dlnd_face_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Face Generation\n",
    "\n",
    "In this project, you'll define and train a DCGAN on a dataset of faces. Your goal is to get a generator network to generate *new* images of faces that look as realistic as possible!\n",
    "\n",
    "The project will be broken down into a series of tasks from **loading in data to defining and training adversarial networks**. At the end of the notebook, you'll be able to visualize the results of your trained Generator to see how it performs; your generated samples should look like fairly realistic faces with small amounts of noise.\n",
    "\n",
    "### Get the Data\n",
    "\n",
    "You'll be using the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train your adversarial networks.\n",
    "\n",
    "This dataset is more complex than the number datasets (like MNIST or SVHN) you've been working with, and so, you should prepare to define deeper networks and train them for a longer time to get good results. It is suggested that you utilize a GPU for training.\n",
    "\n",
    "### Pre-processed Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. Some sample data is show below.\n",
    "\n",
    "<img src='assets/processed_face_data.png' width=60% />\n",
    "\n",
    "> If you are working locally, you can download this data [by clicking here](https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip)\n",
    "\n",
    "This is a zip file that you'll need to extract in the home directory of this notebook for further loading and processing. After extracting the data, you should be left with a directory of data `processed_celeba_small/`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "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",
    "    \n",
    "    # TODO: Implement function and return a dataloader\n",
    "    #define compose transforms\n",
    "    transform = transforms.Compose([transforms.Resize(image_size),\n",
    "                                transforms.ToTensor()])\n",
    "    dataset = datasets.ImageFolder(data_dir,transform = transform)\n",
    "    dataloader = torch.utils.data.DataLoader(dataset, batch_size= batch_size, shuffle=True)\n",
    "    \n",
    "    return dataloader\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 = 20\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 0x7fb88b1c8f28>"
      ]
     },
     "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",
    "    # assume x is scaled to (0, 1)\n",
    "    # scale to feature_range and return scaled x\n",
    "    x = x*2\n",
    "    x = x-1\n",
    "    \n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.8431)\n",
      "Max:  tensor(0.8510)\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, out_channels, \n",
    "                           kernel_size, stride, padding, bias=False)\n",
    "    \n",
    "    # append conv layer\n",
    "    layers.append(conv_layer)\n",
    "\n",
    "    if batch_norm:\n",
    "        # append batchnorm layer\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "     \n",
    "    # using Sequential container\n",
    "    return nn.Sequential(*layers)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tests Passed\n"
     ]
    }
   ],
   "source": [
    "class Discriminator(nn.Module):\n",
    "\n",
    "    def __init__(self, conv_dim):\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",
    "        # complete init function\n",
    "        self.conv_dim = conv_dim\n",
    "\n",
    "        # 32x32 input\n",
    "        self.conv1 = conv(3, conv_dim, 4, batch_norm=False) # first layer, no batch_norm\n",
    "        # 16x16 out\n",
    "        self.conv2 = conv(conv_dim, conv_dim*2, 4)\n",
    "        # 8x8 out\n",
    "        self.conv3 = conv(conv_dim*2, conv_dim*4, 4)\n",
    "        # 4x4 out\n",
    "        \n",
    "        # final, fully-connected layer\n",
    "        self.fc = nn.Linear(conv_dim*4*4*4, 1)\n",
    "\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 hidden layers + leaky relu activation\n",
    "        out = F.leaky_relu(self.conv1(x), 0.2)\n",
    "        out = F.leaky_relu(self.conv2(out), 0.2)\n",
    "        out = F.leaky_relu(self.conv3(out), 0.2)\n",
    "        \n",
    "        # flatten\n",
    "        out = out.view(-1, self.conv_dim*4*4*4)\n",
    "        \n",
    "        # final output layer\n",
    "        out = self.fc(out)        \n",
    "        return out\n",
    "        return x\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "tests.test_discriminator(Discriminator)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generator\n",
    "\n",
    "The generator should upsample an input and generate a *new* image of the same size as our training data `32x32x3`. This should be mostly transpose convolutional layers with normalization applied to the outputs.\n",
    "\n",
    "#### Exercise: Complete the Generator class\n",
    "* The inputs to the generator are vectors of some length `z_size`\n",
    "* The output should be a image of shape `32x32x3`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 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 transposed-convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    # create a sequence of transpose + optional batch norm layers\n",
    "    layers = []\n",
    "    transpose_conv_layer = nn.ConvTranspose2d(in_channels, out_channels, \n",
    "                                              kernel_size, stride, padding, bias=False)\n",
    "    # append transpose convolutional layer\n",
    "    layers.append(transpose_conv_layer)\n",
    "    \n",
    "    if batch_norm:\n",
    "        # append batchnorm layer\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "        \n",
    "    return nn.Sequential(*layers)\n"
   ]
  },
  {
   "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",
    "        # complete init function\n",
    "        super(Generator, self).__init__()\n",
    "\n",
    "        # complete init function\n",
    "        \n",
    "        self.conv_dim = conv_dim\n",
    "        \n",
    "        # first, fully-connected layer\n",
    "        self.fc = nn.Linear(z_size, conv_dim*4*4*4)\n",
    "\n",
    "        # transpose conv layers\n",
    "        self.t_conv1 = deconv(conv_dim*4, conv_dim*2, 4)\n",
    "        self.t_conv2 = deconv(conv_dim*2, conv_dim, 4)\n",
    "        self.t_conv3 = deconv(conv_dim, 3, 4, batch_norm=False)\n",
    "\n",
    "    def forward(self, x):\n",
    "        \"\"\"\n",
    "        Forward propagation of the neural network\n",
    "        :param x: The input to the neural network     \n",
    "        :return: A 32x32x3 Tensor image as output\n",
    "        \"\"\"\n",
    "        # define feedforward behavior\n",
    "        # fully-connected + reshape \n",
    "        out = self.fc(x)\n",
    "        out = out.view(-1, self.conv_dim*4, 4, 4) # (batch_size, depth, 4, 4)\n",
    "        \n",
    "        # hidden transpose conv layers + relu\n",
    "        out = F.relu(self.t_conv1(out))\n",
    "        out = F.relu(self.t_conv2(out))\n",
    "        \n",
    "        # last layer + tanh activation\n",
    "        out = self.t_conv3(out)\n",
    "        out = F.tanh(out)\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": [
    "def weights_init_normal(m):\n",
    "    \"\"\"\n",
    "    Applies initial weights to certain layers in a model .\n",
    "    The weights are taken from a normal distribution \n",
    "    with mean = 0, std dev = 0.02.\n",
    "    :param m: A module or layer in a network    \n",
    "    \"\"\"\n",
    "    # classname will be something like:\n",
    "    # `Conv`, `BatchNorm2d`, `Linear`, etc.\n",
    "    classname = m.__class__.__name__\n",
    "    \n",
    "    # TODO: Apply initial weights to convolutional and linear layers\n",
    "    if classname.find(\"Conv\") != -1:\n",
    "        torch.nn.init.normal_(m.weight.data, 0.0, 0.02)\n",
    "    elif classname.find(\"Linear\") != -1:\n",
    "        torch.nn.init.normal_(m.weight.data, 0.0, 0.02)\n",
    "        torch.nn.init.constant_(m.bias.data, 0.0)\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",
      "  (fc): Linear(in_features=2048, out_features=1, bias=True)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fc): Linear(in_features=100, out_features=2048, bias=True)\n",
      "  (t_conv1): 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_conv2): Sequential(\n",
      "    (0): ConvTranspose2d(64, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv3): Sequential(\n",
      "    (0): ConvTranspose2d(32, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 32\n",
    "g_conv_dim = 32\n",
    "z_size = 100\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE\n",
    "\"\"\"\n",
    "D, G = build_network(d_conv_dim, g_conv_dim, z_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Training on GPU\n",
    "\n",
    "Check if you can train on GPU. Here, we'll set this as a boolean variable `train_on_gpu`. Later, you'll be responsible for making sure that \n",
    ">* Models,\n",
    "* Model inputs, and\n",
    "* Loss function arguments\n",
    "\n",
    "Are moved to GPU, where appropriate."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 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 - 0.9)**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",
    "lr = 0.0002\n",
    "beta1 = 0.5\n",
    "beta2 = 0.999\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": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(D, G, n_epochs, print_every=50):\n",
    "    '''Trains adversarial networks for some number of epochs\n",
    "       param, D: the discriminator network\n",
    "       param, G: the generator network\n",
    "       param, n_epochs: number of epochs to train for\n",
    "       param, print_every: when to print and record the models' losses\n",
    "       return: D and G losses'''\n",
    "    \n",
    "    # move models to GPU\n",
    "    if train_on_gpu:\n",
    "        D.cuda()\n",
    "        G.cuda()\n",
    "\n",
    "    # keep track of loss and generated, \"fake\" samples\n",
    "    samples = []\n",
    "    losses = []\n",
    "\n",
    "    # Get some fixed data for sampling. These are images that are held\n",
    "    # constant throughout training, and allow us to inspect the model's performance\n",
    "    sample_size=16\n",
    "    fixed_z = np.random.uniform(-1, 1, size=(sample_size, z_size))\n",
    "    fixed_z = torch.from_numpy(fixed_z).float()\n",
    "    # move z to GPU if available\n",
    "    if train_on_gpu:\n",
    "        fixed_z = fixed_z.cuda()\n",
    "\n",
    "    # epoch training loop\n",
    "    for epoch in range(n_epochs):\n",
    "\n",
    "        # batch training loop\n",
    "        for batch_i, (real_images, _) in enumerate(celeba_train_loader):\n",
    "\n",
    "            batch_size = real_images.size(0)\n",
    "            real_images = scale(real_images)\n",
    "\n",
    "            # ===============================================\n",
    "            #         YOUR CODE HERE: TRAIN THE NETWORKS\n",
    "            # ===============================================\n",
    "            d_optimizer.zero_grad()\n",
    "            # 1. Train the discriminator on real and fake images\n",
    "            if train_on_gpu:\n",
    "                real_images = real_images.cuda()\n",
    "\n",
    "            D_real = D(real_images)\n",
    "            d_real_loss = real_loss(D_real)\n",
    "\n",
    "            # 2. Train with fake images\n",
    "\n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            # move x to GPU, if available\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "            fake_images = G(z)\n",
    "\n",
    "            # Compute the discriminator losses on fake images            \n",
    "            D_fake = D(fake_images)\n",
    "            d_fake_loss = fake_loss(D_fake)\n",
    "\n",
    "            # add up loss and perform backprop\n",
    "            d_loss = d_real_loss + d_fake_loss\n",
    "            d_loss.backward()\n",
    "            d_optimizer.step()\n",
    "\n",
    "            # 2. Train the generator with an adversarial loss\n",
    "            g_optimizer.zero_grad()\n",
    "        \n",
    "            # 1. Train with fake images and flipped labels\n",
    "\n",
    "            # Generate fake images\n",
    "            z = np.random.uniform(-1, 1, size=(batch_size, z_size))\n",
    "            z = torch.from_numpy(z).float()\n",
    "            if train_on_gpu:\n",
    "                z = z.cuda()\n",
    "            fake_images = G(z)\n",
    "\n",
    "            # Compute the discriminator losses on fake images \n",
    "            # using flipped labels!\n",
    "            D_fake = D(fake_images)\n",
    "            g_loss = real_loss(D_fake) # use real loss to flip labels\n",
    "\n",
    "            # perform backprop\n",
    "            g_loss.backward()\n",
    "            g_optimizer.step()\n",
    "            \n",
    "            \n",
    "            # ===============================================\n",
    "            #              END OF YOUR CODE\n",
    "            # ===============================================\n",
    "\n",
    "            # Print some loss stats\n",
    "            if batch_i % print_every == 0:\n",
    "                # append discriminator loss and generator loss\n",
    "                losses.append((d_loss.item(), g_loss.item()))\n",
    "                # print discriminator and generator loss\n",
    "                print('Epoch [{:5d}/{:5d}] | d_loss: {:6.4f} | g_loss: {:6.4f}'.format(\n",
    "                        epoch+1, n_epochs, d_loss.item(), g_loss.item()))\n",
    "\n",
    "\n",
    "        ## AFTER EACH EPOCH##    \n",
    "        # this code assumes your generator is named G, feel free to change the name\n",
    "        # generate and save sample, fake images\n",
    "        G.eval() # for generating samples\n",
    "        samples_z = G(fixed_z)\n",
    "        samples.append(samples_z)\n",
    "        G.train() # back to training mode\n",
    "\n",
    "    # Save training generator samples\n",
    "    with open('train_samples.pkl', 'wb') as f:\n",
    "        pkl.dump(samples, f)\n",
    "    \n",
    "    # finally return losses\n",
    "    return losses"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set your number of training epochs and train your GAN!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   20] | d_loss: 0.1920 | g_loss: 0.2697\n",
      "Epoch [    1/   20] | d_loss: 0.2080 | g_loss: 0.5062\n",
      "Epoch [    1/   20] | d_loss: 0.1850 | g_loss: 0.3347\n",
      "Epoch [    1/   20] | d_loss: 0.2009 | g_loss: 0.6497\n",
      "Epoch [    1/   20] | d_loss: 0.3008 | g_loss: 0.7407\n",
      "Epoch [    1/   20] | d_loss: 0.3199 | g_loss: 0.3658\n",
      "Epoch [    1/   20] | d_loss: 0.1701 | g_loss: 0.8126\n",
      "Epoch [    1/   20] | d_loss: 0.3034 | g_loss: 0.6900\n",
      "Epoch [    1/   20] | d_loss: 0.3944 | g_loss: 0.2622\n",
      "Epoch [    1/   20] | d_loss: 0.1906 | g_loss: 0.4043\n",
      "Epoch [    1/   20] | d_loss: 0.4321 | g_loss: 0.7866\n",
      "Epoch [    1/   20] | d_loss: 0.3944 | g_loss: 0.3619\n",
      "Epoch [    1/   20] | d_loss: 0.4740 | g_loss: 0.4860\n",
      "Epoch [    1/   20] | d_loss: 0.2215 | g_loss: 0.5381\n",
      "Epoch [    1/   20] | d_loss: 0.2321 | g_loss: 0.6080\n",
      "Epoch [    1/   20] | d_loss: 0.2463 | g_loss: 0.4353\n",
      "Epoch [    1/   20] | d_loss: 0.3537 | g_loss: 0.4975\n",
      "Epoch [    1/   20] | d_loss: 0.3121 | g_loss: 0.4079\n",
      "Epoch [    1/   20] | d_loss: 0.3065 | g_loss: 0.3397\n",
      "Epoch [    1/   20] | d_loss: 0.3576 | g_loss: 0.2946\n",
      "Epoch [    1/   20] | d_loss: 0.3161 | g_loss: 0.0973\n",
      "Epoch [    1/   20] | d_loss: 0.3442 | g_loss: 0.7850\n",
      "Epoch [    1/   20] | d_loss: 0.1241 | g_loss: 0.6660\n",
      "Epoch [    1/   20] | d_loss: 0.3239 | g_loss: 0.2760\n",
      "Epoch [    1/   20] | d_loss: 0.2930 | g_loss: 0.2440\n",
      "Epoch [    1/   20] | d_loss: 0.2681 | g_loss: 0.5035\n",
      "Epoch [    1/   20] | d_loss: 0.1580 | g_loss: 0.4935\n",
      "Epoch [    1/   20] | d_loss: 0.2975 | g_loss: 0.2800\n",
      "Epoch [    1/   20] | d_loss: 0.3961 | g_loss: 0.2894\n",
      "Epoch [    1/   20] | d_loss: 0.3029 | g_loss: 0.2475\n",
      "Epoch [    1/   20] | d_loss: 0.2198 | g_loss: 0.4520\n",
      "Epoch [    1/   20] | d_loss: 0.3540 | g_loss: 0.3101\n",
