dlnd_face_generation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Face Generation\n",
    "\n",
    "In this project, you'll define and train a DCGAN on a dataset of faces. Your goal is to get a generator network to generate *new* images of faces that look as realistic as possible!\n",
    "\n",
    "The project will be broken down into a series of tasks from **loading in data to defining and training adversarial networks**. At the end of the notebook, you'll be able to visualize the results of your trained Generator to see how it performs; your generated samples should look like fairly realistic faces with small amounts of noise.\n",
    "\n",
    "### Get the Data\n",
    "\n",
    "You'll be using the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train your adversarial networks.\n",
    "\n",
    "This dataset is more complex than the number datasets (like MNIST or SVHN) you've been working with, and so, you should prepare to define deeper networks and train them for a longer time to get good results. It is suggested that you utilize a GPU for training.\n",
    "\n",
    "### Pre-processed Data\n",
    "\n",
    "Since the project's main focus is on building the GANs, we've done *some* of the pre-processing for you. Each of the CelebA images has been cropped to remove parts of the image that don't include a face, then resized down to 64x64x3 NumPy images. Some sample data is show below.\n",
    "\n",
    "<img src='assets/processed_face_data.png' width=60% />\n",
    "\n",
    "> If you are working locally, you can download this data [by clicking here](https://s3.amazonaws.com/video.udacity-data.com/topher/2018/November/5be7eb6f_processed-celeba-small/processed-celeba-small.zip)\n",
    "\n",
    "This is a zip file that you'll need to extract in the home directory of this notebook for further loading and processing. After extracting the data, you should be left with a directory of data `processed_celeba_small/`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Archive:  processed_celeba_small.zip\r\n",
      "replace processed_celeba_small/.DS_Store? [y]es, [n]o, [A]ll, [N]one, [r]ename: "
     ]
    }
   ],
   "source": [
    "# can comment out after executing\n",
    "!unzip processed_celeba_small.zip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# necessary imports\n",
    "import torch\n",
    "from torchvision import datasets\n",
    "from torchvision import transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.utils.data import DataLoader\n",
    "\n",
    "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",
    "    image_path = './' + data_dir\n",
    "    transform = transforms.Compose([transforms.Resize(image_size),\n",
    "                                    transforms.ToTensor()])\n",
    "    \n",
    "    dset = datasets.ImageFolder(image_path, transform)\n",
    "    dloader = DataLoader(dataset=dset, batch_size=batch_size, shuffle=True)\n",
    "    return dloader\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": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define function hyperparameters\n",
    "batch_size = 16\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": 9,
   "metadata": {},
   "outputs": [
    {
     "ename": "IndexError",
     "evalue": "index 16 is out of bounds for dimension 0 with size 16",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-9-13b4d9c7a4be>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     16\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0midx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0marange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplot_size\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     17\u001b[0m     \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfig\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_subplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplot_size\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0midx\u001b[0m\u001b[0;34m+\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mxticks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0myticks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 18\u001b[0;31m     \u001b[0mimshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimages\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0midx\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mIndexError\u001b[0m: index 16 is out of bounds for dimension 0 with size 16"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f23af73d668>"
      ]
     },
     "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": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# TODO: Complete the scale function\n",
    "def scale(x, feature_range=(-1, 1)):\n",
    "    ''' Scale takes in an image x and returns that image, scaled\n",
    "       with a feature_range of pixel values from -1 to 1. \n",
    "       This function assumes that the input x is already scaled from 0-1.'''\n",
    "    # assume x is scaled to (0, 1)\n",
    "    # scale to feature_range and return scaled x\n",
    "    min, max = feature_range\n",
    "    x = x * (max - min) + min\n",
    "    return x\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Min:  tensor(-0.9608)\n",
      "Max:  tensor(0.8118)\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": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn as nn\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# helper conv function\n",
    "def conv(in_channels, out_channels, kernel_size, stride=2, padding=1, batch_norm=True):\n",
    "    \"\"\"Creates a convolutional layer, with optional batch normalization.\n",
    "    \"\"\"\n",
    "    layers = []\n",
    "    conv_layer = nn.Conv2d(in_channels=in_channels, out_channels=out_channels, \n",
    "                           kernel_size=kernel_size, stride=stride, padding=padding, bias=False)\n",
    "    \n",
    "    layers.append(conv_layer)\n",
    "\n",
    "    if batch_norm:\n",
    "        layers.append(nn.BatchNorm2d(out_channels))\n",
    "    return nn.Sequential(*layers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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",
    "    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",
    "        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",
    "\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": 16,
   "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)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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",
    "        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",
    "        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": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.nn import init\n",
    "def weights_init_normal(m):\n",
    "    \"\"\"\n",
    "    Applies initial weights to certain layers in a model .\n",
    "    The weights are taken from a normal distribution \n",
    "    with mean = 0, std dev = 0.02.\n",
    "    :param m: A module or layer in a network    \n",
    "    \"\"\"\n",
    "    # classname will be something like:\n",
    "    # `Conv`, `BatchNorm2d`, `Linear`, etc.\n",
    "    classname = m.__class__.__name__\n",
    "    init_gain = 0.02\n",
    "    \n",
    "    # TODO: Apply initial weights to convolutional and linear layers\n",
    "    if hasattr(m, 'weight') and (classname.find('Linear') != -1 or classname.find('Conv') != -1):\n",
    "        init.normal_(m.weight.data, 0.0, init_gain)\n",
    "    elif (classname.find('BatchNorm2d') != -1):\n",
    "        init.normal_(m.weight.data, 1.0, init_gain)\n",
    "        init.constant_(m.bias.data, 0.0)\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Build complete network\n",
    "\n",
    "Define your models' hyperparameters and instantiate the discriminator and generator from the classes defined above. Make sure you've passed in the correct input arguments."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Discriminator(\n",
      "  (conv1): Sequential(\n",
      "    (0): Conv2d(3, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      "  (conv2): Sequential(\n",
      "    (0): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (conv3): Sequential(\n",
      "    (0): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (fc): Linear(in_features=4096, out_features=1, bias=True)\n",
      ")\n",
      "\n",
      "Generator(\n",
      "  (fc): Linear(in_features=100, out_features=4096, bias=True)\n",
      "  (t_conv1): Sequential(\n",
      "    (0): ConvTranspose2d(256, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv2): Sequential(\n",
      "    (0): ConvTranspose2d(128, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "    (1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
      "  )\n",
      "  (t_conv3): Sequential(\n",
      "    (0): ConvTranspose2d(64, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "# Define model hyperparams\n",
    "d_conv_dim = 64\n",
    "g_conv_dim = 64\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": 21,
   "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": 22,
   "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",
    "    batch_size = D_out.size(0)\n",
    "    labels = torch.ones(batch_size) # real labels = 1\n",
    "    # move labels to GPU if available     \n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "    # binary cross entropy with logits loss\n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    # calculate loss\n",
    "    loss = criterion(D_out.squeeze(), labels)\n",
    "    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",
    "    batch_size = D_out.size(0)\n",
    "    labels = torch.zeros(batch_size) # fake labels = 0\n",
    "    if train_on_gpu:\n",
    "        labels = labels.cuda()\n",
    "    criterion = nn.BCEWithLogitsLoss()\n",
    "    # calculate loss\n",
    "    loss = criterion(D_out.squeeze(), labels)\n",
    "    return loss"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Optimizers\n",
    "\n",
    "#### Exercise: Define optimizers for your Discriminator (D) and Generator (G)\n",
    "\n",
    "Define optimizers for your models with appropriate hyperparameters."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.optim as optim\n",
    "\n",
    "# params\n",
    "lr = 0.0002\n",
    "beta1=0.5\n",
    "beta2=0.999 # default value\n",
    "\n",
    "# Create optimizers for the discriminator and generator\n",
    "d_optimizer = optim.Adam(D.parameters(), lr, [beta1, beta2])\n",
    "g_optimizer = optim.Adam(G.parameters(), lr, [beta1, beta2])"
   ]
  },
  {
   "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": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(D, G, n_epochs, print_every=50):\n",
    "    '''Trains adversarial networks for some number of epochs\n",
    "       param, D: the discriminator network\n",
    "       param, G: the generator network\n",
    "       param, n_epochs: number of epochs to train for\n",
    "       param, print_every: when to print and record the models' losses\n",
    "       return: D and G losses'''\n",
    "    \n",
    "    # move models to GPU\n",
    "    if train_on_gpu:\n",
    "        D.cuda()\n",
    "        G.cuda()\n",
    "\n",
    "    # keep track of loss and generated, \"fake\" samples\n",
    "    samples = []\n",
    "    losses = []\n",
    "\n",
    "    # Get some fixed data for sampling. These are images that are held\n",
    "    # constant throughout training, and allow us to inspect the model's performance\n",
    "    sample_size=16\n",
    "    fixed_z = np.random.uniform(-1, 1, size=(sample_size, z_size))\n",
    "    fixed_z = torch.from_numpy(fixed_z).float()\n",
    "    # move z to GPU if available\n",
    "    if train_on_gpu:\n",
    "        fixed_z = fixed_z.cuda()\n",
    "\n",
    "    # epoch training loop\n",
    "    for epoch in range(n_epochs):\n",
    "\n",
    "        # batch training loop\n",
    "        for batch_i, (real_images, _) in enumerate(celeba_train_loader):\n",
    "\n",
    "            batch_size = real_images.size(0)\n",
    "            real_images = scale(real_images)\n",
    "\n",
    "            # ===============================================\n",
    "            #         YOUR CODE HERE: TRAIN THE NETWORKS\n",
    "            # ===============================================\n",
    "                   \n",
    "            # ============================================\n",
    "            #            TRAIN THE DISCRIMINATOR\n",
    "            # ============================================\n",
    "            # 1. Train with real images\n",
    "\n",
    "            # Compute the discriminator losses on real images \n",
    "            d_optimizer.zero_grad()\n",
    "            \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",
    "            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",
    "            # =========================================\n",
    "            #            TRAIN THE GENERATOR\n",
    "            # =========================================\n",
    "            g_optimizer.zero_grad()\n",
    "        \n",
    "            # Train with fake images and flipped labels\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": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [    1/   20] | d_loss: 1.4777 | g_loss: 1.5157\n",
      "Epoch [    1/   20] | d_loss: 0.4415 | g_loss: 3.3586\n",
      "Epoch [    1/   20] | d_loss: 0.0733 | g_loss: 4.7740\n",
      "Epoch [    1/   20] | d_loss: 0.6938 | g_loss: 6.8058\n",
      "Epoch [    1/   20] | d_loss: 0.7328 | g_loss: 3.0228\n",
      "Epoch [    1/   20] | d_loss: 0.3948 | g_loss: 4.1952\n",
      "Epoch [    1/   20] | d_loss: 0.8656 | g_loss: 5.7619\n",
      "Epoch [    1/   20] | d_loss: 0.4471 | g_loss: 3.7083\n",
      "Epoch [    1/   20] | d_loss: 1.1179 | g_loss: 2.1673\n",
      "Epoch [    1/   20] | d_loss: 0.5208 | g_loss: 3.7416\n",
      "Epoch [    1/   20] | d_loss: 0.3045 | g_loss: 3.7195\n",
      "Epoch [    1/   20] | d_loss: 0.8108 | g_loss: 2.5451\n",
      "Epoch [    1/   20] | d_loss: 0.8815 | g_loss: 1.6931\n",
      "Epoch [    1/   20] | d_loss: 0.8063 | g_loss: 4.2489\n",
      "Epoch [    1/   20] | d_loss: 0.5658 | g_loss: 4.1678\n",
      "Epoch [    1/   20] | d_loss: 0.7892 | g_loss: 2.2672\n",
      "Epoch [    1/   20] | d_loss: 0.5417 | g_loss: 3.2999\n",
      "Epoch [    1/   20] | d_loss: 0.6023 | g_loss: 3.5494\n",
      "Epoch [    1/   20] | d_loss: 0.6759 | g_loss: 2.8558\n",
      "Epoch [    1/   20] | d_loss: 0.4514 | g_loss: 3.0397\n",
      "Epoch [    1/   20] | d_loss: 0.6466 | g_loss: 3.0409\n",
      "Epoch [    1/   20] | d_loss: 0.9513 | g_loss: 1.4503\n",
      "Epoch [    1/   20] | d_loss: 0.9193 | g_loss: 3.6769\n",
      "Epoch [    1/   20] | d_loss: 2.1867 | g_loss: 4.5528\n",
      "Epoch [    1/   20] | d_loss: 1.0542 | g_loss: 1.6488\n",
      "Epoch [    1/   20] | d_loss: 1.1745 | g_loss: 3.8571\n",
      "Epoch [    1/   20] | d_loss: 0.4477 | g_loss: 2.2139\n",
      "Epoch [    1/   20] | d_loss: 1.2159 | g_loss: 1.3266\n",
      "Epoch [    1/   20] | d_loss: 1.0323 | g_loss: 2.5236\n",
      "Epoch [    1/   20] | d_loss: 1.2479 | g_loss: 1.0255\n",
      "Epoch [    1/   20] | d_loss: 1.0769 | g_loss: 1.9269\n",
      "Epoch [    1/   20] | d_loss: 0.8853 | g_loss: 2.1145\n",
      "Epoch [    1/   20] | d_loss: 1.4458 | g_loss: 2.0155\n",
      "Epoch [    1/   20] | d_loss: 0.9440 | g_loss: 1.5067\n",
      "Epoch [    1/   20] | d_loss: 0.9786 | g_loss: 1.0066\n",
      "Epoch [    1/   20] | d_loss: 1.2723 | g_loss: 0.8240\n",
      "Epoch [    1/   20] | d_loss: 1.1045 | g_loss: 1.2109\n",
      "Epoch [    1/   20] | d_loss: 0.8502 | g_loss: 2.2847\n",
      "Epoch [    1/   20] | d_loss: 0.8318 | g_loss: 1.9493\n",
      "Epoch [    1/   20] | d_loss: 0.9653 | g_loss: 1.2198\n",
      "Epoch [    1/   20] | d_loss: 1.0771 | g_loss: 2.0263\n",
      "Epoch [    1/   20] | d_loss: 0.7742 | g_loss: 1.8988\n",
      "Epoch [    1/   20] | d_loss: 0.7565 | g_loss: 1.7326\n",
      "Epoch [    1/   20] | d_loss: 0.6696 | g_loss: 2.4515\n",
      "Epoch [    1/   20] | d_loss: 0.9743 | g_loss: 1.7829\n",
      "Epoch [    1/   20] | d_loss: 1.4720 | g_loss: 1.0650\n",
      "Epoch [    1/   20] | d_loss: 0.6699 | g_loss: 1.0722\n",
      "Epoch [    1/   20] | d_loss: 0.8163 | g_loss: 1.2337\n",
      "Epoch [    1/   20] | d_loss: 1.0703 | g_loss: 1.6100\n",
      "Epoch [    1/   20] | d_loss: 1.0166 | g_loss: 1.6612\n",
      "Epoch [    1/   20] | d_loss: 0.9801 | g_loss: 1.3315\n",
      "Epoch [    1/   20] | d_loss: 1.2090 | g_loss: 0.9605\n",
      "Epoch [    1/   20] | d_loss: 1.0008 | g_loss: 1.3961\n",
      "Epoch [    1/   20] | d_loss: 0.9908 | g_loss: 1.2258\n",
      "Epoch [    1/   20] | d_loss: 0.9714 | g_loss: 1.4015\n",
      "Epoch [    1/   20] | d_loss: 1.1533 | g_loss: 1.2158\n",
      "Epoch [    1/   20] | d_loss: 1.3283 | g_loss: 1.3479\n",
      "Epoch [    1/   20] | d_loss: 0.7140 | g_loss: 1.7451\n",
      "Epoch [    1/   20] | d_loss: 1.1320 | g_loss: 1.4984\n",
      "Epoch [    1/   20] | d_loss: 1.0096 | g_loss: 1.1666\n",
      "Epoch [    1/   20] | d_loss: 1.0246 | g_loss: 0.9931\n",
      "Epoch [    1/   20] | d_loss: 0.8398 | g_loss: 1.4911\n",
      "Epoch [    1/   20] | d_loss: 1.0800 | g_loss: 1.1134\n",
      "Epoch [    1/   20] | d_loss: 0.7152 | g_loss: 2.8241\n",
      "Epoch [    1/   20] | d_loss: 1.0178 | g_loss: 1.3533\n",
      "Epoch [    1/   20] | d_loss: 0.8344 | g_loss: 1.4086\n",
      "Epoch [    1/   20] | d_loss: 0.8032 | g_loss: 1.3948\n",
      "Epoch [    1/   20] | d_loss: 1.0622 | g_loss: 1.2387\n",
      "Epoch [    1/   20] | d_loss: 0.6254 | g_loss: 2.1288\n",
      "Epoch [    1/   20] | d_loss: 1.5222 | g_loss: 1.1269\n",
      "Epoch [    1/   20] | d_loss: 0.8782 | g_loss: 0.9695\n",
      "Epoch [    1/   20] | d_loss: 1.1181 | g_loss: 1.0581\n",
      "Epoch [    1/   20] | d_loss: 0.6504 | g_loss: 1.1951\n",
      "Epoch [    1/   20] | d_loss: 0.9252 | g_loss: 1.1622\n",
      "Epoch [    1/   20] | d_loss: 1.0392 | g_loss: 1.3716\n",
      "Epoch [    1/   20] | d_loss: 0.7270 | g_loss: 1.2148\n",
      "Epoch [    1/   20] | d_loss: 1.1733 | g_loss: 1.3319\n",
      "Epoch [    1/   20] | d_loss: 0.8778 | g_loss: 1.6382\n",
      "Epoch [    1/   20] | d_loss: 0.8071 | g_loss: 2.3006\n",
      "Epoch [    1/   20] | d_loss: 0.9136 | g_loss: 1.5967\n",
      "Epoch [    1/   20] | d_loss: 0.9650 | g_loss: 2.0949\n",
      "Epoch [    1/   20] | d_loss: 0.9253 | g_loss: 1.4488\n",
      "Epoch [    1/   20] | d_loss: 0.6803 | g_loss: 1.2663\n",
      "Epoch [    1/   20] | d_loss: 0.8747 | g_loss: 1.2186\n",
      "Epoch [    1/   20] | d_loss: 1.1484 | g_loss: 1.1492\n",
      "Epoch [    1/   20] | d_loss: 1.1330 | g_loss: 1.3957\n",
      "Epoch [    1/   20] | d_loss: 0.9566 | g_loss: 1.6957\n",
      "Epoch [    1/   20] | d_loss: 0.5038 | g_loss: 1.6089\n",
      "Epoch [    1/   20] | d_loss: 0.9796 | g_loss: 1.4819\n",
      "Epoch [    1/   20] | d_loss: 0.9971 | g_loss: 1.5431\n",
      "Epoch [    1/   20] | d_loss: 1.0305 | g_loss: 1.9305\n",
      "Epoch [    1/   20] | d_loss: 1.1495 | g_loss: 0.7072\n",
      "Epoch [    1/   20] | d_loss: 1.0034 | g_loss: 1.5504\n",
      "Epoch [    1/   20] | d_loss: 0.6994 | g_loss: 1.9565\n",
      "Epoch [    1/   20] | d_loss: 0.9113 | g_loss: 1.6623\n",
      "Epoch [    1/   20] | d_loss: 1.2654 | g_loss: 1.0955\n",
      "Epoch [    1/   20] | d_loss: 1.5653 | g_loss: 1.2936\n",
      "Epoch [    1/   20] | d_loss: 0.8947 | g_loss: 1.8061\n",
      "Epoch [    1/   20] | d_loss: 0.9942 | g_loss: 1.7848\n",
      "Epoch [    1/   20] | d_loss: 1.4952 | g_loss: 1.1190\n",
      "Epoch [    1/   20] | d_loss: 0.7884 | g_loss: 1.1000\n",
      "Epoch [    1/   20] | d_loss: 0.9258 | g_loss: 0.6609\n",
      "Epoch [    1/   20] | d_loss: 0.5815 | g_loss: 1.3477\n",
      "Epoch [    1/   20] | d_loss: 0.8587 | g_loss: 2.0199\n",
      "Epoch [    1/   20] | d_loss: 1.2847 | g_loss: 1.3064\n",
