Created
May 10, 2022 05:09
-
-
Save XinyueZ/0d7f17fa5aa0329913c674170b084db4 to your computer and use it in GitHub Desktop.
HOG_SVM_Paper_Scissors_Rock.ipynb
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| { | |
| "cells": [ | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "view-in-github", | |
| "colab_type": "text" | |
| }, | |
| "source": [ | |
| "<a href=\"https://colab.research.google.com/gist/XinyueZ/0d7f17fa5aa0329913c674170b084db4/hog_svm_paper_scissors_rock.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "shOHlg0bQ8s2" | |
| }, | |
| "source": [ | |
| "# Objectives\n", | |
| "\n", | |
| "### Histogram of Oriented Gradients by using OpenCV\n", | |
| "\n", | |
| "Deep learning is a part of machine learning, but she is not the whole of machine learning. Image classification can be solved very well by convolutional layers (FCN: Tensorflow, Pytorch are excellent, and of course, you can write them by hand 🤣), and as an applied ML developer, I can start with a simple approach when building a fast solution, and then choose what kind of model by model-selection. \n", | |
| "\n", | |
| " \n", | |
| "\n", | |
| "It turns out that the Histogram of Oriented Gradients(H.O.G.) combined with SVM was one of the ways of image-classification in the computer vision area was done before more advanced methods like Deep Learning became popular.\n", | |
| "\n", | |
| "### Architecture\n", | |
| "\n", | |
| "" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "X9aFSKo5Q8s7" | |
| }, | |
| "source": [ | |
| "## Import important libraries and define auxilary functions " | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "RdmQ3Va2Q8s8" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "import os\n", | |
| "import sys\n", | |
| "from IPython.display import clear_output" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "n5rYWERXQ8s9" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "import numpy as np\n", | |
| "import pandas as pd\n", | |
| "import seaborn as sns\n", | |
| "import matplotlib.pyplot as plt" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "D0B0Y-sdQ8s-" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "import cv2\n", | |
| "from skimage.feature import hog\n", | |
| "from sklearn.metrics import accuracy_score\n", | |
| "from sklearn.metrics import r2_score\n", | |
| "from sklearn.metrics import classification_report\n", | |
| "from sklearn.svm import SVC\n", | |
| "from sklearn.metrics import confusion_matrix\n", | |
| "from sklearn.model_selection import GridSearchCV" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "LllEUcY_Q8s_" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "from tensorflow.keras.utils import *" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "source": [ | |
| "np.random.seed(4)" | |
| ], | |
| "metadata": { | |
| "id": "N_5JWq9KTipP" | |
| }, | |
| "execution_count": null, | |
| "outputs": [] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "OAsLRf0wQ8s_" | |
| }, | |
| "source": [ | |
| "## Download images \"paper scissors and rock\"\n", | |
| "\n", | |
| "Origin: https://laurencemoroney.com/datasets.html" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "yce80ky7Q8tA" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def download_dataset(src, name):\n", | |
| " download_dir = os.getcwd()\n", | |
| "\n", | |
| " dataset_root = os.getcwd()\n", | |
| " dataset_subdir = \"my_datasets\"\n", | |
| "\n", | |
| " # Location of downloaded file will be \"$dataset_root/$name.zip\".\n", | |
| " # Location of all extracted data will be in \"$dataset_root/$dataset_subdir/$name\".\n", | |
| "\n", | |
| " download_file = os.path.join(download_dir, f\"{name}.zip\")\n", | |
| " dataset_dir = os.path.join(dataset_root, dataset_subdir, name)\n", | |
| "\n", | |
| " get_file(\n", | |
| " fname=download_file,\n", | |
| " cache_dir=dataset_root,\n", | |
| " cache_subdir=dataset_subdir,\n", | |
| " origin=src,\n", | |
| " extract=True)\n", | |
| "\n", | |
| " assert os.path.exists(download_file)\n", | |
| " assert os.path.exists(dataset_dir)\n", | |
| "\n", | |
| " return download_file, dataset_dir\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "JhC5HNWQQ8tB", | |
| "outputId": "579f7860-b916-4a97-b1b1-5aca86246652" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Downloading data from https://dl.dropbox.com/s/w19k944w35af101/train.zip\n", | |
| "201834496/201833815 [==============================] - 2s 0us/step\n", | |
| "201842688/201833815 [==============================] - 2s 0us/step\n", | |
| "/content/train.zip\n", | |
| "['scissors', 'rock', 'paper', '.DS_Store']\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "train_download_file, train_dir = download_dataset(\"https://dl.dropbox.com/s/w19k944w35af101/train.zip\", \"train\")\n", | |
| "print(train_download_file)\n", | |
| "print(os.listdir(train_dir))" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "ZKTeMRpCQ8tC", | |
| "outputId": "c49938db-cf25-49f8-d452-c73ec8678ddc" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Downloading data from https://dl.dropbox.com/s/tyha48tjw4gkiy8/cv.zip\n", | |
| "29687808/29683601 [==============================] - 0s 0us/step\n", | |
| "29696000/29683601 [==============================] - 0s 0us/step\n", | |
| "/content/valid.zip\n", | |
| "['scissors', 'rock', 'paper', '.DS_Store']\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "cv_download_file, cv_dir = download_dataset(\"https://dl.dropbox.com/s/tyha48tjw4gkiy8/cv.zip\", \"valid\")\n", | |
| "print(cv_download_file)\n", | |
| "print(os.listdir(cv_dir))" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "aN9CsPUEQ8tD", | |
| "outputId": "bc4582cf-0f82-48f1-9096-28c1e20fce8d" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Downloading data from https://dl.dropbox.com/s/1x0tt0j53tymn1b/test.zip\n", | |
| "6643712/6638511 [==============================] - 0s 0us/step\n", | |
| "6651904/6638511 [==============================] - 0s 0us/step\n", | |
| "/content/test.zip\n", | |
| "['rock5.png', 'scissors-hires2.png', 'rock2.png', 'paper-hires2.png', 'paper8.png', 'scissors2.png', 'paper2.png', 'scissors6.png', 'scissors7.png', 'scissors5.png', 'paper5.png', 'paper9.png', 'paper6.png', 'rock8.png', 'scissors4.png', 'rock4.png', 'paper1.png', 'rock6.png', 'paper3.png', 'paper-hires1.png', 'rock1.png', 'paper7.png', 'rock7.png', 'scissors-hires1.png', 'rock3.png', 'rock-hires2.png', 'rock9.png', 'scissors1.png', 'scissors3.png', 'scissors8.png', 'paper4.png', 'rock-hires1.png', 'scissors9.png']\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "test_download_file, test_dir = download_dataset(\"https://dl.dropbox.com/s/1x0tt0j53tymn1b/test.zip\", \"test\")\n", | |
| "print(test_download_file)\n", | |
| "print(os.listdir(test_dir))" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "yzlddOvFQ8tD" | |
| }, | |
| "source": [ | |
| "## Create datasets\n", | |
| "\n", | |
| "We will load and process every image. Let's go over some concepts:\n", | |
| "\n", | |
| "<ul>\n", | |
| " <ul>\n", | |
| " <li><code>cv2.resize()</code> to resize the image </li>\n", | |
| " <li><code>cv2.COLOR_BGR2GRAY()</code> will convert the images to greyscale image</li>\n", | |
| " <li><code>hog()</code> will extract the H.O.G. features from the image </li>\n", | |
| " </ul>\n", | |
| " \n", | |
| "</ul>\n", | |
| "\n", | |
| "Return\n", | |
| "\n", | |
| "feature_images: feature-wised image in visual\n", | |
| "\n", | |
| "images: origin images in grayscale\n", | |
| "\n", | |
| "X, the feature-wised image list\n", | |
| "\n", | |
| "y, the label/target of images\n", | |
| "\n", | |
| "<li>scissors: 0\n", | |
| "<li>paper: 1\n", | |
| "<li>rock:2\n", | |
| "\n", | |
| " \n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "### Augmentation helper\n", | |
| "\n", | |
| "We replicate for the same image with some different augmentation versions.\n", | |
| "\n", | |
| "Return\n", | |
| "- origin image\n", | |
| "- Image with augmentation: fliped, plus noise, rotated ...." | |
| ], | |
| "metadata": { | |
| "id": "p5gpufBBoaSy" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "source": [ | |
| "def aug_image(image):\n", | |
| " cols, rows = image.shape\n", | |
| "\n", | |
| " image_list = [image]\n", | |
| "\n", | |
| " rotated_45 = cv2.warpAffine(image, cv2.getRotationMatrix2D(center=(cols // 2 - 1, rows // 2 - 1), angle=45, scale=1), (cols, rows))\n", | |
| " image_list.append(rotated_45)\n", | |
| "\n", | |
| " rotated_neg_45 = cv2.warpAffine(image, cv2.getRotationMatrix2D(center=(cols // 2 - 1, rows // 2 - 1), angle=-45, scale=1), (cols, rows))\n", | |
| " image_list.append(rotated_neg_45)\n", | |
| "\n", | |
| " flip_top_down = cv2.flip(image,0)\n", | |
| " image_list.append(flip_top_down)\n", | |
| "\n", | |
| " #\n", | |
| " # When the training is slow, because the following augmentations bring too many replicates.\n", | |
| " # For good performance machine, we can comment out them.\n", | |
| " #\n", | |
| " #flip_left_right = cv2.flip(image,1 )\n", | |