      "Epoch [    1/   20] | d_loss: 0.3330 | g_loss: 0.4568\n",
      "Epoch [    1/   20] | d_loss: 0.3160 | g_loss: 0.2132\n",
      "Epoch [    1/   20] | d_loss: 0.1771 | g_loss: 0.5352\n",
      "Epoch [    1/   20] | d_loss: 0.2474 | g_loss: 0.4390\n",
      "Epoch [    1/   20] | d_loss: 0.3880 | g_loss: 0.3173\n",
      "Epoch [    1/   20] | d_loss: 0.3475 | g_loss: 0.5520\n",
      "Epoch [    1/   20] | d_loss: 0.2812 | g_loss: 0.2335\n",
      "Epoch [    1/   20] | d_loss: 0.2398 | g_loss: 0.6658\n",
      "Epoch [    1/   20] | d_loss: 0.2231 | g_loss: 0.3742\n",
      "Epoch [    1/   20] | d_loss: 0.2603 | g_loss: 0.6178\n",
      "Epoch [    1/   20] | d_loss: 0.2915 | g_loss: 0.3052\n",
      "Epoch [    1/   20] | d_loss: 0.2684 | g_loss: 0.4374\n",
      "Epoch [    1/   20] | d_loss: 0.3777 | g_loss: 0.3329\n",
      "Epoch [    1/   20] | d_loss: 0.4359 | g_loss: 0.4253\n",
      "Epoch [    1/   20] | d_loss: 0.4528 | g_loss: 0.3575\n",
      "Epoch [    1/   20] | d_loss: 0.2878 | g_loss: 0.4890\n",
      "Epoch [    1/   20] | d_loss: 0.4714 | g_loss: 0.2570\n",
      "Epoch [    1/   20] | d_loss: 0.2315 | g_loss: 0.4838\n",
      "Epoch [    1/   20] | d_loss: 0.2755 | g_loss: 0.2510\n",
      "Epoch [    1/   20] | d_loss: 0.2250 | g_loss: 0.3388\n",
      "Epoch [    1/   20] | d_loss: 0.2449 | g_loss: 0.7008\n",
      "Epoch [    1/   20] | d_loss: 0.3256 | g_loss: 0.4192\n",
      "Epoch [    1/   20] | d_loss: 0.3387 | g_loss: 0.2159\n",
      "Epoch [    1/   20] | d_loss: 0.3790 | g_loss: 0.3307\n",
      "Epoch [    1/   20] | d_loss: 0.2252 | g_loss: 0.4427\n",
      "Epoch [    1/   20] | d_loss: 0.0797 | g_loss: 0.4848\n",
      "Epoch [    1/   20] | d_loss: 0.1569 | g_loss: 0.3678\n",
      "Epoch [    1/   20] | d_loss: 0.3135 | g_loss: 0.5922\n",
      "Epoch [    1/   20] | d_loss: 0.1639 | g_loss: 0.5072\n",
      "Epoch [    1/   20] | d_loss: 0.2024 | g_loss: 0.3906\n",
      "Epoch [    1/   20] | d_loss: 0.2609 | g_loss: 0.4083\n",
      "Epoch [    1/   20] | d_loss: 0.2699 | g_loss: 0.3804\n",
      "Epoch [    1/   20] | d_loss: 0.4043 | g_loss: 0.3050\n",
      "Epoch [    1/   20] | d_loss: 0.3682 | g_loss: 0.1139\n",
      "Epoch [    1/   20] | d_loss: 0.2947 | g_loss: 0.3908\n",
      "Epoch [    1/   20] | d_loss: 0.2479 | g_loss: 0.5316\n",
      "Epoch [    1/   20] | d_loss: 0.3405 | g_loss: 0.3110\n",
      "Epoch [    1/   20] | d_loss: 0.2098 | g_loss: 0.5447\n",
      "Epoch [    1/   20] | d_loss: 0.3275 | g_loss: 0.4644\n",
      "Epoch [    1/   20] | d_loss: 0.2524 | g_loss: 0.3782\n",
      "Epoch [    1/   20] | d_loss: 0.2690 | g_loss: 0.6576\n",
      "Epoch [    1/   20] | d_loss: 0.2207 | g_loss: 0.6448\n",
      "Epoch [    1/   20] | d_loss: 0.2452 | g_loss: 0.4058\n",
      "Epoch [    1/   20] | d_loss: 0.2741 | g_loss: 0.3409\n",
      "Epoch [    1/   20] | d_loss: 0.2053 | g_loss: 0.2412\n",
      "Epoch [    1/   20] | d_loss: 0.3020 | g_loss: 0.7818\n",
      "Epoch [    1/   20] | d_loss: 0.2114 | g_loss: 0.5382\n",
      "Epoch [    1/   20] | d_loss: 0.1400 | g_loss: 0.3841\n",
      "Epoch [    1/   20] | d_loss: 0.2950 | g_loss: 0.3183\n",
      "Epoch [    1/   20] | d_loss: 0.1215 | g_loss: 0.4917\n",
      "Epoch [    1/   20] | d_loss: 0.2987 | g_loss: 0.3680\n",
      "Epoch [    1/   20] | d_loss: 0.3746 | g_loss: 0.2980\n",
      "Epoch [    1/   20] | d_loss: 0.2471 | g_loss: 0.3359\n",
      "Epoch [    1/   20] | d_loss: 0.2152 | g_loss: 0.2610\n",
      "Epoch [    1/   20] | d_loss: 0.2182 | g_loss: 0.5617\n",
      "Epoch [    1/   20] | d_loss: 0.2611 | g_loss: 0.6616\n",
      "Epoch [    1/   20] | d_loss: 0.3253 | g_loss: 0.4335\n",
      "Epoch [    1/   20] | d_loss: 0.2032 | g_loss: 0.1248\n",
      "Epoch [    2/   20] | d_loss: 0.2238 | g_loss: 0.4030\n",
      "Epoch [    2/   20] | d_loss: 0.2689 | g_loss: 0.4021\n",
      "Epoch [    2/   20] | d_loss: 0.1473 | g_loss: 0.5126\n",
      "Epoch [    2/   20] | d_loss: 0.1892 | g_loss: 0.4067\n",
      "Epoch [    2/   20] | d_loss: 0.2859 | g_loss: 0.3738\n",
      "Epoch [    2/   20] | d_loss: 0.1894 | g_loss: 0.4494\n",
      "Epoch [    2/   20] | d_loss: 0.2785 | g_loss: 0.4251\n",
      "Epoch [    2/   20] | d_loss: 0.2882 | g_loss: 0.3134\n",
      "Epoch [    2/   20] | d_loss: 0.3200 | g_loss: 0.4878\n",
      "Epoch [    2/   20] | d_loss: 0.2388 | g_loss: 0.7577\n",
      "Epoch [    2/   20] | d_loss: 0.2223 | g_loss: 0.6472\n",
      "Epoch [    2/   20] | d_loss: 0.3346 | g_loss: 0.4058\n",
      "Epoch [    2/   20] | d_loss: 0.1634 | g_loss: 0.6115\n",
      "Epoch [    2/   20] | d_loss: 0.2260 | g_loss: 0.4416\n",
      "Epoch [    2/   20] | d_loss: 0.3428 | g_loss: 0.3941\n",
      "Epoch [    2/   20] | d_loss: 0.1857 | g_loss: 0.4656\n",
      "Epoch [    2/   20] | d_loss: 0.2214 | g_loss: 0.2811\n",
      "Epoch [    2/   20] | d_loss: 0.4126 | g_loss: 0.4108\n",
      "Epoch [    2/   20] | d_loss: 0.1854 | g_loss: 0.4883\n",
      "Epoch [    2/   20] | d_loss: 0.2116 | g_loss: 0.2924\n",
      "Epoch [    2/   20] | d_loss: 0.1592 | g_loss: 0.6016\n",
      "Epoch [    2/   20] | d_loss: 0.3067 | g_loss: 0.2312\n",
      "Epoch [    2/   20] | d_loss: 0.1952 | g_loss: 0.4582\n",
      "Epoch [    2/   20] | d_loss: 0.2930 | g_loss: 0.4329\n",
      "Epoch [    2/   20] | d_loss: 0.2388 | g_loss: 0.3734\n",
      "Epoch [    2/   20] | d_loss: 0.2312 | g_loss: 0.2385\n",
      "Epoch [    2/   20] | d_loss: 0.3127 | g_loss: 0.3287\n",
      "Epoch [    2/   20] | d_loss: 0.3539 | g_loss: 0.3389\n",
      "Epoch [    2/   20] | d_loss: 0.3108 | g_loss: 0.0835\n",
      "Epoch [    2/   20] | d_loss: 0.3052 | g_loss: 0.5659\n",
      "Epoch [    2/   20] | d_loss: 0.3038 | g_loss: 0.3176\n",
      "Epoch [    2/   20] | d_loss: 0.2460 | g_loss: 0.5629\n",
      "Epoch [    2/   20] | d_loss: 0.4203 | g_loss: 0.9074\n",
      "Epoch [    2/   20] | d_loss: 0.2449 | g_loss: 0.4505\n",
      "Epoch [    2/   20] | d_loss: 0.2824 | g_loss: 0.5563\n",
      "Epoch [    2/   20] | d_loss: 0.1130 | g_loss: 0.4775\n",
      "Epoch [    2/   20] | d_loss: 0.2080 | g_loss: 0.4781\n",
      "Epoch [    2/   20] | d_loss: 0.1787 | g_loss: 0.4022\n",
      "Epoch [    2/   20] | d_loss: 0.2092 | g_loss: 0.4080\n",
      "Epoch [    2/   20] | d_loss: 0.2513 | g_loss: 0.2056\n",
      "Epoch [    2/   20] | d_loss: 0.2578 | g_loss: 0.7976\n",
      "Epoch [    2/   20] | d_loss: 0.1360 | g_loss: 0.7410\n",
      "Epoch [    2/   20] | d_loss: 0.1991 | g_loss: 0.2946\n",
      "Epoch [    2/   20] | d_loss: 0.4246 | g_loss: 0.2216\n",
      "Epoch [    2/   20] | d_loss: 0.2187 | g_loss: 0.3596\n",
      "Epoch [    2/   20] | d_loss: 0.2467 | g_loss: 0.5038\n",
      "Epoch [    2/   20] | d_loss: 0.2072 | g_loss: 0.6510\n",
      "Epoch [    2/   20] | d_loss: 0.4226 | g_loss: 0.2876\n",
      "Epoch [    2/   20] | d_loss: 0.2603 | g_loss: 0.1728\n",
      "Epoch [    2/   20] | d_loss: 0.4138 | g_loss: 0.7257\n",
      "Epoch [    2/   20] | d_loss: 0.1925 | g_loss: 0.5011\n",
      "Epoch [    2/   20] | d_loss: 0.2208 | g_loss: 0.4644\n",
      "Epoch [    2/   20] | d_loss: 0.3400 | g_loss: 0.5898\n",
      "Epoch [    2/   20] | d_loss: 0.3613 | g_loss: 0.2693\n",
      "Epoch [    2/   20] | d_loss: 0.2042 | g_loss: 0.4762\n",
      "Epoch [    2/   20] | d_loss: 0.2359 | g_loss: 0.2221\n",
      "Epoch [    2/   20] | d_loss: 0.1590 | g_loss: 0.2902\n",
      "Epoch [    2/   20] | d_loss: 0.2528 | g_loss: 0.1972\n",
      "Epoch [    2/   20] | d_loss: 0.1822 | g_loss: 0.4094\n",
      "Epoch [    2/   20] | d_loss: 0.2796 | g_loss: 0.6304\n",
      "Epoch [    2/   20] | d_loss: 0.3476 | g_loss: 0.4663\n",
      "Epoch [    2/   20] | d_loss: 0.2654 | g_loss: 0.6522\n",
      "Epoch [    2/   20] | d_loss: 0.1426 | g_loss: 0.5357\n",
      "Epoch [    2/   20] | d_loss: 0.1817 | g_loss: 0.5273\n",
      "Epoch [    2/   20] | d_loss: 0.3139 | g_loss: 0.2599\n",
      "Epoch [    2/   20] | d_loss: 0.1629 | g_loss: 0.4860\n",
      "Epoch [    2/   20] | d_loss: 0.2356 | g_loss: 0.7852\n",
      "Epoch [    2/   20] | d_loss: 0.2155 | g_loss: 0.4339\n",
      "Epoch [    2/   20] | d_loss: 0.2051 | g_loss: 0.3752\n",
      "Epoch [    2/   20] | d_loss: 0.1167 | g_loss: 0.4686\n",
      "Epoch [    2/   20] | d_loss: 0.1666 | g_loss: 0.9150\n",
      "Epoch [    2/   20] | d_loss: 0.3741 | g_loss: 0.3642\n",
      "Epoch [    2/   20] | d_loss: 0.1882 | g_loss: 0.3376\n",
      "Epoch [    2/   20] | d_loss: 0.2208 | g_loss: 0.5289\n",
      "Epoch [    2/   20] | d_loss: 0.2706 | g_loss: 0.3077\n",
      "Epoch [    2/   20] | d_loss: 0.3270 | g_loss: 0.4382\n",
      "Epoch [    2/   20] | d_loss: 0.1683 | g_loss: 0.2782\n",
      "Epoch [    2/   20] | d_loss: 0.3052 | g_loss: 0.6319\n",
      "Epoch [    2/   20] | d_loss: 0.3005 | g_loss: 0.4323\n",
      "Epoch [    2/   20] | d_loss: 0.4268 | g_loss: 0.3677\n",
      "Epoch [    2/   20] | d_loss: 0.2067 | g_loss: 0.3114\n",
      "Epoch [    2/   20] | d_loss: 0.2729 | g_loss: 0.4739\n",
      "Epoch [    2/   20] | d_loss: 0.2658 | g_loss: 0.5516\n",
      "Epoch [    2/   20] | d_loss: 0.1569 | g_loss: 0.4859\n",
      "Epoch [    2/   20] | d_loss: 0.1381 | g_loss: 0.6347\n",
      "Epoch [    2/   20] | d_loss: 0.1900 | g_loss: 0.2630\n",
      "Epoch [    2/   20] | d_loss: 0.2784 | g_loss: 0.4085\n",
      "Epoch [    2/   20] | d_loss: 0.1521 | g_loss: 0.4774\n",
      "Epoch [    2/   20] | d_loss: 0.3423 | g_loss: 0.3856\n",
      "Epoch [    2/   20] | d_loss: 0.3388 | g_loss: 0.5774\n",
      "Epoch [    3/   20] | d_loss: 0.3227 | g_loss: 0.5859\n",
      "Epoch [    3/   20] | d_loss: 0.1579 | g_loss: 0.4061\n",
      "Epoch [    3/   20] | d_loss: 0.1649 | g_loss: 0.6018\n",
      "Epoch [    3/   20] | d_loss: 0.2453 | g_loss: 0.3244\n",
      "Epoch [    3/   20] | d_loss: 0.2254 | g_loss: 0.2116\n",
      "Epoch [    3/   20] | d_loss: 0.2084 | g_loss: 0.5118\n",
      "Epoch [    3/   20] | d_loss: 0.2963 | g_loss: 0.3246\n",
      "Epoch [    3/   20] | d_loss: 0.0852 | g_loss: 0.5354\n",
      "Epoch [    3/   20] | d_loss: 0.2085 | g_loss: 0.4312\n",
      "Epoch [    3/   20] | d_loss: 0.2693 | g_loss: 0.3378\n",
      "Epoch [    3/   20] | d_loss: 0.1762 | g_loss: 0.6213\n",
      "Epoch [    3/   20] | d_loss: 0.2622 | g_loss: 0.8617\n",
      "Epoch [    3/   20] | d_loss: 0.2472 | g_loss: 0.4649\n",
      "Epoch [    3/   20] | d_loss: 0.2697 | g_loss: 0.4778\n",
      "Epoch [    3/   20] | d_loss: 0.4793 | g_loss: 0.3357\n",
      "Epoch [    3/   20] | d_loss: 0.2285 | g_loss: 0.4628\n",
      "Epoch [    3/   20] | d_loss: 0.1473 | g_loss: 0.4537\n",
      "Epoch [    3/   20] | d_loss: 0.1692 | g_loss: 0.4676\n",
      "Epoch [    3/   20] | d_loss: 0.1561 | g_loss: 0.5414\n",
      "Epoch [    3/   20] | d_loss: 0.1560 | g_loss: 0.4608\n",
      "Epoch [    3/   20] | d_loss: 0.3845 | g_loss: 0.4267\n",
      "Epoch [    3/   20] | d_loss: 0.3983 | g_loss: 0.1704\n",
      "Epoch [    3/   20] | d_loss: 0.1435 | g_loss: 0.7246\n",
      "Epoch [    3/   20] | d_loss: 0.1842 | g_loss: 0.7717\n",
      "Epoch [    3/   20] | d_loss: 0.2567 | g_loss: 0.4426\n",
      "Epoch [    3/   20] | d_loss: 0.1385 | g_loss: 0.4329\n",
      "Epoch [    3/   20] | d_loss: 0.1473 | g_loss: 0.5563\n",
      "Epoch [    3/   20] | d_loss: 0.2522 | g_loss: 0.6330\n",
      "Epoch [    3/   20] | d_loss: 0.3246 | g_loss: 0.5485\n",
      "Epoch [    3/   20] | d_loss: 0.1418 | g_loss: 0.4196\n",
      "Epoch [    3/   20] | d_loss: 0.1821 | g_loss: 0.7096\n",
      "Epoch [    3/   20] | d_loss: 0.2274 | g_loss: 0.3679\n",
      "Epoch [    3/   20] | d_loss: 0.3419 | g_loss: 0.2898\n",
      "Epoch [    3/   20] | d_loss: 0.2586 | g_loss: 0.4811\n",
      "Epoch [    3/   20] | d_loss: 0.1681 | g_loss: 0.8167\n",
      "Epoch [    3/   20] | d_loss: 0.3788 | g_loss: 0.3636\n",
      "Epoch [    3/   20] | d_loss: 0.3607 | g_loss: 0.4339\n",
      "Epoch [    3/   20] | d_loss: 0.2748 | g_loss: 0.2965\n",
      "Epoch [    3/   20] | d_loss: 0.2588 | g_loss: 0.3764\n",
      "Epoch [    3/   20] | d_loss: 0.2829 | g_loss: 0.2786\n",
      "Epoch [    3/   20] | d_loss: 0.1772 | g_loss: 0.4500\n",
      "Epoch [    3/   20] | d_loss: 0.3515 | g_loss: 0.5236\n",
      "Epoch [    3/   20] | d_loss: 0.2454 | g_loss: 0.4287\n",
      "Epoch [    3/   20] | d_loss: 0.1338 | g_loss: 0.2036\n",
      "Epoch [    3/   20] | d_loss: 0.2010 | g_loss: 0.4830\n",
      "Epoch [    3/   20] | d_loss: 0.1914 | g_loss: 0.5285\n",
      "Epoch [    3/   20] | d_loss: 0.1844 | g_loss: 0.4618\n",
      "Epoch [    3/   20] | d_loss: 0.1560 | g_loss: 0.2408\n",
      "Epoch [    3/   20] | d_loss: 0.2120 | g_loss: 0.3913\n",
      "Epoch [    3/   20] | d_loss: 0.2574 | g_loss: 0.5852\n",
      "Epoch [    3/   20] | d_loss: 0.1885 | g_loss: 0.3538\n",
      "Epoch [    3/   20] | d_loss: 0.2265 | g_loss: 0.3787\n",
      "Epoch [    3/   20] | d_loss: 0.2356 | g_loss: 0.3406\n",
      "Epoch [    3/   20] | d_loss: 0.2425 | g_loss: 0.6925\n",
      "Epoch [    3/   20] | d_loss: 0.3089 | g_loss: 0.6378\n",
      "Epoch [    3/   20] | d_loss: 0.1377 | g_loss: 0.5451\n",
      "Epoch [    3/   20] | d_loss: 0.3326 | g_loss: 0.5119\n",
      "Epoch [    3/   20] | d_loss: 0.4024 | g_loss: 0.6130\n",
      "Epoch [    3/   20] | d_loss: 0.2084 | g_loss: 0.4039\n",
      "Epoch [    3/   20] | d_loss: 0.1259 | g_loss: 0.4498\n",
      "Epoch [    3/   20] | d_loss: 0.2470 | g_loss: 0.4012\n",
      "Epoch [    3/   20] | d_loss: 0.1473 | g_loss: 0.3416\n",
      "Epoch [    3/   20] | d_loss: 0.1477 | g_loss: 0.4087\n",
      "Epoch [    3/   20] | d_loss: 0.1039 | g_loss: 0.7792\n",
      "Epoch [    3/   20] | d_loss: 0.4519 | g_loss: 0.4490\n",
      "Epoch [    3/   20] | d_loss: 0.4251 | g_loss: 0.3018\n",
      "Epoch [    3/   20] | d_loss: 0.1528 | g_loss: 0.4839\n",
      "Epoch [    3/   20] | d_loss: 0.2498 | g_loss: 0.4877\n",
      "Epoch [    3/   20] | d_loss: 0.2030 | g_loss: 0.5754\n",
      "Epoch [    3/   20] | d_loss: 0.1944 | g_loss: 0.6462\n",
      "Epoch [    3/   20] | d_loss: 0.3120 | g_loss: 0.5418\n",
      "Epoch [    3/   20] | d_loss: 0.3071 | g_loss: 0.3408\n",
      "Epoch [    3/   20] | d_loss: 0.0970 | g_loss: 0.4627\n",
      "Epoch [    3/   20] | d_loss: 0.2190 | g_loss: 0.1832\n",
      "Epoch [    3/   20] | d_loss: 0.3466 | g_loss: 0.3738\n",
      "Epoch [    3/   20] | d_loss: 0.2766 | g_loss: 0.3912\n",
      "Epoch [    3/   20] | d_loss: 0.1441 | g_loss: 0.4595\n",
      "Epoch [    3/   20] | d_loss: 0.3133 | g_loss: 0.6706\n",
      "Epoch [    3/   20] | d_loss: 0.2161 | g_loss: 0.8301\n",
      "Epoch [    3/   20] | d_loss: 0.2935 | g_loss: 0.7453\n",
      "Epoch [    3/   20] | d_loss: 0.4277 | g_loss: 0.3697\n",
      "Epoch [    3/   20] | d_loss: 0.1803 | g_loss: 0.5591\n",
      "Epoch [    3/   20] | d_loss: 0.2453 | g_loss: 0.2448\n",
      "Epoch [    3/   20] | d_loss: 0.2617 | g_loss: 0.5888\n",
      "Epoch [    3/   20] | d_loss: 0.2754 | g_loss: 0.2472\n",
      "Epoch [    3/   20] | d_loss: 0.2934 | g_loss: 0.6428\n",
      "Epoch [    3/   20] | d_loss: 0.3667 | g_loss: 0.5118\n",
      "Epoch [    3/   20] | d_loss: 0.1717 | g_loss: 0.5186\n",
      "Epoch [    3/   20] | d_loss: 0.1565 | g_loss: 0.5562\n",
      "Epoch [    3/   20] | d_loss: 0.1889 | g_loss: 0.5644\n",
      "Epoch [    4/   20] | d_loss: 0.2137 | g_loss: 0.5586\n",
      "Epoch [    4/   20] | d_loss: 0.1311 | g_loss: 0.4624\n",
      "Epoch [    4/   20] | d_loss: 0.1860 | g_loss: 0.2963\n",
      "Epoch [    4/   20] | d_loss: 0.1211 | g_loss: 0.4316\n",
      "Epoch [    4/   20] | d_loss: 0.1370 | g_loss: 0.2524\n",
      "Epoch [    4/   20] | d_loss: 0.2225 | g_loss: 0.5888\n",
      "Epoch [    4/   20] | d_loss: 0.2465 | g_loss: 0.3264\n",
      "Epoch [    4/   20] | d_loss: 0.2913 | g_loss: 0.9324\n",
      "Epoch [    4/   20] | d_loss: 0.3178 | g_loss: 0.5522\n",
      "Epoch [    4/   20] | d_loss: 0.2415 | g_loss: 0.3925\n",
      "Epoch [    4/   20] | d_loss: 0.2365 | g_loss: 0.6584\n",
      "Epoch [    4/   20] | d_loss: 0.1594 | g_loss: 0.2478\n",
      "Epoch [    4/   20] | d_loss: 0.1447 | g_loss: 0.6613\n",
      "Epoch [    4/   20] | d_loss: 0.0838 | g_loss: 0.4398\n",
      "Epoch [    4/   20] | d_loss: 0.1964 | g_loss: 0.5366\n",
      "Epoch [    4/   20] | d_loss: 0.2382 | g_loss: 0.4031\n",
      "Epoch [    4/   20] | d_loss: 0.1650 | g_loss: 0.3962\n",
      "Epoch [    4/   20] | d_loss: 0.1930 | g_loss: 0.8025\n",
      "Epoch [    4/   20] | d_loss: 0.3171 | g_loss: 0.2614\n",
      "Epoch [    4/   20] | d_loss: 0.2294 | g_loss: 0.5612\n",
      "Epoch [    4/   20] | d_loss: 0.1423 | g_loss: 0.6330\n",
      "Epoch [    4/   20] | d_loss: 0.1925 | g_loss: 0.7853\n",
      "Epoch [    4/   20] | d_loss: 0.2782 | g_loss: 1.0913\n",
      "Epoch [    4/   20] | d_loss: 0.1440 | g_loss: 0.3922\n",
      "Epoch [    4/   20] | d_loss: 0.2681 | g_loss: 0.4029\n",
      "Epoch [    4/   20] | d_loss: 0.1957 | g_loss: 0.3744\n",
      "Epoch [    4/   20] | d_loss: 0.2814 | g_loss: 0.4909\n",
      "Epoch [    4/   20] | d_loss: 0.1632 | g_loss: 0.6244\n",
      "Epoch [    4/   20] | d_loss: 0.1312 | g_loss: 0.5902\n",
      "Epoch [    4/   20] | d_loss: 0.1756 | g_loss: 0.3667\n",
      "Epoch [    4/   20] | d_loss: 0.1564 | g_loss: 0.1729\n",
      "Epoch [    4/   20] | d_loss: 0.1594 | g_loss: 0.6339\n",
      "Epoch [    4/   20] | d_loss: 0.3030 | g_loss: 0.5181\n",
      "Epoch [    4/   20] | d_loss: 0.1778 | g_loss: 0.3444\n",
      "Epoch [    4/   20] | d_loss: 0.3606 | g_loss: 0.4738\n",
      "Epoch [    4/   20] | d_loss: 0.2728 | g_loss: 0.4361\n",
      "Epoch [    4/   20] | d_loss: 0.3576 | g_loss: 0.6783\n",
      "Epoch [    4/   20] | d_loss: 0.2760 | g_loss: 0.4988\n",
      "Epoch [    4/   20] | d_loss: 0.4756 | g_loss: 0.7976\n",
      "Epoch [    4/   20] | d_loss: 0.1888 | g_loss: 0.4267\n",
      "Epoch [    4/   20] | d_loss: 0.2090 | g_loss: 0.5607\n",
      "Epoch [    4/   20] | d_loss: 0.3082 | g_loss: 0.3525\n",
      "Epoch [    4/   20] | d_loss: 0.0917 | g_loss: 0.8526\n",
      "Epoch [    4/   20] | d_loss: 0.2869 | g_loss: 0.3250\n",
      "Epoch [    4/   20] | d_loss: 0.3172 | g_loss: 0.4933\n",
      "Epoch [    4/   20] | d_loss: 0.1758 | g_loss: 0.7058\n",
      "Epoch [    4/   20] | d_loss: 0.2264 | g_loss: 0.5420\n",
      "Epoch [    4/   20] | d_loss: 0.1467 | g_loss: 0.3318\n",
      "Epoch [    4/   20] | d_loss: 0.2441 | g_loss: 0.3961\n",
      "Epoch [    4/   20] | d_loss: 0.1755 | g_loss: 0.3252\n",
      "Epoch [    4/   20] | d_loss: 0.1412 | g_loss: 0.3696\n",
      "Epoch [    4/   20] | d_loss: 0.2616 | g_loss: 0.4793\n",
      "Epoch [    4/   20] | d_loss: 0.2034 | g_loss: 0.4833\n",
      "Epoch [    4/   20] | d_loss: 0.2460 | g_loss: 0.5766\n",
      "Epoch [    4/   20] | d_loss: 0.1973 | g_loss: 0.6258\n",
      "Epoch [    4/   20] | d_loss: 0.3049 | g_loss: 0.6227\n",
      "Epoch [    4/   20] | d_loss: 0.1899 | g_loss: 0.3500\n",
      "Epoch [    4/   20] | d_loss: 0.2201 | g_loss: 0.4409\n",
      "Epoch [    4/   20] | d_loss: 0.1951 | g_loss: 1.0443\n",
      "Epoch [    4/   20] | d_loss: 0.1916 | g_loss: 0.6099\n",
      "Epoch [    4/   20] | d_loss: 0.1168 | g_loss: 0.3117\n",
      "Epoch [    4/   20] | d_loss: 0.1342 | g_loss: 0.3572\n",
      "Epoch [    4/   20] | d_loss: 0.2598 | g_loss: 0.6248\n",
      "Epoch [    4/   20] | d_loss: 0.1803 | g_loss: 0.4869\n",
      "Epoch [    4/   20] | d_loss: 0.2093 | g_loss: 0.5659\n",
      "Epoch [    4/   20] | d_loss: 0.0808 | g_loss: 0.5263\n",
      "Epoch [    4/   20] | d_loss: 0.1970 | g_loss: 0.1838\n",
      "Epoch [    4/   20] | d_loss: 0.1416 | g_loss: 0.4027\n",
      "Epoch [    4/   20] | d_loss: 0.2182 | g_loss: 0.3301\n",
      "Epoch [    4/   20] | d_loss: 0.2172 | g_loss: 0.3638\n",
      "Epoch [    4/   20] | d_loss: 0.1119 | g_loss: 0.5803\n",
      "Epoch [    4/   20] | d_loss: 0.1394 | g_loss: 0.4860\n",
      "Epoch [    4/   20] | d_loss: 0.1842 | g_loss: 0.2964\n",
      "Epoch [    4/   20] | d_loss: 0.0976 | g_loss: 0.3681\n",
      "Epoch [    4/   20] | d_loss: 0.2075 | g_loss: 0.9774\n",
      "Epoch [    4/   20] | d_loss: 0.0725 | g_loss: 0.7446\n",
      "Epoch [    4/   20] | d_loss: 0.1489 | g_loss: 0.4188\n",
      "Epoch [    4/   20] | d_loss: 0.2192 | g_loss: 0.4456\n",
      "Epoch [    4/   20] | d_loss: 0.2284 | g_loss: 0.5737\n",
      "Epoch [    4/   20] | d_loss: 0.3361 | g_loss: 0.5570\n",
      "Epoch [    4/   20] | d_loss: 0.1539 | g_loss: 0.5081\n",
      "Epoch [    4/   20] | d_loss: 0.1426 | g_loss: 0.7473\n",
      "Epoch [    4/   20] | d_loss: 0.2935 | g_loss: 0.5852\n",
      "Epoch [    4/   20] | d_loss: 0.1375 | g_loss: 0.3551\n",
      "Epoch [    4/   20] | d_loss: 0.1128 | g_loss: 0.7194\n",
      "Epoch [    4/   20] | d_loss: 0.0974 | g_loss: 0.8630\n",
      "Epoch [    4/   20] | d_loss: 0.1985 | g_loss: 0.9921\n",
      "Epoch [    4/   20] | d_loss: 0.1317 | g_loss: 0.5534\n",
      "Epoch [    4/   20] | d_loss: 0.2147 | g_loss: 0.6146\n",
      "Epoch [    4/   20] | d_loss: 0.1378 | g_loss: 0.6004\n",
      "Epoch [    5/   20] | d_loss: 0.2189 | g_loss: 0.7244\n",
      "Epoch [    5/   20] | d_loss: 0.2410 | g_loss: 0.2793\n",
      "Epoch [    5/   20] | d_loss: 0.1836 | g_loss: 0.4922\n",
      "Epoch [    5/   20] | d_loss: 0.1690 | g_loss: 0.4516\n",
      "Epoch [    5/   20] | d_loss: 0.1033 | g_loss: 0.7920\n",
      "Epoch [    5/   20] | d_loss: 0.0895 | g_loss: 0.4991\n",
      "Epoch [    5/   20] | d_loss: 0.1525 | g_loss: 0.5577\n",
      "Epoch [    5/   20] | d_loss: 0.1572 | g_loss: 0.5128\n",
      "Epoch [    5/   20] | d_loss: 0.0887 | g_loss: 0.5323\n",
      "Epoch [    5/   20] | d_loss: 0.2777 | g_loss: 0.6208\n",
      "Epoch [    5/   20] | d_loss: 0.1127 | g_loss: 1.0578\n",
      "Epoch [    5/   20] | d_loss: 0.0890 | g_loss: 0.8415\n",
      "Epoch [    5/   20] | d_loss: 0.1431 | g_loss: 0.3161\n",
      "Epoch [    5/   20] | d_loss: 0.1490 | g_loss: 0.4686\n",
      "Epoch [    5/   20] | d_loss: 0.2011 | g_loss: 0.4195\n",
      "Epoch [    5/   20] | d_loss: 0.0521 | g_loss: 0.7069\n",
      "Epoch [    5/   20] | d_loss: 0.2648 | g_loss: 0.3348\n",
      "Epoch [    5/   20] | d_loss: 0.1624 | g_loss: 0.3321\n",
      "Epoch [    5/   20] | d_loss: 0.2948 | g_loss: 0.9317\n",
      "Epoch [    5/   20] | d_loss: 0.2231 | g_loss: 0.4105\n",
      "Epoch [    5/   20] | d_loss: 0.2788 | g_loss: 0.6057\n",
      "Epoch [    5/   20] | d_loss: 0.1111 | g_loss: 0.5931\n",
      "Epoch [    5/   20] | d_loss: 0.2674 | g_loss: 0.3344\n",
      "Epoch [    5/   20] | d_loss: 0.1237 | g_loss: 0.4515\n",
      "Epoch [    5/   20] | d_loss: 0.1337 | g_loss: 0.6889\n",
      "Epoch [    5/   20] | d_loss: 0.2086 | g_loss: 0.2864\n",
      "Epoch [    5/   20] | d_loss: 0.1414 | g_loss: 0.5207\n",
      "Epoch [    5/   20] | d_loss: 0.1220 | g_loss: 0.6146\n",
      "Epoch [    5/   20] | d_loss: 0.2134 | g_loss: 0.5338\n",
      "Epoch [    5/   20] | d_loss: 0.2497 | g_loss: 0.5113\n",
      "Epoch [    5/   20] | d_loss: 0.1193 | g_loss: 0.6088\n",
      "Epoch [    5/   20] | d_loss: 0.1262 | g_loss: 0.8524\n",
      "Epoch [    5/   20] | d_loss: 0.2246 | g_loss: 0.5516\n",
      "Epoch [    5/   20] | d_loss: 0.1665 | g_loss: 0.6005\n",
      "Epoch [    5/   20] | d_loss: 0.1551 | g_loss: 0.5817\n",
      "Epoch [    5/   20] | d_loss: 0.1067 | g_loss: 0.5192\n",
      "Epoch [    5/   20] | d_loss: 0.2021 | g_loss: 1.0623\n",
      "Epoch [    5/   20] | d_loss: 0.1822 | g_loss: 0.3244\n",
      "Epoch [    5/   20] | d_loss: 0.1940 | g_loss: 0.3963\n",
      "Epoch [    5/   20] | d_loss: 0.1363 | g_loss: 0.3530\n",
      "Epoch [    5/   20] | d_loss: 0.2123 | g_loss: 0.2735\n",
      "Epoch [    5/   20] | d_loss: 0.1130 | g_loss: 0.4594\n",
      "Epoch [    5/   20] | d_loss: 0.1835 | g_loss: 0.7783\n",
      "Epoch [    5/   20] | d_loss: 0.0857 | g_loss: 0.6576\n",
      "Epoch [    5/   20] | d_loss: 0.1933 | g_loss: 0.5461\n",
      "Epoch [    5/   20] | d_loss: 0.2984 | g_loss: 0.6699\n",
      "Epoch [    5/   20] | d_loss: 0.1973 | g_loss: 0.4498\n",
      "Epoch [    5/   20] | d_loss: 0.1880 | g_loss: 0.6314\n",
      "Epoch [    5/   20] | d_loss: 0.0802 | g_loss: 1.1287\n",
      "Epoch [    5/   20] | d_loss: 0.1127 | g_loss: 0.9667\n",
      "Epoch [    5/   20] | d_loss: 0.1718 | g_loss: 1.0803\n",
      "Epoch [    5/   20] | d_loss: 0.2381 | g_loss: 0.2259\n",
      "Epoch [    5/   20] | d_loss: 0.2117 | g_loss: 0.5955\n",
      "Epoch [    5/   20] | d_loss: 0.2398 | g_loss: 0.3703\n",
      "Epoch [    5/   20] | d_loss: 0.2057 | g_loss: 1.0615\n",
      "Epoch [    5/   20] | d_loss: 0.1640 | g_loss: 0.8634\n",
      "Epoch [    5/   20] | d_loss: 0.2064 | g_loss: 0.6276\n",
      "Epoch [    5/   20] | d_loss: 0.1684 | g_loss: 0.5265\n",
      "Epoch [    5/   20] | d_loss: 0.1400 | g_loss: 0.5038\n",
      "Epoch [    5/   20] | d_loss: 0.2000 | g_loss: 0.5247\n",
      "Epoch [    5/   20] | d_loss: 0.1401 | g_loss: 0.3312\n",
      "Epoch [    5/   20] | d_loss: 0.1214 | g_loss: 0.4692\n",
      "Epoch [    5/   20] | d_loss: 0.2420 | g_loss: 0.7212\n",
      "Epoch [    5/   20] | d_loss: 0.1795 | g_loss: 0.6689\n",
      "Epoch [    5/   20] | d_loss: 0.1468 | g_loss: 0.4239\n",
      "Epoch [    5/   20] | d_loss: 0.2527 | g_loss: 0.8563\n",
      "Epoch [    5/   20] | d_loss: 0.0873 | g_loss: 0.2942\n",
      "Epoch [    5/   20] | d_loss: 0.0917 | g_loss: 0.3677\n",
      "Epoch [    5/   20] | d_loss: 0.0709 | g_loss: 0.6029\n",
      "Epoch [    5/   20] | d_loss: 0.2171 | g_loss: 0.4128\n",
      "Epoch [    5/   20] | d_loss: 0.2449 | g_loss: 0.3528\n",
      "Epoch [    5/   20] | d_loss: 0.1477 | g_loss: 0.4358\n",
      "Epoch [    5/   20] | d_loss: 0.1836 | g_loss: 0.5929\n",
      "Epoch [    5/   20] | d_loss: 0.1670 | g_loss: 0.4760\n",
      "Epoch [    5/   20] | d_loss: 0.1104 | g_loss: 0.8363\n",
      "Epoch [    5/   20] | d_loss: 0.1031 | g_loss: 0.4192\n",
      "Epoch [    5/   20] | d_loss: 0.1486 | g_loss: 0.5612\n",
      "Epoch [    5/   20] | d_loss: 0.2680 | g_loss: 0.8245\n",
      "Epoch [    5/   20] | d_loss: 0.1521 | g_loss: 0.6140\n",
      "Epoch [    5/   20] | d_loss: 0.2001 | g_loss: 0.3220\n",
      "Epoch [    5/   20] | d_loss: 0.1632 | g_loss: 0.4425\n",
      "Epoch [    5/   20] | d_loss: 0.2426 | g_loss: 0.8539\n",
      "Epoch [    5/   20] | d_loss: 0.1947 | g_loss: 0.5821\n",
      "Epoch [    5/   20] | d_loss: 0.2483 | g_loss: 0.6631\n",
      "Epoch [    5/   20] | d_loss: 0.1342 | g_loss: 0.2555\n",
      "Epoch [    5/   20] | d_loss: 0.3583 | g_loss: 0.6989\n",
      "Epoch [    5/   20] | d_loss: 0.1142 | g_loss: 0.6073\n",
      "Epoch [    5/   20] | d_loss: 0.1348 | g_loss: 0.6672\n",
      "Epoch [    5/   20] | d_loss: 0.2151 | g_loss: 0.3868\n",
      "Epoch [    5/   20] | d_loss: 0.1958 | g_loss: 0.2922\n",
      "Epoch [    6/   20] | d_loss: 0.1719 | g_loss: 0.5336\n",
      "Epoch [    6/   20] | d_loss: 0.2080 | g_loss: 0.5141\n",
      "Epoch [    6/   20] | d_loss: 0.1045 | g_loss: 0.5576\n",
      "Epoch [    6/   20] | d_loss: 0.0736 | g_loss: 0.6705\n",
      "Epoch [    6/   20] | d_loss: 0.1747 | g_loss: 0.4678\n",
      "Epoch [    6/   20] | d_loss: 0.1691 | g_loss: 0.6773\n",
      "Epoch [    6/   20] | d_loss: 0.1368 | g_loss: 0.6559\n",
      "Epoch [    6/   20] | d_loss: 0.1681 | g_loss: 0.7726\n",
      "Epoch [    6/   20] | d_loss: 0.1183 | g_loss: 0.4506\n",
      "Epoch [    6/   20] | d_loss: 0.1653 | g_loss: 0.4186\n",
      "Epoch [    6/   20] | d_loss: 0.2683 | g_loss: 0.4964\n",
      "Epoch [    6/   20] | d_loss: 0.1354 | g_loss: 0.2685\n",
      "Epoch [    6/   20] | d_loss: 0.1977 | g_loss: 0.5439\n",
      "Epoch [    6/   20] | d_loss: 0.1948 | g_loss: 0.4200\n",
      "Epoch [    6/   20] | d_loss: 0.1423 | g_loss: 0.7028\n",
      "Epoch [    6/   20] | d_loss: 0.1169 | g_loss: 0.6110\n",
      "Epoch [    6/   20] | d_loss: 0.1828 | g_loss: 0.5800\n",
      "Epoch [    6/   20] | d_loss: 0.0952 | g_loss: 0.6191\n",
      "Epoch [    6/   20] | d_loss: 0.1988 | g_loss: 0.9778\n",
      "Epoch [    6/   20] | d_loss: 0.1906 | g_loss: 0.6241\n",
      "Epoch [    6/   20] | d_loss: 0.1680 | g_loss: 0.5138\n",
      "Epoch [    6/   20] | d_loss: 0.1037 | g_loss: 0.6142\n",
      "Epoch [    6/   20] | d_loss: 0.1869 | g_loss: 0.9688\n",
      "Epoch [    6/   20] | d_loss: 0.2067 | g_loss: 0.6725\n",
      "Epoch [    6/   20] | d_loss: 0.1143 | g_loss: 0.6946\n",
      "Epoch [    6/   20] | d_loss: 0.3083 | g_loss: 0.4072\n",
      "Epoch [    6/   20] | d_loss: 0.1805 | g_loss: 0.3826\n",
      "Epoch [    6/   20] | d_loss: 0.0911 | g_loss: 0.5833\n",
      "Epoch [    6/   20] | d_loss: 0.2135 | g_loss: 0.8875\n",
      "Epoch [    6/   20] | d_loss: 0.2366 | g_loss: 0.6359\n",
      "Epoch [    6/   20] | d_loss: 0.1778 | g_loss: 0.4748\n",
      "Epoch [    6/   20] | d_loss: 0.0949 | g_loss: 0.6246\n",
      "Epoch [    6/   20] | d_loss: 0.2112 | g_loss: 0.3953\n",
      "Epoch [    6/   20] | d_loss: 0.2856 | g_loss: 0.9147\n",
      "Epoch [    6/   20] | d_loss: 0.1172 | g_loss: 0.4376\n",
      "Epoch [    6/   20] | d_loss: 0.1545 | g_loss: 0.7101\n",
      "Epoch [    6/   20] | d_loss: 0.2492 | g_loss: 0.4141\n",
      "Epoch [    6/   20] | d_loss: 0.1824 | g_loss: 0.4561\n",
      "Epoch [    6/   20] | d_loss: 0.2263 | g_loss: 0.3428\n",
      "Epoch [    6/   20] | d_loss: 0.1714 | g_loss: 0.3287\n",
      "Epoch [    6/   20] | d_loss: 0.1849 | g_loss: 0.5457\n",
      "Epoch [    6/   20] | d_loss: 0.2474 | g_loss: 0.5715\n",
      "Epoch [    6/   20] | d_loss: 0.2563 | g_loss: 0.5447\n",
      "Epoch [    6/   20] | d_loss: 0.1417 | g_loss: 0.4635\n",
      "Epoch [    6/   20] | d_loss: 0.5300 | g_loss: 0.6574\n",
      "Epoch [    6/   20] | d_loss: 0.0560 | g_loss: 0.5926\n",
      "Epoch [    6/   20] | d_loss: 0.0966 | g_loss: 0.5822\n",
      "Epoch [    6/   20] | d_loss: 0.1208 | g_loss: 0.5540\n",
      "Epoch [    6/   20] | d_loss: 0.1555 | g_loss: 0.7089\n",
      "Epoch [    6/   20] | d_loss: 0.0932 | g_loss: 0.4885\n",
      "Epoch [    6/   20] | d_loss: 0.0939 | g_loss: 0.7146\n",
      "Epoch [    6/   20] | d_loss: 0.1918 | g_loss: 0.2937\n",
      "Epoch [    6/   20] | d_loss: 0.2126 | g_loss: 0.4939\n",
      "Epoch [    6/   20] | d_loss: 0.1516 | g_loss: 0.7835\n",
      "Epoch [    6/   20] | d_loss: 0.1442 | g_loss: 0.6442\n",
      "Epoch [    6/   20] | d_loss: 0.1980 | g_loss: 0.5623\n",
      "Epoch [    6/   20] | d_loss: 0.1252 | g_loss: 0.8376\n",
      "Epoch [    6/   20] | d_loss: 0.1679 | g_loss: 0.3832\n",
      "Epoch [    6/   20] | d_loss: 0.1423 | g_loss: 0.7479\n",
      "Epoch [    6/   20] | d_loss: 0.0858 | g_loss: 0.3852\n",
      "Epoch [    6/   20] | d_loss: 0.1239 | g_loss: 0.7395\n",
      "Epoch [    6/   20] | d_loss: 0.0932 | g_loss: 0.6176\n",
      "Epoch [    6/   20] | d_loss: 0.2246 | g_loss: 0.4276\n",
      "Epoch [    6/   20] | d_loss: 0.2121 | g_loss: 0.3420\n",
      "Epoch [    6/   20] | d_loss: 0.0939 | g_loss: 0.5239\n",
      "Epoch [    6/   20] | d_loss: 0.2025 | g_loss: 0.4010\n",
      "Epoch [    6/   20] | d_loss: 0.0931 | g_loss: 0.9255\n",
      "Epoch [    6/   20] | d_loss: 0.2203 | g_loss: 0.3475\n",
      "Epoch [    6/   20] | d_loss: 0.2337 | g_loss: 0.3642\n",
      "Epoch [    6/   20] | d_loss: 0.1526 | g_loss: 0.5078\n",
      "Epoch [    6/   20] | d_loss: 0.1080 | g_loss: 0.4723\n",
      "Epoch [    6/   20] | d_loss: 0.2032 | g_loss: 0.4364\n",
      "Epoch [    6/   20] | d_loss: 0.1773 | g_loss: 0.5640\n",
      "Epoch [    6/   20] | d_loss: 0.2024 | g_loss: 0.5408\n",
      "Epoch [    6/   20] | d_loss: 0.3500 | g_loss: 0.2603\n",
      "Epoch [    6/   20] | d_loss: 0.1576 | g_loss: 0.4730\n",
      "Epoch [    6/   20] | d_loss: 0.2290 | g_loss: 0.3947\n",
      "Epoch [    6/   20] | d_loss: 0.1393 | g_loss: 0.8137\n",
      "Epoch [    6/   20] | d_loss: 0.1417 | g_loss: 0.3653\n",
      "Epoch [    6/   20] | d_loss: 0.0917 | g_loss: 0.4302\n",
      "Epoch [    6/   20] | d_loss: 0.1275 | g_loss: 0.4964\n",
      "Epoch [    6/   20] | d_loss: 0.2776 | g_loss: 0.4275\n",
      "Epoch [    6/   20] | d_loss: 0.1485 | g_loss: 0.6961\n",
      "Epoch [    6/   20] | d_loss: 0.1414 | g_loss: 0.3253\n",
      "Epoch [    6/   20] | d_loss: 0.1262 | g_loss: 0.3034\n",
      "Epoch [    6/   20] | d_loss: 0.1246 | g_loss: 0.4274\n",
      "Epoch [    6/   20] | d_loss: 0.3753 | g_loss: 0.4414\n",
      "Epoch [    6/   20] | d_loss: 0.1925 | g_loss: 0.4248\n",
      "Epoch [    6/   20] | d_loss: 0.2166 | g_loss: 0.4802\n",
      "Epoch [    6/   20] | d_loss: 0.1514 | g_loss: 0.3956\n",
      "Epoch [    7/   20] | d_loss: 0.1139 | g_loss: 0.4831\n",
      "Epoch [    7/   20] | d_loss: 0.0961 | g_loss: 0.6801\n",
      "Epoch [    7/   20] | d_loss: 0.2414 | g_loss: 1.0096\n",
      "Epoch [    7/   20] | d_loss: 0.1612 | g_loss: 0.6014\n",
      "Epoch [    7/   20] | d_loss: 0.2324 | g_loss: 0.2516\n",
