Face_Generator.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Face Generator.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/GokulDas027/Generative-Adversarial-Networks-GANs-/blob/master/Face_Generator.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qu1EQayTTBAL",
        "colab_type": "text"
      },
      "source": [
        "# Face Generation\n",
        "\n",
        "Face Generation using DCGAN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o5ncQMTES52c",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch\n",
        "import numpy as np\n",
        "import pickle as pkl\n",
        "import matplotlib.pyplot as plt\n",
        "% matplotlib inline"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hB0KqjvrTAwV",
        "colab_type": "code",
        "outputId": "3025e401-c4c8-4415-b42b-1ef26da3c1b5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "train_on_gpu = torch.cuda.is_available()\n",
        "if train_on_gpu:\n",
        "  print(\"Training on GPU\")\n",
        "else:\n",
        "  print(\"Training NOT on GPU\")"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training on GPU\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3PvNowtUUGfc",
        "colab_type": "text"
      },
      "source": [
        "## Collecting DATA\n",
        "\n",
        "We are using [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train the adversarial networks. This dataset contains over 200,000 celebrity images in RGB Color channel. \n",
        "\n",
        "It can be downloaded from [here](https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip).\n",
        "\n",
        "\n",
        "### Preprocessing\n",
        "Since we are building a GAN, 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."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wCC8laeBTmZn",
        "colab_type": "code",
        "outputId": "46154576-c057-4273-fffe-b5c9766e3e1f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 237
        }
      },
      "source": [
        "!wget https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2019-05-08 13:11:56--  https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip\n",
            "Resolving s3.amazonaws.com (s3.amazonaws.com)... 52.216.10.181\n",
            "Connecting to s3.amazonaws.com (s3.amazonaws.com)|52.216.10.181|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 247694507 (236M) [application/zip]\n",
            "Saving to: ‘processed-celeba-small.zip’\n",
            "\n",
            "processed-celeba-sm 100%[===================>] 236.22M  35.1MB/s    in 7.3s    \n",
            "\n",
            "2019-05-08 13:12:04 (32.4 MB/s) - ‘processed-celeba-small.zip’ saved [247694507/247694507]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h6efHnwLXbnr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from zipfile import ZipFile\n",
        "\n",
        "with ZipFile('processed-celeba-small.zip','r') as dataset:\n",
        "  dataset.extractall()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d0zfsoqUvVwr",
        "colab_type": "text"
      },
      "source": [
        "### Data Loader\n",
        "\n",
        "Defining a DataLoader for loading images into the network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fhEU47MIX3n1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from torchvision import datasets\n",
        "from torchvision import transforms"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jzfc5H2MXfhD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_dataloader(batch_size, image_size, data_dir='processed_celeba_small/'):\n",
        "  \n",
        "  data_transforms = transforms.Compose([transforms.Resize(image_size),\n",
        "                                       transforms.ToTensor(),\n",
        "                                       transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]) \n",
        "  \n",
        "  data = datasets.ImageFolder(data_dir, transform=data_transforms)\n",
        "  \n",
        "  data_loader = torch.utils.data.DataLoader(data, batch_size=batch_size)\n",
        "  \n",
        "  return data_loader"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HYGrmr6lyAVe",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "batch_size = 128\n",
        "img_size = 32\n",
        "\n",
        "celeba_train_dataloader = get_dataloader(batch_size, img_size)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L3PCjfn99qEw",
        "colab_type": "text"
      },
      "source": [
        "## View Image\n",
        "\n",
        "Let's Visualise Some Images from the Train Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PCU-Vsau4NB_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def imshow(img):\n",
        "  npimg = img.numpy()\n",
        "  plt.imshow(np.transpose(npimg,(1,2,0)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zIOvbZboClWM",
        "colab_type": "code",
        "outputId": "4cfaeeaf-0491-44c4-c9fb-2400debe8ecd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 435
        }
      },
      "source": [
        "data_iter = iter(celeba_train_dataloader)\n",
        "images,_ = data_iter.next()\n",
        "\n",
        "plot_size = 10\n",
        "fig = plt.figure(figsize=(plot_size,4))\n",
        "\n",
        "for i in np.arange(plot_size):\n",
        "  ax = fig.add_subplot(2, plot_size/2, i+1, xticks=[], yticks=[])\n",
        "  imshow(images[i])"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkMAAADsCAYAAAB37KKJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGGv0MSSEEEKIseb/Bygl1jtpWy0UAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x288 with 10 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-gUApe6KGR35",
        "colab_type": "text"
      },
      "source": [
        "## Defining the Models\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aPnGeQS1HFO1",
        "colab_type": "text"
      },
      "source": [
        "### Discriminator Model  \n",
        "The Discriminator model is a Batch Normalised, Deep Covolutional Classifier without Pooling.\n",
        "\n",
        "* The inputs are 32x32x3 tensor images.\n",
        "* The output is a single value that will indicate whether a given image is real or fake."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ePCJ8LXvENRz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "\n",
        "def conv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
        "  layers=[]\n",
        "  \n",
        "  conv_layer = nn.Conv2d(in_channels, out_channels, kernel_size, \n",
        "                         stride, padding, bias=False)\n",
        "  \n",
        "  layers.append(conv_layer)\n",
        "\n",
        "  if batch_norm:\n",
        "    layers.append(nn.BatchNorm2d(out_channels))\n",
        "    \n",
        "  return nn.Sequential(*layers)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uxBPswWPH0-Y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Discriminator(nn.Module):\n",
        "  def __init__(self, conv_dim=32):\n",
        "    super(Discriminator,self).__init__()\n",
        "    self.conv_dim = conv_dim\n",
        "    \n",
        "    # 32x32 --> 16x16\n",
        "    self.conv1 = conv(3, conv_dim, 4, batch_norm=False)\n",
        "    # 16x16 --> 8x8\n",
        "    self.conv2 = conv(conv_dim, conv_dim*2, 4)\n",
        "    # 8x8 --> 4x4\n",
        "    self.conv3 = conv(conv_dim*2, conv_dim*4, 4)\n",
        "    # Fully connected layer : flattened img(depth*height*width)\n",
        "    self.fc = nn.Linear(conv_dim*4*4*4, 1)\n",
        "    \n",
        "  def forward(self, x):\n",
        "    # Passing to Convolutional Network\n",
        "    out = F.leaky_relu(self.conv1(x), 0.2)\n",
        "    out = F.leaky_relu(self.conv2(out), 0.2)\n",
        "    out = F.leaky_relu(self.conv3(out), 0.2)\n",
        "    # Flattening\n",
        "    out = out.view(-1, self.conv_dim*4*4*4)\n",
        "    # Passing to Fully Connected Layer\n",
        "    out = self.fc(out)   \n",
        "    return out\n",
        "    "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gb_dSY8mB25t",
        "colab_type": "text"
      },
      "source": [
        "### Generator Model\n",
        "It is a Batch Normalized DeConvolution Network to Upsample the Input and Generate new Images.\n",
        "*  The Generated Images have the same size as of our Training Images (32x32x3)\n",
        "* The inputs to the generator are vectors of some length z_size"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xwtJ817GEpUb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def deconv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
        "  layers=[]\n",
        "  \n",
        "  deconv_layer = nn.ConvTranspose2d(in_channels, out_channels, kernel_size, \n",
        "                         stride, padding, bias=False)\n",
        "  \n",
        "  layers.append(deconv_layer)\n",
        "\n",
        "  if batch_norm:\n",
        "    layers.append(nn.BatchNorm2d(out_channels))\n",
