diff --git "a/torch_binary2_v1.ipynb" "b/torch_binary2_v1.ipynb" new file mode 100644--- /dev/null +++ "b/torch_binary2_v1.ipynb" @@ -0,0 +1,1750 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 1, + "id": "cdd1e20f-136f-4203-b053-a346a16a7c8f", + "metadata": {}, + "outputs": [], + "source": [ + "import torch\n", + "from torch.utils.data import Dataset, DataLoader\n", + "import numpy as np\n", + "import os\n", + "\n", + "np_convert_dir = 'np_convert'\n", + "width_threshold = 350\n", + "\n", + "# Function to pad a matrix\n", + "def pad_matrix(matrix, target_height=50, target_width=350):\n", + " padding_height = max(target_height - matrix.shape[0], 0)\n", + " padding_width = max(target_width - matrix.shape[1], 0)\n", + " return np.pad(matrix, \n", + " pad_width=((0, padding_height), (0, padding_width)), \n", + " mode='constant', \n", + " constant_values=0)\n", + "\n", + "# Custom Dataset Class\n", + "class MyDataset(Dataset):\n", + " def __init__(self, np_convert_dir, width_threshold):\n", + " self.data = []\n", + " for file in os.listdir(np_convert_dir):\n", + " if file.endswith('.npy'):\n", + " file_path = os.path.join(np_convert_dir, file)\n", + " matrix = np.load(file_path)\n", + " \n", + " # Filter matrices with width less than the threshold\n", + " if matrix.shape[1] < width_threshold:\n", + " padded_matrix = pad_matrix(matrix)\n", + " self.data.append(padded_matrix)\n", + "\n", + " # Convert list of numpy arrays to a numpy array and then to a torch Tensor\n", + " self.data = torch.tensor(np.array(self.data), dtype=torch.float32).unsqueeze(1) # Add channel dimension\n", + "\n", + " def __len__(self):\n", + " return len(self.data)\n", + "\n", + " def __getitem__(self, idx):\n", + " return self.data[idx]\n", + "\n", + "# Initialize Dataset\n", + "dataset = MyDataset(np_convert_dir, width_threshold)\n", + "\n", + "# Create a DataLoader\n", + "batch_size = 16 # You can adjust the batch size\n", + "dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "id": "bcef16f6-1458-4661-8749-62f4bf424120", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Total number of matrices in the dataset after projection: 356\n" + ] + } + ], + "source": [ + "# Initialize Dataset\n", + "dataset = MyDataset(np_convert_dir, width_threshold)\n", + "\n", + "# Print out the total number of matrices in the dataset\n", + "total_matrices = len(dataset)\n", + "print(f\"Total number of matrices in the dataset after projection: {total_matrices}\")\n" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "id": "d5239aed-4c8a-4d3e-8087-5ab278ff9ba3", + "metadata": {}, + "outputs": [], + "source": [ + "import torch\n", + "import torch.nn as nn\n", + "import torch.optim as optim\n", + "\n", + "import torch.nn.functional as F\n", + "\n", + "\n", + "class ResidualBlock(nn.Module):\n", + " def __init__(self, channels):\n", + " super(ResidualBlock, self).__init__()\n", + " self.conv1 = nn.ConvTranspose2d(channels, channels, kernel_size=3, stride=1, padding=1)\n", + " self.bn1 = nn.BatchNorm2d(channels)\n", + " self.relu = nn.LeakyReLU(0.01)\n", + " self.dropout = nn.Dropout(0.3)\n", + "\n", + " def forward(self, x):\n", + " identity = x\n", + " out = self.conv1(x)\n", + " out = self.bn1(out)\n", + " out = self.relu(out)\n", + " out = self.dropout(out) \n", + " out = out + identity\n", + " return out\n", + "\n", + "class Generator(nn.Module):\n", + " def __init__(self, z_dim):\n", + " super(Generator, self).__init__()\n", + " self.fc = nn.Linear(z_dim, 25 * 175 * 256)\n", + " self.unflatten = nn.Unflatten(1, (256, 25, 175))\n", + "\n", + " self.upsample1 = nn.Upsample(scale_factor=(2, 1))\n", + " self.conv1 = nn.Conv2d(256, 256, kernel_size=(3, 3), padding=1)\n", + " self.bn1 = nn.BatchNorm2d(256)\n", + " self.relu1 = nn.LeakyReLU(0.01)\n", + " self.dropout1 = nn.Dropout(0.3)\n", + "\n", + " self.resblock1 = ResidualBlock(256)\n", + " self.resblock2 = ResidualBlock(256)\n", + "\n", + " self.upsample2 = nn.Upsample(scale_factor=(1, 2))\n", + " self.conv2 = nn.Conv2d(512, 128, kernel_size=(3, 3), padding=1)\n", + " self.bn2 = nn.BatchNorm2d(128)\n", + " self.relu2 = nn.LeakyReLU(0.01)\n", + " self.dropout2 = nn.Dropout(0.3)\n", + "\n", + " self.resblock3 = ResidualBlock(128)\n", + "\n", + " self.conv3 = nn.Conv2d(128, 1, kernel_size=(3, 3), padding=1)\n", + " self.sigmoid = nn.Sigmoid()\n", + "\n", + " def forward(self, x):\n", + " x = self.fc(x)\n", + " x = self.unflatten(x)\n", + "\n", + " # First upsample and convolution\n", + " x1 = self.upsample1(x)\n", + " x1 = self.conv1(x1)\n", + " x1 = self.bn1(x1)\n", + " x1 = self.relu1(x1)\n", + " x1 = self.dropout1(x1) # Applying dropout\n", + "\n", + " # Residual blocks\n", + " x2 = self.resblock1(x1)\n", + " x2 = self.resblock2(x2)\n", + "\n", + " # Concatenate skip connection\n", + " x3 = torch.cat([x2, x1], dim=1)\n", + "\n", + " # Second upsample and convolution\n", + " x3 = self.upsample2(x3)\n", + " x3 = self.conv2(x3)\n", + " x3 = self.bn2(x3)\n", + " x3 = self.relu2(x3)\n", + " x3 = self.dropout2(x3) # Applying dropout\n", + "\n", + " x3 = self.resblock3(x3)\n", + "\n", + " x3 = self.conv3(x3)\n", + " x3 = nn.Sigmoid()(x3)\n", + " return x3\n", + "\n", + "\n", + "# Discriminator Class\n", + "class Discriminator(nn.Module):\n", + " def __init__(self, img_shape):\n", + " super(Discriminator, self).__init__()\n", + " self.model = nn.Sequential(\n", + " nn.Conv2d(1, 64, kernel_size=3, stride=2, padding=1),\n", + " nn.LeakyReLU(0.01),\n", + " nn.Dropout(0.3),\n", + " nn.Conv2d(64, 128, kernel_size=3, stride=2, padding=1),\n", + " nn.LeakyReLU(0.01),\n", + " nn.Dropout(0.3),\n", + " nn.Flatten(),\n", + " nn.Linear(128 * 13 * 88, 1), # Updated input size\n", + " nn.Sigmoid()\n", + " )\n", + "\n", + " def forward(self, x):\n", + " return self.model(x)\n", + "\n", + "# Initialize models\n", + "z_dim = 100\n", + "img_shape = (1, 50, 350)\n", + "generator = Generator(z_dim)\n", + "discriminator = Discriminator(img_shape)\n", + "\n", + "# Loss and Optimizers\n", + "criterion = nn.BCELoss()\n", + "optimizer_g = optim.Adam(generator.parameters(), lr=0.0002)\n", + "optimizer_d = optim.Adam(discriminator.parameters(), lr=0.0002)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "id": "89b55806-a237-4729-8f02-70d32b90a604", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Using GPU: NVIDIA GeForce RTX 3080\n", + "Epoch [0/1000] Batch 0/23 Loss D: 1.3566068410873413, Loss G: 2.3029024600982666\n", + "Epoch [0/1000] Real Acc: 63.20%, Fake Acc: 37.64%\n", + "Epoch [1/1000] Batch 0/23 Loss D: 1.226878046989441, Loss G: 0.9252141714096069\n", + "Epoch [2/1000] Batch 0/23 Loss D: 1.219167709350586, Loss G: 0.6277535557746887\n", + "Epoch [3/1000] Batch 0/23 Loss D: 0.8834527134895325, Loss G: 0.9767117500305176\n", + "Epoch [4/1000] Batch 0/23 Loss D: 0.566815972328186, Loss G: 1.530390977859497\n", + "Epoch [5/1000] Batch 0/23 Loss D: 0.2712917923927307, Loss G: 2.2228713035583496\n", + "Epoch [5/1000] Real Acc: 99.16%, Fake Acc: 0.00%\n", + "Epoch [6/1000] Batch 0/23 Loss D: 0.1255910024046898, Loss G: 2.994349479675293\n", + "Epoch [7/1000] Batch 0/23 Loss D: 0.0527533832937479, Loss G: 3.243907928466797\n", + "Epoch [8/1000] Batch 0/23 Loss D: 18.40264318138361, Loss G: 2.3011603355407715\n", + "Epoch [9/1000] Batch 0/23 Loss D: 10.423457533121109, Loss G: 6.025820732116699\n", + "Epoch [10/1000] Batch 0/23 Loss D: 4.388331413269043, Loss G: 0.4863603115081787\n", + "Epoch [10/1000] Real Acc: 62.92%, Fake Acc: 69.94%\n", + "Epoch [11/1000] Batch 0/23 Loss D: 0.7153336703777313, Loss G: 1.1453564167022705\n", + "Epoch [12/1000] Batch 0/23 Loss D: 0.7368938326835632, Loss G: 1.3003307580947876\n", + "Epoch [13/1000] Batch 0/23 Loss D: 0.826795756816864, Loss G: 0.9249053001403809\n", + "Epoch [14/1000] Batch 0/23 Loss D: 0.6275086402893066, Loss G: 1.3920291662216187\n", + "Epoch [15/1000] Batch 0/23 Loss D: 0.6243380606174469, Loss G: 1.2248034477233887\n", + "Epoch [15/1000] Real Acc: 92.70%, Fake Acc: 0.00%\n", + "Epoch [16/1000] Batch 0/23 Loss D: 0.6928907036781311, Loss G: 1.2834453582763672\n", + "Epoch [17/1000] Batch 0/23 Loss D: 1.0854501128196716, Loss G: 1.0296900272369385\n", + "Epoch [18/1000] Batch 0/23 Loss D: 1.447881042957306, Loss G: 0.9744443297386169\n", + "Epoch [19/1000] Batch 0/23 Loss D: 1.8444455862045288, Loss G: 0.9577564001083374\n", + "Epoch [20/1000] Batch 0/23 Loss D: 1.2638452649116516, Loss G: 1.1851105690002441\n", + "Epoch [20/1000] Real Acc: 76.12%, Fake Acc: 3.37%\n", + "Epoch [21/1000] Batch 0/23 Loss D: 0.666571170091629, Loss G: 1.395964503288269\n", + "Epoch [22/1000] Batch 0/23 Loss D: 1.21759432554245, Loss G: 1.084275484085083\n", + "Epoch [23/1000] Batch 0/23 Loss D: 1.184739887714386, Loss G: 0.7926622033119202\n", + "Epoch [24/1000] Batch 0/23 Loss D: 1.0500310361385345, Loss G: 0.7580081224441528\n", + "Epoch [25/1000] Batch 0/23 Loss D: 1.1168898940086365, Loss G: 1.0688450336456299\n", + "Epoch [25/1000] Real Acc: 65.73%, Fake Acc: 30.34%\n", + "Epoch [26/1000] Batch 0/23 Loss D: 1.6720695495605469, Loss G: 0.8123888969421387\n", + "Epoch [27/1000] Batch 0/23 Loss D: 2.319933235645294, Loss G: 0.6497201919555664\n", + "Epoch [28/1000] Batch 0/23 Loss D: 2.0067341327667236, Loss G: 0.6201144456863403\n", + "Epoch [29/1000] Batch 0/23 Loss D: 1.4577693343162537, Loss G: 0.6872408390045166\n", + "Epoch [30/1000] Batch 0/23 Loss D: 1.506293773651123, Loss G: 0.6879397630691528\n", + "Epoch [30/1000] Real Acc: 63.20%, Fake Acc: 83.99%\n", + "Epoch [31/1000] Batch 0/23 Loss D: 1.5096220970153809, Loss G: 0.7091506719589233\n", + "Epoch [32/1000] Batch 0/23 Loss D: 1.3889896869659424, Loss G: 0.6823877096176147\n", + "Epoch [33/1000] Batch 0/23 Loss D: 1.6059531569480896, Loss G: 0.7258355617523193\n", + "Epoch [34/1000] Batch 0/23 Loss D: 1.85062974691391, Loss G: 0.7253603935241699\n", + "Epoch [35/1000] Batch 0/23 Loss D: 1.5816133618354797, Loss G: 0.7223132848739624\n", + "Epoch [35/1000] Real Acc: 42.70%, Fake Acc: 65.45%\n", + "Epoch [36/1000] Batch 0/23 Loss D: 1.4388436675071716, Loss G: 0.6061205863952637\n", + "Epoch [37/1000] Batch 0/23 Loss D: 1.4074071049690247, Loss G: 0.6255651712417603\n", + "Epoch [38/1000] Batch 0/23 Loss D: 1.2833685874938965, Loss G: 0.7322970628738403\n", + "Epoch [39/1000] Batch 0/23 Loss D: 1.2170986533164978, Loss G: 0.695654034614563\n", + "Epoch [40/1000] Batch 0/23 Loss D: 1.4146353602409363, Loss G: 0.7670130729675293\n", + "Epoch [40/1000] Real Acc: 39.33%, Fake Acc: 62.64%\n", + "Epoch [41/1000] Batch 0/23 Loss D: 1.745685636997223, Loss G: 0.7407842874526978\n", + "Epoch [42/1000] Batch 0/23 Loss D: 1.4437546133995056, Loss G: 0.7447655200958252\n", + "Epoch [43/1000] Batch 0/23 Loss D: 1.1502766609191895, Loss G: 0.8917789459228516\n", + "Epoch [44/1000] Batch 0/23 Loss D: 1.5302927494049072, Loss G: 0.7247343063354492\n", + "Epoch [45/1000] Batch 0/23 Loss D: 2.063424587249756, Loss G: 0.8739227652549744\n", + "Epoch [45/1000] Real Acc: 13.20%, Fake Acc: 31.18%\n", + "Epoch [46/1000] Batch 0/23 Loss D: 1.5677115321159363, Loss G: 0.5014032125473022\n", + "Epoch [47/1000] Batch 0/23 Loss D: 1.5039459466934204, Loss G: 0.6526497006416321\n", + "Epoch [48/1000] Batch 0/23 Loss D: 1.2569304704666138, Loss G: 0.7620984315872192\n", + "Epoch [49/1000] Batch 0/23 Loss D: 1.30226069688797, Loss G: 0.6161617040634155\n", + "Epoch [50/1000] Batch 0/23 Loss D: 1.2216981947422028, Loss G: 0.9881203174591064\n", + "Epoch [50/1000] Real Acc: 55.90%, Fake Acc: 2.81%\n", + "Epoch [51/1000] Batch 0/23 Loss D: 1.3714171648025513, Loss G: 0.7383532524108887\n", + "Epoch [52/1000] Batch 0/23 Loss D: 1.432130217552185, Loss G: 0.7340185642242432\n", + "Epoch [53/1000] Batch 0/23 Loss D: 1.2120231986045837, Loss G: 0.9031325578689575\n", + "Epoch [54/1000] Batch 0/23 Loss D: 1.2832903861999512, Loss G: 0.7105805277824402\n", + "Epoch [55/1000] Batch 0/23 Loss D: 1.4068577885627747, Loss G: 0.7484405040740967\n", + "Epoch [55/1000] Real Acc: 21.63%, Fake Acc: 32.87%\n", + "Epoch [56/1000] Batch 0/23 Loss D: 1.5624384880065918, Loss G: 0.729198694229126\n", + "Epoch [57/1000] Batch 0/23 Loss D: 1.5176562070846558, Loss G: 0.6474739909172058\n", + "Epoch [58/1000] Batch 0/23 Loss D: 1.4286177158355713, Loss G: 0.7095400094985962\n", + "Epoch [59/1000] Batch 0/23 Loss D: 1.3993130326271057, Loss G: 0.6769618988037109\n", + "Epoch [60/1000] Batch 0/23 Loss D: 1.379546880722046, Loss G: 0.6700743436813354\n", + "Epoch [60/1000] Real Acc: 53.65%, Fake Acc: 58.71%\n", + "Epoch [61/1000] Batch 0/23 Loss D: 1.4229402542114258, Loss G: 0.6694915294647217\n", + "Epoch [62/1000] Batch 0/23 Loss D: 1.3771455883979797, Loss G: 0.7156953811645508\n", + "Epoch [63/1000] Batch 0/23 Loss D: 1.4198797941207886, Loss G: 0.7217038869857788\n", + "Epoch [64/1000] Batch 0/23 Loss D: 1.3461449146270752, Loss G: 0.697999119758606\n", + "Epoch [65/1000] Batch 0/23 Loss D: 1.3742748498916626, Loss G: 0.7329258918762207\n", + "Epoch [65/1000] Real Acc: 52.25%, Fake Acc: 27.81%\n", + "Epoch [66/1000] Batch 0/23 Loss D: 1.3376386165618896, Loss G: 0.6893775463104248\n", + "Epoch [67/1000] Batch 0/23 Loss D: 1.307963252067566, Loss G: 0.7097020149230957\n", + "Epoch [68/1000] Batch 0/23 Loss D: 1.3002374768257141, Loss G: 0.7487878799438477\n", + "Epoch [69/1000] Batch 0/23 Loss D: 1.3220126628875732, Loss G: 0.71403968334198\n", + "Epoch [70/1000] Batch 0/23 Loss D: 1.3009387254714966, Loss G: 0.7862895727157593\n", + "Epoch [70/1000] Real Acc: 53.09%, Fake Acc: 8.99%\n", + "Epoch [71/1000] Batch 0/23 Loss D: 1.349991500377655, Loss G: 0.7581857442855835\n", + "Epoch [72/1000] Batch 0/23 Loss D: 1.2814869284629822, Loss G: 0.7385209798812866\n", + "Epoch [73/1000] Batch 0/23 Loss D: 1.3236351609230042, Loss G: 0.760710597038269\n", + "Epoch [74/1000] Batch 0/23 Loss D: 1.2219421863555908, Loss G: 0.8949530124664307\n", + "Epoch [75/1000] Batch 0/23 Loss D: 1.4474236965179443, Loss G: 0.7034531831741333\n", + "Epoch [75/1000] Real Acc: 26.12%, Fake Acc: 20.79%\n", + "Epoch [76/1000] Batch 0/23 Loss D: 1.3983941674232483, Loss G: 0.805158257484436\n", + "Epoch [77/1000] Batch 0/23 Loss D: 1.3281259536743164, Loss G: 0.6709473133087158\n", + "Epoch [78/1000] Batch 0/23 Loss D: 1.3358154892921448, Loss G: 0.6951057314872742\n", + "Epoch [79/1000] Batch 0/23 Loss D: 1.4058600664138794, Loss G: 0.7173740863800049\n", + "Epoch [80/1000] Batch 0/23 Loss D: 1.3928048610687256, Loss G: 0.8206164240837097\n", + "Epoch [80/1000] Real Acc: 46.91%, Fake Acc: 15.17%\n", + "Epoch [81/1000] Batch 0/23 Loss D: 1.3404240608215332, Loss G: 0.7439298629760742\n", + "Epoch [82/1000] Batch 0/23 Loss D: 1.3886755108833313, Loss G: 0.85617995262146\n", + "Epoch [83/1000] Batch 0/23 Loss D: 1.5947150588035583, Loss G: 0.6265963912010193\n", + "Epoch [84/1000] Batch 0/23 Loss D: 1.4986259937286377, Loss G: 0.6464243531227112\n", + "Epoch [85/1000] Batch 0/23 Loss D: 1.3422194719314575, Loss G: 0.7002148032188416\n", + "Epoch [85/1000] Real Acc: 90.73%, Fake Acc: 11.52%\n", + "Epoch [86/1000] Batch 0/23 Loss D: 1.0118408799171448, Loss G: 0.865095317363739\n", + "Epoch [87/1000] Batch 0/23 Loss D: 0.7166506052017212, Loss G: 1.0873935222625732\n", + "Epoch [88/1000] Batch 0/23 Loss D: 0.5162577033042908, Loss G: 1.4939069747924805\n", + "Epoch [89/1000] Batch 0/23 Loss D: 0.36952637135982513, Loss G: 1.807910680770874\n", + "Epoch [90/1000] Batch 0/23 Loss D: 0.20267826318740845, Loss G: 2.368997097015381\n", + "Epoch [90/1000] Real Acc: 99.16%, Fake Acc: 0.00%\n", + "Epoch [91/1000] Batch 0/23 Loss D: 0.19870954751968384, Loss G: 2.5363473892211914\n", + "Epoch [92/1000] Batch 0/23 Loss D: 0.09119760617613792, Loss G: 2.9885475635528564\n", + "Epoch [93/1000] Batch 0/23 Loss D: 0.06161687895655632, Loss G: 3.3063831329345703\n", + "Epoch [94/1000] Batch 0/23 Loss D: 0.06570183858275414, Loss G: 3.69757080078125\n", + "Epoch [95/1000] Batch 0/23 Loss D: 0.09424623474478722, Loss G: 3.272230625152588\n", + "Epoch [95/1000] Real Acc: 100.00%, Fake Acc: 0.00%\n", + "Epoch [96/1000] Batch 0/23 Loss D: 0.059447672218084335, Loss G: 3.5765810012817383\n", + "Epoch [97/1000] Batch 0/23 Loss D: 0.029303663410246372, Loss G: 4.008110046386719\n", + "Epoch [98/1000] Batch 0/23 Loss D: 0.02905808761715889, Loss G: 3.97983980178833\n", + "Epoch [99/1000] Batch 0/23 Loss D: 0.02982372883707285, Loss G: 4.175451278686523\n", + "Epoch [100/1000] Batch 0/23 Loss D: 0.024879186414182186, Loss G: 4.338191986083984\n", + "Epoch [100/1000] Real Acc: 99.72%, Fake Acc: 0.00%\n", + "Epoch [101/1000] Batch 0/23 Loss D: 0.017482892610132694, Loss G: 4.62394905090332\n", + "Epoch [102/1000] Batch 0/23 Loss D: 0.02218971773982048, Loss G: 4.364886283874512\n", + "Epoch [103/1000] Batch 0/23 Loss D: 0.026660026982426643, Loss G: 4.326655387878418\n", + "Epoch [104/1000] Batch 0/23 Loss D: 0.03558478783816099, Loss G: 4.548123836517334\n", + "Epoch [105/1000] Batch 0/23 Loss D: 0.015114059671759605, Loss G: 4.655581474304199\n", + "Epoch [105/1000] Real Acc: 100.00%, Fake Acc: 0.00%\n", + "Epoch [106/1000] Batch 0/23 Loss D: 0.013976460322737694, Loss G: 4.830254554748535\n", + "Epoch [107/1000] Batch 0/23 Loss D: 0.010250312509015203, Loss G: 4.9672393798828125\n", + "Epoch [108/1000] Batch 0/23 Loss D: 0.010016631567850709, Loss G: 5.005836486816406\n", + "Epoch [109/1000] Batch 0/23 Loss D: 0.012382036540657282, Loss G: 4.701274871826172\n", + "Epoch [110/1000] Batch 0/23 Loss D: 0.01343381218612194, Loss G: 4.743256568908691\n", + "Epoch [110/1000] Real Acc: 100.00%, Fake Acc: 0.00%\n", + "Epoch [111/1000] Batch 0/23 Loss D: 0.010444696061313152, Loss G: 5.016976356506348\n", + "Epoch [112/1000] Batch 0/23 Loss D: 0.01886047702282667, Loss G: 5.3093767166137695\n", + "Epoch [113/1000] Batch 0/23 Loss D: 0.030861418694257736, Loss G: 5.338855743408203\n", + "Epoch [114/1000] Batch 0/23 Loss D: 0.009353988338261843, Loss G: 5.27968692779541\n", + "Epoch [115/1000] Batch 0/23 Loss D: 15.126532942056656, Loss G: 0.0003169609699398279\n", + "Epoch [115/1000] Real Acc: 13.20%, Fake Acc: 40.73%\n", + "Epoch [116/1000] Batch 0/23 Loss D: 9.654601015150547, Loss G: 2.4091765880584717\n", + "Epoch [117/1000] Batch 0/23 Loss D: 2.685072898864746, Loss G: 0.27516454458236694\n", + "Epoch [118/1000] Batch 0/23 Loss D: 1.0563113689422607, Loss G: 1.0655291080474854\n", + "Epoch [119/1000] Batch 0/23 Loss D: 1.221098005771637, Loss G: 0.6903866529464722\n", + "Epoch [120/1000] Batch 0/23 Loss D: 1.2958286702632904, Loss G: 0.5611453056335449\n", + "Epoch [120/1000] Real Acc: 96.35%, Fake Acc: 89.89%\n", + "Epoch [121/1000] Batch 0/23 Loss D: 1.3144029378890991, Loss G: 0.722085177898407\n", + "Epoch [122/1000] Batch 0/23 Loss D: 1.429170310497284, Loss G: 0.7452791929244995\n", + "Epoch [123/1000] Batch 0/23 Loss D: 1.3921171426773071, Loss G: 0.7668530941009521\n", + "Epoch [124/1000] Batch 0/23 Loss D: 1.4415764808654785, Loss G: 0.7708078622817993\n", + "Epoch [125/1000] Batch 0/23 Loss D: 1.4451816082000732, Loss G: 0.6978389024734497\n", + "Epoch [125/1000] Real Acc: 32.58%, Fake Acc: 47.19%\n", + "Epoch [126/1000] Batch 0/23 Loss D: 1.5461853742599487, Loss G: 0.7282048463821411\n", + "Epoch [127/1000] Batch 0/23 Loss D: 1.543860673904419, Loss G: 0.717547595500946\n", + "Epoch [128/1000] Batch 0/23 Loss D: 1.4576064348220825, Loss G: 0.6977700591087341\n", + "Epoch [129/1000] Batch 0/23 Loss D: 1.345940113067627, Loss G: 0.6912244558334351\n", + "Epoch [130/1000] Batch 0/23 Loss D: 1.3006392121315002, Loss G: 0.6990196704864502\n", + "Epoch [130/1000] Real Acc: 87.64%, Fake Acc: 62.08%\n", + "Epoch [131/1000] Batch 0/23 Loss D: 1.2649101614952087, Loss G: 0.7083954811096191\n", + "Epoch [132/1000] Batch 0/23 Loss D: 1.439742088317871, Loss G: 0.6573926210403442\n", + "Epoch [133/1000] Batch 0/23 Loss D: 1.4850196838378906, Loss G: 0.7886487245559692\n", + "Epoch [134/1000] Batch 0/23 Loss D: 1.5995767712593079, Loss G: 0.6766701340675354\n", + "Epoch [135/1000] Batch 0/23 Loss D: 1.3689296245574951, Loss G: 0.7524153590202332\n", + "Epoch [135/1000] Real Acc: 47.19%, Fake Acc: 9.27%\n", + "Epoch [136/1000] Batch 0/23 Loss D: 1.3079585433006287, Loss G: 0.7687126398086548\n", + "Epoch [137/1000] Batch 0/23 Loss D: 1.2042319774627686, Loss G: 0.6798654198646545\n", + "Epoch [138/1000] Batch 0/23 Loss D: 1.4110753536224365, Loss G: 0.6637183427810669\n", + "Epoch [139/1000] Batch 0/23 Loss D: 1.619062900543213, Loss G: 0.76114422082901\n", + "Epoch [140/1000] Batch 0/23 Loss D: 1.5873793959617615, Loss G: 0.7028335332870483\n", + "Epoch [140/1000] Real Acc: 36.80%, Fake Acc: 71.07%\n", + "Epoch [141/1000] Batch 0/23 Loss D: 1.4690414667129517, Loss G: 0.6414059400558472\n", + "Epoch [142/1000] Batch 0/23 Loss D: 1.3533061742782593, Loss G: 0.6881110668182373\n", + "Epoch [143/1000] Batch 0/23 Loss D: 1.2614710330963135, Loss G: 0.7877911925315857\n", + "Epoch [144/1000] Batch 0/23 Loss D: 1.2504699230194092, Loss G: 0.7473196983337402\n", + "Epoch [145/1000] Batch 0/23 Loss D: 1.3405001759529114, Loss G: 0.688910961151123\n", + "Epoch [145/1000] Real Acc: 49.44%, Fake Acc: 60.67%\n", + "Epoch [146/1000] Batch 0/23 Loss D: 1.4874876141548157, Loss G: 0.7156736850738525\n", + "Epoch [147/1000] Batch 0/23 Loss D: 1.62969309091568, Loss G: 0.7271918654441833\n", + "Epoch [148/1000] Batch 0/23 Loss D: 1.5202275514602661, Loss G: 0.7801498174667358\n", + "Epoch [149/1000] Batch 0/23 Loss D: 1.3525118827819824, Loss G: 0.8041303753852844\n", + "Epoch [150/1000] Batch 0/23 Loss D: 1.3013441562652588, Loss G: 0.7462747097015381\n", + "Epoch [150/1000] Real Acc: 84.83%, Fake Acc: 10.11%\n", + "Epoch [151/1000] Batch 0/23 Loss D: 1.2914479970932007, Loss G: 0.7209265828132629\n", + "Epoch [152/1000] Batch 0/23 Loss D: 1.269711971282959, Loss G: 0.6821837425231934\n", + "Epoch [153/1000] Batch 0/23 Loss D: 1.4374898076057434, Loss G: 0.6531413793563843\n", + "Epoch [154/1000] Batch 0/23 Loss D: 1.4936625957489014, Loss G: 0.7063117027282715\n", + "Epoch [155/1000] Batch 0/23 Loss D: 1.4789442420005798, Loss G: 0.7052031755447388\n", + "Epoch [155/1000] Real Acc: 29.49%, Fake Acc: 54.49%\n", + "Epoch [156/1000] Batch 0/23 Loss D: 1.4558178186416626, Loss G: 0.7067527770996094\n", + "Epoch [157/1000] Batch 0/23 Loss D: 1.4458813071250916, Loss G: 0.7357759475708008\n", + "Epoch [158/1000] Batch 0/23 Loss D: 1.2954081892967224, Loss G: 0.7930381298065186\n", + "Epoch [159/1000] Batch 0/23 Loss D: 1.174744576215744, Loss G: 0.7140494585037231\n", + "Epoch [160/1000] Batch 0/23 Loss D: 0.9668097496032715, Loss G: 0.8018391132354736\n", + "Epoch [160/1000] Real Acc: 100.00%, Fake Acc: 0.00%\n", + "Epoch [161/1000] Batch 0/23 Loss D: 0.8165137469768524, Loss G: 0.9103101491928101\n", + "Epoch [162/1000] Batch 0/23 Loss D: 0.6408390998840332, Loss G: 1.240312933921814\n", + "Epoch [163/1000] Batch 0/23 Loss D: 0.43970873951911926, Loss G: 1.6831663846969604\n", + "Epoch [164/1000] Batch 0/23 Loss D: 0.5612220913171768, Loss G: 1.3824470043182373\n", + "Epoch [165/1000] Batch 0/23 Loss D: 0.30290810763835907, Loss G: 2.516434669494629\n", + "Epoch [165/1000] Real Acc: 99.16%, Fake Acc: 0.00%\n", + "Epoch [166/1000] Batch 0/23 Loss D: 0.18201866000890732, Loss G: 2.3400139808654785\n", + "Epoch [167/1000] Batch 0/23 Loss D: 0.11312942579388618, Loss G: 2.9021730422973633\n", + "Epoch [168/1000] Batch 0/23 Loss D: 0.09749546274542809, Loss G: 3.3150806427001953\n", + "Epoch [169/1000] Batch 0/23 Loss D: 0.08048674091696739, Loss G: 3.4571080207824707\n", + "Epoch [170/1000] Batch 0/23 Loss D: 0.033622995018959045, Loss G: 3.6312475204467773\n", + "Epoch [170/1000] Real Acc: 99.72%, Fake Acc: 0.00%\n", + "Epoch [171/1000] Batch 0/23 Loss D: 0.03941074572503567, Loss G: 3.8955190181732178\n", + "Epoch [172/1000] Batch 0/23 Loss D: 0.0403500497341156, Loss G: 3.955806255340576\n", + "Epoch [173/1000] Batch 0/23 Loss D: 0.027411023154854774, Loss G: 4.0483903884887695\n", + "Epoch [174/1000] Batch 0/23 Loss D: 0.09216432645916939, Loss G: 3.9731907844543457\n", + "Epoch [175/1000] Batch 0/23 Loss D: 0.024159252643585205, Loss G: 4.0227556228637695\n", + "Epoch [175/1000] Real Acc: 99.72%, Fake Acc: 0.00%\n", + "Epoch [176/1000] Batch 0/23 Loss D: 0.01949064712971449, Loss G: 4.532188892364502\n", + "Epoch [177/1000] Batch 0/23 Loss D: 0.013826525304466486, Loss G: 4.88571310043335\n", + "Epoch [178/1000] Batch 0/23 Loss D: 0.03585527837276459, Loss G: 4.719040870666504\n", + "Epoch [179/1000] Batch 0/23 Loss D: 0.016620633425191045, Loss G: 4.286524772644043\n", + "Epoch [180/1000] Batch 0/23 Loss D: 0.01652768114581704, Loss G: 4.526773452758789\n", + "Epoch [180/1000] Real Acc: 100.00%, Fake Acc: 0.00%\n", + "Epoch [181/1000] Batch 0/23 Loss D: 0.015424922574311495, Loss G: 4.892144203186035\n", + "Epoch [182/1000] Batch 0/23 Loss D: 0.010965888854116201, Loss G: 4.835153102874756\n", + "Epoch [183/1000] Batch 0/23 Loss D: 5.55500864982605, Loss G: 4.1642231941223145\n", + "Epoch [184/1000] Batch 0/23 Loss D: 5.587210655212402, Loss G: 0.18979988992214203\n", + "Epoch [185/1000] Batch 0/23 Loss D: 2.375686287879944, Loss G: 0.755836009979248\n", + "Epoch [185/1000] Real Acc: 27.81%, Fake Acc: 46.63%\n", + "Epoch [186/1000] Batch 0/23 Loss D: 1.370305597782135, Loss G: 0.5502208471298218\n", + "Epoch [187/1000] Batch 0/23 Loss D: 1.3045029640197754, Loss G: 0.966076135635376\n", + "Epoch [188/1000] Batch 0/23 Loss D: 1.3889209032058716, Loss G: 0.7424423694610596\n", + "Epoch [189/1000] Batch 0/23 Loss D: 1.3626781105995178, Loss G: 0.7018923759460449\n", + "Epoch [190/1000] Batch 0/23 Loss D: 1.2460302114486694, Loss G: 0.8551412224769592\n", + "Epoch [190/1000] Real Acc: 73.60%, Fake Acc: 42.13%\n", + "Epoch [191/1000] Batch 0/23 Loss D: 1.7644831538200378, Loss G: 0.5315945744514465\n", + "Epoch [192/1000] Batch 0/23 Loss D: 1.3180468082427979, Loss G: 1.0869333744049072\n", + "Epoch [193/1000] Batch 0/23 Loss D: 1.0861778259277344, Loss G: 0.8827741146087646\n", + "Epoch [194/1000] Batch 0/23 Loss D: 1.4483409225940704, Loss G: 0.6043944954872131\n", + "Epoch [195/1000] Batch 0/23 Loss D: 1.6117122769355774, Loss G: 0.8957263827323914\n", + "Epoch [195/1000] Real Acc: 1.40%, Fake Acc: 20.22%\n", + "Epoch [196/1000] Batch 0/23 Loss D: 1.7304674983024597, Loss G: 0.784781277179718\n", + "Epoch [197/1000] Batch 0/23 Loss D: 1.543292224407196, Loss G: 0.8542263507843018\n", + "Epoch [198/1000] Batch 0/23 Loss D: 1.4271036982536316, Loss G: 0.7822456955909729\n", + "Epoch [199/1000] Batch 0/23 Loss D: 1.401270568370819, Loss G: 0.6748315095901489\n", + "Epoch [200/1000] Batch 0/23 Loss D: 1.3728432059288025, Loss G: 0.6913474798202515\n", + "Epoch [200/1000] Real Acc: 62.08%, Fake Acc: 58.99%\n", + "Epoch [201/1000] Batch 0/23 Loss D: 1.3522489070892334, Loss G: 0.707693874835968\n", + "Epoch [202/1000] Batch 0/23 Loss D: 1.3759379982948303, Loss G: 0.7198203802108765\n", + "Epoch [203/1000] Batch 0/23 Loss D: 1.4625535607337952, Loss G: 0.6221769452095032\n", + "Epoch [204/1000] Batch 0/23 Loss D: 1.5144935846328735, Loss G: 0.6385952234268188\n", + "Epoch [205/1000] Batch 0/23 Loss D: 1.5278437733650208, Loss G: 0.5942342877388\n", + "Epoch [205/1000] Real Acc: 42.13%, Fake Acc: 79.21%\n", + "Epoch [206/1000] Batch 0/23 Loss D: 1.3697686791419983, Loss G: 0.6917495727539062\n", + "Epoch [207/1000] Batch 0/23 Loss D: 1.3964524269104004, Loss G: 0.7334721088409424\n", + "Epoch [208/1000] Batch 0/23 Loss D: 1.2901886105537415, Loss G: 0.8217331171035767\n", + "Epoch [209/1000] Batch 0/23 Loss D: 1.2945688962936401, Loss G: 0.7687621116638184\n", + "Epoch [210/1000] Batch 0/23 Loss D: 1.2915552854537964, Loss G: 0.7788785696029663\n", + "Epoch [210/1000] Real Acc: 70.51%, Fake Acc: 31.18%\n", + "Epoch [211/1000] Batch 0/23 Loss D: 1.3309801816940308, Loss G: 0.7140809297561646\n", + "Epoch [212/1000] Batch 0/23 Loss D: 1.400968313217163, Loss G: 0.7420631647109985\n", + "Epoch [213/1000] Batch 0/23 Loss D: 1.3814700245857239, Loss G: 0.7816038131713867\n", + "Epoch [214/1000] Batch 0/23 Loss D: 1.3514155149459839, Loss G: 0.7331911325454712\n", + "Epoch [215/1000] Batch 0/23 Loss D: 1.361789882183075, Loss G: 0.7253904342651367\n", + "Epoch [215/1000] Real Acc: 66.85%, Fake Acc: 28.93%\n", + "Epoch [216/1000] Batch 0/23 Loss D: 1.3556568622589111, Loss G: 0.7130385041236877\n", + "Epoch [217/1000] Batch 0/23 Loss D: 1.350392460823059, Loss G: 0.6799065470695496\n", + "Epoch [218/1000] Batch 0/23 Loss D: 1.389225959777832, Loss G: 0.7036751508712769\n", + "Epoch [219/1000] Batch 0/23 Loss D: 1.3088120222091675, Loss G: 0.7051156163215637\n", + "Epoch [220/1000] Batch 0/23 Loss D: 1.3800670504570007, Loss G: 0.6785389184951782\n", + "Epoch [220/1000] Real Acc: 50.28%, Fake Acc: 61.52%\n", + "Epoch [221/1000] Batch 0/23 Loss D: 1.4220040440559387, Loss G: 0.7182072401046753\n", + "Epoch [222/1000] Batch 0/23 Loss D: 1.4317166805267334, Loss G: 0.7338818311691284\n", + "Epoch [223/1000] Batch 0/23 Loss D: 1.3449783325195312, Loss G: 0.8024376630783081\n", + "Epoch [224/1000] Batch 0/23 Loss D: 1.3373899459838867, Loss G: 0.7694451808929443\n", + "Epoch [225/1000] Batch 0/23 Loss D: 1.2976144552230835, Loss G: 0.7450990676879883\n", + "Epoch [225/1000] Real Acc: 60.67%, Fake Acc: 32.58%\n", + "Epoch [226/1000] Batch 0/23 Loss D: 1.4234784245491028, Loss G: 0.761986494064331\n", + "Epoch [227/1000] Batch 0/23 Loss D: 1.4522876739501953, Loss G: 0.7521270513534546\n", + "Epoch [228/1000] Batch 0/23 Loss D: 1.4390273690223694, Loss G: 0.6927119493484497\n", + "Epoch [229/1000] Batch 0/23 Loss D: 1.3521159887313843, Loss G: 0.7132792472839355\n", + "Epoch [230/1000] Batch 0/23 Loss D: 1.3011763095855713, Loss G: 0.7601574659347534\n", + "Epoch [230/1000] Real Acc: 81.18%, Fake Acc: 7.87%\n", + "Epoch [231/1000] Batch 0/23 Loss D: 1.2228386402130127, Loss G: 0.7773152589797974\n", + "Epoch [232/1000] Batch 0/23 Loss D: 1.330244541168213, Loss G: 0.6687941551208496\n", + "Epoch [233/1000] Batch 0/23 Loss D: 1.4438072443008423, Loss G: 0.6913561224937439\n", + "Epoch [234/1000] Batch 0/23 Loss D: 1.6272246837615967, Loss G: 0.7044727206230164\n", + "Epoch [235/1000] Batch 0/23 Loss D: 1.42865389585495, Loss G: 0.7155943512916565\n", + "Epoch [235/1000] Real Acc: 33.99%, Fake Acc: 24.44%\n", + "Epoch [236/1000] Batch 0/23 Loss D: 1.2904733419418335, Loss G: 0.7700707912445068\n", + "Epoch [237/1000] Batch 0/23 Loss D: 1.2540103197097778, Loss G: 0.817524790763855\n", + "Epoch [238/1000] Batch 0/23 Loss D: 1.3122942447662354, Loss G: 0.7656417489051819\n", + "Epoch [239/1000] Batch 0/23 Loss D: 1.6114913821220398, Loss G: 0.6785897612571716\n", + "Epoch [240/1000] Batch 0/23 Loss D: 1.3558874130249023, Loss G: 0.7651538848876953\n", + "Epoch [240/1000] Real Acc: 70.51%, Fake Acc: 3.93%\n", + "Epoch [241/1000] Batch 0/23 Loss D: 1.1711786389350891, Loss G: 0.7237446904182434\n", + "Epoch [242/1000] Batch 0/23 Loss D: 1.565301537513733, Loss G: 0.4845144748687744\n", + "Epoch [243/1000] Batch 0/23 Loss D: 1.424892008304596, Loss G: 1.0897858142852783\n", + "Epoch [244/1000] Batch 0/23 Loss D: 1.2372891306877136, Loss G: 0.8486006259918213\n", + "Epoch [245/1000] Batch 0/23 Loss D: 1.3684066534042358, Loss G: 0.6701468825340271\n", + "Epoch [245/1000] Real Acc: 74.72%, Fake Acc: 74.72%\n", + "Epoch [246/1000] Batch 0/23 Loss D: 1.2928868532180786, Loss G: 0.7291732430458069\n", + "Epoch [247/1000] Batch 0/23 Loss D: 1.3358047008514404, Loss G: 0.9177525043487549\n", + "Epoch [248/1000] Batch 0/23 Loss D: 1.4746721982955933, Loss G: 0.718148946762085\n", + "Epoch [249/1000] Batch 0/23 Loss D: 1.6012911200523376, Loss G: 0.5602088570594788\n", + "Epoch [250/1000] Batch 0/23 Loss D: 1.4221833944320679, Loss G: 0.7781375646591187\n", + "Epoch [250/1000] Real Acc: 28.09%, Fake Acc: 2.81%\n", + "Epoch [251/1000] Batch 0/23 Loss D: 1.1773136854171753, Loss G: 0.9588329792022705\n", + "Epoch [252/1000] Batch 0/23 Loss D: 1.0303861498832703, Loss G: 0.8851476907730103\n", + "Epoch [253/1000] Batch 0/23 Loss D: 1.2420106530189514, Loss G: 0.5316321849822998\n", + "Epoch [254/1000] Batch 0/23 Loss D: 1.4970580041408539, Loss G: 0.5522177219390869\n", + "Epoch [255/1000] Batch 0/23 Loss D: 1.6203969717025757, Loss G: 0.9947744011878967\n", + "Epoch [255/1000] Real Acc: 1.40%, Fake Acc: 8.71%\n", + "Epoch [256/1000] Batch 0/23 Loss D: 1.5631073713302612, Loss G: 0.7025097608566284\n", + "Epoch [257/1000] Batch 0/23 Loss D: 1.4718459844589233, Loss G: 0.6865710020065308\n", + "Epoch [258/1000] Batch 0/23 Loss D: 1.3314658403396606, Loss G: 0.8155374526977539\n", + "Epoch [259/1000] Batch 0/23 Loss D: 1.3119566440582275, Loss G: 0.8188046216964722\n", + "Epoch [260/1000] Batch 0/23 Loss D: 1.3646873235702515, Loss G: 0.7353818416595459\n", + "Epoch [260/1000] Real Acc: 68.54%, Fake Acc: 28.93%\n", + "Epoch [261/1000] Batch 0/23 Loss D: 1.354027509689331, Loss G: 0.6919835209846497\n", + "Epoch [262/1000] Batch 0/23 Loss D: 1.3677789568901062, Loss G: 0.6608819961547852\n", + "Epoch [263/1000] Batch 0/23 Loss D: 1.3392975330352783, Loss G: 0.6319231986999512\n", + "Epoch [264/1000] Batch 0/23 Loss D: 1.3768346309661865, Loss G: 0.6658700704574585\n", + "Epoch [265/1000] Batch 0/23 Loss D: 1.312526524066925, Loss G: 0.7157108783721924\n", + "Epoch [265/1000] Real Acc: 55.90%, Fake Acc: 25.56%\n", + "Epoch [266/1000] Batch 0/23 Loss D: 1.3385170698165894, Loss G: 0.7418551445007324\n", + "Epoch [267/1000] Batch 0/23 Loss D: 1.3174339532852173, Loss G: 0.749839186668396\n", + "Epoch [268/1000] Batch 0/23 Loss D: 1.3761246800422668, Loss G: 0.6918556690216064\n", + "Epoch [269/1000] Batch 0/23 Loss D: 1.415559709072113, Loss G: 0.7676464319229126\n", + "Epoch [270/1000] Batch 0/23 Loss D: 1.3464385271072388, Loss G: 0.8621096611022949\n", + "Epoch [270/1000] Real Acc: 26.69%, Fake Acc: 4.78%\n", + "Epoch [271/1000] Batch 0/23 Loss D: 1.3184148669242859, Loss G: 0.7921794652938843\n", + "Epoch [272/1000] Batch 0/23 Loss D: 1.3017588257789612, Loss G: 0.731876015663147\n", + "Epoch [273/1000] Batch 0/23 Loss D: 1.4345589876174927, Loss G: 0.6691548228263855\n", + "Epoch [274/1000] Batch 0/23 Loss D: 1.3798001408576965, Loss G: 0.6829518675804138\n", + "Epoch [275/1000] Batch 0/23 Loss D: 1.4245054721832275, Loss G: 0.6755024194717407\n", + "Epoch [275/1000] Real Acc: 57.02%, Fake Acc: 58.71%\n", + "Epoch [276/1000] Batch 0/23 Loss D: 1.3221465349197388, Loss G: 0.7538926601409912\n", + "Epoch [277/1000] Batch 0/23 Loss D: 1.2718742489814758, Loss G: 0.7836689949035645\n", + "Epoch [278/1000] Batch 0/23 Loss D: 1.2611377239227295, Loss G: 0.8119997978210449\n", + "Epoch [279/1000] Batch 0/23 Loss D: 1.3081843852996826, Loss G: 0.7660112977027893\n", + "Epoch [280/1000] Batch 0/23 Loss D: 1.4262670278549194, Loss G: 0.694740891456604\n", + "Epoch [280/1000] Real Acc: 39.89%, Fake Acc: 51.12%\n", + "Epoch [281/1000] Batch 0/23 Loss D: 1.4343042373657227, Loss G: 0.6790332794189453\n", + "Epoch [282/1000] Batch 0/23 Loss D: 1.4270336627960205, Loss G: 0.683728814125061\n", + "Epoch [283/1000] Batch 0/23 Loss D: 1.3157293200492859, Loss G: 0.7704383134841919\n", + "Epoch [284/1000] Batch 0/23 Loss D: 1.3112504482269287, Loss G: 0.8011195659637451\n", + "Epoch [285/1000] Batch 0/23 Loss D: 1.3752664923667908, Loss G: 0.7126722931861877\n", + "Epoch [285/1000] Real Acc: 71.35%, Fake Acc: 48.31%\n", + "Epoch [286/1000] Batch 0/23 Loss D: 1.3485132455825806, Loss G: 0.6747778654098511\n", + "Epoch [287/1000] Batch 0/23 Loss D: 1.394045114517212, Loss G: 0.6708738803863525\n", + "Epoch [288/1000] Batch 0/23 Loss D: 1.3657416105270386, Loss G: 0.7238250970840454\n", + "Epoch [289/1000] Batch 0/23 Loss D: 1.3235328197479248, Loss G: 0.762077808380127\n", + "Epoch [290/1000] Batch 0/23 Loss D: 1.3914575576782227, Loss G: 0.6930338740348816\n", + "Epoch [290/1000] Real Acc: 64.61%, Fake Acc: 46.63%\n", + "Epoch [291/1000] Batch 0/23 Loss D: 1.356391727924347, Loss G: 0.7165313959121704\n", + "Epoch [292/1000] Batch 0/23 Loss D: 1.3088539838790894, Loss G: 0.7854912877082825\n", + "Epoch [293/1000] Batch 0/23 Loss D: 1.295613408088684, Loss G: 0.8009235858917236\n", + "Epoch [294/1000] Batch 0/23 Loss D: 1.339880347251892, Loss G: 0.7536688446998596\n", + "Epoch [295/1000] Batch 0/23 Loss D: 1.2979359030723572, Loss G: 0.7457753419876099\n", + "Epoch [295/1000] Real Acc: 66.57%, Fake Acc: 40.17%\n", + "Epoch [296/1000] Batch 0/23 Loss D: 1.4111105799674988, Loss G: 0.7368536591529846\n", + "Epoch [297/1000] Batch 0/23 Loss D: 1.2939008474349976, Loss G: 0.7459163069725037\n", + "Epoch [298/1000] Batch 0/23 Loss D: 1.3324761390686035, Loss G: 0.7441473007202148\n", + "Epoch [299/1000] Batch 0/23 Loss D: 1.3119795322418213, Loss G: 0.7962045669555664\n", + "Epoch [300/1000] Batch 0/23 Loss D: 1.3261666297912598, Loss G: 0.7669582962989807\n", + "Epoch [300/1000] Real Acc: 47.47%, Fake Acc: 37.36%\n", + "Epoch [301/1000] Batch 0/23 Loss D: 1.4214507937431335, Loss G: 0.6994760036468506\n", + "Epoch [302/1000] Batch 0/23 Loss D: 1.2508520483970642, Loss G: 0.8084912300109863\n", + "Epoch [303/1000] Batch 0/23 Loss D: 1.2779242992401123, Loss G: 0.8254171013832092\n", + "Epoch [304/1000] Batch 0/23 Loss D: 1.3525198698043823, Loss G: 0.7731606960296631\n", + "Epoch [305/1000] Batch 0/23 Loss D: 1.3578516244888306, Loss G: 0.7032941579818726\n", + "Epoch [305/1000] Real Acc: 66.29%, Fake Acc: 59.55%\n", + "Epoch [306/1000] Batch 0/23 Loss D: 1.310894787311554, Loss G: 0.6663728356361389\n", + "Epoch [307/1000] Batch 0/23 Loss D: 1.2888153791427612, Loss G: 0.6947363615036011\n", + "Epoch [308/1000] Batch 0/23 Loss D: 1.4376325607299805, Loss G: 0.7225799560546875\n", + "Epoch [309/1000] Batch 0/23 Loss D: 1.2327757477760315, Loss G: 0.8945830464363098\n", + "Epoch [310/1000] Batch 0/23 Loss D: 1.3161768317222595, Loss G: 0.6416352987289429\n", + "Epoch [310/1000] Real Acc: 67.13%, Fake Acc: 77.81%\n", + "Epoch [311/1000] Batch 0/23 Loss D: 1.4153742790222168, Loss G: 0.8408330678939819\n", + "Epoch [312/1000] Batch 0/23 Loss D: 1.4068452715873718, Loss G: 0.9034563899040222\n", + "Epoch [313/1000] Batch 0/23 Loss D: 1.2867833971977234, Loss G: 0.8179123401641846\n", + "Epoch [314/1000] Batch 0/23 Loss D: 1.2880684733390808, Loss G: 0.7766455411911011\n", + "Epoch [315/1000] Batch 0/23 Loss D: 1.3962634801864624, Loss G: 0.6182116270065308\n", + "Epoch [315/1000] Real Acc: 75.00%, Fake Acc: 91.85%\n", + "Epoch [316/1000] Batch 0/23 Loss D: 1.4369109272956848, Loss G: 0.6151957511901855\n", + "Epoch [317/1000] Batch 0/23 Loss D: 1.3636146783828735, Loss G: 0.6818787455558777\n", + "Epoch [318/1000] Batch 0/23 Loss D: 1.298567771911621, Loss G: 0.7446627020835876\n", + "Epoch [319/1000] Batch 0/23 Loss D: 1.20333331823349, Loss G: 0.7857669591903687\n", + "Epoch [320/1000] Batch 0/23 Loss D: 1.2861254215240479, Loss G: 0.7790824770927429\n", + "Epoch [320/1000] Real Acc: 66.57%, Fake Acc: 16.85%\n", + "Epoch [321/1000] Batch 0/23 Loss D: 1.2562110424041748, Loss G: 0.7834905385971069\n", + "Epoch [322/1000] Batch 0/23 Loss D: 1.4311972856521606, Loss G: 0.6828222274780273\n", + "Epoch [323/1000] Batch 0/23 Loss D: 1.3527123928070068, Loss G: 0.7685633897781372\n", + "Epoch [324/1000] Batch 0/23 Loss D: 1.362745761871338, Loss G: 0.8335771560668945\n", + "Epoch [325/1000] Batch 0/23 Loss D: 1.2614567279815674, Loss G: 0.8228909969329834\n", + "Epoch [325/1000] Real Acc: 70.51%, Fake Acc: 24.16%\n", + "Epoch [326/1000] Batch 0/23 Loss D: 1.2806981801986694, Loss G: 0.7307409644126892\n", + "Epoch [327/1000] Batch 0/23 Loss D: 1.3081914186477661, Loss G: 0.7567379474639893\n", + "Epoch [328/1000] Batch 0/23 Loss D: 1.274626612663269, Loss G: 0.7532313466072083\n", + "Epoch [329/1000] Batch 0/23 Loss D: 1.3378113508224487, Loss G: 0.7384880781173706\n", + "Epoch [330/1000] Batch 0/23 Loss D: 1.358560562133789, Loss G: 0.7692633271217346\n", + "Epoch [330/1000] Real Acc: 65.45%, Fake Acc: 24.16%\n", + "Epoch [331/1000] Batch 0/23 Loss D: 1.3424810767173767, Loss G: 0.6997121572494507\n", + "Epoch [332/1000] Batch 0/23 Loss D: 1.363957405090332, Loss G: 0.8109839558601379\n", + "Epoch [333/1000] Batch 0/23 Loss D: 1.2627729177474976, Loss G: 0.8369511365890503\n", + "Epoch [334/1000] Batch 0/23 Loss D: 1.2497701048851013, Loss G: 0.8467363119125366\n", + "Epoch [335/1000] Batch 0/23 Loss D: 1.2288262248039246, Loss G: 0.8028345108032227\n", + "Epoch [335/1000] Real Acc: 80.06%, Fake Acc: 37.08%\n", + "Epoch [336/1000] Batch 0/23 Loss D: 1.2764519453048706, Loss G: 0.6818119287490845\n", + "Epoch [337/1000] Batch 0/23 Loss D: 1.3081479668617249, Loss G: 0.7347421646118164\n", + "Epoch [338/1000] Batch 0/23 Loss D: 1.2091394066810608, Loss G: 0.8403337001800537\n", + "Epoch [339/1000] Batch 0/23 Loss D: 1.171121060848236, Loss G: 0.8115042448043823\n", + "Epoch [340/1000] Batch 0/23 Loss D: 1.2556217908859253, Loss G: 0.8026614189147949\n", + "Epoch [340/1000] Real Acc: 60.11%, Fake Acc: 10.96%\n", + "Epoch [341/1000] Batch 0/23 Loss D: 1.352381706237793, Loss G: 0.9089492559432983\n", + "Epoch [342/1000] Batch 0/23 Loss D: 1.3179346919059753, Loss G: 0.7896602153778076\n", + "Epoch [343/1000] Batch 0/23 Loss D: 1.4864150881767273, Loss G: 0.8688857555389404\n", + "Epoch [344/1000] Batch 0/23 Loss D: 1.0868425369262695, Loss G: 0.9929030537605286\n", + "Epoch [345/1000] Batch 0/23 Loss D: 0.8406593501567841, Loss G: 0.9015517830848694\n", + "Epoch [345/1000] Real Acc: 97.75%, Fake Acc: 0.00%\n", + "Epoch [346/1000] Batch 0/23 Loss D: 0.6599875092506409, Loss G: 1.3837169408798218\n", + "Epoch [347/1000] Batch 0/23 Loss D: 0.45195671916007996, Loss G: 1.6828309297561646\n", + "Epoch [348/1000] Batch 0/23 Loss D: 0.2615986466407776, Loss G: 1.9399809837341309\n", + "Epoch [349/1000] Batch 0/23 Loss D: 0.1901330128312111, Loss G: 2.5106658935546875\n", + "Epoch [350/1000] Batch 0/23 Loss D: 0.11721460893750191, Loss G: 2.8737974166870117\n", + "Epoch [350/1000] Real Acc: 99.44%, Fake Acc: 0.00%\n", + "Epoch [351/1000] Batch 0/23 Loss D: 0.12774625048041344, Loss G: 3.073418378829956\n", + "Epoch [352/1000] Batch 0/23 Loss D: 0.10945721715688705, Loss G: 3.3865392208099365\n", + "Epoch [353/1000] Batch 0/23 Loss D: 0.18019407242536545, Loss G: 2.346719980239868\n", + "Epoch [354/1000] Batch 0/23 Loss D: 5.540263444185257, Loss G: 2.039374589920044\n", + "Epoch [355/1000] Batch 0/23 Loss D: 2.6611390113830566, Loss G: 0.5966695547103882\n", + "Epoch [355/1000] Real Acc: 30.62%, Fake Acc: 44.66%\n", + "Epoch [356/1000] Batch 0/23 Loss D: 1.3811465501785278, Loss G: 0.9092278480529785\n", + "Epoch [357/1000] Batch 0/23 Loss D: 1.4600948095321655, Loss G: 0.7026504874229431\n", + "Epoch [358/1000] Batch 0/23 Loss D: 1.3406109809875488, Loss G: 0.6505447626113892\n", + "Epoch [359/1000] Batch 0/23 Loss D: 1.1824104189872742, Loss G: 0.8602385520935059\n", + "Epoch [360/1000] Batch 0/23 Loss D: 1.1206841468811035, Loss G: 0.8674906492233276\n", + "Epoch [360/1000] Real Acc: 76.40%, Fake Acc: 4.78%\n", + "Epoch [361/1000] Batch 0/23 Loss D: 1.1788605451583862, Loss G: 0.8754480481147766\n", + "Epoch [362/1000] Batch 0/23 Loss D: 1.9273212552070618, Loss G: 0.9193512797355652\n", + "Epoch [363/1000] Batch 0/23 Loss D: 1.4654971957206726, Loss G: 0.6608273983001709\n", + "Epoch [364/1000] Batch 0/23 Loss D: 0.8757962584495544, Loss G: 1.2734451293945312\n", + "Epoch [365/1000] Batch 0/23 Loss D: 1.5091677904129028, Loss G: 1.0769860744476318\n", + "Epoch [365/1000] Real Acc: 33.43%, Fake Acc: 35.39%\n", + "Epoch [366/1000] Batch 0/23 Loss D: 1.657019853591919, Loss G: 0.4704629182815552\n", + "Epoch [367/1000] Batch 0/23 Loss D: 1.432002604007721, Loss G: 0.9653964042663574\n", + "Epoch [368/1000] Batch 0/23 Loss D: 1.3005805015563965, Loss G: 0.6715984344482422\n", + "Epoch [369/1000] Batch 0/23 Loss D: 1.3603782653808594, Loss G: 0.8127942085266113\n", + "Epoch [370/1000] Batch 0/23 Loss D: 1.3688434958457947, Loss G: 0.6848119497299194\n", + "Epoch [370/1000] Real Acc: 44.10%, Fake Acc: 24.72%\n", + "Epoch [371/1000] Batch 0/23 