dlnd_face_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Face Generation\n",
    "\n",
    "In this project, you'll define and train a DCGAN on a dataset of faces. Your goal is to get a generator network to generate *new* images of faces that look as realistic as possible!\n",
    "\n",
    "The project will be broken down into a series of tasks from **loading in data to defining and training adversarial networks**. At the end of the notebook, you'll be able to visualize the results of your trained Generator to see how it performs; your generated samples should look like fairly realistic faces with small amounts of noise.\n",
    "\n",
    "### Get the Data\n",
    "\n",
    "You'll be using the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train your adversarial networks.\n",
    "\n",
    "This dataset is more complex than the number datasets (like MNIST or SVHN) you've been working with, and so, you should prepare to define deeper networks and train them for a longer time to get good results. It is suggested that you utilize a GPU for training.\n",
    "\n",
    "### Pre-processed Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. Some sample data is show below.\n",
    "\n",
    "<img src='assets/processed_face_data.png' width=60% />\n",
    "\n",
    "> If you are working locally, you can download this data [by clicking here](https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip)\n",
    "\n",
    "This is a zip file that you'll need to extract in the home directory of this notebook for further loading and processing. After extracting the data, you should be left with a directory of data `processed_celeba_small/`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!unzip processed_celeba_small.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = 'processed_celeba_small/'\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "import pickle as pkl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import problem_unittests as tests\n",
    "#import helper\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize the CelebA Data\n",
    "\n",
    "The [CelebA](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations, you'll only need the images. Note that these are color images with [3 color channels (RGB)](https://en.wikipedia.org/wiki/Channel_(digital_image)#RGB_Images) each.\n",
    "\n",
    "### Pre-process and Load the Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. This *pre-processed* dataset is a smaller subset of the very large CelebA data.\n",
    "\n",
    "> There are a few other steps that you'll need to **transform** this data and create a **DataLoader**.\n",
    "\n",
    "#### Exercise: Complete the following `get_dataloader` function, such that it satisfies these requirements:\n",
    "\n",
    "* Your images should be square, Tensor images of size `image_size x image_size` in the x and y dimension.\n",
    "* Your function should return a DataLoader that shuffles and batches these Tensor images.\n",
    "\n",
    "#### ImageFolder\n",
    "\n",
    "To create a dataset given a directory of images, it's recommended that you use PyTorch's [ImageFolder](https://pytorch.org/docs/stable/torchvision/datasets.html#imagefolder) wrapper, with a root directory `processed_celeba_small/` and data transformation passed in."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# necessary imports\n",
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision import transforms\n",
    "from torch.utils.data import DataLoader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dataloader(batch_size, image_size, data_dir='processed_celeba_small/'):\n",
    "    \"\"\"\n",
    "    Batch the neural network data using DataLoader\n",
    "    :param batch_size: The size of each batch; the number of images in a batch\n",
    "    :param img_size: The square size of the image data (x, y)\n",
    "    :param data_dir: Directory where image data is located\n",
    "    :return: DataLoader with batched data\n",
    "    \"\"\"\n",
    "    \n",
    "    # resize and normalize the images\n",
    "    transform = transforms.Compose([transforms.Resize(image_size), \n",
    "                                    transforms.ToTensor()])\n",
    "\n",
    "    # get training and test directories\n",
    "    image_path = './' + data_dir\n",
    "    \n",
    "\n",
    "    # define datasets using ImageFolder\n",
    "    train_dataset = datasets.ImageFolder(image_path, transform)\n",
    "\n",
    "    # create and return DataLoaders\n",
    "    train_loader = DataLoader(dataset=train_dataset, batch_size=batch_size, shuffle=True, num_workers=0)\n",
    "\n",
    "    return train_loader\n",
    "    \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a DataLoader\n",
    "\n",
    "#### Exercise: Create a DataLoader `celeba_train_loader` with appropriate hyperparameters.\n",
    "\n",
    "Call the above function and create a dataloader to view images. \n",
    "* You can decide on any reasonable `batch_size` parameter\n",
    "* Your `image_size` **must be** `32`. Resizing the data to a smaller size will make for faster training, while still creating convincing images of faces!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define function hyperparameters\n",
    "batch_size = 32\n",
    "img_size = 32\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# Call your function and get a dataloader\n",
    "celeba_train_loader = get_dataloader(batch_size, img_size)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, you can view some images! You should seen square images of somewhat-centered faces.\n",
    "\n",
    "Note: You'll need to convert the Tensor images into a NumPy type and transpose the dimensions to correctly display an image, suggested `imshow` code is below, but it may not be perfect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7efc8bbd8d68>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper display function\n",
    "def imshow(img):\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# obtain one batch of training images\n",
    "dataiter = iter(celeba_train_loader)\n",
    "images, _ = dataiter.next() # _ for no labels\n",
    "\n",
    "# plot the images in the batch, along with the corresponding labels\n",
    "fig = plt.figure(figsize=(20, 4))\n",
    "plot_size=20\n",
    "for idx in np.arange(plot_size):\n",
    "    ax = fig.add_subplot(2, plot_size/2, idx+1, xticks=[], yticks=[])\n",
    "    imshow(images[idx])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Pre-process your image data and scale it to a pixel range of -1 to 1\n",
    "\n",
    "You need to do a bit of pre-processing; you know that the output of a `tanh` activated generator will contain pixel values in a range from -1 to 1, and so, we need to rescale our training images to a range of -1 to 1. (Right now, they are in a range from 0-1.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Complete the scale function\n",
    "def scale(x, feature_range=(-1, 1)):\n",
    "    ''' Scale takes in an image x and returns that image, scaled\n",
    "       with a feature_range of pixel values from -1 to 1. \n",
    "       This function assumes that the input x is already scaled from 0-1.'''\n",
    "    # assume x is scaled to (0, 1)\n",
    "    # scale to feature_range and return scaled x\n",
    "    min, max = feature_range\n",
    "    x = x * (max - min) + min\n",
    "    \n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.8118)\n",
      "Max:  tensor(0.8745)\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# check scaled range\n",
    "# should be close to -1 to 1\n",
    "img = images[0]\n",
    "scaled_img = scale(img)\n",
    "\n",
    "print('Min: ', scaled_img.min())\n",
    "print('Max: ', scaled_img.max())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Define the Model\n",
    "\n",
    "A GAN is comprised of two adversarial networks, a discriminator and a generator.\n",
    "\n",
    "## Discriminator\n",
    "\n",
    "Your first task will be to define the discriminator. This is a convolutional classifier like you've built before, only without any maxpooling layers. To deal with this complex data, it's suggested you use a deep network with **normalization**. You are also allowed to create any helper functions that may be useful.\n",
    "\n",
    "#### Exercise: Complete the Discriminator class\n",
    "* The inputs to the discriminator are 32x32x3 tensor images\n",
    "* The output should be a single value that will indicate whether a given image is real or fake\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper conv function\n",
    "def conv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
    "    \"\"\"Creates a convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    layers = []\n",
    "    conv_layer = nn.Conv2d(in_channels=in_channels, out_channels=out_channels, \n",
    "                           kernel_size=kernel_size, stride=stride, padding=padding, bias=False)\n",
    "    \n",
    "    layers.append(conv_layer)\n",
    "\n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "    return nn.Sequential(*layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Discriminator(nn.Module):\n",
    "\n",
    "    def __init__(self, conv_dim):\n",
    "        \"\"\"\n",
    "        Initialize the Discriminator Module\n",
    "        :param conv_dim: The depth of the first convolutional layer\n",
    "        \"\"\"\n",
    "        super(Discriminator, self).__init__()\n",
    "        self.conv_dim = conv_dim\n",
    "\n",
    "        # Define all convolutional layers\n",
    "        # Should accept an RGB image as input and output a single value\n",
    "\n",
    "        # Convolutional layers, increasing in depth\n",
    "        # first layer has *no* batchnorm\n",
    "        self.conv1 = conv(3, conv_dim, 4, batch_norm=False) # 3 channels input, x, y = 32, conv_dim=depth= 32\n",
    "        self.conv2 = conv(conv_dim, conv_dim*2, 4) # \n",
    "        self.conv3 = conv(conv_dim*2, conv_dim*4, 4) # \n",
    "        self.conv4 = conv(conv_dim*4, conv_dim*8, 4) # \n",
    "        \n",
    "        # Classification layer\n",
    "        #self.conv5 = conv(conv_dim*8, 1, 4, stride=1, batch_norm=False)\n",
    "        self.fc1 = nn.Linear(conv_dim*8*2*2,1024)\n",
    "        self.fc2 = nn.Linear(1024,512)\n",
    "        self.fc3 = nn.Linear(512,1)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: Discriminator logits; the output of the neural network\n",
    "        \"\"\"\n",
    "        # define feedforward behavior\n",
    "        # relu applied to all conv layers but last\n",
    "        x = F.leaky_relu(self.conv1(x),0.2)\n",
    "        x = F.leaky_relu(self.conv2(x),0.2)\n",
    "        x = F.leaky_relu(self.conv3(x),0.2)\n",
    "        x = F.relu(self.conv4(x))\n",
    "        # last, classification layer\n",
    "        #x = self.conv5(x))\n",
    "        x = x.view(-1, self.conv_dim*8*2*2)\n",
    "        x = self.fc1(x)\n",
    "        x= self.fc2(x)\n",
    "        x = self.fc3(x)\n",
    "        \n",
    "\n",
    "        return x\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_discriminator(Discriminator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generator\n",
    "\n",
    "The generator should upsample an input and generate a *new* image of the same size as our training data `32x32x3`. This should be mostly transpose convolutional layers with normalization applied to the outputs.\n",
    "\n",
    "#### Exercise: Complete the Generator class\n",
    "* The inputs to the generator are vectors of some length `z_size`\n",
    "* The output should be a image of shape `32x32x3`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper deconv function\n",
    "def deconv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
    "    \"\"\"Creates a transpose convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    layers = []\n",
    "    # append transpose conv layer\n",
    "    layers.append(nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, padding, bias=False))\n",
    "    # optional batch norm layer\n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "    return nn.Sequential(*layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Generator(nn.Module):\n",
    "    \n",
    "    def __init__(self, z_size, conv_dim):\n",
    "        \"\"\"\n",
    "        Initialize the Generator Module\n",
    "        :param z_size: The length of the input latent vector, z\n",
    "        :param conv_dim: The depth of the inputs to the *last* transpose convolutional layer\n",
    "        \"\"\"\n",
    "        super(Generator, self).__init__()\n",
    "        self.conv_dim = conv_dim\n",
    "        \n",
    "        # first, fully-connected layer\n",
    "        self.fc = nn.Linear(z_size, conv_dim*8*4*4)\n",
    "\n",
    "        # transpose conv layers\n",
    "        self.t_conv1 = deconv(conv_dim*8, conv_dim*4, 4)\n",
    "        self.t_conv2 = deconv(conv_dim*4, conv_dim*2, 4)\n",
