fashionMNIST-GAN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Seed:  999\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<torch._C.Generator at 0x7efd350faa50>"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from __future__ import print_function\n",
    "#%matplotlib inline\n",
    "import argparse\n",
    "\n",
    "import random\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.parallel\n",
    "import torch.backends.cudnn as cudnn\n",
    "import torch.optim as optim\n",
    "import torch.utils.data as data\n",
    "import torchvision.datasets as dset\n",
    "import torchvision.transforms as transforms\n",
    "import torchvision\n",
    "import torchvision.utils as vutils\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.animation as animation\n",
    "from IPython.display import HTML\n",
    "from PIL import Image\n",
    "import wandb\n",
    "import os\n",
    "\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=1\n",
    "\n",
    "# Set random seed for reproducibility\n",
    "manualSeed = 999\n",
    "#manualSeed = random.randint(1, 10000) # use if you want new results\n",
    "print(\"Random Seed: \", manualSeed)\n",
    "random.seed(manualSeed)\n",
    "torch.manual_seed(manualSeed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "wandb.init(project=\"fashion-mnist-gan\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "workers = 2\n",
    "batch=4\n",
    "im_size = 28\n",
    "nc = 1\n",
    "nz = 100\n",
    "ngf = 64\n",
    "ndf = 64\n",
    "lr = 0.0002\n",
    "lr = 0.0002\n",
    "ngpu = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "#train_transform = \n",
    "transform = transforms.Compose([\n",
    "    transforms.Scale(64),\n",
    "    transforms.ToTensor(),\n",
    "])\n",
    "train_set = dset.FashionMNIST('./data', train=True, transform=transform, target_transform=None, download=True)\n",
    "test_set = dset.FashionMNIST('./data', train=False, transform=transform, target_transform=None, download=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset FashionMNIST\n",
      "    Number of datapoints: 60000\n",
      "    Split: train\n",
      "    Root Location: ./data\n",
      "    Transforms (if any): Compose(\n",
      "                             Scale(size=64, interpolation=PIL.Image.BILINEAR)\n",
      "                             ToTensor()\n",
      "                         )\n",
      "    Target Transforms (if any): None\n"
     ]
    }
   ],
   "source": [
    "print(train_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loader = data.DataLoader(train_set, batch_size=batch, shuffle=True,num_workers=workers)\n",
    "test_loader = data.DataLoader(test_set, batch_size=batch, shuffle=True,num_workers=workers)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "class Generator(nn.Module):\n",
    "    def __init__(self, ngpu):\n",
    "        super(Generator, self).__init__()\n",
    "        self.ngpu = ngpu\n",
    "        self.main = nn.Sequential(\n",
    "            nn.ConvTranspose2d( nz, ngf * 8, 4, 1, 0, bias=False),\n",
    "            nn.BatchNorm2d(ngf * 8),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d(ngf * 8, ngf * 4, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ngf * 4),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d( ngf * 4, ngf * 2, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ngf * 2),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d( ngf * 2, ngf, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ngf),\n",
    "            nn.ReLU(True),\n",
    "            nn.ConvTranspose2d( ngf, nc, 4, 2, 1, bias=False),\n",
    "            nn.Tanh()\n",
    "        )\n",
    "\n",
    "    def forward(self, input):\n",
    "        return self.main(input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Discriminator(nn.Module):\n",
    "    def __init__(self, ngpu):\n",
    "        super(Discriminator, self).__init__()\n",
    "        self.ngpu = ngpu\n",
    "        self.main = nn.Sequential(\n",
    "            # input is (nc) x 64 x 64\n",
    "            nn.Conv2d(nc, ndf, 4, 2, 1, bias=False),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            # state size. (ndf) x 32 x 32\n",
    "            nn.Conv2d(ndf, ndf * 2, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ndf * 2),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            # state size. (ndf*2) x 16 x 16\n",
    "            nn.Conv2d(ndf * 2, ndf * 4, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ndf * 4),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            # state size. (ndf*4) x 8 x 8\n",
    "            nn.Conv2d(ndf * 4, ndf * 8, 4, 2, 1, bias=False),\n",
    "            nn.BatchNorm2d(ndf * 8),\n",
    "            nn.LeakyReLU(0.2, inplace=True),\n",
    "            # state size. (ndf*8) x 4 x 4\n",
    "            nn.Conv2d(ndf * 8, 1, 4, 1, 0, bias=False),\n",
    "            nn.Sigmoid()\n",
    "        )\n",
    "\n",
    "    def forward(self, input):\n",
    "        return self.main(input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "netG = Generator(ngpu)\n",
    "netD = Discriminator(ngpu)\n",