        "    \n",
        "  return nn.Sequential(*layers)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ea5yE100FI1u",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class Generator(nn.Module):\n",
        "  def __init__(self, z_size, conv_dim=32):\n",
        "    super(Generator, self).__init__()\n",
        "    self.conv_dim = conv_dim\n",
        "    \n",
        "    self.fc = nn.Linear(z_size,conv_dim*4*4*4)\n",
        "    # 4x4 --> 8x8\n",
        "    self.deconv1 = deconv(conv_dim*4, conv_dim*2, 4)\n",
        "    # 8x8 --> 16x16\n",
        "    self.deconv2 = deconv(conv_dim*2, conv_dim, 4)\n",
        "    # 16x16 --> 32x32\n",
        "    self.deconv3 = deconv(conv_dim, 3, 4, batch_norm=False)\n",
        "    \n",
        "  def forward(self, x):\n",
        "    # load x to fully connected layer\n",
        "    out = self.fc(x)\n",
        "    # transform linear to (batch_size,depth,height,width) dimension\n",
        "    out = out.view(-1,conv_dim*4,4,4) \n",
        "    # pass the transformed data to Deconv Network\n",
        "    out = F.relu(self.deconv1(out))\n",
        "    out = F.relu(self.deconv2(out))\n",
        "    out = F.tanh(self.deconv3(out))\n",
        "    return out"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nJ8HELl4KK00",
        "colab_type": "text"
      },
      "source": [
        "## Initialise Weights to the Network\n",
        "We should initialize the Weights of the convolutional and linear layers in the model to help it converge.\n",
        "\n",
        "From the original DCGAN paper, they say:\n",
        "> #### All weights were initialized from a zero-centered Normal distribution with standard deviation 0.02.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bz8TC101RaXd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from torch.nn import init\n",
        "\n",
        "def weights_init_normal(m):\n",
        "  classname = m.__class__.__name__\n",
        "  if hasattr(m, 'weight') and (classname.find('Conv') != -1 or classname.find('Linear') != -1):\n",
        "    init.normal_(m.weight.data, 0.0, 0.02)\n",
        "    if hasattr(m, 'bias') and m.bias is not None:\n",
        "      init.constant_(m.bias.data, 0.0)\n",
        "  elif classname.find('BatchNorm2d') != -1:  # BatchNorm Layer's weight is not a matrix; only normal distribution applies.\n",
        "    init.normal_(m.weight.data, 1.0, 0.02)\n",
        "    init.constant_(m.bias.data, 0.0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Iuk6Ox-0UHxl",
        "colab_type": "text"
      },
      "source": [
        "## Building the Complete network"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "41HeDAD_UBxm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def build_network(conv_dim, z_size):\n",
        "\n",
        "  D = Discriminator(conv_dim)\n",
        "  G = Generator(z_size=z_size, conv_dim=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"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0jZQ6xAtUqgL",
        "colab_type": "code",
        "outputId": "a3083843-047b-4bb1-f87c-03bd07337d6c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 563
        }
      },
      "source": [
        "conv_dim = 32\n",
        "z_size = 100\n",
        "\n",
        "D, G = build_network(conv_dim, z_size)\n",
        "\n",
        "if train_on_gpu:\n",
        "  D.cuda()\n",
        "  G.cuda()\n",
        "  print(\"Model moved to GPU\")\n",
        "else:\n",
        "  print(\"Model in CPU\")"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Discriminator(\n",
            "  (conv1): Sequential(\n",
            "    (0): Conv2d(3, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
            "  )\n",
            "  (conv2): Sequential(\n",
            "    (0): Conv2d(32, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
            "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  )\n",
            "  (conv3): Sequential(\n",
            "    (0): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
            "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "  )\n",
            "  (fc): Linear(in_features=2048, out_features=1, bias=True)\n",
            ")\n",
            "\n",
            "Generator(\n",
            "  (fc): Linear(in_features=100, out_features=2048, bias=True)\n",
            "  (deconv1): 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",
            "  (deconv2): 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",
            "  (deconv3): Sequential(\n",
            "    (0): ConvTranspose2d(32, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
            "  )\n",
            ")\n",
            "Model moved to GPU\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fUTg0m2Y81CC",
        "colab_type": "text"
      },
      "source": [
        "## Discriminator and Generator Loss"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7XVr6mqW8_A6",
        "colab_type": "text"
      },
      "source": [
        "### Discriminator Loss\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",
        "* The Discriminator outputs 1 for real images and 0 for fake images, so we need to set up the losses to reflect that\n",
        "\n",
        "### Generator Loss\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."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yPEtxVqPX0Yx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def real_loss(D_out, smooth=True):\n",
        "  batch_size = D_out.size(0)\n",
        "  if smooth:\n",
        "    # smooth, real labels = 0.9\n",
        "    labels = torch.ones(batch_size)*0.9\n",
        "  else:\n",
        "    # real labels = 1\n",
        "    labels = torch.ones(batch_size)\n",
        "  \n",
        "  if train_on_gpu:\n",
        "    labels = labels.cuda()\n",
        "  \n",
        "  # Binary Crossentropy with LogitsLoss\n",
        "  criterion = nn.BCEWithLogitsLoss()\n",
        "  loss = criterion(D_out.squeeze(), labels)    \n",
        "  return loss\n",
        "\n",
        "def fake_loss(D_out):\n",
        "  batch_size = D_out.size(0)\n",
        "  labels = torch.zeros(batch_size)\n",
        "  \n",
        "  if train_on_gpu:\n",
        "    labels = labels.cuda()\n",
        "  \n",
        "  # Binary Crossentropy with LogitsLoss\n",
        "  criterion = nn.BCEWithLogitsLoss()\n",
        "  loss = criterion(D_out.squeeze(), labels)    \n",
        "  return loss"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tekd9KQxG016",
        "colab_type": "text"
      },
      "source": [
        "## Optimizers\n",
        "Define Optimizers for Generator and Discriminator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ClsalEQzGvXK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import torch.optim as optim\n",
        "\n",
        "# hyperparameters\n",
        "lr = 0.0001\n",
        "beta1 = 0.5\n",
        "beta2 = 0.999\n",
        "\n",
        "# optimizers\n",
        "d_optimizer = optim.Adam(D.parameters(), lr, betas=(beta1,beta2))\n",
        "g_optimizer = optim.Adam(G.parameters(), lr, betas=(beta1,beta2))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6EFN4FnSJ1fR",
        "colab_type": "text"
      },
      "source": [
        "## Training\n",
        "\n",
        "Training involves alternating between training Discriminator and Generator.\n",
        "* Train Discriminator by alternating on Real and Fake images.\n",
        "* Then the Generator, which tries to fool the Discriminator have Opposing loss function ie: by computing the Discriminator loss on fake images using Flipped labels.!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BQg9Dq_aJjUJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def train(D,G, num_epochs, print_interval=50):\n",
        "  \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",
        "  \n",
        "  if train_on_gpu:\n",
        "    fixed_z=fixed_z.cuda()\n",
        "    \n",
        "  for epoch in range(num_epochs):\n",
        "    # batch loop\n",
        "    for batch_i,(real_images,_) in enumerate(celeba_train_dataloader):\n",
        "      \n",
        "      batch_size = real_images.size(0)\n",
        "      \n",
        "      #################### Discriminator Training ####################\n",
        "      d_optimizer.zero_grad()\n",
        "      \n",
        "      if train_on_gpu:\n",
        "        real_images = real_images.cuda()\n",
        "      \n",
        "      # calculating the Discriminator loss on real images\n",
        "      D_real = D(real_images)\n",
        "      D_real_loss = real_loss(D_real)\n",
        "      \n",
        "      # Generating fake image\n",
        "      z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
        "      z = torch.from_numpy(z).float()\n",
        "      if train_on_gpu:\n",
        "        z = z.cuda()\n",
        "        fake_images = G(z)\n",
        "      \n",
        "      # calculating the Discriminator loss on fake images\n",