| " #image_list.append(flip_left_right)\n", | |
| "\n", | |
| " #noise = np.random.normal(0, 20, (rows, cols)).astype(np.uint8)\n", | |
| " #image_list.append(image + noise)\n", | |
| "\n", | |
| " return image_list" | |
| ], | |
| "metadata": { | |
| "id": "1VE4o2FAoZ3u" | |
| }, | |
| "execution_count": null, | |
| "outputs": [] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "a3AeVw11Q8tF" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def load_images(path, pixels_per_cell=(8, 8)):\n", | |
| " X = []\n", | |
| " y = []\n", | |
| " feature_images = []\n", | |
| " images = []\n", | |
| " labels = ['scissors', 'paper', 'rock']\n", | |
| "\n", | |
| " for label_idx, label in enumerate(labels):\n", | |
| " path_label_dir = os.path.join(path, label)\n", | |
| " for image_path in os.listdir(path_label_dir):\n", | |
| " image_path = os.path.join(path_label_dir, image_path)\n", | |
| " #\n", | |
| " #read image\n", | |
| " image = cv2.imread(image_path, cv2.IMREAD_GRAYSCALE)\n", | |
| " image = cv2.resize(image, (64, 64))\n", | |
| "\n", | |
| " image_list = aug_image(image)\n", | |
| " for image in image_list:\n", | |
| " #\n", | |
| " # About pixels_per_cell:\n", | |
| " # But why 8×8 patch ? Why not 32×32 ? It is a design choice informed by the scale of features we are looking for.\n", | |
| " # HOG was used for pedestrian detection initially. 8×8 cells in a photo of a pedestrian scaled to 64×128 are big enough\n", | |
| " # to capture interesting features ( e.g. the face, the top of the head etc. ).\n", | |
| " features, hog_image = hog(\n", | |
| " image, visualize=True, block_norm='L2-Hys', pixels_per_cell=pixels_per_cell)\n", | |
| "\n", | |
| " X.append(features)\n", | |
| " y.append(label_idx)\n", | |
| " feature_images.append(hog_image)\n", | |
| " images.append(image)\n", | |
| "\n", | |
| " return np.array(X), np.array(y), np.array(feature_images), np.array(images)\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "3xm3m6j8Q8tG" | |
| }, | |
| "source": [ | |
| "#### Plot helper " | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "rLorDxn7Q8tG" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def plot_images(images, labels, names, nrows=1, ncols=5, figsize=(20, 10)):\n", | |
| " _, axes = plt.subplots(nrows=nrows, ncols=ncols, figsize=figsize)\n", | |
| " for ax, image, label in zip(axes, images, labels):\n", | |
| " ax.set_axis_off()\n", | |
| " ax.imshow(image, cmap=plt.cm.gray, interpolation='nearest')\n", | |
| " ax.set_title('{}: {}'.format(label, names[int(label)]))" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "mbiThfR6Q8tG" | |
| }, | |
| "source": [ | |
| "#### Training set" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "7FpqnrUJQ8tH" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "X_train, y_train, feature_images_train, images_train = load_images(train_dir)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "hqCMtx6qQ8tH" | |
| }, | |
| "source": [ | |
| "#### Some examples in train set" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "MtJWINPbQ8tH", | |
| "outputId": "010a4dec-7e47-4a68-f261-f84fd65708e5" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "execute_result", | |
| "data": { | |
| "text/plain": [ | |
| "[5835, 7072, 5167, 6036, 2842]" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "execution_count": 14 | |
| } | |
| ], | |
| "source": [ | |
| "shuffle_indices = list(np.random.permutation(len(feature_images_train))) \n", | |
| "shuffle_indices[:5]" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 461 | |
| }, | |
| "id": "kUxtB9qAQ8tI", | |
| "outputId": "b111a443-59f3-4dc7-9d31-634df98c5ccb" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 1440x720 with 5 Axes>" | |
| ] | |
| }, | |
| "metadata": { | |
| "needs_background": "light" | |
| } | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 1440x720 with 5 Axes>" | |
| ] | |
| }, | |
| "metadata": { | |
| "needs_background": "light" | |
| } | |
| } | |
| ], | |
| "source": [ | |
| "samples_feature_images_train = feature_images_train[shuffle_indices]\n", | |
| "samples_images_train = images_train[shuffle_indices]\n", | |
| "sample_labels_train = y_train[shuffle_indices]\n", | |
| "plot_images(samples_feature_images_train, sample_labels_train, ['scissors', 'paper', 'rock'], )\n", | |
| "plot_images(samples_images_train, sample_labels_train, ['scissors', 'paper', 'rock'], )\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "Rf-QOBKgQ8tI" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def shuffle_train_set(): \n", | |
| " shuffle_indices = list(np.random.permutation(len(X_train))) \n", | |
| " return X_train[shuffle_indices], y_train[shuffle_indices], feature_images_train[shuffle_indices], images_train[shuffle_indices]" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "t7IW3xQ8Q8tI" | |
| }, | |
| "source": [ | |
| "#### Cross-validation set" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "LJXzJFZnQ8tI" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "X_cv, y_cv, feature_images_cv, images_cv = load_images(cv_dir)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "MpMYMT2oQ8tJ" | |
| }, | |
| "source": [ | |
| "#### Some examples in cv set" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "9iXd69EXQ8tJ", | |
| "outputId": "25517ad9-1829-4c53-a316-d5defb1cdbd0" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "execute_result", | |
| "data": { | |
| "text/plain": [ | |
| "[483, 1479, 375, 649, 635]" | |
| ] | |
| }, | |
| "metadata": {}, | |
| "execution_count": 18 | |
| } | |
| ], | |
| "source": [ | |
| "shuffle_indices = list(np.random.permutation(len(feature_images_cv))) \n", | |
| "shuffle_indices[:5]" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 461 | |
| }, | |
| "id": "8OY8FfCmQ8tJ", | |
| "outputId": "6cdd2df2-fa44-4cec-dd9b-bc2dcd9818ed" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 1440x720 with 5 Axes>" | |
| ] | |
| }, | |
| "metadata": { | |
| "needs_background": "light" | |
| } | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 1440x720 with 5 Axes>" | |
| ] | |
| }, | |
| "metadata": { | |
| "needs_background": "light" | |
| } | |
| } | |
| ], | |
| "source": [ | |
| "sample_feature_images_cv = feature_images_cv[shuffle_indices]\n", | |
| "sample_images_cv = images_cv[shuffle_indices]\n", | |
| "sample_labels_cv = y_cv[shuffle_indices]\n", | |
| "plot_images(sample_feature_images_cv, sample_labels_cv, ['scissors', 'paper', 'rock'],)\n", | |
| "plot_images(sample_images_cv, sample_labels_cv, ['scissors', 'paper', 'rock'],)\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "wnx1FDNyQ8tJ" | |
| }, | |
| "source": [ | |
| "## Train model via SVM" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "NmLK2MqsQ8tJ" | |
| }, | |
| "source": [ | |
| "#### Result validation tooling" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "SsudgTgHQ8tK" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def compare_accuracy(y_true_train, y_pred_train, y_true_cv, y_pred_cv):\n", | |
| " acc_train = accuracy_score(y_true_train, y_pred_train).round(3)\n", | |
| " acc_cv = accuracy_score(y_true_cv, y_pred_cv).round(3)\n", | |
| "\n", | |
| " r2_train = r2_score(y_true_train, y_pred_train).round(3)\n", | |
| " r2_cv = r2_score(y_true_cv, y_pred_cv).round(3)\n", | |
| "\n", | |
| " plt.figure(figsize=(5, 2))\n", | |
| "\n", | |
| " data = [\n", | |
| " ([\"accuracy train\", \"accuracy cv\"], [acc_train, acc_cv]),\n", | |
| " ([\"r2 train\", \"r2 cv\"], [r2_train, r2_cv])\n", | |
| " ]\n", | |
| "\n", | |
| " for i, (x, y) in enumerate(data):\n", | |
| " fig = plt.figure(figsize=(4, 5))\n", | |
| " fig.add_subplot(1, 1, 1)\n", | |
| "\n", | |
| " plt.bar(x, y, color=['b', 'g'], width=0.1, align=\"center\")\n", | |
| " plt.grid(linestyle=\"-.\", axis='y', alpha=0.4)\n", | |
| " plt.tight_layout()\n", | |
| "\n", | |
| " plt.title(\"Score between train and cv\")\n", | |
| " plt.show()\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "HpR7HF58Q8tK" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def print_result(y_true, y_pred, clear_out=True):\n", | |
| " # Just UI\n", | |
| " if clear_out:\n", | |
| " sys.stdout.flush()\n", | |
| " clear_output(wait=True)\n", | |
| "\n", | |
| " print(\"Accuracy: \"+str(accuracy_score(y_true, y_pred).round(3)))\n", | |
| " print(\"R2: \"+str(r2_score(y_true, y_pred).round(3)))\n", | |
| " print(\"Report: \"+classification_report(y_true, y_pred))\n", | |
| "\n", | |
| " label_names = [0, 1, 2]\n", | |
| " cmx = confusion_matrix(y_true, y_pred, labels=label_names)\n", | |
| " \n", | |
| " sns.set(font_scale=1.4)\n", | |
| " sns.heatmap(pd.DataFrame(cmx), annot=True, annot_kws={\"size\": 16}, fmt=\".1f\",)\n", | |
| " \n", | |
| " plt.title(\"Confusion Matrix for SVM results\")\n", | |
| " plt.show()" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "M8hJrSrrQ8tK" | |
| }, | |
| "source": [ | |
| "#### Duration calculation helper" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "m3LP133LQ8tK" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def process_in_duration(message, proc):\n", | |
