dlnd_face_generation3.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": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# can comment out after executing\n",
    "#!unzip processed_celeba_small.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = 'processed_celeba_small/'\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "import pickle as pkl\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import problem_unittests as tests\n",
    "#import helper\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualize the CelebA Data\n",
    "\n",
    "The [CelebA](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations, you'll only need the images. Note that these are color images with [3 color channels (RGB)](https://en.wikipedia.org/wiki/Channel_(digital_image)#RGB_Images) each.\n",
    "\n",
    "### Pre-process and Load the Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. This *pre-processed* dataset is a smaller subset of the very large CelebA data.\n",
    "\n",
    "> There are a few other steps that you'll need to **transform** this data and create a **DataLoader**.\n",
    "\n",
    "#### Exercise: Complete the following `get_dataloader` function, such that it satisfies these requirements:\n",
    "\n",
    "* Your images should be square, Tensor images of size `image_size x image_size` in the x and y dimension.\n",
    "* Your function should return a DataLoader that shuffles and batches these Tensor images.\n",
    "\n",
    "#### ImageFolder\n",
    "\n",
    "To create a dataset given a directory of images, it's recommended that you use PyTorch's [ImageFolder](https://pytorch.org/docs/stable/torchvision/datasets.html#imagefolder) wrapper, with a root directory `processed_celeba_small/` and data transformation passed in."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# necessary imports\n",
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision import transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_dataloader(batch_size, image_size, data_dir='processed_celeba_small/'):\n",
    "    \"\"\"\n",
    "    Batch the neural network data using DataLoader\n",
    "    :param batch_size: The size of each batch; the number of images in a batch\n",
    "    :param img_size: The square size of the image data (x, y)\n",
    "    :param data_dir: Directory where image data is located\n",
    "    :return: DataLoader with batched data\n",
    "    \"\"\"\n",
    "    # Define image augmentation methods\n",
    "    data_transforms = transforms.Compose([transforms.Resize(image_size),\n",
    "                                          transforms.ToTensor()])\n",
    "    \n",
    "    # Load the dataset with ImageFolder\n",
    "    dataset = datasets.ImageFolder(data_dir, transform=data_transforms)\n",
    "    \n",
    "    # Define the dataloader\n",
    "    data_loader = torch.utils.data.DataLoader(dataset, batch_size, shuffle=True)\n",
    "    \n",
    "    return data_loader\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Create a DataLoader\n",
    "\n",
    "#### Exercise: Create a DataLoader `celeba_train_loader` with appropriate hyperparameters.\n",
    "\n",
    "Call the above function and create a dataloader to view images. \n",
    "* You can decide on any reasonable `batch_size` parameter\n",
    "* Your `image_size` **must be** `32`. Resizing the data to a smaller size will make for faster training, while still creating convincing images of faces!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define function hyperparameters\n",
    "batch_size = 28\n",
    "img_size = 32\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# Call your function and get a dataloader\n",
    "celeba_train_loader = get_dataloader(batch_size, img_size)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, you can view some images! You should seen square images of somewhat-centered faces.\n",
    "\n",
    "Note: You'll need to convert the Tensor images into a NumPy type and transpose the dimensions to correctly display an image, suggested `imshow` code is below, but it may not be perfect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff38a4c1c18>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# helper display function\n",
    "def imshow(img):\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# obtain one batch of training images\n",
    "dataiter = iter(celeba_train_loader)\n",
    "images, _ = dataiter.next() # _ for no labels\n",
    "\n",
    "# plot the images in the batch, along with the corresponding labels\n",
    "fig = plt.figure(figsize=(20, 4))\n",
    "plot_size=20\n",
    "for idx in np.arange(plot_size):\n",
    "    ax = fig.add_subplot(2, plot_size/2, idx+1, xticks=[], yticks=[])\n",
    "    imshow(images[idx])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Pre-process your image data and scale it to a pixel range of -1 to 1\n",
    "\n",
    "You need to do a bit of pre-processing; you know that the output of a `tanh` activated generator will contain pixel values in a range from -1 to 1, and so, we need to rescale our training images to a range of -1 to 1. (Right now, they are in a range from 0-1.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Complete the scale function\n",
    "def scale(x, feature_range=(-1, 1)):\n",
    "    ''' Scale takes in an image x and returns that image, scaled\n",
    "       with a feature_range of pixel values from -1 to 1. \n",
    "       This function assumes that the input x is already scaled from 0-1.'''\n",
    "\n",
    "    max_scale = max(feature_range)\n",
    "    min_scale = min(feature_range)\n",
    "    \n",
    "    # Scale to feature range\n",
    "    x = x * (max_scale-min_scale) + min_scale\n",
    "    \n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.9373)\n",
      "Max:  tensor(0.7882)\n"
     ]
    }
   ],
   "source": [
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "# check scaled range\n",
    "# should be close to -1 to 1\n",
    "img = images[0]\n",
    "scaled_img = scale(img)\n",
    "\n",
    "print('Min: ', scaled_img.min())\n",
    "print('Max: ', scaled_img.max())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "# Define the Model\n",
    "\n",
    "A GAN is comprised of two adversarial networks, a discriminator and a generator.\n",
    "\n",
    "## Discriminator\n",
    "\n",
    "Your first task will be to define the discriminator. This is a convolutional classifier like you've built before, only without any maxpooling layers. To deal with this complex data, it's suggested you use a deep network with **normalization**. You are also allowed to create any helper functions that may be useful.\n",
    "\n",
    "#### Exercise: Complete the Discriminator class\n",
    "* The inputs to the discriminator are 32x32x3 tensor images\n",
    "* The output should be a single value that will indicate whether a given image is real or fake\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 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": 11,
   "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",
    "\n",
    "        # define all convolutional layers\n",
    "        self.conv1 = conv(3, conv_dim, 4, batch_norm=False)\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",
    "        # define last classification layer\n",
    "        self.conv5 = conv(conv_dim*8, 1, 4, stride=1, batch_norm=False)\n",
    "        \n",
    "        # define activation function\n",
    "        self.leakyrelu = nn.LeakyReLU()\n",
    "        \n",
    "        # define dropout layer\n",
    "        self.dropout = nn.Dropout(0.3)\n",
    "        \n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: Discriminator logits; the output of the neural network\n",
    "        \"\"\"\n",
    "        ## define feedforward behavior\n",
    "        # all convolutional layers\n",
    "        x = self.leakyrelu(self.conv1(x))\n",
    "        x = self.leakyrelu(self.conv2(x))\n",
    "        x = self.leakyrelu(self.conv3(x))\n",
    "        x = self.dropout(x)\n",
    "        x = self.leakyrelu(self.conv4(x))\n",
    "        x = self.dropout(x)\n",
    "        \n",
    "        # last classification layer\n",
    "        out = self.conv5(x).view(-1, 1)\n",
    "        \n",
    "        return out\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_discriminator(Discriminator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generator\n",
    "\n",
    "The generator should upsample an input and generate a *new* image of the same size as our training data `32x32x3`. This should be mostly transpose convolutional layers with normalization applied to the outputs.\n",
    "\n",
    "#### Exercise: Complete the Generator class\n",
    "* The inputs to the generator are vectors of some length `z_size`\n",
    "* The output should be a image of shape `32x32x3`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "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": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Generator(nn.Module):\n",
    "    \n",
    "    def __init__(self, z_size, conv_dim):\n",
    "        \"\"\"\n",
    "        Initialize the Generator Module\n",
    "        :param z_size: The length of the input latent vector, z\n",
    "        :param conv_dim: The depth of the inputs to the *last* transpose convolutional layer\n",
    "        \"\"\"\n",
    "        super(Generator, self).__init__()\n",
    "        \n",
    "        self.conv_dim = conv_dim\n",
    "\n",
    "        ## Define input fully connected layer\n",
    "        self.fcin = nn.Linear(z_size, 4*4*conv_dim*4)\n",
    "        \n",
    "        ## Define transpose convolutional layers\n",
    "        self.t_conv1 = deconv(conv_dim*4, conv_dim*2, 4, batch_norm=True)\n",
    "        self.t_conv2 = deconv(conv_dim*2, conv_dim, 4, batch_norm=True)\n",
    "        self.t_conv3 = deconv(conv_dim, 3, 4, batch_norm=False)\n",
    "        \n",
    "        ## Define activation function\n",
    "        self.relu = nn.ReLU()\n",
    "        self.tanh = nn.Tanh()\n",
    "        \n",
    "        \n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: A 32x32x3 Tensor image as output\n",
    "        \"\"\"\n",
    "        ## input fully connected layer\n",
    "        x = self.fcin(x)\n",
    "        \n",
    "        ## reshape the tensor for transpose concolutional layer\n",
    "        x = x.view(-1, self.conv_dim*4, 4, 4)\n",
    "        \n",
    "        ## convolutional layers\n",
    "        x = self.relu(self.t_conv1(x))\n",
    "        x = self.relu(self.t_conv2(x))\n",
    "        \n",
    "        ## last convolutional layer with tanh act function\n",
    "        out = self.tanh(self.t_conv3(x))\n",
    "        \n",
    "        return out\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_generator(Generator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Initialize the weights of your networks\n",
    "\n",
    "To help your models converge, you should initialize the weights of the convolutional and linear layers in your model. From reading the [original DCGAN paper](https://arxiv.org/pdf/1511.06434.pdf), they say:\n",
    "> All weights were initialized from a zero-centered Normal distribution with standard deviation 0.02.\n",
    "\n",
    "So, your next task will be to define a weight initialization function that does just this!\n",
    "\n",
    "You can refer back to the lesson on weight initialization or even consult existing model code, such as that from [the `networks.py` file in CycleGAN Github repository](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix/blob/master/models/networks.py) to help you complete this function.\n",
    "\n",
    "#### Exercise: Complete the weight initialization function\n",
    "\n",
    "* This should initialize only **convolutional** and **linear** layers\n",
    "* Initialize the weights to a normal distribution, centered around 0, with a standard deviation of 0.02.\n",
    "* The bias terms, if they exist, may be left alone or set to 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.nn import init\n",
    "\n",
    "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 = m.__class__.__name__\n",
    "    \n",
    "    if hasattr(m, 'weight') and (classname.find('Conv') != -1 or classname.find('Linear') != -1):\n",
    "        ## Initialize conv and linear layer\n",
    "        init.normal_(m.weight.data, 0.0, 0.02)\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": 15,
   "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": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Discriminator(\n",
      "  (conv1): Sequential(\n",
      "    (0): Conv2d(3, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (conv2): Sequential(\n",
      "    (0): Conv2d(32, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv3): Sequential(\n",
      "    (0): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv4): Sequential(\n",
      "    (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv5): Sequential(\n",
      "    (0): Conv2d(256, 1, kernel_size=(4, 4), stride=(1, 1), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (leakyrelu): LeakyReLU(negative_slope=0.01)\n",
      "  (dropout): Dropout(p=0.3)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fcin): Linear(in_features=256, out_features=4096, bias=True)\n",
      "  (t_conv1): Sequential(\n",
