Task1.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "950ec8d9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "from torchvision import datasets, transforms\n",
    "from torch.utils.data import DataLoader, random_split\n",
    "import numpy as np\n",
    "from torch.optim.lr_scheduler import StepLR\n",
    "#import torch.nn.functional as F\n",
    "\n",
    "torch.manual_seed(0)\n",
    "if torch.cuda.is_available():\n",
    "    torch.backends.cudnn.deterministic = True\n",
    "\n",
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2c8ed6bc",
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 64\n",
    "test_batch_size = 64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "9256b225",
   "metadata": {},
   "outputs": [],
   "source": [
    "img_transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize([0.1307], [0.3081])\n",
    "])\n",
    "\n",
    "train_dataset = datasets.MNIST('./data', train=True, download=True, transform=img_transform)\n",
    "train_dataset, val_dataset = random_split(train_dataset, [50000,10000])\n",
    "train_loader = DataLoader(train_dataset, batch_size=batch_size,\n",
    "                          shuffle=True, num_workers=2)\n",
    "val_loader = DataLoader(val_dataset, batch_size=test_batch_size, \n",
    "                        shuffle=False)\n",
    "\n",
    "test_dataset = datasets.MNIST('./data', train=False, download=True, transform=img_transform)\n",
    "test_loader = DataLoader(test_dataset, batch_size=test_batch_size,\n",
    "                             shuffle=True, num_workers=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "18d802e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "class autoencoder(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(autoencoder, self).__init__()\n",
    "        self.encoder = nn.Sequential(\n",
    "            nn.Linear(28 * 28, 128),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(128, 64),\n",
    "            nn.ReLU(True), \n",
    "            nn.Linear(64, 32), \n",
    "            nn.ReLU(True), \n",
    "            nn.Linear(32, 2))\n",
    "        \n",
    "        self.decoder = nn.Sequential(\n",
    "            nn.Linear(2, 32),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(32, 64),\n",
    "            nn.ReLU(True),\n",
    "            nn.Linear(64, 128),\n",
    "            nn.ReLU(True), \n",
    "            nn.Linear(128, 28 * 28), \n",
    "            nn.Tanh())\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.encoder(x)\n",
    "        x = self.decoder(x)\n",
    "        return x\n",
    "\n",
    "ae = autoencoder().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "84e99d90",
   "metadata": {},
   "outputs": [],
   "source": [
    "ae_criterion = nn.MSELoss()\n",
    "\n",
    "epochs = 1000\n",
    "learning_rate = 0.01\n",
    "#ae_optimizer = torch.optim.Adam(ae.parameters(), lr=0.001, weight_decay=learning_rate)\n",
    "ae_optimizer = torch.optim.SGD(ae.parameters(), lr=0.5, momentum=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "1cd355cc",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [1/1000], train loss:0.6733, val loss:0.6215\n",
      "Epoch [2/1000], train loss:0.6103, val loss:0.6074\n",
      "Epoch [3/1000], train loss:0.5978, val loss:0.5931\n",
      "Epoch [4/1000], train loss:0.5896, val loss:0.5872\n",
      "Epoch [5/1000], train loss:0.5827, val loss:0.5808\n",
      "Epoch [6/1000], train loss:0.5769, val loss:0.5751\n",
      "Epoch [7/1000], train loss:0.5724, val loss:0.5721\n",
      "Epoch [8/1000], train loss:0.5682, val loss:0.5670\n",
      "Epoch [9/1000], train loss:0.5646, val loss:0.5634\n",
      "Epoch [10/1000], train loss:0.5609, val loss:0.5605\n",
      "Epoch [11/1000], train loss:0.5590, val loss:0.5586\n",
      "Epoch [12/1000], train loss:0.5563, val loss:0.5598\n",
      "Epoch [13/1000], train loss:0.5546, val loss:0.5564\n",
      "Epoch [14/1000], train loss:0.5530, val loss:0.5543\n",
      "Epoch [15/1000], train loss:0.5511, val loss:0.5526\n",
      "Epoch [16/1000], train loss:0.5505, val loss:0.5532\n",
      "Epoch [17/1000], train loss:0.5488, val loss:0.5495\n",
      "Epoch [18/1000], train loss:0.5471, val loss:0.5554\n",
      "Epoch [19/1000], train loss:0.5465, val loss:0.5497\n",
      "Epoch [20/1000], train loss:0.5441, val loss:0.5475\n",
      "Epoch [21/1000], train loss:0.5427, val loss:0.5514\n",
      "Epoch [22/1000], train loss:0.5419, val loss:0.5488\n",
      "Epoch [23/1000], train loss:0.5406, val loss:0.5429\n",
      "Epoch [24/1000], train loss:0.5396, val loss:0.5450\n",
      "Epoch [25/1000], train loss:0.5403, val loss:0.5642\n",
      "Epoch [26/1000], train loss:0.5385, val loss:0.5429\n",
      "Epoch [27/1000], train loss:0.5371, val loss:0.5437\n",
      "Epoch [28/1000], train loss:0.5369, val loss:0.5389\n",
      "Epoch [29/1000], train loss:0.5370, val loss:0.5484\n",
      "Epoch [30/1000], train loss:0.5362, val loss:0.5424\n",
      "Epoch [31/1000], train loss:0.5361, val loss:0.5378\n",
      "Epoch [32/1000], train loss:0.5351, val loss:0.5530\n",
      "Epoch [33/1000], train loss:0.5347, val loss:0.5368\n",
      "Epoch [34/1000], train loss:0.5336, val loss:0.5363\n",
      "Epoch [35/1000], train loss:0.5330, val loss:0.5345\n",
      "Epoch [36/1000], train loss:0.5321, val loss:0.5343\n",
      "Epoch [37/1000], train loss:0.5303, val loss:0.5374\n",
      "Epoch [38/1000], train loss:0.5315, val loss:0.5326\n",
      "Epoch [39/1000], train loss:0.5304, val loss:0.5377\n",
      "Epoch [40/1000], train loss:0.5305, val loss:0.5317\n",
      "Epoch [41/1000], train loss:0.5298, val loss:0.5334\n",
      "Epoch [42/1000], train loss:0.5309, val loss:0.5337\n",
      "Epoch [43/1000], train loss:0.5302, val loss:0.5375\n",
      "Epoch [44/1000], train loss:0.5297, val loss:0.5360\n",
      "Epoch [45/1000], train loss:0.5291, val loss:0.5330\n",
      "Epoch [46/1000], train loss:0.5281, val loss:0.5340\n",
      "Epoch [47/1000], train loss:0.5273, val loss:0.5302\n",
      "Epoch [48/1000], train loss:0.5285, val loss:0.5397\n",
      "Epoch [49/1000], train loss:0.5269, val loss:0.5274\n",
      "Epoch [50/1000], train loss:0.5266, val loss:0.5271\n",
      "Epoch [51/1000], train loss:0.5261, val loss:0.5316\n",
      "Epoch [52/1000], train loss:0.5255, val loss:0.5288\n",
      "Epoch [53/1000], train loss:0.5256, val loss:0.5309\n",
      "Epoch [54/1000], train loss:0.5240, val loss:0.5260\n",
      "Epoch [55/1000], train loss:0.5249, val loss:0.5276\n",
      "Epoch [56/1000], train loss:0.5247, val loss:0.5263\n",
      "Epoch [57/1000], train loss:0.5232, val loss:0.5351\n",
      "Epoch [58/1000], train loss:0.5248, val loss:0.5328\n",
      "Epoch [59/1000], train loss:0.5248, val loss:0.5263\n",
      "Epoch [60/1000], train loss:0.5233, val loss:0.5469\n",
      "Epoch [61/1000], train loss:0.5261, val loss:0.5314\n",
      "Epoch [62/1000], train loss:0.5256, val loss:0.5292\n",
      "Epoch [63/1000], train loss:0.5270, val loss:0.5524\n",
      "Epoch [64/1000], train loss:0.5261, val loss:0.5275\n",
      "Epoch [65/1000], train loss:0.5246, val loss:0.5268\n",
      "Epoch [66/1000], train loss:0.5250, val loss:0.5299\n",
      "Epoch [67/1000], train loss:0.5240, val loss:0.5404\n",
      "Epoch [68/1000], train loss:0.5249, val loss:0.5257\n",
      "Epoch [69/1000], train loss:0.5228, val loss:0.5346\n",
      "Epoch [70/1000], train loss:0.5234, val loss:0.5261\n",
      "Epoch [71/1000], train loss:0.5230, val loss:0.5269\n",
      "Epoch [72/1000], train loss:0.5224, val loss:0.5284\n",
      "Epoch [73/1000], train loss:0.5232, val loss:0.5240\n",
      "Epoch [74/1000], train loss:0.5230, val loss:0.5257\n",
      "Epoch [75/1000], train loss:0.5221, val loss:0.5266\n",
      "Epoch [76/1000], train loss:0.5226, val loss:0.5309\n",
      "Epoch [77/1000], train loss:0.5224, val loss:0.5227\n",
      "Epoch [78/1000], train loss:0.5226, val loss:0.5311\n",
      "Epoch [79/1000], train loss:0.5239, val loss:0.5265\n",
      "Epoch [80/1000], train loss:0.5215, val loss:0.5233\n",
      "Epoch [81/1000], train loss:0.5229, val loss:0.5238\n",
      "Epoch [82/1000], train loss:0.5214, val loss:0.5260\n",
      "Epoch [83/1000], train loss:0.5200, val loss:0.5234\n",
      "Epoch [84/1000], train loss:0.5183, val loss:0.5234\n",
      "Epoch [85/1000], train loss:0.5190, val loss:0.5199\n",
      "Epoch [86/1000], train loss:0.5207, val loss:0.5224\n",
      "Epoch [87/1000], train loss:0.5186, val loss:0.5229\n",
      "Epoch [88/1000], train loss:0.5201, val loss:0.5209\n",
      "Epoch [89/1000], train loss:0.5188, val loss:0.5212\n",
      "Epoch [90/1000], train loss:0.5198, val loss:0.5282\n",
      "Epoch [91/1000], train loss:0.5193, val loss:0.5254\n",
      "Epoch [92/1000], train loss:0.5196, val loss:0.5202\n",
      "Epoch [93/1000], train loss:0.5178, val loss:0.5233\n",
      "Epoch [94/1000], train loss:0.5182, val loss:0.5232\n",
      "Epoch [95/1000], train loss:0.5176, val loss:0.5214\n",
      "Epoch [96/1000], train loss:0.5187, val loss:0.5247\n",
      "Epoch [97/1000], train loss:0.5212, val loss:0.5234\n",
      "Epoch [98/1000], train loss:0.5202, val loss:0.5219\n",
      "Epoch [99/1000], train loss:0.5206, val loss:0.5224\n",
      "Epoch [100/1000], train loss:0.5183, val loss:0.5215\n",
      "Epoch [101/1000], train loss:0.5183, val loss:0.5190\n",
      "Epoch [102/1000], train loss:0.5188, val loss:0.5315\n",
      "Epoch [103/1000], train loss:0.5184, val loss:0.5242\n",
      "Epoch [104/1000], train loss:0.5176, val loss:0.5194\n",
      "Epoch [105/1000], train loss:0.5174, val loss:0.5223\n",
      "Epoch [106/1000], train loss:0.5182, val loss:0.5219\n",
      "Epoch [107/1000], train loss:0.5168, val loss:0.5193\n",
      "Epoch [108/1000], train loss:0.5165, val loss:0.5170\n",
      "Epoch [109/1000], train loss:0.5175, val loss:0.5319\n",
      "Epoch [110/1000], train loss:0.5195, val loss:0.5197\n",
      "Epoch [111/1000], train loss:0.5197, val loss:0.5182\n",
      "Epoch [112/1000], train loss:0.5162, val loss:0.5265\n",
      "Epoch [113/1000], train loss:0.5195, val loss:0.5292\n",
      "Epoch [114/1000], train loss:0.5164, val loss:0.5214\n",
      "Epoch [115/1000], train loss:0.5172, val loss:0.5174\n",
      "Epoch [116/1000], train loss:0.5156, val loss:0.5177\n",
      "Epoch [117/1000], train loss:0.5143, val loss:0.5182\n",
      "Epoch [118/1000], train loss:0.5169, val loss:0.5172\n",
      "Epoch [119/1000], train loss:0.5169, val loss:0.5203\n",
      "Epoch [120/1000], train loss:0.5159, val loss:0.5198\n",
      "Epoch [121/1000], train loss:0.5148, val loss:0.5300\n",
      "Epoch [122/1000], train loss:0.5136, val loss:0.5178\n",
      "Epoch [123/1000], train loss:0.5124, val loss:0.5153\n",
      "Epoch [124/1000], train loss:0.5116, val loss:0.5177\n",
      "Epoch [125/1000], train loss:0.5124, val loss:0.5155\n",
      "Epoch [126/1000], train loss:0.5117, val loss:0.5211\n",
      "Epoch [127/1000], train loss:0.5110, val loss:0.5151\n",
      "Epoch [128/1000], train loss:0.5116, val loss:0.5149\n",
      "Epoch [129/1000], train loss:0.5113, val loss:0.5149\n",
      "Epoch [130/1000], train loss:0.5113, val loss:0.5170\n",
      "Epoch [131/1000], train loss:0.5109, val loss:0.5151\n",
      "Epoch [132/1000], train loss:0.5131, val loss:0.5357\n",
      "Epoch [133/1000], train loss:0.5121, val loss:0.5149\n",
      "Epoch [134/1000], train loss:0.5112, val loss:0.5227\n",
      "Epoch [135/1000], train loss:0.5114, val loss:0.5149\n",
      "Epoch [136/1000], train loss:0.5126, val loss:0.5133\n",
      "Epoch [137/1000], train loss:0.5124, val loss:0.5134\n",
      "Epoch [138/1000], train loss:0.5122, val loss:0.5159\n",
      "Epoch [139/1000], train loss:0.5097, val loss:0.5128\n",
      "Epoch [140/1000], train loss:0.5105, val loss:0.5173\n",
      "Epoch [141/1000], train loss:0.5101, val loss:0.5154\n",
      "Epoch [142/1000], train loss:0.5126, val loss:0.5217\n",
      "Epoch [143/1000], train loss:0.5130, val loss:0.5267\n",
      "Epoch [144/1000], train loss:0.5127, val loss:0.5299\n",
      "Epoch [145/1000], train loss:0.5113, val loss:0.5146\n",
      "Epoch [146/1000], train loss:0.5102, val loss:0.5179\n",
      "Epoch [147/1000], train loss:0.5114, val loss:0.5230\n",
      "Epoch [148/1000], train loss:0.5091, val loss:0.5108\n",
      "Epoch [149/1000], train loss:0.5084, val loss:0.5161\n",
      "Epoch [150/1000], train loss:0.5089, val loss:0.5163\n",
      "Epoch [151/1000], train loss:0.5096, val loss:0.5151\n",
      "Epoch [152/1000], train loss:0.5091, val loss:0.5120\n",
      "Epoch [153/1000], train loss:0.5078, val loss:0.5173\n",
      "Epoch [154/1000], train loss:0.5078, val loss:0.5152\n",
      "Epoch [155/1000], train loss:0.5082, val loss:0.5140\n",
      "Epoch [156/1000], train loss:0.5073, val loss:0.5104\n",
      "Epoch [157/1000], train loss:0.5100, val loss:0.5220\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [158/1000], train loss:0.5119, val loss:0.5275\n",
      "Epoch [159/1000], train loss:0.5107, val loss:0.5306\n",
      "Epoch [160/1000], train loss:0.5145, val loss:0.5171\n",
      "Epoch [161/1000], train loss:0.5120, val loss:0.5189\n",
      "Epoch [162/1000], train loss:0.5137, val loss:0.5322\n",
      "Epoch [163/1000], train loss:0.5106, val loss:0.5112\n",
      "Epoch [164/1000], train loss:0.5089, val loss:0.5188\n",
      "Epoch [165/1000], train loss:0.5081, val loss:0.5135\n",
      "Epoch [166/1000], train loss:0.5079, val loss:0.5124\n",
      "Epoch [167/1000], train loss:0.5082, val loss:0.5192\n",
      "Epoch [168/1000], train loss:0.5104, val loss:0.5120\n",
      "Epoch [169/1000], train loss:0.5115, val loss:0.5170\n",
      "Epoch [170/1000], train loss:0.5099, val loss:0.5121\n",
      "Epoch [171/1000], train loss:0.5083, val loss:0.5105\n",
      "Epoch [172/1000], train loss:0.5078, val loss:0.5110\n",
      "Epoch [173/1000], train loss:0.5084, val loss:0.5301\n",
      "Epoch [174/1000], train loss:0.5124, val loss:0.5142\n",
      "Epoch [175/1000], train loss:0.5089, val loss:0.5130\n",
      "Epoch [176/1000], train loss:0.5094, val loss:0.5132\n",
      "Epoch [177/1000], train loss:0.5093, val loss:0.5112\n",
      "Epoch [178/1000], train loss:0.5074, val loss:0.5254\n",
      "Epoch [179/1000], train loss:0.5068, val loss:0.5118\n",
      "Epoch [180/1000], train loss:0.5055, val loss:0.5170\n",
      "Epoch [181/1000], train loss:0.5059, val loss:0.5122\n",
      "Epoch [182/1000], train loss:0.5062, val loss:0.5114\n",
      "Epoch [183/1000], train loss:0.5056, val loss:0.5124\n",
      "Epoch [184/1000], train loss:0.5053, val loss:0.5077\n",
      "Epoch [185/1000], train loss:0.5073, val loss:0.5149\n",
      "Epoch [186/1000], train loss:0.5071, val loss:0.5167\n",
      "Epoch [187/1000], train loss:0.5063, val loss:0.5151\n",
      "Epoch [188/1000], train loss:0.5066, val loss:0.5150\n",
      "Epoch [189/1000], train loss:0.5074, val loss:0.5106\n",
      "Epoch [190/1000], train loss:0.5081, val loss:0.5222\n",
      "Epoch [191/1000], train loss:0.5070, val loss:0.5096\n",
      "Epoch [192/1000], train loss:0.5062, val loss:0.5076\n",
      "Epoch [193/1000], train loss:0.5058, val loss:0.5114\n",
      "Epoch [194/1000], train loss:0.5068, val loss:0.5235\n",
      "Epoch [195/1000], train loss:0.5107, val loss:0.5121\n",
      "Epoch [196/1000], train loss:0.5091, val loss:0.5170\n",
      "Epoch [197/1000], train loss:0.5077, val loss:0.5130\n",
      "Epoch [198/1000], train loss:0.5056, val loss:0.5137\n",
      "Epoch [199/1000], train loss:0.5047, val loss:0.5160\n",
      "Epoch [200/1000], train loss:0.5048, val loss:0.5114\n",
      "Epoch [201/1000], train loss:0.5055, val loss:0.5175\n",
      "Epoch [202/1000], train loss:0.5039, val loss:0.5076\n",
      "Epoch [203/1000], train loss:0.5057, val loss:0.5074\n",
      "Epoch [204/1000], train loss:0.5046, val loss:0.5108\n",
      "Epoch [205/1000], train loss:0.5050, val loss:0.5101\n",
      "Epoch [206/1000], train loss:0.5035, val loss:0.5075\n",
      "Epoch [207/1000], train loss:0.5063, val loss:0.5125\n",
      "Epoch [208/1000], train loss:0.5059, val loss:0.5154\n",
      "Epoch [209/1000], train loss:0.5049, val loss:0.5057\n",
