbfYzZAV8JBiONQ3+pvbA/17oM4DA2HmpdjmLFDg9jtURu2XLSxHzpoB2xXAviePHPoNUYE8dgMBgMBoPBYDAYDAaD4RHAvsQxGAwGg8FgMBgMBoPBYHgEeG/rnHv3CJ/0TdAGgyrE56Bkk9zkiXoUlnJKODuC3Kgr6u0MEhkPlKQxJTqgk9OpJOhsaVEJaFAFJFwVqlAP4BCT9CHXAVWxWpDCrkd6c6t7/PGtKHwRKmQ/PxKd8QiSnrhLSwA1b6+bvqn3Rxv3A9/1h9u+GI/U37c3klNdw8nioAe6Y0CKve65hKtTHOsZO5BZffISrhqh6GxeneB40ezAwt5JO6ZzUQdrVAAn/b0Etfzrr5/s2qORxr/G+OegPlaOsgWdZwapSxzrHilnyEm/xnN/NHcq55x3T6sE9S8I9Zx+hL4FDbuggw4o8ZQehV09pwcHkxzOLZjabjyAdGYgKmQI948w2vZvD5+Nemp3D0Vb7vZ0zbNPNY4+JEVFTocFjVeAeFSnkBrA3ekWlOO3V4pBo2PdO+UTO6ucPXJVa+dc0VSpv7sTZXq51H1SBnByqjV0eCTqfQUJT1Vqrp4eSFr1F//mz3ftv/uN3GP+j//yv+/ab+/gCFBiXoDGnm0aJxU6vGBudTFvPnv5Ytf+5S9+vmu/eKL7OoSsNF1rfVdYT5QnRonOH0AqRAcJuoIUK/Xr1Y2eb5+o69rlTX8t4UgVgGaMcOWmCz3nHNTo76Yax1+/lhTmSaSxvltoXEKshQh9FHe0dkLEgGQAF46m7yJM8SEkX3lI6QXcIUD93cApJV1BTrzRPf7NRhLJpKO2W/9m1yzg+OOBT+ylkFpWjNPNZxv19YejdnXjhOPjeUPMbQ8y7QyuL6/hMHV5q3E7gwxjAznHAK4WTyaSE/fhWBHAZaqGE9wGeU7Z7D8B0oM15tZyJRp6BHnU5KniSH9Eya2edY1YWWBfoPPG3VTPOoX8kXHzAM4yk7H6o/pITnF1Xbuy/MNzv0v+X8H9iKGdxlMhcp0AeUGFDbCA88h8o+P/8XdyPLpFHjuAs86XX/zMOefc4ZnkHiOs4dFTOKKgnACTxRpzlq5oAd3VkFu+axdjP+1VO/yeuB+Xmk6TGCAPeZ5fYRwQB6+xj76CtHCGuffyQOPwAvl5gTwgzyFjLjQuE7jsxo3DbAdddnOpuJDDvdObKc6PsJ+9eCJJUBRjbBGk4zVcjeBINVvq/NM7xdkBZO5HX0i6fFPquoOxnvW/7lFOVVe1WzdxL8G9FsgtIjxbF25fa8S/c+Rnk4mOucZ6GnX1TnH0ifpx/ivFw/Bc5xk+01qbnGl/HR1u2z7GJUUczeBEmC30+QBukpNnynsOn6hvS+TLG0i+W8pArNEUewzlnTnkyis4Ot07pPl7tJVbl6X7u8bZLYU09Avk7Q7mydkc9zxS7GFemOFZxmPtDwGlfwOtxZMj7ZEZ3sUClGXp4D31/h0hQrxI0fcpvvZY4N4rvIdT2j/GdwclnKqvka9OIVf2MNdDyr9w/gLv9Bx/xrKfCmPiGAwGg8FgMBgMBoPBYDA8AtiXOAaDwWAwGAwGg8FgMBgMjwDvpfWonXNF0FC5Hjakch4qdndAJUqcqEd+JZrhL471+b/9X//9rj2AI8Zf/+2vd+3Xv/tu145HomiRcvjjW1WW/+03v3fOOTcegzb5XJTUQ9DzIlTKzpzoUatCtL0bVLy/udU93t6pnRVqHx2ICnZ8oGv1EtEy/VDPUVSia502FcCj4P0pVu+C5zkXdrbPeTBG1W3IYW6moh2OE9EOg5bbg6hqBeQLfsiK+ajG3dW8iENQyUDVXpOedqd+vpdLsdJ+DmpfnoJiB/eaCeR2CZyqWlRGVOmHgMTN5qBQ5qTEPUyp70AGEIEK7cGtYO9ouJchJV6xxtSHnCOh9KSkDAxOAaCW+4gMHVByF6HWRYGxzpeQM6LvnsCFZND0xRDucBGvWas/R1grfkefU8IWgpZJ2mIGiUqK+7p4I9eUH378QfcOqecRdHyYvtLD7JFhXtfOZY1Mi3MMS8KFcPLqJhrbDLK2AP0wn0nWcIoK/4eIQ5+e6QI3rxWLh4eiIY9wPOdX3rh0LVdaqw6x4Ag09OcvnuJzUc8nIz1HF7Tx2Vw0czq5UE5FedkG677GPPIgrbq9E/15DvrzPlHVlds0kjZKOeKunnONeDVfIrZgLazhsvcdJLmvITM+SdVfXcyZxNcY9bBeaw/9GKidDLb3FmE9VYXuq0R/JnBP9ECFLwods4HM6wZyqvRQTodnB2pnCIuFD/kgni9HkuGn2CemTd9UDzvJ/XNx72oUIQZAYeZWa7r66Rnv4KTy4yvFmD95LhnMZgV5Ul/zk/1/ALe1bA1Xo5XmcEpXr+Y+w0ix0nUwPyCtHeCYEI6BUEK7CmObZbhf9HOW0QVN9+LhObpw30qRMyyWyAE/0u+Hnuc5v0lIqQyqkaNQlkPpgfeONhXRJZx1mLtA3ePyDK6lkD4UTpO+8rRel83cTiBVe/GpXJBe/vzPdu0ohoMn+pASKs9DLIRUkfrnlnQMz0pnx3qfNjfvCf9e7ol7K3hvKINA8QjdNld0T7tSPL18pc9PhhqTGrL0Xk/js7rQWpj/o+Rx3oXi5bBxXyzwd9mVZE0VpIeo9uAOnku24fU0PpS+eUhMmYMvsC6v4exbIZ95eYIyBj3swShX0TmUvHmfqOp65xRZUnqI+47glEq3wRzOjpRWLSHzXGD+Fxg770jx5+BUz5x+q9h880Z5T3Uquf6mcf/rovzCYiWp2gbxd3wkCdeTl8927eGRrukhHmd0roSbUYwJkeJZGWu5Epd4B51Dvn14ut1f/T2+L9bOc6W37dscucpiCXelnsZhBtnuDPtDv6tnpHwoR58kMdwzkVtwj3IJSl4wp0r1XnLvvBnjnYfurCnqoMSQbB/B4Woc8T1Pc24JCff5Uu/JKfaXCHnZuEC5iCHyMuSoEdys/Or990Vj4hgMBoPBYDAYDAaDwWAwPALYlzgGg8FgMBgMBoPBYDAYDI8A3vtQJj3Pe+uc++6PHmjYNz6t6/rkjx/2x2Fj+C8KG8fHDxvDfx2wcXz8sDH81wEbx8cPG8N/HbBxfPywMfzXgZ80ju/1JY7BYDAYDAaDwWAwGAwGg+FfBianMhgMBoPBYDAYDAaDwWB4BLAvcQwGg8FgMBgMBoPBYDAYHgHsSxyDwWAwGAwGg8FgMBgMhkcA+xLHYDAYDAaDwWAwGAwGg+ERwL7EMRgMBoPBYDAYDAaDwWB4BLAvcQwGg8FgMBgMBoPBYDAYHgHsSxyDwWAwGAwGg8FgMBgMhkcA+xLHYDAYDAaDwWAwGAwGg+ERwL7EMRgMBoPBYDAYDAaDwWB4BPj/AM5/dcfA2K6KAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa9012f9e48>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper display function\n",
    "def imshow(img):\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# obtain one batch of training images\n",
    "dataiter = iter(celeba_train_loader)\n",
    "images, _ = dataiter.next() # _ for no labels\n",
    "\n",
    "# plot the images in the batch, along with the corresponding labels\n",
    "fig = plt.figure(figsize=(20, 4))\n",
    "plot_size=20\n",
    "for idx in np.arange(plot_size):\n",
    "    ax = fig.add_subplot(2, plot_size/2, idx+1, xticks=[], yticks=[])\n",
    "    imshow(images[idx])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Pre-process your image data and scale it to a pixel range of -1 to 1\n",
    "\n",
    "You need to do a bit of pre-processing; you know that the output of a `tanh` activated generator will contain pixel values in a range from -1 to 1, and so, we need to rescale our training images to a range of -1 to 1. (Right now, they are in a range from 0-1.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Complete the scale function\n",
    "def scale(x, feature_range=(-1, 1)):\n",
    "    ''' Scale takes in an image x and returns that image, scaled\n",
    "       with a feature_range of pixel values from -1 to 1. \n",
    "       This function assumes that the input x is already scaled from 0-1.'''\n",
    "    # assume x is scaled to (0, 1)\n",
    "    # scale to feature_range and return scaled x\n",
    "    min, max = feature_range\n",
    "    x = x * (max-min) + min\n",
    "    return x\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.9922)\n",
      "Max:  tensor(0.9608)\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# check scaled range\n",
    "# should be close to -1 to 1\n",
    "img = images[0]\n",
    "scaled_img = scale(img)\n",
    "\n",
    "print('Min: ', scaled_img.min())\n",
    "print('Max: ', scaled_img.max())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Define the Model\n",
    "\n",
    "A GAN is comprised of two adversarial networks, a discriminator and a generator.\n",
    "\n",
    "## Discriminator\n",
    "\n",
    "Your first task will be to define the discriminator. This is a convolutional classifier like you've built before, only without any maxpooling layers. To deal with this complex data, it's suggested you use a deep network with **normalization**. You are also allowed to create any helper functions that may be useful.\n",
    "\n",
    "#### Exercise: Complete the Discriminator class\n",
    "* The inputs to the discriminator are 32x32x3 tensor images\n",
    "* The output should be a single value that will indicate whether a given image is real or fake\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create helper function to do batch normalization for both conv layers and deonv layers\n",
    "def conv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
    "    \"\"\"Creates a convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    layers = []\n",
    "    conv_layer = nn.Conv2d(in_channels=in_channels, out_channels=out_channels, \n",
    "                           kernel_size=kernel_size, stride=stride, padding=padding, bias=False)\n",
    "    \n",
    "    layers.append(conv_layer)\n",
    "\n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "    return nn.Sequential(*layers)\n",
    "\n",
    "\n",
    "def deconv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
    "    \"\"\"Creates a transpose convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    layers = []\n",
    "    # append transpose conv layer\n",
    "    layers.append(nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, padding, bias=False))\n",
    "    # optional batch norm layer\n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "    return nn.Sequential(*layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n",
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Discriminator(nn.Module):\n",
    "\n",
    "    def __init__(self, conv_dim):\n",
    "        \"\"\"\n",
    "        Initialize the Discriminator Module\n",
    "        :param conv_dim: The depth of the first convolutional layer\n",
    "        \"\"\"\n",
    "        super(Discriminator, self).__init__()\n",
    "\n",
    "        # complete init function\n",
    "        # Define all convolutional layers\n",
    "        # Should accept an RGB image as input and output a single value\n",
    "\n",
    "        # Convolutional layers, increasing in depth\n",
    "        # first layer has *no* batchnorm\n",
    "        self.conv_dim = conv_dim\n",
    "        \n",
    "        self.conv1 = conv(3, conv_dim, 4, batch_norm=False) # x, y = 32, depth 32\n",
    "        self.conv2 = conv(conv_dim, conv_dim*2, 4) # (16, 16, 64)\n",
    "        self.conv3 = conv(conv_dim*2, conv_dim*4, 4) # (8, 8, 128)\n",
    "        \n",
    "        # FC layer, real or fake\n",
    "        self.fc = nn.Linear(conv_dim*4*4*4, 1)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: Discriminator logits; the output of the neural network\n",
    "        \"\"\"\n",
    "        # define feedforward behavior\n",
    "        # all hidden layers + leaky relu activation\n",
    "        out = F.leaky_relu(self.conv1(x), 0.2)\n",
    "        out = F.leaky_relu(self.conv2(out), 0.2)\n",
    "        out = F.leaky_relu(self.conv3(out), 0.2)\n",
    "        \n",
    "        # flatten\n",
    "        out = out.view(-1, self.conv_dim*4*4*4)\n",
    "        \n",
    "        # final output layer\n",
    "        out = self.fc(out)   \n",
    "        \n",
    "        return out\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_discriminator(Discriminator)\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_discriminator(Discriminator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generator\n",
    "\n",
    "The generator should upsample an input and generate a *new* image of the same size as our training data `32x32x3`. This should be mostly transpose convolutional layers with normalization applied to the outputs.\n",
    "\n",
    "#### Exercise: Complete the Generator class\n",
    "* The inputs to the generator are vectors of some length `z_size`\n",
    "* The output should be a image of shape `32x32x3`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Generator(nn.Module):\n",
    "    \n",
    "    def __init__(self, z_size, conv_dim):\n",
    "        \"\"\"\n",
    "        Initialize the Generator Module\n",
    "        :param z_size: The length of the input latent vector, z\n",
    "        :param conv_dim: The depth of the inputs to the *last* transpose convolutional layer\n",
    "        \"\"\"\n",
    "        super(Generator, self).__init__()\n",
    "\n",
    "        # complete init function\n",
    "        # complete init function\n",
    "        self.conv_dim = conv_dim\n",
    "        \n",
    "        # first, fully-connected layer\n",
    "        self.fc = nn.Linear(z_size, conv_dim*4*4*4)\n",
    "        \n",
    "        # transpose conv layers\n",
    "        self.t_conv1 = deconv(conv_dim*4, conv_dim*2, 4)\n",
    "        self.t_conv2 = deconv(conv_dim*2, conv_dim, 4)\n",
    "        self.t_conv3 = deconv(conv_dim, 3, 4, batch_norm=False)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: A 32x32x3 Tensor image as output\n",
    "        \"\"\"\n",
    "        # define feedforward behavior\n",
    "        # fully-connected + reshape \n",
    "        out = self.fc(x)\n",
    "        out = out.view(-1, self.conv_dim*4, 4, 4) # (batch_size, depth, 4, 4)\n",
    "        \n",
    "        # hidden transpose conv layers + relu\n",
    "        out = F.relu(self.t_conv1(out))\n",
    "        out = F.relu(self.t_conv2(out))\n",
    "        \n",
    "        # last layer + tanh activation\n",
    "        out = self.t_conv3(out)\n",
    "        out = F.tanh(out)\n",
    "        \n",
    "        return out\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_generator(Generator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize the weights of your networks\n",
    "\n",
    "To help your models converge, you should initialize the weights of the convolutional and linear layers in your model. From reading the [original DCGAN paper](https://arxiv.org/pdf/1511.06434.pdf), they say:\n",
    "> All weights were initialized from a zero-centered Normal distribution with standard deviation 0.02.\n",
    "\n",
    "So, your next task will be to define a weight initialization function that does just this!\n",
    "\n",
    "You can refer back to the lesson on weight initialization or even consult existing model code, such as that from [the `networks.py` file in CycleGAN Github repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/blob/master/models/networks.py) to help you complete this function.\n",
    "\n",
    "#### Exercise: Complete the weight initialization function\n",
    "\n",
    "* This should initialize only **convolutional** and **linear** layers\n",
    "* Initialize the weights to a normal distribution, centered around 0, with a standard deviation of 0.02.\n",
    "* The bias terms, if they exist, may be left alone or set to 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def weights_init_normal(m):\n",
    "    \"\"\"\n",
    "    Applies initial weights to certain layers in a model .\n",
    "    The weights are taken from a normal distribution \n",
    "    with mean = 0, std dev = 0.02.\n",
    "    :param m: A module or layer in a network    \n",
    "    \"\"\"\n",
    "    # classname will be something like:\n",
    "    # `Conv`, `BatchNorm2d`, `Linear`, etc.\n",
    "    classname = m.__class__.__name__\n",
    "    \n",
    "    # TODO: Apply initial weights to convolutional and linear layers\n",
    "    if classname.find('Linear') != -1:\n",
    "        # apply a uniform distribution to the weights and a bias=0\n",
    "        m.weight.data.normal_(0.0, 0.02)\n",
    "        m.bias.data.fill_(0)\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Build complete network\n",
    "\n",
    "Define your models' hyperparameters and instantiate the discriminator and generator from the classes defined above. Make sure you've passed in the correct input arguments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "def build_network(d_conv_dim, g_conv_dim, z_size):\n",
    "    # define discriminator and generator\n",
    "    D = Discriminator(d_conv_dim)\n",
    "    G = Generator(z_size=z_size, conv_dim=g_conv_dim)\n",
    "\n",
    "    # initialize model weights\n",
    "    D.apply(weights_init_normal)\n",
    "    G.apply(weights_init_normal)\n",
    "\n",
    "    print(D)\n",
    "    print()\n",
    "    print(G)\n",
    "    \n",
    "    return D, G\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Define model hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Discriminator(\n",
      "  (conv1): Sequential(\n",
      "    (0): Conv2d(3, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (conv2): Sequential(\n",
      "    (0): Conv2d(32, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv3): Sequential(\n",
      "    (0): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (fc): Linear(in_features=2048, out_features=1, bias=True)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fc): Linear(in_features=100, out_features=2048, bias=True)\n",
      "  (t_conv1): Sequential(\n",
      "    (0): ConvTranspose2d(128, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv2): Sequential(\n",
      "    (0): ConvTranspose2d(64, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv3): Sequential(\n",
      "    (0): ConvTranspose2d(32, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 32\n",
    "g_conv_dim = 32\n",
    "z_size = 100\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "D, G = build_network(d_conv_dim, g_conv_dim, z_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training on GPU\n",
    "\n",
    "Check if you can train on GPU. Here, we'll set this as a boolean variable `train_on_gpu`. Later, you'll be responsible for making sure that \n",
    ">* Models,\n",
    "* Model inputs, and\n",
    "* Loss function arguments\n",
    "\n",
    "Are moved to GPU, where appropriate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training on GPU!\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "import torch\n",
    "\n",
    "# Check for a GPU\n",
    "train_on_gpu = torch.cuda.is_available()\n",
    "if not train_on_gpu:\n",
    "    print('No GPU found. Please use a GPU to train your neural network.')\n",
    "else:\n",
    "    print('Training on GPU!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Discriminator and Generator Losses\n",
    "\n",
    "Now we need to calculate the losses for both types of adversarial networks.\n",
    "\n",
    "### Discriminator Losses\n",
    "\n",
    "> * For the discriminator, the total loss is the sum of the losses for real and fake images, `d_loss = d_real_loss + d_fake_loss`. \n",
    "* Remember that we want the discriminator to output 1 for real images and 0 for fake images, so we need to set up the losses to reflect that.\n",
    "\n",
    "\n",
    "### Generator Loss\n",
    "\n",
    "The generator loss will look similar only with flipped labels. The generator's goal is to get the discriminator to *think* its generated images are *real*.\n",
    "\n",
    "#### Exercise: Complete real and fake loss functions\n",
    "\n",
    "**You may choose to use either cross entropy or a least squares error loss to complete the following `real_loss` and `fake_loss` functions.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def real_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being real.\n",
    "       param, D_out: discriminator logits\n",
    "       return: real loss'''\n",
    "    loss = torch.mean((D_out-1)**2)\n",
    "    return loss\n",
    "\n",
    "def fake_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being fake.\n",
    "       param, D_out: discriminator logits\n",
    "       return: fake loss'''\n",
    "    loss = torch.mean(D_out**2)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizers\n",
    "\n",
    "#### Exercise: Define optimizers for your Discriminator (D) and Generator (G)\n",
    "\n",
    "Define optimizers for your models with appropriate hyperparameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "# params\n",
    "lr = 0.0002\n",
    "beta1=0.5\n",
    "beta2=0.999 # default value\n",
    "\n",
    "# Create optimizers for the discriminator D and generator G\n",
    "d_optimizer = optim.Adam(D.parameters(), lr, [beta1, beta2])\n",
    "g_optimizer = optim.Adam(G.parameters(), lr, [beta1, beta2])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Training\n",
    "\n",
    "Training will involve alternating between training the discriminator and the generator. You'll use your functions `real_loss` and `fake_loss` to help you calculate the discriminator losses.\n",
    "\n",
    "* You should train the discriminator by alternating on real and fake images\n",
    "* Then the generator, which tries to trick the discriminator and should have an opposing loss function\n",
    "\n",
    "\n",
    "#### Saving Samples\n",
    "\n",
    "You've been given some code to print out some loss statistics and save some generated \"fake\" samples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Complete the training function\n",
    "\n",
    "Keep in mind that, if you've moved your models to GPU, you'll also have to move any model inputs to GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(D, G, n_epochs, print_every=50):\n",
    "    '''Trains adversarial networks for some number of epochs\n",
    "       param, D: the discriminator network\n",
    "       param, G: the generator network\n",
    "       param, n_epochs: number of epochs to train for\n",
    "       param, print_every: when to print and record the models' losses\n",
    "       return: D and G losses'''\n",
    "    \n",
    "    # move models to GPU\n",
    "    if train_on_gpu:\n",
    "        D.cuda()\n",
    "        G.cuda()\n",
    "\n",
    "    # keep track of loss and generated, \"fake\" samples\n",
    "    samples = []\n",
    "    losses = []\n",
    "\n",
    "    # Get some fixed data for sampling. These are images that are held\n",
    "    # constant throughout training, and allow us to inspect the model's performance\n",
    "    sample_size=16\n",
    "    fixed_z = np.random.uniform(-1, 1, size=(sample_size, z_size))\n",
    "    fixed_z = torch.from_numpy(fixed_z).float()\n",
    "    # move z to GPU if available\n",
    "    if train_on_gpu:\n",
    "        fixed_z = fixed_z.cuda()\n",
    "\n",
    "    # epoch training loop\n",
    "    for epoch in range(n_epochs):\n",
    "\n",
    "        # batch training loop\n",
    "        for batch_i, (real_images, _) in enumerate(celeba_train_loader):\n",
    "\n",
    "            batch_size = real_images.size(0)\n",
    "            real_images = scale(real_images)\n",
    "\n",
    "            # ===============================================\n",
    "            #         YOUR CODE HERE: TRAIN THE NETWORKS\n",
    "            # ===============================================\n",
    "            \n",
    "            \n",
    "            # ============================================\n",
    "            #            TRAIN THE DISCRIMINATOR\n",
    "            # ============================================\n",
    "            \n",
    "            d_optimizer.zero_grad()\n",
    "            \n",
    "            # Train the discriminator on real and fake images\n",
    "            # 1. Train with real images\n",
    "\n",
    "            # Compute the discriminator losses on real images \n",
    "            if train_on_gpu:\n",
    "                real_images = real_images.cuda()\n",
    "        \n",
    "            D_real = D(real_images)\n",
    "            d_real_loss = real_loss(D_real)\n",
    "            \n",
    "            # 2. Train with fake images\n",
    "        \n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            # move x to GPU, if available\n",
    "            if train_on_gpu:\n",
    "                    z = z.cuda()\n",
    "                    fake_images = G(z)\n",
    "        \n",
    "            # Compute the discriminator losses on fake images            \n",
    "            D_fake = D(fake_images)\n",
    "            d_fake_loss = fake_loss(D_fake)\n",
    "        \n",
    "            # add up loss and perform backprop\n",
    "            d_loss = d_real_loss + d_fake_loss\n",
    "            d_loss.backward()\n",
    "            d_optimizer.step()\n",
    "            \n",
    "            \n",
    "            # =========================================\n",
    "            #            TRAIN THE GENERATOR\n",
    "            # =========================================\n",
    "            g_optimizer.zero_grad()\n",
    "\n",
    "            # 1. Train with fake images and flipped labels\n",
    "        \n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "                fake_images = G(z)\n",
    "        \n",
    "            # Compute the discriminator losses on fake images \n",
    "            # using flipped labels!\n",
    "            D_fake = D(fake_images)\n",
    "            g_loss = real_loss(D_fake) # use real loss to flip labels\n",
    "        \n",
    "            # perform backprop\n",
    "            g_loss.backward()\n",
    "            g_optimizer.step()\n",
    "            \n",
    "            \n",
    "            # ===============================================\n",
    "            #              END OF YOUR CODE\n",
    "            # ===============================================\n",
    "\n",
    "            # Print some loss stats\n",
    "            if batch_i % print_every == 0:\n",
    "                # append discriminator loss and generator loss\n",
    "                losses.append((d_loss.item(), g_loss.item()))\n",
    "                # print discriminator and generator loss\n",
    "                print('Epoch [{:5d}/{:5d}] | d_loss: {:6.4f} | g_loss: {:6.4f}'.format(\n",
    "                        epoch+1, n_epochs, d_loss.item(), g_loss.item()))\n",
    "\n",
    "\n",
    "        ## AFTER EACH EPOCH##    \n",
    "        # this code assumes your generator is named G, feel free to change the name\n",
    "        # generate and save sample, fake images\n",
    "        G.eval() # for generating samples\n",
    "        with torch.no_grad():\n",
    "            samples_z = G(fixed_z)\n",
    "            samples_z = samples_z.detach().cpu()\n",
    "            samples.append(samples_z)\n",
    "        G.train() # back to training mode\n",
    "\n",
    "    # Save training generator samples\n",
    "    with open('train_samples.pkl', 'wb') as f:\n",
    "        pkl.dump(samples, f)\n",
    "    \n",
    "    # finally return losses\n",
    "    return losses"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set your number of training epochs and train your GAN!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   50] | d_loss: 2.1538 | g_loss: 1.6550\n",
      "Epoch [    1/   50] | d_loss: 0.1547 | g_loss: 0.8719\n",
      "Epoch [    1/   50] | d_loss: 0.1220 | g_loss: 0.6862\n",
      "Epoch [    1/   50] | d_loss: 0.1213 | g_loss: 1.1777\n",
      "Epoch [    1/   50] | d_loss: 0.1422 | g_loss: 0.7543\n",
      "Epoch [    1/   50] | d_loss: 0.1813 | g_loss: 0.6576\n",
      "Epoch [    1/   50] | d_loss: 0.3093 | g_loss: 1.2246\n",
      "Epoch [    1/   50] | d_loss: 0.2323 | g_loss: 0.7434\n",
      "Epoch [    1/   50] | d_loss: 0.3970 | g_loss: 0.3324\n",
      "Epoch [    1/   50] | d_loss: 0.3972 | g_loss: 0.5576\n",
      "Epoch [    1/   50] | d_loss: 0.3261 | g_loss: 0.8415\n",
      "Epoch [    1/   50] | d_loss: 0.3149 | g_loss: 0.5875\n",
      "Epoch [    1/   50] | d_loss: 0.2812 | g_loss: 0.4573\n",
      "Epoch [    1/   50] | d_loss: 0.2629 | g_loss: 0.5618\n",
      "Epoch [    1/   50] | d_loss: 0.2905 | g_loss: 0.4950\n",
      "Epoch [    1/   50] | d_loss: 0.3589 | g_loss: 0.5875\n",
      "Epoch [    1/   50] | d_loss: 0.3376 | g_loss: 0.3683\n",
      "Epoch [    1/   50] | d_loss: 0.3602 | g_loss: 1.0504\n",
      "Epoch [    1/   50] | d_loss: 0.5187 | g_loss: 0.4443\n",
      "Epoch [    1/   50] | d_loss: 0.3124 | g_loss: 0.4970\n",
      "Epoch [    1/   50] | d_loss: 0.3535 | g_loss: 0.5073\n",
      "Epoch [    1/   50] | d_loss: 0.3588 | g_loss: 0.3689\n",
      "Epoch [    1/   50] | d_loss: 0.3803 | g_loss: 0.3889\n",
      "Epoch [    1/   50] | d_loss: 0.3450 | g_loss: 0.4960\n",
      "Epoch [    1/   50] | d_loss: 0.4167 | g_loss: 0.3100\n",
      "Epoch [    1/   50] | d_loss: 0.3644 | g_loss: 0.4655\n",
      "Epoch [    1/   50] | d_loss: 0.3921 | g_loss: 0.4062\n",
      "Epoch [    1/   50] | d_loss: 0.4427 | g_loss: 0.2565\n",
      "Epoch [    1/   50] | d_loss: 0.4754 | g_loss: 0.4356\n",
      "Epoch [    1/   50] | d_loss: 0.3122 | g_loss: 0.5712\n",
      "Epoch [    1/   50] | d_loss: 0.3911 | g_loss: 0.3024\n",
      "Epoch [    1/   50] | d_loss: 0.3165 | g_loss: 0.4461\n",
      "Epoch [    1/   50] | d_loss: 0.4400 | g_loss: 0.3749\n",
      "Epoch [    1/   50] | d_loss: 0.5132 | g_loss: 0.2417\n",
      "Epoch [    1/   50] | d_loss: 0.4066 | g_loss: 0.3816\n",
      "Epoch [    1/   50] | d_loss: 0.4149 | g_loss: 0.3509\n",
      "Epoch [    1/   50] | d_loss: 0.4437 | g_loss: 0.6675\n",
      "Epoch [    1/   50] | d_loss: 0.4117 | g_loss: 0.6618\n",
      "Epoch [    1/   50] | d_loss: 0.2145 | g_loss: 0.6637\n",
      "Epoch [    1/   50] | d_loss: 0.3447 | g_loss: 0.5682\n",
      "Epoch [    1/   50] | d_loss: 0.2591 | g_loss: 0.4933\n",
      "Epoch [    1/   50] | d_loss: 0.4610 | g_loss: 0.3350\n",
      "Epoch [    1/   50] | d_loss: 0.3356 | g_loss: 0.4177\n",
      "Epoch [    1/   50] | d_loss: 0.3211 | g_loss: 0.3297\n",
      "Epoch [    1/   50] | d_loss: 0.3595 | g_loss: 0.3836\n",
      "Epoch [    1/   50] | d_loss: 0.3633 | g_loss: 0.3877\n",
      "Epoch [    1/   50] | d_loss: 0.3715 | g_loss: 0.7847\n",
      "Epoch [    1/   50] | d_loss: 0.2698 | g_loss: 0.6842\n",
      "Epoch [    1/   50] | d_loss: 0.3188 | g_loss: 0.6705\n",
      "Epoch [    1/   50] | d_loss: 0.3090 | g_loss: 0.9065\n",
      "Epoch [    1/   50] | d_loss: 0.2985 | g_loss: 0.2636\n",
      "Epoch [    1/   50] | d_loss: 0.5050 | g_loss: 0.2135\n",
      "Epoch [    1/   50] | d_loss: 0.3441 | g_loss: 0.6004\n",
      "Epoch [    1/   50] | d_loss: 0.4544 | g_loss: 0.5770\n",
      "Epoch [    1/   50] | d_loss: 0.4277 | g_loss: 0.4089\n",
      "Epoch [    1/   50] | d_loss: 0.3524 | g_loss: 0.4869\n",
      "Epoch [    1/   50] | d_loss: 0.4116 | g_loss: 0.3321\n",
      "Epoch [    2/   50] | d_loss: 0.4698 | g_loss: 0.5105\n",
      "Epoch [    2/   50] | d_loss: 0.4184 | g_loss: 0.2735\n",
      "Epoch [    2/   50] | d_loss: 0.3058 | g_loss: 0.4195\n",
      "Epoch [    2/   50] | d_loss: 0.3336 | g_loss: 0.5005\n",
      "Epoch [    2/   50] | d_loss: 0.3532 | g_loss: 0.4344\n",
      "Epoch [    2/   50] | d_loss: 0.3884 | g_loss: 0.5882\n",
      "Epoch [    2/   50] | d_loss: 0.3119 | g_loss: 0.7189\n",
      "Epoch [    2/   50] | d_loss: 0.3363 | g_loss: 0.4228\n",
      "Epoch [    2/   50] | d_loss: 0.3836 | g_loss: 0.5664\n",
      "Epoch [    2/   50] | d_loss: 0.3531 | g_loss: 0.4505\n",
      "Epoch [    2/   50] | d_loss: 0.4047 | g_loss: 0.5407\n",
      "Epoch [    2/   50] | d_loss: 0.2949 | g_loss: 0.3862\n",
      "Epoch [    2/   50] | d_loss: 0.3004 | g_loss: 0.3592\n",
      "Epoch [    2/   50] | d_loss: 0.3121 | g_loss: 0.6047\n",
      "Epoch [    2/   50] | d_loss: 0.3528 | g_loss: 0.5127\n",
      "Epoch [    2/   50] | d_loss: 0.3691 | g_loss: 0.2896\n",
      "Epoch [    2/   50] | d_loss: 0.6116 | g_loss: 0.2219\n",
      "Epoch [    2/   50] | d_loss: 0.2675 | g_loss: 0.4094\n",
      "Epoch [    2/   50] | d_loss: 0.3314 | g_loss: 0.5255\n",
      "Epoch [    2/   50] | d_loss: 0.3890 | g_loss: 0.7042\n",
      "Epoch [    2/   50] | d_loss: 0.2704 | g_loss: 0.6710\n",
      "Epoch [    2/   50] | d_loss: 0.2694 | g_loss: 0.6990\n",
      "Epoch [    2/   50] | d_loss: 0.3531 | g_loss: 0.3099\n",
      "Epoch [    2/   50] | d_loss: 0.4939 | g_loss: 0.2031\n",
      "Epoch [    2/   50] | d_loss: 0.4084 | g_loss: 0.2770\n",
      "Epoch [    2/   50] | d_loss: 0.2858 | g_loss: 0.5711\n",
      "Epoch [    2/   50] | d_loss: 0.3952 | g_loss: 0.5490\n",