Loss D: 1.3328911066055298, Loss G: 0.7667127847671509\n", + "Epoch [372/1000] Batch 0/23 Loss D: 1.3347409963607788, Loss G: 0.8152565956115723\n", + "Epoch [373/1000] Batch 0/23 Loss D: 1.3598440885543823, Loss G: 0.7369388341903687\n", + "Epoch [374/1000] Batch 0/23 Loss D: 1.3551753759384155, Loss G: 0.7460918426513672\n", + "Epoch [375/1000] Batch 0/23 Loss D: 1.331753671169281, Loss G: 0.7577112913131714\n", + "Epoch [375/1000] Real Acc: 60.67%, Fake Acc: 20.51%\n", + "Epoch [376/1000] Batch 0/23 Loss D: 1.2603557705879211, Loss G: 0.8199979066848755\n", + "Epoch [377/1000] Batch 0/23 Loss D: 1.295685052871704, Loss G: 0.7899924516677856\n", + "Epoch [378/1000] Batch 0/23 Loss D: 1.3814122080802917, Loss G: 0.719822108745575\n", + "Epoch [379/1000] Batch 0/23 Loss D: 1.4341994524002075, Loss G: 0.7545137405395508\n", + "Epoch [380/1000] Batch 0/23 Loss D: 1.3911518454551697, Loss G: 0.684313178062439\n", + "Epoch [380/1000] Real Acc: 58.99%, Fake Acc: 58.15%\n", + "Epoch [381/1000] Batch 0/23 Loss D: 1.4136781692504883, Loss G: 0.7787370681762695\n", + "Epoch [382/1000] Batch 0/23 Loss D: 1.256838619709015, Loss G: 0.6814044117927551\n", + "Epoch [383/1000] Batch 0/23 Loss D: 1.1362226605415344, Loss G: 0.8655753135681152\n", + "Epoch [384/1000] Batch 0/23 Loss D: 1.1290709972381592, Loss G: 0.8368204236030579\n", + "Epoch [385/1000] Batch 0/23 Loss D: 1.2821193933486938, Loss G: 0.8325232267379761\n", + "Epoch [385/1000] Real Acc: 61.52%, Fake Acc: 21.63%\n", + "Epoch [386/1000] Batch 0/23 Loss D: 1.366173267364502, Loss G: 0.8255764245986938\n", + "Epoch [387/1000] Batch 0/23 Loss D: 1.3865147829055786, Loss G: 0.7674987316131592\n", + "Epoch [388/1000] Batch 0/23 Loss D: 1.2600998878479004, Loss G: 0.7661492824554443\n", + "Epoch [389/1000] Batch 0/23 Loss D: 1.3112045526504517, Loss G: 0.7910493612289429\n", + "Epoch [390/1000] Batch 0/23 Loss D: 1.2730742692947388, Loss G: 0.836269736289978\n", + "Epoch [390/1000] Real Acc: 72.19%, Fake Acc: 16.01%\n", + "Epoch [391/1000] Batch 0/23 Loss D: 1.162348747253418, Loss G: 0.8483536243438721\n", + "Epoch [392/1000] Batch 0/23 Loss D: 1.247029721736908, Loss G: 0.9071640968322754\n", + "Epoch [393/1000] Batch 0/23 Loss D: 1.3250564336776733, Loss G: 0.7809126377105713\n", + "Epoch [394/1000] Batch 0/23 Loss D: 1.2396668195724487, Loss G: 0.8573703169822693\n", + "Epoch [395/1000] Batch 0/23 Loss D: 1.2780097126960754, Loss G: 0.8579076528549194\n", + "Epoch [395/1000] Real Acc: 71.35%, Fake Acc: 20.79%\n", + "Epoch [396/1000] Batch 0/23 Loss D: 1.2734202146530151, Loss G: 0.8070710897445679\n", + "Epoch [397/1000] Batch 0/23 Loss D: 1.3492753505706787, Loss G: 0.8487285375595093\n", + "Epoch [398/1000] Batch 0/23 Loss D: 1.5039272904396057, Loss G: 0.6899596452713013\n", + "Epoch [399/1000] Batch 0/23 Loss D: 1.3094558715820312, Loss G: 1.031321406364441\n", + "Epoch [400/1000] Batch 0/23 Loss D: 0.9797741770744324, Loss G: 0.9351087212562561\n", + "Epoch [400/1000] Real Acc: 85.67%, Fake Acc: 12.08%\n", + "Epoch [401/1000] Batch 0/23 Loss D: 1.20387601852417, Loss G: 0.7705533504486084\n", + "Epoch [402/1000] Batch 0/23 Loss D: 1.6382405161857605, Loss G: 0.6999721527099609\n", + "Epoch [403/1000] Batch 0/23 Loss D: 1.561136245727539, Loss G: 0.8496423959732056\n", + "Epoch [404/1000] Batch 0/23 Loss D: 1.0840027928352356, Loss G: 0.9115134477615356\n", + "Epoch [405/1000] Batch 0/23 Loss D: 1.2510522603988647, Loss G: 0.8895469903945923\n", + "Epoch [405/1000] Real Acc: 53.65%, Fake Acc: 14.04%\n", + "Epoch [406/1000] Batch 0/23 Loss D: 1.278551697731018, Loss G: 0.8093497157096863\n", + "Epoch [407/1000] Batch 0/23 Loss D: 1.3472267985343933, Loss G: 0.9026796817779541\n", + "Epoch [408/1000] Batch 0/23 Loss D: 1.3231485486030579, Loss G: 0.7732565402984619\n", + "Epoch [409/1000] Batch 0/23 Loss D: 1.1535760760307312, Loss G: 1.0017199516296387\n", + "Epoch [410/1000] Batch 0/23 Loss D: 0.9944570660591125, Loss G: 0.7664782404899597\n", + "Epoch [410/1000] Real Acc: 70.22%, Fake Acc: 12.92%\n", + "Epoch [411/1000] Batch 0/23 Loss D: 1.0754416584968567, Loss G: 1.006103277206421\n", + "Epoch [412/1000] Batch 0/23 Loss D: 1.4624862670898438, Loss G: 0.7418809533119202\n", + "Epoch [413/1000] Batch 0/23 Loss D: 1.185081958770752, Loss G: 0.8577172160148621\n", + "Epoch [414/1000] Batch 0/23 Loss D: 1.2195565104484558, Loss G: 1.0890872478485107\n", + "Epoch [415/1000] Batch 0/23 Loss D: 1.1234807968139648, Loss G: 1.0721936225891113\n", + "Epoch [415/1000] Real Acc: 64.61%, Fake Acc: 10.39%\n", + "Epoch [416/1000] Batch 0/23 Loss D: 1.3129889965057373, Loss G: 0.784188985824585\n", + "Epoch [417/1000] Batch 0/23 Loss D: 1.2912843227386475, Loss G: 0.7815068960189819\n", + "Epoch [418/1000] Batch 0/23 Loss D: 1.366939663887024, Loss G: 0.7386006116867065\n", + "Epoch [419/1000] Batch 0/23 Loss D: 1.150592565536499, Loss G: 0.8540475368499756\n", + "Epoch [420/1000] Batch 0/23 Loss D: 1.1070425510406494, Loss G: 0.8077794313430786\n", + "Epoch [420/1000] Real Acc: 75.00%, Fake Acc: 14.61%\n", + "Epoch [421/1000] Batch 0/23 Loss D: 1.2771663069725037, Loss G: 0.8861724138259888\n", + "Epoch [422/1000] Batch 0/23 Loss D: 1.281216025352478, Loss G: 0.9013100862503052\n", + "Epoch [423/1000] Batch 0/23 Loss D: 1.2646365761756897, Loss G: 0.7997046709060669\n", + "Epoch [424/1000] Batch 0/23 Loss D: 1.259267508983612, Loss G: 0.7673053741455078\n", + "Epoch [425/1000] Batch 0/23 Loss D: 1.3174649477005005, Loss G: 0.8132942318916321\n", + "Epoch [425/1000] Real Acc: 61.52%, Fake Acc: 23.60%\n", + "Epoch [426/1000] Batch 0/23 Loss D: 1.168865978717804, Loss G: 0.8112320899963379\n", + "Epoch [427/1000] Batch 0/23 Loss D: 1.107279121875763, Loss G: 0.8777565360069275\n", + "Epoch [428/1000] Batch 0/23 Loss D: 1.0606170892715454, Loss G: 0.8930749893188477\n", + "Epoch [429/1000] Batch 0/23 Loss D: 1.0848719477653503, Loss G: 0.9650166034698486\n", + "Epoch [430/1000] Batch 0/23 Loss D: 1.3222037553787231, Loss G: 0.9611206650733948\n", + "Epoch [430/1000] Real Acc: 54.49%, Fake Acc: 39.89%\n", + "Epoch [431/1000] Batch 0/23 Loss D: 1.365467131137848, Loss G: 0.8602046966552734\n", + "Epoch [432/1000] Batch 0/23 Loss D: 1.1631988286972046, Loss G: 1.084519624710083\n", + "Epoch [433/1000] Batch 0/23 Loss D: 0.9708845317363739, Loss G: 1.0877723693847656\n", + "Epoch [434/1000] Batch 0/23 Loss D: 1.0876606106758118, Loss G: 0.7631200551986694\n", + "Epoch [435/1000] Batch 0/23 Loss D: 1.0344958901405334, Loss G: 0.8547250628471375\n", + "Epoch [435/1000] Real Acc: 98.88%, Fake Acc: 0.00%\n", + "Epoch [436/1000] Batch 0/23 Loss D: 0.5920694172382355, Loss G: 1.3238489627838135\n", + "Epoch [437/1000] Batch 0/23 Loss D: 0.3873443007469177, Loss G: 1.9076546430587769\n", + "Epoch [438/1000] Batch 0/23 Loss D: 1.2906304001808167, Loss G: 0.9767224192619324\n", + "Epoch [439/1000] Batch 0/23 Loss D: 1.728247880935669, Loss G: 0.41026726365089417\n", + "Epoch [440/1000] Batch 0/23 Loss D: 1.4512644112110138, Loss G: 1.2177095413208008\n", + "Epoch [440/1000] Real Acc: 47.47%, Fake Acc: 29.49%\n", + "Epoch [441/1000] Batch 0/23 Loss D: 1.228048026561737, Loss G: 0.7522767186164856\n", + "Epoch [442/1000] Batch 0/23 Loss D: 1.260136902332306, Loss G: 1.0656487941741943\n", + "Epoch [443/1000] Batch 0/23 Loss D: 1.421867549419403, Loss G: 0.5304980278015137\n", + "Epoch [444/1000] Batch 0/23 Loss D: 1.2645386457443237, Loss G: 0.9255527257919312\n", + "Epoch [445/1000] Batch 0/23 Loss D: 1.0954344272613525, Loss G: 0.8573797941207886\n", + "Epoch [445/1000] Real Acc: 80.90%, Fake Acc: 23.03%\n", + "Epoch [446/1000] Batch 0/23 Loss D: 1.1177095770835876, Loss G: 0.8744339346885681\n", + "Epoch [447/1000] Batch 0/23 Loss D: 1.0181353688240051, Loss G: 0.9546855092048645\n", + "Epoch [448/1000] Batch 0/23 Loss D: 1.2210609316825867, Loss G: 0.888063371181488\n", + "Epoch [449/1000] Batch 0/23 Loss D: 1.1406387090682983, Loss G: 0.8644047975540161\n", + "Epoch [450/1000] Batch 0/23 Loss D: 1.5136751532554626, Loss G: 0.8990533351898193\n", + "Epoch [450/1000] Real Acc: 51.40%, Fake Acc: 12.92%\n", + "Epoch [451/1000] Batch 0/23 Loss D: 1.1139764189720154, Loss G: 1.0710115432739258\n", + "Epoch [452/1000] Batch 0/23 Loss D: 0.9158174097537994, Loss G: 1.0976462364196777\n", + "Epoch [453/1000] Batch 0/23 Loss D: 1.0174291729927063, Loss G: 1.0780203342437744\n", + "Epoch [454/1000] Batch 0/23 Loss D: 1.1716626286506653, Loss G: 0.7104790210723877\n", + "Epoch [455/1000] Batch 0/23 Loss D: 1.399678885936737, Loss G: 0.773475706577301\n", + "Epoch [455/1000] Real Acc: 58.99%, Fake Acc: 23.03%\n", + "Epoch [456/1000] Batch 0/23 Loss D: 1.25884211063385, Loss G: 0.7799403071403503\n", + "Epoch [457/1000] Batch 0/23 Loss D: 1.0788463950157166, Loss G: 0.9289555549621582\n", + "Epoch [458/1000] Batch 0/23 Loss D: 1.0595552325248718, Loss G: 1.0237176418304443\n", + "Epoch [459/1000] Batch 0/23 Loss D: 1.2509711384773254, Loss G: 0.917733907699585\n", + "Epoch [460/1000] Batch 0/23 Loss D: 1.387130856513977, Loss G: 0.9984040856361389\n", + "Epoch [460/1000] Real Acc: 60.11%, Fake Acc: 17.42%\n", + "Epoch [461/1000] Batch 0/23 Loss D: 1.265813171863556, Loss G: 0.8879887461662292\n", + "Epoch [462/1000] Batch 0/23 Loss D: 1.0094071626663208, Loss G: 1.0902113914489746\n", + "Epoch [463/1000] Batch 0/23 Loss D: 0.9788961708545685, Loss G: 1.1445553302764893\n", + "Epoch [464/1000] Batch 0/23 Loss D: 1.0460201501846313, Loss G: 1.0568711757659912\n", + "Epoch [465/1000] Batch 0/23 Loss D: 1.3314964771270752, Loss G: 0.7670637369155884\n", + "Epoch [465/1000] Real Acc: 72.47%, Fake Acc: 54.49%\n", + "Epoch [466/1000] Batch 0/23 Loss D: 1.330409824848175, Loss G: 0.8221986293792725\n", + "Epoch [467/1000] Batch 0/23 Loss D: 0.8026895523071289, Loss G: 1.135987639427185\n", + "Epoch [468/1000] Batch 0/23 Loss D: 0.8229768872261047, Loss G: 0.9653142690658569\n", + "Epoch [469/1000] Batch 0/23 Loss D: 0.9729407131671906, Loss G: 1.3714710474014282\n", + "Epoch [470/1000] Batch 0/23 Loss D: 1.276807427406311, Loss G: 0.6765173673629761\n", + "Epoch [470/1000] Real Acc: 56.46%, Fake Acc: 28.09%\n", + "Epoch [471/1000] Batch 0/23 Loss D: 1.2839890718460083, Loss G: 1.0330133438110352\n", + "Epoch [472/1000] Batch 0/23 Loss D: 1.2812111377716064, Loss G: 0.9248455762863159\n", + "Epoch [473/1000] Batch 0/23 Loss D: 1.162864625453949, Loss G: 1.0080493688583374\n", + "Epoch [474/1000] Batch 0/23 Loss D: 1.0879143476486206, Loss G: 1.00362229347229\n", + "Epoch [475/1000] Batch 0/23 Loss D: 0.9017760753631592, Loss G: 1.3030459880828857\n", + "Epoch [475/1000] Real Acc: 92.42%, Fake Acc: 8.15%\n", + "Epoch [476/1000] Batch 0/23 Loss D: 0.9668020606040955, Loss G: 0.8468431830406189\n", + "Epoch [477/1000] Batch 0/23 Loss D: 0.9625434875488281, Loss G: 0.9302283525466919\n", + "Epoch [478/1000] Batch 0/23 Loss D: 1.162643849849701, Loss G: 0.9057276248931885\n", + "Epoch [479/1000] Batch 0/23 Loss D: 1.1874220073223114, Loss G: 0.91496342420578\n", + "Epoch [480/1000] Batch 0/23 Loss D: 1.1754281520843506, Loss G: 1.0420626401901245\n", + "Epoch [480/1000] Real Acc: 55.34%, Fake Acc: 13.76%\n", + "Epoch [481/1000] Batch 0/23 Loss D: 1.358399510383606, Loss G: 0.9548121690750122\n", + "Epoch [482/1000] Batch 0/23 Loss D: 1.3396866917610168, Loss G: 1.1311930418014526\n", + "Epoch [483/1000] Batch 0/23 Loss D: 1.1191985607147217, Loss G: 0.9543262720108032\n", + "Epoch [484/1000] Batch 0/23 Loss D: 0.9591310620307922, Loss G: 1.3152275085449219\n", + "Epoch [485/1000] Batch 0/23 Loss D: 1.0177749395370483, Loss G: 1.0224469900131226\n", + "Epoch [485/1000] Real Acc: 83.71%, Fake Acc: 13.20%\n", + "Epoch [486/1000] Batch 0/23 Loss D: 1.1464133858680725, Loss G: 0.9472783803939819\n", + "Epoch [487/1000] Batch 0/23 Loss D: 1.2749159336090088, Loss G: 0.6958783864974976\n", + "Epoch [488/1000] Batch 0/23 Loss D: 1.1002248525619507, Loss G: 0.9002771377563477\n", + "Epoch [489/1000] Batch 0/23 Loss D: 1.1084023118019104, Loss G: 0.9247973561286926\n", + "Epoch [490/1000] Batch 0/23 Loss D: 1.0141696333885193, Loss G: 0.9191477298736572\n", + "Epoch [490/1000] Real Acc: 71.07%, Fake Acc: 4.21%\n", + "Epoch [491/1000] Batch 0/23 Loss D: 0.8091048002243042, Loss G: 1.105660080909729\n", + "Epoch [492/1000] Batch 0/23 Loss D: 0.9420835971832275, Loss G: 0.9029860496520996\n", + "Epoch [493/1000] Batch 0/23 Loss D: 1.277845859527588, Loss G: 1.0856034755706787\n", + "Epoch [494/1000] Batch 0/23 Loss D: 1.1741573214530945, Loss G: 0.9698554277420044\n", + "Epoch [495/1000] Batch 0/23 Loss D: 0.9498358964920044, Loss G: 1.339543342590332\n", + "Epoch [495/1000] Real Acc: 75.84%, Fake Acc: 0.56%\n", + "Epoch [496/1000] Batch 0/23 Loss D: 0.8526542782783508, Loss G: 1.1228290796279907\n", + "Epoch [497/1000] Batch 0/23 Loss D: 0.9144222736358643, Loss G: 1.135936975479126\n", + "Epoch [498/1000] Batch 0/23 Loss D: 1.2216740846633911, Loss G: 0.7513027191162109\n", + "Epoch [499/1000] Batch 0/23 Loss D: 1.1786751747131348, Loss G: 0.8993327021598816\n", + "Epoch [500/1000] Batch 0/23 Loss D: 0.96193528175354, Loss G: 1.0376603603363037\n", + "Epoch [500/1000] Real Acc: 88.76%, Fake Acc: 1.69%\n", + "Epoch [501/1000] Batch 0/23 Loss D: 0.8398411870002747, Loss G: 1.0996264219284058\n", + "Epoch [502/1000] Batch 0/23 Loss D: 0.9127479195594788, Loss G: 1.1277613639831543\n", + "Epoch [503/1000] Batch 0/23 Loss D: 1.3189143240451813, Loss G: 1.1191145181655884\n", + "Epoch [504/1000] Batch 0/23 Loss D: 1.3004080057144165, Loss G: 1.0063133239746094\n", + "Epoch [505/1000] Batch 0/23 Loss D: 0.8124498128890991, Loss G: 1.4017832279205322\n", + "Epoch [505/1000] Real Acc: 87.08%, Fake Acc: 2.81%\n", + "Epoch [506/1000] Batch 0/23 Loss D: 0.827344685792923, Loss G: 1.383245587348938\n", + "Epoch [507/1000] Batch 0/23 Loss D: 0.9692409634590149, Loss G: 1.1490839719772339\n", + "Epoch [508/1000] Batch 0/23 Loss D: 1.381986379623413, Loss G: 0.7167010307312012\n", + "Epoch [509/1000] Batch 0/23 Loss D: 1.351906657218933, Loss G: 0.6972231864929199\n", + "Epoch [510/1000] Batch 0/23 Loss D: 0.9888405799865723, Loss G: 1.016531229019165\n", + "Epoch [510/1000] Real Acc: 80.90%, Fake Acc: 3.65%\n", + "Epoch [511/1000] Batch 0/23 Loss D: 0.8512818813323975, Loss G: 1.084357500076294\n", + "Epoch [512/1000] Batch 0/23 Loss D: 0.9066343903541565, Loss G: 1.1589806079864502\n", + "Epoch [513/1000] Batch 0/23 Loss D: 1.2294602990150452, Loss G: 0.8922837972640991\n", + "Epoch [514/1000] Batch 0/23 Loss D: 1.136206328868866, Loss G: 0.9287998676300049\n", + "Epoch [515/1000] Batch 0/23 Loss D: 1.1613564491271973, Loss G: 0.9486202001571655\n", + "Epoch [515/1000] Real Acc: 66.57%, Fake Acc: 