| " import datetime\n", | |
| " \n", | |
| " tm_start = datetime.datetime.now()\n", | |
| "\n", | |
| " result = proc()\n", | |
| "\n", | |
| " tm_end = datetime.datetime.now() \n", | |
| " delta = tm_end - tm_start\n", | |
| " \n", | |
| " print(f\"\\033[92m{message} used {delta}\")\n", | |
| " return result" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "_rPltH6_Q8tK" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def train_model():\n", | |
| " model = SVC()\n", | |
| "\n", | |
| " X_shuffled_train, y_shuffled_train, _, _ = shuffle_train_set()\n", | |
| " model.fit(X_shuffled_train, y_shuffled_train)\n", | |
| " \n", | |
| " y_pred_train = model.predict(X_train)\n", | |
| " y_pred_cv = model.predict(X_cv)\n", | |
| "\n", | |
| " print_result(y_cv, y_pred_cv)\n", | |
| " compare_accuracy(y_train, y_pred_train, y_cv, y_pred_cv)\n", | |
| " \n", | |
| " return model" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 1000 | |
| }, | |
| "id": "KSbUwz_PQ8tL", | |
| "outputId": "202b997f-d075-413f-f5dc-8405503087ff" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Accuracy: 0.832\n", | |
| "R2: 0.515\n", | |
| "Report: precision recall f1-score support\n", | |
| "\n", | |
| " 0 0.92 0.82 0.87 496\n", | |
| " 1 0.97 0.67 0.79 496\n", | |
| " 2 0.71 1.00 0.83 496\n", | |
| "\n", | |
| " accuracy 0.83 1488\n", | |
| " macro avg 0.87 0.83 0.83 1488\n", | |
| "weighted avg 0.87 0.83 0.83 1488\n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 2 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "text/plain": [ | |
| "<Figure size 360x144 with 0 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 288x360 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 288x360 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "\u001b[92mTraining has done. used 0:01:36.466468\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "trained_model = process_in_duration(\"Training has done.\", train_model)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "5DAtfXs1Q8tL" | |
| }, | |
| "source": [ | |
| "## Model selection" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "k47CRGknQ8tL" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "def select_model_and_train():\n", | |
| " base_estimator = SVC(gamma='scale')\n", | |
| " param_grid = {'kernel': ('linear', 'rbf'),'C': [1, 10, 100]}\n", | |
| "\n", | |
| " model = GridSearchCV(base_estimator, param_grid, cv=5, return_train_score=True)\n", | |
| " \n", | |
| " X_shuffled_train, y_shuffled_train, _, _ = shuffle_train_set()\n", | |
| " model.fit(X_shuffled_train, y_shuffled_train)\n", | |
| " \n", | |
| " best_parameters = model.best_params_\n", | |
| "\n", | |
| " y_pred_train = model.predict(X_train)\n", | |
| " y_pred_cv = model.predict(X_cv)\n", | |
| " \n", | |
| " print_result(y_cv, y_pred_cv)\n", | |
| " compare_accuracy(y_train, y_pred_train, y_cv, y_pred_cv)\n", | |
| " print(best_parameters)\n", | |
| " \n", | |
| " return model" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 1000 | |
| }, | |
| "id": "F2mNTYYIQ8tL", | |
| "outputId": "3290ec4d-6efa-4157-99e8-4eb0eeb7d494" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Accuracy: 0.832\n", | |
| "R2: 0.515\n", | |
| "Report: precision recall f1-score support\n", | |
| "\n", | |
| " 0 0.92 0.82 0.87 496\n", | |
| " 1 0.97 0.68 0.79 496\n", | |
| " 2 0.71 1.00 0.83 496\n", | |
| "\n", | |
| " accuracy 0.83 1488\n", | |
| " macro avg 0.86 0.83 0.83 1488\n", | |
| "weighted avg 0.86 0.83 0.83 1488\n", | |
| "\n" | |
| ] | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 2 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "text/plain": [ | |
| "<Figure size 360x144 with 0 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 288x360 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 288x360 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "{'C': 10, 'kernel': 'rbf'}\n", | |
| "\u001b[92mModel has been selected. used 0:24:10.521172\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "selected_model = process_in_duration(\"Model has been selected.\", select_model_and_train)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "metadata": { | |
| "id": "L5mJHDsHQ8tM" | |
| }, | |
| "source": [ | |
| "# Bouns, use DL fully connection layers (FCN) instead SVM" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "## Architecture \n", | |
| "\n", | |
| "Instead SVM the non-linear mapping can be replaced by deep neural network (DNN).\n", | |
| "\n", | |
| "" | |
| ], | |
| "metadata": { | |
| "id": "NG5c5Knvn-Em" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "b-q2rnppQ8tM" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "import pandas as pd\n", | |
| "import numpy as np\n", | |
| "import tensorflow as tf\n", | |
| "import keras\n", | |
| "from keras.regularizers import L2\n", | |
| "from keras.models import Sequential\n", | |
| "from keras.layers import InputLayer\n", | |
| "from keras.layers import Dense\n", | |
| "\n", | |
| "from sklearn.metrics import mean_squared_error\n", | |
| "from sklearn.model_selection import train_test_split\n", | |
| "\n", | |
| "import sys\n", | |
| "from IPython.display import clear_output" | |
| ] | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "id": "iknMDOZjQ8tM" | |
| }, | |
| "outputs": [], | |
| "source": [ | |
| "class ImageClassifier:\n", | |
| " def __init__(self,\n", | |
| " X, y, cv,\n", | |
| " name,\n", | |
| " use_mean_norm, hidden_layers, output_units, epochs, L2_lambda=None,\n", | |
| " hidden_activation=\"relu\", output_activation=\"softmax\",\n", | |
| " evaluation_times=50, optimizer=\"adam\", loss=\"sparse_categorical_crossentropy\", verbose=0, debug=True):\n", | |
| " self.X = X\n", | |
| " self.y = y\n", | |
| " self.cv = cv\n", | |
| "\n", | |
| " self.name = name\n", | |
| " # logging meean-square-error between y_test and y_test_pred, y_train and y_train_pred\n", | |
| " # mean of report\n", | |
| " # std of report\n", | |
| " # logging training history\n", | |
| " # {\"train\" : (report, mean, std, histories), \"test\" : (report, mean, std)}\n", | |
| " self.result = {}\n", | |
| "\n", | |
| " self.use_mean_norm = use_mean_norm\n", | |
| " self.hidden_layers = hidden_layers\n", | |
| " self.output_units = output_units\n", | |
| " self.L2_lambdas = L2_lambda\n", | |
| " self.epochs = epochs\n", | |
| " self.evaluation_times = evaluation_times\n", | |
| " self.optimizer = optimizer\n", | |
| " self.hidden_activation = hidden_activation\n", | |
| " self.output_activation = output_activation\n", | |
| " self.loss = loss\n", | |
| " self.verbose = verbose\n", | |
| " self.debug = debug\n", | |
| "\n", | |
| " def preprocessing(self):\n", | |
| " if self.use_mean_norm:\n", | |
| " self.X = (self.X - self.X.mean()) / self.X.std()\n", | |
| " self.y = (self.y - self.y.mean()) / self.y.std()\n", | |
| "\n", | |
| " def build_model(self):\n", | |
| " def check_index_in_list(a_list, index):\n", | |
| " return a_list is not None and index < len(a_list)\n", | |
| "\n", | |
| " n_cols = self.X.shape[1]\n", | |
| "\n", | |
| " self.model = Sequential()\n", | |
| " self.model.add(InputLayer(input_shape=(n_cols,)))\n", | |
| "\n", | |
| " for idx, units in enumerate(self.hidden_layers):\n", | |
| " if check_index_in_list(self.L2_lambdas, idx):\n", | |
| " L2_lambda = self.L2_lambdas[idx]\n", | |
| " self.model.add(Dense(units,\n", | |
| " activation=self.hidden_activation,\n", | |
| " kernel_regularizer=L2(l2=L2_lambda)))\n", | |
| " else:\n", | |
| " self.model.add(Dense(units,\n", | |
| " activation=self.hidden_activation))\n", | |
| "\n", | |
| " self.model.add(Dense(self.output_units, activation=self.output_activation))\n", | |
| " self.model.compile(optimizer=self.optimizer, loss=self.loss)\n", | |
| "\n", | |
| " if self.debug:\n", | |
| " self.model.summary()\n", | |
| "\n", | |
| " def shuffle_train_set(self):\n", | |
| " shuffle_indices = list(np.random.permutation(len(self.X)))\n", | |
| " return self.X[shuffle_indices], self.y[shuffle_indices]\n", | |
| "\n", | |
| " def train(self):\n", | |
| " train_report = [] # logging meean-square-error between y_train and y_train_pred\n", | |
| " test_report = [] # logging meean-square-error between y_test and y_test_pred\n", | |
| " histories = [] # logging training history\n", | |
| "\n", | |
| " for idx in range(self.evaluation_times):\n", | |
| " X_shuffled_train, y_shuffled_train = self.shuffle_train_set()\n", | |
| "\n", | |
| " # Fit model\n", | |
| " history = self.model.fit(X_shuffled_train, y_shuffled_train, validation_data=self.cv, epochs=self.epochs,\n", | |
| " verbose=self.verbose)\n", | |
| "\n", | |
| " # logging history\n", | |
| " histories.append(history)\n", | |
| "\n", | |
| " # Logging the evaltion to report\n", | |
| " y_train_pred = self.model.predict(X_train)\n", | |
| " y_train_pred = np.asarray([y.argmax() for y in y_train_pred])\n", | |
| " train_accuracy = np.round(accuracy_score(y_train, y_train_pred), decimals=3)\n", | |
| " train_report.append(train_accuracy)\n", | |
| "\n", | |
| " y_test_pred = self.model.predict(self.cv[0])\n", | |
| " y_test_pred = np.asarray([y.argmax() for y in y_test_pred])\n", | |
| " test_accuracy = np.round(accuracy_score(\n", | |
| " self.cv[1], y_test_pred), decimals=3)\n", | |
| " test_report.append(test_accuracy)\n", | |
| "\n", | |
| " # Just UI\n", | |
| " sys.stdout.write(\n", | |
| " f\"Complete evaluation {idx+1}, train accuracy: {train_accuracy}, test accuracy: {test_accuracy}\\n\")\n", | |
| " sys.stdout.flush()\n", | |
| " clear_output(wait=True)\n", | |
| "\n", | |
| " self.result[self.name] = {}\n", | |
| "\n", | |
| " mean = np.round(np.mean(train_report), decimals=3)\n", | |
| " std = np.round(np.std(train_report), decimals=3)\n", | |
| " self.result[self.name][\"train\"] = (train_report, mean, std, histories)\n", | |
| "\n", | |
| " mean = np.round(np.mean(test_report), decimals=3)\n", | |
| " std = np.round(np.std(test_report), decimals=3)\n", | |
| " self.result[self.name][\"test\"] = (test_report, mean, std)\n", | |
| "\n", | |
| " def plot_result(self, history_range=None):\n", | |
| " train_report, train_mean, train_std, histories = self.result[self.name][\"train\"]\n", | |
| " test_report, test_mean, test_std = self.result[self.name][\"test\"]\n", | |
| "\n", | |
| " # Plot MSE, mean and std of MSE\n", | |
| " import matplotlib.pyplot as plt\n", | |
| " plt.plot(train_report)\n", | |
| " plt.plot(test_report)\n", | |
| " plt.legend([\"Train Accuracy\", 'Test Accuracy'])\n", | |
| " plt.title(\n", | |
| " f\"Accuracy train vs. test dataset, Eval.{self.evaluation_times} times, {self.epochs} epochs/time\")\n", | |
| " plt.xlabel(\"Evaluation times\")\n", | |
| " plt.ylabel(\"Accuracy\")\n", | |
| " plt.show()\n", | |
| "\n", | |
| " print('Train Mean: ', train_mean)\n", | |
| " print('Train Standard Deviation: ', train_std)\n", | |
| " print(\"Train Report for {}\\n{}\".format(self.name, train_report))\n", | |
| "\n", | |
| " print('Test Mean: ', test_mean)\n", | |
| " print('Test Standard Deviation: ', test_std)\n", | |
| " print(\"Test Report for {}\\n{}\".format(self.name, test_report))\n", | |
| "\n", | |
| " # Plot loss of train and cross-validation\n", | |
| " if history_range is not None:\n", | |
| " for hist_idx in history_range:\n", | |
| " history = histories[hist_idx]\n", | |
| " string = \"loss\"\n", | |
| "\n", | |
| " plt.title(f\"Loss train vs. cross-validation, eval.#{hist_idx+1}\")\n", | |
| " plt.plot(history.history[string])\n", | |
| " plt.plot(history.history['val_'+string])\n", | |
| " plt.xlabel(\"Epochs\")\n", | |
| " plt.ylabel(string)\n", | |
| " plt.legend([string, 'val_'+string])\n", | |
| " plt.show()\n" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "## Deep network\n", | |
| "\n", | |
| "A few layers of neural networks are added and some penalties are applied (regularization) to mitigate the overfitting, but the results of the attempts were not very satisfactory.\n", | |
| "\n", | |
| "The bias and variance between training and cross-validationg is quite uncool." | |
| ], | |
| "metadata": { | |
| "id": "wPST12pYpK0-" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "M3VzDLvVQ8tO", | |
| "outputId": "2d7647d2-3375-41ad-834d-b31443e3c7e8" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "\u001b[92mTraining deep NN has done. used 0:25:16.756477\n" | |
| ] | |
| } | |
| ], | |
| "source": [ | |
| "classifer = ImageClassifier(\n", | |
| " X_train, y_train,\n", | |
| " (X_cv, y_cv),\n", | |
| " \"paper scissors and rock\",\n", | |
| " use_mean_norm=False, \n", | |
| " hidden_layers=[2**3, 2**3, 2**2], \n", | |
| " L2_lambda=[.1, .1, .1 ], \n", | |
| " output_units=3,\n", | |
| " epochs=50, \n", | |
| " evaluation_times=50,\n", | |
| ")\n", | |
| "classifer.build_model()\n", | |
| "process_in_duration(\"Training deep NN has done.\", classifer.train)" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "### Plot\n", | |
| "\n", | |
| "1. Accuracy of train vs. cross-validation\n", | |
| "2. Loss convergence trend of train and cross-validation in 1. training iteration." | |
| ], | |
| "metadata": { | |
| "id": "W8vJvZ-Trnv5" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "execution_count": null, | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 772 | |
| }, | |
| "id": "8qL4-KL5Q8tQ", | |
| "outputId": "0c83e74c-fcc4-4c76-8f47-b5e55487f402" | |
| }, | |
| "outputs": [ | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Train Mean: 0.981\n", | |
| "Train Standard Deviation: 0.012\n", | |
| "Train Report for paper scissors and rock\n", | |
| "[0.968, 0.967, 0.978, 0.978, 0.978, 0.965, 0.983, 0.986, 0.979, 0.929, 0.992, 0.981, 0.964, 0.992, 0.989, 0.972, 0.982, 0.967, 0.991, 0.954, 0.985, 0.989, 0.993, 0.985, 0.988, 0.985, 0.987, 0.978, 0.981, 0.969, 0.99, 0.988, 0.995, 0.989, 0.983, 0.989, 0.991, 0.981, 0.979, 0.989, 0.985, 0.979, 0.983, 0.979, 0.984, 0.992, 0.985, 0.997, 0.993, 0.98]\n", | |
| "Test Mean: 0.77\n", | |
| "Test Standard Deviation: 0.021\n", | |
| "Test Report for paper scissors and rock\n", | |
| "[0.766, 0.755, 0.75, 0.765, 0.775, 0.74, 0.765, 0.769, 0.767, 0.712, 0.784, 0.749, 0.784, 0.78, 0.776, 0.712, 0.79, 0.797, 0.777, 0.71, 0.788, 0.782, 0.768, 0.765, 0.794, 0.767, 0.784, 0.778, 0.792, 0.782, 0.794, 0.789, 0.793, 0.788, 0.751, 0.759, 0.779, 0.751, 0.771, 0.768, 0.79, 0.739, 0.779, 0.772, 0.784, 0.763, 0.739, 0.795, 0.781, 0.784]\n" | |
| ] | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZUAAAEqCAYAAADdx82bAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3xTVf/A8c/NapIOuhktFAqU1YIglL3KBpUhPC5ABMUNigvl8SeP+gAO9GE4QFBxLxARoWyKgDJklDJLKaNld5ekmff3R2wktKW7ael5v16+JOeuc5I033vmlWRZlhEEQRCECqBwdwYEQRCEW4cIKoIgCEKFEUFFEARBqDAiqAiCIAgVRgQVQRAEocKIoCIIgiBUGBFUhJtasGABLVq0cHc2hEpQ2Gc7btw4xo0bV+yxKSkptGjRghUrVlRonqZPn05MTEyFnrMmqqz3tyqIoFKEFStW0KJFCw4cOODurNzUvn37WLBgAdnZ2e7OiiCUyMmTJ1mwYAEpKSnuzsotaeDAgSxduhSAEydO0KJFi0Lf62+//ZZnnnmGmJgYWrRowaRJkyrk+qoKOYvgNvv372fhwoWMHDkSHx+fCj//448/zuTJkyv8vEL1lP9jVJlOnjzJwoULiY6OJjQ01GXbG2+8gZiPXXaZmZmcOXOGdu3aAXDw4EECAgIKvM8An3zyCTk5OURFRZGRkVFheRBBpZYxGo3odLoS769SqVCpau/XxGAwoNfr3Z2NKqPRaNx6fbVa7dbr13Tx8fGoVCoiIyOdr6Oiogrd98svv6RBgwZIklShTY6i+aucjh07xiOPPEKHDh247bbbGDduHHv37nXZx2q18uGHHzJo0CDatm1LdHQ0Y8aMYf369c59rl69yowZM+jduzeRkZF069aNhx9+mMTExCKvvWDBAt5++20A+vXrR4sWLWjRogW7du0CICYmhkmTJvHHH38wZswYoqKiWLJkCQCbNm3iscceo1evXkRGRtK3b1/eeustTCZTgWvc2O6ef969e/cyevRooqKi6NevHytXriz2/brjjju4//77C902cOBAl/b8NWvWcPfdd9OhQwfat2/PkCFD+OCDD4q9RlEOHTrEY489RnR0NO3ateOOO+5g0aJFzu3Tp08nKiqKlJQUHnvsMTp06MCjjz4KgM1m46OPPmLAgAFERkbSp08f3n77bfLy8lyucfjwYR555BG6dOlCVFQUffv25YUXXnDZryzlev3112nXrh3Xrl0rsG3GjBncdtttGAwGoOSfbWEK61PJzs5m+vTp3H777XTs2JGXXnqJnJycAsceO3aMl19+mf79+xMVFUXnzp159tlnOX/+vHOfFStWMHXqVADGjx/v/M7m9x0U1qdS0ve+PN/Lm0lOTmbq1Kl07tyZqKgoRowYQWxsrHP7oUOHaNGiBT/++GOBY/fu3etSvtTUVP7zn/8wZMgQ2rVrR8eOHXn00Uc5fvx4mfNnNBpJT08nPT2dv/76iyZNmmAwGEhPT+fAgQM0bdrUuf16ISEhSJJU5usWpfbeglaApKQk7r//fvR6PZMmTcLDw4Mff/yRCRMm8Nlnn9GpUycAFi5cyMcff8zo0aNp27YtBoOBo0ePEh8fz8CBAwGYMmUKJ06c4IEHHiA0NJSMjAx2797N6dOnad68eaHXHzBgAKdPn2b16tW8/PLL+Pn5AdC0aVPnPmfPnmXKlCmMGTOG0aNHU79+fcDxx63RaBg3bhze3t4cPHiQZcuWcfHiRd5///1iy56SksLUqVMZPXo0I0eOZPny5UyfPp02bdoUmV+AoUOHMn/+fC5evEi9evWc6QkJCZw5c4aHHnoIgJ07dzJt2jS6dOnCc889h0KhIDk5mX379hWbt8L88ccfTJ48mYCAAMaOHUtwcDDJycls3rzZGTgAZFlm0qRJREVF8eKLL6JUKgH4v//7P3766ScGDhzIhAkTSEhIYOnSpSQmJrJ48WIkSSI9PZ2JEyfi6+vLww8/TJ06dbhw4QKbN2/GYDCg1WrLXK6hQ4fy9ddfs3nzZu68805nusViYePGjfTt29dZoyrvZ3s9WZZ54okn+Ouvv7jnnnto1qwZmzZt4qWXXiqw786dO0lOTmbEiBEEBwdz9uxZvvvuO+Lj41m9ejU6nY5OnToxbtw4vvzySx577DHCw8MB6NChQ5F5KMl7n6+s38uiJCUlce+99xIYGMikSZPw9PRkw4YNTJ06lbfffpvhw4cTFRVFWFgYa9euZcyYMS7Hr127Fo1Gw4ABAwBHANq7dy+DBg2iQYMGXL58me+++45x48axevVqgoODS53HJUuWsHDhQpe0rl27Ov994sQJZ7NmeYJXiclCoZYvXy5HRETI+/fvL3KfJ598Um7Tpo2cnJzsTEtLS5Ojo6PlkSNHOtOGDx8uT548ucjzZGVlyREREfKSJUtKnc8lS5bIERER8rlz5wps69u3rxwRESFv3LixwDaDwVAg7aOPPpJbtGghnz9/3pk2f/58OSIiotDz7t6925mWlpYmR0ZGynPmzLlpfk+fPi1HRETIn332mUv6W2+9Jbdu3VpOS0uTZVmW//vf/8odOnSQrVbrTc9XEjabTe7Xr5/cq1cvOSMjw2Wb3W53/vull16SIyIi5FmzZrnsc/ToUTkiIkKePn26S3r+e7N582ZZlmV5w4YNckREhBwfH19kXspaLrvdLvfu3Vt+7LHHXNK3bt0qR0REyBs2bHCmleezHTt2rDx27Fjn6/wyLV682JlmtVrlsWPHyhEREfLy5ctvet2//vpLjoiIkFeuXOlMW7t2rRwRESH/+eefBfZ/6aWX5L59+zpfl/S9l+XyfS+L8tBDD8lDhw6VjUZjgfSePXs6vz/vvfee3KpVK+f3V5Yd37vu3bvLjz/+uDPtxvPIsiyfPXtWjoyMlD/44ANn2rlz5wq8v0U5e/asvGPHDnnbtm1ymzZt5Hnz5sk7duyQFyxYILdu3VqOi4uTd+zYIe/YsaPIc/Tt21eeOHFisdcqCdH8VUY2m43t27fTt29fGjdu7Ez39/dn1KhRHD58mKtXrwLg7e1NYmIiycnJhZ5Lq9WiVqvZvXs3mZmZFZrPevXq0a9fvwLp+f0qdrudnJwc0tPTuf3225FlmcOHDxd73saNGztrYuAod5MmTTh37txNjwsLC6NNmzasWbPGJT02NpYuXbrg7+8PON4zo9HI9u3bi81LcQ4fPsy5c+cYP348vr6+LtsKq/7f2DwXFxcHwIQJE1zSJ0yYgFKpZOvWrc48A2zduhWLxVJoXspaLkmSGDJkCNu3b3dpelqzZg3e3t706tXLmVbez/Z627ZtQ6FQcN999znTlEolDzzwQIF9r++ru3btGhkZGTRu3BgfH59SXzdfSd/7fGX9XhYmMzOTnTt3MmTIEGdzUv5/PXv25NKlS86/6WHDhmGz2Vi3bp3z+N27d3PlyhWGDRvmTNNqtc5/G41GMjIy8PT0pEmTJmV+jxo2bEi3bt3w9fXFYrFwzz330K1bN8xmM5GRkfTq1Ytu3brRrVu3Mp2/tETzVxmlp6djNBpp0qRJgW35VfrU1FQCAwOZMmUKTz75JIMHD6ZZs2b06NGDO+64w9mBptFoeP7553n77bfp3r07bdu2pXfv3gwfPtzZXFVWDRs2LDT9xIkTvPPOO+zevbtA23Rubm6x523QoEGBtDp16pCVlVXsscOGDePtt98mJSWF0NBQ9u/fT2pqKk899ZRzn/vvv5/Y2FgmT55McHAw3bp1Y+DAgcTExJS6HTj/B6UkzR8KhYKQkBCXtNTUVCRJKvBZe3t7ExQURGpqKgDR0dEMGjSIhQsX8tlnnxEdHU1MTAx33HGHs2mqJOW6cuVKgetotVqGDh3Kp59+yoYNGxg1ahRms5mNGzcyYMAAlw728n62N5Y9MDAQLy8vl/Trb6TyZWVlMXfuXNatW1fg5qiwPpiSXr8k732+8nwvb3T27FlkWWbBggUsWLCg0H3S09MJDw8nIiKC5s2bs2bNGmcAXrNmDTqdjr59+zr3N5lMzJs3j1WrVhX4nG+84SkJo9GI0WgEHDcAISEhqNVq0tPT2bVrF1FRUc6+lPwbtsomgkoV6NSpExs2bGDLli1s376dlStXsmzZMp577jkeeeQRwHHn1a9fPzZt2sTOnTv58MMPWbRoER9//DGdO3cu87U9PDwKpOXk5DB+/Hj0ej3PPvssjRo1QqvVcunSJaZPn47dbi/2vApF2Su5Q4cO5Z133mHNmjVMnjyZNWvWuLQ7AwQEBPDzzz/zxx9/sG3bNn7//XdWrlxJ3759+eijjyqlgxHKN9pNkiTmz59PfHw8W7ZsYefOnbz66qssWrSIH374gYCAgBKVq0ePHi7nnT17NqNGjXK23a9Zs4ZRo0axbds2cnNzGTp0qHPfivhsy+qZZ55h3759PPTQQ7Ru3RpPT08kSeLZZ5+tsmHC5fle3ij/vZowYQK9e/cudJ/rb1SGDh3KggULuHz5Mv7+/qxfv96lrwscQ6aXL1/O2LFjad++PT4+PigUCmbNmlWm96i4/pQDBw7w5ZdfAlXUn4IIKmXm7++PTqcrtEnr1KlTAC53vHXq1GHEiBGMGDGCvLw8HnnkERYsWMDEiROdncENGzZkwoQJTJgwgYsXLzJ8+PBig0pZflx37dpFRkYG8+fPJzo62pm+Y8eOUp+rLOrXr0/79u1Zu3YtDz/8MLGxsfTs2dPZfJRPo9HQu3dvevfujSzLzJ07l08++YR9+/Zx++23l/h6+bW1xMREl2aikgoJCUGWZZKTk11GwuXm5nLlyhX69Onjsn/btm1p27YtU6dOJS4ujsmTJ/PDDz/w+OOPl6hcn332mcv5mjVr5vz3kCFDWLJkCRkZGaxZswY/Pz+XZo2K/mxDQkLYuXMnubm5LrWV06dPu+yXlZXFzp07efrpp11qnCaTqcDE3NJ8Z0v73lek/O+NUqksUdPRsGHDmDdvHrGxsTRp0oSMjAyXpi9wNPOOGDGCGTNmuKRnZWU5B9qUxogRI5xNm48//jgTJ04kOjqaQ4cO8f7777No0aIqH6Yt+lTKSKlU0rNnT7Zs2cLZs2ed6ZmZmaxcuZLIyEgCAwMBCkws0mq1hIeHYzKZyMvLw2g0FmimqFevHgEBAcXOlM9vxy7NjPr8u7nr74zsdnuBH7PKNGzYMI4cOcLy5cu5fPlygT++G98zSZJo3bo18E9TisViISkpicuXL9/0Wm3atKFhw4Z88cUXBZplSnJ3mH+XumzZMpf0ZcuWYbPZnM0bWVlZBc7Xpk0b4J/PpyTlym//zv/v+hFBw4YNw2q1smrVKrZs2cLAgQNdalYV/dn26tULu93Ot99+63K+r7/+2mW//BujG8v/+eefF6gdleY7W9L3vjIEBATQuXNnfvzxRy5dulRg+41DdK/vLyysrwsc79ON79Hq1auL/Q6Do6krKSnJ5br5/Sn16tXDZDIxYsQIunXrhizLNGvWjN69e1dpfwqImkqxVqxYwc6dOwuk33fffTzzzDNs376d+++/n/vvvx8PDw9++OEHcnJymD59unPfoUOH0qlTJyIjI/Hz8+P48eP89NNP9O3bF09PT44ePcqDDz7o7HPRaDTExcWRlJRU6NDN6+VPcpo7dy533nknarWaLl26EBAQUOQxHTp0wNfXl+nTpzN27FhUKhXr1q1zznOoCoMHD2bWrFnMnj27QLszwL///W8yMjLo2rUr9erV49KlS3z99dcEBQXRsWNHAC5dusTQoUMZOXIkc+bMKfJaCoWC//znPzz66KOMGDGCu+++m+DgYM6cOcO+ffv47rvvbprXli1bMnr0aH766Sdyc3Pp3LmzMyD27NnT+cP3888/880339C/f38aNWpEXl4eK1asQKlUMnjw4BKX62by2+7nzZuHwWAoEIwr+rONiYmhQ4cOvPfee6SmptK8eXM2btxYoI/Cy8uL6OholixZgsVioUGDBvz111/s2bOnQF9B69atUSqVLFq0iOzsbLRaLW3bti20/6+k731ZygWwefPmm+43c+ZM7rvvPu666y7GjBlDo0aNSEtL4+DBgyQlJbFhwwaX/YcNG8Y777zD8ePHGTRoUIHJpDExMaxcuRIvLy+aN2/O0aNHWbt2bZF9n9eLj49n/PjxPPXUUzz99NMu2/bv34+fn5+zr+vAgQO0b9/+pufbvHkzx44dAxw3NCkpKXz44YfOfLZs2bLYPBVGBJVifP/994WmDxo0iKZNm/Ltt986my9kWSYyMpI333zTZQTKgw8+yObNm/nzzz/Jy8ujfv36PPLII87+lHr16nHnnXfyxx9/8OuvvyJJEo0bN+a///0vo0ePvmn+oqKieO655/jmm294+eWXsdvtfPHFFzcNKr6+vixevJg5c+awYMEC9Ho9AwcOdP7xVIXAwECio6P5448/GDp0aIFZ63fddRc//vgj3333HVlZWQQGBtK7d2+efPLJAp3GJdG9e3e+/PJLZye63W6nYcOGLnM+bub1118nNDSU5cuXs3nzZgICApg4cSJTpkxxNufkNzvExsZy5coVvLy8aN26Na+++qpz2YyKKNfQoUOZN28ewcHBLt8zqPjPVqFQ8NFHHzFr1izndzMmJoYXX3yRESNGuOw7d+5c/vvf//L9999jsVjo1KkTy5Ytc849yhcYGMgbb7zBokWLePXVV7HZbMyePbvIH9aSvPelZTAYCAsLK3a/8PBwli9fzgcffMDKlSvJzMzEz8+Pli1bOidxXi+/v9BgMLj0deWbMWMGKpWKNWvWYDAYiIyM5JNPPuGdd94pUznyHThwwPkdy3/94osv3vSY9evX8/PPPztfZ2dnM2/ePMDxm1TWoCLJVdWDJgiCUA2cPHmSYcOGsWjRokrtk6mtRJ+KIAi1yq5du2jfvr0IKJXErTWVM2fOsHTpUg4ePEhiYiLh4eGsXr262OMMBgMffvihs5mhbt263HXXXUyePNntC+IJgiDUZm7tU0lMTCQuLo527dpht9tLPE575syZbNy4kWeffZbmzZsTHx/P/Pnzyc7O5pVXXqnkXAuCIAhFcWtNxW63O4dATp8+nYSEhGJrKlarlQ4dOvDwww8zZcoUZ/rMmTNZv359oSO1BEEQhKrh1j6Vssx+lWUZm81WYKKcj4+PeLiPIAiCm9W4jnq1Ws3w4cP58ssvOXjwINeuXePPP//khx9+KHSRO0EQBKHq1Mh5Kq+//jqvvfYa//rXv5xpEyZMcFkeoqQyMq5ht5e+hhMQ4EVaWukW57sV1NZyQ+0tuyh37VJcuRUKCT8/zyK318igMnfuXOLi4njzzTdp3LgxBw4c4IMPPiAwMNA5obCkbvbmFCcgoPST8G4FtbXcUHvLLspdu5Sn3DUuqJw4cYJPP/2UDz/80PmckE6dOmG1Wpk/fz733XdfqWZcp6XllqmmEhTkzZUrZVvOuyarreWG2lt2Ue7apbhyKxTSTYNOjetTOXnyJACtWrVySW/dujVms7nQhd8EQRCEqlHjgkr+cvI3PiUtISEBSZIKfUiPIAiCUDXc2vxlNBqdjwtNTU0lNzeX2NhYwLFQYkhICK+88gorV67kyJEjgGNV3rZt2/Laa6+RlpZGWFgY8fHxLF68mLvvvtvlkaaCIAhC1XJrUElLSyuw0mf+6/yn3dntdmw2m3O7Uqnk448/Zt68eSxevJirV69Sv359Jk6cyKOPPlql+RcEQRBc1fpVikVHfenU1nJD7S17ScttNF4jNzcTm81aBbmqfAqFolIfv1wdKZUqQkLqYzYX3TNSXEd9jRv9VR0cPHmVVV/s5eUHOqBS1rhuKUGocEbjNXJyMvD1DUKt1pT5OSfViUqlwGqtPUFFlmUsFjMXLlxEr6+DTle26RbiF7EMrmblkXw+m2t5t8YdmSCUV25uJr6+QWg0HrdEQKmNJElCo/HA1zeQ3NzM4g8ogggqZaDXOip4hjyLm3MiCNWDzWZFrRaPnbgVaDQe5WrCFEGlDPQejqBiNNmK2VMQag9RQ7k1lPdzFEGlDHR/BxWDSdRUBEEQrieCShn80/wl+lQE4VazdOkiBgzo6e5s1FgiqJTBP81fIqgIgiBcTwSVMvin+UsEFUEQhOuJeSploNUoUUii+UsQaoOLFy/ywQf/Y8+eP7FYLLRuHckTT0yhVas2zn127Pidzz//hNOnk1EoFISEhPLgg5Po3TumRNtvJSKolIEkSei1atH8JQi3OIPhGk8/PRlZlnn22RfR6/V8++1XPP30oyxd+hVhYY1JTU1hxowX6N9/EJMnP4Esw8mTieTkZAMUu/1WI4JKGXnq1KL5SxCKsePQBbbHX3DLtXu0rU/3qPrlOsdvv/3KxYsXWLbsW8LDmwHQsWNnxoy5k6+++pwZM2Zy4sQxrFYr06a9iF7vmIUeHd3FeY7itt9qRJ9KGXnq1BhF85cg3NIOHtxPkybhzoACoNPp6NatJwcP7gegadPmKJVKZs78N9u3x5GT47pOWnHbbzWiplJGnlq1mKciCMXoHlX+2oI75eTk4O8fUCDdz8+f7GxH81WjRmG89db7f9dcXgQcNZFnnnmBkJDQYrffakRNpYw8dSrR/CUItzgfHx8yMtILpGdkpOPj4+N83aVLNxYuXMzatVt4/fXZnD6dzH/+8+8Sb7+ViKBSRp460VEvCLe6tm1v49SpJJKTTznT8vLy2LlzO+3atS+wv16vp3fvGAYNGsrp08ml3n4rEM1fZeSpVYshxYJwixs27E5++OEbXnzxWR555HH0eh3ffvsVeXlGxo6dAMDKlcs5dOggXbp0IzAwiEuXLrJ69S906hRdou23GhFUyshTpybPbMNul1EoxEJ6gnAr0us9WbBgMQsXvs/cubOxWq20atWG+fM/JiysMQDNmjVn587tfPDBPLKyMvHz86dXrz5MnvxkibbfasSTH8v45McdRy6zdFUCC57piadWXQk5q55q69MPofaWvSTlvnjxDPXqhVVRjqpGbXtIVz6VSkFKSnKRn2dxT34UfSpl5KUTi0oKgiDcyK3NX2fOnGHp0qUcPHiQxMREwsPDWb16dYmOzcnJYf78+axbt4709HSCg4MZPnw4U6dOreRcO3jqHLUT0VkvCILwD7cGlcTEROLi4mjXrh12u52StsQZDAbGjh2LJEm88MILBAcHc+7cOS5evFjJOf6H/u8mL1FTEQRB+Idbg0pMTAz9+/cHYPr06SQkJJTouMWLF5OTk8Ovv/6Kp6dj2YPOnTtXWj4Lk19TEXNVBEEQ/uHWPhWFomyX/+mnnxg9erQzoLiDl2j+EgRBKKDGddSnpKRw5coV/Pz8eOyxx4iKiqJjx468+OKLZGVlVVk+RPOXIAhCQTUuqFy9ehWAt99+G09PTxYtWsRLL73Etm3bmDZtWpXlw1Mrnv4oCIJwoxo3+dFud4wbDwsL491330WSHBMPvb29mTp1KvHx8bRt27bE57vZeOviaDVKZIWCoCDvMp+jJqpt5b1ebS17ceW+fFmBSlXj7lGLdSuWqSQU5fhdq3FBpU6dOgB07drVGVDyX4NjRFlpgkpZJz8GBXmj81CRlmGoVRPiausEQKi9ZS9Jue12+y03UbA2T3602+1Ffua33OTHhg0botFoitxuMpmqLC96D5Vo/hIEQbhOjQsqGo2G7t27s3PnTpd5LTt27AAgMjKyyvKi8xDL3wuCULgePTryzTdflnj/0aPv5L333qrEHFUNtzZ/GY1G4uLiAEhNTSU3N5fY2FgAoqKiCAkJ4ZVXXmHlypUcOXLEedxTTz3Fvffey7Rp0xg1ahTnz5/nvffeo0ePHqVq+iovvVZF1jVzlV1PEAShunNrUElLSyuwrEr+69mzZzNq1Cjsdjs2m81ln8jISJYsWcLcuXN54okn8PLyYujQoTz//PNVlndwNH9dTDNU6TUFQRCqM7cGldDQUI4fP37TfebMmcOcOXMKpHfp0oUff/yxsrJWIqL5SxBuLWvXrmb27NdZseI36tWr60w3Go3ceecAJkx4mF69+rB06WIOHTpIZmYm9evXZ/DgYdx33zhUqor9Sd22bSuff/4Jp08n4+npRe/eMTzxxBT0ej0AVquVxYs/ZNOm9aSnp+Ht7UNERAtmzJiJn59/sdsrQ40b/VWd6LWOjnpZll1GogmCUDP16tWHd96ZzaZN63nggXHO9N9/j8NkMtG//2DOn0+hYcNGDBgwCL3ek1OnTrJ06WKys7N58smKW9B2+/Y4Zsx4gT59+vHII09w4cJ5Fi1ayLlzZ5k370MAvvrqc37++Scef/xpmjQJJzs7i7179zgHLBW3vTKIoFIOOg8VNruM2WLHQ6N0d3YEodqxnNiB5fg2t1xb3aIX6ojupTrG09OLbt26s3HjOpegsnHjOqKi2lGvXj3q1atHhw4dAZBlmbZtb8Nul/n000U88cSUCrvB/PTTxbRs2Zo33vinpaZOHV9ee+1l9u3bS4cOHTlyJIHo6M6MGjXGuU/v3jHOfxe3vTLUuNFf1Yne4+9nqogmMEG4ZQwYMJijR4+QknIOgOzsLHbv/oMBAwYDjmkLS5cu4p57RhAT040+fbowf/5ccnNzSU9Pq5A8GAwGEhNPEBMzwCW9T58YlEol8fEHAIiIaMkff+xk6dJFHDmSUKD/ubjtlUHUVMpBr/0nqPh5e7g5N4JQ/agjupe6tuBuXbv2wMvLi/Xr1zF+/ES2bt2MLMv07etYUf2jjxawatXPPPTQw7Rs2QovL2/27NnF4sUfYjZXzGjQ3NwcZFkmICDAJV2pVFKnji/Z2Y51DsePn4hCoSA29jc+++wT6tSpw/DhdzNp0qMolcpit1cGEVTKQfd3TcUoFpUUhFuGRqOhV6++bNgQy/jxE9m4cR2dOnXG19cXgC1bNjJ8+CjGjXvIecyBA/srNA9eXt5IklSg5mOz2cjKysTHp44zrxMnTmbixMmcP5/KunVr+PTTxQQHBzNixOhit1cG0fxVDqL5SxBuTQMGDCI5+RR//rmTAwf2OZu+wNH8pVarna9lWWbDhtgKvb5er6d58wg2b97okv0WJB8AACAASURBVB4XtwWbzUbbtrcVOKZBgxAeeugRgoKCOX06udTbK4qoqZTDP81fFjfnRBCEitShQycCAgKZM+cNNBoNPXv2cW7r1Kkzq1atICysMf7+/qxatZKcnIpfE27ixMm8/PLzvPbaKwwZcgcXL57n448Xcvvt0c6BAi+//BwRES1p3rwFer2eXbv+4PLlS3TsGF2i7ZVBBJVyEM1fgnBrUiqV9Os3gB9++JZ+/QY454UATJv2Iu+8M5v58+eiVmsYOHAwffv24/XXX63QPPTo0Zs333ybzz//hJdffg5PTy/69x/ME0887dwnKuo2tmzZwPfff43NZiM0tCEzZsykR4/eJdpeGSS5pA+Gv0WVZ5Xi1POZPDY3jrt7hzOsa+OKz1w1VFtX6oXaW/aSlPvixTPUqxdWRTmqGrV5leKUlOQiP89bbpXi6kStUqBUSKJPRRAE4W+i+ascJEn6e1Z95Y/9FgShZpJl+abzQyRJqrThve4ggko56T1UGPJER70gCIXbv/8vpkx5rMjtt93WgYULF1dhjiqXCCrlJBaVFAThZlq2bMWSJV8Uuf36QQC3AhFUyil/UUlBEITC6PWetGzZ2t3ZqDKio76cHM1fIqgIgiCACCrlJpq/BMGhls9OuGWU93MUQaWcRPOXIIBSqcJiEY/WvhWYzSaUyrL3jIigUk46DxVmix2rrfZNkhKEfF5evmRmXsFsNokaSw0lyzJms4nMzKt4efmW+Tyio76c8heVNJqseOs1bs6NILiHTucJQFbWVWy2W6PmrlAosNtr182iUqkiJKQ+ZnPZ6xtuDSpnzpxh6dKlHDx4kMTERMLDw1m9enWpzrFhwwaeeuopmjdvXupjK8L1z1QRQUWozXQ6T2dwuRXU1mV56tQpX7ndGlQSExOJi4ujXbt22O32UlebjUYjs2bNIjAwsJJyWLz8RSXFCDBBEAQ396nExMQQFxfH/PnzadOmTamP//DDDwkNDaVnz56VkLuSub75SxAEobZza1BRKMp++aSkJL788ktefbVil5suLb3W8bAeUVMRBEGowaO/Xn/9dUaPHk1ERIRb86HzcCwEJ2oqgiAINXT012+//caJEydYsGCBu7OC3uPvmooIKoIgCDUvqOTm5jJnzhymTZuGj49Puc93s4fNFCcoyBu7XUaSAKWCoCDvcuenJqgt5SxMbS27KHftUp5y17ig8vHHH+Pr68uAAQPIzs4GwGKxYLfbyc7ORqvVotGUfGhveZ78mD/sTqtRkZZuqBXDD2vrMEuovWUX5a5diit3cU9+rHFB5dSpU5w4cYLOnTsX2NapUydefvllJkyYUKV50ov1vwRBEIAaGFSeeeYZHnzwQZe0xYsXk5yczOzZswkLq/rnZOvESsWCIAiAm4OK0WgkLi4OgNTUVHJzc4mNjQUgKiqKkJAQXnnlFVauXMmRI0cACh3t9fPPP3Pp0qVCay9VQSwqKQiC4ODWoJKWlsbUqVNd0vJfz549m1GjRmG322/6fOfqQO+hIi07z93ZEARBcDu3BpXQ0FCOHz9+033mzJnDnDlzit3HnXQeoqYiCIIANXjyY3Wi14o+FUEQBBBBpULk11Ts4jkSgiDUciKoVAC9hwoZMJmrd9+PIAhCZRNBpQI4n6kimsAEQajlRFCpAPnL34sJkIIg1HYiqFQAnVY8U0UQBAFEUCkz+brncOvF0x8FQRAAEVTKxHJqN2fmP4JsNQHi6Y+CIAj5RFApEwm7IRt75gXgn+Yv0aciCEJtJ4JKGSj8QgCwZ5wHrm/+srgtT4IgCNWBCCploKgTDAqlM6iolAo0KgVGk5inIghC7SaCShlIChVq//rYM1KdaTqtCoNJ1FQEQajdRFApI01gQ2yZ552vHQ/qEjUVQRBqNxFUykgdGIqcfRnZagYcQcUo+lQEQajlRFApI01gKMgy9qxLwN9PfxSjvwRBqOVEUCkjdWAoAPa/m8D0WtH8JQiCIIJKGakDGoAkuQwrFs1fgiDUdiKolJFCpUHyCXaOABPNX4IgCCKolIvSt4FL85fVJmOxiiYwQRBqL7c+o/7MmTMsXbqUgwcPkpiYSHh4OKtXr77pMbm5uXz22Wds27aN5ORkVCoVbdq0Ydq0abRp06aKcu6g8AvBejYe2W51WVSyjpeySvMhCIJQXbi1ppKYmEhcXBxhYWE0bdq0RMecP3+e77//nm7duvH+++8ze/Zs7HY79957L4cPH67kHLtS+DUA2YY96zI68UwVQRAE99ZUYmJi6N+/PwDTp08nISGh2GNCQ0PZsGEDOp3OmdatWzf69evHV199xezZsystvzdS+DUAwJ6Ril7bBBBBRRCE2s2tNRWFovSX1+v1LgEFwMPDg6ZNm3L58uWKylqJKHzrAxL2zPPOmopRPFNFEIRarNS/6mazmczMTJe09PR0Fi1axDvvvEN8fHyFZa6kDAYDR48eJTw8vEqvK6k8kLwDsWecF48UFgRBoAzNX//+9785efIkK1asAMBoNHLPPfdw7tw5AJYtW8ayZcu4/fbbKzanN/G///0Po9HI2LFjq+ya+RR+DRxBRasGRFARBKF2K3VQ2bt3L6NGjXK+Xr16NefOneOTTz6hVatWTJo0iY8++oglS5ZUaEaL8uuvv7Js2TL+7//+j7CwsFIfHxDgVeZrBwV5k9agMdl7jhBa3xsApUpJUJB3mc9ZE9zq5buZ2lp2Ue7apTzlLnVQuXr1KvXq1XO+3rRpE7fddhs9e/YEYNSoUSxevLjMGSqNHTt28PLLLzNp0iQeeOCBMp0jLS0Xu10u9XFBQd5cuZKDxSMQ2WYhO+UMCknicto1rlzJKVNeaoL8ctdGtbXsoty1S3HlViikm96Ml7pPxdPTk+zsbACsVit79uyhe/fuzu1arZbc3NzSnrbU4uPjeeqppxgyZAgvvPBCpV+vKPlPgZQzL/69/pdo/hIEofYqdU0lMjKSH3/8kS5durB582YMBgN9+vRxbj979iyBgYEVmccCkpKSeOSRR+jQoQOzZs1CkqRKvd7NOEaAgS0zFZ2HP0YRVARBqMVKHVSeffZZHnroIe6++25kWWbw4MFERUU5t2/cuJEOHTqU6FxGo5G4uDgAUlNTyc3NJTY2FoCoqChCQkJ45ZVXWLlyJUeOHAEgLS2NSZMmoVarefjhh10mPGo0Glq3bl3aIpWLpNEhefr/PQKsLgYxpFgQhFqs1EGldevWrF27lv379+Pt7U10dLRzW3Z2Nvfff79L2s2kpaUxdepUl7T817Nnz2bUqFHY7XZstn/W0zp58iQXLlwAYMKECS7HhoSEsHnz5tIWqdzyR4DpPDqK5i9BEGo1SZbl0vdS30LK21EPkLfzGyzHtrKszlNcyjTyxqTOFZ3NaqO2dl5C7S27KHftUuUd9UlJSWzcuNElbc+ePUyaNIkxY8bw+eefl/aUNZ7CrwFYzQSqDKL5SxCEWq3UzV9z5sxBkiTnml0XL15k8uTJeHh4EBAQwFtvvUWdOnUYOXJkhWe2usofARYkZWA06d2cG0EQBPcpdU3lyJEjdOrUyfn6l19+QZZlfvnlF3777Tf69OnD119/XaGZrO6Uf48A87Onk2e2YbPb3ZwjQRAE9yh1UMnOzsbf39/5Oi4uji5dulC3bl0A+vTpw+nTpyssgzWBpPVC0tWhjjUNAKN4Vr0gCLVUqYNKQEAAqamOR+hmZWURHx9Pt27dnNvNZjO1se9f4dcAT/MVADFXRRCEWqvUfSrdu3fnq6++wtvbm127dgHQr18/5/bExETq169fcTmsIRR+DdBe2g7IorNeEIRaq9RBZdq0aSQnJ/PWW2+hVqt54YUXCAlxdFSbTCbWrVvHnXfeWeEZre4UfiEobSbqSEayrpndnR1BEAS3KHVQCQgI4JtvviEnJwcPDw80Go1zmyzLLFu2zGXBydpC4et4CmQDdRbHzmTQtmmAm3MkCIJQ9cr85Edvb2+XgAKOxSRbtmyJr69vuTNW0+Q/WjgqwMSh5DQ350YQBME9yvSM+szMTBYvXkxcXJyz0z4kJIS+ffvy8MMP186govNB0noTrr/GDyevkZ6dh7+P1t3ZEgRBqFKlrqlcuHCBkSNH8umnn6LVahk4cCADBw5Ep9OxZMkSRo4c6Vybq7ZR+NYnSHI8ajkhOd3NuREEQah6pa6pvPvuu2RlZfHFF18UWDhy7969PProo7z77rvMnTu3wjJZUyj8GqA6tQdfLzUJp9Lo1a6Bu7MkCIJQpUpdU9m+fTvjx48vdCXijh07MnbsWLZv314hmatpFH4hYLrG7WE6Dp/OEDPrBUGodUodVPLy8lxm1N8oICCAvLy8cmWqplL4NwSgh+IgRpOFU+ez3ZwjQRCEqlXqoNKsWTN+/fVXzOaCczHMZjOrVq2iefPmFZK5mkZZPwJ16xj8U3/nbv1uEpLEKDBBEGqXUvepTJ48malTp3L33Xdz77330qRJEwCSk5P57rvvOHnyJPPnz6/wjNYEkqTAo/s4UGnoFR9LQuJPyD1fQFKUeeS2IAhCjVLqoDJo0CDeeust3nnnHd544w3n8+FlWSYwMJA5c+YwYMCACs9oTSFJEh6d7+HEBQORV7aRs/FjvPtPRlKUafS2IAhCjVKmX7rhw4czbNgwEhISOH/+PAANGjQgMjISlUr8eEqShHeX0az6Ppe7Tu8mb6MVbb/HkZRqd2dNEAShUhUbAfKDRmGCg4MJDg52vr58+bLz3w0a1O7htGF1vXlfak+Ijx+3n96Ecf18dAOeQlJ5uDtrgiAIlabYoBITE+Ns4iqNo0ePFrvPmTNnWLp0KQcPHiQxMZHw8HBWr15dovOvXLmSjz/+mNTUVBo1asSTTz7J0KFDS53PyqJQSLRp4s/K09B1wEOYf/+cvK1L0PZ7okzvpyBUJ+YjW1D41kfVoKW7syJUM8UGlVmzZlXaj2BiYiJxcXG0a9cOu91e4uewxMbG8tJLLzF58mS6d+/Oxo0bmTZtGp6envTu3btS8loWkU382XXkEhf9OlG/8zVMu37AErQWTbvqE/wEobTshixMO74AtRbPUa+j8Alyd5aEaqTYoDJq1KhKu3hMTIzzWffTp08nISGhRMfNmzePwYMH89xzzwHQpUsXTp06xYIFC6pdUAE4dCqNRl2HYLuSjGn3jygCG6MKae3m3N0aZFnGcngTyvoRKAMauTs7tYI1eQ/IMthtGDd9iP6uGUhK0ZcqOLh1rKuiDENtz507x6lTpxg2bJhL+h133MGhQ4dIT68+a27V8fKgUV0vEk6lIUkS2t6TUPg2IG/jh9hzrro7e7cE++UkTDu/wvDz65jj1yHLYhWDymZN2o3CLxRt38nYryRj2vW9u7MkVCM1bgLFqVOnAGjatKlLerNmzVy2VxdR4QGcTM3GkGdFUmvRDXwa2W7DuGEBslU8zKu8LKf2gEKFMrQNpj+/xbj2PeyGTHdn65Zlz03DdvEEqqbRqJt0RB05AEvCBiyn/3J31oRqosYFlaysLAB8fHxc0uvUqeOyvbqIbOKPXZY5eiYDAEWdeuhiHsV+9Qx525eVuB9JKEiWZazJe1GGtkE36Bk8eozHduEEhp9exXr2YOnPZ7dhz60+Nd3qyHpqNwDqpp0B8Oj8LxRBTcjbuhR79pVCj5FlGevZePJ2foPpr5WYEzZiOfkn1pTD2K6ewZ6XU2X5FypfrW8IDQjwKvOxQUHexe7j5++JbvkhTl7IZnCP8L8P7Em64TyZv/+AR5PW1Ok4uMx5cIeSlLsq5J0/SW5uGv5978M72AeCh2Nu3YHLK/+HMfZ9fDoOxb/fOBQqTbHnspsMXPzxXfLOHSX04bloghoWul91KXtVyy936pm9eNRvSt2/WwYALGOeJ3XpC1jjFtHgwTdd5mMZzx4hY+s35J07iqTSFF47V6iod88r6MPbVXo5Squ2f95lUeOCSn6NJDs7m6Cgf0ad5NdQ8reXVFpaLnZ76WsLQUHeXLlSsjusVmF+7D1ykcuXs/9ZgaDlYJRnjpO2/lMMHsGo6tWM9dJKU+7KZtoXB5ISo38r8px58kVzxyvIu38ie+8ack8dQtv/cZS+Rc+bshuzMa6diz0tBRRKLm78Gl3/JwvsV53KXpXyy23PuoTpQhIeXe654X3wRNNrInkbFpKy+lO03e7HdvUMpj3LsZ2LR9L74tFjPOoWvUCSkE3XkPNykfNykPNyMe3+gUu/LcJz9BtIJbgBqCql+bztOVewZ11CFRpZybmqfMWVW6GQbnozXuOav8LDHXf7N/adJCUluWyvTiLD/UnLNnEhzeBMkyQFur6TkbwDMa77H7arZ9yYw5pHlmUsp/agDGmF5OHpsk1SadB2ux/d4GnIhkwMK2ZiOf57oU2N9pwrGFb9F3vGBXSDpqJpNxjrqT3i8yiEJWkXAKrwgo+9+Kd/ZT2G397GsOI1bJeT0ET/C89730LTOgZJqUJSKFHofFD6NUBVvwXqJrej7T4OOfsS5vi1VV2kCiFbTBjWvItxzbtlanYtL3v2ZSzHtiHbrFV+7cLUuKDSsGFDwsPDWbNmjUv66tWriYqKuumy/O6SP7R420HX1QkkD0/0w55HUmsx/PY2trRz7shejWRPO4uccwVVeKci91E1aov+7tdRBjclL24peVsWI5uNzu229BQMv/wX2ZiDftgLqBq1RRM1CDR6THtXVEUxahRr0m6U9SJQeAUUut3RvxKO7VISmg534XXfO3jcNrTYVSRUoZGowqMx719dZL9MdWb681vkrMtI3kEYtyzGnlt1q5PLNgvGdfPI2/YphhUzsV06WWXXLopbg4rRaCQ2NpbY2FhSU1PJzc11eQ3wyiuv0Lq165yOKVOmsHbtWt5//3127drFrFmz2LFjB08//bQ7ilGswDo6erVrwPo95ziQ6DqUWOEdhP6Ol5BUGoy/vY0tPcVNuaxZrKf2gKRA1bjDTfdTePqhG/oCmo6jsCb9ybUVM7FdPY3tYiKGVbMA0N/1Csq/mx8lD0807YZgO3uwWvyBVhe29BTsGSmomhaspeSTlGr0d07Ha+z/8Og4CkmjL/H5PbreBwoleTu/qlGDV6xn9mM5uhV128Hohz4PdjvGjR+UqdYg263Yrp7GcmxbiaccmPf+jD0jFU2Hu5DNBgy//Je87V8gmw3FH1xJlDNnzpzprotfunSJ++67r9Cg0rJlS1q1asWGDRs4cuSIS8Bo3rw5ISEhfPfdd3z11VdkZWXx73//u0yrIxuNZsryHfb09MBgKPmQ4DZN/DiUlM72+At0bBWMp/afzkzJwxNV2G1YEndiPbEdZaN2KHTu6yA0HViDadf3qJtGF5jUVtpyVwZZlsnb8QXKgEZoWscUu78kSajqt0DZoBXWU7uxJKzHkvgHkmcd9HdOR+lb32V/ZWAYlmNx2DMvoI7o7kyvDmV3B09PDzJ3/4btUiLaXpOQ1Noi95UUyjItnCppdEhKJZbDm1AEhhX4TNyhuM/bbsjCuHYuijp10fV7DIXOB4VvPSyH1iFbjKgatr3p+e2GLKwph7Ge2I55/ypMO77GcmQT1jP7sSbvRdU0GkmjK/J426WT5G37FHXLXmi73o+6ZS9kqwnL4c1YTuxA8g5C6Vf6NRiLK7ckSej1Rfd9SXJNui2oBFXRUZ/vSqaR/3y2h0BfLTPG3Y5apXTZbs+8gOHX2YCE/s7pKNzwh2W9mIjx11kgy6hbx6DtMd5le3XorLaln8Pw06t49BhfoqByPTkvl7zfP0c2ZqMd8BQKnU+h+5nj12H681t0d7yEqkEroHqU3R0CA704vfAJR6162AuVdh3ZbsWw/DVkSx6e/5rl9sVXb/Z5y7Ls6AtNPYx+5EyU/qHObXk7v8GSsB5t/ydQF9L/JJuNmPb+jOXwBsfKBAqlI5AGN0UZ3BRJ64Vxw0IUPkHo73yl0MAiW01cW/5/YLPiOfpNl31sl0+R9/tn2NPOoWrcAW3fR5HUJX8vy9tR79aaSnVQVTUVAE+tmgZBnqzfc47sa2Zua+66ZpKk9UbZqC3WE9uxJO5EFdYeSVv2Ic+lJZuNGNe8i6TRow6PxnJ0i6MN/bq1narD3brl8GZsF06g7TXxpnfNhZFUGtRNo1G36HnTPzRFQCMsx393/GFG9ECSpEotu2wxYT17EIV3QLV79o4q6xzZu35B0/4OlIGNK+06kqRA4R+KJWE9yBS5lJEs20u8HqFss2L68zvyti4BjQ5FYKMSH3uzz9tyLA7LoXV4dLkXdeP2LtuUDVphTT2M5dg21E06Ov+GZVnGemr338HoCOqWvdF2H4tHtwfQtI5B1bAtSv9QFD7BKIMaYzm0HtvV044ai+TaU2H683ts5+LRDZxSoFan8PRD3bIXqDywHN6EbDaialTy4drlramIoFKFQQWgnr8eq83Oxr0p+Pt4EFbXtZlLofNB2TAK6/HfMR/ZjJyXg6JO3QIjnCqDafuX2C4cQzfoGdQtemI5tQdr8l7ULXo5mzSqQ1Ax7fgShV8DNJH9K+0akkIJKg2WI5tR1m2Kok7dAmW3Z14k78/vsZ7YgfXMAWznDmFNPYLtwglsl04iaXQo9MUPcZdlmbwtizD/9TOWo3HINitK/9ByD6+VZTv2zAvYzsVjObYN075fMO//Fcup3dhSErBdTsKekYr9WgbYbUXm1ZKwHtOFk+h6T6r0Ib8K70Ds2VewHNuKKrwTCq3j70O227ClHsG07xfyti7BmvQnCt8GKLwDizyXPfsyxtj3sJ3+C4WXP9YT27GdP4oiOLzIGur1ivqu2zMvYtwwH2X9lnh0H1sgSEkKBarQSKzHtmFNOYQ6ogdy9hXyNn+M5eAaR3PZwCloWvdF4eXv+K7d+D74BKPw9HM0pV3LQBnW3nkd6/mjmLZ/gbpNfzRt+hWad0lSoKoXgWw2Yjm8AWXdZih8ggvdt6Tl/ufcIqjcVFUHFYCWjfxITMli64HztGsaQB0v1ztmhc4HVVh7ZEMm1uPbsSRswH71DJLOB8k7sFJWjbac3od59w9obrsDTYueSAql424pYT1yXg6qMMfdmLuDii3jPOa/VqJpOwRlcOUOH1f4N8SSuAPb5VOoW/Z2lt0xt+LHv2eRXwRk7DlXsaenYruSjO3CMWznj2BJ+hN1447F1jYtx7dhOfAb6tYxoPLAemyr44bClIvCL+Sm7eoAsumaI3hcSsJ27hCWk39iPrgG086vsRxah/X0PuwZqSg8/VD4h4DNgj3rIrbUI9jOHsR6ag+Wo1uwXT6FMqiJS35l2Y5hy1IUdZujadmrIt7WYinrNsNydAv2tHMofOs7yhK31JGWfRVV49uxZ1/CkrAee9YllHWbFaixWpL3Yox9D9mUizbmcTy6j0XhHYjl5B9YEjaA3eZoairkBz1fYd912W7FGPs/ZIsR/dDnURQxGEHS6FEEhGI5tB5rSgLmv1YiG7Px6Hov2p4PofAqfpSqMjAMZNlRc5MUqBq0/Kc1QefjeD5TMbVaZf0WWJP/wnpqt6N2XoKbAtGnUk5V2adyvexrZmZ+thu1SsFrEzqh1xbeuWm/loHlyGYsR7c6ai1+Iahb9kbh1wDJ0w+Fp3+xPzrFsRuyMPz0byRPf/QjXnXpnDft+gHzwTXohkxD1bBthfQryKZrGLcsRjZk4XH7CJSN2pU4UJr2rcK8dwWeD7yPwtOvXPkoCcvx38mLW4p24NPU79Cd83G/YNq3CswG1C16o+k4stA7fHv2FQwrX0fSejne0yJ+fGyZ5zGsmImybjN0Q59HkhTY0s5iPrgGa9JukCRUTTsj6XzAbEQ2G5Etec5/26+lw40jfdRaFL71UQY1QRnUBEVQExS+DZBuWMBVlmUwXcN+LR1bSoKjXDYLmqhBaNrfiaTRYb14AuOqWWhjHkXdrGuFva/FMR/ehGnHl44XChWqRu1QNeuCqlE758x884HfMB/8DRQqPG4fiTqyH8iy4xETCRtQBDVB1+8Jl+ZbuzEb0x/fYj35B1Kdemh7PujsM7tRYd910+4fMR/4DW3/J1HfZDi7c/+9KzDvW4WqWVc8utyDQu9bqvdBlmVHzSxxB9o+D2O7mIjl+Db0d/4zYrE4tivJGFa+gapZF3R9Jxe7f3n7VERQcVNQATiZksVb3+yjaUgdnhoVhZeu6FEzstWM9dRuzIc2YE+7YWKeRofCMwCFTxCaTqNR+oeUOA8uHY6j/oPSz/VY2WrG8PNMZLMRz9FvEhxar1zltmWkYlw3Hzn3KpKnH3LOVRR1m+HRaXSJHvh0bfmrSCot+uEzypyH0pDtNq79OANkO0qlAmvGRZQhbfDoei9K/8KXcslnPX8M42/voGwYiW7g1II/6jYLhpVvIOemox/9RoEgac++gjk+FsuJ7cDfI6TUWtDo/v63Dknvi8I7EMk7EIV3kKM5yMOzTLVZuyET0+6fsJ7Y7pgF3/lf2C6dxHpiO57j5pe6/6o8ZLsd875fkLz8Hf0SRTT/2rMukbfza2zn4lH4hYBSjf