        "      D_fake = D(fake_images)\n",
        "      D_fake_loss = fake_loss(D_fake)\n",
        "      \n",
        "      # computing total loss\n",
        "      D_loss = D_real_loss + D_fake_loss\n",
        "      \n",
        "      D_loss.backward()\n",
        "      d_optimizer.step()\n",
        "      \n",
        "      #################### Generator Training ####################\n",
        "      for i in range(2):\n",
        "        g_optimizer.zero_grad()\n",
        "\n",
        "        # Generating Fake image\n",
        "        z = np.random.uniform(-1,1,size=(batch_size, z_size))\n",
        "        z = torch.from_numpy(z).float()\n",
        "        if train_on_gpu:\n",
        "          z = z.cuda()\n",
        "        fake_images = G(z)\n",
        "\n",
        "        # calculating the Discriminator loss on fake images \n",
        "        # on flipped labels\n",
        "        D_fake = D(fake_images)\n",
        "        G_loss = real_loss(D_fake)\n",
        "\n",
        "        G_loss.backward()\n",
        "        g_optimizer.step()\n",
        "      \n",
        "      # Print some loss stats\n",
        "      if batch_i % print_interval == 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, num_epochs, D_loss.item(), G_loss.item()))\n",
        "          \n",
        "      # After Each Epoch Generate Images\n",
        "      G.eval() # for Generating Images\n",
        "      sample_z = G(fixed_z)\n",
        "      samples.append(sample_z)\n",
        "      G.train()\n",
        "      \n",
        "  # Save training generator samples\n",
        "  with open('train_samples.pkl', 'wb') as f:\n",
        "      pkl.dump(samples, f)\n",
        "      \n",
        "  return losses"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iV0OMnm5VEgG",
        "colab_type": "code",
        "outputId": "7e8e48d2-d53c-4907-f74b-11d907b1fae0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 5508
        }
      },
      "source": [
        "for i in range(4):\n",
        "  num_epochs = 5\n",
        "\n",
        "  losses = train(D,G,num_epochs)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/functional.py:1374: UserWarning: nn.functional.tanh is deprecated. Use torch.tanh instead.\n",
            "  warnings.warn(\"nn.functional.tanh is deprecated. Use torch.tanh instead.\")\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Epoch [    1/    5] | d_loss: 1.5147 | g_loss: 0.6498\n",
            "Epoch [    1/    5] | d_loss: 0.7037 | g_loss: 1.5957\n",
            "Epoch [    1/    5] | d_loss: 0.8483 | g_loss: 1.4117\n",
            "Epoch [    1/    5] | d_loss: 0.8429 | g_loss: 1.6714\n",
            "Epoch [    1/    5] | d_loss: 0.9675 | g_loss: 1.1393\n",
            "Epoch [    1/    5] | d_loss: 1.0689 | g_loss: 1.0600\n",
            "Epoch [    1/    5] | d_loss: 1.0899 | g_loss: 1.0619\n",
            "Epoch [    1/    5] | d_loss: 1.1517 | g_loss: 1.0746\n",
            "Epoch [    1/    5] | d_loss: 1.1444 | g_loss: 0.9668\n",
            "Epoch [    1/    5] | d_loss: 1.3426 | g_loss: 1.0676\n",
            "Epoch [    1/    5] | d_loss: 1.4124 | g_loss: 0.9565\n",
            "Epoch [    1/    5] | d_loss: 1.3138 | g_loss: 0.9822\n",
            "Epoch [    1/    5] | d_loss: 1.2568 | g_loss: 0.9971\n",
            "Epoch [    1/    5] | d_loss: 1.2868 | g_loss: 0.9648\n",
            "Epoch [    1/    5] | d_loss: 1.3361 | g_loss: 0.8703\n",
            "Epoch [    2/    5] | d_loss: 1.3884 | g_loss: 0.7731\n",
            "Epoch [    2/    5] | d_loss: 1.3578 | g_loss: 0.8597\n",
            "Epoch [    2/    5] | d_loss: 1.3655 | g_loss: 0.8372\n",
            "Epoch [    2/    5] | d_loss: 1.3633 | g_loss: 0.7764\n",
            "Epoch [    2/    5] | d_loss: 1.4712 | g_loss: 0.7782\n",
            "Epoch [    2/    5] | d_loss: 1.3411 | g_loss: 0.8406\n",
            "Epoch [    2/    5] | d_loss: 1.3578 | g_loss: 0.8489\n",
            "Epoch [    2/    5] | d_loss: 1.3651 | g_loss: 0.8954\n",
            "Epoch [    2/    5] | d_loss: 1.4248 | g_loss: 0.7915\n",
            "Epoch [    2/    5] | d_loss: 1.3362 | g_loss: 0.8501\n",
            "Epoch [    2/    5] | d_loss: 1.2993 | g_loss: 0.8365\n",
            "Epoch [    2/    5] | d_loss: 1.3919 | g_loss: 0.7857\n",
            "Epoch [    2/    5] | d_loss: 1.4014 | g_loss: 0.7789\n",
            "Epoch [    2/    5] | d_loss: 1.3062 | g_loss: 0.8951\n",
            "Epoch [    2/    5] | d_loss: 1.3736 | g_loss: 0.8277\n",
            "Epoch [    3/    5] | d_loss: 1.3581 | g_loss: 0.8128\n",
            "Epoch [    3/    5] | d_loss: 1.3742 | g_loss: 0.7888\n",
            "Epoch [    3/    5] | d_loss: 1.2978 | g_loss: 1.1223\n",
            "Epoch [    3/    5] | d_loss: 1.4631 | g_loss: 0.7489\n",
            "Epoch [    3/    5] | d_loss: 1.4946 | g_loss: 0.7259\n",
            "Epoch [    3/    5] | d_loss: 1.3856 | g_loss: 0.8803\n",
            "Epoch [    3/    5] | d_loss: 1.3086 | g_loss: 0.9036\n",
            "Epoch [    3/    5] | d_loss: 1.3415 | g_loss: 0.8785\n",
            "Epoch [    3/    5] | d_loss: 1.4296 | g_loss: 0.7942\n",
            "Epoch [    3/    5] | d_loss: 1.3759 | g_loss: 0.8537\n",
            "Epoch [    3/    5] | d_loss: 1.2500 | g_loss: 0.7856\n",
            "Epoch [    3/    5] | d_loss: 1.3603 | g_loss: 0.7711\n",
            "Epoch [    3/    5] | d_loss: 1.3804 | g_loss: 0.7365\n",
            "Epoch [    3/    5] | d_loss: 1.3894 | g_loss: 0.8398\n",
            "Epoch [    3/    5] | d_loss: 1.4060 | g_loss: 0.7954\n",
            "Epoch [    4/    5] | d_loss: 1.4426 | g_loss: 0.8358\n",
            "Epoch [    4/    5] | d_loss: 1.3510 | g_loss: 0.7262\n",
            "Epoch [    4/    5] | d_loss: 1.4112 | g_loss: 0.8777\n",
            "Epoch [    4/    5] | d_loss: 1.4005 | g_loss: 0.8393\n",
            "Epoch [    4/    5] | d_loss: 1.4417 | g_loss: 0.7134\n",
            "Epoch [    4/    5] | d_loss: 1.3410 | g_loss: 0.7969\n",
            "Epoch [    4/    5] | d_loss: 1.4378 | g_loss: 0.8567\n",
            "Epoch [    4/    5] | d_loss: 1.3838 | g_loss: 0.8090\n",
            "Epoch [    4/    5] | d_loss: 1.4765 | g_loss: 0.7779\n",
            "Epoch [    4/    5] | d_loss: 1.4157 | g_loss: 0.8333\n",
            "Epoch [    4/    5] | d_loss: 1.3017 | g_loss: 0.8421\n",
            "Epoch [    4/    5] | d_loss: 1.3776 | g_loss: 0.8849\n",
            "Epoch [    4/    5] | d_loss: 1.3505 | g_loss: 0.8102\n",
            "Epoch [    4/    5] | d_loss: 1.3464 | g_loss: 0.8225\n",
            "Epoch [    4/    5] | d_loss: 1.3458 | g_loss: 0.7838\n",
            "Epoch [    5/    5] | d_loss: 1.3906 | g_loss: 0.7850\n",
            "Epoch [    5/    5] | d_loss: 1.4138 | g_loss: 0.7853\n",
            "Epoch [    5/    5] | d_loss: 1.3286 | g_loss: 0.8476\n",
            "Epoch [    5/    5] | d_loss: 1.3719 | g_loss: 0.8647\n",
            "Epoch [    5/    5] | d_loss: 1.4369 | g_loss: 0.6971\n",
            "Epoch [    5/    5] | d_loss: 1.3686 | g_loss: 0.8020\n",
            "Epoch [    5/    5] | d_loss: 1.3735 | g_loss: 0.8189\n",
            "Epoch [    5/    5] | d_loss: 1.3674 | g_loss: 0.8233\n",
            "Epoch [    5/    5] | d_loss: 1.3695 | g_loss: 0.8156\n",
            "Epoch [    5/    5] | d_loss: 1.3402 | g_loss: 0.8052\n",
            "Epoch [    5/    5] | d_loss: 1.3477 | g_loss: 0.8307\n",
            "Epoch [    5/    5] | d_loss: 1.3791 | g_loss: 0.7939\n",
            "Epoch [    5/    5] | d_loss: 1.4304 | g_loss: 0.8296\n",
            "Epoch [    5/    5] | d_loss: 1.4139 | g_loss: 0.8084\n",
            "Epoch [    5/    5] | d_loss: 1.3661 | g_loss: 0.7465\n",
            "Epoch [    1/    5] | d_loss: 1.4390 | g_loss: 0.7868\n",
            "Epoch [    1/    5] | d_loss: 1.3912 | g_loss: 0.8165\n",
            "Epoch [    1/    5] | d_loss: 1.4003 | g_loss: 0.8214\n",
            "Epoch [    1/    5] | d_loss: 1.3833 | g_loss: 0.7966\n",
            "Epoch [    1/    5] | d_loss: 1.4232 | g_loss: 0.7545\n",
            "Epoch [    1/    5] | d_loss: 1.3441 | g_loss: 0.7682\n",
            "Epoch [    1/    5] | d_loss: 1.3795 | g_loss: 0.7914\n",
            "Epoch [    1/    5] | d_loss: 1.3697 | g_loss: 0.7842\n",
            "Epoch [    1/    5] | d_loss: 1.3966 | g_loss: 0.7864\n",
            "Epoch [    1/    5] | d_loss: 1.3669 | g_loss: 0.7760\n",
            "Epoch [    1/    5] | d_loss: 1.3579 | g_loss: 0.7821\n",
            "Epoch [    1/    5] | d_loss: 1.4043 | g_loss: 0.8169\n",
            "Epoch [    1/    5] | d_loss: 1.4089 | g_loss: 0.7919\n",
            "Epoch [    1/    5] | d_loss: 1.3651 | g_loss: 0.8231\n",
            "Epoch [    1/    5] | d_loss: 1.3554 | g_loss: 0.7812\n",
            "Epoch [    2/    5] | d_loss: 1.3737 | g_loss: 0.8257\n",
            "Epoch [    2/    5] | d_loss: 1.3663 | g_loss: 0.8395\n",
            "Epoch [    2/    5] | d_loss: 1.3597 | g_loss: 0.8401\n",