      "Epoch [    1/   20] | d_loss: 0.9308 | g_loss: 1.3464\n",
      "Epoch [    1/   20] | d_loss: 1.1691 | g_loss: 1.1481\n",
      "Epoch [    1/   20] | d_loss: 1.0927 | g_loss: 0.9118\n",
      "Epoch [    1/   20] | d_loss: 1.1278 | g_loss: 1.3703\n",
      "Epoch [    1/   20] | d_loss: 1.2852 | g_loss: 0.8726\n",
      "Epoch [    1/   20] | d_loss: 1.5128 | g_loss: 1.7699\n",
      "Epoch [    1/   20] | d_loss: 0.5483 | g_loss: 1.9377\n",
      "Epoch [    1/   20] | d_loss: 1.2600 | g_loss: 1.0076\n",
      "Epoch [    2/   20] | d_loss: 1.6156 | g_loss: 1.3357\n",
      "Epoch [    2/   20] | d_loss: 0.6572 | g_loss: 1.1303\n",
      "Epoch [    2/   20] | d_loss: 0.8334 | g_loss: 1.0393\n",
      "Epoch [    2/   20] | d_loss: 0.6332 | g_loss: 1.2391\n",
      "Epoch [    2/   20] | d_loss: 1.5597 | g_loss: 1.4616\n",
      "Epoch [    2/   20] | d_loss: 1.1252 | g_loss: 0.3370\n",
      "Epoch [    2/   20] | d_loss: 0.7147 | g_loss: 2.2091\n",
      "Epoch [    2/   20] | d_loss: 0.5934 | g_loss: 1.4825\n",
      "Epoch [    2/   20] | d_loss: 0.8950 | g_loss: 1.6707\n",
      "Epoch [    2/   20] | d_loss: 0.9064 | g_loss: 1.5536\n",
      "Epoch [    2/   20] | d_loss: 0.9955 | g_loss: 2.0992\n",
      "Epoch [    2/   20] | d_loss: 0.9206 | g_loss: 1.6816\n",
      "Epoch [    2/   20] | d_loss: 1.0988 | g_loss: 2.0489\n",
      "Epoch [    2/   20] | d_loss: 1.3044 | g_loss: 1.1565\n",
      "Epoch [    2/   20] | d_loss: 0.7341 | g_loss: 1.8287\n",
      "Epoch [    2/   20] | d_loss: 1.1612 | g_loss: 1.3407\n",
      "Epoch [    2/   20] | d_loss: 0.9675 | g_loss: 1.3635\n",
      "Epoch [    2/   20] | d_loss: 0.8197 | g_loss: 1.5979\n",
      "Epoch [    2/   20] | d_loss: 0.7773 | g_loss: 2.1433\n",
      "Epoch [    2/   20] | d_loss: 1.5251 | g_loss: 1.5652\n",
      "Epoch [    2/   20] | d_loss: 0.8554 | g_loss: 1.1817\n",
      "Epoch [    2/   20] | d_loss: 1.4832 | g_loss: 0.7530\n",
      "Epoch [    2/   20] | d_loss: 1.1606 | g_loss: 2.0541\n",
      "Epoch [    2/   20] | d_loss: 1.6846 | g_loss: 1.7820\n",
      "Epoch [    2/   20] | d_loss: 0.9029 | g_loss: 2.1606\n",
      "Epoch [    2/   20] | d_loss: 1.0608 | g_loss: 1.4605\n",
      "Epoch [    2/   20] | d_loss: 0.5164 | g_loss: 2.7689\n",
      "Epoch [    2/   20] | d_loss: 1.0570 | g_loss: 1.6238\n",
      "Epoch [    2/   20] | d_loss: 1.0192 | g_loss: 1.5053\n",
      "Epoch [    2/   20] | d_loss: 0.9418 | g_loss: 1.3760\n",
      "Epoch [    2/   20] | d_loss: 0.8363 | g_loss: 1.6546\n",
      "Epoch [    2/   20] | d_loss: 0.8882 | g_loss: 0.8331\n",
      "Epoch [    2/   20] | d_loss: 1.0409 | g_loss: 0.8547\n",
      "Epoch [    2/   20] | d_loss: 1.0544 | g_loss: 1.6383\n",
      "Epoch [    2/   20] | d_loss: 0.6929 | g_loss: 2.0705\n",
      "Epoch [    2/   20] | d_loss: 1.0784 | g_loss: 2.0320\n",
      "Epoch [    2/   20] | d_loss: 1.2180 | g_loss: 0.7608\n",
      "Epoch [    2/   20] | d_loss: 1.0202 | g_loss: 1.0917\n",
      "Epoch [    2/   20] | d_loss: 0.9872 | g_loss: 1.0568\n",
      "Epoch [    2/   20] | d_loss: 1.1704 | g_loss: 1.4230\n",
      "Epoch [    2/   20] | d_loss: 1.1315 | g_loss: 2.5125\n",
      "Epoch [    2/   20] | d_loss: 0.7533 | g_loss: 1.9335\n",
      "Epoch [    2/   20] | d_loss: 1.0210 | g_loss: 1.3888\n",
      "Epoch [    2/   20] | d_loss: 1.0177 | g_loss: 0.6011\n",
      "Epoch [    2/   20] | d_loss: 1.1001 | g_loss: 1.3558\n",
      "Epoch [    2/   20] | d_loss: 0.8750 | g_loss: 0.8951\n",
      "Epoch [    2/   20] | d_loss: 1.1535 | g_loss: 1.3446\n",
      "Epoch [    2/   20] | d_loss: 1.0214 | g_loss: 1.2335\n",
      "Epoch [    2/   20] | d_loss: 0.9728 | g_loss: 1.6902\n",
      "Epoch [    2/   20] | d_loss: 1.0444 | g_loss: 1.3284\n",
      "Epoch [    2/   20] | d_loss: 0.8699 | g_loss: 1.6953\n",
      "Epoch [    2/   20] | d_loss: 0.9324 | g_loss: 2.1900\n",
      "Epoch [    2/   20] | d_loss: 1.1566 | g_loss: 1.6342\n",
      "Epoch [    2/   20] | d_loss: 0.9766 | g_loss: 1.8316\n",
      "Epoch [    2/   20] | d_loss: 1.8029 | g_loss: 0.4307\n",
      "Epoch [    2/   20] | d_loss: 0.9325 | g_loss: 1.4338\n",
      "Epoch [    2/   20] | d_loss: 0.9506 | g_loss: 1.0798\n",
      "Epoch [    2/   20] | d_loss: 0.9632 | g_loss: 1.1326\n",
      "Epoch [    2/   20] | d_loss: 1.1725 | g_loss: 0.6408\n",
      "Epoch [    2/   20] | d_loss: 1.0620 | g_loss: 0.6603\n",
      "Epoch [    2/   20] | d_loss: 0.9886 | g_loss: 1.6419\n",
      "Epoch [    2/   20] | d_loss: 1.9524 | g_loss: 0.5842\n",
      "Epoch [    2/   20] | d_loss: 1.2588 | g_loss: 1.4296\n",
      "Epoch [    2/   20] | d_loss: 0.8272 | g_loss: 1.6987\n",
      "Epoch [    2/   20] | d_loss: 0.6265 | g_loss: 1.8254\n",
      "Epoch [    2/   20] | d_loss: 0.6757 | g_loss: 0.9390\n",
      "Epoch [    2/   20] | d_loss: 1.0086 | g_loss: 0.8579\n",
      "Epoch [    2/   20] | d_loss: 0.5721 | g_loss: 1.1890\n",
      "Epoch [    2/   20] | d_loss: 0.8045 | g_loss: 1.8476\n",
      "Epoch [    2/   20] | d_loss: 1.2043 | g_loss: 1.4723\n",
      "Epoch [    2/   20] | d_loss: 0.6314 | g_loss: 0.5717\n",
      "Epoch [    2/   20] | d_loss: 0.5975 | g_loss: 1.2328\n",
      "Epoch [    2/   20] | d_loss: 1.0635 | g_loss: 1.7794\n",
      "Epoch [    2/   20] | d_loss: 0.9814 | g_loss: 1.8314\n",
      "Epoch [    2/   20] | d_loss: 0.7892 | g_loss: 1.2923\n",
      "Epoch [    2/   20] | d_loss: 0.8464 | g_loss: 2.8471\n",
      "Epoch [    2/   20] | d_loss: 1.4999 | g_loss: 1.4253\n",
      "Epoch [    2/   20] | d_loss: 0.8732 | g_loss: 1.7191\n",
      "Epoch [    2/   20] | d_loss: 1.1530 | g_loss: 1.1523\n",
      "Epoch [    2/   20] | d_loss: 1.3289 | g_loss: 1.4776\n",
      "Epoch [    2/   20] | d_loss: 0.7908 | g_loss: 1.2057\n",
      "Epoch [    2/   20] | d_loss: 0.7666 | g_loss: 1.5258\n",
      "Epoch [    2/   20] | d_loss: 1.1626 | g_loss: 1.6095\n",
      "Epoch [    2/   20] | d_loss: 1.3021 | g_loss: 0.8984\n",
      "Epoch [    2/   20] | d_loss: 1.0911 | g_loss: 1.3232\n",
      "Epoch [    2/   20] | d_loss: 1.2244 | g_loss: 1.1339\n",
      "Epoch [    2/   20] | d_loss: 1.1292 | g_loss: 0.5904\n",
      "Epoch [    2/   20] | d_loss: 1.2622 | g_loss: 1.9802\n",
      "Epoch [    2/   20] | d_loss: 0.6771 | g_loss: 2.7513\n",
      "Epoch [    2/   20] | d_loss: 0.9746 | g_loss: 1.0931\n",
      "Epoch [    2/   20] | d_loss: 1.1647 | g_loss: 1.5737\n",
      "Epoch [    2/   20] | d_loss: 1.2225 | g_loss: 1.8828\n",
      "Epoch [    2/   20] | d_loss: 1.5665 | g_loss: 1.6007\n",
      "Epoch [    2/   20] | d_loss: 0.4274 | g_loss: 1.5202\n",
      "Epoch [    2/   20] | d_loss: 0.7413 | g_loss: 2.0996\n",
      "Epoch [    2/   20] | d_loss: 0.8210 | g_loss: 1.1783\n",
      "Epoch [    2/   20] | d_loss: 0.8915 | g_loss: 1.3369\n",
      "Epoch [    2/   20] | d_loss: 0.8441 | g_loss: 1.0809\n",
      "Epoch [    2/   20] | d_loss: 1.0860 | g_loss: 1.6481\n",
      "Epoch [    2/   20] | d_loss: 0.9326 | g_loss: 2.3616\n",
      "Epoch [    2/   20] | d_loss: 1.0217 | g_loss: 0.8528\n",
      "Epoch [    2/   20] | d_loss: 0.9120 | g_loss: 0.9319\n",
      "Epoch [    2/   20] | d_loss: 1.3132 | g_loss: 1.5855\n",
      "Epoch [    2/   20] | d_loss: 0.8477 | g_loss: 1.3178\n",
      "Epoch [    2/   20] | d_loss: 0.9545 | g_loss: 1.3472\n",
      "Epoch [    2/   20] | d_loss: 0.6293 | g_loss: 2.2820\n",
      "Epoch [    2/   20] | d_loss: 1.2196 | g_loss: 1.0970\n",
      "Epoch [    2/   20] | d_loss: 0.7652 | g_loss: 1.0330\n",
      "Epoch [    2/   20] | d_loss: 1.2759 | g_loss: 1.7284\n",
      "Epoch [    2/   20] | d_loss: 1.1616 | g_loss: 0.7455\n",
      "Epoch [    2/   20] | d_loss: 0.8687 | g_loss: 1.1901\n",
      "Epoch [    2/   20] | d_loss: 1.0862 | g_loss: 1.0488\n",
      "Epoch [    2/   20] | d_loss: 0.8321 | g_loss: 1.3744\n",
      "Epoch [    3/   20] | d_loss: 1.3474 | g_loss: 2.1845\n",
      "Epoch [    3/   20] | d_loss: 0.9996 | g_loss: 1.4297\n",
      "Epoch [    3/   20] | d_loss: 0.9676 | g_loss: 1.3517\n",
      "Epoch [    3/   20] | d_loss: 0.8353 | g_loss: 1.6171\n",
      "Epoch [    3/   20] | d_loss: 1.2130 | g_loss: 1.3593\n",
      "Epoch [    3/   20] | d_loss: 0.8545 | g_loss: 0.9084\n",
      "Epoch [    3/   20] | d_loss: 1.0496 | g_loss: 2.0556\n",
      "Epoch [    3/   20] | d_loss: 0.5366 | g_loss: 2.1696\n",
      "Epoch [    3/   20] | d_loss: 1.0686 | g_loss: 0.9500\n",
      "Epoch [    3/   20] | d_loss: 0.8688 | g_loss: 1.8069\n",
      "Epoch [    3/   20] | d_loss: 0.8280 | g_loss: 1.7275\n",
      "Epoch [    3/   20] | d_loss: 0.8162 | g_loss: 1.0777\n",
      "Epoch [    3/   20] | d_loss: 0.9014 | g_loss: 0.8338\n",
      "Epoch [    3/   20] | d_loss: 0.5038 | g_loss: 1.1248\n",
      "Epoch [    3/   20] | d_loss: 0.6035 | g_loss: 1.9586\n",
      "Epoch [    3/   20] | d_loss: 0.8575 | g_loss: 1.7462\n",
      "Epoch [    3/   20] | d_loss: 1.0140 | g_loss: 1.2060\n",
      "Epoch [    3/   20] | d_loss: 0.9213 | g_loss: 0.6519\n",
      "Epoch [    3/   20] | d_loss: 1.2395 | g_loss: 1.7390\n",
      "Epoch [    3/   20] | d_loss: 0.8258 | g_loss: 1.5598\n",
      "Epoch [    3/   20] | d_loss: 1.0998 | g_loss: 1.0777\n",
      "Epoch [    3/   20] | d_loss: 1.3186 | g_loss: 1.6317\n",
      "Epoch [    3/   20] | d_loss: 0.8641 | g_loss: 1.1782\n",
      "Epoch [    3/   20] | d_loss: 0.5972 | g_loss: 2.4035\n",
      "Epoch [    3/   20] | d_loss: 0.8832 | g_loss: 1.4641\n",
      "Epoch [    3/   20] | d_loss: 0.7156 | g_loss: 2.1394\n",
      "Epoch [    3/   20] | d_loss: 1.1670 | g_loss: 1.5620\n",
      "Epoch [    3/   20] | d_loss: 0.7857 | g_loss: 1.8491\n",
      "Epoch [    3/   20] | d_loss: 1.0859 | g_loss: 1.0550\n",
      "Epoch [    3/   20] | d_loss: 0.6526 | g_loss: 1.7303\n",
      "Epoch [    3/   20] | d_loss: 1.3315 | g_loss: 1.0889\n",
      "Epoch [    3/   20] | d_loss: 0.6869 | g_loss: 2.9560\n",
      "Epoch [    3/   20] | d_loss: 1.1313 | g_loss: 1.5017\n",
      "Epoch [    3/   20] | d_loss: 0.9912 | g_loss: 1.6158\n",
      "Epoch [    3/   20] | d_loss: 0.7903 | g_loss: 2.3653\n",
      "Epoch [    3/   20] | d_loss: 0.9295 | g_loss: 1.5672\n",
      "Epoch [    3/   20] | d_loss: 1.2005 | g_loss: 0.2562\n",
      "Epoch [    3/   20] | d_loss: 0.3957 | g_loss: 1.6349\n",
      "Epoch [    3/   20] | d_loss: 0.6477 | g_loss: 0.6294\n",
      "Epoch [    3/   20] | d_loss: 0.7256 | g_loss: 1.5065\n",
      "Epoch [    3/   20] | d_loss: 0.7727 | g_loss: 0.7893\n",
      "Epoch [    3/   20] | d_loss: 1.0617 | g_loss: 1.7202\n",
      "Epoch [    3/   20] | d_loss: 0.8497 | g_loss: 2.0127\n",
      "Epoch [    3/   20] | d_loss: 1.2661 | g_loss: 1.8585\n",
      "Epoch [    3/   20] | d_loss: 0.8158 | g_loss: 2.4153\n",
      "Epoch [    3/   20] | d_loss: 1.1921 | g_loss: 0.3099\n",
      "Epoch [    3/   20] | d_loss: 2.3587 | g_loss: 0.6003\n",
      "Epoch [    3/   20] | d_loss: 0.6939 | g_loss: 1.6331\n",
      "Epoch [    3/   20] | d_loss: 0.7337 | g_loss: 2.0107\n",
      "Epoch [    3/   20] | d_loss: 0.8113 | g_loss: 1.5662\n",
      "Epoch [    3/   20] | d_loss: 1.2904 | g_loss: 1.2980\n",
      "Epoch [    3/   20] | d_loss: 1.0202 | g_loss: 0.8176\n",
      "Epoch [    3/   20] | d_loss: 1.3699 | g_loss: 1.5700\n",
      "Epoch [    3/   20] | d_loss: 0.9292 | g_loss: 1.3270\n",
      "Epoch [    3/   20] | d_loss: 1.2046 | g_loss: 2.0592\n",
      "Epoch [    3/   20] | d_loss: 0.5646 | g_loss: 1.2715\n",
      "Epoch [    3/   20] | d_loss: 0.8716 | g_loss: 2.2971\n",
      "Epoch [    3/   20] | d_loss: 0.9094 | g_loss: 1.6395\n",
      "Epoch [    3/   20] | d_loss: 0.5421 | g_loss: 1.1858\n",
      "Epoch [    3/   20] | d_loss: 0.9045 | g_loss: 1.5977\n",
      "Epoch [    3/   20] | d_loss: 1.1183 | g_loss: 0.6818\n",
      "Epoch [    3/   20] | d_loss: 0.9307 | g_loss: 1.4424\n",
      "Epoch [    3/   20] | d_loss: 0.8110 | g_loss: 1.6739\n",
      "Epoch [    3/   20] | d_loss: 0.7033 | g_loss: 0.4984\n",
      "Epoch [    3/   20] | d_loss: 0.6197 | g_loss: 2.0154\n",
      "Epoch [    3/   20] | d_loss: 0.8018 | g_loss: 1.4128\n",
      "Epoch [    3/   20] | d_loss: 0.9780 | g_loss: 1.2174\n",
      "Epoch [    3/   20] | d_loss: 1.1289 | g_loss: 2.2319\n",
      "Epoch [    3/   20] | d_loss: 1.2040 | g_loss: 1.0113\n",
      "Epoch [    3/   20] | d_loss: 1.3409 | g_loss: 1.4448\n",
      "Epoch [    3/   20] | d_loss: 0.8973 | g_loss: 1.7528\n",
      "Epoch [    3/   20] | d_loss: 0.8165 | g_loss: 1.5157\n",
      "Epoch [    3/   20] | d_loss: 1.2606 | g_loss: 1.8351\n",
      "Epoch [    3/   20] | d_loss: 0.9174 | g_loss: 1.7852\n",
      "Epoch [    3/   20] | d_loss: 0.7058 | g_loss: 0.6478\n",
      "Epoch [    3/   20] | d_loss: 1.1341 | g_loss: 1.4929\n",
      "Epoch [    3/   20] | d_loss: 0.7125 | g_loss: 1.1944\n",
      "Epoch [    3/   20] | d_loss: 0.8146 | g_loss: 0.9573\n",
      "Epoch [    3/   20] | d_loss: 0.6337 | g_loss: 0.5959\n",
      "Epoch [    3/   20] | d_loss: 0.8833 | g_loss: 1.4152\n",
      "Epoch [    3/   20] | d_loss: 1.0608 | g_loss: 1.9546\n",
      "Epoch [    3/   20] | d_loss: 0.8662 | g_loss: 1.5531\n",
      "Epoch [    3/   20] | d_loss: 1.0249 | g_loss: 0.9477\n",
      "Epoch [    3/   20] | d_loss: 0.8605 | g_loss: 1.3126\n",
      "Epoch [    3/   20] | d_loss: 0.8380 | g_loss: 1.8131\n",
      "Epoch [    3/   20] | d_loss: 1.1483 | g_loss: 0.9329\n",
      "Epoch [    3/   20] | d_loss: 0.5445 | g_loss: 1.6186\n",
      "Epoch [    3/   20] | d_loss: 0.4845 | g_loss: 2.5907\n",
      "Epoch [    3/   20] | d_loss: 1.0815 | g_loss: 1.3184\n",
      "Epoch [    3/   20] | d_loss: 1.0814 | g_loss: 1.2270\n",
      "Epoch [    3/   20] | d_loss: 1.1602 | g_loss: 0.6880\n",
      "Epoch [    3/   20] | d_loss: 0.9606 | g_loss: 1.2931\n",
      "Epoch [    3/   20] | d_loss: 1.5508 | g_loss: 0.8577\n",
      "Epoch [    3/   20] | d_loss: 0.9386 | g_loss: 0.8663\n",
      "Epoch [    3/   20] | d_loss: 0.9239 | g_loss: 1.6778\n",
      "Epoch [    3/   20] | d_loss: 0.5434 | g_loss: 1.0148\n",
      "Epoch [    3/   20] | d_loss: 0.7268 | g_loss: 1.0638\n",
      "Epoch [    3/   20] | d_loss: 0.3634 | g_loss: 1.9019\n",
      "Epoch [    3/   20] | d_loss: 0.5627 | g_loss: 1.6756\n",
      "Epoch [    3/   20] | d_loss: 0.8117 | g_loss: 1.6787\n",
      "Epoch [    3/   20] | d_loss: 0.3865 | g_loss: 2.3867\n",
      "Epoch [    3/   20] | d_loss: 0.8452 | g_loss: 1.5989\n",
      "Epoch [    3/   20] | d_loss: 0.6131 | g_loss: 1.8166\n",
      "Epoch [    3/   20] | d_loss: 1.0034 | g_loss: 1.2006\n",
      "Epoch [    3/   20] | d_loss: 1.3376 | g_loss: 0.9212\n",
      "Epoch [    3/   20] | d_loss: 0.9887 | g_loss: 0.6135\n",
      "Epoch [    3/   20] | d_loss: 0.9416 | g_loss: 1.2031\n",
      "Epoch [    3/   20] | d_loss: 0.6056 | g_loss: 2.3440\n",
      "Epoch [    3/   20] | d_loss: 0.4501 | g_loss: 2.0260\n",
      "Epoch [    3/   20] | d_loss: 0.7279 | g_loss: 2.8882\n",
      "Epoch [    3/   20] | d_loss: 0.6058 | g_loss: 1.9048\n",
      "Epoch [    3/   20] | d_loss: 0.6658 | g_loss: 1.7005\n",
      "Epoch [    3/   20] | d_loss: 1.1999 | g_loss: 1.7882\n",
      "Epoch [    4/   20] | d_loss: 0.7836 | g_loss: 1.8419\n",
      "Epoch [    4/   20] | d_loss: 1.1390 | g_loss: 1.4071\n",
      "Epoch [    4/   20] | d_loss: 0.4602 | g_loss: 1.4095\n",
      "Epoch [    4/   20] | d_loss: 0.8485 | g_loss: 2.4934\n",
      "Epoch [    4/   20] | d_loss: 1.1004 | g_loss: 1.9481\n",
      "Epoch [    4/   20] | d_loss: 1.2274 | g_loss: 1.5584\n",
      "Epoch [    4/   20] | d_loss: 0.8179 | g_loss: 1.8614\n",
      "Epoch [    4/   20] | d_loss: 0.9655 | g_loss: 2.9815\n",
      "Epoch [    4/   20] | d_loss: 0.7399 | g_loss: 1.9113\n",
      "Epoch [    4/   20] | d_loss: 0.6910 | g_loss: 1.5243\n",
      "Epoch [    4/   20] | d_loss: 1.3446 | g_loss: 1.6706\n",
      "Epoch [    4/   20] | d_loss: 0.7744 | g_loss: 1.2720\n",
      "Epoch [    4/   20] | d_loss: 0.6486 | g_loss: 1.9076\n",
      "Epoch [    4/   20] | d_loss: 0.7542 | g_loss: 1.4067\n",
      "Epoch [    4/   20] | d_loss: 0.5149 | g_loss: 1.6872\n",
      "Epoch [    4/   20] | d_loss: 0.5610 | g_loss: 1.0342\n",
      "Epoch [    4/   20] | d_loss: 0.8526 | g_loss: 1.9006\n",
      "Epoch [    4/   20] | d_loss: 0.4675 | g_loss: 1.9184\n",
      "Epoch [    4/   20] | d_loss: 0.8995 | g_loss: 1.0850\n",
      "Epoch [    4/   20] | d_loss: 0.8241 | g_loss: 1.5510\n",
      "Epoch [    4/   20] | d_loss: 0.7482 | g_loss: 0.8046\n",
      "Epoch [    4/   20] | d_loss: 0.4978 | g_loss: 1.5913\n",
      "Epoch [    4/   20] | d_loss: 0.6423 | g_loss: 1.6849\n",
      "Epoch [    4/   20] | d_loss: 0.8961 | g_loss: 1.8454\n",
      "Epoch [    4/   20] | d_loss: 1.1177 | g_loss: 1.2981\n",
      "Epoch [    4/   20] | d_loss: 0.4485 | g_loss: 2.1032\n",
      "Epoch [    4/   20] | d_loss: 1.1268 | g_loss: 1.1399\n",
      "Epoch [    4/   20] | d_loss: 0.7074 | g_loss: 2.2233\n",
      "Epoch [    4/   20] | d_loss: 0.8535 | g_loss: 1.0586\n",
      "Epoch [    4/   20] | d_loss: 1.5314 | g_loss: 2.1558\n",
      "Epoch [    4/   20] | d_loss: 0.8581 | g_loss: 1.2207\n",
      "Epoch [    4/   20] | d_loss: 0.5489 | g_loss: 1.8098\n",
      "Epoch [    4/   20] | d_loss: 0.7494 | g_loss: 1.5933\n",
      "Epoch [    4/   20] | d_loss: 0.7713 | g_loss: 1.1533\n",
      "Epoch [    4/   20] | d_loss: 0.8042 | g_loss: 1.4845\n",
      "Epoch [    4/   20] | d_loss: 0.7255 | g_loss: 1.7393\n",
      "Epoch [    4/   20] | d_loss: 0.5763 | g_loss: 3.2007\n",
      "Epoch [    4/   20] | d_loss: 1.7668 | g_loss: 0.9491\n",
      "Epoch [    4/   20] | d_loss: 0.7954 | g_loss: 1.1826\n",
      "Epoch [    4/   20] | d_loss: 0.7707 | g_loss: 1.1684\n",
      "Epoch [    4/   20] | d_loss: 0.7596 | g_loss: 2.0910\n",
      "Epoch [    4/   20] | d_loss: 1.0046 | g_loss: 1.6453\n",
      "Epoch [    4/   20] | d_loss: 0.8711 | g_loss: 2.3151\n",
      "Epoch [    4/   20] | d_loss: 1.3138 | g_loss: 2.2001\n",
      "Epoch [    4/   20] | d_loss: 0.7927 | g_loss: 1.4628\n",
      "Epoch [    4/   20] | d_loss: 0.8690 | g_loss: 1.0444\n",
      "Epoch [    4/   20] | d_loss: 0.9231 | g_loss: 1.2763\n",
      "Epoch [    4/   20] | d_loss: 0.8063 | g_loss: 1.7540\n",
      "Epoch [    4/   20] | d_loss: 0.7829 | g_loss: 1.5905\n",
      "Epoch [    4/   20] | d_loss: 0.9035 | g_loss: 1.8467\n",
      "Epoch [    4/   20] | d_loss: 0.9757 | g_loss: 1.3164\n",
      "Epoch [    4/   20] | d_loss: 0.8350 | g_loss: 1.6629\n",
      "Epoch [    4/   20] | d_loss: 1.0410 | g_loss: 2.0380\n",
      "Epoch [    4/   20] | d_loss: 0.4795 | g_loss: 1.2602\n",
      "Epoch [    4/   20] | d_loss: 0.3864 | g_loss: 1.5173\n",
      "Epoch [    4/   20] | d_loss: 1.0713 | g_loss: 1.3665\n",
      "Epoch [    4/   20] | d_loss: 1.0445 | g_loss: 1.6389\n",
      "Epoch [    4/   20] | d_loss: 0.8597 | g_loss: 2.2557\n",
      "Epoch [    4/   20] | d_loss: 1.2039 | g_loss: 1.5047\n",
      "Epoch [    4/   20] | d_loss: 0.9167 | g_loss: 2.1765\n",
      "Epoch [    4/   20] | d_loss: 0.9488 | g_loss: 1.7160\n",
      "Epoch [    4/   20] | d_loss: 1.2342 | g_loss: 1.5873\n",
      "Epoch [    4/   20] | d_loss: 0.8540 | g_loss: 2.6019\n",
      "Epoch [    4/   20] | d_loss: 0.5629 | g_loss: 3.6029\n",
      "Epoch [    4/   20] | d_loss: 0.9878 | g_loss: 1.3239\n",
      "Epoch [    4/   20] | d_loss: 0.4433 | g_loss: 1.5597\n",
      "Epoch [    4/   20] | d_loss: 0.6694 | g_loss: 1.8914\n",
      "Epoch [    4/   20] | d_loss: 0.4744 | g_loss: 2.1749\n",
      "Epoch [    4/   20] | d_loss: 0.8778 | g_loss: 2.1694\n",
      "Epoch [    4/   20] | d_loss: 1.2174 | g_loss: 2.1614\n",
      "Epoch [    4/   20] | d_loss: 0.8123 | g_loss: 1.9935\n",
      "Epoch [    4/   20] | d_loss: 0.9762 | g_loss: 1.1527\n",
      "Epoch [    4/   20] | d_loss: 0.6684 | g_loss: 1.6424\n",
      "Epoch [    4/   20] | d_loss: 0.3159 | g_loss: 2.0010\n",
      "Epoch [    4/   20] | d_loss: 0.7784 | g_loss: 0.9291\n",
      "Epoch [    4/   20] | d_loss: 0.2920 | g_loss: 2.9970\n",