      "Epoch [    2/   50] | d_loss: 0.2885 | g_loss: 0.5276\n",
      "Epoch [    2/   50] | d_loss: 0.4056 | g_loss: 0.5731\n",
      "Epoch [    2/   50] | d_loss: 0.3613 | g_loss: 0.7697\n",
      "Epoch [    2/   50] | d_loss: 0.4550 | g_loss: 0.2218\n",
      "Epoch [    2/   50] | d_loss: 0.5122 | g_loss: 0.3182\n",
      "Epoch [    2/   50] | d_loss: 0.2823 | g_loss: 0.4594\n",
      "Epoch [    2/   50] | d_loss: 0.3981 | g_loss: 0.4980\n",
      "Epoch [    2/   50] | d_loss: 0.3131 | g_loss: 0.4045\n",
      "Epoch [    2/   50] | d_loss: 0.3432 | g_loss: 0.4973\n",
      "Epoch [    2/   50] | d_loss: 0.3588 | g_loss: 0.3790\n",
      "Epoch [    2/   50] | d_loss: 0.4357 | g_loss: 0.6759\n",
      "Epoch [    2/   50] | d_loss: 0.7097 | g_loss: 0.5083\n",
      "Epoch [    2/   50] | d_loss: 0.4071 | g_loss: 0.3317\n",
      "Epoch [    2/   50] | d_loss: 0.4776 | g_loss: 0.3387\n",
      "Epoch [    2/   50] | d_loss: 0.3803 | g_loss: 0.3835\n",
      "Epoch [    2/   50] | d_loss: 0.3765 | g_loss: 0.5438\n",
      "Epoch [    2/   50] | d_loss: 0.2866 | g_loss: 0.3386\n",
      "Epoch [    2/   50] | d_loss: 0.3376 | g_loss: 0.2880\n",
      "Epoch [    2/   50] | d_loss: 0.3174 | g_loss: 0.4911\n",
      "Epoch [    2/   50] | d_loss: 0.2852 | g_loss: 1.0361\n",
      "Epoch [    2/   50] | d_loss: 0.4329 | g_loss: 0.7630\n",
      "Epoch [    2/   50] | d_loss: 0.4985 | g_loss: 0.5395\n",
      "Epoch [    2/   50] | d_loss: 0.4344 | g_loss: 0.2527\n",
      "Epoch [    2/   50] | d_loss: 0.3397 | g_loss: 0.5736\n",
      "Epoch [    2/   50] | d_loss: 0.4513 | g_loss: 0.3810\n",
      "Epoch [    2/   50] | d_loss: 0.2679 | g_loss: 0.4677\n",
      "Epoch [    2/   50] | d_loss: 0.3888 | g_loss: 0.5822\n",
      "Epoch [    2/   50] | d_loss: 0.3397 | g_loss: 0.2316\n",
      "Epoch [    2/   50] | d_loss: 0.3045 | g_loss: 0.4815\n",
      "Epoch [    2/   50] | d_loss: 0.5240 | g_loss: 0.4199\n",
      "Epoch [    3/   50] | d_loss: 0.3580 | g_loss: 0.5697\n",
      "Epoch [    3/   50] | d_loss: 0.3777 | g_loss: 0.4326\n",
      "Epoch [    3/   50] | d_loss: 0.5146 | g_loss: 0.1056\n",
      "Epoch [    3/   50] | d_loss: 0.4862 | g_loss: 0.6740\n",
      "Epoch [    3/   50] | d_loss: 0.4398 | g_loss: 0.4359\n",
      "Epoch [    3/   50] | d_loss: 0.3546 | g_loss: 0.5389\n",
      "Epoch [    3/   50] | d_loss: 0.4378 | g_loss: 0.5855\n",
      "Epoch [    3/   50] | d_loss: 0.5494 | g_loss: 0.4326\n",
      "Epoch [    3/   50] | d_loss: 0.3618 | g_loss: 0.3394\n",
      "Epoch [    3/   50] | d_loss: 0.4425 | g_loss: 0.2593\n",
      "Epoch [    3/   50] | d_loss: 0.3188 | g_loss: 0.4124\n",
      "Epoch [    3/   50] | d_loss: 0.2697 | g_loss: 0.6602\n",
      "Epoch [    3/   50] | d_loss: 0.3847 | g_loss: 0.5166\n",
      "Epoch [    3/   50] | d_loss: 0.3582 | g_loss: 0.3665\n",
      "Epoch [    3/   50] | d_loss: 0.3230 | g_loss: 0.4060\n",
      "Epoch [    3/   50] | d_loss: 0.4792 | g_loss: 0.3016\n",
      "Epoch [    3/   50] | d_loss: 0.3539 | g_loss: 0.5227\n",
      "Epoch [    3/   50] | d_loss: 0.3373 | g_loss: 0.4351\n",
      "Epoch [    3/   50] | d_loss: 0.2003 | g_loss: 0.4747\n",
      "Epoch [    3/   50] | d_loss: 0.3337 | g_loss: 0.5402\n",
      "Epoch [    3/   50] | d_loss: 0.3582 | g_loss: 0.4288\n",
      "Epoch [    3/   50] | d_loss: 0.3070 | g_loss: 0.7507\n",
      "Epoch [    3/   50] | d_loss: 0.3983 | g_loss: 0.4085\n",
      "Epoch [    3/   50] | d_loss: 0.3802 | g_loss: 0.5516\n",
      "Epoch [    3/   50] | d_loss: 0.2456 | g_loss: 0.5130\n",
      "Epoch [    3/   50] | d_loss: 0.3897 | g_loss: 0.3936\n",
      "Epoch [    3/   50] | d_loss: 0.3113 | g_loss: 0.4504\n",
      "Epoch [    3/   50] | d_loss: 0.3282 | g_loss: 0.3865\n",
      "Epoch [    3/   50] | d_loss: 0.5152 | g_loss: 0.7590\n",
      "Epoch [    3/   50] | d_loss: 0.3600 | g_loss: 0.6536\n",
      "Epoch [    3/   50] | d_loss: 0.3553 | g_loss: 0.5659\n",
      "Epoch [    3/   50] | d_loss: 0.4036 | g_loss: 0.5119\n",
      "Epoch [    3/   50] | d_loss: 0.3458 | g_loss: 0.3743\n",
      "Epoch [    3/   50] | d_loss: 0.1631 | g_loss: 0.8458\n",
      "Epoch [    3/   50] | d_loss: 0.1675 | g_loss: 0.7109\n",
      "Epoch [    3/   50] | d_loss: 0.2855 | g_loss: 0.4034\n",
      "Epoch [    3/   50] | d_loss: 0.2950 | g_loss: 0.4653\n",
      "Epoch [    3/   50] | d_loss: 0.2970 | g_loss: 0.6143\n",
      "Epoch [    3/   50] | d_loss: 0.5313 | g_loss: 0.3043\n",
      "Epoch [    3/   50] | d_loss: 0.4278 | g_loss: 0.5875\n",
      "Epoch [    3/   50] | d_loss: 0.1989 | g_loss: 0.7067\n",
      "Epoch [    3/   50] | d_loss: 0.3028 | g_loss: 0.5749\n",
      "Epoch [    3/   50] | d_loss: 0.2810 | g_loss: 0.5692\n",
      "Epoch [    3/   50] | d_loss: 0.3298 | g_loss: 0.5445\n",
      "Epoch [    3/   50] | d_loss: 0.4263 | g_loss: 0.4210\n",
      "Epoch [    3/   50] | d_loss: 0.2940 | g_loss: 0.5284\n",
      "Epoch [    3/   50] | d_loss: 0.4659 | g_loss: 0.5682\n",
      "Epoch [    3/   50] | d_loss: 0.4095 | g_loss: 0.7301\n",
      "Epoch [    3/   50] | d_loss: 0.2366 | g_loss: 0.4431\n",
      "Epoch [    3/   50] | d_loss: 0.3113 | g_loss: 0.5992\n",
      "Epoch [    3/   50] | d_loss: 0.3575 | g_loss: 0.6635\n",
      "Epoch [    3/   50] | d_loss: 0.3710 | g_loss: 0.4875\n",
      "Epoch [    3/   50] | d_loss: 0.4582 | g_loss: 0.7407\n",
      "Epoch [    3/   50] | d_loss: 0.4362 | g_loss: 0.4922\n",
      "Epoch [    3/   50] | d_loss: 0.3719 | g_loss: 0.3671\n",
      "Epoch [    3/   50] | d_loss: 0.2342 | g_loss: 0.6863\n",
      "Epoch [    3/   50] | d_loss: 0.2095 | g_loss: 0.3103\n",
      "Epoch [    4/   50] | d_loss: 0.4348 | g_loss: 0.5129\n",
      "Epoch [    4/   50] | d_loss: 0.2941 | g_loss: 0.3123\n",
      "Epoch [    4/   50] | d_loss: 0.3379 | g_loss: 0.7563\n",
      "Epoch [    4/   50] | d_loss: 0.3062 | g_loss: 0.5034\n",
      "Epoch [    4/   50] | d_loss: 0.3177 | g_loss: 0.2644\n",
      "Epoch [    4/   50] | d_loss: 0.3450 | g_loss: 0.3503\n",
      "Epoch [    4/   50] | d_loss: 0.3360 | g_loss: 0.8724\n",
      "Epoch [    4/   50] | d_loss: 0.2673 | g_loss: 0.5944\n",
      "Epoch [    4/   50] | d_loss: 0.2239 | g_loss: 1.0163\n",
      "Epoch [    4/   50] | d_loss: 0.2609 | g_loss: 0.6532\n",
      "Epoch [    4/   50] | d_loss: 0.3607 | g_loss: 0.5774\n",
      "Epoch [    4/   50] | d_loss: 0.4080 | g_loss: 0.6754\n",
      "Epoch [    4/   50] | d_loss: 0.2462 | g_loss: 0.3962\n",
      "Epoch [    4/   50] | d_loss: 0.3671 | g_loss: 0.3546\n",
      "Epoch [    4/   50] | d_loss: 0.2395 | g_loss: 0.3702\n",
      "Epoch [    4/   50] | d_loss: 0.2613 | g_loss: 0.5019\n",
      "Epoch [    4/   50] | d_loss: 0.2079 | g_loss: 0.8886\n",
      "Epoch [    4/   50] | d_loss: 0.4611 | g_loss: 0.3319\n",
      "Epoch [    4/   50] | d_loss: 0.1992 | g_loss: 0.4001\n",
      "Epoch [    4/   50] | d_loss: 0.2511 | g_loss: 0.4822\n",
      "Epoch [    4/   50] | d_loss: 0.3747 | g_loss: 0.4216\n",
      "Epoch [    4/   50] | d_loss: 0.2916 | g_loss: 0.3680\n",
      "Epoch [    4/   50] | d_loss: 0.3135 | g_loss: 0.5629\n",
      "Epoch [    4/   50] | d_loss: 0.3427 | g_loss: 0.4324\n",
      "Epoch [    4/   50] | d_loss: 0.1892 | g_loss: 0.6931\n",
      "Epoch [    4/   50] | d_loss: 0.3108 | g_loss: 0.7164\n",
      "Epoch [    4/   50] | d_loss: 0.2880 | g_loss: 0.4718\n",
      "Epoch [    4/   50] | d_loss: 0.2157 | g_loss: 0.5258\n",
      "Epoch [    4/   50] | d_loss: 0.4188 | g_loss: 0.4689\n",
      "Epoch [    4/   50] | d_loss: 0.4954 | g_loss: 0.7988\n",
      "Epoch [    4/   50] | d_loss: 0.2402 | g_loss: 1.0528\n",
      "Epoch [    4/   50] | d_loss: 0.4156 | g_loss: 0.3328\n",
      "Epoch [    4/   50] | d_loss: 0.3024 | g_loss: 0.6297\n",
      "Epoch [    4/   50] | d_loss: 0.3927 | g_loss: 0.6919\n",
      "Epoch [    4/   50] | d_loss: 0.2845 | g_loss: 0.5541\n",
      "Epoch [    4/   50] | d_loss: 0.2542 | g_loss: 0.5788\n",
      "Epoch [    4/   50] | d_loss: 0.3212 | g_loss: 0.4139\n",
      "Epoch [    4/   50] | d_loss: 0.3477 | g_loss: 0.5475\n",
      "Epoch [    4/   50] | d_loss: 0.3634 | g_loss: 0.3933\n",
      "Epoch [    4/   50] | d_loss: 0.4073 | g_loss: 0.5509\n",
      "Epoch [    4/   50] | d_loss: 0.3817 | g_loss: 0.5094\n",
      "Epoch [    4/   50] | d_loss: 0.3461 | g_loss: 0.2468\n",
      "Epoch [    4/   50] | d_loss: 0.2996 | g_loss: 0.5625\n",
      "Epoch [    4/   50] | d_loss: 0.3468 | g_loss: 0.6297\n",
      "Epoch [    4/   50] | d_loss: 0.2782 | g_loss: 0.7841\n",
      "Epoch [    4/   50] | d_loss: 0.2606 | g_loss: 0.3217\n",
      "Epoch [    4/   50] | d_loss: 0.2179 | g_loss: 0.6720\n",
      "Epoch [    4/   50] | d_loss: 0.2619 | g_loss: 0.7258\n",
      "Epoch [    4/   50] | d_loss: 0.3462 | g_loss: 0.3262\n",
      "Epoch [    4/   50] | d_loss: 0.4353 | g_loss: 0.2997\n",
      "Epoch [    4/   50] | d_loss: 0.3270 | g_loss: 0.6421\n",
      "Epoch [    4/   50] | d_loss: 0.1778 | g_loss: 1.1375\n",
      "Epoch [    4/   50] | d_loss: 0.3586 | g_loss: 0.4911\n",
      "Epoch [    4/   50] | d_loss: 0.3199 | g_loss: 0.5082\n",
      "Epoch [    4/   50] | d_loss: 0.3905 | g_loss: 0.6363\n",
      "Epoch [    4/   50] | d_loss: 0.3233 | g_loss: 0.4748\n",
      "Epoch [    4/   50] | d_loss: 0.4242 | g_loss: 0.1938\n",
      "Epoch [    5/   50] | d_loss: 0.1789 | g_loss: 0.8748\n",
      "Epoch [    5/   50] | d_loss: 0.4035 | g_loss: 0.4457\n",
      "Epoch [    5/   50] | d_loss: 0.3090 | g_loss: 0.9533\n",
      "Epoch [    5/   50] | d_loss: 0.1988 | g_loss: 0.6931\n",
      "Epoch [    5/   50] | d_loss: 0.3181 | g_loss: 0.3848\n",
      "Epoch [    5/   50] | d_loss: 0.1944 | g_loss: 0.6826\n",
      "Epoch [    5/   50] | d_loss: 0.2463 | g_loss: 1.1706\n",
      "Epoch [    5/   50] | d_loss: 0.3281 | g_loss: 0.5305\n",
      "Epoch [    5/   50] | d_loss: 0.2774 | g_loss: 0.6442\n",
      "Epoch [    5/   50] | d_loss: 0.3332 | g_loss: 0.5630\n",
      "Epoch [    5/   50] | d_loss: 0.3455 | g_loss: 0.3701\n",
      "Epoch [    5/   50] | d_loss: 0.1969 | g_loss: 0.4290\n",
      "Epoch [    5/   50] | d_loss: 0.2987 | g_loss: 0.4755\n",
      "Epoch [    5/   50] | d_loss: 0.1766 | g_loss: 0.4837\n",
      "Epoch [    5/   50] | d_loss: 0.4478 | g_loss: 0.5553\n",
      "Epoch [    5/   50] | d_loss: 0.2450 | g_loss: 0.6580\n",
      "Epoch [    5/   50] | d_loss: 0.3021 | g_loss: 0.6345\n",
      "Epoch [    5/   50] | d_loss: 0.2203 | g_loss: 0.3986\n",
      "Epoch [    5/   50] | d_loss: 0.2834 | g_loss: 0.5598\n",
      "Epoch [    5/   50] | d_loss: 0.2568 | g_loss: 0.5977\n",
      "Epoch [    5/   50] | d_loss: 0.3423 | g_loss: 0.3507\n",
      "Epoch [    5/   50] | d_loss: 0.2538 | g_loss: 0.5857\n",
      "Epoch [    5/   50] | d_loss: 0.2801 | g_loss: 0.4368\n",
      "Epoch [    5/   50] | d_loss: 0.2609 | g_loss: 0.6433\n",
      "Epoch [    5/   50] | d_loss: 0.3996 | g_loss: 0.6331\n",
      "Epoch [    5/   50] | d_loss: 0.2915 | g_loss: 0.5154\n",
      "Epoch [    5/   50] | d_loss: 0.3716 | g_loss: 0.6244\n",
      "Epoch [    5/   50] | d_loss: 0.2846 | g_loss: 0.5741\n",
      "Epoch [    5/   50] | d_loss: 0.2705 | g_loss: 0.4315\n",
      "Epoch [    5/   50] | d_loss: 0.2435 | g_loss: 0.4758\n",
      "Epoch [    5/   50] | d_loss: 0.2927 | g_loss: 0.7385\n",
      "Epoch [    5/   50] | d_loss: 0.2084 | g_loss: 0.3919\n",
      "Epoch [    5/   50] | d_loss: 0.2358 | g_loss: 0.5350\n",
      "Epoch [    5/   50] | d_loss: 0.3143 | g_loss: 0.6824\n",
      "Epoch [    5/   50] | d_loss: 0.3866 | g_loss: 0.2727\n",
      "Epoch [    5/   50] | d_loss: 0.2747 | g_loss: 0.9293\n",
      "Epoch [    5/   50] | d_loss: 0.1764 | g_loss: 0.8882\n",
      "Epoch [    5/   50] | d_loss: 0.3462 | g_loss: 0.5514\n",
      "Epoch [    5/   50] | d_loss: 0.4141 | g_loss: 0.9897\n",
      "Epoch [    5/   50] | d_loss: 0.3374 | g_loss: 0.4688\n",
      "Epoch [    5/   50] | d_loss: 0.1923 | g_loss: 0.6706\n",
      "Epoch [    5/   50] | d_loss: 0.3055 | g_loss: 0.7215\n",
      "Epoch [    5/   50] | d_loss: 0.3311 | g_loss: 0.4625\n",
      "Epoch [    5/   50] | d_loss: 0.2703 | g_loss: 0.4067\n",
      "Epoch [    5/   50] | d_loss: 0.2021 | g_loss: 0.7023\n",
      "Epoch [    5/   50] | d_loss: 0.1585 | g_loss: 0.5504\n",
      "Epoch [    5/   50] | d_loss: 0.2803 | g_loss: 0.3644\n",
      "Epoch [    5/   50] | d_loss: 0.2915 | g_loss: 0.5292\n",
      "Epoch [    5/   50] | d_loss: 0.3067 | g_loss: 0.3314\n",
      "Epoch [    5/   50] | d_loss: 0.1936 | g_loss: 0.5022\n",
      "Epoch [    5/   50] | d_loss: 0.3564 | g_loss: 0.5618\n",
      "Epoch [    5/   50] | d_loss: 0.2183 | g_loss: 0.4460\n",
      "Epoch [    5/   50] | d_loss: 0.2967 | g_loss: 0.7381\n",
      "Epoch [    5/   50] | d_loss: 0.3958 | g_loss: 0.2915\n",
      "Epoch [    5/   50] | d_loss: 0.2569 | g_loss: 0.7787\n",
      "Epoch [    5/   50] | d_loss: 0.3863 | g_loss: 0.2884\n",
      "Epoch [    5/   50] | d_loss: 0.3037 | g_loss: 0.3963\n",
      "Epoch [    6/   50] | d_loss: 0.5117 | g_loss: 0.4482\n",
      "Epoch [    6/   50] | d_loss: 0.9938 | g_loss: 0.4496\n",
      "Epoch [    6/   50] | d_loss: 0.2917 | g_loss: 0.5731\n",
      "Epoch [    6/   50] | d_loss: 0.1947 | g_loss: 0.5201\n",
      "Epoch [    6/   50] | d_loss: 0.3108 | g_loss: 0.4472\n",
      "Epoch [    6/   50] | d_loss: 0.2953 | g_loss: 0.2702\n",
      "Epoch [    6/   50] | d_loss: 0.1154 | g_loss: 0.6325\n",
      "Epoch [    6/   50] | d_loss: 0.5589 | g_loss: 0.4624\n",
      "Epoch [    6/   50] | d_loss: 0.2849 | g_loss: 0.5045\n",
      "Epoch [    6/   50] | d_loss: 0.2475 | g_loss: 0.6022\n",
      "Epoch [    6/   50] | d_loss: 0.1624 | g_loss: 0.9585\n",
      "Epoch [    6/   50] | d_loss: 0.4091 | g_loss: 0.3677\n",
      "Epoch [    6/   50] | d_loss: 0.4322 | g_loss: 0.4215\n",
      "Epoch [    6/   50] | d_loss: 0.2871 | g_loss: 0.5893\n",
      "Epoch [    6/   50] | d_loss: 0.2431 | g_loss: 0.5576\n",
      "Epoch [    6/   50] | d_loss: 0.3147 | g_loss: 0.6424\n",
      "Epoch [    6/   50] | d_loss: 0.3119 | g_loss: 0.3621\n",
      "Epoch [    6/   50] | d_loss: 0.2929 | g_loss: 0.6890\n",
      "Epoch [    6/   50] | d_loss: 0.2483 | g_loss: 0.7878\n",
      "Epoch [    6/   50] | d_loss: 0.2450 | g_loss: 0.5049\n",
      "Epoch [    6/   50] | d_loss: 0.3681 | g_loss: 0.4430\n",
      "Epoch [    6/   50] | d_loss: 0.2064 | g_loss: 0.5089\n",
      "Epoch [    6/   50] | d_loss: 0.3575 | g_loss: 0.2480\n",
      "Epoch [    6/   50] | d_loss: 0.2237 | g_loss: 0.5056\n",
      "Epoch [    6/   50] | d_loss: 0.4236 | g_loss: 0.7047\n",
      "Epoch [    6/   50] | d_loss: 0.3510 | g_loss: 0.5727\n",
      "Epoch [    6/   50] | d_loss: 0.3290 | g_loss: 0.4665\n",
      "Epoch [    6/   50] | d_loss: 0.2098 | g_loss: 0.3708\n",
      "Epoch [    6/   50] | d_loss: 0.3485 | g_loss: 0.4887\n",
      "Epoch [    6/   50] | d_loss: 0.2133 | g_loss: 0.6257\n",
      "Epoch [    6/   50] | d_loss: 0.3147 | g_loss: 0.7953\n",
      "Epoch [    6/   50] | d_loss: 0.2839 | g_loss: 0.2826\n",
      "Epoch [    6/   50] | d_loss: 0.2261 | g_loss: 0.8452\n",
      "Epoch [    6/   50] | d_loss: 0.3625 | g_loss: 0.4331\n",
      "Epoch [    6/   50] | d_loss: 0.1781 | g_loss: 0.8424\n",
      "Epoch [    6/   50] | d_loss: 0.2458 | g_loss: 0.2205\n",
      "Epoch [    6/   50] | d_loss: 0.2541 | g_loss: 0.6716\n",
      "Epoch [    6/   50] | d_loss: 0.2079 | g_loss: 0.3899\n",
      "Epoch [    6/   50] | d_loss: 0.2672 | g_loss: 0.2886\n",
      "Epoch [    6/   50] | d_loss: 0.2639 | g_loss: 1.0951\n",
      "Epoch [    6/   50] | d_loss: 0.2960 | g_loss: 0.3713\n",
      "Epoch [    6/   50] | d_loss: 0.2834 | g_loss: 0.4363\n",
      "Epoch [    6/   50] | d_loss: 0.2212 | g_loss: 1.0838\n",
      "Epoch [    6/   50] | d_loss: 0.3476 | g_loss: 0.5194\n",
      "Epoch [    6/   50] | d_loss: 0.1328 | g_loss: 0.7130\n",
      "Epoch [    6/   50] | d_loss: 0.1351 | g_loss: 0.5394\n",
      "Epoch [    6/   50] | d_loss: 0.2810 | g_loss: 0.6205\n",
      "Epoch [    6/   50] | d_loss: 0.2540 | g_loss: 0.3539\n",
      "Epoch [    6/   50] | d_loss: 0.2878 | g_loss: 0.3775\n",
      "Epoch [    6/   50] | d_loss: 0.4117 | g_loss: 0.5818\n",
      "Epoch [    6/   50] | d_loss: 0.3419 | g_loss: 0.3438\n",
      "Epoch [    6/   50] | d_loss: 0.2500 | g_loss: 0.2826\n",
      "Epoch [    6/   50] | d_loss: 0.2191 | g_loss: 0.6014\n",
      "Epoch [    6/   50] | d_loss: 0.4370 | g_loss: 0.7056\n",
      "Epoch [    6/   50] | d_loss: 0.2002 | g_loss: 0.4966\n",
      "Epoch [    6/   50] | d_loss: 0.1897 | g_loss: 0.7955\n",
      "Epoch [    6/   50] | d_loss: 0.2293 | g_loss: 0.8318\n",
      "Epoch [    7/   50] | d_loss: 0.1914 | g_loss: 0.7363\n",
      "Epoch [    7/   50] | d_loss: 0.2343 | g_loss: 0.7236\n",
      "Epoch [    7/   50] | d_loss: 0.3126 | g_loss: 0.6053\n",
      "Epoch [    7/   50] | d_loss: 0.2502 | g_loss: 0.7816\n",
      "Epoch [    7/   50] | d_loss: 0.3067 | g_loss: 0.2727\n",
      "Epoch [    7/   50] | d_loss: 0.3232 | g_loss: 0.9133\n",
      "Epoch [    7/   50] | d_loss: 0.2528 | g_loss: 0.8105\n",
      "Epoch [    7/   50] | d_loss: 0.2374 | g_loss: 0.4807\n",
      "Epoch [    7/   50] | d_loss: 0.2242 | g_loss: 1.1864\n",
      "Epoch [    7/   50] | d_loss: 0.2903 | g_loss: 0.4641\n",
      "Epoch [    7/   50] | d_loss: 0.3793 | g_loss: 0.7957\n",
      "Epoch [    7/   50] | d_loss: 0.2564 | g_loss: 0.5899\n",
      "Epoch [    7/   50] | d_loss: 0.1670 | g_loss: 0.3898\n",
      "Epoch [    7/   50] | d_loss: 0.3214 | g_loss: 0.5482\n",
      "Epoch [    7/   50] | d_loss: 0.2508 | g_loss: 0.8934\n",
      "Epoch [    7/   50] | d_loss: 0.1962 | g_loss: 1.0764\n",
      "Epoch [    7/   50] | d_loss: 0.2225 | g_loss: 0.7463\n",
      "Epoch [    7/   50] | d_loss: 0.2031 | g_loss: 0.3525\n",
      "Epoch [    7/   50] | d_loss: 0.3823 | g_loss: 0.5131\n",
      "Epoch [    7/   50] | d_loss: 0.1660 | g_loss: 0.5130\n",
      "Epoch [    7/   50] | d_loss: 0.2308 | g_loss: 0.4029\n",
      "Epoch [    7/   50] | d_loss: 0.3295 | g_loss: 0.5074\n",
      "Epoch [    7/   50] | d_loss: 0.3422 | g_loss: 0.4204\n",
      "Epoch [    7/   50] | d_loss: 0.4015 | g_loss: 0.3846\n",
      "Epoch [    7/   50] | d_loss: 0.2216 | g_loss: 0.6788\n",
      "Epoch [    7/   50] | d_loss: 0.2545 | g_loss: 0.5615\n",
      "Epoch [    7/   50] | d_loss: 0.2234 | g_loss: 0.4425\n",
      "Epoch [    7/   50] | d_loss: 0.3212 | g_loss: 0.5137\n",
      "Epoch [    7/   50] | d_loss: 0.2215 | g_loss: 0.4042\n",
      "Epoch [    7/   50] | d_loss: 0.3892 | g_loss: 0.6122\n",
      "Epoch [    7/   50] | d_loss: 0.3475 | g_loss: 0.4815\n",
      "Epoch [    7/   50] | d_loss: 0.2476 | g_loss: 0.5433\n",
      "Epoch [    7/   50] | d_loss: 0.2717 | g_loss: 0.5743\n",
      "Epoch [    7/   50] | d_loss: 0.2452 | g_loss: 0.7444\n",
      "Epoch [    7/   50] | d_loss: 0.2817 | g_loss: 0.4856\n",
      "Epoch [    7/   50] | d_loss: 0.2164 | g_loss: 0.4241\n",
      "Epoch [    7/   50] | d_loss: 0.2597 | g_loss: 0.5368\n",
      "Epoch [    7/   50] | d_loss: 0.3722 | g_loss: 0.3107\n",
      "Epoch [    7/   50] | d_loss: 0.2448 | g_loss: 0.5729\n",
      "Epoch [    7/   50] | d_loss: 0.1363 | g_loss: 0.7465\n",
      "Epoch [    7/   50] | d_loss: 0.2830 | g_loss: 0.6352\n",
      "Epoch [    7/   50] | d_loss: 0.4016 | g_loss: 1.3607\n",
      "Epoch [    7/   50] | d_loss: 0.3200 | g_loss: 0.5023\n",
      "Epoch [    7/   50] | d_loss: 0.2344 | g_loss: 0.6398\n",
      "Epoch [    7/   50] | d_loss: 0.1608 | g_loss: 0.8285\n",
      "Epoch [    7/   50] | d_loss: 0.2905 | g_loss: 0.4378\n",
      "Epoch [    7/   50] | d_loss: 0.2051 | g_loss: 0.7011\n",
      "Epoch [    7/   50] | d_loss: 0.2278 | g_loss: 0.5716\n",
      "Epoch [    7/   50] | d_loss: 0.2691 | g_loss: 0.3525\n",
      "Epoch [    7/   50] | d_loss: 0.2298 | g_loss: 0.8977\n",
      "Epoch [    7/   50] | d_loss: 0.1605 | g_loss: 0.6337\n",
      "Epoch [    7/   50] | d_loss: 0.3425 | g_loss: 0.8859\n",
      "Epoch [    7/   50] | d_loss: 0.2590 | g_loss: 0.8128\n",
      "Epoch [    7/   50] | d_loss: 0.1333 | g_loss: 0.7585\n",
      "Epoch [    7/   50] | d_loss: 0.2908 | g_loss: 0.4132\n",
      "Epoch [    7/   50] | d_loss: 0.2207 | g_loss: 0.8560\n",
      "Epoch [    7/   50] | d_loss: 0.1693 | g_loss: 0.6814\n",
      "Epoch [    8/   50] | d_loss: 0.2681 | g_loss: 0.5150\n",
      "Epoch [    8/   50] | d_loss: 0.2821 | g_loss: 0.6139\n",
      "Epoch [    8/   50] | d_loss: 0.2362 | g_loss: 0.5663\n",
      "Epoch [    8/   50] | d_loss: 0.1614 | g_loss: 0.5572\n",
      "Epoch [    8/   50] | d_loss: 0.2086 | g_loss: 0.6945\n",
      "Epoch [    8/   50] | d_loss: 0.1382 | g_loss: 0.4460\n",
      "Epoch [    8/   50] | d_loss: 0.2076 | g_loss: 0.6668\n",
      "Epoch [    8/   50] | d_loss: 0.3254 | g_loss: 0.9130\n",
      "Epoch [    8/   50] | d_loss: 0.1634 | g_loss: 0.5499\n",
      "Epoch [    8/   50] | d_loss: 0.2631 | g_loss: 0.7407\n",
      "Epoch [    8/   50] | d_loss: 0.4032 | g_loss: 1.0346\n",
      "Epoch [    8/   50] | d_loss: 0.2160 | g_loss: 0.4375\n",
      "Epoch [    8/   50] | d_loss: 0.3119 | g_loss: 0.3247\n",
      "Epoch [    8/   50] | d_loss: 0.2173 | g_loss: 0.5468\n",
      "Epoch [    8/   50] | d_loss: 0.2709 | g_loss: 0.4234\n",
      "Epoch [    8/   50] | d_loss: 0.1278 | g_loss: 0.5063\n",
      "Epoch [    8/   50] | d_loss: 0.1932 | g_loss: 0.4133\n",
      "Epoch [    8/   50] | d_loss: 0.1229 | g_loss: 0.3389\n",
      "Epoch [    8/   50] | d_loss: 0.1735 | g_loss: 0.6958\n",
      "Epoch [    8/   50] | d_loss: 0.2639 | g_loss: 0.3212\n",
      "Epoch [    8/   50] | d_loss: 0.4858 | g_loss: 0.4225\n",
      "Epoch [    8/   50] | d_loss: 0.1533 | g_loss: 0.6367\n",
      "Epoch [    8/   50] | d_loss: 0.4407 | g_loss: 0.6952\n",
      "Epoch [    8/   50] | d_loss: 0.2437 | g_loss: 0.3832\n",
      "Epoch [    8/   50] | d_loss: 0.2075 | g_loss: 0.5524\n",
      "Epoch [    8/   50] | d_loss: 0.2709 | g_loss: 0.6962\n",
      "Epoch [    8/   50] | d_loss: 0.2773 | g_loss: 0.7096\n",
      "Epoch [    8/   50] | d_loss: 0.2158 | g_loss: 0.4828\n",
      "Epoch [    8/   50] | d_loss: 0.2367 | g_loss: 0.5372\n",
      "Epoch [    8/   50] | d_loss: 0.2222 | g_loss: 0.5128\n",
      "Epoch [    8/   50] | d_loss: 0.2875 | g_loss: 0.4943\n",
      "Epoch [    8/   50] | d_loss: 0.1681 | g_loss: 0.5357\n",
      "Epoch [    8/   50] | d_loss: 0.2278 | g_loss: 0.6137\n",
      "Epoch [    8/   50] | d_loss: 0.3071 | g_loss: 0.4479\n",
      "Epoch [    8/   50] | d_loss: 0.2101 | g_loss: 0.5586\n",
      "Epoch [    8/   50] | d_loss: 0.2713 | g_loss: 0.6986\n",
      "Epoch [    8/   50] | d_loss: 0.2810 | g_loss: 0.5603\n",
      "Epoch [    8/   50] | d_loss: 0.2206 | g_loss: 0.6004\n",
      "Epoch [    8/   50] | d_loss: 0.2418 | g_loss: 0.4441\n",
      "Epoch [    8/   50] | d_loss: 0.2749 | g_loss: 0.7219\n",
      "Epoch [    8/   50] | d_loss: 0.2044 | g_loss: 0.7233\n",
      "Epoch [    8/   50] | d_loss: 0.2807 | g_loss: 0.5282\n",
      "Epoch [    8/   50] | d_loss: 0.1408 | g_loss: 0.5668\n",
      "Epoch [    8/   50] | d_loss: 0.3198 | g_loss: 1.0838\n",
      "Epoch [    8/   50] | d_loss: 0.1291 | g_loss: 0.6898\n",
      "Epoch [    8/   50] | d_loss: 0.1922 | g_loss: 0.6790\n",
      "Epoch [    8/   50] | d_loss: 0.3098 | g_loss: 0.8639\n",
      "Epoch [    8/   50] | d_loss: 0.2706 | g_loss: 0.5054\n",
      "Epoch [    8/   50] | d_loss: 0.3066 | g_loss: 0.2531\n",
      "Epoch [    8/   50] | d_loss: 0.3567 | g_loss: 0.3607\n",
      "Epoch [    8/   50] | d_loss: 0.3042 | g_loss: 0.5170\n",
      "Epoch [    8/   50] | d_loss: 0.2752 | g_loss: 0.7940\n",
      "Epoch [    8/   50] | d_loss: 0.7834 | g_loss: 0.3238\n",
      "Epoch [    8/   50] | d_loss: 0.3979 | g_loss: 0.7196\n",
      "Epoch [    8/   50] | d_loss: 0.2946 | g_loss: 0.6274\n",
      "Epoch [    8/   50] | d_loss: 0.1583 | g_loss: 1.0451\n",
      "Epoch [    8/   50] | d_loss: 0.2584 | g_loss: 0.7098\n",
      "Epoch [    9/   50] | d_loss: 0.2311 | g_loss: 0.7190\n",
      "Epoch [    9/   50] | d_loss: 0.2157 | g_loss: 0.5910\n",
      "Epoch [    9/   50] | d_loss: 0.2854 | g_loss: 0.7729\n",
      "Epoch [    9/   50] | d_loss: 0.1804 | g_loss: 0.4222\n",
      "Epoch [    9/   50] | d_loss: 0.1636 | g_loss: 0.6684\n",
      "Epoch [    9/   50] | d_loss: 0.3629 | g_loss: 0.5556\n",
      "Epoch [    9/   50] | d_loss: 0.2042 | g_loss: 0.4070\n",
      "Epoch [    9/   50] | d_loss: 0.2240 | g_loss: 0.9952\n",
      "Epoch [    9/   50] | d_loss: 0.2524 | g_loss: 0.6223\n",
      "Epoch [    9/   50] | d_loss: 0.2314 | g_loss: 0.6464\n",
      "Epoch [    9/   50] | d_loss: 0.3490 | g_loss: 0.4936\n",
      "Epoch [    9/   50] | d_loss: 0.2067 | g_loss: 0.8415\n",
      "Epoch [    9/   50] | d_loss: 0.3055 | g_loss: 0.4352\n",
      "Epoch [    9/   50] | d_loss: 0.1748 | g_loss: 0.7443\n",
      "Epoch [    9/   50] | d_loss: 0.2333 | g_loss: 0.3690\n",
      "Epoch [    9/   50] | d_loss: 0.1191 | g_loss: 0.3984\n",
      "Epoch [    9/   50] | d_loss: 0.3315 | g_loss: 0.3366\n",
      "Epoch [    9/   50] | d_loss: 0.2064 | g_loss: 0.6036\n",
      "Epoch [    9/   50] | d_loss: 0.2670 | g_loss: 0.3729\n",
      "Epoch [    9/   50] | d_loss: 0.2574 | g_loss: 0.5856\n",
      "Epoch [    9/   50] | d_loss: 0.2472 | g_loss: 1.1399\n",
      "Epoch [    9/   50] | d_loss: 0.1524 | g_loss: 1.0249\n",
      "Epoch [    9/   50] | d_loss: 0.1115 | g_loss: 0.3788\n",
      "Epoch [    9/   50] | d_loss: 0.2218 | g_loss: 1.0198\n",
      "Epoch [    9/   50] | d_loss: 0.2495 | g_loss: 0.6645\n",
      "Epoch [    9/   50] | d_loss: 0.2326 | g_loss: 0.4370\n",
      "Epoch [    9/   50] | d_loss: 0.2669 | g_loss: 0.4722\n",
      "Epoch [    9/   50] | d_loss: 0.2639 | g_loss: 0.5182\n",
      "Epoch [    9/   50] | d_loss: 0.4835 | g_loss: 0.3550\n",
      "Epoch [    9/   50] | d_loss: 0.2586 | g_loss: 0.6308\n",
      "Epoch [    9/   50] | d_loss: 0.2690 | g_loss: 0.6720\n",
      "Epoch [    9/   50] | d_loss: 0.1076 | g_loss: 0.9498\n",
      "Epoch [    9/   50] | d_loss: 0.2795 | g_loss: 0.6528\n",
      "Epoch [    9/   50] | d_loss: 0.1900 | g_loss: 0.8314\n",
      "Epoch [    9/   50] | d_loss: 0.2218 | g_loss: 0.5275\n",
      "Epoch [    9/   50] | d_loss: 0.3577 | g_loss: 0.3347\n",
      "Epoch [    9/   50] | d_loss: 0.4156 | g_loss: 1.1076\n",
      "Epoch [    9/   50] | d_loss: 0.2170 | g_loss: 0.4513\n",
      "Epoch [    9/   50] | d_loss: 0.2266 | g_loss: 0.6912\n",
      "Epoch [    9/   50] | d_loss: 0.3097 | g_loss: 0.4162\n",
      "Epoch [    9/   50] | d_loss: 0.1466 | g_loss: 0.9508\n",
      "Epoch [    9/   50] | d_loss: 0.2698 | g_loss: 0.8824\n",
      "Epoch [    9/   50] | d_loss: 0.2648 | g_loss: 0.4159\n",
      "Epoch [    9/   50] | d_loss: 0.1545 | g_loss: 0.9647\n",
      "Epoch [    9/   50] | d_loss: 0.3579 | g_loss: 0.9156\n",
      "Epoch [    9/   50] | d_loss: 0.1197 | g_loss: 1.0428\n",
      "Epoch [    9/   50] | d_loss: 0.2160 | g_loss: 0.6632\n",
      "Epoch [    9/   50] | d_loss: 0.2419 | g_loss: 0.4646\n",
      "Epoch [    9/   50] | d_loss: 0.2735 | g_loss: 0.9291\n",
      "Epoch [    9/   50] | d_loss: 0.2074 | g_loss: 0.8521\n",
      "Epoch [    9/   50] | d_loss: 0.1742 | g_loss: 0.6878\n",
      "Epoch [    9/   50] | d_loss: 0.2343 | g_loss: 0.4842\n",
      "Epoch [    9/   50] | d_loss: 0.2505 | g_loss: 0.6253\n",
      "Epoch [    9/   50] | d_loss: 0.2922 | g_loss: 0.6786\n",
      "Epoch [    9/   50] | d_loss: 0.1976 | g_loss: 0.8340\n",
      "Epoch [    9/   50] | d_loss: 0.1373 | g_loss: 0.7941\n",
      "Epoch [    9/   50] | d_loss: 0.4216 | g_loss: 0.4716\n",
      "Epoch [   10/   50] | d_loss: 0.4288 | g_loss: 0.8141\n",
      "Epoch [   10/   50] | d_loss: 0.2329 | g_loss: 0.5177\n",
      "Epoch [   10/   50] | d_loss: 0.2402 | g_loss: 0.6731\n",
      "Epoch [   10/   50] | d_loss: 0.1428 | g_loss: 0.8584\n",
      "Epoch [   10/   50] | d_loss: 0.2518 | g_loss: 0.5999\n",
      "Epoch [   10/   50] | d_loss: 0.3870 | g_loss: 0.6541\n",
      "Epoch [   10/   50] | d_loss: 0.2552 | g_loss: 1.0219\n",
      "Epoch [   10/   50] | d_loss: 0.1162 | g_loss: 0.2819\n",
      "Epoch [   10/   50] | d_loss: 0.1305 | g_loss: 0.6311\n",
      "Epoch [   10/   50] | d_loss: 0.2106 | g_loss: 0.5149\n",
      "Epoch [   10/   50] | d_loss: 0.1470 | g_loss: 0.7037\n",
      "Epoch [   10/   50] | d_loss: 0.2728 | g_loss: 0.6692\n",
      "Epoch [   10/   50] | d_loss: 0.3075 | g_loss: 0.9300\n",
      "Epoch [   10/   50] | d_loss: 0.1377 | g_loss: 0.5974\n",
      "Epoch [   10/   50] | d_loss: 0.2363 | g_loss: 0.7422\n",
      "Epoch [   10/   50] | d_loss: 0.1461 | g_loss: 0.8910\n",
      "Epoch [   10/   50] | d_loss: 0.2076 | g_loss: 0.4937\n",
      "Epoch [   10/   50] | d_loss: 0.1080 | g_loss: 0.7985\n",
      "Epoch [   10/   50] | d_loss: 0.2271 | g_loss: 0.4820\n",
      "Epoch [   10/   50] | d_loss: 0.2944 | g_loss: 0.5917\n",
      "Epoch [   10/   50] | d_loss: 0.1447 | g_loss: 0.6631\n",
      "Epoch [   10/   50] | d_loss: 0.1362 | g_loss: 0.4417\n",
      "Epoch [   10/   50] | d_loss: 0.2480 | g_loss: 0.4525\n",
      "Epoch [   10/   50] | d_loss: 0.1194 | g_loss: 0.8488\n",
      "Epoch [   10/   50] | d_loss: 0.2179 | g_loss: 0.7560\n",
      "Epoch [   10/   50] | d_loss: 0.2152 | g_loss: 1.0037\n",
      "Epoch [   10/   50] | d_loss: 0.0890 | g_loss: 1.2502\n",
      "Epoch [   10/   50] | d_loss: 0.1805 | g_loss: 0.6669\n",
      "Epoch [   10/   50] | d_loss: 0.1776 | g_loss: 0.6508\n",
      "Epoch [   10/   50] | d_loss: 0.2324 | g_loss: 0.5152\n",
      "Epoch [   10/   50] | d_loss: 0.2032 | g_loss: 0.6519\n",
      "Epoch [   10/   50] | d_loss: 0.2735 | g_loss: 0.8088\n",
      "Epoch [   10/   50] | d_loss: 0.2204 | g_loss: 0.5590\n",
      "Epoch [   10/   50] | d_loss: 0.1453 | g_loss: 0.5857\n",