      "Epoch [    7/   20] | d_loss: 0.2071 | g_loss: 0.4734\n",
      "Epoch [    7/   20] | d_loss: 0.2703 | g_loss: 0.3370\n",
      "Epoch [    7/   20] | d_loss: 0.2588 | g_loss: 0.5075\n",
      "Epoch [    7/   20] | d_loss: 0.1108 | g_loss: 0.5877\n",
      "Epoch [    7/   20] | d_loss: 0.1736 | g_loss: 0.4731\n",
      "Epoch [    7/   20] | d_loss: 0.1619 | g_loss: 0.4299\n",
      "Epoch [    7/   20] | d_loss: 0.1539 | g_loss: 0.4599\n",
      "Epoch [    7/   20] | d_loss: 0.1420 | g_loss: 0.4775\n",
      "Epoch [    7/   20] | d_loss: 0.2286 | g_loss: 0.4147\n",
      "Epoch [    7/   20] | d_loss: 0.1162 | g_loss: 0.3370\n",
      "Epoch [    7/   20] | d_loss: 0.1657 | g_loss: 0.1809\n",
      "Epoch [    7/   20] | d_loss: 0.3210 | g_loss: 0.9103\n",
      "Epoch [    7/   20] | d_loss: 0.2154 | g_loss: 0.9899\n",
      "Epoch [    7/   20] | d_loss: 0.2838 | g_loss: 0.2284\n",
      "Epoch [    7/   20] | d_loss: 0.1056 | g_loss: 0.5205\n",
      "Epoch [    7/   20] | d_loss: 0.1219 | g_loss: 0.6049\n",
      "Epoch [    7/   20] | d_loss: 0.2063 | g_loss: 0.5702\n",
      "Epoch [    7/   20] | d_loss: 0.0930 | g_loss: 0.6224\n",
      "Epoch [    7/   20] | d_loss: 0.1168 | g_loss: 0.5857\n",
      "Epoch [    7/   20] | d_loss: 0.2310 | g_loss: 0.5650\n",
      "Epoch [    7/   20] | d_loss: 0.2404 | g_loss: 0.4473\n",
      "Epoch [    7/   20] | d_loss: 0.3391 | g_loss: 0.3027\n",
      "Epoch [    7/   20] | d_loss: 0.0991 | g_loss: 0.7649\n",
      "Epoch [    7/   20] | d_loss: 0.1787 | g_loss: 0.2588\n",
      "Epoch [    7/   20] | d_loss: 0.1274 | g_loss: 0.6243\n",
      "Epoch [    7/   20] | d_loss: 0.0946 | g_loss: 0.6544\n",
      "Epoch [    7/   20] | d_loss: 0.2827 | g_loss: 0.3764\n",
      "Epoch [    7/   20] | d_loss: 0.3246 | g_loss: 0.7195\n",
      "Epoch [    7/   20] | d_loss: 0.1465 | g_loss: 0.5529\n",
      "Epoch [    7/   20] | d_loss: 0.0977 | g_loss: 0.7693\n",
      "Epoch [    7/   20] | d_loss: 0.1778 | g_loss: 0.4948\n",
      "Epoch [    7/   20] | d_loss: 0.3681 | g_loss: 0.3515\n",
      "Epoch [    7/   20] | d_loss: 0.0973 | g_loss: 0.3144\n",
      "Epoch [    7/   20] | d_loss: 0.0696 | g_loss: 0.7325\n",
      "Epoch [    7/   20] | d_loss: 0.1222 | g_loss: 0.6210\n",
      "Epoch [    7/   20] | d_loss: 0.1171 | g_loss: 0.3777\n",
      "Epoch [    7/   20] | d_loss: 0.1191 | g_loss: 0.4092\n",
      "Epoch [    7/   20] | d_loss: 0.1664 | g_loss: 0.8147\n",
      "Epoch [    7/   20] | d_loss: 0.0822 | g_loss: 0.6732\n",
      "Epoch [    7/   20] | d_loss: 0.1021 | g_loss: 0.5243\n",
      "Epoch [    7/   20] | d_loss: 0.2509 | g_loss: 0.1672\n",
      "Epoch [    7/   20] | d_loss: 0.2306 | g_loss: 0.5042\n",
      "Epoch [    7/   20] | d_loss: 0.1008 | g_loss: 0.3163\n",
      "Epoch [    7/   20] | d_loss: 0.1680 | g_loss: 0.2697\n",
      "Epoch [    7/   20] | d_loss: 0.3466 | g_loss: 0.6417\n",
      "Epoch [    7/   20] | d_loss: 0.1010 | g_loss: 0.6129\n",
      "Epoch [    7/   20] | d_loss: 0.1419 | g_loss: 0.2781\n",
      "Epoch [    7/   20] | d_loss: 0.2525 | g_loss: 0.4943\n",
      "Epoch [    7/   20] | d_loss: 0.1818 | g_loss: 0.5523\n",
      "Epoch [    7/   20] | d_loss: 0.1524 | g_loss: 0.5943\n",
      "Epoch [    7/   20] | d_loss: 0.0906 | g_loss: 0.5311\n",
      "Epoch [    7/   20] | d_loss: 0.1722 | g_loss: 0.5491\n",
      "Epoch [    7/   20] | d_loss: 0.1143 | g_loss: 0.2853\n",
      "Epoch [    7/   20] | d_loss: 0.1303 | g_loss: 0.9560\n",
      "Epoch [    7/   20] | d_loss: 0.1380 | g_loss: 0.5577\n",
      "Epoch [    7/   20] | d_loss: 0.0863 | g_loss: 0.6349\n",
      "Epoch [    7/   20] | d_loss: 0.2038 | g_loss: 0.7184\n",
      "Epoch [    7/   20] | d_loss: 0.1978 | g_loss: 0.7398\n",
      "Epoch [    7/   20] | d_loss: 0.1881 | g_loss: 0.8318\n",
      "Epoch [    7/   20] | d_loss: 0.1558 | g_loss: 0.4807\n",
      "Epoch [    7/   20] | d_loss: 0.1534 | g_loss: 0.7965\n",
      "Epoch [    7/   20] | d_loss: 0.0867 | g_loss: 0.6499\n",
      "Epoch [    7/   20] | d_loss: 0.1245 | g_loss: 0.8982\n",
      "Epoch [    7/   20] | d_loss: 0.1419 | g_loss: 0.3924\n",
      "Epoch [    7/   20] | d_loss: 0.2671 | g_loss: 0.5421\n",
      "Epoch [    7/   20] | d_loss: 0.1193 | g_loss: 0.7028\n",
      "Epoch [    7/   20] | d_loss: 0.2584 | g_loss: 0.5083\n",
      "Epoch [    7/   20] | d_loss: 0.0795 | g_loss: 0.4222\n",
      "Epoch [    7/   20] | d_loss: 0.1654 | g_loss: 0.4439\n",
      "Epoch [    7/   20] | d_loss: 0.2357 | g_loss: 0.4899\n",
      "Epoch [    7/   20] | d_loss: 0.1432 | g_loss: 0.5765\n",
      "Epoch [    7/   20] | d_loss: 0.1537 | g_loss: 0.8748\n",
      "Epoch [    7/   20] | d_loss: 0.1635 | g_loss: 0.5430\n",
      "Epoch [    7/   20] | d_loss: 0.1166 | g_loss: 0.4957\n",
      "Epoch [    7/   20] | d_loss: 0.1528 | g_loss: 0.8244\n",
      "Epoch [    7/   20] | d_loss: 0.1555 | g_loss: 0.6436\n",
      "Epoch [    7/   20] | d_loss: 0.1039 | g_loss: 0.8712\n",
      "Epoch [    7/   20] | d_loss: 0.3202 | g_loss: 0.5599\n",
      "Epoch [    7/   20] | d_loss: 0.0923 | g_loss: 0.5254\n",
      "Epoch [    7/   20] | d_loss: 0.1651 | g_loss: 0.7000\n",
      "Epoch [    7/   20] | d_loss: 0.1717 | g_loss: 0.7865\n",
      "Epoch [    7/   20] | d_loss: 0.1011 | g_loss: 0.7223\n",
      "Epoch [    7/   20] | d_loss: 0.1255 | g_loss: 0.5167\n",
      "Epoch [    7/   20] | d_loss: 0.1906 | g_loss: 0.4848\n",
      "Epoch [    7/   20] | d_loss: 0.0791 | g_loss: 0.6097\n",
      "Epoch [    8/   20] | d_loss: 0.0881 | g_loss: 1.0202\n",
      "Epoch [    8/   20] | d_loss: 0.0878 | g_loss: 0.4724\n",
      "Epoch [    8/   20] | d_loss: 0.1329 | g_loss: 0.6940\n",
      "Epoch [    8/   20] | d_loss: 0.1212 | g_loss: 0.4504\n",
      "Epoch [    8/   20] | d_loss: 0.1322 | g_loss: 0.4200\n",
      "Epoch [    8/   20] | d_loss: 0.1465 | g_loss: 0.6660\n",
      "Epoch [    8/   20] | d_loss: 0.1305 | g_loss: 0.3755\n",
      "Epoch [    8/   20] | d_loss: 0.2071 | g_loss: 0.6437\n",
      "Epoch [    8/   20] | d_loss: 0.1244 | g_loss: 0.6384\n",
      "Epoch [    8/   20] | d_loss: 0.0737 | g_loss: 0.8474\n",
      "Epoch [    8/   20] | d_loss: 0.0944 | g_loss: 0.6932\n",
      "Epoch [    8/   20] | d_loss: 0.0792 | g_loss: 0.4741\n",
      "Epoch [    8/   20] | d_loss: 0.1652 | g_loss: 0.5297\n",
      "Epoch [    8/   20] | d_loss: 0.1964 | g_loss: 0.7194\n",
      "Epoch [    8/   20] | d_loss: 0.0963 | g_loss: 1.0155\n",
      "Epoch [    8/   20] | d_loss: 0.2358 | g_loss: 0.6674\n",
      "Epoch [    8/   20] | d_loss: 0.1727 | g_loss: 0.4525\n",
      "Epoch [    8/   20] | d_loss: 0.1510 | g_loss: 0.6010\n",
      "Epoch [    8/   20] | d_loss: 0.3314 | g_loss: 0.3547\n",
      "Epoch [    8/   20] | d_loss: 0.0913 | g_loss: 0.3597\n",
      "Epoch [    8/   20] | d_loss: 0.1051 | g_loss: 0.4719\n",
      "Epoch [    8/   20] | d_loss: 0.2812 | g_loss: 0.9494\n",
      "Epoch [    8/   20] | d_loss: 0.1316 | g_loss: 0.9826\n",
      "Epoch [    8/   20] | d_loss: 0.0763 | g_loss: 0.4413\n",
      "Epoch [    8/   20] | d_loss: 0.1675 | g_loss: 0.7648\n",
      "Epoch [    8/   20] | d_loss: 0.0798 | g_loss: 0.6048\n",
      "Epoch [    8/   20] | d_loss: 0.1145 | g_loss: 0.8206\n",
      "Epoch [    8/   20] | d_loss: 0.2208 | g_loss: 0.4377\n",
      "Epoch [    8/   20] | d_loss: 0.3663 | g_loss: 0.3314\n",
      "Epoch [    8/   20] | d_loss: 0.2031 | g_loss: 0.6146\n",
      "Epoch [    8/   20] | d_loss: 0.1762 | g_loss: 0.6014\n",
      "Epoch [    8/   20] | d_loss: 0.3181 | g_loss: 0.3222\n",
      "Epoch [    8/   20] | d_loss: 0.1452 | g_loss: 0.6265\n",
      "Epoch [    8/   20] | d_loss: 0.1626 | g_loss: 0.5001\n",
      "Epoch [    8/   20] | d_loss: 0.1586 | g_loss: 0.6431\n",
      "Epoch [    8/   20] | d_loss: 0.1253 | g_loss: 0.7075\n",
      "Epoch [    8/   20] | d_loss: 0.2093 | g_loss: 0.5956\n",
      "Epoch [    8/   20] | d_loss: 0.1035 | g_loss: 0.4034\n",
      "Epoch [    8/   20] | d_loss: 0.2452 | g_loss: 0.2913\n",
      "Epoch [    8/   20] | d_loss: 0.1123 | g_loss: 0.5194\n",
      "Epoch [    8/   20] | d_loss: 0.1304 | g_loss: 0.6473\n",
      "Epoch [    8/   20] | d_loss: 0.1470 | g_loss: 0.5815\n",
      "Epoch [    8/   20] | d_loss: 0.1459 | g_loss: 0.9307\n",
      "Epoch [    8/   20] | d_loss: 0.1222 | g_loss: 0.5024\n",
      "Epoch [    8/   20] | d_loss: 0.1608 | g_loss: 0.4167\n",
      "Epoch [    8/   20] | d_loss: 0.2010 | g_loss: 0.7037\n",
      "Epoch [    8/   20] | d_loss: 0.1281 | g_loss: 0.8252\n",
      "Epoch [    8/   20] | d_loss: 0.1767 | g_loss: 0.6521\n",
      "Epoch [    8/   20] | d_loss: 0.3328 | g_loss: 0.2815\n",
      "Epoch [    8/   20] | d_loss: 0.0874 | g_loss: 0.6971\n",
      "Epoch [    8/   20] | d_loss: 0.0793 | g_loss: 0.4245\n",
      "Epoch [    8/   20] | d_loss: 0.1648 | g_loss: 0.4172\n",
      "Epoch [    8/   20] | d_loss: 0.1505 | g_loss: 0.6111\n",
      "Epoch [    8/   20] | d_loss: 0.2057 | g_loss: 0.4376\n",
      "Epoch [    8/   20] | d_loss: 0.1428 | g_loss: 0.4073\n",
      "Epoch [    8/   20] | d_loss: 0.1146 | g_loss: 0.7887\n",
      "Epoch [    8/   20] | d_loss: 0.1693 | g_loss: 0.3322\n",
      "Epoch [    8/   20] | d_loss: 0.1545 | g_loss: 0.7990\n",
      "Epoch [    8/   20] | d_loss: 0.1290 | g_loss: 0.4703\n",
      "Epoch [    8/   20] | d_loss: 0.2561 | g_loss: 0.7100\n",
      "Epoch [    8/   20] | d_loss: 0.2553 | g_loss: 0.4015\n",
      "Epoch [    8/   20] | d_loss: 0.1166 | g_loss: 0.8408\n",
      "Epoch [    8/   20] | d_loss: 0.2197 | g_loss: 0.6137\n",
      "Epoch [    8/   20] | d_loss: 0.1234 | g_loss: 0.3821\n",
      "Epoch [    8/   20] | d_loss: 0.0961 | g_loss: 0.6622\n",
      "Epoch [    8/   20] | d_loss: 0.1553 | g_loss: 0.6498\n",
      "Epoch [    8/   20] | d_loss: 0.2923 | g_loss: 0.9731\n",
      "Epoch [    8/   20] | d_loss: 0.1534 | g_loss: 0.4876\n",
      "Epoch [    8/   20] | d_loss: 0.1532 | g_loss: 0.5124\n",
      "Epoch [    8/   20] | d_loss: 0.1421 | g_loss: 0.7906\n",
      "Epoch [    8/   20] | d_loss: 0.1335 | g_loss: 0.5105\n",
      "Epoch [    8/   20] | d_loss: 0.1346 | g_loss: 0.8381\n",
      "Epoch [    8/   20] | d_loss: 0.1179 | g_loss: 0.5572\n",
      "Epoch [    8/   20] | d_loss: 0.1418 | g_loss: 0.2623\n",
      "Epoch [    8/   20] | d_loss: 0.2203 | g_loss: 0.8829\n",
      "Epoch [    8/   20] | d_loss: 0.1569 | g_loss: 0.3023\n",
      "Epoch [    8/   20] | d_loss: 0.2659 | g_loss: 0.4448\n",
      "Epoch [    8/   20] | d_loss: 0.1156 | g_loss: 0.7597\n",
      "Epoch [    8/   20] | d_loss: 0.1178 | g_loss: 0.3246\n",
      "Epoch [    8/   20] | d_loss: 0.0753 | g_loss: 0.8125\n",
      "Epoch [    8/   20] | d_loss: 0.1768 | g_loss: 0.8335\n",
      "Epoch [    8/   20] | d_loss: 0.1758 | g_loss: 0.3168\n",
      "Epoch [    8/   20] | d_loss: 0.1635 | g_loss: 0.5098\n",
      "Epoch [    8/   20] | d_loss: 0.0849 | g_loss: 0.6821\n",
      "Epoch [    8/   20] | d_loss: 0.1853 | g_loss: 0.5277\n",
      "Epoch [    8/   20] | d_loss: 0.2162 | g_loss: 0.3961\n",
      "Epoch [    8/   20] | d_loss: 0.1741 | g_loss: 0.2654\n",
      "Epoch [    8/   20] | d_loss: 0.1737 | g_loss: 0.1883\n",
      "Epoch [    8/   20] | d_loss: 0.1854 | g_loss: 0.6025\n",
      "Epoch [    8/   20] | d_loss: 0.1456 | g_loss: 0.7239\n",
      "Epoch [    9/   20] | d_loss: 0.1108 | g_loss: 0.4188\n",
      "Epoch [    9/   20] | d_loss: 0.0866 | g_loss: 0.4410\n",
      "Epoch [    9/   20] | d_loss: 0.0868 | g_loss: 0.8065\n",
      "Epoch [    9/   20] | d_loss: 0.2929 | g_loss: 0.7441\n",
      "Epoch [    9/   20] | d_loss: 0.1005 | g_loss: 0.4774\n",
      "Epoch [    9/   20] | d_loss: 0.2248 | g_loss: 0.6080\n",
      "Epoch [    9/   20] | d_loss: 0.0711 | g_loss: 0.4455\n",
      "Epoch [    9/   20] | d_loss: 0.1645 | g_loss: 0.2778\n",
      "Epoch [    9/   20] | d_loss: 0.1778 | g_loss: 0.5579\n",
      "Epoch [    9/   20] | d_loss: 0.2032 | g_loss: 0.1716\n",
      "Epoch [    9/   20] | d_loss: 0.2174 | g_loss: 0.4186\n",
      "Epoch [    9/   20] | d_loss: 0.1645 | g_loss: 0.7121\n",
      "Epoch [    9/   20] | d_loss: 0.1060 | g_loss: 0.7499\n",
      "Epoch [    9/   20] | d_loss: 0.1521 | g_loss: 0.7219\n",
      "Epoch [    9/   20] | d_loss: 0.1199 | g_loss: 0.6682\n",
      "Epoch [    9/   20] | d_loss: 0.1510 | g_loss: 0.5868\n",
      "Epoch [    9/   20] | d_loss: 0.1780 | g_loss: 0.7071\n",
      "Epoch [    9/   20] | d_loss: 0.0949 | g_loss: 0.9546\n",
      "Epoch [    9/   20] | d_loss: 0.1441 | g_loss: 0.4843\n",
      "Epoch [    9/   20] | d_loss: 0.1811 | g_loss: 0.7670\n",
      "Epoch [    9/   20] | d_loss: 0.1315 | g_loss: 0.6273\n",
      "Epoch [    9/   20] | d_loss: 0.2029 | g_loss: 0.6494\n",
      "Epoch [    9/   20] | d_loss: 0.0643 | g_loss: 0.4444\n",
      "Epoch [    9/   20] | d_loss: 0.0910 | g_loss: 0.6207\n",
      "Epoch [    9/   20] | d_loss: 0.1228 | g_loss: 1.1521\n",
      "Epoch [    9/   20] | d_loss: 0.1313 | g_loss: 0.3963\n",
      "Epoch [    9/   20] | d_loss: 0.1191 | g_loss: 0.6542\n",
      "Epoch [    9/   20] | d_loss: 0.2965 | g_loss: 0.5306\n",
      "Epoch [    9/   20] | d_loss: 0.1888 | g_loss: 0.4892\n",
      "Epoch [    9/   20] | d_loss: 0.0718 | g_loss: 0.7493\n",
      "Epoch [    9/   20] | d_loss: 0.1027 | g_loss: 0.4831\n",
      "Epoch [    9/   20] | d_loss: 0.1109 | g_loss: 0.7830\n",
      "Epoch [    9/   20] | d_loss: 0.1110 | g_loss: 0.5414\n",
      "Epoch [    9/   20] | d_loss: 0.0916 | g_loss: 0.6129\n",
      "Epoch [    9/   20] | d_loss: 0.1752 | g_loss: 0.5270\n",
      "Epoch [    9/   20] | d_loss: 0.1522 | g_loss: 0.5666\n",
      "Epoch [    9/   20] | d_loss: 0.2317 | g_loss: 0.4942\n",
      "Epoch [    9/   20] | d_loss: 0.1610 | g_loss: 0.4586\n",
      "Epoch [    9/   20] | d_loss: 0.1861 | g_loss: 0.3588\n",
      "Epoch [    9/   20] | d_loss: 0.0955 | g_loss: 0.5601\n",
      "Epoch [    9/   20] | d_loss: 0.3741 | g_loss: 0.4034\n",
      "Epoch [    9/   20] | d_loss: 0.1661 | g_loss: 0.4862\n",
      "Epoch [    9/   20] | d_loss: 0.1465 | g_loss: 0.6713\n",
      "Epoch [    9/   20] | d_loss: 0.1851 | g_loss: 0.4650\n",
      "Epoch [    9/   20] | d_loss: 0.1163 | g_loss: 0.9798\n",
      "Epoch [    9/   20] | d_loss: 0.1656 | g_loss: 0.9495\n",
      "Epoch [    9/   20] | d_loss: 0.0797 | g_loss: 0.7411\n",
      "Epoch [    9/   20] | d_loss: 0.2190 | g_loss: 0.6243\n",
      "Epoch [    9/   20] | d_loss: 0.2050 | g_loss: 0.9136\n",
      "Epoch [    9/   20] | d_loss: 0.0751 | g_loss: 0.9552\n",
      "Epoch [    9/   20] | d_loss: 0.1364 | g_loss: 0.7632\n",
      "Epoch [    9/   20] | d_loss: 0.1325 | g_loss: 0.4143\n",
      "Epoch [    9/   20] | d_loss: 0.1065 | g_loss: 0.4563\n",
      "Epoch [    9/   20] | d_loss: 0.1385 | g_loss: 0.8512\n",
      "Epoch [    9/   20] | d_loss: 0.1445 | g_loss: 0.8958\n",
      "Epoch [    9/   20] | d_loss: 0.1977 | g_loss: 0.9610\n",
      "Epoch [    9/   20] | d_loss: 0.0504 | g_loss: 0.6756\n",
      "Epoch [    9/   20] | d_loss: 0.1630 | g_loss: 0.3988\n",
      "Epoch [    9/   20] | d_loss: 0.1060 | g_loss: 0.6733\n",
      "Epoch [    9/   20] | d_loss: 0.1256 | g_loss: 0.6420\n",
      "Epoch [    9/   20] | d_loss: 0.2470 | g_loss: 0.6126\n",
      "Epoch [    9/   20] | d_loss: 0.1624 | g_loss: 0.4962\n",
      "Epoch [    9/   20] | d_loss: 0.0813 | g_loss: 0.8334\n",
      "Epoch [    9/   20] | d_loss: 0.2089 | g_loss: 0.8349\n",
      "Epoch [    9/   20] | d_loss: 0.1014 | g_loss: 0.7711\n",
      "Epoch [    9/   20] | d_loss: 0.2025 | g_loss: 0.3710\n",
      "Epoch [    9/   20] | d_loss: 0.1722 | g_loss: 0.3302\n",
      "Epoch [    9/   20] | d_loss: 0.1481 | g_loss: 0.7957\n",
      "Epoch [    9/   20] | d_loss: 0.0988 | g_loss: 0.3466\n",
      "Epoch [    9/   20] | d_loss: 0.1257 | g_loss: 0.7384\n",
      "Epoch [    9/   20] | d_loss: 0.3032 | g_loss: 0.4686\n",
      "Epoch [    9/   20] | d_loss: 0.1448 | g_loss: 0.8147\n",
      "Epoch [    9/   20] | d_loss: 0.1120 | g_loss: 0.4843\n",
      "Epoch [    9/   20] | d_loss: 0.1685 | g_loss: 0.6907\n",
      "Epoch [    9/   20] | d_loss: 0.0540 | g_loss: 0.8110\n",
      "Epoch [    9/   20] | d_loss: 0.0974 | g_loss: 0.8960\n",
      "Epoch [    9/   20] | d_loss: 0.2197 | g_loss: 0.8218\n",
      "Epoch [    9/   20] | d_loss: 0.2626 | g_loss: 0.6728\n",
      "Epoch [    9/   20] | d_loss: 0.1266 | g_loss: 0.9024\n",
      "Epoch [    9/   20] | d_loss: 0.0863 | g_loss: 0.5594\n",
      "Epoch [    9/   20] | d_loss: 0.1796 | g_loss: 0.6032\n",