    "wandb.watch(netG)\n",
    "wandb.watch(netD)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.BCELoss()\n",
    "fixed_noise = torch.randn(64, nz, 1, 1)\n",
    "real_label = 1\n",
    "fake_label = 0\n",
    "optimizerD = optim.Adam(netD.parameters(), lr=lr, betas=(0.5, 0.999))\n",
    "optimizerG = optim.Adam(netG.parameters(), lr=lr, betas=(0.5, 0.999))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Starting Training Loop...\n",
      "[0/5][0/15000]\tLoss_D: 0.5332\tLoss_G: 5.5197\tD(x): 0.8314\tD(G(z)): 0.2915 / 0.0042\n",
      "[0/5][50/15000]\tLoss_D: 0.0236\tLoss_G: 11.3495\tD(x): 0.9777\tD(G(z)): 0.0002 / 0.0000\n",
      "[0/5][100/15000]\tLoss_D: 0.0001\tLoss_G: 10.4113\tD(x): 0.9999\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][150/15000]\tLoss_D: 0.2713\tLoss_G: 13.5018\tD(x): 0.9946\tD(G(z)): 0.2304 / 0.0000\n",
      "[0/5][200/15000]\tLoss_D: 0.0482\tLoss_G: 9.7880\tD(x): 0.9996\tD(G(z)): 0.0461 / 0.0001\n",
      "[0/5][250/15000]\tLoss_D: 0.0210\tLoss_G: 8.1661\tD(x): 0.9798\tD(G(z)): 0.0004 / 0.0004\n",
      "[0/5][300/15000]\tLoss_D: 0.4193\tLoss_G: 2.8705\tD(x): 0.7869\tD(G(z)): 0.1060 / 0.0757\n",
      "[0/5][350/15000]\tLoss_D: 0.4157\tLoss_G: 3.5655\tD(x): 0.8112\tD(G(z)): 0.1572 / 0.0501\n",
      "[0/5][400/15000]\tLoss_D: 0.4722\tLoss_G: 3.2999\tD(x): 0.9008\tD(G(z)): 0.2707 / 0.0930\n",
      "[0/5][450/15000]\tLoss_D: 0.3884\tLoss_G: 3.9007\tD(x): 0.9641\tD(G(z)): 0.2821 / 0.0256\n",
      "[0/5][500/15000]\tLoss_D: 0.3016\tLoss_G: 2.9878\tD(x): 0.8475\tD(G(z)): 0.0912 / 0.0533\n",
      "[0/5][550/15000]\tLoss_D: 1.1411\tLoss_G: 1.9345\tD(x): 0.6076\tD(G(z)): 0.1477 / 0.1573\n",
      "[0/5][600/15000]\tLoss_D: 0.4057\tLoss_G: 3.1689\tD(x): 0.9794\tD(G(z)): 0.2809 / 0.0453\n",
      "[0/5][650/15000]\tLoss_D: 0.2192\tLoss_G: 3.2015\tD(x): 0.9927\tD(G(z)): 0.1691 / 0.0701\n",
      "[0/5][700/15000]\tLoss_D: 0.4859\tLoss_G: 3.2999\tD(x): 0.6945\tD(G(z)): 0.0552 / 0.0538\n",
      "[0/5][750/15000]\tLoss_D: 0.3702\tLoss_G: 3.9965\tD(x): 0.9063\tD(G(z)): 0.2037 / 0.0300\n",
      "[0/5][800/15000]\tLoss_D: 0.2180\tLoss_G: 3.9511\tD(x): 0.8932\tD(G(z)): 0.0948 / 0.0235\n",
      "[0/5][850/15000]\tLoss_D: 2.1503\tLoss_G: 3.8001\tD(x): 0.9853\tD(G(z)): 0.8488 / 0.0327\n",
      "[0/5][900/15000]\tLoss_D: 0.9272\tLoss_G: 1.6213\tD(x): 0.4910\tD(G(z)): 0.1573 / 0.2521\n",
      "[0/5][950/15000]\tLoss_D: 0.5912\tLoss_G: 4.2082\tD(x): 0.9811\tD(G(z)): 0.3148 / 0.0249\n",
      "[0/5][1000/15000]\tLoss_D: 1.9631\tLoss_G: 2.9288\tD(x): 0.9869\tD(G(z)): 0.8391 / 0.0725\n",
      "[0/5][1050/15000]\tLoss_D: 0.1990\tLoss_G: 2.8914\tD(x): 0.9936\tD(G(z)): 0.1644 / 0.0616\n",
      "[0/5][1100/15000]\tLoss_D: 0.7854\tLoss_G: 3.8867\tD(x): 0.8292\tD(G(z)): 0.3325 / 0.0324\n",
      "[0/5][1150/15000]\tLoss_D: 1.9169\tLoss_G: 6.3172\tD(x): 0.9437\tD(G(z)): 0.7723 / 0.0020\n",
      "[0/5][1200/15000]\tLoss_D: 0.1275\tLoss_G: 4.2877\tD(x): 0.9202\tD(G(z)): 0.0376 / 0.0171\n",
      "[0/5][1250/15000]\tLoss_D: 0.2740\tLoss_G: 2.7985\tD(x): 0.8445\tD(G(z)): 0.0698 / 0.0675\n",
      "[0/5][1300/15000]\tLoss_D: 0.3822\tLoss_G: 3.5938\tD(x): 0.9570\tD(G(z)): 0.2710 / 0.0293\n",
      "[0/5][1350/15000]\tLoss_D: 0.5158\tLoss_G: 1.7334\tD(x): 0.7073\tD(G(z)): 0.0867 / 0.2141\n",
      "[0/5][1400/15000]\tLoss_D: 0.2060\tLoss_G: 5.3426\tD(x): 0.8611\tD(G(z)): 0.0461 / 0.0091\n",
      "[0/5][1450/15000]\tLoss_D: 0.6206\tLoss_G: 4.1317\tD(x): 0.9871\tD(G(z)): 0.3786 / 0.0179\n",
      "[0/5][1500/15000]\tLoss_D: 0.6129\tLoss_G: 3.5619\tD(x): 0.5549\tD(G(z)): 0.0091 / 0.0471\n",
      "[0/5][1550/15000]\tLoss_D: 0.2084\tLoss_G: 3.4925\tD(x): 0.8671\tD(G(z)): 0.0426 / 0.0656\n",
      "[0/5][1600/15000]\tLoss_D: 0.1151\tLoss_G: 3.3806\tD(x): 0.9563\tD(G(z)): 0.0658 / 0.0394\n",
      "[0/5][1650/15000]\tLoss_D: 0.1763\tLoss_G: 7.9094\tD(x): 0.8467\tD(G(z)): 0.0014 / 0.0004\n",
      "[0/5][1700/15000]\tLoss_D: 0.1583\tLoss_G: 3.4869\tD(x): 0.9276\tD(G(z)): 0.0747 / 0.0370\n",
      "[0/5][1750/15000]\tLoss_D: 0.5458\tLoss_G: 5.0565\tD(x): 0.9908\tD(G(z)): 0.2733 / 0.0092\n",
      "[0/5][1800/15000]\tLoss_D: 0.0592\tLoss_G: 6.0262\tD(x): 0.9467\tD(G(z)): 0.0034 / 0.0027\n",
      "[0/5][1850/15000]\tLoss_D: 0.3873\tLoss_G: 4.8268\tD(x): 0.7028\tD(G(z)): 0.0141 / 0.0111\n",
      "[0/5][1900/15000]\tLoss_D: 0.0321\tLoss_G: 5.0617\tD(x): 0.9833\tD(G(z)): 0.0149 / 0.0082\n",
      "[0/5][1950/15000]\tLoss_D: 0.0629\tLoss_G: 6.6615\tD(x): 0.9478\tD(G(z)): 0.0086 / 0.0045\n",
      "[0/5][2000/15000]\tLoss_D: 0.6270\tLoss_G: 5.2672\tD(x): 0.5958\tD(G(z)): 0.0031 / 0.0092\n",
      "[0/5][2050/15000]\tLoss_D: 0.2622\tLoss_G: 7.6127\tD(x): 0.9642\tD(G(z)): 0.1500 / 0.0015\n",
      "[0/5][2100/15000]\tLoss_D: 0.0182\tLoss_G: 6.6508\tD(x): 0.9869\tD(G(z)): 0.0050 / 0.0020\n",
      "[0/5][2150/15000]\tLoss_D: 0.2553\tLoss_G: 4.1126\tD(x): 0.9980\tD(G(z)): 0.2028 / 0.0247\n",
      "[0/5][2200/15000]\tLoss_D: 2.1905\tLoss_G: 1.0301\tD(x): 0.2918\tD(G(z)): 0.3192 / 0.3927\n",
      "[0/5][2250/15000]\tLoss_D: 0.0707\tLoss_G: 5.9255\tD(x): 0.9477\tD(G(z)): 0.0128 / 0.0047\n",
      "[0/5][2300/15000]\tLoss_D: 0.1274\tLoss_G: 4.1542\tD(x): 0.9204\tD(G(z)): 0.0385 / 0.0197\n",
      "[0/5][2350/15000]\tLoss_D: 0.7052\tLoss_G: 5.2950\tD(x): 0.9763\tD(G(z)): 0.4550 / 0.0061\n",
      "[0/5][2400/15000]\tLoss_D: 0.0268\tLoss_G: 4.6935\tD(x): 0.9938\tD(G(z)): 0.0203 / 0.0223\n",
      "[0/5][2450/15000]\tLoss_D: 0.3329\tLoss_G: 3.4476\tD(x): 0.7476\tD(G(z)): 0.0262 / 0.0432\n",
      "[0/5][2500/15000]\tLoss_D: 0.0623\tLoss_G: 5.2521\tD(x): 0.9672\tD(G(z)): 0.0276 / 0.0085\n",
      "[0/5][2550/15000]\tLoss_D: 0.0286\tLoss_G: 6.7308\tD(x): 0.9952\tD(G(z)): 0.0231 / 0.0026\n",
      "[0/5][2600/15000]\tLoss_D: 0.1251\tLoss_G: 5.4638\tD(x): 0.9900\tD(G(z)): 0.0998 / 0.0089\n",
      "[0/5][2650/15000]\tLoss_D: 1.3938\tLoss_G: 7.2277\tD(x): 0.9974\tD(G(z)): 0.5339 / 0.0012\n",
      "[0/5][2700/15000]\tLoss_D: 0.5055\tLoss_G: 2.5631\tD(x): 0.7103\tD(G(z)): 0.1284 / 0.1089\n",
      "[0/5][2750/15000]\tLoss_D: 0.0130\tLoss_G: 6.5607\tD(x): 0.9954\tD(G(z)): 0.0084 / 0.0021\n",
      "[0/5][2800/15000]\tLoss_D: 0.0871\tLoss_G: 3.6531\tD(x): 0.9925\tD(G(z)): 0.0753 / 0.0314\n",
      "[0/5][2850/15000]\tLoss_D: 0.3005\tLoss_G: 6.0175\tD(x): 0.9497\tD(G(z)): 0.1923 / 0.0033\n",
      "[0/5][2900/15000]\tLoss_D: 0.4178\tLoss_G: 6.1694\tD(x): 0.6941\tD(G(z)): 0.0059 / 0.0132\n",
      "[0/5][2950/15000]\tLoss_D: 0.0412\tLoss_G: 4.1944\tD(x): 0.9988\tD(G(z)): 0.0379 / 0.0261\n",
      "[0/5][3000/15000]\tLoss_D: 0.0443\tLoss_G: 4.9598\tD(x): 0.9882\tD(G(z)): 0.0315 / 0.0104\n",
      "[0/5][3050/15000]\tLoss_D: 0.1126\tLoss_G: 3.8386\tD(x): 0.9996\tD(G(z)): 0.1029 / 0.0227\n",
      "[0/5][3100/15000]\tLoss_D: 0.0895\tLoss_G: 4.8167\tD(x): 0.9425\tD(G(z)): 0.0288 / 0.0174\n",
      "[0/5][3150/15000]\tLoss_D: 0.2961\tLoss_G: 3.3007\tD(x): 0.9978\tD(G(z)): 0.2252 / 0.0378\n",
      "[0/5][3200/15000]\tLoss_D: 0.0198\tLoss_G: 5.3437\tD(x): 0.9911\tD(G(z)): 0.0107 / 0.0096\n",
      "[0/5][3250/15000]\tLoss_D: 0.2049\tLoss_G: 10.8650\tD(x): 0.8367\tD(G(z)): 0.0002 / 0.0001\n",
      "[0/5][3300/15000]\tLoss_D: 0.1377\tLoss_G: 4.8059\tD(x): 0.9998\tD(G(z)): 0.1152 / 0.0116\n",
      "[0/5][3350/15000]\tLoss_D: 0.0249\tLoss_G: 4.3236\tD(x): 0.9834\tD(G(z)): 0.0079 / 0.0703\n",
      "[0/5][3400/15000]\tLoss_D: 0.0305\tLoss_G: 4.7595\tD(x): 0.9887\tD(G(z)): 0.0186 / 0.0130\n",
      "[0/5][3450/15000]\tLoss_D: 0.0494\tLoss_G: 6.6356\tD(x): 0.9545\tD(G(z)): 0.0023 / 0.0018\n",
      "[0/5][3500/15000]\tLoss_D: 0.1533\tLoss_G: 3.5764\tD(x): 0.8856\tD(G(z)): 0.0213 / 0.0367\n",
      "[0/5][3550/15000]\tLoss_D: 0.1349\tLoss_G: 3.5445\tD(x): 0.9999\tD(G(z)): 0.1094 / 0.0381\n",
      "[0/5][3600/15000]\tLoss_D: 0.2893\tLoss_G: 3.5390\tD(x): 0.8571\tD(G(z)): 0.1135 / 0.0321\n",
      "[0/5][3650/15000]\tLoss_D: 0.1403\tLoss_G: 5.1328\tD(x): 0.9758\tD(G(z)): 0.0988 / 0.0072\n",
      "[0/5][3700/15000]\tLoss_D: 0.9317\tLoss_G: 5.9702\tD(x): 0.7444\tD(G(z)): 0.3711 / 0.0034\n",
      "[0/5][3750/15000]\tLoss_D: 0.1072\tLoss_G: 3.9350\tD(x): 0.9563\tD(G(z)): 0.0593 / 0.0226\n",
      "[0/5][3800/15000]\tLoss_D: 0.1045\tLoss_G: 5.1096\tD(x): 0.9106\tD(G(z)): 0.0025 / 0.0069\n",
      "[0/5][3850/15000]\tLoss_D: 0.0762\tLoss_G: 4.3062\tD(x): 0.9869\tD(G(z)): 0.0607 / 0.0141\n",
      "[0/5][3900/15000]\tLoss_D: 0.8869\tLoss_G: 6.6733\tD(x): 1.0000\tD(G(z)): 0.4273 / 0.0017\n",
      "[0/5][3950/15000]\tLoss_D: 0.1273\tLoss_G: 3.5072\tD(x): 0.9231\tD(G(z)): 0.0383 / 0.0323\n",
      "[0/5][4000/15000]\tLoss_D: 0.8341\tLoss_G: 2.8642\tD(x): 0.8672\tD(G(z)): 0.4069 / 0.0612\n",
      "[0/5][4050/15000]\tLoss_D: 1.6585\tLoss_G: 8.4445\tD(x): 0.9891\tD(G(z)): 0.5490 / 0.0003\n",
      "[0/5][4100/15000]\tLoss_D: 0.0190\tLoss_G: 5.5890\tD(x): 0.9998\tD(G(z)): 0.0183 / 0.0103\n",
      "[0/5][4150/15000]\tLoss_D: 0.1266\tLoss_G: 5.8083\tD(x): 0.9809\tD(G(z)): 0.0898 / 0.0057\n",
      "[0/5][4200/15000]\tLoss_D: 0.0945\tLoss_G: 6.9154\tD(x): 0.9199\tD(G(z)): 0.0091 / 0.0063\n",
      "[0/5][4250/15000]\tLoss_D: 0.3167\tLoss_G: 13.2783\tD(x): 0.7542\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][4300/15000]\tLoss_D: 0.2602\tLoss_G: 4.2678\tD(x): 0.8382\tD(G(z)): 0.0127 / 0.0206\n",
      "[0/5][4350/15000]\tLoss_D: 0.2390\tLoss_G: 5.4135\tD(x): 0.7947\tD(G(z)): 0.0016 / 0.0095\n",
      "[0/5][4400/15000]\tLoss_D: 0.0054\tLoss_G: 5.5187\tD(x): 0.9994\tD(G(z)): 0.0047 / 0.0041\n",
      "[0/5][4450/15000]\tLoss_D: 0.0099\tLoss_G: 5.7084\tD(x): 0.9987\tD(G(z)): 0.0085 / 0.0034\n",
      "[0/5][4500/15000]\tLoss_D: 0.0369\tLoss_G: 4.4759\tD(x): 0.9859\tD(G(z)): 0.0224 / 0.0123\n",
      "[0/5][4550/15000]\tLoss_D: 0.3676\tLoss_G: 3.9369\tD(x): 0.7383\tD(G(z)): 0.0037 / 0.0443\n",
      "[0/5][4600/15000]\tLoss_D: 0.4159\tLoss_G: 3.0055\tD(x): 0.7082\tD(G(z)): 0.0467 / 0.1142\n",
      "[0/5][4650/15000]\tLoss_D: 0.0080\tLoss_G: 5.4016\tD(x): 0.9994\tD(G(z)): 0.0074 / 0.0055\n",
      "[0/5][4700/15000]\tLoss_D: 0.1103\tLoss_G: 4.7928\tD(x): 0.9174\tD(G(z)): 0.0197 / 0.0172\n",
      "[0/5][4750/15000]\tLoss_D: 0.0313\tLoss_G: 4.9589\tD(x): 0.9759\tD(G(z)): 0.0065 / 0.0094\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0/5][4800/15000]\tLoss_D: 0.1216\tLoss_G: 6.5412\tD(x): 0.9132\tD(G(z)): 0.0182 / 0.0092\n",
      "[0/5][4850/15000]\tLoss_D: 0.0168\tLoss_G: 4.3842\tD(x): 0.9992\tD(G(z)): 0.0158 / 0.0208\n",
      "[0/5][4900/15000]\tLoss_D: 0.8006\tLoss_G: 0.7144\tD(x): 0.5112\tD(G(z)): 0.0250 / 0.5030\n",
      "[0/5][4950/15000]\tLoss_D: 0.7907\tLoss_G: 6.7680\tD(x): 0.5912\tD(G(z)): 0.0003 / 0.0022\n",
      "[0/5][5000/15000]\tLoss_D: 0.0321\tLoss_G: 4.2775\tD(x): 0.9927\tD(G(z)): 0.0244 / 0.0163\n",
      "[0/5][5050/15000]\tLoss_D: 0.1242\tLoss_G: 4.9255\tD(x): 0.9910\tD(G(z)): 0.1003 / 0.0115\n",
      "[0/5][5100/15000]\tLoss_D: 0.0056\tLoss_G: 6.0634\tD(x): 0.9989\tD(G(z)): 0.0045 / 0.0029\n",
      "[0/5][5150/15000]\tLoss_D: 0.5753\tLoss_G: 9.6544\tD(x): 0.9993\tD(G(z)): 0.4015 / 0.0001\n",
      "[0/5][5200/15000]\tLoss_D: 0.3276\tLoss_G: 6.2327\tD(x): 0.9994\tD(G(z)): 0.2405 / 0.0022\n",
      "[0/5][5250/15000]\tLoss_D: 0.0571\tLoss_G: 4.0866\tD(x): 0.9594\tD(G(z)): 0.0150 / 0.0202\n",
      "[0/5][5300/15000]\tLoss_D: 0.5414\tLoss_G: 2.9027\tD(x): 0.6555\tD(G(z)): 0.0433 / 0.1096\n",
      "[0/5][5350/15000]\tLoss_D: 0.0261\tLoss_G: 7.7177\tD(x): 0.9935\tD(G(z)): 0.0189 / 0.0015\n",
      "[0/5][5400/15000]\tLoss_D: 0.3752\tLoss_G: 8.4078\tD(x): 0.7537\tD(G(z)): 0.0002 / 0.0002\n",
      "[0/5][5450/15000]\tLoss_D: 0.0094\tLoss_G: 8.5784\tD(x): 0.9910\tD(G(z)): 0.0003 / 0.0002\n",
      "[0/5][5500/15000]\tLoss_D: 0.3016\tLoss_G: 2.9122\tD(x): 0.9414\tD(G(z)): 0.1972 / 0.0871\n",
      "[0/5][5550/15000]\tLoss_D: 0.9398\tLoss_G: 4.9958\tD(x): 0.9998\tD(G(z)): 0.2706 / 0.0102\n",
      "[0/5][5600/15000]\tLoss_D: 0.0310\tLoss_G: 4.9769\tD(x): 0.9995\tD(G(z)): 0.0297 / 0.0128\n",
      "[0/5][5650/15000]\tLoss_D: 0.0224\tLoss_G: 5.3369\tD(x): 0.9958\tD(G(z)): 0.0179 / 0.0131\n",
      "[0/5][5700/15000]\tLoss_D: 1.3230\tLoss_G: 8.5300\tD(x): 0.8948\tD(G(z)): 0.6726 / 0.0003\n",
      "[0/5][5750/15000]\tLoss_D: 0.3404\tLoss_G: 3.3886\tD(x): 0.8017\tD(G(z)): 0.0540 / 0.0397\n",
      "[0/5][5800/15000]\tLoss_D: 0.0008\tLoss_G: 7.8004\tD(x): 1.0000\tD(G(z)): 0.0007 / 0.0006\n",
      "[0/5][5850/15000]\tLoss_D: 0.2291\tLoss_G: 6.3587\tD(x): 0.8501\tD(G(z)): 0.0038 / 0.0078\n",
      "[0/5][5900/15000]\tLoss_D: 0.1397\tLoss_G: 4.4322\tD(x): 0.9992\tD(G(z)): 0.1200 / 0.0154\n",
      "[0/5][5950/15000]\tLoss_D: 0.0039\tLoss_G: 6.4052\tD(x): 0.9997\tD(G(z)): 0.0036 / 0.0032\n",
      "[0/5][6000/15000]\tLoss_D: 0.0528\tLoss_G: 8.6670\tD(x): 0.9528\tD(G(z)): 0.0014 / 0.0006\n",
      "[0/5][6050/15000]\tLoss_D: 0.0542\tLoss_G: 4.5920\tD(x): 0.9717\tD(G(z)): 0.0249 / 0.0135\n",
      "[0/5][6100/15000]\tLoss_D: 0.0300\tLoss_G: 5.1193\tD(x): 0.9997\tD(G(z)): 0.0283 / 0.0128\n",
      "[0/5][6150/15000]\tLoss_D: 0.0773\tLoss_G: 4.5117\tD(x): 0.9413\tD(G(z)): 0.0152 / 0.0250\n",
      "[0/5][6200/15000]\tLoss_D: 1.3221\tLoss_G: 2.0150\tD(x): 0.5311\tD(G(z)): 0.0386 / 0.2473\n",
      "[0/5][6250/15000]\tLoss_D: 0.0241\tLoss_G: 5.7959\tD(x): 0.9849\tD(G(z)): 0.0087 / 0.0093\n",
      "[0/5][6300/15000]\tLoss_D: 0.3463\tLoss_G: 5.4106\tD(x): 0.7479\tD(G(z)): 0.0075 / 0.0130\n",
      "[0/5][6350/15000]\tLoss_D: 0.0044\tLoss_G: 8.4485\tD(x): 0.9989\tD(G(z)): 0.0033 / 0.0005\n",
      "[0/5][6400/15000]\tLoss_D: 0.0031\tLoss_G: 10.7048\tD(x): 0.9971\tD(G(z)): 0.0002 / 0.0000\n",
      "[0/5][6450/15000]\tLoss_D: 0.2030\tLoss_G: 10.0611\tD(x): 0.8391\tD(G(z)): 0.0000 / 0.0001\n",
      "[0/5][6500/15000]\tLoss_D: 1.3052\tLoss_G: 4.0333\tD(x): 0.3826\tD(G(z)): 0.0064 / 0.0678\n",
      "[0/5][6550/15000]\tLoss_D: 0.0551\tLoss_G: 6.5548\tD(x): 0.9551\tD(G(z)): 0.0063 / 0.0017\n",
      "[0/5][6600/15000]\tLoss_D: 0.0296\tLoss_G: 6.4841\tD(x): 0.9930\tD(G(z)): 0.0218 / 0.0052\n",
      "[0/5][6650/15000]\tLoss_D: 0.0095\tLoss_G: 7.4454\tD(x): 0.9919\tD(G(z)): 0.0014 / 0.0012\n",
      "[0/5][6700/15000]\tLoss_D: 0.2274\tLoss_G: 5.0844\tD(x): 0.8511\tD(G(z)): 0.0069 / 0.0144\n",
      "[0/5][6750/15000]\tLoss_D: 0.5717\tLoss_G: 9.2453\tD(x): 0.9903\tD(G(z)): 0.3913 / 0.0001\n",
      "[0/5][6800/15000]\tLoss_D: 1.1679\tLoss_G: 10.4162\tD(x): 0.9976\tD(G(z)): 0.4058 / 0.0000\n",
      "[0/5][6850/15000]\tLoss_D: 0.0267\tLoss_G: 9.8073\tD(x): 0.9739\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][6900/15000]\tLoss_D: 0.3133\tLoss_G: 8.8068\tD(x): 0.8095\tD(G(z)): 0.0010 / 0.0010\n",
      "[0/5][6950/15000]\tLoss_D: 0.0123\tLoss_G: 6.0452\tD(x): 0.9892\tD(G(z)): 0.0013 / 0.0038\n",
      "[0/5][7000/15000]\tLoss_D: 0.0278\tLoss_G: 11.6361\tD(x): 0.9728\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][7050/15000]\tLoss_D: 0.1267\tLoss_G: 4.5362\tD(x): 0.9037\tD(G(z)): 0.0211 / 0.0195\n",
      "[0/5][7100/15000]\tLoss_D: 0.0527\tLoss_G: 4.9276\tD(x): 1.0000\tD(G(z)): 0.0500 / 0.0091\n",
      "[0/5][7150/15000]\tLoss_D: 0.4126\tLoss_G: 6.8626\tD(x): 0.9851\tD(G(z)): 0.2925 / 0.0012\n",
      "[0/5][7200/15000]\tLoss_D: 0.2058\tLoss_G: 7.9883\tD(x): 0.8532\tD(G(z)): 0.0112 / 0.0014\n",
      "[0/5][7250/15000]\tLoss_D: 1.2684\tLoss_G: 9.7694\tD(x): 0.9867\tD(G(z)): 0.5834 / 0.0001\n",
      "[0/5][7300/15000]\tLoss_D: 0.0007\tLoss_G: 9.3894\tD(x): 0.9998\tD(G(z)): 0.0005 / 0.0002\n",
      "[0/5][7350/15000]\tLoss_D: 0.2772\tLoss_G: 4.9794\tD(x): 0.9428\tD(G(z)): 0.1681 / 0.0085\n",
      "[0/5][7400/15000]\tLoss_D: 0.0288\tLoss_G: 5.4924\tD(x): 0.9865\tD(G(z)): 0.0148 / 0.0077\n",
      "[0/5][7450/15000]\tLoss_D: 0.0819\tLoss_G: 3.4577\tD(x): 1.0000\tD(G(z)): 0.0741 / 0.0389\n",
      "[0/5][7500/15000]\tLoss_D: 0.4240\tLoss_G: 6.1162\tD(x): 1.0000\tD(G(z)): 0.3051 / 0.0023\n",
      "[0/5][7550/15000]\tLoss_D: 0.0691\tLoss_G: 4.3015\tD(x): 0.9929\tD(G(z)): 0.0589 / 0.0224\n",
      "[0/5][7600/15000]\tLoss_D: 0.0176\tLoss_G: 5.6260\tD(x): 0.9886\tD(G(z)): 0.0060 / 0.0059\n",
      "[0/5][7650/15000]\tLoss_D: 0.0241\tLoss_G: 11.7736\tD(x): 0.9769\tD(G(z)): 0.0002 / 0.0000\n",
      "[0/5][7700/15000]\tLoss_D: 0.1607\tLoss_G: 5.6870\tD(x): 0.8591\tD(G(z)): 0.0020 / 0.0049\n",
      "[0/5][7750/15000]\tLoss_D: 0.0047\tLoss_G: 4.6491\tD(x): 1.0000\tD(G(z)): 0.0047 / 0.0114\n",
      "[0/5][7800/15000]\tLoss_D: 1.5149\tLoss_G: 3.7403\tD(x): 0.3749\tD(G(z)): 0.0134 / 0.1146\n",
      "[0/5][7850/15000]\tLoss_D: 0.0462\tLoss_G: 6.6814\tD(x): 0.9580\tD(G(z)): 0.0012 / 0.0020\n",
      "[0/5][7900/15000]\tLoss_D: 0.0816\tLoss_G: 5.2105\tD(x): 0.9998\tD(G(z)): 0.0745 / 0.0098\n",
      "[0/5][7950/15000]\tLoss_D: 0.0563\tLoss_G: 5.4056\tD(x): 0.9691\tD(G(z)): 0.0242 / 0.0086\n",
      "[0/5][8000/15000]\tLoss_D: 0.0041\tLoss_G: 7.1836\tD(x): 0.9972\tD(G(z)): 0.0013 / 0.0012\n",
      "[0/5][8050/15000]\tLoss_D: 0.6585\tLoss_G: 8.6661\tD(x): 0.9999\tD(G(z)): 0.4063 / 0.0002\n",
      "[0/5][8100/15000]\tLoss_D: 0.4778\tLoss_G: 4.3646\tD(x): 0.9955\tD(G(z)): 0.3653 / 0.0140\n",
      "[0/5][8150/15000]\tLoss_D: 0.1735\tLoss_G: 6.3763\tD(x): 0.8646\tD(G(z)): 0.0020 / 0.0023\n",
      "[0/5][8200/15000]\tLoss_D: 0.4440\tLoss_G: 9.4071\tD(x): 0.9741\tD(G(z)): 0.3322 / 0.0001\n",
      "[0/5][8250/15000]\tLoss_D: 0.0429\tLoss_G: 4.5127\tD(x): 0.9999\tD(G(z)): 0.0416 / 0.0126\n",
      "[0/5][8300/15000]\tLoss_D: 0.0054\tLoss_G: 7.7915\tD(x): 0.9952\tD(G(z)): 0.0006 / 0.0005\n",
      "[0/5][8350/15000]\tLoss_D: 0.0714\tLoss_G: 6.2692\tD(x): 0.9992\tD(G(z)): 0.0657 / 0.0040\n",
      "[0/5][8400/15000]\tLoss_D: 0.0054\tLoss_G: 16.9603\tD(x): 0.9946\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][8450/15000]\tLoss_D: 0.2831\tLoss_G: 7.2420\tD(x): 0.9994\tD(G(z)): 0.2029 / 0.0015\n",
      "[0/5][8500/15000]\tLoss_D: 0.0120\tLoss_G: 12.8923\tD(x): 0.9881\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][8550/15000]\tLoss_D: 0.9944\tLoss_G: 7.8952\tD(x): 1.0000\tD(G(z)): 0.4696 / 0.0006\n",
      "[0/5][8600/15000]\tLoss_D: 0.0029\tLoss_G: 7.4687\tD(x): 0.9983\tD(G(z)): 0.0011 / 0.0007\n",
      "[0/5][8650/15000]\tLoss_D: 0.0693\tLoss_G: 6.7607\tD(x): 0.9380\tD(G(z)): 0.0028 / 0.0014\n",
      "[0/5][8700/15000]\tLoss_D: 0.0162\tLoss_G: 7.0651\tD(x): 0.9850\tD(G(z)): 0.0009 / 0.0013\n",
      "[0/5][8750/15000]\tLoss_D: 0.0163\tLoss_G: 12.6965\tD(x): 0.9839\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][8800/15000]\tLoss_D: 0.0482\tLoss_G: 9.6280\tD(x): 0.9561\tD(G(z)): 0.0006 / 0.0003\n",
      "[0/5][8850/15000]\tLoss_D: 0.0609\tLoss_G: 5.4117\tD(x): 0.9896\tD(G(z)): 0.0473 / 0.0081\n",
      "[0/5][8900/15000]\tLoss_D: 0.1782\tLoss_G: 10.9423\tD(x): 0.8622\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][8950/15000]\tLoss_D: 0.0051\tLoss_G: 6.2731\tD(x): 0.9980\tD(G(z)): 0.0031 / 0.0027\n",
      "[0/5][9000/15000]\tLoss_D: 0.0210\tLoss_G: 4.6134\tD(x): 0.9992\tD(G(z)): 0.0199 / 0.0144\n",
      "[0/5][9050/15000]\tLoss_D: 0.0684\tLoss_G: 4.4966\tD(x): 0.9999\tD(G(z)): 0.0659 / 0.0113\n",
      "[0/5][9100/15000]\tLoss_D: 0.0060\tLoss_G: 5.3081\tD(x): 0.9999\tD(G(z)): 0.0059 / 0.0051\n",
      "[0/5][9150/15000]\tLoss_D: 0.0014\tLoss_G: 7.6433\tD(x): 0.9995\tD(G(z)): 0.0009 / 0.0006\n",
      "[0/5][9200/15000]\tLoss_D: 0.0172\tLoss_G: 5.6516\tD(x): 0.9927\tD(G(z)): 0.0097 / 0.0058\n",
      "[0/5][9250/15000]\tLoss_D: 1.0499\tLoss_G: 9.7223\tD(x): 0.4121\tD(G(z)): 0.0000 / 0.0001\n",
      "[0/5][9300/15000]\tLoss_D: 0.0324\tLoss_G: 10.6180\tD(x): 0.9686\tD(G(z)): 0.0001 / 0.0000\n",
      "[0/5][9350/15000]\tLoss_D: 0.0022\tLoss_G: 9.1758\tD(x): 0.9989\tD(G(z)): 0.0010 / 0.0011\n",
      "[0/5][9400/15000]\tLoss_D: 0.0172\tLoss_G: 5.1997\tD(x): 0.9999\tD(G(z)): 0.0169 / 0.0140\n",
      "[0/5][9450/15000]\tLoss_D: 0.4066\tLoss_G: 7.8064\tD(x): 1.0000\tD(G(z)): 0.2305 / 0.0008\n",
      "[0/5][9500/15000]\tLoss_D: 0.0932\tLoss_G: 7.3026\tD(x): 0.9770\tD(G(z)): 0.0615 / 0.0050\n",
      "[0/5][9550/15000]\tLoss_D: 0.2737\tLoss_G: 4.4560\tD(x): 0.7831\tD(G(z)): 0.0133 / 0.0197\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0/5][9600/15000]\tLoss_D: 0.1365\tLoss_G: 10.0485\tD(x): 0.8798\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][9650/15000]\tLoss_D: 0.1268\tLoss_G: 4.8202\tD(x): 0.9982\tD(G(z)): 0.1019 / 0.0140\n",
      "[0/5][9700/15000]\tLoss_D: 0.0147\tLoss_G: 2.9295\tD(x): 0.9997\tD(G(z)): 0.0142 / 0.1219\n",
      "[0/5][9750/15000]\tLoss_D: 0.0018\tLoss_G: 7.7904\tD(x): 0.9988\tD(G(z)): 0.0006 / 0.0006\n",
      "[0/5][9800/15000]\tLoss_D: 0.3702\tLoss_G: 5.3048\tD(x): 0.7873\tD(G(z)): 0.0042 / 0.0080\n",
      "[0/5][9850/15000]\tLoss_D: 0.0045\tLoss_G: 7.1336\tD(x): 0.9967\tD(G(z)): 0.0011 / 0.0014\n",
      "[0/5][9900/15000]\tLoss_D: 0.4741\tLoss_G: 8.0587\tD(x): 0.6319\tD(G(z)): 0.0000 / 0.0012\n",
      "[0/5][9950/15000]\tLoss_D: 0.0080\tLoss_G: 8.8409\tD(x): 0.9925\tD(G(z)): 0.0004 / 0.0003\n",
      "[0/5][10000/15000]\tLoss_D: 0.0125\tLoss_G: 4.9969\tD(x): 0.9955\tD(G(z)): 0.0080 / 0.0071\n",
      "[0/5][10050/15000]\tLoss_D: 0.0037\tLoss_G: 11.0571\tD(x): 0.9963\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][10100/15000]\tLoss_D: 0.0034\tLoss_G: 9.6274\tD(x): 0.9970\tD(G(z)): 0.0004 / 0.0004\n",
      "[0/5][10150/15000]\tLoss_D: 0.0517\tLoss_G: 3.8176\tD(x): 0.9941\tD(G(z)): 0.0443 / 0.0251\n",
      "[0/5][10200/15000]\tLoss_D: 0.0269\tLoss_G: 4.5906\tD(x): 0.9930\tD(G(z)): 0.0194 / 0.0162\n",
      "[0/5][10250/15000]\tLoss_D: 0.0006\tLoss_G: 9.8745\tD(x): 0.9995\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][10300/15000]\tLoss_D: 0.0349\tLoss_G: 5.9802\tD(x): 0.9978\tD(G(z)): 0.0311 / 0.0081\n",
      "[0/5][10350/15000]\tLoss_D: 0.3858\tLoss_G: 3.0872\tD(x): 0.7407\tD(G(z)): 0.0018 / 0.0865\n",
      "[0/5][10400/15000]\tLoss_D: 1.5086\tLoss_G: 10.0760\tD(x): 0.9996\tD(G(z)): 0.7328 / 0.0000\n",
      "[0/5][10450/15000]\tLoss_D: 0.0719\tLoss_G: 4.7290\tD(x): 0.9998\tD(G(z)): 0.0679 / 0.0129\n",
      "[0/5][10500/15000]\tLoss_D: 0.2044\tLoss_G: 5.1591\tD(x): 1.0000\tD(G(z)): 0.1806 / 0.0063\n",
      "[0/5][10550/15000]\tLoss_D: 1.8997\tLoss_G: 7.3011\tD(x): 0.4877\tD(G(z)): 0.0000 / 0.0014\n",
      "[0/5][10600/15000]\tLoss_D: 0.0071\tLoss_G: 8.2357\tD(x): 0.9937\tD(G(z)): 0.0008 / 0.0013\n",
      "[0/5][10650/15000]\tLoss_D: 0.0240\tLoss_G: 11.5939\tD(x): 0.9767\tD(G(z)): 0.0001 / 0.0000\n",
      "[0/5][10700/15000]\tLoss_D: 0.0721\tLoss_G: 3.9046\tD(x): 0.9962\tD(G(z)): 0.0640 / 0.0675\n",
      "[0/5][10750/15000]\tLoss_D: 0.0411\tLoss_G: 6.6323\tD(x): 0.9613\tD(G(z)): 0.0014 / 0.0021\n",
      "[0/5][10800/15000]\tLoss_D: 0.0128\tLoss_G: 5.4355\tD(x): 0.9931\tD(G(z)): 0.0059 / 0.0045\n",
      "[0/5][10850/15000]\tLoss_D: 0.0113\tLoss_G: 4.2603\tD(x): 0.9954\tD(G(z)): 0.0067 / 0.0180\n",
      "[0/5][10900/15000]\tLoss_D: 0.0359\tLoss_G: 5.4292\tD(x): 0.9820\tD(G(z)): 0.0173 / 0.0106\n",
      "[0/5][10950/15000]\tLoss_D: 0.0063\tLoss_G: 7.2498\tD(x): 0.9987\tD(G(z)): 0.0049 / 0.0023\n",
      "[0/5][11000/15000]\tLoss_D: 0.0058\tLoss_G: 5.0274\tD(x): 0.9999\tD(G(z)): 0.0057 / 0.0077\n",
      "[0/5][11050/15000]\tLoss_D: 0.1500\tLoss_G: 7.0147\tD(x): 0.8707\tD(G(z)): 0.0016 / 0.0032\n",
      "[0/5][11100/15000]\tLoss_D: 0.0028\tLoss_G: 6.6762\tD(x): 0.9989\tD(G(z)): 0.0016 / 0.0016\n",
      "[0/5][11150/15000]\tLoss_D: 0.0693\tLoss_G: 8.5707\tD(x): 0.9370\tD(G(z)): 0.0002 / 0.0002\n",
      "[0/5][11200/15000]\tLoss_D: 0.0120\tLoss_G: 8.9945\tD(x): 0.9884\tD(G(z)): 0.0002 / 0.0002\n",
      "[0/5][11250/15000]\tLoss_D: 0.0205\tLoss_G: 4.5284\tD(x): 0.9964\tD(G(z)): 0.0166 / 0.0120\n",
      "[0/5][11300/15000]\tLoss_D: 0.0062\tLoss_G: 5.4820\tD(x): 0.9999\tD(G(z)): 0.0061 / 0.0045\n",
      "[0/5][11350/15000]\tLoss_D: 0.0078\tLoss_G: 7.1793\tD(x): 0.9953\tD(G(z)): 0.0031 / 0.0023\n",
      "[0/5][11400/15000]\tLoss_D: 0.6816\tLoss_G: 8.2321\tD(x): 1.0000\tD(G(z)): 0.4134 / 0.0003\n",
      "[0/5][11450/15000]\tLoss_D: 0.0120\tLoss_G: 8.3643\tD(x): 0.9892\tD(G(z)): 0.0010 / 0.0003\n",
      "[0/5][11500/15000]\tLoss_D: 1.1114\tLoss_G: 4.4679\tD(x): 0.4461\tD(G(z)): 0.0022 / 0.0196\n",
      "[0/5][11550/15000]\tLoss_D: 0.0115\tLoss_G: 8.7066\tD(x): 0.9901\tD(G(z)): 0.0014 / 0.0008\n",
      "[0/5][11600/15000]\tLoss_D: 0.0009\tLoss_G: 9.2821\tD(x): 0.9995\tD(G(z)): 0.0003 / 0.0002\n",
      "[0/5][11650/15000]\tLoss_D: 0.0522\tLoss_G: 5.3082\tD(x): 0.9979\tD(G(z)): 0.0469 / 0.0111\n",
      "[0/5][11700/15000]\tLoss_D: 0.2896\tLoss_G: 2.4873\tD(x): 0.7980\tD(G(z)): 0.0306 / 0.0878\n",
      "[0/5][11750/15000]\tLoss_D: 0.1151\tLoss_G: 5.9312\tD(x): 0.9985\tD(G(z)): 0.1028 / 0.0067\n",
      "[0/5][11800/15000]\tLoss_D: 0.1515\tLoss_G: 5.2038\tD(x): 0.8862\tD(G(z)): 0.0031 / 0.0061\n",
      "[0/5][11850/15000]\tLoss_D: 0.0089\tLoss_G: 8.6306\tD(x): 0.9921\tD(G(z)): 0.0008 / 0.0004\n",
      "[0/5][11900/15000]\tLoss_D: 0.0004\tLoss_G: 9.9772\tD(x): 0.9998\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][11950/15000]\tLoss_D: 0.3999\tLoss_G: 4.2702\tD(x): 0.7373\tD(G(z)): 0.0157 / 0.0385\n",
      "[0/5][12000/15000]\tLoss_D: 0.0387\tLoss_G: 4.9414\tD(x): 0.9695\tD(G(z)): 0.0075 / 0.0086\n",
      "[0/5][12050/15000]\tLoss_D: 0.0255\tLoss_G: 4.5733\tD(x): 0.9994\tD(G(z)): 0.0244 / 0.0115\n",
      "[0/5][12100/15000]\tLoss_D: 0.0839\tLoss_G: 4.1266\tD(x): 0.9993\tD(G(z)): 0.0789 / 0.0183\n",
      "[0/5][12150/15000]\tLoss_D: 0.1720\tLoss_G: 5.7278\tD(x): 0.9924\tD(G(z)): 0.1387 / 0.0053\n",
      "[0/5][12200/15000]\tLoss_D: 0.0005\tLoss_G: 9.9572\tD(x): 0.9996\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][12250/15000]\tLoss_D: 0.0016\tLoss_G: 7.6548\tD(x): 0.9998\tD(G(z)): 0.0015 / 0.0006\n",
      "[0/5][12300/15000]\tLoss_D: 0.0023\tLoss_G: 6.5532\tD(x): 0.9997\tD(G(z)): 0.0020 / 0.0036\n",
      "[0/5][12350/15000]\tLoss_D: 0.0002\tLoss_G: 10.1499\tD(x): 1.0000\tD(G(z)): 0.0002 / 0.0001\n",
      "[0/5][12400/15000]\tLoss_D: 0.0173\tLoss_G: 7.8956\tD(x): 0.9873\tD(G(z)): 0.0045 / 0.0016\n",
      "[0/5][12450/15000]\tLoss_D: 0.0102\tLoss_G: 6.3223\tD(x): 0.9977\tD(G(z)): 0.0078 / 0.0049\n",
      "[0/5][12500/15000]\tLoss_D: 0.0005\tLoss_G: 11.7288\tD(x): 0.9995\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][12550/15000]\tLoss_D: 0.0007\tLoss_G: 12.3131\tD(x): 0.9994\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][12600/15000]\tLoss_D: 0.0335\tLoss_G: 7.6661\tD(x): 0.9707\tD(G(z)): 0.0027 / 0.0015\n",
      "[0/5][12650/15000]\tLoss_D: 0.3027\tLoss_G: 7.6991\tD(x): 0.7649\tD(G(z)): 0.0002 / 0.0007\n",
      "[0/5][12700/15000]\tLoss_D: 0.0231\tLoss_G: 5.0882\tD(x): 1.0000\tD(G(z)): 0.0226 / 0.0133\n",
      "[0/5][12750/15000]\tLoss_D: 0.0099\tLoss_G: 6.5497\tD(x): 0.9968\tD(G(z)): 0.0067 / 0.0025\n",
      "[0/5][12800/15000]\tLoss_D: 0.1321\tLoss_G: 4.5220\tD(x): 0.9967\tD(G(z)): 0.1105 / 0.0215\n",
      "[0/5][12850/15000]\tLoss_D: 0.0146\tLoss_G: 5.0981\tD(x): 0.9956\tD(G(z)): 0.0100 / 0.0101\n",
      "[0/5][12900/15000]\tLoss_D: 0.0040\tLoss_G: 6.1814\tD(x): 0.9983\tD(G(z)): 0.0022 / 0.0024\n",
      "[0/5][12950/15000]\tLoss_D: 0.0059\tLoss_G: 6.7898\tD(x): 0.9987\tD(G(z)): 0.0046 / 0.0036\n",
      "[0/5][13000/15000]\tLoss_D: 0.0733\tLoss_G: 4.9348\tD(x): 0.9998\tD(G(z)): 0.0677 / 0.0078\n",
      "[0/5][13050/15000]\tLoss_D: 0.0007\tLoss_G: 8.2542\tD(x): 0.9995\tD(G(z)): 0.0003 / 0.0003\n",
      "[0/5][13100/15000]\tLoss_D: 0.0023\tLoss_G: 11.3084\tD(x): 0.9977\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][13150/15000]\tLoss_D: 0.0310\tLoss_G: 3.9539\tD(x): 0.9995\tD(G(z)): 0.0295 / 0.0520\n",
      "[0/5][13200/15000]\tLoss_D: 0.0040\tLoss_G: 6.1979\tD(x): 0.9997\tD(G(z)): 0.0037 / 0.0028\n",
      "[0/5][13250/15000]\tLoss_D: 0.0029\tLoss_G: 9.4357\tD(x): 0.9988\tD(G(z)): 0.0017 / 0.0006\n",
      "[0/5][13300/15000]\tLoss_D: 0.0543\tLoss_G: 17.2390\tD(x): 0.9487\tD(G(z)): 0.0000 / 0.0000\n",
      "[0/5][13350/15000]\tLoss_D: 2.4800\tLoss_G: 1.2894\tD(x): 0.0985\tD(G(z)): 0.0015 / 0.4694\n",
      "[0/5][13400/15000]\tLoss_D: 0.0237\tLoss_G: 3.4001\tD(x): 0.9992\tD(G(z)): 0.0222 / 0.0520\n",
      "[0/5][13450/15000]\tLoss_D: 0.0010\tLoss_G: 7.7012\tD(x): 0.9999\tD(G(z)): 0.0009 / 0.0011\n",
      "[0/5][13500/15000]\tLoss_D: 0.0131\tLoss_G: 5.5526\tD(x): 0.9993\tD(G(z)): 0.0123 / 0.0083\n",
      "[0/5][13550/15000]\tLoss_D: 0.0006\tLoss_G: 9.3714\tD(x): 0.9998\tD(G(z)): 0.0003 / 0.0002\n",
      "[0/5][13600/15000]\tLoss_D: 0.0537\tLoss_G: 5.2191\tD(x): 0.9999\tD(G(z)): 0.0510 / 0.0103\n",
      "[0/5][13650/15000]\tLoss_D: 0.0524\tLoss_G: 5.5320\tD(x): 0.9997\tD(G(z)): 0.0493 / 0.0169\n",
      "[0/5][13700/15000]\tLoss_D: 0.0001\tLoss_G: 11.5134\tD(x): 1.0000\tD(G(z)): 0.0001 / 0.0000\n",
      "[0/5][13750/15000]\tLoss_D: 0.0154\tLoss_G: 6.6395\tD(x): 0.9885\tD(G(z)): 0.0037 / 0.0035\n",
      "[0/5][13800/15000]\tLoss_D: 0.0512\tLoss_G: 8.4880\tD(x): 0.9508\tD(G(z)): 0.0001 / 0.0002\n",
      "[0/5][13850/15000]\tLoss_D: 0.3186\tLoss_G: 7.8557\tD(x): 0.8552\tD(G(z)): 0.1032 / 0.0013\n",
      "[0/5][13900/15000]\tLoss_D: 0.0165\tLoss_G: 5.5847\tD(x): 0.9923\tD(G(z)): 0.0087 / 0.0051\n",
      "[0/5][13950/15000]\tLoss_D: 0.0006\tLoss_G: 8.1734\tD(x): 1.0000\tD(G(z)): 0.0006 / 0.0005\n",
      "[0/5][14000/15000]\tLoss_D: 0.3957\tLoss_G: 4.5160\tD(x): 1.0000\tD(G(z)): 0.3068 / 0.0180\n",
      "[0/5][14050/15000]\tLoss_D: 0.0057\tLoss_G: 5.1254\tD(x): 0.9996\tD(G(z)): 0.0053 / 0.0088\n",
      "[0/5][14100/15000]\tLoss_D: 0.7670\tLoss_G: 10.1050\tD(x): 0.9994\tD(G(z)): 0.3848 / 0.0001\n",
      "[0/5][14150/15000]\tLoss_D: 0.1050\tLoss_G: 6.9537\tD(x): 0.9123\tD(G(z)): 0.0007 / 0.0014\n",
      "[0/5][14200/15000]\tLoss_D: 0.0056\tLoss_G: 10.2556\tD(x): 0.9946\tD(G(z)): 0.0002 / 0.0001\n",
      "[0/5][14250/15000]\tLoss_D: 0.0003\tLoss_G: 10.0510\tD(x): 0.9998\tD(G(z)): 0.0001 / 0.0001\n",
      "[0/5][14300/15000]\tLoss_D: 0.2140\tLoss_G: 5.6976\tD(x): 0.9987\tD(G(z)): 0.1810 / 0.0051\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0/5][14350/15000]\tLoss_D: 0.0714\tLoss_G: 5.5320\tD(x): 0.9980\tD(G(z)): 0.0653 / 0.0050\n",
      "[0/5][14400/15000]\tLoss_D: 0.0015\tLoss_G: 8.7854\tD(x): 0.9987\tD(G(z)): 0.0002 / 0.0003\n",
      "[0/5][14450/15000]\tLoss_D: 0.0539\tLoss_G: 4.7020\tD(x): 0.9991\tD(G(z)): 0.0508 / 0.0119\n",
      "[0/5][14500/15000]\tLoss_D: 0.0002\tLoss_G: 8.8556\tD(x): 1.0000\tD(G(z)): 0.0002 / 0.0002\n",
      "[0/5][14550/15000]\tLoss_D: 0.0120\tLoss_G: 5.4714\tD(x): 1.0000\tD(G(z)): 0.0119 / 0.0059\n",
      "[0/5][14600/15000]\tLoss_D: 0.4987\tLoss_G: 4.1404\tD(x): 0.7577\tD(G(z)): 0.0243 / 0.0195\n",
      "[0/5][14650/15000]\tLoss_D: 0.0049\tLoss_G: 8.2816\tD(x): 0.9978\tD(G(z)): 0.0027 / 0.0016\n",
      "[0/5][14700/15000]\tLoss_D: 0.0282\tLoss_G: 10.7621\tD(x): 0.9731\tD(G(z)): 0.0003 / 0.0001\n",
      "[0/5][14750/15000]\tLoss_D: 0.0868\tLoss_G: 4.7484\tD(x): 0.9251\tD(G(z)): 0.0056 / 0.0112\n",
      "[0/5][14800/15000]\tLoss_D: 0.6371\tLoss_G: 6.3339\tD(x): 0.6021\tD(G(z)): 0.0001 / 0.0027\n",
      "[0/5][14850/15000]\tLoss_D: 0.2680\tLoss_G: 6.0632\tD(x): 0.8243\tD(G(z)): 0.0028 / 0.0058\n",
      "[0/5][14900/15000]\tLoss_D: 0.3864\tLoss_G: 3.7019\tD(x): 0.7573\tD(G(z)): 0.0569 / 0.0590\n",
      "[0/5][14950/15000]\tLoss_D: 0.0568\tLoss_G: 7.5533\tD(x): 0.9484\tD(G(z)): 0.0004 / 0.0007\n",
      "[1/5][0/15000]\tLoss_D: 0.0831\tLoss_G: 8.7316\tD(x): 0.9272\tD(G(z)): 0.0004 / 0.0006\n",
      "[1/5][50/15000]\tLoss_D: 0.0040\tLoss_G: 11.0126\tD(x): 0.9963\tD(G(z)): 0.0003 / 0.0003\n",
      "[1/5][100/15000]\tLoss_D: 0.0090\tLoss_G: 3.3237\tD(x): 1.0000\tD(G(z)): 0.0089 / 0.0439\n",
      "[1/5][150/15000]\tLoss_D: 0.1714\tLoss_G: 4.8982\tD(x): 0.9985\tD(G(z)): 0.1438 / 0.0101\n",
      "[1/5][200/15000]\tLoss_D: 0.0034\tLoss_G: 6.2217\tD(x): 1.0000\tD(G(z)): 0.0033 / 0.0029\n",
      "[1/5][250/15000]\tLoss_D: 0.0018\tLoss_G: 9.7686\tD(x): 0.9998\tD(G(z)): 0.0016 / 0.0001\n",
      "[1/5][300/15000]\tLoss_D: 0.0468\tLoss_G: 5.5571\tD(x): 0.9996\tD(G(z)): 0.0435 / 0.0106\n",
      "[1/5][350/15000]\tLoss_D: 0.3914\tLoss_G: 9.8798\tD(x): 0.9979\tD(G(z)): 0.2736 / 0.0001\n",
      "[1/5][400/15000]\tLoss_D: 0.0328\tLoss_G: 6.5388\tD(x): 0.9731\tD(G(z)): 0.0046 / 0.0045\n",
      "[1/5][450/15000]\tLoss_D: 0.0004\tLoss_G: 13.6921\tD(x): 0.9996\tD(G(z)): 0.0000 / 0.0000\n",
      "[1/5][500/15000]\tLoss_D: 0.0136\tLoss_G: 7.5483\tD(x): 0.9874\tD(G(z)): 0.0007 / 0.0008\n",
      "[1/5][550/15000]\tLoss_D: 0.0063\tLoss_G: 10.9935\tD(x): 0.9938\tD(G(z)): 0.0000 / 0.0000\n",
      "[1/5][600/15000]\tLoss_D: 1.0597\tLoss_G: 6.9798\tD(x): 0.6569\tD(G(z)): 0.0000 / 0.0037\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-36-543c061c9dcd>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     41\u001b[0m         \u001b[0merrD_fake\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcriterion\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     42\u001b[0m         \u001b[0;31m# Calculate the gradients for this batch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 43\u001b[0;31m         \u001b[0merrD_fake\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbackward\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[0m\u001b[1;32m     44\u001b[0m         \u001b[0mD_G_z1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0moutput\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmean\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitem\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[1;32m     45\u001b[0m         \u001b[0;31m# Add the gradients from the all-real and all-fake batches\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/Downloads/yes/envs/pykyle36/lib/python3.8/site-packages/torch-1.4.0-py3.8-linux-x86_64.egg/torch/tensor.py\u001b[0m in \u001b[0;36mbackward\u001b[0;34m(self, gradient, retain_graph, create_graph)\u001b[0m\n\u001b[1;32m    193\u001b[0m                 \u001b[0mproducts\u001b[0m\u001b[0;34m.\u001b[0m \u001b[0mDefaults\u001b[0m \u001b[0mto\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;31m`\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    194\u001b[0m         \"\"\"\n\u001b[0;32m--> 195\u001b[0;31m         \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbackward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgradient\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    196\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    197\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mregister_hook\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhook\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~/Downloads/yes/envs/pykyle36/lib/python3.8/site-packages/torch-1.4.0-py3.8-linux-x86_64.egg/torch/autograd/__init__.py\u001b[0m in \u001b[0;36mbackward\u001b[0;34m(tensors, grad_tensors, retain_graph, create_graph, grad_variables)\u001b[0m\n\u001b[1;32m     95\u001b[0m         \u001b[0mretain_graph\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 97\u001b[0;31m     Variable._execution_engine.run_backward(\n\u001b[0m\u001b[1;32m     98\u001b[0m         \u001b[0mtensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgrad_tensors\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mretain_graph\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcreate_graph\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     99\u001b[0m         allow_unreachable=True)  # allow_unreachable flag\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "device = torch.device('cpu')\n",
    "# Lists to keep track of progress\n",
    "img_list = []\n",
    "G_losses = []\n",
    "D_losses = []\n",
    "iters = 0\n",
    "num_epochs=5\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(train_loader, 0):\n",
    "\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[0].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",
    "        # (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 % 50 == 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(train_loader),\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"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "for item in img_list:\n",
    "    im = transforms.ToPILImage()(item)\n",
    "    plt.imshow(im)"
   ]
  },
  {
   "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}