      "    (0): ConvTranspose2d(256, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv2): Sequential(\n",
      "    (0): ConvTranspose2d(128, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv3): Sequential(\n",
      "    (0): ConvTranspose2d(64, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (relu): ReLU()\n",
      "  (tanh): Tanh()\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 32\n",
    "g_conv_dim = 64\n",
    "z_size = 256\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": 17,
   "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": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "def real_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being real.\n",
    "       param, D_out: discriminator logits\n",
    "       return: real loss'''\n",
    "    loss = torch.mean(D_out-1)**2\n",
    "    return loss\n",
    "\n",
    "def fake_loss(D_out):\n",
    "    '''Calculates how close discriminator outputs are to being fake.\n",
    "       param, D_out: discriminator logits\n",
    "       return: fake loss'''\n",
    "    loss = torch.mean(D_out**2)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizers\n",
    "\n",
    "#### Exercise: Define optimizers for your Discriminator (D) and Generator (G)\n",
    "\n",
    "Define optimizers for your models with appropriate hyperparameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "lr = 0.0002\n",
    "beta1 = 0.3\n",
    "beta2 = 0.999\n",
    "\n",
    "# Create optimizers for the discriminator D and generator G\n",
    "d_optimizer = optim.Adam(D.parameters(), lr, [beta1, beta2])\n",
    "g_optimizer = optim.Adam(G.parameters(), lr, [beta1, beta2])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---\n",
    "## Training\n",
    "\n",
    "Training will involve alternating between training the discriminator and the generator. You'll use your functions `real_loss` and `fake_loss` to help you calculate the discriminator losses.\n",
    "\n",
    "* You should train the discriminator by alternating on real and fake images\n",
    "* Then the generator, which tries to trick the discriminator and should have an opposing loss function\n",
    "\n",
    "\n",
    "#### Saving Samples\n",
    "\n",
    "You've been given some code to print out some loss statistics and save some generated \"fake\" samples."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exercise: Complete the training function\n",
    "\n",
    "Keep in mind that, if you've moved your models to GPU, you'll also have to move any model inputs to GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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",
    "            # 1. Train the discriminator on real and fake images\n",
    "            # ==================================================\n",
    "            \n",
    "            d_optimizer.zero_grad()\n",
    "            \n",
    "            ## Compute real loss on discriminator\n",
    "            \n",
    "            # move real images to GPU if available\n",
    "            if train_on_gpu:\n",
    "                real_images = real_images.cuda()\n",
    "            \n",
    "            # pass real images to descriminator\n",
    "            D_real = D(real_images)\n",
    "            \n",
    "            # compute real loss of descriminator\n",
    "            d_real_loss = real_loss(D_real)\n",
    "            \n",
    "            \n",
    "            ## Compute fake loss on discriminator\n",
    "            \n",
    "            # Compute latent vector z\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "                \n",
    "            # generate fake images with generator\n",
    "            fake_images = G(z)\n",
    "            \n",
    "            # pass fake images to descriminator\n",
    "            D_fake = D(fake_images)\n",
    "            \n",
    "            # compute fake loss of descrimintor\n",
    "            d_fake_loss = fake_loss(D_fake)\n",
    "            \n",
    "            \n",
    "            ## Compute total descriminator loss\n",
    "            d_loss = d_real_loss + d_fake_loss\n",
    "            \n",
    "            ## Perform backprop to the descriminator\n",
    "            d_loss.backward()\n",
    "            d_optimizer.step()\n",
    "\n",
    "            # ===============================================\n",
    "            # 2. Train the generator with an adversarial loss\n",
    "            # ===============================================\n",
    "            \n",
    "            g_optimizer.zero_grad()\n",
    "            \n",
    "            # generate latent vector z\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "                \n",
    "            # generate fake images from latent vector z\n",
    "            fake_images = G(z)\n",
    "            \n",
    "            # compute generator loss\n",
    "            D_fake = D(fake_images)\n",
    "            g_loss = real_loss(D_fake)\n",
    "            \n",
    "            # perform backprop on generator\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": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "import signal\n",
    "\n",
    "from contextlib import contextmanager\n",
    "\n",
    "import requests\n",
    "\n",
    "\n",
    "DELAY = INTERVAL = 4 * 60  # interval time in seconds\n",
    "MIN_DELAY = MIN_INTERVAL = 2 * 60\n",
    "KEEPALIVE_URL = \"https://nebula.udacity.com/api/v1/remote/keep-alive\"\n",
    "TOKEN_URL = \"http://metadata.google.internal/computeMetadata/v1/instance/attributes/keep_alive_token\"\n",
    "TOKEN_HEADERS = {\"Metadata-Flavor\":\"Google\"}\n",
    "\n",
    "\n",
    "def _request_handler(headers):\n",
    "    def _handler(signum, frame):\n",
    "        requests.request(\"POST\", KEEPALIVE_URL, headers=headers)\n",
    "    return _handler\n",
    "\n",
    "\n",
    "@contextmanager\n",
    "def active_session(delay=DELAY, interval=INTERVAL):\n",
    "    \"\"\"\n",
    "    Example:\n",
    "\n",
    "    from workspace_utils import active session\n",
    "\n",
    "    with active_session():\n",
    "        # do long-running work here\n",
    "    \"\"\"\n",
    "    token = requests.request(\"GET\", TOKEN_URL, headers=TOKEN_HEADERS).text\n",
    "    headers = {'Authorization': \"STAR \" + token}\n",
    "    delay = max(delay, MIN_DELAY)\n",
    "    interval = max(interval, MIN_INTERVAL)\n",
    "    original_handler = signal.getsignal(signal.SIGALRM)\n",
    "    try:\n",
    "        signal.signal(signal.SIGALRM, _request_handler(headers))\n",
    "        signal.setitimer(signal.ITIMER_REAL, delay, interval)\n",
    "        yield\n",
    "    finally:\n",
    "        signal.signal(signal.SIGALRM, original_handler)\n",
    "        signal.setitimer(signal.ITIMER_REAL, 0)\n",
    "\n",
    "\n",
    "def keep_awake(iterable, delay=DELAY, interval=INTERVAL):\n",
    "    \"\"\"\n",
    "    Example:\n",
    "\n",
    "    from workspace_utils import keep_awake\n",
    "\n",
    "    for i in keep_awake(range(5)):\n",
    "        # do iteration with lots of work here\n",
    "    \"\"\"\n",
    "    with active_session(delay, interval): yield from iterable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   20] | d_loss: 1.4303 | g_loss: 0.5118\n",
      "Epoch [    1/   20] | d_loss: 0.2810 | g_loss: 0.6996\n",
      "Epoch [    1/   20] | d_loss: 0.1093 | g_loss: 1.1517\n",
      "Epoch [    1/   20] | d_loss: 0.0542 | g_loss: 1.1368\n",
      "Epoch [    1/   20] | d_loss: 0.1100 | g_loss: 0.8566\n",
      "Epoch [    1/   20] | d_loss: 0.1286 | g_loss: 0.9310\n",
      "Epoch [    1/   20] | d_loss: 0.0383 | g_loss: 0.9415\n",
      "Epoch [    1/   20] | d_loss: 0.2627 | g_loss: 0.3938\n",
      "Epoch [    1/   20] | d_loss: 0.3162 | g_loss: 0.5311\n",
      "Epoch [    1/   20] | d_loss: 0.1257 | g_loss: 0.6899\n",
      "Epoch [    1/   20] | d_loss: 0.0508 | g_loss: 1.1854\n",
      "Epoch [    1/   20] | d_loss: 0.0373 | g_loss: 1.0965\n",
      "Epoch [    1/   20] | d_loss: 0.0962 | g_loss: 0.7352\n",
      "Epoch [    1/   20] | d_loss: 0.1484 | g_loss: 0.5252\n",
      "Epoch [    1/   20] | d_loss: 0.0807 | g_loss: 0.8797\n",
      "Epoch [    1/   20] | d_loss: 0.1822 | g_loss: 1.0472\n",
      "Epoch [    1/   20] | d_loss: 0.1090 | g_loss: 1.0264\n",
      "Epoch [    1/   20] | d_loss: 0.1143 | g_loss: 0.6624\n",
      "Epoch [    1/   20] | d_loss: 0.0995 | g_loss: 0.6434\n",
      "Epoch [    1/   20] | d_loss: 0.1582 | g_loss: 0.5709\n",
      "Epoch [    1/   20] | d_loss: 0.3542 | g_loss: 1.3626\n",
      "Epoch [    1/   20] | d_loss: 0.2566 | g_loss: 0.2688\n",
      "Epoch [    1/   20] | d_loss: 0.2547 | g_loss: 0.3523\n",
      "Epoch [    1/   20] | d_loss: 0.2102 | g_loss: 0.5710\n",
      "Epoch [    1/   20] | d_loss: 0.2103 | g_loss: 1.0097\n",
      "Epoch [    1/   20] | d_loss: 0.1685 | g_loss: 1.0314\n",
      "Epoch [    1/   20] | d_loss: 0.1886 | g_loss: 0.4620\n",
      "Epoch [    1/   20] | d_loss: 0.2415 | g_loss: 0.2494\n",
      "Epoch [    1/   20] | d_loss: 0.2150 | g_loss: 0.6302\n",
      "Epoch [    1/   20] | d_loss: 0.0567 | g_loss: 1.1554\n",
      "Epoch [    1/   20] | d_loss: 0.2576 | g_loss: 0.3900\n",
      "Epoch [    1/   20] | d_loss: 1.0758 | g_loss: 0.8775\n",
      "Epoch [    1/   20] | d_loss: 0.1296 | g_loss: 0.3070\n",
      "Epoch [    1/   20] | d_loss: 0.1991 | g_loss: 0.6780\n",
      "Epoch [    1/   20] | d_loss: 0.4002 | g_loss: 0.3446\n",
      "Epoch [    1/   20] | d_loss: 0.1353 | g_loss: 0.4219\n",
      "Epoch [    1/   20] | d_loss: 0.4507 | g_loss: 1.4340\n",
      "Epoch [    1/   20] | d_loss: 0.3955 | g_loss: 0.4255\n",
      "Epoch [    1/   20] | d_loss: 0.6559 | g_loss: 0.5108\n",
      "Epoch [    1/   20] | d_loss: 0.2823 | g_loss: 0.4742\n",
      "Epoch [    1/   20] | d_loss: 0.0724 | g_loss: 0.7248\n",
      "Epoch [    1/   20] | d_loss: 0.4010 | g_loss: 0.5157\n",
      "Epoch [    1/   20] | d_loss: 0.2808 | g_loss: 0.3029\n",
      "Epoch [    1/   20] | d_loss: 0.2409 | g_loss: 0.2673\n",
      "Epoch [    1/   20] | d_loss: 0.0653 | g_loss: 0.6856\n",
      "Epoch [    1/   20] | d_loss: 0.1895 | g_loss: 0.5919\n",
      "Epoch [    1/   20] | d_loss: 0.3922 | g_loss: 0.3870\n",
      "Epoch [    1/   20] | d_loss: 0.3647 | g_loss: 0.4773\n",
      "Epoch [    1/   20] | d_loss: 0.1970 | g_loss: 0.5501\n",
      "Epoch [    1/   20] | d_loss: 0.3190 | g_loss: 0.8911\n",
      "Epoch [    1/   20] | d_loss: 0.2488 | g_loss: 0.9870\n",
      "Epoch [    1/   20] | d_loss: 0.1705 | g_loss: 0.4856\n",
      "Epoch [    1/   20] | d_loss: 0.3993 | g_loss: 0.6334\n",
      "Epoch [    1/   20] | d_loss: 0.2727 | g_loss: 0.6621\n",
      "Epoch [    1/   20] | d_loss: 0.4546 | g_loss: 1.2963\n",
      "Epoch [    1/   20] | d_loss: 0.3843 | g_loss: 0.2693\n",
      "Epoch [    1/   20] | d_loss: 0.4913 | g_loss: 0.4466\n",
      "Epoch [    1/   20] | d_loss: 0.2648 | g_loss: 0.6624\n",
      "Epoch [    1/   20] | d_loss: 0.2597 | g_loss: 0.9295\n",
      "Epoch [    1/   20] | d_loss: 0.3751 | g_loss: 0.5266\n",
      "Epoch [    1/   20] | d_loss: 0.2005 | g_loss: 0.6241\n",
      "Epoch [    1/   20] | d_loss: 0.2876 | g_loss: 0.6437\n",
      "Epoch [    1/   20] | d_loss: 0.7633 | g_loss: 0.5923\n",
      "Epoch [    1/   20] | d_loss: 0.1811 | g_loss: 0.5455\n",
      "Epoch [    1/   20] | d_loss: 0.2656 | g_loss: 0.4764\n",
      "Epoch [    2/   20] | d_loss: 0.5248 | g_loss: 1.3922\n",
      "Epoch [    2/   20] | d_loss: 0.3432 | g_loss: 0.3245\n",
      "Epoch [    2/   20] | d_loss: 0.2984 | g_loss: 0.3735\n",
      "Epoch [    2/   20] | d_loss: 0.2070 | g_loss: 0.7123\n",
      "Epoch [    2/   20] | d_loss: 0.2810 | g_loss: 0.4155\n",
      "Epoch [    2/   20] | d_loss: 0.2970 | g_loss: 0.6078\n",
      "Epoch [    2/   20] | d_loss: 0.3142 | g_loss: 0.8216\n",
      "Epoch [    2/   20] | d_loss: 0.4749 | g_loss: 0.8251\n",
      "Epoch [    2/   20] | d_loss: 0.2569 | g_loss: 0.7321\n",
      "Epoch [    2/   20] | d_loss: 0.2105 | g_loss: 0.4902\n",
      "Epoch [    2/   20] | d_loss: 0.4245 | g_loss: 0.9470\n",
      "Epoch [    2/   20] | d_loss: 0.2387 | g_loss: 0.5140\n",
      "Epoch [    2/   20] | d_loss: 0.3058 | g_loss: 0.3026\n",
      "Epoch [    2/   20] | d_loss: 0.2699 | g_loss: 0.5884\n",
      "Epoch [    2/   20] | d_loss: 0.3382 | g_loss: 0.1900\n",
      "Epoch [    2/   20] | d_loss: 0.3974 | g_loss: 0.1642\n",
      "Epoch [    2/   20] | d_loss: 0.4457 | g_loss: 0.4556\n",
      "Epoch [    2/   20] | d_loss: 0.2091 | g_loss: 0.4107\n",
      "Epoch [    2/   20] | d_loss: 0.2577 | g_loss: 0.5823\n",
      "Epoch [    2/   20] | d_loss: 0.4625 | g_loss: 0.4344\n",
      "Epoch [    2/   20] | d_loss: 0.4147 | g_loss: 0.4085\n",
      "Epoch [    2/   20] | d_loss: 0.2317 | g_loss: 0.6261\n",
      "Epoch [    2/   20] | d_loss: 0.2402 | g_loss: 0.4723\n",
      "Epoch [    2/   20] | d_loss: 0.2126 | g_loss: 0.6345\n",
      "Epoch [    2/   20] | d_loss: 0.2307 | g_loss: 0.6884\n",
      "Epoch [    2/   20] | d_loss: 0.2831 | g_loss: 0.3776\n",
      "Epoch [    2/   20] | d_loss: 0.1077 | g_loss: 0.6944\n",
      "Epoch [    2/   20] | d_loss: 0.3105 | g_loss: 0.5278\n",
      "Epoch [    2/   20] | d_loss: 0.5118 | g_loss: 0.8970\n",
      "Epoch [    2/   20] | d_loss: 0.1435 | g_loss: 0.5446\n",