            "Epoch [    2/    5] | d_loss: 1.3527 | g_loss: 0.8276\n",
            "Epoch [    2/    5] | d_loss: 1.4286 | g_loss: 0.7787\n",
            "Epoch [    2/    5] | d_loss: 1.3836 | g_loss: 0.8233\n",
            "Epoch [    2/    5] | d_loss: 1.3648 | g_loss: 0.8351\n",
            "Epoch [    2/    5] | d_loss: 1.3834 | g_loss: 0.8459\n",
            "Epoch [    2/    5] | d_loss: 1.3851 | g_loss: 0.7723\n",
            "Epoch [    2/    5] | d_loss: 1.3742 | g_loss: 0.8144\n",
            "Epoch [    2/    5] | d_loss: 1.3419 | g_loss: 0.7685\n",
            "Epoch [    2/    5] | d_loss: 1.3791 | g_loss: 0.8372\n",
            "Epoch [    2/    5] | d_loss: 1.3760 | g_loss: 0.8188\n",
            "Epoch [    2/    5] | d_loss: 1.3690 | g_loss: 0.8367\n",
            "Epoch [    2/    5] | d_loss: 1.3900 | g_loss: 0.7932\n",
            "Epoch [    3/    5] | d_loss: 1.4062 | g_loss: 0.8150\n",
            "Epoch [    3/    5] | d_loss: 1.3680 | g_loss: 0.8411\n",
            "Epoch [    3/    5] | d_loss: 1.3712 | g_loss: 0.8360\n",
            "Epoch [    3/    5] | d_loss: 1.3613 | g_loss: 0.8186\n",
            "Epoch [    3/    5] | d_loss: 1.3781 | g_loss: 0.7794\n",
            "Epoch [    3/    5] | d_loss: 1.3622 | g_loss: 0.8004\n",
            "Epoch [    3/    5] | d_loss: 1.3541 | g_loss: 0.8438\n",
            "Epoch [    3/    5] | d_loss: 1.3548 | g_loss: 0.8589\n",
            "Epoch [    3/    5] | d_loss: 1.3548 | g_loss: 0.7674\n",
            "Epoch [    3/    5] | d_loss: 1.3431 | g_loss: 0.8232\n",
            "Epoch [    3/    5] | d_loss: 1.3342 | g_loss: 0.8173\n",
            "Epoch [    3/    5] | d_loss: 1.3673 | g_loss: 0.8939\n",
            "Epoch [    3/    5] | d_loss: 1.3472 | g_loss: 0.8399\n",
            "Epoch [    3/    5] | d_loss: 1.3284 | g_loss: 0.8497\n",
            "Epoch [    3/    5] | d_loss: 1.3459 | g_loss: 0.8003\n",
            "Epoch [    4/    5] | d_loss: 1.3733 | g_loss: 0.8059\n",
            "Epoch [    4/    5] | d_loss: 1.3445 | g_loss: 0.8686\n",
            "Epoch [    4/    5] | d_loss: 1.3393 | g_loss: 0.8516\n",
            "Epoch [    4/    5] | d_loss: 1.3559 | g_loss: 0.8226\n",
            "Epoch [    4/    5] | d_loss: 1.3867 | g_loss: 0.8279\n",
            "Epoch [    4/    5] | d_loss: 1.3375 | g_loss: 0.8510\n",
            "Epoch [    4/    5] | d_loss: 1.3485 | g_loss: 0.8317\n",
            "Epoch [    4/    5] | d_loss: 1.3666 | g_loss: 0.8647\n",
            "Epoch [    4/    5] | d_loss: 1.3279 | g_loss: 0.7349\n",
            "Epoch [    4/    5] | d_loss: 1.3330 | g_loss: 0.8276\n",
            "Epoch [    4/    5] | d_loss: 1.3138 | g_loss: 0.7871\n",
            "Epoch [    4/    5] | d_loss: 1.3629 | g_loss: 0.8560\n",
            "Epoch [    4/    5] | d_loss: 1.3572 | g_loss: 0.8898\n",
            "Epoch [    4/    5] | d_loss: 1.3414 | g_loss: 0.8705\n",
            "Epoch [    4/    5] | d_loss: 1.3359 | g_loss: 0.8185\n",
            "Epoch [    5/    5] | d_loss: 1.3229 | g_loss: 0.8030\n",
            "Epoch [    5/    5] | d_loss: 1.3327 | g_loss: 0.8668\n",
            "Epoch [    5/    5] | d_loss: 1.3748 | g_loss: 0.8639\n",
            "Epoch [    5/    5] | d_loss: 1.3183 | g_loss: 0.7856\n",
            "Epoch [    5/    5] | d_loss: 1.3731 | g_loss: 0.8008\n",
            "Epoch [    5/    5] | d_loss: 1.3374 | g_loss: 0.8604\n",
            "Epoch [    5/    5] | d_loss: 1.3723 | g_loss: 0.7998\n",
            "Epoch [    5/    5] | d_loss: 1.3454 | g_loss: 0.8773\n",
            "Epoch [    5/    5] | d_loss: 1.3135 | g_loss: 0.7840\n",
            "Epoch [    5/    5] | d_loss: 1.3480 | g_loss: 0.8831\n",
            "Epoch [    5/    5] | d_loss: 1.3097 | g_loss: 0.7947\n",
            "Epoch [    5/    5] | d_loss: 1.3117 | g_loss: 0.8709\n",
            "Epoch [    5/    5] | d_loss: 1.2663 | g_loss: 0.9097\n",
            "Epoch [    5/    5] | d_loss: 1.3265 | g_loss: 0.7936\n",
            "Epoch [    5/    5] | d_loss: 1.3333 | g_loss: 0.8444\n",
            "Epoch [    1/    5] | d_loss: 1.3480 | g_loss: 0.8463\n",
            "Epoch [    1/    5] | d_loss: 1.3180 | g_loss: 0.8171\n",
            "Epoch [    1/    5] | d_loss: 1.3271 | g_loss: 0.8415\n",
            "Epoch [    1/    5] | d_loss: 1.3221 | g_loss: 0.8098\n",
            "Epoch [    1/    5] | d_loss: 1.4108 | g_loss: 0.8064\n",
            "Epoch [    1/    5] | d_loss: 1.3110 | g_loss: 0.8202\n",
            "Epoch [    1/    5] | d_loss: 1.3082 | g_loss: 0.8766\n",
            "Epoch [    1/    5] | d_loss: 1.3027 | g_loss: 0.8626\n",
            "Epoch [    1/    5] | d_loss: 1.3341 | g_loss: 0.8135\n",
            "Epoch [    1/    5] | d_loss: 1.3548 | g_loss: 0.8585\n",
            "Epoch [    1/    5] | d_loss: 1.3058 | g_loss: 0.8248\n",
            "Epoch [    1/    5] | d_loss: 1.3392 | g_loss: 0.8456\n",
            "Epoch [    1/    5] | d_loss: 1.3389 | g_loss: 0.9206\n",
            "Epoch [    1/    5] | d_loss: 1.3215 | g_loss: 0.8136\n",
            "Epoch [    1/    5] | d_loss: 1.3022 | g_loss: 0.8838\n",
            "Epoch [    2/    5] | d_loss: 1.2865 | g_loss: 0.8523\n",
            "Epoch [    2/    5] | d_loss: 1.2997 | g_loss: 0.8500\n",
            "Epoch [    2/    5] | d_loss: 1.2654 | g_loss: 0.8850\n",
            "Epoch [    2/    5] | d_loss: 1.3128 | g_loss: 0.8249\n",
            "Epoch [    2/    5] | d_loss: 1.3940 | g_loss: 0.8291\n",
            "Epoch [    2/    5] | d_loss: 1.3519 | g_loss: 0.8927\n",
            "Epoch [    2/    5] | d_loss: 1.2959 | g_loss: 0.9137\n",
            "Epoch [    2/    5] | d_loss: 1.2626 | g_loss: 0.9123\n",
            "Epoch [    2/    5] | d_loss: 1.3306 | g_loss: 0.8340\n",
            "Epoch [    2/    5] | d_loss: 1.2626 | g_loss: 0.8724\n",
            "Epoch [    2/    5] | d_loss: 1.3047 | g_loss: 0.9198\n",
            "Epoch [    2/    5] | d_loss: 1.2642 | g_loss: 0.9290\n",
            "Epoch [    2/    5] | d_loss: 1.3008 | g_loss: 0.9245\n",
            "Epoch [    2/    5] | d_loss: 1.2705 | g_loss: 0.8678\n",
            "Epoch [    2/    5] | d_loss: 1.3000 | g_loss: 0.8949\n",
            "Epoch [    3/    5] | d_loss: 1.3003 | g_loss: 0.8636\n",
            "Epoch [    3/    5] | d_loss: 1.2692 | g_loss: 0.9102\n",
            "Epoch [    3/    5] | d_loss: 1.2625 | g_loss: 0.9684\n",
            "Epoch [    3/    5] | d_loss: 1.2736 | g_loss: 0.8718\n",
            "Epoch [    3/    5] | d_loss: 1.3302 | g_loss: 0.7668\n",
            "Epoch [    3/    5] | d_loss: 1.2735 | g_loss: 0.8879\n",
            "Epoch [    3/    5] | d_loss: 1.2723 | g_loss: 0.9569\n",
            "Epoch [    3/    5] | d_loss: 1.2446 | g_loss: 0.9217\n",
            "Epoch [    3/    5] | d_loss: 1.3575 | g_loss: 0.9697\n",
            "Epoch [    3/    5] | d_loss: 1.2716 | g_loss: 0.9337\n",
            "Epoch [    3/    5] | d_loss: 1.2266 | g_loss: 0.8856\n",
            "Epoch [    3/    5] | d_loss: 1.2046 | g_loss: 0.9203\n",
            "Epoch [    3/    5] | d_loss: 1.2296 | g_loss: 0.9487\n",
            "Epoch [    3/    5] | d_loss: 1.2093 | g_loss: 0.7665\n",
            "Epoch [    3/    5] | d_loss: 1.2688 | g_loss: 0.9763\n",
            "Epoch [    4/    5] | d_loss: 1.2800 | g_loss: 0.9879\n",
            "Epoch [    4/    5] | d_loss: 1.2236 | g_loss: 0.9594\n",
            "Epoch [    4/    5] | d_loss: 1.3056 | g_loss: 0.8818\n",
            "Epoch [    4/    5] | d_loss: 1.2968 | g_loss: 0.9805\n",
            "Epoch [    4/    5] | d_loss: 1.2874 | g_loss: 0.8108\n",
            "Epoch [    4/    5] | d_loss: 1.2471 | g_loss: 1.0957\n",
            "Epoch [    4/    5] | d_loss: 1.2078 | g_loss: 0.9621\n",
            "Epoch [    4/    5] | d_loss: 1.1943 | g_loss: 0.9726\n",
            "Epoch [    4/    5] | d_loss: 1.2433 | g_loss: 0.8155\n",
            "Epoch [    4/    5] | d_loss: 1.3307 | g_loss: 1.0887\n",
            "Epoch [    4/    5] | d_loss: 1.2195 | g_loss: 0.8077\n",
            "Epoch [    4/    5] | d_loss: 1.2298 | g_loss: 0.9929\n",
            "Epoch [    4/    5] | d_loss: 1.2381 | g_loss: 1.0350\n",
            "Epoch [    4/    5] | d_loss: 1.2039 | g_loss: 0.8392\n",
            "Epoch [    4/    5] | d_loss: 1.2149 | g_loss: 1.1351\n",
            "Epoch [    5/    5] | d_loss: 1.2456 | g_loss: 0.9961\n",
            "Epoch [    5/    5] | d_loss: 1.2322 | g_loss: 1.0136\n",
            "Epoch [    5/    5] | d_loss: 1.1651 | g_loss: 1.0210\n",
            "Epoch [    5/    5] | d_loss: 1.1757 | g_loss: 0.9780\n",
            "Epoch [    5/    5] | d_loss: 1.2008 | g_loss: 0.8929\n",
            "Epoch [    5/    5] | d_loss: 1.1686 | g_loss: 0.9279\n",
            "Epoch [    5/    5] | d_loss: 1.1784 | g_loss: 0.9949\n",
            "Epoch [    5/    5] | d_loss: 1.1476 | g_loss: 0.9815\n",
            "Epoch [    5/    5] | d_loss: 1.2497 | g_loss: 1.1323\n",
            "Epoch [    5/    5] | d_loss: 1.2199 | g_loss: 1.0053\n",
            "Epoch [    5/    5] | d_loss: 1.1935 | g_loss: 0.8833\n",
            "Epoch [    5/    5] | d_loss: 1.2932 | g_loss: 0.9103\n",
            "Epoch [    5/    5] | d_loss: 1.1451 | g_loss: 0.9965\n",
            "Epoch [    5/    5] | d_loss: 1.1761 | g_loss: 0.9769\n",
            "Epoch [    5/    5] | d_loss: 1.2178 | g_loss: 0.9832\n",
            "Epoch [    1/    5] | d_loss: 1.2558 | g_loss: 0.8585\n",
            "Epoch [    1/    5] | d_loss: 1.1781 | g_loss: 1.0425\n",
            "Epoch [    1/    5] | d_loss: 1.1846 | g_loss: 1.1579\n",
            "Epoch [    1/    5] | d_loss: 1.1468 | g_loss: 1.0110\n",
            "Epoch [    1/    5] | d_loss: 1.2036 | g_loss: 0.8556\n",
            "Epoch [    1/    5] | d_loss: 1.1756 | g_loss: 1.0616\n",
            "Epoch [    1/    5] | d_loss: 1.1264 | g_loss: 1.0470\n",
            "Epoch [    1/    5] | d_loss: 1.0645 | g_loss: 1.0787\n",
            "Epoch [    1/    5] | d_loss: 1.4933 | g_loss: 1.6557\n",
            "Epoch [    1/    5] | d_loss: 1.1808 | g_loss: 1.0454\n",
            "Epoch [    1/    5] | d_loss: 1.1820 | g_loss: 1.0700\n",
            "Epoch [    1/    5] | d_loss: 1.1267 | g_loss: 1.0927\n",
            "Epoch [    1/    5] | d_loss: 1.1570 | g_loss: 0.9211\n",
            "Epoch [    1/    5] | d_loss: 1.1067 | g_loss: 0.9394\n",
            "Epoch [    1/    5] | d_loss: 1.1240 | g_loss: 1.1577\n",
            "Epoch [    2/    5] | d_loss: 1.1891 | g_loss: 1.1706\n",
            "Epoch [    2/    5] | d_loss: 1.1373 | g_loss: 1.0154\n",
            "Epoch [    2/    5] | d_loss: 1.1038 | g_loss: 1.1985\n",
            "Epoch [    2/    5] | d_loss: 1.0626 | g_loss: 0.9856\n",
            "Epoch [    2/    5] | d_loss: 1.1826 | g_loss: 0.8139\n",
            "Epoch [    2/    5] | d_loss: 1.0569 | g_loss: 1.2713\n",
            "Epoch [    2/    5] | d_loss: 1.1324 | g_loss: 0.8904\n",
            "Epoch [    2/    5] | d_loss: 1.0819 | g_loss: 1.0412\n",
            "Epoch [    2/    5] | d_loss: 1.1804 | g_loss: 1.1035\n",
            "Epoch [    2/    5] | d_loss: 1.4654 | g_loss: 0.9972\n",
            "Epoch [    2/    5] | d_loss: 1.1172 | g_loss: 0.9824\n",
            "Epoch [    2/    5] | d_loss: 1.0834 | g_loss: 1.0037\n",
            "Epoch [    2/    5] | d_loss: 1.1478 | g_loss: 1.2069\n",
            "Epoch [    2/    5] | d_loss: 1.1052 | g_loss: 0.9352\n",
            "Epoch [    2/    5] | d_loss: 1.1710 | g_loss: 1.3207\n",
            "Epoch [    3/    5] | d_loss: 1.2331 | g_loss: 1.0970\n",
            "Epoch [    3/    5] | d_loss: 1.0750 | g_loss: 1.0815\n",
            "Epoch [    3/    5] | d_loss: 1.0990 | g_loss: 1.2892\n",
            "Epoch [    3/    5] | d_loss: 1.0565 | g_loss: 1.0109\n",
            "Epoch [    3/    5] | d_loss: 1.2696 | g_loss: 0.8061\n",
            "Epoch [    3/    5] | d_loss: 1.0858 | g_loss: 1.0255\n",
            "Epoch [    3/    5] | d_loss: 1.0361 | g_loss: 1.0398\n",
            "Epoch [    3/    5] | d_loss: 1.0664 | g_loss: 1.2805\n",
            "Epoch [    3/    5] | d_loss: 1.1135 | g_loss: 1.0645\n",
            "Epoch [    3/    5] | d_loss: 1.0640 | g_loss: 1.1383\n",
            "Epoch [    3/    5] | d_loss: 1.0882 | g_loss: 1.0949\n",
            "Epoch [    3/    5] | d_loss: 1.2650 | g_loss: 0.9288\n",
            "Epoch [    3/    5] | d_loss: 1.0940 | g_loss: 1.1489\n",
            "Epoch [    3/    5] | d_loss: 1.0555 | g_loss: 1.2722\n",
            "Epoch [    3/    5] | d_loss: 1.0530 | g_loss: 1.2082\n",
            "Epoch [    4/    5] | d_loss: 1.1896 | g_loss: 1.3418\n",
            "Epoch [    4/    5] | d_loss: 0.9941 | g_loss: 1.0610\n",
            "Epoch [    4/    5] | d_loss: 1.0248 | g_loss: 1.0328\n",
            "Epoch [    4/    5] | d_loss: 1.0425 | g_loss: 1.2570\n",
            "Epoch [    4/    5] | d_loss: 2.2272 | g_loss: 0.7256\n",
            "Epoch [    4/    5] | d_loss: 1.0183 | g_loss: 1.1507\n",
            "Epoch [    4/    5] | d_loss: 1.0640 | g_loss: 1.0510\n",
            "Epoch [    4/    5] | d_loss: 0.9983 | g_loss: 1.2124\n",
            "Epoch [    4/    5] | d_loss: 1.0811 | g_loss: 1.1205\n",
            "Epoch [    4/    5] | d_loss: 1.0593 | g_loss: 0.9293\n",
            "Epoch [    4/    5] | d_loss: 1.0018 | g_loss: 1.0990\n",
            "Epoch [    4/    5] | d_loss: 0.9538 | g_loss: 1.1971\n",
            "Epoch [    4/    5] | d_loss: 1.0625 | g_loss: 1.2414\n",
            "Epoch [    4/    5] | d_loss: 0.9840 | g_loss: 1.2959\n",
            "Epoch [    4/    5] | d_loss: 1.5008 | g_loss: 3.5524\n",
            "Epoch [    5/    5] | d_loss: 1.4995 | g_loss: 0.9392\n",
            "Epoch [    5/    5] | d_loss: 0.9847 | g_loss: 1.0821\n",
            "Epoch [    5/    5] | d_loss: 1.0285 | g_loss: 1.3189\n",
            "Epoch [    5/    5] | d_loss: 0.9758 | g_loss: 1.1566\n",
            "Epoch [    5/    5] | d_loss: 1.0084 | g_loss: 0.9321\n",
            "Epoch [    5/    5] | d_loss: 0.9223 | g_loss: 1.1210\n",
            "Epoch [    5/    5] | d_loss: 0.9085 | g_loss: 1.0563\n",
            "Epoch [    5/    5] | d_loss: 0.9755 | g_loss: 1.4606\n",
            "Epoch [    5/    5] | d_loss: 1.2039 | g_loss: 1.5792\n",
            "Epoch [    5/    5] | d_loss: 1.0906 | g_loss: 1.3794\n",
            "Epoch [    5/    5] | d_loss: 0.9271 | g_loss: 0.9405\n",
            "Epoch [    5/    5] | d_loss: 0.9153 | g_loss: 1.1027\n",
            "Epoch [    5/    5] | d_loss: 0.9518 | g_loss: 1.3008\n",
            "Epoch [    5/    5] | d_loss: 0.9900 | g_loss: 1.1066\n",
            "Epoch [    5/    5] | d_loss: 0.9710 | g_loss: 1.2052\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nFSlMbeYUK8z",
        "colab_type": "text"
      },
      "source": [
        "# Training Loss\n",
        "Plotting the loss of Discriminator and Generator "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "13GY2TUyUcmu",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "b887d696-ef98-48e4-81bd-b0849b71fea8"
      },
      "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()"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f5f6052a710>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsvXd8o1ed7/8+6rYlN7mMx/a4TMt0\nz4ynJJNGCiGBJMAmG1oIbIHsLrCUZeHe3cvN8ruv/V12YXeB8ApkgYSSJYFACiGBJb1NJlMyvWSa\n24x7k2Vb9Tn3jyPJkiy3GcmWnfN+vfyy/ejRoyPpeT7P93zO93yPkFKi0Wg0moWFaa4boNFoNJr0\no8Vdo9FoFiBa3DUajWYBosVdo9FoFiBa3DUajWYBosVdo9FoFiBa3DXzBiGEWQjhFUIsSee+Gs1C\nROg8d02mEEJ44/7NBfxAOPL/p6WUD81+qy4eIcT/AaqklJ+Y67ZoNBNhmesGaBYuUkpn9G8hRBPw\nF1LKZyfaXwhhkVKGZqNtGs1CR9symjlDCPF/hBCPCCF+IYQYAj4mhLhUCPGGEGJACNEuhPiOEMIa\n2d8ihJBCiNrI/z+PPP6MEGJICLFTCFE3030jj98ohHhbCDEohPiuEOI1IcQnLuA9rRFCvBRp/yEh\nxHvjHnufEOJY5PXbhBBfiGwvE0I8HXlOnxDi5bjnVAkhHhNCdAshzgoh/ibuse1CiH1CCI8QolMI\n8a8zba9m4aLFXTPXfAD4L6AAeAQIAX8LlAA7gPcAn57k+R8B/hdQDLQA/99M9xVClAG/BL4ced2z\nwNaZvhEhhA14CvgdUAp8AXhECLEssssDwJ9LKV3AeuClyPYvA2ciz1kE/GPkeKbI8XYDlcD1wJeF\nENdGnvdd4F+llPnAMuDRmbZZs3DR4q6Za16VUv5WSmlIKUellLullLuklCEp5RngfuCqSZ7/qJRy\nj5QyCDwENFzAvu8D9kspn4g89u9AzwW8lx2ADSW4wYgF9QzwocjjQWC1EMIlpeyTUu6L274YWCKl\nDEgpo5H7pUC+lPKfI9tPAT9KOt5yIYRbSjkkpdx1AW3WLFC0uGvmmtb4f4QQlwghfieE6BBCeICv\no6LpieiI+3sEcE604yT7Lo5vh1RZBm3TaHsyi4EWmZil0IyKukH1Um4BWoQQLwohtkW2/9/Ifs8J\nIU4LIb4c2V4DLInYNQNCiAHg71HRPcAngdXACSHEm0KImy6gzZoFihZ3zVyTnK71A+AwsCxiN3wN\nEBluQztQFf1HCCEYE+SZcB6ojjw/yhLgHECkR3ILUIayWx6ObPdIKb8gpawF3g98RQhxFeqGc1JK\nWRj345JS3hx53gkp5Ycix/sW8GshhOMC2q1ZgGhx12QbLmAQGBZCrGJyvz1dPAVsEkLcLISwoDz/\n0imeYxZCOOJ+7MDrqDGDLwkhrEKIa4CbUL57jhDiI0KI/Ij1MwQYAJHXXRq5KQyi0kUNYCcQEEJ8\nKfIaZiHEOiHE5sjz7hRClEgpjcjzZPSYGo0Wd0228SXgLpT4/QA1yJpRpJSdwB3AvwG9wFLgLVRe\n/kR8DBiN+zkhpfQDNwO3ojz77wAfkVKejDznLqA5Yjf9eeQYACuB5wEv8BrwbSnlK5G00JtQg7tN\nkWP+AMiPPO8m4Fgk0+ibwB1SysCFfxKahYSexKTRJCGEMKMsltuklK/MdXs0mgtBR+4aDSCEeI8Q\nojBir/wvVCbKm3PcLI3mgtHirtEoLkflmncDNwAfiNgsGs28RNsyGo1GswDRkbtGo9EsQOascFhJ\nSYmsra2dq5fXaDSaecnevXt7pJRTperOnbjX1tayZ8+euXp5jUajmZcIIZqns5+2ZTQajWYBosVd\no9FoFiBa3DUajWYBklUrMQWDQdra2vD5fHPdFE0SDoeDqqoqrFbrXDdFo9FMg6wS97a2NlwuF7W1\ntSQW1tPMJVJKent7aWtro66ubuonaDSaOSerbBmfz4fb7dbCnmUIIXC73bpHpdHMI7JK3AEt7FmK\n/l40mvlF1om7RqN5h+DthoGWuW7FgkWLexJms5mGhgbWrFnDhg0b+Na3voVhqPUP9uzZw+c+97mL\nfo3vf//7/PSnP53Rcy677LILfr0HH3yQ8+fPX/DzNZqM0PwqvP2HuW7FgmXKAdXIsl0vA/bI/o9K\nKf930j6fAP6VyHJiwL1Syh+mt6mzQ05ODvv37wegq6uLj3zkI3g8Hv7pn/6JxsZGGhsbL+r4oVCI\nu+++e8bPe/311y/4NR988EHWrl3L4sWLp/2ccDiM2Wy+4NfUaKYkHAIjNNetWLBMJ3L3A9dIKTeg\nVot/jxBie4r9HpFSNkR+5qWwJ1NWVsb999/Pvffei5SSF198kfe9730AvPTSSzQ0NNDQ0MDGjRsZ\nGhoC4Bvf+Abr1q1jw4YNfPWrXwXg6quv5vOf/zyNjY18+9vf5p577uGb3/xm7LEvfOELNDY2smrV\nKnbv3s0HP/hBli9fzj/+4z/G2uJ0qrWcX3zxRa6++mpuu+02LrnkEj760Y8Srez59a9/nS1btrB2\n7Vo+9alPIaXk0UcfZc+ePXz0ox+loaGB0dFRnnvuOTZu3Mi6dev4sz/7M/x+Vdm2traWr3zlK2za\ntIlf/epXs/Mha965yDAY4bluxYJlysg9spK7N/KvNfKT8TrBL57oonsoveW0S112rl5ZNqPn1NfX\nEw6H6erqStj+zW9+k+9973vs2LEDr9eLw+HgmWee4YknnmDXrl3k5ubS19cX2z8QCMRq6dxzzz0J\nx7LZbOzZs4dvf/vb3Hrrrezdu5fi4mKWLl3KF77wBdxud8L+b731FkeOHGHx4sXs2LGD1157jcsv\nv5zPfOYzfO1rXwPgzjvv5KmnnuK2227j3nvv5Zvf/CaNjY34fD4+8YlP8Nxzz7FixQo+/vGPc999\n9/H5z38eALfbzb59+2b0GWk0F4QRAqmXfM0U0/LcIwvz7ge6gD9KKXel2O1PhBAHhRCPCiGqJzjO\np4QQe4QQe7q7uy+i2XPPjh07+OIXv8h3vvMdBgYGsFgsPPvss3zyk58kNzcXgOLi4tj+d9xxx4TH\nuuWWWwBYt24da9asoaKiArvdTn19Pa2treP237p1K1VVVZhMJhoaGmhqagLghRdeYNu2baxbt47n\nn3+eI0eOjHvuiRMnqKurY8WKFQDcddddvPzyy9Nqp0aTVoywit41GWFak5iklGGgQQhRCDwmhFgr\npTwct8tvgV9IKf1CiE8DPwGuSXGc+4H7ARobGyeN/mcaYWeKM2fOYDabKSsr49ixY7HtX/3qV3nv\ne9/L008/zY4dO/jDHyYfGMrLy5vwMbvdDoDJZIr9Hf0/FBrvScbvYzabCYVC+Hw+/vqv/5o9e/ZQ\nXV3NPffcc0F56ZO1U6NJK9qWySgzypaRUg4ALwDvSdreG7ck2Q+Bzelp3tzS3d3N3XffzWc+85lx\ned6nT59m3bp1fOUrX2HLli0cP36c66+/ngceeICRkRGABFsm00SFvKSkBK/Xy6OPPhp7zOVyxcYE\nVq5cSVNTE6dOnQLgZz/7GVddddWstVOjiWGEtS2TQaaTLVMKBKWUA0KIHOB64BtJ+1RIKdsj/94C\nHGOeMjo6SkNDA8FgEIvFwp133skXv/jFcfv9x3/8By+88AImk4k1a9Zw4403Yrfb2b9/P42Njdhs\nNm666Sb++Z//eVbaXVhYyF/+5V+ydu1aFi1axJYtW2KPfeITn+Duu+8mJyeHnTt38sADD3D77bcT\nCoXYsmXLBWXvaDQXjRGJ3KUEPUku7Uy5hqoQYj3KZjGjIv1fSim/LoT4OrBHSvmkEOL/R4l6COgD\n/kpKeXyy4zY2NsrkxTqOHTvGqlWrLvjNaDKL/n40aWXn98Dngav+Hkw67Xa6CCH2SimnzMmeTrbM\nQWBjiu1fi/v7fwD/Y6aN1Gg072CiOe5GWIt7BtAzVDUazdwQHUzVGTMZQYu7RqOZG6KirjNmMoIW\nd41GMzfEInedMZMJtLhrNJrZxzBUlgxoWyZDaHHXaDSzT7ygGzpyzwRa3FPQ2dnJRz7yEerr69m8\neTOXXnopjz322Jy05cUXX7yoipAaTVYSXw1SR+4ZQYt7ElJK3v/+93PllVdy5swZ9u7dy8MPP0xb\nW1vGXjNViYEoFyLukx1Po8kK4gdR9YBqRtDinsTzzz+PzWZLmLVZU1PDZz/7WcLhMF/+8pfZsmUL\n69ev5wc/+AEweRnevXv3ctVVV7F582ZuuOEG2tvVRN7kMsC//e1v2bZtGxs3buS6666js7OTpqYm\nvv/97/Pv//7vNDQ08Morr9DU1MQ111zD+vXrufbaa2lpUSvZRGehbtu2jb//+7+f5U9No5kh8dG6\njtwzwrQKh80JJ58Fb2d6j+ksh+XXTbrLkSNH2LRpU8rHfvSjH1FQUMDu3bvx+/3s2LGDd7/73UDq\nMrzbtm3js5/9LE888QSlpaU88sgj/MM//AM//vGPgcQywP39/bzxxhsIIfjhD3/Iv/zLv/Ctb32L\nu+++G6fTyd/93d8BcPPNN3PXXXdx11138eMf/5jPfe5zPP744wC0tbXx+uuv60U2NNmPjtwzTvaK\ne5bwN3/zN7z66qvYbDZqamo4ePBgrCjX4OAgJ0+exGazxcrwArEyvIWFhRw+fJjrr78eUKsbVVRU\nxI4dX163ra2NO+64g/b2dgKBAHV1dSnbs3PnTn7zm98AqmZ7fJR+++23a2HXzA8MHblnmuwV9yki\n7EyxZs0afv3rX8f+/973vkdPTw+NjY0sWbKE7373u9xwww0Jz3nxxRdTluGVUrJmzRp27tyZ8rXi\ny+t+9rOf5Ytf/CK33HILL7744rgFPaaDLtermTck2DIZX/vnHYn23JO45ppr8Pl83HfffbFt0RK+\nN9xwA/fddx/BYBCAt99+m+Hh4QmPtXLlSrq7u2PiHgwGUy6gAaoXUFlZCcBPfvKT2Pb4cr2gFsp+\n+OGHAXjooYe44oorLuRtajRzS3y2jLZlMoIW9ySEEDz++OO89NJL1NXVsXXrVu666y6+8Y1v8Bd/\n8ResXr2aTZs2sXbtWj796U9Pmplis9l49NFH+cpXvsKGDRtoaGiYMPPlnnvu4fbbb2fz5s2UlJTE\ntt9888089thjsQHV7373uzzwwAOsX7+en/3sZ3z7299O+2eg0WQcbctknClL/mYKXfJ3/qG/H03a\n6DsDBx5Rf6+6GRatndv2zCOmW/JXR+4ajWb20ZF7xtHirtFoZh+dCplxsk7c58om0kyO/l40aSUh\nW0bXlskEWSXuDoeD3t5eLSRZhpSS3t5eHA7HXDdFs1DQ2TIZJ6vy3Kuqqmhra6O7u3uum6JJwuFw\nxCZpaTQXjaEj90yTVeJutVonnJmp0WgWEPGCrgdUM0JW2TIajeYdgrZlMo4Wd41GM/tEBV2YdOSe\nIbS4azSa2Scq6GaLXokpQ2hx12g0s48RApMFhFlH7hlCi7tGo5l9DANMJjCZteeeIbS4azSa2UeG\nVdSuI/eMMaW4CyEcQog3hRAHhBBHhBD/lGIfuxDiESHEKSHELiFEbSYaq9FoFghRW0ZH7hljOpG7\nH7hGSrkBaADeI4TYnrTPnwP9UsplwL8D30hvMzUazYLCCCthFyY9iSlDTCnuUuGN/GuN/CTXB7gV\niK4w8ShwrRBCpK2VGo1mYRG1ZUxmLe4ZYlqeuxDCLITYD3QBf5RS7krapRJoBZBShoBBwJ3iOJ8S\nQuwRQuzRJQY0mncwschd2zKZYlriLqUMSykbgCpgqxDigirrSynvl1I2SikbS0tLL+QQGo1mIZBg\ny2hxzwQzypaRUg4ALwDvSXroHFANIISwAAVAbzoaqNFoFiDxtoyO3DPCdLJlSoUQhZG/c4DrgeNJ\nuz0J3BX5+zbgeanr9mo0monQk5gyznSqQlYAPxFCmFE3g19KKZ8SQnwd2COlfBL4EfAzIcQpoA/4\nUMZarNFo5j9GGCwOZcvoyD0jTCnuUsqDwMYU278W97cPuD29TdNoNAsWGVbCrlMhM0ZW1XPXaDTv\nEIxwxJYRWtwzhBZ3jUYz+0SzZRDalskQWtw1Gs3sE6stI/SAaobQ4q7RaGafaLYMUkfuGUKLu0aj\nmX2MsCr5K9GRe4bQ4q7RaGYfaURsGalXYsoQWtw1Gs3sIuVYtow0dOSeIfRiHRqNZnaJeuymaFVI\nHb1nAh25azSa2SUaqYtIKiREct11rJlO9Kep0WhmFyOkfkdXYgJtzWQAHblrNJrZJWbLmMBI2qZJ\nG1rcNRrN7BJvy5iStmnShrZlNBrN7BKL3C2qcFj8Nk3a0JG7RqOZXeKzZaJFw3TxsLSjxV2j0cwu\nCbZMRNR15J52tLhrNJrZJZYtEx+5a3FPN1rcNRrN7JLKltGRe9rRA6oajWZ2iQq6MI8NqGrPPe3o\nyF2j0cwu8ZOYonaMtmXSjo7cNRrN7BJvywhz4jZN2tDirtFoZhcZL+7alskUWtw1Gs3sErVlhHms\ntoyO3NOOFneNRjO7RMv7xtsy2nNPO1rcNRrN7JJQFVKXH8gUWtw1Gs3sEj9DVUfuGUOLu0ajmV2S\nV2ICPaCaAaYUdyFEtRDiBSHEUSHEESHE36bY52ohxKAQYn/k52uZaa5Go5n3GKGI3y7iUiG1uKeb\n6UxiCgFfklLuE0K4gL1CiD9KKY8m7feKlPJ96W+iRqNZUMjwWAqkXokpY0wZuUsp26WU+yJ/DwHH\ngMpMN0yj0SxQDGNM1HU994wxI89dCFELbAR2pXj4UiHEASHEM0KINRM8/1NCiD1CiD3d3d0zbqxG\no1kAGCGVKQNxk5i0uKebaYu7EMIJ/Br4vJTSk/TwPqBGSrkB+C7weKpjSCnvl1I2SikbS0tLL7TN\nGo1mPiPDY167ECqK15F72pmWuAshrChhf0hK+Zvkx6WUHimlN/L304BVCFGS1pZqNJqFgREes2VA\nRe86ck8708mWEcCPgGNSyn+bYJ9Fkf0QQmyNHLc3nQ3VaDQLhGi2TBSTWWfLZIDpZMvsAO4EDgkh\n9ke2/U9gCYCU8vvAbcBfCSFCwCjwISmlzEB7NRrNfEcaY7YMqL915J52phR3KeWrgJhin3uBe9PV\nKI1Gs4BJtmXiV2TSpA09Q1Wj0cwu8dkyoCJ3PaCadrS4azSa2SU+WwYikbsW93SjxV2j0cwu47Jl\nhI7cM4AWd41GM7uME3ftuWcCLe4ajWZ2SWXL6Mg97Whx12g0s0vKyF2Le7rR4q7RaGaX5GwZHbln\nBC3uGo1mdkm2ZbTnnhG0uGs0mtnFMMbWTgX1t7Zl0o4Wd41GM3tIOcEkJh25pxst7hqNZvaI2i9C\nV4XMNFrcNRrN7BG/OHYUPaCaEbS4azQLkN1NfRxsG5jrZozHCKnfybaMjtzTjhZ3jWYBcrzdw8lO\n71w3YzxRERfxA6o6cs8E06nnrtFo5hmBsMQSzsJBylS2jPbcM4IWd41mARIIGVjNky7DMDfExD1p\nEpNe2yftaFtGo1mABMMGgVAWRu4xWyZpEpMR1gKfZrS4azQLjFDYIGxIguEsFMuJsmVAz1JNM1rc\nNZoFRlTUAyGDrFvKOJYtkxS5gx5UTTNa3DWaBUYgMpBqSEnYyDJxT2nLmBIf06QFLe4azQIj3mvP\nOmsmpS1jSnxMkxa0uGs0C4xgXApkINvSIVNly0SjeB25pxUt7hrNAiNB3LMtYyaVLWPSnnsm0OKu\n0SwwEm2ZLBP3lJOYdLZMJtDirtEsMOKtmOwT9xTZMjoVMiNocddoFhjxg6hZJ+4TTWICbcukmSnF\nXQhRLYR4QQhxVAhxRAjxtyn2EUKI7wghTgkhDgohNmWmuRqNZiribZlAKNuyZSJtS5UtowdU08p0\nasuEgC9JKfcJIVzAXiHEH6WUR+P2uRFYHvnZBtwX+a3RaGaZ7M6WSVXyV6dCZoIpI3cpZbuUcl/k\n7yHgGFCZtNutwE+l4g2gUAhRkfbWajSaKQmEDSwmVTRsXtkyOnJPKzPy3IUQtcBGYFfSQ5VAa9z/\nbYy/ASCE+JQQYo8QYk93d/fMWqrRaKZFIGSQYzMjBASzLRXSCIMQSQtkRz33LGvrPGfa4i6EcAK/\nBj4vpfRcyItJKe+XUjZKKRtLS0sv5BAajWYKgmEDm8WE1WzKTlsm3m8HHblniGmJuxDCihL2h6SU\nv0mxyzmgOu7/qsg2jUYzywTDBlazCZvZlIWTmIxESwb0JKYMMZ1sGQH8CDgmpfy3CXZ7Evh4JGtm\nOzAopWxPYzs1Gs00CYQMbGYTNospO2vLjIvco9kyWXYjmudMJ1tmB3AncEgIsT+y7X8CSwCklN8H\nngZuAk4BI8An099UjUYzHQJhSY7NhDVkyr4BVSOUmCkDcZOYdOSeTqYUdynlq8Ck63VJVTT6b9LV\nKI1Gc+GoyF1gNYvs89xleLwtoycxZQQ9Q1WjWWBEB1SVLZNl4q5tmVlDi7tGs8AIhrJ4QDWVuOsB\n1YygxV2jWUCEDUnIkFjNKhUy6yL3yWwZ7bmnFS3uGs0CIirmNosJ63zJltGRe0bQ4q7RLCCiA6g2\ns0kNqGbbItmpsmWEUL67jtzTihZ3jWYBEfXYbRYTdou6vLMqYyaVLQOqHIEeUE0rWtw1mgVE1JaJ\neu5qWzZF7uHEujJRhFnXlkkzWtw1mgVEMFK/3WoWY+KeTRkzRni8LQPKd9e2TFrR4q7RLCACYSWQ\n0cJhkGVlfyeyZYRZD6imGS3uGs0CIrryki2S5w7gz7rIPZW46wHVdKPFXaNZQMR77jZLFkbuqbJl\nQAm+jtzTihZ3jWYBEYjPczdHV2PKogFVGR4rNxCPjtzTjhZ3jWYBEQwZCAEWk8CalZG7kdqWMels\nmXSjxV2jWUD4Iwt1CCFinnvW5LkbhsplT2XLCJ0tk260uGs0C4hgyIhNXopmy2RN8bBUi2NHMZn1\nJKY0o8Vdo1lABMMyJupmk8BiEtljy0QHTFNmy+gB1XSjxV2jWUAEwuGYuAOR4mHZIu4h9Xsiz13b\nMmlFi/uFMNgGB3+pIw1N1hEMyVgKJChrJpr7PudMZssIk76e0owW9wuh7yz0ngbf4Fy3RKNJIBA2\nYimQALZsWmpvUltGFw5LN1rcL4SAN/J7eG7bodEkodZPHbusbRZT9tSWiYm7nsQ0G2hxvxD8Wtw1\n2Ul0/dQoWbUa06S2jPbc040W9wshMBT57Z3bdmg0SQQjee5RskrcJ7NldOSedrS4XwixyF2LuyZ7\nMAxJMDx+QDVrCodNli0jsiTP3dOuxtMWAFrcZ4phQHBE/a1tGU0WEYgrGhbFnk3rqE46iSkLassY\nBhx9Ao48tiB6EVrcZ0rAC9E1Kf06ctdkD8G49VOjZJctE2nHhJOY5ridXUdgtB/CQRhomdu2pAEt\n7jMlasWYzNqW0WQV8eunRrGaBWFDEjayIHqP2TJZuBKTYUDz65BXotrXN/+tmSnFXQjxYyFElxDi\n8ASPXy2EGBRC7I/8fC39zcwiotF6Xom2ZTRZRdR+ic9zz6rKkFNNYpJy7qL3rqMw0gd1V0LhkgXh\nu08ncn8QeM8U+7wipWyI/Hz94puVxUQzZVwVynuf666kRhMhmMJzz6rVmGLZMqpNUkqCYQPDkMho\njfe5iN6jUbuzFEpWgHuZEvqRvtlvSxpJ0T9KREr5shCiNvNNmSf4vQQMOOfLpU5KCA6D3TXXrdJo\nYgJutyROYoIsidyTbJlXTvawt7kfgMVD56kd6KU1t5nbti2b3XZ1H4ORXljzARAC3PVwEhW95xbP\nblvSSLo890uFEAeEEM8IIdZMtJMQ4lNCiD1CiD3d3d1peulZJuClw2fm5SYfvmBYWzOarCFV5J5V\ni2RHUx0jtkzHoI/CXCuXLnWzanERLoeV9v7h2W1rvNdeulJtyylS//eemr12ZIB0iPs+oEZKuQH4\nLvD4RDtKKe+XUjZKKRtLS0vT8NJzgN/LkOEgYM5lJBB6R2fMdHl8nOoamutmaCJEB1StSQOqoAqK\nzTlJk5g8viAVBQ6217tZv6SYUqcdIQ28vtDstan7OAz3QO3lKmqP4l4Kg60Q8s9eW9LMRYu7lNIj\npfRG/n4asAohSi66ZdlKYIghw07QnMuwP/yOzph5/XQvfzjSiZRZIByalKmQY6sxZUHedpwtYxiS\nYX8Yl8OqtgkzNosJEwZDsynu59+CXDeUXpK4vXipuhn1N81eW9LMRYu7EGKREOqWJ4TYGjlm78Ue\nN2vxexk0HATMOQwHQu9YW0ZKSafHRyBkMOSfxYtRMyGBsFo/NaEqpCW6GlMW3IBj2TImhgMhDClx\n2iPDfiYzdosJIcN4fMHZa9NIDxRUJUbtoLZZ7PPamplyQFUI8QvgaqBECNEG/G/ACiCl/D5wG/BX\nQogQMAp8SC7UUC4cwgiM4DEWIYUFT8jyjo3ch/whRgLqYu3zBsiPRmCaOSMQGls/NUpWee5GWFky\nQuCNBAQuR0SCIpG7CBqxxzJO0AeBkdSDpiYzFNerQVUpx4v/PGA62TIfnuLxe4F709aibCYwhC8Y\nxm/Ko9Rlx9tuJzDiwTbX7ZoDOgd9sb97h/3UluTNYWs0oPLc4y0ZyFJxh5j1MmbLmDAJQZ6V2bNl\nRlWmDjkTZMS4l0LXMRjqgPyK2WlTGtEzVGeC34s/ZBAw57K8zEnQnMuQ5525YEenx49JCHJsZnq8\ngblujoZoRcjECNNqFghBdizYIcOxTJkxcR+zZQCcNhNe/yzZMqORPPacotSPF9eriH2eWjNa3GdC\nwIsvGCZozmVZmVNlzHgH5rpVc0KHx0eJy0ap007fsBb3bCAQMhIyZQCEEJGl9rJA3BMi9yA2i2ks\nJz8yicllF7MbuQsxsbjb8iB/sRb3dwR+L76gAXYnxXk2LA4Xo8OesUJi7xCklHQN+ViU76DYaaNv\nOKAzZrKAQNgYZ8uAypjJisqQRig2gcnrD+G0W8bGByKin2dT4j4r59NIn5qAaJ7EnS6sAW8nhOdf\n0oAW95kQGMIXhrw8F0II8lwFjPr88zoX9kIYGAniDxqU5zsoybMTCBl4Ruffyb/QSF6FKYrVLLLD\nc0+yZWKWDMS2u6yCQMiYnXIsHzkfAAAgAElEQVQJo30T++1RcotV8Ob3ZL49aUaL+0zwexky7BTm\nqSHU/PwiRgNhgqPz74u/GDo8ajC1LN+O26k+i97hd9YNLqMM98LuH44N+E2T5PVTo1gtWVL21wjH\n6soM+YJjaZAQF7knevIZZbR/6vICjsKxfecZWtxngOEfwiMdFOSoEf6CwkIkMDA4/774i6HT48Nq\nFpTk2SnOi4p7mnz33tMqRe2dTOsb4O2GvjMzelryEntRsmY1JiMMJgthQzISiJvABLHIPS+yaShV\nrruU4DmfnrYERtR5NpHfHiUnIu6++Te2psV9BviHPfhNuRTmKEFzF6m7fn///PviL4ZOj49Slx2T\nSeCwmnE5LPSmI2PG74WDv1Ti9k4lMAydR9TfnvYZPTUYluMGVCG6GlMWiHvElvH6QkhJoi0Tjdyt\nyoNPmevedRT2/iQ9C2nEMmWmiNxtTjVOoCP3hY1/eJCAOZfCXBVeuAoKMZsEHs/8Lg06EwxD0j3k\npzzfEdtWnGdLjy0zEpnYPJMp39GJKAuFc/tUhJtXAkPTF3cp5cS2jNlEcK4jdykhOAomM0ORVMdE\nz12Jeo4FTGKCjJnoedGRcmmJmRHLcZ8ichdCRe+j0wzgRvrglW/BQOvFtS8NaHGfLqEAft8oAXMe\n+RFbRlgc5DjseN9Bue69wwGCYZkg7m6nnT5vAONiV/uJivtQx/StmaNPwOFfX9zrxmOE4cxLmS8I\nlyobJByC8/vU5JnSS9TnMc3B+mgeu80yfialda6zZQwDTjwN3i4oqhs/gQlitowJgzy7ObW4RyP2\n7mNqKbyLYaRPpV9GbZfJcBRO35bpPgGhgOplzDHzUtznJGc3oCYwhaxOXNGBICFw5LoY9XrmXypg\nOAhv/ueMV5zpjAymJoh7no2QIS++Jki0qyzl9LreIb+K5jznJ0xVO9szzBP7z03/xjPQrErAth+Y\n3v4pODcwytmeSWoOndsHO++Foc7E7V1HVS+kqlHlV0upbnTTYGwVptTZMnM2iSkcgiO/gfaDqvLi\nku0xyyXVgCqGQb7DOt5z9w2q6LlkuRLPi809H+0DR0Hq9VyTiUbu07nGo+3qPTXnKdLzTtxPdXn5\n8WtnZ3/ijF+VHrDn5mMyjUVHuc5CTEEvAyOzWOwoHYz2q1Kn/Wdn9LROjw+bxURR7ljUFc2YueiZ\nqiN9yo4wW6ZnzfQ3qxrh0gBvahE82DbAme5hzg2MTq8N0e70DD+XKFJKnj3aye8Pd6RetzQchKZX\nI+MLj4zZA1JC25vq/RfVqZW+YPrWzP7/YtHQ4ZSpkLaI5z7rAUjID4d+CT0nYfn1UHcFCMGQL4jD\nak5sa3TpPRnG6bCMj9yj30vt5WB3jo1LXCij/VNbMlFyitT3FpzC/guMgOec2t/nUT2VOWTeiXt5\nvh2A3x1qn91BooAXXyhMrjOxG+fML8AaHqHbO89SAX0RK2lkZgNFHR4f5fmOhOJUsYyZi/0MouJe\nsERF0ED/cGDinlrfmbHFllMMPoYNSVu/EvVT3dO0WQYjIjJ47oLmL/R4A/QNB/AFwzT3pojeOw6q\nQdMV71YDjAceVkI/0KwyZKq2KJ/Xlqsixulkh4z2Q38z+b7zKSN3m9mkLO/ZtGZG+2H/Q0qUV92s\neiMRhnwhnI6kiUOxyD2My2HB60+ayDTQAlYHOMuhfI3qcV5oRVYp1bk23VWWYumQU1gzfWfUsZdd\nFylbcPLC2pcm5p24uxxW3rNmEb1ePy8cn707o/QP4Qsa5LoKErY7XQXYjFG6h+aZuEdP1JHpV2cO\nhQ16hgIsirNkAOwWlTFzUb2pcEj5mrluKKqF4R5Ot3Xw053NvHAixfcspVqhvrhOzTIcGi+C7YOj\nBEIGDquZ013eqSPXcFCJqWuR6g1cwKDYya4hpc0WE293Ji1kYoShZRcUVMLiTbDuT5VAHXwEmneC\nNUcJVxRXxfRsmf4mwobEFh6ecEAVZrF4WPcJ2POAOsfW/gksWpvw8JAvRH6yuMcidwOXwxpLl4wx\n0AIF1Uo0y9eq76fr+IW1LzCsvuupMmWiTDcdsveUKlngXqq+ux4t7jOmtiSPrbXFHDnv4cj5mQ9m\nDo4GGQ3MbPEC/4iHgLSQ70ysfmh25OMyh+gZnGd13aORu29g2lOre7wBDCljvad4Spx2eqYS93BI\nedmpFhUf7VeCnVMMRTUMjAbYtXcvACc7h/CHkr6vkV7V9XUvVRX7UkTuzb0jmIRge30xQ74QnZ4p\nbsCe80qAl1wasYZmZs1IKTnZ6aWqKJcV5S5OdyctGdd5RH3uSy5TIlVQCWs/qN5LfxNUbgJz3CCj\nq0LtP1WE2t9MWErsYe8EM1SnIe69p2H/f13cbGsjDCefhcO/UVFx4yehZPx6qEO+UKLfDurzEELZ\nMpHHYtaMz6POj8Il6n9nmVrMuvMCs2ZiaZDjB1N9wfB4v98RCegmS4c0wipydy9V76Nkubox+1Os\nVNZ+cOz6yyDzUtwBtte7qS7O5YXjXfTMwA5o7h3mZzub+NGrZ3jxRNe0BwFHhwYImnMpyE0q8GvL\nI89mYWAgs7nugdAUda77m2Hfz9Rg03SIRiFSTjuHNzoztbzAMe4xt9NG//AUGTNdR+D406lFM3rB\n5bo5F8rnSHeQxXRxS8NigmElmglEJ/gU14NrsXoPwURfvbl3hIoCB6sq8jEJwamuKayZwVZ1YRbV\nqJoifTMT96gls7zMycpyF4GQQVN0YNUwoOUNJUrupWNPKq6HVbeoAdTFm2Kbu4f8SNci9c9k+e5S\nwkAzhiGxhn1YxfjPf2zBjknEvfVNdQ6deXG6bzcRI6xuDm27lQXT8LGUnnYwbOALJk1ggoi4m8EI\nx6L6WHXI6OB6VNwBytepm/HIBaQhR5+TwpZ57lgX/7WrJTGYMFuVzz+ZLTPYpm6M7sjNzL1c/U4e\n+PV2qcyh1jdn3u4ZMm/F3WQS3Lh2ETaLiacOnOe5Y508sf8cP3+jmR++cobXTvUQSopUznR7eXL/\neQpybSwvd3GgdZAHX2vij0c7GZxiQHTUG8lxz0k6KW1O8uxmAqMenth/jsffOsev97bx+Fvn6PKk\nZ6alLxjmkd0tPPja2YmzMM7vUyfYBFkmI4EQT+w/R380uvYNKDsDxoR1CjoGfeTazGPZQnEU59kI\nG5KB0Uk+x2hmjudcigYqe6grlMvj+88TdFZxafEQtcU5FOVaOdaeVOKh74zy5x0FY7W24/zpkUCI\nriEfNe5cHFYzVUU5nOoamtyaGWiFvFJljxTVRnoH04+wopbMsjInVUU55NnNnIhaMz1vq+NFo/Z4\nyi6BzXcpAQFOdQ3x8zeaaQ7kq31TWE4xhrshMMJobgUgsYbHnx9jS+1NIO4+j/L8bXkqk+dCcrTb\ndqvz75L3qsHTCYpxjSv1G4/JFBtQBfBEI/fBVrUqUl7Z2L7lq9VncyHR+2i/8vjtiRarYUha+kYY\nCYTZ35Ik5FOlQ/aeUscsqlX/55WonkFPnLhLCW//ASwONTCcYeatuAPk2S3cuLaC0aDByS5vrBhR\nqdPGm2f7+PkbzbT2qRHuU11DPHWwHbfTzu2bq7hhzSI+saOWdZUFHG/38NCbzakHwCIERjwJOe4x\n7E4Kc20sygnh8anViYJhg06Pj1/tbZt+lkYyhgHeLoJt+9jz+59Tfeoh1g88y5P7z3OoLUlwwqEx\n4Zxgyvre5n7OdA/zxpneSLQ+oKJGmJbvPjgS5GTnEDXuvITB1CglTmXVTDioaoTHIvZUg4QjfYyI\nHH5zqAeH1Uzjpk3YQ8OI0X5WVeTT1j86dgMOBdRNrLhO/e+qiIjgWITb0jeClFDjVjbasjIn/SPB\nicskGGHwtI1Fh0WRY08zeo9aMpWFOeTZLZhMguVlLs52D+MPhqDldRUpJq/VmYJ9EWE51RtQIjFZ\n5B7JKhoqXAWALYW4WyO57xMOqHYdVefE+jvUzfLEMzOrgujzqAygkuVQsX7SXaOLX4+zZUDlnRsG\nOVYzFpMYWyh7oEV9L6Y4ubK7lJB2Hpl5yuFonxJrU6L89Xj9+IJhHFYze1v68QXjovecoskj997T\nqo2WiGUphIri+5vGetOdR9QNsP5qFUBkmHkt7gDVxbncfVU9d1+1lI9tr+HWlbnc6n2EO2q8hCU8\nureNJw+c53cHOyjPt/PBTZU4rGrwpiDHyrsuKePjl9Xiclh5/K3z7G9N8QVKSWDEgyXHNT4bwZZH\njtXMzZfkc+f2Gj6ybQkf2rqED29bQp7NzGP72sa65qhu6d7mfv7z5TP88Whn6nQ5gGNPIN/8Iadf\n+RWy9ySrFheyo6Cf5Xlenj3Wyeunesai0P6zaoDImpPS8hgNhDnYNojVLDjROcSgx6P2zytVF8k0\nxP2lk92YTIIdy9wpHy/KtakEgYnEc7ANI+inP2zHGDw/7oL0e7rZ0wUC+JNNlTjLI93bgSZWLc5H\nCDjWEYneB1qUGBdH7A2LXQ3Exolgc+8IDquZMpe62JaWORGCia2ZoXYlaFFxzytRkfQ0fffeYWXJ\nrCh3xbatWOQiZEhaTx9ROe1Lto8TlGS6PD7O9Y9iNQvO9gwjXRWqbRMJWH8T5BYzbC8HwBoc//6s\nZhNOf4fqPSQjJXQcUv6/qxxW3KDOh5bXp/W+ATj9XCRL5Nopd43aoCmXZRRmkGGEELii6ZD+IWWj\nxFsyUcrXKMFN1ROcjAkyZVr7VSB407pFBELqOo2RU6jakuqmN9KnPjN30viCe5kqc9zfpCyb08+r\nXmbFhpm19wKZ9+IOJEaSza9DKMDioUPcub2GTTVFnOn2srjQwfs3jgl7PAU5Vv60sYraEuXhP388\nSXRDfoIBP7a8/PEvbs1Td+mkQa98h5U/3VJNYa6NJw+c50THEAdaB3jwtSZefrubHJuZw+cGeeyt\nc4kRAkA4hOw5ycHgYp4v+CCua75E+RWfxGyxcIO7h7WVBew628fvD3eogeHuE0rglmxXJ1pShLGv\npZ9g2OCWDZWYhODQ6Sb1gKNAiWJE3AdGAuMHLlETgU53edlaVzzeK41gs5jId1gnrDEje09zumeU\n33tqOdrahadvbAKPPxji6OkmPKZ83r+xksJcm4qUHPnQ30S+w0p1US5Hz0cmi/WdUd3+guqxF3BV\nKPtCSqSUtPSOUOPOjc1JcNotVBQ4Jhb3qBVRUKV+C6Gi9/6mxAFgKdUM1o5DCYL7dueYJRNlcYED\nl8NC/+k9yvIoT8waScX+1gGsZsFly0rw+kP0mdxqLCGVJWCE1Y2uqJZRcx5mAaYUa/pazSaqB/eS\ne/K34z1qb6ea7xBtm3upEs2WSPGyqeg7q7JWai6dVt541JbJs6eYPGRSnjuAMzqRKWozxn/XUdzL\n1Pc0kwJr0V5risHU1r5R3E4bNe48VpS7eKuln+HoOFc0HTKVTRftNcePpUAkkrcpy6bpFZUnv/yG\nWVuPdUGIe4yRPlV3wu4Cz3lsvh6uWlHKn11exwc3VWG3TDwbzW4xc/P6xTTWFnGgdZAnD5wb8+wj\nE5hy8grGP9FkAmtuyunquTYLt22uYlG+g6cPtfP88S4KcqzctrmKj22v4YY1izg/MMoju1tjlkMo\nbNDa9DanOgbY56+kYUUdDUuKVN6zexnmriNct9LNZUvdvN3p5cHXTtF26iBh97KxyCHuZPcFw+xv\nHWB5mYsl7lxWVeTT3BaZIxAT9z7aB0b46c5mfrGrhYGRMYEOhQ1ePNFFcZ6NTUsmv3jdztQ1ZqSU\nnDx+gNPBYsqXrGAkEOaZ1/dyomOIsCF5Zt8Z/KPDbF65dGzmqxCq2z3QAobBqop8BkeDyubqOwOF\ntYm+bn6FmkTiG6THG8DrD7GkODehHcvKnHQP+VOPrwy0qGjdFpcNVVynyiB442aSNr+mAohjT8Gx\n30LIj5SSU11jlkwUIQQrypwEus8QKKidcjbkSCDEiY4hVi/O55JFLoSAs/7IOZfKmvGcVz2wwhoC\n0qpu8CnqjtstJhyhQYxwCE7+MbEX0HlYtatsVdwHdR2YbWrgbzLLwwir4+UUQvX2Sd9bFK8/RJ7d\njCVFymY0cgdiue4MtCiBdJaP39+ao7b3N0/8gr2nE4Mdv0dF00lpkGFDcm5glKoiZZdcWu8mbMDu\npqTMmlTJB72nIh570vVhMqveZfdxaNurIvZZXIt1YYl782tKbDd8SE1uOb8fUFG02TT13dJkElyx\nvJTrV5fT1DPC7w61EzaUJRMIS3JdE9ShsOVNmK7msJp5/8ZKttUV84GNldzeWEV1RHRWL87ng5sq\nGQmE+cXuFp7Yf47vv3SaN/ftpXckzLLlq7l0aZwNsmg9BEcRfafZVu/mY9uXUG/to7WrjyfanJwe\ntkei3TEr4a2WAQIhg6116mRurCnCGvTQMehTJ2xuMUH/KH946wy5NjOjQYOHd7fGxgr2tQwwMBLk\n6pWlmMP+SYt0lbkc9HoD/PeRjrGIB9j3djO9necoq13NuzatZn1NGYtEP08faudnO5vo6e6gvjSP\nisVViQcsqo2J67IyJzaLiVNNLeoCi44XRHEtVr+H2mnpU99FjTtJ3EuVZXKqOzn/3FB+e3J0WFTL\nkC/I71/dqWywnpPKWy5fo2Zbdh2FvQ/S19lKrzfRkomyyjWKKeyn2Sgb91gyh9oGCRmSDVWF5NpU\nT+PtYbs6l1MNqg40x7J7gmGDsM2ZUtytQmIPDROwONWNMZrBYRjQeVRFnPEesC0Xlr5L3TwGJhHO\ntt2q17ds4gHUZFQd99S9P0xmlb/OmLgb/c1qUttEdlZRrWpnqiwxn0dVGd3/kPobJiwY1unxEQgZ\nVBepc6Yoz8aqCheH2gZVDyIWuSf1oEJ+dQNKjtqjuJepfSw2qLsq9T4ZYuGI+3CvGrBYvEndRUtX\nQuehCyowtLaygGsuKeNM9zB/ONKBd0idEHkTirsTUnSHYw+Hhrgsp5nafDFuMLKqKJcPbanGaVdl\nc9csLuCq0mG2rF/LFasqE/cvrlc+cMchQBXsuqHMw8rKYry51Tx5oJ29Q4UEe5vAUClnb7X2s7TM\nSWnEey7Ks1HrDNI2LPFjwXAUc7LLCyO93LJhMR/aUo3dYuI3e9vY19LPm2d7WV7upCY3ALv/E976\nWazrnMzmmiI21xRxrH2IB19vYndTHwdaBzhxZD+lThsbGzYjTCYcxVVcUR5gW10xA6NBti9SN4Zx\nkU9hjfp9+nlsXQe5pNCgp+UYYSkJFtZx+Nwgj+xu4bG32hg0FykR9JynuXcEt9M2zkIqyLVS6rKP\nt2a8nUocknzdQ10hdvdasQ42ceDtMxx//iFCOaWw8kaV7dDwEQgHGd75I0pGTiVYMlFKgudwWE0c\nHC4cl70VT9iQHGwbpMadizsyOF1f6qRzKITPUZJ6MlN/k8r5tuYQCBtImytlXrU54MEkJH2lW9W1\ncepZ5R33n1VBSfm68ccuXaU+z55T4x8DdZOPDqKmyGWfCK8/lDpTBiIDqpHI3W7FEhwmONST2m+P\nUlSrbgiDKTJ8ojNEgyNqolhwdMI0yGjiRVXRWECwrd6NBHad6VMBnNkyflC197R6/WS/PYp7qbpm\nl16rbpqzyMIR96ZX1Mm4JNI9XNxwUdXZNlQXcvnyEk50DLH3mPLU8gsmsCXsKcQ9HFJR0YGH4Y37\n4O3/hpadKZ9elGfjY9tr+LPL63hXfR4lsh+zu378jibT2NRrv1d1mbtPUFx1CR+9bBnb690c9ZVw\nqKmT9tZTHGgdwB802F6XeCKvyDcYFk4OnxvkzS7B4GiQyysFZfkOivJsfGjLEsrzHbx0QnmuV9Y4\n1PsI+dTFcW5fyvdhs5i4ckUpH7+0hqqiHF492cPzx7tYbu2kfkkVIjfSC8mvwDTczWV1Bfz11ctY\nWxRSUZsj6eZpdyoRHemFE8/Q2P4wi3t3crjfzH/u6eOPRzvxBQ3OD/h4aHcbHUY+ocHznOsfjWXJ\nJLOszEn7oC9R4KPCUKgi91DY4NmjnTx7rBNb6VKuLA/wXtNOekZC/Ca4laGgsppajRKecdzEyeFc\n1gX2x1YRikf0N1OyqJoWr4mHdrVwfoLsqVNdXrz+EA3VY59BXYl6D+eNIjWoGu/9hwIqYo2k3gVC\nBsYE4s5oPyYhGLUUKctldABadylLxupIGXX2++HgSCGdZw8RSjEOQ/dxFTjNIKVPSjl+eb14kiL3\nfH+7WmRkMnEvqFLXfaqB755TSsTX3a4i9kO/UqmjJgvYE8fPWvtHKXXZyYn7DgtyrKytzOfIeQ+j\nQSN1OmTnEWUDpxoTANUjuuyzU2YRZYLp9aWyHW+3OtmWbB/zTAuqVZRyfv/Uo9OjAyqCcS1K8EW3\nlIQoOPkKvZ3HGLaVUOCc4M5ry1ORjGGAf1CJX8chFSk48qFmh+redh+HpddMPqAS9Q+j+bLJLNqg\nprB3HlETXwLDULoSs0lw6VI3tYVb6Pjd87y2ew8dxVuoL82jLKlcQKFpBGdhCW+c6SMQDHNzgZNl\nzrFubY7NzAc3VfL66V7KHAb5J36tXqfho2owsflVWLROCUMKivJs3NpQSXPvMC09Hja2D2JyN4y9\n7/xKdREPdWArrFY3jJyi1F3vuiuUgAz3kN/fRGB0L02WKurceayrKqCyMAfPaIhnDrezs8nGso4T\nhNxbqClO/V01VBdytmeYpw6e5/JlJWyuKUIMtKjXt7vo9Ph44XgX7YM+ttYVc2nRFkwHT7PEMQqX\nf5iDrXYefrMVu9VErzeAw2pmy/LLWD30skpzK4y7yMNBGGyjun4jHyis5NljnfxyTysN1YVctrQk\nYTbp/tZ+CnOtMUEHVW2zIMdKcyCfeiOkhMkV8Z4HW1WUG+ndBMMSbC4ItEdWPIq70Yz2YzYJBsxO\nNY5QuoLhk6/gC4ZxL2scNxYQNiRPH24HXxnDfSd4+vk9VFfVsqYyn4qCiH3TdUyN16TywiN0DfnI\ntVliaY/+kEEgZEwSuccPqFooHX4bX04O+ZO8BmaryvRJ9t2DPnXNVW5W19KqW+Do46pmUF5JwjUY\nChu0D4yyvnp8z3xtZQEHWgd5u3OIDTlFiZ57YFjZXNVbZ22QdCYsDHFvekV9yVVbx7YJARUNqgvq\n7VLd14k4/Gu1j9kC+VUqUhjtg84jLHfYGL3kOloty1Nm2gDKlpGG6voNNAOR6ceLG9TAn8kEHYVq\nEM5zbiwjIxX9Z9XAmGuCgZc8txL1joMQGFIXZlyXsMJdTOmqVdg7e2mXsK0uKXVRSvANUl+1jp09\nBuUFOdQVVo3LorCYTVxZ54ID/6Vufuv/VL3u0mtg7wMqVW7pNRO/D1SOeQ0dcD6cGB3mR/xxz3kl\nhiN96n1NhBDgLEU4S9lavhGBSIywcq3c3ljNIbGc4YNHcBUOUFm0IuWhHFYzt22u4r+PdPLKyR76\nvH6uG2jF66rnpQPnOdXlxWE18971FcpDDxcqMVi8kSVVa/jTcj9PHVQFut69ppwV5S6ssgpef0P1\nEuPFfbBNDd4V1VHrzuPOS2t47VQPb7UMcLTdQ47VjNVswmIStA/6uHplaYINJ4SgrjSPt5sKudIm\nMB98WJ3jlZuUJWMyxyLGQCgMDheEpepFOuIG/0cHEBYbPtTN+GjuFgbb3gQjTGF9NXFDqQC8frqH\nLo+fW7Zup+TAAUzWTnZ3FHPo3CDb6ou5tNKKGGxVN90JRM3jC/LIm624HBY+vG0Jdos5dR33hO/Z\nNDagKoco9J2jP/9dlE2RPkphDZx9WYltNLjrO6NuFCWR86DsEgi+W00iSrJk2gd9hAxJddH43PNS\np50Sp43jHR42FBWq61NK9b67jqnrfhpZUHPB/Bf34R6VCli7Y7yntWitmk59fr+qwpcK36AS9kVr\n1cyxgWZ1opgsUL0VseRSNlhzmDT2j3rFw91Qc5m6qTiS0iZLVqhjdh2bWNylVBdtUc3k+dAV6+HE\n71Xbi2rHJk5EsJQsZfXwa6y8tBKzPSm69g+BEaa0tIz3Lq6gsjAH86kJVv059qTqFa27TbUJVORY\nvhba9qjxjakWO4hWbozvWtvyIhUPz6kLcLRf3QynQa4t9SlrNgka1qxmqOd5qmpkyuqIUaxmEzet\nyKXSf4bmwy9xINDOkfzlDBaMsL3ezaaawrHMKrMFtv5l7LmlLjuf3FGXdES7mm7edUzZHtFIOEmA\n7RYz11yibgjH24cIhA2CYYNgWFJfmsfqxeNTbZeWONnf4qS15gPUDu1X53PLTjCZCToraOr1c6an\nD48vhHAWwAhq8DBB3PsJ2QoZDRn895EOjpz30bDkKlzDLfyh1Yy12BsbL2jtG2Fvcz/rqwpYWlkO\nHTVso4+GDfW8dKKbXWf6yOtsYr00EGWrJ/yM3zjdq9YXHg3y3LEubly7KHUd93hM5lhtG3vXQcxm\nQadzJSvjdpFSjp9EV1Srrtn+ZjVzFVROvy1X9RKjVG5S13gKv90kBJUpxF0IwSUV+bx6sgdvsQtn\nOKRunnaX6p1H69xkIVOKuxDix8D7gC4p5bhblFCf9LeBm1Cn1ieklKlN2UzQdUzdRePqcsSw5owN\nrNZfrUask4mmDVZvH/uSAiPqmNOdRVZcD5s+Ps7WScBiB3e9au/Sa1OL90ifujCXXDr565WuigyK\nBdX7G9eeOmh6FfNgc2KKG8TydEVOISvckeyOXHfEQw2NZT0M9yhvv/6q8Z5s3ZXqfZx9CVbfOnlb\ne0+rG4M5KVpzVShx9w2q6Cd3ksh9uuQU4XI6cZknmUk43AOHf40Y6aMBWJwvOTxcQ/3KBjYurUjo\nEcyI8jXqM+k7OzbA2H9W9VKSzruqotyEgbvJqCzKwWYxcdLnonbDHeBpZ+D4S3ScPsBBRyXn+tpx\nWM0sL3NySVkOHGW87z7aj2Ev4Fz/KOf6R9lWX8z2uhsJGZJT+9p45lA7H9xchTvPxh+OdFCUa+OK\n5ZFroWQ5nHkJe3iE69sGO80AABIzSURBVFeXYzYJuva8xcliJ8tzikgVt/d4/Rxt97BxSRE5VjOv\nneqhuigXiUqrnNSWkeFIgbmD+PLr6Q+r4MQwJG+c6eXguUGuW1XGsrK4zCRXhfqMByLiHg6piqGl\nq8ZfZ+Xjb0it/SOU59snTJVeucjFqyd7ODVkpgFUTzYUUIPc05i4NVdMJ3J/ELgX+OkEj98ILI/8\nbAPui/yeHXpOqLuzfXymAqCskc4jammuVN573xkVZeeVjG2b6ah2tMLfVJSthu63lV8ajYTjiS5Q\nMZHfHsXqgJKVSkzcKSJe12J1M+k7m0LcI8IXP3iZ645M7ugbs6/aD6hu8qIUA0GOfKjeosrUVm0Z\ns1mSGe5RUXnVlvGP5Veq9kffczrEXQjVlslmLLbvVzfQZddCQTVlznKumarbPx2K69X30nVEiXtg\nRPUIL7KGiNkkqHXncbbHS/dQITvPSk4PbsJVvpbllaVcWuqisjBHTdaKLk0YL+6GAb4BTHlLcZjN\n3LCmnPpSda3YTIJbGyp5ZHcLT+4/T5nLzkggzIe2LB4bDyhZocZZek4iKjdxTY2VpiND7PLVc+5E\n9zgrCeC1Uz3YLCa21hbjsJpo6x/hxRNdLHHnYhKCvAl6X5gi2TLdxyE4iq9Mrdo05AvyzOEOzvWP\n4rRbeOpgO+9aGWZD1CM3maCwBl/XafaZe1iX04srFBizZCbBHwrTMeinsXbiORz5DitVRTkc6/ey\nAYnwDSjdEGL89ZVFTHlWSylfBiarLHUr8FOpeAMoFELMTqb+SJ+yDVJFr1EKqpVwtB8c/5gRVuJS\nXD87AyLFS1UE23Us9eP9Z1V3ejorxCy7Fho+nPpGZDKpG0R08YB4ojPs4rvtUWGN+u5GWHU5S5ZN\nfNNccql67aNPRLrETaonYYRVlsKRx1VNb5M5dQ5w9IYQLfw03YUTpqKgWp0TE62B2ndW2WLVW9WE\nknQIO6j3WbpK2QGhgIoipRyrUXMR1JXkMewP8/M3mmnrH+GypW7uvHIVV60sp7p4bBYuFrs6v+LF\nPaBsuIaVdXxyR21M2KPk2Mx8YFMVZpOqx3PZUnfiAHyuW52Pkdx40XWMWncu1Ss3s791gGcOdySU\nEm7rH+FM9zBbaovJsZkRQvCetYtwWM2c6R4mz25OWMksARHJljm/D3KLMRfX0j8c4KFdLXQP+WP1\noOpK8nj+eBevRUpw+IJhDngLOHCyif0nm9n31m4MkzV1AJXE+QEfhpSx/PaJWFWRT2fQgddvqICl\n84j6bu3j5zbEI6WkfXB08oquGSIdnnslEJ9k2hbZNs7EFUJ8CvgUwJIlk6Q3TZdoMfzJ7tBCqMyO\nMy+Orykx2KYuxOIU4pMJLDbVze0+rirnxVs4RlgJQtnq6d1obHmJsymTKa5TYxHDPYme4OiAEuz4\nSSfRm0m0xkzPSZXpU9EwyXuxwyXvU4PZza9D02vq/ZitKoK05qieUsX61L68s1ztP3hO3STSVUjJ\nvTRyszmrvvd4/N7Eqfbppnw1nH9LCXy0kuFEA+MzoL40j8WFDioLc2msLZp4YF8IleIXP5Epkt1h\nzS3COsHzCnKs/MmmKpp6h9lYnRRYCKFu8uf2KT+86yiioJrL1tRhd/bz6qke+kcC3LxhMS67hVdP\n9uByWBJSOnNtFm5ct4hH97ZNbMmAOh98HnWdLrsWV8hGMCwpyLXw3nUVsRW/bl6/mOePd/Hm2T66\nhnyqTr/XxXV5Nq4pH6Hz4Nuczl3G8mQrMAWtfSOYTYKKwtSZX1GWlTl54biVjlErrs6jKkiqu3LS\n53h8QZ4/1sXZnmGV31HgYFmZi2VlTgqSCxBmgFkdUJVS3g/cD9DY2Hjxa371nFADfFMN6i1aq/zh\njoPKe4/SdzpSpnPqO3zaKFut8t/7mxIj2ugsu6ksmeniXg7iD8qOihd33+D4fHKLTVktUXFvPxCp\nujdF1Oleqn5C/ki54WZlR5SuVL2hyabbmy3KAvK0T39FnOngLFc3vd7T48V9urbXhVJQrT7HziPK\n4kquZHiBOKxm7tgyzWDInpTrPsGMzGTcTnts8tQ4SlZA625Vg3y4B1a8GyEEjbXFuJ12nj7Uzi92\ntbC2soD2QR/Xry4fN6BdVZTLe9dVpFxMJIYwq+wikwUWrWOttGKzmFizOD/heCaT4NpVZeTZLbxx\nppfq4lyuaNhA+eE3IXgMc06INz1uCjy+cWnA8fQPBzja7qGyMGfSAXhQ30F9qZO2bhtLR/owWWwT\nBpVSqglpr0Z6FlcsLyFsSE52eXn57W5efrubLbXFXL68JOXz00U6xP0cEJ/BXxXZlln8Qyrqm+Lu\nCagTvrhe1Z2pvXLsgus7o7rolglO6kxQVKder+tYorj3N6koqTBNNxq7U4lL1zGovWKsN+AbSD3h\nIlpAzDeoot6ay6YvTBb7mNDPhPxKJe7p8NujCKG+696Tym+Ofw/9Z1UPIboIRroRQt28W3cpSyY+\nNXe2cOQnFtIaHYjULk9R9G665Fepz635dTUOE1e2uK4kjw9vXcKT+8/x5tk+3E4bqytSv9byFOUZ\nEogGA2WrwJpDLiT0AOIRQs3rWF9VQG7E/qGoBjqPUlvq4hVRxx+OdPDhrUtS1rEZCYR4fL+SqWtX\nTV0aAuCSCheHjzoZHPVQVL8uZYJGr9fPc8e7ONc/ypLiXK5bXR6L0rfVuxkYCXCqyztWQ+n/tXd2\nsXFUVxz/HXv9kTjGjj/ihATjALHzUcAJNAktQYEKBBSBVPWBqFJRRUUfqARSpQpUqVUf+9KWh6pV\n1Za+oLQq/YrSlpZSpLa0Ck0gQIibOCGmJCE4jmOCwXbs+PTh3sFjZ21vnLHn7uz5SaOdmV17/5q7\nc+bec849dx5JwuG4C/iiOLYC76tqgUu2XwZR+dKZ/O1xll/vHgjRTLbhc843O7VGyXxTnnNP/L5D\nLqp//kOXqnnqDd/rTHCK8rJ1Pi7h1yAdv+CuQb6RzuJG19uMYhP5AqlJE/ndk/K3RzRe61xD8Xos\n8TTT+YyvtGyYiHPM1whhJqpq3W8qKhExdDZv7fJLoqzMXVMdd9dvijuwoaaSBze30tlaz53rW6b3\nqc9GtI7qyjyZb9NQU5WbCOj6613RcDW3X381fYPn+fdbF5ezHr0wzq79JxkcHuOBzitdFdICaGus\nQavr3Mpv8bVu/f/815E+ntnzP/oGR7hzfQuf27TyIvdL/eJKbm5r+Li+1HxSSCrkTmA70CQix4Fv\nARUAqvoj4I+4NMgjuFTIL82X2EmcPuwMUk2BQ5vGNS6b4dQb7ofa78t0LpS/Pc6ydU7Hvqddb1nV\nBTjbtiX7PU0druxB70Hnvhp+f+K7prKowbmFTux1N8lsrq4kqG91xmim6eVzYWmb62GeOToxp+Cj\nM87nPt8GN8p7HhtJ/qFVCFW1ro2jiUxDZ5Npy6Z2N/KdJre9uqKc2zsK6wFPS3O7n0g4TfbVbCxt\n84HttaxuquH6lXXse/ss1RXltDYsptm7nf504BSnzg1z3w0rJmbcFkB5mdBwzUZe7x7nzZ4Klp3t\nczWRgH90n2bgo1HWrajltvbmaedjLCSzKlDVHbO8r8CjiSkqhNEhV4mt9RIyLstzsGyD8yePDuVP\ngVwolrZ5P7i40gRN7c4oJN2jrFzsAqu9XS7WkC9TJiIyRKPDC7aYAFW18KmvJv9/KxY5A9Hv8/Rh\nYkWlBLJXZmXdAzA+ms6U9Mj9MvKB2x8eSObh2dTuFvPOl3qbFPWtl6e1ug62fOXja7CtvYnTgyP8\ns7sPiNYcyNE3eJ7tHc2Tc+UL5KaO1bxU7spU9BzrnxikLXZB6dbG+e+RF0r6j5e50NfthohNBbpk\nIlbcACf2uYDX2Z7CM1OSpqwcPvnlhfmu5rVw5g8uYJsvxz0i8ntXLJrfG3ihaLzW5WePDPoVlXpc\nUHEhRiRpzliMUvNGPnDVEMfOF5ZaOxsihbtA0yTWcanKlbNjcysfDI9ycmCYkwNDnHx/iC2rG9g4\ny9oE01FTleOuDS5mc35snL7BET4cGWN1U03+GvUpUqTG/bDrdV9qYGxJi7vxjv19YVMg06SpHcqe\nc7338pxzV+QLrlXVum35JwquzR00Dd6497/l/KMDbwdbAyRR4sY9ypTJ9zAvIWqrK+hYXkHH8kvv\nqc9EZa6MK+vnfy3UuRLWo6YQxs77Kd4dl97rFnFVFcdGFj4FMi0qql3Q+HSXy5yoviJ/cE0ENj/i\nsomywJJlrsfef3RixaKGBXDJpE2u2s01GD5XcBqkkU2Kz7j3H3W5sM2zTy3OS8t613td6BTINFm2\n3rknzhzJ72+PyFUmN2MzbaKUyP5j7jcjknzgNkTiE5mGBtzxTG1uZJbiG3/XrYI1d7nc27lQWQPr\n7iut3kzjdc7VcmG0tIboDde61M4TrzgXXlKzYEMnmsgULUqRBTebcckUXzetqhZW3XR5PcyWDXNP\ntypGcpUTQdJS6sVFKZFjIwuTJRMKkXFPKg3SKEqKz7gbcyPKTy6lm72ieiLPPY0JRWlRVevy3If6\nS2uEakzCxmulQuN1sPazBZVBzRQtG5z/+YoCSjJnhWgi0+iwGfcSxox7qVBWlsoivalzZafbSom4\n662UYizGJMwtYxhZI15j3HruJYsZd8PIGpOMu/XcSxVzyxhG1shVu/TH8srSmcthXIQZd8PIGtFE\nplLJ6zfyYsbdMLJI2zbXczdKFjPuhpFFWvLXXTdKBwuoGoZhZBAz7oZhGBnEjLthGEYGMeNuGIaR\nQcy4G4ZhZBAz7oZhGBnEjLthGEYGMeNuGIaRQURV0/likdPA23P88yagL0E580Ux6DSNyWAak8E0\nzs7Vqto824dSM+6Xg4jsVdWb09YxG8Wg0zQmg2lMBtOYHOaWMQzDyCBm3A3DMDJIsRr3H6ctoECK\nQadpTAbTmAymMSGK0uduGIZhzEyx9twNwzCMGTDjbhiGkUGKzriLyN0ickhEjojIE2nrARCRn4lI\nr4gciJ1rEJHnRaTbv6a6DL2IXCUiL4rIQRF5U0QeC02niFSLyMsi8prX+G1/frWI7PFt/ksRSX2J\nIREpF5FXRWR3wBp7ROQNEdkvInv9uWDa2+upF5FnReS/ItIlIreEpFFEOvz1i7ZzIvJ4SBqno6iM\nu4iUAz8A7gHWAztEJIQlZ34O3D3l3BPAC6q6BnjBH6fJGPA1VV0PbAUe9dcuJJ0jwB2qeiPQCdwt\nIluB7wDfU9XrgLPAwylqjHgM6Iodh6gR4HZV7YzlZYfU3gBPAc+p6lrgRtw1DUajqh7y168TuAn4\nCPhtSBqnRVWLZgNuAf4cO34SeDJtXV5LG3AgdnwIWOH3VwCH0tY4Re/vgTtD1QksBl4BtuBmA+by\n/QZS0rYKd0PfAewGJDSNXkcP0DTlXDDtDdQBx/CJHSFqnKLrLuClkDXGt6LquQMrgXdix8f9uRBp\nUdV3/f4poCVNMXFEpA3YCOwhMJ3e3bEf6AWeB44CA6o65j8SQpt/H/g6MO6PGwlPI4ACfxGRfSLy\niD8XUnuvBk4DT3sX109EpIawNMZ5ENjp90PV+DHFZtyLEnWP9yByTkVkCfBr4HFVPRd/LwSdqnpB\n3RB4FbAZWJumnqmIyH1Ar6ruS1tLAdyqqptwbsxHReS2+JsBtHcO2AT8UFU3Ah8yxb0RgEYAfAzl\nfuBXU98LReNUis24nwCuih2v8udC5D0RWQHgX3tT1oOIVOAM+zOq+ht/OjidAKo6ALyIc3HUi0jO\nv5V2m38auF9EeoBf4FwzTxGWRgBU9YR/7cX5iTcTVnsfB46r6h5//CzO2IekMeIe4BVVfc8fh6hx\nEsVm3P8DrPGZCZW4YdKulDVNxy7gIb//EM7HnRoiIsBPgS5V/W7srWB0ikiziNT7/UW4mEAXzsh/\n3n8sVY2q+qSqrlLVNtzv72+q+gUC0gggIjUiUhvt4/zFBwiovVX1FPCOiHT4U58BDhKQxhg7mHDJ\nQJgaJ5O2038OQY17gcM4X+w30tbjNe0E3gVGcb2Rh3F+2BeAbuCvQEPKGm/FDR1fB/b77d6QdAI3\nAK96jQeAb/rz1wAvA0dww+KqtNvc69oO7A5Ro9fzmt/ejO6VkNrb6+kE9vo2/x2wNECNNcAZoC52\nLiiN+TYrP2AYhpFBis0tYxiGYRSAGXfDMIwMYsbdMAwjg5hxNwzDyCBm3A3DMDKIGXfDMIwMYsbd\nMAwjg/wfzQ7ME4r071cAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nGyfOOPDUeqd",
        "colab_type": "text"
      },
      "source": [
        "# Generating Samples from Training\n",
        "view the sample images generated during training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jM0Es0whVsCw",
        "colab_type": "code",
        "colab": {}
      },
      "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)))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NnHlv_7ZVzzG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load samples from generator, taken while training\n",
        "with open('train_samples.pkl', 'rb') as f:\n",
        "    samples = pkl.load(f)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pqTEJsCnV274",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 245
        },
        "outputId": "c8bf5a6a-74b9-461f-af77-e79e384226b0"
      },
      "source": [
        "_ = view_samples(-1, samples)"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4sAAADjCAYAAAAsXIHJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 1152x288 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j7uq2KuElMqd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# END"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}