gandet.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision.transforms as transforms\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import datasets\n",
    "from torchvision.utils import save_image\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define constants\n",
    "epochs = 200\n",
    "batch_size = 64\n",
    "lr = 0.0002\n",
    "nworkers = 8\n",
    "latent_dim = 100\n",
    "sample_interval = 400\n",
    "img_shape = (1, 28, 28)\n",
    "adversarial_loss = torch.nn.BCELoss()\n",
    "device = 'cuda'\n",
    "valid = torch.ones(batch_size, 1, device=device)\n",
    "fake = torch.zeros(batch_size, 1, device=device)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Configure data loader\n",
    "os.makedirs('./data/mnist', exist_ok=True)\n",
    "transform = transforms.Compose([\n",
    "                        transforms.ToTensor(),\n",
    "                        transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))\n",
    "                   ])\n",
    "\n",
    "dataset = datasets.MNIST('./data/mnist', train=True, download=True, transform=transform)\n",
    "\n",
    "dataloader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, shuffle=True, drop_last=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Generator1(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Generator1, self).__init__()\n",
    "\n",
    "        def block(in_feat, out_feat, normalize=True):\n",
    "            layers = [nn.Linear(in_feat, out_feat)]\n",
    "            if normalize:\n",
    "                layers.append(nn.BatchNorm1d(out_feat, 0.8))\n",
    "            layers.append(nn.LeakyReLU(0.2, inplace=True))\n",
    "            return layers\n",
    "\n",
    "        self.model = nn.Sequential(\n",
    "            *block(latent_dim, 128, normalize=False),\n",
    "            *block(128, 256),\n",
    "            *block(256, 512),\n",
    "            *block(512, 1024),\n",
    "            nn.Linear(1024, int(np.prod(img_shape))),\n",
    "            nn.Tanh()\n",
    "        )\n",
    "\n",
    "    def forward(self, z):\n",
    "        img = self.model(z)\n",
    "        img = img.view(img.size(0), *img_shape)\n",
    "        return img\n",
    "\n",
    "\n",
    "class Discriminator1(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Discriminator1, self).__init__()\n",
    "\n",
    "        self.model = nn.Sequential(\n",
    "            nn.Linear(int(np.prod(img_shape)), 512),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Linear(512, 256),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Linear(256, 1),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "\n",
    "    def forward(self, img):\n",
    "        img_flat = img.view(img.size(0), -1)\n",
    "        validity = self.model(img_flat)\n",
    "\n",
    "        return validity\n",
    "\n",
    "class Generator2(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Generator2, self).__init__()\n",
    "\n",
    "        def block(in_feat, out_feat, normalize=True):\n",
    "            layers = [nn.Linear(in_feat, out_feat)]\n",
    "            if normalize:\n",
    "                layers.append(nn.BatchNorm1d(out_feat, 0.8))\n",
    "            layers.append(nn.LeakyReLU(0.2, inplace=True))\n",
    "            return layers\n",
    "\n",
    "        self.model = nn.Sequential(\n",
    "            *block(latent_dim, 128, normalize=False),\n",
    "            *block(128, 256),\n",
    "            *block(256, 512),\n",
    "            *block(512, 1024),\n",
    "            nn.Linear(1024, int(np.prod(img_shape))),\n",
    "            nn.Tanh()\n",
    "        )\n",
    "\n",
    "    def forward(self, z):\n",
    "        img = self.model(z)\n",
    "        img = img.view(img.size(0), *img_shape)\n",
    "        return img\n",
    "\n",
    "\n",
    "class Discriminator2(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(Discriminator2, self).__init__()\n",
    "\n",
    "        self.model = nn.Sequential(\n",
    "            nn.Linear(int(np.prod(img_shape)), 512),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Linear(512, 256),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Linear(256, 1),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "\n",
    "    def forward(self, img):\n",
    "        img_flat = img.view(img.size(0), -1)\n",
    "        validity = self.model(img_flat)\n",
    "\n",
    "        return validity    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "class GAN1(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.discriminator = Discriminator1().to(device)\n",
    "        self.generator = Generator1().to(device)\n",
    "\n",
    "    def forward(self, real_imgs):\n",
    "        # Generator Forward\n",
    "        z = torch.Tensor(np.random.normal(0, 1, (real_imgs.shape[0], latent_dim))).to(device)\n",
    "        self.gen_img = self.generator(z)\n",
    "        self.disc_gen = self.discriminator(self.gen_img)\n",
    "        # This clears the function graph built up for the discriminator\n",
    "        self.discriminator.zero_grad()\n",
    "\n",
    "        # Discriminator Forward\n",