      "Epoch [    2/   20] | d_loss: 0.3702 | g_loss: 0.2462\n",
      "Epoch [    2/   20] | d_loss: 0.2309 | g_loss: 0.4247\n",
      "Epoch [    2/   20] | d_loss: 0.2059 | g_loss: 0.4123\n",
      "Epoch [    2/   20] | d_loss: 0.8164 | g_loss: 0.6037\n",
      "Epoch [    2/   20] | d_loss: 0.1926 | g_loss: 0.7637\n",
      "Epoch [    2/   20] | d_loss: 0.5913 | g_loss: 1.3121\n",
      "Epoch [    2/   20] | d_loss: 0.1997 | g_loss: 0.5613\n",
      "Epoch [    2/   20] | d_loss: 0.2217 | g_loss: 0.3558\n",
      "Epoch [    2/   20] | d_loss: 0.2873 | g_loss: 1.0640\n",
      "Epoch [    2/   20] | d_loss: 0.4121 | g_loss: 0.3422\n",
      "Epoch [    2/   20] | d_loss: 0.5363 | g_loss: 0.4174\n",
      "Epoch [    2/   20] | d_loss: 0.1054 | g_loss: 0.5379\n",
      "Epoch [    2/   20] | d_loss: 0.2743 | g_loss: 0.7749\n",
      "Epoch [    2/   20] | d_loss: 0.1773 | g_loss: 0.8359\n",
      "Epoch [    2/   20] | d_loss: 0.2123 | g_loss: 0.8792\n",
      "Epoch [    2/   20] | d_loss: 0.3221 | g_loss: 0.4720\n",
      "Epoch [    2/   20] | d_loss: 0.1823 | g_loss: 0.7135\n",
      "Epoch [    2/   20] | d_loss: 0.2436 | g_loss: 0.7235\n",
      "Epoch [    2/   20] | d_loss: 0.2751 | g_loss: 0.6358\n",
      "Epoch [    2/   20] | d_loss: 0.3766 | g_loss: 0.3267\n",
      "Epoch [    2/   20] | d_loss: 0.1308 | g_loss: 0.1215\n",
      "Epoch [    2/   20] | d_loss: 0.4049 | g_loss: 0.3257\n",
      "Epoch [    2/   20] | d_loss: 0.5378 | g_loss: 0.4099\n",
      "Epoch [    2/   20] | d_loss: 0.2001 | g_loss: 0.4606\n",
      "Epoch [    2/   20] | d_loss: 0.1462 | g_loss: 0.4479\n",
      "Epoch [    2/   20] | d_loss: 0.1956 | g_loss: 0.4525\n",
      "Epoch [    2/   20] | d_loss: 0.2758 | g_loss: 0.5466\n",
      "Epoch [    2/   20] | d_loss: 0.1584 | g_loss: 0.6379\n",
      "Epoch [    2/   20] | d_loss: 0.1768 | g_loss: 1.0927\n",
      "Epoch [    2/   20] | d_loss: 0.3963 | g_loss: 0.4448\n",
      "Epoch [    2/   20] | d_loss: 0.2067 | g_loss: 0.7656\n",
      "Epoch [    2/   20] | d_loss: 0.4599 | g_loss: 0.5780\n",
      "Epoch [    2/   20] | d_loss: 0.2409 | g_loss: 0.3433\n",
      "Epoch [    2/   20] | d_loss: 0.0720 | g_loss: 0.5871\n",
      "Epoch [    2/   20] | d_loss: 0.4926 | g_loss: 0.4406\n",
      "Epoch [    3/   20] | d_loss: 0.2442 | g_loss: 0.7100\n",
      "Epoch [    3/   20] | d_loss: 0.2869 | g_loss: 1.0866\n",
      "Epoch [    3/   20] | d_loss: 0.1548 | g_loss: 0.5884\n",
      "Epoch [    3/   20] | d_loss: 0.1829 | g_loss: 0.1527\n",
      "Epoch [    3/   20] | d_loss: 0.2196 | g_loss: 0.6688\n",
      "Epoch [    3/   20] | d_loss: 0.3402 | g_loss: 0.4146\n",
      "Epoch [    3/   20] | d_loss: 0.3044 | g_loss: 0.3755\n",
      "Epoch [    3/   20] | d_loss: 0.2840 | g_loss: 0.7321\n",
      "Epoch [    3/   20] | d_loss: 0.1761 | g_loss: 0.2563\n",
      "Epoch [    3/   20] | d_loss: 0.1783 | g_loss: 0.9698\n",
      "Epoch [    3/   20] | d_loss: 0.2280 | g_loss: 0.4123\n",
      "Epoch [    3/   20] | d_loss: 0.3045 | g_loss: 0.7167\n",
      "Epoch [    3/   20] | d_loss: 0.4137 | g_loss: 0.4682\n",
      "Epoch [    3/   20] | d_loss: 0.1567 | g_loss: 0.6956\n",
      "Epoch [    3/   20] | d_loss: 0.2999 | g_loss: 0.6281\n",
      "Epoch [    3/   20] | d_loss: 0.2391 | g_loss: 0.4856\n",
      "Epoch [    3/   20] | d_loss: 0.2506 | g_loss: 0.3300\n",
      "Epoch [    3/   20] | d_loss: 0.2382 | g_loss: 0.3191\n",
      "Epoch [    3/   20] | d_loss: 0.5397 | g_loss: 0.3858\n",
      "Epoch [    3/   20] | d_loss: 0.3812 | g_loss: 1.3223\n",
      "Epoch [    3/   20] | d_loss: 0.1741 | g_loss: 0.7368\n",
      "Epoch [    3/   20] | d_loss: 0.3110 | g_loss: 0.4825\n",
      "Epoch [    3/   20] | d_loss: 0.2929 | g_loss: 0.9283\n",
      "Epoch [    3/   20] | d_loss: 0.5686 | g_loss: 0.2931\n",
      "Epoch [    3/   20] | d_loss: 0.2044 | g_loss: 0.5755\n",
      "Epoch [    3/   20] | d_loss: 0.1919 | g_loss: 0.4009\n",
      "Epoch [    3/   20] | d_loss: 0.1988 | g_loss: 0.4588\n",
      "Epoch [    3/   20] | d_loss: 0.2121 | g_loss: 0.1832\n",
      "Epoch [    3/   20] | d_loss: 0.2818 | g_loss: 0.8979\n",
      "Epoch [    3/   20] | d_loss: 0.3226 | g_loss: 0.8815\n",
      "Epoch [    3/   20] | d_loss: 0.1147 | g_loss: 0.5331\n",
      "Epoch [    3/   20] | d_loss: 0.4021 | g_loss: 0.4280\n",
      "Epoch [    3/   20] | d_loss: 0.1378 | g_loss: 0.4344\n",
      "Epoch [    3/   20] | d_loss: 0.3142 | g_loss: 0.4408\n",
      "Epoch [    3/   20] | d_loss: 0.1120 | g_loss: 0.7561\n",
      "Epoch [    3/   20] | d_loss: 0.1591 | g_loss: 0.4976\n",
      "Epoch [    3/   20] | d_loss: 0.2780 | g_loss: 0.6684\n",
      "Epoch [    3/   20] | d_loss: 0.2946 | g_loss: 1.0432\n",
      "Epoch [    3/   20] | d_loss: 0.2213 | g_loss: 0.4455\n",
      "Epoch [    3/   20] | d_loss: 0.4229 | g_loss: 0.6071\n",
      "Epoch [    3/   20] | d_loss: 0.3213 | g_loss: 0.6280\n",
      "Epoch [    3/   20] | d_loss: 0.1569 | g_loss: 0.8113\n",
      "Epoch [    3/   20] | d_loss: 0.3928 | g_loss: 0.3738\n",
      "Epoch [    3/   20] | d_loss: 0.3277 | g_loss: 0.9122\n",
      "Epoch [    3/   20] | d_loss: 0.1351 | g_loss: 0.7589\n",
      "Epoch [    3/   20] | d_loss: 0.1049 | g_loss: 0.5682\n",
      "Epoch [    3/   20] | d_loss: 0.3409 | g_loss: 0.1956\n",
      "Epoch [    3/   20] | d_loss: 0.3687 | g_loss: 0.4859\n",
      "Epoch [    3/   20] | d_loss: 0.2462 | g_loss: 0.7306\n",
      "Epoch [    3/   20] | d_loss: 0.1405 | g_loss: 0.7881\n",
      "Epoch [    3/   20] | d_loss: 0.4744 | g_loss: 0.3635\n",
      "Epoch [    3/   20] | d_loss: 0.2346 | g_loss: 0.3667\n",
      "Epoch [    3/   20] | d_loss: 0.2300 | g_loss: 0.6821\n",
      "Epoch [    3/   20] | d_loss: 0.4166 | g_loss: 0.4864\n",
      "Epoch [    3/   20] | d_loss: 0.3778 | g_loss: 0.6886\n",
      "Epoch [    3/   20] | d_loss: 0.2805 | g_loss: 0.2273\n",
      "Epoch [    3/   20] | d_loss: 0.3344 | g_loss: 0.6530\n",
      "Epoch [    3/   20] | d_loss: 0.3744 | g_loss: 0.5841\n",
      "Epoch [    3/   20] | d_loss: 0.5421 | g_loss: 0.8372\n",
      "Epoch [    3/   20] | d_loss: 0.3143 | g_loss: 0.2929\n",
      "Epoch [    3/   20] | d_loss: 0.1743 | g_loss: 0.4973\n",
      "Epoch [    3/   20] | d_loss: 0.2717 | g_loss: 0.8352\n",
      "Epoch [    3/   20] | d_loss: 1.0670 | g_loss: 0.6527\n",
      "Epoch [    3/   20] | d_loss: 0.0617 | g_loss: 1.1701\n",
      "Epoch [    3/   20] | d_loss: 0.3670 | g_loss: 0.3839\n",
      "Epoch [    4/   20] | d_loss: 0.4352 | g_loss: 0.7510\n",
      "Epoch [    4/   20] | d_loss: 0.2140 | g_loss: 0.8606\n",
      "Epoch [    4/   20] | d_loss: 0.3012 | g_loss: 0.3463\n",
      "Epoch [    4/   20] | d_loss: 0.3093 | g_loss: 0.2125\n",
      "Epoch [    4/   20] | d_loss: 0.2084 | g_loss: 0.6151\n",
      "Epoch [    4/   20] | d_loss: 0.4173 | g_loss: 0.2551\n",
      "Epoch [    4/   20] | d_loss: 0.1370 | g_loss: 1.1584\n",
      "Epoch [    4/   20] | d_loss: 0.2868 | g_loss: 0.9188\n",
      "Epoch [    4/   20] | d_loss: 0.3474 | g_loss: 0.3443\n",
      "Epoch [    4/   20] | d_loss: 0.1704 | g_loss: 0.7352\n",
      "Epoch [    4/   20] | d_loss: 0.4081 | g_loss: 0.6054\n",
      "Epoch [    4/   20] | d_loss: 0.4215 | g_loss: 0.6170\n",
      "Epoch [    4/   20] | d_loss: 0.1719 | g_loss: 0.5104\n",
      "Epoch [    4/   20] | d_loss: 0.2015 | g_loss: 0.2082\n",
      "Epoch [    4/   20] | d_loss: 0.1093 | g_loss: 0.6605\n",
      "Epoch [    4/   20] | d_loss: 0.3182 | g_loss: 0.5705\n",
      "Epoch [    4/   20] | d_loss: 0.2129 | g_loss: 0.8570\n",
      "Epoch [    4/   20] | d_loss: 0.3324 | g_loss: 0.8340\n",
      "Epoch [    4/   20] | d_loss: 0.3469 | g_loss: 0.6188\n",
      "Epoch [    4/   20] | d_loss: 0.2265 | g_loss: 0.2933\n",
      "Epoch [    4/   20] | d_loss: 0.2516 | g_loss: 0.3365\n",
      "Epoch [    4/   20] | d_loss: 0.4457 | g_loss: 1.0082\n",
      "Epoch [    4/   20] | d_loss: 0.4521 | g_loss: 0.5302\n",
      "Epoch [    4/   20] | d_loss: 0.2506 | g_loss: 0.5056\n",
      "Epoch [    4/   20] | d_loss: 0.3336 | g_loss: 0.5701\n",
      "Epoch [    4/   20] | d_loss: 0.1184 | g_loss: 0.4242\n",
      "Epoch [    4/   20] | d_loss: 0.3646 | g_loss: 0.9272\n",
      "Epoch [    4/   20] | d_loss: 0.1736 | g_loss: 0.8714\n",
      "Epoch [    4/   20] | d_loss: 0.2965 | g_loss: 0.6246\n",
      "Epoch [    4/   20] | d_loss: 0.1407 | g_loss: 0.2684\n",
      "Epoch [    4/   20] | d_loss: 0.3255 | g_loss: 0.8074\n",
      "Epoch [    4/   20] | d_loss: 0.1604 | g_loss: 0.3000\n",
      "Epoch [    4/   20] | d_loss: 0.3067 | g_loss: 0.5941\n",
      "Epoch [    4/   20] | d_loss: 0.4143 | g_loss: 0.5929\n",
      "Epoch [    4/   20] | d_loss: 0.1169 | g_loss: 0.6682\n",
      "Epoch [    4/   20] | d_loss: 0.1108 | g_loss: 0.4003\n",
      "Epoch [    4/   20] | d_loss: 0.2503 | g_loss: 0.4668\n",
      "Epoch [    4/   20] | d_loss: 0.2465 | g_loss: 0.2999\n",
      "Epoch [    4/   20] | d_loss: 0.4670 | g_loss: 0.7622\n",
      "Epoch [    4/   20] | d_loss: 0.1248 | g_loss: 0.5569\n",
      "Epoch [    4/   20] | d_loss: 0.1835 | g_loss: 0.4283\n",
      "Epoch [    4/   20] | d_loss: 0.3038 | g_loss: 0.6180\n",
      "Epoch [    4/   20] | d_loss: 0.2165 | g_loss: 0.4590\n",
      "Epoch [    4/   20] | d_loss: 0.2442 | g_loss: 0.4907\n",
      "Epoch [    4/   20] | d_loss: 0.5282 | g_loss: 0.4689\n",
      "Epoch [    4/   20] | d_loss: 0.4251 | g_loss: 0.4986\n",
      "Epoch [    4/   20] | d_loss: 0.2339 | g_loss: 0.2901\n",
      "Epoch [    4/   20] | d_loss: 0.2151 | g_loss: 0.3965\n",
      "Epoch [    4/   20] | d_loss: 0.2128 | g_loss: 0.3264\n",
      "Epoch [    4/   20] | d_loss: 0.3635 | g_loss: 0.5540\n",
      "Epoch [    4/   20] | d_loss: 0.1198 | g_loss: 0.4893\n",
      "Epoch [    4/   20] | d_loss: 0.1211 | g_loss: 0.7032\n",
      "Epoch [    4/   20] | d_loss: 0.4073 | g_loss: 0.8798\n",
      "Epoch [    4/   20] | d_loss: 0.2767 | g_loss: 0.1609\n",
      "Epoch [    4/   20] | d_loss: 0.2150 | g_loss: 0.6336\n",
      "Epoch [    4/   20] | d_loss: 0.2869 | g_loss: 0.6758\n",
      "Epoch [    4/   20] | d_loss: 0.2619 | g_loss: 0.8592\n",
      "Epoch [    4/   20] | d_loss: 0.2228 | g_loss: 0.5156\n",
      "Epoch [    4/   20] | d_loss: 0.3355 | g_loss: 0.2484\n",
      "Epoch [    4/   20] | d_loss: 0.3025 | g_loss: 0.5261\n",
      "Epoch [    4/   20] | d_loss: 0.2200 | g_loss: 0.9227\n",
      "Epoch [    4/   20] | d_loss: 0.1080 | g_loss: 0.8673\n",
      "Epoch [    4/   20] | d_loss: 0.3293 | g_loss: 0.3172\n",
      "Epoch [    4/   20] | d_loss: 0.1505 | g_loss: 1.0949\n",
      "Epoch [    4/   20] | d_loss: 0.2490 | g_loss: 1.1157\n",
      "Epoch [    5/   20] | d_loss: 0.2711 | g_loss: 0.3968\n",
      "Epoch [    5/   20] | d_loss: 0.1845 | g_loss: 0.5570\n",
      "Epoch [    5/   20] | d_loss: 0.4694 | g_loss: 0.1843\n",
      "Epoch [    5/   20] | d_loss: 0.7380 | g_loss: 0.4290\n",
      "Epoch [    5/   20] | d_loss: 0.1427 | g_loss: 0.2576\n",
      "Epoch [    5/   20] | d_loss: 0.3223 | g_loss: 0.4022\n",
      "Epoch [    5/   20] | d_loss: 0.1036 | g_loss: 0.3063\n",
      "Epoch [    5/   20] | d_loss: 0.1787 | g_loss: 0.7045\n",
      "Epoch [    5/   20] | d_loss: 0.2248 | g_loss: 0.9476\n",
      "Epoch [    5/   20] | d_loss: 0.1433 | g_loss: 0.4474\n",
      "Epoch [    5/   20] | d_loss: 0.1255 | g_loss: 0.8816\n",
      "Epoch [    5/   20] | d_loss: 0.1438 | g_loss: 1.0027\n",
      "Epoch [    5/   20] | d_loss: 0.2242 | g_loss: 0.8308\n",
      "Epoch [    5/   20] | d_loss: 0.3017 | g_loss: 0.7034\n",
      "Epoch [    5/   20] | d_loss: 0.6212 | g_loss: 0.8291\n",
      "Epoch [    5/   20] | d_loss: 0.1625 | g_loss: 0.6850\n",
      "Epoch [    5/   20] | d_loss: 0.3303 | g_loss: 0.6968\n",
      "Epoch [    5/   20] | d_loss: 0.2798 | g_loss: 0.3625\n",
      "Epoch [    5/   20] | d_loss: 0.3061 | g_loss: 0.8678\n",
      "Epoch [    5/   20] | d_loss: 0.2277 | g_loss: 0.7986\n",
      "Epoch [    5/   20] | d_loss: 0.4013 | g_loss: 0.3662\n",
      "Epoch [    5/   20] | d_loss: 0.2225 | g_loss: 0.5717\n",
      "Epoch [    5/   20] | d_loss: 0.0719 | g_loss: 0.8927\n",
      "Epoch [    5/   20] | d_loss: 0.3575 | g_loss: 1.2211\n",
      "Epoch [    5/   20] | d_loss: 0.4042 | g_loss: 0.5277\n",
      "Epoch [    5/   20] | d_loss: 0.1491 | g_loss: 1.2069\n",
      "Epoch [    5/   20] | d_loss: 0.3507 | g_loss: 0.9339\n",
      "Epoch [    5/   20] | d_loss: 0.2923 | g_loss: 0.3908\n",
      "Epoch [    5/   20] | d_loss: 0.2359 | g_loss: 0.5397\n",
      "Epoch [    5/   20] | d_loss: 0.2320 | g_loss: 0.5019\n",
      "Epoch [    5/   20] | d_loss: 0.1092 | g_loss: 0.2233\n",
      "Epoch [    5/   20] | d_loss: 0.6415 | g_loss: 0.4384\n",
      "Epoch [    5/   20] | d_loss: 0.2310 | g_loss: 0.1807\n",
      "Epoch [    5/   20] | d_loss: 0.1340 | g_loss: 1.1926\n",
      "Epoch [    5/   20] | d_loss: 0.1085 | g_loss: 1.2122\n",
      "Epoch [    5/   20] | d_loss: 0.1140 | g_loss: 0.3664\n",
      "Epoch [    5/   20] | d_loss: 0.2510 | g_loss: 0.9141\n",
      "Epoch [    5/   20] | d_loss: 0.2382 | g_loss: 0.3788\n",
      "Epoch [    5/   20] | d_loss: 0.2974 | g_loss: 0.4310\n",
      "Epoch [    5/   20] | d_loss: 0.5813 | g_loss: 0.8317\n",
      "Epoch [    5/   20] | d_loss: 0.2483 | g_loss: 0.3609\n",
      "Epoch [    5/   20] | d_loss: 0.1439 | g_loss: 0.9814\n",