      "Epoch [    4/   20] | d_loss: 0.7850 | g_loss: 1.0986\n",
      "Epoch [    4/   20] | d_loss: 0.7048 | g_loss: 1.5352\n",
      "Epoch [    4/   20] | d_loss: 0.5997 | g_loss: 1.3628\n",
      "Epoch [    4/   20] | d_loss: 0.8470 | g_loss: 1.9214\n",
      "Epoch [    4/   20] | d_loss: 1.0398 | g_loss: 2.1533\n",
      "Epoch [    4/   20] | d_loss: 0.7606 | g_loss: 1.9052\n",
      "Epoch [    4/   20] | d_loss: 0.6284 | g_loss: 2.2422\n",
      "Epoch [    4/   20] | d_loss: 0.5375 | g_loss: 0.8917\n",
      "Epoch [    4/   20] | d_loss: 0.9173 | g_loss: 0.8551\n",
      "Epoch [    4/   20] | d_loss: 0.6704 | g_loss: 1.9433\n",
      "Epoch [    4/   20] | d_loss: 1.7906 | g_loss: 1.6566\n",
      "Epoch [    4/   20] | d_loss: 0.7019 | g_loss: 1.6526\n",
      "Epoch [    4/   20] | d_loss: 0.8563 | g_loss: 1.3154\n",
      "Epoch [    4/   20] | d_loss: 1.1542 | g_loss: 1.2803\n",
      "Epoch [    4/   20] | d_loss: 0.6131 | g_loss: 1.9081\n",
      "Epoch [    4/   20] | d_loss: 0.8353 | g_loss: 3.6216\n",
      "Epoch [    4/   20] | d_loss: 1.0068 | g_loss: 1.1475\n",
      "Epoch [    4/   20] | d_loss: 0.5869 | g_loss: 1.4516\n",
      "Epoch [    4/   20] | d_loss: 1.2558 | g_loss: 2.2839\n",
      "Epoch [    4/   20] | d_loss: 1.0474 | g_loss: 1.1252\n",
      "Epoch [    4/   20] | d_loss: 1.6731 | g_loss: 0.5285\n",
      "Epoch [    4/   20] | d_loss: 0.7006 | g_loss: 2.0897\n",
      "Epoch [    4/   20] | d_loss: 0.7906 | g_loss: 1.5134\n",
      "Epoch [    4/   20] | d_loss: 0.8976 | g_loss: 1.5199\n",
      "Epoch [    4/   20] | d_loss: 0.9550 | g_loss: 1.3397\n",
      "Epoch [    4/   20] | d_loss: 0.9758 | g_loss: 2.0884\n",
      "Epoch [    4/   20] | d_loss: 0.8225 | g_loss: 1.8248\n",
      "Epoch [    4/   20] | d_loss: 0.7954 | g_loss: 1.4284\n",
      "Epoch [    4/   20] | d_loss: 0.3026 | g_loss: 3.3482\n",
      "Epoch [    4/   20] | d_loss: 0.4828 | g_loss: 1.8478\n",
      "Epoch [    4/   20] | d_loss: 1.1528 | g_loss: 1.3965\n",
      "Epoch [    4/   20] | d_loss: 0.7364 | g_loss: 2.1588\n",
      "Epoch [    4/   20] | d_loss: 0.5439 | g_loss: 2.4483\n",
      "Epoch [    4/   20] | d_loss: 0.8608 | g_loss: 1.4007\n",
      "Epoch [    4/   20] | d_loss: 0.5863 | g_loss: 1.6798\n",
      "Epoch [    4/   20] | d_loss: 0.9394 | g_loss: 2.7878\n",
      "Epoch [    5/   20] | d_loss: 0.9450 | g_loss: 1.8686\n",
      "Epoch [    5/   20] | d_loss: 0.8605 | g_loss: 2.7244\n",
      "Epoch [    5/   20] | d_loss: 0.7824 | g_loss: 2.2748\n",
      "Epoch [    5/   20] | d_loss: 0.4649 | g_loss: 2.8900\n",
      "Epoch [    5/   20] | d_loss: 0.4723 | g_loss: 2.0231\n",
      "Epoch [    5/   20] | d_loss: 0.7258 | g_loss: 1.4618\n",
      "Epoch [    5/   20] | d_loss: 0.7880 | g_loss: 1.8745\n",
      "Epoch [    5/   20] | d_loss: 1.1654 | g_loss: 1.2903\n",
      "Epoch [    5/   20] | d_loss: 0.4241 | g_loss: 1.9407\n",
      "Epoch [    5/   20] | d_loss: 0.9011 | g_loss: 1.3931\n",
      "Epoch [    5/   20] | d_loss: 1.0083 | g_loss: 1.4173\n",
      "Epoch [    5/   20] | d_loss: 0.6164 | g_loss: 1.7375\n",
      "Epoch [    5/   20] | d_loss: 0.5156 | g_loss: 2.1360\n",
      "Epoch [    5/   20] | d_loss: 0.6466 | g_loss: 1.8523\n",
      "Epoch [    5/   20] | d_loss: 0.6906 | g_loss: 1.3963\n",
      "Epoch [    5/   20] | d_loss: 0.8061 | g_loss: 1.8290\n",
      "Epoch [    5/   20] | d_loss: 0.7819 | g_loss: 2.3547\n",
      "Epoch [    5/   20] | d_loss: 0.6227 | g_loss: 1.7161\n",
      "Epoch [    5/   20] | d_loss: 1.0629 | g_loss: 3.3199\n",
      "Epoch [    5/   20] | d_loss: 0.9093 | g_loss: 1.6409\n",
      "Epoch [    5/   20] | d_loss: 1.1026 | g_loss: 1.2063\n",
      "Epoch [    5/   20] | d_loss: 0.5375 | g_loss: 1.4481\n",
      "Epoch [    5/   20] | d_loss: 0.8268 | g_loss: 1.8320\n",
      "Epoch [    5/   20] | d_loss: 0.4080 | g_loss: 3.2149\n",
      "Epoch [    5/   20] | d_loss: 0.9366 | g_loss: 0.4512\n",
      "Epoch [    5/   20] | d_loss: 0.6856 | g_loss: 1.4295\n",
      "Epoch [    5/   20] | d_loss: 0.6820 | g_loss: 1.4001\n",
      "Epoch [    5/   20] | d_loss: 0.8656 | g_loss: 0.8021\n",
      "Epoch [    5/   20] | d_loss: 0.3765 | g_loss: 2.0937\n",
      "Epoch [    5/   20] | d_loss: 0.2955 | g_loss: 2.6498\n",
      "Epoch [    5/   20] | d_loss: 0.8968 | g_loss: 2.1061\n",
      "Epoch [    5/   20] | d_loss: 0.5134 | g_loss: 2.3738\n",
      "Epoch [    5/   20] | d_loss: 0.4387 | g_loss: 1.7518\n",
      "Epoch [    5/   20] | d_loss: 0.7853 | g_loss: 1.2655\n",
      "Epoch [    5/   20] | d_loss: 0.5262 | g_loss: 1.8744\n",
      "Epoch [    5/   20] | d_loss: 0.4775 | g_loss: 1.0492\n",
      "Epoch [    5/   20] | d_loss: 0.3057 | g_loss: 2.1034\n",
      "Epoch [    5/   20] | d_loss: 0.5298 | g_loss: 2.0261\n",
      "Epoch [    5/   20] | d_loss: 0.8006 | g_loss: 1.9103\n",
      "Epoch [    5/   20] | d_loss: 0.5750 | g_loss: 2.0999\n",
      "Epoch [    5/   20] | d_loss: 0.5979 | g_loss: 1.0660\n",
      "Epoch [    5/   20] | d_loss: 0.7581 | g_loss: 2.2059\n",
      "Epoch [    5/   20] | d_loss: 0.5393 | g_loss: 2.4158\n",
      "Epoch [    5/   20] | d_loss: 0.7892 | g_loss: 2.9342\n",
      "Epoch [    5/   20] | d_loss: 0.6857 | g_loss: 3.0092\n",
      "Epoch [    5/   20] | d_loss: 0.7047 | g_loss: 2.8398\n",
      "Epoch [    5/   20] | d_loss: 0.7053 | g_loss: 1.2375\n",
      "Epoch [    5/   20] | d_loss: 0.9963 | g_loss: 1.7142\n",
      "Epoch [    5/   20] | d_loss: 0.6710 | g_loss: 2.3043\n",
      "Epoch [    5/   20] | d_loss: 0.5680 | g_loss: 1.4124\n",
      "Epoch [    5/   20] | d_loss: 0.6154 | g_loss: 1.6625\n",
      "Epoch [    5/   20] | d_loss: 0.7752 | g_loss: 2.3686\n",
      "Epoch [    5/   20] | d_loss: 0.8963 | g_loss: 1.8213\n",
      "Epoch [    5/   20] | d_loss: 0.6527 | g_loss: 1.0756\n",
      "Epoch [    5/   20] | d_loss: 0.6545 | g_loss: 2.0532\n",
      "Epoch [    5/   20] | d_loss: 0.6062 | g_loss: 2.9464\n",
      "Epoch [    5/   20] | d_loss: 0.9319 | g_loss: 2.9003\n",
      "Epoch [    5/   20] | d_loss: 0.4677 | g_loss: 1.5727\n",
      "Epoch [    5/   20] | d_loss: 0.6457 | g_loss: 3.1739\n",
      "Epoch [    5/   20] | d_loss: 0.8871 | g_loss: 1.5912\n",
      "Epoch [    5/   20] | d_loss: 0.7736 | g_loss: 2.3318\n",
      "Epoch [    5/   20] | d_loss: 0.7329 | g_loss: 2.3561\n",
      "Epoch [    5/   20] | d_loss: 0.6185 | g_loss: 1.4525\n",
      "Epoch [    5/   20] | d_loss: 0.6676 | g_loss: 1.5938\n",
      "Epoch [    5/   20] | d_loss: 0.6459 | g_loss: 2.3217\n",
      "Epoch [    5/   20] | d_loss: 0.3385 | g_loss: 2.2097\n",
      "Epoch [    5/   20] | d_loss: 0.9431 | g_loss: 2.1293\n",
      "Epoch [    5/   20] | d_loss: 0.6980 | g_loss: 2.4295\n",
      "Epoch [    5/   20] | d_loss: 0.6163 | g_loss: 1.6744\n",
      "Epoch [    5/   20] | d_loss: 0.5102 | g_loss: 2.6268\n",
      "Epoch [    5/   20] | d_loss: 0.5728 | g_loss: 1.1626\n",
      "Epoch [    5/   20] | d_loss: 1.6939 | g_loss: 1.5585\n",
      "Epoch [    5/   20] | d_loss: 1.0402 | g_loss: 1.4429\n",
      "Epoch [    5/   20] | d_loss: 0.7622 | g_loss: 2.1439\n",
      "Epoch [    5/   20] | d_loss: 0.8458 | g_loss: 0.9093\n",
      "Epoch [    5/   20] | d_loss: 0.6915 | g_loss: 1.7789\n",
      "Epoch [    5/   20] | d_loss: 0.6407 | g_loss: 2.3744\n",
      "Epoch [    5/   20] | d_loss: 0.4202 | g_loss: 1.2655\n",
      "Epoch [    5/   20] | d_loss: 0.6934 | g_loss: 1.8968\n",
      "Epoch [    5/   20] | d_loss: 0.4123 | g_loss: 1.5058\n",
      "Epoch [    5/   20] | d_loss: 0.8358 | g_loss: 2.6238\n",
      "Epoch [    5/   20] | d_loss: 0.8824 | g_loss: 2.2216\n",
      "Epoch [    5/   20] | d_loss: 0.6430 | g_loss: 1.6580\n",
      "Epoch [    5/   20] | d_loss: 0.7204 | g_loss: 1.8886\n",
      "Epoch [    5/   20] | d_loss: 0.2692 | g_loss: 2.6712\n",
      "Epoch [    5/   20] | d_loss: 0.3051 | g_loss: 1.8385\n",
      "Epoch [    5/   20] | d_loss: 1.2175 | g_loss: 1.1323\n",
      "Epoch [    5/   20] | d_loss: 0.6522 | g_loss: 1.7090\n",
      "Epoch [    5/   20] | d_loss: 0.5152 | g_loss: 1.1707\n",
      "Epoch [    5/   20] | d_loss: 0.6932 | g_loss: 2.5993\n",
      "Epoch [    5/   20] | d_loss: 0.5232 | g_loss: 2.0236\n",
      "Epoch [    5/   20] | d_loss: 0.6056 | g_loss: 1.1341\n",
      "Epoch [    5/   20] | d_loss: 0.9634 | g_loss: 1.3224\n",
      "Epoch [    5/   20] | d_loss: 0.8479 | g_loss: 1.8316\n",
      "Epoch [    5/   20] | d_loss: 0.9399 | g_loss: 2.0240\n",
      "Epoch [    5/   20] | d_loss: 0.8661 | g_loss: 2.0968\n",
      "Epoch [    5/   20] | d_loss: 1.0798 | g_loss: 1.2390\n",
      "Epoch [    5/   20] | d_loss: 0.7031 | g_loss: 2.1598\n",
      "Epoch [    5/   20] | d_loss: 0.9152 | g_loss: 0.7745\n",
      "Epoch [    5/   20] | d_loss: 0.7156 | g_loss: 2.2995\n",
      "Epoch [    5/   20] | d_loss: 0.5269 | g_loss: 2.5970\n",
      "Epoch [    5/   20] | d_loss: 0.3944 | g_loss: 2.2267\n",
      "Epoch [    5/   20] | d_loss: 1.0081 | g_loss: 3.1051\n",
      "Epoch [    5/   20] | d_loss: 0.5446 | g_loss: 2.7600\n",
      "Epoch [    5/   20] | d_loss: 0.7076 | g_loss: 3.0517\n",
      "Epoch [    5/   20] | d_loss: 1.0182 | g_loss: 1.5654\n",
      "Epoch [    5/   20] | d_loss: 0.3370 | g_loss: 1.5001\n",
      "Epoch [    5/   20] | d_loss: 0.5937 | g_loss: 1.6381\n",
      "Epoch [    5/   20] | d_loss: 0.5464 | g_loss: 1.5106\n",
      "Epoch [    5/   20] | d_loss: 0.4997 | g_loss: 1.2252\n",
      "Epoch [    5/   20] | d_loss: 0.6501 | g_loss: 1.8319\n",
      "Epoch [    5/   20] | d_loss: 1.3006 | g_loss: 2.2019\n",
      "Epoch [    5/   20] | d_loss: 0.7036 | g_loss: 1.1379\n",
      "Epoch [    6/   20] | d_loss: 0.6286 | g_loss: 2.3646\n",
      "Epoch [    6/   20] | d_loss: 0.5782 | g_loss: 2.2913\n",
      "Epoch [    6/   20] | d_loss: 0.5746 | g_loss: 2.5978\n",
      "Epoch [    6/   20] | d_loss: 1.1669 | g_loss: 1.8819\n",
      "Epoch [    6/   20] | d_loss: 0.6279 | g_loss: 1.9326\n",
      "Epoch [    6/   20] | d_loss: 0.6590 | g_loss: 2.8053\n",
      "Epoch [    6/   20] | d_loss: 0.6548 | g_loss: 1.7478\n",
      "Epoch [    6/   20] | d_loss: 0.3863 | g_loss: 2.3196\n",
      "Epoch [    6/   20] | d_loss: 0.2619 | g_loss: 2.0320\n",
      "Epoch [    6/   20] | d_loss: 0.3305 | g_loss: 1.7490\n",
      "Epoch [    6/   20] | d_loss: 0.4609 | g_loss: 2.5037\n",
      "Epoch [    6/   20] | d_loss: 0.5873 | g_loss: 2.3120\n",
      "Epoch [    6/   20] | d_loss: 0.6720 | g_loss: 2.5640\n",
      "Epoch [    6/   20] | d_loss: 0.5682 | g_loss: 1.5528\n",
      "Epoch [    6/   20] | d_loss: 0.5501 | g_loss: 2.5536\n",
      "Epoch [    6/   20] | d_loss: 1.2038 | g_loss: 2.0243\n",
      "Epoch [    6/   20] | d_loss: 0.8454 | g_loss: 2.9110\n",
      "Epoch [    6/   20] | d_loss: 0.3262 | g_loss: 2.2674\n",
      "Epoch [    6/   20] | d_loss: 0.8597 | g_loss: 2.2606\n",
      "Epoch [    6/   20] | d_loss: 0.9169 | g_loss: 2.4685\n",
      "Epoch [    6/   20] | d_loss: 1.2237 | g_loss: 0.6383\n",
      "Epoch [    6/   20] | d_loss: 0.7078 | g_loss: 1.9396\n",
      "Epoch [    6/   20] | d_loss: 0.6366 | g_loss: 2.1628\n",
      "Epoch [    6/   20] | d_loss: 0.4680 | g_loss: 1.4411\n",
      "Epoch [    6/   20] | d_loss: 0.6734 | g_loss: 2.4537\n",
      "Epoch [    6/   20] | d_loss: 1.7643 | g_loss: 3.6217\n",
      "Epoch [    6/   20] | d_loss: 0.9438 | g_loss: 3.3940\n",
      "Epoch [    6/   20] | d_loss: 0.5107 | g_loss: 1.5289\n",
      "Epoch [    6/   20] | d_loss: 0.6952 | g_loss: 2.2803\n",
      "Epoch [    6/   20] | d_loss: 0.5006 | g_loss: 2.9065\n",
      "Epoch [    6/   20] | d_loss: 0.7736 | g_loss: 1.7702\n",
      "Epoch [    6/   20] | d_loss: 0.8356 | g_loss: 1.5958\n",
      "Epoch [    6/   20] | d_loss: 1.3405 | g_loss: 1.5291\n",
      "Epoch [    6/   20] | d_loss: 0.6077 | g_loss: 2.5628\n",
      "Epoch [    6/   20] | d_loss: 0.8011 | g_loss: 2.1249\n",
      "Epoch [    6/   20] | d_loss: 0.5515 | g_loss: 1.4230\n",
      "Epoch [    6/   20] | d_loss: 0.4868 | g_loss: 2.9034\n",
      "Epoch [    6/   20] | d_loss: 0.3465 | g_loss: 1.7409\n",
      "Epoch [    6/   20] | d_loss: 1.3010 | g_loss: 1.0932\n",
      "Epoch [    6/   20] | d_loss: 0.8526 | g_loss: 2.8130\n",
      "Epoch [    6/   20] | d_loss: 0.4209 | g_loss: 2.4381\n",
      "Epoch [    6/   20] | d_loss: 0.6761 | g_loss: 2.5652\n",
      "Epoch [    6/   20] | d_loss: 0.5801 | g_loss: 1.3022\n",
      "Epoch [    6/   20] | d_loss: 0.4066 | g_loss: 1.8752\n",
      "Epoch [    6/   20] | d_loss: 0.3552 | g_loss: 3.0301\n",
      "Epoch [    6/   20] | d_loss: 1.0686 | g_loss: 1.3832\n",
      "Epoch [    6/   20] | d_loss: 0.8324 | g_loss: 3.3466\n",
      "Epoch [    6/   20] | d_loss: 1.4650 | g_loss: 1.4172\n",
      "Epoch [    6/   20] | d_loss: 0.7144 | g_loss: 2.7917\n",
      "Epoch [    6/   20] | d_loss: 0.5494 | g_loss: 2.0183\n",
      "Epoch [    6/   20] | d_loss: 0.4606 | g_loss: 3.9682\n",
      "Epoch [    6/   20] | d_loss: 0.8085 | g_loss: 2.2794\n",
      "Epoch [    6/   20] | d_loss: 0.7051 | g_loss: 1.2979\n",
      "Epoch [    6/   20] | d_loss: 0.7317 | g_loss: 2.1068\n",
      "Epoch [    6/   20] | d_loss: 0.5519 | g_loss: 1.9580\n",
      "Epoch [    6/   20] | d_loss: 0.3716 | g_loss: 1.3115\n",
      "Epoch [    6/   20] | d_loss: 0.8479 | g_loss: 2.0423\n",
      "Epoch [    6/   20] | d_loss: 0.4751 | g_loss: 3.4793\n",
      "Epoch [    6/   20] | d_loss: 0.5751 | g_loss: 2.8538\n",
      "Epoch [    6/   20] | d_loss: 0.8443 | g_loss: 2.0700\n",
      "Epoch [    6/   20] | d_loss: 1.1391 | g_loss: 2.3698\n",
      "Epoch [    6/   20] | d_loss: 0.9437 | g_loss: 1.9428\n",
      "Epoch [    6/   20] | d_loss: 0.7887 | g_loss: 3.1890\n",
      "Epoch [    6/   20] | d_loss: 1.1231 | g_loss: 2.2518\n",
      "Epoch [    6/   20] | d_loss: 0.5503 | g_loss: 1.5090\n",
      "Epoch [    6/   20] | d_loss: 0.4286 | g_loss: 1.6920\n",
      "Epoch [    6/   20] | d_loss: 1.4754 | g_loss: 1.8607\n",
      "Epoch [    6/   20] | d_loss: 1.2254 | g_loss: 1.7257\n",
      "Epoch [    6/   20] | d_loss: 1.0863 | g_loss: 1.8908\n",
      "Epoch [    6/   20] | d_loss: 0.9074 | g_loss: 0.8697\n",
      "Epoch [    6/   20] | d_loss: 0.4145 | g_loss: 2.1921\n",
      "Epoch [    6/   20] | d_loss: 0.5757 | g_loss: 3.0469\n",
      "Epoch [    6/   20] | d_loss: 0.4437 | g_loss: 2.2237\n",
      "Epoch [    6/   20] | d_loss: 0.3749 | g_loss: 0.9148\n",
      "Epoch [    6/   20] | d_loss: 0.5262 | g_loss: 1.8475\n",
      "Epoch [    6/   20] | d_loss: 0.8521 | g_loss: 2.5906\n",
      "Epoch [    6/   20] | d_loss: 0.6792 | g_loss: 2.3173\n",
      "Epoch [    6/   20] | d_loss: 0.4900 | g_loss: 3.0090\n",
      "Epoch [    6/   20] | d_loss: 0.6154 | g_loss: 1.7243\n",
      "Epoch [    6/   20] | d_loss: 1.0489 | g_loss: 0.7181\n",
      "Epoch [    6/   20] | d_loss: 0.5731 | g_loss: 3.9241\n",
      "Epoch [    6/   20] | d_loss: 0.5521 | g_loss: 2.3375\n",
      "Epoch [    6/   20] | d_loss: 0.3687 | g_loss: 3.5083\n",
      "Epoch [    6/   20] | d_loss: 1.2184 | g_loss: 0.8319\n",
      "Epoch [    6/   20] | d_loss: 1.5042 | g_loss: 1.5611\n",
      "Epoch [    6/   20] | d_loss: 0.5490 | g_loss: 2.4900\n",
      "Epoch [    6/   20] | d_loss: 0.2899 | g_loss: 2.7642\n",
      "Epoch [    6/   20] | d_loss: 0.8645 | g_loss: 3.3248\n",
      "Epoch [    6/   20] | d_loss: 0.9340 | g_loss: 1.2315\n",
      "Epoch [    6/   20] | d_loss: 1.0149 | g_loss: 2.1842\n",
      "Epoch [    6/   20] | d_loss: 0.3612 | g_loss: 1.7111\n",
      "Epoch [    6/   20] | d_loss: 1.3856 | g_loss: 0.9159\n",
      "Epoch [    6/   20] | d_loss: 0.9642 | g_loss: 0.9418\n",
      "Epoch [    6/   20] | d_loss: 1.0717 | g_loss: 1.3989\n",
      "Epoch [    6/   20] | d_loss: 0.7304 | g_loss: 2.4054\n",
      "Epoch [    6/   20] | d_loss: 0.5795 | g_loss: 3.2128\n",
      "Epoch [    6/   20] | d_loss: 0.6304 | g_loss: 3.1234\n",
      "Epoch [    6/   20] | d_loss: 0.3918 | g_loss: 2.3756\n",
      "Epoch [    6/   20] | d_loss: 0.7151 | g_loss: 3.1554\n",
      "Epoch [    6/   20] | d_loss: 0.9671 | g_loss: 1.2328\n",
      "Epoch [    6/   20] | d_loss: 0.4687 | g_loss: 1.6897\n",
      "Epoch [    6/   20] | d_loss: 0.5001 | g_loss: 3.5112\n",
      "Epoch [    6/   20] | d_loss: 0.6824 | g_loss: 2.5973\n",
      "Epoch [    6/   20] | d_loss: 0.3264 | g_loss: 1.6576\n",
      "Epoch [    6/   20] | d_loss: 0.2079 | g_loss: 3.0751\n",
      "Epoch [    6/   20] | d_loss: 0.8110 | g_loss: 2.6357\n",
      "Epoch [    6/   20] | d_loss: 0.4129 | g_loss: 2.2702\n",
      "Epoch [    6/   20] | d_loss: 0.8259 | g_loss: 2.5443\n",
      "Epoch [    6/   20] | d_loss: 0.1603 | g_loss: 2.3190\n",
      "Epoch [    6/   20] | d_loss: 0.5126 | g_loss: 1.9200\n",
      "Epoch [    6/   20] | d_loss: 1.2494 | g_loss: 1.0327\n",
      "Epoch [    6/   20] | d_loss: 0.4637 | g_loss: 1.1648\n",
      "Epoch [    6/   20] | d_loss: 0.2997 | g_loss: 3.0175\n",
      "Epoch [    7/   20] | d_loss: 0.3185 | g_loss: 3.1440\n",
      "Epoch [    7/   20] | d_loss: 1.0884 | g_loss: 1.9726\n",
      "Epoch [    7/   20] | d_loss: 1.2200 | g_loss: 1.9864\n",
      "Epoch [    7/   20] | d_loss: 0.2011 | g_loss: 2.6756\n",
      "Epoch [    7/   20] | d_loss: 0.5414 | g_loss: 2.4989\n",
      "Epoch [    7/   20] | d_loss: 0.3912 | g_loss: 1.4328\n",
      "Epoch [    7/   20] | d_loss: 0.4344 | g_loss: 2.2290\n",
      "Epoch [    7/   20] | d_loss: 0.2101 | g_loss: 3.1433\n",
      "Epoch [    7/   20] | d_loss: 1.2235 | g_loss: 2.5982\n",
      "Epoch [    7/   20] | d_loss: 0.2724 | g_loss: 3.1527\n",
      "Epoch [    7/   20] | d_loss: 0.4637 | g_loss: 2.5778\n",
      "Epoch [    7/   20] | d_loss: 1.1473 | g_loss: 2.9349\n",
      "Epoch [    7/   20] | d_loss: 0.4759 | g_loss: 1.3934\n",
      "Epoch [    7/   20] | d_loss: 0.4732 | g_loss: 3.0503\n",
      "Epoch [    7/   20] | d_loss: 0.3749 | g_loss: 3.7502\n",
      "Epoch [    7/   20] | d_loss: 1.4540 | g_loss: 0.7274\n",
      "Epoch [    7/   20] | d_loss: 0.6421 | g_loss: 0.9025\n",
      "Epoch [    7/   20] | d_loss: 0.3481 | g_loss: 1.6520\n",
      "Epoch [    7/   20] | d_loss: 0.4316 | g_loss: 3.5122\n",
      "Epoch [    7/   20] | d_loss: 0.4792 | g_loss: 1.7792\n",
      "Epoch [    7/   20] | d_loss: 0.6468 | g_loss: 2.5784\n",
      "Epoch [    7/   20] | d_loss: 0.7096 | g_loss: 2.0793\n",
      "Epoch [    7/   20] | d_loss: 0.5323 | g_loss: 1.5626\n",
      "Epoch [    7/   20] | d_loss: 0.6952 | g_loss: 1.6796\n",
      "Epoch [    7/   20] | d_loss: 1.3012 | g_loss: 0.8232\n",
      "Epoch [    7/   20] | d_loss: 0.9522 | g_loss: 2.4470\n",
      "Epoch [    7/   20] | d_loss: 0.4734 | g_loss: 1.9695\n",
      "Epoch [    7/   20] | d_loss: 0.4727 | g_loss: 1.4363\n",
      "Epoch [    7/   20] | d_loss: 0.8799 | g_loss: 2.0756\n",
      "Epoch [    7/   20] | d_loss: 0.5379 | g_loss: 2.9475\n",
      "Epoch [    7/   20] | d_loss: 0.3956 | g_loss: 3.1555\n",
      "Epoch [    7/   20] | d_loss: 0.5860 | g_loss: 1.9249\n",
      "Epoch [    7/   20] | d_loss: 0.4369 | g_loss: 3.0385\n",
      "Epoch [    7/   20] | d_loss: 0.4029 | g_loss: 1.7637\n",
      "Epoch [    7/   20] | d_loss: 0.4806 | g_loss: 1.5192\n",
      "Epoch [    7/   20] | d_loss: 0.4569 | g_loss: 0.7131\n",
      "Epoch [    7/   20] | d_loss: 0.6442 | g_loss: 2.0848\n",
      "Epoch [    7/   20] | d_loss: 1.1109 | g_loss: 4.3732\n",
      "Epoch [    7/   20] | d_loss: 0.6085 | g_loss: 1.4523\n",
      "Epoch [    7/   20] | d_loss: 0.5553 | g_loss: 1.2030\n",
      "Epoch [    7/   20] | d_loss: 0.4688 | g_loss: 2.5677\n",
      "Epoch [    7/   20] | d_loss: 0.5809 | g_loss: 1.8818\n",
      "Epoch [    7/   20] | d_loss: 0.4079 | g_loss: 3.2033\n",
      "Epoch [    7/   20] | d_loss: 1.0286 | g_loss: 3.3878\n",
      "Epoch [    7/   20] | d_loss: 0.9459 | g_loss: 4.0727\n",
      "Epoch [    7/   20] | d_loss: 0.6759 | g_loss: 2.5785\n",
      "Epoch [    7/   20] | d_loss: 0.6270 | g_loss: 1.5129\n",
      "Epoch [    7/   20] | d_loss: 0.8505 | g_loss: 1.8814\n",