      "Epoch [   10/   50] | d_loss: 0.3183 | g_loss: 0.3895\n",
      "Epoch [   10/   50] | d_loss: 0.1245 | g_loss: 0.8378\n",
      "Epoch [   10/   50] | d_loss: 0.1320 | g_loss: 0.4691\n",
      "Epoch [   10/   50] | d_loss: 0.2188 | g_loss: 0.4660\n",
      "Epoch [   10/   50] | d_loss: 0.1589 | g_loss: 0.3279\n",
      "Epoch [   10/   50] | d_loss: 0.1797 | g_loss: 0.6504\n",
      "Epoch [   10/   50] | d_loss: 0.3564 | g_loss: 0.6061\n",
      "Epoch [   10/   50] | d_loss: 0.1772 | g_loss: 0.6453\n",
      "Epoch [   10/   50] | d_loss: 0.1651 | g_loss: 0.7867\n",
      "Epoch [   10/   50] | d_loss: 0.0797 | g_loss: 0.4840\n",
      "Epoch [   10/   50] | d_loss: 0.2202 | g_loss: 0.3060\n",
      "Epoch [   10/   50] | d_loss: 0.2306 | g_loss: 0.5339\n",
      "Epoch [   10/   50] | d_loss: 0.2321 | g_loss: 0.8233\n",
      "Epoch [   10/   50] | d_loss: 0.2214 | g_loss: 0.8235\n",
      "Epoch [   10/   50] | d_loss: 0.1569 | g_loss: 0.3843\n",
      "Epoch [   10/   50] | d_loss: 0.2062 | g_loss: 0.5830\n",
      "Epoch [   10/   50] | d_loss: 0.2149 | g_loss: 0.9323\n",
      "Epoch [   10/   50] | d_loss: 0.3052 | g_loss: 0.6535\n",
      "Epoch [   10/   50] | d_loss: 0.2874 | g_loss: 0.4999\n",
      "Epoch [   10/   50] | d_loss: 0.1373 | g_loss: 1.0751\n",
      "Epoch [   10/   50] | d_loss: 0.1590 | g_loss: 1.0229\n",
      "Epoch [   10/   50] | d_loss: 0.0974 | g_loss: 1.0585\n",
      "Epoch [   10/   50] | d_loss: 0.2686 | g_loss: 0.4483\n",
      "Epoch [   11/   50] | d_loss: 0.2832 | g_loss: 0.6870\n",
      "Epoch [   11/   50] | d_loss: 0.1945 | g_loss: 0.6711\n",
      "Epoch [   11/   50] | d_loss: 0.1308 | g_loss: 1.1170\n",
      "Epoch [   11/   50] | d_loss: 0.2564 | g_loss: 0.8781\n",
      "Epoch [   11/   50] | d_loss: 0.1774 | g_loss: 0.7217\n",
      "Epoch [   11/   50] | d_loss: 0.2332 | g_loss: 0.7637\n",
      "Epoch [   11/   50] | d_loss: 0.1734 | g_loss: 0.3120\n",
      "Epoch [   11/   50] | d_loss: 0.1410 | g_loss: 0.6658\n",
      "Epoch [   11/   50] | d_loss: 0.1882 | g_loss: 0.6349\n",
      "Epoch [   11/   50] | d_loss: 0.1130 | g_loss: 0.5646\n",
      "Epoch [   11/   50] | d_loss: 0.1165 | g_loss: 0.9706\n",
      "Epoch [   11/   50] | d_loss: 0.2378 | g_loss: 0.6922\n",
      "Epoch [   11/   50] | d_loss: 0.2543 | g_loss: 0.4005\n",
      "Epoch [   11/   50] | d_loss: 0.2200 | g_loss: 0.8264\n",
      "Epoch [   11/   50] | d_loss: 0.1582 | g_loss: 0.6452\n",
      "Epoch [   11/   50] | d_loss: 0.2589 | g_loss: 0.7405\n",
      "Epoch [   11/   50] | d_loss: 0.3778 | g_loss: 0.4409\n",
      "Epoch [   11/   50] | d_loss: 0.4138 | g_loss: 0.7161\n",
      "Epoch [   11/   50] | d_loss: 0.1482 | g_loss: 0.5404\n",
      "Epoch [   11/   50] | d_loss: 0.1483 | g_loss: 0.8470\n",
      "Epoch [   11/   50] | d_loss: 0.2454 | g_loss: 0.4586\n",
      "Epoch [   11/   50] | d_loss: 0.1136 | g_loss: 0.7730\n",
      "Epoch [   11/   50] | d_loss: 0.2448 | g_loss: 0.5978\n",
      "Epoch [   11/   50] | d_loss: 0.3079 | g_loss: 0.7949\n",
      "Epoch [   11/   50] | d_loss: 0.2341 | g_loss: 0.9931\n",
      "Epoch [   11/   50] | d_loss: 0.2449 | g_loss: 0.4084\n",
      "Epoch [   11/   50] | d_loss: 0.2286 | g_loss: 0.6212\n",
      "Epoch [   11/   50] | d_loss: 0.3003 | g_loss: 0.5417\n",
      "Epoch [   11/   50] | d_loss: 0.2983 | g_loss: 0.5827\n",
      "Epoch [   11/   50] | d_loss: 0.1830 | g_loss: 0.3926\n",
      "Epoch [   11/   50] | d_loss: 0.2696 | g_loss: 0.7233\n",
      "Epoch [   11/   50] | d_loss: 0.2920 | g_loss: 0.4060\n",
      "Epoch [   11/   50] | d_loss: 0.2050 | g_loss: 0.7584\n",
      "Epoch [   11/   50] | d_loss: 0.2476 | g_loss: 0.6053\n",
      "Epoch [   11/   50] | d_loss: 0.2022 | g_loss: 0.3930\n",
      "Epoch [   11/   50] | d_loss: 0.1724 | g_loss: 0.6953\n",
      "Epoch [   11/   50] | d_loss: 0.1889 | g_loss: 0.9832\n",
      "Epoch [   11/   50] | d_loss: 0.2746 | g_loss: 0.5781\n",
      "Epoch [   11/   50] | d_loss: 0.1469 | g_loss: 0.6719\n",
      "Epoch [   11/   50] | d_loss: 0.1659 | g_loss: 0.7538\n",
      "Epoch [   11/   50] | d_loss: 0.1825 | g_loss: 0.9473\n",
      "Epoch [   11/   50] | d_loss: 0.1989 | g_loss: 0.5479\n",
      "Epoch [   11/   50] | d_loss: 0.0723 | g_loss: 0.8022\n",
      "Epoch [   11/   50] | d_loss: 0.2633 | g_loss: 0.7664\n",
      "Epoch [   11/   50] | d_loss: 0.2420 | g_loss: 0.4987\n",
      "Epoch [   11/   50] | d_loss: 0.1119 | g_loss: 0.7888\n",
      "Epoch [   11/   50] | d_loss: 0.1508 | g_loss: 0.3650\n",
      "Epoch [   11/   50] | d_loss: 0.1304 | g_loss: 0.8352\n",
      "Epoch [   11/   50] | d_loss: 0.2464 | g_loss: 0.3240\n",
      "Epoch [   11/   50] | d_loss: 0.2834 | g_loss: 0.5108\n",
      "Epoch [   11/   50] | d_loss: 0.2444 | g_loss: 0.6323\n",
      "Epoch [   11/   50] | d_loss: 0.2245 | g_loss: 0.6771\n",
      "Epoch [   11/   50] | d_loss: 0.3051 | g_loss: 0.6308\n",
      "Epoch [   11/   50] | d_loss: 0.1511 | g_loss: 0.9623\n",
      "Epoch [   11/   50] | d_loss: 0.1340 | g_loss: 0.5567\n",
      "Epoch [   11/   50] | d_loss: 0.0944 | g_loss: 0.5185\n",
      "Epoch [   11/   50] | d_loss: 0.1614 | g_loss: 0.7151\n",
      "Epoch [   12/   50] | d_loss: 0.2884 | g_loss: 0.5838\n",
      "Epoch [   12/   50] | d_loss: 0.1747 | g_loss: 1.0545\n",
      "Epoch [   12/   50] | d_loss: 0.2675 | g_loss: 0.4104\n",
      "Epoch [   12/   50] | d_loss: 0.1286 | g_loss: 0.7979\n",
      "Epoch [   12/   50] | d_loss: 0.2251 | g_loss: 0.3811\n",
      "Epoch [   12/   50] | d_loss: 0.2765 | g_loss: 0.5512\n",
      "Epoch [   12/   50] | d_loss: 0.3231 | g_loss: 0.4804\n",
      "Epoch [   12/   50] | d_loss: 0.2374 | g_loss: 0.4487\n",
      "Epoch [   12/   50] | d_loss: 0.1327 | g_loss: 0.7421\n",
      "Epoch [   12/   50] | d_loss: 0.1233 | g_loss: 0.6883\n",
      "Epoch [   12/   50] | d_loss: 0.1468 | g_loss: 0.7172\n",
      "Epoch [   12/   50] | d_loss: 0.1877 | g_loss: 0.5256\n",
      "Epoch [   12/   50] | d_loss: 0.3186 | g_loss: 0.8448\n",
      "Epoch [   12/   50] | d_loss: 0.1904 | g_loss: 0.6619\n",
      "Epoch [   12/   50] | d_loss: 0.2346 | g_loss: 1.0760\n",
      "Epoch [   12/   50] | d_loss: 0.2322 | g_loss: 0.5665\n",
      "Epoch [   12/   50] | d_loss: 0.3301 | g_loss: 0.7966\n",
      "Epoch [   12/   50] | d_loss: 0.1251 | g_loss: 0.7837\n",
      "Epoch [   12/   50] | d_loss: 0.3471 | g_loss: 0.4264\n",
      "Epoch [   12/   50] | d_loss: 0.2307 | g_loss: 0.5768\n",
      "Epoch [   12/   50] | d_loss: 0.3218 | g_loss: 0.6441\n",
      "Epoch [   12/   50] | d_loss: 0.1968 | g_loss: 0.5437\n",
      "Epoch [   12/   50] | d_loss: 0.2598 | g_loss: 0.6658\n",
      "Epoch [   12/   50] | d_loss: 0.2110 | g_loss: 0.8080\n",
      "Epoch [   12/   50] | d_loss: 0.1369 | g_loss: 0.8182\n",
      "Epoch [   12/   50] | d_loss: 0.1543 | g_loss: 0.4297\n",
      "Epoch [   12/   50] | d_loss: 0.2383 | g_loss: 1.0925\n",
      "Epoch [   12/   50] | d_loss: 0.0855 | g_loss: 1.1166\n",
      "Epoch [   12/   50] | d_loss: 0.0850 | g_loss: 0.9980\n",
      "Epoch [   12/   50] | d_loss: 0.1491 | g_loss: 0.5777\n",
      "Epoch [   12/   50] | d_loss: 0.1071 | g_loss: 0.6224\n",
      "Epoch [   12/   50] | d_loss: 0.1629 | g_loss: 0.7088\n",
      "Epoch [   12/   50] | d_loss: 0.1507 | g_loss: 0.7634\n",
      "Epoch [   12/   50] | d_loss: 0.2358 | g_loss: 0.4478\n",
      "Epoch [   12/   50] | d_loss: 0.2164 | g_loss: 0.5150\n",
      "Epoch [   12/   50] | d_loss: 0.2330 | g_loss: 0.5660\n",
      "Epoch [   12/   50] | d_loss: 0.2301 | g_loss: 0.7762\n",
      "Epoch [   12/   50] | d_loss: 0.1532 | g_loss: 0.9158\n",
      "Epoch [   12/   50] | d_loss: 0.2343 | g_loss: 0.6974\n",
      "Epoch [   12/   50] | d_loss: 0.1927 | g_loss: 0.6456\n",
      "Epoch [   12/   50] | d_loss: 0.1567 | g_loss: 0.3998\n",
      "Epoch [   12/   50] | d_loss: 0.1990 | g_loss: 0.8487\n",
      "Epoch [   12/   50] | d_loss: 0.2932 | g_loss: 0.7502\n",
      "Epoch [   12/   50] | d_loss: 0.3733 | g_loss: 0.3972\n",
      "Epoch [   12/   50] | d_loss: 0.2829 | g_loss: 0.5119\n",
      "Epoch [   12/   50] | d_loss: 0.1335 | g_loss: 0.8690\n",
      "Epoch [   12/   50] | d_loss: 0.1738 | g_loss: 1.1877\n",
      "Epoch [   12/   50] | d_loss: 0.2219 | g_loss: 0.7713\n",
      "Epoch [   12/   50] | d_loss: 0.1092 | g_loss: 1.0930\n",
      "Epoch [   12/   50] | d_loss: 0.1148 | g_loss: 0.7366\n",
      "Epoch [   12/   50] | d_loss: 0.2084 | g_loss: 0.6220\n",
      "Epoch [   12/   50] | d_loss: 0.3887 | g_loss: 0.6635\n",
      "Epoch [   12/   50] | d_loss: 0.2118 | g_loss: 0.7340\n",
      "Epoch [   12/   50] | d_loss: 0.1209 | g_loss: 0.5860\n",
      "Epoch [   12/   50] | d_loss: 0.1450 | g_loss: 0.7586\n",
      "Epoch [   12/   50] | d_loss: 0.2158 | g_loss: 0.7763\n",
      "Epoch [   12/   50] | d_loss: 0.1868 | g_loss: 0.9231\n",
      "Epoch [   13/   50] | d_loss: 0.2703 | g_loss: 0.6055\n",
      "Epoch [   13/   50] | d_loss: 0.2327 | g_loss: 0.6666\n",
      "Epoch [   13/   50] | d_loss: 0.1863 | g_loss: 0.6280\n",
      "Epoch [   13/   50] | d_loss: 0.1385 | g_loss: 0.7784\n",
      "Epoch [   13/   50] | d_loss: 0.1355 | g_loss: 0.8457\n",
      "Epoch [   13/   50] | d_loss: 0.0952 | g_loss: 0.8785\n",
      "Epoch [   13/   50] | d_loss: 0.1440 | g_loss: 0.7083\n",
      "Epoch [   13/   50] | d_loss: 0.1043 | g_loss: 0.8127\n",
      "Epoch [   13/   50] | d_loss: 0.1460 | g_loss: 0.5079\n",
      "Epoch [   13/   50] | d_loss: 0.2074 | g_loss: 1.1570\n",
      "Epoch [   13/   50] | d_loss: 0.7563 | g_loss: 1.2470\n",
      "Epoch [   13/   50] | d_loss: 0.1184 | g_loss: 0.9073\n",
      "Epoch [   13/   50] | d_loss: 0.1892 | g_loss: 0.8435\n",
      "Epoch [   13/   50] | d_loss: 0.1794 | g_loss: 1.1289\n",
      "Epoch [   13/   50] | d_loss: 0.0899 | g_loss: 1.0440\n",
      "Epoch [   13/   50] | d_loss: 0.2597 | g_loss: 0.2843\n",
      "Epoch [   13/   50] | d_loss: 0.2459 | g_loss: 0.2699\n",
      "Epoch [   13/   50] | d_loss: 0.2029 | g_loss: 0.7902\n",
      "Epoch [   13/   50] | d_loss: 0.2317 | g_loss: 0.6576\n",
      "Epoch [   13/   50] | d_loss: 0.2457 | g_loss: 0.5146\n",
      "Epoch [   13/   50] | d_loss: 0.2273 | g_loss: 0.6947\n",
      "Epoch [   13/   50] | d_loss: 0.2348 | g_loss: 0.9888\n",
      "Epoch [   13/   50] | d_loss: 0.1468 | g_loss: 0.5260\n",
      "Epoch [   13/   50] | d_loss: 0.2195 | g_loss: 0.8576\n",
      "Epoch [   13/   50] | d_loss: 0.1253 | g_loss: 1.3099\n",
      "Epoch [   13/   50] | d_loss: 0.1644 | g_loss: 0.5300\n",
      "Epoch [   13/   50] | d_loss: 0.1688 | g_loss: 0.5274\n",
      "Epoch [   13/   50] | d_loss: 0.2177 | g_loss: 0.5262\n",
      "Epoch [   13/   50] | d_loss: 0.1616 | g_loss: 0.6629\n",
      "Epoch [   13/   50] | d_loss: 0.2916 | g_loss: 0.6486\n",
      "Epoch [   13/   50] | d_loss: 0.2130 | g_loss: 0.8019\n",
      "Epoch [   13/   50] | d_loss: 0.1244 | g_loss: 1.2513\n",
      "Epoch [   13/   50] | d_loss: 0.3546 | g_loss: 0.5997\n",
      "Epoch [   13/   50] | d_loss: 0.3953 | g_loss: 0.5639\n",
      "Epoch [   13/   50] | d_loss: 0.2543 | g_loss: 0.5963\n",
      "Epoch [   13/   50] | d_loss: 0.1781 | g_loss: 0.5755\n",
      "Epoch [   13/   50] | d_loss: 0.2061 | g_loss: 0.6038\n",
      "Epoch [   13/   50] | d_loss: 0.1667 | g_loss: 0.8861\n",
      "Epoch [   13/   50] | d_loss: 0.1723 | g_loss: 1.1758\n",
      "Epoch [   13/   50] | d_loss: 0.1582 | g_loss: 0.6572\n",
      "Epoch [   13/   50] | d_loss: 0.2375 | g_loss: 0.7137\n",
      "Epoch [   13/   50] | d_loss: 0.2697 | g_loss: 0.6438\n",
      "Epoch [   13/   50] | d_loss: 0.1557 | g_loss: 0.9929\n",
      "Epoch [   13/   50] | d_loss: 0.2105 | g_loss: 0.8627\n",
      "Epoch [   13/   50] | d_loss: 0.2500 | g_loss: 0.8559\n",
      "Epoch [   13/   50] | d_loss: 0.1364 | g_loss: 0.9313\n",
      "Epoch [   13/   50] | d_loss: 0.1441 | g_loss: 0.7070\n",
      "Epoch [   13/   50] | d_loss: 0.2573 | g_loss: 0.3690\n",
      "Epoch [   13/   50] | d_loss: 0.3740 | g_loss: 0.6903\n",
      "Epoch [   13/   50] | d_loss: 0.2283 | g_loss: 0.5065\n",
      "Epoch [   13/   50] | d_loss: 0.3377 | g_loss: 0.5699\n",
      "Epoch [   13/   50] | d_loss: 0.2681 | g_loss: 1.3374\n",
      "Epoch [   13/   50] | d_loss: 0.1686 | g_loss: 0.5858\n",
      "Epoch [   13/   50] | d_loss: 0.1796 | g_loss: 0.6685\n",
      "Epoch [   13/   50] | d_loss: 0.1874 | g_loss: 0.5846\n",
      "Epoch [   13/   50] | d_loss: 0.1895 | g_loss: 0.5005\n",
      "Epoch [   13/   50] | d_loss: 0.3004 | g_loss: 0.8447\n",
      "Epoch [   14/   50] | d_loss: 0.1314 | g_loss: 0.4933\n",
      "Epoch [   14/   50] | d_loss: 0.2565 | g_loss: 0.3203\n",
      "Epoch [   14/   50] | d_loss: 0.1287 | g_loss: 0.6259\n",
      "Epoch [   14/   50] | d_loss: 0.2115 | g_loss: 0.5114\n",
      "Epoch [   14/   50] | d_loss: 0.3883 | g_loss: 0.9013\n",
      "Epoch [   14/   50] | d_loss: 0.1721 | g_loss: 1.1022\n",
      "Epoch [   14/   50] | d_loss: 0.2292 | g_loss: 0.5390\n",
      "Epoch [   14/   50] | d_loss: 0.1966 | g_loss: 0.6327\n",
      "Epoch [   14/   50] | d_loss: 0.1254 | g_loss: 0.9700\n",
      "Epoch [   14/   50] | d_loss: 0.2826 | g_loss: 0.4222\n",
      "Epoch [   14/   50] | d_loss: 0.1681 | g_loss: 0.6921\n",
      "Epoch [   14/   50] | d_loss: 0.1708 | g_loss: 0.8101\n",
      "Epoch [   14/   50] | d_loss: 0.2253 | g_loss: 1.0723\n",
      "Epoch [   14/   50] | d_loss: 0.1714 | g_loss: 0.7859\n",
      "Epoch [   14/   50] | d_loss: 0.1456 | g_loss: 0.7151\n",
      "Epoch [   14/   50] | d_loss: 0.1800 | g_loss: 0.9866\n",
      "Epoch [   14/   50] | d_loss: 0.2167 | g_loss: 0.5726\n",
      "Epoch [   14/   50] | d_loss: 0.1335 | g_loss: 0.4542\n",
      "Epoch [   14/   50] | d_loss: 0.1476 | g_loss: 0.4020\n",
      "Epoch [   14/   50] | d_loss: 0.1906 | g_loss: 0.9395\n",
      "Epoch [   14/   50] | d_loss: 0.1938 | g_loss: 0.7225\n",
      "Epoch [   14/   50] | d_loss: 0.2396 | g_loss: 0.9566\n",
      "Epoch [   14/   50] | d_loss: 0.0603 | g_loss: 1.2475\n",
      "Epoch [   14/   50] | d_loss: 0.3215 | g_loss: 0.6302\n",
      "Epoch [   14/   50] | d_loss: 0.2513 | g_loss: 0.7930\n",
      "Epoch [   14/   50] | d_loss: 0.1356 | g_loss: 0.9709\n",
      "Epoch [   14/   50] | d_loss: 0.0989 | g_loss: 0.5740\n",
      "Epoch [   14/   50] | d_loss: 0.2754 | g_loss: 1.1676\n",
      "Epoch [   14/   50] | d_loss: 0.1184 | g_loss: 0.9319\n",
      "Epoch [   14/   50] | d_loss: 0.2348 | g_loss: 1.2243\n",
      "Epoch [   14/   50] | d_loss: 0.1327 | g_loss: 0.4368\n",
      "Epoch [   14/   50] | d_loss: 0.2212 | g_loss: 0.7281\n",
      "Epoch [   14/   50] | d_loss: 0.1944 | g_loss: 0.8933\n",
      "Epoch [   14/   50] | d_loss: 0.3052 | g_loss: 0.7265\n",
      "Epoch [   14/   50] | d_loss: 0.1695 | g_loss: 1.0401\n",
      "Epoch [   14/   50] | d_loss: 0.1216 | g_loss: 0.8436\n",
      "Epoch [   14/   50] | d_loss: 0.3737 | g_loss: 0.7854\n",
      "Epoch [   14/   50] | d_loss: 0.1479 | g_loss: 1.1177\n",
      "Epoch [   14/   50] | d_loss: 0.3562 | g_loss: 0.6616\n",
      "Epoch [   14/   50] | d_loss: 0.1620 | g_loss: 0.9378\n",
      "Epoch [   14/   50] | d_loss: 0.1932 | g_loss: 0.4709\n",
      "Epoch [   14/   50] | d_loss: 0.5088 | g_loss: 0.4213\n",
      "Epoch [   14/   50] | d_loss: 0.1712 | g_loss: 0.9552\n",
      "Epoch [   14/   50] | d_loss: 0.0962 | g_loss: 1.6271\n",
      "Epoch [   14/   50] | d_loss: 0.1254 | g_loss: 0.6355\n",
      "Epoch [   14/   50] | d_loss: 0.1675 | g_loss: 0.6543\n",
      "Epoch [   14/   50] | d_loss: 0.1313 | g_loss: 0.6917\n",
      "Epoch [   14/   50] | d_loss: 0.3334 | g_loss: 0.7115\n",
      "Epoch [   14/   50] | d_loss: 0.1323 | g_loss: 0.8219\n",
      "Epoch [   14/   50] | d_loss: 0.1316 | g_loss: 0.6987\n",
      "Epoch [   14/   50] | d_loss: 0.2128 | g_loss: 0.6592\n",
      "Epoch [   14/   50] | d_loss: 0.1250 | g_loss: 0.8081\n",
      "Epoch [   14/   50] | d_loss: 0.1394 | g_loss: 1.1721\n",
      "Epoch [   14/   50] | d_loss: 0.3319 | g_loss: 0.2380\n",
      "Epoch [   14/   50] | d_loss: 0.2926 | g_loss: 0.6582\n",
      "Epoch [   14/   50] | d_loss: 0.2276 | g_loss: 0.4371\n",
      "Epoch [   14/   50] | d_loss: 0.1640 | g_loss: 0.6431\n",
      "Epoch [   15/   50] | d_loss: 0.1193 | g_loss: 0.8902\n",
      "Epoch [   15/   50] | d_loss: 0.1467 | g_loss: 0.6102\n",
      "Epoch [   15/   50] | d_loss: 0.1041 | g_loss: 0.7769\n",
      "Epoch [   15/   50] | d_loss: 0.2225 | g_loss: 0.6300\n",
      "Epoch [   15/   50] | d_loss: 0.2888 | g_loss: 0.6752\n",
      "Epoch [   15/   50] | d_loss: 0.1144 | g_loss: 1.0544\n",
      "Epoch [   15/   50] | d_loss: 0.2003 | g_loss: 0.7841\n",
      "Epoch [   15/   50] | d_loss: 0.4063 | g_loss: 0.5329\n",
      "Epoch [   15/   50] | d_loss: 0.2038 | g_loss: 0.9335\n",
      "Epoch [   15/   50] | d_loss: 0.2569 | g_loss: 0.9110\n",
      "Epoch [   15/   50] | d_loss: 0.2631 | g_loss: 1.7181\n",
      "Epoch [   15/   50] | d_loss: 0.2142 | g_loss: 0.7632\n",
      "Epoch [   15/   50] | d_loss: 0.1806 | g_loss: 0.9534\n",
      "Epoch [   15/   50] | d_loss: 0.2356 | g_loss: 0.6295\n",
      "Epoch [   15/   50] | d_loss: 0.2280 | g_loss: 1.0940\n",
      "Epoch [   15/   50] | d_loss: 0.1037 | g_loss: 0.7859\n",
      "Epoch [   15/   50] | d_loss: 0.1656 | g_loss: 0.7204\n",
      "Epoch [   15/   50] | d_loss: 0.2097 | g_loss: 1.1073\n",
      "Epoch [   15/   50] | d_loss: 0.1250 | g_loss: 0.8506\n",
      "Epoch [   15/   50] | d_loss: 0.1170 | g_loss: 0.6540\n",
      "Epoch [   15/   50] | d_loss: 0.2327 | g_loss: 0.7559\n",
      "Epoch [   15/   50] | d_loss: 0.1409 | g_loss: 0.6461\n",
      "Epoch [   15/   50] | d_loss: 0.2577 | g_loss: 0.6966\n",
      "Epoch [   15/   50] | d_loss: 0.1550 | g_loss: 0.3953\n",
      "Epoch [   15/   50] | d_loss: 0.1904 | g_loss: 0.5973\n",
      "Epoch [   15/   50] | d_loss: 0.0987 | g_loss: 0.6508\n",
      "Epoch [   15/   50] | d_loss: 0.1452 | g_loss: 0.6687\n",
      "Epoch [   15/   50] | d_loss: 0.1514 | g_loss: 0.9857\n",
      "Epoch [   15/   50] | d_loss: 0.1255 | g_loss: 0.7133\n",
      "Epoch [   15/   50] | d_loss: 0.1501 | g_loss: 0.4415\n",
      "Epoch [   15/   50] | d_loss: 0.2276 | g_loss: 0.7940\n",
      "Epoch [   15/   50] | d_loss: 0.1038 | g_loss: 0.9146\n",
      "Epoch [   15/   50] | d_loss: 0.2218 | g_loss: 0.6834\n",
      "Epoch [   15/   50] | d_loss: 0.1922 | g_loss: 1.2672\n",
      "Epoch [   15/   50] | d_loss: 0.1554 | g_loss: 0.7544\n",
      "Epoch [   15/   50] | d_loss: 0.2081 | g_loss: 0.3842\n",
      "Epoch [   15/   50] | d_loss: 0.1409 | g_loss: 0.8384\n",
      "Epoch [   15/   50] | d_loss: 0.2094 | g_loss: 0.7824\n",
      "Epoch [   15/   50] | d_loss: 0.1236 | g_loss: 1.0088\n",
      "Epoch [   15/   50] | d_loss: 0.2689 | g_loss: 0.5950\n",
      "Epoch [   15/   50] | d_loss: 0.3397 | g_loss: 0.7030\n",
      "Epoch [   15/   50] | d_loss: 0.1052 | g_loss: 1.1023\n",
      "Epoch [   15/   50] | d_loss: 0.2016 | g_loss: 0.7652\n",
      "Epoch [   15/   50] | d_loss: 0.1839 | g_loss: 0.7482\n",
      "Epoch [   15/   50] | d_loss: 0.1288 | g_loss: 0.8233\n",
      "Epoch [   15/   50] | d_loss: 0.0972 | g_loss: 1.0157\n",
      "Epoch [   15/   50] | d_loss: 0.2683 | g_loss: 1.0353\n",
      "Epoch [   15/   50] | d_loss: 0.1182 | g_loss: 0.8096\n",
      "Epoch [   15/   50] | d_loss: 0.1848 | g_loss: 1.4849\n",
      "Epoch [   15/   50] | d_loss: 0.1638 | g_loss: 0.7148\n",
      "Epoch [   15/   50] | d_loss: 0.1230 | g_loss: 0.8008\n",
      "Epoch [   15/   50] | d_loss: 0.1434 | g_loss: 0.7033\n",
      "Epoch [   15/   50] | d_loss: 0.2327 | g_loss: 0.5749\n",
      "Epoch [   15/   50] | d_loss: 0.2172 | g_loss: 0.5882\n",
      "Epoch [   15/   50] | d_loss: 0.1030 | g_loss: 1.1295\n",
      "Epoch [   15/   50] | d_loss: 0.1137 | g_loss: 1.3709\n",
      "Epoch [   15/   50] | d_loss: 0.2465 | g_loss: 0.7694\n",
      "Epoch [   16/   50] | d_loss: 0.4674 | g_loss: 0.7406\n",
      "Epoch [   16/   50] | d_loss: 0.2320 | g_loss: 1.0963\n",
      "Epoch [   16/   50] | d_loss: 0.0999 | g_loss: 0.9103\n",
      "Epoch [   16/   50] | d_loss: 0.2320 | g_loss: 0.4154\n",
      "Epoch [   16/   50] | d_loss: 0.3339 | g_loss: 0.6025\n",
      "Epoch [   16/   50] | d_loss: 0.1658 | g_loss: 0.5561\n",
      "Epoch [   16/   50] | d_loss: 0.1657 | g_loss: 0.4550\n",
      "Epoch [   16/   50] | d_loss: 0.1021 | g_loss: 0.8809\n",
      "Epoch [   16/   50] | d_loss: 0.1691 | g_loss: 1.7004\n",
      "Epoch [   16/   50] | d_loss: 0.1756 | g_loss: 0.5791\n",
      "Epoch [   16/   50] | d_loss: 0.3878 | g_loss: 0.7052\n",
      "Epoch [   16/   50] | d_loss: 0.2032 | g_loss: 0.7159\n",
      "Epoch [   16/   50] | d_loss: 0.1164 | g_loss: 0.6155\n",
      "Epoch [   16/   50] | d_loss: 0.1836 | g_loss: 0.4723\n",
      "Epoch [   16/   50] | d_loss: 0.1374 | g_loss: 0.3487\n",
      "Epoch [   16/   50] | d_loss: 0.2524 | g_loss: 0.5491\n",
      "Epoch [   16/   50] | d_loss: 0.3503 | g_loss: 0.4353\n",
      "Epoch [   16/   50] | d_loss: 0.1783 | g_loss: 0.5809\n",
      "Epoch [   16/   50] | d_loss: 0.1161 | g_loss: 1.2721\n",
      "Epoch [   16/   50] | d_loss: 0.2418 | g_loss: 0.6191\n",
      "Epoch [   16/   50] | d_loss: 0.1972 | g_loss: 0.4702\n",
      "Epoch [   16/   50] | d_loss: 0.1751 | g_loss: 1.0038\n",
      "Epoch [   16/   50] | d_loss: 0.1160 | g_loss: 0.7401\n",
      "Epoch [   16/   50] | d_loss: 0.2089 | g_loss: 0.9424\n",
      "Epoch [   16/   50] | d_loss: 0.1809 | g_loss: 0.8546\n",
      "Epoch [   16/   50] | d_loss: 0.1582 | g_loss: 0.2745\n",
      "Epoch [   16/   50] | d_loss: 0.2283 | g_loss: 0.4465\n",
      "Epoch [   16/   50] | d_loss: 0.2190 | g_loss: 0.8564\n",
      "Epoch [   16/   50] | d_loss: 0.1774 | g_loss: 0.7317\n",
      "Epoch [   16/   50] | d_loss: 0.0693 | g_loss: 0.5243\n",
      "Epoch [   16/   50] | d_loss: 0.2173 | g_loss: 0.9554\n",
      "Epoch [   16/   50] | d_loss: 0.1081 | g_loss: 0.6305\n",
      "Epoch [   16/   50] | d_loss: 0.1862 | g_loss: 0.7268\n",
      "Epoch [   16/   50] | d_loss: 0.2102 | g_loss: 0.6751\n",
      "Epoch [   16/   50] | d_loss: 0.1538 | g_loss: 0.4494\n",
      "Epoch [   16/   50] | d_loss: 0.1670 | g_loss: 0.6105\n",
      "Epoch [   16/   50] | d_loss: 0.2644 | g_loss: 1.1829\n",
      "Epoch [   16/   50] | d_loss: 0.1603 | g_loss: 0.7660\n",
      "Epoch [   16/   50] | d_loss: 0.0793 | g_loss: 0.5427\n",
      "Epoch [   16/   50] | d_loss: 0.1415 | g_loss: 0.5572\n",
      "Epoch [   16/   50] | d_loss: 0.1418 | g_loss: 0.9167\n",
      "Epoch [   16/   50] | d_loss: 0.1215 | g_loss: 0.4677\n",
      "Epoch [   16/   50] | d_loss: 0.1510 | g_loss: 0.7001\n",
      "Epoch [   16/   50] | d_loss: 0.1649 | g_loss: 0.6564\n",
      "Epoch [   16/   50] | d_loss: 0.1719 | g_loss: 0.8416\n",
      "Epoch [   16/   50] | d_loss: 0.2874 | g_loss: 0.4984\n",
      "Epoch [   16/   50] | d_loss: 0.2087 | g_loss: 0.6762\n",
      "Epoch [   16/   50] | d_loss: 0.0510 | g_loss: 0.7960\n",
      "Epoch [   16/   50] | d_loss: 0.1648 | g_loss: 0.6628\n",
      "Epoch [   16/   50] | d_loss: 0.1116 | g_loss: 0.7430\n",
      "Epoch [   16/   50] | d_loss: 0.3147 | g_loss: 0.4802\n",
      "Epoch [   16/   50] | d_loss: 0.1660 | g_loss: 0.9728\n",
      "Epoch [   16/   50] | d_loss: 0.2061 | g_loss: 0.8819\n",
      "Epoch [   16/   50] | d_loss: 0.1659 | g_loss: 0.8405\n",
      "Epoch [   16/   50] | d_loss: 0.3802 | g_loss: 0.3583\n",
      "Epoch [   16/   50] | d_loss: 0.1604 | g_loss: 0.6645\n",
      "Epoch [   16/   50] | d_loss: 0.1482 | g_loss: 0.9530\n",
      "Epoch [   17/   50] | d_loss: 0.2204 | g_loss: 0.8851\n",
      "Epoch [   17/   50] | d_loss: 0.1531 | g_loss: 0.9688\n",
      "Epoch [   17/   50] | d_loss: 0.1766 | g_loss: 0.5875\n",
      "Epoch [   17/   50] | d_loss: 0.1702 | g_loss: 0.6139\n",
      "Epoch [   17/   50] | d_loss: 0.1017 | g_loss: 0.8040\n",
      "Epoch [   17/   50] | d_loss: 0.2257 | g_loss: 0.6021\n",
      "Epoch [   17/   50] | d_loss: 0.1833 | g_loss: 0.7676\n",
      "Epoch [   17/   50] | d_loss: 0.1174 | g_loss: 1.0908\n",
      "Epoch [   17/   50] | d_loss: 0.2351 | g_loss: 0.5691\n",
      "Epoch [   17/   50] | d_loss: 0.2440 | g_loss: 0.5381\n",
      "Epoch [   17/   50] | d_loss: 0.1242 | g_loss: 0.7314\n",
      "Epoch [   17/   50] | d_loss: 0.1555 | g_loss: 0.8029\n",
      "Epoch [   17/   50] | d_loss: 0.0552 | g_loss: 0.9999\n",
      "Epoch [   17/   50] | d_loss: 0.3636 | g_loss: 0.2561\n",
      "Epoch [   17/   50] | d_loss: 0.2260 | g_loss: 0.4665\n",
      "Epoch [   17/   50] | d_loss: 0.2557 | g_loss: 0.7782\n",
      "Epoch [   17/   50] | d_loss: 0.1554 | g_loss: 0.3310\n",
      "Epoch [   17/   50] | d_loss: 0.1729 | g_loss: 0.5186\n",
      "Epoch [   17/   50] | d_loss: 0.0909 | g_loss: 0.6112\n",
      "Epoch [   17/   50] | d_loss: 0.1238 | g_loss: 0.4426\n",
      "Epoch [   17/   50] | d_loss: 0.2207 | g_loss: 0.9353\n",
      "Epoch [   17/   50] | d_loss: 0.1334 | g_loss: 0.7399\n",
      "Epoch [   17/   50] | d_loss: 0.2044 | g_loss: 1.0445\n",
      "Epoch [   17/   50] | d_loss: 0.3988 | g_loss: 0.5734\n",
      "Epoch [   17/   50] | d_loss: 0.1517 | g_loss: 0.7263\n",
      "Epoch [   17/   50] | d_loss: 0.1310 | g_loss: 0.6352\n",
      "Epoch [   17/   50] | d_loss: 0.1709 | g_loss: 0.7890\n",
      "Epoch [   17/   50] | d_loss: 0.1774 | g_loss: 1.3558\n",
      "Epoch [   17/   50] | d_loss: 0.1136 | g_loss: 0.4260\n",
      "Epoch [   17/   50] | d_loss: 0.0972 | g_loss: 0.8044\n",
      "Epoch [   17/   50] | d_loss: 0.1037 | g_loss: 0.5680\n",
      "Epoch [   17/   50] | d_loss: 0.2027 | g_loss: 0.8655\n",
      "Epoch [   17/   50] | d_loss: 0.2078 | g_loss: 0.6004\n",
      "Epoch [   17/   50] | d_loss: 0.1544 | g_loss: 0.9064\n",
      "Epoch [   17/   50] | d_loss: 0.3581 | g_loss: 1.3572\n",
      "Epoch [   17/   50] | d_loss: 0.1421 | g_loss: 0.9028\n",
      "Epoch [   17/   50] | d_loss: 0.1117 | g_loss: 0.8022\n",
      "Epoch [   17/   50] | d_loss: 0.0908 | g_loss: 0.8091\n",
      "Epoch [   17/   50] | d_loss: 0.1340 | g_loss: 0.5548\n",
      "Epoch [   17/   50] | d_loss: 0.0987 | g_loss: 0.3145\n",
      "Epoch [   17/   50] | d_loss: 0.2108 | g_loss: 0.5856\n",
      "Epoch [   17/   50] | d_loss: 0.1538 | g_loss: 0.7568\n",
      "Epoch [   17/   50] | d_loss: 0.1176 | g_loss: 1.1907\n",
      "Epoch [   17/   50] | d_loss: 0.1742 | g_loss: 0.5819\n",
      "Epoch [   17/   50] | d_loss: 0.2470 | g_loss: 0.4801\n",
      "Epoch [   17/   50] | d_loss: 0.0861 | g_loss: 1.0696\n",
      "Epoch [   17/   50] | d_loss: 0.2215 | g_loss: 0.5530\n",
      "Epoch [   17/   50] | d_loss: 0.2744 | g_loss: 0.9146\n",
      "Epoch [   17/   50] | d_loss: 0.3011 | g_loss: 0.9541\n",
      "Epoch [   17/   50] | d_loss: 0.2021 | g_loss: 0.4258\n",
      "Epoch [   17/   50] | d_loss: 0.1328 | g_loss: 0.5411\n",
      "Epoch [   17/   50] | d_loss: 0.0920 | g_loss: 0.7324\n",
      "Epoch [   17/   50] | d_loss: 0.1719 | g_loss: 0.6538\n",
      "Epoch [   17/   50] | d_loss: 0.1524 | g_loss: 0.8321\n",
      "Epoch [   17/   50] | d_loss: 0.3953 | g_loss: 0.2984\n",
      "Epoch [   17/   50] | d_loss: 0.1011 | g_loss: 0.8042\n",
      "Epoch [   17/   50] | d_loss: 0.1555 | g_loss: 0.8265\n",
      "Epoch [   18/   50] | d_loss: 0.1652 | g_loss: 1.0174\n",
      "Epoch [   18/   50] | d_loss: 0.1891 | g_loss: 0.8885\n",
      "Epoch [   18/   50] | d_loss: 0.1716 | g_loss: 0.8538\n",
      "Epoch [   18/   50] | d_loss: 0.1901 | g_loss: 0.6766\n",
      "Epoch [   18/   50] | d_loss: 0.2268 | g_loss: 0.4757\n",
      "Epoch [   18/   50] | d_loss: 0.1465 | g_loss: 0.5540\n",
      "Epoch [   18/   50] | d_loss: 0.0921 | g_loss: 0.8788\n",
      "Epoch [   18/   50] | d_loss: 0.0925 | g_loss: 0.8068\n",
      "Epoch [   18/   50] | d_loss: 0.0520 | g_loss: 0.9143\n",
      "Epoch [   18/   50] | d_loss: 0.2516 | g_loss: 0.8084\n",
      "Epoch [   18/   50] | d_loss: 0.0968 | g_loss: 0.8497\n",
      "Epoch [   18/   50] | d_loss: 0.1147 | g_loss: 0.7159\n",
      "Epoch [   18/   50] | d_loss: 0.2073 | g_loss: 0.7637\n",
      "Epoch [   18/   50] | d_loss: 0.1665 | g_loss: 0.4547\n",
      "Epoch [   18/   50] | d_loss: 0.3356 | g_loss: 0.9590\n",
      "Epoch [   18/   50] | d_loss: 0.1992 | g_loss: 0.5791\n",
      "Epoch [   18/   50] | d_loss: 0.1085 | g_loss: 0.7625\n",
      "Epoch [   18/   50] | d_loss: 0.1479 | g_loss: 0.8881\n",
      "Epoch [   18/   50] | d_loss: 0.1909 | g_loss: 0.6680\n",
      "Epoch [   18/   50] | d_loss: 0.0885 | g_loss: 0.6748\n",
      "Epoch [   18/   50] | d_loss: 0.1274 | g_loss: 0.9651\n",
      "Epoch [   18/   50] | d_loss: 0.3384 | g_loss: 0.5386\n",
      "Epoch [   18/   50] | d_loss: 0.2389 | g_loss: 0.9476\n",
      "Epoch [   18/   50] | d_loss: 0.0996 | g_loss: 0.7550\n",
      "Epoch [   18/   50] | d_loss: 0.1806 | g_loss: 0.5776\n",
      "Epoch [   18/   50] | d_loss: 0.1337 | g_loss: 0.4823\n",
      "Epoch [   18/   50] | d_loss: 0.2062 | g_loss: 1.0806\n",
      "Epoch [   18/   50] | d_loss: 0.1712 | g_loss: 0.6130\n",
      "Epoch [   18/   50] | d_loss: 0.1387 | g_loss: 0.8565\n",
      "Epoch [   18/   50] | d_loss: 0.3246 | g_loss: 1.1436\n",
      "Epoch [   18/   50] | d_loss: 0.0766 | g_loss: 1.0237\n",
      "Epoch [   18/   50] | d_loss: 0.1134 | g_loss: 0.8525\n",
      "Epoch [   18/   50] | d_loss: 0.2189 | g_loss: 0.4962\n",
      "Epoch [   18/   50] | d_loss: 0.1071 | g_loss: 0.4871\n",
      "Epoch [   18/   50] | d_loss: 0.1657 | g_loss: 0.6007\n",
      "Epoch [   18/   50] | d_loss: 0.1718 | g_loss: 1.0627\n",
      "Epoch [   18/   50] | d_loss: 0.1348 | g_loss: 0.7668\n",
      "Epoch [   18/   50] | d_loss: 0.2055 | g_loss: 1.0626\n",
      "Epoch [   18/   50] | d_loss: 0.1374 | g_loss: 1.0353\n",
      "Epoch [   18/   50] | d_loss: 0.1507 | g_loss: 0.7606\n",
      "Epoch [   18/   50] | d_loss: 0.1133 | g_loss: 0.7226\n",