      "Epoch [    5/   20] | d_loss: 0.2610 | g_loss: 0.6818\n",
      "Epoch [    5/   20] | d_loss: 0.0892 | g_loss: 0.9579\n",
      "Epoch [    5/   20] | d_loss: 0.2924 | g_loss: 0.7340\n",
      "Epoch [    5/   20] | d_loss: 0.3153 | g_loss: 0.5705\n",
      "Epoch [    5/   20] | d_loss: 0.1117 | g_loss: 0.8416\n",
      "Epoch [    5/   20] | d_loss: 0.2742 | g_loss: 0.9437\n",
      "Epoch [    5/   20] | d_loss: 0.3363 | g_loss: 0.6882\n",
      "Epoch [    5/   20] | d_loss: 0.0713 | g_loss: 1.0719\n",
      "Epoch [    5/   20] | d_loss: 0.1772 | g_loss: 0.4346\n",
      "Epoch [    5/   20] | d_loss: 0.2463 | g_loss: 0.2048\n",
      "Epoch [    5/   20] | d_loss: 0.1107 | g_loss: 0.3950\n",
      "Epoch [    5/   20] | d_loss: 0.2692 | g_loss: 0.8274\n",
      "Epoch [    5/   20] | d_loss: 0.2757 | g_loss: 0.3206\n",
      "Epoch [    5/   20] | d_loss: 0.2982 | g_loss: 0.5249\n",
      "Epoch [    5/   20] | d_loss: 0.4258 | g_loss: 0.3377\n",
      "Epoch [    5/   20] | d_loss: 0.1841 | g_loss: 0.4953\n",
      "Epoch [    5/   20] | d_loss: 0.3006 | g_loss: 0.3217\n",
      "Epoch [    5/   20] | d_loss: 0.3529 | g_loss: 0.3836\n",
      "Epoch [    5/   20] | d_loss: 0.2445 | g_loss: 0.3687\n",
      "Epoch [    5/   20] | d_loss: 0.1729 | g_loss: 0.4267\n",
      "Epoch [    5/   20] | d_loss: 0.0543 | g_loss: 0.7868\n",
      "Epoch [    5/   20] | d_loss: 0.1945 | g_loss: 0.0162\n",
      "Epoch [    5/   20] | d_loss: 0.1148 | g_loss: 0.4567\n",
      "Epoch [    6/   20] | d_loss: 0.3913 | g_loss: 1.0254\n",
      "Epoch [    6/   20] | d_loss: 0.2174 | g_loss: 0.4835\n",
      "Epoch [    6/   20] | d_loss: 0.3043 | g_loss: 0.7038\n",
      "Epoch [    6/   20] | d_loss: 0.1010 | g_loss: 1.1385\n",
      "Epoch [    6/   20] | d_loss: 0.1523 | g_loss: 0.3718\n",
      "Epoch [    6/   20] | d_loss: 0.0855 | g_loss: 0.5153\n",
      "Epoch [    6/   20] | d_loss: 0.2572 | g_loss: 0.6496\n",
      "Epoch [    6/   20] | d_loss: 0.0816 | g_loss: 0.9178\n",
      "Epoch [    6/   20] | d_loss: 0.0908 | g_loss: 0.3863\n",
      "Epoch [    6/   20] | d_loss: 0.1963 | g_loss: 0.5500\n",
      "Epoch [    6/   20] | d_loss: 0.2028 | g_loss: 0.4776\n",
      "Epoch [    6/   20] | d_loss: 0.3765 | g_loss: 0.5226\n",
      "Epoch [    6/   20] | d_loss: 0.1669 | g_loss: 0.5704\n",
      "Epoch [    6/   20] | d_loss: 0.6390 | g_loss: 0.8397\n",
      "Epoch [    6/   20] | d_loss: 0.2550 | g_loss: 0.4626\n",
      "Epoch [    6/   20] | d_loss: 0.2643 | g_loss: 0.3808\n",
      "Epoch [    6/   20] | d_loss: 0.1341 | g_loss: 0.8392\n",
      "Epoch [    6/   20] | d_loss: 0.1849 | g_loss: 0.3436\n",
      "Epoch [    6/   20] | d_loss: 0.2385 | g_loss: 0.3045\n",
      "Epoch [    6/   20] | d_loss: 0.4284 | g_loss: 0.3885\n",
      "Epoch [    6/   20] | d_loss: 0.2564 | g_loss: 0.2381\n",
      "Epoch [    6/   20] | d_loss: 0.4330 | g_loss: 0.6399\n",
      "Epoch [    6/   20] | d_loss: 0.0733 | g_loss: 0.5994\n",
      "Epoch [    6/   20] | d_loss: 0.2933 | g_loss: 0.4602\n",
      "Epoch [    6/   20] | d_loss: 0.1345 | g_loss: 0.6676\n",
      "Epoch [    6/   20] | d_loss: 0.3106 | g_loss: 0.9058\n",
      "Epoch [    6/   20] | d_loss: 0.1437 | g_loss: 0.7422\n",
      "Epoch [    6/   20] | d_loss: 0.2783 | g_loss: 0.6791\n",
      "Epoch [    6/   20] | d_loss: 0.0887 | g_loss: 0.3075\n",
      "Epoch [    6/   20] | d_loss: 0.2422 | g_loss: 0.6865\n",
      "Epoch [    6/   20] | d_loss: 0.1311 | g_loss: 0.9265\n",
      "Epoch [    6/   20] | d_loss: 0.2196 | g_loss: 0.4950\n",
      "Epoch [    6/   20] | d_loss: 0.1290 | g_loss: 0.4205\n",
      "Epoch [    6/   20] | d_loss: 0.2322 | g_loss: 0.3439\n",
      "Epoch [    6/   20] | d_loss: 0.1535 | g_loss: 0.2116\n",
      "Epoch [    6/   20] | d_loss: 0.3485 | g_loss: 1.2038\n",
      "Epoch [    6/   20] | d_loss: 0.0941 | g_loss: 0.5390\n",
      "Epoch [    6/   20] | d_loss: 0.4623 | g_loss: 0.5252\n",
      "Epoch [    6/   20] | d_loss: 0.4070 | g_loss: 0.8025\n",
      "Epoch [    6/   20] | d_loss: 0.0732 | g_loss: 0.6679\n",
      "Epoch [    6/   20] | d_loss: 0.3764 | g_loss: 0.6001\n",
      "Epoch [    6/   20] | d_loss: 0.0915 | g_loss: 0.5974\n",
      "Epoch [    6/   20] | d_loss: 0.2939 | g_loss: 0.3851\n",
      "Epoch [    6/   20] | d_loss: 0.3845 | g_loss: 0.3607\n",
      "Epoch [    6/   20] | d_loss: 0.4332 | g_loss: 0.6602\n",
      "Epoch [    6/   20] | d_loss: 0.2250 | g_loss: 0.5800\n",
      "Epoch [    6/   20] | d_loss: 0.5303 | g_loss: 0.6003\n",
      "Epoch [    6/   20] | d_loss: 0.2233 | g_loss: 0.9781\n",
      "Epoch [    6/   20] | d_loss: 0.2964 | g_loss: 0.4293\n",
      "Epoch [    6/   20] | d_loss: 0.1582 | g_loss: 0.4692\n",
      "Epoch [    6/   20] | d_loss: 0.1287 | g_loss: 0.6999\n",
      "Epoch [    6/   20] | d_loss: 0.8122 | g_loss: 0.1758\n",
      "Epoch [    6/   20] | d_loss: 0.3393 | g_loss: 0.3035\n",
      "Epoch [    6/   20] | d_loss: 0.3181 | g_loss: 1.0612\n",
      "Epoch [    6/   20] | d_loss: 0.1749 | g_loss: 0.4612\n",
      "Epoch [    6/   20] | d_loss: 0.1158 | g_loss: 0.7670\n",
      "Epoch [    6/   20] | d_loss: 0.4658 | g_loss: 0.7455\n",
      "Epoch [    6/   20] | d_loss: 0.1457 | g_loss: 0.9387\n",
      "Epoch [    6/   20] | d_loss: 0.1653 | g_loss: 0.6832\n",
      "Epoch [    6/   20] | d_loss: 0.2019 | g_loss: 1.1859\n",
      "Epoch [    6/   20] | d_loss: 0.1569 | g_loss: 0.5403\n",
      "Epoch [    6/   20] | d_loss: 0.1958 | g_loss: 0.6563\n",
      "Epoch [    6/   20] | d_loss: 0.1124 | g_loss: 0.4704\n",
      "Epoch [    6/   20] | d_loss: 0.2750 | g_loss: 1.1212\n",
      "Epoch [    6/   20] | d_loss: 0.2943 | g_loss: 0.7107\n",
      "Epoch [    7/   20] | d_loss: 0.2750 | g_loss: 0.4543\n",
      "Epoch [    7/   20] | d_loss: 0.2316 | g_loss: 0.8978\n",
      "Epoch [    7/   20] | d_loss: 0.2688 | g_loss: 0.6992\n",
      "Epoch [    7/   20] | d_loss: 0.1369 | g_loss: 0.2500\n",
      "Epoch [    7/   20] | d_loss: 0.4336 | g_loss: 0.5545\n",
      "Epoch [    7/   20] | d_loss: 0.0817 | g_loss: 0.6236\n",
      "Epoch [    7/   20] | d_loss: 0.1327 | g_loss: 0.2108\n",
      "Epoch [    7/   20] | d_loss: 0.3270 | g_loss: 0.3257\n",
      "Epoch [    7/   20] | d_loss: 0.5066 | g_loss: 0.6026\n",
      "Epoch [    7/   20] | d_loss: 0.1880 | g_loss: 1.0036\n",
      "Epoch [    7/   20] | d_loss: 0.0952 | g_loss: 0.4873\n",
      "Epoch [    7/   20] | d_loss: 0.7584 | g_loss: 1.1044\n",
      "Epoch [    7/   20] | d_loss: 0.2546 | g_loss: 0.3937\n",
      "Epoch [    7/   20] | d_loss: 0.2006 | g_loss: 0.5855\n",
      "Epoch [    7/   20] | d_loss: 0.2160 | g_loss: 0.4693\n",
      "Epoch [    7/   20] | d_loss: 0.2964 | g_loss: 0.4337\n",
      "Epoch [    7/   20] | d_loss: 0.1217 | g_loss: 0.4729\n",
      "Epoch [    7/   20] | d_loss: 0.1013 | g_loss: 0.3992\n",
      "Epoch [    7/   20] | d_loss: 0.0621 | g_loss: 0.6677\n",
      "Epoch [    7/   20] | d_loss: 0.3252 | g_loss: 0.4162\n",
      "Epoch [    7/   20] | d_loss: 0.2261 | g_loss: 0.7270\n",
      "Epoch [    7/   20] | d_loss: 0.2070 | g_loss: 0.9501\n",
      "Epoch [    7/   20] | d_loss: 0.4285 | g_loss: 0.7528\n",
      "Epoch [    7/   20] | d_loss: 0.3149 | g_loss: 0.6490\n",
      "Epoch [    7/   20] | d_loss: 0.1277 | g_loss: 0.5156\n",
      "Epoch [    7/   20] | d_loss: 0.1921 | g_loss: 0.4774\n",
      "Epoch [    7/   20] | d_loss: 0.2352 | g_loss: 0.0435\n",
      "Epoch [    7/   20] | d_loss: 0.2538 | g_loss: 0.8976\n",
      "Epoch [    7/   20] | d_loss: 0.1741 | g_loss: 0.3984\n",
      "Epoch [    7/   20] | d_loss: 0.2649 | g_loss: 0.6223\n",
      "Epoch [    7/   20] | d_loss: 0.2290 | g_loss: 0.6050\n",
      "Epoch [    7/   20] | d_loss: 0.1676 | g_loss: 0.6976\n",
      "Epoch [    7/   20] | d_loss: 0.1993 | g_loss: 1.1399\n",
      "Epoch [    7/   20] | d_loss: 0.2114 | g_loss: 0.7536\n",
      "Epoch [    7/   20] | d_loss: 0.2371 | g_loss: 0.1286\n",
      "Epoch [    7/   20] | d_loss: 0.2814 | g_loss: 0.3174\n",
      "Epoch [    7/   20] | d_loss: 0.2416 | g_loss: 0.5407\n",
      "Epoch [    7/   20] | d_loss: 0.2678 | g_loss: 0.6921\n",
      "Epoch [    7/   20] | d_loss: 0.1073 | g_loss: 0.7153\n",
      "Epoch [    7/   20] | d_loss: 0.2169 | g_loss: 0.3387\n",
      "Epoch [    7/   20] | d_loss: 0.1922 | g_loss: 0.2941\n",
      "Epoch [    7/   20] | d_loss: 0.1586 | g_loss: 0.6147\n",
      "Epoch [    7/   20] | d_loss: 0.0892 | g_loss: 0.5725\n",
      "Epoch [    7/   20] | d_loss: 0.0498 | g_loss: 0.9938\n",
      "Epoch [    7/   20] | d_loss: 0.1694 | g_loss: 0.6073\n",
      "Epoch [    7/   20] | d_loss: 0.0554 | g_loss: 1.0394\n",
      "Epoch [    7/   20] | d_loss: 0.0819 | g_loss: 0.7968\n",
      "Epoch [    7/   20] | d_loss: 0.2854 | g_loss: 0.2295\n",
      "Epoch [    7/   20] | d_loss: 0.4055 | g_loss: 0.6208\n",
      "Epoch [    7/   20] | d_loss: 0.1841 | g_loss: 0.8409\n",
      "Epoch [    7/   20] | d_loss: 0.2789 | g_loss: 0.9543\n",
      "Epoch [    7/   20] | d_loss: 0.1371 | g_loss: 0.6920\n",
      "Epoch [    7/   20] | d_loss: 0.2749 | g_loss: 0.6376\n",
      "Epoch [    7/   20] | d_loss: 0.0807 | g_loss: 0.9135\n",
      "Epoch [    7/   20] | d_loss: 0.1712 | g_loss: 0.8299\n",
      "Epoch [    7/   20] | d_loss: 0.0759 | g_loss: 0.7126\n",
      "Epoch [    7/   20] | d_loss: 0.2661 | g_loss: 0.6425\n",
      "Epoch [    7/   20] | d_loss: 0.3565 | g_loss: 0.9000\n",
      "Epoch [    7/   20] | d_loss: 0.4212 | g_loss: 0.4828\n",
      "Epoch [    7/   20] | d_loss: 0.1363 | g_loss: 0.8171\n",
      "Epoch [    7/   20] | d_loss: 0.1552 | g_loss: 1.0565\n",
      "Epoch [    7/   20] | d_loss: 0.1025 | g_loss: 0.7165\n",
      "Epoch [    7/   20] | d_loss: 0.3271 | g_loss: 1.0209\n",
      "Epoch [    7/   20] | d_loss: 0.0723 | g_loss: 0.6571\n",
      "Epoch [    7/   20] | d_loss: 0.3099 | g_loss: 0.6137\n",
      "Epoch [    8/   20] | d_loss: 0.3169 | g_loss: 0.5728\n",
      "Epoch [    8/   20] | d_loss: 0.0593 | g_loss: 0.3630\n",
      "Epoch [    8/   20] | d_loss: 0.2199 | g_loss: 1.0665\n",
      "Epoch [    8/   20] | d_loss: 0.2668 | g_loss: 0.2394\n",
      "Epoch [    8/   20] | d_loss: 0.1448 | g_loss: 0.4497\n",
      "Epoch [    8/   20] | d_loss: 0.0844 | g_loss: 0.6153\n",
      "Epoch [    8/   20] | d_loss: 0.1786 | g_loss: 0.5982\n",
      "Epoch [    8/   20] | d_loss: 0.1231 | g_loss: 0.7005\n",
      "Epoch [    8/   20] | d_loss: 0.2886 | g_loss: 0.4840\n",
      "Epoch [    8/   20] | d_loss: 0.0539 | g_loss: 0.7130\n",
      "Epoch [    8/   20] | d_loss: 0.1688 | g_loss: 0.6586\n",
      "Epoch [    8/   20] | d_loss: 0.1966 | g_loss: 0.6710\n",
      "Epoch [    8/   20] | d_loss: 0.1115 | g_loss: 0.5323\n",
      "Epoch [    8/   20] | d_loss: 0.4087 | g_loss: 0.5412\n",
      "Epoch [    8/   20] | d_loss: 0.1565 | g_loss: 0.9257\n",
      "Epoch [    8/   20] | d_loss: 0.2077 | g_loss: 1.0019\n",
      "Epoch [    8/   20] | d_loss: 0.2842 | g_loss: 1.1357\n",
      "Epoch [    8/   20] | d_loss: 0.1319 | g_loss: 0.6569\n",
      "Epoch [    8/   20] | d_loss: 0.2927 | g_loss: 0.6375\n",
      "Epoch [    8/   20] | d_loss: 0.2323 | g_loss: 0.7548\n",
      "Epoch [    8/   20] | d_loss: 0.1189 | g_loss: 0.7247\n",
      "Epoch [    8/   20] | d_loss: 0.1150 | g_loss: 0.8492\n",
      "Epoch [    8/   20] | d_loss: 0.2199 | g_loss: 0.3876\n",
      "Epoch [    8/   20] | d_loss: 0.2632 | g_loss: 0.7253\n",
      "Epoch [    8/   20] | d_loss: 0.5152 | g_loss: 0.7103\n",
      "Epoch [    8/   20] | d_loss: 0.1710 | g_loss: 1.0870\n",
      "Epoch [    8/   20] | d_loss: 0.1305 | g_loss: 0.4307\n",
      "Epoch [    8/   20] | d_loss: 0.0912 | g_loss: 0.4576\n",
      "Epoch [    8/   20] | d_loss: 0.1624 | g_loss: 0.5658\n",
      "Epoch [    8/   20] | d_loss: 0.2826 | g_loss: 0.3920\n",
      "Epoch [    8/   20] | d_loss: 0.1711 | g_loss: 1.0431\n",
      "Epoch [    8/   20] | d_loss: 0.1787 | g_loss: 0.8397\n",
      "Epoch [    8/   20] | d_loss: 0.4148 | g_loss: 0.9849\n",
      "Epoch [    8/   20] | d_loss: 0.4378 | g_loss: 0.9815\n",
      "Epoch [    8/   20] | d_loss: 0.0691 | g_loss: 0.8324\n",
      "Epoch [    8/   20] | d_loss: 0.0883 | g_loss: 0.8376\n",
      "Epoch [    8/   20] | d_loss: 0.4442 | g_loss: 0.2118\n",
      "Epoch [    8/   20] | d_loss: 0.0737 | g_loss: 0.7411\n",
      "Epoch [    8/   20] | d_loss: 0.4882 | g_loss: 0.6061\n",
      "Epoch [    8/   20] | d_loss: 0.0963 | g_loss: 0.2708\n",
      "Epoch [    8/   20] | d_loss: 0.3262 | g_loss: 0.6234\n",
      "Epoch [    8/   20] | d_loss: 0.1951 | g_loss: 0.4161\n",
      "Epoch [    8/   20] | d_loss: 0.1476 | g_loss: 0.6220\n",
      "Epoch [    8/   20] | d_loss: 0.0983 | g_loss: 0.3976\n",
      "Epoch [    8/   20] | d_loss: 0.2025 | g_loss: 0.9421\n",
      "Epoch [    8/   20] | d_loss: 0.5307 | g_loss: 0.1182\n",
      "Epoch [    8/   20] | d_loss: 0.3757 | g_loss: 0.4103\n",
      "Epoch [    8/   20] | d_loss: 0.1669 | g_loss: 0.6127\n",
      "Epoch [    8/   20] | d_loss: 0.3343 | g_loss: 0.3390\n",
      "Epoch [    8/   20] | d_loss: 0.0930 | g_loss: 0.9028\n",
      "Epoch [    8/   20] | d_loss: 0.0390 | g_loss: 0.6946\n",
      "Epoch [    8/   20] | d_loss: 0.6430 | g_loss: 0.5910\n",
      "Epoch [    8/   20] | d_loss: 0.2643 | g_loss: 0.3261\n",
      "Epoch [    8/   20] | d_loss: 0.1208 | g_loss: 0.5880\n",
      "Epoch [    8/   20] | d_loss: 0.1627 | g_loss: 0.7979\n",
      "Epoch [    8/   20] | d_loss: 0.1640 | g_loss: 0.6045\n",
      "Epoch [    8/   20] | d_loss: 0.0973 | g_loss: 0.6133\n",
      "Epoch [    8/   20] | d_loss: 0.0573 | g_loss: 0.5007\n",
      "Epoch [    8/   20] | d_loss: 0.1916 | g_loss: 0.2286\n",
      "Epoch [    8/   20] | d_loss: 0.2070 | g_loss: 0.5470\n",
      "Epoch [    8/   20] | d_loss: 0.2322 | g_loss: 0.4936\n",
      "Epoch [    8/   20] | d_loss: 0.2291 | g_loss: 1.0315\n",
      "Epoch [    8/   20] | d_loss: 0.3692 | g_loss: 0.3557\n",
      "Epoch [    8/   20] | d_loss: 0.3082 | g_loss: 0.4302\n",
      "Epoch [    8/   20] | d_loss: 0.2337 | g_loss: 0.4481\n",
      "Epoch [    9/   20] | d_loss: 0.1857 | g_loss: 0.6872\n",
      "Epoch [    9/   20] | d_loss: 0.2270 | g_loss: 0.7560\n",
      "Epoch [    9/   20] | d_loss: 0.2178 | g_loss: 0.5808\n",
      "Epoch [    9/   20] | d_loss: 0.1282 | g_loss: 0.6165\n",
      "Epoch [    9/   20] | d_loss: 0.1296 | g_loss: 0.5340\n",
      "Epoch [    9/   20] | d_loss: 0.1851 | g_loss: 0.5221\n",
      "Epoch [    9/   20] | d_loss: 0.4225 | g_loss: 0.4681\n",
      "Epoch [    9/   20] | d_loss: 0.1207 | g_loss: 0.7847\n",
      "Epoch [    9/   20] | d_loss: 0.2497 | g_loss: 0.5466\n",
      "Epoch [    9/   20] | d_loss: 0.1679 | g_loss: 0.9059\n",
      "Epoch [    9/   20] | d_loss: 0.1362 | g_loss: 0.7532\n",
      "Epoch [    9/   20] | d_loss: 0.2609 | g_loss: 1.1600\n",
      "Epoch [    9/   20] | d_loss: 0.1675 | g_loss: 1.2682\n",
      "Epoch [    9/   20] | d_loss: 0.0820 | g_loss: 0.4469\n",
      "Epoch [    9/   20] | d_loss: 0.2544 | g_loss: 0.3613\n",
      "Epoch [    9/   20] | d_loss: 0.1929 | g_loss: 0.6501\n",
      "Epoch [    9/   20] | d_loss: 0.2178 | g_loss: 0.8456\n",
      "Epoch [    9/   20] | d_loss: 0.0536 | g_loss: 1.2328\n",
      "Epoch [    9/   20] | d_loss: 0.1166 | g_loss: 1.1094\n",
      "Epoch [    9/   20] | d_loss: 0.1432 | g_loss: 0.4225\n",
      "Epoch [    9/   20] | d_loss: 0.2942 | g_loss: 1.3287\n",
      "Epoch [    9/   20] | d_loss: 0.4999 | g_loss: 0.2661\n",
      "Epoch [    9/   20] | d_loss: 0.1277 | g_loss: 0.7831\n",
      "Epoch [    9/   20] | d_loss: 0.4574 | g_loss: 0.5679\n",
      "Epoch [    9/   20] | d_loss: 0.3098 | g_loss: 0.4661\n",
      "Epoch [    9/   20] | d_loss: 0.3340 | g_loss: 0.3908\n",
      "Epoch [    9/   20] | d_loss: 0.3616 | g_loss: 0.4567\n",
      "Epoch [    9/   20] | d_loss: 0.2717 | g_loss: 0.7199\n",
      "Epoch [    9/   20] | d_loss: 0.2143 | g_loss: 0.9724\n",
      "Epoch [    9/   20] | d_loss: 0.4568 | g_loss: 0.6595\n",
      "Epoch [    9/   20] | d_loss: 0.6079 | g_loss: 1.3115\n",
      "Epoch [    9/   20] | d_loss: 0.1508 | g_loss: 0.6195\n",
      "Epoch [    9/   20] | d_loss: 0.0883 | g_loss: 0.4320\n",
      "Epoch [    9/   20] | d_loss: 0.1415 | g_loss: 0.6618\n",
      "Epoch [    9/   20] | d_loss: 0.3278 | g_loss: 0.8060\n",
      "Epoch [    9/   20] | d_loss: 0.4213 | g_loss: 0.4434\n",
      "Epoch [    9/   20] | d_loss: 0.0794 | g_loss: 0.7180\n",
      "Epoch [    9/   20] | d_loss: 0.2532 | g_loss: 0.6708\n",
      "Epoch [    9/   20] | d_loss: 0.1370 | g_loss: 0.4311\n",
      "Epoch [    9/   20] | d_loss: 0.1021 | g_loss: 0.5170\n",
      "Epoch [    9/   20] | d_loss: 0.0737 | g_loss: 0.5680\n",
      "Epoch [    9/   20] | d_loss: 0.1043 | g_loss: 0.5635\n",
      "Epoch [    9/   20] | d_loss: 0.0704 | g_loss: 0.3378\n",
      "Epoch [    9/   20] | d_loss: 0.1689 | g_loss: 1.0264\n",
      "Epoch [    9/   20] | d_loss: 0.1797 | g_loss: 0.8020\n",
      "Epoch [    9/   20] | d_loss: 0.1803 | g_loss: 0.6056\n",
      "Epoch [    9/   20] | d_loss: 0.2896 | g_loss: 0.5168\n",
      "Epoch [    9/   20] | d_loss: 0.0469 | g_loss: 0.4375\n",
      "Epoch [    9/   20] | d_loss: 0.1253 | g_loss: 0.6493\n",
      "Epoch [    9/   20] | d_loss: 0.1006 | g_loss: 0.7010\n",
      "Epoch [    9/   20] | d_loss: 0.2272 | g_loss: 0.5493\n",
      "Epoch [    9/   20] | d_loss: 0.3796 | g_loss: 0.3985\n",
      "Epoch [    9/   20] | d_loss: 0.1408 | g_loss: 0.7002\n",
      "Epoch [    9/   20] | d_loss: 0.1578 | g_loss: 0.4955\n",
      "Epoch [    9/   20] | d_loss: 0.1989 | g_loss: 0.5698\n",
      "Epoch [    9/   20] | d_loss: 0.2068 | g_loss: 1.2035\n",
      "Epoch [    9/   20] | d_loss: 0.1174 | g_loss: 1.1542\n",
      "Epoch [    9/   20] | d_loss: 0.1107 | g_loss: 0.8035\n",
      "Epoch [    9/   20] | d_loss: 0.2957 | g_loss: 0.8965\n",
      "Epoch [    9/   20] | d_loss: 0.1183 | g_loss: 0.7632\n",
      "Epoch [    9/   20] | d_loss: 0.1061 | g_loss: 0.6348\n",
      "Epoch [    9/   20] | d_loss: 0.1685 | g_loss: 0.6916\n",
      "Epoch [    9/   20] | d_loss: 0.2575 | g_loss: 0.5559\n",
      "Epoch [    9/   20] | d_loss: 0.1101 | g_loss: 0.5258\n",
      "Epoch [    9/   20] | d_loss: 0.3215 | g_loss: 0.3547\n",
      "Epoch [   10/   20] | d_loss: 0.0687 | g_loss: 0.3609\n",
      "Epoch [   10/   20] | d_loss: 0.3846 | g_loss: 0.6912\n",
      "Epoch [   10/   20] | d_loss: 0.1796 | g_loss: 0.8349\n",
      "Epoch [   10/   20] | d_loss: 0.0431 | g_loss: 1.0919\n",
      "Epoch [   10/   20] | d_loss: 0.0679 | g_loss: 0.9303\n",
      "Epoch [   10/   20] | d_loss: 0.2099 | g_loss: 0.6676\n",
      "Epoch [   10/   20] | d_loss: 0.1113 | g_loss: 0.8467\n",
      "Epoch [   10/   20] | d_loss: 0.0700 | g_loss: 0.9860\n",
      "Epoch [   10/   20] | d_loss: 0.3572 | g_loss: 0.7514\n",
      "Epoch [   10/   20] | d_loss: 0.1734 | g_loss: 0.7752\n",
      "Epoch [   10/   20] | d_loss: 0.3156 | g_loss: 0.8985\n",
      "Epoch [   10/   20] | d_loss: 0.2267 | g_loss: 0.4837\n",
      "Epoch [   10/   20] | d_loss: 0.1250 | g_loss: 0.6589\n",
      "Epoch [   10/   20] | d_loss: 0.1505 | g_loss: 0.5900\n",
      "Epoch [   10/   20] | d_loss: 0.0799 | g_loss: 0.3165\n",
      "Epoch [   10/   20] | d_loss: 0.2007 | g_loss: 1.2146\n",
      "Epoch [   10/   20] | d_loss: 0.2119 | g_loss: 0.7990\n",
      "Epoch [   10/   20] | d_loss: 0.1435 | g_loss: 0.7629\n",
      "Epoch [   10/   20] | d_loss: 0.0636 | g_loss: 0.6340\n",
      "Epoch [   10/   20] | d_loss: 0.1747 | g_loss: 0.4845\n",
      "Epoch [   10/   20] | d_loss: 0.1064 | g_loss: 0.7170\n",
      "Epoch [   10/   20] | d_loss: 0.2131 | g_loss: 0.5402\n",
      "Epoch [   10/   20] | d_loss: 0.2434 | g_loss: 1.1586\n",
      "Epoch [   10/   20] | d_loss: 0.1051 | g_loss: 0.5023\n",
      "Epoch [   10/   20] | d_loss: 0.1175 | g_loss: 0.3117\n",
      "Epoch [   10/   20] | d_loss: 0.1426 | g_loss: 0.5181\n",
      "Epoch [   10/   20] | d_loss: 0.0446 | g_loss: 0.6223\n",
      "Epoch [   10/   20] | d_loss: 0.1010 | g_loss: 0.4064\n",
      "Epoch [   10/   20] | d_loss: 0.1750 | g_loss: 0.3050\n",
      "Epoch [   10/   20] | d_loss: 0.0712 | g_loss: 0.5281\n",
      "Epoch [   10/   20] | d_loss: 0.2541 | g_loss: 0.5824\n",
      "Epoch [   10/   20] | d_loss: 0.1702 | g_loss: 0.5536\n",
      "Epoch [   10/   20] | d_loss: 0.1382 | g_loss: 0.7080\n",
      "Epoch [   10/   20] | d_loss: 0.1385 | g_loss: 0.5917\n",
      "Epoch [   10/   20] | d_loss: 0.0806 | g_loss: 0.8606\n",
      "Epoch [   10/   20] | d_loss: 0.0459 | g_loss: 0.8115\n",
      "Epoch [   10/   20] | d_loss: 0.0302 | g_loss: 1.0354\n",
      "Epoch [   10/   20] | d_loss: 0.0591 | g_loss: 0.6755\n",
      "Epoch [   10/   20] | d_loss: 0.1549 | g_loss: 0.4713\n",
      "Epoch [   10/   20] | d_loss: 0.0915 | g_loss: 0.4790\n",
      "Epoch [   10/   20] | d_loss: 0.1055 | g_loss: 0.7552\n",
      "Epoch [   10/   20] | d_loss: 0.1529 | g_loss: 0.4262\n",
      "Epoch [   10/   20] | d_loss: 0.1871 | g_loss: 0.3927\n",
      "Epoch [   10/   20] | d_loss: 0.2026 | g_loss: 0.6799\n",
      "Epoch [   10/   20] | d_loss: 0.1876 | g_loss: 0.4341\n",
      "Epoch [   10/   20] | d_loss: 0.1770 | g_loss: 1.1812\n",
      "Epoch [   10/   20] | d_loss: 0.0935 | g_loss: 0.6017\n",
      "Epoch [   10/   20] | d_loss: 0.1318 | g_loss: 0.5908\n",
      "Epoch [   10/   20] | d_loss: 0.4247 | g_loss: 0.8405\n",
      "Epoch [   10/   20] | d_loss: 0.2448 | g_loss: 0.3731\n",
      "Epoch [   10/   20] | d_loss: 0.1308 | g_loss: 0.8499\n",
      "Epoch [   10/   20] | d_loss: 0.1162 | g_loss: 0.9050\n",
      "Epoch [   10/   20] | d_loss: 0.0419 | g_loss: 1.0500\n",
      "Epoch [   10/   20] | d_loss: 0.3415 | g_loss: 0.6610\n",
      "Epoch [   10/   20] | d_loss: 0.2239 | g_loss: 0.2681\n",
      "Epoch [   10/   20] | d_loss: 0.4121 | g_loss: 0.5588\n",
      "Epoch [   10/   20] | d_loss: 0.5013 | g_loss: 0.8772\n",
      "Epoch [   10/   20] | d_loss: 0.2477 | g_loss: 1.3014\n",
      "Epoch [   10/   20] | d_loss: 0.2946 | g_loss: 0.4151\n",
      "Epoch [   10/   20] | d_loss: 0.1191 | g_loss: 0.9998\n",
      "Epoch [   10/   20] | d_loss: 0.1895 | g_loss: 0.4348\n",
      "Epoch [   10/   20] | d_loss: 0.2086 | g_loss: 0.3411\n",
      "Epoch [   10/   20] | d_loss: 0.2139 | g_loss: 0.6426\n",
      "Epoch [   10/   20] | d_loss: 0.1422 | g_loss: 0.7362\n",
      "Epoch [   10/   20] | d_loss: 0.1223 | g_loss: 0.6225\n",
      "Epoch [   11/   20] | d_loss: 0.1070 | g_loss: 0.6857\n",
      "Epoch [   11/   20] | d_loss: 0.3574 | g_loss: 0.7638\n",
      "Epoch [   11/   20] | d_loss: 0.0480 | g_loss: 0.8496\n",
      "Epoch [   11/   20] | d_loss: 0.1049 | g_loss: 1.0427\n",
      "Epoch [   11/   20] | d_loss: 0.1006 | g_loss: 1.1883\n",
      "Epoch [   11/   20] | d_loss: 0.0692 | g_loss: 0.9293\n",
      "Epoch [   11/   20] | d_loss: 0.1890 | g_loss: 0.3487\n",
      "Epoch [   11/   20] | d_loss: 0.0950 | g_loss: 0.6862\n",
      "Epoch [   11/   20] | d_loss: 0.0906 | g_loss: 0.6574\n",
      "Epoch [   11/   20] | d_loss: 0.1261 | g_loss: 1.1196\n",
      "Epoch [   11/   20] | d_loss: 0.1462 | g_loss: 0.4979\n",
      "Epoch [   11/   20] | d_loss: 0.2373 | g_loss: 0.4368\n",
      "Epoch [   11/   20] | d_loss: 0.2341 | g_loss: 0.7194\n",
      "Epoch [   11/   20] | d_loss: 0.1959 | g_loss: 0.5847\n",
      "Epoch [   11/   20] | d_loss: 0.0377 | g_loss: 0.8325\n",
      "Epoch [   11/   20] | d_loss: 0.2180 | g_loss: 0.8037\n",
      "Epoch [   11/   20] | d_loss: 0.0812 | g_loss: 0.8814\n",
      "Epoch [   11/   20] | d_loss: 0.2222 | g_loss: 1.2281\n",
      "Epoch [   11/   20] | d_loss: 0.1183 | g_loss: 1.0200\n",
      "Epoch [   11/   20] | d_loss: 0.2869 | g_loss: 0.2648\n",
      "Epoch [   11/   20] | d_loss: 0.1239 | g_loss: 0.2633\n",
      "Epoch [   11/   20] | d_loss: 0.9230 | g_loss: 0.6163\n",
      "Epoch [   11/   20] | d_loss: 0.2584 | g_loss: 0.8750\n",
      "Epoch [   11/   20] | d_loss: 0.0933 | g_loss: 0.8991\n",
      "Epoch [   11/   20] | d_loss: 0.2617 | g_loss: 0.5428\n",
      "Epoch [   11/   20] | d_loss: 0.0939 | g_loss: 1.0611\n",
      "Epoch [   11/   20] | d_loss: 0.1613 | g_loss: 1.0022\n",
      "Epoch [   11/   20] | d_loss: 0.2303 | g_loss: 0.3102\n",
      "Epoch [   11/   20] | d_loss: 0.1157 | g_loss: 1.0064\n",
      "Epoch [   11/   20] | d_loss: 0.1167 | g_loss: 0.8914\n",
      "Epoch [   11/   20] | d_loss: 0.1470 | g_loss: 1.1067\n",
      "Epoch [   11/   20] | d_loss: 0.2077 | g_loss: 1.3483\n",
      "Epoch [   11/   20] | d_loss: 0.1563 | g_loss: 0.7470\n",
      "Epoch [   11/   20] | d_loss: 0.1570 | g_loss: 0.5115\n",
      "Epoch [   11/   20] | d_loss: 0.2954 | g_loss: 0.6220\n",
      "Epoch [   11/   20] | d_loss: 0.2957 | g_loss: 0.5655\n",
      "Epoch [   11/   20] | d_loss: 0.1152 | g_loss: 0.8733\n",
      "Epoch [   11/   20] | d_loss: 0.3546 | g_loss: 0.6500\n",
      "Epoch [   11/   20] | d_loss: 0.0592 | g_loss: 0.6482\n",
      "Epoch [   11/   20] | d_loss: 0.1949 | g_loss: 0.9009\n",
      "Epoch [   11/   20] | d_loss: 0.1948 | g_loss: 0.5876\n",
      "Epoch [   11/   20] | d_loss: 0.1388 | g_loss: 0.6750\n",
      "Epoch [   11/   20] | d_loss: 0.1767 | g_loss: 0.5526\n",
      "Epoch [   11/   20] | d_loss: 0.2672 | g_loss: 0.8134\n",
      "Epoch [   11/   20] | d_loss: 0.2094 | g_loss: 0.4686\n",
      "Epoch [   11/   20] | d_loss: 0.2035 | g_loss: 0.6790\n",
      "Epoch [   11/   20] | d_loss: 0.0418 | g_loss: 0.8091\n",
      "Epoch [   11/   20] | d_loss: 0.2926 | g_loss: 0.9299\n",
      "Epoch [   11/   20] | d_loss: 0.3133 | g_loss: 0.7772\n",
      "Epoch [   11/   20] | d_loss: 0.2817 | g_loss: 0.4992\n",
      "Epoch [   11/   20] | d_loss: 0.0550 | g_loss: 0.6271\n",
      "Epoch [   11/   20] | d_loss: 0.2440 | g_loss: 0.9816\n",
      "Epoch [   11/   20] | d_loss: 0.0914 | g_loss: 0.4921\n",
      "Epoch [   11/   20] | d_loss: 0.2012 | g_loss: 0.4446\n",
      "Epoch [   11/   20] | d_loss: 0.2156 | g_loss: 0.4459\n",
      "Epoch [   11/   20] | d_loss: 0.0987 | g_loss: 0.1813\n",
      "Epoch [   11/   20] | d_loss: 0.0432 | g_loss: 1.0306\n",
      "Epoch [   11/   20] | d_loss: 0.0634 | g_loss: 0.7320\n",
      "Epoch [   11/   20] | d_loss: 0.1760 | g_loss: 0.1844\n",
      "Epoch [   11/   20] | d_loss: 0.1222 | g_loss: 1.1830\n",
      "Epoch [   11/   20] | d_loss: 0.1897 | g_loss: 0.8966\n",
      "Epoch [   11/   20] | d_loss: 0.1141 | g_loss: 0.4776\n",
      "Epoch [   11/   20] | d_loss: 0.2920 | g_loss: 0.9651\n",
      "Epoch [   11/   20] | d_loss: 0.3750 | g_loss: 0.9580\n",
      "Epoch [   11/   20] | d_loss: 0.3721 | g_loss: 0.3639\n",
      "Epoch [   12/   20] | d_loss: 0.1410 | g_loss: 1.0211\n",
