homeworks/mlp_mnist.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "GSN/DNN 2021/22 - Homework 1 (student version).ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GFHAg_dRZgO7"
      },
      "source": [
        "# Task Overview\n",
        "\n",
        "In this task, your goal is to verify the impact of data noise level in neural network training.\n",
        "You should use MLP architecture trained on MNIST dataset (like in previous lab exercises).\n",
        "\n",
        "\n",
        "We will experiment with two setups:\n",
        "1. Pick X. Take X% of training examples and reassign their labels to random ones. Note that we don't change anything in the test set.\n",
        "2. Pick X. During each training step, for each sample, change values of X% randomly selected pixels to random values. Note that we don't change anything in the test set.\n",
        "\n",
        "For both setups, check the impact of various levels of noise (various values of X%) on model performance. Show plots comparing crossentropy (log-loss) and accuracy with varying X%, and also comparing two setups with each other.\n",
        "Prepare short report briefly explaining the results and observed trends. Consider questions like \"why accuracy/loss increases/decreases so quickly/slowly\", \"why Z is higher in setup 1/2\" and any potentially surprising things you see on charts."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vjmV92bxago7"
      },
      "source": [
        "### Potential questions, clarifications\n",
        "* Q: Can I still use sigmoid/MSE loss?\n",
        "  * You should train your network with softmax and crossentropy loss (log-loss), especially since you should report crossentropy loss.\n",
        "* Q: When I pick X% of pixels/examples, does it have to be exactly X% or can it be X% in expectation?\n",
        "  * A: It's fine either way.\n",
        "* Q: When I randomize pixels, should I randomize them again each time a particular example is drawn (each training step/epoch) or only once before training?\n",
        "  * A: Each training step/epoch.\n",
        "* Q: When I randomize labels, should I randomize them again each time a particular example is drawn (each training step/epoch) or only once before training?\n",
        "  * A: Only once before training.\n",
        "* Q: What is the expected length of report/explanation?\n",
        "  * A: There is no minimum/maximum, but between 5 (concise) and 20 sentences should be good. Don't forget about plots.\n",
        "* Q: When I replace labels/pixels with random values, what random distribution should I use?\n",
        "  * A: A distribution reasonably similar to the data. However, you don't need to match dataset's distribution exactly - approximation will be totally fine, especially if it's faster or easier to get.\n",
        "* Q: Can I use something different than Colab/Jupyter Notebook? E.g. just Python files.\n",
        "  * A: Yes, although notebook is encouraged; please include in you solution code and pdf."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tt2zpwKUZiiW"
      },
      "source": [
        "# Model definition and training."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LBwBXLwmBUHp"
      },
      "source": [
        "import numpy as np\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "from torchvision import datasets, transforms"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FtxJU7RjDBZW"
      },
      "source": [
        "**Model definition**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7adLgBkfBUJx"
      },
      "source": [
        "class Network(nn.Module):\n",
        "    def __init__(self, sizes):\n",
        "        super(Network, self).__init__()\n",
        "        self.layers = nn.ModuleList(\n",
        "           [nn.Linear(sizes[i], sizes[i + 1]) for i in range(len(sizes) - 1)]\n",
        "        )\n",
        "    \n",
        "    def forward(self, x):\n",
        "        x = torch.flatten(x, 1)\n",
        "\n",
        "        for layer in self.layers[:-1]:\n",
        "            x = layer(x)\n",
        "            x = F.relu(x)\n",
        "        \n",
        "        x = self.layers[-1](x)\n",
        "        return F.log_softmax(x, dim=1)"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oUk8JD4XBUMO"
      },
      "source": [
        "def train(model, device, train_loader, optimizer, epoch, log_interval):\n",
        "    model.train()\n",
        "    train_loss = []\n",
        "    correct = 0\n",
        "    \n",
        "    for batch_idx, (data, target) in enumerate(train_loader):\n",
        "        data, target = data.to(device), target.to(device)\n",
        "        optimizer.zero_grad()\n",
        "        output = model(data)\n",
        "        pred = output.argmax(dim=1, keepdim=True)\n",
        "        correct += pred.eq(target.view_as(pred)).sum().item()\n",
        "        loss = F.nll_loss(output, target)\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        if batch_idx % log_interval == 0:\n",
        "            print('Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n",
        "                epoch, batch_idx * len(data), len(train_loader.dataset),\n",
        "                100. * batch_idx / len(train_loader), loss.item()))\n",
        "        train_loss.append(loss.item())\n",
        "    \n",
        "    return np.array([\n",
        "        np.mean(train_loss), correct / len(train_loader.dataset)\n",
        "    ])"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6la7cF5ZBUO0"
      },
      "source": [
        "def test(model, device, test_loader):\n",
        "    model.eval()\n",
        "    test_loss = 0\n",
        "    correct = 0\n",
        "\n",
        "    with torch.no_grad():\n",
        "        for data, target in test_loader:\n",
        "            data, target = data.to(device), target.to(device)\n",
        "            output = model(data)\n",
        "            test_loss += F.nll_loss(output, target, reduction='sum').item()\n",
        "            pred = output.argmax(dim=1, keepdim=True)\n",
        "            correct += pred.eq(target.view_as(pred)).sum().item()\n",
        "\n",
        "    test_loss /= len(test_loader.dataset)\n",
        "\n",
        "    print('\\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\\n'.format(\n",
        "        test_loss, correct, len(test_loader.dataset),\n",
        "        100. * correct / len(test_loader.dataset)))\n",
        "\n",
        "    return np.array([test_loss, correct / len(test_loader.dataset)])"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CvpUsIO9DGcT"
      },
      "source": [
        "batch_size = 256\n",
        "test_batch_size = 1000\n",
        "epochs = 8\n",
        "lr = 1e-3\n",
        "use_cuda = False\n",
        "seed = 256\n",
        "log_interval = 64\n",
        "sizes = [784, 32, 32, 10]"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-YpSqbPiDRMq"
      },
      "source": [
        "use_cuda = not use_cuda and torch.cuda.is_available()\n",
        "\n",
        "torch.manual_seed(seed)\n",
        "device = torch.device(\"cuda\" if use_cuda else \"cpu\")\n",
        "\n",
        "train_kwargs = {'batch_size': batch_size}\n",
        "test_kwargs = {'batch_size': test_batch_size}\n",
        "if use_cuda:\n",
        "    cuda_kwargs = {'num_workers': 1,\n",
        "                    'pin_memory': True,\n",
        "                    'shuffle': True}\n",
        "    train_kwargs.update(cuda_kwargs)\n",
        "    test_kwargs.update(cuda_kwargs)"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fkJrT3xyDRPS"
      },
      "source": [
        "transform=transforms.Compose([\n",
        "    transforms.ToTensor(),\n",
        "    transforms.Normalize((0.1307,), (0.3081,))\n",
        "    ])\n",
        "dataset1 = datasets.MNIST('../data', train=True, download=True,\n",
        "                    transform=transform)\n",
        "dataset2 = datasets.MNIST('../data', train=False,\n",
        "                    transform=transform)\n",
        "train_loader = torch.utils.data.DataLoader(dataset1,**train_kwargs)\n",
        "test_loader = torch.utils.data.DataLoader(dataset2, **test_kwargs)"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ArUGCvMHEBEh"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def plot_history(history):\n",
        "    \"\"\"Visualize model's performance.\"\"\"\n",
        "    fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(8, 5))\n",
        "    fig.tight_layout(pad=3.0)\n",
        "\n",
        "    ax1.plot(history['train'][:, 0], label='train_loss', c='deepskyblue')\n",
        "    ax1.plot(history['test'][:, 0], label='test_loss', c='springgreen')\n",
        "    ax1.set_ylabel('loss')\n",
        "    ax1.set_xlabel('epoch')\n",
        "    ax1.legend()\n",
        "        \n",
        "    ax2.plot(history['train'][:, 1], label='train_acc', c='gold')\n",
        "    ax2.plot(history['test'][:, 1], label='test_acc', c='mediumpurple')\n",
        "    ax2.set_ylabel('accuracy')\n",
        "    ax2.set_xlabel('epoch')\n",
        "    ax2.legend()\n",
        "\n",
        "    plt.suptitle(\"Model's performance\")\n",
        "    plt.show()"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SUBQNqAeDRRu"
      },
      "source": [
        "model = Network(sizes).to(device)\n",
        "optimizer = optim.Adam(model.parameters(), lr=lr)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZR2kU8VEDRUh",
        "outputId": "4aafdea0-fd79-496c-c1f2-31773bf7bc94"
      },
      "source": [
        " history1 = {k: np.zeros((epochs, 2)) for k in ['train', 'test']}\n",
        "\n",
        "for epoch in range(1, epochs + 1):\n",
        "    history1['train'][epoch - 1] = train(\n",
        "        model, device, train_loader, optimizer, epoch, log_interval\n",
        "        )\n",
        "    history1['test'][epoch - 1] = test(model, device, test_loader)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.328902\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.534325\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.352584\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.268161\n",
            "\n",
            "Test set: Average loss: 0.2907, Accuracy: 9151/10000 (92%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.287796\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.301576\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.253809\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.212114\n",
            "\n",
            "Test set: Average loss: 0.2305, Accuracy: 9314/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.212889\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.257574\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.208472\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.180120\n",
            "\n",
            "Test set: Average loss: 0.1987, Accuracy: 9400/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.189153\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.225081\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.180559\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.155340\n",
            "\n",
            "Test set: Average loss: 0.1782, Accuracy: 9461/10000 (95%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.176182\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.193178\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.163919\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.134077\n",
            "\n",
            "Test set: Average loss: 0.1658, Accuracy: 9499/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.168631\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.169393\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.150226\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.116368\n",
            "\n",
            "Test set: Average loss: 0.1577, Accuracy: 9526/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.163236\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.152741\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.139195\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.106281\n",
            "\n",
            "Test set: Average loss: 0.1510, Accuracy: 9541/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.156437\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.136869\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.129180\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.099771\n",
            "\n",
            "Test set: Average loss: 0.1460, Accuracy: 9564/10000 (96%)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 368
        },
        "id": "YQWpznHSDRW2",
        "outputId": "9c16d1cb-3058-40ba-9696-eb5e3da8917d"
      },
      "source": [
        "plot_history(history1)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o9g0-_NhZrNI"
      },
      "source": [
        "# Training models in setup 1: with randomized labels."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l4O4_tD2n3hP"
      },
      "source": [
        "from torch.distributions.categorical import Categorical\n",
        "\n",
        "\n",
        "class RandomizeLabels:\n",
        "    def __init__(self, percent, distribution):\n",
        "        self.percent = percent\n",
        "        self.distribution = distribution\n",
        "        \n",
        "    def __call__(self, labels):\n",
        "        mask = Categorical(torch.tensor(\n",
        "            [self.percent, 1 - self.percent])\n",
        "        ).sample(labels.size())\n",
        "        return labels * mask + (1 - mask) * Categorical(\n",
        "            self.distribution).sample(labels.size())\n",
        "    \n",
        "    def __repr__(self):\n",
        "        return self.__class__.__name__ + self.percent"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vtWOi_z-wik_"
      },
      "source": [
        "# load training data to ensure no data contamination\n",
        "dataset1 = datasets.MNIST('../data', train=True, download=True,\n",
        "                    transform=transform)\n",
        "number_values, number_counts = dataset1.targets.unique(return_counts=True)\n",
        "numbers_dist = {\n",