      "Epoch [   18/   50] | d_loss: 0.2454 | g_loss: 0.5056\n",
      "Epoch [   18/   50] | d_loss: 0.1031 | g_loss: 0.6616\n",
      "Epoch [   18/   50] | d_loss: 0.2052 | g_loss: 0.7079\n",
      "Epoch [   18/   50] | d_loss: 0.1688 | g_loss: 0.8118\n",
      "Epoch [   18/   50] | d_loss: 0.0863 | g_loss: 0.7306\n",
      "Epoch [   18/   50] | d_loss: 0.0855 | g_loss: 0.8845\n",
      "Epoch [   18/   50] | d_loss: 0.2131 | g_loss: 0.5341\n",
      "Epoch [   18/   50] | d_loss: 0.2487 | g_loss: 0.3618\n",
      "Epoch [   18/   50] | d_loss: 0.0876 | g_loss: 1.0138\n",
      "Epoch [   18/   50] | d_loss: 0.1895 | g_loss: 0.5663\n",
      "Epoch [   18/   50] | d_loss: 0.2157 | g_loss: 0.7169\n",
      "Epoch [   18/   50] | d_loss: 0.1160 | g_loss: 0.6519\n",
      "Epoch [   18/   50] | d_loss: 0.1162 | g_loss: 1.1405\n",
      "Epoch [   18/   50] | d_loss: 0.1709 | g_loss: 1.0122\n",
      "Epoch [   18/   50] | d_loss: 0.2077 | g_loss: 0.5585\n",
      "Epoch [   18/   50] | d_loss: 0.1220 | g_loss: 1.2153\n",
      "Epoch [   19/   50] | d_loss: 0.1310 | g_loss: 1.0145\n",
      "Epoch [   19/   50] | d_loss: 0.0679 | g_loss: 0.8935\n",
      "Epoch [   19/   50] | d_loss: 0.1376 | g_loss: 0.7445\n",
      "Epoch [   19/   50] | d_loss: 0.2219 | g_loss: 0.8868\n",
      "Epoch [   19/   50] | d_loss: 0.0903 | g_loss: 0.6097\n",
      "Epoch [   19/   50] | d_loss: 0.1343 | g_loss: 0.6412\n",
      "Epoch [   19/   50] | d_loss: 0.0816 | g_loss: 1.1630\n",
      "Epoch [   19/   50] | d_loss: 0.2243 | g_loss: 0.7141\n",
      "Epoch [   19/   50] | d_loss: 0.1067 | g_loss: 0.5733\n",
      "Epoch [   19/   50] | d_loss: 0.0810 | g_loss: 0.7550\n",
      "Epoch [   19/   50] | d_loss: 0.1104 | g_loss: 1.1718\n",
      "Epoch [   19/   50] | d_loss: 0.1177 | g_loss: 0.9199\n",
      "Epoch [   19/   50] | d_loss: 0.1910 | g_loss: 0.7894\n",
      "Epoch [   19/   50] | d_loss: 0.1071 | g_loss: 0.8686\n",
      "Epoch [   19/   50] | d_loss: 0.0813 | g_loss: 1.1912\n",
      "Epoch [   19/   50] | d_loss: 0.1144 | g_loss: 0.8835\n",
      "Epoch [   19/   50] | d_loss: 0.1807 | g_loss: 0.8167\n",
      "Epoch [   19/   50] | d_loss: 0.1349 | g_loss: 0.7806\n",
      "Epoch [   19/   50] | d_loss: 0.0906 | g_loss: 0.7055\n",
      "Epoch [   19/   50] | d_loss: 0.1767 | g_loss: 1.3355\n",
      "Epoch [   19/   50] | d_loss: 0.2404 | g_loss: 0.9196\n",
      "Epoch [   19/   50] | d_loss: 0.1324 | g_loss: 0.4325\n",
      "Epoch [   19/   50] | d_loss: 0.1013 | g_loss: 1.0236\n",
      "Epoch [   19/   50] | d_loss: 0.2335 | g_loss: 0.9775\n",
      "Epoch [   19/   50] | d_loss: 0.1468 | g_loss: 0.4372\n",
      "Epoch [   19/   50] | d_loss: 0.0969 | g_loss: 0.8509\n",
      "Epoch [   19/   50] | d_loss: 0.1108 | g_loss: 0.5408\n",
      "Epoch [   19/   50] | d_loss: 0.1871 | g_loss: 0.7131\n",
      "Epoch [   19/   50] | d_loss: 0.2060 | g_loss: 0.4830\n",
      "Epoch [   19/   50] | d_loss: 0.2310 | g_loss: 1.8066\n",
      "Epoch [   19/   50] | d_loss: 0.1694 | g_loss: 0.8530\n",
      "Epoch [   19/   50] | d_loss: 0.0918 | g_loss: 0.6896\n",
      "Epoch [   19/   50] | d_loss: 0.3830 | g_loss: 0.4018\n",
      "Epoch [   19/   50] | d_loss: 0.1044 | g_loss: 0.6764\n",
      "Epoch [   19/   50] | d_loss: 0.1585 | g_loss: 0.7384\n",
      "Epoch [   19/   50] | d_loss: 0.2192 | g_loss: 0.6072\n",
      "Epoch [   19/   50] | d_loss: 0.1126 | g_loss: 0.6314\n",
      "Epoch [   19/   50] | d_loss: 0.4041 | g_loss: 0.7950\n",
      "Epoch [   19/   50] | d_loss: 0.1209 | g_loss: 1.2605\n",
      "Epoch [   19/   50] | d_loss: 0.1965 | g_loss: 0.7208\n",
      "Epoch [   19/   50] | d_loss: 0.1479 | g_loss: 1.0729\n",
      "Epoch [   19/   50] | d_loss: 0.1496 | g_loss: 1.1909\n",
      "Epoch [   19/   50] | d_loss: 0.1821 | g_loss: 0.6744\n",
      "Epoch [   19/   50] | d_loss: 0.1581 | g_loss: 0.8585\n",
      "Epoch [   19/   50] | d_loss: 0.1117 | g_loss: 1.2584\n",
      "Epoch [   19/   50] | d_loss: 0.1253 | g_loss: 0.6938\n",
      "Epoch [   19/   50] | d_loss: 0.3206 | g_loss: 0.5978\n",
      "Epoch [   19/   50] | d_loss: 0.1090 | g_loss: 0.9803\n",
      "Epoch [   19/   50] | d_loss: 0.2191 | g_loss: 0.7663\n",
      "Epoch [   19/   50] | d_loss: 0.3171 | g_loss: 0.4154\n",
      "Epoch [   19/   50] | d_loss: 0.1693 | g_loss: 0.5947\n",
      "Epoch [   19/   50] | d_loss: 0.1448 | g_loss: 0.7699\n",
      "Epoch [   19/   50] | d_loss: 0.0732 | g_loss: 0.7942\n",
      "Epoch [   19/   50] | d_loss: 0.3387 | g_loss: 0.8888\n",
      "Epoch [   19/   50] | d_loss: 0.2583 | g_loss: 0.6921\n",
      "Epoch [   19/   50] | d_loss: 0.4441 | g_loss: 0.7341\n",
      "Epoch [   19/   50] | d_loss: 0.2615 | g_loss: 1.3814\n",
      "Epoch [   20/   50] | d_loss: 0.2676 | g_loss: 1.3167\n",
      "Epoch [   20/   50] | d_loss: 0.2328 | g_loss: 0.7025\n",
      "Epoch [   20/   50] | d_loss: 0.0968 | g_loss: 0.5724\n",
      "Epoch [   20/   50] | d_loss: 0.1833 | g_loss: 0.3735\n",
      "Epoch [   20/   50] | d_loss: 0.2282 | g_loss: 0.7082\n",
      "Epoch [   20/   50] | d_loss: 0.0838 | g_loss: 0.9231\n",
      "Epoch [   20/   50] | d_loss: 0.1172 | g_loss: 0.6535\n",
      "Epoch [   20/   50] | d_loss: 0.1958 | g_loss: 0.9364\n",
      "Epoch [   20/   50] | d_loss: 0.2277 | g_loss: 1.0056\n",
      "Epoch [   20/   50] | d_loss: 0.1578 | g_loss: 0.5321\n",
      "Epoch [   20/   50] | d_loss: 0.2784 | g_loss: 0.5821\n",
      "Epoch [   20/   50] | d_loss: 0.1052 | g_loss: 0.5631\n",
      "Epoch [   20/   50] | d_loss: 0.2028 | g_loss: 0.5218\n",
      "Epoch [   20/   50] | d_loss: 0.2074 | g_loss: 0.8011\n",
      "Epoch [   20/   50] | d_loss: 0.1115 | g_loss: 0.6773\n",
      "Epoch [   20/   50] | d_loss: 0.0992 | g_loss: 0.5418\n",
      "Epoch [   20/   50] | d_loss: 0.1605 | g_loss: 1.2567\n",
      "Epoch [   20/   50] | d_loss: 0.0986 | g_loss: 0.7782\n",
      "Epoch [   20/   50] | d_loss: 0.2515 | g_loss: 0.8075\n",
      "Epoch [   20/   50] | d_loss: 0.1486 | g_loss: 0.6732\n",
      "Epoch [   20/   50] | d_loss: 0.1292 | g_loss: 0.6617\n",
      "Epoch [   20/   50] | d_loss: 0.2190 | g_loss: 0.9130\n",
      "Epoch [   20/   50] | d_loss: 0.2370 | g_loss: 0.7667\n",
      "Epoch [   20/   50] | d_loss: 0.2497 | g_loss: 0.8480\n",
      "Epoch [   20/   50] | d_loss: 0.1635 | g_loss: 0.9114\n",
      "Epoch [   20/   50] | d_loss: 0.1850 | g_loss: 0.9099\n",
      "Epoch [   20/   50] | d_loss: 0.1744 | g_loss: 0.8371\n",
      "Epoch [   20/   50] | d_loss: 0.2407 | g_loss: 0.7131\n",
      "Epoch [   20/   50] | d_loss: 0.3160 | g_loss: 1.1323\n",
      "Epoch [   20/   50] | d_loss: 0.1148 | g_loss: 0.9384\n",
      "Epoch [   20/   50] | d_loss: 0.1287 | g_loss: 0.8571\n",
      "Epoch [   20/   50] | d_loss: 0.1125 | g_loss: 0.7716\n",
      "Epoch [   20/   50] | d_loss: 0.1114 | g_loss: 1.1187\n",
      "Epoch [   20/   50] | d_loss: 0.0908 | g_loss: 1.0150\n",
      "Epoch [   20/   50] | d_loss: 0.2293 | g_loss: 0.8626\n",
      "Epoch [   20/   50] | d_loss: 0.4621 | g_loss: 0.5428\n",
      "Epoch [   20/   50] | d_loss: 0.1330 | g_loss: 0.8388\n",
      "Epoch [   20/   50] | d_loss: 0.1134 | g_loss: 0.7443\n",
      "Epoch [   20/   50] | d_loss: 0.1837 | g_loss: 0.9013\n",
      "Epoch [   20/   50] | d_loss: 0.1278 | g_loss: 1.1350\n",
      "Epoch [   20/   50] | d_loss: 0.1394 | g_loss: 0.3523\n",
      "Epoch [   20/   50] | d_loss: 0.0851 | g_loss: 1.0333\n",
      "Epoch [   20/   50] | d_loss: 0.1503 | g_loss: 0.9126\n",
      "Epoch [   20/   50] | d_loss: 0.1519 | g_loss: 1.0101\n",
      "Epoch [   20/   50] | d_loss: 0.1543 | g_loss: 0.7232\n",
      "Epoch [   20/   50] | d_loss: 0.1144 | g_loss: 0.9099\n",
      "Epoch [   20/   50] | d_loss: 0.1789 | g_loss: 0.8681\n",
      "Epoch [   20/   50] | d_loss: 0.1029 | g_loss: 1.0524\n",
      "Epoch [   20/   50] | d_loss: 0.2959 | g_loss: 0.9372\n",
      "Epoch [   20/   50] | d_loss: 0.0484 | g_loss: 1.0435\n",
      "Epoch [   20/   50] | d_loss: 0.1633 | g_loss: 0.8272\n",
      "Epoch [   20/   50] | d_loss: 0.3213 | g_loss: 0.4260\n",
      "Epoch [   20/   50] | d_loss: 0.1765 | g_loss: 0.6349\n",
      "Epoch [   20/   50] | d_loss: 0.2170 | g_loss: 0.6994\n",
      "Epoch [   20/   50] | d_loss: 0.1501 | g_loss: 0.6131\n",
      "Epoch [   20/   50] | d_loss: 0.1938 | g_loss: 0.5587\n",
      "Epoch [   20/   50] | d_loss: 0.1035 | g_loss: 1.2041\n",
      "Epoch [   21/   50] | d_loss: 0.2430 | g_loss: 0.8597\n",
      "Epoch [   21/   50] | d_loss: 0.2814 | g_loss: 1.0365\n",
      "Epoch [   21/   50] | d_loss: 0.0772 | g_loss: 1.2229\n",
      "Epoch [   21/   50] | d_loss: 0.1357 | g_loss: 0.8012\n",
      "Epoch [   21/   50] | d_loss: 0.1661 | g_loss: 0.8336\n",
      "Epoch [   21/   50] | d_loss: 0.2065 | g_loss: 0.5001\n",
      "Epoch [   21/   50] | d_loss: 0.1182 | g_loss: 0.7329\n",
      "Epoch [   21/   50] | d_loss: 0.1433 | g_loss: 0.6428\n",
      "Epoch [   21/   50] | d_loss: 0.2770 | g_loss: 0.6713\n",
      "Epoch [   21/   50] | d_loss: 0.1926 | g_loss: 0.8761\n",
      "Epoch [   21/   50] | d_loss: 0.1423 | g_loss: 0.8203\n",
      "Epoch [   21/   50] | d_loss: 0.1351 | g_loss: 0.7421\n",
      "Epoch [   21/   50] | d_loss: 0.3502 | g_loss: 1.2178\n",
      "Epoch [   21/   50] | d_loss: 0.1507 | g_loss: 0.6961\n",
      "Epoch [   21/   50] | d_loss: 0.1956 | g_loss: 0.6590\n",
      "Epoch [   21/   50] | d_loss: 0.2224 | g_loss: 0.4106\n",
      "Epoch [   21/   50] | d_loss: 0.2196 | g_loss: 1.3170\n",
      "Epoch [   21/   50] | d_loss: 0.1096 | g_loss: 0.2526\n",
      "Epoch [   21/   50] | d_loss: 0.1643 | g_loss: 0.7053\n",
      "Epoch [   21/   50] | d_loss: 0.1231 | g_loss: 0.9559\n",
      "Epoch [   21/   50] | d_loss: 0.1886 | g_loss: 0.5261\n",
      "Epoch [   21/   50] | d_loss: 0.1899 | g_loss: 0.9762\n",
      "Epoch [   21/   50] | d_loss: 0.0726 | g_loss: 1.0739\n",
      "Epoch [   21/   50] | d_loss: 0.1794 | g_loss: 0.4660\n",
      "Epoch [   21/   50] | d_loss: 0.2984 | g_loss: 0.9503\n",
      "Epoch [   21/   50] | d_loss: 0.0886 | g_loss: 1.1396\n",
      "Epoch [   21/   50] | d_loss: 0.1799 | g_loss: 1.1011\n",
      "Epoch [   21/   50] | d_loss: 0.2079 | g_loss: 0.8304\n",
      "Epoch [   21/   50] | d_loss: 0.1501 | g_loss: 0.5114\n",
      "Epoch [   21/   50] | d_loss: 0.2860 | g_loss: 0.6261\n",
      "Epoch [   21/   50] | d_loss: 0.1094 | g_loss: 0.9025\n",
      "Epoch [   21/   50] | d_loss: 0.0879 | g_loss: 1.0118\n",
      "Epoch [   21/   50] | d_loss: 0.2661 | g_loss: 0.8829\n",
      "Epoch [   21/   50] | d_loss: 0.1587 | g_loss: 0.3876\n",
      "Epoch [   21/   50] | d_loss: 0.2415 | g_loss: 0.7415\n",
      "Epoch [   21/   50] | d_loss: 0.1452 | g_loss: 0.5605\n",
      "Epoch [   21/   50] | d_loss: 0.2962 | g_loss: 0.6605\n",
      "Epoch [   21/   50] | d_loss: 0.1400 | g_loss: 0.6229\n",
      "Epoch [   21/   50] | d_loss: 0.1724 | g_loss: 0.6754\n",
      "Epoch [   21/   50] | d_loss: 0.1496 | g_loss: 0.7312\n",
      "Epoch [   21/   50] | d_loss: 0.1715 | g_loss: 0.5661\n",
      "Epoch [   21/   50] | d_loss: 0.1150 | g_loss: 0.8791\n",
      "Epoch [   21/   50] | d_loss: 0.1921 | g_loss: 0.7146\n",
      "Epoch [   21/   50] | d_loss: 0.3572 | g_loss: 0.3386\n",
      "Epoch [   21/   50] | d_loss: 0.2204 | g_loss: 0.7413\n",
      "Epoch [   21/   50] | d_loss: 0.1056 | g_loss: 0.5047\n",
      "Epoch [   21/   50] | d_loss: 0.2449 | g_loss: 0.7040\n",
      "Epoch [   21/   50] | d_loss: 0.0831 | g_loss: 0.8752\n",
      "Epoch [   21/   50] | d_loss: 0.1117 | g_loss: 0.6817\n",
      "Epoch [   21/   50] | d_loss: 0.2683 | g_loss: 0.6528\n",
      "Epoch [   21/   50] | d_loss: 0.2025 | g_loss: 0.9899\n",
      "Epoch [   21/   50] | d_loss: 0.1880 | g_loss: 0.7949\n",
      "Epoch [   21/   50] | d_loss: 0.2237 | g_loss: 0.8300\n",
      "Epoch [   21/   50] | d_loss: 0.1280 | g_loss: 0.5051\n",
      "Epoch [   21/   50] | d_loss: 0.1075 | g_loss: 0.9542\n",
      "Epoch [   21/   50] | d_loss: 0.0712 | g_loss: 1.3467\n",
      "Epoch [   21/   50] | d_loss: 0.1307 | g_loss: 0.7986\n",
      "Epoch [   22/   50] | d_loss: 0.2257 | g_loss: 0.5699\n",
      "Epoch [   22/   50] | d_loss: 0.0886 | g_loss: 1.2121\n",
      "Epoch [   22/   50] | d_loss: 0.1364 | g_loss: 0.8903\n",
      "Epoch [   22/   50] | d_loss: 0.0816 | g_loss: 1.1152\n",
      "Epoch [   22/   50] | d_loss: 0.1460 | g_loss: 0.9601\n",
      "Epoch [   22/   50] | d_loss: 0.2859 | g_loss: 0.9771\n",
      "Epoch [   22/   50] | d_loss: 0.1981 | g_loss: 0.4691\n",
      "Epoch [   22/   50] | d_loss: 0.2860 | g_loss: 0.7324\n",
      "Epoch [   22/   50] | d_loss: 0.0836 | g_loss: 0.5644\n",
      "Epoch [   22/   50] | d_loss: 0.1022 | g_loss: 0.9147\n",
      "Epoch [   22/   50] | d_loss: 0.1377 | g_loss: 0.7668\n",
      "Epoch [   22/   50] | d_loss: 0.1468 | g_loss: 0.6175\n",
      "Epoch [   22/   50] | d_loss: 0.2111 | g_loss: 0.9313\n",
      "Epoch [   22/   50] | d_loss: 0.1251 | g_loss: 1.0041\n",
      "Epoch [   22/   50] | d_loss: 0.1323 | g_loss: 0.8193\n",
      "Epoch [   22/   50] | d_loss: 0.1090 | g_loss: 1.1179\n",
      "Epoch [   22/   50] | d_loss: 0.1506 | g_loss: 0.5058\n",
      "Epoch [   22/   50] | d_loss: 0.2118 | g_loss: 0.6691\n",
      "Epoch [   22/   50] | d_loss: 0.1134 | g_loss: 0.5219\n",
      "Epoch [   22/   50] | d_loss: 0.1783 | g_loss: 0.7608\n",
      "Epoch [   22/   50] | d_loss: 0.1025 | g_loss: 1.0239\n",
      "Epoch [   22/   50] | d_loss: 0.1806 | g_loss: 0.7079\n",
      "Epoch [   22/   50] | d_loss: 0.2577 | g_loss: 0.9509\n",
      "Epoch [   22/   50] | d_loss: 0.1728 | g_loss: 0.6626\n",
      "Epoch [   22/   50] | d_loss: 0.1389 | g_loss: 0.6198\n",
      "Epoch [   22/   50] | d_loss: 0.3532 | g_loss: 0.6497\n",
      "Epoch [   22/   50] | d_loss: 0.3459 | g_loss: 0.6760\n",
      "Epoch [   22/   50] | d_loss: 0.1876 | g_loss: 0.6897\n",
      "Epoch [   22/   50] | d_loss: 0.1483 | g_loss: 1.5658\n",
      "Epoch [   22/   50] | d_loss: 0.1616 | g_loss: 0.6997\n",
      "Epoch [   22/   50] | d_loss: 0.0793 | g_loss: 0.5408\n",
      "Epoch [   22/   50] | d_loss: 0.1321 | g_loss: 0.6247\n",
      "Epoch [   22/   50] | d_loss: 0.2050 | g_loss: 0.3742\n",
      "Epoch [   22/   50] | d_loss: 0.1311 | g_loss: 0.8837\n",
      "Epoch [   22/   50] | d_loss: 0.0808 | g_loss: 0.5782\n",
      "Epoch [   22/   50] | d_loss: 0.1630 | g_loss: 0.7352\n",
      "Epoch [   22/   50] | d_loss: 0.0944 | g_loss: 0.8015\n",
      "Epoch [   22/   50] | d_loss: 0.2153 | g_loss: 0.4902\n",
      "Epoch [   22/   50] | d_loss: 0.1311 | g_loss: 0.5609\n",
      "Epoch [   22/   50] | d_loss: 0.0947 | g_loss: 1.3199\n",
      "Epoch [   22/   50] | d_loss: 0.1116 | g_loss: 1.2103\n",
      "Epoch [   22/   50] | d_loss: 0.1749 | g_loss: 0.7134\n",
      "Epoch [   22/   50] | d_loss: 0.1001 | g_loss: 0.8474\n",
      "Epoch [   22/   50] | d_loss: 0.0776 | g_loss: 0.8652\n",
      "Epoch [   22/   50] | d_loss: 0.0424 | g_loss: 0.8941\n",
      "Epoch [   22/   50] | d_loss: 0.1304 | g_loss: 0.4013\n",
      "Epoch [   22/   50] | d_loss: 0.0849 | g_loss: 1.0502\n",
      "Epoch [   22/   50] | d_loss: 0.1712 | g_loss: 0.6106\n",
      "Epoch [   22/   50] | d_loss: 0.1411 | g_loss: 0.8187\n",
      "Epoch [   22/   50] | d_loss: 0.1215 | g_loss: 0.7104\n",
      "Epoch [   22/   50] | d_loss: 0.0863 | g_loss: 0.9307\n",
      "Epoch [   22/   50] | d_loss: 0.0758 | g_loss: 1.2214\n",
      "Epoch [   22/   50] | d_loss: 0.1218 | g_loss: 0.8446\n",
      "Epoch [   22/   50] | d_loss: 0.1900 | g_loss: 0.8627\n",
      "Epoch [   22/   50] | d_loss: 0.1188 | g_loss: 0.4764\n",
      "Epoch [   22/   50] | d_loss: 0.1255 | g_loss: 0.5675\n",
      "Epoch [   22/   50] | d_loss: 0.1103 | g_loss: 0.8168\n",
      "Epoch [   23/   50] | d_loss: 0.3327 | g_loss: 1.0904\n",
      "Epoch [   23/   50] | d_loss: 0.1186 | g_loss: 0.8344\n",
      "Epoch [   23/   50] | d_loss: 0.1457 | g_loss: 0.8886\n",
      "Epoch [   23/   50] | d_loss: 0.1180 | g_loss: 0.6791\n",
      "Epoch [   23/   50] | d_loss: 0.0918 | g_loss: 0.9238\n",
      "Epoch [   23/   50] | d_loss: 0.1887 | g_loss: 0.7126\n",
      "Epoch [   23/   50] | d_loss: 0.1169 | g_loss: 0.8403\n",
      "Epoch [   23/   50] | d_loss: 0.2466 | g_loss: 0.3178\n",
      "Epoch [   23/   50] | d_loss: 0.2221 | g_loss: 1.0497\n",
      "Epoch [   23/   50] | d_loss: 0.0687 | g_loss: 0.8669\n",
      "Epoch [   23/   50] | d_loss: 0.0927 | g_loss: 1.2496\n",
      "Epoch [   23/   50] | d_loss: 0.3422 | g_loss: 0.4956\n",
      "Epoch [   23/   50] | d_loss: 0.0993 | g_loss: 1.3514\n",
      "Epoch [   23/   50] | d_loss: 0.1214 | g_loss: 0.7440\n",
      "Epoch [   23/   50] | d_loss: 0.1006 | g_loss: 0.6024\n",
      "Epoch [   23/   50] | d_loss: 0.0899 | g_loss: 0.6974\n",
      "Epoch [   23/   50] | d_loss: 0.2121 | g_loss: 0.9816\n",
      "Epoch [   23/   50] | d_loss: 0.1278 | g_loss: 0.9947\n",
      "Epoch [   23/   50] | d_loss: 0.2170 | g_loss: 0.5411\n",
      "Epoch [   23/   50] | d_loss: 0.1308 | g_loss: 0.5383\n",
      "Epoch [   23/   50] | d_loss: 0.1664 | g_loss: 0.9325\n",
      "Epoch [   23/   50] | d_loss: 0.2005 | g_loss: 1.1480\n",
      "Epoch [   23/   50] | d_loss: 0.2417 | g_loss: 0.5181\n",
      "Epoch [   23/   50] | d_loss: 0.1136 | g_loss: 0.5772\n",
      "Epoch [   23/   50] | d_loss: 0.0934 | g_loss: 0.5570\n",
      "Epoch [   23/   50] | d_loss: 0.2092 | g_loss: 0.9958\n",
      "Epoch [   23/   50] | d_loss: 0.1233 | g_loss: 0.8427\n",
      "Epoch [   23/   50] | d_loss: 0.1034 | g_loss: 0.8626\n",
      "Epoch [   23/   50] | d_loss: 0.1811 | g_loss: 0.4980\n",
      "Epoch [   23/   50] | d_loss: 0.1122 | g_loss: 1.2327\n",
      "Epoch [   23/   50] | d_loss: 0.1215 | g_loss: 0.8618\n",
      "Epoch [   23/   50] | d_loss: 0.0813 | g_loss: 1.1867\n",
      "Epoch [   23/   50] | d_loss: 0.3430 | g_loss: 0.7143\n",
      "Epoch [   23/   50] | d_loss: 0.2000 | g_loss: 0.7745\n",
      "Epoch [   23/   50] | d_loss: 0.1215 | g_loss: 0.9330\n",
      "Epoch [   23/   50] | d_loss: 0.0712 | g_loss: 0.5818\n",
      "Epoch [   23/   50] | d_loss: 0.1304 | g_loss: 0.7142\n",
      "Epoch [   23/   50] | d_loss: 0.2265 | g_loss: 0.7246\n",
      "Epoch [   23/   50] | d_loss: 0.1572 | g_loss: 0.9614\n",
      "Epoch [   23/   50] | d_loss: 0.1168 | g_loss: 0.6674\n",
      "Epoch [   23/   50] | d_loss: 0.2125 | g_loss: 0.5518\n",
      "Epoch [   23/   50] | d_loss: 0.1016 | g_loss: 1.0597\n",
      "Epoch [   23/   50] | d_loss: 0.1828 | g_loss: 1.1951\n",
      "Epoch [   23/   50] | d_loss: 0.3927 | g_loss: 0.5725\n",
      "Epoch [   23/   50] | d_loss: 0.2124 | g_loss: 0.7181\n",
      "Epoch [   23/   50] | d_loss: 0.0874 | g_loss: 0.7483\n",
      "Epoch [   23/   50] | d_loss: 0.3793 | g_loss: 0.8390\n",
      "Epoch [   23/   50] | d_loss: 0.1440 | g_loss: 0.5226\n",
      "Epoch [   23/   50] | d_loss: 0.0861 | g_loss: 0.6690\n",
      "Epoch [   23/   50] | d_loss: 0.0715 | g_loss: 1.0755\n",
      "Epoch [   23/   50] | d_loss: 0.1301 | g_loss: 0.6481\n",
      "Epoch [   23/   50] | d_loss: 0.1335 | g_loss: 0.6057\n",
      "Epoch [   23/   50] | d_loss: 0.1138 | g_loss: 0.7448\n",
      "Epoch [   23/   50] | d_loss: 0.0612 | g_loss: 0.5966\n",
      "Epoch [   23/   50] | d_loss: 0.1554 | g_loss: 0.7753\n",
      "Epoch [   23/   50] | d_loss: 0.1598 | g_loss: 1.0420\n",
      "Epoch [   23/   50] | d_loss: 0.4421 | g_loss: 0.4298\n",
      "Epoch [   24/   50] | d_loss: 0.2655 | g_loss: 0.8142\n",
      "Epoch [   24/   50] | d_loss: 0.1303 | g_loss: 0.3842\n",
      "Epoch [   24/   50] | d_loss: 0.1178 | g_loss: 0.7596\n",
      "Epoch [   24/   50] | d_loss: 0.1270 | g_loss: 0.6356\n",
      "Epoch [   24/   50] | d_loss: 0.2879 | g_loss: 1.0043\n",
      "Epoch [   24/   50] | d_loss: 0.0800 | g_loss: 0.8087\n",
      "Epoch [   24/   50] | d_loss: 0.2657 | g_loss: 0.9094\n",
      "Epoch [   24/   50] | d_loss: 0.2449 | g_loss: 0.5358\n",
      "Epoch [   24/   50] | d_loss: 0.1338 | g_loss: 0.7565\n",
      "Epoch [   24/   50] | d_loss: 0.0835 | g_loss: 0.9353\n",
      "Epoch [   24/   50] | d_loss: 0.1393 | g_loss: 0.7348\n",
      "Epoch [   24/   50] | d_loss: 0.0960 | g_loss: 0.5199\n",
      "Epoch [   24/   50] | d_loss: 0.1438 | g_loss: 0.9532\n",
      "Epoch [   24/   50] | d_loss: 0.1423 | g_loss: 0.4692\n",
      "Epoch [   24/   50] | d_loss: 0.1416 | g_loss: 0.8428\n",
      "Epoch [   24/   50] | d_loss: 0.0917 | g_loss: 1.2810\n",
      "Epoch [   24/   50] | d_loss: 0.1634 | g_loss: 0.8038\n",
      "Epoch [   24/   50] | d_loss: 0.1367 | g_loss: 0.8286\n",
      "Epoch [   24/   50] | d_loss: 0.0745 | g_loss: 0.7273\n",
      "Epoch [   24/   50] | d_loss: 0.3191 | g_loss: 0.6656\n",
      "Epoch [   24/   50] | d_loss: 0.1014 | g_loss: 0.7575\n",
      "Epoch [   24/   50] | d_loss: 0.1604 | g_loss: 0.6611\n",
      "Epoch [   24/   50] | d_loss: 0.1942 | g_loss: 0.6618\n",
      "Epoch [   24/   50] | d_loss: 0.1685 | g_loss: 0.7058\n",
      "Epoch [   24/   50] | d_loss: 0.0637 | g_loss: 0.9285\n",
      "Epoch [   24/   50] | d_loss: 0.1616 | g_loss: 0.5358\n",
      "Epoch [   24/   50] | d_loss: 0.1167 | g_loss: 0.7638\n",
      "Epoch [   24/   50] | d_loss: 0.1366 | g_loss: 0.8266\n",
      "Epoch [   24/   50] | d_loss: 0.0963 | g_loss: 1.0693\n",
      "Epoch [   24/   50] | d_loss: 0.4122 | g_loss: 0.5509\n",
      "Epoch [   24/   50] | d_loss: 0.1663 | g_loss: 0.3909\n",
      "Epoch [   24/   50] | d_loss: 0.0934 | g_loss: 0.9604\n",
      "Epoch [   24/   50] | d_loss: 0.1602 | g_loss: 0.8109\n",
      "Epoch [   24/   50] | d_loss: 0.1378 | g_loss: 0.9871\n",
      "Epoch [   24/   50] | d_loss: 0.1558 | g_loss: 0.4837\n",
      "Epoch [   24/   50] | d_loss: 0.1355 | g_loss: 0.7328\n",
      "Epoch [   24/   50] | d_loss: 0.1425 | g_loss: 0.8630\n",
      "Epoch [   24/   50] | d_loss: 0.0637 | g_loss: 1.2210\n",
      "Epoch [   24/   50] | d_loss: 0.0740 | g_loss: 0.7223\n",
      "Epoch [   24/   50] | d_loss: 0.1197 | g_loss: 0.9644\n",
      "Epoch [   24/   50] | d_loss: 0.0940 | g_loss: 0.4309\n",
      "Epoch [   24/   50] | d_loss: 0.0774 | g_loss: 0.9779\n",
      "Epoch [   24/   50] | d_loss: 0.1612 | g_loss: 0.3690\n",
      "Epoch [   24/   50] | d_loss: 0.1689 | g_loss: 0.3700\n",
      "Epoch [   24/   50] | d_loss: 0.0805 | g_loss: 0.8126\n",
      "Epoch [   24/   50] | d_loss: 0.2537 | g_loss: 0.8230\n",
      "Epoch [   24/   50] | d_loss: 0.1337 | g_loss: 0.9616\n",
      "Epoch [   24/   50] | d_loss: 0.1016 | g_loss: 1.0035\n",
      "Epoch [   24/   50] | d_loss: 0.0978 | g_loss: 0.8689\n",
      "Epoch [   24/   50] | d_loss: 0.1553 | g_loss: 0.4782\n",
      "Epoch [   24/   50] | d_loss: 0.1403 | g_loss: 0.8149\n",
      "Epoch [   24/   50] | d_loss: 0.1828 | g_loss: 0.3783\n",
      "Epoch [   24/   50] | d_loss: 0.0947 | g_loss: 0.9028\n",
      "Epoch [   24/   50] | d_loss: 0.1116 | g_loss: 0.8244\n",
      "Epoch [   24/   50] | d_loss: 0.1340 | g_loss: 0.7135\n",
      "Epoch [   24/   50] | d_loss: 0.0729 | g_loss: 1.1072\n",
      "Epoch [   24/   50] | d_loss: 0.1032 | g_loss: 0.9909\n",
      "Epoch [   25/   50] | d_loss: 0.2555 | g_loss: 0.6539\n",
      "Epoch [   25/   50] | d_loss: 0.0813 | g_loss: 0.5788\n",
      "Epoch [   25/   50] | d_loss: 0.0996 | g_loss: 0.7851\n",
      "Epoch [   25/   50] | d_loss: 0.1106 | g_loss: 0.8913\n",
      "Epoch [   25/   50] | d_loss: 0.0676 | g_loss: 0.8021\n",
      "Epoch [   25/   50] | d_loss: 0.0858 | g_loss: 0.7110\n",
      "Epoch [   25/   50] | d_loss: 0.1094 | g_loss: 1.2874\n",
      "Epoch [   25/   50] | d_loss: 0.1021 | g_loss: 0.8820\n",
      "Epoch [   25/   50] | d_loss: 0.0953 | g_loss: 0.5568\n",
      "Epoch [   25/   50] | d_loss: 0.0856 | g_loss: 0.9664\n",
      "Epoch [   25/   50] | d_loss: 0.1834 | g_loss: 0.8329\n",
      "Epoch [   25/   50] | d_loss: 0.1064 | g_loss: 0.5574\n",
      "Epoch [   25/   50] | d_loss: 0.1196 | g_loss: 0.8489\n",
      "Epoch [   25/   50] | d_loss: 0.0684 | g_loss: 0.9552\n",
      "Epoch [   25/   50] | d_loss: 0.1168 | g_loss: 0.6670\n",
      "Epoch [   25/   50] | d_loss: 0.1657 | g_loss: 0.7904\n",
      "Epoch [   25/   50] | d_loss: 0.3619 | g_loss: 0.4053\n",
      "Epoch [   25/   50] | d_loss: 0.1919 | g_loss: 1.1128\n",
      "Epoch [   25/   50] | d_loss: 0.1008 | g_loss: 0.5920\n",
      "Epoch [   25/   50] | d_loss: 0.1953 | g_loss: 0.7142\n",
      "Epoch [   25/   50] | d_loss: 0.2314 | g_loss: 0.4992\n",
      "Epoch [   25/   50] | d_loss: 0.0858 | g_loss: 0.8661\n",
      "Epoch [   25/   50] | d_loss: 0.1003 | g_loss: 0.5924\n",
      "Epoch [   25/   50] | d_loss: 0.1084 | g_loss: 0.6132\n",
      "Epoch [   25/   50] | d_loss: 0.2477 | g_loss: 0.2471\n",
      "Epoch [   25/   50] | d_loss: 0.1053 | g_loss: 0.4001\n",
      "Epoch [   25/   50] | d_loss: 0.0730 | g_loss: 1.1192\n",
      "Epoch [   25/   50] | d_loss: 0.0640 | g_loss: 0.8905\n",
      "Epoch [   25/   50] | d_loss: 0.1869 | g_loss: 0.5421\n",
      "Epoch [   25/   50] | d_loss: 0.1186 | g_loss: 1.0196\n",
      "Epoch [   25/   50] | d_loss: 0.1551 | g_loss: 0.7307\n",
      "Epoch [   25/   50] | d_loss: 0.1653 | g_loss: 0.6400\n",
      "Epoch [   25/   50] | d_loss: 0.1434 | g_loss: 0.4666\n",
      "Epoch [   25/   50] | d_loss: 0.1783 | g_loss: 0.7430\n",
      "Epoch [   25/   50] | d_loss: 0.1170 | g_loss: 0.7877\n",
      "Epoch [   25/   50] | d_loss: 0.1640 | g_loss: 0.8798\n",
      "Epoch [   25/   50] | d_loss: 0.1291 | g_loss: 0.5851\n",
      "Epoch [   25/   50] | d_loss: 0.1826 | g_loss: 0.6092\n",
      "Epoch [   25/   50] | d_loss: 0.0855 | g_loss: 1.4856\n",
      "Epoch [   25/   50] | d_loss: 0.0865 | g_loss: 0.4134\n",
      "Epoch [   25/   50] | d_loss: 0.1123 | g_loss: 1.1239\n",
      "Epoch [   25/   50] | d_loss: 0.1355 | g_loss: 0.8685\n",
      "Epoch [   25/   50] | d_loss: 0.1298 | g_loss: 0.9660\n",
      "Epoch [   25/   50] | d_loss: 0.1049 | g_loss: 0.5322\n",
      "Epoch [   25/   50] | d_loss: 0.1868 | g_loss: 0.9267\n",
      "Epoch [   25/   50] | d_loss: 0.1143 | g_loss: 1.1119\n",
      "Epoch [   25/   50] | d_loss: 0.1839 | g_loss: 0.7221\n",
      "Epoch [   25/   50] | d_loss: 0.1042 | g_loss: 0.9370\n",
      "Epoch [   25/   50] | d_loss: 0.1300 | g_loss: 0.7102\n",
      "Epoch [   25/   50] | d_loss: 0.1550 | g_loss: 0.4654\n",
      "Epoch [   25/   50] | d_loss: 0.1075 | g_loss: 0.7915\n",
      "Epoch [   25/   50] | d_loss: 0.0947 | g_loss: 0.7745\n",
      "Epoch [   25/   50] | d_loss: 0.2461 | g_loss: 0.9563\n",
      "Epoch [   25/   50] | d_loss: 0.1615 | g_loss: 0.8980\n",
      "Epoch [   25/   50] | d_loss: 0.1698 | g_loss: 0.4520\n",
      "Epoch [   25/   50] | d_loss: 0.1443 | g_loss: 0.9596\n",
      "Epoch [   25/   50] | d_loss: 0.1217 | g_loss: 0.7120\n",
      "Epoch [   26/   50] | d_loss: 0.1834 | g_loss: 0.5778\n",
      "Epoch [   26/   50] | d_loss: 0.1100 | g_loss: 0.6576\n",
      "Epoch [   26/   50] | d_loss: 0.1893 | g_loss: 0.7879\n",
      "Epoch [   26/   50] | d_loss: 0.1522 | g_loss: 1.0419\n",
      "Epoch [   26/   50] | d_loss: 0.1859 | g_loss: 0.6298\n",
      "Epoch [   26/   50] | d_loss: 0.0689 | g_loss: 1.0572\n",
      "Epoch [   26/   50] | d_loss: 0.2413 | g_loss: 0.6679\n",
      "Epoch [   26/   50] | d_loss: 0.1027 | g_loss: 0.9914\n",
      "Epoch [   26/   50] | d_loss: 0.0964 | g_loss: 1.2447\n",
      "Epoch [   26/   50] | d_loss: 0.2177 | g_loss: 0.7375\n",
      "Epoch [   26/   50] | d_loss: 0.1059 | g_loss: 1.1319\n",
      "Epoch [   26/   50] | d_loss: 0.0952 | g_loss: 1.0607\n",
      "Epoch [   26/   50] | d_loss: 0.0888 | g_loss: 0.9508\n",
      "Epoch [   26/   50] | d_loss: 0.1124 | g_loss: 0.8971\n",
      "Epoch [   26/   50] | d_loss: 0.1105 | g_loss: 1.0140\n",
      "Epoch [   26/   50] | d_loss: 0.1725 | g_loss: 0.7287\n",
      "Epoch [   26/   50] | d_loss: 0.0739 | g_loss: 0.4883\n",
      "Epoch [   26/   50] | d_loss: 0.1907 | g_loss: 0.5839\n",
      "Epoch [   26/   50] | d_loss: 0.0803 | g_loss: 0.7905\n",
      "Epoch [   26/   50] | d_loss: 0.2892 | g_loss: 0.8385\n",
      "Epoch [   26/   50] | d_loss: 0.1826 | g_loss: 0.7292\n",
      "Epoch [   26/   50] | d_loss: 0.1257 | g_loss: 1.1875\n",
      "Epoch [   26/   50] | d_loss: 0.1901 | g_loss: 0.4848\n",
      "Epoch [   26/   50] | d_loss: 0.1071 | g_loss: 0.6133\n",
      "Epoch [   26/   50] | d_loss: 0.1441 | g_loss: 0.9192\n",
      "Epoch [   26/   50] | d_loss: 0.1101 | g_loss: 0.7950\n",
      "Epoch [   26/   50] | d_loss: 0.1987 | g_loss: 0.5327\n",
      "Epoch [   26/   50] | d_loss: 0.0876 | g_loss: 0.8462\n",
      "Epoch [   26/   50] | d_loss: 0.0759 | g_loss: 1.1977\n",
      "Epoch [   26/   50] | d_loss: 0.1250 | g_loss: 0.7570\n",
      "Epoch [   26/   50] | d_loss: 0.2010 | g_loss: 0.4706\n",
      "Epoch [   26/   50] | d_loss: 0.1438 | g_loss: 0.7523\n",
      "Epoch [   26/   50] | d_loss: 0.1138 | g_loss: 0.8672\n",
      "Epoch [   26/   50] | d_loss: 0.1207 | g_loss: 0.6762\n",
      "Epoch [   26/   50] | d_loss: 0.0816 | g_loss: 0.5244\n",
      "Epoch [   26/   50] | d_loss: 0.0838 | g_loss: 0.7807\n",
      "Epoch [   26/   50] | d_loss: 0.0989 | g_loss: 1.0129\n",
      "Epoch [   26/   50] | d_loss: 0.1573 | g_loss: 0.6971\n",
      "Epoch [   26/   50] | d_loss: 0.1025 | g_loss: 0.7270\n",
      "Epoch [   26/   50] | d_loss: 0.1493 | g_loss: 0.8250\n",
      "Epoch [   26/   50] | d_loss: 0.2147 | g_loss: 0.5515\n",
      "Epoch [   26/   50] | d_loss: 0.2143 | g_loss: 1.0221\n",
      "Epoch [   26/   50] | d_loss: 0.0958 | g_loss: 0.4331\n",
      "Epoch [   26/   50] | d_loss: 0.1549 | g_loss: 0.5589\n",
      "Epoch [   26/   50] | d_loss: 0.1223 | g_loss: 0.9230\n",
      "Epoch [   26/   50] | d_loss: 0.1832 | g_loss: 0.6944\n",
      "Epoch [   26/   50] | d_loss: 0.1297 | g_loss: 0.7155\n",
      "Epoch [   26/   50] | d_loss: 0.1049 | g_loss: 0.8100\n",
      "Epoch [   26/   50] | d_loss: 0.1556 | g_loss: 1.6938\n",
