Federated learning on CIFAR-10 using CNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "failing-middle",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision\n",
    "import syft as sy\n",
    "from torch import nn\n",
    "import torch.optim as optim\n",
    "import torch.nn.functional as F\n",
    "from torchvision import datasets, transforms\n",
    "from torchvision.datasets import ImageFolder\n",
    "from torchvision.transforms import ToTensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "beginning-cassette",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = './data/cifar10'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "enhanced-obligation",
   "metadata": {},
   "outputs": [],
   "source": [
    "hook = sy.TorchHook(torch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "turkish-turning",
   "metadata": {},
   "outputs": [],
   "source": [
    "jake = sy.VirtualWorker(hook, id=\"jake\")\n",
    "john = sy.VirtualWorker(hook, id=\"john\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "stupid-metallic",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = ImageFolder(data_dir + '/train', transform = ToTensor())\n",
    "test_dataset = ImageFolder(data_dir + '/test', transform = ToTensor())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ranging-clearing",
   "metadata": {},
   "outputs": [],
   "source": [
    "federated_train_loader = sy.FederatedDataLoader(\n",
    "    train_dataset.federate((jake, john)), batch_size=64, shuffle=True)\n",
    "\n",
    "test_loader = torch.utils.data.DataLoader(\n",
    "    test_dataset, batch_size=64, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "future-pennsylvania",
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_batch(dl):\n",
    "    for images, labels in dl:\n",
    "        print(images.size())\n",
    "        print(labels.size())\n",
    "        break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "complete-disability",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([64, 3, 32, 32])\n",
      "torch.Size([64])\n"
     ]
    }
   ],
   "source": [
    "show_batch(test_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "transsexual-seeking",
   "metadata": {},
   "outputs": [],
   "source": [
    "# define the args\n",
    "args = {\n",
    "    'use_cuda' : True,\n",
    "    'batch_size' : 64,\n",
    "    'test_batch_size' : 1000,\n",
    "    'lr' : 0.01,\n",
    "    'log_interval' : 10,\n",
    "    'epochs' : 10\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "covered-messenger",
   "metadata": {},
   "outputs": [],
   "source": [
    "# create a simple CNN net\n",
    "class Net(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        \n",
    "        self.conv = nn.Sequential(\n",
    "            nn.Conv2d(in_channels = 3, out_channels = 32, kernel_size = 3, stride = 1,padding=1),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(2,2), # output bs*32*16*16\n",
    "            nn.Conv2d(in_channels=32,out_channels = 64, kernel_size = 3, stride = 1),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(2,2) # output bs*64*2*2\n",
    "        )\n",
    "        \n",
    "        self.fc = nn.Sequential(\n",
    "            nn.Linear(in_features=576, out_features=128),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(in_features=128, out_features=10),\n",
    "        )\n",
    "\n",
    "        self.dropout = nn.Dropout2d(0.25)\n",
    "    \n",
    "    def forward(self, x):\n",
    "        x = self.conv(x)\n",
    "        x = F.max_pool2d(x,2)\n",
    "        x = x.view(-1, 576)\n",
    "        x = self.fc(x)\n",
    "        x = F.log_softmax(x, dim=1)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "decimal-sport",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train(args, model, device, train_loader, optimizer, epoch):\n",
    "    model.train()\n",
    "\n",
    "    # iterate over federated data\n",
    "    for batch_idx, (data, target) in enumerate(train_loader):\n",
    "\n",
    "        # send the model to the remote location \n",
    "        model = model.send(data.location)\n",
    "\n",
    "        # the same torch code that we are use to\n",
    "        data, target = data.to(device), target.to(device)\n",
    "        optimizer.zero_grad()\n",
    "        output = model(data)\n",
    "\n",
    "        # this loss is a ptr to the tensor loss \n",
    "        # at the remote location\n",
    "        loss = F.nll_loss(output, target)\n",
    "\n",
    "        # call backward() on the loss ptr,\n",
    "        # that will send the command to call\n",
    "        # backward on the actual loss tensor\n",
    "        # present on the remote machine\n",
    "        loss.backward()\n",
    "\n",
    "        optimizer.step()\n",
    "\n",
    "        # get back the updated model\n",
    "        model.get()\n",
    "\n",
    "        if batch_idx % args['log_interval'] == 0:\n",
    "\n",
    "            # a thing to note is the variable loss was\n",
    "            # also created at remote worker, so we need to\n",
    "            # explicitly get it back\n",
    "            loss = loss.get()\n",
    "\n",
    "            print('Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n",
    "                    epoch, \n",
    "                    batch_idx * args['batch_size'], # no of images done\n",
    "                    len(train_loader) * args['batch_size'], # total images left\n",
    "                    100. * batch_idx / len(train_loader), \n",
    "                    loss.item()\n",
    "                )\n",
    "            )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "linear-inspector",
   "metadata": {},
   "outputs": [],
   "source": [
    "def test(model, device, test_loader):\n",
    "    model.eval()\n",
    "    test_loss = 0\n",
    "    correct = 0\n",
    "    with torch.no_grad():\n",
    "        for data, target in test_loader:\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "\n",
    "            # add losses together\n",
    "            test_loss += F.nll_loss(output, target, reduction='sum').item() \n",
    "\n",
    "            # get the index of the max probability class\n",
    "            pred = output.argmax(dim=1, keepdim=True)  \n",
    "            correct += pred.eq(target.view_as(pred)).sum().item()\n",
    "\n",
    "    test_loss /= len(test_loader.dataset)\n",
    "\n",
    "    print('\\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\\n'.format(\n",
    "        test_loss, correct, len(test_loader.dataset),\n",
    "        100. * correct / len(test_loader.dataset)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "committed-disorder",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Epoch: 1 [0/50048 (0%)]\tLoss: 2.304238\n",
      "Train Epoch: 1 [640/50048 (1%)]\tLoss: 2.250695\n",
      "Train Epoch: 1 [1280/50048 (3%)]\tLoss: 2.210575\n",
      "Train Epoch: 1 [1920/50048 (4%)]\tLoss: 2.157618\n",
      "Train Epoch: 1 [2560/50048 (5%)]\tLoss: 2.097458\n",
      "Train Epoch: 1 [3200/50048 (6%)]\tLoss: 2.014193\n",
      "Train Epoch: 1 [3840/50048 (8%)]\tLoss: 1.903906\n",
      "Train Epoch: 1 [4480/50048 (9%)]\tLoss: 1.790621\n",
      "Train Epoch: 1 [5120/50048 (10%)]\tLoss: 1.699381\n",
      "Train Epoch: 1 [5760/50048 (12%)]\tLoss: 1.668174\n",
      "Train Epoch: 1 [6400/50048 (13%)]\tLoss: 1.640497\n",
      "Train Epoch: 1 [7040/50048 (14%)]\tLoss: 1.620039\n",
      "Train Epoch: 1 [7680/50048 (15%)]\tLoss: 1.621258\n",
      "Train Epoch: 1 [8320/50048 (17%)]\tLoss: 1.632844\n",
      "Train Epoch: 1 [8960/50048 (18%)]\tLoss: 1.612172\n",
      "Train Epoch: 1 [9600/50048 (19%)]\tLoss: 1.626764\n",
      "Train Epoch: 1 [10240/50048 (20%)]\tLoss: 1.602517\n",
      "Train Epoch: 1 [10880/50048 (22%)]\tLoss: 1.606276\n",
      "Train Epoch: 1 [11520/50048 (23%)]\tLoss: 1.624046\n",
      "Train Epoch: 1 [12160/50048 (24%)]\tLoss: 1.611804\n",
      "Train Epoch: 1 [12800/50048 (26%)]\tLoss: 1.598474\n",
      "Train Epoch: 1 [13440/50048 (27%)]\tLoss: 1.609613\n",
      "Train Epoch: 1 [14080/50048 (28%)]\tLoss: 1.611302\n",
      "Train Epoch: 1 [14720/50048 (29%)]\tLoss: 1.596290\n",
      "Train Epoch: 1 [15360/50048 (31%)]\tLoss: 1.604798\n",
      "Train Epoch: 1 [16000/50048 (32%)]\tLoss: 1.611980\n",
      "Train Epoch: 1 [16640/50048 (33%)]\tLoss: 1.579297\n",
      "Train Epoch: 1 [17280/50048 (35%)]\tLoss: 1.628582\n",
      "Train Epoch: 1 [17920/50048 (36%)]\tLoss: 1.606954\n",
      "Train Epoch: 1 [18560/50048 (37%)]\tLoss: 1.599662\n",
      "Train Epoch: 1 [19200/50048 (38%)]\tLoss: 1.607485\n",
      "Train Epoch: 1 [19840/50048 (40%)]\tLoss: 1.630230\n",
      "Train Epoch: 1 [20480/50048 (41%)]\tLoss: 1.577389\n",
      "Train Epoch: 1 [21120/50048 (42%)]\tLoss: 1.566212\n",
      "Train Epoch: 1 [21760/50048 (43%)]\tLoss: 1.590440\n",
      "Train Epoch: 1 [22400/50048 (45%)]\tLoss: 1.578603\n",
      "Train Epoch: 1 [23040/50048 (46%)]\tLoss: 1.584203\n",
      "Train Epoch: 1 [23680/50048 (47%)]\tLoss: 1.606847\n",
      "Train Epoch: 1 [24320/50048 (49%)]\tLoss: 1.593112\n",
      "Train Epoch: 1 [24960/50048 (50%)]\tLoss: 1.635808\n",
      "Train Epoch: 1 [25600/50048 (51%)]\tLoss: 2.665464\n",
      "Train Epoch: 1 [26240/50048 (52%)]\tLoss: 2.221110\n",
      "Train Epoch: 1 [26880/50048 (54%)]\tLoss: 2.012409\n",
      "Train Epoch: 1 [27520/50048 (55%)]\tLoss: 1.855983\n",
      "Train Epoch: 1 [28160/50048 (56%)]\tLoss: 1.722438\n",
      "Train Epoch: 1 [28800/50048 (58%)]\tLoss: 1.668542\n",
      "Train Epoch: 1 [29440/50048 (59%)]\tLoss: 1.625824\n",
      "Train Epoch: 1 [30080/50048 (60%)]\tLoss: 1.622784\n",
      "Train Epoch: 1 [30720/50048 (61%)]\tLoss: 1.620621\n",
      "Train Epoch: 1 [31360/50048 (63%)]\tLoss: 1.634574\n",
      "Train Epoch: 1 [32000/50048 (64%)]\tLoss: 1.589327\n",
      "Train Epoch: 1 [32640/50048 (65%)]\tLoss: 1.599848\n",
      "Train Epoch: 1 [33280/50048 (66%)]\tLoss: 1.598874\n",
      "Train Epoch: 1 [33920/50048 (68%)]\tLoss: 1.642654\n",
      "Train Epoch: 1 [34560/50048 (69%)]\tLoss: 1.575226\n",
      "Train Epoch: 1 [35200/50048 (70%)]\tLoss: 1.585621\n",
      "Train Epoch: 1 [35840/50048 (72%)]\tLoss: 1.575716\n",
      "Train Epoch: 1 [36480/50048 (73%)]\tLoss: 1.568301\n",
      "Train Epoch: 1 [37120/50048 (74%)]\tLoss: 1.561391\n",
      "Train Epoch: 1 [37760/50048 (75%)]\tLoss: 1.556981\n",
      "Train Epoch: 1 [38400/50048 (77%)]\tLoss: 1.511768\n",
      "Train Epoch: 1 [39040/50048 (78%)]\tLoss: 1.525153\n",
      "Train Epoch: 1 [39680/50048 (79%)]\tLoss: 1.509150\n",
      "Train Epoch: 1 [40320/50048 (81%)]\tLoss: 1.537958\n",
      "Train Epoch: 1 [40960/50048 (82%)]\tLoss: 1.560348\n",
      "Train Epoch: 1 [41600/50048 (83%)]\tLoss: 1.583375\n",
      "Train Epoch: 1 [42240/50048 (84%)]\tLoss: 1.568494\n",
      "Train Epoch: 1 [42880/50048 (86%)]\tLoss: 1.541867\n",
      "Train Epoch: 1 [43520/50048 (87%)]\tLoss: 1.515552\n",
      "Train Epoch: 1 [44160/50048 (88%)]\tLoss: 1.523981\n",
      "Train Epoch: 1 [44800/50048 (90%)]\tLoss: 1.533148\n",
      "Train Epoch: 1 [45440/50048 (91%)]\tLoss: 1.510229\n",
      "Train Epoch: 1 [46080/50048 (92%)]\tLoss: 1.481247\n",
      "Train Epoch: 1 [46720/50048 (93%)]\tLoss: 1.520732\n",
      "Train Epoch: 1 [47360/50048 (95%)]\tLoss: 1.509005\n",
      "Train Epoch: 1 [48000/50048 (96%)]\tLoss: 1.505210\n",
      "Train Epoch: 1 [48640/50048 (97%)]\tLoss: 1.448666\n",
      "Train Epoch: 1 [49280/50048 (98%)]\tLoss: 1.453850\n",
      "Train Epoch: 1 [49920/50048 (100%)]\tLoss: 1.561555\n",
      "\n",
      "Test set: Average loss: 5.1593, Accuracy: 1357/10000 (14%)\n",
      "\n",
      "Train Epoch: 2 [0/50048 (0%)]\tLoss: 8.706122\n",
      "Train Epoch: 2 [640/50048 (1%)]\tLoss: 2.104597\n",
      "Train Epoch: 2 [1280/50048 (3%)]\tLoss: 1.787595\n",
      "Train Epoch: 2 [1920/50048 (4%)]\tLoss: 1.664378\n",
      "Train Epoch: 2 [2560/50048 (5%)]\tLoss: 1.675770\n",
      "Train Epoch: 2 [3200/50048 (6%)]\tLoss: 1.600927\n",
      "Train Epoch: 2 [3840/50048 (8%)]\tLoss: 1.636781\n",
      "Train Epoch: 2 [4480/50048 (9%)]\tLoss: 1.622145\n",
      "Train Epoch: 2 [5120/50048 (10%)]\tLoss: 1.570470\n",
      "Train Epoch: 2 [5760/50048 (12%)]\tLoss: 1.631598\n",
      "Train Epoch: 2 [6400/50048 (13%)]\tLoss: 1.563763\n",
      "Train Epoch: 2 [7040/50048 (14%)]\tLoss: 1.606031\n",
      "Train Epoch: 2 [7680/50048 (15%)]\tLoss: 1.580330\n",
      "Train Epoch: 2 [8320/50048 (17%)]\tLoss: 1.575917\n",
      "Train Epoch: 2 [8960/50048 (18%)]\tLoss: 1.623463\n",
      "Train Epoch: 2 [9600/50048 (19%)]\tLoss: 1.514594\n",
      "Train Epoch: 2 [10240/50048 (20%)]\tLoss: 1.563731\n",
      "Train Epoch: 2 [10880/50048 (22%)]\tLoss: 1.594806\n",
      "Train Epoch: 2 [11520/50048 (23%)]\tLoss: 1.570984\n",
      "Train Epoch: 2 [12160/50048 (24%)]\tLoss: 1.542361\n",
      "Train Epoch: 2 [12800/50048 (26%)]\tLoss: 1.475902\n",
      "Train Epoch: 2 [13440/50048 (27%)]\tLoss: 1.504755\n",
      "Train Epoch: 2 [14080/50048 (28%)]\tLoss: 1.506225\n",
      "Train Epoch: 2 [14720/50048 (29%)]\tLoss: 1.505505\n",
      "Train Epoch: 2 [15360/50048 (31%)]\tLoss: 1.513716\n",
      "Train Epoch: 2 [16000/50048 (32%)]\tLoss: 1.546614\n",
      "Train Epoch: 2 [16640/50048 (33%)]\tLoss: 1.539091\n",
      "Train Epoch: 2 [17280/50048 (35%)]\tLoss: 1.462804\n",
      "Train Epoch: 2 [17920/50048 (36%)]\tLoss: 1.571598\n",
      "Train Epoch: 2 [18560/50048 (37%)]\tLoss: 1.503482\n",
      "Train Epoch: 2 [19200/50048 (38%)]\tLoss: 1.567404\n",
      "Train Epoch: 2 [19840/50048 (40%)]\tLoss: 1.548216\n",
      "Train Epoch: 2 [20480/50048 (41%)]\tLoss: 1.444099\n",
      "Train Epoch: 2 [21120/50048 (42%)]\tLoss: 1.488720\n",
      "Train Epoch: 2 [21760/50048 (43%)]\tLoss: 1.518271\n",
      "Train Epoch: 2 [22400/50048 (45%)]\tLoss: 1.521405\n",
      "Train Epoch: 2 [23040/50048 (46%)]\tLoss: 1.549800\n",
      "Train Epoch: 2 [23680/50048 (47%)]\tLoss: 1.593049\n",
      "Train Epoch: 2 [24320/50048 (49%)]\tLoss: 1.509757\n",
      "Train Epoch: 2 [24960/50048 (50%)]\tLoss: 1.519997\n",
      "Train Epoch: 2 [25600/50048 (51%)]\tLoss: 1.833127\n",
      "Train Epoch: 2 [26240/50048 (52%)]\tLoss: 1.615171\n",
      "Train Epoch: 2 [26880/50048 (54%)]\tLoss: 1.536045\n",
