S12/EVA4S12.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.6"
    },
    "colab": {
      "name": "EVA4S12.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fj2Pu3zBSJy1",
        "colab_type": "text"
      },
      "source": [
        "# Installing Packages"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "eHrZbdWRjalh",
        "colab": {}
      },
      "source": [
        "!pip install --no-cache-dir -r requirements.txt"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aO-7t1Y7-hV4",
        "colab_type": "text"
      },
      "source": [
        "# Imports\n",
        "Importing necessary packages and modules"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-yVpodAMpoCi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "\n",
        "import random\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from tensornet.data.datasets import TinyImageNet\n",
        "from tensornet.model import ResNet18\n",
        "from tensornet.model.utils.loss import cross_entropy_loss\n",
        "from tensornet.model.utils.optimizers import sgd\n",
        "from tensornet.engine.ops.lr_scheduler import reduce_lr_on_plateau\n",
        "from tensornet.utils import initialize_cuda, plot_metric"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ve_XDp27GSN_",
        "colab_type": "text"
      },
      "source": [
        "## Set Seed and Get GPU Availability"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CxVy79mmfQ1e",
        "colab_type": "code",
        "outputId": "cbb208d7-0966-4824-b7e0-a49a72488a84",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# Initialize CUDA and set random seed\n",
        "cuda, device = initialize_cuda(1)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "GPU Available? True\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cKi2t8X-CKJ5",
        "colab_type": "text"
      },
      "source": [
        "## Download Dataset\n",
        "\n",
        "Importing the TinyImagenet class to download dataset and create data loader"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QZjDtZVLGXh-",
        "colab_type": "code",
        "outputId": "13addccd-b8cd-4f06-9b34-9ac461b21be4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 70
        }
      },
      "source": [
        "dataset = TinyImageNet(\n",
        "    train_batch_size=128,\n",
        "    val_batch_size=128,\n",
        "    cuda=cuda,\n",
        "    num_workers=16,\n",
        "    horizontal_flip_prob=0.3,\n",
        "    vertical_flip_prob=0.1,\n",
        "    rotate_degree=10,\n",
        "    cutout_prob=0.3,\n",
        "    cutout_dim=(16, 16),\n",
        ")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Files already downloaded.\n",
            "Files already downloaded.\n",
            "Files already downloaded.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jugzRPudr2MW",
        "colab_type": "text"
      },
      "source": [
        "## Data Visualization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RUusUZ9emk2M",
        "colab_type": "code",
        "outputId": "4f250d15-e784-417f-d9a8-71cc2fb416c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 146
        }
      },
      "source": [
        "# Fetch data\n",
        "classes = dataset.classes\n",
        "\n",
        "# Set number of images to display\n",
        "num_images = 4\n",
        "\n",
        "# Display images with labels\n",
        "fig, axs = plt.subplots(1, 4, figsize=(8, 8))\n",
        "fig.tight_layout()\n",
        "\n",
        "for i in range(num_images):\n",
        "    idx = random.randint(0, len(dataset.val_data))\n",
        "    axs[i].axis('off')\n",
        "    axs[i].set_title(f'Label: {classes[dataset.val_data[idx][1]].split(\",\")[0]}')\n",
        "    axs[i].imshow(dataset.unnormalize(dataset.val_data[idx][0], transpose=True))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 4 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qFcWRAeY_MsB",
        "colab_type": "text"
      },
      "source": [
        "## Training and Validation Dataloaders\n",
        "This is the final step in data preparation. It sets the dataloader arguments and then creates the dataloader"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "D8Tjigo2poCl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Create train data loader\n",
        "train_loader = dataset.loader(train=True)\n",
        "\n",
        "# Create val data loader\n",
        "val_loader = dataset.loader(train=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HSeDIqBq-6pC",
        "colab_type": "text"
      },
      "source": [
        "# Model Architecture and Summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z_0Bgttg6MhZ",
        "colab_type": "code",
        "outputId": "25dabbf8-f0a8-4479-af8d-3463b5214f1f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "model = ResNet18(num_classes=len(dataset.classes)).to(device)\n",
        "model.summary(dataset.image_size)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "----------------------------------------------------------------\n",
            "        Layer (type)               Output Shape         Param #\n",
            "================================================================\n",
            "            Conv2d-1           [-1, 64, 64, 64]           1,728\n",
            "       BatchNorm2d-2           [-1, 64, 64, 64]             128\n",
            "            Conv2d-3           [-1, 64, 64, 64]          36,864\n",
            "       BatchNorm2d-4           [-1, 64, 64, 64]             128\n",
            "            Conv2d-5           [-1, 64, 64, 64]          36,864\n",
            "       BatchNorm2d-6           [-1, 64, 64, 64]             128\n",
            "        BasicBlock-7           [-1, 64, 64, 64]               0\n",
            "            Conv2d-8           [-1, 64, 64, 64]          36,864\n",
            "       BatchNorm2d-9           [-1, 64, 64, 64]             128\n",
            "           Conv2d-10           [-1, 64, 64, 64]          36,864\n",
            "      BatchNorm2d-11           [-1, 64, 64, 64]             128\n",
            "       BasicBlock-12           [-1, 64, 64, 64]               0\n",
            "           Conv2d-13          [-1, 128, 32, 32]          73,728\n",
            "      BatchNorm2d-14          [-1, 128, 32, 32]             256\n",
            "           Conv2d-15          [-1, 128, 32, 32]         147,456\n",
            "      BatchNorm2d-16          [-1, 128, 32, 32]             256\n",
            "           Conv2d-17          [-1, 128, 32, 32]           8,192\n",
            "      BatchNorm2d-18          [-1, 128, 32, 32]             256\n",
            "       BasicBlock-19          [-1, 128, 32, 32]               0\n",
            "           Conv2d-20          [-1, 128, 32, 32]         147,456\n",
            "      BatchNorm2d-21          [-1, 128, 32, 32]             256\n",
            "           Conv2d-22          [-1, 128, 32, 32]         147,456\n",
            "      BatchNorm2d-23          [-1, 128, 32, 32]             256\n",
            "       BasicBlock-24          [-1, 128, 32, 32]               0\n",
            "           Conv2d-25          [-1, 256, 16, 16]         294,912\n",
            "      BatchNorm2d-26          [-1, 256, 16, 16]             512\n",
            "           Conv2d-27          [-1, 256, 16, 16]         589,824\n",
            "      BatchNorm2d-28          [-1, 256, 16, 16]             512\n",
            "           Conv2d-29          [-1, 256, 16, 16]          32,768\n",
            "      BatchNorm2d-30          [-1, 256, 16, 16]             512\n",
            "       BasicBlock-31          [-1, 256, 16, 16]               0\n",
            "           Conv2d-32          [-1, 256, 16, 16]         589,824\n",
            "      BatchNorm2d-33          [-1, 256, 16, 16]             512\n",
            "           Conv2d-34          [-1, 256, 16, 16]         589,824\n",
            "      BatchNorm2d-35          [-1, 256, 16, 16]             512\n",
            "       BasicBlock-36          [-1, 256, 16, 16]               0\n",
            "           Conv2d-37            [-1, 512, 8, 8]       1,179,648\n",
            "      BatchNorm2d-38            [-1, 512, 8, 8]           1,024\n",
            "           Conv2d-39            [-1, 512, 8, 8]       2,359,296\n",
            "      BatchNorm2d-40            [-1, 512, 8, 8]           1,024\n",
            "           Conv2d-41            [-1, 512, 8, 8]         131,072\n",
            "      BatchNorm2d-42            [-1, 512, 8, 8]           1,024\n",
            "       BasicBlock-43            [-1, 512, 8, 8]               0\n",
            "           Conv2d-44            [-1, 512, 8, 8]       2,359,296\n",
            "      BatchNorm2d-45            [-1, 512, 8, 8]           1,024\n",
            "           Conv2d-46            [-1, 512, 8, 8]       2,359,296\n",
            "      BatchNorm2d-47            [-1, 512, 8, 8]           1,024\n",
            "       BasicBlock-48            [-1, 512, 8, 8]               0\n",
            "AdaptiveAvgPool2d-49            [-1, 512, 1, 1]               0\n",
            "           Linear-50                  [-1, 200]         102,600\n",
            "================================================================\n",
            "Total params: 11,271,432\n",
            "Trainable params: 11,271,432\n",
            "Non-trainable params: 0\n",
            "----------------------------------------------------------------\n",
            "Input size (MB): 0.05\n",
            "Forward/backward pass size (MB): 45.01\n",
            "Params size (MB): 43.00\n",
            "Estimated Total Size (MB): 88.05\n",
            "----------------------------------------------------------------\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HcyKrcQEGcdp",
        "colab_type": "text"
      },
      "source": [
        "# Model Training and Validation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YLB26HeHpCZ3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "optimizer = sgd(model, learning_rate=0.01, momentum=0.9)  # Create optimizer\n",
        "criterion = cross_entropy_loss()  # Create loss function"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "5EBoL93BVb4Z",
        "outputId": "e420760b-f51e-40c2-fce8-65d12925ed77",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "epochs = 50\n",
        "\n",
        "# Setup Reduce LR on Plateau\n",
        "scheduler = reduce_lr_on_plateau(optimizer, patience=2, verbose=True, min_lr=1e-6)\n",
        "\n",
        "# Model Training\n",
        "model.fit(\n",
        "    train_loader, optimizer, criterion, device=device, epochs=epochs,\n",
        "    val_loader=val_loader, callbacks=[scheduler]\n",
        ")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1:\n",
            "602/602 [========] - 149s 247ms/step - loss: 4.4573 - accuracy: 4.7045\n",
            "Validation set: Average loss: 0.0331, Accuracy: 11.21%\n",
            "\n",
            "Epoch 2:\n",
            "602/602 [========] - 148s 247ms/step - loss: 3.6814 - accuracy: 10.5993\n",
            "Validation set: Average loss: 0.0295, Accuracy: 16.54%\n",
            "\n",
            "Epoch 3:\n",
            "602/602 [========] - 148s 246ms/step - loss: 3.2787 - accuracy: 14.9748\n",
            "Validation set: Average loss: 0.0278, Accuracy: 20.88%\n",
            "\n",
            "Epoch 4:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.9940 - accuracy: 18.5150\n",
            "Validation set: Average loss: 0.0251, Accuracy: 27.20%\n",
            "\n",
            "Epoch 5:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.7735 - accuracy: 21.4928\n",
            "Validation set: Average loss: 0.0221, Accuracy: 33.30%\n",
            "\n",
            "Epoch 6:\n",
            "602/602 [========] - 148s 245ms/step - loss: 2.5921 - accuracy: 24.1308\n",
            "Validation set: Average loss: 0.0211, Accuracy: 35.80%\n",
            "\n",
            "Epoch 7:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.4434 - accuracy: 26.4772\n",
            "Validation set: Average loss: 0.0210, Accuracy: 37.10%\n",
            "\n",
            "Epoch 8:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.3087 - accuracy: 28.5611\n",
            "Validation set: Average loss: 0.0214, Accuracy: 36.38%\n",
            "\n",
            "Epoch 9:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.1871 - accuracy: 30.4464\n",
            "Validation set: Average loss: 0.0186, Accuracy: 42.92%\n",
            "\n",
            "Epoch 10:\n",
            "602/602 [========] - 148s 246ms/step - loss: 2.0793 - accuracy: 32.1968\n",
            "Validation set: Average loss: 0.0193, Accuracy: 41.51%\n",
            "\n",
            "Epoch 11:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.9801 - accuracy: 33.8205\n",
            "Validation set: Average loss: 0.0198, Accuracy: 41.33%\n",
            "\n",
            "Epoch 12:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.8828 - accuracy: 35.3500\n",
            "Validation set: Average loss: 0.0189, Accuracy: 43.06%\n",
            "\n",
            "Epoch    12: reducing learning rate of group 0 to 1.0000e-03.\n",
            "Epoch 13:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.5230 - accuracy: 37.0570\n",
            "Validation set: Average loss: 0.0149, Accuracy: 53.13%\n",
            "\n",
            "Epoch 14:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.4370 - accuracy: 38.9765\n",
            "Validation set: Average loss: 0.0148, Accuracy: 53.71%\n",
            "\n",
            "Epoch 15:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.3978 - accuracy: 40.7305\n",
            "Validation set: Average loss: 0.0148, Accuracy: 53.66%\n",
            "\n",
            "Epoch 16:\n",
            "602/602 [========] - 148s 247ms/step - loss: 1.3710 - accuracy: 42.3051\n",
            "Validation set: Average loss: 0.0147, Accuracy: 53.72%\n",
            "\n",
            "Epoch 17:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.3431 - accuracy: 43.7315\n",
            "Validation set: Average loss: 0.0147, Accuracy: 54.03%\n",
            "\n",
            "Epoch 18:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.3177 - accuracy: 45.0288\n",
            "Validation set: Average loss: 0.0147, Accuracy: 53.81%\n",
            "\n",
            "Epoch 19:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.2914 - accuracy: 46.2251\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.08%\n",
            "\n",
            "Epoch 20:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.2702 - accuracy: 47.3214\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.12%\n",
            "\n",
            "Epoch 21:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.2480 - accuracy: 48.3362\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.19%\n",
            "\n",
            "Epoch 22:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.2217 - accuracy: 49.2835\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.37%\n",
            "\n",
            "Epoch 23:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.1980 - accuracy: 50.1770\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.38%\n",
            "\n",
            "Epoch 24:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.1777 - accuracy: 51.0138\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.40%\n",
            "\n",
            "Epoch 25:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.1545 - accuracy: 51.8104\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.60%\n",
            "\n",
            "Epoch 26:\n",
            "602/602 [========] - 148s 246ms/step - loss: 1.1344 - accuracy: 52.5702\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.56%\n",
            "\n",
            "Epoch 27:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.1105 - accuracy: 53.2898\n",
            "Validation set: Average loss: 0.0146, Accuracy: 54.53%\n",
            "\n",
            "Epoch 28:\n",
            "602/602 [========] - 148s 247ms/step - loss: 1.0900 - accuracy: 53.9797\n",
            "Validation set: Average loss: 0.0147, Accuracy: 54.30%\n",
            "\n",
            "Epoch    28: reducing learning rate of group 0 to 1.0000e-04.\n",
            "Epoch 29:\n",
            "602/602 [========] - 148s 247ms/step - loss: 1.0218 - accuracy: 54.6586\n",
            "Validation set: Average loss: 0.0144, Accuracy: 55.20%\n",
            "\n",
            "Epoch 30:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.0063 - accuracy: 55.3448\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.37%\n",
            "\n",
            "Epoch 31:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.0078 - accuracy: 55.9931\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.36%\n",
            "\n",
            "Epoch 32:\n",
            "602/602 [========] - 149s 247ms/step - loss: 1.0011 - accuracy: 56.6004\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.34%\n",
            "\n",
            "Epoch 33:\n",
            "602/602 [========] - 149s 247ms/step - loss: 0.9956 - accuracy: 57.1813\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.34%\n",
            "\n",
            "Epoch 34:\n",
            "602/602 [========] - 148s 247ms/step - loss: 0.9877 - accuracy: 57.7291\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.35%\n",
            "\n",
            "Epoch 35:\n",
            "602/602 [========] - 148s 247ms/step - loss: 0.9905 - accuracy: 58.2484\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.30%\n",
            "\n",
            "Epoch    35: reducing learning rate of group 0 to 1.0000e-05.\n",
            "Epoch 36:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9816 - accuracy: 58.7375\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.26%\n",
            "\n",
            "Epoch 37:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9768 - accuracy: 59.2100\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.37%\n",
            "\n",
            "Epoch 38:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9748 - accuracy: 59.6557\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.31%\n",
            "\n",
            "Epoch 39:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9760 - accuracy: 60.0803\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.35%\n",
            "\n",
            "Epoch 40:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9816 - accuracy: 60.4804\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.28%\n",
            "\n",
            "Epoch    40: reducing learning rate of group 0 to 1.0000e-06.\n",
            "Epoch 41:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9782 - accuracy: 60.8632\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.35%\n",
            "\n",
            "Epoch 42:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9785 - accuracy: 61.2288\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.33%\n",
            "\n",
            "Epoch 43:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9821 - accuracy: 61.5730\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.39%\n",
            "\n",
            "Epoch 44:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9742 - accuracy: 61.9053\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.33%\n",
            "\n",
            "Epoch 45:\n",
            "602/602 [========] - 148s 247ms/step - loss: 0.9775 - accuracy: 62.2206\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.42%\n",
            "\n",
            "Epoch 46:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9771 - accuracy: 62.5191\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.36%\n",
            "\n",
            "Epoch 47:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9772 - accuracy: 62.8079\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.31%\n",
            "\n",
            "Epoch 48:\n",
            "602/602 [========] - 148s 246ms/step - loss: 0.9740 - accuracy: 63.0871\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.31%\n",
            "\n",
            "Epoch 49:\n",
            "602/602 [========] - 148s 247ms/step - loss: 0.9743 - accuracy: 63.3537\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.31%\n",
            "\n",
            "Epoch 50:\n",
            "602/602 [========] - 149s 247ms/step - loss: 0.9750 - accuracy: 63.6093\n",
            "Validation set: Average loss: 0.0143, Accuracy: 55.38%\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "10araB7Y6wVS",
        "colab_type": "text"
      },
      "source": [
        "# Plotting Results\n",
        "\n",
        "Plot changes in validation accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lBXFeKGvXJba",
        "colab_type": "code",
        "outputId": "d3bcce9d-f522-441a-94dd-b806a84e60c7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 350
        }
      },
      "source": [
        "plot_metric({\n",
        "    'Training': model.learner.train_accuracies,\n",
        "    'Validation': model.learner.val_accuracies\n",
        "}, 'Accuracy')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbYAAAFNCAYAAABsXEqqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOzdd3xUVfr48c9JbyQhpBBCIAUCoSNFEGwIikpRd0WxgQ1d9WvflXVZXdd119V1xfLbdbGigCgCggI2RBBBIKHXACkQSCW9JzPn98edhAQCmZBMZpJ53q/Xfd2Z286TIcyTU+65SmuNEEII0VG42DsAIYQQojVJYhNCCNGhSGITQgjRoUhiE0II0aFIYhNCCNGhSGITQgjRoUhiE8KJKKU+Ukr9zd5xCGFLktiE01BK/aSUyldKedo7FltRhkeVUnuVUqVKqXSl1BKl1EB7xyZEW5HEJpyCUioKuBTQwJQ2LtutDYt7A3gMeBQIAuKAL4Hr2zAGIexKEptwFncBvwIfATPq71BKeSulXlNKpSmlCpVSG5VS3pZ9Y5VSm5RSBUqp40qpmZbtPyml7qt3jZlKqY313mul1MNKqcPAYcu2NyzXKFJKJSqlLq13vKtS6lml1FGlVLFlf6RS6v8ppV47I96VSqknzvwBlVK9gYeB6VrrH7XWlVrrMq31Qq31y/UO7ayUWmUpZ4tSKrbeNc4X41+UUp8rpT62nLtPKTW83v6LlFI7LPuWKKU+q9/sqZSapJTaafksNymlBjXxbybEBZHEJpzFXcBCy3KNUiqs3r5/AcOASzBqOX8AzEqpnsAa4C0gBBgC7GxGmTcAFwP9LO+3Wa4RBCwCliilvCz7ngSmA9cB/sA9QBkwH5iulHIBUEoFA+Mt55/pKiBda721ibhuBV4AOgNHgJfq7TtfjGDUdhcDgcBK4G1LXB7Acow/HIKAT4Eba09SSg0FPgAeALoA/wNWduRmYWE/kthEh6eUGgv0BD7XWicCR4HbLPtcMJLIY1rrE1prk9Z6k9a60nLMD1rrT7XW1VrrU1rr5iS2f2it87TW5QBa6wWWa9RorV8DPIE+lmPvA+ZorQ9pwy7LsVuBQoykBUZS+klrndVIeV2ADCviWq613qq1rsFI9ENqdzQRI8BGrfVqrbUJ+AQYbNk+CnAD3rR8VsuA+gl2FvA/rfUWy2c8H6i0nCdEq5LEJpzBDOA7rXWu5f0iTjdHBgNeGMnuTJHn2G6t4/XfKKWeVkodsDR3FgABlvKbKms+cIfl9R0YCaUxp4BwK+LKrPe6DPCzMsbGzvWy9CF2A07ohrOq1//5ewJPWZohCyzXjrScJ0SrksQmOjRLX9k04HKlVKZSKhN4AhislBoM5AIVQGwjpx8/x3aAUsCn3vuujRxT9yVv6av6gyWWzlrrQIyamLKirAXAVEu88RiDQRqzFuhev9+rOayI8XwygAilVP1jI+u9Pg68pLUOrLf4aK0/vZBYhTgfSWyio7sBMGH0cw2xLPHAz8BdWmszRt/Pv5VS3SyDOEZb+n4WAuOVUtOUUm5KqS5Kqdpmu53ATUopH6VUL+DeJuLoBNQAOYCbUuo5jL60Wu8BLyqleluG7A9SSnUB0FqnY/R9fQIsrW3aPJPW+jDwH+BTpdQVSikPpZSXUupWpdRsKz6rpmI8n80Yn/Mjls9qKjCy3v53gQeVUhdbfj5fpdT1SqlOVl5fCKtJYhMd3QzgQ631Ma11Zu2CMejhdksz2tPAHozkkQf8E3DRWh/DGMzxlGX7Tk73Kb0OVAFZGE2FC5uI41vgGyAJSMOoJdZvqvs38DnwHVAEvA9419s/HxjIuZshaz1q+dn+H1CA0bx5I/BVE+dZE+M5aa2rgJswEnwBRpPp1xj9aGitE4D7LbHlYwxamWnNtYVoLiUPGhXC8SmlLsNokuyp28l/WqXUFuAdrfWH9o5FOBepsQnh4JRS7hg3Xb/nyElNKXW5UqqrpSlyBjAIowYoRJtqyxkRhBDNpJSKBxKAXcDddg6nKX0wmlN9gWTgt1pra24/EKJVSVOkEEKIDkWaIoUQQnQoktiEEEJ0KO2ijy04OFhHRUXZOwwhhBAOIjExMVdrHdLYvnaR2KKiokhISLB3GEIIIRyEUirtXPukKVIIIUSHIolNCCFEhyKJTQghRIciiU0IIUSHIolNCCFEhyKJTQghRIciiU0IIUSHIolNCCFEhyKJTQghRIfSLmYeEUII0b7ll1ZxNKeE5JxShvYIpHdYJ5uVJYlNCCFEq6gxmUnLK+NodglHc0pJzikhOddY55dV1x33p+viJbEJIYRwHOVVJo7mlHAk+/RyNKeE1FOlVJtOP+MzpJMnMcG+TBwQTmyILzEhvsQE+9G9s7dN45PEJoQQolG1CSwpq5ikrBIOZxVzOLuE4/ll1D6j2tVF0TPIh9hQP8b3C6NXiB+xoX7EhPji7+Vul7glsQkhhJOrNplJyS3lUGYxSVnFHLSsj+WdTmDuroroYF8Gdg/gNxd1p3eYH71C/ejZxQdPN1f7/gBnkMQmhBBOQmtNZlEFBzOKOZBZxKHMYg5lFnM0p6SuCdHVxUhgA7oFcOPQCOLCOhEX5kfPLr64u7aPgfSS2IQQogMqrzJxKKuY/SeLOJRZxIHMYg5mFFFUUVN3TESgN326duKKPqH07dqJuLBOxIT44uXuWDWw5pLEJoQQ7VxOcSX7M4rYf7LIsi4kJbcUs6UZ0dfDlb7h/kwa3I34rp3oG+5PXFgnArzt0wdma5LYhBCiHckuqmDPiUJjSTfW2cWVdfsjAr3p182f6wd1o1+4P/3C/ene2RsXF2XHqNuWJDYhhHBQp0oq2Z1eyO70QvacKGDPiUKyiowkphTEhvgxplcw/bv5079bAP3C/Qnw6Zi1sOaQxCaEEA6gsLyavScK2ZVewB5LMjtRUA6cTmKXxAYzMCKAgd2NJObrKV/hjZFPRQgh2liNycyhrGJ2HCtg5/ECdhzL52hOad3+nl18GNojkJmXRDGwewADIgLwkyRmNfmkhBDCxnJLKklMy2d7Wj47jhs1svJqEwBdfD0YEhnIDUMiGBwZyKDuAQT6eNg54vZNEpsQQrQis1lzNKeEhLR8ElLz2X4sn5Rcozbm4epCv27+3DoykiGRgVzUozPdO3ujlPMM7GgLNk1sSqlA4D1gAKCBe4BDwGdAFJAKTNNa59syDiGEsJWqGjN7ThSyNSWPbal5JKblU1huTPgb5OvBsJ6duXVEJMOjOjMgIsDhZunoiGxdY3sD+EZr/VullAfgAzwLrNVav6yUmg3MBp6xcRxCCNEqSitr2H4sn20peWxNzWPn8QIqqs0AxIT4MrF/V4ZFdWZ4z85EB/tKbcwObJbYlFIBwGXATACtdRVQpZSaClxhOWw+8BOS2IQQDqqsqobEtHw2Hz3Fr8mn2J1eSI1Z46KgXzd/bhvZk5HRnRkeFUSwn6e9wxXYtsYWDeQAHyqlBgOJwGNAmNY6w3JMJhBmwxiEEKJZKqpNbE/LZ3PyKTYfPcWu9AKqTRpXF8Wg7gHMuiyGi2O6cFGPQDrZafZ6cX62TGxuwEXA/2mttyil3sBodqyjtdZKKd3YyUqpWcAsgB49etgwTCGEMzObNfszith4JJdfjuSyNSWPyhozLgoGdg/knrHRjI7pwoioILlvrJ2w5b9SOpCutd5ief8FRmLLUkqFa60zlFLhQHZjJ2ut5wHzAIYPH95o8hNCiAtxPK+Mnw8bieyXo7kUWJ7u3CesE7df3JMxvbowMjpIamTtlM0Sm9Y6Uyl1XCnVR2t9CLgK2G9ZZgAvW9YrbBWDEEKA0bz4a/Ip1iflsD4ph2TLzdDhAV6Mjw9jbK9gLontQqi/l50jFa3B1vXq/wMWWkZEJgN3Ay7A50qpe4E0YJqNYxBCOBmtNUdzSvnpUDbrk3LYkpJHVY0ZTzcXRsV04Y6Le3JZXAixITJqsSOyaWLTWu8Ehjey6ypbliuEcD4V1Sa2pOSx7mA2Px7M5lheGQC9Qv24c1RPLo8LYWR0ULt/1phomvSECiHarczCCn60JLJfjuRSXm3C082FMb2Cuf+yGK7sE0L3zj72DlO0MUlsQoh2Q2tNUlYJ3+/P5Lv9WexOLwSMZ5D9dlh3xvUNZXRsF6mVOTlJbEIIh1ZjMpOYls93+7P4fn9WXRPjkMhAfn9NHyb0C6N3qJ/0lYk6ktiEEA6nqsbMpqO5rNmTyfcHssgrrcLD1YVLenXhgctjGB8fRpiMYBTnIIlNCOEQKqpN/Hw4lzV7M/hhfxZFFTX4eboxrm8oEwd05bK4EHkmmbCK/JYIIeymotrE+qQcVu3OYO2BLEqrTPh7uTGhX1euG9iVMb2Cpb9MNJskNiFEm6o2mfnlSC5f7crgu32ZFFfW0NnHncmDu3HtwHBGx3TBw83F3mGKdkwSmxDC5kxmzdaUPL7afZI1ezLIL6umk6cb1wzoyuTB3bgktgvurpLMROuQxCaEsJmkrGKWbk9nxY6TZBZV4O3uyoR+YUwe3I3L4oLloZvCJiSxCSFaVU5xJSt3nWTZ9nT2nSzC1UVxRVwIf7o+nqviQ/HxkK8dYVvyGyaEaLGKahM/HMhiaWI6Gw7nYjJrBkYE8Pzkfkwe3E0ewCnalCQ2IcQF0Vqz72QRSxKO8+XOkxSWVxMe4MWsy2K4aWgEvcM62TtE4aQksQkhmiWvtIovd5xgSWI6BzKK8HBzYWL/rkwbHsklsV1wcZEZQIR9SWITQjTJbNZsTj7Fwi1pfL8/i2qTZlD3AF68YQBTBnUjwEceyCkchyQ2IcQ55ZdWsXzrEdZu24WpIJ1ozyL+1cufETFd6BZYDOokHFGgFCgXcHEFFzfL4gou7g3fA2gN6HqvMd5rs7GYTadf1y6u7uDmBW6elnW9164epxeXRm4Z0BqqSqGiAMoLTq+rSoxruPvUW7zBwwfcvI1yTZVgqoaayoavXVyNY2vPqV27eRv7aiqguvz0UlO7rrR8Vq7GccrViFm5Gp+fqdJyTgVUl1leW9ZKnVFmvdcoqCyCiiLLuvD066pS43PyqD3P1/Jz+p6Oue5zPWOtXOr9++iG6zqWGnrdXJ3K+OxqyqGq9mcoNdZVlnX3ERDat/V+Uc8giU0IcVrWPvS2DyjKOEJp7nG8K7K4R5VwD4AHxvdZmmVxRMrV+FJ2dTcSndZGIjPX2Dsy+3D3NRJaTaWRVLTJ3hEZrn1FEpsQwsaKTlKz9m+47PqUSjxIM4eTq7rgEzKImOjehEbEgH84dAo3Eoe21LDAUqvSxpemNhtJxFwDpprTr83VYDY3/Ku+rivO8sLFUmtR9WowtYu52qgF1VQ2XFdXgKnKslRbalW1r6uMuLwDwSvw7LVnJ+OYqrJ6taPS07Uk5QKunvUSpSe4WWqG2tywNlW3Ljd+3tqakHu9xc3bOL/2szKbLevaGqrJKKPRWpllwue6WuAZ5ZpN4OUPXgHgWbvuZMRdX01Vw3OrSo3PoLY2WVNR7/MtP12jbvDvpk6v69e4od57jLg9zqwNW2qJPl1a47f2nCSxCeHMKoooXPsvfBLfQZtNfFgzkW+D7uDGMQOYOiRCJh12NN6dW3a+m4exeAe2TjwOSn5rhWgrWlv+Ojc1XFeXQcFxKDgGhceM14XHjXVxhvGXsaunpR/J0sRWW3Oo7cNydTv7tU8QBMWcXgJ71v3lr2uqSP72/xGSOJcAcwErTGNIiH2ISZeN5r7oIHm2mWjXJLEJcT5m8+mmm8oiKM6EohOW5SQUWl4XZxjNYmcOeqi/NOhwPw/vIAiMhJA4iLnC2FY7cKF+s1tN5emmvqoyS3NfbRNgNZTkQGVhvQsrzP7dyXLrhrngGLHmDLbRnwMD5zL+qolMDfRu3c9OCDuRxCbal4oiyNoLp46AbygERUNgD8vIsEZoDaU5kL0fsg9CzgEozT2dHMw1lnX16cRRVWb0tVSVGf0M5+IVAP7dwb8bhA82+g+oN0KwblGn+41q+5HqRsS5GqPPAiKNZBYQCZ5+rfNZaQ3l+ZCXzInkfRzct5PSzMNE6gy8PAJIGf4cw8dPY4RMcSU6GPmNFraldb2O52acY6qGslzI3AOZuyFjt/E6P6XxczqFQ+coYwnsYSSvnIOQfQDK804f590ZOnU73VxX26zn4mtp5nM/PZLM3cfS2W1Ze/hBp67gH2Eks9ZKQDZSaTKz5lA5C36tIiEtGE+3q5k8eCZ3jupJfGTH7mMRzk0Sm2h9ZjOk/QJ7lsD+FUYTXu39Rm5elg5sL6PfCCz3+FQ0XNeOuKvVORrCB8HQ26HrYAjuBaWnID+14ZKywWgi9PQ3hhPHT4bQeGMJiQe/0OYn2nbmZEE5C35N47NtxzlVWkV0sC9zro/nt8O6E+jjYe/whLA5SWyidWgNGbuMZLZ3GRSfNGo+fa+Hzj0tQ4irTg8lNlUaazCSnLv3GWsvY0h22AAI628MZT5TUAxEjjh7u6naGETRwRNYfVprEtPy+fCXVL7Zl4nWmvHxYdw5uidjYoNlmivhVCSxiZbRGja9BTs+gdwko3mv9wQY+DeIu9Zo0mtrZ96704FV1pj4elcGH21KZc+JQvy93LhvbDR3ju5J9852+OyFcACS2ETLpP0C3/8Zuo+EyW9A/BRjmLmwqdySSj7ZnMbCLWnkllTRO9SPl24cwI1DI+R5Z8Lpyf8A0TLZB4z1tPnGgAphU0eyi3nv5xSW7ThBtcnMlX1CuXtMFGN7Bcu9Z0JYSGITLZNz0Bio0Snc3pF0WFprNh89xbs/J7PuUA6ebi7cPKw7946NJibEsUdmCmEPkthEy+QcguA4pxqo0VaqTWZW7c7g3Z+T2XeyiGA/D56cEMcdo3oS5CujG4U4F0lsomVyDkHvq+0dRYdSUW3ii8R03ll/lPT8cnqF+vHyTQO5YWgEXu6u9g5PCIcniU1cuLI8KM2GkD72jqRDKK2sYeGWNN79OYWc4kqG9gjkL5P7M65vqAzXF6IZJLGJC5ebZKwlsbVIQVkVH21K5aNNqRSUVTOmVxfeuHUIo2O6yIAQIS6AJDZx4XIOGWtJbBfkVEkl7/6cwiebUymtMjE+PoyHr4xlaI8WPppECCcniU1cuJxDxsMTA3rYO5J2Jbekknc3JPPJr2mUV5uYNKgbD18ZS9+ujcyuIoRoNkls4sLlHITg3uDiYu9I2oWc4kre/TmZTzanUVljYsrgbjwyrje9QmXIvhCtyaaJTSmVChQDJqBGaz1cKRUEfAZEAanANK11vi3jEDaSmwQ9Rtk7CoeXU1zJ/9YfZcGWNKpqzNwwJIKHx/UiVu5BE8Im2qLGdqXWOrfe+9nAWq31y0qp2Zb3z7RBHKI1VZYYT3kOmWHvSBxWYXk18zYc5YONqVSZzEwd0o1HruwlN1ULYWP2aIqcClxheT0f+AlJbO1P3YjIvvaNwwGVV5n4aFMq76w/SmF5NVMGd+OJCXFEB/vaOzQhnIKtE5sGvlNKaeB/Wut5QJjWOsOyPxMIs3EMwhZqR0QGy4jIWlU1Zj5LOM6baw+TU1zJuL6hPH11H/p1k0EhQrQlWye2sVrrE0qpUOB7pdTB+ju11tqS9M6ilJoFzALo0UNG3Tmc3EPGI2qCou0did2ZzZqvdp/kte+SOJZXxsioIP5z+0WMiJKnHAhhDzZNbFrrE5Z1tlJqOTASyFJKhWutM5RS4UD2Oc6dB8wDGD58eKPJT9hRziHoEutUzz5rzNaUPP62aj+70wvpF+7Ph3eP4Iq4ELmxWgg7slliU0r5Ai5a62LL66uBvwIrgRnAy5b1ClvFIGwo5yB0HWjvKOwm7VQpL685yJq9mXT19+Lf0wZzw5AImfpKCAdgyxpbGLDc8perG7BIa/2NUmob8LlS6l4gDZhmwxiELVRXQH4qDPitvSNpc4Xl1bz942E+2pSKu6sLT02I475LY/D2kMmJhXAUNktsWutkYHAj208BV9mqXNEGTh0BbXaqqbRqTGYWbjnG3B+SKCivZtqwSJ66Oo5Qfy97hyaEOIPMPCKaL7d2jkjnGOq/NSWP51bs5WBmMZfEduFP18fTv1uAvcMSQpyDJDbRfDmHQLlAl172jsSmsosr+MfqgyzfcYKIQG/eueMirunfVQaGCOHgJLGJ5ss5CJ2jwL1jNsPVmMx8vDmN179PorLGzCNX9uLhK3tJP5oQ7YQkNtF8OUkd9sbs+s2Ol/YO5oUp/WUKLCHaGUlsonlMNcbgkbhr7B1Jqyooq+KlVQdYkphOtwAvaXYUoh2TxCaaJz8FzNUdZuCI1pqvdmfw16/2kV9WzQOXx/DYVb3x8ZD/GkK0V/K/VzRPjmVWtJA4+8bRCk4UlPPnL/fy48FsBnUPYP49I2W0oxAdgCQ20Tx1kx+338RmMms+2ZzKq98ewqxhzvXxzLwkCjdXeWCqEB2BJDbRPDmHwL87eHaydyQXJCmrmGeW7mbHsQIuiwvhpRsGEBnkY++whBCtSBKbaJ6cg+1yxhGzWfPhplT++c1B/DzdmHvLEKYO6SaDQ4TogCSxCeuZzZB7GKLG2juSZsksrODpJbvYeCSX8fFhvPybgQT7edo7LCGEjUhiE9YrPAY15e2qxrZqdwbPLt9DVY2Zf9w0kFtHREotTYgOThKbsF5OkrFuB0P9iyuqeX7lPpZtP8HgyEDm3jKE6GBfe4clhGgDktiE9WqH+jv4iMhtqXk88dlOThaU8+hVvfm/cb1wlxGPQjgNSWzCejmHwDcUfILsHUmjzGbNf9cf5bXvDtG9sw9LHryEYT072zssIUQbk8QmrJd7yGH71/JLq3jy852sO5TDpEHhvPybQfh5yq+3EM5I/ucL62ht1NgGOd4Dz3ceL+DhhdvJLq7gr1P7c+eonjJARAgnJolNWKc4EyqLHGpWf6018zel8tLqA4T5e/HFg5cwODLQ3mEJIexMEpuwTt0ckY6R2Iorqpm9dA+r9mQwPj6U124eQoCPu73DEkI4AElswjq5jjPU/3BWMbM+SeRYXhl/vLYv918ag4uLND0KIQyS2IR1cg6CVwD4hdo1jHUHs/m/T3fg5e7Kp/ePYmS0Y47QFELYjyQ2YZ2cQ0ZtzU6DMrTWvPdzCn9fc4B+4f68e9dwugV62yUWIYRjk8QmrJNzCPpca5eiK2tM/Gn5Xr5ITOe6gV35182D5UGgQohzkm8H0bTSU1CWa5f+tZziSh5ckEhiWj6Pj+/No+N6S3+aEOK8JLGJpuVaHi7axiMi958s4v6PEzhVWsl/br+I6waGt2n5Qoj2SSbQ62gKTxiPl2lNaZuMdRsmth/2Z/Gb/27CrDVfPHiJJDUhhNUksXUkGbvg9f6w9F6ormidax74Ctb9HWKugIDI1rlmE5YkHOeBBYnEhfmx4pExDIgIaJNyhRAdgyS2juTIWkDDvmXw8VSjb6yl1/viHoi4CG5Z2CYjIudtOMrvv9jNJbFdWHT/KEI7edm8TCFExyKJrSNJ3Qgh8XDzR3ByB7w/Hk4dvbBrpW2Gxbcbj6i5fQl4+rVqqGfSWvOPNQf4++qDXD8onPdmDMdXJjEWQlwASWwdhakajv0KUWOh/40w4yuoKIT3rjKSVHOc3AmLpkFABNy5HLxt++iXGpOZZ5bu5n/rk7ljVA/evHUonm6uNi1TCNFxSWLrKE7uhOpSI7EB9LgY7vsBfLrAx1NgzxfWXSf7IHxyozHLyF0rbD7TSEW1iYcWbufzhHQeu6o3L04dgKsM5xdCtIAkto4i9Wdj3XPM6W1BMXDv9xAx3BhQsuFV4/Ez55KXAp/cAK7uRlIL6G7TkIsqqpnxwVa+P5DFC1P688SEOHncjBCixaQTo6Oo7V/zC2m43ScI7voSVjwCP/4NtvwP/COMJSAC/LuBf3fw6QxfPwk1FTBzNXSJtWm4heXV3Pn+FvafLGLuLUOYOiTCpuUJIZyHJLaOoLZ/bchtje9384Sb5hnNlCcSoegk5KdA2kajH66WRyeYsQLC+tk03KKKau76YCsHMor4353DuCo+zKblCSGciyS2juDM/rXGKAXDZhhLfZUlRqIrSocuvSCwh01DLamsYeYHW9l3opD/3iFJTQjR+iSxdQSN9a9Zy9MPQuKMxcbKqmq458Nt7Eov5O3pQ5nQT5KaEKL12XzwiFLKVSm1Qyn1teV9tFJqi1LqiFLqM6WUh61j6PDO1b/mQMqrTNzz0TYS0vKYe8sQrpUpsoQQNtJkYlNK/Z9SqiU3Mj0GHKj3/p/A61rrXkA+cG8Lri3q37/moCqqTdz/cQJbU/J4/ZYhTB7czd4hCSE6MGtqbGHANqXU50qpiaoZ47GVUt2B64H3LO8VMA6ovalqPnBD80IWDVjTv2ZHFdUmZn2SyC9Hc3n1t4Nl9KMQwuaaTGxa6zlAb+B9YCZwWCn1d6WUNePB5wJ/AGqnm+8CFGitayzv04FGv+mUUrOUUglKqYScnBwrinJSLelfs7Fqk5mHF25nQ1IO/7xpEL8ZZtv74oQQAqzsY9NaayDTstQAnYEvlFKvnOscpdQkIFtrnXghgWmt52mth2uth4eEOG7fkd05aP+a1prZS/ew9mA2f7thANNGtM2TAYQQoslRkUqpx4C7gFyMJsXfa62rlVIuwGGMGlljxgBTlFLXAV6AP/AGEKiUcrPU2roDJ1r+Yzippu5fs6PXvkti6fZ0Hh/fmztG9bR3OEIIJ2JNjS0IuElrfY3WeonWuhpAa20GJp3rJK31H7XW3bXWUcCtwI9a69uBdcBvLYfNAFa05Adwag7av/bJr2m8ve4I00dG8thVve0djhDCyViT2NYAebVvlFL+SqmLAbTWB8551rk9AzyplDqC0ef2/gVcQ4BD9q99szeT51bsZXx8KC9OHSBzPwoh2pw1N2j/F7io3vuSRradl9b6J+Any+tkYKTVEYpzc7D+tW2peTy6eAeDuwfy1vSLcHOVObaFEG3Pmm8eZRk8AtQ1QcqMJfbmYPevHc4q5r75CXQP9OaDmSPw9pDnqQkh7MOaxJaslHpUKeVuWR4Dkm0dmGhCxi6H6V/LLKxgxgdb8XBzYf49IwnylclkhBD2Y01ie7rcE10AACAASURBVBC4BGP0YjpwMTDLlkEJKzhI/1pxRTUzP9xKYXk1H84cQWSQj13jEUKIJpsUtdbZGKMahSNxgP41k1nz6Kc7OJJdwod3j2BARIDdYhFCiFrW3MfmhTGfY3+M+9EA0FrfY8O4xPnU9q8Nnm7XMP6x+gDrDuXwtxsGcGlvxxjAIoQQ1jRFfgJ0Ba4B1mPcVF1sy6BEEzJ2QVWJXfvXPtt2jPc2pjBjdE+5AVsI4VCsSWy9tNZ/Bkq11vMxJjW+2LZhObnj22D/CqNm1hg7969tST7FnC/3cmnvYP48ybZP2xZCiOayZth+7bdrgVJqAMZ8kaG2C8mJaQ2//ge++zNoE/h3h4sfgIvuAu/A08fZsX/teF4Zv1u4ncjOPrwt96oJIRyQNd9K8yzPY5sDrAT2YzxTTbSmqjJYdj98+yz0uRZuWQBB0fD9n+H1/rDmGchLsev9a8UV1dw7fxs1JjPvzRhOgI97m8cghBBNOW+NzTLRcZHWOh/YAMS0SVTOJj8VFt8BWXth3J9h7JPg4gLxkyFjt1GL2/Y+bJ0HPUbbpX/NZNY8vngnR3NKmX/3SGJC/Nq0fCGEsNZ5a2yWWUbONXu/aA1H1sK8K6DwGNy+BC572khqtcIHwY3vwON7YOwTkL0fXD3avH/tlW8OsvZgNs9P7sfY3sFtWrYQQjSHNX1sPyilngY+A0prN2qt8859imiS1rDxdfjxRaO/7NYFEHSeCrF/OFz1HFz6NJRmt2n/2ke/pPC/DcncMaoHd42OarNyhRDiQliT2G6xrB+ut00jzZIXTmujP23PEhjwG5jyFnj4Wneuhw94RNk0vFpaa/79fRJv/XiECf3CeH5y/zYpVwghWsKamUei2yIQp7LjEyOpXT4brpgNDvhoF5NZM+fLvXy69Ri3DI/kpRsHyAhIIUS7YM3MI3c1tl1r/XHrh+MEijLg2znQcyxc/oxDJrWKahOPLd7Bt/uyePjKWJ6+uo88V00I0W5Y0xQ5ot5rL+AqYDsgia25tIavnwBTFUx5s+EgEQdRVFHN/fMT2JKSx/OT+3H3GKmwCyHaF2uaIv+v/nulVCCw2GYRdWR7l0LSGrj6JegSa+9ozpJdVMGMD7dxJLuYN6cPZcrgbvYOSQghmu1CHhhaCsif8c1Vmgtr/gARw2HU7+wdzVlSc0u584MtnCqp4oOZI2RSYyFEu2VNH9tXGKMgwbjvrR/wuS2D6pBW/x4qimDq2+DiWE+XTsoq5rZ3t2DWmsWzRjGoe2DTJwkhhIOypsb2r3qva4A0rXW6jeLpmA58DfuWwZVzIDTe3tE0sO9kIXe+vxU3F8XiWaPoFdrJ3iEJIUSLWJPYjgEZWusKAKWUt1IqSmudatPIOoryfFj1JIQNhLGP2zuaBnYdL+CuD7bi6+HKovtHERVs5b10QgjhwKwZlrcEMNd7b7JsE9b4do7Rvzb1bXB1nEmDE1LzuOO9Lfh7u/HZA6MlqQkhOgxrEpub1rqq9o3ltYftQupAjqyFnQtgzGPQbYi9o6mz+egp7vpgKyGdPPn8gdFEBvnYOyQhhGg11iS2HKXUlNo3SqmpQK7tQuogqsrgq8cgOM64EdtBbEjKYeaHW4kI9GbxA6MID/C2d0hCCNGqrOljexBYqJR62/I+HWh0NhJRz9EfofA43L4U3L3sHQ0Aaw9k8bsF24kN9WPBvSPp4udp75CEEKLVWXOD9lFglFLKz/K+xOZRdQTJ68DdF6Ivs3ckAPx4MIsHFyQSH+7Px/eMJNBHWpOFEB1Tk02RSqm/K6UCtdYlWusSpVRnpdTf2iK4du3oOogaA272TyDrk3J48JPt9O3qzyf3XixJTQjRoVnTx3at1rqg9o3ladrX2S6kDqDgGOQdhZgr7R0Jm47kMuvjBGJD/fjk3pEEeDvOyEwhhLAFaxKbq1KqrjNGKeUNSOfM+ST/ZKxj7ZvYtqbkce/8BKK6+LLwPqmpCSGcgzWDRxYCa5VSHwIKmAnMt2VQ7d7RdeDXFUL62i2ExLR87v5wK90CvVhw38UE+UpSE0I4B2sGj/xTKbULGI8xZ+S3QE9bB9Zumc2Qsh56TbDbs9Z2HS9gpuU+tUX3jyKkk1SwhRDOw9oHgmVhJLWbgXHAAZtF1N5l7YGyU3Zrhtx7opA7399CoK87i+4fRZi/Y9xqIIQQbeWcNTalVBww3bLkAp8BSmtt/xERjuzoOmMdfXmbF30ku5g7399CJy93Ft03im6BcvO1EML5nK8p8iDwMzBJa30EQCn1RJtE1Z4lr4OQePAPb9Nis4oqmPHBNlxdXFh438UyTZYQwmmdrynyJiADWKeUelcpdRXG4BFxLtXlkLa5zZshiyqqmfHBVgrKqvjo7hEyobEQwqmdM7Fprb/UWt8K9AXWAY8DoUqp/yqlrm7qwkopL6XUVqXULqXUPqXUC5bt0UqpLUqpI0qpz5RSHWe43rFfwVQJMVe0WZGVNSZmfZzAkewS3rlzGAMiAtqsbCGEcERNDh7RWpdqrRdprScD3YEdgDWz+lYC47TWg4EhwESl1Cjgn8DrWuteQD5w7wVH72iS14GLO/Qc0ybFmc2aJz/fxa/Jefzr5sFc2jukTcoVQghHZu2oSMCYdURrPU9rfZUVx+p680q6WxaNMaryC8v2+cANzYnBoR1dB5EjwdPP5kVprXlx1X5W7c7g2ev6csPQCJuXKYQQ7UGzEltzKaVclVI7gWzge+AoUKC1rrEckg50jG/k0lOQubvNptGatyGZD39J5Z4x0dx/aUyblCmEEO2BTROb1tqktR6C0YQ5EqO/zipKqVlKqQSlVEJOTo7NYmw1KT8Z65grbF7U8h3p/GPNQSYNCmfO9fEoO90ILoQQjsimia2WZRLldcBoIFApVXubQXfgxDnOmae1Hq61Hh4S0g76jo6uA88A6DbUpsX8ciSX3y/ZzeiYLrw2bTAuLpLUhBCiPpslNqVUiFIq0PLaG5iAMWPJOuC3lsNmACtsFUOb0dqY+Dj6UnC1ZvrNC5OcU8LvFiQSG+LH/+4ahqebq83KEkKI9sqWNbZwjHvgdgPbgO+11l9jjKh8Uil1BOgCvG/DGNpGXrLxtOyYK2xWRGFZNffNT8DN1YX3ZgzH30sePyOEEI2xWfVCa70bOKtdTmudjNHf1nEc/dFYx46zyeVrTGYe+XQ7x/PLWHjfKJlVRAghzsN27WbOJPknCOgBQbYZnfi3VQf4+XAur/xmECOjg2xShhBCdBRtMnikQzPVQMrPEHuFTR5Ts3BLGh9tSuW+sdFMGxHZ6tcXQoiORhJbS53cAZWFNulf23Q0l+dX7OOKPiH88br4Vr++EEJ0RJLYWip5HaAg+opWvWxqbikPLdxOVLAvb04fiqsM6xdCCKtIYmupo+sgfBD4dmm1SxZVVHPfxwkAvC8jIIUQolkksbVEZQmkb23VZkitNU9+tpPU3FL+e/swenaRR9AIIURzSGJribRfwFzTqvNDvr8xhR8OZDPn+nhGx7ZeLVAIIZyFJLaWSF4Prp7QY1SrXG7n8QJeXnOQa/qHMeOSqFa5phBCOBtJbC2RsgF6XAzu3i2+VGF5NY8s2k6Yvxev/GawTGwshBAXSBLbhSrNhaw9EH1Ziy+ltWb20t1kFlbw1m1DCfCRwSJCCHGhJLFdqNSfjXX05S2+1Ce/prFmbyZ/mNiHi3p0bvH1hBDCmUliu1ApG8DDr8WPqdl7opC/fX2AK/uEcN9YeWCoEEK0lCS2C5WyAXqOAdcLbzYsqazhkUXbCfL14LVpQ+TZakII0QoksV2IwhNw6kiL+te01jy7bA/H8sp4c/pQgnw9WjFAIYRwXpLYLkTKBmPdgsT2ecJxVu46yZMT4mTGfiGEaEWS2C5Eygbw7gxhAy7o9NTcUp5fuY+xvYL53RW9Wjk4IYRwbpLYmktrI7FFXQouzf/4zGbNM0t34+7iwqs3D5LJjYUQopVJYmuuvGQoSoeYCxvmv3BLGltS8pgzKZ7wgJbf2C2EEKIhSWzNVde/1vzEdjyvjH+sOcilvYOZNlweGiqEELYgia25UtZDp3Do0ry+Ma01f1y2BwW8/JtBMmWWEELYiCS25jCbIeVnYzRkMxPT4m3H2Xgklz9eF09EoDRBCiGErUhia46cA1CW2+xmyJMF5by06gCjY7pw28geNgpOCCEESGJrnrr+tUutPqW2CdKsNa/8dpDMLiKEEDYmia05ktdD52gItL7W9UViOuuTcnhmYl8ig3xsGJwQQgiQxGY9U43xxOxmzDaSVVTBi1/vZ2R0EHeO6mnD4IQQQtSSxGatjF1QWWR1YtNa86fle6gymXnlN9IEKYQQbUUSm7VS1htrKxPbt/sy+eFANk9f3YeoYF8bBiaEEKI+SWzWStkAof3AL7TJQyuqTbz49QH6du3EzEuibB+bEEKIOpLYrFFTCcd+tbq2Nm9DMicKynl+cn/cXOUjFkKItiTfutZI3wY15VYlthMF5fznpyNcPzCc0bFd2iA4IYQQ9Ulis0bKBlAuxhOzm/CP1QfQGv54Xd82CEwIIcSZJLFZI2UDhA8B78DzHrYl+RRf787gd1fE0r2z3LMmhBD2IImtKZUlRlNkE82QJrPmL1/tJyLQmwcui22j4IQQQpxJEltTjv0K5pomE9unW49xIKOIZ6+Lx9vDtY2CE0IIcSZJbE1JWQ8u7tBj1DkPKSir4l/fHeLi6CCuG9i1DYMTQghxJpslNqVUpFJqnVJqv1Jqn1LqMcv2IKXU90qpw5Z1Z1vF0GJmMxxYCVFjwOPcN1m//n0SReXV/GVKf3nOmhBC2Jkta2w1wFNa637AKOBhpVQ/YDawVmvdG1hree+Yjm2C/FQYfNs5DzmYWcSCLce4/eKexIf7t11sQgghGmWzxKa1ztBab7e8LgYOABHAVGC+5bD5wA22iqHFdi4Cj04QP7nR3VprXli5Hz9PN56cENfGwQkhhGhMm/SxKaWigKHAFiBMa51h2ZUJhLVFDM1WWQL7voQBN4JH40P3v92XxebkUzx9dRydfT3aOEAhhBCNsXliU0r5AUuBx7XWRfX3aa01oM9x3iylVIJSKiEnJ8fWYZ5t/5dQXQpD7mh0t9ms+ff3h4gJ8WW6PBVbCCEchk0Tm1LKHSOpLdRaL7NszlJKhVv2hwPZjZ2rtZ6ntR6utR4eEhJiyzAbt3MRdOkFkSMb3b1qTwZJWSU8Pj5O5oMUQggHYstRkQp4Hzigtf53vV0rgRmW1zOAFbaK4YLlJRsPFR1yGzQyytFk1ryx9jC9Q/24fmC4HQIUQghxLrasaowB7gTGKaV2WpbrgJeBCUqpw8B4y3vHsnORMTfk4OmN7v5690mOZBu1NVd5gKgQQjgUN1tdWGu9ETjXt/5Vtiq3xcxm2PkpxFwJ/t3O2l1jMvPGD4fp27UT1w6Qm7GFEMLRSOfQmVLWQ1E6DL290d0rd50kObeUx8f3xkVqa0II4XAksZ1p50LwCoA+15+1q8Zk5s21h+kX7s/V/aS2JoQQjkgSW30VhXDgKxjwW3D3Omv38h0nSD1VJrU1IYRwYJLY6tu7DGoqGm2GrDaZefPHwwyI8GdCP8e8p1wIIYQNB4+0SzsXQkg8dLvorF3LtqdzPK+cF2bKRMdCtIW8vDxOnDhBVVWVvUMRbczDw4OIiAiCgoIu6HxJbLVykowHik548ax716pqzLy59giDIwO5sk+onQIUwnnk5eVx/PhxYmNj8fHxwcVFGpechdlspqysjMOHD5OdnU3fvn2bfQ35bam1cyEoVxh0y1m7vkhM50RBOU+M7y21NSHawIkTJ4iNjcXPz0+SmpNxcXHBz8+P3r17U1hYyN69e5t/DRvE1f6YamDXYuh9NXRq2H9WWWPi7R8Pc1GPQC6Ps8PUXkI4oaqqKnx8Gp98XDgHHx8f3Nzc+OGHHzh16lSzzpXEBpC8DkoyjSm0zvD5tuOcLKzgiQlxUlsTog1JTc251f77K6UoKCho3rm2CKjd2bEAfLpA3MQGm81mzXsbU7ioRyBjewXbKTghhHBuNTU1zTpeEltlCRxaY9y75tbwmWobDueQdqqMmWOipbYmhLCaUqrJ5aeffmpxOV27dmXOnDnNOqeiogKlFO+9916Ly3dUMiry6I9gqmz0KdmfbE4j2M+Tif1llhEhhPU2b95c97q8vJxx48YxZ84crr/+9IxG/fr1a3E5q1evJjS0eSO1PT092bx5M7GxsS0u31FJYju0BrwCocfoBpuP55Xx46FsHrmyFx5uUrEVQlhv1KhRda9LSkoAiI2NbbD9XCoqKvDyOnvmo8ZcdNHZ99w2RSllVRztmXN/Y5tqIOkbiLsGXBvm+IVbjuGiFLddLE/HFkLYxjvvvINSiu3bt3PppZfi7e3NW2+9hdaap556igEDBuDr60tkZCQzZswgJyenwflnNkXeeuutjB07ltWrV9O/f3/8/Py4/PLLOXToUN0xjTVFjho1ijvuuIP58+cTExODv78/kydPJjMzs0F5ycnJTJgwAW9vb2JjY1m0aBGTJk1i4sSG4xPszblrbOlboTwP+lzbYHNFtYnPth1jfHwo4QHedgpOCOEsbrnlFh5++GH++te/EhQUhNlsJi8vjzlz5hAeHk5WVhavvvoqEyZMYMeOHeft8z9y5Ahz5szhL3/5C+7u7jz55JPcdtttJCYmnjeGDRs2cOzYMebOnUtRURGPP/44Dz30EMuWLQOMG6cnTZpEVVUVH330EW5ubrzwwgvk5eUxYMCAVv08Wsq5E9vBVeDqAb3GN9i8ancG+WXV3DU6yj5xCSGcytNPP80DDzzQYNuHH35Y99pkMjFs2DB69erFtm3bGDly5DmvlZeXx5YtW+jZsydg1NCmT59OamoqUVFR5zyvtLSUVatW0alTJwDS09OZM2cONTU1uLm5sXz5cg4cOMCuXbsYNGgQYDSF9urVSxKbw9AaDq2GqEvBs1ODXZ/8mkZsiC+XxHaxU3BCiDO98NU+9p8sskvZ/br58/zk/ja7fv1BJbVWrlzJ3//+dw4cOEBR0emfOykp6byJLS4uri6pwelBKunp6edNbKNHj65LarXnmUwmMjMz6d69O9u2bSMqKqouqQFER0czcOBAq37GtuS8fWy5SZCXDH2va7B5T3ohO48XcOeonjLEXwjRJsLCGs549Msvv3DjjTcSGxvLggUL2Lx5Mxs2bACMGtj5BAYGNnjv4eHRKudlZmYSEnL27EuNbbM3562xHVxlrOMa9q99vDkVHw9XbhrWve1jEkKcky1rTPZ25h/RS5cupUePHixcuLBuW/0BIPbQtWtX1q9ff9b2nJwcunZ1rFuinLfGdmg1hA+BgIi6TfmlVazcdZIbhkbg7+Vux+CEEM6svLy8rsZUq36Ss4cRI0aQmprK7t2767alpKSwZ88eO0bVOOdMbMVZkJ4AfRu2a3+RmE5ljZk7R/U8x4lCCGF7EyZMICkpid///vesXbuW559/nsWLF9s1phtvvJG+ffty00038fnnn7Ns2TKmTp1K165dHW5eT8eKpq0kfQPoBsP8zWbNgi1pjIjqTHy4v/1iE0I4vZtuuokXX3yRhQsXMmXKFLZs2cKXX35p15hcXFxYtWoVUVFR3HXXXTz55JM88cQTxMbG4u/vWN+ZSmtt7xiaNHz4cJ2QkNB6F1x0K2Ttg8d31z1UdN2hbO7+cBtvTh/KlMHdWq8sIUSzJSYmMmzYMHuHIZpw6tQpYmJimD17Nn/84x9b/fqJiYls3LiRiRMn0qdPnwb7lFKJWuvhjZ3nfINHqkqNx9RcNKPBk7IXyLyQQghxXm+//TZeXl706tWr7qZxgBkzZtg5soacL7EdXQc1FQ2G+cu8kEII0TQPDw9effVVjh07hqurKxdffDFr166lWzfHauVyvsR2aA14BkDPMXWbFmxJk3khhRCiCbNmzWLWrFn2DqNJzlU9MZuMgSO9J4CrMZy/xmRmSUK6zAsphBAdhHMltvRtUJbboBlyc/Ip8kqruHGo3JAthBAdgXMltoOrwMW9waTHq/dk4OvhyhV9HG9aGCGEEM3nXInt0GqIGgteAYDRDPntvizGxYfh5e5q5+CEEEK0BudJbLmH4dQR6HO6GXJLSh55pVVcP1CG+AshREfhPImtdtLjerONrNqTgY+HK1f0CbVTUEIIIVqb8yS2Q2ug6yAIjAQszZB7MxnXN1SaIYUQrWry5MnnfU7ZI488QmBgIJWVlU1e68iRIyil+Oabb+q2de/endmzZ5/3vJ07d6KUYuPGjdYHDrzzzjusXLnyrO3WlOkonOM+tpIcOL4FLn+mbtPWlDxOlVZx/cBwOwYmhOiIpk+fzu23387+/fvrHvRZy2Qy8cUXX3DTTTfh6el5Qdf/6quvCA4Obo1Qz/LOO+8wfPhwpkyZ0mZltjbnqLFVFELcNRA/qW7Tqj0ZeLtLM6QQovVNnToVHx8fPv3007P2rVu3jqysLKZPn37B1x86dCiRkZEtCbFdlHmhnCOxBfeC2z6DrkbTgMms+Xaf0Qzp7SHNkEKI1uXr68vkyZP57LPPztq3ePFiQkNDGTduHCdOnODuu+8mOjoab29v4uLieP7556murj7v9RtrFnzrrbeIjIzE19eXqVOnkpmZedZ5r776KsOHD8ff35+wsDCmTp3K0aNH6/aPHTuWXbt28f7776OUQinFggULzlnm4sWLGTBgAJ6envTo0YPnnnsOk8lUt/+9995DKcW+ffsYP348vr6+xMfHs2LFiqY/xBawWWJTSn2glMpWSu2tty1IKfW9UuqwZd3ZVuWfz5aUU+SWVHGdNEMKIWxk+vTpHD58mMTExLpt1dXVLFu2jGnTpuHq6kpOTg7BwcHMnTuXb775hqeeeop3332Xxx9/vFllLV26lEcffZSpU6eybNky4uPjuf/++886Lj09nUcffZSVK1cyb948KisrGTNmDMXFxQDMmzeP3r17M2XKFDZv3szmzZuZOHFio2WuXr2a6dOnM3LkSFasWMFDDz3Eyy+/zGOPPdboZ3HDDTewfPlyoqOjueWWW8jIyGjWz9gsWmubLMBlwEXA3nrbXgFmW17PBv5pzbWGDRumW9Oc5Xt0nzmrdWlldateVwjROhISEuwdQotVVlbqwMBA/fTTT9dt++qrrzSgf/nll0bPqa6u1vPnz9fe3t66utr4fjp8+LAG9Jo1a+qOi4iI0M8880zd+6FDh+pJkyY1uNbMmTM1oH/++edGy6qpqdGlpaXax8dHL1y4sG774MGD9b333nvW8WeWOWzYMD1+/PgGx7z00kva1dVVnzx5Umut9bvvvqsBPX/+/LpjsrKytFJKv/vuu43GVV9CQoKeO3euPnjw4Fn7gAR9jpxhs8EjWusNSqmoMzZPBa6wvJ4P/AQ8QxsymTVrLKMhfTycY+yMEB3CmtmQucc+ZXcdCNe+3KxTPDw86p42/corr6CU4rPPPqNnz56MHj0aALPZzOuvv857771HamoqFRUVdeenp6cTFRXVZDlVVVXs2rWLhx56qMH2m266iY8++qjBtk2bNvHcc8+xY8cO8vLy6rYnJSU162errq5m586d/Oc//2mw/ZZbbuFPf/oTv/76KzfeeGPd9quvvrrudWhoKMHBwaSnpzerzOZo6z62MK11bf0zEwg714FKqVlKqQSlVEJOTk6rBbAtNY/ckkpphhRC2Nz06dM5duwYmzdvpqKighUrVnDrrbeiLM+CfO2113jmmWe4+eabWblyJVu3buXNN98EaJDkzic7Oxuz2UxoaMOBcGe+T0lJ4ZprrsHV1ZV58+bxyy+/sG3bNoKCgqwuq36ZJpOJsLCGX+G17+snTYDAwMAG7z08PJpdZnPYrcqitdZKqXM+vltrPQ+YB8YTtFur3NV7MvByd2FcXxkNKUS70swakyO48sorCQsLY/HixWRkZFBcXNxgNOSSJUu49dZb+etf/1q3bffu3c0qIzQ0FBcXF7KzsxtsP/P9mjVrqKys5Msvv8Tb23iSSVVVFQUFBc39sQgNDcXV1fWsMrKysgAICgpq9jVbU1vX2LKUUuEAlnV2E8e3qtpmyCv7SDOkEML2XF1dmTZtGkuWLGHRokXEx8czePDguv3l5eVn3cu2cOHCZpXh4eHBoEGDzhppuGzZsgbvy8vLcXV1xc3t9Hff4sWLMZvNZ12vqdqUu7s7Q4cOZcmSJQ22f/7557i6ujJq1Khm/Qytra0T20qg9hniMwDbjvk8Q0JqHjnF0gwphGg706dPJzMzk+XLl59179qECRNYtGgR//3vf/n222+5/fbbSU1NbXYZzz77LF9//TWPPPII3333HbNnz+aHH35ocMxVV11FVVUVd999N2vXrmXu3Ln8+c9/xt/fv8Fxffv2Zf369Xz33XckJCSc1axY64UXXuD777/nvvvu49tvv+WVV17hL3/5Cw8++CDh4fb9jrXlcP9Pgc1AH6VUulLqXuBlYIJS6jAw3vK+zazek4GnmzRDCiHazujRo4mKikJrfVZie+GFF5g2bRrPPvss06dPx9fXl9dff73ZZdx8883MnTuX5cuXc8MNN7B3717efffdBscMGTKE999/n02bNjFp0iQ+//xzli5dSqdOnRoc99xzzxEXF8fNN9/MiBEjWL16daNlXnfddSxatIhff/2VyZMn8+abb/KHP/yBN954o9nxtzZljJp0bMOHD9cJCQktuobZrBn1j7UM7RHI/+4c3kqRCSFsITExkWHDhtk7DGFniYmJbNy4kYkTJ9KnT58G+5RSiVrrRr/MnWPmESAhLZ9saYYUQogOz2kS2+o9GXi4uXBV/DnvMBBCCNEBOEViM5s1a/ZmcEVcCH6eMhpSCCE6MqdIbDvTC8gqquT6QdIMKYQQHZ1TVF+GRgby5cNj6B3qeuw68AAABmFJREFUZ+9QhBBC2JhTJDalFEMiA5s+UAjhMMxmMy4uTtGoJBpx5o3jzSG/NUIIh+Ph4UFZWZm9wxB2VFZWdsHJTRKbEMLhREREcOTIEUpKSlr0l7tof8xmMyUlJSQlJdU9LNXVtXkPhHaKpkghRPsSFBREQUEB+/btazC3oXAOZrOZzMxM8vPz0VqfNe1XU+Q3RgjhkKKjo8nOzmb9+vX4+vpKgnMyNTU1lJaWMnbsWEJCQpp17v9v785i7ZriOI5/f7kq2hJD0YiiGk2kzaVE1PRAE1JDkBBDSEQkEhGpxOxFCA88mL2Y+2AMingQTTVIiHmeYmiFptw21JCKoX4e9rp1XHVvuZe9796/T3Jy9l7n5uR//rnr/s/aa9+18psSEY0kiblz5zJ58mSWLVvG2rVrGQ9LAMboSWLSpEnMmDGD2bNnr9+/bmOlsEVEY0miv7+f/v7+ukOJcSQ3j0RERKuksEVERKuksEVERKuksEVERKuMi41GJa0CPhuDt9oWWD0G79NWyc/IkqPhJT/DS35GtrE52sX2Bv8PYFwUtrEi6ZW/23E1kp+NkRwNL/kZXvIzsrHIUS5FRkREq6SwRUREq3StsN1adwANl/yMLDkaXvIzvORnZKPOUafm2CIiov26NmKLiIiW60RhkzRf0oeSPpZ0cd3xNIGkOyUNSHqnp20bSYslfVSet64zxjpJ2knSUknvSXpX0oLSnhwBkjaT9JKkN0t+Li/tu0p6sfS1ByRtWnesdZPUJ+l1SU+U8+SokLRc0tuS3pD0SmkbdR9rfWGT1AfcAhwOzAJOljSr3qga4W5g/pC2i4EltmcCS8p5V/0KnGd7FrAfcHb5vUmOKj8B82zvCcwB5kvaD7gauM72bsA3wBk1xtgUC4D3e86Toz87xPacnlv8R93HWl/YgH2Bj21/avtn4H7gmJpjqp3tZ4GvhzQfAywsxwuBY//XoBrE9krbr5Xj76n+MO1IcgSAKz+U0wnlYWAe8FBp72x+BkmaBhwJ3F7ORXI0klH3sS4Uth2Bz3vOvyht8VdTba8sx18CU+sMpikkTQf2Al4kOVqvXGJ7AxgAFgOfAGts/1p+JH0NrgcuBH4r51NIjnoZeErSq5LOLG2j7mPZjy02yLYldf6WWUmbAw8D59r+rnfDw67nyPY6YI6krYBFwO41h9Qoko4CBmy/KunguuNpqINsr5C0PbBY0ge9L/7bPtaFEdsKYKee82mlLf7qK0k7AJTngZrjqZWkCVRF7R7bj5Tm5GgI22uApcD+wFaSBr8wd72vHQgcLWk51RTIPOAGkqP1bK8ozwNUX472ZQz6WBcK28vAzHIn0qbAScDjNcfUVI8Dp5Xj04DHaoylVmUu5A7gfdvX9ryUHAGStisjNSRNBA6lmodcChxffqyz+QGwfYntabanU/3dedr2KSRHAEiaLGmLwWPgMOAdxqCPdeIftCUdQXWtuw+40/ZVNYdUO0n3AQdTraT9FXAZ8CjwILAz1W4KJ9geeoNJJ0g6CHgOeJs/5kcupZpn63yOJO1BNbHfR/UF+UHbV0iaQTU62QZ4HTjV9k/1RdoM5VLk+baPSo4qJQ+LyukmwL22r5I0hVH2sU4UtoiI6I4uXIqMiIgOSWGLiIhWSWGLiIhWSWGLiIhWSWGLiIhWSWGLqIGkdWVF88HHmC2mLGl6764NEV2TJbUi6vGj7Tl1BxHRRhmxRTRI2Z/qmrJH1UuSdivt0yU9LektSUsk7Vzap0paVPZFe1PSAeWt+iTdVvZKe6qsDhLRCSlsEfWYOORS5Ik9r31rux+4mWrFHICbgIW29wDuAW4s7TcCz5R90fYG3i3tM4FbbM8G1gDH/cefJ6IxsvJIRA0k/WB78w20L6fawPPTsgjzl7anSFoN7GD7l9K+0va2klYB03qXZCrb7CwuGzUi6SJggu0r//tPFlG/jNgimsd/c/xP9K49uI7Mp0eHpLBFNM+JPc8vlOPnqVaIBziFaoFmgCXAWbB+488t/68gI5oq3+Ii6jGx7D496Enbg7f8by3pLapR18ml7RzgLkkXAKuA00v7AuBWSWdQjczOAlYS0WGZY4tokDLHto/t1XXHEjFe5VJkRES0SkZsERHRKhmxRUREq6SwRUREq6SwRUREq6SwRUREq6SwRUREq6SwRUREq/wOZDofD5v9PewAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 504x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b4_5PYFG65iY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}