demo_CrossDatasetOpenSet_training.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "OpenGAN: Open-Set Recognition via Open Data Generation\n",
    "================\n",
    "**Supplemental Material for ICCV2021 Submission**\n",
    "\n",
    "\n",
    "In this notebook, we demonstrate how we train the GAN-fea model on the TinyImageNet train-set, providing the closed-set images."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "import packages\n",
    "------------------\n",
    "\n",
    "Some packages are installed automatically through Anaconda. PyTorch should be also installed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.7.4 (default, Aug 13 2019, 20:35:49) \n",
      "[GCC 7.3.0]\n",
      "1.4.0+cu92\n",
      "Random Seed:  999\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<torch._C.Generator at 0x7f28dc7c32f0>"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from __future__ import print_function, division\n",
    "import os, random, time, copy\n",
    "from skimage import io, transform\n",
    "import numpy as np\n",
    "import libmr\n",
    "import pandas as pd\n",
    "import os.path as path\n",
    "import scipy.io as sio\n",
    "from scipy import misc\n",
    "from scipy import ndimage, signal\n",
    "import scipy\n",
    "import pickle\n",
    "import sys\n",
    "import math\n",
    "import matplotlib.pyplot as plt\n",
    "import PIL.Image\n",
    "from io import BytesIO\n",
    "from skimage import data, img_as_float\n",
    "from skimage.measure import compare_ssim as ssim\n",
    "from skimage.measure import compare_psnr as psnr\n",
    "\n",
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torch.optim import lr_scheduler \n",
    "import torch.nn.functional as F\n",
    "from torch.autograd import Variable\n",
    "import torchvision\n",
    "from torchvision import datasets, models, transforms\n",
    "import torchvision.utils as vutils\n",
    "\n",
    "from utils.eval_funcs import *\n",
    "from utils.dataset_tinyimagenet import *\n",
    "\n",
    "import warnings # ignore warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "print(sys.version)\n",
    "print(torch.__version__)\n",
    "\n",
    "\n",
    "manualSeed = 999\n",
    "print(\"Random Seed: \", manualSeed)\n",
    "random.seed(manualSeed)\n",
    "torch.manual_seed(manualSeed)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Setup config parameters\n",
    " -----------------\n",
    " \n",
    " There are several things to setup, like which GPU to use, where to read images and save files, etc. Please read and understand this. By default, you should be able to run this script smoothly by changing nothing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./exp/step002_GANfea_v1_Res18pt\n"
     ]
    }
   ],
   "source": [
    "# set the random seed\n",
    "torch.manual_seed(0)\n",
    "\n",
    "################## set attributes for this project/experiment ##################\n",
    "# config result folder\n",
    "exp_dir = './exp' # experiment directory, used for reading the init model\n",
    "\n",
    "modelFlag = 'Res18sc'\n",
    "\n",
    "project_name = 'step002_GANfea_v1_' + modelFlag   # we save all the checkpoints in this directory\n",
    "\n",
    "# set device, which gpu to use.\n",
    "device ='cpu'\n",
    "if torch.cuda.is_available(): \n",
    "    device='cuda:1'\n",
    "\n",
    "\n",
    "total_epoch_num = 100 # total number of epoch in training\n",
    "batch_size = 128    \n",
    "\n",
    "newsize = (64, 64)\n",
    "\n",
    "\n",
    "path_to_feats = './feats' # the path to cached off-the-shelf features\n",
    "pklName = path.join(path_to_feats, modelFlag.lower()+'.pkl')\n",
    "\n",
    "\n",
    "\n",
    "# For GAN-fea, we set the hyper-parameters as below.\n",
    "\n",
    "# Number of channels in the training images. For color images this is 3\n",
    "nc = 512\n",
    "\n",
    "# Size of z latent vector (i.e. size of generator input)\n",
    "nz = 100\n",
    "\n",
    "# Size of feature maps in generator\n",
    "ngf = 64\n",
    "\n",
    "# Size of feature maps in discriminator\n",
    "ndf = 64\n",
    "\n",
    "# Beta1 hyperparam for Adam optimizers\n",
    "beta1 = 0.5\n",
    "\n",
    "# Number of GPUs available. Use 0 for CPU mode.\n",
    "ngpu = 1\n",
    "\n",
    "\n",
    "\n",
    "# TinyImageNet has 200 classes.\n",
    "nClassTotal = 200\n",
    "nClassCloseset = nClassTotal\n",
    "\n",
    "#project_name += '_K{}run{}'.format(nClassCloseset, runIdx)\n",
    "if not os.path.exists(exp_dir): os.makedirs(exp_dir)\n",
    "\n",
    "lr = 0.0001 # learning rate\n",
    "\n",
    "num_epochs = total_epoch_num\n",
    "torch.cuda.device_count()\n",
    "torch.cuda.empty_cache()\n",
    "\n",
    "save_dir = os.path.join(exp_dir, project_name)\n",
    "print(save_dir)    \n",
    "if not os.path.exists(save_dir): os.makedirs(save_dir)\n",
    "\n",
    "log_filename = os.path.join(save_dir, 'train.log')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define model architecture\n",
    "---------\n",
    "\n",
    "Here is the definition of the model architectures for generator and discriminator. Both are multi-layer perceptron (MLP) networks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def weights_init(m):\n",
    "    classname = m.__class__.__name__\n",
    "    if classname.find('Conv') != -1:\n",
    "        nn.init.normal_(m.weight.data, 0.0, 0.02)\n",
    "    elif classname.find('BatchNorm') != -1:\n",
    "        nn.init.normal_(m.weight.data, 1.0, 0.02)\n",
    "        nn.init.constant_(m.bias.data, 0)     \n",
    "        \n",
    "\n",
    "class Generator(nn.Module):\n",
    "    def __init__(self, ngpu=1, nz=100, ngf=64, nc=512):\n",
    "        super(Generator, self).__init__()\n",
    "        self.ngpu = ngpu\n",
    "        self.nz = nz\n",
    "        self.ngf = ngf\n",
    "        self.nc = nc\n",
    "        \n",
    "        self.main = nn.Sequential(\n",
    "            # input is Z, going into a convolution\n",
    "            # Conv2d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True, padding_mode='zeros')\n",
    "            nn.Conv2d( self.nz, self.ngf * 8, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ngf * 8),\n",
    "            nn.ReLU(True),\n",
    "            # state size. (self.ngf*8) x 4 x 4\n",
    "            nn.Conv2d(self.ngf * 8, self.ngf * 4, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ngf * 4),\n",
    "            nn.ReLU(True),\n",
    "            # state size. (self.ngf*4) x 8 x 8\n",
    "            nn.Conv2d( self.ngf * 4, self.ngf * 2, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ngf * 2),\n",
    "            nn.ReLU(True),\n",
    "            # state size. (self.ngf*2) x 16 x 16\n",
    "            nn.Conv2d( self.ngf * 2, self.ngf*4, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ngf*4),\n",
    "            nn.ReLU(True),\n",
    "            # state size. (self.ngf) x 32 x 32\n",
    "            nn.Conv2d( self.ngf*4, self.nc, 1, 1, 0, bias=True),\n",
    "            #nn.Tanh()\n",