3qadSRA/Ho/K8il+S3piSQ9/sy5JwraDoMR3P7iALv2fV/47IsY96zHPOB1Y6+kF4PlawcsoxsyCzXDZBsszqa8c4fA9mOpt1QPDr/q1TnMO39GfO+X9AOeAp1k4433VcElXJyZ1AB+PPIRT797Sj+3lqmjmlL/YCb953Isoycm4Y9Nw35Wjr23HTk3HTka+nYLiYCoBv2fIk7VM1Ht2L6/XM8ut6PJmpgofvYLp/C8MubqJp3pdGYaWUut/X0foxbFiGpNOgGPI0iuAmW49sx7/vF0W4cGolHp7tRBjUp9Hh71kWufT8dj673OSYpVhFL0i7yNn2EOjAUVcd/oWwYVeIf7fw7bs1td+ARPdplW/4oId2gZ1CF3VbkOWRZrtKnhdounSRv59fYryQD4Nmq7Ga1VAAAIABJREFUK4qej1bZ9UtLlmWsZ/Zj2vk1stmIts8k1I1vL/44q4m87V9hPfE7qqZd0Pae6NI8lP83Lst2TDu+wnJkM+qWffDoMb7ADUJlk82Gf0aGjni11MO2ZbsVw8o3kXPT0I9+86Z9fWL0Vzm5o0/leqFBXrQK82dHwgW2HjhPWD0vgv2KnjQmSRKShx6FdyBK/4ao6kWgatQOdbMuqJrcjiVpF5YjW1HVb1HkzOd89qxLGNcX3eGYT+HpB3YrlsMb0dQLx6IrunO0MLIsY97/K6bfl6EIaIh+2EuOjmhJgTKoMerWMUgenliTdmNJ2IDtSjJICscd+HV3muZjW7GlHkHb66FSTawrL6V/KKqw9jQYOJY8jX+pfuAVQU2QDZmO5hjf+s6hp9azBzHt/NrR2VpEMM9X1Y+fVnj5o27ZC4VXAPbsywT0uRezunRr6lUlSZJQ+tZH3ToGTWR/lAFhJTtOoXL0FarUWBLWYz1/FFXYbc5RgZ6eHlzLNTian47/7qghdLu/ygMKOCaWqlv2Rt2yV9nmAUkKlPWaYzm8ATnzgqNJtYjvleioLyd3BxUAfx8tnVoGcygpjQ17UvDUqQlvUPzolBtJWi9UTTpiTd7rGLUU3LTQR73KdjvWxJ3kxS0Bu93R4ehx8x9pZb0IrGf2Yzi6HdmYg2w2OJpD1Lqb/ujJFhN5WxZhObwJVbOu6AY+jeKGjmtJoURZtxnq1n1BqcF2Zr9jQtih9X839ckovAMx7foBycsfj7ZDSv3elJdC74unl7bUn7kkSShDo7CdP4rlyBZUjRwT4oxr5qLwqYuu/xM37Sx2F0mSUAaGoWnTjzoNQt0+4q8kyjLxUpIkVPUiUPiFYDmyGWvSLpShbVDofNBpIO2X97El70UTPRqPjiOrPMC75lVRru+KQueDpNJgObzRcVMaWHjwFR315eTu5q/rGU1WPvn1CAdOXqVv+xDu698clbL0d0V2QybG397Gnn0F3YCnnGPUZVnGduYApj3LsWekoAgMQ9t9HMq6zYo5o4MtPQXbH19iupAEdscyFJLOB0VQE+dIFTkvF9mU+/f/rzma5szX8Ij+F+q2g0v0Rynb7dguJWJN2o01eS+yMQuUarBZ8Oh8D5p2VR9UoHyfud2QheHn/4AkofAOwnb5FPpRrxXow6qOasukT9vlUxjX/Q/ZakHb52E4sZm8M0fw6DGu1JNsqytZtmNc/Ra2q2fwHDOr0FFook+lnKpTUAGw22WWxyWxdtdZmofWYdzAFoQGl34CpD0vB+Nv72LPSEEb8xiSzgfT7h/5//buPLypMm38+DdNk7RNm+77QunK0pa1UkCoIgKyiAP6CiqIg8wLVXQWGGYcxw1EZwZ0EERERn6i78j4c2GwKgrqUAUBQRBafEtp6QotbVOa0iVp0rx/FCKlZSmEpsv9ua5caZ6ck953Tps75zznPE9T2XEUnoFohk7HOWpoq4uqrsTf34PTpXqaKouaT58tz6Op/ARNVadAoUCh0TafkqrRotC4o3DRoooZgXNY/3bnABcUmLx9WCoKcB37SIec9dWW693mlvITzeONWRrRjJqDuu8t9gvuBuopRQWaZ7as3/YyTfrmaRBcbpmHKibF0WHZVZOhnPptL+Fyy8MoA6JbPS9F5Tp1tqJy3u7MU7y7I4c6o5nUASHcNSoKnbZ9F55ZjbXUbXuZprJcwIrCzQv1kLtQxd98zVdtXypvq8UMTkqHHh640eyxzc1Fh7FUFKAeOLnLvFc9qajA+WFUPsQ3YRi1uqvbi+9OpKhcp85aVADO1jeyddcJvv6hBLXKiSkjenPbkDBUzle/d2FtbMC4+584eQWh6j/2uq+I7mkfMBfqqblL3j3L9RaVzjXIkGjB3VXFfWPjuHVQKP/66jjvfX2c/xws4e5bohkU54fyKs5CUahccEn9ZQdEK4QQUlS6hGBfLb++ZwCZeZVs/uo4a7dkotOqGdY3kJT+gUQGeXSZQylCiO5NikoXkhDly7OR3hzKqWRPVilfHyxm+/4ignzcGN4/kJT+Qfh7Xd+QLUIIcT2kqHQxSicnhsT7MyTen9qGRvb/72m+yyrjo29O8NE3J0jpH8i00VH4eUpxEUJ0PCkqXZjWRUXqwFBSB4ZSUV3P1wdL2LG/mP3/W87tQ8OYNLzXJQeqFEKIG0GKSjfh5+nKPbfEMGZQGB9m5PHZ3kK+OXyKO0dGcsug0Gu6iFIIIdpLPmm6GV9PF+ZN6cfTc5IJD3DnnztyeHLDXv5zqITahkZHhyeE6OYcWlTy8/OZO3cugwYNIiUlhaVLl1JfX3/F9erq6lixYgVjx45lwIABjBs3jjVr1mAydf7xiTpKryAPFs0YyK/vSULtrGTTtmx+s/pbXv3oCAePlWO2NDk6RCFEN+Sww18Gg4HZs2cTEhLCqlWr0Ov1vPDCC+j1el5++eXLrvvMM8+wY8cOfvOb3xAbG8vhw4d55ZVXMBgMPPHEEx2UQeenUChIivYjMcqXgrIadmeWsu9oGQeyy3F3VZHcN4BRScFEBrV/8EohhGiLw4rK5s2bMRgMbNmyBR+f5kHNlEolixYtIi0tjdjYtmc1M5vNbNu2jYcffphZs2YBkJKSwsmTJ0lPT5ei0gaFQkFkkI7IIB3/dWsMR/P17M4s5dvDp/j6hxKiQ3TcNiSMoX0CpO9FCHFdHPYJkpGRQUpKiq2gAIwfPx61Wk1GRsYl17NarVgsFjw8PFq063Q6eviIM1fFWelEUrQf86cm8PKjNzNzbCxn6xtZ//FRFq/dzb+/PUH1WaOjwxRCdFEO21PJzc1l+vTpLdrUajURERHk5eVdcj2VSsXUqVN5++23GTx4MDExMRw5coT33nuPBx544EaH3a24uThz+9BwbhsSRtYJPTv2F/Pvb0+QvjuflP6B3HNrDLrLzJsghBAXc2ifik7X+li+Tqejurr6sus+99xzPP300/zXf/08T/OcOXN49NFH7R5nT+CkUJAY5UtilC9l+jq+PFDM1wdL+PF4JQ+MiyO5T4AMAyOEuCpd8jqVlStXsnPnTpYtW0ZkZCSHDh3i1Vdfxc/Pj3nz5rXrtS432uaV+Pt7XHmhLsbf34OE+EDuujWWVf86yLp/Z/Fjnp4F05Lw1rnYlumpemruknfPcj15O6yo6HQ6DAZDq3aDwUBUVNQl1zt27Bhvvvkma9eu5bbbbgMgOTkZs9nMK6+8wsyZM3F3v/pC0ZmHvnckN2cFv585kC++L+KjjBMczvmSGbfFMvXWWCoqzjo6PIfo7tv8UiTvnuV6h753WEd9dHQ0ubm5LdpMJhOFhYWXLSrHjx8HoG/fvi3a+/Xrh8lkoqyszP7B9lBKJyfuGNaLZ3+ZTLCfln988hPPbthDSUWto0MTQnRSDisqo0ePZs+ePVRVVdnatm/fjslkIjU19ZLrhYY2z+mdlZXVoj0zMxOFQkFISMiNCbgHC/bV8of7BjPztliOntDz1Ia9rP84izJ9naNDE0J0Mg47/DVjxgzeeecd0tLSSEtLo7KykhdffJGJEycSE/PzFJ5PPPEEW7Zs4ejRowAkJCSQlJTE008/TWVlJb169eLw4cOsX7+e6dOn4+oqo/PeCE5OCm5PDmdyagzvfJrFlweK2Xf0NCMSg7hzRCR+MuS+EAIH96m89dZbLFu2jIULF6LRaJg0aRKLFy9usVxTUxMWi8X2WKlUsm7dOlatWsX69eupqKggODiYX/7yl/z3f/93R6fR4+i0au65JYZxyRF8+l0BXx8s4bvMUkYPCGHcTeEEers5OkQhhAPJHPXSUd8uF+etNzTwyXcFZPx4EkuTlb69vEkdGMLgOP9ud3W+bPOeRfJum8xRL24oH50Ls8bHM2VkJN8cPkXGoZOs+3cWOjcVI5OCSR0QQoDsvQjRY0hREXbh5a5hyohIJqX0Iitfz38OlvD53iI+21PIwBg/pt7cm15BPfOcfyF6Eikqwq6cnH6+Or+qxsjOQ82zUT77/75nUGxzcYkIlOIiRHclRUXcMN4eGu4aFcW45Ah27C/i8++LOLjxe4bE+XPnzb0JD7j20QyEEJ2TFBVxw7m5OHPnzb0ZOzSML74vYvv+Ig4cKyehtw8DY/1IivbFz1NOSRaiO5CiIjqMm4uKu0ZFcXtyONu/L2JPVhnvfHEMgFA/LUkxvgyI9iM6VIfSqXudOSZETyFFRXQ47bniMvXm3pTq6zicW8nh3Eq+2Nfcse+mcaZfpDf9e/uQ0NsXX08XR4cshLhKUlSEwygUCoJ9tQT7ahl/UwR1DWaO5us5nFdJ1gk9+7PLAQjycaN/bx8So3xJiPLBSYbhF6LTkqIiOg03F2eG9glgaJ8ArFYrpyrryDyhJ+uEnm9+PMmXB4qJCHDnnjEx9I/0ufILCiE6nBQV0SkpFApC/LSE+GkZlxxOo9nC/uxyPsrIY+XmQyT09uGeW2PkDDIhOhkpKqJLUDkrGd4/iKHxAXx5oJj03fk88+Y+RiYGc9eo3vjopN9FiM5AioroUlTOTkwYFsHNScF88l0+Xx4oZu9PZUSH6Ajy1RLs40awrxtBvm746Fyk/0WIDiZFRXRJ7q4q7h0Ty22Dw9i2r5CCshr2HS2jzmi2LaN2diIuwovh/YMYHOuPRq10YMRC9AxSVESX5uflygPj4gGwWq3U1DVyqrKWU/o6SsprOZRTzhsfH0WjVjIkzp8RCUH0ifDGyUn2YIS4EaSoiG5DoVCg06rRadXER3gDMHNsLDlFZ9idWcr+7NPszizF20PDoFg/egV5EBHgQai/ttsN0y+Eo0hREd2ak0JBfIQ38RHe3H97HIeOV/BdZim7jpTy1Q8lACidFIT6aYkI9CAy2IP+vX1ksjEhrpEUFdFjqFVKbuobyE19A2lqslJWVUdh2VkKT9dQWHaWH3Mr+PbIKQACvF1JivIlMdqX+HAv1CrpjxHiakhRET2Sk9PPV/MP6xcINPfJnD5Tz5HcSo7k6dn540l2HCi2dfj3i/LD201FqL+WIB83OWQmRBukqAhxjkKhINDbjcChbowdGo6p0UJ20RmO5FbyU0EV73+VY5t6WumkIMjHzVZggnzcCPRxI9DbDTcX+bcSPZdD//rz8/NZunQpP/zwAxqNhkmTJrFo0SJcXa88DHpNTQ2vvPIKn3/+OXq9noCAAKZOncrjjz/eAZGLnkCtUtomHAPw8nbjSPZpSsrPUlJRS0l5LXknDXz/02msF6ync1MR6ONGyLl+mohAd8L83dHIITTRAzisqBgMBmbPnk1ISAirVq1Cr9fzwgsvoNfrefnlly+7bl1dHQ888AAKhYLFixcTEBBAUVERpaWlHRS96IlUzkrCA9xbDQ3TaLZwuqqesqp6yvR1lJ67ff/TaXYeOgmAQgHBvloiAt0J8dXiqVXj6d58pprOrfleDqeJ7sBhRWXz5s0YDAa2bNmCj0/z4IBKpZJFixaRlpZGbGzsJdddv349NTU1fPzxx2i1WgCGDRvWIXELcTGVs5JQf3dC/VsWG6vVSmV1AwVlZyksq6GwrIbswjPsySpr83U83FQEn9vDCT437lmIrxYvdzUKGRlAdBEOKyoZGRmkpKTYCgrA+PHjeeKJJ8jIyLhsUXn//fe57777bAVFiM5IoVDg5+WKn5crQ+L9be3GRguGWhOGWhPVF9xX1TRwsrKOfT+dbjEygKtGiY+HC57uarzcNc33Wg1eHhp6B3vIrJmiU3FYUcnNzWX69Okt2tRqNREREeTl5V1yveLiYsrLy/H29mb+/Pns2rULjUbDmDFj+NOf/oSnp+eNDl2I66JRKfH3csXfq+1iYLVaMdSaOFlRy8nKOk5V1lJVY6S61kR2YRVnzpqwNP3cixMe4M6AGD/bBZ0y3plwJIf2qeh0ulbtOp2O6urqS65XUVEBwF//+lfGjBnD66+/TklJCStXrqSyspJ//OMfNyxmITqCQqHA012Dp7uGvm3MG9NktVJb34jeYOSngioOHa/gk+/ySd+dj6e7moExfkSHeKLTqvBwU+Ph2nwvY5+JjtDlzn1samoCoFevXqxYscJ2rNnDw4PHH3+cw4cPk5SUdNWv5+t77fNx+Pt7XPO6XVlPzRs6V+5RwNDEEGYBhloT+38qY19WKft+KrOdIHAhtUqJn6cLvUM86R2qIzrUi94hOnx0Llfss+lMeXckybv9HFZUdDodBoOhVbvBYCAqKuqS650/vDV8+PAW/wjDhw8HICcnp11FpbLyrO3ag/bw9/egvLym3et1dT01b+j8uSf28iKxlxcPjo9Db2igpq7x3M1ETX3zfUV1AzmFVew6/HPR8XBTEeqnxVXjjMrZCbWzEpXKCfW5n8OCdagV4KNzwVenwVXj3CNOHOjs2/tGuVLeTk6Ky34Zd1hRiY6OJjc3t0WbyWSisLCQadOmXXK98PBw1Gr1JZ83Go12i1GIrshZ6USAtxsB3pdept5opuj0+bPSznJKX8vZMw00mi2YzE00mpswmS00NjZx8VcuF7XyXIFxwc/LBT9PF/w8XfHzdMHX0wVXtTPGRgtGk4WGc/dGkxmVs5LeIR4oneTU6e7MYUVl9OjRvPbaa1RVVeHt3fzXv337dkwmE6mpqZdcT61WM3LkSHbv3o3VarV9Y9q1axcACQkJNz54Ibo4V40zceFexIV7XXa5piYrzi4qck5Uoq8xUlndgL6mofneYCTvZDW1DebLvsaFPNxUDInzZ0ifAPpEeEmB6YYUVqu1/cd+7MBgMDB58mRCQ0NJS0ujsrKSF198keHDh7e4+PGJJ55gy5YtHD161NaWmZnJjBkzuP3225k2bRonT57kpZdeIiEhod0d9XL4q316at7Qc3O/Ut51DWYqDQ1UVNdTcaYBY6MFjVqJi0rZfK9WolEpMdQ1ciD7ND8er8TYaMHdVcXgOD8Gxfrj7aHBTeOMq4szrmrnTjHfjWzvtnXaw186nY633nqLZcuWsXDhQtswLYsXL26xXFNTExaLpUVbQkICGzZsYOXKlaSlpeHu7s7EiRNZtGhRR6YghADcXJxxc2k90kBbkvsEYGq0cCRPz4Hs0+z76TQZP55qtZxGrcRN44yHq8o2R47nuXudVo2r2hlnpQKl0glnpQJnpRNKJwUatRJPrRoXdZc7B6nbcNieSmcheyrt01Pzhp6b+43Mu9FsIbfEwNn6RuqNZuqNZuqMZuqNFuqMzScaGGpNGOqaLxI1W67uf1WjUqLTqvDUas4NhaPCVeOMi8YZN40zLmolrud+DvB2xdtD0+rkA9nebeu0eypCCKFyVtKn12XOKLiA1Wql3mimutaEsdGC2WLFYmnCbLFiPnffYDK3GqmgVF9HTrGJeqMFs6WpzdfWqJQE+rg2T4fg40aQrxthZxqorzOiUTUfylOfO4zXaG6ipq7liAiGOhMqpRODYv0J9dde9uy4s/WNfP+/pzmar6dvL29GJgZ3q8FGpagIIboEhUKBm4sKNxfVNb9Go7mJBpP53F6RhdqGRsqq6imtrOOUvpbckmr2HS1rdcbblWMDrPDRNycI9HFjaLw/Q+MDiAh0R6FQYGq0cOh4BXuyyjiSV4mlyYqHm4oD2eVs+eYEtw4KZcyQMDy1lz6ztauQoiKE6DFUzk6onNV4uP384d0vsuUypsbmUaedXVScLq/B2NjUfFp0Y/PNWemETquyjS6t06pxd1VRU9fIwWPl7M8+zWd7CvnkuwL8PF3oFeRB1gk9DSYLXu5qxg4NI6VfEBGB7uQUV/P5vkLSd+fz2d5CRiQEMi45ghC/rjuuofSpSJ9Ku/TUvKHn5i55t9/Z+vMFppzC0zUkRvkyvF8g8RHebZ7Zdqqylu3fF7Ers5RGcxOBPm4EeLkS4N18C/R2JcDbDXdXFRceWTv/o7GxiUpD86neF97XGc0MjvVnRGIQOrer2wu63j4VKSpSVNqlp+YNPTd3ybvjGGpNfHP4JAWlNc1z9Jypx2iyXHnFi2hdnPH1bB5+p6C0BqWTgiHx/qQODKVPhNdl+3yko14IIboJnVbNpOGRtsdWq5WausZzk8DVUXfBhaYXfhVWKRXNoxx4No904Kr5+aO9pPwsO388ye4jpez76TSB3q6MHhjCuOTwG3LxqRQVIYTopBQKha3fJibs2qb1CPV3576xcdydGs3+7ObZSP//17lEBHrQv41RsK+XFBUhhOgB1ColIxKCGZEQTF2DGVfNjTmNWYqKEEL0MG4uN+6jX0ZzE0IIYTdSVIQQQtiNFBUhhBB2I0VFCCGE3UhREUIIYTdSVIQQQthNjz+l+HpmmOsMs9M5Qk/NG3pu7pJ3z3K5vK/0nvT4sb+EEELYjxz+EkIIYTdSVIQQQtiNFBUhhBB2I0VFCCGE3UhREUIIYTdSVIQQQtiNFBUhhBB2I0VFCCGE3UhREUIIYTdSVNohPz+fuXPnMmjQIFJSUli6dCn19fWODsuuCgoKeOqpp5g6dSr9+vVj8uTJbS63c+dOfvGLX5CYmMjYsWN5++23OzhS+/rss89IS0sjNTWVgQMHMmXKFP75z3/S1NTUYrnulvcXX3zBzJkzGTZsmC2nv/zlL9TU1LRYrrvlfbHa2lpGjx5NfHw8R44cafHcli1bmDBhAomJiUyaNIlPP/3UQVFevw8//JD4+PhWt+eee67FctezvXv82F9Xy2AwMHv2bEJCQli1ahV6vZ4XXngBvV7Pyy+/7Ojw7CYnJ4edO3cyYMAAmpqaaGsUn4MHD5KWlsbUqVNZsmQJP/zwA8uXL8fZ2ZmZM2c6IOrrt3HjRkJCQvj973+Pr68ve/fu5fnnn6eoqIglS5YA3TPv6upqkpOTeeihh/D09CQ7O5s1a9aQnZ3Nm2++CXTPvC+2Zs0aLBZLq/Zt27axZMkSfvWrXzFy5Eh27NjBb3/7W7RaLampqQ6I1D42bNiAh4eH7bGfn5/t5+ve3lZxVV5//XXrgAEDrJWVlba2rVu3WuPi4qzHjh1zYGT2ZbFYbD8vWbLEOmnSpFbLzJ0713r33Xe3aHvyySetI0eObLF+V3Lhdj1v+fLl1sTERKvRaLRard0z77Zs3rzZGhcXZy0tLbVard0/7+zsbOvAgQNteR8+fNj23IQJE6yPPfZYi+Ufeugh6/Tp0zs6TLv44IMPrHFxcW3+vZ93vdtbDn9dpYyMDFJSUvDx8bG1jR8/HrVaTUZGhgMjsy8np8v/SZhMJvbs2cPEiRNbtE+ePJny8nKysrJuZHg3zIXb9by+fftiNBo5c+ZMt827Ld7e3gA0Njb2iLyfe+457r//fiIjI1u0FxUVkZeXx6RJk1q0T548mSNHjqDX6zswyo5hj+0tReUq5ebmEhMT06JNrVYTERFBXl6eg6LqeIWFhTQ2NhIdHd2iPTY2FqBbvRcHDhzAy8sLX1/fbp+3xWLBaDSSmZnJq6++ypgxYwgLC+v2eW/ZsoWCggIWLFjQ6rnzuV2c+/nPga6c+5QpU+jbty9jxoxhzZo1mM1mwD7/39KncpUMBgM6na5Vu06no7q62gEROcb5XC9+L84/7i7vxZEjR/jwww955JFHUCqV3T7vYcOG2TrnR40axcqVK4Huvb1ramr429/+xpIlS9Bqta2ev1Tunp6eLZ7vSvz9/Vm4cCFJSUkolUoyMjJYu3YtxcXFvPjii3bZ3lJUhLhIeXk5jz32GImJicybN8/R4XSIt99+m/r6enJycnjttdeYP38+GzdudHRYN9Tf//53evXqxZ133unoUDrMqFGjGDVqlO3xyJEj8fDwYPXq1aSlpdnld8jhr6uk0+kwGAyt2g0Gg+2bS09wPteL34vzj7v6e1FTU8O8efNwcXHhtddeQ6VSAd0/7759+zJ48GDuvfde1qxZw969e9m+fXu3zTsnJ4fNmzfz+OOPYzAYMBgM1NXVAVBXV8fZs2cvmfv5b+tdNfeL3XHHHQBkZWXZZXtLUblK0dHR5ObmtmgzmUwUFhYSFRXloKg6XkREBCqVqtWx1ePHjwN06ffCaDSyYMECKisr2bBhg63DGrp33hfr27cvTk5OFBYWdtu8CwoKMJvNzJ49m+TkZJKTk5k/fz4As2fP5v7777fldnHu5z8Humrul2OP7S1F5SqNHj2aPXv2UFVVZWvbvn07JpOpS5+v3l5qtZqUlBQ+++yzFu3p6en4+/vTv39/B0V2fcxmM48//jjZ2dm88cYbhIaGtni+u+bdloMHD9LU1ERYWFi3zXvw4MFs2rSpxe2Pf/wjAM8++yzLli0jPDycqKioVhc7pqenk5iY2OYZg13RJ598gkKhICEhwS7bW/nMM888c4Ni7VZiY2P54IMP+OabbwgMDOTgwYMsX76cMWPGcN999zk6PLupr6/nyy+/5Pjx4+zatYuKigqCgoI4fvw4rq6u6HQ6wsPDWbduHadOnUKr1fLxxx+zceNGFi9eTFJSkqNTuCbPPPMM6enpPPbYYwQEBFBaWmq7ubu7o1aru2Xec+fO5fTp09TU1HDq1Cm2b9/O8uXLCQ8P5w9/+ANKpbJb5u3q6kpYWFiLm9Fo5KOPPuLRRx8lMTERAF9fX9asWUNjYyNOTk5s2rSJ9PR0li5d2uoU5K5g7ty5lJWVUVNTQ0FBAe+88w4bN25k+vTp/OIXvwC47u2tsFrbuGRatOnEiRMsW7aMAwcOoNFomDRpEosXL8bV1dXRodlNcXExt912W5vPvfDCC0ybNg1oHsbhpZdeIjc3l4CAAObMmcPs2bM7MlS7GjNmDCUlJW0+t2nTJoYNGwZ0v7z//ve/8+WXX1JcXAxAWFgY48aN46GHHsLd3d22XHfLuy179+5l9uzZvP/++7aiAvDRRx+xbt06SkpKiIiI4JFHHml17UpX8fzzz5ORkUFZWRlms5nIyEimTZvGgw8+iFKptC13PdtbiooQQgi5RQPCAAAFWElEQVS7kT4VIYQQdiNFRQghhN1IURFCCGE3UlSEEELYjRQVIYQQdiNFRQghhN1IURGiiysuLiY+Pp7169c7OhQhZJRiIa7Ghx9+aBvGoy1vvPEGo0eP7sCIhOicpKgI0Q4LFy4kPDy8VXufPn0cEI0QnY8UFSHa4eabb2bgwIGODkOITkv6VISwo/j4eJ566ik+/fRTJk6cSGJiInfeeScZGRmtli0uLubXv/41w4YNIykpienTp7Njx45Wy5lMJtauXcuECRNISEhgxIgRLFiwgJycnFbLvvfee4wdO5aEhASmT5/O4cOHWzxfUVHBn/70J1JTU22v9fDDD7f5WkJcC9lTEaIdampq0Ov1rdovHAb94MGDbNu2jVmzZqHVavnXv/7FggULeOuttxg6dCgAlZWVzJgxg9raWmbPno23tzdbt27l0UcfZcWKFUyePBmApqYmFixYwLfffsuECROYNWsWdXV17N27l6ysLNvc4QCffvopdXV13HvvvSgUCjZs2MDChQvZsWOHbbKxxx57jGPHjnH//fcTFhZGVVUV+/btIz8/v8VrCXHNrEKIK/rggw+scXFxl7w1NDRYrVar7fEPP/xgW1ev11uTk5OtM2bMsLUtX77cGhcXZ92zZ4+trb6+3nrHHXdYR44caTWZTC1+7xtvvNEqpqamJqvVarUWFRVZ4+LirDfddJP1zJkztud37NhhjYuLs3711VdWq9Vqra6utsbFxVk3bNhgx3dGiJZkT0WIdnjyySeJjo5u1X5+TwAgISGBQYMG2R57e3szefJk/ud//ofq6mo8PT35z3/+Q//+/W1D6gO4uLhw3333sXTpUo4ePcqAAQP4/PPP8fT0bHPYcYVC0eLx+PHjW0z3en6vqKioyPb6KpWKffv2MX36dLy8vK7xXRDi0qSoCNEOiYmJV+yob2vypvNtJ0+exNPTk5MnTzJu3LhWy50vWCUlJQwYMIDCwkIiIyNRq9VXjC04OLjF44vnG1er1SxatIi//vWvjBw5kqSkJFJTU5k6dWqrdYW4VtJRL0Q3ceEkSxeyXjBl0pw5c/j8889ZvHgxHh4erF27lokTJ7J3796OClN0c1JUhLCz/Pz8S7aFhITY7k+cONFquby8PABCQ0MBiIiIID8/H5PJZLf4wsPDmTNnDuvXr+eLL75ArVazbt06u72+6NmkqAhhZ5mZmRw8eND2uKqqivT0dAYNGmQ7JHXrrbeSlZXF/v37bcsZjUbeffdd/P396d+/P9DcT1JdXc2mTZta/R5rOydtra+vp6GhoUVbUFAQvr6+tkNkQlwv6VMRoh2+/fZbCgoKWrUnJSXRu3dvAOLi4pg/fz4PPPCA7ZTi2tpafve739mWnzdvHunp6fzqV79i1qxZ+Pj4sHXrVo4fP86KFStwdm7+15w6dSpbt27lb3/7G5mZmSQnJ2M0Gtm7dy933HEHd91111XHnp+fz4MPPsiECROIiYlBrVazc+dOcnNzWbJkyXW+M0I0k6IiRDusXr26zfY///nPtqIyaNAgUlJSWL16NUVFRfTu3ZtXX32V5ORk2/K+vr68++67rFixgnfffZf6+npiY2NZvXo1t99+u205pVLJ66+/zrp160hPT2fHjh14enoyYMAAEhIS2hV7UFAQU6ZM4bvvvuPjjz9GoVAQGRnJ888/z913330N74YQrSms7d2HFkJcUnx8PPfeey/PPfeco0MRwiGkT0UIIYTdSFERQghhN1JUhBBC2I30qQghhLAb2VMRQghhN1JUhBBC2I0UFSGEEHYjRUUIIYTdSFERQghhN1JUhBBC2M3/AXVunGgGj07qAAAAAElFTkSuQmCC\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| } | |
| ], | |
| "source": [ | |
| "classifer.plot_result(history_range=np.arange(1))" | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "## Shallow network\n", | |
| "\n", | |
| "In fact, even if use a shallow bit of network and reduce the penalty, the result is still similar with tiny improving. \n", | |
| "\n", | |
| "#### Conclusion\n", | |
| "\n", | |
| "1. Deep or shallow network doesn't help so much.\n", | |
| "2. `evaluation_times` is not the key-point, when extract the plot of later iterations, the loss went worse rather than the #1 round.\n", | |
| "3. Bias, variance happens highly.\n", | |
| "\n", | |
| "The intuitive reason for this is that the input is a 300+ dimensional FEATURE vector, and this is clearly too coarse. A conjecture would be to leave H.O.G and let me enter the world of convolutional networks." | |
| ], | |
| "metadata": { | |
| "id": "cAK4OkMap7oj" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "source": [ | |
| "classifer = ImageClassifier(\n", | |
| " X_train, y_train,\n", | |
| " (X_cv, y_cv),\n", | |
| " \"paper scissors and rock\",\n", | |
| " use_mean_norm=False, \n", | |
| " hidden_layers=[2**4], \n", | |
| " L2_lambda=[.2], \n", | |
| " output_units=3,\n", | |
| " epochs=50, \n", | |
| " evaluation_times=50,\n", | |
| ")\n", | |
| "classifer.build_model()\n", | |
| "process_in_duration(\"Training shallow NN has done.\", classifer.train)" | |
| ], | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/" | |
| }, | |
| "id": "nhtgb98qqEco", | |
| "outputId": "74a48365-4c01-4123-8562-606b4f8dc101" | |
| }, | |
| "execution_count": null, | |
| "outputs": [ | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "\u001b[92mTraining shallow NN has done. used 0:22:39.030534\n" | |
| ] | |
| } | |
| ] | |
| }, | |
| { | |
| "cell_type": "markdown", | |
| "source": [ | |
| "### Plot\n", | |
| "\n", | |
| "1. Accuracy of train vs. cross-validation\n", | |
| "2. Loss convergence trend of train and cross-validation in 1. training iteration." | |
| ], | |
| "metadata": { | |
| "id": "6D5W3q-jsFJm" | |
| } | |
| }, | |
| { | |
| "cell_type": "code", | |
| "source": [ | |
| "classifer.plot_result(history_range=np.arange(1))" | |
| ], | |
| "metadata": { | |
| "colab": { | |
| "base_uri": "https://localhost:8080/", | |
| "height": 772 | |
| }, | |
| "id": "AHdT5KyuqJmv", | |
| "outputId": "07e0bbc5-c313-4148-e246-93f0d4b19e20" | |
| }, | |
| "execution_count": null, | |
| "outputs": [ | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| }, | |
| { | |
| "output_type": "stream", | |
| "name": "stdout", | |
| "text": [ | |
| "Train Mean: 0.975\n", | |
| "Train Standard Deviation: 0.023\n", | |
| "Train Report for paper scissors and rock\n", | |
| "[0.959, 0.916, 0.978, 0.975, 0.945, 0.993, 0.96, 0.948, 0.976, 0.945, 0.985, 0.962, 0.996, 0.991, 0.984, 0.988, 0.995, 0.987, 0.978, 0.987, 0.96, 0.996, 0.995, 0.904, 0.991, 0.982, 0.946, 0.978, 0.982, 0.981, 0.972, 0.98, 0.992, 0.979, 0.985, 0.981, 0.979, 0.996, 0.99, 0.987, 0.994, 0.985, 0.995, 0.966, 0.981, 0.99, 0.891, 0.985, 0.984, 0.984]\n", | |
| "Test Mean: 0.78\n", | |
| "Test Standard Deviation: 0.044\n", | |
| "Test Report for paper scissors and rock\n", | |
| "[0.726, 0.745, 0.767, 0.875, 0.819, 0.865, 0.77, 0.801, 0.821, 0.664, 0.743, 0.801, 0.851, 0.829, 0.739, 0.776, 0.837, 0.833, 0.778, 0.841, 0.794, 0.768, 0.769, 0.747, 0.821, 0.802, 0.755, 0.769, 0.782, 0.786, 0.707, 0.761, 0.794, 0.786, 0.81, 0.708, 0.737, 0.798, 0.759, 0.799, 0.802, 0.761, 0.779, 0.702, 0.757, 0.798, 0.679, 0.819, 0.796, 0.763]\n" | |
| ] | |
| }, | |
| { | |
| "output_type": "display_data", | |
| "data": { | |
| "image/png": "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\n", | |
| "text/plain": [ | |
| "<Figure size 432x288 with 1 Axes>" | |
| ] | |
| }, | |
| "metadata": {} | |
| } | |
| ] | |
| } | |
| ], | |
| "metadata": { | |
| "interpreter": { | |
| "hash": "20bf69066c0dd38d51965b69d5e1b6e387082e3198ba56e97997ac55f4e50ad0" | |
| }, | |
| "kernelspec": { | |
| "display_name": "Python 3.8.9 64-bit", | |
| "language": "python", | |
| "name": "python3" | |
| }, | |
| "language_info": { | |
| "codemirror_mode": { | |
| "name": "ipython", | |
| "version": 3 | |
| }, | |
| "file_extension": ".py", | |
| "mimetype": "text/x-python", | |
| "name": "python", | |
| "nbconvert_exporter": "python", | |
| "pygments_lexer": "ipython3", | |
| "version": "3.8.9" | |
| }, | |
| "orig_nbformat": 4, | |
| "colab": { | |
| "name": "HOG_SVM_Paper_Scissors_Rock.ipynb", | |
| "provenance": [], | |
| "toc_visible": true, | |
| "machine_shape": "hm", | |
| "include_colab_link": true | |
| }, | |
| "accelerator": "GPU" | |
| }, | |
| "nbformat": 4, | |
| "nbformat_minor": 0 | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment