plate_effects.ipynb

plate_effects.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "plate_effects.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/Swarchal/d7e058104fb191c8d89062bca116c926/plate_effects.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "jpJPJ5OruruF",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CNNs to detect plate effects\n",
        "\n",
        "Scott Warchal  \n",
        "29-01-2019\n",
        "\n",
        "**Aim:** create and train CNN models to flag potential plate effects by using plate-maps as images\n"
      ]
    },
    {
      "metadata": {
        "id": "Tv_1O9uEBNkg",
        "colab_type": "code",
        "outputId": "6cbaa605-cd87-4911-d6a4-b93c534eaab0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 119
        }
      },
      "cell_type": "code",
      "source": [
        "import os\n",
        "import sys\n",
        "import platform\n",
        "from random import sample\n",
        "from string import ascii_uppercase as LETTERS\n",
        "\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "import torch.nn.functional as F\n",
        "from torchvision import datasets, transforms\n",
        "\n",
        "import numpy as np\n",
        "import matplotlib\n",
        "import matplotlib.pyplot as plt\n",
        "from tqdm import tqdm\n",
        "\n",
        "%matplotlib inline\n",
        "plt.style.use(\"seaborn-ticks\")\n",
        "\n",
        "print(f\"python = {platform.python_version()}\")\n",
        "\n",
        "for library in [torch, np, matplotlib]:\n",
        "  print(f\"{library.__name__} = {library.__version__}\")\n",
        "\n",
        "if torch.cuda.is_available():\n",
        "    device = \"cuda\"\n",
        "else:\n",
        "    device = \"cpu\"\n",
        "    \n",
        "print(f\"\\ndevice = {device}\")\n",
        "\n",
        "torch.backends.cudnn.benchmark = True"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "python = 3.6.7\n",
            "torch = 1.0.0\n",
            "numpy = 1.14.6\n",
            "matplotlib = 3.0.2\n",
            "\n",
            "device = cuda\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "IwqaMaU0BTQT",
        "colab_type": "code",
        "outputId": "86f049ce-9c40-4313-a83d-6a28b1034c44",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 337
        }
      },
      "cell_type": "code",
      "source": [
        "! nvidia-smi"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Wed Jan 30 14:24:26 2019       \n",
            "+-----------------------------------------------------------------------------+\n",
            "| NVIDIA-SMI 396.44                 Driver Version: 396.44                    |\n",
            "|-------------------------------+----------------------+----------------------+\n",
            "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\n",
            "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\n",
            "|===============================+======================+======================|\n",
            "|   0  Tesla K80           Off  | 00000000:00:04.0 Off |                    0 |\n",
            "| N/A   34C    P8    28W / 149W |     11MiB / 11441MiB |      0%      Default |\n",
            "+-------------------------------+----------------------+----------------------+\n",
            "                                                                               \n",
            "+-----------------------------------------------------------------------------+\n",
            "| Processes:                                                       GPU Memory |\n",
            "|  GPU       PID   Type   Process name                             Usage      |\n",
            "|=============================================================================|\n",
            "|  No running processes found                                                 |\n",
            "+-----------------------------------------------------------------------------+\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "Dc3sRbS6oJFX",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Creating a simple CNN model"
      ]
    },
    {
      "metadata": {
        "id": "np_Flsv-NSgE",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# create simple CNN model similar to AlexNet architecture\n",
        "\n",
        "class SimpleNet(nn.Module):\n",
        "    \"\"\"\n",
        "    5 conv blocks -> 2 dense layers\n",
        "    Hugely overkill for a 1 channel 16*24 image.\n",
        "    \"\"\"\n",
        "    \n",
        "    def __init__(self, num_classes=6, n_channels=1):\n",
        "        super(SimpleNet, self).__init__()\n",
        "        self.features = nn.Sequential( \n",
        "            ## conv block 1\n",
        "            nn.Conv2d(n_channels, 64, kernel_size=11, stride=4, padding=2),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=1),\n",
        "            ## conv block 2 (no pooling)\n",
        "            nn.Conv2d(64, 192, kernel_size=5, padding=2),\n",
        "            nn.ReLU(inplace=True),\n",
        "            ## conv block 3 (no pooling)\n",
        "            nn.Conv2d(192, 384, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            ## conv block 4 (no pooling)\n",
        "            nn.Conv2d(384, 256, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            ## conv block 5\n",
        "            nn.Conv2d(256, 256, kernel_size=3, padding=1),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.MaxPool2d(kernel_size=3, stride=2)\n",
        "        )\n",
        "        self.classifier = nn.Sequential(\n",
        "            nn.Dropout(),\n",
        "            nn.Linear(256, 1024),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Dropout(),\n",
        "            nn.Linear(1024, 1024),\n",
        "            nn.ReLU(inplace=True),\n",
        "            nn.Linear(1024, num_classes)\n",
        "        )\n",
        "        \n",
        "    def forward(self, x):\n",
        "        x = self.features(x)\n",
        "        x = x.view(x.size(0), 256)\n",
        "        x = self.classifier(x)\n",
        "        return F.log_softmax(x, dim=1)\n",
        "      "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "mT8SrfJODB1B",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "Quick test of `SimpleNet`..."
      ]
    },
    {
      "metadata": {
        "id": "IvoN_s_r2PLb",
        "colab_type": "code",
        "outputId": "7480a5aa-5b6d-4f67-a9d1-1a05872a3c0a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "cell_type": "code",
      "source": [
        "# create a random tensor in the shape of a 384-well plate\n",
        "x = np.random.randn(1, 16, 24)\n",
        "# have to add a batch dimension to the tensor\n",
        "x = torch.Tensor(x).unsqueeze(0)\n",
        "model = SimpleNet()\n",
        "\n",
        "x = x.to(device)\n",
        "model = model.to(device)\n",
        "\n",
        "model(x)"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "tensor([[-1.7703, -1.8112, -1.8187, -1.7747, -1.7647, -1.8124]],\n",
              "       device='cuda:0', grad_fn=<LogSoftmaxBackward>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "metadata": {
        "id": "4q23S6m4Dnw0",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Generating a synthetic datasets\n",
        "\n",
        "Functions to generate numpy arrays of 384-well plates, either random values, or with typical plate effects."
      ]
    },
    {
      "metadata": {
        "id": "R_F6KmxDDrDk",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def normalise(x):\n",
        "    return (x - x.mean()) / x.std()\n",
        "\n",
        "\n",
        "def add_noise(x, sigma=None):\n",
        "    \"\"\"\n",
        "    add random (Gaussian) noise to a plate\n",
        "    \n",
        "    Parameters:\n",
        "    -----------\n",
        "    x: numpy.array\n",
        "        plate\n",
        "    sigma: float\n",
        "        std dev for Gausian noise.\n",
        "        if None then sigma will be randomly sampled from [0.1, 5].\n",
        "    \n",
        "    Returns:\n",
        "    --------\n",
        "    numpy.array\n",
        "    \"\"\"\n",
        "    if sigma is None:\n",
        "        sigma = np.random.uniform(0.01, 2)\n",
        "    noise = np.random.normal(scale=sigma, size=(16, 24))\n",
        "    assert x.shape == noise.shape\n",
        "    return x + noise\n",
        "    \n",
        "\n",
        "def normal_random_plate(sigma=None, norm=False):\n",
        "    \"\"\"\n",
        "    Generate a random plate from a Gaussian distribution\n",
        "    \"\"\"\n",
        "    if sigma is None:\n",
        "        sigma = np.random.uniform(1, 3)\n",
        "    output =  np.random.normal(scale=sigma, size=(16, 24))\n",
        "    if norm:\n",
        "        output = normalise(output)\n",
        "    return output\n",
        "\n",
        "\n",
        "def uniform_random_plate(norm=True, **kwargs):\n",
        "    \"\"\"\n",
        "    Generate a random plate from a uniform distribution\n",
        "    \"\"\"\n",
        "    output = np.random.uniform(**kwargs, size=(16, 24))\n",
        "    if norm:\n",
        "        output = normalise(output)\n",
        "    return output\n",
        "\n",
        "\n",
        "def edge_1(diam=25, center=None, invert=None, norm=True):\n",
        "    \"\"\"\n",
        "    Generate a plate with an edge effect by creating a bi-variate\n",
        "    Gaussian distribution and adding random noise.\n",
        "    \"\"\"\n",
        "    if invert is None:\n",
        "        # randomly invert or not\n",
        "        invert = sample([True, False], 1)[0]\n",
        "    size = 24\n",
        "    # generate a 24*24 bi-variate gaussian distribution\n",
        "    x = np.arange(0, size, 1, float)\n",
        "    y = x[:,np.newaxis]\n",
        "    if center is None:\n",
        "        x0 = y0 = size // 2\n",
        "    else:\n",
        "        x0 = center[0]\n",
        "        y0 = center[1]\n",
        "    output = np.exp(-4*np.log(2) * ((x-x0)**2 + (y-y0)**2) / diam**2)\n",
        "    # generate and add random noise to gaussian distribution\n",
        "    noise = np.random.normal(loc=1, scale=0.2, size=(24, 24))\n",
        "    noise = np.clip(noise, a_min=-2, a_max=2)\n",
        "    output *= noise\n",
        "    # trim edges as original data is generated as a square\n",
        "    output = output[4:20, :]\n",
        "    if invert:\n",
        "        output = 1 / output\n",
        "    assert output.shape == (16, 24)\n",
        "    if norm:\n",
        "        output = normalise(output)\n",
        "    return output\n",
        "\n",
        "\n",
        "def edge_2(sigma=None, invert=None, norm=True):\n",
        "    \"\"\"\n",
        "    docstring\n",
        "    \"\"\"\n",
        "    if sigma is None:\n",
        "        sigma = np.random.uniform(1, 5)\n",
        "    if invert is None:\n",
        "        invert = sample([True, False], 1)[0]\n",
        "    edge_plate = np.zeros(shape=(16, 24))\n",
        "    # outer\n",
        "    edge_plate[0, :]     = 5 # top edge\n",
        "    edge_plate[-1, :]    = 5 # bottom edge\n",
        "    edge_plate[:, 0]     = 5 # left edge\n",
        "    edge_plate[1:-1, -1] = 5 # right edge\n",
        "    # inner\n",
        "    edge_plate[1, 1:-1]  = 2.5 # top inner\n",
        "    edge_plate[-2, 1:-1] = 2.5 # bottom inner\n",
        "    edge_plate[1:-1, 1]  = 2.5 # left inner\n",
        "    edge_plate[1:-1, -2] = 2.5 # right inner\n",
        "    output =  normal_random_plate() + edge_plate\n",
        "    if invert:\n",
        "        output = 1 / output\n",
        "    if norm:\n",
        "        output = normalise(output)\n",
        "    return output\n",
        "\n",
        "\n",
        "def edge_effect_plate():\n",
        "    f = sample([edge_1, edge_2], 1)[0]\n",
        "    return f()\n",
        "    \n",
        "\n",
        "\n",
        "def row_effect_plate(norm=True):\n",
        "    \"\"\"\n",
        "    Generate a plate with row effects by generating alternate stipes of 0,1\n",
        "    on rows, and adding random noise.\n",
        "    \"\"\"\n",
        "    random_plate = normal_random_plate()\n",
        "    # column effects\n",
        "    effect = np.ones((16, 24)) * 1.5\n",
        "    effect[::2, :] = 0\n",
        "    random_plate += effect\n",
        "    if norm:\n",
        "        random_plate = normalise(random_plate)\n",
        "    return random_plate\n",
        "\n",
        "\n",
        "def column_effect_plate(norm=True):\n",
        "    \"\"\"\n",
        "    Generate a plate with column effects by generating alternte stripes of 0,1\n",
        "    on column, and adding random noise.\n",
        "    \"\"\"\n",
        "    random_plate = normal_random_plate()\n",
        "    # column effects\n",
        "    effect = np.ones((16, 24)) * 1.5\n",
        "    effect[:, ::2] = 0\n",
        "    random_plate += effect\n",
        "    if norm:\n",
        "        random_plate = normalise(random_plate)\n",
        "    return random_plate\n",
        "\n",
        "\n",
        "def single_checker_effect_plate(norm=True):\n",
        "    random_plate = normal_random_plate()\n",
        "    # column effects\n",
        "    effect = np.ones((16, 24)) * 5\n",
        "    effect[::2, ::2] = 0\n",
        "    random_plate += effect\n",
        "    if norm:\n",
        "        random_plate = normalise(random_plate)\n",
        "    return random_plate\n",
        "\n",
        "\n",
        "def quad_checker_effect_plate(sigma=1, norm=True):\n",
        "    random_plate = normal_random_plate()\n",
        "    checker_mask = (np.random.randn(8, 12)\n",
        "                        .repeat(2, axis=0)\n",
        "                        .repeat(2, axis=1)) * (sigma * 2)\n",
        "    random_plate += checker_mask\n",
        "    if norm:\n",
        "        random_plate = normalise(random_plate)\n",
        "    return random_plate\n",
        "    \n",
        "    \n",
        "def add_effects(*plates):\n",
        "    raise NotImplementedError()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "kOvQURjxNGTg",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Example images of synthetic plates"
      ]
    },
    {
      "metadata": {
        "id": "aGQQ8mxbFpoE",
        "colab_type": "code",
        "outputId": "c08c1b87-9611-4bcf-f749-a8b2abe78938",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 2169
        }
      },
      "cell_type": "code",
      "source": [
        "def plot_plate(arr, title=None):\n",
        "    plt.imshow(arr, cmap=plt.cm.viridis)\n",
        "    plt.grid(False)\n",
        "    plt.xticks(ticks=range(24), labels=range(1, 25))\n",
        "    plt.yticks(ticks=range(16), labels=LETTERS[:16])\n",
        "    if title is not None:\n",
        "        plt.title(title)\n",
        "    plt.show()\n",
        "\n",
        "    \n",
        "funcs = [normal_random_plate, uniform_random_plate, edge_1, edge_2,\n",
        "         row_effect_plate, column_effect_plate, quad_checker_effect_plate,\n",
        "         single_checker_effect_plate]\n",
        "\n",
        "names = [\"normal\", \"uniform\", \"edge_1\", \"edge_2\", \"row\", \"column\", \"96 => 384 checker\",\n",
        "         \"single checker\"]\n",
        "\n",
        "for name, f in zip(names, funcs):\n",
        "    plot_plate(f(), title=name)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ncCIinWIDJyLSKTZwIiKdYgMnItKp/wNADbjcRqaA9AAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3XlYlWX+P/D34bDJ5sLmAgoaruSS\nlnsamttEM9pULtDiWGMpOjaiuKVpuY1WimaazpVijbsz/tTKkEpN4xtmKs2oXwVEFBBEBNk55/7+\n4SU/l3M4z33LAR/n/bourytOz+d87vM89/nwcJ/nfB6DEEKAiIh0x6GuB0BERGpYwImIdIoFnIhI\np1jAiYh0igWciEinWMCJiHSKBZweSbNmzUJsbKxdnlsIgfXr16NDhw5ISkqySw4iLRzregBEejN3\n7lyYzWY0atSorodC/+V4Bk66Ex8fj/DwcAwYMABjx45FXl4erl+/jrFjxyIsLAxvvvkmCgsLq7Y/\nfPgw+vXrh6FDh2Lr1q144oknkJGRAQDYunUrhgwZgrCwMLzzzjsoLS21mX/48OF4//334eTkZLfX\nSKQFCzjpyqVLlzBt2jQsX74cBw8eRPfu3TFv3jx89tlnaNiwIRISEvDuu+/iyJEjAACTyYSYmBjM\nnz8fX331FdLS0lBSUgIASEpKwooVK7Bx40YkJCTAw8MDK1assDmGLl262PU1EmnFAk66cujQITz1\n1FNo3bo1AGDkyJFISEhAYmIihg4dCgAICAjAU089BQBIS0tDeXk5+vXrBwCIjIyE2WwGACQkJGDY\nsGHw9/cHAIwaNQoHDhyo7ZdEpIxr4KQrhYWFSEpKwpAhQ6oe8/DwQH5+Pjw9Pase8/LyAgDcuHGj\n6r8BwM/P767n+vbbb6vO1oUQqKiosPdLIKoxLOCkK35+fujVqxdWrlx51+MvvvjiXeveeXl5CAwM\nhIeHB4qLi6sez83Nveu5hg8fjunTp9t/4ER2wCUU0pU+ffogKSkJly5dAgCcOnUK77//Pjp37oz4\n+HgAQHp6Oo4fPw4ACAoKQmVlJRITEwEA//jHP2AwGAAAYWFhOHDgAPLy8gDc+nB03bp1tf2SiJTx\nDJx0xc/PDwsWLMCECRNQUVEBd3d3zJw5E82bN8eUKVMQFhaGVq1aYdCgQQAAZ2dnzJs3DzNmzICn\npydef/11ODg4wGAwoEOHDhg/fnzVuri3tzfee+89m2N47rnnUFlZiezsbERHR8PFxQVLly5Fx44d\n7f3yie5iYD9w+m9SXFyMLl26ICkp6a41cyI94hIKPfJeeOEF7N+/HwCwf/9+tGrVisWbHgk8A6dH\nXlJSEubPn4+ysjK4u7tj3rx51S53zJ8/H0ePHrX4/+bOnYuePXvaa6hEUljAiYh0qtY/xCwtLUVy\ncjJ8fX1hNBprOz0R0UPHZDIhJycHoaGhcHV11RxX6wU8OTkZY8aMqe20REQPvS+++ALdunXTvH2t\nF3BfX18AQJPiJ+EotP+mAQAY5T9zNfvWl44BgFLfetIxQvEj4atPqP0lYnI3S8e03FGilCuzl7tS\nHAzyIWbFWWlyk49xuqGWq6SZSSmu3hX5Y10vRykVKjzU4gpbyb82n1/UJn+5u/wE8bxSqZQLlWqr\nxUVN5JuWuWfKfaPXVFmE65n7q+qjVrVewG8vmzgKVzgJySIp5Ce/yUFtFhud5KuBwvBu8VQM9JAv\n4I6Ky1YGd8WrNhQKuEF1VnrIv0ENlQoDBCDqqxURww35F+dQqFZ4DG6199ocXNQKuEM9+TEanRT3\nvUFtPzq4yhdwRye1lgyyy8q8jJCISKekCvjevXvRoUOHqq8eExFR3ZEu4IGBgfjmm2/sNR4iItJI\ncwHPz8/HqVOnEBMTg3379tlzTEREpIHmAv7111+jf//+6Nu3L9LS0pCdnW0zJjY2Fm3atLnr34AB\nAx5owEREdIvmAr53714899xzMBqNGDJkSFVviepERUXh7Nmzd/07ePDgAw2YiIhu0XRNU1ZWFk6e\nPInFixfDYDCgtLS0qjUnERHVDU0FfO/evRgzZgxiYmIA3Lr11KBBg5Ceno7mzZvbdYBERGSZpiWU\nffv2YcSIEVU/GwwG/OEPf+CHmUREdUjTGfju3bvve2zChAk1PhgiItKO38QkItKpOrsnZmWAD2CU\n61NS6Snfk8A1o0A6BgDK6sv3UHHNU2tw1PSIWtzNZvJ9TbJ6qPU0cSq0vY0lJvmeYCj3UutZUeEr\n33/CKUStuZfDJbX92P65s9IxP/+npVKuJgfV+t6Ue8qXBbNj7d1WoNhXrWwJtdYwSv18hFEyyKw2\nOJ6BExHpFAs4EZFOsYATEemU5sWkjIwMhIeHIzQ0FABQXl6O6OhoqbtHEBFRzZH6NCA4OBhxcXEA\ngJ9//hlr1qzBhg0b7DIwIiKqnvISSm5uLvz8/GpyLEREJEHqDDw1NRWRkZEoKytDdna2zbPv2NhY\nrFq16oEGSERElikvoVy4cAF/+ctfsHv3bjg6Wn6aqKgoREVF3fVYRkYGW8oSEdUA5SWUVq1awcXF\nBZmZmTU5HiIi0ki5gOfn5yMnJwf+/v41OR4iItJIaQ0cAMrKyjBnzhw4OzvbZWBERFQ9zQU8ICAA\nJ06csOdYiIhIQp01sxJGg3TDF5cf/yOdJz/8cekYAKh0k28uU+ihtjvzOqk1s3Iol28g5BKg1pXK\n/JuXWlybm9IxIs1dKZfXafm/Bo2l8g3SAKBC8T4mJy83k45xS1EbY25HtQZTKo3LXArMSrnMjvKr\nuGa1Hl1wUHubwbFYfj+WNpQbpLlM7UXxq/RERDrFAk5EpFMs4EREOqV50TYtLQ0LFy5EXl4ezGYz\nunTpgunTp/MqFCKiOqLpDNxkMiEqKgrjxo3Djh07sHPnTgDA6tWr7To4IiKyTtMZ+I8//oiWLVvi\nqaeeAnDrrvTR0dFwcOAKDBFRXdFUwFNSUtCuXbu7HnN1dbUZx2ZWRET2o6mAGwwGmEzyF1GymRUR\nkf1oWgNp2bIlTp8+fddj5eXlOHfunF0GRUREtmkq4L1798bly5eRkJAAADCbzfjb3/6G/fv323Vw\nRERknaYC7uDggA0bNmDbtm0YMWIERo8eDU9PT0yaNMne4yMiIis0Xwfu5+eHTz/91J5jISIiCbwO\nkIhIp+qsG6GxsAyODnLp817oKJ3H7WqldAwAFDaX7wBX3FitI1uLfWpd4zJ7yHdMLCl0Ucrlc1Ft\njIZuJdIxuU3UpqX7CflvBecOL1bKVXmtnlKcuG778tt7jRz5g1Kub/7WVynOYJI/1rmPq3XT80yT\nz2VWrFpeaWVKcQVB8u+Zhp8fk9q+0rEchcHSaXgGTkSkVyzgREQ6xQJORKRTmlaTMjIyEB4ejtDQ\nUAghYDQaMX78ePTs2dPe4yMiIis0fxwQHByMuLg4AEB6ejrGjx+PDz/8EG3btrXb4IiIyDqlJZTm\nzZtj/Pjx+PLLL2t6PEREpJHyZYShoaHYsmVLtduwGyERkf0oF/CioiIYjdVf+8luhERE9qN8FUpy\ncvJ9PcKJiKj2KBXw9PR0fP7553jttddqeDhERKSV5iWU1NRUREZGory8HCaTCe+++y6aNm1qz7ER\nEVE1NBXwgIAAnDhxwt5jISIiCXXWzEq4OMJslGsYZTDLN74xlsnfCg4AXHPld43JWe0jhZyOanEV\ngfLNeeofV2tmVRCs1szKWCKfz2BUy5XVSz6mfeMcpVxhoWeV4v5+Tv7Lb0+5X1DKta1VP6U40aFQ\nOsZrn4dSrmtd5I91w9PyTdwA4Eaw2twvaiqfz+uZJ6S2N1UWAllnpPPwq/RERDrFAk5EpFMs4ERE\nOqV5offOhlZ3io2NRYMGDWp8YEREVD2pT+rubGhFRER1i0soREQ6xQJORKRTUksot7+NeVtwcDDm\nz59vdXt2IyQish+7roGzGyERkf1wCYWISKceaAkFAKKjo9GxY8caHRQREdmmuYCzoRUR0cOFSyhE\nRDpVZ90IHYrLYTTIddNzuVFPOs/NZs7SMQBQ6Srfgcyg1vgQ7plq3fdKmsofvoI2aoMUjmpjdD7r\nKR3jaFZKBYe2N6VjzhxvoZTr367NleKEQX4//vX0WKVc5d5qx7pz4yzpmBO9gpRyeZ2R60gKANf7\nlyjl8vxZvn4AgPe/5fdjqbfc6zKVOwHyu51n4EREesUCTkSkUyzgREQ69UDdCNu2bYtZs2bZZWBE\nRFQ9diMkItIpLqEQEemUXS8jZDMrIiL7eaCv0vfq1QtvvfWW1e3ZzIqIyH64Bk5EpFNcAyci0ikW\ncCIindJcwAMCArBr1y57joWIiCTwDJyISKfqrBthXueGcHCV61QnjPJ5yhrIdxUEAJd8+a5xTjfV\ncpnUGiZCOMmPMbybWk93D6Nc58jbdu7vLR3j1eWaUq7+Tf9XOmbHja5KuQxlCpMRwJMdL0jHbGt5\nUCnXY9+9rhR35qq/dExgUK5Srhv/biId0/j/uSjlym+lFAaPCzekYwra1JfaXq3XJ8/AiYh0iwWc\niEinWMCJiHRK0xr4xYsXsWjRIly7dmttsmnTppg7dy4aNWpk18EREZF1Ns/ATSYToqKiMG7cOGzf\nvh3bt29Hhw4d8MEHH9TG+IiIyAqbZ+A//vgjQkJC0K1bt6rHxo0bByFUPzclIqKaYLOAp6SkoE2b\nNnc95uCgbemc3QiJiOzHZgF3cHBAZWVl1c9vvfUWbt68iaysLOzZswf16lm/0zO7ERIR2Y/NU+mQ\nkBCcPn266uc1a9YgLi4OJpMJZrPZroMjIiLrbBbwHj16ICsrCwkJCVWP/fbbbygqKoLRqPZtNCIi\nenA2l1AMBgPWr1+P+fPnY/Xq1XBycoKbmxvWrFkDV1fX2hgjERFZoOk6cG9vb6xYscLeYyEiIgl1\n1szKWCZglGzh0vB4jnSe4sfUvmx0rb2TdExRC5NSrvpn1Jai6v8mf/j2VjyhlMslV22Mf4v4XDom\neturSrl+rVciHWO8Ln+cAcA9JF8p7vj/hEjHPF3kpZTrmcfOKcXF/9JBOqbgkNoYK+R6PgEArjZV\naxrnlqkUhqs9G0rHOBfK1TazWe018av0REQ6xQJORKRTLOBERDqlqYBnZGRgxIgRdz0WGxuLzZs3\n22VQRERkG8/AiYh0igWciEinNF+HlpqaisjIyKqfL1++jLFjx1Ybw2ZWRET2o7mABwcHIy4ururn\n2NhYmzFsZkVEZD9cQiEi0ikWcCIinWIBJyLSKU1r4AEBAdi1a9ddj927tk1ERLWLZ+BERDpVZ90I\nzc4GwEWuA1elr6d0ntKGal30ShrL323I/ZJarsIgtRtEu+TJdzDr1uW8Uq6fzwYrxX2d31E6xv9n\nta6Oqa7NpGOcgm4q5TILte5xwkX+WBfsbaKU6+Bj/kpx7oGF0jFFhWrdCN0y5fejU1e1TpC4It9V\nEAAani2TjrnWQe5eCeYSdiMkIvqvwgJORKRTLOBERDqlXMAtdSgkIqLawzNwIiKdYgEnItIpu15G\nyG6ERET2Y9cCzm6ERET2wyUUIiKdYgEnItIpFnAiIp1SLuCWOhQSEVHt4Rk4EZFO1V03QkcATnIx\nOZ3c5BMp/opyqJDvGueaq9ZVsLxrkVKc9xG5jmcAcLKzfMc+ADDcVJsq36WFSMeUDVDrzNaq/WXp\nmLytAUq54ucsV4ob9b8vSsdca+GulMvphLdSXPlZ+c6CzoVqx8zU94Z0THmJs1IuB2+192deOxfp\nGGO5XC5DuXQKADwDJyLSLRZwIiKdUirg48aNQ+/evfHdd9/V9HiIiEgjpQK+fv169O3bt6bHQkRE\nEriEQkSkU2xmRUSkU2xmRUSkU1JLKAUFBSgvv3XBotlshtGodhd2IiJ6cFIF/L333kN8fDyEEEhJ\nSUFwcLC9xkVERDZIFfCoqChs3LgRo0aNQr9+/RAYGGivcRERkQ1Sa+BBQUHYunWrvcZCREQSeBkh\nEZFO1VkzK5/DmXCCXCObnP7yjYeKG6s12XHJk48rUusThfJCteY82d3kP0Q2GNQa+vi3zlGKK/7G\nXzqmvGuJUq60JPn50fP135RydT/yllJcM5986Zjcq/LNpQDA0MCsFCfqmaRjXK6rzeGyc/KvrV6W\n2ntamZB/zxT7yY1RqPWz4xk4EZFesYATEekUCzgRkU5pLuAZGRkYMWJE1c/x8fEYM2ZM1Rd7iIio\ndil9iHn27FmsXLkSn3/+OZyd1T68ICKiByO9hJKXl4fp06fjo48+QqNGjewxJiIi0kDqDLyyshKT\nJk3C0KFD0apVK5vbsxshEZH9SJ2Bp6amYujQodi5cyeysrJsbh8VFYWzZ8/e9e/gwYPKgyUiov9P\nqoCHhIRgzJgxeOeddzB16lSYTPIX/BMRUc1QuoxwyJAhCAwMxOrVq2t6PEREpJHydeCzZ8/G/v37\nkZiYWJPjISIijTR/iBkQEIBdu3ZV/ezu7o6vv/7aLoMiIiLb+E1MIiKdqrNuhJfDm8DgLteJzOui\nfHe1Zj+otfm6+oSHfJBikzRDkdphcCyWT1hc6KKUq/SQp1Kcy1D5LoYFlxoq5TIEyncx/J9vOyjl\nMra7qRSXeV2++179X9W+LOf/+3SlOBWXzzdXinO6IT+HzYrfHayspxbX4Kp8N0LhKPe6zCWAynfa\neQZORKRTLOBERDolVcDvbWhFRER1h2fgREQ6xQJORKRTLOBERDpl18sI2Y2QiMh+7FrAo6KiEBUV\ndddjGRkZGDBggD3TEhH9V+ASChGRTkmfgaempiIyMrLq5+joaHTs2LFGB0VERLZJFfCAgACcOHHC\nXmMhIiIJXEIhItKpOmtm1SDFBKNzpVRMkZ9ROk9ZR3fpGAAoaSzfwMZYotbNyiDfowsAUOkuP0ZU\nqP3Ozg+VO1a3eZQ7Scc45ssfZwAwe8jv/3qdrivlMnyl1nCr02vJ0jFH/UKVcqUmBirFVTSukI5p\nlKcwFwGUeym8ZxSbxjU8q/ZGK/aVf884VMrtD4Pizc14Bk5EpFMs4EREOsUCTkSkUzYLeEZGBtq0\naYNff/31rsdfeOEFxMTE2G1gRERUPU1n4IGBgdi7d2/VzxcvXkRBQYHdBkVERLZpKuCdOnXC0aNH\nYTLd+qh037596N27t10HRkRE1dNUwJ2cnNCpUyckJiYCAA4ePIh+/frZjIuNjUWbNm3u+sc+KERE\nNUPzdeBDhgzB3r174ePjA39/f7i5udmMYTMrIiL70XwVSs+ePZGYmIh9+/Zh8ODB9hwTERFpoLmA\nOzs748knn8TOnTsRFhZmzzEREZEGUl+lHzJkCPLy8uDp6Wmv8RARkUY2C3hAQAAWL14MAOjfvz/6\n9+8PAOjevTu6d+9u18EREZF1/CYmEZFO1Vk3Qpfccjgay6RiivxtX/lyL/+DWdIxAHA5vIl0jMlF\nKRW8W19TinNZ30g65soI+U5zAGAuku8qCAAVpxpIx5h81FqzNdkjfwDMr6t9Ia3IXa0l3g9J7aVj\nhL9aJ8hGP6u9vUtal0rHGBW6TgLAzRD5+eh5Ti1Xibfa+apLvnynRffMcqntKyvLkSOdhWfgRES6\nxQJORKRTLOBERDqlqYBnZGSgXbt2OHPmTNVju3btwq5du+w2MCIiqp7mM/DHHnsMy5cvt+dYiIhI\nguYC3qFDB7i5ueHYsWP2HA8REWkkdZ3RlClTMH36dPTo0UPT9rGxsVi1apXSwIiIqHpSBTwoKAjt\n27fH/v37NW3PboRERPYjfRXKhAkTsG7dOlRWqn25gIiIaoZ0Affx8cHAgQOxZcsWe4yHiIg0UroO\nfOzYscjKUvuKOhER1QxNa+B3diQEAHd3dxw9etRugyIiItv4TUwiIp2qs26ExY1dYHRxlYppeE6u\neyEAXH2msXQMADgrdCAreEwpFYybvZXinAvkO7l5JNVTymVW7LQoFJr2OWQalXJl9jZLx7gWqe2P\nSX/6l1Lcp5/+XjrGLVut82FBsFpcwEfyZaHEX37fA0C9S/KdBeunqnWr9Dx1VSkur4e/dIxDpVz9\nkN2+Kk4pioiI6hwLOBGRTmn6Wyk9PR0LFy5ETk4OzGYznnjiCURHR8PVVW4JhIiIao7NM3Cz2Yyo\nqCi8+uqr2LlzJ3bv3o1mzZphzpw5tTE+IiKywmYBP3LkCIKCgtCzZ8+qx15//XWcOnUK166p3QqM\niIgenM0llJSUFLRvf/d9/AwGA0JCQpCWlgZvb+tXULCZFRGR/dgs4AaDASbT/ZftCCFgNFZ/uReb\nWRER2Y/NJZSWLVsiOTn5rseEEDh//jyCgoLsNS4iIrLBZgHv3bs3MjIy8MMPP1Q99vnnn6Nr165o\n0KCBXQdHRETW2VxCcXBwwIYNGzB37lysWLECQgiEhoZi9uzZtTE+IiKyQtN14L6+vvjkk0/sPRYi\nIpLAb2ISEelUnTWzqnetAo6O5XbPU+mq1tDH44p8wxyfX5VSodhfrXlTQZB8hymj4i4v9VZrtmPy\nkG9yZChTO2aO/sXyuY57KeVaeuU5pbj6A/KkY0w7Gyrl8vtF7WDfDJT/hvWNYLVzQZX56Jyvdjew\n9BeaKMX5nJZvGueUKzkXRal0DoBn4EREusUCTkSkUyzgREQ6pXkNPCMjA+Hh4QgNDYUQAuXl5Xjj\njTfw7LPP2nN8RERkhdSHmMHBwYiLiwMA5OfnY/jw4ejbty/byhIR1QHlJZQGDRrA19cXOTk5NTke\nIiLSSPkywoyMDOTn56NJE+uX5rAbIRGR/UgV8NTUVERGRkIIARcXFyxZsgSOjtafgt0IiYjsR3kN\nnIiI6hYvIyQi0ikWcCIindJcwAMCArBr1y57joWIiCTUejOr27dnqzQV1Uo+USzfiAYATOXyTZiE\nWg8mmEvU/hASCmFCbXcAhYpxJvn9iHLFHVmvRDpEFCmMDwAK1OLMefLz3lyq9jatrFCc+2Xy+USR\n2jET8j3jUFmp1qRLFJUpxZkq5JtnVQi541wpbjW/snT7yurUegG/fd147rWvayfhVbUw+/dJvMN/\nai+VWh839YlSZ+0uNVLrsQjI94G8RWVeqc7Fm4pxyFCISba9iSUK9Vv1LQ1kqoXlquZTkJOTgxYt\nWmje3iCEUJ3DSkpLS5GcnAxfX9/7boo8YMAAHDx4UPo5azPuUc2lGveo5lKNe1RzqcYxl7Y4k8mE\nnJwchIaGSn2zvdZPkFxdXdGtWzer/z8gIEDpeWsz7lHNpRr3qOZSjXtUc6nGMZe2OJkz79t4FQoR\nkU6xgBMR6RQLOBGRThnnzZs3r64Hcafu3bs/9HGPai7VuEc1l2rco5pLNY65aibOklq/CoWIiGoG\nl1CIiHSKBZyISKdYwImIdIoFnIhIp1jAiYh0igWciEinHpoCfu7cOQwcOBCbN2+Wilu6dClefvll\nvPDCCzhw4IDN7UtKSjB58mRERETgxRdfxHfffSeVr7S0FAMHDtTUGz0xMRE9evRAZGQkIiMjsWDB\nAs159uzZg+effx4jRozA999/rylm+/btVbkiIyPRpUsXmzFFRUWYOHEiIiMjMXLkSBw+fFhTLrPZ\njDlz5mDkyJGIjIzEhQsXqt3+3uObmZmJyMhIjB49GpMnT0Z5ueWee5bmxaZNm9ChQwcUFVlu2Wkp\n12uvvYaIiAi89tprVR0xbcWdOHECo0aNQmRkJP70pz8hLy9P0/gA4PDhw2jTpo3m/RETE4Pw8PCq\nY2fpmN8bU1FRgb/+9a/44x//iFdffRU3btzQlGvSpElVecLDwzFnzhxNcT///HPV/vjzn/9sMd+9\nMRcuXMCYMWMQERGB2bNno7LScj/Me9/HWuaHpfe+rblhLZet+XFvjJa5YW2MgO35oZl4CBQVFYmI\niAgxe/ZsERcXpznu2LFjYty4cUIIIfLy8kS/fv1sxuzbt0+sW7dOCCFERkaGGDRokNRYP/zwQzFi\nxAixc+dOm9v+9NNPIioqSur5hbj1WgYNGiQKCwtFdna2mD17tvRzJCYminnz5tncLi4uTixbtkwI\nIURWVpYYPHiwpuc/cOCAmDx5shBCiIsXL4o333zT6raWjm9MTIzYv3+/EEKI5cuXiy+++EJT3O7d\nu8WHH34o+vfvL27evKkpZtq0aWLfvn1CCCE2b94slixZoikuKipKpKenCyGEiI2NFWvWrLEZI4QQ\npaWlIiIiQvTu3Vvz/pg+fbpISEiwtgstxmzevFksWLBACCHEli1bRHx8vKa4O8XExIiTJ09qihs+\nfLi4cOGCEEKINWvWiLVr19qMGT9+vPj++++FEEKsWrVK7Nmz575clt7HtuaHpRhbc8NanK35YSnG\n1tywFieE7fkh46E4A3d2dsZnn30GPz8/qbgnn3wSK1asAAB4eXmhpKTEZkP0YcOG4Y033gBw68zM\n399fc74LFy7g/Pnz6N+/v9Q4ZR07dgw9e/aEh4cH/Pz8pM7cb1u9ejXefvttm9s1bNgQ+fn5AICC\nggI0bNhQ0/OnpaWhY8eOAIDmzZvjypUrVve9peObmJiIAQMGAACeeeYZHDt2TFPcwIEDMWXKFBgM\nlm8gYClm7ty5GDx48H2v11bcypUrERgYCCEEsrOz0bhxY5sxAPDpp59i9OjRcHZ21jxGWyzFfPfd\nd3j++ecBAC+//HLV/tSaKyUlBYWFhVXH0Vbcnfvuxo0b980VSzEXL16sev6+ffvixx9/vC+Xpfex\nrflhKWbAgAHVzg1rcbbmh6WYjz76qNq5YS3OZDLZnB8yHooC7ujoKNUD9zaj0Qg3NzcAwI4dO/D0\n00/f12PcmpEjR2Lq1KmYOXOm5nxLlixBTEyM1BjPnz+P8ePHY9SoURYnryUZGRkoLS3F+PHjMXr0\naIvFrTqnTp1CkyZN4Ovra3Pb3/3ud7hy5QqeffZZREREYPr06ZpytG7dGkeOHIHJZEJKSgouXbqE\n69evW9zW0vEtKSmpmsDe3t4W/2y1FOfh4VHtuCzFuLm5wWg0wmQy4csvv0R4eLimOAA4dOgQhgwZ\ngtzc3KpiWV1Mamoqzpw5g6FDh0qNEQA2b96MV155BVOmTLnvT3JLMZcvX8ahQ4cQGRmJKVOmWPzF\nVN17a9OmTYiIiNA8xpkzZ2L6WP/9AAAF2klEQVTChAkYPHgwjh8/juHDh9uMad26NX744QcAt5YN\ncnPvvz2CpfexrflhKcbT09Pia7EVZ2t+WKsz1c0Na3Hp6ek254eMh6KAP6j4+Hjs2LED7777ruaY\nLVu2YM2aNYiOjobQ0E3gn//8Jzp37ozAwEDNOYKCgjBx4kSsWbMGS5YswaxZs6yu9d4rPz8fq1at\nwuLFizFjxgxNY7xtx44d9725rPnXv/6Fpk2b4ttvv8XGjRsxf/58TXH9+vXD448/jjFjxmDjxo1o\n2bKl1BjvpBonw2QyYdq0aejRowd69uypOe7pp5/G119/jZYtW2LdunU2t1+0aBFmzJghPb7f//73\nmDp1KjZt2oR27dph1apVNmOEEAgODkZcXBxCQkKwdu1azfnKy8tx/Phx9OjRQ3PMggULsGrVKnzz\nzTfo2rUrvvzyS5sx06dPx1dffYVXXnkFQohqj7W197FKjC33xmmZH/fGaJ0bd8apzg9rdF/ADx8+\njE8//RSfffaZpt/AycnJyMy8dW+ldu3awWQyWf0A4k7ff/89Dh48iJdeegnbt2/HJ598gqNHj1Yb\n4+/vj2HDhsFgMKB58+bw8fFBdna2zVze3t7o0qULHB0d0bx5c7i7u2sa422JiYmaPsAEgF9++QV9\n+vQBALRt2xZXr17VfF++KVOmYMuWLXjvvfdQUFAAb29vzWN0c3NDaWkpACA7O1t6+UzWjBkz0KJF\nC0ycOFFzzLfffgsAMBgMVWed1cnOzkZKSgqmTp2Kl156CVevXrV6hnuvnj17ol27dgCAsLAwnDt3\nzmaMj48PnnzySQBAnz59cP78eU25gFsfSFpaOqnO2bNn0bVrVwBAr169kJxs+z5qTZo0wdq1a7Fp\n0yZ06tQJzZo1s7jdve9jLfND9r1fXZyt+XFvjNa5cWdccXGx8vywRtcFvLCwEEuXLsXatWvRoEED\nTTFJSUn4+9//DgDIzc1FcXGxpnXfjz/+GDt37sS2bdvw4osv4u2330avXr2qjdmzZw82bNgA4Na9\n7q5du6Zpzb1Pnz746aefYDabcf36dc1jBG5Ndnd3d83ray1atMDJkycB3PqT3N3dXdMy1JkzZ6rO\nJA4dOoT27dvDwUH7dOrVqxe++eYbAMCBAwfQt29fzbGy9uzZAycnJ0yaNEkqLjY2Fv/5z60blp48\neRLBwcHVbu/v74/4+Hhs27YN27Ztg5+fn+arqqKionDp0iUAt34Bh4SE2Ix5+umnq64a+u2332yO\n706nT59G27ZtNW8P3PqFcfuXxOnTpzXdQWblypVVV9Ts2rULYWFh921j6X1sa36ovPetxdmaH5Zi\ntMyNe+MeZH5Y81B0I0xOTsaSJUtw+fJlODo6wt/fH7GxsTYPzNatWxEbG3vXzluyZAmaNm1qNaa0\ntBSzZs1CZmYmSktLMXHiRIuTqjqxsbFo1qwZRowYUe12N2/exNSpU1FQUICKigpMnDgR/fr105Rj\ny5Yt2LFjBwDgrbfesvgBlSXJycn4+OOPsX79ek3bFxUVYebMmbh27RoqKysxefJkTUsMZrMZM2fO\nxPnz5+Hi4oJly5ahSZMmVsd07/FdtmwZYmJiUFZWhqZNm2LRokVwcnKyGderVy8cPXoUv/76Kx5/\n/HF07twZ06ZNqzbm2rVrcHFxqVo/b9WqFe7tomwpLjo6GgsXLoTRaISrqyuWLl16118ZtuZtWFgY\nEhISNO2PiIgIrFu3DvXq1YObmxsWLVpkM9eyZcvwwQcfICcnB25ubliyZAl8fHxs5oqNjUVsbCy6\ndu2KYcOGaT5mU6ZMwdKlS+Hk5IT69etj4cKF8PLyqjZm6tSpWLBgAYQQ6Natm8XlA0vv48WLF2P2\n7NlW54elmO7duyMxMdHq3LAWd+XKFXh5eVmdH5ZiJk2ahOXLl1udG9bi7qxP1uaHjIeigBMRkTxd\nL6EQEf03YwEnItIpFnAiIp1iASci0ikWcCIinWIBJyLSKRZwIiKd+j9dPd7pSvygaAAAAABJRU5E\nrkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEZFOsYATEenU/weWqGNjti71aQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0qn/BfPimyElXvUcAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/wclMsedxoeHKwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAENCAYAAAAWpT4gAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n7fH+AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "LqYeodNDozsI",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Generating a large-ish dataset of synthetic data and labels"
      ]
    },
    {
      "metadata": {
        "id": "5QvAqWijoxuy",
        "colab_type": "code",
        "outputId": "e6166df8-d03d-4a94-ded4-f43c488ef22b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "cell_type": "code",
      "source": [
        "label_dict = {\n",
        "    0: \"random\",\n",
        "    1: \"edge\",\n",
        "    2: \"column\",\n",
        "    3: \"row\",\n",
        "    4: \"quad checkers\",\n",
        "    5: \"single checkers\"\n",
        "}\n",
        "\n",
        "\n",
        "label_func_dict = {\n",
        "    0: normal_random_plate,\n",
        "    1: edge_effect_plate,\n",
        "    2: column_effect_plate,\n",
        "    3: row_effect_plate,\n",
        "    4: quad_checker_effect_plate,\n",
        "    5: single_checker_effect_plate\n",
        "}\n",
        "\n",
        "\n",
        "labels  = []\n",
        "dataset = []\n",
        "\n",
        "print(\"Creating synthetic data...\", flush=True)\n",
        "num_samples = int(500000)\n",
        "for i in tqdm(range(num_samples)):\n",
        "    label = sample([0, 1, 2, 3, 4], 1)[0]\n",
        "    # [np.newaxis] adds a channel dimension needed for pytorch\n",
        "    data = label_func_dict[label]()[np.newaxis, ...]\n",
        "    labels.append(label)\n",
        "    dataset.append(data)\n",
        "    \n",
        "\n",
        "# convert to torch tensors and delete originals\n",
        "arr_tensor = torch.from_numpy(np.stack(dataset))\n",
        "del dataset\n",
        "labels_tensor = torch.from_numpy(np.stack(labels))\n",
        "del labels"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Creating synthetic data...\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 500000/500000 [01:02<00:00, 8034.25it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "94Bn1UN4wlnm",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "### Train neural network"
      ]
    },
    {
      "metadata": {
        "id": "3fiKmMcAq7Pg",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "class PlateDataset(torch.utils.data.Dataset):\n",
        "    \"\"\"Plate dataset class\"\"\"\n",
        "    def __init__(self, data_tensor, label_tensor, transform=None):\n",
        "        assert data_tensor.shape[0] == label_tensor.shape[0]\n",
        "        self.data_tensor = data_tensor\n",
        "        self.label_tensor = label_tensor\n",
        "        self.transform = transform\n",
        "        \n",
        "    def __len__(self):\n",
        "        return self.label_tensor.shape[0]\n",
        "    \n",
        "    def __getitem__(self, index):\n",
        "        sample = self.data_tensor[index, ...].type(torch.FloatTensor)\n",
        "        label = self.label_tensor[index]\n",
        "        if self.transform:\n",
        "            sample = self.transform(sample)\n",
        "        return sample, label\n",
        "\n",
        "    \n",
        "def make_dataloader(dataset, batch_size, train=True):\n",
        "    return torch.utils.data.DataLoader(\n",
        "        dataset,\n",
        "        batch_size=batch_size,\n",
        "        shuffle=train,\n",
        "        num_workers=2,\n",
        "        pin_memory=True if device == \"cuda\" else False\n",
        "    )\n",
        " \n",
        "# split into training and test sets\n",
        "# already in a random order so no need to shuffle\n",
        "n = arr_tensor.shape[0]\n",
        "div = n // 3\n",
        "\n",
        "# TODO: check this isn't copying data\n",
        "# hopefully these are just references to the tensor\n",
        "train_data = PlateDataset(\n",
        "    arr_tensor[div:],\n",
        "    labels_tensor[div:]\n",
        ")\n",
        "test_data = PlateDataset(\n",
        "    arr_tensor[:div],\n",
        "    labels_tensor[:div]\n",
        ")\n",
        " \n",
        "def main():\n",
        "    batch_size = 2048\n",
        "    test_batch_size = len(test_data)\n",
        "    epochs = 20\n",
        "    lr = 0.05\n",
        "    momentum = 0.5\n",
        "    \n",
        "    history = {\n",
        "        \"train_loss\": [],\n",
        "        \"test_loss\" : [],\n",
        "        \"train_acc\" : [],\n",
        "        \"test_acc\"  : []\n",
        "    }\n",
        "    \n",
        "    model = SimpleNet().to(device)\n",
        "    optimizer = optim.SGD(model.parameters(), lr=lr, momentum=momentum)\n",
        "    \n",
        "    train_loader = make_dataloader(train_data, batch_size=batch_size, train=True)\n",
        "    test_loader  = make_dataloader(test_data,  batch_size=batch_size, train=False)\n",
        "    \n",
        "    for epoch in tqdm(range(1, epochs + 1)):\n",
        "        running_corrects = 0\n",
        "        running_length = 0\n",
        "        running_loss = 0.0\n",
        "        model.train()\n",
        "        for batch_idx, (data, target) in enumerate(train_loader):\n",
        "            running_length += len(target)\n",
        "            data, target = data.to(device), target.to(device)\n",
        "            optimizer.zero_grad()\n",
        "            output = model(data)\n",
        "            loss = F.nll_loss(output, target)\n",
        "            loss.backward()\n",
        "            optimizer.step()\n",
        "            running_loss =+ loss.item()\n",
        "            pred = output.argmax(dim=1, keepdim=True)\n",
        "            running_corrects += pred.eq(target.view_as(pred)).sum().item()\n",
        "        epoch_train_acc = float(running_corrects) / float(running_length)\n",
        "        epoch_train_loss = float(running_loss) / float(running_length)\n",
        "        history[\"train_acc\"].append(epoch_train_acc)\n",
        "        history[\"train_loss\"].append(epoch_train_loss)\n",
        "       \n",
        "        model.eval()\n",
        "        test_loss = 0\n",
        "        correct = 0\n",
        "        # dont calculate gradients when evaluating as there is no need for backprop\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",
        "        test_loss /= len(test_loader.dataset)\n",
        "        epoch_test_acc = correct / len(test_loader.dataset)\n",
        "        history[\"test_acc\"].append(epoch_test_acc)\n",
        "        history[\"test_loss\"].append(test_loss)\n",
        "        \n",
        "    return history"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "6xFXL2ST9cj8",
        "colab_type": "code",
        "outputId": "a9cfb622-a2d1-479b-f407-e6b7baa817ab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "cell_type": "code",
      "source": [
        "history = main()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 20/20 [11:49<00:00, 34.27s/it]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "metadata": {
        "id": "sEtkfuDCW6vW",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def plot_subplot(hist, name):\n",
        "    plt.plot(\n",
        "        range(1, len(hist[name]) + 1),\n",
        "        hist[name],\n",
        "        label=name\n",
        "    )\n",
        "    plt.grid(linestyle=\":\")\n",
        "    \n",
        "\n",
        "def plot_history(history):\n",
        "    plt.figure(figsize=[8, 8])\n",
        "    ##\n",
        "    plt.subplot(211)\n",
        "    plt.title(\"Accuracy\", fontweight=\"bold\", loc=\"left\")\n",
        "    plot_subplot(history, \"train_acc\")\n",
        "    plot_subplot(history, \"test_acc\")\n",
        "    plt.ylabel(\"Accuracy\")\n",
        "    plt.legend()\n",
        "    ##\n",
        "    plt.subplot(212)\n",
        "    plt.title(\"Loss\", fontweight=\"bold\", loc=\"left\")\n",
        "    plot_subplot(history, \"train_loss\")\n",
        "    plot_subplot(history, \"test_loss\")\n",
        "    plt.grid(linestyle=\":\")\n",
        "    plt.xlabel(\"Epochs\")\n",
        "    plt.ylabel(\"Loss\")\n",
        "    plt.yscale(\"log\")\n",
        "    plt.legend()\n",
        "    plt.tight_layout()\n",
        "    plt.show()\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Djp4nEwTsqrH",
        "colab_type": "code",
        "outputId": "8a952981-cd25-4693-86fc-432adbb7219d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 585
        }
      },
      "cell_type": "code",
      "source": [
        "plot_history(history)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjgAAAI4CAYAAABndZP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}