    "        self.disc_gen_det = self.discriminator(self.gen_img.detach())\n",
    "        self.disc_real = self.discriminator(real_imgs)\n",
    "        \n",
    "        return self.gen_img, self.disc_gen, self.disc_gen_det, self.disc_real\n",
    "\n",
    "class GAN2(nn.Module):\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        self.discriminator = Discriminator2().to(device)\n",
    "        self.generator = Generator2().to(device)\n",
    "\n",
    "    def forward(self, real_imgs):\n",
    "        # Generator Forward\n",
    "        z = torch.Tensor(np.random.normal(0, 1, (real_imgs.shape[0], latent_dim))).to(device)\n",
    "        self.gen_img = self.generator(z)\n",
    "        self.disc_gen = self.discriminator(self.gen_img)\n",
    "        # This clears the function graph built up for the discriminator\n",
    "        self.discriminator.zero_grad()\n",
    "\n",
    "        # Discriminator Forward\n",
    "        self.disc_gen_det = self.discriminator(self.gen_img.detach())\n",
    "        self.disc_real = self.discriminator(real_imgs)\n",
    "        \n",
    "        return self.gen_img, self.disc_gen, self.disc_gen_det, self.disc_real\n",
    "\n",
    "class GAN3(nn.Module):\n",
    "    def __init__(self, disc_epoch, gen_epoch, disc_gan, gen_gan):\n",
    "        super().__init__()\n",
    "        gan1 = torch.load(\"gan%d/models/gan%d.pt\" %(gen_gan, gen_epoch))\n",
    "        gan2 = torch.load(\"gan%d/models/gan%d.pt\" %(disc_gan, disc_epoch))\n",
    "        self.generator = gan1.generator.to(device)\n",
    "        self.discriminator = gan2.discriminator.to(device)\n",
    "\n",
    "    def forward(self, real_imgs):\n",
    "        # Generator Forward\n",
    "        z = torch.Tensor(np.random.normal(0, 1, (real_imgs.shape[0], latent_dim))).to(device)\n",
    "        self.gen_img = self.generator(z)\n",
    "        self.disc_gen = self.discriminator(self.gen_img)\n",
    "        # This clears the function graph built up for the discriminator\n",
    "        self.discriminator.zero_grad()\n",
    "\n",
    "        # Discriminator Forward\n",
    "        self.disc_gen_det = self.discriminator(self.gen_img.detach())\n",
    "        self.disc_real = self.discriminator(real_imgs)\n",
    "        \n",
    "        return self.gen_img, self.disc_gen, self.disc_gen_det, self.disc_real    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def run_gan(loader, gan, optimizer_g, optimizer_d, epoch, gan_number):\n",
    "    for (img, _) in loader:\n",
    "        img = img.to(device)\n",
    "        gen_img, disc_gen, disc_gen_det, disc_real = gan(img)\n",
    "        optimizer_g.zero_grad()\n",
    "        fake_loss = adversarial_loss(disc_gen_det, fake)\n",
    "        real_loss = adversarial_loss(disc_real, valid)\n",
    "        g_loss = adversarial_loss(disc_gen, valid)\n",
    "        g_loss.backward()\n",
    "        optimizer_g.step()\n",
    "        \n",
    "        optimizer_d.zero_grad()\n",
    "        d_loss = (real_loss + fake_loss) / 2\n",
    "        d_loss.backward()\n",
    "        optimizer_d.step()\n",
    "#     print (\"[D loss: %f] [G loss: %f]\" % (d_loss.item(), g_loss.item()))\n",
    "    if epoch % 10 == 0 :\n",
    "        torch.cuda.empty_cache()\n",
    "        print (\"[D loss: %f] [G loss: %f]\" % (d_loss.item(), g_loss.item()))\n",
    "        model_file = \"gan%d/models/gan%d.pt\" %(gan_number,epoch)\n",
    "        img_file = \"gan%d/images/%d.png\" %(gan_number,epoch)\n",
    "        torch.save(gan, model_file)\n",
    "        save_image(gen_img.data[:25], img_file, nrow=5, normalize=True)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[D loss: 0.299966] [G loss: 3.164120]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type GAN1. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n",
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type Discriminator1. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n",
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type Generator1. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[D loss: 0.115951] [G loss: 3.328443]\n",
      "[D loss: 0.161005] [G loss: 3.308055]\n",
      "[D loss: 0.037031] [G loss: 3.642890]\n",
      "[D loss: 0.094922] [G loss: 4.218652]\n",
      "[D loss: 0.048096] [G loss: 3.359873]\n",
      "[D loss: 0.109202] [G loss: 5.441256]\n",
      "[D loss: 0.120905] [G loss: 4.200471]\n",
      "[D loss: 0.307409] [G loss: 7.524408]\n",
      "[D loss: 0.263560] [G loss: 4.578220]\n",
      "[D loss: 0.161660] [G loss: 3.900024]\n",
      "[D loss: 0.033263] [G loss: 4.226141]\n",
      "[D loss: 0.029113] [G loss: 6.682710]\n",
      "[D loss: 0.037671] [G loss: 5.838984]\n",
      "[D loss: 0.077551] [G loss: 3.722403]\n",
      "[D loss: 0.062109] [G loss: 3.572708]\n",
      "[D loss: 0.045669] [G loss: 7.441618]\n",
      "[D loss: 0.034755] [G loss: 5.344816]\n",
      "[D loss: 0.101658] [G loss: 5.577047]\n",
      "[D loss: 0.002658] [G loss: 7.635423]\n",
      "[D loss: 0.070197] [G loss: 5.292394]\n",
      "[D loss: 0.007477] [G loss: 5.460470]\n",
      "[D loss: 0.232115] [G loss: 9.060591]\n",
      "[D loss: 0.129756] [G loss: 6.455109]\n",
      "[D loss: 0.067354] [G loss: 4.747743]\n",
      "[D loss: 0.062026] [G loss: 5.958092]\n",
      "[D loss: 0.043424] [G loss: 4.833499]\n",
      "[D loss: 0.089330] [G loss: 2.694772]\n",
      "[D loss: 0.081356] [G loss: 4.710995]\n",
      "[D loss: 0.026783] [G loss: 4.588550]\n",
      "[D loss: 0.033145] [G loss: 4.353066]\n",
      "[D loss: 0.085445] [G loss: 2.573732]\n",
      "[D loss: 0.078955] [G loss: 3.735849]\n",
      "[D loss: 0.047535] [G loss: 3.673229]\n",
      "[D loss: 0.088010] [G loss: 8.274862]\n",
      "[D loss: 0.050465] [G loss: 2.860980]\n",
      "[D loss: 0.071583] [G loss: 3.277780]\n",
      "[D loss: 0.301336] [G loss: 4.987052]\n",
      "[D loss: 0.092636] [G loss: 4.743896]\n",
      "[D loss: 0.107820] [G loss: 4.356790]\n",
      "[D loss: 0.056139] [G loss: 5.361749]\n",
      "[D loss: 0.273139] [G loss: 3.292062]\n",
      "[D loss: 0.143383] [G loss: 3.603232]\n",
      "[D loss: 0.197245] [G loss: 4.035206]\n",
      "[D loss: 0.142864] [G loss: 3.398161]\n",
      "[D loss: 0.229786] [G loss: 4.353006]\n",
      "[D loss: 0.151884] [G loss: 7.462816]\n",
      "[D loss: 0.040618] [G loss: 4.275157]\n",
      "[D loss: 0.127610] [G loss: 3.811103]\n",
      "[D loss: 0.143395] [G loss: 4.004935]\n"
     ]
    }
   ],
   "source": [
    "gan1 = GAN1()\n",
    "optimizer_g = torch.optim.Adam(gan1.generator.parameters(), lr=lr)\n",
    "optimizer_d = torch.optim.Adam(gan1.discriminator.parameters(), lr=lr)\n",
    "\n",
    "for i in range(500):\n",
    "    run_gan(dataloader, gan1, optimizer_g, optimizer_d, i, 3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[D loss: 0.097600] [G loss: 3.067646]\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type GAN2. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n",
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type Discriminator2. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n",
      "/jet/var/python/lib/python3.6/site-packages/torch/serialization.py:241: UserWarning: Couldn't retrieve source code for container of type Generator2. It won't be checked for correctness upon loading.\n",
      "  \"type \" + obj.__name__ + \". It won't be checked \"\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[D loss: 0.167209] [G loss: 2.883319]\n",
      "[D loss: 0.180146] [G loss: 2.690252]\n",
      "[D loss: 0.102366] [G loss: 2.881555]\n",
      "[D loss: 0.073220] [G loss: 4.795817]\n",
      "[D loss: 0.185620] [G loss: 4.231509]\n",
      "[D loss: 0.018788] [G loss: 5.192479]\n",
      "[D loss: 0.227863] [G loss: 6.585853]\n",
      "[D loss: 0.072446] [G loss: 5.536769]\n",
      "[D loss: 0.049745] [G loss: 4.304088]\n",
      "[D loss: 0.163303] [G loss: 3.461035]\n",
      "[D loss: 0.079035] [G loss: 4.695574]\n",
      "[D loss: 0.026474] [G loss: 6.172735]\n",
      "[D loss: 0.109875] [G loss: 6.399917]\n",
      "[D loss: 0.066072] [G loss: 4.646819]\n",
      "[D loss: 0.166195] [G loss: 4.012285]\n",
      "[D loss: 0.073047] [G loss: 7.679369]\n",
      "[D loss: 0.113361] [G loss: 8.163986]\n",
      "[D loss: 0.202540] [G loss: 6.109647]\n",
      "[D loss: 0.066149] [G loss: 3.045276]\n",
      "[D loss: 0.144828] [G loss: 4.141537]\n",
      "[D loss: 0.011151] [G loss: 7.149577]\n",
      "[D loss: 0.025199] [G loss: 3.939393]\n",
      "[D loss: 0.223856] [G loss: 4.794902]\n",
      "[D loss: 0.103146] [G loss: 4.092116]\n",
      "[D loss: 0.009292] [G loss: 5.005027]\n",
      "[D loss: 0.043777] [G loss: 5.915321]\n",
      "[D loss: 0.026891] [G loss: 7.510965]\n",
      "[D loss: 0.143149] [G loss: 6.103864]\n",
      "[D loss: 0.168935] [G loss: 4.585710]\n",
      "[D loss: 0.150702] [G loss: 3.938515]\n",
      "[D loss: 0.043047] [G loss: 5.591506]\n",
      "[D loss: 0.063134] [G loss: 3.955415]\n",
      "[D loss: 0.023231] [G loss: 6.081146]\n",
      "[D loss: 0.218902] [G loss: 4.478435]\n",
      "[D loss: 0.111325] [G loss: 7.093885]\n",
      "[D loss: 0.194092] [G loss: 5.083550]\n",
      "[D loss: 0.126693] [G loss: 3.193168]\n",
      "[D loss: 0.082014] [G loss: 4.455947]\n",
      "[D loss: 0.102939] [G loss: 3.328770]\n",
      "[D loss: 0.105304] [G loss: 4.265505]\n",
      "[D loss: 0.042508] [G loss: 5.003506]\n",
      "[D loss: 0.047066] [G loss: 3.457253]\n",
      "[D loss: 0.044456] [G loss: 5.176131]\n",
      "[D loss: 0.108836] [G loss: 3.596637]\n",
      "[D loss: 0.209554] [G loss: 5.375165]\n",
      "[D loss: 0.144332] [G loss: 5.118526]\n",
      "[D loss: 0.123963] [G loss: 5.806181]\n",
      "[D loss: 0.108956] [G loss: 3.589151]\n",
      "[D loss: 0.218346] [G loss: 3.649876]\n"
     ]
    }
   ],
   "source": [
    "gan2 = GAN2()\n",
    "optimizer_g = torch.optim.Adam(gan2.generator.parameters(), lr=lr)\n",
    "optimizer_d = torch.optim.Adam(gan2.discriminator.parameters(), lr=lr)\n",
    "\n",
    "for i in range(500):\n",
    "    run_gan(dataloader, gan2, optimizer_g, optimizer_d, i, 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def eval_gan(loader, gan):\n",
    "    g_losses = 0\n",
    "    d_losses = 0\n",
    "    for (img, _) in loader:\n",
    "        img = img.to(device)\n",
    "        gen_img, disc_gen, disc_gen_det, disc_real = gan(img)\n",
    "        fake_loss = adversarial_loss(disc_gen_det, fake)\n",
    "        real_loss = adversarial_loss(disc_real, valid)\n",
    "        g_loss = adversarial_loss(disc_gen, valid)\n",
    "        d_loss = (real_loss + fake_loss) / 2\n",
    "        g_losses += g_loss.item()\n",
    "        d_losses += d_loss.item()\n",
    "#         torch.cuda.empty_cache()\n",
    "    return g_losses, d_losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "g_losses = []\n",
    "d_losses = []\n",
    "for i in range(0, 100, 10):\n",
    "    gan3 = GAN3(disc_epoch=i, gen_epoch=90,disc_gan=1, gen_gan=2)\n",
    "    g_loss, d_loss = eval_gan(dataloader, gan3)\n",
    "    g_losses.append(g_loss)\n",
    "    d_losses.append(d_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAaEAAAEWCAYAAADPZygPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkpAxxpjA/D/eqsu/exkl7gAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots( nrows=1, ncols=1 )\n",
    "ax.plot(d_losses)\n",
    "plt.ylabel(\"dicriminator loss\")\n",
    "plt.xlabel(\"discriminator epochs for fixed generator\")\n",
    "plt.title(\"Discriminator loss agains various epoch with a fixed generator\")\n",
    "plt.show()\n",
    "fig.savefig('disc_loss_fixed_gen.png')\n",
    "plt.close(fig) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZcAAAEWCAYAAACqitpwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4xLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvDW2N/gAAIABJREFUeJzt3Xd4FdXWwOHfCqH3Lr0XQXpoIoig\nFBv2LkUQC3bs189+vXr1XhG9NhQFKQqogIogIoqFFnqREnontBA6Sdb3x+zIISbkJJyTSVnv8+TJ\nnD1tzZw5s2b2zOwRVcUYY4wJpQi/AzDGGJP7WHIxxhgTcpZcjDHGhJwlF2OMMSFnycUYY0zIWXIx\nxhgTcpZcchAR2SgiF/sdx9kSkUMiUtvvONIiItVdjPn8jiWcwrE9icjTIvLRGfr3FZHfQjnPgGmL\niHwiIvtFZJ6IdBSR1WGYT00RURGJDPW0c5MclVxE5CYRmSsih0Vkt+u+V0TE79hSEpGfRWSA33Fk\nR6paTFXX+x1HWlR1s4sx0e9YchpVfUVVB4AvO+ELgEuAqqraRlV/VdUGWTTvHCeciR5yUHIRkcHA\nW8DrwDlAReBuoANQIItjCeuPxR2B5ZjvJjexo9EcrQawUVUP+x2I37JiO053Hqqa7f+AksBh4Np0\nhisIvAFsBnYB7wOFXb/OwFZgMLAb2AH0y+C4TwA7gc+A0sC3QCyw33VXdcP/E0gEjgGHgHdc+fnA\nfCDO/T8/YP4/u/F+B44CdVNZvo3AxQHxDgG2u78hQEHXr5yL5wCwD/gViHD9ngC2AfHAaqBrGuvy\nMmARcBDYAjyfon9vYBOwF/i/FLG1AWa7+e8A3gEKBIyrycsHfAr8D/jOxTQXqOP6CfCm+74OAsuA\n81KJ9UYgOkXZw8Dk9JYFqOni6e+++1kBZZFumMrAZLcuY4A7A8b/FHg54HNnYGvA52DXdzDb39PA\nHreub03x+xiJty1uAp5J/r5d/zuBP10MK4GWAdvTo8BSvG3yC6BQGvFtAlq57lvd+mnsPvcHJrru\n54FRrnuzG+6Q+2sP9AV+c8u6H9gA9DzDb/pJYF1A7FenMVx/vN9bopvXC4HfBVDHfX/Jy17Zra/O\nAevwY7ztdRvwMpDP9cvn4t0DrAcGBW4fqcTSEm97iwfGu/UauI1cDizG+338ATRN8RtP8zsJYtwn\n3LjHgci01h9wbor1dSC9bcl9d7/j/Sb3Bi5TqushVAkgnH9ADyAhrS8zYLg38XYCZYDiwDfAvwJ+\noAnAi0B+4FLgCFA6A+O+hrcTKAyUBa4Firjhx+N+YG6cn4EBAZ/L4P2Ybndf+s3uc9mA4TcDjV3/\n/Kks30ZO7cBfBOYAFYDybkN7yfX7F97OKb/764i3o26At3OtHLBjrZPGuuwMNME7u22Kt8O7yvVr\n5DbIC/DOGt8ATgbE1gpo55ajJt6O7aGAaadMLnvxElIkMBr43PXrDiwASrn4zwUqpRJrEbwfT72A\nsvnATUEsS00Xz0igqPtuk8uSk8ss4F2gENAc74fXJSD+VJNLBtd3MNvff/G2vwvxDrYauP4jgUlu\nvJrAGqC/63c93s6ytVuHdYEaAdvTPLwdbRn3Pd2dRnwjgcGu+0O8HdY9Af0edt3Pcyq5nLYeA3ZQ\nJ/ESXj7gHryDI0ljvte7+CLwDiIOp7YNBEz7t9S+C/f5TrwdbBFgGvBGQL+vgQ/cNlDBrZe7XL+7\ngVVANbeeZqZcroDpFMDbKT+I99u7BjiRvI0ALfAOltq65e/jvofkA8M0v5Mgx13s4iyc3vpLub6C\n2Jb64m2H9+P9VgufcX8croQQyj/gNmBnirI/8LL3UaAT3g/nMAE/XrwjpQ0BG9pRTt/Qd+PtBIMZ\n9wRpHNW5YZoD+wM+/8zpyeV2YF6KcWYDfQOGfzGd9bCRUzvwdcClAf2641UJgJd4JpHi7Advx7Ib\nuJhUklc68x4CvOm6nwXGBvQr4tbPxWmM+xDwdcDnlMnlo4B+lwKrXHcXt3G3I+BIPI15jAKedd31\n8JJNkSCWpaaLp3ZA/+SySLwfaiJQPKD/v4BPA+JPK7kEtb6D3P4SgKIB/cfhnTHmc+u+UUC/u4Cf\nXfc04MEzbE+3BXz+N/B+GsP259SZ4J/AAE4dBGzi1BnB86SfXGJSbDsKnBPkdrgY6JVGv76cIbm4\nssl4Z8BLObVTroh3pF84YLibgZmu+ycCki7QLeVyBfTrhJfMJaDsN04ll/dwB4EB/VcDF6b3nQQ5\n7h3Brr9U1ld621JfYHMw35Oq5phrLnuBcoF1fKp6vqqWcv0i8I7eiwALROSAiBwAprryv6ajqgkB\nn48AxYIcN1ZVjyV/EJEiIvKBiGwSkYN4R7elznCHUWW8H2GgTUCVgM9bzrwazji9Ta4MvOtSMcAP\nIrJeRJ4EUNUYvB3988BuEflcRCqTChFpKyIzRSRWROLwjt7KBcz7r1hV9Qje95A8bn0R+VZEdrp1\n80rAuKnZGdCd/J2gqj/hVan9z8X7oYiUSGMaY/B2CAC34J1FHgliWZKlte4rA/tUNT6gLOX3lqoM\nrO9gtr/9evq1hOTvuxzeEXLKbSE5vmp4ByJpSXXdp+IXoKOIVMLbCY0DOohITbyqlMVnmEea80z+\njtKar4j0FpHFAevlPM68LaVnmJvG26p63JXVwFuHOwLm8wHeGQyk2N75++84UGVgm7q9sRM4bg1g\ncPJ83Lyqceq3C2l/J8GMe9p2nMH1l9629Lfpn0lOSS6z8Y4sep1hmD14ZyaNVbWU+yupqmn9WDI6\nrqYYZzBetUdbVS2Bd8QC3lFoasNvx9s4AlXHO8pJax5nknJ61V0ZqhqvqoNVtTZwJfCIiHR1/cao\n6gVuXMWr6kvNGLyjvGqqWhKvmi152XYAVZMHFJHkasJk7+FVI9Rz6+bpgHEzRFWHqmorvKq4+sBj\naQw6HSgvIs3xksyYIJflr1mlMd3tQBkRKR5QFvi9HcZLDMnOSRF/MOs7mO2vtIgUTRHDdjfuSf6+\nLSTHtwXvesNZcYnyCF6VyCxVPYi3ExyId/SblNpoZzNPEamBlwzuw6s+LgUsJ5PbkogUwztr/Rh4\nXkTKuF5b8PYv5QLWfwlVbez678DbiSerfobZ7ACqpLiDNXDcLcA/A+ZTSlWLqOrYIBYhmHH/WudB\nrL+U309621Jq46QpRyQXVT2Ad4HuXRG5TkSKi0iE25EUdcMk4a3IN0WkAoCIVBGR7kFMPzPjFsfb\nIRxwG+lzKfrvAgKf5ZgC1BeRW0QkUkRuxNthfpvuCkjdWOAZESkvIuXwqqpGudgvF5G6bgOPw6vW\nSRKRBiLSRUQK4l3MOwqktlNIXr59qnpMRNrgnQ0kmwBcISLni0gBvCNzSTHuQeCQiDTEq1fPMBFp\n7c468uPtxI+lFa+qnsS77vU6Xl319CCX5YxUdQteFey/RKSQiDTFqyIa5QZZDFwqImVE5By8M5Xk\n+INa3xnY/l4QkQIi0hHvwu549W6XHgf80/0uagCPBMT3EfCoiLRydyHWdcNkxi94O6pf3OefU3xO\nKdYtb2afaSqKtzOLBRCRfnhH3pn1Ft6NHwPwbiB5H0BVdwA/AP8RkRJu31JHRC50440DHhCRqiJS\nGu8ieVpm4/3e7nO/81541xOTDQPudtu1iEhREbksxcFLWjI6bnrrbxdQ1f2GCWJbypAckVwAVPXf\neAv6ON5K2YV36voE3o8f1x0DzHHVMT/inV0EI6PjDsG7+LsH78L61BT93wKuE++BrqGquhdvhzAY\nrwrpceByVd0TZHwpvQxE49UdLwMWujLwrjn8iHfRfTbwrqrOxLsY/KqLeSfeaf9TaUz/XuBFEYnH\nS1zjknuo6gq8I9jP8Y7UDuFdW0iuZngUbwcej/eD+CKTy1jCjb+fU3emvX6G4cfgXd8Yn6L6M81l\nCdLNeNcPtuNd+H1OVX90/T4DluDVd//A6cuakfWd3va3E289bMe76eFuVV3l+t2Pl3zX49XvjwGG\nA6jqeLy7EMfgfR8T8ZJvZvyCl6hnpfH5NK7K65/A765apl1GZqaqK4H/4G3Du/Buyvg9M4G7nXwP\nTh3oPAK0FJFb3efeeBfjV+Kt5wlAJddvGN61qyV4v7OvzhDzCbyL+P3xrgnfhncAedz1j8a7seAd\nN58YvGsZ6crouEGsv5+AFcBOEUneD6W5LWWUnF41aEzGueqGA3jVYBv8jie3EZHOeBfJq6Y3rMl+\nRGQu3kX5T/yOJSvlmDMXk72IyBXi3dRQFO9W5GV4R+/G5GkicqGInOOqxfrg3f6esmYj17PkYjKr\nF6ce4KyH90yJnQYb41VnLsE7mx8MXOeu6+QpVi1mjDEm5OzMxRhjTMjlykb6ypUrpzVr1vQ7DGOM\nyVEWLFiwR1XLpz9k+nJlcqlZsybR0dF+h2GMMTmKiJyp9YEMCVu1mHuAbHHA30ERecg9bDZdRNa6\n/6Xd8CIiQ0UkRkSWikjLgGn1ccOvdXdfGGOMycbCllxUdbWqNlfV5nit5B7BewDtSWCGqtYDZnDq\nadeeeHcd1cNrUuI9gICn39viPen6XHJCMsYYkz1l1QX9rsA6Vd2EdwvrCFc+ArjKdfcCRqpnDl4j\nkJXwWvudrqr7VHU/XrMePbIobmOMMZmQVcnlJry2sAAqBtzzvROvuWvwWt4MbHFzqytLq/w0IjJQ\nRKJFJDo2NjaUsRtjjMmgsCcX1yjalXiNCp7GPXQXkgdtVPVDVY1S1ajy5UNys4MxxphMyoozl57A\nQlXd5T7vctVduP+7Xfk2Tm+auqorS6vcGGNMNpUVyeVmTlWJgfdejeQ7vvrgvTExuby3u2usHRDn\nqs+mAd1EpLS7kN/NlRljjMmmwppcXKOGl3B6E9WvApeIyFq85tFfdeVT8Jp5jsFr4vpeAFXdB7yE\n9070+XivAt4XzriNMTnDuthDjI/egjVjlf2E9SFK91rWsinK9uLdPZZyWAUGpTGd4WTynQLGmNzp\n26XbeXzCUo6cSGTxlgO81Os8IiIy9ZJKEwa58gl9Y0zudSIhiX99/yef/L6RVjVK06RKST79YyNJ\nCv+8yhJMdmHJxRiTY+yMO8agMQtZsGk/d3SoxVOXNiQyQihWMJJ3ZsagqrxydRNLMNmAJRdjTI7w\nR8weHvh8EUdOJPLOLS24vGnlv/oN7lafCIGhP8WQmKS8dm1TSzA+s+RijMnWkpKU92et441pq6ld\nvhifD2xJ3QrFTxtGRHikWwMiIoQhP65FgdeubUo+SzC+seRijMm24o6eZPC4Jfz45y4ub1qJ165t\nStGCae+2Hrq4PoLw5o9rSEpSXr++mSUYn1hyMcZkSyu2x3Hv6IVs23+U569oRJ/zayKSfqJ48OJ6\n5IuAN35YQ5Iq/7mhuSUYH1hyMcZkO+Ojt/DMxOWULlKAL+5qR6saZTI0/n1d6iEivD5tNUkK/72h\nGZH57MW7WcmSizEm2zh2MpEXvlnB2HlbaF+7LG/f0oJyxQpmalqDLqpLvgjh1e9XkaTKkBubW4LJ\nQpZcjDHZwpZ9R7hn9AKWbzvIvZ3r8Mgl9c86Gdx9YR0iBF6ZsgpVGHJTc/Ln4gSz99BxElWpULyQ\n36FYcjHG+G/mqt089MViklQZ1juKSxpVTH+kIA3sVIcIEV7+7k+SVBl6c4tcmWB+WLGTp79eRovq\npRnWO8rvcCy5GGP8k5ikvPXjGob+FMO5lUrw/m0tqVG2aMjnM6BjbUSEl75dyX1jFvL2zS0pEJk7\nEszBYyd5YfJKvly4lUaVSjC4W32/QwIsuRhjfLLv8Ake/HwRv67dw3WtqvLyVedRKH++sM2v/wW1\nyCfw/DcrGTRmIf+7JecnmN9j9vDY+CXsij/O/V3qcn+XetlmmSy5GGOy3OItB7h31AL2HD7Bq9c0\n4cbW1YK6zfhs9e1Qi4gI4dlJK7h39AL+d2tLCkaGL6GFy5ETCbz6/SpGzt5E7fJF+fKe82lerZTf\nYZ3GkosxJsuoKqPmbubFb1ZQsUQhvrz7fJpULZmlMfRu7z0v838Tl3PPqIW8d1vOSjALNu1n8LjF\nbNx7hDs61OLxHg3CesaXWZZcjDFZ4siJBP7x9XK+XrSNzg3KM+TG5pQqUsCXWG5vV4N8Ijz99TLu\n/mwB793WKlvuoAMdT0hkyI9r+eCXdVQqWZixd7ajfZ2y6Y/oE0suxpiwWx97iHtGLWTN7ngeuaQ+\n911U1/eGJW9pW50IgSe/WsZdny3gg9uzb4JZuf0gj4xbzKqd8dwYVY1nLj+X4oXy+x3WGVlyMcaE\n1dTlO3h0/FLy5xNG9GtDp/rl/Q7pLze1qU6ECE98tZQ7R0YzrHdUtkowCYlJfDBrPUN+XEOpIgUY\n3jeKLg1Dd5t2OFlyMcaERUJiEv+etpoPZ62nWbVSvHtrS6qUKux3WH9zQ+tqiMDjXy5lwAgvwRQu\n4H+CWRd7iMHjlrB4ywEua1qJl3udR+mi/lQjZoYlF2NMyO0+eIz7xi5i3oZ93N6uBs9cfm62vmh+\nfVQ1IkR4dMIS+o+Yz8d9WvuWYJKSlBGzN/La1FUUyp+Pt29uwRXNKqc7XnZjycUYE1Jz1+/lvrGL\niD92kjdvbMbVLar6HVJQrm1VlYgIGDxuCf0+ncfwvq0pUiBrd5Fb9x/hsfFLmb1+Lxc1KM9r1zal\nQgn/m3LJjLA+bSMipURkgoisEpE/RaS9iJQRkekistb9L+2GFREZKiIxIrJURFoGTKePG36tiPQJ\nZ8zGmMxRVYbNWs8tH82lWMFIJg7qkGMSS7KrW1TlzRubM2/DPvp+Mp/DxxOyZL6qyrj5W+gx5FeW\nbj3Aq9c0YXjf1jk2sUD4z1zeAqaq6nUiUgAoAjwNzFDVV0XkSeBJ4AmgJ1DP/bUF3gPaikgZ4Dkg\nClBggYhMVtX9YY7dGBOk+GMneWz8Uqau2En3xhV5/fpmlMjmdzOlpVfzKogID3+xmH6fzGd4v9YU\nO8MLys7W7vhjPPXlMmas2k3bWmV44/pmVCtTJGzzyyphW2MiUhLoBPQFUNUTwAkR6QV0doONAH7G\nSy69gJGqqsAcd9ZTyQ07XVX3uelOB3oAY8MVuzEmeKt3xnP3qAVs3neEpy9tyJ2uHa+c7MpmlYkQ\nePDzxfQdPo9P72gTlgTz3dIdPDNxGUdOJPJ/lzei3/k1fb9FO1TCeeZSC4gFPhGRZsAC4EGgoqru\ncMPsBJLvq6sCbAkYf6srS6v8NCIyEBgIUL169dAthTEmTRMXbeOpr5ZRtGAkowe0pV3t7PtQX0Zd\n3rQyESI8MHYRfYbP49N+rUP2bMmBIyd4dtIKJi/ZTrOqJfnPDc2oW6F4SKadXYTzmksk0BJ4T1Vb\nAIfxqsD+4s5SNBQzU9UPVTVKVaPKl88+99EbkxudTEzi2UnLeeiLxTSpUpIpD1yQqxJLskubVOKd\nW1qwZMsBeg+fx8FjJ896mjNX76bbm7OYsmwHj1xSny/vOT/XJRYIb3LZCmxV1bnu8wS8ZLPLVXfh\n/u92/bcB1QLGr+rK0io3xvjk0983MnL2JgZcUIvRd7bN0Ree09PjvEr879aWLN8Wx+0fzyPuaOYS\nzKHjCTz11VL6fTKfUkXyM3FQBx7oWi/Xvh0zbEulqjuBLSLSwBV1BVYCk4HkO776AJNc92Sgt7tr\nrB0Q56rPpgHdRKS0u7Osmyszxvgg7uhJ3pkZQ6f65Xnm8ka58sVbKXVvfA7v3tqKldvj6P3x3Awn\nmLnr99LzrVl8Pn8Ld11Ym2/uv4DzqmRtg51ZLdx3i90PjHZ3iq0H+uEltHEi0h/YBNzghp0CXArE\nAEfcsKjqPhF5CZjvhnsx+eK+MSbrffDLOuKOnuSJHg3SHzgXuaRRRd6/rRX3jFrIbR/NZVT/tpQs\ncuZrMMdOJvLGtNV8/PsGqpcpwvi72hNVs0wWRewv8S575C5RUVEaHR3tdxjG5Do7447R+Y2Z9Gh8\nDkNuauF3OL6YuWo3d322gPrnFGNU/7Zptuy8dOsBHhm3hJjdh7itXXWe6nkuRcN4S3MoiMgCVQ3J\nO5Jz//msMSZk3pqxhsQkZXC3vHXWEuiihhX4sHcr1uw6xC3D5rL/8InT+p9MTOK/09dw9bt/cOhY\nAiPvaMPLVzXJ9okl1Cy5GGOCErP7EOOit3Jr2xq54iG/s9G5QQWG9Y5iXewhbvloLvtcglmzK56r\n3/2doTPW0qtZZaY91ClbtQKdlfJWKjXGZNob01ZTOH8+7u9S1+9QsoUL65fnoz5RDBgRzS3D5nBF\ns8q8NWMtxQtG8v5tLelxXiW/Q/SVnbkYY9K1cPN+pq7YyZ0da1O2WEG/w8k2OtYrz/C+rdm49zCv\nT1tN5/rlmfZwpzyfWMDOXIwx6VBVXv1+FeWKFWBAx1p+h5PtdKhbjnF3tWdH3DG6NaqY45u+CRVL\nLsaYM/p5dSzzNuzjpV6N89xF6WA1rVqKpjmrAeiws2oxY0yaEpOU16auokbZItzUxtrsM8Gz5GKM\nSdPERdtYtTOeR7s1yBNP4pvQsa3FGJOqYycT+e/0NTSpUpLLmtgFapMxllyMMakaNWcT2w4c5cme\nDXPNO0ZM1rHkYoz5m4PHvMYpO9YrR4e65fwOx+RAllyMMX/z4S/rOXDkJE/0aOh3KCaHsuRijDnN\n7oPH+Oi39VzZrHKubxbehI8lF2PMaYbMWEtCojK4W32/QzE5mCUXY8xf1sUe4ov5W7i1bXVqlC3q\ndzgmB7PkYoz5y39+WE2hyAju71rP71BMDmfJxRgDwKLN+5mybCd3dqpNOWuc0pwlSy7GGFS9Zl7K\nFi3AgI61/Q7H5AKWXIwx/LImljnr9/FA13oUs8YpTQhYcjEmj0tK8prUr16mCDdb45QmRCy5GJPH\nTVriNU45uFt9CkTaLsGERli3JBHZKCLLRGSxiES7sjIiMl1E1rr/pV25iMhQEYkRkaUi0jJgOn3c\n8GtFpE84YzYmLzmekMh/flhD48oluKJpZb/DMblIVhymXKSqzVU1yn1+EpihqvWAGe4zQE+gnvsb\nCLwHXjICngPaAm2A55ITkjHm7Iyes5mt+61xShN6fpwD9wJGuO4RwFUB5SPVMwcoJSKVgO7AdFXd\np6r7gelAj6wO2pjc5uCxk7z901ouqFuOjvXK+x2OyWXCnVwU+EFEFojIQFdWUVV3uO6dQEXXXQXY\nEjDuVleWVvlpRGSgiESLSHRsbGwol8GYXGnYrPXst8YpTZiE+57DC1R1m4hUAKaLyKrAnqqqIqKh\nmJGqfgh8CBAVFRWSaRqTW+2OP8ZHv27g8qaVaFLVGqc0oRfWMxdV3eb+7wa+xrtmsstVd+H+73aD\nbwOqBYxe1ZWlVW6MyaShM9ZyMjGJR7s18DsUk0uFLbmISFERKZ7cDXQDlgOTgeQ7vvoAk1z3ZKC3\nu2usHRDnqs+mAd1EpLS7kN/NlRljMmHDnsOMnbeFm9tUp2Y5a5zShEc4q8UqAl+LSPJ8xqjqVBGZ\nD4wTkf7AJuAGN/wU4FIgBjgC9ANQ1X0i8hIw3w33oqruC2PcxuRqb/ywmoKREdzfta7foZhcLGzJ\nRVXXA81SKd8LdE2lXIFBaUxrODA81DEak9cs2XKA75bu4IGu9ahQvJDf4ZhczB7HNSaPSG6cskzR\nAtzZsZbf4ZhczpKLMXnEr2v38Me6vdzfpS7FC+X3OxyTy1lyMSYPSG6cslqZwtzS1hqnNOFnycWY\nPOCbpdtZueMgj3ZrQMHIfH6HY/IASy7G5HInEpJ444fVNKpkjVOarJOh5OKeNWkarmCMMaE3Zu4m\ntuw7yhPWOKXJQukmFxH5WURKuNaJFwLDROS/4Q/NGHO24o+dZOhPMZxfpyyd6pXzOxyThwRz5lJS\nVQ8C1+C1WtwWuDi8YRljQmHYrxvYd/gET/RoiHug2ZgsEUxyiXRtgN0AfBvmeIwxIRIbf5yPfl3P\nZU0q0axaKb/DMXlMMMnlRby2vGJUdb6I1AbWhjcsY8zZevuntRxPSOLR7tY4pcl66Tb/oqrjgfEB\nn9cD14YzKGPM2dm45zBj5m7m5jbVqGWNUxofBHNB/9/ugn5+EZkhIrEicltWBGeMyZz/TF9D/nwR\nPNC1nt+hmDwqmGqxbu6C/uXARqAu8Fg4gzLGZN6yrXF8s2Q7AzrWssYpjW+CuqDv/l8GjFfVuDDG\nY4w5S69NXUXpIvkZ2Km236GYPCyY5PKtez1xK2CGiJQHjoU3LGNMZvy6NpbfYvZwX5d61jil8VW6\nyUVVnwTOB6JU9SRwGOgV7sCMMRmTlOQ1qV+lVGFua2eNUxp/pXu3mIjkB24DOrmHsH4B3g9zXMaY\nDPp22Q6WbzvImzc2s8Ypje+CeRPle0B+4F33+XZXNiBcQRljMuZEQhJvTFtNw3OK06tZFb/DMSao\n5NJaVQNfV/yTiCwJV0DGmIz7fP5mNu87wif9WlvjlCZbCOaCfqKI1En+4J7QTwxfSMaYjDh0PIGh\nM9bSrnYZOtcv73c4xgDBJZfHgJmudeRfgJ+AwcHOQETyicgiEfnWfa4lInNFJEZEvhCRAq68oPsc\n4/rXDJjGU658tYh0z8gCGpPbffTrevYcOsGTPc+1xilNthHM3WIzgHrAA8D9QANVnZmBeTwI/Bnw\n+TXgTVWtC+wH+rvy/sB+V/6mGw4RaQTcBDQGegDviohdrTQG2HPoOMNmrafneefQ3BqnNNlImslF\nRK5J/sN7gLKu+7vMlaVLRKq6cT9ynwXoAkxwg4wArnLdvdxnXP+ubvhewOeqelxVNwAxQJvgF9GY\n3Oudn2I4Zo1TmmzoTBf0rzhDPwW+CmL6Q4DHgeLuc1nggKomuM9bgeRbW6oAWwBUNUFE4tzwVYA5\nAdMMHOcvIjIQGAhQvbrd429yv017DzN67iZubF2NOuWL+R2OMadJM7moar+zmbCIXA7sVtUFItL5\nbKYVDFX9EPgQICoqSsM9P2P89p8f1pAvQnjQGqc02VAwtyJnVgfgShG5FCgElADeAkqJSKQ7e6kK\nbHPDbwOqAVtFJBIoCewNKE8WOI4xedLybXFMXrKdQRfVoWIJa5zSZD/B3C2WKar6lKpWVdWaeBfk\nf1LVW4GZwHVusD7AJNc92X0SdePRAAAgAElEQVTG9f9JVdWV3+TuJquFd3PBvHDFbUxO8NrUVZQq\nkp+7LqyT/sDG+OCMyUVEIkTk/BDP8wngERGJwbum8rEr/xgo68ofAZ4EUNUVwDhgJTAVGKSq9pyN\nybN+j9nDr2v3cN9FdSlhjVOabEq8k4MzDCCySFVbZFE8IREVFaXR0dF+h2FMyJ1ISOLa9/5g3+ET\nzBh8IYXy2135JnREZIGqRoViWsFUi80QkWvFns4yxldb9h3h+g9ms2xbHI/3aGCJxWRrwVzQvwuv\nmipRRI4CAqiqlghrZMaYv0xdvpPHJnhN+r13a0t6Nqnkc0TGnFm6yUVVi6c3jDEmPI4nJPKvKav4\n9I+NNK1akndubkn1skX8DsuYdAV1K7KIXAl0ch9/VtVvwxeSMQZg894jDBqzkGXb4ujXoSZP9mxo\n72kxOUYwLwt7FWgNjHZFD4pIB1V9KqyRGZOHfb9sB49PWIoIfHB7K7o3PsfvkIzJkGDOXC4Fmqtq\nEoCIjAAWAZZcjAmx4wmJvPLdn4yYvYlm1Urxzs0tqFbGqsFMzhPsE/qlgH2uu2SYYjEmT9u45zD3\njV3I8m0HGXBBLR7v0ZACkWF7ztmYsAomufwLWCQiM/HuFOuEnbUYE1LfLt3Ok18uI1+EMKx3FJc0\nquh3SMaclWDuFhsrIj/jXXcBeEJVd4Y1KmPyiGMnE3n5u5WMmrOZFtVL8fbNLaha2qrBTM4XzAX9\nGaraFa+Nr5RlxphM2rDnMINGL2TljoMM7FSbx7o3IH8+qwYzuUOayUVECgFFgHIiUhqvSgy81o3/\n9j4VY0zwJi3extNfLSN/ZATD+0bRpaFVg5nc5UxnLncBDwGVgQWcSi4HgXfCHJcxudKxk4m88M1K\nxs7bTKsapXn75hZULlXY77CMCbkzvSzsLeAtEblfVd/OwpiMyZXWxR5i0OiFrNoZz90X1mFwt/pW\nDWZyrWAu6L8tIucBjfBe+pVcPjKcgRmTm0xctI2nv15GwcgIPunXmosaVPA7JGPCKpgL+s8BnfGS\nyxSgJ/AbYMnFmHQcPZHI85NX8EX0FlrXLM3Qm1tQqaRVg5ncL5jnXK4DmgGLVLWfiFQERoU3LGNy\nvpjd8QwavYjVu+IZdFEdHr64PpFWDWbyiGCSy1FVTRKRBBEpAezm9HfaG2NS+HLBVp6ZuJwiBfIx\n4o42XFi/vN8hGZOlgkku0SJSChiGd9fYIWB2WKMyJoc6ciKBZyetYMKCrbStVYahN7egYolC6Y9o\nTC5zxuTi3j75L1U9ALwvIlOBEqq6NEuiMyYHWbMrnkGjFxITe4j7u9Tlwa71rBrM5FlnTC6qqiIy\nBWjiPm/MiqCMyWnGR2/h/yYtp1jBSD67oy0X1Cvnd0jG+CqYarGFItJaVeeHPRpjcpjDxxP4v0nL\n+WrhNtrXLstbNzWnglWDGUMw5+xtgdkisk5ElorIMhFJt1pMRAqJyDwRWSIiK0TkBVdeS0TmikiM\niHwhIgVceUH3Ocb1rxkwradc+WoR6Z65RTUmtFbtPMiV7/zG14u28WDXeowa0NYSizFOMGcumd2Z\nHwe6qOohEckP/CYi3wOPAG+q6uci8j7QH3jP/d+vqnVF5CbgNeBGEWkE3AQ0xmuK5kcRqa+qiZmM\ny5izoqqMi97Cs5NWULxQfkb3b8v5da0azJhA6Z65qOomvFuPu7juI0GOp6p6yH3M7/4U6AJMcOUj\ngKtcdy/3Gde/q7uhoBfwuaoeV9UNQAzQJohlMybkDh1P4OEvFvPEl8uIqlmaKQ9eYInFmFQE+4R+\nFNAA+AQvSYwCOgQxbj6825frAv8D1gEHVDXBDbKVUy0sVwG2AKhqgojEAWVd+ZyAyQaOEzivgcBA\ngOrVq6cXmjEZ9ueOgwwavZCNew/zyCX1GXRRXfJFSPojGpMHBVMtdjXQAlgIoKrbRaR4MBN3VVfN\n3XMyXwMNMxtoEPP6EPgQICoqSsM1H5M37Yg7yg0fzKZQ/nyMHtCO9nXK+h2SMdlaMMnlhLslWQFE\npGhGZ6KqB9xrktsDpUQk0p29VAW2ucG24VW/bRWRSKAksDegPFngOMaEnary9FfLOJmYxDf3XUDN\nchn+CRiT5wRzt9g4EfkALyncCfyI97T+GYlIeXfGgogUBi4B/gRm4rVXBtAHmOS6J7vPuP4/qaq6\n8pvc3WS1gHrAvGAWzphQ+GrhNmaujuXx7g0tsRgTpGCa3H9DRC7Be0lYA+BZVZ0exLQrASPcdZcI\nYJyqfisiK4HPReRlYBHwsRv+Y+AzEYkB9uHdIYaqrhCRccBKIAEYZHeKmayy++AxXvhmBVE1StP3\n/Jp+h2NMjiHeyUHuEhUVpdHR0X6HYXI4VWXgZwuYtSaW7x/sSO3yxfwOyZiwEpEFqhoVimmlWy0m\nIteIyFoRiRORgyISLyIHQzFzY7KzyUu2M33lLgZ3q2+JxZgMCuaC/r+BK1T1z3AHY0x2ERt/nOcm\nr6B5tVL0v6C23+EYk+MEc0F/lyUWk9c8O2k5R44n8vp1Te1ZFmMyIdj3uXwBTMRr0gUAVf0qbFEZ\n46Pvlu7g++U7eax7A+pVDOqRLmNMCsEklxJ4Tb50CyhTwJKLyXX2HjrOs5OW06RKSe7qZNVhxmRW\nMLci98uKQIzJDp7/ZiUHj51k9PVt7UVfxpyFYO4Wqy8iM0RkufvcVESeCX9oxmStaSt28s2S7dzf\npR4NzynhdzjG5GjBHJoNA54CTgK4VxzfFM6gjMlqB46c4B9fL6dRpRLc07mO3+EYk+MFk1yKqGrK\n5lYSUh3