      "Epoch [    7/   20] | d_loss: 0.3418 | g_loss: 3.4320\n",
      "Epoch [    7/   20] | d_loss: 0.9471 | g_loss: 3.7407\n",
      "Epoch [    7/   20] | d_loss: 0.6355 | g_loss: 2.2508\n",
      "Epoch [    7/   20] | d_loss: 0.3334 | g_loss: 3.6225\n",
      "Epoch [    7/   20] | d_loss: 1.4642 | g_loss: 0.5570\n",
      "Epoch [    7/   20] | d_loss: 0.8436 | g_loss: 2.0541\n",
      "Epoch [    7/   20] | d_loss: 0.8849 | g_loss: 0.6699\n",
      "Epoch [    7/   20] | d_loss: 0.0958 | g_loss: 4.8318\n",
      "Epoch [    7/   20] | d_loss: 0.5897 | g_loss: 2.0830\n",
      "Epoch [    7/   20] | d_loss: 1.1062 | g_loss: 2.4654\n",
      "Epoch [    7/   20] | d_loss: 0.5463 | g_loss: 2.6268\n",
      "Epoch [    7/   20] | d_loss: 0.4880 | g_loss: 2.8655\n",
      "Epoch [    7/   20] | d_loss: 0.3685 | g_loss: 2.3414\n",
      "Epoch [    7/   20] | d_loss: 0.2303 | g_loss: 4.2323\n",
      "Epoch [    7/   20] | d_loss: 0.6743 | g_loss: 2.2473\n",
      "Epoch [    7/   20] | d_loss: 0.5033 | g_loss: 1.9285\n",
      "Epoch [    7/   20] | d_loss: 0.5317 | g_loss: 3.0303\n",
      "Epoch [    7/   20] | d_loss: 0.3177 | g_loss: 1.4515\n",
      "Epoch [    7/   20] | d_loss: 0.6655 | g_loss: 2.3635\n",
      "Epoch [    7/   20] | d_loss: 0.7652 | g_loss: 1.4136\n",
      "Epoch [    7/   20] | d_loss: 0.3068 | g_loss: 3.6382\n",
      "Epoch [    7/   20] | d_loss: 0.9263 | g_loss: 1.0274\n",
      "Epoch [    7/   20] | d_loss: 0.4538 | g_loss: 2.2119\n",
      "Epoch [    7/   20] | d_loss: 0.9785 | g_loss: 2.3416\n",
      "Epoch [    7/   20] | d_loss: 0.1284 | g_loss: 3.2726\n",
      "Epoch [    7/   20] | d_loss: 0.7457 | g_loss: 3.4599\n",
      "Epoch [    7/   20] | d_loss: 0.2529 | g_loss: 1.9799\n",
      "Epoch [    7/   20] | d_loss: 0.1908 | g_loss: 2.0237\n",
      "Epoch [    7/   20] | d_loss: 0.4534 | g_loss: 2.7233\n",
      "Epoch [    7/   20] | d_loss: 0.4839 | g_loss: 3.4965\n",
      "Epoch [    7/   20] | d_loss: 0.9886 | g_loss: 4.4312\n",
      "Epoch [    7/   20] | d_loss: 0.7460 | g_loss: 1.9869\n",
      "Epoch [    7/   20] | d_loss: 0.4854 | g_loss: 2.2096\n",
      "Epoch [    7/   20] | d_loss: 0.3504 | g_loss: 1.9228\n",
      "Epoch [    7/   20] | d_loss: 0.1443 | g_loss: 2.3784\n",
      "Epoch [    7/   20] | d_loss: 0.6043 | g_loss: 2.3427\n",
      "Epoch [    7/   20] | d_loss: 1.2325 | g_loss: 3.7429\n",
      "Epoch [    7/   20] | d_loss: 0.2064 | g_loss: 2.4201\n",
      "Epoch [    7/   20] | d_loss: 0.5500 | g_loss: 2.2828\n",
      "Epoch [    7/   20] | d_loss: 0.2565 | g_loss: 1.1424\n",
      "Epoch [    7/   20] | d_loss: 0.3961 | g_loss: 2.0740\n",
      "Epoch [    7/   20] | d_loss: 0.8193 | g_loss: 1.8837\n",
      "Epoch [    7/   20] | d_loss: 0.2711 | g_loss: 3.6494\n",
      "Epoch [    7/   20] | d_loss: 0.5560 | g_loss: 1.8927\n",
      "Epoch [    7/   20] | d_loss: 0.9326 | g_loss: 2.1127\n",
      "Epoch [    7/   20] | d_loss: 0.6880 | g_loss: 0.8255\n",
      "Epoch [    7/   20] | d_loss: 0.6537 | g_loss: 2.5526\n",
      "Epoch [    7/   20] | d_loss: 0.3889 | g_loss: 1.5060\n",
      "Epoch [    7/   20] | d_loss: 0.8719 | g_loss: 2.2616\n",
      "Epoch [    7/   20] | d_loss: 0.7800 | g_loss: 2.8147\n",
      "Epoch [    7/   20] | d_loss: 0.3462 | g_loss: 1.6301\n",
      "Epoch [    7/   20] | d_loss: 0.3134 | g_loss: 3.5288\n",
      "Epoch [    7/   20] | d_loss: 0.2711 | g_loss: 2.1535\n",
      "Epoch [    7/   20] | d_loss: 0.6865 | g_loss: 2.8842\n",
      "Epoch [    7/   20] | d_loss: 0.5225 | g_loss: 1.9548\n",
      "Epoch [    7/   20] | d_loss: 0.7034 | g_loss: 1.1369\n",
      "Epoch [    7/   20] | d_loss: 1.2108 | g_loss: 2.2171\n",
      "Epoch [    7/   20] | d_loss: 0.8568 | g_loss: 2.0962\n",
      "Epoch [    7/   20] | d_loss: 0.4163 | g_loss: 3.1514\n",
      "Epoch [    7/   20] | d_loss: 0.4438 | g_loss: 1.4448\n",
      "Epoch [    7/   20] | d_loss: 0.6753 | g_loss: 2.2215\n",
      "Epoch [    7/   20] | d_loss: 0.3328 | g_loss: 2.8005\n",
      "Epoch [    7/   20] | d_loss: 0.9338 | g_loss: 2.9418\n",
      "Epoch [    7/   20] | d_loss: 0.1753 | g_loss: 2.8579\n",
      "Epoch [    7/   20] | d_loss: 0.5733 | g_loss: 1.1335\n",
      "Epoch [    8/   20] | d_loss: 0.2716 | g_loss: 2.3648\n",
      "Epoch [    8/   20] | d_loss: 0.2394 | g_loss: 2.9877\n",
      "Epoch [    8/   20] | d_loss: 0.1255 | g_loss: 3.2871\n",
      "Epoch [    8/   20] | d_loss: 0.5243 | g_loss: 1.0110\n",
      "Epoch [    8/   20] | d_loss: 0.2202 | g_loss: 1.9604\n",
      "Epoch [    8/   20] | d_loss: 0.4327 | g_loss: 2.1677\n",
      "Epoch [    8/   20] | d_loss: 0.6165 | g_loss: 4.2027\n",
      "Epoch [    8/   20] | d_loss: 0.6621 | g_loss: 2.6435\n",
      "Epoch [    8/   20] | d_loss: 0.2994 | g_loss: 2.7173\n",
      "Epoch [    8/   20] | d_loss: 0.5395 | g_loss: 2.1070\n",
      "Epoch [    8/   20] | d_loss: 0.5626 | g_loss: 4.0950\n",
      "Epoch [    8/   20] | d_loss: 0.3591 | g_loss: 2.9391\n",
      "Epoch [    8/   20] | d_loss: 0.2959 | g_loss: 2.1905\n",
      "Epoch [    8/   20] | d_loss: 0.6060 | g_loss: 3.0365\n",
      "Epoch [    8/   20] | d_loss: 0.6056 | g_loss: 0.6505\n",
      "Epoch [    8/   20] | d_loss: 0.5754 | g_loss: 1.4427\n",
      "Epoch [    8/   20] | d_loss: 0.8347 | g_loss: 1.5290\n",
      "Epoch [    8/   20] | d_loss: 0.5425 | g_loss: 2.0525\n",
      "Epoch [    8/   20] | d_loss: 1.0642 | g_loss: 1.0410\n",
      "Epoch [    8/   20] | d_loss: 0.3042 | g_loss: 2.3431\n",
      "Epoch [    8/   20] | d_loss: 1.0906 | g_loss: 3.2700\n",
      "Epoch [    8/   20] | d_loss: 0.4055 | g_loss: 2.4360\n",
      "Epoch [    8/   20] | d_loss: 0.5278 | g_loss: 2.6295\n",
      "Epoch [    8/   20] | d_loss: 0.4649 | g_loss: 2.0600\n",
      "Epoch [    8/   20] | d_loss: 0.7409 | g_loss: 1.8701\n",
      "Epoch [    8/   20] | d_loss: 0.2547 | g_loss: 3.2603\n",
      "Epoch [    8/   20] | d_loss: 0.9324 | g_loss: 2.3107\n",
      "Epoch [    8/   20] | d_loss: 0.3788 | g_loss: 3.6590\n",
      "Epoch [    8/   20] | d_loss: 0.8907 | g_loss: 1.9694\n",
      "Epoch [    8/   20] | d_loss: 0.2653 | g_loss: 2.5508\n",
      "Epoch [    8/   20] | d_loss: 0.5107 | g_loss: 0.5063\n",
      "Epoch [    8/   20] | d_loss: 0.3012 | g_loss: 3.7100\n",
      "Epoch [    8/   20] | d_loss: 0.4462 | g_loss: 3.6636\n",
      "Epoch [    8/   20] | d_loss: 0.4011 | g_loss: 1.7619\n",
      "Epoch [    8/   20] | d_loss: 0.6824 | g_loss: 2.4157\n",
      "Epoch [    8/   20] | d_loss: 0.3741 | g_loss: 1.5713\n",
      "Epoch [    8/   20] | d_loss: 0.5317 | g_loss: 2.6647\n",
      "Epoch [    8/   20] | d_loss: 1.0151 | g_loss: 1.0020\n",
      "Epoch [    8/   20] | d_loss: 0.0685 | g_loss: 2.4033\n",
      "Epoch [    8/   20] | d_loss: 0.4851 | g_loss: 2.1705\n",
      "Epoch [    8/   20] | d_loss: 0.3658 | g_loss: 3.2743\n",
      "Epoch [    8/   20] | d_loss: 0.6893 | g_loss: 2.9148\n",
      "Epoch [    8/   20] | d_loss: 0.8964 | g_loss: 0.8015\n",
      "Epoch [    8/   20] | d_loss: 0.3566 | g_loss: 3.6518\n",
      "Epoch [    8/   20] | d_loss: 0.5494 | g_loss: 2.3795\n",
      "Epoch [    8/   20] | d_loss: 0.5527 | g_loss: 3.1018\n",
      "Epoch [    8/   20] | d_loss: 0.5922 | g_loss: 1.7728\n",
      "Epoch [    8/   20] | d_loss: 0.4310 | g_loss: 3.8940\n",
      "Epoch [    8/   20] | d_loss: 0.4546 | g_loss: 3.3209\n",
      "Epoch [    8/   20] | d_loss: 0.2772 | g_loss: 2.4808\n",
      "Epoch [    8/   20] | d_loss: 0.8261 | g_loss: 1.7325\n",
      "Epoch [    8/   20] | d_loss: 0.8238 | g_loss: 2.2490\n",
      "Epoch [    8/   20] | d_loss: 0.1868 | g_loss: 4.5551\n",
      "Epoch [    8/   20] | d_loss: 0.6868 | g_loss: 3.1066\n",
      "Epoch [    8/   20] | d_loss: 0.4501 | g_loss: 2.0389\n",
      "Epoch [    8/   20] | d_loss: 0.6329 | g_loss: 4.8129\n",
      "Epoch [    8/   20] | d_loss: 0.9927 | g_loss: 2.2279\n",
      "Epoch [    8/   20] | d_loss: 0.3434 | g_loss: 2.6461\n",
      "Epoch [    8/   20] | d_loss: 1.4120 | g_loss: 1.1695\n",
      "Epoch [    8/   20] | d_loss: 1.5678 | g_loss: 2.4450\n",
      "Epoch [    8/   20] | d_loss: 0.4183 | g_loss: 3.2718\n",
      "Epoch [    8/   20] | d_loss: 0.5367 | g_loss: 1.5200\n",
      "Epoch [    8/   20] | d_loss: 1.0597 | g_loss: 1.6041\n",
      "Epoch [    8/   20] | d_loss: 0.5258 | g_loss: 3.3499\n",
      "Epoch [    8/   20] | d_loss: 0.8026 | g_loss: 2.5008\n",
      "Epoch [    8/   20] | d_loss: 0.4904 | g_loss: 2.4570\n",
      "Epoch [    8/   20] | d_loss: 0.5745 | g_loss: 3.9548\n",
      "Epoch [    8/   20] | d_loss: 0.1582 | g_loss: 2.4827\n",
      "Epoch [    8/   20] | d_loss: 0.8146 | g_loss: 2.0839\n",
      "Epoch [    8/   20] | d_loss: 0.5346 | g_loss: 1.7872\n",
      "Epoch [    8/   20] | d_loss: 0.9356 | g_loss: 2.3405\n",
      "Epoch [    8/   20] | d_loss: 0.9648 | g_loss: 1.1613\n",
      "Epoch [    8/   20] | d_loss: 1.0167 | g_loss: 2.4819\n",
      "Epoch [    8/   20] | d_loss: 0.9654 | g_loss: 1.7985\n",
      "Epoch [    8/   20] | d_loss: 0.3588 | g_loss: 3.7332\n",
      "Epoch [    8/   20] | d_loss: 0.7009 | g_loss: 3.0617\n",
      "Epoch [    8/   20] | d_loss: 0.2566 | g_loss: 3.4964\n",
      "Epoch [    8/   20] | d_loss: 0.6201 | g_loss: 2.9333\n",
      "Epoch [    8/   20] | d_loss: 0.5456 | g_loss: 1.3001\n",
      "Epoch [    8/   20] | d_loss: 0.2495 | g_loss: 4.2464\n",
      "Epoch [    8/   20] | d_loss: 0.3601 | g_loss: 2.8353\n",
      "Epoch [    8/   20] | d_loss: 0.6679 | g_loss: 2.5102\n",
      "Epoch [    8/   20] | d_loss: 0.8156 | g_loss: 2.6748\n",
      "Epoch [    8/   20] | d_loss: 0.3842 | g_loss: 1.5930\n",
      "Epoch [    8/   20] | d_loss: 0.1584 | g_loss: 2.5857\n",
      "Epoch [    8/   20] | d_loss: 0.6924 | g_loss: 1.4500\n",
      "Epoch [    8/   20] | d_loss: 0.6598 | g_loss: 3.0289\n",
      "Epoch [    8/   20] | d_loss: 0.4308 | g_loss: 2.0904\n",
      "Epoch [    8/   20] | d_loss: 0.8208 | g_loss: 2.1510\n",
      "Epoch [    8/   20] | d_loss: 0.5859 | g_loss: 3.9524\n",
      "Epoch [    8/   20] | d_loss: 0.5898 | g_loss: 2.7251\n",
      "Epoch [    8/   20] | d_loss: 0.3131 | g_loss: 2.5466\n",
      "Epoch [    8/   20] | d_loss: 0.6809 | g_loss: 2.4343\n",
      "Epoch [    8/   20] | d_loss: 0.7854 | g_loss: 2.2688\n",
      "Epoch [    8/   20] | d_loss: 0.7531 | g_loss: 1.5330\n",
      "Epoch [    8/   20] | d_loss: 0.5086 | g_loss: 3.7746\n",
      "Epoch [    8/   20] | d_loss: 0.5366 | g_loss: 2.8246\n",
      "Epoch [    8/   20] | d_loss: 0.4404 | g_loss: 2.4588\n",
      "Epoch [    8/   20] | d_loss: 0.2175 | g_loss: 2.0546\n",
      "Epoch [    8/   20] | d_loss: 0.6645 | g_loss: 2.2573\n",
      "Epoch [    8/   20] | d_loss: 0.8789 | g_loss: 3.3170\n",
      "Epoch [    8/   20] | d_loss: 0.4211 | g_loss: 3.3998\n",
      "Epoch [    8/   20] | d_loss: 0.6537 | g_loss: 1.0024\n",
      "Epoch [    8/   20] | d_loss: 0.3108 | g_loss: 2.9588\n",
      "Epoch [    8/   20] | d_loss: 0.3658 | g_loss: 3.0228\n",
      "Epoch [    8/   20] | d_loss: 0.4552 | g_loss: 3.5083\n",
      "Epoch [    8/   20] | d_loss: 0.8409 | g_loss: 2.3513\n",
      "Epoch [    8/   20] | d_loss: 0.3837 | g_loss: 2.0518\n",
      "Epoch [    8/   20] | d_loss: 0.5497 | g_loss: 1.4573\n",
      "Epoch [    8/   20] | d_loss: 0.1129 | g_loss: 2.7302\n",
      "Epoch [    8/   20] | d_loss: 0.6164 | g_loss: 1.9639\n",
      "Epoch [    8/   20] | d_loss: 0.3645 | g_loss: 3.4954\n",
      "Epoch [    8/   20] | d_loss: 0.6859 | g_loss: 1.8480\n",
      "Epoch [    9/   20] | d_loss: 0.7571 | g_loss: 2.3399\n",
      "Epoch [    9/   20] | d_loss: 0.2790 | g_loss: 3.1188\n",
      "Epoch [    9/   20] | d_loss: 0.4174 | g_loss: 2.1881\n",
      "Epoch [    9/   20] | d_loss: 0.2558 | g_loss: 3.6654\n",
      "Epoch [    9/   20] | d_loss: 0.6052 | g_loss: 2.0602\n",
      "Epoch [    9/   20] | d_loss: 0.3439 | g_loss: 3.0681\n",
      "Epoch [    9/   20] | d_loss: 0.2858 | g_loss: 4.1072\n",
      "Epoch [    9/   20] | d_loss: 0.3390 | g_loss: 1.8100\n",
      "Epoch [    9/   20] | d_loss: 0.7031 | g_loss: 1.2437\n",
      "Epoch [    9/   20] | d_loss: 0.2112 | g_loss: 2.4800\n",
      "Epoch [    9/   20] | d_loss: 0.8357 | g_loss: 3.1858\n",
      "Epoch [    9/   20] | d_loss: 0.6637 | g_loss: 1.2015\n",
      "Epoch [    9/   20] | d_loss: 0.4438 | g_loss: 1.5257\n",
      "Epoch [    9/   20] | d_loss: 0.4129 | g_loss: 2.9047\n",
      "Epoch [    9/   20] | d_loss: 0.6924 | g_loss: 2.1557\n",
      "Epoch [    9/   20] | d_loss: 1.0715 | g_loss: 1.0364\n",
      "Epoch [    9/   20] | d_loss: 0.5828 | g_loss: 2.2737\n",
      "Epoch [    9/   20] | d_loss: 0.1643 | g_loss: 1.7594\n",
      "Epoch [    9/   20] | d_loss: 1.1879 | g_loss: 4.7614\n",
      "Epoch [    9/   20] | d_loss: 0.1556 | g_loss: 2.2247\n",
      "Epoch [    9/   20] | d_loss: 0.7911 | g_loss: 2.5476\n",
      "Epoch [    9/   20] | d_loss: 1.0231 | g_loss: 1.7727\n",
      "Epoch [    9/   20] | d_loss: 0.2600 | g_loss: 2.1373\n",
      "Epoch [    9/   20] | d_loss: 0.5500 | g_loss: 1.4825\n",
      "Epoch [    9/   20] | d_loss: 0.2963 | g_loss: 3.2831\n",
      "Epoch [    9/   20] | d_loss: 1.0379 | g_loss: 3.1680\n",
      "Epoch [    9/   20] | d_loss: 0.4114 | g_loss: 0.8967\n",
      "Epoch [    9/   20] | d_loss: 0.2735 | g_loss: 3.9160\n",
      "Epoch [    9/   20] | d_loss: 0.4194 | g_loss: 4.4019\n",
      "Epoch [    9/   20] | d_loss: 0.6105 | g_loss: 3.1351\n",
      "Epoch [    9/   20] | d_loss: 0.2405 | g_loss: 4.6779\n",
      "Epoch [    9/   20] | d_loss: 0.1111 | g_loss: 2.2916\n",
      "Epoch [    9/   20] | d_loss: 0.5613 | g_loss: 2.5521\n",
      "Epoch [    9/   20] | d_loss: 1.8136 | g_loss: 1.7330\n",
      "Epoch [    9/   20] | d_loss: 0.3393 | g_loss: 2.5786\n",
      "Epoch [    9/   20] | d_loss: 0.7559 | g_loss: 1.9401\n",
      "Epoch [    9/   20] | d_loss: 0.7652 | g_loss: 1.7166\n",
      "Epoch [    9/   20] | d_loss: 0.6109 | g_loss: 3.0786\n",
      "Epoch [    9/   20] | d_loss: 0.2620 | g_loss: 2.2976\n",
      "Epoch [    9/   20] | d_loss: 1.4291 | g_loss: 2.9920\n",
      "Epoch [    9/   20] | d_loss: 0.2626 | g_loss: 2.1063\n",
      "Epoch [    9/   20] | d_loss: 0.2214 | g_loss: 1.4302\n",
      "Epoch [    9/   20] | d_loss: 0.7814 | g_loss: 1.7380\n",
      "Epoch [    9/   20] | d_loss: 0.2772 | g_loss: 3.4138\n",
      "Epoch [    9/   20] | d_loss: 0.6253 | g_loss: 3.2446\n",
      "Epoch [    9/   20] | d_loss: 0.4909 | g_loss: 5.0731\n",
      "Epoch [    9/   20] | d_loss: 0.1805 | g_loss: 3.7975\n",
      "Epoch [    9/   20] | d_loss: 0.3861 | g_loss: 2.9762\n",
      "Epoch [    9/   20] | d_loss: 0.6035 | g_loss: 3.7165\n",
      "Epoch [    9/   20] | d_loss: 0.6940 | g_loss: 1.7840\n",
      "Epoch [    9/   20] | d_loss: 0.5003 | g_loss: 3.0363\n",
      "Epoch [    9/   20] | d_loss: 0.1314 | g_loss: 2.5658\n",
      "Epoch [    9/   20] | d_loss: 0.6023 | g_loss: 2.6398\n",
      "Epoch [    9/   20] | d_loss: 0.4126 | g_loss: 3.9531\n",
      "Epoch [    9/   20] | d_loss: 0.3711 | g_loss: 3.3242\n",
      "Epoch [    9/   20] | d_loss: 0.8681 | g_loss: 2.5412\n",
      "Epoch [    9/   20] | d_loss: 0.0936 | g_loss: 2.6311\n",
      "Epoch [    9/   20] | d_loss: 0.4244 | g_loss: 3.3363\n",
      "Epoch [    9/   20] | d_loss: 0.2376 | g_loss: 2.7156\n",
      "Epoch [    9/   20] | d_loss: 0.3962 | g_loss: 3.0223\n",
      "Epoch [    9/   20] | d_loss: 0.6290 | g_loss: 3.2891\n",
      "Epoch [    9/   20] | d_loss: 0.4390 | g_loss: 2.7633\n",
      "Epoch [    9/   20] | d_loss: 0.3393 | g_loss: 2.6327\n",
      "Epoch [    9/   20] | d_loss: 0.8694 | g_loss: 1.8585\n",
      "Epoch [    9/   20] | d_loss: 0.3508 | g_loss: 3.1713\n",
      "Epoch [    9/   20] | d_loss: 0.8379 | g_loss: 1.1310\n",
      "Epoch [    9/   20] | d_loss: 0.2913 | g_loss: 1.4191\n",
      "Epoch [    9/   20] | d_loss: 0.6200 | g_loss: 2.8816\n",
      "Epoch [    9/   20] | d_loss: 1.0407 | g_loss: 1.1866\n",
      "Epoch [    9/   20] | d_loss: 0.4747 | g_loss: 3.4388\n",
      "Epoch [    9/   20] | d_loss: 0.9793 | g_loss: 2.6750\n",
      "Epoch [    9/   20] | d_loss: 0.4132 | g_loss: 3.9722\n",
      "Epoch [    9/   20] | d_loss: 0.4068 | g_loss: 3.3895\n",
      "Epoch [    9/   20] | d_loss: 0.4562 | g_loss: 2.7868\n",
      "Epoch [    9/   20] | d_loss: 0.6820 | g_loss: 3.2910\n",
      "Epoch [    9/   20] | d_loss: 1.0435 | g_loss: 1.6231\n",
      "Epoch [    9/   20] | d_loss: 0.3307 | g_loss: 2.3718\n",
      "Epoch [    9/   20] | d_loss: 0.3392 | g_loss: 3.1151\n",
      "Epoch [    9/   20] | d_loss: 0.4349 | g_loss: 4.2200\n",
      "Epoch [    9/   20] | d_loss: 0.6203 | g_loss: 2.6163\n",
      "Epoch [    9/   20] | d_loss: 0.3085 | g_loss: 2.6100\n",
      "Epoch [    9/   20] | d_loss: 0.1772 | g_loss: 1.0058\n",
      "Epoch [    9/   20] | d_loss: 0.3836 | g_loss: 2.3073\n",
      "Epoch [    9/   20] | d_loss: 0.2850 | g_loss: 1.8775\n",
      "Epoch [    9/   20] | d_loss: 0.8660 | g_loss: 2.2974\n",
      "Epoch [    9/   20] | d_loss: 0.5821 | g_loss: 1.7087\n",
      "Epoch [    9/   20] | d_loss: 0.5562 | g_loss: 3.0479\n",
      "Epoch [    9/   20] | d_loss: 0.8219 | g_loss: 2.8849\n",
      "Epoch [    9/   20] | d_loss: 0.3972 | g_loss: 2.6092\n",
      "Epoch [    9/   20] | d_loss: 0.3984 | g_loss: 2.6709\n",
      "Epoch [    9/   20] | d_loss: 0.2474 | g_loss: 3.1100\n",
      "Epoch [    9/   20] | d_loss: 0.2796 | g_loss: 2.9767\n",
      "Epoch [    9/   20] | d_loss: 0.4632 | g_loss: 2.4292\n",
      "Epoch [    9/   20] | d_loss: 0.3751 | g_loss: 1.9784\n",
      "Epoch [    9/   20] | d_loss: 0.2558 | g_loss: 3.6304\n",
      "Epoch [    9/   20] | d_loss: 0.4780 | g_loss: 3.6236\n",
      "Epoch [    9/   20] | d_loss: 0.2408 | g_loss: 2.5893\n",
      "Epoch [    9/   20] | d_loss: 0.8654 | g_loss: 2.7032\n",
      "Epoch [    9/   20] | d_loss: 0.6925 | g_loss: 1.7322\n",
      "Epoch [    9/   20] | d_loss: 0.4236 | g_loss: 3.0437\n",
      "Epoch [    9/   20] | d_loss: 0.3283 | g_loss: 3.0949\n",
      "Epoch [    9/   20] | d_loss: 0.3482 | g_loss: 2.6347\n",
      "Epoch [    9/   20] | d_loss: 0.2141 | g_loss: 3.0115\n",
      "Epoch [    9/   20] | d_loss: 1.3315 | g_loss: 1.6740\n",
      "Epoch [    9/   20] | d_loss: 1.4434 | g_loss: 3.9476\n",
      "Epoch [    9/   20] | d_loss: 1.0824 | g_loss: 1.0611\n",
      "Epoch [    9/   20] | d_loss: 0.4189 | g_loss: 1.9291\n",
      "Epoch [    9/   20] | d_loss: 0.1239 | g_loss: 5.9181\n",
      "Epoch [    9/   20] | d_loss: 0.6950 | g_loss: 3.8676\n",
      "Epoch [    9/   20] | d_loss: 0.5063 | g_loss: 1.6343\n",
      "Epoch [    9/   20] | d_loss: 0.4206 | g_loss: 2.3776\n",
      "Epoch [    9/   20] | d_loss: 1.9759 | g_loss: 0.8133\n",
      "Epoch [    9/   20] | d_loss: 0.9053 | g_loss: 2.0364\n",
      "Epoch [   10/   20] | d_loss: 0.5560 | g_loss: 2.3290\n",
      "Epoch [   10/   20] | d_loss: 0.2435 | g_loss: 2.9923\n",
      "Epoch [   10/   20] | d_loss: 0.3469 | g_loss: 2.9959\n",
      "Epoch [   10/   20] | d_loss: 0.8836 | g_loss: 1.8388\n",
      "Epoch [   10/   20] | d_loss: 0.6937 | g_loss: 3.4912\n",
      "Epoch [   10/   20] | d_loss: 0.1308 | g_loss: 3.3047\n",
      "Epoch [   10/   20] | d_loss: 0.1025 | g_loss: 3.0384\n",
      "Epoch [   10/   20] | d_loss: 0.3583 | g_loss: 3.2360\n",
      "Epoch [   10/   20] | d_loss: 0.2031 | g_loss: 5.2700\n",
      "Epoch [   10/   20] | d_loss: 0.6281 | g_loss: 2.3606\n",
      "Epoch [   10/   20] | d_loss: 0.4444 | g_loss: 1.1235\n",
      "Epoch [   10/   20] | d_loss: 0.3636 | g_loss: 3.4812\n",
      "Epoch [   10/   20] | d_loss: 0.8433 | g_loss: 4.0042\n",
      "Epoch [   10/   20] | d_loss: 0.4918 | g_loss: 2.4604\n",
      "Epoch [   10/   20] | d_loss: 0.3413 | g_loss: 2.0862\n",
      "Epoch [   10/   20] | d_loss: 0.9792 | g_loss: 3.2700\n",
      "Epoch [   10/   20] | d_loss: 0.4233 | g_loss: 2.4998\n",
      "Epoch [   10/   20] | d_loss: 1.3136 | g_loss: 1.7847\n",
      "Epoch [   10/   20] | d_loss: 0.2475 | g_loss: 2.8947\n",
      "Epoch [   10/   20] | d_loss: 0.3713 | g_loss: 2.1894\n",
      "Epoch [   10/   20] | d_loss: 0.3995 | g_loss: 2.9155\n",
      "Epoch [   10/   20] | d_loss: 0.6671 | g_loss: 1.6057\n",
      "Epoch [   10/   20] | d_loss: 0.4838 | g_loss: 1.1925\n",
      "Epoch [   10/   20] | d_loss: 0.4662 | g_loss: 2.7291\n",
      "Epoch [   10/   20] | d_loss: 1.1450 | g_loss: 2.8197\n",
      "Epoch [   10/   20] | d_loss: 0.4008 | g_loss: 2.2925\n",
      "Epoch [   10/   20] | d_loss: 1.0537 | g_loss: 2.6058\n",
      "Epoch [   10/   20] | d_loss: 0.1367 | g_loss: 2.5053\n",
      "Epoch [   10/   20] | d_loss: 0.2324 | g_loss: 3.7206\n",
      "Epoch [   10/   20] | d_loss: 0.7819 | g_loss: 1.7826\n",
      "Epoch [   10/   20] | d_loss: 0.4175 | g_loss: 2.2946\n",
      "Epoch [   10/   20] | d_loss: 0.7830 | g_loss: 2.8147\n",
      "Epoch [   10/   20] | d_loss: 0.3472 | g_loss: 3.0451\n",
      "Epoch [   10/   20] | d_loss: 0.3247 | g_loss: 4.5306\n",
      "Epoch [   10/   20] | d_loss: 0.6365 | g_loss: 2.9057\n",
      "Epoch [   10/   20] | d_loss: 1.0362 | g_loss: 3.5477\n",
      "Epoch [   10/   20] | d_loss: 0.4340 | g_loss: 4.5345\n",
      "Epoch [   10/   20] | d_loss: 0.6313 | g_loss: 1.7910\n",
      "Epoch [   10/   20] | d_loss: 0.3438 | g_loss: 2.3135\n",
      "Epoch [   10/   20] | d_loss: 0.6933 | g_loss: 2.4995\n",
      "Epoch [   10/   20] | d_loss: 0.4639 | g_loss: 4.9997\n",
      "Epoch [   10/   20] | d_loss: 1.1680 | g_loss: 2.7525\n",
      "Epoch [   10/   20] | d_loss: 0.4383 | g_loss: 1.9668\n",
      "Epoch [   10/   20] | d_loss: 0.3260 | g_loss: 2.4003\n",
      "Epoch [   10/   20] | d_loss: 0.3856 | g_loss: 0.6852\n",
      "Epoch [   10/   20] | d_loss: 0.4592 | g_loss: 2.3675\n",
      "Epoch [   10/   20] | d_loss: 0.4532 | g_loss: 4.8669\n",
      "Epoch [   10/   20] | d_loss: 0.2253 | g_loss: 2.8302\n",
      "Epoch [   10/   20] | d_loss: 0.5326 | g_loss: 2.3699\n",
      "Epoch [   10/   20] | d_loss: 0.4851 | g_loss: 3.4943\n",
      "Epoch [   10/   20] | d_loss: 0.3547 | g_loss: 1.5797\n",
      "Epoch [   10/   20] | d_loss: 0.2851 | g_loss: 2.2806\n",
      "Epoch [   10/   20] | d_loss: 0.3539 | g_loss: 1.7941\n",
      "Epoch [   10/   20] | d_loss: 0.5694 | g_loss: 1.8957\n",
      "Epoch [   10/   20] | d_loss: 0.4591 | g_loss: 3.1174\n",
      "Epoch [   10/   20] | d_loss: 0.3730 | g_loss: 3.2241\n",
      "Epoch [   10/   20] | d_loss: 0.3483 | g_loss: 3.2057\n",
      "Epoch [   10/   20] | d_loss: 0.4813 | g_loss: 3.2193\n",
      "Epoch [   10/   20] | d_loss: 0.6159 | g_loss: 4.6406\n",
      "Epoch [   10/   20] | d_loss: 0.4271 | g_loss: 1.4679\n",
      "Epoch [   10/   20] | d_loss: 0.7309 | g_loss: 2.9356\n",
      "Epoch [   10/   20] | d_loss: 0.8351 | g_loss: 2.0198\n",
      "Epoch [   10/   20] | d_loss: 0.3569 | g_loss: 2.9897\n",
      "Epoch [   10/   20] | d_loss: 0.2735 | g_loss: 1.1696\n",
      "Epoch [   10/   20] | d_loss: 0.4553 | g_loss: 2.7996\n",
      "Epoch [   10/   20] | d_loss: 0.6934 | g_loss: 1.1416\n",
      "Epoch [   10/   20] | d_loss: 0.3956 | g_loss: 2.3125\n",
      "Epoch [   10/   20] | d_loss: 0.3137 | g_loss: 3.5216\n",
      "Epoch [   10/   20] | d_loss: 0.4183 | g_loss: 2.3349\n",
      "Epoch [   10/   20] | d_loss: 0.7402 | g_loss: 2.0998\n",
      "Epoch [   10/   20] | d_loss: 0.6121 | g_loss: 2.5240\n",
      "Epoch [   10/   20] | d_loss: 0.3709 | g_loss: 3.2731\n",
      "Epoch [   10/   20] | d_loss: 0.4914 | g_loss: 1.3365\n",
      "Epoch [   10/   20] | d_loss: 0.4738 | g_loss: 3.7252\n",
      "Epoch [   10/   20] | d_loss: 0.7012 | g_loss: 1.3158\n",
      "Epoch [   10/   20] | d_loss: 0.6266 | g_loss: 3.7386\n",
      "Epoch [   10/   20] | d_loss: 0.1352 | g_loss: 3.8657\n",
      "Epoch [   10/   20] | d_loss: 0.5034 | g_loss: 5.2459\n",
      "Epoch [   10/   20] | d_loss: 0.8595 | g_loss: 4.3247\n",
      "Epoch [   10/   20] | d_loss: 0.1435 | g_loss: 1.1959\n",
      "Epoch [   10/   20] | d_loss: 0.2661 | g_loss: 3.3550\n",
      "Epoch [   10/   20] | d_loss: 0.9170 | g_loss: 1.0548\n",
      "Epoch [   10/   20] | d_loss: 0.1519 | g_loss: 2.2491\n",
      "Epoch [   10/   20] | d_loss: 0.5610 | g_loss: 1.6882\n",
      "Epoch [   10/   20] | d_loss: 0.9825 | g_loss: 1.0291\n",
      "Epoch [   10/   20] | d_loss: 0.0625 | g_loss: 2.3928\n",
      "Epoch [   10/   20] | d_loss: 0.8607 | g_loss: 1.7539\n",
      "Epoch [   10/   20] | d_loss: 0.5999 | g_loss: 2.3644\n",
      "Epoch [   10/   20] | d_loss: 0.8017 | g_loss: 1.9331\n",
      "Epoch [   10/   20] | d_loss: 0.6029 | g_loss: 3.2028\n",
      "Epoch [   10/   20] | d_loss: 0.1475 | g_loss: 5.5820\n",
      "Epoch [   10/   20] | d_loss: 0.0956 | g_loss: 4.2175\n",
      "Epoch [   10/   20] | d_loss: 0.3380 | g_loss: 1.9780\n",
      "Epoch [   10/   20] | d_loss: 0.2930 | g_loss: 4.5077\n",
      "Epoch [   10/   20] | d_loss: 0.4721 | g_loss: 3.0668\n",
      "Epoch [   10/   20] | d_loss: 0.4594 | g_loss: 3.9695\n",
      "Epoch [   10/   20] | d_loss: 0.5305 | g_loss: 1.8789\n",
      "Epoch [   10/   20] | d_loss: 0.3347 | g_loss: 2.3678\n",
      "Epoch [   10/   20] | d_loss: 0.5848 | g_loss: 3.0735\n",
      "Epoch [   10/   20] | d_loss: 0.8902 | g_loss: 1.8473\n",
      "Epoch [   10/   20] | d_loss: 0.3070 | g_loss: 4.1379\n",
      "Epoch [   10/   20] | d_loss: 0.2871 | g_loss: 3.6296\n",
      "Epoch [   10/   20] | d_loss: 0.3352 | g_loss: 3.2685\n",
      "Epoch [   10/   20] | d_loss: 0.4027 | g_loss: 3.9810\n",
      "Epoch [   10/   20] | d_loss: 0.6623 | g_loss: 1.9900\n",
      "Epoch [   10/   20] | d_loss: 0.6495 | g_loss: 3.3655\n",
      "Epoch [   10/   20] | d_loss: 0.8398 | g_loss: 3.1625\n",
      "Epoch [   10/   20] | d_loss: 0.9270 | g_loss: 3.5731\n",
      "Epoch [   10/   20] | d_loss: 0.4768 | g_loss: 2.6430\n",
      "Epoch [   10/   20] | d_loss: 0.3238 | g_loss: 3.6027\n",
      "Epoch [   10/   20] | d_loss: 0.3508 | g_loss: 3.9942\n",
      "Epoch [   10/   20] | d_loss: 0.2112 | g_loss: 4.0213\n",
      "Epoch [   10/   20] | d_loss: 0.5417 | g_loss: 3.0512\n",
      "Epoch [   11/   20] | d_loss: 0.3153 | g_loss: 2.3115\n",
      "Epoch [   11/   20] | d_loss: 0.5800 | g_loss: 3.7031\n",
      "Epoch [   11/   20] | d_loss: 0.7695 | g_loss: 2.9084\n",
      "Epoch [   11/   20] | d_loss: 0.0762 | g_loss: 2.7842\n",
      "Epoch [   11/   20] | d_loss: 0.2852 | g_loss: 3.7965\n",
      "Epoch [   11/   20] | d_loss: 1.2887 | g_loss: 2.7026\n",
      "Epoch [   11/   20] | d_loss: 0.7655 | g_loss: 3.6318\n",
      "Epoch [   11/   20] | d_loss: 0.6272 | g_loss: 1.2852\n",
      "Epoch [   11/   20] | d_loss: 0.7867 | g_loss: 3.7062\n",
      "Epoch [   11/   20] | d_loss: 0.1791 | g_loss: 3.8359\n",
      "Epoch [   11/   20] | d_loss: 0.6489 | g_loss: 1.7466\n",
      "Epoch [   11/   20] | d_loss: 0.4874 | g_loss: 2.5264\n",
      "Epoch [   11/   20] | d_loss: 0.4515 | g_loss: 3.2891\n",
      "Epoch [   11/   20] | d_loss: 0.4281 | g_loss: 1.0876\n",
      "Epoch [   11/   20] | d_loss: 0.8062 | g_loss: 0.9027\n",
      "Epoch [   11/   20] | d_loss: 0.6745 | g_loss: 3.7615\n",
      "Epoch [   11/   20] | d_loss: 0.2895 | g_loss: 2.5439\n",
      "Epoch [   11/   20] | d_loss: 0.3638 | g_loss: 3.6383\n",
      "Epoch [   11/   20] | d_loss: 1.0469 | g_loss: 4.1634\n",
      "Epoch [   11/   20] | d_loss: 0.3772 | g_loss: 3.9481\n",
      "Epoch [   11/   20] | d_loss: 0.9312 | g_loss: 3.8308\n",
      "Epoch [   11/   20] | d_loss: 0.1227 | g_loss: 2.3303\n",
      "Epoch [   11/   20] | d_loss: 0.3707 | g_loss: 2.9253\n",
      "Epoch [   11/   20] | d_loss: 0.4595 | g_loss: 3.2367\n",
      "Epoch [   11/   20] | d_loss: 1.7802 | g_loss: 0.9127\n",
      "Epoch [   11/   20] | d_loss: 0.4939 | g_loss: 3.2480\n",
      "Epoch [   11/   20] | d_loss: 0.1975 | g_loss: 2.2539\n",
      "Epoch [   11/   20] | d_loss: 0.5538 | g_loss: 3.5935\n",
      "Epoch [   11/   20] | d_loss: 0.6027 | g_loss: 3.3723\n",
      "Epoch [   11/   20] | d_loss: 0.1351 | g_loss: 3.9319\n",
      "Epoch [   11/   20] | d_loss: 0.1891 | g_loss: 3.5949\n",
      "Epoch [   11/   20] | d_loss: 0.6295 | g_loss: 1.2802\n",
      "Epoch [   11/   20] | d_loss: 0.3794 | g_loss: 2.3438\n",
      "Epoch [   11/   20] | d_loss: 0.5698 | g_loss: 2.3605\n",
      "Epoch [   11/   20] | d_loss: 0.5287 | g_loss: 4.1172\n",
      "Epoch [   11/   20] | d_loss: 0.5836 | g_loss: 3.9989\n",
      "Epoch [   11/   20] | d_loss: 0.2349 | g_loss: 3.9708\n",
      "Epoch [   11/   20] | d_loss: 0.3461 | g_loss: 2.2572\n",
      "Epoch [   11/   20] | d_loss: 0.2359 | g_loss: 2.3674\n",
      "Epoch [   11/   20] | d_loss: 0.1657 | g_loss: 4.3939\n",
      "Epoch [   11/   20] | d_loss: 0.6489 | g_loss: 2.6723\n",
      "Epoch [   11/   20] | d_loss: 0.1779 | g_loss: 3.3646\n",
      "Epoch [   11/   20] | d_loss: 0.1604 | g_loss: 3.5704\n",
      "Epoch [   11/   20] | d_loss: 0.2050 | g_loss: 4.7949\n",
      "Epoch [   11/   20] | d_loss: 0.3026 | g_loss: 2.9916\n",
      "Epoch [   11/   20] | d_loss: 0.4337 | g_loss: 3.6583\n",
      "Epoch [   11/   20] | d_loss: 0.4451 | g_loss: 1.8397\n",
      "Epoch [   11/   20] | d_loss: 0.6063 | g_loss: 3.9543\n",
      "Epoch [   11/   20] | d_loss: 0.3706 | g_loss: 1.5397\n",
      "Epoch [   11/   20] | d_loss: 1.0293 | g_loss: 2.9182\n",
      "Epoch [   11/   20] | d_loss: 0.3875 | g_loss: 2.7729\n",
      "Epoch [   11/   20] | d_loss: 0.1661 | g_loss: 3.9448\n",
      "Epoch [   11/   20] | d_loss: 0.2066 | g_loss: 4.5437\n",
      "Epoch [   11/   20] | d_loss: 0.4978 | g_loss: 3.3058\n",
      "Epoch [   11/   20] | d_loss: 0.4081 | g_loss: 2.7256\n",
      "Epoch [   11/   20] | d_loss: 0.1873 | g_loss: 1.2049\n",
      "Epoch [   11/   20] | d_loss: 0.4616 | g_loss: 1.6971\n",
      "Epoch [   11/   20] | d_loss: 0.3844 | g_loss: 3.3442\n",
      "Epoch [   11/   20] | d_loss: 0.1091 | g_loss: 4.1900\n",
      "Epoch [   11/   20] | d_loss: 0.2574 | g_loss: 2.2545\n",
      "Epoch [   11/   20] | d_loss: 0.3428 | g_loss: 2.5696\n",
      "Epoch [   11/   20] | d_loss: 0.1172 | g_loss: 5.0693\n",
      "Epoch [   11/   20] | d_loss: 0.5084 | g_loss: 3.1348\n",
      "Epoch [   11/   20] | d_loss: 0.5586 | g_loss: 3.6089\n",
      "Epoch [   11/   20] | d_loss: 0.1913 | g_loss: 2.4122\n",
      "Epoch [   11/   20] | d_loss: 0.1390 | g_loss: 3.2718\n",
      "Epoch [   11/   20] | d_loss: 0.4271 | g_loss: 3.3143\n",
      "Epoch [   11/   20] | d_loss: 0.4699 | g_loss: 2.0227\n",
      "Epoch [   11/   20] | d_loss: 0.5676 | g_loss: 3.0327\n",
      "Epoch [   11/   20] | d_loss: 0.4062 | g_loss: 2.8329\n",
      "Epoch [   11/   20] | d_loss: 0.7452 | g_loss: 3.2940\n",
      "Epoch [   11/   20] | d_loss: 1.2199 | g_loss: 2.7371\n",
      "Epoch [   11/   20] | d_loss: 0.4143 | g_loss: 3.3867\n",
      "Epoch [   11/   20] | d_loss: 0.1565 | g_loss: 4.3635\n",
      "Epoch [   11/   20] | d_loss: 0.2244 | g_loss: 2.4860\n",
      "Epoch [   11/   20] | d_loss: 0.7439 | g_loss: 3.5411\n",
      "Epoch [   11/   20] | d_loss: 0.2351 | g_loss: 3.3182\n",
      "Epoch [   11/   20] | d_loss: 0.2056 | g_loss: 1.2038\n",
      "Epoch [   11/   20] | d_loss: 0.3300 | g_loss: 2.2831\n",
      "Epoch [   11/   20] | d_loss: 0.3365 | g_loss: 2.1964\n",
      "Epoch [   11/   20] | d_loss: 0.2229 | g_loss: 3.5356\n",
      "Epoch [   11/   20] | d_loss: 0.1165 | g_loss: 3.6402\n",
      "Epoch [   11/   20] | d_loss: 0.0473 | g_loss: 3.3084\n",
      "Epoch [   11/   20] | d_loss: 0.1457 | g_loss: 2.9663\n",
      "Epoch [   11/   20] | d_loss: 0.6206 | g_loss: 3.2748\n",
      "Epoch [   11/   20] | d_loss: 0.9568 | g_loss: 1.4789\n",
      "Epoch [   11/   20] | d_loss: 0.1580 | g_loss: 2.2262\n",
      "Epoch [   11/   20] | d_loss: 0.6203 | g_loss: 2.3206\n",
      "Epoch [   11/   20] | d_loss: 0.3738 | g_loss: 2.1209\n",
      "Epoch [   11/   20] | d_loss: 0.2467 | g_loss: 3.2925\n",
      "Epoch [   11/   20] | d_loss: 0.5210 | g_loss: 3.4371\n",
      "Epoch [   11/   20] | d_loss: 0.4998 | g_loss: 2.2802\n",
      "Epoch [   11/   20] | d_loss: 0.4138 | g_loss: 4.3857\n",
      "Epoch [   11/   20] | d_loss: 0.3244 | g_loss: 1.7970\n",
      "Epoch [   11/   20] | d_loss: 0.7516 | g_loss: 3.1398\n",
      "Epoch [   11/   20] | d_loss: 0.8620 | g_loss: 4.5165\n",
      "Epoch [   11/   20] | d_loss: 0.0306 | g_loss: 2.6306\n",
      "Epoch [   11/   20] | d_loss: 1.0069 | g_loss: 4.0098\n",
      "Epoch [   11/   20] | d_loss: 0.4638 | g_loss: 3.7118\n",
      "Epoch [   11/   20] | d_loss: 0.2974 | g_loss: 2.9442\n",
      "Epoch [   11/   20] | d_loss: 0.8559 | g_loss: 1.3161\n",
      "Epoch [   11/   20] | d_loss: 0.3565 | g_loss: 1.9601\n",
      "Epoch [   11/   20] | d_loss: 0.3646 | g_loss: 1.8994\n",
      "Epoch [   11/   20] | d_loss: 0.7843 | g_loss: 2.1393\n",
      "Epoch [   11/   20] | d_loss: 0.5118 | g_loss: 1.7898\n",
      "Epoch [   11/   20] | d_loss: 0.5723 | g_loss: 2.0700\n",
      "Epoch [   11/   20] | d_loss: 0.6800 | g_loss: 0.9144\n",
      "Epoch [   11/   20] | d_loss: 0.3329 | g_loss: 2.1820\n",
      "Epoch [   11/   20] | d_loss: 0.6217 | g_loss: 3.5265\n",
      "Epoch [   11/   20] | d_loss: 0.3778 | g_loss: 2.4641\n",
      "Epoch [   11/   20] | d_loss: 0.4227 | g_loss: 4.2394\n",
      "Epoch [   11/   20] | d_loss: 0.7578 | g_loss: 2.7547\n",
      "Epoch [   11/   20] | d_loss: 0.3874 | g_loss: 2.0849\n",
      "Epoch [   12/   20] | d_loss: 0.7396 | g_loss: 2.8692\n",
      "Epoch [   12/   20] | d_loss: 0.8614 | g_loss: 1.9709\n",
      "Epoch [   12/   20] | d_loss: 0.6882 | g_loss: 2.7717\n",
      "Epoch [   12/   20] | d_loss: 0.3047 | g_loss: 3.1186\n",
      "Epoch [   12/   20] | d_loss: 0.4071 | g_loss: 3.1273\n",
      "Epoch [   12/   20] | d_loss: 0.6581 | g_loss: 3.5696\n",
      "Epoch [   12/   20] | d_loss: 0.6012 | g_loss: 2.6619\n",
      "Epoch [   12/   20] | d_loss: 0.7398 | g_loss: 1.8005\n",
      "Epoch [   12/   20] | d_loss: 0.7969 | g_loss: 3.3431\n",
      "Epoch [   12/   20] | d_loss: 1.6768 | g_loss: 0.8867\n",
      "Epoch [   12/   20] | d_loss: 0.1775 | g_loss: 3.1070\n",
      "Epoch [   12/   20] | d_loss: 0.1249 | g_loss: 4.2389\n",
      "Epoch [   12/   20] | d_loss: 1.0637 | g_loss: 1.6490\n",
      "Epoch [   12/   20] | d_loss: 0.4183 | g_loss: 2.9249\n",
      "Epoch [   12/   20] | d_loss: 0.2265 | g_loss: 4.0155\n",
      "Epoch [   12/   20] | d_loss: 0.8616 | g_loss: 2.3050\n",
      "Epoch [   12/   20] | d_loss: 0.7683 | g_loss: 3.2775\n",
      "Epoch [   12/   20] | d_loss: 0.3521 | g_loss: 2.1881\n",
      "Epoch [   12/   20] | d_loss: 0.8015 | g_loss: 2.2392\n",
      "Epoch [   12/   20] | d_loss: 0.8713 | g_loss: 3.0151\n",
      "Epoch [   12/   20] | d_loss: 0.4848 | g_loss: 1.8936\n",
      "Epoch [   12/   20] | d_loss: 0.3754 | g_loss: 4.1255\n",
      "Epoch [   12/   20] | d_loss: 0.1196 | g_loss: 4.2008\n",
      "Epoch [   12/   20] | d_loss: 0.4971 | g_loss: 2.7441\n",
      "Epoch [   12/   20] | d_loss: 0.5551 | g_loss: 2.1634\n",
      "Epoch [   12/   20] | d_loss: 0.3505 | g_loss: 2.7322\n",
      "Epoch [   12/   20] | d_loss: 1.2382 | g_loss: 6.6103\n",
      "Epoch [   12/   20] | d_loss: 0.4900 | g_loss: 1.8277\n",
      "Epoch [   12/   20] | d_loss: 0.5229 | g_loss: 2.7245\n",
      "Epoch [   12/   20] | d_loss: 0.6709 | g_loss: 4.4918\n",
      "Epoch [   12/   20] | d_loss: 0.1626 | g_loss: 3.7895\n",
      "Epoch [   12/   20] | d_loss: 0.1741 | g_loss: 1.9912\n",
      "Epoch [   12/   20] | d_loss: 0.5096 | g_loss: 5.1358\n",
      "Epoch [   12/   20] | d_loss: 0.3974 | g_loss: 2.0267\n",
      "Epoch [   12/   20] | d_loss: 0.3005 | g_loss: 2.6188\n",
      "Epoch [   12/   20] | d_loss: 0.2907 | g_loss: 2.8378\n",
      "Epoch [   12/   20] | d_loss: 0.3084 | g_loss: 3.0339\n",
      "Epoch [   12/   20] | d_loss: 0.1664 | g_loss: 3.3163\n",
      "Epoch [   12/   20] | d_loss: 0.8491 | g_loss: 2.8036\n",
      "Epoch [   12/   20] | d_loss: 0.4657 | g_loss: 4.1586\n",
      "Epoch [   12/   20] | d_loss: 0.3453 | g_loss: 4.5190\n",
      "Epoch [   12/   20] | d_loss: 0.0962 | g_loss: 5.8406\n",
      "Epoch [   12/   20] | d_loss: 0.2597 | g_loss: 2.8551\n",
      "Epoch [   12/   20] | d_loss: 0.4205 | g_loss: 3.5901\n",
      "Epoch [   12/   20] | d_loss: 0.5627 | g_loss: 4.7427\n",
      "Epoch [   12/   20] | d_loss: 0.4897 | g_loss: 2.7678\n",
      "Epoch [   12/   20] | d_loss: 0.2699 | g_loss: 3.3727\n",
      "Epoch [   12/   20] | d_loss: 0.1964 | g_loss: 2.7240\n",
      "Epoch [   12/   20] | d_loss: 0.5218 | g_loss: 4.5912\n",
      "Epoch [   12/   20] | d_loss: 0.2862 | g_loss: 2.0451\n",
      "Epoch [   12/   20] | d_loss: 0.4139 | g_loss: 3.2969\n",
      "Epoch [   12/   20] | d_loss: 0.8467 | g_loss: 3.0371\n",
      "Epoch [   12/   20] | d_loss: 0.1546 | g_loss: 4.7226\n",
      "Epoch [   12/   20] | d_loss: 0.2782 | g_loss: 4.8648\n",
      "Epoch [   12/   20] | d_loss: 0.1060 | g_loss: 1.9697\n",
      "Epoch [   12/   20] | d_loss: 0.7065 | g_loss: 2.7042\n",
      "Epoch [   12/   20] | d_loss: 0.9230 | g_loss: 1.9808\n",
      "Epoch [   12/   20] | d_loss: 0.3758 | g_loss: 1.2901\n",
      "Epoch [   12/   20] | d_loss: 0.3774 | g_loss: 2.9806\n",
      "Epoch [   12/   20] | d_loss: 0.6252 | g_loss: 4.7772\n",
      "Epoch [   12/   20] | d_loss: 0.3842 | g_loss: 3.6032\n",
      "Epoch [   12/   20] | d_loss: 0.4079 | g_loss: 2.4064\n",
      "Epoch [   12/   20] | d_loss: 0.9378 | g_loss: 2.8368\n",
      "Epoch [   12/   20] | d_loss: 1.1905 | g_loss: 4.7838\n",
      "Epoch [   12/   20] | d_loss: 0.2133 | g_loss: 2.5750\n",
      "Epoch [   12/   20] | d_loss: 0.1368 | g_loss: 2.4747\n",
      "Epoch [   12/   20] | d_loss: 0.5431 | g_loss: 1.3299\n",
      "Epoch [   12/   20] | d_loss: 0.6797 | g_loss: 2.6117\n",
      "Epoch [   12/   20] | d_loss: 0.7566 | g_loss: 0.8906\n",
      "Epoch [   12/   20] | d_loss: 0.2944 | g_loss: 3.3997\n",
      "Epoch [   12/   20] | d_loss: 0.4228 | g_loss: 3.0504\n",
      "Epoch [   12/   20] | d_loss: 0.3499 | g_loss: 3.6148\n",
      "Epoch [   12/   20] | d_loss: 0.1581 | g_loss: 2.1792\n",
      "Epoch [   12/   20] | d_loss: 0.3234 | g_loss: 3.3272\n",
      "Epoch [   12/   20] | d_loss: 0.7096 | g_loss: 2.5433\n",
      "Epoch [   12/   20] | d_loss: 0.2964 | g_loss: 2.9313\n",
      "Epoch [   12/   20] | d_loss: 0.1709 | g_loss: 2.2696\n",
      "Epoch [   12/   20] | d_loss: 0.1235 | g_loss: 2.3503\n",
      "Epoch [   12/   20] | d_loss: 0.9741 | g_loss: 3.5871\n",
      "Epoch [   12/   20] | d_loss: 0.2953 | g_loss: 2.4933\n",
      "Epoch [   12/   20] | d_loss: 0.2107 | g_loss: 2.7633\n",
      "Epoch [   12/   20] | d_loss: 0.8288 | g_loss: 3.8709\n",
      "Epoch [   12/   20] | d_loss: 0.2296 | g_loss: 2.3948\n",
      "Epoch [   12/   20] | d_loss: 0.3670 | g_loss: 2.0199\n",
      "Epoch [   12/   20] | d_loss: 0.5310 | g_loss: 4.5991\n",
      "Epoch [   12/   20] | d_loss: 0.3293 | g_loss: 4.4375\n",
      "Epoch [   12/   20] | d_loss: 0.3013 | g_loss: 2.5588\n",
      "Epoch [   12/   20] | d_loss: 0.3053 | g_loss: 3.0622\n",
      "Epoch [   12/   20] | d_loss: 0.5303 | g_loss: 1.6012\n",
      "Epoch [   12/   20] | d_loss: 0.9532 | g_loss: 3.5220\n",
      "Epoch [   12/   20] | d_loss: 0.9052 | g_loss: 3.1232\n",
      "Epoch [   12/   20] | d_loss: 0.2409 | g_loss: 1.8787\n",
      "Epoch [   12/   20] | d_loss: 0.2885 | g_loss: 3.7020\n",
      "Epoch [   12/   20] | d_loss: 0.4328 | g_loss: 3.6578\n",
      "Epoch [   12/   20] | d_loss: 1.0763 | g_loss: 1.8873\n",
      "Epoch [   12/   20] | d_loss: 0.2268 | g_loss: 4.2207\n",
      "Epoch [   12/   20] | d_loss: 0.4679 | g_loss: 2.6046\n",
      "Epoch [   12/   20] | d_loss: 0.6222 | g_loss: 3.2229\n",
      "Epoch [   12/   20] | d_loss: 0.3901 | g_loss: 3.3383\n",
      "Epoch [   12/   20] | d_loss: 0.1884 | g_loss: 2.8848\n",
      "Epoch [   12/   20] | d_loss: 0.3895 | g_loss: 3.7045\n",
      "Epoch [   12/   20] | d_loss: 0.1858 | g_loss: 3.2100\n",
      "Epoch [   12/   20] | d_loss: 0.5335 | g_loss: 2.6395\n",
      "Epoch [   12/   20] | d_loss: 0.1499 | g_loss: 2.7355\n",
      "Epoch [   12/   20] | d_loss: 0.5415 | g_loss: 2.8653\n",
      "Epoch [   12/   20] | d_loss: 0.4525 | g_loss: 4.3314\n",
      "Epoch [   12/   20] | d_loss: 0.2784 | g_loss: 2.4652\n",
      "Epoch [   12/   20] | d_loss: 0.3211 | g_loss: 5.2845\n",
      "Epoch [   12/   20] | d_loss: 0.2520 | g_loss: 4.1569\n",
      "Epoch [   12/   20] | d_loss: 0.6582 | g_loss: 2.8731\n",
      "Epoch [   12/   20] | d_loss: 0.1961 | g_loss: 4.9684\n",
      "Epoch [   12/   20] | d_loss: 0.2523 | g_loss: 3.2521\n",
      "Epoch [   12/   20] | d_loss: 0.6949 | g_loss: 0.9887\n",
      "Epoch [   13/   20] | d_loss: 0.3021 | g_loss: 3.0645\n",
      "Epoch [   13/   20] | d_loss: 0.6020 | g_loss: 3.9801\n",
      "Epoch [   13/   20] | d_loss: 0.4241 | g_loss: 2.6583\n",
      "Epoch [   13/   20] | d_loss: 0.2872 | g_loss: 3.0256\n",
      "Epoch [   13/   20] | d_loss: 0.1054 | g_loss: 3.5760\n",
      "Epoch [   13/   20] | d_loss: 0.8643 | g_loss: 1.0371\n",
      "Epoch [   13/   20] | d_loss: 0.6483 | g_loss: 2.9058\n",
      "Epoch [   13/   20] | d_loss: 0.1307 | g_loss: 2.5079\n",
      "Epoch [   13/   20] | d_loss: 0.4289 | g_loss: 2.9062\n",
      "Epoch [   13/   20] | d_loss: 0.9646 | g_loss: 3.4728\n",
      "Epoch [   13/   20] | d_loss: 0.2878 | g_loss: 3.4247\n",
      "Epoch [   13/   20] | d_loss: 0.1556 | g_loss: 0.9050\n",
      "Epoch [   13/   20] | d_loss: 0.4359 | g_loss: 3.7715\n",
      "Epoch [   13/   20] | d_loss: 0.2177 | g_loss: 2.8614\n",
      "Epoch [   13/   20] | d_loss: 0.7876 | g_loss: 3.2701\n",
      "Epoch [   13/   20] | d_loss: 1.0518 | g_loss: 4.4766\n",
      "Epoch [   13/   20] | d_loss: 0.0790 | g_loss: 4.2935\n",
      "Epoch [   13/   20] | d_loss: 0.1648 | g_loss: 3.7856\n",
      "Epoch [   13/   20] | d_loss: 0.5599 | g_loss: 3.7532\n",
      "Epoch [   13/   20] | d_loss: 0.2883 | g_loss: 4.0255\n",
      "Epoch [   13/   20] | d_loss: 0.2113 | g_loss: 3.8087\n",
      "Epoch [   13/   20] | d_loss: 0.2926 | g_loss: 2.1862\n",
      "Epoch [   13/   20] | d_loss: 0.2928 | g_loss: 3.7145\n",
      "Epoch [   13/   20] | d_loss: 0.1979 | g_loss: 3.1783\n",
      "Epoch [   13/   20] | d_loss: 0.3997 | g_loss: 3.3061\n",
      "Epoch [   13/   20] | d_loss: 0.4190 | g_loss: 4.9468\n",
      "Epoch [   13/   20] | d_loss: 0.7182 | g_loss: 2.3486\n",
      "Epoch [   13/   20] | d_loss: 0.4551 | g_loss: 2.4502\n",
      "Epoch [   13/   20] | d_loss: 0.3281 | g_loss: 2.7527\n",
      "Epoch [   13/   20] | d_loss: 0.3135 | g_loss: 4.2525\n",
      "Epoch [   13/   20] | d_loss: 0.1359 | g_loss: 3.4594\n",
      "Epoch [   13/   20] | d_loss: 0.2158 | g_loss: 3.9263\n",
      "Epoch [   13/   20] | d_loss: 0.3694 | g_loss: 4.0670\n",
      "Epoch [   13/   20] | d_loss: 0.5323 | g_loss: 2.9199\n",
      "Epoch [   13/   20] | d_loss: 0.4095 | g_loss: 4.3366\n",
      "Epoch [   13/   20] | d_loss: 0.5831 | g_loss: 2.8489\n",
      "Epoch [   13/   20] | d_loss: 0.2776 | g_loss: 2.6462\n",
      "Epoch [   13/   20] | d_loss: 0.4579 | g_loss: 2.9552\n",
      "Epoch [   13/   20] | d_loss: 0.5702 | g_loss: 2.5059\n",
      "Epoch [   13/   20] | d_loss: 0.2054 | g_loss: 2.3029\n",
      "Epoch [   13/   20] | d_loss: 0.3981 | g_loss: 4.4517\n",
      "Epoch [   13/   20] | d_loss: 0.2597 | g_loss: 5.2379\n",
      "Epoch [   13/   20] | d_loss: 0.0870 | g_loss: 4.8842\n",
      "Epoch [   13/   20] | d_loss: 0.2448 | g_loss: 3.8112\n",
      "Epoch [   13/   20] | d_loss: 0.5353 | g_loss: 3.1397\n",
      "Epoch [   13/   20] | d_loss: 0.5305 | g_loss: 3.5527\n",
      "Epoch [   13/   20] | d_loss: 0.2083 | g_loss: 4.3388\n",
      "Epoch [   13/   20] | d_loss: 0.1701 | g_loss: 3.2427\n",
      "Epoch [   13/   20] | d_loss: 0.2234 | g_loss: 3.2661\n",
      "Epoch [   13/   20] | d_loss: 0.2030 | g_loss: 2.2424\n",
      "Epoch [   13/   20] | d_loss: 0.4297 | g_loss: 3.8172\n",
      "Epoch [   13/   20] | d_loss: 0.6439 | g_loss: 5.4731\n",
      "Epoch [   13/   20] | d_loss: 0.2079 | g_loss: 2.6886\n",
      "Epoch [   13/   20] | d_loss: 0.3995 | g_loss: 3.7047\n",
      "Epoch [   13/   20] | d_loss: 0.5719 | g_loss: 2.3762\n",
      "Epoch [   13/   20] | d_loss: 0.4043 | g_loss: 2.3248\n",
      "Epoch [   13/   20] | d_loss: 0.6065 | g_loss: 1.8044\n",
      "Epoch [   13/   20] | d_loss: 0.4466 | g_loss: 2.3328\n",
      "Epoch [   13/   20] | d_loss: 0.4499 | g_loss: 3.2569\n",
      "Epoch [   13/   20] | d_loss: 1.2711 | g_loss: 3.4944\n",
      "Epoch [   13/   20] | d_loss: 0.4817 | g_loss: 0.9699\n",
      "Epoch [   13/   20] | d_loss: 1.0595 | g_loss: 2.2906\n",
      "Epoch [   13/   20] | d_loss: 0.6853 | g_loss: 2.8940\n",
      "Epoch [   13/   20] | d_loss: 0.2258 | g_loss: 4.1569\n",
      "Epoch [   13/   20] | d_loss: 0.4667 | g_loss: 1.5277\n",
      "Epoch [   13/   20] | d_loss: 0.6449 | g_loss: 1.6154\n",
      "Epoch [   13/   20] | d_loss: 0.2934 | g_loss: 3.9413\n",
      "Epoch [   13/   20] | d_loss: 0.4614 | g_loss: 4.0925\n",
      "Epoch [   13/   20] | d_loss: 0.3432 | g_loss: 6.2048\n",
      "Epoch [   13/   20] | d_loss: 0.2590 | g_loss: 2.1941\n",
      "Epoch [   13/   20] | d_loss: 0.2770 | g_loss: 3.5425\n",
      "Epoch [   13/   20] | d_loss: 0.4140 | g_loss: 3.7640\n",
      "Epoch [   13/   20] | d_loss: 0.0854 | g_loss: 2.8914\n",
      "Epoch [   13/   20] | d_loss: 0.1681 | g_loss: 3.2555\n",
      "Epoch [   13/   20] | d_loss: 0.1343 | g_loss: 4.5682\n",
      "Epoch [   13/   20] | d_loss: 0.3341 | g_loss: 4.7774\n",
      "Epoch [   13/   20] | d_loss: 0.2841 | g_loss: 3.2584\n",
      "Epoch [   13/   20] | d_loss: 0.1224 | g_loss: 3.2856\n",
      "Epoch [   13/   20] | d_loss: 0.2031 | g_loss: 0.7938\n",
      "Epoch [   13/   20] | d_loss: 0.0829 | g_loss: 4.2620\n",
      "Epoch [   13/   20] | d_loss: 0.3417 | g_loss: 4.3164\n",
      "Epoch [   13/   20] | d_loss: 0.4258 | g_loss: 2.8480\n",
      "Epoch [   13/   20] | d_loss: 0.7984 | g_loss: 5.1558\n",
      "Epoch [   13/   20] | d_loss: 0.3331 | g_loss: 1.8913\n",
      "Epoch [   13/   20] | d_loss: 0.4198 | g_loss: 3.3054\n",
      "Epoch [   13/   20] | d_loss: 0.9941 | g_loss: 2.3049\n",
      "Epoch [   13/   20] | d_loss: 0.3764 | g_loss: 4.0253\n",
      "Epoch [   13/   20] | d_loss: 0.2888 | g_loss: 1.2786\n",
      "Epoch [   13/   20] | d_loss: 0.6256 | g_loss: 4.3941\n",
      "Epoch [   13/   20] | d_loss: 0.8636 | g_loss: 2.8919\n",
      "Epoch [   13/   20] | d_loss: 0.5719 | g_loss: 1.0033\n",
      "Epoch [   13/   20] | d_loss: 0.2415 | g_loss: 2.4017\n",
      "Epoch [   13/   20] | d_loss: 0.2750 | g_loss: 3.9170\n",
      "Epoch [   13/   20] | d_loss: 0.1773 | g_loss: 3.5672\n",
      "Epoch [   13/   20] | d_loss: 1.1530 | g_loss: 0.8377\n",
      "Epoch [   13/   20] | d_loss: 0.1986 | g_loss: 3.0791\n",
      "Epoch [   13/   20] | d_loss: 0.6788 | g_loss: 4.9153\n",
      "Epoch [   13/   20] | d_loss: 0.1733 | g_loss: 4.0892\n",
      "Epoch [   13/   20] | d_loss: 0.4854 | g_loss: 2.4885\n",
      "Epoch [   13/   20] | d_loss: 0.3685 | g_loss: 2.5486\n",
      "Epoch [   13/   20] | d_loss: 0.3525 | g_loss: 3.6848\n",
      "Epoch [   13/   20] | d_loss: 0.2129 | g_loss: 2.5652\n",
      "Epoch [   13/   20] | d_loss: 0.3517 | g_loss: 3.4877\n",
      "Epoch [   13/   20] | d_loss: 0.4645 | g_loss: 3.1119\n",
      "Epoch [   13/   20] | d_loss: 0.4091 | g_loss: 1.5787\n",
      "Epoch [   13/   20] | d_loss: 0.1880 | g_loss: 2.5508\n",
      "Epoch [   13/   20] | d_loss: 0.1917 | g_loss: 3.0530\n",
      "Epoch [   13/   20] | d_loss: 0.4395 | g_loss: 4.1296\n",
      "Epoch [   13/   20] | d_loss: 0.4851 | g_loss: 4.2408\n",
      "Epoch [   13/   20] | d_loss: 0.2641 | g_loss: 4.5335\n",
      "Epoch [   13/   20] | d_loss: 0.4641 | g_loss: 1.7874\n",
      "Epoch [   13/   20] | d_loss: 0.1087 | g_loss: 4.0839\n",
      "Epoch [   13/   20] | d_loss: 0.3106 | g_loss: 3.9323\n",
      "Epoch [   14/   20] | d_loss: 0.2495 | g_loss: 3.4102\n",
      "Epoch [   14/   20] | d_loss: 0.0888 | g_loss: 4.4380\n",
      "Epoch [   14/   20] | d_loss: 0.2198 | g_loss: 4.0994\n",
      "Epoch [   14/   20] | d_loss: 0.3575 | g_loss: 3.3111\n",
      "Epoch [   14/   20] | d_loss: 0.8411 | g_loss: 5.3756\n",
      "Epoch [   14/   20] | d_loss: 0.5977 | g_loss: 1.2186\n",
      "Epoch [   14/   20] | d_loss: 0.5055 | g_loss: 5.8209\n",
      "Epoch [   14/   20] | d_loss: 0.2812 | g_loss: 4.1859\n",
      "Epoch [   14/   20] | d_loss: 0.2804 | g_loss: 2.1543\n",
      "Epoch [   14/   20] | d_loss: 1.0721 | g_loss: 2.9410\n",
      "Epoch [   14/   20] | d_loss: 0.4477 | g_loss: 2.4622\n",
      "Epoch [   14/   20] | d_loss: 0.0951 | g_loss: 3.8619\n",
      "Epoch [   14/   20] | d_loss: 0.4616 | g_loss: 4.3924\n",
      "Epoch [   14/   20] | d_loss: 0.6721 | g_loss: 4.7133\n",
      "Epoch [   14/   20] | d_loss: 0.4237 | g_loss: 2.4625\n",
      "Epoch [   14/   20] | d_loss: 0.1917 | g_loss: 1.8981\n",
      "Epoch [   14/   20] | d_loss: 0.1495 | g_loss: 3.3174\n",
      "Epoch [   14/   20] | d_loss: 0.1038 | g_loss: 3.6160\n",
      "Epoch [   14/   20] | d_loss: 0.3643 | g_loss: 6.0291\n",
      "Epoch [   14/   20] | d_loss: 0.4983 | g_loss: 3.4283\n",
      "Epoch [   14/   20] | d_loss: 0.2816 | g_loss: 4.2483\n",
      "Epoch [   14/   20] | d_loss: 0.2938 | g_loss: 2.6261\n",
      "Epoch [   14/   20] | d_loss: 0.2675 | g_loss: 2.8417\n",
      "Epoch [   14/   20] | d_loss: 0.2922 | g_loss: 3.4421\n",
      "Epoch [   14/   20] | d_loss: 0.3936 | g_loss: 3.7185\n",
      "Epoch [   14/   20] | d_loss: 0.7147 | g_loss: 4.1232\n",
      "Epoch [   14/   20] | d_loss: 0.5519 | g_loss: 2.2910\n",
      "Epoch [   14/   20] | d_loss: 0.1508 | g_loss: 3.1038\n",
      "Epoch [   14/   20] | d_loss: 0.4109 | g_loss: 3.9323\n",
      "Epoch [   14/   20] | d_loss: 0.7917 | g_loss: 3.0007\n",
      "Epoch [   14/   20] | d_loss: 0.4837 | g_loss: 4.1160\n",
      "Epoch [   14/   20] | d_loss: 0.1633 | g_loss: 3.6866\n",
      "Epoch [   14/   20] | d_loss: 0.1899 | g_loss: 5.0415\n",
      "Epoch [   14/   20] | d_loss: 0.5187 | g_loss: 3.8335\n",
      "Epoch [   14/   20] | d_loss: 0.3687 | g_loss: 2.6631\n",
      "Epoch [   14/   20] | d_loss: 0.5183 | g_loss: 2.5362\n",
      "Epoch [   14/   20] | d_loss: 0.2034 | g_loss: 2.6242\n",
      "Epoch [   14/   20] | d_loss: 0.3600 | g_loss: 2.5474\n",
      "Epoch [   14/   20] | d_loss: 0.8451 | g_loss: 5.2086\n",
      "Epoch [   14/   20] | d_loss: 0.1507 | g_loss: 2.2004\n",
      "Epoch [   14/   20] | d_loss: 0.5700 | g_loss: 3.2387\n",
      "Epoch [   14/   20] | d_loss: 0.2764 | g_loss: 5.7349\n",
      "Epoch [   14/   20] | d_loss: 0.2221 | g_loss: 1.8330\n",
      "Epoch [   14/   20] | d_loss: 0.2332 | g_loss: 3.6575\n",
      "Epoch [   14/   20] | d_loss: 0.3316 | g_loss: 3.3526\n",
      "Epoch [   14/   20] | d_loss: 0.9263 | g_loss: 2.2045\n",
      "Epoch [   14/   20] | d_loss: 0.3153 | g_loss: 1.7247\n",
      "Epoch [   14/   20] | d_loss: 0.1815 | g_loss: 2.1528\n",
      "Epoch [   14/   20] | d_loss: 0.4161 | g_loss: 3.9651\n",
      "Epoch [   14/   20] | d_loss: 0.8161 | g_loss: 2.9525\n",
      "Epoch [   14/   20] | d_loss: 0.1798 | g_loss: 2.1600\n",
      "Epoch [   14/   20] | d_loss: 0.6051 | g_loss: 3.7086\n",
      "Epoch [   14/   20] | d_loss: 0.7633 | g_loss: 2.2723\n",
      "Epoch [   14/   20] | d_loss: 0.5609 | g_loss: 2.6659\n",
      "Epoch [   14/   20] | d_loss: 0.5078 | g_loss: 3.9080\n",
      "Epoch [   14/   20] | d_loss: 0.3924 | g_loss: 2.9669\n",
      "Epoch [   14/   20] | d_loss: 0.3489 | g_loss: 3.7267\n",
      "Epoch [   14/   20] | d_loss: 0.8619 | g_loss: 2.9292\n",
      "Epoch [   14/   20] | d_loss: 0.3885 | g_loss: 1.9100\n",
      "Epoch [   14/   20] | d_loss: 0.0341 | g_loss: 3.8814\n",
      "Epoch [   14/   20] | d_loss: 0.1003 | g_loss: 3.6114\n",
      "Epoch [   14/   20] | d_loss: 0.1299 | g_loss: 2.8845\n",
      "Epoch [   14/   20] | d_loss: 0.3838 | g_loss: 2.6151\n",
      "Epoch [   14/   20] | d_loss: 0.3418 | g_loss: 5.2837\n",
      "Epoch [   14/   20] | d_loss: 0.6716 | g_loss: 2.3845\n",
      "Epoch [   14/   20] | d_loss: 0.7599 | g_loss: 2.8300\n",
      "Epoch [   14/   20] | d_loss: 0.3381 | g_loss: 3.6679\n",
      "Epoch [   14/   20] | d_loss: 0.5377 | g_loss: 2.4719\n",
      "Epoch [   14/   20] | d_loss: 0.6239 | g_loss: 3.7932\n",
      "Epoch [   14/   20] | d_loss: 0.6503 | g_loss: 2.0127\n",
      "Epoch [   14/   20] | d_loss: 0.5043 | g_loss: 2.0637\n",
      "Epoch [   14/   20] | d_loss: 0.4636 | g_loss: 5.7998\n",
      "Epoch [   14/   20] | d_loss: 0.1605 | g_loss: 1.5072\n",
      "Epoch [   14/   20] | d_loss: 0.2701 | g_loss: 4.3636\n",
      "Epoch [   14/   20] | d_loss: 0.3728 | g_loss: 3.4395\n",
      "Epoch [   14/   20] | d_loss: 0.3227 | g_loss: 2.9351\n",
      "Epoch [   14/   20] | d_loss: 0.2957 | g_loss: 3.5391\n",
      "Epoch [   14/   20] | d_loss: 0.1709 | g_loss: 4.4713\n",
      "Epoch [   14/   20] | d_loss: 0.7945 | g_loss: 3.4544\n",
      "Epoch [   14/   20] | d_loss: 0.3282 | g_loss: 1.8838\n",
      "Epoch [   14/   20] | d_loss: 0.3287 | g_loss: 1.9155\n",
      "Epoch [   14/   20] | d_loss: 0.5282 | g_loss: 3.3469\n",
      "Epoch [   14/   20] | d_loss: 0.2911 | g_loss: 3.6122\n",
      "Epoch [   14/   20] | d_loss: 0.3074 | g_loss: 2.0614\n",
      "Epoch [   14/   20] | d_loss: 0.4400 | g_loss: 2.7386\n",
      "Epoch [   14/   20] | d_loss: 0.3861 | g_loss: 3.8211\n",
      "Epoch [   14/   20] | d_loss: 0.4775 | g_loss: 3.8484\n",
      "Epoch [   14/   20] | d_loss: 0.4826 | g_loss: 2.7772\n",
      "Epoch [   14/   20] | d_loss: 0.1128 | g_loss: 3.3882\n",
      "Epoch [   14/   20] | d_loss: 1.1562 | g_loss: 1.7471\n",
      "Epoch [   14/   20] | d_loss: 0.4275 | g_loss: 1.8203\n",
      "Epoch [   14/   20] | d_loss: 0.1203 | g_loss: 3.5274\n",
      "Epoch [   14/   20] | d_loss: 0.4382 | g_loss: 4.1847\n",
      "Epoch [   14/   20] | d_loss: 0.0661 | g_loss: 2.7534\n",
      "Epoch [   14/   20] | d_loss: 0.6633 | g_loss: 2.4073\n",
      "Epoch [   14/   20] | d_loss: 0.2010 | g_loss: 3.0830\n",
      "Epoch [   14/   20] | d_loss: 0.3236 | g_loss: 2.7942\n",
      "Epoch [   14/   20] | d_loss: 0.5433 | g_loss: 4.2811\n",
      "Epoch [   14/   20] | d_loss: 0.4048 | g_loss: 4.4221\n",
      "Epoch [   14/   20] | d_loss: 0.0182 | g_loss: 4.6246\n",
      "Epoch [   14/   20] | d_loss: 0.3341 | g_loss: 2.5313\n",
      "Epoch [   14/   20] | d_loss: 0.3324 | g_loss: 3.9113\n",
      "Epoch [   14/   20] | d_loss: 0.7908 | g_loss: 2.1098\n",
      "Epoch [   14/   20] | d_loss: 0.7007 | g_loss: 1.8359\n",
      "Epoch [   14/   20] | d_loss: 0.3556 | g_loss: 3.0637\n",
      "Epoch [   14/   20] | d_loss: 0.4812 | g_loss: 3.0206\n",
      "Epoch [   14/   20] | d_loss: 0.4576 | g_loss: 4.9966\n",
      "Epoch [   14/   20] | d_loss: 0.2629 | g_loss: 2.3465\n",
      "Epoch [   14/   20] | d_loss: 0.5067 | g_loss: 3.9864\n",
      "Epoch [   14/   20] | d_loss: 0.6026 | g_loss: 2.5666\n",
      "Epoch [   14/   20] | d_loss: 0.4268 | g_loss: 2.7615\n",
      "Epoch [   14/   20] | d_loss: 0.2684 | g_loss: 3.3561\n",
      "Epoch [   14/   20] | d_loss: 0.3361 | g_loss: 3.4486\n",
      "Epoch [   15/   20] | d_loss: 0.1372 | g_loss: 2.1789\n",
      "Epoch [   15/   20] | d_loss: 0.1212 | g_loss: 3.6173\n",
      "Epoch [   15/   20] | d_loss: 0.0741 | g_loss: 6.3333\n",
      "Epoch [   15/   20] | d_loss: 0.1077 | g_loss: 4.1774\n",
      "Epoch [   15/   20] | d_loss: 0.0950 | g_loss: 1.5996\n",
      "Epoch [   15/   20] | d_loss: 0.2614 | g_loss: 4.4424\n",
      "Epoch [   15/   20] | d_loss: 0.5814 | g_loss: 2.1862\n",
      "Epoch [   15/   20] | d_loss: 0.2728 | g_loss: 2.7554\n",
      "Epoch [   15/   20] | d_loss: 0.4587 | g_loss: 3.7126\n",
      "Epoch [   15/   20] | d_loss: 0.4015 | g_loss: 3.3817\n",
      "Epoch [   15/   20] | d_loss: 0.1352 | g_loss: 1.4054\n",
      "Epoch [   15/   20] | d_loss: 0.2834 | g_loss: 2.8542\n",
      "Epoch [   15/   20] | d_loss: 0.1870 | g_loss: 3.4082\n",
      "Epoch [   15/   20] | d_loss: 0.1708 | g_loss: 3.9387\n",
      "Epoch [   15/   20] | d_loss: 0.1372 | g_loss: 3.2530\n",
      "Epoch [   15/   20] | d_loss: 0.3450 | g_loss: 3.0568\n",
      "Epoch [   15/   20] | d_loss: 0.2587 | g_loss: 4.1032\n",
      "Epoch [   15/   20] | d_loss: 0.2418 | g_loss: 3.5011\n",
      "Epoch [   15/   20] | d_loss: 0.5026 | g_loss: 2.4490\n",
      "Epoch [   15/   20] | d_loss: 0.4588 | g_loss: 2.4791\n",
      "Epoch [   15/   20] | d_loss: 0.3364 | g_loss: 2.7080\n",
      "Epoch [   15/   20] | d_loss: 0.3590 | g_loss: 4.3300\n",
      "Epoch [   15/   20] | d_loss: 0.2725 | g_loss: 0.8924\n",
      "Epoch [   15/   20] | d_loss: 0.5650 | g_loss: 3.1276\n",
      "Epoch [   15/   20] | d_loss: 0.6709 | g_loss: 3.4227\n",
      "Epoch [   15/   20] | d_loss: 0.7253 | g_loss: 3.3660\n",
      "Epoch [   15/   20] | d_loss: 0.6472 | g_loss: 3.0265\n",
      "Epoch [   15/   20] | d_loss: 0.3684 | g_loss: 4.2027\n",
      "Epoch [   15/   20] | d_loss: 0.6449 | g_loss: 2.9631\n",
      "Epoch [   15/   20] | d_loss: 0.1234 | g_loss: 2.9507\n",
      "Epoch [   15/   20] | d_loss: 0.4124 | g_loss: 5.7789\n",
      "Epoch [   15/   20] | d_loss: 0.1595 | g_loss: 1.8839\n",
      "Epoch [   15/   20] | d_loss: 0.3175 | g_loss: 4.4592\n",
      "Epoch [   15/   20] | d_loss: 0.1302 | g_loss: 3.2962\n",
      "Epoch [   15/   20] | d_loss: 0.3745 | g_loss: 4.9706\n",
      "Epoch [   15/   20] | d_loss: 0.3670 | g_loss: 3.9526\n",
      "Epoch [   15/   20] | d_loss: 0.3146 | g_loss: 4.4180\n",
      "Epoch [   15/   20] | d_loss: 0.7137 | g_loss: 2.9320\n",
      "Epoch [   15/   20] | d_loss: 0.4818 | g_loss: 3.8108\n",
      "Epoch [   15/   20] | d_loss: 0.4420 | g_loss: 1.6361\n",
      "Epoch [   15/   20] | d_loss: 0.7371 | g_loss: 1.7533\n",
      "Epoch [   15/   20] | d_loss: 0.5103 | g_loss: 3.9984\n",
      "Epoch [   15/   20] | d_loss: 0.6132 | g_loss: 3.4319\n",
      "Epoch [   15/   20] | d_loss: 1.3401 | g_loss: 2.8343\n",
      "Epoch [   15/   20] | d_loss: 1.0663 | g_loss: 1.6245\n",
      "Epoch [   15/   20] | d_loss: 0.3685 | g_loss: 5.0126\n",
      "Epoch [   15/   20] | d_loss: 0.3181 | g_loss: 3.5063\n",
      "Epoch [   15/   20] | d_loss: 0.1516 | g_loss: 2.2166\n",
      "Epoch [   15/   20] | d_loss: 0.1413 | g_loss: 6.2728\n",
      "Epoch [   15/   20] | d_loss: 0.1274 | g_loss: 3.3616\n",
      "Epoch [   15/   20] | d_loss: 0.4230 | g_loss: 2.6157\n",
      "Epoch [   15/   20] | d_loss: 0.3729 | g_loss: 4.3563\n",
      "Epoch [   15/   20] | d_loss: 0.3153 | g_loss: 1.5798\n",
      "Epoch [   15/   20] | d_loss: 0.1909 | g_loss: 4.0139\n",
      "Epoch [   15/   20] | d_loss: 0.0830 | g_loss: 3.5584\n",
      "Epoch [   15/   20] | d_loss: 0.3657 | g_loss: 3.6839\n",
      "Epoch [   15/   20] | d_loss: 0.2368 | g_loss: 2.1761\n",
      "Epoch [   15/   20] | d_loss: 0.2007 | g_loss: 4.2163\n",
      "Epoch [   15/   20] | d_loss: 0.6011 | g_loss: 3.6773\n",
      "Epoch [   15/   20] | d_loss: 0.1523 | g_loss: 4.2817\n",
      "Epoch [   15/   20] | d_loss: 0.1584 | g_loss: 6.7074\n",
      "Epoch [   15/   20] | d_loss: 0.3513 | g_loss: 4.5597\n",
      "Epoch [   15/   20] | d_loss: 0.6473 | g_loss: 3.1780\n",
      "Epoch [   15/   20] | d_loss: 0.5488 | g_loss: 5.0328\n",
      "Epoch [   15/   20] | d_loss: 0.1278 | g_loss: 4.9817\n",
      "Epoch [   15/   20] | d_loss: 0.4174 | g_loss: 4.0056\n",
      "Epoch [   15/   20] | d_loss: 0.3591 | g_loss: 1.8634\n",
      "Epoch [   15/   20] | d_loss: 0.1574 | g_loss: 2.3647\n",
      "Epoch [   15/   20] | d_loss: 1.2613 | g_loss: 2.3370\n",
      "Epoch [   15/   20] | d_loss: 0.5989 | g_loss: 1.4446\n",
      "Epoch [   15/   20] | d_loss: 0.4621 | g_loss: 2.1895\n",
      "Epoch [   15/   20] | d_loss: 0.1958 | g_loss: 5.0312\n",
      "Epoch [   15/   20] | d_loss: 0.2178 | g_loss: 3.0131\n",
      "Epoch [   15/   20] | d_loss: 0.0620 | g_loss: 4.2108\n",
      "Epoch [   15/   20] | d_loss: 1.0677 | g_loss: 1.7500\n",
      "Epoch [   15/   20] | d_loss: 0.4680 | g_loss: 3.4305\n",
      "Epoch [   15/   20] | d_loss: 0.1379 | g_loss: 4.9455\n",
      "Epoch [   15/   20] | d_loss: 0.3152 | g_loss: 3.1794\n",
      "Epoch [   15/   20] | d_loss: 0.3441 | g_loss: 3.7339\n",
      "Epoch [   15/   20] | d_loss: 0.4156 | g_loss: 3.2509\n",
      "Epoch [   15/   20] | d_loss: 0.3107 | g_loss: 4.7336\n",
      "Epoch [   15/   20] | d_loss: 0.4573 | g_loss: 1.9790\n",
      "Epoch [   15/   20] | d_loss: 0.0951 | g_loss: 5.8502\n",
      "Epoch [   15/   20] | d_loss: 0.2736 | g_loss: 3.2926\n",
      "Epoch [   15/   20] | d_loss: 0.9210 | g_loss: 3.0105\n",
      "Epoch [   15/   20] | d_loss: 0.4116 | g_loss: 3.3807\n",
      "Epoch [   15/   20] | d_loss: 0.4565 | g_loss: 3.4322\n",
      "Epoch [   15/   20] | d_loss: 0.1723 | g_loss: 4.8051\n",
      "Epoch [   15/   20] | d_loss: 0.3202 | g_loss: 3.9188\n",
      "Epoch [   15/   20] | d_loss: 0.5696 | g_loss: 2.5113\n",
      "Epoch [   15/   20] | d_loss: 0.6428 | g_loss: 4.2494\n",
      "Epoch [   15/   20] | d_loss: 0.1839 | g_loss: 2.5527\n",
      "Epoch [   15/   20] | d_loss: 0.1665 | g_loss: 3.2268\n",
      "Epoch [   15/   20] | d_loss: 0.1810 | g_loss: 2.4965\n",
      "Epoch [   15/   20] | d_loss: 0.9211 | g_loss: 4.9909\n",
      "Epoch [   15/   20] | d_loss: 0.3544 | g_loss: 3.2853\n",
      "Epoch [   15/   20] | d_loss: 0.0526 | g_loss: 2.0993\n",
      "Epoch [   15/   20] | d_loss: 0.0758 | g_loss: 2.6698\n",
      "Epoch [   15/   20] | d_loss: 0.4004 | g_loss: 1.7276\n",
      "Epoch [   15/   20] | d_loss: 0.2887 | g_loss: 2.7432\n",
      "Epoch [   15/   20] | d_loss: 0.1675 | g_loss: 3.9879\n",
      "Epoch [   15/   20] | d_loss: 0.7553 | g_loss: 3.4397\n",
      "Epoch [   15/   20] | d_loss: 0.2716 | g_loss: 3.5347\n",
      "Epoch [   15/   20] | d_loss: 0.7650 | g_loss: 2.1323\n",
      "Epoch [   15/   20] | d_loss: 0.1777 | g_loss: 3.8176\n",
      "Epoch [   15/   20] | d_loss: 0.1143 | g_loss: 4.2812\n",
      "Epoch [   15/   20] | d_loss: 1.0762 | g_loss: 0.7047\n",
      "Epoch [   15/   20] | d_loss: 0.3530 | g_loss: 3.1327\n",
      "Epoch [   15/   20] | d_loss: 0.2344 | g_loss: 4.0660\n",
      "Epoch [   15/   20] | d_loss: 0.2077 | g_loss: 1.7869\n",
      "Epoch [   15/   20] | d_loss: 0.3622 | g_loss: 2.2625\n",
      "Epoch [   15/   20] | d_loss: 0.2204 | g_loss: 3.6337\n",
      "Epoch [   15/   20] | d_loss: 0.4297 | g_loss: 4.8326\n",
      "Epoch [   16/   20] | d_loss: 0.6865 | g_loss: 5.5818\n",
      "Epoch [   16/   20] | d_loss: 0.1891 | g_loss: 4.8078\n",
      "Epoch [   16/   20] | d_loss: 0.2161 | g_loss: 6.0034\n",
      "Epoch [   16/   20] | d_loss: 0.2381 | g_loss: 3.6575\n",
      "Epoch [   16/   20] | d_loss: 0.3694 | g_loss: 3.9015\n",
      "Epoch [   16/   20] | d_loss: 0.2431 | g_loss: 3.1535\n",
      "Epoch [   16/   20] | d_loss: 1.3999 | g_loss: 3.4146\n",
      "Epoch [   16/   20] | d_loss: 0.2059 | g_loss: 4.5305\n",
      "Epoch [   16/   20] | d_loss: 0.2500 | g_loss: 2.4958\n",
      "Epoch [   16/   20] | d_loss: 0.0735 | g_loss: 2.4528\n",
      "Epoch [   16/   20] | d_loss: 0.1264 | g_loss: 3.3134\n",
      "Epoch [   16/   20] | d_loss: 0.1658 | g_loss: 2.9290\n",
      "Epoch [   16/   20] | d_loss: 0.3939 | g_loss: 2.9181\n",
      "Epoch [   16/   20] | d_loss: 0.4216 | g_loss: 3.2930\n",
      "Epoch [   16/   20] | d_loss: 0.3972 | g_loss: 5.3377\n",
      "Epoch [   16/   20] | d_loss: 0.1371 | g_loss: 2.7428\n",
      "Epoch [   16/   20] | d_loss: 0.1432 | g_loss: 3.8643\n",
      "Epoch [   16/   20] | d_loss: 0.1650 | g_loss: 3.2667\n",
      "Epoch [   16/   20] | d_loss: 0.7852 | g_loss: 3.1042\n",
      "Epoch [   16/   20] | d_loss: 0.3320 | g_loss: 3.2234\n",
      "Epoch [   16/   20] | d_loss: 0.1853 | g_loss: 3.2217\n",
      "Epoch [   16/   20] | d_loss: 0.1370 | g_loss: 3.2382\n",
      "Epoch [   16/   20] | d_loss: 0.3987 | g_loss: 4.5718\n",
      "Epoch [   16/   20] | d_loss: 0.2405 | g_loss: 2.8117\n",
      "Epoch [   16/   20] | d_loss: 0.0521 | g_loss: 2.9403\n",
      "Epoch [   16/   20] | d_loss: 0.0548 | g_loss: 5.1889\n",
      "Epoch [   16/   20] | d_loss: 0.2498 | g_loss: 2.3507\n",
      "Epoch [   16/   20] | d_loss: 0.1263 | g_loss: 3.1487\n",
      "Epoch [   16/   20] | d_loss: 0.3661 | g_loss: 3.1730\n",
      "Epoch [   16/   20] | d_loss: 0.1608 | g_loss: 3.2988\n",
      "Epoch [   16/   20] | d_loss: 0.5030 | g_loss: 4.3030\n",
      "Epoch [   16/   20] | d_loss: 0.6538 | g_loss: 4.7371\n",
      "Epoch [   16/   20] | d_loss: 0.2782 | g_loss: 5.4347\n",
      "Epoch [   16/   20] | d_loss: 0.1771 | g_loss: 3.8887\n",
      "Epoch [   16/   20] | d_loss: 0.0431 | g_loss: 3.7536\n",
      "Epoch [   16/   20] | d_loss: 0.6400 | g_loss: 3.4067\n",
      "Epoch [   16/   20] | d_loss: 0.2278 | g_loss: 3.0883\n",
      "Epoch [   16/   20] | d_loss: 0.1155 | g_loss: 3.2051\n",
      "Epoch [   16/   20] | d_loss: 0.5424 | g_loss: 3.3160\n",
      "Epoch [   16/   20] | d_loss: 0.3686 | g_loss: 3.0438\n",
      "Epoch [   16/   20] | d_loss: 0.3866 | g_loss: 4.5600\n",
      "Epoch [   16/   20] | d_loss: 0.2834 | g_loss: 4.5421\n",
      "Epoch [   16/   20] | d_loss: 0.5173 | g_loss: 3.6166\n",
      "Epoch [   16/   20] | d_loss: 0.4183 | g_loss: 3.0108\n",
      "Epoch [   16/   20] | d_loss: 0.4165 | g_loss: 3.0470\n",
      "Epoch [   16/   20] | d_loss: 0.4046 | g_loss: 4.8360\n",
      "Epoch [   16/   20] | d_loss: 0.5426 | g_loss: 2.7225\n",
      "Epoch [   16/   20] | d_loss: 0.4660 | g_loss: 2.8646\n",
      "Epoch [   16/   20] | d_loss: 0.1059 | g_loss: 3.7596\n",
      "Epoch [   16/   20] | d_loss: 0.5403 | g_loss: 2.0248\n",
      "Epoch [   16/   20] | d_loss: 0.1942 | g_loss: 1.8727\n",
      "Epoch [   16/   20] | d_loss: 0.0951 | g_loss: 4.8502\n",
      "Epoch [   16/   20] | d_loss: 0.1125 | g_loss: 2.9599\n",
      "Epoch [   16/   20] | d_loss: 0.3915 | g_loss: 3.6664\n",
      "Epoch [   16/   20] | d_loss: 0.1042 | g_loss: 3.6127\n",
      "Epoch [   16/   20] | d_loss: 1.0541 | g_loss: 4.6377\n",
      "Epoch [   16/   20] | d_loss: 0.3697 | g_loss: 2.9583\n",
      "Epoch [   16/   20] | d_loss: 0.3684 | g_loss: 4.2946\n",
      "Epoch [   16/   20] | d_loss: 0.4011 | g_loss: 2.6332\n",
      "Epoch [   16/   20] | d_loss: 0.9136 | g_loss: 1.8356\n",
      "Epoch [   16/   20] | d_loss: 0.1720 | g_loss: 3.0980\n",
      "Epoch [   16/   20] | d_loss: 0.5771 | g_loss: 5.7094\n",
      "Epoch [   16/   20] | d_loss: 0.1553 | g_loss: 4.6427\n",
      "Epoch [   16/   20] | d_loss: 0.1841 | g_loss: 3.3533\n",
      "Epoch [   16/   20] | d_loss: 0.5837 | g_loss: 1.8418\n",
      "Epoch [   16/   20] | d_loss: 0.7291 | g_loss: 5.1453\n",
      "Epoch [   16/   20] | d_loss: 0.2767 | g_loss: 2.8353\n",
      "Epoch [   16/   20] | d_loss: 0.3141 | g_loss: 3.7415\n",
      "Epoch [   16/   20] | d_loss: 0.2590 | g_loss: 3.3210\n",
      "Epoch [   16/   20] | d_loss: 0.3758 | g_loss: 2.9890\n",
      "Epoch [   16/   20] | d_loss: 0.2578 | g_loss: 4.6562\n",
      "Epoch [   16/   20] | d_loss: 0.2463 | g_loss: 5.3737\n",
      "Epoch [   16/   20] | d_loss: 0.5280 | g_loss: 3.2587\n",
      "Epoch [   16/   20] | d_loss: 0.1112 | g_loss: 1.6529\n",
      "Epoch [   16/   20] | d_loss: 0.2332 | g_loss: 3.1031\n",
      "Epoch [   16/   20] | d_loss: 0.4796 | g_loss: 2.8872\n",
      "Epoch [   16/   20] | d_loss: 0.0486 | g_loss: 3.2443\n",
      "Epoch [   16/   20] | d_loss: 0.2552 | g_loss: 4.0361\n",
      "Epoch [   16/   20] | d_loss: 0.3537 | g_loss: 3.2921\n",
      "Epoch [   16/   20] | d_loss: 0.0377 | g_loss: 3.5340\n",
      "Epoch [   16/   20] | d_loss: 1.8967 | g_loss: 6.0478\n",
      "Epoch [   16/   20] | d_loss: 0.2697 | g_loss: 2.8376\n",
      "Epoch [   16/   20] | d_loss: 0.4767 | g_loss: 3.0326\n",
      "Epoch [   16/   20] | d_loss: 0.3630 | g_loss: 5.0858\n",
      "Epoch [   16/   20] | d_loss: 0.0267 | g_loss: 5.3897\n",
      "Epoch [   16/   20] | d_loss: 0.2823 | g_loss: 2.8237\n",
      "Epoch [   16/   20] | d_loss: 0.0793 | g_loss: 3.0083\n",
      "Epoch [   16/   20] | d_loss: 0.4176 | g_loss: 3.8931\n",
      "Epoch [   16/   20] | d_loss: 0.1324 | g_loss: 5.7824\n",
      "Epoch [   16/   20] | d_loss: 0.3554 | g_loss: 2.9703\n",
      "Epoch [   16/   20] | d_loss: 0.7462 | g_loss: 3.7860\n",
      "Epoch [   16/   20] | d_loss: 0.1732 | g_loss: 3.4758\n",
      "Epoch [   16/   20] | d_loss: 0.3934 | g_loss: 4.8301\n",
      "Epoch [   16/   20] | d_loss: 0.0616 | g_loss: 3.9872\n",
      "Epoch [   16/   20] | d_loss: 0.9802 | g_loss: 4.3816\n",
      "Epoch [   16/   20] | d_loss: 0.4133 | g_loss: 2.3264\n",
      "Epoch [   16/   20] | d_loss: 0.0939 | g_loss: 5.5789\n",
      "Epoch [   16/   20] | d_loss: 0.0610 | g_loss: 4.1795\n",
      "Epoch [   16/   20] | d_loss: 0.1327 | g_loss: 4.5378\n",
      "Epoch [   16/   20] | d_loss: 0.4003 | g_loss: 2.2029\n",
      "Epoch [   16/   20] | d_loss: 0.2898 | g_loss: 2.4617\n",
      "Epoch [   16/   20] | d_loss: 0.5281 | g_loss: 2.9867\n",
      "Epoch [   16/   20] | d_loss: 0.2304 | g_loss: 3.9322\n",
      "Epoch [   16/   20] | d_loss: 0.4750 | g_loss: 2.1810\n",
      "Epoch [   16/   20] | d_loss: 0.2882 | g_loss: 7.0891\n",
      "Epoch [   16/   20] | d_loss: 0.7777 | g_loss: 3.3352\n",
      "Epoch [   16/   20] | d_loss: 0.4439 | g_loss: 2.8039\n",
      "Epoch [   16/   20] | d_loss: 1.3308 | g_loss: 1.6838\n",
      "Epoch [   16/   20] | d_loss: 0.6290 | g_loss: 0.7117\n",
      "Epoch [   16/   20] | d_loss: 0.0457 | g_loss: 4.2690\n",
      "Epoch [   16/   20] | d_loss: 0.2492 | g_loss: 2.1016\n",
      "Epoch [   16/   20] | d_loss: 0.5687 | g_loss: 4.5957\n",
      "Epoch [   16/   20] | d_loss: 0.1990 | g_loss: 2.6115\n",
      "Epoch [   17/   20] | d_loss: 0.4579 | g_loss: 2.5124\n",
      "Epoch [   17/   20] | d_loss: 0.2394 | g_loss: 3.6432\n",
      "Epoch [   17/   20] | d_loss: 1.1741 | g_loss: 1.5985\n",
      "Epoch [   17/   20] | d_loss: 0.1865 | g_loss: 2.8019\n",
      "Epoch [   17/   20] | d_loss: 0.2973 | g_loss: 4.3172\n",
      "Epoch [   17/   20] | d_loss: 0.1758 | g_loss: 2.0447\n",
      "Epoch [   17/   20] | d_loss: 0.9075 | g_loss: 3.1920\n",
      "Epoch [   17/   20] | d_loss: 0.2608 | g_loss: 4.0483\n",
      "Epoch [   17/   20] | d_loss: 0.3379 | g_loss: 1.5314\n",
      "Epoch [   17/   20] | d_loss: 0.2394 | g_loss: 4.4589\n",
      "Epoch [   17/   20] | d_loss: 0.3033 | g_loss: 3.7634\n",
      "Epoch [   17/   20] | d_loss: 0.4083 | g_loss: 3.7473\n",
      "Epoch [   17/   20] | d_loss: 0.3384 | g_loss: 2.6459\n",
      "Epoch [   17/   20] | d_loss: 0.3670 | g_loss: 3.0686\n",
      "Epoch [   17/   20] | d_loss: 0.5225 | g_loss: 3.0417\n",
      "Epoch [   17/   20] | d_loss: 0.3062 | g_loss: 3.5345\n",
      "Epoch [   17/   20] | d_loss: 0.5909 | g_loss: 2.8979\n",
      "Epoch [   17/   20] | d_loss: 0.9252 | g_loss: 5.8993\n",
      "Epoch [   17/   20] | d_loss: 0.0474 | g_loss: 3.4592\n",
      "Epoch [   17/   20] | d_loss: 0.5298 | g_loss: 3.0772\n",
      "Epoch [   17/   20] | d_loss: 0.3199 | g_loss: 0.7575\n",
      "Epoch [   17/   20] | d_loss: 0.5905 | g_loss: 3.7344\n",
      "Epoch [   17/   20] | d_loss: 0.4056 | g_loss: 1.4698\n",
      "Epoch [   17/   20] | d_loss: 0.2514 | g_loss: 2.3959\n",
      "Epoch [   17/   20] | d_loss: 0.2220 | g_loss: 4.6872\n",
      "Epoch [   17/   20] | d_loss: 0.3140 | g_loss: 4.5171\n",
      "Epoch [   17/   20] | d_loss: 0.1782 | g_loss: 2.2489\n",
      "Epoch [   17/   20] | d_loss: 0.6304 | g_loss: 3.7311\n",
      "Epoch [   17/   20] | d_loss: 0.1750 | g_loss: 4.1095\n",
      "Epoch [   17/   20] | d_loss: 0.1213 | g_loss: 4.7643\n",
      "Epoch [   17/   20] | d_loss: 0.0986 | g_loss: 4.0768\n",
      "Epoch [   17/   20] | d_loss: 0.2829 | g_loss: 1.8201\n",
      "Epoch [   17/   20] | d_loss: 0.2993 | g_loss: 2.2563\n",
      "Epoch [   17/   20] | d_loss: 0.4875 | g_loss: 3.7315\n",
      "Epoch [   17/   20] | d_loss: 0.3805 | g_loss: 2.5894\n",
      "Epoch [   17/   20] | d_loss: 0.5178 | g_loss: 3.9617\n",
      "Epoch [   17/   20] | d_loss: 0.2456 | g_loss: 2.8532\n",
      "Epoch [   17/   20] | d_loss: 0.1621 | g_loss: 3.6046\n",
      "Epoch [   17/   20] | d_loss: 0.2411 | g_loss: 3.3542\n",
      "Epoch [   17/   20] | d_loss: 0.4586 | g_loss: 2.0347\n",
      "Epoch [   17/   20] | d_loss: 0.2299 | g_loss: 3.9248\n",
      "Epoch [   17/   20] | d_loss: 0.2869 | g_loss: 0.6119\n",
      "Epoch [   17/   20] | d_loss: 0.5850 | g_loss: 4.5225\n",
      "Epoch [   17/   20] | d_loss: 0.5083 | g_loss: 3.2699\n",
      "Epoch [   17/   20] | d_loss: 0.1753 | g_loss: 5.9854\n",
      "Epoch [   17/   20] | d_loss: 0.0584 | g_loss: 4.5156\n",
      "Epoch [   17/   20] | d_loss: 0.4184 | g_loss: 3.2364\n",
      "Epoch [   17/   20] | d_loss: 0.1447 | g_loss: 4.1323\n",
      "Epoch [   17/   20] | d_loss: 0.1412 | g_loss: 6.3691\n",
      "Epoch [   17/   20] | d_loss: 0.3670 | g_loss: 3.7454\n",
      "Epoch [   17/   20] | d_loss: 0.0618 | g_loss: 7.5563\n",
      "Epoch [   17/   20] | d_loss: 0.1746 | g_loss: 3.8206\n",
      "Epoch [   17/   20] | d_loss: 0.3641 | g_loss: 4.4296\n",
      "Epoch [   17/   20] | d_loss: 0.5568 | g_loss: 4.2328\n",
      "Epoch [   17/   20] | d_loss: 0.2995 | g_loss: 3.1581\n",
      "Epoch [   17/   20] | d_loss: 0.1034 | g_loss: 3.4795\n",
      "Epoch [   17/   20] | d_loss: 0.3307 | g_loss: 4.0866\n",
      "Epoch [   17/   20] | d_loss: 0.1337 | g_loss: 1.8650\n",
      "Epoch [   17/   20] | d_loss: 0.7767 | g_loss: 3.8930\n",
      "Epoch [   17/   20] | d_loss: 0.2233 | g_loss: 2.2600\n",
      "Epoch [   17/   20] | d_loss: 0.5142 | g_loss: 2.1597\n",
      "Epoch [   17/   20] | d_loss: 0.1424 | g_loss: 4.4144\n",
      "Epoch [   17/   20] | d_loss: 0.1086 | g_loss: 2.9492\n",
      "Epoch [   17/   20] | d_loss: 0.5293 | g_loss: 4.4672\n",
      "Epoch [   17/   20] | d_loss: 0.7168 | g_loss: 1.8145\n",
      "Epoch [   17/   20] | d_loss: 0.2360 | g_loss: 1.5426\n",
      "Epoch [   17/   20] | d_loss: 0.2436 | g_loss: 2.6379\n",
      "Epoch [   17/   20] | d_loss: 0.4311 | g_loss: 4.1469\n",
      "Epoch [   17/   20] | d_loss: 0.5781 | g_loss: 3.7492\n",
      "Epoch [   17/   20] | d_loss: 0.4939 | g_loss: 2.1714\n",
      "Epoch [   17/   20] | d_loss: 0.4308 | g_loss: 2.0595\n",
      "Epoch [   17/   20] | d_loss: 0.4577 | g_loss: 4.2093\n",
      "Epoch [   17/   20] | d_loss: 0.4248 | g_loss: 2.9759\n",
      "Epoch [   17/   20] | d_loss: 0.2522 | g_loss: 3.9824\n",
      "Epoch [   17/   20] | d_loss: 0.1496 | g_loss: 4.3987\n",
      "Epoch [   17/   20] | d_loss: 0.2312 | g_loss: 3.8625\n",
      "Epoch [   17/   20] | d_loss: 0.1168 | g_loss: 4.2314\n",
      "Epoch [   17/   20] | d_loss: 0.2511 | g_loss: 1.0646\n",
      "Epoch [   17/   20] | d_loss: 0.4544 | g_loss: 1.6973\n",
      "Epoch [   17/   20] | d_loss: 0.2132 | g_loss: 3.0572\n",
      "Epoch [   17/   20] | d_loss: 0.1331 | g_loss: 3.7779\n",
      "Epoch [   17/   20] | d_loss: 0.1997 | g_loss: 4.3508\n",
      "Epoch [   17/   20] | d_loss: 0.1694 | g_loss: 5.0189\n",
      "Epoch [   17/   20] | d_loss: 0.0796 | g_loss: 3.5665\n",
      "Epoch [   17/   20] | d_loss: 0.8962 | g_loss: 3.9776\n",
      "Epoch [   17/   20] | d_loss: 0.3433 | g_loss: 3.7564\n",
      "Epoch [   17/   20] | d_loss: 0.0516 | g_loss: 6.4256\n",
      "Epoch [   17/   20] | d_loss: 0.3949 | g_loss: 3.0529\n",
      "Epoch [   17/   20] | d_loss: 0.4355 | g_loss: 3.6198\n",
      "Epoch [   17/   20] | d_loss: 0.1395 | g_loss: 6.2581\n",
      "Epoch [   17/   20] | d_loss: 0.1231 | g_loss: 4.9383\n",
      "Epoch [   17/   20] | d_loss: 0.9381 | g_loss: 5.6960\n",
      "Epoch [   17/   20] | d_loss: 0.6346 | g_loss: 2.0319\n",
      "Epoch [   17/   20] | d_loss: 0.3052 | g_loss: 2.4198\n",
      "Epoch [   17/   20] | d_loss: 0.7557 | g_loss: 2.1931\n",
      "Epoch [   17/   20] | d_loss: 0.2475 | g_loss: 3.6771\n",
      "Epoch [   17/   20] | d_loss: 0.9936 | g_loss: 1.9088\n",
      "Epoch [   17/   20] | d_loss: 0.0629 | g_loss: 2.5099\n",
      "Epoch [   17/   20] | d_loss: 0.2308 | g_loss: 2.5676\n",
      "Epoch [   17/   20] | d_loss: 0.4541 | g_loss: 3.2493\n",
      "Epoch [   17/   20] | d_loss: 0.2510 | g_loss: 3.6749\n",
      "Epoch [   17/   20] | d_loss: 0.2006 | g_loss: 3.5099\n",
      "Epoch [   17/   20] | d_loss: 0.1557 | g_loss: 3.4150\n",
      "Epoch [   17/   20] | d_loss: 0.1718 | g_loss: 4.4859\n",
      "Epoch [   17/   20] | d_loss: 0.1425 | g_loss: 3.8457\n",
      "Epoch [   17/   20] | d_loss: 0.0671 | g_loss: 3.2751\n",
      "Epoch [   17/   20] | d_loss: 0.2173 | g_loss: 5.5249\n",
      "Epoch [   17/   20] | d_loss: 0.0347 | g_loss: 4.1831\n",
      "Epoch [   17/   20] | d_loss: 0.1510 | g_loss: 2.9875\n",
      "Epoch [   17/   20] | d_loss: 0.4873 | g_loss: 3.5308\n",
      "Epoch [   17/   20] | d_loss: 0.2553 | g_loss: 3.4247\n",
      "Epoch [   17/   20] | d_loss: 0.1719 | g_loss: 3.4679\n",
      "Epoch [   17/   20] | d_loss: 0.3965 | g_loss: 4.2024\n",
      "Epoch [   18/   20] | d_loss: 0.0265 | g_loss: 4.6178\n",
      "Epoch [   18/   20] | d_loss: 0.0271 | g_loss: 3.9280\n",
      "Epoch [   18/   20] | d_loss: 0.0648 | g_loss: 2.8881\n",
      "Epoch [   18/   20] | d_loss: 0.2324 | g_loss: 3.5920\n",
      "Epoch [   18/   20] | d_loss: 0.2042 | g_loss: 3.4072\n",
      "Epoch [   18/   20] | d_loss: 0.1647 | g_loss: 4.7169\n",
      "Epoch [   18/   20] | d_loss: 0.2279 | g_loss: 1.2387\n",
      "Epoch [   18/   20] | d_loss: 0.5288 | g_loss: 1.9852\n",
      "Epoch [   18/   20] | d_loss: 0.4472 | g_loss: 4.5714\n",
      "Epoch [   18/   20] | d_loss: 0.2766 | g_loss: 3.1421\n",
      "Epoch [   18/   20] | d_loss: 0.1940 | g_loss: 3.9253\n",
      "Epoch [   18/   20] | d_loss: 0.2666 | g_loss: 2.6128\n",
      "Epoch [   18/   20] | d_loss: 0.5677 | g_loss: 1.4719\n",
      "Epoch [   18/   20] | d_loss: 0.1129 | g_loss: 4.2347\n",
      "Epoch [   18/   20] | d_loss: 0.4635 | g_loss: 1.9145\n",
      "Epoch [   18/   20] | d_loss: 0.6168 | g_loss: 4.5518\n",
      "Epoch [   18/   20] | d_loss: 0.2947 | g_loss: 5.5876\n",
      "Epoch [   18/   20] | d_loss: 0.0414 | g_loss: 4.3461\n",
      "Epoch [   18/   20] | d_loss: 0.1932 | g_loss: 1.5117\n",
      "Epoch [   18/   20] | d_loss: 0.2243 | g_loss: 2.8161\n",
      "Epoch [   18/   20] | d_loss: 0.8462 | g_loss: 5.9382\n",
      "Epoch [   18/   20] | d_loss: 1.4549 | g_loss: 1.9504\n",
      "Epoch [   18/   20] | d_loss: 0.3529 | g_loss: 4.1728\n",
      "Epoch [   18/   20] | d_loss: 0.1913 | g_loss: 2.9773\n",
      "Epoch [   18/   20] | d_loss: 0.2402 | g_loss: 4.7320\n",
      "Epoch [   18/   20] | d_loss: 0.6476 | g_loss: 2.8145\n",
      "Epoch [   18/   20] | d_loss: 0.7530 | g_loss: 3.3859\n",
      "Epoch [   18/   20] | d_loss: 1.0795 | g_loss: 4.2948\n",
      "Epoch [   18/   20] | d_loss: 0.1920 | g_loss: 1.1562\n",
      "Epoch [   18/   20] | d_loss: 0.2382 | g_loss: 2.2562\n",
      "Epoch [   18/   20] | d_loss: 0.3620 | g_loss: 2.1297\n",
      "Epoch [   18/   20] | d_loss: 0.4905 | g_loss: 4.5051\n",
      "Epoch [   18/   20] | d_loss: 0.4471 | g_loss: 3.3374\n",
      "Epoch [   18/   20] | d_loss: 0.4733 | g_loss: 3.7292\n",
      "Epoch [   18/   20] | d_loss: 0.5087 | g_loss: 4.3010\n",
      "Epoch [   18/   20] | d_loss: 0.0538 | g_loss: 4.2061\n",
      "Epoch [   18/   20] | d_loss: 0.1937 | g_loss: 2.6255\n",
      "Epoch [   18/   20] | d_loss: 0.3056 | g_loss: 1.8706\n",
      "Epoch [   18/   20] | d_loss: 0.3401 | g_loss: 1.9563\n",
      "Epoch [   18/   20] | d_loss: 0.4243 | g_loss: 3.5540\n",
      "Epoch [   18/   20] | d_loss: 0.3537 | g_loss: 5.6437\n",
      "Epoch [   18/   20] | d_loss: 0.4131 | g_loss: 4.7842\n",
      "Epoch [   18/   20] | d_loss: 0.0952 | g_loss: 3.7385\n",
      "Epoch [   18/   20] | d_loss: 0.4411 | g_loss: 1.8257\n",
      "Epoch [   18/   20] | d_loss: 0.0701 | g_loss: 1.7857\n",
      "Epoch [   18/   20] | d_loss: 0.1581 | g_loss: 5.0752\n",
      "Epoch [   18/   20] | d_loss: 0.6695 | g_loss: 2.0335\n",
      "Epoch [   18/   20] | d_loss: 0.6008 | g_loss: 1.1781\n",
      "Epoch [   18/   20] | d_loss: 0.3619 | g_loss: 5.3500\n",
      "Epoch [   18/   20] | d_loss: 0.7489 | g_loss: 3.1274\n",
      "Epoch [   18/   20] | d_loss: 0.3489 | g_loss: 5.3812\n",
      "Epoch [   18/   20] | d_loss: 0.7589 | g_loss: 1.5818\n",
      "Epoch [   18/   20] | d_loss: 0.4527 | g_loss: 3.2286\n",
      "Epoch [   18/   20] | d_loss: 0.1227 | g_loss: 2.9706\n",
      "Epoch [   18/   20] | d_loss: 0.7137 | g_loss: 2.4436\n",
      "Epoch [   18/   20] | d_loss: 0.1964 | g_loss: 2.3794\n",
      "Epoch [   18/   20] | d_loss: 0.1875 | g_loss: 3.9590\n",
      "Epoch [   18/   20] | d_loss: 0.1477 | g_loss: 4.9025\n",
      "Epoch [   18/   20] | d_loss: 0.1991 | g_loss: 3.4024\n",
      "Epoch [   18/   20] | d_loss: 0.0296 | g_loss: 3.0906\n",
      "Epoch [   18/   20] | d_loss: 0.3221 | g_loss: 3.4785\n",
      "Epoch [   18/   20] | d_loss: 0.8777 | g_loss: 3.4128\n",
      "Epoch [   18/   20] | d_loss: 0.1172 | g_loss: 4.2389\n",
      "Epoch [   18/   20] | d_loss: 0.5133 | g_loss: 4.2371\n",
      "Epoch [   18/   20] | d_loss: 0.3428 | g_loss: 2.7830\n",
      "Epoch [   18/   20] | d_loss: 0.4419 | g_loss: 5.6109\n",
      "Epoch [   18/   20] | d_loss: 0.0972 | g_loss: 6.2326\n",
      "Epoch [   18/   20] | d_loss: 1.1358 | g_loss: 2.2482\n",
      "Epoch [   18/   20] | d_loss: 1.1887 | g_loss: 2.3554\n",
      "Epoch [   18/   20] | d_loss: 0.0927 | g_loss: 4.8233\n",
      "Epoch [   18/   20] | d_loss: 0.1433 | g_loss: 3.9455\n",
      "Epoch [   18/   20] | d_loss: 0.6280 | g_loss: 2.9213\n",
      "Epoch [   18/   20] | d_loss: 0.3869 | g_loss: 4.8215\n",
      "Epoch [   18/   20] | d_loss: 0.4026 | g_loss: 2.3528\n",
      "Epoch [   18/   20] | d_loss: 0.0292 | g_loss: 4.8936\n",
      "Epoch [   18/   20] | d_loss: 0.0853 | g_loss: 3.2883\n",
      "Epoch [   18/   20] | d_loss: 0.3474 | g_loss: 4.1842\n",
      "Epoch [   18/   20] | d_loss: 0.3862 | g_loss: 7.2515\n",
      "Epoch [   18/   20] | d_loss: 0.1129 | g_loss: 4.5559\n",
      "Epoch [   18/   20] | d_loss: 0.2973 | g_loss: 5.3219\n",
      "Epoch [   18/   20] | d_loss: 0.1854 | g_loss: 1.6279\n",
      "Epoch [   18/   20] | d_loss: 0.2079 | g_loss: 1.7124\n",
      "Epoch [   18/   20] | d_loss: 0.5175 | g_loss: 1.5356\n",
      "Epoch [   18/   20] | d_loss: 0.1969 | g_loss: 1.9602\n",
      "Epoch [   18/   20] | d_loss: 0.1314 | g_loss: 3.5660\n",
      "Epoch [   18/   20] | d_loss: 0.2378 | g_loss: 6.0185\n",
      "Epoch [   18/   20] | d_loss: 0.1481 | g_loss: 2.5137\n",
      "Epoch [   18/   20] | d_loss: 0.1583 | g_loss: 5.8755\n",
      "Epoch [   18/   20] | d_loss: 0.3004 | g_loss: 5.1247\n",
      "Epoch [   18/   20] | d_loss: 0.8278 | g_loss: 5.0506\n",
      "Epoch [   18/   20] | d_loss: 1.2696 | g_loss: 5.0675\n",
      "Epoch [   18/   20] | d_loss: 0.2340 | g_loss: 3.4072\n",
      "Epoch [   18/   20] | d_loss: 0.2100 | g_loss: 3.4976\n",
      "Epoch [   18/   20] | d_loss: 1.0666 | g_loss: 4.7039\n",
      "Epoch [   18/   20] | d_loss: 0.0597 | g_loss: 3.3731\n",
      "Epoch [   18/   20] | d_loss: 0.6271 | g_loss: 5.6595\n",
      "Epoch [   18/   20] | d_loss: 0.5851 | g_loss: 2.2595\n",
      "Epoch [   18/   20] | d_loss: 0.5815 | g_loss: 2.2095\n",
      "Epoch [   18/   20] | d_loss: 0.3629 | g_loss: 4.0634\n",
      "Epoch [   18/   20] | d_loss: 0.5925 | g_loss: 2.9485\n",
      "Epoch [   18/   20] | d_loss: 0.1034 | g_loss: 4.4440\n",
      "Epoch [   18/   20] | d_loss: 0.2984 | g_loss: 2.6862\n",
      "Epoch [   18/   20] | d_loss: 0.3871 | g_loss: 2.7999\n",
      "Epoch [   18/   20] | d_loss: 0.1798 | g_loss: 3.1695\n",
      "Epoch [   18/   20] | d_loss: 0.4366 | g_loss: 2.6318\n",
      "Epoch [   18/   20] | d_loss: 0.2206 | g_loss: 3.9405\n",
      "Epoch [   18/   20] | d_loss: 0.2069 | g_loss: 3.3524\n",
      "Epoch [   18/   20] | d_loss: 0.2885 | g_loss: 3.8165\n",
      "Epoch [   18/   20] | d_loss: 0.3291 | g_loss: 2.5018\n",
      "Epoch [   18/   20] | d_loss: 0.8326 | g_loss: 2.1960\n",
      "Epoch [   18/   20] | d_loss: 0.5776 | g_loss: 3.4731\n",
      "Epoch [   18/   20] | d_loss: 0.3715 | g_loss: 4.0428\n",
      "Epoch [   18/   20] | d_loss: 0.5048 | g_loss: 3.5086\n",
      "Epoch [   19/   20] | d_loss: 1.2450 | g_loss: 1.9785\n",
      "Epoch [   19/   20] | d_loss: 0.5784 | g_loss: 2.8497\n",
      "Epoch [   19/   20] | d_loss: 0.2813 | g_loss: 4.1553\n",
      "Epoch [   19/   20] | d_loss: 0.1294 | g_loss: 3.5399\n",
      "Epoch [   19/   20] | d_loss: 0.6738 | g_loss: 4.7395\n",
      "Epoch [   19/   20] | d_loss: 0.0711 | g_loss: 5.6721\n",
      "Epoch [   19/   20] | d_loss: 0.2362 | g_loss: 3.1460\n",
      "Epoch [   19/   20] | d_loss: 0.2317 | g_loss: 2.5083\n",
      "Epoch [   19/   20] | d_loss: 0.6166 | g_loss: 4.4117\n",
      "Epoch [   19/   20] | d_loss: 0.2133 | g_loss: 3.2339\n",
      "Epoch [   19/   20] | d_loss: 1.3876 | g_loss: 0.5195\n",
      "Epoch [   19/   20] | d_loss: 1.2801 | g_loss: 2.7425\n",
      "Epoch [   19/   20] | d_loss: 0.2140 | g_loss: 4.1231\n",
      "Epoch [   19/   20] | d_loss: 1.6646 | g_loss: 4.2806\n",
      "Epoch [   19/   20] | d_loss: 0.2077 | g_loss: 3.9067\n",
      "Epoch [   19/   20] | d_loss: 0.6619 | g_loss: 5.4127\n",
      "Epoch [   19/   20] | d_loss: 0.3860 | g_loss: 2.8227\n",
      "Epoch [   19/   20] | d_loss: 0.2888 | g_loss: 2.6167\n",
      "Epoch [   19/   20] | d_loss: 0.1658 | g_loss: 2.1361\n",
      "Epoch [   19/   20] | d_loss: 1.1850 | g_loss: 5.5053\n",
      "Epoch [   19/   20] | d_loss: 0.1545 | g_loss: 2.6787\n",
      "Epoch [   19/   20] | d_loss: 0.2670 | g_loss: 1.4585\n",
      "Epoch [   19/   20] | d_loss: 0.1713 | g_loss: 2.5073\n",
      "Epoch [   19/   20] | d_loss: 0.3664 | g_loss: 3.7767\n",
      "Epoch [   19/   20] | d_loss: 0.0311 | g_loss: 5.1434\n",
      "Epoch [   19/   20] | d_loss: 0.0908 | g_loss: 3.8220\n",
      "Epoch [   19/   20] | d_loss: 0.3523 | g_loss: 4.8772\n",
      "Epoch [   19/   20] | d_loss: 0.3355 | g_loss: 2.4516\n",
      "Epoch [   19/   20] | d_loss: 0.6540 | g_loss: 2.0120\n",
      "Epoch [   19/   20] | d_loss: 0.0983 | g_loss: 5.0068\n",
      "Epoch [   19/   20] | d_loss: 0.0584 | g_loss: 4.7142\n",
      "Epoch [   19/   20] | d_loss: 0.0589 | g_loss: 4.7230\n",
      "Epoch [   19/   20] | d_loss: 0.3114 | g_loss: 5.0181\n",
      "Epoch [   19/   20] | d_loss: 0.3484 | g_loss: 4.1043\n",
      "Epoch [   19/   20] | d_loss: 0.1765 | g_loss: 5.1688\n",
      "Epoch [   19/   20] | d_loss: 0.2454 | g_loss: 4.0702\n",
      "Epoch [   19/   20] | d_loss: 0.6595 | g_loss: 2.7170\n",
      "Epoch [   19/   20] | d_loss: 0.2199 | g_loss: 4.2043\n",
      "Epoch [   19/   20] | d_loss: 0.0324 | g_loss: 2.8351\n",
      "Epoch [   19/   20] | d_loss: 0.3488 | g_loss: 6.4697\n",
      "Epoch [   19/   20] | d_loss: 0.0606 | g_loss: 5.2742\n",
      "Epoch [   19/   20] | d_loss: 0.6447 | g_loss: 2.0730\n",
      "Epoch [   19/   20] | d_loss: 0.8982 | g_loss: 3.9780\n",
      "Epoch [   19/   20] | d_loss: 0.1127 | g_loss: 6.7021\n",
      "Epoch [   19/   20] | d_loss: 1.3693 | g_loss: 2.0139\n",
      "Epoch [   19/   20] | d_loss: 0.4804 | g_loss: 1.2217\n",
      "Epoch [   19/   20] | d_loss: 0.1487 | g_loss: 3.4244\n",
      "Epoch [   19/   20] | d_loss: 0.4217 | g_loss: 2.9287\n",
      "Epoch [   19/   20] | d_loss: 0.1499 | g_loss: 2.6960\n",
      "Epoch [   19/   20] | d_loss: 0.7788 | g_loss: 2.9905\n",
      "Epoch [   19/   20] | d_loss: 0.3154 | g_loss: 2.2796\n",
      "Epoch [   19/   20] | d_loss: 0.2502 | g_loss: 3.7420\n",
      "Epoch [   19/   20] | d_loss: 0.0687 | g_loss: 3.4553\n",
      "Epoch [   19/   20] | d_loss: 0.2802 | g_loss: 3.7817\n",
      "Epoch [   19/   20] | d_loss: 0.1534 | g_loss: 2.3694\n",
      "Epoch [   19/   20] | d_loss: 0.3530 | g_loss: 2.5481\n",
      "Epoch [   19/   20] | d_loss: 0.2414 | g_loss: 4.0359\n",
      "Epoch [   19/   20] | d_loss: 0.1735 | g_loss: 6.0172\n",
      "Epoch [   19/   20] | d_loss: 0.4438 | g_loss: 5.7676\n",
      "Epoch [   19/   20] | d_loss: 0.2018 | g_loss: 4.8877\n",
      "Epoch [   19/   20] | d_loss: 0.1869 | g_loss: 6.3499\n",
      "Epoch [   19/   20] | d_loss: 0.2831 | g_loss: 3.1779\n",
      "Epoch [   19/   20] | d_loss: 0.0769 | g_loss: 3.3381\n",
      "Epoch [   19/   20] | d_loss: 0.2984 | g_loss: 4.0797\n",
      "Epoch [   19/   20] | d_loss: 0.3138 | g_loss: 6.1274\n",
      "Epoch [   19/   20] | d_loss: 0.1597 | g_loss: 2.5621\n",
      "Epoch [   19/   20] | d_loss: 0.0606 | g_loss: 5.8277\n",
      "Epoch [   19/   20] | d_loss: 0.3973 | g_loss: 2.5645\n",
      "Epoch [   19/   20] | d_loss: 0.5182 | g_loss: 1.5693\n",
      "Epoch [   19/   20] | d_loss: 0.2942 | g_loss: 4.5481\n",
      "Epoch [   19/   20] | d_loss: 0.0490 | g_loss: 1.3559\n",
      "Epoch [   19/   20] | d_loss: 0.1175 | g_loss: 6.0418\n",
      "Epoch [   19/   20] | d_loss: 0.2519 | g_loss: 2.2551\n",
      "Epoch [   19/   20] | d_loss: 0.3092 | g_loss: 4.1333\n",
      "Epoch [   19/   20] | d_loss: 0.1268 | g_loss: 4.8720\n",
      "Epoch [   19/   20] | d_loss: 0.0507 | g_loss: 4.0791\n",
      "Epoch [   19/   20] | d_loss: 0.4282 | g_loss: 3.7147\n",
      "Epoch [   19/   20] | d_loss: 0.2416 | g_loss: 2.8063\n",
      "Epoch [   19/   20] | d_loss: 0.4095 | g_loss: 3.9985\n",
      "Epoch [   19/   20] | d_loss: 0.4039 | g_loss: 4.1981\n",
      "Epoch [   19/   20] | d_loss: 0.3261 | g_loss: 1.8163\n",
      "Epoch [   19/   20] | d_loss: 0.4710 | g_loss: 3.9749\n",
      "Epoch [   19/   20] | d_loss: 0.2770 | g_loss: 4.5083\n",
      "Epoch [   19/   20] | d_loss: 0.2208 | g_loss: 5.1562\n",
      "Epoch [   19/   20] | d_loss: 0.1290 | g_loss: 1.3612\n",
      "Epoch [   19/   20] | d_loss: 0.0443 | g_loss: 6.0440\n",
      "Epoch [   19/   20] | d_loss: 0.0716 | g_loss: 5.3099\n",
      "Epoch [   19/   20] | d_loss: 0.3162 | g_loss: 3.2862\n",
      "Epoch [   19/   20] | d_loss: 0.1950 | g_loss: 3.9454\n",
      "Epoch [   19/   20] | d_loss: 0.1722 | g_loss: 5.3902\n",
      "Epoch [   19/   20] | d_loss: 0.1373 | g_loss: 3.5585\n",
      "Epoch [   19/   20] | d_loss: 0.3175 | g_loss: 4.0666\n",
      "Epoch [   19/   20] | d_loss: 0.1298 | g_loss: 2.4728\n",
      "Epoch [   19/   20] | d_loss: 0.0540 | g_loss: 3.9403\n",
      "Epoch [   19/   20] | d_loss: 0.1417 | g_loss: 3.1668\n",
      "Epoch [   19/   20] | d_loss: 0.4624 | g_loss: 5.5705\n",
      "Epoch [   19/   20] | d_loss: 0.2812 | g_loss: 4.4479\n",
      "Epoch [   19/   20] | d_loss: 0.0284 | g_loss: 3.0225\n",
      "Epoch [   19/   20] | d_loss: 0.0595 | g_loss: 6.5414\n",
      "Epoch [   19/   20] | d_loss: 0.4144 | g_loss: 5.8881\n",
      "Epoch [   19/   20] | d_loss: 0.2347 | g_loss: 2.4115\n",
      "Epoch [   19/   20] | d_loss: 0.1089 | g_loss: 4.4573\n",
      "Epoch [   19/   20] | d_loss: 0.7453 | g_loss: 3.4182\n",
      "Epoch [   19/   20] | d_loss: 0.2111 | g_loss: 4.2285\n",
      "Epoch [   19/   20] | d_loss: 0.0818 | g_loss: 2.0974\n",
      "Epoch [   19/   20] | d_loss: 0.0820 | g_loss: 5.4574\n",
      "Epoch [   19/   20] | d_loss: 0.4238 | g_loss: 3.8211\n",
      "Epoch [   19/   20] | d_loss: 0.2330 | g_loss: 3.4029\n",
      "Epoch [   19/   20] | d_loss: 0.4256 | g_loss: 3.6292\n",
      "Epoch [   19/   20] | d_loss: 0.0578 | g_loss: 5.5164\n",
      "Epoch [   19/   20] | d_loss: 0.1680 | g_loss: 3.0753\n",
      "Epoch [   19/   20] | d_loss: 0.5043 | g_loss: 2.7954\n",
      "Epoch [   19/   20] | d_loss: 0.3739 | g_loss: 5.5988\n",
      "Epoch [   20/   20] | d_loss: 0.0587 | g_loss: 2.6085\n",
      "Epoch [   20/   20] | d_loss: 0.1865 | g_loss: 2.9512\n",
      "Epoch [   20/   20] | d_loss: 0.6187 | g_loss: 3.9844\n",
      "Epoch [   20/   20] | d_loss: 0.7818 | g_loss: 2.7299\n",
      "Epoch [   20/   20] | d_loss: 0.6116 | g_loss: 2.9356\n",
      "Epoch [   20/   20] | d_loss: 0.0390 | g_loss: 4.1925\n",
      "Epoch [   20/   20] | d_loss: 0.1960 | g_loss: 4.1173\n",
      "Epoch [   20/   20] | d_loss: 0.1890 | g_loss: 4.1090\n",
      "Epoch [   20/   20] | d_loss: 0.0997 | g_loss: 4.1312\n",
      "Epoch [   20/   20] | d_loss: 0.1866 | g_loss: 1.0258\n",
      "Epoch [   20/   20] | d_loss: 0.3313 | g_loss: 4.6203\n",
      "Epoch [   20/   20] | d_loss: 0.1156 | g_loss: 3.7816\n",
      "Epoch [   20/   20] | d_loss: 0.4622 | g_loss: 2.6807\n",
      "Epoch [   20/   20] | d_loss: 0.4520 | g_loss: 3.4024\n",
      "Epoch [   20/   20] | d_loss: 0.0378 | g_loss: 3.4190\n",
      "Epoch [   20/   20] | d_loss: 0.3427 | g_loss: 2.6026\n",
      "Epoch [   20/   20] | d_loss: 0.1344 | g_loss: 3.1891\n",
      "Epoch [   20/   20] | d_loss: 0.0206 | g_loss: 5.1162\n",
      "Epoch [   20/   20] | d_loss: 0.1165 | g_loss: 3.8263\n",
      "Epoch [   20/   20] | d_loss: 0.4542 | g_loss: 2.8750\n",
      "Epoch [   20/   20] | d_loss: 0.1489 | g_loss: 4.3295\n",
      "Epoch [   20/   20] | d_loss: 0.3214 | g_loss: 5.0584\n",
      "Epoch [   20/   20] | d_loss: 0.1441 | g_loss: 4.2097\n",
      "Epoch [   20/   20] | d_loss: 0.7591 | g_loss: 2.9741\n",
      "Epoch [   20/   20] | d_loss: 0.1975 | g_loss: 3.7248\n",
      "Epoch [   20/   20] | d_loss: 0.3276 | g_loss: 3.1182\n",
      "Epoch [   20/   20] | d_loss: 0.2292 | g_loss: 3.6033\n",
      "Epoch [   20/   20] | d_loss: 0.2384 | g_loss: 4.9433\n",
      "Epoch [   20/   20] | d_loss: 0.8294 | g_loss: 2.1334\n",
      "Epoch [   20/   20] | d_loss: 0.1640 | g_loss: 4.2953\n",
      "Epoch [   20/   20] | d_loss: 1.9554 | g_loss: 4.0406\n",
      "Epoch [   20/   20] | d_loss: 0.2305 | g_loss: 1.7981\n",
      "Epoch [   20/   20] | d_loss: 0.1664 | g_loss: 4.9968\n",
      "Epoch [   20/   20] | d_loss: 0.1432 | g_loss: 3.4852\n",
      "Epoch [   20/   20] | d_loss: 0.0729 | g_loss: 3.6848\n",
      "Epoch [   20/   20] | d_loss: 0.2645 | g_loss: 4.7941\n",
      "Epoch [   20/   20] | d_loss: 0.2505 | g_loss: 3.4737\n",
      "Epoch [   20/   20] | d_loss: 0.0424 | g_loss: 3.5121\n",
      "Epoch [   20/   20] | d_loss: 0.2750 | g_loss: 4.2043\n",
      "Epoch [   20/   20] | d_loss: 0.0920 | g_loss: 4.7810\n",
      "Epoch [   20/   20] | d_loss: 0.0335 | g_loss: 3.6082\n",
      "Epoch [   20/   20] | d_loss: 0.1534 | g_loss: 4.4133\n",
      "Epoch [   20/   20] | d_loss: 0.3552 | g_loss: 1.8233\n",
      "Epoch [   20/   20] | d_loss: 0.2475 | g_loss: 3.8644\n",
      "Epoch [   20/   20] | d_loss: 0.4575 | g_loss: 3.7688\n",
      "Epoch [   20/   20] | d_loss: 0.2236 | g_loss: 2.5018\n",
      "Epoch [   20/   20] | d_loss: 0.2349 | g_loss: 3.7509\n",
      "Epoch [   20/   20] | d_loss: 0.2791 | g_loss: 2.7559\n",
      "Epoch [   20/   20] | d_loss: 0.4328 | g_loss: 7.0739\n",
      "Epoch [   20/   20] | d_loss: 0.3123 | g_loss: 2.1347\n",
      "Epoch [   20/   20] | d_loss: 0.1873 | g_loss: 3.1571\n",
      "Epoch [   20/   20] | d_loss: 0.1549 | g_loss: 4.1612\n",
      "Epoch [   20/   20] | d_loss: 0.1650 | g_loss: 4.9926\n",
      "Epoch [   20/   20] | d_loss: 0.1850 | g_loss: 2.8773\n",
      "Epoch [   20/   20] | d_loss: 0.5824 | g_loss: 3.7504\n",
      "Epoch [   20/   20] | d_loss: 0.3320 | g_loss: 3.2909\n",
      "Epoch [   20/   20] | d_loss: 0.3778 | g_loss: 2.3677\n",
      "Epoch [   20/   20] | d_loss: 0.0822 | g_loss: 4.7431\n",
      "Epoch [   20/   20] | d_loss: 0.1043 | g_loss: 2.7812\n",
      "Epoch [   20/   20] | d_loss: 0.1029 | g_loss: 4.7488\n",
      "Epoch [   20/   20] | d_loss: 0.2452 | g_loss: 6.1446\n",
      "Epoch [   20/   20] | d_loss: 0.4337 | g_loss: 3.1830\n",
      "Epoch [   20/   20] | d_loss: 0.4776 | g_loss: 3.5760\n",
      "Epoch [   20/   20] | d_loss: 0.0986 | g_loss: 3.6100\n",
      "Epoch [   20/   20] | d_loss: 0.3387 | g_loss: 2.2061\n",
      "Epoch [   20/   20] | d_loss: 0.5293 | g_loss: 5.3874\n",
      "Epoch [   20/   20] | d_loss: 0.2684 | g_loss: 3.1504\n",
      "Epoch [   20/   20] | d_loss: 0.1234 | g_loss: 2.6872\n",
      "Epoch [   20/   20] | d_loss: 1.1567 | g_loss: 3.9713\n",
      "Epoch [   20/   20] | d_loss: 0.2640 | g_loss: 4.7056\n",
      "Epoch [   20/   20] | d_loss: 0.1640 | g_loss: 3.1511\n",
      "Epoch [   20/   20] | d_loss: 0.1787 | g_loss: 4.0280\n",
      "Epoch [   20/   20] | d_loss: 0.3137 | g_loss: 1.2719\n",
      "Epoch [   20/   20] | d_loss: 0.2430 | g_loss: 3.1714\n",
      "Epoch [   20/   20] | d_loss: 0.0371 | g_loss: 4.6429\n",
      "Epoch [   20/   20] | d_loss: 0.1417 | g_loss: 4.0651\n",
      "Epoch [   20/   20] | d_loss: 0.1668 | g_loss: 4.6798\n",
      "Epoch [   20/   20] | d_loss: 0.7471 | g_loss: 7.2178\n",
      "Epoch [   20/   20] | d_loss: 0.1097 | g_loss: 3.7236\n",
      "Epoch [   20/   20] | d_loss: 0.2440 | g_loss: 2.4877\n",
      "Epoch [   20/   20] | d_loss: 0.1245 | g_loss: 6.2474\n",
      "Epoch [   20/   20] | d_loss: 0.1097 | g_loss: 3.9484\n",
      "Epoch [   20/   20] | d_loss: 0.3025 | g_loss: 2.7205\n",
      "Epoch [   20/   20] | d_loss: 0.2288 | g_loss: 2.1863\n",
      "Epoch [   20/   20] | d_loss: 0.4928 | g_loss: 3.3285\n",
      "Epoch [   20/   20] | d_loss: 0.1453 | g_loss: 2.9679\n",
      "Epoch [   20/   20] | d_loss: 0.2959 | g_loss: 3.5091\n",
      "Epoch [   20/   20] | d_loss: 0.0549 | g_loss: 4.3946\n",
      "Epoch [   20/   20] | d_loss: 0.2631 | g_loss: 4.5699\n",
      "Epoch [   20/   20] | d_loss: 0.5417 | g_loss: 2.9341\n",
      "Epoch [   20/   20] | d_loss: 0.1672 | g_loss: 6.2068\n",
      "Epoch [   20/   20] | d_loss: 0.2507 | g_loss: 5.1171\n",
      "Epoch [   20/   20] | d_loss: 0.0474 | g_loss: 4.1707\n",
      "Epoch [   20/   20] | d_loss: 0.2797 | g_loss: 3.9881\n",
      "Epoch [   20/   20] | d_loss: 0.2848 | g_loss: 3.5116\n",
      "Epoch [   20/   20] | d_loss: 0.2399 | g_loss: 3.4947\n",
      "Epoch [   20/   20] | d_loss: 0.1270 | g_loss: 6.0808\n",
      "Epoch [   20/   20] | d_loss: 0.2817 | g_loss: 5.2550\n",
      "Epoch [   20/   20] | d_loss: 0.0478 | g_loss: 4.7783\n",
      "Epoch [   20/   20] | d_loss: 0.4252 | g_loss: 5.0222\n",
      "Epoch [   20/   20] | d_loss: 0.0747 | g_loss: 4.5580\n",
      "Epoch [   20/   20] | d_loss: 1.2044 | g_loss: 3.0120\n",
      "Epoch [   20/   20] | d_loss: 0.4331 | g_loss: 2.2022\n",
      "Epoch [   20/   20] | d_loss: 0.3590 | g_loss: 4.3832\n",
      "Epoch [   20/   20] | d_loss: 0.4347 | g_loss: 3.5932\n",
      "Epoch [   20/   20] | d_loss: 1.5043 | g_loss: 2.3318\n",
      "Epoch [   20/   20] | d_loss: 0.2251 | g_loss: 2.8542\n",
      "Epoch [   20/   20] | d_loss: 0.5982 | g_loss: 4.2939\n",
      "Epoch [   20/   20] | d_loss: 0.0851 | g_loss: 4.6154\n",
      "Epoch [   20/   20] | d_loss: 0.2500 | g_loss: 4.5671\n",
      "Epoch [   20/   20] | d_loss: 0.0584 | g_loss: 3.8822\n",
      "Epoch [   20/   20] | d_loss: 0.6952 | g_loss: 4.3122\n",
      "Epoch [   20/   20] | d_loss: 0.2074 | g_loss: 4.9202\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": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f23ae782cc0>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f23ae5c6ba8>"
      ]
     },
     "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": 27,
   "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": 28,
   "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": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f23ae75ef60>"
      ]
     },
     "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:** The model is generating faces of complexion similar to the training set (i.e., more white faces and not diverse). The image resolution is poor because of which I cannot see if there are minute things such as moles or birth marks or dimples in the faces. There is not much variety in the hair type in the generated images. I think that if the model is trained for more number of epochs and with a much diverse training set, the outputs generated will be improved."
   ]
  },
  {
   "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
}