        "    k: v/len(dataset1.targets) for k, v in zip(number_values, number_counts)\n",
        "}\n",
        "numbers_dist = dict(sorted(numbers_dist.items(), key=lambda x: x[0]))"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tTXFgQJEr3eV"
      },
      "source": [
        "xs = np.arange(0, 1.1, 0.1)"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G7gnAXOCr3nU"
      },
      "source": [
        "history2 = {\n",
        "    x: {k: np.zeros((epochs, 2)) for k in ['train', 'test']} for x in xs\n",
        "}"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c-iioSjqr3px",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c07ff2f5-d5e2-4ee7-844b-140c18665402"
      },
      "source": [
        "for x in xs:\n",
        "    print(f'Randomizing {x * 100:.2f}% of training labels')\n",
        "    dataset1 = datasets.MNIST('../data', train=True, download=True,\n",
        "                    transform=transform)\n",
        "    randomizer = RandomizeLabels(x, torch.tensor(list(numbers_dist.values())))\n",
        "    dataset1.targets = randomizer(dataset1.targets)\n",
        "    train_loader = torch.utils.data.DataLoader(dataset1, **train_kwargs)\n",
        "\n",
        "    # forces same weight initialization for each iteration\n",
        "    torch.manual_seed(seed)\n",
        "    model = Network(sizes).to(device)\n",
        "    optimizer = optim.Adam(model.parameters(), lr=lr)\n",
        "\n",
        "    for epoch in range(1, epochs + 1):\n",
        "        history2[x]['train'][epoch - 1] = train(\n",
        "            model, device, train_loader, optimizer, epoch, log_interval\n",
        "            )\n",
        "        history2[x]['test'][epoch - 1] = test(model, device, test_loader)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Randomizing 0.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.328902\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.534325\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.352584\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.268161\n",
            "\n",
            "Test set: Average loss: 0.2907, Accuracy: 9151/10000 (92%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.287796\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.301576\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.253809\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.212114\n",
            "\n",
            "Test set: Average loss: 0.2305, Accuracy: 9314/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.212889\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.257574\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.208472\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.180120\n",
            "\n",
            "Test set: Average loss: 0.1987, Accuracy: 9400/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.189153\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.225081\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.180559\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.155340\n",
            "\n",
            "Test set: Average loss: 0.1782, Accuracy: 9461/10000 (95%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.176182\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.193178\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.163919\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.134077\n",
            "\n",
            "Test set: Average loss: 0.1658, Accuracy: 9499/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.168631\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.169393\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.150226\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.116368\n",
            "\n",
            "Test set: Average loss: 0.1577, Accuracy: 9526/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.163236\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.152741\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.139195\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.106281\n",
            "\n",
            "Test set: Average loss: 0.1510, Accuracy: 9541/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.156437\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.136869\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.129180\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.099771\n",
            "\n",
            "Test set: Average loss: 0.1460, Accuracy: 9564/10000 (96%)\n",
            "\n",
            "Randomizing 10.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.325947\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.762495\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.716529\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.782498\n",
            "\n",
            "Test set: Average loss: 0.3961, Accuracy: 9149/10000 (91%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.745877\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.600989\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.611060\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.715492\n",
            "\n",
            "Test set: Average loss: 0.3456, Accuracy: 9297/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.699363\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.550732\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.569955\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.677003\n",
            "\n",
            "Test set: Average loss: 0.3171, Accuracy: 9386/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.666017\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.521490\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.538104\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.650258\n",
            "\n",
            "Test set: Average loss: 0.2967, Accuracy: 9447/10000 (94%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.635462\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.505149\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.515341\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.629647\n",
            "\n",
            "Test set: Average loss: 0.2832, Accuracy: 9477/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.616482\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.495089\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.504734\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.614771\n",
            "\n",
            "Test set: Average loss: 0.2716, Accuracy: 9488/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.599115\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.483897\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.493084\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.606498\n",
            "\n",
            "Test set: Average loss: 0.2639, Accuracy: 9521/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.587591\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.477818\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.483158\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.597146\n",
            "\n",
            "Test set: Average loss: 0.2572, Accuracy: 9533/10000 (95%)\n",
            "\n",
            "Randomizing 20.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.323029\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.276812\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.057619\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 1.224443\n",
            "\n",
            "Test set: Average loss: 0.5127, Accuracy: 9094/10000 (91%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.989084\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.086656\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.948947\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.158216\n",
            "\n",
            "Test set: Average loss: 0.4590, Accuracy: 9253/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.918959\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.043654\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.907226\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.127602\n",
            "\n",
            "Test set: Average loss: 0.4318, Accuracy: 9345/10000 (93%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.874681\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.023639\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.887691\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.108379\n",
            "\n",
            "Test set: Average loss: 0.4060, Accuracy: 9412/10000 (94%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.847752\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.007598\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.874939\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.094441\n",
            "\n",
            "Test set: Average loss: 0.3914, Accuracy: 9428/10000 (94%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.832319\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.001162\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.865996\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.077156\n",
            "\n",
            "Test set: Average loss: 0.3790, Accuracy: 9449/10000 (94%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.818742\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.993884\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.857673\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.067156\n",
            "\n",
            "Test set: Average loss: 0.3709, Accuracy: 9488/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.813462\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.987727\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.852912\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.060736\n",
            "\n",
            "Test set: Average loss: 0.3642, Accuracy: 9502/10000 (95%)\n",
            "\n",
            "Randomizing 30.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.314164\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.562473\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.379949\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 1.475342\n",
            "\n",
            "Test set: Average loss: 0.6495, Accuracy: 8988/10000 (90%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.321536\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.410755\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.312833\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.438270\n",
            "\n",
            "Test set: Average loss: 0.5998, Accuracy: 9207/10000 (92%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.260158\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.380995\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.292821\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.420185\n",
            "\n",
            "Test set: Average loss: 0.5715, Accuracy: 9303/10000 (93%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.231688\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.364090\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.286523\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.400336\n",
            "\n",
            "Test set: Average loss: 0.5532, Accuracy: 9358/10000 (94%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.214254\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.351837\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.279221\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.388725\n",
            "\n",
            "Test set: Average loss: 0.5391, Accuracy: 9393/10000 (94%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.201946\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.346605\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.269902\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.377760\n",
            "\n",
            "Test set: Average loss: 0.5274, Accuracy: 9423/10000 (94%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.190445\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.340662\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.264797\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.364429\n",
            "\n",
            "Test set: Average loss: 0.5166, Accuracy: 9448/10000 (94%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.184439\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.335725\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.257322\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.357777\n",
            "\n",
            "Test set: Average loss: 0.5108, Accuracy: 9477/10000 (95%)\n",
            "\n",
            "Randomizing 40.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.313284\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.727501\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.593222\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 1.702773\n",
            "\n",
            "Test set: Average loss: 0.7626, Accuracy: 9029/10000 (90%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.505301\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.614388\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.538819\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.682313\n",
            "\n",
            "Test set: Average loss: 0.7156, Accuracy: 9183/10000 (92%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.477636\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.586022\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.519618\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.662130\n",
            "\n",
            "Test set: Average loss: 0.6860, Accuracy: 9265/10000 (93%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.457735\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.573105\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.504929\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.652195\n",
            "\n",
            "Test set: Average loss: 0.6657, Accuracy: 9335/10000 (93%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.439599\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.562701\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.493396\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.639833\n",
            "\n",
            "Test set: Average loss: 0.6560, Accuracy: 9371/10000 (94%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.431350\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.558275\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.490711\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.635130\n",
            "\n",
            "Test set: Average loss: 0.6484, Accuracy: 9397/10000 (94%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.428163\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.558701\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.487588\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.627026\n",
            "\n",
            "Test set: Average loss: 0.6452, Accuracy: 9410/10000 (94%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.425418\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.551756\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.486379\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.617285\n",
            "\n",
            "Test set: Average loss: 0.6377, Accuracy: 9413/10000 (94%)\n",
            "\n",
            "Randomizing 50.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.308566\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.893411\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.780773\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 1.900007\n",
            "\n",
            "Test set: Average loss: 0.9299, Accuracy: 8936/10000 (89%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.684312\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.794777\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.729943\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.873433\n",
            "\n",
            "Test set: Average loss: 0.8721, Accuracy: 9119/10000 (91%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.643436\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.775599\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.705976\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.867930\n",
            "\n",
            "Test set: Average loss: 0.8417, Accuracy: 9215/10000 (92%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.625820\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.770384\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.694281\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.858729\n",
            "\n",
            "Test set: Average loss: 0.8209, Accuracy: 9277/10000 (93%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.614125\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.761892\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.683380\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.850657\n",
            "\n",
            "Test set: Average loss: 0.8036, Accuracy: 9326/10000 (93%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.605888\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.754110\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.677640\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.846147\n",
            "\n",
            "Test set: Average loss: 0.7983, Accuracy: 9346/10000 (93%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.599677\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.743600\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.671136\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.841012\n",
            "\n",
            "Test set: Average loss: 0.7916, Accuracy: 9357/10000 (94%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.597569\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.737518\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.668566\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.834983\n",
            "\n",
            "Test set: Average loss: 0.7871, Accuracy: 9381/10000 (94%)\n",
            "\n",
            "Randomizing 60.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.304800\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.026261\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.935245\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.016407\n",
            "\n",
            "Test set: Average loss: 1.0976, Accuracy: 8954/10000 (90%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.930332\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.936470\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.897468\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.006931\n",
            "\n",
            "Test set: Average loss: 1.0450, Accuracy: 9129/10000 (91%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.895282\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.918835\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.885800\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.995566\n",
            "\n",
            "Test set: Average loss: 1.0144, Accuracy: 9232/10000 (92%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.877960\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.906115\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.876982\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.983394\n",
            "\n",
            "Test set: Average loss: 0.9895, Accuracy: 9266/10000 (93%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.865422\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.894368\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.866470\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.970619\n",
            "\n",
            "Test set: Average loss: 0.9777, Accuracy: 9303/10000 (93%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.855883\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.889908\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.857980\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.961311\n",
            "\n",
            "Test set: Average loss: 0.9672, Accuracy: 9323/10000 (93%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.847492\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.881740\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.854479\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.948547\n",
            "\n",
            "Test set: Average loss: 0.9612, Accuracy: 9338/10000 (93%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.840422\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.873748\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.848473\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.937809\n",
            "\n",
            "Test set: Average loss: 0.9540, Accuracy: 9349/10000 (93%)\n",
            "\n",
            "Randomizing 70.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.298957\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.149345\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.112147\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.131457\n",
            "\n",
            "Test set: Average loss: 1.3427, Accuracy: 8848/10000 (88%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 2.117981\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 2.092885\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 2.077103\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.124719\n",
            "\n",
            "Test set: Average loss: 1.2693, Accuracy: 9009/10000 (90%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 2.092846\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 2.078644\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 2.067947\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 2.120093\n",
            "\n",
            "Test set: Average loss: 1.2477, Accuracy: 9090/10000 (91%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 2.082892\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 2.072261\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 2.054134\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 2.114606\n",
            "\n",
            "Test set: Average loss: 1.2238, Accuracy: 9154/10000 (92%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 2.077456\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 2.072340\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 2.043447\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 2.110800\n",
            "\n",
            "Test set: Average loss: 1.2149, Accuracy: 9191/10000 (92%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 2.074721\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 2.068922\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 2.041489\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 2.107907\n",
            "\n",
            "Test set: Average loss: 1.2005, Accuracy: 9231/10000 (92%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 2.072017\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 2.067109\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 2.039727\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 2.104081\n",
            "\n",
            "Test set: Average loss: 1.2023, Accuracy: 9231/10000 (92%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 2.067654\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 2.060757\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 2.037241\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 2.101448\n",
            "\n",
            "Test set: Average loss: 1.1937, Accuracy: 9208/10000 (92%)\n",
            "\n",
            "Randomizing 80.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.299507\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.239496\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.227427\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.230522\n",
            "\n",
            "Test set: Average loss: 1.5948, Accuracy: 8586/10000 (86%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 2.206819\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 2.184236\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 2.214674\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.226802\n",
            "\n",
            "Test set: Average loss: 1.5110, Accuracy: 8853/10000 (89%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 2.186515\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 2.178867\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 2.206042\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 2.224342\n",
            "\n",
            "Test set: Average loss: 1.4733, Accuracy: 8970/10000 (90%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 2.179813\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 2.173324\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 2.202831\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 2.221626\n",
            "\n",
            "Test set: Average loss: 1.4608, Accuracy: 8990/10000 (90%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 2.176965\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 2.170818\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 2.201909\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 2.216875\n",
            "\n",
            "Test set: Average loss: 1.4518, Accuracy: 9002/10000 (90%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 2.175259\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 2.166720\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 2.196356\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 2.211627\n",
            "\n",
            "Test set: Average loss: 1.4423, Accuracy: 9010/10000 (90%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 2.170785\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 2.164003\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 2.193127\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 2.205057\n",
            "\n",
            "Test set: Average loss: 1.4394, Accuracy: 9018/10000 (90%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 2.163508\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 2.162902\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 2.193487\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 2.202833\n",
            "\n",
            "Test set: Average loss: 1.4327, Accuracy: 8995/10000 (90%)\n",
            "\n",
            "Randomizing 90.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.301287\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.290424\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.291232\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.261701\n",
            "\n",
            "Test set: Average loss: 1.9681, Accuracy: 6654/10000 (67%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 2.257023\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 2.273906\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 2.287647\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.260416\n",
            "\n",
            "Test set: Average loss: 1.8879, Accuracy: 8157/10000 (82%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 2.253490\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 2.267440\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 2.287446\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 2.259617\n",
            "\n",
            "Test set: Average loss: 1.8365, Accuracy: 8330/10000 (83%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 2.245714\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 2.261583\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 2.282949\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 2.265498\n",
            "\n",
            "Test set: Average loss: 1.8267, Accuracy: 8323/10000 (83%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 2.244542\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 2.259521\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 2.282988\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 2.261157\n",
            "\n",
            "Test set: Average loss: 1.8081, Accuracy: 8274/10000 (83%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 2.243749\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 2.254463\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 2.283870\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 2.259453\n",
            "\n",
            "Test set: Average loss: 1.7891, Accuracy: 8200/10000 (82%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 2.242261\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 2.248787\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 2.282795\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 2.256696\n",
            "\n",
            "Test set: Average loss: 1.7935, Accuracy: 8132/10000 (81%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 2.242680\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 2.244848\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 2.284296\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 2.250443\n",
            "\n",
            "Test set: Average loss: 1.7860, Accuracy: 8086/10000 (81%)\n",
            "\n",
            "Randomizing 100.00% of training labels\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.305882\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.298209\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.294037\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.304117\n",
            "\n",
            "Test set: Average loss: 2.3053, Accuracy: 1143/10000 (11%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 2.299474\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 2.303161\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 2.293689\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.302136\n",
            "\n",
            "Test set: Average loss: 2.3043, Accuracy: 1166/10000 (12%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 2.297042\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 2.301128\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 2.294818\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 2.300568\n",
            "\n",
            "Test set: Average loss: 2.3025, Accuracy: 1176/10000 (12%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 2.295652\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 2.298683\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 2.292837\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 2.301674\n",
            "\n",
            "Test set: Average loss: 2.2947, Accuracy: 1177/10000 (12%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 2.298493\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 2.300623\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 2.288707\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 2.301648\n",
            "\n",
            "Test set: Average loss: 2.2977, Accuracy: 1261/10000 (13%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 2.294862\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 2.301797\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 2.289674\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 2.300495\n",
            "\n",
            "Test set: Average loss: 2.3038, Accuracy: 1294/10000 (13%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 2.295073\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 2.301357\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 2.283385\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 2.296354\n",
            "\n",
            "Test set: Average loss: 2.3000, Accuracy: 1288/10000 (13%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 2.293318\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 2.299173\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 2.283817\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 2.295442\n",
            "\n",
            "Test set: Average loss: 2.3033, Accuracy: 1286/10000 (13%)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ieho04CuZ06r"
      },
      "source": [
        "# Training models in setup 2: with randomized pixels."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MGkhsf7sXoHM"
      },
      "source": [
        "class RandomizePixels:\n",
        "    def __init__(self, percent, distribution):\n",
        "        self.percent = percent\n",
        "        self.distribution = distribution\n",
        "        \n",
        "    def __call__(self, data):\n",
        "        mask = Categorical(torch.tensor(\n",
        "            [self.percent, 1 - self.percent])\n",
        "        ).sample(data.size())\n",
        "        randomized = Categorical(self.distribution).sample(data.size())\n",
        "        return (data * mask) + ((1 - mask) * randomized / 255)\n",
        "    \n",
        "    def __repr__(self):\n",
        "        return self.__class__.__name__ + self.percent"
      ],
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kGI1C3o67JZT"
      },
      "source": [
        "# load training data to ensure no data contamination\n",
        "dataset1 = datasets.MNIST('../data', train=True, download=True,\n",
        "                    transform=transform)\n",
        "pixel_values, pixel_counts = dataset1.data.flatten().unique(return_counts=True)\n",
        "pixels_dist = {\n",
        "    k: v/len(dataset1.data) for k, v in zip(pixel_values, pixel_counts)\n",
        "}\n",
        "pixels_dist = dict(sorted(pixels_dist.items(), key=lambda x: x[0]))"
      ],
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sLmeEhglXoJ1"
      },
      "source": [
        "xs = np.arange(0.0, 1.1, 0.1)"
      ],
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_hdsmha7XoMc"
      },
      "source": [
        "history3 = {\n",
        "    x: {k: np.zeros((epochs, 2)) for k in ['train', 'test']} for x in xs\n",
        "}"
      ],
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OhuClXQM4dvv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "66b834d8-ec51-42b8-8fb5-2827efaa55f5"
      },
      "source": [
        "for x in xs:\n",
        "    print(f'Randomizing {x * 100}% of pixels in each training image')\n",
        "    train_transform = transforms.Compose([\n",
        "        transforms.ToTensor(),\n",
        "        RandomizePixels(x, torch.tensor(list(pixels_dist.values()))),\n",
        "        transforms.Normalize((0.1307,), (0.3081,)),\n",
        "    ])\n",
        "    dataset1 = datasets.MNIST('../data', train=True, download=True,\n",
        "                    transform=train_transform)\n",
        "    train_loader = torch.utils.data.DataLoader(dataset1, **train_kwargs)\n",
        "\n",
        "    # forces same weight initialization for each iteration\n",
        "    torch.manual_seed(seed)\n",
        "    model = Network(sizes).to(device)\n",
        "    optimizer = optim.Adam(model.parameters(), lr=lr)\n",
        "\n",
        "    for epoch in range(1, epochs + 1):\n",
        "        history3[x]['train'][epoch - 1] = train(\n",
        "            model, device, train_loader, optimizer, epoch, log_interval\n",
        "            )\n",
        "        history3[x]['test'][epoch - 1] = test(model, device, test_loader)"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Randomizing 0.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.328902\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.534325\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.352584\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.268161\n",
            "\n",
            "Test set: Average loss: 0.2907, Accuracy: 9151/10000 (92%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.287796\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.301576\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.253809\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.212114\n",
            "\n",
            "Test set: Average loss: 0.2305, Accuracy: 9314/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.212889\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.257574\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.208472\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.180120\n",
            "\n",
            "Test set: Average loss: 0.1987, Accuracy: 9400/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.189153\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.225081\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.180559\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.155340\n",
            "\n",
            "Test set: Average loss: 0.1782, Accuracy: 9461/10000 (95%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.176182\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.193178\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.163919\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.134077\n",
            "\n",
            "Test set: Average loss: 0.1658, Accuracy: 9499/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.168631\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.169393\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.150226\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.116368\n",
            "\n",
            "Test set: Average loss: 0.1577, Accuracy: 9526/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.163236\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.152741\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.139195\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.106281\n",
            "\n",
            "Test set: Average loss: 0.1510, Accuracy: 9541/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.156437\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.136869\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.129180\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.099771\n",
            "\n",
            "Test set: Average loss: 0.1460, Accuracy: 9564/10000 (96%)\n",
            "\n",
            "Randomizing 10.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.327550\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.554894\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.383521\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.296973\n",
            "\n",
            "Test set: Average loss: 0.2911, Accuracy: 9159/10000 (92%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.318016\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.333028\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.269934\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.247653\n",
            "\n",
            "Test set: Average loss: 0.2360, Accuracy: 9288/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.248273\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.246450\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.225802\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.221924\n",
            "\n",
            "Test set: Average loss: 0.1982, Accuracy: 9409/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.206265\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.245215\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.217363\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.189932\n",
            "\n",
            "Test set: Average loss: 0.1762, Accuracy: 9472/10000 (95%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.216724\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.226184\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.179225\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.164002\n",
            "\n",
            "Test set: Average loss: 0.1630, Accuracy: 9506/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.177303\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.206483\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.163122\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.155231\n",
            "\n",
            "Test set: Average loss: 0.1518, Accuracy: 9535/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.184404\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.165895\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.164712\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.136828\n",
            "\n",
            "Test set: Average loss: 0.1461, Accuracy: 9560/10000 (96%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.165250\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.150221\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.155521\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.123068\n",
            "\n",
            "Test set: Average loss: 0.1364, Accuracy: 9587/10000 (96%)\n",
            "\n",
            "Randomizing 20.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.327556\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.592166\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.427049\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.336862\n",
            "\n",
            "Test set: Average loss: 0.3027, Accuracy: 9127/10000 (91%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.353733\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.341079\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.303318\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.294014\n",
            "\n",
            "Test set: Average loss: 0.2468, Accuracy: 9269/10000 (93%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.292917\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.261940\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.232048\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.271170\n",
            "\n",
            "Test set: Average loss: 0.2063, Accuracy: 9398/10000 (94%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.221631\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.269234\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.241315\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.204733\n",
            "\n",
            "Test set: Average loss: 0.1858, Accuracy: 9445/10000 (94%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.207841\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.206987\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.204042\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.190277\n",
            "\n",
            "Test set: Average loss: 0.1668, Accuracy: 9507/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.189720\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.189893\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.177424\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.173226\n",
            "\n",
            "Test set: Average loss: 0.1558, Accuracy: 9551/10000 (96%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.198490\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.180439\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.182374\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.131283\n",
            "\n",
            "Test set: Average loss: 0.1468, Accuracy: 9566/10000 (96%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.191353\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.165439\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.194551\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.158250\n",
            "\n",
            "Test set: Average loss: 0.1338, Accuracy: 9601/10000 (96%)\n",
            "\n",
            "Randomizing 30.000000000000004% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.324857\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.664406\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.467528\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.380290\n",
            "\n",
            "Test set: Average loss: 0.3241, Accuracy: 9088/10000 (91%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.401631\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.371494\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.344859\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.357972\n",
            "\n",
            "Test set: Average loss: 0.2744, Accuracy: 9233/10000 (92%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.319473\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.317598\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.293737\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.303306\n",
            "\n",
            "Test set: Average loss: 0.2348, Accuracy: 9316/10000 (93%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.268464\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.319580\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.293767\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.256042\n",
            "\n",
            "Test set: Average loss: 0.2106, Accuracy: 9373/10000 (94%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.248651\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.268031\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.289756\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.247546\n",
            "\n",
            "Test set: Average loss: 0.1884, Accuracy: 9461/10000 (95%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.250133\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.236948\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.253861\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.253694\n",
            "\n",
            "Test set: Average loss: 0.1759, Accuracy: 9487/10000 (95%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.245538\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.212676\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.173219\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.216879\n",
            "\n",
            "Test set: Average loss: 0.1692, Accuracy: 9485/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.269529\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.218824\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.225187\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.196305\n",
            "\n",
            "Test set: Average loss: 0.1518, Accuracy: 9538/10000 (95%)\n",
            "\n",
            "Randomizing 40.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.325504\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.752592\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.573982\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.462296\n",
            "\n",
            "Test set: Average loss: 0.3657, Accuracy: 9022/10000 (90%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.468136\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.435210\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.404323\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.413756\n",
            "\n",
            "Test set: Average loss: 0.3195, Accuracy: 9142/10000 (91%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.389228\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.381504\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.371182\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.375584\n",
            "\n",
            "Test set: Average loss: 0.2722, Accuracy: 9254/10000 (93%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.301976\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.399986\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.341333\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.347020\n",
            "\n",
            "Test set: Average loss: 0.2419, Accuracy: 9325/10000 (93%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.320055\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.373526\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.353653\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.332060\n",
            "\n",
            "Test set: Average loss: 0.2226, Accuracy: 9370/10000 (94%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.300302\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.291929\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.311696\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.309905\n",
            "\n",
            "Test set: Average loss: 0.2081, Accuracy: 9413/10000 (94%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.278418\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.275161\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.229506\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.272214\n",
            "\n",
            "Test set: Average loss: 0.1955, Accuracy: 9454/10000 (95%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.313047\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.287017\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.281731\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.286852\n",
            "\n",
            "Test set: Average loss: 0.1837, Accuracy: 9491/10000 (95%)\n",
            "\n",
            "Randomizing 50.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.320609\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 0.877406\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.647368\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.574071\n",
            "\n",
            "Test set: Average loss: 0.4178, Accuracy: 8941/10000 (89%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.529796\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.499013\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.504452\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.534914\n",
            "\n",
            "Test set: Average loss: 0.3734, Accuracy: 9066/10000 (91%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.460373\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.425427\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.486905\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.462991\n",
            "\n",
            "Test set: Average loss: 0.3271, Accuracy: 9172/10000 (92%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.398207\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.502818\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.424618\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.471041\n",
            "\n",
            "Test set: Average loss: 0.2932, Accuracy: 9255/10000 (93%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.374762\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.482718\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.447944\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.437248\n",
            "\n",
            "Test set: Average loss: 0.2674, Accuracy: 9308/10000 (93%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.360324\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.369691\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.407311\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.414662\n",
            "\n",
            "Test set: Average loss: 0.2480, Accuracy: 9374/10000 (94%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.359789\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.371377\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.345138\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.352250\n",
            "\n",
            "Test set: Average loss: 0.2316, Accuracy: 9393/10000 (94%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.337919\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.338066\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.365058\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.353102\n",
            "\n",
            "Test set: Average loss: 0.2144, Accuracy: 9438/10000 (94%)\n",
            "\n",
            "Randomizing 60.00000000000001% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.316271\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.066506\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 0.762913\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.721051\n",
            "\n",
            "Test set: Average loss: 0.5005, Accuracy: 8790/10000 (88%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.659562\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.630483\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.663535\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.664421\n",
            "\n",
            "Test set: Average loss: 0.4588, Accuracy: 8893/10000 (89%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.642537\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.591993\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.679063\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.644985\n",
            "\n",
            "Test set: Average loss: 0.4192, Accuracy: 8996/10000 (90%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.554878\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.585459\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.557197\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.607700\n",
            "\n",
            "Test set: Average loss: 0.3795, Accuracy: 9098/10000 (91%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.559012\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.624544\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.610248\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.566962\n",
            "\n",
            "Test set: Average loss: 0.3608, Accuracy: 9118/10000 (91%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.495284\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.573927\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.538805\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.578065\n",
            "\n",
            "Test set: Average loss: 0.3204, Accuracy: 9234/10000 (92%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.492357\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.478266\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.460910\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.489237\n",
            "\n",
            "Test set: Average loss: 0.3083, Accuracy: 9262/10000 (93%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.472812\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.491732\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.514947\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.536121\n",
            "\n",
            "Test set: Average loss: 0.2786, Accuracy: 9343/10000 (93%)\n",
            "\n",
            "Randomizing 70.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.314399\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.384819\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.000425\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 0.911899\n",
            "\n",
            "Test set: Average loss: 0.5916, Accuracy: 8624/10000 (86%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 0.944414\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 0.844198\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 0.895084\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 0.898377\n",
            "\n",
            "Test set: Average loss: 0.5670, Accuracy: 8691/10000 (87%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 0.875155\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 0.776506\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 0.873691\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 0.876880\n",
            "\n",
            "Test set: Average loss: 0.5312, Accuracy: 8765/10000 (88%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 0.800242\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 0.886596\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 0.794533\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 0.912240\n",
            "\n",
            "Test set: Average loss: 0.4929, Accuracy: 8840/10000 (88%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 0.838248\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 0.860749\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 0.883240\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 0.823150\n",
            "\n",
            "Test set: Average loss: 0.4831, Accuracy: 8874/10000 (89%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 0.788341\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 0.854443\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 0.777437\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 0.894795\n",
            "\n",
            "Test set: Average loss: 0.4345, Accuracy: 8998/10000 (90%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 0.748227\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 0.749928\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 0.764836\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 0.836283\n",
            "\n",
            "Test set: Average loss: 0.4137, Accuracy: 9043/10000 (90%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 0.690414\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 0.751785\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 0.748271\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 0.857041\n",
            "\n",
            "Test set: Average loss: 0.3903, Accuracy: 9073/10000 (91%)\n",
            "\n",
            "Randomizing 80.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.313452\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 1.864867\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 1.398016\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 1.340170\n",
            "\n",
            "Test set: Average loss: 0.7390, Accuracy: 8211/10000 (82%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.358205\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.265771\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.317828\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.310793\n",
            "\n",
            "Test set: Average loss: 0.7056, Accuracy: 8328/10000 (83%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.260353\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.162324\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.207993\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.263701\n",
            "\n",
            "Test set: Average loss: 0.6779, Accuracy: 8400/10000 (84%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.136686\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.252024\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.196596\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.290496\n",
            "\n",
            "Test set: Average loss: 0.6451, Accuracy: 8489/10000 (85%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.304245\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.277148\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.317411\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.226330\n",
            "\n",
            "Test set: Average loss: 0.6268, Accuracy: 8498/10000 (85%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.249121\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.230312\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.172113\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.374517\n",
            "\n",
            "Test set: Average loss: 0.6086, Accuracy: 8532/10000 (85%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.234259\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.147063\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.209098\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.285892\n",
            "\n",
            "Test set: Average loss: 0.5706, Accuracy: 8586/10000 (86%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.176986\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.217118\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.195169\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.296640\n",
            "\n",
            "Test set: Average loss: 0.5510, Accuracy: 8654/10000 (87%)\n",
            "\n",
            "Randomizing 90.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.313836\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.261100\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.038630\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.004978\n",
            "\n",
            "Test set: Average loss: 0.9057, Accuracy: 7145/10000 (71%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 1.977757\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 1.932691\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 1.937171\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 1.939338\n",
            "\n",
            "Test set: Average loss: 0.7754, Accuracy: 7678/10000 (77%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 1.931579\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 1.855919\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 1.840933\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 1.907681\n",
            "\n",
            "Test set: Average loss: 0.7329, Accuracy: 7903/10000 (79%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 1.813069\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 1.867911\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 1.854746\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 1.911260\n",
            "\n",
            "Test set: Average loss: 0.7410, Accuracy: 7902/10000 (79%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 1.970093\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 1.881352\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 1.885836\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 1.892188\n",
            "\n",
            "Test set: Average loss: 0.7339, Accuracy: 7915/10000 (79%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 1.889848\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 1.839094\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 1.853345\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 1.919769\n",
            "\n",
            "Test set: Average loss: 0.7214, Accuracy: 7963/10000 (80%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 1.926596\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 1.785716\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 1.897847\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 1.944160\n",
            "\n",
            "Test set: Average loss: 0.7350, Accuracy: 7911/10000 (79%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 1.856061\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 1.894748\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 1.872686\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 1.916900\n",
            "\n",
            "Test set: Average loss: 0.7246, Accuracy: 7905/10000 (79%)\n",
            "\n",
            "Randomizing 100.0% of pixels in each training image\n",
            "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.312356\n",
            "Train Epoch: 1 [16384/60000 (27%)]\tLoss: 2.310473\n",
            "Train Epoch: 1 [32768/60000 (54%)]\tLoss: 2.302153\n",
            "Train Epoch: 1 [49152/60000 (82%)]\tLoss: 2.304182\n",
            "\n",
            "Test set: Average loss: 2.3092, Accuracy: 1046/10000 (10%)\n",
            "\n",
            "Train Epoch: 2 [0/60000 (0%)]\tLoss: 2.295433\n",
            "Train Epoch: 2 [16384/60000 (27%)]\tLoss: 2.305795\n",
            "Train Epoch: 2 [32768/60000 (54%)]\tLoss: 2.301555\n",
            "Train Epoch: 2 [49152/60000 (82%)]\tLoss: 2.296545\n",
            "\n",
            "Test set: Average loss: 2.3083, Accuracy: 715/10000 (7%)\n",
            "\n",
            "Train Epoch: 3 [0/60000 (0%)]\tLoss: 2.299117\n",
            "Train Epoch: 3 [16384/60000 (27%)]\tLoss: 2.304365\n",
            "Train Epoch: 3 [32768/60000 (54%)]\tLoss: 2.294991\n",
            "Train Epoch: 3 [49152/60000 (82%)]\tLoss: 2.305091\n",
            "\n",
            "Test set: Average loss: 2.3019, Accuracy: 886/10000 (9%)\n",
            "\n",
            "Train Epoch: 4 [0/60000 (0%)]\tLoss: 2.293969\n",
            "Train Epoch: 4 [16384/60000 (27%)]\tLoss: 2.308291\n",
            "Train Epoch: 4 [32768/60000 (54%)]\tLoss: 2.302571\n",
            "Train Epoch: 4 [49152/60000 (82%)]\tLoss: 2.297899\n",
            "\n",
            "Test set: Average loss: 2.3033, Accuracy: 882/10000 (9%)\n",
            "\n",
            "Train Epoch: 5 [0/60000 (0%)]\tLoss: 2.293067\n",
            "Train Epoch: 5 [16384/60000 (27%)]\tLoss: 2.303445\n",
            "Train Epoch: 5 [32768/60000 (54%)]\tLoss: 2.301106\n",
            "Train Epoch: 5 [49152/60000 (82%)]\tLoss: 2.299173\n",
            "\n",
            "Test set: Average loss: 2.3019, Accuracy: 1092/10000 (11%)\n",
            "\n",
            "Train Epoch: 6 [0/60000 (0%)]\tLoss: 2.295146\n",
            "Train Epoch: 6 [16384/60000 (27%)]\tLoss: 2.308686\n",
            "Train Epoch: 6 [32768/60000 (54%)]\tLoss: 2.298102\n",
            "Train Epoch: 6 [49152/60000 (82%)]\tLoss: 2.298945\n",
            "\n",
            "Test set: Average loss: 2.3023, Accuracy: 591/10000 (6%)\n",
            "\n",
            "Train Epoch: 7 [0/60000 (0%)]\tLoss: 2.298870\n",
            "Train Epoch: 7 [16384/60000 (27%)]\tLoss: 2.303475\n",
            "Train Epoch: 7 [32768/60000 (54%)]\tLoss: 2.305031\n",
            "Train Epoch: 7 [49152/60000 (82%)]\tLoss: 2.299059\n",
            "\n",
            "Test set: Average loss: 2.3021, Accuracy: 878/10000 (9%)\n",
            "\n",
            "Train Epoch: 8 [0/60000 (0%)]\tLoss: 2.295788\n",
            "Train Epoch: 8 [16384/60000 (27%)]\tLoss: 2.306657\n",
            "Train Epoch: 8 [32768/60000 (54%)]\tLoss: 2.301784\n",
            "Train Epoch: 8 [49152/60000 (82%)]\tLoss: 2.302490\n",
            "\n",
            "Test set: Average loss: 2.3028, Accuracy: 979/10000 (10%)\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rXaJMQyYZ3i5"
      },
      "source": [
        "# Plots and report."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FecW3w5ygiWp"
      },
      "source": [
        "So-called \"base model\" was a 3-layered feed forward network, trained for 8 epochs using Adam optimizer. It performed well both on training as well as on testing data (acc 0.963, 0.956). Its learning curve suggested no overfitting."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 368
        },
        "id": "bKo-WCbRhXUj",
        "outputId": "f9a9119c-b19c-4475-efef-8ade8fe6537d"
      },
      "source": [
        "plot_history(history1)"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZRVcUaauhXeY"
      },
      "source": [
        "Reassigning labels resulted in significant decrease in accuracy on training data, and slight decrease in accuracy on testing data, but it's not surprising with unbiased mislabeling. Let's say we randomize label for 80% of data. That means there is 20% chance that we didn't change correct label, and 8% that we randomly chose correct label. Which means that for each class k we have 28% chance that image was labeled as k, and 8% for any other class. So our network has expected training accuracy at 28%, but that doesn't mean it won't perform well on test data.\n",
        "\n",
        "With 100% randomization our network will simply guess class with largest representation in data, what leaves us with ~10% accuracy.\n",
        "\n",
        "Why we have such good generalization? Because neural networks \"move\" our input into high-dimensional space, creating high-dimensional features, and making separation easy. But that still holds for mislabeled data - images don't change in this setup. On the other hand, one could argue that 80% accuracy on MNIST isn't acceptable, since MNIST is often called a \"machine learning unitest\"."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "g-dEAJo0joKZ",
        "outputId": "bb068f81-788c-4f45-cbaa-41793616693e"
      },
      "source": [
        "for x in xs[1:]:\n",
        "    print(f'{x * 100:.2f}% labels reassigned')\n",
        "    plot_history(history2[x])\n",
        "    print()"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhgAAAFfCAYAAAAMMbVZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXyU5bn/8c+VZLJBFkgAZQcFBMGNCFq0bnVDKmpP3WpbrK1d1GqrVjzH2ta2Pz09Hru5tNZqW3vqhlXpkR6sW8FWVqUuyK5IACUEQgjZk+v3xzOBIYQQYCbPJPN9v17zysw8y1wTyOSb+76f+zZ3R0RERCSe0sIuQERERLofBQwRERGJOwUMERERiTsFDBEREYk7BQwRERGJOwUMERERiTsFDJEuyMyGmpmbWUYH9p1mZq91Rl0dZYFHzGyrmS0Iux4RiT8FDJEEM7MPzKzezIpbPf9mNCQM7eR6kmHym5OAM4GB7j4h7GJEJP4UMEQ6x/vAZS0PzGwckBteOeGJtroMAT5w9x0HeLyIJDkFDJHO8SjwhZjHXwT+ELuDmRWY2R/MrMzM1prZbWaWFt2WbmZ3m9lmM1sDnNfGsb81s41mtt7MfmRm6fsqKtp9ssbMtpvZ+2b2ub3s930zm2FmT0T3fcPMjo7Z3t/Mno7W/r6ZfbONY/9oZpXAVcBDwIlmVmVmP4ju9xUzW2VmW8xsppn1jzmHm9k1ZrYSWGlmp5pZqZl9x8w2Rd/3BWY22cxWRM/x7zHHTzCz182sIrrvvWaW2er8XzOzldF97jMzi9n+FTN7L/rel5rZcft63yKpTgFDpHPMA/LNbHT0F/+lwB9b7fNLoAAYDpxCEEiujG77CjAFOBYoAf6t1bG/AxqBw6P7nAV8ua1C3N0AzKwH8AvgXHfPAz4BLGnnPUwFngJ6A38CnjWzSDQE/QX4FzAAOAO4wczObnXsDKCQIFh9DXjd3Xu6+/fM7HTgTuBi4FBgLfB4q9e/AJgIjIk+PgTIjr7m7cBvgCuA8cDJwHfNbFh03ybgW0AxcGK0xm+0Ov8U4HjgqGgdZ0e/T58Fvk/w75EPnA+Ud/B9i6QsBQyRztPSinEm8B6wvmVDTOi41d23u/sHwH8Dn4/ucjHwM3df5+5bCH4ZtxzbD5gM3ODuO9x9E/DT6Pn2pRkYa2Y57r7R3d9tZ9/F7j7D3RuAewh+uZ9A8Eu5j7vf4e717r6G4Jd97Ou/7u7Punuzu9e0ce7PAQ+7+xvuXgfcStDCMTRmnzvdfUvM8Q3Aj6P1PE4QHn4e/f69CywFjgZw98XuPs/dG6Pf218ThLhYd7l7hbt/CLwCHBN9/svAT9x9oQdWufvaDr5vkZSlvkyRzvMoMAcYRqvuEYJfjhGCv9xbrCX4yxigP7Cu1bYWQ6LHboxp1U9rtf8e3H2HmV0C3AT81sz+Adzo7sv2csi6mGObzaw0WpcD/c2sImbfdGBuW8fuRX/gjZjzV5lZOcH7/2Av5yh396bo/ZbQ8XHM9hqgJ4CZjSQIRSUEY18ygMWtzvdRzP3qlmOBQcDqNmoewr7ft0jKUsAQ6STuvtbM3idobbiq1ebNBH+RDyH4yxtgMLtaOTYS/KIjZluLdUAdUOzujftZ02xgtpnlAD8i+Av85L3svvP1o90DA4ENBF0z77v7iPZeah+lbCB47y3n7wEUEdPK04FztOcB4E3gMnffbmY3sGc3096sAw7by/P7et8iKUtdJCKd6yrg9NZXT0T/En8S+LGZ5ZnZEODb7Bqn8STwTTMbaGa9gOkxx24EXgD+28zyzSzNzA4zs9ZdALsxs35mNjX6y7wOqCLoMtmb8WZ2UfQqjhuix8wDFgDbzewWM8uJDkgda2bHd/i7Ao8BV5rZMWaWBfw/YH60OyMe8oBKoMrMjgC+vh/HPgTcZGbjLXB49N8nHu9bpNtSwBDpRO6+2t0X7WXzdcAOYA3wGsFAyoej234DzCYYUPgG8OdWx34ByCRo/dhKMKDy0H2Uk0YQYjYAWwjGJLT3i/c54JLo+T8PXOTuDdFwNIVgzML7BK0xDxEMWO0Qd38R+C7wNEFrzWHEdyzDTcDlwHaC7+UT+1HbU8CPCf49tgPPAr3j8b5FujNzT4Y5d0QkmZnZ94HD3f2KsGsRka5BLRgiIiISdwoYIiIiEnfqIhEREZG4UwuGiIiIxJ0ChoiIiMSdAoaIiIjEnQKGiIiIxJ0ChoiIiMSdAoaIiIjEnQKGiIiIxJ0ChoiIiMSdAoaIiIjEnQKGiIiIxJ0ChoiIiMSdAoaIiIjEnQKGiIiIxF1G2AXsr+LiYh86dGjYZYh0C4sXL97s7n3CrqMz6TNEJH7a+wzpcgFj6NChLFq0KOwyRLoFM1sbdg2dTZ8hIvHT3meIukhEREQk7hQwREREJO4UMERERCTuutwYDEktDQ0NlJaWUltbG3YpXVp2djYDBw4kEomEXYqIpAgFDElqpaWl5OXlMXToUMws7HK6JHenvLyc0tJShg0bFnY5IpIi1EUiSa22tpaioiKFi4NgZhQVFakVSEQ6lQKGJD2Fi4On76GIdDYFDBEREYk7BQyRdlRUVHD//ffv93GTJ0+moqJiv4+bNm0aM2bM2O/jRESSTbcLGE0Ocyvgg5qwK5HuYG8Bo7Gxsd3jZs2aRWFhYaLKEhGJG2926mubqapoZMvH9Xy0tpZ1y6up2tb+59y+dLurSLY2wKlL4N+HwA81YL5buWElLKmK7zmP6Qk/G7H37dOnT2f16tUcc8wxRCIRsrOz6dWrF8uWLWPFihVccMEFrFu3jtraWq6//nquvvpqYNd01FVVVZx77rmcdNJJ/POf/2TAgAE899xz5OTk7LO2l156iZtuuonGxkaOP/54HnjgAbKyspg+fTozZ84kIyODs846i7vvvpunnnqKH/zgB6Snp1NQUMCcOXPi9S0SkSTkzU5DvVNf10xDbTP1dc3U1zbTEP3a8jh4ztvZFmxvyxmX9WX0hPwDrrHbBYziTDipAJ7brIAhB++uu+7inXfeYcmSJbz66qucd955vPPOOzsv93z44Yfp3bs3NTU1HH/88XzmM5+hqKhot3OsXLmSxx57jN/85jdcfPHFPP3001xxxRXtvm5tbS3Tpk3jpZdeYuTIkXzhC1/ggQce4POf/zzPPPMMy5Ytw8x2dsPccccdzJ49mwEDBhxQ10yYzOwc4OdAOvCQu9/VavsQ4GGgD7AFuMLdS6PbmoC3o7t+6O7nd1rhIgnS1OhUljdQUdZARVk9FWUNbNvcwI7Kxl2Boa4Z2s4Fu0lLg0h2GplZaWRmB7fs3HTye0fIzEojkm07t0Va9ol+7dUv86DeR7cLGABTi+HG1bCmBobv+w9F6SLaa2noLBMmTNhtLolf/OIXPPPMMwCsW7eOlStX7hEwhg0bxjHHHAPA+PHj+eCDD/b5OsuXL2fYsGGMHDkSgC9+8Yvcd999XHvttWRnZ3PVVVcxZcoUpkyZAsCkSZOYNm0aF198MRdddFE83mqnMLN04D7gTKAUWGhmM919acxudwN/cPffm9npwJ3A56Pbatz9mE4tWiQOmpuc7VsbdwaIlhBRsame7Vsb8ZjwkJWbRmGfCL37ZZKZk7ZnIGgjHESyjMzsNNIzLLSryBIWMMzsYWAKsMndx7ax3Qj+apkMVAPT3P2NeLx2S8B4bjN8a1A8zigS6NGjx877r776Ki+++CKvv/46ubm5nHrqqW3ONZGVlbXzfnp6OjU1Bz5AKCMjgwULFvDSSy8xY8YM7r33Xl5++WV+9atfMX/+fJ5//nnGjx/P4sWL9wg6SWoCsMrd1wCY2ePAVCA2YIwBvh29/wrwbKdWKHKAvNnZXtHItrIGKjbHtEaUNVC5pYHmpl37ZmanUVAcod+QbEaVRCgojlDYN5PC4gjZPdLDexMHIZEtGL8D7gX+sJft5wIjoreJwAPRrwftsBwY20MBQw5eXl4e27dvb3Pbtm3b6NWrF7m5uSxbtox58+bF7XVHjRrFBx98wKpVqzj88MN59NFHOeWUU6iqqqK6uprJkyczadIkhg8fDsDq1auZOHEiEydO5K9//Svr1q3rKgFjALAu5nEpe34O/Au4iOAPkguBPDMrcvdyINvMFgGNwF3u3mb4MLOrgasBBg8eHN93ICnN3dlR2RSEiJjujIqyBraVN9DUsKspIiPTKCiOUHRoJocd3TMIEX2CW07P9G43X03CAoa7zzGzoe3sMpWg2dOBeWZWaGaHuvvGeLz+1CK480Mob4AiLb8gB6ioqIhJkyYxduxYcnJy6Nev385t55xzDr/61a8YPXo0o0aN4oQTTojb62ZnZ/PII4/w2c9+ducgz6997Wts2bKFqVOnUltbi7tzzz33AHDzzTezcuVK3J0zzjiDo48+Om61JIGbgHvNbBowB1gPtPztN8Td15vZcOBlM3vb3Ve3PoG7Pwg8CFBSUtKBnmvp7tyd5iaCQZLRW310wGND3a7Bj7HbG+qchvro4Mj6Zuqqm6ksb6Chftd/qbR0osEhkyGjc2NCRCY9CrpfiGiPuSfuZy0aMP53L10k/0vwF8dr0ccvAbe4+6L2zllSUuKLFrW7CwALK2HCG/D7I+ALhxxI9ZIM3nvvPUaPHh12Gd1CW99LM1vs7iUhlYSZnQh8393Pjj6+FcDd79zL/j2BZe4+sI1tvyP4vGl3IpGOfoZI19DU6FRuaaCyvIHK8kZqdjTtOzDUB1deNDd38EWMYEBklhHJSiOSGdzPykknv3cGBdEAUdgnQs/CDNLSUihEtPMZ0iUGeR5I8+b4PBiQCc9uVsAQSWILgRFmNoygZeJS4PLYHcysGNji7s3ArQRXlGBmvYBqd6+L7jMJ+ElnFi+J5+7UVDWxLRogKssboveDx1XbGve4miIjEg0CMYEgKzednoVGJDt4nJmdRiSzZb+03QJEZvS5lscZkfAGSnZlYQaM9UDsCImB0ef2cCDNm2kG5xfD7z+CmibI6ZpjZKSbuuaaa/jHP/6x23PXX389V155ZUgVhcPdG83sWmA2wWWqD7v7u2Z2B7DI3WcCpwJ3mpkTdJFcEz18NPBrM2smmDTwrlZXn0gX0VDfzPYtjTHBIRoitgSBorF+94/9HgXBZZYDDs8hvyiDgqII+dFbbs900tIVBpJBmAFjJnBtdNT4RGBbvMZftJhaDA9sgJe2wpTieJ5Z5ODcd999YZeQNNx9FjCr1XO3x9yfAezR7eHu/wTGJbxAOWjeHAyEbN360HK/envTbvtHMo38oggFRREGjcwJAkTvCPnFEfJ7ZZCR2e0moe6WEnmZ6mMEf3kUm1kp8D0gAuDuvyL4QJkMrCK4TDXuf7qdWgh56fBcuQKGiEhnqalqonRlNeuW17DxgxoqyxtpaoxphTDIK8wgvyjCkDG5u1ogekfIL8rolldUpKJEXkVy2T62O7uaOhMiKw0m94aZm+FXI0GtZiIi8dfY0MzGNbWsWxGEirL1dUAwt0P/w7IZOqZHTDdGBnm9IqRn6AO5u+sSgzwPxtRieKIM5lfCJwrCrkZEpOvzZmfzxnrWLa9m3fJqNrxfS1ODk5YOhwzNZuK5vRk0Kpe+A7M0HiKFdfuAcW5vyLBg0i0FDBGRA7N9awPrVtSwbnk1pStrqKkKxk307pfJ2BPzGTQyl/6H55CZpfEREuj2AaMwEozFeG4z/OdhYVcjXU1FRQV/+tOf+MY3vrHfx/7sZz/j6quvJjc3d6/7tKy6WlysQUKSXOprm1m/qoYPl1ezbkU1FZsaAMjNS2fwqFwGjcph4MhcehZ0+18jcoBS4n/GBcVw7UpYtgOO6LHv/UVaVFRUcP/99x9wwLjiiivaDRgiyaKpydm0tjZopVhRzUdra/HmYE6J/oflcGS0laLo0EwNwJQOSYmAcX5REDCeK1fA6Mpu4K8s4aO4nvMYDuFnnLvX7dOnT2f16tUcc8wxnHnmmfTt25cnn3ySuro6LrzwQn7wgx+wY8cOLr74YkpLS2lqauK73/0uH3/8MRs2bOC0006juLiYV155ZZ+13HPPPTz88MMAfPnLX+aGG25o89yXXHIJ06dPZ+bMmWRkZHDWWWdx9913x+17IqnB3akoawjGUayooXRlNQ11DgZ9B2Vx3Om9GDQqh0OH5mhAphyQlAgYg7LhuJ5BN8ktWudI9sNdd93FO++8w5IlS3jhhReYMWMGCxYswN05//zzmTNnDmVlZfTv35/nn38eCBZBKygo4J577uGVV17pUPfH4sWLeeSRR5g/fz7uzsSJEznllFNYs2bNHucuLy/nmWeeYdmyZZgZFRUVCf0eSPfR1Oh8sHQHHyzdwbrlNVRVNAKQ3zuDkcflMWhkLgNH5HTZ1TsluaREwIDgapLvfwAf1cEhWfvcXZJQey0NneGFF17ghRde4NhjjwWgqqqKlStXcvLJJ3PjjTdyyy23MGXKFE4++eT9Pvdrr73GhRdeuHM5+Isuuoi5c+dyzjnn7HHuxsZGsrOzueqqq5gyZQpTpkyJ6/uU7mfrx/W8O6+SZQsrqd3RTFZOGgNH5FByZi8GjQwW5BKJt5QJGBcUw/c+gL+Uw1f6h12NdEXuzq233spXv/rVPba98cYbzJo1i9tuu40zzjiD22+/vY0z7L+RI0e2ee4FCxbw0ksvMWPGDO69915efvnluLyedB8N9c2sWlLF0nmVbHy/lrQ0GDa2B2NOyGfQqNyUWpBLwpEyAWNcDxiaHXSTKGBIR+Xl5bF9+3YAzj77bL773e/yuc99jp49e7J+/XoikQiNjY307t2bK664gsLCQh566KHdju1IF8nJJ5/MtGnTmD59Ou7OM888w6OPPsqGDRv2OHdVVRXV1dVMnjyZSZMmMXz48IR+D6RrKSut491521ixuIr62mYK+0T4xKeLOOL4PHLzUuYjX5JAyvxvM4OpRfCrDVDVCD1T5p3LwSgqKmLSpEmMHTuWc889l8svv5wTTzwRgJ49e/LHP/6RVatWcfPNN5OWlkYkEuGBBx4A4Oqrr+acc86hf//++xzkedxxxzFt2jQmTJgABIM8jz32WGbPnr3Hubdv387UqVOpra3F3bnnnnsS+02QpFdX08TKN6p4d14lZaV1pEeMw4/qyZgT8+k/PFtXfUgoLJixu+soKSnxRYsWHdCxr2yF0/8FTx8JF/WJc2GSEO+99x6jR48Ou4xuoa3vpZktdveSkEoKxcF8hiQTd+ej92t5d34lq5ZU0VjvFPXP5MgT8hk5Po/sXA3UlMRr7zMkpf6OP7kAemXAs5sVMESka6qpamLZokqWzqtk68cNRLKMUePzGHNCPn0HZam1QpJGSgWMjDSYUgTPl0Njc/BYpDNMnDiRurq63Z579NFHGTdOq43Lvnmzs25lDUvnVbLm7Sqam4I1P06/tBeHH9NT03NLUkqpgAHB5aqPfgyvbYNTe4VdjaSK+fPnh12CdEFVFY28t6CSpfMr2b6lkazcNMZNKmDMCfkUHarr7SW5pVzAOLsXZFnQTaKA0TW4u5p9D1JXG2uVypqbnLXvVfPu69tY+1417jBwRA4nnlfE8HE9yIiotUK6hpQLGD0z4FO9gmnDf3p4cHWJJK/s7GzKy8spKipSyDhA7k55eTnZ2dlhlyLt2La5gaXzK3lvQSXVlU3k5qVz3Bm9GDMxXxNhSZeUcgEDgm6S51fA2zvgqJ5hVyPtGThwIKWlpZSVlYVdSpeWnZ3NwIEDwy5D2rBhdQ0LZm+hdGUNZjBkdC5jTshn6JgepKUrVEvXlZIB49NFYASTbilgJLdIJMKwYcPCLkMk7uprm3n9+XLefm0bPQszmHhub0ZPyKdnYUp+LEs3lJL/kw/JghPyg3EY3x0adjUikmrWLtvBq0+Wsb2ikaM/WcAJk4uI6EoQ6WZSMmBA0E0yfQ2sqw1WWxURSbTaHU289txmli3cTq++ET5z3QAOHZYTdlkiCZG6AaMoCBgzy+GaAWFXIyLd3eq3qvj7jDJqdjQx/lO9OP6sXroiRLq1lA0YR/SAkTnBOAwFDBFJlOrtjcx5ejOr/lVF8YBMPn11f/oM1BwW0v2lbMCAYAn3e0qhogEKdRWYiMSRu7NicRVznimjoa6ZEyb35tjTe5GuK0MkRaR0wJhaDD9ZB3/dApf1C7saEekuqioaeeWpTaxdWh2d0rsvvftlhl2WSKdK6YAxMR/6RoJuEgUMETlY3uy8O6+Sf/6lnOZm56QLijnq5ALS0tRqIaknoQHDzM4Bfg6kAw+5+12ttg8BHgb6AFuAK9y9NJE1xUq3YE6MJ8ugrhl0lZiIHKhtmxt4+YlNrF9Vw8AROZx2cV/NwCkpLWG/Us0sHbgPOBcYA1xmZmNa7XY38Ad3Pwq4A7gzUfXszQXFsL0JXq3o7FcWke6gudlZ8moFj/3kQ8pK6zjtkj5M/Xp/hQtJeYlswZgArHL3NQBm9jgwFVgas88Y4NvR+68Azyawnjad0Qty04JukrN7d/ari0hXtuWjel56/GM+XlvH0DG5nPrZvpqJUyQqkZ0CA4B1MY9Lo8/F+hdwUfT+hUCemRW1PpGZXW1mi8xsUbzXpMhJD4LFzM2gBSdFpCOampyFL2zh8bs/ZNvmBs68oh/nfflQhQuRGGGPOrgJOMXM3gROAdYDTa13cvcH3b3E3Uv69OkT9yKmFsP6eli8Pe6nFpFupqy0jqd+uo75f93CYUf15PJbBjNqfJ5W+xVpJZFxez0wKObxwOhzO7n7BqItGGbWE/iMu3f6aIgpRUHSenYzlOR39quLSFfQ2NDMwhe28sbLW8npmc7kLx3C8HFaLVFkbxIZMBYCI8xsGEGwuBS4PHYHMysGtrh7M3ArwRUlna4oAicXwHPl8KPhYVQgIsls4/s1vPz4JrZuamD0xDwmnV9Mdm562GWJJLWEdZG4eyNwLTAbeA940t3fNbM7zOz86G6nAsvNbAXQD/hxourZl6nF8M4OWFMTVgUikmzq65qZ8+cynv7lehobnPO/1p8zLu2ncCHSAQkdkeTus4BZrZ67Peb+DGBGImvoqKnF8O3VwdUk3xq07/1FpHtbt7yal5/cxPatjRx1UgEnnFdEpibLEekwDXmOGp4D43oE4zAUMERSV31tM689t5ml8yop7BPhomsH0H+4llQX2V8KGDGmFsP/Wwub66FYywaIpCR3Z93yao47vZAJZ/cmI1OtFiIHQj85MaYWQTPw/JawKxGRsGTlpHP59MF84tPFChciB0E/PTHG58GAzKCbRERSV0TBQuSg6acohlnQTfLCFqjZY7ovERER6SgFjFamFkN1M7y4NexKREREui4FjFZOLYT89OByVRERETkwChitZKbB5CKYWQ5NWvxMRETkgChgtGFqEZQ1wLzKsCsRERHpmhQw2nBuEURM3SQiIiIHSgGjDQUZwViMZzeDq5tERERkvylg7MUFxbCyBpZVh12JiIhI16OAsRfnFwVf1U0iIiKy/xQw9mJgNozvCc+Vh12JiIhI16OA0Y6pxTC/EjbWhV2JiIhI16KA0Y4LisGBv6gVQyShzOwcM1tuZqvMbHob24eY2Utm9paZvWpmA2O2fdHMVkZvX+zcykVkbxQw2jG2BwzL1jgMkUQys3TgPuBcYAxwmZmNabXb3cAf3P0o4A7gzuixvYHvAROBCcD3zKxXZ9UuInungNGOlsXPXtoKVY1hVyPSbU0AVrn7GnevBx4HprbaZwzwcvT+KzHbzwb+5u5b3H0r8DfgnE6oWUT2QQFjH6YWQZ3DbC1+JpIoA4B1MY9Lo8/F+hdwUfT+hUCemRV18FgRCYECxj6cVAC9M4JJt0QkNDcBp5jZm8ApwHqgqaMHm9nVZrbIzBaVlZUlqkYRiaGAsQ8ZaTClCJ4vh4bmsKsR6ZbWA4NiHg+MPreTu29w94vc/VjgP6LPVXTk2Oi+D7p7ibuX9OnTJ971i0gbFDA6YGoxbG2E17aFXYlIt7QQGGFmw8wsE7gUmBm7g5kVm1nL59WtwMPR+7OBs8ysV3Rw51nR50QkZAoYHXB2b8hO09UkIong7o3AtQTB4D3gSXd/18zuMLPzo7udCiw3sxVAP+DH0WO3AD8kCCkLgTuiz4lIyDLCLqAr6JEOn+oVjMP46eHB1SUiEj/uPguY1eq522PuzwBm7OXYh9nVoiEiSSKhLRgdmDxnsJm9YmZvRifQmZzIeg7G1CJYWwdv7Qi7EhERkeSXsIDRwclzbiNoDj2WoN/1/kTVc7A+XQyGuklEREQ6IpEtGB2ZPMeB/Oj9AmBDAus5KP0y4cR8Xa4qIiLSEYkMGB2ZAOf7wBVmVkrQ/3pdAus5aFOL4c0q+LA27EpERESSW9hXkVwG/M7dBwKTgUdjLkXbKVkmyZlaHHydqVYMERGRdiUyYHRkApyrgCcB3P11IBsobn2iZJkkZ1QujMqB57S6qoiISLsSGTD2OXkO8CFwBoCZjSYIGEk9j+8FxfBqBVQ0hF2JiIhI8kpYwOjg5Dk3Al8xs38BjwHT3N0TVVM8TC2GRodZmspHRERkrxI60VYHJs9ZCkxKZA3xNjEf+kWCy1Uv7xd2NSIiIskp7EGeXU6aBXNi/HUL1GnxMxERkTYpYByAC4phexO8sjXsSkRERJKTAsYBOKMQeqTpahIREZG9UcA4ANnpwQqrMzdDc1IPSRUREQmHAsYBmloMG+ph8fawKxEREUk+ChgH6LwiSEdrk4iIiLRFAeMAFUXg5EKtrioiItIWBYyDMLUI3q2G1TVhVyIiIpJcFDAOQsviZ2rFEBER2Z0CxkEYlgNH9dA4DBERkdYUMA7SZ/vA3G1w3lvwdlXY1YiIiCQHBYyD9J3B8JPh8M9KOHoRXLkM1tWGXZWIiEi4FDAOUmYa3DwYVk+Ebw+EP30MIxfA9NVa0l1ERFKXAkac9I7A3YfDiolBt8lP1sFh8+GedVoUTUREUo8CRpwNyYY/jIY3xkNJHty4Go5YAP/zsaYVFxGR1KGAkSDH5MHso+FvR0GvDLjiPShZDH/bEnZlIiIiiZcRdgHd3ad6w6Je8NgmuO19OOstOKsX/OfwIISIiIh0Ondo3gKNGzr7bJIAACAASURBVKFxQ6uv0ftFt0PPsw/4JRQwOkGawef6wb/1gfvXw4/WwnGLg+d+NCzoVhERETlo3gxN5XuGhdb3mzaC1+95fFoBZBwa3MwOqhQFjE6UlQbfGgRXHgJ3fQg/Xw9PboLrBsC/DwkGioqIiOzBm6GpbC8tDrEBYiPQuOfxaYWQ0T8IDrmfjIaI/rvCRMv9tNy4layAEYLCCNx1GFwzAL73AdxTCr/9CG4dHISNnPSwKxQRkU7lTUE4aPgAGtZGv8bcGj/cS4tDb4j0h/RDIXf0noEhoz9kHAJpOZ36dkABI1SDsuHhI+BbA2H6GrhlDdy7Hn44DK7oB+kH1zolIiLJwpuClobdgkNskPgQaDV5Uno/iAyB7OMgchFEBu0eHNIPgbSszn4nHaaAkQTG9YTnj4JXt8J31sC0ZfDf64KBoOf0PuhuMBERSTRvbCNAfBATJD5kj66L9EMgMhSyj4e8zwZhIjI0ehsc1+6KMChgJJFTe8H84+CpMrh1DUx+G04vhJ8cBuN1xYmISLjcoekjqH0L6t6G+qXQ8H40RKwDmnbfP6N/EBZyToD8S2LCw1DIGAxp3XuEvwJGkjGDi/vCBcXw6w1wx9pg/oxL+8KPh8Hwzu9GExFJPc1VUPdOECRib03lu/bJOBQiwyFnEuQPjQkQQ6IBInm7LzpDQgOGmZ0D/BxIBx5y97tabf8pcFr0YS7Q190LE1lTV5GZBtcNhC8eAj/5MBgI+nQZfL1/sMDagNT+fysiEh/eCPUr9wwSDWt27ZPWEzLHQs+LIGscZB8FWWMhvSi8uruAhAUMM0sH7gPOBEqBhWY2092Xtuzj7t+K2f864NhE1dNV5WfAj4bDNwbA9z8IBoH+cj2cXACX9IXP9IF+mWFXKSKS5NyDqzRaB4n6peB10Z3SIXMkZJdAwZWQdVQQKCJDwDTx9f5KZAvGBGCVu68BMLPHganA0r3sfxnwvQTW06X1z4IHR8F3BsGfNsETm+CalXDdSjitMAgbF/WBIs2lISKprq3ujdq3gpkrW2T0D8JDjzN2BYnMI7r9uIjOlMiAMQBYF/O4FJjY1o5mNgQYBrycwHq6hcNz4fahwe2dKniiLAgbV6+Ab6yET/WCS/oEYzgKFTZEpLtrroHaxVAzD2rnQe0bwcDLFi3dG3mfiXZtjFP3RidJlkGelwIz3L2prY1mdjVwNcDgwYM7s66kNrZncLtjKCypCoLGE2Vw5XL46go4u3fQsnF+EeQly7+0dFtm9mfgt8Bf3b057HqkG3IPwkPNPKh5PRoolrDz8s/IYcElnwVXRYOEujfClMhfO+uBQTGPB0afa8ulwDV7O5G7Pwg8CFBSUqJFz1sxg2Pzgtudw2Hh9iBsPFkGfymH7DSYHA0b5xVBD80UKolxP3Al8Aszewp4xN2Xh1yTdGXNVVCzKAgSNa8HwaJpU7DNekDOBCi6GbJPhJyJkNE33HplN4kMGAuBEWY2jCBYXApc3nonMzsC6AW8nsBaUoYZTMgPbv91GLxeGYSNp8rgz5shNw0+XRSEjXN7Q7bChsSJu78IvGhmBQRjql40s3XAb4A/untDuyeQ1OYODSt3tU7UzIO6t4BoY1jmKOhxLuScGMwrkXUkmJpmk1nC/nXcvdHMrgVmE1ym+rC7v2tmdwCL3H1mdNdLgcfdXS0TcZZmMKkguP30cJhbEXShzCgLvualw9TiYMzGWb2DS2NFDoaZFQFXAJ8H3gT+BzgJ+CJwaniVSdJpqoTaBbvCRM28XYMw0/IheyIU/Uc0UEzQmIkuyLra7/WSkhJftGhR2GV0aY3N8EpF0LLx582wtREKM+DC4qBl4/RCiChspAQzW+zuJXE61zPAKOBR4HfuvjFm26J4vc7B0mdICLwZ6pftPnai7l3AAYPMMUGrREvrROZojZvoItr7DFH7UgrKSIMzewe3+5vhxa1B2Hi6DB75CIoygvk1LukLpxRq0TXpsF+4+yttbUiWcCGdqGkLVD4JVc9GWye2Bc+n9QpCRN5ng0CRPQHSC8KtVRJCASPFZabB5KLgVtsEs6Nh438+hgc3QnEEPhntZjmpAI7tqdYN2asxZvamu1cAmFkv4DJ3vz/kuqSzeD1U/RW2/QF2/G/wOHMU5F8K2dEWiswRap1IEQoYslN2dEzG1GKoboJZ5cFVKK9tC7pSAHLSYGI+TMoPQseJ+ZpvQ3b6irvf1/LA3bea2VcIri6R7sodahcGoWL748FaHel9ofAbUPAFyDpGS0KnKAUMaVNuOvxb3+AGsLEO/rEN/lEZBI67PgzWDTRgbI+gdaOllWNwlj5PUlS6mVnLgO3ocgGayL67algL2/4IlX+A+hVg2dBzahAqepwJpr88Up0ChnTIoVm7B44dTTC/Mggdr22DP34MD2wItg3I3BU2JhXAUT2CcR/S7f0f8ISZ/Tr6+KvR56S7aNoG22fAtkeh5u/BczmnwCHfgbx/01gK2Y0ChhyQHulweq/gBtDk8HbVrhaOf2wLJvoC6JkOJ0S7VU4qCLpYNLNot3QLQaj4evTx34CHwitH4sIbYMcLQaioeg68NlgQrPhHkP85yBwadoWSpPQxL3GRbnBMXnC7ZkDw3Ie10W6VaCvHHWuDi9LSgGN67t7KoeXnu77o9OAPRG/SlblD3ZtBqKj8UzB7ZlrvYArugs8HV36oH1T2QQFDEmZwdnC7rF/weFsjzIvpVvntxmDpeYAhWUHYmJAP43oEt2L13ncpZjYCuBMYA+xcktLdh4dWlOyfhlKo/J8gWNS/C5YJPT8N+Z+HnucGj0U6qEMBw8yuBx4BthM0eR4LTHf3FxJYm3QzBRnBAmxn9w4eNzQHi7S1BI6XKuB/Nu3a/9DMIGgc1TP6tQeM7gFZGs+RrB4Bvgf8FDiNYF0S/Wslu+Yq2P7n4CqQ6pcBh5xPQL8HIP9iSO8ddoXSRXW0BeNL7v5zMzubYN2QzxPM1qeAIQcskgbH5we3GwYFrbIf1cPbO+CtquDr2zvgF6VQH51wNh0YlRtt5egZhI5xPWBItlpsk0COu78UvZJkLfB9M1sM3B52YdKKN0H1S0FLxfY/g1dDZBgU3Q4FV0Dm4WFXKN1ARwNGy0f3ZODR6Joi+jiXuDILrlY5NCtYG6VFYzOsrIG3dgQDSd/aAfO3B+uptMhPDy6XjW3tGNtDc3R0sjozSwNWRtchWg/07MiBZnYO8HOCDPmQu9/Vavtg4PdAYXSf6e4+y8yGAu8BLau2znP3r8XhvXRPTdthy09g28PQuAHSCoNAkf+FoNVCH+sSRx0NGIvN7AVgGHCrmeWxc4k7kcTKSAu6Rkb3CKYvb1HZCO/s2L3F4/FN8KvGXfsMyoq2csS0dozK1WykCXI9kAt8E/ghQTfJF/d1UHS+jPuAM4FSYKGZzXT3pTG73QY86e4PmNkYYBYwNLpttbsfE7d30R25B2Mryr4DjRuhxxTo+3PoOQXSsvd9vMgB6GjAuAo4Bljj7tVm1pugf1UkNPkZ8ImC4NbCHUrrdg8db+0IpkBvjHazRAxGR7tZRubCiJzoLTcYJyL7LxoSLnH3m4Aq9u/zYQKwyt3XRM/1ODAViA0YDuRH7xcAGw666FRRuwQ+vhZq/gHZx8OAZyBnYthVSQro6MfpicASd99hZlcAxxE0Z4okFTMYlB3cJses7lzfDMurd+9mmbNt90GlAH0iuweOlvuH52jujva4e5OZnXSAhw8A1sU8LgVa/wb8PvCCmV0H9AA+FbNtmJm9CVQCt7n73AOso3tp2gJlt0HFr4OBmoc8BAVXah0Q6TQd/ch8ADjazI4GbiS4kuQPwCmJKkwknjLTgm6ScT2Bfruer2mC1TXBGI+dt2r421b4/ce7n+OQzJjwEQ0dI3KDrz3SO/XtJKs3zWwm8BSwo+VJd/9zHM59GcES8P9tZicCj5rZWGAjMNjdy81sPPCsmR3p7pWxB5vZ1cDVAIMHD45DOUnMm6DiIdj8H9C0FXpdA8U/gPReYVcmKaajAaPR3d3MpgL3uvtvzeyqRBYm0hly0mFsz+DW2o7Y8FG9K4DM2hJc7RKrf+aerR4jcuCwnOA1UkQ2UA6cHvOcA/sKGOuBQTGPB0afi3UVcA6Au79uZtlAsbtvAuqizy82s9XASGBR7MHu/iDwIEBJSYnvx3vqWqr/CR9fB3VvQM4nod8vIfuosKuSFNXRgLHdzG4luDz15OhIcY3Pl26tR3pwVcpRbYSP7Y2wqlWrx6oamLkZNjXsvu+grF1hY3BWcEnt4OxgcrEBWd1nwKm7H+i4rIXACDMbRhAsLgUub7XPh8AZwO/MbDRBmCkzsz7AlmgXzXBgBLDmAOvouho3wqZboPJRyBgA/R+DvEt0VYiEqqMB4xKCH/gvuftH0UvG/itxZYkkt7wMODYvuLW2rSV8VO/e9TJzM3zcKnykAf2zYoJHqwAyODsYzNoVmNkjBC0Wu3H3L7V3nLs3Ri9rnU1wCerD0Uvh7wAWuftMgq7Z35jZt6KvMS3aqvpJ4A4zayC4su1r7r4lvu8siXkDbPkFlP8AvA6KboWif4e0Dl0dLJJQFl1Zed87mvUDjo8+XBBtmux0JSUlvmjRon3vKJKEaptgXR2srYUPW3+N3m9o9SNZmNF+ADkkE9IO8A9VM1vs7iUH/87AzD4T8zAbuBDY4O7fjMf546XbfIbs+Bt8/E2oXwY9JkO/n0HmiLCrkhTT3mdIR6cKv5igxeJVgkm3fmlmN7v7jLhVKZICstOj4zRy297e7PBx/d4DyNxtUNG4+zERC7phWgeQ0wthWE7i31MLd3869rGZPQa81nkVpIiGtfDxt6HqzxA5DAb+JZjPQiTJdLTx9T+A41taLaL9ni8CChgicZQWM5vpCXvZp7IxCBtr66JfY0LISxWwoS7oK/jj6M4NGG0YAfTd517SMc01wSyc5XcBFiyX3vtGTZQlSaujASOtVZdIOVrESCQU+Rl7v/IFgkXk1tdBr04ehm1m29l9DMZHwC2dW0U35A5Vz8Gmb0HDB5B3MfS9GyKD9nmoSJg6GjD+z8xmA49FH19CMFWviCSZSBoMDaHlwt3bGPIqB6VuOWz6Jux4ATKPhEEvQ4/Twq5KpEM61Arh7jcTXEN+VPT2oLvrLxMR2cnMLjSzgpjHhWZ2QZg1dVlN22HTd+D9cVAzD/r+DIa9qXAhXUqHuznc/Wl3/3b09kxHjjGzc8xsuZmtMrPpe9nnYjNbambvmtmfOlqPiCSd77n7tpYH7l4BfC/Eeroed9j2R3h/FGz5r2Cl0+EroPf1YJp6SLqWdrtI2uhT3bkJcHfPb2Nby7H7XCHRzEYAtwKT3H2rmWlAmEjX1dYfLF1kFo8koEXJpJtp94f/IPtUO7JC4leA+9x9a/T1QplbQ0TiYpGZ3UPwhwXANcDiEOvpGpqrglk4K36lRcmkW0nk/+C2Vkgc0GqfkcBIM/uHmc0zs3MSWI+IJNZ1QD3wBPA4UEsQMqQ95f8FFQ8Ei5INXwGFVylcSLcQdvNlBsG18qcSLHA0x8zGRftud0qplRBFuih33wG0OdZK2lH9CmSXQL9fhF2JSFwlMiZ3ZIXEUmCmuze4+/vACoLAsRt3f9DdS9y9pE+fPgkrWEQOnJn9zcwKYx73il7eLnvTXAu1CyDn5LArEYm7RAaMnSskmlkmwQqJM1vt8yxB6wVmVkzQZZJ6KyGKdA/Fsa2P0bFVGrjdntqFwSJluZ8MuxKRuEtYwHD3RqBlhcT3gCdbVkg0s/Oju80Gys1sKfAKcLO7lyeqJhFJqOboSssAmNlQ2r4KTVpUzw2+5p4Ubh0iCZDQMRjuPotWM366++0x9x34dvQmIl3bfwCvmdnfCS5lP5no2CnZi5o5wQyd6UVhVyISdxqqLCJx4e7/B5QAywmWFbgRqAm1qGTmjVDzT8jV+AvpnsK+ikREugkz+zJwPcGA7iUEC8K+DpweZl1Jq+5f0Lxd4y+k21ILhojEy/XA8cBadz8NOBaoaP+QFNYy/kJXkEg3pYAhIvFS6+61AGaW5e7LgFEh15S8qudAZBhEBoZdiUhCqItEROKlNDoPxrPA38xsK7A25JqSkzvUzIUek8OuRCRhFDBEJC7c/cLo3e+b2StAAfB/IZaUvOqXQdNmjb+Qbk0BQ0Tizt3/HnYNSa2mZf4Ljb+Q7ktjMEREOlv1HEjvB5E9VkYQ6TYUMEREOlv13KD1wizsSkQSRgFDRKQzNayFxg8hR+MvpHtTwBAR6UzVGn8hqUEBQ0SkM1XPgbQCyBoXdiUiCaWAISLSmWrmQs4ksPSwKxFJKAUMEZHO0rgpmAND819IClDAEBHpLDWvBV+1/oikAAUMEZHOUj0HLBtySsKuRCThFDBERDpL9VzIOREsM+xKRBJOAUNEpDM0VULdEnWPSMpQwBAR6Qw1/wSaNcBTUoYChohIZ6iZA2RAzglhVyLSKRQwREQ6Q/VcyB4PaT3CrkSkUyhgiIgkWnMt1C7Q9OCSUhQwREQSrXYBeL0WOJOUooAhIpJo1XOCr7mTwq1DpBMpYIiIJFrN3GBxs/TeYVci0mkSGjDM7BwzW25mq8xsehvbp5lZmZktid6+nMh6REQ6nTcGl6hq/gtJMRmJOrGZpQP3AWcCpcBCM5vp7ktb7fqEu1+bqDpEREJVuwSaqzT/haScRLZgTABWufsad68HHgemJvD1RESST010/IVaMCTFJDJgDADWxTwujT7X2mfM7C0zm2Fmg9o6kZldbWaLzGxRWVlZImoVEUmM6rkQOQwi/cOuRKRThT3I8y/AUHc/Cvgb8Pu2dnL3B929xN1L+vTp06kFiogcMG8OBnhq/gtJQYkMGOuB2BaJgdHndnL3cneviz58CBifwHpERDpX/TJoKtf8F5KSEhkwFgIjzGyYmWUClwIzY3cws0NjHp4PvJfAekREOtfO+S/UgiGpJ2FXkbh7o5ldC8wG0oGH3f1dM7sDWOTuM4Fvmtn5QCOwBZiWqHpERDpdzVzIODQYgyGSYhIWMADcfRYwq9Vzt8fcvxW4NZE1iIiEwj1owcg5GczCrkak04U9yFNEpHtqWAuNpZr/QlKWAoaISCJo/gtJcQoYIiKJUD0X0goha2zYlYiEQgFDRCQRauZAzklg+piV1KT/+SISug4sjDjYzF4xszejM/9Ojtl2a/S45WZ2dudWvheNH0P9Co2/kJSW0KtIRET2pYMLI94GPOnuD5jZGIKr04ZG718KHAn0B140s5Hu3tS576KV6rnBV81/ISlMLRgiEraOLIzoQH70fgGwIXp/KvC4u9e5+/vAquj5wlUzFywXso8LuxKR0ChgiEjYOrIw4veBK8yslKD14rr9OLbzVc+BnBPAMsOuRCQ0Chgi0hVcBvzO3QcCk4FHzTo+erJTV2Ru2gZ1/9L4C0l5ChgiErZ9LowIXAU8CeDurwPZQHEHj+3cFZlr/gG45r+QlKeAISJh2+fCiMCHwBkAZjaaIGCURfe71MyyzGwYMAJY0GmVt6V6DpARdJGIpDBdRSIioergwog3Ar8xs28RDPic5u4OvGtmTwJLCRZNvCb0K0hq5kJ2CaTlhlqGSNgUMEQkdB1YGHEpMGkvx/4Y+HFCC+yo5hqoWQi9vxV2JSKhUxeJiEi81MwHGjT/hQgKGCIi8VMzBzDIabOxRSSlKGCIiMRL9VzIGgfpvcKuRCR0ChgiIvHgDVDzT8jR/BcioIAhIhIftW+CV2v8hUiUAoaISDxUzwm+KmCIAAoYIiLxUTMXIodDxqFhVyKSFBQwREQOljcHAzy1/ojITgoYIiIHq34pNG/V+iMiMRQwREQO1s7xF2rBEGmhgCEicrCq50JGf4gMC7sSkaSR0IBhZueY2XIzW2Vm09vZ7zNm5mZWksh6RETizj2YwTPnk2AWdjUiSSNhAcPM0oH7gHOBMcBlZjamjf3ygOuB+YmqRUQkYRreh8YNujxVpJVEtmBMAFa5+xp3rwceB6a2sd8Pgf8EahNYi4hIYmj8hUibEhkwBgDrYh6XRp/bycyOAwa5+/PxfOFGmuJ5OhGRvauZC2m9IXOPBlqRlBbaIE8zSwPuAW7swL5Xm9kiM1tUVlbW7r7V1DOOB/ghf6eOxjhVKyKyF9VzIPckMI2ZF4mVyJ+I9cCgmMcDo8+1yAPGAq+a2QfACcDMtgZ6uvuD7l7i7iV9+vRp90VraOQo+nE7rzCO+/kbqw/2fYiItK1xIzSs0gJnIm1IZMBYCIwws2FmlglcCsxs2eju29y92N2HuvtQYB5wvrsvOpgXLSKXJ/gss7kCB87iUS7hKdZTeTCnFRHZU/Xc4KsGeIrsIWEBw90bgWuB2cB7wJPu/q6Z3WFm5yfqdVucxeG8zde5g9N4jmUcwb38lNc1PkNE4qd6LlgPyD427EpEkk5GIk/u7rOAWa2eu30v+54a79fPJsJ3OYXLGcd1zOLbzOZ3LOEBzuMTDI73y4lIqqmZAzkngkXCrkQk6aTEqKTD6M3zfI6nuZgt1DCJh/kyz7GZHWGXJiJdVdNWqHtbl6eK7EVKBAwAw7iIMbzHNdzMJ/g9/2IU9/IQi2mmOezyRKSrqfkH4FrgTGQvUiZgtOhJFj/hLN7kqxxJH77CX5jEwyxhY9iliUhXUj0XiEDOxLArEUlKKRcwWoylH3/nSn7PBaxmC+N5kBv4K5WaUFREOqJ6DuQcD2k5YVcikpRSNmBA0G3yBY5hGddyNeP5BfM5gnt5gndwPOzyRCRZNVdD7SLNfyHSjpQOGC16k8sDTGEeX+ZQ8riUGZzFo6xgc9iliUgyqpkHNGr+C5F2KGDEmMBAFvAV7mUyC1nPOB7gu7xMDQ1hlyYiyaRmLmCQMynsSkSSlgJGK+mkcQ0TWMa1XMyR/Ig5HMl9zGJF2KWJSLKongNZR0N6QdiViCQtBYy9OIQ8HuUiXuaLZJHBefyJi3icD6kIuzQRCZPXQ83rmv9CZB8UMPbhNIbxL77GnZzB/7GK0dzHT3iNeq3UKpKaat8Ar9H8FyL7oIDRAZlkMJ2TeY9rOZPh3MKLHMuv+TsfhF2aiHQ2LXAm0iEKGPthCIU8y2XM5DKqaeBUfscXeYaPqQq7NBHpLNVzIHMkZPQLuxKRpKaAcQA+zSje5Rv8ByfzGG8znJ9zGTN4jmXUqetEpPvyZqh5TfNfiHRAQldT7c5yyeRHnMEVHMVPmcfTLOVx3qGALC5iNJcyltMZRgbpYZcqIvFS9w40V6h7RKQDFDAO0hH04dd8mnuZzIus4XHeYQZLeYQl9CGXz3IklzKWSQwiTQ1GIl1bTXT8hVowRPZJASNOIqRzLiM4lxH8min8lVU8zjs8wpvcz0IGkMcljOUyxjKe/hgWdskisr+q50DGQIgMCbsS2U8NDQ2UlpZSW6v1pg5EdnY2AwcOJBKJdPgYBYwEyCbChYzmQkaznTr+wnIe5x1+yXzu4XUOoxeXMpZLGctYNFBMpEtwD1owck8D0x8IXU1paSl5eXkMHToU07/ffnF3ysvLKS0tZdiwYR0+TgEjwfLI4nKO4nKOYis1PMN7PMY73Mlr/Ji5jKUvlzKWSziSwykKu1wR2ZuG1dC4UfNfdFG1tbUKFwfIzCgqKqKsrGy/jlPA6ES9yOFLHMeXOI6PqWJGdGDobbzMbbxMCf13ho2BaApikaSyc/4Ljb/oqhQuDtyBfO806jAk/ejJNUxgLl9iLTfwX5yJ49zECwzip3ySh7mfBWzSHBsiyaFmDqQXQebosCsR6RIUMJLAYAq5iUks4qus4Dp+yGmUU8M1zKI//83ZPMojvEkFNWGXKpK6qucG3SP6K1gOQEVFBffff/9+Hzd58mQqKrrmGlgKGElmBEXcxim8wzd4i69zCyexii18iefox91M5TEeYjErKcfxsMsVSQ0NG4IxGJr/Qg7Q3gJGY2P7kzPOmjWLwsLCRJWVUBqDkaQMYxz9GEc/fsTpLGIDj/MOT/AOM1kOwKH05JMM+f/t3XlwVOWax/Hvk87WWQgJSVjCKijgwqKERWTYLgOUXlcEL+JcrJGIC+J4y0FH8A5XvUVNUV61yopsDoKoKIpYI3qZyCqLBhlUtssmSBBjAIEkhADhmT+6abuzQIDT6XTn+VSluvuc06efE5KXX95zzvvSn7b0pw2dybDbX40JBhv/IrIUPgmnNju7z/hu0PSVGlc/88wz7Nmzh27duhETE0N8fDypqans2LGDnTt3cuedd3LgwAFOnTrFxIkTycnJAaBt27Zs3LiRkpIShg8fzi233MK6devIyspiyZIluN3uaj9v1qxZzJw5k9OnT9OhQwfmz59PQkIChYWFjB8/nr179wKQm5vLzTffzLx585g+fToiQpcuXZg/f/4Vf0ssYIQBQcgmi2yymM4/8w8Os4r9rGIfq9jPQrYCkEFCQOC4nkwb3MsYJ5xcDVFJnv9EjLkM06ZNY8uWLWzevJmVK1dy6623smXLFt9tn2+++SZpaWmUlZWRnZ3NPffcQ5MmgXcW7tq1i3fffZdZs2YxcuRIPvzwQ8aMGVPt5919992MGzcOgMmTJzNnzhwmTJjAE088Qf/+/Vm8eDEVFRWUlJSwdetWXnzxRdatW0d6ejpHjx515JiDGjBEZBjwKuACZqvqtErrxwOPARVACZCjqtuCWVO4E4ROZNCJDB6mB4qyh6MBgeNDtgOQhpt+tPYFjq40w2WBw5hLV7YG3DeD2N9kEeECPQ11pWfPngFjSrz22mssXrwYgAMHDrBr164qAaNdu3Z06+YJuTfddBP79u2rcf9btmxh8uTJHDt2jJKSEoYOHQrA8uXLmTdvHgAul4uUlBTmzZvHvffeS3p6OgBpaWmOHGPQfltExAW8DgwBCoB8EfmkUoB4R1Xf8G5/O/AyMCxYNUUiQehAEzrQhH/lRgD27g4jawAADy5JREFUc8wXNlaxjyXeUyopxHELrX29HDfSnBibK8WYC6s4CuXfQ/LIUFdiIkhiYqLv+cqVK8nLy2P9+vUkJCQwYMCAakccjYuL8z13uVyUldV84f/YsWP5+OOP6dq1K3PnzmXlypWO1l8bwfxztiewW1X3qupp4D3gDv8NVPWE38tEsKsWndCGxvwL3ZjDHexmIgU8xQLuZhTXs5ujTCKP3swmlWkMZT5/ZTVr+ZHTNhOsMVWdXOt5tPEvzBVITk6muLi42nXHjx8nNTWVhIQEduzYwYYNG67484qLi2nevDlnzpxhwYIFvuWDBw8mNzcXgIqKCo4fP86gQYP44IMPOHLkCEBYnCLJAg74vS4AelXeSEQeA54CYoFBQaynwcqikW80UYCfKWYNP/p6OZ5jOQBuoulDK/p7ezh6kUU8tR933piIVLYaJBbie4a6EhPGmjRpQt++fbn++utxu900bfrbNBHDhg3jjTfeoHPnznTs2JHevXtf8ee98MIL9OrVi4yMDHr16uULN6+++io5OTnMmTMHl8tFbm4uffr04bnnnqN///64XC66d+/O3Llzr7gGUQ1Op4GIjACGqepD3tcPAL1U9fEath8NDFXVP1azLgfIAWjduvVN+/fvD0rNDdVhSgMCx7f8jAJxuMgmixvI5Foy6EwG15JBM5LsbpUIISLfqGqPUNdRl3r06KEbN26s/Rv29QaJgTZrgleUCbrt27fTubMNknYlqvseXqgNCWYPxkGgld/rlt5lNXkPyK1uharOBGaCp3FwqkDjkU6ib3I2gF8p40tv4FhPAe+yhWP8dj6wMfFc6w0bnuCRzrVk0IoUCx4mspwrhVPfQJOnQ12JMWEnmAEjH7haRNrhCRb3AaP9NxCRq1V1l/flrcAuTMil4ub3dOT3dARAUQopYRtFAV9L2MFsNvnel0SsL2x09gsgbWlsd6+Y8FS2AThr41+Yeuuxxx5j7dq1AcsmTpzIgw8+GKKKfhO0gKGqZ0XkceDveG5TfVNVt4rIX4CNqvoJ8LiI/A44A/wKVDk9YkJPEJqRTDOSGcRVAesOU8p2DgcEj/9lL2/xrW+beKLp5Ase6b7g0Z40u4vF1G8nVwNRnltUjamHXn/99VCXUKOg3tStqkuBpZWWPe/3fGIwP98EXzqJ9CORfrQJWH6cU2yv1OOxlh95h+9928QQxTU0Cejt6EgT2pNGMnGVP8pEqFqMl/M3YKD3ZQKQqaqNvesqwPdD9aOq3u5ocWVrIK4buBo5ultjGgIbNcYERQrx9KYVvQMuw4ESytnB4YBej838zEds55zfXcpNSaQDaXQgjfbex/NfqVQ/NK4JP7UZL0dV/81v+wlAd79dlKlqcIbX1NNQth4aPxyU3RsT6SxgmDqVRBw9yKIHWQHLyzjDTo6wiyPs5qjvK6/S6RbwjFDantSA0HE+iGSSaBeahhffeDkAInJ+vJyaRvT9A/DnOqns1Degp2z8C2MukwUMUy+4iaErzehKsyrryjjDXn71hY493scNFLCQrQE9H0nEVhM8PGGkBck2N0v9U6vxcgBEpA3QDrwDt3jEi8hG4CwwTVU/dqyyk6s9j+5bHNulabiOHTvGO++8w6OPPnrJ733llVfIyckhISEhCJUFjwUMU++5ieE6MrmOzCrrTnOWfRxjj18A2c1RvqOQJezgDOd828YT7Qsb7b3BowXJNCOJpiTRlEQSiK3LQzOX5j5gkapW+C1ro6oHReQqYLmIfK+qeyq/sdJYOrX7tJNrILYTRFf9uTPmUp2frv1yA8aYMWMsYBhTl2KJ5hrSuYb0KusqOMcBjvv1fPwWQpaxh7JqhkZPJpZmJPlCx2/PEwOWNyWRWPv1ccKljJdzH57JEX1U9aD3ca+IrMRzfUaVgHHJY+loBZR9CY1GXfwIjKkF/+nahwwZQmZmJu+//z7l5eXcddddTJ06ldLSUkaOHElBQQEVFRVMmTKFwsJCfvrpJwYOHEh6ejorVqyodv+PPPII+fn5lJWVMWLECKZOnQpAfn4+EydOpLS0lLi4OL744gsSEhKYNGkSn3/+OVFRUYwbN44JEyY4fszWQpqI5SKKtqTSllR+R/uAdec4RyGl/EyJ76sw4HkpW/iFPPYGDDLmLw23L3hcKIxkkEC03Y5bk4uOlwMgIp2AVGC937JU4KSqlotIOtAX+C9HqirfAueOg7ufI7sz9cuaxUUUHSx3dJ8ZWXH0uyujxvX+07UvW7aMRYsW8fXXX6Oq3H777axevZqioiJatGjBp59+CnjmKElJSeHll19mxYoVvtlOq/PSSy+RlpZGRUUFgwcP5rvvvqNTp06MGjWKhQsXkp2dzYkTJ3C73cycOZN9+/axefNmoqOjHZt7pDILGKZBiiKK5iTTnOSLbnuKM/ziDSP+ocQ/kORzkEJKKeF0lfcLkEEimSSSQQLpJJBBovex6vMmuBtM70gtx8sBT/B4TwPnNugMzBCRc3gmbpxWabbmy3f++gu7wNMEwbJly1i2bBndu3tuiCopKWHXrl3069ePP/3pT0yaNInbbruNfv1qH3Dff/99Zs6cydmzZzl06BDbtm1DRGjevDnZ2dkANGrkud06Ly+P8ePHEx3taWecmp69sobRihlzBeKJoTWNaU3ji25bQjmFlFbpDTlEMUWcpIhSvqOQIk5ylJqnWk4hLiB4nA8lNT1PIjZs75652Hg53tf/Wc371gE3BKWosjUQ3Rpianm9hgkrF+ppqAuqyrPPPsvDD1e9BXrTpk0sXbqUyZMnM3jwYJ5//vlq9hDohx9+YPr06eTn55OamsrYsWOrne69rlnAMMZBScSRRBztufhfBGep4ChlvuBxmJNVnh/mJD9ynG84RBGlARet+ovDVaVn5BGyqwyAZmpB1dODkTgk1JWYCOI/XfvQoUOZMmUK999/P0lJSRw8eJCYmBjOnj1LWloaY8aMoXHjxsyePTvgvTWdIjlx4gSJiYmkpKRQWFjIZ599xoABA+jYsSOHDh0iPz+f7OxsiouLcbvdDBkyhBkzZjBw4EDfKZJg9GJYwDAmRKJxkUkSmSTVantFKabcFzxqCiVFlPIDvzKS64J8BBHq3FFwpdnpEeMo/+nahw8fzujRo+nTpw8ASUlJvP322+zevZunn36aqKgoYmJiyM31zP+Zk5PDsGHDaNGiRbUXeXbt2pXu3bvTqVMnWrVqRd++fQGIjY1l4cKFTJgwgbKyMtxuN3l5eTz00EPs3LmTLl26EBMTw7hx43j88WonOr8iQZuuPVgueaplY0yNbLr2C1AFCc/TTqYqm679yl3qdO026pAxxlTHwoUxV8ROkRhjjDFholevXpSXB95iO3/+fG64ITjXO18JCxjGGGNMmPjqq69CXUKt2SkSY4wxxjjOAoYxxpgGIdxuaqhPLud7ZwHDGGNMxIuPj+fIkSMWMi6DqnLkyBHi4+Mv6X12DYYxxpiI17JlSwoKCigqKgp1KWEpPj6eli1bXtJ7LGAYY4yJeDExMbRr1y7UZTQodorEGGOMMY6zgGGMMcYYx1nAMMYYY4zjwm4uEhEpAvbXYtN04HCQywkVO7bwVB+PrY2qhnbu6jpmbQhgxxau6uOx1diGhF3AqC0R2RipkzjZsYWnSD62SBTJ/152bOEp3I7NTpEYY4wxxnEWMIwxxhjjuEgOGDNDXUAQ2bGFp0g+tkgUyf9edmzhKayOLWKvwTDGGGNM6ERyD4YxxhhjQiTiAoaIDBORf4jIbhF5JtT1OEVEWonIChHZJiJbRWRiqGtymoi4ROT/ROR/Ql2L00SksYgsEpEdIrJdRPqEuiZTPWtDwpe1IfVLRJ0iEREXsBMYAhQA+cAfVHVbSAtzgIg0B5qr6iYRSQa+Ae6MhGM7T0SeAnoAjVT1tlDX4yQReQtYo6qzRSQWSFDVY6GuywSyNiS8WRtSv0RaD0ZPYLeq7lXV08B7wB0hrskRqnpIVTd5nxcD24Gs0FblHBFpCdwKzA51LU4TkRTgn4A5AKp6ur43DA2YtSFhytqQ+ifSAkYWcMDvdQER9At0noi0BboDX4W2Eke9Avw7cC7UhQRBO6AI+G9v9+1sEUkMdVGmWtaGhC9rQ+qZSAsYEU9EkoAPgSdV9USo63GCiNwG/KKq34S6liCJBm4EclW1O1AKRMy5fRNerA0JS2HZhkRawDgItPJ73dK7LCKISAyehmGBqn4U6noc1Be4XUT24emSHiQib4e2JEcVAAWqev6vxUV4GgtT/1gbEp6sDamHIi1g5ANXi0g770Uw9wGfhLgmR4iI4Dn/tl1VXw51PU5S1WdVtaWqtsXzb7ZcVceEuCzHqOrPwAER6ehdNBiImAvrIoy1IWHI2pD6KTrUBThJVc+KyOPA3wEX8Kaqbg1xWU7pCzwAfC8im73L/kNVl4awJlN7E4AF3v+09gIPhrgeUw1rQ0w9FnZtSETdpmqMMcaY+iHSTpEYY4wxph6wgGGMMcYYx1nAMMYYY4zjLGAYY4wxxnEWMIwxxhjjOAsYJmREZEAkznpojKkb1obUbxYwjDHGGOM4CxjmokRkjIh8LSKbRWSGiLhEpERE/iYiW0XkCxHJ8G7bTUQ2iMh3IrJYRFK9yzuISJ6IfCsim0SkvXf3SSKySER2iMgC72iDxpgIYm1Iw2QBw1yQiHQGRgF9VbUbUAHcDyQCG1X1OmAV8GfvW+YBk1S1C/C93/IFwOuq2hW4GTjkXd4deBK4FrgKz2iDxpgIYW1IwxVRQ4WboBgM3ATke/8wcAO/4JkSeaF3m7eBj0QkBWisqqu8y98CPhCRZCBLVRcDqOopAO/+vlbVAu/rzUBb4MvgH5Yxpo5YG9JAWcAwFyPAW6r6bMBCkSmVtrvcMefL/Z5XYD+TxkQaa0MaKDtFYi7mC2CEiGQCiEiaiLTB87MzwrvNaOBLVT0O/Coi/bzLHwBWqWoxUCAid3r3ESciCXV6FMaYULE2pIGypGcuSFW3ichkYJmIRAFngMeAUqCnd90veM6xAvwReMP7y+8/498DwAwR+Yt3H/fW4WEYY0LE2pCGy2ZTNZdFREpUNSnUdRhjwpO1IZHPTpEYY4wxxnHWg2GMMcYYx1kPhjHGGGMcZwHDGGOMMY6zgGGMMcYYx1nAMMYYY4zjLGAYY4wxxnEWMIwxxhjjuP8H3aj89wch9SMAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "20.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "30.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "40.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "50.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "60.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "70.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "80.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "90.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "100.00% labels reassigned\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yG7AcIX7joj8"
      },
      "source": [
        "Randomizing pixels means that whatever signal there is in our data, it will be noisy, so it will be harder for our model to learn good data representation. Still, it does not have a huge impact, because we draw pixels from distribution we have in our data, and each image consists mostly of 0s. So truly threatening change would be 1->0 or 0->1, where 0 is white and 1 black. Since we don't perform adversarial attack, it doesn't happen so often, so our accuracy is obviously worse, but still rather high."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PmBD1PLsadrH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "bd961018-8666-4ec6-d9c8-03baafec8fad"
      },
      "source": [
        "for x in xs[1:]:\n",
        "    print(f'{x * 100:.2f}% pixels randomized')\n",
        "    plot_history(history3[x])\n",
        "    print()"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "10.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "20.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "30.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "40.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "50.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "60.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "70.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "80.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "90.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "100.00% pixels randomized\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hO7aQn6KoP1x"
      },
      "source": [
        "Let's visualize, how probable is to set pixel value to 0, so to \"erase a pixel\"."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "43Hqjuafm8er",
        "outputId": "2fa55eec-bc84-47b1-805f-8c07a803bb9f"
      },
      "source": [
        "fig, ax = plt.subplots()\n",
        "ax.plot([x/784 for x in pixels_dist.values()], 'o')\n",
        "ax.set_title('Probability of setting pixel value to x')\n",
        "ax.set_xlabel('x')\n",
        "ax.set_ylabel('prob')\n",
        "plt.show()"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gxl_u3Bkobbi"
      },
      "source": [
        "81% of all pixels in all images are 0s, which also mean that we will set any pixel value to 0 with 81% probability. If we added noise with more aggresive way (for example changing some 0s to 255s or another way around), network's performance would decrease more significantly.\n",
        "\n",
        "Let's visualize differences between experiments."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 308
        },
        "id": "RMirQpnBoe1E",
        "outputId": "1293c5ee-94bb-4da2-8131-98b9afcac45a"
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "test_labels = pd.DataFrame.from_dict(\n",
        "    {f'{k:.2f}': np.mean(v['test'], axis=0) for k, v in history2.items()},\n",
        "    orient='index', columns=['loss', 'acc']\n",
        ")\n",
        "test_labels['randomized'] = 'labels'\n",
        "test_pixels = pd.DataFrame.from_dict(\n",
        "    {f'{k:.2f}': np.mean(v['test'], axis=0) for k, v in history3.items()},\n",
        "    orient='index', columns=['loss', 'acc'],\n",
        ")\n",
        "test_pixels['randomized'] = 'pixels'\n",
        "df = pd.concat([test_labels, test_pixels])\n",
        "\n",
        "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(15, 4))\n",
        "for g, c in zip(['labels', 'pixels'], ['deepskyblue', 'orange']):\n",
        "    subset = df[df.randomized == g]\n",
        "    ax1.scatter(df.index.unique(), subset.acc, c=c, label=g)\n",
        "    ax2.scatter(df.index.unique(), subset.loss, c=c, label=g)\n",
        "\n",
        "ax1.legend(loc='lower left')\n",
        "ax1.set_xlabel('% of randomization')\n",
        "ax1.set_ylabel('accuracy')\n",
        "\n",
        "ax2.legend(loc='upper left')\n",
        "ax2.set_xlabel('% of randomization')\n",
        "ax2.set_ylabel('loss')\n",
        "\n",
        "plt.suptitle(\"Model's performance on test set with different noise colored by type of noise\")\n",
        "plt.show()"
      ],
      "execution_count": 148,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1080x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FTD90wpIwFq9"
      },
      "source": [
        "Label randomization means higher loss, since we use cross entropy loss. On the other hand, accuracy on test set decreases slower, so we have trade-off scenario.\n",
        "\n",
        "To sum up, neural networks are good at generalization, and don't bother with some unbiased noise, at least when training on such easy dataset as MNIST."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jIg5cPqApktb"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}