      "Epoch [    9/   20] | d_loss: 0.0878 | g_loss: 0.5268\n",
      "Epoch [    9/   20] | d_loss: 0.1698 | g_loss: 0.6315\n",
      "Epoch [    9/   20] | d_loss: 0.1112 | g_loss: 0.5933\n",
      "Epoch [    9/   20] | d_loss: 0.1116 | g_loss: 0.2918\n",
      "Epoch [    9/   20] | d_loss: 0.1493 | g_loss: 0.5545\n",
      "Epoch [    9/   20] | d_loss: 0.0649 | g_loss: 0.6698\n",
      "Epoch [    9/   20] | d_loss: 0.1200 | g_loss: 0.7504\n",
      "Epoch [    9/   20] | d_loss: 0.0771 | g_loss: 0.3267\n",
      "Epoch [    9/   20] | d_loss: 0.4160 | g_loss: 0.7331\n",
      "Epoch [   10/   20] | d_loss: 0.1191 | g_loss: 0.7123\n",
      "Epoch [   10/   20] | d_loss: 0.0951 | g_loss: 0.8770\n",
      "Epoch [   10/   20] | d_loss: 0.1823 | g_loss: 0.4195\n",
      "Epoch [   10/   20] | d_loss: 0.1291 | g_loss: 0.6426\n",
      "Epoch [   10/   20] | d_loss: 0.2651 | g_loss: 0.4353\n",
      "Epoch [   10/   20] | d_loss: 0.1997 | g_loss: 0.8070\n",
      "Epoch [   10/   20] | d_loss: 0.1303 | g_loss: 0.7581\n",
      "Epoch [   10/   20] | d_loss: 0.1248 | g_loss: 0.2799\n",
      "Epoch [   10/   20] | d_loss: 0.1473 | g_loss: 0.7041\n",
      "Epoch [   10/   20] | d_loss: 0.2146 | g_loss: 1.0391\n",
      "Epoch [   10/   20] | d_loss: 0.0886 | g_loss: 0.7239\n",
      "Epoch [   10/   20] | d_loss: 0.1411 | g_loss: 0.5891\n",
      "Epoch [   10/   20] | d_loss: 0.1720 | g_loss: 0.4467\n",
      "Epoch [   10/   20] | d_loss: 0.2347 | g_loss: 0.1932\n",
      "Epoch [   10/   20] | d_loss: 0.1424 | g_loss: 0.8711\n",
      "Epoch [   10/   20] | d_loss: 0.0811 | g_loss: 0.6151\n",
      "Epoch [   10/   20] | d_loss: 0.1442 | g_loss: 0.7636\n",
      "Epoch [   10/   20] | d_loss: 0.2365 | g_loss: 0.2642\n",
      "Epoch [   10/   20] | d_loss: 0.1511 | g_loss: 0.6174\n",
      "Epoch [   10/   20] | d_loss: 0.2494 | g_loss: 0.3750\n",
      "Epoch [   10/   20] | d_loss: 0.1109 | g_loss: 0.7041\n",
      "Epoch [   10/   20] | d_loss: 0.1601 | g_loss: 0.6255\n",
      "Epoch [   10/   20] | d_loss: 0.1082 | g_loss: 1.0811\n",
      "Epoch [   10/   20] | d_loss: 0.0708 | g_loss: 0.9558\n",
      "Epoch [   10/   20] | d_loss: 0.1603 | g_loss: 0.9264\n",
      "Epoch [   10/   20] | d_loss: 0.0825 | g_loss: 0.8628\n",
      "Epoch [   10/   20] | d_loss: 0.1675 | g_loss: 0.8469\n",
      "Epoch [   10/   20] | d_loss: 0.1483 | g_loss: 0.6853\n",
      "Epoch [   10/   20] | d_loss: 0.1853 | g_loss: 0.7565\n",
      "Epoch [   10/   20] | d_loss: 0.0702 | g_loss: 0.7371\n",
      "Epoch [   10/   20] | d_loss: 0.2226 | g_loss: 0.2916\n",
      "Epoch [   10/   20] | d_loss: 0.0604 | g_loss: 1.2345\n",
      "Epoch [   10/   20] | d_loss: 0.1687 | g_loss: 0.5941\n",
      "Epoch [   10/   20] | d_loss: 0.0785 | g_loss: 0.4874\n",
      "Epoch [   10/   20] | d_loss: 0.1450 | g_loss: 0.7610\n",
      "Epoch [   10/   20] | d_loss: 0.1538 | g_loss: 0.7262\n",
      "Epoch [   10/   20] | d_loss: 0.1189 | g_loss: 0.9359\n",
      "Epoch [   10/   20] | d_loss: 0.1069 | g_loss: 0.5752\n",
      "Epoch [   10/   20] | d_loss: 0.1826 | g_loss: 0.3609\n",
      "Epoch [   10/   20] | d_loss: 0.0815 | g_loss: 0.7463\n",
      "Epoch [   10/   20] | d_loss: 0.1965 | g_loss: 0.4685\n",
      "Epoch [   10/   20] | d_loss: 0.1369 | g_loss: 0.7369\n",
      "Epoch [   10/   20] | d_loss: 0.1198 | g_loss: 0.8090\n",
      "Epoch [   10/   20] | d_loss: 0.1230 | g_loss: 0.6622\n",
      "Epoch [   10/   20] | d_loss: 0.1946 | g_loss: 0.8167\n",
      "Epoch [   10/   20] | d_loss: 0.1318 | g_loss: 0.7282\n",
      "Epoch [   10/   20] | d_loss: 0.0819 | g_loss: 0.3867\n",
      "Epoch [   10/   20] | d_loss: 0.1005 | g_loss: 0.3775\n",
      "Epoch [   10/   20] | d_loss: 0.2188 | g_loss: 0.3191\n",
      "Epoch [   10/   20] | d_loss: 0.2727 | g_loss: 1.3329\n",
      "Epoch [   10/   20] | d_loss: 0.1063 | g_loss: 0.7518\n",
      "Epoch [   10/   20] | d_loss: 0.0622 | g_loss: 0.7575\n",
      "Epoch [   10/   20] | d_loss: 0.1893 | g_loss: 0.4563\n",
      "Epoch [   10/   20] | d_loss: 0.1581 | g_loss: 0.7401\n",
      "Epoch [   10/   20] | d_loss: 0.0530 | g_loss: 0.4566\n",
      "Epoch [   10/   20] | d_loss: 0.1226 | g_loss: 0.8365\n",
      "Epoch [   10/   20] | d_loss: 0.1228 | g_loss: 0.7768\n",
      "Epoch [   10/   20] | d_loss: 0.0886 | g_loss: 0.7080\n",
      "Epoch [   10/   20] | d_loss: 0.0765 | g_loss: 0.5195\n",
      "Epoch [   10/   20] | d_loss: 0.1699 | g_loss: 1.2560\n",
      "Epoch [   10/   20] | d_loss: 0.3042 | g_loss: 0.7202\n",
      "Epoch [   10/   20] | d_loss: 0.1902 | g_loss: 0.4778\n",
      "Epoch [   10/   20] | d_loss: 0.2755 | g_loss: 0.5319\n",
      "Epoch [   10/   20] | d_loss: 0.0686 | g_loss: 0.5684\n",
      "Epoch [   10/   20] | d_loss: 0.0778 | g_loss: 0.6223\n",
      "Epoch [   10/   20] | d_loss: 0.1317 | g_loss: 0.3742\n",
      "Epoch [   10/   20] | d_loss: 0.0766 | g_loss: 0.6013\n",
      "Epoch [   10/   20] | d_loss: 0.0905 | g_loss: 0.4448\n",
      "Epoch [   10/   20] | d_loss: 0.1206 | g_loss: 0.5988\n",
      "Epoch [   10/   20] | d_loss: 0.0902 | g_loss: 0.4440\n",
      "Epoch [   10/   20] | d_loss: 0.1639 | g_loss: 0.6589\n",
      "Epoch [   10/   20] | d_loss: 0.0917 | g_loss: 0.5482\n",
      "Epoch [   10/   20] | d_loss: 0.0898 | g_loss: 0.6670\n",
      "Epoch [   10/   20] | d_loss: 0.0614 | g_loss: 0.5565\n",
      "Epoch [   10/   20] | d_loss: 0.1985 | g_loss: 0.6025\n",
      "Epoch [   10/   20] | d_loss: 0.1407 | g_loss: 0.5884\n",
      "Epoch [   10/   20] | d_loss: 0.1473 | g_loss: 0.3955\n",
      "Epoch [   10/   20] | d_loss: 0.0986 | g_loss: 0.7658\n",
      "Epoch [   10/   20] | d_loss: 0.1829 | g_loss: 0.5340\n",
      "Epoch [   10/   20] | d_loss: 0.1246 | g_loss: 0.5715\n",
      "Epoch [   10/   20] | d_loss: 0.1391 | g_loss: 0.6473\n",
      "Epoch [   10/   20] | d_loss: 0.1275 | g_loss: 0.5426\n",
      "Epoch [   10/   20] | d_loss: 0.0766 | g_loss: 0.6540\n",
      "Epoch [   10/   20] | d_loss: 0.1106 | g_loss: 0.7084\n",
      "Epoch [   10/   20] | d_loss: 0.0918 | g_loss: 0.4380\n",
      "Epoch [   10/   20] | d_loss: 0.1489 | g_loss: 0.6330\n",
      "Epoch [   10/   20] | d_loss: 0.2705 | g_loss: 0.3240\n",
      "Epoch [   10/   20] | d_loss: 0.0719 | g_loss: 0.8159\n",
      "Epoch [   10/   20] | d_loss: 0.1221 | g_loss: 0.5408\n",
      "Epoch [   10/   20] | d_loss: 0.0411 | g_loss: 0.4558\n",
      "Epoch [   11/   20] | d_loss: 0.1132 | g_loss: 0.5912\n",
      "Epoch [   11/   20] | d_loss: 0.1162 | g_loss: 1.0574\n",
      "Epoch [   11/   20] | d_loss: 0.1940 | g_loss: 0.6070\n",
      "Epoch [   11/   20] | d_loss: 0.1222 | g_loss: 0.4279\n",
      "Epoch [   11/   20] | d_loss: 0.2410 | g_loss: 0.6897\n",
      "Epoch [   11/   20] | d_loss: 0.1914 | g_loss: 0.3858\n",
      "Epoch [   11/   20] | d_loss: 0.0629 | g_loss: 0.3237\n",
      "Epoch [   11/   20] | d_loss: 0.0834 | g_loss: 0.5780\n",
      "Epoch [   11/   20] | d_loss: 0.1280 | g_loss: 0.5242\n",
      "Epoch [   11/   20] | d_loss: 0.0601 | g_loss: 0.8991\n",
      "Epoch [   11/   20] | d_loss: 0.0486 | g_loss: 0.4326\n",
      "Epoch [   11/   20] | d_loss: 0.1732 | g_loss: 0.4950\n",
      "Epoch [   11/   20] | d_loss: 0.3071 | g_loss: 0.3398\n",
      "Epoch [   11/   20] | d_loss: 0.1769 | g_loss: 0.6925\n",
      "Epoch [   11/   20] | d_loss: 0.1007 | g_loss: 0.7754\n",
      "Epoch [   11/   20] | d_loss: 0.0687 | g_loss: 0.3239\n",
      "Epoch [   11/   20] | d_loss: 0.0980 | g_loss: 0.9579\n",
      "Epoch [   11/   20] | d_loss: 0.1499 | g_loss: 0.3611\n",
      "Epoch [   11/   20] | d_loss: 0.1578 | g_loss: 0.4174\n",
      "Epoch [   11/   20] | d_loss: 0.2165 | g_loss: 0.8115\n",
      "Epoch [   11/   20] | d_loss: 0.0861 | g_loss: 0.6706\n",
      "Epoch [   11/   20] | d_loss: 0.1495 | g_loss: 0.4918\n",
      "Epoch [   11/   20] | d_loss: 0.1633 | g_loss: 0.3037\n",
      "Epoch [   11/   20] | d_loss: 0.1409 | g_loss: 0.2925\n",
      "Epoch [   11/   20] | d_loss: 0.1011 | g_loss: 0.7490\n",
      "Epoch [   11/   20] | d_loss: 0.0658 | g_loss: 0.4486\n",
      "Epoch [   11/   20] | d_loss: 0.1284 | g_loss: 0.6945\n",
      "Epoch [   11/   20] | d_loss: 0.1162 | g_loss: 0.3473\n",
      "Epoch [   11/   20] | d_loss: 0.0897 | g_loss: 0.4329\n",
      "Epoch [   11/   20] | d_loss: 0.1540 | g_loss: 0.8336\n",
      "Epoch [   11/   20] | d_loss: 0.2236 | g_loss: 0.4731\n",
      "Epoch [   11/   20] | d_loss: 0.0973 | g_loss: 0.6612\n",
      "Epoch [   11/   20] | d_loss: 0.1650 | g_loss: 0.7233\n",
      "Epoch [   11/   20] | d_loss: 0.1217 | g_loss: 0.8576\n",
      "Epoch [   11/   20] | d_loss: 0.0921 | g_loss: 0.5656\n",
      "Epoch [   11/   20] | d_loss: 0.1386 | g_loss: 0.4646\n",
      "Epoch [   11/   20] | d_loss: 0.2259 | g_loss: 0.4643\n",
      "Epoch [   11/   20] | d_loss: 0.0786 | g_loss: 0.7406\n",
      "Epoch [   11/   20] | d_loss: 0.0896 | g_loss: 0.7412\n",
      "Epoch [   11/   20] | d_loss: 0.2127 | g_loss: 0.5559\n",
      "Epoch [   11/   20] | d_loss: 0.1468 | g_loss: 0.5385\n",
      "Epoch [   11/   20] | d_loss: 0.1240 | g_loss: 0.5954\n",
      "Epoch [   11/   20] | d_loss: 0.0656 | g_loss: 0.6580\n",
      "Epoch [   11/   20] | d_loss: 0.1141 | g_loss: 0.7572\n",
      "Epoch [   11/   20] | d_loss: 0.3902 | g_loss: 0.4920\n",
      "Epoch [   11/   20] | d_loss: 0.1358 | g_loss: 0.6122\n",
      "Epoch [   11/   20] | d_loss: 0.0768 | g_loss: 0.3420\n",
      "Epoch [   11/   20] | d_loss: 0.4258 | g_loss: 0.6126\n",
      "Epoch [   11/   20] | d_loss: 0.2282 | g_loss: 0.9461\n",
      "Epoch [   11/   20] | d_loss: 0.0585 | g_loss: 0.5924\n",
      "Epoch [   11/   20] | d_loss: 0.3784 | g_loss: 1.1196\n",
      "Epoch [   11/   20] | d_loss: 0.0768 | g_loss: 0.5636\n",
      "Epoch [   11/   20] | d_loss: 0.1063 | g_loss: 0.6096\n",
      "Epoch [   11/   20] | d_loss: 0.1565 | g_loss: 0.4319\n",
      "Epoch [   11/   20] | d_loss: 0.1244 | g_loss: 0.5725\n",
      "Epoch [   11/   20] | d_loss: 0.0720 | g_loss: 0.6117\n",
      "Epoch [   11/   20] | d_loss: 0.1078 | g_loss: 0.1339\n",
      "Epoch [   11/   20] | d_loss: 0.2133 | g_loss: 1.0251\n",
      "Epoch [   11/   20] | d_loss: 0.2727 | g_loss: 0.5841\n",
      "Epoch [   11/   20] | d_loss: 0.1219 | g_loss: 0.4327\n",
      "Epoch [   11/   20] | d_loss: 0.2041 | g_loss: 0.5856\n",
      "Epoch [   11/   20] | d_loss: 0.1628 | g_loss: 0.6349\n",
      "Epoch [   11/   20] | d_loss: 0.1025 | g_loss: 0.7279\n",
      "Epoch [   11/   20] | d_loss: 0.3424 | g_loss: 0.1528\n",
      "Epoch [   11/   20] | d_loss: 0.1449 | g_loss: 0.8521\n",
      "Epoch [   11/   20] | d_loss: 0.1536 | g_loss: 0.5344\n",
      "Epoch [   11/   20] | d_loss: 0.1722 | g_loss: 0.5436\n",
      "Epoch [   11/   20] | d_loss: 0.1364 | g_loss: 0.5136\n",
      "Epoch [   11/   20] | d_loss: 0.0601 | g_loss: 1.0189\n",
      "Epoch [   11/   20] | d_loss: 0.1626 | g_loss: 0.7389\n",
      "Epoch [   11/   20] | d_loss: 0.1483 | g_loss: 0.6102\n",
      "Epoch [   11/   20] | d_loss: 0.2277 | g_loss: 0.8197\n",
      "Epoch [   11/   20] | d_loss: 0.0722 | g_loss: 0.5991\n",
      "Epoch [   11/   20] | d_loss: 0.2015 | g_loss: 0.7629\n",
      "Epoch [   11/   20] | d_loss: 0.1926 | g_loss: 0.5937\n",
      "Epoch [   11/   20] | d_loss: 0.1470 | g_loss: 0.4294\n",
      "Epoch [   11/   20] | d_loss: 0.1965 | g_loss: 1.0957\n",
      "Epoch [   11/   20] | d_loss: 0.0691 | g_loss: 0.9148\n",
      "Epoch [   11/   20] | d_loss: 0.0761 | g_loss: 0.3966\n",
      "Epoch [   11/   20] | d_loss: 0.0817 | g_loss: 0.6251\n",
      "Epoch [   11/   20] | d_loss: 0.0993 | g_loss: 0.6214\n",
      "Epoch [   11/   20] | d_loss: 0.0937 | g_loss: 0.8223\n",
      "Epoch [   11/   20] | d_loss: 0.2012 | g_loss: 0.6058\n",
      "Epoch [   11/   20] | d_loss: 0.0766 | g_loss: 0.7019\n",
      "Epoch [   11/   20] | d_loss: 0.2897 | g_loss: 0.7024\n",
      "Epoch [   11/   20] | d_loss: 0.1480 | g_loss: 0.6339\n",
      "Epoch [   11/   20] | d_loss: 0.4262 | g_loss: 0.6449\n",
      "Epoch [   11/   20] | d_loss: 0.1280 | g_loss: 0.7405\n",
      "Epoch [   11/   20] | d_loss: 0.1314 | g_loss: 0.7610\n",
      "Epoch [   11/   20] | d_loss: 0.1915 | g_loss: 0.5023\n",
      "Epoch [   12/   20] | d_loss: 0.1556 | g_loss: 0.6268\n",
      "Epoch [   12/   20] | d_loss: 0.2643 | g_loss: 0.6699\n",
      "Epoch [   12/   20] | d_loss: 0.0795 | g_loss: 0.7233\n",
      "Epoch [   12/   20] | d_loss: 0.1955 | g_loss: 0.8634\n",
      "Epoch [   12/   20] | d_loss: 0.2303 | g_loss: 0.4170\n",
      "Epoch [   12/   20] | d_loss: 0.1530 | g_loss: 0.6618\n",
      "Epoch [   12/   20] | d_loss: 0.0739 | g_loss: 0.6714\n",
      "Epoch [   12/   20] | d_loss: 0.1064 | g_loss: 0.3334\n",
      "Epoch [   12/   20] | d_loss: 0.1614 | g_loss: 0.4904\n",
      "Epoch [   12/   20] | d_loss: 0.1808 | g_loss: 0.7113\n",
      "Epoch [   12/   20] | d_loss: 0.1308 | g_loss: 1.0585\n",
      "Epoch [   12/   20] | d_loss: 0.2610 | g_loss: 0.5158\n",
      "Epoch [   12/   20] | d_loss: 0.1385 | g_loss: 0.7184\n",
      "Epoch [   12/   20] | d_loss: 0.0846 | g_loss: 0.8928\n",
      "Epoch [   12/   20] | d_loss: 0.1162 | g_loss: 0.6307\n",
      "Epoch [   12/   20] | d_loss: 0.0570 | g_loss: 0.6072\n",
      "Epoch [   12/   20] | d_loss: 0.1504 | g_loss: 0.7116\n",
      "Epoch [   12/   20] | d_loss: 0.1670 | g_loss: 0.5739\n",
      "Epoch [   12/   20] | d_loss: 0.1071 | g_loss: 0.7617\n",
      "Epoch [   12/   20] | d_loss: 0.1012 | g_loss: 0.6144\n",
      "Epoch [   12/   20] | d_loss: 0.1113 | g_loss: 0.5337\n",
      "Epoch [   12/   20] | d_loss: 0.1252 | g_loss: 0.6227\n",
      "Epoch [   12/   20] | d_loss: 0.2499 | g_loss: 0.7138\n",
      "Epoch [   12/   20] | d_loss: 0.2029 | g_loss: 0.4968\n",
      "Epoch [   12/   20] | d_loss: 0.1130 | g_loss: 0.6677\n",
      "Epoch [   12/   20] | d_loss: 0.0883 | g_loss: 0.6556\n",
      "Epoch [   12/   20] | d_loss: 0.0636 | g_loss: 0.6551\n",
      "Epoch [   12/   20] | d_loss: 0.1771 | g_loss: 0.5784\n",
      "Epoch [   12/   20] | d_loss: 0.1158 | g_loss: 0.6788\n",
      "Epoch [   12/   20] | d_loss: 0.1376 | g_loss: 0.8546\n",
      "Epoch [   12/   20] | d_loss: 0.1134 | g_loss: 0.4894\n",
      "Epoch [   12/   20] | d_loss: 0.0907 | g_loss: 0.8757\n",
      "Epoch [   12/   20] | d_loss: 0.0995 | g_loss: 0.6207\n",
      "Epoch [   12/   20] | d_loss: 0.0948 | g_loss: 0.5398\n",
      "Epoch [   12/   20] | d_loss: 0.0911 | g_loss: 0.8845\n",
      "Epoch [   12/   20] | d_loss: 0.0953 | g_loss: 0.6468\n",
      "Epoch [   12/   20] | d_loss: 0.1735 | g_loss: 0.6415\n",
      "Epoch [   12/   20] | d_loss: 0.1049 | g_loss: 0.9227\n",
      "Epoch [   12/   20] | d_loss: 0.1055 | g_loss: 0.6506\n",
      "Epoch [   12/   20] | d_loss: 0.0394 | g_loss: 0.2918\n",
      "Epoch [   12/   20] | d_loss: 0.1414 | g_loss: 0.3236\n",
      "Epoch [   12/   20] | d_loss: 0.1381 | g_loss: 0.6251\n",
      "Epoch [   12/   20] | d_loss: 0.0569 | g_loss: 0.5790\n",
      "Epoch [   12/   20] | d_loss: 0.0829 | g_loss: 0.8712\n",
      "Epoch [   12/   20] | d_loss: 0.0772 | g_loss: 0.9234\n",
      "Epoch [   12/   20] | d_loss: 0.1345 | g_loss: 0.6265\n",
      "Epoch [   12/   20] | d_loss: 0.1818 | g_loss: 0.5796\n",
      "Epoch [   12/   20] | d_loss: 0.0867 | g_loss: 0.6031\n",
      "Epoch [   12/   20] | d_loss: 0.0999 | g_loss: 0.2786\n",
      "Epoch [   12/   20] | d_loss: 0.0831 | g_loss: 0.4502\n",
      "Epoch [   12/   20] | d_loss: 0.0830 | g_loss: 0.3450\n",
      "Epoch [   12/   20] | d_loss: 0.1095 | g_loss: 0.5699\n",
      "Epoch [   12/   20] | d_loss: 0.0943 | g_loss: 0.7377\n",
      "Epoch [   12/   20] | d_loss: 0.0801 | g_loss: 0.6267\n",
      "Epoch [   12/   20] | d_loss: 0.1574 | g_loss: 0.6399\n",
      "Epoch [   12/   20] | d_loss: 0.0931 | g_loss: 0.2809\n",
      "Epoch [   12/   20] | d_loss: 0.1316 | g_loss: 0.6722\n",
      "Epoch [   12/   20] | d_loss: 0.1458 | g_loss: 0.5810\n",
      "Epoch [   12/   20] | d_loss: 0.1521 | g_loss: 0.7085\n",
      "Epoch [   12/   20] | d_loss: 0.0609 | g_loss: 0.4072\n",
      "Epoch [   12/   20] | d_loss: 0.2822 | g_loss: 1.1781\n",
      "Epoch [   12/   20] | d_loss: 0.1001 | g_loss: 0.5586\n",
      "Epoch [   12/   20] | d_loss: 0.0969 | g_loss: 0.7131\n",
      "Epoch [   12/   20] | d_loss: 0.1257 | g_loss: 0.4832\n",
      "Epoch [   12/   20] | d_loss: 0.1315 | g_loss: 1.1273\n",
      "Epoch [   12/   20] | d_loss: 0.1473 | g_loss: 0.6768\n",
      "Epoch [   12/   20] | d_loss: 0.1953 | g_loss: 0.6248\n",
      "Epoch [   12/   20] | d_loss: 0.2330 | g_loss: 0.4809\n",
