TRAIN_DCGAN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "#importing some stuff..\n",
    "import os, time, sys\n",
    "import matplotlib.pyplot as plt\n",
    "import itertools\n",
    "import pickle\n",
    "import imageio\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from torchvision import datasets, transforms\n",
    "from torch.autograd import Variable"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "==> Preparing data..\n"
     ]
    }
   ],
   "source": [
    "print('==> Preparing data..')\n",
    "\n",
    "\n",
    "img_size = 64\n",
    "#Just performing some transfomations on the image\n",
    "transform_train = transforms.Compose([\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize(mean=(0.5, 0.5, 0.5),\n",
    "                             std=(0.5, 0.5, 0.5))\n",
    "])\n",
    "\n",
    "dataset_dir = 'data/resized'          # Change path according to your machine!\n",
    "dataset = datasets.ImageFolder(dataset_dir, transform_train)\n",
    "# data_loader\n",
    "train_loader = torch.utils.data.DataLoader(dataset, batch_size=128, shuffle=True)\n",
    "temp = plt.imread(train_loader.dataset.imgs[0][0])\n",
    "if (temp.shape[0] != img_size) or (temp.shape[0] != img_size):\n",
    "    print ('Error! image size is not 64 x 64, run \"lfw_preprocess.py\" !!')\n",
    "    sys.exit(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Hyper-Parameters--\n",
    "\n",
    "#setting the learning rate\n",
    "lr = 0.0002\n",
    "\n",
    "#Number of samples to take a random distribution from\n",
    "rand_num = 9\n",
    "\n",
    "#Number of epochs\n",
    "num_epoch = 20\n",
    "\n",
    "#select min batch size\n",
    "batch_size = 128\n",
    "\n",
    "#Number of batches per Epoch\n",
    "num_batches = len(train_loader)\n",
    "\n",
    "\n",
    "if not os.path.isdir('LFW_DCGAN_results'):\n",
    "    os.mkdir('LFW_DCGAN_results')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Generator Network!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "class generator(nn.Module):\n",
    "    def __init__(self, d=128):\n",
    "        super(generator, self).__init__()\n",
    "        self.deconv1 = nn.ConvTranspose2d(100, d*8, 4, 1, 0)\n",
    "        self.deconv1_bn = nn.BatchNorm2d(d*8)\n",
    "        self.deconv2 = nn.ConvTranspose2d(d*8, d*4, 4, 2, 1)\n",
    "        self.deconv2_bn = nn.BatchNorm2d(d*4)\n",
    "        self.deconv3 = nn.ConvTranspose2d(d*4, d*2, 4, 2, 1)\n",
    "        self.deconv3_bn = nn.BatchNorm2d(d*2)\n",
    "        self.deconv4 = nn.ConvTranspose2d(d*2, d, 4, 2, 1)\n",
    "        self.deconv4_bn = nn.BatchNorm2d(d)\n",
    "        self.deconv5 = nn.ConvTranspose2d(d, 3, 4, 2, 1)\n",
    "\n",
    "    \n",
    "    def weight_init(self, mean, std):\n",
    "        for m in self._modules:\n",
    "            normal_init(self._modules[m], mean, std)\n",
    "\n",
    "    # Connecting the layers\n",
    "    def forward(self, input):\n",
    "        x = F.relu(self.deconv1_bn(self.deconv1(input)))\n",
    "        x = F.relu(self.deconv2_bn(self.deconv2(x)))\n",
    "        x = F.relu(self.deconv3_bn(self.deconv3(x)))\n",
    "        x = F.relu(self.deconv4_bn(self.deconv4(x)))\n",
    "        x = F.tanh(self.deconv5(x))\n",
    "\n",
    "        return x\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Discriminator Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "class discriminator(nn.Module):\n",
    "    def __init__(self, d=128):\n",
    "        super(discriminator, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(3, d, 4, 2, 1)\n",
    "        self.conv2 = nn.Conv2d(d, d*2, 4, 2, 1)\n",
    "        self.conv2_bn = nn.BatchNorm2d(d*2)\n",
    "        self.conv3 = nn.Conv2d(d*2, d*4, 4, 2, 1)\n",
    "        self.conv3_bn = nn.BatchNorm2d(d*4)\n",
    "        self.conv4 = nn.Conv2d(d*4, d*8, 4, 2, 1)\n",
    "        self.conv4_bn = nn.BatchNorm2d(d*8)\n",
    "        self.conv5 = nn.Conv2d(d*8, 1, 4, 1, 0)\n",
    "\n",
    "    def weight_init(self, mean, std):\n",
    "        for m in self._modules:\n",
    "            normal_init(self._modules[m], mean, std)\n",
    "\n",
    "    def forward(self, input):\n",
    "        x = F.leaky_relu(self.conv1(input), 0.2)\n",
    "        x = F.leaky_relu(self.conv2_bn(self.conv2(x)), 0.2)\n",
    "        x = F.leaky_relu(self.conv3_bn(self.conv3(x)), 0.2)\n",
    "        x = F.leaky_relu(self.conv4_bn(self.conv4(x)), 0.2)\n",
    "        x = F.sigmoid(self.conv5(x))\n",
    "\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# network\n",
    "G = generator(128)\n",
    "D = discriminator(128)\n",
    "\n",
    "def normal_init(m, mean, std):\n",
    "    if isinstance(m, nn.ConvTranspose2d) or isinstance(m, nn.Conv2d):\n",
    "        m.weight.data.normal_(mean, std)\n",
    "        m.bias.data.zero_()\n",
    "\n",
    "#Initializing weights\n",
    "G.weight_init(mean=0.0, std=0.02)\n",
    "D.weight_init(mean=0.0, std=0.02)\n",
    "\n",
    "# Binary Cross Entropy loss\n",
    "BCE_loss = nn.BCELoss()\n",
    "\n",
    "# Define optimizers\n",
    "G_optimizer = optim.Adam(G.parameters(), lr=lr, betas=(0.5, 0.999))\n",
    "D_optimizer = optim.Adam(D.parameters(), lr=lr, betas=(0.5, 0.999))\n",
    "\n",
    "\n",
    "\n",
    "fixed_z_ = torch.randn((5 * 5, 100)).view(-1, 100, 1, 1)    #noise\n",
    "fixed_z_ = Variable(fixed_z_, volatile=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize(num_epoch, path):\n",
    "    \n",
    "    z_ = torch.randn((5*5, 100)).view(-1, 100, 1, 1)\n",
    "    z_ = Variable(z_, volatile=True)\n",
    "\n",
    "    G.eval()\n",
    "    test_images = G(fixed_z_)\n",
    "    G.train()\n",
    "\n",
    "    size_figure_grid = 5\n",
    "    fig, ax = plt.subplots(size_figure_grid, size_figure_grid, figsize=(5, 5))\n",
    "    for i, j in itertools.product(range(size_figure_grid), range(size_figure_grid)):\n",
    "        ax[i, j].get_xaxis().set_visible(False)\n",
    "        ax[i, j].get_yaxis().set_visible(False)\n",
    "\n",
    "    for k in range(5*5):\n",
    "        i = k // 5\n",
    "        j = k % 5\n",
    "        ax[i, j].cla()\n",
    "        ax[i, j].imshow((test_images[k].cpu().data.numpy().transpose(1, 2, 0) + 1) / 2)\n",
    "\n",
    "    label = 'Epoch {0}'.format(num_epoch)\n",
    "    fig.text(0.5, 0.04, label, ha='center')\n",
    "    plt.savefig(path)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training..\n",
      "[1/20] - loss_d: 0.146, loss_g: 7.503\n",
      "[2/20] - loss_d: 0.330, loss_g: 8.765\n",
      "[3/20] - loss_d: 0.182, loss_g: 8.815\n",
      "[4/20] - loss_d: 0.167, loss_g: 9.085\n",
      "[5/20] - loss_d: 0.082, loss_g: 8.060\n",
      "[6/20] - loss_d: 0.133, loss_g: 10.367\n",
      "[7/20] - loss_d: 0.040, loss_g: 9.163\n",
      "[8/20] - loss_d: 0.095, loss_g: 10.032\n",
      "[9/20] - loss_d: 0.049, loss_g: 9.562\n",
      "[10/20] - loss_d: 0.099, loss_g: 12.124\n",
      "[11/20] - loss_d: 0.033, loss_g: 10.301\n",
      "[12/20] - loss_d: 0.090, loss_g: 12.472\n",
      "[13/20] - loss_d: 0.026, loss_g: 10.734\n",
      "[14/20] - loss_d: 0.073, loss_g: 12.554\n",
      "[15/20] - loss_d: 0.027, loss_g: 10.811\n",
      "[16/20] - loss_d: 0.064, loss_g: 12.977\n",
      "[17/20] - loss_d: 0.022, loss_g: 11.102\n",
      "[18/20] - loss_d: 0.058, loss_g: 13.241\n",
      "[19/20] - loss_d: 0.019, loss_g: 10.907\n",
      "[20/20] - loss_d: 0.131, loss_g: 20.500\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/lib/python2.7/dist-packages/matplotlib/pyplot.py:516: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).\n",
      "  max_open_warning, RuntimeWarning)\n"
     ]
    }
   ],
   "source": [
    "train_hist = {}\n",
    "train_hist['D_losses'] = []                                #storing losses for each batch\n",
    "train_hist['G_losses'] = []\n",
    "\n",
    "print('Training..')\n",
    "\n",
    "for epoch in range(num_epoch):\n",
    "    D_losses = []                                           #storing losses for each batch\n",
    "    G_losses = []\n",
    "    \n",
    "    num_iter = 0\n",
    "    for images, _ in train_loader:\n",
    "    \n",
    "        mini_batch = images.size()[0]\n",
    "        labels_real = Variable(torch.ones(mini_batch))       # Labels for real images - all ones\n",
    "        labels_fake = Variable(torch.zeros(mini_batch))      # Labels for fake images - all zeros\n",
    "\n",
    "        images = Variable(images)\n",
    "        \n",
    "        # Train the discriminator, it tries to discriminate between real and fake samples\n",
    "        D.zero_grad()\n",
    "        D_result = D(images).squeeze()\n",
    "        D_real_loss = BCE_loss(D_result, labels_real)\n",
    "\n",
    "        z_ = Variable(torch.randn((mini_batch, 100)).view(-1, 100, 1, 1))\n",
    "        G_result = G(z_)\n",
    " \n",
    "        D_result = D(G_result).squeeze()\n",
    "        D_fake_loss = BCE_loss(D_result, labels_fake)\n",
    "        D_fake_score = D_result.data.mean()\n",
    "\n",
    "        D_train_loss = D_real_loss + D_fake_loss        # Calculate the total loss\n",
    "\n",
    "        D_train_loss.backward()                         #Backpropagation\n",
    "        D_optimizer.step()                              #Update the weights\n",
    "\n",
    "        D_losses.append(D_train_loss.data[0])\n",
    "\n",
    "        # Train the generator, it tries to fool the discriminator\n",
    "        G.zero_grad()\n",
    "        \n",
    "        # Draw samples from the input distribution and pass to generator\n",
    "        z_ = Variable(torch.randn((mini_batch, 100)).view(-1, 100, 1, 1))\n",
    "        \n",
    "        G_result = G(z_)\n",
    "        D_result = D(G_result).squeeze()\n",
    "        G_train_loss = BCE_loss(D_result, labels_real)\n",
    "        G_train_loss.backward()\n",
    "        G_optimizer.step()\n",
    "\n",
    "        G_losses.append(G_train_loss.data[0])\n",
    "\n",
    "        num_iter += 1\n",
    "\n",
    "\n",
    "    print('[%d/%d] - loss_d: %.3f, loss_g: %.3f' % ((epoch + 1), num_epoch, torch.mean(torch.FloatTensor(D_losses)),\n",
    "                                                              torch.mean(torch.FloatTensor(G_losses))))\n",
    "    p = 'LFW_DCGAN_results/LFW_DCGAN_' + str(epoch + 1) + '.png'\n",
    "    visualize((epoch+1), path=p)\n",
    "    train_hist['D_losses'].append(torch.mean(torch.FloatTensor(D_losses)))\n",
    "    train_hist['G_losses'].append(torch.mean(torch.FloatTensor(G_losses)))\n",
    "    \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "with open(\"LFW_DCGAN_results/genloss.pkl\",\"wb\") as f:\n",
    "    pickle.dump(train_hist['G_losses'],f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"LFW_DCGAN_results/disloss.pkl\",\"wb\") as f:\n",
    "    pickle.dump(train_hist['D_losses'],f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6f2d3436d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_hist['G_losses'])\n",
    "plt.title('Generator Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEKCAYAAADpfBXhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQEruIiIVSMldRKQCKbmLiFSg/w9vBBM63u7/aQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f6f7c55d6d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_hist['D_losses'])\n",
    "plt.title('Discriminator Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}