GNR_638_Kaggle.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "GNR-638-Kaggle.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "wtth5dr-Rzwx"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "from torch.autograd import Variable\n",
        "\n",
        "import torchvision\n",
        "import torchvision.transforms as transforms\n",
        "from torch.utils.data import Dataset, DataLoader\n",
        "from sklearn.metrics import confusion_matrix\n",
        "%matplotlib inline\n",
        "%config InlineBackend.figure_format = 'retina'\n",
        "import matplotlib.pyplot as plt\n",
        "import torch\n",
        "from torchvision import datasets, transforms\n",
        "import helper"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8xh7fNVN5aKK",
        "outputId": "e8f717c9-a52b-40e2-98ac-1b69d57f268c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "dir = \"/content/drive/My Drive/GNR 638 Kaggle\""
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GTCXef-Y2rHI",
        "outputId": "76f780db-7b55-47a9-f72b-a3d5eb16f53a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "if torch.cuda.is_available():\n",
        "  device = torch.device(\"cuda\")\n",
        "  print(\"GPU activated\")\n",
        "else:\n",
        "  device = torch.device(\"cpu\")"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "GPU activated\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ELQIpGfj2u3n"
      },
      "source": [
        "#device = torch.device(\"cpu\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RmCPxbVJZKPB"
      },
      "source": [
        "import os\n",
        "os.getcwd()\n",
        "os.chdir(dir)"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8-5E6oOY-av-"
      },
      "source": [
        "os.environ['CUDA_LAUNCH_BLOCKING'] = \"1\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zT_Olp6mXybn"
      },
      "source": [
        "\n",
        "train_set = datasets.ImageFolder(dir+\"/data2\",transform = transforms.Compose([transforms.Resize(255),\n",
        "                                transforms.CenterCrop(224),\n",
        "                                #transforms.Grayscale(),\n",
        "                                transforms.ToTensor()]))\n",
        "\n",
        "train_set, val_set = torch.utils.data.random_split(train_set, [2541,539]) "
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cdWKQCj_WUGp"
      },
      "source": [
        "train_loader = DataLoader(train_set, batch_size=77, shuffle = True)\n",
        "test_loader = DataLoader(val_set, batch_size=77, shuffle = True)"
      ],
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jffTUGXeWXrz"
      },
      "source": [
        "next(iter(test_loader))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dHpio0WZl_h4"
      },
      "source": [
        "class LocationCNN(nn.Module):\n",
        "    \n",
        "    def __init__(self):\n",
        "        super(LocationCNN, self).__init__()\n",
        "        \n",
        "        \n",
        "        self.layer1 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=3, out_channels=48, kernel_size=11, stride = 4, padding = 2),\n",
        "            nn.BatchNorm2d(48),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "        self.layer2 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels=48, out_channels=128, kernel_size=5, padding = 2),\n",
        "            nn.MaxPool2d(2),\n",
        "            nn.BatchNorm2d(128),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "        self.layer3 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels = 128, out_channels = 192, kernel_size = 3, padding = 1),\n",
        "            nn.MaxPool2d(2),\n",
        "            nn.BatchNorm2d(192),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "        self.layer4 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels = 192, out_channels = 192, kernel_size  = 3, padding = 1),\n",
        "            nn.BatchNorm2d(192),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "        self.layer5 = nn.Sequential(\n",
        "            nn.Conv2d(in_channels = 192, out_channels = 128, kernel_size  = 1),\n",
        "            nn.MaxPool2d(2),\n",
        "            nn.BatchNorm2d(128),\n",
        "            nn.ReLU()\n",
        "        )\n",
        "        self.flatten = nn.Flatten()\n",
        "        self.layer6 = nn.Sequential(\n",
        "            nn.Linear(in_features=6*6*128, out_features=98),\n",
        "            nn.Dropout2d(0.25),\n",
        "            nn.Linear(in_features=98,out_features=7),\n",
        "            nn.Softmax()\n",
        "        )\n",
        "        \n",
        "    def forward(self, x):\n",
        "        out = self.layer1(x)\n",
        "        out = self.layer2(out)\n",
        "        out = self.layer3(out)\n",
        "        out = self.layer4(out)\n",
        "        out = self.layer5(out)\n",
        "        out = self.flatten(out)\n",
        "        out = self.layer6(out)\n",
        "        return out\n",
        "\n"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YF1rHNKTWiUj",
        "outputId": "8e19a779-eeba-438e-d8b2-6e84f7dfec58",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 674
        }
      },
      "source": [
        "model = LocationCNN()\n",
        "model.to(device)\n",
        "\n",
        "error = nn.CrossEntropyLoss()\n",
        "\n",
        "learning_rate = 0.001\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
        "print(model)"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "LocationCNN(\n",
            "  (layer1): Sequential(\n",
            "    (0): Conv2d(3, 48, kernel_size=(11, 11), stride=(4, 4), padding=(2, 2))\n",
            "    (1): BatchNorm2d(48, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "    (2): ReLU()\n",
            "  )\n",
            "  (layer2): Sequential(\n",
            "    (0): Conv2d(48, 128, kernel_size=(5, 5), stride=(1, 1), padding=(2, 2))\n",
            "    (1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "    (3): ReLU()\n",
            "  )\n",
            "  (layer3): Sequential(\n",
            "    (0): Conv2d(128, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (2): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "    (3): ReLU()\n",
            "  )\n",
            "  (layer4): Sequential(\n",
            "    (0): Conv2d(192, 192, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
            "    (1): BatchNorm2d(192, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "    (2): ReLU()\n",
            "  )\n",
            "  (layer5): Sequential(\n",
            "    (0): Conv2d(192, 128, kernel_size=(1, 1), stride=(1, 1))\n",
            "    (1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n",
            "    (3): ReLU()\n",
            "  )\n",
            "  (flatten): Flatten()\n",
            "  (layer6): Sequential(\n",
            "    (0): Linear(in_features=4608, out_features=98, bias=True)\n",
            "    (1): Dropout2d(p=0.25, inplace=False)\n",
            "    (2): Linear(in_features=98, out_features=7, bias=True)\n",
            "    (3): Softmax(dim=None)\n",
            "  )\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S2kD7ZqTWkW6",
        "outputId": "7051b29b-c3ab-4e43-eadd-ae2d8870ade7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "num_epochs = 20\n",
        "count = 0\n",
        "# Lists for visualization of loss and accuracy \n",
        "loss_list = []\n",
        "iteration_list = []\n",
        "accuracy_train_list = []\n",
        "accuracy_list = []\n",
        "test_loss_list = []\n",
        "# Lists for knowing classwise accuracy\n",
        "predictions_list = []\n",
        "predictions_train_list = []\n",
        "labels_list = []\n",
        "\n",
        "for epoch in range(num_epochs):\n",
        "    for images, labels in train_loader:\n",
        "        # Transfering images and labels to GPU if available\n",
        "        images, labels = images.to(device), labels.to(device)\n",
        "\n",
        "        train = Variable(images.view(77,3,224,224))\n",
        "        labels = Variable(labels)\n",
        "        # Forward pass \n",
        "        correct = 0\n",
        "        outputs = model(train)\n",
        "        predictions_train = torch.max(outputs, 1)[1].to(device)\n",
        "        predictions_train_list.append(predictions_train)\n",
        "        correct += (predictions_train == labels).sum()\n",
        "        total += len(labels)\n",
        "        \n",
        "        \n",
        "        loss = error(outputs,  labels)\n",
        "        # Initializing a gradient as 0 so there is no mixing of gradient among the batches\n",
        "        optimizer.zero_grad()\n",
        "        \n",
        "        #Propagating the error backward\n",
        "        loss.backward()\n",
        "        \n",