      "Epoch [   12/   20] | d_loss: 0.1244 | g_loss: 0.4248\n",
      "Epoch [   12/   20] | d_loss: 0.0869 | g_loss: 0.5698\n",
      "Epoch [   12/   20] | d_loss: 0.0500 | g_loss: 0.7337\n",
      "Epoch [   12/   20] | d_loss: 0.1385 | g_loss: 0.4443\n",
      "Epoch [   12/   20] | d_loss: 0.2632 | g_loss: 0.4028\n",
      "Epoch [   12/   20] | d_loss: 0.5218 | g_loss: 0.4241\n",
      "Epoch [   12/   20] | d_loss: 0.1395 | g_loss: 0.6748\n",
      "Epoch [   12/   20] | d_loss: 0.1630 | g_loss: 0.7085\n",
      "Epoch [   12/   20] | d_loss: 0.1606 | g_loss: 0.3840\n",
      "Epoch [   12/   20] | d_loss: 0.1650 | g_loss: 0.8384\n",
      "Epoch [   12/   20] | d_loss: 0.0725 | g_loss: 0.3446\n",
      "Epoch [   12/   20] | d_loss: 0.2236 | g_loss: 0.4404\n",
      "Epoch [   12/   20] | d_loss: 0.1203 | g_loss: 0.9957\n",
      "Epoch [   12/   20] | d_loss: 0.0650 | g_loss: 0.7250\n",
      "Epoch [   12/   20] | d_loss: 0.0833 | g_loss: 0.6890\n",
      "Epoch [   12/   20] | d_loss: 0.1708 | g_loss: 0.5034\n",
      "Epoch [   12/   20] | d_loss: 0.1186 | g_loss: 0.5581\n",
      "Epoch [   12/   20] | d_loss: 0.0845 | g_loss: 0.6917\n",
      "Epoch [   12/   20] | d_loss: 0.1311 | g_loss: 0.4369\n",
      "Epoch [   12/   20] | d_loss: 0.2333 | g_loss: 0.7286\n",
      "Epoch [   12/   20] | d_loss: 0.1135 | g_loss: 0.5918\n",
      "Epoch [   12/   20] | d_loss: 0.0474 | g_loss: 0.7538\n",
      "Epoch [   13/   20] | d_loss: 0.0787 | g_loss: 0.4525\n",
      "Epoch [   13/   20] | d_loss: 0.1839 | g_loss: 1.1511\n",
      "Epoch [   13/   20] | d_loss: 0.0665 | g_loss: 0.9641\n",
      "Epoch [   13/   20] | d_loss: 0.1108 | g_loss: 0.6658\n",
      "Epoch [   13/   20] | d_loss: 0.0503 | g_loss: 0.7089\n",
      "Epoch [   13/   20] | d_loss: 0.0891 | g_loss: 0.6773\n",
      "Epoch [   13/   20] | d_loss: 0.0918 | g_loss: 0.4238\n",
      "Epoch [   13/   20] | d_loss: 0.0875 | g_loss: 1.1348\n",
      "Epoch [   13/   20] | d_loss: 0.1073 | g_loss: 0.2980\n",
      "Epoch [   13/   20] | d_loss: 0.1733 | g_loss: 0.8189\n",
      "Epoch [   13/   20] | d_loss: 0.1136 | g_loss: 0.4556\n",
      "Epoch [   13/   20] | d_loss: 0.1043 | g_loss: 0.5311\n",
      "Epoch [   13/   20] | d_loss: 0.0536 | g_loss: 0.5566\n",
      "Epoch [   13/   20] | d_loss: 0.0810 | g_loss: 0.7376\n",
      "Epoch [   13/   20] | d_loss: 0.0649 | g_loss: 0.8694\n",
      "Epoch [   13/   20] | d_loss: 0.0533 | g_loss: 0.9362\n",
      "Epoch [   13/   20] | d_loss: 0.0670 | g_loss: 0.9210\n",
      "Epoch [   13/   20] | d_loss: 0.0446 | g_loss: 0.8332\n",
      "Epoch [   13/   20] | d_loss: 0.1164 | g_loss: 1.1613\n",
      "Epoch [   13/   20] | d_loss: 0.1040 | g_loss: 0.6093\n",
      "Epoch [   13/   20] | d_loss: 0.1308 | g_loss: 0.6688\n",
      "Epoch [   13/   20] | d_loss: 0.1637 | g_loss: 0.5523\n",
      "Epoch [   13/   20] | d_loss: 0.1437 | g_loss: 0.4173\n",
      "Epoch [   13/   20] | d_loss: 0.1667 | g_loss: 0.6844\n",
      "Epoch [   13/   20] | d_loss: 0.0912 | g_loss: 0.7226\n",
      "Epoch [   13/   20] | d_loss: 0.1180 | g_loss: 0.8798\n",
      "Epoch [   13/   20] | d_loss: 0.0845 | g_loss: 0.8903\n",
      "Epoch [   13/   20] | d_loss: 0.1504 | g_loss: 0.7999\n",
      "Epoch [   13/   20] | d_loss: 0.1088 | g_loss: 0.7717\n",
      "Epoch [   13/   20] | d_loss: 0.0964 | g_loss: 0.9940\n",
      "Epoch [   13/   20] | d_loss: 0.1076 | g_loss: 0.7374\n",
      "Epoch [   13/   20] | d_loss: 0.1434 | g_loss: 0.6848\n",
      "Epoch [   13/   20] | d_loss: 0.0853 | g_loss: 0.9779\n",
      "Epoch [   13/   20] | d_loss: 0.1966 | g_loss: 0.5111\n",
      "Epoch [   13/   20] | d_loss: 0.0915 | g_loss: 0.6139\n",
      "Epoch [   13/   20] | d_loss: 0.1547 | g_loss: 0.8375\n",
      "Epoch [   13/   20] | d_loss: 0.2134 | g_loss: 0.6793\n",
      "Epoch [   13/   20] | d_loss: 0.1171 | g_loss: 0.5403\n",
      "Epoch [   13/   20] | d_loss: 0.0600 | g_loss: 0.7804\n",
      "Epoch [   13/   20] | d_loss: 0.1559 | g_loss: 0.5006\n",
      "Epoch [   13/   20] | d_loss: 0.4597 | g_loss: 0.6317\n",
      "Epoch [   13/   20] | d_loss: 0.1479 | g_loss: 0.5067\n",
      "Epoch [   13/   20] | d_loss: 0.0708 | g_loss: 0.8838\n",
      "Epoch [   13/   20] | d_loss: 0.1261 | g_loss: 0.3815\n",
      "Epoch [   13/   20] | d_loss: 0.2309 | g_loss: 0.6543\n",
      "Epoch [   13/   20] | d_loss: 0.0970 | g_loss: 0.7243\n",
      "Epoch [   13/   20] | d_loss: 0.1805 | g_loss: 0.5713\n",
      "Epoch [   13/   20] | d_loss: 0.1513 | g_loss: 0.9355\n",
      "Epoch [   13/   20] | d_loss: 0.1062 | g_loss: 0.4349\n",
      "Epoch [   13/   20] | d_loss: 0.1169 | g_loss: 1.2189\n",
      "Epoch [   13/   20] | d_loss: 0.1464 | g_loss: 0.5389\n",
      "Epoch [   13/   20] | d_loss: 0.1935 | g_loss: 1.1404\n",
      "Epoch [   13/   20] | d_loss: 0.1091 | g_loss: 1.1468\n",
      "Epoch [   13/   20] | d_loss: 0.0590 | g_loss: 0.5142\n",
      "Epoch [   13/   20] | d_loss: 0.0616 | g_loss: 0.6576\n",
      "Epoch [   13/   20] | d_loss: 0.1036 | g_loss: 0.6998\n",
      "Epoch [   13/   20] | d_loss: 0.1356 | g_loss: 0.5282\n",
      "Epoch [   13/   20] | d_loss: 0.1326 | g_loss: 0.8104\n",
      "Epoch [   13/   20] | d_loss: 0.1763 | g_loss: 0.6045\n",
      "Epoch [   13/   20] | d_loss: 0.0703 | g_loss: 0.7887\n",
      "Epoch [   13/   20] | d_loss: 0.0729 | g_loss: 0.5243\n",
      "Epoch [   13/   20] | d_loss: 0.2179 | g_loss: 0.5749\n",
      "Epoch [   13/   20] | d_loss: 0.0828 | g_loss: 0.4636\n",
      "Epoch [   13/   20] | d_loss: 0.3407 | g_loss: 0.6607\n",
      "Epoch [   13/   20] | d_loss: 0.0616 | g_loss: 0.8618\n",
      "Epoch [   13/   20] | d_loss: 0.1206 | g_loss: 0.7928\n",
      "Epoch [   13/   20] | d_loss: 0.2325 | g_loss: 0.7163\n",
      "Epoch [   13/   20] | d_loss: 0.1528 | g_loss: 0.5825\n",
      "Epoch [   13/   20] | d_loss: 0.1953 | g_loss: 0.4924\n",
      "Epoch [   13/   20] | d_loss: 0.0465 | g_loss: 0.5016\n",
      "Epoch [   13/   20] | d_loss: 0.0987 | g_loss: 0.5969\n",
      "Epoch [   13/   20] | d_loss: 0.1090 | g_loss: 0.7228\n",
      "Epoch [   13/   20] | d_loss: 0.0888 | g_loss: 0.9206\n",
      "Epoch [   13/   20] | d_loss: 0.1000 | g_loss: 0.7632\n",
      "Epoch [   13/   20] | d_loss: 0.1230 | g_loss: 0.6441\n",
      "Epoch [   13/   20] | d_loss: 0.1228 | g_loss: 0.3894\n",
      "Epoch [   13/   20] | d_loss: 0.3848 | g_loss: 1.6419\n",
      "Epoch [   13/   20] | d_loss: 0.2007 | g_loss: 0.7385\n",
      "Epoch [   13/   20] | d_loss: 0.1744 | g_loss: 0.3490\n",
      "Epoch [   13/   20] | d_loss: 0.0635 | g_loss: 0.9140\n",
      "Epoch [   13/   20] | d_loss: 0.1088 | g_loss: 0.7055\n",
      "Epoch [   13/   20] | d_loss: 0.0378 | g_loss: 0.5904\n",
      "Epoch [   13/   20] | d_loss: 0.1480 | g_loss: 0.5067\n",
      "Epoch [   13/   20] | d_loss: 0.0976 | g_loss: 0.4255\n",
      "Epoch [   13/   20] | d_loss: 0.0580 | g_loss: 0.4716\n",
      "Epoch [   13/   20] | d_loss: 0.1137 | g_loss: 0.7541\n",
      "Epoch [   13/   20] | d_loss: 0.0733 | g_loss: 0.4025\n",
      "Epoch [   13/   20] | d_loss: 0.1686 | g_loss: 0.7800\n",
      "Epoch [   13/   20] | d_loss: 0.0546 | g_loss: 0.5376\n",
      "Epoch [   13/   20] | d_loss: 0.1620 | g_loss: 0.3495\n",
      "Epoch [   14/   20] | d_loss: 0.3026 | g_loss: 0.9534\n",
      "Epoch [   14/   20] | d_loss: 0.0911 | g_loss: 0.6014\n",
      "Epoch [   14/   20] | d_loss: 0.0881 | g_loss: 0.7542\n",
      "Epoch [   14/   20] | d_loss: 0.1076 | g_loss: 0.6520\n",
      "Epoch [   14/   20] | d_loss: 0.1780 | g_loss: 0.6238\n",
      "Epoch [   14/   20] | d_loss: 0.0645 | g_loss: 0.7487\n",
      "Epoch [   14/   20] | d_loss: 0.1788 | g_loss: 0.7205\n",
      "Epoch [   14/   20] | d_loss: 0.0804 | g_loss: 0.8784\n",
      "Epoch [   14/   20] | d_loss: 0.2450 | g_loss: 0.6830\n",
      "Epoch [   14/   20] | d_loss: 0.1553 | g_loss: 0.5055\n",
      "Epoch [   14/   20] | d_loss: 0.1455 | g_loss: 0.6088\n",
      "Epoch [   14/   20] | d_loss: 0.1775 | g_loss: 0.4947\n",
      "Epoch [   14/   20] | d_loss: 0.0523 | g_loss: 0.9053\n",
      "Epoch [   14/   20] | d_loss: 0.0522 | g_loss: 0.7780\n",
      "Epoch [   14/   20] | d_loss: 0.1268 | g_loss: 0.5576\n",
      "Epoch [   14/   20] | d_loss: 0.0997 | g_loss: 0.8352\n",
      "Epoch [   14/   20] | d_loss: 0.0919 | g_loss: 0.8264\n",
      "Epoch [   14/   20] | d_loss: 0.0819 | g_loss: 0.6570\n",
      "Epoch [   14/   20] | d_loss: 0.0622 | g_loss: 0.5401\n",
      "Epoch [   14/   20] | d_loss: 0.1224 | g_loss: 1.1241\n",
      "Epoch [   14/   20] | d_loss: 0.1687 | g_loss: 0.4953\n",
      "Epoch [   14/   20] | d_loss: 0.1376 | g_loss: 0.8342\n",
      "Epoch [   14/   20] | d_loss: 0.1936 | g_loss: 0.5882\n",
      "Epoch [   14/   20] | d_loss: 0.1052 | g_loss: 0.5879\n",
      "Epoch [   14/   20] | d_loss: 0.0727 | g_loss: 0.5580\n",
      "Epoch [   14/   20] | d_loss: 0.3785 | g_loss: 0.5207\n",
      "Epoch [   14/   20] | d_loss: 0.4152 | g_loss: 0.3157\n",
      "Epoch [   14/   20] | d_loss: 0.1289 | g_loss: 0.6639\n",
      "Epoch [   14/   20] | d_loss: 0.0773 | g_loss: 0.5963\n",
      "Epoch [   14/   20] | d_loss: 0.0644 | g_loss: 0.8875\n",
      "Epoch [   14/   20] | d_loss: 0.0424 | g_loss: 0.8984\n",
      "Epoch [   14/   20] | d_loss: 0.1771 | g_loss: 0.5965\n",
      "Epoch [   14/   20] | d_loss: 0.1170 | g_loss: 0.4498\n",
      "Epoch [   14/   20] | d_loss: 0.1937 | g_loss: 0.4601\n",
      "Epoch [   14/   20] | d_loss: 0.1336 | g_loss: 0.4415\n",
      "Epoch [   14/   20] | d_loss: 0.2212 | g_loss: 0.6957\n",
      "Epoch [   14/   20] | d_loss: 0.0505 | g_loss: 0.8987\n",
      "Epoch [   14/   20] | d_loss: 0.0572 | g_loss: 0.7523\n",
      "Epoch [   14/   20] | d_loss: 0.1130 | g_loss: 0.4130\n",
      "Epoch [   14/   20] | d_loss: 0.3247 | g_loss: 0.8203\n",
      "Epoch [   14/   20] | d_loss: 0.1221 | g_loss: 0.5483\n",
      "Epoch [   14/   20] | d_loss: 0.1896 | g_loss: 0.3230\n",
      "Epoch [   14/   20] | d_loss: 0.0549 | g_loss: 0.6218\n",
      "Epoch [   14/   20] | d_loss: 0.2119 | g_loss: 0.4719\n",
      "Epoch [   14/   20] | d_loss: 0.1690 | g_loss: 0.4169\n",
      "Epoch [   14/   20] | d_loss: 0.1003 | g_loss: 0.5171\n",
      "Epoch [   14/   20] | d_loss: 0.0493 | g_loss: 0.6458\n",
      "Epoch [   14/   20] | d_loss: 0.2326 | g_loss: 0.5567\n",
      "Epoch [   14/   20] | d_loss: 0.0945 | g_loss: 0.4387\n",
      "Epoch [   14/   20] | d_loss: 0.1716 | g_loss: 0.6492\n",
      "Epoch [   14/   20] | d_loss: 0.0907 | g_loss: 0.5431\n",
      "Epoch [   14/   20] | d_loss: 0.1913 | g_loss: 0.5746\n",
      "Epoch [   14/   20] | d_loss: 0.0879 | g_loss: 0.4320\n",
      "Epoch [   14/   20] | d_loss: 0.0948 | g_loss: 0.6953\n",
      "Epoch [   14/   20] | d_loss: 0.3280 | g_loss: 0.7343\n",
      "Epoch [   14/   20] | d_loss: 0.1260 | g_loss: 0.7504\n",
      "Epoch [   14/   20] | d_loss: 0.2496 | g_loss: 0.5621\n",
      "Epoch [   14/   20] | d_loss: 0.0397 | g_loss: 0.7747\n",
      "Epoch [   14/   20] | d_loss: 0.0847 | g_loss: 0.5600\n",
      "Epoch [   14/   20] | d_loss: 0.2006 | g_loss: 0.4096\n",
      "Epoch [   14/   20] | d_loss: 0.1910 | g_loss: 0.6212\n",
      "Epoch [   14/   20] | d_loss: 0.1105 | g_loss: 0.6920\n",
      "Epoch [   14/   20] | d_loss: 0.0983 | g_loss: 0.9409\n",
      "Epoch [   14/   20] | d_loss: 0.0806 | g_loss: 0.6277\n",
      "Epoch [   14/   20] | d_loss: 0.0725 | g_loss: 0.5997\n",
      "Epoch [   14/   20] | d_loss: 0.0761 | g_loss: 0.8027\n",
      "Epoch [   14/   20] | d_loss: 0.1506 | g_loss: 0.5693\n",
      "Epoch [   14/   20] | d_loss: 0.1698 | g_loss: 0.2963\n",
      "Epoch [   14/   20] | d_loss: 0.1650 | g_loss: 0.6136\n",
      "Epoch [   14/   20] | d_loss: 0.1324 | g_loss: 0.4909\n",
      "Epoch [   14/   20] | d_loss: 0.1157 | g_loss: 0.3191\n",
      "Epoch [   14/   20] | d_loss: 0.0840 | g_loss: 1.0181\n",
      "Epoch [   14/   20] | d_loss: 0.2222 | g_loss: 0.4313\n",
      "Epoch [   14/   20] | d_loss: 0.0976 | g_loss: 0.6670\n",
      "Epoch [   14/   20] | d_loss: 0.1696 | g_loss: 0.4635\n",
      "Epoch [   14/   20] | d_loss: 0.2476 | g_loss: 0.6084\n",
      "Epoch [   14/   20] | d_loss: 0.1719 | g_loss: 0.3624\n",
      "Epoch [   14/   20] | d_loss: 0.1446 | g_loss: 0.6315\n",
      "Epoch [   14/   20] | d_loss: 0.0731 | g_loss: 0.7069\n",
      "Epoch [   14/   20] | d_loss: 0.0660 | g_loss: 0.7236\n",
      "Epoch [   14/   20] | d_loss: 0.1333 | g_loss: 0.4696\n",
      "Epoch [   14/   20] | d_loss: 0.1479 | g_loss: 0.4890\n",
      "Epoch [   14/   20] | d_loss: 0.1923 | g_loss: 0.4332\n",
      "Epoch [   14/   20] | d_loss: 0.0748 | g_loss: 0.7867\n",
      "Epoch [   14/   20] | d_loss: 0.1171 | g_loss: 0.8340\n",
      "Epoch [   14/   20] | d_loss: 0.0495 | g_loss: 0.4310\n",
      "Epoch [   14/   20] | d_loss: 0.1394 | g_loss: 0.3658\n",
      "Epoch [   14/   20] | d_loss: 0.1765 | g_loss: 0.5364\n",
      "Epoch [   14/   20] | d_loss: 0.1775 | g_loss: 0.7254\n",
      "Epoch [   14/   20] | d_loss: 0.0994 | g_loss: 0.7100\n",
      "Epoch [   15/   20] | d_loss: 0.0739 | g_loss: 1.0066\n",
      "Epoch [   15/   20] | d_loss: 0.1628 | g_loss: 0.6393\n",
      "Epoch [   15/   20] | d_loss: 0.1822 | g_loss: 0.7475\n",
      "Epoch [   15/   20] | d_loss: 0.1114 | g_loss: 0.7964\n",
      "Epoch [   15/   20] | d_loss: 0.1263 | g_loss: 0.5554\n",
      "Epoch [   15/   20] | d_loss: 0.2338 | g_loss: 0.9616\n",
      "Epoch [   15/   20] | d_loss: 0.0620 | g_loss: 0.8224\n",
      "Epoch [   15/   20] | d_loss: 0.1361 | g_loss: 0.6363\n",
      "Epoch [   15/   20] | d_loss: 0.0814 | g_loss: 1.1376\n",
      "Epoch [   15/   20] | d_loss: 0.1030 | g_loss: 0.7175\n",
      "Epoch [   15/   20] | d_loss: 0.0924 | g_loss: 0.4854\n",
      "Epoch [   15/   20] | d_loss: 0.1604 | g_loss: 0.6086\n",
      "Epoch [   15/   20] | d_loss: 0.0944 | g_loss: 0.5009\n",
      "Epoch [   15/   20] | d_loss: 0.1351 | g_loss: 0.6457\n",
      "Epoch [   15/   20] | d_loss: 0.1501 | g_loss: 0.2551\n",
      "Epoch [   15/   20] | d_loss: 0.0968 | g_loss: 0.4047\n",
      "Epoch [   15/   20] | d_loss: 0.0766 | g_loss: 0.7770\n",
      "Epoch [   15/   20] | d_loss: 0.0379 | g_loss: 0.9815\n",
      "Epoch [   15/   20] | d_loss: 0.0843 | g_loss: 0.5298\n",
      "Epoch [   15/   20] | d_loss: 0.1438 | g_loss: 0.5743\n",
      "Epoch [   15/   20] | d_loss: 0.1338 | g_loss: 0.9292\n",
      "Epoch [   15/   20] | d_loss: 0.0977 | g_loss: 0.6408\n",
      "Epoch [   15/   20] | d_loss: 0.0784 | g_loss: 0.8227\n",
      "Epoch [   15/   20] | d_loss: 0.0979 | g_loss: 0.8938\n",
      "Epoch [   15/   20] | d_loss: 0.1248 | g_loss: 0.5750\n",
      "Epoch [   15/   20] | d_loss: 0.2508 | g_loss: 0.7159\n",
      "Epoch [   15/   20] | d_loss: 0.1782 | g_loss: 0.8130\n",
      "Epoch [   15/   20] | d_loss: 0.1456 | g_loss: 0.2653\n",
      "Epoch [   15/   20] | d_loss: 0.0428 | g_loss: 0.9914\n",
      "Epoch [   15/   20] | d_loss: 0.1812 | g_loss: 0.7662\n",
      "Epoch [   15/   20] | d_loss: 0.1333 | g_loss: 0.5359\n",
      "Epoch [   15/   20] | d_loss: 0.1181 | g_loss: 0.4417\n",
      "Epoch [   15/   20] | d_loss: 0.0946 | g_loss: 0.5562\n",
      "Epoch [   15/   20] | d_loss: 0.0634 | g_loss: 0.5982\n",
      "Epoch [   15/   20] | d_loss: 0.1124 | g_loss: 0.6274\n",
      "Epoch [   15/   20] | d_loss: 0.1518 | g_loss: 1.0054\n",
      "Epoch [   15/   20] | d_loss: 0.2403 | g_loss: 0.4478\n",
      "Epoch [   15/   20] | d_loss: 0.1652 | g_loss: 0.5061\n",
      "Epoch [   15/   20] | d_loss: 0.2028 | g_loss: 1.1303\n",
      "Epoch [   15/   20] | d_loss: 0.0992 | g_loss: 1.0397\n",
      "Epoch [   15/   20] | d_loss: 0.0635 | g_loss: 0.7077\n",
      "Epoch [   15/   20] | d_loss: 0.2415 | g_loss: 0.5680\n",
      "Epoch [   15/   20] | d_loss: 0.1773 | g_loss: 0.6021\n",
      "Epoch [   15/   20] | d_loss: 0.1809 | g_loss: 0.8299\n",
      "Epoch [   15/   20] | d_loss: 0.1648 | g_loss: 1.0977\n",
      "Epoch [   15/   20] | d_loss: 0.1550 | g_loss: 0.9582\n",
      "Epoch [   15/   20] | d_loss: 0.1045 | g_loss: 0.7712\n",
      "Epoch [   15/   20] | d_loss: 0.0733 | g_loss: 0.8085\n",
      "Epoch [   15/   20] | d_loss: 0.0628 | g_loss: 0.8539\n",
      "Epoch [   15/   20] | d_loss: 0.2144 | g_loss: 0.3190\n",
      "Epoch [   15/   20] | d_loss: 0.1385 | g_loss: 0.5064\n",
      "Epoch [   15/   20] | d_loss: 0.1353 | g_loss: 0.8808\n",
      "Epoch [   15/   20] | d_loss: 0.1353 | g_loss: 0.8747\n",
      "Epoch [   15/   20] | d_loss: 0.1118 | g_loss: 0.6480\n",