6.18%\n", + "Epoch [516/1000] Batch 0/23 Loss D: 1.046595811843872, Loss G: 1.1357020139694214\n", + "Epoch [517/1000] Batch 0/23 Loss D: 0.895269513130188, Loss G: 1.0263099670410156\n", + "Epoch [518/1000] Batch 0/23 Loss D: 0.854153960943222, Loss G: 1.0511419773101807\n", + "Epoch [519/1000] Batch 0/23 Loss D: 1.09624582529068, Loss G: 0.8383673429489136\n", + "Epoch [520/1000] Batch 0/23 Loss D: 1.2033076882362366, Loss G: 0.9303722381591797\n", + "Epoch [520/1000] Real Acc: 91.85%, Fake Acc: 12.08%\n", + "Epoch [521/1000] Batch 0/23 Loss D: 0.7682222127914429, Loss G: 1.1101475954055786\n", + "Epoch [522/1000] Batch 0/23 Loss D: 0.7903246283531189, Loss G: 1.5575205087661743\n", + "Epoch [523/1000] Batch 0/23 Loss D: 0.8473615944385529, Loss G: 1.4123138189315796\n", + "Epoch [524/1000] Batch 0/23 Loss D: 1.0549888610839844, Loss G: 1.0058766603469849\n", + "Epoch [525/1000] Batch 0/23 Loss D: 1.0763972997665405, Loss G: 1.1479051113128662\n", + "Epoch [525/1000] Real Acc: 67.42%, Fake Acc: 23.60%\n", + "Epoch [526/1000] Batch 0/23 Loss D: 0.9940019845962524, Loss G: 1.3726279735565186\n", + "Epoch [527/1000] Batch 0/23 Loss D: 0.7920177578926086, Loss G: 1.3562382459640503\n", + "Epoch [528/1000] Batch 0/23 Loss D: 0.9098484814167023, Loss G: 1.1452527046203613\n", + "Epoch [529/1000] Batch 0/23 Loss D: 1.0238491296768188, Loss G: 0.861614465713501\n", + "Epoch [530/1000] Batch 0/23 Loss D: 0.8432283103466034, Loss G: 1.0478837490081787\n", + "Epoch [530/1000] Real Acc: 84.83%, Fake Acc: 6.74%\n", + "Epoch [531/1000] Batch 0/23 Loss D: 0.8502391576766968, Loss G: 1.270294427871704\n", + "Epoch [532/1000] Batch 0/23 Loss D: 0.7793549001216888, Loss G: 1.234632968902588\n", + "Epoch [533/1000] Batch 0/23 Loss D: 0.6083102226257324, Loss G: 1.5051263570785522\n", + "Epoch [534/1000] Batch 0/23 Loss D: 1.05707448720932, Loss G: 1.1760027408599854\n", + "Epoch [535/1000] Batch 0/23 Loss D: 1.21476548910141, Loss G: 1.2839593887329102\n", + "Epoch [535/1000] Real Acc: 58.71%, Fake Acc: 33.15%\n", + "Epoch [536/1000] Batch 0/23 Loss D: 1.0009171962738037, Loss G: 1.2492523193359375\n", + "Epoch [537/1000] Batch 0/23 Loss D: 0.899721086025238, Loss G: 1.299705147743225\n", + "Epoch [538/1000] Batch 0/23 Loss D: 0.7263862788677216, Loss G: 1.3483233451843262\n", + "Epoch [539/1000] Batch 0/23 Loss D: 0.8326920568943024, Loss G: 1.1883724927902222\n", + "Epoch [540/1000] Batch 0/23 Loss D: 1.26783287525177, Loss G: 0.8542253971099854\n", + "Epoch [540/1000] Real Acc: 76.40%, Fake Acc: 17.42%\n", + "Epoch [541/1000] Batch 0/23 Loss D: 0.9796928465366364, Loss G: 0.9668920040130615\n", + "Epoch [542/1000] Batch 0/23 Loss D: 0.8427854776382446, Loss G: 1.1145529747009277\n", + "Epoch [543/1000] Batch 0/23 Loss D: 0.917818546295166, Loss G: 1.4526606798171997\n", + "Epoch [544/1000] Batch 0/23 Loss D: 1.2132918238639832, Loss G: 1.1038093566894531\n", + "Epoch [545/1000] Batch 0/23 Loss D: 1.30362069606781, Loss G: 0.9951394200325012\n", + "Epoch [545/1000] Real Acc: 57.87%, Fake Acc: 14.04%\n", + "Epoch [546/1000] Batch 0/23 Loss D: 1.0301354229450226, Loss G: 1.1759815216064453\n", + "Epoch [547/1000] Batch 0/23 Loss D: 0.7074394226074219, Loss G: 1.499385118484497\n", + "Epoch [548/1000] Batch 0/23 Loss D: 0.6802792251110077, Loss G: 1.4506016969680786\n", + "Epoch [549/1000] Batch 0/23 Loss D: 0.8536123633384705, Loss G: 1.0345278978347778\n", + "Epoch [550/1000] Batch 0/23 Loss D: 0.8390980958938599, Loss G: 0.9676164984703064\n", + "Epoch [550/1000] Real Acc: 87.92%, Fake Acc: 5.90%\n", + "Epoch [551/1000] Batch 0/23 Loss D: 0.8567375838756561, Loss G: 1.2562099695205688\n", + "Epoch [552/1000] Batch 0/23 Loss D: 0.7144043743610382, Loss G: 1.2866206169128418\n", + "Epoch [553/1000] Batch 0/23 Loss D: 0.9831603765487671, Loss G: 1.1864463090896606\n", + "Epoch [554/1000] Batch 0/23 Loss D: 1.4158393144607544, Loss G: 1.1492021083831787\n", + "Epoch [555/1000] Batch 0/23 Loss D: 0.8303086757659912, Loss G: 1.4137241840362549\n", + "Epoch [555/1000] Real Acc: 73.31%, Fake Acc: 8.15%\n", + "Epoch [556/1000] Batch 0/23 Loss D: 0.9895890653133392, Loss G: 1.3444364070892334\n", + "Epoch [557/1000] Batch 0/23 Loss D: 0.8140131533145905, Loss G: 1.2868233919143677\n", + "Epoch [558/1000] Batch 0/23 Loss D: 0.6904152631759644, Loss G: 1.531409502029419\n", + "Epoch [559/1000] Batch 0/23 Loss D: 0.7611165642738342, Loss G: 1.1771949529647827\n", + "Epoch [560/1000] Batch 0/23 Loss D: 0.7733664512634277, Loss G: 1.1706809997558594\n", + "Epoch [560/1000] Real Acc: 91.29%, Fake Acc: 8.71%\n", + "Epoch [561/1000] Batch 0/23 Loss D: 0.9163267612457275, Loss G: 1.1867077350616455\n", + "Epoch [562/1000] Batch 0/23 Loss D: 0.786279171705246, Loss G: 1.1529465913772583\n", + "Epoch [563/1000] Batch 0/23 Loss D: 0.6995255351066589, Loss G: 1.451488971710205\n", + "Epoch [564/1000] Batch 0/23 Loss D: 0.6066789478063583, Loss G: 1.663501262664795\n", + "Epoch [565/1000] Batch 0/23 Loss D: 0.7337348461151123, Loss G: 1.3290562629699707\n", + "Epoch [565/1000] Real Acc: 66.29%, Fake Acc: 10.96%\n", + "Epoch [566/1000] Batch 0/23 Loss D: 1.2217023372650146, Loss G: 0.9173364043235779\n", + "Epoch [567/1000] Batch 0/23 Loss D: 1.1192271411418915, Loss G: 1.3516664505004883\n", + "Epoch [568/1000] Batch 0/23 Loss D: 0.648033082485199, Loss G: 1.5992666482925415\n", + "Epoch [569/1000] Batch 0/23 Loss D: 0.6062504649162292, Loss G: 1.7530279159545898\n", + "Epoch [570/1000] Batch 0/23 Loss D: 0.6433566212654114, Loss G: 1.3978195190429688\n", + "Epoch [570/1000] Real Acc: 85.39%, Fake Acc: 7.87%\n", + "Epoch [571/1000] Batch 0/23 Loss D: 0.9266109466552734, Loss G: 1.31773841381073\n", + "Epoch [572/1000] Batch 0/23 Loss D: 0.9332485795021057, Loss G: 1.0625450611114502\n", + "Epoch [573/1000] Batch 0/23 Loss D: 0.8571532368659973, Loss G: 0.9684953689575195\n", + "Epoch [574/1000] Batch 0/23 Loss D: 0.6828047633171082, Loss G: 1.2882194519042969\n", + "Epoch [575/1000] Batch 0/23 Loss D: 0.465501606464386, Loss G: 1.5250849723815918\n", + "Epoch [575/1000] Real Acc: 82.02%, Fake Acc: 0.56%\n", + "Epoch [576/1000] Batch 0/23 Loss D: 0.7590071856975555, Loss G: 1.644336223602295\n", + "Epoch [577/1000] Batch 0/23 Loss D: 0.6488112807273865, Loss G: 1.6278586387634277\n", + "Epoch [578/1000] Batch 0/23 Loss D: 0.9610456228256226, Loss G: 1.172910213470459\n", + "Epoch [579/1000] Batch 0/23 Loss D: 0.7875046133995056, Loss G: 1.6723692417144775\n", + "Epoch [580/1000] Batch 0/23 Loss D: 0.6704318225383759, Loss G: 1.6290020942687988\n", + "Epoch [580/1000] Real Acc: 95.79%, Fake Acc: 1.69%\n", + "Epoch [581/1000] Batch 0/23 Loss D: 0.584953248500824, Loss G: 1.9524844884872437\n", + "Epoch [582/1000] Batch 0/23 Loss D: 0.7293239533901215, Loss G: 1.4461321830749512\n", + "Epoch [583/1000] Batch 0/23 Loss D: 0.9100567996501923, Loss G: 1.1282732486724854\n", + "Epoch [584/1000] Batch 0/23 Loss D: 1.0879571437835693, Loss G: 0.9308192133903503\n", + "Epoch [585/1000] Batch 0/23 Loss D: 0.6393142342567444, Loss G: 1.492064356803894\n", + "Epoch [585/1000] Real Acc: 98.31%, Fake Acc: 0.00%\n", + "Epoch [586/1000] Batch 0/23 Loss D: 0.47181840240955353, Loss G: 1.8473970890045166\n", + "Epoch [587/1000] Batch 0/23 Loss D: 0.7433572113513947, Loss G: 2.171815872192383\n", + "Epoch [588/1000] Batch 0/23 Loss D: 1.2099753618240356, Loss G: 1.7080864906311035\n", + "Epoch [589/1000] Batch 0/23 Loss D: 0.9833909869194031, Loss G: 1.3592580556869507\n", + "Epoch [590/1000] Batch 0/23 Loss D: 0.6448595225811005, Loss G: 1.9290361404418945\n", + "Epoch [590/1000] Real Acc: 80.62%, Fake Acc: 5.34%\n", + "Epoch [591/1000] Batch 0/23 Loss D: 0.6428126990795135, Loss G: 1.3631060123443604\n", + "Epoch [592/1000] Batch 0/23 Loss D: 0.5963557958602905, Loss G: 1.5087494850158691\n", + "Epoch [593/1000] Batch 0/23 Loss D: 0.8823137283325195, Loss G: 1.0898923873901367\n", + "Epoch [594/1000] Batch 0/23 Loss D: 0.8664138317108154, Loss G: 0.963871955871582\n", + "Epoch [595/1000] Batch 0/23 Loss D: 0.7429839372634888, Loss G: 1.2912752628326416\n", + "Epoch [595/1000] Real Acc: 87.36%, Fake Acc: 2.25%\n", + "Epoch [596/1000] Batch 0/23 Loss D: 0.6443332731723785, Loss G: 1.270743489265442\n", + "Epoch [597/1000] Batch 0/23 Loss D: 0.7205969989299774, Loss G: 1.8037986755371094\n", + "Epoch [598/1000] Batch 0/23 Loss D: 0.6905708909034729, Loss G: 1.435646414756775\n", + "Epoch [599/1000] Batch 0/23 Loss D: 0.923624575138092, Loss G: 1.3067071437835693\n", + "Epoch [600/1000] Batch 0/23 Loss D: 1.0034796595573425, Loss G: 1.345103144645691\n", + "Epoch [600/1000] Real Acc: 85.67%, Fake Acc: 8.99%\n", + "Epoch [601/1000] Batch 0/23 Loss D: 0.5592558979988098, Loss G: 1.7526612281799316\n", + "Epoch [602/1000] Batch 0/23 Loss D: 0.6019476652145386, Loss G: 1.7992433309555054\n", + "Epoch [603/1000] Batch 0/23 Loss D: 0.5446206033229828, Loss G: 1.4333922863006592\n", + "Epoch [604/1000] Batch 0/23 Loss D: 0.6724929809570312, Loss G: 1.4425253868103027\n", + "Epoch [605/1000] Batch 0/23 Loss D: 0.8339780569076538, Loss G: 1.1659283638000488\n", + "Epoch [605/1000] Real Acc: 91.01%, Fake Acc: 10.67%\n", + "Epoch [606/1000] Batch 0/23 Loss D: 0.7369318902492523, Loss G: 1.2196327447891235\n", + "Epoch [607/1000] Batch 0/23 Loss D: 0.6135860085487366, Loss G: 1.5305159091949463\n", + "Epoch [608/1000] Batch 0/23 Loss D: 0.48384763300418854, Loss G: 1.5334725379943848\n", + "Epoch [609/1000] Batch 0/23 Loss D: 0.5851092636585236, Loss G: 1.701306939125061\n", + "Epoch [610/1000] Batch 0/23 Loss D: 0.8001765608787537, Loss G: 1.2714124917984009\n", + "Epoch [610/1000] Real Acc: 79.49%, Fake Acc: 14.33%\n", + "Epoch [611/1000] Batch 0/23 Loss D: 0.8686279654502869, Loss G: 1.6079680919647217\n", + "Epoch [612/1000] Batch 0/23 Loss D: 0.9254225492477417, Loss G: 1.3769323825836182\n", + "Epoch [613/1000] Batch 0/23 Loss D: 0.6632502675056458, Loss G: 1.7352259159088135\n", + "Epoch [614/1000] Batch 0/23 Loss D: 0.42070770263671875, Loss G: 2.104358673095703\n", + "Epoch [615/1000] Batch 0/23 Loss D: 0.5198249071836472, Loss G: 1.5741581916809082\n", + "Epoch [615/1000] Real Acc: 94.66%, Fake Acc: 4.21%\n", + "Epoch [616/1000] Batch 0/23 Loss D: 0.4678592383861542, Loss G: 1.784447431564331\n", + "Epoch [617/1000] Batch 0/23 Loss D: 0.7438032925128937, Loss G: 1.1954259872436523\n", + "Epoch [618/1000] Batch 0/23 Loss D: 0.8836023211479187, Loss G: 1.093039631843567\n", + "Epoch [619/1000] Batch 0/23 Loss D: 0.6692802309989929, Loss G: 1.3944647312164307\n", + "Epoch [620/1000] Batch 0/23 Loss D: 0.49361804127693176, Loss G: 1.76767897605896\n", + "Epoch [620/1000] Real Acc: 92.70%, Fake Acc: 0.00%\n", + "Epoch [621/1000] Batch 0/23 Loss D: 0.4496869146823883, Loss G: 1.9203524589538574\n", + "Epoch [622/1000] Batch 0/23 Loss D: 0.5445290058851242, Loss G: 1.8525354862213135\n", + "Epoch [623/1000] Batch 0/23 Loss D: 1.0467582643032074, Loss G: 1.574391484260559\n", + "Epoch [624/1000] Batch 0/23 Loss D: 0.925912469625473, Loss G: 1.2842776775360107\n", + "Epoch [625/1000] Batch 0/23 Loss D: 0.675659716129303, Loss G: 2.346728563308716\n", + "Epoch [625/1000] Real Acc: 92.42%, Fake Acc: 2.53%\n", + "Epoch [626/1000] Batch 0/23 Loss D: 0.5326738655567169, Loss G: 1.4802238941192627\n", + "Epoch [627/1000] Batch 0/23 Loss D: 0.48331043124198914, Loss G: 1.8977540731430054\n", + "Epoch [628/1000] Batch 0/23 Loss D: 0.6718797087669373, Loss G: 1.223353385925293\n", + "Epoch [629/1000] Batch 0/23 Loss D: 0.9016595482826233, Loss G: 1.0866467952728271\n", + "Epoch [630/1000] Batch 0/23 Loss D: 0.7563532590866089, Loss G: 1.164564847946167\n", + "Epoch [630/1000] Real Acc: 93.26%, Fake Acc: 2.53%\n", + "Epoch [631/1000] Batch 0/23 Loss D: 0.5883781313896179, Loss G: 1.7277402877807617\n", + "Epoch [632/1000] Batch 0/23 Loss D: 0.4498717486858368, Loss G: 1.4269880056381226\n", + "Epoch [633/1000] Batch 0/23 Loss D: 0.40355993807315826, Loss G: 2.090630531311035\n", + "Epoch [634/1000] Batch 0/23 Loss D: 0.8329779803752899, Loss G: 1.1771671772003174\n", + "Epoch [635/1000] Batch 0/23 Loss D: 0.6922571063041687, Loss G: 1.666759729385376\n", + "Epoch [635/1000] Real Acc: 80.06%, Fake Acc: 9.83%\n", + "Epoch [636/1000] Batch 0/23 Loss D: 0.747355043888092, Loss G: 1.5961406230926514\n", + "Epoch [637/1000] Batch 0/23 Loss D: 0.6246569454669952, Loss G: 1.8203967809677124\n", + "Epoch [638/1000] Batch 0/23 Loss D: 0.6380771398544312, Loss G: 1.7720553874969482\n", + "Epoch [639/1000] Batch 0/23 Loss D: 0.5769774168729782, Loss G: 1.6565502882003784\n", + "Epoch [640/1000] Batch 0/23 Loss D: 0.5730944871902466, Loss G: 1.7599544525146484\n", + "Epoch [640/1000] Real Acc: 94.10%, Fake Acc: 3.09%\n", + "Epoch [641/1000] Batch 0/23 Loss D: 0.6916592121124268, Loss G: 1.3258917331695557\n", + "Epoch [642/1000] Batch 0/23 Loss D: 0.6088141798973083, Loss G: 1.3935219049453735\n", + "Epoch [643/1000] Batch 0/23 Loss D: 0.5516484379768372, Loss G: 1.3452067375183105\n", + "Epoch [644/1000] Batch 0/23 Loss D: 0.5107370913028717, Loss G: 1.8384137153625488\n", + "Epoch [645/1000] Batch 0/23 Loss D: 0.265736386179924, Loss G: 2.1873226165771484\n", + "Epoch [645/1000] Real Acc: 94.94%, Fake Acc: 0.56%\n", + "Epoch [646/1000] Batch 0/23 Loss D: 0.4214561730623245, Loss G: 1.8164873123168945\n", + "Epoch [647/1000] Batch 0/23 Loss D: 0.696767657995224, Loss G: 2.0747504234313965\n", + "Epoch [648/1000] Batch 0/23 Loss D: 0.8753452003002167, Loss G: 1.3925683498382568\n", + "Epoch [649/1000] Batch 0/23 Loss D: 0.755436360836029, Loss G: 1.6707382202148438\n", + "Epoch [650/1000] Batch 0/23 Loss D: 0.7063145041465759, Loss G: 1.664191484451294\n", + "Epoch [650/1000] Real Acc: 90.17%, Fake Acc: 3.65%\n", + "Epoch [651/1000] Batch 0/23 Loss D: 0.41027264297008514, Loss G: 1.8925267457962036\n", + "Epoch [652/1000] Batch 0/23 Loss D: 0.47225430607795715, Loss G: 2.0790858268737793\n", + "Epoch [653/1000] Batch 0/23 Loss D: 0.3732316642999649, Loss G: 1.829578161239624\n", + "Epoch [654/1000] Batch 0/23 Loss D: 0.6464215517044067, Loss G: 1.49284029006958\n", + "Epoch [655/1000] Batch 0/23 Loss D: 0.5504963099956512, Loss G: 1.820786714553833\n", + "Epoch [655/1000] Real Acc: 93.54%, Fake Acc: 2.53%\n", + "Epoch [656/1000] Batch 0/23 Loss D: 0.6373735964298248, Loss G: 1.3126215934753418\n", + "Epoch [657/1000] Batch 0/23 Loss D: 0.5444552004337311, Loss G: 1.6463463306427002\n", + "Epoch [658/1000] Batch 0/23 Loss D: 0.40984682738780975, Loss G: 1.6370964050292969\n", + "Epoch [659/1000] Batch 0/23 Loss D: 0.4207684248685837, Loss G: 1.9740104675292969\n", + "Epoch [660/1000] Batch 0/23 Loss D: 0.4369654059410095, Loss G: 2.005077838897705\n", + "Epoch [660/1000] Real Acc: 85.67%, Fake Acc: 4.49%\n", + "Epoch [661/1000] Batch 0/23 Loss D: 0.4078576862812042, Loss G: 2.028688907623291\n", + "Epoch [662/1000] Batch 0/23 Loss D: 0.5862630009651184, Loss G: 1.496462106704712\n", + "Epoch [663/1000] Batch 0/23 Loss D: 0.6505922377109528, Loss G: 2.0850486755371094\n", + "Epoch [664/1000] Batch 0/23 Loss D: 0.5681969374418259, Loss G: 1.6413692235946655\n", + "Epoch [665/1000] Batch 0/23 Loss D: 0.4073810577392578, Loss G: 2.3936541080474854\n", + "Epoch [665/1000] Real Acc: 96.63%, Fake Acc: 5.62%\n", + "Epoch [666/1000] Batch 0/23 Loss D: 0.4790714681148529, Loss G: 1.6331710815429688\n", + "Epoch [667/1000] Batch 0/23 Loss D: 0.6248458027839661, Loss G: 1.7683165073394775\n", + "Epoch [668/1000] Batch 0/23 Loss D: 0.7225433588027954, Loss G: 1.4833409786224365\n", + "Epoch [669/1000] Batch 0/23 Loss D: 0.5647903382778168, Loss G: 1.575711727142334\n", + "Epoch [670/1000] Batch 0/23 Loss D: 0.5364974141120911, Loss G: 1.494586706161499\n", + "Epoch [670/1000] Real Acc: 97.75%, Fake Acc: 2.53%\n", + "Epoch [671/1000] Batch 0/23 Loss D: 0.395054429769516, Loss G: 1.9541083574295044\n", + "Epoch [672/1000] Batch 0/23 Loss D: 0.3073514401912689, Loss G: 2.0750296115875244\n", + "Epoch [673/1000] Batch 0/23 Loss D: 0.4368128776550293, Loss G: 1.8805391788482666\n", + "Epoch [674/1000] Batch 0/23 Loss D: 0.40857505798339844, Loss G: 1.6597763299942017\n", + "Epoch [675/1000] Batch 0/23 Loss D: 0.3938868045806885, Loss G: 2.0070629119873047\n", + "Epoch [675/1000] Real Acc: 81.74%, Fake Acc: 12.64%\n", + "Epoch [676/1000] Batch 0/23 Loss D: 0.49831071496009827, Loss G: 2.0357131958007812\n", + "Epoch [677/1000] Batch 0/23 Loss D: 0.6706366240978241, Loss G: 1.5840790271759033\n", + "Epoch [678/1000] Batch 0/23 Loss D: 0.4689032584428787, Loss G: 2.3528635501861572\n", + "Epoch [679/1000] Batch 0/23 Loss D: 0.35451893508434296, Loss G: 2.1359715461730957\n", + "Epoch [680/1000] Batch 0/23 Loss D: 0.45559224486351013, Loss G: 1.8422263860702515\n", + "Epoch [680/1000] Real Acc: 96.91%, Fake Acc: 1.12%\n", + "Epoch [681/1000] Batch 0/23 Loss D: 0.45947614312171936, Loss G: 1.6690003871917725\n", + "Epoch [682/1000] Batch 0/23 Loss D: 0.4738115221261978, Loss G: 1.6514534950256348\n", + "Epoch [683/1000] Batch 0/23 Loss D: 0.49108219146728516, Loss G: 1.8508507013320923\n", + "Epoch [684/1000] Batch 0/23 Loss D: 0.3053670674562454, Loss G: 2.0611233711242676\n", + "Epoch [685/1000] Batch 0/23 Loss D: 0.27861763536930084, Loss G: 1.9463157653808594\n", + "Epoch [685/1000] Real Acc: 97.47%, Fake Acc: 1.40%\n", + "Epoch [686/1000] Batch 0/23 Loss D: 0.41502389311790466, Loss G: 2.248588800430298\n", + "Epoch [687/1000] Batch 0/23 Loss D: 0.5757710486650467, Loss G: 1.9125847816467285\n", + "Epoch [688/1000] Batch 0/23 Loss D: 1.016851618885994, Loss G: 1.8594303131103516\n", + "Epoch [689/1000] Batch 0/23 Loss D: 0.538707971572876, Loss G: 2.3420052528381348\n", + "Epoch [690/1000] Batch 0/23 Loss D: 0.5565034300088882, Loss G: 2.3646109104156494\n", + "Epoch [690/1000] Real Acc: 93.26%, Fake Acc: 2.25%\n", + "Epoch [691/1000] Batch 0/23 Loss D: 0.42304034531116486, Loss G: 2.0903587341308594\n", + "Epoch [692/1000] Batch 0/23 Loss D: 0.40439555048942566, Loss G: 2.459674835205078\n", + "Epoch [693/1000] Batch 0/23 Loss D: 0.4161361753940582, Loss G: 2.0646142959594727\n", + "Epoch [694/1000] Batch 0/23 Loss D: 0.5657832622528076, Loss G: 2.0898244380950928\n", + "Epoch [695/1000] Batch 0/23 Loss D: 0.617393285036087, Loss G: 1.525062084197998\n", + "Epoch [695/1000] Real Acc: 98.88%, Fake Acc: 11.80%\n", + "Epoch [696/1000] Batch 0/23 Loss D: 0.5455068945884705, Loss G: 1.7085018157958984\n", + "Epoch [697/1000] Batch 0/23 Loss D: 0.46626943349838257, Loss G: 1.5640250444412231\n", + "Epoch [698/1000] Batch 0/23 Loss D: 0.4670231342315674, Loss G: 2.2775232791900635\n", + "Epoch [699/1000] Batch 0/23 Loss D: 0.3681516945362091, Loss G: 1.9306716918945312\n", + "Epoch [700/1000] Batch 0/23 Loss D: 0.5577767938375473, Loss G: 1.6950385570526123\n", + "Epoch [700/1000] Real Acc: 87.36%, Fake Acc: 7.58%\n", + "Epoch [701/1000] Batch 0/23 Loss D: 0.5427952706813812, Loss G: 2.0050573348999023\n", + "Epoch [702/1000] Batch 0/23 Loss D: 0.6183153390884399, Loss G: 1.7897576093673706\n", + "Epoch [703/1000] Batch 0/23 Loss D: 0.599407821893692, Loss G: 2.3556928634643555\n", + "Epoch [704/1000] Batch 0/23 Loss D: 0.5119924247264862, Loss G: 2.0047173500061035\n", + "Epoch [705/1000] Batch 0/23 Loss D: 0.39800673723220825, Loss G: 2.2286899089813232\n", + "Epoch [705/1000] Real Acc: 97.75%, Fake Acc: 1.69%\n", + "Epoch [706/1000] Batch 0/23 Loss D: 0.38441258668899536, Loss G: 2.2046737670898438\n", + "Epoch [707/1000] Batch 0/23 Loss D: 0.5313562452793121, Loss G: 1.4880666732788086\n", + "Epoch [708/1000] Batch 0/23 Loss D: 0.43445271253585815, Loss G: 1.7866506576538086\n", + "Epoch [709/1000] Batch 0/23 Loss D: 0.4926695227622986, Loss G: 1.6710944175720215\n", + "Epoch [710/1000] Batch 0/23 Loss D: 0.4078086167573929, Loss G: 1.7837939262390137\n", + "Epoch [710/1000] Real Acc: 97.47%, Fake Acc: 0.00%\n", + "Epoch [711/1000] Batch 0/23 Loss D: 0.28621676564216614, Loss G: 2.1750917434692383\n", + "Epoch [712/1000] Batch 0/23 Loss D: 0.23197770863771439, Loss G: 2.121762275695801\n", + "Epoch [713/1000] Batch 0/23 Loss D: 0.26723408699035645, Loss G: 2.6382195949554443\n", + "Epoch [714/1000] Batch 0/23 Loss D: 0.4748719036579132, Loss G: 1.921700358390808\n", + "Epoch [715/1000] Batch 0/23 Loss D: 0.5776992440223694, Loss G: 1.859175205230713\n", + "Epoch [715/1000] Real Acc: 79.78%, Fake Acc: 7.87%\n", + "Epoch [716/1000] Batch 0/23 Loss D: 0.3694310933351517, Loss G: 2.2715187072753906\n", + "Epoch [717/1000] Batch 0/23 Loss D: 0.5222395658493042, Loss G: 2.226367473602295\n", + "Epoch [718/1000] Batch 0/23 Loss D: 0.26056408882141113, Loss G: 2.624802827835083\n", + "Epoch [719/1000] Batch 0/23 Loss D: 0.4095437079668045, Loss G: 2.105677604675293\n", + "Epoch [720/1000] Batch 0/23 Loss D: 0.4998094290494919, Loss G: 2.190855026245117\n", + "Epoch [720/1000] Real Acc: 91.01%, Fake Acc: 3.37%\n", + "Epoch [721/1000] Batch 0/23 Loss D: 0.47007372975349426, Loss G: 1.7278811931610107\n", + "Epoch [722/1000] Batch 0/23 Loss D: 0.5202972143888474, Loss G: 1.225940465927124\n", + "Epoch [723/1000] Batch 0/23 Loss D: 0.3766924887895584, Loss G: 1.9878731966018677\n", + "Epoch [724/1000] Batch 0/23 Loss D: 0.3198724538087845, Loss G: 1.8506534099578857\n", + "Epoch [725/1000] Batch 0/23 Loss D: 0.27296310663223267, Loss G: 2.192688226699829\n", + "Epoch [725/1000] Real Acc: 91.29%, Fake Acc: 0.28%\n", + "Epoch [726/1000] Batch 0/23 Loss D: 0.4102349579334259, Loss G: 1.911916732788086\n", + "Epoch [727/1000] Batch 0/23 Loss D: 0.3120775520801544, Loss G: 2.0978260040283203\n", + "Epoch [728/1000] Batch 0/23 Loss D: 0.516964003443718, Loss G: 2.0101397037506104\n", + "Epoch [729/1000] Batch 0/23 Loss D: 0.5505951195955276, Loss G: 1.575822353363037\n", + "Epoch [730/1000] Batch 0/23 Loss D: 0.4277323782444, Loss G: 2.9986555576324463\n", + "Epoch [730/1000] Real Acc: 86.24%, Fake Acc: 4.78%\n", + "Epoch [731/1000] Batch 0/23 Loss D: 0.5542042478919029, Loss G: 1.9992533922195435\n", + "Epoch [732/1000] Batch 0/23 Loss D: 0.31105512380599976, Loss G: 2.0793018341064453\n", + "Epoch [733/1000] Batch 0/23 Loss D: 0.3799158185720444, Loss G: 2.3727757930755615\n", + "Epoch [734/1000] Batch 0/23 Loss D: 0.3364306539297104, Loss G: 2.0846590995788574\n", + "Epoch [735/1000] Batch 0/23 Loss D: 0.5612092614173889, Loss G: 1.8319580554962158\n", + "Epoch [735/1000] Real Acc: 96.07%, Fake Acc: 2.53%\n", + "Epoch [736/1000] Batch 0/23 Loss D: 0.39485564827919006, Loss G: 1.6760692596435547\n", + "Epoch [737/1000] Batch 0/23 Loss D: 0.49780070781707764, Loss G: 1.9869725704193115\n", + "Epoch [738/1000] Batch 0/23 Loss D: 0.40727993845939636, Loss G: 1.9389435052871704\n", + "Epoch [739/1000] Batch 0/23 Loss D: 0.4445662796497345, Loss G: 1.8105010986328125\n", + "Epoch [740/1000] Batch 0/23 Loss D: 0.3699406534433365, Loss G: 1.6921947002410889\n", + "Epoch [740/1000] Real Acc: 97.47%, Fake Acc: 0.56%\n", + "Epoch [741/1000] Batch 0/23 Loss D: 0.45340389013290405, Loss G: 2.106980323791504\n", + "Epoch [742/1000] Batch 0/23 Loss D: 0.3331945687532425, Loss G: 2.1161506175994873\n", + "Epoch [743/1000] Batch 0/23 Loss D: 0.2192007228732109, Loss G: 2.207183361053467\n", + "Epoch [744/1000] Batch 0/23 Loss D: 0.38123421370983124, Loss G: 2.5318892002105713\n", + "Epoch [745/1000] Batch 0/23 Loss D: 0.2628899812698364, Loss G: 1.958256721496582\n", + "Epoch [745/1000] Real Acc: 87.64%, Fake Acc: 3.65%\n", + "Epoch [746/1000] Batch 0/23 Loss D: 0.668101042509079, Loss G: 2.422340154647827\n", + "Epoch [747/1000] Batch 0/23 Loss D: 0.47939226031303406, Loss G: 2.1966443061828613\n", + "Epoch [748/1000] Batch 0/23 Loss D: 0.40049228072166443, Loss G: 2.769843339920044\n", + "Epoch [749/1000] Batch 0/23 Loss D: 0.4379987120628357, Loss G: 2.6643924713134766\n", + "Epoch [750/1000] Batch 0/23 Loss D: 0.41154783964157104, Loss G: 2.137369155883789\n", + "Epoch [750/1000] Real Acc: 95.51%, Fake Acc: 2.25%\n", + "Epoch [751/1000] Batch 0/23 Loss D: 0.4446513056755066, Loss G: 2.0730953216552734\n", + "Epoch [752/1000] Batch 0/23 Loss D: 0.4092576205730438, Loss G: 2.299288272857666\n", + "Epoch [753/1000] Batch 0/23 Loss D: 0.2849167734384537, Loss G: 1.9391170740127563\n", + "Epoch [754/1000] Batch 0/23 Loss D: 0.5208230912685394, Loss G: 2.1716158390045166\n", + "Epoch [755/1000] Batch 0/23 Loss D: 0.32266412675380707, Loss G: 2.2717125415802\n", + "Epoch [755/1000] Real Acc: 99.16%, Fake Acc: 1.69%\n", + "Epoch [756/1000] Batch 0/23 Loss D: 0.2486598826944828, Loss G: 1.684791922569275\n", + "Epoch [757/1000] Batch 0/23 Loss D: 0.2627113610506058, Loss G: 2.2583999633789062\n", + "Epoch [758/1000] Batch 0/23 Loss D: 0.3872673809528351, Loss G: 2.0979745388031006\n", + "Epoch [759/1000] Batch 0/23 Loss D: 0.2943151742219925, Loss G: 2.598573684692383\n", + "Epoch [760/1000] Batch 0/23 Loss D: 0.26166434586048126, Loss G: 2.299146890640259\n", + "Epoch [760/1000] Real Acc: 98.03%, Fake Acc: 3.93%\n", + "Epoch [761/1000] Batch 0/23 Loss D: 0.32942503690719604, Loss G: 2.5791683197021484\n", + "Epoch [762/1000] Batch 0/23 Loss D: 0.6599729359149933, Loss G: 1.7470852136611938\n", + "Epoch [763/1000] Batch 0/23 Loss D: 0.36727993190288544, Loss G: 2.950119972229004\n", + "Epoch [764/1000] Batch 0/23 Loss D: 0.4546245038509369, Loss G: 2.026840925216675\n", + "Epoch [765/1000] Batch 0/23 Loss D: 0.5075939893722534, Loss G: 2.837717056274414\n", + "Epoch [765/1000] Real Acc: 96.91%, Fake Acc: 1.97%\n", + "Epoch [766/1000] Batch 0/23 Loss D: 0.3978602886199951, Loss G: 2.432551383972168\n", + "Epoch [767/1000] Batch 0/23 Loss D: 0.26262734085321426, Loss G: 2.6047911643981934\n", + "Epoch [768/1000] Batch 0/23 Loss D: 0.3338627517223358, Loss G: 2.739806652069092\n", + "Epoch [769/1000] Batch 0/23 Loss D: 0.43135911226272583, Loss G: 2.3162877559661865\n", + "Epoch [770/1000] Batch 0/23 Loss D: 0.40528682619333267, Loss G: 2.584047794342041\n", + "Epoch [770/1000] Real Acc: 96.91%, Fake Acc: 2.81%\n", + "Epoch [771/1000] Batch 0/23 Loss D: 0.5685635209083557, Loss G: 2.3242921829223633\n", + "Epoch [772/1000] Batch 0/23 Loss D: 0.3906610310077667, Loss G: 2.015174388885498\n", + "Epoch [773/1000] Batch 0/23 Loss D: 0.33574172854423523, Loss G: 2.521000385284424\n", + "Epoch [774/1000] Batch 0/23 Loss D: 0.35465314984321594, Loss G: 1.938570499420166\n", + "Epoch [775/1000] Batch 0/23 Loss D: 0.341297447681427, Loss G: 2.160167932510376\n", + "Epoch [775/1000] Real Acc: 97.19%, Fake Acc: 1.12%\n", + "Epoch [776/1000] Batch 0/23 Loss D: 0.33708329498767853, Loss G: 2.279932737350464\n", + "Epoch [777/1000] Batch 0/23 Loss D: 0.2845924496650696, Loss G: 2.702007532119751\n", + "Epoch [778/1000] Batch 0/23 Loss D: 0.24666361510753632, Loss G: 2.3755502700805664\n", + "Epoch [779/1000] Batch 0/23 Loss D: 0.30508939176797867, Loss G: 2.236236810684204\n", + "Epoch [780/1000] Batch 0/23 Loss D: 0.34169401228427887, Loss G: 2.177642822265625\n", + "Epoch [780/1000] Real Acc: 88.20%, Fake Acc: 1.40%\n", + "Epoch [781/1000] Batch 0/23 Loss D: 0.45319247245788574, Loss G: 1.7297025918960571\n", + "Epoch [782/1000] Batch 0/23 Loss D: 0.4721265435218811, Loss G: 2.782609224319458\n", + "Epoch [783/1000] Batch 0/23 Loss D: 0.5456440150737762, Loss G: 1.9321858882904053\n", + "Epoch [784/1000] Batch 0/23 Loss D: 0.45069437474012375, Loss G: 3.3206937313079834\n", + "Epoch [785/1000] Batch 0/23 Loss D: 0.23938192427158356, Loss G: 2.0431485176086426\n", + "Epoch [785/1000] Real Acc: 99.44%, Fake Acc: 5.06%\n", + "Epoch [786/1000] Batch 0/23 Loss D: 0.3409685343503952, Loss G: 3.2893428802490234\n", + "Epoch [787/1000] Batch 0/23 Loss D: 0.3074025511741638, Loss G: 2.286190986633301\n", + "Epoch [788/1000] Batch 0/23 Loss D: 0.38436073064804077, Loss G: 2.2126293182373047\n", + "Epoch [789/1000] Batch 0/23 Loss D: 0.2768014743924141, Loss G: 2.0464305877685547\n", + "Epoch [790/1000] Batch 0/23 Loss D: 0.5100202858448029, Loss G: 2.0845603942871094\n", + "Epoch [790/1000] Real Acc: 98.60%, Fake Acc: 5.62%\n", + "Epoch [791/1000] Batch 0/23 Loss D: 0.45906277000904083, Loss G: 2.1275787353515625\n", + "Epoch [792/1000] Batch 0/23 Loss D: 0.3624872416257858, Loss G: 2.2780303955078125\n", + "Epoch [793/1000] Batch 0/23 Loss D: 0.177254781126976, Loss G: 2.639557361602783\n", + "Epoch [794/1000] Batch 0/23 Loss D: 0.2617115080356598, Loss G: 2.1705777645111084\n", + "Epoch [795/1000] Batch 0/23 Loss D: 0.38156653940677643, Loss G: 1.9594659805297852\n", + "Epoch [795/1000] Real Acc: 97.47%, Fake Acc: 1.12%\n", + "Epoch [796/1000] Batch 0/23 Loss D: 0.34074393659830093, Loss G: 2.612037181854248\n", + "Epoch [797/1000] Batch 0/23 Loss D: 0.2259068638086319, Loss G: 2.0103085041046143\n", + "Epoch [798/1000] Batch 0/23 Loss D: 0.2306382805109024, Loss G: 2.752187967300415\n", + "Epoch [799/1000] Batch 0/23 Loss D: 0.22339308261871338, Loss G: 2.3246235847473145\n", + "Epoch [800/1000] Batch 0/23 Loss D: 0.449050597846508, Loss G: 3.260932445526123\n", + "Epoch [800/1000] Real Acc: 87.36%, Fake Acc: 4.78%\n", + "Epoch [801/1000] Batch 0/23 Loss D: 0.5084923803806305, Loss G: 2.493833303451538\n", + "Epoch [802/1000] Batch 0/23 Loss D: 0.2602807730436325, Loss G: 3.0771851539611816\n", + "Epoch [803/1000] Batch 0/23 Loss D: 0.2940438985824585, Loss G: 2.434727430343628\n", + "Epoch [804/1000] Batch 0/23 Loss D: 0.30049796402454376, Loss G: 2.3504369258880615\n", + "Epoch [805/1000] Batch 0/23 Loss D: 