      "Epoch [   26/   50] | d_loss: 0.1400 | g_loss: 0.8264\n",
      "Epoch [   26/   50] | d_loss: 0.1182 | g_loss: 0.8931\n",
      "Epoch [   26/   50] | d_loss: 0.0829 | g_loss: 0.8779\n",
      "Epoch [   26/   50] | d_loss: 0.0695 | g_loss: 1.0157\n",
      "Epoch [   26/   50] | d_loss: 0.1500 | g_loss: 0.7012\n",
      "Epoch [   26/   50] | d_loss: 0.1819 | g_loss: 0.7990\n",
      "Epoch [   26/   50] | d_loss: 0.1463 | g_loss: 0.8884\n",
      "Epoch [   26/   50] | d_loss: 0.1931 | g_loss: 0.3503\n",
      "Epoch [   27/   50] | d_loss: 0.2186 | g_loss: 0.5432\n",
      "Epoch [   27/   50] | d_loss: 0.0753 | g_loss: 0.6816\n",
      "Epoch [   27/   50] | d_loss: 0.3333 | g_loss: 0.6823\n",
      "Epoch [   27/   50] | d_loss: 0.1639 | g_loss: 0.5371\n",
      "Epoch [   27/   50] | d_loss: 0.2243 | g_loss: 1.3610\n",
      "Epoch [   27/   50] | d_loss: 0.1353 | g_loss: 0.6137\n",
      "Epoch [   27/   50] | d_loss: 0.0682 | g_loss: 1.3620\n",
      "Epoch [   27/   50] | d_loss: 0.1984 | g_loss: 0.4537\n",
      "Epoch [   27/   50] | d_loss: 0.1393 | g_loss: 0.8644\n",
      "Epoch [   27/   50] | d_loss: 0.0961 | g_loss: 0.6061\n",
      "Epoch [   27/   50] | d_loss: 0.1281 | g_loss: 0.6114\n",
      "Epoch [   27/   50] | d_loss: 0.1041 | g_loss: 1.0557\n",
      "Epoch [   27/   50] | d_loss: 0.1729 | g_loss: 0.6190\n",
      "Epoch [   27/   50] | d_loss: 0.0807 | g_loss: 0.6590\n",
      "Epoch [   27/   50] | d_loss: 0.1949 | g_loss: 0.4791\n",
      "Epoch [   27/   50] | d_loss: 0.1073 | g_loss: 0.6720\n",
      "Epoch [   27/   50] | d_loss: 0.1243 | g_loss: 0.9215\n",
      "Epoch [   27/   50] | d_loss: 0.4969 | g_loss: 0.6549\n",
      "Epoch [   27/   50] | d_loss: 0.1895 | g_loss: 0.5449\n",
      "Epoch [   27/   50] | d_loss: 0.1126 | g_loss: 0.8428\n",
      "Epoch [   27/   50] | d_loss: 0.3569 | g_loss: 0.5172\n",
      "Epoch [   27/   50] | d_loss: 0.1694 | g_loss: 0.7547\n",
      "Epoch [   27/   50] | d_loss: 0.1062 | g_loss: 0.5651\n",
      "Epoch [   27/   50] | d_loss: 0.2653 | g_loss: 0.5076\n",
      "Epoch [   27/   50] | d_loss: 0.1284 | g_loss: 0.6988\n",
      "Epoch [   27/   50] | d_loss: 0.1099 | g_loss: 0.7310\n",
      "Epoch [   27/   50] | d_loss: 0.0553 | g_loss: 0.9353\n",
      "Epoch [   27/   50] | d_loss: 0.0988 | g_loss: 0.6308\n",
      "Epoch [   27/   50] | d_loss: 0.0948 | g_loss: 1.0282\n",
      "Epoch [   27/   50] | d_loss: 0.1424 | g_loss: 0.7312\n",
      "Epoch [   27/   50] | d_loss: 0.1684 | g_loss: 0.6486\n",
      "Epoch [   27/   50] | d_loss: 0.0928 | g_loss: 0.7988\n",
      "Epoch [   27/   50] | d_loss: 0.1769 | g_loss: 0.3797\n",
      "Epoch [   27/   50] | d_loss: 0.0827 | g_loss: 0.7104\n",
      "Epoch [   27/   50] | d_loss: 0.1576 | g_loss: 0.7408\n",
      "Epoch [   27/   50] | d_loss: 0.0710 | g_loss: 0.7914\n",
      "Epoch [   27/   50] | d_loss: 0.1587 | g_loss: 0.8540\n",
      "Epoch [   27/   50] | d_loss: 0.2365 | g_loss: 0.6991\n",
      "Epoch [   27/   50] | d_loss: 0.1320 | g_loss: 0.8113\n",
      "Epoch [   27/   50] | d_loss: 0.1430 | g_loss: 1.0271\n",
      "Epoch [   27/   50] | d_loss: 0.0440 | g_loss: 1.2450\n",
      "Epoch [   27/   50] | d_loss: 0.1510 | g_loss: 0.8099\n",
      "Epoch [   27/   50] | d_loss: 0.1568 | g_loss: 0.3986\n",
      "Epoch [   27/   50] | d_loss: 0.0886 | g_loss: 0.8511\n",
      "Epoch [   27/   50] | d_loss: 0.1426 | g_loss: 0.6820\n",
      "Epoch [   27/   50] | d_loss: 0.1967 | g_loss: 0.6063\n",
      "Epoch [   27/   50] | d_loss: 0.1307 | g_loss: 0.7501\n",
      "Epoch [   27/   50] | d_loss: 0.1897 | g_loss: 0.9004\n",
      "Epoch [   27/   50] | d_loss: 0.1745 | g_loss: 0.4313\n",
      "Epoch [   27/   50] | d_loss: 0.1012 | g_loss: 1.0554\n",
      "Epoch [   27/   50] | d_loss: 0.1376 | g_loss: 0.7161\n",
      "Epoch [   27/   50] | d_loss: 0.0928 | g_loss: 0.7949\n",
      "Epoch [   27/   50] | d_loss: 0.0707 | g_loss: 0.7460\n",
      "Epoch [   27/   50] | d_loss: 0.0958 | g_loss: 1.1857\n",
      "Epoch [   27/   50] | d_loss: 0.0902 | g_loss: 1.0319\n",
      "Epoch [   27/   50] | d_loss: 0.0574 | g_loss: 0.4798\n",
      "Epoch [   27/   50] | d_loss: 0.0946 | g_loss: 1.2161\n",
      "Epoch [   28/   50] | d_loss: 0.2357 | g_loss: 0.7475\n",
      "Epoch [   28/   50] | d_loss: 0.0769 | g_loss: 0.6681\n",
      "Epoch [   28/   50] | d_loss: 0.1061 | g_loss: 0.9143\n",
      "Epoch [   28/   50] | d_loss: 0.0404 | g_loss: 0.8754\n",
      "Epoch [   28/   50] | d_loss: 0.0664 | g_loss: 0.8378\n",
      "Epoch [   28/   50] | d_loss: 0.0796 | g_loss: 0.6131\n",
      "Epoch [   28/   50] | d_loss: 0.0679 | g_loss: 1.2869\n",
      "Epoch [   28/   50] | d_loss: 0.1191 | g_loss: 1.0296\n",
      "Epoch [   28/   50] | d_loss: 0.2114 | g_loss: 0.4776\n",
      "Epoch [   28/   50] | d_loss: 0.1038 | g_loss: 0.6674\n",
      "Epoch [   28/   50] | d_loss: 0.1037 | g_loss: 1.1536\n",
      "Epoch [   28/   50] | d_loss: 0.1244 | g_loss: 0.4814\n",
      "Epoch [   28/   50] | d_loss: 0.1684 | g_loss: 0.4344\n",
      "Epoch [   28/   50] | d_loss: 0.2532 | g_loss: 0.8645\n",
      "Epoch [   28/   50] | d_loss: 0.1791 | g_loss: 0.7374\n",
      "Epoch [   28/   50] | d_loss: 0.1264 | g_loss: 0.8054\n",
      "Epoch [   28/   50] | d_loss: 0.1332 | g_loss: 0.8612\n",
      "Epoch [   28/   50] | d_loss: 0.1336 | g_loss: 0.8019\n",
      "Epoch [   28/   50] | d_loss: 0.1296 | g_loss: 0.9674\n",
      "Epoch [   28/   50] | d_loss: 0.1426 | g_loss: 0.8468\n",
      "Epoch [   28/   50] | d_loss: 0.0714 | g_loss: 0.9788\n",
      "Epoch [   28/   50] | d_loss: 0.0498 | g_loss: 0.9301\n",
      "Epoch [   28/   50] | d_loss: 0.0935 | g_loss: 1.0713\n",
      "Epoch [   28/   50] | d_loss: 0.1557 | g_loss: 0.4073\n",
      "Epoch [   28/   50] | d_loss: 0.1053 | g_loss: 0.7950\n",
      "Epoch [   28/   50] | d_loss: 0.2062 | g_loss: 0.7584\n",
      "Epoch [   28/   50] | d_loss: 0.1338 | g_loss: 0.5679\n",
      "Epoch [   28/   50] | d_loss: 0.0873 | g_loss: 0.7686\n",
      "Epoch [   28/   50] | d_loss: 0.0986 | g_loss: 0.3736\n",
      "Epoch [   28/   50] | d_loss: 0.1942 | g_loss: 1.5288\n",
      "Epoch [   28/   50] | d_loss: 0.0690 | g_loss: 0.9002\n",
      "Epoch [   28/   50] | d_loss: 0.1487 | g_loss: 0.4243\n",
      "Epoch [   28/   50] | d_loss: 0.2734 | g_loss: 1.6411\n",
      "Epoch [   28/   50] | d_loss: 0.0754 | g_loss: 0.8791\n",
      "Epoch [   28/   50] | d_loss: 0.0689 | g_loss: 0.6560\n",
      "Epoch [   28/   50] | d_loss: 0.1548 | g_loss: 1.1106\n",
      "Epoch [   28/   50] | d_loss: 0.3918 | g_loss: 0.8867\n",
      "Epoch [   28/   50] | d_loss: 0.1252 | g_loss: 1.0918\n",
      "Epoch [   28/   50] | d_loss: 0.0882 | g_loss: 0.9966\n",
      "Epoch [   28/   50] | d_loss: 0.1661 | g_loss: 0.7939\n",
      "Epoch [   28/   50] | d_loss: 0.0926 | g_loss: 0.5684\n",
      "Epoch [   28/   50] | d_loss: 0.1571 | g_loss: 0.5265\n",
      "Epoch [   28/   50] | d_loss: 0.1527 | g_loss: 0.4178\n",
      "Epoch [   28/   50] | d_loss: 0.0993 | g_loss: 0.7230\n",
      "Epoch [   28/   50] | d_loss: 0.1708 | g_loss: 1.1897\n",
      "Epoch [   28/   50] | d_loss: 0.1440 | g_loss: 0.7829\n",
      "Epoch [   28/   50] | d_loss: 0.0913 | g_loss: 0.8453\n",
      "Epoch [   28/   50] | d_loss: 0.1944 | g_loss: 1.0464\n",
      "Epoch [   28/   50] | d_loss: 0.1627 | g_loss: 0.9637\n",
      "Epoch [   28/   50] | d_loss: 0.1528 | g_loss: 0.6849\n",
      "Epoch [   28/   50] | d_loss: 0.0802 | g_loss: 1.0961\n",
      "Epoch [   28/   50] | d_loss: 0.1648 | g_loss: 0.7448\n",
      "Epoch [   28/   50] | d_loss: 0.1378 | g_loss: 0.3334\n",
      "Epoch [   28/   50] | d_loss: 0.1468 | g_loss: 0.9302\n",
      "Epoch [   28/   50] | d_loss: 0.1051 | g_loss: 0.9602\n",
      "Epoch [   28/   50] | d_loss: 0.2584 | g_loss: 0.7493\n",
      "Epoch [   28/   50] | d_loss: 0.1755 | g_loss: 1.0019\n",
      "Epoch [   29/   50] | d_loss: 0.0822 | g_loss: 0.9100\n",
      "Epoch [   29/   50] | d_loss: 0.0762 | g_loss: 0.9819\n",
      "Epoch [   29/   50] | d_loss: 0.4568 | g_loss: 0.8069\n",
      "Epoch [   29/   50] | d_loss: 0.0645 | g_loss: 0.9396\n",
      "Epoch [   29/   50] | d_loss: 0.0808 | g_loss: 0.9360\n",
      "Epoch [   29/   50] | d_loss: 0.0953 | g_loss: 0.7846\n",
      "Epoch [   29/   50] | d_loss: 0.1814 | g_loss: 0.6684\n",
      "Epoch [   29/   50] | d_loss: 0.0631 | g_loss: 1.0319\n",
      "Epoch [   29/   50] | d_loss: 0.0686 | g_loss: 0.6933\n",
      "Epoch [   29/   50] | d_loss: 0.1262 | g_loss: 0.8230\n",
      "Epoch [   29/   50] | d_loss: 0.0829 | g_loss: 0.6129\n",
      "Epoch [   29/   50] | d_loss: 0.1179 | g_loss: 0.8837\n",
      "Epoch [   29/   50] | d_loss: 0.1425 | g_loss: 0.9228\n",
      "Epoch [   29/   50] | d_loss: 0.1286 | g_loss: 0.7267\n",
      "Epoch [   29/   50] | d_loss: 0.2406 | g_loss: 0.7195\n",
      "Epoch [   29/   50] | d_loss: 0.1597 | g_loss: 0.9673\n",
      "Epoch [   29/   50] | d_loss: 0.0781 | g_loss: 0.6272\n",
      "Epoch [   29/   50] | d_loss: 0.1896 | g_loss: 0.9472\n",
      "Epoch [   29/   50] | d_loss: 0.3225 | g_loss: 0.6125\n",
      "Epoch [   29/   50] | d_loss: 0.1020 | g_loss: 0.8372\n",
      "Epoch [   29/   50] | d_loss: 0.1330 | g_loss: 0.5288\n",
      "Epoch [   29/   50] | d_loss: 0.1574 | g_loss: 1.2910\n",
      "Epoch [   29/   50] | d_loss: 0.0916 | g_loss: 0.6746\n",
      "Epoch [   29/   50] | d_loss: 0.0907 | g_loss: 0.5631\n",
      "Epoch [   29/   50] | d_loss: 0.1468 | g_loss: 0.6768\n",
      "Epoch [   29/   50] | d_loss: 0.0680 | g_loss: 0.9252\n",
      "Epoch [   29/   50] | d_loss: 0.1510 | g_loss: 0.9899\n",
      "Epoch [   29/   50] | d_loss: 0.1951 | g_loss: 0.8040\n",
      "Epoch [   29/   50] | d_loss: 0.0617 | g_loss: 0.5616\n",
      "Epoch [   29/   50] | d_loss: 0.1199 | g_loss: 0.9955\n",
      "Epoch [   29/   50] | d_loss: 0.1354 | g_loss: 0.8882\n",
      "Epoch [   29/   50] | d_loss: 0.1545 | g_loss: 1.1597\n",
      "Epoch [   29/   50] | d_loss: 0.1387 | g_loss: 0.3291\n",
      "Epoch [   29/   50] | d_loss: 0.2415 | g_loss: 0.6525\n",
      "Epoch [   29/   50] | d_loss: 0.1631 | g_loss: 0.4601\n",
      "Epoch [   29/   50] | d_loss: 0.0816 | g_loss: 1.2459\n",
      "Epoch [   29/   50] | d_loss: 0.1405 | g_loss: 0.8419\n",
      "Epoch [   29/   50] | d_loss: 0.2011 | g_loss: 0.4966\n",
      "Epoch [   29/   50] | d_loss: 0.2436 | g_loss: 0.7157\n",
      "Epoch [   29/   50] | d_loss: 0.1461 | g_loss: 0.8611\n",
      "Epoch [   29/   50] | d_loss: 0.1269 | g_loss: 0.8380\n",
      "Epoch [   29/   50] | d_loss: 0.2576 | g_loss: 0.5973\n",
      "Epoch [   29/   50] | d_loss: 0.1053 | g_loss: 0.8617\n",
      "Epoch [   29/   50] | d_loss: 0.0976 | g_loss: 0.7653\n",
      "Epoch [   29/   50] | d_loss: 0.0836 | g_loss: 0.8842\n",
      "Epoch [   29/   50] | d_loss: 0.0852 | g_loss: 0.8816\n",
      "Epoch [   29/   50] | d_loss: 0.1209 | g_loss: 0.8475\n",
      "Epoch [   29/   50] | d_loss: 0.1437 | g_loss: 0.7689\n",
      "Epoch [   29/   50] | d_loss: 0.0905 | g_loss: 0.6014\n",
      "Epoch [   29/   50] | d_loss: 0.1280 | g_loss: 0.8306\n",
      "Epoch [   29/   50] | d_loss: 0.1999 | g_loss: 0.7299\n",
      "Epoch [   29/   50] | d_loss: 0.1506 | g_loss: 0.7769\n",
      "Epoch [   29/   50] | d_loss: 0.1995 | g_loss: 0.6543\n",
      "Epoch [   29/   50] | d_loss: 0.1266 | g_loss: 0.9117\n",
      "Epoch [   29/   50] | d_loss: 0.0682 | g_loss: 1.0592\n",
      "Epoch [   29/   50] | d_loss: 0.1229 | g_loss: 0.9113\n",
      "Epoch [   29/   50] | d_loss: 0.2063 | g_loss: 0.3260\n",
      "Epoch [   30/   50] | d_loss: 0.1946 | g_loss: 0.6214\n",
      "Epoch [   30/   50] | d_loss: 0.1731 | g_loss: 0.6796\n",
      "Epoch [   30/   50] | d_loss: 0.1916 | g_loss: 0.6947\n",
      "Epoch [   30/   50] | d_loss: 0.0915 | g_loss: 0.8481\n",
      "Epoch [   30/   50] | d_loss: 0.0562 | g_loss: 0.7974\n",
      "Epoch [   30/   50] | d_loss: 0.0964 | g_loss: 0.7104\n",
      "Epoch [   30/   50] | d_loss: 0.1509 | g_loss: 1.0847\n",
      "Epoch [   30/   50] | d_loss: 0.2231 | g_loss: 0.7035\n",
      "Epoch [   30/   50] | d_loss: 0.2060 | g_loss: 0.5445\n",
      "Epoch [   30/   50] | d_loss: 0.0891 | g_loss: 0.9308\n",
      "Epoch [   30/   50] | d_loss: 0.1216 | g_loss: 0.9775\n",
      "Epoch [   30/   50] | d_loss: 0.1679 | g_loss: 0.4539\n",
      "Epoch [   30/   50] | d_loss: 0.1776 | g_loss: 1.2040\n",
      "Epoch [   30/   50] | d_loss: 0.1226 | g_loss: 1.1065\n",
      "Epoch [   30/   50] | d_loss: 0.1121 | g_loss: 0.5462\n",
      "Epoch [   30/   50] | d_loss: 0.0857 | g_loss: 0.5999\n",
      "Epoch [   30/   50] | d_loss: 0.0825 | g_loss: 0.6465\n",
      "Epoch [   30/   50] | d_loss: 0.2070 | g_loss: 1.0225\n",
      "Epoch [   30/   50] | d_loss: 0.1247 | g_loss: 0.9666\n",
      "Epoch [   30/   50] | d_loss: 0.1002 | g_loss: 0.7540\n",
      "Epoch [   30/   50] | d_loss: 0.1584 | g_loss: 0.8741\n",
      "Epoch [   30/   50] | d_loss: 0.0549 | g_loss: 0.9441\n",
      "Epoch [   30/   50] | d_loss: 0.1330 | g_loss: 0.7752\n",
      "Epoch [   30/   50] | d_loss: 0.1953 | g_loss: 0.6500\n",
      "Epoch [   30/   50] | d_loss: 0.1823 | g_loss: 0.3855\n",
      "Epoch [   30/   50] | d_loss: 0.1276 | g_loss: 0.5990\n",
      "Epoch [   30/   50] | d_loss: 0.0909 | g_loss: 1.0216\n",
      "Epoch [   30/   50] | d_loss: 0.1010 | g_loss: 0.8038\n",
      "Epoch [   30/   50] | d_loss: 0.0977 | g_loss: 0.4387\n",
      "Epoch [   30/   50] | d_loss: 0.0966 | g_loss: 0.9960\n",
      "Epoch [   30/   50] | d_loss: 0.1764 | g_loss: 0.8316\n",
      "Epoch [   30/   50] | d_loss: 0.1028 | g_loss: 0.8211\n",
      "Epoch [   30/   50] | d_loss: 0.1244 | g_loss: 0.9731\n",
      "Epoch [   30/   50] | d_loss: 0.0881 | g_loss: 0.8490\n",
      "Epoch [   30/   50] | d_loss: 0.1159 | g_loss: 1.0745\n",
      "Epoch [   30/   50] | d_loss: 0.1724 | g_loss: 0.5961\n",
      "Epoch [   30/   50] | d_loss: 0.0914 | g_loss: 0.8571\n",
      "Epoch [   30/   50] | d_loss: 0.1006 | g_loss: 0.7739\n",
      "Epoch [   30/   50] | d_loss: 0.1567 | g_loss: 1.2735\n",
      "Epoch [   30/   50] | d_loss: 0.1255 | g_loss: 1.1334\n",
      "Epoch [   30/   50] | d_loss: 0.1351 | g_loss: 1.4120\n",
      "Epoch [   30/   50] | d_loss: 0.1729 | g_loss: 0.5499\n",
      "Epoch [   30/   50] | d_loss: 0.1157 | g_loss: 0.6203\n",
      "Epoch [   30/   50] | d_loss: 0.0994 | g_loss: 0.6057\n",
      "Epoch [   30/   50] | d_loss: 0.1118 | g_loss: 0.6856\n",
      "Epoch [   30/   50] | d_loss: 0.0955 | g_loss: 0.6318\n",
      "Epoch [   30/   50] | d_loss: 0.2325 | g_loss: 0.5423\n",
      "Epoch [   30/   50] | d_loss: 0.0988 | g_loss: 0.6956\n",
      "Epoch [   30/   50] | d_loss: 0.1842 | g_loss: 0.8786\n",
      "Epoch [   30/   50] | d_loss: 0.1387 | g_loss: 1.2793\n",
      "Epoch [   30/   50] | d_loss: 0.1732 | g_loss: 0.8097\n",
      "Epoch [   30/   50] | d_loss: 0.2371 | g_loss: 0.9917\n",
      "Epoch [   30/   50] | d_loss: 0.1199 | g_loss: 0.9581\n",
      "Epoch [   30/   50] | d_loss: 0.0826 | g_loss: 0.6809\n",
      "Epoch [   30/   50] | d_loss: 0.1147 | g_loss: 0.8234\n",
      "Epoch [   30/   50] | d_loss: 0.0950 | g_loss: 0.9395\n",
      "Epoch [   30/   50] | d_loss: 0.1409 | g_loss: 0.9788\n",
      "Epoch [   31/   50] | d_loss: 0.2560 | g_loss: 1.1240\n",
      "Epoch [   31/   50] | d_loss: 0.1153 | g_loss: 0.9991\n",
      "Epoch [   31/   50] | d_loss: 0.2344 | g_loss: 0.6564\n",
      "Epoch [   31/   50] | d_loss: 0.1865 | g_loss: 1.1977\n",
      "Epoch [   31/   50] | d_loss: 0.0986 | g_loss: 0.8539\n",
      "Epoch [   31/   50] | d_loss: 0.1439 | g_loss: 0.5830\n",
      "Epoch [   31/   50] | d_loss: 0.0740 | g_loss: 0.8977\n",
      "Epoch [   31/   50] | d_loss: 0.1026 | g_loss: 0.8547\n",
      "Epoch [   31/   50] | d_loss: 0.0797 | g_loss: 0.8832\n",
      "Epoch [   31/   50] | d_loss: 0.1237 | g_loss: 0.9938\n",
      "Epoch [   31/   50] | d_loss: 0.0870 | g_loss: 0.6582\n",
      "Epoch [   31/   50] | d_loss: 0.0866 | g_loss: 0.9017\n",
      "Epoch [   31/   50] | d_loss: 0.0751 | g_loss: 1.0999\n",
      "Epoch [   31/   50] | d_loss: 0.1265 | g_loss: 0.7791\n",
      "Epoch [   31/   50] | d_loss: 0.1340 | g_loss: 0.8097\n",
      "Epoch [   31/   50] | d_loss: 0.0964 | g_loss: 0.9189\n",
      "Epoch [   31/   50] | d_loss: 0.0694 | g_loss: 0.6905\n",
      "Epoch [   31/   50] | d_loss: 0.1995 | g_loss: 1.1916\n",
      "Epoch [   31/   50] | d_loss: 0.1603 | g_loss: 1.3289\n",
      "Epoch [   31/   50] | d_loss: 0.1324 | g_loss: 0.5299\n",
      "Epoch [   31/   50] | d_loss: 0.0579 | g_loss: 0.6012\n",
      "Epoch [   31/   50] | d_loss: 0.1586 | g_loss: 0.4007\n",
      "Epoch [   31/   50] | d_loss: 0.1038 | g_loss: 0.4545\n",
      "Epoch [   31/   50] | d_loss: 0.0575 | g_loss: 1.1437\n",
      "Epoch [   31/   50] | d_loss: 0.0553 | g_loss: 0.7147\n",
      "Epoch [   31/   50] | d_loss: 0.1128 | g_loss: 0.5195\n",
      "Epoch [   31/   50] | d_loss: 0.0923 | g_loss: 1.0950\n",
      "Epoch [   31/   50] | d_loss: 0.1443 | g_loss: 0.7929\n",
      "Epoch [   31/   50] | d_loss: 0.0651 | g_loss: 0.9038\n",
      "Epoch [   31/   50] | d_loss: 0.0808 | g_loss: 0.8183\n",
      "Epoch [   31/   50] | d_loss: 0.0859 | g_loss: 1.0798\n",
      "Epoch [   31/   50] | d_loss: 0.0842 | g_loss: 0.8337\n",
      "Epoch [   31/   50] | d_loss: 0.3217 | g_loss: 0.6119\n",
      "Epoch [   31/   50] | d_loss: 0.1439 | g_loss: 0.8630\n",
      "Epoch [   31/   50] | d_loss: 0.1233 | g_loss: 0.6900\n",
      "Epoch [   31/   50] | d_loss: 0.1820 | g_loss: 1.0895\n",
      "Epoch [   31/   50] | d_loss: 0.1594 | g_loss: 0.8204\n",
      "Epoch [   31/   50] | d_loss: 0.0933 | g_loss: 0.8118\n",
      "Epoch [   31/   50] | d_loss: 0.1158 | g_loss: 0.5280\n",
      "Epoch [   31/   50] | d_loss: 0.1259 | g_loss: 0.9437\n",
      "Epoch [   31/   50] | d_loss: 0.0847 | g_loss: 0.7228\n",
      "Epoch [   31/   50] | d_loss: 0.1770 | g_loss: 0.1663\n",
      "Epoch [   31/   50] | d_loss: 0.1109 | g_loss: 1.0219\n",
      "Epoch [   31/   50] | d_loss: 0.0770 | g_loss: 0.9298\n",
      "Epoch [   31/   50] | d_loss: 0.0695 | g_loss: 1.0416\n",
      "Epoch [   31/   50] | d_loss: 0.1735 | g_loss: 0.6986\n",
      "Epoch [   31/   50] | d_loss: 0.0770 | g_loss: 0.6124\n",
      "Epoch [   31/   50] | d_loss: 0.0776 | g_loss: 0.6247\n",
      "Epoch [   31/   50] | d_loss: 0.2124 | g_loss: 0.7358\n",
      "Epoch [   31/   50] | d_loss: 0.0544 | g_loss: 1.1215\n",
      "Epoch [   31/   50] | d_loss: 0.2329 | g_loss: 0.4200\n",
      "Epoch [   31/   50] | d_loss: 0.5250 | g_loss: 0.6925\n",
      "Epoch [   31/   50] | d_loss: 0.1280 | g_loss: 0.4377\n",
      "Epoch [   31/   50] | d_loss: 0.0914 | g_loss: 0.9209\n",
      "Epoch [   31/   50] | d_loss: 0.2438 | g_loss: 1.3087\n",
      "Epoch [   31/   50] | d_loss: 0.1699 | g_loss: 0.7624\n",
      "Epoch [   31/   50] | d_loss: 0.1630 | g_loss: 0.6879\n",
      "Epoch [   32/   50] | d_loss: 0.2292 | g_loss: 0.8321\n",
      "Epoch [   32/   50] | d_loss: 0.0702 | g_loss: 0.8521\n",
      "Epoch [   32/   50] | d_loss: 0.0875 | g_loss: 1.4459\n",
      "Epoch [   32/   50] | d_loss: 0.1962 | g_loss: 0.6184\n",
      "Epoch [   32/   50] | d_loss: 0.0818 | g_loss: 0.9953\n",
      "Epoch [   32/   50] | d_loss: 0.1412 | g_loss: 0.9225\n",
      "Epoch [   32/   50] | d_loss: 0.0897 | g_loss: 0.6893\n",
      "Epoch [   32/   50] | d_loss: 0.2819 | g_loss: 1.1271\n",
      "Epoch [   32/   50] | d_loss: 0.1628 | g_loss: 0.5995\n",
      "Epoch [   32/   50] | d_loss: 0.0854 | g_loss: 0.9975\n",
      "Epoch [   32/   50] | d_loss: 0.1291 | g_loss: 0.5684\n",
      "Epoch [   32/   50] | d_loss: 0.1372 | g_loss: 0.7861\n",
      "Epoch [   32/   50] | d_loss: 0.1427 | g_loss: 0.7459\n",
      "Epoch [   32/   50] | d_loss: 0.0783 | g_loss: 0.8527\n",
      "Epoch [   32/   50] | d_loss: 0.0943 | g_loss: 0.5216\n",
      "Epoch [   32/   50] | d_loss: 0.1349 | g_loss: 0.7398\n",
      "Epoch [   32/   50] | d_loss: 0.1539 | g_loss: 0.7065\n",
      "Epoch [   32/   50] | d_loss: 0.0879 | g_loss: 0.8222\n",
      "Epoch [   32/   50] | d_loss: 0.1026 | g_loss: 0.5818\n",
      "Epoch [   32/   50] | d_loss: 0.2315 | g_loss: 0.8981\n",
      "Epoch [   32/   50] | d_loss: 0.0635 | g_loss: 1.1186\n",
      "Epoch [   32/   50] | d_loss: 0.0776 | g_loss: 0.8691\n",
      "Epoch [   32/   50] | d_loss: 0.1668 | g_loss: 0.9589\n",
      "Epoch [   32/   50] | d_loss: 0.0664 | g_loss: 0.6886\n",
      "Epoch [   32/   50] | d_loss: 0.1253 | g_loss: 0.9789\n",
      "Epoch [   32/   50] | d_loss: 0.0867 | g_loss: 0.8259\n",
      "Epoch [   32/   50] | d_loss: 0.1116 | g_loss: 0.6498\n",
      "Epoch [   32/   50] | d_loss: 0.1085 | g_loss: 0.6758\n",
      "Epoch [   32/   50] | d_loss: 0.0768 | g_loss: 0.8550\n",
      "Epoch [   32/   50] | d_loss: 0.0659 | g_loss: 0.5327\n",
      "Epoch [   32/   50] | d_loss: 0.1348 | g_loss: 1.0788\n",
      "Epoch [   32/   50] | d_loss: 0.0970 | g_loss: 0.9964\n",
      "Epoch [   32/   50] | d_loss: 0.2044 | g_loss: 0.7251\n",
      "Epoch [   32/   50] | d_loss: 0.1515 | g_loss: 1.0124\n",
      "Epoch [   32/   50] | d_loss: 0.1088 | g_loss: 0.4216\n",
      "Epoch [   32/   50] | d_loss: 0.1767 | g_loss: 1.2810\n",
      "Epoch [   32/   50] | d_loss: 0.2536 | g_loss: 0.8083\n",
      "Epoch [   32/   50] | d_loss: 0.1020 | g_loss: 0.7664\n",
      "Epoch [   32/   50] | d_loss: 0.1813 | g_loss: 0.9621\n",
      "Epoch [   32/   50] | d_loss: 0.0485 | g_loss: 0.9582\n",
      "Epoch [   32/   50] | d_loss: 0.1292 | g_loss: 0.4368\n",
      "Epoch [   32/   50] | d_loss: 0.2462 | g_loss: 1.2203\n",
      "Epoch [   32/   50] | d_loss: 0.1487 | g_loss: 0.6933\n",
      "Epoch [   32/   50] | d_loss: 0.0832 | g_loss: 0.9621\n",
      "Epoch [   32/   50] | d_loss: 0.2921 | g_loss: 0.4746\n",
      "Epoch [   32/   50] | d_loss: 0.0917 | g_loss: 0.6037\n",
      "Epoch [   32/   50] | d_loss: 0.2409 | g_loss: 1.1857\n",
      "Epoch [   32/   50] | d_loss: 0.0575 | g_loss: 0.9439\n",
      "Epoch [   32/   50] | d_loss: 0.1056 | g_loss: 0.7869\n",
      "Epoch [   32/   50] | d_loss: 0.1784 | g_loss: 0.7148\n",
      "Epoch [   32/   50] | d_loss: 0.3005 | g_loss: 0.5911\n",
      "Epoch [   32/   50] | d_loss: 0.1816 | g_loss: 0.6715\n",
      "Epoch [   32/   50] | d_loss: 0.2427 | g_loss: 0.7507\n",
      "Epoch [   32/   50] | d_loss: 0.1121 | g_loss: 1.2118\n",
      "Epoch [   32/   50] | d_loss: 0.0863 | g_loss: 1.0131\n",
      "Epoch [   32/   50] | d_loss: 0.0694 | g_loss: 0.7223\n",
      "Epoch [   32/   50] | d_loss: 0.1210 | g_loss: 0.9435\n",
      "Epoch [   33/   50] | d_loss: 0.1384 | g_loss: 0.7378\n",
      "Epoch [   33/   50] | d_loss: 0.1476 | g_loss: 0.6810\n",
      "Epoch [   33/   50] | d_loss: 0.1690 | g_loss: 1.2074\n",
      "Epoch [   33/   50] | d_loss: 0.1099 | g_loss: 0.8573\n",
      "Epoch [   33/   50] | d_loss: 0.1466 | g_loss: 0.4888\n",
      "Epoch [   33/   50] | d_loss: 0.1320 | g_loss: 0.4525\n",
      "Epoch [   33/   50] | d_loss: 0.0999 | g_loss: 0.5546\n",
      "Epoch [   33/   50] | d_loss: 0.0926 | g_loss: 0.9667\n",
      "Epoch [   33/   50] | d_loss: 0.0824 | g_loss: 0.7753\n",
      "Epoch [   33/   50] | d_loss: 0.1032 | g_loss: 0.7210\n",
      "Epoch [   33/   50] | d_loss: 0.0779 | g_loss: 0.7081\n",
      "Epoch [   33/   50] | d_loss: 0.0879 | g_loss: 1.0271\n",
      "Epoch [   33/   50] | d_loss: 0.0884 | g_loss: 1.3369\n",
      "Epoch [   33/   50] | d_loss: 0.0634 | g_loss: 0.6757\n",
      "Epoch [   33/   50] | d_loss: 0.1260 | g_loss: 0.8176\n",
      "Epoch [   33/   50] | d_loss: 0.0767 | g_loss: 0.5839\n",
      "Epoch [   33/   50] | d_loss: 0.1311 | g_loss: 0.9701\n",
      "Epoch [   33/   50] | d_loss: 0.1084 | g_loss: 0.7480\n",
      "Epoch [   33/   50] | d_loss: 0.1451 | g_loss: 0.7904\n",
      "Epoch [   33/   50] | d_loss: 0.0625 | g_loss: 0.8223\n",
      "Epoch [   33/   50] | d_loss: 0.1430 | g_loss: 0.7693\n",
      "Epoch [   33/   50] | d_loss: 0.0623 | g_loss: 1.1320\n",
      "Epoch [   33/   50] | d_loss: 0.0687 | g_loss: 1.0200\n",
      "Epoch [   33/   50] | d_loss: 0.1717 | g_loss: 0.8039\n",
      "Epoch [   33/   50] | d_loss: 0.0703 | g_loss: 0.5238\n",
      "Epoch [   33/   50] | d_loss: 0.2606 | g_loss: 0.6107\n",
      "Epoch [   33/   50] | d_loss: 0.1004 | g_loss: 1.2301\n",
      "Epoch [   33/   50] | d_loss: 0.0711 | g_loss: 0.9048\n",
      "Epoch [   33/   50] | d_loss: 0.1085 | g_loss: 1.1320\n",
      "Epoch [   33/   50] | d_loss: 0.0846 | g_loss: 0.9861\n",
      "Epoch [   33/   50] | d_loss: 0.1676 | g_loss: 0.7274\n",
      "Epoch [   33/   50] | d_loss: 0.2660 | g_loss: 1.1597\n",
      "Epoch [   33/   50] | d_loss: 0.1857 | g_loss: 0.5871\n",
      "Epoch [   33/   50] | d_loss: 0.1991 | g_loss: 0.9992\n",
      "Epoch [   33/   50] | d_loss: 0.0983 | g_loss: 1.0234\n",
      "Epoch [   33/   50] | d_loss: 0.2611 | g_loss: 0.9089\n",
      "Epoch [   33/   50] | d_loss: 0.0685 | g_loss: 0.8842\n",
      "Epoch [   33/   50] | d_loss: 0.0911 | g_loss: 1.1033\n",
      "Epoch [   33/   50] | d_loss: 0.1331 | g_loss: 0.4998\n",
      "Epoch [   33/   50] | d_loss: 0.1358 | g_loss: 0.8254\n",
      "Epoch [   33/   50] | d_loss: 0.1103 | g_loss: 0.6724\n",
      "Epoch [   33/   50] | d_loss: 0.0826 | g_loss: 0.7443\n",
      "Epoch [   33/   50] | d_loss: 0.1103 | g_loss: 0.5945\n",
      "Epoch [   33/   50] | d_loss: 0.0793 | g_loss: 1.0950\n",
      "Epoch [   33/   50] | d_loss: 0.0456 | g_loss: 1.2326\n",
      "Epoch [   33/   50] | d_loss: 0.1329 | g_loss: 0.7224\n",
      "Epoch [   33/   50] | d_loss: 0.2980 | g_loss: 0.5205\n",
      "Epoch [   33/   50] | d_loss: 0.0652 | g_loss: 1.7270\n",
      "Epoch [   33/   50] | d_loss: 0.0832 | g_loss: 1.3685\n",
      "Epoch [   33/   50] | d_loss: 0.2069 | g_loss: 0.9008\n",
      "Epoch [   33/   50] | d_loss: 0.0587 | g_loss: 0.8701\n",
      "Epoch [   33/   50] | d_loss: 0.1543 | g_loss: 0.5913\n",
      "Epoch [   33/   50] | d_loss: 0.1019 | g_loss: 0.3920\n",
      "Epoch [   33/   50] | d_loss: 0.1968 | g_loss: 1.0571\n",
      "Epoch [   33/   50] | d_loss: 0.1138 | g_loss: 0.7333\n",
      "Epoch [   33/   50] | d_loss: 0.0985 | g_loss: 0.7875\n",
      "Epoch [   33/   50] | d_loss: 0.0850 | g_loss: 0.4998\n",
      "Epoch [   34/   50] | d_loss: 0.2305 | g_loss: 0.9663\n",
      "Epoch [   34/   50] | d_loss: 0.1826 | g_loss: 0.7270\n",
      "Epoch [   34/   50] | d_loss: 0.1049 | g_loss: 0.5345\n",
      "Epoch [   34/   50] | d_loss: 0.0646 | g_loss: 0.9766\n",
      "Epoch [   34/   50] | d_loss: 0.1284 | g_loss: 0.9990\n",
      "Epoch [   34/   50] | d_loss: 0.1063 | g_loss: 0.6442\n",
      "Epoch [   34/   50] | d_loss: 0.1115 | g_loss: 0.7093\n",
      "Epoch [   34/   50] | d_loss: 0.1708 | g_loss: 1.1212\n",
      "Epoch [   34/   50] | d_loss: 0.1122 | g_loss: 1.1732\n",
      "Epoch [   34/   50] | d_loss: 0.0648 | g_loss: 1.0164\n",
      "Epoch [   34/   50] | d_loss: 0.0646 | g_loss: 0.9959\n",
      "Epoch [   34/   50] | d_loss: 0.1600 | g_loss: 0.4232\n",
      "Epoch [   34/   50] | d_loss: 0.1510 | g_loss: 0.6230\n",
      "Epoch [   34/   50] | d_loss: 0.1223 | g_loss: 1.0453\n",
      "Epoch [   34/   50] | d_loss: 0.1337 | g_loss: 0.7856\n",
      "Epoch [   34/   50] | d_loss: 0.2084 | g_loss: 0.6566\n",
      "Epoch [   34/   50] | d_loss: 0.1338 | g_loss: 0.7985\n",
      "Epoch [   34/   50] | d_loss: 0.0484 | g_loss: 0.9160\n",
      "Epoch [   34/   50] | d_loss: 0.1236 | g_loss: 0.7190\n",
      "Epoch [   34/   50] | d_loss: 0.0988 | g_loss: 1.4325\n",
      "Epoch [   34/   50] | d_loss: 0.1284 | g_loss: 0.8203\n",
      "Epoch [   34/   50] | d_loss: 0.1112 | g_loss: 0.7643\n",
      "Epoch [   34/   50] | d_loss: 0.1640 | g_loss: 0.9015\n",
      "Epoch [   34/   50] | d_loss: 0.2028 | g_loss: 0.8225\n",
      "Epoch [   34/   50] | d_loss: 0.4785 | g_loss: 0.8704\n",
      "Epoch [   34/   50] | d_loss: 0.1922 | g_loss: 0.6780\n",
      "Epoch [   34/   50] | d_loss: 0.1447 | g_loss: 0.4604\n",
      "Epoch [   34/   50] | d_loss: 0.1475 | g_loss: 0.7273\n",
      "Epoch [   34/   50] | d_loss: 0.0904 | g_loss: 0.6518\n",
      "Epoch [   34/   50] | d_loss: 0.0581 | g_loss: 0.6682\n",
      "Epoch [   34/   50] | d_loss: 0.2351 | g_loss: 0.7581\n",
      "Epoch [   34/   50] | d_loss: 0.1485 | g_loss: 0.6510\n",
      "Epoch [   34/   50] | d_loss: 0.1543 | g_loss: 1.1278\n",
      "Epoch [   34/   50] | d_loss: 0.1092 | g_loss: 1.2268\n",
      "Epoch [   34/   50] | d_loss: 0.0693 | g_loss: 1.1238\n",
      "Epoch [   34/   50] | d_loss: 0.0502 | g_loss: 1.0983\n",
      "Epoch [   34/   50] | d_loss: 0.2017 | g_loss: 0.6431\n",
      "Epoch [   34/   50] | d_loss: 0.0982 | g_loss: 0.6387\n",
      "Epoch [   34/   50] | d_loss: 0.2270 | g_loss: 1.0838\n",
      "Epoch [   34/   50] | d_loss: 0.1184 | g_loss: 0.7006\n",
      "Epoch [   34/   50] | d_loss: 0.1044 | g_loss: 0.7455\n",
      "Epoch [   34/   50] | d_loss: 0.1055 | g_loss: 0.6161\n",
      "Epoch [   34/   50] | d_loss: 0.3051 | g_loss: 0.9172\n",
      "Epoch [   34/   50] | d_loss: 0.1528 | g_loss: 0.7053\n",
      "Epoch [   34/   50] | d_loss: 0.0788 | g_loss: 0.9268\n",
      "Epoch [   34/   50] | d_loss: 0.1392 | g_loss: 0.8243\n",
      "Epoch [   34/   50] | d_loss: 0.0617 | g_loss: 0.7479\n",
      "Epoch [   34/   50] | d_loss: 0.1400 | g_loss: 0.6883\n",
      "Epoch [   34/   50] | d_loss: 0.1846 | g_loss: 0.8864\n",
      "Epoch [   34/   50] | d_loss: 0.1392 | g_loss: 0.7185\n",
      "Epoch [   34/   50] | d_loss: 0.2350 | g_loss: 0.6388\n",
      "Epoch [   34/   50] | d_loss: 0.1270 | g_loss: 0.7658\n",
      "Epoch [   34/   50] | d_loss: 0.0761 | g_loss: 0.5992\n",
      "Epoch [   34/   50] | d_loss: 0.2169 | g_loss: 0.5152\n",
      "Epoch [   34/   50] | d_loss: 0.1423 | g_loss: 0.7714\n",
      "Epoch [   34/   50] | d_loss: 0.2570 | g_loss: 1.0169\n",
      "Epoch [   34/   50] | d_loss: 0.0844 | g_loss: 0.8172\n",
      "Epoch [   35/   50] | d_loss: 0.0946 | g_loss: 0.8797\n",
      "Epoch [   35/   50] | d_loss: 0.1482 | g_loss: 1.0598\n",
      "Epoch [   35/   50] | d_loss: 0.1519 | g_loss: 1.0450\n",
      "Epoch [   35/   50] | d_loss: 0.1015 | g_loss: 0.7694\n",
      "Epoch [   35/   50] | d_loss: 0.0439 | g_loss: 0.7944\n",
      "Epoch [   35/   50] | d_loss: 0.1176 | g_loss: 0.9027\n",
      "Epoch [   35/   50] | d_loss: 0.1104 | g_loss: 0.7317\n",
      "Epoch [   35/   50] | d_loss: 0.1602 | g_loss: 0.4994\n",
      "Epoch [   35/   50] | d_loss: 0.1237 | g_loss: 0.5130\n",
      "Epoch [   35/   50] | d_loss: 0.0922 | g_loss: 0.6872\n",
      "Epoch [   35/   50] | d_loss: 0.0948 | g_loss: 0.8720\n",
      "Epoch [   35/   50] | d_loss: 0.2279 | g_loss: 0.4905\n",
      "Epoch [   35/   50] | d_loss: 0.1488 | g_loss: 1.1077\n",
      "Epoch [   35/   50] | d_loss: 0.2249 | g_loss: 0.8068\n",
      "Epoch [   35/   50] | d_loss: 0.0742 | g_loss: 0.6733\n",
      "Epoch [   35/   50] | d_loss: 0.1744 | g_loss: 0.4931\n",
      "Epoch [   35/   50] | d_loss: 0.0953 | g_loss: 0.5844\n",
      "Epoch [   35/   50] | d_loss: 0.1712 | g_loss: 0.5724\n",
      "Epoch [   35/   50] | d_loss: 0.0635 | g_loss: 1.1394\n",
      "Epoch [   35/   50] | d_loss: 0.1710 | g_loss: 0.6226\n",
      "Epoch [   35/   50] | d_loss: 0.0720 | g_loss: 0.5944\n",
      "Epoch [   35/   50] | d_loss: 0.1255 | g_loss: 0.9762\n",
      "Epoch [   35/   50] | d_loss: 0.2066 | g_loss: 0.5244\n",
      "Epoch [   35/   50] | d_loss: 0.2457 | g_loss: 0.7599\n",
      "Epoch [   35/   50] | d_loss: 0.2011 | g_loss: 1.1531\n",
      "Epoch [   35/   50] | d_loss: 0.0611 | g_loss: 1.3744\n",
      "Epoch [   35/   50] | d_loss: 0.1619 | g_loss: 0.6439\n",
      "Epoch [   35/   50] | d_loss: 0.6117 | g_loss: 0.4227\n",
      "Epoch [   35/   50] | d_loss: 0.2591 | g_loss: 0.9147\n",
      "Epoch [   35/   50] | d_loss: 0.1076 | g_loss: 0.9729\n",
      "Epoch [   35/   50] | d_loss: 0.1277 | g_loss: 0.8495\n",
      "Epoch [   35/   50] | d_loss: 0.0694 | g_loss: 0.8979\n",
      "Epoch [   35/   50] | d_loss: 0.1922 | g_loss: 0.5342\n",
      "Epoch [   35/   50] | d_loss: 0.1155 | g_loss: 0.8993\n",
      "Epoch [   35/   50] | d_loss: 0.0729 | g_loss: 0.7148\n",
      "Epoch [   35/   50] | d_loss: 0.1316 | g_loss: 0.5252\n",
      "Epoch [   35/   50] | d_loss: 0.0686 | g_loss: 0.7535\n",
      "Epoch [   35/   50] | d_loss: 0.0602 | g_loss: 1.3311\n",
      "Epoch [   35/   50] | d_loss: 0.0914 | g_loss: 0.8523\n",
      "Epoch [   35/   50] | d_loss: 0.0640 | g_loss: 0.8867\n",
      "Epoch [   35/   50] | d_loss: 0.1186 | g_loss: 0.6905\n",
      "Epoch [   35/   50] | d_loss: 0.0862 | g_loss: 0.5933\n",
      "Epoch [   35/   50] | d_loss: 0.1726 | g_loss: 0.8292\n",
      "Epoch [   35/   50] | d_loss: 0.0923 | g_loss: 0.5728\n",
      "Epoch [   35/   50] | d_loss: 0.0741 | g_loss: 1.0296\n",
      "Epoch [   35/   50] | d_loss: 0.1156 | g_loss: 0.6220\n",
      "Epoch [   35/   50] | d_loss: 0.1684 | g_loss: 0.3923\n",
      "Epoch [   35/   50] | d_loss: 0.0496 | g_loss: 1.0217\n",
      "Epoch [   35/   50] | d_loss: 0.2746 | g_loss: 0.6831\n",
      "Epoch [   35/   50] | d_loss: 0.1377 | g_loss: 1.1958\n",
      "Epoch [   35/   50] | d_loss: 0.1289 | g_loss: 0.4890\n",
      "Epoch [   35/   50] | d_loss: 0.0920 | g_loss: 0.7325\n",
      "Epoch [   35/   50] | d_loss: 0.0615 | g_loss: 0.8300\n",
      "Epoch [   35/   50] | d_loss: 0.2115 | g_loss: 0.6833\n",
      "Epoch [   35/   50] | d_loss: 0.0851 | g_loss: 0.9252\n",
      "Epoch [   35/   50] | d_loss: 0.1043 | g_loss: 0.5641\n",
      "Epoch [   35/   50] | d_loss: 0.1487 | g_loss: 0.8078\n",
      "Epoch [   36/   50] | d_loss: 0.0828 | g_loss: 0.8601\n",
      "Epoch [   36/   50] | d_loss: 0.1129 | g_loss: 0.8099\n",
      "Epoch [   36/   50] | d_loss: 0.1215 | g_loss: 1.0265\n",
      "Epoch [   36/   50] | d_loss: 0.1504 | g_loss: 0.7917\n",
      "Epoch [   36/   50] | d_loss: 0.1174 | g_loss: 0.6191\n",
      "Epoch [   36/   50] | d_loss: 0.1186 | g_loss: 0.8216\n",
      "Epoch [   36/   50] | d_loss: 0.1182 | g_loss: 0.9714\n",
      "Epoch [   36/   50] | d_loss: 0.1062 | g_loss: 1.1275\n",
      "Epoch [   36/   50] | d_loss: 0.0712 | g_loss: 0.6526\n",
      "Epoch [   36/   50] | d_loss: 0.0846 | g_loss: 1.1065\n",
      "Epoch [   36/   50] | d_loss: 0.1043 | g_loss: 0.6381\n",
      "Epoch [   36/   50] | d_loss: 0.0711 | g_loss: 0.7106\n",
      "Epoch [   36/   50] | d_loss: 0.0722 | g_loss: 0.6194\n",
      "Epoch [   36/   50] | d_loss: 0.2378 | g_loss: 1.0856\n",
      "Epoch [   36/   50] | d_loss: 0.0599 | g_loss: 0.9665\n",
      "Epoch [   36/   50] | d_loss: 0.0914 | g_loss: 0.7933\n",
      "Epoch [   36/   50] | d_loss: 0.1929 | g_loss: 0.5184\n",
      "Epoch [   36/   50] | d_loss: 0.1936 | g_loss: 0.5647\n",
      "Epoch [   36/   50] | d_loss: 0.1723 | g_loss: 0.8379\n",
      "Epoch [   36/   50] | d_loss: 0.0757 | g_loss: 0.6165\n",
      "Epoch [   36/   50] | d_loss: 0.0855 | g_loss: 0.6825\n",
      "Epoch [   36/   50] | d_loss: 0.2095 | g_loss: 0.6337\n",
      "Epoch [   36/   50] | d_loss: 0.1363 | g_loss: 0.8960\n",
      "Epoch [   36/   50] | d_loss: 0.0739 | g_loss: 0.8683\n",
      "Epoch [   36/   50] | d_loss: 0.1037 | g_loss: 0.6789\n",
      "Epoch [   36/   50] | d_loss: 0.1163 | g_loss: 0.9054\n",
      "Epoch [   36/   50] | d_loss: 0.2148 | g_loss: 0.6312\n",
      "Epoch [   36/   50] | d_loss: 0.1039 | g_loss: 1.0020\n",
      "Epoch [   36/   50] | d_loss: 0.0855 | g_loss: 0.9962\n",
      "Epoch [   36/   50] | d_loss: 0.1339 | g_loss: 0.4587\n",
      "Epoch [   36/   50] | d_loss: 0.1709 | g_loss: 0.4200\n",
      "Epoch [   36/   50] | d_loss: 0.0652 | g_loss: 0.8980\n",
      "Epoch [   36/   50] | d_loss: 0.1917 | g_loss: 0.6304\n",
      "Epoch [   36/   50] | d_loss: 0.0996 | g_loss: 1.0558\n",
      "Epoch [   36/   50] | d_loss: 0.3365 | g_loss: 0.8044\n",
      "Epoch [   36/   50] | d_loss: 0.1747 | g_loss: 0.6893\n",
      "Epoch [   36/   50] | d_loss: 0.1435 | g_loss: 0.8543\n",
      "Epoch [   36/   50] | d_loss: 0.0976 | g_loss: 1.0106\n",
      "Epoch [   36/   50] | d_loss: 0.1178 | g_loss: 1.5922\n",
      "Epoch [   36/   50] | d_loss: 0.2157 | g_loss: 0.7489\n",
      "Epoch [   36/   50] | d_loss: 0.1215 | g_loss: 1.1207\n",
      "Epoch [   36/   50] | d_loss: 0.1271 | g_loss: 0.8499\n",
      "Epoch [   36/   50] | d_loss: 0.0642 | g_loss: 1.1155\n",
      "Epoch [   36/   50] | d_loss: 0.0969 | g_loss: 1.1822\n",
      "Epoch [   36/   50] | d_loss: 0.2060 | g_loss: 0.6349\n",
      "Epoch [   36/   50] | d_loss: 0.1454 | g_loss: 0.6150\n",
      "Epoch [   36/   50] | d_loss: 0.0592 | g_loss: 0.7354\n",
      "Epoch [   36/   50] | d_loss: 0.1732 | g_loss: 1.0022\n",
      "Epoch [   36/   50] | d_loss: 0.0942 | g_loss: 0.6606\n",
      "Epoch [   36/   50] | d_loss: 0.2012 | g_loss: 1.1448\n",
      "Epoch [   36/   50] | d_loss: 0.0935 | g_loss: 0.6940\n",
      "Epoch [   36/   50] | d_loss: 0.5171 | g_loss: 0.3084\n",
      "Epoch [   36/   50] | d_loss: 0.1690 | g_loss: 0.7261\n",
      "Epoch [   36/   50] | d_loss: 0.1146 | g_loss: 0.7821\n",
      "Epoch [   36/   50] | d_loss: 0.2537 | g_loss: 0.4034\n",
      "Epoch [   36/   50] | d_loss: 0.1592 | g_loss: 1.1023\n",
      "Epoch [   36/   50] | d_loss: 0.1715 | g_loss: 0.6335\n",
      "Epoch [   37/   50] | d_loss: 0.0712 | g_loss: 0.8540\n",
      "Epoch [   37/   50] | d_loss: 0.1057 | g_loss: 1.0785\n",
      "Epoch [   37/   50] | d_loss: 0.2074 | g_loss: 0.9125\n",
      "Epoch [   37/   50] | d_loss: 0.1179 | g_loss: 0.7677\n",
      "Epoch [   37/   50] | d_loss: 0.0762 | g_loss: 0.7216\n",
      "Epoch [   37/   50] | d_loss: 0.0811 | g_loss: 0.8136\n",
      "Epoch [   37/   50] | d_loss: 0.0910 | g_loss: 0.9738\n",
      "Epoch [   37/   50] | d_loss: 0.1515 | g_loss: 0.3503\n",
      "Epoch [   37/   50] | d_loss: 0.0703 | g_loss: 0.8309\n",
      "Epoch [   37/   50] | d_loss: 0.2068 | g_loss: 0.7694\n",
      "Epoch [   37/   50] | d_loss: 0.3058 | g_loss: 1.2728\n",
      "Epoch [   37/   50] | d_loss: 0.1665 | g_loss: 0.5974\n",
      "Epoch [   37/   50] | d_loss: 0.1374 | g_loss: 0.6198\n",
      "Epoch [   37/   50] | d_loss: 0.0754 | g_loss: 0.7815\n",
      "Epoch [   37/   50] | d_loss: 0.0913 | g_loss: 0.7803\n",
      "Epoch [   37/   50] | d_loss: 0.1205 | g_loss: 0.9258\n",
      "Epoch [   37/   50] | d_loss: 0.0640 | g_loss: 0.8679\n",
      "Epoch [   37/   50] | d_loss: 0.2042 | g_loss: 0.5356\n",
      "Epoch [   37/   50] | d_loss: 0.1367 | g_loss: 1.0059\n",
      "Epoch [   37/   50] | d_loss: 0.0736 | g_loss: 0.8021\n",
      "Epoch [   37/   50] | d_loss: 0.0537 | g_loss: 0.8414\n",
      "Epoch [   37/   50] | d_loss: 0.0978 | g_loss: 0.8729\n",
      "Epoch [   37/   50] | d_loss: 0.0947 | g_loss: 0.7945\n",
      "Epoch [   37/   50] | d_loss: 0.1105 | g_loss: 0.5656\n",
      "Epoch [   37/   50] | d_loss: 0.0737 | g_loss: 0.9007\n",
      "Epoch [   37/   50] | d_loss: 0.0773 | g_loss: 0.7181\n",
      "Epoch [   37/   50] | d_loss: 0.1464 | g_loss: 0.9033\n",
      "Epoch [   37/   50] | d_loss: 0.1017 | g_loss: 0.7815\n",
      "Epoch [   37/   50] | d_loss: 0.0730 | g_loss: 1.1330\n",
      "Epoch [   37/   50] | d_loss: 0.1051 | g_loss: 0.7745\n",
      "Epoch [   37/   50] | d_loss: 0.1520 | g_loss: 0.6644\n",
      "Epoch [   37/   50] | d_loss: 0.1774 | g_loss: 0.5306\n",
      "Epoch [   37/   50] | d_loss: 0.1205 | g_loss: 0.3623\n",
      "Epoch [   37/   50] | d_loss: 0.2975 | g_loss: 0.8699\n",
      "Epoch [   37/   50] | d_loss: 0.1129 | g_loss: 0.6775\n",
      "Epoch [   37/   50] | d_loss: 0.0925 | g_loss: 0.7915\n",
      "Epoch [   37/   50] | d_loss: 0.1679 | g_loss: 0.7139\n",
      "Epoch [   37/   50] | d_loss: 0.0978 | g_loss: 0.9491\n",
      "Epoch [   37/   50] | d_loss: 0.0706 | g_loss: 0.5817\n",
      "Epoch [   37/   50] | d_loss: 0.0741 | g_loss: 0.7463\n",
      "Epoch [   37/   50] | d_loss: 0.0849 | g_loss: 0.5266\n",
      "Epoch [   37/   50] | d_loss: 0.2385 | g_loss: 1.1517\n",
      "Epoch [   37/   50] | d_loss: 0.0877 | g_loss: 0.7306\n",
      "Epoch [   37/   50] | d_loss: 0.0731 | g_loss: 0.9725\n",
      "Epoch [   37/   50] | d_loss: 0.1077 | g_loss: 0.5464\n",
      "Epoch [   37/   50] | d_loss: 0.0674 | g_loss: 1.2701\n",
      "Epoch [   37/   50] | d_loss: 0.2256 | g_loss: 0.7005\n",
      "Epoch [   37/   50] | d_loss: 0.1168 | g_loss: 0.8049\n",
      "Epoch [   37/   50] | d_loss: 0.1105 | g_loss: 0.8132\n",
      "Epoch [   37/   50] | d_loss: 0.0811 | g_loss: 0.6350\n",
      "Epoch [   37/   50] | d_loss: 0.1009 | g_loss: 1.0288\n",
      "Epoch [   37/   50] | d_loss: 0.0920 | g_loss: 0.9017\n",
      "Epoch [   37/   50] | d_loss: 0.0712 | g_loss: 0.7870\n",
      "Epoch [   37/   50] | d_loss: 0.1428 | g_loss: 0.5975\n",
      "Epoch [   37/   50] | d_loss: 0.2035 | g_loss: 0.7364\n",
      "Epoch [   37/   50] | d_loss: 0.1699 | g_loss: 0.9479\n",
      "Epoch [   37/   50] | d_loss: 0.0531 | g_loss: 1.0863\n",
      "Epoch [   38/   50] | d_loss: 0.2470 | g_loss: 0.9474\n",
      "Epoch [   38/   50] | d_loss: 0.1263 | g_loss: 0.9689\n",
      "Epoch [   38/   50] | d_loss: 0.1237 | g_loss: 0.7961\n",
      "Epoch [   38/   50] | d_loss: 0.1998 | g_loss: 0.5760\n",
      "Epoch [   38/   50] | d_loss: 0.1851 | g_loss: 0.6368\n",
      "Epoch [   38/   50] | d_loss: 0.0874 | g_loss: 0.6671\n",
      "Epoch [   38/   50] | d_loss: 0.0806 | g_loss: 0.7802\n",
      "Epoch [   38/   50] | d_loss: 0.0610 | g_loss: 1.0893\n",
      "Epoch [   38/   50] | d_loss: 0.1756 | g_loss: 0.4664\n",
      "Epoch [   38/   50] | d_loss: 0.0697 | g_loss: 0.7657\n",
      "Epoch [   38/   50] | d_loss: 0.1384 | g_loss: 0.7458\n",
      "Epoch [   38/   50] | d_loss: 0.1876 | g_loss: 0.5698\n",
      "Epoch [   38/   50] | d_loss: 0.1238 | g_loss: 0.7612\n",
      "Epoch [   38/   50] | d_loss: 0.1819 | g_loss: 0.7002\n",
      "Epoch [   38/   50] | d_loss: 0.0788 | g_loss: 0.9978\n",
      "Epoch [   38/   50] | d_loss: 0.1244 | g_loss: 0.5400\n",
      "Epoch [   38/   50] | d_loss: 0.0710 | g_loss: 1.0286\n",
      "Epoch [   38/   50] | d_loss: 0.0922 | g_loss: 0.7430\n",
      "Epoch [   38/   50] | d_loss: 0.0615 | g_loss: 1.1494\n",
      "Epoch [   38/   50] | d_loss: 0.0937 | g_loss: 1.0322\n",
      "Epoch [   38/   50] | d_loss: 0.0960 | g_loss: 0.9262\n",
      "Epoch [   38/   50] | d_loss: 0.0798 | g_loss: 0.7650\n",
      "Epoch [   38/   50] | d_loss: 0.0633 | g_loss: 0.7545\n",
      "Epoch [   38/   50] | d_loss: 0.0677 | g_loss: 1.0526\n",
      "Epoch [   38/   50] | d_loss: 0.1625 | g_loss: 0.9476\n",
      "Epoch [   38/   50] | d_loss: 0.0934 | g_loss: 1.0536\n",
      "Epoch [   38/   50] | d_loss: 0.0677 | g_loss: 1.0713\n",
      "Epoch [   38/   50] | d_loss: 0.0664 | g_loss: 1.1214\n",
      "Epoch [   38/   50] | d_loss: 0.0841 | g_loss: 0.9960\n",
      "Epoch [   38/   50] | d_loss: 0.0663 | g_loss: 1.0130\n",
      "Epoch [   38/   50] | d_loss: 0.1032 | g_loss: 0.6355\n",
      "Epoch [   38/   50] | d_loss: 0.0632 | g_loss: 0.7309\n",
      "Epoch [   38/   50] | d_loss: 0.0664 | g_loss: 1.0273\n",
      "Epoch [   38/   50] | d_loss: 0.0682 | g_loss: 1.0593\n",
      "Epoch [   38/   50] | d_loss: 0.0924 | g_loss: 0.7405\n",
      "Epoch [   38/   50] | d_loss: 0.1050 | g_loss: 0.8530\n",
      "Epoch [   38/   50] | d_loss: 0.1387 | g_loss: 0.7542\n",
      "Epoch [   38/   50] | d_loss: 0.1055 | g_loss: 0.8251\n",
      "Epoch [   38/   50] | d_loss: 0.0744 | g_loss: 1.0208\n",
      "Epoch [   38/   50] | d_loss: 0.1181 | g_loss: 0.4743\n",
      "Epoch [   38/   50] | d_loss: 0.1117 | g_loss: 0.5055\n",
      "Epoch [   38/   50] | d_loss: 0.0471 | g_loss: 1.0197\n",
      "Epoch [   38/   50] | d_loss: 0.0738 | g_loss: 0.5827\n",
      "Epoch [   38/   50] | d_loss: 0.1638 | g_loss: 0.7982\n",
      "Epoch [   38/   50] | d_loss: 0.0926 | g_loss: 0.5910\n",
      "Epoch [   38/   50] | d_loss: 0.2582 | g_loss: 0.4334\n",
      "Epoch [   38/   50] | d_loss: 0.1192 | g_loss: 1.1342\n",
      "Epoch [   38/   50] | d_loss: 0.0768 | g_loss: 0.6664\n",
      "Epoch [   38/   50] | d_loss: 0.0843 | g_loss: 0.8273\n",
      "Epoch [   38/   50] | d_loss: 0.1389 | g_loss: 0.7463\n",
      "Epoch [   38/   50] | d_loss: 0.2396 | g_loss: 1.4041\n",
      "Epoch [   38/   50] | d_loss: 0.1327 | g_loss: 0.4649\n",
      "Epoch [   38/   50] | d_loss: 0.0769 | g_loss: 0.9303\n",
      "Epoch [   38/   50] | d_loss: 0.1050 | g_loss: 1.0935\n",
      "Epoch [   38/   50] | d_loss: 0.0563 | g_loss: 0.8542\n",
      "Epoch [   38/   50] | d_loss: 0.1782 | g_loss: 1.1066\n",
      "Epoch [   38/   50] | d_loss: 0.1951 | g_loss: 1.0134\n",
      "Epoch [   39/   50] | d_loss: 0.1120 | g_loss: 0.7566\n",
      "Epoch [   39/   50] | d_loss: 0.1307 | g_loss: 0.8839\n",
      "Epoch [   39/   50] | d_loss: 0.0556 | g_loss: 1.1146\n",
      "Epoch [   39/   50] | d_loss: 0.1264 | g_loss: 0.8711\n",
      "Epoch [   39/   50] | d_loss: 0.1078 | g_loss: 0.6667\n",
      "Epoch [   39/   50] | d_loss: 0.1382 | g_loss: 1.3381\n",
      "Epoch [   39/   50] | d_loss: 0.1432 | g_loss: 0.5261\n",
      "Epoch [   39/   50] | d_loss: 0.1474 | g_loss: 0.5140\n",
      "Epoch [   39/   50] | d_loss: 0.0573 | g_loss: 0.3572\n",
      "Epoch [   39/   50] | d_loss: 0.1733 | g_loss: 0.8115\n",
      "Epoch [   39/   50] | d_loss: 0.0473 | g_loss: 0.7551\n",
      "Epoch [   39/   50] | d_loss: 0.2287 | g_loss: 0.5274\n",
      "Epoch [   39/   50] | d_loss: 0.0815 | g_loss: 0.8041\n",
      "Epoch [   39/   50] | d_loss: 0.0759 | g_loss: 0.6535\n",
      "Epoch [   39/   50] | d_loss: 0.0326 | g_loss: 1.8252\n",
      "Epoch [   39/   50] | d_loss: 0.0501 | g_loss: 1.0355\n",
      "Epoch [   39/   50] | d_loss: 0.0850 | g_loss: 1.1206\n",
      "Epoch [   39/   50] | d_loss: 0.0698 | g_loss: 0.7902\n",
      "Epoch [   39/   50] | d_loss: 0.0700 | g_loss: 0.7401\n",
      "Epoch [   39/   50] | d_loss: 0.1112 | g_loss: 1.0091\n",
      "Epoch [   39/   50] | d_loss: 0.0820 | g_loss: 0.6787\n",
      "Epoch [   39/   50] | d_loss: 0.1462 | g_loss: 0.5964\n",
      "Epoch [   39/   50] | d_loss: 0.0997 | g_loss: 1.1544\n",
      "Epoch [   39/   50] | d_loss: 0.1239 | g_loss: 1.0392\n",
      "Epoch [   39/   50] | d_loss: 0.1872 | g_loss: 0.5673\n",
      "Epoch [   39/   50] | d_loss: 0.0790 | g_loss: 0.9261\n",
      "Epoch [   39/   50] | d_loss: 0.0937 | g_loss: 0.8233\n",
      "Epoch [   39/   50] | d_loss: 0.0809 | g_loss: 0.9234\n",
      "Epoch [   39/   50] | d_loss: 0.1077 | g_loss: 0.8431\n",
      "Epoch [   39/   50] | d_loss: 0.0452 | g_loss: 1.0456\n",
      "Epoch [   39/   50] | d_loss: 0.1447 | g_loss: 0.6810\n",
      "Epoch [   39/   50] | d_loss: 0.1210 | g_loss: 1.0876\n",
      "Epoch [   39/   50] | d_loss: 0.0849 | g_loss: 0.7532\n",
      "Epoch [   39/   50] | d_loss: 0.0856 | g_loss: 0.5274\n",
      "Epoch [   39/   50] | d_loss: 0.0806 | g_loss: 0.9869\n",
      "Epoch [   39/   50] | d_loss: 0.1046 | g_loss: 0.8889\n",
      "Epoch [   39/   50] | d_loss: 0.1278 | g_loss: 0.8421\n",
      "Epoch [   39/   50] | d_loss: 0.0492 | g_loss: 1.1942\n",
      "Epoch [   39/   50] | d_loss: 0.1049 | g_loss: 1.0253\n",
      "Epoch [   39/   50] | d_loss: 0.1648 | g_loss: 0.8367\n",
      "Epoch [   39/   50] | d_loss: 0.2699 | g_loss: 0.6721\n",
      "Epoch [   39/   50] | d_loss: 0.1136 | g_loss: 0.7367\n",
      "Epoch [   39/   50] | d_loss: 0.0965 | g_loss: 0.9860\n",
      "Epoch [   39/   50] | d_loss: 0.2699 | g_loss: 0.6076\n",
      "Epoch [   39/   50] | d_loss: 0.1511 | g_loss: 0.6485\n",
      "Epoch [   39/   50] | d_loss: 0.0756 | g_loss: 1.1037\n",
      "Epoch [   39/   50] | d_loss: 0.3131 | g_loss: 0.8408\n",
      "Epoch [   39/   50] | d_loss: 0.0791 | g_loss: 1.0726\n",
      "Epoch [   39/   50] | d_loss: 0.0788 | g_loss: 0.7035\n",
      "Epoch [   39/   50] | d_loss: 0.1111 | g_loss: 0.6844\n",
      "Epoch [   39/   50] | d_loss: 0.1721 | g_loss: 0.7477\n",
      "Epoch [   39/   50] | d_loss: 0.0929 | g_loss: 0.7936\n",
      "Epoch [   39/   50] | d_loss: 0.1168 | g_loss: 0.9196\n",
      "Epoch [   39/   50] | d_loss: 0.0733 | g_loss: 1.1310\n",
      "Epoch [   39/   50] | d_loss: 0.2425 | g_loss: 0.3864\n",
      "Epoch [   39/   50] | d_loss: 0.1374 | g_loss: 0.5370\n",
      "Epoch [   39/   50] | d_loss: 0.1715 | g_loss: 0.8006\n",
      "Epoch [   40/   50] | d_loss: 0.1238 | g_loss: 0.5989\n",
      "Epoch [   40/   50] | d_loss: 0.0830 | g_loss: 1.0055\n",
      "Epoch [   40/   50] | d_loss: 0.1047 | g_loss: 0.6626\n",
      "Epoch [   40/   50] | d_loss: 0.0952 | g_loss: 1.2452\n",
      "Epoch [   40/   50] | d_loss: 0.0572 | g_loss: 0.6416\n",
      "Epoch [   40/   50] | d_loss: 0.1361 | g_loss: 0.7439\n",
      "Epoch [   40/   50] | d_loss: 0.0472 | g_loss: 0.9963\n",
      "Epoch [   40/   50] | d_loss: 0.0527 | g_loss: 1.1663\n",
      "Epoch [   40/   50] | d_loss: 0.0995 | g_loss: 0.5444\n",
      "Epoch [   40/   50] | d_loss: 0.0754 | g_loss: 1.0555\n",
      "Epoch [   40/   50] | d_loss: 0.0562 | g_loss: 1.2909\n",
      "Epoch [   40/   50] | d_loss: 0.1194 | g_loss: 0.4202\n",
      "Epoch [   40/   50] | d_loss: 0.1295 | g_loss: 0.8556\n",
      "Epoch [   40/   50] | d_loss: 0.1628 | g_loss: 1.2863\n",
      "Epoch [   40/   50] | d_loss: 0.1283 | g_loss: 1.0428\n",
      "Epoch [   40/   50] | d_loss: 0.0628 | g_loss: 1.0686\n",
      "Epoch [   40/   50] | d_loss: 0.1868 | g_loss: 0.6174\n",
      "Epoch [   40/   50] | d_loss: 0.0709 | g_loss: 0.9771\n",
      "Epoch [   40/   50] | d_loss: 0.1365 | g_loss: 0.6699\n",
      "Epoch [   40/   50] | d_loss: 0.1227 | g_loss: 0.6863\n",
      "Epoch [   40/   50] | d_loss: 0.2040 | g_loss: 0.3846\n",
      "Epoch [   40/   50] | d_loss: 0.0832 | g_loss: 0.9678\n",
      "Epoch [   40/   50] | d_loss: 0.1806 | g_loss: 1.0921\n",
      "Epoch [   40/   50] | d_loss: 0.0972 | g_loss: 0.7939\n",
      "Epoch [   40/   50] | d_loss: 0.0685 | g_loss: 0.9331\n",
      "Epoch [   40/   50] | d_loss: 0.0872 | g_loss: 0.7507\n",
      "Epoch [   40/   50] | d_loss: 0.0954 | g_loss: 1.1462\n",
      "Epoch [   40/   50] | d_loss: 0.0758 | g_loss: 0.8152\n",
      "Epoch [   40/   50] | d_loss: 0.2049 | g_loss: 0.8784\n",
      "Epoch [   40/   50] | d_loss: 0.0655 | g_loss: 1.0889\n",
      "Epoch [   40/   50] | d_loss: 0.0648 | g_loss: 0.7913\n",
      "Epoch [   40/   50] | d_loss: 0.0778 | g_loss: 0.6209\n",
      "Epoch [   40/   50] | d_loss: 0.0557 | g_loss: 0.9532\n",
      "Epoch [   40/   50] | d_loss: 0.0559 | g_loss: 1.0402\n",
      "Epoch [   40/   50] | d_loss: 0.1894 | g_loss: 0.6882\n",
      "Epoch [   40/   50] | d_loss: 0.1295 | g_loss: 0.6376\n",
      "Epoch [   40/   50] | d_loss: 0.2217 | g_loss: 0.7707\n",
      "Epoch [   40/   50] | d_loss: 0.0811 | g_loss: 0.7565\n",
      "Epoch [   40/   50] | d_loss: 0.1018 | g_loss: 0.7805\n",
      "Epoch [   40/   50] | d_loss: 0.1047 | g_loss: 0.8336\n",
      "Epoch [   40/   50] | d_loss: 0.1131 | g_loss: 1.0991\n",
      "Epoch [   40/   50] | d_loss: 0.1204 | g_loss: 1.2202\n",
      "Epoch [   40/   50] | d_loss: 0.0510 | g_loss: 0.9929\n",
      "Epoch [   40/   50] | d_loss: 0.0731 | g_loss: 0.4603\n",
      "Epoch [   40/   50] | d_loss: 0.1769 | g_loss: 1.1963\n",
      "Epoch [   40/   50] | d_loss: 0.0438 | g_loss: 0.7020\n",
      "Epoch [   40/   50] | d_loss: 0.1526 | g_loss: 1.0480\n",
      "Epoch [   40/   50] | d_loss: 0.0757 | g_loss: 0.9580\n",
      "Epoch [   40/   50] | d_loss: 0.2161 | g_loss: 0.9110\n",
      "Epoch [   40/   50] | d_loss: 0.1592 | g_loss: 0.7748\n",
      "Epoch [   40/   50] | d_loss: 0.1161 | g_loss: 1.5144\n",
      "Epoch [   40/   50] | d_loss: 0.1281 | g_loss: 0.9761\n",
      "Epoch [   40/   50] | d_loss: 0.0535 | g_loss: 0.8557\n",
      "Epoch [   40/   50] | d_loss: 0.0644 | g_loss: 0.8275\n",
      "Epoch [   40/   50] | d_loss: 0.1183 | g_loss: 0.6205\n",
      "Epoch [   40/   50] | d_loss: 0.0860 | g_loss: 0.8618\n",
      "Epoch [   40/   50] | d_loss: 0.0865 | g_loss: 1.0367\n",
      "Epoch [   41/   50] | d_loss: 0.4496 | g_loss: 1.2143\n",
      "Epoch [   41/   50] | d_loss: 0.0466 | g_loss: 0.8591\n",
      "Epoch [   41/   50] | d_loss: 0.1366 | g_loss: 0.4275\n",
      "Epoch [   41/   50] | d_loss: 0.1885 | g_loss: 0.5852\n",
      "Epoch [   41/   50] | d_loss: 0.3569 | g_loss: 0.6165\n",
      "Epoch [   41/   50] | d_loss: 0.0550 | g_loss: 0.7750\n",
      "Epoch [   41/   50] | d_loss: 0.0683 | g_loss: 0.8694\n",
      "Epoch [   41/   50] | d_loss: 0.1610 | g_loss: 0.7259\n",
      "Epoch [   41/   50] | d_loss: 0.0455 | g_loss: 0.7370\n",
      "Epoch [   41/   50] | d_loss: 0.0693 | g_loss: 0.7000\n",
      "Epoch [   41/   50] | d_loss: 0.1377 | g_loss: 0.7118\n",
      "Epoch [   41/   50] | d_loss: 0.1373 | g_loss: 0.8387\n",
      "Epoch [   41/   50] | d_loss: 0.2024 | g_loss: 0.6887\n",
      "Epoch [   41/   50] | d_loss: 0.1373 | g_loss: 0.4731\n",
      "Epoch [   41/   50] | d_loss: 0.0786 | g_loss: 1.0296\n",
      "Epoch [   41/   50] | d_loss: 0.0684 | g_loss: 0.6425\n",
      "Epoch [   41/   50] | d_loss: 0.2502 | g_loss: 1.2216\n",
      "Epoch [   41/   50] | d_loss: 0.1122 | g_loss: 0.5017\n",
      "Epoch [   41/   50] | d_loss: 0.0803 | g_loss: 0.7268\n",
      "Epoch [   41/   50] | d_loss: 0.1018 | g_loss: 0.4792\n",
      "Epoch [   41/   50] | d_loss: 0.1844 | g_loss: 0.8396\n",
      "Epoch [   41/   50] | d_loss: 0.0692 | g_loss: 1.1378\n",
      "Epoch [   41/   50] | d_loss: 0.2006 | g_loss: 0.5210\n",
      "Epoch [   41/   50] | d_loss: 0.0986 | g_loss: 1.1697\n",
      "Epoch [   41/   50] | d_loss: 0.0544 | g_loss: 0.8926\n",
      "Epoch [   41/   50] | d_loss: 0.0736 | g_loss: 0.9170\n",
      "Epoch [   41/   50] | d_loss: 0.1107 | g_loss: 0.6858\n",
      "Epoch [   41/   50] | d_loss: 0.1833 | g_loss: 0.5922\n",
      "Epoch [   41/   50] | d_loss: 0.0842 | g_loss: 0.7628\n",
      "Epoch [   41/   50] | d_loss: 0.1822 | g_loss: 0.9408\n",
      "Epoch [   41/   50] | d_loss: 0.2423 | g_loss: 0.8913\n",
      "Epoch [   41/   50] | d_loss: 0.1098 | g_loss: 0.5501\n",
      "Epoch [   41/   50] | d_loss: 0.0952 | g_loss: 0.5866\n",
      "Epoch [   41/   50] | d_loss: 0.0864 | g_loss: 0.7442\n",
      "Epoch [   41/   50] | d_loss: 0.2368 | g_loss: 0.6172\n",
      "Epoch [   41/   50] | d_loss: 0.1795 | g_loss: 0.5910\n",
      "Epoch [   41/   50] | d_loss: 0.1742 | g_loss: 0.7060\n",
      "Epoch [   41/   50] | d_loss: 0.2570 | g_loss: 1.2317\n",
      "Epoch [   41/   50] | d_loss: 0.0815 | g_loss: 0.7098\n",
      "Epoch [   41/   50] | d_loss: 0.1469 | g_loss: 1.3368\n",
      "Epoch [   41/   50] | d_loss: 0.0964 | g_loss: 0.9285\n",
      "Epoch [   41/   50] | d_loss: 0.1052 | g_loss: 0.5472\n",
      "Epoch [   41/   50] | d_loss: 0.2171 | g_loss: 0.4065\n",
      "Epoch [   41/   50] | d_loss: 0.1425 | g_loss: 0.6621\n",
      "Epoch [   41/   50] | d_loss: 0.0879 | g_loss: 0.7451\n",
      "Epoch [   41/   50] | d_loss: 0.1792 | g_loss: 0.8470\n",
      "Epoch [   41/   50] | d_loss: 0.0924 | g_loss: 0.8315\n",
      "Epoch [   41/   50] | d_loss: 0.1461 | g_loss: 0.6174\n",
      "Epoch [   41/   50] | d_loss: 0.0856 | g_loss: 0.9626\n",
      "Epoch [   41/   50] | d_loss: 0.1114 | g_loss: 0.5812\n",
      "Epoch [   41/   50] | d_loss: 0.0539 | g_loss: 1.0913\n",
      "Epoch [   41/   50] | d_loss: 0.0772 | g_loss: 0.5764\n",
      "Epoch [   41/   50] | d_loss: 0.0477 | g_loss: 0.8105\n",
      "Epoch [   41/   50] | d_loss: 0.0784 | g_loss: 0.5407\n",
      "Epoch [   41/   50] | d_loss: 0.1348 | g_loss: 0.8201\n",
      "Epoch [   41/   50] | d_loss: 0.1139 | g_loss: 1.0953\n",
      "Epoch [   41/   50] | d_loss: 0.1068 | g_loss: 0.7557\n",
      "Epoch [   42/   50] | d_loss: 0.1974 | g_loss: 0.6739\n",
      "Epoch [   42/   50] | d_loss: 0.0629 | g_loss: 0.9897\n",
      "Epoch [   42/   50] | d_loss: 0.0507 | g_loss: 0.8463\n",
      "Epoch [   42/   50] | d_loss: 0.2631 | g_loss: 0.7262\n",
      "Epoch [   42/   50] | d_loss: 0.1536 | g_loss: 0.3045\n",
      "Epoch [   42/   50] | d_loss: 0.0854 | g_loss: 0.5443\n",
      "Epoch [   42/   50] | d_loss: 0.0787 | g_loss: 1.4358\n",
      "Epoch [   42/   50] | d_loss: 0.2126 | g_loss: 1.6493\n",
      "Epoch [   42/   50] | d_loss: 0.0785 | g_loss: 0.9434\n",
      "Epoch [   42/   50] | d_loss: 0.0730 | g_loss: 0.6170\n",
      "Epoch [   42/   50] | d_loss: 0.2768 | g_loss: 0.8875\n",
      "Epoch [   42/   50] | d_loss: 0.1152 | g_loss: 1.0590\n",
      "Epoch [   42/   50] | d_loss: 0.1717 | g_loss: 0.9140\n",
      "Epoch [   42/   50] | d_loss: 0.0625 | g_loss: 0.4839\n",
      "Epoch [   42/   50] | d_loss: 0.1627 | g_loss: 0.8732\n",
      "Epoch [   42/   50] | d_loss: 0.0579 | g_loss: 0.8176\n",
      "Epoch [   42/   50] | d_loss: 0.0728 | g_loss: 0.7232\n",
      "Epoch [   42/   50] | d_loss: 0.0724 | g_loss: 0.4458\n",
      "Epoch [   42/   50] | d_loss: 0.2546 | g_loss: 0.5493\n",
      "Epoch [   42/   50] | d_loss: 0.2066 | g_loss: 0.3744\n",
      "Epoch [   42/   50] | d_loss: 0.0567 | g_loss: 0.6975\n",
      "Epoch [   42/   50] | d_loss: 0.0625 | g_loss: 1.0635\n",
      "Epoch [   42/   50] | d_loss: 0.0503 | g_loss: 0.9411\n",
      "Epoch [   42/   50] | d_loss: 0.0776 | g_loss: 0.8102\n",
      "Epoch [   42/   50] | d_loss: 0.3510 | g_loss: 0.4170\n",
      "Epoch [   42/   50] | d_loss: 0.1035 | g_loss: 0.7006\n",
      "Epoch [   42/   50] | d_loss: 0.0830 | g_loss: 0.8451\n",
      "Epoch [   42/   50] | d_loss: 0.3052 | g_loss: 0.4633\n",
      "Epoch [   42/   50] | d_loss: 0.0624 | g_loss: 0.9964\n",
      "Epoch [   42/   50] | d_loss: 0.0600 | g_loss: 1.0167\n",
      "Epoch [   42/   50] | d_loss: 0.0934 | g_loss: 0.8245\n",
      "Epoch [   42/   50] | d_loss: 0.1953 | g_loss: 1.1422\n",
      "Epoch [   42/   50] | d_loss: 0.1460 | g_loss: 0.6954\n",
      "Epoch [   42/   50] | d_loss: 0.0578 | g_loss: 0.5293\n",
      "Epoch [   42/   50] | d_loss: 0.1808 | g_loss: 0.9787\n",
      "Epoch [   42/   50] | d_loss: 0.0838 | g_loss: 1.3154\n",
      "Epoch [   42/   50] | d_loss: 0.1645 | g_loss: 0.5752\n",
      "Epoch [   42/   50] | d_loss: 0.0806 | g_loss: 0.7533\n",
      "Epoch [   42/   50] | d_loss: 0.1218 | g_loss: 1.2563\n",
      "Epoch [   42/   50] | d_loss: 0.1114 | g_loss: 0.4026\n",
      "Epoch [   42/   50] | d_loss: 0.0896 | g_loss: 0.8588\n",
      "Epoch [   42/   50] | d_loss: 0.0565 | g_loss: 0.5668\n",
      "Epoch [   42/   50] | d_loss: 0.1025 | g_loss: 0.6824\n",
      "Epoch [   42/   50] | d_loss: 0.1647 | g_loss: 0.6109\n",
      "Epoch [   42/   50] | d_loss: 0.0484 | g_loss: 0.9285\n",
      "Epoch [   42/   50] | d_loss: 0.0771 | g_loss: 0.7471\n",
      "Epoch [   42/   50] | d_loss: 0.1053 | g_loss: 0.9654\n",
      "Epoch [   42/   50] | d_loss: 0.0652 | g_loss: 0.9820\n",
      "Epoch [   42/   50] | d_loss: 0.1414 | g_loss: 0.8395\n",
      "Epoch [   42/   50] | d_loss: 0.1069 | g_loss: 0.9615\n",
      "Epoch [   42/   50] | d_loss: 0.0681 | g_loss: 0.6593\n",
      "Epoch [   42/   50] | d_loss: 0.1541 | g_loss: 0.7520\n",
      "Epoch [   42/   50] | d_loss: 0.2189 | g_loss: 0.8154\n",
      "Epoch [   42/   50] | d_loss: 0.1404 | g_loss: 0.8564\n",
      "Epoch [   42/   50] | d_loss: 0.2219 | g_loss: 0.3582\n",
      "Epoch [   42/   50] | d_loss: 0.1643 | g_loss: 1.1581\n",
      "Epoch [   42/   50] | d_loss: 0.1447 | g_loss: 0.9964\n",
      "Epoch [   43/   50] | d_loss: 0.1383 | g_loss: 0.4216\n",
      "Epoch [   43/   50] | d_loss: 0.5391 | g_loss: 0.2155\n",
      "Epoch [   43/   50] | d_loss: 0.2424 | g_loss: 1.2903\n",
      "Epoch [   43/   50] | d_loss: 0.1785 | g_loss: 0.2792\n",
      "Epoch [   43/   50] | d_loss: 0.0846 | g_loss: 0.5062\n",
      "Epoch [   43/   50] | d_loss: 0.1641 | g_loss: 0.4225\n",
      "Epoch [   43/   50] | d_loss: 0.0727 | g_loss: 1.0042\n",
      "Epoch [   43/   50] | d_loss: 0.0817 | g_loss: 0.9691\n",
      "Epoch [   43/   50] | d_loss: 0.1029 | g_loss: 0.6958\n",
      "Epoch [   43/   50] | d_loss: 0.0887 | g_loss: 1.0232\n",
      "Epoch [   43/   50] | d_loss: 0.0604 | g_loss: 0.9286\n",
      "Epoch [   43/   50] | d_loss: 0.1128 | g_loss: 0.9296\n",
      "Epoch [   43/   50] | d_loss: 0.1333 | g_loss: 0.8830\n",
      "Epoch [   43/   50] | d_loss: 0.1307 | g_loss: 0.6803\n",
      "Epoch [   43/   50] | d_loss: 0.0751 | g_loss: 1.2415\n",
      "Epoch [   43/   50] | d_loss: 0.1512 | g_loss: 1.0023\n",
      "Epoch [   43/   50] | d_loss: 0.1243 | g_loss: 0.4096\n",
      "Epoch [   43/   50] | d_loss: 0.0530 | g_loss: 0.8546\n",
      "Epoch [   43/   50] | d_loss: 0.1275 | g_loss: 0.8891\n",
      "Epoch [   43/   50] | d_loss: 0.2040 | g_loss: 0.6784\n",
      "Epoch [   43/   50] | d_loss: 0.1527 | g_loss: 1.1501\n",
      "Epoch [   43/   50] | d_loss: 0.1029 | g_loss: 0.5768\n",
      "Epoch [   43/   50] | d_loss: 0.0846 | g_loss: 0.7203\n",
      "Epoch [   43/   50] | d_loss: 0.0829 | g_loss: 0.8968\n",
      "Epoch [   43/   50] | d_loss: 0.2161 | g_loss: 1.0627\n",
      "Epoch [   43/   50] | d_loss: 0.1396 | g_loss: 0.6799\n",
      "Epoch [   43/   50] | d_loss: 0.1527 | g_loss: 0.6710\n",
      "Epoch [   43/   50] | d_loss: 0.1282 | g_loss: 0.9495\n",
      "Epoch [   43/   50] | d_loss: 0.0997 | g_loss: 1.2063\n",
      "Epoch [   43/   50] | d_loss: 0.0674 | g_loss: 0.9010\n",
      "Epoch [   43/   50] | d_loss: 0.1159 | g_loss: 1.1794\n",
      "Epoch [   43/   50] | d_loss: 0.0879 | g_loss: 0.4761\n",
      "Epoch [   43/   50] | d_loss: 0.0785 | g_loss: 0.6767\n",
      "Epoch [   43/   50] | d_loss: 0.0428 | g_loss: 0.6607\n",
      "Epoch [   43/   50] | d_loss: 0.1496 | g_loss: 0.8663\n",
      "Epoch [   43/   50] | d_loss: 0.3012 | g_loss: 0.4288\n",
      "Epoch [   43/   50] | d_loss: 0.1068 | g_loss: 0.8909\n",
      "Epoch [   43/   50] | d_loss: 0.1099 | g_loss: 0.5371\n",
      "Epoch [   43/   50] | d_loss: 0.1160 | g_loss: 0.7486\n",
      "Epoch [   43/   50] | d_loss: 0.0769 | g_loss: 1.1986\n",
      "Epoch [   43/   50] | d_loss: 0.1473 | g_loss: 0.5845\n",
      "Epoch [   43/   50] | d_loss: 0.0804 | g_loss: 0.6391\n",
      "Epoch [   43/   50] | d_loss: 0.0527 | g_loss: 0.9799\n",
      "Epoch [   43/   50] | d_loss: 0.0543 | g_loss: 1.0388\n",
      "Epoch [   43/   50] | d_loss: 0.1202 | g_loss: 0.7780\n",
      "Epoch [   43/   50] | d_loss: 0.2449 | g_loss: 0.4105\n",
      "Epoch [   43/   50] | d_loss: 0.1404 | g_loss: 0.9794\n",
      "Epoch [   43/   50] | d_loss: 0.0811 | g_loss: 0.9588\n",
      "Epoch [   43/   50] | d_loss: 0.1269 | g_loss: 0.5335\n",
      "Epoch [   43/   50] | d_loss: 0.1274 | g_loss: 0.7580\n",
      "Epoch [   43/   50] | d_loss: 0.0512 | g_loss: 0.7771\n",
      "Epoch [   43/   50] | d_loss: 0.1792 | g_loss: 1.3044\n",
      "Epoch [   43/   50] | d_loss: 0.0347 | g_loss: 0.9973\n",
      "Epoch [   43/   50] | d_loss: 0.0779 | g_loss: 0.7963\n",
      "Epoch [   43/   50] | d_loss: 0.1032 | g_loss: 0.9473\n",
      "Epoch [   43/   50] | d_loss: 0.0860 | g_loss: 0.7471\n",
      "Epoch [   43/   50] | d_loss: 0.1280 | g_loss: 0.8400\n",
      "Epoch [   44/   50] | d_loss: 0.2529 | g_loss: 0.8273\n",
      "Epoch [   44/   50] | d_loss: 0.0870 | g_loss: 0.7037\n",
      "Epoch [   44/   50] | d_loss: 0.0871 | g_loss: 1.0453\n",
      "Epoch [   44/   50] | d_loss: 0.1514 | g_loss: 0.4755\n",
      "Epoch [   44/   50] | d_loss: 0.2108 | g_loss: 1.0700\n",
      "Epoch [   44/   50] | d_loss: 0.1027 | g_loss: 0.8462\n",
      "Epoch [   44/   50] | d_loss: 0.0832 | g_loss: 0.6564\n",
      "Epoch [   44/   50] | d_loss: 0.2593 | g_loss: 0.4610\n",
      "Epoch [   44/   50] | d_loss: 0.0650 | g_loss: 0.9721\n",
      "Epoch [   44/   50] | d_loss: 0.2210 | g_loss: 0.6876\n",
      "Epoch [   44/   50] | d_loss: 0.0330 | g_loss: 0.8241\n",
      "Epoch [   44/   50] | d_loss: 0.0855 | g_loss: 1.1496\n",
      "Epoch [   44/   50] | d_loss: 0.0587 | g_loss: 1.0423\n",
      "Epoch [   44/   50] | d_loss: 0.1194 | g_loss: 1.0528\n",
      "Epoch [   44/   50] | d_loss: 0.0534 | g_loss: 1.0141\n",
      "Epoch [   44/   50] | d_loss: 0.1102 | g_loss: 0.8313\n",
      "Epoch [   44/   50] | d_loss: 0.1263 | g_loss: 0.5798\n",
      "Epoch [   44/   50] | d_loss: 0.0709 | g_loss: 0.7018\n",
      "Epoch [   44/   50] | d_loss: 0.3311 | g_loss: 1.4922\n",
      "Epoch [   44/   50] | d_loss: 0.0959 | g_loss: 1.0438\n",
      "Epoch [   44/   50] | d_loss: 0.0964 | g_loss: 0.9759\n",
      "Epoch [   44/   50] | d_loss: 0.1075 | g_loss: 0.5498\n",
      "Epoch [   44/   50] | d_loss: 0.0568 | g_loss: 0.9957\n",
      "Epoch [   44/   50] | d_loss: 0.1690 | g_loss: 0.8069\n",
      "Epoch [   44/   50] | d_loss: 0.0689 | g_loss: 1.2542\n",
      "Epoch [   44/   50] | d_loss: 0.0728 | g_loss: 1.0934\n",
      "Epoch [   44/   50] | d_loss: 0.1262 | g_loss: 0.7021\n",
      "Epoch [   44/   50] | d_loss: 0.0934 | g_loss: 0.6812\n",
      "Epoch [   44/   50] | d_loss: 0.0486 | g_loss: 0.7154\n",
      "Epoch [   44/   50] | d_loss: 0.0737 | g_loss: 0.5170\n",
      "Epoch [   44/   50] | d_loss: 0.0720 | g_loss: 0.7360\n",
      "Epoch [   44/   50] | d_loss: 0.2591 | g_loss: 0.6714\n",
      "Epoch [   44/   50] | d_loss: 0.0475 | g_loss: 0.8768\n",
      "Epoch [   44/   50] | d_loss: 0.0794 | g_loss: 0.6062\n",
      "Epoch [   44/   50] | d_loss: 0.0708 | g_loss: 0.7828\n",
      "Epoch [   44/   50] | d_loss: 0.1878 | g_loss: 1.2227\n",
      "Epoch [   44/   50] | d_loss: 0.1247 | g_loss: 0.7317\n",
      "Epoch [   44/   50] | d_loss: 0.0964 | g_loss: 0.7304\n",
      "Epoch [   44/   50] | d_loss: 0.0818 | g_loss: 0.7856\n",
      "Epoch [   44/   50] | d_loss: 0.1503 | g_loss: 0.8513\n",
      "Epoch [   44/   50] | d_loss: 0.1226 | g_loss: 0.5689\n",
      "Epoch [   44/   50] | d_loss: 0.1267 | g_loss: 0.8939\n",
      "Epoch [   44/   50] | d_loss: 0.0496 | g_loss: 0.8183\n",
      "Epoch [   44/   50] | d_loss: 0.1053 | g_loss: 1.0039\n",
      "Epoch [   44/   50] | d_loss: 0.1836 | g_loss: 1.0807\n",
      "Epoch [   44/   50] | d_loss: 0.0871 | g_loss: 1.0717\n",
      "Epoch [   44/   50] | d_loss: 0.1327 | g_loss: 1.0178\n",
      "Epoch [   44/   50] | d_loss: 0.1282 | g_loss: 0.7525\n",
      "Epoch [   44/   50] | d_loss: 0.0941 | g_loss: 1.0776\n",
      "Epoch [   44/   50] | d_loss: 0.0555 | g_loss: 0.8967\n",
      "Epoch [   44/   50] | d_loss: 0.0601 | g_loss: 0.7502\n",
      "Epoch [   44/   50] | d_loss: 0.1005 | g_loss: 0.6085\n",
      "Epoch [   44/   50] | d_loss: 0.2860 | g_loss: 0.8335\n",
      "Epoch [   44/   50] | d_loss: 0.2273 | g_loss: 0.5357\n",
      "Epoch [   44/   50] | d_loss: 0.1388 | g_loss: 1.1882\n",
      "Epoch [   44/   50] | d_loss: 0.0808 | g_loss: 1.0293\n",
      "Epoch [   44/   50] | d_loss: 0.0840 | g_loss: 0.9575\n",
      "Epoch [   45/   50] | d_loss: 0.3186 | g_loss: 0.9003\n",
      "Epoch [   45/   50] | d_loss: 0.2038 | g_loss: 0.4887\n",
      "Epoch [   45/   50] | d_loss: 0.0668 | g_loss: 0.8986\n",
      "Epoch [   45/   50] | d_loss: 0.1355 | g_loss: 0.6009\n",
      "Epoch [   45/   50] | d_loss: 0.0537 | g_loss: 0.9710\n",
      "Epoch [   45/   50] | d_loss: 0.2083 | g_loss: 0.5562\n",
      "Epoch [   45/   50] | d_loss: 0.1483 | g_loss: 1.0845\n",
      "Epoch [   45/   50] | d_loss: 0.0561 | g_loss: 1.2501\n",
      "Epoch [   45/   50] | d_loss: 0.0472 | g_loss: 0.7138\n",
      "Epoch [   45/   50] | d_loss: 0.1262 | g_loss: 0.9208\n",
      "Epoch [   45/   50] | d_loss: 0.1602 | g_loss: 1.2819\n",
      "Epoch [   45/   50] | d_loss: 0.2193 | g_loss: 0.6232\n",
      "Epoch [   45/   50] | d_loss: 0.1301 | g_loss: 0.9890\n",
      "Epoch [   45/   50] | d_loss: 0.1136 | g_loss: 1.1168\n",
      "Epoch [   45/   50] | d_loss: 0.1589 | g_loss: 0.3214\n",
      "Epoch [   45/   50] | d_loss: 0.0929 | g_loss: 0.8417\n",
      "Epoch [   45/   50] | d_loss: 0.2019 | g_loss: 0.6689\n",
      "Epoch [   45/   50] | d_loss: 0.1600 | g_loss: 0.6366\n",
      "Epoch [   45/   50] | d_loss: 0.0919 | g_loss: 0.7709\n",
      "Epoch [   45/   50] | d_loss: 0.0463 | g_loss: 0.6742\n",
      "Epoch [   45/   50] | d_loss: 0.0484 | g_loss: 0.8252\n",
      "Epoch [   45/   50] | d_loss: 0.0663 | g_loss: 0.6792\n",
      "Epoch [   45/   50] | d_loss: 0.0853 | g_loss: 0.5845\n",
      "Epoch [   45/   50] | d_loss: 0.0520 | g_loss: 0.5721\n",
      "Epoch [   45/   50] | d_loss: 0.0896 | g_loss: 0.6798\n",
      "Epoch [   45/   50] | d_loss: 0.1189 | g_loss: 0.5991\n",
      "Epoch [   45/   50] | d_loss: 0.0907 | g_loss: 0.8753\n",
      "Epoch [   45/   50] | d_loss: 0.2320 | g_loss: 0.9811\n",
      "Epoch [   45/   50] | d_loss: 0.0575 | g_loss: 0.7266\n",
      "Epoch [   45/   50] | d_loss: 0.1185 | g_loss: 1.1209\n",
      "Epoch [   45/   50] | d_loss: 0.0601 | g_loss: 1.0349\n",
      "Epoch [   45/   50] | d_loss: 0.0666 | g_loss: 1.0488\n",
      "Epoch [   45/   50] | d_loss: 0.0556 | g_loss: 0.6261\n",
      "Epoch [   45/   50] | d_loss: 0.0552 | g_loss: 1.2359\n",
      "Epoch [   45/   50] | d_loss: 0.1972 | g_loss: 1.0998\n",
      "Epoch [   45/   50] | d_loss: 0.0893 | g_loss: 0.6452\n",
      "Epoch [   45/   50] | d_loss: 0.2381 | g_loss: 0.5436\n",
      "Epoch [   45/   50] | d_loss: 0.2242 | g_loss: 1.0905\n",
      "Epoch [   45/   50] | d_loss: 0.1316 | g_loss: 0.4010\n",
      "Epoch [   45/   50] | d_loss: 0.1981 | g_loss: 0.3747\n",
      "Epoch [   45/   50] | d_loss: 0.4248 | g_loss: 1.3696\n",
      "Epoch [   45/   50] | d_loss: 0.1782 | g_loss: 1.3279\n",
      "Epoch [   45/   50] | d_loss: 0.0637 | g_loss: 0.7353\n",
      "Epoch [   45/   50] | d_loss: 0.0654 | g_loss: 0.5541\n",
      "Epoch [   45/   50] | d_loss: 0.0483 | g_loss: 0.4315\n",
      "Epoch [   45/   50] | d_loss: 0.1079 | g_loss: 0.5961\n",
      "Epoch [   45/   50] | d_loss: 0.2364 | g_loss: 0.9762\n",
      "Epoch [   45/   50] | d_loss: 0.1404 | g_loss: 0.7655\n",
      "Epoch [   45/   50] | d_loss: 0.1057 | g_loss: 1.5287\n",
      "Epoch [   45/   50] | d_loss: 0.0975 | g_loss: 0.7914\n",
      "Epoch [   45/   50] | d_loss: 0.2358 | g_loss: 1.3412\n",
      "Epoch [   45/   50] | d_loss: 0.1936 | g_loss: 0.2884\n",
      "Epoch [   45/   50] | d_loss: 0.0685 | g_loss: 0.7902\n",
      "Epoch [   45/   50] | d_loss: 0.1272 | g_loss: 0.5839\n",
      "Epoch [   45/   50] | d_loss: 0.1303 | g_loss: 0.6148\n",
      "Epoch [   45/   50] | d_loss: 0.0535 | g_loss: 0.7606\n",
      "Epoch [   45/   50] | d_loss: 0.1297 | g_loss: 0.8340\n",
      "Epoch [   46/   50] | d_loss: 0.2338 | g_loss: 0.6259\n",
      "Epoch [   46/   50] | d_loss: 0.1412 | g_loss: 0.5220\n",
      "Epoch [   46/   50] | d_loss: 0.1021 | g_loss: 0.6660\n",
      "Epoch [   46/   50] | d_loss: 0.0846 | g_loss: 1.0937\n",
      "Epoch [   46/   50] | d_loss: 0.1586 | g_loss: 0.9622\n",
      "Epoch [   46/   50] | d_loss: 0.1301 | g_loss: 0.9011\n",
      "Epoch [   46/   50] | d_loss: 0.1089 | g_loss: 0.7386\n",
      "Epoch [   46/   50] | d_loss: 0.1701 | g_loss: 0.6420\n",
      "Epoch [   46/   50] | d_loss: 0.1448 | g_loss: 0.9449\n",
      "Epoch [   46/   50] | d_loss: 0.0964 | g_loss: 0.9761\n",
      "Epoch [   46/   50] | d_loss: 0.1500 | g_loss: 0.6340\n",
      "Epoch [   46/   50] | d_loss: 0.0987 | g_loss: 0.6131\n",
      "Epoch [   46/   50] | d_loss: 0.0746 | g_loss: 0.9495\n",
      "Epoch [   46/   50] | d_loss: 0.1179 | g_loss: 0.9001\n",
      "Epoch [   46/   50] | d_loss: 0.0493 | g_loss: 0.3917\n",
      "Epoch [   46/   50] | d_loss: 0.0939 | g_loss: 0.9579\n",
      "Epoch [   46/   50] | d_loss: 0.2896 | g_loss: 1.1612\n",
      "Epoch [   46/   50] | d_loss: 0.0570 | g_loss: 1.4020\n",
      "Epoch [   46/   50] | d_loss: 0.0990 | g_loss: 0.7594\n",
      "Epoch [   46/   50] | d_loss: 0.0673 | g_loss: 0.8831\n",
      "Epoch [   46/   50] | d_loss: 0.0892 | g_loss: 1.1771\n",
      "Epoch [   46/   50] | d_loss: 0.1709 | g_loss: 0.8906\n",
      "Epoch [   46/   50] | d_loss: 0.1503 | g_loss: 0.9900\n",
      "Epoch [   46/   50] | d_loss: 0.0585 | g_loss: 1.0810\n",
      "Epoch [   46/   50] | d_loss: 0.1775 | g_loss: 0.3463\n",
      "Epoch [   46/   50] | d_loss: 0.1161 | g_loss: 0.9134\n",
      "Epoch [   46/   50] | d_loss: 0.0886 | g_loss: 0.7909\n",
      "Epoch [   46/   50] | d_loss: 0.0306 | g_loss: 0.9351\n",
      "Epoch [   46/   50] | d_loss: 0.0492 | g_loss: 1.1278\n",
      "Epoch [   46/   50] | d_loss: 0.1229 | g_loss: 0.4967\n",
      "Epoch [   46/   50] | d_loss: 0.1228 | g_loss: 0.6047\n",
      "Epoch [   46/   50] | d_loss: 0.1074 | g_loss: 0.9153\n",
      "Epoch [   46/   50] | d_loss: 0.0775 | g_loss: 0.5386\n",
      "Epoch [   46/   50] | d_loss: 0.1067 | g_loss: 1.1586\n",
      "Epoch [   46/   50] | d_loss: 0.0785 | g_loss: 0.7141\n",
      "Epoch [   46/   50] | d_loss: 0.0863 | g_loss: 0.8752\n",
      "Epoch [   46/   50] | d_loss: 0.1584 | g_loss: 0.6474\n",
      "Epoch [   46/   50] | d_loss: 0.0541 | g_loss: 1.0251\n",
      "Epoch [   46/   50] | d_loss: 0.1062 | g_loss: 1.5087\n",
      "Epoch [   46/   50] | d_loss: 0.0723 | g_loss: 0.9293\n",
      "Epoch [   46/   50] | d_loss: 0.0729 | g_loss: 1.0219\n",
      "Epoch [   46/   50] | d_loss: 0.0742 | g_loss: 0.7492\n",
      "Epoch [   46/   50] | d_loss: 0.1497 | g_loss: 0.5904\n",
      "Epoch [   46/   50] | d_loss: 0.1062 | g_loss: 0.8804\n",
      "Epoch [   46/   50] | d_loss: 0.1054 | g_loss: 0.7189\n",
      "Epoch [   46/   50] | d_loss: 0.0895 | g_loss: 0.6622\n",
      "Epoch [   46/   50] | d_loss: 0.2455 | g_loss: 1.1565\n",
      "Epoch [   46/   50] | d_loss: 0.1333 | g_loss: 0.7353\n",
      "Epoch [   46/   50] | d_loss: 0.2301 | g_loss: 0.7515\n",
      "Epoch [   46/   50] | d_loss: 0.0429 | g_loss: 0.7875\n",
      "Epoch [   46/   50] | d_loss: 0.0518 | g_loss: 0.9910\n",
      "Epoch [   46/   50] | d_loss: 0.1312 | g_loss: 0.7143\n",
      "Epoch [   46/   50] | d_loss: 0.2039 | g_loss: 0.7052\n",
      "Epoch [   46/   50] | d_loss: 0.0934 | g_loss: 0.6203\n",
      "Epoch [   46/   50] | d_loss: 0.1064 | g_loss: 0.9273\n",
      "Epoch [   46/   50] | d_loss: 0.1364 | g_loss: 0.6145\n",
      "Epoch [   46/   50] | d_loss: 0.1461 | g_loss: 1.0332\n",
      "Epoch [   47/   50] | d_loss: 0.3043 | g_loss: 0.6868\n",
      "Epoch [   47/   50] | d_loss: 0.1079 | g_loss: 0.8817\n",
      "Epoch [   47/   50] | d_loss: 0.0632 | g_loss: 1.3532\n",
      "Epoch [   47/   50] | d_loss: 0.0728 | g_loss: 0.6157\n",
      "Epoch [   47/   50] | d_loss: 0.2675 | g_loss: 1.4897\n",
      "Epoch [   47/   50] | d_loss: 0.0624 | g_loss: 1.0526\n",
      "Epoch [   47/   50] | d_loss: 0.1427 | g_loss: 0.7719\n",
      "Epoch [   47/   50] | d_loss: 0.0986 | g_loss: 1.1110\n",
      "Epoch [   47/   50] | d_loss: 0.0385 | g_loss: 1.0676\n",
      "Epoch [   47/   50] | d_loss: 0.0860 | g_loss: 0.7194\n",
      "Epoch [   47/   50] | d_loss: 0.1335 | g_loss: 0.8836\n",
      "Epoch [   47/   50] | d_loss: 0.1381 | g_loss: 0.6325\n",
      "Epoch [   47/   50] | d_loss: 0.2621 | g_loss: 1.1095\n",
      "Epoch [   47/   50] | d_loss: 0.1482 | g_loss: 0.8993\n",
      "Epoch [   47/   50] | d_loss: 0.0992 | g_loss: 0.6534\n",
      "Epoch [   47/   50] | d_loss: 0.0780 | g_loss: 0.8822\n",
      "Epoch [   47/   50] | d_loss: 0.1496 | g_loss: 0.7978\n",
      "Epoch [   47/   50] | d_loss: 0.0998 | g_loss: 0.6454\n",
      "Epoch [   47/   50] | d_loss: 0.0955 | g_loss: 1.4551\n",
      "Epoch [   47/   50] | d_loss: 0.0423 | g_loss: 1.1353\n",
      "Epoch [   47/   50] | d_loss: 0.0809 | g_loss: 0.8592\n",
      "Epoch [   47/   50] | d_loss: 0.0515 | g_loss: 0.7380\n",
      "Epoch [   47/   50] | d_loss: 0.0552 | g_loss: 0.6990\n",
      "Epoch [   47/   50] | d_loss: 0.0971 | g_loss: 1.0552\n",
      "Epoch [   47/   50] | d_loss: 0.0703 | g_loss: 0.8593\n",
      "Epoch [   47/   50] | d_loss: 0.1765 | g_loss: 0.9405\n",
      "Epoch [   47/   50] | d_loss: 0.0957 | g_loss: 0.9219\n",
      "Epoch [   47/   50] | d_loss: 0.1017 | g_loss: 0.7867\n",
      "Epoch [   47/   50] | d_loss: 0.0909 | g_loss: 0.7897\n",
      "Epoch [   47/   50] | d_loss: 0.0615 | g_loss: 0.8907\n",
      "Epoch [   47/   50] | d_loss: 0.0478 | g_loss: 0.4339\n",
      "Epoch [   47/   50] | d_loss: 0.1325 | g_loss: 0.8740\n",
      "Epoch [   47/   50] | d_loss: 0.1047 | g_loss: 1.4289\n",
      "Epoch [   47/   50] | d_loss: 0.1910 | g_loss: 0.4545\n",
      "Epoch [   47/   50] | d_loss: 0.1002 | g_loss: 0.8423\n",
      "Epoch [   47/   50] | d_loss: 0.1981 | g_loss: 0.4502\n",
      "Epoch [   47/   50] | d_loss: 0.0851 | g_loss: 1.0064\n",
      "Epoch [   47/   50] | d_loss: 0.1250 | g_loss: 0.5806\n",
      "Epoch [   47/   50] | d_loss: 0.2805 | g_loss: 0.6114\n",
      "Epoch [   47/   50] | d_loss: 0.1373 | g_loss: 0.5824\n",
      "Epoch [   47/   50] | d_loss: 0.0962 | g_loss: 0.9593\n",
      "Epoch [   47/   50] | d_loss: 0.0660 | g_loss: 0.8210\n",
      "Epoch [   47/   50] | d_loss: 0.0647 | g_loss: 0.8523\n",
      "Epoch [   47/   50] | d_loss: 0.0683 | g_loss: 1.1810\n",
      "Epoch [   47/   50] | d_loss: 0.1296 | g_loss: 0.8857\n",
      "Epoch [   47/   50] | d_loss: 0.0821 | g_loss: 0.5077\n",
      "Epoch [   47/   50] | d_loss: 0.1703 | g_loss: 0.8532\n",
      "Epoch [   47/   50] | d_loss: 0.1340 | g_loss: 0.3932\n",
      "Epoch [   47/   50] | d_loss: 0.1695 | g_loss: 0.9871\n",
      "Epoch [   47/   50] | d_loss: 0.0675 | g_loss: 0.8129\n",
      "Epoch [   47/   50] | d_loss: 0.1758 | g_loss: 0.4269\n",
      "Epoch [   47/   50] | d_loss: 0.1039 | g_loss: 1.0447\n",
      "Epoch [   47/   50] | d_loss: 0.0760 | g_loss: 0.8517\n",
      "Epoch [   47/   50] | d_loss: 0.0888 | g_loss: 0.7360\n",
      "Epoch [   47/   50] | d_loss: 0.1066 | g_loss: 1.1775\n",
      "Epoch [   47/   50] | d_loss: 0.1942 | g_loss: 0.8403\n",
      "Epoch [   47/   50] | d_loss: 0.1644 | g_loss: 1.0075\n",
      "Epoch [   48/   50] | d_loss: 0.1055 | g_loss: 0.7615\n",
      "Epoch [   48/   50] | d_loss: 0.0979 | g_loss: 0.9359\n",
      "Epoch [   48/   50] | d_loss: 0.0481 | g_loss: 1.0855\n",
      "Epoch [   48/   50] | d_loss: 0.1754 | g_loss: 1.1154\n",
      "Epoch [   48/   50] | d_loss: 0.0834 | g_loss: 0.9841\n",
      "Epoch [   48/   50] | d_loss: 0.0754 | g_loss: 0.9151\n",
      "Epoch [   48/   50] | d_loss: 0.1032 | g_loss: 1.1107\n",
      "Epoch [   48/   50] | d_loss: 0.1134 | g_loss: 0.7587\n",
      "Epoch [   48/   50] | d_loss: 0.0775 | g_loss: 0.7472\n",
      "Epoch [   48/   50] | d_loss: 0.0715 | g_loss: 0.7553\n",
      "Epoch [   48/   50] | d_loss: 0.0522 | g_loss: 0.8233\n",
      "Epoch [   48/   50] | d_loss: 0.0810 | g_loss: 1.0358\n",
      "Epoch [   48/   50] | d_loss: 0.0542 | g_loss: 0.8973\n",
      "Epoch [   48/   50] | d_loss: 0.1345 | g_loss: 0.8912\n",
      "Epoch [   48/   50] | d_loss: 0.0852 | g_loss: 1.2879\n",
      "Epoch [   48/   50] | d_loss: 0.1288 | g_loss: 0.6346\n",
      "Epoch [   48/   50] | d_loss: 0.0580 | g_loss: 0.9972\n",
      "Epoch [   48/   50] | d_loss: 0.0968 | g_loss: 1.0840\n",
      "Epoch [   48/   50] | d_loss: 0.0813 | g_loss: 0.6292\n",
      "Epoch [   48/   50] | d_loss: 0.1057 | g_loss: 0.5940\n",
      "Epoch [   48/   50] | d_loss: 0.1063 | g_loss: 0.7801\n",
      "Epoch [   48/   50] | d_loss: 0.0952 | g_loss: 0.8215\n",
      "Epoch [   48/   50] | d_loss: 0.0791 | g_loss: 0.6594\n",
      "Epoch [   48/   50] | d_loss: 0.0523 | g_loss: 1.5978\n",
      "Epoch [   48/   50] | d_loss: 0.0971 | g_loss: 0.8159\n",
      "Epoch [   48/   50] | d_loss: 0.0815 | g_loss: 1.0230\n",
      "Epoch [   48/   50] | d_loss: 0.1141 | g_loss: 0.7138\n",
      "Epoch [   48/   50] | d_loss: 0.1502 | g_loss: 0.3671\n",
      "Epoch [   48/   50] | d_loss: 0.1131 | g_loss: 0.6163\n",
      "Epoch [   48/   50] | d_loss: 0.0877 | g_loss: 0.9830\n",
      "Epoch [   48/   50] | d_loss: 0.1041 | g_loss: 0.8468\n",
      "Epoch [   48/   50] | d_loss: 0.1250 | g_loss: 1.6611\n",
      "Epoch [   48/   50] | d_loss: 0.1425 | g_loss: 0.4913\n",
      "Epoch [   48/   50] | d_loss: 0.0637 | g_loss: 0.8267\n",
      "Epoch [   48/   50] | d_loss: 0.0824 | g_loss: 0.7556\n",
      "Epoch [   48/   50] | d_loss: 0.0786 | g_loss: 0.7403\n",
      "Epoch [   48/   50] | d_loss: 0.1948 | g_loss: 0.9442\n",
      "Epoch [   48/   50] | d_loss: 0.1192 | g_loss: 0.7454\n",
      "Epoch [   48/   50] | d_loss: 0.1092 | g_loss: 0.7192\n",
      "Epoch [   48/   50] | d_loss: 0.1108 | g_loss: 0.6050\n",
      "Epoch [   48/   50] | d_loss: 0.1199 | g_loss: 0.5923\n",
      "Epoch [   48/   50] | d_loss: 0.0529 | g_loss: 0.9502\n",
      "Epoch [   48/   50] | d_loss: 0.1573 | g_loss: 1.0736\n",
      "Epoch [   48/   50] | d_loss: 0.1127 | g_loss: 0.4067\n",
      "Epoch [   48/   50] | d_loss: 0.0781 | g_loss: 0.8109\n",
      "Epoch [   48/   50] | d_loss: 0.0580 | g_loss: 1.1493\n",
      "Epoch [   48/   50] | d_loss: 0.1457 | g_loss: 0.9081\n",
      "Epoch [   48/   50] | d_loss: 0.0719 | g_loss: 0.9026\n",
      "Epoch [   48/   50] | d_loss: 0.1810 | g_loss: 1.0079\n",
      "Epoch [   48/   50] | d_loss: 0.0871 | g_loss: 1.0071\n",
      "Epoch [   48/   50] | d_loss: 0.1100 | g_loss: 1.4026\n",
      "Epoch [   48/   50] | d_loss: 0.2583 | g_loss: 1.1671\n",
      "Epoch [   48/   50] | d_loss: 0.0387 | g_loss: 1.1952\n",
      "Epoch [   48/   50] | d_loss: 0.2772 | g_loss: 0.2518\n",
      "Epoch [   48/   50] | d_loss: 0.0673 | g_loss: 0.7050\n",
      "Epoch [   48/   50] | d_loss: 0.1350 | g_loss: 0.8977\n",
      "Epoch [   48/   50] | d_loss: 0.0793 | g_loss: 0.6565\n",
      "Epoch [   49/   50] | d_loss: 0.1383 | g_loss: 1.0662\n",
      "Epoch [   49/   50] | d_loss: 0.1222 | g_loss: 1.3994\n",
      "Epoch [   49/   50] | d_loss: 0.0419 | g_loss: 0.8077\n",
      "Epoch [   49/   50] | d_loss: 0.0789 | g_loss: 0.7598\n",
      "Epoch [   49/   50] | d_loss: 0.2443 | g_loss: 0.5568\n",
      "Epoch [   49/   50] | d_loss: 0.0914 | g_loss: 0.9335\n",
      "Epoch [   49/   50] | d_loss: 0.0739 | g_loss: 0.9304\n",
      "Epoch [   49/   50] | d_loss: 0.1925 | g_loss: 0.5870\n",
      "Epoch [   49/   50] | d_loss: 0.1689 | g_loss: 0.8093\n",
      "Epoch [   49/   50] | d_loss: 0.1287 | g_loss: 0.5180\n",
      "Epoch [   49/   50] | d_loss: 0.1092 | g_loss: 0.5304\n",
      "Epoch [   49/   50] | d_loss: 0.0549 | g_loss: 0.7750\n",
      "Epoch [   49/   50] | d_loss: 0.1537 | g_loss: 1.0550\n",
      "Epoch [   49/   50] | d_loss: 0.0874 | g_loss: 0.8265\n",
      "Epoch [   49/   50] | d_loss: 0.0582 | g_loss: 0.8404\n",
      "Epoch [   49/   50] | d_loss: 0.0887 | g_loss: 0.8410\n",
      "Epoch [   49/   50] | d_loss: 0.2930 | g_loss: 0.3669\n",
      "Epoch [   49/   50] | d_loss: 0.0565 | g_loss: 1.0133\n",
      "Epoch [   49/   50] | d_loss: 0.0732 | g_loss: 1.0263\n",
      "Epoch [   49/   50] | d_loss: 0.0983 | g_loss: 0.9341\n",
      "Epoch [   49/   50] | d_loss: 0.0386 | g_loss: 1.3751\n",
      "Epoch [   49/   50] | d_loss: 0.0761 | g_loss: 0.8914\n",
      "Epoch [   49/   50] | d_loss: 0.1311 | g_loss: 0.9613\n",
      "Epoch [   49/   50] | d_loss: 0.0957 | g_loss: 0.8853\n",
      "Epoch [   49/   50] | d_loss: 0.0383 | g_loss: 0.9014\n",
      "Epoch [   49/   50] | d_loss: 0.1107 | g_loss: 0.6690\n",
      "Epoch [   49/   50] | d_loss: 0.1725 | g_loss: 0.6185\n",
      "Epoch [   49/   50] | d_loss: 0.0774 | g_loss: 0.5763\n",
      "Epoch [   49/   50] | d_loss: 0.1331 | g_loss: 0.6863\n",
      "Epoch [   49/   50] | d_loss: 0.1146 | g_loss: 0.5413\n",
      "Epoch [   49/   50] | d_loss: 0.0634 | g_loss: 0.7257\n",
      "Epoch [   49/   50] | d_loss: 0.0830 | g_loss: 0.7760\n",
      "Epoch [   49/   50] | d_loss: 0.0985 | g_loss: 0.7116\n",
      "Epoch [   49/   50] | d_loss: 0.1378 | g_loss: 0.6103\n",
      "Epoch [   49/   50] | d_loss: 0.2318 | g_loss: 0.5345\n",
      "Epoch [   49/   50] | d_loss: 0.0740 | g_loss: 0.6172\n",
      "Epoch [   49/   50] | d_loss: 0.0692 | g_loss: 0.6381\n",
      "Epoch [   49/   50] | d_loss: 0.1399 | g_loss: 0.7767\n",
      "Epoch [   49/   50] | d_loss: 0.2494 | g_loss: 0.5893\n",
      "Epoch [   49/   50] | d_loss: 0.1652 | g_loss: 0.9079\n",
      "Epoch [   49/   50] | d_loss: 0.1111 | g_loss: 0.8732\n",
      "Epoch [   49/   50] | d_loss: 0.1575 | g_loss: 0.8673\n",
      "Epoch [   49/   50] | d_loss: 0.1047 | g_loss: 0.8617\n",
      "Epoch [   49/   50] | d_loss: 0.0567 | g_loss: 0.5741\n",
      "Epoch [   49/   50] | d_loss: 0.1322 | g_loss: 0.6106\n",
      "Epoch [   49/   50] | d_loss: 0.0785 | g_loss: 0.5913\n",
      "Epoch [   49/   50] | d_loss: 0.1385 | g_loss: 0.7222\n",
      "Epoch [   49/   50] | d_loss: 0.1045 | g_loss: 0.9524\n",
      "Epoch [   49/   50] | d_loss: 0.0720 | g_loss: 0.6692\n",
      "Epoch [   49/   50] | d_loss: 0.0934 | g_loss: 0.9957\n",
      "Epoch [   49/   50] | d_loss: 0.1322 | g_loss: 0.7895\n",
      "Epoch [   49/   50] | d_loss: 0.0861 | g_loss: 0.9688\n",
      "Epoch [   49/   50] | d_loss: 0.1260 | g_loss: 0.8470\n",
      "Epoch [   49/   50] | d_loss: 0.3282 | g_loss: 0.7270\n",
      "Epoch [   49/   50] | d_loss: 0.1281 | g_loss: 0.8400\n",
      "Epoch [   49/   50] | d_loss: 0.0467 | g_loss: 1.0525\n",
      "Epoch [   49/   50] | d_loss: 0.0982 | g_loss: 0.9895\n",
      "Epoch [   50/   50] | d_loss: 0.1401 | g_loss: 0.5018\n",
      "Epoch [   50/   50] | d_loss: 0.1118 | g_loss: 1.0712\n",
      "Epoch [   50/   50] | d_loss: 0.1079 | g_loss: 0.6255\n",
      "Epoch [   50/   50] | d_loss: 0.1726 | g_loss: 1.0594\n",
      "Epoch [   50/   50] | d_loss: 0.0843 | g_loss: 0.9527\n",
      "Epoch [   50/   50] | d_loss: 0.0914 | g_loss: 0.6765\n",
      "Epoch [   50/   50] | d_loss: 0.0805 | g_loss: 0.7942\n",
      "Epoch [   50/   50] | d_loss: 0.1350 | g_loss: 0.5723\n",
      "Epoch [   50/   50] | d_loss: 0.1249 | g_loss: 0.9475\n",
      "Epoch [   50/   50] | d_loss: 0.0637 | g_loss: 1.2183\n",
      "Epoch [   50/   50] | d_loss: 0.0696 | g_loss: 0.6833\n",
      "Epoch [   50/   50] | d_loss: 0.1645 | g_loss: 0.7164\n",
      "Epoch [   50/   50] | d_loss: 0.0956 | g_loss: 1.2856\n",
      "Epoch [   50/   50] | d_loss: 0.3283 | g_loss: 0.6904\n",
      "Epoch [   50/   50] | d_loss: 0.0434 | g_loss: 0.5518\n",
      "Epoch [   50/   50] | d_loss: 0.1009 | g_loss: 0.5225\n",
      "Epoch [   50/   50] | d_loss: 0.0679 | g_loss: 0.9034\n",
      "Epoch [   50/   50] | d_loss: 0.0786 | g_loss: 0.7456\n",
      "Epoch [   50/   50] | d_loss: 0.0816 | g_loss: 0.8636\n",
      "Epoch [   50/   50] | d_loss: 0.0465 | g_loss: 0.5964\n",
      "Epoch [   50/   50] | d_loss: 0.0831 | g_loss: 1.2296\n",
      "Epoch [   50/   50] | d_loss: 0.1212 | g_loss: 1.1688\n",
      "Epoch [   50/   50] | d_loss: 0.1139 | g_loss: 0.4769\n",
      "Epoch [   50/   50] | d_loss: 0.0648 | g_loss: 1.0010\n",
      "Epoch [   50/   50] | d_loss: 0.0773 | g_loss: 0.7422\n",
      "Epoch [   50/   50] | d_loss: 0.0955 | g_loss: 0.8399\n",
      "Epoch [   50/   50] | d_loss: 0.0634 | g_loss: 0.8882\n",
      "Epoch [   50/   50] | d_loss: 0.2817 | g_loss: 0.8395\n",
      "Epoch [   50/   50] | d_loss: 0.0727 | g_loss: 0.7106\n",
      "Epoch [   50/   50] | d_loss: 0.1269 | g_loss: 0.5298\n",
      "Epoch [   50/   50] | d_loss: 0.0749 | g_loss: 1.1243\n",
      "Epoch [   50/   50] | d_loss: 0.0996 | g_loss: 0.9504\n",
      "Epoch [   50/   50] | d_loss: 0.0902 | g_loss: 1.1887\n",
      "Epoch [   50/   50] | d_loss: 0.0955 | g_loss: 0.7242\n",
      "Epoch [   50/   50] | d_loss: 0.0650 | g_loss: 1.2950\n",
      "Epoch [   50/   50] | d_loss: 0.0756 | g_loss: 0.9113\n",
      "Epoch [   50/   50] | d_loss: 0.1131 | g_loss: 0.6853\n",
      "Epoch [   50/   50] | d_loss: 0.0399 | g_loss: 0.9434\n",
      "Epoch [   50/   50] | d_loss: 0.1289 | g_loss: 0.6608\n",
      "Epoch [   50/   50] | d_loss: 0.0475 | g_loss: 1.1754\n",
      "Epoch [   50/   50] | d_loss: 0.0961 | g_loss: 0.9177\n",
      "Epoch [   50/   50] | d_loss: 0.0834 | g_loss: 0.9897\n",
      "Epoch [   50/   50] | d_loss: 0.1907 | g_loss: 0.9076\n",
      "Epoch [   50/   50] | d_loss: 0.1246 | g_loss: 0.6444\n",
      "Epoch [   50/   50] | d_loss: 0.1147 | g_loss: 1.2864\n",
      "Epoch [   50/   50] | d_loss: 0.1422 | g_loss: 0.9766\n",
      "Epoch [   50/   50] | d_loss: 0.0652 | g_loss: 0.7566\n",
      "Epoch [   50/   50] | d_loss: 0.0537 | g_loss: 1.3228\n",
      "Epoch [   50/   50] | d_loss: 0.0387 | g_loss: 0.8687\n",
      "Epoch [   50/   50] | d_loss: 0.1364 | g_loss: 0.9976\n",
      "Epoch [   50/   50] | d_loss: 0.0828 | g_loss: 1.0534\n",
      "Epoch [   50/   50] | d_loss: 0.0500 | g_loss: 0.9451\n",
      "Epoch [   50/   50] | d_loss: 0.1013 | g_loss: 0.9106\n",
      "Epoch [   50/   50] | d_loss: 0.0525 | g_loss: 0.9230\n",
      "Epoch [   50/   50] | d_loss: 0.0618 | g_loss: 0.8427\n",
      "Epoch [   50/   50] | d_loss: 0.1989 | g_loss: 0.5289\n",
      "Epoch [   50/   50] | d_loss: 0.0608 | g_loss: 0.5143\n"
     ]
    }
   ],
   "source": [
    "# set number of epochs \n",
    "n_epochs = 50\n",
    "\n",
    "from workspace_utils import active_session\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "# call training function\n",
    "with active_session():\n",
    "           losses = train(D, G, n_epochs=n_epochs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Training loss\n",
    "\n",
    "Plot the training losses for the generator and discriminator, recorded after each epoch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fa8f5fc1a20>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJztnXecHNWV739nemZ68owmKY2kkYRyYCSNkISQRBLRhLUJNpjkgHm2wQbbC17v82N577O7trExGExYDBiMDTYYbDDYBqFAUBrlLI1Gk3OOPZ3u+6Oququ7q6qruqtDte5XH32mu+Kp0Oeee+455xJjDBwOh8NJLdISLQCHw+FwzIcrdw6Hw0lBuHLncDicFIQrdw6Hw0lBuHLncDicFIQrdw6Hw0lBuHLnWAYishHRMBFNN3NbDicVIR7nzokVRDQs+5oDYByAR/z+DcbYq/GXKnqI6P8BqGCM3ZloWTgcNdITLQAndWGM5UmfiagewNcYYx+qbU9E6Ywxdzxk43BSHe6W4SQMIvp/RPQ6Ef2BiIYAfJmI1hDRDiLqJ6I2InqCiDLE7dOJiBFRpfj9d+L694loiIi2E9FMo9uK668kopNENEBEvyKiT4nozgiuaRERbRXlP0REV8vWfY6Ijonnbyai+8Xl5UT0nrhPLxFtk+1TQURvEVEXEZ0hom/J1q0mor1ENEhEHUT0M6PyclIXrtw5ieZfAPweQCGA1wG4AXwHQCmAtQCuAPANjf1vAfC/ARQDaATwf41uS0TlAP4I4Afiec8AOM/ohRBRJoB3AfwNQBmA+wG8TkTniJu8COCrjLF8AEsBbBWX/wBAnbjPJFFGEJFNPN5uAFMBbATwAyK6RNzvVwB+xhgrAHAOgDeMysxJXbhy5ySaTxhj7zDGvIyxMcbYbsbYTsaYmzFWB+A5ABs09n+DMVbDGHMBeBVAVQTbfg7AfsbYX8R1jwHojuBa1gLIhKBwXaIL6n0AXxTXuwAsJKJ8xlgvY2yvbPkUANMZY07GmKT0VwMoYIz9p7i8FsBvgo43h4hKGGNDjLGdEcjMSVG4cuckmib5FyKaT0R/I6J2IhoE8AgEa1qNdtnnUQB5ahtqbDtFLgcTogyadcgezBQAjSwwSqEBgtUNCL2UawE0EtEWIlolLv9vcbtNRHSaiH4gLp8BYLroruknon4A/wrBugeAuwAsBHCCiHYR0VURyMxJUbhy5ySa4HCtZwEcBnCO6G74MQCKsQxtACqkL0RE8CtkI7QCmCbuLzEdQAsAiD2SawGUQ3C3vCYuH2SM3c8YqwRwPYAHiWgDhAbnFGOsSPY/nzF2jbjfCcbYF8Xj/RzAm0SUFYHcnBSEK3dOspEPYADACBEtgLa/3SzeBbCciK4honQIPv+yMPvYiChL9t8O4DMIYwbfI6IMIroYwFUA/khE2UR0CxEViK6fIYhhoeJ5Z4uNwoC43ANgOwAnEX1PPIeNiJYQ0Qpxv9uIqJQx5hX3YwC8Jt8bjkXhyp2TbHwPwB0QlN+zEAZZYwpjrAPAzQB+AaAHwGwA+yDE5avxZQBjsv8nGGPjAK4BcB0En/0TAG5hjJ0U97kDQIPobvoqgNvE5fMAfARgGMCnAB5njH0ihoVeBWFwt1485rMACsT9rgJwTIw0ehTAzYwxZ+R3gpNK8CQmDicIMUqlFcANjLGPEy0PhxMJ3HLncAAQ0RVEVCi6V/43BPfKrgSLxeFEDFfuHI7ABRBizbshxNZfL7pZOBxLwt0yHA6Hk4Jwy53D4XBSkIQVDistLWWVlZWJOj2Hw+FYkj179nQzxsKF6iZOuVdWVqKmpiZRp+dwOBxLQkQNerbjbhkOh8NJQbhy53A4nBSEK3cOh8NJQfhMTBzOWY7L5UJzczMcDkeiReHIyMrKQkVFBTIyMiLanyt3Ducsp7m5Gfn5+aisrERgQUtOomCMoaenB83NzZg5c2b4HRTgbhkO5yzH4XCgpKSEK/YkgohQUlISVW+KK3cOh8MVexIS7TOxnHLvHh7HZ6e7Mep0J1oUDofDSVosp9x7R5zYWdeLUacn0aJwOByTsNlsqKqqwqJFi3DuuefiF7/4BbxeYd6Rmpoa3HfffVGf45lnnsHLL79saJ/zzz8/4vO99NJLaG1tjXj/aOEDqhwOJ+FkZ2dj//79AIDOzk7ccsstGBgYwH/8x3+guroa1dXVUR3f7XbjnnvuMbzfZ599FvE5X3rpJSxevBhTpkzRvY/H44HNZov4nHIsZ7lzOJzUpry8HM899xyefPJJMMawZcsWfO5znwMAbN26FVVVVaiqqsKyZcswNDQEAPjpT3+KJUuW4Nxzz8VDDz0EALjwwgvxb//2b9iwYQMef/xxPPzww3j00Ud96+6//36sX78eCxYswO7du/H5z38ec+bMwb//+7/7ZMnLE+ZQ37JlCy688ELccMMNmD9/Pm699VZIFXUfeeQRrFy5EosXL8bdd98NxhjeeOMN1NTU4NZbb0VVVRXGxsawadMmLFu2DEuWLMFXvvIVjI8LFaUrKyvxyCOP4IILLsCf/vQn0+6j5Sx3PuzD4cSOLSc60TVkbhn7snw7LpxXbmifWbNmwev1orOzM2D5o48+iqeeegpr167F8PAwsrKy8P777+Ptt9/Gzp07kZOTg97eXt/2/f392Lp1KwDg4YcfDjhWZmYmtm3bhscffxzXXXcd9uzZg+LiYsyePRv3338/SkpKArbft28fjhw5gilTpmDt2rX49NNPccEFF+Db3/42fvzjHwMAbrvtNrz77ru44YYb8OSTT+LRRx9FdXU1HA4H7rzzTmzatAlz587F7bffjqeffhrf/e53AQgx7Z988omhexQObrlzOJykRGmuibVr1+KBBx7AE088gf7+fqSnp+PDDz/EXXfdhZycHABAcXGxb/ubb75Z9fjXXnstAGDJkiVYtGgRJk+eDLvdjlmzZqGpqSlk+/POOw8VFRVIS0tDVVUV6uvrAQCbN2/GqlWrsGTJEnz00Uc4cuRIyL4nTpzAzJkzMXfuXADAHXfcgW3btumSM1IsZ7lzOJzYYdTCjhV1dXWw2WwoLy/HsWPHfMsfeughXH311XjvvfewevVqfPjhh2CMqYYN5ubmqp7DbrcDANLS0nyfpe9ud2g0nnwbm80Gt9sNh8OBb37zm6ipqcG0adPw8MMPK8amh5sUSUvOSLGs5c4nkOJwUpOuri7cc889+Pa3vx2itE+fPo0lS5bgwQcfRHV1NY4fP47LLrsML7zwAkZHRwEgwC0TayRFXlpaiuHhYbzxxhu+dfn5+b4xgfnz56O+vh61tbUAgFdeeQUbNmyIqWzccudwOAlnbGwMVVVVcLlcSE9Px2233YYHHnggZLtf/vKX2Lx5M2w2GxYuXIgrr7wSdrsd+/fvR3V1NTIzM3HVVVfhP//zP+Mid1FREb7+9a9jyZIlqKysxMqVK33r7rzzTtxzzz3Izs7G9u3b8eKLL+LGG2+E2+3GypUrI4reMULC5lCtrq5mkUzWcapjCO8ebMOXV89AWb49/A4cDkeTY8eOYcGCBYkWg6OA0rMhoj2MsbCxoZZ1y3A4HA5HHcspd14Cg8PhcMITVrkT0TQi2kxEx4joCBF9R2EbIqIniKiWiA4S0fLYiMvhcDgcPegZUHUD+B5jbC8R5QPYQ0QfMMaOyra5EsAc8f8qAE+LfzkcDoeTAMJa7oyxNsbYXvHzEIBjAKYGbXYdgJeZwA4ARUQ02XRp5XKBx0JyOByOGoZ87kRUCWAZgJ1Bq6YCkKd0NSO0AeBwOBxOnNCt3IkoD8CbAL7LGBsMXq2wS4hpTUR3E1ENEdV0dXUZk5TD4aQsHR0duOWWWzBr1iysWLECa9aswVtvvZUQWbZs2RJVNchkQZdyJ6IMCIr9VcbYnxU2aQYwTfa9AkBIIWPG2HOMsWrGWHVZWVkk8nI4nBSDMYbrr78e69evR11dHfbs2YPXXnsNzc3NMTunUnkBiUiUu9bxEoWeaBkC8BsAxxhjv1DZ7K8AbhejZlYDGGCMtZkop1yi2ByWw+EkhI8++giZmZkBGZszZszAvffeC4/Hgx/84AdYuXIlli5dimeffRaAdgnePXv2YMOGDVixYgUuv/xytLUJqii4BPA777yDVatWYdmyZbj00kvR0dGB+vp6PPPMM3jsscdQVVWFjz/+GA0NDbjkkkuwdOlSXHLJJWhsbAQgZKA+8MADuOiii/Dggw/G+a6FR0+0zFoAtwE4RET7xWX/BmA6ADDGngHwHoCrANQCGAVwl/micjicmHPqQ2C4w9xj5k0E5lyquvrIkSNYvlw5evo3v/kNCgsLsXv3boyPj2Pt2rW47LLLACiX4F21ahXuvfde/OUvf0FZWRlef/11/OhHP8ILL7wAILAEcF9fH3bs2AEiwvPPP4+f/vSn+PnPf4577rkHeXl5+P73vw8AuOaaa3D77bfjjjvuwAsvvID77rsPb7/9NgDg5MmT+PDDD02bYMNMwip3xtgnCGMuM6HJ/JZZQnE4nLOXb33rW/jkk0+QmZmJGTNm4ODBg76CXAMDAzh16hQyMzN9JXgB+ErwFhUV4fDhw9i4cSMAYWajyZP9gXvy0rrNzc24+eab0dbWBqfTiZkzZyrKs337dvz5z4I3+rbbbsO//uu/+tbdeOONSanYASsXDuORkByO+WhY2LFi0aJFePPNN33fn3rqKXR3d6O6uhrTp0/Hr371K1x++eUB+2zZskWxBC9jDIsWLcL27dsVzyUvrXvvvffigQcewLXXXostW7aETOahhrxSZSxK9ZqF5coPcDic1OLiiy+Gw+HA008/7Vsmle+9/PLL8fTTT8PlcgEQ3CAjIyOqx5o3bx66urp8yt3lcilOngEIvYCpU4WI7d/+9re+5fJSvYAwSfZrr70GAHj11VdxwQUXRHKZcYcrdw6Hk1CICG+//Ta2bt2KmTNn4rzzzsMdd9yBn/zkJ/ja176GhQsXYvny5Vi8eDG+8Y1vaEamZGZm4o033sCDDz6Ic889F1VVVaqRLw8//DBuvPFGrFu3DqWlpb7l11xzDd566y3fgOoTTzyBF198EUuXLsUrr7yCxx9/3PR7EAssV/K3tnMY7xxoxa2rpqO8ICsGknE4Zxe85G/yclaV/OVVITkcDic8llPuHA6HwwkPV+4cDifsBM6c+BPtM7GscuevIodjDllZWejp6eEKPolgjKGnpwdZWZGPK1o3zp3D4ZhCRUUFmpubwYv5JRdZWVm+JK1I4MqdwznLycjIUM3O5FgXy7plOBxOktJ9CvjsScDrSbQkZzWWU+48EjIF8HoAtzPRUnBixal/AuNDwn9OwrCccuekAHtfBj7+eaKl4HBSGq7cOfFnqD3REnA4KY9llTuP2uJwOBx1LKvcORwOh6MOV+4cDoeTgnDlzuFwTIbHtCUDllPuxMtCcjhJDh8QSwYsp9w5HI4GQx3Aib/ziAOOdZU749aBuTAGfPoE0HYg0ZJwouHg60DrPsA5nEAheO86GbCscueYjNcDOEeAk/8IXTfSDXi98ZeJYxzJbZkUlnsyyHD2wpU7R5vhLmDX/wANnyZaksTStEtweXA4FoErd44244PC38GWxMqRaGo3ATUvJFoKHUgukWSwmrl7JpFw5c7hcGJEMjQwZy+WU+7k9cDmcQCM+4DjC7fCOBwfXSeAvoZES6GJ5ZR7Rl8tVra8jLSxnkSLwuEkH0k1oBolPaeB0d5ES6HM4T8D+3+faCk0sZxyl0iFd5eTggw0A6c+SLQUqcHBPwI7n020FJbFesqdZ6iai2MA6G9MtBSpw95XgOaaBArAfx8cAespdwluuZvDzueAfa8mWorUI+Fdy0Sfn5NorKvcOebgdSdaAg6HEwMsqNx5t5PDUcWsAdXx4STofSQp3acSLYEurKfcfbqdv3ixgTeepmBlxTjSA3z2K6B5d6IlSU4OvZFoCXRhPeXOiTEWVkoccxjrE/721Ue2Pw96SAosq9y5CjIbC97RoXYh7NDKVrLpJIFi5c8jKbCscrekMrIESaAc9LL/90LYoduRaEn8cKuVkyRYWLlz4gpXWgZJkPGRDM8pGWTgcOXOsTDJqESSxSVhihxJeH/NgDFguDPRUsScsMqdiF4gok4iOqyy/kIiGiCi/eL/H5svZsAJY3p4DoeT4nQcBnb/xjIhjZGix3J/CcAVYbb5mDFWJf5/JHqx1CGkUGEkK8Dvc2Qk7L4lUz33MDTXAFt/Fv/zSlb7qKz4IGNCpccUet/DKnfG2DYASVqajWMaVn6prSx7KhPuuZz6IHkypNv2C5Ue2/YnWhLTMMvnvoaIDhDR+0S0SG0jIrqbiGqIqKarqyvCUwmWCeM/6NjjHAWYJ9FSxI+RHj5XrNVprgH2vGR8v/Eh8e8w0HFUKDdsccxQ7nsBzGCMnQvgVwDeVtuQMfYcY6yaMVZdVlYW5Wm5co85nz4uWDNnAyM9wK7ngPptJh0wwdEyURk/cZbdTEPt1AfAYFt0xzj6F6HcsMWJWrkzxgYZY8Pi5/cAZBBRadSScTjxxClabgNn+VyxnJQhauVORJOIBHOBiM4Tj8mnSeJwOPpIJherPBqv83hyyWYQPaGQfwCwHcA8Imomoq8S0T1EdI+4yQ0ADhPRAQBPAPgii6VDXLz5ZOGbnpzw+5m0uBxA406DioY/z6g58hbQfTLRUkRMergNGGNfCrP+SQBPmiZRGHiYe6JIxhufjDKJmOpH/ifQcQTIKwOKZ+nfz+sFhtqAwqnmyRITkrghco4kWoKI4RmqHE44Et1LdI8Lf706IpfkA6r124C9L0c/wHg2oPaMT/4DOPin+MpiEhZU7lIoZILF4CQRFn8ZXGNA7xnzjysl6yS79ZnsP+ae2kRLEBEWVO4SSf5CcM5yDLyfh94ADrzmt9CjglQ+x4HuWqDtgP7tk8HHOj4EdJ2MXJbeM8LAq9tprlwmENbnzuFwYsyImNDHzEygYiqfY8gh0X2RXWRwxwQaavteFSYnmbEmsv0PvCb8zZ8EVN9lnlwmYDnLnSXzIJqVSVTXuK8+Ni4JSyHdexPebTOt4WSwrGONo9+c4wy1m3McE7GccvfD3TIpwf4/+K2fZMAshZZoP3K489dtiSxNXxdJ3ih0HvN/NvM5ffwL4OQ/zTtelFhYuScxjgHgzMeJ/4HHC+co0N+UaCnMQemZyZdt/i/AE6NiV2EbFj1KU6dibdgewygag+99PH8nQ+3+OjJm4x4HWvbE5tgRwJV7LDj8Z6D+k8CSoqmK1yvUoNn3O6FkaqwZHwJGVYqUer3A4Tdj30UeMLkhi4lyYxZwqyRAPrUqlCloiFlPuZtSGCnGSC9Qssk40AIMR1iNU01RyKtGjvWFP060YXmfPQnsfDZQJuk+j/YIkQ/H3on8+AlViBGcW6pDrhUDH+/3cMyoH1smn2MwtvV90mQxJH0NCqIk2W82Cqyn3DmRs/dlYPfzKivj9FJ/+kR8zpNwoigV0HVCyEjVc6ye00JPseFT4btR4+fjXwhutQBxEqjgdj4rvKexgmz+z711sTuPGl5P3MpKc+XOSW66T4W35HY9K9ThTno3hE4O/xk4+ld927rEnpBjMLJzucfNdzPpRakhUnOb9DdFfo0B51RTeSzob4zY+lPgSHzKaFtQuafIDzhenPi7tWtTH3ojvCXndgJdx/3fU6hrDUDfxBGaDVsK3I99vxPq7avRczp2A91mE6e5Wy2XxGQp4ywZhG3dl2gJ4kO0Cl1z/wiOHWCNeoT/6ZmmHV7xPACSeg5VSfnagtWOTlk9LuGv1wukyezSoXbBgJmyLPwxEvWb9LiA9oNxPaUFLXcLkWoWZDgSer1hzu0eF2q4yOk8JkTXRMv4EHDg9dDjSxx8Hfj459GfR8I1ZlKpgjiz7WfCf997EoGibTsAbP2JEG4sId33MVkUlfxdZMy8CKqG7ca27z0DtB8W8griHANvPeWeDNbw2Uh/Q1LWz/ARrmHZ/hTwyS8Dlx15W4iuAaJ7rxp3CoNz7YeU1ytFZciR/Mx6G4BPfglsV6myLb8Os34r3bX+QVfGAhVrJLhGw2+jRsdR4a+RMOPGHUDNi8BAs/o2eg2Tui36zwsICXrH3onumiPEespdhHnPMqs41shf7obtQsihHLcTOKYwyBeNtT4SbR6ATHkxLzRdEuEsXSPXEaw0faeNYz0XX0OrYgWb1YtyjQk1Yw6/IXxvPwhs/7W2ojRKLHp88mMOi1a7GQOyERN/o9Syyj3lGevTFzeul+AX2+tVD8mq26KcxTcSYYy8GvIBsk+fMG4VBRDHxt5q7rZo5JXi56V3UVLqI93RyRRr+uv9n8eHhb96ejJWe7YacOUeS6LpFu94RvgfKSPdQqq8RHB8+9afADW/ifz4ZuMcMe7PlMMY0LpX+KyWwaoXx6Bw78Ml45z6QIhG0svm/wqsa6JE1MolSQdUW/f7P4dcYwSytu7XTog78Lr/s65eRpLdLxOwsHK3wMNIpBUQ/EIruSWS3foyBPNPTgEAx96NfNCx/aBgqUq1yYOfo9RoN9eI0UgGGnEp2chsRrqEZ66kyDzjiQ8TPPF+6LJoPBVdJ4Q5To0iZTfHm7EoDY4IsJxyJ/GHlXfcmlNfxY2EDDxrNGZabiAlRnqMVS2s/xQYkSn39kNAy97Q7VxjgvtHjyx676HbIX4IitCINW6n/zyDrcDeV4I2EOU//l5y9dIAYVDWIyvTcXpzoLGh5/4FZ9YG0/BZ5PKZTQKmOrRcnLulSLXInmgU1tafCBM866V+m/YPIjhyh3kBl0N5Wzm1mwTFnzdRfRuj19kWp/jl4PdJK7om+Bq0XFVx72GywDLPO34t+PYDSi7owOPUjtyp2wrMOF+2QOP3yEI+WB7LWe48QzVWxOGlNlK0rPO49vr+MOGFgJCaH9wIGC3q1nlMcGuYhXTehu3K5YONKjgzGO3RcNtEOZGIx6XcS5L7y6VBW6MzUY0PCZE7Z0uinkEsqNwFWLK1sAPNsasTbYTxYdFXHOGPMdxuIT9U2XOo26pdnTDeNO0Gal4IWmjgvWneLcTCH39P3/byBuP439STmgCgeZfw1x3U24imomWknNkGnNB5jUbZ9ihw4m+hy81MwtIqNR0uxyCFsaxyTzr2vuJXJIkcSD38pqCMjDY0emR2DAjuFS03RLSRKkbQ4/aKJpxUsvrHI0ja6a0TxgFUiXFUC/OENhxaBBQPE2UyqwZK++HA73reNbN+Q/t/7/+s9b54wiToffakMIBuhASHVVpPuZs6ibAKfQ1Cl9moggw3wBMPpO6u0fv02a/0b9ul4TKJ1zhDUsYjG0mEEn96sbqOI2/FyGqN4fONttZ/NEiRUWqMDwmhrxbCeso9HkhTZUWahZcMA6nJIEMyIinTWFRRlJJl9OArdxsjYyVWpSLCWbhyYlFKuK/e/GPKiSbXIhhuuRuDrKC09D5UxhL+AkREcFEmNUZ6gKGO2MuTLBiaP1OHWyau4xcqv6vgRC55EpbLoZ2YpTZWcWarDnnkVTV1NoD9jWE2sIDuMBHLKfeY07QrPnOBAkDdZmDLf8dtZpa4s+s5YRzCiB++64RQ6CkYK9ekV0LPjEmxSngyEpGz73f+z1KRNQlpPKNxuzEjJdgHH46GT/Rtl3SGErfck4vaTbIvMW7pJUtPbfaZqEgiK8WIcj/8ZyGhJZiQCSvMStOHoBSk6I3hDn/dcPl6s9Hjcx8fCpSl9kNzrHkl/7JjANj6M/XjO0dCyyNLDdRQh5BEFSsSkACUClg3iSnZGumIkHXNmcZs9aO9Qtp09V1A/iTlbdoPA6f+CdhUJoRIJLFwpR2OdKoyhRfnk8f8yt3lUN7GdHRGy8hDFEd7hWnaokWtQfG6jUXYyBvI0x9FJVJcMKN2v4XglnsikZTeJ78EdmoUCZNC0jo0urO1HwgKSkq40VKoWm4nwzVILNrKymfFCY65lteoMZvgAV0ty50x82XxerV908MGxkjk75iR+up6SDoXSwTwAVULc+xdYY7PiJFFTGhVIPQpAh3HkqIktBJoIrZ6ZTAmZAYaiZ5IJrTmJTXrmpp369hI46G2HzI/MS5cdE6wX10Nrycw1FJ14ukI2fYzc493FmJdt0wyIM28E+yjVcPtFAo4zb8aKJpunrvC4w790TbtMufYijAhJO3E34GyueE3T6YIp3gm0KjhHtdfkdPMTE7nKJBu196mdR9QPDP8sU4FTRmXTM84aeCWuyGS8h2S/1DVploDhC7vWL9sUgqdFxMwyKvAtp/Ff05NqUELTtySphMLKOKV4IcmryUep5nnNZEPMGv1sMymty52xzbbck8F5O9dAuBPxGy0/JnBflbDLVUE08fFm6PiVHzynoS8AmAi6DxqbPtwmZLR3nN5oat41vlmXn2zXRkaVBWJ5RR2sWyUYknthwk9vbWVu5mzie94OjTpIlbdBOeIaLFFefy+euDjX5ghUeQEh3HG0xKNFeH83NGMWYwPhk91jxU9tfrcdXoLpXHUSYIigpZV7gwInxHo9egfIBrrN+lHF6SwXWOymtPiOseAUIgouPHY+7L2oYN9wWZOUmwEuRxD7YmRwQjJVBkw+BnGM6IizRa/c53tDLQkWgLrKnddnNkmxLb2nkmcDDueFmpOB+N1I6QhGGgRQtVSIQwsKQdHIkTpeZzZpm9f+Ty2iseOZ3ZyCj2TZCcJ3v+wyp2IXiCiTiJSDLImgSeIqJaIDhLRcvPFjBCHGF6o5irweoWpuPRGu0SC3D8b/MDdCnJt/UnifdTR4HUnx6ClmSjFcGuW8zVAPCsNmh2LzlEnCQaY9UjwEoArNNZfCWCO+P9uAE9HL5ZZhMkC7DgkTDBR/3HcJApALWFItfKdRSz6qGL/Neiujc1xzxas4EJLGSxguTPGtgHQGtK/DsDLTGAHgCIimmyWgKGYeNOkwUBVy108l+gqaRsYQ+dgBJEEZjLYpp3wFBcS1MgMhKv6x+EkCV0qlTLjiBlJTFMByAs3N4vLQqr9ENHdEKx7TJ8+3YRTh8Esv9fWnwClc/Ba+1IAwP0bNRJ3As4ZfH4T5NnzkvB36U3RH8tdH3PtAAAgAElEQVRqcMuTYxWMVr6MAWY4hpQ0lqJpxxh7jjFWzRirLisrM+HUJqFnADMiP7LJFq688l4iS+CO9QsurXgTy8qDHE48iUPQhBmWezOAabLvFQCS61eoeiONW9I277hQRiA9guqL0fYkgi3XWM9Ko4ZjQBbeqUJwejpHxCLjJpzYolUF1iTMsNz/CuB2MWpmNYABxljKFmBe2fxbYMdTOreO8aBKfwymMTOLDoNZoWcLqRDmyomeOITAhrXciegPAC4EUEpEzQD+D4AMAGCMPQPgPQBXAagFMArgrlgJGzsM/uBcwYOqevdP/Ag6J9Fw5Z7y6Jm/NhmUO2PsS2HWMwDfMk0io2h2b8IoUz3dokRPgReriY45iYFb7mcBeiqPxl6vJD7S3iCRuami+EHJJ3XQhYaAkQjfddz4PhwOJ7kZj2GhNRHLKXdVBppDk4L0KlM1a4pI38QNPESPw+FI6KllpDVZjElYVrkHqOOxPmDvK8Cpf6hsHIXlrqfYUrdKcbIQq5v73KODuzQ4FkDPXK0FU2IuhmWVuw/G/AOcRuZ/BOBTtlrVIElPJb1gpS0qIb3FpTQPzRsEDodjHOsr92jQM6ihpwAQV8DxwfDk3RxOksIHVI1iMN3fDMsa0F8BjjcCHA4H4MpdP+F8sSrrU2HWIA6Hw1HAcso91PY1OMjWcxporolekBFeG5vD4URIMiQxWRolN4hUcKuiWs8B1F0pu57zfx7u1D5nNBx719zjcTicxMPdMhooGexqijXWWYHBRbRco4mRg8PhWIM46ALrKvdkJqT2DIfD4cjhyj08MW8BeYQLh8OxHtZX7oCp9dpjC3fLcDgccLeMbqQMU9UKihHeyCAffslI9BM0uzxe7KrvxZBDbd5WDoeT+nDlHoqSMS4p93FxYJMxIUFJbWAzQub0fKRjK+3ewpDDDY+XoaWfx9hzOGctFplmL3mQblhvHVD/aej6RNdmB+LqKeofc6IwKwPEM2M5nCSDW+4asKC/8MeOBreK0vdtP425VJrEMRSyb9SJY21DaB3gkTscTtLBfe4GYQoKX3E99NeV6ThsTAaj28cIl0do6BwuT4Il4XA4oXDlrg+9hbvkKLltlOg9Y+y47nH1dV530sXvcDicBMAtd52kyS5j3+8UblyShCCeeC/REnA4nKSAD6iGhzEEjFL2NwFkw5jLg/1N/Zg/KR8TEiacjM3/lWgJOBzOWYTlLHdScmwEu2X66n1x5D3DOuZA5XA4nHjC3TIKKE17F4nPPcEkiaOIw+EkBO6WCcGbNxmA/NYwhL1RLXsBe0EMpeJwOBwDxMG6s57JG5yQc/qj8FUYR7qBQ28YP9f4oPF9wsCjZTgcDg+F1EPr/tgd+8TfY3ds7pfhcM5euM9dB3nliZaAw+FwDMKVe3h43RQOh2MV5mwU/nLLPZRQXW4x5S6Ky70ynJRh6U2JliA2VH8ldFm01zppqfiBK3fDtA86sL+pP9FicDhJTd+oE7vqe+HxmqBksoqM75Ps4cs5xUBWYejyzNzojitZp9xy10GQKX+mewRjvFgWJ1mYfVGiJRBY//2Ar429o/B4GRxuld9KrN2d01fH9vjRsOIOYOXXlNdFrZTj52mwvnK3GIoZtpzUJbcs0RII2DICvoZ9C820rCU/s1UomAKk2ZQbOLMsd+6W0YGHT1fH4SQ1cZzHIGpW3Km9PqsAmLo8ihNwt4x+Rrq115+lhrIr2wSLsWha9MfgGGPNt+J7PjUdYyWFbCYFk8NvE631DoBb7gpEoqsZGLwmvqy9I2dJMbKKlYmWIHHY8xNz3qwElskIct1oMnOdgQMnqKGYtDjKA0RhGSpF2gB8QNVs2vod2Hmm1zc7UbR0DEY/dR2Lt2VUOie+51Ni9T2JloCjRZq8KF+491OH4sstjUYaRRhjcOudC3lKNO6TGMF97ubSNSzMjuQUlXvfqDOqELCzp8Nqsk8rO0xl/QkzzD2fLZq6eAl4ymVz43cupXwLr4Eos5wS/+eMLO1tlQwZtWicMA1Cfc8odtf3affEtSJ9MnM0jx8XiJLHcieiK4joBBHVEtFDCuvvJKIuItov/leJIzIPL9M3P6jb48WoU9iOIOxzvH0Ip7uGYyxhkpGmUCo5LHFWcNNWAcWzzDueUnlovcS7Z3XRD4HFXwhZ7PJ4sb+pPz7hvUHXzBjz9zDzJynvUzBF3QctRQoZaWQlA0AlWkcy1DSVu1zWYGWuZc0H924jCQetXAtMWaYcJeVrEAlJYbkTkQ3AUwCuBLAQwJeIaKHCpq8zxqrE/8+bLGcIjT0j2NfUj8beUXQPh85b2jEoLOsb9UfTuDwM+8QEp7MuFl4l0cTh8mBcinWetSGOAimQr2Mwywgzo7geZo4LDwBwwXeNbS8b1OsdcWLM5UFb/xiQlg4sul55HzOeXek5AV9rGvqwp7Ev9Pjn3Q2fclIbm8jMAeZdBVR9yViSk5Q4pOL/16Vu8yYKfzOygbL5QQfQsurz9Bxdm5wSYN4VgVN/Skh++CSy3M8DUMsYq2OMOQG8BuC62IoVHqdHuDkt/WM41TmMvY19ONPtt8aHx90h+0izMynBxH96iLu/PAKchTP9X9b/QPWl3tfUj72NUkZvAkOLFl4rWlkm3tuofL7qcjjcHt3vCgBByRhB6nHkBVl/G34AlC9Q3mfG+eGvd/ntocvkPbr51wSMi7i9DC6PwnXmloQuCyZ7ApCeCUyoDL+tHFum8DeaJKpzNgLLbxOyTI0QzTn1uHukHszEReo9IRPRo9ynAmiSfW8WlwXzBSI6SERvEJFiDB0R3U1ENURU09XVFYG46oy7vWgfDLXgg86vum5HXS/qukZMlUnCKxsEileds+HJ5/u/2NIFqy/VqPqS8nIp4iQ7grR4CZUGfMjhwr7GfnQNab9rUeF7SQy+LJUXaK8vnOpTtorJdLZ0IFO0xKM1YOT7GzqWXgNLY6UtHSisMHBOkWh6jkrjSVOXK9eimX+1eiNtInqUu9IbFnxr3wFQyRhbCuBDAL9VOhBj7DnGWDVjrLqsLLI4bC0FHQ4l942czhj9YI+1DWJ3fZ/iurouoddhOiE+yyQP+Pf9WmVyhvsB5KhYkKu/KaTbK9UGCUat0VNJVJHceUOO0J6hLsINKgPw3wODCrZ8geC7lzEw5sKIvBe75IaA9b7bLo11GP19JXtV1omL9G87aYkJJ5Tdj7mXAyWzTThmZOhR7s0A5JZ4BYBW+QaMsR7GmKQZ/wfACnPEMxdpYFX6bKhrHQWDGoqgY2gc424T/bvRoPBD7Rt1+qKMDDN5KbD6fxncyYRnQmQsZluJiYs1QzcjljI9M9I9I+Jo2yAOtgz4F9gygIt+iDG76MKRHnnlWmMHDmuNM5XPGgS7oRQw3JYUVoQ0eOYdPGBn9VUJUvB6lPtuAHOIaCYRZQL4IoC/yjcgInl/5loAx8wTMXYcbR2Ew+XBrjM9iRYlaTnePoTDcuVghGmrDbtGNh3v8I+dqP3Ypq0EzrkEhnsj01cZ2z57AjD3spDFTYXVunYfc3nQ2h+UE6GnZoteJWNUIctPEcGaACSfcZnYu5qyzKAAQZFMiz8vFOuK+XhWAnoaS24ENjwY99OGfdMYY24A3wbwDwhK+4+MsSNE9AgRXStudh8RHSGiAwDuA3BnrAQGAC+Z4z8edLjRNTwOpTGjUacbLf1jivuZ8fopHsPowFuMkQaOQ3oWJecobK2AjutxuDwByWUuN5ONnaj8EM+5FJh2ni4RGBg8jAH2PGDSubr2AeC3/KeGdkJHM/QN1O1v6sffj7QLX6TBTiMRPOEUXYY4iKeR1TqaUYLTxVrnjDBjMqdYsIjL5/u/A35XWTif+7TzgOmr0NQ3iqa+UaBsnrJcRqj6EnDuF43vF2uIlKNnYowuLckYew/Ae0HLfiz7/EMAOvs+0cMozbSgiuBkpkGHC7Y0wtHWQbi9DJMKs2ALsqTcMmXkZQxNvaOYOiEb6WEeYFj3RmYu4FJuUOKDjh+ULUNINuqpVd+msEIYSEq3hz3cvqZ+EIDVi2NjsbUPOFDfM4pl5+QgTKqNn3NvVvXXs7QM9NkrAcdWY4JIUSnpMilmXQjUbVE4ifie6LHyq7+iWSphyF6OrrxgxalBxK4JcT9fgx7medoygNkXo3mLoFaMVjFSPLqeyJxkHyMwkbMiQ1WLtoHAbvOR1kEcbB6AWyODdczlRd+oUF+mY9CB1gFHaPdbgZ5i7a5ra/8YDjaLYYmJsECCBjAVjblZFyFsI0CkrtgVyiCo3ulwP0QdP9SeESH81cFUSrgGL7KlhyZSyeLOPUox27Z0/b0ZOXYxrjrYD6+g3PvLFIaxyhcA+RPDhOGp3SPhrjOzlJ10jxQVrPnRMpHCGEPnkENfCQPJdaRn/CYJGw2LKvf4xpn3jTpDkp6Otw+hf8zpiwPWU5iMyS2xoM17R5043j6IEWnQt3gmghlzeTAwphKrn1Wo6B/WZNaFgd+D3CjqVxTF/V/0L/oHmKafH36bMLROvBAA4MkRJlIfdLiM5ylUfVkWdqnwk1n9TWDpjeHT8AFlJVAS1OBJyl0Wgz6WExR9nD9JV3XCvuzpissH8oXGiGUEHUOSL39i0B5hlFdhBbDuAX/jrfMeH5p4PRqL9LnYFKVYcqPufQGgZ8SJ010jONOtI+w5LU3wlc++WPgu3W8L5LkAFlTu8W4f3R4vjrcPKU7d53KzAL88YwxDDhc8jGFoXFDCo05/pAzTkP5E+xBOlGor5/1N/TjaNqi8cs03gaJKzf1DYB7smfplAILv+829zQFuKsYY6iesCdwn6BIYGLzFBqIB0mxAZi76R53ozg5twHwsvUlfskwYvGl+q6u1fxRHWgdVx1IAAJUK1Q5t6YBN7IkoFX6SfvQZGha0EYUQZLkPZFVgLL/Sv/78bwNVt4Y/zoUPoV9FuXdPWI5dFXeBKcm8/DbgXJUcAi3S7VAM45SuXaGc7oi9HK0FVf4FkrUcxiXla6BLjfWYpPfbqTdCTe5qTYbiewawnHKPN/7szVDkoZQMQHPfGA63DmJ/Yx8Otwxi3O3Bjg79zZEjI4qkG8DnCvGwwBLHB5r6cbJzSJBTXi9kQiVcNuHH3dg7iqa+MfSP+csZMwDt+dqxvy19Y/j7kQ54vMI5T3YMBcZVK1E6D8fah/BGTyW6co1ZekZo7vOXpmAMGBZ7RQG9MLkSu+iHGtO/CfJpPk3RldY36oSzUCqCprWHyjrJxSOmw/dnBSXk2PP1hVRquQqI4E3LCHqDRQorIh/c1zqnnjLK09cIUU3BeQZBJYa9RZVCIwcA598LrL3PmJxmoFToLYncM9ZU7knYK3J7vGgWLUKpNEKo397/4F3p6laelmV5tPxz6kLY84AFn8OLI2vx0fFO3+KPjndiv9hI1TT04ZMWt6DIFLL4Buf5k1zk0p8SGweAwBjDwJgLDAwdg+NgILi9Xgw6XOgZceJY+xC0GC+aiR3Tvo6xTIPp4SGI9zPdLpRYCOJPNc3wynsi6QoKS28mo9T4iD9e31FX3OHfJqsQ426hMJ1v7ERNZsDvdgn26VauE5RVpLVObOlAhb5wTedk0SVi9nSA8sZaur5gF5AS6ZmCGyQ4VLLyAjFeXbz/OSX+xsKep38CjWDla9dxj0vnCO9YcORUbllyVJlUwZrKPUmQv7/dw+En8HCU+a1gYurWbWPvqOq6wawpoUkZspfuqHcGhrKU06iPtwsRQFLXdFCh1o6rYIZgEU0+Fyj2d3nl19fcN4ajbYPoGByPyFD5rLbHeIp91S2hy6Rj2DJVKw/25lSiM3ceHDMuBkvPwo5psoKlEVXKFM45lis2CgVTAtZK78ToeHBhOgWLpHSeYJHOCppEmyi62X7W/0D3vKWe4nOE98m0MFyF51k8S8jWnH2xUPlygYaBEu7ovlLFEVrIwWHU598bfh97vjCekFce2TkTxFmn3JsKzZtdqE7HoAzJWgBvRp7vx9/jygrwxytiSw8/qj/vioCB1H9IcdVBDNknBVTIBIBPTgVPUSj+YOz5wPyrwJRCO4l8bpe2gTEhBl62TA/yuPaGolXozJ0HlCsVGpUxYQacHm/gwHVGtlCtUEnxizBKR13JBrDMXKGeCqVhNGeq4NM9/170jji1ffD+Iwl/xFtUX3GtUOIg+Noq1+NI+TX63ExpaYJFqmMg1uzOasy8B4oRSSS4WdIzBVeGkTT/wqnKjUGEN8Q5sQrNhVEm0Esx/dnF/uecKyr+RM3gpcBZp9z7svVF1A5nmtBKs9Cv9T1CgzCeno8DzdqZn32jbuyu71OOkJFF03i8LOzUf4NZU3BwUmC9cKXKmQHyBsvPGACCM1foGThcgpJmIJzoGEafT4agH/icjQHZofL6QG5bNupKNvgsb59FRgS3x4uW/jGfQt/T0IdjsgHlz2q78Zl7ruHqfy3TrgHWfQ/IyMafaprQ2DtqYPIWQb7TXaOKIXKeitWqPSdh9+TxyRpm/tViZrAGUg9ggsZguQKdQ8GhxGKoZuE0fDYyOaSiK4v0NqbZolfu5QuFCpvyujUzzheWRVKwLEZYUrlTXJzuTLAoDdCfVYFDkz7v+/7WWFXINvL6NuGQ3CZK7hPYpaxEwubjnfjtZ/VhrWdPWqAy0ip+dbpr2GfRSj8GqbCaO38KdlXc5duWia/RmMtvkbf2j6GxR3QvVVT7w8nCcLp4PdryFwNFldjbKNTrl09rKK/Ts/NML3bW9fq+u71eHGzuR9+IE6/saAg4LmN+vepNS/c1JtK4SNg6Q7nlgD0fo1O1U/6laoumW9omH09CSU6Hy4P+UZmxYM8TKiZOXho+MzgzV4jc0vm8JfpHlUN8+0ed2FnXi/cPBfZIIx1/N6VtJRJ6FPKDScuSCMvVgSUCnLYcZLlVQgJNpK5kAzK8Y5gw1qhr+9aCczGS6a+p3ZU3D1MH9/m+K72Qg/bJKBhvC1G8Alpvov9gh8TaL7rDuxToGXGCFfvP99f9/tpwzYUrYHcPoZI1AySoL3mIIQsKW+sZceLj3UKV6G9eNBv2dP2+bVd6LhomnA+kpcHp9sIGYdYtQGic2vKXKu7X2DOKFrHy5u76XnSbXeEzPROvpF2D7gZ9x3Vm5CMTsgY1Em3kexzxKnEn8PruJvSOOOGLG9Ljl5ajpxpnEOF0rivEPam8R23nEOq6RnDZIgP10td+R/+2IShUM4XQQNZ2DmNqUTZ6Rpw4p9yEiUAMYjnlDggRIwXjrTinZ4spx2sqXIEh+2Qs7HzXt4xUHppRBrKmImtYaIjkXf/RjAk4VXIxGKXB7h6COy00o/Ng5jKUYovygXOERuSdkyMAhBH78CWLJasyDZtPdGJQLSEqDMG1ZhqKVqFs5KRictCvN5/G/Rv1zQ0q73o/s/U0JhVkQZ5Ks1vWWwhmd30vpqiuFUo/bDnRqbFFeIw0GE0la2FLm4xh+0Q8tbkWy/NXYE3uHsFPC4SdWGNg1IWP9rVgGXliYrYr1nMXCefiM4rHy5BG2uW6VVeprFDLrH3nQBsA+JT7kMOFI62DWAWmfs0xiHj58FgHTnX4Jw/S+xswE0sqd2d6HrrT50ak3JUeb0u0PjjfsUOVW/2E8zGcWQ53mh19R9p91lBTYbUvUUMpvt3t8aI2Yy5KsQWDYy54ivzH7h4eR+m089BjK0PtUb+bIlz1xnFbHloKlqEzdy7GNeL3lfE3dttPB1bRDG6YIo1keP7jM77PY04PznSPIDhPUovmwhUocLQonn1vYx9GxAiWeCQYsrQM9OYIZQycbi929OVhzcavCys1ZsaSONU5BMbEOUPV64LppmtoHDmZNuTaA3/yerJ167tHMKkwC1kZxqOLPF6GJzadwrLpRbhwntY4VnB2XKBcIeNOOh/i+4fb0dI3hjmzr0JJ2WRAx7h/55ADZXn2qOaOCI2Wij+W9LlHg1NM2ol6UEWDE6WX4dBEYa5LRjZ05c1DX05l4EZhMvDePyz4GMfT81BPFdicezUOTBZSrV/Z3gAvA14+anDCCCI0Fa3EeIZat1nHy6z4whN2V9yJmryLFNZFT2PvKPY09IbdrrlwBY5OvFZxXdymRtSjD2zpYcMw/SF/ynQNjeOxD06iczBwILKlfwy7zoTeq9/taMCLn54JWa5FQ9EqNBStxlv7WvDXA61o7R/DAYVMbS2kwfBwhkfIayUOTHrzhQCIcXE8Jzj/9fmP6/DOgVao4RZzTlwl8xVKKoRS1zWMV3c04kirTrcvk8ZsEDhOkQTj5pa03CVctmxMzPL4upFleXbf7OhKjGYUw2XLwe6KO+ChTFQM7NF9Li+lI00jNl1OiCKPgNpOoUu3b4oY5jeIgFjkXfXhlR0g+P7MQOqVuFUqW3rSMtGbPQMdeQvRUhA6kBwt2076wzY/ONqBwmxjk3HIC7vJFaZPiQZpUbfHi5c+q8fF88sxq8y4vzT6xiRQO4y7PXC6vchMF4yC34kDxkfbBlFe4A+l/KM41nHezNAIIsX5UDVoK/CXSO4bceJ18djnTjOeSR3udoTowgkzgHUPwDPGAISOeUm9wyGHG0OO4ZD1Eh2D2gX9PjvdjdOdw7htTSUA+MKF5bO2McbCWvH7mgawtW4It6yajokFoaGto043stJtSEuLn9a3vOUuv1UF2enIylC/pGG70C30pNkND5vXlsTGKo2UYNeIGkaic/RQ36sQEy7dS0rDmeIL4Ew3rgxf361v0BoQrMBPa4Nj9M1lZNyDIYcbm08Ym+vXrMbUh6gVT3UM4/lP6sw9tnj4/U39YQfjI22qJKUebn8l5dkw4EaNyvSULF13AWfx/IEStJavB6rvws663oAEveAeU/+oE7/88BROqGRdDzrcqO0aRtuA8LuQ3EfBV/Ps1jp8HON3NhjLK3c5melpyJAl3qycGVR4yoA1JSUf9Yh+05GMYkWLtD+rwpcYNZYeZW0YADVTb8PuqQqz1OsguDLleHq+MQuSSLUxaM8Tkoxq+mMz6q+nZHIsCblLPovemFp7Y0+zKfIouWXGXeoKeHjcjT/VNKmuDz2B8KehZxSbj3di20lzJ6yXeGqzUPc/XB5BsDJ0ebz4896WEKXKxDlvXRNDo6bkpSa8KueTGpHeosX+2aQU2N/Yj1Gn22fBn+hQVu5HWgfQNTSuK1LtdKd6DyMWWM4tIx/xnlJSBIz53RNF2ZnImZSO+u4RzCzNRbrNhsVTC+DxMnQOjqMxaOCyqXAl8pzaERTduXPQkzMLjGxoKqxGX/YMLO74S4BELQVVaC1YChZcDyMC3LbI08Cb+wKt6kMT/wWZHmMv1ObjnajrCt1n2D4Juyq+gbYuf1auEH8QXvl5vYLdZItjl1SLgTEX9jT0om3A4VOYwUrc6FjawKgLHUMO349cyQVytHUQ/zjSjnsvPgfptlC76lDzAHpHndgwt0zTZdvcF1qe4n+26bPqf/tZPe44v9L3XQoxHA3T4wjXxvUMj+Pl7Q24eeU0TCky/g4POlzweBn+frgduXab6vl8UTJBWUzD4+6Ae/D4plP43FJ/MlkkXrLf72zEhfO0a+54vEK4rrx5crg8SVEV2HLKXc45l34VDX97FABQMUF4oTJtaZg7UUwBpjTk2wXf7Oi4B625ga19S6HW5BmyglOyMqTD9okYT8+H3S1ryYnAoE+xNxVWo2zklK5to8Vty4LbZqz7CgjWnBLBEQuHJn0ehY6WsMd7fJNwvfdvnIsxpwe9I/pDCvuzpsHGjIVshisL0T00jm1BYY17G/uxBsCuM7040z2MRVOEQWe9P9Lf7WwIsN7G3aHKUioN4XB7kaeg3D881gEAgnLXaF3+VOPvHWhNvq6END7VIhoCaVIhNMZUrd1gxt0e/HrzaVyyoBxLKwSDSaqHtLexT1W5y8cMgtl0rBPH24Z8iXOzw8SFB0uq1LjVKljKWj2x9gEHtsrccEKSn//+KNGavxTTBvagod/pK3b29JbTmrLHC0srd2QXwZlVgj5nOmaQQreS0oQUcY+oHMKYY4cmfR75422o7Nuuvd3E6zF56HBAgpJeWgqXo6VwefgN48iQfSLyxzvCbtcTVBxtNLMEo5nGaq7/5pM6QwN7x8uvNHR8QAijNML+yTch0zOCNYDPly+5ibTShx774CQuXTARB5pDfdZajYKSovjouP/+e73CcCEDhRyntjPQPaDW1a/rGlYdCJbXNPJFnzChES7NUy4lLL8Pv94sKK9tJ7t8yl3iVMew7/5nZwYaPNtOduHSheoRK/IaP+E6TkYtYymaRWoYW/sd+NWmQCNLXklVLy2FKwJCqY2+e7HE8j73trlfxomyy5XfBiJgjv7ZiUbtpRi0S6kw6m+P25aNIbu1KsRpcbzsypC6M7HgsQ9OGo7YiAeOjCIMZimnjodTIjvP9KDLYDYsA/DxqS60ypTZgSZ/qODQuFvVDpGSdMLxl/2tqnI9u9Vv5fp9+8KF6qluKiE39OU9jWe2nsYzW0OtV3kd/QNN/ZrJUmrX7y9xwQyNh9R3j+D3uwIH7YNLcitF1hh1z0XSQMQKyyt3u9jN801iveh6/8oF1wr1MHRiJORNKizWFmYyCyvgScs0bIEnE499cDJmx2YA/n64HX8/rE+p6sHp9qKmvg+v725SHohj5tSoCRcGCPjPU9+tXmYaUG7k5MuM6EDGGD4S6yGp0dIXGpXVN+LEaIYQ4ukF4Zcfars35Yp515lezQFp1WOIf5PBh24Ua7tlAFRXFsOeYUNZ+g1CdcDCqcCRt4WVsrk61Z6NPSPN99CvWjwJzzfrC8lz27KxY/rd0YjOMRE9c9hGwpjTE1CJMhit4mtqvLLdX9Ts1Z0NuGttYAXFMz0j6JHFWUc6DP3B0fCuNr2WqVIjZPSeMwDvHWpTDSuUoxS1Neby4Hjp5chzdaOtIfwxzEAqtaHknk8vWo0AAA8eSURBVNNfSTSQkx1DmF6cE1HGrxEsb7nb0ghV04qQNnmJvyrb0puAZYFzTJbl2zEhx5/4smhKAe7fOBd3nT8z4FheMdRqPF3I+f78cuXuelUEiRyc2BEcKRQJpseo66B/1BVST37z8U4cDFMO2iz2NijHketFyspUaiQaFQbm9Sh29XO54LFlYSCrQnXQX45Rd5kSfxczxeu7R/HYByd9QQX13SN4YpOxwIiBMRfePdiKvx1sw3uHzOsJqmE95a7H0iiZDRQFTgycaUvDnTILaaM4sJOdacOVSyahNF+oJTGeXoATpZf5kpbUig3FutXlxJ9ERTnsOqMvIS0WhE4FaYwXP60HoPw7eXNvYMx/cKkEo4QrYRCMkfGD9gF9stWLkVjSvAxGkYqJRVq0zwjWU+4mIR8Amj+pALetngEAqK6cgL6cSpA42XSxSvTA/En+GVemFSfvPIqc5CecvzscSmGX8cTrZbpcNJG4sOTomzErMv6wS587VhowTYuiqFi8sLzP3SileZmaLfq6OWVYNycwceGqJZNDulETcjPxheUVPutkw7wybD3RhfICOzoHTa4lzjkr6c2ZiSlDB9GfpT27z37DFT7N5fFNpyyh7MwkWRLytLCsco/0ZbphxTTDls7sslwsmlIAIsLhlgEsmhJag3X59AlYPn0CAPXojVWzin0zBy2fMSFqfycntRm2T9Q1aN86EDuLVi+xGtBORnqGxxUrbyYblnXLpNsiU+7ZmTYU5Si7WtTPlYbLFk1CnlgLOz9LGJg1Oj9OSa6/7vmaWZGHHpblh07swTl7idatwzHGJ3EuABYpllXumQrp2/FCUupSIzF3YmB8/MaFExXTp6Xt5FE7kVAeoXIvivK8HA4HqOuKbDA13ljOLSMlLVVNj38oos8TJBrshdkZinOELp5aiMVTC30ZdL/88BSmTsgGEeG6qimYWJCFaFx2kcwQs2FeGZZPn4BNxzpwumvYNysRh8NJTSxnuWdl2HD3+lk+/3Y8qZpWhDkT87B8hv/cWpM/ExGICPdvnIubqoUZZWaV5SHXno50WxpWzfJPqLBuTim+tm4mrlwSfmLfcysCZ1K6f+NcZGi4qe7fONd3vy5ZMBF3r5+tui2Hw0kNLGe5AwiZBzIsczYCBVrTJ+sjK8OGzy2N/jgS588uxYoZE3C6cwQLJueDiDB/UgbK87MUU7O1JtmdOiEb9d2juHrpZOTa030z8uilanoRxpwe1SSTW1dPx6s79E+ooYepRdlhw9sm5GT4ZsfhcFKFeLzTllTuhqmoTrQEqtjTbVgYFH1TnBs64HvLqukhywDggjmlvuMAQFa6DVOLsnH9sql4e18LcjK1k63uOL8SvSNOnCOOEVy2cCL6Rl0oyc1EbdcwZpTkgECqpVoBYEZJjq6MwWDmT84PUO63rJqO3+8MbEBy7emqP4TVs0qwoy5xCUAcTjJjObfM2cI3L5qNb14kuE8WTilQnJcRAFZWCq6di+eXY92cUkwrFmppzyzNxX2XzMHX1s3SPE9xbqZPsQNCZFBZvh1paYS5E/NhT7dpKnYA+PzyCty/cS6uXjoZ39gQer7LFgWWeb1tzQzcv3FuQJbvV9bOxMSCLNx3yZyAbaW5UotyMnDTymm+5bPKcn3XGiqPcsmIcEwvzsHX12vfr1hw08ppuO+SOZhZmhv3c3NSF67ckxR7ug32dBvu3zgXly8K9cOvmDEBX1juT27JyrChurI4YLDVlkamJ1vMmaheOXPuxHxkyxT23etn4bY1M7BoSqGvcbp/41yU5gnRPrPL8rBsehG+sWEWCsVIHlsa+ba9dMFEXDS/HAunFODmldMwtSgba2aXYMHkAlxXNRUVE3ICGgNpDKMs3471c0tRWerPHP5fFwaOM1RXho7ZLJtehDx7OhZMFnpSF8wpxfXLpuJbF52DFbJxlnOnFYbsG8ytKj0tqcGWM6kgC7Y0wvXLpmKd2BNT4mxIFJJnfnOi4+xwy6Qg6+dqT/8VjvmT8n1WsV4kn7+UpGVLo5DKePLGJdee7hsfuam6Ap6gRBdbGuHCeaF18aVQ0/ICOzJsaQGN2+qg/AB547VmVglWzyxBWhphxYxirJhRjP1N/XC4PMjKsOGmldMw5nRjYMyFZdMm4LyZxWBMiII62jros5yvWDwJVywObFDXzy3DujmlcHq8aOwZDai/Hsx3L50DIsLX1s3E3w624aqlk3G4ZQCleXbY021YN6cUH5/yx0rL29/qymLUdg6jbcCBWWW5SCNCbecwvnLBTNjSCFtPdOGkbD7P7146B/882oGjrYO4Z8NsXx316cU5vtmRJKqmFWF/k5DNWpqXiXRbmq+mSmF2BmaU5GDE6cHKygmo7RxGTX0f7lpbCXu6Dc9sPY1cuw1fXzcLJzuGfRnbmelpuuYPLc23o1ss5KUkm0ShjnBdpYzxWHNd1RT8ZX+rrm0LszNCZi0LJs/ouGEEcOV+lnLlksnhN1LhvJnFqCzNRVF2hmIlxQvnlWFq0FRr6bY03S9b9YxivHeozXDjI0QnBS6TV+8Mlsme5u9lLNMRfUVEsKfbcE55Hm6srgiY7g4Azp9dgoExl6+By8/KwBfPmy6u81vkK2ZMwKyyPORk2tA36gwJbb155TS09I9hYkEWvIxh8dRC3724eulkzOnIw98OtuHr62eBiHD5okkhvbsvrKjAx6e6cKCpH+vmlKF3xImL5pf7lPttayoBCDXS/1jThBuqK1CQ5b/fkwuzA8pwXFs1BWVicb25E/Mw5irH/En5PtfaH3c3oaV/TPG+XLKgHFOLsvHy9gZcumAi5k/Ox7G2QWw65p/YoignA+vmlGJWaR4qS3Lx+u4m5Gel49qqKfjHkQ7csLwCz2w9jdWzSjBvUj5yMm0Yc3nwt4OhSr4gO8NXmKtqehFOtA/5ZkhaNKUA3cNODDlcuGTBRBxtG/TNZrVoSgGOtIaWdz5/dglmleWBSKjrvrKyGOUFdmRn2BQnRL919XR0Do4HrAsuS3LX2sqQ/cyGjM7ubhbV1dWspqYmIefmpBbShNEVE+JbwK1vxAkGYMjhQsfgOM6bWRx2n1gz5vSASL1qaeegA1mZtgBFbgaMMTAGpKUR3tzTjMbeUXzpvOkozM7wTbfn8bKAnlZT7yhK8jLhdHuRnWnTDCtW40T7EKYUZeH5j8/4ln3lgpkBhoEkm8PtQU5mqInx2q5GtA04cPuaGegcGkeGjfCPIx1YNq0I1ZXFmmNO+xr7sOVEF25aOc0XofadS+YgLY18hk9j7yhml+XhI3Hy+SUVhQGNvVGIaA9jLGyUiC7lTkRXAHgcgA3A84yx/w5abwfwMoAVAHoA3MwYq9c6JlfuHE5q4nB50NI/htkGZjaLloExFzJtaegccmBGibGB6YExF461DWLVzOKIEgTjjV7lHnZAlYhsAJ4CcCWAhQC+REQLgzb7KoA+xtg5AB4D8BPjInM4nFQgK8MWV8UOwNdDMKrYpX1XzyqxhGI3gp5omfMA1DLG6hhjTgCvAbguaJvrAPxW/PwGgEso1e4Uh8PhWAg9yn0qAHm6Y7O4THEbxpgbwACAkLKHRHQ3EdUQUU1XV1dkEnM4HA4nLHqUu5IFHuyo17MNGGPPMcaqGWPVZWXRhfJxOBwORx09yr0ZwDTZ9woAwQGfvm2IKB1AIYDkr2bP4XA4KYoe5b4bwBwimklEmQC+COCvQdv8FcAd4ucbAHzEEhVjyeFwOJzweSWMMTcRfRvAPyCEQr7AGDtCRI8AqGGM/RXAbwC8QkS1ECz2L8ZSaA6Hw+FooytpkDH2HoD3gpb9WPbZAeBGc0XjcDgcTqTwwmEcDoeTgiSs/AARdQFoiHD3UgDWmKXWGKl4Xal4TUBqXhe/JmswgzEWNtwwYco9GoioRk/6rdVIxetKxWsCUvO6+DWlFtwtw+FwOCkIV+4cDoeTglhVuT+XaAFiRCpeVypeE5Ca18WvKYWwpM+dw+FwONpY1XLncDgcjgZcuXM4HE4KYjnlTkRXENEJIqoloocSLY8RiKieiA4R0X4iqhGXFRPRB0R0Svw7QVxORPSEeJ0HiWh5YqX3Q0QvEFEnER2WLTN8HUR0h7j9KSK6Q+lc8ULlmh4mohbxee0noqtk634oXtMJIrpctjxp3k8imkZEm4noGBEdIaLviMut/qzUrsvSz8t0hPkFrfEfQm2b0wBmAcgEcADAwkTLZUD+egClQct+CuAh8fNDAH4ifr4KwPsQyimvBrAz0fLLZF4PYDmAw5FeB4BiAHXi3wni5wlJdk0PA/i+wrYLxXfPDmCm+E7aku39BDAZwHLxcz6Ak6LsVn9Watdl6edl9n+rWe56ZoWyGvJZrH4L4HrZ8peZwA4ARUQ0ORECBsMY24bQks5Gr+NyAB8wxnoZY30APgBwReylV0blmtS4DsBrjLFxxtgZALUQ3s2kej8ZY22Msb3i5yEAxyBMrGP1Z6V2XWpY4nmZjdWUu55ZoZIZBuCfRLSHiO4Wl01kjLUBwksLoFxcbrVrNXodVrm+b4suihck9wUseE1EVAlgGYCdSKFnFXRdQIo8LzOwmnLXNeNTErOWMbYcwmTj3yKi9RrbWv1aJdSuwwrX9zSA2QCqALQB+Lm43FLXRER5AN4E8F3G2KDWpgrLrHRdKfG8zMJqyl3PrFBJC2OsVfzbCeAtCN3CDsndIv7tFDe32rUavY6kvz7GWAdjzMMY8wL4HwjPC7DQNRFRBgQF+Cpj7M/iYss/K6XrSoXnZSZWU+56ZoVKSogol4jypc8ALgNwGIGzWN0B4C/i578CuF2MYFgNYEDqSicpRq/jHwAuI6IJYvf5MnFZ0hA0xvEvEJ4XIFzTF4nITkQzAcwBsAtJ9n4SEUGYSOcYY+wXslWWflZq12X152U6iR7RNfofwoj+SQij3D9KtDwG5J4FYTT+AIAjkuwASgBsAnBK/FssLicAT4nXeQhAdaKvQXYtf4DQ7XVBsH6+Gsl1APgKhMGtWgB3JeE1vSLKfBDCj36ybPsfidd0AsCVyfh+ArgAgpvhIID94v+rUuBZqV2XpZ+X2f95+QEOh8NJQazmluFwOByODrhy53A4nBSEK3cOh8NJQbhy53A4nBSEK3cOh8NJQbhy53A4nBSEK3cOh8NJQf4/WbcZCEnG1dIAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa8f5fc1c50>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "losses = np.array(losses)\n",
    "plt.plot(losses.T[0], label='Discriminator', alpha=0.5)\n",
    "plt.plot(losses.T[1], label='Generator', alpha=0.5)\n",
    "plt.title(\"Training Losses\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Generator samples from training\n",
    "\n",
    "View samples of images from the generator, and answer a question about the strengths and weaknesses of your trained models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper function for viewing a list of passed in sample images\n",
    "def view_samples(epoch, samples):\n",
    "    fig, axes = plt.subplots(figsize=(16,4), nrows=2, ncols=8, sharey=True, sharex=True)\n",
    "    for ax, img in zip(axes.flatten(), samples[epoch]):\n",
    "        img = img.detach().cpu().numpy()\n",
    "        img = np.transpose(img, (1, 2, 0))\n",
    "        img = ((img + 1)*255 / (2)).astype(np.uint8)\n",
    "        ax.xaxis.set_visible(False)\n",
    "        ax.yaxis.set_visible(False)\n",
    "        im = ax.imshow(img.reshape((32,32,3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load samples from generator, taken while training\n",
    "with open('train_samples.pkl', 'rb') as f:\n",
    "    samples = pkl.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fa8f5eb9128>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "_ = view_samples(-1, samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question: What do you notice about your generated samples and how might you improve this model?\n",
    "When you answer this question, consider the following factors:\n",
    "* The dataset is biased; it is made of \"celebrity\" faces that are mostly white\n",
    "* Model size; larger models have the opportunity to learn more features in a data feature space\n",
    "* Optimization strategy; optimizers and number of epochs affect your final result\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:** (Write your answer in this cell)\n",
    "\n",
    "The generated face pictures look like all have low low quality, like they were taken by a camara from 30 years ago. \n",
    "\n",
    "1. Unfortuanately, the generated face does not have darker skin faces, so it is less real in a way bacause the data is biased.\n",
    "\n",
    "2. Model size matters, in general deeper nets have better ability to dicsover more feautures. If I add one more layer to the model, the generator may be could generate more clear faces, like more details of eyes, etc\n",
    "\n",
    "3. Optimizer, we could test out different beta2 and learning rate to find relatively the best combo for this model. As for number of epochs, when I ran 10 epoch and get a little bit blurry faces comapred to epoch=50. So increase number of epoch may help, but before that I think frist should try tune the hyperparameters then move to model size adjustment for this case"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting This Project\n",
    "When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as \"dlnd_face_generation.ipynb\" and save it as a HTML file under \"File\" -> \"Download as\". Include the \"problem_unittests.py\" files in your submission."
   ]
  }
 ],
 "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": 1
}