    "            # state size. (self.nc) x 64 x 64\n",
    "        )\n",
    "\n",
    "    def forward(self, input):\n",
    "        return self.main(input)\n",
    "\n",
    "    \n",
    "class Discriminator(nn.Module):\n",
    "    def __init__(self, ngpu=1, nc=512, ndf=64):\n",
    "        super(Discriminator, self).__init__()\n",
    "        self.ngpu = ngpu\n",
    "        self.nc = nc\n",
    "        self.ndf = ndf\n",
    "        self.main = nn.Sequential(\n",
    "            nn.Conv2d(self.nc, self.ndf*8, 1, 1, 0, bias=False),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Conv2d(self.ndf*8, self.ndf*4, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ndf*4),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Conv2d(self.ndf*4, self.ndf*2, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ndf*2),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Conv2d(self.ndf*2, self.ndf, 1, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(self.ndf),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            nn.Conv2d(self.ndf, 1, 1, 1, 0, bias=False),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "\n",
    "    def forward(self, input):\n",
    "        return self.main(input)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## initialize the networks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda:1\n"
     ]
    }
   ],
   "source": [
    "netG = Generator(ngpu=ngpu, nz=nz, ngf=ngf, nc=nc).to(device)\n",
    "netD = Discriminator(ngpu=ngpu, nc=nc, ndf=ndf).to(device)\n",
    "\n",
    "# Handle multi-gpu if desired\n",
    "if ('cuda' in device) and (ngpu > 1):\n",
    "    netD = nn.DataParallel(netD, list(range(ngpu)))\n",
    "\n",
    "    # Apply the weights_init function to randomly initialize all weights\n",
    "#  to mean=0, stdev=0.2.\n",
    "netD.apply(weights_init)\n",
    "\n",
    "\n",
    "if ('cuda' in device) and (ngpu > 1):\n",
    "    netG = nn.DataParallel(netG, list(range(ngpu)))\n",
    "netG.apply(weights_init)\n",
    "\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Sanity Check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([128, 100, 1, 1]) torch.Size([128, 512, 1, 1]) torch.Size([128, 1, 1, 1])\n"
     ]
    }
   ],
   "source": [
    "noise = torch.randn(batch_size, nz, 1, 1, device=device)\n",
    "# Generate fake image batch with G\n",
    "fake = netG(noise)\n",
    "predLabel = netD(fake)\n",
    "\n",
    "print(noise.shape, fake.shape, predLabel.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## setup dataset of TinyImageNet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['WholeFeatVec', 'ClassVectors'])\n",
      "torch.Size([100000, 512, 1, 1])\n"
     ]
    }
   ],
   "source": [
    "with open(pklName, \"rb\") as fn:\n",
    "    wholeData = pickle.load(fn)\n",
    "print(wholeData.keys())\n",
    "\n",
    "WholeFeatVec = torch.cat(wholeData['WholeFeatVec'],0)\n",
    "del wholeData\n",
    "\n",
    "WholeFeatVec.unsqueeze_(-1).unsqueeze_(-1)\n",
    "print(WholeFeatVec.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## a helper function to read cached off-the-shelf features per closed images"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "class FeatDataset(Dataset):\n",
    "    def __init__(self, data):\n",
    "        self.data = data\n",
    "        self.current_set_len = data.shape[0]        \n",
    "        \n",
    "    def __len__(self):        \n",
    "        return self.current_set_len\n",
    "    \n",
    "    def __getitem__(self, idx):\n",
    "        curdata = self.data[idx]        \n",
    "        return curdata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "trainset_closeset = FeatDataset(data=WholeFeatVec)\n",
    "dataloader = DataLoader(trainset_closeset, batch_size=batch_size, shuffle=True, num_workers=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([128, 512, 1, 1])\n"
     ]
    }
   ],
   "source": [
    "data_sampler = iter(dataloader)\n",
    "feaList = next(data_sampler)\n",
    "print(feaList.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "setup training\n",
    "-----------"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize BCELoss function\n",
    "criterion = nn.BCELoss()\n",
    "\n",
    "# Create batch of latent vectors that we will use to visualize\n",
    "#  the progression of the generator\n",
    "fixed_noise = torch.randn(64, nz, 1, 1, device=device)\n",
    "\n",
    "# Establish convention for real and fake labels during training\n",
    "real_label = 1\n",
    "fake_label = 0\n",
    "\n",
    "# Setup Adam optimizers for both G and D\n",
    "optimizerD = optim.Adam(netD.parameters(), lr=lr/1.5, betas=(beta1, 0.999))\n",
    "optimizerG = optim.Adam(netG.parameters(), lr=lr, betas=(beta1, 0.999))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting Training Loop...\n",
      "[0/100][0/782]\tLoss_D: 1.3904\tLoss_G: 0.6754\tD(x): 0.5113\tD(G(z)): 0.5118 / 0.5095\n",
      "[0/100][200/782]\tLoss_D: 1.3841\tLoss_G: 0.6978\tD(x): 0.5003\tD(G(z)): 0.4989 / 0.4978\n",
      "[0/100][400/782]\tLoss_D: 1.3646\tLoss_G: 0.7121\tD(x): 0.5040\tD(G(z)): 0.4927 / 0.4908\n",
      "[0/100][600/782]\tLoss_D: 1.3590\tLoss_G: 0.7106\tD(x): 0.5104\tD(G(z)): 0.4959 / 0.4916\n",
      "[1/100][0/782]\tLoss_D: 1.3572\tLoss_G: 0.7193\tD(x): 0.5085\tD(G(z)): 0.4927 / 0.4876\n",
      "[1/100][200/782]\tLoss_D: 1.3525\tLoss_G: 0.7310\tD(x): 0.5071\tD(G(z)): 0.4888 / 0.4820\n",
      "[1/100][400/782]\tLoss_D: 1.3414\tLoss_G: 0.7508\tD(x): 0.5082\tD(G(z)): 0.4831 / 0.4730\n",
      "[1/100][600/782]\tLoss_D: 1.3239\tLoss_G: 0.7887\tD(x): 0.5070\tD(G(z)): 0.4691 / 0.4572\n",
      "[2/100][0/782]\tLoss_D: 1.2931\tLoss_G: 0.8090\tD(x): 0.5126\tD(G(z)): 0.4590 / 0.4483\n",
      "[2/100][200/782]\tLoss_D: 1.3013\tLoss_G: 0.8057\tD(x): 0.5174\tD(G(z)): 0.4687 / 0.4486\n",
      "[2/100][400/782]\tLoss_D: 1.2147\tLoss_G: 0.8561\tD(x): 0.5507\tD(G(z)): 0.4503 / 0.4304\n",
      "[2/100][600/782]\tLoss_D: 1.2575\tLoss_G: 0.8736\tD(x): 0.5205\tD(G(z)): 0.4429 / 0.4227\n",
      "[3/100][0/782]\tLoss_D: 1.3315\tLoss_G: 0.7914\tD(x): 0.5346\tD(G(z)): 0.4954 / 0.4582\n",
      "[3/100][200/782]\tLoss_D: 1.1178\tLoss_G: 0.9569\tD(x): 0.5663\tD(G(z)): 0.4125 / 0.3916\n",
      "[3/100][400/782]\tLoss_D: 0.9573\tLoss_G: 0.9924\tD(x): 0.6429\tD(G(z)): 0.3933 / 0.3795\n",
      "[3/100][600/782]\tLoss_D: 0.8703\tLoss_G: 0.9968\tD(x): 0.6921\tD(G(z)): 0.3874 / 0.3763\n",
      "[4/100][0/782]\tLoss_D: 0.9672\tLoss_G: 1.1836\tD(x): 0.6022\tD(G(z)): 0.3536 / 0.3242\n",
      "[4/100][200/782]\tLoss_D: 0.7990\tLoss_G: 1.2265\tD(x): 0.6662\tD(G(z)): 0.3111 / 0.3040\n",
      "[4/100][400/782]\tLoss_D: 0.7093\tLoss_G: 1.3085\tD(x): 0.7206\tD(G(z)): 0.3024 / 0.2874\n",
      "[4/100][600/782]\tLoss_D: 0.6811\tLoss_G: 1.2815\tD(x): 0.7289\tD(G(z)): 0.2964 / 0.2857\n",
      "[5/100][0/782]\tLoss_D: 0.5554\tLoss_G: 1.4250\tD(x): 0.7798\tD(G(z)): 0.2587 / 0.2493\n",
      "[5/100][200/782]\tLoss_D: 0.6702\tLoss_G: 1.5426\tD(x): 0.6875\tD(G(z)): 0.2322 / 0.2232\n",
      "[5/100][400/782]\tLoss_D: 0.4905\tLoss_G: 1.6099\tD(x): 0.7917\tD(G(z)): 0.2207 / 0.2090\n",
      "[5/100][600/782]\tLoss_D: 0.4391\tLoss_G: 1.6976\tD(x): 0.8141\tD(G(z)): 0.2039 / 0.1926\n",
      "[6/100][0/782]\tLoss_D: 0.4218\tLoss_G: 1.7800\tD(x): 0.8072\tD(G(z)): 0.1834 / 0.1750\n",
      "[6/100][200/782]\tLoss_D: 0.3564\tLoss_G: 1.9529\tD(x): 0.8355\tD(G(z)): 0.1594 / 0.1490\n",
      "[6/100][400/782]\tLoss_D: 0.3058\tLoss_G: 2.0635\tD(x): 0.8573\tD(G(z)): 0.1385 / 0.1330\n",
      "[6/100][600/782]\tLoss_D: 0.3928\tLoss_G: 2.0055\tD(x): 0.8461\tD(G(z)): 0.1945 / 0.1566\n",
      "[7/100][0/782]\tLoss_D: 0.3034\tLoss_G: 2.0319\tD(x): 0.8657\tD(G(z)): 0.1444 / 0.1383\n",
      "[7/100][200/782]\tLoss_D: 0.3425\tLoss_G: 2.0925\tD(x): 0.8326\tD(G(z)): 0.1393 / 0.1319\n",
      "[7/100][400/782]\tLoss_D: 0.2814\tLoss_G: 2.1810\tD(x): 0.8702\tD(G(z)): 0.1299 / 0.1208\n",
      "[7/100][600/782]\tLoss_D: 0.2758\tLoss_G: 2.2406\tD(x): 0.8702\tD(G(z)): 0.1251 / 0.1148\n",
      "[8/100][0/782]\tLoss_D: 0.2619\tLoss_G: 2.2583\tD(x): 0.8829\tD(G(z)): 0.1240 / 0.1112\n",
      "[8/100][200/782]\tLoss_D: 0.2365\tLoss_G: 2.3658\tD(x): 0.8984\tD(G(z)): 0.1174 / 0.1046\n",
      "[8/100][400/782]\tLoss_D: 0.2175\tLoss_G: 2.8759\tD(x): 0.8829\tD(G(z)): 0.0868 / 0.0652\n",
      "[8/100][600/782]\tLoss_D: 0.1606\tLoss_G: 2.9307\tD(x): 0.9173\tD(G(z)): 0.0707 / 0.0596\n",
      "[9/100][0/782]\tLoss_D: 0.1487\tLoss_G: 2.8618\tD(x): 0.9301\tD(G(z)): 0.0706 / 0.0624\n",
      "[9/100][200/782]\tLoss_D: 0.2059\tLoss_G: 2.6264\tD(x): 0.9199\tD(G(z)): 0.1114 / 0.0812\n",
      "[9/100][400/782]\tLoss_D: 0.1625\tLoss_G: 2.7200\tD(x): 0.9269\tD(G(z)): 0.0813 / 0.0731\n",
      "[9/100][600/782]\tLoss_D: 0.1584\tLoss_G: 2.6587\tD(x): 0.9280\tD(G(z)): 0.0791 / 0.0747\n",
      "[10/100][0/782]\tLoss_D: 0.1630\tLoss_G: 2.6617\tD(x): 0.9339\tD(G(z)): 0.0881 / 0.0798\n",
      "[10/100][200/782]\tLoss_D: 0.1448\tLoss_G: 2.7724\tD(x): 0.9368\tD(G(z)): 0.0754 / 0.0681\n",
      "[10/100][400/782]\tLoss_D: 0.1493\tLoss_G: 2.9715\tD(x): 0.9355\tD(G(z)): 0.0773 / 0.0637\n",
      "[10/100][600/782]\tLoss_D: 0.1960\tLoss_G: 3.3951\tD(x): 0.9050\tD(G(z)): 0.0873 / 0.0467\n",
      "[11/100][0/782]\tLoss_D: 0.0865\tLoss_G: 3.6107\tD(x): 0.9534\tD(G(z)): 0.0377 / 0.0359\n",
      "[11/100][200/782]\tLoss_D: 0.0991\tLoss_G: 3.3803\tD(x): 0.9460\tD(G(z)): 0.0422 / 0.0384\n",
      "[11/100][400/782]\tLoss_D: 0.0916\tLoss_G: 3.6341\tD(x): 0.9413\tD(G(z)): 0.0302 / 0.0286\n",
      "[11/100][600/782]\tLoss_D: 0.0698\tLoss_G: 3.6555\tD(x): 0.9679\tD(G(z)): 0.0361 / 0.0308\n",
      "[12/100][0/782]\tLoss_D: 0.1055\tLoss_G: 3.5694\tD(x): 0.9534\tD(G(z)): 0.0543 / 0.0455\n",
      "[12/100][200/782]\tLoss_D: 0.0653\tLoss_G: 3.5562\tD(x): 0.9750\tD(G(z)): 0.0390 / 0.0332\n",
      "[12/100][400/782]\tLoss_D: 0.0583\tLoss_G: 3.7421\tD(x): 0.9746\tD(G(z)): 0.0319 / 0.0280\n",
      "[12/100][600/782]\tLoss_D: 0.0563\tLoss_G: 3.6488\tD(x): 0.9743\tD(G(z)): 0.0297 / 0.0283\n",
      "[13/100][0/782]\tLoss_D: 0.0557\tLoss_G: 3.6732\tD(x): 0.9767\tD(G(z)): 0.0314 / 0.0276\n",
      "[13/100][200/782]\tLoss_D: 0.0589\tLoss_G: 3.7285\tD(x): 0.9802\tD(G(z)): 0.0375 / 0.0285\n",
      "[13/100][400/782]\tLoss_D: 0.0473\tLoss_G: 3.9641\tD(x): 0.9803\tD(G(z)): 0.0269 / 0.0224\n",
      "[13/100][600/782]\tLoss_D: 0.0433\tLoss_G: 3.8549\tD(x): 0.9849\tD(G(z)): 0.0276 / 0.0242\n",
      "[14/100][0/782]\tLoss_D: 0.0620\tLoss_G: 3.9298\tD(x): 0.9753\tD(G(z)): 0.0273 / 0.0245\n",
      "[14/100][200/782]\tLoss_D: 0.0337\tLoss_G: 4.4437\tD(x): 0.9853\tD(G(z)): 0.0179 / 0.0162\n",
      "[14/100][400/782]\tLoss_D: 0.0392\tLoss_G: 4.0136\tD(x): 0.9859\tD(G(z)): 0.0246 / 0.0202\n",
      "[14/100][600/782]\tLoss_D: 0.0292\tLoss_G: 4.2641\tD(x): 0.9867\tD(G(z)): 0.0157 / 0.0151\n",
      "[15/100][0/782]\tLoss_D: 0.0396\tLoss_G: 4.1230\tD(x): 0.9815\tD(G(z)): 0.0206 / 0.0188\n",
      "[15/100][200/782]\tLoss_D: 0.0433\tLoss_G: 4.0878\tD(x): 0.9875\tD(G(z)): 0.0300 / 0.0233\n",
      "[15/100][400/782]\tLoss_D: 0.0338\tLoss_G: 4.0411\tD(x): 0.9875\tD(G(z)): 0.0210 / 0.0190\n",
      "[15/100][600/782]\tLoss_D: 0.0345\tLoss_G: 4.1141\tD(x): 0.9904\tD(G(z)): 0.0245 / 0.0226\n",
      "[16/100][0/782]\tLoss_D: 0.0759\tLoss_G: 4.6480\tD(x): 0.9731\tD(G(z)): 0.0108 / 0.0108\n",
      "[16/100][200/782]\tLoss_D: 0.0306\tLoss_G: 4.3304\tD(x): 0.9898\tD(G(z)): 0.0201 / 0.0152\n",
      "[16/100][400/782]\tLoss_D: 0.0229\tLoss_G: 4.3554\tD(x): 0.9912\tD(G(z)): 0.0140 / 0.0145\n",
      "[16/100][600/782]\tLoss_D: 0.0208\tLoss_G: 4.6815\tD(x): 0.9914\tD(G(z)): 0.0120 / 0.0109\n",
      "[17/100][0/782]\tLoss_D: 0.0544\tLoss_G: 3.3216\tD(x): 0.9858\tD(G(z)): 0.0373 / 0.0535\n",
      "[17/100][200/782]\tLoss_D: 0.0274\tLoss_G: 4.2352\tD(x): 0.9916\tD(G(z)): 0.0188 / 0.0163\n",
      "[17/100][400/782]\tLoss_D: 0.0188\tLoss_G: 4.7128\tD(x): 0.9919\tD(G(z)): 0.0106 / 0.0100\n",
      "[17/100][600/782]\tLoss_D: 0.0318\tLoss_G: 3.9441\tD(x): 0.9911\tD(G(z)): 0.0225 / 0.0292\n",
      "[18/100][0/782]\tLoss_D: 0.0868\tLoss_G: 4.4143\tD(x): 0.9725\tD(G(z)): 0.0143 / 0.0133\n",
      "[18/100][200/782]\tLoss_D: 0.0230\tLoss_G: 4.5645\tD(x): 0.9931\tD(G(z)): 0.0159 / 0.0125\n",
      "[18/100][400/782]\tLoss_D: 0.0192\tLoss_G: 4.4251\tD(x): 0.9942\tD(G(z)): 0.0133 / 0.0130\n",
      "[18/100][600/782]\tLoss_D: 0.0190\tLoss_G: 4.6114\tD(x): 0.9938\tD(G(z)): 0.0127 / 0.0122\n",
      "[19/100][0/782]\tLoss_D: 0.0204\tLoss_G: 4.7220\tD(x): 0.9910\tD(G(z)): 0.0113 / 0.0109\n",
      "[19/100][200/782]\tLoss_D: 0.0166\tLoss_G: 4.7593\tD(x): 0.9936\tD(G(z)): 0.0101 / 0.0094\n",
      "[19/100][400/782]\tLoss_D: 0.0197\tLoss_G: 4.6643\tD(x): 0.9930\tD(G(z)): 0.0125 / 0.0108\n",
      "[19/100][600/782]\tLoss_D: 0.0374\tLoss_G: 4.4364\tD(x): 0.9829\tD(G(z)): 0.0195 / 0.0160\n",
      "[20/100][0/782]\tLoss_D: 0.0359\tLoss_G: 4.1808\tD(x): 0.9840\tD(G(z)): 0.0195 / 0.0173\n",
      "[20/100][200/782]\tLoss_D: 0.0460\tLoss_G: 4.1985\tD(x): 0.9761\tD(G(z)): 0.0214 / 0.0177\n",
      "[20/100][400/782]\tLoss_D: 0.0449\tLoss_G: 4.3162\tD(x): 0.9792\tD(G(z)): 0.0231 / 0.0162\n",
      "[20/100][600/782]\tLoss_D: 0.2668\tLoss_G: 3.9403\tD(x): 0.9005\tD(G(z)): 0.1264 / 0.0291\n",
      "[21/100][0/782]\tLoss_D: 0.2924\tLoss_G: 3.7324\tD(x): 0.8799\tD(G(z)): 0.1224 / 0.0530\n",
      "[21/100][200/782]\tLoss_D: 0.5420\tLoss_G: 3.4312\tD(x): 0.7180\tD(G(z)): 0.1095 / 0.0787\n",
      "[21/100][400/782]\tLoss_D: 0.4584\tLoss_G: 3.1301\tD(x): 0.8107\tD(G(z)): 0.1338 / 0.0875\n",
      "[21/100][600/782]\tLoss_D: 0.3228\tLoss_G: 3.0243\tD(x): 0.9452\tD(G(z)): 0.2164 / 0.0744\n",
      "[22/100][0/782]\tLoss_D: 0.2594\tLoss_G: 4.3158\tD(x): 0.8373\tD(G(z)): 0.0552 / 0.0281\n",
      "[22/100][200/782]\tLoss_D: 0.3857\tLoss_G: 3.9615\tD(x): 0.7507\tD(G(z)): 0.0468 / 0.0379\n",
      "[22/100][400/782]\tLoss_D: 0.5498\tLoss_G: 2.5622\tD(x): 0.8132\tD(G(z)): 0.2093 / 0.1056\n",
      "[22/100][600/782]\tLoss_D: 1.9112\tLoss_G: 2.3628\tD(x): 0.3300\tD(G(z)): 0.1175 / 0.1178\n",
      "[23/100][0/782]\tLoss_D: 1.2555\tLoss_G: 2.1183\tD(x): 0.6040\tD(G(z)): 0.3329 / 0.1759\n",
      "[23/100][200/782]\tLoss_D: 0.4627\tLoss_G: 2.8147\tD(x): 0.8142\tD(G(z)): 0.1582 / 0.1047\n",
      "[23/100][400/782]\tLoss_D: 0.3618\tLoss_G: 2.7796\tD(x): 0.9001\tD(G(z)): 0.1897 / 0.0927\n",
      "[23/100][600/782]\tLoss_D: 0.4410\tLoss_G: 2.9853\tD(x): 0.8317\tD(G(z)): 0.1451 / 0.0750\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[24/100][0/782]\tLoss_D: 0.9603\tLoss_G: 1.6438\tD(x): 0.8191\tD(G(z)): 0.4721 / 0.2392\n",
      "[24/100][200/782]\tLoss_D: 0.5893\tLoss_G: 2.2162\tD(x): 0.7623\tD(G(z)): 0.2139 / 0.1451\n",
      "[24/100][400/782]\tLoss_D: 0.6262\tLoss_G: 1.8236\tD(x): 0.8004\tD(G(z)): 0.2883 / 0.2010\n",
      "[24/100][600/782]\tLoss_D: 0.4712\tLoss_G: 2.4267\tD(x): 0.8240\tD(G(z)): 0.2108 / 0.1072\n",
      "[25/100][0/782]\tLoss_D: 0.8062\tLoss_G: 1.8400\tD(x): 0.8761\tD(G(z)): 0.4055 / 0.2258\n",
      "[25/100][200/782]\tLoss_D: 0.2982\tLoss_G: 2.4565\tD(x): 0.8734\tD(G(z)): 0.1401 / 0.1124\n",
      "[25/100][400/782]\tLoss_D: 0.6117\tLoss_G: 2.1834\tD(x): 0.7291\tD(G(z)): 0.2074 / 0.1488\n",
      "[25/100][600/782]\tLoss_D: 0.4739\tLoss_G: 3.0330\tD(x): 0.7751\tD(G(z)): 0.1671 / 0.0902\n",
      "[26/100][0/782]\tLoss_D: 0.9526\tLoss_G: 1.4353\tD(x): 0.8050\tD(G(z)): 0.4629 / 0.2639\n",
      "[26/100][200/782]\tLoss_D: 0.6708\tLoss_G: 2.2559\tD(x): 0.7325\tD(G(z)): 0.2665 / 0.1286\n",
      "[26/100][400/782]\tLoss_D: 0.8308\tLoss_G: 2.4378\tD(x): 0.6378\tD(G(z)): 0.2483 / 0.1257\n",
      "[26/100][600/782]\tLoss_D: 0.6356\tLoss_G: 2.6632\tD(x): 0.7557\tD(G(z)): 0.2302 / 0.1069\n",
      "[27/100][0/782]\tLoss_D: 0.3972\tLoss_G: 2.5720\tD(x): 0.8445\tD(G(z)): 0.1768 / 0.1094\n",
      "[27/100][200/782]\tLoss_D: 0.5179\tLoss_G: 3.1351\tD(x): 0.7604\tD(G(z)): 0.1554 / 0.0760\n",
      "[27/100][400/782]\tLoss_D: 1.1515\tLoss_G: 2.0239\tD(x): 0.4906\tD(G(z)): 0.1780 / 0.1630\n",
      "[27/100][600/782]\tLoss_D: 0.7603\tLoss_G: 2.0446\tD(x): 0.6798\tD(G(z)): 0.2154 / 0.1804\n",
      "[28/100][0/782]\tLoss_D: 0.6148\tLoss_G: 3.0979\tD(x): 0.7490\tD(G(z)): 0.1520 / 0.0851\n",
      "[28/100][200/782]\tLoss_D: 0.5762\tLoss_G: 2.3047\tD(x): 0.7309\tD(G(z)): 0.1831 / 0.1424\n",
      "[28/100][400/782]\tLoss_D: 0.4324\tLoss_G: 2.3854\tD(x): 0.8372\tD(G(z)): 0.1882 / 0.1330\n",
      "[28/100][600/782]\tLoss_D: 0.4060\tLoss_G: 2.6960\tD(x): 0.7894\tD(G(z)): 0.1388 / 0.0871\n",
      "[29/100][0/782]\tLoss_D: 0.9907\tLoss_G: 2.0583\tD(x): 0.6268\tD(G(z)): 0.3012 / 0.1725\n",
      "[29/100][200/782]\tLoss_D: 0.4021\tLoss_G: 2.8136\tD(x): 0.8357\tD(G(z)): 0.1722 / 0.0923\n",
      "[29/100][400/782]\tLoss_D: 0.4970\tLoss_G: 2.2470\tD(x): 0.8795\tD(G(z)): 0.2842 / 0.1382\n",
      "[29/100][600/782]\tLoss_D: 0.3839\tLoss_G: 2.4180\tD(x): 0.8428\tD(G(z)): 0.1542 / 0.1204\n",
      "[30/100][0/782]\tLoss_D: 0.7949\tLoss_G: 2.0987\tD(x): 0.8904\tD(G(z)): 0.4177 / 0.1742\n",
      "[30/100][200/782]\tLoss_D: 0.3913\tLoss_G: 3.2517\tD(x): 0.8261\tD(G(z)): 0.1391 / 0.0841\n",
      "[30/100][400/782]\tLoss_D: 0.3612\tLoss_G: 2.4059\tD(x): 0.8994\tD(G(z)): 0.2032 / 0.1226\n",
      "[30/100][600/782]\tLoss_D: 0.5552\tLoss_G: 2.4255\tD(x): 0.8060\tD(G(z)): 0.2603 / 0.1214\n",
      "[31/100][0/782]\tLoss_D: 0.8371\tLoss_G: 2.1091\tD(x): 0.8998\tD(G(z)): 0.4046 / 0.1738\n",
      "[31/100][200/782]\tLoss_D: 0.5639\tLoss_G: 2.7324\tD(x): 0.7969\tD(G(z)): 0.2361 / 0.0965\n",
      "[31/100][400/782]\tLoss_D: 0.4135\tLoss_G: 3.1174\tD(x): 0.7928\tD(G(z)): 0.1262 / 0.0752\n",
      "[31/100][600/782]\tLoss_D: 0.3766\tLoss_G: 2.9660\tD(x): 0.8741\tD(G(z)): 0.1946 / 0.0748\n",
      "[32/100][0/782]\tLoss_D: 1.0841\tLoss_G: 1.4882\tD(x): 0.8589\tD(G(z)): 0.5536 / 0.2699\n",
      "[32/100][200/782]\tLoss_D: 0.5596\tLoss_G: 2.7119\tD(x): 0.8512\tD(G(z)): 0.2810 / 0.0983\n",
      "[32/100][400/782]\tLoss_D: 0.2312\tLoss_G: 3.4703\tD(x): 0.9090\tD(G(z)): 0.1068 / 0.0559\n",
      "[32/100][600/782]\tLoss_D: 0.5126\tLoss_G: 3.2978\tD(x): 0.8158\tD(G(z)): 0.2244 / 0.0720\n",
      "[33/100][0/782]\tLoss_D: 0.3693\tLoss_G: 2.6876\tD(x): 0.8919\tD(G(z)): 0.1988 / 0.1001\n",
      "[33/100][200/782]\tLoss_D: 0.2912\tLoss_G: 3.2748\tD(x): 0.8605\tD(G(z)): 0.1056 / 0.0556\n",
      "[33/100][400/782]\tLoss_D: 0.3652\tLoss_G: 2.8270\tD(x): 0.8306\tD(G(z)): 0.1362 / 0.0889\n",
      "[33/100][600/782]\tLoss_D: 0.3712\tLoss_G: 2.9180\tD(x): 0.8473\tD(G(z)): 0.1608 / 0.0801\n",
      "[34/100][0/782]\tLoss_D: 0.2759\tLoss_G: 2.7850\tD(x): 0.8929\tD(G(z)): 0.1368 / 0.0976\n",
      "[34/100][200/782]\tLoss_D: 0.1831\tLoss_G: 3.2952\tD(x): 0.9029\tD(G(z)): 0.0727 / 0.0544\n",
      "[34/100][400/782]\tLoss_D: 0.2291\tLoss_G: 4.4706\tD(x): 0.8699\tD(G(z)): 0.0484 / 0.0356\n",
      "[34/100][600/782]\tLoss_D: 0.3431\tLoss_G: 2.9715\tD(x): 0.7902\tD(G(z)): 0.0786 / 0.0693\n",
      "[35/100][0/782]\tLoss_D: 0.3522\tLoss_G: 4.0497\tD(x): 0.8660\tD(G(z)): 0.1415 / 0.0319\n",
      "[35/100][200/782]\tLoss_D: 0.2614\tLoss_G: 3.4283\tD(x): 0.8965\tD(G(z)): 0.1173 / 0.0600\n",
      "[35/100][400/782]\tLoss_D: 0.3291\tLoss_G: 3.8935\tD(x): 0.7910\tD(G(z)): 0.0556 / 0.0388\n",
      "[35/100][600/782]\tLoss_D: 0.4621\tLoss_G: 3.0499\tD(x): 0.7741\tD(G(z)): 0.1356 / 0.0783\n",
      "[36/100][0/782]\tLoss_D: 0.3125\tLoss_G: 2.5174\tD(x): 0.9496\tD(G(z)): 0.1871 / 0.1107\n",
      "[36/100][200/782]\tLoss_D: 0.3350\tLoss_G: 3.5374\tD(x): 0.8565\tD(G(z)): 0.1462 / 0.0462\n",
      "[36/100][400/782]\tLoss_D: 0.2217\tLoss_G: 3.4167\tD(x): 0.9343\tD(G(z)): 0.1339 / 0.0513\n",
      "[36/100][600/782]\tLoss_D: 0.4665\tLoss_G: 4.5204\tD(x): 0.8728\tD(G(z)): 0.2418 / 0.0314\n",
      "[37/100][0/782]\tLoss_D: 0.7641\tLoss_G: 2.1503\tD(x): 0.7904\tD(G(z)): 0.3632 / 0.1448\n",
      "[37/100][200/782]\tLoss_D: 0.2560\tLoss_G: 3.6196\tD(x): 0.9495\tD(G(z)): 0.1400 / 0.0509\n",
      "[37/100][400/782]\tLoss_D: 0.2619\tLoss_G: 3.9914\tD(x): 0.8695\tD(G(z)): 0.0975 / 0.0411\n",
      "[37/100][600/782]\tLoss_D: 0.4380\tLoss_G: 3.5408\tD(x): 0.9191\tD(G(z)): 0.2646 / 0.0539\n",
      "[38/100][0/782]\tLoss_D: 0.2927\tLoss_G: 2.9977\tD(x): 0.9422\tD(G(z)): 0.1905 / 0.0642\n",
      "[38/100][200/782]\tLoss_D: 0.1567\tLoss_G: 3.4213\tD(x): 0.9242\tD(G(z)): 0.0696 / 0.0503\n",
      "[38/100][400/782]\tLoss_D: 0.1864\tLoss_G: 3.9690\tD(x): 0.9109\tD(G(z)): 0.0814 / 0.0435\n",
      "[38/100][600/782]\tLoss_D: 0.2828\tLoss_G: 3.9180\tD(x): 0.9253\tD(G(z)): 0.1679 / 0.0408\n",
      "[39/100][0/782]\tLoss_D: 0.3888\tLoss_G: 3.5439\tD(x): 0.9152\tD(G(z)): 0.2283 / 0.0561\n",
      "[39/100][200/782]\tLoss_D: 0.2460\tLoss_G: 4.3083\tD(x): 0.9118\tD(G(z)): 0.1245 / 0.0383\n",
      "[39/100][400/782]\tLoss_D: 0.1955\tLoss_G: 3.7719\tD(x): 0.9512\tD(G(z)): 0.1218 / 0.0404\n",
      "[39/100][600/782]\tLoss_D: 0.3347\tLoss_G: 3.1815\tD(x): 0.9350\tD(G(z)): 0.2096 / 0.0719\n",
      "[40/100][0/782]\tLoss_D: 0.7714\tLoss_G: 2.9818\tD(x): 0.8883\tD(G(z)): 0.4391 / 0.0905\n",
      "[40/100][200/782]\tLoss_D: 0.2945\tLoss_G: 2.9953\tD(x): 0.9158\tD(G(z)): 0.1560 / 0.0646\n",
      "[40/100][400/782]\tLoss_D: 0.2490\tLoss_G: 3.5530\tD(x): 0.8998\tD(G(z)): 0.1114 / 0.0442\n",
      "[40/100][600/782]\tLoss_D: 0.0996\tLoss_G: 3.9781\tD(x): 0.9440\tD(G(z)): 0.0395 / 0.0301\n",
      "[41/100][0/782]\tLoss_D: 1.4089\tLoss_G: 3.1874\tD(x): 0.8829\tD(G(z)): 0.6521 / 0.0919\n",