    "        self.t_conv3 = deconv(conv_dim*2, 3, 4, batch_norm=False)\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: A 32x32x3 Tensor image as output\n",
    "        \"\"\"\n",
    "        # define feedforward behavior\n",
    "        x = self.fc(x)\n",
    "        x = x.view(-1, self.conv_dim*8, 4, 4) # (batch_size, depth, 4, 4)\n",
    "        \n",
    "        # hidden transpose conv layers + relu\n",
    "        x = F.relu(self.t_conv1(x))\n",
    "        x = F.relu(self.t_conv2(x))\n",
    "        \n",
    "        # last layer + tanh activation\n",
    "        x = self.t_conv3(x)\n",
    "        x = F.tanh(x)\n",
    "        \n",
    "        return x\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_generator(Generator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize the weights of your networks\n",
    "\n",
    "To help your models converge, you should initialize the weights of the convolutional and linear layers in your model. From reading the [original DCGAN paper](https://arxiv.org/pdf/1511.06434.pdf), they say:\n",
    "> All weights were initialized from a zero-centered Normal distribution with standard deviation 0.02.\n",
    "\n",
    "So, your next task will be to define a weight initialization function that does just this!\n",
    "\n",
    "You can refer back to the lesson on weight initialization or even consult existing model code, such as that from [the `networks.py` file in CycleGAN Github repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/blob/master/models/networks.py) to help you complete this function.\n",
    "\n",
    "#### Exercise: Complete the weight initialization function\n",
    "\n",
    "* This should initialize only **convolutional** and **linear** layers\n",
    "* Initialize the weights to a normal distribution, centered around 0, with a standard deviation of 0.02.\n",
    "* The bias terms, if they exist, may be left alone or set to 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "def weights_init_normal(m):\n",
    "    \"\"\"\n",
    "    Applies initial weights to certain layers in a model .\n",
    "    The weights are taken from a normal distribution \n",
    "    with mean = 0, std dev = 0.02.\n",
    "    :param m: A module or layer in a network    \n",
    "    \"\"\"\n",
    "    # classname will be something like:\n",
    "    # `Conv`, `BatchNorm2d`, `Linear`, etc.\n",
    "    classname = m.__class__.__name__\n",
    "    \n",
    "    # TODO: Apply initial weights to convolutional and linear layers\n",
    "    if hasattr(m, 'weight') and (classname.find('Conv') != -1 or classname.find('Linear') != -1):\n",
    "        nn.init.normal(m.weight.data, 0.0, 0.02)\n",
    "    \n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Build complete network\n",
    "\n",
    "Define your models' hyperparameters and instantiate the discriminator and generator from the classes defined above. Make sure you've passed in the correct input arguments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "def build_network(d_conv_dim, g_conv_dim, z_size):\n",
    "    # define discriminator and generator\n",
    "    D = Discriminator(d_conv_dim)\n",
    "    G = Generator(z_size=z_size, conv_dim=g_conv_dim)\n",
    "\n",
    "    # initialize model weights\n",
    "    D.apply(weights_init_normal)\n",
    "    G.apply(weights_init_normal)\n",
    "\n",
    "    print(D)\n",
    "    print()\n",
    "    print(G)\n",
    "    \n",
    "    return D, G\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Define model hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Discriminator(\n",
      "  (conv1): Sequential(\n",
      "    (0): Conv2d(3, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (conv2): 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",
      "  (conv3): Sequential(\n",
      "    (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv4): Sequential(\n",
      "    (0): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (fc1): Linear(in_features=2048, out_features=1024, bias=True)\n",
      "  (fc2): Linear(in_features=1024, out_features=512, bias=True)\n",
      "  (fc3): Linear(in_features=512, out_features=1, bias=True)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fc): Linear(in_features=100, out_features=16384, bias=True)\n",
      "  (t_conv1): Sequential(\n",
      "    (0): ConvTranspose2d(1024, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv2): Sequential(\n",
      "    (0): ConvTranspose2d(512, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv3): Sequential(\n",
      "    (0): ConvTranspose2d(256, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      ")\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py:14: UserWarning: nn.init.normal is now deprecated in favor of nn.init.normal_.\n",
      "  \n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 64\n",
    "g_conv_dim = 128\n",
    "z_size = 100\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "D, G = build_network(d_conv_dim, g_conv_dim, z_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training on GPU\n",
    "\n",
    "Check if you can train on GPU. Here, we'll set this as a boolean variable `train_on_gpu`. Later, you'll be responsible for making sure that \n",
    ">* Models,\n",
    "* Model inputs, and\n",
    "* Loss function arguments\n",
    "\n",
    "Are moved to GPU, where appropriate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training on GPU!\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "import torch\n",
    "\n",
    "# Check for a GPU\n",
    "train_on_gpu = torch.cuda.is_available()\n",
    "if not train_on_gpu:\n",
    "    print('No GPU found. Please use a GPU to train your neural network.')\n",
    "else:\n",
    "    print('Training on GPU!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Discriminator and Generator Losses\n",
    "\n",
    "Now we need to calculate the losses for both types of adversarial networks.\n",
    "\n",
    "### Discriminator Losses\n",
    "\n",
    "> * For the discriminator, the total loss is the sum of the losses for real and fake images, `d_loss = d_real_loss + d_fake_loss`. \n",
    "* Remember that we want the discriminator to output 1 for real images and 0 for fake images, so we need to set up the losses to reflect that.\n",
    "\n",
    "\n",
    "### Generator Loss\n",
    "\n",
    "The generator loss will look similar only with flipped labels. The generator's goal is to get the discriminator to *think* its generated images are *real*.\n",
    "\n",
    "#### Exercise: Complete real and fake loss functions\n",
    "\n",
    "**You may choose to use either cross entropy or a least squares error loss to complete the following `real_loss` and `fake_loss` functions.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def real_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being real.\n",
    "       param, D_out: discriminator logits\n",
    "       return: real loss'''\n",
    "    labels = torch.ones(D_out.size(0))\n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "    # binary cross entropy with logits loss\n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    # calculate loss\n",
    "    loss = criterion(D_out.squeeze(), labels)\n",
    "    \n",
    "    return loss\n",
    "    \n",
    "\n",
    "def fake_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being fake.\n",
    "       param, D_out: discriminator logits\n",
    "       return: fake loss'''\n",
    "    batch_size = D_out.size(0)\n",
    "    labels = torch.zeros(batch_size) # fake labels = 0\n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    # calculate loss\n",
    "    loss = criterion(D_out.squeeze(), labels)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizers\n",
    "\n",
    "#### Exercise: Define optimizers for your Discriminator (D) and Generator (G)\n",
    "\n",
    "Define optimizers for your models with appropriate hyperparameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "# params\n",
    "lr = 0.0002\n",
    "beta1=0.5\n",
    "beta2=0.999 # default value\n",
    "\n",
    "# Create optimizers for the discriminator and generator\n",
    "d_optimizer = optim.Adam(D.parameters(), lr, [beta1, beta2])\n",
    "g_optimizer = optim.Adam(G.parameters(), lr, [beta1, beta2])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Training\n",
    "\n",
    "Training will involve alternating between training the discriminator and the generator. You'll use your functions `real_loss` and `fake_loss` to help you calculate the discriminator losses.\n",
    "\n",
    "* You should train the discriminator by alternating on real and fake images\n",
    "* Then the generator, which tries to trick the discriminator and should have an opposing loss function\n",
    "\n",
    "\n",
    "#### Saving Samples\n",
    "\n",
    "You've been given some code to print out some loss statistics and save some generated \"fake\" samples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Complete the training function\n",
    "\n",
    "Keep in mind that, if you've moved your models to GPU, you'll also have to move any model inputs to GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(D, G, n_epochs, print_every=50):\n",
    "    '''Trains adversarial networks for some number of epochs\n",
    "       param, D: the discriminator network\n",
    "       param, G: the generator network\n",
    "       param, n_epochs: number of epochs to train for\n",
    "       param, print_every: when to print and record the models' losses\n",
    "       return: D and G losses'''\n",
    "    \n",
    "    # move models to GPU\n",
    "    if train_on_gpu:\n",
    "        D.cuda()\n",
    "        G.cuda()\n",
    "\n",
    "    # keep track of loss and generated, \"fake\" samples\n",
    "    samples = []\n",
    "    losses = []\n",
    "\n",
    "    # Get some fixed data for sampling. These are images that are held\n",
    "    # constant throughout training, and allow us to inspect the model's performance\n",
    "    sample_size=16\n",
    "    fixed_z = np.random.uniform(-1, 1, size=(sample_size, z_size))\n",
    "    fixed_z = torch.from_numpy(fixed_z).float()\n",
    "    # move z to GPU if available\n",
    "    if train_on_gpu:\n",
    "        fixed_z = fixed_z.cuda()\n",
    "\n",
    "    # epoch training loop\n",
    "    for epoch in range(n_epochs):\n",
    "\n",
    "        # batch training loop\n",
    "        for batch_i, (real_images, _) in enumerate(celeba_train_loader):\n",
    "\n",
    "            batch_size = real_images.size(0)\n",
    "            real_images = scale(real_images)\n",
    "\n",
    "            # ===============================================\n",
    "            #         YOUR CODE HERE: TRAIN THE NETWORKS\n",
    "            # ===============================================\n",
    "            \n",
    "            # 1. Train the discriminator on real and fake images\n",
    "            d_optimizer.zero_grad()\n",
    "            # Compute the discriminator losses on real images \n",
    "            if train_on_gpu:\n",
    "                real_images = real_images.cuda()\n",
    "        \n",
    "            D_real = D(real_images)\n",
    "            d_real_loss = real_loss(D_real)\n",
    "        \n",
    "            # 2. Train with fake images\n",
    "        \n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            # move x to GPU, if available\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "            fake_images = G(z)\n",
    "        \n",
    "            # Compute the discriminator losses on fake images            \n",
    "            D_fake = D(fake_images)\n",
    "            d_fake_loss = fake_loss(D_fake)\n",
    "        \n",
    "            # add up loss and perform backprop\n",
    "            d_loss = d_real_loss + d_fake_loss\n",
    "            d_loss.backward()\n",
    "            d_optimizer.step() \n",
    "\n",
    "            # 3. Train the generator with an adversarial loss\n",
    "            G.train()\n",
    "            g_optimizer.zero_grad()\n",
    "            \n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            # move x to GPU, if available\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "            fake_images = G(z)\n",
    "        \n",
    "            # Train with fake images and flipped labels\n",
    "        \n",
    "            # Compute the discriminator losses on fake images \n",
    "            # using flipped labels!\n",
    "            D_fake = D(fake_images)\n",
    "            g_loss = real_loss(D_fake) # use real loss to flip labels\n",
    "        \n",
    "            # perform backprop\n",
    "            g_loss.backward()\n",
    "            g_optimizer.step()\n",
    "            \n",
    "            \n",
    "            # ===============================================\n",
    "            #              END OF YOUR CODE\n",
    "            # ===============================================\n",
    "\n",
    "            # Print some loss stats\n",
    "            if batch_i % print_every == 0:\n",