        "        # Optimizing the parameters\n",
        "        optimizer.step()\n",
        "        \n",
        "        count += 1\n",
        "        \n",
        "        accuracy_train = correct * 100 / float(total)\n",
        "        iteration_list.append(count)\n",
        "        accuracy_train_list.append(accuracy)\n",
        "    # Testing the model\n",
        "    \n",
        "        if count % 50000 != 0:\n",
        "            total = 0\n",
        "            correct = 0\n",
        "        \n",
        "            for images, labels in test_loader:\n",
        "                images, labels = images.to(device), labels.to(device)\n",
        "                labels_list.append(labels)\n",
        "            \n",
        "                test = Variable(images.view(77, 3, 224, 224))\n",
        "                \n",
        "                outputs = model(test)\n",
        "                predictions = torch.max(outputs, 1)[1].to(device)\n",
        "                predictions_list.append(predictions)\n",
        "                correct += (predictions == labels).sum()\n",
        "                total += len(labels)\n",
        "                labels = Variable(labels)\n",
        "                \n",
        "            accuracy = correct * 100 / float(total)\n",
        "            loss_list.append(loss.data)\n",
        "            accuracy_list.append(accuracy)\n",
        "        \n",
        "        if count % 50000!=0:\n",
        "            print(\"Iteration: {}, Training Loss: {}, Validation Accuracy: {}%\".format(count, loss.data, accuracy))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/container.py:117: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
            "  input = module(input)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Iteration: 1, Training Loss: 1.9537831544876099, Validation Accuracy: 16.32653045654297%\n",
            "Iteration: 2, Training Loss: 1.9252780675888062, Validation Accuracy: 33.95176315307617%\n",
            "Iteration: 3, Training Loss: 1.71213698387146, Validation Accuracy: 50.46382141113281%\n",
            "Iteration: 4, Training Loss: 1.634764313697815, Validation Accuracy: 55.10204315185547%\n",
            "Iteration: 5, Training Loss: 1.6290886402130127, Validation Accuracy: 59.18367385864258%\n",
            "Iteration: 6, Training Loss: 1.593295693397522, Validation Accuracy: 62.33766174316406%\n",
            "Iteration: 7, Training Loss: 1.4970428943634033, Validation Accuracy: 68.46011352539062%\n",
            "Iteration: 8, Training Loss: 1.4590469598770142, Validation Accuracy: 72.17068481445312%\n",
            "Iteration: 9, Training Loss: 1.5225374698638916, Validation Accuracy: 69.7588119506836%\n",
            "Iteration: 10, Training Loss: 1.5244532823562622, Validation Accuracy: 75.69573211669922%\n",
            "Iteration: 11, Training Loss: 1.4230575561523438, Validation Accuracy: 74.95362091064453%\n",
            "Iteration: 12, Training Loss: 1.366692304611206, Validation Accuracy: 74.39703369140625%\n",
            "Iteration: 13, Training Loss: 1.3964176177978516, Validation Accuracy: 73.84044647216797%\n",
            "Iteration: 14, Training Loss: 1.384026288986206, Validation Accuracy: 74.7680892944336%\n",
            "Iteration: 15, Training Loss: 1.4884766340255737, Validation Accuracy: 73.84044647216797%\n",
            "Iteration: 16, Training Loss: 1.4285608530044556, Validation Accuracy: 74.21150207519531%\n",
            "Iteration: 17, Training Loss: 1.3833141326904297, Validation Accuracy: 75.32467651367188%\n",
            "Iteration: 18, Training Loss: 1.4164708852767944, Validation Accuracy: 75.13914489746094%\n",
            "Iteration: 19, Training Loss: 1.3614957332611084, Validation Accuracy: 75.32467651367188%\n",
            "Iteration: 20, Training Loss: 1.3862496614456177, Validation Accuracy: 77.36549377441406%\n",
            "Iteration: 21, Training Loss: 1.469557762145996, Validation Accuracy: 77.55101776123047%\n",
            "Iteration: 22, Training Loss: 1.3573486804962158, Validation Accuracy: 77.92208099365234%\n",
            "Iteration: 23, Training Loss: 1.3570255041122437, Validation Accuracy: 78.29313659667969%\n",
            "Iteration: 24, Training Loss: 1.3309063911437988, Validation Accuracy: 79.7773666381836%\n",
            "Iteration: 25, Training Loss: 1.4047871828079224, Validation Accuracy: 81.44712829589844%\n",
            "Iteration: 26, Training Loss: 1.3197723627090454, Validation Accuracy: 79.22077941894531%\n",
            "Iteration: 27, Training Loss: 1.3550870418548584, Validation Accuracy: 76.80890655517578%\n",
            "Iteration: 28, Training Loss: 1.4156686067581177, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 29, Training Loss: 1.3709452152252197, Validation Accuracy: 80.33395385742188%\n",
            "Iteration: 30, Training Loss: 1.386741042137146, Validation Accuracy: 77.7365493774414%\n",
            "Iteration: 31, Training Loss: 1.4076849222183228, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 32, Training Loss: 1.3434919118881226, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 33, Training Loss: 1.350374460220337, Validation Accuracy: 78.66419219970703%\n",
            "Iteration: 34, Training Loss: 1.3559635877609253, Validation Accuracy: 79.40631103515625%\n",
            "Iteration: 35, Training Loss: 1.371533989906311, Validation Accuracy: 80.51947784423828%\n",
            "Iteration: 36, Training Loss: 1.3413883447647095, Validation Accuracy: 79.7773666381836%\n",
            "Iteration: 37, Training Loss: 1.3492432832717896, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 38, Training Loss: 1.4140186309814453, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 39, Training Loss: 1.4008100032806396, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 40, Training Loss: 1.3368258476257324, Validation Accuracy: 80.51947784423828%\n",
            "Iteration: 41, Training Loss: 1.3711766004562378, Validation Accuracy: 79.03524780273438%\n",
            "Iteration: 42, Training Loss: 1.3142063617706299, Validation Accuracy: 79.40631103515625%\n",
            "Iteration: 43, Training Loss: 1.4003819227218628, Validation Accuracy: 79.7773666381836%\n",
            "Iteration: 44, Training Loss: 1.355129599571228, Validation Accuracy: 81.07606506347656%\n",
            "Iteration: 45, Training Loss: 1.3953496217727661, Validation Accuracy: 82.00370788574219%\n",
            "Iteration: 46, Training Loss: 1.3552117347717285, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 47, Training Loss: 1.3340542316436768, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 48, Training Loss: 1.3299791812896729, Validation Accuracy: 81.07606506347656%\n",
            "Iteration: 49, Training Loss: 1.3311271667480469, Validation Accuracy: 79.59183502197266%\n",
            "Iteration: 50, Training Loss: 1.3854155540466309, Validation Accuracy: 79.59183502197266%\n",
            "Iteration: 51, Training Loss: 1.3585926294326782, Validation Accuracy: 80.89054107666016%\n",
            "Iteration: 52, Training Loss: 1.32658052444458, Validation Accuracy: 82.00370788574219%\n",
            "Iteration: 53, Training Loss: 1.320422649383545, Validation Accuracy: 82.7458267211914%\n",
            "Iteration: 54, Training Loss: 1.2821847200393677, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 55, Training Loss: 1.2446904182434082, Validation Accuracy: 81.2615966796875%\n",
            "Iteration: 56, Training Loss: 1.2582963705062866, Validation Accuracy: 80.33395385742188%\n",
            "Iteration: 57, Training Loss: 1.3437104225158691, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 58, Training Loss: 1.3579779863357544, Validation Accuracy: 82.00370788574219%\n",
            "Iteration: 59, Training Loss: 1.3203026056289673, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 60, Training Loss: 1.2072231769561768, Validation Accuracy: 84.7866439819336%\n",
            "Iteration: 61, Training Loss: 1.2637732028961182, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 62, Training Loss: 1.3112190961837769, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 63, Training Loss: 1.3651789426803589, Validation Accuracy: 85.15769958496094%\n",
            "Iteration: 64, Training Loss: 1.2758313417434692, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 65, Training Loss: 1.3497098684310913, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 66, Training Loss: 1.3456084728240967, Validation Accuracy: 83.4879379272461%\n",
            "Iteration: 67, Training Loss: 1.3302947282791138, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 68, Training Loss: 1.3286856412887573, Validation Accuracy: 84.23005676269531%\n",
            "Iteration: 69, Training Loss: 1.30303955078125, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 70, Training Loss: 1.2676597833633423, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 71, Training Loss: 1.2941652536392212, Validation Accuracy: 84.7866439819336%\n",
            "Iteration: 72, Training Loss: 1.2837409973144531, Validation Accuracy: 83.11688232421875%\n",
            "Iteration: 73, Training Loss: 1.296205759048462, Validation Accuracy: 83.30241394042969%\n",
            "Iteration: 74, Training Loss: 1.3382513523101807, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 75, Training Loss: 1.2848732471466064, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 76, Training Loss: 1.324831485748291, Validation Accuracy: 82.18923950195312%\n",
            "Iteration: 77, Training Loss: 1.3752936124801636, Validation Accuracy: 80.14842224121094%\n",
            "Iteration: 78, Training Loss: 1.356689453125, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 79, Training Loss: 1.3107749223709106, Validation Accuracy: 79.59183502197266%\n",
            "Iteration: 80, Training Loss: 1.4376980066299438, Validation Accuracy: 79.96289825439453%\n",
            "Iteration: 81, Training Loss: 1.3577144145965576, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 82, Training Loss: 1.326814889907837, Validation Accuracy: 78.84972381591797%\n",
            "Iteration: 83, Training Loss: 1.3277784585952759, Validation Accuracy: 79.96289825439453%\n",
            "Iteration: 84, Training Loss: 1.3344277143478394, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 85, Training Loss: 1.265794277191162, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 86, Training Loss: 1.2734827995300293, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 87, Training Loss: 1.273010015487671, Validation Accuracy: 84.04452514648438%\n",
            "Iteration: 88, Training Loss: 1.2943029403686523, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 89, Training Loss: 1.3528892993927002, Validation Accuracy: 80.33395385742188%\n",
            "Iteration: 90, Training Loss: 1.295226812362671, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 91, Training Loss: 1.3411318063735962, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 92, Training Loss: 1.3325245380401611, Validation Accuracy: 84.04452514648438%\n",
            "Iteration: 93, Training Loss: 1.2603509426116943, Validation Accuracy: 81.81818389892578%\n",
            "Iteration: 94, Training Loss: 1.3711950778961182, Validation Accuracy: 82.7458267211914%\n",
            "Iteration: 95, Training Loss: 1.3557733297348022, Validation Accuracy: 81.44712829589844%\n",
            "Iteration: 96, Training Loss: 1.3270542621612549, Validation Accuracy: 81.2615966796875%\n",
            "Iteration: 97, Training Loss: 1.3457976579666138, Validation Accuracy: 80.70500946044922%\n",
            "Iteration: 98, Training Loss: 1.315621256828308, Validation Accuracy: 82.00370788574219%\n",
            "Iteration: 99, Training Loss: 1.287087321281433, Validation Accuracy: 81.07606506347656%\n",
            "Iteration: 100, Training Loss: 1.3855112791061401, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 101, Training Loss: 1.372514009475708, Validation Accuracy: 82.37477111816406%\n",
            "Iteration: 102, Training Loss: 1.2838715314865112, Validation Accuracy: 81.81818389892578%\n",
            "Iteration: 103, Training Loss: 1.3527029752731323, Validation Accuracy: 80.51947784423828%\n",
            "Iteration: 104, Training Loss: 1.3880432844161987, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 105, Training Loss: 1.3125556707382202, Validation Accuracy: 82.7458267211914%\n",
            "Iteration: 106, Training Loss: 1.333235740661621, Validation Accuracy: 82.00370788574219%\n",
            "Iteration: 107, Training Loss: 1.3106287717819214, Validation Accuracy: 80.89054107666016%\n",
            "Iteration: 108, Training Loss: 1.2934690713882446, Validation Accuracy: 80.89054107666016%\n",
            "Iteration: 109, Training Loss: 1.2787357568740845, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 110, Training Loss: 1.3056774139404297, Validation Accuracy: 84.23005676269531%\n",
            "Iteration: 111, Training Loss: 1.2869817018508911, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 112, Training Loss: 1.3437421321868896, Validation Accuracy: 84.23005676269531%\n",
            "Iteration: 113, Training Loss: 1.358159065246582, Validation Accuracy: 86.2708740234375%\n",
            "Iteration: 114, Training Loss: 1.2663946151733398, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 115, Training Loss: 1.2380825281143188, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 116, Training Loss: 1.305432915687561, Validation Accuracy: 83.85900115966797%\n",
            "Iteration: 117, Training Loss: 1.3177924156188965, Validation Accuracy: 84.04452514648438%\n",
            "Iteration: 118, Training Loss: 1.373531699180603, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 119, Training Loss: 1.351359486579895, Validation Accuracy: 81.63265228271484%\n",
            "Iteration: 120, Training Loss: 1.3072404861450195, Validation Accuracy: 82.18923950195312%\n",
            "Iteration: 121, Training Loss: 1.3124911785125732, Validation Accuracy: 80.14842224121094%\n",
            "Iteration: 122, Training Loss: 1.295076608657837, Validation Accuracy: 79.03524780273438%\n",
            "Iteration: 123, Training Loss: 1.40622079372406, Validation Accuracy: 81.81818389892578%\n",
            "Iteration: 124, Training Loss: 1.3093358278274536, Validation Accuracy: 82.7458267211914%\n",
            "Iteration: 125, Training Loss: 1.2972782850265503, Validation Accuracy: 81.81818389892578%\n",
            "Iteration: 126, Training Loss: 1.3458125591278076, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 127, Training Loss: 1.338187575340271, Validation Accuracy: 80.14842224121094%\n",
            "Iteration: 128, Training Loss: 1.3623408079147339, Validation Accuracy: 78.10760498046875%\n",
            "Iteration: 129, Training Loss: 1.288814663887024, Validation Accuracy: 81.2615966796875%\n",
            "Iteration: 130, Training Loss: 1.3084720373153687, Validation Accuracy: 82.18923950195312%\n",
            "Iteration: 131, Training Loss: 1.2666521072387695, Validation Accuracy: 83.30241394042969%\n",
            "Iteration: 132, Training Loss: 1.3124014139175415, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 133, Training Loss: 1.2955944538116455, Validation Accuracy: 84.97216796875%\n",
            "Iteration: 134, Training Loss: 1.3651190996170044, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 135, Training Loss: 1.2346850633621216, Validation Accuracy: 85.15769958496094%\n",
            "Iteration: 136, Training Loss: 1.2913706302642822, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 137, Training Loss: 1.3129383325576782, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 138, Training Loss: 1.3073444366455078, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 139, Training Loss: 1.3087972402572632, Validation Accuracy: 83.4879379272461%\n",
            "Iteration: 140, Training Loss: 1.3845000267028809, Validation Accuracy: 82.56029510498047%\n",
            "Iteration: 141, Training Loss: 1.3748923540115356, Validation Accuracy: 81.44712829589844%\n",
            "Iteration: 142, Training Loss: 1.3401321172714233, Validation Accuracy: 82.18923950195312%\n",
            "Iteration: 143, Training Loss: 1.2989311218261719, Validation Accuracy: 82.7458267211914%\n",
            "Iteration: 144, Training Loss: 1.300986409187317, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 145, Training Loss: 1.2455124855041504, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 146, Training Loss: 1.2520434856414795, Validation Accuracy: 86.4563980102539%\n",
            "Iteration: 147, Training Loss: 1.2994728088378906, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 148, Training Loss: 1.222357153892517, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 149, Training Loss: 1.2904592752456665, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 150, Training Loss: 1.3051261901855469, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 151, Training Loss: 1.2400352954864502, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 152, Training Loss: 1.2682546377182007, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 153, Training Loss: 1.244132161140442, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 154, Training Loss: 1.3097244501113892, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 155, Training Loss: 1.3229169845581055, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 156, Training Loss: 1.2526793479919434, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 157, Training Loss: 1.2572098970413208, Validation Accuracy: 86.2708740234375%\n",
            "Iteration: 158, Training Loss: 1.25676429271698, Validation Accuracy: 85.52875518798828%\n",
            "Iteration: 159, Training Loss: 1.3694366216659546, Validation Accuracy: 83.67346954345703%\n",
            "Iteration: 160, Training Loss: 1.2997171878814697, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 161, Training Loss: 1.2574039697647095, Validation Accuracy: 83.4879379272461%\n",
            "Iteration: 162, Training Loss: 1.2740674018859863, Validation Accuracy: 84.04452514648438%\n",
            "Iteration: 163, Training Loss: 1.3375099897384644, Validation Accuracy: 82.93135833740234%\n",
            "Iteration: 164, Training Loss: 1.4012266397476196, Validation Accuracy: 85.52875518798828%\n",
            "Iteration: 165, Training Loss: 1.3131450414657593, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 166, Training Loss: 1.2823994159698486, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 167, Training Loss: 1.337752103805542, Validation Accuracy: 86.64192962646484%\n",
            "Iteration: 168, Training Loss: 1.2190577983856201, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 169, Training Loss: 1.2720428705215454, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 170, Training Loss: 1.2778359651565552, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 171, Training Loss: 1.3595216274261475, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 172, Training Loss: 1.2646851539611816, Validation Accuracy: 85.89981842041016%\n",
            "Iteration: 173, Training Loss: 1.317016839981079, Validation Accuracy: 85.15769958496094%\n",
            "Iteration: 174, Training Loss: 1.2863744497299194, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 175, Training Loss: 1.3269284963607788, Validation Accuracy: 86.2708740234375%\n",
            "Iteration: 176, Training Loss: 1.3853684663772583, Validation Accuracy: 85.15769958496094%\n",
            "Iteration: 177, Training Loss: 1.2526482343673706, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 178, Training Loss: 1.2693589925765991, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 179, Training Loss: 1.2586268186569214, Validation Accuracy: 86.4563980102539%\n",
            "Iteration: 180, Training Loss: 1.3130239248275757, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 181, Training Loss: 1.2686352729797363, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 182, Training Loss: 1.2762504816055298, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 183, Training Loss: 1.3138118982315063, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 184, Training Loss: 1.3499492406845093, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 185, Training Loss: 1.3057975769042969, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 186, Training Loss: 1.3206062316894531, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 187, Training Loss: 1.281071424484253, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 188, Training Loss: 1.2991222143173218, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 189, Training Loss: 1.2702018022537231, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 190, Training Loss: 1.2730753421783447, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 191, Training Loss: 1.2236682176589966, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 192, Training Loss: 1.3602467775344849, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 193, Training Loss: 1.3174846172332764, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 194, Training Loss: 1.2664233446121216, Validation Accuracy: 84.97216796875%\n",
            "Iteration: 195, Training Loss: 1.2814539670944214, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 196, Training Loss: 1.2316319942474365, Validation Accuracy: 83.67346954345703%\n",
            "Iteration: 197, Training Loss: 1.2914804220199585, Validation Accuracy: 83.4879379272461%\n",
            "Iteration: 198, Training Loss: 1.2544167041778564, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 199, Training Loss: 1.2607654333114624, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 200, Training Loss: 1.2644984722137451, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 201, Training Loss: 1.310340404510498, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 202, Training Loss: 1.2726497650146484, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 203, Training Loss: 1.2923160791397095, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 204, Training Loss: 1.208917260169983, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 205, Training Loss: 1.2467225790023804, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 206, Training Loss: 1.2301914691925049, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 207, Training Loss: 1.2558525800704956, Validation Accuracy: 88.31169128417969%\n",
            "Iteration: 208, Training Loss: 1.2965530157089233, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 209, Training Loss: 1.3083924055099487, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 210, Training Loss: 1.3066807985305786, Validation Accuracy: 86.4563980102539%\n",
            "Iteration: 211, Training Loss: 1.2920323610305786, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 212, Training Loss: 1.3146593570709229, Validation Accuracy: 84.41558837890625%\n",
            "Iteration: 213, Training Loss: 1.271395206451416, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 214, Training Loss: 1.3381736278533936, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 215, Training Loss: 1.2822942733764648, Validation Accuracy: 85.52875518798828%\n",
            "Iteration: 216, Training Loss: 1.3211479187011719, Validation Accuracy: 85.34323120117188%\n",
            "Iteration: 217, Training Loss: 1.3461589813232422, Validation Accuracy: 86.2708740234375%\n",
            "Iteration: 218, Training Loss: 1.2887691259384155, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 219, Training Loss: 1.2716751098632812, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 220, Training Loss: 1.282823920249939, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 221, Training Loss: 1.3202704191207886, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 222, Training Loss: 1.3085907697677612, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 223, Training Loss: 1.2930335998535156, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 224, Training Loss: 1.2907360792160034, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 225, Training Loss: 1.2741895914077759, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 226, Training Loss: 1.2549102306365967, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 227, Training Loss: 1.2054206132888794, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 228, Training Loss: 1.2049082517623901, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 229, Training Loss: 1.2465381622314453, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 230, Training Loss: 1.224644422531128, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 231, Training Loss: 1.3199172019958496, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 232, Training Loss: 1.2864717245101929, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 233, Training Loss: 1.2194132804870605, Validation Accuracy: 86.64192962646484%\n",
            "Iteration: 234, Training Loss: 1.2351751327514648, Validation Accuracy: 86.82746124267578%\n",
            "Iteration: 235, Training Loss: 1.2642288208007812, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 236, Training Loss: 1.2702975273132324, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 237, Training Loss: 1.2474404573440552, Validation Accuracy: 87.38404846191406%\n",
            "Iteration: 238, Training Loss: 1.312038540840149, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 239, Training Loss: 1.3402941226959229, Validation Accuracy: 88.31169128417969%\n",
            "Iteration: 240, Training Loss: 1.2685304880142212, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 241, Training Loss: 1.2835359573364258, Validation Accuracy: 88.31169128417969%\n",
            "Iteration: 242, Training Loss: 1.2697559595108032, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 243, Training Loss: 1.3147414922714233, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 244, Training Loss: 1.2951388359069824, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 245, Training Loss: 1.219914436340332, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 246, Training Loss: 1.2211285829544067, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 247, Training Loss: 1.2747575044631958, Validation Accuracy: 88.4972152709961%\n",
            "Iteration: 248, Training Loss: 1.2706655263900757, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 249, Training Loss: 1.2502949237823486, Validation Accuracy: 84.7866439819336%\n",
            "Iteration: 250, Training Loss: 1.2855679988861084, Validation Accuracy: 86.08534240722656%\n",
            "Iteration: 251, Training Loss: 1.3778514862060547, Validation Accuracy: 86.64192962646484%\n",
            "Iteration: 252, Training Loss: 1.2397139072418213, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 253, Training Loss: 1.272625207901001, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 254, Training Loss: 1.2589240074157715, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 255, Training Loss: 1.3049863576889038, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 256, Training Loss: 1.28713858127594, Validation Accuracy: 88.4972152709961%\n",
            "Iteration: 257, Training Loss: 1.2344934940338135, Validation Accuracy: 88.4972152709961%\n",
            "Iteration: 258, Training Loss: 1.2180097103118896, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 259, Training Loss: 1.2552660703659058, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 260, Training Loss: 1.3012471199035645, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 261, Training Loss: 1.2210516929626465, Validation Accuracy: 86.64192962646484%\n",
            "Iteration: 262, Training Loss: 1.2933558225631714, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 263, Training Loss: 1.3224674463272095, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 264, Training Loss: 1.3076221942901611, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 265, Training Loss: 1.2692954540252686, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 266, Training Loss: 1.2984330654144287, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 267, Training Loss: 1.2187409400939941, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 268, Training Loss: 1.2497555017471313, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 269, Training Loss: 1.3228189945220947, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 270, Training Loss: 1.2525073289871216, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 271, Training Loss: 1.2611485719680786, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 272, Training Loss: 1.2170627117156982, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 273, Training Loss: 1.1835588216781616, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 274, Training Loss: 1.2427982091903687, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 275, Training Loss: 1.2575643062591553, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 276, Training Loss: 1.244837999343872, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 277, Training Loss: 1.2358745336532593, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 278, Training Loss: 1.225940465927124, Validation Accuracy: 88.12615966796875%\n",
            "Iteration: 279, Training Loss: 1.2421640157699585, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 280, Training Loss: 1.280775547027588, Validation Accuracy: 88.31169128417969%\n",
            "Iteration: 281, Training Loss: 1.2455140352249146, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 282, Training Loss: 1.267112374305725, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 283, Training Loss: 1.2996751070022583, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 284, Training Loss: 1.2964394092559814, Validation Accuracy: 90.35250854492188%\n",
            "Iteration: 285, Training Loss: 1.232173204421997, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 286, Training Loss: 1.2599862813949585, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 287, Training Loss: 1.2600072622299194, Validation Accuracy: 88.31169128417969%\n",
            "Iteration: 288, Training Loss: 1.2361196279525757, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 289, Training Loss: 1.228588342666626, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 290, Training Loss: 1.2564297914505005, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 291, Training Loss: 1.243573784828186, Validation Accuracy: 86.4563980102539%\n",
            "Iteration: 292, Training Loss: 1.2554899454116821, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 293, Training Loss: 1.330143928527832, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 294, Training Loss: 1.3018956184387207, Validation Accuracy: 87.56957244873047%\n",
            "Iteration: 295, Training Loss: 1.269566535949707, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 296, Training Loss: 1.3241204023361206, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 297, Training Loss: 1.2784863710403442, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 298, Training Loss: 1.2343690395355225, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 299, Training Loss: 1.2438075542449951, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 300, Training Loss: 1.2287334203720093, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 301, Training Loss: 1.2460423707962036, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 302, Training Loss: 1.2284338474273682, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 303, Training Loss: 1.263474702835083, Validation Accuracy: 88.4972152709961%\n",
            "Iteration: 304, Training Loss: 1.2289924621582031, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 305, Training Loss: 1.256547451019287, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 306, Training Loss: 1.2463393211364746, Validation Accuracy: 90.35250854492188%\n",
            "Iteration: 307, Training Loss: 1.2782313823699951, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 308, Training Loss: 1.2992581129074097, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 309, Training Loss: 1.2136222124099731, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 310, Training Loss: 1.253178596496582, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 311, Training Loss: 1.165642261505127, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 312, Training Loss: 1.2141236066818237, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 313, Training Loss: 1.255417823791504, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 314, Training Loss: 1.2821418046951294, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 315, Training Loss: 1.2268564701080322, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 316, Training Loss: 1.233109712600708, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 317, Training Loss: 1.1762723922729492, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 318, Training Loss: 1.2222143411636353, Validation Accuracy: 88.68274688720703%\n",
            "Iteration: 319, Training Loss: 1.2087783813476562, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 320, Training Loss: 1.201077938079834, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 321, Training Loss: 1.3079665899276733, Validation Accuracy: 90.35250854492188%\n",
            "Iteration: 322, Training Loss: 1.3059377670288086, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 323, Training Loss: 1.1930760145187378, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 324, Training Loss: 1.2432702779769897, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 325, Training Loss: 1.297032356262207, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 326, Training Loss: 1.2943507432937622, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 327, Training Loss: 1.2174142599105835, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 328, Training Loss: 1.2427626848220825, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 329, Training Loss: 1.2779074907302856, Validation Accuracy: 89.7959213256836%\n",
            "Iteration: 330, Training Loss: 1.2683357000350952, Validation Accuracy: 90.35250854492188%\n",
            "Iteration: 331, Training Loss: 1.2798986434936523, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 332, Training Loss: 1.261365294456482, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 333, Training Loss: 1.2806330919265747, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 334, Training Loss: 1.2698827981948853, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 335, Training Loss: 1.2333606481552124, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 336, Training Loss: 1.2618496417999268, Validation Accuracy: 87.01298522949219%\n",
            "Iteration: 337, Training Loss: 1.2941886186599731, Validation Accuracy: 85.71428680419922%\n",
            "Iteration: 338, Training Loss: 1.3343520164489746, Validation Accuracy: 85.89981842041016%\n",
            "Iteration: 339, Training Loss: 1.2482244968414307, Validation Accuracy: 85.15769958496094%\n",
            "Iteration: 340, Training Loss: 1.3080527782440186, Validation Accuracy: 83.4879379272461%\n",
            "Iteration: 341, Training Loss: 1.2924244403839111, Validation Accuracy: 83.11688232421875%\n",
            "Iteration: 342, Training Loss: 1.3830105066299438, Validation Accuracy: 83.30241394042969%\n",
            "Iteration: 343, Training Loss: 1.3371515274047852, Validation Accuracy: 84.60111236572266%\n",
            "Iteration: 344, Training Loss: 1.3110017776489258, Validation Accuracy: 86.64192962646484%\n",
            "Iteration: 345, Training Loss: 1.2581102848052979, Validation Accuracy: 87.94062805175781%\n",
            "Iteration: 346, Training Loss: 1.249882459640503, Validation Accuracy: 87.19851684570312%\n",
            "Iteration: 347, Training Loss: 1.2651287317276, Validation Accuracy: 87.7551040649414%\n",
            "Iteration: 348, Training Loss: 1.2454994916915894, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 349, Training Loss: 1.2439461946487427, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 350, Training Loss: 1.269405722618103, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 351, Training Loss: 1.2981131076812744, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 352, Training Loss: 1.2349613904953003, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 353, Training Loss: 1.2555062770843506, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 354, Training Loss: 1.251718521118164, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 355, Training Loss: 1.2474955320358276, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 356, Training Loss: 1.2328380346298218, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 357, Training Loss: 1.2647281885147095, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 358, Training Loss: 1.2570751905441284, Validation Accuracy: 89.23933410644531%\n",
            "Iteration: 359, Training Loss: 1.2540645599365234, Validation Accuracy: 88.86827850341797%\n",
            "Iteration: 360, Training Loss: 1.2769126892089844, Validation Accuracy: 89.05380249023438%\n",
            "Iteration: 361, Training Loss: 1.2391383647918701, Validation Accuracy: 89.42485809326172%\n",
            "Iteration: 362, Training Loss: 1.2848918437957764, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 363, Training Loss: 1.2533271312713623, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 364, Training Loss: 1.2413712739944458, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 365, Training Loss: 1.2089242935180664, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 366, Training Loss: 1.1784073114395142, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 367, Training Loss: 1.2707921266555786, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 368, Training Loss: 1.2304985523223877, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 369, Training Loss: 1.2159818410873413, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 370, Training Loss: 1.2822047472000122, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 371, Training Loss: 1.2037726640701294, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 372, Training Loss: 1.2563292980194092, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 373, Training Loss: 1.2042120695114136, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 374, Training Loss: 1.2437825202941895, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 375, Training Loss: 1.2353079319000244, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 376, Training Loss: 1.3162091970443726, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 377, Training Loss: 1.1788569688796997, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 378, Training Loss: 1.2389631271362305, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 379, Training Loss: 1.2178243398666382, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 380, Training Loss: 1.2534005641937256, Validation Accuracy: 89.61038970947266%\n",
            "Iteration: 381, Training Loss: 1.25856351852417, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 382, Training Loss: 1.204140543937683, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 383, Training Loss: 1.2401319742202759, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 384, Training Loss: 1.1915583610534668, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 385, Training Loss: 1.2325046062469482, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 386, Training Loss: 1.2350637912750244, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 387, Training Loss: 1.229472041130066, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 388, Training Loss: 1.2869588136672974, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 389, Training Loss: 1.2279572486877441, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 390, Training Loss: 1.2174087762832642, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 391, Training Loss: 1.1992548704147339, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 392, Training Loss: 1.1795042753219604, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 393, Training Loss: 1.3069393634796143, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 394, Training Loss: 1.2198787927627563, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 395, Training Loss: 1.3115465641021729, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 396, Training Loss: 1.2174389362335205, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 397, Training Loss: 1.2041218280792236, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 398, Training Loss: 1.2497620582580566, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 399, Training Loss: 1.2517259120941162, Validation Accuracy: 90.53803253173828%\n",
            "Iteration: 400, Training Loss: 1.2676469087600708, Validation Accuracy: 90.16697692871094%\n",
            "Iteration: 401, Training Loss: 1.2961113452911377, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 402, Training Loss: 1.3074578046798706, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 403, Training Loss: 1.265071153640747, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 404, Training Loss: 1.2044408321380615, Validation Accuracy: 90.35250854492188%\n",
            "Iteration: 405, Training Loss: 1.2107892036437988, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 406, Training Loss: 1.2551007270812988, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 407, Training Loss: 1.239438772201538, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 408, Training Loss: 1.18081796169281, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 409, Training Loss: 1.2044521570205688, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 410, Training Loss: 1.2056019306182861, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 411, Training Loss: 1.2322185039520264, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 412, Training Loss: 1.2434126138687134, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 413, Training Loss: 1.2723731994628906, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 414, Training Loss: 1.2984209060668945, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 415, Training Loss: 1.2094712257385254, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 416, Training Loss: 1.2272483110427856, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 417, Training Loss: 1.2823762893676758, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 418, Training Loss: 1.2482290267944336, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 419, Training Loss: 1.217861294746399, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 420, Training Loss: 1.2268528938293457, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 421, Training Loss: 1.268175482749939, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 422, Training Loss: 1.2812049388885498, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 423, Training Loss: 1.214714765548706, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 424, Training Loss: 1.2036912441253662, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 425, Training Loss: 1.193245530128479, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 426, Training Loss: 1.2355202436447144, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 427, Training Loss: 1.2161941528320312, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 428, Training Loss: 1.2548271417617798, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 429, Training Loss: 1.246140956878662, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 430, Training Loss: 1.2091646194458008, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 431, Training Loss: 1.2015002965927124, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 432, Training Loss: 1.2478816509246826, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 433, Training Loss: 1.240647315979004, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 434, Training Loss: 1.1903841495513916, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 435, Training Loss: 1.2449171543121338, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 436, Training Loss: 1.260107398033142, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 437, Training Loss: 1.2414052486419678, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 438, Training Loss: 1.2569571733474731, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 439, Training Loss: 1.212172508239746, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 440, Training Loss: 1.2274147272109985, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 441, Training Loss: 1.2954747676849365, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 442, Training Loss: 1.2101449966430664, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 443, Training Loss: 1.204401969909668, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 444, Training Loss: 1.1999754905700684, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 445, Training Loss: 1.2511707544326782, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 446, Training Loss: 1.2700556516647339, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 447, Training Loss: 1.2214114665985107, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 448, Training Loss: 1.2762116193771362, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 449, Training Loss: 1.1920952796936035, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 450, Training Loss: 1.1914546489715576, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 451, Training Loss: 1.2178376913070679, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 452, Training Loss: 1.1926798820495605, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 453, Training Loss: 1.2055481672286987, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 454, Training Loss: 1.2355741262435913, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 455, Training Loss: 1.2097543478012085, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 456, Training Loss: 1.2002497911453247, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 457, Training Loss: 1.2695131301879883, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 458, Training Loss: 1.2119684219360352, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 459, Training Loss: 1.2348219156265259, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 460, Training Loss: 1.1908332109451294, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 461, Training Loss: 1.230573296546936, Validation Accuracy: 90.90908813476562%\n",
            "Iteration: 462, Training Loss: 1.2034764289855957, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 463, Training Loss: 1.2157313823699951, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 464, Training Loss: 1.2278441190719604, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 465, Training Loss: 1.2197574377059937, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 466, Training Loss: 1.2173922061920166, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 467, Training Loss: 1.250687599182129, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 468, Training Loss: 1.2613412141799927, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 469, Training Loss: 1.178411841392517, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 470, Training Loss: 1.2041983604431152, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 471, Training Loss: 1.199035882949829, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 472, Training Loss: 1.1785593032836914, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 473, Training Loss: 1.2120318412780762, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 474, Training Loss: 1.1989163160324097, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 475, Training Loss: 1.2047288417816162, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 476, Training Loss: 1.2022790908813477, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 477, Training Loss: 1.2004460096359253, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 478, Training Loss: 1.2185918092727661, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 479, Training Loss: 1.2695670127868652, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 480, Training Loss: 1.2176990509033203, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 481, Training Loss: 1.2558823823928833, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 482, Training Loss: 1.2174738645553589, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 483, Training Loss: 1.2679299116134644, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 484, Training Loss: 1.2002524137496948, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 485, Training Loss: 1.2034740447998047, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 486, Training Loss: 1.217432975769043, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 487, Training Loss: 1.1714520454406738, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 488, Training Loss: 1.237074613571167, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 489, Training Loss: 1.2305442094802856, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 490, Training Loss: 1.2305794954299927, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 491, Training Loss: 1.2019398212432861, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 492, Training Loss: 1.2284739017486572, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 493, Training Loss: 1.197562336921692, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 494, Training Loss: 1.2408630847930908, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 495, Training Loss: 1.2550938129425049, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 496, Training Loss: 1.2047141790390015, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 497, Training Loss: 1.2400121688842773, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 498, Training Loss: 1.270944356918335, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 499, Training Loss: 1.1959060430526733, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 500, Training Loss: 1.1789741516113281, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 501, Training Loss: 1.2421555519104004, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 502, Training Loss: 1.1780604124069214, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 503, Training Loss: 1.179569959640503, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 504, Training Loss: 1.1925088167190552, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 505, Training Loss: 1.2563438415527344, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 506, Training Loss: 1.2593363523483276, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 507, Training Loss: 1.215045690536499, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 508, Training Loss: 1.2106951475143433, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 509, Training Loss: 1.1908949613571167, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 510, Training Loss: 1.1920554637908936, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 511, Training Loss: 1.2043894529342651, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 512, Training Loss: 1.2674639225006104, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 513, Training Loss: 1.2154796123504639, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 514, Training Loss: 1.2039978504180908, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 515, Training Loss: 1.191959261894226, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 516, Training Loss: 1.195400357246399, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 517, Training Loss: 1.2173582315444946, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 518, Training Loss: 1.2002848386764526, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 519, Training Loss: 1.1913232803344727, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 520, Training Loss: 1.1737127304077148, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 521, Training Loss: 1.1916710138320923, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 522, Training Loss: 1.2301353216171265, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 523, Training Loss: 1.1844648122787476, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 524, Training Loss: 1.2512545585632324, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 525, Training Loss: 1.2312043905258179, Validation Accuracy: 91.83673858642578%\n",
            "Iteration: 526, Training Loss: 1.2087994813919067, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 527, Training Loss: 1.196226716041565, Validation Accuracy: 89.9814453125%\n",
            "Iteration: 528, Training Loss: 1.184184193611145, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 529, Training Loss: 1.2784615755081177, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 530, Training Loss: 1.2169333696365356, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 531, Training Loss: 1.2079848051071167, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 532, Training Loss: 1.2284629344940186, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 533, Training Loss: 1.2122858762741089, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 534, Training Loss: 1.2007012367248535, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 535, Training Loss: 1.2251660823822021, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 536, Training Loss: 1.1751914024353027, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 537, Training Loss: 1.1966973543167114, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 538, Training Loss: 1.2118910551071167, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 539, Training Loss: 1.230385422706604, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 540, Training Loss: 1.1887117624282837, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 541, Training Loss: 1.1916142702102661, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 542, Training Loss: 1.204393982887268, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 543, Training Loss: 1.2557088136672974, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 544, Training Loss: 1.2611016035079956, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 545, Training Loss: 1.1785914897918701, Validation Accuracy: 94.8051986694336%\n",
            "Iteration: 546, Training Loss: 1.2177729606628418, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 547, Training Loss: 1.2115076780319214, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 548, Training Loss: 1.2303276062011719, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 549, Training Loss: 1.2311769723892212, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 550, Training Loss: 1.2094429731369019, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 551, Training Loss: 1.1784181594848633, Validation Accuracy: 94.61966705322266%\n",
            "Iteration: 552, Training Loss: 1.1654222011566162, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 553, Training Loss: 1.1971185207366943, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 554, Training Loss: 1.2446846961975098, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 555, Training Loss: 1.1919875144958496, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 556, Training Loss: 1.2048524618148804, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 557, Training Loss: 1.2312372922897339, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 558, Training Loss: 1.2175296545028687, Validation Accuracy: 90.72356414794922%\n",
            "Iteration: 559, Training Loss: 1.1914159059524536, Validation Accuracy: 91.2801513671875%\n",
            "Iteration: 560, Training Loss: 1.2278555631637573, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 561, Training Loss: 1.2080541849136353, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 562, Training Loss: 1.2855663299560547, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 563, Training Loss: 1.1815584897994995, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 564, Training Loss: 1.2300342321395874, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 565, Training Loss: 1.2381354570388794, Validation Accuracy: 91.4656753540039%\n",
            "Iteration: 566, Training Loss: 1.2363368272781372, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 567, Training Loss: 1.2166948318481445, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 568, Training Loss: 1.1784098148345947, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 569, Training Loss: 1.2382524013519287, Validation Accuracy: 91.09461975097656%\n",
            "Iteration: 570, Training Loss: 1.2201869487762451, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 571, Training Loss: 1.2167353630065918, Validation Accuracy: 92.94990539550781%\n",
            "Iteration: 572, Training Loss: 1.2245005369186401, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 573, Training Loss: 1.1789675951004028, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 574, Training Loss: 1.268685221672058, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 575, Training Loss: 1.2187577486038208, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 576, Training Loss: 1.2470186948776245, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 577, Training Loss: 1.2173917293548584, Validation Accuracy: 92.02226257324219%\n",
            "Iteration: 578, Training Loss: 1.2082077264785767, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 579, Training Loss: 1.2364397048950195, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 580, Training Loss: 1.193967342376709, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 581, Training Loss: 1.2446449995040894, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 582, Training Loss: 1.2110228538513184, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 583, Training Loss: 1.209725260734558, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 584, Training Loss: 1.2177778482437134, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 585, Training Loss: 1.2438673973083496, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 586, Training Loss: 1.2001358270645142, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 587, Training Loss: 1.2263644933700562, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 588, Training Loss: 1.2283836603164673, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 589, Training Loss: 1.2336211204528809, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 590, Training Loss: 1.217767357826233, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 591, Training Loss: 1.1814982891082764, Validation Accuracy: 92.20779418945312%\n",
            "Iteration: 592, Training Loss: 1.1997276544570923, Validation Accuracy: 91.65120697021484%\n",
            "Iteration: 593, Training Loss: 1.2243387699127197, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 594, Training Loss: 1.1976512670516968, Validation Accuracy: 93.69202423095703%\n",
            "Iteration: 595, Training Loss: 1.2177538871765137, Validation Accuracy: 93.13543701171875%\n",
            "Iteration: 596, Training Loss: 1.1963022947311401, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 597, Training Loss: 1.2195522785186768, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 598, Training Loss: 1.1918357610702515, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 599, Training Loss: 1.1912299394607544, Validation Accuracy: 94.43413543701172%\n",
            "Iteration: 600, Training Loss: 1.1785439252853394, Validation Accuracy: 93.5064926147461%\n",
            "Iteration: 601, Training Loss: 1.1821749210357666, Validation Accuracy: 93.87754821777344%\n",
            "Iteration: 602, Training Loss: 1.1823902130126953, Validation Accuracy: 94.06307983398438%\n",
            "Iteration: 603, Training Loss: 1.230440616607666, Validation Accuracy: 92.39331817626953%\n",
            "Iteration: 604, Training Loss: 1.2411417961120605, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 605, Training Loss: 1.2119897603988647, Validation Accuracy: 92.7643814086914%\n",
            "Iteration: 606, Training Loss: 1.282494068145752, Validation Accuracy: 93.32096862792969%\n",
            "Iteration: 607, Training Loss: 1.1779816150665283, Validation Accuracy: 92.57884979248047%\n",
            "Iteration: 608, Training Loss: 1.2584269046783447, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 609, Training Loss: 1.2041150331497192, Validation Accuracy: 94.24861145019531%\n",
            "Iteration: 610, Training Loss: 1.2200254201889038, Validation Accuracy: 94.24861145019531%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WKfCVoxDYamp",
        "outputId": "2fe193b4-04a9-49bf-dce9-5ea22cf9d67c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        }
      },
      "source": [
        "plt.plot(iteration_list,loss_list)\n",
        "plt.title(\"Train Loss\")\n",
        "plt.xlabel(\"Iterations\")\n",
        "plt.ylabel(\"Training Loss\")\n",
        "plt.show()"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 385,
              "height": 277
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U2GXbbsZ1QZw",
        "outputId": "d9dd78be-3833-48cf-8b33-c4c7c35a75fe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        }
      },
      "source": [
        "plt.plot(iteration_list,accuracy_list)\n",
        "plt.title(\"Validation Accuracy\")\n",
        "plt.xlabel(\"Iterations\")\n",
        "plt.ylabel(\"Validation Accuracy\")\n",
        "plt.show()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "image/png": {
              "width": 382,
              "height": 277
            },
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4MKr0QcW8q4X"
      },
      "source": [
        "\"\"\" To save the model \"\"\"\n",
        "torch.save(model.state_dict(),\"model_name\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wclzSe3TOcRK"
      },
      "source": [
        "\"\"\" To load the model \"\"\"\n",
        "model = LocationCNN()\n",
        "model.load_state_dict(torch.load(\"model_name\"))\n",
        "model.eval()\n",
        "model.to(device)"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}