      "Train Epoch: 2 [27520/50048 (55%)]\tLoss: 1.466292\n",
      "Train Epoch: 2 [28160/50048 (56%)]\tLoss: 1.383344\n",
      "Train Epoch: 2 [28800/50048 (58%)]\tLoss: 1.374533\n",
      "Train Epoch: 2 [29440/50048 (59%)]\tLoss: 1.418957\n",
      "Train Epoch: 2 [30080/50048 (60%)]\tLoss: 1.420178\n",
      "Train Epoch: 2 [30720/50048 (61%)]\tLoss: 1.481364\n",
      "Train Epoch: 2 [31360/50048 (63%)]\tLoss: 1.287103\n",
      "Train Epoch: 2 [32000/50048 (64%)]\tLoss: 1.412184\n",
      "Train Epoch: 2 [32640/50048 (65%)]\tLoss: 1.313678\n",
      "Train Epoch: 2 [33280/50048 (66%)]\tLoss: 1.560602\n",
      "Train Epoch: 2 [33920/50048 (68%)]\tLoss: 1.374440\n",
      "Train Epoch: 2 [34560/50048 (69%)]\tLoss: 1.377987\n",
      "Train Epoch: 2 [35200/50048 (70%)]\tLoss: 1.355363\n",
      "Train Epoch: 2 [35840/50048 (72%)]\tLoss: 1.362227\n",
      "Train Epoch: 2 [36480/50048 (73%)]\tLoss: 1.439060\n",
      "Train Epoch: 2 [37120/50048 (74%)]\tLoss: 1.262917\n",
      "Train Epoch: 2 [37760/50048 (75%)]\tLoss: 1.318050\n",
      "Train Epoch: 2 [38400/50048 (77%)]\tLoss: 1.316417\n",
      "Train Epoch: 2 [39040/50048 (78%)]\tLoss: 1.479026\n",
      "Train Epoch: 2 [39680/50048 (79%)]\tLoss: 1.412493\n",
      "Train Epoch: 2 [40320/50048 (81%)]\tLoss: 1.225443\n",
      "Train Epoch: 2 [40960/50048 (82%)]\tLoss: 1.330498\n",
      "Train Epoch: 2 [41600/50048 (83%)]\tLoss: 1.275422\n",
      "Train Epoch: 2 [42240/50048 (84%)]\tLoss: 1.194156\n",
      "Train Epoch: 2 [42880/50048 (86%)]\tLoss: 1.306020\n",
      "Train Epoch: 2 [43520/50048 (87%)]\tLoss: 1.240214\n",
      "Train Epoch: 2 [44160/50048 (88%)]\tLoss: 1.375097\n",
      "Train Epoch: 2 [44800/50048 (90%)]\tLoss: 1.417889\n",
      "Train Epoch: 2 [45440/50048 (91%)]\tLoss: 1.337082\n",
      "Train Epoch: 2 [46080/50048 (92%)]\tLoss: 1.231147\n",
      "Train Epoch: 2 [46720/50048 (93%)]\tLoss: 1.280269\n",
      "Train Epoch: 2 [47360/50048 (95%)]\tLoss: 1.297153\n",
      "Train Epoch: 2 [48000/50048 (96%)]\tLoss: 1.224588\n",
      "Train Epoch: 2 [48640/50048 (97%)]\tLoss: 1.221909\n",
      "Train Epoch: 2 [49280/50048 (98%)]\tLoss: 1.267006\n",
      "Train Epoch: 2 [49920/50048 (100%)]\tLoss: 1.281871\n",
      "\n",
      "Test set: Average loss: 4.4333, Accuracy: 2130/10000 (21%)\n",
      "\n",
      "Train Epoch: 3 [0/50048 (0%)]\tLoss: 7.495867\n",
      "Train Epoch: 3 [640/50048 (1%)]\tLoss: 1.568213\n",
      "Train Epoch: 3 [1280/50048 (3%)]\tLoss: 1.503882\n",
      "Train Epoch: 3 [1920/50048 (4%)]\tLoss: 1.481274\n",
      "Train Epoch: 3 [2560/50048 (5%)]\tLoss: 1.551211\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Epoch: 3 [3200/50048 (6%)]\tLoss: 1.394798\n",
      "Train Epoch: 3 [3840/50048 (8%)]\tLoss: 1.520364\n",
      "Train Epoch: 3 [4480/50048 (9%)]\tLoss: 1.551208\n",
      "Train Epoch: 3 [5120/50048 (10%)]\tLoss: 1.353395\n",
      "Train Epoch: 3 [5760/50048 (12%)]\tLoss: 1.410080\n",
      "Train Epoch: 3 [6400/50048 (13%)]\tLoss: 1.439280\n",
      "Train Epoch: 3 [7040/50048 (14%)]\tLoss: 1.365974\n",
      "Train Epoch: 3 [7680/50048 (15%)]\tLoss: 1.405336\n",
      "Train Epoch: 3 [8320/50048 (17%)]\tLoss: 1.416731\n",
      "Train Epoch: 3 [8960/50048 (18%)]\tLoss: 1.295581\n",
      "Train Epoch: 3 [9600/50048 (19%)]\tLoss: 1.426612\n",
      "Train Epoch: 3 [10240/50048 (20%)]\tLoss: 1.327028\n",
      "Train Epoch: 3 [10880/50048 (22%)]\tLoss: 1.553579\n",
      "Train Epoch: 3 [11520/50048 (23%)]\tLoss: 1.457275\n",
      "Train Epoch: 3 [12160/50048 (24%)]\tLoss: 1.487447\n",
      "Train Epoch: 3 [12800/50048 (26%)]\tLoss: 1.392838\n",
      "Train Epoch: 3 [13440/50048 (27%)]\tLoss: 1.276032\n",
      "Train Epoch: 3 [14080/50048 (28%)]\tLoss: 1.387896\n",
      "Train Epoch: 3 [14720/50048 (29%)]\tLoss: 1.339422\n",
      "Train Epoch: 3 [15360/50048 (31%)]\tLoss: 1.464117\n",
      "Train Epoch: 3 [16000/50048 (32%)]\tLoss: 1.348402\n",
      "Train Epoch: 3 [16640/50048 (33%)]\tLoss: 1.359259\n",
      "Train Epoch: 3 [17280/50048 (35%)]\tLoss: 1.348100\n",
      "Train Epoch: 3 [17920/50048 (36%)]\tLoss: 1.308972\n",
      "Train Epoch: 3 [18560/50048 (37%)]\tLoss: 1.263574\n",
      "Train Epoch: 3 [19200/50048 (38%)]\tLoss: 1.416123\n",
      "Train Epoch: 3 [19840/50048 (40%)]\tLoss: 1.327332\n",
      "Train Epoch: 3 [20480/50048 (41%)]\tLoss: 1.394418\n",
      "Train Epoch: 3 [21120/50048 (42%)]\tLoss: 1.370494\n",
      "Train Epoch: 3 [21760/50048 (43%)]\tLoss: 1.351350\n",
      "Train Epoch: 3 [22400/50048 (45%)]\tLoss: 1.412218\n",
      "Train Epoch: 3 [23040/50048 (46%)]\tLoss: 1.322882\n",
      "Train Epoch: 3 [23680/50048 (47%)]\tLoss: 1.319573\n",
      "Train Epoch: 3 [24320/50048 (49%)]\tLoss: 1.247164\n",
      "Train Epoch: 3 [24960/50048 (50%)]\tLoss: 1.250493\n",
      "Train Epoch: 3 [25600/50048 (51%)]\tLoss: 1.514456\n",
      "Train Epoch: 3 [26240/50048 (52%)]\tLoss: 1.224237\n",
      "Train Epoch: 3 [26880/50048 (54%)]\tLoss: 1.155195\n",
      "Train Epoch: 3 [27520/50048 (55%)]\tLoss: 1.348619\n",
      "Train Epoch: 3 [28160/50048 (56%)]\tLoss: 1.226561\n",
      "Train Epoch: 3 [28800/50048 (58%)]\tLoss: 1.061426\n",
      "Train Epoch: 3 [29440/50048 (59%)]\tLoss: 1.161469\n",
      "Train Epoch: 3 [30080/50048 (60%)]\tLoss: 1.286068\n",
      "Train Epoch: 3 [30720/50048 (61%)]\tLoss: 1.127025\n",
      "Train Epoch: 3 [31360/50048 (63%)]\tLoss: 1.163910\n",
      "Train Epoch: 3 [32000/50048 (64%)]\tLoss: 1.165989\n",
      "Train Epoch: 3 [32640/50048 (65%)]\tLoss: 1.389572\n",
      "Train Epoch: 3 [33280/50048 (66%)]\tLoss: 1.035576\n",
      "Train Epoch: 3 [33920/50048 (68%)]\tLoss: 1.180991\n",
      "Train Epoch: 3 [34560/50048 (69%)]\tLoss: 1.051112\n",
      "Train Epoch: 3 [35200/50048 (70%)]\tLoss: 1.225350\n",
      "Train Epoch: 3 [35840/50048 (72%)]\tLoss: 1.207252\n",
      "Train Epoch: 3 [36480/50048 (73%)]\tLoss: 1.271439\n",
      "Train Epoch: 3 [37120/50048 (74%)]\tLoss: 1.239046\n",
      "Train Epoch: 3 [37760/50048 (75%)]\tLoss: 1.205631\n",
      "Train Epoch: 3 [38400/50048 (77%)]\tLoss: 1.092178\n",
      "Train Epoch: 3 [39040/50048 (78%)]\tLoss: 1.119560\n",
      "Train Epoch: 3 [39680/50048 (79%)]\tLoss: 1.215153\n",
      "Train Epoch: 3 [40320/50048 (81%)]\tLoss: 1.249548\n",
      "Train Epoch: 3 [40960/50048 (82%)]\tLoss: 1.108177\n",
      "Train Epoch: 3 [41600/50048 (83%)]\tLoss: 1.153845\n",
      "Train Epoch: 3 [42240/50048 (84%)]\tLoss: 1.205850\n",
      "Train Epoch: 3 [42880/50048 (86%)]\tLoss: 0.994531\n",
      "Train Epoch: 3 [43520/50048 (87%)]\tLoss: 1.135702\n",
      "Train Epoch: 3 [44160/50048 (88%)]\tLoss: 1.058622\n",
      "Train Epoch: 3 [44800/50048 (90%)]\tLoss: 1.012974\n",
      "Train Epoch: 3 [45440/50048 (91%)]\tLoss: 1.255570\n",
      "Train Epoch: 3 [46080/50048 (92%)]\tLoss: 1.297675\n",
      "Train Epoch: 3 [46720/50048 (93%)]\tLoss: 1.275774\n",
      "Train Epoch: 3 [47360/50048 (95%)]\tLoss: 1.067070\n",
      "Train Epoch: 3 [48000/50048 (96%)]\tLoss: 1.293804\n",
      "Train Epoch: 3 [48640/50048 (97%)]\tLoss: 1.313698\n",
      "Train Epoch: 3 [49280/50048 (98%)]\tLoss: 1.128651\n",
      "Train Epoch: 3 [49920/50048 (100%)]\tLoss: 1.025034\n",
      "\n",
      "Test set: Average loss: 4.3404, Accuracy: 2489/10000 (25%)\n",
      "\n",
      "Train Epoch: 4 [0/50048 (0%)]\tLoss: 7.458377\n",
      "Train Epoch: 4 [640/50048 (1%)]\tLoss: 1.552834\n",
      "Train Epoch: 4 [1280/50048 (3%)]\tLoss: 1.316447\n",
      "Train Epoch: 4 [1920/50048 (4%)]\tLoss: 1.341854\n",
      "Train Epoch: 4 [2560/50048 (5%)]\tLoss: 1.346247\n",
      "Train Epoch: 4 [3200/50048 (6%)]\tLoss: 1.206120\n",
      "Train Epoch: 4 [3840/50048 (8%)]\tLoss: 1.194810\n",
      "Train Epoch: 4 [4480/50048 (9%)]\tLoss: 1.269112\n",
      "Train Epoch: 4 [5120/50048 (10%)]\tLoss: 1.198100\n",
      "Train Epoch: 4 [5760/50048 (12%)]\tLoss: 1.217504\n",
      "Train Epoch: 4 [6400/50048 (13%)]\tLoss: 1.339452\n",
      "Train Epoch: 4 [7040/50048 (14%)]\tLoss: 1.151266\n",
      "Train Epoch: 4 [7680/50048 (15%)]\tLoss: 1.336123\n",
      "Train Epoch: 4 [8320/50048 (17%)]\tLoss: 1.242143\n",
      "Train Epoch: 4 [8960/50048 (18%)]\tLoss: 1.151185\n",
      "Train Epoch: 4 [9600/50048 (19%)]\tLoss: 1.171404\n",
      "Train Epoch: 4 [10240/50048 (20%)]\tLoss: 1.184507\n",
      "Train Epoch: 4 [10880/50048 (22%)]\tLoss: 1.322774\n",
      "Train Epoch: 4 [11520/50048 (23%)]\tLoss: 1.240797\n",
      "Train Epoch: 4 [12160/50048 (24%)]\tLoss: 1.239441\n",
      "Train Epoch: 4 [12800/50048 (26%)]\tLoss: 1.208549\n",
      "Train Epoch: 4 [13440/50048 (27%)]\tLoss: 1.241014\n",
      "Train Epoch: 4 [14080/50048 (28%)]\tLoss: 1.235329\n",
      "Train Epoch: 4 [14720/50048 (29%)]\tLoss: 1.243130\n",
      "Train Epoch: 4 [15360/50048 (31%)]\tLoss: 1.399240\n",
      "Train Epoch: 4 [16000/50048 (32%)]\tLoss: 1.211406\n",
      "Train Epoch: 4 [16640/50048 (33%)]\tLoss: 1.289153\n",
      "Train Epoch: 4 [17280/50048 (35%)]\tLoss: 1.111959\n",
      "Train Epoch: 4 [17920/50048 (36%)]\tLoss: 1.208973\n",
      "Train Epoch: 4 [18560/50048 (37%)]\tLoss: 1.143233\n",
      "Train Epoch: 4 [19200/50048 (38%)]\tLoss: 1.229719\n",
      "Train Epoch: 4 [19840/50048 (40%)]\tLoss: 1.260606\n",
      "Train Epoch: 4 [20480/50048 (41%)]\tLoss: 1.086356\n",
      "Train Epoch: 4 [21120/50048 (42%)]\tLoss: 1.255345\n",
      "Train Epoch: 4 [21760/50048 (43%)]\tLoss: 1.218208\n",
      "Train Epoch: 4 [22400/50048 (45%)]\tLoss: 1.103989\n",
      "Train Epoch: 4 [23040/50048 (46%)]\tLoss: 1.046784\n",
      "Train Epoch: 4 [23680/50048 (47%)]\tLoss: 1.213380\n",
      "Train Epoch: 4 [24320/50048 (49%)]\tLoss: 1.157421\n",
      "Train Epoch: 4 [24960/50048 (50%)]\tLoss: 1.005803\n",
      "Train Epoch: 4 [25600/50048 (51%)]\tLoss: 1.095242\n",
      "Train Epoch: 4 [26240/50048 (52%)]\tLoss: 1.047616\n",
      "Train Epoch: 4 [26880/50048 (54%)]\tLoss: 1.060352\n",
      "Train Epoch: 4 [27520/50048 (55%)]\tLoss: 1.202096\n",
      "Train Epoch: 4 [28160/50048 (56%)]\tLoss: 1.065372\n",
      "Train Epoch: 4 [28800/50048 (58%)]\tLoss: 0.996695\n",
      "Train Epoch: 4 [29440/50048 (59%)]\tLoss: 0.848151\n",
      "Train Epoch: 4 [30080/50048 (60%)]\tLoss: 0.878699\n",
      "Train Epoch: 4 [30720/50048 (61%)]\tLoss: 1.083167\n",
      "Train Epoch: 4 [31360/50048 (63%)]\tLoss: 1.092551\n",
      "Train Epoch: 4 [32000/50048 (64%)]\tLoss: 0.931217\n",
      "Train Epoch: 4 [32640/50048 (65%)]\tLoss: 1.122574\n",
      "Train Epoch: 4 [33280/50048 (66%)]\tLoss: 1.084466\n",
      "Train Epoch: 4 [33920/50048 (68%)]\tLoss: 0.956313\n",
      "Train Epoch: 4 [34560/50048 (69%)]\tLoss: 1.197949\n",
      "Train Epoch: 4 [35200/50048 (70%)]\tLoss: 1.125830\n",
      "Train Epoch: 4 [35840/50048 (72%)]\tLoss: 1.062725\n",
      "Train Epoch: 4 [36480/50048 (73%)]\tLoss: 0.888798\n",
      "Train Epoch: 4 [37120/50048 (74%)]\tLoss: 1.150556\n",
      "Train Epoch: 4 [37760/50048 (75%)]\tLoss: 0.904847\n",
      "Train Epoch: 4 [38400/50048 (77%)]\tLoss: 1.076064\n",
      "Train Epoch: 4 [39040/50048 (78%)]\tLoss: 1.160296\n",
      "Train Epoch: 4 [39680/50048 (79%)]\tLoss: 0.935522\n",
      "Train Epoch: 4 [40320/50048 (81%)]\tLoss: 0.982619\n",
      "Train Epoch: 4 [40960/50048 (82%)]\tLoss: 0.898586\n",
      "Train Epoch: 4 [41600/50048 (83%)]\tLoss: 0.922304\n",
      "Train Epoch: 4 [42240/50048 (84%)]\tLoss: 0.945023\n",
      "Train Epoch: 4 [42880/50048 (86%)]\tLoss: 0.871002\n",
      "Train Epoch: 4 [43520/50048 (87%)]\tLoss: 1.019295\n",
      "Train Epoch: 4 [44160/50048 (88%)]\tLoss: 1.029593\n",
      "Train Epoch: 4 [44800/50048 (90%)]\tLoss: 1.020393\n",
      "Train Epoch: 4 [45440/50048 (91%)]\tLoss: 0.935931\n",
      "Train Epoch: 4 [46080/50048 (92%)]\tLoss: 0.864334\n",
      "Train Epoch: 4 [46720/50048 (93%)]\tLoss: 0.977505\n",
      "Train Epoch: 4 [47360/50048 (95%)]\tLoss: 1.009822\n",
      "Train Epoch: 4 [48000/50048 (96%)]\tLoss: 1.007580\n",
      "Train Epoch: 4 [48640/50048 (97%)]\tLoss: 1.020479\n",
      "Train Epoch: 4 [49280/50048 (98%)]\tLoss: 0.862538\n",
      "Train Epoch: 4 [49920/50048 (100%)]\tLoss: 0.897779\n",
      "\n",
      "Test set: Average loss: 4.1115, Accuracy: 3084/10000 (31%)\n",
      "\n",
      "Train Epoch: 5 [0/50048 (0%)]\tLoss: 7.250389\n",
      "Train Epoch: 5 [640/50048 (1%)]\tLoss: 1.245219\n",
      "Train Epoch: 5 [1280/50048 (3%)]\tLoss: 1.272605\n",
      "Train Epoch: 5 [1920/50048 (4%)]\tLoss: 1.204453\n",
      "Train Epoch: 5 [2560/50048 (5%)]\tLoss: 1.300871\n",
      "Train Epoch: 5 [3200/50048 (6%)]\tLoss: 1.115127\n",
      "Train Epoch: 5 [3840/50048 (8%)]\tLoss: 1.187306\n",
      "Train Epoch: 5 [4480/50048 (9%)]\tLoss: 1.404221\n",
      "Train Epoch: 5 [5120/50048 (10%)]\tLoss: 1.265236\n",
      "Train Epoch: 5 [5760/50048 (12%)]\tLoss: 1.185681\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Epoch: 5 [6400/50048 (13%)]\tLoss: 1.277691\n",
      "Train Epoch: 5 [7040/50048 (14%)]\tLoss: 1.119711\n",
      "Train Epoch: 5 [7680/50048 (15%)]\tLoss: 1.317215\n",
      "Train Epoch: 5 [8320/50048 (17%)]\tLoss: 1.253503\n",
      "Train Epoch: 5 [8960/50048 (18%)]\tLoss: 1.501366\n",
      "Train Epoch: 5 [9600/50048 (19%)]\tLoss: 1.174926\n",
      "Train Epoch: 5 [10240/50048 (20%)]\tLoss: 1.251156\n",
      "Train Epoch: 5 [10880/50048 (22%)]\tLoss: 1.292401\n",
      "Train Epoch: 5 [11520/50048 (23%)]\tLoss: 1.294362\n",
      "Train Epoch: 5 [12160/50048 (24%)]\tLoss: 1.139868\n",
      "Train Epoch: 5 [12800/50048 (26%)]\tLoss: 1.162836\n",
      "Train Epoch: 5 [13440/50048 (27%)]\tLoss: 1.218704\n",
      "Train Epoch: 5 [14080/50048 (28%)]\tLoss: 1.022158\n",
      "Train Epoch: 5 [14720/50048 (29%)]\tLoss: 1.164927\n",
      "Train Epoch: 5 [15360/50048 (31%)]\tLoss: 1.236923\n",
      "Train Epoch: 5 [16000/50048 (32%)]\tLoss: 1.005167\n",
      "Train Epoch: 5 [16640/50048 (33%)]\tLoss: 1.138367\n",
      "Train Epoch: 5 [17280/50048 (35%)]\tLoss: 1.187400\n",
      "Train Epoch: 5 [17920/50048 (36%)]\tLoss: 1.226922\n",
      "Train Epoch: 5 [18560/50048 (37%)]\tLoss: 1.288286\n",
      "Train Epoch: 5 [19200/50048 (38%)]\tLoss: 1.078293\n",
      "Train Epoch: 5 [19840/50048 (40%)]\tLoss: 0.964325\n",
      "Train Epoch: 5 [20480/50048 (41%)]\tLoss: 1.265334\n",
      "Train Epoch: 5 [21120/50048 (42%)]\tLoss: 1.183573\n",
      "Train Epoch: 5 [21760/50048 (43%)]\tLoss: 1.222590\n",
      "Train Epoch: 5 [22400/50048 (45%)]\tLoss: 1.089137\n",
      "Train Epoch: 5 [23040/50048 (46%)]\tLoss: 1.065711\n",
      "Train Epoch: 5 [23680/50048 (47%)]\tLoss: 0.907890\n",
      "Train Epoch: 5 [24320/50048 (49%)]\tLoss: 1.003827\n",
      "Train Epoch: 5 [24960/50048 (50%)]\tLoss: 0.995070\n",
      "Train Epoch: 5 [25600/50048 (51%)]\tLoss: 1.045151\n",
      "Train Epoch: 5 [26240/50048 (52%)]\tLoss: 0.987600\n",
      "Train Epoch: 5 [26880/50048 (54%)]\tLoss: 1.139112\n",
      "Train Epoch: 5 [27520/50048 (55%)]\tLoss: 1.077723\n",
      "Train Epoch: 5 [28160/50048 (56%)]\tLoss: 1.048534\n",
      "Train Epoch: 5 [28800/50048 (58%)]\tLoss: 0.822858\n",
      "Train Epoch: 5 [29440/50048 (59%)]\tLoss: 0.994931\n",
      "Train Epoch: 5 [30080/50048 (60%)]\tLoss: 1.004534\n",
      "Train Epoch: 5 [30720/50048 (61%)]\tLoss: 0.968374\n",
      "Train Epoch: 5 [31360/50048 (63%)]\tLoss: 1.035579\n",
      "Train Epoch: 5 [32000/50048 (64%)]\tLoss: 1.110539\n",
      "Train Epoch: 5 [32640/50048 (65%)]\tLoss: 0.900825\n",
      "Train Epoch: 5 [33280/50048 (66%)]\tLoss: 1.050558\n",
      "Train Epoch: 5 [33920/50048 (68%)]\tLoss: 0.800423\n",
      "Train Epoch: 5 [34560/50048 (69%)]\tLoss: 0.827204\n",
      "Train Epoch: 5 [35200/50048 (70%)]\tLoss: 1.215323\n",
      "Train Epoch: 5 [35840/50048 (72%)]\tLoss: 0.945803\n",
      "Train Epoch: 5 [36480/50048 (73%)]\tLoss: 0.981583\n",
      "Train Epoch: 5 [37120/50048 (74%)]\tLoss: 0.845216\n",
      "Train Epoch: 5 [37760/50048 (75%)]\tLoss: 0.863435\n",
      "Train Epoch: 5 [38400/50048 (77%)]\tLoss: 0.964212\n",
      "Train Epoch: 5 [39040/50048 (78%)]\tLoss: 1.025246\n",
      "Train Epoch: 5 [39680/50048 (79%)]\tLoss: 0.762530\n",
      "Train Epoch: 5 [40320/50048 (81%)]\tLoss: 0.872988\n",
      "Train Epoch: 5 [40960/50048 (82%)]\tLoss: 0.963711\n",
      "Train Epoch: 5 [41600/50048 (83%)]\tLoss: 0.815751\n",
      "Train Epoch: 5 [42240/50048 (84%)]\tLoss: 0.994837\n",
      "Train Epoch: 5 [42880/50048 (86%)]\tLoss: 0.916181\n",
      "Train Epoch: 5 [43520/50048 (87%)]\tLoss: 0.973030\n",
      "Train Epoch: 5 [44160/50048 (88%)]\tLoss: 1.061229\n",
      "Train Epoch: 5 [44800/50048 (90%)]\tLoss: 1.007452\n",
      "Train Epoch: 5 [45440/50048 (91%)]\tLoss: 0.871213\n",
      "Train Epoch: 5 [46080/50048 (92%)]\tLoss: 0.906693\n",
      "Train Epoch: 5 [46720/50048 (93%)]\tLoss: 1.073641\n",
      "Train Epoch: 5 [47360/50048 (95%)]\tLoss: 0.769357\n",
      "Train Epoch: 5 [48000/50048 (96%)]\tLoss: 0.850959\n",
      "Train Epoch: 5 [48640/50048 (97%)]\tLoss: 0.875434\n",
      "Train Epoch: 5 [49280/50048 (98%)]\tLoss: 0.855935\n",
      "Train Epoch: 5 [49920/50048 (100%)]\tLoss: 0.968043\n",
      "\n",
      "Test set: Average loss: 4.0143, Accuracy: 3158/10000 (32%)\n",
      "\n",
      "Train Epoch: 6 [0/50048 (0%)]\tLoss: 6.689412\n",
      "Train Epoch: 6 [640/50048 (1%)]\tLoss: 1.189732\n",
      "Train Epoch: 6 [1280/50048 (3%)]\tLoss: 1.271396\n",
      "Train Epoch: 6 [1920/50048 (4%)]\tLoss: 1.038033\n",
      "Train Epoch: 6 [2560/50048 (5%)]\tLoss: 1.036162\n",
      "Train Epoch: 6 [3200/50048 (6%)]\tLoss: 0.896135\n",
      "Train Epoch: 6 [3840/50048 (8%)]\tLoss: 1.328245\n",
      "Train Epoch: 6 [4480/50048 (9%)]\tLoss: 0.945587\n",
      "Train Epoch: 6 [5120/50048 (10%)]\tLoss: 1.183580\n",
      "Train Epoch: 6 [5760/50048 (12%)]\tLoss: 1.100551\n",
      "Train Epoch: 6 [6400/50048 (13%)]\tLoss: 1.210645\n",
      "Train Epoch: 6 [7040/50048 (14%)]\tLoss: 1.022842\n",
      "Train Epoch: 6 [7680/50048 (15%)]\tLoss: 1.201876\n",
      "Train Epoch: 6 [8320/50048 (17%)]\tLoss: 1.046741\n",
      "Train Epoch: 6 [8960/50048 (18%)]\tLoss: 1.083195\n",
      "Train Epoch: 6 [9600/50048 (19%)]\tLoss: 1.227070\n",
      "Train Epoch: 6 [10240/50048 (20%)]\tLoss: 0.985303\n",
      "Train Epoch: 6 [10880/50048 (22%)]\tLoss: 1.225209\n",
      "Train Epoch: 6 [11520/50048 (23%)]\tLoss: 1.104527\n",
      "Train Epoch: 6 [12160/50048 (24%)]\tLoss: 0.932569\n",
      "Train Epoch: 6 [12800/50048 (26%)]\tLoss: 0.954081\n",
      "Train Epoch: 6 [13440/50048 (27%)]\tLoss: 1.129096\n",
      "Train Epoch: 6 [14080/50048 (28%)]\tLoss: 1.128204\n",
      "Train Epoch: 6 [14720/50048 (29%)]\tLoss: 1.195575\n",
      "Train Epoch: 6 [15360/50048 (31%)]\tLoss: 1.181290\n",
      "Train Epoch: 6 [16000/50048 (32%)]\tLoss: 1.022578\n",
      "Train Epoch: 6 [16640/50048 (33%)]\tLoss: 1.067373\n",
      "Train Epoch: 6 [17280/50048 (35%)]\tLoss: 1.269623\n",
      "Train Epoch: 6 [17920/50048 (36%)]\tLoss: 1.028492\n",
      "Train Epoch: 6 [18560/50048 (37%)]\tLoss: 1.060621\n",
      "Train Epoch: 6 [19200/50048 (38%)]\tLoss: 1.068495\n",
      "Train Epoch: 6 [19840/50048 (40%)]\tLoss: 1.095670\n",
      "Train Epoch: 6 [20480/50048 (41%)]\tLoss: 1.165449\n",
      "Train Epoch: 6 [21120/50048 (42%)]\tLoss: 1.105961\n",
      "Train Epoch: 6 [21760/50048 (43%)]\tLoss: 1.035641\n",
      "Train Epoch: 6 [22400/50048 (45%)]\tLoss: 1.217671\n",
      "Train Epoch: 6 [23040/50048 (46%)]\tLoss: 1.166234\n",
      "Train Epoch: 6 [23680/50048 (47%)]\tLoss: 1.116490\n",
      "Train Epoch: 6 [24320/50048 (49%)]\tLoss: 0.851672\n",
      "Train Epoch: 6 [24960/50048 (50%)]\tLoss: 1.179117\n",
      "Train Epoch: 6 [25600/50048 (51%)]\tLoss: 0.898060\n",
      "Train Epoch: 6 [26240/50048 (52%)]\tLoss: 1.046812\n",
      "Train Epoch: 6 [26880/50048 (54%)]\tLoss: 1.043768\n",
      "Train Epoch: 6 [27520/50048 (55%)]\tLoss: 0.813748\n",
      "Train Epoch: 6 [28160/50048 (56%)]\tLoss: 1.043913\n",
      "Train Epoch: 6 [28800/50048 (58%)]\tLoss: 1.128653\n",
      "Train Epoch: 6 [29440/50048 (59%)]\tLoss: 0.819731\n",
      "Train Epoch: 6 [30080/50048 (60%)]\tLoss: 1.006939\n",
      "Train Epoch: 6 [30720/50048 (61%)]\tLoss: 0.853436\n",
      "Train Epoch: 6 [31360/50048 (63%)]\tLoss: 0.740101\n",
      "Train Epoch: 6 [32000/50048 (64%)]\tLoss: 1.029333\n",
      "Train Epoch: 6 [32640/50048 (65%)]\tLoss: 0.809053\n",
      "Train Epoch: 6 [33280/50048 (66%)]\tLoss: 0.983381\n",
      "Train Epoch: 6 [33920/50048 (68%)]\tLoss: 0.698948\n",
      "Train Epoch: 6 [34560/50048 (69%)]\tLoss: 0.776045\n",
      "Train Epoch: 6 [35200/50048 (70%)]\tLoss: 0.902267\n",
      "Train Epoch: 6 [35840/50048 (72%)]\tLoss: 0.885379\n",
      "Train Epoch: 6 [36480/50048 (73%)]\tLoss: 0.912060\n",
      "Train Epoch: 6 [37120/50048 (74%)]\tLoss: 0.757270\n",
      "Train Epoch: 6 [37760/50048 (75%)]\tLoss: 0.983837\n",
      "Train Epoch: 6 [38400/50048 (77%)]\tLoss: 0.906937\n",
      "Train Epoch: 6 [39040/50048 (78%)]\tLoss: 0.750300\n",
      "Train Epoch: 6 [39680/50048 (79%)]\tLoss: 0.930464\n",
      "Train Epoch: 6 [40320/50048 (81%)]\tLoss: 0.608516\n",
      "Train Epoch: 6 [40960/50048 (82%)]\tLoss: 0.853243\n",
      "Train Epoch: 6 [41600/50048 (83%)]\tLoss: 0.782429\n",
      "Train Epoch: 6 [42240/50048 (84%)]\tLoss: 0.905208\n",
      "Train Epoch: 6 [42880/50048 (86%)]\tLoss: 0.902006\n",
      "Train Epoch: 6 [43520/50048 (87%)]\tLoss: 0.683391\n",
      "Train Epoch: 6 [44160/50048 (88%)]\tLoss: 0.854627\n",
      "Train Epoch: 6 [44800/50048 (90%)]\tLoss: 0.693592\n",
      "Train Epoch: 6 [45440/50048 (91%)]\tLoss: 1.158527\n",
      "Train Epoch: 6 [46080/50048 (92%)]\tLoss: 0.778129\n",
      "Train Epoch: 6 [46720/50048 (93%)]\tLoss: 0.909242\n",
      "Train Epoch: 6 [47360/50048 (95%)]\tLoss: 0.881523\n",
      "Train Epoch: 6 [48000/50048 (96%)]\tLoss: 0.949505\n",
      "Train Epoch: 6 [48640/50048 (97%)]\tLoss: 0.811165\n",
      "Train Epoch: 6 [49280/50048 (98%)]\tLoss: 0.861244\n",
      "Train Epoch: 6 [49920/50048 (100%)]\tLoss: 0.904871\n",
      "\n",
      "Test set: Average loss: 3.9007, Accuracy: 3313/10000 (33%)\n",
      "\n",
      "Train Epoch: 7 [0/50048 (0%)]\tLoss: 7.003454\n",
      "Train Epoch: 7 [640/50048 (1%)]\tLoss: 1.153734\n",
      "Train Epoch: 7 [1280/50048 (3%)]\tLoss: 1.072940\n",
      "Train Epoch: 7 [1920/50048 (4%)]\tLoss: 1.100422\n",
      "Train Epoch: 7 [2560/50048 (5%)]\tLoss: 1.216830\n",
      "Train Epoch: 7 [3200/50048 (6%)]\tLoss: 1.004197\n",
      "Train Epoch: 7 [3840/50048 (8%)]\tLoss: 1.017943\n",
      "Train Epoch: 7 [4480/50048 (9%)]\tLoss: 0.966454\n",
      "Train Epoch: 7 [5120/50048 (10%)]\tLoss: 1.202149\n",
      "Train Epoch: 7 [5760/50048 (12%)]\tLoss: 1.013268\n",
      "Train Epoch: 7 [6400/50048 (13%)]\tLoss: 1.011777\n",
      "Train Epoch: 7 [7040/50048 (14%)]\tLoss: 0.835577\n",
      "Train Epoch: 7 [7680/50048 (15%)]\tLoss: 1.097483\n",
      "Train Epoch: 7 [8320/50048 (17%)]\tLoss: 0.992198\n",
      "Train Epoch: 7 [8960/50048 (18%)]\tLoss: 0.928440\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Epoch: 7 [9600/50048 (19%)]\tLoss: 1.164559\n",
      "Train Epoch: 7 [10240/50048 (20%)]\tLoss: 0.923925\n",
      "Train Epoch: 7 [10880/50048 (22%)]\tLoss: 0.882428\n",
      "Train Epoch: 7 [11520/50048 (23%)]\tLoss: 1.062252\n",
      "Train Epoch: 7 [12160/50048 (24%)]\tLoss: 0.724704\n",
      "Train Epoch: 7 [12800/50048 (26%)]\tLoss: 0.962698\n",
      "Train Epoch: 7 [13440/50048 (27%)]\tLoss: 1.036913\n",
      "Train Epoch: 7 [14080/50048 (28%)]\tLoss: 1.017223\n",
      "Train Epoch: 7 [14720/50048 (29%)]\tLoss: 0.883209\n",
      "Train Epoch: 7 [15360/50048 (31%)]\tLoss: 1.070466\n",
      "Train Epoch: 7 [16000/50048 (32%)]\tLoss: 0.935717\n",
      "Train Epoch: 7 [16640/50048 (33%)]\tLoss: 1.224811\n",
      "Train Epoch: 7 [17280/50048 (35%)]\tLoss: 1.102363\n",
      "Train Epoch: 7 [17920/50048 (36%)]\tLoss: 1.117949\n",
      "Train Epoch: 7 [18560/50048 (37%)]\tLoss: 0.959182\n",
      "Train Epoch: 7 [19200/50048 (38%)]\tLoss: 1.107941\n",
      "Train Epoch: 7 [19840/50048 (40%)]\tLoss: 0.972160\n",
      "Train Epoch: 7 [20480/50048 (41%)]\tLoss: 0.997495\n",
      "Train Epoch: 7 [21120/50048 (42%)]\tLoss: 1.077318\n",
      "Train Epoch: 7 [21760/50048 (43%)]\tLoss: 0.996902\n",
      "Train Epoch: 7 [22400/50048 (45%)]\tLoss: 1.047771\n",
      "Train Epoch: 7 [23040/50048 (46%)]\tLoss: 1.107562\n",
      "Train Epoch: 7 [23680/50048 (47%)]\tLoss: 0.917270\n",
      "Train Epoch: 7 [24320/50048 (49%)]\tLoss: 0.947407\n",
      "Train Epoch: 7 [24960/50048 (50%)]\tLoss: 0.881521\n",
      "Train Epoch: 7 [25600/50048 (51%)]\tLoss: 0.980076\n",
      "Train Epoch: 7 [26240/50048 (52%)]\tLoss: 0.790501\n",
      "Train Epoch: 7 [26880/50048 (54%)]\tLoss: 0.992699\n",
      "Train Epoch: 7 [27520/50048 (55%)]\tLoss: 0.803924\n",
      "Train Epoch: 7 [28160/50048 (56%)]\tLoss: 0.921170\n",
      "Train Epoch: 7 [28800/50048 (58%)]\tLoss: 0.807585\n",
      "Train Epoch: 7 [29440/50048 (59%)]\tLoss: 0.985587\n",
      "Train Epoch: 7 [30080/50048 (60%)]\tLoss: 0.981542\n",
      "Train Epoch: 7 [30720/50048 (61%)]\tLoss: 0.832521\n",
      "Train Epoch: 7 [31360/50048 (63%)]\tLoss: 0.646493\n",
      "Train Epoch: 7 [32000/50048 (64%)]\tLoss: 0.694892\n",
      "Train Epoch: 7 [32640/50048 (65%)]\tLoss: 0.832676\n",
      "Train Epoch: 7 [33280/50048 (66%)]\tLoss: 0.722012\n",
      "Train Epoch: 7 [33920/50048 (68%)]\tLoss: 0.893614\n",
      "Train Epoch: 7 [34560/50048 (69%)]\tLoss: 0.870295\n",
      "Train Epoch: 7 [35200/50048 (70%)]\tLoss: 0.794396\n",
      "Train Epoch: 7 [35840/50048 (72%)]\tLoss: 0.912132\n",
      "Train Epoch: 7 [36480/50048 (73%)]\tLoss: 0.961560\n",
      "Train Epoch: 7 [37120/50048 (74%)]\tLoss: 1.031184\n",
      "Train Epoch: 7 [37760/50048 (75%)]\tLoss: 0.726516\n",
      "Train Epoch: 7 [38400/50048 (77%)]\tLoss: 0.743402\n",
      "Train Epoch: 7 [39040/50048 (78%)]\tLoss: 0.793790\n",
      "Train Epoch: 7 [39680/50048 (79%)]\tLoss: 1.027673\n",
      "Train Epoch: 7 [40320/50048 (81%)]\tLoss: 0.749850\n",
      "Train Epoch: 7 [40960/50048 (82%)]\tLoss: 0.693859\n",
      "Train Epoch: 7 [41600/50048 (83%)]\tLoss: 0.935352\n",
      "Train Epoch: 7 [42240/50048 (84%)]\tLoss: 0.883940\n",
      "Train Epoch: 7 [42880/50048 (86%)]\tLoss: 0.866116\n",
      "Train Epoch: 7 [43520/50048 (87%)]\tLoss: 0.808718\n",
      "Train Epoch: 7 [44160/50048 (88%)]\tLoss: 0.660728\n",
      "Train Epoch: 7 [44800/50048 (90%)]\tLoss: 0.913370\n",
      "Train Epoch: 7 [45440/50048 (91%)]\tLoss: 0.715326\n",
      "Train Epoch: 7 [46080/50048 (92%)]\tLoss: 0.834272\n",
      "Train Epoch: 7 [46720/50048 (93%)]\tLoss: 0.720671\n",
      "Train Epoch: 7 [47360/50048 (95%)]\tLoss: 0.672570\n",
      "Train Epoch: 7 [48000/50048 (96%)]\tLoss: 0.946262\n",
      "Train Epoch: 7 [48640/50048 (97%)]\tLoss: 0.765957\n",
      "Train Epoch: 7 [49280/50048 (98%)]\tLoss: 0.818125\n",
      "Train Epoch: 7 [49920/50048 (100%)]\tLoss: 0.862601\n",
      "\n",
      "Test set: Average loss: 3.9971, Accuracy: 3376/10000 (34%)\n",
      "\n",
      "Train Epoch: 8 [0/50048 (0%)]\tLoss: 7.078762\n",
      "Train Epoch: 8 [640/50048 (1%)]\tLoss: 1.179980\n",
      "Train Epoch: 8 [1280/50048 (3%)]\tLoss: 1.051056\n",
      "Train Epoch: 8 [1920/50048 (4%)]\tLoss: 1.064773\n",
      "Train Epoch: 8 [2560/50048 (5%)]\tLoss: 0.944019\n",
      "Train Epoch: 8 [3200/50048 (6%)]\tLoss: 1.026255\n",
      "Train Epoch: 8 [3840/50048 (8%)]\tLoss: 1.001453\n",
      "Train Epoch: 8 [4480/50048 (9%)]\tLoss: 0.848291\n",
      "Train Epoch: 8 [5120/50048 (10%)]\tLoss: 1.154771\n",
      "Train Epoch: 8 [5760/50048 (12%)]\tLoss: 1.004940\n",
      "Train Epoch: 8 [6400/50048 (13%)]\tLoss: 0.942660\n",
      "Train Epoch: 8 [7040/50048 (14%)]\tLoss: 0.912112\n",
      "Train Epoch: 8 [7680/50048 (15%)]\tLoss: 1.047237\n",
      "Train Epoch: 8 [8320/50048 (17%)]\tLoss: 1.037336\n",
      "Train Epoch: 8 [8960/50048 (18%)]\tLoss: 1.035128\n",
      "Train Epoch: 8 [9600/50048 (19%)]\tLoss: 0.947861\n",
      "Train Epoch: 8 [10240/50048 (20%)]\tLoss: 1.109934\n",
      "Train Epoch: 8 [10880/50048 (22%)]\tLoss: 1.111900\n",
      "Train Epoch: 8 [11520/50048 (23%)]\tLoss: 1.031485\n",
      "Train Epoch: 8 [12160/50048 (24%)]\tLoss: 1.094681\n",
      "Train Epoch: 8 [12800/50048 (26%)]\tLoss: 1.121153\n",
      "Train Epoch: 8 [13440/50048 (27%)]\tLoss: 1.268618\n",
      "Train Epoch: 8 [14080/50048 (28%)]\tLoss: 1.133764\n",
      "Train Epoch: 8 [14720/50048 (29%)]\tLoss: 1.077894\n",
      "Train Epoch: 8 [15360/50048 (31%)]\tLoss: 0.931048\n",
      "Train Epoch: 8 [16000/50048 (32%)]\tLoss: 0.980039\n",
      "Train Epoch: 8 [16640/50048 (33%)]\tLoss: 1.036700\n",
      "Train Epoch: 8 [17280/50048 (35%)]\tLoss: 1.014424\n",
      "Train Epoch: 8 [17920/50048 (36%)]\tLoss: 1.122661\n",
      "Train Epoch: 8 [18560/50048 (37%)]\tLoss: 0.904136\n",
      "Train Epoch: 8 [19200/50048 (38%)]\tLoss: 0.809993\n",
      "Train Epoch: 8 [19840/50048 (40%)]\tLoss: 0.964570\n",
      "Train Epoch: 8 [20480/50048 (41%)]\tLoss: 0.926501\n",
      "Train Epoch: 8 [21120/50048 (42%)]\tLoss: 1.054118\n",
      "Train Epoch: 8 [21760/50048 (43%)]\tLoss: 1.119211\n",
      "Train Epoch: 8 [22400/50048 (45%)]\tLoss: 1.036441\n",
      "Train Epoch: 8 [23040/50048 (46%)]\tLoss: 1.001253\n",
      "Train Epoch: 8 [23680/50048 (47%)]\tLoss: 0.864020\n",
      "Train Epoch: 8 [24320/50048 (49%)]\tLoss: 1.017286\n",
      "Train Epoch: 8 [24960/50048 (50%)]\tLoss: 0.886495\n",
      "Train Epoch: 8 [25600/50048 (51%)]\tLoss: 0.988105\n",
      "Train Epoch: 8 [26240/50048 (52%)]\tLoss: 0.844942\n",
      "Train Epoch: 8 [26880/50048 (54%)]\tLoss: 0.800151\n",
      "Train Epoch: 8 [27520/50048 (55%)]\tLoss: 0.960836\n",
      "Train Epoch: 8 [28160/50048 (56%)]\tLoss: 0.885608\n",
      "Train Epoch: 8 [28800/50048 (58%)]\tLoss: 0.830720\n",
      "Train Epoch: 8 [29440/50048 (59%)]\tLoss: 0.627387\n",
      "Train Epoch: 8 [30080/50048 (60%)]\tLoss: 0.877364\n",
      "Train Epoch: 8 [30720/50048 (61%)]\tLoss: 0.823594\n",
      "Train Epoch: 8 [31360/50048 (63%)]\tLoss: 0.732282\n",
      "Train Epoch: 8 [32000/50048 (64%)]\tLoss: 0.793128\n",
      "Train Epoch: 8 [32640/50048 (65%)]\tLoss: 0.715032\n",
      "Train Epoch: 8 [33280/50048 (66%)]\tLoss: 0.924907\n",
      "Train Epoch: 8 [33920/50048 (68%)]\tLoss: 0.833574\n",
      "Train Epoch: 8 [34560/50048 (69%)]\tLoss: 0.682612\n",
      "Train Epoch: 8 [35200/50048 (70%)]\tLoss: 0.831133\n",
      "Train Epoch: 8 [35840/50048 (72%)]\tLoss: 0.892532\n",
      "Train Epoch: 8 [36480/50048 (73%)]\tLoss: 0.812054\n",
      "Train Epoch: 8 [37120/50048 (74%)]\tLoss: 0.535791\n",
      "Train Epoch: 8 [37760/50048 (75%)]\tLoss: 0.821478\n",
      "Train Epoch: 8 [38400/50048 (77%)]\tLoss: 0.792334\n",
      "Train Epoch: 8 [39040/50048 (78%)]\tLoss: 0.962863\n",
      "Train Epoch: 8 [39680/50048 (79%)]\tLoss: 0.871479\n",
      "Train Epoch: 8 [40320/50048 (81%)]\tLoss: 0.792229\n",
      "Train Epoch: 8 [40960/50048 (82%)]\tLoss: 0.744101\n",
      "Train Epoch: 8 [41600/50048 (83%)]\tLoss: 0.751952\n",
      "Train Epoch: 8 [42240/50048 (84%)]\tLoss: 0.760790\n",
      "Train Epoch: 8 [42880/50048 (86%)]\tLoss: 0.888301\n",
      "Train Epoch: 8 [43520/50048 (87%)]\tLoss: 0.847174\n",
      "Train Epoch: 8 [44160/50048 (88%)]\tLoss: 0.770799\n",
      "Train Epoch: 8 [44800/50048 (90%)]\tLoss: 0.928981\n",
      "Train Epoch: 8 [45440/50048 (91%)]\tLoss: 0.759207\n",
      "Train Epoch: 8 [46080/50048 (92%)]\tLoss: 0.833093\n",
      "Train Epoch: 8 [46720/50048 (93%)]\tLoss: 0.945813\n",
      "Train Epoch: 8 [47360/50048 (95%)]\tLoss: 1.037133\n",
      "Train Epoch: 8 [48000/50048 (96%)]\tLoss: 0.943005\n",
      "Train Epoch: 8 [48640/50048 (97%)]\tLoss: 1.014674\n",
      "Train Epoch: 8 [49280/50048 (98%)]\tLoss: 0.901574\n",
      "Train Epoch: 8 [49920/50048 (100%)]\tLoss: 0.815766\n",
      "\n",
      "Test set: Average loss: 4.0229, Accuracy: 3409/10000 (34%)\n",
      "\n",
      "Train Epoch: 9 [0/50048 (0%)]\tLoss: 6.966828\n",
      "Train Epoch: 9 [640/50048 (1%)]\tLoss: 0.898334\n",
      "Train Epoch: 9 [1280/50048 (3%)]\tLoss: 0.918711\n",
      "Train Epoch: 9 [1920/50048 (4%)]\tLoss: 0.803167\n",
      "Train Epoch: 9 [2560/50048 (5%)]\tLoss: 1.142138\n",
      "Train Epoch: 9 [3200/50048 (6%)]\tLoss: 1.020735\n",
      "Train Epoch: 9 [3840/50048 (8%)]\tLoss: 0.983048\n",
      "Train Epoch: 9 [4480/50048 (9%)]\tLoss: 1.171086\n",
      "Train Epoch: 9 [5120/50048 (10%)]\tLoss: 1.055621\n",
      "Train Epoch: 9 [5760/50048 (12%)]\tLoss: 0.921691\n",
      "Train Epoch: 9 [6400/50048 (13%)]\tLoss: 0.893748\n",
      "Train Epoch: 9 [7040/50048 (14%)]\tLoss: 1.029617\n",
      "Train Epoch: 9 [7680/50048 (15%)]\tLoss: 0.942249\n",
      "Train Epoch: 9 [8320/50048 (17%)]\tLoss: 1.045128\n",
      "Train Epoch: 9 [8960/50048 (18%)]\tLoss: 1.013988\n",
      "Train Epoch: 9 [9600/50048 (19%)]\tLoss: 1.223285\n",
      "Train Epoch: 9 [10240/50048 (20%)]\tLoss: 0.866921\n",
      "Train Epoch: 9 [10880/50048 (22%)]\tLoss: 0.927101\n",
      "Train Epoch: 9 [11520/50048 (23%)]\tLoss: 1.036976\n",
      "Train Epoch: 9 [12160/50048 (24%)]\tLoss: 0.973240\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Epoch: 9 [12800/50048 (26%)]\tLoss: 1.055329\n",
      "Train Epoch: 9 [13440/50048 (27%)]\tLoss: 0.862610\n",
      "Train Epoch: 9 [14080/50048 (28%)]\tLoss: 0.956215\n",
      "Train Epoch: 9 [14720/50048 (29%)]\tLoss: 1.052734\n",
      "Train Epoch: 9 [15360/50048 (31%)]\tLoss: 0.901689\n",
      "Train Epoch: 9 [16000/50048 (32%)]\tLoss: 0.869646\n",
      "Train Epoch: 9 [16640/50048 (33%)]\tLoss: 1.046412\n",
      "Train Epoch: 9 [17280/50048 (35%)]\tLoss: 0.819834\n",
      "Train Epoch: 9 [17920/50048 (36%)]\tLoss: 1.090525\n",
      "Train Epoch: 9 [18560/50048 (37%)]\tLoss: 0.954119\n",
      "Train Epoch: 9 [19200/50048 (38%)]\tLoss: 0.968223\n",
      "Train Epoch: 9 [19840/50048 (40%)]\tLoss: 0.815727\n",
      "Train Epoch: 9 [20480/50048 (41%)]\tLoss: 1.004352\n",
      "Train Epoch: 9 [21120/50048 (42%)]\tLoss: 0.874320\n",
      "Train Epoch: 9 [21760/50048 (43%)]\tLoss: 0.884369\n",
      "Train Epoch: 9 [22400/50048 (45%)]\tLoss: 0.894453\n",
      "Train Epoch: 9 [23040/50048 (46%)]\tLoss: 0.971830\n",
      "Train Epoch: 9 [23680/50048 (47%)]\tLoss: 0.878662\n",
      "Train Epoch: 9 [24320/50048 (49%)]\tLoss: 0.978615\n",
      "Train Epoch: 9 [24960/50048 (50%)]\tLoss: 0.862984\n",
      "Train Epoch: 9 [25600/50048 (51%)]\tLoss: 0.840022\n",
      "Train Epoch: 9 [26240/50048 (52%)]\tLoss: 0.824416\n",
      "Train Epoch: 9 [26880/50048 (54%)]\tLoss: 0.905669\n",
      "Train Epoch: 9 [27520/50048 (55%)]\tLoss: 0.872424\n",
      "Train Epoch: 9 [28160/50048 (56%)]\tLoss: 0.898437\n",
      "Train Epoch: 9 [28800/50048 (58%)]\tLoss: 0.928985\n",
      "Train Epoch: 9 [29440/50048 (59%)]\tLoss: 0.620692\n",
      "Train Epoch: 9 [30080/50048 (60%)]\tLoss: 0.775219\n",
      "Train Epoch: 9 [30720/50048 (61%)]\tLoss: 0.828597\n",
      "Train Epoch: 9 [31360/50048 (63%)]\tLoss: 0.998602\n",
      "Train Epoch: 9 [32000/50048 (64%)]\tLoss: 0.604097\n",
      "Train Epoch: 9 [32640/50048 (65%)]\tLoss: 1.069160\n",
      "Train Epoch: 9 [33280/50048 (66%)]\tLoss: 0.701461\n",
      "Train Epoch: 9 [33920/50048 (68%)]\tLoss: 0.920835\n",
      "Train Epoch: 9 [34560/50048 (69%)]\tLoss: 1.140749\n",
      "Train Epoch: 9 [35200/50048 (70%)]\tLoss: 0.871730\n",
      "Train Epoch: 9 [35840/50048 (72%)]\tLoss: 0.794151\n",
      "Train Epoch: 9 [36480/50048 (73%)]\tLoss: 0.637525\n",
      "Train Epoch: 9 [37120/50048 (74%)]\tLoss: 0.755643\n",
      "Train Epoch: 9 [37760/50048 (75%)]\tLoss: 0.660773\n",
      "Train Epoch: 9 [38400/50048 (77%)]\tLoss: 1.018779\n",
      "Train Epoch: 9 [39040/50048 (78%)]\tLoss: 0.715554\n",
      "Train Epoch: 9 [39680/50048 (79%)]\tLoss: 0.702211\n",
      "Train Epoch: 9 [40320/50048 (81%)]\tLoss: 0.728644\n",
      "Train Epoch: 9 [40960/50048 (82%)]\tLoss: 0.857449\n",
      "Train Epoch: 9 [41600/50048 (83%)]\tLoss: 0.712136\n",
      "Train Epoch: 9 [42240/50048 (84%)]\tLoss: 0.587116\n",
      "Train Epoch: 9 [42880/50048 (86%)]\tLoss: 0.887169\n",
      "Train Epoch: 9 [43520/50048 (87%)]\tLoss: 0.954561\n",
      "Train Epoch: 9 [44160/50048 (88%)]\tLoss: 0.733142\n",
      "Train Epoch: 9 [44800/50048 (90%)]\tLoss: 0.769831\n",
      "Train Epoch: 9 [45440/50048 (91%)]\tLoss: 0.780675\n",
      "Train Epoch: 9 [46080/50048 (92%)]\tLoss: 0.678151\n",
      "Train Epoch: 9 [46720/50048 (93%)]\tLoss: 0.843365\n",
      "Train Epoch: 9 [47360/50048 (95%)]\tLoss: 0.835203\n",
      "Train Epoch: 9 [48000/50048 (96%)]\tLoss: 1.046026\n",
      "Train Epoch: 9 [48640/50048 (97%)]\tLoss: 0.670450\n",
      "Train Epoch: 9 [49280/50048 (98%)]\tLoss: 0.724072\n",
      "Train Epoch: 9 [49920/50048 (100%)]\tLoss: 0.851139\n",
      "\n",
      "Test set: Average loss: 4.1723, Accuracy: 3412/10000 (34%)\n",
      "\n",
      "Train Epoch: 10 [0/50048 (0%)]\tLoss: 7.376417\n",
      "Train Epoch: 10 [640/50048 (1%)]\tLoss: 0.920805\n",
      "Train Epoch: 10 [1280/50048 (3%)]\tLoss: 0.975556\n",
      "Train Epoch: 10 [1920/50048 (4%)]\tLoss: 1.048972\n",
      "Train Epoch: 10 [2560/50048 (5%)]\tLoss: 0.889359\n",
      "Train Epoch: 10 [3200/50048 (6%)]\tLoss: 1.047701\n",
      "Train Epoch: 10 [3840/50048 (8%)]\tLoss: 0.924349\n",
      "Train Epoch: 10 [4480/50048 (9%)]\tLoss: 0.973696\n",
      "Train Epoch: 10 [5120/50048 (10%)]\tLoss: 0.892823\n",
      "Train Epoch: 10 [5760/50048 (12%)]\tLoss: 1.006023\n",
      "Train Epoch: 10 [6400/50048 (13%)]\tLoss: 0.914831\n",
      "Train Epoch: 10 [7040/50048 (14%)]\tLoss: 0.924422\n",
      "Train Epoch: 10 [7680/50048 (15%)]\tLoss: 0.936027\n",
      "Train Epoch: 10 [8320/50048 (17%)]\tLoss: 0.995034\n",
      "Train Epoch: 10 [8960/50048 (18%)]\tLoss: 0.957949\n",
      "Train Epoch: 10 [9600/50048 (19%)]\tLoss: 0.960802\n",
      "Train Epoch: 10 [10240/50048 (20%)]\tLoss: 1.141215\n",
      "Train Epoch: 10 [10880/50048 (22%)]\tLoss: 0.938021\n",
      "Train Epoch: 10 [11520/50048 (23%)]\tLoss: 1.056182\n",
      "Train Epoch: 10 [12160/50048 (24%)]\tLoss: 0.935830\n",
      "Train Epoch: 10 [12800/50048 (26%)]\tLoss: 0.857422\n",
      "Train Epoch: 10 [13440/50048 (27%)]\tLoss: 0.907074\n",
      "Train Epoch: 10 [14080/50048 (28%)]\tLoss: 0.868992\n",
      "Train Epoch: 10 [14720/50048 (29%)]\tLoss: 0.921464\n",
      "Train Epoch: 10 [15360/50048 (31%)]\tLoss: 0.907111\n",
      "Train Epoch: 10 [16000/50048 (32%)]\tLoss: 0.929403\n",
      "Train Epoch: 10 [16640/50048 (33%)]\tLoss: 0.955201\n",
      "Train Epoch: 10 [17280/50048 (35%)]\tLoss: 0.945036\n",
      "Train Epoch: 10 [17920/50048 (36%)]\tLoss: 0.912729\n",
      "Train Epoch: 10 [18560/50048 (37%)]\tLoss: 0.958529\n",
      "Train Epoch: 10 [19200/50048 (38%)]\tLoss: 0.948019\n",
      "Train Epoch: 10 [19840/50048 (40%)]\tLoss: 0.852258\n",
      "Train Epoch: 10 [20480/50048 (41%)]\tLoss: 0.768854\n",
      "Train Epoch: 10 [21120/50048 (42%)]\tLoss: 0.855356\n",
      "Train Epoch: 10 [21760/50048 (43%)]\tLoss: 0.968566\n",
      "Train Epoch: 10 [22400/50048 (45%)]\tLoss: 0.894655\n",
      "Train Epoch: 10 [23040/50048 (46%)]\tLoss: 0.966674\n",
      "Train Epoch: 10 [23680/50048 (47%)]\tLoss: 0.953977\n",
      "Train Epoch: 10 [24320/50048 (49%)]\tLoss: 0.935820\n",
      "Train Epoch: 10 [24960/50048 (50%)]\tLoss: 0.979175\n",
      "Train Epoch: 10 [25600/50048 (51%)]\tLoss: 0.887442\n",
      "Train Epoch: 10 [26240/50048 (52%)]\tLoss: 0.765029\n",
      "Train Epoch: 10 [26880/50048 (54%)]\tLoss: 0.813957\n",
      "Train Epoch: 10 [27520/50048 (55%)]\tLoss: 0.781726\n",
      "Train Epoch: 10 [28160/50048 (56%)]\tLoss: 0.690525\n",
      "Train Epoch: 10 [28800/50048 (58%)]\tLoss: 0.643892\n",
      "Train Epoch: 10 [29440/50048 (59%)]\tLoss: 0.787196\n",
      "Train Epoch: 10 [30080/50048 (60%)]\tLoss: 0.830959\n",
      "Train Epoch: 10 [30720/50048 (61%)]\tLoss: 0.670646\n",
      "Train Epoch: 10 [31360/50048 (63%)]\tLoss: 0.791856\n",
      "Train Epoch: 10 [32000/50048 (64%)]\tLoss: 0.987663\n",
      "Train Epoch: 10 [32640/50048 (65%)]\tLoss: 0.708090\n",
      "Train Epoch: 10 [33280/50048 (66%)]\tLoss: 0.925546\n",
      "Train Epoch: 10 [33920/50048 (68%)]\tLoss: 0.650317\n",
      "Train Epoch: 10 [34560/50048 (69%)]\tLoss: 0.902596\n",
      "Train Epoch: 10 [35200/50048 (70%)]\tLoss: 0.789095\n",
      "Train Epoch: 10 [35840/50048 (72%)]\tLoss: 0.753626\n",
      "Train Epoch: 10 [36480/50048 (73%)]\tLoss: 0.675085\n",
      "Train Epoch: 10 [37120/50048 (74%)]\tLoss: 0.661397\n",
      "Train Epoch: 10 [37760/50048 (75%)]\tLoss: 0.833203\n",
      "Train Epoch: 10 [38400/50048 (77%)]\tLoss: 0.727932\n",
      "Train Epoch: 10 [39040/50048 (78%)]\tLoss: 0.607417\n",
      "Train Epoch: 10 [39680/50048 (79%)]\tLoss: 0.901553\n",
      "Train Epoch: 10 [40320/50048 (81%)]\tLoss: 0.560631\n",
      "Train Epoch: 10 [40960/50048 (82%)]\tLoss: 0.900231\n",
      "Train Epoch: 10 [41600/50048 (83%)]\tLoss: 0.828146\n",
      "Train Epoch: 10 [42240/50048 (84%)]\tLoss: 0.777608\n",
      "Train Epoch: 10 [42880/50048 (86%)]\tLoss: 0.946823\n",
      "Train Epoch: 10 [43520/50048 (87%)]\tLoss: 0.767128\n",
      "Train Epoch: 10 [44160/50048 (88%)]\tLoss: 0.705562\n",
      "Train Epoch: 10 [44800/50048 (90%)]\tLoss: 1.090590\n",
      "Train Epoch: 10 [45440/50048 (91%)]\tLoss: 0.614446\n",
      "Train Epoch: 10 [46080/50048 (92%)]\tLoss: 1.071108\n",
      "Train Epoch: 10 [46720/50048 (93%)]\tLoss: 0.644776\n",
      "Train Epoch: 10 [47360/50048 (95%)]\tLoss: 0.778956\n",
      "Train Epoch: 10 [48000/50048 (96%)]\tLoss: 0.777609\n",
      "Train Epoch: 10 [48640/50048 (97%)]\tLoss: 0.721396\n",
      "Train Epoch: 10 [49280/50048 (98%)]\tLoss: 0.832028\n",
      "Train Epoch: 10 [49920/50048 (100%)]\tLoss: 0.598625\n",
      "\n",
      "Test set: Average loss: 4.0293, Accuracy: 3408/10000 (34%)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = Net()\n",
    "optimizer = optim.SGD(model.parameters(), lr=args['lr'])\n",
    "\n",
    "for epoch in range(1, args['epochs'] + 1):\n",
    "        train(args, model, device, federated_train_loader, optimizer, epoch)\n",
    "        test(model, device, test_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "visible-sixth",
   "metadata": {},
   "outputs": [],
   "source": [
    "torch.save(model.state_dict(), 'CIFER10CNN.pth')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "numerical-reproduction",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "alleged-gateway",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}