    "                # append discriminator loss and generator loss\n",
    "                losses.append((d_loss.item(), g_loss.item()))\n",
    "                # print discriminator and generator loss\n",
    "                print('Epoch [{:5d}/{:5d}] | d_loss: {:6.4f} | g_loss: {:6.4f}'.format(\n",
    "                        epoch+1, n_epochs, d_loss.item(), g_loss.item()))\n",
    "\n",
    "\n",
    "        ## AFTER EACH EPOCH##    \n",
    "        # this code assumes your generator is named G, feel free to change the name\n",
    "        # generate and save sample, fake images\n",
    "        G.eval() # for generating samples\n",
    "        samples_z = G(fixed_z)\n",
    "        samples.append(samples_z)\n",
    "        G.train() # back to training mode\n",
    "\n",
    "    # Save training generator samples\n",
    "    with open('train_samples.pkl', 'wb') as f:\n",
    "        pkl.dump(samples, f)\n",
    "    \n",
    "    # finally return losses\n",
    "    return losses"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set your number of training epochs and train your GAN!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   20] | d_loss: 0.9676 | g_loss: 1.6911\n",
      "Epoch [    1/   20] | d_loss: 0.9860 | g_loss: 1.1953\n",
      "Epoch [    1/   20] | d_loss: 0.9688 | g_loss: 1.0010\n",
      "Epoch [    1/   20] | d_loss: 0.9352 | g_loss: 1.4345\n",
      "Epoch [    1/   20] | d_loss: 1.0321 | g_loss: 1.7775\n",
      "Epoch [    1/   20] | d_loss: 0.7592 | g_loss: 1.2627\n",
      "Epoch [    1/   20] | d_loss: 1.1674 | g_loss: 0.5903\n",
      "Epoch [    1/   20] | d_loss: 1.1821 | g_loss: 1.6513\n",
      "Epoch [    1/   20] | d_loss: 1.0050 | g_loss: 1.8502\n",
      "Epoch [    1/   20] | d_loss: 0.9425 | g_loss: 1.0142\n",
      "Epoch [    1/   20] | d_loss: 0.9585 | g_loss: 1.6174\n",
      "Epoch [    1/   20] | d_loss: 1.1472 | g_loss: 0.9037\n",
      "Epoch [    1/   20] | d_loss: 1.2717 | g_loss: 1.3622\n",
      "Epoch [    1/   20] | d_loss: 1.0742 | g_loss: 1.2437\n",
      "Epoch [    1/   20] | d_loss: 1.1471 | g_loss: 3.4696\n",
      "Epoch [    1/   20] | d_loss: 0.8888 | g_loss: 1.2734\n",
      "Epoch [    1/   20] | d_loss: 1.2778 | g_loss: 2.0298\n",
      "Epoch [    1/   20] | d_loss: 1.0229 | g_loss: 2.2200\n",
      "Epoch [    1/   20] | d_loss: 1.3712 | g_loss: 0.8051\n",
      "Epoch [    1/   20] | d_loss: 0.8543 | g_loss: 2.1028\n",
      "Epoch [    1/   20] | d_loss: 1.0874 | g_loss: 1.5676\n",
      "Epoch [    1/   20] | d_loss: 1.1777 | g_loss: 1.4777\n",
      "Epoch [    1/   20] | d_loss: 1.2725 | g_loss: 0.6924\n",
      "Epoch [    1/   20] | d_loss: 0.9709 | g_loss: 1.4517\n",
      "Epoch [    1/   20] | d_loss: 1.0857 | g_loss: 1.4868\n",
      "Epoch [    1/   20] | d_loss: 1.1832 | g_loss: 1.8501\n",
      "Epoch [    1/   20] | d_loss: 1.1831 | g_loss: 0.8374\n",
      "Epoch [    1/   20] | d_loss: 1.0603 | g_loss: 1.5839\n",
      "Epoch [    1/   20] | d_loss: 0.9327 | g_loss: 2.1773\n",
      "Epoch [    1/   20] | d_loss: 0.8657 | g_loss: 1.2284\n",
      "Epoch [    1/   20] | d_loss: 1.0408 | g_loss: 1.1643\n",
      "Epoch [    1/   20] | d_loss: 1.6540 | g_loss: 0.4778\n",
      "Epoch [    1/   20] | d_loss: 1.0414 | g_loss: 1.3571\n",
      "Epoch [    1/   20] | d_loss: 0.8539 | g_loss: 2.4779\n",
      "Epoch [    1/   20] | d_loss: 1.1442 | g_loss: 2.0349\n",
      "Epoch [    1/   20] | d_loss: 1.1088 | g_loss: 0.9981\n",
      "Epoch [    1/   20] | d_loss: 1.2711 | g_loss: 0.9799\n",
      "Epoch [    1/   20] | d_loss: 1.2830 | g_loss: 1.1603\n",
      "Epoch [    1/   20] | d_loss: 0.9275 | g_loss: 1.1676\n",
      "Epoch [    1/   20] | d_loss: 0.8627 | g_loss: 1.7610\n",
      "Epoch [    1/   20] | d_loss: 0.9376 | g_loss: 1.7327\n",
      "Epoch [    1/   20] | d_loss: 0.9983 | g_loss: 1.1854\n",
      "Epoch [    1/   20] | d_loss: 1.0333 | g_loss: 1.1860\n",
      "Epoch [    1/   20] | d_loss: 0.8442 | g_loss: 1.3893\n",
      "Epoch [    1/   20] | d_loss: 1.0878 | g_loss: 1.1780\n",
      "Epoch [    1/   20] | d_loss: 1.0006 | g_loss: 1.6181\n",
      "Epoch [    1/   20] | d_loss: 1.1676 | g_loss: 1.2778\n",
      "Epoch [    1/   20] | d_loss: 1.3777 | g_loss: 0.7239\n",
      "Epoch [    1/   20] | d_loss: 1.1260 | g_loss: 1.3190\n",
      "Epoch [    1/   20] | d_loss: 0.5934 | g_loss: 1.6914\n",
      "Epoch [    1/   20] | d_loss: 1.1558 | g_loss: 1.5048\n",
      "Epoch [    1/   20] | d_loss: 1.1989 | g_loss: 1.0432\n",
      "Epoch [    1/   20] | d_loss: 1.5392 | g_loss: 0.8887\n",
      "Epoch [    1/   20] | d_loss: 1.0632 | g_loss: 1.3007\n",
      "Epoch [    1/   20] | d_loss: 1.3703 | g_loss: 1.0068\n",
      "Epoch [    1/   20] | d_loss: 1.2809 | g_loss: 1.4082\n",
      "Epoch [    1/   20] | d_loss: 0.9057 | g_loss: 2.4200\n",
      "Epoch [    2/   20] | d_loss: 0.9239 | g_loss: 0.9849\n",
      "Epoch [    2/   20] | d_loss: 1.0439 | g_loss: 1.4708\n",
      "Epoch [    2/   20] | d_loss: 1.1669 | g_loss: 1.0527\n",
      "Epoch [    2/   20] | d_loss: 1.1020 | g_loss: 0.9281\n",
      "Epoch [    2/   20] | d_loss: 1.1750 | g_loss: 0.9878\n",
      "Epoch [    2/   20] | d_loss: 1.0319 | g_loss: 0.7305\n",
      "Epoch [    2/   20] | d_loss: 1.1734 | g_loss: 1.1882\n",
      "Epoch [    2/   20] | d_loss: 1.3102 | g_loss: 1.5521\n",
      "Epoch [    2/   20] | d_loss: 1.1864 | g_loss: 1.9953\n",
      "Epoch [    2/   20] | d_loss: 1.0386 | g_loss: 1.5964\n",
      "Epoch [    2/   20] | d_loss: 1.2860 | g_loss: 1.1661\n",
      "Epoch [    2/   20] | d_loss: 1.3812 | g_loss: 1.3430\n",
      "Epoch [    2/   20] | d_loss: 1.1646 | g_loss: 0.9080\n",
      "Epoch [    2/   20] | d_loss: 1.0613 | g_loss: 1.1944\n",
      "Epoch [    2/   20] | d_loss: 0.8722 | g_loss: 1.8393\n",
      "Epoch [    2/   20] | d_loss: 1.4602 | g_loss: 1.6249\n",
      "Epoch [    2/   20] | d_loss: 1.1054 | g_loss: 1.1940\n",
      "Epoch [    2/   20] | d_loss: 0.9470 | g_loss: 1.3032\n",
      "Epoch [    2/   20] | d_loss: 1.9301 | g_loss: 0.5296\n",
      "Epoch [    2/   20] | d_loss: 0.9336 | g_loss: 1.1525\n",
      "Epoch [    2/   20] | d_loss: 0.9220 | g_loss: 1.3188\n",
      "Epoch [    2/   20] | d_loss: 1.2046 | g_loss: 0.7775\n",
      "Epoch [    2/   20] | d_loss: 0.7781 | g_loss: 1.0723\n",
      "Epoch [    2/   20] | d_loss: 1.0466 | g_loss: 1.2607\n",
      "Epoch [    2/   20] | d_loss: 1.1515 | g_loss: 1.6720\n",
      "Epoch [    2/   20] | d_loss: 1.0878 | g_loss: 0.9470\n",
      "Epoch [    2/   20] | d_loss: 1.1321 | g_loss: 1.0949\n",
      "Epoch [    2/   20] | d_loss: 1.1350 | g_loss: 0.8842\n",
      "Epoch [    2/   20] | d_loss: 1.3014 | g_loss: 0.8909\n",
      "Epoch [    2/   20] | d_loss: 0.9576 | g_loss: 1.4282\n",
      "Epoch [    2/   20] | d_loss: 1.0951 | g_loss: 2.0283\n",
      "Epoch [    2/   20] | d_loss: 1.1247 | g_loss: 1.1074\n",
      "Epoch [    2/   20] | d_loss: 1.3633 | g_loss: 0.8818\n",
      "Epoch [    2/   20] | d_loss: 1.1324 | g_loss: 1.2894\n",
      "Epoch [    2/   20] | d_loss: 1.7054 | g_loss: 1.7729\n",
      "Epoch [    2/   20] | d_loss: 1.1873 | g_loss: 0.8309\n",
      "Epoch [    2/   20] | d_loss: 0.8900 | g_loss: 1.5166\n",
      "Epoch [    2/   20] | d_loss: 1.1386 | g_loss: 1.3945\n",
      "Epoch [    2/   20] | d_loss: 1.0967 | g_loss: 1.3424\n",
      "Epoch [    2/   20] | d_loss: 1.1168 | g_loss: 1.5903\n",
      "Epoch [    2/   20] | d_loss: 0.8248 | g_loss: 0.9432\n",
      "Epoch [    2/   20] | d_loss: 1.0324 | g_loss: 0.6572\n",
      "Epoch [    2/   20] | d_loss: 0.9445 | g_loss: 1.6199\n",
      "Epoch [    2/   20] | d_loss: 1.5977 | g_loss: 1.1900\n",
      "Epoch [    2/   20] | d_loss: 1.1761 | g_loss: 1.2716\n",
      "Epoch [    2/   20] | d_loss: 1.1663 | g_loss: 0.7714\n",
      "Epoch [    2/   20] | d_loss: 1.1038 | g_loss: 1.3761\n",
      "Epoch [    2/   20] | d_loss: 1.2215 | g_loss: 1.0630\n",
      "Epoch [    2/   20] | d_loss: 0.9145 | g_loss: 1.2163\n",
      "Epoch [    2/   20] | d_loss: 1.2774 | g_loss: 1.1692\n",
      "Epoch [    2/   20] | d_loss: 1.0369 | g_loss: 1.7349\n",
      "Epoch [    2/   20] | d_loss: 1.0475 | g_loss: 1.3084\n",
      "Epoch [    2/   20] | d_loss: 1.4143 | g_loss: 1.3519\n",
      "Epoch [    2/   20] | d_loss: 1.2023 | g_loss: 0.3948\n",
      "Epoch [    2/   20] | d_loss: 0.9738 | g_loss: 1.2780\n",
      "Epoch [    2/   20] | d_loss: 1.1619 | g_loss: 0.9810\n",
      "Epoch [    2/   20] | d_loss: 0.9390 | g_loss: 1.2847\n",
      "Epoch [    3/   20] | d_loss: 1.0328 | g_loss: 1.9148\n",
      "Epoch [    3/   20] | d_loss: 1.3063 | g_loss: 1.4358\n",
      "Epoch [    3/   20] | d_loss: 1.3498 | g_loss: 0.8947\n",
      "Epoch [    3/   20] | d_loss: 1.0354 | g_loss: 0.6688\n",
      "Epoch [    3/   20] | d_loss: 1.5199 | g_loss: 0.9587\n",
      "Epoch [    3/   20] | d_loss: 0.9617 | g_loss: 2.0722\n",
      "Epoch [    3/   20] | d_loss: 1.0108 | g_loss: 1.5090\n",
      "Epoch [    3/   20] | d_loss: 1.0953 | g_loss: 1.3565\n",
      "Epoch [    3/   20] | d_loss: 1.2205 | g_loss: 1.5646\n",
      "Epoch [    3/   20] | d_loss: 1.0531 | g_loss: 2.2791\n",
      "Epoch [    3/   20] | d_loss: 1.1034 | g_loss: 1.1759\n",
      "Epoch [    3/   20] | d_loss: 1.4383 | g_loss: 1.2136\n",
      "Epoch [    3/   20] | d_loss: 1.1330 | g_loss: 1.2957\n",
      "Epoch [    3/   20] | d_loss: 0.8083 | g_loss: 1.5258\n",
      "Epoch [    3/   20] | d_loss: 1.0363 | g_loss: 1.8944\n",
      "Epoch [    3/   20] | d_loss: 0.9837 | g_loss: 1.6195\n",
      "Epoch [    3/   20] | d_loss: 1.2455 | g_loss: 1.2697\n",
      "Epoch [    3/   20] | d_loss: 1.7405 | g_loss: 1.9458\n",
      "Epoch [    3/   20] | d_loss: 0.9624 | g_loss: 0.9298\n",
      "Epoch [    3/   20] | d_loss: 1.2584 | g_loss: 1.0317\n",
      "Epoch [    3/   20] | d_loss: 1.4192 | g_loss: 1.0850\n",
      "Epoch [    3/   20] | d_loss: 0.9554 | g_loss: 0.6941\n",
      "Epoch [    3/   20] | d_loss: 1.1444 | g_loss: 0.9526\n",
      "Epoch [    3/   20] | d_loss: 0.9631 | g_loss: 1.5522\n",
      "Epoch [    3/   20] | d_loss: 1.1614 | g_loss: 1.0263\n",
      "Epoch [    3/   20] | d_loss: 0.7437 | g_loss: 1.7347\n",
      "Epoch [    3/   20] | d_loss: 1.1424 | g_loss: 1.2079\n",
      "Epoch [    3/   20] | d_loss: 1.0357 | g_loss: 1.4021\n",
      "Epoch [    3/   20] | d_loss: 1.1560 | g_loss: 1.2063\n",
      "Epoch [    3/   20] | d_loss: 1.0453 | g_loss: 0.6883\n",
      "Epoch [    3/   20] | d_loss: 0.8989 | g_loss: 0.7247\n",
      "Epoch [    3/   20] | d_loss: 1.3839 | g_loss: 1.6007\n",
      "Epoch [    3/   20] | d_loss: 0.8300 | g_loss: 1.4742\n",
      "Epoch [    3/   20] | d_loss: 0.9700 | g_loss: 0.9643\n",
      "Epoch [    3/   20] | d_loss: 0.9834 | g_loss: 1.1084\n",
      "Epoch [    3/   20] | d_loss: 1.0030 | g_loss: 1.2973\n",
      "Epoch [    3/   20] | d_loss: 0.9891 | g_loss: 1.4591\n",
      "Epoch [    3/   20] | d_loss: 1.1212 | g_loss: 0.8704\n",
      "Epoch [    3/   20] | d_loss: 1.0000 | g_loss: 1.4955\n",
      "Epoch [    3/   20] | d_loss: 1.1902 | g_loss: 1.0028\n",
      "Epoch [    3/   20] | d_loss: 1.2422 | g_loss: 1.1933\n",
      "Epoch [    3/   20] | d_loss: 0.9040 | g_loss: 1.8872\n",
      "Epoch [    3/   20] | d_loss: 1.2597 | g_loss: 1.2357\n",
      "Epoch [    3/   20] | d_loss: 1.4773 | g_loss: 0.5417\n",
      "Epoch [    3/   20] | d_loss: 0.7828 | g_loss: 1.2412\n",
      "Epoch [    3/   20] | d_loss: 0.9975 | g_loss: 1.4744\n",
      "Epoch [    3/   20] | d_loss: 1.1769 | g_loss: 1.2349\n",
      "Epoch [    3/   20] | d_loss: 1.1391 | g_loss: 1.2103\n",
      "Epoch [    3/   20] | d_loss: 1.4663 | g_loss: 1.3329\n",
      "Epoch [    3/   20] | d_loss: 1.1591 | g_loss: 0.9857\n",
      "Epoch [    3/   20] | d_loss: 1.2824 | g_loss: 0.7639\n",
      "Epoch [    3/   20] | d_loss: 0.9622 | g_loss: 1.6525\n",
      "Epoch [    3/   20] | d_loss: 0.9667 | g_loss: 1.1696\n",
      "Epoch [    3/   20] | d_loss: 1.4024 | g_loss: 1.0886\n",
      "Epoch [    3/   20] | d_loss: 0.8513 | g_loss: 1.1318\n",
      "Epoch [    3/   20] | d_loss: 1.1803 | g_loss: 0.8075\n",
      "Epoch [    3/   20] | d_loss: 1.2652 | g_loss: 1.5446\n",
      "Epoch [    4/   20] | d_loss: 1.3087 | g_loss: 1.8058\n",
      "Epoch [    4/   20] | d_loss: 0.6959 | g_loss: 2.0149\n",
      "Epoch [    4/   20] | d_loss: 1.5716 | g_loss: 1.0894\n",
      "Epoch [    4/   20] | d_loss: 0.9976 | g_loss: 0.7657\n",
      "Epoch [    4/   20] | d_loss: 0.4512 | g_loss: 2.4834\n",
      "Epoch [    4/   20] | d_loss: 1.0121 | g_loss: 1.1337\n",
      "Epoch [    4/   20] | d_loss: 1.1067 | g_loss: 1.3577\n",
      "Epoch [    4/   20] | d_loss: 0.3571 | g_loss: 2.2963\n",
      "Epoch [    4/   20] | d_loss: 1.0851 | g_loss: 2.2813\n",
      "Epoch [    4/   20] | d_loss: 1.0985 | g_loss: 1.5526\n",
      "Epoch [    4/   20] | d_loss: 1.0049 | g_loss: 1.0899\n",
      "Epoch [    4/   20] | d_loss: 0.8128 | g_loss: 1.7503\n",
      "Epoch [    4/   20] | d_loss: 1.1313 | g_loss: 1.4549\n",
      "Epoch [    4/   20] | d_loss: 1.0903 | g_loss: 1.1621\n",
      "Epoch [    4/   20] | d_loss: 0.8459 | g_loss: 1.0197\n",
      "Epoch [    4/   20] | d_loss: 1.2415 | g_loss: 1.2209\n",
      "Epoch [    4/   20] | d_loss: 0.8710 | g_loss: 1.0359\n",
      "Epoch [    4/   20] | d_loss: 0.8072 | g_loss: 1.2857\n",
      "Epoch [    4/   20] | d_loss: 1.0647 | g_loss: 1.4233\n",
      "Epoch [    4/   20] | d_loss: 1.5690 | g_loss: 1.7885\n",
      "Epoch [    4/   20] | d_loss: 1.0503 | g_loss: 0.8441\n",
      "Epoch [    4/   20] | d_loss: 1.3750 | g_loss: 0.9643\n",
      "Epoch [    4/   20] | d_loss: 0.8214 | g_loss: 1.3345\n",
      "Epoch [    4/   20] | d_loss: 1.5729 | g_loss: 0.8468\n",
      "Epoch [    4/   20] | d_loss: 1.0153 | g_loss: 1.5555\n",
      "Epoch [    4/   20] | d_loss: 1.2443 | g_loss: 0.6661\n",
      "Epoch [    4/   20] | d_loss: 2.0231 | g_loss: 0.8048\n",
      "Epoch [    4/   20] | d_loss: 1.1299 | g_loss: 1.1418\n",
      "Epoch [    4/   20] | d_loss: 0.8635 | g_loss: 1.1165\n",
      "Epoch [    4/   20] | d_loss: 0.7671 | g_loss: 1.9636\n",
      "Epoch [    4/   20] | d_loss: 0.9436 | g_loss: 1.4486\n",
      "Epoch [    4/   20] | d_loss: 1.0455 | g_loss: 1.5447\n",
      "Epoch [    4/   20] | d_loss: 1.0410 | g_loss: 1.4838\n",
      "Epoch [    4/   20] | d_loss: 1.6003 | g_loss: 1.2929\n",
      "Epoch [    4/   20] | d_loss: 1.1193 | g_loss: 1.5500\n",
      "Epoch [    4/   20] | d_loss: 0.8044 | g_loss: 2.2028\n",
      "Epoch [    4/   20] | d_loss: 1.0129 | g_loss: 0.9133\n",
      "Epoch [    4/   20] | d_loss: 0.9698 | g_loss: 1.2201\n",
      "Epoch [    4/   20] | d_loss: 1.7115 | g_loss: 1.7519\n",
      "Epoch [    4/   20] | d_loss: 1.0223 | g_loss: 0.9940\n",
      "Epoch [    4/   20] | d_loss: 0.8791 | g_loss: 1.6753\n",
      "Epoch [    4/   20] | d_loss: 0.9210 | g_loss: 1.4160\n",
      "Epoch [    4/   20] | d_loss: 1.0588 | g_loss: 1.6381\n",
      "Epoch [    4/   20] | d_loss: 0.8291 | g_loss: 1.4366\n",
      "Epoch [    4/   20] | d_loss: 1.0828 | g_loss: 0.9133\n",
      "Epoch [    4/   20] | d_loss: 0.8686 | g_loss: 1.5388\n",
      "Epoch [    4/   20] | d_loss: 1.1838 | g_loss: 1.2919\n",
      "Epoch [    4/   20] | d_loss: 0.7069 | g_loss: 2.4253\n",
      "Epoch [    4/   20] | d_loss: 1.0946 | g_loss: 2.0147\n",
      "Epoch [    4/   20] | d_loss: 1.0602 | g_loss: 1.5941\n",
      "Epoch [    4/   20] | d_loss: 1.2722 | g_loss: 1.2642\n",
      "Epoch [    4/   20] | d_loss: 1.1791 | g_loss: 1.1483\n",
      "Epoch [    4/   20] | d_loss: 0.9217 | g_loss: 1.5286\n",
      "Epoch [    4/   20] | d_loss: 0.8790 | g_loss: 1.7156\n",
      "Epoch [    4/   20] | d_loss: 0.9759 | g_loss: 2.6021\n",
      "Epoch [    4/   20] | d_loss: 1.4057 | g_loss: 0.8089\n",
      "Epoch [    4/   20] | d_loss: 1.0420 | g_loss: 0.8927\n",
      "Epoch [    5/   20] | d_loss: 1.1094 | g_loss: 1.3148\n",
      "Epoch [    5/   20] | d_loss: 0.9507 | g_loss: 1.5993\n",
      "Epoch [    5/   20] | d_loss: 0.8967 | g_loss: 1.4169\n",
      "Epoch [    5/   20] | d_loss: 1.0801 | g_loss: 1.3829\n",
      "Epoch [    5/   20] | d_loss: 1.1382 | g_loss: 1.1361\n",
      "Epoch [    5/   20] | d_loss: 0.8629 | g_loss: 1.4870\n",
      "Epoch [    5/   20] | d_loss: 0.9248 | g_loss: 1.6427\n",
      "Epoch [    5/   20] | d_loss: 0.6632 | g_loss: 3.0418\n",
      "Epoch [    5/   20] | d_loss: 0.8634 | g_loss: 1.7034\n",
      "Epoch [    5/   20] | d_loss: 0.7737 | g_loss: 0.8661\n",
      "Epoch [    5/   20] | d_loss: 0.9826 | g_loss: 1.1529\n",
      "Epoch [    5/   20] | d_loss: 1.2141 | g_loss: 2.0955\n",
      "Epoch [    5/   20] | d_loss: 1.4827 | g_loss: 1.1785\n",
      "Epoch [    5/   20] | d_loss: 1.3415 | g_loss: 2.3738\n",
      "Epoch [    5/   20] | d_loss: 0.8975 | g_loss: 1.4388\n",
      "Epoch [    5/   20] | d_loss: 1.3934 | g_loss: 0.9560\n",
      "Epoch [    5/   20] | d_loss: 0.7881 | g_loss: 1.6608\n",
      "Epoch [    5/   20] | d_loss: 1.2886 | g_loss: 1.8347\n",
      "Epoch [    5/   20] | d_loss: 0.6970 | g_loss: 1.5875\n",
      "Epoch [    5/   20] | d_loss: 1.7485 | g_loss: 0.6054\n",
      "Epoch [    5/   20] | d_loss: 1.3297 | g_loss: 1.1724\n",
      "Epoch [    5/   20] | d_loss: 1.3013 | g_loss: 1.8028\n",
      "Epoch [    5/   20] | d_loss: 1.5977 | g_loss: 1.0622\n",
      "Epoch [    5/   20] | d_loss: 0.9979 | g_loss: 1.2578\n",
      "Epoch [    5/   20] | d_loss: 1.0086 | g_loss: 0.9203\n",
      "Epoch [    5/   20] | d_loss: 0.9900 | g_loss: 1.2321\n",
      "Epoch [    5/   20] | d_loss: 1.1634 | g_loss: 0.9603\n",
      "Epoch [    5/   20] | d_loss: 0.9883 | g_loss: 0.9070\n",
      "Epoch [    5/   20] | d_loss: 1.0820 | g_loss: 1.9432\n",
      "Epoch [    5/   20] | d_loss: 0.5877 | g_loss: 1.9160\n",
      "Epoch [    5/   20] | d_loss: 0.8722 | g_loss: 1.5271\n",
      "Epoch [    5/   20] | d_loss: 1.0861 | g_loss: 2.1907\n",
      "Epoch [    5/   20] | d_loss: 0.8459 | g_loss: 0.9291\n",
      "Epoch [    5/   20] | d_loss: 0.9976 | g_loss: 1.2151\n",
      "Epoch [    5/   20] | d_loss: 0.8541 | g_loss: 1.9494\n",
      "Epoch [    5/   20] | d_loss: 0.9290 | g_loss: 1.3254\n",
      "Epoch [    5/   20] | d_loss: 0.7237 | g_loss: 2.3764\n",
      "Epoch [    5/   20] | d_loss: 1.8937 | g_loss: 2.6593\n",
      "Epoch [    5/   20] | d_loss: 1.5365 | g_loss: 1.3025\n",
      "Epoch [    5/   20] | d_loss: 1.1102 | g_loss: 1.1398\n",
      "Epoch [    5/   20] | d_loss: 0.7749 | g_loss: 1.8841\n",
      "Epoch [    5/   20] | d_loss: 1.0922 | g_loss: 1.7576\n",
      "Epoch [    5/   20] | d_loss: 1.4762 | g_loss: 1.5943\n",
      "Epoch [    5/   20] | d_loss: 1.0039 | g_loss: 0.7832\n",
      "Epoch [    5/   20] | d_loss: 1.0046 | g_loss: 1.4315\n",
      "Epoch [    5/   20] | d_loss: 1.2094 | g_loss: 1.0694\n",
      "Epoch [    5/   20] | d_loss: 1.5234 | g_loss: 1.1031\n",
      "Epoch [    5/   20] | d_loss: 1.1601 | g_loss: 1.5399\n",
      "Epoch [    5/   20] | d_loss: 1.0492 | g_loss: 1.1243\n",
      "Epoch [    5/   20] | d_loss: 1.2543 | g_loss: 1.0055\n",
      "Epoch [    5/   20] | d_loss: 1.3408 | g_loss: 1.7392\n",
      "Epoch [    5/   20] | d_loss: 1.0297 | g_loss: 1.7380\n",
      "Epoch [    5/   20] | d_loss: 1.2770 | g_loss: 2.1537\n",
      "Epoch [    5/   20] | d_loss: 1.3121 | g_loss: 1.6702\n",
      "Epoch [    5/   20] | d_loss: 0.7287 | g_loss: 1.3146\n",
      "Epoch [    5/   20] | d_loss: 0.7601 | g_loss: 1.2951\n",
      "Epoch [    5/   20] | d_loss: 1.9371 | g_loss: 3.4024\n",
      "Epoch [    6/   20] | d_loss: 0.6118 | g_loss: 2.2641\n",
      "Epoch [    6/   20] | d_loss: 1.0464 | g_loss: 1.4070\n",
      "Epoch [    6/   20] | d_loss: 1.7725 | g_loss: 1.0107\n",
      "Epoch [    6/   20] | d_loss: 1.1468 | g_loss: 1.2552\n",
      "Epoch [    6/   20] | d_loss: 0.5828 | g_loss: 2.3569\n",
      "Epoch [    6/   20] | d_loss: 1.3057 | g_loss: 0.8626\n",
      "Epoch [    6/   20] | d_loss: 1.5239 | g_loss: 1.9876\n",
      "Epoch [    6/   20] | d_loss: 0.9016 | g_loss: 1.0744\n",
      "Epoch [    6/   20] | d_loss: 0.9330 | g_loss: 1.7048\n",
      "Epoch [    6/   20] | d_loss: 1.0795 | g_loss: 1.6778\n",
      "Epoch [    6/   20] | d_loss: 0.9494 | g_loss: 1.7986\n",
      "Epoch [    6/   20] | d_loss: 0.4280 | g_loss: 1.7800\n",
      "Epoch [    6/   20] | d_loss: 1.5305 | g_loss: 0.9172\n",
      "Epoch [    6/   20] | d_loss: 0.8794 | g_loss: 1.3766\n",
      "Epoch [    6/   20] | d_loss: 0.6095 | g_loss: 1.8507\n",
      "Epoch [    6/   20] | d_loss: 0.8133 | g_loss: 1.1778\n",
      "Epoch [    6/   20] | d_loss: 1.4801 | g_loss: 1.7993\n",
      "Epoch [    6/   20] | d_loss: 1.4140 | g_loss: 1.5908\n",
      "Epoch [    6/   20] | d_loss: 0.7982 | g_loss: 1.4776\n",
      "Epoch [    6/   20] | d_loss: 0.9030 | g_loss: 1.5081\n",
      "Epoch [    6/   20] | d_loss: 1.3154 | g_loss: 1.2648\n",
      "Epoch [    6/   20] | d_loss: 1.2010 | g_loss: 0.9912\n",
      "Epoch [    6/   20] | d_loss: 0.5992 | g_loss: 1.0930\n",
      "Epoch [    6/   20] | d_loss: 1.1247 | g_loss: 0.6702\n",
      "Epoch [    6/   20] | d_loss: 1.2469 | g_loss: 2.0698\n",
      "Epoch [    6/   20] | d_loss: 0.6192 | g_loss: 1.4602\n",
      "Epoch [    6/   20] | d_loss: 0.8937 | g_loss: 1.0934\n",
      "Epoch [    6/   20] | d_loss: 2.6042 | g_loss: 0.2336\n",
      "Epoch [    6/   20] | d_loss: 1.4668 | g_loss: 2.1021\n",
      "Epoch [    6/   20] | d_loss: 0.4584 | g_loss: 1.8937\n",
      "Epoch [    6/   20] | d_loss: 0.8363 | g_loss: 1.5189\n",
      "Epoch [    6/   20] | d_loss: 0.8990 | g_loss: 2.0183\n",
      "Epoch [    6/   20] | d_loss: 0.8304 | g_loss: 1.6838\n",
      "Epoch [    6/   20] | d_loss: 0.8285 | g_loss: 2.1685\n",
      "Epoch [    6/   20] | d_loss: 0.5588 | g_loss: 1.0222\n",
      "Epoch [    6/   20] | d_loss: 1.2632 | g_loss: 2.3068\n",
      "Epoch [    6/   20] | d_loss: 1.7389 | g_loss: 2.2775\n",
      "Epoch [    6/   20] | d_loss: 0.7034 | g_loss: 1.1364\n",
      "Epoch [    6/   20] | d_loss: 0.5790 | g_loss: 1.7592\n",
      "Epoch [    6/   20] | d_loss: 1.1485 | g_loss: 0.9530\n",
      "Epoch [    6/   20] | d_loss: 0.8262 | g_loss: 1.0661\n",
      "Epoch [    6/   20] | d_loss: 1.3428 | g_loss: 1.5669\n",
      "Epoch [    6/   20] | d_loss: 0.7067 | g_loss: 1.7621\n",
      "Epoch [    6/   20] | d_loss: 0.8970 | g_loss: 1.1749\n",
      "Epoch [    6/   20] | d_loss: 1.5447 | g_loss: 0.7284\n",
      "Epoch [    6/   20] | d_loss: 0.7359 | g_loss: 1.9570\n",
      "Epoch [    6/   20] | d_loss: 0.7925 | g_loss: 2.0573\n",
      "Epoch [    6/   20] | d_loss: 0.7242 | g_loss: 2.0485\n",
      "Epoch [    6/   20] | d_loss: 0.5213 | g_loss: 2.8658\n",
      "Epoch [    6/   20] | d_loss: 0.5958 | g_loss: 1.9390\n",
      "Epoch [    6/   20] | d_loss: 1.0804 | g_loss: 1.0812\n",
      "Epoch [    6/   20] | d_loss: 1.3501 | g_loss: 0.9436\n",
      "Epoch [    6/   20] | d_loss: 0.7925 | g_loss: 1.8153\n",
      "Epoch [    6/   20] | d_loss: 1.2995 | g_loss: 1.4686\n",
      "Epoch [    6/   20] | d_loss: 1.2454 | g_loss: 2.0964\n",
      "Epoch [    6/   20] | d_loss: 1.0540 | g_loss: 1.4097\n",
      "Epoch [    6/   20] | d_loss: 0.7904 | g_loss: 1.6525\n",
      "Epoch [    7/   20] | d_loss: 0.8717 | g_loss: 2.8630\n",
      "Epoch [    7/   20] | d_loss: 0.7339 | g_loss: 1.6814\n",
      "Epoch [    7/   20] | d_loss: 0.8679 | g_loss: 1.4938\n",
      "Epoch [    7/   20] | d_loss: 0.7728 | g_loss: 1.3124\n",
      "Epoch [    7/   20] | d_loss: 0.4902 | g_loss: 2.4550\n",
      "Epoch [    7/   20] | d_loss: 0.8405 | g_loss: 1.4939\n",
      "Epoch [    7/   20] | d_loss: 1.4061 | g_loss: 1.1689\n",
      "Epoch [    7/   20] | d_loss: 0.7908 | g_loss: 2.4657\n",
      "Epoch [    7/   20] | d_loss: 1.4125 | g_loss: 0.7517\n",
      "Epoch [    7/   20] | d_loss: 0.8796 | g_loss: 1.9068\n",
      "Epoch [    7/   20] | d_loss: 0.9285 | g_loss: 3.3219\n",
      "Epoch [    7/   20] | d_loss: 1.1326 | g_loss: 1.7701\n",
      "Epoch [    7/   20] | d_loss: 0.7950 | g_loss: 1.7453\n",
      "Epoch [    7/   20] | d_loss: 1.5497 | g_loss: 1.5432\n",
      "Epoch [    7/   20] | d_loss: 0.9262 | g_loss: 2.8292\n",
      "Epoch [    7/   20] | d_loss: 0.6481 | g_loss: 1.7707\n",
      "Epoch [    7/   20] | d_loss: 0.6949 | g_loss: 2.4039\n",
      "Epoch [    7/   20] | d_loss: 1.5661 | g_loss: 1.4982\n",
      "Epoch [    7/   20] | d_loss: 0.5018 | g_loss: 1.1197\n",
      "Epoch [    7/   20] | d_loss: 0.5870 | g_loss: 2.3856\n",
      "Epoch [    7/   20] | d_loss: 1.5710 | g_loss: 1.4458\n",
      "Epoch [    7/   20] | d_loss: 0.7075 | g_loss: 1.1079\n",
      "Epoch [    7/   20] | d_loss: 1.1308 | g_loss: 1.0685\n",
      "Epoch [    7/   20] | d_loss: 0.8746 | g_loss: 1.2686\n",
      "Epoch [    7/   20] | d_loss: 0.5317 | g_loss: 2.5475\n",
      "Epoch [    7/   20] | d_loss: 0.5612 | g_loss: 2.2726\n",
      "Epoch [    7/   20] | d_loss: 0.6283 | g_loss: 2.1699\n",
      "Epoch [    7/   20] | d_loss: 1.4281 | g_loss: 1.5790\n",
      "Epoch [    7/   20] | d_loss: 0.5504 | g_loss: 1.8412\n",
      "Epoch [    7/   20] | d_loss: 1.2070 | g_loss: 3.2317\n",
      "Epoch [    7/   20] | d_loss: 1.2012 | g_loss: 0.7157\n",
      "Epoch [    7/   20] | d_loss: 1.5599 | g_loss: 1.9504\n",
      "Epoch [    7/   20] | d_loss: 1.4183 | g_loss: 2.2880\n",
      "Epoch [    7/   20] | d_loss: 1.0790 | g_loss: 1.9067\n",
      "Epoch [    7/   20] | d_loss: 1.0769 | g_loss: 1.3515\n",
      "Epoch [    7/   20] | d_loss: 0.7015 | g_loss: 1.7393\n",
      "Epoch [    7/   20] | d_loss: 0.7694 | g_loss: 1.4655\n",
      "Epoch [    7/   20] | d_loss: 0.4779 | g_loss: 2.6088\n",
      "Epoch [    7/   20] | d_loss: 0.4980 | g_loss: 1.1883\n",
      "Epoch [    7/   20] | d_loss: 0.5700 | g_loss: 1.2382\n",
      "Epoch [    7/   20] | d_loss: 1.4944 | g_loss: 1.8992\n",
      "Epoch [    7/   20] | d_loss: 1.0946 | g_loss: 1.9482\n",
      "Epoch [    7/   20] | d_loss: 0.4925 | g_loss: 2.0184\n",
      "Epoch [    7/   20] | d_loss: 0.3645 | g_loss: 1.8294\n",
      "Epoch [    7/   20] | d_loss: 0.6061 | g_loss: 1.1327\n",
      "Epoch [    7/   20] | d_loss: 0.8919 | g_loss: 1.6409\n",
      "Epoch [    7/   20] | d_loss: 1.3803 | g_loss: 1.0459\n",
      "Epoch [    7/   20] | d_loss: 0.9754 | g_loss: 1.3113\n",
      "Epoch [    7/   20] | d_loss: 1.0077 | g_loss: 3.0849\n",
      "Epoch [    7/   20] | d_loss: 0.5401 | g_loss: 1.0100\n",
      "Epoch [    7/   20] | d_loss: 1.1102 | g_loss: 2.2529\n",
      "Epoch [    7/   20] | d_loss: 0.8575 | g_loss: 2.4222\n",
      "Epoch [    7/   20] | d_loss: 0.2461 | g_loss: 2.1346\n",
      "Epoch [    7/   20] | d_loss: 0.9401 | g_loss: 1.5053\n",
      "Epoch [    7/   20] | d_loss: 0.6856 | g_loss: 2.8767\n",
      "Epoch [    7/   20] | d_loss: 0.7536 | g_loss: 1.3758\n",
      "Epoch [    7/   20] | d_loss: 0.8906 | g_loss: 2.0579\n",
      "Epoch [    8/   20] | d_loss: 0.8703 | g_loss: 2.0282\n",
      "Epoch [    8/   20] | d_loss: 0.8932 | g_loss: 1.6999\n",
      "Epoch [    8/   20] | d_loss: 0.9698 | g_loss: 3.2545\n",
      "Epoch [    8/   20] | d_loss: 1.2696 | g_loss: 2.5314\n",
      "Epoch [    8/   20] | d_loss: 0.4084 | g_loss: 2.3250\n",
      "Epoch [    8/   20] | d_loss: 0.5238 | g_loss: 2.2543\n",
      "Epoch [    8/   20] | d_loss: 0.6131 | g_loss: 0.9970\n",
      "Epoch [    8/   20] | d_loss: 1.0356 | g_loss: 1.2371\n",
      "Epoch [    8/   20] | d_loss: 0.7150 | g_loss: 1.3427\n",
      "Epoch [    8/   20] | d_loss: 0.6263 | g_loss: 2.5008\n",
      "Epoch [    8/   20] | d_loss: 1.2383 | g_loss: 1.9489\n",
      "Epoch [    8/   20] | d_loss: 1.1681 | g_loss: 1.8416\n",
      "Epoch [    8/   20] | d_loss: 0.5688 | g_loss: 1.6886\n",
      "Epoch [    8/   20] | d_loss: 1.4399 | g_loss: 2.5763\n",
      "Epoch [    8/   20] | d_loss: 0.3214 | g_loss: 1.7616\n",
      "Epoch [    8/   20] | d_loss: 1.5989 | g_loss: 1.5555\n",
      "Epoch [    8/   20] | d_loss: 0.6546 | g_loss: 2.8573\n",
      "Epoch [    8/   20] | d_loss: 0.6804 | g_loss: 1.1102\n",
      "Epoch [    8/   20] | d_loss: 0.6123 | g_loss: 2.2260\n",
      "Epoch [    8/   20] | d_loss: 0.7022 | g_loss: 1.1903\n",
      "Epoch [    8/   20] | d_loss: 0.6963 | g_loss: 1.8496\n",
      "Epoch [    8/   20] | d_loss: 0.8324 | g_loss: 1.3430\n",
      "Epoch [    8/   20] | d_loss: 0.9490 | g_loss: 3.1164\n",
      "Epoch [    8/   20] | d_loss: 0.6805 | g_loss: 1.9548\n",
      "Epoch [    8/   20] | d_loss: 0.6239 | g_loss: 1.5233\n",
      "Epoch [    8/   20] | d_loss: 1.1506 | g_loss: 1.5385\n",
      "Epoch [    8/   20] | d_loss: 0.8424 | g_loss: 1.5123\n",
      "Epoch [    8/   20] | d_loss: 1.1239 | g_loss: 2.0171\n",
      "Epoch [    8/   20] | d_loss: 0.7298 | g_loss: 1.3593\n",
      "Epoch [    8/   20] | d_loss: 0.7258 | g_loss: 1.7922\n",
      "Epoch [    8/   20] | d_loss: 0.9257 | g_loss: 0.8839\n",
      "Epoch [    8/   20] | d_loss: 0.3315 | g_loss: 2.8789\n",
      "Epoch [    8/   20] | d_loss: 0.8967 | g_loss: 1.9475\n",
      "Epoch [    8/   20] | d_loss: 0.5694 | g_loss: 1.4063\n",
      "Epoch [    8/   20] | d_loss: 0.8306 | g_loss: 1.3653\n",
      "Epoch [    8/   20] | d_loss: 1.2211 | g_loss: 1.3684\n",
      "Epoch [    8/   20] | d_loss: 1.0889 | g_loss: 2.4473\n",
      "Epoch [    8/   20] | d_loss: 0.9203 | g_loss: 1.3038\n",
      "Epoch [    8/   20] | d_loss: 0.3769 | g_loss: 1.6235\n",
      "Epoch [    8/   20] | d_loss: 0.3974 | g_loss: 1.7755\n",
      "Epoch [    8/   20] | d_loss: 0.8263 | g_loss: 1.1948\n",
      "Epoch [    8/   20] | d_loss: 1.1427 | g_loss: 2.8934\n",
      "Epoch [    8/   20] | d_loss: 0.9728 | g_loss: 1.8252\n",
      "Epoch [    8/   20] | d_loss: 1.0308 | g_loss: 1.9317\n",
      "Epoch [    8/   20] | d_loss: 0.9127 | g_loss: 2.6934\n",
      "Epoch [    8/   20] | d_loss: 0.8376 | g_loss: 1.5793\n",
      "Epoch [    8/   20] | d_loss: 0.6112 | g_loss: 1.9696\n",
      "Epoch [    8/   20] | d_loss: 0.7025 | g_loss: 1.3507\n",
      "Epoch [    8/   20] | d_loss: 0.8644 | g_loss: 1.0123\n",
      "Epoch [    8/   20] | d_loss: 0.9051 | g_loss: 1.9862\n",
      "Epoch [    8/   20] | d_loss: 1.3088 | g_loss: 1.4905\n",
      "Epoch [    8/   20] | d_loss: 1.4628 | g_loss: 2.8177\n",
      "Epoch [    8/   20] | d_loss: 0.4249 | g_loss: 1.8853\n",
      "Epoch [    8/   20] | d_loss: 0.7284 | g_loss: 2.6959\n",
      "Epoch [    8/   20] | d_loss: 1.1466 | g_loss: 0.9781\n",
      "Epoch [    8/   20] | d_loss: 0.3971 | g_loss: 2.2957\n",
      "Epoch [    8/   20] | d_loss: 0.7355 | g_loss: 1.5477\n",
      "Epoch [    9/   20] | d_loss: 0.9513 | g_loss: 2.4511\n",
      "Epoch [    9/   20] | d_loss: 1.0136 | g_loss: 1.8882\n",
      "Epoch [    9/   20] | d_loss: 0.6202 | g_loss: 2.8028\n",
      "Epoch [    9/   20] | d_loss: 0.4144 | g_loss: 1.5227\n",
      "Epoch [    9/   20] | d_loss: 0.4758 | g_loss: 2.0083\n",
      "Epoch [    9/   20] | d_loss: 0.7435 | g_loss: 2.5732\n",
      "Epoch [    9/   20] | d_loss: 0.5920 | g_loss: 2.6930\n",
      "Epoch [    9/   20] | d_loss: 0.8838 | g_loss: 0.8093\n",
      "Epoch [    9/   20] | d_loss: 0.7508 | g_loss: 1.7880\n",
      "Epoch [    9/   20] | d_loss: 1.9656 | g_loss: 2.1470\n",
      "Epoch [    9/   20] | d_loss: 0.3830 | g_loss: 1.9213\n",
      "Epoch [    9/   20] | d_loss: 0.7992 | g_loss: 2.3753\n",
      "Epoch [    9/   20] | d_loss: 1.1619 | g_loss: 3.5968\n",
      "Epoch [    9/   20] | d_loss: 0.6684 | g_loss: 3.5705\n",
      "Epoch [    9/   20] | d_loss: 0.8368 | g_loss: 1.4438\n",
      "Epoch [    9/   20] | d_loss: 0.3654 | g_loss: 2.9936\n",
      "Epoch [    9/   20] | d_loss: 1.8275 | g_loss: 2.2845\n",
      "Epoch [    9/   20] | d_loss: 0.2698 | g_loss: 4.3254\n",
      "Epoch [    9/   20] | d_loss: 0.7467 | g_loss: 2.0069\n",
      "Epoch [    9/   20] | d_loss: 1.1271 | g_loss: 1.5915\n",
      "Epoch [    9/   20] | d_loss: 1.2376 | g_loss: 2.1593\n",
      "Epoch [    9/   20] | d_loss: 0.5939 | g_loss: 2.1384\n",
      "Epoch [    9/   20] | d_loss: 0.8134 | g_loss: 2.7277\n",
      "Epoch [    9/   20] | d_loss: 0.3417 | g_loss: 3.3261\n",
      "Epoch [    9/   20] | d_loss: 0.6136 | g_loss: 3.4338\n",
      "Epoch [    9/   20] | d_loss: 0.4077 | g_loss: 2.1269\n",
      "Epoch [    9/   20] | d_loss: 0.5255 | g_loss: 2.1108\n",
      "Epoch [    9/   20] | d_loss: 0.8813 | g_loss: 1.7974\n",
      "Epoch [    9/   20] | d_loss: 0.8785 | g_loss: 2.4114\n",
      "Epoch [    9/   20] | d_loss: 0.6489 | g_loss: 2.1109\n",
      "Epoch [    9/   20] | d_loss: 0.4346 | g_loss: 2.9434\n",
      "Epoch [    9/   20] | d_loss: 0.2341 | g_loss: 3.0525\n",
      "Epoch [    9/   20] | d_loss: 0.9234 | g_loss: 2.4840\n",
      "Epoch [    9/   20] | d_loss: 0.7262 | g_loss: 2.4091\n",
      "Epoch [    9/   20] | d_loss: 0.8963 | g_loss: 2.1522\n",
      "Epoch [    9/   20] | d_loss: 1.9601 | g_loss: 1.7448\n",
      "Epoch [    9/   20] | d_loss: 0.5081 | g_loss: 4.1171\n",
      "Epoch [    9/   20] | d_loss: 0.8657 | g_loss: 2.5320\n",
      "Epoch [    9/   20] | d_loss: 0.6745 | g_loss: 1.7733\n",
      "Epoch [    9/   20] | d_loss: 0.7220 | g_loss: 1.3106\n",
      "Epoch [    9/   20] | d_loss: 1.1174 | g_loss: 3.1463\n",
      "Epoch [    9/   20] | d_loss: 0.7108 | g_loss: 3.3757\n",
      "Epoch [    9/   20] | d_loss: 0.7834 | g_loss: 2.6014\n",
      "Epoch [    9/   20] | d_loss: 0.9256 | g_loss: 0.9729\n",
      "Epoch [    9/   20] | d_loss: 1.1063 | g_loss: 1.3874\n",
      "Epoch [    9/   20] | d_loss: 0.9014 | g_loss: 2.3423\n",
      "Epoch [    9/   20] | d_loss: 0.7811 | g_loss: 2.8633\n",
      "Epoch [    9/   20] | d_loss: 0.3844 | g_loss: 1.7175\n",
      "Epoch [    9/   20] | d_loss: 0.6734 | g_loss: 2.9366\n",
      "Epoch [    9/   20] | d_loss: 0.9014 | g_loss: 3.6766\n",
      "Epoch [    9/   20] | d_loss: 0.5495 | g_loss: 1.2338\n",
      "Epoch [    9/   20] | d_loss: 0.6902 | g_loss: 1.5362\n",
      "Epoch [    9/   20] | d_loss: 1.3349 | g_loss: 3.0147\n",
      "Epoch [    9/   20] | d_loss: 0.9111 | g_loss: 1.7636\n",
      "Epoch [    9/   20] | d_loss: 0.8814 | g_loss: 2.0065\n",
      "Epoch [    9/   20] | d_loss: 0.6270 | g_loss: 1.6694\n",
      "Epoch [    9/   20] | d_loss: 0.5960 | g_loss: 1.0082\n",
      "Epoch [   10/   20] | d_loss: 0.8962 | g_loss: 2.5883\n",
      "Epoch [   10/   20] | d_loss: 1.4487 | g_loss: 3.6937\n",
      "Epoch [   10/   20] | d_loss: 0.4487 | g_loss: 3.0803\n",
      "Epoch [   10/   20] | d_loss: 0.2265 | g_loss: 2.4965\n",
      "Epoch [   10/   20] | d_loss: 0.6191 | g_loss: 3.0589\n",
      "Epoch [   10/   20] | d_loss: 0.5672 | g_loss: 3.5241\n",
      "Epoch [   10/   20] | d_loss: 0.7541 | g_loss: 2.1373\n",
      "Epoch [   10/   20] | d_loss: 0.6786 | g_loss: 2.6289\n",
      "Epoch [   10/   20] | d_loss: 0.3178 | g_loss: 1.6843\n",
      "Epoch [   10/   20] | d_loss: 1.0586 | g_loss: 2.4313\n",
      "Epoch [   10/   20] | d_loss: 0.4933 | g_loss: 2.5107\n",
      "Epoch [   10/   20] | d_loss: 0.6514 | g_loss: 3.4134\n",
      "Epoch [   10/   20] | d_loss: 0.4534 | g_loss: 3.0303\n",
      "Epoch [   10/   20] | d_loss: 0.6982 | g_loss: 2.9748\n",
      "Epoch [   10/   20] | d_loss: 0.4254 | g_loss: 3.7385\n",
      "Epoch [   10/   20] | d_loss: 0.9130 | g_loss: 2.3831\n",
      "Epoch [   10/   20] | d_loss: 0.4928 | g_loss: 2.4497\n",
      "Epoch [   10/   20] | d_loss: 0.7176 | g_loss: 1.8748\n",
      "Epoch [   10/   20] | d_loss: 1.0815 | g_loss: 1.9816\n",
      "Epoch [   10/   20] | d_loss: 0.8198 | g_loss: 2.9593\n",
      "Epoch [   10/   20] | d_loss: 0.5418 | g_loss: 0.9023\n",
      "Epoch [   10/   20] | d_loss: 0.8581 | g_loss: 1.4673\n",
      "Epoch [   10/   20] | d_loss: 0.7236 | g_loss: 2.1944\n",
      "Epoch [   10/   20] | d_loss: 0.8027 | g_loss: 3.3100\n",
      "Epoch [   10/   20] | d_loss: 0.8402 | g_loss: 1.9348\n",
      "Epoch [   10/   20] | d_loss: 0.8260 | g_loss: 1.7117\n",
      "Epoch [   10/   20] | d_loss: 0.5643 | g_loss: 3.2319\n",
      "Epoch [   10/   20] | d_loss: 0.9556 | g_loss: 2.2960\n",
      "Epoch [   10/   20] | d_loss: 1.4973 | g_loss: 2.6114\n",
      "Epoch [   10/   20] | d_loss: 1.3882 | g_loss: 0.5927\n",
      "Epoch [   10/   20] | d_loss: 0.4324 | g_loss: 2.7364\n",
      "Epoch [   10/   20] | d_loss: 1.0592 | g_loss: 0.7354\n",
      "Epoch [   10/   20] | d_loss: 0.4941 | g_loss: 2.2411\n",
      "Epoch [   10/   20] | d_loss: 0.6788 | g_loss: 1.9482\n",
      "Epoch [   10/   20] | d_loss: 1.0207 | g_loss: 2.6383\n",
      "Epoch [   10/   20] | d_loss: 0.5999 | g_loss: 3.5493\n",
      "Epoch [   10/   20] | d_loss: 0.4652 | g_loss: 1.6469\n",
      "Epoch [   10/   20] | d_loss: 0.9069 | g_loss: 3.9916\n",
      "Epoch [   10/   20] | d_loss: 0.7019 | g_loss: 1.4417\n",
      "Epoch [   10/   20] | d_loss: 0.9414 | g_loss: 2.7940\n",
      "Epoch [   10/   20] | d_loss: 0.9996 | g_loss: 1.0967\n",
      "Epoch [   10/   20] | d_loss: 0.4583 | g_loss: 2.8184\n",
      "Epoch [   10/   20] | d_loss: 0.4541 | g_loss: 3.5711\n",
      "Epoch [   10/   20] | d_loss: 0.2322 | g_loss: 4.0011\n",
      "Epoch [   10/   20] | d_loss: 1.0877 | g_loss: 1.3509\n",
      "Epoch [   10/   20] | d_loss: 1.2866 | g_loss: 1.5930\n",
      "Epoch [   10/   20] | d_loss: 0.7432 | g_loss: 2.3075\n",
      "Epoch [   10/   20] | d_loss: 1.1241 | g_loss: 1.5193\n",
      "Epoch [   10/   20] | d_loss: 1.1107 | g_loss: 2.4805\n",
      "Epoch [   10/   20] | d_loss: 0.2834 | g_loss: 3.3277\n",
      "Epoch [   10/   20] | d_loss: 0.6089 | g_loss: 2.9198\n",
      "Epoch [   10/   20] | d_loss: 0.8156 | g_loss: 2.2737\n",
      "Epoch [   10/   20] | d_loss: 0.8935 | g_loss: 0.9857\n",
      "Epoch [   10/   20] | d_loss: 0.7454 | g_loss: 2.4182\n",
      "Epoch [   10/   20] | d_loss: 0.3432 | g_loss: 2.0112\n",
      "Epoch [   10/   20] | d_loss: 0.7236 | g_loss: 1.7448\n",
      "Epoch [   10/   20] | d_loss: 0.4891 | g_loss: 4.2802\n",
      "Epoch [   11/   20] | d_loss: 1.0688 | g_loss: 1.6381\n",
      "Epoch [   11/   20] | d_loss: 0.3975 | g_loss: 3.9436\n",
      "Epoch [   11/   20] | d_loss: 0.9535 | g_loss: 2.5536\n",
      "Epoch [   11/   20] | d_loss: 0.4120 | g_loss: 3.9490\n",
      "Epoch [   11/   20] | d_loss: 0.4544 | g_loss: 3.5483\n",
      "Epoch [   11/   20] | d_loss: 1.3209 | g_loss: 1.2965\n",
      "Epoch [   11/   20] | d_loss: 0.8827 | g_loss: 3.4834\n",
      "Epoch [   11/   20] | d_loss: 1.1972 | g_loss: 1.6067\n",
      "Epoch [   11/   20] | d_loss: 0.8473 | g_loss: 3.2889\n",
      "Epoch [   11/   20] | d_loss: 0.7363 | g_loss: 2.8775\n",
      "Epoch [   11/   20] | d_loss: 0.8106 | g_loss: 1.2219\n",
      "Epoch [   11/   20] | d_loss: 1.0735 | g_loss: 3.5773\n",
      "Epoch [   11/   20] | d_loss: 1.2265 | g_loss: 2.4657\n",
      "Epoch [   11/   20] | d_loss: 0.7664 | g_loss: 1.9185\n",
      "Epoch [   11/   20] | d_loss: 0.9715 | g_loss: 2.3568\n",
      "Epoch [   11/   20] | d_loss: 1.5295 | g_loss: 1.2279\n",
      "Epoch [   11/   20] | d_loss: 1.0653 | g_loss: 3.4844\n",
      "Epoch [   11/   20] | d_loss: 0.4984 | g_loss: 0.9576\n",
      "Epoch [   11/   20] | d_loss: 0.4030 | g_loss: 1.6235\n",
      "Epoch [   11/   20] | d_loss: 0.8749 | g_loss: 2.0441\n",
      "Epoch [   11/   20] | d_loss: 1.0243 | g_loss: 2.4140\n",
      "Epoch [   11/   20] | d_loss: 0.6407 | g_loss: 6.0751\n",
      "Epoch [   11/   20] | d_loss: 0.9826 | g_loss: 4.6070\n",
      "Epoch [   11/   20] | d_loss: 0.6176 | g_loss: 2.8263\n",
      "Epoch [   11/   20] | d_loss: 1.0231 | g_loss: 1.3613\n",
      "Epoch [   11/   20] | d_loss: 0.5942 | g_loss: 2.0016\n",
      "Epoch [   11/   20] | d_loss: 0.1197 | g_loss: 2.7854\n",
      "Epoch [   11/   20] | d_loss: 0.7986 | g_loss: 3.0095\n",
      "Epoch [   11/   20] | d_loss: 0.6368 | g_loss: 2.8020\n",
      "Epoch [   11/   20] | d_loss: 0.5031 | g_loss: 3.8317\n",
      "Epoch [   11/   20] | d_loss: 0.8399 | g_loss: 2.7373\n",
      "Epoch [   11/   20] | d_loss: 0.2863 | g_loss: 4.1479\n",
      "Epoch [   11/   20] | d_loss: 0.2550 | g_loss: 0.8374\n",
      "Epoch [   11/   20] | d_loss: 1.6788 | g_loss: 3.3872\n",
      "Epoch [   11/   20] | d_loss: 1.4942 | g_loss: 2.4205\n",
      "Epoch [   11/   20] | d_loss: 0.3417 | g_loss: 3.1123\n",
      "Epoch [   11/   20] | d_loss: 0.8372 | g_loss: 2.7881\n",
      "Epoch [   11/   20] | d_loss: 0.9031 | g_loss: 1.7544\n",
      "Epoch [   11/   20] | d_loss: 0.2847 | g_loss: 2.7250\n",
      "Epoch [   11/   20] | d_loss: 0.5772 | g_loss: 3.1671\n",
      "Epoch [   11/   20] | d_loss: 0.8525 | g_loss: 3.1422\n",
      "Epoch [   11/   20] | d_loss: 0.3510 | g_loss: 2.9819\n",
      "Epoch [   11/   20] | d_loss: 0.7076 | g_loss: 3.4821\n",
      "Epoch [   11/   20] | d_loss: 0.3398 | g_loss: 2.6610\n",
      "Epoch [   11/   20] | d_loss: 0.5163 | g_loss: 2.2092\n",
      "Epoch [   11/   20] | d_loss: 0.6521 | g_loss: 1.5599\n",
      "Epoch [   11/   20] | d_loss: 0.7904 | g_loss: 2.4259\n",
      "Epoch [   11/   20] | d_loss: 0.6536 | g_loss: 2.5251\n",
      "Epoch [   11/   20] | d_loss: 0.2831 | g_loss: 2.0834\n",
      "Epoch [   11/   20] | d_loss: 0.4752 | g_loss: 1.2127\n",
      "Epoch [   11/   20] | d_loss: 0.1971 | g_loss: 2.0126\n",
      "Epoch [   11/   20] | d_loss: 0.7302 | g_loss: 1.2967\n",
      "Epoch [   11/   20] | d_loss: 0.4150 | g_loss: 1.5557\n",
      "Epoch [   11/   20] | d_loss: 1.3552 | g_loss: 2.0483\n",
      "Epoch [   11/   20] | d_loss: 1.0617 | g_loss: 2.1195\n",
      "Epoch [   11/   20] | d_loss: 0.4656 | g_loss: 2.8916\n",
      "Epoch [   11/   20] | d_loss: 1.4274 | g_loss: 1.5999\n",
      "Epoch [   12/   20] | d_loss: 0.6041 | g_loss: 3.3310\n",
      "Epoch [   12/   20] | d_loss: 0.6714 | g_loss: 3.0603\n",
      "Epoch [   12/   20] | d_loss: 0.9199 | g_loss: 2.2670\n",
      "Epoch [   12/   20] | d_loss: 0.3882 | g_loss: 3.0453\n",
      "Epoch [   12/   20] | d_loss: 0.5817 | g_loss: 2.8146\n",
      "Epoch [   12/   20] | d_loss: 0.5643 | g_loss: 4.3994\n",
      "Epoch [   12/   20] | d_loss: 0.3647 | g_loss: 4.4195\n",
      "Epoch [   12/   20] | d_loss: 0.5783 | g_loss: 2.4066\n",
      "Epoch [   12/   20] | d_loss: 0.5304 | g_loss: 1.6959\n",
      "Epoch [   12/   20] | d_loss: 0.3883 | g_loss: 2.2023\n",
      "Epoch [   12/   20] | d_loss: 0.2580 | g_loss: 3.6110\n",
      "Epoch [   12/   20] | d_loss: 0.4830 | g_loss: 2.6878\n",
      "Epoch [   12/   20] | d_loss: 1.2203 | g_loss: 1.4317\n",
      "Epoch [   12/   20] | d_loss: 0.6074 | g_loss: 2.0169\n",
      "Epoch [   12/   20] | d_loss: 0.2617 | g_loss: 3.2253\n",
      "Epoch [   12/   20] | d_loss: 1.3529 | g_loss: 3.1919\n",
      "Epoch [   12/   20] | d_loss: 1.6510 | g_loss: 1.1275\n",
      "Epoch [   12/   20] | d_loss: 0.2543 | g_loss: 2.8911\n",
      "Epoch [   12/   20] | d_loss: 0.5921 | g_loss: 1.6429\n",
      "Epoch [   12/   20] | d_loss: 0.4468 | g_loss: 3.0435\n",
      "Epoch [   12/   20] | d_loss: 0.5879 | g_loss: 2.3736\n",
      "Epoch [   12/   20] | d_loss: 1.1931 | g_loss: 2.7515\n",
      "Epoch [   12/   20] | d_loss: 1.2731 | g_loss: 1.4507\n",
      "Epoch [   12/   20] | d_loss: 0.3779 | g_loss: 1.9607\n",
      "Epoch [   12/   20] | d_loss: 0.5712 | g_loss: 2.8909\n",
      "Epoch [   12/   20] | d_loss: 0.7770 | g_loss: 2.6055\n",
      "Epoch [   12/   20] | d_loss: 1.3312 | g_loss: 3.3124\n",
      "Epoch [   12/   20] | d_loss: 0.3503 | g_loss: 3.8364\n",
      "Epoch [   12/   20] | d_loss: 0.9441 | g_loss: 1.2781\n",
      "Epoch [   12/   20] | d_loss: 1.2548 | g_loss: 3.2425\n",
      "Epoch [   12/   20] | d_loss: 0.6446 | g_loss: 2.3463\n",
      "Epoch [   12/   20] | d_loss: 0.5448 | g_loss: 2.5524\n",
      "Epoch [   12/   20] | d_loss: 0.4846 | g_loss: 2.9703\n",
      "Epoch [   12/   20] | d_loss: 0.5503 | g_loss: 2.3170\n",
      "Epoch [   12/   20] | d_loss: 0.8863 | g_loss: 2.4576\n",
      "Epoch [   12/   20] | d_loss: 0.3365 | g_loss: 3.4713\n",
      "Epoch [   12/   20] | d_loss: 0.5134 | g_loss: 2.5023\n",
      "Epoch [   12/   20] | d_loss: 1.4123 | g_loss: 4.0007\n",
      "Epoch [   12/   20] | d_loss: 0.9978 | g_loss: 1.9064\n",
      "Epoch [   12/   20] | d_loss: 0.5720 | g_loss: 2.2358\n",
      "Epoch [   12/   20] | d_loss: 1.0557 | g_loss: 1.2569\n",
      "Epoch [   12/   20] | d_loss: 1.0945 | g_loss: 2.3807\n",
      "Epoch [   12/   20] | d_loss: 0.3258 | g_loss: 2.5396\n",
      "Epoch [   12/   20] | d_loss: 0.4549 | g_loss: 2.2589\n",
      "Epoch [   12/   20] | d_loss: 0.7356 | g_loss: 2.1307\n",
      "Epoch [   12/   20] | d_loss: 0.6116 | g_loss: 1.6630\n",
      "Epoch [   12/   20] | d_loss: 0.6858 | g_loss: 2.1604\n",
      "Epoch [   12/   20] | d_loss: 0.7160 | g_loss: 2.7666\n",
      "Epoch [   12/   20] | d_loss: 0.7516 | g_loss: 3.5462\n",
      "Epoch [   12/   20] | d_loss: 0.5047 | g_loss: 1.8274\n",
      "Epoch [   12/   20] | d_loss: 0.5319 | g_loss: 2.6025\n",
      "Epoch [   12/   20] | d_loss: 0.7343 | g_loss: 3.8656\n",
      "Epoch [   12/   20] | d_loss: 0.6359 | g_loss: 2.5419\n",
      "Epoch [   12/   20] | d_loss: 0.4487 | g_loss: 2.7902\n",
      "Epoch [   12/   20] | d_loss: 0.5970 | g_loss: 4.0275\n",
      "Epoch [   12/   20] | d_loss: 0.6996 | g_loss: 1.8538\n",
      "Epoch [   12/   20] | d_loss: 0.3501 | g_loss: 1.5296\n",
      "Epoch [   13/   20] | d_loss: 0.5052 | g_loss: 2.9901\n",
      "Epoch [   13/   20] | d_loss: 0.1582 | g_loss: 5.3477\n",
      "Epoch [   13/   20] | d_loss: 0.6463 | g_loss: 1.6270\n",
      "Epoch [   13/   20] | d_loss: 0.5469 | g_loss: 2.2765\n",
      "Epoch [   13/   20] | d_loss: 0.6757 | g_loss: 1.6049\n",
      "Epoch [   13/   20] | d_loss: 0.4375 | g_loss: 3.0992\n",
      "Epoch [   13/   20] | d_loss: 0.3305 | g_loss: 3.2590\n",
      "Epoch [   13/   20] | d_loss: 0.3477 | g_loss: 3.0666\n",
      "Epoch [   13/   20] | d_loss: 0.8737 | g_loss: 2.0248\n",
      "Epoch [   13/   20] | d_loss: 0.3752 | g_loss: 3.3782\n",
      "Epoch [   13/   20] | d_loss: 1.5638 | g_loss: 1.2529\n",
      "Epoch [   13/   20] | d_loss: 0.3498 | g_loss: 3.0690\n",
      "Epoch [   13/   20] | d_loss: 0.9158 | g_loss: 2.6608\n",
      "Epoch [   13/   20] | d_loss: 0.3883 | g_loss: 2.0789\n",
      "Epoch [   13/   20] | d_loss: 0.2083 | g_loss: 4.0340\n",
      "Epoch [   13/   20] | d_loss: 0.3718 | g_loss: 4.0630\n",
      "Epoch [   13/   20] | d_loss: 0.5698 | g_loss: 1.2422\n",
      "Epoch [   13/   20] | d_loss: 0.3638 | g_loss: 1.1744\n",
      "Epoch [   13/   20] | d_loss: 0.4988 | g_loss: 3.1592\n",
      "Epoch [   13/   20] | d_loss: 0.6284 | g_loss: 4.0819\n",
      "Epoch [   13/   20] | d_loss: 0.7868 | g_loss: 3.6473\n",
      "Epoch [   13/   20] | d_loss: 0.3143 | g_loss: 3.9764\n",
      "Epoch [   13/   20] | d_loss: 0.1165 | g_loss: 2.5389\n",
      "Epoch [   13/   20] | d_loss: 0.8271 | g_loss: 4.6028\n",
      "Epoch [   13/   20] | d_loss: 0.3570 | g_loss: 2.5796\n",
      "Epoch [   13/   20] | d_loss: 0.1756 | g_loss: 4.5303\n",
      "Epoch [   13/   20] | d_loss: 0.6774 | g_loss: 2.2200\n",
      "Epoch [   13/   20] | d_loss: 1.0409 | g_loss: 2.6852\n",
      "Epoch [   13/   20] | d_loss: 0.3528 | g_loss: 3.3596\n",
      "Epoch [   13/   20] | d_loss: 0.5760 | g_loss: 3.2863\n",
      "Epoch [   13/   20] | d_loss: 0.7227 | g_loss: 3.4669\n",
      "Epoch [   13/   20] | d_loss: 0.4321 | g_loss: 2.9353\n",
      "Epoch [   13/   20] | d_loss: 0.7759 | g_loss: 2.5830\n",
      "Epoch [   13/   20] | d_loss: 0.3190 | g_loss: 3.6379\n",
      "Epoch [   13/   20] | d_loss: 0.4547 | g_loss: 3.0247\n",
      "Epoch [   13/   20] | d_loss: 0.2889 | g_loss: 3.0395\n",
      "Epoch [   13/   20] | d_loss: 0.7075 | g_loss: 2.2341\n",
      "Epoch [   13/   20] | d_loss: 0.5808 | g_loss: 2.4365\n",
      "Epoch [   13/   20] | d_loss: 0.3051 | g_loss: 1.8147\n",
      "Epoch [   13/   20] | d_loss: 0.9958 | g_loss: 2.5580\n",
      "Epoch [   13/   20] | d_loss: 0.1889 | g_loss: 1.4287\n",
      "Epoch [   13/   20] | d_loss: 0.8145 | g_loss: 4.1667\n",
      "Epoch [   13/   20] | d_loss: 0.5287 | g_loss: 1.9039\n",
      "Epoch [   13/   20] | d_loss: 0.5339 | g_loss: 1.4242\n",
      "Epoch [   13/   20] | d_loss: 0.4609 | g_loss: 3.1339\n",
      "Epoch [   13/   20] | d_loss: 0.4225 | g_loss: 2.1911\n",
      "Epoch [   13/   20] | d_loss: 0.3411 | g_loss: 2.7802\n",
      "Epoch [   13/   20] | d_loss: 0.7396 | g_loss: 3.0535\n",
      "Epoch [   13/   20] | d_loss: 0.3947 | g_loss: 2.1389\n",
      "Epoch [   13/   20] | d_loss: 0.5575 | g_loss: 2.6971\n",
      "Epoch [   13/   20] | d_loss: 1.1731 | g_loss: 3.5585\n",
      "Epoch [   13/   20] | d_loss: 1.0326 | g_loss: 3.3862\n",
      "Epoch [   13/   20] | d_loss: 0.8286 | g_loss: 4.3268\n",
      "Epoch [   13/   20] | d_loss: 0.2822 | g_loss: 4.4288\n",
      "Epoch [   13/   20] | d_loss: 0.9210 | g_loss: 2.6945\n",
      "Epoch [   13/   20] | d_loss: 0.2793 | g_loss: 2.9160\n",
      "Epoch [   13/   20] | d_loss: 0.3276 | g_loss: 1.7343\n",
      "Epoch [   14/   20] | d_loss: 0.2150 | g_loss: 2.5084\n",
      "Epoch [   14/   20] | d_loss: 0.4751 | g_loss: 2.9144\n",
      "Epoch [   14/   20] | d_loss: 0.4036 | g_loss: 3.1603\n",
      "Epoch [   14/   20] | d_loss: 1.4364 | g_loss: 2.1780\n",
      "Epoch [   14/   20] | d_loss: 0.2745 | g_loss: 3.1182\n",
      "Epoch [   14/   20] | d_loss: 1.0153 | g_loss: 3.0353\n",
      "Epoch [   14/   20] | d_loss: 0.4414 | g_loss: 2.8629\n",
      "Epoch [   14/   20] | d_loss: 0.6750 | g_loss: 3.4449\n",
      "Epoch [   14/   20] | d_loss: 0.3522 | g_loss: 3.5629\n",
      "Epoch [   14/   20] | d_loss: 0.4980 | g_loss: 2.2709\n",
      "Epoch [   14/   20] | d_loss: 0.2709 | g_loss: 3.6476\n",
      "Epoch [   14/   20] | d_loss: 0.4454 | g_loss: 4.6222\n",
      "Epoch [   14/   20] | d_loss: 0.4479 | g_loss: 1.9565\n",
      "Epoch [   14/   20] | d_loss: 0.5015 | g_loss: 3.7076\n",
      "Epoch [   14/   20] | d_loss: 0.1415 | g_loss: 4.6265\n",
      "Epoch [   14/   20] | d_loss: 0.1170 | g_loss: 2.9106\n",
      "Epoch [   14/   20] | d_loss: 0.3726 | g_loss: 3.7515\n",
      "Epoch [   14/   20] | d_loss: 0.7650 | g_loss: 2.1120\n",
      "Epoch [   14/   20] | d_loss: 0.4260 | g_loss: 4.2321\n",
      "Epoch [   14/   20] | d_loss: 0.5546 | g_loss: 2.9335\n",
      "Epoch [   14/   20] | d_loss: 0.5699 | g_loss: 2.0892\n",
      "Epoch [   14/   20] | d_loss: 0.9664 | g_loss: 3.6838\n",
      "Epoch [   14/   20] | d_loss: 0.6951 | g_loss: 1.6049\n",
      "Epoch [   14/   20] | d_loss: 1.7363 | g_loss: 3.1671\n",
      "Epoch [   14/   20] | d_loss: 0.8345 | g_loss: 0.9223\n",
      "Epoch [   14/   20] | d_loss: 0.2090 | g_loss: 3.4690\n",
      "Epoch [   14/   20] | d_loss: 0.1456 | g_loss: 3.7507\n",
      "Epoch [   14/   20] | d_loss: 1.3355 | g_loss: 3.1235\n",
      "Epoch [   14/   20] | d_loss: 1.2849 | g_loss: 2.6276\n",
      "Epoch [   14/   20] | d_loss: 0.3135 | g_loss: 3.4690\n",
      "Epoch [   14/   20] | d_loss: 0.5842 | g_loss: 1.7042\n",
      "Epoch [   14/   20] | d_loss: 0.5112 | g_loss: 3.9145\n",
      "Epoch [   14/   20] | d_loss: 1.6617 | g_loss: 2.9703\n",
      "Epoch [   14/   20] | d_loss: 0.6670 | g_loss: 2.6640\n",
      "Epoch [   14/   20] | d_loss: 0.4874 | g_loss: 3.8132\n",
      "Epoch [   14/   20] | d_loss: 0.4909 | g_loss: 2.4523\n",
      "Epoch [   14/   20] | d_loss: 0.5023 | g_loss: 3.5520\n",
      "Epoch [   14/   20] | d_loss: 0.3923 | g_loss: 2.0686\n",
      "Epoch [   14/   20] | d_loss: 1.0642 | g_loss: 1.5426\n",
      "Epoch [   14/   20] | d_loss: 0.5764 | g_loss: 3.0971\n",
      "Epoch [   14/   20] | d_loss: 0.7699 | g_loss: 5.4917\n",
      "Epoch [   14/   20] | d_loss: 0.4178 | g_loss: 3.0044\n",
      "Epoch [   14/   20] | d_loss: 0.3653 | g_loss: 2.5822\n",
      "Epoch [   14/   20] | d_loss: 0.6308 | g_loss: 1.7270\n",
      "Epoch [   14/   20] | d_loss: 0.4135 | g_loss: 2.0344\n",
      "Epoch [   14/   20] | d_loss: 0.3799 | g_loss: 1.4233\n",
      "Epoch [   14/   20] | d_loss: 0.4880 | g_loss: 3.4976\n",
      "Epoch [   14/   20] | d_loss: 0.5197 | g_loss: 3.3757\n",
      "Epoch [   14/   20] | d_loss: 0.3750 | g_loss: 2.1773\n",
