CNN/mnist_cnn_pytorch.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "mnist_cnn_pytorch.ipynb",
      "version": "0.3.2",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "metadata": {
        "id": "PH1jL1zFoxn_",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# CNN\n",
        "\n",
        "In this notebook, we will go through basics of CNN using MNIST dataset.\n",
        "\n",
        "Here we will use [PyTorch](https://pytorch.org/tutorials/  \"PyTorch Tutorial\").\n",
        "\n",
        "\n",
        "Hey yo, but what is CNN?\n",
        "\n",
        "Everything is explained in-detail in [blog post](https://dudeperf3ct.github.io/cnn/mnist/2018/10/17/Force-of-Convolutional-Neural-Networks/). This is notebook which replicates the result of blog and runs in colab. Enjoy!\n",
        "\n",
        "\n",
        "#### Run in Colab\n",
        "\n",
        "You can run this notebook in google colab.\n",
        "\n",
        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/dudeperf3ct/DL_notebooks/blob/master/CNN/mnist_cnn_pytorch.ipynb)"
      ]
    },
    {
      "metadata": {
        "id": "TGU-my-xquin",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# Install dependencies"
      ]
    },
    {
      "metadata": {
        "id": "X2-3LGhFqssP",
        "colab_type": "code",
        "outputId": "0965d7cb-91c2-45a9-edbb-476fff76b577",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "cell_type": "code",
      "source": [
        "!pip install --upgrade torch torchvision"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already up-to-date: torch in /usr/local/lib/python3.6/dist-packages (1.0.0)\n",
            "Requirement already up-to-date: torchvision in /usr/local/lib/python3.6/dist-packages (0.2.1)\n",
            "Requirement already satisfied, skipping upgrade: six in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.11.0)\n",
            "Requirement already satisfied, skipping upgrade: numpy in /usr/local/lib/python3.6/dist-packages (from torchvision) (1.15.4)\n",
            "Requirement already satisfied, skipping upgrade: pillow>=4.1.1 in /usr/local/lib/python3.6/dist-packages (from torchvision) (5.3.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "EqtWax9oq2t9",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# PyTorch"
      ]
    },
    {
      "metadata": {
        "id": "LfvSJBZRoxoa",
        "colab_type": "code",
        "outputId": "f0cf58a2-1836-4ae0-8c6a-6b146473c066",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "cell_type": "code",
      "source": [
        "# load all the required libraries\n",
        "import copy\n",
        "import time\n",
        "import numpy as np                                    # package for computing\n",
        "from sklearn.model_selection import train_test_split  # split dataset\n",
        "import torch                                          # import keras with tensorflow as backend\n",
        "from torchvision import datasets                      # import mnist dataset from keras \n",
        "from torchvision import transforms            # sequential and functional api keras \n",
        "from torch.utils.data.sampler import SubsetRandomSampler\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "from torch import optim\n",
        "import matplotlib.pyplot as plt             # matplotlib library for plotting\n",
        "\n",
        "# display plots inline (in notebook itself)\n",
        "%matplotlib inline              \n",
        "\n",
        "# Detect if we have a GPU available\n",
        "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
        "\n",
        "print('Training on %s ...'%format(device))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training on cuda:0 ...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "21saVR_Foxoo",
        "colab_type": "code",
        "outputId": "22a06529-bcd9-474d-ed76-c020b8ede986",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 72
        }
      },
      "cell_type": "code",
      "source": [
        "# load mnist data\n",
        "\n",
        "batch_size = 256\n",
        "num_workers = 0\n",
        "val_size = 0.2\n",
        "\n",
        "transform = transforms.Compose([transforms.ToTensor()])\n",
        "\n",
        "# choose the training and test datasets\n",
        "train_data = datasets.MNIST('data', train=True,\n",
        "                              download=True, transform=transform)\n",
        "test_data = datasets.MNIST('data', train=False,\n",
        "                             download=True, transform=transform)\n",
        "\n",
        "# obtain training indices that will be used for validation\n",
        "num_test = len(test_data)\n",
        "indices = list(range(num_test))\n",
        "np.random.shuffle(indices)\n",
        "split = int(np.floor(val_size * num_test))\n",
        "test_idx, val_idx = indices[split:], indices[:split]\n",
        "\n",
        "# define samplers for obtaining training and validation batches\n",
        "test_sampler = SubsetRandomSampler(test_idx)\n",
        "valid_sampler = SubsetRandomSampler(val_idx)\n",
        "\n",
        "# prepare data loaders (combine dataset and sampler)\n",
        "train_loader = torch.utils.data.DataLoader(train_data, batch_size=batch_size,\n",
        "    num_workers=num_workers)\n",
        "val_loader = torch.utils.data.DataLoader(test_data, batch_size=batch_size, \n",
        "    sampler=valid_sampler, num_workers=num_workers)\n",
        "test_loader = torch.utils.data.DataLoader(test_data, batch_size=batch_size, \n",
        "    sampler=test_sampler, num_workers=num_workers)\n",
        "\n",
        "print('Training samples:', len(train_loader.dataset))\n",
        "print ('Validation samples:', len(val_loader.dataset))\n",
        "print('Test samples:', len(test_loader.dataset))\n",
        "\n",
        "# Create training and validation dataloaders\n",
        "dataloaders_dict = {'train': train_loader, \n",
        "                    'val': val_loader}\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Training samples: 60000\n",
            "Validation samples: 10000\n",
            "Test samples: 10000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "rUvtxpRXoxo7",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "# \n",
        "\n",
        "From the previous post, we learned how to split dev and test data.\n",
        "Here, we keep the dimensions of the data same.\n",
        "\n",
        "\n"
      ]
    },
    {
      "metadata": {
        "id": "kZ71Q_Hfoxo-",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Visualization of data\n",
        "\n",
        "Enough talk, show me the data!"
      ]
    },
    {
      "metadata": {
        "id": "7t2g8PbroxpB",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# helper function to un-normalize and display an image\n",
        "def imshow(img):\n",
        "    img = img / 2 + 0.5  # unnormalize\n",
        "    plt.imshow(np.squeeze(np.transpose(img, (1, 2, 0))), cmap='gray')  # convert from Tensor image"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "XLSKBfA5oxpN",
        "colab_type": "code",
        "outputId": "8c6c2100-cb1c-4c63-967d-31af1b6454b7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "cell_type": "code",
      "source": [
        "dataiter = iter(train_loader)\n",
        "images, labels = dataiter.next()\n",
        "images = images.numpy()\n",
        "\n",
        "# plot the images in the batch, along with the corresponding labels\n",
        "fig = plt.figure(figsize=(25, 4))\n",
        "#display 20 images\n",
        "for idx in np.arange(20):\n",
        "    ax = fig.add_subplot(2, 20/2, idx+1, xticks=[], yticks=[])\n",
        "    imshow(images[idx])\n",
        "    # print out the correct label for each image\n",
        "    ax.set_title(str(labels[idx]))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABW4AAAD2CAYAAABGOR6XAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzs3XncTeX6x/HvE5nKkKH5oMKSWUp0\nlJLUUWlWokGdxlOaSHOdJJImqZDS5FSiaCS/UEmDkk7CCoWISMqYIc/vj73d3fc69tO27eHej8/7\n9fJ6XZdr77VuXNbeez1rXbugsLBQAAAAAAAAAAB/7JLrBQAAAAAAAAAAXJy4BQAAAAAAAADPcOIW\nAAAAAAAAADzDiVsAAAAAAAAA8AwnbgEAAAAAAADAM5y4BQAAAAAAAADPlMz1ArYKguCSMAyfzPEa\nykuaIulsSTdKOl7Sb9ZDzpdUWVJPSceGYbgl64uE4WHPzJX0uKSjJP0h6YkwDAcEQXCC6Blv+NY3\nYRjOjP/eLpI+ljQrDMMLgyC4TFKrMAzPy+FSEedj3wRB0EzSCEkTwzD8Z/wxHG884lvfSFohaZCk\n+pIKJV0VhuF4jjd+8bBvFin2/uZQxS76eCkMwzs43vjFt77hdSo/+NY3khZLGiypiWLHm5fDMLyd\nvvGHbz3DsSY/+Ng38d/Li8/gXlxxGwRBCUn353odkvpJ+s/Wf0RJN4dhWNf69VkYhmMlLZR0de6W\nCU975nrFTuzXlXS4pGuDIDiUnvGHp32z1RWS9tqahGE4WFL1IAhOyfbi4PKxb4IgaC3paUmf2Q/g\neOMPH/tG0gBJ88IwrCPpTEkvBEFQnuONPzztm3slbZRUT1IzSZ2DIDiO440/fOwbXqf852PfxOMl\nYRjWldRcseNNe/rGDz72DMca//nYN9bv5cVncF+uuB0vqWIQBLMlXaLY1a5BvHZNGIbvBEFQU7Ez\n4X3ij6ks6fowDF8OgmA/Sc9J2kdSacWuBLg1fva8l6Qz4tv6RNK/wjBcGwTBJEkfSTpd0sWK/ac+\nS9KBSay3r6T3giB4IgzDjTv2R0eKfOyZsyTdGv+J3qogCEbGf+9z0TO+8LFvFATBPoq9qXhIUlNr\nvX0l3SNpTDr/ErDdfOyb5ZKOVOwHRvtH1svxxg8+9s1xklpJUhiGXwdB8IWkYyWNFscbX/jYN69K\nmhN/f7M6CIKvFLtqe7w43vjCx77hdcp/PvbNKEn/laQwDH8NgmBafE1vi77xgY89w7HGfz72TV59\nBvfiiltJF0n6I/6TtbskTY9fDdJesatBqsQfV1XSljAMG0q6VrG/TMXjD8IwrCepoaQD4/8IHSX9\nQ7GrA+pLqiTpOmu/zSTVD8NwiqQOkiaHYbjKqp8bBMHUIAhmBkFwSxAEBZIUhuEsxUYoHJnWvwVs\nDx97po6kedZj5yl29S094w8f+0aSHpb0b7mjWaTYi1ydIAgO2uE/OXaEd30ThuHMSA8ZHG+84V3f\nKDYeoYT12DWSasVjjjd+8K5vwjCcEIbhD5IUBEEFSUdI+jRe43jjBx/7htcp//nYN++GYbhUkoIg\nqCPpMEnvxmv0Te752DMca/znXd/E5c1ncF9O3EqSgiDYTdIxip3xVhiGcyV9KOnE+ENKShoWj6dJ\nqh6Pl0k6PgiCVpI2hGHYKQzDJfHnPRuG4dowDP+IP7edtcu3wz/nnTSXNNWqvS/pZUktFJt1e74k\ne87Fp5Ja7tifGDvKs54pJ+l3K18vaTcrp2c84VPfxOcv7RGG4YvRdYZhuFnSF6JvvOBT3ySB440n\nPOub8YqN8SkRBEEjSW0klYmvi+ONRzzrm61rKiXpP5JeD8PwY6vE8cYTPvZNEegbT/jWN/HXqLmS\nvpTULwzDb6wyfeMB33rmL9AznvCpb/LtM7gvoxK2qiipQNKUINh65bR2lzQhHv8RhuHarbH+vGrk\noXj8uKR9gyB4TLEz+dUkrbS2v1LSnlb+ixXvqdil1JKkMAyHWbUfgiAYIukkxS7RlmLNY28LueFN\nz0haq/gH4Lhyil3NtBU94w8v+iYIgrKS+ks6tYi10jf+8KJvkkTf+MOnvukm6QlJsxT7QDxW0q9W\nnb7xh099oyAIdldsZMIiSZdH1krf+MOrvvkL9I0/vOqb+AmYWkEQVJM0OgiCP8IwHBQv0zd+8Kpn\n/gI94w8v+iYfP4P7duJ2mWL/QIeGYWif8FJ85sU2xc+I95XUN35LxTuSJkv6SVIV66FV4r+3LQWR\n/TVQbJ7XhvhvlZS0Kek/CbLFm56RNFuxW07nxPPakmYKPvKlb5opNotpcvzFq6ykUkEQVAvD8MRt\nPRk55UvfIL940zdhGC7Tn3PAFATBBElfJ/sHQVZ50zdBEJSU9JqkGWEYXrftp8AT3vQN8oo3fRME\nwXmS3gjD8NcwDJcHQfCSpBMkDdrms5Er3vQM8oovfZN3n8F9GZWwSbG1lJX0luI/yQ+CoFwQBE8H\nQfC3op4cBMHgIAiOi6fzJC1VbI7bm5K6xLdTUrGhxG8l2Mwyxc7YbzVEsStTFATBHoqNSrCfW02x\nQdjIDR97ZoSkq+O39+wj6RzFxm1sRc/knld9E4bh5DAMK4VhuHcYhntLukbSy5EXDPom97zqmyTR\nN7nnXd8EQTAwCILr4vHRkvZT7I3vVvRN7nnXN4q9H15dxElb+ib3fOybv0Lf5J6PfdNVsXmWCoJg\nV8VGFv7XqtM3ueVjz/wVeib3vOqbfPwM7suJ2yWKfXBYKOk+Sa2D2DfOTZP0XRj/QoQiDJLUO/6c\nmYp9G917kkYq9g2UX0iaIekHSQMSbOMzxYafb3W+pBOCIPhW0hTFZnrZ8y8Oj+8HueFjzzwi6UdJ\noaSJku4Ow/Arq07P5J6PfZNQEAQlFPuJIH2TW971TRAEveLbu0rSmUEQzA6CoI/1eI43uedd30ga\nKKlzEATzFLtF7Mz4Lakcb/zhY99cJql5/Diz9Vcvq87xJve86xtep/KCd32j2Inb5vFtzpK0Ir62\nreib3PKuZzjW5AXv+qYoPr4nLigsLMz1GrwQP8s/TdIB0cu2t/HYQNIkSTWtUQrYydAzSMV29k07\nxb6UoUlWFgdvcbxBKjjeIBUcb5AK+gapoG+wvegZpCLf3xP7csVtzsXP8o+WdEUSD79RUn/+8+/c\n6BmkYjv7pqekXn/5KBR7HG+QCo43SAXHG6SCvkEq6BtsL3oGqcj398ScuHXdIOm8IAgOTvSA+Nn3\nWpIeztqq4DN6BqlIpm8ukbQ0DMNR2VsWPMfxBqngeINUcLxBKugbpIK+wfaiZ5CKvH1PzKgEAAAA\nAAAAAPAMV9wCAAAAAAAAgGdKFlUsKCjgclxPFBYWFuR6Dcmib/xB3yAV9A1SQd8gFfQNUpEvfUPP\n+CNfekaib3xC3yAV9A1SUVTfcMUtAAAAAAAAAHiGE7cAAAAAAAAA4BlO3AIAAAAAAACAZzhxCwAA\nAAAAAACe4cQtAAAAAAAAAHiGE7cAAAAAAAAA4BlO3AIAAAAAAACAZzhxCwAAAAAAAACe4cQtAAAA\nAAAAAHiGE7cAAAAAAAAA4BlO3AIAAAAAAACAZzhxCwAAAAAAAACe4cQtAAAAAAAAAHimZK4XAOSD\nZs2aOflVV11l4vPPP9+pPffccyZ+9NFHndq0adMysDoAAAAgNY888oiTd+vWzcQzZsxwaieddJKT\nL1iwIHMLAwDAQ++9956JCwoKnFqbNm3Svj+uuAUAAAAAAAAAz3DiFgAAAAAAAAA8w4lbAAAAAAAA\nAPBMsZhxW6JECSevWLFi0s+1Z5WWK1fOqQVBYOJ//etfTq1///4m7tSpk1P7/fffTdy3b1+n9u9/\n/zvptSF3mjRp4uTjx4938goVKpi4sLDQqZ133nkm7tChg1OrUqVKupaIncixxx5r4uHDhzu11q1b\nmzgMw6ytCX647bbbTBx9fdlllz9/Nnv00Uc7tffffz+j6wKQH8qXL+/ku+++u4lPPPFEp1atWjUT\nP/jgg05tw4YNGVgdMqlmzZom7tKli1PbsmWLiQ8++GCnVrduXSdnxu3OpU6dOibeddddndpRRx1l\n4scff9yp2T21I8aMGWPic845x6lt3LgxLftAZkX75ogjjjDxvffe69T+/ve/Z2VNwF956KGHnNzu\nW/s7jjKFK24BAAAAAAAAwDOcuAUAAAAAAAAAz3g1KqF69epOXqpUKRPblyJLUqtWrUxcqVIlp3bG\nGWekZT2LFi0y8YABA5zaaaedZuLVq1c7ta+++srE3I6aP5o3b27iUaNGObXo+A17PEL039++TSc6\nGqFFixYmnjZtWsLnIXn2bVmS+3f+2muvZXs5GXHYYYeZeOrUqTlcCXLtwgsvdPKePXuauKjbEKMj\nXQDsPOxb4u1jhiS1bNnSyRs0aJDUNvfZZx8n79atW2qLQ84sX77cxB988IFTi476ws6lfv36Jo6+\n7zjrrLNMbI9kkqR9993XxNH3JOl6H2L35qBBg5zatddea+JVq1alZX9Iv+jn6okTJ5p46dKlTm3v\nvfd28mgdyCR77Onll1/u1DZt2mTi9957L+Nr4YpbAAAAAAAAAPAMJ24BAAAAAAAAwDOcuAUAAAAA\nAAAAz+R8xm2TJk1MPGHCBKcWnX+SadFZPLfddpuJ16xZ49SGDx9u4iVLlji1lStXmjgMw3QuETuo\nXLlyJj7kkEOc2gsvvGDi6Oy2osyZM8fJ+/XrZ+KXXnrJqX300