      "Epoch [210/1000], train loss:0.5057, val loss:0.5167\n",
      "Epoch [211/1000], train loss:0.5046, val loss:0.5079\n",
      "Epoch [212/1000], train loss:0.5038, val loss:0.5083\n",
      "Epoch [213/1000], train loss:0.5058, val loss:0.5169\n",
      "Epoch [214/1000], train loss:0.5061, val loss:0.5112\n",
      "Epoch [215/1000], train loss:0.5055, val loss:0.5094\n",
      "Epoch [216/1000], train loss:0.5053, val loss:0.5070\n",
      "Epoch [217/1000], train loss:0.5070, val loss:0.5115\n",
      "Epoch [218/1000], train loss:0.5037, val loss:0.5108\n",
      "Epoch [219/1000], train loss:0.5025, val loss:0.5073\n",
      "Epoch [220/1000], train loss:0.5028, val loss:0.5059\n",
      "Epoch [221/1000], train loss:0.5022, val loss:0.5092\n",
      "Epoch [222/1000], train loss:0.5017, val loss:0.5045\n",
      "Epoch [223/1000], train loss:0.5024, val loss:0.5080\n",
      "Epoch [224/1000], train loss:0.5030, val loss:0.5078\n",
      "Epoch [225/1000], train loss:0.5043, val loss:0.5080\n",
      "Epoch [226/1000], train loss:0.5027, val loss:0.5035\n",
      "Epoch [227/1000], train loss:0.5030, val loss:0.5199\n",
      "Epoch [228/1000], train loss:0.5037, val loss:0.5045\n",
      "Epoch [229/1000], train loss:0.5011, val loss:0.5031\n",
      "Epoch [230/1000], train loss:0.5016, val loss:0.5035\n",
      "Epoch [231/1000], train loss:0.5019, val loss:0.5124\n",
      "Epoch [232/1000], train loss:0.5026, val loss:0.5052\n",
      "Epoch [233/1000], train loss:0.5043, val loss:0.5065\n",
      "Epoch [234/1000], train loss:0.5028, val loss:0.5050\n",
      "Epoch [235/1000], train loss:0.5020, val loss:0.5090\n",
      "Epoch [236/1000], train loss:0.5024, val loss:0.5089\n",
      "Epoch [237/1000], train loss:0.5035, val loss:0.5058\n",
      "Epoch [238/1000], train loss:0.5016, val loss:0.5113\n",
      "Epoch [239/1000], train loss:0.5025, val loss:0.5086\n",
      "Epoch [240/1000], train loss:0.5018, val loss:0.5047\n",
      "Epoch [241/1000], train loss:0.5029, val loss:0.5191\n",
      "Epoch [242/1000], train loss:0.5030, val loss:0.5132\n",
      "Epoch [243/1000], train loss:0.5025, val loss:0.5045\n",
      "Epoch [244/1000], train loss:0.5009, val loss:0.5121\n",
      "Epoch [245/1000], train loss:0.5019, val loss:0.5073\n",
      "Epoch [246/1000], train loss:0.5029, val loss:0.5063\n",
      "Epoch [247/1000], train loss:0.5018, val loss:0.5060\n",
      "Epoch [248/1000], train loss:0.5044, val loss:0.5071\n",
      "Epoch [249/1000], train loss:0.5041, val loss:0.5091\n",
      "Epoch [250/1000], train loss:0.5026, val loss:0.5090\n",
      "Epoch [251/1000], train loss:0.5057, val loss:0.5151\n",
      "Epoch [252/1000], train loss:0.5036, val loss:0.5151\n",
      "Epoch [253/1000], train loss:0.5020, val loss:0.5065\n",
      "Epoch [254/1000], train loss:0.5028, val loss:0.5152\n",
      "Epoch [255/1000], train loss:0.5038, val loss:0.5087\n",
      "Epoch [256/1000], train loss:0.5022, val loss:0.5089\n",
      "Epoch [257/1000], train loss:0.5016, val loss:0.5252\n",
      "Epoch [258/1000], train loss:0.5012, val loss:0.5234\n",
      "Epoch [259/1000], train loss:0.5023, val loss:0.5164\n",
      "Epoch [260/1000], train loss:0.5021, val loss:0.5058\n",
      "Epoch [261/1000], train loss:0.5008, val loss:0.5069\n",
      "Epoch [262/1000], train loss:0.5010, val loss:0.5046\n",
      "Epoch [263/1000], train loss:0.5010, val loss:0.5085\n",
      "Epoch [264/1000], train loss:0.5019, val loss:0.5076\n",
      "Epoch [265/1000], train loss:0.5037, val loss:0.5084\n",
      "Epoch [266/1000], train loss:0.5019, val loss:0.5020\n",
      "Epoch [267/1000], train loss:0.4996, val loss:0.5085\n",
      "Epoch [268/1000], train loss:0.5009, val loss:0.5049\n",
      "Epoch [269/1000], train loss:0.5009, val loss:0.5045\n",
      "Epoch [270/1000], train loss:0.4996, val loss:0.5050\n",
      "Epoch [271/1000], train loss:0.5001, val loss:0.5089\n",
      "Epoch [272/1000], train loss:0.5024, val loss:0.5203\n",
      "Epoch [273/1000], train loss:0.5027, val loss:0.5330\n",
      "Epoch [274/1000], train loss:0.5027, val loss:0.5175\n",
      "Epoch [275/1000], train loss:0.5015, val loss:0.5213\n",
      "Epoch [276/1000], train loss:0.5003, val loss:0.5028\n",
      "Epoch [277/1000], train loss:0.5003, val loss:0.5191\n",
      "Epoch [278/1000], train loss:0.5006, val loss:0.5023\n",
      "Epoch [279/1000], train loss:0.5007, val loss:0.5037\n",
      "Epoch [280/1000], train loss:0.5025, val loss:0.5074\n",
      "Epoch [281/1000], train loss:0.5019, val loss:0.5082\n",
      "Epoch [282/1000], train loss:0.5007, val loss:0.5130\n",
      "Epoch [283/1000], train loss:0.4987, val loss:0.5016\n",
      "Epoch [284/1000], train loss:0.4997, val loss:0.5160\n",
      "Epoch [285/1000], train loss:0.5001, val loss:0.5077\n",
      "Epoch [286/1000], train loss:0.5013, val loss:0.5102\n",
      "Epoch [287/1000], train loss:0.5015, val loss:0.5039\n",
      "Epoch [288/1000], train loss:0.4978, val loss:0.5071\n",
      "Epoch [289/1000], train loss:0.4980, val loss:0.5044\n",
      "Epoch [290/1000], train loss:0.4981, val loss:0.5048\n",
      "Epoch [291/1000], train loss:0.4986, val loss:0.5067\n",
      "Epoch [292/1000], train loss:0.5006, val loss:0.5066\n",
      "Epoch [293/1000], train loss:0.5002, val loss:0.5113\n",
      "Epoch [294/1000], train loss:0.4999, val loss:0.5000\n",
      "Epoch [295/1000], train loss:0.5012, val loss:0.5103\n",
      "Epoch [296/1000], train loss:0.4998, val loss:0.5084\n",
      "Epoch [297/1000], train loss:0.4986, val loss:0.5035\n",
      "Epoch [298/1000], train loss:0.5016, val loss:0.5032\n",
      "Epoch [299/1000], train loss:0.5010, val loss:0.5050\n",
      "Epoch [300/1000], train loss:0.5021, val loss:0.5074\n",
      "Epoch [301/1000], train loss:0.4998, val loss:0.5043\n",
      "Epoch [302/1000], train loss:0.4999, val loss:0.5137\n",
      "Epoch [303/1000], train loss:0.4985, val loss:0.5104\n",
      "Epoch [304/1000], train loss:0.4989, val loss:0.5005\n",
      "Epoch [305/1000], train loss:0.5002, val loss:0.5131\n",
      "Epoch [306/1000], train loss:0.5002, val loss:0.5022\n",
      "Epoch [307/1000], train loss:0.4977, val loss:0.5002\n",
      "Epoch [308/1000], train loss:0.4974, val loss:0.5021\n",
      "Epoch [309/1000], train loss:0.4978, val loss:0.5032\n",
      "Epoch [310/1000], train loss:0.4984, val loss:0.5074\n",
      "Epoch [311/1000], train loss:0.4982, val loss:0.5013\n",
      "Epoch [312/1000], train loss:0.5016, val loss:0.5072\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [313/1000], train loss:0.5015, val loss:0.5024\n",
      "Epoch [314/1000], train loss:0.4986, val loss:0.5121\n",
      "Epoch [315/1000], train loss:0.4993, val loss:0.5057\n",
      "Epoch [316/1000], train loss:0.4982, val loss:0.5009\n",
      "Epoch [317/1000], train loss:0.4983, val loss:0.5171\n",
      "Epoch [318/1000], train loss:0.4966, val loss:0.5161\n",
      "Epoch [319/1000], train loss:0.4966, val loss:0.5016\n",
      "Epoch [320/1000], train loss:0.4979, val loss:0.5007\n",
      "Epoch [321/1000], train loss:0.4972, val loss:0.5031\n",
      "Epoch [322/1000], train loss:0.4983, val loss:0.5113\n",
      "Epoch [323/1000], train loss:0.5019, val loss:0.5166\n",
      "Epoch [324/1000], train loss:0.5007, val loss:0.5099\n",
      "Epoch [325/1000], train loss:0.5031, val loss:0.5073\n",
      "Epoch [326/1000], train loss:0.5006, val loss:0.5081\n",
      "Epoch [327/1000], train loss:0.5009, val loss:0.5133\n",
      "Epoch [328/1000], train loss:0.5011, val loss:0.5254\n",
      "Epoch [329/1000], train loss:0.4992, val loss:0.5042\n",
      "Epoch [330/1000], train loss:0.4981, val loss:0.5019\n",
      "Epoch [331/1000], train loss:0.4987, val loss:0.5038\n",
      "Epoch [332/1000], train loss:0.4982, val loss:0.5053\n",
      "Epoch [333/1000], train loss:0.4985, val loss:0.5049\n",
      "Epoch [334/1000], train loss:0.4977, val loss:0.5001\n",
      "Epoch [335/1000], train loss:0.4970, val loss:0.5089\n",
      "Epoch [336/1000], train loss:0.4961, val loss:0.5014\n",
      "Epoch [337/1000], train loss:0.4967, val loss:0.5035\n",
      "Epoch [338/1000], train loss:0.5008, val loss:0.5012\n",
      "Epoch [339/1000], train loss:0.5028, val loss:0.5289\n",
      "Epoch [340/1000], train loss:0.5033, val loss:0.5078\n",
      "Epoch [341/1000], train loss:0.4993, val loss:0.5009\n",
      "Epoch [342/1000], train loss:0.4990, val loss:0.5080\n",
      "Epoch [343/1000], train loss:0.4997, val loss:0.5013\n",
      "Epoch [344/1000], train loss:0.4981, val loss:0.5032\n",
      "Epoch [345/1000], train loss:0.4993, val loss:0.5088\n",
      "Epoch [346/1000], train loss:0.4965, val loss:0.5034\n",
      "Epoch [347/1000], train loss:0.4967, val loss:0.4995\n",
      "Epoch [348/1000], train loss:0.4976, val loss:0.5086\n",
      "Epoch [349/1000], train loss:0.4966, val loss:0.5025\n",
      "Epoch [350/1000], train loss:0.4958, val loss:0.5060\n",
      "Epoch [351/1000], train loss:0.4956, val loss:0.5063\n",
      "Epoch [352/1000], train loss:0.4964, val loss:0.5088\n",
      "Epoch [353/1000], train loss:0.4963, val loss:0.4987\n",
      "Epoch [354/1000], train loss:0.4953, val loss:0.5001\n",
      "Epoch [355/1000], train loss:0.4975, val loss:0.5107\n",
      "Epoch [356/1000], train loss:0.4969, val loss:0.5216\n",
      "Epoch [357/1000], train loss:0.4966, val loss:0.4975\n",
      "Epoch [358/1000], train loss:0.4963, val loss:0.5031\n",
      "Epoch [359/1000], train loss:0.4953, val loss:0.5010\n",
      "Epoch [360/1000], train loss:0.4958, val loss:0.5034\n",
      "Epoch [361/1000], train loss:0.4951, val loss:0.4983\n",
      "Epoch [362/1000], train loss:0.4966, val loss:0.5274\n",
      "Epoch [363/1000], train loss:0.4976, val loss:0.5016\n",
      "Epoch [364/1000], train loss:0.4963, val loss:0.5062\n",
      "Epoch [365/1000], train loss:0.4972, val loss:0.5040\n",
      "Epoch [366/1000], train loss:0.4963, val loss:0.5027\n",
      "Epoch [367/1000], train loss:0.4967, val loss:0.5070\n",
      "Epoch [368/1000], train loss:0.4950, val loss:0.5078\n",
      "Epoch [369/1000], train loss:0.4950, val loss:0.4998\n",
      "Epoch [370/1000], train loss:0.4947, val loss:0.4984\n",
      "Epoch [371/1000], train loss:0.4949, val loss:0.5096\n",
      "Epoch [372/1000], train loss:0.4946, val loss:0.4965\n",
      "Epoch [373/1000], train loss:0.4945, val loss:0.5013\n",
      "Epoch [374/1000], train loss:0.4941, val loss:0.5034\n",
      "Epoch [375/1000], train loss:0.4951, val loss:0.5024\n",
      "Epoch [376/1000], train loss:0.4958, val loss:0.5007\n",
      "Epoch [377/1000], train loss:0.4965, val loss:0.5047\n",
      "Epoch [378/1000], train loss:0.4951, val loss:0.5025\n",
      "Epoch [379/1000], train loss:0.4957, val loss:0.5019\n",
      "Epoch [380/1000], train loss:0.4964, val loss:0.5022\n",
      "Epoch [381/1000], train loss:0.4948, val loss:0.5085\n",
      "Epoch [382/1000], train loss:0.4944, val loss:0.4990\n",
      "Epoch [383/1000], train loss:0.4938, val loss:0.4990\n",
      "Epoch [384/1000], train loss:0.4944, val loss:0.5045\n",
      "Epoch [385/1000], train loss:0.4939, val loss:0.4987\n",
      "Epoch [386/1000], train loss:0.4954, val loss:0.4999\n",
      "Epoch [387/1000], train loss:0.4942, val loss:0.4990\n",
      "Epoch [388/1000], train loss:0.4931, val loss:0.5050\n",
      "Epoch [389/1000], train loss:0.4945, val loss:0.5014\n",
      "Epoch [390/1000], train loss:0.4967, val loss:0.5020\n",
      "Epoch [391/1000], train loss:0.4957, val loss:0.5089\n",
      "Epoch [392/1000], train loss:0.4956, val loss:0.5000\n",
      "Epoch [393/1000], train loss:0.4953, val loss:0.5020\n",
      "Epoch [394/1000], train loss:0.4955, val loss:0.5041\n",
      "Epoch [395/1000], train loss:0.4983, val loss:0.5062\n",
      "Epoch [396/1000], train loss:0.4963, val loss:0.4988\n",
      "Epoch [397/1000], train loss:0.4960, val loss:0.5035\n",
      "Epoch [398/1000], train loss:0.4980, val loss:0.5210\n",
      "Epoch [399/1000], train loss:0.4985, val loss:0.5100\n",
      "Epoch [400/1000], train loss:0.4978, val loss:0.5029\n",
      "Epoch [401/1000], train loss:0.4979, val loss:0.5009\n",
      "Epoch [402/1000], train loss:0.4961, val loss:0.5009\n",
      "Epoch [403/1000], train loss:0.4957, val loss:0.5014\n",
      "Epoch [404/1000], train loss:0.4946, val loss:0.4989\n",
      "Epoch [405/1000], train loss:0.4940, val loss:0.4989\n",
      "Epoch [406/1000], train loss:0.4935, val loss:0.5042\n",
      "Epoch [407/1000], train loss:0.4944, val loss:0.5028\n",
      "Epoch [408/1000], train loss:0.4939, val loss:0.4975\n",
      "Epoch [409/1000], train loss:0.4946, val loss:0.5028\n",
      "Epoch [410/1000], train loss:0.4934, val loss:0.5043\n",
      "Epoch [411/1000], train loss:0.4948, val loss:0.5208\n",
      "Epoch [412/1000], train loss:0.4976, val loss:0.5050\n",
      "Epoch [413/1000], train loss:0.4952, val loss:0.4977\n",
      "Epoch [414/1000], train loss:0.4930, val loss:0.5084\n",
      "Epoch [415/1000], train loss:0.4935, val loss:0.5029\n",
      "Epoch [416/1000], train loss:0.4927, val loss:0.5059\n",
      "Epoch [417/1000], train loss:0.4938, val loss:0.5029\n",
      "Epoch [418/1000], train loss:0.4960, val loss:0.5134\n",
      "Epoch [419/1000], train loss:0.4944, val loss:0.5072\n",
      "Epoch [420/1000], train loss:0.4943, val loss:0.4971\n",
      "Epoch [421/1000], train loss:0.4930, val loss:0.5009\n",
      "Epoch [422/1000], train loss:0.4934, val loss:0.5030\n",
      "Epoch [423/1000], train loss:0.4937, val loss:0.4986\n",
      "Epoch [424/1000], train loss:0.4922, val loss:0.5014\n",
      "Epoch [425/1000], train loss:0.4926, val loss:0.4988\n",
      "Epoch [426/1000], train loss:0.4928, val loss:0.4998\n",
      "Epoch [427/1000], train loss:0.4929, val loss:0.4956\n",
      "Epoch [428/1000], train loss:0.4923, val loss:0.4999\n",
      "Epoch [429/1000], train loss:0.4936, val loss:0.4976\n",
      "Epoch [430/1000], train loss:0.4930, val loss:0.4942\n",
      "Epoch [431/1000], train loss:0.4921, val loss:0.4974\n",
      "Epoch [432/1000], train loss:0.4945, val loss:0.5064\n",
      "Epoch [433/1000], train loss:0.4942, val loss:0.4960\n",
      "Epoch [434/1000], train loss:0.4937, val loss:0.5058\n",
      "Epoch [435/1000], train loss:0.4922, val loss:0.5186\n",
      "Epoch [436/1000], train loss:0.4940, val loss:0.4970\n",
      "Epoch [437/1000], train loss:0.4925, val loss:0.4975\n",
      "Epoch [438/1000], train loss:0.4929, val loss:0.4982\n",
      "Epoch [439/1000], train loss:0.4929, val loss:0.5029\n",
      "Epoch [440/1000], train loss:0.4927, val loss:0.4997\n",
      "Epoch [441/1000], train loss:0.4922, val loss:0.4967\n",
      "Epoch [442/1000], train loss:0.4935, val loss:0.5005\n",
      "Epoch [443/1000], train loss:0.4937, val loss:0.5124\n",
      "Epoch [444/1000], train loss:0.4920, val loss:0.4949\n",
      "Epoch [445/1000], train loss:0.4921, val loss:0.4962\n",
      "Epoch [446/1000], train loss:0.4916, val loss:0.4963\n",
      "Epoch [447/1000], train loss:0.4933, val loss:0.4962\n",
      "Epoch [448/1000], train loss:0.4940, val loss:0.5034\n",
      "Epoch [449/1000], train loss:0.4958, val loss:0.5023\n",
      "Epoch [450/1000], train loss:0.4944, val loss:0.5066\n",
      "Epoch [451/1000], train loss:0.4959, val loss:0.4987\n",
      "Epoch [452/1000], train loss:0.4954, val loss:0.5047\n",
      "Epoch [453/1000], train loss:0.4947, val loss:0.5023\n",
      "Epoch [454/1000], train loss:0.4938, val loss:0.4993\n",
      "Epoch [455/1000], train loss:0.4943, val loss:0.4960\n",
      "Epoch [456/1000], train loss:0.4926, val loss:0.5045\n",
      "Epoch [457/1000], train loss:0.4933, val loss:0.4968\n",
      "Epoch [458/1000], train loss:0.4942, val loss:0.5045\n",
      "Epoch [459/1000], train loss:0.4950, val loss:0.4965\n",
      "Epoch [460/1000], train loss:0.4966, val loss:0.5012\n",
      "Epoch [461/1000], train loss:0.4956, val loss:0.5057\n",
      "Epoch [462/1000], train loss:0.4946, val loss:0.4955\n",
      "Epoch [463/1000], train loss:0.4929, val loss:0.4978\n",
      "Epoch [464/1000], train loss:0.4926, val loss:0.4973\n",
      "Epoch [465/1000], train loss:0.4927, val loss:0.4990\n",
      "Epoch [466/1000], train loss:0.4933, val loss:0.5018\n",
      "Epoch [467/1000], train loss:0.4916, val loss:0.5029\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [468/1000], train loss:0.4925, val loss:0.4991\n",
      "Epoch [469/1000], train loss:0.4931, val loss:0.4998\n",
      "Epoch [470/1000], train loss:0.4956, val loss:0.5077\n",
      "Epoch [471/1000], train loss:0.4940, val loss:0.4988\n",
      "Epoch [472/1000], train loss:0.4922, val loss:0.4989\n",
      "Epoch [473/1000], train loss:0.4923, val loss:0.5254\n",
      "Epoch [474/1000], train loss:0.4921, val loss:0.5015\n",
      "Epoch [475/1000], train loss:0.4926, val loss:0.5043\n",
      "Epoch [476/1000], train loss:0.4948, val loss:0.4983\n",
      "Epoch [477/1000], train loss:0.4937, val loss:0.4956\n",
      "Epoch [478/1000], train loss:0.4919, val loss:0.5037\n",
      "Epoch [479/1000], train loss:0.4936, val loss:0.5079\n",
      "Epoch [480/1000], train loss:0.4915, val loss:0.4968\n",
      "Epoch [481/1000], train loss:0.4938, val loss:0.4997\n",
      "Epoch [482/1000], train loss:0.4943, val loss:0.5024\n",
      "Epoch [483/1000], train loss:0.4919, val loss:0.4959\n",
      "Epoch [484/1000], train loss:0.4929, val loss:0.5121\n",
      "Epoch [485/1000], train loss:0.4947, val loss:0.5004\n",
      "Epoch [486/1000], train loss:0.4924, val loss:0.5036\n",
      "Epoch [487/1000], train loss:0.4911, val loss:0.4957\n",
      "Epoch [488/1000], train loss:0.4908, val loss:0.4973\n",
      "Epoch [489/1000], train loss:0.4901, val loss:0.4952\n",
      "Epoch [490/1000], train loss:0.4934, val loss:0.4958\n",
      "Epoch [491/1000], train loss:0.4931, val loss:0.5009\n",
      "Epoch [492/1000], train loss:0.4915, val loss:0.4979\n",
      "Epoch [493/1000], train loss:0.4912, val loss:0.4990\n",
      "Epoch [494/1000], train loss:0.4912, val loss:0.4969\n",
      "Epoch [495/1000], train loss:0.4923, val loss:0.4958\n",
      "Epoch [496/1000], train loss:0.4929, val loss:0.4947\n",
      "Epoch [497/1000], train loss:0.4910, val loss:0.4993\n",
      "Epoch [498/1000], train loss:0.4914, val loss:0.4971\n",
      "Epoch [499/1000], train loss:0.4901, val loss:0.4994\n",
      "Epoch [500/1000], train loss:0.4905, val loss:0.4963\n",
      "Epoch [501/1000], train loss:0.4922, val loss:0.5014\n",
      "Epoch [502/1000], train loss:0.4933, val loss:0.5041\n",
      "Epoch [503/1000], train loss:0.4946, val loss:0.4999\n",
      "Epoch [504/1000], train loss:0.4948, val loss:0.5006\n",
      "Epoch [505/1000], train loss:0.4932, val loss:0.4978\n",
      "Epoch [506/1000], train loss:0.4931, val loss:0.4974\n",
      "Epoch [507/1000], train loss:0.4930, val loss:0.4946\n",
      "Epoch [508/1000], train loss:0.4937, val loss:0.4956\n",
      "Epoch [509/1000], train loss:0.4933, val loss:0.5086\n",
      "Epoch [510/1000], train loss:0.5021, val loss:0.5070\n",
      "Epoch [511/1000], train loss:0.5011, val loss:0.5029\n",
      "Epoch [512/1000], train loss:0.4966, val loss:0.5101\n",
      "Epoch [513/1000], train loss:0.4951, val loss:0.4969\n",
      "Epoch [514/1000], train loss:0.4952, val loss:0.4995\n",
      "Epoch [515/1000], train loss:0.4981, val loss:0.4947\n",
      "Epoch [516/1000], train loss:0.4953, val loss:0.5170\n",
      "Epoch [517/1000], train loss:0.4938, val loss:0.4985\n",
      "Epoch [518/1000], train loss:0.4933, val loss:0.5041\n",
      "Epoch [519/1000], train loss:0.4945, val loss:0.5006\n",
      "Epoch [520/1000], train loss:0.4923, val loss:0.4946\n",
      "Epoch [521/1000], train loss:0.4913, val loss:0.4948\n",
      "Epoch [522/1000], train loss:0.4905, val loss:0.4954\n",
      "Epoch [523/1000], train loss:0.4898, val loss:0.4975\n",
      "Epoch [524/1000], train loss:0.4913, val loss:0.4956\n",
      "Epoch [525/1000], train loss:0.4916, val loss:0.5028\n",
      "Epoch [526/1000], train loss:0.4960, val loss:0.4964\n",
      "Epoch [527/1000], train loss:0.4969, val loss:0.5024\n",
      "Epoch [528/1000], train loss:0.5032, val loss:0.5308\n",
      "Epoch [529/1000], train loss:0.4982, val loss:0.5001\n",
      "Epoch [530/1000], train loss:0.4963, val loss:0.5040\n",
      "Epoch [531/1000], train loss:0.4982, val loss:0.5109\n",
      "Epoch [532/1000], train loss:0.4999, val loss:0.5102\n",
      "Epoch [533/1000], train loss:0.4973, val loss:0.5001\n",
      "Epoch [534/1000], train loss:0.4973, val loss:0.5047\n",
      "Epoch [535/1000], train loss:0.4953, val loss:0.4982\n",
      "Epoch [536/1000], train loss:0.4960, val loss:0.4990\n",
      "Epoch [537/1000], train loss:0.4984, val loss:0.5031\n",
      "Epoch [538/1000], train loss:0.5005, val loss:0.4982\n",
      "Epoch [539/1000], train loss:0.4972, val loss:0.5122\n",
      "Epoch [540/1000], train loss:0.5013, val loss:0.5003\n",
      "Epoch [541/1000], train loss:0.4967, val loss:0.5108\n",
      "Epoch [542/1000], train loss:0.4962, val loss:0.5075\n",
      "Epoch [543/1000], train loss:0.4956, val loss:0.4988\n",
      "Epoch [544/1000], train loss:0.4936, val loss:0.5066\n",
      "Epoch [545/1000], train loss:0.4958, val loss:0.5000\n",
      "Epoch [546/1000], train loss:0.4952, val loss:0.4975\n",
      "Epoch [547/1000], train loss:0.5007, val loss:0.5154\n",
      "Epoch [548/1000], train loss:0.5007, val loss:0.4987\n",
      "Epoch [549/1000], train loss:0.4965, val loss:0.5070\n",
      "Epoch [550/1000], train loss:0.4990, val loss:0.4995\n",
      "Epoch [551/1000], train loss:0.4970, val loss:0.5055\n",
      "Epoch [552/1000], train loss:0.4939, val loss:0.4982\n",
      "Epoch [553/1000], train loss:0.4946, val loss:0.4989\n",
      "Epoch [554/1000], train loss:0.4957, val loss:0.5076\n",
      "Epoch [555/1000], train loss:0.4972, val loss:0.5022\n",
      "Epoch [556/1000], train loss:0.4954, val loss:0.4986\n",
      "Epoch [557/1000], train loss:0.4938, val loss:0.5063\n",
      "Epoch [558/1000], train loss:0.4949, val loss:0.4970\n",
      "Epoch [559/1000], train loss:0.4929, val loss:0.4945\n",
      "Epoch [560/1000], train loss:0.4927, val loss:0.4973\n",
      "Epoch [561/1000], train loss:0.4920, val loss:0.4977\n",
      "Epoch [562/1000], train loss:0.4934, val loss:0.4938\n",
      "Epoch [563/1000], train loss:0.4943, val loss:0.4985\n",
      "Epoch [564/1000], train loss:0.4922, val loss:0.4996\n",
      "Epoch [565/1000], train loss:0.4928, val loss:0.5004\n",
      "Epoch [566/1000], train loss:0.4908, val loss:0.4935\n",
      "Epoch [567/1000], train loss:0.4918, val loss:0.5054\n",
      "Epoch [568/1000], train loss:0.4904, val loss:0.4923\n",
      "Epoch [569/1000], train loss:0.4904, val loss:0.5033\n",
      "Epoch [570/1000], train loss:0.4919, val loss:0.5020\n",
      "Epoch [571/1000], train loss:0.4926, val loss:0.4972\n",
      "Epoch [572/1000], train loss:0.4920, val loss:0.4956\n",
      "Epoch [573/1000], train loss:0.4906, val loss:0.5045\n",
      "Epoch [574/1000], train loss:0.4938, val loss:0.5104\n",
      "Epoch [575/1000], train loss:0.4946, val loss:0.5050\n",
      "Epoch [576/1000], train loss:0.4933, val loss:0.5091\n",
      "Epoch [577/1000], train loss:0.4920, val loss:0.4986\n",
      "Epoch [578/1000], train loss:0.4913, val loss:0.4953\n",
      "Epoch [579/1000], train loss:0.4910, val loss:0.4959\n",
      "Epoch [580/1000], train loss:0.4929, val loss:0.4962\n",
      "Epoch [581/1000], train loss:0.4933, val loss:0.4986\n",
      "Epoch [582/1000], train loss:0.4923, val loss:0.5047\n",
      "Epoch [583/1000], train loss:0.4914, val loss:0.5002\n",
      "Epoch [584/1000], train loss:0.4926, val loss:0.4989\n",
      "Epoch [585/1000], train loss:0.4910, val loss:0.4928\n",
      "Epoch [586/1000], train loss:0.4907, val loss:0.4957\n",
      "Epoch [587/1000], train loss:0.4905, val loss:0.5006\n",
      "Epoch [588/1000], train loss:0.4906, val loss:0.4946\n",
      "Epoch [589/1000], train loss:0.4910, val loss:0.4977\n",
      "Epoch [590/1000], train loss:0.4918, val loss:0.4940\n",
      "Epoch [591/1000], train loss:0.4910, val loss:0.4961\n",
      "Epoch [592/1000], train loss:0.4902, val loss:0.4988\n",
      "Epoch [593/1000], train loss:0.4912, val loss:0.4965\n",
      "Epoch [594/1000], train loss:0.4918, val loss:0.4957\n",
      "Epoch [595/1000], train loss:0.4904, val loss:0.4953\n",
      "Epoch [596/1000], train loss:0.4924, val loss:0.4986\n",
      "Epoch [597/1000], train loss:0.4912, val loss:0.4996\n",
      "Epoch [598/1000], train loss:0.4932, val loss:0.4956\n",
      "Epoch [599/1000], train loss:0.4902, val loss:0.5056\n",
      "Epoch [600/1000], train loss:0.4904, val loss:0.5148\n",
      "Epoch [601/1000], train loss:0.4919, val loss:0.4943\n",
      "Epoch [602/1000], train loss:0.4932, val loss:0.4977\n",
      "Epoch [603/1000], train loss:0.4918, val loss:0.4987\n",
      "Epoch [604/1000], train loss:0.4914, val loss:0.4944\n",
      "Epoch [605/1000], train loss:0.4900, val loss:0.4936\n",
      "Epoch [606/1000], train loss:0.4932, val loss:0.4996\n",
      "Epoch [607/1000], train loss:0.4916, val loss:0.4976\n",
      "Epoch [608/1000], train loss:0.4917, val loss:0.5101\n",
      "Epoch [609/1000], train loss:0.4944, val loss:0.4936\n",
      "Epoch [610/1000], train loss:0.4920, val loss:0.4948\n",
      "Epoch [611/1000], train loss:0.4912, val loss:0.5017\n",
      "Epoch [612/1000], train loss:0.4928, val loss:0.4994\n",
      "Epoch [613/1000], train loss:0.4923, val loss:0.4969\n",
      "Epoch [614/1000], train loss:0.4937, val loss:0.4959\n",
      "Epoch [615/1000], train loss:0.4945, val loss:0.5000\n",
      "Epoch [616/1000], train loss:0.4936, val loss:0.5014\n",
      "Epoch [617/1000], train loss:0.4922, val loss:0.4970\n",
      "Epoch [618/1000], train loss:0.4914, val loss:0.5032\n",
      "Epoch [619/1000], train loss:0.4927, val loss:0.5010\n",
      "Epoch [620/1000], train loss:0.4923, val loss:0.4933\n",
      "Epoch [621/1000], train loss:0.4939, val loss:0.4961\n",
      "Epoch [622/1000], train loss:0.4928, val loss:0.5147\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [623/1000], train loss:0.4917, val loss:0.4923\n",
      "Epoch [624/1000], train loss:0.4927, val loss:0.5063\n",
      "Epoch [625/1000], train loss:0.4929, val loss:0.4954\n",
      "Epoch [626/1000], train loss:0.4913, val loss:0.4990\n",
      "Epoch [627/1000], train loss:0.4909, val loss:0.4955\n",
      "Epoch [628/1000], train loss:0.4912, val loss:0.5127\n",
      "Epoch [629/1000], train loss:0.4900, val loss:0.4930\n",
      "Epoch [630/1000], train loss:0.4914, val loss:0.4973\n",
      "Epoch [631/1000], train loss:0.4894, val loss:0.4967\n",
      "Epoch [632/1000], train loss:0.4895, val loss:0.5048\n",
      "Epoch [633/1000], train loss:0.4900, val loss:0.4929\n",
      "Epoch [634/1000], train loss:0.4891, val loss:0.4913\n",
      "Epoch [635/1000], train loss:0.4944, val loss:0.4960\n",
      "Epoch [636/1000], train loss:0.4907, val loss:0.5023\n",
      "Epoch [637/1000], train loss:0.4920, val loss:0.4978\n",
      "Epoch [638/1000], train loss:0.4924, val loss:0.4970\n",
      "Epoch [639/1000], train loss:0.4914, val loss:0.4973\n",
      "Epoch [640/1000], train loss:0.4908, val loss:0.4945\n",
      "Epoch [641/1000], train loss:0.4913, val loss:0.4963\n",
      "Epoch [642/1000], train loss:0.4914, val loss:0.5055\n",
      "Epoch [643/1000], train loss:0.4983, val loss:0.5130\n",
      "Epoch [644/1000], train loss:0.4926, val loss:0.4942\n",
      "Epoch [645/1000], train loss:0.4910, val loss:0.5055\n",
      "Epoch [646/1000], train loss:0.4938, val loss:0.5014\n",
      "Epoch [647/1000], train loss:0.4928, val loss:0.5065\n",
      "Epoch [648/1000], train loss:0.4918, val loss:0.5000\n",
      "Epoch [649/1000], train loss:0.4928, val loss:0.5084\n",
      "Epoch [650/1000], train loss:0.4939, val loss:0.5046\n",
      "Epoch [651/1000], train loss:0.4950, val loss:0.5020\n",
      "Epoch [652/1000], train loss:0.4937, val loss:0.4957\n",
      "Epoch [653/1000], train loss:0.4928, val loss:0.5028\n",
      "Epoch [654/1000], train loss:0.4951, val loss:0.5004\n",
      "Epoch [655/1000], train loss:0.4933, val loss:0.4980\n",
      "Epoch [656/1000], train loss:0.4916, val loss:0.4970\n",
      "Epoch [657/1000], train loss:0.4937, val loss:0.4965\n",
      "Epoch [658/1000], train loss:0.4916, val loss:0.4941\n",
      "Epoch [659/1000], train loss:0.4912, val loss:0.4984\n",
      "Epoch [660/1000], train loss:0.4918, val loss:0.4953\n",
      "Epoch [661/1000], train loss:0.4905, val loss:0.4942\n",
      "Epoch [662/1000], train loss:0.4932, val loss:0.4962\n",
      "Epoch [663/1000], train loss:0.4902, val loss:0.5029\n",
      "Epoch [664/1000], train loss:0.4932, val loss:0.5024\n",
      "Epoch [665/1000], train loss:0.4911, val loss:0.4986\n",
      "Epoch [666/1000], train loss:0.4918, val loss:0.4952\n",
      "Epoch [667/1000], train loss:0.4911, val loss:0.4989\n",
      "Epoch [668/1000], train loss:0.4904, val loss:0.4924\n",
      "Epoch [669/1000], train loss:0.4914, val loss:0.4971\n",
      "Epoch [670/1000], train loss:0.4915, val loss:0.4921\n",
      "Epoch [671/1000], train loss:0.4909, val loss:0.4952\n",
      "Epoch [672/1000], train loss:0.4956, val loss:0.5071\n",
      "Epoch [673/1000], train loss:0.4917, val loss:0.4946\n",
      "Epoch [674/1000], train loss:0.4916, val loss:0.4975\n",
      "Epoch [675/1000], train loss:0.4920, val loss:0.4994\n",
      "Epoch [676/1000], train loss:0.4903, val loss:0.4939\n",
      "Epoch [677/1000], train loss:0.4900, val loss:0.4917\n",
      "Epoch [678/1000], train loss:0.4898, val loss:0.4973\n",
      "Epoch [679/1000], train loss:0.4902, val loss:0.4943\n",
      "Epoch [680/1000], train loss:0.4914, val loss:0.4930\n",
      "Epoch [681/1000], train loss:0.4908, val loss:0.5040\n",
      "Epoch [682/1000], train loss:0.4904, val loss:0.4954\n",
      "Epoch [683/1000], train loss:0.4913, val loss:0.5047\n",
      "Epoch [684/1000], train loss:0.4906, val loss:0.4915\n",
      "Epoch [685/1000], train loss:0.4919, val loss:0.4961\n",
      "Epoch [686/1000], train loss:0.4919, val loss:0.5260\n",
      "Epoch [687/1000], train loss:0.4923, val loss:0.5024\n",
      "Epoch [688/1000], train loss:0.4920, val loss:0.4935\n",
      "Epoch [689/1000], train loss:0.4940, val loss:0.5027\n",
      "Epoch [690/1000], train loss:0.4923, val loss:0.5090\n",
      "Epoch [691/1000], train loss:0.4911, val loss:0.5073\n",
      "Epoch [692/1000], train loss:0.4926, val loss:0.4965\n",
      "Epoch [693/1000], train loss:0.4930, val loss:0.4949\n",
      "Epoch [694/1000], train loss:0.4912, val loss:0.5010\n",
      "Epoch [695/1000], train loss:0.4917, val loss:0.4935\n",
      "Epoch [696/1000], train loss:0.4908, val loss:0.4963\n",
      "Epoch [697/1000], train loss:0.4916, val loss:0.5009\n",
      "Epoch [698/1000], train loss:0.4922, val loss:0.4979\n",
      "Epoch [699/1000], train loss:0.4959, val loss:0.4961\n",
      "Epoch [700/1000], train loss:0.4919, val loss:0.5105\n",
      "Epoch [701/1000], train loss:0.4931, val loss:0.5027\n",
      "Epoch [702/1000], train loss:0.4933, val loss:0.5044\n",
      "Epoch [703/1000], train loss:0.4917, val loss:0.4926\n",
      "Epoch [704/1000], train loss:0.4942, val loss:0.5233\n",
      "Epoch [705/1000], train loss:0.4967, val loss:0.4965\n",
      "Epoch [706/1000], train loss:0.4930, val loss:0.5015\n",
      "Epoch [707/1000], train loss:0.4946, val loss:0.4958\n",
      "Epoch [708/1000], train loss:0.4939, val loss:0.4966\n",
      "Epoch [709/1000], train loss:0.4923, val loss:0.4994\n",
      "Epoch [710/1000], train loss:0.4917, val loss:0.4931\n",
      "Epoch [711/1000], train loss:0.4939, val loss:0.5008\n",
      "Epoch [712/1000], train loss:0.4926, val loss:0.4997\n",
      "Epoch [713/1000], train loss:0.4929, val loss:0.4974\n",
      "Epoch [714/1000], train loss:0.4899, val loss:0.5007\n",
      "Epoch [715/1000], train loss:0.4921, val loss:0.4979\n",
      "Epoch [716/1000], train loss:0.4899, val loss:0.4997\n",
      "Epoch [717/1000], train loss:0.4904, val loss:0.5052\n",
      "Epoch [718/1000], train loss:0.4941, val loss:0.4991\n",
      "Epoch [719/1000], train loss:0.4922, val loss:0.4958\n",
      "Epoch [720/1000], train loss:0.4929, val loss:0.5015\n",
      "Epoch [721/1000], train loss:0.5008, val loss:0.5147\n",