      "Epoch [   14/   20] | d_loss: 1.2197 | g_loss: 3.0680\n",
      "Epoch [   14/   20] | d_loss: 0.5284 | g_loss: 2.6431\n",
      "Epoch [   14/   20] | d_loss: 0.6339 | g_loss: 2.8344\n",
      "Epoch [   14/   20] | d_loss: 0.6328 | g_loss: 2.5003\n",
      "Epoch [   14/   20] | d_loss: 0.6863 | g_loss: 3.3747\n",
      "Epoch [   14/   20] | d_loss: 0.5109 | g_loss: 3.8936\n",
      "Epoch [   14/   20] | d_loss: 0.1285 | g_loss: 3.1910\n",
      "Epoch [   14/   20] | d_loss: 0.2712 | g_loss: 3.4546\n",
      "Epoch [   15/   20] | d_loss: 1.8745 | g_loss: 2.0604\n",
      "Epoch [   15/   20] | d_loss: 0.8666 | g_loss: 2.1968\n",
      "Epoch [   15/   20] | d_loss: 0.2661 | g_loss: 2.1316\n",
      "Epoch [   15/   20] | d_loss: 0.6632 | g_loss: 2.7431\n",
      "Epoch [   15/   20] | d_loss: 0.1692 | g_loss: 3.1074\n",
      "Epoch [   15/   20] | d_loss: 0.3487 | g_loss: 2.9478\n",
      "Epoch [   15/   20] | d_loss: 0.2881 | g_loss: 4.0488\n",
      "Epoch [   15/   20] | d_loss: 0.6090 | g_loss: 3.6049\n",
      "Epoch [   15/   20] | d_loss: 0.4803 | g_loss: 2.6483\n",
      "Epoch [   15/   20] | d_loss: 0.2600 | g_loss: 4.5556\n",
      "Epoch [   15/   20] | d_loss: 0.5773 | g_loss: 2.7321\n",
      "Epoch [   15/   20] | d_loss: 0.2925 | g_loss: 3.0850\n",
      "Epoch [   15/   20] | d_loss: 0.3672 | g_loss: 2.2643\n",
      "Epoch [   15/   20] | d_loss: 0.3367 | g_loss: 2.9413\n",
      "Epoch [   15/   20] | d_loss: 0.8611 | g_loss: 2.0228\n",
      "Epoch [   15/   20] | d_loss: 0.2366 | g_loss: 1.7442\n",
      "Epoch [   15/   20] | d_loss: 0.5192 | g_loss: 3.8010\n",
      "Epoch [   15/   20] | d_loss: 0.5787 | g_loss: 3.7651\n",
      "Epoch [   15/   20] | d_loss: 1.0771 | g_loss: 3.1846\n",
      "Epoch [   15/   20] | d_loss: 1.3017 | g_loss: 1.8112\n",
      "Epoch [   15/   20] | d_loss: 0.5342 | g_loss: 1.2120\n",
      "Epoch [   15/   20] | d_loss: 0.4864 | g_loss: 3.6415\n",
      "Epoch [   15/   20] | d_loss: 1.8285 | g_loss: 1.6970\n",
      "Epoch [   15/   20] | d_loss: 0.1617 | g_loss: 3.8537\n",
      "Epoch [   15/   20] | d_loss: 0.3465 | g_loss: 2.0365\n",
      "Epoch [   15/   20] | d_loss: 0.3282 | g_loss: 2.3535\n",
      "Epoch [   15/   20] | d_loss: 0.5605 | g_loss: 5.4110\n",
      "Epoch [   15/   20] | d_loss: 0.3868 | g_loss: 4.3645\n",
      "Epoch [   15/   20] | d_loss: 1.3304 | g_loss: 2.5072\n",
      "Epoch [   15/   20] | d_loss: 0.1904 | g_loss: 2.9486\n",
      "Epoch [   15/   20] | d_loss: 0.2447 | g_loss: 2.4846\n",
      "Epoch [   15/   20] | d_loss: 1.3208 | g_loss: 2.5365\n",
      "Epoch [   15/   20] | d_loss: 0.7969 | g_loss: 3.6624\n",
      "Epoch [   15/   20] | d_loss: 0.2877 | g_loss: 3.0152\n",
      "Epoch [   15/   20] | d_loss: 0.5462 | g_loss: 2.7660\n",
      "Epoch [   15/   20] | d_loss: 0.6912 | g_loss: 4.1604\n",
      "Epoch [   15/   20] | d_loss: 0.3643 | g_loss: 2.2048\n",
      "Epoch [   15/   20] | d_loss: 0.1246 | g_loss: 1.3622\n",
      "Epoch [   15/   20] | d_loss: 0.2685 | g_loss: 3.2298\n",
      "Epoch [   15/   20] | d_loss: 0.6315 | g_loss: 2.5257\n",
      "Epoch [   15/   20] | d_loss: 0.4473 | g_loss: 3.3083\n",
      "Epoch [   15/   20] | d_loss: 0.8399 | g_loss: 4.1693\n",
      "Epoch [   15/   20] | d_loss: 1.0466 | g_loss: 2.7749\n",
      "Epoch [   15/   20] | d_loss: 0.2834 | g_loss: 2.2464\n",
      "Epoch [   15/   20] | d_loss: 0.5668 | g_loss: 2.6733\n",
      "Epoch [   15/   20] | d_loss: 0.1549 | g_loss: 2.9009\n",
      "Epoch [   15/   20] | d_loss: 0.8974 | g_loss: 1.5959\n",
      "Epoch [   15/   20] | d_loss: 0.1873 | g_loss: 3.2647\n",
      "Epoch [   15/   20] | d_loss: 0.2624 | g_loss: 2.7383\n",
      "Epoch [   15/   20] | d_loss: 0.6819 | g_loss: 4.3051\n",
      "Epoch [   15/   20] | d_loss: 0.3068 | g_loss: 4.0918\n",
      "Epoch [   15/   20] | d_loss: 0.7916 | g_loss: 3.8525\n",
      "Epoch [   15/   20] | d_loss: 1.2716 | g_loss: 2.4977\n",
      "Epoch [   15/   20] | d_loss: 1.0534 | g_loss: 2.4354\n",
      "Epoch [   15/   20] | d_loss: 0.3834 | g_loss: 3.3407\n",
      "Epoch [   15/   20] | d_loss: 1.8051 | g_loss: 1.8855\n",
      "Epoch [   15/   20] | d_loss: 0.5393 | g_loss: 3.2239\n",
      "Epoch [   16/   20] | d_loss: 1.1367 | g_loss: 3.0839\n",
      "Epoch [   16/   20] | d_loss: 0.1870 | g_loss: 2.2303\n",
      "Epoch [   16/   20] | d_loss: 0.5383 | g_loss: 3.6566\n",
      "Epoch [   16/   20] | d_loss: 0.5540 | g_loss: 2.3569\n",
      "Epoch [   16/   20] | d_loss: 0.2753 | g_loss: 4.3035\n",
      "Epoch [   16/   20] | d_loss: 1.0860 | g_loss: 4.8740\n",
      "Epoch [   16/   20] | d_loss: 0.2325 | g_loss: 4.8985\n",
      "Epoch [   16/   20] | d_loss: 0.7863 | g_loss: 4.1471\n",
      "Epoch [   16/   20] | d_loss: 1.5021 | g_loss: 2.6553\n",
      "Epoch [   16/   20] | d_loss: 1.4822 | g_loss: 1.7313\n",
      "Epoch [   16/   20] | d_loss: 0.8421 | g_loss: 2.5684\n",
      "Epoch [   16/   20] | d_loss: 0.3936 | g_loss: 2.8833\n",
      "Epoch [   16/   20] | d_loss: 0.8793 | g_loss: 3.4795\n",
      "Epoch [   16/   20] | d_loss: 0.2924 | g_loss: 5.3465\n",
      "Epoch [   16/   20] | d_loss: 0.8217 | g_loss: 2.9469\n",
      "Epoch [   16/   20] | d_loss: 0.2830 | g_loss: 5.3882\n",
      "Epoch [   16/   20] | d_loss: 0.5410 | g_loss: 2.9535\n",
      "Epoch [   16/   20] | d_loss: 1.1822 | g_loss: 2.7612\n",
      "Epoch [   16/   20] | d_loss: 0.4151 | g_loss: 2.4524\n",
      "Epoch [   16/   20] | d_loss: 0.8099 | g_loss: 2.5348\n",
      "Epoch [   16/   20] | d_loss: 0.4076 | g_loss: 4.6127\n",
      "Epoch [   16/   20] | d_loss: 0.5299 | g_loss: 3.3582\n",
      "Epoch [   16/   20] | d_loss: 0.4243 | g_loss: 3.7712\n",
      "Epoch [   16/   20] | d_loss: 0.7053 | g_loss: 3.2301\n",
      "Epoch [   16/   20] | d_loss: 0.1635 | g_loss: 3.7155\n",
      "Epoch [   16/   20] | d_loss: 0.1631 | g_loss: 3.2329\n",
      "Epoch [   16/   20] | d_loss: 0.5871 | g_loss: 3.8967\n",
      "Epoch [   16/   20] | d_loss: 1.3536 | g_loss: 2.1056\n",
      "Epoch [   16/   20] | d_loss: 0.1749 | g_loss: 5.2145\n",
      "Epoch [   16/   20] | d_loss: 0.1603 | g_loss: 3.5665\n",
      "Epoch [   16/   20] | d_loss: 0.3129 | g_loss: 2.6796\n",
      "Epoch [   16/   20] | d_loss: 0.3769 | g_loss: 1.4772\n",
      "Epoch [   16/   20] | d_loss: 0.3721 | g_loss: 4.6049\n",
      "Epoch [   16/   20] | d_loss: 0.7433 | g_loss: 3.3960\n",
      "Epoch [   16/   20] | d_loss: 0.5157 | g_loss: 2.7458\n",
      "Epoch [   16/   20] | d_loss: 0.1865 | g_loss: 2.6377\n",
      "Epoch [   16/   20] | d_loss: 0.2248 | g_loss: 2.9934\n",
      "Epoch [   16/   20] | d_loss: 0.9851 | g_loss: 3.6754\n",
      "Epoch [   16/   20] | d_loss: 0.2095 | g_loss: 2.2703\n",
      "Epoch [   16/   20] | d_loss: 0.4689 | g_loss: 2.0052\n",
      "Epoch [   16/   20] | d_loss: 0.2471 | g_loss: 1.8647\n",
      "Epoch [   16/   20] | d_loss: 0.4751 | g_loss: 2.4651\n",
      "Epoch [   16/   20] | d_loss: 0.2895 | g_loss: 5.4340\n",
      "Epoch [   16/   20] | d_loss: 0.2174 | g_loss: 4.1344\n",
      "Epoch [   16/   20] | d_loss: 1.5366 | g_loss: 1.4559\n",
      "Epoch [   16/   20] | d_loss: 0.8584 | g_loss: 3.6011\n",
      "Epoch [   16/   20] | d_loss: 0.4116 | g_loss: 3.4179\n",
      "Epoch [   16/   20] | d_loss: 0.4849 | g_loss: 3.7818\n",
      "Epoch [   16/   20] | d_loss: 0.2928 | g_loss: 3.9792\n",
      "Epoch [   16/   20] | d_loss: 0.6461 | g_loss: 0.9797\n",
      "Epoch [   16/   20] | d_loss: 0.5278 | g_loss: 3.8382\n",
      "Epoch [   16/   20] | d_loss: 0.2479 | g_loss: 2.9809\n",
      "Epoch [   16/   20] | d_loss: 0.6783 | g_loss: 2.2310\n",
      "Epoch [   16/   20] | d_loss: 0.5333 | g_loss: 3.3267\n",
      "Epoch [   16/   20] | d_loss: 0.2436 | g_loss: 3.6854\n",
      "Epoch [   16/   20] | d_loss: 0.5647 | g_loss: 2.2477\n",
      "Epoch [   16/   20] | d_loss: 0.4890 | g_loss: 4.4853\n",
      "Epoch [   17/   20] | d_loss: 0.5791 | g_loss: 2.4526\n",
      "Epoch [   17/   20] | d_loss: 0.3977 | g_loss: 2.3548\n",
      "Epoch [   17/   20] | d_loss: 0.1539 | g_loss: 4.1298\n",
      "Epoch [   17/   20] | d_loss: 0.4158 | g_loss: 4.0752\n",
      "Epoch [   17/   20] | d_loss: 0.6086 | g_loss: 3.1411\n",
      "Epoch [   17/   20] | d_loss: 0.2098 | g_loss: 3.0539\n",
      "Epoch [   17/   20] | d_loss: 0.1936 | g_loss: 4.3995\n",
      "Epoch [   17/   20] | d_loss: 1.3422 | g_loss: 4.2124\n",
      "Epoch [   17/   20] | d_loss: 0.4468 | g_loss: 1.6185\n",
      "Epoch [   17/   20] | d_loss: 0.5136 | g_loss: 4.8884\n",
      "Epoch [   17/   20] | d_loss: 0.2964 | g_loss: 3.6467\n",
      "Epoch [   17/   20] | d_loss: 0.4011 | g_loss: 5.4907\n",
      "Epoch [   17/   20] | d_loss: 0.1338 | g_loss: 1.6553\n",
      "Epoch [   17/   20] | d_loss: 0.8503 | g_loss: 1.0404\n",
      "Epoch [   17/   20] | d_loss: 0.6597 | g_loss: 2.6653\n",
      "Epoch [   17/   20] | d_loss: 0.2098 | g_loss: 2.5211\n",
      "Epoch [   17/   20] | d_loss: 0.4750 | g_loss: 2.0425\n",
      "Epoch [   17/   20] | d_loss: 0.6994 | g_loss: 2.7088\n",
      "Epoch [   17/   20] | d_loss: 0.3278 | g_loss: 3.4429\n",
      "Epoch [   17/   20] | d_loss: 0.4648 | g_loss: 3.7240\n",
      "Epoch [   17/   20] | d_loss: 0.0772 | g_loss: 4.8854\n",
      "Epoch [   17/   20] | d_loss: 0.6631 | g_loss: 3.5875\n",
      "Epoch [   17/   20] | d_loss: 0.5326 | g_loss: 3.6622\n",
      "Epoch [   17/   20] | d_loss: 1.9927 | g_loss: 4.6967\n",
      "Epoch [   17/   20] | d_loss: 0.5979 | g_loss: 3.3989\n",
      "Epoch [   17/   20] | d_loss: 0.3342 | g_loss: 3.1233\n",
      "Epoch [   17/   20] | d_loss: 1.0004 | g_loss: 3.2585\n",
      "Epoch [   17/   20] | d_loss: 0.3587 | g_loss: 4.4636\n",
      "Epoch [   17/   20] | d_loss: 0.7425 | g_loss: 3.1432\n",
      "Epoch [   17/   20] | d_loss: 0.4988 | g_loss: 2.7217\n",
      "Epoch [   17/   20] | d_loss: 0.1072 | g_loss: 3.1718\n",
      "Epoch [   17/   20] | d_loss: 0.8891 | g_loss: 3.5053\n",
      "Epoch [   17/   20] | d_loss: 0.1762 | g_loss: 4.6673\n",
      "Epoch [   17/   20] | d_loss: 0.8806 | g_loss: 3.2503\n",
      "Epoch [   17/   20] | d_loss: 1.2844 | g_loss: 5.1032\n",
      "Epoch [   17/   20] | d_loss: 0.1670 | g_loss: 4.4831\n",
      "Epoch [   17/   20] | d_loss: 0.5446 | g_loss: 3.9622\n",
      "Epoch [   17/   20] | d_loss: 0.6424 | g_loss: 3.3785\n",
      "Epoch [   17/   20] | d_loss: 0.5902 | g_loss: 3.7560\n",
      "Epoch [   17/   20] | d_loss: 0.1688 | g_loss: 4.6554\n",
      "Epoch [   17/   20] | d_loss: 0.5734 | g_loss: 4.4842\n",
      "Epoch [   17/   20] | d_loss: 0.3221 | g_loss: 4.0637\n",
      "Epoch [   17/   20] | d_loss: 0.6296 | g_loss: 2.8143\n",
      "Epoch [   17/   20] | d_loss: 0.3576 | g_loss: 1.9698\n",
      "Epoch [   17/   20] | d_loss: 0.5395 | g_loss: 3.2650\n",
      "Epoch [   17/   20] | d_loss: 0.3544 | g_loss: 3.3467\n",
      "Epoch [   17/   20] | d_loss: 0.8010 | g_loss: 3.4289\n",
      "Epoch [   17/   20] | d_loss: 0.6473 | g_loss: 2.4779\n",
      "Epoch [   17/   20] | d_loss: 0.2805 | g_loss: 3.4608\n",
      "Epoch [   17/   20] | d_loss: 0.2314 | g_loss: 4.2833\n",
      "Epoch [   17/   20] | d_loss: 1.0289 | g_loss: 6.3001\n",
      "Epoch [   17/   20] | d_loss: 0.3752 | g_loss: 1.5592\n",
      "Epoch [   17/   20] | d_loss: 1.7678 | g_loss: 0.9382\n",
      "Epoch [   17/   20] | d_loss: 0.8249 | g_loss: 2.1459\n",
      "Epoch [   17/   20] | d_loss: 0.6236 | g_loss: 3.2759\n",
      "Epoch [   17/   20] | d_loss: 0.3898 | g_loss: 4.3260\n",
      "Epoch [   17/   20] | d_loss: 0.3394 | g_loss: 6.0246\n",
      "Epoch [   18/   20] | d_loss: 0.3372 | g_loss: 2.3228\n",
      "Epoch [   18/   20] | d_loss: 0.5600 | g_loss: 3.5574\n",
      "Epoch [   18/   20] | d_loss: 0.7562 | g_loss: 6.5689\n",
      "Epoch [   18/   20] | d_loss: 0.7085 | g_loss: 3.4843\n",
      "Epoch [   18/   20] | d_loss: 0.4516 | g_loss: 4.2816\n",
      "Epoch [   18/   20] | d_loss: 0.4466 | g_loss: 2.7104\n",
      "Epoch [   18/   20] | d_loss: 0.2561 | g_loss: 2.5692\n",
      "Epoch [   18/   20] | d_loss: 1.0714 | g_loss: 2.4108\n",
      "Epoch [   18/   20] | d_loss: 0.9909 | g_loss: 4.6758\n",
      "Epoch [   18/   20] | d_loss: 0.2873 | g_loss: 1.5321\n",
      "Epoch [   18/   20] | d_loss: 0.1454 | g_loss: 2.5499\n",
      "Epoch [   18/   20] | d_loss: 0.2713 | g_loss: 1.3939\n",
      "Epoch [   18/   20] | d_loss: 0.4256 | g_loss: 3.7501\n",
      "Epoch [   18/   20] | d_loss: 0.3309 | g_loss: 2.9787\n",
      "Epoch [   18/   20] | d_loss: 0.4518 | g_loss: 2.8266\n",
      "Epoch [   18/   20] | d_loss: 0.5340 | g_loss: 1.3738\n",
      "Epoch [   18/   20] | d_loss: 0.4166 | g_loss: 5.6747\n",
      "Epoch [   18/   20] | d_loss: 1.2734 | g_loss: 2.4567\n",
      "Epoch [   18/   20] | d_loss: 0.1843 | g_loss: 6.1309\n",
      "Epoch [   18/   20] | d_loss: 0.3986 | g_loss: 6.7937\n",
      "Epoch [   18/   20] | d_loss: 2.1894 | g_loss: 4.2207\n",
      "Epoch [   18/   20] | d_loss: 0.8674 | g_loss: 4.4210\n",
      "Epoch [   18/   20] | d_loss: 1.1122 | g_loss: 2.8779\n",
      "Epoch [   18/   20] | d_loss: 0.4538 | g_loss: 3.0053\n",
      "Epoch [   18/   20] | d_loss: 0.5457 | g_loss: 4.8049\n",
      "Epoch [   18/   20] | d_loss: 0.8015 | g_loss: 3.3846\n",
      "Epoch [   18/   20] | d_loss: 0.3323 | g_loss: 3.0479\n",
      "Epoch [   18/   20] | d_loss: 0.2919 | g_loss: 3.6163\n",
      "Epoch [   18/   20] | d_loss: 0.1969 | g_loss: 5.7533\n",
      "Epoch [   18/   20] | d_loss: 0.7426 | g_loss: 2.0302\n",
      "Epoch [   18/   20] | d_loss: 0.6758 | g_loss: 2.4861\n",
      "Epoch [   18/   20] | d_loss: 0.5966 | g_loss: 2.1081\n",
      "Epoch [   18/   20] | d_loss: 0.4752 | g_loss: 4.8552\n",
      "Epoch [   18/   20] | d_loss: 0.5251 | g_loss: 3.5692\n",
      "Epoch [   18/   20] | d_loss: 0.8170 | g_loss: 2.1353\n",
      "Epoch [   18/   20] | d_loss: 0.7981 | g_loss: 2.6688\n",
      "Epoch [   18/   20] | d_loss: 0.2946 | g_loss: 1.4616\n",
      "Epoch [   18/   20] | d_loss: 0.4190 | g_loss: 5.1772\n",
      "Epoch [   18/   20] | d_loss: 0.7084 | g_loss: 1.3351\n",
      "Epoch [   18/   20] | d_loss: 0.1718 | g_loss: 3.3993\n",
      "Epoch [   18/   20] | d_loss: 0.2914 | g_loss: 4.5578\n",
      "Epoch [   18/   20] | d_loss: 0.3680 | g_loss: 4.8393\n",
      "Epoch [   18/   20] | d_loss: 0.3229 | g_loss: 4.9326\n",
      "Epoch [   18/   20] | d_loss: 0.7616 | g_loss: 2.2741\n",
      "Epoch [   18/   20] | d_loss: 0.0797 | g_loss: 2.2263\n",
      "Epoch [   18/   20] | d_loss: 0.6586 | g_loss: 3.8987\n",
      "Epoch [   18/   20] | d_loss: 0.2659 | g_loss: 4.3480\n",
      "Epoch [   18/   20] | d_loss: 0.3999 | g_loss: 2.7061\n",
      "Epoch [   18/   20] | d_loss: 0.1940 | g_loss: 3.1596\n",
      "Epoch [   18/   20] | d_loss: 0.6110 | g_loss: 3.6212\n",
      "Epoch [   18/   20] | d_loss: 0.8273 | g_loss: 3.3628\n",
      "Epoch [   18/   20] | d_loss: 1.2120 | g_loss: 1.9435\n",
      "Epoch [   18/   20] | d_loss: 0.3773 | g_loss: 2.6033\n",
      "Epoch [   18/   20] | d_loss: 0.5303 | g_loss: 3.4905\n",
      "Epoch [   18/   20] | d_loss: 0.4487 | g_loss: 1.7776\n",
      "Epoch [   18/   20] | d_loss: 0.4145 | g_loss: 4.1932\n",
      "Epoch [   18/   20] | d_loss: 0.3013 | g_loss: 3.1579\n",
      "Epoch [   19/   20] | d_loss: 0.4948 | g_loss: 2.4684\n",
      "Epoch [   19/   20] | d_loss: 0.2236 | g_loss: 5.0732\n",
      "Epoch [   19/   20] | d_loss: 0.0984 | g_loss: 4.1145\n",
      "Epoch [   19/   20] | d_loss: 0.5497 | g_loss: 2.3979\n",
      "Epoch [   19/   20] | d_loss: 0.5860 | g_loss: 4.9707\n",
      "Epoch [   19/   20] | d_loss: 0.2866 | g_loss: 6.1648\n",
      "Epoch [   19/   20] | d_loss: 0.3430 | g_loss: 5.1710\n",
      "Epoch [   19/   20] | d_loss: 0.5842 | g_loss: 3.9019\n",
      "Epoch [   19/   20] | d_loss: 0.4113 | g_loss: 3.6769\n",
      "Epoch [   19/   20] | d_loss: 0.6004 | g_loss: 3.6218\n",
      "Epoch [   19/   20] | d_loss: 0.3982 | g_loss: 3.5737\n",
      "Epoch [   19/   20] | d_loss: 0.2159 | g_loss: 3.4249\n",
      "Epoch [   19/   20] | d_loss: 0.2559 | g_loss: 5.5986\n",
      "Epoch [   19/   20] | d_loss: 1.0554 | g_loss: 5.3484\n",
      "Epoch [   19/   20] | d_loss: 0.2667 | g_loss: 4.1424\n",
      "Epoch [   19/   20] | d_loss: 0.2787 | g_loss: 6.0069\n",
      "Epoch [   19/   20] | d_loss: 0.4357 | g_loss: 3.9492\n",
      "Epoch [   19/   20] | d_loss: 0.4475 | g_loss: 4.9536\n",
      "Epoch [   19/   20] | d_loss: 0.2469 | g_loss: 2.8115\n",
      "Epoch [   19/   20] | d_loss: 0.4499 | g_loss: 6.1471\n",
      "Epoch [   19/   20] | d_loss: 0.1631 | g_loss: 3.3068\n",
      "Epoch [   19/   20] | d_loss: 0.1893 | g_loss: 4.9132\n",
      "Epoch [   19/   20] | d_loss: 0.8203 | g_loss: 4.8630\n",
      "Epoch [   19/   20] | d_loss: 0.4024 | g_loss: 3.1362\n",
      "Epoch [   19/   20] | d_loss: 0.5925 | g_loss: 4.4068\n",
      "Epoch [   19/   20] | d_loss: 0.4223 | g_loss: 2.4464\n",
      "Epoch [   19/   20] | d_loss: 0.6343 | g_loss: 6.3159\n",
      "Epoch [   19/   20] | d_loss: 0.2495 | g_loss: 5.4443\n",
      "Epoch [   19/   20] | d_loss: 0.2499 | g_loss: 3.6054\n",
      "Epoch [   19/   20] | d_loss: 0.4256 | g_loss: 3.1441\n",
      "Epoch [   19/   20] | d_loss: 1.8616 | g_loss: 1.7940\n",
      "Epoch [   19/   20] | d_loss: 0.3122 | g_loss: 2.9019\n",
      "Epoch [   19/   20] | d_loss: 0.1841 | g_loss: 4.4494\n",
      "Epoch [   19/   20] | d_loss: 0.4356 | g_loss: 3.8039\n",
      "Epoch [   19/   20] | d_loss: 1.1426 | g_loss: 3.6802\n",
      "Epoch [   19/   20] | d_loss: 0.2204 | g_loss: 4.1163\n",
      "Epoch [   19/   20] | d_loss: 0.3729 | g_loss: 5.6955\n",
      "Epoch [   19/   20] | d_loss: 0.1305 | g_loss: 5.5874\n",
      "Epoch [   19/   20] | d_loss: 0.2777 | g_loss: 2.0761\n",
      "Epoch [   19/   20] | d_loss: 1.0718 | g_loss: 2.5484\n",
      "Epoch [   19/   20] | d_loss: 0.4874 | g_loss: 4.2115\n",
      "Epoch [   19/   20] | d_loss: 0.8677 | g_loss: 4.1490\n",
      "Epoch [   19/   20] | d_loss: 0.0909 | g_loss: 4.3333\n",
      "Epoch [   19/   20] | d_loss: 0.3893 | g_loss: 2.9384\n",
      "Epoch [   19/   20] | d_loss: 0.3660 | g_loss: 2.9343\n",
      "Epoch [   19/   20] | d_loss: 0.8736 | g_loss: 3.7674\n",
      "Epoch [   19/   20] | d_loss: 0.1557 | g_loss: 2.3395\n",
      "Epoch [   19/   20] | d_loss: 0.2781 | g_loss: 3.4364\n",
      "Epoch [   19/   20] | d_loss: 0.5974 | g_loss: 2.8561\n",
      "Epoch [   19/   20] | d_loss: 0.9828 | g_loss: 3.5645\n",
      "Epoch [   19/   20] | d_loss: 0.3398 | g_loss: 3.0752\n",
      "Epoch [   19/   20] | d_loss: 0.4315 | g_loss: 2.9873\n",
      "Epoch [   19/   20] | d_loss: 0.0744 | g_loss: 7.4433\n",
      "Epoch [   19/   20] | d_loss: 0.3304 | g_loss: 5.2491\n",
      "Epoch [   19/   20] | d_loss: 0.2050 | g_loss: 2.6203\n",
      "Epoch [   19/   20] | d_loss: 0.0415 | g_loss: 6.3392\n",
      "Epoch [   19/   20] | d_loss: 0.1261 | g_loss: 4.1938\n",
      "Epoch [   20/   20] | d_loss: 0.1126 | g_loss: 2.0411\n",
      "Epoch [   20/   20] | d_loss: 0.4336 | g_loss: 5.6304\n",
      "Epoch [   20/   20] | d_loss: 0.3595 | g_loss: 4.3063\n",
      "Epoch [   20/   20] | d_loss: 0.4526 | g_loss: 2.9371\n",
      "Epoch [   20/   20] | d_loss: 0.0543 | g_loss: 3.2896\n",
      "Epoch [   20/   20] | d_loss: 0.1970 | g_loss: 2.6817\n",
      "Epoch [   20/   20] | d_loss: 0.8817 | g_loss: 3.8067\n",
      "Epoch [   20/   20] | d_loss: 0.0776 | g_loss: 5.0047\n",
      "Epoch [   20/   20] | d_loss: 0.1381 | g_loss: 4.4845\n",
      "Epoch [   20/   20] | d_loss: 0.3101 | g_loss: 3.0619\n",
      "Epoch [   20/   20] | d_loss: 0.2772 | g_loss: 3.4917\n",
      "Epoch [   20/   20] | d_loss: 0.3407 | g_loss: 4.2575\n",
      "Epoch [   20/   20] | d_loss: 0.3067 | g_loss: 4.8697\n",
      "Epoch [   20/   20] | d_loss: 0.7416 | g_loss: 3.8586\n",
      "Epoch [   20/   20] | d_loss: 0.1356 | g_loss: 3.8204\n",
      "Epoch [   20/   20] | d_loss: 0.4814 | g_loss: 5.6185\n",
      "Epoch [   20/   20] | d_loss: 0.6992 | g_loss: 2.6296\n",
      "Epoch [   20/   20] | d_loss: 0.1766 | g_loss: 2.5708\n",
      "Epoch [   20/   20] | d_loss: 0.3053 | g_loss: 4.7382\n",
      "Epoch [   20/   20] | d_loss: 0.5595 | g_loss: 4.7389\n",
      "Epoch [   20/   20] | d_loss: 0.5097 | g_loss: 4.4475\n",
      "Epoch [   20/   20] | d_loss: 0.4714 | g_loss: 3.6301\n",
      "Epoch [   20/   20] | d_loss: 0.3857 | g_loss: 5.8430\n",
      "Epoch [   20/   20] | d_loss: 0.0415 | g_loss: 4.0293\n",
      "Epoch [   20/   20] | d_loss: 0.8463 | g_loss: 3.5246\n",
      "Epoch [   20/   20] | d_loss: 0.1628 | g_loss: 4.0457\n",
      "Epoch [   20/   20] | d_loss: 0.4506 | g_loss: 0.0717\n",
      "Epoch [   20/   20] | d_loss: 0.4321 | g_loss: 2.9666\n",
      "Epoch [   20/   20] | d_loss: 0.4729 | g_loss: 4.1431\n",
      "Epoch [   20/   20] | d_loss: 0.6503 | g_loss: 2.3635\n",
      "Epoch [   20/   20] | d_loss: 0.3945 | g_loss: 3.7511\n",
      "Epoch [   20/   20] | d_loss: 0.2956 | g_loss: 4.1623\n",
      "Epoch [   20/   20] | d_loss: 0.1887 | g_loss: 1.7419\n",
      "Epoch [   20/   20] | d_loss: 0.2655 | g_loss: 4.6362\n",
      "Epoch [   20/   20] | d_loss: 0.1297 | g_loss: 2.5390\n",
      "Epoch [   20/   20] | d_loss: 0.5024 | g_loss: 3.8328\n",
      "Epoch [   20/   20] | d_loss: 0.4999 | g_loss: 2.2111\n",
      "Epoch [   20/   20] | d_loss: 0.2859 | g_loss: 2.7456\n",
      "Epoch [   20/   20] | d_loss: 0.2163 | g_loss: 4.4723\n",
      "Epoch [   20/   20] | d_loss: 1.1351 | g_loss: 3.1254\n",
      "Epoch [   20/   20] | d_loss: 0.2346 | g_loss: 2.6424\n",
      "Epoch [   20/   20] | d_loss: 0.6077 | g_loss: 4.5969\n",
      "Epoch [   20/   20] | d_loss: 1.2993 | g_loss: 2.6371\n",
      "Epoch [   20/   20] | d_loss: 0.4270 | g_loss: 3.9657\n",
      "Epoch [   20/   20] | d_loss: 0.6211 | g_loss: 2.7425\n",
      "Epoch [   20/   20] | d_loss: 0.2486 | g_loss: 2.8781\n",
      "Epoch [   20/   20] | d_loss: 0.6076 | g_loss: 2.5456\n",
      "Epoch [   20/   20] | d_loss: 0.9996 | g_loss: 2.7555\n",
      "Epoch [   20/   20] | d_loss: 0.2093 | g_loss: 4.2714\n",
      "Epoch [   20/   20] | d_loss: 0.2416 | g_loss: 4.0230\n",
      "Epoch [   20/   20] | d_loss: 0.1992 | g_loss: 4.5071\n",
      "Epoch [   20/   20] | d_loss: 0.5021 | g_loss: 4.9852\n",
      "Epoch [   20/   20] | d_loss: 0.1270 | g_loss: 7.5787\n",
      "Epoch [   20/   20] | d_loss: 0.9633 | g_loss: 4.4384\n",
      "Epoch [   20/   20] | d_loss: 0.3523 | g_loss: 5.5049\n",
      "Epoch [   20/   20] | d_loss: 0.3199 | g_loss: 3.1769\n",
      "Epoch [   20/   20] | d_loss: 0.4456 | g_loss: 2.8004\n"
     ]
    }
   ],
   "source": [
    "# set number of epochs \n",
    "n_epochs = 20\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "# call training function\n",
    "losses = train(D, G, n_epochs=n_epochs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Training loss\n",
    "\n",
    "Plot the training losses for the generator and discriminator, recorded after each epoch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7efc8553f550>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7efc88050a58>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "losses = np.array(losses)\n",
    "plt.plot(losses.T[0], label='Discriminator', alpha=0.5)\n",
    "plt.plot(losses.T[1], label='Generator', alpha=0.5)\n",
    "plt.title(\"Training Losses\")\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Generator samples from training\n",
    "\n",
    "View samples of images from the generator, and answer a question about the strengths and weaknesses of your trained models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper function for viewing a list of passed in sample images\n",
    "def view_samples(epoch, samples):\n",
    "    fig, axes = plt.subplots(figsize=(16,4), nrows=2, ncols=8, sharey=True, sharex=True)\n",
    "    for ax, img in zip(axes.flatten(), samples[epoch]):\n",
    "        img = img.detach().cpu().numpy()\n",
    "        img = np.transpose(img, (1, 2, 0))\n",
    "        img = ((img + 1)*255 / (2)).astype(np.uint8)\n",
    "        ax.xaxis.set_visible(False)\n",
    "        ax.yaxis.set_visible(False)\n",
    "        im = ax.imshow(img.reshape((32,32,3)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load samples from generator, taken while training\n",
    "with open('train_samples.pkl', 'rb') as f:\n",
    "    samples = pkl.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7efc8a2fffd0>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "_ = view_samples(-1, samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question: What do you notice about your generated samples and how might you improve this model?\n",
    "When you answer this question, consider the following factors:\n",
    "* The dataset is biased; it is made of \"celebrity\" faces that are mostly white\n",
    "* Model size; larger models have the opportunity to learn more features in a data feature space\n",
    "* Optimization strategy; optimizers and number of epochs affect your final result\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:** We see that the majority of the generated faces consists of people of white color. This obviously has to do with the fact that the generator is eventually trained aiming at features that the discriminator considers real. The real images dataset predominantly consists of white faces. Also, a more complex generator network would be able to create more complex features that would resemble the ones identified in the real faces. A longer training session would also increase the generators capabilities, since as we can see the faces currently look blurry and of low resolution. This would be fixed by a more complex network and more training epochs. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting This Project\n",
    "When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as \"dlnd_face_generation.ipynb\" and save it as a HTML file under \"File\" -> \"Download as\". Include the \"problem_unittests.py\" files in your submission."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}