dlnd_face_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Face Generation\n",
    "\n",
    "In this project, we will define and train a DCGAN on a dataset of faces. The 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, we will be able to visualize the results of the trained Generator to see how it performs; The generated samples should look like fairly realistic faces with small amounts of noise.\n",
    "\n",
    "### Get the Data\n",
    "\n",
    "We will be using the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train the adversarial networks.\n",
    "\n",
    "This dataset is more complex than the number datasets (like MNIST or SVHN) we have been working with in other projects, and so, we 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, *Some* of the pre-processing has been done. 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",
    "> You can download the 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": [],
   "source": [
    "# can comment out after executing\n",
    "#!unzip processed_celeba_small.zip #unzipped"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First let's load the required libraries. It is often good practice to load all the libraries which are used in the project in one cell usually at the top"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = 'processed_celeba_small/'\n",
    "\n",
    "#importing required libraries\n",
    "import pickle as pkl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision import transforms\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "\n",
    "#import helper\n",
    "import problem_unittests as tests\n",
    "\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 we are going to be generating faces, we don't need the annotations, we will 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, *Some* of the pre-processing has been done. 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 we will need to **transform** this data and create a **DataLoader**.\n",
    "\n",
    "#### Defining `get_dataloader` function, such that it satisfies these requirements:\n",
    "\n",
    "* The images should be square, Tensor images of size `image_size x image_size` in the x and y dimension.\n",
    "* The 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 to use PyTorch's [ImageFolder](https://pytorch.org/docs/stable/torchvision/datasets.html#imagefolder) wrapper, with a root directory `processed_celeba_small/` and data transformation passed in."
   ]
  },
  {
   "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",
    "    transform = transforms.Compose([\n",
    "        transforms.Resize(image_size),\n",
    "        transforms.ToTensor()\n",
    "    ])\n",
    "    dataset = datasets.ImageFolder(data_dir, transform=transform)\n",
    "    trainloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=True)\n",
    "    \n",
    "    return trainloader\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a DataLoader\n",
    "\n",
    "#### Defining a DataLoader `celeba_train_loader` with appropriate hyperparameters.\n",
    "\n",
    "Call the above function and create a dataloader to view images. \n",
    "* We can decide on any reasonable `batch_size` parameter\n",
    "* Our `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",
    "# Call your function and get a dataloader\n",
    "celeba_train_loader = get_dataloader(batch_size, img_size)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, we can view some images! We should be able to see square images of somewhat-centered faces.\n",
    "\n",
    "Note: We will 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+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsvcmvJVme53VsvvP0ZncPH2LKsbOqslGjlFpVAlagWsOGv4FhjYAdLNghsWghISQWINRILIAdtLrUtICiqKIysyoyRg8PH577m9+d77WBxbV3vx/LuK8zPOM69EO/7yZOXLdnduwMv3PM7Pv9fb2iKJzBYDAYDAaDwWAwGAwGg+Gfb/j/X1fAYDAYDAaDwWAwGAwGg8Hwu2EvcQwGg8FgMBgMBoPBYDAY7gDsJY7BYDAYDAaDwWAwGAwGwx2AvcQxGAwGg8FgMBgMBoPBYLgDsJc4BoPBYDAYDAaDwWAwGAx3APYSx2AwGAwGg8FgMBgMBoPhDsBe4hgMBoPBYDAYDAaDwWAw3AHYSxyDwWAwGAwGg8FgMBgMhjsAe4ljMBgMBoPBYDAYDAaDwXAHEL7NwVEzLJJB5JxzruA/eCxW/mUjPP7BLYd7hf7B5/F473TbG6jcK1C+uczmC/FX75Z/4fl4lFf4G36t1qzII50njdflbJngd3RDzjOtrpvNTly+HFYv8XsiDIMiiaNv/Y7mrvRPwX9Am/AYz9/cnzn+tqiMEcHH37KfeYznrf7P977dNr9dDoNgXY4j3WcUqI2zPFc5y9blZZauy7PFUr+nOt55t41dHVPcMqbnaXpaFMXe5n99O4RhWETRajxlqDcv7nkah7iDSr/4GNsBjvF9zDPcc8D+qvSp/ofty6a4OSd7MS90bGXcVY5ROSvYd7gOr4n7yzgGGTHYThw/ocZJv7+zLjfqdeecc6dnp2442s5c9DyvuBlPt41stkkQqP5xrFgScZxHqn8QqkfZhznGC9uNfRug/wu0eV62uRdsPjfbNXCbJz1jMcdKBd4tTYwxHaAOaYr5uljo+Jz9r/Lp1WRrczGO/KIRf3tsu1vun/B93QPjRo56ezgmR7ziOGdbcN5X5hHa+qbM+Bvg71jfSv+izDa/fZ0AKsPkln6vrD2oO2NWOTZni8wt0nx7c/Hm/Lz3W+Kg590SByvHbF4vfbSz52/eQxS3LCKVdfrmPPjDPNf4uO181TMjLuBfUowzxtn8lnpV9gO39UixsejyotjaXPQ9vwjW84VzSJXiHoH9y767bT25/eY2g+fM881r180e5La2/W64fSf7/wYKV7iiKLZy4X6zWdwfDJxzzmXYqgZz7BWuFOMLxMq8gXjaQkwMNz8teLf0J2NoFZvj3Hq8FJo3eWXefDv2/vbvvI/bYt9iqbV7scA6jr1NnGgP0KrV12UfoTKe6LrRVL//1eJ0a3Nxtb+5qRfmIu4nRr8k2LukKeIP2svH3OXanrJ9b3mWuX1L8e3nDu+2qXjrSVS8re+qcQR/e0t9g1vWoerzFtfR1e/zZeqW2XbWxWYtKvrN1bNqdT3jZdl+v/uyKdqB7ROGwabDK2Mey1Ll2a2C4lvVqsw5orL/wdji3pl3lKWsi84Zxhq7SaJne+7NC49jAbEM52ELfP7y8jvNxbd6iZMMIvezf/exc865tPoEhxPyoQwbCjyA+AUaa4GHP/xthJlYd2qIMNPDS4Lqe76uNQ50nkl5SOZpw5n5mPB8KcQxgeMX0XxdLpyuHy8VID0+JhfNdTGdHqpeZ4/X5eFrlWdn6qdsgi4pB+nFX/57bltI4sj98KNHZT25adExHn7PlnzgUwOFCL71mgaty/T7aD5dl/meig+jDbxQauBFS4QxFZe/N0Ks6B47S+V+u7UuPz5S2x/ggXw0nqzLl9fX6/Lr68t1+ZNnr9blF6fjdbnwVQcfC3aRalNRWZgxOT49fv212xKiKHYfPPnIOefcxfW5/gEPVlFUW5enWChmmY6poR07iCCtRGO7kegfei2dM8EqE+JBczhSv3MBrrdW84KBczbVsRFeoCwR3Kd4iXaF4y+v1C/zudp/udB8Hc81fpe+xmklAOf6232Mk3/93/g31+U/+unPnHPO/Qf/0X/otgbPc2EZ5L1bHl65mHTaiitPHj9al48OD9bl/QPFkv5Ob12uRZpzM4zz5XS4Lnfrim2thvp5gXk8KedOvdnVues61scLlBZiK6a8m3lq+yHOzU2sh7FQ8KE3UB27Pd3f6embdfn4+bN1OZ0hdqPP/8H/8H9sbS42Yt/9yU9XcSfAMuy7zS9FHDZktUZ7XV5qWrrZTPMyquuY4ZXiVYZNfLvd1/FxQ8dgMI0mo3V5Plu1e6Om9mzXseHHDihOEPMwjk7O1ObpfLYuB5WXS+hTrMtprn6pPtToWinWkhnGVVJbjbf/8/MLt03cLGkxNnOdptqkjnaIsW4laJMYC1eIcR57auc6HqxqNcwdtPkC8czDmFouNS5qN+fBMJuMr3R9xOcF1oWUL8U9zbMZNqjnGGfnQ8TZdPPLnRjzNfGx2N/yQn2BPp/Ml1ubi4EfuEGtjAto/wSxbWdHsatex4Yb9zCfanxy4+6H3DLzBc3mDXoS6boTrF0X12rTy+EqBi8YAIhi8/8U/HBU2X8y7vAl4W0vBvnh53c/hG06ZpbONxz5++H+YOD+23/733HOOTc6VP27X6n9gv/xy3U5nan+45+rb7NfdNblaEcxsfARn7C+Bhi3YYg9rceX6Ooj3+HDwWTVh/lSMXY20T5zMkV5pvJ0oXuazTHmlgV+1/h7+kox99lzrOOIL48e7q/Lv/jxT9fl1hTH/IWue/jX2N89+8+3Nhed5zsX3sQ3tVVcqPyor355/0j1fnOuPe015k27o/5NJ2qvs5mOmeYaMxGCI/erDutShDUwKOdOkG1+YcGXSPyQxVcmjJEzzOnKx008gzr0dehpXeniGavV1N4vamGNXGr/5rnVmP3lU42R74t+M3H/1r+22vvGvtYkri0F9mQB1pMAMQlLubvAB/IxYuve7mBd5ofG5y+1zp9fY+840lzzPTynlsMiwPPkZKJ4y/eztTrau6tnx50d7adC9O3lhcblFdbIHYzd9z/+YF0+PNLePMMLhjTVeA1HqlsXA+lP//3/7jvNRZNTGQwGg8FgMBgMBoPBYDDcAbwVE6dwzmXlW+kCb5X4FsznV+QCb7bBpvHwljPEm7h4ot9rc7whc3oL2Q7wRn0K+QveejdQiaKzOs+0o/MNa6r7AseSlTPDF5ccb+JDvCmj/KfwVRfP0xu6uKFrBYHeZEZ1HTNq6m3d+ETlxXX5JcHbTAX7fZBlmbu6XH2tC8CeqIH2FePrUVbovkifz/iVFF8NU7B1+AqTTA0fTJ823oLvt9TPA7x53t9bsQt2BnoLn+Mt7wJfM1r4sjZo6wt2iK9TTVyz09JX0XZXX25yT+e5GH22Lo+X6At8Uc9xfn6VyzZ//PreyLPMjUfXN/+z/j3GF8fKVzuwj2J8Ndtv6/4f7+rt84+eiOlx/1BvmcEadAm+YuVgy7x8dbwuj8Z6yxyUcyqJ9Xfdtq5DCc/VUG/ZF/hSOMeX6Gvc05uTs3X561N9jX5zqS8VQzAA9ZdVFtr5pd60/9P/7X9dl++Vb9o5B7aCctx4t8iXWDl+ET47PVmXdzEvGmRNgIUyu8ZXvrn6pE6ZAaWF+Moxw1eubjmnag2NmwpTA+drgznHrwV+jjiCeTxEn1PaE4OemiJ2zHB/e3uKm00wSl481ceMk+PtfaEi8jx3w+EqBjLO8asUJVEeGeG4/xRf5Hx8WZ1fabSGCx3TB4unQXldvPlrYT9Wmy4WqzYtKvTtzWMwmym+97piqrlc5dMTtG2+OUbGqFc64ZoGujzlclibfbAt9QV8e8HVc+qjGtaHPthvHYzz3W57Y7kJ1mIdsoYIf5uAudrGOkeZD7/yjcDamE7FeErLmNtuiZG2XOp8ZKhWZHCYQ16kulxPNM+fR6oLZQ7nQ7AI5prHXPMW+NJKdpEPOl7wuwkfvxc8J0o6WTPNhtqFv5N9NJyxbSEVJAN5yTEHFvdtsjt8cQ8xNpodxaibrdTlub44cwptUNiXxYrofHO5cvz2NyM3rJy3VJn9znPGtZs+UkMssVdJu5pPywNd/PqH+pruddXPDbRJBGZbSBYB9/wV2Tafbyg/Bpuj3HcuQx1LtjrZ5wSZHS2wKSdTjb/JG+1tEjCE3rt/pGPmqssO2CohWLLcp19AXdDYkF5hKyhy52XlNSElyTBYzoeqh5drDVli/rW7mrtkeY7AfufIDjB5uBtuVJ53oOgIVW7Eq/Zto90Yl4OKrEm/T8H+niOmLMFWHIFRNANbZ4rn5wzndxg/M0qHlroux3J+w/r5XrLMKjzPOb8c5xXpb4Xht1kSzDGfYn04P9Xa9upU+/PXL7Vv7+9oTT0+PlX5lY6hFO9gR89u+wervXoD8WIMtsvFhc7BZ/i9ms5x1BErqNXUnJv3NLdevda8fHqs/fg/ev7n6/LOvtQgDx8/UH0PdK1gBuZm+PaB1Jg4BoPBYDAYDAaDwWAwGAx3APYSx2AwGAwGg8FgMBgMBoPhDuCt5FSe55xXUqt8SHw8lG8T/kQLXSoWg8rlb0RDmxyDVjbV+6Ulkl+eLvTHOeiBISjCrbbogb2DVbmFTPUNsRbdfE/XmYLuGoKyNs+R8R4JukKHTPEVfjDaplB9g5rq20iUlCxoSsIRt3TO8ZtVW1/Ft2Th/r3gOb9MzluvgeINGlcX9LEM7TAc6b5mcAogZdALVNc6KP4dJEptg775oC/a2g/ee29dfu9QNLRumayYGcDTHGMlZRJIUPOdkEOGEzDhKmR+EXQO6X2Vn70RzfmLV6/X5QKUc0rxlpRcvSPaeFHkblnKyGK0i4dxmCPxcpyLKv6DA7X533l8b11+ck9JcR/sSyqxM9CEqTMJHOikzKv+waH+drFEQtJSUsDEoU3IcgpQRudI8rcAnXkCaccQcqrTM/XRZ8eia/7VJ5+vy19c6HgPSc/Gmc7Ja/36N79al3f/yaoNhiMEr+8Jz3lrWqqHsUcXAOaPjyBfYxK4GHN3eAVKPiivJEwXmcZCwS6E/GeRqf19kJVv5IqUIdINKsFY7CSa83TNWYAq3oJUqsAcHUM2EtdFke93NBaXaJs66tPtgk6OcbmYbVkKd3ONMHQ7u6sxX3EsLJhYT/ezSDkvKUlkv+s8vZokxL1d3eduT5TcBiQ6t7oooc43dPUlZbGQp00hoZuAKu4hmfLBYHddnjOJJ6XNkAvNFpB0or5JDbJrxK8MSUQnE1233oi+dez3xoo3vqoP1qcWJFEPQNn+6D3Fzf2+fu+10FeQ89aRcJ9JKulkQQ3NFG04grSRUqV52V/5gnsxOn8g2eMtcqoi0L0yafxeV+OsA6r4U1DbX18oFk7nmN+Q61aSldMFZouUf8L3PdcopUpRxcEPSfNh1rBkXblvw9CqYQz0mNwafUd5Qbbc7NpGSUFFMVsm15xh3M0oVaMkqqKUqqRcRcnb/Ae3Oi7hkO+gi6qada39fH7n331X5IFz49JlgfWZJypfPVLfervYq+9qHLaYVgErYIDUATz/AnFuMkcMgzyXEqmI8pKyuygVWlC2CMlOC2tYF+OSe4Czc+1h0gWS8/paUzlfGR8HTcWaJYwrMqRzGOO557q3vXQNv42g3KfnkNbQSWuBDcgF7qEWc2OCuUWZI+J/gmM66N8DSI6PIIXpd9RGDaRguHHzavLZiEnsES8Zr1OWK6Ycqu/JhZ75zq7Vd8fXkvecLvX7opK+AikFkOia2ROScmxuW6p6ozL0bzNtAPhegEGU9Z8hIXW2wLOtlk6XIOY2ajrPUU99dW9f8+j9h1qP7x+tniMp8x9BTnV+BiMYxLL79/X8eQSpYqdH8wnNodPXWgt/+dXTdflvn73UMZBCf/a5jnFO12rHlPS+/fOFMXEMBoPBYDAYDAaDwWAwGO4A7CWOwWAwGAwGg8FgMBgMBsMdwFvJqZxzzi+z7ZMGHmWbqZwFjgnH+n36G1EFr36tjOTFmahK9Ux/G4NP6ofIGs9rgZZ3FSMr9dWKrtXbF+W5doqs8i9EZWrviqpY9EXbuqrr+MtEdZwhEz2pY16FwsoyHKxCZJxPRJcOY91frbmiYr2ORT/7vgiC0HW6KylGBKplG5S1+6CppaRsJpRWqR3ozNRqqw27oJA3ILPoN8Wb+/ChaGUP4IJEmc2NOwap2WkKqh5GcQ45TAGK45IZ4yEzCuAeEYN+S9eswz1JV56dq9+mkIhgJDiP1NF3ZE/led5aXhZAGkSq8GQoac2jviiBf/yTj9blHz5Umw966pd2XZT/ep3SB/WvR/chX+VBT+MnwPi/qaZHfjj6qMg2O6HNUZ7OJUtpgS7ZBC05qcNlbKnjrz55ti5fQlpFMjF7awq3khunqtF45LaFopCxGF2dIrhk1EDrHfRFB65hPsWQWUV0gMHvMcck6P45JC4+5HH1usZCTGpxSXP1IWFMoA3ogcqdYFwu4KrUhMxhAnpqFzKTMVzNWN8E9xGB2k75SQb6c4hys0nnnu3B85xLotKlBUtqXlke4LyI+++2NC+zGMfAhWQXrnm9puZfHxKyiAoKyLUq0irUOS2bLofkK4NjxwRxZIg5N4X0sI4zvncg+evL15KcxnD5SOq61uWlJDqM03TuCuCAGaEcB6vj2abbwM0cDDAXOy31w4N7ko89eCDp6V5f7lADUPbbLcyFUGOP7o8h5gjVLt0WaOb7cHaBdHk6XY35dIK5UmyWClGQTbfNFN/xepCw1tFXCaTQAeLsFNK6eYr5V9kLoY9Qd7r/bBN+GLhO6WyS0dEF5Rj9228rznVaus8eKPS76N8u1j/2Y0WFhHuriuzhrIeqHZ+s6P1//clX69++fPZiXZ5AZprdIlvi9YPKggZ59Zbcqb6L5Or7oPA8l5VrYEHnHigPJ0/Ub35b99VAIPTR5xFtNbHiT7DOzCmhwpgP6SAImRvX7BsJSxjo72rYN0VNyKXpCAlJ1mKsvXZcqC5NOBV6NcWRKEEc6aOMdXyJPdXSaX5fNPR7a6A6bx3+t93LKu0GaX3FwQrjfAH54x7iq4/+ihHHHu3IqfLjQ8lsDjqQt3JvhH1EfOOgGvGZVqA7ls+gihvMKanD2BlBXns11jr6ZqTn4adXer54cSq5ztkYe2M8d+CRzLXKdvW3OD89T2ObcqqKOxXl/2gUOmPmTL+C/m9hP9Os0XVL8+zxPtbXh4/X5SfvSfJ0/5H6eXd/5QKVRJordJgbTyBPhmS7v6v1vQ9XqRBzLsUa9vBa7y76D5RCovUr3dPzEz2DXU0gZ5xINuXlmrupQsB3hjFxDAaDwWAwGAwGg8FgMBjuAOwljsFgMBgMBoPBYDAYDAbDHcDbuVMVnkvmK5p1dw55zEJ0oBj07NlU9L2zb0Sfdp+pfDgVFetwXxTJh8hCfdgW/apR03unHLQsXMqdXumcry9Wmb+Hz0WhihJdh7Swq09FfSr24CBxX7T15hHcTvZAW4eciy4HdIXwQJ0sPDgqeaLThbWvVc9gRRf0Q9HZvy8KJ9eaFNT4ZkPU78Md0dfCXO2wGInr5SHbf72mvw1BU6yBkr/T0zl34ObRboFmTucj0hnL81CmFqS6Dh20UsgzFksMCtTLo+SKDGPIqcJQ5V6X0iLVcQbHmYpjA10JcnIutwjPc155T5xnDVAY6xhvP3n4YF3++L7opoMu5A4xXBDwfjcB/T7xGDLoKgOaOV1BIA26ofoXcOGhhCAr4IKDujvQXanEorRgCQlJBzKWj0Cz/PJcMeDzM9Go85ySk83lG3ehPN+uhKMoxwolJZQsUFa4uyvKJiVuIeZKTg0PxkINjnNLzOMMMqcQtNgZZGNFqr9txKt+Jj2226BsEm5T6J9awvGhYwJI1uagtrYhbSgg3SoWqnsSIzZBGkqXpazQOZuQH20TnvNcUA7kipzYRz/WSD+GywNknk3IgPttxEjExQamRQyZYVJximP80a+Uo8YlNTpDrFpS4hJ8O/6u/g4uGRP1RRvygz3E+uNLOUG0IK9tNyW5pONjkdNpUPCd2qDIyvoU26ON+763diEixXuAteroSBKqg31Rr3cHkttyHQ0R+wJS0WEf4sFZLsR8DeB8xPMUaP/GcjX+s6muOZsprnH9y/LNMvQZpDrs/w7W5XuUuTrhGO5Up4gpPupL57o0e0drIeC5wnllLSk97CCO7kOWerircdiIKUtVOwdcijB32V8VIS4k1D5iKuN6u69Y/q/88S+cc8797ItX69/+63/436/Ln3zxdF2eY62iO1VFtc/fK3ZWdGui5ZV7K2yWU21TYuW5vJwvS1+xYZZrbC9bGquNJuSzt7h0FdApXEG+wrWlAxeqCHNhOZJsJ4Sczo9UjkrZro/2DvF8QokN5ZoLuF0Oz+ReVGBd7GHsdjuKR1OHWJkhjsOF8RTnKbAXHQb6fdh+66wa3xm6b8wJtG3oUfpLZyPu+SDbhztUFzKXw6akSj+6pz3fEdqrDflZA3vUBPtLSWr1G1vHoxMvhfiUU3GNRCyII8X0GrYinRB1hBywizb4KtTYP4GDUS1BapDiW1X5/iikcqvsyVE33+OqwHVG8XGG2D9ZaOwFIfa3iIn7Hc2thztK+fDo4OG6vLOnNbh1oHLcXT3fhIHalWtugZicYiz6kMcG6KAiwO9O4Bh6iKVtdqW+8vBc0kLMOoecbjTWOn15+vbpGoyJYzAYDAaDwWAwGAwGg8FwB2AvcQwGg8FgMBgMBoPBYDAY7gDeikcXLDzXfLais0VDUfbipahKMehD2VD0wIOhKIyHPVHinnwoKvDBQPSkQR2OJ3Wdn1nDyeAkVXyR6raG09Xx52Kgua9fi1r5m3Pdx/MrUZwuP1Pdva9V3+wQrg0/FhW3+Z7oX3kd1H66aYGK60F+5Re6rgO90EVlpvIKXe37ocgLN5uu6GxJoGvVQROtuHOAVlYjrbMhqlqzISqjR9kEnDcos0iYVR/uMaTt+6Q4luehe46HdvU90dFIMfaZSZ4UY8gQ/IyUV2qidP4EKeATuhyMmZUdmdgh6fK3q75ZoygKtyzvIwFNcwqHgz7qfa+vsdqCtMaR2o97CECJp0TOx/E8JoZTVES5DF2RynBDRZYH6ukC7Y/p4dIUUiMQGgPMc1LV6dy0AyncvT3JPBJQ12EE4SrccsSUG9nTlgw+1vBLiiop6gXGbQxZRaMBiWeNoVttXAclfHcg2UAjgtwTjmw55Bw19FUKtycP0sIb8wHSU2sYfxHozl5MKZ1qm8JtottULLiCFKQFOcfVuWKxz7mLedbkvIREJF/QFeEdTUbnOT8vJZ+e7p+OiSEGPR3rapgrA1DCW2jHFta5JmjMDVD344L05lukVehrv+zrAudbgm5PeeIc7TxdbHYWmyzUL7vsu7H6LoArIKW207nGQxjq/LyPZQCXmhtrrS3KqVbXXp23Aclsp6176cB5rVmTJLsOWUUMmbGfwykO8yUKKclhnEUMpYNVRDmVEJTumJTtkfIeQQa5xLxZIp5W3KzQxjWu3YgLHcg4+x1IobFez1GmiySGVFXGvEUkceQ+vL+SU7x3Tw4j9/YlhatDsoCw6HK4eS1A+aezJd1JlpCrUSLiU9Kaq91hDuTmS8XXdnc1rv7+L36u3+Ck81/8V//NuvyrT7/U9SmtwlzwKXsoNu+Hior+2/1zhdwVblGuURn65HIkt54iwmBCrOTmgvLgyUiyzgKiwAbW15wOpjh9s6O9U8Q1GM57631sxRGVsiGsv3D7mw0h54oVU7p7WsdDxIIc9zfL1Tbjkc55vbzC+TXO0rn2hpQInidad7eNmxbgPo/PFzHWpABrUYj9JPd2lxe6t31IeB9BinPUUDv2KEXmXgplxq6bbVjg8/lM9+NTKgkpE58/6QjmcK+OexfGIKwTBVxhwx1azuk+3BWcxeCcnN/IyLc4nz2nZyeG7ByyqbASbtRuMDB2L071/H9yqYfxXguSfzhyPtyV89RHR0oFcW9fbsY1yGI9zFGvuZJfFZhPBeaNxxjhNkvEqs8BCgZcLz1f52/V9WzxeP/+unzyRqljcuyRah1JsK+uNf/GU5NTGQwGg8FgMBgMBoPBYDD8/xL2EsdgMBgMBoPBYDAYDAaD4Q7greRU+SR1s79a0aOLJeijpHWD4reXiIb04UD0wI8O9O5oH5SxVh10Xkpu6nCTglyHlNAC1G5myO6WlMqjuY599ECUtf3XojKFn4vK9Fef6T5enYvaNzzTeepv9LeHf1cZtLs/FLUr6IHG7KnskUKXkw4NqcM6g/z2uMeFy11aUlR7rRb+BZnEkd2+AypjHfTRNqRVNch2MnDrEvxtDNlUxZ0GtMaC7hyglq8lJ6B4F6AmhhH5g5BHYXxkyI5O1VREhwlQYUNwvyNQIluQnXlOtMCc7h8FM9i/KwlHsXYbiCGtuR6JblpviZLdhETAhySRbRp6CY6B5Ar96KFMymtYg6QGfeoHlDyt2pqOGZSeMfW7B4nmEnILH3TGALKEmGONDmk4ZxP33Uh0PB1EODb8d0wzDwJ/LRta0oHJ3eLiggrRgYmxsglXoyakVR0ckxWk+6qY8IYxnmtw3LkZL5S9hCGlkorVDnI7zkUPrhoB5t8ccTvOdM50CUkCKO8O9FTKCWpwmcnBG6eMZZsoCufS0nGN1yggd6i0F+IVY2Sd4xNxtAHpUQNxtAGKdQRJTYBrOUgrAlDXb5jDdKmhbCQDPbyJvpjP1eZTjNlLSOFmGF+DlsYDJVcB4sII1024HuguXMh4dNPXW7Th8D3P1UvZQhN7jJASXlwvusWBj/JROpxQTuXTqQbjnxKqgGWshexPr5RDFxVKuM6dwX0yJJV/Tvk24g7kbpRezNBXTdSFLmRNxJcx3BKrMlE6KxHbWyNrSeI+/vCJc865PiRJg55iWIQgz9hfq8vhhJKIyVRSlQXGf0Y54VDOprO5JCwBnNc4XGHC5l6+fLaq4z1JBf7+L/6FdbnRVEz/z/7Bf7ku/+o3n6u+OecCxyzGS0Fdva+VAAAgAElEQVQZHdt8G+2/vbnoucKF5XPEBE5S85nGJMcb5RwLtPfwXBIquoTWsV7NIXlK4HbU35cUz4Msg+5UIaSFftmGdJXyMS/pyJnNkSqiIUmIDylNCldFusallecunT9TtdzVRHtAxqM5pMUu0e+zzrvb6Nzsf71bhpuHdatWgxQZfxAi5uWYix2kCNhva653KZvCepxg7UwQv2s4/sZxivv2SiqGnGkccEtYLz1ISP0cew48gwaQ9KW3aEvrhep40FH8WvrYD0cqny1v5sp2+3PtL8b7Rf9wr5BBqnR8qj3BVy8UH/1EfdXuavwnIfc56EM8W/gxKoFyUVFCrfqoIn3iOoR5mVdSbiA+QgZbZBpzXsG9KBwZsV60GrqPXlex4/Wx5KAZnv/HGNNRUzHgu8KYOAaDwWAwGAwGg8FgMBgMdwD2EsdgMBgMBoPBYDAYDAaD4Q7greRURVq4tHRzyhbI5J6LDrSbiEp7NBCV6IcH4vvd24ebSg90OtD2aw1RdQPIAvKATimQHoHCm2YqByX1rGACe180qB/koHAVynDuzySV+V/+VjKrKySPvpyJEzteyu3mPhhfrT+EFKyl4z28P/NRuZxUL7d9+r/v+a5Z0kBJKUwhNyLVq4DbShBCShQw2zdo45Sm5KSnoQy6X0bqH6iVPhwHipLSTgo+GKYVJym2a0h5FinsuCYz/1cIyQu0B+QclEvEOP8U1PUM7hSUS2wTQRC4QXdFs/YdHXEgA4N8Ige1kC5cdERJPEiugmRjOYSEI4bEoQYJQpRovnp0KyolBX6l/wWPkhuPLg+azznkDXRlyTA2ElCIF+jUBhyM6LISQupQoT1jTtzIToptShsL5/KS5u1XpBeQJIJKulhsloxWJGtoH0pT2J4JKOEJ3BtiyEI8/G2E+HvTRS30cQuuSgl+zzA/SN8PKq49cOqBPIfzkg4+nLsZJGWTqf426ak+7C7KW7aJoijW8TOHnLheVz06LdGGU7iTtCALqEMGllTcjCBbxPzz6eTEecZ+5NiuhKKScpyR264i5T901vIr7aky6dXFQnKS/cHBuvz8+LUu5avu7abW+ulCa3OOCsWx9gY3V/K87fWn7/muXsYzyqNiyqYwJ3jzKRolCDgvvy1fc845H25vAaR1IWTJYUw5FWWJOk9eXrfIIJ/Bls6DJDhAHWOP6wJdVTR26UgY+ZvLfUg3O6CBX44gxZqCfk5UnMW2F1P9IHCt0u2pO9C4ojMd58RsAko8lJqttvaCULm4PIMkEH16dqz+ungDCRWkFdyn+Ix7N+dB7PAgt/jFvyhpFbbd7j/+T/7TdfnFa1H1c/QvZQQVCX9F38Lj9Wtxm4XYhp+3KeAoityl6SqGDCeSU9FdL8Jano4l25hcyg0vX+j3GqQ0M0hQ2j25xDTgGJNhzxPEWi/9Gh1v4KxZyg8DPIcsYXEVwcGuhnMkTA+x1Fh0c6xzcEHzKF2ByyO7KsPeIKrINXWtBSTNi9o7lFNtGEPYVrkF1585pYc6qIX1rweJZA8uVC246CaIl3zGiel+y+cXOtqW8c3HzrSSrCGo6Hb0O+VUfDbBg0oEubTvID9FLBwjHl3A0ZnOfm3sGWY4/7Kc38WW5VSKA9yf499xvdFUFf0EDrBXkJIdHigut9sq19CHAfb2dM2c0+0P5RDPXDeupR40VnQzLuBmXDGhYkzMmfoEQRey5HwCy+up+iovsI4maBv057TQovLs6zfr8tNXarPvCmPiGAwGg8FgMBgMBoPBYDDcAdhLHIPBYDAYDAaDwWAwGAyGO4C30nqkWe5en68oQQncYw4gE/rgQPS9Hz8SZWh3TxSwFiRUcQsuRy3RGWM6UkV0/AC1yUP26QakA6AiBiVVLY8hjwHdOwa1eb+ld1o/fyyq8JsrHTP+UhSqN6nuY3Gs83/z56KNP9i5vy63PhCdkbKjrKKaAgV6TVPbogtH4LtGs/Gts2agfkL55FLQHTPIESiDIiMyLDY7I1CGkqMPI0jJKgxf0BP9ktpNyUkBS4IlHAZSOIBVXH4o8wJVvCILAzWRdL7C0/EL0ENjHBPhb+dwSKhB5rBNRGHoDvZXjmhnVxrPdIprQZLCNk8hQ6T7Ct1r/IgUU12XUqgA7eJT5oRj2KV+OVBySL4yDJ6cmgP2C20AmGW+Ul9SaNXmV6HiC+UqrYR1hywLNHYftS9KijKZst8XeZG5yXQVTwLQqgdwfYlAAb4EVXxvb29dphtXB/TNCBKC5Ug60AR9FUMq5SDdimP9TkckV0qBuqAy0+0jgCSPbnNLusnRSYnucJSUQaoXYpxNJqLI57w/mhbkolwv4eaRBLDw2DZK6u4ckrd2S5VqwHkjgBSgjnur457rcGdoIoZEmBecXBncEShdpQSBThnrqUYnPa4B4LwvcEwOCj8p5DX0XYXC39AeYKevOHV8IflHpwc3x4nu+/RM7jI3Lm7OOVeUVHtvi/K4whVraj/XoVrFeYplyJ0YB3FOOq9UnPdYb4zbirT0Fro65bLFjYQPf1mRyVBK4zbLAJzHuA36OeroV9oA0j7I+TycJ8V6nFdWgG07In0bgR+4XnMlXXz48PH6925fMWEBl6PhuVxTLk/P1uXlXHGmIklEv4wRU2djxeYIUhDqDyt9A9ldUO5BF3AEG080VzKsix+9LxfUn//BD9flN//of1+X55i7Ph2p2L8VCdXm/WXl54LFTVKs7e1R8yx3k4vV/c8RY6IGUg3M1D7Dc/VbhnYLPcYw7NWaSAsAiQulitlS/9NoQtqLtl2O1f9Z2XcpnF3jBOkhIsUvyoM9SsUh4ci492D8xdyawR2twLpDWaAHiRLdd3ymRniH3/JvHOm4l6K0cUmXJkjZW9jP7cHB6D4knDvYl1ScJykzRqqBAPKagHt9hOmwfNak5CurrHmb9790aOI+1sNaGGMfS6lwjfI3tM0CGqoF9+9wSIJKyY1nq2PyfNuxdVVXpsfwKJ9Gyo1XF4phX11pTC6RzuEB2r6Gfgs8ygPhQuujD7G/DBxSPkAq6krJa1YZB5xPvDWurZRTQfKMsZDT4Q3pVBzmcZpuHi+U/7/BM9sZcrScYg36rjAmjsFgMBgMBoPBYDAYDAbDHYC9xDEYDAaDwWAwGAwGg8FguAN4O3eqzLnleEUP6kaiG/3wSFTBvwvJ0JMDOFC0VE5iUfH9EK4TyEKeklycMVM4pRV0MwI9FBRJr5SOBOAzBxW3HVGiIk8U2sOm7u/vvSdpweWl6GJXJ6JWTae65uxr1ffqL0WV6u7oXl0X1DEfNHbIhKJsVb6F7fr7oShcVkoMaJyQhqCJglZPKrcHGcaSkhxQ6xJIVujsQRckHxneqxQ2UNVw/qIsF2gI36MGbTNtnHIuUsJD0DArmeczumxBloKjlrc4BJGSHIGK3ojejZwqCALXbq0o4jQA8VEPujQFnDaUhKFdIrQL3UxIya7QYnGeHP1F/qFPCmlJUfRBa/UZgnCdLKMkEvMcv8cFpThwq0jY13DCq0FOVaOcStelmxXp5+t5UFD08P3gOW8tJ6q4OORsP9VhZyC5Kd02fBzfhMwpAFU5S+WI5NOxAbE4Q1fwukvE4lrpZhVElCRSPgfqNyn4lIGA5k5HgILOPv5mqR4dwyj5mUGu5Y0Ux0PQxts9UbG3Cc95zi/XIkoyHejhuDXXAM03RMwjPTuOuLZR8gJ5Kx3/6MTgM05udhEsNpQYlym5XIAezDantJL0cF4znSteHu1JCpLCwmMKiUANzmlz0JWvLiVjbt5IsLfJGi8KV9yMIUrTcL9hRT7KGOY2lj22Ldcu/m3FSYjVwTzK6MhBx8esPHaz9KLay5uPqf4tJQ9wzYoRi0OOP66F+p108uIWav+78aZauW1NSwunp0+frX9vnmDPh05qYX1uwbFwNLrCWRFn4J42o5xmofhKeRtClwtwrRjS1SRZ/b6Yi24/n+o6dG2pw531h48k1f/lniS4X78SJT9nXKhsdrghULGiNNhsYLVZU7zFPWqap+50upJS5njOoLvW9bnkmOk1HKnowokY2k4U+yn95N41hTtU0mxt/H05ggvpWJKIG9eaGNKfpEVZNNJDcK3nfoLPM3xewf7HY7oF9GGGeJ1CQjVDHVPMxaSt+jTgdvTOUIltKFfWP/3egrx8py0p5GFf0tuHkJ3X6RzI5xqcv3KXfAZhO5ZtRElSDg0lJcQM9ky5wEcTrhm8vwxxtIH1fV9bPOdDXn0BGfnrK439gpK69djfooOqcy4r28dz7CvMG6RE+OaFUolcXis+xgn24dCvRZjTccB3BIh5DfV/XNO7BjrfZngWuHEeK+hwjGf+vGoVqTLmou8pnjunMtd0xpoU+8/rkfYqM+x/YjwDX19KKp7hbynf/q4wJo7BYDAYDAaDwWAwGAwGwx2AvcQxGAwGg8FgMBgMBoPBYLgDeCvujucVrlZKD/bqohW9vyMa1P22yh3Q9JKG6FEeXHNcRLo3KGk5aWhwR4CMgE4BFXcO0OPyUv4C9r+Lm8gUD8p7AhZaqxCF8kFflKi/80DXfzrWMc+uQX+eiqp49inkVD8QtbGDDPkugowENOagpJoVW+SqFrlz6TrDtuhjc3D9cvAdI7ikVNw2UObvBZ2E6MIBDrMfUVpFx5RbyjcURlIZcZmQDlCRxllOGRDaOKpQaEFVr5wTrjEh3QR0/IJURsiPIlD7muG7oapmWeaGwxVtjxS/CDTRGlxwSJklDZftH/A+KZuisxelLaBVx3QWYt9toGQXdBKghI7uY6BZ0uWINNiK+Rjp/5jsnPf9psbGw11RcXtP36zLJ6A/pt4mocQW/ak8SVjokne4K8rwBx+8vy63m4ory7Ho/h3IANgp2S1ua5Q2kuKbI4YuZqKQJogBYSnXyivubZjbFZkJaOsYWxWpViYJASWxM0hpWi1R1DsJ3LquzlAWpT5EfO/t76zLXvb2VNXvBM9buxU14ZLYa4Naj/byU9KSMW6Dzd9U6F7ItqO8IARdOIm1tjB+B1h3sxvqMCQxBfrfR1xMp1q3FrPNLgyVdQt06Slo4AGktod7B+vyybXoxy+ON7szzKaqZ7+7mh+3uer8PiiKwhWlw4QPmnQd+w1KTytxrSIbpAsjZeCcc5vXOd9jDL1F7sJ4WdYzg9yq+u+bpVK3Sbg4trgWhNFmqcBirhiRLjUuqpIu3AY2bFXC//bo/1m6dOdnx845585fq12WM43DTltSmfeffLAuU/o7QfxZ4N6Gw+t1eYYYmVYkh7pugrWuC3lEM1aMSsrf86XON7mWaxaXvxZkWB8/PlqX/+jHWidOL7Q2DGeUw1cGqurrCA6IjYe7gmGq2F7fqT65m3irvXUH7pKUL83Hihk+5E4ZpNG9nmJ/vSE5VQzpDcc/2zlbaryM59rDz6/hBDrXtXa6q2vV64r5cUJHKjy3YA+ZV9wZaWG32T0zCOHyCPl/yniBGJBhXM4R3+mClMTvaF10bqPMriLbxPBp1rVu9Vuao02sW124aXbgfBgydjG1AeZfgj1whOdOhiulY1DFQj6jYlLw+aKyFlZirc5NZz9KxKFWdT3K0VGvOZyqCsRRyuga5d7D36pro5yIA7d5PaH0bA4nryVkVjXKunFjTUioek3N0WYDzm5wqg6RCsBhjcqDb+9BmXLDx3NYDrcr/h3d4ZzDeoZ4nuWa/8tUcWeK9WU8psugYvp8kW485vUryePy3yOeGhPHYDAYDAaDwWAwGAwGg+EOwF7iGAwGg8FgMBgMBoPBYDDcAbwVjy70nRu0VpSgAzH+3aCBDOCgUvt0PgHFG8nWnUfbjgrNF9TvJWi4S9CNQhzPjNe8VraiP/mQR1Wo/ZQWQIoTNJDZviuq1EdisLo/vFZdLhY65nQhKtYEMqvXnyhzd/veQ9UBDjoR6KDeb/13GyhcsXYSmoEmPIfrQQgKYoVKGJBirf4hVc3hbwtmgPdvk8dQZgFpz4ZyQHkLG4WZx5k8vmBb3sJ/Zxn0Py8ETRE0zAbozLMlHBJucXDy3wHd2Dnn0ixz5+cryvUCsiI6YzVBG06izVOdkihm2KdrUMbfQUV0IccAqKLeZtr8DQXTAz3xNtlWxU2LbjukUMIcpeLaA753gBDXhrTkUb+LsqRV44lOOoZL240p1rZ788bNaX9X1O+dnqikpOafnSp+DOqQ7NWUvZ/c4CBQPItq6p8QjbuYiOI5mcFtA7TeIqZb3eq6BaRdFSorJmbFmYZjCD8vIC2aw1WDsaA/UNsEoLzWFro/yh8rvnU4T5ptM5IKQRC6bn/XOedcDhlSxPnBP2CsoJyUshmM+ZScc5yohvkdJ6Qfq1wHRTlE7MrLkXwjIXLOuQx1n0FCtaTDXkwJCforB40aMhu6bFF+s8jpfqRj5guu06RIq50WZT2LLcZW35Oksd9V++1B2lhD+2WQSi1x7x42NwHqR2o5lngXIz5xzaMk3LvFWSot5RFVt6nNcioiZ8y9Jbb6WJcdZAApYusY0rrJVOWcDlpVv6ON9dkm8nzpZsOVPNanBAFyhyUCxAVkmLWO1oSzse7nm2++WZePj7Xmn19KWrXAtbh/6rUUow53NBef3ONat4oTni8Z1DxVOz8/Pl6XGSPiQPvcn/30o3X5b776el3+zVPKE//Z6/JvHVKR8ReV3789rra5Lnqe56JodUEfi/xiAheqJeJTrr7qYa2oY43y6LyHmMv4m2PfPoV0iy5klI7s7h6uy3Fn1c8h4pqP9cyFnKNcT/E71mUfMbFwHMe3SBKxN46ZAoF7ZsRTjqMw5or5bsCZz/0h50qrJXlUUKkrpU1YE4LN8YRyuSbO2YBzJxVH2Rwys1L2U5Gw0R0SjT7H+jeHFJJpFlKMqQTS3HpTe7YIz3817FESdh2dILGWLDDGmqWj2jblVJ7DlqPiKqYyH+f6fexdv5YD02yBFBZYL+vYl9JZ1ce+YQLnvwDtHKPdOKby6aqcY/8e0uExxPzAcwtdC6sBkmMOezq6oMG1eDRSvBgNKa3SMdOZ6jaepvidjo/fDcbEMRgMBoPBYDAYDAaDwWC4A7CXOAaDwWAwGAwGg8FgMBgMdwBvJaeKfM/db67oR493xSXqtUj3Fk1oUeGHI6M2srovkL15Mhf1bIbszUvQdulIQ9eEwa6osI+Rtd8r60P6UgbK0uW5KKwvnssR4M0INH/cXoTs2B8fiWr99Er3d3UqOhWSdbvx53A2+FD3VO+Kcpv4cDlYt9/2yKpFnrvpfEUVDUHrjCF9a8LFJ6BjCvqTebxnE1FPC7RbDPp5HVQ5JBt3rYb6sIY07UHxbe5eDukF6bHk83lO/UNq+QI0/RGo3xnkGXm++f7oiNbrSnqzzJ+pjLrVKSPzti3AWSHPMjccrsZZhrkV4HrNGh1rKHdS/WZsU8wLH9zCBH2RQnaXeBjcU83jWg3yOtz+Dc0+IHsb9c3ploZ6TUAxHYNOPFmQVgoXBvztAlRID1KwNuq40xHltn5CuRxo1MXN7WyvPz0nqm4d7n2kWL85gQSzIfp83BFt1aO/CGQedGGjU1mIe7g6U/zLUvV/r6dx7kGglE/LmVHXdehqxez6pKrPQW2l28todo1jNEcT0Fw99P9ijvsDbbwNNwuwjd1kinPWdE/bhO/7Li77b3ypezu7FJ04xVyk2wbH/5yymEz36aN/Y7pc3UJLjyH/9BDLY7RRdHMM9JQzyOmWV3DHoSyVFPbF5jXdLTgG4ESWKkaMSDPvSmbSH4hmTslSuwVZWCn1qUgFvieCIHA7/dU1Pnj/0fr3g4NdHQMq/wJj3ltgr4A5lCP25HDBDEHnbnW1h2jDNanRoBQPjiUVF6TV70XFuYRlyrCc/o7UfLRhAQkV9zzDma4/nIIGTko4XP04/yiddLethVtcIvMsc6PhSnYTY340ECty0PmHI8Wfb96crMuffv7Funx9JSckSi/mS7XLCL+nmFPXI8WfBdoo3CCX603V55NU1/z8hSRcz56/XJf7HcWzjz7+eF1+7/69dfmr55IgLTLK3IRKHxUbi1U9TEUuvf39TeCca5XjcjGBZG2ieLpYaG2uI34kkKl4uLGlo2wU0mtISCeIYfORYiEfkg7vP1iX23AfvEkLkM3V3zPIR9l+lN5zz1NgL+phnhfotxSxgC5IzDLB9Abs80mm8dfgmo04u23c3DbXKv8WJ0NqnChVKrBhpPvYJdYrCH4raTYagWJql881aK8k1v/cOBDO0M5LrMXPX2r+vT6TVDFF3OUzUw5pFSVRrabm98GO6hij7k04cTbx3FmPtA/0nNogXzs6b1e2qvUfe2kGeTyz1iKmktAhKY4P6rqXGHLTOebCCOPfH9EpTu1Zb6h9ajVIJ/3V+QvIgCP8e4Bn3SDBmEAaFjplu4LpQiD5i/msqUOWyG8xhavm5SVkVhPFozHWVzyafmcYE8dgMBgMBoPBYDAYDAaD4Q7AXuIYDAaDwWAwGAwGg8FgMNwBvJWcqhZ67sPeinL0sC/KULcp+lYEGjwdGRZjUQtnyNJcgKrk56K8ugkoRleiUw1BSQpBLY8yvY+aDkSv9Eva/2ws+tr4WBTTl1/pfJ98Lmra87HqfolXXUkkvlO3Lqr1B/vKiv/VUMcsppCinKj8+nPV5+gDOIvUSG8GRX1LyIvCzUqKX5NyNDj0NOuimFEGN5mpD59982JdvoAkbQ7pG4m2ESj+tbpoaw/uifr7/kPR2Pd7cKQp/5uRYke3JVAyF8hCfgmq9MtXqu/Z