      "Epoch [722/1000], train loss:0.4954, val loss:0.4994\n",
      "Epoch [723/1000], train loss:0.4934, val loss:0.5322\n",
      "Epoch [724/1000], train loss:0.4927, val loss:0.5116\n",
      "Epoch [725/1000], train loss:0.4919, val loss:0.5040\n",
      "Epoch [726/1000], train loss:0.4922, val loss:0.4973\n",
      "Epoch [727/1000], train loss:0.4922, val loss:0.4945\n",
      "Epoch [728/1000], train loss:0.4911, val loss:0.4946\n",
      "Epoch [729/1000], train loss:0.4934, val loss:0.5102\n",
      "Epoch [730/1000], train loss:0.4929, val loss:0.5052\n",
      "Epoch [731/1000], train loss:0.4898, val loss:0.4930\n",
      "Epoch [732/1000], train loss:0.4891, val loss:0.4929\n",
      "Epoch [733/1000], train loss:0.4904, val loss:0.4934\n",
      "Epoch [734/1000], train loss:0.4902, val loss:0.4974\n",
      "Epoch [735/1000], train loss:0.4894, val loss:0.4942\n",
      "Epoch [736/1000], train loss:0.4914, val loss:0.4943\n",
      "Epoch [737/1000], train loss:0.4916, val loss:0.4945\n",
      "Epoch [738/1000], train loss:0.4901, val loss:0.4962\n",
      "Epoch [739/1000], train loss:0.4928, val loss:0.5066\n",
      "Epoch [740/1000], train loss:0.4899, val loss:0.4928\n",
      "Epoch [741/1000], train loss:0.4933, val loss:0.4954\n",
      "Epoch [742/1000], train loss:0.4927, val loss:0.4954\n",
      "Epoch [743/1000], train loss:0.4938, val loss:0.4931\n",
      "Epoch [744/1000], train loss:0.4936, val loss:0.5069\n",
      "Epoch [745/1000], train loss:0.5024, val loss:0.4946\n",
      "Epoch [746/1000], train loss:0.4913, val loss:0.4938\n",
      "Epoch [747/1000], train loss:0.4905, val loss:0.5048\n",
      "Epoch [748/1000], train loss:0.4910, val loss:0.4976\n",
      "Epoch [749/1000], train loss:0.4918, val loss:0.4956\n",
      "Epoch [750/1000], train loss:0.4937, val loss:0.4971\n",
      "Epoch [751/1000], train loss:0.4925, val loss:0.4986\n",
      "Epoch [752/1000], train loss:0.4944, val loss:0.5082\n",
      "Epoch [753/1000], train loss:0.4925, val loss:0.4966\n",
      "Epoch [754/1000], train loss:0.4944, val loss:0.5096\n",
      "Epoch [755/1000], train loss:0.4939, val loss:0.5013\n",
      "Epoch [756/1000], train loss:0.4930, val loss:0.4964\n",
      "Epoch [757/1000], train loss:0.4924, val loss:0.5032\n",
      "Epoch [758/1000], train loss:0.4925, val loss:0.5025\n",
      "Epoch [759/1000], train loss:0.4929, val loss:0.4961\n",
      "Epoch [760/1000], train loss:0.4923, val loss:0.5010\n",
      "Epoch [761/1000], train loss:0.4923, val loss:0.4955\n",
      "Epoch [762/1000], train loss:0.4925, val loss:0.5057\n",
      "Epoch [763/1000], train loss:0.4966, val loss:0.4992\n",
      "Epoch [764/1000], train loss:0.4936, val loss:0.4975\n",
      "Epoch [765/1000], train loss:0.4919, val loss:0.4988\n",
      "Epoch [766/1000], train loss:0.4954, val loss:0.5002\n",
      "Epoch [767/1000], train loss:0.4946, val loss:0.4999\n",
      "Epoch [768/1000], train loss:0.4940, val loss:0.4978\n",
      "Epoch [769/1000], train loss:0.4933, val loss:0.5010\n",
      "Epoch [770/1000], train loss:0.4924, val loss:0.4992\n",
      "Epoch [771/1000], train loss:0.4916, val loss:0.4916\n",
      "Epoch [772/1000], train loss:0.4918, val loss:0.5096\n",
      "Epoch [773/1000], train loss:0.4926, val loss:0.4974\n",
      "Epoch [774/1000], train loss:0.4902, val loss:0.4977\n",
      "Epoch [775/1000], train loss:0.4914, val loss:0.5015\n",
      "Epoch [776/1000], train loss:0.4903, val loss:0.4985\n",
      "Epoch [777/1000], train loss:0.4906, val loss:0.4967\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [778/1000], train loss:0.4901, val loss:0.5014\n",
      "Epoch [779/1000], train loss:0.4902, val loss:0.4997\n",
      "Epoch [780/1000], train loss:0.4907, val loss:0.4963\n",
      "Epoch [781/1000], train loss:0.4900, val loss:0.5001\n",
      "Epoch [782/1000], train loss:0.4910, val loss:0.5002\n",
      "Epoch [783/1000], train loss:0.4906, val loss:0.4980\n",
      "Epoch [784/1000], train loss:0.4894, val loss:0.4935\n",
      "Epoch [785/1000], train loss:0.4902, val loss:0.4992\n",
      "Epoch [786/1000], train loss:0.4909, val loss:0.4967\n",
      "Epoch [787/1000], train loss:0.4907, val loss:0.4998\n",
      "Epoch [788/1000], train loss:0.4927, val loss:0.4944\n",
      "Epoch [789/1000], train loss:0.4906, val loss:0.4985\n",
      "Epoch [790/1000], train loss:0.4894, val loss:0.4943\n",
      "Epoch [791/1000], train loss:0.4898, val loss:0.4949\n",
      "Epoch [792/1000], train loss:0.4894, val loss:0.4990\n",
      "Epoch [793/1000], train loss:0.4894, val loss:0.4957\n",
      "Epoch [794/1000], train loss:0.4922, val loss:0.5130\n",
      "Epoch [795/1000], train loss:0.4938, val loss:0.4950\n",
      "Epoch [796/1000], train loss:0.4900, val loss:0.4918\n",
      "Epoch [797/1000], train loss:0.4899, val loss:0.5032\n",
      "Epoch [798/1000], train loss:0.4923, val loss:0.5054\n",
      "Epoch [799/1000], train loss:0.4902, val loss:0.4917\n",
      "Epoch [800/1000], train loss:0.4894, val loss:0.4957\n",
      "Epoch [801/1000], train loss:0.4907, val loss:0.4990\n",
      "Epoch [802/1000], train loss:0.4917, val loss:0.4947\n",
      "Epoch [803/1000], train loss:0.4914, val loss:0.4958\n",
      "Epoch [804/1000], train loss:0.4916, val loss:0.4954\n",
      "Epoch [805/1000], train loss:0.4898, val loss:0.4920\n",
      "Epoch [806/1000], train loss:0.4893, val loss:0.4952\n",
      "Epoch [807/1000], train loss:0.4932, val loss:0.4971\n",
      "Epoch [808/1000], train loss:0.4911, val loss:0.5019\n",
      "Epoch [809/1000], train loss:0.4936, val loss:0.5007\n",
      "Epoch [810/1000], train loss:0.4909, val loss:0.4959\n",
      "Epoch [811/1000], train loss:0.4907, val loss:0.5024\n",
      "Epoch [812/1000], train loss:0.4912, val loss:0.4949\n",
      "Epoch [813/1000], train loss:0.4936, val loss:0.4963\n",
      "Epoch [814/1000], train loss:0.4920, val loss:0.5126\n",
      "Epoch [815/1000], train loss:0.4926, val loss:0.5006\n",
      "Epoch [816/1000], train loss:0.4918, val loss:0.5018\n",
      "Epoch [817/1000], train loss:0.4950, val loss:0.5004\n",
      "Epoch [818/1000], train loss:0.4963, val loss:0.5059\n",
      "Epoch [819/1000], train loss:0.4950, val loss:0.4964\n",
      "Epoch [820/1000], train loss:0.4939, val loss:0.5026\n",
      "Epoch [821/1000], train loss:0.4980, val loss:0.5040\n",
      "Epoch [822/1000], train loss:0.4975, val loss:0.5037\n",
      "Epoch [823/1000], train loss:0.4940, val loss:0.4997\n",
      "Epoch [824/1000], train loss:0.4921, val loss:0.4977\n",
      "Epoch [825/1000], train loss:0.4919, val loss:0.4968\n",
      "Epoch [826/1000], train loss:0.4919, val loss:0.4960\n",
      "Epoch [827/1000], train loss:0.4914, val loss:0.4999\n",
      "Epoch [828/1000], train loss:0.4936, val loss:0.4967\n",
      "Epoch [829/1000], train loss:0.4913, val loss:0.5020\n",
      "Epoch [830/1000], train loss:0.4909, val loss:0.5035\n",
      "Epoch [831/1000], train loss:0.4966, val loss:0.5099\n",
      "Epoch [832/1000], train loss:0.4942, val loss:0.5008\n",
      "Epoch [833/1000], train loss:0.4924, val loss:0.5116\n",
      "Epoch [834/1000], train loss:0.4951, val loss:0.5037\n",
      "Epoch [835/1000], train loss:0.4947, val loss:0.4986\n",
      "Epoch [836/1000], train loss:0.4945, val loss:0.4985\n",
      "Epoch [837/1000], train loss:0.4975, val loss:0.5041\n",
      "Epoch [838/1000], train loss:0.4965, val loss:0.5103\n",
      "Epoch [839/1000], train loss:0.4963, val loss:0.5036\n",
      "Epoch [840/1000], train loss:0.4925, val loss:0.4950\n",
      "Epoch [841/1000], train loss:0.4941, val loss:0.4967\n",
      "Epoch [842/1000], train loss:0.4938, val loss:0.4977\n",
      "Epoch [843/1000], train loss:0.4952, val loss:0.4996\n",
      "Epoch [844/1000], train loss:0.4956, val loss:0.5028\n",
      "Epoch [845/1000], train loss:0.4936, val loss:0.4941\n",
      "Epoch [846/1000], train loss:0.4920, val loss:0.4928\n",
      "Epoch [847/1000], train loss:0.4988, val loss:0.5008\n",
      "Epoch [848/1000], train loss:0.4959, val loss:0.5373\n",
      "Epoch [849/1000], train loss:0.4977, val loss:0.4987\n",
      "Epoch [850/1000], train loss:0.4943, val loss:0.4994\n",
      "Epoch [851/1000], train loss:0.4945, val loss:0.4945\n",
      "Epoch [852/1000], train loss:0.4933, val loss:0.5041\n",
      "Epoch [853/1000], train loss:0.4942, val loss:0.4981\n",
      "Epoch [854/1000], train loss:0.4980, val loss:0.5074\n",
      "Epoch [855/1000], train loss:0.4980, val loss:0.5064\n",
      "Epoch [856/1000], train loss:0.4965, val loss:0.4980\n",
      "Epoch [857/1000], train loss:0.4928, val loss:0.5025\n",
      "Epoch [858/1000], train loss:0.4932, val loss:0.5026\n",
      "Epoch [859/1000], train loss:0.4941, val loss:0.4984\n",
      "Epoch [860/1000], train loss:0.4927, val loss:0.4919\n",
      "Epoch [861/1000], train loss:0.4941, val loss:0.4934\n",
      "Epoch [862/1000], train loss:0.4909, val loss:0.4915\n",
      "Epoch [863/1000], train loss:0.4904, val loss:0.4977\n",
      "Epoch [864/1000], train loss:0.4900, val loss:0.4945\n",
      "Epoch [865/1000], train loss:0.4895, val loss:0.4968\n",
      "Epoch [866/1000], train loss:0.4941, val loss:0.4955\n",
      "Epoch [867/1000], train loss:0.4934, val loss:0.5081\n",
      "Epoch [868/1000], train loss:0.4933, val loss:0.5011\n",
      "Epoch [869/1000], train loss:0.4914, val loss:0.4953\n",
      "Epoch [870/1000], train loss:0.4910, val loss:0.4933\n",
      "Epoch [871/1000], train loss:0.4895, val loss:0.5014\n",
      "Epoch [872/1000], train loss:0.4919, val loss:0.5136\n",
      "Epoch [873/1000], train loss:0.4903, val loss:0.4970\n",
      "Epoch [874/1000], train loss:0.4913, val loss:0.4997\n",
      "Epoch [875/1000], train loss:0.4910, val loss:0.4946\n",
      "Epoch [876/1000], train loss:0.4967, val loss:0.5031\n",
      "Epoch [877/1000], train loss:0.4953, val loss:0.4971\n",
      "Epoch [878/1000], train loss:0.4925, val loss:0.4998\n",
      "Epoch [879/1000], train loss:0.4913, val loss:0.4991\n",
      "Epoch [880/1000], train loss:0.4950, val loss:0.4977\n",
      "Epoch [881/1000], train loss:0.4926, val loss:0.4951\n",
      "Epoch [882/1000], train loss:0.4915, val loss:0.4963\n",
      "Epoch [883/1000], train loss:0.4908, val loss:0.5186\n",
      "Epoch [884/1000], train loss:0.4905, val loss:0.4971\n",
      "Epoch [885/1000], train loss:0.4932, val loss:0.5154\n",
      "Epoch [886/1000], train loss:0.4920, val loss:0.5081\n",
      "Epoch [887/1000], train loss:0.4915, val loss:0.4909\n",
      "Epoch [888/1000], train loss:0.4918, val loss:0.5037\n",
      "Epoch [889/1000], train loss:0.5080, val loss:0.5005\n",
      "Epoch [890/1000], train loss:0.4952, val loss:0.4987\n",
      "Epoch [891/1000], train loss:0.4947, val loss:0.5006\n",
      "Epoch [892/1000], train loss:0.4942, val loss:0.5078\n",
      "Epoch [893/1000], train loss:0.4906, val loss:0.4964\n",
      "Epoch [894/1000], train loss:0.4910, val loss:0.5415\n",
      "Epoch [895/1000], train loss:0.4970, val loss:0.5115\n",
      "Epoch [896/1000], train loss:0.4941, val loss:0.4957\n",
      "Epoch [897/1000], train loss:0.4918, val loss:0.4968\n",
      "Epoch [898/1000], train loss:0.4949, val loss:0.5167\n",
      "Epoch [899/1000], train loss:0.4949, val loss:0.4966\n",
      "Epoch [900/1000], train loss:0.4953, val loss:0.5068\n",
      "Epoch [901/1000], train loss:0.4950, val loss:0.5005\n",
      "Epoch [902/1000], train loss:0.4969, val loss:0.4980\n",
      "Epoch [903/1000], train loss:0.5080, val loss:0.5196\n",
      "Epoch [904/1000], train loss:0.5089, val loss:0.5244\n",
      "Epoch [905/1000], train loss:0.5077, val loss:0.5078\n",
      "Epoch [906/1000], train loss:0.5039, val loss:0.5135\n",
      "Epoch [907/1000], train loss:0.5021, val loss:0.5030\n",
      "Epoch [908/1000], train loss:0.5006, val loss:0.5019\n",
      "Epoch [909/1000], train loss:0.4958, val loss:0.4979\n",
      "Epoch [910/1000], train loss:0.4949, val loss:0.5027\n",
      "Epoch [911/1000], train loss:0.4950, val loss:0.5020\n",
      "Epoch [912/1000], train loss:0.4947, val loss:0.5048\n",
      "Epoch [913/1000], train loss:0.4949, val loss:0.4987\n",
      "Epoch [914/1000], train loss:0.4963, val loss:0.5010\n",
      "Epoch [915/1000], train loss:0.4962, val loss:0.5036\n",
      "Epoch [916/1000], train loss:0.4949, val loss:0.4968\n",
      "Epoch [917/1000], train loss:0.4950, val loss:0.4992\n",
      "Epoch [918/1000], train loss:0.5012, val loss:0.5258\n",
      "Epoch [919/1000], train loss:0.4954, val loss:0.4983\n",
      "Epoch [920/1000], train loss:0.4949, val loss:0.4971\n",
      "Epoch [921/1000], train loss:0.4996, val loss:0.5390\n",
      "Epoch [922/1000], train loss:0.5014, val loss:0.5001\n",
      "Epoch [923/1000], train loss:0.4958, val loss:0.5022\n",
      "Epoch [924/1000], train loss:0.4942, val loss:0.5025\n",
      "Epoch [925/1000], train loss:0.4923, val loss:0.5122\n",
      "Epoch [926/1000], train loss:0.4936, val loss:0.5015\n",
      "Epoch [927/1000], train loss:0.4919, val loss:0.5112\n",
      "Epoch [928/1000], train loss:0.4909, val loss:0.5016\n",
      "Epoch [929/1000], train loss:0.4935, val loss:0.4959\n",
      "Epoch [930/1000], train loss:0.4939, val loss:0.5195\n",
      "Epoch [931/1000], train loss:0.4955, val loss:0.4983\n",
      "Epoch [932/1000], train loss:0.4924, val loss:0.5573\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [933/1000], train loss:0.4935, val loss:0.5364\n",
      "Epoch [934/1000], train loss:0.4930, val loss:0.4958\n",
      "Epoch [935/1000], train loss:0.4929, val loss:0.4990\n",
      "Epoch [936/1000], train loss:0.4941, val loss:0.5169\n",
      "Epoch [937/1000], train loss:0.4922, val loss:0.5216\n",
      "Epoch [938/1000], train loss:0.4930, val loss:0.4987\n",
      "Epoch [939/1000], train loss:0.4942, val loss:0.5065\n",
      "Epoch [940/1000], train loss:0.4915, val loss:0.4965\n",
      "Epoch [941/1000], train loss:0.4923, val loss:0.5035\n",
      "Epoch [942/1000], train loss:0.4932, val loss:0.4940\n",
      "Epoch [943/1000], train loss:0.4920, val loss:0.4987\n",
      "Epoch [944/1000], train loss:0.4927, val loss:0.5010\n",
      "Epoch [945/1000], train loss:0.4927, val loss:0.5008\n",
      "Epoch [946/1000], train loss:0.4918, val loss:0.5047\n",
      "Epoch [947/1000], train loss:0.4917, val loss:0.4962\n",
      "Epoch [948/1000], train loss:0.4915, val loss:0.5142\n",
      "Epoch [949/1000], train loss:0.4907, val loss:0.4947\n",
      "Epoch [950/1000], train loss:0.4933, val loss:0.5091\n",
      "Epoch [951/1000], train loss:0.4922, val loss:0.4971\n",
      "Epoch [952/1000], train loss:0.4908, val loss:0.4948\n",
      "Epoch [953/1000], train loss:0.4909, val loss:0.5027\n",
      "Epoch [954/1000], train loss:0.4933, val loss:0.4975\n",
      "Epoch [955/1000], train loss:0.4920, val loss:0.4970\n",
      "Epoch [956/1000], train loss:0.4926, val loss:0.5118\n",
      "Epoch [957/1000], train loss:0.4941, val loss:0.5022\n",
      "Epoch [958/1000], train loss:0.4998, val loss:0.5022\n",
      "Epoch [959/1000], train loss:0.4968, val loss:0.4972\n",
      "Epoch [960/1000], train loss:0.4918, val loss:0.4956\n",
      "Epoch [961/1000], train loss:0.4911, val loss:0.4961\n",
      "Epoch [962/1000], train loss:0.4917, val loss:0.5165\n",
      "Epoch [963/1000], train loss:0.4961, val loss:0.4966\n",
      "Epoch [964/1000], train loss:0.4938, val loss:0.4976\n",
      "Epoch [965/1000], train loss:0.4926, val loss:0.4967\n",
      "Epoch [966/1000], train loss:0.5043, val loss:0.5462\n",
      "Epoch [967/1000], train loss:0.5140, val loss:0.5193\n",
      "Epoch [968/1000], train loss:0.4997, val loss:0.4996\n",
      "Epoch [969/1000], train loss:0.4960, val loss:0.4965\n",
      "Epoch [970/1000], train loss:0.4934, val loss:0.4990\n",
      "Epoch [971/1000], train loss:0.4923, val loss:0.4944\n",
      "Epoch [972/1000], train loss:0.4934, val loss:0.5033\n",
      "Epoch [973/1000], train loss:0.4927, val loss:0.5048\n",
      "Epoch [974/1000], train loss:0.4940, val loss:0.4977\n",
      "Epoch [975/1000], train loss:0.4922, val loss:0.5025\n",
      "Epoch [976/1000], train loss:0.4909, val loss:0.5078\n",