SmBzqxW9WcuDICV6/vin5rTrMmLMWzK9oj4jUwbuIj4hcB+wIa1TGZKGfVu3iq0XbuLdz\nHRpXLul3OMbkCsHcLTYIryn7hiKyDdgA3BbWqIzJInFHT/LUV8toULE493Wp53c4xuQawdwtth64\n2DW1H6Gq8eEPy5is8c/vVrLn0AmG9Y6iQKRVhxkTKsG8ifKRFJ8B4oAFqro4THEZE3Y/r97NuOit\n3NO5Dk2rlvI7HGNylWAO1aKAu/FeLVwFuAvoAQwTkcfDGJsxYRN/zKsOq1uhGA92teowY0ItmGsu\nVYGWqnoIQESeA74DOgEL8Bq2NCZHeWXKKnYdPMaEe86nUP58fodjTK4TzJlLBQLaFMN73qWiqh5N\nUW5MjvB7zB7GztvMgI61aVm9tN/hGJMrBXPmMhqYKyLJryO+AhjjLvCvDFtkxoTB4eMJPPHlUmqX\nK8ojl9T3Oxxjcq1g7hZ7SUS+Bzq4ortVNfk1j7eGLTJjwuC1qavYduAo4+9qb9VhxoRRMGcuuGRi\n7w02Odqc9XsZOXsT/TrUJKpmGb/DMSZXsxv7TZ5w9EQiT3y5lOplivBY9wZ+h2NMrhe25CIi1URk\npoisFJEVIvKgKy8jItNFZK37X9qVi4gMFZEYEVkqIi0DptXHDb9WRPqEK2aTe70+bTWb9h7htWub\nUqRAUCfsxpizEM4zlwRgsKo2AtoBg0SkEfAkMENV6wEz3GeAnkA99zcQeA+8ZAQ8B7QF2gDPJSck\nY4IRvXEfn/yxgdvb1aB9nbJ+h2NMnhC25KKqO1R1oeuOB/7EewizFzDCDTYCuMp19wJGqmcOUEpE\nKgHdgemquk9V9wPT8R7iNCZdx04m8viEpVQuWZgnezb0Oxxj8owsueYiIjWBFsBcvGdkkltV3glU\ndN1VgC0Bo23lVKsAqZWnnMdAEYkWkejY2NiQxm9yrjenr2H9nsO8dm1Tiha06jBjskrYk4uIFAO+\nBB5S1YOB/VRVcU35ny1V/VBVo1Q1qnz58qGYpMnhFm3ez7Bf13Nzm2pcUK+c3+EYk6eENbmISH68\nxDJaVb9yxbtcdRfu/25Xvg2oFjB6VVeWVrkxaTp2MpHHJiylYolCPHXpuX6HY0yeE867xQT4GPhT\nVf8b0GsykHzHVx9gUkB5b3fXWDsgzlWfTQO6iUhpdyG/myszYbYz7hh/xOzBO8HMWYbOWEvM7kP8\n65omlCiU3+9wjMlzwlkJ3QG4HVgmIslN8z8NvAqME5H+wCbgBtdvCnApEAMcAfoBqOo+EXkJmO+G\ne1FV94UxbgOs3H6QPp/MIzb+OO1rl+XZKxpxbqUSfocVlGVb4/hg1nqub1WVzg0q+B2OMXmS5MSj\n0vRERUVpdLQ1KJBZc9bv5c4R0RQrFMlt7Wrw0a/riTt6khtbV2dwt/qUK1bQ7xDTdCIhiSvf+Y39\nR07ww8MXUrKwnbUYEywRWaCqUaGYlt0+Y04zdfkOHvh8MdXLFGHkHW2oXKowt7WtwZAZa/hs9ia+\nXbKdB7rWo8/5NbPlmxvfmRnDqp3xfNwnyhKLMT7KfnsH45sxczdz7+iFNK5cgvF3tadyqcIAlCyS\nn+euaMzUhzoRVbM0/5zyJ93e/IXpK3dlq+sxK7bH8e7MGK5uUYWu51ZMfwRjTNhYcjGoKkNnrOXp\nr5dxYf3yjB7QltJFC/xtuLoVivFJvzZ82q81+SKEO0dGc/vH81i9M96HqE93MjGJx8YvpVSRAjx3\nRSO/wzEmz7PkksclJinPTV7Bf6ev4ZqWVfiwd1S6bW91blCBqQ914rkrGrFsWxw935rF/01czr7D\nJ7Io6r97/+d1rNxxkJevOo9SRf6eGI0xWcuSSx52PCGRBz5fxMjZmxjYqTZvXNeM/PmC2yTy54ug\nX4da/PxoZ25vV4Mx8zbT+fWZfPzbBk4mJoU58tOt3hnP0J/WcnnTSvQ475wsnbcxJnWWXPKo+GMn\n6ffJfL5buoOnL23I05eeS0SEZHg6pYsW4IVe5/H9gx1pVq0UL327ku5DZjFz1e70Rw6BhMQkHpuw\nhBKF8vPClY2zZJ7GmPRZcsmDYuOPc/OwOczdsI//XN+MgZ3qnPU061cszsg72vBxnyhUod+n8+kz\nfB4xu8N7PebDX9ezdGscL/RqTNlsfIu0MXmNJZc8ZvPeI1z//h/E7D7ER72juLZV1ZBNW0Toem5F\npj3UiWcuO5eFm/fTfcivPD95BQeOhP56TMzueIZMX0uPxudwWZNKIZ++MSbzLLnkISu2x3Ht+39w\n4OhJxtzZjosahufp9QKREQzoWJufH+3MTa2rMXL2Rjq/8TMj/thIQoiuxyQmKY9NWEqRgvl46arz\n8FobMsZkF5Zc8ojZ6/Zy0wdziIwQJtzdnpbVw/++tbLFCvLPq5vw3QMdaVSpBM9NXkHPt35l1pqz\nfyXC8N82sGjzAV64sjHli1t1mDHZjSWXPGDq8h30+WQeFUsW4st7zqduheJZOv9zK5Vg9IC2fHB7\nK04kJtF7+Dz6fzqfdbGHMjW99bGHeOOH1Vx8bkWubFY5xNEaY0LBkksuN3ruJu4dvZDzKpdgwt2n\nnrrPaiJC98bn8MPDnXiqZ0PmbthH9zdn8dK3K4k7ejLo6SQlKU98uZSCkRG8crVVhxmTXVlyyaVU\nlbd+XMs/vl7unrpvly0eLiwYmY+7LqzDzEc7c12rqgz/fQMXvfEzo+ZsCup6zIjZG5m/cT/PXtGY\nCiUKhT9gY0ymWHLJhZKfun/zx1NP3RcukM/vsE5TvnhBXr22Kd/cdwF1KxTjmYnLuWzob/wesyfN\ncTbtPcy/p66mc4PyXNvyb2+6NsZkI5ZccpnjCYk8MNZ76v6uTrX5z/XBP3Xvh/OqlOSLge1479aW\nHD6RwK0fzeXOkdFs3HP4tOGSq8MiI4R/XdPEqsOMyeay717HZNhfT90v8566f+rSc3PETlhE6Nmk\nEj8+ciGPdW/AHzF7uOTNX3hlyp8cPOZdjxk9bzNz1u/jH5edS6WS/lw3MsYEz14WlkvExh+n7yde\nC8X/vq4p17QM3cORWW33wWO8Pm01ExZupWzRAgzsVJu3flxLyxqlGXlHmxyRMI3JiUL5sjA7c8kF\nNu89wnXv/8G62EMM6xOVoxMLQIUShXj9+mZMHnQBNcsW5ZUpqwCsOsyYHMTeRJnDrdgeR5/h80lI\nSmLMne2y5OHIrNKkaknG392e6St3UbJwfqqWLuJ3SMaYIFlyycFmr9vLnSOjKVEoks8Hts/yhyOz\ngojQrbE1o29MTmPJJYf6ftkOHvx8MdXLnnrXvTHGZBdhu+YiIsNFZLeILA8oKyMi00Vkrftf2pWL\niAwVkRgRWSoiLQPG6eOGXysifcIVb04yas4m7h2zkPOq+PvUvTHGpCWcF/Q/BXqkKHsSmKGq9YAZ\n7jNAT6Ce+xsIvAdeMgKeA9oCbYDnkhNSXqSqDPlxDc9MXM5FDSpkm6fujTEmpbAlF1WdBexLUdwL\nGOG6RwBXBZSPVM8coJSIVAK6A9NVdZ+q7gem8/eElSckJinPTlrBkB/Xcm3Lqnxwe6ts99S9McYk\ny+prLhVVdYfr3glUdN1VgC0Bw211ZWmV/42IDMQ766F69eohDNl/xxMSefiLxUxZtpO7OtXmyZ4N\n7ZZcY0y25ttzLuo9vRmyJzhV9UNVjVLVqPLly4dqsr6LP3aSvsPnM2XZTv5x6bk55ql7Y0zeltXJ\nZZer7sL93+3KtwHVAoar6srSKs8TYuOPc9OHc5i/cR//vaEZd3aq7XdIxhgTlKxOLpOB5Du++gCT\nAsp7u7vG2gFxrvpsGtBNREq7C/ndXFmut2nvYa57/w/Wxx7OFU/dG2PylrBdcxGRsUBnoJyIbMW7\n6+tVYJyI9Ac2ATe4wacAlwIxwBGgH4Cq7hORl4D5brgXVTXlTQK5zh/r9vDA2MUkJCUx+s62ueqp\ne2NM3mANV2YjsfHHeWXKn3y9aBs1yhbh4z5RufKpe2NM9hTKhivtCf1sIClJGTt/M699v4qjJxO5\nv0tdBl1Ul0L57VZjY0zOZMnFZyu2x/GPr5ezeMsB2tcuy0tXnUfdCsX8DssYY86KJRefHDqewH9/\nWMOnf2ygTNECDLmxOb2aV7bbjI0xuYIllyymqny/fCcvfrOSXfHHuKVNdR7v3pCSRfL7HZoxxoSM\nJZcstHnvEZ6dvJyfV8fSqFIJ3rutJS3sTjBjTC5kySULHE9IZNis9bz9UwyREcL/Xd6IPu1rEJnP\nXgRqjMmdLLmE2R/r9vDMxOWsjz3MpU3O4dnLG3NOyUJ+h2WMMWFlySVMAp9ZqVamMJ/0a81FDSr4\nHZYxxmQJSy4hZs+sGGOMJZeQCnxmpV3tMrx8VRN7ZsUYkydZcgmBwGdWShcpwJs3NuOq5lXsmRVj\nTJ5lyeUs2DMrxhiTOksumWTPrBhjTNosuWSQPbNijDHps+SSAbPX7eWZictYZ8+sGGPMGVlyCcKe\nQ8d55bs/+cqeWTHGmKBYcjkDe2bFGGMyx5JLGlZsj+OZictZtNmeWTHGmIyy5JLCoeMJvDl9DZ/8\nbs+sGGNMZllyCbB06wEGjlxgz6wYY8xZyjHJRUR6AG8B+YCPVPXVUM+jepki1KtYzJ5ZMcaYs5Qj\nkouI5AP+B1wCbAXmi8hkVV0ZyvmUKlKAz/q3DeUkjTEmT8opT/61AWJUdb2qngA+B3r5HJMxxpg0\n5JTkUgXYEvB5qyv7i4gMFJFoEYmOjY3N0uCMMcacLqckl3Sp6oeqGqWqUeXLl/c7HGOMydNySnLZ\nBlQL+FzVlRljjMmGckpymQ/UE5FaIlIAuAmY7HNMxhhj0pAj7hZT1QQRuQ+Yhncr8nBVXeFzWMYY\nY9KQI5ILgKpOAab4HYcxxpj05ZRqMWOMMTmIqKrfMYSciMQCm85iEuWAPSEKJ6ezdXE6Wx+n2Lo4\nXW5YHzVUNSS32+bK5HK2RCRaVaP8jiM7sHVxOlsfp9i6OJ2tj9NZtZgxxpiQs+RijDEm5Cy5pO5D\nvwPIRmxdnM7Wxym2Lk5n6yOAXXMxxhgTcnbmYowxJuQsuRhjjAk5Sy4BRKSHiKwWkRgRedLvePwk\nItVEZKaIrBSRFSLyoN8x+U1E8onIIhH51u9Y/CYipURkgoisEpE/RaS93zH5SUQedr+T5SIyVkQK\n+R2T3yy5OAFvu+wJNAJuFpFG/kblqwRgsKo2AtoBg/L4+gB4EPjT7yCyibeAqaraEGhGHl4vIlIF\neACIUtXz8No/vMnfqPxnyeUUe9tlAFXdoaoLXXc83s6jypnHyr1EpCpwGfCR37H4TURKAp2AjwFU\n9YSqHvA3Kt9FAoVFJBIoAmz3OR7fWXI5Jd23XeZVIlITaAHM9TcSXw0BHgeS/A4kG6gFxAKfuGrC\nj0SkqN9B+UVVtwFvAJuBHUCcqv7gb1T+s+RizkhEigFfAg+p6kG/4/GDiFwO7FbVBX7Hkk1EAi2B\n91S1BXAYyLPXKEWkNF4tRy2gMlBURG7zNyr/WXI5xd52mYKI5MdLLKNV9Su/4/FRB+BKEdmIV13a\nRURG+RuSr7YCW1U1+Ux2Al6yyasuBjaoaqyqngS+As73OSbfWXI5xd52GUBEBK9O/U9V/a/f8fhJ\nVZ9S1aqqWhNvu/hJVfPskamq7gS2iEgDV9QVWOljSH7bDLQTkSLud9OVPHyDQ7Ic87KwcLO3Xf5N\nB+B2YJmILHZlT7uXthlzPzDaHYitB/r5HI9vVHWuiEwAFuLdZbkIawrGmn8xxhgTelYtZowxJuQs\nuRhjjAk5Sy7GGGNCzpKLMcaYkLPkYowxJuQsueQxIvK8iDzqul8UkYtDMM0pIlIqA8NfmdlWp11r\nvPdmZtysIiI1RWT5WYw/VkSWisjDmRy/oIj8KCKLReRG1zzLWTc6KiJ9ReSds51OOLl1f4vfcRh7\nziVPU9Vnz2Z898CYqOqlGZzvZDL/gGop4F7g3WBHEJFIVU3I5PyylIicA7RW1boZGCfl8rUAUNXm\n7vMXIQzRd+l8nzWBW4AxIZqeySQ7c8kDROQfIrJGRH4DGgSUfyoi17nuV927W5aKyBuurKKIfC0i\nS9zf+e7IcLWIjASWA9VEZKOIlHP9VrnprhGR0SJysYj8LiJrRaSNm+5fR8Bu2KEi8oeIrA+Ip5iI\nzBCRhSKyTESSW6h+FajjjspfF8/r7j0ay0TkRjd+ZxH5VUQmk8rT4yLSTURmu+mPd22o4Zbl325a\n80SkriuvKSI/ufUzQ0Sqp7WO3Czyicgw8d7x8YOIFHbDPxCwnj9P5ev6Aajilq+jiDQXkTlu+K/F\na8cKEflZRIaISDTeqwCSl6sCMApo7aZRxw0bJSI13PdQTkQi3Prp5sa7zS3vYhH5QLxXUCAi/dx3\nOQ/vwdrUtq/yIjLdLetHIrJJRMqlM91DIvJPt87miEjFgGl9KSLz3V8HV/68iHwmIr8Dn7nv41f3\n/S0MWO+vAh3d/B4WkUIi8on7PheJyEUB2+BkEfkJmJHacpmzpKr2l4v/gFbAMrxmwEsAMcCjrt+n\nwHVAWWA1px6qLeX+f4HXYCV4rRaUxDsyTALaBcxjI1DO9UsAmuAduCwAhgOC17DfRDd8X+CdgBjG\nu+Eb4b32ALyz6hKuu5yLW9w8lgfM+1pguouvIl5THJWAzngNKtZKZZ2UA2YBRd3nJ4BnA5blH667\nN/Ct6/4G6OO67whYlrTWUQLQ3JWPA25z3duBgoHrOUVsKZdvKXCh634RGOK6fwbeTeM775wcd8Cw\nUa57gFvfjwEfuLJz3fLld5/fdcteya3P8kAB4Pfk7y3F/N4BnnLdPQB16zjV6bpuBa5w3f8GnnHd\nY4ALXHd1vOaHAJ7H254Ku89FgEKuux4QncayD8ZrbQOgoVueQnjb4FagjN+/0dz6Z9ViuV9H4GtV\nPQLgjuRTigOOAR+L95bF5DctdsHbyaCqiUCcO3LepKpz0pjfBlVd5ua1Apihqioiy/B2nKmZqKpJ\nwMrkI1i8RPKKiHTCS2ZV8JJHShcAY118u0TkF6A1cBCYp6obUhmnHV4i+11EwNtxzg7oPzbg/5uu\nuz1wjev+DG+HCGmvow2qmtxszoKAZV+K12zKRGBiGuvDWwHee1NKqeovrmgEXmJIluHqLlX9SESu\nB+4GkqvNuuIdhMx366MwsBtoC/ysqrEuni+A+qlM9gLgajf9qSKyP53pApzg1Ha2ALjEdV8MNHLD\nA5RIPqsEJqvqUdedH3hHRJoDiWnElRzb2y62VSKyKWDY6aq6L43xzFmy5GJQr121Nng7g+uA+/B2\nmmk5fIZ+xwO6kwI+J5H29hY4TvJe5Va8I+ZWqnpSvBaJM/rq2LTiFLwdy81p9Nc0ujMicJkS8Xas\n4L1wrBNwBfAPEWmima/vP9P3kCoRKYLX4jdAMSAeb32MUNWnUgx7VSbj+msSqU3XOanudAJv/SRv\nGxF4Z8XHUsQCpy/vw8AuvLdgRuAdHGVUhtefCZ5dc8n9ZgFXiUhhESmOt1M7jTsyLKleo5QP4/1g\nwauLvscNk88dSWeVknjvUDnp6slruPJ4oHjAcL8CN7r4yuPtuOelM+05QIeA6ylFRSTwyPfGgP/J\nZzR/cOrVtbe6+UIG1pGIRADVVHUmXlVcSbwdfKpUNQ7YLyIdXdHtwC9pDR+k14DRwLPAsIBluM5d\nr0FEyohIDbyXw10oImXFe/3C9WlM83fgBjduN6B0OtM9kx/wGsXEjdM8jeFKAjvcGe/teFWSkPr2\ncaubVn28qrbV6cRgQsDOXHI5VV3oqjOW4FVJzE9lsOLAJBEphHe0+YgrfxD4UET64x1d3oP3pr2s\nMBr4xlWnRQOrAFR1r3g3CCwHvsd7O2R7vOVT4HFV3SkiDdOasKrGikhfYKyIFHTFzwBrXHdpEVmK\nd/aRfHZzP96bFx/DewtjcivAGVlH+YBRLgEJMFTTfz1wH+B9d8ZxVq0Pi8iFeFWGHVQ1UUSuFZF+\nqvqJiDwD/OAS4ElgkKrOEZHn8RLsAWBxGpN+AW9d3u6G3QnEq+qe1KYLbDpDmA8A/3PrPxLv4Oju\nVIZ7F/hSRHoDUzl1FrIUSBSRJXjX894F3nPbUQLQV1WPB1S7mTCxVpGNCeCq36JUdY/fseQULkEn\nuurV9nhvqEzrjMPkEWC2Y5EAAABDSURBVHbmYow5W9WBce7s5ARwp8/xmGzAzlyMMcaEnF3QN8YY\nE3KWXIwxxoScJRdjjDEhZ8nFGGNMyFlyMcYYE3L/D6VepKUZ6RSyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots( nrows=1, ncols=1 )\n",
    "ax.plot(g_losses)\n",
    "plt.ylabel(\"generator loss\")\n",
    "plt.xlabel(\"discriminator epochs for fixed generator\")\n",
    "plt.title(\"Generator loss agains various epoch with a fixed generator\")\n",
    "plt.show()\n",
    "fig.savefig('gen_loss_fixed_gen.png')   # save the figure to file\n",
    "plt.close(fig) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "d_losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "g_losses"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "g_losses = []\n",
    "d_losses = []\n",
    "for i in range(0, 100, 10):\n",
    "    gan3 = GAN3(disc_epoch=90, gen_epoch=i, disc_gan=2, gen_gan=1)\n",
    "    g_loss, d_loss = eval_gan(dataloader, gan3)\n",
    "    g_losses.append(g_loss)\n",
    "    d_losses.append(d_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAasAAAEWCAYAAADYRbjGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots( nrows=1, ncols=1 )\n",
    "ax.plot(d_losses)\n",
    "plt.ylabel(\"dicriminator loss\")\n",
    "plt.xlabel(\"generator epochs for fixed discriminator\")\n",
    "plt.title(\"Discriminator loss agains various epoch with a fixed discriminator\")\n",
    "plt.show()\n",
    "fig.savefig('disc_loss_fixed_disc.png')\n",
    "plt.close(fig) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAacAAAEWCAYAAADCeVhIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nuZzjyck551zO8eTknHMu5/x/AFV7DeMYfqgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots( nrows=1, ncols=1 )\n",
    "ax.plot(g_losses)\n",
    "plt.ylabel(\"generator loss\")\n",
    "plt.xlabel(\"generator epochs for fixed discriminator\")\n",
    "plt.title(\"Generator loss agains various epoch with a fixed discriminator\")\n",
    "plt.show()\n",
    "fig.savefig('gen_loss_fixed_disc.png')\n",
    "plt.close(fig) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}