CNN_mnist.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0W3t1OdRFD1r"
      },
      "outputs": [],
      "source": [
        "'''\n",
        "Instructions: To run, upload notebook to Google Colab and change runtime type to T4 GPU\n",
        "\n",
        "Credit: https://github.com/rasbt/machine-learning-book/blob/main/ch14/ch14_part1.ipynb\n",
        "'''"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "x-NeYPCPDMyO"
      },
      "outputs": [],
      "source": [
        "import time\n",
        "import torchvision\n",
        "from torchvision import datasets\n",
        "from torchvision import transforms\n",
        "from torch.utils.data import DataLoader\n",
        "import torch.nn.functional as F\n",
        "import torch\n",
        "\n",
        "import sys\n",
        "import numpy as np\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import pandas as pd\n",
        "import sklearn\n",
        "import sklearn.model_selection\n",
        "from torch.nn.functional import one_hot\n",
        "from torch.utils.data import DataLoader, TensorDataset\n",
        "import torchvision\n",
        "from torchvision import transforms\n",
        "\n",
        "image_path = './'\n",
        "transform = transforms.Compose([transforms.ToTensor()])\n",
        "\n",
        "mnist_dataset = torchvision.datasets.MNIST(root=image_path,\n",
        "                                           train=True,\n",
        "                                           transform=transform,\n",
        "                                           download=True)\n",
        "\n",
        "from torch.utils.data import Subset\n",
        "mnist_valid_dataset = Subset(mnist_dataset, torch.arange(10000))\n",
        "mnist_train_dataset = Subset(mnist_dataset, torch.arange(10000, len(mnist_dataset)))\n",
        "mnist_test_dataset = torchvision.datasets.MNIST(root=image_path,\n",
        "                                           train=False,\n",
        "                                           transform=transform,\n",
        "                                           download=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jrM5_LUHEdl7"
      },
      "outputs": [],
      "source": [
        "batch_size = 64\n",
        "torch.manual_seed(1)\n",
        "train_dl = DataLoader(mnist_train_dataset, batch_size, shuffle=True)\n",
        "valid_dl = DataLoader(mnist_valid_dataset, batch_size, shuffle=False)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CodyTLjvEkhd",
        "outputId": "fed11729-c756-4d3d-c629-4497368816de"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "torch.Size([4, 64, 7, 7])"
            ]
          },
          "execution_count": 6,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model = nn.Sequential()\n",
        "model.add_module('conv1', nn.Conv2d(in_channels=1, out_channels=32, kernel_size=5, padding=2))\n",
        "model.add_module('relu1', nn.ReLU())\n",
        "model.add_module('pool1', nn.MaxPool2d(kernel_size=2))\n",
        "model.add_module('conv2', nn.Conv2d(in_channels=32, out_channels=64, kernel_size=5, padding=2))\n",
        "model.add_module('relu2', nn.ReLU())\n",
        "model.add_module('pool2', nn.MaxPool2d(kernel_size=2))\n",
        "\n",
        "x = torch.ones((4, 1, 28, 28))\n",
        "model(x).shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7DOimaiXEnUR",
        "outputId": "0e49cff2-eaf0-41be-9cc6-df1aa69eae70"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "torch.Size([4, 3136])"
            ]
          },
          "execution_count": 7,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model.add_module('flatten', nn.Flatten())\n",
        "\n",
        "x = torch.ones((4, 1, 28, 28))\n",
        "model(x).shape"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kcEv2y5XEp-Z"
      },
      "outputs": [],
      "source": [
        "model.add_module('fc1', nn.Linear(3136, 1024))\n",
        "model.add_module('relu3', nn.ReLU())\n",
        "model.add_module('dropout', nn.Dropout(p=0.5))\n",
        "\n",
        "model.add_module('fc2', nn.Linear(1024, 10))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KT7mGrtXEsTe"
      },
      "outputs": [],
      "source": [
        "device = torch.device(\"cuda:0\")\n",
        "# device = torch.device(\"cpu\")\n",
        "\n",
        "model = model.to(device)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_3VgUiH8EuFV",
        "outputId": "80fa1e32-1856-4c83-fca7-228912a28bc6"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1 accuracy: 0.9501 val_accuracy: 0.9810\n",
            "Epoch 2 accuracy: 0.9837 val_accuracy: 0.9860\n",
            "Epoch 3 accuracy: 0.9892 val_accuracy: 0.9852\n",
            "Epoch 4 accuracy: 0.9919 val_accuracy: 0.9891\n",
            "Epoch 5 accuracy: 0.9930 val_accuracy: 0.9905\n",
            "Epoch 6 accuracy: 0.9946 val_accuracy: 0.9894\n",
            "Epoch 7 accuracy: 0.9954 val_accuracy: 0.9895\n",
            "Epoch 8 accuracy: 0.9965 val_accuracy: 0.9899\n",
            "Epoch 9 accuracy: 0.9962 val_accuracy: 0.9902\n",
            "Epoch 10 accuracy: 0.9965 val_accuracy: 0.9908\n",
            "Epoch 11 accuracy: 0.9969 val_accuracy: 0.9907\n",
            "Epoch 12 accuracy: 0.9972 val_accuracy: 0.9906\n",
            "Epoch 13 accuracy: 0.9980 val_accuracy: 0.9888\n",
            "Epoch 14 accuracy: 0.9979 val_accuracy: 0.9888\n",
            "Epoch 15 accuracy: 0.9982 val_accuracy: 0.9904\n",
            "Epoch 16 accuracy: 0.9979 val_accuracy: 0.9898\n",
            "Epoch 17 accuracy: 0.9977 val_accuracy: 0.9918\n",
            "Epoch 18 accuracy: 0.9980 val_accuracy: 0.9899\n",
            "Epoch 19 accuracy: 0.9985 val_accuracy: 0.9910\n",
            "Epoch 20 accuracy: 0.9981 val_accuracy: 0.9915\n"
          ]
        }
      ],
      "source": [
        "loss_fn = nn.CrossEntropyLoss()\n",
        "optimizer = torch.optim.Adam(model.parameters(), lr=0.001)\n",
        "\n",
        "def train(model, num_epochs, train_dl, valid_dl):\n",
        "    loss_hist_train = [0] * num_epochs\n",
        "    accuracy_hist_train = [0] * num_epochs\n",
        "    loss_hist_valid = [0] * num_epochs\n",
        "    accuracy_hist_valid = [0] * num_epochs\n",
        "    for epoch in range(num_epochs):\n",
        "        model.train()\n",
        "        for x_batch, y_batch in train_dl:\n",
        "            x_batch = x_batch.to(device)\n",
        "            y_batch = y_batch.to(device)\n",
        "            pred = model(x_batch)\n",
        "            loss = loss_fn(pred, y_batch)\n",
        "            loss.backward()\n",
        "            optimizer.step()\n",
        "            optimizer.zero_grad()\n",
        "            loss_hist_train[epoch] += loss.item()*y_batch.size(0)\n",
        "            is_correct = (torch.argmax(pred, dim=1) == y_batch).float()\n",
        "            accuracy_hist_train[epoch] += is_correct.sum().cpu()\n",
        "\n",
        "        loss_hist_train[epoch] /= len(train_dl.dataset)\n",
        "        accuracy_hist_train[epoch] /= len(train_dl.dataset)\n",
        "\n",
        "        model.eval()\n",
        "        with torch.no_grad():\n",
        "            for x_batch, y_batch in valid_dl:\n",
        "                x_batch = x_batch.to(device)\n",
        "                y_batch = y_batch.to(device)\n",
        "                pred = model(x_batch)\n",
        "                loss = loss_fn(pred, y_batch)\n",
        "                loss_hist_valid[epoch] += loss.item()*y_batch.size(0)\n",
        "                is_correct = (torch.argmax(pred, dim=1) == y_batch).float()\n",
        "                accuracy_hist_valid[epoch] += is_correct.sum().cpu()\n",
        "\n",
        "        loss_hist_valid[epoch] /= len(valid_dl.dataset)\n",
        "        accuracy_hist_valid[epoch] /= len(valid_dl.dataset)\n",
        "\n",
        "        print(f'Epoch {epoch+1} accuracy: {accuracy_hist_train[epoch]:.4f} val_accuracy: {accuracy_hist_valid[epoch]:.4f}')\n",
        "    return loss_hist_train, loss_hist_valid, accuracy_hist_train, accuracy_hist_valid\n",
        "\n",
        "torch.manual_seed(1)\n",
        "num_epochs = 20\n",
        "hist = train(model, num_epochs, train_dl, valid_dl)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 394
        },
        "id": "4acJbI-lE0cS",
        "outputId": "c6c67372-1652-481e-da6e-cfde1bdc25b0"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x400 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "\n",
        "x_arr = np.arange(len(hist[0])) + 1\n",
        "\n",
        "fig = plt.figure(figsize=(12, 4))\n",
        "ax = fig.add_subplot(1, 2, 1)\n",
        "ax.plot(x_arr, hist[0], '-o', label='Train loss')\n",
        "ax.plot(x_arr, hist[1], '--<', label='Validation loss')\n",
        "ax.set_xlabel('Epoch', size=15)\n",
        "ax.set_ylabel('Loss', size=15)\n",
        "ax.legend(fontsize=15)\n",
        "ax = fig.add_subplot(1, 2, 2)\n",
        "ax.plot(x_arr, hist[2], '-o', label='Train acc.')\n",
        "ax.plot(x_arr, hist[3], '--<', label='Validation acc.')\n",
        "ax.legend(fontsize=15)\n",
        "ax.set_xlabel('Epoch', size=15)\n",
        "ax.set_ylabel('Accuracy', size=15)\n",
        "\n",
        "#plt.savefig('figures/14_13.png')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r7QJf6UWE2pC",
        "outputId": "9ce395f1-6c00-4573-d44f-99ef865fa669"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Test accuracy: 0.9927\n"
          ]
        }
      ],
      "source": [
        "torch.cuda.synchronize()\n",
        "model_cpu = model.cpu()\n",
        "pred = model(mnist_test_dataset.data.unsqueeze(1) / 255.)\n",
        "is_correct = (torch.argmax(pred, dim=1) == mnist_test_dataset.targets).float()\n",
        "print(f'Test accuracy: {is_correct.mean():.4f}')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        },
        "id": "n_vTNApoE5kW",
        "outputId": "341192af-4e43-4f77-b7d7-b7ef5add2fbf"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA7YAAAFACAYAAACIruC1AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAA3VUlEQVR4nO3de5xN9f7H8c9gMBgjuUTj8jvkllzzE7kdlaYUkkulkm4o3SSd44eEJKSr45Lq9DuOQxx3SRdxlIhcSzdyiVEuTYz7bf/+WL8pa30We8+etfaa7+zX8/HwGN/3fPeaT+csa+Yze6/PTgiFQiEBAAAAAMBQ+YIuAAAAAACAnKCxBQAAAAAYjcYWAAAAAGA0GlsAAAAAgNFobAEAAAAARqOxBQAAAAAYjcYWAAAAAGC0ApFsOnv2rKSnp0tycrIkJCT4XRMMEAqFJDMzU8qXLy/58vn3+xHOPTjF6twT4fyDxrUPQeHahyBx7UNQsnPuRdTYpqenS4UKFTwpDnnLTz/9JKmpqb4dn3MP5+P3uSfC+Yfz49qHoHDtQ5C49iEokZx7ETW2ycnJvx+wePHiOa8Mxjt06JBUqFDh93PDL5x7cIrVuSfC+QeNax+CwrUPQeLah6Bk59yLqLHNeilA8eLFOclg4/fLRDj3cD6xeIkS5x/Oh2sfgsK1D0Hi2oegRHLuMTwKAAAAAGA0GlsAAAAAgNFobAEAAAAARqOxBQAAAAAYjcYWAAAAAGA0GlsAAAAAgNFobAEAAAAARqOxBQAAAAAYjcYWAAAAAGA0GlsAAAAAgNFobAEAAAAARisQdAFAXjZmzBiVHTt2TGUbN25U2cyZM8Mev3fv3ipr0qSJyu66666wxwIAAABMxTO2AAAAAACj0dgCAAAAAIxGYwsAAAAAMBqNLQAAAADAaAyPAjzUtWtX23rGjBlRHyshISHsngkTJqjso48+UlnLli1t64oVK0ZdF3Ah33//vcqqV69uW7/66qtqzyOPPOJbTci9jhw5orKnnnrKtna7zl155ZUqc7veVqpUKQfVAQBMwjO2AAAAAACj0dgCAAAAAIxGYwsAAAAAMBqNLQAAAADAaAyPAqLkHBQlEv2wqBo1aqgsLS3Ntv7xxx/Vnnnz5qlsy5YtKpsyZYptPWDAgOyWCERk3bp1KsuXz/471EsvvTRW5SCXS09PV9kbb7xhW+fPn1/tWbNmjcrmz5+vsj59+uSgOphq7dq1KuvYsaPKtm/fHoNqLuyDDz5QWc2aNVVWoUKFWJQDA7ld+9q1a2dbv/baa2pP7969VeZ2vTUJz9gCAAAAAIxGYwsAAAAAMBqNLQAAAADAaNxjC0TA7X6u2bNnh31c7dq1VeZ2X2ypUqVUVqxYMdv65MmTak/jxo1VtmHDBpUdOHDggnUCXlm/fr3KnOey271uyPv27dunsu7duwdQCfK6xYsXq+zEiRMBVBKe288Eb731lsqmTZsWi3KQy7n9POd2r6zTI488orL77rtPZUlJSdEVlkvwjC0AAAAAwGg0tgAAAAAAo9HYAgAAAACMRmMLAAAAADBarh0eNXPmTJU537RdRKR8+fIqK1y4sG3drVs3teeSSy5RWdWqVbNTIuLInj17VBYKhVTmHBblNsCiXLlyUdUwZswYlX3zzTcRPfamm26K6msCF7Jp0yaVub0J/N133x2LcpCLvPrqqyqbM2eOylavXu3Z11y+fLnKnNfpunXrqj0tWrTwrAYE4/Tp07b1e++9F1Al2XfllVeqbOzYsSo7cuSIbV20aFHfakLu9Z///Edlu3fvDvu422+/XWXOfikv4BlbAAAAAIDRaGwBAAAAAEajsQUAAAAAGI3GFgAAAABgtFw7POqpp55S2fbt26M61oQJE1RWvHhxldWqVSuq4/utQoUKKuvfv79t7TZ8AN65+eabVbZlyxaVJScn29YlS5b0rIbp06er7OTJk54dH8iu7777TmXOASciIl27do1FOchFHn/8cZXlz5/f1685a9assFnFihXVnnfffVdlDRs29K4w+O6TTz6xrVesWKH2PP3007EqJ1t+/fVXlX399dcqO3r0qG3N8Ki878SJEyobPnx4VMe66667VJaQkBDVsXIznrEFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGy7XDoyZPnqyyDRs2qMxt4NPmzZtt63Xr1qk9S5cuVdnKlStV5hw0sXPnTrUnUomJibZ1qVKl1J49e/ZEVJdzoBTDo2KvUqVKvh5/9OjRtvX3338f0eMaN24cUQbk1KhRo1RWuXJllXF9yvtuvPFG2zoUCqk9Z86c8ezruX3/dBums2PHDtt627Ztak+jRo1Udvbs2RxUBz9t2rRJZbfddpttXbVqVbVnwIABvtWUE/PmzQu6BORSGzduVNnatWsjemyBAvYW74YbbvCkptyOZ2wBAAAAAEajsQUAAAAAGI3GFgAAAABgtFx7j+0111wTUeYmLS0t7J6MjAyVud2L67w3bPXq1RHV4KZQoUK2dfXq1dWeGjVqqMztzburVKkSdR3IfRYsWKCywYMH29Zub9RdtmxZlY0cOVJlRYoUyUF1gMj27dtV5nY9dLuuud37CHMtW7ZMZd9++61tnZCQoPbkz58/qq/Xq1cvlbVp00ZlKSkpKluyZIlt/dxzz0X0NcePH6+y3r17R/RY+Mvt/8OjR4/a1lOmTFF7ihUr5ltNkXL7ec7t35Pbvx/En1mzZkX92Ouuu87DSszBM7YAAAAAAKPR2AIAAAAAjEZjCwAAAAAwGo0tAAAAAMBouXZ4lN8uuugilbVu3Trs4yIdYBWJf//73ypzG2pVp04dlTnfjBxmW7NmjcrchkU5de3aVWUtW7b0pCbgXG4DTtyULl3a50oQS25Dw9y+/+zfvz+q41esWFFlnTp1sq2feeYZtSfSgXiVKlWyrSdOnKj2uNXev39/lR0/fty27tOnj9qTmJgYUV2IzMyZM1X23nvvqaxq1aq2daNGjXyrKSeGDx+uMrdBUa1atVJZiRIlfKgIuVmk33cLFiyoshEjRnhdjhF4xhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABgtbodHBWHv3r229UMPPaT2hEIhlQ0ePFhlJUuW9K4wxFSHDh1Utnjx4rCP6969u8rcBlEAfti4cWNE+9yG7sBcp06dUlm0g6JatGihsunTp6usVKlSUR3fjXN41IABA9Sevn37quzIkSMqc57b7dq1U3uqVKmS3RJxATNmzFCZ2/83vXv3jkU52eYcvjZ16lS1p0AB/aP4wIEDVcZgsrxvxYoVtvXnn38e0ePchunVq1fPi5KMwzO2AAAAAACj0dgCAAAAAIxGYwsAAAAAMBqNLQAAAADAaAyPiqFx48bZ1s5hUiIiJUqUUFn16tX9Kgk+27Nnj8qcwwFERE6cOKGy0qVL29ZuwySKFSuWg+qA83MOrXj77bfVnvr166vsuuuu860mmKNRo0YqczuHvBwUFQm3gU///Oc/VfbFF1/Eohyc4+DBgypbuXJlRI91G8aZG0yaNMm23rdvn9pTq1YtlbVu3dq3mpB7rV69OqrH5dbhaUHgGVsAAAAAgNFobAEAAAAARqOxBQAAAAAYjXtsffLpp5+qbOTIkWEfN3fuXJXVrl3bk5oQex07dlTZ/v37I3pst27dbOsqVap4UhMQiY8//ti2zsjIUHvS0tJUVrhwYd9qQu5w5syZsHtWrVoVg0qyLxQKqezs2bMR7XP+dz/zzDNqz5QpU3JQXXxzmzWxa9culd1+++2xKMcTW7duDbuHn/GQJZJ7bN1m8eTWe8yDwDO2AAAAAACj0dgCAAAAAIxGYwsAAAAAMBqNLQAAAADAaAyP8sl7772nspMnT9rW1157rdrTpEkT32qC/+bNm2dbr1u3LqLHtWrVSmVDhw71oiQgKhs2bAi7p3PnzjGoBEGaMGGCyvLnzx9AJd6YP3++ytyu0wkJCSpz/nc/++yz3hUGSU5OVlm9evVUtmnTJpX9+uuvtnXJkiU9qytSe/fuVdmMGTPCPu7qq6/2oxzkcm5DZqdOnRr2cSkpKSpLTU31pKa8gGdsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0Rge5YFjx46p7P3331dZoUKFbGu3wROJiYneFQZfHThwQGUjRoywrZ0Dw87HbUBGsWLFoqoLyK6ff/5ZZcuXL7eta9SoofbccsstvtWE3GHBggVBlxCxffv2qWzz5s22tfManR2lSpWyrfl+7a2kpCSVVa1aVWUzZ85UWdu2bW3rvn37elbXV199pbKtW7eqbMeOHSpzG0LmlC8fzzHFI7efIUOhUNjHXXfddX6Uk2fwrwkAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiN4VEeGD16tMrWrVunshtuuMG2btq0qW81wX8vvviiyr744ouwj+vQoYPKhg4d6kVJQFT+/ve/q+yXX36xrZ3XLyC3ee6551Q2bty4qI5VuXJllb3zzju2dcWKFaM6NiI3ZMgQlbkN2HEOObvttts8q6F06dIqcxsKtX///qiO36NHj6geB7PNmDEj7J4SJUqo7MEHH/ShmryDZ2wBAAAAAEajsQUAAAAAGI3GFgAAAABgNO6xzSa3N6sfNmyYylJSUlQ2aNAgX2pCMMaOHRvV49zu+SpWrFhOywGitmPHjrB7LrroohhUAkTmxhtvVNm3337r2fFr1aqlsubNm3t2fESmZs2aKnv33XdV5pxrsnXrVs9q6NSpU0T7unfvrrIpU6aEfVxSUlK2a4JZdu3apbKpU6eGfVxqaqrKGjVq5ElNeRXP2AIAAAAAjEZjCwAAAAAwGo0tAAAAAMBoNLYAAAAAAKMxPCqMAwcO2NaPPvqo2nP69GmVuQ22aNKkiXeFwVjOc0pEJDEx0bPjuw0ucx7/1KlTas/BgwcjOn5GRoZt/dJLL2WjOrv8+fPb1i+88ILaU6RIkaiPj8jMnz8/7J6bbropBpUgtwmFQio7c+ZM2MctWrQoouM/8MADKktPT4+qroSEhIi+ZiTcBkUi96pfv/4F17Hwpz/9KarHbdq0SWVXXHFFTstBLrJixQqVuV3DnNq3b+9HOXkaz9gCAAAAAIxGYwsAAAAAMBqNLQAAAADAaDS2AAAAAACjMTzqHG4DMdLS0mzrbdu2qT1Vq1ZV2bBhw7wrDHlKnTp1fD1+ly5dVFauXDnb+pdfflF7pk2b5ltNkSpbtqzKBg4cGEAledfy5ctV5nY+ACIivXv3Vln//v3DPq5t27Yqcw6LO59I9rl9v470+E69evWK6nHAudyGAUUyIIhBUXmf29BQN6VKlbKtH3/8cR+qydt4xhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiN4VHn2Lp1q8rWrFkT9nFjx45VWZUqVTypCbnXjTfeqLI5c+bEvhCHd99917NjJSYmqixfvvC/D2vXrp3KrrzyyrCPa9asWWSFIWqzZ89W2enTp1VWv35927ply5a+1YTcq2PHjiobNWqUyvbv3x+Lci7IOXhFRKRmzZq29RtvvKH2OIfrAdFISEiIKEP8Wbx4cUT7KlSoYFunpKT4UU6exjO2AAAAAACj0dgCAAAAAIxGYwsAAAAAMFrc3mO7Y8cOlbVp0ybs48aMGaOym266yZOaYJZZs2apzHnv2cmTJ6M+/ubNm23radOmRX2s++67z7auVKlSRI+79dZbVea8Zw2519GjR1W2aNGiiB7buXNn2zp//vye1ASzuF0rpk+frjLnfIGXX37Zp4rO73/+539U1qdPn5jXgfh0/PjxsHuSkpJiUAmCdOrUKZVt2bIloscWLlzYtnabc4IL4xlbAAAAAIDRaGwBAAAAAEajsQUAAAAAGI3GFgAAAABgtLgdHjVx4kSVuQ2UcmrZsqXKeANuZOnfv79vx546dapvx0be5DZ4okSJEipr3769yh577DE/SkIe0KJFi7CZ2zDGSZMmqWz+/Pkqu/nmm23rnj17qj2hUEhltWrV0sUCMfL222+rzHm9HTx4cIyqQVDy5dPPGTZq1EhlX3/9tcouu+wyX2qKJzxjCwAAAAAwGo1tHFq6VCQhIfyfoUODrhR5zdGjInPmiNx3n0j16iKFC4sULSpSt651vh0+HHSFyMu+/FJk5EiRjh1FUlP/uNYBsXLsmMjgwSLVqlnXv/LlRe69V2T37qArQzw5cECkTBnr+le1atDVIC+L9ffduH0pcjy75BKR7t3dP3fmjMiUKdbfmzePXU2ID1OnijzwgPX3mjVF2rUTOXRIZMUKkWeeEfnXv0SWLbO+4QJeGzZMZO7coKtAvDp+XKR1a5GVK0XKlRNp315k+3aRt98WWbDAyv/0p6CrRDx48kmR/fuDrgLxINbfd2ls41CNGiJ//7v75xYtshrbChVEWrWKZVWIB4mJIg8+KPL441Zjm2XPHpG2bUXWrbM+x+3E8EOTJiJ16og0amT9qVxZ5MSJoKtCvBg+3GpemzQR+eADkWLFrHzsWKvRuPde6xVVgJ8+/ljknXes78Uut70Dnor19924aGyXL1+ustdffz2ASnK/rGdru3XjJXrwXvfu7q8WKFdOZNw4kaZNRWbNEjl5UqRgwdjXl9e4DY/6/PPPA6gkd3j66aAriB9paWkRZfHi5EmRrB87xo37o6kVEenb12o0li2zXrbXsGEwNSJn3AYEPfHEE7Z169atY1WOq2PHRHr2FKlVS6RfPxpbP+TPn19lzz33nMrcBs82aNDAl5qCFOvvu9xji98dOfLHywXuuivYWhB/6ta1Pp44Yd3/AwB5xWefiRw8KFKlikj9+vrznTpZH12GRAOeefZZkR9/FJkwwXoFFZDX0Njid7NmWc1t/frWb/OAWPrxR+tjYqJIyZLB1gIAXtqwwfp4vidksvKNG2NTD+LPxo0iL74o0qMHM1SQd9HY4ndZL0Pm2VoE4ZVXrI9paSKFCgVbCwB4aedO62Nqqvvns/IdO2JTD+LL2bMi998vUqKEyKhRQVcD+Ccu7rFFeHv2WAMF8ucXuf32oKtBvHnvPZE337SerR02LOhqAMBbWW9lVqSI++eLFrU+ZmbGph7El9deE1m92prAffHFQVcD+CcuGttPP/1UZZkRfveo6niDr2LnTnzIQ/71L+utftLSrLcDAmLl229F7rxTJBQSGT36j3ttAQAwxfxceoP0zp0iAweKtGwpcs89QVcTn8qXL6+yt956K4BK8j5eigwR4WXICMbu3dYvUzIyrMmgjz0WdEUA4L2s34kfPer++SNHrI/JybGpB/Hj4YetqdwTJgRdCeC/uHjGFhf2zTfW+4cWKybSoUPQ1SBe/PqrSJs21j1lPXqIjBkTdEUA4I+KFa2Pu3a5fz4rr1QpNvUgfixYYN1b26uXPT9+3Pq4e7dIq1bW36dN41V7MBuNLeQf/7A+dux4/vt/AC8dPixyww0imzdb590bb/C+yQDyrqxbLNaudf98Vl6nTmzqQXz57TfrfZLdHD/+x+eyml3AVDS256hXr57KPv74Y9u6ZB57H5JQSGTqVOvvvAwZsXDihEj79iJffCFy/fXW/d0u72cOAHnG1VeLpKSIbN0qsn69iPPHjZkzrY833xzrypDXhULu+fbtIv/1X9Z7K2/ZEtOSAN9wj22cW77ceinopZeKtG4ddDXI686csaZuL1livY/erFkiBQsGXRUA+KtgQZE+fay/P/zwH/fUioiMHWu9x2jLliINGwZTHwDkBTxjG+eyhkbdcYdIPn7NAZ+9/rrI7NnW30uVEnnoIfd9Y8ZYnwe8tHCh/e2kTp60Pl511R/ZoEEibdvGti7Eh4EDRT76SGTFCpHLLrN+ubdjh8iqVSKlS4swJBVAXhPr77s0tnHsxIk/Xv50553B1oL4kJHxx9+zGlw3Q4bQ2MJ7+/ZZTYTTudm+fbGrB/GlcGGRTz4Ref556xagOXNESpa03oJl2DCR1NSgKwQAb8X6+y6NbRwrVMiaTAvEypAh1h8gCPfcw/s4IlhJSSJDh1p/gCBVrnz++28Br8T6+25cNLZ//etfI8oAAAAAAObhrkoAAAAAgNFobAEAAAAARqOxBQAAAAAYLaJ7bEP/f3f5oUOHfC0G5sg6F0I+Tx7g3INTrM69c78G5x+ycO1DULj2IUhc+xCU7Jx7ETW2mZmZIiJSoUKFHJSFvCgzM1NSUlJ8Pb4I5x40v8+9rK8hwvkHjWsfgsK1D0Hi2oegRHLuJYQiaH/Pnj0r6enpkpycLAkJCZ4VCHOFQiHJzMyU8uXLS758/r2inXMPTrE690Q4/6Bx7UNQuPYhSFz7EJTsnHsRNbYAAAAAAORWDI8CAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0QpEsuns2bOSnp4uycnJkpCQ4HdNMEAoFJLMzEwpX7685Mvn3+9HOPfgFKtzT4TzDxrXPgSFax+CxLUPQcnOuRdRY5ueni4VKlTwpDjkLT/99JOkpqb6dnzOPZyP3+eeCOcfzo9rH4LCtQ9B4tqHoERy7kXU2CYnJ/9+wOLFi+e8Mhjv0KFDUqFChd/PDb9w7sEpVueeCOcfNK59CArXPgSJax+Ckp1zL6LGNuulAMWLF+ckg43fLxPh3MP5xOIlSpx/OB+ufQgK1z4EiWsfghLJucfwKAAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDQaWwAAAACA0WhsAQAAAABGo7EFAAAAABitQNAFAAAAAIDpTpw4obKmTZuqbN26dSpr166dbT1nzhzP6ooXPGMLAAAAADAajS0AAAAAwGg0tgAAAAAAo9HYAgAAAACMxvAoAICvMjIybOudO3dGfaxKlSrZ1i+99JLaU7t2bZVVq1ZNZXXr1o26DgBmWb58ucrchvp89913tvWCBQvUnoULF6qsbdu2YWto0qSJypo3bx72cci9nMOinnjiCbVn/fr1KktISFBZw4YNPasrXvGMLQAAAADAaDS2AAAAAACj0dgCAAAAAIxGYwsAAAAAMFqeHB61d+9e27pLly5qj9vAgAcffFBllStX9qwuLx08eNC2/s9//qP2pKWlqSwxMdG3mgDEF7ehKvPnz1fZ0qVLbesffvgh6q9ZvXp123r79u1qj3OYx/mcPXs26joA5B6HDh2yrbt166b2fPzxxypLSkpS2alTp2zrzMzMiGpw+zkskq9XtGhRlY0fP15lnTp1iqgOxNarr75qW0+cOFHtueaaa1Q2dOhQlV111VXeFRaneMYWAAAAAGA0GlsAAAAAgNFobAEAAAAARjP+HtuMjAyVXX755ba1835UEZGyZcuqzJT7aUVEGjRoYFvv379f7VmzZo3KLrvsMu8KQ1Sc9wKJiPzlL39R2ddff21bf/TRR2oP90zDC1u3blXZuHHjbOtJkyapPceOHVNZKBTyrjAX3333na/HB2Cep59+2rZ2u//fjds1rGbNmrZ1mTJl1J7ixYtHdHznffwLFy6MqIb77rtPZdWqVbOt69SpE1EN8NeePXvC7rn22mtVxv20/uAZWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDSjhke5DUjq0qWLyg4cOGBbP/zww2rPa6+95l1hPhs+fLjKtm3bZlu7DXZhUFTwpkyZorKBAweqbOfOnWGP5TZ06uKLL46uMOAcu3btUtnLL78c+0IcatSoobLatWsHUAliacuWLba12/f+2bNnq2zp0qUqy5fP/vv7Xr16qT1NmzZVGd8/c6+vvvpKZTNnzgz7uAoVKqjsf//3f1VWtWpV27pEiRJqT7FixcJ+PRE9PGro0KFqz7Bhw1Tm9v1+yJAhtvWbb76p9lx00UUR1QXvHD582LYuWLCg2uM2PAr+4BlbAAAAAIDRaGwBAAAAAEajsQUAAAAAGI3GFgAAAABgNKOGR61du1ZlbsMinAYPHuxDNf5wG4owZswYld1yyy22ddeuXX2rCZFzDuF54okn1B63QSgJCQlhj/3II4+o7PXXX1dZyZIlwx4L5nM7j5wDn5o1a6b2pKWlqcxt2EVKSopt7TYsxTk0Q0Tk+uuvV5lz4FPjxo3Vnvr166ssKSlJZUWLFlUZzLBp0yaVjRs3TmWzZs2yrfft2+dZDStXrlRZYmKiyqpXr64y57+nV155Re1x+7cEb7ldd5zXQ7fvqf3791dZq1atPKvLjXN4mXMAlIjIyZMnVeb2c59zYNq9996r9tx0003ZrBDZkZ6errLJkyfb1m7D6Bo0aOBbTbDjGVsAAAAAgNFobAEAAAAARqOxBQAAAAAYjcYWAAAAAGC0XDs8au/evSr797//HdFj33rrLdu6dOnSntTkNbdBUdddd11Ej+3YsaNtnZyc7ElNyBnnwIcDBw54duxp06apbNGiRSobOHCgypyDpxhwYpYjR46ozO1asWHDBtt6zpw5ER2/SZMmKlu3bp1tXblyZbVn586dKktNTVWZc4AK8p6NGzfa1m5DoaZPn66ygwcPhj222znVvHlzlbmdo6NHj7atGzZsqPasWrVKZW7X7vfee8+2rlu3rtrTq1cvlcFbJ06cCLvnnnvuUVmfPn18qCbnRowYoTK37/fbtm2zrZ1D1kQYHuW34cOHB11C1D7//HOVOQeeno/zWletWjVPavIDP20AAAAAAIxGYwsAAAAAMBqNLQAAAADAaLn2Htsnn3xSZVOmTFGZ25sed+7c2ZeavPbpp5+q7Oeff1ZZjx49VHbnnXf6UhMit2PHDpW9/fbbYR/ndl9W2bJlVfbhhx+GPZbb/Wlub+zerVs32/qSSy4Je2wE4+TJkyq74447VOa8n1ZEZMCAAbb1tddeG3UdbvcrOlWsWDHq48NcPXv2VNns2bNt63379kV0LLdz9IorrrCt3e5BLFy4cETHd95XNn78eLXH7Xvs+vXrVea8bj700ENqz6233qqy3Drnw1SDBg0Ku6dx48YxqMQ/aWlpKnOeuytXroxVOfh/CxcuDLvn/vvvj0Eldr1791aZs9aMjAy15+jRoxEdv3jx4rZ137591Z5I/l3GAs/YAgAAAACMRmMLAAAAADAajS0AAAAAwGg0tgAAAAAAo+Xa4VEJCQkRZZdeeqnKChYs6EtN2XHs2DGVOQdguL2Bvdt/41tvveVdYfCM23CRQ4cO2dYtWrRQe5YtW6ay48ePq2zq1Km29fPPP6/2bNmyRWVuA8jat29vWy9atEjtKVmypMrgv8OHD9vWboNy5s+frzK3gTRPPfWUbV2kSJEcVod44nYdGjVqlMreeOMNlYVCIdu6TJkyao/bgBPnOSsiUrRo0QvWmR0HDhywrU+fPq32PPvssyq7/vrrVbZ9+3bP6kJkfvzxR5Xt3r1bZSVKlLCtnQPITNO6dWuVuQ0+g3/cBiudOnVKZampqbb1PffcE/XXdF6f1q5dq/Z06NBBZW4/9zmvyW4/M7gN73P7mjt37rStJ06cqPbcfffdKqtUqZLK/MYztgAAAAAAo9HYAgAAAACMRmMLAAAAADAajS0AAAAAwGi5dnhUpBYsWKCyNm3a2NbOoQIi7kMsorV06dKIspUrV4Y9VufOnT2oCLFw4sQJlTmHfz3xxBMRHatw4cIqu/fee23rmTNnqj1bt25VmXNggIgeIpQbBqzBMmfOHNt65MiRao/bAIbly5erLCUlxbO6EH/cvm+NHj1aZW7XGOcgx1mzZqk9//3f/x19cQ5nzpxR2U8//aQy50CTtm3bqj0ZGRlR1XDXXXepzO3nDURvypQpKnMbKNWpUyfbumnTpr7VhPgwefJklf3yyy8q69mzZ1THT09PV9mkSZNs62HDhkV0LLdBus7r00MPPaT2OAdfnU+7du1s64ULF6o9e/bsURnDowAAAAAAyCYaWwAAAACA0WhsAQAAAABGo7EFAAAAABgt1w6Peuyxx1S2ZMkSlbndfL1s2TLb2m3Qxdy5c3NQnZ3b8Z1DhNxUqVJFZSNGjPCkJvjvX//6V9g9bjfYd+jQIaqvt2bNmqgeJyJy1VVX2dbFihWL+ljw1ooVK8LuqV+/vsoiHfoAROr06dMqy58/f0SPTUxMtK1XrVql9rgNwPv222/DHjspKUll33zzTURZqVKlbOuff/457Nc7n7Jly9rWAwcOVHuc/zsgZ9y+z7oN6HL7mRHIiXXr1kW077LLLovq+MOHD1fZhAkTbGu3XuKaa65R2dixY1VWu3btqOpyU7VqVc+O5TeesQUAAAAAGI3GFgAAAABgNBpbAAAAAIDRcu09tg0bNlTZpk2bVLZ+/XqVvf/++7b1qFGj1J4yZcqorHv37tmo8A9ub9Jep06dsI9zewNxt/tukTvdfvvtKnPeu7169Wq1x+2eMrdze/bs2bZ1RkaG2uN2r5HbPuebfruds7Vq1VIZ/Od236HTokWLVPbss8+qzPkm6m735gLn43bv1p///GeVffjhhyrbsWOHbf3oo49GXUeBAvYfTdzu/Y1UJPfU5sunf8ffsWNHlb366qu2dbly5aKuC9GrUaOGypo1axZAJcjL3Gb4ROv7779X2bRp08I+7sEHH1TZK6+8orKCBQtGV1iU3Hq0Bg0axLSG8+EZWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYLRcOzzKzUUXXaQyt8EWzuyFF17wrSYRkR9//FFloVBIZfXq1bOtx4wZ41dJiIFrr71WZSkpKbb1xo0b1Z6aNWuqzO1NuJ2uu+46lY0bN05lN910k8qcgwucQ1BE9BuDIzb27dtnW7udCydOnFCZ2/Ao5xu+9+rVS+1p3Lixyn766SeVOd+Q/fLLL1d73Hz99dcqa9KkiW2dmpoa0bEQW0lJSSpzDrETEfntt99UNnLkSNv6s88+U3suvvhilVWsWFFlzvN9w4YNas+qVatUFq2ePXuqbMSIESpzG9YH7xw5ckRlORkcBuTEoUOHVOb2s71b5vTaa6+pzO062q1bN9t6/PjxYY8dC4cPH7atnQP+RGI/wOp8eMYWAAAAAGA0GlsAAAAAgNFobAEAAAAARqOxBQAAAAAYzajhUbnV0KFDVeY2AGbUqFG2denSpX2rCf4rWbKkymbMmGFbd+rUSe05ePCgytyGDzz66KO2tdsQtMKFC6usY8eOKnv++edt68WLF6s9W7duVVmVKlVUBm/169fPtn7xxRejPtaZM2dsa7fhYm6Z38qUKWNbt2rVSu2ZNm1ajKpBTrkNUXIOj/LS3XffrbJIh0cVL17cth47dqzac88996gsf/78kRUHz0yfPl1lW7ZsUVmpUqViUU6g5s2bF3ZPYmJiDCqJX24/x0eaOaWnp0f0OLd9seZWw+TJk23rW2+9NVblZBvP2AIAAAAAjEZjCwAAAAAwGo0tAAAAAMBoNLYAAAAAAKMxPCqbnMOBRETeeecdlTkHVoiIXHzxxb7UhNzj2muvta1nzpyp9kydOlVlbsNYnEPJ3AZFuRk0aJDKvvnmG9t67ty5Yb+eiPu5DW85h+506dJF7enWrZvKTp06pbJdu3bZ1s5hUkHZu3evbe12Ha1du7bKBg4c6FtNyL2cgxZzMlhs/PjxtvUdd9wR9bEAP3z55Zcqmz9/ftjHPffcc36UAx9MmjRJZStWrAibjRgxQu3p2bOnyrzsL9wGkBYpUsS2fvLJJz37el7jGVsAAAAAgNFobAEAAAAARqOxBQAAAAAYjXtss2nRokUR7Wvbtq3KGjRo4HU5yOWc99yeL/NSUlKSyrp27Wpbu91j+8knn6js119/VVnJkiVzUB2c8ufPb1s3atRI7fn+++8jOtbHH39sW7vdhztkyBCVffHFFxEd3yuhUEhlbveZIe+bPHmyyoYPH25bu53Hbtzu07711lujKwzwgdt17sUXX1TZb7/9prJmzZrZ1mlpaZ7VFe/S09NVtmfPHs+O73YP7Nq1a1XWrl0729ptZsrixYtVtmDBApUlJyeH3eO81oqIrFu3TmXOeRdXXXWV2pNb8IwtAAAAAMBoNLYAAAAAAKPR2AIAAAAAjEZjCwAAAAAwGsOjsslteFTRokVV1q9fv1iUA0SkS5cutvW8efPUnmnTpqns9ddfV9ngwYO9Kwyeuuaaa8LuWb9+vcrchkclJiba1j169FB7HnjgAZW99NJLKps6dWrYupD3uZ1nTz75pMoyMzPDHss5GEVEZPz48SorVKhQhNUhaJUrV1ZZ8eLFY1+Ih86cOWNbjxkzRu1x+96bmpqqMudjCxTgR3ivlC9fXmXVqlVT2Y4dO1S2ZMkS27pnz55qT5EiRVRWrlw5la1evdq2dhv4VLNmTZW5DRtzXlvdBvW51eUcFCXiPsQqt+IZWwAAAACA0WhsAQAAAABGo7EFAAAAABiNxhYAAAAAYDTuPA9jwoQJtvXPP/+s9pQtW1ZlDRo08K0mILvy5bP/Dqt///5qz5w5c1Q2ZMgQld122222tduABeRebdq0UdmAAQNUdurUKdt60qRJas8PP/ygsqVLl0ZV16WXXhrV42CO+fPnq+zQoUNhH+c2oNFtAF6zZs2iKwy5QuvWrVXmNtTn4MGDKtu/f79tXapUKe8Kc7Fx40aV/e1vf1PZ2rVrbWvncKDzmTJlisoaN24cYXXwwptvvqmytm3bqmzhwoW2tdv32L59+6rMbXiU06pVq1Q2YsSIiPaFQiHbunr16hEd65ZbbglbV27GM7YAAAAAAKPR2AIAAAAAjEZjCwAAAAAwGo0tAAAAAMBoDI8Kwzk8KiEhQe258cYbIzpWZmambZ2RkaH2VKxYMRvVAdGpV6+eyoYNG6ayfv36qeyvf/2rbe025CIpKSn64uCrmjVrqqxr164qmz59ethjffLJJxF9zQIF7N9q3AZwvPDCCxEdC2Zwfr8TERk1alRUx7rzzjtV1qpVq6iOBfN98803Krv++utt60gG8+SE27Ae5wArN6VLl1bZzTffrLJGjRpFVxg8k5qaqrL3339fZX/+859t688//1zt6dy5c0Rf0znwya3niFSPHj1sa7fr78UXXxz18XMrnrEFAAAAABiNxhYAAAAAYDQaWwAAAACA0bjH1gPO+8dE3O87fOmll2zr2rVrqz3vvPOOd4UB2XD33XerbOLEiSqbNWuWbf3DDz+oPXXq1PGuMHjK7f7nl19+WWXOeyS//PJLteeXX35RWeXKlVXmPLeGDBly4SJhnMOHD9vWbvdynzx5MqJj1a1b17Z2Oz8RH0aMGKEyt3kQa9eujUU5F5Qvn36uyHkPY9++fdWev/zlL77VBG+53bu9cuVK29ptPsWWLVtU9sYbb6jsvvvus63dzik3zseJiNSoUSOix+Y1PGMLAAAAADAajS0AAAAAwGg0tgAAAAAAo9HYAgAAAACMxvAoD7jdAD558mSV3X///bb1oEGDfKsJyC63N47/6KOPVFapUiXbeuTIkWrP1KlTvSsMvitbtqzKFixYYFv/4x//UHvc3ojebTBUmTJloi8ORliyZIltvXv37qiPNXbsWNu6cOHCUR8LZrvllltU1rhxY5WlpaXZ1ps2bfKtJhGRBx98UGX169dXWa9evXytA8ErUaKEbd2zZ8+IHjd69GgfqgHP2AIAAAAAjEZjCzlwQKRMGZGEBJGqVYOuBvHgm29EunUTKVdOpFAhkcqVRfr0Edm/P+jKkJe1amVd58735/33g64QeR3XPgTl2DGRwYNFqlUTKVxYpHx5kXvvFcnBiyuAbNm3T6RfP5Hq1UWSkkRKlhRp0EDkqae8+xq8FBny5JN8U0XsLFkicvPNIkePitSoIdK0qchXX4mMGycyd67I55+LpKYGXSXysltvFSlWTOeXXhr7WhA/uPYhKMePi7RuLbJypfVLlfbtRbZvF3n7bZEFC6z8T38KukrkZV9+KXL99daTaZdfbp2Dhw6JbN4s8tJLIl69MpvGNs59/LHIO++IPPigyKRJQVeDvO7oUZE77rA+Dh4s8uyzVh4KifTvLzJmjMh994ksXhxsncjbxoyxnikDYoVrH4I0fLjVvDZpIvLBB3/8Ym/sWOvJjXvvFVm6NNASkYft2yeSlma9amDuXJF27eyf/+IL775WQigUCoXbdOjQIUlJSZGDBw9K8eLFvfvqBli+fLlt/cwzz6g9LVq0UFnv3r1VdtFFF9nWBQsWzGF1OXPsmMgVV1gvh5ozx3p5SpUqIlu2hH9srM6JeD73cqs2bdrY1itWrFB7vnC5StWqVUumTBG56y7rZSibN4vkO+dmiFOnrHNw+3aR9etF6tZ1//qxPCc4//KWVq1Eli0T2bYt+saWa5+7uo5/sBs3bozocf3791fZCy+84ElNuQnXPgTl5EnrdrODB0XWrhVxzriqW1dk40aRNWtEGjY8/3G49iFaDz0kMn689eqUhx7K/uOzc05wj20ce/ZZkR9/FJkwQSQxMehqEA++/NL62KKF/Qc7EescvPpq6+9z58a2LgDwE9c+BOWzz6ymtkoV3dSKiHTqZH2cPz+2dSE+HDtm/WKvaFGRHj38/3q8FDlObdwo8uKL1knWvLn1m2LAb0eOWB8dL1743cUXWx83bIhNPYhPb75p3eeTL5/1TFmHDiIVKwZdFfIyrn0IStY51aCB++ez8ghfZAFky5o1IpmZIs2aWQOjFi0S+fBD677vatVEunSxBpl5hcY2Dp09K3L//SIlSoiMGhV0NYgnWW+Vu2OH++e3bbvw5wEvDB9uX/frJzJokPUH8APXPgRl507r4/kGk2XlnHvww+bN1scyZaxfIjtflTJggPXL5ttv9+br8VLkOPTaayKrV1sTyLJ+SwzEQtbt6AsX6kncu3dbv8UTsX67B3itRQuRf/xDZOtWa4jPd9+JPPecSIEC1kCfV14JukLkVVz7EJTDh62PRYq4f75oUesj5x78kJFhfZw3z3pLvXHjRPbutV4p2q+f9VLl7t2t+QJe4BnbMJo3b25bL1myJKBKvLFzp8jAgSItW4rcc0/Q1cBEM2fOtK2dQ2NERLa4TCCrVauWtGljvexp7VqRG26wLnC1aols2iTSs6fI6dPWXuc9aIAXhg61r6tVs35bfOWV1tsQDBliTYhPSgqkPKP9+uuvYfeUKVNGZY8//rgP1eQ+XPsAxKOzZ62Pp09bv0g+d3jU6NHWKwVmzLD+/s9/5vzrcQmNMw8/bE3ImzAh6EoQjxISRGbNst7DbM0akcaNRZKTrfdz3LvXaixEzn8fGuCHNm2s5va330RWrQq6GuRFXPsQlKy39jl61P3zWfd/JyfHph7El3PfM95teFRWtmyZN1+PZ2zjzIIF1r21vXrZ8+PHrY+7d1tviSEiMm2ayCWXxLI6xINKlayXnMyeLbJihfUylMsvF+nWzfrBT8RaA7F02WVWw7FnT9CVIK/i2ocgZA3G27XL/fNZeaVKsakH8SXrvCpS5I9ZA+fKeuu9vXu9+Xo0tnHot9/O/5uR48f/+FxWswt4rUABkc6drT/nynpL3KxfrgCxknUfUNb9ZoAfuPYh1rLuFlq71v3zWXmdOrGpB/El6y2mjh0TOXFCpFAh++ez7mI595ndnKCxjTOhkHu+fbvIf/2X9T5nLrdHAr9zvjn2tqxxnjn0888iM2daA806dvTkkEBE9u0TWb7c+vv53hIDF9a3b98LrkVEBrmMnS5XrpxvNZmCax/8dPXVIikp1tC89etF6tWzfz5rbMbNN8e6MsSDihWtX65s2GA9cdamjf3zWU+mub3HcjS4xxZATH31lX41wK5dIu3bW1MZX3yR4T3w3ooVInPmiJw5Y8+3bxe55RbrPrN27c7/lhhATnHtQxAKFhTp08f6+8MP/3FPrYjI2LHW+9e2bCnSsGEw9SHv69/f+tivn/12n/XrreueiL5FMlo8YwsgpsaMse4xa9BApFw5676KTz+1XqIyaJA19h3w2vffW0MqLrnEOvdKlLCmMX75pdVsXH65yBtvBF0l8jKufQjKwIEiH31k/YLvsstEmje3rn+rVln3Pb71VtAVIi+74w6RDz4Qeecdaxp806bWS5NXrLCufw88oG/PiBaNLYCY6tDBeundhg0in31mTQFNSxN5/HHuL4N/GjcW6d3b+kFu9WrrntqiRa2X5XXubH2OZ8vgJ659CErhwiKffCLy/PMiU6dar14pWdJ628dhw3ilCvz39tvWy+InThRZutSaFN+ggfV2Z17+Uo/GFiJiTSU73/23gJc6dLD+ALFUs6bI3/4WdBWIZ1z7EKSkJOu9vJ3v5w3EQkKC9czsAw/4+3VobAEAQI488cQTF1wDAOA3hkcBAAAAAIxGYwsAAAAAMBqNLQAAAADAaBHdYxv6/6lChw4d8rUYmCPrXAj5PHGKcw9OsTr3zv0anH/IwrUPQeHahyBx7UNQsnPuRdTYZmZmiohIhQoVclAW8qLMzExJSUnx9fginHvQ/D73sr6GCOcfNK59CArXPgSJax+CEsm5lxCKoP09e/aspKenS3JysiQkJHhWIMwVCoUkMzNTypcvL/ny+feKds49OMXq3BPh/IPGtQ9B4dqHIHHtQ1Cyc+5F1NgCAAAAAJBbMTwKAAAAAGA0GlsAAAAAgNFobAEAAAAARqOxBQAAAAAYjcYWAAAAAGA0GlsAAAAAgNFobAEAAAAARvs/Ga2z9ku2piQAAAAASUVORK5CYII=",
            "text/plain": [
              "<Figure size 1200x400 with 12 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(12, 4))\n",
        "for i in range(12):\n",
        "    ax = fig.add_subplot(2, 6, i+1)\n",
        "    ax.set_xticks([]); ax.set_yticks([])\n",
        "    img = mnist_test_dataset[i][0][0, :, :]\n",
        "    pred = model(img.unsqueeze(0).unsqueeze(1))\n",
        "    y_pred = torch.argmax(pred)\n",
        "    ax.imshow(img, cmap='gray_r')\n",
        "    ax.text(0.9, 0.1, y_pred.item(),\n",
        "            size=15, color='blue',\n",
        "            horizontalalignment='center',\n",
        "            verticalalignment='center',\n",
        "            transform=ax.transAxes)\n",
        "\n",
        "\n",
        "#plt.savefig('figures/14_14.png')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "f7K8TPOXF9zA"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}