      "Epoch [   15/   20] | d_loss: 0.0525 | g_loss: 1.0061\n",
      "Epoch [   15/   20] | d_loss: 0.0836 | g_loss: 0.3305\n",
      "Epoch [   15/   20] | d_loss: 0.2036 | g_loss: 0.3987\n",
      "Epoch [   15/   20] | d_loss: 0.1179 | g_loss: 0.5498\n",
      "Epoch [   15/   20] | d_loss: 0.1335 | g_loss: 0.4904\n",
      "Epoch [   15/   20] | d_loss: 0.0767 | g_loss: 0.5938\n",
      "Epoch [   15/   20] | d_loss: 0.1112 | g_loss: 0.3123\n",
      "Epoch [   15/   20] | d_loss: 0.1658 | g_loss: 0.8011\n",
      "Epoch [   15/   20] | d_loss: 0.0810 | g_loss: 0.5673\n",
      "Epoch [   15/   20] | d_loss: 0.0992 | g_loss: 1.0696\n",
      "Epoch [   15/   20] | d_loss: 0.1614 | g_loss: 0.6692\n",
      "Epoch [   15/   20] | d_loss: 0.1964 | g_loss: 0.5616\n",
      "Epoch [   15/   20] | d_loss: 0.0573 | g_loss: 0.7279\n",
      "Epoch [   15/   20] | d_loss: 0.0388 | g_loss: 0.4665\n",
      "Epoch [   15/   20] | d_loss: 0.1116 | g_loss: 0.5192\n",
      "Epoch [   15/   20] | d_loss: 0.1320 | g_loss: 0.6833\n",
      "Epoch [   15/   20] | d_loss: 0.1723 | g_loss: 0.4181\n",
      "Epoch [   15/   20] | d_loss: 0.0766 | g_loss: 0.7719\n",
      "Epoch [   15/   20] | d_loss: 0.0390 | g_loss: 0.5248\n",
      "Epoch [   15/   20] | d_loss: 0.0604 | g_loss: 0.6115\n",
      "Epoch [   15/   20] | d_loss: 0.0932 | g_loss: 0.7873\n",
      "Epoch [   15/   20] | d_loss: 0.0607 | g_loss: 0.4751\n",
      "Epoch [   15/   20] | d_loss: 0.3719 | g_loss: 0.7782\n",
      "Epoch [   15/   20] | d_loss: 0.2122 | g_loss: 0.7552\n",
      "Epoch [   15/   20] | d_loss: 0.0908 | g_loss: 0.4854\n",
      "Epoch [   15/   20] | d_loss: 0.3590 | g_loss: 0.4402\n",
      "Epoch [   15/   20] | d_loss: 0.0769 | g_loss: 0.6144\n",
      "Epoch [   15/   20] | d_loss: 0.1447 | g_loss: 0.4241\n",
      "Epoch [   15/   20] | d_loss: 0.1304 | g_loss: 0.5864\n",
      "Epoch [   15/   20] | d_loss: 0.1558 | g_loss: 0.1431\n",
      "Epoch [   15/   20] | d_loss: 0.1643 | g_loss: 0.4595\n",
      "Epoch [   15/   20] | d_loss: 0.0673 | g_loss: 1.0946\n",
      "Epoch [   15/   20] | d_loss: 0.2102 | g_loss: 0.7599\n",
      "Epoch [   15/   20] | d_loss: 0.1047 | g_loss: 0.7773\n",
      "Epoch [   15/   20] | d_loss: 0.1537 | g_loss: 0.7542\n",
      "Epoch [   15/   20] | d_loss: 0.0631 | g_loss: 0.6393\n",
      "Epoch [   16/   20] | d_loss: 0.0799 | g_loss: 0.6332\n",
      "Epoch [   16/   20] | d_loss: 0.0466 | g_loss: 0.5335\n",
      "Epoch [   16/   20] | d_loss: 0.1387 | g_loss: 0.3855\n",
      "Epoch [   16/   20] | d_loss: 0.0829 | g_loss: 0.4983\n",
      "Epoch [   16/   20] | d_loss: 0.0722 | g_loss: 0.5046\n",
      "Epoch [   16/   20] | d_loss: 0.1173 | g_loss: 0.3456\n",
      "Epoch [   16/   20] | d_loss: 0.0561 | g_loss: 0.4662\n",
      "Epoch [   16/   20] | d_loss: 0.0695 | g_loss: 0.5464\n",
      "Epoch [   16/   20] | d_loss: 0.0563 | g_loss: 0.4720\n",
      "Epoch [   16/   20] | d_loss: 0.3684 | g_loss: 0.8605\n",
      "Epoch [   16/   20] | d_loss: 0.1489 | g_loss: 0.7852\n",
      "Epoch [   16/   20] | d_loss: 0.0897 | g_loss: 0.6197\n",
      "Epoch [   16/   20] | d_loss: 0.0647 | g_loss: 0.7502\n",
      "Epoch [   16/   20] | d_loss: 0.1892 | g_loss: 0.5788\n",
      "Epoch [   16/   20] | d_loss: 0.0613 | g_loss: 0.9838\n",
      "Epoch [   16/   20] | d_loss: 0.2418 | g_loss: 0.3389\n",
      "Epoch [   16/   20] | d_loss: 0.0958 | g_loss: 0.6337\n",
      "Epoch [   16/   20] | d_loss: 0.0857 | g_loss: 0.8155\n",
      "Epoch [   16/   20] | d_loss: 0.1054 | g_loss: 0.8601\n",
      "Epoch [   16/   20] | d_loss: 0.0501 | g_loss: 0.8380\n",
      "Epoch [   16/   20] | d_loss: 0.1237 | g_loss: 0.6013\n",
      "Epoch [   16/   20] | d_loss: 0.1662 | g_loss: 0.5815\n",
      "Epoch [   16/   20] | d_loss: 0.0888 | g_loss: 0.7877\n",
      "Epoch [   16/   20] | d_loss: 0.0658 | g_loss: 0.7610\n",
      "Epoch [   16/   20] | d_loss: 0.0669 | g_loss: 0.8468\n",
      "Epoch [   16/   20] | d_loss: 0.0905 | g_loss: 0.5314\n",
      "Epoch [   16/   20] | d_loss: 0.2086 | g_loss: 0.3828\n",
      "Epoch [   16/   20] | d_loss: 0.0864 | g_loss: 0.9317\n",
      "Epoch [   16/   20] | d_loss: 0.1350 | g_loss: 0.5227\n",
      "Epoch [   16/   20] | d_loss: 0.0593 | g_loss: 0.5515\n",
      "Epoch [   16/   20] | d_loss: 0.0768 | g_loss: 0.8570\n",
      "Epoch [   16/   20] | d_loss: 0.0914 | g_loss: 0.7104\n",
      "Epoch [   16/   20] | d_loss: 0.1895 | g_loss: 0.7300\n",
      "Epoch [   16/   20] | d_loss: 0.0992 | g_loss: 0.8433\n",
      "Epoch [   16/   20] | d_loss: 0.0979 | g_loss: 0.7816\n",
      "Epoch [   16/   20] | d_loss: 0.1947 | g_loss: 0.4769\n",
      "Epoch [   16/   20] | d_loss: 0.0547 | g_loss: 0.8777\n",
      "Epoch [   16/   20] | d_loss: 0.0384 | g_loss: 0.6262\n",
      "Epoch [   16/   20] | d_loss: 0.1805 | g_loss: 0.6216\n",
      "Epoch [   16/   20] | d_loss: 0.0665 | g_loss: 0.5345\n",
      "Epoch [   16/   20] | d_loss: 0.0966 | g_loss: 0.5990\n",
      "Epoch [   16/   20] | d_loss: 0.0495 | g_loss: 0.8997\n",
      "Epoch [   16/   20] | d_loss: 0.2119 | g_loss: 0.5653\n",
      "Epoch [   16/   20] | d_loss: 0.0337 | g_loss: 0.6929\n",
      "Epoch [   16/   20] | d_loss: 0.0649 | g_loss: 1.0181\n",
      "Epoch [   16/   20] | d_loss: 0.0248 | g_loss: 0.6852\n",
      "Epoch [   16/   20] | d_loss: 0.0797 | g_loss: 0.5057\n",
      "Epoch [   16/   20] | d_loss: 0.2061 | g_loss: 0.4018\n",
      "Epoch [   16/   20] | d_loss: 0.1069 | g_loss: 0.5080\n",
      "Epoch [   16/   20] | d_loss: 0.2476 | g_loss: 0.5986\n",
      "Epoch [   16/   20] | d_loss: 0.2003 | g_loss: 0.6324\n",
      "Epoch [   16/   20] | d_loss: 0.0645 | g_loss: 0.6254\n",
      "Epoch [   16/   20] | d_loss: 0.0708 | g_loss: 0.8889\n",
      "Epoch [   16/   20] | d_loss: 0.1022 | g_loss: 0.7136\n",
      "Epoch [   16/   20] | d_loss: 0.0518 | g_loss: 0.8248\n",
      "Epoch [   16/   20] | d_loss: 0.0363 | g_loss: 0.4231\n",
      "Epoch [   16/   20] | d_loss: 0.1514 | g_loss: 0.4657\n",
      "Epoch [   16/   20] | d_loss: 0.1352 | g_loss: 0.7468\n",
      "Epoch [   16/   20] | d_loss: 0.0865 | g_loss: 0.8039\n",
      "Epoch [   16/   20] | d_loss: 0.1245 | g_loss: 0.5848\n",
      "Epoch [   16/   20] | d_loss: 0.0680 | g_loss: 0.6243\n",
      "Epoch [   16/   20] | d_loss: 0.1454 | g_loss: 0.5757\n",
      "Epoch [   16/   20] | d_loss: 0.1127 | g_loss: 0.8093\n",
      "Epoch [   16/   20] | d_loss: 0.1347 | g_loss: 0.4495\n",
      "Epoch [   16/   20] | d_loss: 0.0939 | g_loss: 0.4897\n",
      "Epoch [   16/   20] | d_loss: 0.2184 | g_loss: 0.3576\n",
      "Epoch [   16/   20] | d_loss: 0.2661 | g_loss: 0.5175\n",
      "Epoch [   16/   20] | d_loss: 0.1874 | g_loss: 0.5583\n",
      "Epoch [   16/   20] | d_loss: 0.0956 | g_loss: 0.4908\n",
      "Epoch [   16/   20] | d_loss: 0.0626 | g_loss: 0.4071\n",
      "Epoch [   16/   20] | d_loss: 0.1188 | g_loss: 0.4202\n",
      "Epoch [   16/   20] | d_loss: 0.2459 | g_loss: 0.9406\n",
      "Epoch [   16/   20] | d_loss: 0.1597 | g_loss: 0.3357\n",
      "Epoch [   16/   20] | d_loss: 0.1123 | g_loss: 0.7544\n",
      "Epoch [   16/   20] | d_loss: 0.2103 | g_loss: 0.4957\n",
      "Epoch [   16/   20] | d_loss: 0.0868 | g_loss: 0.6633\n",
      "Epoch [   16/   20] | d_loss: 0.0843 | g_loss: 0.6107\n",
      "Epoch [   16/   20] | d_loss: 0.0731 | g_loss: 0.9325\n",
      "Epoch [   16/   20] | d_loss: 0.0706 | g_loss: 0.6343\n",
      "Epoch [   16/   20] | d_loss: 0.0876 | g_loss: 0.7303\n",
      "Epoch [   16/   20] | d_loss: 0.1574 | g_loss: 0.6961\n",
      "Epoch [   16/   20] | d_loss: 0.5539 | g_loss: 0.4058\n",
      "Epoch [   16/   20] | d_loss: 0.0734 | g_loss: 0.4068\n",
      "Epoch [   16/   20] | d_loss: 0.1638 | g_loss: 0.5538\n",
      "Epoch [   16/   20] | d_loss: 0.0684 | g_loss: 0.9393\n",
      "Epoch [   16/   20] | d_loss: 0.1907 | g_loss: 0.6513\n",
      "Epoch [   16/   20] | d_loss: 0.0731 | g_loss: 0.3429\n",
      "Epoch [   16/   20] | d_loss: 0.1713 | g_loss: 0.4165\n",
      "Epoch [   16/   20] | d_loss: 0.0487 | g_loss: 0.5798\n",
      "Epoch [   16/   20] | d_loss: 0.1567 | g_loss: 0.6236\n",
      "Epoch [   17/   20] | d_loss: 0.0623 | g_loss: 0.6391\n",
      "Epoch [   17/   20] | d_loss: 0.1882 | g_loss: 0.6172\n",
      "Epoch [   17/   20] | d_loss: 0.0674 | g_loss: 0.6096\n",
      "Epoch [   17/   20] | d_loss: 0.0833 | g_loss: 0.5348\n",
      "Epoch [   17/   20] | d_loss: 0.0550 | g_loss: 0.6576\n",
      "Epoch [   17/   20] | d_loss: 0.1211 | g_loss: 0.6017\n",
      "Epoch [   17/   20] | d_loss: 0.0803 | g_loss: 0.8931\n",
      "Epoch [   17/   20] | d_loss: 0.1359 | g_loss: 0.3255\n",
      "Epoch [   17/   20] | d_loss: 0.0614 | g_loss: 0.6705\n",
      "Epoch [   17/   20] | d_loss: 0.0728 | g_loss: 0.6321\n",
      "Epoch [   17/   20] | d_loss: 0.1644 | g_loss: 0.4343\n",
      "Epoch [   17/   20] | d_loss: 0.0873 | g_loss: 0.7881\n",
      "Epoch [   17/   20] | d_loss: 0.0740 | g_loss: 0.9931\n",
      "Epoch [   17/   20] | d_loss: 0.1553 | g_loss: 0.6640\n",
      "Epoch [   17/   20] | d_loss: 0.1410 | g_loss: 0.4212\n",
      "Epoch [   17/   20] | d_loss: 0.1830 | g_loss: 0.4740\n",
      "Epoch [   17/   20] | d_loss: 0.0895 | g_loss: 0.7204\n",
      "Epoch [   17/   20] | d_loss: 0.1568 | g_loss: 0.5903\n",
      "Epoch [   17/   20] | d_loss: 0.0761 | g_loss: 0.5754\n",
      "Epoch [   17/   20] | d_loss: 0.0868 | g_loss: 0.6300\n",
      "Epoch [   17/   20] | d_loss: 0.0671 | g_loss: 0.5893\n",
      "Epoch [   17/   20] | d_loss: 0.0804 | g_loss: 0.6921\n",
      "Epoch [   17/   20] | d_loss: 0.1243 | g_loss: 0.9331\n",
      "Epoch [   17/   20] | d_loss: 0.0376 | g_loss: 0.8828\n",
      "Epoch [   17/   20] | d_loss: 0.1292 | g_loss: 0.8645\n",
      "Epoch [   17/   20] | d_loss: 0.0916 | g_loss: 0.6882\n",
      "Epoch [   17/   20] | d_loss: 0.1495 | g_loss: 0.7129\n",
      "Epoch [   17/   20] | d_loss: 0.0650 | g_loss: 0.7155\n",
      "Epoch [   17/   20] | d_loss: 0.0695 | g_loss: 0.4177\n",
      "Epoch [   17/   20] | d_loss: 0.0635 | g_loss: 0.4409\n",
      "Epoch [   17/   20] | d_loss: 0.0596 | g_loss: 0.5648\n",
      "Epoch [   17/   20] | d_loss: 0.2014 | g_loss: 0.5440\n",
      "Epoch [   17/   20] | d_loss: 0.0932 | g_loss: 0.6050\n",
      "Epoch [   17/   20] | d_loss: 0.2305 | g_loss: 0.4772\n",
      "Epoch [   17/   20] | d_loss: 0.2021 | g_loss: 0.3146\n",
      "Epoch [   17/   20] | d_loss: 0.0378 | g_loss: 0.5599\n",
      "Epoch [   17/   20] | d_loss: 0.0387 | g_loss: 0.3342\n",
      "Epoch [   17/   20] | d_loss: 0.0707 | g_loss: 0.9766\n",
      "Epoch [   17/   20] | d_loss: 0.0747 | g_loss: 0.4703\n",
      "Epoch [   17/   20] | d_loss: 0.1731 | g_loss: 0.6250\n",
      "Epoch [   17/   20] | d_loss: 0.0744 | g_loss: 0.6258\n",
      "Epoch [   17/   20] | d_loss: 0.0589 | g_loss: 0.7282\n",
      "Epoch [   17/   20] | d_loss: 0.2438 | g_loss: 0.4978\n",
      "Epoch [   17/   20] | d_loss: 0.0906 | g_loss: 0.5228\n",
      "Epoch [   17/   20] | d_loss: 0.3107 | g_loss: 0.5122\n",
      "Epoch [   17/   20] | d_loss: 0.1274 | g_loss: 0.5279\n",
      "Epoch [   17/   20] | d_loss: 0.0917 | g_loss: 1.1187\n",
      "Epoch [   17/   20] | d_loss: 0.4059 | g_loss: 0.3602\n",
      "Epoch [   17/   20] | d_loss: 0.1425 | g_loss: 0.4214\n",
      "Epoch [   17/   20] | d_loss: 0.1634 | g_loss: 0.5044\n",
      "Epoch [   17/   20] | d_loss: 0.1176 | g_loss: 0.5857\n",
      "Epoch [   17/   20] | d_loss: 0.0646 | g_loss: 0.6240\n",
      "Epoch [   17/   20] | d_loss: 0.2024 | g_loss: 0.5211\n",
      "Epoch [   17/   20] | d_loss: 0.0509 | g_loss: 0.7970\n",
      "Epoch [   17/   20] | d_loss: 0.1732 | g_loss: 0.5051\n",
      "Epoch [   17/   20] | d_loss: 0.0721 | g_loss: 0.7334\n",
      "Epoch [   17/   20] | d_loss: 0.1202 | g_loss: 0.6868\n",
      "Epoch [   17/   20] | d_loss: 0.2994 | g_loss: 1.1553\n",
      "Epoch [   17/   20] | d_loss: 0.3136 | g_loss: 0.4833\n",
      "Epoch [   17/   20] | d_loss: 0.1043 | g_loss: 0.1945\n",
      "Epoch [   17/   20] | d_loss: 0.0922 | g_loss: 0.5896\n",
      "Epoch [   17/   20] | d_loss: 0.1794 | g_loss: 0.2523\n",
      "Epoch [   17/   20] | d_loss: 0.0817 | g_loss: 1.0061\n",
      "Epoch [   17/   20] | d_loss: 0.0841 | g_loss: 0.7436\n",
      "Epoch [   17/   20] | d_loss: 0.1337 | g_loss: 0.6015\n",
      "Epoch [   17/   20] | d_loss: 0.1904 | g_loss: 0.5291\n",
      "Epoch [   17/   20] | d_loss: 0.1017 | g_loss: 0.3173\n",
      "Epoch [   17/   20] | d_loss: 0.0594 | g_loss: 0.6935\n",
      "Epoch [   17/   20] | d_loss: 0.1024 | g_loss: 0.6756\n",
      "Epoch [   17/   20] | d_loss: 0.2278 | g_loss: 0.2241\n",
      "Epoch [   17/   20] | d_loss: 0.1087 | g_loss: 0.7821\n",
      "Epoch [   17/   20] | d_loss: 0.0950 | g_loss: 0.7142\n",
      "Epoch [   17/   20] | d_loss: 0.0927 | g_loss: 0.5502\n",
      "Epoch [   17/   20] | d_loss: 0.1094 | g_loss: 0.7467\n",
      "Epoch [   17/   20] | d_loss: 0.0658 | g_loss: 0.7385\n",
      "Epoch [   17/   20] | d_loss: 0.1559 | g_loss: 0.7944\n",
      "Epoch [   17/   20] | d_loss: 0.1040 | g_loss: 0.5931\n",
      "Epoch [   17/   20] | d_loss: 0.0498 | g_loss: 0.5946\n",
      "Epoch [   17/   20] | d_loss: 0.0513 | g_loss: 0.4291\n",
      "Epoch [   17/   20] | d_loss: 0.0867 | g_loss: 0.9445\n",
      "Epoch [   17/   20] | d_loss: 0.0850 | g_loss: 0.7701\n",
      "Epoch [   17/   20] | d_loss: 0.1753 | g_loss: 0.3754\n",
      "Epoch [   17/   20] | d_loss: 0.1196 | g_loss: 0.7229\n",
      "Epoch [   17/   20] | d_loss: 0.1105 | g_loss: 0.3771\n",
      "Epoch [   17/   20] | d_loss: 0.1155 | g_loss: 0.5812\n",
      "Epoch [   17/   20] | d_loss: 0.1292 | g_loss: 0.6337\n",
      "Epoch [   17/   20] | d_loss: 0.0544 | g_loss: 0.8690\n",
      "Epoch [   17/   20] | d_loss: 0.1615 | g_loss: 0.7411\n",
      "Epoch [   17/   20] | d_loss: 0.1230 | g_loss: 0.5029\n",
      "Epoch [   17/   20] | d_loss: 0.0434 | g_loss: 0.7851\n",
      "Epoch [   18/   20] | d_loss: 0.1288 | g_loss: 0.4225\n",
      "Epoch [   18/   20] | d_loss: 0.0917 | g_loss: 0.6905\n",
      "Epoch [   18/   20] | d_loss: 0.2003 | g_loss: 0.7261\n",
      "Epoch [   18/   20] | d_loss: 0.1118 | g_loss: 0.4857\n",
      "Epoch [   18/   20] | d_loss: 0.1161 | g_loss: 0.2856\n",
      "Epoch [   18/   20] | d_loss: 0.1040 | g_loss: 0.8246\n",
      "Epoch [   18/   20] | d_loss: 0.1400 | g_loss: 0.4277\n",
      "Epoch [   18/   20] | d_loss: 0.0371 | g_loss: 0.8342\n",
      "Epoch [   18/   20] | d_loss: 0.0862 | g_loss: 1.0221\n",
      "Epoch [   18/   20] | d_loss: 0.0825 | g_loss: 0.8224\n",
      "Epoch [   18/   20] | d_loss: 0.1929 | g_loss: 0.5800\n",
      "Epoch [   18/   20] | d_loss: 0.0856 | g_loss: 0.5429\n",
      "Epoch [   18/   20] | d_loss: 0.1211 | g_loss: 0.7071\n",
      "Epoch [   18/   20] | d_loss: 0.1885 | g_loss: 0.9349\n",
      "Epoch [   18/   20] | d_loss: 0.1364 | g_loss: 0.6888\n",
      "Epoch [   18/   20] | d_loss: 0.1840 | g_loss: 0.4499\n",
      "Epoch [   18/   20] | d_loss: 0.1259 | g_loss: 1.2726\n",
      "Epoch [   18/   20] | d_loss: 0.1676 | g_loss: 0.8574\n",
      "Epoch [   18/   20] | d_loss: 0.0431 | g_loss: 0.6708\n",
      "Epoch [   18/   20] | d_loss: 0.0642 | g_loss: 0.5170\n",
      "Epoch [   18/   20] | d_loss: 0.1040 | g_loss: 0.8319\n",
      "Epoch [   18/   20] | d_loss: 0.1085 | g_loss: 0.5929\n",
      "Epoch [   18/   20] | d_loss: 0.0959 | g_loss: 0.5682\n",
      "Epoch [   18/   20] | d_loss: 0.0434 | g_loss: 0.8292\n",
      "Epoch [   18/   20] | d_loss: 0.0767 | g_loss: 0.5447\n",
      "Epoch [   18/   20] | d_loss: 0.0918 | g_loss: 0.6925\n",
      "Epoch [   18/   20] | d_loss: 0.1870 | g_loss: 0.4850\n",
      "Epoch [   18/   20] | d_loss: 0.1560 | g_loss: 0.3580\n",
      "Epoch [   18/   20] | d_loss: 0.0666 | g_loss: 0.8625\n",
      "Epoch [   18/   20] | d_loss: 0.1736 | g_loss: 0.4247\n",
      "Epoch [   18/   20] | d_loss: 0.1068 | g_loss: 0.4330\n",
      "Epoch [   18/   20] | d_loss: 0.0342 | g_loss: 0.8473\n",
      "Epoch [   18/   20] | d_loss: 0.0363 | g_loss: 1.0081\n",
      "Epoch [   18/   20] | d_loss: 0.1552 | g_loss: 0.3126\n",
      "Epoch [   18/   20] | d_loss: 0.0631 | g_loss: 0.6359\n",
      "Epoch [   18/   20] | d_loss: 0.2907 | g_loss: 0.5055\n",
      "Epoch [   18/   20] | d_loss: 0.2428 | g_loss: 0.6786\n",
      "Epoch [   18/   20] | d_loss: 0.0973 | g_loss: 0.8788\n",
      "Epoch [   18/   20] | d_loss: 0.1219 | g_loss: 0.5692\n",
      "Epoch [   18/   20] | d_loss: 0.2746 | g_loss: 0.7197\n",
      "Epoch [   18/   20] | d_loss: 0.1379 | g_loss: 0.6229\n",