UcmtvtLkvr06ZP0PvGno48+2slr\n165t4nydcRudGXbAAQeYuEaNGk6toKAgK2uCH6L//mXKlMnRSpANhx9+uJN36dLFxK1bt3Zq9kzC\nqO7duzv5jz/+aGL7OwMk97Xw008/TX6xyKm6deua2J7zKEmdO3c2cdmyZZ1a9DXkhx9+MHF0hv/B\nBx9s4o4dOzq1xx9/3MSzZ89OdtnIobVr15p4wYIFOVwJfGN/Jmnfvn0OV1K0888/38mfeuopE9uf\nuZA/ojNtmXGLXLK/n2jXXXd1apMnTzbxiBEjMr4WrrgFAAAAAAAAAM9w4hYAAAAAAAAAPJPzUQkL\nFy408YoVK5xaOkYlRG/z+/XXX538mGOOMfHGjRud2vPPP7/D+4dfBg8ebOJOnTqlZZvRkQu77767\nid9//32nZt/W36hRo7Tsf2cXvU3q448/ztFK0ic6quOSSy4xsX0bs8QtqTuDtm3bmvjqq69O+Lho\nL5x00kkm/umnn9K/MGTE2WefbeJHHnnEqVWtWtXE0VvcJ02a5OTVqlUz8f33359wf9Ht2M8755xz\n/nrByBr7ffF9993n1Oy+KV++fNLbjI57Ov74400cvS3QPsbYvbitHP6rVKmSiRs3bpzDlcA348eP\nN3FRoxKWLVvm5PaogujYr+hIQtsRRxzh5NFRQNh5MAIOiRx11FFOfuutt5o4el7nl19+SWkf0e00\naNDAxPPmzXNq0TFkmcYVtwAAAAAAAADgGU7cAgAAAAAAAIBnOHELAAAAAAAAAJ7J+Yxbe/5Ejx49\nnJo9n+/LL790agMGDEi4zenTp5v4uOOOc2pr16518vr165v4mmuuSWLFyCfNmjVz8hNPPNHERc3Q\nic6mfeONN5y8f//+Jv7xxx+dmt2rK1eudGpt2rRJav9IXnSGVnEwdOjQhLXoPEIUP61atXLyYcOG\nmbio2e/ROaYLFixI78KQNiVL/vn269BDD3VqTz75pInLlSvn1D744AMT9+rVy6lNnjzZyUuXLm3i\nESNGOLV27dolXNvnn3+esIbcOu2000z8z3/+M6VtRGe0Rd8n//DDDyauVatWSvtAfrCPL9WrV0/6\neYcddpiT27OPed0pHp544gkTjx49OuHjNm3a5ORLly5NaX8VKlRw8hkzZph43333Tfi86Np4/cp/\nhYWFTl6mTJkcrQS+GTJkiJPXrl3bxPXq1XNq0ffEybrlllucvEqVKia2v3NGkr766quU9pGq4nfG\nAwAAAAAAAADyHCduAQAAAAAAAMAzOR+VYIve7jBhwgQTr1692qk1btzYxBdffLFTs29jj45GiPrm\nm29MfOmllya/WHirSZMmJh4/frxTs2/Fid6K8c4775i4U6dOTq1169ZOftttt5k4elv78uXLTRy9\nhH7Lli0mtsc2SNIhhxxi4mnTpgmJNWrUyMR77bVXDleSGUXdDh/taRQ/F1xwgZMXdZvgpEmTTPzc\nc89laklIsy5dupi4qNEo0f/vZ599tolXrVpV5D7sxxY1GmHRokVO/uyzzxa5XeTOWWedldTj5s+f\n7+RTp041cc+ePZ2aPRoh6uCDD05+ccg79qivZ555xqndddddCZ8Xrf36668mHjhwYDqWhhzbvHmz\niYs6RqTL8ccf7+R77LFHUs+Lvn5t2LAhbWuCH6LjpD755JMcrQS5tm7dOie3z+XsyEgN+9xRjRo1\nnJp97ibXYzu44hYAAAAAAAAAPMOJWwAAAAAAAADwDCduAQAAAAAAAMAzXs24jSpqfttvv/2WsHbJ\nJZeY+OWXX3Zq9pwKFA916tRx8h49epg4Oiv0559/NvGSJUucmj3Xb82aNU7trbfeKjJPRdmyZZ38\nhhtuMHHnzp13ePvFWfv27U0c/XvMV/as3gMOOCDh4xYvXpyN5SCLqlat6uQXXXSRk9uvW/YsQUm6\n5557MrcwpE2vXr2c/JZbbjFxdN76448/bmJ7nrr013NtbbfeemtSj+vWrZuT23Pa4Rf7/W30exne\nffddE8+dO9epLVu2LKX9FccZ8ti26DGqqBm3QDqcc845JraPbVLy7+3vuOOOtK4J2WHPUJbc8zrR\nz+4HHXRQVtYEP9mvTQ0bNnRqs2bNMnH0e4WKsttuuzm5Pfu/XLlyTs2eqTxy5Mik95EJXHELAAAA\nAAAAAJ7hxC0AAAAAAAAAeMbrUQlFsW/hadasmVNr3bq1idu2bevU7FvJkL9Kly5t4v79+zs1+zb6\n1atXO7Xzzz/fxJ9//rlTy/Ut99WrV8/p/vNJEAQJa998800WV5I+dh9Hb0/99ttvTRztaeSnmjVr\nmnjUqFFJP+/RRx918okTJ6ZrSUgz+xZOezSCJG3cuNHE48aNc2r2LVvr169PuP0yZco4ebt27Zzc\nfk0pKChwavaIjTFjxiTcB/zy448/mjgbt7K3bNky4/uAn3bZ5c9rexgzh1REx77ddNNNTl6rVi0T\n77rrrklvd/r06SbetGlTiqtDLkXHfn344YcmPumkk7K9HHjkb3/7m5PbY1SiIzauuuoqE2/PmK8H\nH3zQyc866ywT2++zJOnvf/970tvNNK64BQAAAAAAAADPcOIWAAAAAAAAADzDiVsAAAAAAAAA8Eze\nzrhdu3atie3ZF5I0bdo0Ez/55JNOLToP0J5z+thjjzm1wsLCHV4nMqNp06YmtmfaRp1yyilO/v77\n72dsTfDD1KlTc70Eo0KFCk5+wgknmLhLly5OLTqf0tarVy8TR+dCIT/ZvdCoUaMiH/vee++Z+JFH\nHsnYmrBjKlWq5ORXXnmliaPvJ+y5tqeeemrS+7BnAg4fPtypRef920aOHOnk/fr1S3qfyH/dunUz\n8W677Zb08xo2bJiwNmXKFCf/+OOPt39h8JY915bPQzsfew7/eeed59Si3x+TSKtWrZx8e/po1apV\nJo7Oxn377bdNXNQceAD5oUGDBiZ+7bXXnFrVqlVNHP2ej+05r9O9e3cTX3jhhQkf17t376S3mW1c\ncQsAAAAAAAAAnuHELQAAAAAAAAB4Jm9HJdjmzZvn5Pblz8OGDXNq0ds97Dx6+9hzzz1n4iVLluzo\nMpFGDz74oIkLCgqcmn3ZvG+jEXbZ5c+fldi3oSF9KleunNLzGjdubOJoT9m3he2///5OrVSpUibu\n3LmzU7P/vSX3lq5PP/3UqW3YsMHEJUu6h+YvvviiyLXDf9Hb4fv27ZvwsZMnT3byCy64wMS//fZb\neheGtLGPBZJ7e1eUfev6nnvu6dS6du1q4g4dOjg1+3ay3Xff3alFb0O18xdeeMGp2eOmkJ/KlSvn\n5PXq1TPxnXfe6dSKGikVfZ0q6r3Jjz/+aGK7TyXpjz/+SLxYAF6zX1sk6fXXXzdx9erVs70cffjh\nhyYeMmRI1vcPf1SpUiXXS0Aa2J9to+MCn3rqKRMX9Z6kZcuWTu3mm282sX1uSPrf8wFnnXWWiaOf\n8+1zfoMHD972H8ADXHELAAAAAAAAAJ7hxC0AAAAAAAAAeIYTtwAAAAAAAADgmWIx4zbqtddeM/Gc\nOXOcWnT+xbHHHmvie++916nVqFHDxL1793Zqixcv3uF1InknnXSSkzdp0sTE0bl+9lwm39hzWqLr\nnj59eraXk7fsWbHRv8dBgwaZ+JZbbkl6m40aNTJxdPbN5s2bTbxu3TqnNnPmTBM//fTTTu3zzz93\ncnvm8k8//eTUFi1aZOKyZcs6tdmzZxe5dvipZs2aJh41alTSz/vuu++cPNor8NPGjRudfPny5Sau\nVq2aU/v+++9NHD2GFcWeMbpq1Sqnts8++zj5zz//bOI33ngj6X3AH7vuuquTN23a1MTRY4r972+/\nRkpu33z88cdO7YQTTnDy6Oxcmz2j7vTTT3dqjzzyiImj/xcA5Bf7fXD0PXGytmd+dpT9ue8f//iH\nU3vnnXdSWg/yU3TWP/LTOeecY+KhQ4c6Nft9cPQ4MXfuXBMfeuihTs3OTznlFKe23377Obn9Hsl+\nfy5JF110UZFr9wVX3AIAAAAAAACAZzhxCwAAAAAAAACe4cQtAAAAAAAAAHimWM64tc2YMcPJO3bs\n6OQnn3yyiYcNG+bULrvsMhPXrl3bqR133HHpWiKSEJ35WapUKRMvW7bMqb388stZWVMipUuXNvFd\nd92V8HETJkxw8ptvvjlTSyp2rrzyShMvWLDAqR1xxBEpbXPhwoUmHj16tFObNWuWiT/55JOUth91\n6aWXOrk9AzM64xT5qWfPnibentluffv2zcRykGG//vqrk5966qkmfvPNN51a5cqVTTxv3jynNmbM\nGBM/88wzTu2XX34x8UsvveTUojNuo3XkB/v9TXT+7Kuvvprwef/+979NHH1/8dFHH5nY7r1tPbZB\ngwYJ92G/TvXp08epFfUaumHDhoTbhJ/s+aR/9fp11FFHmXjgwIEZWxMyJ/p5+eijjzZxly5dnNq4\nceNM/Pvvv6e8z4svvtjEV199dcrbQf6bOHGiiaPfa4P8dPbZZzu5fZ5t06ZNTs1+/3zuuec6tZUr\nV5r4gQcecGqtW7c2cXT+bXQ2tz1Ht2rVqk7thx9+MLF97JP+9z16LnHFLQAAAAAAAAB4hhO3AAAA\nAAAAAOCZYj8qISp6K+Pzzz9v4qFDhzq1kiX//OuxbwOS3MuoJ02alL4FYrtFb8FbsmRJVvdvj0aQ\npNtuu83EPXr0cGqLFi0ycfRy/zVr1mRgdcXffffdl+slpOTYY49NWBs1alQWV4J0adKkiZO3a9cu\nqefZt8ZLUhiGaVsTcufTTz81sX2L+Y6w34vYt4hJ/3s7MyNX8sOuu+7q5PbIg+h7CNs777zj5I8+\n+qiJo+917f57++23nVrDhg2dfOPGjSbu16+fU7PHKJxyyilObfjw4Sb+v//7P6dmv07btz1GTZ8+\nPWEN2WUfT+xbTLfl9NNPN3G9evWc2syZM9O7MGSFPYasd+/eGdmHPU6OUQk7N3vUTlT0NbJGjRom\njo7Lgz/skaOS+298zz33OLXouNJEoseJwYMHm7hly5ZJry06RsEe1eHTaIQorrgFAAAAAAAAAM9w\n4hYAAAAAAAAAPMOJWwAAAAAAAADwTLGfcduoUSMnP/PMM538sMMOM7E90zYqOqPpgw8+SMPqkA6v\nv/561vdpz7KMzqA7++yzTRydXXnGGWdkdmEoFl577bVcLwEpePfdd518jz32SPjYTz75xMQXXnhh\nppaEYqZs2bImjs60jc6hfOmll7KyJmy/EiVKmLhXr15OrXv37iZeu3atU7vppptMHP33tefaHnro\noU5t4MCBJm7atKlTmzNnjpNfccUVJrbnvklShQoVTHzEEUc4tc6dO5u4Q4cOTm38+PFK5IcffjDx\nAQcckPBxyK5BgwaZODqrsCiXXnqpk1977bVpWxOKl+OPPz7XS4AnNm/enLAWnUca/W4Z+Cl6DuTV\nV181sf26vz2qVq3q5Pbc/ahOnTo5+YwZMxI+1v4OIp9xxS0AAAAAAAAAeIYTtwAAAAAAAADgmWIx\nKiEIAie/6qqrTHz66ac7tb333jvp7f7xxx8mXrJkiVOL3qKIzIreJmHnp556qlO75ppr0r7/6667\nzslvv/12E1esWNGpDR8+3MTnn39+2tcCwE9VqlRx8qJeJx5//HETr1mzJmNrQvEybty4XC8BaWDf\nTm6PRpCkdevWmTh6i7o9jqVFixZOrWvXrib+xz/+4dTsERt33323Uxs2bJiTF3UL46pVq0w8duxY\np2bn0VsUzz333ITbjL6/gh9mz56d6yUgzXbddVcnb9eunYknTJjg1NavX5/2/dvHKEl65JFH0r4P\n5Cf7tvrosadu3bpObo9fufLKKzO7MKQsXf+/7fMsZ511llOzxzfNmzfPqY0YMSIt+/cJV9wCAAAA\nAAAAgGc4cQsAAAAAAAAAnuHELQAAAAAAAAB4Jm9m3EZn09rzs+yZtpJUs2bNlPbx+eefO3nv3r1N\n/Prrr6e0TaRHYWFhwjzaGwMGDDDx008/7dRWrFhh4uh8uPPOO8/EjRs3dmr777+/ky9cuNDE0ZmD\n9uxKIFn23OY6deo4tU8++STby0GS7PmQu+yS/M9Cp0yZkonloJg7/vjjc70EpMEdd9yRsFaiRAkT\n9+jRw6ndddddJq5Vq1bS+7Of16dPH6dmf59Durz44otF5vDfo48+auKrr77aqR100EEJnxf9ngl7\nO9EZhMi8Vq1amfjWW291ascdd5yJDzjgAKdW1KzrolSuXNnE7du3d2oPPvigk5crVy7hduwZu7//\n/ntKa0F+sme5S9J+++3n5Ndff302l4Mcs+cYX3HFFU5t2bJlJm7Tpk3W1pQrXHELAAAAAAAAAJ7h\nxC0AAAAAAAAAeMarUQl77bXzPwa0AAAgAElEQVSXk9erV8/EAwcOdGp169ZNaR+ffvqpk99///0m\nHjNmjFPbsmVLSvtAdtm3FUruJfVnnHGGU1u1apWJa9eunfQ+orc1T5w40cRF3fIIJMse/7E9t9wj\nu5o0aeLkbdu2NXH0NWPjxo0mfuyxx5zaTz/9lIHVobg78MADc70EpMHSpUtNXK1aNadWunRpE0fH\nNtnefvttJ//ggw9MPHr0aKc2f/58E2diNAKKt2+++cbJizoO8dnJL/bn5wYNGiR83I033ujkq1ev\nTml/9viFQw45xKlFx97ZJk2a5ORPPPGEie3PXNj5RPvGfm+N4qdGjRpO/s9//tPE0V4YMmSIiRct\nWpTZhXmAswMAAAAAAAAA4BlO3AIAAAAAAACAZzhxCwAAAAAAAACeyfqM28qVKzv54MGDTRydHZjq\nLDd7HukDDzzg1MaNG+fk69evT2kfyK6PP/7YyadOnWriww47LOHz9t57byePzlG2rVixwsQvvfSS\nU7vmmmuSWieQDi1btnTyZ555JjcLwf+oVKmSk0ePMbbFixebuHv37hlbE3YeH374oYmjs7CZLZk/\njjrqKBOfeuqpTs2eC7ls2TKn9vTTT5t45cqVTo25f8gUe46gJJ188sk5Wgky5Yorrsj4PqLHszfe\neMPE0c9Zv//+e8bXg/xQoUIFJz/llFNM/Nprr2V7Ociw8ePHO7k98/aFF15wanfeeWdW1uQLrrgF\nAAAAAAAAAM9w4hYAAAAAAAAAPJORUQmHH364k/fo0cPEzZs3d2r77bdfSvtYt26diQcMGODU7r33\nXhOvXbs2pe3DL4sWLXLy008/3cSXXXaZU7vtttuS2uYjjzzi5E888YSJ586du71LBHZIQUFBrpcA\nwHMzZsww8Zw5c5xadLzUQQcdZOLly5dndmHYLqtXrzbx888/79SiOZBrM2fOdPJZs2Y5+cEHH5zN\n5WA7XHjhhSa++uqrndoFF1yww9ufN2+ek9ufz+3RPtL/jtywX8+ArTp27OjkGzZscPLo8QfFy7Bh\nw5y8V69eJh4zZky2l+MVrrgFAAAAAAAAAM9w4hYAAAAAAAAAPMOJWwAAAAAAAADwTEFhYWHiYkFB\n4mIR+vbt6+T2jNuiRGcovfnmmybevHmzU3vggQdM/Ouvv27vEvNOYWFh3gzATLVvkH70jb/suWOS\n9PTTT5v4ySefdGrROc6ZRt8ktvfeezv5yy+/bOJWrVo5te+//97EtWrVyuzCPEDfZFf0GDJ06FAn\nf//9900cnW0Yfb+VS/QNUpEvfUPP+CNfekbKTN+ULl3aye3XkHvuucep7bHHHiYePXq0Uxs/fryJ\nozMnly5duqPL9M7O3jfZ9tJLLzl5dIZ2hw4dTLxgwYKsrCkV9A1SUVTfcMUtAAAAAAAAAHiGE7cA\nAAAAAAAA4JmMjEpA+nG5PVJB3yAV9A1SQd9kV4UKFZx8xIgRTt62bVsTv/rqq06ta9euJl67dm0G\nVpc8+gapyJe+oWf8kS89I9E3PqFvkAr6BqlgVAIAAAAAAAAA5BFO3AIAAAAAAACAZzhxCwAAAAAA\nAACeYcZtnmBOClJB3yAV9A1SQd/kVnTmbe/evU18xRVXOLVGjRqZeObMmZld2F+gb5CKfOkbesYf\n+dIzEn3jE/oGqaBvkApm3AIAAAAAAABAHuHELQAAAAAAAAB4hlEJeYLL7ZEK+gapoG+QCvoGqaBv\nkIp86Rt6xh/50jMSfeMT+gapoG+QCkYlAAAAAAAAAEAe4cQtAAAAAAAAAHiGE7cAAAAAAAAA4Jki\nZ9wCAAAAAAAAALKPK24BAAAAAAAAwDOcuAUAAAAAAAAAz3DiFgAAAAAAAAA8w4lbAAAAAAAAAPAM\nJ24BAAAAAAAAwDMlc72ArYIguCQMwydzvIbykqZIOlvSYkmDJTVR7AT3y2EY3h4EwQmSeko6NgzD\nLTlbLHzsmRslHS/pN+sh50uqLHrGGx72DceaPOBb34RhODMIgisUO+5I0ruSrpJ0rOgbb3jaN80k\njZA0MQzDf8Yfw/HGI771jaTHJe1tlatKelbSt5JahWF4XvZXiCgP+2aRYr1zqGLvb14Kw/AOjjd+\n8bBvflHsfXFdSX9IejYMw/uCILhMHG+84FvP8J44P/jWN5JWSBokqb6kQklXhWE43tdjjRdX3AZB\nUELS/bleh6R+kv4ThuHMeLwkDMO6kppL6hwEQfswDMdKWijp6hyuc6fnac9I0s1hGNa1fn1Gz/jD\n077hWOM5H/smCIJWkq5XrGdqSyov6e/0jT887ZvWkp6W9Jn9APrGHz72TRiGR299X6PYB5wfJD0X\nhuFgSdWDIDglpyuFl30j6V5JGyXVk9RMsfc3x3G88YenffOApDAMw0BSC0kXB0HQluONH3zsGd4T\n+8/HvpE0QNK8MAzrSDpT0gtBEJT39VjjyxW34yVVDIJgtqRLFPtpSRCvXROG4TtBENSU9LGkPvHH\nVJZ0fRiGLwdBsJ+k5yTtI6m0Yj/RvTUIgl0k9ZJ0Rnxbn0j6VxiGa4MgmCTpI0mnS7pYsf/UZ0k6\nMP7YUZL+K0lhGP4aBMG0+JreltRX0ntBEDwRhuHGTPyF4C/52DNFoWf84GPfcKzxn49901XS4DAM\nl8fzc6310jd+8LFvlks6UrEPOPtH1kvf+MHHvrFdKmlaGIZfxfO+ku6RNCZNf36kxse+eVXSnPiV\nbquDIPhKsRP/48Xxxhc+9k3D+DYVhuGqIAg+l9RA0v+J440PfOwZ3hP7z8e+OU5SK0kKw/DrIAi+\nUOwq7dHy8FjjxRW3ki6S9Ef8J/l3SZoeP/PdXrEz31Xij6sqaUsYhg0lXavYX6bi8QdhGNZT7GB/\nYBAE+0jqKOkfiv2Ut76kSpKus/bbTFL9MAynSOogaXIYhqskKQzDd8MwXCpJQRDUkXSYYpfdKwzD\nWYrdDn9kuv8ikDTveibu3CAIpgZBMDMIgluCICiQ6BmPeNc3HGvygnd9I6mxpN2DIPgwCIIwCIJ7\n4z/Npm/84V3fxK+etF+zDPrGG971zVZBEJSSdJOk3tZvj5dUJwiCg9Lxh0fKvOubMAwnhGH4gyQF\nQVBB0hGSPo3XON74wbu+kfSepI5BEJQMgmBfxa6inBivcbzJPR97hvfE/vOxbwollbAeu0ZSrXjs\n3bHGlxO3kqQgCHaTdIykhyQpDMO5kj6UdGL8ISUlDYvH0yRVj8fLJB0fv0x+QxiGncIwXBJ/3rNh\nGK4Nw/CP+HPbWbt8O/xz3klzSVMj6ykRBMFcSV9K6heG4TdW+VNJLXf0z4wd41nPvC/pZcVu6zle\nsfm29mwUesYTnvUNx5o84VnfVFLsp8TtJf1d0kmKXXGwFX3jCc/65q/QN57wtG86S/osDMPvtv5G\nGIabJX0h+sYLPvZN/IT/fyS9Hobhx1aJ440nPOubuxS7iGGFYlfIjQzjV/hzvPGHZz3De+I84Vnf\njJd0bfxzeCNJbSSVia/Lu2ONL6MStqooqUDSlCDYeuW0dpc0IR7/EYbh2q2x/jxD/lA8flzSvkEQ\nPKbYQb+apJXW9ldK2tPKf7HiPRW7lNqI/+PXCoKgmqTRQRD8EYbhoHh5WWRbyA1veiYMw2FW7Ycg\nCIYo9sLxXPz36Bl/eNM3EseaPOJT3/wm6cUwDFcrdgvqM4q9URkar9M3/vCpb/4KfeMPH/vmXElP\nbOP36Rt/eNU3QRDsrtjIhEWSLo+slb7xh099M0yxMWJ3S9pD0tggCDqGYTgiXqdv/OBTz/CeOH/4\n1DfdFHtPM0uxi6fGSvrVqnvVN76duF2m2D/QoWEYrrEL8ZkX2xQ/I95XUt/4rcbvSJos6SdJVayH\nVon/3rYURPZ3nqQ3wjD8NQzD5UEQvCTpBMW+eQ7+8KlnGig2y2tD/LdKStqU9J8E2eRT33CsyR/e\n9I2kBYq9+dnqj/gv+MenvkH+8Kpvgtg3MbdUbFYc/OVN3wRBUFLSa5JmhGF43bafAk940zeKnXDr\nGYZhoaRfgiB4V1JrSSP+55nIJZ96hvfE+cObvgnDcJn+nI2rIAgmSPo62T9ItvkyKmGTYmspK+kt\nxX8iGwRBuSAIng6C4G9FPTkIgsFBEBwXT+dJWqrYzIo3JXWJb6ekYkOJ30qwmWWKnbHfqqtiszQU\nBMGuit36/l+rXk2xL/lAbvjYM0MU+8mNgiDYQ7FRCfZz6Znc87FvONb4z8e+eVnSJUEQVAyCoKyk\nLop9ccdW9E3u+dg3f4W+yT1f++ZgScvjVzRF0Te552PfdJO0uoiTtvRN7vnYN6Gkk+PbL6vY7csz\nrDp9k1s+9gzvif3nXd8EQTAwCILr4vHRkvZT7GTwVl71jS8nbpco9pe0UNJ9kloHsW+cmybpuzA+\n2L4IgyT1jj9npmLfRveepJGKfTP7F4od8H+QNCDBNj5TbJ7OVl0lNY9vc5Zic3bus+qHx/eD3PCx\nZ86XdEIQBN9KmqLYPK8XrTo9k3s+9g3HGv951zdhGL4s6ZX4876SNEnSM9bj6Zvc865vgiDoFd/e\nVZLODIJgdhAEfazH0ze5513fxO2v2AclRxD7Aphmom9yzce+uUzx9zfWr15WneNN7vnYNxdI6hD/\nPPXf+DaelDjeeMK7nuE9cV7wrm8kDZTUOQiCeZL6SzozPr7Qy2NNQWFhYa7X4IX4Wf5pkg6IXra9\njccGih0Qalq3xWMnQ88gFfQNUkHfIBX0DVKxnX3TTrEv1WySlcXBWxxvkAqON9heHGuQinw/1vhy\nxW3Oxc/yj5Z0RRIPv1FSf/7z79zoGaSCvkEq6Bukgr5BKrazb3pK6vWXj0Kxx/EGqeB4g+3FsQap\nyPdjDSduXTdIOi8IgoMTPSB+9r2WpIeztir4jJ5BKugbpIK+QSroG6Qimb65RNLSMAxHZW9Z8BzH\nG6SC4w22F8capCJvjzWMSgAAAAAAAAAAz3DFLQAAAAAAAAB4pmRRxYKCAi7H9URhYWFBrteQLPrG\nH/QNUkHfIBX0DVJB3yAV+dI39Iw/8qVnJPrGJ/QNUkHfIBVF9Q1X3AIAAAAAAACAZzhxCwAAAAAA\nAACe4cQtAAAAAAAAAHiGE7cAAAAAAAAA4BlO3AIAAAAAAACAZzhxCwAAAAAAAACeKZnrBQAAgMTq\n1Klj4rFjxzq1EiVKmLhGjRpZWxMAAAAAIPO44hYAAAAAAAAAPMOJWwAAAAAAAADwDCduAQAAAAAA\nAMAzzLgFAMAjjz76qJOfffbZJq5cubJTe/PNN7OyJgAAACCXDjzwQBP36dPHqZ122mkmbtSokVOb\nPXt2ZhcGZBhX3AIAAAAAAACAZzhxCwAAAAAAAACeydtRCfXq1TPxSSed5NQuvfRSE0+dOtWpffnl\nlwm3+fDDDzv5xo0bd2SJAABs01577WXiV1991am1aNHCyQsLC008Y8YMp3bxxRdnYHUAAABAbh1x\nxBFOPnbsWBMvX77cqT322GMm/umnnzK7MCDLuOIWAAAAAAAAADzDiVsAAAAAAAAA8AwnbgEAAAAA\nAADAMwX27Lz/KRYUJC5m2WWXXebk/fv3N/Huu++eln20adPGySdOnJiW7aZDYWFhQa7XkCyf+mZn\nV1z7xv4/f/bZZzu133//3cTNmjVzauXLlzdx586dndqkSZNMvHjx4mSX4li6dKmTjxkzxsk///zz\nlLabbcW1b3KpTp06Tm6/hrVv396pFRS4f/033XSTiaM9xOtUanzuG/vf/8UXX3Rqdq/Ys/4ladGi\nRZldWIbQN0hFvvSNzz1z3nnnmbhdu3ZOrUmTJiYOgqDI7XzyyScmPvnkk53ab7/9tiNLTKt86RnJ\n777Jtt12283J7ffr++67r1P7+9//buL58+enZf/0TXadeOKJTj5y5EgnHzRokIlvvfVWp7Zu3brM\nLWw70TdIRVF9wxW3AAAAAAAAAOAZTtwCAAAAAAAAgGfyZlRC5cqVnXzWrFkm3nPPPdOyj19//dXJ\n7Vuw33333bTsI1Vcbo9UFNe+6devn4m7d++ekfWkw5YtW5x85syZJo7eAm3n6bq9K1XFtW9yqUWL\nFk4+efLkhI+Njkro0qWLiaN94xP6Jj3KlStn4jAMndp+++1n4ksvvdSpDR06NLMLyxD6BqnIl77J\ndc9UrVrVxNFjhD3WIPoZaMqUKQm3efTRRzu5fSv77NmznVp0pEsu5UvPSLnvm0yIjjWoVq1awseu\nXLnSxMccc4xTGzZsmImjr5HNmzc38erVq1NaZxR9k3m1atUy8VdffeXUPvzwQye3R0ZFP2f5hL5B\nKhiVAAAAAAAAAAB5hBO3AAAAAAAAAOAZTtwCAAAAAAAAgGdK5noByfrll1+c/M477zTxAw884NTs\n+XALFy50atWrV0+4j0qVKjn5CSecYOJcz7hF8VCjRg0Tly1b1ql16tTJxFdccUXCbbz11ltO3rVr\n1zStLn+cfvrpKT1vxYoVJv7vf/+b0jai87SCIDBx9BjStGlTJ2/QoIGJe/fu7dTs9eR6xi3So06d\nOib+z3/+49Sic2xt0f4eM2ZMehcGr61bt87Ec+bMcWr2jNui5gMCybrhhhucvFSpUiY++OCDnVrn\nzp0TbseebVq/fv00rQ47auzYsSauWbOmU7O/L+D+++93atHPXba6des6+WeffWZi+3VPku644w4T\n33333X+9YHjPfi/brVs3p2Z/zomK9kZRn8n79u1r4uicZPv90+LFi52affyCv8qUKePk9vztr7/+\n2ql17NjRyX2ea4vssr8Dy/5uKkm65ZZbnDw6Y9t22223mbhPnz5pWl36ccUtAAAAAAAAAHiGE7cA\nAAAAAAAA4JmCwsLCxMWCgsRFj0yfPt3JGzdubOIZM2Y4Nfv2jr9y0EEHmfi7775LcXXpUVhYmPi+\nWs/kS99kStu2bU0cveXZHodQsWJFp1bU/0Xbt99+6+TRWxkj2yyWfWP/34zeehX9+7HZtyAvWbJk\ne5aXlPLlyzt59Hafom4Le/LJJ0182WWXpXdh26m49k229erVy8Q333yzU3vnnXdMfPnllzu16K1/\n+YK+Sb8zzjjDyV955RUTv/DCC07t/PPPz8qa0o2+Sb/WrVs7uf3eN1o77bTTnLyoMS5FsW9fnTt3\nrlOL3uqcDvnSN9numeOOO87J7VEJI0aMcGr2e9IdYY9AsG85laQFCxaY+IADDkjL/lKVLz0j+X2s\nsccjPPTQQ0k/b8OGDU5uv561adPGqRV1W7N9jIq+7kVfF9OBvkm/6GiWq666ysS1a9d2aosWLcrK\nmtKNvkm/Fi1aOLl9/GnevLlTS/a8StTzzz/v5NkeSVlU33DFLQAAAAAAAAB4hhO3AAAAAAAAAOAZ\nTtwCAAAAAAAAgGdK5noB6XDPPfc4+a233mriJk2apLzdUqVKpfxcFG9Dhw41ccOGDZ3aYYcdltQ2\nVq9e7eTDhw838dSpU53aiy++aOLff/896XUWV/PmzdtmnGsnnXSSkxc10zY668uecYv8NGXKFCe3\nX3/mz5/v1K677joT5+tMW2TeZ599lrDWsWNHJ+/Zs6eTZ2KON7Jrn332cXL7vcCBBx6Y8HnRGfq7\n7babiaMzbL/44gsnP+SQQ7Z7nZK0yy5/Xgti7w/ZVbKk+9HOnjf80ksvZWSfI0eONHF0xm2ZMmVM\nXKFCBae2atWqjKwH6XXXXXc5eY8ePRI+9tlnnzXx8uXLnVr//v2d3K5HP6+PGzfOxFWrVk34PLv3\n4LfSpUubuEuXLk5t0qRJJs7XmbbIDPv/f/Szsv09P9HjzejRo518zJgxJo7Oxj7rrLNMHJ2ja58P\n3LhxY7LLzgiuuAUAAAAAAAAAz3DiFgAAAAAAAAA8UyxGJURvk5g8ebKJ3333XacWva29KPYIhjPP\nPDPF1SEfValSxcn79Onj5BdddJGJf/nlF6dm33bYt29fpzZjxgwTr1+/3qktXLgwtcUiq6IjVAYM\nGGDi6K0XRWnZsqWTT58+fccWhpw45ZRTTHz44Yc7tcLCQhO/8sorTo2RJ0iFfZt79FjUoUMHJx88\neHBW1oT0atu2rYmjtwX+7W9/2+Ht16tXz8l//vlnJ7dvS9x3332d2rBhw0y8//77J9zHzJkzd2SJ\n2AETJ0508qZNm5p43bp1GdlndPSTba+99jLxueee69QGDRqUkfUgvaKjT8qWLWviBQsWODV7XOFf\njeupVauWiW+55RanVq1aNROvXbvWqdmjG3gvlT9uvPFGE+++++5Oze4bwGaPOLBHI0jueb727dsn\nvc05c+Y4uf2+K/rext7nV199lfQ+MoErbgEAAAAAAADAM5y4BQAAAAAAAADPcOIWAAAAAAAAADxT\nLGbcdu7c2ckbN25s4gYNGqS8XXtWLnYut99+u5NffPHFTv7oo4+aODqXZ82aNZlbGHLimGOOMfF5\n553n1C688MKEz9u0aZOTd+vWzcSzZ89Oz+KQVZUqVXLyI488MqnnrVy50skXLVqU0v6vueYaJy9q\n5mX37t1T2gf8Zc9NjorOvEV+sucAbs9MW3vOaM+ePZ3aJ598YuIwDIvczooVK0wcPd4UNdd2/vz5\nJo6+TiJ7cjHz87vvvjPxN99849Tq169v4tq1a2dtTUif6HfJnHDCCSaOzsy2v9vjyiuvdGoVK1Z0\n8gcffNDEJ554olOzvz+kd+/eTu2JJ55IZtnwTLt27Uz80UcfObVp06ZleznIE9HvBLLZ82/TZdWq\nVU4e/R6AXOKKWwAAAAAAAADwDCduAQAAAAAAAMAzeTMqoW7duk7+2muvmbhWrVpOrWTJ9PyxXn/9\n9bRsB/4oV66ciaO3Etq39l177bVObeLEiU4+btw4E+fitjRkVvPmzZ383XffNXGJEiWS3k70tuaF\nCxea+I8//khxdcil6L9bs2bNTLzLLu7PQrds2WLiDz74IOl9XHfddQlrV199tZPXqFEj4WNvuOEG\nE0dvcV68eHHS6wGQOfbto5LUokWLpJ5nv55I7nuY6G2oqSpqNEKUfcuiT7cWIvPssVCbN2/O4UqQ\nCdOnT3dye/RKdFRCmzZtTHzcccc5tYceesjJq1evnnCf//73v01sj6dD/mjVqpWT269tDRs2THm7\nRx99tImXL1/u1KKjWpD/CgoKthlL7hi6MmXKOLWDDjrIye3RhvZnN0launSpiTt16uTUfPq8xBW3\nAAAAAAAAAOAZTtwCAAAAAAAAgGc4cQsAAAAAAAAAnsmbGbcHH3ywkx9wwAEmTtdM2yh7zmB0riDy\n02233Wbi6IzbESNGmNieaSoxx3Zn07FjRyffnrm2tlKlSjn5W2+9ZeLPP//cqb3xxhsmtmd4S9KM\nGTNS2j/Sr3Xr1k5+5JFHmtieaSu5MyiLmvnYpEmThNuUpA4dOiR87tq1a028aNEipxYEgYlHjhzp\n1M455xwTL1iwIOH2AWSWPYtacmfxR02ZMsXE9gxIKfW5tnvssYeTn3DCCSY+6qijklqLJL399tsp\n7R/5r3Tp0iaOzhm0rV69OhvLQZpt2LDByVetWpXwsfvuu6+JR40a5dSi8ynt74F46qmnnNro0aO3\ne53wS5cuXZx81qxZJv7+++8TPs+eRSpJDzzwgJPbr1nR3uzevbuJH3vssaTXCn/Vr1/fxNHvjrn+\n+utNHH0vFZ1ja7M/A0n/+xnJV1xxCwAAAAAAAACe4cQtAAAAAAAAAHgmb0YlRG8dvvHGG0183333\nObWibtPZHvvss09atgN/3HzzzSaOXm7/4osvmpjRCDu3V1991cntUS2HHXaYU6tatWpK+zj00EMT\n5nfeeadTe/jhh03cr18/p7Zs2bKU9o/klS9f3sT2mJ6oH3/80cmff/55E8+dO9ep1alTx8Q9evRw\naqeccoqT22MWomNc7FvIKlas6NQmTJiQsIb8ZN9qGn0NQ34aMmSIk9uvKb/99ptTO/fcc028dOnS\ntOz/8ssvd/JevXolfOw333xj4uhIoXStB/mnZs2aJrZH9ESNHTs26W3a/w8aN27s1Fq2bGniV155\nxamFYZj0PpCadI1Xsser9O/f36n98MMPadkHcueiiy5ycvv1KzriwB4tF/0MdNlllzn5uHHjTNy+\nfXunNmzYMBPPmzfPqW3P8Qf+WLFihYntz2OS+9m5qFEskrRu3ToTz5w5M51LzBquuAUAAAAAAAAA\nz3DiFgAAAAAAAAA8w4lbAAAAAAAAAPBM3sy4jRowYICJ58yZ49QqVaqU8HklS7p/5IEDB5q4QoUK\naVodfPXZZ5+ZODpj1O6F9evXO7Xx48dndmHwypQpU5z8xBNPNHH16tWdmj2Hba+99nJqp59+upPb\n856is3hsu+zi/kzt+uuvN3GzZs2c2rHHHmviLVu2JNwmUteqVSsTP/TQQwkf9+STTzr53XffbeJo\nb9jz3KIzulavXu3kI0aMMHH37t2dWu3atU08aNCghNt57733nFq6ZtQhu5hrW/yMGjWqyDzdTj75\nZCe/4447Ej528+bNTm4fY5hpu/MoXbq0k++///5OfsQRRyS1nehr1BdffGHiQw45xKlVrlzZxH/7\n29+cmv3aVqtWLad24YUXJrUWJK9EiRJOfuSRR5q4qPeyUW+99ZaTR49FyH/169c3cfScS/T1xGb/\n/4/Ooh05cmTC57388stObr9ft7/XZlvbRX6we6pFixZOzX4tivZClP39Ncy4BQAAAAAAAACkBSdu\nAQAAAAAAAMAznLgFAAAAAAAAAM8UFDUvraCgoNgNU4vO4rnrrrtMHJ3zNW/ePBPbcySl7M8HLCws\nTH6IUI5lu28OP/xwJ//yyy9NvHHjRqdmz8zq1q2bU7v99ttNvGbNmiL3MXv27NQWm2X0jV86d+5s\n4quvvtqpNW/ePKVt3nTTTSbu169faguLoG9cPXv2NHHv3r0TPi46z8v20UcfOXn0mGKLvt68//77\nJo7Od5o8eXLC7Tz88Jdzz/QAAAxrSURBVMMmjs7GzQT6Jv2isx2Leu9xzDHHOLndNz6jb7Lrjz/+\ncPKiPgdceeWVTj5kyJCMrCkV+dI32e6ZsmXLOvmee+5p4ugcWfv1pE2bNgm3WaZMGSe3Zw5uj2jv\nLVq0KOFjn3nmGRNHZ6P+/PPPJp4/f37S+8+XnpH8Ota88sorTh79/oZkRf8dO3TokPKasom+SZ79\n/jX6/TD16tUzcfRzdPny5U1cqlQpp7ZixYqk92/v4+uvv3Zq0VnNmUbfZF6DBg1M/NVXXzm16Hsb\nuze+/fbbzC5sBxTVN1xxCwAAAAAAAACe4cQtAAAAAAAAAHgm8X2dxVT08vvoeATbpk2bTBy9vQfZ\ntc8++zj5m2++aeLq1as7teuuu87EL7zwglP75ZdfTDxw4ECnZo9K2H333Z2aPWIBSNXw4cNN/PLL\nLzu1//u//zPxUUcdlfQ2a9WqteMLQ5EqVapk4ui4nTFjxiR8XpMmTUxcs2ZNp2Zv54YbbnBq0Vvc\n69SpY+L//Oc/SW/HHpWA4s8e7wTY7r33XhPvsot7zcaWLVsSPi9fxm3sbKLjEOyxbyeffLJTq1u3\nbkr7WLVqlYlXr17t1DZv3uzkRY0JGjp0qIkHDRrk1KZNm5bS2pB+++67r5N37drVxGeccYZTs29B\njv4b2rcr29uQ3LEd2PksXrw4YS16jElVUeNXUPw0bNjQxNvz3iZfccUtAAAAAAAAAHiGE7cAAAAA\nAAAA4BlO3AIAAAAAAACAZ3a6Gbf33HNP0o996qmnTMzMlNyKzlCqUKGCiXv27OnUonNtE7nmmmsS\n1ux5o5I0Y8aMpLYJJCs6I+6LL74w8fbMuP3222/Ttib8NXu227byRKKzluznNWrUyKktXLjQycuU\nKWPi77//3qkdeeSRJv7tt9+SWguA4i36fQ5NmzY1cVHHIsl9bzRnzpwMrA47avTo0U5+3HHHmXjD\nhg1O7a233jJx9PXDntEefd78+fNNHP0MNHv2bCe357B/9913Tu3666838Zo1awQ/HXvssU5+9913\nJ3zsbbfdZuLo94WceuqpJo7OuJ05c+aOLBF5wP7eheh3QmRD69atTZyuubnw1/r1600cfW8zadIk\nJ9+4cWM2lpRRXHELAAAAAAAAAJ7hxC0AAAAAAAAAeCbroxKqVKni5MOGDTPxiy++6NSieSr22Wcf\nJ7/00kuTfu6rr766w/tHegwYMMDJ7dt0orVobrNv+6tdu7ZTW7BggYlvvvlmp7Zq1arkF4u8ED02\nXHLJJSaO3gY4YsSItO+/RIkSTt64ceOknhcdsfDJJ5+kbU3YNvt20h49eji1U045xcQtWrRwak2a\nNDFx+fLlE27//PPPd/Lo7WU///yzie+66y6ntnjx4oTbxc6ldOnSuV4CcqhcuXIm7tKli1Ozb6WP\nir7XHj58uImjtx7CD+3atXNyewTC6aef7tSmT5+e0j5KlvzzI+J9993n1Pbbbz8nX7ZsmYk7duzo\n1BiP4K+jjz7axEV9durQoYOT2+Pk9t57b6d2xx13JNyOPX4DxZM9eifZUWI7Ytddd3Xyyy+/3MTP\nP/98xveP7Kpbt66TX3zxxSZevny5U3viiSecvDgcf7jiFgAAAAAAAAA8w4lbAAAAAAAAAPAMJ24B\nAAAAAAAAwDNZn3EbnaFz8sknm7hOnTpO7ccffzRxdI7f3LlzTdysWTOnZm/nxhtvdGoVKlRIuLYH\nHngg4f6RW3369HHyTZs2mbhp06ZOrW3btgm3s8cee5j4rbfecmrdu3c3sd1fKD7sWVxjx451ag0b\nNjSx3SfptNdee5n4+uuvd2pt2rRJahuzZs1y8smTJ+/4wlAk+3izbt06p2bPlfzoo4+cWqrzvVav\nXu3k9ozld955J6Vtovhr3769kz/66KM5WgmyITo3+8knnzTxmWeemfB51113nZMPHDjQyZlr67/o\na8uvv/5q4hkzZqS0zTJlyjj5K6+8YuITTzzRqW3YsMHJzznnHBNPmzYtpf0j++zZ1xUrVnRq77//\nvonffPNNp2bPFT3ppJOcmr2d6Lz+6AxKFD8zZ8408ZIlS5yaPXs9On90e9j9F91OzZo1TXzBBRek\nvA/4wz6mjBs3zqnZ89Z79uzp1EaOHJnZheUAV9wCAAAAAAAAgGc4cQsAAAAAAAAAnsn6qITorXsH\nHHCAiVu2bOnUJk2aZOL58+c7NftS/COPPNKpRW8fs0VvL5o9e7aJ77zzTqf2+++/J9wOcqt///65\nXgLy0MMPP2xiezRClH1ckqQwDE28fv36hM8rW7ask0dHtdjjEYo6TkVvL7Nvne/WrVvC5yEzvvji\nCxN36tTJqdn/pkcffXTS23z22WdN/PXXXzu1L7/80sntWxaxc/npp5+c/JtvvjFx/fr1s70ceMS+\nRVAqejzCvHnzTBwdWYb88+233zp5kyZNTDxkyBCnVqVKFRN/9dVXTu27774zcY8ePZxaEAQm/vTT\nT53aFVdc4eTTp09PZtnwjD0WJfr52M7tW9Ml6dRTTzXxI4884tRWrlxp4qFDhzq1Hbk9HvnBHo9w\n7733OrXoSErb8OHDTXzggQc6tcaNGzv5LbfcYuLouZp27dqZ+Oeff05ixfBdv379TBx93/Piiy+a\nuKj+Ki644hYAAAAAAAAAPMOJWwAAAAAAAADwDCduAQAAAAAAAMAzBdGZNk6xoCBxMU3seRRz5851\nao8//nja9/fLL784uT37yWeFhYUFf/0oP2Sjb5Ac+sZ1ySWXmHjw4MFJP8+eOfrbb78lfFzFihWd\nvGnTptuxuj+tWbPGyU877TQTv/feeyltc3vQN0gFfZN5U6dONXGzZs2c2ptvvunkHTp0yMqadhR9\nk7y6deua+IYbbnBqXbt2NXF0Buo//vEPEy9YsCBDq8uufOmbbPRMr169TNy9e3entssuyV2j8/rr\nrzv5U089ZeKxY8fuwOr8kS89I2Wnb+z3wf/85z+d2siRI0281157ObXod8vY7Pm3b7zxxo4u0Qv0\nTXr861//MvH999/v1EqXLp3wefb3fEjunPZ77rnHqW3cuHFHlphW9E1q2rZt6+RjxowxsT2XW5K6\ndOmyzcfls6L6hituAQAAAAAAAMAznLgFAAAAAAAAAM/kfFSCLXqZfLdu3RI+1r4FuVOnTgkfF72t\nuU2bNk4+bdq07VliznC5PVJB37hq1qxp4nvvvdepnXPOOZnevWPz5s1O/vDDD5t41KhRTu3TTz/N\nypq2om+QCvom85588kkTX3zxxU5t0qRJTh59v+Mr+iZ5w4cPN/HZZ5+d8HFXX321kz/xxBMZW1Ou\n5Evf5Lpn8Kd86RkpO31z7bXXmtgeXbiNtTi5PXbwsccec2p9+/Y18fr163d0iV6gb5AK+iZ59ufz\nL774wqmVKVPGxPZoBEl67bXXMrquXGBUAgAAAAAAAADkEU7cAgAAAAAAAIBnOHELAAAAAAAAAJ4p\nmesF2DZs2ODk999/f1LPO/fcczOxHADFzPz5803ctWtXp/b666+bODob8ttvvzVxhw4dEm5/9uzZ\nRe5/woQJCR87ffr0Ip8LAL179zZxgwYNnNqIESOyvRxkWP369Z28QoUKCR87ZMgQE9uvNQCwLc8+\n+6yJS5Uq5dRuv/12E3/++edOzX6//NBDD2VodQCKq7Jlyzr5DTfcYOKKFSs6Nft7X4rjTNvtwRW3\nAAAAAAAAAOAZTtwCAPD/7dwxioNAFIBhvVQuYKq0nsKjpE8XyA1SWOYCuYOl4D3cTnYWDMuDZN8s\n31eNTDPFj8VDBwAAAJJp13Xd32zb/U0+al3X9q/P8Fu6yUM3ROiGCN0QoZt95/O5eP7+O+E8z8Xe\n6XTa1tM0vfdgCdTSjXdNHrU00zS6yUQ3ROhm3zAMxfPlctnWz+ez2Dsej9v657Wq/9GrbnxxCwAA\nAACQjMEtAAAAAEAyBrcAAAAAAMm447YS7kkhQjdE6IYI3RChm31d1xXPj8djW/d9X+yN4/iRM2VR\nSzfeNXnU0kzT6CYT3RChm9LhcNjW9/u92Lvdbtv6er0We8uyvPdgybjjFgAAAACgIga3AAAAAADJ\nuCqhEj63J0I3ROiGCN0QoRsiaulGM3nU0kzT6CYT3RChGyJclQAAAAAAUBGDWwAAAACAZAxuAQAA\nAACSeXnHLQAAAAAAn+eLWwAAAACAZAxuAQAAAACSMbgFAAAAAEjG4BYAAAAAIBmDWwAAAACAZAxu\nAQAAAACS+QIWA5hHw/w4vQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f449c803278>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "2M3zflCxoxpa",
        "colab_type": "code",
        "outputId": "516f7e5f-0bc4-4903-c390-f4bfd0f3f325",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 700
        }
      },
      "cell_type": "code",
      "source": [
        "img = np.squeeze(images[0])\n",
        "\n",
        "fig = plt.figure(figsize = (12,12)) \n",
        "ax = fig.add_subplot(111)\n",
        "ax.imshow(img, cmap='gray')\n",
        "width, height = img.shape\n",
        "thresh = img.max()/2.5\n",
        "for x in range(width):\n",
        "    for y in range(height):\n",
        "        val = round(img[x][y],2) if img[x][y] !=0 else 0\n",
        "        ax.annotate(str(val), xy=(y,x),\n",
        "                    horizontalalignment='center',\n",
        "                    verticalalignment='center',\n",
        "                    color='white' if img[x][y]<thresh else 'black')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqwAAAKrCAYAAAA0x9+IAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nEEIIYWj/HyyJ9cftDnEJAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f443746f940>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "IF9hePDooxqG",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Sequential API"
      ]
    },
    {
      "metadata": {
        "id": "sJ_YL3H6oxqJ",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# [0-9] unique labels\n",
        "num_classes = 10\n",
        "epochs = 5"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "CpksX4ZvoxqS",
        "colab_type": "code",
        "outputId": "7117bbba-76e5-4c97-ff78-58a087142e51",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        }
      },
      "cell_type": "code",
      "source": [
        "class Flatten(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Flatten, self).__init__()\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = x.view(x.size(0), -1)\n",
        "        return x\n",
        "\n",
        "## Define the NN architecture\n",
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Net, self).__init__()\n",
        "        D_in = 784\n",
        "        H = 16\n",
        "        D_out = num_classes\n",
        "        self.classifier = torch.nn.Sequential(\n",
        "            nn.Conv2d(1, 32, 3, padding=0),\n",
        "            nn.ReLU(),\n",
        "            nn.Conv2d(32, 64, 3, padding=0),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(2, 2),\n",
        "            nn.Dropout(p=0.25),\n",
        "            Flatten(),\n",
        "            nn.Linear(12*12*64, 128),\n",
        "            nn.Dropout(p=0.5),\n",
        "            nn.Linear(128, num_classes),\n",
        "            nn.LogSoftmax(dim=1)\n",
        "        )\n",
        "        \n",
        "    def forward(self, x):\n",
        "        \n",
        "        # 2 conv layer, with relu activation function\n",
        "        x = self.classifier(x)\n",
        "        return x\n",
        "\n",
        "# initialize the NN\n",
        "model = Net()\n",
        "model = model.to(device)\n",
        "print(model)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Net(\n",
            "  (classifier): Sequential(\n",
            "    (0): Conv2d(1, 32, kernel_size=(3, 3), stride=(1, 1))\n",
            "    (1): ReLU()\n",
            "    (2): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1))\n",
            "    (3): ReLU()\n",
            "    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "    (5): Dropout(p=0.25)\n",
            "    (6): Flatten()\n",
            "    (7): Linear(in_features=9216, out_features=128, bias=True)\n",
            "    (8): Dropout(p=0.5)\n",
            "    (9): Linear(in_features=128, out_features=10, bias=True)\n",
            "    (10): LogSoftmax()\n",
            "  )\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "HMS2s2oVoxqc",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# specify loss function\n",
        "criterion = nn.NLLLoss()\n",
        "\n",
        "# specify optimizer\n",
        "optimizer = optim.Adam(model.parameters(), lr=1e-3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "CEyDMgWDoxqk",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_model(model, dataloaders, criterion, optimizer, num_epochs):\n",
        "    since = time.time()\n",
        "\n",
        "    history = dict()\n",
        "\n",
        "    best_model_wts = copy.deepcopy(model.state_dict())\n",
        "    best_acc = 0.0\n",
        "\n",
        "    for epoch in range(num_epochs):\n",
        "        print('Epoch {}/{}'.format(epoch, num_epochs - 1))\n",
        "        print('-' * 10)\n",
        "\n",
        "        # Each epoch has a training and validation phase\n",
        "        for phase in ['train', 'val']:\n",
        "            if phase == 'train':\n",
        "                model.train()  # Set model to training mode\n",
        "            else:\n",
        "                model.eval()   # Set model to evaluate mode\n",
        "\n",
        "            running_loss = 0.0\n",
        "            running_corrects = 0\n",
        "\n",
        "            # Iterate over data.\n",
        "            for inputs, labels in dataloaders[phase]:\n",
        "                inputs = inputs.to(device)\n",
        "                labels = labels.to(device)\n",
        "\n",
        "                # zero the parameter gradients\n",
        "                optimizer.zero_grad()\n",
        "\n",
        "                # forward\n",
        "                # track history if only in train\n",
        "                with torch.set_grad_enabled(phase == 'train'):\n",
        "                    # Get model outputs and calculate loss\n",
        "\n",
        "                    # backward + optimize only if in training phase\n",
        "                    if phase == 'train':\n",
        "                        outputs = model(inputs)\n",
        "                        loss = criterion(outputs, labels)\n",
        "                        loss.backward()\n",
        "                        optimizer.step()\n",
        "                        \n",
        "                    else:\n",
        "                        outputs = model(inputs)\n",
        "                        loss = criterion(outputs, labels)\n",
        "                    \n",
        "                    _, preds = torch.max(outputs, 1)\n",
        "\n",
        "                # statistics\n",
        "                running_loss += loss.item() * inputs.size(0)\n",
        "                running_corrects += torch.sum(preds == labels.data)\n",
        "\n",
        "            epoch_loss = running_loss / (len(dataloaders[phase])*batch_size)\n",
        "            epoch_acc = running_corrects.double() / (len(dataloaders[phase])*batch_size)\n",
        "\n",
        "            print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))\n",
        "            print (len(dataloaders[phase].dataset))\n",
        "\n",
        "            # deep copy the model\n",
        "            if phase == 'val' and epoch_acc > best_acc:\n",
        "                best_acc = epoch_acc\n",
        "                best_model_wts = copy.deepcopy(model.state_dict())\n",
        "            \n",
        "            if phase+'_acc' in history:\n",
        "                # append the new number to the existing array at this slot\n",
        "                history[phase+'_acc'].append(epoch_acc)\n",
        "            else:\n",
        "                # create a new array in this slot\n",
        "                history[phase+'_acc'] = [epoch_acc]\n",
        "            \n",
        "            if phase+'_loss' in history:\n",
        "                # append the new number to the existing array at this slot\n",
        "                history[phase+'_loss'].append(epoch_loss)\n",
        "            else:\n",
        "                # create a new array in this slot\n",
        "                history[phase+'_loss'] = [epoch_loss]            \n",
        "\n",
        "    time_elapsed = time.time() - since\n",
        "    print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))\n",
        "    print('Best val Acc: {:4f}'.format(best_acc))\n",
        "\n",
        "    # load best model weights\n",
        "    model.load_state_dict(best_model_wts)\n",
        "    return model, history"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "4jxYPzqUoxqw",
        "colab_type": "code",
        "outputId": "cbd31d24-c8ee-4db5-c2ee-6d039f24c588",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 605
        }
      },
      "cell_type": "code",
      "source": [
        "model, history = train_model(model, dataloaders_dict, criterion, optimizer, epochs)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 0/4\n",
            "----------\n",
            "train Loss: 0.2784 Acc: 0.9138\n",
            "60000\n",
            "val Loss: 0.0693 Acc: 0.9517\n",
            "10000\n",
            "Epoch 1/4\n",
            "----------\n",
            "train Loss: 0.0877 Acc: 0.9710\n",
            "60000\n",
            "val Loss: 0.0472 Acc: 0.9609\n",
            "10000\n",
            "Epoch 2/4\n",
            "----------\n",
            "train Loss: 0.0665 Acc: 0.9777\n",
            "60000\n",
            "val Loss: 0.0423 Acc: 0.9624\n",
            "10000\n",
            "Epoch 3/4\n",
            "----------\n",
            "train Loss: 0.0560 Acc: 0.9801\n",
            "60000\n",
            "val Loss: 0.0396 Acc: 0.9634\n",
            "10000\n",
            "Epoch 4/4\n",
            "----------\n",
            "train Loss: 0.0485 Acc: 0.9818\n",
            "60000\n",
            "val Loss: 0.0356 Acc: 0.9634\n",
            "10000\n",
            "Training complete in 1m 9s\n",
            "Best val Acc: 0.963379\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "yB8CJAS8oxq5",
        "colab_type": "code",
        "outputId": "126a5aec-e34e-4901-990f-f7376c7b1da6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        }
      },
      "cell_type": "code",
      "source": [
        "# summarize history for accuracy\n",
        "plt.plot(history['train_acc'])\n",
        "plt.plot(history['val_acc'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()\n",
        "# summarize history for loss\n",
        "plt.plot(history['train_loss'])\n",
        "plt.plot(history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEVCAYAAAAPRfkLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f4434aea748>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEVCAYAAAAPRfkLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f443751d080>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "a4wNdBENoxrE",
        "colab_type": "code",
        "outputId": "885269d5-a6ed-4edc-b2ca-b29c02d0eea4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 274
        }
      },
      "cell_type": "code",
      "source": [
        "# initialize lists to monitor test loss and accuracy\n",
        "test_loss = 0.0\n",
        "class_correct = list(0. for i in range(10))\n",
        "class_total = list(0. for i in range(10))\n",
        "\n",
        "model.eval() # prep model for *evaluation*\n",
        "\n",
        "for data, target in test_loader:\n",
        "    # forward pass: compute predicted outputs by passing inputs to the model\n",
        "    data = data.to(device)\n",
        "    target = target.to(device)\n",
        "    \n",
        "    output = model(data)\n",
        "    # calculate the loss\n",
        "    loss = criterion(output, target)\n",
        "    # update test loss \n",
        "    test_loss += loss.item()*data.size(0)\n",
        "    # convert output probabilities to predicted class\n",
        "    _, pred = torch.max(output, 1)\n",
        "    # compare predictions to true label\n",
        "    correct = np.squeeze(pred.eq(target.data.view_as(pred)))\n",
        "    # calculate test accuracy for each object class\n",
        "    for i in range(len(data)):\n",
        "        label = target.data[i]\n",
        "        class_correct[label] += correct[i].item()\n",
        "        class_total[label] += 1\n",
        "\n",
        "# calculate and print avg test loss\n",
        "test_loss = test_loss/len(test_loader.dataset)\n",
        "print('Test Loss: {:.6f}\\n'.format(test_loss))\n",
        "\n",
        "for i in range(10):\n",
        "    if class_total[i] > 0:\n",
        "        print('Test Accuracy of %5s: %2d%% (%2d/%2d)' % (\n",
        "            str(i), 100 * class_correct[i] / class_total[i],\n",
        "            np.sum(class_correct[i]), np.sum(class_total[i])))\n",
        "    else:\n",
        "        print('Test Accuracy of %5s: N/A (no training examples)' % (classes[i]))\n",
        "\n",
        "print('\\nTest Accuracy (Overall): %2d%% (%2d/%2d)' % (\n",
        "    100. * np.sum(class_correct) / np.sum(class_total),\n",
        "    np.sum(class_correct), np.sum(class_total)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test Loss: 0.034638\n",
            "\n",
            "Test Accuracy of     0: 99% (782/785)\n",
            "Test Accuracy of     1: 99% (881/883)\n",
            "Test Accuracy of     2: 99% (827/835)\n",
            "Test Accuracy of     3: 97% (800/824)\n",
            "Test Accuracy of     4: 99% (787/793)\n",
            "Test Accuracy of     5: 99% (722/727)\n",
            "Test Accuracy of     6: 97% (724/740)\n",
            "Test Accuracy of     7: 97% (802/820)\n",
            "Test Accuracy of     8: 97% (768/785)\n",
            "Test Accuracy of     9: 97% (789/808)\n",
            "\n",
            "Test Accuracy (Overall): 98% (7882/8000)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "KAmbST80oxrU",
        "colab_type": "code",
        "outputId": "9759fb72-28c3-4bdd-ad81-50aefc6490e5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "cell_type": "code",
      "source": [
        "# obtain one batch of test images\n",
        "dataiter = iter(test_loader)\n",
        "images, labels = dataiter.next()\n",
        "\n",
        "images = images.to(device)\n",
        "labels = labels.to(device)\n",
        "\n",
        "# get sample outputs\n",
        "output = model(images)\n",
        "# convert output probabilities to predicted class\n",
        "_, preds = torch.max(output, 1)\n",
        "# prep images for display\n",
        "images = images.cpu().numpy()\n",
        "\n",
        "# plot the images in the batch, along with predicted and true labels\n",
        "fig = plt.figure(figsize=(25, 4))\n",
        "for idx in np.arange(20):\n",
        "    ax = fig.add_subplot(2, 20/2, idx+1, xticks=[], yticks=[])\n",
        "    ax.imshow(np.squeeze(images[idx]), cmap='gray')\n",
        "    ax.set_title(\"{} ({})\".format(str(preds[idx].item()), str(labels[idx].item())),\n",
        "                 color=(\"green\" if preds[idx]==labels[idx] else \"red\"))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABW4AAAD2CAYAAABGOR6XAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCJj/B6UR2HVRWnjWAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f4434bf2160>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Jf99fnR_qBw_",
        "colab_type": "code",
        "outputId": "824fb1ba-60a0-4c51-f9fd-851420b4b616",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        }
      },
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "random_idx = random.randint(0, len(test_loader.dataset))\n",
        "# get sample outputs\n",
        "image = test_loader.dataset[random_idx][0].to(device)\n",
        "\n",
        "output = model(image[None, :])\n",
        "# convert output probabilities to predicted class\n",
        "_, preds = torch.max(output, 1)\n",
        "label = test_loader.dataset[random_idx][1]\n",
        "label = label.to(device)\n",
        "\n",
        "# plot the images one example, along with predicted and true labels and probabilities\n",
        "fig = plt.figure(figsize=(25, 4))\n",
        "ax = fig.add_subplot(1, 1, 1, xticks=[], yticks=[])\n",
        "ax.imshow(test_loader.dataset[random_idx][0].detach().squeeze(), cmap='gray')\n",
        "ax.set_title(\"{} ({})\".format(str(preds.cpu().numpy()[0]), str(label.cpu().numpy())),\n",
        "             color=(\"green\" if preds==label else \"red\"))\n",
        "ax = fig.add_subplot(1, 2, 2, xticks=[], yticks=[])\n",
        "x_labels = torch.arange(10)\n",
        "ax.bar(torch.arange(10), height=torch.exp(output).cpu().detach().squeeze(), tick_label=x_labels)\n",
        "ax.set_title(\"{:.2f} % accuracy\".format(float(torch.exp(output[:, preds])*100)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5,1,'100.00 % accuracy')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzQAAAEHCAYAAAB4AwNdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/x8wfKWQ9r1ScgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f44319abe80>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "Jg9QDRIpoxrs",
        "colab_type": "text"
      },
      "cell_type": "markdown",
      "source": [
        "## Functional API"
      ]
    },
    {
      "metadata": {
        "id": "DID5C_ppoxru",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# [0-9] unique labels\n",
        "num_classes = 10\n",
        "epochs = 5"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "qeFncs6Eoxr9",
        "colab_type": "code",
        "outputId": "100debef-bccb-4a2d-ce67-d8725ebb4efa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 182
        }
      },
      "cell_type": "code",
      "source": [
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "\n",
        "# define the CNN architecture\n",
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Net, self).__init__()\n",
        "        # convolutional layer\n",
        "        self.conv1 = nn.Conv2d(1, 32, 3, padding=0)\n",
        "        self.conv2 = nn.Conv2d(32, 64, 3, padding=0)\n",
        "        # max pooling layer\n",
        "        self.pool = nn.MaxPool2d(2, 2)\n",
        "        self.dropout1 = nn.Dropout(p=0.25)\n",
        "        self.dropout2 = nn.Dropout(p=0.5)\n",
        "        self.fc1 = nn.Linear(12 * 12 * 64, 128)\n",
        "        self.output = nn.Linear(128, num_classes)\n",
        "\n",
        "    def forward(self, x):\n",
        "        # add sequence of convolutional and max pooling layers\n",
        "        x = F.relu(self.conv1(x))\n",
        "        x = F.relu(self.conv2(x))\n",
        "        x = self.pool(x)\n",
        "        x = self.dropout1(x)\n",
        "        x = x.view(-1, 12 * 12 * 64)\n",
        "        x = F.relu(self.fc1(x))\n",
        "        x = self.dropout2(x)\n",
        "        x = F.log_softmax(self.output(x), dim=1)\n",
        "        return x\n",
        "\n",
        "# create a complete CNN\n",
        "model = Net()\n",
        "model = model.to(device)\n",
        "print(model)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Net(\n",
            "  (conv1): Conv2d(1, 32, kernel_size=(3, 3), stride=(1, 1))\n",
            "  (conv2): Conv2d(32, 64, kernel_size=(3, 3), stride=(1, 1))\n",
            "  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
            "  (dropout1): Dropout(p=0.25)\n",
            "  (dropout2): Dropout(p=0.5)\n",
            "  (fc1): Linear(in_features=9216, out_features=128, bias=True)\n",
            "  (output): Linear(in_features=128, out_features=10, bias=True)\n",
            ")\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "5ttwozDLoxsd",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# specify loss function\n",
        "criterion = nn.NLLLoss()\n",
        "\n",
        "# specify optimizer\n",
        "optimizer = optim.Adam(model.parameters(), lr=1e-3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "Xr2OljTqoxsl",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "def train_model(model, dataloaders, criterion, optimizer, num_epochs):\n",
        "    since = time.time()\n",
        "\n",
        "    history = dict()\n",
        "\n",
        "    best_model_wts = copy.deepcopy(model.state_dict())\n",
        "    best_acc = 0.0\n",
        "\n",
        "    for epoch in range(num_epochs):\n",
        "        print('Epoch {}/{}'.format(epoch, num_epochs - 1))\n",
        "        print('-' * 10)\n",
        "\n",
        "        # Each epoch has a training and validation phase\n",
        "        for phase in ['train', 'val']:\n",
        "            if phase == 'train':\n",
        "                model.train()  # Set model to training mode\n",
        "            else:\n",
        "                model.eval()   # Set model to evaluate mode\n",
        "\n",
        "            running_loss = 0.0\n",
        "            running_corrects = 0\n",
        "\n",
        "            # Iterate over data.\n",
        "            for inputs, labels in dataloaders[phase]:\n",
        "                inputs = inputs.to(device)\n",
        "                labels = labels.to(device)\n",
        "\n",
        "                # zero the parameter gradients\n",
        "                optimizer.zero_grad()\n",
        "\n",
        "                # forward\n",
        "                # track history if only in train\n",
        "                with torch.set_grad_enabled(phase == 'train'):\n",
        "                    # Get model outputs and calculate loss\n",
        "\n",
        "                    # backward + optimize only if in training phase\n",
        "                    if phase == 'train':\n",
        "                        outputs = model(inputs)\n",
        "                        loss = criterion(outputs, labels)\n",
        "                        loss.backward()\n",
        "                        optimizer.step()\n",
        "                        \n",
        "                    else:\n",
        "                        outputs = model(inputs)\n",
        "                        loss = criterion(outputs, labels)\n",
        "                    \n",
        "                    _, preds = torch.max(outputs, 1)\n",
        "\n",
        "                # statistics\n",
        "                running_loss += loss.item() * inputs.size(0)\n",
        "                running_corrects += torch.sum(preds == labels.data)\n",
        "\n",
        "            epoch_loss = running_loss / (len(dataloaders[phase])*batch_size)\n",
        "            epoch_acc = running_corrects.double() / (len(dataloaders[phase])*batch_size)\n",
        "\n",
        "            print('{} Loss: {:.4f} Acc: {:.4f}'.format(phase, epoch_loss, epoch_acc))\n",
        "            print (len(dataloaders[phase].dataset))\n",
        "\n",
        "            # deep copy the model\n",
        "            if phase == 'val' and epoch_acc > best_acc:\n",
        "                best_acc = epoch_acc\n",
        "                best_model_wts = copy.deepcopy(model.state_dict())\n",
        "            \n",
        "            if phase+'_acc' in history:\n",
        "                # append the new number to the existing array at this slot\n",
        "                history[phase+'_acc'].append(epoch_acc)\n",
        "            else:\n",
        "                # create a new array in this slot\n",
        "                history[phase+'_acc'] = [epoch_acc]\n",
        "            \n",
        "            if phase+'_loss' in history:\n",
        "                # append the new number to the existing array at this slot\n",
        "                history[phase+'_loss'].append(epoch_loss)\n",
        "            else:\n",
        "                # create a new array in this slot\n",
        "                history[phase+'_loss'] = [epoch_loss]            \n",
        "\n",
        "    time_elapsed = time.time() - since\n",
        "    print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))\n",
        "    print('Best val Acc: {:4f}'.format(best_acc))\n",
        "\n",
        "    # load best model weights\n",
        "    model.load_state_dict(best_model_wts)\n",
        "    return model, history"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "EsXuO63loxsr",
        "colab_type": "code",
        "outputId": "c3b4b893-f33d-4bd7-f91c-1f15a36c7477",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 605
        }
      },
      "cell_type": "code",
      "source": [
        "model, history = train_model(model, dataloaders_dict, criterion, optimizer, epochs)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 0/4\n",
            "----------\n",
            "train Loss: 0.3495 Acc: 0.8918\n",
            "60000\n",
            "val Loss: 0.0586 Acc: 0.9565\n",
            "10000\n",
            "Epoch 1/4\n",
            "----------\n",
            "train Loss: 0.1026 Acc: 0.9675\n",
            "60000\n",
            "val Loss: 0.0391 Acc: 0.9639\n",
            "10000\n",
            "Epoch 2/4\n",
            "----------\n",
            "train Loss: 0.0758 Acc: 0.9748\n",
            "60000\n",
            "val Loss: 0.0404 Acc: 0.9658\n",
            "10000\n",
            "Epoch 3/4\n",
            "----------\n",
            "train Loss: 0.0616 Acc: 0.9787\n",
            "60000\n",
            "val Loss: 0.0311 Acc: 0.9653\n",
            "10000\n",
            "Epoch 4/4\n",
            "----------\n",
            "train Loss: 0.0532 Acc: 0.9813\n",
            "60000\n",
            "val Loss: 0.0329 Acc: 0.9663\n",
            "10000\n",
            "Training complete in 1m 9s\n",
            "Best val Acc: 0.966309\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "a0LIJT0xoxs0",
        "colab_type": "code",
        "outputId": "82f80ef4-8925-4832-bb31-2841a6c1aa72",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        }
      },
      "cell_type": "code",
      "source": [
        "# summarize history for accuracy\n",
        "plt.plot(history['train_acc'])\n",
        "plt.plot(history['val_acc'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()\n",
        "# summarize history for loss\n",
        "plt.plot(history['train_loss'])\n",
        "plt.plot(history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEVCAYAAAAPRfkLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3XmYXHd15/93dfW+SS2ppe7Wvvl4\nw8Z2JFsYyza2sVnMEgPhGRLsGIaQQMYB5scPZiCBAeJAYEwyzvzCZgiDg0nAEH6ALa+SN7Bs2Rhs\nywdka2mptbTU+961zB/3Vqla6qVa6urq7vq8nqeerrpbnb5Sf8+93++950aSySQiIiIARfkOQERE\nZg4lBRERSVNSEBGRNCUFERFJU1IQEZE0JQUREUlTUhABzOybZvaZCZa5ycwemKaQRPJCSUFERNKK\n8x2AyGSZ2Srgl8BtwPuACPBe4NPAq4Et7n5zuOw7gb8h+L/eAvxnd3/ZzBYC3wfWAy8CfcD+cJ2z\ngf8PaAQGgT9196cniOnTwB+H37MT+GN37zCzCuBrwGXAAPAFd//eONO/A+xy98+H201/NrM9wB3A\ne4BrgArgW8BCoAT4tLt/P1zvOuAr4fTfhfvna8CT7v7lcJlzgYeBRnePZbf3Za7TmYLMVouAQ+5u\nwG+AHwA3AucB/8nM1prZCuAbwNvc/Uzg5wQNI8D/C7S6+2rgQ8C1AGZWBPwE+K67nwF8EPgPMxvz\nAMrMLgI+DGwgSDJl4WeAjwGl4fdcA9xuZk3jTJ/IMnc3d98HfBn4mbufBdwMfMvMSsysCrgT+KPw\nd9gFfI4gCf6njG29HfiREoJkUlKQ2aoY+Pfw/W+Bp9z9qLsfAw4CTQSN7cPuvitc7pvAlWEDvxn4\nNwB33wNsC5c5E1hMcESOuz8OtAKvGSsQd98BLHf3LndPAE8Aa8LZbwTuCpfbT9Cot4wzfSI/y3j/\nVuDvw/ePAeUEZzeXAs3u/nw47+PAR4BfAGvNzMLpbydIpiJp6j6S2Sru7v2p90BP5jwgCtQD7amJ\n7t5pZhGCs4wFQGfGOqnl5gOVwM7jbSe1BF00ozKzSuA2M7sinLSA4KyE8Ls6MmLomWD6RNoy3l8L\nfMrM6oEEQTda0SjbHsqI9ccEZ1LfIkgg2xDJoKQgc9lhYFPqg5nVETSeRwmSwLyMZeuBVwjGHbrC\n7qYRzOymMb7nrwi6jS5y9x4z+wKwNJx3lKCRTm1jGUHDPtb0VEJLqRvtC82shOBM6V3u/gszKwNS\nSfLEbVcCC8Izku8TjMV0Aj8Mz2xE0tR9JHPZ/cBmM0t15XwQuC/sQ/8lQfcJZrYWeG24zF5gv5m9\nI5y3yMy+H/bTj2Ux8FKYEFYSdA1Vh/N+CrzXzCJm1gA8S9BgjzX9IHB++N1rMuI6UVX4Sg2A3wIM\nhd/7GNBgZhvCeZ8G/jp8/wDBWc9/QV1HMgolBZmzwiPj9xMMFL9EMI7wZ+HsW4GVZrYb+F/A3eE6\nSeDdwIfDdR4BHnT33nG+6p+By83MCa74+ShwlZn9FcFR+RGCZLMV+K/hIPFY078BrDKz34cx/nCM\n360D+BLwrJk9C7xMMED+M4JupBuA75nZ7wgG3/9buF6c4AwjCjw+8V6UQhPR8xRECouZfRxY5O4f\nz3csMvNoTEGkgISD0h8AXp/vWGRmUveRSIEwsz8jGIP4oru/ku94ZGZS95GIiKTpTEFERNJm/ZhC\na2v3KZ/q1NVV0t7eN5XhTAnFNTmKa3IU1+TM1bjq62sio00v6DOF4uLoxAvlgeKaHMU1OYprcgot\nroJOCiIiMpKSgoiIpCkpiIhImpKCiIikKSmIiEiakoKIiKQpKYiISNqsv3lNRGQuGo7F6emP0d03\nRE//8PFXX/CzqqqMN2xYRmnJ1N6vkNOkYGa3AZcASeAWd38qY95bgU8Bg8Bd7n67mVUD3yV42lQZ\n8Fl335LLGHNl69YHueKKqyZc7h/+4Su8853vpqlp6YTLisjsNDQcH9mw9w/T3TdMb/8w3aM0+D39\nwwwOx8fdZrQowkZbROPC8Z7/NHk5Swpmdjmw3t03mdlZBA9C3xTOKwJuBy4EjgH3mNlPgLcB7u6f\nNLMm4CGCB6nPKgcPtvDAA1uySgq33PKxaYhIRKZKZgPffUJDfrxxH6KnP0ZP/xDd/cMMDWf31NPS\nkiJqKkpYsqCCmooSqipKqKkopbqyhOqK8FVZQnV5CetXL2Sof2jijU5SLs8UriJ4EhTuvtPM6sys\n1t27CB8s7u6tAGb2IHA1wbNlzwvXrws/zzr/839+kZ07X+Cyyzbw+te/gYMHW/jqV/83t976P2ht\nPUJ/fz833/wBLr30Mj784Q/w0Y9+nIcffpDe3h727dvLoUMtfOhDH2HTpkvz/auIzGmDw/HgaH3U\nhn2Y7v4hhuJJ2jr709OzbeDLSqJUVxTTuKBqZKMevmoqU43+8WmT6QqaV11G6yxLCg3AjozPreG0\nrvB9jZmtB/YAVwJb3f2LZnaTme0iSApvmuhL6uoqx60Bcsf//wKPP3fglH+J0Vx6/lJuvv6cMef/\n+Z//GXfeeSfr16/nlVde4d///QccO3aMq666gre//e00Nzdzyy238La3vZHS0mLq6qqoqiqjpWUf\n//Iv3+aRRx7hrrvu4i1vuW5K454K9fU1+Q5hVIprcuZiXANDMbp6h+jqHaI79bNv6KRpXb1DdIXT\nhybookkpL41SW1XKiiU11FSWUltVRk1VCbVVZdRWlVJbWRr8rC6lprKUmqpSyqa4r380ufh3nM6B\n5nRFPndPmtmNBF1KncBuIGJmfwzsc/frzOx84FvAH4y30YmqBPb3DRGPj15INRqNjDlvom22tnaP\nOb+jo4/BwWF6ewdZs+YMWlu7icWK2L59B3fe+a9EIkUcO9ZGa2s3Q0Mx2tt76e0dxOwcWlu7aWho\noK2tY9zvyIf6+poZFxMorsma6XElk0mGhhN09w/R2x+ju3/opP720frkh2NZHsGXRqmpKKFxYWVw\nlB52x2Qezae7bipLWbW8js6O7KuRJodjdHXETnU3ZO10/x3HSii5TAotBGcGKU3AwdQHd98GXAZg\nZrcSnDFcDmwJ5z9nZk1mFg0fNn5K3vW6dbzrdetGnTcdfxwlJSUA3H//vXR1dfFP//RNurq6eP/7\n/+SkZaPR40cWeviRzEXDsTht3YO0dQ3S1jVAW/cg7V0DHOsapGdgmI7uwUk18OWlUaorSli6qOqk\nRj3ofy+lurw4+BlOKyme3JX4U311z0yXy6RwH/BZ4GtmdiHQ4u7pFtjM7gFuBHqB64GvAEuBi4Ef\nmdlKoOd0EkK+FBUVEY+PDLujo4PGxiaKiorYtu0hhoeH8xSdSG7EEwk6uodo6x4IGv3Uz7Dxb+sa\noLtv7P/3FWVRqspLWFZfldHXXkp1xchGPXUUfyoNvEwsZ0nB3Z8wsx1m9gSQAD5kZjcBne7+Y+Ab\nBIkjCdzq7kfN7GvAHWa2LYztg7mKL5dWrlyN+0s0NjYxf/58AK644nV84hMf5cUXn+dNb3oLixcv\n5tvf/kaeIxXJTiKZpLtvOGjgu05o9MOfHT2DjHWCW1JcxIKaMpbVV7Ogpoy62nIW1paxoLacBTXB\nzxXL6mZkt1ahmfXPaD6dJ6/N9L7VmUZxTc5siSuZTNI3GDvpqD6z8W/vHiQ2xvhbUSRCXU0ZC05o\n5NM/a8uorighEhn1QV9jxjVTzNW4xnrymu5oFpnjBofiI7pyBuNJmg91ZSSAwTFvlIoAtVWlLF9c\nHTb05Sc1/vOqSikqGr/Bl9lDSUFkFovFE+nB2rH68XsHxr4Spqq8mMV1FSOO6jMb/LqaMoqj6rcv\nJEoKIjNUIpGks3eItq4BjmV25WQ0/l29Q4zVf1pWGmVBTRmrGmtHdOmsWVFHUSLBgppyykoL68oa\nmZiSgkgeJJNJuvuHgwY+PKo/lurHD4/8O3qGiCdGb/KLoxEW1JRjK+ZTN0qXzsLaMirKikftx5+p\nfeQyMygpiOTAcCzB3kNd7NrTlj6qb08d8XcP0t49OOa1+JEIzK8uY1VjDQvDfvy62rIR/fk1lSUU\nTTBwK3IqlBRETlMimeRwWx+7D3axu6WbVw520Xyke8yrdWorS2haVJVxVB829mGjP6+6lGiR+vEl\nP5QU8ugd77ie7373B1RWVuY7FJmE9u7BIAGkX930Dx4fzI0WRVixpJozVi6gMuzXTzX8dTVllIxT\nq0sk35QURMbRPxhjz8EuXgkb/90Hu2jvHhyxTMOCSl69bhFrmmpZ3VjL8sXVlBQXqe9eZiUlhRy4\n+eb38Ld/+xUaGho4dOggn/zkx6ivX0x/fz8DAwN85CP/D2effW6+w5QTxOIJmo/0hN1AQSI4dKxv\nxNU986pKuWD9IlY31rK6qZbVDTVUlpfkLWaRqTbnk8Ldu37Gs0d+O+q8aFFkzKs7xnPB4lfxh+ve\nPOb8zZuv5PHHH+GGG97Fo49uY/PmK1m7dj2bN1/Bjh1Pceed/8IXvvD3k/5emTqJZJIj7f3pxn/3\nwS72HR45DlBeGsVWzGd1Uy1rGoOzgLqasgnvzBWZzeZ8UsiHzZuv5Pbbv8oNN7yLxx7bxoc//BHu\nuuv/8P3v/x+Gh4cpLy/Pd4gFp7NnMN34724JuoL6ThgHWLa4Ot34r26qpXFBpe7UlYIz55PCH657\n85hH9bnq812zZi3HjrVy+PAhuru7efTRrSxatJhPf/pzvPTSi9x++1en/DvluP7BGHsPBf3/qUTQ\n1jVyHGBJXQXnrVvI6sbgLGDFkmoNAItQAEkhXzZtei1f//r/5rLLLqejo521a9cDsG3bw8RiuX8A\nR6GIxRPs2t/BMy8eCs8Aumg52jtiHKC2qpRXr1vE6sYaVjfVsqqhluoKjQOIjEZJIUcuv/xKPvjB\nm/nOd77PwEA/n//83/Dwww9www3v4oEH7uPnP/9pvkOcdZLJJEc6Ro4D7D3UQyx+/CawstQ4QNgN\ntKZJ4wAik6GkkCNnnXUO27Y9mf58550/TL9/7WsvB+BNb3rLtMc1m3T2DmWMAQSvzOJu0aIIy+qr\nOXvNQhrqylnTWEvjwiqNA4icBiUFmREGhlLjAMEdwbtbujjWNTBimcV1FbxqzcL0QPCKxdWUlkR1\nP4DIFFJSkGkXiyc40No7YiC45WjviKd21VSWcP7ahenLQVc1ahxAZDooKUhOJZNJWjv6w6P/4Iqg\nvYe7RxSDKy0pYv2y+cHloE21rA4LwWkcQGT6KSnIlOpKjQOkzgJaRo4DFEUiLKuvChv/4CygcVGl\nCsCJzBBKCnLKBofi7D3czSsZA8FHO0eOA9TPL+ec1QvSZwErltRQVqL7AURmKiUFyUoimWR3Syc7\nXjgYnAW0dHPgaM+IcYDqihLOW7swfTno6sYaaipL8xe0iEyakoJk5ds/38njzx9Kfy4tLmLd0nnp\newFWN9ayaJ7GAURmOyUFmdCBo708/vwhli+p5nUXLGV1Yy1L66s0DiAyBykpyITu274PgD95w1ms\na6jJczQis18imSCWiBNLxIglY8QSMYYTwc/gFWc4MRzOj48yP8aizlrOqzmfaNHUjtEpKci4OnoG\n+eULh1hSV8HGcxppO9aT75BmlOFEjIHYAP2xAQZiAwzEU+8HKeuM0ts7SFEkShERopEiioqC90WR\nolFf0XGmRygiWhT+TM+LUBSJBvMjwXZldIlk4qSGNf05eUJDnDjeEFd0Rmnv6iUWjzGcHLl+sEw8\nXH/4eEN/YiOeHLnteDI+Jb/Tf9vYyNLqxinZVoqSgozrwR37icWTXLtxBdE5VD4ikUyEjflgRkMe\nvPrjQaPef0Ij3x/rZyA+OGLZ2BT9cU+VSDrhpH5Gj78nSCSlxcUkE4yeeE74mZl4gm0cT0Kjfsco\n2zopyTH6d9T2V9LW0Z1uoE9uWE8+mk41xMMZR9yjN/pxEsnExDtwihRHohQXFY94lReXU1JUTHGk\nmOKiYH5JUUn6fXGkmJJoan5xsGz6FaW4qISSjO2uXNJAdWz+1Mc+5VuUOWNgKMbWZw9QU1nCa85t\nyHc4QHAz3FBiOGigw4b7YLyIg8faGIgNMhDrpz8+OOLoPWjcjzfm/fEBhuJDp/T9pUUlVBSXU1VS\nycKKBVREyykvLqe8uIyK4nLKo+XBz+IyFs6rpbOrj0QyEbxIEE8mSCaTxMNGKpFMkhjxPpF+xcN1\nEonj6584P5labsQrtZ04CZLh/Hh6eiwRZzgeG3NbSSb/4KmZIN2Iho1qWbSMqpKqdIOaalxLIsUn\nNdgj142OaKwXzq+hr2d41Ib7+LJhwx0pJloUnZYztvoFuSnvoqQgY3r0NwfpHYjxtteupnQK7i2I\nJWLHj7rjxxv1zKPz4+9Hb9AH4oOndMQXjUTDRruM2tJFYUM+shE/sVGvKK6gPBpOD9edTP/tTK3J\nNFFcyczkRJC0xk5myfET1CiJ7OQEFmyzqrqMgb7YGEfK4zfc0Ug0Z1e+zdR/x1xRUpBRxRMJ7n+q\nmdLiIq68cCnxRJyjvW209Bw7oVtllIY7NhAerR9v+PvjA8QSk3+ORIQIZWHDPK+sloYTGu7y4nIW\n1s4jMRihvDiYHhy9H2/MK6LlFBcV63LZLEUiEaKRKFGm9ybDQmt8ZyolBTlJMplk2wu7aI/uZuWr\n43x954s0dx9gODE8qe2UhF0tFSXlLCivG/VovDzdiJenG/XMo/OyaOmEp+JqTESmjpKC0Dfcx56u\nZvZ2NbOnax97uprpGe6ldB0cBCKdEZqqG1i1YClF8ZKTGu7U5/LiiuNdL9HyKb9UTkRyT0mhwAwn\nYuzvbkkngL1dzRzpPzpimZriWmLHGmiqbOI9l17M8pqllEVLdUQuUgCUFOawRDJBa9/R4Cygu5k9\nnc3s72kZcY10RXE5Z9atZ1XtclbWLmdl7Qq+/dNXOPLyMf74Ty5i7fx5efwNRGS6KSnMIV1D3eEZ\nQDN7Ovext3s//bH+9PxoJMrS6kZW1a5gVe1yVtUup75y0Yg++wOtPfzm5WOsXzaPtUuVEEQKjZLC\nLDUUH2Jf94H0GMDermbaBtpHLFNfsZBzF57JytrlrKpdwbLqRkqi4z+9bMv2ZgCu27giZ7GLyMyV\n06RgZrcBlwBJ4BZ3fypj3luBTwGDwF3ufns4/T3Ax4EY8Nfu/vNcxjgbJJIJDvYeHjEQfLD38Ijr\n9atLqkYkgJW1y6kqqZzU97R3ByUtGhZUcv76RVP9a4jILJCzpGBmlwPr3X2TmZ0F3AFsCucVAbcD\nFwLHgHvM7CdAP/A3wEVANfBZoKCSQjKZ5GhfG88eeen4YHD3/hF34JYUFae7gFJJYGF53Wlfh//g\njv3EE0mu3bicIl3TL1KQcnmmcBXwEwB332lmdWZW6+5dwCKgw91bAczsQeBqgqTwgLt3A93AB3IY\n34zQH+tnb9f+9FjA3q59dA4dv8InQoQlVYvTYwCralfQVNUw5Zd79g/GePjZA9TOoJIWIjL9cpkU\nGoAdGZ9bw2ld4fsaM1sP7AGuBLaGy1Wa2U+BOuAz7v7geF9SV1dJcfGpN5D19dNXCjqWiLOvYz+7\n2vbw+2N72NW2h5auwyNqzdRVzGPj0lezbuEq1i1YxZoFK6gsqch5bP/xyMv0D8a44bozaWocu8jW\ndO6vyVBck6O4JqeQ4prOgeZ0f4S7J83sRoIupU5gd8b8hcDbgZXAw2a20t3HrNDV3t53ygHl8rr7\nZDLJ0f429oZjAHu6mmnuOTCi1ENZtJT189eEXUDLWTVvBfPL5o2Iq7cjRi+5vTcgFk/w44d/T2lJ\nERutfsx9MlPvU1Bck6O4JmeuxjVWQsllUmghODNIaSK4QRYAd98GXAZgZrcSnDFUAE+4ewx42cy6\ngXrgSA7jnBI9Q73BvQAZN4X1Dh9PWEWRIpqqGsJxgGA8oKFq8Yyof/+0H+FY1yBXXbiM6orxr04S\nkbktl0nhPoKB4q+Z2YVASzhWAICZ3QPcCPQC1wNfAcqA75jZFwm6j6qBoyduON+G4sPs72kZcTXQ\n0f5jI5ZZWF7HmXXr0wPBy2uaKI3OvIfYJ5NJ7n1yH5EIXLNxeb7DEZE8y1lScPcnzGyHmT0BJIAP\nmdlNQKe7/xj4BkHiSAK3uvtRADP7IfCrcDN/6e7T92SMUSSSCY70taa7gPZ27WN/z8ERl4NWFldw\n1oIzRlwNVFNanceos7dzbzv7DvfwB2cuZvH83I9diMjMltMxBXf/xAmTnsuYdzdw9yjrfA34Wi7j\nGk/nYNeILqC9XfsZiA+k5xdHoiyvWTryruCKRbO2LPO94fOXdbOaiECB39E8MDzA79tfzjgLaKZ9\nsGPEMosrF3Fe7dnpweCl1U2UFM2N3bb/SA/Pv9LGGcvns6apNt/hiMgMMDdat1Nw966f8VDzoyST\nxy9sCu4KPivjxrBlVE7yruDZZEvqLOFinSWISKBgk0JlcQVnLlpHU3ljehxgQfn8WdsNNFnt3YP8\n6sXDNC6s5Ly1C/MdjojMEAWbFK5bdRX19W+bkdcfT4cHnm4OS1qsUEkLEUnL/0XyMu36B2Ns/fUB\naqtK2XTOknyHIyIziJJCAXrkuRb6B+NcfdEySk6jRIiIzD1KCgUmFk9w/9PNlJVEueKCpfkOR0Rm\nGCWFAvPUziO0dQ1y2XmNKmkhIidRUiggyWSSe7eHJS02qKSFiJxMSaGAvLinneYjPWw4czH1Kmkh\nIqNQUiggqZIW16qkhYiMQUmhQOw73M0Lu9s4c8V8VjeqpIWIjE5JoUBs2d4MqKSFiIxPSaEAtHUN\nsH3nYZoWVXHuGpW0EJGxKSkUgAee3h+WtFiukhYiMi4lhTmubyAoaTGvupRLzm6YeAURKWhKCnPc\ntucOMDCUKmmhf24RGZ9aiTksFk/wwNP7VdJCRLKmpDCHPfniYdq7B9l8fhNV5SppISITU1KYo5LJ\nJFu276MoEuGaDcvyHY6IzBJKCnPUC7vb2N/ay4azFrNonkpaiEh2lBTmqFRJi+tU0kJEJkFJYQ7a\ne6ibF/e0c9bKOlY21OQ7HBGZRZQU5qAtT4VnCSppISKTpKQwxxzrHGD7i0dYWl/FuasX5DscEZll\nlBTmmPufbiaRTHLdxhVEVNJCRCYpq6RgZmpdZoG+gWG2PdfC/OpSLj57Sb7DEZFZKNszhb1m9nkz\nW5PTaOS0bP11C4NDca75g+UUR3USKCKTV5zlchuBdwB3mNkw8G3gh+4+lLPIZFJi8QT3P91MWWmU\ny1/dlO9wRGSWyupw0t0Pufvt7n4F8Ofh62B49lCeywAlO7964TCdPUNcfn4TlSppISKnKOs+BjPb\nbGZ3APcAjwOvBTqAf89RbJKlVEmLaFGE129Ynu9wRGQWy6r7yMx2AXuArwN/5u7D4aydZva2HMUm\nWfrtK20cONrLJecsYUGtTtxE5NRlO6ZwHRBx998DmNkF7v5sOO+ynEQmWduikhYiMkWy7T66Cfhk\nxudPmNnfAbh7cqqDkuztPdTNzr3tnL2qjhVLVNJCRE5PtknhSne/OfXB3f+IYExB8uyeJ/cCKmkh\nIlMj2+6jUjMrTV2CambVwISXuJjZbcAlQBK4xd2fypj3VuBTwCBwl7vfnjGvAnge+Jy7fyfLGAvO\n0Y5+nn6plWX11ZyzSiUtROT0ZXum8M8Eg8o/MLMfAi+E08ZkZpcD6919E/A+4B8z5hUBtwNvBDYD\n15tZ5pNgPgW0Zf1bFKj7UiUtLl6ukhYiMiWyvU/hW8CVwL8B/0rQdXT3BKtdBfwkXH8nUGdmteG8\nRUCHu7e6ewJ4ELgawMzOBM4Gfj65X6Ww9A4M8+hzB6mrKWPjWSppISJTYzK1EKqBVuAocCbwqwmW\nbwiXT2kNp6Xe15jZejMrIUg4qZbtK8BHJxFXQdr67AEGh1XSQkSmVrb3KfwD8HqCRn0XsBb48iS/\nK92/4e5JM7sRuAPoBHYDETN7L/BLd99tZllttK6ukuLi6CRDOa6+fmZesTNeXMOxOA89c4DK8mJu\nuPqMab2DeTbur3xSXJOjuCYnF3FlXfvI3c8ys4fd/Uozuwh4+wTrtHD8zACgCTiY+uDu2wjvcTCz\nWwlujns7sMbM3gwsAwbNbL+7PzDWl7S392X5K5ysvr6G1tbuU14/VyaK69HnWmjvHuS6jSvo7R6g\nt3tgRsSVL4prchTX5MzVuMZKKNkmhcHwZ5mZRdx9h5lNdKZwH/BZ4GtmdiHQ4u7p38DM7gFuBHqB\n64GvuPtdGfM/A+wZLyEUokQyyb1hSYur/2DZxCuIiExCtknBzewvgEeA+83MgfnjruD+hJntMLMn\ngATwITO7Ceh09x8D3yBIHEngVnc/eqq/RCH57cvHOHisj03nNKikhYhMuWyTwgeBOoICeO8mGBS+\ndaKV3P0TJ0x6LmPe3YxzBZO7fybL2ArKvU/q+csikjvZJoXb3P2vwvf/mqtgZHy7D3bhzR2cs3oB\nyxdX5zscEZmDsk0KcTN7HfAEkH6wTniPgUwTnSWISK5le4H7+4H7gT4gFr6Gx11DplRrRz9P+xFW\nLK7m7JV1+Q5HROaorM4U3H1ergOR8d33VDPJJFx78QqVtBCRnMn25rX/Mdp0d//rqQ1HRtPTP8yj\nv2lhQW0ZG85cnO9wRGQOy7b7KJ7xihKUpdDZwzR5+NkDDA0nVNJCRHIu2+6jz2Z+NrMo8KOcRCQj\nDMfiPLhjPxVlxWw+vynf4YjIHHeqh50lwLqpDERG98sXDtPVO8QVr26ioizbi8VERE5NtmMKzQR3\nHqcsAL6Ti4DkuEQyyZZ0SYvl+Q5HRApAtoeemY/eTAJd7t6Rg3gkw292BSUtLj23gbqasnyHIyIF\nINvuoyrgg+6+1933AbeZ2Tk5jEuAe8PnL1+rm9VEZJpkmxT+CfhFxudvhdMkR15u6eR3+zs5d80C\nltWrpIWITI9sk0Kxuz+a+uDuj5Hx0ByZelvCkhZv2KizBBGZPtmOKXSa2Z8DWwkSyXXAzHvqxBxx\n8GgvO37XysolNZypkhYiMo2yPVP4U+Ai4N+A7xNcjvqnuQqq0P3HIy+HJS2Wq6SFiEyrrJKCu7cC\nX3T3V7n7ecDXw2kyxXr6h7nOd8d2AAAPt0lEQVR/+z4WqqSFiORBVknBzL4AfDJj0ifM7O9yE1Jh\ne+iZ/QwNx7lmwwqiRSppISLTK9tW5wp3vzn1wd3/iJH3LsgUSJW0qKoo4bLzGvMdjogUoGyTQqmZ\nlaY+mFk1QakLmUKPP3+I7r5h3rBplUpaiEheZNvy/DOw08yeJqiSugH4as6iKkBBSYtmokURrr9s\nDfFBPcNIRKZftgPN3yK42ugHwJ3Ap4EP5DCugvPc749yuK2PTec0sKC2PN/hiEiByrYg3leBa4EG\nYBewFvhyDuMqOPdsD25Wu3ajCt+JSP5kO6ZwsbufBfza3TcA1wCVuQursOw60Mmu/Z2ct3YhS1XS\nQkTyKNukMBj+LDOziLvvAC7NUUwFJ1XS4jqVtBCRPMt2oNnN7C+AR4D7zcyB+bkLq3Acbu/jmd+1\nsqqhBluhXSoi+ZVtUvggUAd0AO8GlgC35iqoQnLf9maSwHUXr1BJCxHJu2yf0ZwE2sKP/5q7cApL\nV98Qj/32IIvmlXOR1ec7HBGRU35Gs0yBh585wHAswTUblqukhYjMCGqJ8mRwOCxpUV6skhYiMmMo\nKeTJE88foqd/mCsuWEp5qUpaiMjMoKSQB4lEki3b91EcjXD1RcvyHY6ISJqSQh48+/ujHGnvZ9M5\nDcyrLst3OCIiaUoKeXDv9r0AXKub1URkhlFSmGa79nfy8oEuXr1uEU2LqvIdjojICEoK0+yeJ1Nn\nCSp8JyIzj5LCNDrU1sevf3+U1Y21nLFcJS1EZObJ6bWQZnYbcAmQBG5x96cy5r0V+BRBsb273P32\ncPqXgMvC2G5197tzGeN0um/7PpW0EJEZLWdnCmZ2ObDe3TcB7wP+MWNeEXA78EZgM3C9mS0zsyuB\nc8N1rmMOPd2tq3eIx357KChpcYZKWojIzJTL7qOrgJ8AuPtOoM7MasN5i4AOd2919wTwIHA1QRXW\nd4bLdABVZhbNYYzT5qFn9hOLJ7h24wqKinSWICIzUy67jxqAHRmfW8NpXeH7GjNbD+wBrgS2unsc\n6A2Xfx/wi3DamOrqKikuPvW8UV9fc8rrZmtgKMbDz7ZQU1nC265cT3nZxLt9OuI6FYprchTX5Ciu\nyclFXNNZXyF9eOzuSTO7EbgD6AR2Z84PxxveB7x+oo22t/edckD19TW0tnaf8vrZeuiZ/XT3DfHm\n16yiu6ufib5xuuKaLMU1OYprchTX5JxuXGMllFwmhRaCM4OUJuBg6oO7byMYUMbMbiU4Y8DMrgX+\nO3Cdu3fmML5pkUgkuW97M8XRIq5SSQsRmeFyOaZwH/AOADO7EGhx93RaM7N7zGyxmVUB1wMPmNk8\n4O+BN7t722gbnW2e+V0rRzr6ufRVDcyrKs13OCIi48rZmYK7P2FmO8zsCSABfMjMbgI63f3HwDcI\nEkeS4NLTo2b2AYJB6H8zs9Sm3uvu+3IVZy4lk0nu3b6PCPD6DbpZTURmvpyOKbj7J06Y9FzGvLuB\nu09Y/uvA13MZ03T6/f5OXmnp4oL1i2hcqJIWIjLz6Y7mHLr3yeAER4XvRGS2UFLIkYPHevn1rqOs\nbapl/bJ5+Q5HRCQrSgo5smV7MxCcJaikhYjMFkoKOdDZO8QTzx9i8fwKLlRJCxGZRZQUcuDBHUFJ\ni9dvXK6SFiIyqygpTLHBoTgPP7Of6ooSLn1VY77DERGZFCWFKfbYbw/SOxDjdRcupaxkTtTyE5EC\noqQwheKJBFu276OkuIjXqaSFiMxCSgpT6JnfHeVo5wCXvqqR2kqVtBCR2UdJYYokk0nufXIvEeBa\nlbQQkVlKSWGK/K65g90Hu7ngjHqWLKjMdzgiIqdESWGKpEpaXKeSFiIyiykpTIEDR3t57uVjrFs6\nj3UqaSEis5iSwhS4b7sK34nI3KCkcJo6egb55QuHWFJXwQXrF+U7HBGR06KkcJqCkhZJrt24QiUt\nRGTWU1I4DQNDMbY+e4CayhJec27DxCuIiMxwSgqn4dHfBCUtrrpwGaUqaSEic4CSwimKJxLc/1Qz\npcVFXHnh0nyHIyIyJZQUTtEObw1KWpzXSI1KWojIHKGkcAqSyST3PLmPCPB6lbQQkTlESeEUvLSv\ng72HurnQ6llSp5IWIjJ3KCmcgi3bVdJCROYmJYVJOtDaw29ePsb6ZfNYu1QlLURkblFSmKQt25sB\nnSWIyNykpDAJ7d1BSYuGBZWcr5IWIjIHKSlMwoM79hNPJLl243KKIippISJzj5JClvoHYzz87AFq\nVdJCROYwJYUsPfqbg/QPxrjqomWUFKukhYjMTUoKWYjFE9z/1D5KS4q48sJl+Q5HRCRnlBSy8PRL\nRzjWNchlr2qiuqIk3+GIiOSMksIEkskk927fRyQC12xUSQsRmduUFCawc287+w73cJEtZvH8inyH\nIyKSU0oKE7hXJS1EpIAoKYxj/5Eenn+ljTOWz2dNU22+wxERybniXG7czG4DLgGSwC3u/lTGvLcC\nnwIGgbvc/faJ1plu6cJ3F+ssQUQKQ87OFMzscmC9u28C3gf8Y8a8IuB24I3AZuB6M1s23jrTrb17\nkF+9eJjGhZWct3ZhvsIQEZlWuew+ugr4CYC77wTqzCzVB7MI6HD3VndPAA8CV0+wzrR64OnmsKTF\nCpW0EJGCkcvuowZgR8bn1nBaV/i+xszWA3uAK4GtE6wzqrq6SopP4w7j+vqak6b1DQyz7bkW5teU\ncf3l6ygtmf47mEeLayZQXJOjuCZHcU1OLuLK6ZjCCdKH2+6eNLMbgTuATmB35vzR1hlLe3vfKQdU\nX19Da2v3SdPvfXIffQMxrtu4gs6OU9/+VMeVb4prchTX5CiuyTnduMZKKLlMCi0ER/kpTcDB1Ad3\n3wZcBmBmtxKcMZSPt850iMUT3P90M2UlUa64YOl0frWISN7lckzhPuAdAGZ2IdDi7um0Zmb3mNli\nM6sCrgcemGid6fDUziO0dw9y2XmNKmkhIgUnZ2cK7v6Eme0wsyeABPAhM7sJ6HT3HwPfIEgCSeBW\ndz8KHD1xnVzFN5oRJS02qKSFiBSenI4puPsnTpj0XMa8u4G7s1hn2ry4p53mIz1sPGsx9SppISIF\nSHc0Z0iVtLhWJS1EpEApKYT2He7mhd1tnLliPqsbVdJCRAqTkkJoy/ZmQCUtRKSwKSkAbV0DbN95\nmKZFVZy7RiUtRKRwKSkA96dLWixXSQsRKWgFnxT6BmJs+3UL86pLueTsholXEBGZwwo+KWx77gAD\nQ3GuvmgZJcUFvztEpMAVdCs4HEvwwNP7VdJCRCRU0Enh0V/vp717kM3nN1FVrpIWIiIFmxSSySQ/\n3voyRZEI12xYlu9wRERmhIJNCi/sbmPPwS42nLWYRfNU0kJEBAo4KTz0zAEArlNJCxGRtOl8yM6M\nYivms35lHSsbZuYTlURE8qFgk8K1G1fM2CcqiYjkS8F2H4mIyMmUFEREJE1JQURE0pQUREQkTUlB\nRETSlBRERCRNSUFERNKUFEREJC2STCbzHYOIiMwQOlMQEZE0JQUREUlTUhARkTQlBRERSVNSEBGR\nNCUFERFJU1IQEZG0gnjIjpndBlwCJIFb3P2pjHlXA38LxIFfuPvnZkhce4DmMC6A97j7gWmM7Vzg\nP4Db3P32E+blc5+NF9ce8rTPzOxLwGUEf1O3uvvdGfPyub/Gi2sPedhfZlYJfAdYApQDn3P3n2XM\nz8v+yiKuPeT3b7ICeD6M6zsZ06d0f835pGBmlwPr3X2TmZ0F3AFsyljkH4FrgQPANjP7kbu/OAPi\nAniDu/fkOpZRYqsC/hfw4BiL5GufTRQX5GGfmdmVwLnhv+VC4Fng7oxF8rW/JooL8vN/7HrgaXf/\nkpmtBO4HfpYxPy/7K4u4IE9/k6FPAW2jTJ/S/VUI3UdXAT8BcPedQJ2Z1QKY2Rqgzd2b3T0B/CJc\nPq9xzQCDwBuBlhNn5HmfjRlXnj0CvDN83wFUmVkU8r6/xowrn9z9B+7+pfDjcmB/al4+99d4ceWb\nmZ0JnA38/ITpU76/5vyZAtAA7Mj43BpO6wp/tmbMOwKsnQFxpfyzma0CHgM+6e7TUpPE3WNAzMxG\nm523fTZBXCnTvs/cPQ70hh/fR3AKn+piyOf+Gi+ulLz8HwMwsyeAZcCbMybn829yvLhS8rW/vgJ8\nGLjxhOlTvr8K4UzhRJFTnJdrJ373XwMfBa4AzgVumO6AspTPfXaivO4zM3srQeP74XEWm/b9NU5c\ned1f7v4a4C3A98xsrP0y7ftrnLjysr/M7L3AL919dxaLn/b+KoSk0EKQTVOagINjzFvK9HVNjBcX\n7v5ddz8SHh3/AnjVNMU1kXzus3Hlc5+Z2bXAfyfoc+7MmJXX/TVOXHnbX2Z2kZktD2P4NUGPRX04\nO2/7a4K48vn/603AW83sV8D7gU+Hg8uQg/1VCEnhPuAdAGZ2IdDi7t0A7r4HqDWzVWZWTHC6eF++\n4zKzeWa2xcxKw2UvJ7jqIO/yvM/GlM99ZmbzgL8H3uzuIwYC87m/xosrz//HNgMfC+NYAlQDRyHv\n/7/GjCuf+8vd/8jdN7j7JcA3Ca4+eiCct4cp3l8FUTrbzP6O4B88AXwIuADodPcfm9lm4Ivhoj9y\n9y/PkLhuIeg/7Ce4auQvp6v/0swuIujDXAUME1zV8FNgdz73WRZx5WWfmdkHgM8Av8uY/BDw2zzv\nr4niytf+qgC+RTCYWwF8FlhInv8ms4grb3+TGTF+BtgTfszJ/iqIpCAiItkphO4jERHJkpKCiIik\nKSmIiEiakoKIiKQpKYiISJqSgkgemdlNZva9fMchkqKkICIiabpPQSQLZvaXwLsISh+8BHyJoKzy\nPcD54WLvdvcDZvYmgjo5feHrA+H0i4GvAkMEJZDfS1A/5w8JCiGeDewF/nC6b4oSSdGZgsgEzGwj\n8HZgs7tvIihDfTWwBvi2u18GbAU+Fj6o5ZvADe5+JUHS+Hy4qe8B/9ndLwe2EdS0ATgH+ABwEUGh\ntQun4/cSGU0hlM4WOV1XAOuAh8Oy3VUEhceOuXuq/PnjwF8BZwCH3T1Vi38r8EEzWwTMd/fnAdz9\nqxCMKQBPuXtf+PkAMD/3v5LI6JQURCY2CPzU3dOlp8Oa+s9kLBMheKzqid0+mdPHOjOPjbKOSF6o\n+0hkYo8DbzCzagAz+wugkeBpeReEy7wW+A1B8bnFZrYinH418Ct3PwYcNbMN4TY+Fm5HZEZRUhCZ\ngLs/DfwTsNXMHiPoTuokqNJ6k5k9BFwK3Obu/QQPtPmBmW0leDTip8JN/QnwD2a2jaA6ri5FlRlH\nVx+JnILUIxndfVm+YxGZSjpTEBGRNJ0piIhIms4UREQkTUlBRETSlBRERCRNSUFERNKUFEREJO3/\nAvleCG8obt5yAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f4437524320>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEVCAYAAAAPRfkLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xl8nPV57/3PjEbbaLMkS5bl3cZc\nGBMWsxoDhgAhBAgnCUnb8DSlCaFNSR5O2/P05OQkfUJ70qRtWtK0SU/StOXkJOkheUIIYQlLAg7g\nAMZhNfYFeLdlLbZlSZYsa5vnj/vWeCRrGckajZbv+/Xya+b+3fdv5prb0ly6fr97iSQSCURERACi\n2Q5ARESmDiUFERFJUlIQEZEkJQUREUlSUhARkSQlBRERSVJSEDkFZvYdM/viKNvcZmZPptsukk1K\nCiIikhTLdgAik8XMlgK/Bu4BPgFEgI8BXwDOBR5z94+H234Y+H8JfkfqgE+6+3YzqwT+A1gJvAl0\nAPvCPmcC/wzMB44Dv+/uL6UZWwXwP4FzgF7gf7n7X4fr/gfw4TDefcD/5e51w7WPd/+IgCoFmX3m\nAvXubsBrwH3A7wFnAx81sxVmthj4F+A/ufsZwMPAt8L+/xVocvdlwJ3AdQBmFgUeAL7r7qcDfwj8\n1MzS/cPrr4DmMK7LgD8ys8vMbDXwEeCs8HV/AlwzXPv4d4tIQElBZpsY8KPw+evAJnc/6O6HgANA\nLXAt8JS7vxNu9x3gqvAL/grghwDuvgvYEG5zBlAN/Fu47jmgCbg0zbhuAL4Z9j0M3A+8BzgCVAG3\nmlm5u/+ju393hHaRU6KkILNNr7sf638OHE1dB+QQfNk29ze6ewvBEM1coAJoSenTv90cIA5sNbNt\nZraNIElUphnXgPcMn1e7+37ggwTDRHvM7GEzWzRce5rvJTIszSmInKwBWNu/YGblQB9wkODLuixl\n2ypgB8G8Q2s43DSAmd2W5ntWAnvC5cqwDXd/CnjKzIqArwJfAW4drj3tTykyBFUKIid7ArjCzJaH\ny38IPO7uPQQT1R8AMLMVBOP/ALuBfWZ2S7hurpn9R/iFnY6HgDv6+xJUAQ+b2XvM7BtmFnX3duBV\nIDFc+6l+cBElBZFB3H0fcDvBRPE2gnmEPwhXfxlYYmY7gX8kGPvH3RPAbwOfDvv8CvhF+IWdjs8D\n5Sl9v+LuL4bP48BbZrYF+C3gz0doFzklEd1PQURE+qlSEBGRJCUFERFJUlIQEZEkJQUREUma9ucp\nNDW1jXumvLw8TnNzx0SGMyEU19gorrFRXGMzU+OqqiqJDNU+qyuFWCwn2yEMSXGNjeIaG8U1NrMt\nrlmdFEREZCAlBRERSVJSEBGRJCUFERFJUlIQEZGkjB6Samb3AJcQXL3xLnfflLLukwS3ROwluMLj\nncB6ghugbAk3e93dP5PJGEVE5ISMJQUzWw+sdPe1ZraK4I5Ua8N1cYIrSl7u7t1m9ktOXL9+g7vf\nkqm4RERkeJkcPrqa4J61uPtWgssCl4bLHe5+dZgQ4gQ3LanPYCwn2eyNPPni7sl8SxGRKS+Tw0c1\nwOaU5aawrbW/wcw+C9wFfM3dd4Q3TD/TzB4kuO3h3e7+xEhvUl4eH9dJHM/d/zqvvt3EN/6fd7No\nXsmY+4/mscce47rrrht1uy996Ut87GMfY9GigXdSrKqa+JgmguIaG8U1NoprbDIR12Re5uKkU6rd\n/Stm9g/AI2b2LPA2cDfBjdGXE9xq8DR37xruRcd7mve61TW88lYT331oC3e8f/W4XmM4Bw7Ucf/9\nD7Bmzej3bL/jjv8bgKamtmRbVVXJgOWpQnGNjeIaG8U1Nqca13AJJZNJoY6gMuhXCxwAMLMK4Cx3\n/5W7HzOzR4F17v4ccF+4/XYzqwcWADsnOrjzTp/L0vmlvLC1gZvWLWV+Zbp3TRzd3//9X7N16xYu\nv/xC3vOe6zlwoI6vfe2bfPnLf0FTUyPHjh3j4x+/g3XrLufTn76DP/mTP+Opp35Be/tR9uzZTX19\nHXfe+cesXbtuwmISEUlHJpPC4wR/9X/LzNYAde7en9ZygXvN7Gx3PwpcBPxvM7sVmO/uXzWzGmAe\nsP9UgvjhL99h07bGIdd19/aRSMBf3LuJ4sK8tF/zwjOq+ci7Txt2/e/8zu9y//0/ZNmyFezZs4tv\nfvM7NDcf5qKLLuH6629k//59fOELn2XdussH9GtsbOCrX/06W7e+zHe/+z0lBRGZdBlLCu6+0cw2\nm9lGoA+408xuA1rc/Sdm9hcEw0M9BIekPggUAz8ws5uBPOBTIw0dnaqCvBzaj0U43t1HYX4fOdGJ\nn3dftSoYmiopKWXr1i08+OD9RCJRWltbTtr27LPPBaCmpoajR49OeCwiIqPJ6JyCu392UNOrKevu\nBe4dtL4NuGkiY/jIu08b9q/6qqoSHn1mO9984A3OXFrB7TeeOZFvDUBubi4ATzzxc1pbW/nGN75D\na2srt9/+uydtm5NzYsJc984WkWyY9Wc0r7EqFlQV8fyWBhom6Jrp0WiU3t7eAW1Hjhxh/vxaotEo\nGzb8ku7u7gl5LxGRiTTrk0I0EuGmS5fSl0jw0HO7JuQ1lyxZhvs22ttPDAFdeeW72bjxGe6661MU\nFhZSXV3Nv//7v0zI+4mITJTIdB+mOJU7r/Uf0tWXSPDn//oi9Yc6+Ks7Lqa6PD6RIY47rqlGcY2N\n4hobxTU2E3BIqu68NpwB1cJGneUsIrOXkkLowjOqmV8ZZ+Mb9TQeOZbtcEREskJJIRSNnqgWHt64\nK9vhiIhkhZJCiotWzaOmIqgWDqpaEJFZSEkhRTQa4aZ1S+ntS/DQrzW3ICKzj5LCIBevmse8ijjP\nvX6Agy2qFkRkdlFSGCSYW1hCb1+CRzJcLdxyy010dEzMCXMiIhNBSWEIF585j+ryQp557QCHWjqz\nHY6IyKRRUhhCTjTKTZcGcwsPPz/2auHjH7+V+vrgRnL19Qf4/d//KH/2Z/+Zz3zmD/jkJ3+PN998\nY6JDFhGZEJN5k52suP+dh3i58fUh1+VEI/T2DXNCdCJB/Lwunu9L4M/mE42eOPnvvOp38cHTbhz2\nPa+44iqee+5XfOhDH+GZZzZwxRVXsWLFSq644ko2b97E97//v/jSl/72lD6XiEgmqFIYTiRCYX4O\nCRJ0HO8ZU9cgKTwDwLPPbuCyy9azYcMv+NSnPsE///M/0tJy8mWzRUSmghlfKXzwtBuH/at+tGuH\n9PT28blvP8+Ro8f5wh+spaK0IK33XL58BYcONdHQUE9bWxvPPPM0c+dW84Uv/CXbtr3JP/3T18b1\nWUREMk2VwghiOVFuvHQpPb0JHn1+z5j6rl17Gd/+9je5/PL1tLQcYcGChQBs2PAUPT1jqzxERCaL\nksIoLj2rhrllBWx4tY7mtuNp91u//iqefPIxrrzyat773hu4777v88d/fCerV5/FoUOHePjhBzMY\ntYjI+Mz44aNT1V8t3PvoNh59fjcfvfb0tPqtWrWaDRteSC5///v/X/L5ZZetB+CGG94/scGKiJwi\nVQppuPSsGipLg2rhyNH0qwURkelGSSENsZwoN1y6hO6evjHPLYiITCdKCmm67F3zqSzN5+lX9tOi\nakFEZiglhTTFcqK8b+3SoFp4QdWCiMxMGZ1oNrN7gEuABHCXu29KWfdJ4BNAL/AqcKe7J0bqk22X\nvWs+D23cxdMv7+f6S5ZQVpSX7ZBERCZUxioFM1sPrHT3tQRf/l9PWRcHfhu43N3XAWcAa0fqMxXk\nxqLcsHYJXT19/PwF3W9BRGaeTA4fXQ08AODuW4FyMysNlzvc/Wp37w4TRBlQP1KfqeLys2spL8nn\nqd/sp7W9K9vhiIhMqEwOH9UAm1OWm8K21v4GM/sscBfwNXffYWaj9hmsvDxOLJYz7iCrqkrG3Ocj\n15zOt37yOr96vZ7fv2n1uN97JOOJazIorrFRXGOjuMYmE3FN5slrkcEN7v4VM/sH4BEzezadPoM1\nN4//JjWjXftoOGtWVDCnOI+HntvBFWfXUBqf2LmF8caVaYprbBTX2CiusTnVuIZLKJkcPqoj+Cu/\nXy1wAMDMKszsCgB3PwY8Cqwbqc9UkhvL4X2XLKGru4/HXtSRSCIyc2QyKTwO3AJgZmuAOnfvT2u5\nwL1mVhwuXwT4KH2mlPXn1lJWnMcvN++nrUNzCyIyM2QsKbj7RmCzmW0kOIroTjO7zcw+4O4NwF8A\nT5nZr4GDwIND9clUfKcqN5bD+y5ewvHuXh7ftDfb4YiITIiMzim4+2cHNb2asu5e4N40+kxZ68+t\n5ZHnd/Pk5n1cd9Fiigtzsx2SiMgp0RnNpyAvN4frL17M8a5eHt+kuQURmf6UFE7R+vMWUFqUx5Mv\n7ePose5shyMickqUFE5Rfm4O771oMZ1dmlsQkelPSWECXHXeAkriufxi817aO1UtiMj0paQwAfLz\ncnjvxYs5dryXJ1QtiMg0pqQwQd593kKKC3N54qV9dKhaEJFpSklhguTnBUciHTvewxMv7ct2OCIi\n46KkMIGuWrMgqBY27aWjsyfb4YiIjJmSwgQqyItx3UWL6Djew5ObNbcgItOPksIEe/eahRQVxHhi\n016OHVe1ICLTi5LCBCvMj3HdRYtp7+zhyc2aWxCR6UVJIQOuPj+oFh5/cY+qBRGZVpQUMqAwP8Z7\nLlxEe2cPv1C1ICLTiJJChlx9/iLi+TEeU7UgItOIkkKGxAtOVAu//I2qBRGZHpQUMuiaCxaG1cJe\nOrtULYjI1KekkEHxglyuvXARR49189Rv9mc7HBGRUSkpZNi1FyykMD/Gz1/cw/Gu3myHIyIyIiWF\nDIsX5HLtBQtp6+jmqZdVLYjI1KakMAmuvXARhfk5/PyF3RzvVrUgIlOXksIkKCrI5erzF9Ha0c3T\nqhZEZApTUpgk77lwEQV5OTz6vKoFEZm6Ypl8cTO7B7gESAB3ufumlHVXAV8GegEHbgeuAH4EbAk3\ne93dP5PJGCdLcWEuV5+/kId/vZsNL+/nPRctznZIIiInyVilYGbrgZXuvhb4BPD1QZt8G7jF3dcB\nJcB7w/YN7n5l+G9GJIR+1120mPy8HB59YQ9dqhZEZArK5PDR1cADAO6+FSg3s9KU9ee7e/+pvk1A\nZQZjmRKKC3O5es1CWtq72PBKXbbDERE5SSSRSGTkhc3s28DD7v7TcPkZ4BPu/tag7eYDzwAXA+8C\nvgm8A1QAd7v7EyO9T09PbyIWy8nAJ8iMlqPHuf1LTxAviPEvn7uWvNzpE7uIzCiRoRozOqcwWgBm\nVg38DPgjdz9kZm8DdwM/BJYDT5nZae7eNdyLNjd3jDugqqoSmpraxt1/vK5as4BHn9/Dj590rrlg\n0ZSJazSKa2wU19gorrE51biqqkqGbM/k8FEdUJOyXAsc6F8Ih5IeBT7v7o8DuPt+d7/P3RPuvh2o\nBxZkMMasuO6ixeTlRnnk+d1092huQUSmjkwmhceBWwDMbA1Q5+6pae3vgHvc/ef9DWZ2q5n9l/B5\nDTAPmHEH9pfG83j3moUcOdrFr149MHoHEZFJkrHhI3ffaGabzWwj0AfcaWa3AS3AY8DHgJVmdnvY\n5QfAfwA/MLObgTzgUyMNHU1n771oMb/8zT4eeX43V5xTS25Mp4yISPZldE7B3T87qOnVlOf5w3S7\nKUPhTCmlRXlcdd4CHntxL8++VsdVaxZmOyQREZ3RnE3vvXgJebEoDz+/m+6evmyHIyKipJBNZUV5\nXHneAg63HufZ1zW3ICLZp6SQZddfvJjcWJRHfr2Lnl5VCyKSXUoKWVZWnM/6c2s5pGpBRKYAJYUp\n4H2XLCE3FuXhjbtVLYhIVikpTAFzivNZf04th1o72fhGfbbDEZFZTElhirj+kiXEcqI8tFFzCyKS\nPUoKU0R5SVAtHGzp5KmX9mY7HBGZpZQUppDrL1lMLCfCD3/xlqoFEckKJYUppKK0gMvPqaX+UAfP\nb2nIdjgiMgspKUwxN1yyhFhOhIc27qK3T9WCiEwuJYUppqK0gGsvWkLjkWOqFkRk0ikpTEG3XL2S\nnGiEn6laEJFJpqQwBVWXx7n87Pk0Nh/jhTdVLYjI5FFSmKLet3ZJWC3spq8vM/fRFhEZTElhippb\nVsi6d82n4XAHL2xVtSAik0NJYQq7sb9aeG6XqgURmRRKClPY3DmFXHpWDfWHO3hxm6oFEck8JYUp\n7oZLlxKNqFoQkcmhpDDFVYfVwoFDHbzkjdkOR0RmOCWFaeDGS5cQjUR48Lld9CVULYhI5igpTAPV\n5XHWrp5H3cF2XtqmakFEMieWyRc3s3uAS4AEcJe7b0pZdxXwZaAXcOB2d+8bqc9sduOlS9m4pZ6f\nPbeLC86oJhqJZDskEZmBMlYpmNl6YKW7rwU+AXx90CbfBm5x93VACfDeNPrMWvMq4qxdXcP+g+38\nxpuyHY6IzFBjTgpmlm9mi9LY9GrgAQB33wqUm1lpyvrz3X1f+LwJqEyjz6x206VLiUTgwed2am5B\nRDIireEjM/tvwFHgX4GXgDYze9zdvzBCtxpgc8pyU9jWCuDureFrzwfeA3yBYDhp2D5DKS+PE4vl\npPMxhlRVVTLuvpk0VFxVVSWsX7OQpzfvY3v9US49u3ZKxDUVKK6xUVxjM5viSndO4SZgHfAx4Gfu\n/l/N7JdjfK+TBsHNrBr4GfBH7n7IzEbtM1hzc8cYwzihqqqEpqa2cffPlJHiunbNAjb8Zh/fe3Qr\nK2qKJ3VuYTrur2xSXGOjuMbmVOMaLqGkO3zU7e4J4HrC4R1gtD/P6wj+yu9XCxzoXwiHhR4FPu/u\nj6fTR2B+ZREXr5rH3sajvPL2wWyHIyIzTLpJ4YiZPQyscvdfm9mNwGgX+n8cuAXAzNYAde6emtb+\nDrjH3X8+hj5CcCRShGBuIaG5BRGZQOkOH30UuBZ4LlzuBH5vpA7uvtHMNpvZRoIEcqeZ3Qa0AI8R\nDEWtNLPbwy4/cPdvD+4zpk8zS9TOLeLCVdW8uLWRV945yHkrq7IdkojMEOkmhSqgyd2bzOyTBOcR\nfHW0Tu7+2UFNr6Y8z0+zjwzhpnXL2LS1kQef3cW5p80lovMWRGQCpDt89O9Al5mdB9wO/BidQ5BV\nC+YWccEZ1exuaOPVdw5lOxwRmSHSTQqJ8MziDwD/5O6PkMaRQZJZN61bCsBPNbcgIhMk3aRQbGYX\nEkwC/9zM8oHyzIUl6VhYVRxUC/VtvLZd1YKInLp0k8LfAf8CfMvdm4AvAj/IVFCSvvdfuhTQkUgi\nMjHSmmh29/uA+8yswszKgc+F5y1Ili2sLuZ8q2KzN/H6jsOcvaIy2yGJyDSWVqVgZuvMbDuwDXgb\n2GpmF2Q0Mknb+9ctA1QtiMipS3f46MvAze5e7e5zgd8B/j5zYclYLKouZs3pVeyoa2XLzsPZDkdE\nprF0k0Kvu7/Rv+DuLwM9mQlJxuP9OhJJRCZAuiev9ZnZh4AnwuX3EtwcR6aIxfNKOG/lXF5++yBv\n7mpm9bKKbIckItNQupXCHwKfBHYBOwkucfEHGYpJxql/bkHVgoiM14iVgpk9Q3BbTAhOVtsSPi8F\n7gWuyFhkMmZLako497S5vPLOQbbububMpaoWRGRsRhs++vykRCET5v2XLeWVdw7y02d3smpJua6J\nJCJjMmJScPcNkxWITIylNaWcs6KSV7cfYtvuZlapWhCRMRjzPZpl6nv/Zf1zC7uyG4iITDtKCjPQ\nsvmlnL2ikrf2HmHb7uZshyMi04iSwgyVepaziEi6lBRmqOW1pZy1vIJte47ge1QtiEh6lBRmsJuT\n1cKu7AYiItOGksIMtmJBGauXVbB1dzNv7T2S7XBEZBpQUpjhbtbcgoiMgZLCDHfawjLOXFrOm7ua\neXufqgURGZmSwiyQPBLpWVULIjKydK+SOi5mdg9wCcH1k+5y900p6wqAbwGr3f2CsO1K4EecuMbS\n6+7+mUzGOBucvmgOq5aUs2VXM+/sb+G0BWXZDklEpqiMVQpmth5Y6e5rgU8AXx+0yd8CrwzRdYO7\nXxn+U0KYIDdfpmpBREaXyeGjq4EHANx9K1BuZqUp6z8H/CSD7y8pTl80hzMWz+GNnYfZXteS7XBE\nZIrK5PBRDbA5ZbkpbGsFcPc2MxvqLvNnmtmDQAVwt7s/McQ2SeXlcWKxnHEHWVVVMu6+mZSJuD52\n42o+983n+PmLe/niJxeO6zVm0/6aCIprbBTX2GQirozOKQySzjWc3wbuBn4ILAeeMrPT3L1ruA7N\nzR3jDqiqqoSmprZx98+UTMVVU5qPLZrD5m2NvPDqfpbXlo7eaRLiOlWKa2wU19jM1LiGSyiZHD6q\nI6gM+tUCB0bq4O773f0+d0+4+3agHliQwRhnnf4rqOq8BREZSiaTwuPALQBmtgaoc/cR05qZ3Wpm\n/yV8XgPMA/ZnMMZZ54zFczh9YRmvbT/EzgOt2Q5HRKaYjCUFd98IbDazjQRHHt1pZreZ2QcAzOxH\nwP8JntrTZvZR4EFgfXgb0J8Cnxpp6EjGLhKJJKuFn+maSCIySEbnFNz9s4OaXk1Z9+Fhut2UuYgE\nYNWSck5bWMYr7xxkd30bS2qm5iSaiEw+ndE8C0UikeR5Cz/VeQsikkJJYZY6c0k5py04US2IiICS\nwqwVzC0sBXQkkoicoKQwi61eWsGK2lJefvsgexpULYiIksKspiORRGQwJYVZ7qxlFSybX8rmt5rY\n23g02+GISJYpKcxywZFISwH4meYWRGY9JQXhXcsrWVpTwkvexL4mVQsis5mSgmhuQUSSlBQEgHNW\nVLKkpoSXtjWy/2B7tsMRkSxRUhAgnFtYt4wEmlsQmc2UFCTpnNMqWTKvhE1bVS2IzFZKCpIUiUR4\n/7qlJICHNu7KdjgikgVKCjLAuSvnsri6mBffbODAIVULIrONkoIMEIlEuKl/bkHVgsiso6QgJznv\n9LksrCrmhTcbqD88/ntgi8j0o6QgJ4n2zy0kdN6CyGyjpCBDWmNVLKgq4vk362lQtSAyaygpyJCC\namEZiYSORBKZTZQUZFjnWxUL5hbx6y0NNDarWhCZDZQUZFjRSISb1i2lL5HgoY27sx2OiEwCJQUZ\n0QVnVFM7t4iNb9RzQGc5i8x4sUy+uJndA1wCJIC73H1TyroC4FvAane/IJ0+MvmikQg3XbqUbz24\nhe8+8iYfunwZZcX52Q5LRDIkY0nBzNYDK919rZmtAv4NWJuyyd8CrwCrx9BHsuDCM6r52cZdPPtq\nHc++WkdlaQErFpSyvLaMFbWlLJ5XQm5MRafITJDJSuFq4AEAd99qZuVmVurureH6zwGVwK1j6CNZ\nEI1G+NPfOpeXdxzi9bea2F7XyotbG3lxayMAsZwIi+eVsLy2lBVhoqgsKyASiWQ5chEZq0wmhRpg\nc8pyU9jWCuDubWZWOZY+QykvjxOL5Yw7yKqqknH3zaSpFldVVQmnL5/Lb11jJBIJ6g914LsP47ub\n2banmZ37W9hR18qT7ANgTkk+tricM5ZWYEvKWblwDgX5mftxm2r7q5/iGhvFNTaZiCujcwqDjOfP\nxlH7NJ/CoZJVVSU0NbWNu3+mTIe4YsDqxXNYvXgOsIyu7l52N7SxfX8rO+pa2F7Xygtb6nlhSz0A\nkQgsqipm+YKgklheW8q8ijjRCagmpsP+mkoU19jM1LiGSyiZTAp1BH/l96sFDmSgj0wBebk5rFw4\nh5UL5yTbDrd2sqOule1hkthd38aexqM8/fJ+AIoKYixLGXJaVltKUUFutj6CiJDZpPA4cDfwLTNb\nA9S5+2hpbTx9ZIqqKC2gorSAC86oBqCnt4+9jUeTiWLH/lbe2HGYN3YcTvaZXxlPzk0sry1lYVUx\n0ajmJkQmS8aSgrtvNLPNZrYR6APuNLPbgBZ3/4mZ/QhYBJiZPQ18291/MLhPpuKTyRfLibJsfinL\n5pdy9fkLAWjt6GJHXTjktL+VHQdaOfB6Pc+9Hgw75efmsGx+CSsWBElieW0ZZUV52fwYIjNaRucU\n3P2zg5peTVn34TT7yAxWGs/j3NPmcu5pcwHo60tQd6idHXWtvBNOXm/bc4Rte44k+8wtK0gmiRW1\nZcwpL8pW+CIzzmRONIuMKhqNsLCqmIVVxVxxTi0AHZ3d7DzQFgw51bWyfX8LL7zZwAtvNgCQG4uy\neF5xMDcRTmSXl+TrkFiRcVBSkCkvXpDL6mUVrF5WAUAikaCx+VhyAntP41F27m9l+/5W2LQXgDnF\necG8xIKgmlhSU0J+7vgPXRaZLZQUZNqJRCLMq4gzryLOpWfNp6qqhH11R9hd35acwH6nroXNbzWx\n+a0mILhcx6Lq4jBJBImiurxQ1YTIIEoKMiPk5+Zw+qI5nL4oOCQ2kUhwuPX4iSGnuhZ217exu6GN\np34THBJbXJgbTl4HSWLZ/FLiBfqVkNlNvwEyI0UiESrLCqgsK+CiVfMA6O4JDolNnZt4bfshXtt+\nKOgDzJ9bFFQS4UR2bWWRDomVWUVJQWaN3Fg0WRn0a2nvYkdKkth5oI26g+0881pwzmRBXg7L5pcm\nLwC4vLaU0rgOiZWZS0lBZrWyojzOW1nFeSurAOjt66PuYEcwiR0eErt1dzNbdzcn+1TPKUxOYC+v\nLWVRdTGxHF0lVmYGJQWRFDnRKIuqi1lUXcyV5y4AoL2zm511rWxPORP7+S0NPL/lxCGxS2tKkkli\nxYIyykt0zwmZnpQUREZRVJDLWcsrOWt5cFHfvkSChsMd4QR2Kzv2t/DO/hbe3teS7FNeks/pi8sp\nKYiFl/vIpzK87EdZUZ7mKWTKUlIQGaNoJML8yiLmVxax7l3zAejs6gkPiQ3mJvqvEjuUnGiE8pJ8\nKkryqSgrCJJFST4VpQXJxKGjoCRb9JMnMgEK8mLY4nJscTkQHBKbH8/n7Z2HONTayaHWTppbj3Oo\ntZPDrZ0cbjvO2/taSKRUFwNfL4fK0gLKUyqMipLweVkB5cX5utudZISSgkgGRCIRyorzWVJTwpKa\noa9b39Pbx5G24xxuS0kWrQOf7z/YPux7lBXlnTQ0VVGST2VZ8Lwknjsh96uQ2UVJQSRLYjlR5s4p\nZO6cwmG3OXa8J1lZpCaLw2FtwE+OAAAQAElEQVT1sbexjZ0Hhr4xYSwnQkVJkDT6L2NeWZrPskXl\n5CQSVJTkU5jBu+HJ9KSfCJEprDA/xoKqYhZUFQ+5vi+RoK29i0NhogiSxXEOt51IIKlXmB0snh9L\nJouTqo7SfOYU5+tw21lGSUFkGouGw1RlxfkDTspL1d3TR3PbiaGp470J9ta3hpXHcZpajrGv6eiQ\nfSMRmFOcfyJZlJycOIoLc3UNqRlESUFkhsuNRakuj1NdHgdOvrdvIpHg2PGeoauNluD5rgPB/beH\nkheLUp5abaRMiPcfVaUr1E4fSgois1wkEiFekEu8IJdF1cMMU/UlaGnvGnZC/HBbJw2HO4Z9j+LC\n3JMqjNTJ8TnFOtlvqlBSEJFRRcNzK8pL8mFB2ZDbHO/upbntxCT44UGVR/3hDvY0DD1MFY1EqCgN\nJr6LCmIUFeQSL4hRVJg77HJRQYyC/JiOsJpgszop9CX6sh2CyIyRn5tDTUWcmor4kOsTiQTtnT0c\naukMJ8IHVhutHV0cbDnG3sbetN8zEgkmy0dKHvGCGMUD2oN1eRrSGtKsTQr3bvk/vNz0GtWFc5lX\nVE1NvIqaeDXziqqZF68iL0dXwhSZSJFIhOLCXIoLc4c8d6N/rqO3r4+Ozh46Ons42tlNR2cP7ce6\nae/soX2Y5aOd3RxuPU5Pb/p/6MVyohQVnqg6+h/jBbkD2ufP66Cnq/tEwimIkROduUdkzdqkcHr5\nCg53HWJf6wHq2gdejiBChIqCOcyLV1MTJomaonnUxKspztNN4kUyKScapSSeR0k8j3lj7NvV3Tts\n8hiuvbW9iwOH2kkk0n+fgrycE8mksD9ZDFxOVicp7QV5OVP+SK1ZmxQurb2Qm895N42NrRw53kJ9\nRyMN7U3UdzRS395AQ0cTbx523jzsA/oV5capSSaLE48VBXOIRmbuXw8i00Febg55uTljvkptXyJB\n5/HeAZVHf/IgJ0rjoXY6OrtpP3YiwXR0dtNw5BjHG4eeJxlKNBJJVhsjJY/By0UFMXJjkzPcldGk\nYGb3AJcACeAud9+Usu4a4K+AXuARd/9LM7sS+BGwJdzsdXf/TCZjjEQilBfMobxgDqsqTh+wrqO7\ng/qOJhraG8Nk0UhDRyM7WnazvWXXgG1zo7nMi1eFVUU1NUXzmBevojpeRW501uZekWmh/8t6qAsR\nDj6Ed7Ce3mC4KzVZpCaP9nC5o3Ng1XKwpZPevvTLk7xY9MS8SH6MhTWlfOjyZRN+VnrGvq3MbD2w\n0t3Xmtkq4N+AtSmbfB24DtgPbDCzH4ftG9z9lkzFNRbx3DjLy5awvGzJgPbu3m6ajh0aUFUECaOJ\nfUfrBmwbIUJlYcVJ1UVNvIp47tATciIyfcRyopQW5VFaNLZ5yEQiQVd334BkcnSI5BEkmRNtR9qO\nU9fUzq76Nq49f8H0SQrA1cADAO6+1czKzazU3VvNbDlw2N33ApjZI+H2r2cwngmTm5NLbXENtcU1\nA9r7En00dx4Jh6JSq4sm3ji0lTcObR2wfUlecXJyuyaZLKqZmxj6WHERmTkikQj5eTnk5+VQMfTJ\n6MPq60tQUVnMkebhL5g4XplMCjXA5pTlprCtNXxsSlnXCKwgSApnmtmDQAVwt7s/MdKblJfHiZ3C\nWFtV1dBXsByveZRxBktOam89fpS61nr2tdazv7Weurbg+TtHdvL2kR0Dti2I5bOgpIba0nksKK1J\n/qspriYWze5hdBO9vyaK4hobxTU2symuyRzsHmnKvX/d28DdwA+B5cBTZnaau3cN17G5efizKEcz\n2ljhRKtkHpWl8zin9JxkW1dvFw0dB2lob6C+I5joPnj8ILuP7GN78+4B/aORKFWFlSdVF9XxKgpj\nBRmPf7L3V7oU19gorrGZqXENl1AymRTqCCqCfrXAgWHWLQDq3H0/cF/Ytt3M6sN1OzMYZ1bl5eSx\nqKSWRSW1ybaqqhIaGls4eOwwDeEQVOrRUQ0dTXBwy4DXmZNfFiaL8HyLMGGU5pVM+UPgRGTqyGRS\neJzgr/5vmdkagi/9NgB332VmpWa2FNgH3Ajcama3AvPd/atmVgPMI5iInnWikSjV8blUx+fyrrln\nJtsTiQStXUcHJYvgcVvz22xrfnvA6xTGCoIEMeCci2oqCyrIyfJQlIhMPRlLCu6+0cw2m9lGoA+4\n08xuA1rc/SfAp4D/CDe/z93fMrMDwA/M7GYgD/jUSENHs1EkEqEsv4Sy/BJOL18xYF1nT+eAI6H6\nJ7r3tO1jV+ueAdvGIjlUxecOORSVr7O5RWatSGIsp/FNQU1NbeP+ADN1rHCw3r5emo4dGmIoqoHj\nvSfn3IqC8hPnW6QMRS1fMD8ZVyKRoC/RR4IEfSnPE4k++kiE6xMk6AvWDVge2DfByesTifCR1G0H\ntYXbFZcU0NLaMWC7wXGd2D4Rxtc3KIYT2/QlYwie97/Pic/VN2j7gX3736cgP5dYIo/i3CKKcuPJ\nx6LcIopyiygOn+fl5E7Y/3U6ZsvP/USZqXFVVZUMOa6ss6pmgZxoTnhCXTXnVJ1oTyQSHDneklJd\nnEgaWw+/xdbDb530Ov1fijJxcqO5QySO/uQRH3Jdfk6+5ookI5QUZrHUs7nPqFg5YF1H97EgSSTP\n6G6gM9FJb0+CCBGikQiRSJQoEaKRKJFI2EY0fDzRnmyLRIgmnw/aLvV1CNb3b9//2gO2S8YQobQk\nTvvRrpS+wTYDthvw+gPjjUQGvubQ8fS3n/gs/c8HxhNNbl9eGWfPgUaOdrfT3t1Be3c7Rwc9prY3\nHTt40smPw4lFcgYkjtTKo3hQe39CKYwVKJHIqJQUZEjx3EKWlS1hWcrZ3DO1jM6Uglh+Mummq7uv\n56RkcWK5I5lg+h+bjx856YKOw4lGohTF4pQWFlMQKQyTxxAVSV4RRbGgPZ5bqGt6zTJKCiJTSG40\nxpz8MubkD30jm6H09vXS0XNsUMJop72rg6M9JxJM/7rWzjbquhpIMPp0XIQI8VghRXlximJFFIeP\nA4azkknkRELRkW0TJxHOefWG82D9/0p6MnNAiJKCyDSXE82hJK+Ykrz0Lo/Sfx5MR88x2rvaae8J\nq5Dw+dGuQZVKTwftXe0cPHY47fmkwlhBkCjywsSRklCK84YY5oqduA5YImXivy/Rd9KX4VBtwXJv\n8uCCvkQvveHkf2+id8B2icH9GbTcN7Ct4ECMo+2dI8TS//rjeO+hXnNwPMPs81g0xn+/6I+pjlcN\nuX68lBREZqFoJEpxbhHFuenfHySRSHCsp3NgNTKoCkl9bO/uYH9bHT2J9O6kFovGRvwSnK7656ei\nkSg54WPqv5xIlFg096S2yIDtc8K2SNiWQ1XJHErzxnjRpDQoKYhIWiKRCPHcQuK5hVRRmVafRCLB\n8d6ukxLHwOfBY3eki74ewi/BCDmRnBG+JKOjftH2f5n2Hwxw4vUiA75oh30tgsfKimJaWzqTyye2\nyxnmtSIDvsQzNSeTqbk0JQURyZhIJEJBLJ+CWD6VhRUjbjtVDxiomltCU2LqxZUpOqxARESSlBRE\nRCRJSUFERJKUFEREJElJQUREkpQUREQkSUlBRESSlBRERCRp2t9kR0REJo4qBRERSVJSEBGRJCUF\nERFJUlIQEZEkJQUREUlSUhARkSQlBRERSZoVN9kxs3uAS4AEcJe7b0pZdw3wV0Av8Ii7/+UUiWsX\nsDeMC+BWd98/ibGdBfwUuMfd/2nQumzus5Hi2kWW9pmZ/Q1wOcHv1Jfd/f6UddncXyPFtYss7C8z\niwP3AvOAAuAv3f2hlPVZ2V9pxLWL7P5OFgJvhHHdm9I+oftrxicFM1sPrHT3tWa2Cvg3YG3KJl8H\nrgP2AxvM7Mfu/uYUiAvgenc/mulYhoitCPhH4BfDbJKtfTZaXJCFfWZmVwFnhf+XlcDLwP0pm2Rr\nf40WF2TnZ+wm4CV3/xszWwI8ATyUsj4r+yuNuCBLv5OhzwOHh2if0P01G4aPrgYeAHD3rUC5mZUC\nmNly4LC773X3PuCRcPusxjUFHAfeB9QNXpHlfTZsXFn2K+DD4fMjQJGZ5UDW99ewcWWTu9/n7n8T\nLi4C9vWvy+b+GimubDOzM4AzgYcHtU/4/prxlQJQA2xOWW4K21rDx6aUdY3AiikQV7//aWZLgWeB\n/+buk3JNEnfvAXrMbKjVWdtno8TVb9L3mbv3Au3h4icISvj+IYZs7q+R4uqXlZ8xADPbCCwEbkxp\nzubv5Ehx9cvW/vo74NPA7w1qn/D9NRsqhcEi41yXaYPf+8+BPwGuBM4CPjTZAaUpm/tssKzuMzO7\nmeDL99MjbDbp+2uEuLK6v9z9UuD9wPfMbLj9Mun7a4S4srK/zOxjwK/dfWcam5/y/poNSaGOIJv2\nqwUODLNuAZM3NDFSXLj7d929Mfzr+BHgXZMU12iyuc9GlM19ZmbXAf+dYMy5JWVVVvfXCHFlbX+Z\n2flmtiiM4RWCEYuqcHXW9tcocWXz5+sG4GYzex64HfhCOLkMGdhfsyEpPA7cAmBma4A6d28DcPdd\nQKmZLTWzGEG5+Hi24zKzMjN7zMzywm3XExx1kHVZ3mfDyuY+M7My4G+BG919wERgNvfXSHFl+Wfs\nCuBPwzjmAcXAQcj6z9ewcWVzf7n7b7n7he5+CfAdgqOPngzX7WKC99esuHS2mX2F4D+8D7gTOA9o\ncfefmNkVwF+Hm/7Y3b86ReK6i2D88BjBUSOfmazxSzM7n2AMcynQTXBUw4PAzmzuszTiyso+M7M7\ngC8Cb6U0/xJ4Pcv7a7S4srW/CoF/JZjMLQTuBirJ8u9kGnFl7XcyJcYvArvCxYzsr1mRFEREJD2z\nYfhIRETSpKQgIiJJSgoiIpKkpCAiIklKCiIikqSkIJJFZnabmX0v23GI9FNSEBGRJJ2nIJIGM/sM\n8BGCSx9sA/6G4LLKjwLnhJv9trvvN7MbCK6T0xH+uyNsvxj4GtBFcAnkjxFcP+eDBBdCPBPYDXxw\nsk+KEumnSkFkFGZ2EfAB4Ap3X0twGeprgOXAv7v75cDTwJ+GN2r5DvAhd7+KIGn8j/Clvgd80t3X\nAxsIrmkDsBq4Azif4EJraybjc4kMZTZcOlvkVF0JnAY8FV62u4jgwmOH3L3/8ufPAf8ZOB1ocPf+\na/E/Dfyhmc0F5rj7GwDu/jUI5hSATe7eES7vB+Zk/iOJDE1JQWR0x4EH3T156enwmvq/SdkmQnBb\n1cHDPqntw1XmPUP0EckKDR+JjO454HozKwYwsz8C5hPcLe+8cJvLgNcILj5XbWaLw/ZrgOfd/RBw\n0MwuDF/jT8PXEZlSlBRERuHuLwHfAJ42s2cJhpNaCK7SepuZ/RJYB9zj7scIbmhzn5k9TXBrxM+H\nL/W7wD+Y2QaCq+PqUFSZcnT0kcg49N+S0d0XZjsWkYmkSkFERJJUKYiISJIqBRERSVJSEBGRJCUF\nERFJUlIQEZEkJQUREUn6/wGOH/jbXBquqQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f4437842fd0>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "3UmhBPfHoxs6",
        "colab_type": "code",
        "outputId": "21871686-5177-4f55-d67c-829b199f7d1d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 274
        }
      },
      "cell_type": "code",
      "source": [
        "# initialize lists to monitor test loss and accuracy\n",
        "test_loss = 0.0\n",
        "class_correct = list(0. for i in range(10))\n",
        "class_total = list(0. for i in range(10))\n",
        "\n",
        "model.eval() # prep model for *evaluation*\n",
        "\n",
        "for data, target in test_loader:\n",
        "  \n",
        "    data = data.to(device)\n",
        "    target = target.to(device)  \n",
        "  \n",
        "    # forward pass: compute predicted outputs by passing inputs to the model\n",
        "    output = model(data)\n",
        "    # calculate the loss\n",
        "    loss = criterion(output, target)\n",
        "    # update test loss \n",
        "    test_loss += loss.item()*data.size(0)\n",
        "    # convert output probabilities to predicted class\n",
        "    _, pred = torch.max(output, 1)\n",
        "    # compare predictions to true label\n",
        "    correct = np.squeeze(pred.eq(target.data.view_as(pred)))\n",
        "    # calculate test accuracy for each object class\n",
        "    for i in range(len(data)):\n",
        "        label = target.data[i]\n",
        "        class_correct[label] += correct[i].item()\n",
        "        class_total[label] += 1\n",
        "\n",
        "# calculate and print avg test loss\n",
        "test_loss = test_loss/len(test_loader.dataset)\n",
        "print('Test Loss: {:.6f}\\n'.format(test_loss))\n",
        "\n",
        "for i in range(10):\n",
        "    if class_total[i] > 0:\n",
        "        print('Test Accuracy of %5s: %2d%% (%2d/%2d)' % (\n",
        "            str(i), 100 * class_correct[i] / class_total[i],\n",
        "            np.sum(class_correct[i]), np.sum(class_total[i])))\n",
        "    else:\n",
        "        print('Test Accuracy of %5s: N/A (no training examples)' % (classes[i]))\n",
        "\n",
        "print('\\nTest Accuracy (Overall): %2d%% (%2d/%2d)' % (\n",
        "    100. * np.sum(class_correct) / np.sum(class_total),\n",
        "    np.sum(class_correct), np.sum(class_total)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test Loss: 0.026044\n",
            "\n",
            "Test Accuracy of     0: 99% (782/785)\n",
            "Test Accuracy of     1: 99% (881/883)\n",
            "Test Accuracy of     2: 98% (825/835)\n",
            "Test Accuracy of     3: 99% (820/824)\n",
            "Test Accuracy of     4: 98% (784/793)\n",
            "Test Accuracy of     5: 98% (714/727)\n",
            "Test Accuracy of     6: 98% (732/740)\n",
            "Test Accuracy of     7: 99% (812/820)\n",
            "Test Accuracy of     8: 98% (774/785)\n",
            "Test Accuracy of     9: 97% (787/808)\n",
            "\n",
            "Test Accuracy (Overall): 98% (7911/8000)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "i2ZEW93toxtL",
        "colab_type": "code",
        "outputId": "85b664a0-d41b-404e-d5f7-cf66afc05f90",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        }
      },
      "cell_type": "code",
      "source": [
        "# obtain one batch of test images\n",
        "dataiter = iter(test_loader)\n",
        "images, labels = dataiter.next()\n",
        "\n",
        "images = images.to(device)\n",
        "labels = labels.to(device)\n",
        "\n",
        "# get sample outputs\n",
        "output = model(images)\n",
        "# convert output probabilities to predicted class\n",
        "_, preds = torch.max(output, 1)\n",
        "# prep images for display\n",
        "images = images.cpu().numpy()\n",
        "\n",
        "# plot the images in the batch, along with predicted and true labels\n",
        "fig = plt.figure(figsize=(25, 4))\n",
        "for idx in np.arange(20):\n",
        "    ax = fig.add_subplot(2, 20/2, idx+1, xticks=[], yticks=[])\n",
        "    ax.imshow(np.squeeze(images[idx]), cmap='gray')\n",
        "    ax.set_title(\"{} ({})\".format(str(preds[idx].item()), str(labels[idx].item())),\n",
        "                 color=(\"green\" if preds[idx]==labels[idx] else \"red\"))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABW4AAAD2CAYAAABGOR6XAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nmb8BBdUSI2IGR8oAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f44378c1eb8>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "metadata": {
        "id": "QbxfMw5_p_lK",
        "colab_type": "code",
        "outputId": "db0f280b-e551-4fc9-b665-916c083614e1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        }
      },
      "cell_type": "code",
      "source": [
        "import random\n",
        "\n",
        "random_idx = random.randint(0, len(test_loader.dataset))\n",
        "# get sample outputs\n",
        "image = test_loader.dataset[random_idx][0].to(device)\n",
        "\n",
        "output = model(image[None, :])\n",
        "# convert output probabilities to predicted class\n",
        "_, preds = torch.max(output, 1)\n",
        "label = test_loader.dataset[random_idx][1]\n",
        "label = label.to(device)\n",
        "\n",
        "# plot the images one example, along with predicted and true labels and probabilities\n",
        "fig = plt.figure(figsize=(25, 4))\n",
        "ax = fig.add_subplot(1, 1, 1, xticks=[], yticks=[])\n",
        "ax.imshow(test_loader.dataset[random_idx][0].detach().squeeze(), cmap='gray')\n",
        "ax.set_title(\"{} ({})\".format(str(preds.cpu().numpy()[0]), str(label.cpu().numpy())),\n",
        "             color=(\"green\" if preds==label else \"red\"))\n",
        "ax = fig.add_subplot(1, 2, 2, xticks=[], yticks=[])\n",
        "x_labels = torch.arange(10)\n",
        "ax.bar(torch.arange(10), height=torch.exp(output).cpu().detach().squeeze(), tick_label=x_labels)\n",
        "ax.set_title(\"{:.2f} % accuracy\".format(float(torch.exp(output[:, preds])*100)))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5,1,'100.00 % accuracy')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAzQAAAEHCAYAAAB4AwNdAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQLMEGgAAoFkCDQAA0Kz/B4iVLAOhEouEAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<matplotlib.figure.Figure at 0x7f4434a9b9e8>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}