      "Epoch [977/1000], train loss:0.4917, val loss:0.5022\n",
      "Epoch [978/1000], train loss:0.4925, val loss:0.4966\n",
      "Epoch [979/1000], train loss:0.4928, val loss:0.4943\n",
      "Epoch [980/1000], train loss:0.4937, val loss:0.4984\n",
      "Epoch [981/1000], train loss:0.4952, val loss:0.5016\n",
      "Epoch [982/1000], train loss:0.5010, val loss:0.4971\n",
      "Epoch [983/1000], train loss:0.4949, val loss:0.5025\n",
      "Epoch [984/1000], train loss:0.4948, val loss:0.5024\n",
      "Epoch [985/1000], train loss:0.4935, val loss:0.4953\n",
      "Epoch [986/1000], train loss:0.4936, val loss:0.5045\n",
      "Epoch [987/1000], train loss:0.4915, val loss:0.4957\n",
      "Epoch [988/1000], train loss:0.4911, val loss:0.5009\n",
      "Epoch [989/1000], train loss:0.4921, val loss:0.5125\n",
      "Epoch [990/1000], train loss:0.4913, val loss:0.4984\n",
      "Epoch [991/1000], train loss:0.4912, val loss:0.4942\n",
      "Epoch [992/1000], train loss:0.4903, val loss:0.4943\n",
      "Epoch [993/1000], train loss:0.4909, val loss:0.4963\n",
      "Epoch [994/1000], train loss:0.4927, val loss:0.4987\n",
      "Epoch [995/1000], train loss:0.4912, val loss:0.5184\n",
      "Epoch [996/1000], train loss:0.4897, val loss:0.4981\n",
      "Epoch [997/1000], train loss:0.4907, val loss:0.4949\n",
      "Epoch [998/1000], train loss:0.4916, val loss:0.5090\n",
      "Epoch [999/1000], train loss:0.4930, val loss:0.5062\n",
      "Epoch [1000/1000], train loss:0.4911, val loss:0.5022\n"
     ]
    }
   ],
   "source": [
    "train_loss  = []\n",
    "val_loss = []\n",
    "for epoch in range(epochs):\n",
    "    ae.train()\n",
    "    batch_loss  = []\n",
    "    for data in train_loader:\n",
    "        img, _ = data\n",
    "        img = img.view(img.size(0), -1).to(device)\n",
    "        output = ae(img)\n",
    "        \n",
    "        scheduler = StepLR(ae_optimizer, step_size=50, gamma=0.9)\n",
    "        loss = ae_criterion(output, img)\n",
    "        ae_optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        ae_optimizer.step()\n",
    "        # ===================log========================\n",
    "        batch_loss.append(loss.item())\n",
    "    train_loss.append(np.mean(batch_loss))\n",
    "    \n",
    "    \n",
    "    ae.eval()\n",
    "    batch_loss  = []\n",
    "    with torch.no_grad():\n",
    "        for img, y in val_loader:\n",
    "            img = img.view(img.size(0), -1)\n",
    "            img, y = img.to(device), y.to(device)\n",
    "            output = ae(img)\n",
    "            loss = ae_criterion(output, img)\n",
    "            batch_loss.append(loss.item())\n",
    "            val_loss.append(np.mean(batch_loss))\n",
    "        print(f'Epoch [{epoch + 1}/{epochs}], train loss:{train_loss[-1]:.4f}, val loss:{val_loss[-1]:.4f}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "985fa8ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss:0.5085482958016122\n"
     ]
    }
   ],
   "source": [
    "ae.eval()\n",
    "batch_loss = []\n",
    "with torch.no_grad():\n",
    "    for img, y in test_loader:\n",
    "        img = img.view(img.size(0), -1)\n",
    "        img, y = img.cuda(), y.cuda()\n",
    "        output = ae(img)\n",
    "        loss = ae_criterion(output, img)\n",
    "        batch_loss.append(loss.item())\n",
    "    print(f\"Test loss:{np.mean(batch_loss)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "b17733d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier = nn.Sequential(\n",
    "                nn.Linear(2, 64),\n",
    "                nn.ReLU(True),\n",
    "                nn.Linear(64, 128),\n",
    "                nn.ReLU(True),\n",
    "                nn.Linear(128, 10)).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "0c306f33",
   "metadata": {},
   "outputs": [],
   "source": [
    "classifier_criterion = nn.CrossEntropyLoss()\n",
    "classifier_optimizer = torch.optim.Adam(classifier.parameters(), lr=learning_rate)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "04584600",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Acc:77.206%\n",
      "Val accuracy:83.34%\n",
      "Train Acc:83.182%\n",
      "Val accuracy:81.6%\n",
      "Train Acc:81.982%\n",
      "Val accuracy:83.57%\n",
      "Train Acc:82.588%\n",
      "Val accuracy:82.95%\n",
      "Train Acc:82.994%\n",
      "Val accuracy:80.84%\n",
      "Train Acc:84.086%\n",
      "Val accuracy:82.8%\n",
      "Train Acc:83.51%\n",
      "Val accuracy:83.81%\n",
      "Train Acc:83.728%\n",
      "Val accuracy:81.37%\n",
      "Train Acc:83.798%\n",
      "Val accuracy:83.11%\n",
      "Train Acc:84.132%\n",
      "Val accuracy:83.48%\n",
      "Train Acc:83.448%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:84.444%\n",
      "Val accuracy:84.82%\n",
      "Train Acc:84.314%\n",
      "Val accuracy:82.7%\n",
      "Train Acc:84.43%\n",
      "Val accuracy:79.93%\n",
      "Train Acc:84.442%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:84.644%\n",
      "Val accuracy:81.72%\n",
      "Train Acc:84.188%\n",
      "Val accuracy:84.17%\n",
      "Train Acc:84.592%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:84.816%\n",
      "Val accuracy:83.78%\n",
      "Train Acc:84.65%\n",
      "Val accuracy:82.87%\n",
      "Train Acc:84.8%\n",
      "Val accuracy:84.03%\n",
      "Train Acc:84.876%\n",
      "Val accuracy:85.41%\n",
      "Train Acc:84.478%\n",
      "Val accuracy:81.91%\n",
      "Train Acc:84.732%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.78%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:84.856%\n",
      "Val accuracy:83.4%\n",
      "Train Acc:84.872%\n",
      "Val accuracy:85.08%\n",
      "Train Acc:84.886%\n",
      "Val accuracy:85.48%\n",
      "Train Acc:84.824%\n",
      "Val accuracy:84.35%\n",
      "Train Acc:84.74%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:83.978%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:84.646%\n",
      "Val accuracy:83.97%\n",
      "Train Acc:84.902%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:84.964%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:84.972%\n",
      "Val accuracy:81.34%\n",
      "Train Acc:83.646%\n",
      "Val accuracy:85.18%\n",
      "Train Acc:84.99%\n",
      "Val accuracy:84.67%\n",
      "Train Acc:84.97%\n",
      "Val accuracy:83.68%\n",
      "Train Acc:84.974%\n",
      "Val accuracy:85.08%\n",
      "Train Acc:84.954%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:84.544%\n",
      "Val accuracy:81.42%\n",
      "Train Acc:84.078%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:84.212%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:84.966%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:84.862%\n",
      "Val accuracy:84.57%\n",
      "Train Acc:84.828%\n",
      "Val accuracy:83.43%\n",
      "Train Acc:84.892%\n",
      "Val accuracy:84.19%\n",
      "Train Acc:83.882%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:84.862%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:84.854%\n",
      "Val accuracy:84.34%\n",
      "Train Acc:85.28%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.064%\n",
      "Val accuracy:82.82%\n",
      "Train Acc:85.14%\n",
      "Val accuracy:85.22%\n",
      "Train Acc:85.318%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:85.172%\n",
      "Val accuracy:83.58%\n",
      "Train Acc:85.222%\n",
      "Val accuracy:85.4%\n",
      "Train Acc:84.99%\n",
      "Val accuracy:84.45%\n",
      "Train Acc:85.324%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:85.366%\n",
      "Val accuracy:85.69%\n",
      "Train Acc:85.19%\n",
      "Val accuracy:84.13%\n",
      "Train Acc:85.042%\n",
      "Val accuracy:83.67%\n",
      "Train Acc:85.228%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.134%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.158%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.17%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.37%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.156%\n",
      "Val accuracy:82.15%\n",
      "Train Acc:85.284%\n",
      "Val accuracy:84.71%\n",
      "Train Acc:84.922%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:84.678%\n",
      "Val accuracy:80.55%\n",
      "Train Acc:85.156%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:85.01%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:85.072%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:85.236%\n",
      "Val accuracy:84.27%\n",
      "Train Acc:84.71%\n",
      "Val accuracy:85.41%\n",
      "Train Acc:85.19%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:84.562%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:85.35%\n",
      "Train Acc:84.952%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:85.046%\n",
      "Val accuracy:83.4%\n",
      "Train Acc:84.924%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:85.304%\n",
      "Val accuracy:83.52%\n",
      "Train Acc:84.72%\n",
      "Val accuracy:82.93%\n",
      "Train Acc:85.264%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.162%\n",
      "Val accuracy:82.9%\n",
      "Train Acc:84.984%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:85.246%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:85.28%\n",
      "Val accuracy:84.67%\n",
      "Train Acc:85.032%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:84.46%\n",
      "Val accuracy:85.33%\n",
      "Train Acc:84.654%\n",
      "Val accuracy:84.67%\n",
      "Train Acc:84.982%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:85.38%\n",
      "Val accuracy:84.9%\n",
      "Train Acc:85.132%\n",
      "Val accuracy:85.28%\n",
      "Train Acc:85.262%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:84.98%\n",
      "Val accuracy:81.1%\n",
      "Train Acc:85.066%\n",
      "Val accuracy:84.79%\n",
      "Train Acc:85.19%\n",
      "Val accuracy:84.85%\n",
      "Train Acc:84.796%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.246%\n",
      "Val accuracy:84.13%\n",
      "Train Acc:85.272%\n",
      "Val accuracy:82.09%\n",
      "Train Acc:85.13%\n",
      "Val accuracy:81.86%\n",
      "Train Acc:84.86%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:84.676%\n",
      "Val accuracy:84.27%\n",
      "Train Acc:85.016%\n",
      "Val accuracy:83.26%\n",
      "Train Acc:85.146%\n",
      "Val accuracy:85.64%\n",
      "Train Acc:84.764%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:85.226%\n",
      "Val accuracy:83.08%\n",
      "Train Acc:84.406%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.136%\n",
      "Val accuracy:84.34%\n",
      "Train Acc:85.326%\n",
      "Val accuracy:83.95%\n",
      "Train Acc:84.394%\n",
      "Val accuracy:80.51%\n",
      "Train Acc:84.888%\n",
      "Val accuracy:84.2%\n",
      "Train Acc:85.252%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:84.968%\n",
      "Val accuracy:85.46%\n",
      "Train Acc:85.244%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:85.3%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:82.63%\n",
      "Train Acc:85.222%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:85.294%\n",
      "Val accuracy:84.63%\n",
      "Train Acc:85.088%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.234%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:85.406%\n",
      "Val accuracy:83.9%\n",
      "Train Acc:85.222%\n",
      "Val accuracy:84.57%\n",
      "Train Acc:85.62%\n",
      "Val accuracy:85.46%\n",
      "Train Acc:84.778%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.492%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.394%\n",
      "Val accuracy:84.15%\n",
      "Train Acc:85.504%\n",
      "Val accuracy:84.1%\n",
      "Train Acc:85.232%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:85.294%\n",
      "Val accuracy:85.66%\n",
      "Train Acc:84.828%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.144%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:84.852%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:84.948%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:84.622%\n",
      "Val accuracy:85.82%\n",
      "Train Acc:85.45%\n",
      "Val accuracy:84.54%\n",
      "Train Acc:85.208%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.572%\n",
      "Val accuracy:85.77%\n",
      "Train Acc:85.49%\n",
      "Val accuracy:85.33%\n",
      "Train Acc:85.252%\n",
      "Val accuracy:85.44%\n",
      "Train Acc:84.5%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.174%\n",
      "Val accuracy:84.61%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:85.76%\n",
      "Train Acc:85.49%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.168%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:85.48%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:83.878%\n",
      "Val accuracy:83.52%\n",
      "Train Acc:84.288%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:84.558%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:84.77%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.166%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.246%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.566%\n",
      "Val accuracy:84.61%\n",
      "Train Acc:85.248%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.36%\n",
      "Val accuracy:85.69%\n",
      "Train Acc:85.39%\n",
      "Val accuracy:85.66%\n",
      "Train Acc:85.582%\n",
      "Val accuracy:84.49%\n",
      "Train Acc:85.382%\n",
      "Val accuracy:83.1%\n",
      "Train Acc:85.47%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.46%\n",
      "Val accuracy:83.92%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.592%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.354%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.436%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:85.588%\n",
      "Val accuracy:85.08%\n",
      "Train Acc:85.322%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:85.502%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.196%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.098%\n",
      "Val accuracy:85.4%\n",
      "Train Acc:85.652%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:85.378%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.452%\n",
      "Val accuracy:85.78%\n",
      "Train Acc:85.292%\n",
      "Val accuracy:83.27%\n",
      "Train Acc:85.37%\n",
      "Val accuracy:85.57%\n",
      "Train Acc:85.39%\n",
      "Val accuracy:84.84%\n",
      "Train Acc:85.294%\n",
      "Val accuracy:82.14%\n",
      "Train Acc:84.908%\n",
      "Val accuracy:84.86%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:83.63%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:85.536%\n",
      "Val accuracy:84.01%\n",
      "Train Acc:85.29%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.102%\n",
      "Val accuracy:85.13%\n",
      "Train Acc:85.444%\n",
      "Val accuracy:84.18%\n",
      "Train Acc:85.356%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.492%\n",
      "Val accuracy:85.54%\n",
      "Train Acc:85.482%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.392%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.202%\n",
      "Val accuracy:81.75%\n",
      "Train Acc:85.316%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.678%\n",
      "Val accuracy:84.25%\n",
      "Train Acc:85.004%\n",
      "Val accuracy:84.9%\n",
      "Train Acc:85.674%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.378%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.19%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.27%\n",
      "Val accuracy:85.06%\n",
      "Train Acc:85.53%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.518%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:85.25%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:85.454%\n",
      "Val accuracy:85.11%\n",
      "Train Acc:85.466%\n",
      "Val accuracy:83.64%\n",
      "Train Acc:85.528%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:84.86%\n",
      "Train Acc:85.194%\n",
      "Val accuracy:84.35%\n",
      "Train Acc:85.278%\n",
      "Val accuracy:83.77%\n",
      "Train Acc:85.272%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.452%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.248%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:85.424%\n",
      "Val accuracy:84.56%\n",