2Zt1eTaDbA8ynGZNdWw1QHOvgX5Hyl1OyQF+fkevSbMsc5dXq3YvIHc5bKHeLdEG6d4zQzuevni1Lk/gyEGqLtj6FZcYOnUdYM794P331+Xdvvrx5m9TSivRXzNILH752efr8mfPVccRxxdo8b4HNzGMrwC02FYCKQrGT1LXffi4JzdV3Fmn49+qDYfvvHA1nqZLDZQJ3PhSNP79o8G6XIPrAvszA/V3PhL9PEzUJgv0IcdnE3KObkf96Xlwaiip2klEmZr6MAVle4I4vywo4dA1U8hP6ohHHinhpKXju0OIYxLIJVK4KaVojwTjYpvIi8LNSweFJV0Q4Py0QEMHmE85aNj+nC5AdG/S77MMDiq41CXaugnHih4lSeibehnH6Ig0gwvENWRoY7T/aKI5QUeOigufR1o8xuOSUmEd3oHMixLcOeqQ4lr7eytJAyVk3xdhELhBKZXd39U8q6M+aCo3gzz39ETrDB3ZWpC7BJByDsfaf1x/ISchh7HQaWpsv//o/rp8uH+wLnvlGMlT7qcgoUKwolSYDlND0NOfwoXp5WvJBq7PtVeha9M19nRZTokNXUad2/g/t0i9vi+Kwrms3HdmlPJRcop5cP1S9/z6VBLCFmj2j3744bqcYE92gX3pNWRmb87VvzPsmQJIJW4k7c45d1ZK8dOlrnl6Dnesl5JTneCa5xc6xwnqXmtJ5tPGXu5sTPdBgbLESt9VDno3/bURRe5cOd+nQzjcTDUOE7ja1OBCVXHGhFQqCyHnvVI7jEc65vxMzwLpTL//4P0P1uV6pFgVLuHcWl53nql/JmOV37xRHx5DtkeHNzqI9iH9qWHc+C3FoxrG6BLzz4NbK10AT840RrrY7/vJu5NTreuEwUS5EdMizCE9CiA/nWN/8/pM7XiNfX8N0pZrOCsuMG4j7Jna2N+3mypHZZDPECMWmNshJO1dX3tbpnSgXCxjihC4Fk0mKp8WihG7A/ULZdEd7OUHPe3NTjF+Lkup+Tbnp+ec88pnGzpmFogZTOFw0FMb77Q1Dk+Hmk9xQ/EpQz+8xh7mKlUMbeCZO4RbXaeuOvQQA4LSQSql/Ivu1dgLRw3VN8Y5uL9mKpCAllRogxx7ZLrjeqgD/3aJMT3Ffi0t3n5PY0wcg8FgMBgMBoPBYDAYDIY7AHuJYzAYDAaDwWAwGAwGg8FwB/BWcqrQd26vuaJs9VuQL0ACREeG6wtRxhZQKTx6KHpiqyNK2gI0uNfHoFHCBaCGTM5hBjrvRJnCr89FSerdW0me5qCVv3wl+vPfwjHq1bloTe22KFTvHSjjfbMBShRcmVyma+6BCn02hdQBVOuLr3V/3VeilwU9/W0erGhk23QAKFzhlmVmeh+U1G5HdWD28wKSuMuR6KwTZODuH0gSFWBILUDzPrvQ/Z5e6jyjsVwgxkONl48fP9H5S7p9xSUHrx8vx/q741NRVV++kbPPyalomB6y4MeQXuTonyu4E+z31ed1UFhrgfqK2ekpraKLyTaRF4VblHRhSlIc6LY+7pN81jdn6ouLicb8Q8zLBpwXrtEvJ28kRRudq02zheZRt64x0IIUpJ6s5pRHVwW4OVxcSpb36SefrsvPQedvQp51/9576/JuH04RTm3+zdMv1+VT0NzJFV9A8jFDd5Ey6rntU46DMHLt/mru5LHG1fVC45wOU6mn8fbqXO3WQ5b+6zFpqzrPEE4ydEQKQ0gnMRcKfzM9OColT3Q4m8Ntagl5xhh6nxwU74S05q76s4A7VYrzsLM8SGgpHWvCfWCOYb8AjXZ74psqijxzs9mKOl9gnNA9MYTTRAqnLsqPPbgpBAlo8w3RpwP2Eeb0cIjYDMr5DGtqBjcPVzq6+GifMzhSvTo+1vkgC5il7NPN7j9N9HXha43OC/VpjPvm1GpG+tsO6M3fPFd9/IpGZzvwfc81y3toIMbH7EM49oEN7eKKe6HGc6sj56kQ/daDe9fFtfrt6kp7mPlEsfAccbbf1l4kKvsuxzygI2OGdYgSqrMrxfO/+OtfrcvPjyVb5Xg96ks2nsXqz/Sl1oIKhx9l77a+qvy+Xce/qLxtGt+kkN4uEJfoJDJA23Iucj2fYp0/v9aaN4WdVw3uc3SngZrXdSB77nVX160lGu/Z8fN1OYCr2y4kCpQFODgbpXD867ZUl8sJ4nwcUSsAACAASURBVGtFLMV1rqJ/W5fomOhBuszxti3kRe5my9W8mCNmxYgx9SYdXeF2RGkxJKkjSP8uz3XO4xPNv69faP75lHw7SG86mgthCJm9v6rbEjLL18eSJP76b36zLn/6pSR8Ezzb7O7q3O/tS9JJGXjU1LjcHeh4v+Ksoz4PQ8iosbkZncP1y8ceYMu42WtUJHvc/y3gDoZntEaT9YbMH/Krg4eHKu/ur8vtuuYx00GMJpvlv7VY56wlq3if5Ir7Geq+i+ekw0O5IEeQBdUwL/MF3IlOFMcvThVrxyOtu9wz1dDv9ZrKO13FieQCz0Tz7T9rFE7udbnP88PtFG21D9fUHz1W//zFJ5oL6Ab3l59KTnx2oXagC2k70X7xwY769smh9kWPjuAiWeazePlG8/mbF0qnMYJUMoMMLmwovuwhPcqD+5Iw1yFx60MW1oArd4Z4mmAft7sjqVy/q/W9VtNczH0df3mL/PW3YUwcg8FgMBgMBoPBYDAYDIY7AHuJYzAYDAaDwWAwGAwGg8FwB/BWcirfd67RXL33abSY+VnHTGai6g5BgxuA+paC2P7JU1Fyf/mZKFeffSHq2fha9Kd9MP8+FpvNBdc6vnUhCt3gcEWLCkEZPD8Xxe3lhShi8e7eurw70P3VQzhOgNIZICN66kSj3Gvp/j4/1bXmoPDFlzp+9rloX51HD3RT7ZvzbJc+fkPgDMA39j3KJ1RewDFoDhnA7h6phJAywL1iCZZ0vS3qW8JM4qAnX4Fa/uKlxkJ4f+XOsbuj/hnBVeo1ZFPfgBI+hAyg3RWV7ehIbVwD3S2FVODqWtS+fEGKo6iSnbruwyGb/s5AlD8oXbaKonBuWcqAKAMjtZ5yE0ovAkgZDu9zXmo8PH0h+cIpaI6UFxzuqz8e7Oqe63U4Cy0UD1qlC1RegNY71bmHU9EK3/vg4brsDUQrPB+Ji/nFc7ipgLp8b0dU5EFXkoYxptEVqPBjUK0XyBTv0SGojFm52x5lNYpq7ui9j8rz4nfEm8TTfLq4Uj2H15KGvXcAKmmqWPICVN6roeZCF3PhMfqfLlThtfptfyDaaC2+cW9QfJxB4rpYqPzqRHU8H8FVDPKTna6osoco1zxKLGhtAXkUPkEEkEss4bLkQc4YvAPqv3Ml5bikgk8Rl3Z7arceXCdGFxp7dD5otTCH4IgX1lWuyKlQhwBOXY4STkjdUkirvNJ1KQ80RqZDUXwLSDTpVthpaKxFjOP4HpRjPqFbXEoaPWLAJZx1lnDfKrDeFJA8np+uxjjdCr8vfM9by2BiSPZCxAAaU9BpIoYEbLFQnZ49U0w6h1RtCgmPj7/td9SePYydOFR9JkONnWY5RigvS7FWcf27HGtcfv1S6+UVjhkcKBZwyl3DkWw41BjidT3/lj0KTcs2H7FV+M65ehlNKeXKMPenU40xH9KTKZykTi60L80p/cNe5xyS1iEkak3sEeoxpKjYY3UhBzrYWa1REeb5xYni+Os3qm+rDRnWrta2FuRZl3BfuppKXu4jzYCDTJ7rHCXr+NlBQeWqUXT7VlVZkbnhfDXOk5r6h86g/D2EDGc5h3sX9grHp4ptL96o367hHps0JcmJ4Xz3zWv1xc7Xkn8M4GLXiVex25vBkRe6EQ9yxnv35cI6yuDaCPn7CaRPb65U93t7WkfaTR0T+eqVCHK+BKkREkiuIko+3tG3fM9zzi8lqFRbci7SpSdAG0WOknLFv8f3tS/8g49+sC7v71Lyon3EeKo+ePNGe9qTE82LYgHJcekCVaBNMkSuBUb/JWJxCvfjHOPRxzxrYP7vDLR37iPWj+iIChln7DN9ic4Tw7lsXP4tr78NZOv/wg2WUssCTq9YBx4f6L5+g+f833zydF2eZZqL55CEp9hfdtCfX7URr3+s57jHTx6vy81yX3L11bP1b89e6/3A89fal77BnnoGd+IWJP8HeJ473NX8OzxU/H3/nuRXB5BNtdo6z1GuePHepcbcV2dwKhzSifW7wZg4BoPBYDAYDAaDwWAwGAx3AG/HxAmcq3dXb9rqSF7q+0iw5utLUB1f6qZXesv1Z79S0tK/+lRfqM6XesM4muv8r9/oTfQe3j7Xavpy0e0iyRC+bobZ6otiHUmiAiQtDiO9tUxxzU+PkXB1rLepw5Hq0qzrreMOEjoxiWEU6QvNZIa3l6nen02+0RtId6L289s39dnuN6ygvB1+zVim+tpzeqnEt17MLzNIpHlJJoPecM/AZJjgrbKf6H4bSLYX4+sDk4I+e61zBmXCv7ApNs9oqv55BVbCMeoeIvtxlqofvkn1hrYLhlAbb035uekCCazjrtqj1UEirAqjQOe8QlLKbSPfUJrNmEhUx4ah6tpA0sZff/F0Xf78a70tf41k0R6SfzXwleejR0os3GzrrXQbX5QvUZ9Of/mtip2NkDQVdZwhMn324qt1+YtvxLQiA2kKlkGMcfqzR/pC8wRfzoIQSSwxTnaRlPw6ULsO11/VtjcXC893hV8me/YU13JP8yCpq03CQnNr4dBuSCyX4l6On6utekg+Po/UP1+dah63mvdVxvv9wtd48aNVx8zmYOiBJXFyplj21Qt98T/GMZ8jwV8dX3f+4EhfQv/oo/fX5SbYch6Sy0b8HUl2AzI4wDQoGkhsvUXkhXPTMllt6tRWCIUuRLLfDlg2u0jUHSEJ5QSMxotTJC3GfKphndnr6+sPv/Izsd4Sc2QWr/oxw1xZoq26TcWzJbYJ1xMwaK7UnvlSN4tQ7xIwqtp1nfNiCGbKSLEmxVfEJphJrbbqNi2ToefF9lgAvu+7VmPV/mRM+DmZFPjqifs9PdOe4BqJ4j/7WskUT8ZiNo2QgLKFdv7BI83R93f1e32g8TIGo8YrkyB2W/r3EZLavnymde4crIBvwAT7FAmjw0jjJkLM/+LLz9ZlJq3mepyCkcHEjtWEx9tPDv/bKDzPZSVDOsJX6hyM0Agxh8yp4VB9NwJbJ4Bxx/sfPF6XmzXNy8tY7bjTVTuOLhTrDgYwRUiQxLW+msf1mtr24FBx4QuYCZxjD/n0UzG9Bn2Nly7ovwn2vR4/2eLreQjGbu4xoat+DzCPg7xyopuzu20h8APXqK/2836M/T7YMQG+mg+H6qsvn6u9P3+uveCbC8W+MTLfZ2Dd9QaKvw+QWLjTURy6BqN1DtbbzR4ppSwBa9L+AVjkL8XsGSEmP30ldsh8hoTYSCi7RF919jRGBiH2DEwyD6OLARLi0gwh9t4NQ9U554Jy0Hlg39AUgc8UZAK3kFz+/SOx/u/viwk6Qxz97Buxsq+hAAmYDBtJaSdgzGZLMHHmq+uGPpheYNwvMV5mYA5NsKf+688VLzOwS570tEY/6Wt8HXJfChZ7ASb1bKw1phOpDm3sAYKbttxi4v/C5S51q3r4S6o1dEwKllCAsdSHCVAfSYOHMPuZQq6R5WCKNfQ8kdQ0/64WYO6DAf7ekw/X5V5tFQs/+1xjIqyr7MHQp+lpDfMLraM5GN9jrBdnl2DKeOrbBPsE7hkOj/Re4iFiymik+frlSygNJooB3xXGxDEYDAaDwWAwGAwGg8FguAOwlzgGg8FgMBgMBoPBYDAYDHcAbyen8jzXLpPsxZBweJDB+JBEnL0R3fDFC1E/56loUz9+Igr/AImCzs5FiXv2FFIoT7TYDx7p+HsPRXlrQvKSljKaJNVv9/d07GBPlKhxAanI45+uy3kimtU//J/+bF1++rWoXfWOypRrMUGaDzpxluqe5peia42/EUV952hFv/Pz7dHGA+e5Vknz7DfUD4NOE8eonmfnoqS+PhZVtRmK7lgHTZ6EaR//164jIXBDfxuDbu3V1VYnSHL8P/+Tf+ycc27vl6KkLpF4881Q9NSHj5Q0rgl6+PBK5zs/VxuPcB3W8f4DJc26cDr+6VeS9rA/d3uiyjVrSDKYvyOqauHcTW7QHFTmIWi4E9BN44h9rfF5NdUc/eqp6P8fPFICuZ0dUVjHl2qvH32gxHK1UOfMFkwoCblGOaULUC4XoEXHGI9Q+rkfffiBzgEK8wAJjJmA++VzJR88RPLH/UNJq06QoHO3JXrlY8z7z0409v/8ky+dc86lMySe+54o8twtyvNlDjJKtE8A2WW/DfkQKLMhpA9HoBtHqGsBGRIY9q4D6Q2UkxVyPJNdZmU1F5B2zXKNOcqaHj2Q3G4Xc3GUS06bz9UPcwSPswvNuRD9XMP6wgR+bA8f4ZJJyZeQYG4Tvue7epnwdIlE1Jz6TKa609R6koAO7TAXEiZBbItOPOiI/suEt/UEEgpmfEay9jGSeRfzFS24APV3igS2vQMkLUbS5AITM4NEZQx53SLTNSnJ6PR1HzVQ26/Gum/S63d3MZaRXPnkdLUO+bcl0/094HmeC0sZTuFVR/8NYrTDoAfZHJKQzhcatwNIbAfvPVmXn3MdRcx79Eh7ofmF9kt5rmvVkfy2cPOyhrrOYqm6TDD2A0gCdrCfacE0ImlobPUhCZ6MFQd7MdZ6rCnF36APKZ24ZetSkfZscYn0fW8tIfR9Sjg0fgokk44heWx10BeQ6p9hzZtfSiLqQYrVqeuG4kC/Hx4odnW7bZRFs2+3V/u8FAnJ3/9A+5hnF5pn4Znm2atXX67LI0gbezDWCOuaQx4MK3JIonKPfYeO8TkvYWQCLYVfru9bVHA43wWuVcqd80LyhTnkngvEckp4X7zSXvD0HDLNTLFyD4lKB/tK5p00NLZ/+IHmqzfHHjFCu/nfLvv4bYdrcQPxEVKdR1gLjp4zLQESzo/1bPHRE62pB/vazzQhP/KxXvc7kvDM9jQW+0i+vd/XvnrbyMtFkA+abKMAZT4rHSL2P7wnmekMMqjnZ2qXDPubCaSuBweSYtUg6ckRJ3PM45sRP0Ng8BHPwgZMS5Dw3AMf4sND1Tfa1XkOEYNbt6xdlaT6kHEy+XIdErEenr1u5GL+Nidj4Tl3EysYtLFv4H4+Z1thH8LE63PsDxaZ+vwh9uf9Pa2Fux3Fymyutejxkcb2oKXxEpbPy7225taHH0hu1d5F4nc8eyd1zY8W4nMcaRG7OtW6PJsq5s4XGpdz7GmTSM8fHaRGYILk/a7q2Xwh6ex3hTFxDAaDwWAwGAwGg8FgMBjuAOwljsFgMBgMBoPBYDAYDAbDHcBbyam8wnPBDf0JNCjPQ1ZpSFL6O6Jc7e2K1rYzgHSnJrqknysz82xfFKY/fKhz+oFogzs7Ok8L2dYd6P2upBRTkrTfF53x5z/SOebIjt1qIWP0SDTHP/mRKFFfIWv9l2/gFnOpchsZ1y9ALV9kop3VFmrLsxeiCHZHK0pXsUW6cRwG7uHeisq1C5eoJkbCTk9UssGuyk04E83gZ9/bETXsRlbgnHNtUDwrDP+F/jaGq4WLcX60T7N0OfvZDyTfIX3t15+LEvdgT9TQCNTgewNR5TKwDbMlaZOqSxOygS5cqz7/+ot1eQD6+fv3RaE8filZUpqDr71l+CXNkfczW6rdnr8RtbjA+9odOKL8/Icfrcv7oLC+eilaH10DOodyDerB7WDQAxUf1MnFFM46peSiYJ9DNtKCq5wHqn5EiunHj3W+kc6dRzrP3/uXfrEuH8AFYJ5qnFxPRbX++Qeibu4jy/3xP/2/1uWkpOAv5tuT5BR56mbDlbQig/SwgBPIAuXQwQ1mCXo2XOCaiSiptT3QUBH/BgP9fnSk433Ik/xcbUXJzbx0J/JArQ0hM+n21W/xBO4K12q3n8LVLEG/tTzNlS5kWQniTg1uSznlqaw7xpcPTdMcLkjbhOfJGQJmCy6JWW+sYRjzlHs14BpWr3Mu6Dw+5Fc+ZHSUNOeQmp6fS0K1oFNXKUdIETxGiKkhYsT+ocZdK9b190EtH8PBbwY6cbbQNVPMvwB1r8HFcgltTZDQRQKOd1er370t0sY9z3NBKU/L4KS2hCQ1gEylCRli65Hq9gBOKo+u5CgygYPOD/qiT9P5r0VXr4GcZw67cP6rqS9uYv0MjlgFYnUTkqgW3Mt24d7WgYRyOlH/jyCh+skTSU7u9XV/n3wjaREl2B4l0g7A/+C23XKL+xvPeS4q+9FH7IywF+mgXfoDrXk1uMbRWeg1nDKffik59flcYzuHy2kXEpk9rKn9nn7fGahP47h0RfMZc9UoP/nhx+vy8M9/sy7/4LEkPxEo/69efr4uU36cIo5XXJTQTtwn5G6zhKqARPpmbSi26qBaOK+8fx/npWRzvlSMC+DWs4txXq9DytCX82Et0TkP7mstGuxrbO9i3+BjrZ3TnRYx8iYWBxg3TchgOwPNob0jSdXH2B8dDVT3BbTFCWQ1+3C48dE/mapYiVm7PcWRJaSePlwOfbhPbhfeWtoVVOgC2NNgGHbbmqNP3tN+uoN9YYS17aBPKZz+1sN+kfuU89eQQk51nhgOXmk5FwLEMMYRplBoYci3MZ/uHWg/Secv1dBB8OZcCmn6EuPaS5mcQvCwl+tDAtgp0zgEW5QZr3BT202udFWXWA/rdwrJGvcnzYb6Jwr0tx08hPZr+ttBorHqxbq3+5AHhnjWy8v98BHGTXJf6TEGba1zb0I9b+eIL0Gqee5Dfrv3RDLXbldpHmqFztlv4z1CA1L0lI5zTsfj3cURUib8rftuMCaOwWAwGAwGg8FgMBgMBsMdgL3EMRgMBoPBYDAYDAaDwWC4A3grOVXhnFsUK2rVDLzxCC4ZnYHoQLv3RfWqBaI2haBcLQpRjNIFHDYaolP1DnWeBC44Cbh4HjL7j8aip0W1lXQgSEShbQai1R/CpebiTPKTVgDKsdQHrnEkOtcDyEY+2le9XkBa9Zsr3esxHEpmdPQq1H7TK7XH6GJ1DJQN3xtJErmPHt0ry3TOAo0PEqBeSxTg9hPRQIeg6Q92RNmMA7VJUsDVBlR6HzT8EMy/HFT9C/TLRw9WFLYPDkWJXYCyP7mWDG8ffdKKkbEftMqMBgwquhTuMDNkUK+BbfrRY9Hy+g0NDJhTuPlEFFnnbZEr/lvwSp0dDUDmoDY+h7vSaKKxF4JW3UDG/g7698OHooTmoE/XQN9sttSPHiiceYYxfCxa4rSkjfL6ZH5yPne7yN7eFy360QHowXAqcBizNVK/0fynE837AP37k8eP1+UJ5KBfPZPL1bqRt2cU59Ll3J2+ero6LcaJB9eDOiQ52QSOTRiU1xe6l9ipDx8+0VgNMdJjyHM4/6ZokzrawQOdOCvp+RH0EE24JTBUJZAQdfqaKznozj4kplGuMUqXBlKPczobYL6moB5HoDbXb6nbNpGlqbssnTIog2jBgSlGPSihIpU9QYxKMOd8xE4HmncIuVEKmvcELlTpBFRktNfNFOG4yHGd6VhzZT5UPOvUNef9FJRq9Avp3EVN9z1e6PgAQbgBWcLZWHWfwVkrgoysWY5NupxsA0G5L0nnarN0pusWcDEJdFuuBulnCzK4dgsyMcg3F9g7ZVgjMc0qDo4tOJBMIZVYy9MgPaZLCqUEbciDD6Gh2vfVxtMxZJNLHTMPJVEuIP1+Bo1ERPeignR+1gdzFG4ys+Vm2cDvA89zLirrFWPMNLBuHR0pLu5A7kR55nSmcbjXU//W35eEY3KkdSkIFK8SjHmuY3SNSXHPo+vVfjGu6xyk3tOd5fE91feT62c6HnLSPq55ij7NEUlTyO5oSOXlOoay3owKaEic5OS2RWmj77mo3FBNINtP4fwX1HS9Tk1zhe6FcahYdXgkGUQDEod6W23V6MAFEfK7xRgOtnSe8bh//vYGgfGJaQbquH4T0oscjRw5jd0Ijlge5CcLOFjNkBYgxxoRoz2OkNIixH6pCSe2bcJzzgVrRzTGKD476t4OkJbhaFd90YaEugeJWj2EbD9DqopU42QM2dj4WvunRl190OthTSv3rhljPdZ0uunGaPMQWRN8rO9rC1nnXIopUtDxEVJIpv1gHbi/KTBOWtgrdsu4E3jbXReLmzrBapApEehcl2Nvv5hzT6BjDve1hw/xrJCgPxO4gzV8zb8G9qUDxMUIUvTUX53zAfacfcTe/QNI7Pb1rD7Evony0QTy28GR4v8A8q+6p/solnouubxQWorhWLGDToQe2qAbv30cNSaOwWAwGAwGg8FgMBgMBsMdgL3EMRgMBoPBYDAYDAaDwWC4A/g95FQrahmYXs7LQc93mx0zKCuJIEMJI1FSSWENc1HfEk/nDzK4ScFNIQUNLS1EjxtOVxQzfy4aFliz7uJaVKmvvpH85FEimtXHj0Unvn8gyt+HcPaYTHT9kzeiR/3jT3QfT+GO8nKsSgRLSH2u1bCzV6s65FukG/uec7V4dT06RCQx3EKQ7b8GynQNVPFWReQg7PfUPj24N9HpYAGa/AS0/dNLtVUNfPX7h6sxEoAyGYMed7Aj6mWxEJWy09XvdHvJA/IaNS6XyE5+MVRdOnRJuS863XIsqqZDVnlSKL13JqcqXFEKRAJIMugucT5U255dqrycqf1boKp6GAMNSKXiuqjFMSQyDn1AOvd4KtqgB/p/UbpvFKRlgjaeL+Fa5ovi24AUpVfT9Zdw06FDUw45ZTYDBRuSsgz95UN29+XXcjA4ORONc1bKP/ItWsXleebmk/PyvKDUYuyHdF2gDhAxow3fgxzOb42ogzLO6SueUp6RLSB36NAFSfdcBKu2pfSDc9WFlOegb3PSzHU4ZTIR4o4POnkK2vg8owsTXYTgwrJkWfe39LeohQOKQjIqusGMxxqfZyeS6u6iT31IcSiLySEtI8E2Z+hKKYnQP9BZI8Rfp5A/BSVFP8L5apDBhOgvh36k2xTp/1nGvuYcwVzH/VF2liC+OMzj03Otx204TdzIZPwtulMFvudapbQ0wxqSzjS2M8TNLILUC2shZTXNmvYqGZyPCp9rJ1zeKC1FeTHVOFouFVuLUk4VQjLRjOHqF2ien59pb9W+J+eXPuTvzab6cIk1ejrX2nF5opgY0lgS8gDnYJVzi6NJVDl+e45/znlr17KiKCq/34Djky5HPuID3XRi7JLblGs36T6HdAER12Nd99WxqPVhQFnyavx7kKpliMuNRHHuX/4TOS9++ETuKL/89a/W5auR6jJ9gfECB5UcEogCsqAAMisGnhx9R5mel7+D/U1RuKKU+yznHPvqq/6uJBkZZJptyEsGLe3b9w8lQ2tAhhHWNS8ZiwusM3PEP24dE65XxbclJz6eDxgjQkirAsxRjpU45zMP1rBM/UmJUppr3V8wXiPmt+Bs1+A2bpsa8W9hde7K+o9GpNtkF7JaRoclZChzPAstPd3zEvuyDJuT6RRyYrTXoK/x08R67Mo96ALtmS757KqxTylTNlMfcenMIFtboI4enDgdtk+MOwXkz1xruY7yOah3I6faojtVURSSSlN3WaF/YB/CS+N4xvtdtH031n1xj5og6IZ8hsKzTqcNWVaosePKtbkGGVQb+9IdSCXvv6eJMIUjHJ0vI6QdoeSVKTRC2MNdXUHaN1X5eqx5eYVnkSXm98GO3oc4h3QO/wwYE8dgMBgMBoPBYDAYDAaD4Q7AXuIYDAaDwWAwGAwGg8FgMNwBvJ2cqijWDhb5HM4UU0hS4MaUgQ4WdECna4gS5bWQWX6sd0rLc7naFJcqX5yKknoyggtHTdSqF6c6z/OXq2OShLRl1eX//kL04Oenou39MWilD97T3/Z7olZ1cpwTGbQz1PejrtrgJ3uSqFzBhcNlao8c9KtJ6eyzTTmVc25Nk01he0VSbABOXwzKYAOuNt2eqG8juATNIOdIQSefLkUfu4SEagZJ3AzuKN0uMo+XdF+PUj30TweuAlcT0fHPL9Une3Q1q9G1SmNhAhcWh7rstiHho4MFaJOkzpLN6BXvjqp64x7B/spAtZxDbsI2J8W0V1c/BoH6uhaCssnxl6u/fLhJkXp9PVI7FpACBCVFMoSbnYf2XM4gwwLFMIJcgDRnJPV3BaibS9Bf5zO430H+RvrrNY7/m+eSU40hxVGzbq8/izx3i3K+FBW6se4xi0Cxh2PTAjTpSzgDXV+pn3M4efnMeo8xSde4BLK1ekJqufqw8Mu6QYsZoSMoCfFBYaUKzQOFPICDExRULkO/UQXFrw6UdGUY6wtofUlR9/3NEtDviygM3d7uirp/ciKnPNKep5BWFbdIEhmFKTn00dacO1UTIFDxIefIclLuQcku26Kg/A3XSUnZBuU9hOQyW0BmgKalXGXJa+Z02NAxISRINcglR4gHGc7TKR1KttmffhC4Vrnm5BgzGeLNkm4hkB5mcIxhO7DdahEo+xW6O+YOBzfkaVPIA4scLpKlnNbHNi5OtFZ1W1pDX71WXLuG40+vA6dIrIs5pCs1xIgQ0icPMddDX9waIfEPlJ1tE3mRr/cgjBsR1qrra+3PMsy5qAaqPOJuFDMWwlkVLiQp2miG2HV+pWuNsLdjDPzq+JvVOSCh8nON/dZAfbT35mJdPjySE+Gf/um/ui5/DVfF0Z/9UvfxG/0t46KrrKNMh0AHGjrTcK+zfXeqPM/c7Hp1//NrtXEC2W4NjkrXM+35IkiiEsiWfMREH3IKjnnKHDPsjWeQeTP+JZThlBphH+tcDkmiw36GTj0RpNwZ9rc+pNMeYlBGORFd9Cgtv0VOTJlRjvuYLNV+20XhijKlho+YR6kK5XsJ5Ulw8k2xt8b0cyFdPLEHTpA6YY79UBtjY68veV2/q7E0vFrNEQ+CLg9jYYLnlTrllHCko5PbApZUKdI15MvN8m8PGlW/kpoE7afT/z/svcmvbNmV3rdPf6KP2777msz3siWZLJIqSuWSqmRpUCXIhg1pIkATGYZHBjzxP2BAQ08Mj6yJJx4asGG7BkIBBlyw3KhkVEeyWKzsyczXN7eJG+3ptwdx4qzfYcY18+WLZ+ES6xuQO+OdOLHPbtZe+9zv259xMUcHdS7l7VBmbIwxEzBxqAAAIABJREFUZZ2jc/1uZWKUZtLFjjJNl5JjHAWAdwGDCPEX+5ISUvkcTs6jofSbh6NbsjoWT1dwbcX+P4Y0u4e9Xb+P9xJ4d+GUkCrC4TbBvCwgMz+DDPwS8Wu2lOtnyB9c5N37cG7+ulAmjkKhUCgUCoVCoVAoFArFNYC+xFEoFAqFQqFQKBQKhUKhuAZ4aU7rhlGVQzazmgtNc4mjtn3IhFxHqGQZTvgnlXo5F+rRsy9O5fqnIj26OBPaUgFXi48fPWzKP/lQ6rYxlOgNcLo+jIQ+nAht6gVOpx58LL/z3XflC12ceB1a0OBAjzJwxHHhaHFjKG3Tx4nXM1C3ikTqefl0XYcyp9jpVeE07gKkjBaoQw6phoP3fC7KPFW+vyeOVA5pqwXeEcK1hYeNe6Di9dCfi6XQiTcuN4Ev9HByFknf7/eFYhdheCeg/pcr+c1uR8biAK4Fw57UZb6ES9FM6NE8hj7NKL2hE8/rkXBYY8ymmzxwTB1oLBzQdldwXJlDcpZ2hRJKyqsL5wU/xBgAhdUB5TwDnXwJam8Kp66ydiXgKfQBKNIV5pAHG54YJ9JXLsYs5TR0k4AUYFHIc88wpkhPLfGsnz9/IvcBBbqh5u9QHue4rglraidp+m3ZIB0BKGGSNvHQD6uUfG7Qs126jsCFKKJ0UsYCqc10qtkoBZyIYw5VbEmfqK3aXvbwm5Wh3IYuVPnWzylhoPsFu8jDc7jh65FwOI7TOGt1IAcKKTFhTCCdHPobF1IpL4RrF+8J2jbnN2VWZ6fihFWgMaI+pDYHa6ldBCp5+ULW3PMLWdPBCDcDuCcUoPNnC5lnK6zjBWS0LekYhunsUtb3S0ikE0iKOqAcL6uN48nu1sXSGjOt408US9tnGMOUhjHGV1g7LdYlF7E4pCUbmwH905IH0k0JOVIBinpZL6Q0rQtAVe934faC36cjoIs557FTsEhbSB4CSG45dule1HYy4rinRPD1/P2wqiozrx3OItRvseTaJp+nGGNxT2j2OWSeHp6HeVKG3G6CnG+1knvScfP5pXz+lx/9qCk/rp3rhsh/3jyReRYjX3bMJ015byjX/Ef/7J825W9/+92m/Nc/F3lnFHwh98mpe2AR+YrlGKCc6qtf3aVovCxKc1kfSbCayZp9fHLclFNIzyhzNJAN022R8ddBbsH1yuIpSuQNKeRaERyUOF83qnAPVcmQc+aQW0RW8uXAwXwylJkjz4Gcg5LzFPlMspRrKCG3kOqUlMchNqQ73V/8MuqGoesZcy/GTpRLOFIa5AiMqX1I0bojkaEUlIo9lXwug+T6FPJmxtr5ZN1PCzgJc/75WMNSrD8+12LsjboWbqo+fhN95CGRTeBKWbXmJSVSyN+Qu45qeR/leq8Ka6zkVldJv3EkAtuBeVgFKVYCB8oC8iEfx3WE2HOssCeOeA1kzGUqeYNT1O0cSm4zxxw+eyRj4ugNGSsduGbzmIeikHm2SmQvOENsWixw5MsFypfyu0vE3DneNSwSujkigHxNKBNHoVAoFAqFQqFQKBQKheIaQF/iKBQKhUKhUCgUCoVCoVBcA7wkv9wxtqblp6AGLZeghkGiE0Fh5FihHgUeJB+XQjGbz4V6dPpYrr98IZTAu2+915T3bgqdNDqQ777/XaFoWW9NVUoSoTjNngt96fCF0LA+fCg0dBcuEKdPhEI12cNp1qFQGOcX8vuXl1Lf8zlcJgpQJ3EKuQP5B2muy9P1fapil2RVcZUgdZYuJgmojCtQvUJQPwPKAEC7pXgohTPRHBSzBfo5xyn0XZw2TuehDRU8hhwigsxkIkOlReW2cO3JErgagW4ZBzjhPJS+5YH9Hlw4KHvJ4RTw+Km4fySgWRv7+t6TblxUHFD46bw2Gsn88EC7XKxAA1zSnQrjASf8U/IYdaS9KlAnKa90SpxEb0F/dNd0z8ATimkMpx6OtZZ7DeU3odwvo8wH5awCtZgysoU89/BI3AkyyCfOJnKyfNudo/5/szv0Bn3zt/7u3zPGGPP5x582n1egN1PCSMeiHG1i0W88kZ9OIz6kPZjGZg4qug+3PcoV6X7jO3V/IU6t4LqQOaQVQ04SIXYEco0HSmyBuEPno5z3R3/mkGjSla1F98fYpRRwl6hsZZJkPeZIG25JuUhFbpWlSAljBGlHCBcU0qdZrvj8dKqCRKNknKjp9J2erJXxDE5yHtazmZT3T2TedBCvDSSSOcZsiNicQQrnrKSOs6lQ188nIuNyEKcvXTgK1dKqsuWW8WpI08x8/MXaJWj4rffltyBTK1uSQHy5JRmi/Piq67fLyjh22J90pyzg1LEZ6RVlsIgFXbjzdDuytiWYT5TJEFe5KlZ4ELZ+0bKf43e3f75b/xSBtbaR9gZolwISMmvpXof1HPLZZYIa0h2owNqCdkwW8t3ziYzn55BTffSFOKt+8kByzY3Ra25l/vkXcI2DU+sMjpvH++LK+RgSgU4Ixxo46PQgHTd4Dg7Olm8axil7lO5U1uxeilPaykxW67zcx5oXdORZZgnqj7GaIbdZVhJ/h8iFyki+W9HtD3M0h+R8gdx1/kLyA98wn1/HpOWpSFL9DvMWiY9RKnUMISt0MP/LnHIu6fMc3y0SrpFYCwu6nMHlk7LMnBLpHbvfbuCIRL/CHKKrYl4w5rVs0poi42iE/G84FFlafyxSmEu4xsVdWaNyuLI+pHR4Rnm/U38PDlec/xjvDuridzG3IF1nfSPIwlJHYkSOvQljEOX8BeXldDTEnmjwGuRUxjrG1nlh1QraWJ+4/nEv2JLMSv3ncHiaYZ/Rj5E7uTI+U8ipSuytvvziflPOZ5BT1VJLH06NlF5NZ5JjcI968/bNphwjn1msZP5doDyHbHyGYzYu4QS8wBqxwrycIKfKkesO4Zr2daFMHIVCoVAoFAqFQqFQKBSKawB9iaNQKBQKhUKhUCgUCoVCcQ3wUnKqqqpMWstfYnw1g2PM1IHDE74b+cLFWgRw+AmEHnj2XOhul+dCVXrz/beb8lvv35PK+/JbhydCrbMGtLXVms6aLKWOi1tCOX7judCz3h2L89RqKXXseEJ9XT7FqdKx3HMGCu2ZXG6eTIQqdT6js468PytcuSfpaM6yvmaHLhzWGCOHwMN5A/TthO4+OAE/AE2PpjkF6HTVXK6/AA04w2ncZE2ezoSe1iWFHM8c1U4tfVDfyGxPV7i3I/0ZQYaQgT46e/K0KVtQ1QcDoURaSjUgZ8shI5lB/nMJqqYbQi7kvC7iuMh9XFAF9/vyzH/nb36/KY97QitNF/Js01ieIYC0qQd5VEn6fQQHD9DSlzjtP8/hGkQabd2/LsZRgX5egSo85fyLt7smlRUkVKDLp6AtJqBXr9Cne2NxungCZ50p5H2UKfi1UNBxKGd4NXS6XfMbf/M3jTHGfPqhuI5MLoWCT3e4EHTokA46iOIvQOWcoa43ScFF7E4Rh04vZS5ay2ena8e6P0MfwklH2psSomwl3/Mg7ep4kP7kcN6AhpF08iKnywfkYjmp2KQ8k4oNZwPv9TjFVZU1y9qpaQSnvkO43VWcQ5QhUlkDmTGp/XQfckBRTuECt7yUfk8YJ/HMGeLraS1hmkBiOIVLFPUTJdr84pksbtWePF8XbieRLKOmwBx1EFNIqbeIAZTsbiQKxhgTwImkt7eWkXje7tzGXM8zvcG670oqgDDOK0r/IPMgzdzC1amEe6Vn4eqHOZ1jnGcpJSJwrERsC+CstFEc0Jmt7d4B2ToW7EdwbOn0JZ6PISEwkFlSqpGgPE0YcynJsFuLXAmLZHdxlCgra6bLjZultDkURsZDn1JCnOGZC4rFIGtw0C8VnFhXdM2DdDWOJQs+GMq8/MF7Mp7dus+GA/ksQB2fn8p6UCL/OLlx1JR7kFXQKS7GGjbqQ94J6WSFYOMgf8/pKERZE7t6c/sd5jnWViatx3ynD0dSzHcHbkw+xvaCMqRM2q0/E+lZAKkE1y4HMt8ccq2LM4mLn3/yWVN+9lTkcf3Beu5MnklueXjjRlP+oCfrQol5nmBstWOuzK0E7qjpAs6iKco4OqDAWktzwCiGnAprxLNLcTDbOepx0ZLp4TlXkMEvs+0SuRAxL+phTYBro/VxvALymArxuBvJ/NqDvK6HGFjV44FjqkBub7CPWCJXDFGXDhxfDeR6ThcycsT0EkcOeJTLok/5HJWhfFbuv5ne7k73HNbYOuZZuJ3SVZHqLUotfcTQHtymLiD572EujOAa7CIvPL2Q3OazByJn/OiLX8h3MS6mm714KXV5/97dpvy3vvedplxOJf95XMj9Dm9IvKCsrTX/0P9zbGpnWI8TrBdTHDtzeir7G7qydaKXz2mUiaNQKBQKhUKhUCgUCoVCcQ2gL3EUCoVCoVAoFAqFQqFQKK4BXoq7Y601ZU15q0BfK0EfgmrKOKT+ggZe4GTrDBKKZSEUv5M7hygLPzsthCKZg4ZkPKGk5XBZSVe1DCITKl0Oumk1kd/cRx0zUFurEpKMhfx+spT7XKB8tpDnPpsLpexyKXTWvELTg/4W4bVaVVP3nB17OVR2I2tx8BkkQ6CPrUB3o8MNaXMt+cocsiVQXgPQ7VOMl3PQEz/5XKiqFejZvZrm/WwmtLoYNMXVXD6PIDM5gqzJRzmFC86jU5EHHFVCeY3RD3OeME4HC8iGlmiDFBQ6a3fv3rDBZpi7mGd/+7f/dlP+h7//D5ryi5/+uCmXOB1+kQg9MMzgZJGCtrqSazK6eYD/y6d0XbgroR1PT9e03SyVObeYC/U16shvVlb6f7mS+qb8IfRjBnpwBpkbndAKUE/jQ5FTzacPmvIqhbODI+0a1I4DBfr8VZEsl+avf/QXxhhjLk/F3WwJSSJjQwGpRFEg3oVyzRRyqvlU2qToQm5KBwE4JhQGLkFTmRd99Eu3u5YHhHAVYB0DDzEOMZFxxKPMA7E6B32YriclJQwtFxwp0l3PdbfLJXx/d/IbwnFE/nB0KOvW0Z5QtqeIM5TfFJhDOcZ8Qio1nHKWkLfOpkLzT+B2UCWQhdB9DPKkqB7PJX4zjCAPSWTsTCbyOxmltpAqHh7Ic8cYG3QHIw2c05hrSQ+yTwN3Kro1BbUrl7ND2ngUxea9d75ljDGmxNinwwbd6ihxyuFGkkL7BjWVyZdwZkK96Z5XYG3pwEmIzifniIUbh5lihbUH6ybzss/uy9rqY/6t4MJjQeV3kFutFtL/51PJfy4uhRKeZpRHbXdfo8MRc4xdoiitOZuu+yaOJW+kw2JaQrIHmVtJSSbiiY91gHkEJWp+JuUIrkVxT556jDmygpxs45rItTKCPOrkWGRTSyyA79x9sykHEZybIFE3cFntwd3JxZgtLfPA7fG1LZf6qmvjLmGtqAl7vR4+p7vZdp1eCYn3ApKxS0iUOy5kZXBv8ruQWSFCxWOpg9+XNnzy/HFTPrFr+cXxzVvNZ8Nj6e9oX8aiRZKfLqWODvIZi/Uyh4RjBclRCq1Uha3cdClj63//V/9nU94fSt3fuCf5j9t5PTJjxziNLMnBXiNHP/IZlhld8zC34OxFt0XXk/k6mUvMfnYmclGDGNjHutQJZQz4LR3buhxFcN/FGMwgraKEe4G9Dp3IHA/5JHVk2D87mOtOgb7A2lMxaWCg4N6xdsJydupOZYzdtGFLJbvdoY7GhJSED3oSN1+cyTifzXD8Bl3LkNMWvvT56ULa/CeffdiUj/ZEtnp0sI6Xt/axnxtL+ebb7zRlz5Ex9/ChHGmQTsS9jGtxgUU9R2zlcQ4ZXcUw/igXnGHdPT6QufjGW1K3rwtl4igUCoVCoVAoFAqFQqFQXAPoSxyFQqFQKBQKhUKhUCgUimuAl+KXu8aYDcvMhcOIB+pTfwzKWldu7+LU5aAn1/AUcLpndCkdsEJVWoGWnM62u+MUOehd5brCOWioz38u9OAnj+S068yAvj0SSlx/D/RwI6D8I4zlXzyfVM/tp4oXdH8hXRA9sjlc2012ST22jYNG1ZL9UCqFOpD6CPpggDIp3DyB26I98yTf+vn+vlBOK7xTvITM5ryWxCyfyCn6h5AqdAKhu7lwWFpAGmPonoGxlYC26oIqPuoK3XJlZGytXHmO+2fiTvDJg4dyT0huSO/dNTatxT5KcEr64+dCMS3gZDP24XyCuUVXshblFZRjOgiREkrHnd5Q6I8hnKVW9RxdQPLTh8yu25e+CwO4bcCpoWpRjoUKWUJ+kubyrAnkelPIBf7kw4+l/NHn8t1MxomHceptKPU7pI/PZzPzr//oj4wxxpR0N0Pbk5JaoW8TyBc6XaFqL/G8nz6V+XLUp5OQtG0CN5IR3PmG0MX24LwQ1JKcOJR+M5AscSyWqO8KEjqDsehbfheuZpD5FIyh7va2oWyKzhY+xpG7S5rxL2HTWtOpSEwi0qFBOaYKuCWhwvN7dH6bQ5YGOdVqDke4BcYPqNcxJFS9GOtu3RYF6NghYl4J56+ziayRSS7Pl5BCjPi6h9hMenUOajHbgHHfh4RqBYeKDtbajXy2qmiT82pYLpfmL370E2OMMe8cS/3dm7I++XSqakmOJZ7RnSpDH/p0OAEyrBUu7h+C7h9CwrO3d9CUN1LRwpF+ODuXvjq9kPgfQ0pz40CcNzzMFbrAMZ9azqXPJ5DiUlrbShpItb8iluXtL+wMeVmZZ5P1fCms5Hkn+xL/Sk+eIUQOydhOl1C6cnaGkPtBlvPoDFLUCd2kJB77kJTw87RZ3+DOYqReQSh9Nwb1foR+DPsiJ8jRXxbrog93NQedYeFCRXdP6/B69CNlka9BTuW6nul21u0cIjemDNilIyrmoov1gW5vdHwMMLdySHLiSuaZgdzs+FDmXPeHP2jKBfKPQS25ODq+2XzWP5DveZABlahvhbWwQGzPKTnHkRMp8uiSLkVop6eIAT/9uTjuvPf2SVO+1xW3niB8PeuiY4wJ6jHvY7z5WHMy7CNzxHPKbemaWmE9PzdwTcS6m2H/F0WSp7jY7hYz6TsvgHSydvAKIXmj25OPNZTy1xLjkTknnSXpLuohFljkLhZrAKVVrWiJ77ZUjlucwF4V1jjGbuIS5xz2r5brMCVjuKaL3MI7l3m5gKzoMmBCB0kaNsU+10XI4wdwIr5Zx8W7d2S8Hx7JOj7FPBsMcFTAWK5ZLC/kMeDCyOej+2QBl7UU+4wMx2yUGNWDkcRrBy6KP/rpR+ZloUwchUKhUCgUCoVCoVAoFIprAH2Jo1AoFAqFQqFQKBQKhUJxDfBycirXmLimmcUdUJn2hJK0fySUfMqKSCuKenJNFAk9LV0JbakbCW11Ppf7fPyJyFYmj3EKfwZHACPf7QdrmhWpmLMnoD/Ppe6zTH7n4QuhfI3vyL2jW3Kftw+FqrffFQp2CWePvalQt2LQ/MxcyiEodD3Q+MKaangx307F/iYoy8rMajeELijWdGOoIHEr8J6PErCMxmCgG9tI2ioDrcwFa5nPO8bJ78MuqIogEKY1LdYFJZOuBUWZoQzpTwWHF9D9KzxrGOCUetDaDBxuKjifpVbG/ZdPhML59IK/K7cpX5OEw3GE5m3RR2dwY/r00aOmfAjpw+0BXJcCUKMhkcvQdgXcpuhQRqemElJID+3ox3Q2WvevLUFPhWwnL4TimsBhirI/UkVzR37fSheZFJ9XcGJJ8O2HZ0Kv/vSL+1ufg/Kbytk9/d9WFSjZdAXZ/lsVpG+BK+3agTOJg7l1CueHhwsZqweOfDczkFNh/pGebXDPfCOhgzyWbjuteUYpGK7hyfweHQz53CWlnuh/SJQoP+Hnvstr6MJiXgvKqjIXs7V0YwnHptPnIre8dUMcZhy4ZFDCGJWQ1eKZl5AVtSSPpHOjPh30XdSRtTDAXLT1d/2W6xPc4bpC96Wb3yKDUx36eoV6uXScoysg5FQJfjdDPE4gHa7wVBwDDX17h/05XyzNH//JnxtjjBn97m81n1NWSveuEJT2FM4kCSQWHl3gSDNHEAsjoYFTNkPtkYv43oM8IKqlcs5Q7t2FPHUwFAnREjKoooR7HxxB8Ugm6IP6j3FpO5DfYr0oKFduOZfQwRBObDuUwhFlZc2kdsVLodm7hPRweC4yq7bcEjIFdhf6kddbyKku4LJyCRfNApKeGLKAAL/l1WsqHd7499UEdP6nT582ZcrYL+EU9saxyKyYs8WQzeAEhFYuV7XsxKQR3O3Dd8e+qWt4nmf6tfyXZlmUKjroIEsZKvITB2vqarVdqpTnMrZz7FEcSMxiSGj2B5BWnchc69RyjiCW+ezAqbFMcfwA6lLAWdXAnalCvuZjrnC9dh2JFxXqe3Yp8dfD0RXvfu+Dpnx8W1y0ljMZO7uE4xgT1VKhLmKbh3zlPJPnLyHbZzDiOp9BHkj58QxzwYHjXD6Bu26AWDSQ9c32Id0ZrNtrhfG1RE69QtnFPomuuC7iHN3PitZghiyHbm8tZ7/tLlR0TW05RtXrKCWsrwwH9Wg57W6f+S2XWlzjI4CEkHUucSzAEvlBznb2ZC700Vff/bY4Od0+lHcKR8N1+XCf+1sZ4x9++pOmbLG3CbD37sKeuA/Jv2e375Eq1N1B/5SGY1o+70P++uKZSLc+/5k4ZH1dKBNHoVAoFAqFQqFQKBQKheIaQF/iKBQKhUKhUCgUCoVCoVBcA7yUnMoaobx7OG16sC9Upn5f6HthB3TTqdCjliuhIfVA9+t1hU5qC6FinZ4KFWq1FHrSAJS4AWiMw65QuEc1/bjA/XpdoT6dncId61Roju5MaFOTmVDyZnC5igdCWx3vi5zq6dNnTTkCRSuGvMzBadY+aPFDT57D1DKeB87uSKtFWZnzyzWd2A+k/j5cebqwHCCjLwFlMXZIPQOF1JdrXoA2+ov7X6AWcv8RKN8d9iH69rgeXz7oyKRjzxNQ3FZClZ0vhVa6AMX5CI5Y4578fh8yBxcn3xcJ6dGg7kKfcXMsbRlAQlDA/efZ81OzKzjGMaG/Hh8VJGG9nsgnOnAkiuA243VRbp0ID4kApDsFnYJoMgQnMg9jpqrk8xjU2aAuW0iBLOQT6UTm+WwulHc6+EQjyAUQdzJIBDIXkh664ln53b2utI0LmiOdCAJIfZzaNcdJdkkgd2SCeS1e7NarXbgYDCAnPIZLybe//X5TpguVCSgVgJTlQtr89JG4WVUDuX88kDHlx+vPOwP0PRwDSrtdBuV3IflK4YgDCjmVhw7p1JAteIg71oF0BfGUzglFxbZ8fU5xG8nZAvKJkNLDirISQY55mSO+FpQ5weGNtN0Kus2c8wjSJltKDJzBwcPWvxXa7fTtFaQyCeSqYO2bGI4rEWJnhedO8RwJ5C2UU/GaKaR2YST3idA2Gwmss8N10Vpr0nodptxvAWl05YybchdxljJfMtkp5bMtZw/MP8gjKBFxMG7B+DY+XHO8eszRVaobwkkT7iaXCRw5L2Sehz255vAA+dRY5v+qgMvcubTNF49F2sNx3LKh+v8ZjmOMV8dqSrLPZ1LvyRLueJBqsOxgLjh0Y8ooZ5TftRjzdJDzQLm/MbrRlI/3ZSxtpDMeco4IuU5RcD7Jjz74hTgP/eLnIgl+887tpnwTriwR1rMx5hbX8fIKKQYdE9tyqt1rVB3HMX4t7eS4iiHTppTFg5SW/RbhuACLPIyyOcq9zycyR57flxx+WMjvjl3EuY7kH+5w3Z8l8n0LqXiKnOSygLQLTqHjAzhIov+5XlN6UxgcXYBrFisZ6ye3pP/fee+e3B9OoLbcnm/sAlU9bkYjHKGB/VlyjjWcAwvJgIN2rCCnCh06Fcl4fjQXqbyLMXP3TZl/+3fvNOWDYzjk1uNnOpEYUT6TeGkxdhZYT6cTWWd5KkM3kuf28YAV5HXtnAnrRMWckPkQZP5YU8viNUhUrTGbalBu67QkswTlmFJ0kcN3sOeYLZAH5PL5xUracwnJYZXLvOjjKJYeGj2tncfOkdsGOObDQ+wzGFsuXMrGI8ltTjAve/huithEaXleVFuvSZCv8eiYOZzzTm7IewTzFMfF/H9AmTgKhUKhUCgUCoVCoVAoFNcA+hJHoVAoFAqFQqFQKBQKheIa4KXkVFVlzaymAXk41T2jMw2dFOCU4PpCoZtMhE64WgqdzgflqaiEEnjrplCXb50I3SgAhdUDna4wQlVazta0uAWoktn4vCnbQqhaB6FQnz4wQqeqIF04HEl9LU7czkGVWkI65uDE+9jnKdd03OHJ1jjZvKbN7ZI2XllrFsmakpjmpGxLPZegeg1BKysNZQpSTw+ODVBtmPFI6NmLsdAK80x+6wC04g6kdQ6oh/OafupAvtQdSp+MB3LveAX5DlhzHVBubx6K/CQM4AgCat0CFOME7iwW1PWbI8hzbgnlbm8o9eGJ9P/mpz81u4I1xqT1vU8OZU4cH4p7wkc/+7gp/+63hT5qQBt0O3RzAD0X44FymXQmdNazJyIPcyFzpNSuBF15406SV+gYyGBy0NATOPJ4OJF+DAePXiAx5RQUYroQlZA0XEBq96c/+1zuj9Pk6b4VQAoW1hTgJWj5rw7bOIIxljgtSjvlXdIPd27dbMq/+f3vNeUO2mp+KXGuwnwZjmWMVJiLj5590ZQHY5m7MSSyZU1bnaN/lpOzppxAQllBJro/kvnBuNDtgpJOlzusKXRNM4g7GaQNlJoxNhVoyyX6f5dY0//X7UJ3rg4kRk7LaQKOf3hOlnOU+5At9RGPUw9SAPRjspC+sQ7lvHK9V08LOkgwzuVwR8nR13yOblfq0uvLGk0XhtUKMke6rCAuUuRGuch8InPNx7qYD9fjZ5cuHI5jTFhLG3o9mferBDJwSKvGkBiyn3OOScQel7JVrOdLSHuePnoi10CeGsM1LICcajPWKsRQm0NC9EJcL56ciTwk8aWON3oST10EzgLOkpeI7Z8+FAnVfcjGW9R5OlJd5XCXSg3oAAAgAElEQVT0OmyNzHpM5FlS/wYlSZAdwBGFf8YM4aDjYAyXaBeHkkfkdpSCdnz5j7s3Tpry97/1dlN+80Tc6nq165lzhdscnUNLNNwD9O//+n/8SVP+q9UXTXkCZ9MCdoEHfTgOVjJHpytKOuGQ1vpzL/KHpn93Nxerypqs7iPmW13kmWVBCRDyUswFz6e0A7IGK9c4kKoOcCTC5CkkObDv6iBfpJx1UbufuXC4rRDZUvzmwpFn2rsrEp9oIPGUsa1ELHbRBo6hhE/6cJWIDOPte+JCdeuGrPuLqeyHIsjCdomysmZW1+s8lDp1IM816Me9geQILUdM3LOH5487ktNUyCkryjkh3XqG4xWqZxLH6FabrtbzJZ0g/5/CQQs5RA/jK+xJ33GcYtthAuTRCdZrxhrmMdahZEnKZSs3orNj+pXn2QU28l/mpZQn2ta1DOzb87bYY57HXAX7O+y5VzMZqwn21gHc9p7PpT279TEkwxHWthCjyJMcNRpIn3TgyBjCCS/pbXeKXOJYgBVixApxYYJ95MWljJ3NewljjCkTaYM3KKf6mlAmjkKhUCgUCoVCoVAoFArFNYC+xFEoFAqFQqFQKBQKhUKhuAZ4KTlVUVkzma+pQlUsNKH+udDN9mM45YAe2IngSNMV+vF8hhOkQ0g7XLifOEKV8j2hKoWQaLlwofE8qYNT0/I8cHld3s/HCfYLoWp1QZ3OoP7oO3I9DB9Mkcr1yRIuA7g/GdWkypXgqq4sKc1rmlV5hVvNN4LjGKdutxTUTwvSouvJs/RHQu/yQSXNQHMtQIHnc3Ui6f/37wmVOJnL/ekMFEAe04WcI+zUY4eOLaAszuZCk0ymQk2LQQc+PhJac4B6FaAGL0HFncPtJYE7QZaTCgiXJyu/G8HBwC1fjyOONcaUNXXxxrFQsx/eF8eKLz//sim/fSyU2ZOuSAEGoPz7kFNZ1JuqBR9ztD+EzAbPT5lTCKeujfMFDBlMyBP7R1KvyULiggN5Q46T/2eQOiR0BQroDiZ9+uRM6Od/9ZFIzcoSbhWg1HchB23kPTs249hQVO0VEirKuwaQcLx1782mPBzI815MRdq0grRpjnJOmwFIxjpwhJvAKczvQIZTyzlCULlDyGToWmAQ87uQSgaQ2IWgD1Nu0XLhgJyhtDIuSRtm8zEGLVOhmV8upA12Cdd1Tbd27QogzaOuLwN9uojk8wTP4yVSVwfzMsaYpOSMMXDvQGSpWR/Oh1jiu4HEPbce8zlowzEaMUq2Sxst1t+I8hOznfqdo4/oFJYhdrI8B109x5o67MkYn9f9WO6QNu67jtmL1211c0/mGd2FLs9lbo06Up9BLHGCMmAHa4tnKNlE3w6k3zysV/NLyakyrJcJ3C78epzTlSSHDKrIQf3vwNUjplMc5DNLqe/ZUn7/k1+I9OCP/+Ivm/IF6kXJt9NyMqJMtCm23IXafm2vjk26RBewtiKT8RXupJQT5li36YTEareeE65IiHsdxM4JJCwLSATCWqLaxbWlQ8dJictH+yI5uXksUpzbkNd+9MXDpvwQjpgRHFwiT8bsgRQN05UZuqVqiR4ppzI7h7XWZPl6HFe5jGfK2krK+dmHjFUYexbzOEOukEHO2ImkIW7dOG7KqSPSqg5kyYfHIlUKa7e6vIKT7Uy+FyDOj+Hs2z2ElAnynBXiYImJY7G3SSGdfHYu/VzAQvRb735X6o7zBSZYj/pjyR93CWuMSWpp03PI0HvYO4SIG6sFJEzI9VcYlCPGGcgWuRaM4PCbPJHfjRfITRLsH7CP69a5YwdSyQLxlcd2BNjQBdx/0lmTznaURFFmSnkRvltxj4jxTql1nlOusx5vZbk7lyprrKmaoxXoDIr9Ada2duxnXggZdsicD/tIrFeHR7Iu7nfkKIhlX+bXuCNyqRs4ImBUy+yGQ7mHH6BegdyjO4I8ErnNxUz2Ci5ypAyyqXki5Rkct8/ghPzkXGLA+Zl8bnFkCWV5w97L82qUiaNQKBQKhUKhUCgUCoVCcQ2gL3EUCoVCoVAoFAqFQqFQKK4BXtKdypjZYk2RygOhW8eQPu15QqGCksLEA6Eh5jjVeTmX+3iOfKHXuYKGRooZTjavQJcsXcil8jV1ywXtrAQ9r/LgThRKXZwKTiZw6klWoLzKFcZGcs2qJCVOmjgvUUcyiAvQ7injqaUAVbk7zqq1xmR1u/GAfwd0vTnp26Bnd3BiuIPT/i2+6+N5S8gsSCH30T85JQSgiiegm7n9+ndx7wR0xwXkVCUcjmLQh4Mu3GvQ/zmobHQIoqtUCjrlijIAjIAM9MIMz7cqeLb+brHpsy7kKeegT/O3LyErOZPmMvtoly5cUNhfyZK0eflu0ENf53AT8OTzDtzHNpILD+4gOWQbdEXLYlBVBzJHF5D9UdJDGQNlYRnaIKMTD+jEWSrlDtrSgHa/mq7rWZW7pf67NTWdFH9KE1yU98bSlhEdwxKhaaa5tGeA79JJZY7+rEBF34OzWRev92NInsJ6DkZwJvLw714IWjHcBj04a1WIHVfZ1LQd+6QylFBVLbcE+S6lEJOpDPYIMt5doqqs0PIRC1PIM5IALk0R1gHEHw/SgYDSzgxSQUrtKE8BtZtyQh+yDGPh1pOu2zFD23JM9WLIfzLGzu0OVsVcxiDXvCWkC0mGNsCzplyb0ad0h5tD6nzjQNwFd4XAc82t/bVs5ea+SBz24cyWQb55eSk0aQ+DLwA1nss2o0aOuci+6g9FNtOSeSM/qLAeb9bdBPHULaSNCwz3fAUJDLK+JZwxqqX85jMsEv/Pn/24Kd9//Fyux5zzDOMXpAJySUta4EEm2vYn2wXWv2rhUuNC5miRz1XIOegCVFEiSrdFUO4pbXRcaTvmhRdwDZogr8rQZ/P5ely13OmQmlMGdTKS8XL7hsgJunDhiXCkQYrYweeL4FDJ+3P8Oq3xy6eCs+qmjXcoq3Icp5GY2ALSTMr27fY1krLkAvl2DtlgD7FnibjSCWXCxDgKYom5/uDF/ab87FzmQlxLsXw45viUTR1ITPGH2BhFkKgYSma4n5HLW+59iVzzxcMHTXkAWfobt+UYgQKxmErUADKyXcJap3FES5jfYOMRIi+gw9Z8BfdLzBWuGzYRyXeykngVeog/PN4D68kejlQYsN/j9ecujgeoAsipEDtacQTxghLACrlIbrg2QArOPI2ujdxr4P6Umi8gu5su1msw912vDiuyVIwZiz0OHTDR9KZlyInPI/R5jLg5m0gOMRrLGA6RIzlIes6XMi+n6P9+uM7hD/oyzw72pb+PT+TejiP5Pp3BIrj2JlhTuR+eLaQNLuCa9fhU9mBPzmSMTi9lTO2FUp+4gyNXgpffXygTR6FQKBQKhUKhUCgUCoXiGkBf4igUCoVCoVAoFAqFQqFQXAM4bVeUX3Gx47wwxnz5Ky9U7Bp3rbU7OUJe+/DfKrQfrz+0D389oP14/aF9+OsB7cfrD+3DXw9oP15/aB/+euBr9eNLvcRRKBQKhUKhUCgUCoVCoVD824HKqRQKhUKhUCgUCoVCoVAorgH0JY5CoVAoFAqFQqFQKBQKxTWAvsRRKBQKhUKhUCgUCoVCobgG0Jc4CoVCoVAoFAqFQqFQKBTXAPoSR6FQKBQKhUKhUCgUCoXiGkBf4igUCoVCoVAoFAqFQqFQXAPoSxyFQqFQKBQKhUKhUCgUimsAfYmjUCgUCoVCoVAoFAqFQnENoC9xFAqFQqFQKBQKhUKhUCiuAfQljkKhUCgUCoVCoVAoFArFNYC+xFEoFAqFQqFQKBQKhUKhuAbQlzgKhUKhUCgUCoVCoVAoFNcA+hJHoVAoFAqFQqFQKBQKheIaQF/iKBQKhUKhUCgUCoVCoVBcA+hLHIVCoVAoFAqFQqFQKBSKawB9iaNQKBQKhUKhUCgUCoVCcQ2gL3EUCoVCoVAoFAqFQqFQKK4B9CWOQqFQKBQKhUKhUCgUCsU1gL7EUSgUCoVCoVAoFAqFQqG4BtCXOAqFQqFQKBQKhUKhUCgU1wD6EkehUCgUCoVCoVAoFAqF4hpAX+IoFAqFQqFQKBQKhUKhUFwD+C9zcbfbsaPxyBhjTBjgq7aUYiXlspSyg/u4joPrq6bs4XMH3+CbJsdK2fPkX3xf6uM4/LX6dyzL8h+VrbZ+XqJeeSnlEtfglsYLAqmv5+E++K2ikO9mBT5H+7nyXROu7zmdzcxqtfrqQ30DhFHHdnuDdd1Qn7LMf+V32d5sY/YzG5rXeO72vmJbsUV5z7Jc17NCn7Bs2catGmMMoV199JUXSl0qP5OvutIeGCKm4m/lUseq2P7cjiPPvZokp9baI7MDhI5j47pNS45hXMM5ZDjnnNZFW8H2sui71vOjD1o/xbmGGnl12UObcEy56D3OeV7DeZzmGL/43Pc47uQ5UPXWPG6NTRft1Jr3675e5pVJy2onc/Hw8NDeu3fvJb6B2IM4u5zPmvL0ctKUs0zGsONu73TGABftwLFTlIhVdSO66JOA8xkdxzmcI8bx9zleGX+LcntcZrxm3w4H/aYcBvI5n5r3uf/4xc7mYjzYt72DO8YYY/Jc2jxPV/LbWdKUi3TelKtSPm8PKrQRY2prDLSjXfP5VZO61dq/Cq2W2/oxxwhjHutLuFesAb4v/cU+ZewouHbWH+d5boqi2Mlc9H3PRnVOUyBQ2HZAFdjt5avyCT6L3d6cV2JbPsNvt/+dY8X55Uu/AnvFf/ByH/cPXem3wGNs3f7cGddpX9bd3nDYlE+fPd7ZXPRc1wZ1bPK4DrRyEeaWUq7YX/aKMUC0ltfta1Rr7vJ3UZ/NXCgR8/IC+Qf7BfcIMG94DedKdUXlr5q75RVtwPbjPN6Ul1lmsnw3c3H/8NC+cffe+j8wfk6fP2vKl5MLqTOu4Zpj2x0hdeazoMb8vLUXcbe3ldvq//r/r1jb2nkT56igvZ+RvnUxjq8clxh//eGoKS8TyWmDKG7Ke/uHZhv+8sd/sbO5OBr27I2jfWOMMWUh9bCIFe325Lja3uZEqy2+xt6AMbhqxWzcc9MjzBVb9/7V+S/HLNdrzhsPsdMLoqbc7Uvf+egvc+Ua8NUqPHzwwJyfne1oXfRtFIXGmHb9WzkZ25vtdsVcrKpy6/WtRt9ebIExNw6lPbvRep3JscdOkX8yE7WtuIDktfUuQsA5yn9pt8FV7xfkm+4V/cmPs7z4WnPxpV7ijMYj85/8p//MGGPM3ZtybyeVTUS2vGzKlxdnTRlxyHSwmNtMJncPiV3sSKd0sGcKsFke9SVxPzyUoBSGcv/NBhfrY2uBXGWSaGfYFJ7PJdF+MV3IM+GaFIN6fHJD6t7fa8qzpTzf/Py8KRcPZUFKXkibFfhudfumMcaY//5//B/MrtDtDczf/4f/xBhjzMXZi+bzyZnUx3W2B55+v9eUA7zEm8+mUuc8lWvQz8PBoCkfHOzjntKHDn73ciqb0fN6HKXZsvlsNZM+WWGhSismJDKGBgNp170bx1K+K+MmOXws342eNuUCwSCZytiZPpHnXp1LaAjx3FHUaco/+p//+kuzI8Sua/6dftcYY8wkkTFc4JrQwzxD8CkjJN94h8bFMoqlv4qO9PtqJW29TKUPmGg4JYJ9Kdf369qNsDiN+1KOHVzrVrim25Qz9O9nTyS+zHO5fn8gDzXqyqK4wHvKKRaYYU+e7zgMm3K6kP6drNblf/VQYt2r4t69e+bP/uxP6/+6IlFBucKL1hIx90f/9x815f/tD/+gKX/5UMZwGEo7Mym9PJd51kW/uNisnV1IwryqXwxFPRnXt27IfO7E0vaXlxLXnp/K71SO3Nti4cxy6f+zS3m+FC9Ll4m89BgPJXb8/t/7naZ8B/XB+xxT4AXLf/Zf/IudzcXewR3z7//zPzTGGPP8yYPm8yef/HVTzh593JRffPavm/LyUj73HbxQK2Uc5uh3t8IYwIuh1os5vAItOa6s3N+pkw7beuOKl18WyYqLemGeB74kny7W66JIcT0SrVjGzBHW6/39cVMeDiXulHh5+OKFzPXNx5//4udmV4gC33z7rVvGGGMuFlhPLF9mo60wJg3e/ae59ElWyT/kSPIKJJQuN/l4u84k3+HLVeer1wT4o8TVL3G2v9Bpb3Sl6OOiPfTz3Z7MrRt4EZNm6dbyoyXWphsnTfm3fu/3m/J/+1/9853NxcBzzb39dVwYDiSeDRGXxh15ngjjNk2Q26F/s5KbMrxERr7qoY04NPgyOkBeGqF8cHhgjDFmOpWY9/S55GZ8sRJgs3LjWPqiwHr2/LnMlSTF3MUYCPhCKZBYM0U+zhf8g1CerxNh01m/PP+//uoTsyu8cfee+cM/Xq+LTiI533/33/zXTflf/sH/JHVGrj5bSE6SY60wyENitH0/4J5D2mTAMRJJ+wShzMUO/gDo17HBZ56FPGixxNhKEc/RtyOM16Mbkrd0e1Jf/kE5yzlhJbb+zj/4D5ryjz+TqXXyzgdN+Z/80/9Yvov1+M4o2NlcvHG0b/7Ff/mfG2OMmUzuN5+nqfRp6MkzdzDGel15nqgV36QdOeYTxJnlTMZDgfGcIe5mWN8WmbRjXifEVSX/PpnK/ZYpcpGVPIfPPCORMRhgE9+L5fn2uOadvNWUf/PvSt8dvyX9VXHP7Gz/Y+dmfv+j3/s9sytEUWi++8H7xhhj4lj6imtYgnmWI2ZMJpI/Z+iHBeYotuKmsvgPu/1FT4V4HfgyL9+7c9CUf/Pt9Trz9JHE0M9PJbZe4i++ZSX1jZD7O46UfazF3K/wD90LjAW2wTKRtTDL5Jk6GAv847WPhPXLh0+/1lx8qZc4URSYd+6u/+JYrKSDfCQ385V0UIwKBQhuHt5mkgkRorFiBJZhKBM6wHf38Ma5g6DLt7iBWd/TYjErkBRFWMAwl02AhNNHsIysDNgpRmCCv4b39mRAHR5LstqNsVhyZHZlYCwceY4iXD+H3b6/+0Yoi9xc1i9sJqcyyBfYsMaRBAz+xbQs5fPAk/p3IyxmXblmbywvTm4hgbt98yaukUSQC+BiLvU5rV82nV+cNp9NsLGcLWQCPb+Q7z1/IddMXsiGlol2FckEGg3lZUHYkXolAV5SBjKOQvRbOpXJypc+ppC67RKVMWZebwxybMQ4hhcIID76NOqAfcRNWSDPk2MuLrBApnO8uHFk/Pf51wQkRg42QXf21gvXt25L/986lrniW/y1Bi+ISIU6nUl7Pnwh4yHDfBr3JC6Egczjs6UsxpdIDgushF0kyS6fI9203w4no5G//Di/xKG66uoNVjN5QXL/i8+bcooFtR9jvvrykOd4mRxjzu2BzbJcydjhX66LOhle4WVmEMi/b/7yZowxCRatgo+H52C9mJIM8JKonMv4418TU8yzzz6XDf3hSOaxDbazP3YJ13FM1123dQd9t4+XpU+XMlZNIi+0HLy4KrCelPicLw8ivOg57Mrz7I+k7yq8gHt6IWN+ugBzrWnrKxiEV4xHazkWkHz4eLngcyMjY2CFROfpM/mjwXIp8XV/X9aMIV4SDPBHgPl8+ZX6vjIcx/ibXAQvcVogUxdJbIHcIiNrDX/zc5G3BHb7y5XWi/DWH1KQL6Gd/Xpsc928ik3Dl2ltJuZVjFopXmLjdFZJXOjgpWCJl4seWQRIdPnHno8/lpeXu4TruqYTr1/K86/zi7nExQE2RAU3CyvETuSu/bHEky7yzJB/KXexyXIlz0vwRz9uIskG3q9zoCP8kayD/jq9kHhRIOGfYbN6NpX1YDLHhgJ/9CDLLUD/Dnr8K7Y8X4U8uYs83WFfNxuvl2H5/QpYY/x6o8U/vjpYy3/v3/sPm/JgIP3zL//gf2nKDz/7tCkHGOdDrFf9jjx7DzlSjHYwZJ2i3wYdqY9Tb9zcAi9rMc9SvAhP8MeqJeL88kJynnPkKsc3JEc6GEu9Bh0ZZyFeUk4ey9wKS5mvyeRRU7alxFxr8Kw7RFWWZl6/YCNbPELbxnj5xJdrjGlUMKTJHGV5NlKxu9jcB5GMjQoRcYU3rXGGOVX/0R3vasywK/OyJ7c2KV7cXl7KXiNPpS4JGcVgoBgPf/iOZX4/eih/BBqe3JXL8YfUimvDFvXKlczBb4jNOkuG34ovSNsXbyu2WX3YW1i8RGHz8I97ZMSTIRdhvQyx1vbrHP7uHcklUqxV/ZXcb4pYSRKHh/yz05Ux5GOvy+fw8QebFf+4hr5ysI4XYKZ18K4hR/7wdaFn4igUCoVCoVAoFAqFQqFQXAPoSxyFQqFQKBQKhUKhUCgUimuAl+OXV5Wx2ZryFzrgPkFvakC1HPSEbkp6OKmNPeggA0oiukIPH0Cq4YGu2MdZGTGkA+2Dy2oqGB4jhN6RZ7FRJ+y3yqBTUhbGc3NS6OinQpXr3JDnGIMqbkHjrQZCp8pSaF43VFh3d7TxsszN9OyJMcaYFSRL5LLlkEGUJQ+dlLqFjrR9BxSzgwOhft65dbspv3H7VlN+E3KqXrRdG1jmQp9fHK7pxpdTufcC8rUJaMX3n4ps6mPoVB89hnRsItKGp/dBTexKvQ4g4TNG7k9KJM+T8CC9KTOef/CrD4z+JrDGmLzmxTs8P4GH/IFu6IJCzuNQeSB3BRrqCvTU5VJovqQif/cd0fK+96b0NfXznpHnv1HLXN44FMkNY0SMaOSCZkkN7aMXQj1dQk7z4ZdynhE1zGFX5muK+NI68+BS5sGog3HNM4U2UrMdSjisERawz7NJWgfwgVYMucP58ydyNc6gGA1ESjaE7rYCTXMP4/z4QM6HKhBbHz6R+0exnH82ztb3f34u9P3ZTObiaAjab+uQx+2HJvOQOwf904dsYYEzBHIsWYyhz148b8ovnkv51g2Rs2aUi+wQnueZvVrOlJxL3HC78pwPZkJldyqcDwS6O89+dtHvHUjC3j6Wvvj+Xem7m4ciPaogm/j0kZyP8cc/lnMrJsm6r32Ph1Lz7A/Mj5z9KNdHGF8ZzpbjgbLtM52oc+fZDjiHA+NhjhjPM5qymq7Me7wqXMcxYX32hTUSbzg+C9Szap2ZgvWSp2RekUPwT2c8NJ4yJILtyf7anLESepSvyvdomMDD2w3lbvgdHjzMeq1wJtNZBrn8StaL23s4hwr093CJA71BwV8sdne2GFFVVSMF7WCt4joDJbiJEM6PIMMcYDFiLBriHK4+Ym3Ykc8LI/24xDgpEPcCnFG2oeuTVn+I3PbnX4rEIoGkIcPY9HhQMeboBFLoOc4Eq1BHgzz2COcVhjgL5nAk8YWxIS1W9We71HBYU9VjiGtShec6xLpV4azEYoG1n/kZ1u0kwXzFdzMcvcL9RAfns5gScl7oErv1sQzjDmXrUj4spe4J4kiKXHsBDc90gTOOnsl6UeYyFvdHkPlgXDyCvNr0ZM9R4OxEnmPi4ayTncKR85Z5ngoUoSbA2SPsowqmDCnyP54L2LvizM1eLHOHklseSpwiRs0xHqaddexYcDGmjwV04VwDLhEXnp7JmjtZSJtXqEteyOc0o5hO5LvpCmP5ithc8TDo+v52p3NRkKaSq3Atb+0hEGPsVVJtqolbOQddMSC/w557DOnkjbH0+dvHEotvH6yvGWFNOhjIOHv2TNr12ZnEwSdTiQUvINursL8tcSYPzYAqHjnAfWxLIyafJ3hfwEPJ/db5T18PysRRKBQKhUKhUCgUCoVCobgG0Jc4CoVCoVAoFAqFQqFQKBTXAC8lp3Jdx3Sj9XsfGMmYhCct05ovpAsALIMhWejD1ssDFWsQbbeBjGlDhxOkW/Zg9GWoKXQ+KOY5fseFC0HLVQFlB+USJ0xXsHpL50I1W57C+hiuBeOb4tC0dyA0RwM5TA4nlry2zt6hmsrYqjRJ7QZS4BRtl5Ic/B6p3Dw9PobV4h4obid4rjdOhPJ6BzKAURcOD6AGupQCgSnv19TWrif0uYxUVUg4xrBp7NE1C9KLnz8WOdUMffXkE7kmCIWKVwlT0pSgU1pQfV0f4wJtU+a7o/wT1lpT1aepkyqeY555kA16nGfedvu+tOTp87Dk9OSef+cH323Kv/uD7zTlI4wBUtR90CUHdT/24KoQw7Ft0JHnCECFzuFCEITyO5dLSNUwz376i4dN+Sms6leQqHigk1vIdRaXQqmM9oSiWXq11MLZ7XvvjX2s16Keyr+TkpqCDr2aihtCB1I+H+2wSkXydIh5eRNOcaOeDO4FTur3QJkvKMur4/Ln90Ue9PgxXC9ap+7LPJ9cCiWac9uCehogzsag8ncT6Z+MMlT05xJt8/CxSMH24X63c9sG3HdD8w6sjNXFqYzD2am4RRYpLFBJ28bfVGJMog/uifztt78tEsb3YUM77pPyK/fZH4mV6dMnEvc+ergux1ivPVfqTgej2RzOjviHEG5fjpFxR2ti05I8wZ0BVGTPk3WcNO2QDmWIu5u8YpfuVI7rmnjjQuHJvCkQBxmHqpIyj+1SKZdrGyVPtHcGTZ5UajoJxfg8wH2i+rtDrKd0F8rhtkR3EErLZynsYSl3u8LWnC5btE9dwCEyxH2YYHpY33eZ07QhGo4AcZFWv33IyW5SkoJcEGY/Zg+298ORzLleX2JLF3KOIIakFPHYgROqw7lT9ynDEy2W90Zy7xI5xxSy4TM4WD3rSPn5BM43kBoxL6kg+bY4GsHH2kw75S7W781cdHc4FytrTFK7ySSI6xmkYY/ui2z+o5/+eVO+hJS2giQ8QZx1sG4MkKv1cKzByb60+cGelLuQBQXImTt1eX8kbTOEHXFApzr0YY5+uIAN+fOZzK0zSBIvU+QAc8g7YUvqruBieSF9G6VSnxlcC0c3sBfZIRzHMUHtFFTBbp2yaQ97O5drOyTiEebKoC/5SgfzKYY8MYQMnsqiCr8bQWrZxWQfdtd1SCDnLyDtypZSrxy22bEPZ1us3V1IdC5mkIWhYiHiVALX5xW/EsQAACAASURBVNlEjoAY4vkMXGQ57Ro51S6N4qxt7MHpnOQgxwiwVtDB6pfutPVTB9/1EaNhEmveviF7sffvyrEcN49k3L59U47auLm3XgM9R9o7yqXNbvVkfCxvS3705Qtp+598Ke6Zj6bi5lpVEv99rN051jbbklzBQQ3Px7w3Q87sfgMHVWXiKBQKhUKhUCgUCoVCoVBcA+hLHIVCoVAoFAqFQqFQKBSKa4CX4u44jlCorQWVCLTCGCfsh5DfdCCDikHFojtRDBeqEO+XfFzf7wm1MQK10QX9mA4Nfk3Rt6At+nQ4wfPZEI4ToBk7oEcVoApnoNONcX8H5ZCvySDFGoCKGzugopci71nV9LVd0sattaao6VseTv2mbINOFqQ9+5BVjAZCGb59Q6RSd05uNOWbcDoYdeFC5G2nUNL1wML9zK3LlES5qDtpv64n7eq6Qr0rQYlcrITidv8UJ8PD2epxIH07fAMua5gxVUGpAGQRoGqW7nbHkVeFY6zx63YpMQ5DUOuNJ+2Sg77nY14moAEucBq+A7nADz94ryn/zg++1ZRv43R4yiLjUBopJkW9llPRBY5UUp+SPiu0zAgxgtTmu7fEeciBg4/x5fof/1xkLOcvRILkYs7FcL5ZzWXezyPp96q+/64FOV7dFl9rhoPia0GBf/ZEqJ/Pn0v8eOuOyBnfhZPYeABKMiihfcrZYCFRoGl7tTzpcE/m2YehtMoFnL4MZCBTyHbOZ3BswByypLmD8jrqS6xZ5TJGE7hDgDlvTi9EDkOnMsr1dglblaZcrJ+7WAj19uLhz5vyEHNiMufaBtkY5J/v3BJ68G9/515T/s4dkcLdGCKmxpQCy3060nTmh9+Refz8fB33HMhAu5BlFpBVOGBILxL0F9wqhvguXSwyrF0FnBoomypyOO4glq0gnwiQG9hNbNrxZLR1LCwR+7KMUpPqK9ca05ZYUpId0VUK6wBlU1HMuCnlToTx35NOpMy8W8sVe7wHJAZ0LOKYgHliy0nl8TOh78/Q9pRruV25f7mQfrtcyfWHyMt6KIeQE1DqtVvYxhEkhwwlgszlCLlXH8oLumzuwdnk6FhymsFQ6Pf9gdynBzlViBx4m2zKmLYrWFEHrxUcS3z08x5kvZT3pQt5vgnkVM/Zp89lbRifSWy6XMp9ZpAsGeRdEfr9ciJxdIHrO3WCW1bMhV4NtqrMqh4f03N5ltm5PMvP/vzfNOUn93/RlOkgR+mZjxz+5ljGJKUab+3JunhjJH047GFfgjESII/a7DNiSM1C5DkepCg+9cRwProxlDhy60iumUC2SJkVXVnpWrZCmWtk4sj8TuDQenQkedRuYY2t3YrpJOgjL6ULYwGpWBdOqUMcvxEhR2y7BkOK6vBoCLhTQWfEbRldkdw6zwsLuV+GvUaCPWLqcL9COTPyDNQlRGye5xILA8aLQub09FykgZ2R5AN0n+NmdzPc7A71VNZak9fjiVIp5uREiXXdddn2iHh0p6IsC9e8A7fNf/zv/o2m/K23xQU3RnziEQ1hLRiulnC42pN1vBig3yrpn8MR5jmaOLgvMegBjgVwY6kj3wsUJd2xMS6qFGU6MUNO9Q32i8rEUSgUCoVCoVAoFAqFQqG4BtCXOAqFQqFQKBQKhUKhUCgU1wAvdxSyFRpxABp+B3QwyqxId4tAa4tA9wpBHwrBswrweYTTuDt9ugCAhkYaUrXl9HM4SbkB6Iw0sqrgWoXnM6BI55R5wN2pB+pmAtpUkQtF0AH1zYeMJMaJ62PIuGb1fbwdWjlYa0xR96EPmmAIineIo8F9PHsvpiMUZC23Rbb0xk0pDyG5ooSKzkOkWZZof7utjLZ38T3erwMK+xh0u3feEBnC+QRyi8WsKZ+CMvziAcZIKBRqmFa1HFbKlguL9O0uXRtacIwx3uYkfVBV4UhlW9Yg0v4Vygmc2jLQ+t49FinO73xPXKhO6OZBtzLMRVItQzq+1dKqEA43lEpazNsK8i8HNMsAMoPxvsybDNRWv2QskOIcLlT3z0X2Y32JL5QVOpAURMNu/e+7g2MMPLK2jxMXD5Auhcr59KlQy59ATnWwJ7Tbd98V6dseXJoicEU7cIHqQbbUxdyd4XfjWha5B0lWsZI59NkvhNqegFbawW/6iGcFWiArGQukpekwtYJUp5zCOQi05clCPn/6QiQEvTvi8rRLWFuaNF23wcWL+1LXhdDXo5hOBvL8/aFIJU72ZA595+07TfkdyAZPRtIWY/SX36XzDSSMPfmtH35fnOU+e7R2FHsKd74h7leWMoeditIXKaeIFz7W/S7uYxGPKW1bQOqyWlE2BbkC1iT+vSmv5Ve7po1vHHsoocrhauNhTLZtQVDE2A4Ql0eQZh4eyiJyuC/938G624/hyIK+HUG2tpFTRYy9oOzTPSOguwnkpmdTWa8/+1ykpx999nlTfpFjzkHylSDmpnQvwnN0UYeQrh2QBe0U1hpTSwE7Hfm9g7600QCxaNgXqcatN95oym/ce1euH0l/9frSXxH6lI5jnk+9gAyOlvga/7GRgnbggkn3GkqEikxyoHwFN9c9yVE6WKMj6MUiSI6/fCJxMcQcXUCuSqkzHe84X7OslnSXFIi9GmxVmaqODxdPxWnw7Lm4/T16InE2yRN+uSlGyAvvHEjc/I3bkruewJHqCJLU/RhrVwS5JOalD/fFTT7vQ55hWzaMiGVw7HTglEaXyRCf9woZZ4OxPNP5TMbus0vpt3Osfwnkr7NE1vHzM2nXt771jnkdsFVlyrpvLOyMiwLHDcB1kFJRynMji/0Ic0fEFg9yKpdyKuRVdMWinLflltXsI1FHQ8m/1JE5iudRLob9CM7TsHgOnrMBRZnxoV1uuVhCaua6yFdd7jU2v7PbdXEjd64oIcazJ5DG8iiOEmsks1uu8UUq/dD1JJ795ruyLv3Wd95sysfHEuc8OKsa7P+z+ToXywo4nMVwwyxkrcogYR13pZZv7Uk/z2aQmcMpdYL5RKksZWd47WA89JXF3pF74Cx5eZmxMnEUCoVCoVAoFAqFQqFQKK4B9CWOQqFQKBQKhUKhUCgUCsU1wEu7UwU1XYoyiSrYTmXrkm4KGvyAdG9QrnzQ76NIaI69rtBDO3AH8HF/jyf/gwLt1LSvwhH6EpihxoGkyMUJ8g6oWDmochFogX4qdKoOqJMxJCrzTO7jgWZV4RoPdOjBUJ51L1/TvnYppzLGNrQ4H9KU4UComf2ulHtwlTkaC5X43btCPX77DTkxvA+KW0uiBcouZVEGLigsW5zS7tRyKg80RepaKljTBHAAi0D3HoIe+zZce568ECnK5YVIbJap9GeBcWEKnkIO+QfcglAF4+6OZdyC4zom6q2p7Rlomg7a3MGP+3QnAZ01WUq9Y1B4P3j7XlO+sy9zrotrKKk0oJlWHh2yQCGs61bBGaqiVLGUuWLQ5LQeKqARC0OM0y6+sC/Xv2/EWeTZXGiUj0E/nhZw4kHcoRtatHNfqvr3mvvSKQ7/biHtgNQyhTRsAaeR7/2GUKOPD0RaFWEek85KuSTlr4xJTijlzYiHks4cHYk73TO4oVRL6c+ATnhw2SogLeI8m0NuE0L+MO7KD6cZJFRLabUZ2uYCTlikP+8SVVmaxWwt0Xz8WGQo1sgYy0t5nszSHQHzL5TYebK/15THkNsO4M7YRdmgXXy0l4P58lYga8v3f/D9db1+/JPmM8+F/AZz23XlOSoDhze442QrWQv7HUitUXfes4ILYw6aNinbK7gIFqDd57WEo9qhI05VVmZ2WdOwIaeqMGZIu3db8hm6jck9+5A23T2SGHrrhszLAzgPDXtwp8KNYlDv6cLRraWNPn4nQJz38HmnCzfMoYytt+Gkcu+m0Nb7gcz5P/30o6Z8CiexWSLxtEJshYLP3BmJFJDxZbp8Pe5Unus0MuohZDB95IpjSBg/+ODbTfnW3XtNeTCSNgojWWcM3EbaY0BiESUFzC8p+a8Q76taIsLhTCe3jFR9H2u6B6kfJHJOawxADs8jEPryfA+ei+zzdIJcF/eh61mB+ZGmy/oZdhdbq7I0i/N1nJmdS90uz6ScYc3O8dMhVs9jzKcPTqTPb/fl8z4ULjH6ivsJB25KxsOexqPbYe1eyXyPjrnIXStoLLBdau9F0PZdSJ7DHvIfOL85kDnmVmJ0Ukk7Mac9OxPnowJr0C5hrTVFfeSEg4QuRSzvBrKGdZB7eXSMgt4owFEA7KOr5FQU8rQcePF5hbm2SXB4FAG2EcZFvuK13LEgA4ZcroCEMeB34U/n4HqDvljNxXEuQ6wNAllTKcWqNvLBncqpRL7sXrEPLbCXpQMXpUoe5gKdhR0cQzJEUvnGscSnQQdjvoO8FNJWF/FvEwsL7Bt4/IaXoh+wn/eQow1iqeOdodTLFpJDfX4m1y8gMa2Qx5awgqTjpIuxw31k9Q1kqcrEUSgUCoVCoVAoFAqFQqG4BtCXOAqFQqFQKBQKhUKhUCgU1wAvJadyXdf0ums6U4kT80kNClzyuyA3AeWKtEFSbDuQTbk4ybk7ECqwTycsUNgMJV00tajpWqS+ESSIVS6p0zxVHG5aoC32e6DZggU1h/PCCvRjJ4G8YAg6LU7R73bl+caLddlzdvmuzTFO3RYxTtcfw23mALIp0r3vnIgM6S3IqUZwsglALW83OeQuoC+6kFZRZuFU5VfKFie3t12f0G8Ycy7u4cMbYq8vNNS7N0UK8uCR0BdJbe4NpZ9XrlBVU9AFTYo6ZJDnlLuj/BOe55peLb1zM2nojJIRSHFIW8whN7HgIh/vCYXx3om4edFZzm/RIum4ANkU5iVpjqamJSegJ3ugDZegqtuC3HLpxxLPWpWUf0gf+UNp/z3Mnbcwfn/2QGQ/vzgH/RiSxy5kKWVNIbc7l+TY+n+dX/pk8x/SDpSwRrHMuTfffKsp37ktrkZdSC+CCH3SktzBQYBOgZAfDAJp27RxqpEx1HIMhMNEci7zyWm5jckTljmlVZAQ0O0P7gdDOM4lmGdQBZplCsnHQmiuc7hs7RJFXpjzx+vxtDgXmvoALPzJXBzxDOVkWCv2OuLIcNCBwwzcGYMAsjj0kUunFNDvfUgXnY7E8u99by2nOp9Jm1yeS5n055D3hrTnDP27nMt381So38OxyIgOIRFb4bnncIGj4whZ4XPcf8M43qWEgy4clGPThYrzo+Vih/tQgnLrWGRTb9yUdtgbQtpDaUcH/exSTiVtQqnzRmYIdaRxEe8ijJVeFw6YY5E4dftSx6Mx1mKoneZzkdP95IE4BJ2Brp4hbuZc3yntwbqQZq9HwuE6TuNQ2cM82B/L2Lt9S2RjeweyJvhYzyrEJetiPYdEmSu7U0IiQEdGuBI5lGigf5tvwnnHw7gLuZ5SOu1jDEI25cBlxcF61UG9hseyTuwfi7Txrz+U52buip8yHcT4snbKcd3dyeOqqjJJLc98dF9cqJYzGYclpEGGLohY2948lnh3KGm1iWg8SncfzO8Sf98uIc9xne190ch/EB/pRssNQtBac+UKS2eflmMSXGQhJwqR8wTof0qO7CNI0CDVoUzNFK9JTlWVJqnXXDreBliTvCtchS1isI8+DTzOJ5TpSIVGda6SU9GtFXXeuMJV2GvwqAQHUhkXscCDRA7bW+O0pHM49sFI3UscI5AnshYuJpKXrhYy9t2uxOwWD8PZ5JK7zFFtIyluHcmA3IvOdHHriBO4LcKxkmC+vQep4HiA3IbHNkDa6sTyXiDA0Qphvb9IcW8LnWOFMWRdyVUo2Q4wj1kvB3M+wdh6/FAcWimJohy7hHNwa99Dp6pvsL9QJo5CoVAoFAqFQqFQKBQKxTWAvsRRKBQKhUKhUCgUCoVCobgGeCk5lXGchipq4XBD+hplUzncmyJDaRWkSqBcxZAvGNCWgpY+CqeKg3KYgebWvn79W663/VEt5S6gU9E9ym37xcjvgMIVd0CFXQhFy6ANyhUocWARumgDHzTaYS1xIi3tVeF6nun21tRuyql6HXmWw5FQv28fi9zo1onQ+PqQJIWk+INbXkIKUllSD0G3NnQCAW2Rla7HF2UYFlRVi9PdnRy/g9PgyWz1QWs82JfnPj6UcZOcg2KaCi2bw6XEb9kMFL0Edctos7Q7WMcxeX0SfAEXl4ROWqDh+vg4hWNB5csD3T6S/j0EhbA0lLyAHmgpeaPjGJwsIK/zNrRndG6V4yR/0LcryLwoBbOgKpIiSbmOi/nvd+X3Tw6kr2+PhGr95VTm5QrPainvKnc3B7fBaZURhxhvIrj0dUXC9PZbd5vyGPPSj0ADv8I9pT3RcJo/OMExrre1RI9yCLoWLJYS+yYX57g3nerkfgn6LUU/k9o6B3W+C6eCA7iP0B1gFctvXUBCdXop9ORdoigy8+LFI2OMMZERWYGF3GS+pNOSXNODk9M+HAJj9B1lHsannII0ezofSDx24TIUU1p4c+0o+OStbzWffeF+2ZRnoG9HHciZA1KLpd/PERfoGkUXEMqPx/tyzxWcIEkF529ROZXlq/qzXcqpKpMV9X0ZYyDVKOge48D5CZKMQ9jd3DmCXLkjdY09uHlApxJC/hNBQhXCqYpyNrf+3IUM3Efs60JyOYYkfYRy3JNyGMi68M574jj55Oy9pvzlhcyh3lwo5Jh+pmpJqvFMcFNycjgR7hDWWpPXsYPzZtyTthj14CoEKUOF2OVUkKTgGegGQylM6XD8y+dBS0MlxcpSxOFuuQfKJSXnlPEhz0Ae5aKdA5/rhzwHHcRuwsXw7FDkZQ9On8hFlFZButIfrseP54mz56vCloVZTdcSr+VM1pAcaw7XBxjJmNtY/272JPZ1KIOq2J5MxPk3bTrYoA1RpmuSVztyWqYJdGHF/bjkumb7PsrDWKDbKcdN4EMuOKA7j4x7uuPOjfTRxanIqWYXMo93CWutyWvZbBRJnOFxGtwLOIi7LiVXiCcO1kXDfRFzQbQ716jWlEObtvYV9UU0jKJEtWoNEWfrNW0VHY+OoPObfHcGK6wKgdQtZQ3O4ATZO2AMkPtvttW79+Dc/AblVDIXfTq54VsdxBuv5dIHyS/KxxGcOiFhzdCflDbSBSzgUKgdwbrIN6pC5MQ5pHI+4n8AJzHKnSLskwfYH5wUdP6DbBh5bIlYk8L9isdMUKJpEbvN19w6KhNHoVAoFAqFQqFQKBQKheIaQF/iKBQKhUKhUCgUCoVCoVBcA7ycnMrI6d0kbfmg06eZUIacHNRyUMkK8E3jntCceMp1N4bzE38fp5Y7oBaTb0069+bEcxfSGtLg6AjQ4kKiTKcshzw7Zzv9ilIE/lSWCnWrpPwLtEA6UW1oXJQrvDIcce/qdOV3x0Ohnp4ciJzqDhw2juE0sg+3ixiU/bIkNVHocSW7ChKKytneLwG5yht3KtyE3cY3kRVdHzxQVVtlqeMAp6C/eVekY0+m4oowvRAXFjuAVAA0SJ48Tplfbl96in0tVMYxq5q8WOKZ85yyIrgtZPI5qZA9uBa9CTlVTMrrFZTftj8YHKRAf0xKuNjV/e7RQQxUzKqiVIpUcfwO6bH4nPXyQX8PPKlLH2P2AJLB+Nmp1BcxqyopC1mXdzgTr0T7uSilgJNNX57Fg6tYDOcQuoq5Zjvdt93OLKMv8NBBPaZS0Nnp2BBx2sK9zWKM7o1ECrbMxbWpTKS950uheFtPrk/RJ3uQSOzBkaqApnN2KXP38lLoyTuFrYwpaxcOyBpWK5FnZDlp1fLVbij9NYJDIB2JGOCo6mvJLxxSrAGyz7FGjWsHwrfeFNnMciFt7oPufQnZTH8gfWExvvh8OZwr6Dji060CLkuUWVGuGsGVpAu572bMzma7609rjMlreYRFLKfcgRJeD3lAB/Tskz1pnwEkqZQNeHRJoeSVrp2QzdF5xt3q7EmJFd3LIM+j8yfu3ZI5o89HY7n+e7/x7ab86aPHTfnZRObuLJDYRImsi5yug4HvlZCc7xDWiNsMpScce3TBLCA3KUBrrxzJC3IP4xkSP8o82I+UcDhX5ZT8rc0YowwEz9SSczjUeaAY8H50xOE6LqDU3cMa0I3luQsejUA3Wv+rbmlXub9+ExRFZs5O1+OMcpE5HP58zMVxLGPvGJLUHh6Y61I7/WdbYb0MKKeDhJzX0xWunjsO9goGew7b6lu6U6F/KB9lP/M3+V2sIwHmNGUsYzjHjpZShwWkxU8fPDCvA461xq/nWgnXJT+k26/sHXiURIC1ihJilkOOaLY1nW2d7fligTGPVKbZm9LZNmBDI4+xmP80+OJvMgf30KV+RXmf3GeVMe+ROZcUEi8tdDatcFC3gbNDmbHjOI0rFffnBZyf6FrlI2+JO7IWhdhnRMhRc0hY+wXk7laevcDxDwX2E3kuY6rlVFlunGQpz2s9FYoMDHRwhdwpkt/0ETe7iDtd5Oase8tsCsmNSztr7Jm5Hn9dKBNHoVAoFAqFQqFQKBQKheIaQF/iKBQKhUKhUCgUCoVCoVBcA7yU1sMxjvHqE/9LuHCU4JI5oDtZSKt4qn7gCfXIxwnvdL4JQOtsUZ5AVfUMJSxyiUu+ZM2jsuS7VdtFEQ7oVA5oyT7oqZSOeaCUOdkVzkoAKWikRdJ9io47QU07o1TkVeEaazq1O8atfaF+v3/3pCm/fftWU755dNiU+zjtP46kf+jU0aL1gxnGNikpNyKtDPRIF05VXk0fBKm55VZCqQalH2Yr9bwt1epiLJ4ciZxqOBZXhAs8iO8JPdUDXa/EqeIr0CAtaXY7hLXGZPU4bknYeA1aPaPbBU54p2vHyZ6UQ48UXrqJ4XP0NZilxnO20wYbmjnmPGmOHqc5qM20BLCg/HutuCDFErK7yttO4xxCxjmIZVwniFktWUhU00R3SFVdY1u0oHxNyqz/YChz14cEjI4+bJSraJ28xiNtuWI/f7XNAzi/ROj8fgRXCdBvc7jRDA+l7qczuGAVElOSTGi2Caj8s0TKI/ZhV/owA/02BJ18NZd22ilsZZyilvYk4gCSrcQZiy5jHpw3IsSfDiTEHboXcGJQtkEJ7xUyj6ucUOJaxkVHumQu0qosxTygaxzmXA/Ob4MB3MFAkebaxnU5BuW/25G+yyHLoqQo7sjYWNXU/F3KjCtrzbKe7y3JAucQafr4fNyRMXY4lHIAanmGfnYD6ecU0k+6KXqI0ZTWuQHzhnWf+5jcHh0hIXNcJtInOSj+EfowwFrl4HcOj2Se/Y3v3mvKf/nRX8n1dPTCGA3hSteBvDpsxaDdwVpjiuqrth4tqSglpFwXIdUvC6zhhTwDHZ48uAO1JPSU1KAfHTiHtcqb8ca6cN3EPKCEkbkWJc+sO9dFv7VGShvlkLowl0/glOMgL+1AAtML13Vr5dyvCmuNLde/fTk5az5ezEU+GSDH6CJXgDmc6SA+BMxPWrESZeQKIeaCD6mg63OPgJtu5FfcW7SkVdslOa2kx9ueF1Ee15JaYxyXJfYf6KuQN8K4mE+lXT/52Y/N64C1tpErxmi3TiCxnPlCB/JZH07FDtaQAs9D2T5zUddn/MbeAHkyHYdbMbPup5wrJ8YF95wWkigfMbVVDiBB4jESJfsasinIwnMr8+9yKvuRIaTmoSdzcZNj2GqHOaqVlHeFeMD5wTWe+WeMPXzcCfG5lHP0ob/k2rbdqc3CybLEmpYZuB3W49+pIJXlewlyVzyueTiGoQv35WL7d+MV8hnm3Zy7lNbRZQ55TgU3q8pROZVCoVAoFAqFQqFQKBQKxa8l9CWOQqFQKBQKhUKhUCgUCsU1wEvJqaxpmEo8SN+4oG+RghmAfhySegpK9nIptO3R0VCu8enOATppizYKOh3oV3Q/2lDoKpy0H/DedLsBDZiMNAd1p8tECGptBpp/W63FE+pBf6zK7dfgU3dDU9shUzX0PfPm/tqZ54fvv9N8/hvvvd2Ux+P9phyE2yn7BSi+KajUFzOh+p2fv2jK86nIDOh0YCgfA4V4DGePcW/dX1FLykJqopRL3C/D+EhyykOk/0NwjLuxfD4eyVhIIQUxaI+y4InkMo4jPEeEsSseSK8Oa42x2bru2Qp0Px5XjzlBKqcBbXU0FmlLv4c5Cpqrg3HuhEIzpOTQoyuKJ9d4KDu1BMeCWhtQ+oG+a502j88d56tze/2bdOeBjBP01A441XsD1B30e7rsrEBRzjcOaWbX+BWTG3TMltMIYpgLOivlNrkLCSkow5SEeRifId0xyu1PurmeUhE6DwwGQu8dwclo8VzowD5i6HhACRXdDGXMLTG+XVfWlwP8Vo8UXbgAjjoiBcmzr8osdgFbFiadrZ0V/FTqx7+Q+HAKovTApyyVchZI/KJYJDotGQYkbQHnn7+d/u+2aOFr7PdlzXXevNeUV4m01eRnH8rnK7hPYjwOhrgPxyzGFyVlASj1PTgnFa1xR6dBULY37bRDCYe11pR1nCdNv6wYV+CeCUnJrRsiSRtizE8h33u+EGmd5wsN/ORInuFoH1IZzMUQaVqI+Z3VkmO3pdSQMTGDnG9+inyjg/XMlzquMM+KQsbxfk9+883bIrt+80Tkxx9/+VDqfoUELWjJrF7X3w+tOHyQ1t6Swm0fN1xzOAwD0Ox9uMn5WEe51lLmaDHn2nNRvtuMsYpy1v+3vTdrkivLrvTOHf36GB4DAoE5p6rMrIFks0iTtcwkmUkyk970E/Uj9AMkU7NFU1uLZJE15JyJBBBAjB4ePt5ZD37D1+ddADtBOLotaHs9HThu3OEM++zrvtZeODWlxRsKYuS/dDihdBPzN4dEls4tWar5OJvR8U3niZEPDHqQfTbzdJtyKs9zLmm66mqkfSPLX+8SE2MBtILXywMD7A8xYu6GnAprOgjVjuBUw3gWooxA2JRioLR/U+35eqnUf2JDtm7WGxaSXCt0zcI7D/OlDbUW5USaL+lc3ShlTAAAIABJREFUMeDLP/yjey/wfOeaON+JIdsf6P2i12I+SRcq3XeB94tsovvO09fHaboltSHpaVEWx7IY6N/yRvGP4g0bzmt0mMLHNNPyUFohxdqawAnPcyjLQCdCzLvZDBLYqf72+Ee5ie0fSgK9bGQ/Rbm9PKeqK7dcrq5Npy2OSYFYkuIZKWFdpnSSgmQa7+UHCZ2ttI+2IVdOmOsiPy/oKntTigPnLpH7cWnR8bV0atc5XMVKjYmHuRKFcPijmx3mLu1E+R0EbcXoFFj/C8o1GBPHYDAYDAaDwWAwGAwGg+EWwL7EMRgMBoPBYDAYDAaDwWC4BXhLdyrnoqbavgc6/5JyFlCG2qDcs8L7Ai4AHVK8N6hNAiv/kzbnQOmKWLUfNKuyuVZAdwL8/wZ9ySetDjRq8OZquA1sUmj1rAVvEc8R4pgAdFrSz32fkgn/T/7/XZG0YvfzDx8755z77CNJqO7sixLub4yDHqYA7Wt0IanUN0+P1+2TE4mGpo3EwDnnJleixeYL0QQLSCgqVAG/c7Czbv/ys58755z7EE4qrF5OOdfJhWRb3x/rHk9GcI0B9Thuaxx6XY1tpwvK7VS0dFbHL0iJBgU/Aj2Szj3bRF1VbtFI+Cgb23BsoWMIJIwh+m4IajRZ25RtOF99UfqiFlds06HMp5wjwTHNWudapWRiw1VO1688uhCAFo9jQh9uFaCZlzlccyCdOzqQzObOUO1vzuTakGIcF01fVlt3p3oLYAwZE5dLSCJA00w6dLl6vUxlY3rWb3KM4Xg1n4SUtyiGD4aS1bThZkfXPQ/yNbr5ZEvd4wSTcTYSxX861XjOIP+4c1fXjThHArpcYE5vEVVdrZ2oAlJm8RtJDHowXRiTvu57jHF8ORqv2xHcm4oAa70N10RsOj7meUBpE+dPEycGO7s6Fu5Yf9kf6r5S9fO//w//QefD5PFBJw4ZUxz7n3I9PUcCudiGipLxADRqr4nHb5LF/ItQ1a5crvqTMYkOQC0cznk73NH4nI8VPxzcS/Z21J9ppvn5/NmP6/YrtO8eaF48uS+HSEqYlo30JcS8vnPnns734lTtK+3Fo6nGM+nq3EeHcqXcRUy8eqU1R/njp59Ijv03//j7dTsHpb7CfO304ESTIY/aIjzPuaiZ23RSDDF/KLPxIYmqsJ9M4TZ6fqY84hqyMbpsJl3N+e5QucsA4767q7XWblFytrrPAnk0FRzM/2rEzgr7fk03QYRxH/lqBffV2ocsBXJxlkNgLkE5VbsNedE679nevuhDTjW5VhzM4OLiv8mNCJKPDPFuAb1LinVZpTrPHBKazNMzDkONbRIyCjD/WH1OqbiH/suR2y7gepmhTZkJrV2jWM/UhctPC9I+yugDxFbKRRgvUzj7hHNK6LaIIHBBfyWd2t1RnBlAKkypUgm5O3PaGvGy31VcDLDPTBeaw6+wXl+e6b2j21F/7fTgxAq53Pp9h++0mNsZ1t8STm6XU83Tl6eKu5coI+HjHbjb0fkL5NFJF3kC5D1Jos/x+rohh/c33ja3g7qqXdrs/5Q0U57K0iBzOFMS1UapBD17G9Lvxx8c6px8Xqy/mnJyvucjzy8baZUHWb3P6yN20G0zhVRxnGsdvzzXuU+PX+n4Ed4tuOfRWZA5F2Jxgfvhdw3Bm1LwfwbGxDEYDAaDwWAwGAwGg8FguAWwL3EMBoPBYDAYDAaDwWAwGG4B3kpO5era1Q0tsAXq2xL0vQ12PihpKeip3bYocXTKmYKKVYJ61No4K6QAZFtTOgJqaZ6S3r5ChvsF28n5kH8VHqv36/gZqIdjVM5fzESFzyHXoctBSKeODScEusWganojWdgmbTwMQ3dwd0VtHOyIssaK/SVoc5SeLeE8NbsQPXuQgDb++K6OX4g2ObqEm9W1JFdBQTqgKI4RXJASt7qHClyzHFTGc9zLH774ft3+/ljU9jncqUiny0rRWVuojj4KNeZjup05XZfzLwx1Hq8lal3Zfj/ym6qu3aKhDdJ5KvbhgAHZRhxQQgOXGKzR8bn60cO68SNRRVtw+9nbk7ztEHI81wZ1kTTrhvaPrt2gFeaQ1i3gkpFBzlEWihFeJTpj6GEcY9CrQd2saj1TK9R1H9wRzd37TnT5JWiii7xozvFfWk7lvaa1SQHOEFufnrxct3voh11Q/LtduB2BHV4hWMG0b6N9M9XKlJJIXb830HWGB+rX3rnWIqnwB7s6Pl9gbUEuNoOD4TVcACdT3UO2p2dqg/rfhaQrL97P2K0ox6t7DzBnfJ/Ueh1fbkgbFf9OQMNOsG5GcDaiFLEHOvmTh5LR/PyDh+v2PsYjplNOa9UvpEjPIX89v1SMboOGHkS6/umlxjRI6FQH+j/2VA9rZwkZwY37hXOb9HpKsCPkCetjtqimcs5zrmwcZt7gFtFCXx1CblSX2hfjlm7q48eSG/UhVWM+cXamPp9hnGu4gF2+1Lzoon/2GukiJTsB7tGHa8+dXe31B/uQFQxEYf/w0ZN1+/EDzacsU3/88Yu/w/W1tijLvbjQvWfoyx7lBHAn2yZ857mkkfklmHt0yqPbECVGJ2NJHyYzzc94CCkc9q6zS43d2VffrttzzO1eR8/5+Scfrtu/QLu7zjvgQoZ7zzHRKSnKEM/miJ1jPMeMUnBIDrKFjrmaql1ALsJoSfcVOtMsy9Xaraotxta6ci5fxXaWZ2AphXqjqgEcGVGq4exK67KGhL+G9DsvKUnT3/bams+P7sqR7ehAsqBdXNdrxsXDe1GJfhpDBnyKvfD45cm6vUy1bigR2z/Q/LuH/fIOPu8NdL8Ocb6FuBPjmRYL5Qne7PUSmHeH7+pGTh/BnTHccFjU0cwzmCO0Ibdl3ChQwiDG+9puX9eawymqyLQWZhONh+spllc375oZXN3wRHPMl0uss/NT7ZeUfz3oyYkraSlezvHgU6zvHG5aUaDnaHV1j7NSxy+RG9U37zJvlMW/Paq6cotG5r3hOol2AelbBmkmj6mqDY3nurmMFCufnapPfvedpMUp9o3HRzoPlL2uquDQWzduWphDRYk9Ce3xTOf79qnG85uvtT7meO8ddnQvHey73VR/Gzq9L5UslYK9OS/fJH1jUvPTYqoxcQwGg8FgMBgMBoPBYDAYbgHsSxyDwWAwGAwGg8FgMBgMhluAt5RTVa4sVjQ0H5Stm2rQzm1KOFgmv4WK9iEquF9eSJJEKlFnR7TBqCM6XUEpVK52r6Xz7/b/lEJeofJ0Ctoi6XzjK1HiTq5EiZotSBEDRQsyjws8xzUqvzs4UnVRiZv00wxVtLvopygKm2tuV061t7eSObRB3QtA8S/x3V6AdjfSuLX31MdT0IevLtUPp6OrdfvsVJ/XC9DTApEV6RK0u6d2d2dFK4RKwo2nOsfkVNS3LiQzH+1LtudjfiQdPUcOGt6LC1S4B4Xdj0RrpMvIgJRwUOpnoNkuQ9LmONffDXVduzJf3fsGeXKDwgjHLNB8E8gdepDW7N+9h+PVXxdj0VCfnsmJLPzxxbr92UeP1+2H90U/7vbUv368opPSVWNyrTly/FLnO4GrQA23q25f97s7VP9325CugEJeLiTLSslOhFtEB1TxGLGpgKShaGja/8W9qeh0B6o4qdoHoHhnifrqh2++Wre/+fbpun33UMd/8rFo/R3EHudTWqX2dLKijf7w1Zfrz8ZTjdXRIzncPH6iOXGOWED5zLAHudOu2nRnqSo4NiCOzCE/GY21Rv0uYhZkPpTDbBWe57ybc/MaGw4bagcY01242nz+4Qfr9gePHq3bIdbiBHLe0zNR8WdffLFutzzFwPiJJDID8NWTRto7nep8o1Otv+PvNV9ySIX/7GeSCD19qXH5+gcdP8F+VvXpFPb6/k/pVog+i+G+0qODQ9PeppqqrmtXVPnNP9af+3C64355Z6g4NGipfw6wthz2+DO43s3ndAaidEt/GkASQQebKdxO7u6vqPoP70rCHGCuBJDQUjJzdqnc5uz42bodQrriLbWf3YEr5OP7lBbpmKNdyQYmkNnRZbCuMP/c+3GK8zznWs3yP9jR/n+wp/uLkAvQwcjvKOfYhzvYDDPt/JVyxDHS5yzR8ctScWlyCanS7/6wbvcwr+41eUoLsaqkQyj6bVnqfn88lZzrt19IzvWCeyfygSHk4p1Ac9ZBApHmkIVD0sp9bzLX8/WT4E/+/51RV841kgdKrLne+etzTEfXWu1WrP0khpyw1dLaDVhCodL8zxHDpni/Ob2U85Hn/tRBCsO24eCTpchJkHvQjbYF16Yu5InDoWTJCebuAhK6IIbjXAcynBZlxprfZa4RO0ds2ibqunZl4yBZI3dxMZ1P4RSEY/Ba4Ar012iOfRT9myPWlngvC5EnxVjrMZwVY2zIN6U2aoxtCQn37Br55LXG9A5iRw/vLryXa8jWJnAzSlE6ooZ0rERmXyFeUJbDmHHjnMa9411R18o1KS0ONpyqUOYixxiyfAilkBvuVLrXLvbX45daZz/8qHeOOwPF2V9+rFzzwyd6d7lJG2rsMculrnn8Srno7/6ovOXpK+2LYaC1+OFD7a/9tp7pao6SE3juNp41peMc+iPI3xQxGdl+mtuYMXEMBoPBYDAYDAaDwWAwGG4B7Escg8FgMBgMBoPBYDAYDIZbgLfiXdV15fKmwv0SFOgcdEO6K3VQGZ3V3pegp52jUnsI6ZEHatXSY2V8SERAvfbhVhSBNr67s6Iisro+aV50m7ociZ56PUMlfxA5W6R4oyJ8DnrUFfqG12q1RdGiRIryriSGg8h7cOEIfN8NmvsI8R2ej3bEav9gdFWgW1+PNG7fPRcl+zs4Qn3zXJS4qyko513Q0kFl7Lb07EMPVdqj1f3GMWQVoHKfnYkGN5+jGn2osaKDzgx/G0Wi2R2CcvubSA4vnR1R5O89klQooUQwUZ+dTtUH3x+LCvj3ThXX3xVeXTu/kRNSbuHTzQgSBIdK/gEo7gHoud++koTs2YXmcIo5UGJCHIKu3j5Tn5awjvjgicax18jS5pBBnk90ncuFzn1Vai6cXYm2Oj8WVZzU5XuHkvd9vA9aMhyC5ktU8gcBfMPlDv0HIxJRSbetp7o5n/e6Dzed9haQEF6PNedPnksGQwnVH7/9bt2m883Dozvrtoe4+OFj0VN7bXVuDZr5dLSSE6S4fgm5Bx2m7h2J4np0T3PrBPMshOTjzr7GkKrckm5Bge59hLlzMZJsoROAztwBLfon0lPfFp7nubBxT/IhVayXeggPD9Rra008ONTztEFRfvVccWMGSraPPfLgQDKXNuLYGG59p7ifQKFrzWLPZqIWtz3N+KMdyCzh3jEv4JoygCMdPv/t736/bl9m2gMoo46xv+ep9to5KOdLPCulCev1sUWnuCAM3O7B6jnzC8xtjFu3q1jZh2PXAT6vQPd/Nn61bj8f6xlTSOtu8hPnnGtDrpxPNYa9CHMHUu2yoWrvUHoea9y+f6Hr/8Mfvlm3j+GI1etp/pWxpFIuVqyZTzRH7t3XtR5AxnW0p/N8+YXGfA7H0SRGjuYxum4Pge+5fpPfHO5LQtXdyL10fILPY8i5v30p2dT/+0dJFf/unxRfryGL8yCV70Oi/Ne/+Pm6/fBQ/Uunz3kjf/SRTxbIwWYLOOJgj/y7ryQF+P0zjfUYuWinrz26N9B91ej/MtMY0bmTDioZJC0BShPUb1mR4aehdkETq9uQBzNnjiD16sFZ9QDSoxacwUZzxJhrjVtnoDFvQ6pdQwa/hHysgCzRL5RPDDpRc79wD8O+WKZ6z1hOtZ4yyL0jyKkS7Ms+xoEOkgHiBbc2f0M2pGM6cIjy+E6VvZ990ZWlKxppe3WgtV/5ipc15CYV3uEyjNHkWv04xr0uN1xmddkRXOP4TnE01PPv9nQPHUjXokaWf51rjK4nGiOWheh0EXe7mneXiO/fwbXq4lprPoe2P4Lz1ADOgTVzeUgew5iyd0geb0ombLH8Ru3qtRvdhpwR16WcasNJma8frPmA9/MYnyeB9vtBh05qWrsLvAs8+1pOxG283+zsrdZ94Ovko0vFuO+/0f50fooYHuj9z0cMfzXR/PvhJeL2UvOiwHvnIcZzihyV8TR8Q+pCN6uf+q5hTByDwWAwGAwGg8FgMBgMhluAt/oavSpLt2gKkeZgjxQohtbBN3QdfGPYAiOlCPC1aaRvnmaFvhV78aUKtS1QWPaXn362brczfSuXoAjUAoVzb74C3N/TN6VFxkJV+qasHYHpUepb4a8uVUAywnPs6nY3Cj33UOQqb+l4L1abxaAzfHNbodCnWxdm2iIVx8M3+PiV1sO4hfiFk9+yzpfqkxP8mvcUhae+O9E32C9QMG0w1K9idx+qaOf4xx/UzjRfHrUOdQ9h8413rDF0HX07O6n0HGdz9es5CqieoTh1D7/u7KDg7138WvAARZHboeZL9lS/aM5QiO3uB3qmj/Hr+of7+vx/d/+H2xZq51zRDE3bI/sGxdDQDlHsrIdfDUs9mvvqTL8adA/EyqjwS/kzFB/e2X2wbic7KmI7S7X+lvgldueGgZWD5Ya1WKMIX41ftyNfx/z4XL/yp/gV6+WZ5t35rp7vzz4SG6QVoICcD/Yg2CA+2jnqsC6bX06qLf7675xzVfMTmoc5zMJvDjEuvdb4XF+IzfLH3/1x3X51rs8//vVv1u1/9//87br93d+jDxGHylzX+vTjD9ZthAk3u16t+7JgEUitgxc/Klb++q/EZrt3T+0xigM6zEsWjt3J1AcFJmkXsbXl6ZebY8xd/Fjp0plYFZXDL1hbxs0vUCF+7czxS33Y1vMc4Bf/Pth/z8bql5MfxW5kUcPSU3/9KvqLdftnT9S/Jdbfq3Oti25X1+p0V/sMC4cWZNTGWgejsdbwl98r1j87UZ+PphxTFHRGgdAUxQXLpeZahjiagXVUgz2Zg4lzU3C1rLb3C3Ic+O5xw1SYjtUPExBGeiiU2k/AKgIl8Aq/2p2gyP+PZ+qfHPv553/1367bR0MxNf7hb/6vdZsxtBXpHm6KQKOWq4tQ6PIV9uK//Y9gkGCB/M//y1/p+p8ot/rhm6/X7ayvAHB0pH1hOFAf7B5qfy8C9cEk13P3IjG3vPfC4FgZNOw1cWQAlg3II85HruiDXfz8THvb/40Y+bdfiNF4eqWxIFv8ySPthUcoND+60lrMwG4LweSaN+MYYw/gr+kXVxrHkyuN3YtrzdMvL5Rr5TWYKb76YHmmOfBwT7FzOIDhBtgdVfGnxcSdcy5iXdrtk8Wd7zzXKlcX6fQVK4sQ+UwE5hNYo92Bbo7My/Gp9oHTkWLSvBKD6R5i6KMj0RaneJ+owNwPEZeX09Xarbq6lwxzazRRvHsJ1ugZ5tN0oTHc6+t+7yIu7O1qDt3fVzyPIn1elbqHVqz9qNPT8R4Y/17xflhxVV2vTVsKsomwP5NFUCP3OkY///032gu/BBP7bK44U9D0AfkqGaW//kCx688/Ub46KPHu1lRFZz3ayyudD0vOjSrNo+NnYDqCxf5qBMOTpc7jMDeO7ipefAy1QRdjdAlm7s5DrekabJOluynKvGF18m6onWg0HpmxYADV6gcydDy8K5MpRoVA6ikOfnGidXFyqX67DxOBh5j/oaf+HF8p7715T2Qh5lcvn6/bcyhlOl2trbpU3z8/073MMZ/2kEMtF+rnR1iL/Vrjdv1Mx5xMdJ6SDCoymeq3z2mMiWMwGAwGg8FgMBgMBoPBcAtgX+IYDAaDwWAwGAwGg8FgMNwCvJ2cqqrdfLbik5GylaBocQyqLoushuDkD1qiHs1B5z5DQcZ9FAcK9yStOYIsqsNCZxUKB1ai2aVXK/pdBgroAgVOl5ArdFGocR+0qSNIPljYOBiJ8nd8JsrfnSeiYoagxLFAp48CzTkKpFWQcbkbuvgWJRy+57ukoWSHLHCLdg26WxCAmolxPjhUUcPPQn0XONhX0aiHoJhRfuFPRWHdH0JOMUTxQRQb89qr8/iR5sTeHcmUBvuiMn/59B/X7VdjyHowV3aHKkg2HqOwZ4yCgJD5jUC/y1C49Qrn//6FqNiPPv3Zuv2rf/PX7r2hKU5X168nM/ugM4ag77Gw6tGh1lOAZw4g/YtAo8z2JDP74JHoqfv7Oo8HiVYOemDRtAsUNo4g7drd0X3Ffc2LPgqhf/uViky2nO7raF+0yA7Gsd8R/bID+vt0qvvyIekhBZSk1Bs5x3bFVLVzVdMXkKTmKDRZLkTrpDQoA8W719b9f/bpx+v20Yeih5+df7Juf4EYc+eu1g7YyW4G6csQhchnTYFyFtfb62hOoJavC0B5pwzj+QtRW30Uwm9jXpZ9FsLXs7ZQ8HwyhfQG0popik8e9XRvFYo5bhtVo9fw/QSf6noJ+nB3T3Fxp6/P+4nmcz+ALLKEJAk05sf3tS8e7Gr+X75QTMvBY15i3WUNvb0ouA+xkCmKjI/VnzEKoReghMcoNPprFHN99p2KDzLWUtK8UT0f180R17gWb+JIVW6PNh4Gvtvvr/afXYxVNlPfJy2NT6uluZ2ggGq31uedtvKDnQ5kKtA/HaKYbQjZIOUrIYqchih+nTdysxQStD4kTv2B5j4jV7erffbJE63LX/1S+1Y1Vy4WZ1p/3R3kZZCK391T7rSHYrqUSPZQRDan/G6LCH3fDTur52uhr1gU10Ox1wjS6t5A47K7A3o8crisi32xo378b/5C0sZ/8/mfrdvPfq985HKkHOHDeyzi3uR/CXMetSenWn8XufaGp+caoxRL4cFDjemHD9R+9h8lyRijcGxZs+iznpVzLUH+HqE/kqaosOdt7/fgsqjctJEcVZDEswhqH/vPDgqvD/ssRyDkyFuSQPn2CAVj+y0PxyguFjDfoBFAwvebdLUWA8hEI6zhFuJ5G/3HPe8CkqtLGKskKPIaXeu+5h31eTtXbIJyd2MMKbWMUfC3U7wfaaPvatdqksGo1l4RUs6I1NVDYdu7d1WS4N6V4tuXMEthsXK+IrVguHHQQ06Bg3xIYGvEg7R5f10iRs8gK2R+M4HRRE6NGGJtG/tEjJyahc0PkKMwvpbIe06+kWQ+Rk774Ne/Wrdv3tV8f8vj2ezDlCdVjnu/BrELkxUHWVHuQ3KFciq/ui8ZLnP4AHPkN3/x63W75VAov1L/9CFjvHNvVUJhisLD0yu145b6O/V1zf37ylu++E5F4ydzXecu3nOuTvR8DyDzOp2ob76HXHKMsgMLmIUUxevnzk/14TAmjsFgMBgMBoPBYDAYDAbDLYB9iWMwGAwGg8FgMBgMBoPBcAvwdnKqunaLdEXnI+2rA0pgDHegEtKqDLTtPjzYD3dEZ+rCeaENydVgoOrdESo50/nJldBLXYv+VDd0zOJa584gJ5jM4J4w0L3cG4pytbsLuQAqVS9AfYpK/W0bVM8Z6Iys9u/Q9kALrytWXF9RsbZZbdx3vmsHq/sLQXcLKCOh2wX5jjXkURgHugD0QJV7kOhE85loZTnmQoS5MLwjeUCvL7pxp73q/wh0y2gAJ6vHqjR/fCp3nM6uxnyes4K6aKtxFzRz0OJzUPxHqJRepjr+Aq4tFaiwTz79he4TUqStwvNccCMD8ljpHPMKEocAxwS+jhl04FLRk0RuBrlR0NJa//PPJMu5f0+StjYoqYsxnM7w+dpxAzKJAaj3QUdjfg3qYRSob/+7v/7LdTudS14U+Tr+EWiZ9/c0plCLuWwhaUdZYC7DqKEkRfZmusGR7l1Rl4WrGre/5UhOP4sUDhiQKhYp1hCoojDncBFchZ5//U/r9scPJHf4+ZP/ad2+O9SarkEFr0qdZzlTTK8bJ70IMW4IaeUCMX+2UKzs9SEhgttGiXVJ6vce4m+W0tlH83W20Ni28NyUaJGoGoTvR05VO2/tGlUypobqowqyXSgI3QFowPs7mv8RXNVy7FFBoni5syvKuQ+5SHyosV5MJLmgVHDRuPWlkCsEdAqDRO/Bfbj5YV12EbsZa9ptjdEv4drz9KmkVd9+JzePCV28PDi6QN9Xw90kuIkp3vY8caLAd4cNJfoQ1OslHDO7cKRiXItg19PvqR+ePJKsOoGsqMIaOfnyD+v2ZCLZjI88YP9Q6yWKda0bCZuPvTiJdf1fQFr53/9bxc3RHG4rJ6KNfynljzvalYTj4aHifB/OalWlYPnoQHPuc+zHGRwtwzesy23C9z3XaSQMHvqlgNNZDblcAinOJ/dFlU+wtu4eKL599YOcqjLkcDuQOZ1+I2erFhyMdo+U3wwQD1266pd+C1Kgju6lt6PYcd/X+X6T6/pHZ3JniRF3rr7/UufMNO5hV/G19uBOhT2PeWzo6bphRGer1XMEwfZ+Dy7yzJ0er6RfMaSCh5BqHKLcQYI1EdAlFrKl1i4kV5BFLmo4YiJB8GdyijpEHuvDKRdhXFahSJ5b2JiHfeTwle5rmGhvfbSHdQ6ZXx+yvT3IWfsDzeM2jklQAsGHpDeCFOwO3t/8NiXA20PoB26vkQoFkOkx/wshW4ohP0uwh/zbzz5Ytx/uI0ecwrmT6xuyH86NO3CCPDzQ+mpDxnzzXhPjXSOGPKmLNXEXLrQpZMCjvtbZBdzpUuQ6rceKBfcOtUd6yHsu4ajmUrXnlypHMYcsazC8kU5uby16znNek9PQ7dbDO1+3pfnzYICyCkvFjKnTOt471Nz7q4+1np5AWhVgrbc7GtvHkIfu01UP97N7tOrPOUpf5JgrfP+JdzUOiwy5dgZXvwXcowv19+NPNP4R4uNojmthp0swp8oQbmoMDXz5/ol6KmPiGAwGg8FgMBgMBoPBYDDcAtiXOAaDwWAwGAwGg8FgMBgMtwBvWcbac16wok51BqI+RaD8h7mocqTBh6AQD0Dr3gP9atgFnRHyjx64t2B0uRT0w+lE8ohqhCrkDb0SLFTngTJawKmqzESlzvoBAAAgAElEQVR3G4J+mYCeN+V5Et3vnXuiE1cdUcGu4W7hQMvzQF1PQfuaw1nl5t4rajzeEZ6rXdzQtHxUyK4hn6jp6oDK4zGq2w880Rp9VO+foqp/BJpdkoKShu8OE7iQ9QY6ZzvR53G0Ok8UotI8nmk4lNzucE/zEkoFV8DhJa/ojqQ5tDvUXOS3m9eXmsdjUO3TAlXZUTG+As25/okVxt8WnnPuZhnRTYwuEQHdeLCGqJBLML77qOrfva/57NOVBRKKNp55CbeRDH2914GjWHs1fwpI67yKDnagIseiWe6C8v4Qzj4ZnJtcqXW8g/PsJhgASBQu8NxlwWPUDED1venK7Qk4nCvTpbv+buW2NRuJIltSmjLQ81YZY9UCbdE3P/lUcjevxjyHDMdDHCoXoCSDyhljTdeF+i1sOqBATAwgLaEMqgJ9uwXK9oMHog9fX0pGRmlVD+50+3BeuprquQ/2QJGHyw8d/mLQn8vg/bhwVM65RTOP27Wu5/t0atD6aMEpboC+O+yQKg+6O+SBdNgKIb+oMTOvc41pjmvFkEG4Zm/BMnAR3MS6WH8Hd3SdNNNa+QAuOCFO5GERFXCw+9XHkuX8/q7i9FffSmb1/bHcR4pS63vO+BWunsPf4mr0XO0Sf3XfBx30MfaHISQOPuY/JawhYvH+QGuuB7ehOkP/YF8ctHStVkcSni6crdrYmFqNPBAqrw1XxU8eS5L3v/2v/+O6/eylYk0Mmde9Q+2jRweiiict3bsPGngNh7FhW9f9xYda389fSs5XIHny3+Co+K4IAt8NG6kIxZMp4kYOZxoPUtpOqPXxKWSAh32Ny1/87KN1+/oaDluQ9vnIn1otzfOdDpw+kdPdyNE72E9rONYwL+p1ML/2dPzFhcbrAq5V/9/p17ovuBzdh3S9g1hzfS3ZRgAJY0mpETp2r5GmhduUUxW5Oz9fyalClC94hL1wF9J3zqScrneIQze5h3PODZHDpMgFKyRrPqRysznklZBrVYgBebPua7gTBnTe7CPmR5BWduCki/ePAjK8wNd+1sGekmCetTF3I+RUjE0dyImfYE9pl+8nSQ3D0N3ZX8WUAGsuo1Mc8tXIZx6LMgd93fcnfc3binEU40X30xpaFboLdvBuwnuo/BtpI+RC2DczyI93IF0NE93jfqTPHw01XjkcNDOMo4+5WSF3nSFGLnG/NRbg6JX2y6SzytnrenvlN5wnmXsOeSrLb7Qwr4Z452rh/fIxYtVf/w+S9v4ZpHItzJHKV5t5zuGBYmHL19qZp8qNi0ZyGELwOETZlqrU+0cPJTd6pfq+EyluZ0vOFZZB0XienWtPXeS6F5bl4J7nVfwc58Q7x09dlcbEMRgMBoPBYDAYDAaDwWC4BbAvcQwGg8FgMBgMBoPBYDAYbgHeil/u+4FL2ivqZRi9gWIECmGO74jmGSpVz0SRDGKdp9+XnCUGtbsFihGrVpOSNIWExYdc4IZm7kEe0AL1MwSFMgVtMr4vGniLMihWPgcNb+lEHbuCo0V2JWq7D5abB3nZFFKUqlLfFA1FrCy2J6eq69KV2arCdl2R7o/K46ieX4ECRtpXDNpfD5Ioj1TJCC5QkKwEGxR/OAhA+kDHj6Bpe5AekCo9AM32EJTwSUTqNxx2oC0KUe2fcrEKlNTZQuMJcxZXYF7WlEjAZaYdv5/K/85zzrvRFnp/KvtZ/UPPUG0wLDn/9Qek+e6gT9uQM3mgThaQrSxSVW2PIL+BWYvrNRThGZyEqlTn8CCJaieiOfq+5gIdJErQb10JNweMe8fXOWczuC9hTaU57gEU7BbkN1Hz3Fs0xHHpdOK++dv/0znnXBDjuvuiDEeg7CKEbVBYS/R3Akpq0tYYBqyMX+h5UzgIlKiw76NvU5y/aPon33D0Un+3sJ5qOungvoZwp1pORd/ndeiaNOjpOZbYR3y6lRyI/vrq+OW67eCKkRfvhzZeO88tGtesKVwSo0AU4jDA+sAibUHG2se9DhA3Qox1hbWeYgw4t+dTXAt/S6fBqqHucz8LIUwIMdE51waQxdFlsIQsoUaw8bGGepRL/vLzdfsuJLBHT4/X7d9+9cO6/d2J3GIKf/tr0Tm5Gd2DY9vdWHFoAZewEm4ofgl3TkgIE9DAC3rZ5JBkQJ6GUOySruZ2USgncHCQK5u4VUCaU6HvW4nkM5/9/NN1ex9uSxmkBztwTKKDaIicrqwh1Vto7YZOcXYAkyXmWpRgVuX7+f0wjmP36MlKtldO6aoHadAM8ewaDpMd0Oz39BCdPUr+1UezhY4vEFMr7CE18kuv1HWzqdo3suQIca7uKYdIBpIVFpA73YH8cidWTI0RR/uUkMD99WNIWgvsu9UC+RtcrhzXfQYZWeOg6rYo4ajq0s2Xq/II3VDnpZTacf9jeQRI3B3kRnRHSjCGdLByGMPFWH2+gCSHLmchJExx08aW53wak2K6e5At+5Df1pDVFB7lmppDCXI0ynAolWVYzBCXO3Bh+uiBcoz07My9DwSB5/aaYFBSfko9P15B+X6EdNrVyMU9vI8U6MeNM0ZYfyjv4HtYi7gH7zWhyEd/dhEXL+ByxDFN2oz7+o+s0OcZ1kiOcV/i5ktIdLIp3JUiySvvfST32wXi2myykqZvs/yGc3J+8pgfBK+fb7sDff75J4phn//iw3X7yS9/vm7v7D1ct6M2SlLg/Z/lHMISstixnIhrlrBoXtKYe3CsKuT7Bd0n4Y6b+IqneaaxyjLGcMgslxrb8QJjgvHhpK49yqbw3QXW/U+FMXEMBoPBYDAYDAaDwWAwGG4B7Escg8FgMBgMBoPBYDAYDIZbgLe066jXuox8LlrTBjsOVKUM1DCf8hRQvxOQsUSmcs6BKupAmydtypuI2hSnojMtQJd1jQSnhrMRHSRiVAkvlvq7HJW1B33RkiNfd7mAvoHOC9kclbJTnbNLeiepVaC/zeB4EDZ9XW6xenxVl26Rrii5RSZ6bRBQ6wX3BlQhZ3VtUh/JGu2EkMG0QBv3RD2mk1C9BC29xv2Alu419+PxHkETjQOdY2cH86bQ+VpOY0jJV0A6LeiW84XO34LManGueR/C2eMu5HdH94/0ty32zrbRPEfA58H1INl7k7SqAN0PJgguoLwOdGKfGoaCUiidJwZVsIXDw2ZNJSGkOpAWlFhzDtTzEHOQoSYA5TWCzCqqtIZqyLVKSHHmkMhdgxZPhwpK4YJ1kNuihqMsXHW1ooSWPV1r+AEcpihHWGjukQKc5rr/Mle/lW1GVGg1YPEXQsJYL7QuF6iqP0cMmzUyJ0orS1BMQ1BYuT54+QHkeWO4LoSgxV9PJBvY2xXNtQNaf4B5uQcJzOWZpDcBXA6WC8gAtojKC1yWrOjOU8jAEjoTODxnDecFbMGU7UboMN9x/mOek3oLyn2J/YSyGFLub+YJ3RMqrjnsl5TEbNB9IbmLQfNvQYLE6DcZg0IM1xnvEE4kHdHGfcg/ltWX6/bZeNrcy/bWYl3XbtnEsHYL7m3YEzIEzhy5QtmFCxxiRgK3IQ/SxnQGlzmsrWDDzU09dzVSvkQ6d9WsrwySvAVifoT9utNVv+4F6tc5cpUAa9dv0R0SEj7IuegEtFxKwhdiXpQ5pV7Il6r3I20MgsANGzenCdxrNnIoxI0Sz59f6xkySIkSuOO0kQtEiJ2UamaIwQXa2QJ9BNlpu4lp7Z4cIWtQ+wd99fkyw5rzdO90s3OHer4+1iLl3w/gPvb8+bfrdoAY0OlJRhbC3XCZam7OFysnrKranpzKOedcE2cGeA9olRqHCeYVZZ1VBRcf5LFlAHkGdU6UGSO0pcg5Fnh3KdFue5BT1TcOPpAWI9+n2my+VOyLESFZWoJrkZJXHxKuoBW/9vMce8QSHjc1kvY++vUQ0qptIvADN2it5s018i1XQCKO/kSY+U/cDgVOs4wuiPgPzsUCTkEVSl54IS8GuU7TDrH+uQ5q73LdLiHnonPuxk1yv0RuWfEVFXl0gfdbnn9wKGfHnXsf6JR4D784fb46R0GHtndF7aqmDIIHGTim5Ma6CRPEmAfqt6MDvX8lgfqBUvGE7y4oneFjnLOZXKiLKzhZIh4EjctwjJIMzD8XyD0cJNIJ9zyWBaAMDu+rKSSxo5HaZ5ea63O6MrMsC8qI1ChLwXzfzX/aHmlMHIPBYDAYDAaDwWAwGAyGWwD7EsdgMBgMBoPBYDAYDAaD4Rbg7eRUde3qhp62vAKdGHy3BNqqsqZjBc5TUdqBStpoh3QnAt2tABUxTSUjmE9VTT4Htao3XDlfBG3RiT3Q2qIYFbHf4LBRg+7me6x8DncQ8ONGoO1PC9H5fLTpiuDQT8UC1NmGmlltUU7lOc8FDWWL7kJBCso86GseZF+he301eFL5C0hiPDhZeLmoxMu5pBLzpfqhuyuKbwcOT3W1oqfVoLDXmf7Om4vi6C01DzyMWwI6cAJKdA1JXJ6j2rhDVf8OHEoyVURntX/S7nPS495SsPhT4TnPhQ1F2KOsCJTcGtIOH+uSVHY6C3lwu6mwzuhSwar0FeiVrAQfQMYTtCipWfV1xNlDlxfI70g93ZDO4WvnNqRqHtZ8CYlInlIip2tNUEF+NMdcgqRh0BEdtHKNm4+nefyu8D3nug2HOIXDW//+k3W7RMwo4Z6XYO4lCajf09G63Wpr3tYe5DxY00FLNNewpf5pYR6lE62prJHCssL/bKD17CNuBqCkMy7QwYouhOVSz8q4U+KcMaQudCfc2VH/8fzEHLK5rcIPXdlZxa48p/Oa4tUgUD/HodoBaP6MqjUp9NDOBZAQUfI7g1MCnf2Gu5JotOHKkjXSvNlMfZJCTkdHkAHcOULIT31QyCPIhbpdOTtUGR3HFFM6iBFFqjhF17NH9yRLXSDJePpqRaN+dilq9bui9nxXNlJgyg5y3LNDXE9xzynGOY90TLwhMdU5ux2NeQaaN6VS0wmkjRgjnMZ1+qucJmbMB7V9iXssxlqjNbjl3NvothOC2l5xX9xwQcN9YS/Y39c+XuaSNsY4Z9F+Pxuj7/mu3ew5S8hNyPlnjOc6o7SN5QKCAPGqR+dTIUcO56MdQNroca+F1K3VyO7ogpNDirozlONmDReUKlPeUzmN0RCx8AB51BnWS4f7Mu6FctwAUhzG6bAjN7mJ3zwfYv12sIp5/bZi5QHythdjPQvneYb9Psc2sKD8A+sygLSngPxuukTeAIlLmKhvC8z5opFuZXAd8gqWT4C0w9P9XqB8QrtLx1fIyCDLcnCKqz26yMKty9H5htZHkBFC3hTG7+e3fK927kY5EyDmMOcssc9XkFCVyFd9SG4og6IrMst4FHDcXCIHZtmKfhvlLCBbutlS6QjMfLIqIAlFPtnexzsFHTTxLuVwnRySJ77X0C3QQzw+fCAXpwB5XR+5zvUPx83zvL3D0Zvhrd2eA8ylALKfDNe7nChuTueIpyhrwP1+udQ6rj2+ZyNG5zj+Ss6jKaRVNeJE1twbS0KwzMQ8g8PVSPlyN9F3BDViTVHq+pSBzyCtupppjry80B6ZYs15eKesqEGrKSnTmk7nKC/xz8CYOAaDwWAwGAwGg8FgMBgMtwD2JY7BYDAYDAaDwWAwGAwGwy3AW3Faq6p06XwlJ2jDYaSNStIpKOte+Xo5VV2CwkipEqhnEahqS9Cprq5EoUohL7hORWEq8FTDhgaadEVByyB98mJRmaYjyQbSGajIcMdyoGJmoM3lcDkoQeGjROD0+LmOx7PuD0m1VDNvnCDqLVb+9wPfdXsNbRc0fdIHa9AUKcOhbK4iDRm0/hpOOTWkUhnkblM4z8xAg3RwOHH5i3Xz9OV5c3315b0DUY9n1+fr9vziWNcvQMkEsxpD7iqPUjCNW4W5m2YYW4wb5+UyhWwuRMX9aNs04wae57zmoQK6LZDii7lag5qZYazncDyibM+j1A6Ll65EdCLaYAfiNnO6PDTOF1lK2iScTzAXfNyjg+QtIOUWtMwaNEc6gqSoDj8G1ZNuCUvQe9tw8+l2MMcaPra3RUecqqrcbL5a43uf/1r3sH9v3Z5caT7ThYxV7IdwZlrORQ+tF3DHgQ4javNzzYtuAgcB0EkryNYOmjg6nuh8F6en6/aTDBI+jGeOmBvCSSwBNXgxlVStjc+5R7QSLeQAMrvdXVFhD/Ykg4vgApBhHW8TtR+6qtVIiHrY2/A8Xql9q6q0JipH+XGNY+imo/YSkpvLkeQUoyuN++7B3XW7Mxjic0mrFo0sZo79qaoglcF+dn6u+Lq3LylFhLWSdCBnw7pZFor7dU3JLiVidOISuqCKPzjSM91INP/d7/7otgbPd3UjS/I8yLExV+nkl2NOLtCHCWQnweL1Mk06z3DvnM60L9JtI4PMJ4X0c9rEsAUkThcnkvv6uBeOSQZHwL0DSXViSDpzxNYK8ud0rmstkCMNIPmZ4bnTJSWaiiOtzuslj++K2jlXNpvRHFLdLiXUkGRUiIs5jvcoFUPyQDljCienyZJyceWllGi16OAG+v+NtKqERJxSHeaENebONeZdUCi+DCHX/ABOmdwX6SxGJ54KMkrmfhUkBQ6OV2EjnfSC7eU5nu+7VrNHRaX6PsE4RJDYTOAMu4BUaQFJWpnreCheXURpD/KSHHlDBjnVGO8ZGy6nzXju03X2nGUbsP6pJoeR0CzTHCrhxLXT1rpkLwf+hkZs3axo84SLteiEBRlvjT17q/Bq58dV04QMELk13ZjSDHlzCeevGFI+yMkoT/GRgFYF3r8KlobQ+a8m6usRXOniRto4R598973eRS4gv9kZKE/ju2uNDJjOrpRZOcj+siXfdV/vNHt4Xw5i3DPayNnu33/snHMuiinRfjd4nu9arWa9452g30cZEnw+mqrfnp9q3zg6UH7GvMHzXuliCzrp6eMM71n5TP0fbLhjSjbaavqE/eCxDZeo6Vj5yXKgeZAXelcoc7UXmDfnV3q+b4/llHWG94wSUlwf8z6AtC7Ai7VfvP27vjFxDAaDwWAwGAwGg8FgMBhuAexLHIPBYDAYDAaDwWAwGAyGW4C3k1OVtZtOVxTOHdBTu6ikXopt6ApW/UYF65LUOlDIa1ACF9eiKk3GkEqgkj6ripdoX03hKBGvzpNBcrIENfT4QlTxYgpnKFDssjaov5DfVLnorDmkQyVoUyUo2NMFqGBdVFyPQTVLRHteXK6oZnW9RTmV57mkkb9R9rPhxrXRRjV2UPxJZaSDDh2RsploxRPIqebgkLb7ouf7vuivZ2dytVg0jl1lpXOcwyVodCFK3tkLtXfgdhUEoIziOXw6f0H+Qcr5Feh0yZ6orSkcxnw4BZDyGZbbGzvCc97aTSSknArypZpuN6D5pgGryYuGPStE2exUojzmkJl5JRwRKj3nkm5PI1ELYyq9mr6+voRjSSIKZW8ouYeDnCDPtS5LyhIwLpx3S1BVlzPNk/FM82eB8c2WkIhB3uXFOubGuGmLaipXe77LG3eo9uFH68+jttaBv1B/tw/vr9tg8rvqQPN8PpHEJoPEzAd9n25PPpwXikz9PKnUb5M5JKxN3/qQBlSgqm5IgijJAOW2Ciir0PhvOJwhhi6xXlug0UegxfZAv93b1xpdQtJU5j+t2v/bw3elt1ovGZyH6h31/zxXH17AHWEOt5kUY+FDhlHCNWUOOcsM8XUOB6PSaQ70EV9boNCnjSxgCVkQpV00uMixh00g3dmQMIIGvGzrvhhTHWn7BenvDmAc1afdRPc27K76hi5r74ra1S5v1oUHaZhf4VkoJ4ZzTwrpJ+VuMWQ1lDUEcA9L0YfcR+nkNJtBcgUJzYuTVZzdO1C8Ta40PoOBZIWf7CF2lFoTdPgMQ8rdMD5w6CowL+mw0YEj2fml5JUVZMZdrPV5AOn0FjGdL9zf/MPvnXPOLRFDfvPJx+t2jTHKYrrdICfDs9XIdZag2S8gT7qewqEPrqUBZBZT5Jc1HP8uzi+bc+jcUQ/9OYarIuJZkSp3dQvFlwKuoD3Eo50BXNGmihEt5njYFxkbwoXuNw4gK2z6mC5f74ogCNxgb5UHF+eQ76VwV8LYZtgMYZrr4lRzrDvTOEeISVGomIRUxVV0GIOGqUb/FPgN/OVo1f85+rINd7QI6z+BxDdGLMiwzi4gleX7UlLpmQLkX4EPx62W1lmAdeYh123hHgqHTtsmfN95ndX1fboxYW3lNfYf5Mr5HJKrFPkZ3P8Kuggiv6DrbkSXL8hzKS3c3VWcvOE1XFwpdtO1eH9XefHOAHkG97mS6wP7O2LhHHN5Rsk/5pdDDOjD8c9hfD3kTP39leSYbs7vCs9z7maKIg1wAd5xYmzU80pz+Ac4r+2+xDtBzO8F1CdtvAtUeM+PIb1NsedPIUNLrrWOO+er9+wrxPNXr3Qvl2d4n3lDCY1iiVz4CrH6he73y+8Vc//4QsdcI5+J6PSMGOmFG8nEGum/4F3fmDgGg8FgMBgMBoPBYDAYDLcA9iWOwWAwGAwGg8FgMBgMBsMtwFvJqcqqcuPJin67A5rmPmhlb9IbVPXrq4TTnaaEhIKuRSggv+G4EOD2DyAp6O6oUvWooZz/8OzZ+rMF6OzZQvSou7uinpe4aAYHnwROLWDNuqognQ60zw0nHrjsoM483YJYOX/SUJq3LshZXwP0LlT69kilxr1REuEqOh2AYgwHmBmo4lANuFZfThY7Bw/WbR8uCWEimtvundU95KAMX12qYjyMVFzldA6oE1xRUp6jtse+B9t/MqM7FarmQ/JVOrhzQF6yE8K9oXw/35N6nufChue7sZ58uqCgDdouKbzXkEfMIX/J0RmkK5cZ5j9omzs9SQIvTuWo9O03P+j4ht4Yt3Rf3UhzbQjJTwgq8hxrdDHRfYXUFEGqkc0xHyGnmhak4qI6PF1c0JeUgoTFjcuC2xq8KHTJnVXc2n34eP05Y5wPSmgAx6YYrgStluZepiXiJnRdaGt8ag+yQayL0VhyTwepWohjkuEq1seQDw7vy6XBwc0nh1TDo0QT7j8tSKuiWM9RQ64SbXhygBIea60v4fxw947iy7MFJAz1FrVwQO35roxW/bJwiPcd7SfTWuP1CgHrGeJMb6rP9xGP6WSYguZfQf7SgRPLYirq8PfffrVun55pn44bOn2C+Z7DETLG2o6w51GCu1joOuczzZd+W9ISD+usAJ2cUr+cLnrQLPmQznGvbTW0a8/bXmz1fN/FjWzap2PaRHP4GjLNXhcOIMh56CqWg7JP6Rkp1pRKs58pY+51NXcGcJBKGhljiDH8/LNfrNvtnmQCSU8y7QjOogXkQRXu12Ge5ZBC5IgFk6XGf3StufMK1PkQkkevRp9RxrxFLLPMffl8levFHa2JfIO9rn9ktAfCXk1JVIY1XUDaSxmED0lpG33t1XrO+YXi68WZXMTmjVwuOZYMLYIj0fUUsRAymJ2B7rGDPb3APr7ItBZLlBG4ONe1agiJCuRDKc5TQg6EygXrLtum5L8Vhu6jxklvChfE0Vj7mQ+HP89RGiTMC0gFkQyWcODKkSO1MFd9rOk2xtaHdCTuaF32+qv9tYXcpoMyDHT1pDsVc7QaMf9irHE7v4IbEVx2wwUlLZCf9DRuXcQp7ql0hyuL7bkZEXlZuBeNa2LgtBbpauqwPhxiS4Xci5LTDfdeXgydGiVw3IQTUkzHzUJtDyddNrnmwR5kyNgPLsdKsPjuOkOOWm68u2pMp1hPE8SOKd5HFrXmw96+ygtEHT2HCzleLCmy/dINnufczfYS0vWM+Rn2rQLvSlPIwH94qXjTggwuxZ7T62k97ewohu6iH/oDlXz4p999vW7/8R+/0P2ET51zzi2W6teTV3KG9pzm068+l8yWjmh0obo4Ufv7E/3t1y9VFuJ0gvFHbkwJI9XfIeTSbXzOd+mfCmPiGAwGg8FgMBgMBoPBYDDcAtiXOAaDwWAwGAwGg8FgMBgMtwBvJafK88KdNq5BCWjzB7uqok3ZFOl7pJVR8hFBOsC/XYLWGUI6UIHCRnoSaabtoe5n+Pihc865+UQ02AL04IoOLqC1sfr9cqnjw5DUfrj/QGeR5q+npIagMOY45wxU9Iy0yxuJis9rvis85wV+c1q6UQikfpaQVrn69TRwjjPH0EEK0gYlvNOX9K2VUOZBGwC4KTWSmAWqkbMYvB+I7hi1IWXD+WaQ2NBhIMTzLQvNoTHkdBXoi+VCY9vfEcXxzoEkfAmpx9sXwznnVkK4G/pkCEkKZRAlZDYJKr9323qeBSRGMzifUBKYZzqmDjE/MX96PfXFRx98oHNCruU3Tglxon7udCgRogucmhHWHMxR3CLjOta9p3CNWyCOlKT2k9oIdy+6iZHSG3rbXIO6ibqR4dURqL6QD/kby0mdUuI/PFC8KcO4PBGFdGeg8Vn6XLu6Vj/WGs3xvC1IdboNnbWGw5lH2QJkGCX6fsNtCtT2mk5cA12/OIfbEvogQND3ESMCTBi6FlEeEMXvhzbu+75rNTT35QLOC07XK9AeLXWvZ3PNsUczOJv4kDLAmYkS0TnkT2EbsaiHeADqet1SO+qvjt/HXunBfsLDmg9A8w8Q3+lwMsP+egVpY0jnMqynErJbygj8+PVyUA8087RezY0tKhud53suaiTiNdw/nv4oKvU5nJ8+hFSCcdCHtKlG7KcUuSh8HCPQEWyTGg9nG/TnYUMz9yATpLsQ99A8Y7ygIxOdQuDUUdC1A3MR9PDjMznofP2tqOWX53T0gmwcEpgtGqhsIAhDN2jklHRyomtQib3Fr14vm/IgpykxnymboEPfTh9yhzcoxbqQC3BuzxoZ5fRC555CF5t0lVvsQB4S1bqvGg46NdyjPOydlN2N4f4atiCzRDx2OL4HCV6GORA3sdbnhv2OSALf/byRpC3va6L8w+W36/bJTH0Fw1gXQ9rBPKO7g32mYG4DRxrmtFgXAUgkIA8AABecSURBVGTeLbzHJJD/Dpv+2Xg9QP/RjSaClIKyrQTOm12MJ52qfMzROtJc8CA/j7vYm0vkdHAEdMi7O3Q+2iLysnQvxqu4cG//0fpzH/GKL6D+hlMxxoU5KiS8Xqw43YHMlP3rIeersYdsxF0kGFkjmWUZjJaHEgLch7D/lhUk6rjfRf76/DpHEFri7au9p7G4/5GkPpRQMbxQrqkNZJs7o/aITgdyX+R5PhzQZthnruAwNsCYHI8Ul1v4HqHGs7R7OuccuUKMd5qPfvHpun3/w5/pHqaruHg1knvfo48kj+v28P5B6TTewydTSN8muq9TWE89u9bxMyapeEdidZkK/6Asfa+vHGx8AfntpmDwjTAmjsFgMBgMBoPBYDAYDAbDLYB9iWMwGAwGg8FgMBgMBoPBcAvwVnKq2tVu2dAJ//DN9+vPz89Uefo3v/xs3UahdlcUrNJMWi0o7hGrw0NOQ5kVnKJilMlvB3CtilFZvqGZ7+yITjWBC8sYVforUGtLUP4WoPb5dKGiKwXosXNQ6E4noq22alG3Wj1IR6aiaVPycbC/uudgm3IqD1RRcr1AvaZsKgAdmnQw0o0jyHk45kECqVwJmjwoqbORaMN0jcoxzun0enXsRMcu59c6FpXePTgtkOGWQooVQGKRgXo4J60Sz723p+rovVA0uziCO1Vf4xb6oLwG78eFo65rVzS0Zh9U8QjzpwL1kBX2Y8hv6BJzDXnSghI5yik8rBF4QVBmQdngADKPIFpRpD3HWKC/K7D+KtA1Cxxf09gA9OMSVP2iwLpETKkx7wqs0RoyEkpxlqicXzWylG0aHHm1c/7a3uP1EgtS+SNQTzM4P5UYnwJ/Pb2QxGF6qJjXQ8ytUUl/AalAPaYjBCjZTZzzIcmop3RMIiUdc6igfEbPkcI9rId1VoDCXuC+6JrDXyBi8NhbmBeU4i0pF9wivDBwye5qbpfYUr0rOIDAmWe61J2P4b4S3Idjhg8ZGCQUV2eKe8c/ii5cL9XXQ0gUd/YkxegfaA/sNHtR5CBVLunaCPcoOBFW1C5AEhPR+aOkXA9rF5JjD2MUxnB8oUQuRQDHOa8b14my3h5t3PO8tZQrh9ZnGarvL5bKFQ5zODiSJu2/PkBsuFcit0HocVlJPZWa+QJuQ7XmcFQ18q87ivPHL+V6dFRBJhrDKQ7OS2Gsa4Z0qMQ68yivxmOcnmMfLeBQWmmO0kMmwznb4XuQpzrnwjB0e/srOdUS/VbxGeiUir+teUuQi84RuyaQ7bZ9rbMMErKSuQb2kAgSGbogHiSrvuugT2ZwsMuXcJC81OdLSt4wplEPkg/INlotrdFrzKMCjmoeXK6WdPFEfhtjvi/XE2J7azGqnHuwWN1TjlhWfSwHx+V3cpsdQRrmIU54lD4h/+lg/Y2v1bcp1vQcY7ixdHHOBfa9bLo6KEa+EUEaGkPCHg+0XqMu3gMw/u2B3HkGyMs6jJUtxe4aTo0l3FGXKcYFFrFxpGOGgyP3PuD5novaq/kUtyFxWmC+QXIaUH4WURKNuY3zJD09A+dnzsQfz1zAXTDH+C5mio1pk2vM4Eo4wVrku12BXLiCy2qK/TIt6eyHZ8U7kw85fwRZWAuOryXmbL2x70GC9h7MNz3Pd61GXhxjTDK4GtL5t/I1D19dILdr6VlQWcMdYAOMUWLEu4CrHnKq3Vz3kEQ60R6krffurORJBSSUVxev1u0Um9hkpvblhfat0bmu/wq58PG15sLZFK696PyAumg60UWvL+fQgeNcUCsHvkRs+udgTByDwWAwGAwGg8FgMBgMhlsA+xLHYDAYDAaDwWAwGAwGg+EW4K3kVM7zXdC43Hz/7On64xyU0T8HrWjYEd1pClkRK79T0ROBKlig2vd8LFrWi+fH6/bkleQCA1CruruoRH135aYSg0o6vZK0oFjCSQFfabUgCYliUaoLUNlIiarAmxpNRb+6gFTqzg7cCXA8KaARpC4vnq3o2xkq6L8zarfm3VUV7oG0PLDBfHQKaX+sJO5BehSDJl+BSjhDn5w9Ey395FQUthhVzhNw9f1G/lRkoD3CbWoGSdAFKplXofrt7h31fdxhRXRQIuEakoAqvQ8nkghykSDS/QwgHaMHzvshjTvnnOe8RmbH8aIkMIyjjeNvUGON5jj+Go40SzICQbfOr7V26kztCK4olKt5Tn2XtJv5AFp5CeeTvAb1dMOFQeNC6UWbDjF0JID7D2PNEtTy67muO4eDXIHzbNDMg7cLlT8Fvu+7XkMXD/BcdOvJM8WPCvfsLTX3YrhW9XZEw77ALV9cas2Rql2ir77+7Rfr9vJK6+jorujW2c1aRMw/fiFqu/Ow5h6IGvrw4wc6JFEfXy30fCGcUQbD3XW7cNo7qkzxmhJdSj4oNYvQN8F7oBuvThy4ariiC4e+nqGe6z4y0ImXM7mNPBupv+aRjrmjpuvu6B8wdnDj9KXODzeNHG5ilGUMR+rrRbqKHVdXlG3gflPIDECdLinFgXZ2MBAN/N7hwbrdhhS4pPsj1hbJ4QXo1TXltRjfi4ZmXJA2/47wfM9FTQxfwMXlGvdznenzEwzEvV3kBzD3iSitwv1TbpZCMnZ+qb3wxrHIOec8yDy6cMSp09UCv4asYzHV/Y7OJT8usS4HQ93vw4da2wPw3D3s+3QEHUGG8OUPiv9nJ5o7XbqToS/LpfosK97PzhiEodtp3FIvL7TOMsSKDHHR2/gZE9IySFdjOAfGCx1z/lznf/rjmc4PeTedGvt95H90JWqcZzKP8iX9/2SiefEK8tNvnyvu+si1Ht1X7Lx/X7lwAIcbB5czCGfd1ZLyXci7IE3rQjoZNnnFNkOrnxWu9eMqtw/7ysPu4SKfH2pv+fFCe9sU0tvOUgOdZ5A89uiIq/3k8lL7zMVI8zlItF7uPZKkK4Fj143MJsO4drtw2IH0bfZS7y35XNffw7ocDhXzd3Y1nnGL706QEKGcRAqpzsVEo7uALORJIJlt8b1KY2wTvu+5XpOj0Y0QKnwHha0L6AgWUm+CvBxhw8c8pMvmbKoYNTpXX9eQ2ReQdKV4N5lPV+N+fa1xySCPyTEHp5BZpiihEMH9qAvHzdCjnE1zY1nq+OGu3KliSMrrjRVGubLwPuRUvu+7Tns1zz3v9fs30n0XhHqucapY9fSl1lYcae0O8S5WeRi3kWTBl8cai0Gk8xQTSNvgRNht1kiN97mrK8XqFJLROfbT6xySWDzry6XaP1zp+lPkXOx6pO8b78k+ckNKW3NI9CgL/KkwJo7BYDAYDAaDwWAwGAwGwy2AfYljMBgMBoPBYDAYDAaDwXAL8FYagbwo3PHpiuZ0jurdPdAKJ6D8PwCtOgXFlq4SeQ6JQCn6WIRjhqAeOdDsjy9Ff6pmoA2WohGnixVVKS9Fa1qimn0GOVXQQRXsI937LiQKHdzLEnS6DLTolyeiGVegsLb6Os9yDpkEZEovzyUX+/qrH5xzzs1mP61K9U/GTd9u0L4gp6r+9NDVPyBTId0abUpBCpyohuvP3uHhuh13RK0LSkhvIKcKGkp7NgOVzRdNMvJFpZtVGvvTa8kNdkBxHPQ0Dikq/6e16G6dznDdXk41R2I8Uw/yj4M90ZZDSG9Im94mPM85v3EzCNC3G9XQ4bLigarq4/5K0FCvIFFbgsI/xHnqhfp69BzuOGO4dkG2FHqgDcbN+KEP8xpSJkg1HNfifcgzQDH14CBSUBoIfiedDcaQTT2HK9oEVGc/BL0TVffXVPTtmXA453vOa62e0wd1nW5ceQ6q75X6OxuLHpp0NT5dSFgHQ8Xl8UTHzxeg7Ho6Zgrp5z99+dW6vXuh64bNrdWQLI3x/x247u3f1RoqISHJcZ0u4unVic5z/2PRvTt97C+XkPlQOolxWcw1RzGcrg+Hkq3C95zXUOdJhw1SxYTyWPtWPnqxbv9wJanEM4T5uzWcqjw6L9xft18d6W8v0Rc//8s/X7eHQ43BoC+K/o1DZARKvsM5ioVi3mSsWDuZQfID15yScQ7xIgJ1mX1DCfGGuwWkw5QzThCbThvZUQ6JzLuidnKZGo31jM+PtYfMICV6daH7uTxQvw7oAgg1K2N0iZiUQ8LKPThAjE5aOufje4/W7fuHd51zm3LX/TvaT9s7kmdMMj1TEaiPl4gvnRSOYbBqmkIG8C0c75YtPeDlXDnP7Frr76ije3DMwcr3tC/6nouafTmB68cIdPuDROMVwm3Ih7SYAYX7fO8OXI4i9Z0HF5zJRPvJHtbf/qOH63brjva0vQerzyv0CXOws5eKF+HX36zbF8gLo1B5zN1HmgP7B4qjr+DEmkHmeIJc/hSOKHvYPzacqjDfkrUsa3vjWWWFmx+v5lm0p1ytRjAPcQ/tjsbnFZ6lDfnjeKzY1oqpyYEjLnKINnLydkv90Otq7uw9lETYb+I/04MEblPjY+X1Du8qmYN8DVqdkA6+lA2hRIBXaMx9uPpVld45nl1I5jfC+9JspPbRexL9V1XlFo2k2Mc+4KeQKlK+zlIVeJ4QcsYaOeLG+wXWSwI5cQx3wXRKib6OzyEBCpq4x1ITKWLEFDJyuqxyGw34rBjHzTHCevE1Xkm7j2N0yMaLWP16GbEkV9tbi57nOT9Y3XcIyd4cjl4ODrPsCA/vvhcopxJDWpVg/bEP7w9wLUiXn33zfN3u1RrbvZaOvypWc7vAOwyl9xOMw9NUrlXLAeR8e4rbz64grbqAKyzfcyB/9aAR3BhDjEsLORJzgzHyrp8KY+IYDAaDwWAwGAwGg8FgMNwC2Jc4BoPBYDAYDAaDwWAwGAy3AG8lp8ry3L04Wcmp5qAEjyaiKl5eiQ5EF5xWwkrbohst4SyUpqKqhqD4BaCP7fRAz30gZ4UZqsl3OqI8tgcrWlQGicWiJfpcmup8fg8UStA4A9AiW6DWjkFJPIY84/mF2r1H99ZtSjuuLuCQVYiiPM9BEb+hmm1RklO72lX1qi88yqbeQCV+E42vxPHVxuegGMKdowMpWQSnlt4QFMMMVElILmaXq/6cF5KE5BXdzkRxDGKdY/+u5AztLhwJMLfmpa4/pUNXhXN6orwmkPBQChJzJaHKepXT+2GL8DznNTRHjmMBynqEZ3OB2qQWlhjrl5fq39OJ5vbBkaiFEQwudrFGFjVokaAFt1ta9zdUZB/zYgY5YwmKa2df1+yCEl7RuSmn/EsxaJnp8+lS7REooJeg1xeYyz4oylWhcQybivfeNmnjdeWmzX0vQI1vwZGjgNRrjr7K8YxBrD7pD9RX7Z7m//m5nvf6lcb57n1RxX/+sw/0t3Cy8K90bzfV9n1U9Q9+9rN1u3NfVP7+B5JN5nB1ux5DigvaeoD1dH4uGvhgT9KDpKO+ySDdzSGVq+FsCGMzd+9A59kqPM95DdU4gGIrwhwuIZ/Ij9V3o7lkLl/9qDH69FB9kdIZBs9/bx/y3K8kRVt+K8px+880H/p0H2o6JsK6Ta90nWvIZsYnohxXiO+Hu5C8QUbXBUWa2wqMipyPQEJHPcqeZ3CaybAvXjXOIWW5RTlVXbu0cYnoQ753F3NmcQW3RcSVU9C9D+DSFUP+EeD3Mt4198skUb+FcKms8OznV5LEtBopJiVcKeJXJ9Q+e7B7d92usT/liHEzULlbnq4/gXvKOaQgyR04kY51zMWV+iOFV2MLEq2iek+/H9ZuLdsadJXbzeDqlO5Q8qb7iJBnUfJfww0mGGrt7vfhWkknpzNdK1gqTwlOcC3IuMt5I/mH9KOCxCKe6NjBVO1fHEgWGw40R/Ye8rnhvgTHnRly0TPkWnOs12CGuENtIPqsHdOLczsoPeeum7HI4fB3gX3uFOujxB7C/HkE15cQEr+opfPEWJd7O4pncQHZLssvPFNsvYbM3mv6p0B+6C01VjX6uI1kcWdPMttoB3tqG+ufOfiyxOdohzomhSMh3ZnGC9wPpLAIB1uGty4VMcY9pZC2xXe4J+M9As+84fJHR0K8X9JljtUFIrhceYg/qYMUDTG41Uhq6KqYzSAzhrOqj7nfxnuhH1LmBbdF5NoznCcYal++5JyCrLfToxScGh3vDe3toKoqlzauwAFkahlkwBWcZJEWuhj7WQ4X1Dne857BQTHBujzoal3c3Yf89RpyNjh1fvhQMuN2k3/QvXKKOPLNS+UzQao+jrCPjjFWr0Yo1QKnKkrlvAgujHw31i26ErK8GvG3hhyt4rv3T4QxcQwGg8FgMBgMBoPBYDAYbgHsSxyDwWAwGAwGg8FgMBgMhluAt5JTVVXlpo2rUgXXmSno3hdXoh7moAxFrDAOQjGdKcZwvhi0RAndkH+AMRb3RUNjlfko6v7JMT7oX8UM1cDhQhSAzpiBEreAg9YC1clPRqIqfnEsB4EL9EcH95Wjwv9+X3SxFFKAfCyOWDta3Y/vbfO7ttpVDZWvhgyKbiFkb4agGN/IsJzbdO4pKzoXYMw3Ko+DdosxJLW79uFUFmsueK2V9CzzJbHI4EgFIyuXRKCwg3IbwUljmWk+ZZhQEejRk4XmsZeJ4pgkeo4Q87LOIQtaipYXvKGS/DZwI1csIRUsQdUNIQ+jC9cUNGOui9FIff7DC7my3L8r2UbP17XCluiE3TuazxlcIXRV56bRqi8iSLvCruQH7a7WX6uvNVwHrOpPqRpcVkgJx3qdpqLCzrAuwwAyHkzICn1Jt5iyGcd6m/ZUta5xdSY5zEEginGJscoRN33M5wqV8VsDUU8pSQt+/HHdPnupsR3sSnLVG6rPP/rzT9ftKaQ1N7GbY9iChCRA20HCmC1FZ41A37480bnvffLRur3A2H7/ww/r9iEkRJScTOB+4GGMeJ8duHxsF76r3J/Gap8U6z1Jy+pEfZRB5vH1t5K5vXysGLLfwXqNFE9aHY17tweXueca65MU8lb0dThY3VunIwlBAYr97FSxNkJ/9kH/bw+0/0JxvEEhz7Ef1LCf4CqiBDRNFTEyxCxaV0imvUWZcV2v9+cS+1y5Ia+EVBV72zJFDgMZags06VakueBhTm64QuK6Pta0j+fMK1G7X56tpB1LyEBaGJMhOjmAhJjnrqHFzSrKrCAhyBVTEsjikx72/ZbW98LpbymRjRLdZzp/PxqOqizcvJG2t/DMC7ieLZZwKvIVE+hORQU753CA2/bgMJLsak37kHCkY82H6yvR+OslXcFW9+YH2POWOkc2lvT+6lr7hDfU9QcHWIsYu+tXkt+l2OvnmMvX2BcrxK851h+lQ71dzYebNVPX29sXqyBwi2YfGRWaV089jeELjCdLONCR5goOii3IHXoXevad/uvLJgzuaK6y9MEELmwlZPyteDWPooBrFbkKZFbtHmQ4eOeIsF/miB0zjGeYQUIUUVKGvXAKx0/EJsbWyxlkX1NmaVuE5zmvee9j6Ykp3nEOdjSXKElkzNt4v4Ccvka5ALp5eZSn1CiLkOj4GMdXeAfxmr0ogPQlwr204TxVYhfLMb4hxqXCOi8Zg8CfONpXvneBNZqXjJEYrw1bYTW9LaamQuWKctWHdMnkS2KOnDmE7DKMsc/hjAFyb9qEnbzUvHiIeXF4X/15/5eSBZ88VY7y77//rc5fruJo4ismjiAVfjFWDI0hWx8cKP6eX2oNTVDqoOI4YI+o0Pl00WPewnIqE0jlErjrBcHbv+sbE8dgMBgMBoPBYDAYDAaD4RbAvsQxGAwGg8FgMBgMBoPBYLgF8N6GBul53plz7un7ux3DG/Ckrus7//nD/vOwMfyvChvH2w8bw38dsHG8/bAx/NcBG8fbDxvDfx2wcbz9sDH814GfNI5v9SWOwWAwGAwGg8FgMBgMBoPhvw5MTmUwGAwGg8FgMBgMBoPBcAtgX+IYDAaDwWAwGAwGg8FgMNwC2Jc4BoPBYDAYDAaDwWAwGAy3APYljsFgMBgMBoPBYDAYDAbDLYB9iWMwGAwGg8FgMBgMBoPBcAtgX+IYDAaDwWAwGAwGg8FgMNwC2Jc4BoPBYDAYDAaDwWAwGAy3APYljsFgMBgMBoPBYDAYDAbDLYB9iWMwGAwGg8FgMBgMBoPBcAvw/wMyPyQp52hNQgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f2150e13e48>"
      ]
     },
     "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",
    "# 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": [
    "#### Pre-processing the image data and scale it to a pixel range of -1 to 1\n",
    "\n",
    "We need to do a bit of pre-processing; 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": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining the scale function to convert images to the range of -1 to 1\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",
    "    x = (feature_range[1]-feature_range[0])*x + feature_range[0]\n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.9843)\n",
      "Max:  tensor(0.9765)\n"
     ]
    }
   ],
   "source": [
    "# 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",
    "First we will define the discriminator. This is a convolutional classifier just like the previous projects like SVHN or MNIST, only without any maxpooling layers. To deal with this complex data, it's suggested you use a deep network with **normalization**. We will create any helper functions that will be useful.\n",
    "\n",
    "#### Defining 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": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def conv(input_size, output_size, kernel_size, stride, padding, batch_norm=False):\n",
    "    \n",
    "    layers = []\n",
    "    conv_layer = nn.Conv2d(input_size, output_size, kernel_size, stride, padding, bias=False)\n",
    "    layers.append(conv_layer)\n",
    "    \n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(output_size))\n",
    "        \n",
    "    return nn.Sequential(*layers)\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Discriminator(nn.Module):\n",
    "\n",
    "    def __init__(self, conv_dim, drop_prob=0.2):\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",
    "        #saving for future use\n",
    "        self.conv_dim = conv_dim\n",
    "\n",
    "        # complete init function\n",
    "        self.conv1 = conv(3, conv_dim, 4, 2, 1, batch_norm=False)\n",
    "        self.conv2 = conv(conv_dim, conv_dim*2, 4, 2, 1, batch_norm=True)\n",
    "        self.conv3 = conv(conv_dim*2, conv_dim*4, 4, 2, 1, batch_norm=True)\n",
    "        self.conv4 = conv(conv_dim*4, conv_dim*8, 4, 2, 1, batch_norm=True)\n",
    "        self.conv5 = conv(conv_dim*8, conv_dim*16, 4, 2, 1, batch_norm=True)\n",
    "        \n",
    "        #fully connected layers\n",
    "        self.fc1 = nn.Linear(conv_dim*16, conv_dim*2)\n",
    "        self.fc2 = nn.Linear(conv_dim*2, 1)\n",
    "        \n",
    "        #dropout\n",
    "        self.dropout = nn.Dropout(drop_prob)\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",
    "        x = F.leaky_relu(self.conv1(x), 0.2)\n",
    "        x = F.leaky_relu(self.conv2(x), 0.2)\n",
    "        x = F.leaky_relu(self.conv3(x), 0.2)\n",
    "        x = F.leaky_relu(self.conv4(x), 0.2)\n",
    "        x = F.leaky_relu(self.conv5(x), 0.2)\n",
    "        \n",
    "        x = x.view(-1, self.conv_dim*16)\n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = self.dropout(x)\n",
    "        x = self.fc2(x)\n",
    "        \n",
    "        return x\n",
    "\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",
    "#### Defining 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": [],
   "source": [
    "def deconv(input_size, output_size, kernel_size, stride, padding, batch_norm=False):\n",
    "    \n",
    "    layers = []\n",
    "    \n",
    "    tconv = nn.ConvTranspose2d(input_size, output_size, kernel_size, stride, padding, bias=False)\n",
    "    layers.append(tconv)\n",
    "    \n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(output_size))\n",
    "        \n",
    "    return nn.Sequential(*layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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, drop_prob=0.2):\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",
    "        #saving for future use\n",
    "        self.conv_dim = conv_dim\n",
    "        \n",
    "        #fully connected layers\n",
    "        self.fc1 = nn.Linear(z_size, conv_dim*4)\n",
    "        self.fc2 = nn.Linear(conv_dim*4, conv_dim*16)\n",
    "        \n",
    "        self.conv1 = deconv(conv_dim*16, conv_dim*8, 4, 2, 1, batch_norm=True)\n",
    "        self.conv2 = deconv(conv_dim*8, conv_dim*4, 4, 2, 1, batch_norm=True)\n",
    "        self.conv3 = deconv(conv_dim*4, conv_dim*2, 4, 2, 1, batch_norm=True)\n",
    "        self.conv4 = deconv(conv_dim*2, conv_dim, 4, 2, 1, batch_norm=True)\n",
    "        self.conv5 = deconv(conv_dim, 3, 4, 2, 1, batch_norm=False)\n",
    "\n",
    "        #dropout\n",
    "        self.dropout = nn.Dropout(drop_prob)\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",
    "        x = self.fc1(x)          #NEED TO SEE IF I HAVE TO USE RELU IN BOTH FC1 AND FC2\n",
    "        x = self.dropout(x)\n",
    "        x = self.fc2(x)\n",
    "        \n",
    "        x = x.view(-1,self.conv_dim*16, 1, 1)\n",
    "        \n",
    "        x = F.relu(self.conv1(x))\n",
    "        x = F.relu(self.conv2(x))\n",
    "        x = F.relu(self.conv3(x))\n",
    "        x = F.relu(self.conv4(x))\n",
    "        x = F.tanh(self.conv5(x))\n",
    "        \n",
    "        return x\n",
    "\n",
    "\n",
    "tests.test_generator(Generator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize the weights of your networks\n",
    "\n",
    "To help our models converge, we should initialize the weights of the convolutional and linear layers in our 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, next we will be to define a weight initialization function that does just this!\n",
    "\n",
    "You can refer the code on initializing weights from [the `networks.py` file in CycleGAN Github repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/blob/master/models/networks.py) to understand this function.\n",
    "\n",
    "#### Defining 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": 2,
   "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",
    "    if hasattr(m, 'weight') and (classname.find('Linear')!=-1 or classname.find('Conv')!=-1):\n",
    "        m.weight.data.normal_(0, 0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Build complete network\n",
    "\n",
    "Define the model's 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": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "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": [
    "#### Defining model hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "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",
      "  (conv4): Sequential(\n",
      "    (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv5): Sequential(\n",
      "    (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (fc1): Linear(in_features=512, out_features=64, bias=True)\n",
      "  (fc2): Linear(in_features=64, out_features=1, bias=True)\n",
      "  (dropout): Dropout(p=0.2)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fc1): Linear(in_features=100, out_features=128, bias=True)\n",
      "  (fc2): Linear(in_features=128, out_features=512, bias=True)\n",
      "  (conv1): Sequential(\n",
      "    (0): ConvTranspose2d(512, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv2): Sequential(\n",
      "    (0): ConvTranspose2d(256, 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",
      "  (conv3): 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",
      "  (conv4): 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",
      "  (conv5): Sequential(\n",
      "    (0): ConvTranspose2d(32, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (dropout): Dropout(p=0.2)\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 32\n",
    "g_conv_dim = 32\n",
    "z_size = 100\n",
    "\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": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training on GPU!\n"
     ]
    }
   ],
   "source": [
    "# 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",
    "#### Defining the real and fake loss functions\n",
    "\n",
    "**We can 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": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def real_loss(D_out, smooth=False):\n",
    "    '''Calculates how close discriminator outputs are to being real.\n",
    "       param, D_out: discriminator logits\n",
    "       return: real loss'''\n",
    "    batch_size = D_out.size(0)\n",
    "    \n",
    "    if smooth:\n",
    "        labels = torch.ones(batch_size)*0.9\n",
    "    else:\n",
    "        labels = torch.ones(batch_size)\n",
    "    \n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "    \n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    loss = criterion(D_out.squeeze(), labels)\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",
    "    batch_size = D_out.size(0)\n",
    "    labels = torch.zeros(batch_size)\n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "        \n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    loss = criterion(D_out.squeeze(), labels)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizers\n",
    "\n",
    "#### Defining optimizers for your Discriminator (D) and Generator (G)\n",
    "\n",
    "Define optimizers for the models with appropriate hyperparameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create optimizers for the discriminator D and generator G\n",
    "lr = 0.0002\n",
    "beta1 = 0.3\n",
    "beta2 = 0.999\n",
    "\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. we will use your functions `real_loss` and `fake_loss` to help calculate the discriminator losses.\n",
    "\n",
    "* We 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",
    "We will print out some loss statistics and save some generated \"fake\" samples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Defining 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": 5,
   "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",
    "            #         TRAIN THE NETWORKS\n",
    "            # ===============================================\n",
    "            \n",
    "            # 1. Train the discriminator on real and fake images\n",
    "            fake_z = torch.from_numpy(np.random.uniform(-1, 1, size=(batch_size, z_size))).float()\n",
    "            if train_on_gpu:\n",
    "                fake_z = fake_z.cuda()\n",
    "                real_images = real_images.cuda()\n",
    "                \n",
    "            fake_images = G.forward(fake_z)\n",
    "            \n",
    "            d_loss = real_loss(D.forward(real_images)) + fake_loss(D.forward(fake_images))\n",
    "            \n",
    "            d_optimizer.zero_grad()\n",
    "            d_loss.backward()\n",
    "            d_optimizer.step()\n",
    "\n",
    "            # 2. Train the generator with an adversarial loss\n",
    "            fake_z = torch.from_numpy(np.random.uniform(-1, 1, size=(batch_size, z_size))).float()\n",
    "            if train_on_gpu:\n",
    "                fake_z = fake_z.cuda()\n",
    "                \n",
    "            \n",
    "            fake_images = G.forward(fake_z)\n",
    "            g_loss = real_loss(D.forward(fake_images))\n",
    "            \n",
    "            g_optimizer.zero_grad()\n",
    "            g_loss.backward()\n",
    "            g_optimizer.step()\n",
    "            \n",
    "            \n",
    "            # ===============================================\n",
    "            #              END OF TRAINING NETWORKS\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",
    "        samples_z = G(fixed_z)\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": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   20] | d_loss: 2.6321 | g_loss: 1.1448\n",
      "Epoch [    1/   20] | d_loss: 0.8535 | g_loss: 1.6881\n",
      "Epoch [    1/   20] | d_loss: 0.7014 | g_loss: 1.8188\n",
      "Epoch [    1/   20] | d_loss: 0.5177 | g_loss: 2.3932\n",
      "Epoch [    1/   20] | d_loss: 0.3770 | g_loss: 2.6380\n",
      "Epoch [    1/   20] | d_loss: 0.3922 | g_loss: 2.8544\n",
      "Epoch [    1/   20] | d_loss: 0.2165 | g_loss: 3.2963\n",
      "Epoch [    1/   20] | d_loss: 0.2296 | g_loss: 3.6472\n",
      "Epoch [    1/   20] | d_loss: 0.1852 | g_loss: 3.1079\n",
      "Epoch [    1/   20] | d_loss: 0.1507 | g_loss: 3.5492\n",
      "Epoch [    1/   20] | d_loss: 0.2034 | g_loss: 3.4404\n",
      "Epoch [    1/   20] | d_loss: 0.1863 | g_loss: 3.5509\n",
      "Epoch [    1/   20] | d_loss: 0.1249 | g_loss: 3.6662\n",
      "Epoch [    1/   20] | d_loss: 0.1613 | g_loss: 4.1597\n",
      "Epoch [    1/   20] | d_loss: 0.1343 | g_loss: 3.7701\n",
      "Epoch [    1/   20] | d_loss: 0.1140 | g_loss: 3.8668\n",
      "Epoch [    1/   20] | d_loss: 0.2482 | g_loss: 3.8251\n",
      "Epoch [    1/   20] | d_loss: 0.1103 | g_loss: 4.3049\n",
      "Epoch [    1/   20] | d_loss: 0.1865 | g_loss: 3.9822\n",
      "Epoch [    1/   20] | d_loss: 0.1117 | g_loss: 3.7069\n",
      "Epoch [    1/   20] | d_loss: 0.0755 | g_loss: 4.4358\n",
      "Epoch [    1/   20] | d_loss: 0.0620 | g_loss: 5.0347\n",
      "Epoch [    1/   20] | d_loss: 0.1322 | g_loss: 4.0754\n",
      "Epoch [    1/   20] | d_loss: 0.1732 | g_loss: 4.0714\n",
      "Epoch [    1/   20] | d_loss: 0.0586 | g_loss: 4.8117\n",
      "Epoch [    1/   20] | d_loss: 0.1073 | g_loss: 4.6473\n",
      "Epoch [    1/   20] | d_loss: 0.0787 | g_loss: 4.5750\n",
      "Epoch [    1/   20] | d_loss: 0.2887 | g_loss: 4.2717\n",
      "Epoch [    1/   20] | d_loss: 0.0902 | g_loss: 5.1367\n",
      "Epoch [    1/   20] | d_loss: 0.1777 | g_loss: 3.5487\n",
      "Epoch [    1/   20] | d_loss: 0.1188 | g_loss: 4.4571\n",
      "Epoch [    1/   20] | d_loss: 0.0740 | g_loss: 3.7102\n",
      "Epoch [    1/   20] | d_loss: 0.0794 | g_loss: 4.4541\n",
      "Epoch [    1/   20] | d_loss: 0.1558 | g_loss: 4.5834\n",
      "Epoch [    1/   20] | d_loss: 0.0962 | g_loss: 3.7376\n",
      "Epoch [    1/   20] | d_loss: 0.3319 | g_loss: 3.9455\n",
      "Epoch [    1/   20] | d_loss: 0.1349 | g_loss: 4.1339\n",
      "Epoch [    1/   20] | d_loss: 0.1941 | g_loss: 3.8846\n",
      "Epoch [    1/   20] | d_loss: 0.0413 | g_loss: 4.5176\n",
      "Epoch [    1/   20] | d_loss: 0.3054 | g_loss: 3.2657\n",
      "Epoch [    1/   20] | d_loss: 0.1029 | g_loss: 4.3890\n",
      "Epoch [    1/   20] | d_loss: 0.1151 | g_loss: 3.7921\n",
      "Epoch [    1/   20] | d_loss: 0.1488 | g_loss: 5.1732\n",
      "Epoch [    1/   20] | d_loss: 1.0345 | g_loss: 0.6311\n",
      "Epoch [    1/   20] | d_loss: 0.1918 | g_loss: 4.1106\n",
      "Epoch [    1/   20] | d_loss: 0.0943 | g_loss: 4.4074\n",
      "Epoch [    1/   20] | d_loss: 0.2352 | g_loss: 3.8845\n",
      "Epoch [    1/   20] | d_loss: 0.2758 | g_loss: 4.0232\n",
      "Epoch [    1/   20] | d_loss: 0.2688 | g_loss: 2.6051\n",
      "Epoch [    1/   20] | d_loss: 0.4689 | g_loss: 3.9864\n",
      "Epoch [    1/   20] | d_loss: 0.3022 | g_loss: 3.2353\n",
      "Epoch [    1/   20] | d_loss: 0.2380 | g_loss: 3.0028\n",
      "Epoch [    1/   20] | d_loss: 0.4157 | g_loss: 4.9667\n",
      "Epoch [    1/   20] | d_loss: 0.3943 | g_loss: 4.2094\n",
      "Epoch [    1/   20] | d_loss: 0.4639 | g_loss: 2.9216\n",
      "Epoch [    1/   20] | d_loss: 0.4262 | g_loss: 4.0152\n",
      "Epoch [    1/   20] | d_loss: 0.1741 | g_loss: 3.5619\n",
      "Epoch [    2/   20] | d_loss: 0.3977 | g_loss: 3.2721\n",
      "Epoch [    2/   20] | d_loss: 0.4932 | g_loss: 2.4025\n",
      "Epoch [    2/   20] | d_loss: 0.3353 | g_loss: 4.4710\n",
      "Epoch [    2/   20] | d_loss: 0.2950 | g_loss: 3.2766\n",
      "Epoch [    2/   20] | d_loss: 0.5527 | g_loss: 2.5489\n",
      "Epoch [    2/   20] | d_loss: 0.5201 | g_loss: 3.6320\n",
      "Epoch [    2/   20] | d_loss: 0.2917 | g_loss: 2.3019\n",
      "Epoch [    2/   20] | d_loss: 0.2845 | g_loss: 2.9903\n",
      "Epoch [    2/   20] | d_loss: 0.2979 | g_loss: 2.7296\n",
      "Epoch [    2/   20] | d_loss: 0.4156 | g_loss: 2.8233\n",
      "Epoch [    2/   20] | d_loss: 0.8990 | g_loss: 1.3843\n",
      "Epoch [    2/   20] | d_loss: 0.4799 | g_loss: 2.3420\n",
      "Epoch [    2/   20] | d_loss: 0.5637 | g_loss: 2.1190\n",
      "Epoch [    2/   20] | d_loss: 0.6097 | g_loss: 1.8563\n",
      "Epoch [    2/   20] | d_loss: 0.6471 | g_loss: 1.9838\n",
      "Epoch [    2/   20] | d_loss: 0.4987 | g_loss: 3.2009\n",
      "Epoch [    2/   20] | d_loss: 0.5852 | g_loss: 3.0278\n",
      "Epoch [    2/   20] | d_loss: 0.4764 | g_loss: 2.3132\n",
      "Epoch [    2/   20] | d_loss: 0.7625 | g_loss: 2.5413\n",
      "Epoch [    2/   20] | d_loss: 0.3639 | g_loss: 2.3991\n",
      "Epoch [    2/   20] | d_loss: 0.4047 | g_loss: 2.4204\n",
      "Epoch [    2/   20] | d_loss: 0.7026 | g_loss: 2.2135\n",
      "Epoch [    2/   20] | d_loss: 0.5082 | g_loss: 2.1424\n",
      "Epoch [    2/   20] | d_loss: 0.4416 | g_loss: 2.4267\n",
      "Epoch [    2/   20] | d_loss: 0.4440 | g_loss: 2.9103\n",
      "Epoch [    2/   20] | d_loss: 0.4876 | g_loss: 2.7510\n",
      "Epoch [    2/   20] | d_loss: 0.5430 | g_loss: 2.4895\n",
      "Epoch [    2/   20] | d_loss: 0.6860 | g_loss: 2.0602\n",
      "Epoch [    2/   20] | d_loss: 0.2569 | g_loss: 2.7133\n",
      "Epoch [    2/   20] | d_loss: 0.5362 | g_loss: 2.2228\n",
      "Epoch [    2/   20] | d_loss: 0.9571 | g_loss: 1.5698\n",
      "Epoch [    2/   20] | d_loss: 0.8479 | g_loss: 3.3252\n",
      "Epoch [    2/   20] | d_loss: 0.6204 | g_loss: 2.4249\n",
      "Epoch [    2/   20] | d_loss: 0.7630 | g_loss: 1.8514\n",
      "Epoch [    2/   20] | d_loss: 0.6227 | g_loss: 1.9779\n",
      "Epoch [    2/   20] | d_loss: 0.3577 | g_loss: 2.3638\n",
      "Epoch [    2/   20] | d_loss: 0.5467 | g_loss: 1.9880\n",
      "Epoch [    2/   20] | d_loss: 0.3708 | g_loss: 2.8827\n",
      "Epoch [    2/   20] | d_loss: 0.7605 | g_loss: 1.9360\n",
      "Epoch [    2/   20] | d_loss: 1.0589 | g_loss: 2.5625\n",
      "Epoch [    2/   20] | d_loss: 0.7242 | g_loss: 2.0284\n",
      "Epoch [    2/   20] | d_loss: 0.8028 | g_loss: 1.7763\n",
      "Epoch [    2/   20] | d_loss: 0.6147 | g_loss: 2.6661\n",
      "Epoch [    2/   20] | d_loss: 0.4612 | g_loss: 2.0790\n",
      "Epoch [    2/   20] | d_loss: 0.7649 | g_loss: 1.6950\n",
      "Epoch [    2/   20] | d_loss: 0.8277 | g_loss: 1.4426\n",
      "Epoch [    2/   20] | d_loss: 0.7180 | g_loss: 2.3234\n",
      "Epoch [    2/   20] | d_loss: 0.6005 | g_loss: 1.4612\n",
      "Epoch [    2/   20] | d_loss: 0.6818 | g_loss: 2.1728\n",
      "Epoch [    2/   20] | d_loss: 0.3993 | g_loss: 1.9835\n",
      "Epoch [    2/   20] | d_loss: 0.7880 | g_loss: 2.2527\n",
      "Epoch [    2/   20] | d_loss: 0.5467 | g_loss: 2.5877\n",
      "Epoch [    2/   20] | d_loss: 0.3899 | g_loss: 2.7746\n",
      "Epoch [    2/   20] | d_loss: 0.4814 | g_loss: 1.9359\n",
      "Epoch [    2/   20] | d_loss: 0.8045 | g_loss: 2.2382\n",
      "Epoch [    2/   20] | d_loss: 0.7569 | g_loss: 2.0976\n",
      "Epoch [    2/   20] | d_loss: 0.5265 | g_loss: 2.4395\n",
      "Epoch [    3/   20] | d_loss: 0.5359 | g_loss: 2.3065\n",
      "Epoch [    3/   20] | d_loss: 0.7906 | g_loss: 1.9342\n",
      "Epoch [    3/   20] | d_loss: 0.4620 | g_loss: 1.5667\n",
      "Epoch [    3/   20] | d_loss: 1.0388 | g_loss: 1.9356\n",
      "Epoch [    3/   20] | d_loss: 1.2436 | g_loss: 1.6928\n",
      "Epoch [    3/   20] | d_loss: 0.8709 | g_loss: 2.3228\n",
      "Epoch [    3/   20] | d_loss: 0.6522 | g_loss: 1.7011\n",
      "Epoch [    3/   20] | d_loss: 0.9595 | g_loss: 1.4782\n",
      "Epoch [    3/   20] | d_loss: 0.6122 | g_loss: 2.3003\n",
      "Epoch [    3/   20] | d_loss: 0.9325 | g_loss: 2.4501\n",
      "Epoch [    3/   20] | d_loss: 0.9460 | g_loss: 2.7668\n",
      "Epoch [    3/   20] | d_loss: 0.9400 | g_loss: 2.1740\n",
      "Epoch [    3/   20] | d_loss: 0.4022 | g_loss: 2.0088\n",
      "Epoch [    3/   20] | d_loss: 0.6864 | g_loss: 2.3983\n",
      "Epoch [    3/   20] | d_loss: 0.6627 | g_loss: 2.2973\n",
      "Epoch [    3/   20] | d_loss: 0.5895 | g_loss: 2.9333\n",
      "Epoch [    3/   20] | d_loss: 0.9593 | g_loss: 1.9088\n",
      "Epoch [    3/   20] | d_loss: 0.8265 | g_loss: 2.2325\n",
      "Epoch [    3/   20] | d_loss: 0.7284 | g_loss: 2.2575\n",
      "Epoch [    3/   20] | d_loss: 0.5880 | g_loss: 1.2802\n",
      "Epoch [    3/   20] | d_loss: 0.9350 | g_loss: 1.9549\n",
      "Epoch [    3/   20] | d_loss: 0.5619 | g_loss: 2.0805\n",
      "Epoch [    3/   20] | d_loss: 0.5875 | g_loss: 2.2176\n",
      "Epoch [    3/   20] | d_loss: 0.9525 | g_loss: 1.5223\n",
      "Epoch [    3/   20] | d_loss: 0.8286 | g_loss: 2.1442\n",
      "Epoch [    3/   20] | d_loss: 0.8077 | g_loss: 2.2202\n",
      "Epoch [    3/   20] | d_loss: 0.9133 | g_loss: 1.9003\n",
      "Epoch [    3/   20] | d_loss: 0.5069 | g_loss: 1.8556\n",
      "Epoch [    3/   20] | d_loss: 1.0154 | g_loss: 2.5290\n",
      "Epoch [    3/   20] | d_loss: 0.6677 | g_loss: 1.5342\n",
      "Epoch [    3/   20] | d_loss: 0.4851 | g_loss: 2.0564\n",
      "Epoch [    3/   20] | d_loss: 0.7000 | g_loss: 0.9337\n",
      "Epoch [    3/   20] | d_loss: 1.1938 | g_loss: 1.8289\n",
      "Epoch [    3/   20] | d_loss: 0.6346 | g_loss: 1.0078\n",
      "Epoch [    3/   20] | d_loss: 0.8603 | g_loss: 2.0691\n",
      "Epoch [    3/   20] | d_loss: 0.5634 | g_loss: 1.8529\n",
      "Epoch [    3/   20] | d_loss: 0.6837 | g_loss: 2.2178\n",
      "Epoch [    3/   20] | d_loss: 0.8627 | g_loss: 1.7888\n",
      "Epoch [    3/   20] | d_loss: 0.9896 | g_loss: 2.6147\n",
      "Epoch [    3/   20] | d_loss: 0.5488 | g_loss: 1.6564\n",
      "Epoch [    3/   20] | d_loss: 0.9339 | g_loss: 1.5232\n",
      "Epoch [    3/   20] | d_loss: 0.9448 | g_loss: 1.7184\n",
      "Epoch [    3/   20] | d_loss: 0.8139 | g_loss: 1.5345\n",
      "Epoch [    3/   20] | d_loss: 0.8875 | g_loss: 1.9971\n",
      "Epoch [    3/   20] | d_loss: 0.8616 | g_loss: 1.7720\n",
      "Epoch [    3/   20] | d_loss: 0.6258 | g_loss: 1.6515\n",
      "Epoch [    3/   20] | d_loss: 0.7068 | g_loss: 1.7338\n",
      "Epoch [    3/   20] | d_loss: 0.7616 | g_loss: 1.3125\n",
      "Epoch [    3/   20] | d_loss: 0.8091 | g_loss: 1.7073\n",
      "Epoch [    3/   20] | d_loss: 0.8184 | g_loss: 2.4601\n",
      "Epoch [    3/   20] | d_loss: 0.7263 | g_loss: 2.0273\n",
      "Epoch [    3/   20] | d_loss: 0.7660 | g_loss: 1.6263\n",
      "Epoch [    3/   20] | d_loss: 0.7371 | g_loss: 1.7783\n",
      "Epoch [    3/   20] | d_loss: 0.9355 | g_loss: 1.1534\n",
      "Epoch [    3/   20] | d_loss: 1.2471 | g_loss: 1.7357\n",
      "Epoch [    3/   20] | d_loss: 1.2681 | g_loss: 1.9886\n",
      "Epoch [    3/   20] | d_loss: 0.8445 | g_loss: 1.4748\n",
      "Epoch [    4/   20] | d_loss: 1.0350 | g_loss: 1.5110\n",
      "Epoch [    4/   20] | d_loss: 0.5397 | g_loss: 2.1232\n",
      "Epoch [    4/   20] | d_loss: 0.7256 | g_loss: 1.8288\n",
      "Epoch [    4/   20] | d_loss: 0.7238 | g_loss: 1.2570\n",
      "Epoch [    4/   20] | d_loss: 0.7773 | g_loss: 1.9650\n",
      "Epoch [    4/   20] | d_loss: 0.6427 | g_loss: 1.7414\n",
      "Epoch [    4/   20] | d_loss: 0.7190 | g_loss: 1.2564\n",
      "Epoch [    4/   20] | d_loss: 0.5337 | g_loss: 1.8722\n",
      "Epoch [    4/   20] | d_loss: 1.0837 | g_loss: 1.2936\n",
      "Epoch [    4/   20] | d_loss: 0.7649 | g_loss: 1.5304\n",
      "Epoch [    4/   20] | d_loss: 0.9731 | g_loss: 1.9793\n",
      "Epoch [    4/   20] | d_loss: 1.1432 | g_loss: 1.7605\n",
      "Epoch [    4/   20] | d_loss: 0.6146 | g_loss: 1.6278\n",
      "Epoch [    4/   20] | d_loss: 0.8613 | g_loss: 2.0567\n",
      "Epoch [    4/   20] | d_loss: 0.7370 | g_loss: 1.6134\n",
      "Epoch [    4/   20] | d_loss: 0.9237 | g_loss: 2.0779\n",
      "Epoch [    4/   20] | d_loss: 0.8913 | g_loss: 1.7242\n",
      "Epoch [    4/   20] | d_loss: 0.9055 | g_loss: 1.6532\n",
      "Epoch [    4/   20] | d_loss: 0.8318 | g_loss: 2.1232\n",
      "Epoch [    4/   20] | d_loss: 0.9252 | g_loss: 1.8001\n",
      "Epoch [    4/   20] | d_loss: 0.7974 | g_loss: 1.8169\n",
      "Epoch [    4/   20] | d_loss: 0.5206 | g_loss: 2.2866\n",
      "Epoch [    4/   20] | d_loss: 0.7403 | g_loss: 1.6823\n",
      "Epoch [    4/   20] | d_loss: 0.4042 | g_loss: 2.1786\n",
      "Epoch [    4/   20] | d_loss: 0.6660 | g_loss: 1.8628\n",
      "Epoch [    4/   20] | d_loss: 1.1587 | g_loss: 1.7020\n",
      "Epoch [    4/   20] | d_loss: 0.8930 | g_loss: 2.0275\n",
      "Epoch [    4/   20] | d_loss: 0.9431 | g_loss: 1.6717\n",
      "Epoch [    4/   20] | d_loss: 0.6943 | g_loss: 1.7014\n",
      "Epoch [    4/   20] | d_loss: 0.7834 | g_loss: 1.2850\n",
      "Epoch [    4/   20] | d_loss: 0.8720 | g_loss: 1.8167\n",
      "Epoch [    4/   20] | d_loss: 1.0884 | g_loss: 2.5070\n",
      "Epoch [    4/   20] | d_loss: 1.1935 | g_loss: 1.3763\n",
      "Epoch [    4/   20] | d_loss: 0.7322 | g_loss: 1.7202\n",
      "Epoch [    4/   20] | d_loss: 0.8188 | g_loss: 1.9680\n",
      "Epoch [    4/   20] | d_loss: 0.8339 | g_loss: 1.4825\n",
      "Epoch [    4/   20] | d_loss: 0.5464 | g_loss: 1.9228\n",
      "Epoch [    4/   20] | d_loss: 0.8137 | g_loss: 1.7299\n",
      "Epoch [    4/   20] | d_loss: 0.7900 | g_loss: 1.9176\n",
      "Epoch [    4/   20] | d_loss: 0.8780 | g_loss: 1.7258\n",
      "Epoch [    4/   20] | d_loss: 0.8689 | g_loss: 2.3880\n",
      "Epoch [    4/   20] | d_loss: 0.6863 | g_loss: 1.8275\n",
      "Epoch [    4/   20] | d_loss: 0.8515 | g_loss: 1.2106\n",
      "Epoch [    4/   20] | d_loss: 0.9572 | g_loss: 1.3939\n",
      "Epoch [    4/   20] | d_loss: 0.7464 | g_loss: 2.0267\n",
      "Epoch [    4/   20] | d_loss: 0.8474 | g_loss: 2.0153\n",
      "Epoch [    4/   20] | d_loss: 0.7790 | g_loss: 2.3663\n",
      "Epoch [    4/   20] | d_loss: 0.9939 | g_loss: 1.4170\n",
      "Epoch [    4/   20] | d_loss: 0.7708 | g_loss: 1.7340\n",
      "Epoch [    4/   20] | d_loss: 0.7970 | g_loss: 1.7654\n",
      "Epoch [    4/   20] | d_loss: 0.9574 | g_loss: 1.7506\n",
      "Epoch [    4/   20] | d_loss: 0.9722 | g_loss: 1.5411\n",
      "Epoch [    4/   20] | d_loss: 0.8870 | g_loss: 1.5232\n",
      "Epoch [    4/   20] | d_loss: 0.6274 | g_loss: 2.1961\n",
      "Epoch [    4/   20] | d_loss: 1.0701 | g_loss: 1.6487\n",
      "Epoch [    4/   20] | d_loss: 0.6818 | g_loss: 1.6282\n",
      "Epoch [    4/   20] | d_loss: 0.6516 | g_loss: 1.6643\n",
      "Epoch [    5/   20] | d_loss: 0.7237 | g_loss: 2.1762\n",
      "Epoch [    5/   20] | d_loss: 0.7958 | g_loss: 1.1483\n",
      "Epoch [    5/   20] | d_loss: 0.6876 | g_loss: 1.5704\n",
      "Epoch [    5/   20] | d_loss: 0.6889 | g_loss: 1.2248\n",
      "Epoch [    5/   20] | d_loss: 0.6553 | g_loss: 1.6406\n",
      "Epoch [    5/   20] | d_loss: 0.7252 | g_loss: 2.1167\n",
      "Epoch [    5/   20] | d_loss: 0.7618 | g_loss: 1.7597\n",
      "Epoch [    5/   20] | d_loss: 0.9587 | g_loss: 1.9171\n",
      "Epoch [    5/   20] | d_loss: 0.7646 | g_loss: 1.6272\n",
      "Epoch [    5/   20] | d_loss: 0.6773 | g_loss: 1.4583\n",
      "Epoch [    5/   20] | d_loss: 1.1751 | g_loss: 1.5394\n",
      "Epoch [    5/   20] | d_loss: 0.7007 | g_loss: 1.6673\n",
      "Epoch [    5/   20] | d_loss: 0.7113 | g_loss: 1.7016\n",
      "Epoch [    5/   20] | d_loss: 0.8320 | g_loss: 1.4895\n",
      "Epoch [    5/   20] | d_loss: 0.8107 | g_loss: 2.4641\n",
      "Epoch [    5/   20] | d_loss: 1.1412 | g_loss: 1.5228\n",
      "Epoch [    5/   20] | d_loss: 0.8839 | g_loss: 1.4978\n",
      "Epoch [    5/   20] | d_loss: 0.8480 | g_loss: 1.9482\n",
      "Epoch [    5/   20] | d_loss: 0.6898 | g_loss: 1.8871\n",
      "Epoch [    5/   20] | d_loss: 1.2072 | g_loss: 1.5817\n",
      "Epoch [    5/   20] | d_loss: 1.0344 | g_loss: 1.7392\n",
      "Epoch [    5/   20] | d_loss: 0.6056 | g_loss: 1.5748\n",
      "Epoch [    5/   20] | d_loss: 1.0129 | g_loss: 2.3278\n",
      "Epoch [    5/   20] | d_loss: 0.8846 | g_loss: 0.9256\n",
      "Epoch [    5/   20] | d_loss: 0.8846 | g_loss: 2.0320\n",
      "Epoch [    5/   20] | d_loss: 0.8806 | g_loss: 1.5381\n",
      "Epoch [    5/   20] | d_loss: 0.6890 | g_loss: 1.5066\n",
      "Epoch [    5/   20] | d_loss: 0.9998 | g_loss: 1.5365\n",
      "Epoch [    5/   20] | d_loss: 0.9454 | g_loss: 1.9572\n",
      "Epoch [    5/   20] | d_loss: 0.9733 | g_loss: 1.5601\n",
      "Epoch [    5/   20] | d_loss: 0.8159 | g_loss: 2.5314\n",
      "Epoch [    5/   20] | d_loss: 0.6465 | g_loss: 2.1690\n",
      "Epoch [    5/   20] | d_loss: 0.9696 | g_loss: 1.4804\n",
      "Epoch [    5/   20] | d_loss: 1.1736 | g_loss: 1.4276\n",
      "Epoch [    5/   20] | d_loss: 0.5068 | g_loss: 1.9342\n",
      "Epoch [    5/   20] | d_loss: 0.7452 | g_loss: 1.6905\n",
      "Epoch [    5/   20] | d_loss: 0.7253 | g_loss: 1.8555\n",
      "Epoch [    5/   20] | d_loss: 0.7572 | g_loss: 1.8809\n",
      "Epoch [    5/   20] | d_loss: 0.4738 | g_loss: 2.2810\n",
      "Epoch [    5/   20] | d_loss: 0.8834 | g_loss: 1.7167\n",
      "Epoch [    5/   20] | d_loss: 0.8019 | g_loss: 2.2950\n",
      "Epoch [    5/   20] | d_loss: 0.5388 | g_loss: 2.0403\n",
      "Epoch [    5/   20] | d_loss: 0.9004 | g_loss: 1.9711\n",
      "Epoch [    5/   20] | d_loss: 0.8954 | g_loss: 1.6157\n",
      "Epoch [    5/   20] | d_loss: 1.0644 | g_loss: 1.5562\n",
      "Epoch [    5/   20] | d_loss: 1.1611 | g_loss: 1.5166\n",
      "Epoch [    5/   20] | d_loss: 0.7791 | g_loss: 1.2976\n",
      "Epoch [    5/   20] | d_loss: 0.8976 | g_loss: 1.6557\n",
      "Epoch [    5/   20] | d_loss: 0.6490 | g_loss: 2.6248\n",
      "Epoch [    5/   20] | d_loss: 0.8828 | g_loss: 2.1774\n",
      "Epoch [    5/   20] | d_loss: 0.6427 | g_loss: 1.8494\n",
      "Epoch [    5/   20] | d_loss: 0.7843 | g_loss: 2.4488\n",
      "Epoch [    5/   20] | d_loss: 0.7134 | g_loss: 1.9503\n",
      "Epoch [    5/   20] | d_loss: 0.5281 | g_loss: 2.1671\n",
      "Epoch [    5/   20] | d_loss: 0.6992 | g_loss: 2.0209\n",
      "Epoch [    5/   20] | d_loss: 0.6305 | g_loss: 2.1558\n",
      "Epoch [    5/   20] | d_loss: 0.6179 | g_loss: 2.3758\n",
      "Epoch [    6/   20] | d_loss: 0.9110 | g_loss: 1.6471\n",
      "Epoch [    6/   20] | d_loss: 0.5945 | g_loss: 1.8821\n",
      "Epoch [    6/   20] | d_loss: 1.0673 | g_loss: 1.9797\n",
      "Epoch [    6/   20] | d_loss: 0.8761 | g_loss: 2.0430\n",
      "Epoch [    6/   20] | d_loss: 0.9055 | g_loss: 1.3121\n",
      "Epoch [    6/   20] | d_loss: 0.8923 | g_loss: 2.1780\n",
      "Epoch [    6/   20] | d_loss: 0.7213 | g_loss: 2.3483\n",
      "Epoch [    6/   20] | d_loss: 1.0584 | g_loss: 1.5423\n",
      "Epoch [    6/   20] | d_loss: 0.7084 | g_loss: 2.1435\n",
      "Epoch [    6/   20] | d_loss: 0.6293 | g_loss: 2.0792\n",
      "Epoch [    6/   20] | d_loss: 1.0880 | g_loss: 1.7714\n",
      "Epoch [    6/   20] | d_loss: 1.0929 | g_loss: 1.2930\n",
      "Epoch [    6/   20] | d_loss: 0.8503 | g_loss: 1.8478\n",
      "Epoch [    6/   20] | d_loss: 1.0179 | g_loss: 1.5686\n",
      "Epoch [    6/   20] | d_loss: 0.6996 | g_loss: 2.2256\n",
      "Epoch [    6/   20] | d_loss: 0.6734 | g_loss: 2.0805\n",
      "Epoch [    6/   20] | d_loss: 0.9848 | g_loss: 1.6859\n",
      "Epoch [    6/   20] | d_loss: 0.7257 | g_loss: 1.7260\n",
      "Epoch [    6/   20] | d_loss: 0.6160 | g_loss: 1.6956\n",
      "Epoch [    6/   20] | d_loss: 0.8353 | g_loss: 2.3081\n",
      "Epoch [    6/   20] | d_loss: 0.5312 | g_loss: 2.2031\n",
      "Epoch [    6/   20] | d_loss: 0.4502 | g_loss: 1.5313\n",
      "Epoch [    6/   20] | d_loss: 0.7570 | g_loss: 1.6344\n",
      "Epoch [    6/   20] | d_loss: 0.6676 | g_loss: 1.8414\n",
      "Epoch [    6/   20] | d_loss: 0.8496 | g_loss: 2.0350\n",
      "Epoch [    6/   20] | d_loss: 0.9641 | g_loss: 2.0317\n",
      "Epoch [    6/   20] | d_loss: 0.8749 | g_loss: 2.1353\n",
      "Epoch [    6/   20] | d_loss: 0.8853 | g_loss: 1.7559\n",
      "Epoch [    6/   20] | d_loss: 0.7900 | g_loss: 1.6574\n",
      "Epoch [    6/   20] | d_loss: 0.6745 | g_loss: 1.9461\n",
      "Epoch [    6/   20] | d_loss: 0.7540 | g_loss: 2.0256\n",
      "Epoch [    6/   20] | d_loss: 0.6550 | g_loss: 1.9909\n",
      "Epoch [    6/   20] | d_loss: 0.7050 | g_loss: 1.3521\n",
      "Epoch [    6/   20] | d_loss: 0.7006 | g_loss: 1.6148\n",
      "Epoch [    6/   20] | d_loss: 0.7932 | g_loss: 2.2926\n",
      "Epoch [    6/   20] | d_loss: 0.5077 | g_loss: 1.9625\n",
      "Epoch [    6/   20] | d_loss: 0.9641 | g_loss: 2.1678\n",
      "Epoch [    6/   20] | d_loss: 0.8641 | g_loss: 1.8500\n",
      "Epoch [    6/   20] | d_loss: 0.5824 | g_loss: 2.2392\n",
      "Epoch [    6/   20] | d_loss: 1.0083 | g_loss: 2.3546\n",
      "Epoch [    6/   20] | d_loss: 0.7015 | g_loss: 1.3639\n",
      "Epoch [    6/   20] | d_loss: 1.1056 | g_loss: 2.2350\n",
      "Epoch [    6/   20] | d_loss: 0.7250 | g_loss: 1.4556\n",
      "Epoch [    6/   20] | d_loss: 0.6161 | g_loss: 2.0826\n",
      "Epoch [    6/   20] | d_loss: 0.6811 | g_loss: 1.9355\n",
      "Epoch [    6/   20] | d_loss: 0.4646 | g_loss: 2.1438\n",
      "Epoch [    6/   20] | d_loss: 0.7799 | g_loss: 2.6916\n",
      "Epoch [    6/   20] | d_loss: 0.7440 | g_loss: 1.1252\n",
      "Epoch [    6/   20] | d_loss: 0.6403 | g_loss: 2.3949\n",
      "Epoch [    6/   20] | d_loss: 0.6617 | g_loss: 1.6259\n",
      "Epoch [    6/   20] | d_loss: 0.8739 | g_loss: 2.3631\n",
      "Epoch [    6/   20] | d_loss: 0.5168 | g_loss: 1.9565\n",
      "Epoch [    6/   20] | d_loss: 1.0165 | g_loss: 1.9775\n",
      "Epoch [    6/   20] | d_loss: 0.6038 | g_loss: 1.9982\n",
      "Epoch [    6/   20] | d_loss: 0.8153 | g_loss: 2.1546\n",
      "Epoch [    6/   20] | d_loss: 1.0292 | g_loss: 1.2542\n",
      "Epoch [    6/   20] | d_loss: 0.5882 | g_loss: 2.0662\n",
      "Epoch [    7/   20] | d_loss: 0.6326 | g_loss: 2.0369\n",
      "Epoch [    7/   20] | d_loss: 0.9990 | g_loss: 1.6397\n",
      "Epoch [    7/   20] | d_loss: 0.8710 | g_loss: 2.4528\n",
      "Epoch [    7/   20] | d_loss: 0.7089 | g_loss: 2.3879\n",
      "Epoch [    7/   20] | d_loss: 0.7342 | g_loss: 1.9462\n",
      "Epoch [    7/   20] | d_loss: 0.9862 | g_loss: 2.0719\n",
      "Epoch [    7/   20] | d_loss: 0.7470 | g_loss: 1.5003\n",
      "Epoch [    7/   20] | d_loss: 0.8157 | g_loss: 1.8041\n",
      "Epoch [    7/   20] | d_loss: 0.6152 | g_loss: 1.9463\n",
      "Epoch [    7/   20] | d_loss: 0.6218 | g_loss: 2.2792\n",
      "Epoch [    7/   20] | d_loss: 0.7386 | g_loss: 3.0547\n",
      "Epoch [    7/   20] | d_loss: 0.5897 | g_loss: 2.1826\n",
      "Epoch [    7/   20] | d_loss: 0.5525 | g_loss: 2.1512\n",
      "Epoch [    7/   20] | d_loss: 0.5702 | g_loss: 2.1088\n",
      "Epoch [    7/   20] | d_loss: 1.1295 | g_loss: 2.0631\n",
      "Epoch [    7/   20] | d_loss: 0.9829 | g_loss: 1.8342\n",
      "Epoch [    7/   20] | d_loss: 0.8681 | g_loss: 1.5120\n",
      "Epoch [    7/   20] | d_loss: 0.6454 | g_loss: 1.9434\n",
      "Epoch [    7/   20] | d_loss: 0.8462 | g_loss: 1.8867\n",
      "Epoch [    7/   20] | d_loss: 0.7206 | g_loss: 1.5093\n",
      "Epoch [    7/   20] | d_loss: 0.7084 | g_loss: 1.9366\n",
      "Epoch [    7/   20] | d_loss: 0.9414 | g_loss: 2.1250\n",
      "Epoch [    7/   20] | d_loss: 0.8948 | g_loss: 2.2899\n",
      "Epoch [    7/   20] | d_loss: 0.7934 | g_loss: 2.3444\n",
      "Epoch [    7/   20] | d_loss: 0.4059 | g_loss: 2.5574\n",
      "Epoch [    7/   20] | d_loss: 0.9506 | g_loss: 1.5071\n",
      "Epoch [    7/   20] | d_loss: 0.8934 | g_loss: 1.4209\n",
      "Epoch [    7/   20] | d_loss: 0.7084 | g_loss: 1.8537\n",
      "Epoch [    7/   20] | d_loss: 0.8096 | g_loss: 3.1313\n",
      "Epoch [    7/   20] | d_loss: 0.5586 | g_loss: 1.7796\n",
      "Epoch [    7/   20] | d_loss: 0.5800 | g_loss: 2.2524\n",
      "Epoch [    7/   20] | d_loss: 0.8081 | g_loss: 2.5044\n",
      "Epoch [    7/   20] | d_loss: 0.5954 | g_loss: 1.8997\n",
      "Epoch [    7/   20] | d_loss: 0.7222 | g_loss: 1.8183\n",
      "Epoch [    7/   20] | d_loss: 0.7980 | g_loss: 1.5147\n",
      "Epoch [    7/   20] | d_loss: 0.6320 | g_loss: 2.3410\n",
      "Epoch [    7/   20] | d_loss: 0.7106 | g_loss: 1.9237\n",
      "Epoch [    7/   20] | d_loss: 0.5388 | g_loss: 2.0260\n",
      "Epoch [    7/   20] | d_loss: 0.5193 | g_loss: 1.9719\n",
      "Epoch [    7/   20] | d_loss: 0.6711 | g_loss: 1.5188\n",
      "Epoch [    7/   20] | d_loss: 0.4230 | g_loss: 2.0590\n",
      "Epoch [    7/   20] | d_loss: 0.7095 | g_loss: 2.2696\n",
      "Epoch [    7/   20] | d_loss: 0.5716 | g_loss: 1.9655\n",
      "Epoch [    7/   20] | d_loss: 0.6559 | g_loss: 2.6743\n",
      "Epoch [    7/   20] | d_loss: 0.7252 | g_loss: 2.8940\n",
      "Epoch [    7/   20] | d_loss: 0.8411 | g_loss: 2.1675\n",
      "Epoch [    7/   20] | d_loss: 0.7507 | g_loss: 2.7877\n",
      "Epoch [    7/   20] | d_loss: 0.7142 | g_loss: 2.7082\n",
      "Epoch [    7/   20] | d_loss: 0.7950 | g_loss: 1.9802\n",
      "Epoch [    7/   20] | d_loss: 0.7176 | g_loss: 2.0712\n",
      "Epoch [    7/   20] | d_loss: 0.4593 | g_loss: 1.4542\n",
      "Epoch [    7/   20] | d_loss: 0.9413 | g_loss: 2.2532\n",
      "Epoch [    7/   20] | d_loss: 0.5745 | g_loss: 1.9787\n",
      "Epoch [    7/   20] | d_loss: 0.9937 | g_loss: 1.7511\n",
      "Epoch [    7/   20] | d_loss: 0.7592 | g_loss: 2.1842\n",
      "Epoch [    7/   20] | d_loss: 0.9542 | g_loss: 1.6462\n",
      "Epoch [    7/   20] | d_loss: 0.3149 | g_loss: 2.2224\n",
      "Epoch [    8/   20] | d_loss: 0.6487 | g_loss: 1.7232\n",
      "Epoch [    8/   20] | d_loss: 1.1066 | g_loss: 2.7220\n",
      "Epoch [    8/   20] | d_loss: 0.7153 | g_loss: 2.1706\n",
      "Epoch [    8/   20] | d_loss: 0.8171 | g_loss: 1.3373\n",
      "Epoch [    8/   20] | d_loss: 0.9675 | g_loss: 2.1002\n",
      "Epoch [    8/   20] | d_loss: 0.8775 | g_loss: 2.1779\n",
      "Epoch [    8/   20] | d_loss: 0.7557 | g_loss: 2.0457\n",
      "Epoch [    8/   20] | d_loss: 0.7803 | g_loss: 2.0007\n",
      "Epoch [    8/   20] | d_loss: 0.9383 | g_loss: 1.7285\n",
      "Epoch [    8/   20] | d_loss: 0.7489 | g_loss: 2.4049\n",
      "Epoch [    8/   20] | d_loss: 0.8285 | g_loss: 2.4017\n",
      "Epoch [    8/   20] | d_loss: 0.8937 | g_loss: 1.8269\n",
      "Epoch [    8/   20] | d_loss: 0.7635 | g_loss: 2.0450\n",
      "Epoch [    8/   20] | d_loss: 0.6874 | g_loss: 2.6678\n",
      "Epoch [    8/   20] | d_loss: 0.5811 | g_loss: 2.5899\n",
      "Epoch [    8/   20] | d_loss: 0.7559 | g_loss: 2.0453\n",
      "Epoch [    8/   20] | d_loss: 0.7660 | g_loss: 1.3035\n",
      "Epoch [    8/   20] | d_loss: 0.7298 | g_loss: 2.2063\n",
      "Epoch [    8/   20] | d_loss: 0.9232 | g_loss: 2.1621\n",
      "Epoch [    8/   20] | d_loss: 0.8701 | g_loss: 2.7518\n",
      "Epoch [    8/   20] | d_loss: 0.4739 | g_loss: 1.4621\n",
      "Epoch [    8/   20] | d_loss: 0.8028 | g_loss: 1.9471\n",
      "Epoch [    8/   20] | d_loss: 0.4038 | g_loss: 2.5062\n",
      "Epoch [    8/   20] | d_loss: 0.8210 | g_loss: 2.9199\n",
      "Epoch [    8/   20] | d_loss: 0.9173 | g_loss: 1.7965\n",
      "Epoch [    8/   20] | d_loss: 1.3344 | g_loss: 2.6441\n",
      "Epoch [    8/   20] | d_loss: 0.4155 | g_loss: 1.7805\n",
      "Epoch [    8/   20] | d_loss: 0.7002 | g_loss: 1.6675\n",
      "Epoch [    8/   20] | d_loss: 0.9177 | g_loss: 1.7954\n",
      "Epoch [    8/   20] | d_loss: 0.6736 | g_loss: 3.0887\n",
      "Epoch [    8/   20] | d_loss: 0.4807 | g_loss: 2.1413\n",
      "Epoch [    8/   20] | d_loss: 0.7109 | g_loss: 1.7329\n",
      "Epoch [    8/   20] | d_loss: 0.6652 | g_loss: 1.5582\n",
      "Epoch [    8/   20] | d_loss: 0.6289 | g_loss: 2.6219\n",
      "Epoch [    8/   20] | d_loss: 0.7985 | g_loss: 2.1838\n",
      "Epoch [    8/   20] | d_loss: 1.1312 | g_loss: 1.6483\n",
      "Epoch [    8/   20] | d_loss: 0.8796 | g_loss: 2.2488\n",
      "Epoch [    8/   20] | d_loss: 0.5583 | g_loss: 2.2067\n",
      "Epoch [    8/   20] | d_loss: 0.6701 | g_loss: 1.7430\n",
      "Epoch [    8/   20] | d_loss: 0.5325 | g_loss: 3.0357\n",
      "Epoch [    8/   20] | d_loss: 1.0061 | g_loss: 1.4745\n",
      "Epoch [    8/   20] | d_loss: 0.5629 | g_loss: 2.2647\n",
      "Epoch [    8/   20] | d_loss: 0.5117 | g_loss: 1.8092\n",
      "Epoch [    8/   20] | d_loss: 0.4497 | g_loss: 2.4396\n",
      "Epoch [    8/   20] | d_loss: 0.4170 | g_loss: 2.5425\n",
      "Epoch [    8/   20] | d_loss: 0.7716 | g_loss: 2.4287\n",
      "Epoch [    8/   20] | d_loss: 0.5422 | g_loss: 1.6672\n",
      "Epoch [    8/   20] | d_loss: 0.9029 | g_loss: 2.0799\n",
      "Epoch [    8/   20] | d_loss: 0.7034 | g_loss: 2.7944\n",
      "Epoch [    8/   20] | d_loss: 0.6906 | g_loss: 3.4887\n",
      "Epoch [    8/   20] | d_loss: 1.0956 | g_loss: 1.8786\n",
      "Epoch [    8/   20] | d_loss: 0.5978 | g_loss: 2.6828\n",
      "Epoch [    8/   20] | d_loss: 1.0057 | g_loss: 1.9945\n",
      "Epoch [    8/   20] | d_loss: 0.7611 | g_loss: 2.1264\n",
      "Epoch [    8/   20] | d_loss: 0.5402 | g_loss: 2.2404\n",
      "Epoch [    8/   20] | d_loss: 0.9958 | g_loss: 1.4530\n",
      "Epoch [    8/   20] | d_loss: 0.7280 | g_loss: 2.6529\n",
      "Epoch [    9/   20] | d_loss: 0.9472 | g_loss: 2.2257\n",
      "Epoch [    9/   20] | d_loss: 0.7146 | g_loss: 1.9886\n",
      "Epoch [    9/   20] | d_loss: 0.7318 | g_loss: 2.2311\n",
      "Epoch [    9/   20] | d_loss: 0.5595 | g_loss: 1.9428\n",
      "Epoch [    9/   20] | d_loss: 0.6983 | g_loss: 2.7658\n",
      "Epoch [    9/   20] | d_loss: 0.6756 | g_loss: 2.7955\n",
      "Epoch [    9/   20] | d_loss: 0.7711 | g_loss: 1.8971\n",
      "Epoch [    9/   20] | d_loss: 0.9047 | g_loss: 1.7128\n",
      "Epoch [    9/   20] | d_loss: 0.8740 | g_loss: 2.5443\n",
      "Epoch [    9/   20] | d_loss: 0.5434 | g_loss: 2.4350\n",
      "Epoch [    9/   20] | d_loss: 0.4482 | g_loss: 1.1222\n",
      "Epoch [    9/   20] | d_loss: 0.6239 | g_loss: 2.6918\n",
      "Epoch [    9/   20] | d_loss: 0.6116 | g_loss: 2.2414\n",
      "Epoch [    9/   20] | d_loss: 0.9339 | g_loss: 2.4594\n",
      "Epoch [    9/   20] | d_loss: 0.8737 | g_loss: 2.2683\n",
      "Epoch [    9/   20] | d_loss: 0.5940 | g_loss: 1.8324\n",
      "Epoch [    9/   20] | d_loss: 0.8593 | g_loss: 1.7352\n",
      "Epoch [    9/   20] | d_loss: 0.6453 | g_loss: 3.2040\n",
      "Epoch [    9/   20] | d_loss: 0.7957 | g_loss: 1.8286\n",
      "Epoch [    9/   20] | d_loss: 0.5518 | g_loss: 2.5256\n",
      "Epoch [    9/   20] | d_loss: 0.6577 | g_loss: 1.4035\n",
      "Epoch [    9/   20] | d_loss: 0.4130 | g_loss: 1.7314\n",
      "Epoch [    9/   20] | d_loss: 0.6999 | g_loss: 2.1014\n",
      "Epoch [    9/   20] | d_loss: 0.8840 | g_loss: 2.7719\n",
      "Epoch [    9/   20] | d_loss: 0.7654 | g_loss: 3.4965\n",
      "Epoch [    9/   20] | d_loss: 0.6716 | g_loss: 1.8595\n",
      "Epoch [    9/   20] | d_loss: 0.4752 | g_loss: 2.4239\n",
      "Epoch [    9/   20] | d_loss: 0.7530 | g_loss: 3.1136\n",
      "Epoch [    9/   20] | d_loss: 0.9083 | g_loss: 1.7715\n",
      "Epoch [    9/   20] | d_loss: 0.3842 | g_loss: 2.7796\n",
      "Epoch [    9/   20] | d_loss: 0.5083 | g_loss: 2.6155\n",
      "Epoch [    9/   20] | d_loss: 0.5639 | g_loss: 3.8912\n",
      "Epoch [    9/   20] | d_loss: 0.8550 | g_loss: 1.2545\n",
      "Epoch [    9/   20] | d_loss: 0.4403 | g_loss: 2.2279\n",
      "Epoch [    9/   20] | d_loss: 0.6482 | g_loss: 1.9314\n",
      "Epoch [    9/   20] | d_loss: 0.5292 | g_loss: 2.0121\n",
      "Epoch [    9/   20] | d_loss: 0.5171 | g_loss: 2.5667\n",
      "Epoch [    9/   20] | d_loss: 0.5790 | g_loss: 2.4186\n",
      "Epoch [    9/   20] | d_loss: 0.8224 | g_loss: 2.0965\n",
      "Epoch [    9/   20] | d_loss: 0.8410 | g_loss: 0.9486\n",
      "Epoch [    9/   20] | d_loss: 0.6720 | g_loss: 2.2673\n",
      "Epoch [    9/   20] | d_loss: 0.8663 | g_loss: 3.0772\n",
      "Epoch [    9/   20] | d_loss: 0.6518 | g_loss: 2.5169\n",
      "Epoch [    9/   20] | d_loss: 0.4204 | g_loss: 1.9353\n",
      "Epoch [    9/   20] | d_loss: 0.6296 | g_loss: 2.5462\n",
      "Epoch [    9/   20] | d_loss: 0.5936 | g_loss: 2.1098\n",
      "Epoch [    9/   20] | d_loss: 0.2490 | g_loss: 3.7316\n",
      "Epoch [    9/   20] | d_loss: 0.6029 | g_loss: 1.9934\n",
      "Epoch [    9/   20] | d_loss: 0.6353 | g_loss: 2.0936\n",
      "Epoch [    9/   20] | d_loss: 1.1859 | g_loss: 1.5664\n",
      "Epoch [    9/   20] | d_loss: 0.5729 | g_loss: 2.9542\n",
      "Epoch [    9/   20] | d_loss: 0.5736 | g_loss: 1.7847\n",
      "Epoch [    9/   20] | d_loss: 1.1863 | g_loss: 2.5929\n",
      "Epoch [    9/   20] | d_loss: 1.1700 | g_loss: 2.4553\n",
      "Epoch [    9/   20] | d_loss: 0.6886 | g_loss: 2.7169\n",
      "Epoch [    9/   20] | d_loss: 0.4861 | g_loss: 2.4411\n",
      "Epoch [    9/   20] | d_loss: 0.7970 | g_loss: 2.4900\n",
      "Epoch [   10/   20] | d_loss: 0.7424 | g_loss: 2.3122\n",
      "Epoch [   10/   20] | d_loss: 0.5416 | g_loss: 2.8017\n",
      "Epoch [   10/   20] | d_loss: 1.0369 | g_loss: 2.0112\n",
      "Epoch [   10/   20] | d_loss: 1.0192 | g_loss: 1.6710\n",
      "Epoch [   10/   20] | d_loss: 0.7068 | g_loss: 2.3431\n",
      "Epoch [   10/   20] | d_loss: 0.5684 | g_loss: 2.7148\n",
      "Epoch [   10/   20] | d_loss: 0.6096 | g_loss: 2.4313\n",
      "Epoch [   10/   20] | d_loss: 0.4925 | g_loss: 1.9204\n",
      "Epoch [   10/   20] | d_loss: 0.8248 | g_loss: 1.6160\n",
      "Epoch [   10/   20] | d_loss: 0.8123 | g_loss: 3.0948\n",
      "Epoch [   10/   20] | d_loss: 0.5369 | g_loss: 1.9119\n",
      "Epoch [   10/   20] | d_loss: 0.7619 | g_loss: 2.9801\n",
      "Epoch [   10/   20] | d_loss: 0.4761 | g_loss: 1.3334\n",
      "Epoch [   10/   20] | d_loss: 0.7573 | g_loss: 1.9222\n",
      "Epoch [   10/   20] | d_loss: 0.5202 | g_loss: 2.5401\n",
      "Epoch [   10/   20] | d_loss: 0.5913 | g_loss: 2.2972\n",
      "Epoch [   10/   20] | d_loss: 0.6955 | g_loss: 1.8585\n",
      "Epoch [   10/   20] | d_loss: 0.6335 | g_loss: 2.4264\n",
      "Epoch [   10/   20] | d_loss: 0.6289 | g_loss: 2.9994\n",
      "Epoch [   10/   20] | d_loss: 0.4621 | g_loss: 2.4665\n",
      "Epoch [   10/   20] | d_loss: 0.6756 | g_loss: 2.3219\n",
      "Epoch [   10/   20] | d_loss: 0.4780 | g_loss: 2.8281\n",
      "Epoch [   10/   20] | d_loss: 0.8495 | g_loss: 1.9711\n",
      "Epoch [   10/   20] | d_loss: 0.6808 | g_loss: 2.7853\n",
      "Epoch [   10/   20] | d_loss: 0.7597 | g_loss: 2.7714\n",
      "Epoch [   10/   20] | d_loss: 0.5847 | g_loss: 1.9486\n",
      "Epoch [   10/   20] | d_loss: 0.9952 | g_loss: 1.7967\n",
      "Epoch [   10/   20] | d_loss: 0.5983 | g_loss: 3.2530\n",
      "Epoch [   10/   20] | d_loss: 0.3953 | g_loss: 3.0882\n",
      "Epoch [   10/   20] | d_loss: 0.6465 | g_loss: 2.6374\n",
      "Epoch [   10/   20] | d_loss: 1.0358 | g_loss: 2.1812\n",
      "Epoch [   10/   20] | d_loss: 0.5637 | g_loss: 1.9710\n",
      "Epoch [   10/   20] | d_loss: 0.5415 | g_loss: 2.5176\n",
      "Epoch [   10/   20] | d_loss: 0.6667 | g_loss: 1.5066\n",
      "Epoch [   10/   20] | d_loss: 0.6563 | g_loss: 2.0316\n",
      "Epoch [   10/   20] | d_loss: 0.6465 | g_loss: 2.7921\n",
      "Epoch [   10/   20] | d_loss: 0.3908 | g_loss: 2.8411\n",
      "Epoch [   10/   20] | d_loss: 0.7693 | g_loss: 3.0705\n",
      "Epoch [   10/   20] | d_loss: 0.9138 | g_loss: 1.8258\n",
      "Epoch [   10/   20] | d_loss: 0.4339 | g_loss: 2.0739\n",
      "Epoch [   10/   20] | d_loss: 0.7731 | g_loss: 2.8300\n",
      "Epoch [   10/   20] | d_loss: 0.8311 | g_loss: 2.2008\n",
      "Epoch [   10/   20] | d_loss: 0.5327 | g_loss: 2.6397\n",
      "Epoch [   10/   20] | d_loss: 0.6647 | g_loss: 2.0727\n",
      "Epoch [   10/   20] | d_loss: 0.6663 | g_loss: 1.6757\n",
      "Epoch [   10/   20] | d_loss: 0.3871 | g_loss: 2.4466\n",
      "Epoch [   10/   20] | d_loss: 0.9391 | g_loss: 3.3735\n",
      "Epoch [   10/   20] | d_loss: 0.3188 | g_loss: 1.4319\n",
      "Epoch [   10/   20] | d_loss: 0.7388 | g_loss: 3.3307\n",
      "Epoch [   10/   20] | d_loss: 1.1890 | g_loss: 1.9934\n",
      "Epoch [   10/   20] | d_loss: 1.0473 | g_loss: 2.1222\n",
      "Epoch [   10/   20] | d_loss: 0.6640 | g_loss: 2.9678\n",
      "Epoch [   10/   20] | d_loss: 0.2630 | g_loss: 3.7222\n",
      "Epoch [   10/   20] | d_loss: 0.8306 | g_loss: 1.6914\n",
      "Epoch [   10/   20] | d_loss: 0.4778 | g_loss: 2.2998\n",
      "Epoch [   10/   20] | d_loss: 0.9667 | g_loss: 1.7335\n",
      "Epoch [   10/   20] | d_loss: 0.4964 | g_loss: 2.1869\n",
      "Epoch [   11/   20] | d_loss: 0.7258 | g_loss: 3.3877\n",
      "Epoch [   11/   20] | d_loss: 0.6698 | g_loss: 2.3998\n",
      "Epoch [   11/   20] | d_loss: 0.4812 | g_loss: 3.0962\n",
      "Epoch [   11/   20] | d_loss: 0.4428 | g_loss: 2.8411\n",
      "Epoch [   11/   20] | d_loss: 0.4657 | g_loss: 1.8599\n",
      "Epoch [   11/   20] | d_loss: 0.6509 | g_loss: 2.2897\n",
      "Epoch [   11/   20] | d_loss: 0.9693 | g_loss: 1.6281\n",
      "Epoch [   11/   20] | d_loss: 0.4792 | g_loss: 3.5783\n",
      "Epoch [   11/   20] | d_loss: 0.5166 | g_loss: 3.6053\n",
      "Epoch [   11/   20] | d_loss: 0.5920 | g_loss: 3.3179\n",
      "Epoch [   11/   20] | d_loss: 0.5138 | g_loss: 2.4669\n",
      "Epoch [   11/   20] | d_loss: 0.3905 | g_loss: 3.8997\n",
      "Epoch [   11/   20] | d_loss: 0.5970 | g_loss: 1.4958\n",
      "Epoch [   11/   20] | d_loss: 0.4844 | g_loss: 3.1959\n",
      "Epoch [   11/   20] | d_loss: 0.7208 | g_loss: 1.8594\n",
      "Epoch [   11/   20] | d_loss: 0.5001 | g_loss: 1.9846\n",
      "Epoch [   11/   20] | d_loss: 0.7075 | g_loss: 2.7171\n",
      "Epoch [   11/   20] | d_loss: 0.9994 | g_loss: 2.0347\n",
      "Epoch [   11/   20] | d_loss: 0.5095 | g_loss: 2.2478\n",
      "Epoch [   11/   20] | d_loss: 0.7857 | g_loss: 3.0577\n",
      "Epoch [   11/   20] | d_loss: 0.6060 | g_loss: 1.7393\n",
      "Epoch [   11/   20] | d_loss: 0.7923 | g_loss: 1.8802\n",
      "Epoch [   11/   20] | d_loss: 1.0450 | g_loss: 2.0202\n",
      "Epoch [   11/   20] | d_loss: 0.6532 | g_loss: 3.0299\n",
      "Epoch [   11/   20] | d_loss: 0.3782 | g_loss: 2.5729\n",
      "Epoch [   11/   20] | d_loss: 0.5119 | g_loss: 2.8642\n",
      "Epoch [   11/   20] | d_loss: 0.6497 | g_loss: 3.1971\n",
      "Epoch [   11/   20] | d_loss: 0.6683 | g_loss: 3.2721\n",
      "Epoch [   11/   20] | d_loss: 0.4829 | g_loss: 2.4464\n",
      "Epoch [   11/   20] | d_loss: 0.2990 | g_loss: 3.3313\n",
      "Epoch [   11/   20] | d_loss: 0.5145 | g_loss: 2.5353\n",
      "Epoch [   11/   20] | d_loss: 0.2730 | g_loss: 2.1636\n",
      "Epoch [   11/   20] | d_loss: 0.4165 | g_loss: 2.5807\n",
      "Epoch [   11/   20] | d_loss: 0.6840 | g_loss: 2.8809\n",
      "Epoch [   11/   20] | d_loss: 0.5286 | g_loss: 2.2769\n",
      "Epoch [   11/   20] | d_loss: 0.5430 | g_loss: 3.1082\n",
      "Epoch [   11/   20] | d_loss: 0.9752 | g_loss: 1.6517\n",
      "Epoch [   11/   20] | d_loss: 0.6340 | g_loss: 2.6133\n",
      "Epoch [   11/   20] | d_loss: 0.8017 | g_loss: 1.5084\n",
      "Epoch [   11/   20] | d_loss: 0.4512 | g_loss: 2.3768\n",
      "Epoch [   11/   20] | d_loss: 0.5016 | g_loss: 2.5238\n",
      "Epoch [   11/   20] | d_loss: 0.7096 | g_loss: 2.5192\n",
      "Epoch [   11/   20] | d_loss: 1.0221 | g_loss: 1.8822\n",
      "Epoch [   11/   20] | d_loss: 0.5167 | g_loss: 2.9361\n",
      "Epoch [   11/   20] | d_loss: 0.8537 | g_loss: 2.0499\n",
      "Epoch [   11/   20] | d_loss: 0.3014 | g_loss: 2.0300\n",
      "Epoch [   11/   20] | d_loss: 0.8130 | g_loss: 3.3273\n",
      "Epoch [   11/   20] | d_loss: 0.5838 | g_loss: 3.6089\n",
      "Epoch [   11/   20] | d_loss: 0.5233 | g_loss: 3.6575\n",
      "Epoch [   11/   20] | d_loss: 0.6088 | g_loss: 2.6130\n",
      "Epoch [   11/   20] | d_loss: 0.4873 | g_loss: 2.7970\n",
      "Epoch [   11/   20] | d_loss: 0.8534 | g_loss: 2.4517\n",
      "Epoch [   11/   20] | d_loss: 0.5663 | g_loss: 2.0756\n",
      "Epoch [   11/   20] | d_loss: 0.8670 | g_loss: 2.1562\n",
      "Epoch [   11/   20] | d_loss: 0.4669 | g_loss: 1.5122\n",
      "Epoch [   11/   20] | d_loss: 0.4361 | g_loss: 3.9215\n",
      "Epoch [   11/   20] | d_loss: 1.2621 | g_loss: 1.9639\n",
      "Epoch [   12/   20] | d_loss: 0.4700 | g_loss: 2.9188\n",
      "Epoch [   12/   20] | d_loss: 0.9687 | g_loss: 1.8719\n",
      "Epoch [   12/   20] | d_loss: 0.5019 | g_loss: 2.5205\n",
      "Epoch [   12/   20] | d_loss: 0.3783 | g_loss: 2.2657\n",
      "Epoch [   12/   20] | d_loss: 0.7945 | g_loss: 1.8859\n",
      "Epoch [   12/   20] | d_loss: 0.7660 | g_loss: 2.4647\n",
      "Epoch [   12/   20] | d_loss: 0.7097 | g_loss: 2.6616\n",
      "Epoch [   12/   20] | d_loss: 0.9436 | g_loss: 3.1111\n",
      "Epoch [   12/   20] | d_loss: 1.1682 | g_loss: 2.7668\n",
      "Epoch [   12/   20] | d_loss: 0.5979 | g_loss: 2.2532\n",
      "Epoch [   12/   20] | d_loss: 0.7433 | g_loss: 2.2865\n",
      "Epoch [   12/   20] | d_loss: 0.2576 | g_loss: 3.6622\n",
      "Epoch [   12/   20] | d_loss: 0.6808 | g_loss: 2.5208\n",
      "Epoch [   12/   20] | d_loss: 0.7638 | g_loss: 3.8084\n",
      "Epoch [   12/   20] | d_loss: 0.6621 | g_loss: 3.4552\n",
      "Epoch [   12/   20] | d_loss: 0.9749 | g_loss: 2.0723\n",
      "Epoch [   12/   20] | d_loss: 0.4830 | g_loss: 2.3177\n",
      "Epoch [   12/   20] | d_loss: 0.4169 | g_loss: 3.0911\n",
      "Epoch [   12/   20] | d_loss: 0.6976 | g_loss: 2.9121\n",
      "Epoch [   12/   20] | d_loss: 0.7818 | g_loss: 3.5836\n",
      "Epoch [   12/   20] | d_loss: 0.6238 | g_loss: 3.0760\n",
      "Epoch [   12/   20] | d_loss: 0.7638 | g_loss: 2.0558\n",
      "Epoch [   12/   20] | d_loss: 0.2723 | g_loss: 2.5812\n",
      "Epoch [   12/   20] | d_loss: 0.7886 | g_loss: 2.5216\n",
      "Epoch [   12/   20] | d_loss: 0.5278 | g_loss: 3.3990\n",
      "Epoch [   12/   20] | d_loss: 0.1591 | g_loss: 3.4859\n",
      "Epoch [   12/   20] | d_loss: 0.7900 | g_loss: 2.7271\n",
      "Epoch [   12/   20] | d_loss: 1.2407 | g_loss: 2.1940\n",
      "Epoch [   12/   20] | d_loss: 0.6260 | g_loss: 2.9989\n",
      "Epoch [   12/   20] | d_loss: 0.4983 | g_loss: 2.4494\n",
      "Epoch [   12/   20] | d_loss: 0.5470 | g_loss: 3.9698\n",
      "Epoch [   12/   20] | d_loss: 0.4149 | g_loss: 2.6681\n",
      "Epoch [   12/   20] | d_loss: 1.0127 | g_loss: 1.9403\n",
      "Epoch [   12/   20] | d_loss: 0.4736 | g_loss: 1.4935\n",
      "Epoch [   12/   20] | d_loss: 0.4798 | g_loss: 3.1465\n",
      "Epoch [   12/   20] | d_loss: 0.3659 | g_loss: 2.5894\n",
      "Epoch [   12/   20] | d_loss: 0.2747 | g_loss: 2.5317\n",
      "Epoch [   12/   20] | d_loss: 0.3250 | g_loss: 2.8626\n",
      "Epoch [   12/   20] | d_loss: 0.4272 | g_loss: 3.0185\n",
      "Epoch [   12/   20] | d_loss: 0.5647 | g_loss: 2.4979\n",
      "Epoch [   12/   20] | d_loss: 0.4449 | g_loss: 3.4624\n",
      "Epoch [   12/   20] | d_loss: 0.2014 | g_loss: 2.1132\n",
      "Epoch [   12/   20] | d_loss: 0.7885 | g_loss: 2.1343\n",
      "Epoch [   12/   20] | d_loss: 0.4731 | g_loss: 1.8981\n",
      "Epoch [   12/   20] | d_loss: 0.9214 | g_loss: 5.0321\n",
      "Epoch [   12/   20] | d_loss: 0.5412 | g_loss: 1.6790\n",
      "Epoch [   12/   20] | d_loss: 0.4025 | g_loss: 3.6188\n",
      "Epoch [   12/   20] | d_loss: 0.7122 | g_loss: 2.5737\n",
      "Epoch [   12/   20] | d_loss: 0.8623 | g_loss: 2.8193\n",
      "Epoch [   12/   20] | d_loss: 0.4771 | g_loss: 2.8924\n",
      "Epoch [   12/   20] | d_loss: 0.4901 | g_loss: 3.8070\n",
      "Epoch [   12/   20] | d_loss: 0.5682 | g_loss: 2.0271\n",
      "Epoch [   12/   20] | d_loss: 0.4904 | g_loss: 2.4501\n",
      "Epoch [   12/   20] | d_loss: 0.4217 | g_loss: 3.4854\n",
      "Epoch [   12/   20] | d_loss: 0.7267 | g_loss: 3.1115\n",
      "Epoch [   12/   20] | d_loss: 0.7368 | g_loss: 1.8658\n",
      "Epoch [   12/   20] | d_loss: 0.4481 | g_loss: 2.8648\n",
      "Epoch [   13/   20] | d_loss: 0.4358 | g_loss: 2.2812\n",
      "Epoch [   13/   20] | d_loss: 0.3197 | g_loss: 2.4547\n",
      "Epoch [   13/   20] | d_loss: 0.6203 | g_loss: 4.1434\n",
      "Epoch [   13/   20] | d_loss: 0.4165 | g_loss: 2.6811\n",
      "Epoch [   13/   20] | d_loss: 0.3398 | g_loss: 2.1130\n",
      "Epoch [   13/   20] | d_loss: 0.2989 | g_loss: 2.1068\n",
      "Epoch [   13/   20] | d_loss: 0.3175 | g_loss: 3.1869\n",
      "Epoch [   13/   20] | d_loss: 0.5113 | g_loss: 2.1509\n",
      "Epoch [   13/   20] | d_loss: 0.3550 | g_loss: 2.9104\n",
      "Epoch [   13/   20] | d_loss: 0.5518 | g_loss: 1.3637\n",
      "Epoch [   13/   20] | d_loss: 0.9540 | g_loss: 1.2058\n",
      "Epoch [   13/   20] | d_loss: 0.5511 | g_loss: 3.6806\n",
      "Epoch [   13/   20] | d_loss: 0.3062 | g_loss: 3.3493\n",
      "Epoch [   13/   20] | d_loss: 0.4362 | g_loss: 3.1318\n",
      "Epoch [   13/   20] | d_loss: 0.3023 | g_loss: 2.3971\n",
      "Epoch [   13/   20] | d_loss: 0.4420 | g_loss: 2.7952\n",
      "Epoch [   13/   20] | d_loss: 0.4421 | g_loss: 3.6219\n",
      "Epoch [   13/   20] | d_loss: 0.8774 | g_loss: 2.2649\n",
      "Epoch [   13/   20] | d_loss: 0.5768 | g_loss: 2.9734\n",
      "Epoch [   13/   20] | d_loss: 0.3426 | g_loss: 2.9280\n",
      "Epoch [   13/   20] | d_loss: 0.6571 | g_loss: 1.9433\n",
      "Epoch [   13/   20] | d_loss: 0.8599 | g_loss: 3.0833\n",
      "Epoch [   13/   20] | d_loss: 0.7083 | g_loss: 2.0686\n",
      "Epoch [   13/   20] | d_loss: 0.3492 | g_loss: 2.0508\n",
      "Epoch [   13/   20] | d_loss: 0.6197 | g_loss: 3.2744\n",
      "Epoch [   13/   20] | d_loss: 0.3276 | g_loss: 2.7039\n",
      "Epoch [   13/   20] | d_loss: 0.7263 | g_loss: 3.3783\n",
      "Epoch [   13/   20] | d_loss: 0.4055 | g_loss: 3.3380\n",
      "Epoch [   13/   20] | d_loss: 0.4968 | g_loss: 2.9997\n",
      "Epoch [   13/   20] | d_loss: 0.9025 | g_loss: 3.5180\n",
      "Epoch [   13/   20] | d_loss: 0.3446 | g_loss: 3.4123\n",
      "Epoch [   13/   20] | d_loss: 0.7342 | g_loss: 2.7378\n",
      "Epoch [   13/   20] | d_loss: 0.3627 | g_loss: 4.6629\n",
      "Epoch [   13/   20] | d_loss: 0.3452 | g_loss: 2.9540\n",
      "Epoch [   13/   20] | d_loss: 0.9906 | g_loss: 3.2333\n",
      "Epoch [   13/   20] | d_loss: 0.5299 | g_loss: 2.5761\n",
      "Epoch [   13/   20] | d_loss: 0.4956 | g_loss: 3.1952\n",
      "Epoch [   13/   20] | d_loss: 0.4007 | g_loss: 3.0401\n",
      "Epoch [   13/   20] | d_loss: 0.6602 | g_loss: 1.7556\n",
      "Epoch [   13/   20] | d_loss: 0.6764 | g_loss: 3.3480\n",
      "Epoch [   13/   20] | d_loss: 0.4964 | g_loss: 3.5425\n",
      "Epoch [   13/   20] | d_loss: 0.3996 | g_loss: 2.4613\n",
      "Epoch [   13/   20] | d_loss: 0.4632 | g_loss: 2.0831\n",
      "Epoch [   13/   20] | d_loss: 0.3709 | g_loss: 2.3637\n",
      "Epoch [   13/   20] | d_loss: 0.3526 | g_loss: 2.6123\n",
      "Epoch [   13/   20] | d_loss: 0.6192 | g_loss: 3.1794\n",
      "Epoch [   13/   20] | d_loss: 0.5085 | g_loss: 2.9854\n",
      "Epoch [   13/   20] | d_loss: 0.5563 | g_loss: 2.1440\n",
      "Epoch [   13/   20] | d_loss: 0.3391 | g_loss: 2.7507\n",
      "Epoch [   13/   20] | d_loss: 0.6578 | g_loss: 1.8880\n",
      "Epoch [   13/   20] | d_loss: 0.3543 | g_loss: 2.9646\n",
      "Epoch [   13/   20] | d_loss: 0.3867 | g_loss: 2.3437\n",
      "Epoch [   13/   20] | d_loss: 0.1882 | g_loss: 2.4777\n",
      "Epoch [   13/   20] | d_loss: 0.6924 | g_loss: 2.6719\n",
      "Epoch [   13/   20] | d_loss: 0.4824 | g_loss: 3.1258\n",
      "Epoch [   13/   20] | d_loss: 0.4586 | g_loss: 2.5618\n",
      "Epoch [   13/   20] | d_loss: 0.5417 | g_loss: 2.6978\n",
      "Epoch [   14/   20] | d_loss: 0.3107 | g_loss: 3.3559\n",
      "Epoch [   14/   20] | d_loss: 0.3233 | g_loss: 3.1780\n",
      "Epoch [   14/   20] | d_loss: 0.4376 | g_loss: 3.1663\n",
      "Epoch [   14/   20] | d_loss: 0.8275 | g_loss: 2.0866\n",
      "Epoch [   14/   20] | d_loss: 0.3231 | g_loss: 2.6453\n",
      "Epoch [   14/   20] | d_loss: 0.3767 | g_loss: 3.5593\n",
      "Epoch [   14/   20] | d_loss: 0.5079 | g_loss: 2.1332\n",
      "Epoch [   14/   20] | d_loss: 0.7824 | g_loss: 2.4864\n",
      "Epoch [   14/   20] | d_loss: 0.9545 | g_loss: 2.1083\n",
      "Epoch [   14/   20] | d_loss: 0.4560 | g_loss: 2.5086\n",
      "Epoch [   14/   20] | d_loss: 0.6543 | g_loss: 1.8180\n",
      "Epoch [   14/   20] | d_loss: 0.7776 | g_loss: 3.9561\n",
      "Epoch [   14/   20] | d_loss: 0.4962 | g_loss: 3.3448\n",
      "Epoch [   14/   20] | d_loss: 0.4992 | g_loss: 3.2838\n",
      "Epoch [   14/   20] | d_loss: 0.3970 | g_loss: 3.0132\n",
      "Epoch [   14/   20] | d_loss: 0.2155 | g_loss: 2.3131\n",
      "Epoch [   14/   20] | d_loss: 0.2454 | g_loss: 2.4425\n",
      "Epoch [   14/   20] | d_loss: 0.5211 | g_loss: 3.2831\n",
      "Epoch [   14/   20] | d_loss: 0.3166 | g_loss: 2.8944\n",
      "Epoch [   14/   20] | d_loss: 0.4754 | g_loss: 2.7228\n",
      "Epoch [   14/   20] | d_loss: 0.7217 | g_loss: 4.5665\n",
      "Epoch [   14/   20] | d_loss: 0.9464 | g_loss: 1.7299\n",
      "Epoch [   14/   20] | d_loss: 0.4454 | g_loss: 2.9277\n",
      "Epoch [   14/   20] | d_loss: 0.2488 | g_loss: 4.3871\n",
      "Epoch [   14/   20] | d_loss: 0.9583 | g_loss: 2.2244\n",
      "Epoch [   14/   20] | d_loss: 0.5408 | g_loss: 2.1330\n",
      "Epoch [   14/   20] | d_loss: 0.3678 | g_loss: 3.1334\n",
      "Epoch [   14/   20] | d_loss: 0.9760 | g_loss: 2.7119\n",
      "Epoch [   14/   20] | d_loss: 0.5277 | g_loss: 3.9787\n",
      "Epoch [   14/   20] | d_loss: 0.5468 | g_loss: 3.0517\n",
      "Epoch [   14/   20] | d_loss: 0.3422 | g_loss: 1.5137\n",
      "Epoch [   14/   20] | d_loss: 0.7137 | g_loss: 3.1761\n",
      "Epoch [   14/   20] | d_loss: 0.2831 | g_loss: 2.2245\n",
      "Epoch [   14/   20] | d_loss: 0.3289 | g_loss: 2.6947\n",
      "Epoch [   14/   20] | d_loss: 0.1647 | g_loss: 4.5814\n",
      "Epoch [   14/   20] | d_loss: 0.4824 | g_loss: 3.0766\n",
      "Epoch [   14/   20] | d_loss: 0.3833 | g_loss: 3.8398\n",
      "Epoch [   14/   20] | d_loss: 0.5485 | g_loss: 1.9828\n",
      "Epoch [   14/   20] | d_loss: 0.3662 | g_loss: 3.1246\n",
      "Epoch [   14/   20] | d_loss: 0.2900 | g_loss: 4.2803\n",
      "Epoch [   14/   20] | d_loss: 0.4200 | g_loss: 3.3942\n",
      "Epoch [   14/   20] | d_loss: 0.2823 | g_loss: 3.7292\n",
      "Epoch [   14/   20] | d_loss: 0.2945 | g_loss: 2.9662\n",
      "Epoch [   14/   20] | d_loss: 0.4040 | g_loss: 3.7483\n",
      "Epoch [   14/   20] | d_loss: 0.8297 | g_loss: 1.7061\n",
      "Epoch [   14/   20] | d_loss: 0.9189 | g_loss: 3.6280\n",
      "Epoch [   14/   20] | d_loss: 0.4887 | g_loss: 2.5903\n",
      "Epoch [   14/   20] | d_loss: 0.6007 | g_loss: 2.7668\n",
      "Epoch [   14/   20] | d_loss: 0.4492 | g_loss: 3.5383\n",
      "Epoch [   14/   20] | d_loss: 0.2540 | g_loss: 3.9588\n",
      "Epoch [   14/   20] | d_loss: 0.2794 | g_loss: 3.6506\n",
      "Epoch [   14/   20] | d_loss: 0.5654 | g_loss: 4.1327\n",
      "Epoch [   14/   20] | d_loss: 0.7566 | g_loss: 3.0870\n",
      "Epoch [   14/   20] | d_loss: 0.5129 | g_loss: 1.9910\n",
      "Epoch [   14/   20] | d_loss: 0.7785 | g_loss: 2.2937\n",
      "Epoch [   14/   20] | d_loss: 0.4570 | g_loss: 2.8970\n",
      "Epoch [   14/   20] | d_loss: 0.5894 | g_loss: 3.9553\n",
      "Epoch [   15/   20] | d_loss: 1.0148 | g_loss: 3.0801\n",
      "Epoch [   15/   20] | d_loss: 0.7859 | g_loss: 2.2547\n",
      "Epoch [   15/   20] | d_loss: 0.4196 | g_loss: 2.3732\n",
      "Epoch [   15/   20] | d_loss: 0.2989 | g_loss: 3.5071\n",
      "Epoch [   15/   20] | d_loss: 0.7513 | g_loss: 2.7610\n",
      "Epoch [   15/   20] | d_loss: 0.8351 | g_loss: 4.7708\n",
      "Epoch [   15/   20] | d_loss: 0.4063 | g_loss: 3.3758\n",
      "Epoch [   15/   20] | d_loss: 0.4423 | g_loss: 2.5997\n",
      "Epoch [   15/   20] | d_loss: 0.6742 | g_loss: 3.8875\n",
      "Epoch [   15/   20] | d_loss: 0.4009 | g_loss: 3.0465\n",
      "Epoch [   15/   20] | d_loss: 0.3023 | g_loss: 3.8977\n",
      "Epoch [   15/   20] | d_loss: 0.1895 | g_loss: 4.2071\n",
      "Epoch [   15/   20] | d_loss: 0.3835 | g_loss: 3.3273\n",
      "Epoch [   15/   20] | d_loss: 0.4407 | g_loss: 4.2538\n",
      "Epoch [   15/   20] | d_loss: 0.9556 | g_loss: 3.8783\n",
      "Epoch [   15/   20] | d_loss: 0.2615 | g_loss: 3.1568\n",
      "Epoch [   15/   20] | d_loss: 0.4745 | g_loss: 4.4394\n",
      "Epoch [   15/   20] | d_loss: 0.7441 | g_loss: 2.7179\n",
      "Epoch [   15/   20] | d_loss: 0.3103 | g_loss: 3.1134\n",
      "Epoch [   15/   20] | d_loss: 0.5401 | g_loss: 2.4173\n",
      "Epoch [   15/   20] | d_loss: 0.2821 | g_loss: 2.4137\n",
      "Epoch [   15/   20] | d_loss: 0.7780 | g_loss: 3.9442\n",
      "Epoch [   15/   20] | d_loss: 0.7028 | g_loss: 3.8544\n",
      "Epoch [   15/   20] | d_loss: 0.5238 | g_loss: 3.9915\n",
      "Epoch [   15/   20] | d_loss: 0.2566 | g_loss: 2.8428\n",
      "Epoch [   15/   20] | d_loss: 0.3602 | g_loss: 4.0753\n",
      "Epoch [   15/   20] | d_loss: 0.5109 | g_loss: 3.2178\n",
      "Epoch [   15/   20] | d_loss: 0.4236 | g_loss: 2.3738\n",
      "Epoch [   15/   20] | d_loss: 0.8490 | g_loss: 3.0562\n",
      "Epoch [   15/   20] | d_loss: 0.4786 | g_loss: 3.8670\n",
      "Epoch [   15/   20] | d_loss: 0.8426 | g_loss: 3.3440\n",
      "Epoch [   15/   20] | d_loss: 0.3294 | g_loss: 2.6043\n",
      "Epoch [   15/   20] | d_loss: 1.0874 | g_loss: 4.2240\n",
      "Epoch [   15/   20] | d_loss: 0.5560 | g_loss: 2.4503\n",
      "Epoch [   15/   20] | d_loss: 0.5757 | g_loss: 3.3818\n",
      "Epoch [   15/   20] | d_loss: 0.3934 | g_loss: 1.9354\n",
      "Epoch [   15/   20] | d_loss: 0.1515 | g_loss: 3.1172\n",
      "Epoch [   15/   20] | d_loss: 0.3977 | g_loss: 2.8836\n",
      "Epoch [   15/   20] | d_loss: 0.1502 | g_loss: 2.2620\n",
      "Epoch [   15/   20] | d_loss: 0.2199 | g_loss: 3.0293\n",
      "Epoch [   15/   20] | d_loss: 0.4755 | g_loss: 3.1006\n",
      "Epoch [   15/   20] | d_loss: 0.6136 | g_loss: 5.1841\n",
      "Epoch [   15/   20] | d_loss: 0.2375 | g_loss: 3.6761\n",
      "Epoch [   15/   20] | d_loss: 0.4207 | g_loss: 3.1561\n",
      "Epoch [   15/   20] | d_loss: 0.5032 | g_loss: 4.2091\n",
      "Epoch [   15/   20] | d_loss: 0.4353 | g_loss: 3.2085\n",
      "Epoch [   15/   20] | d_loss: 0.3894 | g_loss: 2.1050\n",
      "Epoch [   15/   20] | d_loss: 0.8591 | g_loss: 0.9856\n",
      "Epoch [   15/   20] | d_loss: 0.1816 | g_loss: 3.5787\n",
      "Epoch [   15/   20] | d_loss: 0.6259 | g_loss: 2.9897\n",
      "Epoch [   15/   20] | d_loss: 0.4011 | g_loss: 2.7357\n",
      "Epoch [   15/   20] | d_loss: 0.3743 | g_loss: 3.1793\n",
      "Epoch [   15/   20] | d_loss: 0.4883 | g_loss: 3.7459\n",
      "Epoch [   15/   20] | d_loss: 0.6166 | g_loss: 2.8565\n",
      "Epoch [   15/   20] | d_loss: 0.7031 | g_loss: 3.4416\n",
      "Epoch [   15/   20] | d_loss: 0.3815 | g_loss: 3.2094\n",
      "Epoch [   15/   20] | d_loss: 0.2088 | g_loss: 2.5481\n",
      "Epoch [   16/   20] | d_loss: 0.3922 | g_loss: 3.7970\n",
      "Epoch [   16/   20] | d_loss: 0.4911 | g_loss: 3.4016\n",
      "Epoch [   16/   20] | d_loss: 0.1922 | g_loss: 3.4765\n",
      "Epoch [   16/   20] | d_loss: 0.3322 | g_loss: 3.7217\n",
      "Epoch [   16/   20] | d_loss: 0.7854 | g_loss: 3.1743\n",
      "Epoch [   16/   20] | d_loss: 0.3528 | g_loss: 3.1998\n",
      "Epoch [   16/   20] | d_loss: 0.1488 | g_loss: 4.2947\n",
      "Epoch [   16/   20] | d_loss: 0.4237 | g_loss: 3.9137\n",
      "Epoch [   16/   20] | d_loss: 0.3011 | g_loss: 4.5379\n",
      "Epoch [   16/   20] | d_loss: 0.4111 | g_loss: 4.3937\n",
      "Epoch [   16/   20] | d_loss: 0.4001 | g_loss: 2.4847\n",
      "Epoch [   16/   20] | d_loss: 0.3205 | g_loss: 4.1307\n",
      "Epoch [   16/   20] | d_loss: 0.4537 | g_loss: 5.5587\n",
      "Epoch [   16/   20] | d_loss: 0.3122 | g_loss: 2.5318\n",
      "Epoch [   16/   20] | d_loss: 0.2610 | g_loss: 3.6188\n",
      "Epoch [   16/   20] | d_loss: 0.4836 | g_loss: 2.4104\n",
      "Epoch [   16/   20] | d_loss: 0.3855 | g_loss: 3.0679\n",
      "Epoch [   16/   20] | d_loss: 0.3822 | g_loss: 2.9679\n",
      "Epoch [   16/   20] | d_loss: 0.8275 | g_loss: 5.9458\n",
      "Epoch [   16/   20] | d_loss: 0.4521 | g_loss: 2.6898\n",
      "Epoch [   16/   20] | d_loss: 0.4185 | g_loss: 3.0547\n",
      "Epoch [   16/   20] | d_loss: 0.3636 | g_loss: 3.1346\n",
      "Epoch [   16/   20] | d_loss: 0.1974 | g_loss: 3.6161\n",
      "Epoch [   16/   20] | d_loss: 0.7387 | g_loss: 2.0880\n",
      "Epoch [   16/   20] | d_loss: 0.5439 | g_loss: 3.6806\n",
      "Epoch [   16/   20] | d_loss: 0.3569 | g_loss: 4.9322\n",
      "Epoch [   16/   20] | d_loss: 1.2754 | g_loss: 4.7900\n",
      "Epoch [   16/   20] | d_loss: 0.4242 | g_loss: 4.0591\n",
      "Epoch [   16/   20] | d_loss: 0.4532 | g_loss: 3.5432\n",
      "Epoch [   16/   20] | d_loss: 0.1683 | g_loss: 3.1209\n",
      "Epoch [   16/   20] | d_loss: 0.4444 | g_loss: 4.1354\n",
      "Epoch [   16/   20] | d_loss: 0.4734 | g_loss: 4.0735\n",
      "Epoch [   16/   20] | d_loss: 0.1246 | g_loss: 2.9356\n",
      "Epoch [   16/   20] | d_loss: 0.4308 | g_loss: 2.6994\n",
      "Epoch [   16/   20] | d_loss: 0.2110 | g_loss: 4.8743\n",
      "Epoch [   16/   20] | d_loss: 0.4812 | g_loss: 3.6146\n",
      "Epoch [   16/   20] | d_loss: 0.6500 | g_loss: 2.0305\n",
      "Epoch [   16/   20] | d_loss: 0.4278 | g_loss: 3.9102\n",
      "Epoch [   16/   20] | d_loss: 0.5164 | g_loss: 3.7706\n",
      "Epoch [   16/   20] | d_loss: 0.3288 | g_loss: 2.9114\n",
      "Epoch [   16/   20] | d_loss: 0.8480 | g_loss: 2.3524\n",
      "Epoch [   16/   20] | d_loss: 2.1081 | g_loss: 6.2654\n",
      "Epoch [   16/   20] | d_loss: 0.4622 | g_loss: 2.9757\n",
      "Epoch [   16/   20] | d_loss: 0.5332 | g_loss: 2.6180\n",
      "Epoch [   16/   20] | d_loss: 0.6403 | g_loss: 2.5579\n",
      "Epoch [   16/   20] | d_loss: 0.7586 | g_loss: 2.2675\n",
      "Epoch [   16/   20] | d_loss: 0.3792 | g_loss: 3.7026\n",
      "Epoch [   16/   20] | d_loss: 0.4957 | g_loss: 4.9058\n",
      "Epoch [   16/   20] | d_loss: 0.9967 | g_loss: 2.2395\n",
      "Epoch [   16/   20] | d_loss: 0.3603 | g_loss: 4.6311\n",
      "Epoch [   16/   20] | d_loss: 0.3527 | g_loss: 2.0793\n",
      "Epoch [   16/   20] | d_loss: 0.3888 | g_loss: 3.3834\n",
      "Epoch [   16/   20] | d_loss: 0.4834 | g_loss: 4.9803\n",
      "Epoch [   16/   20] | d_loss: 0.4480 | g_loss: 2.7521\n",
      "Epoch [   16/   20] | d_loss: 0.6819 | g_loss: 2.9238\n",
      "Epoch [   16/   20] | d_loss: 0.5055 | g_loss: 3.9631\n",
      "Epoch [   16/   20] | d_loss: 0.3885 | g_loss: 2.5499\n",
      "Epoch [   17/   20] | d_loss: 1.0350 | g_loss: 4.9805\n",
      "Epoch [   17/   20] | d_loss: 0.4504 | g_loss: 2.5165\n",
      "Epoch [   17/   20] | d_loss: 0.2939 | g_loss: 4.5040\n",
      "Epoch [   17/   20] | d_loss: 0.2650 | g_loss: 3.6562\n",
      "Epoch [   17/   20] | d_loss: 0.3880 | g_loss: 3.7257\n",
      "Epoch [   17/   20] | d_loss: 0.3806 | g_loss: 3.1868\n",
      "Epoch [   17/   20] | d_loss: 0.8014 | g_loss: 4.0503\n",
      "Epoch [   17/   20] | d_loss: 0.2033 | g_loss: 2.2631\n",
      "Epoch [   17/   20] | d_loss: 0.4039 | g_loss: 2.6208\n",
      "Epoch [   17/   20] | d_loss: 0.4016 | g_loss: 2.2749\n",
      "Epoch [   17/   20] | d_loss: 0.2542 | g_loss: 1.7054\n",
      "Epoch [   17/   20] | d_loss: 0.6249 | g_loss: 2.4526\n",
      "Epoch [   17/   20] | d_loss: 0.2852 | g_loss: 4.2947\n",
      "Epoch [   17/   20] | d_loss: 0.2362 | g_loss: 3.3473\n",
      "Epoch [   17/   20] | d_loss: 1.0001 | g_loss: 1.5578\n",
      "Epoch [   17/   20] | d_loss: 0.4291 | g_loss: 2.0404\n",
      "Epoch [   17/   20] | d_loss: 0.3575 | g_loss: 2.8947\n",
      "Epoch [   17/   20] | d_loss: 0.4293 | g_loss: 3.0592\n",
      "Epoch [   17/   20] | d_loss: 0.5815 | g_loss: 2.7465\n",
      "Epoch [   17/   20] | d_loss: 0.3972 | g_loss: 2.3075\n",
      "Epoch [   17/   20] | d_loss: 0.5822 | g_loss: 2.8357\n",
      "Epoch [   17/   20] | d_loss: 0.3527 | g_loss: 4.9993\n",
      "Epoch [   17/   20] | d_loss: 0.3586 | g_loss: 3.8557\n",
      "Epoch [   17/   20] | d_loss: 0.4758 | g_loss: 5.1976\n",
      "Epoch [   17/   20] | d_loss: 0.7590 | g_loss: 2.6905\n",
      "Epoch [   17/   20] | d_loss: 0.0883 | g_loss: 3.7203\n",
      "Epoch [   17/   20] | d_loss: 0.3814 | g_loss: 2.3481\n",
      "Epoch [   17/   20] | d_loss: 0.6718 | g_loss: 2.6995\n",
      "Epoch [   17/   20] | d_loss: 0.2674 | g_loss: 2.5724\n",
      "Epoch [   17/   20] | d_loss: 0.4540 | g_loss: 4.1575\n",
      "Epoch [   17/   20] | d_loss: 0.3019 | g_loss: 2.5964\n",
      "Epoch [   17/   20] | d_loss: 0.7640 | g_loss: 2.3763\n",
      "Epoch [   17/   20] | d_loss: 0.1994 | g_loss: 3.8201\n",
      "Epoch [   17/   20] | d_loss: 0.2592 | g_loss: 4.4076\n",
      "Epoch [   17/   20] | d_loss: 0.3103 | g_loss: 3.6627\n",
      "Epoch [   17/   20] | d_loss: 0.1833 | g_loss: 3.6219\n",
      "Epoch [   17/   20] | d_loss: 0.2449 | g_loss: 5.2007\n",
      "Epoch [   17/   20] | d_loss: 0.3469 | g_loss: 3.6389\n",
      "Epoch [   17/   20] | d_loss: 0.1523 | g_loss: 4.0748\n",
      "Epoch [   17/   20] | d_loss: 0.4291 | g_loss: 3.7251\n",
      "Epoch [   17/   20] | d_loss: 1.1755 | g_loss: 5.6707\n",
      "Epoch [   17/   20] | d_loss: 0.2491 | g_loss: 4.3467\n",
      "Epoch [   17/   20] | d_loss: 0.2883 | g_loss: 4.6062\n",
      "Epoch [   17/   20] | d_loss: 0.5637 | g_loss: 3.3528\n",
      "Epoch [   17/   20] | d_loss: 0.2551 | g_loss: 4.9144\n",
      "Epoch [   17/   20] | d_loss: 0.4916 | g_loss: 3.3439\n",
      "Epoch [   17/   20] | d_loss: 0.6386 | g_loss: 3.8452\n",
      "Epoch [   17/   20] | d_loss: 0.3312 | g_loss: 2.9797\n",
      "Epoch [   17/   20] | d_loss: 0.7939 | g_loss: 2.2027\n",
      "Epoch [   17/   20] | d_loss: 0.4064 | g_loss: 2.8521\n",
      "Epoch [   17/   20] | d_loss: 1.6152 | g_loss: 0.6182\n",
      "Epoch [   17/   20] | d_loss: 0.6897 | g_loss: 2.1129\n",
      "Epoch [   17/   20] | d_loss: 0.2455 | g_loss: 1.4335\n",
      "Epoch [   17/   20] | d_loss: 0.7587 | g_loss: 2.2320\n",
      "Epoch [   17/   20] | d_loss: 0.4955 | g_loss: 2.8657\n",
      "Epoch [   17/   20] | d_loss: 0.6828 | g_loss: 2.8898\n",
      "Epoch [   17/   20] | d_loss: 0.4768 | g_loss: 3.7305\n",
      "Epoch [   18/   20] | d_loss: 0.2950 | g_loss: 4.5887\n",
      "Epoch [   18/   20] | d_loss: 0.4932 | g_loss: 1.7932\n",
      "Epoch [   18/   20] | d_loss: 0.2819 | g_loss: 3.7428\n",
      "Epoch [   18/   20] | d_loss: 0.4363 | g_loss: 2.8544\n",
      "Epoch [   18/   20] | d_loss: 1.1929 | g_loss: 2.7221\n",
      "Epoch [   18/   20] | d_loss: 0.7194 | g_loss: 3.2052\n",
      "Epoch [   18/   20] | d_loss: 0.7796 | g_loss: 3.2902\n",
      "Epoch [   18/   20] | d_loss: 0.5215 | g_loss: 2.8850\n",
      "Epoch [   18/   20] | d_loss: 0.2516 | g_loss: 2.1227\n",
      "Epoch [   18/   20] | d_loss: 0.3441 | g_loss: 2.7035\n",
      "Epoch [   18/   20] | d_loss: 0.2586 | g_loss: 3.6039\n",
      "Epoch [   18/   20] | d_loss: 0.8308 | g_loss: 4.4183\n",
      "Epoch [   18/   20] | d_loss: 0.3062 | g_loss: 4.3915\n",
      "Epoch [   18/   20] | d_loss: 0.6103 | g_loss: 2.5689\n",
      "Epoch [   18/   20] | d_loss: 0.3409 | g_loss: 2.6112\n",
      "Epoch [   18/   20] | d_loss: 0.7681 | g_loss: 4.5710\n",
      "Epoch [   18/   20] | d_loss: 0.6360 | g_loss: 3.7395\n",
      "Epoch [   18/   20] | d_loss: 0.4051 | g_loss: 4.1545\n",
      "Epoch [   18/   20] | d_loss: 0.1318 | g_loss: 3.8903\n",
      "Epoch [   18/   20] | d_loss: 0.3855 | g_loss: 2.9327\n",
      "Epoch [   18/   20] | d_loss: 0.3686 | g_loss: 3.3875\n",
      "Epoch [   18/   20] | d_loss: 0.2462 | g_loss: 5.4463\n",
      "Epoch [   18/   20] | d_loss: 0.2177 | g_loss: 3.8273\n",
      "Epoch [   18/   20] | d_loss: 0.8630 | g_loss: 3.2094\n",
      "Epoch [   18/   20] | d_loss: 0.5244 | g_loss: 4.2272\n",
      "Epoch [   18/   20] | d_loss: 0.2804 | g_loss: 3.8474\n",
      "Epoch [   18/   20] | d_loss: 0.3454 | g_loss: 3.5124\n",
      "Epoch [   18/   20] | d_loss: 0.2685 | g_loss: 3.0993\n",
      "Epoch [   18/   20] | d_loss: 0.6338 | g_loss: 3.7771\n",
      "Epoch [   18/   20] | d_loss: 0.0801 | g_loss: 5.7813\n",
      "Epoch [   18/   20] | d_loss: 0.2913 | g_loss: 4.5295\n",
      "Epoch [   18/   20] | d_loss: 0.4393 | g_loss: 4.1686\n",
      "Epoch [   18/   20] | d_loss: 0.3583 | g_loss: 2.9192\n",
      "Epoch [   18/   20] | d_loss: 0.6055 | g_loss: 2.9210\n",
      "Epoch [   18/   20] | d_loss: 0.3301 | g_loss: 3.6700\n",
      "Epoch [   18/   20] | d_loss: 0.3376 | g_loss: 3.9531\n",
      "Epoch [   18/   20] | d_loss: 0.1617 | g_loss: 3.5263\n",
      "Epoch [   18/   20] | d_loss: 0.6627 | g_loss: 3.8280\n",
      "Epoch [   18/   20] | d_loss: 0.8609 | g_loss: 1.9472\n",
      "Epoch [   18/   20] | d_loss: 0.4514 | g_loss: 3.2958\n",
      "Epoch [   18/   20] | d_loss: 0.2307 | g_loss: 3.8284\n",
      "Epoch [   18/   20] | d_loss: 0.8119 | g_loss: 2.8746\n",
      "Epoch [   18/   20] | d_loss: 0.3619 | g_loss: 2.7084\n",
      "Epoch [   18/   20] | d_loss: 0.7506 | g_loss: 4.0571\n",
      "Epoch [   18/   20] | d_loss: 0.9215 | g_loss: 4.0092\n",
      "Epoch [   18/   20] | d_loss: 0.1376 | g_loss: 5.4634\n",
      "Epoch [   18/   20] | d_loss: 0.3782 | g_loss: 4.9934\n",
      "Epoch [   18/   20] | d_loss: 0.4724 | g_loss: 3.7931\n",
      "Epoch [   18/   20] | d_loss: 0.3916 | g_loss: 3.2610\n",
      "Epoch [   18/   20] | d_loss: 0.3399 | g_loss: 3.9056\n",
      "Epoch [   18/   20] | d_loss: 0.3530 | g_loss: 2.9238\n",
      "Epoch [   18/   20] | d_loss: 0.3439 | g_loss: 2.4565\n",
      "Epoch [   18/   20] | d_loss: 0.1379 | g_loss: 3.7898\n",
      "Epoch [   18/   20] | d_loss: 0.5654 | g_loss: 3.5563\n",
      "Epoch [   18/   20] | d_loss: 0.2064 | g_loss: 3.5098\n",
      "Epoch [   18/   20] | d_loss: 0.7514 | g_loss: 4.0498\n",
      "Epoch [   18/   20] | d_loss: 0.5032 | g_loss: 2.2347\n",
      "Epoch [   19/   20] | d_loss: 0.7360 | g_loss: 3.2106\n",
      "Epoch [   19/   20] | d_loss: 0.5303 | g_loss: 2.7830\n",
      "Epoch [   19/   20] | d_loss: 0.4483 | g_loss: 4.5492\n",
      "Epoch [   19/   20] | d_loss: 0.9436 | g_loss: 0.8562\n",
      "Epoch [   19/   20] | d_loss: 0.4988 | g_loss: 3.2192\n",
      "Epoch [   19/   20] | d_loss: 0.3471 | g_loss: 3.6349\n",
      "Epoch [   19/   20] | d_loss: 0.2112 | g_loss: 5.2882\n",
      "Epoch [   19/   20] | d_loss: 0.4837 | g_loss: 2.8045\n",
      "Epoch [   19/   20] | d_loss: 0.6639 | g_loss: 3.0890\n",
      "Epoch [   19/   20] | d_loss: 0.3492 | g_loss: 5.2439\n",
      "Epoch [   19/   20] | d_loss: 0.4449 | g_loss: 3.5472\n",
      "Epoch [   19/   20] | d_loss: 0.2585 | g_loss: 4.2671\n",
      "Epoch [   19/   20] | d_loss: 0.2698 | g_loss: 4.3520\n",
      "Epoch [   19/   20] | d_loss: 0.7672 | g_loss: 3.4388\n",
      "Epoch [   19/   20] | d_loss: 0.7435 | g_loss: 2.7727\n",
      "Epoch [   19/   20] | d_loss: 0.1372 | g_loss: 4.8354\n",
      "Epoch [   19/   20] | d_loss: 0.1430 | g_loss: 4.0643\n",
      "Epoch [   19/   20] | d_loss: 0.7942 | g_loss: 6.6779\n",
      "Epoch [   19/   20] | d_loss: 0.3806 | g_loss: 4.5388\n",
      "Epoch [   19/   20] | d_loss: 0.1868 | g_loss: 4.3215\n",
      "Epoch [   19/   20] | d_loss: 0.7667 | g_loss: 2.6328\n",
      "Epoch [   19/   20] | d_loss: 0.1564 | g_loss: 4.8298\n",
      "Epoch [   19/   20] | d_loss: 0.1606 | g_loss: 3.0118\n",
      "Epoch [   19/   20] | d_loss: 0.0777 | g_loss: 4.3828\n",
      "Epoch [   19/   20] | d_loss: 1.3576 | g_loss: 5.9501\n",
      "Epoch [   19/   20] | d_loss: 0.5100 | g_loss: 2.1065\n",
      "Epoch [   19/   20] | d_loss: 0.8489 | g_loss: 2.1082\n",
      "Epoch [   19/   20] | d_loss: 0.2827 | g_loss: 3.0252\n",
      "Epoch [   19/   20] | d_loss: 0.4045 | g_loss: 3.4336\n",
      "Epoch [   19/   20] | d_loss: 0.5987 | g_loss: 2.4967\n",
      "Epoch [   19/   20] | d_loss: 0.5212 | g_loss: 4.0638\n",
      "Epoch [   19/   20] | d_loss: 0.2225 | g_loss: 4.3818\n",
      "Epoch [   19/   20] | d_loss: 0.3477 | g_loss: 4.5558\n",
      "Epoch [   19/   20] | d_loss: 0.1849 | g_loss: 4.1253\n",
      "Epoch [   19/   20] | d_loss: 0.6907 | g_loss: 4.2096\n",
      "Epoch [   19/   20] | d_loss: 0.8307 | g_loss: 5.1951\n",
      "Epoch [   19/   20] | d_loss: 0.1216 | g_loss: 5.6756\n",
      "Epoch [   19/   20] | d_loss: 0.2811 | g_loss: 3.0390\n",
      "Epoch [   19/   20] | d_loss: 0.1338 | g_loss: 4.8632\n",
      "Epoch [   19/   20] | d_loss: 0.6675 | g_loss: 2.5961\n",
      "Epoch [   19/   20] | d_loss: 0.4173 | g_loss: 4.3009\n",
      "Epoch [   19/   20] | d_loss: 0.4785 | g_loss: 3.4260\n",
      "Epoch [   19/   20] | d_loss: 0.3179 | g_loss: 4.4300\n",
      "Epoch [   19/   20] | d_loss: 0.3576 | g_loss: 3.0891\n",
      "Epoch [   19/   20] | d_loss: 0.2969 | g_loss: 2.8647\n",
      "Epoch [   19/   20] | d_loss: 0.5667 | g_loss: 2.4964\n",
      "Epoch [   19/   20] | d_loss: 0.2973 | g_loss: 4.0874\n",
      "Epoch [   19/   20] | d_loss: 0.5786 | g_loss: 4.3159\n",
      "Epoch [   19/   20] | d_loss: 0.0910 | g_loss: 3.8004\n",
      "Epoch [   19/   20] | d_loss: 0.4088 | g_loss: 3.8359\n",
      "Epoch [   19/   20] | d_loss: 0.1637 | g_loss: 3.0004\n",
      "Epoch [   19/   20] | d_loss: 0.4515 | g_loss: 3.4864\n",
      "Epoch [   19/   20] | d_loss: 0.9373 | g_loss: 3.1515\n",
      "Epoch [   19/   20] | d_loss: 0.4250 | g_loss: 3.9369\n",
      "Epoch [   19/   20] | d_loss: 0.6429 | g_loss: 3.5243\n",
      "Epoch [   19/   20] | d_loss: 1.9543 | g_loss: 2.7574\n",
      "Epoch [   19/   20] | d_loss: 0.5402 | g_loss: 2.9847\n",
      "Epoch [   20/   20] | d_loss: 0.3065 | g_loss: 2.3059\n",
      "Epoch [   20/   20] | d_loss: 0.4688 | g_loss: 4.0638\n",
      "Epoch [   20/   20] | d_loss: 0.5254 | g_loss: 2.5372\n",
      "Epoch [   20/   20] | d_loss: 0.6561 | g_loss: 3.3043\n",
      "Epoch [   20/   20] | d_loss: 0.2983 | g_loss: 3.1666\n",
      "Epoch [   20/   20] | d_loss: 0.3032 | g_loss: 4.0130\n",
      "Epoch [   20/   20] | d_loss: 0.1952 | g_loss: 4.2203\n",
      "Epoch [   20/   20] | d_loss: 0.2135 | g_loss: 5.4475\n",
      "Epoch [   20/   20] | d_loss: 0.4825 | g_loss: 3.5359\n",
      "Epoch [   20/   20] | d_loss: 0.4276 | g_loss: 4.3017\n",
      "Epoch [   20/   20] | d_loss: 0.0989 | g_loss: 4.0315\n",
      "Epoch [   20/   20] | d_loss: 0.5554 | g_loss: 4.2801\n",
      "Epoch [   20/   20] | d_loss: 0.1556 | g_loss: 3.5085\n",
      "Epoch [   20/   20] | d_loss: 0.2741 | g_loss: 3.3816\n",
      "Epoch [   20/   20] | d_loss: 0.2711 | g_loss: 2.9564\n",
      "Epoch [   20/   20] | d_loss: 0.4984 | g_loss: 4.2851\n",
      "Epoch [   20/   20] | d_loss: 0.5974 | g_loss: 4.7122\n",
      "Epoch [   20/   20] | d_loss: 0.2627 | g_loss: 4.6703\n",
      "Epoch [   20/   20] | d_loss: 0.9033 | g_loss: 2.3354\n",
      "Epoch [   20/   20] | d_loss: 0.7362 | g_loss: 3.6585\n",
      "Epoch [   20/   20] | d_loss: 0.0668 | g_loss: 4.1166\n",
      "Epoch [   20/   20] | d_loss: 0.4834 | g_loss: 2.5076\n",
      "Epoch [   20/   20] | d_loss: 0.3580 | g_loss: 2.8174\n",
      "Epoch [   20/   20] | d_loss: 0.1938 | g_loss: 5.1310\n",
      "Epoch [   20/   20] | d_loss: 0.3998 | g_loss: 3.9015\n",
      "Epoch [   20/   20] | d_loss: 0.2419 | g_loss: 4.8259\n",
      "Epoch [   20/   20] | d_loss: 0.3141 | g_loss: 4.2477\n",
      "Epoch [   20/   20] | d_loss: 0.4699 | g_loss: 3.1412\n",
      "Epoch [   20/   20] | d_loss: 0.1514 | g_loss: 5.5998\n",
      "Epoch [   20/   20] | d_loss: 0.7601 | g_loss: 1.8726\n",
      "Epoch [   20/   20] | d_loss: 0.2594 | g_loss: 3.7291\n",
      "Epoch [   20/   20] | d_loss: 0.3146 | g_loss: 4.1485\n",
      "Epoch [   20/   20] | d_loss: 0.3388 | g_loss: 4.0719\n",
      "Epoch [   20/   20] | d_loss: 0.4670 | g_loss: 5.8140\n",
      "Epoch [   20/   20] | d_loss: 0.2756 | g_loss: 3.8420\n",
      "Epoch [   20/   20] | d_loss: 0.1893 | g_loss: 3.2251\n",
      "Epoch [   20/   20] | d_loss: 0.2147 | g_loss: 2.8127\n",
      "Epoch [   20/   20] | d_loss: 1.0951 | g_loss: 2.6373\n",
      "Epoch [   20/   20] | d_loss: 0.6375 | g_loss: 5.3076\n",
      "Epoch [   20/   20] | d_loss: 0.4836 | g_loss: 3.6808\n",
      "Epoch [   20/   20] | d_loss: 0.6335 | g_loss: 2.2382\n",
      "Epoch [   20/   20] | d_loss: 0.1934 | g_loss: 3.0439\n",
      "Epoch [   20/   20] | d_loss: 0.4371 | g_loss: 3.3381\n",
      "Epoch [   20/   20] | d_loss: 0.3143 | g_loss: 2.9332\n",
      "Epoch [   20/   20] | d_loss: 0.5507 | g_loss: 1.7380\n",
      "Epoch [   20/   20] | d_loss: 0.6412 | g_loss: 3.5875\n",
      "Epoch [   20/   20] | d_loss: 0.1672 | g_loss: 4.0503\n",
      "Epoch [   20/   20] | d_loss: 0.1950 | g_loss: 5.4514\n",
      "Epoch [   20/   20] | d_loss: 0.6295 | g_loss: 4.7431\n",
      "Epoch [   20/   20] | d_loss: 0.1615 | g_loss: 3.9483\n",
      "Epoch [   20/   20] | d_loss: 0.8414 | g_loss: 5.1815\n",
      "Epoch [   20/   20] | d_loss: 1.0400 | g_loss: 2.1327\n",
      "Epoch [   20/   20] | d_loss: 0.1917 | g_loss: 4.3371\n",
      "Epoch [   20/   20] | d_loss: 0.4250 | g_loss: 3.0546\n",
      "Epoch [   20/   20] | d_loss: 0.2981 | g_loss: 4.5643\n",
      "Epoch [   20/   20] | d_loss: 0.1726 | g_loss: 4.6914\n",
      "Epoch [   20/   20] | d_loss: 0.3543 | g_loss: 4.7851\n"
     ]
    }
   ],
   "source": [
    "# set number of epochs \n",
    "n_epochs = 20\n",
    "\n",
    "\n",
    "# call training function\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": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f214349aeb8>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f21434ba278>"
      ]
     },
     "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": [
    "We can see that the loss of generator decreases and by the end of training the loss increases. Maybe I should have stopped early\n",
    "\n",
    "---\n",
    "\n",
    "## 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": 27,
   "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": 28,
   "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": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f214f3829b0>"
      ]
     },
     "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:**\n",
    "1. The generated samples are blurred mostly because the size of image is 32x32 images as input.\n",
    "2. The loss is not stable and the model has converged at nearly 10 epochs. I have trained for even more because I wanted to see if there will be any improvement. but I could not find any.\n",
    "3. To improve the model, Firstly I have to take images of bigger size as input\n",
    "4. Next, I need to use a different architecture in GAN for generating more realistic looking images like WGAN or LSGAN.\n",
    "5. Obviously I need to train for more epochs, but I have to choose a different loss function that will give good results apart from BCELoss, I could try MSELoss.\n",
    "6. Also I have to experiment with different beta1, beta2 and maybe in learning rate using STEP LR might work good. Need to experiment with these hyper parameters."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}