      "Epoch [   12/   20] | d_loss: 0.0516 | g_loss: 0.4467\n",
      "Epoch [   12/   20] | d_loss: 0.0641 | g_loss: 0.4297\n",
      "Epoch [   12/   20] | d_loss: 0.3074 | g_loss: 0.5476\n",
      "Epoch [   12/   20] | d_loss: 0.2247 | g_loss: 0.7624\n",
      "Epoch [   12/   20] | d_loss: 0.1030 | g_loss: 0.7711\n",
      "Epoch [   12/   20] | d_loss: 0.0619 | g_loss: 0.6777\n",
      "Epoch [   12/   20] | d_loss: 0.3607 | g_loss: 0.5602\n",
      "Epoch [   12/   20] | d_loss: 0.1518 | g_loss: 0.5067\n",
      "Epoch [   12/   20] | d_loss: 0.2361 | g_loss: 0.7889\n",
      "Epoch [   12/   20] | d_loss: 0.1423 | g_loss: 0.4617\n",
      "Epoch [   12/   20] | d_loss: 0.0791 | g_loss: 0.4727\n",
      "Epoch [   12/   20] | d_loss: 0.1286 | g_loss: 0.5459\n",
      "Epoch [   12/   20] | d_loss: 0.2416 | g_loss: 1.0701\n",
      "Epoch [   12/   20] | d_loss: 0.2676 | g_loss: 0.8073\n",
      "Epoch [   12/   20] | d_loss: 0.1987 | g_loss: 0.5640\n",
      "Epoch [   12/   20] | d_loss: 0.0366 | g_loss: 0.7471\n",
      "Epoch [   12/   20] | d_loss: 0.1575 | g_loss: 0.7957\n",
      "Epoch [   12/   20] | d_loss: 0.1382 | g_loss: 0.8034\n",
      "Epoch [   12/   20] | d_loss: 0.5114 | g_loss: 0.7724\n",
      "Epoch [   12/   20] | d_loss: 0.4221 | g_loss: 1.2116\n",
      "Epoch [   12/   20] | d_loss: 0.2704 | g_loss: 0.5767\n",
      "Epoch [   12/   20] | d_loss: 0.1308 | g_loss: 1.1901\n",
      "Epoch [   12/   20] | d_loss: 0.2359 | g_loss: 0.9848\n",
      "Epoch [   12/   20] | d_loss: 0.0850 | g_loss: 0.6453\n",
      "Epoch [   12/   20] | d_loss: 0.0316 | g_loss: 0.7842\n",
      "Epoch [   12/   20] | d_loss: 0.1017 | g_loss: 0.9783\n",
      "Epoch [   12/   20] | d_loss: 0.2467 | g_loss: 0.4434\n",
      "Epoch [   12/   20] | d_loss: 0.4959 | g_loss: 0.4749\n",
      "Epoch [   12/   20] | d_loss: 0.1039 | g_loss: 0.3640\n",
      "Epoch [   12/   20] | d_loss: 0.1665 | g_loss: 1.0627\n",
      "Epoch [   12/   20] | d_loss: 0.0952 | g_loss: 0.6914\n",
      "Epoch [   12/   20] | d_loss: 0.1146 | g_loss: 0.8935\n",
      "Epoch [   12/   20] | d_loss: 0.2923 | g_loss: 0.7609\n",
      "Epoch [   12/   20] | d_loss: 0.0634 | g_loss: 0.8487\n",
      "Epoch [   12/   20] | d_loss: 0.1631 | g_loss: 0.5436\n",
      "Epoch [   12/   20] | d_loss: 0.4517 | g_loss: 0.6917\n",
      "Epoch [   12/   20] | d_loss: 0.3192 | g_loss: 0.5081\n",
      "Epoch [   12/   20] | d_loss: 0.0907 | g_loss: 0.7494\n",
      "Epoch [   12/   20] | d_loss: 0.0796 | g_loss: 0.6832\n",
      "Epoch [   12/   20] | d_loss: 0.1711 | g_loss: 0.5376\n",
      "Epoch [   12/   20] | d_loss: 0.3264 | g_loss: 0.2901\n",
      "Epoch [   12/   20] | d_loss: 0.1873 | g_loss: 0.9661\n",
      "Epoch [   12/   20] | d_loss: 0.1396 | g_loss: 0.7256\n",
      "Epoch [   12/   20] | d_loss: 0.2017 | g_loss: 0.8349\n",
      "Epoch [   12/   20] | d_loss: 0.0875 | g_loss: 0.8566\n",
      "Epoch [   12/   20] | d_loss: 0.1763 | g_loss: 0.7574\n",
      "Epoch [   12/   20] | d_loss: 0.0699 | g_loss: 1.0734\n",
      "Epoch [   12/   20] | d_loss: 0.1836 | g_loss: 1.0962\n",
      "Epoch [   12/   20] | d_loss: 0.4803 | g_loss: 0.5698\n",
      "Epoch [   12/   20] | d_loss: 0.0796 | g_loss: 0.6550\n",
      "Epoch [   12/   20] | d_loss: 0.0730 | g_loss: 0.8227\n",
      "Epoch [   12/   20] | d_loss: 0.2249 | g_loss: 0.6820\n",
      "Epoch [   12/   20] | d_loss: 0.1591 | g_loss: 0.4762\n",
      "Epoch [   12/   20] | d_loss: 0.2928 | g_loss: 0.8045\n",
      "Epoch [   12/   20] | d_loss: 0.1934 | g_loss: 0.8540\n",
      "Epoch [   12/   20] | d_loss: 0.2385 | g_loss: 0.7962\n",
      "Epoch [   12/   20] | d_loss: 0.0853 | g_loss: 0.5406\n",
      "Epoch [   12/   20] | d_loss: 0.1351 | g_loss: 0.4790\n",
      "Epoch [   12/   20] | d_loss: 0.2817 | g_loss: 1.0139\n",
      "Epoch [   12/   20] | d_loss: 0.1580 | g_loss: 1.1240\n",
      "Epoch [   12/   20] | d_loss: 0.1867 | g_loss: 0.7436\n",
      "Epoch [   12/   20] | d_loss: 0.0975 | g_loss: 0.6272\n",
      "Epoch [   12/   20] | d_loss: 0.2282 | g_loss: 0.9026\n",
      "Epoch [   12/   20] | d_loss: 0.1611 | g_loss: 0.5818\n",
      "Epoch [   13/   20] | d_loss: 0.1488 | g_loss: 0.4914\n",
      "Epoch [   13/   20] | d_loss: 0.1114 | g_loss: 1.0213\n",
      "Epoch [   13/   20] | d_loss: 0.0547 | g_loss: 0.6044\n",
      "Epoch [   13/   20] | d_loss: 0.0684 | g_loss: 0.6738\n",
      "Epoch [   13/   20] | d_loss: 0.0928 | g_loss: 0.9593\n",
      "Epoch [   13/   20] | d_loss: 0.0715 | g_loss: 0.6192\n",
      "Epoch [   13/   20] | d_loss: 0.1544 | g_loss: 0.4400\n",
      "Epoch [   13/   20] | d_loss: 0.1610 | g_loss: 0.2567\n",
      "Epoch [   13/   20] | d_loss: 0.1758 | g_loss: 0.9287\n",
      "Epoch [   13/   20] | d_loss: 0.1714 | g_loss: 0.4665\n",
      "Epoch [   13/   20] | d_loss: 0.1094 | g_loss: 0.9983\n",
      "Epoch [   13/   20] | d_loss: 0.2651 | g_loss: 0.5437\n",
      "Epoch [   13/   20] | d_loss: 0.1191 | g_loss: 0.7260\n",
      "Epoch [   13/   20] | d_loss: 0.0784 | g_loss: 0.6252\n",
      "Epoch [   13/   20] | d_loss: 0.2389 | g_loss: 0.9742\n",
      "Epoch [   13/   20] | d_loss: 0.0979 | g_loss: 0.8851\n",
      "Epoch [   13/   20] | d_loss: 0.1055 | g_loss: 0.5037\n",
      "Epoch [   13/   20] | d_loss: 0.1161 | g_loss: 0.7495\n",
      "Epoch [   13/   20] | d_loss: 0.4066 | g_loss: 0.7613\n",
      "Epoch [   13/   20] | d_loss: 0.0601 | g_loss: 0.4609\n",
      "Epoch [   13/   20] | d_loss: 0.1566 | g_loss: 0.9667\n",
      "Epoch [   13/   20] | d_loss: 0.0871 | g_loss: 0.7737\n",
      "Epoch [   13/   20] | d_loss: 0.2713 | g_loss: 0.6838\n",
      "Epoch [   13/   20] | d_loss: 0.0449 | g_loss: 0.8727\n",
      "Epoch [   13/   20] | d_loss: 0.0350 | g_loss: 0.8765\n",
      "Epoch [   13/   20] | d_loss: 0.0562 | g_loss: 0.7237\n",
      "Epoch [   13/   20] | d_loss: 0.3056 | g_loss: 0.6686\n",
      "Epoch [   13/   20] | d_loss: 0.1481 | g_loss: 0.5605\n",
      "Epoch [   13/   20] | d_loss: 0.2398 | g_loss: 0.6230\n",
      "Epoch [   13/   20] | d_loss: 0.1759 | g_loss: 0.4354\n",
      "Epoch [   13/   20] | d_loss: 0.1055 | g_loss: 0.7258\n",
      "Epoch [   13/   20] | d_loss: 0.1348 | g_loss: 0.6432\n",
      "Epoch [   13/   20] | d_loss: 0.4440 | g_loss: 0.8351\n",
      "Epoch [   13/   20] | d_loss: 0.1712 | g_loss: 0.6553\n",
      "Epoch [   13/   20] | d_loss: 0.3778 | g_loss: 1.0841\n",
      "Epoch [   13/   20] | d_loss: 0.0565 | g_loss: 0.8168\n",
      "Epoch [   13/   20] | d_loss: 0.2653 | g_loss: 0.4832\n",
      "Epoch [   13/   20] | d_loss: 0.0910 | g_loss: 1.1403\n",
      "Epoch [   13/   20] | d_loss: 0.1640 | g_loss: 0.7468\n",
      "Epoch [   13/   20] | d_loss: 0.3056 | g_loss: 0.7381\n",
      "Epoch [   13/   20] | d_loss: 0.1859 | g_loss: 0.6571\n",
      "Epoch [   13/   20] | d_loss: 0.1560 | g_loss: 0.6825\n",
      "Epoch [   13/   20] | d_loss: 0.1818 | g_loss: 1.1569\n",
      "Epoch [   13/   20] | d_loss: 0.0257 | g_loss: 0.8301\n",
      "Epoch [   13/   20] | d_loss: 0.0692 | g_loss: 1.1906\n",
      "Epoch [   13/   20] | d_loss: 0.3709 | g_loss: 0.9566\n",
      "Epoch [   13/   20] | d_loss: 0.2811 | g_loss: 0.7177\n",
      "Epoch [   13/   20] | d_loss: 0.2908 | g_loss: 1.2143\n",
      "Epoch [   13/   20] | d_loss: 0.1067 | g_loss: 0.7036\n",
      "Epoch [   13/   20] | d_loss: 0.0510 | g_loss: 0.9173\n",
      "Epoch [   13/   20] | d_loss: 0.4452 | g_loss: 0.9637\n",
      "Epoch [   13/   20] | d_loss: 0.1337 | g_loss: 0.3673\n",
      "Epoch [   13/   20] | d_loss: 0.1655 | g_loss: 0.5556\n",
      "Epoch [   13/   20] | d_loss: 0.0828 | g_loss: 0.4053\n",
      "Epoch [   13/   20] | d_loss: 0.2216 | g_loss: 0.4567\n",
      "Epoch [   13/   20] | d_loss: 0.1431 | g_loss: 1.2516\n",
      "Epoch [   13/   20] | d_loss: 0.2477 | g_loss: 1.1289\n",
      "Epoch [   13/   20] | d_loss: 0.1579 | g_loss: 0.9532\n",
      "Epoch [   13/   20] | d_loss: 0.1638 | g_loss: 0.7727\n",
      "Epoch [   13/   20] | d_loss: 0.1858 | g_loss: 0.7868\n",
      "Epoch [   13/   20] | d_loss: 0.5599 | g_loss: 0.6947\n",
      "Epoch [   13/   20] | d_loss: 0.0440 | g_loss: 0.5155\n",
      "Epoch [   13/   20] | d_loss: 0.4761 | g_loss: 0.6154\n",
      "Epoch [   13/   20] | d_loss: 0.1722 | g_loss: 0.5623\n",
      "Epoch [   13/   20] | d_loss: 0.2385 | g_loss: 1.0149\n",
      "Epoch [   14/   20] | d_loss: 0.1828 | g_loss: 0.2393\n",
      "Epoch [   14/   20] | d_loss: 0.1190 | g_loss: 0.6442\n",
      "Epoch [   14/   20] | d_loss: 0.1296 | g_loss: 1.3072\n",
      "Epoch [   14/   20] | d_loss: 0.1706 | g_loss: 0.7154\n",
      "Epoch [   14/   20] | d_loss: 0.1484 | g_loss: 0.2813\n",
      "Epoch [   14/   20] | d_loss: 0.5455 | g_loss: 1.4058\n",
      "Epoch [   14/   20] | d_loss: 0.1236 | g_loss: 0.6100\n",
      "Epoch [   14/   20] | d_loss: 0.1678 | g_loss: 0.8762\n",
      "Epoch [   14/   20] | d_loss: 0.1707 | g_loss: 0.7207\n",
      "Epoch [   14/   20] | d_loss: 0.0706 | g_loss: 0.6323\n",
      "Epoch [   14/   20] | d_loss: 0.0531 | g_loss: 1.0402\n",
      "Epoch [   14/   20] | d_loss: 0.1422 | g_loss: 0.6616\n",
      "Epoch [   14/   20] | d_loss: 0.0810 | g_loss: 1.1764\n",
      "Epoch [   14/   20] | d_loss: 0.0578 | g_loss: 0.7218\n",
      "Epoch [   14/   20] | d_loss: 0.4318 | g_loss: 0.6791\n",
      "Epoch [   14/   20] | d_loss: 0.2729 | g_loss: 0.9794\n",
      "Epoch [   14/   20] | d_loss: 0.1489 | g_loss: 0.2415\n",
      "Epoch [   14/   20] | d_loss: 0.3761 | g_loss: 0.6759\n",
      "Epoch [   14/   20] | d_loss: 0.1807 | g_loss: 0.7441\n",
      "Epoch [   14/   20] | d_loss: 0.0415 | g_loss: 0.4133\n",
      "Epoch [   14/   20] | d_loss: 0.1940 | g_loss: 0.7471\n",
      "Epoch [   14/   20] | d_loss: 0.1549 | g_loss: 0.5300\n",
      "Epoch [   14/   20] | d_loss: 0.0985 | g_loss: 0.6029\n",
      "Epoch [   14/   20] | d_loss: 0.1782 | g_loss: 0.7514\n",
      "Epoch [   14/   20] | d_loss: 0.2559 | g_loss: 0.6286\n",
      "Epoch [   14/   20] | d_loss: 0.1149 | g_loss: 0.7716\n",
      "Epoch [   14/   20] | d_loss: 0.1867 | g_loss: 0.6151\n",
      "Epoch [   14/   20] | d_loss: 0.1116 | g_loss: 0.5404\n",
      "Epoch [   14/   20] | d_loss: 0.0579 | g_loss: 0.7232\n",
      "Epoch [   14/   20] | d_loss: 0.1514 | g_loss: 0.7087\n",
      "Epoch [   14/   20] | d_loss: 0.0967 | g_loss: 1.0180\n",
      "Epoch [   14/   20] | d_loss: 0.3191 | g_loss: 0.5227\n",
      "Epoch [   14/   20] | d_loss: 0.3298 | g_loss: 0.4045\n",
      "Epoch [   14/   20] | d_loss: 0.0554 | g_loss: 0.7371\n",
      "Epoch [   14/   20] | d_loss: 0.1291 | g_loss: 0.8731\n",
      "Epoch [   14/   20] | d_loss: 0.0482 | g_loss: 1.0085\n",
      "Epoch [   14/   20] | d_loss: 0.0779 | g_loss: 1.1535\n",
      "Epoch [   14/   20] | d_loss: 0.3009 | g_loss: 0.4636\n",
      "Epoch [   14/   20] | d_loss: 0.0900 | g_loss: 0.8347\n",
      "Epoch [   14/   20] | d_loss: 0.1401 | g_loss: 0.7563\n",
      "Epoch [   14/   20] | d_loss: 0.0559 | g_loss: 0.9821\n",
      "Epoch [   14/   20] | d_loss: 0.2220 | g_loss: 0.6007\n",
      "Epoch [   14/   20] | d_loss: 0.0683 | g_loss: 0.4367\n",
      "Epoch [   14/   20] | d_loss: 0.0584 | g_loss: 0.6878\n",
      "Epoch [   14/   20] | d_loss: 0.0968 | g_loss: 0.5650\n",
      "Epoch [   14/   20] | d_loss: 0.3315 | g_loss: 0.5263\n",
      "Epoch [   14/   20] | d_loss: 0.0716 | g_loss: 0.8502\n",
      "Epoch [   14/   20] | d_loss: 0.0839 | g_loss: 0.5262\n",
      "Epoch [   14/   20] | d_loss: 0.2850 | g_loss: 0.7324\n",
      "Epoch [   14/   20] | d_loss: 0.0331 | g_loss: 0.7885\n",
      "Epoch [   14/   20] | d_loss: 0.1908 | g_loss: 0.6797\n",
      "Epoch [   14/   20] | d_loss: 0.0922 | g_loss: 0.6115\n",
      "Epoch [   14/   20] | d_loss: 0.0989 | g_loss: 1.0802\n",
      "Epoch [   14/   20] | d_loss: 0.1546 | g_loss: 0.6995\n",
      "Epoch [   14/   20] | d_loss: 0.2310 | g_loss: 0.8114\n",
      "Epoch [   14/   20] | d_loss: 0.0413 | g_loss: 0.2556\n",
      "Epoch [   14/   20] | d_loss: 0.0559 | g_loss: 1.0184\n",
      "Epoch [   14/   20] | d_loss: 0.1952 | g_loss: 0.5798\n",
      "Epoch [   14/   20] | d_loss: 0.2386 | g_loss: 0.9101\n",
      "Epoch [   14/   20] | d_loss: 0.3211 | g_loss: 0.4866\n",
      "Epoch [   14/   20] | d_loss: 0.1193 | g_loss: 0.4327\n",
      "Epoch [   14/   20] | d_loss: 0.0859 | g_loss: 0.7410\n",
      "Epoch [   14/   20] | d_loss: 0.0672 | g_loss: 0.9653\n",
      "Epoch [   14/   20] | d_loss: 0.3146 | g_loss: 0.6749\n",
      "Epoch [   14/   20] | d_loss: 0.1102 | g_loss: 0.5890\n",
      "Epoch [   15/   20] | d_loss: 0.0934 | g_loss: 0.6842\n",
      "Epoch [   15/   20] | d_loss: 0.2921 | g_loss: 0.5041\n",
      "Epoch [   15/   20] | d_loss: 0.1167 | g_loss: 0.5570\n",
      "Epoch [   15/   20] | d_loss: 0.3838 | g_loss: 0.6274\n",
      "Epoch [   15/   20] | d_loss: 0.1384 | g_loss: 0.8803\n",
      "Epoch [   15/   20] | d_loss: 0.1433 | g_loss: 0.8101\n",
      "Epoch [   15/   20] | d_loss: 0.0202 | g_loss: 0.7634\n",
      "Epoch [   15/   20] | d_loss: 0.1335 | g_loss: 1.3521\n",
      "Epoch [   15/   20] | d_loss: 0.0774 | g_loss: 0.8829\n",
      "Epoch [   15/   20] | d_loss: 0.3829 | g_loss: 0.5739\n",
      "Epoch [   15/   20] | d_loss: 0.1198 | g_loss: 1.1993\n",
      "Epoch [   15/   20] | d_loss: 0.1896 | g_loss: 0.9711\n",
      "Epoch [   15/   20] | d_loss: 0.0750 | g_loss: 0.8893\n",
      "Epoch [   15/   20] | d_loss: 0.1193 | g_loss: 1.0952\n",
      "Epoch [   15/   20] | d_loss: 0.0745 | g_loss: 0.8723\n",
      "Epoch [   15/   20] | d_loss: 0.1075 | g_loss: 0.5378\n",
      "Epoch [   15/   20] | d_loss: 0.0733 | g_loss: 0.6854\n",
      "Epoch [   15/   20] | d_loss: 0.1124 | g_loss: 0.5445\n",
      "Epoch [   15/   20] | d_loss: 0.0838 | g_loss: 0.7558\n",
      "Epoch [   15/   20] | d_loss: 0.0249 | g_loss: 0.6820\n",
      "Epoch [   15/   20] | d_loss: 0.3600 | g_loss: 0.9241\n",
      "Epoch [   15/   20] | d_loss: 0.0447 | g_loss: 0.8855\n",
      "Epoch [   15/   20] | d_loss: 0.1888 | g_loss: 0.7599\n",
      "Epoch [   15/   20] | d_loss: 0.0837 | g_loss: 0.4075\n",
      "Epoch [   15/   20] | d_loss: 0.1779 | g_loss: 0.8691\n",
      "Epoch [   15/   20] | d_loss: 0.1926 | g_loss: 0.5879\n",
      "Epoch [   15/   20] | d_loss: 0.2802 | g_loss: 0.5089\n",
      "Epoch [   15/   20] | d_loss: 0.1428 | g_loss: 0.4596\n",
      "Epoch [   15/   20] | d_loss: 0.1313 | g_loss: 1.0122\n",
      "Epoch [   15/   20] | d_loss: 0.3380 | g_loss: 1.2201\n",
      "Epoch [   15/   20] | d_loss: 0.1468 | g_loss: 0.4296\n",
      "Epoch [   15/   20] | d_loss: 0.1596 | g_loss: 0.5813\n",
      "Epoch [   15/   20] | d_loss: 0.1330 | g_loss: 0.7372\n",
      "Epoch [   15/   20] | d_loss: 0.0956 | g_loss: 0.8745\n",
      "Epoch [   15/   20] | d_loss: 0.0775 | g_loss: 1.0763\n",
      "Epoch [   15/   20] | d_loss: 0.0758 | g_loss: 0.5811\n",
      "Epoch [   15/   20] | d_loss: 0.1762 | g_loss: 0.6785\n",
      "Epoch [   15/   20] | d_loss: 0.0818 | g_loss: 0.6057\n",
      "Epoch [   15/   20] | d_loss: 0.1340 | g_loss: 0.3747\n",
      "Epoch [   15/   20] | d_loss: 0.1252 | g_loss: 0.8195\n",
      "Epoch [   15/   20] | d_loss: 0.0666 | g_loss: 0.5671\n",
      "Epoch [   15/   20] | d_loss: 0.0730 | g_loss: 0.7802\n",
      "Epoch [   15/   20] | d_loss: 0.1231 | g_loss: 0.5458\n",
      "Epoch [   15/   20] | d_loss: 0.0607 | g_loss: 0.6260\n",
      "Epoch [   15/   20] | d_loss: 0.4999 | g_loss: 0.4824\n",
      "Epoch [   15/   20] | d_loss: 0.1761 | g_loss: 0.5947\n",
      "Epoch [   15/   20] | d_loss: 0.2512 | g_loss: 0.5850\n",
      "Epoch [   15/   20] | d_loss: 0.1099 | g_loss: 0.9232\n",
      "Epoch [   15/   20] | d_loss: 0.3403 | g_loss: 0.3997\n",
      "Epoch [   15/   20] | d_loss: 0.1708 | g_loss: 0.8507\n",
      "Epoch [   15/   20] | d_loss: 0.1363 | g_loss: 0.3203\n",
      "Epoch [   15/   20] | d_loss: 0.1388 | g_loss: 1.0833\n",
      "Epoch [   15/   20] | d_loss: 0.0615 | g_loss: 0.8172\n",
      "Epoch [   15/   20] | d_loss: 0.2703 | g_loss: 0.8641\n",
      "Epoch [   15/   20] | d_loss: 0.2137 | g_loss: 1.2091\n",
      "Epoch [   15/   20] | d_loss: 0.0469 | g_loss: 0.9736\n",
      "Epoch [   15/   20] | d_loss: 0.2198 | g_loss: 0.6514\n",
      "Epoch [   15/   20] | d_loss: 0.2312 | g_loss: 0.7025\n",
      "Epoch [   15/   20] | d_loss: 0.0825 | g_loss: 0.6790\n",
      "Epoch [   15/   20] | d_loss: 0.2186 | g_loss: 0.5353\n",
      "Epoch [   15/   20] | d_loss: 0.0329 | g_loss: 0.8259\n",
      "Epoch [   15/   20] | d_loss: 0.2791 | g_loss: 0.6628\n",
      "Epoch [   15/   20] | d_loss: 0.0886 | g_loss: 0.8422\n",
      "Epoch [   15/   20] | d_loss: 0.3162 | g_loss: 0.5461\n",
      "Epoch [   15/   20] | d_loss: 0.2642 | g_loss: 0.6723\n",
      "Epoch [   16/   20] | d_loss: 0.2430 | g_loss: 0.7024\n",
      "Epoch [   16/   20] | d_loss: 0.0656 | g_loss: 0.6138\n",
      "Epoch [   16/   20] | d_loss: 0.0591 | g_loss: 0.8073\n",
      "Epoch [   16/   20] | d_loss: 0.0729 | g_loss: 1.2085\n",
      "Epoch [   16/   20] | d_loss: 0.0836 | g_loss: 1.1173\n",
      "Epoch [   16/   20] | d_loss: 0.0298 | g_loss: 0.6879\n",
      "Epoch [   16/   20] | d_loss: 0.0447 | g_loss: 0.8397\n",
      "Epoch [   16/   20] | d_loss: 0.0574 | g_loss: 0.9228\n",
      "Epoch [   16/   20] | d_loss: 0.0492 | g_loss: 0.8502\n",
      "Epoch [   16/   20] | d_loss: 0.0269 | g_loss: 0.7902\n",
      "Epoch [   16/   20] | d_loss: 0.1672 | g_loss: 1.0175\n",
      "Epoch [   16/   20] | d_loss: 0.1524 | g_loss: 0.7764\n",
      "Epoch [   16/   20] | d_loss: 0.0510 | g_loss: 1.1561\n",
      "Epoch [   16/   20] | d_loss: 0.0497 | g_loss: 0.7105\n",
      "Epoch [   16/   20] | d_loss: 0.1057 | g_loss: 0.8793\n",
      "Epoch [   16/   20] | d_loss: 0.0623 | g_loss: 0.4112\n",
      "Epoch [   16/   20] | d_loss: 0.1925 | g_loss: 0.6077\n",
      "Epoch [   16/   20] | d_loss: 0.0794 | g_loss: 0.7808\n",
      "Epoch [   16/   20] | d_loss: 0.2343 | g_loss: 1.2303\n",
      "Epoch [   16/   20] | d_loss: 0.2346 | g_loss: 0.9778\n",
      "Epoch [   16/   20] | d_loss: 0.0469 | g_loss: 0.7666\n",
      "Epoch [   16/   20] | d_loss: 0.2291 | g_loss: 0.5439\n",
      "Epoch [   16/   20] | d_loss: 0.0989 | g_loss: 0.3767\n",
      "Epoch [   16/   20] | d_loss: 0.3356 | g_loss: 0.8960\n",
      "Epoch [   16/   20] | d_loss: 0.1956 | g_loss: 0.5567\n",
      "Epoch [   16/   20] | d_loss: 0.0626 | g_loss: 0.8113\n",
      "Epoch [   16/   20] | d_loss: 0.6374 | g_loss: 0.7091\n",
      "Epoch [   16/   20] | d_loss: 0.3246 | g_loss: 0.3967\n",
      "Epoch [   16/   20] | d_loss: 0.0689 | g_loss: 0.7244\n",
      "Epoch [   16/   20] | d_loss: 0.2251 | g_loss: 0.7302\n",
      "Epoch [   16/   20] | d_loss: 0.1126 | g_loss: 1.1795\n",
      "Epoch [   16/   20] | d_loss: 0.1848 | g_loss: 1.0437\n",
      "Epoch [   16/   20] | d_loss: 0.0987 | g_loss: 0.6177\n",
      "Epoch [   16/   20] | d_loss: 0.0592 | g_loss: 0.5439\n",
      "Epoch [   16/   20] | d_loss: 0.1135 | g_loss: 0.5499\n",
      "Epoch [   16/   20] | d_loss: 0.0405 | g_loss: 1.4298\n",
      "Epoch [   16/   20] | d_loss: 0.1594 | g_loss: 0.8776\n",
      "Epoch [   16/   20] | d_loss: 0.2169 | g_loss: 0.6466\n",
      "Epoch [   16/   20] | d_loss: 0.2149 | g_loss: 0.8183\n",
      "Epoch [   16/   20] | d_loss: 0.1515 | g_loss: 0.9199\n",
      "Epoch [   16/   20] | d_loss: 0.1885 | g_loss: 0.7218\n",
      "Epoch [   16/   20] | d_loss: 0.0630 | g_loss: 0.6518\n",
      "Epoch [   16/   20] | d_loss: 0.1039 | g_loss: 1.0070\n",
      "Epoch [   16/   20] | d_loss: 0.0921 | g_loss: 0.9849\n",
      "Epoch [   16/   20] | d_loss: 0.0945 | g_loss: 0.7610\n",
      "Epoch [   16/   20] | d_loss: 0.1743 | g_loss: 0.9171\n",
      "Epoch [   16/   20] | d_loss: 0.1878 | g_loss: 0.5314\n",
      "Epoch [   16/   20] | d_loss: 0.0589 | g_loss: 0.6217\n",
      "Epoch [   16/   20] | d_loss: 0.3567 | g_loss: 0.7622\n",
      "Epoch [   16/   20] | d_loss: 0.3025 | g_loss: 0.6811\n",
      "Epoch [   16/   20] | d_loss: 0.0526 | g_loss: 0.6830\n",
      "Epoch [   16/   20] | d_loss: 0.0683 | g_loss: 0.9942\n",
      "Epoch [   16/   20] | d_loss: 0.3018 | g_loss: 0.6583\n",
      "Epoch [   16/   20] | d_loss: 0.0786 | g_loss: 0.9250\n",
      "Epoch [   16/   20] | d_loss: 0.0634 | g_loss: 0.8564\n",
      "Epoch [   16/   20] | d_loss: 0.0235 | g_loss: 0.7295\n",
      "Epoch [   16/   20] | d_loss: 0.1586 | g_loss: 0.6879\n",
      "Epoch [   16/   20] | d_loss: 0.0667 | g_loss: 0.9544\n",
      "Epoch [   16/   20] | d_loss: 0.0530 | g_loss: 1.4842\n",
      "Epoch [   16/   20] | d_loss: 0.0685 | g_loss: 0.9563\n",
      "Epoch [   16/   20] | d_loss: 0.2242 | g_loss: 0.7038\n",
      "Epoch [   16/   20] | d_loss: 0.2033 | g_loss: 0.2621\n",
      "Epoch [   16/   20] | d_loss: 0.2030 | g_loss: 0.4548\n",
      "Epoch [   16/   20] | d_loss: 0.1521 | g_loss: 0.9368\n",
      "Epoch [   16/   20] | d_loss: 0.1275 | g_loss: 0.2212\n",
      "Epoch [   17/   20] | d_loss: 0.0797 | g_loss: 0.6955\n",
      "Epoch [   17/   20] | d_loss: 0.1433 | g_loss: 0.4733\n",
      "Epoch [   17/   20] | d_loss: 0.1094 | g_loss: 0.8592\n",
      "Epoch [   17/   20] | d_loss: 0.3581 | g_loss: 0.9733\n",
      "Epoch [   17/   20] | d_loss: 0.3152 | g_loss: 0.2374\n",
      "Epoch [   17/   20] | d_loss: 0.2536 | g_loss: 0.6524\n",
      "Epoch [   17/   20] | d_loss: 0.5239 | g_loss: 0.6148\n",
      "Epoch [   17/   20] | d_loss: 0.1343 | g_loss: 0.7957\n",
      "Epoch [   17/   20] | d_loss: 0.0783 | g_loss: 0.9557\n",
      "Epoch [   17/   20] | d_loss: 0.1032 | g_loss: 0.8382\n",
      "Epoch [   17/   20] | d_loss: 0.0671 | g_loss: 0.8568\n",
      "Epoch [   17/   20] | d_loss: 0.0581 | g_loss: 0.5509\n",
      "Epoch [   17/   20] | d_loss: 0.0886 | g_loss: 0.7577\n",
      "Epoch [   17/   20] | d_loss: 0.0818 | g_loss: 0.6938\n",
      "Epoch [   17/   20] | d_loss: 0.5292 | g_loss: 0.2535\n",
      "Epoch [   17/   20] | d_loss: 0.2389 | g_loss: 0.7984\n",
      "Epoch [   17/   20] | d_loss: 0.0705 | g_loss: 0.4099\n",
      "Epoch [   17/   20] | d_loss: 0.2555 | g_loss: 0.2594\n",
      "Epoch [   17/   20] | d_loss: 0.1147 | g_loss: 1.3171\n",
      "Epoch [   17/   20] | d_loss: 0.2400 | g_loss: 0.8205\n",
      "Epoch [   17/   20] | d_loss: 0.0282 | g_loss: 0.6509\n",
      "Epoch [   17/   20] | d_loss: 0.3366 | g_loss: 1.2993\n",
      "Epoch [   17/   20] | d_loss: 0.3776 | g_loss: 0.5825\n",
      "Epoch [   17/   20] | d_loss: 0.0718 | g_loss: 0.4595\n",
      "Epoch [   17/   20] | d_loss: 0.1179 | g_loss: 0.6121\n",
      "Epoch [   17/   20] | d_loss: 0.0531 | g_loss: 0.5637\n",
      "Epoch [   17/   20] | d_loss: 0.0610 | g_loss: 0.9379\n",
      "Epoch [   17/   20] | d_loss: 0.0628 | g_loss: 0.8714\n",
      "Epoch [   17/   20] | d_loss: 0.0707 | g_loss: 0.6772\n",
      "Epoch [   17/   20] | d_loss: 0.1546 | g_loss: 0.9956\n",
      "Epoch [   17/   20] | d_loss: 0.0831 | g_loss: 0.6170\n",
      "Epoch [   17/   20] | d_loss: 0.0553 | g_loss: 1.1718\n",
      "Epoch [   17/   20] | d_loss: 0.1200 | g_loss: 0.7078\n",
      "Epoch [   17/   20] | d_loss: 0.1316 | g_loss: 0.2575\n",
      "Epoch [   17/   20] | d_loss: 0.1839 | g_loss: 0.4670\n",
      "Epoch [   17/   20] | d_loss: 0.1446 | g_loss: 0.8722\n",
      "Epoch [   17/   20] | d_loss: 0.0873 | g_loss: 1.0775\n",
      "Epoch [   17/   20] | d_loss: 0.2188 | g_loss: 0.7071\n",
      "Epoch [   17/   20] | d_loss: 0.0543 | g_loss: 0.2566\n",
      "Epoch [   17/   20] | d_loss: 0.1545 | g_loss: 0.6441\n",
      "Epoch [   17/   20] | d_loss: 0.1386 | g_loss: 0.9608\n",
      "Epoch [   17/   20] | d_loss: 0.4595 | g_loss: 0.5266\n",
      "Epoch [   17/   20] | d_loss: 0.0881 | g_loss: 0.8161\n",
      "Epoch [   17/   20] | d_loss: 0.1391 | g_loss: 0.5701\n",
      "Epoch [   17/   20] | d_loss: 0.0327 | g_loss: 0.6722\n",
      "Epoch [   17/   20] | d_loss: 0.1501 | g_loss: 1.2247\n",
      "Epoch [   17/   20] | d_loss: 0.1119 | g_loss: 0.8627\n",
      "Epoch [   17/   20] | d_loss: 0.1707 | g_loss: 0.8219\n",
      "Epoch [   17/   20] | d_loss: 0.0350 | g_loss: 0.9085\n",
      "Epoch [   17/   20] | d_loss: 0.0869 | g_loss: 0.3745\n",
      "Epoch [   17/   20] | d_loss: 0.0860 | g_loss: 0.3381\n",
      "Epoch [   17/   20] | d_loss: 0.1975 | g_loss: 0.5471\n",
      "Epoch [   17/   20] | d_loss: 0.2201 | g_loss: 1.1627\n",
      "Epoch [   17/   20] | d_loss: 0.1228 | g_loss: 0.5691\n",
      "Epoch [   17/   20] | d_loss: 0.0859 | g_loss: 0.5308\n",
      "Epoch [   17/   20] | d_loss: 0.0811 | g_loss: 0.4082\n",
      "Epoch [   17/   20] | d_loss: 0.2005 | g_loss: 0.5103\n",
      "Epoch [   17/   20] | d_loss: 0.0716 | g_loss: 0.5023\n",
      "Epoch [   17/   20] | d_loss: 0.0275 | g_loss: 0.9279\n",
      "Epoch [   17/   20] | d_loss: 0.0702 | g_loss: 0.7273\n",
      "Epoch [   17/   20] | d_loss: 0.0725 | g_loss: 0.8526\n",
      "Epoch [   17/   20] | d_loss: 0.0431 | g_loss: 0.6478\n",
      "Epoch [   17/   20] | d_loss: 0.1459 | g_loss: 1.0165\n",
      "Epoch [   17/   20] | d_loss: 0.2196 | g_loss: 0.6672\n",
      "Epoch [   17/   20] | d_loss: 0.4764 | g_loss: 0.4465\n",
      "Epoch [   18/   20] | d_loss: 0.1329 | g_loss: 0.9408\n",
      "Epoch [   18/   20] | d_loss: 0.0792 | g_loss: 1.2915\n",
      "Epoch [   18/   20] | d_loss: 0.0928 | g_loss: 0.3167\n",
      "Epoch [   18/   20] | d_loss: 0.2552 | g_loss: 0.6952\n",
      "Epoch [   18/   20] | d_loss: 0.0798 | g_loss: 0.7723\n",
      "Epoch [   18/   20] | d_loss: 0.0335 | g_loss: 1.0345\n",
      "Epoch [   18/   20] | d_loss: 0.3831 | g_loss: 0.6389\n",
      "Epoch [   18/   20] | d_loss: 0.0935 | g_loss: 0.4759\n",
      "Epoch [   18/   20] | d_loss: 0.3185 | g_loss: 1.0724\n",
      "Epoch [   18/   20] | d_loss: 0.0897 | g_loss: 0.9859\n",
      "Epoch [   18/   20] | d_loss: 0.0560 | g_loss: 0.7870\n",
      "Epoch [   18/   20] | d_loss: 0.1564 | g_loss: 0.5739\n",
      "Epoch [   18/   20] | d_loss: 0.0614 | g_loss: 0.5375\n",
      "Epoch [   18/   20] | d_loss: 0.1072 | g_loss: 1.2224\n",
      "Epoch [   18/   20] | d_loss: 0.1984 | g_loss: 1.1800\n",
      "Epoch [   18/   20] | d_loss: 0.1212 | g_loss: 0.7282\n",
      "Epoch [   18/   20] | d_loss: 0.0283 | g_loss: 0.3104\n",
      "Epoch [   18/   20] | d_loss: 0.0555 | g_loss: 0.6243\n",
      "Epoch [   18/   20] | d_loss: 0.2002 | g_loss: 0.3131\n",
      "Epoch [   18/   20] | d_loss: 0.0381 | g_loss: 0.6909\n",
      "Epoch [   18/   20] | d_loss: 0.1319 | g_loss: 0.7332\n",
      "Epoch [   18/   20] | d_loss: 0.1174 | g_loss: 0.9413\n",
      "Epoch [   18/   20] | d_loss: 0.2111 | g_loss: 0.3671\n",
      "Epoch [   18/   20] | d_loss: 0.3544 | g_loss: 0.7575\n",
      "Epoch [   18/   20] | d_loss: 0.0503 | g_loss: 0.8531\n",
      "Epoch [   18/   20] | d_loss: 0.0403 | g_loss: 1.2691\n",
      "Epoch [   18/   20] | d_loss: 0.2732 | g_loss: 0.8095\n",
      "Epoch [   18/   20] | d_loss: 0.1221 | g_loss: 0.8960\n",
      "Epoch [   18/   20] | d_loss: 0.1591 | g_loss: 0.7577\n",
      "Epoch [   18/   20] | d_loss: 0.1900 | g_loss: 0.4040\n",
      "Epoch [   18/   20] | d_loss: 0.4835 | g_loss: 0.9106\n",
      "Epoch [   18/   20] | d_loss: 0.0305 | g_loss: 0.8431\n",
      "Epoch [   18/   20] | d_loss: 0.1501 | g_loss: 0.5098\n",
      "Epoch [   18/   20] | d_loss: 0.0437 | g_loss: 0.4506\n",
      "Epoch [   18/   20] | d_loss: 0.0371 | g_loss: 0.8657\n",
      "Epoch [   18/   20] | d_loss: 0.1207 | g_loss: 1.0943\n",
      "Epoch [   18/   20] | d_loss: 0.1073 | g_loss: 0.9687\n",
      "Epoch [   18/   20] | d_loss: 0.0663 | g_loss: 0.5271\n",
      "Epoch [   18/   20] | d_loss: 0.0792 | g_loss: 0.7251\n",
      "Epoch [   18/   20] | d_loss: 0.0913 | g_loss: 0.8300\n",
      "Epoch [   18/   20] | d_loss: 0.1544 | g_loss: 0.9054\n",
      "Epoch [   18/   20] | d_loss: 0.0580 | g_loss: 0.3825\n",
      "Epoch [   18/   20] | d_loss: 0.1572 | g_loss: 0.8529\n",
      "Epoch [   18/   20] | d_loss: 0.0926 | g_loss: 0.6564\n",
      "Epoch [   18/   20] | d_loss: 0.0537 | g_loss: 0.8822\n",
      "Epoch [   18/   20] | d_loss: 0.0725 | g_loss: 0.5219\n",
      "Epoch [   18/   20] | d_loss: 0.0552 | g_loss: 0.7135\n",
      "Epoch [   18/   20] | d_loss: 0.1204 | g_loss: 0.7997\n",
      "Epoch [   18/   20] | d_loss: 0.1227 | g_loss: 0.9602\n",
      "Epoch [   18/   20] | d_loss: 0.2749 | g_loss: 0.7811\n",
      "Epoch [   18/   20] | d_loss: 0.1597 | g_loss: 0.9488\n",
      "Epoch [   18/   20] | d_loss: 0.3344 | g_loss: 0.8746\n",
      "Epoch [   18/   20] | d_loss: 0.0981 | g_loss: 0.9760\n",
      "Epoch [   18/   20] | d_loss: 0.1206 | g_loss: 1.1623\n",
      "Epoch [   18/   20] | d_loss: 0.2279 | g_loss: 0.8411\n",
      "Epoch [   18/   20] | d_loss: 0.1924 | g_loss: 0.7251\n",
      "Epoch [   18/   20] | d_loss: 0.1459 | g_loss: 0.7328\n",
      "Epoch [   18/   20] | d_loss: 0.1469 | g_loss: 0.8948\n",
      "Epoch [   18/   20] | d_loss: 0.0799 | g_loss: 0.6582\n",
      "Epoch [   18/   20] | d_loss: 0.0711 | g_loss: 0.6221\n",
      "Epoch [   18/   20] | d_loss: 0.1596 | g_loss: 0.4234\n",
      "Epoch [   18/   20] | d_loss: 0.0789 | g_loss: 0.5271\n",
      "Epoch [   18/   20] | d_loss: 0.0677 | g_loss: 0.7436\n",
      "Epoch [   18/   20] | d_loss: 0.2718 | g_loss: 0.6119\n",
      "Epoch [   18/   20] | d_loss: 0.0632 | g_loss: 0.9314\n",
      "Epoch [   19/   20] | d_loss: 0.1191 | g_loss: 1.3996\n",
      "Epoch [   19/   20] | d_loss: 0.1773 | g_loss: 0.5581\n",
      "Epoch [   19/   20] | d_loss: 0.2196 | g_loss: 0.5564\n",
      "Epoch [   19/   20] | d_loss: 0.0494 | g_loss: 0.5194\n",
      "Epoch [   19/   20] | d_loss: 0.1111 | g_loss: 0.6186\n",
      "Epoch [   19/   20] | d_loss: 0.1086 | g_loss: 0.9303\n",
      "Epoch [   19/   20] | d_loss: 0.0675 | g_loss: 0.7796\n",
      "Epoch [   19/   20] | d_loss: 0.3380 | g_loss: 0.2641\n",
      "Epoch [   19/   20] | d_loss: 0.0602 | g_loss: 0.6980\n",
      "Epoch [   19/   20] | d_loss: 0.3141 | g_loss: 0.6261\n",
      "Epoch [   19/   20] | d_loss: 0.1035 | g_loss: 0.7551\n",
      "Epoch [   19/   20] | d_loss: 0.1552 | g_loss: 0.4205\n",
      "Epoch [   19/   20] | d_loss: 0.0900 | g_loss: 1.0715\n",
      "Epoch [   19/   20] | d_loss: 0.0851 | g_loss: 1.1312\n",
      "Epoch [   19/   20] | d_loss: 0.0414 | g_loss: 0.7129\n",
      "Epoch [   19/   20] | d_loss: 0.1446 | g_loss: 0.7449\n",
      "Epoch [   19/   20] | d_loss: 0.0977 | g_loss: 1.3797\n",
      "Epoch [   19/   20] | d_loss: 0.1793 | g_loss: 1.2071\n",
      "Epoch [   19/   20] | d_loss: 0.0475 | g_loss: 0.8982\n",
      "Epoch [   19/   20] | d_loss: 0.1111 | g_loss: 0.4846\n",
      "Epoch [   19/   20] | d_loss: 0.2010 | g_loss: 0.3243\n",
      "Epoch [   19/   20] | d_loss: 0.2079 | g_loss: 0.6163\n",
      "Epoch [   19/   20] | d_loss: 0.1027 | g_loss: 0.6374\n",
      "Epoch [   19/   20] | d_loss: 0.2287 | g_loss: 0.7846\n",
      "Epoch [   19/   20] | d_loss: 0.2066 | g_loss: 0.6471\n",
      "Epoch [   19/   20] | d_loss: 0.0473 | g_loss: 0.4858\n",
      "Epoch [   19/   20] | d_loss: 0.0891 | g_loss: 0.9284\n",
      "Epoch [   19/   20] | d_loss: 0.0401 | g_loss: 0.5049\n",
      "Epoch [   19/   20] | d_loss: 0.0576 | g_loss: 1.1662\n",
      "Epoch [   19/   20] | d_loss: 0.0487 | g_loss: 1.2265\n",
      "Epoch [   19/   20] | d_loss: 0.0313 | g_loss: 0.8916\n",
      "Epoch [   19/   20] | d_loss: 0.3346 | g_loss: 0.3248\n",
      "Epoch [   19/   20] | d_loss: 0.0839 | g_loss: 0.8563\n",
      "Epoch [   19/   20] | d_loss: 0.2152 | g_loss: 0.9495\n",
      "Epoch [   19/   20] | d_loss: 0.0760 | g_loss: 0.3141\n",
      "Epoch [   19/   20] | d_loss: 0.1717 | g_loss: 0.5228\n",
      "Epoch [   19/   20] | d_loss: 0.0923 | g_loss: 0.9705\n",
      "Epoch [   19/   20] | d_loss: 0.1896 | g_loss: 0.5514\n",
      "Epoch [   19/   20] | d_loss: 0.0702 | g_loss: 0.5959\n",
      "Epoch [   19/   20] | d_loss: 0.0329 | g_loss: 1.3047\n",
      "Epoch [   19/   20] | d_loss: 0.1972 | g_loss: 1.0361\n",
      "Epoch [   19/   20] | d_loss: 0.4400 | g_loss: 1.0444\n",
      "Epoch [   19/   20] | d_loss: 0.0351 | g_loss: 0.8020\n",
      "Epoch [   19/   20] | d_loss: 0.1147 | g_loss: 0.5053\n",
      "Epoch [   19/   20] | d_loss: 0.0530 | g_loss: 1.1237\n",
      "Epoch [   19/   20] | d_loss: 0.1526 | g_loss: 1.0399\n",
      "Epoch [   19/   20] | d_loss: 0.0941 | g_loss: 0.9538\n",
      "Epoch [   19/   20] | d_loss: 0.0203 | g_loss: 0.5778\n",
      "Epoch [   19/   20] | d_loss: 0.2994 | g_loss: 0.5588\n",
      "Epoch [   19/   20] | d_loss: 0.1614 | g_loss: 0.6466\n",
      "Epoch [   19/   20] | d_loss: 0.0453 | g_loss: 0.9857\n",
      "Epoch [   19/   20] | d_loss: 0.0739 | g_loss: 1.1934\n",
      "Epoch [   19/   20] | d_loss: 0.1834 | g_loss: 0.4690\n",
      "Epoch [   19/   20] | d_loss: 0.2931 | g_loss: 0.6366\n",
      "Epoch [   19/   20] | d_loss: 0.1085 | g_loss: 0.6609\n",
      "Epoch [   19/   20] | d_loss: 0.0570 | g_loss: 0.6098\n",
      "Epoch [   19/   20] | d_loss: 0.1308 | g_loss: 0.3697\n",
      "Epoch [   19/   20] | d_loss: 0.0397 | g_loss: 0.5135\n",
      "Epoch [   19/   20] | d_loss: 0.0356 | g_loss: 0.9302\n",
      "Epoch [   19/   20] | d_loss: 0.0392 | g_loss: 1.3437\n",
      "Epoch [   19/   20] | d_loss: 0.0434 | g_loss: 1.1177\n",
      "Epoch [   19/   20] | d_loss: 0.1347 | g_loss: 0.8424\n",
      "Epoch [   19/   20] | d_loss: 0.0686 | g_loss: 0.8426\n",
      "Epoch [   19/   20] | d_loss: 0.1174 | g_loss: 0.7692\n",
      "Epoch [   19/   20] | d_loss: 0.0573 | g_loss: 0.9255\n",
      "Epoch [   20/   20] | d_loss: 0.2114 | g_loss: 0.7282\n",
      "Epoch [   20/   20] | d_loss: 0.0878 | g_loss: 0.8178\n",
      "Epoch [   20/   20] | d_loss: 0.2249 | g_loss: 0.4050\n",
      "Epoch [   20/   20] | d_loss: 0.0958 | g_loss: 0.8766\n",
      "Epoch [   20/   20] | d_loss: 0.1935 | g_loss: 0.9132\n",
      "Epoch [   20/   20] | d_loss: 0.0566 | g_loss: 1.4690\n",
      "Epoch [   20/   20] | d_loss: 0.1604 | g_loss: 0.7871\n",
      "Epoch [   20/   20] | d_loss: 0.0908 | g_loss: 0.9756\n",
      "Epoch [   20/   20] | d_loss: 0.0510 | g_loss: 0.6039\n",
      "Epoch [   20/   20] | d_loss: 0.1355 | g_loss: 0.3004\n",
      "Epoch [   20/   20] | d_loss: 0.1771 | g_loss: 0.8254\n",
      "Epoch [   20/   20] | d_loss: 0.2093 | g_loss: 0.8296\n",
      "Epoch [   20/   20] | d_loss: 0.1866 | g_loss: 0.7436\n",
      "Epoch [   20/   20] | d_loss: 0.2111 | g_loss: 0.8757\n",
      "Epoch [   20/   20] | d_loss: 0.0722 | g_loss: 0.5344\n",
      "Epoch [   20/   20] | d_loss: 0.1275 | g_loss: 0.5282\n",
      "Epoch [   20/   20] | d_loss: 0.1960 | g_loss: 1.1327\n",
      "Epoch [   20/   20] | d_loss: 0.0575 | g_loss: 0.9507\n",
      "Epoch [   20/   20] | d_loss: 0.0536 | g_loss: 0.2814\n",
      "Epoch [   20/   20] | d_loss: 0.1853 | g_loss: 0.6936\n",
      "Epoch [   20/   20] | d_loss: 0.1079 | g_loss: 0.9839\n",
      "Epoch [   20/   20] | d_loss: 0.4202 | g_loss: 0.5811\n",
      "Epoch [   20/   20] | d_loss: 0.2819 | g_loss: 0.0196\n",
      "Epoch [   20/   20] | d_loss: 0.1050 | g_loss: 0.4898\n",
      "Epoch [   20/   20] | d_loss: 0.1700 | g_loss: 0.8424\n",
      "Epoch [   20/   20] | d_loss: 0.2399 | g_loss: 0.2922\n",
      "Epoch [   20/   20] | d_loss: 0.1245 | g_loss: 0.9163\n",
      "Epoch [   20/   20] | d_loss: 0.4611 | g_loss: 0.7797\n",
      "Epoch [   20/   20] | d_loss: 0.1145 | g_loss: 0.7530\n",
      "Epoch [   20/   20] | d_loss: 0.1263 | g_loss: 0.6843\n",
      "Epoch [   20/   20] | d_loss: 0.1400 | g_loss: 0.8946\n",
      "Epoch [   20/   20] | d_loss: 0.0453 | g_loss: 0.5388\n",
      "Epoch [   20/   20] | d_loss: 0.1114 | g_loss: 0.6657\n",
      "Epoch [   20/   20] | d_loss: 0.1501 | g_loss: 0.4918\n",
      "Epoch [   20/   20] | d_loss: 0.1776 | g_loss: 1.5967\n",
      "Epoch [   20/   20] | d_loss: 0.0968 | g_loss: 0.7469\n",
      "Epoch [   20/   20] | d_loss: 0.1247 | g_loss: 0.4607\n",
      "Epoch [   20/   20] | d_loss: 0.0456 | g_loss: 0.9833\n",
      "Epoch [   20/   20] | d_loss: 0.1111 | g_loss: 0.8433\n",
      "Epoch [   20/   20] | d_loss: 0.0291 | g_loss: 0.7010\n",
      "Epoch [   20/   20] | d_loss: 0.1511 | g_loss: 0.5782\n",
      "Epoch [   20/   20] | d_loss: 0.1257 | g_loss: 0.4822\n",
      "Epoch [   20/   20] | d_loss: 0.2795 | g_loss: 0.4985\n",
      "Epoch [   20/   20] | d_loss: 0.1701 | g_loss: 0.9609\n",
      "Epoch [   20/   20] | d_loss: 0.0482 | g_loss: 0.6467\n",
      "Epoch [   20/   20] | d_loss: 0.1579 | g_loss: 0.8140\n",
      "Epoch [   20/   20] | d_loss: 0.1333 | g_loss: 1.3345\n",
      "Epoch [   20/   20] | d_loss: 0.1047 | g_loss: 0.7364\n",
      "Epoch [   20/   20] | d_loss: 0.0589 | g_loss: 1.1328\n",
      "Epoch [   20/   20] | d_loss: 0.1003 | g_loss: 0.9778\n",
      "Epoch [   20/   20] | d_loss: 0.0344 | g_loss: 0.9380\n",
      "Epoch [   20/   20] | d_loss: 0.1226 | g_loss: 0.7384\n",
      "Epoch [   20/   20] | d_loss: 0.6456 | g_loss: 0.7904\n",
      "Epoch [   20/   20] | d_loss: 0.4244 | g_loss: 0.8395\n",
      "Epoch [   20/   20] | d_loss: 0.0842 | g_loss: 1.1326\n",
      "Epoch [   20/   20] | d_loss: 0.1566 | g_loss: 1.2514\n",
      "Epoch [   20/   20] | d_loss: 0.1135 | g_loss: 0.6874\n",
      "Epoch [   20/   20] | d_loss: 0.1300 | g_loss: 1.0661\n",
      "Epoch [   20/   20] | d_loss: 0.1871 | g_loss: 0.4388\n",
      "Epoch [   20/   20] | d_loss: 0.1301 | g_loss: 0.9531\n",
      "Epoch [   20/   20] | d_loss: 0.0916 | g_loss: 1.0771\n",
      "Epoch [   20/   20] | d_loss: 0.0456 | g_loss: 0.4594\n",
      "Epoch [   20/   20] | d_loss: 0.2667 | g_loss: 0.7446\n",
      "Epoch [   20/   20] | d_loss: 0.1101 | g_loss: 1.0375\n",
      "Epoch [   20/   20] | d_loss: 0.3462 | g_loss: 0.8918\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",
    "with active_session():\n",
    "    losses = train(D, G, n_epochs=n_epochs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Training loss\n",
    "\n",
    "Plot the training losses for the generator and discriminator, recorded after each epoch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7ff388c06be0>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff386a693c8>"
      ]
     },
     "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": 24,
   "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": 25,
   "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": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff36f14a6a0>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "_ = view_samples(-1, samples)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Question: What do you notice about your generated samples and how might you improve this model?\n",
    "When you answer this question, consider the following factors:\n",
    "* The dataset is biased; it is made of \"celebrity\" faces that are mostly white\n",
    "* Model size; larger models have the opportunity to learn more features in a data feature space\n",
    "* Optimization strategy; optimizers and number of epochs affect your final result\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Answer:** \n",
    "Here are problems along with its potential solutions:\n",
    "1. **Problem  :** Most generated faces are white. Therefore, the model is bias to the white people.\n",
    "   **Solution :** Balance trainset for all races\n",
    "   \n",
    "2. **Problem  :** There are some flaw face features, such as mouth, eyes, nose.\n",
    "   **Solution :** Make the network larger to recognite wider features and train with more epochs.\n",
    "   \n",
    "3. **Problem  :** The image has low resolution\n",
    "   **Solution :** Use higher resolution on training images and potentially use high resolution GAN.\n",
    "   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting This Project\n",
    "When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as \"dlnd_face_generation.ipynb\" and save it as a HTML file under \"File\" -> \"Download as\". Include the \"problem_unittests.py\" files in your submission."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}