0.30113860964775085, Loss G: 2.8282077312469482\n", + "Epoch [805/1000] Real Acc: 96.63%, Fake Acc: 0.00%\n", + "Epoch [806/1000] Batch 0/23 Loss D: 0.27890579402446747, Loss G: 2.5492920875549316\n", + "Epoch [807/1000] Batch 0/23 Loss D: 0.2866232469677925, Loss G: 2.3009092807769775\n", + "Epoch [808/1000] Batch 0/23 Loss D: 0.39869242906570435, Loss G: 2.5716278553009033\n", + "Epoch [809/1000] Batch 0/23 Loss D: 0.3016359657049179, Loss G: 1.8543099164962769\n", + "Epoch [810/1000] Batch 0/23 Loss D: 0.2745816260576248, Loss G: 2.6317224502563477\n", + "Epoch [810/1000] Real Acc: 95.22%, Fake Acc: 0.00%\n", + "Epoch [811/1000] Batch 0/23 Loss D: 0.47969190776348114, Loss G: 2.4773685932159424\n", + "Epoch [812/1000] Batch 0/23 Loss D: 0.3244854658842087, Loss G: 2.8434700965881348\n", + "Epoch [813/1000] Batch 0/23 Loss D: 0.2549198493361473, Loss G: 1.9353375434875488\n", + "Epoch [814/1000] Batch 0/23 Loss D: 0.4308910593390465, Loss G: 2.0691640377044678\n", + "Epoch [815/1000] Batch 0/23 Loss D: 0.2178979068994522, Loss G: 2.317265510559082\n", + "Epoch [815/1000] Real Acc: 97.75%, Fake Acc: 0.00%\n", + "Epoch [816/1000] Batch 0/23 Loss D: 0.286690354347229, Loss G: 2.049811840057373\n", + "Epoch [817/1000] Batch 0/23 Loss D: 0.32257765531539917, Loss G: 2.861771583557129\n", + "Epoch [818/1000] Batch 0/23 Loss D: 0.256879024207592, Loss G: 2.3488330841064453\n", + "Epoch [819/1000] Batch 0/23 Loss D: 0.24206174165010452, Loss G: 2.578643798828125\n", + "Epoch [820/1000] Batch 0/23 Loss D: 0.35975614190101624, Loss G: 2.556269884109497\n", + "Epoch [820/1000] Real Acc: 93.26%, Fake Acc: 3.09%\n", + "Epoch [821/1000] Batch 0/23 Loss D: 0.3224495053291321, Loss G: 2.5737340450286865\n", + "Epoch [822/1000] Batch 0/23 Loss D: 0.2407764121890068, Loss G: 3.5213284492492676\n", + "Epoch [823/1000] Batch 0/23 Loss D: 0.37477223575115204, Loss G: 2.6729226112365723\n", + "Epoch [824/1000] Batch 0/23 Loss D: 0.322219904512167, Loss G: 3.3301455974578857\n", + "Epoch [825/1000] Batch 0/23 Loss D: 0.32546573877334595, Loss G: 2.5282387733459473\n", + "Epoch [825/1000] Real Acc: 97.47%, Fake Acc: 1.97%\n", + "Epoch [826/1000] Batch 0/23 Loss D: 0.3258020579814911, Loss G: 2.6167643070220947\n", + "Epoch [827/1000] Batch 0/23 Loss D: 0.30952006578445435, Loss G: 2.388688802719116\n", + "Epoch [828/1000] Batch 0/23 Loss D: 0.31285932660102844, Loss G: 2.847085475921631\n", + "Epoch [829/1000] Batch 0/23 Loss D: 0.3301183208823204, Loss G: 2.0890464782714844\n", + "Epoch [830/1000] Batch 0/23 Loss D: 0.17198138684034348, Loss G: 2.4792003631591797\n", + "Epoch [830/1000] Real Acc: 98.31%, Fake Acc: 0.28%\n", + "Epoch [831/1000] Batch 0/23 Loss D: 0.206959068775177, Loss G: 2.5451111793518066\n", + "Epoch [832/1000] Batch 0/23 Loss D: 0.24676945805549622, Loss G: 2.344508647918701\n", + "Epoch [833/1000] Batch 0/23 Loss D: 0.4213044345378876, Loss G: 2.4947705268859863\n", + "Epoch [834/1000] Batch 0/23 Loss D: 0.2318100482225418, Loss G: 2.469478130340576\n", + "Epoch [835/1000] Batch 0/23 Loss D: 0.25957415997982025, Loss G: 2.6126041412353516\n", + "Epoch [835/1000] Real Acc: 97.47%, Fake Acc: 2.25%\n", + "Epoch [836/1000] Batch 0/23 Loss D: 0.3431788384914398, Loss G: 2.931692123413086\n", + "Epoch [837/1000] Batch 0/23 Loss D: 0.3049492985010147, Loss G: 2.3051652908325195\n", + "Epoch [838/1000] Batch 0/23 Loss D: 0.38884488493204117, Loss G: 2.389439582824707\n", + "Epoch [839/1000] Batch 0/23 Loss D: 0.2823922485113144, Loss G: 2.2551677227020264\n", + "Epoch [840/1000] Batch 0/23 Loss D: 0.3812674880027771, Loss G: 2.9796626567840576\n", + "Epoch [840/1000] Real Acc: 95.79%, Fake Acc: 0.84%\n", + "Epoch [841/1000] Batch 0/23 Loss D: 0.20645532757043839, Loss G: 2.165639877319336\n", + "Epoch [842/1000] Batch 0/23 Loss D: 0.25912514328956604, Loss G: 3.1003761291503906\n", + "Epoch [843/1000] Batch 0/23 Loss D: 0.2538253143429756, Loss G: 2.4062857627868652\n", + "Epoch [844/1000] Batch 0/23 Loss D: 0.3273898661136627, Loss G: 3.0974349975585938\n", + "Epoch [845/1000] Batch 0/23 Loss D: 0.33049117028713226, Loss G: 2.1955480575561523\n", + "Epoch [845/1000] Real Acc: 98.88%, Fake Acc: 2.25%\n", + "Epoch [846/1000] Batch 0/23 Loss D: 0.2696770280599594, Loss G: 2.860813617706299\n", + "Epoch [847/1000] Batch 0/23 Loss D: 0.28046105802059174, Loss G: 2.6653499603271484\n", + "Epoch [848/1000] Batch 0/23 Loss D: 0.17039768397808075, Loss G: 2.424706220626831\n", + "Epoch [849/1000] Batch 0/23 Loss D: 0.25552429258823395, Loss G: 2.633436679840088\n", + "Epoch [850/1000] Batch 0/23 Loss D: 0.2570285350084305, Loss G: 2.4547855854034424\n", + "Epoch [850/1000] Real Acc: 98.31%, Fake Acc: 1.97%\n", + "Epoch [851/1000] Batch 0/23 Loss D: 0.23369663208723068, Loss G: 2.6814160346984863\n", + "Epoch [852/1000] Batch 0/23 Loss D: 0.24738512933254242, Loss G: 2.529574394226074\n", + "Epoch [853/1000] Batch 0/23 Loss D: 0.2703852355480194, Loss G: 3.323538303375244\n", + "Epoch [854/1000] Batch 0/23 Loss D: 0.25548428297042847, Loss G: 2.4834179878234863\n", + "Epoch [855/1000] Batch 0/23 Loss D: 0.2069792076945305, Loss G: 3.166983127593994\n", + "Epoch [855/1000] Real Acc: 98.31%, Fake Acc: 1.40%\n", + "Epoch [856/1000] Batch 0/23 Loss D: 0.32368743419647217, Loss G: 2.0717761516571045\n", + "Epoch [857/1000] Batch 0/23 Loss D: 0.19805990159511566, Loss G: 2.4600343704223633\n", + "Epoch [858/1000] Batch 0/23 Loss D: 0.26512235403060913, Loss G: 2.547729969024658\n", + "Epoch [859/1000] Batch 0/23 Loss D: 0.14602632820606232, Loss G: 3.082188129425049\n", + "Epoch [860/1000] Batch 0/23 Loss D: 0.23736684024333954, Loss G: 2.715940237045288\n", + "Epoch [860/1000] Real Acc: 98.31%, Fake Acc: 1.12%\n", + "Epoch [861/1000] Batch 0/23 Loss D: 0.20693979412317276, Loss G: 3.2972092628479004\n", + "Epoch [862/1000] Batch 0/23 Loss D: 0.3101959526538849, Loss G: 2.383943557739258\n", + "Epoch [863/1000] Batch 0/23 Loss D: 0.21865353360772133, Loss G: 4.005591869354248\n", + "Epoch [864/1000] Batch 0/23 Loss D: 0.2011248990893364, Loss G: 3.09849214553833\n", + "Epoch [865/1000] Batch 0/23 Loss D: 0.22805733233690262, Loss G: 3.1340842247009277\n", + "Epoch [865/1000] Real Acc: 97.47%, Fake Acc: 0.84%\n", + "Epoch [866/1000] Batch 0/23 Loss D: 0.24684391915798187, Loss G: 2.2790639400482178\n", + "Epoch [867/1000] Batch 0/23 Loss D: 0.37354443967342377, Loss G: 3.0161008834838867\n", + "Epoch [868/1000] Batch 0/23 Loss D: 0.34263794124126434, Loss G: 3.0583455562591553\n", + "Epoch [869/1000] Batch 0/23 Loss D: 0.2385891154408455, Loss G: 2.8178086280822754\n", + "Epoch [870/1000] Batch 0/23 Loss D: 0.14645694196224213, Loss G: 3.1074976921081543\n", + "Epoch [870/1000] Real Acc: 98.88%, Fake Acc: 1.12%\n", + "Epoch [871/1000] Batch 0/23 Loss D: 0.2926277816295624, Loss G: 2.5231614112854004\n", + "Epoch [872/1000] Batch 0/23 Loss D: 0.25309522449970245, Loss G: 2.630669116973877\n", + "Epoch [873/1000] Batch 0/23 Loss D: 0.26575466990470886, Loss G: 3.018551826477051\n", + "Epoch [874/1000] Batch 0/23 Loss D: 0.1526063233613968, Loss G: 2.6986451148986816\n", + "Epoch [875/1000] Batch 0/23 Loss D: 0.19461897015571594, Loss G: 2.809312343597412\n", + "Epoch [875/1000] Real Acc: 96.63%, Fake Acc: 0.56%\n", + "Epoch [876/1000] Batch 0/23 Loss D: 0.20205870270729065, Loss G: 2.6488945484161377\n", + "Epoch [877/1000] Batch 0/23 Loss D: 0.1588442549109459, Loss G: 2.7190206050872803\n", + "Epoch [878/1000] Batch 0/23 Loss D: 0.30875879526138306, Loss G: 3.1383352279663086\n", + "Epoch [879/1000] Batch 0/23 Loss D: 0.20848187804222107, Loss G: 3.0213990211486816\n", + "Epoch [880/1000] Batch 0/23 Loss D: 0.2854010872542858, Loss G: 2.4069042205810547\n", + "Epoch [880/1000] Real Acc: 100.00%, Fake Acc: 3.37%\n", + "Epoch [881/1000] Batch 0/23 Loss D: 0.2198922373354435, Loss G: 3.413283586502075\n", + "Epoch [882/1000] Batch 0/23 Loss D: 0.325573593378067, Loss G: 2.332303762435913\n", + "Epoch [883/1000] Batch 0/23 Loss D: 0.18957790732383728, Loss G: 3.6655454635620117\n", + "Epoch [884/1000] Batch 0/23 Loss D: 0.25293346494436264, Loss G: 2.1194913387298584\n", + "Epoch [885/1000] Batch 0/23 Loss D: 0.22987408936023712, Loss G: 2.596169948577881\n", + "Epoch [885/1000] Real Acc: 97.19%, Fake Acc: 0.56%\n", + "Epoch [886/1000] Batch 0/23 Loss D: 0.2527097836136818, Loss G: 3.1973447799682617\n", + "Epoch [887/1000] Batch 0/23 Loss D: 0.3377048820257187, Loss G: 2.2845475673675537\n", + "Epoch [888/1000] Batch 0/23 Loss D: 0.22306158021092415, Loss G: 3.2654097080230713\n", + "Epoch [889/1000] Batch 0/23 Loss D: 0.22401708364486694, Loss G: 1.9371683597564697\n", + "Epoch [890/1000] Batch 0/23 Loss D: 0.3029610700905323, Loss G: 2.7133402824401855\n", + "Epoch [890/1000] Real Acc: 97.75%, Fake Acc: 2.53%\n", + "Epoch [891/1000] Batch 0/23 Loss D: 0.2819867208600044, Loss G: 1.766751766204834\n", + "Epoch [892/1000] Batch 0/23 Loss D: 0.2509399801492691, Loss G: 3.0118541717529297\n", + "Epoch [893/1000] Batch 0/23 Loss D: 0.22666864097118378, Loss G: 2.997074604034424\n", + "Epoch [894/1000] Batch 0/23 Loss D: 0.20403822138905525, Loss G: 3.6658883094787598\n", + "Epoch [895/1000] Batch 0/23 Loss D: 0.2636972740292549, Loss G: 2.5857341289520264\n", + "Epoch [895/1000] Real Acc: 98.03%, Fake Acc: 2.53%\n", + "Epoch [896/1000] Batch 0/23 Loss D: 0.24335666745901108, Loss G: 2.7981953620910645\n", + "Epoch [897/1000] Batch 0/23 Loss D: 0.1886695809662342, Loss G: 2.1786892414093018\n", + "Epoch [898/1000] Batch 0/23 Loss D: 0.1808682158589363, Loss G: 3.0338850021362305\n", + "Epoch [899/1000] Batch 0/23 Loss D: 0.26953013986349106, Loss G: 2.3070523738861084\n", + "Epoch [900/1000] Batch 0/23 Loss D: 0.45048655569553375, Loss G: 2.8103585243225098\n", + "Epoch [900/1000] Real Acc: 97.47%, Fake Acc: 1.69%\n", + "Epoch [901/1000] Batch 0/23 Loss D: 0.18031931668519974, Loss G: 3.7229087352752686\n", + "Epoch [902/1000] Batch 0/23 Loss D: 0.24374520778656006, Loss G: 3.355424642562866\n", + "Epoch [903/1000] Batch 0/23 Loss D: 0.4248531609773636, Loss G: 2.7784042358398438\n", + "Epoch [904/1000] Batch 0/23 Loss D: 0.19800512120127678, Loss G: 2.3226237297058105\n", + "Epoch [905/1000] Batch 0/23 Loss D: 0.2723765969276428, Loss G: 2.995560646057129\n", + "Epoch [905/1000] Real Acc: 98.03%, Fake Acc: 0.84%\n", + "Epoch [906/1000] Batch 0/23 Loss D: 0.2157057523727417, Loss G: 3.051037073135376\n", + "Epoch [907/1000] Batch 0/23 Loss D: 0.21459923684597015, Loss G: 3.1612894535064697\n", + "Epoch [908/1000] Batch 0/23 Loss D: 0.17645389586687088, Loss G: 3.2754383087158203\n", + "Epoch [909/1000] Batch 0/23 Loss D: 0.19378702342510223, Loss G: 3.891714572906494\n", + "Epoch [910/1000] Batch 0/23 Loss D: 0.17613763362169266, Loss G: 2.6204423904418945\n", + "Epoch [910/1000] Real Acc: 99.72%, Fake Acc: 1.97%\n", + "Epoch [911/1000] Batch 0/23 Loss D: 0.15383370965719223, Loss G: 3.592376470565796\n", + "Epoch [912/1000] Batch 0/23 Loss D: 0.17839837819337845, Loss G: 3.306835651397705\n", + "Epoch [913/1000] Batch 0/23 Loss D: 0.27631013840436935, Loss G: 2.566854953765869\n", + "Epoch [914/1000] Batch 0/23 Loss D: 0.2756144031882286, Loss G: 2.681434154510498\n", + "Epoch [915/1000] Batch 0/23 Loss D: 0.21173498034477234, Loss G: 2.5820412635803223\n", + "Epoch [915/1000] Real Acc: 99.44%, Fake Acc: 1.40%\n", + "Epoch [916/1000] Batch 0/23 Loss D: 0.2905176058411598, Loss G: 2.7166903018951416\n", + "Epoch [917/1000] Batch 0/23 Loss D: 0.2600097581744194, Loss G: 2.6018404960632324\n", + "Epoch [918/1000] Batch 0/23 Loss D: 0.2742656469345093, Loss G: 2.972858428955078\n", + "Epoch [919/1000] Batch 0/23 Loss D: 0.3654281795024872, Loss G: 2.7434275150299072\n", + "Epoch [920/1000] Batch 0/23 Loss D: 0.3063781000673771, Loss G: 3.0240626335144043\n", + "Epoch [920/1000] Real Acc: 98.03%, Fake Acc: 1.97%\n", + "Epoch [921/1000] Batch 0/23 Loss D: 0.2854783982038498, Loss G: 3.232327938079834\n", + "Epoch [922/1000] Batch 0/23 Loss D: 0.18851874768733978, Loss G: 2.81685209274292\n", + "Epoch [923/1000] Batch 0/23 Loss D: 0.20574994385242462, Loss G: 2.8319621086120605\n", + "Epoch [924/1000] Batch 0/23 Loss D: 0.14824939519166946, Loss G: 3.076359510421753\n", + "Epoch [925/1000] Batch 0/23 Loss D: 0.20681221783161163, Loss G: 2.8218369483947754\n", + "Epoch [925/1000] Real Acc: 99.16%, Fake Acc: 2.53%\n", + "Epoch [926/1000] Batch 0/23 Loss D: 0.2289821282029152, Loss G: 3.5913100242614746\n", + "Epoch [927/1000] Batch 0/23 Loss D: 0.12397369369864464, Loss G: 2.776756763458252\n", + "Epoch [928/1000] Batch 0/23 Loss D: 0.26125068962574005, Loss G: 3.483673572540283\n", + "Epoch [929/1000] Batch 0/23 Loss D: 0.33381442353129387, Loss G: 2.409715175628662\n", + "Epoch [930/1000] Batch 0/23 Loss D: 0.1366928108036518, Loss G: 3.3834450244903564\n", + "Epoch [930/1000] Real Acc: 97.47%, Fake Acc: 1.97%\n", + "Epoch [931/1000] Batch 0/23 Loss D: 0.14287569373846054, Loss G: 2.994218349456787\n", + "Epoch [932/1000] Batch 0/23 Loss D: 0.16490013152360916, Loss G: 3.347367286682129\n", + "Epoch [933/1000] Batch 0/23 Loss D: 0.1769627258181572, Loss G: 2.3035130500793457\n", + "Epoch [934/1000] Batch 0/23 Loss D: 0.2006552293896675, Loss G: 3.196596622467041\n", + "Epoch [935/1000] Batch 0/23 Loss D: 0.16395948827266693, Loss G: 3.055840015411377\n", + "Epoch [935/1000] Real Acc: 99.44%, Fake Acc: 1.97%\n", + "Epoch [936/1000] Batch 0/23 Loss D: 0.1976695992052555, Loss G: 2.4289817810058594\n", + "Epoch [937/1000] Batch 0/23 Loss D: 0.16332245618104935, Loss G: 3.212984561920166\n", + "Epoch [938/1000] Batch 0/23 Loss D: 0.13315842300653458, Loss G: 2.586730480194092\n", + "Epoch [939/1000] Batch 0/23 Loss D: 0.18311869353055954, Loss G: 3.5585265159606934\n", + "Epoch [940/1000] Batch 0/23 Loss D: 0.20945944637060165, Loss G: 2.2657833099365234\n", + "Epoch [940/1000] Real Acc: 97.75%, Fake Acc: 1.69%\n", + "Epoch [941/1000] Batch 0/23 Loss D: 0.27115150541067123, Loss G: 2.6978840827941895\n", + "Epoch [942/1000] Batch 0/23 Loss D: 0.2078029066324234, Loss G: 2.5154612064361572\n", + "Epoch [943/1000] Batch 0/23 Loss D: 0.15444830805063248, Loss G: 3.449021816253662\n", + "Epoch [944/1000] Batch 0/23 Loss D: 0.11535946652293205, Loss G: 3.347259283065796\n", + "Epoch [945/1000] Batch 0/23 Loss D: 0.18246818333864212, Loss G: 3.656714916229248\n", + "Epoch [945/1000] Real Acc: 98.31%, Fake Acc: 0.84%\n", + "Epoch [946/1000] Batch 0/23 Loss D: 0.2060379683971405, Loss G: 2.893153429031372\n", + "Epoch [947/1000] Batch 0/23 Loss D: 0.253531277179718, Loss G: 3.4443166255950928\n", + "Epoch [948/1000] Batch 0/23 Loss D: 0.2272874265909195, Loss G: 2.76613187789917\n", + "Epoch [949/1000] Batch 0/23 Loss