      "Epoch [   18/   20] | d_loss: 0.1114 | g_loss: 0.7497\n",
      "Epoch [   18/   20] | d_loss: 0.1627 | g_loss: 0.6730\n",
      "Epoch [   18/   20] | d_loss: 0.1785 | g_loss: 0.7081\n",
      "Epoch [   18/   20] | d_loss: 0.1427 | g_loss: 0.4925\n",
      "Epoch [   18/   20] | d_loss: 0.1951 | g_loss: 0.9050\n",
      "Epoch [   18/   20] | d_loss: 0.0695 | g_loss: 0.6133\n",
      "Epoch [   18/   20] | d_loss: 0.0947 | g_loss: 0.5382\n",
      "Epoch [   18/   20] | d_loss: 0.0687 | g_loss: 1.0236\n",
      "Epoch [   18/   20] | d_loss: 0.0692 | g_loss: 0.5278\n",
      "Epoch [   18/   20] | d_loss: 0.0643 | g_loss: 0.8891\n",
      "Epoch [   18/   20] | d_loss: 0.0440 | g_loss: 0.6269\n",
      "Epoch [   18/   20] | d_loss: 0.0888 | g_loss: 0.2823\n",
      "Epoch [   18/   20] | d_loss: 0.1348 | g_loss: 0.5761\n",
      "Epoch [   18/   20] | d_loss: 0.1106 | g_loss: 0.4606\n",
      "Epoch [   18/   20] | d_loss: 0.0799 | g_loss: 0.5838\n",
      "Epoch [   18/   20] | d_loss: 0.1463 | g_loss: 0.5775\n",
      "Epoch [   18/   20] | d_loss: 0.0588 | g_loss: 0.4597\n",
      "Epoch [   18/   20] | d_loss: 0.2156 | g_loss: 0.4837\n",
      "Epoch [   18/   20] | d_loss: 0.0364 | g_loss: 0.9437\n",
      "Epoch [   18/   20] | d_loss: 0.1520 | g_loss: 0.6348\n",
      "Epoch [   18/   20] | d_loss: 0.1665 | g_loss: 0.4439\n",
      "Epoch [   18/   20] | d_loss: 0.1173 | g_loss: 0.7331\n",
      "Epoch [   18/   20] | d_loss: 0.1051 | g_loss: 0.7531\n",
      "Epoch [   18/   20] | d_loss: 0.0575 | g_loss: 0.2691\n",
      "Epoch [   18/   20] | d_loss: 0.0871 | g_loss: 1.0945\n",
      "Epoch [   18/   20] | d_loss: 0.1566 | g_loss: 0.7421\n",
      "Epoch [   18/   20] | d_loss: 0.0888 | g_loss: 1.1475\n",
      "Epoch [   18/   20] | d_loss: 0.1176 | g_loss: 0.9295\n",
      "Epoch [   18/   20] | d_loss: 0.1888 | g_loss: 0.6595\n",
      "Epoch [   18/   20] | d_loss: 0.0796 | g_loss: 0.6299\n",
      "Epoch [   18/   20] | d_loss: 0.0566 | g_loss: 0.5003\n",
      "Epoch [   18/   20] | d_loss: 0.0964 | g_loss: 0.5577\n",
      "Epoch [   18/   20] | d_loss: 0.1809 | g_loss: 0.5895\n",
      "Epoch [   18/   20] | d_loss: 0.1856 | g_loss: 0.6005\n",
      "Epoch [   18/   20] | d_loss: 0.1353 | g_loss: 0.5516\n",
      "Epoch [   18/   20] | d_loss: 0.2225 | g_loss: 0.3173\n",
      "Epoch [   18/   20] | d_loss: 0.1874 | g_loss: 0.5255\n",
      "Epoch [   18/   20] | d_loss: 0.0598 | g_loss: 0.7255\n",
      "Epoch [   18/   20] | d_loss: 0.0973 | g_loss: 0.4730\n",
      "Epoch [   18/   20] | d_loss: 0.1327 | g_loss: 0.8273\n",
      "Epoch [   18/   20] | d_loss: 0.1866 | g_loss: 0.5715\n",
      "Epoch [   18/   20] | d_loss: 0.0863 | g_loss: 0.5200\n",
      "Epoch [   18/   20] | d_loss: 0.1278 | g_loss: 0.7869\n",
      "Epoch [   18/   20] | d_loss: 0.1570 | g_loss: 0.6535\n",
      "Epoch [   18/   20] | d_loss: 0.1426 | g_loss: 0.5281\n",
      "Epoch [   18/   20] | d_loss: 0.1283 | g_loss: 0.4792\n",
      "Epoch [   18/   20] | d_loss: 0.2792 | g_loss: 0.5718\n",
      "Epoch [   18/   20] | d_loss: 0.1569 | g_loss: 0.6273\n",
      "Epoch [   18/   20] | d_loss: 0.1711 | g_loss: 0.4563\n",
      "Epoch [   19/   20] | d_loss: 0.1088 | g_loss: 0.4684\n",
      "Epoch [   19/   20] | d_loss: 0.0910 | g_loss: 0.3759\n",
      "Epoch [   19/   20] | d_loss: 0.0698 | g_loss: 0.8871\n",
      "Epoch [   19/   20] | d_loss: 0.2391 | g_loss: 0.7513\n",
      "Epoch [   19/   20] | d_loss: 0.1690 | g_loss: 0.7806\n",
      "Epoch [   19/   20] | d_loss: 0.1232 | g_loss: 1.0117\n",
      "Epoch [   19/   20] | d_loss: 0.1218 | g_loss: 0.6869\n",
      "Epoch [   19/   20] | d_loss: 0.1342 | g_loss: 0.5149\n",
      "Epoch [   19/   20] | d_loss: 0.1949 | g_loss: 0.6602\n",
      "Epoch [   19/   20] | d_loss: 0.0614 | g_loss: 0.4095\n",
      "Epoch [   19/   20] | d_loss: 0.1190 | g_loss: 0.8523\n",
      "Epoch [   19/   20] | d_loss: 0.0637 | g_loss: 0.9370\n",
      "Epoch [   19/   20] | d_loss: 0.0637 | g_loss: 0.5204\n",
      "Epoch [   19/   20] | d_loss: 0.1350 | g_loss: 0.5345\n",
      "Epoch [   19/   20] | d_loss: 0.0990 | g_loss: 0.7371\n",
      "Epoch [   19/   20] | d_loss: 0.1928 | g_loss: 0.5915\n",
      "Epoch [   19/   20] | d_loss: 0.1423 | g_loss: 0.4874\n",
      "Epoch [   19/   20] | d_loss: 0.0959 | g_loss: 0.2206\n",
      "Epoch [   19/   20] | d_loss: 0.0807 | g_loss: 0.7791\n",
      "Epoch [   19/   20] | d_loss: 0.2787 | g_loss: 0.5849\n",
      "Epoch [   19/   20] | d_loss: 0.1250 | g_loss: 0.8314\n",
      "Epoch [   19/   20] | d_loss: 0.0842 | g_loss: 0.5571\n",
      "Epoch [   19/   20] | d_loss: 0.0882 | g_loss: 0.6616\n",
      "Epoch [   19/   20] | d_loss: 0.1260 | g_loss: 0.9908\n",
      "Epoch [   19/   20] | d_loss: 0.0793 | g_loss: 0.6572\n",
      "Epoch [   19/   20] | d_loss: 0.2082 | g_loss: 0.6365\n",
      "Epoch [   19/   20] | d_loss: 0.0625 | g_loss: 0.8851\n",
      "Epoch [   19/   20] | d_loss: 0.0656 | g_loss: 0.6711\n",
      "Epoch [   19/   20] | d_loss: 0.0803 | g_loss: 0.9139\n",
      "Epoch [   19/   20] | d_loss: 0.0708 | g_loss: 0.6123\n",
      "Epoch [   19/   20] | d_loss: 0.1638 | g_loss: 0.4742\n",
      "Epoch [   19/   20] | d_loss: 0.0542 | g_loss: 0.7452\n",
      "Epoch [   19/   20] | d_loss: 0.0600 | g_loss: 0.5757\n",
      "Epoch [   19/   20] | d_loss: 0.2256 | g_loss: 1.0756\n",
      "Epoch [   19/   20] | d_loss: 0.3704 | g_loss: 0.3811\n",
      "Epoch [   19/   20] | d_loss: 0.2431 | g_loss: 0.9169\n",
      "Epoch [   19/   20] | d_loss: 0.0791 | g_loss: 0.6660\n",
      "Epoch [   19/   20] | d_loss: 0.0508 | g_loss: 0.5025\n",
      "Epoch [   19/   20] | d_loss: 0.0771 | g_loss: 0.5975\n",
      "Epoch [   19/   20] | d_loss: 0.1290 | g_loss: 0.5796\n",
      "Epoch [   19/   20] | d_loss: 0.1720 | g_loss: 0.5208\n",
      "Epoch [   19/   20] | d_loss: 0.1009 | g_loss: 0.6612\n",
      "Epoch [   19/   20] | d_loss: 0.0521 | g_loss: 0.5297\n",
      "Epoch [   19/   20] | d_loss: 0.0643 | g_loss: 0.5147\n",
      "Epoch [   19/   20] | d_loss: 0.1120 | g_loss: 0.6140\n",
      "Epoch [   19/   20] | d_loss: 0.1215 | g_loss: 0.5902\n",
      "Epoch [   19/   20] | d_loss: 0.1761 | g_loss: 0.7430\n",
      "Epoch [   19/   20] | d_loss: 0.1095 | g_loss: 0.6096\n",
      "Epoch [   19/   20] | d_loss: 0.0871 | g_loss: 0.7665\n",
      "Epoch [   19/   20] | d_loss: 0.0649 | g_loss: 0.5188\n",
      "Epoch [   19/   20] | d_loss: 0.0450 | g_loss: 0.7826\n",
      "Epoch [   19/   20] | d_loss: 0.0801 | g_loss: 0.6745\n",
      "Epoch [   19/   20] | d_loss: 0.0556 | g_loss: 0.5138\n",
      "Epoch [   19/   20] | d_loss: 0.0967 | g_loss: 0.7426\n",
      "Epoch [   19/   20] | d_loss: 0.1232 | g_loss: 0.4044\n",
      "Epoch [   19/   20] | d_loss: 0.1179 | g_loss: 0.4617\n",
      "Epoch [   19/   20] | d_loss: 0.0446 | g_loss: 1.0139\n",
      "Epoch [   19/   20] | d_loss: 0.0959 | g_loss: 0.4913\n",
      "Epoch [   19/   20] | d_loss: 0.1122 | g_loss: 0.6750\n",
      "Epoch [   19/   20] | d_loss: 0.1475 | g_loss: 0.7808\n",
      "Epoch [   19/   20] | d_loss: 0.1117 | g_loss: 0.5921\n",
      "Epoch [   19/   20] | d_loss: 0.0552 | g_loss: 0.7186\n",
      "Epoch [   19/   20] | d_loss: 0.0965 | g_loss: 0.4194\n",
      "Epoch [   19/   20] | d_loss: 0.0516 | g_loss: 0.6838\n",
      "Epoch [   19/   20] | d_loss: 0.1148 | g_loss: 0.8333\n",
      "Epoch [   19/   20] | d_loss: 0.1034 | g_loss: 0.7760\n",
      "Epoch [   19/   20] | d_loss: 0.2307 | g_loss: 0.5441\n",
      "Epoch [   19/   20] | d_loss: 0.1613 | g_loss: 0.7141\n",
      "Epoch [   19/   20] | d_loss: 0.0937 | g_loss: 0.4660\n",
      "Epoch [   19/   20] | d_loss: 0.1090 | g_loss: 0.7719\n",
      "Epoch [   19/   20] | d_loss: 0.0828 | g_loss: 0.7666\n",
      "Epoch [   19/   20] | d_loss: 0.0451 | g_loss: 0.7985\n",
      "Epoch [   19/   20] | d_loss: 0.1580 | g_loss: 1.1343\n",
      "Epoch [   19/   20] | d_loss: 0.0965 | g_loss: 0.6620\n",
      "Epoch [   19/   20] | d_loss: 0.1513 | g_loss: 0.6336\n",
      "Epoch [   19/   20] | d_loss: 0.1561 | g_loss: 0.4799\n",
      "Epoch [   19/   20] | d_loss: 0.0882 | g_loss: 0.8494\n",
      "Epoch [   19/   20] | d_loss: 0.0651 | g_loss: 0.9079\n",
      "Epoch [   19/   20] | d_loss: 0.0773 | g_loss: 0.5934\n",
      "Epoch [   19/   20] | d_loss: 0.0415 | g_loss: 0.9997\n",
      "Epoch [   19/   20] | d_loss: 0.2534 | g_loss: 0.5563\n",
      "Epoch [   19/   20] | d_loss: 0.0837 | g_loss: 0.5892\n",
      "Epoch [   19/   20] | d_loss: 0.1009 | g_loss: 0.6674\n",
      "Epoch [   19/   20] | d_loss: 0.0292 | g_loss: 0.7847\n",
      "Epoch [   19/   20] | d_loss: 0.0732 | g_loss: 0.6625\n",
      "Epoch [   19/   20] | d_loss: 0.1344 | g_loss: 0.6705\n",
      "Epoch [   19/   20] | d_loss: 0.0523 | g_loss: 0.7227\n",
      "Epoch [   19/   20] | d_loss: 0.1067 | g_loss: 0.4220\n",
      "Epoch [   19/   20] | d_loss: 0.0592 | g_loss: 1.1003\n",
      "Epoch [   19/   20] | d_loss: 0.2171 | g_loss: 0.5966\n",
      "Epoch [   20/   20] | d_loss: 0.0670 | g_loss: 0.7597\n",
      "Epoch [   20/   20] | d_loss: 0.0456 | g_loss: 0.5944\n",
      "Epoch [   20/   20] | d_loss: 0.1051 | g_loss: 1.2001\n",
      "Epoch [   20/   20] | d_loss: 0.1391 | g_loss: 0.4510\n",
      "Epoch [   20/   20] | d_loss: 0.0893 | g_loss: 0.6854\n",
      "Epoch [   20/   20] | d_loss: 0.0373 | g_loss: 0.7994\n",
      "Epoch [   20/   20] | d_loss: 0.0863 | g_loss: 0.7343\n",
      "Epoch [   20/   20] | d_loss: 0.0551 | g_loss: 0.5489\n",
      "Epoch [   20/   20] | d_loss: 0.0919 | g_loss: 0.7934\n",
      "Epoch [   20/   20] | d_loss: 0.0912 | g_loss: 0.7262\n",
      "Epoch [   20/   20] | d_loss: 0.0920 | g_loss: 0.7514\n",
      "Epoch [   20/   20] | d_loss: 0.1946 | g_loss: 0.4612\n",
      "Epoch [   20/   20] | d_loss: 0.0965 | g_loss: 0.7669\n",
      "Epoch [   20/   20] | d_loss: 0.1050 | g_loss: 0.7833\n",
      "Epoch [   20/   20] | d_loss: 0.1583 | g_loss: 0.4964\n",
      "Epoch [   20/   20] | d_loss: 0.1338 | g_loss: 0.4706\n",
      "Epoch [   20/   20] | d_loss: 0.0684 | g_loss: 0.3286\n",
      "Epoch [   20/   20] | d_loss: 0.2545 | g_loss: 0.4005\n",
      "Epoch [   20/   20] | d_loss: 0.2041 | g_loss: 0.4131\n",
      "Epoch [   20/   20] | d_loss: 0.0565 | g_loss: 0.6034\n",
      "Epoch [   20/   20] | d_loss: 0.1102 | g_loss: 0.7290\n",
      "Epoch [   20/   20] | d_loss: 0.0540 | g_loss: 0.4611\n",
      "Epoch [   20/   20] | d_loss: 0.0817 | g_loss: 0.4950\n",
      "Epoch [   20/   20] | d_loss: 0.0521 | g_loss: 0.7024\n",
      "Epoch [   20/   20] | d_loss: 0.1296 | g_loss: 0.7671\n",
      "Epoch [   20/   20] | d_loss: 0.2067 | g_loss: 0.2016\n",
      "Epoch [   20/   20] | d_loss: 0.1056 | g_loss: 0.6654\n",
      "Epoch [   20/   20] | d_loss: 0.0562 | g_loss: 0.7711\n",
      "Epoch [   20/   20] | d_loss: 0.1016 | g_loss: 0.6672\n",
      "Epoch [   20/   20] | d_loss: 0.0764 | g_loss: 0.7488\n",
      "Epoch [   20/   20] | d_loss: 0.1742 | g_loss: 0.4546\n",
      "Epoch [   20/   20] | d_loss: 0.1681 | g_loss: 0.6070\n",
      "Epoch [   20/   20] | d_loss: 0.2189 | g_loss: 0.4201\n",
      "Epoch [   20/   20] | d_loss: 0.2130 | g_loss: 0.8567\n",
      "Epoch [   20/   20] | d_loss: 0.1407 | g_loss: 0.8455\n",
      "Epoch [   20/   20] | d_loss: 0.1165 | g_loss: 0.6162\n",
      "Epoch [   20/   20] | d_loss: 0.1035 | g_loss: 1.0042\n",
      "Epoch [   20/   20] | d_loss: 0.2402 | g_loss: 0.5118\n",
      "Epoch [   20/   20] | d_loss: 0.0819 | g_loss: 0.6730\n",
      "Epoch [   20/   20] | d_loss: 0.0668 | g_loss: 0.6897\n",
      "Epoch [   20/   20] | d_loss: 0.1325 | g_loss: 0.5561\n",
      "Epoch [   20/   20] | d_loss: 0.1386 | g_loss: 0.6261\n",
      "Epoch [   20/   20] | d_loss: 0.0759 | g_loss: 0.4483\n",
      "Epoch [   20/   20] | d_loss: 0.1289 | g_loss: 0.3695\n",
      "Epoch [   20/   20] | d_loss: 0.0930 | g_loss: 0.6232\n",
      "Epoch [   20/   20] | d_loss: 0.0921 | g_loss: 0.6064\n",
      "Epoch [   20/   20] | d_loss: 0.0922 | g_loss: 0.6009\n",
      "Epoch [   20/   20] | d_loss: 0.1828 | g_loss: 0.6361\n",
      "Epoch [   20/   20] | d_loss: 0.1432 | g_loss: 0.5876\n",
      "Epoch [   20/   20] | d_loss: 0.0530 | g_loss: 0.9270\n",
      "Epoch [   20/   20] | d_loss: 0.0717 | g_loss: 0.7498\n",
      "Epoch [   20/   20] | d_loss: 0.0956 | g_loss: 0.8567\n",
      "Epoch [   20/   20] | d_loss: 0.0931 | g_loss: 0.5158\n",
      "Epoch [   20/   20] | d_loss: 0.2473 | g_loss: 0.8262\n",
      "Epoch [   20/   20] | d_loss: 0.0711 | g_loss: 0.6588\n",
      "Epoch [   20/   20] | d_loss: 0.1260 | g_loss: 0.6174\n",
      "Epoch [   20/   20] | d_loss: 0.1052 | g_loss: 0.7688\n",
      "Epoch [   20/   20] | d_loss: 0.1311 | g_loss: 0.4173\n",
      "Epoch [   20/   20] | d_loss: 0.0478 | g_loss: 0.7507\n",
      "Epoch [   20/   20] | d_loss: 0.0926 | g_loss: 0.6190\n",
      "Epoch [   20/   20] | d_loss: 0.1537 | g_loss: 0.6735\n",
      "Epoch [   20/   20] | d_loss: 0.0850 | g_loss: 0.4030\n",
      "Epoch [   20/   20] | d_loss: 0.1133 | g_loss: 0.4191\n",
      "Epoch [   20/   20] | d_loss: 0.1880 | g_loss: 0.5811\n",
      "Epoch [   20/   20] | d_loss: 0.1628 | g_loss: 0.6201\n",
      "Epoch [   20/   20] | d_loss: 0.0698 | g_loss: 0.6908\n",
      "Epoch [   20/   20] | d_loss: 0.0871 | g_loss: 0.4610\n",
      "Epoch [   20/   20] | d_loss: 0.0427 | g_loss: 1.0295\n",
      "Epoch [   20/   20] | d_loss: 0.1767 | g_loss: 0.4471\n",
      "Epoch [   20/   20] | d_loss: 0.2097 | g_loss: 0.4389\n",
      "Epoch [   20/   20] | d_loss: 0.0528 | g_loss: 1.0143\n",
      "Epoch [   20/   20] | d_loss: 0.1895 | g_loss: 0.4744\n",
      "Epoch [   20/   20] | d_loss: 0.2292 | g_loss: 0.6413\n",
      "Epoch [   20/   20] | d_loss: 0.1041 | g_loss: 0.4773\n",
      "Epoch [   20/   20] | d_loss: 0.0418 | g_loss: 0.8311\n",
      "Epoch [   20/   20] | d_loss: 0.0925 | g_loss: 0.7987\n",
      "Epoch [   20/   20] | d_loss: 0.2354 | g_loss: 0.5991\n",
      "Epoch [   20/   20] | d_loss: 0.1311 | g_loss: 0.4820\n",
      "Epoch [   20/   20] | d_loss: 0.1019 | g_loss: 0.4249\n",
      "Epoch [   20/   20] | d_loss: 0.0954 | g_loss: 1.0508\n",
      "Epoch [   20/   20] | d_loss: 0.1843 | g_loss: 0.7806\n",
      "Epoch [   20/   20] | d_loss: 0.0739 | g_loss: 0.6237\n",
      "Epoch [   20/   20] | d_loss: 0.1316 | g_loss: 0.5555\n",
      "Epoch [   20/   20] | d_loss: 0.1636 | g_loss: 0.6354\n",
      "Epoch [   20/   20] | d_loss: 0.1110 | g_loss: 1.0556\n",
      "Epoch [   20/   20] | d_loss: 0.0950 | g_loss: 0.6411\n",
      "Epoch [   20/   20] | d_loss: 0.1673 | g_loss: 0.4506\n",
      "Epoch [   20/   20] | d_loss: 0.2331 | g_loss: 0.5166\n",
      "Epoch [   20/   20] | d_loss: 0.1655 | g_loss: 0.6668\n",
      "Epoch [   20/   20] | d_loss: 0.1090 | g_loss: 0.5045\n"
     ]
    }
   ],
   "source": [
    "# set number of epochs \n",
    "n_epochs = 20\n",
    "\n",
    "\n",
    "\"\"\"\n",
    "DON'T MODIFY ANYTHING IN THIS CELL\n",
    "\"\"\"\n",
    "# call training function\n",
    "losses = train(D, G, n_epochs=n_epochs)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Training loss\n",
    "\n",
    "Plot the training losses for the generator and discriminator, recorded after each epoch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fb86c4a65c0>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb86c69f198>"
      ]
     },
     "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": 30,
   "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": 31,
   "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": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb86c4e0da0>"
      ]
     },
     "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",
    "\n",
    "* Diversity in ethnicity would definitly imporove results as more facial features can be introduced.\n",
    "* Increasing model size and dataset resolution will improve results but would be computationnally heavy\n",
    "* The training loss plot is spiky. I would expect lower step size and different optimizers might smooth out the behaviour. number of epochs didn't seem to have a big effect on the results. i've noticed a tiny improvement between 2 and 20 epochs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Submitting This Project\n",
    "When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as \"dlnd_face_generation.ipynb\" and save it as a HTML file under \"File\" -> \"Download as\". Include the \"problem_unittests.py\" files in your submission."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}