      "Train Acc:85.336%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.456%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:85.414%\n",
      "Val accuracy:85.64%\n",
      "Train Acc:85.288%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:85.516%\n",
      "Val accuracy:85.87%\n",
      "Train Acc:84.98%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Val accuracy:84.79%\n",
      "Train Acc:85.504%\n",
      "Val accuracy:85.59%\n",
      "Train Acc:85.402%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:85.462%\n",
      "Val accuracy:85.3%\n",
      "Train Acc:85.302%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:85.48%\n",
      "Val accuracy:84.66%\n",
      "Train Acc:85.218%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:85.354%\n",
      "Val accuracy:83.66%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:83.89%\n",
      "Train Acc:85.24%\n",
      "Val accuracy:84.27%\n",
      "Train Acc:85.102%\n",
      "Val accuracy:85.48%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.712%\n",
      "Val accuracy:85.13%\n",
      "Train Acc:85.078%\n",
      "Val accuracy:85.64%\n",
      "Train Acc:85.41%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:84.864%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.166%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.316%\n",
      "Val accuracy:85.35%\n",
      "Train Acc:85.466%\n",
      "Val accuracy:85.59%\n",
      "Train Acc:85.54%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:82.91%\n",
      "Train Acc:85.296%\n",
      "Val accuracy:84.52%\n",
      "Train Acc:85.44%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:85.326%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.452%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.404%\n",
      "Val accuracy:84.62%\n",
      "Train Acc:85.648%\n",
      "Val accuracy:83.99%\n",
      "Train Acc:85.014%\n",
      "Val accuracy:85.62%\n",
      "Train Acc:85.34%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:82.56%\n",
      "Train Acc:85.266%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.43%\n",
      "Val accuracy:85.08%\n",
      "Train Acc:85.208%\n",
      "Val accuracy:83.48%\n",
      "Train Acc:85.416%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.512%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:85.544%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:84.81%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:83.0%\n",
      "Train Acc:85.284%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.528%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:85.138%\n",
      "Val accuracy:85.18%\n",
      "Train Acc:85.172%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.126%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:85.352%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.39%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:85.268%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.098%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:85.308%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.566%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:85.17%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.646%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.078%\n",
      "Val accuracy:84.64%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:85.3%\n",
      "Train Acc:85.52%\n",
      "Val accuracy:85.18%\n",
      "Train Acc:85.156%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:85.63%\n",
      "Val accuracy:85.5%\n",
      "Train Acc:85.486%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:85.214%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.46%\n",
      "Val accuracy:83.62%\n",
      "Train Acc:84.98%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:85.358%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:85.498%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.21%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:85.322%\n",
      "Val accuracy:84.77%\n",
      "Train Acc:85.254%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.462%\n",
      "Val accuracy:85.68%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:85.608%\n",
      "Val accuracy:84.84%\n",
      "Train Acc:85.532%\n",
      "Val accuracy:83.11%\n",
      "Train Acc:85.314%\n",
      "Val accuracy:85.81%\n",
      "Train Acc:85.416%\n",
      "Val accuracy:85.4%\n",
      "Train Acc:85.474%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:85.13%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:85.334%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:85.656%\n",
      "Val accuracy:85.48%\n",
      "Train Acc:85.488%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:85.168%\n",
      "Val accuracy:83.11%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:85.22%\n",
      "Train Acc:85.318%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.422%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.562%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.3%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.25%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.488%\n",
      "Val accuracy:85.38%\n",
      "Train Acc:85.386%\n",
      "Val accuracy:85.28%\n",
      "Train Acc:85.46%\n",
      "Val accuracy:83.99%\n",
      "Train Acc:85.426%\n",
      "Val accuracy:80.93%\n",
      "Train Acc:85.024%\n",
      "Val accuracy:85.89%\n",
      "Train Acc:85.438%\n",
      "Val accuracy:84.9%\n",
      "Train Acc:85.524%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.494%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.416%\n",
      "Val accuracy:82.7%\n",
      "Train Acc:85.302%\n",
      "Val accuracy:85.66%\n",
      "Train Acc:85.536%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.702%\n",
      "Val accuracy:85.46%\n",
      "Train Acc:85.524%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:85.55%\n",
      "Val accuracy:84.67%\n",
      "Train Acc:85.078%\n",
      "Val accuracy:84.23%\n",
      "Train Acc:85.392%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:85.526%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:85.492%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.174%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:85.336%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:84.84%\n",
      "Val accuracy:84.77%\n",
      "Train Acc:85.174%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:85.232%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.602%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.396%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.11%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.342%\n",
      "Val accuracy:84.74%\n",
      "Train Acc:85.422%\n",
      "Val accuracy:85.62%\n",
      "Train Acc:85.496%\n",
      "Val accuracy:83.87%\n",
      "Train Acc:85.292%\n",
      "Val accuracy:83.8%\n",
      "Train Acc:85.52%\n",
      "Val accuracy:85.32%\n",
      "Train Acc:85.412%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:84.896%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.54%\n",
      "Val accuracy:84.23%\n",
      "Train Acc:85.47%\n",
      "Val accuracy:85.54%\n",
      "Train Acc:85.216%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.576%\n",
      "Val accuracy:85.49%\n",
      "Train Acc:85.282%\n",
      "Val accuracy:85.64%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.6%\n",
      "Val accuracy:83.05%\n",
      "Train Acc:85.144%\n",
      "Val accuracy:83.8%\n",
      "Train Acc:85.366%\n",
      "Val accuracy:84.25%\n",
      "Train Acc:85.346%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.122%\n",
      "Val accuracy:85.5%\n",
      "Train Acc:85.572%\n",
      "Val accuracy:85.79%\n",
      "Train Acc:85.564%\n",
      "Val accuracy:85.5%\n",
      "Train Acc:85.618%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:85.318%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:85.472%\n",
      "Val accuracy:85.38%\n",
      "Train Acc:85.382%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:84.916%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.608%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:85.456%\n",
      "Val accuracy:84.55%\n",
      "Train Acc:85.556%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:85.344%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:85.358%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:85.598%\n",
      "Val accuracy:85.22%\n",
      "Train Acc:85.21%\n",
      "Val accuracy:85.49%\n",
      "Train Acc:85.314%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.512%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:85.464%\n",
      "Val accuracy:85.59%\n",
      "Train Acc:85.592%\n",
      "Val accuracy:84.18%\n",
      "Train Acc:85.396%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:85.344%\n",
      "Val accuracy:85.32%\n",
      "Train Acc:85.406%\n",
      "Val accuracy:84.86%\n",
      "Train Acc:84.922%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.358%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.384%\n",
      "Val accuracy:84.23%\n",
      "Train Acc:85.422%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.254%\n",
      "Val accuracy:84.52%\n",
      "Train Acc:85.226%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:85.202%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:85.438%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.426%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:84.34%\n",
      "Train Acc:85.292%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:85.544%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.456%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:85.082%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.138%\n",
      "Val accuracy:84.71%\n",
      "Train Acc:85.456%\n",
      "Val accuracy:85.35%\n",
      "Train Acc:85.608%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.142%\n",
      "Val accuracy:80.48%\n",
      "Train Acc:84.824%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:85.546%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:85.108%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.478%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.412%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.162%\n",
      "Val accuracy:85.32%\n",
      "Train Acc:85.658%\n",
      "Val accuracy:85.02%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:81.87%\n",
      "Train Acc:85.294%\n",
      "Val accuracy:82.84%\n",
      "Train Acc:85.386%\n",
      "Val accuracy:84.84%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:84.61%\n",
      "Train Acc:85.352%\n",
      "Val accuracy:85.7%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.414%\n",
      "Val accuracy:82.25%\n",
      "Train Acc:85.198%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:85.116%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.478%\n",
      "Val accuracy:84.25%\n",
      "Train Acc:85.604%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:85.628%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:84.998%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:85.49%\n",
      "Val accuracy:85.54%\n",
      "Train Acc:85.244%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.676%\n",
      "Val accuracy:85.49%\n",
      "Train Acc:85.102%\n",
      "Val accuracy:83.5%\n",
      "Train Acc:85.332%\n",
      "Val accuracy:85.71%\n",
      "Train Acc:85.564%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:85.126%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.524%\n",
      "Val accuracy:83.33%\n",
      "Train Acc:85.416%\n",
      "Val accuracy:85.56%\n",
      "Train Acc:85.344%\n",
      "Val accuracy:85.38%\n",
      "Train Acc:85.328%\n",
      "Val accuracy:82.41%\n",
      "Train Acc:85.232%\n",
      "Val accuracy:83.47%\n",
      "Train Acc:85.354%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:85.376%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.174%\n",
      "Val accuracy:85.48%\n",
      "Train Acc:84.874%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:85.486%\n",
      "Val accuracy:84.16%\n",
      "Train Acc:85.302%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.574%\n",
      "Val accuracy:84.3%\n",
      "Train Acc:85.424%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.206%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.244%\n",
      "Val accuracy:84.96%\n",
      "Train Acc:85.52%\n",
      "Val accuracy:83.08%\n",
      "Train Acc:85.576%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.462%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Val accuracy:85.3%\n",
      "Train Acc:85.306%\n",
      "Val accuracy:84.13%\n",
      "Train Acc:85.118%\n",
      "Val accuracy:85.59%\n",
      "Train Acc:85.434%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.492%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.342%\n",
      "Val accuracy:85.11%\n",
      "Train Acc:85.298%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:85.304%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:85.454%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:85.628%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.024%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:84.82%\n",
      "Train Acc:85.034%\n",
      "Val accuracy:84.37%\n",
      "Train Acc:85.67%\n",
      "Val accuracy:85.44%\n",
      "Train Acc:84.898%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.506%\n",
      "Val accuracy:85.69%\n",
      "Train Acc:85.254%\n",
      "Val accuracy:85.73%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:85.61%\n",
      "Train Acc:85.548%\n",
      "Val accuracy:84.73%\n",
      "Train Acc:85.47%\n",
      "Val accuracy:85.56%\n",
      "Train Acc:85.144%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:85.342%\n",
      "Val accuracy:84.84%\n",
      "Train Acc:85.394%\n",
      "Val accuracy:85.61%\n",
      "Train Acc:85.602%\n",
      "Val accuracy:85.44%\n",
      "Train Acc:84.832%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:85.0%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:84.47%\n",
      "Train Acc:85.528%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:85.538%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:85.32%\n",
      "Val accuracy:85.67%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.576%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.028%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.638%\n",
      "Val accuracy:85.06%\n",
      "Train Acc:85.21%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.206%\n",
      "Val accuracy:83.61%\n",
      "Train Acc:85.594%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.438%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.308%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:82.9%\n",
      "Train Acc:85.256%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.056%\n",
      "Val accuracy:85.33%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:85.348%\n",
      "Val accuracy:81.99%\n",
      "Train Acc:85.336%\n",
      "Val accuracy:85.62%\n",
      "Train Acc:85.428%\n",
      "Val accuracy:84.43%\n",
      "Train Acc:85.404%\n",
      "Val accuracy:83.95%\n",
      "Train Acc:85.392%\n",
      "Val accuracy:84.55%\n",
      "Train Acc:85.192%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.496%\n",
      "Val accuracy:80.78%\n",
      "Train Acc:85.58%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:84.29%\n",
      "Train Acc:85.514%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.402%\n",
      "Val accuracy:84.21%\n",
      "Train Acc:85.07%\n",
      "Val accuracy:85.11%\n",
      "Train Acc:85.206%\n",
      "Val accuracy:83.59%\n",
      "Train Acc:85.37%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:85.436%\n",
      "Val accuracy:84.82%\n",
      "Train Acc:85.168%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:85.12%\n",
      "Val accuracy:84.77%\n",
      "Train Acc:85.184%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:85.474%\n",
      "Val accuracy:82.88%\n",
      "Train Acc:85.528%\n",
      "Val accuracy:84.13%\n",
      "Train Acc:85.278%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.27%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:85.428%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:85.662%\n",
      "Val accuracy:83.45%\n",
      "Train Acc:85.45%\n",
      "Val accuracy:82.5%\n",
      "Train Acc:85.446%\n",
      "Val accuracy:83.99%\n",
      "Train Acc:85.34%\n",
      "Val accuracy:85.51%\n",
      "Train Acc:85.296%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.186%\n",
      "Val accuracy:84.68%\n",
      "Train Acc:85.34%\n",
      "Val accuracy:80.03%\n",
      "Train Acc:85.294%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.434%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:85.596%\n",
      "Val accuracy:84.14%\n",
      "Train Acc:85.458%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:85.328%\n",
      "Val accuracy:83.17%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:85.57%\n",
      "Train Acc:85.35%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.128%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.276%\n",
      "Val accuracy:84.86%\n",
      "Train Acc:85.622%\n",
      "Val accuracy:83.5%\n",
      "Train Acc:85.066%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:85.582%\n",
      "Val accuracy:83.22%\n",
      "Train Acc:85.456%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:84.67%\n",
      "Train Acc:85.542%\n",
      "Val accuracy:83.72%\n",
      "Train Acc:85.072%\n",
      "Val accuracy:85.32%\n",
      "Train Acc:85.496%\n",
      "Val accuracy:85.62%\n",
      "Train Acc:85.572%\n",
      "Val accuracy:82.97%\n",
      "Train Acc:85.106%\n",
      "Val accuracy:84.61%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.546%\n",
      "Val accuracy:84.29%\n",
      "Train Acc:85.012%\n",
      "Val accuracy:84.08%\n",
      "Train Acc:85.324%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.452%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:84.716%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.264%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:85.04%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:84.86%\n",
      "Train Acc:85.15%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:85.238%\n",
      "Val accuracy:85.58%\n",
      "Train Acc:85.332%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.442%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:85.488%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.226%\n",
      "Val accuracy:83.91%\n",
      "Train Acc:85.382%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:85.426%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.262%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:85.498%\n",
      "Val accuracy:83.54%\n",
      "Train Acc:85.16%\n",
      "Val accuracy:83.18%\n",
      "Train Acc:85.198%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:84.996%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:84.986%\n",
      "Val accuracy:81.04%\n",
      "Train Acc:85.486%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:85.382%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.466%\n",
      "Val accuracy:83.59%\n",
      "Train Acc:84.996%\n",
      "Val accuracy:85.5%\n",
      "Train Acc:85.4%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.168%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.066%\n",
      "Val accuracy:85.02%\n",
      "Train Acc:85.246%\n",
      "Val accuracy:85.71%\n",
      "Train Acc:85.068%\n",
      "Val accuracy:84.15%\n",
      "Train Acc:85.216%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.354%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.424%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.23%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:85.478%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.506%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:84.608%\n",
      "Val accuracy:82.91%\n",
      "Train Acc:85.324%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:85.374%\n",
      "Val accuracy:84.79%\n",
      "Train Acc:85.36%\n",
      "Val accuracy:84.0%\n",
      "Train Acc:85.426%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:85.356%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:84.912%\n",
      "Val accuracy:82.66%\n",
      "Train Acc:85.26%\n",
      "Val accuracy:83.8%\n",
      "Train Acc:85.128%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:85.316%\n",
      "Val accuracy:84.56%\n",
      "Train Acc:85.124%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.348%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.416%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.16%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:84.33%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:85.102%\n",
      "Val accuracy:81.83%\n",
      "Train Acc:84.982%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.064%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.324%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:85.312%\n",
      "Val accuracy:84.61%\n",
      "Train Acc:84.94%\n",
      "Val accuracy:82.42%\n",
      "Train Acc:85.146%\n",
      "Val accuracy:84.08%\n",
      "Train Acc:84.892%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:83.98%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.028%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:85.208%\n",
      "Val accuracy:84.44%\n",
      "Train Acc:85.202%\n",
      "Val accuracy:83.4%\n",
      "Train Acc:85.274%\n",
      "Val accuracy:85.7%\n",
      "Train Acc:85.274%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.504%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:85.646%\n",
      "Val accuracy:83.77%\n",
      "Train Acc:84.532%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.252%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.278%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:85.322%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.882%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:85.306%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:85.29%\n",
      "Val accuracy:84.95%\n",
      "Train Acc:85.222%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:85.182%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.322%\n",
      "Val accuracy:85.61%\n",
      "Train Acc:85.066%\n",
      "Val accuracy:82.79%\n",
      "Train Acc:84.782%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:85.43%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:85.35%\n",
      "Val accuracy:85.44%\n",
      "Train Acc:85.054%\n",
      "Val accuracy:83.22%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:84.48%\n",
      "Train Acc:85.042%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:85.272%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:84.994%\n",
      "Val accuracy:85.35%\n",
      "Train Acc:85.464%\n",
      "Val accuracy:85.13%\n",
      "Train Acc:84.488%\n",
      "Val accuracy:83.3%\n",
      "Train Acc:84.936%\n",
      "Val accuracy:84.39%\n",
      "Train Acc:84.98%\n",
      "Val accuracy:84.64%\n",
      "Train Acc:85.162%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:85.2%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:85.152%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.472%\n",
      "Val accuracy:85.72%\n",
      "Train Acc:85.12%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:85.23%\n",
      "Val accuracy:83.35%\n",
      "Train Acc:85.474%\n",
      "Val accuracy:83.89%\n",
      "Train Acc:85.132%\n",
      "Val accuracy:84.94%\n",
      "Train Acc:85.338%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:85.21%\n",
      "Val accuracy:83.22%\n",
      "Train Acc:85.484%\n",
      "Val accuracy:83.13%\n",
      "Train Acc:84.986%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:85.324%\n",
      "Val accuracy:85.78%\n",
      "Train Acc:85.058%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:85.132%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:84.084%\n",
      "Val accuracy:84.96%\n",
      "Train Acc:85.166%\n",
      "Val accuracy:83.16%\n",
      "Train Acc:85.552%\n",
      "Val accuracy:85.28%\n",
      "Train Acc:85.174%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:84.91%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:84.97%\n",
      "Val accuracy:84.29%\n",
      "Train Acc:85.376%\n",
      "Val accuracy:84.66%\n",
      "Train Acc:85.344%\n",
      "Val accuracy:85.37%\n",
      "Train Acc:85.528%\n",
      "Val accuracy:84.45%\n",
      "Train Acc:85.244%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:85.108%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:85.534%\n",
      "Val accuracy:84.79%\n",
      "Train Acc:84.876%\n",
      "Val accuracy:85.22%\n",
      "Train Acc:85.052%\n",
      "Val accuracy:84.93%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Acc:85.552%\n",
      "Val accuracy:85.5%\n",
      "Train Acc:85.468%\n",
      "Val accuracy:82.77%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:85.33%\n",
      "Train Acc:85.23%\n",
      "Val accuracy:84.29%\n",
      "Train Acc:85.362%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:85.358%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:85.19%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:85.256%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:84.86%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:85.224%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.3%\n",
      "Val accuracy:83.54%\n",
      "Train Acc:85.21%\n",
      "Val accuracy:82.02%\n",
      "Train Acc:84.902%\n",
      "Val accuracy:85.11%\n",
      "Train Acc:85.41%\n",
      "Val accuracy:83.73%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:85.388%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:85.39%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:85.524%\n",
      "Val accuracy:84.54%\n",
      "Train Acc:85.364%\n",
      "Val accuracy:84.74%\n",
      "Train Acc:85.562%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:85.052%\n",
      "Val accuracy:80.58%\n",
      "Train Acc:85.224%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:85.316%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.074%\n",
      "Val accuracy:85.3%\n",
      "Train Acc:85.122%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.304%\n",
      "Val accuracy:82.89%\n",
      "Train Acc:85.194%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:85.306%\n",
      "Val accuracy:84.48%\n",
      "Train Acc:85.24%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:85.408%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:84.324%\n",
      "Val accuracy:85.53%\n",
      "Train Acc:85.462%\n",
      "Val accuracy:83.55%\n",
      "Train Acc:85.336%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.406%\n",
      "Val accuracy:85.26%\n",
      "Train Acc:85.15%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:85.516%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:84.76%\n",
      "Val accuracy:83.99%\n",
      "Train Acc:84.878%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:85.304%\n",
      "Val accuracy:84.7%\n",
      "Train Acc:85.396%\n",
      "Val accuracy:83.18%\n",
      "Train Acc:85.02%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.27%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:84.986%\n",
      "Val accuracy:85.41%\n",
      "Train Acc:84.726%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:85.194%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:85.698%\n",
      "Val accuracy:85.57%\n",
      "Train Acc:85.444%\n",
      "Val accuracy:85.54%\n",
      "Train Acc:84.912%\n",
      "Val accuracy:85.13%\n",
      "Train Acc:85.334%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:84.932%\n",
      "Val accuracy:85.55%\n",
      "Train Acc:85.474%\n",
      "Val accuracy:84.46%\n",
      "Train Acc:85.598%\n",
      "Val accuracy:84.18%\n",
      "Train Acc:84.89%\n",
      "Val accuracy:85.4%\n",
      "Train Acc:85.366%\n",
      "Val accuracy:85.34%\n",
      "Train Acc:85.308%\n",
      "Val accuracy:85.09%\n",
      "Train Acc:85.376%\n",
      "Val accuracy:85.12%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:83.23%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:84.79%\n",
      "Train Acc:85.282%\n",
      "Val accuracy:85.14%\n",
      "Train Acc:85.584%\n",
      "Val accuracy:85.89%\n",
      "Train Acc:85.448%\n",
      "Val accuracy:84.96%\n",
      "Train Acc:85.432%\n",
      "Val accuracy:85.46%\n",
      "Train Acc:85.428%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:84.912%\n",
      "Val accuracy:85.52%\n",
      "Train Acc:84.734%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.298%\n",
      "Val accuracy:84.81%\n",
      "Train Acc:85.322%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:85.302%\n",
      "Val accuracy:82.45%\n",
      "Train Acc:84.956%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:85.482%\n",
      "Val accuracy:84.38%\n",
      "Train Acc:85.41%\n",
      "Val accuracy:85.3%\n",
      "Train Acc:85.238%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:85.154%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.48%\n",
      "Val accuracy:85.57%\n",
      "Train Acc:85.376%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:85.234%\n",
      "Val accuracy:84.39%\n",
      "Train Acc:85.24%\n",
      "Val accuracy:85.67%\n",
      "Train Acc:85.236%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:85.268%\n",
      "Val accuracy:84.55%\n",
      "Train Acc:85.144%\n",
      "Val accuracy:85.35%\n",
      "Train Acc:85.096%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:85.404%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:85.062%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:85.5%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:85.25%\n",
      "Val accuracy:85.6%\n",
      "Train Acc:85.26%\n",
      "Val accuracy:85.71%\n",
      "Train Acc:85.312%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:85.256%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:85.398%\n",
      "Val accuracy:84.48%\n",
      "Train Acc:84.86%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:85.446%\n",
      "Val accuracy:85.36%\n",
      "Train Acc:85.452%\n",
      "Val accuracy:83.6%\n",
      "Train Acc:85.178%\n",
      "Val accuracy:85.22%\n",
      "Train Acc:85.422%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:83.51%\n",
      "Train Acc:84.978%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:85.286%\n",
      "Val accuracy:83.7%\n",
      "Train Acc:85.312%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:85.362%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:85.434%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:85.326%\n",
      "Val accuracy:85.06%\n",
      "Train Acc:85.368%\n",
      "Val accuracy:80.83%\n",
      "Train Acc:84.658%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:85.064%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:85.418%\n",
      "Val accuracy:85.31%\n",
      "Train Acc:85.446%\n",
      "Val accuracy:85.2%\n",
      "Train Acc:85.486%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:84.964%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:83.712%\n",
      "Val accuracy:82.25%\n",
      "Train Acc:84.088%\n",
      "Val accuracy:84.8%\n",
      "Train Acc:84.612%\n",
      "Val accuracy:85.08%\n",
      "Train Acc:84.484%\n",
      "Val accuracy:83.52%\n",
      "Train Acc:84.622%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:84.692%\n",
      "Val accuracy:84.54%\n",
      "Train Acc:84.378%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:84.54%\n",
      "Val accuracy:84.63%\n",
      "Train Acc:84.648%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:84.838%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:84.72%\n",
      "Val accuracy:84.33%\n",
      "Train Acc:84.664%\n",
      "Val accuracy:83.26%\n",
      "Train Acc:84.686%\n",
      "Val accuracy:85.18%\n",
      "Train Acc:84.766%\n",
      "Val accuracy:84.54%\n",
      "Train Acc:84.502%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:84.69%\n",
      "Val accuracy:84.83%\n",
      "Train Acc:84.788%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.906%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:84.61%\n",
      "Val accuracy:83.89%\n",
      "Train Acc:84.68%\n",
      "Val accuracy:82.95%\n",
      "Train Acc:84.534%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:84.722%\n",
      "Val accuracy:78.42%\n",
      "Train Acc:83.076%\n",
      "Val accuracy:84.15%\n",
      "Train Acc:84.842%\n",
      "Val accuracy:85.21%\n",
      "Train Acc:84.93%\n",
      "Val accuracy:85.04%\n",
      "Train Acc:84.862%\n",
      "Val accuracy:85.17%\n",
      "Train Acc:84.914%\n",
      "Val accuracy:84.96%\n",
      "Train Acc:84.938%\n",
      "Val accuracy:81.79%\n",
      "Train Acc:84.57%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:84.9%\n",
      "Val accuracy:83.67%\n",
      "Train Acc:84.786%\n",
      "Val accuracy:85.07%\n",
      "Train Acc:84.834%\n",
      "Val accuracy:85.06%\n",
      "Train Acc:84.836%\n",
      "Val accuracy:84.49%\n",
      "Train Acc:84.862%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:84.752%\n",
      "Val accuracy:84.44%\n",
      "Train Acc:84.992%\n",
      "Val accuracy:82.74%\n",
      "Train Acc:84.798%\n",
      "Val accuracy:83.46%\n",
      "Train Acc:84.902%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.866%\n",
      "Val accuracy:84.07%\n",
      "Train Acc:84.922%\n",
      "Val accuracy:84.84%\n",
      "Train Acc:84.592%\n",
      "Val accuracy:85.39%\n",
      "Train Acc:84.64%\n",
      "Val accuracy:84.51%\n",
      "Train Acc:85.02%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:84.902%\n",
      "Val accuracy:83.71%\n",
      "Train Acc:84.95%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:84.666%\n",
      "Val accuracy:84.64%\n",
      "Train Acc:84.716%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:84.96%\n",
      "Val accuracy:84.5%\n",
      "Train Acc:84.818%\n",
      "Val accuracy:84.85%\n",
      "Train Acc:84.9%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:84.894%\n",
      "Val accuracy:84.05%\n",
      "Train Acc:84.59%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.724%\n",
      "Val accuracy:83.66%\n",
      "Train Acc:85.03%\n",
      "Val accuracy:83.91%\n",
      "Train Acc:84.954%\n",
      "Val accuracy:84.22%\n",
      "Train Acc:84.908%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.678%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.856%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:84.154%\n",
      "Val accuracy:84.62%\n",
      "Train Acc:84.434%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:84.038%\n",
      "Val accuracy:83.63%\n",
      "Train Acc:84.764%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:84.908%\n",
      "Val accuracy:84.74%\n",
      "Train Acc:84.588%\n",
      "Val accuracy:85.23%\n",
      "Train Acc:84.876%\n",
      "Val accuracy:83.28%\n",
      "Train Acc:84.882%\n",
      "Val accuracy:85.13%\n",
      "Train Acc:84.856%\n",
      "Val accuracy:84.48%\n",
      "Train Acc:84.932%\n",
      "Val accuracy:84.73%\n",
      "Train Acc:84.802%\n",
      "Val accuracy:84.97%\n",
      "Train Acc:84.798%\n",
      "Val accuracy:84.26%\n",
      "Train Acc:84.946%\n",
      "Val accuracy:84.17%\n",
      "Train Acc:84.806%\n",
      "Val accuracy:84.73%\n",
      "Train Acc:84.908%\n",
      "Val accuracy:85.63%\n",
      "Train Acc:84.896%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:84.898%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:84.67%\n",
      "Val accuracy:84.65%\n",
      "Train Acc:85.24%\n",
      "Val accuracy:85.29%\n",
      "Train Acc:84.242%\n",
      "Val accuracy:81.4%\n",
      "Train Acc:83.908%\n",
      "Val accuracy:84.03%\n",
      "Train Acc:84.672%\n",
      "Val accuracy:84.91%\n",
      "Train Acc:84.574%\n",
      "Val accuracy:84.12%\n",
      "Train Acc:84.722%\n",
      "Val accuracy:85.18%\n",
      "Train Acc:84.688%\n",
      "Val accuracy:84.81%\n",
      "Train Acc:84.622%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:84.404%\n",
      "Val accuracy:84.9%\n",
      "Train Acc:84.656%\n",
      "Val accuracy:85.43%\n",
      "Train Acc:84.818%\n",
      "Val accuracy:82.73%\n",
      "Train Acc:84.736%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:84.866%\n",
      "Val accuracy:84.2%\n",
      "Train Acc:85.008%\n",
      "Val accuracy:84.78%\n",
      "Train Acc:84.988%\n",
      "Val accuracy:84.39%\n",
      "Train Acc:84.556%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:84.912%\n",
      "Val accuracy:82.62%\n",
      "Train Acc:84.588%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:84.948%\n",
      "Val accuracy:83.57%\n",
      "Train Acc:84.698%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:84.986%\n",
      "Val accuracy:84.16%\n",
      "Train Acc:84.792%\n",
      "Val accuracy:85.27%\n",
      "Train Acc:84.754%\n",
      "Val accuracy:84.33%\n",
      "Train Acc:84.874%\n",
      "Val accuracy:82.83%\n",
      "Train Acc:84.814%\n",
      "Val accuracy:83.35%\n",
      "Train Acc:84.384%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:84.562%\n",
      "Val accuracy:81.89%\n",
      "Train Acc:84.756%\n",
      "Val accuracy:82.39%\n",
      "Train Acc:84.808%\n",
      "Val accuracy:85.16%\n",
      "Train Acc:84.832%\n",
      "Val accuracy:84.21%\n",
      "Train Acc:84.736%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:84.742%\n",
      "Val accuracy:85.15%\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Acc:84.874%\n",
      "Val accuracy:83.77%\n",
      "Train Acc:84.842%\n",
      "Val accuracy:84.18%\n",
      "Train Acc:84.71%\n",
      "Val accuracy:84.57%\n",
      "Train Acc:85.122%\n",
      "Val accuracy:84.75%\n",
      "Train Acc:84.932%\n",
      "Val accuracy:84.87%\n",
      "Train Acc:84.94%\n",
      "Val accuracy:84.24%\n",
      "Train Acc:85.14%\n",
      "Val accuracy:83.86%\n",
      "Train Acc:84.954%\n",
      "Val accuracy:85.15%\n",
      "Train Acc:84.414%\n",
      "Val accuracy:84.62%\n",
      "Train Acc:84.658%\n",
      "Val accuracy:82.32%\n",
      "Train Acc:84.748%\n",
      "Val accuracy:84.47%\n",
      "Train Acc:84.75%\n",
      "Val accuracy:84.01%\n",
      "Train Acc:84.78%\n",
      "Val accuracy:83.82%\n",
      "Train Acc:84.73%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:84.234%\n",
      "Val accuracy:84.55%\n",
      "Train Acc:84.74%\n",
      "Val accuracy:84.81%\n",
      "Train Acc:84.758%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.682%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.806%\n",
      "Val accuracy:84.69%\n",
      "Train Acc:84.69%\n",
      "Val accuracy:84.0%\n",
      "Train Acc:84.624%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:84.466%\n",
      "Val accuracy:84.26%\n",
      "Train Acc:84.642%\n",
      "Val accuracy:85.25%\n",
      "Train Acc:84.934%\n",
      "Val accuracy:84.47%\n",
      "Train Acc:83.978%\n",
      "Val accuracy:84.55%\n",
      "Train Acc:84.73%\n",
      "Val accuracy:83.36%\n",
      "Train Acc:84.61%\n",
      "Val accuracy:83.01%\n",
      "Train Acc:83.834%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:84.948%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.884%\n",
      "Val accuracy:83.26%\n",
      "Train Acc:84.99%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:84.446%\n",
      "Val accuracy:84.59%\n",
      "Train Acc:84.514%\n",
      "Val accuracy:84.14%\n",
      "Train Acc:84.788%\n",
      "Val accuracy:83.95%\n",
      "Train Acc:84.84%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:84.56%\n",
      "Val accuracy:81.78%\n",
      "Train Acc:84.73%\n",
      "Val accuracy:84.82%\n",
      "Train Acc:84.774%\n",
      "Val accuracy:85.19%\n",
      "Train Acc:84.602%\n",
      "Val accuracy:84.05%\n",
      "Train Acc:84.728%\n",
      "Val accuracy:84.92%\n",
      "Train Acc:84.84%\n",
      "Val accuracy:84.79%\n",
      "Train Acc:84.602%\n",
      "Val accuracy:84.22%\n",
      "Train Acc:84.834%\n",
      "Val accuracy:83.11%\n",
      "Train Acc:84.068%\n",
      "Val accuracy:82.86%\n",
      "Train Acc:84.796%\n",
      "Val accuracy:85.05%\n",
      "Train Acc:84.53%\n",
      "Val accuracy:85.32%\n",
      "Train Acc:84.842%\n",
      "Val accuracy:83.51%\n",
      "Train Acc:84.626%\n",
      "Val accuracy:82.13%\n",
      "Train Acc:84.564%\n",
      "Val accuracy:84.52%\n",
      "Train Acc:84.3%\n",
      "Val accuracy:84.74%\n",
      "Train Acc:84.47%\n",
      "Val accuracy:83.97%\n",
      "Train Acc:84.644%\n",
      "Val accuracy:83.03%\n",
      "Train Acc:84.466%\n",
      "Val accuracy:83.77%\n",
      "Train Acc:84.67%\n",
      "Val accuracy:84.18%\n",
      "Train Acc:84.63%\n",
      "Val accuracy:85.45%\n",
      "Train Acc:84.73%\n",
      "Val accuracy:85.28%\n",
      "Train Acc:84.834%\n",
      "Val accuracy:84.49%\n",
      "Train Acc:84.762%\n",
      "Val accuracy:83.65%\n",
      "Train Acc:84.738%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.686%\n",
      "Val accuracy:84.88%\n",
      "Train Acc:84.71%\n",
      "Val accuracy:84.43%\n",
      "Train Acc:84.744%\n",
      "Val accuracy:81.63%\n",
      "Train Acc:84.562%\n",
      "Val accuracy:85.47%\n",
      "Train Acc:84.682%\n",
      "Val accuracy:81.88%\n",
      "Train Acc:84.028%\n",
      "Val accuracy:78.01%\n",
      "Train Acc:83.944%\n",
      "Val accuracy:84.27%\n",
      "Train Acc:84.788%\n",
      "Val accuracy:84.41%\n",
      "Train Acc:84.57%\n",
      "Val accuracy:84.15%\n",
      "Train Acc:84.606%\n",
      "Val accuracy:83.3%\n",
      "Train Acc:84.938%\n",
      "Val accuracy:83.15%\n",
      "Train Acc:84.844%\n",
      "Val accuracy:84.09%\n",
      "Train Acc:84.664%\n",
      "Val accuracy:83.27%\n",
      "Train Acc:83.824%\n",
      "Val accuracy:84.9%\n",
      "Train Acc:84.596%\n",
      "Val accuracy:83.12%\n",
      "Train Acc:84.506%\n",
      "Val accuracy:82.98%\n",
      "Train Acc:84.664%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:84.762%\n",
      "Val accuracy:85.1%\n",
      "Train Acc:84.51%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:84.956%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:84.45%\n",
      "Val accuracy:84.0%\n",
      "Train Acc:84.82%\n",
      "Val accuracy:85.0%\n",
      "Train Acc:84.542%\n",
      "Val accuracy:84.17%\n",
      "Train Acc:84.928%\n",
      "Val accuracy:84.69%\n",
      "Train Acc:84.45%\n",
      "Val accuracy:83.83%\n",
      "Train Acc:84.882%\n",
      "Val accuracy:84.56%\n",
      "Train Acc:84.886%\n",
      "Val accuracy:84.85%\n",
      "Train Acc:84.504%\n",
      "Val accuracy:84.23%\n",
      "Train Acc:84.882%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:84.664%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:84.57%\n",
      "Val accuracy:85.24%\n",
      "Train Acc:84.89%\n",
      "Val accuracy:84.89%\n",
      "Train Acc:84.714%\n",
      "Val accuracy:84.93%\n",
      "Train Acc:84.562%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:84.84%\n",
      "Val accuracy:84.76%\n",
      "Train Acc:84.906%\n",
      "Val accuracy:85.11%\n",
      "Train Acc:84.69%\n",
      "Val accuracy:84.66%\n",
      "Train Acc:84.796%\n",
      "Val accuracy:84.22%\n",
      "Train Acc:84.51%\n",
      "Val accuracy:83.32%\n",
      "Train Acc:84.662%\n",
      "Val accuracy:83.79%\n",
      "Train Acc:84.584%\n",
      "Val accuracy:84.98%\n",
      "Train Acc:84.804%\n",
      "Val accuracy:83.4%\n",
      "Train Acc:84.678%\n",
      "Val accuracy:83.94%\n",
      "Train Acc:84.528%\n",
      "Val accuracy:84.28%\n",
      "Train Acc:84.76%\n",
      "Val accuracy:85.01%\n",
      "Train Acc:84.544%\n",
      "Val accuracy:84.26%\n",
      "Train Acc:84.618%\n",
      "Val accuracy:83.85%\n",
      "Train Acc:84.8%\n",
      "Val accuracy:84.58%\n",
      "Train Acc:84.582%\n",
      "Val accuracy:81.17%\n",
      "Train Acc:84.784%\n",
      "Val accuracy:84.42%\n",
      "Train Acc:84.718%\n",
      "Val accuracy:84.6%\n",
      "Train Acc:84.59%\n",
      "Val accuracy:83.72%\n",
      "Train Acc:84.772%\n",
      "Val accuracy:84.53%\n",
      "Train Acc:84.696%\n",
      "Val accuracy:84.74%\n",
      "Train Acc:84.192%\n",
      "Val accuracy:81.96%\n",
      "Train Acc:84.114%\n",
      "Val accuracy:84.32%\n",
      "Train Acc:84.266%\n",
      "Val accuracy:85.42%\n",
      "Train Acc:84.672%\n",
      "Val accuracy:81.31%\n",
      "Train Acc:84.87%\n",
      "Val accuracy:83.75%\n",
      "Train Acc:84.34%\n",
      "Val accuracy:84.72%\n",
      "Train Acc:84.602%\n",
      "Val accuracy:82.63%\n",
      "Train Acc:84.606%\n",
      "Val accuracy:84.62%\n",
      "Train Acc:84.828%\n",
      "Val accuracy:82.43%\n",
      "Train Acc:84.802%\n",
      "Val accuracy:83.54%\n",
      "Train Acc:84.474%\n",
      "Val accuracy:84.22%\n",
      "Train Acc:84.574%\n",
      "Val accuracy:84.99%\n",
      "Train Acc:84.86%\n",
      "Val accuracy:84.56%\n",
      "Train Acc:83.782%\n",
      "Val accuracy:85.03%\n",
      "Train Acc:84.696%\n",
      "Val accuracy:83.93%\n",
      "Train Acc:84.486%\n",
      "Val accuracy:82.95%\n",
      "Train Acc:84.148%\n",
      "Val accuracy:84.61%\n"
     ]
    }
   ],
   "source": [
    "for epoch in range(epochs):\n",
    "    classifier.train()\n",
    "    running_loss = 0.0\n",
    "    running_acc = 0\n",
    "    for img, y in train_loader:\n",
    "        img = img.view(img.size(0), -1)\n",
    "        img, y = img.to(device), y.to(device)\n",
    "        # ===================forward=====================\n",
    "        code_vector = ae.encoder(img)\n",
    "        y_hat = classifier(code_vector)\n",
    "        loss = classifier_criterion(y_hat, y)\n",
    "        # ===================backward====================\n",
    "        classifier_optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        classifier_optimizer.step()\n",
    "\n",
    "        # print statistics\n",
    "        running_loss += loss.item() * img.size(0)\n",
    "        out = torch.argmax(y_hat.detach(),dim=1)\n",
    "        assert out.shape==y.shape\n",
    "        running_acc += (y==out).sum().item()\n",
    "    print(f\"Train Acc:{running_acc*100/len(train_dataset)}%\")\n",
    "\n",
    "    correct = 0\n",
    "    classifier.eval()\n",
    "    with torch.no_grad():\n",
    "        for img, y in val_loader:\n",
    "            img = img.view(img.size(0), -1)\n",
    "            img, y = img.cuda(), y.cuda()\n",
    "            code_vector = ae.encoder(img)\n",
    "            y_hat = classifier(code_vector)\n",
    "            y_hat = torch.argmax(y_hat, dim=1)\n",
    "            acc = (y_hat==y).sum().item()\n",
    "            correct += acc\n",
    "    print(f\"Val accuracy:{correct*100/len(val_dataset)}%\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "25d5cb94",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy:84.75%\n"
     ]
    }
   ],
   "source": [
    "# evaluate your model on the never-seen-before test data\n",
    "classifier.eval()\n",
    "correct = 0\n",
    "with torch.no_grad():\n",
    "    for img, y in test_loader:\n",
    "        img = img.view(img.size(0), -1)\n",
    "        img, y = img.cuda(), y.cuda()\n",
    "        code_vector = ae.encoder(img)\n",
    "        y_hat = classifier(code_vector)\n",
    "        y_hat = torch.argmax(y_hat, dim=1)\n",
    "        acc = (y_hat==y).sum().item()\n",
    "        correct += acc\n",
    "    print(f\"Test accuracy:{correct*100/len(test_dataset)}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "edcf9e07",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "336c19bc",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "23476da2",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "106fc325",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e669ad60",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d300180a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2a350570",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a10b66a8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e12e49fd",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2b4b3bed",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "150ff1e5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c3b24a80",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ab41d59a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "34b51d3d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96e105cf",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "828b6320",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f7fbcfc0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a605b756",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}