      "[41/100][200/782]\tLoss_D: 0.2752\tLoss_G: 3.6445\tD(x): 0.9428\tD(G(z)): 0.1833 / 0.0371\n",
      "[41/100][400/782]\tLoss_D: 0.4330\tLoss_G: 3.1167\tD(x): 0.8441\tD(G(z)): 0.2004 / 0.0812\n",
      "[41/100][600/782]\tLoss_D: 0.4758\tLoss_G: 4.0160\tD(x): 0.7580\tD(G(z)): 0.1194 / 0.0403\n",
      "[42/100][0/782]\tLoss_D: 0.3404\tLoss_G: 2.8508\tD(x): 0.8817\tD(G(z)): 0.1714 / 0.0963\n",
      "[42/100][200/782]\tLoss_D: 0.1924\tLoss_G: 3.6508\tD(x): 0.9200\tD(G(z)): 0.0839 / 0.0485\n",
      "[42/100][400/782]\tLoss_D: 0.9746\tLoss_G: 2.5637\tD(x): 0.5415\tD(G(z)): 0.1320 / 0.1147\n",
      "[42/100][600/782]\tLoss_D: 0.1570\tLoss_G: 3.9003\tD(x): 0.9549\tD(G(z)): 0.0957 / 0.0463\n",
      "[43/100][0/782]\tLoss_D: 0.3752\tLoss_G: 3.2147\tD(x): 0.7987\tD(G(z)): 0.1121 / 0.0632\n",
      "[43/100][200/782]\tLoss_D: 0.3947\tLoss_G: 4.0432\tD(x): 0.9418\tD(G(z)): 0.2616 / 0.0370\n",
      "[43/100][400/782]\tLoss_D: 0.1718\tLoss_G: 4.5836\tD(x): 0.9184\tD(G(z)): 0.0376 / 0.0260\n",
      "[43/100][600/782]\tLoss_D: 0.4576\tLoss_G: 3.8445\tD(x): 0.7580\tD(G(z)): 0.1013 / 0.0426\n",
      "[44/100][0/782]\tLoss_D: 0.2415\tLoss_G: 4.3352\tD(x): 0.9435\tD(G(z)): 0.1450 / 0.0363\n",
      "[44/100][200/782]\tLoss_D: 0.2374\tLoss_G: 4.7529\tD(x): 0.8373\tD(G(z)): 0.0431 / 0.0310\n",
      "[44/100][400/782]\tLoss_D: 0.3046\tLoss_G: 3.7366\tD(x): 0.9437\tD(G(z)): 0.2008 / 0.0435\n",
      "[44/100][600/782]\tLoss_D: 0.2934\tLoss_G: 3.9859\tD(x): 0.9173\tD(G(z)): 0.1733 / 0.0329\n",
      "[45/100][0/782]\tLoss_D: 0.2736\tLoss_G: 2.2445\tD(x): 0.9289\tD(G(z)): 0.1620 / 0.1433\n",
      "[45/100][200/782]\tLoss_D: 0.2150\tLoss_G: 4.8028\tD(x): 0.9187\tD(G(z)): 0.1144 / 0.0156\n",
      "[45/100][400/782]\tLoss_D: 0.2887\tLoss_G: 3.5759\tD(x): 0.9238\tD(G(z)): 0.1583 / 0.0516\n",
      "[45/100][600/782]\tLoss_D: 0.3098\tLoss_G: 3.8451\tD(x): 0.9381\tD(G(z)): 0.2022 / 0.0416\n",
      "[46/100][0/782]\tLoss_D: 0.1731\tLoss_G: 3.2382\tD(x): 0.9754\tD(G(z)): 0.1084 / 0.1134\n",
      "[46/100][200/782]\tLoss_D: 0.2956\tLoss_G: 3.5783\tD(x): 0.9270\tD(G(z)): 0.1831 / 0.0536\n",
      "[46/100][400/782]\tLoss_D: 0.4743\tLoss_G: 4.1966\tD(x): 0.8106\tD(G(z)): 0.2002 / 0.0331\n",
      "[46/100][600/782]\tLoss_D: 0.1797\tLoss_G: 3.9974\tD(x): 0.8986\tD(G(z)): 0.0623 / 0.0385\n",
      "[47/100][0/782]\tLoss_D: 0.2008\tLoss_G: 3.0827\tD(x): 0.9094\tD(G(z)): 0.0851 / 0.0757\n",
      "[47/100][200/782]\tLoss_D: 0.1145\tLoss_G: 3.5381\tD(x): 0.9644\tD(G(z)): 0.0707 / 0.0441\n",
      "[47/100][400/782]\tLoss_D: 0.1260\tLoss_G: 3.6979\tD(x): 0.9653\tD(G(z)): 0.0820 / 0.0421\n",
      "[47/100][600/782]\tLoss_D: 0.3307\tLoss_G: 3.9235\tD(x): 0.8074\tD(G(z)): 0.0513 / 0.0383\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[48/100][0/782]\tLoss_D: 0.4249\tLoss_G: 2.9172\tD(x): 0.8899\tD(G(z)): 0.1819 / 0.0993\n",
      "[48/100][200/782]\tLoss_D: 0.2601\tLoss_G: 3.4676\tD(x): 0.8722\tD(G(z)): 0.0969 / 0.0527\n",
      "[48/100][400/782]\tLoss_D: 0.1812\tLoss_G: 3.2121\tD(x): 0.9569\tD(G(z)): 0.1103 / 0.0567\n",
      "[48/100][600/782]\tLoss_D: 0.2052\tLoss_G: 3.8517\tD(x): 0.9014\tD(G(z)): 0.0856 / 0.0372\n",
      "[49/100][0/782]\tLoss_D: 0.1980\tLoss_G: 3.2296\tD(x): 0.9711\tD(G(z)): 0.1384 / 0.0643\n",
      "[49/100][200/782]\tLoss_D: 0.3694\tLoss_G: 2.9240\tD(x): 0.9186\tD(G(z)): 0.2285 / 0.0754\n",
      "[49/100][400/782]\tLoss_D: 0.6300\tLoss_G: 3.6972\tD(x): 0.6730\tD(G(z)): 0.0637 / 0.0711\n",
      "[49/100][600/782]\tLoss_D: 0.2325\tLoss_G: 3.4851\tD(x): 0.8577\tD(G(z)): 0.0616 / 0.0513\n",
      "[50/100][0/782]\tLoss_D: 0.2415\tLoss_G: 2.9846\tD(x): 0.9711\tD(G(z)): 0.1770 / 0.0673\n",
      "[50/100][200/782]\tLoss_D: 0.1269\tLoss_G: 3.5486\tD(x): 0.9625\tD(G(z)): 0.0802 / 0.0446\n",
      "[50/100][400/782]\tLoss_D: 0.1320\tLoss_G: 4.8424\tD(x): 0.9093\tD(G(z)): 0.0306 / 0.0188\n",
      "[50/100][600/782]\tLoss_D: 0.1261\tLoss_G: 3.0995\tD(x): 0.9697\tD(G(z)): 0.0860 / 0.0640\n",
      "[51/100][0/782]\tLoss_D: 0.0860\tLoss_G: 4.2659\tD(x): 0.9591\tD(G(z)): 0.0342 / 0.0287\n",
      "[51/100][200/782]\tLoss_D: 0.2244\tLoss_G: 4.7129\tD(x): 0.8744\tD(G(z)): 0.0418 / 0.0304\n",
      "[51/100][400/782]\tLoss_D: 0.4130\tLoss_G: 3.5444\tD(x): 0.7677\tD(G(z)): 0.0946 / 0.0545\n",
      "[51/100][600/782]\tLoss_D: 0.1393\tLoss_G: 4.8886\tD(x): 0.9447\tD(G(z)): 0.0719 / 0.0205\n",
      "[52/100][0/782]\tLoss_D: 0.2302\tLoss_G: 3.3644\tD(x): 0.9083\tD(G(z)): 0.1083 / 0.0674\n",
      "[52/100][200/782]\tLoss_D: 0.1244\tLoss_G: 3.8341\tD(x): 0.9522\tD(G(z)): 0.0607 / 0.0421\n",
      "[52/100][400/782]\tLoss_D: 0.3606\tLoss_G: 3.7727\tD(x): 0.8629\tD(G(z)): 0.1595 / 0.0433\n",
      "[52/100][600/782]\tLoss_D: 0.1399\tLoss_G: 6.3772\tD(x): 0.8888\tD(G(z)): 0.0149 / 0.0128\n",
      "[53/100][0/782]\tLoss_D: 0.0984\tLoss_G: 3.7684\tD(x): 0.9535\tD(G(z)): 0.0465 / 0.0457\n",
      "[53/100][200/782]\tLoss_D: 0.1347\tLoss_G: 4.1827\tD(x): 0.9221\tD(G(z)): 0.0478 / 0.0276\n",
      "[53/100][400/782]\tLoss_D: 0.8297\tLoss_G: 5.1047\tD(x): 0.9380\tD(G(z)): 0.4533 / 0.0195\n",
      "[53/100][600/782]\tLoss_D: 0.3544\tLoss_G: 3.7096\tD(x): 0.8024\tD(G(z)): 0.0977 / 0.0489\n",
      "[54/100][0/782]\tLoss_D: 0.1889\tLoss_G: 3.4049\tD(x): 0.9409\tD(G(z)): 0.0940 / 0.0635\n",
      "[54/100][200/782]\tLoss_D: 0.1253\tLoss_G: 4.1157\tD(x): 0.9415\tD(G(z)): 0.0589 / 0.0322\n",
      "[54/100][400/782]\tLoss_D: 0.3612\tLoss_G: 4.0000\tD(x): 0.9075\tD(G(z)): 0.1792 / 0.0410\n",
      "[54/100][600/782]\tLoss_D: 0.2272\tLoss_G: 4.1371\tD(x): 0.8837\tD(G(z)): 0.0832 / 0.0286\n",
      "[55/100][0/782]\tLoss_D: 0.2273\tLoss_G: 3.7288\tD(x): 0.9477\tD(G(z)): 0.1476 / 0.0460\n",
      "[55/100][200/782]\tLoss_D: 0.1961\tLoss_G: 2.8901\tD(x): 0.9722\tD(G(z)): 0.1386 / 0.0803\n",
      "[55/100][400/782]\tLoss_D: 0.1158\tLoss_G: 3.5761\tD(x): 0.9674\tD(G(z)): 0.0752 / 0.0429\n",
      "[55/100][600/782]\tLoss_D: 0.1521\tLoss_G: 3.7258\tD(x): 0.9297\tD(G(z)): 0.0680 / 0.0371\n",
      "[56/100][0/782]\tLoss_D: 0.1493\tLoss_G: 3.6182\tD(x): 0.9677\tD(G(z)): 0.0661 / 0.0480\n",
      "[56/100][200/782]\tLoss_D: 0.2678\tLoss_G: 3.9008\tD(x): 0.8439\tD(G(z)): 0.0721 / 0.0320\n",
      "[56/100][400/782]\tLoss_D: 0.0802\tLoss_G: 3.6277\tD(x): 0.9691\tD(G(z)): 0.0466 / 0.0389\n",
      "[56/100][600/782]\tLoss_D: 0.0571\tLoss_G: 3.7999\tD(x): 0.9867\tD(G(z)): 0.0420 / 0.0317\n",
      "[57/100][0/782]\tLoss_D: 0.0836\tLoss_G: 3.7332\tD(x): 0.9724\tD(G(z)): 0.0508 / 0.0340\n",
      "[57/100][200/782]\tLoss_D: 0.0453\tLoss_G: 6.0170\tD(x): 0.9762\tD(G(z)): 0.0205 / 0.0182\n",
      "[57/100][400/782]\tLoss_D: 0.0618\tLoss_G: 4.2110\tD(x): 0.9855\tD(G(z)): 0.0440 / 0.0274\n",
      "[57/100][600/782]\tLoss_D: 0.0465\tLoss_G: 4.3943\tD(x): 0.9869\tD(G(z)): 0.0319 / 0.0202\n",
      "[58/100][0/782]\tLoss_D: 0.0674\tLoss_G: 4.1710\tD(x): 0.9794\tD(G(z)): 0.0438 / 0.0251\n",
      "[58/100][200/782]\tLoss_D: 0.0299\tLoss_G: 5.8383\tD(x): 0.9927\tD(G(z)): 0.0221 / 0.0189\n",
      "[58/100][400/782]\tLoss_D: 0.0363\tLoss_G: 4.3585\tD(x): 0.9900\tD(G(z)): 0.0253 / 0.0194\n",
      "[58/100][600/782]\tLoss_D: 0.1233\tLoss_G: 5.8862\tD(x): 0.9664\tD(G(z)): 0.0659 / 0.0141\n",
      "[59/100][0/782]\tLoss_D: 0.2051\tLoss_G: 4.2588\tD(x): 0.9796\tD(G(z)): 0.1381 / 0.0335\n",
      "[59/100][200/782]\tLoss_D: 0.1392\tLoss_G: 5.0233\tD(x): 0.9794\tD(G(z)): 0.1038 / 0.0169\n",
      "[59/100][400/782]\tLoss_D: 0.0851\tLoss_G: 4.1948\tD(x): 0.9761\tD(G(z)): 0.0236 / 0.0194\n",
      "[59/100][600/782]\tLoss_D: 0.0347\tLoss_G: 4.2305\tD(x): 0.9932\tD(G(z)): 0.0271 / 0.0202\n",
      "[60/100][0/782]\tLoss_D: 0.8730\tLoss_G: 5.7693\tD(x): 0.9364\tD(G(z)): 0.2538 / 0.0067\n",
      "[60/100][200/782]\tLoss_D: 0.0725\tLoss_G: 4.8825\tD(x): 0.9807\tD(G(z)): 0.0505 / 0.0150\n",
      "[60/100][400/782]\tLoss_D: 0.2499\tLoss_G: 5.6862\tD(x): 0.9562\tD(G(z)): 0.1657 / 0.0081\n",
      "[60/100][600/782]\tLoss_D: 0.0427\tLoss_G: 4.4448\tD(x): 0.9812\tD(G(z)): 0.0232 / 0.0182\n",
      "[61/100][0/782]\tLoss_D: 0.1596\tLoss_G: 4.8679\tD(x): 0.9526\tD(G(z)): 0.0328 / 0.0237\n",
      "[61/100][200/782]\tLoss_D: 0.5164\tLoss_G: 4.4632\tD(x): 0.7285\tD(G(z)): 0.1254 / 0.0170\n",
      "[61/100][400/782]\tLoss_D: 0.1913\tLoss_G: 4.7150\tD(x): 0.9325\tD(G(z)): 0.1042 / 0.0264\n",
      "[61/100][600/782]\tLoss_D: 0.3857\tLoss_G: 6.1636\tD(x): 0.7136\tD(G(z)): 0.0058 / 0.0082\n",
      "[62/100][0/782]\tLoss_D: 0.4356\tLoss_G: 4.5409\tD(x): 0.9359\tD(G(z)): 0.2461 / 0.0339\n",
      "[62/100][200/782]\tLoss_D: 1.3176\tLoss_G: 5.9987\tD(x): 0.4644\tD(G(z)): 0.0261 / 0.0160\n",
      "[62/100][400/782]\tLoss_D: 0.4350\tLoss_G: 4.1741\tD(x): 0.7686\tD(G(z)): 0.0966 / 0.0353\n",
      "[62/100][600/782]\tLoss_D: 0.2221\tLoss_G: 5.7660\tD(x): 0.8666\tD(G(z)): 0.0301 / 0.0106\n",
      "[63/100][0/782]\tLoss_D: 0.3647\tLoss_G: 5.4440\tD(x): 0.9145\tD(G(z)): 0.2172 / 0.0143\n",
      "[63/100][200/782]\tLoss_D: 0.2735\tLoss_G: 3.8857\tD(x): 0.8772\tD(G(z)): 0.1149 / 0.0399\n",
      "[63/100][400/782]\tLoss_D: 0.1733\tLoss_G: 4.6218\tD(x): 0.9075\tD(G(z)): 0.0651 / 0.0202\n",
      "[63/100][600/782]\tLoss_D: 0.0698\tLoss_G: 4.2397\tD(x): 0.9774\tD(G(z)): 0.0449 / 0.0299\n",
      "[64/100][0/782]\tLoss_D: 0.5704\tLoss_G: 4.6592\tD(x): 0.9680\tD(G(z)): 0.2895 / 0.0260\n",
      "[64/100][200/782]\tLoss_D: 0.1930\tLoss_G: 4.8750\tD(x): 0.9373\tD(G(z)): 0.1078 / 0.0135\n",
      "[64/100][400/782]\tLoss_D: 0.5224\tLoss_G: 4.9770\tD(x): 0.7881\tD(G(z)): 0.1852 / 0.0368\n",
      "[64/100][600/782]\tLoss_D: 0.0855\tLoss_G: 5.0465\tD(x): 0.9598\tD(G(z)): 0.0410 / 0.0241\n",
      "[65/100][0/782]\tLoss_D: 0.1591\tLoss_G: 3.1808\tD(x): 0.9857\tD(G(z)): 0.1209 / 0.0949\n",
      "[65/100][200/782]\tLoss_D: 0.3324\tLoss_G: 2.9090\tD(x): 0.8422\tD(G(z)): 0.1310 / 0.0720\n",
      "[65/100][400/782]\tLoss_D: 0.1917\tLoss_G: 4.7457\tD(x): 0.8683\tD(G(z)): 0.0387 / 0.0217\n",
      "[65/100][600/782]\tLoss_D: 0.1888\tLoss_G: 3.9230\tD(x): 0.9367\tD(G(z)): 0.1094 / 0.0345\n",
      "[66/100][0/782]\tLoss_D: 0.7120\tLoss_G: 1.7690\tD(x): 0.9892\tD(G(z)): 0.3353 / 0.2222\n",
      "[66/100][200/782]\tLoss_D: 0.1620\tLoss_G: 6.3682\tD(x): 0.9535\tD(G(z)): 0.0996 / 0.0065\n",
      "[66/100][400/782]\tLoss_D: 2.1613\tLoss_G: 2.8819\tD(x): 0.3099\tD(G(z)): 0.1351 / 0.0833\n",
      "[66/100][600/782]\tLoss_D: 0.1789\tLoss_G: 4.2898\tD(x): 0.9446\tD(G(z)): 0.0903 / 0.0271\n",
      "[67/100][0/782]\tLoss_D: 0.3352\tLoss_G: 3.4733\tD(x): 0.8485\tD(G(z)): 0.1295 / 0.0686\n",
      "[67/100][200/782]\tLoss_D: 0.2065\tLoss_G: 3.9689\tD(x): 0.9568\tD(G(z)): 0.1366 / 0.0304\n",
      "[67/100][400/782]\tLoss_D: 0.5009\tLoss_G: 2.8849\tD(x): 0.7197\tD(G(z)): 0.1070 / 0.0800\n",
      "[67/100][600/782]\tLoss_D: 0.4257\tLoss_G: 4.7281\tD(x): 0.9782\tD(G(z)): 0.2776 / 0.0224\n",
      "[68/100][0/782]\tLoss_D: 0.0435\tLoss_G: 4.4290\tD(x): 0.9905\tD(G(z)): 0.0325 / 0.0279\n",
      "[68/100][200/782]\tLoss_D: 0.0519\tLoss_G: 3.6387\tD(x): 0.9911\tD(G(z)): 0.0411 / 0.0349\n",
      "[68/100][400/782]\tLoss_D: 0.0435\tLoss_G: 3.9981\tD(x): 0.9911\tD(G(z)): 0.0330 / 0.0261\n",
      "[68/100][600/782]\tLoss_D: 0.0402\tLoss_G: 3.9714\tD(x): 0.9895\tD(G(z)): 0.0289 / 0.0249\n",
      "[69/100][0/782]\tLoss_D: 0.0318\tLoss_G: 4.6265\tD(x): 0.9918\tD(G(z)): 0.0229 / 0.0185\n",
      "[69/100][200/782]\tLoss_D: 0.0312\tLoss_G: 4.3852\tD(x): 0.9914\tD(G(z)): 0.0220 / 0.0166\n",
      "[69/100][400/782]\tLoss_D: 0.0479\tLoss_G: 3.9456\tD(x): 0.9946\tD(G(z)): 0.0411 / 0.0250\n",
      "[69/100][600/782]\tLoss_D: 0.0163\tLoss_G: 7.8196\tD(x): 0.9886\tD(G(z)): 0.0048 / 0.0045\n",
      "[70/100][0/782]\tLoss_D: 0.0413\tLoss_G: 6.2704\tD(x): 0.9694\tD(G(z)): 0.0090 / 0.0082\n",
      "[70/100][200/782]\tLoss_D: 0.0417\tLoss_G: 4.1227\tD(x): 0.9932\tD(G(z)): 0.0337 / 0.0218\n",
      "[70/100][400/782]\tLoss_D: 0.0496\tLoss_G: 4.3752\tD(x): 0.9810\tD(G(z)): 0.0293 / 0.0192\n",
      "[70/100][600/782]\tLoss_D: 0.0178\tLoss_G: 6.5541\tD(x): 0.9926\tD(G(z)): 0.0102 / 0.0089\n",
      "[71/100][0/782]\tLoss_D: 0.0282\tLoss_G: 5.0549\tD(x): 0.9927\tD(G(z)): 0.0197 / 0.0131\n",
      "[71/100][200/782]\tLoss_D: 0.0199\tLoss_G: 5.3329\tD(x): 0.9906\tD(G(z)): 0.0102 / 0.0074\n",
      "[71/100][400/782]\tLoss_D: 0.0344\tLoss_G: 5.0312\tD(x): 0.9818\tD(G(z)): 0.0156 / 0.0095\n",
      "[71/100][600/782]\tLoss_D: 0.0413\tLoss_G: 6.7467\tD(x): 0.9670\tD(G(z)): 0.0071 / 0.0061\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[72/100][0/782]\tLoss_D: 0.0670\tLoss_G: 4.8736\tD(x): 0.9572\tD(G(z)): 0.0191 / 0.0159\n",
      "[72/100][200/782]\tLoss_D: 0.0225\tLoss_G: 5.2267\tD(x): 0.9916\tD(G(z)): 0.0138 / 0.0109\n",
      "[72/100][400/782]\tLoss_D: 0.0178\tLoss_G: 6.6808\tD(x): 0.9876\tD(G(z)): 0.0051 / 0.0046\n",
      "[72/100][600/782]\tLoss_D: 0.0125\tLoss_G: 6.1366\tD(x): 0.9961\tD(G(z)): 0.0086 / 0.0062\n",
      "[73/100][0/782]\tLoss_D: 0.0150\tLoss_G: 5.2247\tD(x): 0.9959\tD(G(z)): 0.0107 / 0.0084\n",
      "[73/100][200/782]\tLoss_D: 0.0156\tLoss_G: 5.3612\tD(x): 0.9957\tD(G(z)): 0.0110 / 0.0072\n",
      "[73/100][400/782]\tLoss_D: 0.0265\tLoss_G: 7.0934\tD(x): 0.9784\tD(G(z)): 0.0043 / 0.0041\n",
      "[73/100][600/782]\tLoss_D: 0.0068\tLoss_G: 7.4287\tD(x): 0.9969\tD(G(z)): 0.0037 / 0.0035\n",
      "[74/100][0/782]\tLoss_D: 0.0059\tLoss_G: 6.4680\tD(x): 0.9984\tD(G(z)): 0.0043 / 0.0043\n",
      "[74/100][200/782]\tLoss_D: 0.0434\tLoss_G: 6.7597\tD(x): 0.9797\tD(G(z)): 0.0219 / 0.0078\n",
      "[74/100][400/782]\tLoss_D: 0.0748\tLoss_G: 5.1938\tD(x): 0.9507\tD(G(z)): 0.0223 / 0.0101\n",
      "[74/100][600/782]\tLoss_D: 0.0475\tLoss_G: 5.6732\tD(x): 0.9766\tD(G(z)): 0.0220 / 0.0119\n",
      "[75/100][0/782]\tLoss_D: 0.1539\tLoss_G: 4.3596\tD(x): 0.9692\tD(G(z)): 0.0355 / 0.0434\n",
      "[75/100][200/782]\tLoss_D: 0.1291\tLoss_G: 5.7736\tD(x): 0.9181\tD(G(z)): 0.0220 / 0.0135\n",
      "[75/100][400/782]\tLoss_D: 0.1953\tLoss_G: 4.3946\tD(x): 0.9798\tD(G(z)): 0.1486 / 0.0228\n",
      "[75/100][600/782]\tLoss_D: 0.4102\tLoss_G: 5.3251\tD(x): 0.9476\tD(G(z)): 0.2573 / 0.0117\n",
      "[76/100][0/782]\tLoss_D: 1.0912\tLoss_G: 4.5372\tD(x): 0.9533\tD(G(z)): 0.4977 / 0.0345\n",
      "[76/100][200/782]\tLoss_D: 0.0658\tLoss_G: 7.2072\tD(x): 0.9589\tD(G(z)): 0.0215 / 0.0034\n",
      "[76/100][400/782]\tLoss_D: 0.1385\tLoss_G: 5.0189\tD(x): 0.9319\tD(G(z)): 0.0600 / 0.0125\n",
      "[76/100][600/782]\tLoss_D: 0.4788\tLoss_G: 3.8280\tD(x): 0.6990\tD(G(z)): 0.0293 / 0.0515\n",
      "[77/100][0/782]\tLoss_D: 0.6585\tLoss_G: 2.5745\tD(x): 0.9666\tD(G(z)): 0.4069 / 0.1189\n",
      "[77/100][200/782]\tLoss_D: 0.2162\tLoss_G: 4.9395\tD(x): 0.9004\tD(G(z)): 0.0399 / 0.0232\n",
      "[77/100][400/782]\tLoss_D: 0.1958\tLoss_G: 5.2461\tD(x): 0.9321\tD(G(z)): 0.0875 / 0.0149\n",
      "[77/100][600/782]\tLoss_D: 0.2532\tLoss_G: 4.6603\tD(x): 0.8952\tD(G(z)): 0.0373 / 0.0185\n",
      "[78/100][0/782]\tLoss_D: 0.2322\tLoss_G: 3.7070\tD(x): 0.9482\tD(G(z)): 0.1389 / 0.0511\n",
      "[78/100][200/782]\tLoss_D: 0.1669\tLoss_G: 5.6224\tD(x): 0.9604\tD(G(z)): 0.1133 / 0.0077\n",
      "[78/100][400/782]\tLoss_D: 0.1028\tLoss_G: 6.4240\tD(x): 0.9249\tD(G(z)): 0.0095 / 0.0055\n",
      "[78/100][600/782]\tLoss_D: 0.1716\tLoss_G: 4.8358\tD(x): 0.9326\tD(G(z)): 0.0901 / 0.0208\n",
      "[79/100][0/782]\tLoss_D: 0.3924\tLoss_G: 4.2251\tD(x): 0.9798\tD(G(z)): 0.1676 / 0.0329\n",
      "[79/100][200/782]\tLoss_D: 0.9037\tLoss_G: 6.1441\tD(x): 0.9905\tD(G(z)): 0.4968 / 0.0094\n",
      "[79/100][400/782]\tLoss_D: 0.1472\tLoss_G: 4.7189\tD(x): 0.9286\tD(G(z)): 0.0654 / 0.0165\n",
      "[79/100][600/782]\tLoss_D: 0.1792\tLoss_G: 4.8653\tD(x): 0.8891\tD(G(z)): 0.0492 / 0.0153\n",
      "[80/100][0/782]\tLoss_D: 0.1804\tLoss_G: 4.2884\tD(x): 0.9465\tD(G(z)): 0.1048 / 0.0290\n",
      "[80/100][200/782]\tLoss_D: 0.2955\tLoss_G: 3.6724\tD(x): 0.8766\tD(G(z)): 0.1307 / 0.0430\n",
      "[80/100][400/782]\tLoss_D: 0.3039\tLoss_G: 5.4842\tD(x): 0.9622\tD(G(z)): 0.2091 / 0.0088\n",
      "[80/100][600/782]\tLoss_D: 0.1409\tLoss_G: 5.2089\tD(x): 0.9517\tD(G(z)): 0.0631 / 0.0169\n",
      "[81/100][0/782]\tLoss_D: 0.3047\tLoss_G: 2.9738\tD(x): 0.7928\tD(G(z)): 0.0362 / 0.0855\n",
      "[81/100][200/782]\tLoss_D: 0.2598\tLoss_G: 4.1347\tD(x): 0.9406\tD(G(z)): 0.1038 / 0.0325\n",
      "[81/100][400/782]\tLoss_D: 0.0652\tLoss_G: 4.8899\tD(x): 0.9817\tD(G(z)): 0.0442 / 0.0144\n",
      "[81/100][600/782]\tLoss_D: 0.2731\tLoss_G: 4.7479\tD(x): 0.8104\tD(G(z)): 0.0403 / 0.0206\n",
      "[82/100][0/782]\tLoss_D: 0.7157\tLoss_G: 3.9958\tD(x): 0.9287\tD(G(z)): 0.4118 / 0.0346\n",
      "[82/100][200/782]\tLoss_D: 0.2025\tLoss_G: 6.7993\tD(x): 0.8639\tD(G(z)): 0.0267 / 0.0089\n",
      "[82/100][400/782]\tLoss_D: 0.0574\tLoss_G: 4.0874\tD(x): 0.9931\tD(G(z)): 0.0469 / 0.0309\n",
      "[82/100][600/782]\tLoss_D: 0.0327\tLoss_G: 4.1801\tD(x): 0.9965\tD(G(z)): 0.0284 / 0.0231\n",
      "[83/100][0/782]\tLoss_D: 0.0282\tLoss_G: 4.8320\tD(x): 0.9909\tD(G(z)): 0.0185 / 0.0149\n",
      "[83/100][200/782]\tLoss_D: 0.0574\tLoss_G: 5.7002\tD(x): 0.9715\tD(G(z)): 0.0270 / 0.0091\n",
      "[83/100][400/782]\tLoss_D: 0.1745\tLoss_G: 6.5326\tD(x): 0.8650\tD(G(z)): 0.0173 / 0.0046\n",
      "[83/100][600/782]\tLoss_D: 0.0625\tLoss_G: 5.4034\tD(x): 0.9805\tD(G(z)): 0.0401 / 0.0194\n",
      "[84/100][0/782]\tLoss_D: 0.0432\tLoss_G: 6.8786\tD(x): 0.9645\tD(G(z)): 0.0060 / 0.0048\n",
      "[84/100][200/782]\tLoss_D: 0.2404\tLoss_G: 5.4949\tD(x): 0.8192\tD(G(z)): 0.0176 / 0.0106\n",
      "[84/100][400/782]\tLoss_D: 0.1348\tLoss_G: 5.5037\tD(x): 0.9739\tD(G(z)): 0.0352 / 0.0095\n",
      "[84/100][600/782]\tLoss_D: 0.2035\tLoss_G: 4.0912\tD(x): 0.9341\tD(G(z)): 0.0960 / 0.0350\n",
      "[85/100][0/782]\tLoss_D: 0.8562\tLoss_G: 5.7934\tD(x): 0.9608\tD(G(z)): 0.4238 / 0.0266\n",
      "[85/100][200/782]\tLoss_D: 0.3811\tLoss_G: 3.9153\tD(x): 0.9855\tD(G(z)): 0.2861 / 0.0352\n",
      "[85/100][400/782]\tLoss_D: 0.0679\tLoss_G: 4.2057\tD(x): 0.9772\tD(G(z)): 0.0423 / 0.0264\n",
      "[85/100][600/782]\tLoss_D: 0.1458\tLoss_G: 4.2587\tD(x): 0.9417\tD(G(z)): 0.0764 / 0.0332\n",
      "[86/100][0/782]\tLoss_D: 0.0363\tLoss_G: 5.1983\tD(x): 0.9818\tD(G(z)): 0.0174 / 0.0156\n",
      "[86/100][200/782]\tLoss_D: 0.0353\tLoss_G: 4.5638\tD(x): 0.9875\tD(G(z)): 0.0210 / 0.0166\n",
      "[86/100][400/782]\tLoss_D: 0.0431\tLoss_G: 4.4155\tD(x): 0.9781\tD(G(z)): 0.0192 / 0.0169\n",
      "[86/100][600/782]\tLoss_D: 0.0313\tLoss_G: 4.0340\tD(x): 0.9976\tD(G(z)): 0.0282 / 0.0221\n",
      "[87/100][0/782]\tLoss_D: 0.0239\tLoss_G: 4.8656\tD(x): 0.9966\tD(G(z)): 0.0197 / 0.0142\n",
      "[87/100][200/782]\tLoss_D: 0.0371\tLoss_G: 4.1934\tD(x): 0.9937\tD(G(z)): 0.0298 / 0.0207\n",
      "[87/100][400/782]\tLoss_D: 0.7272\tLoss_G: 4.1808\tD(x): 0.7068\tD(G(z)): 0.0243 / 0.0577\n",
      "[87/100][600/782]\tLoss_D: 0.0846\tLoss_G: 6.2698\tD(x): 0.9346\tD(G(z)): 0.0096 / 0.0084\n",
      "[88/100][0/782]\tLoss_D: 0.0325\tLoss_G: 5.8671\tD(x): 0.9822\tD(G(z)): 0.0138 / 0.0107\n",
      "[88/100][200/782]\tLoss_D: 0.0176\tLoss_G: 5.5529\tD(x): 0.9945\tD(G(z)): 0.0119 / 0.0090\n",
      "[88/100][400/782]\tLoss_D: 0.0205\tLoss_G: 5.2158\tD(x): 0.9923\tD(G(z)): 0.0126 / 0.0093\n",
      "[88/100][600/782]\tLoss_D: 0.0284\tLoss_G: 6.7308\tD(x): 0.9829\tD(G(z)): 0.0107 / 0.0103\n",
      "[89/100][0/782]\tLoss_D: 0.0172\tLoss_G: 6.3055\tD(x): 0.9954\tD(G(z)): 0.0124 / 0.0109\n",
      "[89/100][200/782]\tLoss_D: 0.0139\tLoss_G: 5.8623\tD(x): 0.9958\tD(G(z)): 0.0096 / 0.0083\n",
      "[89/100][400/782]\tLoss_D: 0.0098\tLoss_G: 6.2547\tD(x): 0.9972\tD(G(z)): 0.0070 / 0.0062\n",
      "[89/100][600/782]\tLoss_D: 0.0080\tLoss_G: 5.7884\tD(x): 0.9985\tD(G(z)): 0.0065 / 0.0055\n",
      "[90/100][0/782]\tLoss_D: 0.0162\tLoss_G: 5.2933\tD(x): 0.9976\tD(G(z)): 0.0135 / 0.0092\n",
      "[90/100][200/782]\tLoss_D: 0.3313\tLoss_G: 8.7324\tD(x): 0.7794\tD(G(z)): 0.0004 / 0.0005\n",
      "[90/100][400/782]\tLoss_D: 0.0096\tLoss_G: 7.8852\tD(x): 0.9937\tD(G(z)): 0.0033 / 0.0031\n",
      "[90/100][600/782]\tLoss_D: 0.0130\tLoss_G: 8.7941\tD(x): 0.9896\tD(G(z)): 0.0024 / 0.0023\n",
      "[91/100][0/782]\tLoss_D: 0.0199\tLoss_G: 6.5446\tD(x): 0.9934\tD(G(z)): 0.0129 / 0.0129\n",
      "[91/100][200/782]\tLoss_D: 0.0346\tLoss_G: 8.0113\tD(x): 0.9791\tD(G(z)): 0.0121 / 0.0018\n",
      "[91/100][400/782]\tLoss_D: 0.0059\tLoss_G: 6.8553\tD(x): 0.9985\tD(G(z)): 0.0044 / 0.0039\n",
      "[91/100][600/782]\tLoss_D: 0.0196\tLoss_G: 5.9479\tD(x): 0.9956\tD(G(z)): 0.0150 / 0.0104\n",
      "[92/100][0/782]\tLoss_D: 0.0149\tLoss_G: 5.5593\tD(x): 0.9968\tD(G(z)): 0.0115 / 0.0074\n",
      "[92/100][200/782]\tLoss_D: 0.0095\tLoss_G: 7.2440\tD(x): 0.9950\tD(G(z)): 0.0044 / 0.0041\n",
      "[92/100][400/782]\tLoss_D: 0.0060\tLoss_G: 6.2706\tD(x): 0.9987\tD(G(z)): 0.0047 / 0.0038\n",
      "[92/100][600/782]\tLoss_D: 0.0060\tLoss_G: 6.7633\tD(x): 0.9984\tD(G(z)): 0.0044 / 0.0036\n",
      "[93/100][0/782]\tLoss_D: 0.0063\tLoss_G: 6.5984\tD(x): 0.9975\tD(G(z)): 0.0038 / 0.0029\n",
      "[93/100][200/782]\tLoss_D: 0.0090\tLoss_G: 6.1038\tD(x): 0.9973\tD(G(z)): 0.0062 / 0.0037\n",
      "[93/100][400/782]\tLoss_D: 0.0115\tLoss_G: 6.0218\tD(x): 0.9983\tD(G(z)): 0.0097 / 0.0042\n",
      "[93/100][600/782]\tLoss_D: 0.0113\tLoss_G: 8.9963\tD(x): 0.9895\tD(G(z)): 0.0007 / 0.0006\n",
      "[94/100][0/782]\tLoss_D: 0.0201\tLoss_G: 5.3676\tD(x): 0.9961\tD(G(z)): 0.0159 / 0.0076\n",
      "[94/100][200/782]\tLoss_D: 0.0096\tLoss_G: 5.8766\tD(x): 0.9960\tD(G(z)): 0.0055 / 0.0039\n",
      "[94/100][400/782]\tLoss_D: 0.0082\tLoss_G: 5.8200\tD(x): 0.9974\tD(G(z)): 0.0055 / 0.0040\n",
      "[94/100][600/782]\tLoss_D: 0.0096\tLoss_G: 7.1687\tD(x): 0.9978\tD(G(z)): 0.0065 / 0.0019\n",
      "[95/100][0/782]\tLoss_D: 0.0331\tLoss_G: 7.3747\tD(x): 0.9770\tD(G(z)): 0.0055 / 0.0107\n",
      "[95/100][200/782]\tLoss_D: 0.0041\tLoss_G: 9.1222\tD(x): 0.9971\tD(G(z)): 0.0012 / 0.0012\n",
      "[95/100][400/782]\tLoss_D: 0.0104\tLoss_G: 7.0506\tD(x): 0.9979\tD(G(z)): 0.0082 / 0.0058\n",
      "[95/100][600/782]\tLoss_D: 0.0059\tLoss_G: 6.1707\tD(x): 0.9990\tD(G(z)): 0.0049 / 0.0036\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[96/100][0/782]\tLoss_D: 0.0089\tLoss_G: 7.4160\tD(x): 0.9947\tD(G(z)): 0.0035 / 0.0014\n",
      "[96/100][200/782]\tLoss_D: 0.0085\tLoss_G: 5.5493\tD(x): 0.9978\tD(G(z)): 0.0063 / 0.0048\n",
      "[96/100][400/782]\tLoss_D: 0.0222\tLoss_G: 6.5087\tD(x): 0.9850\tD(G(z)): 0.0069 / 0.0037\n",
      "[96/100][600/782]\tLoss_D: 0.0079\tLoss_G: 6.8320\tD(x): 0.9975\tD(G(z)): 0.0053 / 0.0035\n",
      "[97/100][0/782]\tLoss_D: 0.0128\tLoss_G: 5.7338\tD(x): 0.9976\tD(G(z)): 0.0102 / 0.0083\n",
      "[97/100][200/782]\tLoss_D: 0.0189\tLoss_G: 5.6226\tD(x): 0.9922\tD(G(z)): 0.0108 / 0.0082\n",
      "[97/100][400/782]\tLoss_D: 0.1887\tLoss_G: 6.1319\tD(x): 0.9668\tD(G(z)): 0.1051 / 0.0058\n",
      "[97/100][600/782]\tLoss_D: 0.0159\tLoss_G: 8.0507\tD(x): 0.9861\tD(G(z)): 0.0016 / 0.0018\n",
      "[98/100][0/782]\tLoss_D: 2.3761\tLoss_G: 3.7526\tD(x): 0.9826\tD(G(z)): 0.8045 / 0.0542\n",
      "[98/100][200/782]\tLoss_D: 0.1886\tLoss_G: 5.1074\tD(x): 0.9068\tD(G(z)): 0.0708 / 0.0106\n",
      "[98/100][400/782]\tLoss_D: 0.0498\tLoss_G: 6.0213\tD(x): 0.9645\tD(G(z)): 0.0114 / 0.0059\n",
      "[98/100][600/782]\tLoss_D: 0.1292\tLoss_G: 6.7725\tD(x): 0.9006\tD(G(z)): 0.0144 / 0.0044\n",
      "[99/100][0/782]\tLoss_D: 0.0391\tLoss_G: 5.4435\tD(x): 0.9899\tD(G(z)): 0.0257 / 0.0133\n",
      "[99/100][200/782]\tLoss_D: 0.2375\tLoss_G: 6.4299\tD(x): 0.9082\tD(G(z)): 0.1042 / 0.0062\n",
      "[99/100][400/782]\tLoss_D: 0.1512\tLoss_G: 4.1620\tD(x): 0.9306\tD(G(z)): 0.0682 / 0.0261\n",
      "[99/100][600/782]\tLoss_D: 0.0919\tLoss_G: 5.5948\tD(x): 0.9562\tD(G(z)): 0.0406 / 0.0106\n"
     ]
    }
   ],
   "source": [
    "# Training Loop\n",
    "\n",
    "# Lists to keep track of progress\n",
    "img_list = []\n",
    "G_losses = []\n",
    "D_losses = []\n",
    "iters = 0\n",
    "\n",
    "print(\"Starting Training Loop...\")\n",
    "# For each epoch\n",
    "for epoch in range(num_epochs):\n",
    "    # For each batch in the dataloader\n",
    "    for i, data in enumerate(dataloader, 0):\n",
    "        ############################\n",
    "        # (1) Update D network: maximize log(D(x)) + log(1 - D(G(z)))\n",
    "        ###########################\n",
    "        ## Train with all-real batch\n",
    "        netD.zero_grad()\n",
    "        # Format batch\n",
    "        real_cpu = data.to(device)\n",
    "        b_size = real_cpu.size(0)\n",
    "        label = torch.full((b_size,), real_label, device=device)\n",
    "        # Forward pass real batch through D\n",
    "        output = netD(real_cpu).view(-1)\n",
    "        # Calculate loss on all-real batch\n",
    "        errD_real = criterion(output, label)\n",
    "        # Calculate gradients for D in backward pass\n",
    "        errD_real.backward()\n",
    "        D_x = output.mean().item()\n",
    "\n",
    "        ## Train with all-fake batch\n",
    "        # Generate batch of latent vectors\n",
    "        noise = torch.randn(b_size, nz, 1, 1, device=device)\n",
    "        # Generate fake image batch with G\n",
    "        fake = netG(noise)\n",
    "        label.fill_(fake_label)\n",
    "        # Classify all fake batch with D\n",
    "        output = netD(fake.detach()).view(-1)\n",
    "        # Calculate D's loss on the all-fake batch\n",
    "        errD_fake = criterion(output, label)\n",
    "        # Calculate the gradients for this batch\n",
    "        errD_fake.backward()\n",
    "        D_G_z1 = output.mean().item()\n",
    "        # Add the gradients from the all-real and all-fake batches\n",
    "        errD = errD_real + errD_fake\n",
    "        # Update D\n",
    "        optimizerD.step()\n",
    "        \n",
    "\n",
    "        ############################\n",
    "        # (2) Update G network: maximize log(D(G(z)))\n",
    "        ###########################\n",
    "        netG.zero_grad()\n",
    "        label.fill_(real_label)  # fake labels are real for generator cost\n",
    "        # Since we just updated D, perform another forward pass of all-fake batch through D\n",
    "        output = netD(fake).view(-1)\n",
    "        # Calculate G's loss based on this output\n",
    "        errG = criterion(output, label)\n",
    "        # Calculate gradients for G\n",
    "        errG.backward()\n",
    "        D_G_z2 = output.mean().item()\n",
    "        # Update G\n",
    "        optimizerG.step()\n",
    "\n",
    "        # Output training stats\n",
    "        if i % 200 == 0:\n",
    "            print('[%d/%d][%d/%d]\\tLoss_D: %.4f\\tLoss_G: %.4f\\tD(x): %.4f\\tD(G(z)): %.4f / %.4f'\n",
    "                  % (epoch, num_epochs, i, len(dataloader),\n",
    "                     errD.item(), errG.item(), D_x, D_G_z1, D_G_z2))\n",
    "\n",
    "        # Save Losses for plotting later\n",
    "        G_losses.append(errG.item())\n",
    "        D_losses.append(errD.item())\n",
    "\n",
    "        # Check how the generator is doing by saving G's output on fixed_noise\n",
    "        if (iters % 500 == 0) or ((epoch == num_epochs-1) and (i == len(dataloader)-1)):\n",
    "            with torch.no_grad():\n",
    "                fake = netG(fixed_noise).detach().cpu()\n",
    "            img_list.append(vutils.make_grid(fake, padding=2, normalize=True))\n",
    "\n",
    "        iters += 1\n",
    "        \n",
    "        \n",
    "    cur_model_wts = copy.deepcopy(netG.state_dict())\n",
    "    path_to_save_paramOnly = os.path.join(save_dir, 'epoch-{}.GNet'.format(epoch+1))\n",
    "    torch.save(cur_model_wts, path_to_save_paramOnly)\n",
    "    \n",
    "    cur_model_wts = copy.deepcopy(netD.state_dict())\n",
    "    path_to_save_paramOnly = os.path.join(save_dir, 'epoch-{}.DNet'.format(epoch+1))\n",
    "    torch.save(cur_model_wts, path_to_save_paramOnly)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## drawing the error curves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,5))\n",
    "plt.title(\"Generator and Discriminator Loss During Training\")\n",
    "plt.plot(G_losses,label=\"G\")\n",
    "plt.plot(D_losses,label=\"D\")\n",
    "plt.xlabel(\"iterations\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.savefig('learningCurves_{}.png'.format(modelFlag), bbox_inches='tight',transparent=True)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}