D: 0.20406624674797058, Loss G: 3.0956509113311768\n", + "Epoch [950/1000] Batch 0/23 Loss D: 0.1776702105998993, Loss G: 3.3508732318878174\n", + "Epoch [950/1000] Real Acc: 97.19%, Fake Acc: 1.40%\n", + "Epoch [951/1000] Batch 0/23 Loss D: 0.26598262786865234, Loss G: 3.144310474395752\n", + "Epoch [952/1000] Batch 0/23 Loss D: 0.23356162011623383, Loss G: 3.344789981842041\n", + "Epoch [953/1000] Batch 0/23 Loss D: 0.15512552857398987, Loss G: 3.511610984802246\n", + "Epoch [954/1000] Batch 0/23 Loss D: 0.29654766619205475, Loss G: 3.355741024017334\n", + "Epoch [955/1000] Batch 0/23 Loss D: 0.14333780109882355, Loss G: 2.5089993476867676\n", + "Epoch [955/1000] Real Acc: 98.88%, Fake Acc: 1.40%\n", + "Epoch [956/1000] Batch 0/23 Loss D: 0.2020891308784485, Loss G: 3.198739767074585\n", + "Epoch [957/1000] Batch 0/23 Loss D: 0.12147865071892738, Loss G: 2.8845200538635254\n", + "Epoch [958/1000] Batch 0/23 Loss D: 0.16822516545653343, Loss G: 3.821180820465088\n", + "Epoch [959/1000] Batch 0/23 Loss D: 0.29471029341220856, Loss G: 3.0861730575561523\n", + "Epoch [960/1000] Batch 0/23 Loss D: 0.20678913593292236, Loss G: 2.4100584983825684\n", + "Epoch [960/1000] Real Acc: 98.60%, Fake Acc: 1.97%\n", + "Epoch [961/1000] Batch 0/23 Loss D: 0.17619511485099792, Loss G: 3.1557583808898926\n", + "Epoch [962/1000] Batch 0/23 Loss D: 0.1656854897737503, Loss G: 2.8825104236602783\n", + "Epoch [963/1000] Batch 0/23 Loss D: 0.23758909478783607, Loss G: 3.6622655391693115\n", + "Epoch [964/1000] Batch 0/23 Loss D: 0.3515275865793228, Loss G: 2.8364129066467285\n", + "Epoch [965/1000] Batch 0/23 Loss D: 0.13080247677862644, Loss G: 4.495528697967529\n", + "Epoch [965/1000] Real Acc: 99.16%, Fake Acc: 2.53%\n", + "Epoch [966/1000] Batch 0/23 Loss D: 0.3244066685438156, Loss G: 3.241178512573242\n", + "Epoch [967/1000] Batch 0/23 Loss D: 0.23680805042386055, Loss G: 3.3793437480926514\n", + "Epoch [968/1000] Batch 0/23 Loss D: 0.2225012481212616, Loss G: 2.419538974761963\n", + "Epoch [969/1000] Batch 0/23 Loss D: 0.19923000037670135, Loss G: 3.2838656902313232\n", + "Epoch [970/1000] Batch 0/23 Loss D: 0.19154616445302963, Loss G: 2.7181711196899414\n", + "Epoch [970/1000] Real Acc: 98.60%, Fake Acc: 0.84%\n", + "Epoch [971/1000] Batch 0/23 Loss D: 0.1977619007229805, Loss G: 3.0215773582458496\n", + "Epoch [972/1000] Batch 0/23 Loss D: 0.11632193624973297, Loss G: 3.3004980087280273\n", + "Epoch [973/1000] Batch 0/23 Loss D: 0.19640526175498962, Loss G: 3.58303165435791\n", + "Epoch [974/1000] Batch 0/23 Loss D: 0.1615258865058422, Loss G: 2.6869184970855713\n", + "Epoch [975/1000] Batch 0/23 Loss D: 0.12897267937660217, Loss G: 3.2286739349365234\n", + "Epoch [975/1000] Real Acc: 99.44%, Fake Acc: 1.69%\n", + "Epoch [976/1000] Batch 0/23 Loss D: 0.28993314504623413, Loss G: 3.1189634799957275\n", + "Epoch [977/1000] Batch 0/23 Loss D: 0.15834452211856842, Loss G: 3.0074617862701416\n", + "Epoch [978/1000] Batch 0/23 Loss D: 0.14427898079156876, Loss G: 2.6294593811035156\n", + "Epoch [979/1000] Batch 0/23 Loss D: 0.16840355843305588, Loss G: 3.788971424102783\n", + "Epoch [980/1000] Batch 0/23 Loss D: 0.18142416328191757, Loss G: 2.9234983921051025\n", + "Epoch [980/1000] Real Acc: 99.16%, Fake Acc: 1.12%\n", + "Epoch [981/1000] Batch 0/23 Loss D: 0.1488904505968094, Loss G: 3.280503273010254\n", + "Epoch [982/1000] Batch 0/23 Loss D: 0.1452934294939041, Loss G: 3.3900537490844727\n", + "Epoch [983/1000] Batch 0/23 Loss D: 0.2960474267601967, Loss G: 3.0189406871795654\n", + "Epoch [984/1000] Batch 0/23 Loss D: 0.15827138721942902, Loss G: 3.461329460144043\n", + "Epoch [985/1000] Batch 0/23 Loss D: 0.26694902032613754, Loss G: 3.3112001419067383\n", + "Epoch [985/1000] Real Acc: 98.31%, Fake Acc: 0.84%\n", + "Epoch [986/1000] Batch 0/23 Loss D: 0.11519734561443329, Loss G: 2.616959571838379\n", + "Epoch [987/1000] Batch 0/23 Loss D: 0.5341759622097015, Loss G: 3.7079825401306152\n", + "Epoch [988/1000] Batch 0/23 Loss D: 0.19574761390686035, Loss G: 3.403731107711792\n", + "Epoch [989/1000] Batch 0/23 Loss D: 0.28019050508737564, Loss G: 3.63687801361084\n", + "Epoch [990/1000] Batch 0/23 Loss D: 0.18675917387008667, Loss G: 3.1097211837768555\n", + "Epoch [990/1000] Real Acc: 98.60%, Fake Acc: 0.84%\n", + "Epoch [991/1000] Batch 0/23 Loss D: 0.22472655028104782, Loss G: 2.827575922012329\n", + "Epoch [992/1000] Batch 0/23 Loss D: 0.10588303208351135, Loss G: 3.571537971496582\n", + "Epoch [993/1000] Batch 0/23 Loss D: 0.2705617845058441, Loss G: 3.0330653190612793\n", + "Epoch [994/1000] Batch 0/23 Loss D: 0.2050105556845665, Loss G: 2.677123546600342\n", + "Epoch [995/1000] Batch 0/23 Loss D: 0.2237449437379837, Loss G: 3.112083673477173\n", + "Epoch [995/1000] Real Acc: 98.31%, Fake Acc: 2.25%\n", + "Epoch [996/1000] Batch 0/23 Loss D: 0.12193522974848747, Loss G: 3.7434451580047607\n", + "Epoch [997/1000] Batch 0/23 Loss D: 0.15884925797581673, Loss G: 3.6587600708007812\n", + "Epoch [998/1000] Batch 0/23 Loss D: 0.15163548290729523, Loss G: 4.11873197555542\n", + "Epoch [999/1000] Batch 0/23 Loss D: 0.18032533675432205, Loss G: 4.287444114685059\n" + ] + } + ], + "source": [ + "num_epochs = 1000\n", + "print_interval = 50 # Adjust this to control how often you print the loss\n", + "discriminator_update_frequency = 4 # Update discriminator less frequently\n", + "\n", + "# Check for GPU availability\n", + "if torch.cuda.is_available():\n", + " print(\"Using GPU:\", torch.cuda.get_device_name(0))\n", + "else:\n", + " print(\"No GPU available, using the CPU instead.\")\n", + "\n", + "# Define device\n", + "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n", + "\n", + "# Move models to the chosen device\n", + "generator = generator.to(device)\n", + "discriminator = discriminator.to(device)\n", + "\n", + "# Initialize lists for storing metrics\n", + "d_losses = []\n", + "g_losses = []\n", + "real_accuracies = []\n", + "fake_accuracies = []\n", + "\n", + "import matplotlib.pyplot as plt\n", + "\n", + "def plot_generated_images(epoch, generator, fixed_noise, device, num_images=16):\n", + " # Tell the generator we are in eval mode\n", + " generator.eval()\n", + "\n", + " # Generate fake images\n", + " with torch.no_grad():\n", + " fake_images = generator(fixed_noise).detach().cpu()\n", + " \n", + " # Process and plot each image\n", + " fig, axs = plt.subplots(4, 4, figsize=(8, 8))\n", + " fig.suptitle(f\"Generated Images at Epoch {epoch}\")\n", + " cnt = 0\n", + " for i in range(4):\n", + " for j in range(4):\n", + " axs[i, j].imshow(fake_images[cnt, 0, :, :], cmap='gray')\n", + " axs[i, j].axis('off')\n", + " cnt += 1\n", + " plt.show()\n", + "\n", + " # Set the generator back to train mode\n", + " generator.train()\n", + "\n", + "# Create a fixed noise vector for generating images\n", + "fixed_noise = torch.randn(16, z_dim, 1, 1, device=device)\n", + "\n", + "\n", + "# Training loop\n", + "for epoch in range(num_epochs):\n", + " # Initialize variables to accumulate losses and accuracies\n", + " epoch_d_loss = 0.0\n", + " epoch_g_loss = 0.0\n", + " real_correct = 0\n", + " fake_correct = 0\n", + " total_real = 0\n", + " total_fake = 0\n", + "\n", + " for i, real_images in enumerate(dataloader):\n", + " current_batch_size = real_images.size(0)\n", + "\n", + " # Move data to the appropriate device and create labels for real and fake data\n", + " real_images = real_images.to(device)\n", + " real_labels = torch.ones(current_batch_size, 1, device=device)\n", + " fake_labels = torch.zeros(current_batch_size, 1, device=device)\n", + "\n", + " # ---------------------\n", + " # Train Discriminator\n", + " # ---------------------\n", + " discriminator.zero_grad()\n", + "\n", + " # Compute discriminator loss on real images\n", + " outputs = discriminator(real_images)\n", + " d_loss_real = criterion(outputs, real_labels)\n", + " d_loss_real.backward()\n", + " real_predictions = (outputs > 0.5).float()\n", + " real_correct += (real_predictions == real_labels).sum().item()\n", + " total_real += current_batch_size\n", + "\n", + " # Generate fake images\n", + " z = torch.randn(current_batch_size, z_dim, device=device)\n", + " fake_images = generator(z)\n", + "\n", + " # Compute discriminator loss on fake images\n", + " outputs = discriminator(fake_images.detach())\n", + " d_loss_fake = criterion(outputs, fake_labels)\n", + " d_loss_fake.backward()\n", + " fake_predictions = (outputs < 0.5).float()\n", + " fake_correct += (fake_predictions == fake_labels).sum().item()\n", + " total_fake += current_batch_size\n", + "\n", + " # Update discriminator based on update frequency\n", + " if i % discriminator_update_frequency == 0:\n", + " optimizer_d.step()\n", + "\n", + " # -----------------\n", + " # Train Generator\n", + " # -----------------\n", + " generator.zero_grad()\n", + " outputs = discriminator(fake_images)\n", + " g_loss = criterion(outputs, real_labels)\n", + " g_loss.backward()\n", + " optimizer_g.step()\n", + "\n", + " # Print loss stats\n", + " if i % print_interval == 0:\n", + " print(f\"Epoch [{epoch}/{num_epochs}] Batch {i}/{len(dataloader)} \\\n", + " Loss D: {d_loss_real.item() + d_loss_fake.item()}, Loss G: {g_loss.item()}\")\n", + " \n", + " # Accumulate losses for the epoch\n", + " epoch_d_loss += (d_loss_real.item() + d_loss_fake.item())\n", + " epoch_g_loss += g_loss.item()\n", + " \n", + "\n", + " \n", + " # Compute average losses and accuracies for the epoch\n", + " d_losses.append(epoch_d_loss / len(dataloader))\n", + " g_losses.append(epoch_g_loss / len(dataloader))\n", + " real_accuracies.append((real_correct / total_real) * 100)\n", + " fake_accuracies.append((fake_correct / total_fake) * 100)\n", + " \n", + " # Print discriminator's success rate every 10 epochs\n", + " if epoch % 5 == 0:\n", + "\n", + " plot_generated_images(epoch, generator, fixed_noise, device)\n", + " \n", + " real_acc = (real_correct / total_real) * 100\n", + " fake_acc = (fake_correct / total_fake) * 100\n", + " print(f\"Epoch [{epoch}/{num_epochs}] \\\n", + " Real Acc: {real_acc:.2f}%, Fake Acc: {fake_acc:.2f}%\")\n", + "\n", + " # Reset counters after each epoch\n", + " real_correct = 0\n", + " fake_correct = 0\n", + " total_real = 0\n", + " total_fake = 0\n", + "\n", + "# Add any additional code needed for saving models, tracking progress, etc.\n" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "id": "63feec53-528b-451c-a25c-d1d14fc9eaec", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "\n", + "# Plotting\n", + "plt.figure(figsize=(10, 5))\n", + "\n", + "# Plot discriminator and generator loss\n", + "plt.subplot(1, 2, 1)\n", + "plt.plot(d_losses, label='Discriminator Loss')\n", + "plt.plot(g_losses, label='Generator Loss')\n", + "plt.title(\"Loss During Training\")\n", + "plt.xlabel(\"Epoch\")\n", + "plt.ylabel(\"Loss\")\n", + "plt.legend()\n", + "\n", + "# Plot discriminator accuracy\n", + "plt.subplot(1, 2, 2)\n", + "plt.plot(real_accuracies, label='Real Accuracy')\n", + "plt.plot(fake_accuracies, label='Fake Accuracy')\n", + "plt.title(\"Discriminator Accuracy\")\n", + "plt.xlabel(\"Epoch\")\n", + "plt.ylabel(\"Accuracy (%)\")\n", + "plt.legend()\n", + "\n", + "plt.tight_layout()\n", + "plt.show()\n" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "id": "c765397c-ef2b-4a46-8df1-0f2b9bfe6abe", + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Using GPU: NVIDIA GeForce RTX 3080\n" + ] + }, + { + "data": { + "text/plain": [ + "Discriminator(\n", + " (model): Sequential(\n", + " (0): Conv2d(1, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n", + " (1): LeakyReLU(negative_slope=0.01)\n", + " (2): Dropout(p=0.3, inplace=False)\n", + " (3): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))\n", + " (4): LeakyReLU(negative_slope=0.01)\n", + " (5): Dropout(p=0.3, inplace=False)\n", + " (6): Flatten(start_dim=1, end_dim=-1)\n", + " (7): Linear(in_features=146432, out_features=1, bias=True)\n", + " (8): Sigmoid()\n", + " )\n", + ")" + ] + }, + "execution_count": 10, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "# # loading model script for later use\n", + "\n", + "# if torch.cuda.is_available():\n", + "# device = torch.device(\"cuda:0\") # Use the first GPU\n", + "# print(\"Using GPU:\", torch.cuda.get_device_name(0))\n", + "# else:\n", + "# device = torch.device(\"cpu\")\n", + "# print(\"GPU not available, using CPU instead.\")\n", + "\n", + "\n", + "# generator = torch.load('generator_model.pth', map_location=device)\n", + "# discriminator = torch.load('discriminator_model.pth', map_location=device)\n", + "\n", + "# generator.to(device)\n", + "# discriminator.to(device)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "id": "e60c7cc0-e383-4c9f-85d4-19deef59d680", + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "import torch\n", + "import numpy as np\n", + "\n", + "def show_images(images, num_images=5, threshold=0.5):\n", + " plt.figure(figsize=(50, 50)) # Adjust figsize as needed\n", + " binary_images = []\n", + "\n", + " for i in range(num_images):\n", + " img = images[i].squeeze().detach().cpu().numpy() # Remove channel dimension and convert to numpy\n", + " # Convert to binary\n", + " binary_img = (img > threshold).astype(int)\n", + " binary_images.append(binary_img)\n", + "\n", + " plt.subplot(1, num_images, i + 1)\n", + " plt.imshow(binary_img, cmap='gray')\n", + " plt.axis('off')\n", + " \n", + " plt.show()\n", + " return binary_images\n", + "\n", + "# Generate random noise\n", + "z = torch.randn(5, z_dim, device=device) # Adjust 'z_dim' to match your generator's input dimension\n", + "\n", + "# Generate images from the noise\n", + "generator.eval() # Set the generator to eval mode\n", + "with torch.no_grad(): # Disable gradient computation\n", + " fake_images = generator(z)\n", + "\n", + "# Display and get binary images\n", + "binary_images = show_images(fake_images)\n", + "\n", + "# Save binary images as .npy files\n", + "for i, img in enumerate(binary_images):\n", + " np.save(f'binary_image_{i}.npy', img)\n" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "id": "08c432f5-1737-4a54-86ee-64d4e91d09d5", + "metadata": {}, + "outputs": [], + "source": [ + "# Save the entire model\n", + "torch.save(generator, 'generator_model.pth')\n", + "torch.save(discriminator, 'discriminator_model.pth')\n", + "# Save the model state dictionaries\n", + "torch.save(generator.state_dict(), 'generator_state_dict.pth')\n", + "torch.save(discriminator.state_dict(), 'discriminator_state_dict.pth')\n" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "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": "3.11.6" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +}