CVPR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Computer Vision and Pattern Recognition - Project 3 (CNN classifier)\n",
    "#### Gaia Marsich [SM3500600]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- [Introduction](#intro)\n",
    "<br></br>\n",
    "- [Task 1](#task1)\n",
    "<br></br>\n",
    "- [Task 2](#task2)\n",
    "<br></br>\n",
    "- [Task 3](#task3)\n",
    "    - [AlexNet (1)](#alex1)\n",
    "    - [AlexNet (2)](#alex2)\n",
    "<br></br>\n",
    "- [References](#ref)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Introduction <a class=\"anchor\" id=\"intro\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This project requires the implementation of an image classifier based on convolutional neural networks. The provided dataset (from [Lazebnik et al., 2006]), contains 15 categories (office, kitchen, living room, bedroom, store, industrial, tall building, inside city, street, highway, coast, open country, mountain, forest, suburb), and is already divided in training set and test set.\n",
    "\n",
    "First of all, let's do the imports:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch import nn\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import random_split, DataLoader\n",
    "\n",
    "from torchvision.datasets import ImageFolder\n",
    "from torchvision import models\n",
    "from torchvision.models.alexnet import AlexNet_Weights\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import svm\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "from side_code import CNN\n",
    "from side_code import train_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 1 <a class=\"anchor\" id=\"task1\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The aim of this part is to train a shallow network from scratch according to the some provided specifications."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Setting a seed for reproducibility of the results\n",
    "\n",
    "torch.manual_seed(1)\n",
    "\n",
    "#  Load the dataset\n",
    "\n",
    "train_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/train'\n",
    "test_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/test'\n",
    "\n",
    "train = ImageFolder(root=train_path, transform=train_test.transform)\n",
    "dataset_test = ImageFolder(root=test_path, transform=train_test.transform)\n",
    "\n",
    "\n",
    "# Split the provided training set in 85% for the actual training set and 15% for validation set\n",
    "\n",
    "train_size = int(0.85 * len(train))\n",
    "val_size = len(train) - train_size\n",
    "\n",
    "dataset_train, dataset_val = random_split(train, [train_size, val_size])\n",
    "\n",
    "\n",
    "# Create the loaders\n",
    "\n",
    "batch_size = 32 # batch_size = 32 as required by the project\n",
    "train_loader = DataLoader(dataset_train, batch_size, shuffle=True) \n",
    "val_loader = DataLoader(dataset_val, batch_size, shuffle=False)\n",
    "test_loader = DataLoader(dataset_test, batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's see how many images we have in the three sets:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of images in the train set: 1275\n",
      "Number of images in the validation set: 225\n",
      "Number of images in the test set: 2985\n"
     ]
    }
   ],
   "source": [
    "print(f'Number of images in the train set: {len(dataset_train)}')\n",
    "print(f'Number of images in the validation set: {len(dataset_val)}')\n",
    "print(f'Number of images in the test set: {len(dataset_test)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's instantiate a model..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instantiate a model\n",
    "\n",
    "model = CNN.CNN1()\n",
    "\n",
    "# Set parameters for the training\n",
    "\n",
    "learning_rate = 0.0005\n",
    "loss_function = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model.parameters(), momentum=0.9, lr=learning_rate) # the momentum by default is 0, but I need it different from 0"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "...and do the training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH 1:\n",
      "LOSS: train: 2.7081405580043794, validation: 2.7080423831939697; accuracy validation set: 10.666666666666666%\n",
      "\n",
      "EPOCH 2:\n",
      "LOSS: train: 2.7079203367233275, validation: 2.70837140083313; accuracy validation set: 7.555555555555555%\n",
      "\n",
      "EPOCH 3:\n",
      "LOSS: train: 2.7077822148799897, validation: 2.708571195602417; accuracy validation set: 7.111111111111111%\n",
      "\n",
      "EPOCH 4:\n",
      "LOSS: train: 2.7076762557029723, validation: 2.708514928817749; accuracy validation set: 5.777777777777778%\n",
      "\n",
      "EPOCH 5:\n",
      "LOSS: train: 2.7074696362018584, validation: 2.7086305618286133; accuracy validation set: 5.777777777777778%\n",
      "\n",
      "EPOCH 6:\n",
      "LOSS: train: 2.7072836220264436, validation: 2.708883047103882; accuracy validation set: 3.5555555555555554%\n",
      "\n",
      "EPOCH 7:\n",
      "LOSS: train: 2.707004129886627, validation: 2.709536075592041; accuracy validation set: 4.888888888888889%\n",
      "\n",
      "EPOCH 8:\n",
      "LOSS: train: 2.706456732749939, validation: 2.710134983062744; accuracy validation set: 5.333333333333333%\n",
      "\n",
      "EPOCH 9:\n",
      "LOSS: train: 2.705702692270279, validation: 2.7115936279296875; accuracy validation set: 4.0%\n",
      "\n",
      "EPOCH 10:\n",
      "LOSS: train: 2.704133725166321, validation: 2.716362953186035; accuracy validation set: 4.444444444444445%\n",
      "\n",
      "EPOCH 11:\n",
      "LOSS: train: 2.7007834017276764, validation: 2.716226816177368; accuracy validation set: 10.666666666666666%\n",
      "\n",
      "EPOCH 12:\n",
      "LOSS: train: 2.694363522529602, validation: 2.6949126720428467; accuracy validation set: 6.666666666666667%\n",
      "\n",
      "EPOCH 13:\n",
      "LOSS: train: 2.6719179332256315, validation: 2.6344411373138428; accuracy validation set: 8.444444444444445%\n",
      "\n",
      "EPOCH 14:\n",
      "LOSS: train: 2.632832145690918, validation: 2.645655632019043; accuracy validation set: 15.11111111111111%\n",
      "\n",
      "EPOCH 15:\n",
      "LOSS: train: 2.5480454742908476, validation: 2.570164680480957; accuracy validation set: 23.555555555555557%\n",
      "\n",
      "EPOCH 16:\n",
      "LOSS: train: 2.490582549571991, validation: 2.4645698070526123; accuracy validation set: 20.0%\n",
      "\n",
      "EPOCH 17:\n",
      "LOSS: train: 2.397291350364685, validation: 2.308265209197998; accuracy validation set: 25.333333333333332%\n",
      "\n",
      "EPOCH 18:\n",
      "LOSS: train: 2.277894952893257, validation: 2.2490296363830566; accuracy validation set: 23.555555555555557%\n",
      "\n",
      "EPOCH 19:\n",
      "LOSS: train: 2.1200131475925446, validation: 2.1245338916778564; accuracy validation set: 26.666666666666668%\n",
      "\n",
      "EPOCH 20:\n",
      "LOSS: train: 1.99085875749588, validation: 2.048485279083252; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 21:\n",
      "LOSS: train: 1.9475576728582382, validation: 1.9861515760421753; accuracy validation set: 31.11111111111111%\n",
      "\n",
      "EPOCH 22:\n",
      "LOSS: train: 1.750426208972931, validation: 2.0117392539978027; accuracy validation set: 24.88888888888889%\n",
      "\n",
      "EPOCH 23:\n",
      "LOSS: train: 1.641465625166893, validation: 1.9120804071426392; accuracy validation set: 31.11111111111111%\n",
      "\n",
      "EPOCH 24:\n",
      "LOSS: train: 1.4750674039125442, validation: 1.9323476552963257; accuracy validation set: 32.44444444444444%\n",
      "\n",
      "EPOCH 25:\n",
      "LOSS: train: 1.376972734928131, validation: 2.261425256729126; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 26:\n",
      "LOSS: train: 1.1648697555065155, validation: 2.238422393798828; accuracy validation set: 33.333333333333336%\n",
      "\n",
      "EPOCH 27:\n",
      "LOSS: train: 0.9151898838579655, validation: 2.2565534114837646; accuracy validation set: 32.0%\n",
      "\n",
      "EPOCH 28:\n",
      "LOSS: train: 0.8732439123094082, validation: 2.2943201065063477; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 29:\n",
      "LOSS: train: 0.8078067272901535, validation: 2.881394147872925; accuracy validation set: 29.77777777777778%\n",
      "\n",
      "EPOCH 30:\n",
      "LOSS: train: 0.5028398241847754, validation: 2.8497378826141357; accuracy validation set: 33.77777777777778%\n",
      "\n",
      "EPOCH 31:\n",
      "LOSS: train: 0.3600087705999613, validation: 3.3704161643981934; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 32:\n",
      "LOSS: train: 0.2818453563377261, validation: 3.8701999187469482; accuracy validation set: 30.22222222222222%\n",
      "\n",
      "EPOCH 33:\n",
      "LOSS: train: 0.18565535489469767, validation: 4.350548267364502; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 34:\n",
      "LOSS: train: 0.14894565399736165, validation: 4.611318111419678; accuracy validation set: 28.88888888888889%\n",
      "\n",
      "EPOCH 35:\n",
      "LOSS: train: 0.09163083760067821, validation: 5.792628288269043; accuracy validation set: 31.11111111111111%\n",
      "\n",
      "EPOCH 36:\n",
      "LOSS: train: 0.09451204016804696, validation: 5.498600959777832; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 37:\n",
      "LOSS: train: 0.13860716591589153, validation: 4.6075897216796875; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 38:\n",
      "LOSS: train: 0.09226119741797448, validation: 5.274537563323975; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 39:\n",
      "LOSS: train: 0.11552973620127886, validation: 5.6583099365234375; accuracy validation set: 29.333333333333332%\n",
      "\n",
      "EPOCH 40:\n",
      "LOSS: train: 0.09076359994942322, validation: 5.691266059875488; accuracy validation set: 31.11111111111111%\n",
      "\n",
      "EPOCH 41:\n",
      "LOSS: train: 0.12841950422152876, validation: 5.868002891540527; accuracy validation set: 28.444444444444443%\n",
      "\n",
      "EPOCH 42:\n",
      "LOSS: train: 0.10831521053332835, validation: 5.42611837387085; accuracy validation set: 32.888888888888886%\n",
      "\n",
      "EPOCH 43:\n",
      "LOSS: train: 0.08950675355736167, validation: 6.072053909301758; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 44:\n",
      "LOSS: train: 0.057030558557016775, validation: 5.763001441955566; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 45:\n",
      "LOSS: train: 0.031194121006410568, validation: 5.914209842681885; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 46:\n",
      "LOSS: train: 0.014300728458329104, validation: 6.158411502838135; accuracy validation set: 32.44444444444444%\n",
      "\n",
      "EPOCH 47:\n",
      "LOSS: train: 0.007548467782908119, validation: 6.246539115905762; accuracy validation set: 31.555555555555557%\n",
      "\n",
      "EPOCH 48:\n",
      "LOSS: train: 0.0023589952383190394, validation: 6.3560285568237305; accuracy validation set: 33.333333333333336%\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Perform the training\n",
    "model_path, loss_train, loss_val, accuracies_train, accuracies_val = train_test.train_model(\n",
    "    model=model,\n",
    "    EPOCHS=48,\n",
    "    train_loader=train_loader,\n",
    "    val_loader=val_loader,\n",
    "    optimizer=optimizer,\n",
    "    loss_function=loss_function,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here, the training is performed with the support of a validation set so that, during the training, the generalisation capability of the current network can be estimated. The analysis keeps going until the number of epochs that has been set is reached.\n",
    "The algorithm compares the model at the current epoch with the best model previously obtained (the performances are based on the validation loss). At the end, the more recent model will be the optimal one among those that had been tested."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's have a look at the plot of the loss during the training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "epochs = list(range(1, len(loss_train) + 1)) # epochs started from 0\n",
    "\n",
    "plt.plot(epochs, loss_train, label='Training loss')\n",
    "plt.plot(epochs, loss_val, label='Validation loss')\n",
    "plt.title('Training and validation losses over epochs')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As expected, since the network is trained on a specific train set over and over, the loss associated to this set keeps decreasing. On the other hand, the tests done on a separate set, the validation set, report an increase of the loss."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here it is the plot on the accuracy:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(epochs, accuracies_train, label='Train accuracy')\n",
    "plt.plot(epochs, accuracies_val, label='Validation accuracy')\n",
    "plt.title('Training and validation accuracies over epochs')\n",
    "plt.xlabel('Epoch')\n",
    "plt.ylabel('Accuracy [%]')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The more the network is trained the more it is expected to perform correctly. In this graph, one can observe that the accuracies associated to the train and validation sets increase."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's evaluate the accuracy on the test set:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the test images: 25 %\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the trained network on the test set, and collect the infos for the confusion matrix\n",
    "\n",
    "train_test.perform_test(model = model, model_path = model_path, test_loader = test_loader, dataset_test = dataset_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 2 <a class=\"anchor\" id=\"task2\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The previous result can be improved using some techniques. Here, I will use:\n",
    "\n",
    "* data augmentation: the only difference with the original version is the train dataset: \"new\" images are added to the initial training and validation dataset. The \"new\" images are simply the left-to-right reflections of the original train dataset.\n",
    "* batch normalization\n",
    "* different size of convolutional filters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Remark that from Task 1 we already have dataset_train, dataset_val and dataset_test\n",
    "\n",
    "# Data augmentation for the train dataset\n",
    "\n",
    "flipped_dataset_train = train_test.HorizontallyFlippedDataset(dataset_train)\n",
    "data_augmentation_dataset_train = torch.utils.data.ConcatDataset([dataset_train, flipped_dataset_train])\n",
    "\n",
    "\n",
    "# Create the loaders\n",
    "\n",
    "batch_size = 32\n",
    "data_augmentation_train_loader = DataLoader(data_augmentation_dataset_train, batch_size, shuffle=True) \n",
    "val_loader = DataLoader(dataset_val, batch_size, shuffle=False)\n",
    "test_loader = DataLoader(dataset_test, batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now let's instantiate the model and perform the training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH 1:\n",
      "LOSS: train: 2.438126739859581, validation: 2.162184476852417; accuracy validation set: 28.0%\n",
      "\n",
      "EPOCH 2:\n",
      "LOSS: train: 1.9653183922171593, validation: 2.6591739654541016; accuracy validation set: 22.22222222222222%\n",
      "\n",
      "EPOCH 3:\n",
      "LOSS: train: 1.7284159272909165, validation: 2.4800045490264893; accuracy validation set: 24.88888888888889%\n",
      "\n",
      "EPOCH 4:\n",
      "LOSS: train: 1.5653546467423438, validation: 1.7226542234420776; accuracy validation set: 35.111111111111114%\n",
      "\n",
      "EPOCH 5:\n",
      "LOSS: train: 1.432427529990673, validation: 1.5532550811767578; accuracy validation set: 40.44444444444444%\n",
      "\n",
      "EPOCH 6:\n",
      "LOSS: train: 1.3256931141018868, validation: 2.0183494091033936; accuracy validation set: 33.77777777777778%\n",
      "\n",
      "EPOCH 7:\n",
      "LOSS: train: 1.2227705039083958, validation: 1.4687899351119995; accuracy validation set: 46.22222222222222%\n",
      "\n",
      "EPOCH 8:\n",
      "LOSS: train: 1.1274842619895935, validation: 1.8482474088668823; accuracy validation set: 34.22222222222222%\n",
      "\n",
      "EPOCH 9:\n",
      "LOSS: train: 1.0623985588550569, validation: 2.137892007827759; accuracy validation set: 32.0%\n",
      "\n",
      "EPOCH 10:\n",
      "LOSS: train: 0.9962612293660641, validation: 1.6445382833480835; accuracy validation set: 41.333333333333336%\n",
      "\n",
      "EPOCH 11:\n",
      "LOSS: train: 0.9418791778385639, validation: 1.6192799806594849; accuracy validation set: 39.111111111111114%\n",
      "\n",
      "EPOCH 12:\n",
      "LOSS: train: 0.86190185546875, validation: 1.0999605655670166; accuracy validation set: 58.22222222222222%\n",
      "\n",
      "EPOCH 13:\n",
      "LOSS: train: 0.7989915233105421, validation: 1.258381962776184; accuracy validation set: 51.111111111111114%\n",
      "\n",
      "EPOCH 14:\n",
      "LOSS: train: 0.771745228767395, validation: 1.2154011726379395; accuracy validation set: 51.55555555555556%\n",
      "\n",
      "EPOCH 15:\n",
      "LOSS: train: 0.7052845094352961, validation: 1.6583272218704224; accuracy validation set: 43.111111111111114%\n",
      "\n",
      "EPOCH 16:\n",
      "LOSS: train: 0.6727200176566839, validation: 1.7407678365707397; accuracy validation set: 35.111111111111114%\n",
      "\n",
      "EPOCH 17:\n",
      "LOSS: train: 0.6360953181982041, validation: 1.561142921447754; accuracy validation set: 46.22222222222222%\n",
      "\n",
      "EPOCH 18:\n",
      "LOSS: train: 0.5841588243842125, validation: 1.139017105102539; accuracy validation set: 56.0%\n",
      "\n",
      "EPOCH 19:\n",
      "LOSS: train: 0.5481151856482029, validation: 1.4430298805236816; accuracy validation set: 48.888888888888886%\n",
      "\n",
      "EPOCH 20:\n",
      "LOSS: train: 0.5190689321607351, validation: 1.3364644050598145; accuracy validation set: 52.44444444444444%\n",
      "\n",
      "EPOCH 21:\n",
      "LOSS: train: 0.4965594096109271, validation: 1.2978684902191162; accuracy validation set: 55.55555555555556%\n",
      "\n",
      "EPOCH 22:\n",
      "LOSS: train: 0.4880111552774906, validation: 1.171772837638855; accuracy validation set: 56.44444444444444%\n",
      "\n",
      "EPOCH 23:\n",
      "LOSS: train: 0.4257216054946184, validation: 1.2169175148010254; accuracy validation set: 52.888888888888886%\n",
      "\n",
      "EPOCH 24:\n",
      "LOSS: train: 0.4067483788356185, validation: 1.4790843725204468; accuracy validation set: 46.22222222222222%\n",
      "\n",
      "EPOCH 25:\n",
      "LOSS: train: 0.3631345549598336, validation: 1.2428830862045288; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 26:\n",
      "LOSS: train: 0.3705681266263127, validation: 1.1007195711135864; accuracy validation set: 58.22222222222222%\n",
      "\n",
      "EPOCH 27:\n",
      "LOSS: train: 0.3419547103345394, validation: 1.910424828529358; accuracy validation set: 43.111111111111114%\n",
      "\n",
      "EPOCH 28:\n",
      "LOSS: train: 0.3121259329840541, validation: 1.3540211915969849; accuracy validation set: 55.111111111111114%\n",
      "\n",
      "EPOCH 29:\n",
      "LOSS: train: 0.2924040937796235, validation: 1.2025096416473389; accuracy validation set: 59.111111111111114%\n",
      "\n",
      "EPOCH 30:\n",
      "LOSS: train: 0.2911701511591673, validation: 1.585922360420227; accuracy validation set: 45.333333333333336%\n",
      "\n",
      "EPOCH 31:\n",
      "LOSS: train: 0.2652057338505983, validation: 1.1887139081954956; accuracy validation set: 59.111111111111114%\n",
      "\n",
      "EPOCH 32:\n",
      "LOSS: train: 0.2328427605330944, validation: 1.1321569681167603; accuracy validation set: 59.55555555555556%\n",
      "\n",
      "EPOCH 33:\n",
      "LOSS: train: 0.22426460282877086, validation: 1.3653303384780884; accuracy validation set: 53.333333333333336%\n",
      "\n",
      "EPOCH 34:\n",
      "LOSS: train: 0.2160870110616088, validation: 1.1490681171417236; accuracy validation set: 57.333333333333336%\n",
      "\n",
      "EPOCH 35:\n",
      "LOSS: train: 0.20501114847138524, validation: 1.0960686206817627; accuracy validation set: 57.333333333333336%\n",
      "\n",
      "EPOCH 36:\n",
      "LOSS: train: 0.20504265632480384, validation: 1.5419178009033203; accuracy validation set: 50.666666666666664%\n",
      "\n",
      "EPOCH 37:\n",
      "LOSS: train: 0.20494542801752685, validation: 1.6709904670715332; accuracy validation set: 49.77777777777778%\n",
      "\n",
      "EPOCH 38:\n",
      "LOSS: train: 0.1860571475699544, validation: 2.1420652866363525; accuracy validation set: 41.333333333333336%\n",
      "\n",
      "EPOCH 39:\n",
      "LOSS: train: 0.16383140869438648, validation: 1.228216290473938; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 40:\n",
      "LOSS: train: 0.16020800694823265, validation: 1.1941673755645752; accuracy validation set: 57.77777777777778%\n",
      "\n",
      "EPOCH 41:\n",
      "LOSS: train: 0.15438198801130057, validation: 1.1364731788635254; accuracy validation set: 59.55555555555556%\n",
      "\n",
      "EPOCH 42:\n",
      "LOSS: train: 0.1504371203482151, validation: 1.238837718963623; accuracy validation set: 56.0%\n",
      "\n",
      "EPOCH 43:\n",
      "LOSS: train: 0.14084114478901028, validation: 1.6609243154525757; accuracy validation set: 50.22222222222222%\n",
      "\n",
      "EPOCH 44:\n",
      "LOSS: train: 0.13187773199751973, validation: 1.4021066427230835; accuracy validation set: 50.22222222222222%\n",
      "\n",
      "EPOCH 45:\n",
      "LOSS: train: 0.11765429377555847, validation: 1.2204563617706299; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 46:\n",
      "LOSS: train: 0.11867675585672259, validation: 1.207632064819336; accuracy validation set: 59.111111111111114%\n",
      "\n",
      "EPOCH 47:\n",
      "LOSS: train: 0.12175504378974437, validation: 1.4684464931488037; accuracy validation set: 53.333333333333336%\n",
      "\n",
      "EPOCH 48:\n",
      "LOSS: train: 0.13155135326087475, validation: 1.2061530351638794; accuracy validation set: 56.888888888888886%\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Instantiate a model\n",
    "\n",
    "model_data_augmentation = CNN.CNN2()\n",
    "\n",
    "\n",
    "# Set parameters for the training\n",
    "\n",
    "learning_rate = 0.0005\n",
    "loss_function = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(model_data_augmentation.parameters(), momentum=0.9, lr=learning_rate) # the momentum by default is 0, but I need it different from 0\n",
    "\n",
    "# Perform the training\n",
    "task2_model_path, task2_loss_train, task2_loss_val, task2_accuracies_train, task2_accuracies_val = train_test.train_model(model = model_data_augmentation, EPOCHS = 48, train_loader = data_augmentation_train_loader, val_loader = val_loader, optimizer = optimizer, loss_function = loss_function)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the test images: 58 %\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfYAAAH9CAYAAADoLF5/AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAACmB0lEQVR4nOzdeVzN2f8H8Ndtu7e67VKhXKkoSqVFGoSaMJbs26CQGWSZrI2lxZKdmbE0YyvG9h0SX0ykUYixF0ZIJENlSVFpvZ/fH/3cr6vo3vrcae697+c8zuMx9/M5n/fn3Fs695zPWTgMwzAghBBCiEJQaewCEEIIIYQ9VLETQgghCoQqdkIIIUSBUMVOCCGEKBCq2AkhhBAFQhU7IYQQokCoYieEEEIUCFXshBBCiAKhip0QQghRIFSxE0IIIQqEKnZCCCFERjZt2gSBQAAejwd3d3dcvnz5s/k3bNiANm3aQFNTE+bm5vjuu+9QWloq1T3VGlJg0vhiTUexHtP/7Z+sxiupKGM1HgCoq7L/q9uC34TVeH8XvWQ1HgAEmnRiPeYD4VtW4914k8VqPABQV1FlPeazonzWYyojLXUu6zHfFD9kLVbFS/ZiAYB6E0uJ8x44cADBwcGIioqCu7s7NmzYAF9fX9y7dw9NmzatkX/v3r2YP38+duzYgc6dO+P+/fvw9/cHh8PBunXrJL4vtdgJIYQQGVi3bh0CAwMREBAAOzs7REVFQUtLCzt27Kg1/4ULF+Dp6YlRo0ZBIBDgyy+/xMiRI+ts5X+MKnZCCCGKS1jFaiorK8ObN2/EUllZzV7J8vJyXLt2Dd7e3qJjKioq8Pb2xsWLF2staufOnXHt2jVRRf7w4UOcOHECffr0keotU8VOCCFEcTFCVlNkZCT09PTEUmRkZI3bvnz5ElVVVTAxMRE7bmJigtzc3FqLOmrUKEREROCLL76Auro6WrduDS8vL3z//fdSvWWq2AkhhBAJhYSEoLCwUCyFhISwEjspKQnLly/H5s2bcf36dcTGxuL48eNYsmSJVHFo8BwhhBDFJRSyGo7L5YLLrXvAYJMmTaCqqoq8vDyx43l5eTA1Na31mkWLFmHMmDGYOHEiAMDe3h7FxcWYNGkSFixYABUVydriSt1i9/f3h5+fX2MXgxBCiIwwjJDVJCkNDQ107NgRiYmJomNCoRCJiYnw8PCo9ZqSkpIalbeqqur/vw9G4nvLRcX+frj/+2RkZIRevXrh5s2bjV00QgghpFbBwcHYunUrYmJikJ6ejsmTJ6O4uBgBAQEAgLFjx4p14/fr1w9btmzB/v378ejRIyQkJGDRokXo16+fqIKXhNx0xffq1Qs7d+4EAOTm5mLhwoXo27cvsrOzZXbP8vJyaGhoyCw+IYQQGWO5K14aw4cPx4sXL7B48WLk5ubC0dER8fHxogF12dnZYi30hQsXgsPhYOHChXj69CmMjY3Rr18/LFu2TKr7ykWLHah+rmFqagpTU1M4Ojpi/vz5ePLkCV68eAEAePLkCYYNGwZ9fX0YGhpiwIAByMrKEl1fVVWF4OBg6Ovrw8jICHPnzq3RteHl5YWgoCDMnDkTTZo0ga+vLwAgOTkZbm5u4HK5MDMzw/z581FZWSm6rqysDNOnT0fTpk3B4/HwxRdf4MqVK6LzSUlJ4HA4OHnyJJycnKCpqYkePXrg+fPn+P3332FrawtdXV2MGjUKJSUlMvwUCSFEybA8Kl5aQUFBePz4McrKynDp0iW4u7uLziUlJSE6Olr0Wk1NDaGhoXjw4AHevXuH7OxsbNq0Cfr6+lLdU24q9g8VFRXh119/hZWVFYyMjFBRUQFfX1/o6Ojg3LlzSElJAZ/PR69evVBeXg4AWLt2LaKjo7Fjxw6cP38e+fn5OHz4cI3YMTEx0NDQQEpKCqKiovD06VP06dMHrq6uSEtLw5YtW7B9+3YsXbpUdM3cuXNx6NAhxMTE4Pr167CysoKvry/y88VXtgoLC8PGjRtx4cIF0ReRDRs2YO/evTh+/DhOnTqFn376SbYfHiGEEIUmN13xx44dA5/PBwAUFxfDzMwMx44dg4qKCvbu3QuhUIht27aBw+EAAHbu3Al9fX0kJSXhyy+/xIYNGxASEoJBgwYBAKKionDy5Mka97G2tsaqVatErxcsWABzc3Ns3LgRHA4Hbdu2xbNnzzBv3jwsXrwY7969w5YtWxAdHY3evXsDALZu3YqEhARs374dc+bMEcVaunQpPD09AQATJkxASEgIMjMzYWlZvUThkCFDcObMGcybN08GnyAhhCghYVVjl+AfJzcVe/fu3bFlyxYAwOvXr7F582b07t0bly9fRlpaGh48eAAdHR2xa0pLS5GZmYnCwkLk5OSIdYGoqanBxcWlRnd8x44dxV6np6fDw8ND9IUBADw9PVFUVIS///4bBQUFqKioEFXYAKCurg43Nzekp6eLxXJwcBD9v4mJCbS0tESV+vtjn1s6sKysrMYKRxVMFdQ57K+jTQghCqEe3efyTm4qdm1tbVhZWYleb9u2DXp6eti6dSuKiorQsWNH7Nmzp8Z1xsbGUt9HVtTV1UX/z+FwxF6/Pyb8zECPyMhIhIeHix0bpt0ew/n27BaUEEKI3JLLZ+xAdSWooqKCd+/ewdnZGRkZGWjatCmsrKzE0vsl/8zMzHDp0iXR9ZWVlbh27Vqd97G1tcXFixfFWvYpKSnQ0dFBixYt0Lp1a9Ez+fcqKipw5coV2NnZsfqea1vxaJA2u/cghBCFIhSym+SA3FTsZWVlyM3NRW5uLtLT0zFt2jQUFRWhX79+GD16NJo0aYIBAwbg3LlzePToEZKSkjB9+nT8/fffAIAZM2ZgxYoViIuLw927dzFlyhQUFBTUed8pU6bgyZMnmDZtGu7evYsjR44gNDQUwcHBUFFRgba2NiZPnow5c+YgPj4ed+7cQWBgIEpKSjBhwgRWPwMulwtdXV2xRN3whBDyaY21QE1jkpuu+Pj4eJiZmQEAdHR00LZtW/z222/w8vICAJw9exbz5s3DoEGD8PbtWzRv3hw9e/aErq4uAGDWrFnIycnBuHHjoKKigvHjx2PgwIEoLCz87H2bN2+OEydOYM6cOejQoQMMDQ0xYcIELFy4UJRnxYoVEAqFGDNmDN6+fQsXFxecPHkSBgYGsvkwCCGEkE/gMNKsU0f+dWJNR7Ee0//tn6zGK6mouaVhQ6mrsv+dtAW/Cavx/i56yWo8AAg06cR6zAfCt6zGu/Emi9V4AKCuwn7P1LOi/LozkTppqde9brq03hQ/ZC1WWcYF1mIBANe6M6vxZEFuWuyEEEKI1OSk+5xNcvOMnRBCCCF1oxY7IYQQxUUL1BBCCCEKhLriCSGEECLPqMVOCCFEccnJojJsooqdEEKI4qKueEIIIYTIM1qgRs6paTRnPWZBUMe6M0nBcNN1VuMBgFfT9qzHPJN3i9V4yvoPi6umXncmKZVXVrAe87tmXVmNt+7ZWVbjyQtVFfbbh2WlT9iLdbPm9twNwXXwZTWeLFBXPCGEEIXFMMo33Y264gkhhBAFQi12QgghiksJB89RxU4IIURxKeF0N+qKJ4QQQhQItdgJIYQoLiXsiqcWO4Dc3FxMmzYNlpaW4HK5MDc3R79+/ZCYmPiP3N/LywszZ878R+5FCCFKRVjFbpIDSt9iz8rKgqenJ/T19bF69WrY29ujoqICJ0+exNSpU3H37t3GLiIhhBAiMaVvsU+ZMgUcDgeXL1/G4MGDYWNjg3bt2iE4OBh//vknACA7OxsDBgwAn8+Hrq4uhg0bhry8PFGMzMxMDBgwACYmJuDz+XB1dcXp06fF7rN582ZYW1uDx+PBxMQEQ4YMAQD4+/sjOTkZP/zwAzgcDjgcDrKysv6x908IIQqNEbKb5IBSV+z5+fmIj4/H1KlToa2tXeO8vr4+hEIhBgwYgPz8fCQnJyMhIQEPHz7E8OHDRfmKiorQp08fJCYm4saNG+jVqxf69euH7OxsAMDVq1cxffp0RERE4N69e4iPj0fXrtWrXv3www/w8PBAYGAgcnJykJOTA3Nz83/mAyCEEEUnFLKb5IBSd8U/ePAADMOgbdu2n8yTmJiIW7du4dGjR6IKd9euXWjXrh2uXLkCV1dXdOjQAR06dBBds2TJEhw+fBhHjx5FUFAQsrOzoa2tjb59+0JHRwctW7aEk5MTAEBPTw8aGhrQ0tKCqampbN8wIYQQhafULXZJlslPT0+Hubm5WCvazs4O+vr6SE9PB1DdYp89ezZsbW2hr68PPp+P9PR0UYvdx8cHLVu2hKWlJcaMGYM9e/agpKRE6vKWlZXhzZs3YomW+ieEkM+grnjlYm1tDQ6H0+ABcrNnz8bhw4exfPlynDt3DqmpqbC3t0d5eTkAQEdHB9evX8e+fftgZmaGxYsXo0OHDigoKJDqPpGRkdDT0xNLjPBtg8pOCCEKTQm74pW6Yjc0NISvry82bdqE4uLiGucLCgpga2uLJ0+e4MmT/+02dOfOHRQUFMDOzg4AkJKSAn9/fwwcOBD29vYwNTWtMQBOTU0N3t7eWLVqFW7evImsrCz88ccfAAANDQ1UVdU9jSIkJASFhYViiaOi04BPgBBCiKJR6mfsALBp0yZ4enrCzc0NERERcHBwQGVlJRISErBlyxbcuXMH9vb2GD16NDZs2IDKykpMmTIF3bp1g4uLC4Dqln9sbCz69esHDoeDRYsWQfjBN7tjx47h4cOH6Nq1KwwMDHDixAkIhUK0adMGACAQCHDp0iVkZWWBz+fD0NAQKrVshcjlcsHlcsWOcTgcGX46hBAi5+Sklc0mpW6xA4ClpSWuX7+O7t27Y9asWWjfvj18fHyQmJiILVu2gMPh4MiRIzAwMEDXrl3h7e0NS0tLHDhwQBRj3bp1MDAwQOfOndGvXz/4+vrC2dlZdF5fXx+xsbHo0aMHbG1tERUVhX379qFdu3YAqrvyVVVVYWdnB2NjY9GzeUIIIQ3DMFWsJnnAYWj0lVxT02jOesyCoI6sxjPcdJ3VeADg1bQ96zHP5N1iNZ6y/sPiqqmzHrO8soL1mN8168pqvHXPzrIaT16o1tK72FBlpU/qziShd2ejWYsFAJpd/VmNJwtK3xVPCCFEgSlhVzxV7IQQQhSXnExRY5PSP2MnhBBCFAlV7IQQQhRXI89j37RpEwQCAXg8Htzd3XH58uVP5vXy8hLtGfJh+uqrr6S6J1XshBBCFFcjrjx34MABBAcHIzQ0FNevX0eHDh3g6+uL58+f15o/NjZWtGdITk4Obt++DVVVVQwdOlSq+1LFTgghhMjAunXrEBgYiICAANjZ2SEqKgpaWlrYsWNHrfkNDQ1hamoqSgkJCdDS0pK6YqfBc4QQQhQXy6Piy8rKUFZWJnastsXDysvLce3aNYSEhIiOqaiowNvbGxcvXpToXtu3b8eIESNq3X30c6hil3OyWHfOYOM1VuOVPDvHajwA0GzWhfWYbM+/LpPB3Gt50ERTl/WYT9++Yj3mtldXWY2nz5Puj68k3lWWsx6T7d/Lqn/7dDKWR8VHRkYiPDxc7FhoaCjCwsLEjr18+RJVVVUwMTERO25iYiLR/iSXL1/G7du3sX37dqnLSBU7IYQQIqGQkBAEBweLHfu4tc6G7du3w97eHm5ublJfSxU7IYQQxcVyj0Jt3e61adKkCVRVVZGXlyd2PC8vD6ampp+9tri4GPv370dERES9ykiD5wghhCiuRprupqGhgY4dOyIxMfGDogiRmJgIDw+Pz17722+/oaysDF9//XW93jK12AkhhBAZCA4Oxrhx4+Di4gI3Nzds2LABxcXFCAgIAACMHTsWzZs3R2RkpNh127dvh5+fH4yMjOp1X6rYCSGEKK5GXFJ2+PDhePHiBRYvXozc3Fw4OjoiPj5eNKAuOzu7xhbd9+7dw/nz53Hq1Kl635d2d5Nz6jLY3Y1tNCpeuTTXqV8r43NkMSpel6vFajwVDvtzVORhVLwsVJY/ZS3Wu6NrWIsFAJr9Z7MaTxaU8hm7v79/rcv2PXjwoNHK4+fn1yj3JoQQoliUtiu+V69e2Llzp9gxY2NjqeOUl5dDQ0ODrWIRQghhE+3upjy4XK7Y0n2mpqZQVVVFcnIy3NzcwOVyYWZmhvnz56OyslJ0nZeXF4KCgjBz5kw0adIEvr6+AIDbt2+jd+/e4PP5MDExwZgxY/Dy5UvRdQcPHoS9vT00NTVhZGQEb29vFBcXIywsDDExMThy5Iio5yApKemf/jgIIUQxNfImMI1BaSv22jx9+hR9+vSBq6sr0tLSsGXLFmzfvh1Lly4VyxcTEwMNDQ2kpKQgKioKBQUF6NGjB5ycnHD16lXEx8cjLy8Pw4YNAwDk5ORg5MiRGD9+PNLT05GUlIRBgwaBYRjMnj0bw4YNQ69evUQL/3fu3Lkx3j4hhBAFoLRd8ceOHQOfzxe97t27N2xsbGBubo6NGzeCw+Ggbdu2ePbsGebNm4fFixeLRi9aW1tj1apVomuXLl0KJycnLF++XHRsx44dMDc3x/3791FUVITKykoMGjQILVu2BADY29uL8mpqaqKsrKzORQsIIYRISQm74pW2Yu/evTu2bNkieq2trY2pU6fCw8MDnA9Gt3p6eqKoqAh///03LCwsAAAdO3YUi5WWloYzZ86IfVF4LzMzE19++SV69uwJe3t7+Pr64ssvv8SQIUNgYGAgVZlr23yAYRix8hJCCPmAnHSfs0lpK3ZtbW1YWVnV+9oPFRUVoV+/fli5cmWNvGZmZlBVVUVCQgIuXLiAU6dO4aeffsKCBQtw6dIltGrVSuL71rb5AEeFD1VV9jfdIIQQIp/oGfsHbG1tcfHiRXw4tT8lJQU6Ojpo0aLFJ69zdnbGX3/9BYFAACsrK7H0/ksAh8OBp6cnwsPDcePGDWhoaODw4cMAqpcerKqqqrN8ISEhKCwsFEsqKjoNfNeEEKLAaPCccpsyZQqePHmCadOm4e7duzhy5AhCQ0MRHBxcY3WgD02dOhX5+fkYOXIkrly5gszMTJw8eRIBAQGoqqrCpUuXsHz5cly9ehXZ2dmIjY3FixcvYGtrCwAQCAS4efMm7t27h5cvX6KiovYFJLhcLnR1dcUSdcMTQshnMAy7SQ5Qxf6B5s2b48SJE7h8+TI6dOiAb7/9FhMmTMDChQs/e12zZs2QkpKCqqoqfPnll7C3t8fMmTOhr68PFRUV6Orq4uzZs+jTpw9sbGywcOFCrF27Fr179wYABAYGok2bNnBxcYGxsTFSUlL+ibdLCCFEAdGSsnKOlpRlDy0pyw5aUpY9tKRsw73bF8paLADQHBled6ZGprSD5wghhCgBOXkuzibqiieEEEIUCLXYCSGEKC5aoIYQQghRINQVTwghhBB5Ri12QgghiksJJ35RxU4IIURxUVc8IYQQQuQZtdjlnCw6mZyatGY1Hr9FN1bjAcDbHf6sxzQM/JXVeLJY7PdzSxvXVxXLLZq35e9YjQcAqjJ43611zFiNl/oyk9V4gGx+3kpHCVvsVLETQghRXEo43Y2+DhJCCCEKhFrshBBCFBYjpFHxhBBCiOJQwmfs1BVPCCGEKBCFq9ijo6Ohr68v1TX+/v7w8/OTSXkIIYQ0IkbIbpIDclWxf6oCTkpKAofDQUFBAYYPH4779+//84UjhBDy7yNk2E1yQOGesWtqakJTU7Oxi0EIIYQ0CrlqsUuitq74pUuXomnTptDR0cHEiRMxf/58ODo61rh2zZo1MDMzg5GREaZOnYqKigoAwMaNG9G+fXtRvri4OHA4HERFRYmOeXt7Y+HChQCAzMxMDBgwACYmJuDz+XB1dcXp06dFeSMiIsTivefo6IhFixY15O0TQgj5kFDIbpIDClexf2zPnj1YtmwZVq5ciWvXrsHCwgJbtmypke/MmTPIzMzEmTNnEBMTg+joaERHRwMAunXrhjt37uDFixcAgOTkZDRp0gRJSUkAgIqKCly8eBFeXl4AgKKiIvTp0weJiYm4ceMGevXqhX79+iE7OxsAMH78eKSnp+PKlSui+9+4cQM3b95EQECA7D4MQghRNlSx//sdO3YMfD5fLPXu3fuT+X/66SdMmDABAQEBsLGxweLFi2Fvb18jn4GBATZu3Ii2bduib9+++Oqrr5CYmAgAaN++PQwNDZGcnAyg+pn+rFmzRK8vX76MiooKdO7cGQDQoUMHfPPNN2jfvj2sra2xZMkStG7dGkePHgUAtGjRAr6+vti5c6fo/jt37kS3bt1gaWnJzgdFCCFEKcldxd69e3ekpqaKpW3btn0y/7179+Dm5iZ27OPXANCuXTuoqqqKXpuZmeH58+cAAA6Hg65duyIpKQkFBQW4c+cOpkyZgrKyMty9exfJyclwdXWFlpYWgOoW++zZs2Frawt9fX3w+Xykp6eLWuwAEBgYiH379qG0tBTl5eXYu3cvxo8f/9n3XlZWhjdv3oglRgm3JCSEEIkxDLtJDsjd4DltbW1YWVmJHfv7778bHFddXV3sNYfDgfCDbhcvLy/88ssvOHfuHJycnKCrqyuq7JOTk9Gt2/82Opk9ezYSEhKwZs0aWFlZQVNTE0OGDEF5ebkoT79+/cDlcnH48GFoaGigoqICQ4YM+WwZIyMjER4eLl5OFT44qroNeeuEEKK45KT7nE1y12KXVps2bcSeZQOo8VoS75+z//bbb6Jn6V5eXjh9+jRSUlJExwAgJSUF/v7+GDhwIOzt7WFqaoqsrCyxeGpqahg3bhx27tyJnTt3YsSIEXWO5g8JCUFhYaFY4qjoSP1eCCGEKC65a7FLa9q0aQgMDISLiws6d+6MAwcO4ObNm1I/y3ZwcICBgQH27t2LY8eOAaiu2GfPng0OhwNPT09RXmtra8TGxqJfv37gcDhYtGiRWOv/vYkTJ8LW1hZA9ZeBunC5XHC5XLFjHI4sNgclhBAFISdzz9mk8C320aNHIyQkBLNnz4azszMePXoEf39/8Hg8qeJwOBx06dIFHA4HX3zxBYDqyl5XVxcuLi7Q1tYW5V23bh0MDAzQuXNn9OvXD76+vnB2dq4R09raGp07d0bbtm3h7u7esDdKCCGkJiVceY7DKOHoKx8fH5iammL37t2NWg6GYWBtbY0pU6YgODi4XjHUNJqzXCrAqUlrVuPdzH/EajwAKNg2lvWYhoG/shqvsqqS1XgAoKLC/nfxKpafQepytViNBwDFFaWsx3QwbMVqvNSXmazGA+Tj5y0LleVPWYtVsvrzg5KlpTVnh1T5N23ahNWrVyM3NxcdOnTATz/9VOsA7vcKCgqwYMECxMbGIj8/Hy1btsSGDRvQp08fie+p8F3xJSUliIqKgq+vL1RVVbFv3z6cPn0aCQkJjVquFy9eYP/+/cjNzaW564QQIiuN2BV/4MABBAcHIyoqCu7u7tiwYQN8fX1x7949NG3atEb+8vJy+Pj4oGnTpjh48CCaN2+Ox48fS73/icJX7BwOBydOnMCyZctQWlqKNm3a4NChQ/D29m7UcjVt2hRNmjTBL7/8AgMDg0YtCyGEKCqmEXso1q1bh8DAQFHjLSoqCsePH8eOHTswf/78Gvl37NiB/Px8XLhwQTRTSyAQSH1fha/YNTU1xZZz/bdQwicghBCiNMrLy3Ht2jWEhISIjqmoqMDb2xsXL16s9ZqjR4/Cw8MDU6dOxZEjR2BsbIxRo0Zh3rx5Yuus1EXhK3ZCCCFKjOWu+LKyMpSVlYkdq23G0suXL1FVVQUTExOx4yYmJrh7926tsR8+fIg//vgDo0ePxokTJ/DgwQNMmTIFFRUVCA0NlbiMCj8qnhBCiBJjeVR8ZGQk9PT0xFJkZCQrRRUKhWjatCl++eUXdOzYEcOHD8eCBQvENhyTBLXYCSGEEAmFhITUmMX0cWsdAJo0aQJVVVXk5eWJHc/Ly4OpqWmtsc3MzKCuri7W7W5ra4vc3FyUl5dDQ0NDojJSi50QQojiEjKsJi6XC11dXbFUW8WuoaGBjh07ijYTA6pb5ImJifDw8Ki1qJ6ennjw4IHYgmb379+HmZmZxJU6QBU7IYQQRdaI27YGBwdj69atiImJQXp6OiZPnozi4mLRKPmxY8eKDa6bPHky8vPzMWPGDNy/fx/Hjx/H8uXLMXXqVKnuS13xck5H4/Pry9fH/TfsLQ4BAGoqko/mlJTBRPYXF/rLypbVeL3zXrEaDwDyy96wHrO0soLVeGVV7MYDZPM79Kr8Lavx+DL4tygE+7Nn2P6boaeuXXcmJTV8+HC8ePECixcvRm5uLhwdHREfHy8aUJednS22CJG5uTlOnjyJ7777Dg4ODmjevDlmzJiBefPmSXVfpVx5TpEY8K3qziQltv+YVAqrWI0HyGb1LKrY2cHIoDKSBRMtdtePeF3K7hcFQHkr9r/yLrEWq3jxCNZiAYB2xH5W48kCtdgJIYQoLjlZ351N9IydEEIIUSDUYieEEKK4lHDbVqrYCSGEKKzGXCu+sSh0VzyHw0FcXJxc3DssLAyOjo4yKw8hhBDl8K+t2P39/eHn59fYxRBJSkoCh8NBQUGBRPlzcnLQu3dv2RaKEELI57G8QI08oK54lr1f9u9TSwYSQgj5B8lJZcymf22L/UNeXl6YPn065s6dC0NDQ5iamiIsLEwsT0ZGBrp27Qoejwc7OzskJCSIna+txZ2amgoOh4OsrCwAwOPHj9GvXz8YGBhAW1sb7dq1w4kTJ5CVlYXu3bsDAAwMDMDhcODv7y8qW1BQEGbOnIkmTZrA19cXQM2u+Hnz5sHGxgZaWlqwtLTEokWLUFHB/kIehBBClJvctNhjYmIQHByMS5cu4eLFi/D394enpyd8fHwgFAoxaNAgmJiY4NKlSygsLMTMmTOlvsfUqVNRXl6Os2fPQltbG3fu3AGfz4e5uTkOHTqEwYMH4969e9DV1YWm5v8WeYiJicHkyZORkpLyydg6OjqIjo5Gs2bNcOvWLQQGBkJHRwdz586tz8dBCCFEEko4j11uKnYHBwfRfrTW1tbYuHEjEhMT4ePjg9OnT+Pu3bs4efIkmjVrBgBYvny51M+4s7OzMXjwYNjb2wMALC0tRecMDQ0BAE2bNoW+vr7YddbW1li1atVnYy9cuFD0/wKBALNnz8b+/fupYieEEFlSwq54uarYP2RmZobnz58DANLT02Fubi6q1AF8cvecz5k+fTomT56MU6dOwdvbG4MHD65x39p07NixzjwHDhzAjz/+iMzMTBQVFaGyshK6urpSla+srAxlZWVixxiGAYfDkSoOIYQQxSUXz9gBQF1dXew1h8MR29quLu8X2v9wafyPn3FPnDgRDx8+xJgxY3Dr1i24uLjgp59+qjO2tvbn10q+ePEiRo8ejT59+uDYsWO4ceMGFixYgPLyconLDwCRkZHQ09MTS6UVr6WKQQghyoQRMqwmeSA3Ffvn2Nra4smTJ8jJyREd+/PPP8XyGBsbA4BYntTU1BqxzM3N8e233yI2NhazZs3C1q1bAUC0F25VlfQbmly4cAEtW7bEggUL4OLiAmtrazx+/FjqOCEhISgsLBRLPHV2N7IghBCFooTT3RSiYvf29oaNjQ3GjRuHtLQ0nDt3DgsWLBDLY2VlBXNzc4SFhSEjIwPHjx/H2rVrxfLMnDkTJ0+exKNHj3D9+nWcOXMGtrbVO361bNkSHA4Hx44dw4sXL1BUVCRx+aytrZGdnY39+/cjMzMTP/74Iw4fPiz1++RyudDV1RVL1A1PCCHkQwpRsauoqODw4cN49+4d3NzcMHHiRCxbtkwsj7q6Ovbt24e7d+/CwcEBK1euxNKlS8XyVFVVYerUqbC1tUWvXr1gY2ODzZs3AwCaN2+O8PBwzJ8/HyYmJggKCpK4fP3798d3332HoKAgODo64sKFC1i0aFHD3zghhJDPEwrZTXKA9mOXc7QfO3toP3Z20H7s7KH92Bvu7RR2VwDV2fw7q/FkQSFa7IQQQgipJjfT3QghhBCpycmANzZRxU4IIURhKePTZuqKJ4QQQhQItdgJIYQoLuqKJ4QQQhSIElbs1BVPCCGEKBBqscu5ksqyujNJSZo1+BXJ16+lW7u/Ln92Zn9+b6cLrIfE08qXrMYrY3lePADo89j/LF+VsrsmgJYal9V4AFBaye7vpCw8K2Z/vQY2ycv67myiip0QQojiUsKKnbriCSGEEAVCLXZCCCGKSwmfLFLFTgghRGEp4zN26oonhBBCFIjCVuwCgQAbNmz4bB4Oh4O4uLhGuTchhJB/gJBhN8mBf1XF7u/vDz8/P1ZiXblyBZMmTWIl1ofevHmDBQsWoG3btuDxeDA1NYW3tzdiY2NFaxJ/fG9ZfYEghBBSByHLSQ4o7DN2Y2Nj1mMWFBTgiy++QGFhIZYuXQpXV1eoqakhOTkZc+fORY8ePaCvry+TexNCCCGS+Fe12D/k5eWF6dOnY+7cuTA0NISpqSnCwsJE5xmGQVhYGCwsLMDlctGsWTNMnz5ddP7j7vCMjAx07doVPB4PdnZ2SEhIqHHPJ0+eYNiwYdDX14ehoSEGDBiArKws0fnvv/8eWVlZuHTpEsaNGwc7OzvY2NggMDAQqamp4PP5Ne4tEAgAAAMHDgSHw4FAIEBWVhZUVFRw9epVsftv2LABLVu2VNoFYgghhG2MkGE1yYN/bcUOADExMdDW1salS5ewatUqREREiCrkQ4cOYf369fj555+RkZGBuLg42Nvb1xpHKBRi0KBB0NDQwKVLlxAVFYV58+aJ5amoqICvry90dHRw7tw5pKSkgM/no1evXigvL4dQKMT+/fsxevRoNGvWrMY9+Hw+1NRqdoBcuXIFALBz507k5OTgypUrEAgE8Pb2xs6dO8Xy7ty5E/7+/lBR+Vf/WAghRH5QV/y/i4ODA0JDQwEA1tbW2LhxIxITE+Hj44Ps7GzR8211dXVYWFjAzc2t1jinT5/G3bt3cfLkSVGlvHz5cvTu3VuU58CBAxAKhdi2bRs4HA6A6opWX18fSUlJcHR0xOvXr9G2bVup3sP7bnl9fX2YmpqKjk+cOBHffvst1q1bBy6Xi+vXr+PWrVs4cuSIVPEJIYSQD/2rm4YODg5ir83MzPD8+XMAwNChQ/Hu3TtYWloiMDAQhw8fRmVlZa1x0tPTYW5uLtbS9vDwEMuTlpaGBw8eQEdHB3w+H3w+H4aGhigtLUVmZqZoYBxb/Pz8oKqqisOHDwMAoqOj0b17d1HXfW3Kysrw5s0bscR2uQghRJFQV/y/jLq6uthrDocjev5sbm6Oe/fuYfPmzdDU1MSUKVPQtWtXVFTUbwOKoqIidOzYEampqWLp/v37GDVqFIyNjaGvr4+7d+82+H0BgIaGBsaOHYudO3eivLwce/fuxfjx4z97TWRkJPT09MRSVRW7G1kQQohCUcKu+H91xV4XTU1N9OvXDz/++COSkpJw8eJF3Lp1q0Y+W1tbPHnyBDk5OaJjf/75p1geZ2dnZGRkoGnTprCyshJLenp6UFFRwYgRI7Bnzx48e/asxj2Kioo+2WOgrq6OqqqqGscnTpyI06dPY/PmzaisrMSgQYM++35DQkJQWFgollRVdT97DSGEkMazadMmCAQC8Hg8uLu74/Lly5/MGx0dDQ6HI5Z4PJ7U95Tbij06Ohrbt2/H7du38fDhQ/z666/Q1NREy5Yta+T19vaGjY0Nxo0bh7S0NJw7dw4LFiwQyzN69Gg0adIEAwYMwLlz5/Do0SMkJSVh+vTp+PvvvwEAy5Ytg7m5Odzd3bFr1y7cuXMHGRkZ2LFjB5ycnFBUVFRrWQUCARITE5Gbm4vXr1+Ljtva2qJTp06YN28eRo4cCU1Nzc++Zy6XC11dXbH0fjwAIYSQmhghu0kaBw4cQHBwMEJDQ3H9+nV06NABvr6+okfKtdHV1UVOTo4oPX78WOr3LLcVu76+PrZu3QpPT084ODjg9OnT+O9//wsjI6MaeVVUVHD48GG8e/cObm5umDhxIpYtWyaWR0tLC2fPnoWFhQUGDRoEW1tbTJgwAaWlpdDVrW4VGxoa4s8//8TXX3+NpUuXwsnJCV26dMG+ffuwevVq6Onp1VrWtWvXIiEhAebm5nBychI7N2HCBJSXl9fZDU8IIaQeGrErft26dQgMDERAQADs7OwQFRUFLS0t7Nix45PXcDgcmJqaipKJiYl0NwXAYWj0VaNasmQJfvvtN9y8ebNe13N55iyXCEo7j97F2IbVeMdd2P8cO10oZj3m0+KXrMYrq6zfOJfP0edpsx6zQljz8VhDaKlxWY0HAKWV5azH1NaQvmv3c4rLS1mNBwCvix6wFuvVV91YiwUA/NhTKCsrEzvG5XLB5Yr//MvLy6GlpYWDBw+Krag6btw4FBQU1DoDKjo6GhMnTkTz5s0hFArh7OyM5cuXo127dlKVUW5b7PKuqKgIt2/fxsaNGzFt2rTGLg4hhCgktrviaxvEHBkZWeO+L1++RFVVVY0Wt4mJCXJzc2sta5s2bbBjxw4cOXIEv/76K4RCITp37ix6HCypf/U8dkUWFBSEffv2wc/Pj7rhCSFEVljuOAsJCUFwcLDYsY9b6/Xl4eEhNhW7c+fOsLW1xc8//4wlS5ZIHIcq9kYSHR2N6Ojoxi4GIYQQKdTW7V6bJk2aQFVVFXl5eWLH8/LyxBYr+xx1dXU4OTnhwQPpHk1QVzwhhBCF1Vij4jU0NNCxY0ckJiaKjgmFQiQmJtZYIO1TqqqqcOvWLZiZmUn1nqnFTgghRGFJO0WNTcHBwRg3bhxcXFzg5uaGDRs2oLi4GAEBAQCAsWPHonnz5qJn9BEREejUqROsrKxQUFCA1atX4/Hjx5g4caJU96WKnRBCCJGB4cOH48WLF1i8eDFyc3Ph6OiI+Ph40YC67OxssU2/Xr9+jcDAQOTm5sLAwAAdO3bEhQsXYGdnJ9V9abqbnKPpbuyh6W7soOlu7KHpbg2X153d6W4mZ5JZjScL1GInhBCiuBjlW52TKnY552bEbisTANIKHrEar6SirO5MUlKRwVK6z8sKWY1nm1TCajwAuGrTnPWYXo/Y7Vl48vbTy2XWl7Y6u61MAHhewu7PWxadn7JosZdWsdujoqmmwWo80nBUsRNCCFFYjTl4rrFQxU4IIURhMULl64qneeyEEEKIAqEWOyGEEIVFXfGEEEKIAmGUcFS80nXFe3l5YebMmY0egxBCCJEFhavY/f39xfa+BYCDBw+Cx+Nh7dq1iI2NFdslRyAQYMOGDf9sIQkhhPwjGmut+Mak8F3x27Ztw9SpUxEVFSVan5cQQohyoFHxCmbVqlWYNm0a9u/fL6rUP+xG9/LywuPHj/Hdd9+Bw+GA88GiJykpKfDy8oKWlhYMDAzg6+uL169fi84LhULMnTsXhoaGMDU1RVhYmNi9CwoKMHHiRBgbG0NXVxc9evRAWlqa6HxYWBgcHR2xe/duCAQC6OnpYcSIEXj79q3sPhBCCCEKT2Er9nnz5mHJkiU4duwYBg4cWGue2NhYtGjRAhEREcjJyUFOTg4AIDU1FT179oSdnR0uXryI8+fPo1+/fqiq+t/a0jExMdDW1salS5ewatUqREREICEhQXR+6NCheP78OX7//Xdcu3YNzs7O6NmzJ/Lz80V5MjMzERcXh2PHjuHYsWNITk7GihUrZPSJEEKI8mEYdpM8UMiu+N9//x1HjhxBYmIievTo8cl8hoaGUFVVhY6OjtjG96tWrYKLiws2b94sOtauXTuxax0cHBAaGgoAsLa2xsaNG5GYmAgfHx+cP38ely9fxvPnz8HlVm8MsWbNGsTFxeHgwYOYNGkSgOpWf3R0NHR0dAAAY8aMQWJiIpYtW8bOB0EIIUpOGbviFbJid3BwwMuXLxEaGgo3Nzfw+Xyprk9NTcXQoUPrvMeHzMzM8Px59RrZaWlpKCoqgpGRkVied+/eITMzU/RaIBCIKvWPY9SmrKwMZWXi664LGSFUOArb8UIIIURKClmxN2/eHAcPHkT37t3Rq1cv/P7772IVaF00NTXrzKOuri72msPhiLY7LSoqgpmZGZKSkmpcp6+vL1GM2kRGRiI8PFzsmDlfgJa6lnWWlxBClJEyttgVtqnXsmVLJCcnIzc3F7169frkoDQNDQ2xZ+dAdWs8MTGx3vd2dnZGbm4u1NTUYGVlJZaaNGlS77ghISEoLCwUS+Y6gnrHI4QQRaeMz9gVtmIHAHNzcyQlJeH58+fw9fXFmzdvauQRCAQ4e/Ysnj59ipcvXwKorkCvXLmCKVOm4ObNm7h79y62bNkiOl8Xb29veHh4wM/PD6dOnUJWVhYuXLiABQsW4OrVq/V+P1wuF7q6umKJuuEJIYR8SOFrhRYtWiApKQkvX76stXKPiIhAVlYWWrduDWNjYwCAjY0NTp06hbS0NLi5ucHDwwNHjhyBmppkTy44HA5OnDiBrl27IiAgADY2NhgxYgQeP34MExMT1t8jIYSQ2jFCDqtJHnAYRl46F0htujTvyXrMtIJHrMYrqSirO5OUVDjs/wMz12nKaryiihJW4wHAVZvmrMf0evS67kxSePL20wNA68uMb8h6zOclhazGU1dRZTUeAJRWlrMeU5XlcmqqabAaDwDy32awFiuzvS9rsQCg9e2TrMaTBYVvsRNCCCHKRCFHxRNCCCGA/Kzvziaq2AkhhCgsIW3bSgghhBB5Ri12QgghCotRwhY7VeyEEEIUlrxMUWMTdcUTQgghCoRa7IQQQhSWMq7UQhW7nGuppsd6zKvCqrozNTKhDP61qrK8PK8Q7JcxKK/uDYqkNV7btO5MUgh7k8dqPFmprKpkNZ4sFk1iezEZALDUZffnnVOSz2o8tlFXPCGEEELkGrXYCSGEKCyax04IIYQQuUYtdkIIIQqL5rETQgghCkQZR8U3Wlc8h8NBXFycxPkFAgE2bNggs/IQQgghikCmFbu/vz/8/PxqPZeTk4PevXtLHOvKlSuYNGkSSyWrlpSUBA6HI0rGxsbo06cPbt26xep9CCGENA4hw2E1yYNGa7GbmpqCy+VKnN/Y2BhaWloyKcu9e/eQk5ODkydPoqysDF999RXKy8tlci9CCCH/HIbhsJrkwb+iK75z586YN2+e2PkXL15AXV0dZ8+eBVCzK57D4WDbtm0YOHAgtLS0YG1tjaNHj4rFOHr0KKytrcHj8dC9e3fExMSAw+GgoKBALF/Tpk1hamoKZ2dnzJw5E0+ePMHdu3dF5w8dOoR27dqBy+VCIBBg7dq1Yte/fv0aY8eOhYGBAbS0tNC7d29kZGSIzkdHR0NfXx/Hjh1DmzZtoKWlhSFDhqCkpAQxMTEQCAQwMDDA9OnTUVX1718chhBCyL/Xv2K62+jRo7F//34wH4xyOHDgAJo1a4YuXbp88rrw8HAMGzYMN2/eRJ8+fTB69Gjk51evgvTo0SMMGTIEfn5+SEtLwzfffIMFCxZ8thyFhYXYv38/AEBDQwMAcO3aNQwbNgwjRozArVu3EBYWhkWLFiE6Olp0nb+/P65evYqjR4/i4sWLYBgGffr0QUVFhShPSUkJfvzxR+zfvx/x8fFISkrCwIEDceLECZw4cQK7d+/Gzz//jIMHD0r9+RFCCKkdw7CbpLVp0yYIBALweDy4u7vj8uXLEl23f/9+cDicTz7O/px/RcU+bNgwPHv2DOfPnxcd27t3L0aOHAnOZ5Zp9Pf3x8iRI2FlZYXly5ejqKhI9KH9/PPPaNOmDVavXo02bdpgxIgR8Pf3rzVOixYtwOfzoa+vj71796J///5o27YtAGDdunXo2bMnFi1aBBsbG/j7+yMoKAirV68GAGRkZODo0aPYtm0bunTpgg4dOmDPnj14+vSp2ODAiooKbNmyBU5OTujatSuGDBmC8+fPY/v27bCzs0Pfvn3RvXt3nDlzpoGfJiGEkPca8xn7gQMHEBwcjNDQUFy/fh0dOnSAr68vnj9//tnrsrKyMHv27M82bD/nX1GxGxsb48svv8SePXsAVLe2L168iNGjR3/2OgcHB9H/a2trQ1dXV/SB3bt3D66urmL53dzcao1z7tw5XLt2DdHR0bCxsUFUVJToXHp6Ojw9PcXye3p6IiMjA1VVVUhPT4eamhrc3d1F542MjNCmTRukp6eLjmlpaaF169ai1yYmJhAIBODz+WLHPvcDLysrw5s3b8RSFUNd94QQ8m+0bt06BAYGIiAgAHZ2doiKioKWlhZ27NjxyWuqqqowevRohIeHw9LSsl73/VdU7EB1d/zBgwdRUVGBvXv3wt7eHvb29p+9Rl1dXew1h8OBUCiU+t6tWrVCmzZtMG7cOEycOBHDhw+XOkZdaiurtOWPjIyEnp6eWPqr8D7rZSWEEEXB9uC52hpYZWVlNe5bXl6Oa9euwdvbW3RMRUUF3t7euHjx4ifLGxERgaZNm2LChAn1fs//mop9wIABKC0tRXx8PPbu3Vtna70ubdq0wdWrV8WOXblypc7rpk6ditu3b+Pw4cMAAFtbW6SkpIjlSUlJgY2NDVRVVWFra4vKykpcunRJdP7Vq1e4d+8e7OzsGvQePhYSEoLCwkKx1E7PhtV7EEKIImG7K762BlZkZGSN+758+RJVVVUwMTERO25iYoLc3Nxay/r+8ezWrVsb9J5lvvJcYWEhUlNTxY4ZGRnVyKetrQ0/Pz8sWrQI6enpGDlyZIPu+80332DdunWYN28eJkyYgNTUVNGAt889t9fS0kJgYCBCQ0Ph5+eHWbNmwdXVFUuWLMHw4cNx8eJFbNy4EZs3bwYAWFtbY8CAAQgMDMTPP/8MHR0dzJ8/H82bN8eAAQMa9B4+xuVya0wRVOWwv60jIYSQ2oWEhCA4OFjsmDRTtz/l7du3GDNmDLZu3YomTZo0KJbMW+xJSUlwcnISS+Hh4bXmHT16NNLS0tClSxdYWFg06L6tWrXCwYMHERsbCwcHB2zZskU0Kr6uH0JQUBDS09Px22+/wdnZGf/5z3+wf/9+tG/fHosXL0ZERITYQLydO3eiY8eO6Nu3Lzw8PMAwDE6cOFGjq50QQsg/i2E5cblc6OrqiqXa6pQmTZpAVVUVeXl5Ysfz8vJgampaI39mZiaysrLQr18/qKmpQU1NDbt27cLRo0ehpqaGzMxMid8zh2GUZyXdZcuWISoqCk+ePGnsorDm65aDWI956Pk1VuNVVFWyGk9WLPXMWI33uvwtq/EAoLOeNesx3Th6rMYLy0liNR4ANNOp2cvXUM/evmI1noYa+1/khTL482ypW7NSaYicknxW4wFA/tuMujNJ6ILZYNZiAUDnnEMS53V3d4ebmxt++uknAIBQKISFhQWCgoIwf/58sbylpaV48OCB2LGFCxfi7du3+OGHH2BjYyOahl0Xhd4EZvPmzXB1dYWRkRFSUlKwevVqBAUFNXaxCCGEKIHg4GCMGzcOLi4ucHNzw4YNG1BcXIyAgAAAwNixY9G8eXNERkaCx+Ohffv2Ytfr6+sDQI3jdVHoij0jIwNLly5Ffn4+LCwsMGvWLISEhDR2sQghhPxDGnMZ2OHDh+PFixdYvHgxcnNz4ejoiPj4eNGAuuzsbKiosP9EXKm64hURdcWzh7ri2UFd8eyhrviGO2c6hLVYANAl99+/Oui/ZrobIYQQQhpOobviCSGEKDcG8rEjG5uoYieEEKKwhEr4sJm64gkhhBAFQoPn5ByP17CFfGpjwOPXnUkKFUL5GDxXIWR3Q52S8lJW4wHVC2T8273dOob1mLqBu1mP2c1EuilEdUnOu81qPHmhqsL+6pelpdmsxfrDZBhrsQCgR95/WI0nC9QVTwghRGEp4zN26oonhBBCFAi12AkhhCgs6Tfyln9UsRNCCFFY1BVPCCGEELlGFfs/xN/fH35+fo1dDEIIUSpClpM8+FdX7P7+/uBwOPj2229rnJs6dSo4HI7Yvuj/BC8vL8ycOVPq63744QdER0ezXh5CCCGfRhX7v5C5uTn279+Pd+/eiY6VlpZi7969sLBgfw63rOjp6Ym24COEEEJk5V9fsTs7O8Pc3ByxsbGiY7GxsbCwsICTk5PoWFlZGaZPn46mTZuCx+Phiy++wJUrV0Tno6Oja1SscXFx4HD+N7AiLCwMjo6O2L17NwQCAfT09DBixAi8fVu9S5e/vz+Sk5Pxww8/gMPhgMPhICsrC1VVVZgwYQJatWoFTU1NtGnTBj/88IPYvT7uivfy8sL06dMxd+5cGBoawtTUFGFhYSx8YoQQQt5jwGE1yYN/fcUOAOPHj8fOnTtFr3fs2CHaqP69uXPn4tChQ4iJicH169dhZWUFX19f5OdLt6VgZmYm4uLicOzYMRw7dgzJyclYsWIFgOrudA8PDwQGBiInJwc5OTkwNzeHUChEixYt8Ntvv+HOnTtYvHgxvv/+e/znP59foSgmJgba2tq4dOkSVq1ahYiICCQkJEhVXkIIIZ8m5LCb5IFcVOxff/01zp8/j8ePH+Px48dISUnB119/LTpfXFyMLVu2YPXq1ejduzfs7OywdetWaGpqYvv27VLdSygUIjo6Gu3bt0eXLl0wZswYJCYmAqjuTtfQ0ICWlhZMTU1hamoKVVVVqKurIzw8HC4uLmjVqhVGjx6NgICAOit2BwcHhIaGwtraGmPHjoWLi4voXoQQQkh9yMU8dmNjY3z11VeIjo4GwzD46quv0KRJE9H5zMxMVFRUwNPTU3RMXV0dbm5uSE9Pl+peAoEAOjo6otdmZmZ4/vx5nddt2rQJO3bsQHZ2Nt69e4fy8nI4Ojp+9hoHBwex13Xdq6ysDGVlZWLHGIYRe5xACCHkf4Ry0n3OJrlosQPV3fHR0dGIiYnB+PHjpb5eRUUFH+93U1FRUSOfurq62GsOhwOh8PNjIffv34/Zs2djwoQJOHXqFFJTUxEQEIDy8vLPXiftvSIjI6GnpyeWqqrefPYehBCizBiWkzyQm4q9V69eKC8vR0VFBXx9fcXOtW7dGhoaGkhJSREdq6iowJUrV2BnZwegutX/9u1bFBcXi/KkpqZKXQ4NDQ1UVYnvApaSkoLOnTtjypQpcHJygpWVFTIzM6WOXZeQkBAUFhaKJVVVXdbvQwghRH7JRVc8AKiqqoq61VVVxbcJ1NbWxuTJkzFnzhwYGhrCwsICq1atQklJCSZMmAAAcHd3h5aWFr7//ntMnz4dly5dqte8coFAgEuXLiErKwt8Ph+GhoawtrbGrl27cPLkSbRq1Qq7d+/GlStX0KpVqwa/7w9xuVxwuVyxY9QNTwghnyYvc8/ZJDctdgDQ1dWFrm7tLdQVK1Zg8ODBGDNmDJydnfHgwQOcPHkSBgYGAABDQ0P8+uuvOHHiBOzt7bFv3756TS+bPXs2VFVVYWdnB2NjY2RnZ+Obb77BoEGDMHz4cLi7u+PVq1eYMmVKQ94qIYQQFgg5HFaTPOAwHz94JnKFx2N/kR4DHp/VeBXCSlbjyUqFsKruTFIoKS9lNR4gH8/43m4dw3pM3cDdrMfsZtKe1XjJebdZjScvVFVU684kpdLSbNZiHTQbzVosABiSs4fVeLIgN13xhBBCiLTk4csw26hiJ4QQorDoGTshhBBC5Bq12AkhhCgseVkGlk1UsRNCCFFYtPIcIYQQQuQatdgJIYQoLBoVTwghhCgQesZO5I6NfnPWY97JZ29xCHniYMTuEsA57/JZjQcA6jJYDISrqsFqPB0ZLCaT292K9ZimZ9hdUEZFTlYl0+NpsxpPjcP+7yRpGKrYCSGEKCxlnMdOFTshhBCFpYzP2GlUPCGEEKJAqMVOCCFEYdHgOUIIIUSBKOMzduqKl0BJSQkGDx4MXV1dcDgcFBQU1HpMIBBgw4YNjV1cQgghSkzpK/YnT55g/PjxaNasGTQ0NNCyZUvMmDEDr169EuWJiYnBuXPncOHCBeTk5EBPT6/WY1euXMGkSZMa8d0QQgj5kJDlJK1NmzZBIBCAx+PB3d0dly9f/mTe2NhYuLi4QF9fH9ra2nB0dMTu3dJPH1Xqiv3hw4dwcXFBRkYG9u3bhwcPHiAqKgqJiYnw8PBAfn71POTMzEzY2tqiffv2MDU1BYfDqfWYsbExtLS0GvldEUIIeY/hsJukceDAAQQHByM0NBTXr19Hhw4d4Ovri+fPn9ea39DQEAsWLMDFixdx8+ZNBAQEICAgACdPnpTqvkpdsU+dOhUaGho4deoUunXrBgsLC/Tu3RunT5/G06dPsWDBAnh5eWHt2rU4e/YsOBwOvLy8aj0GoEZXfEFBAb755huYmJiAx+Ohffv2OHbsmOj8+fPn0aVLF2hqasLc3BzTp09HcXHxP/wpEEIIkYV169YhMDAQAQEBsLOzQ1RUFLS0tLBjx45a83t5eWHgwIGwtbVF69atMWPGDDg4OOD8+fNS3VdpB8/l5+fj5MmTWLZsGTQ1NcXOmZqaYvTo0Thw4AAyMjIQEhKC27dvIzY2Fhoa1at0zZ8/v8axDwmFQvTu3Rtv377Fr7/+itatW+POnTtQVa1epSkzMxO9evXC0qVLsWPHDrx48QJBQUEICgrCzp07Zf8BEEKIEmB78FxZWRnKysrEjnG5XHC5XLFj5eXluHbtGkJCQkTHVFRU4O3tjYsXL9Z5H4Zh8Mcff+DevXtYuXKlVGVU2oo9IyMDDMPA1ta21vO2trZ4/fo1qqqqoKWlBQ0NDZiamorO13bsQ6dPn8bly5eRnp4OGxsbAIClpaXofGRkJEaPHo2ZM2cCAKytrfHjjz+iW7du2LJlC3g8HkvvlBBClBfbFXtkZCTCw8PFjoWGhiIsLEzs2MuXL1FVVQUTExOx4yYmJrh79+4n4xcWFqJ58+YoKyuDqqoqNm/eDB8fH6nKqLQV+3sMI5t1iVJTU9GiRQtRpf6xtLQ03Lx5E3v27BEri1AoxKNHj2r9wlHbN0UhI4QKR6mfqBBCyD8mJCQEwcHBYsc+bq03hI6ODlJTU1FUVITExEQEBwfD0tJS9MhXEkpbsVtZWYHD4SA9PR0DBw6scT49PR0GBgYwNjauV/yPu/c/VlRUhG+++QbTp0+vcc7CwqLWa2r7pthUuzlM+Ob1KiMhhCg6tptutXW716ZJkyZQVVVFXl6e2PG8vLxP9vQC1d31VlbVmx45OjoiPT0dkZGRUlXsStvUMzIygo+PDzZv3ox3796JncvNzcWePXswfPhwcOq5Y5ODgwP+/vtv3L9/v9bzzs7OuHPnDqysrGqk2p7ZA9XfFAsLC8WSsTb7u7sRQoiiEHLYTZLS0NBAx44dkZiY+L+yCIWiWVcSl18orNFTWxelrdgBYOPGjSgrK4Ovry/Onj2LJ0+eID4+Hj4+PmjevDmWLVtW79jdunVD165dMXjwYCQkJODRo0f4/fffER8fDwCYN28eLly4gKCgIKSmpiIjIwNHjhxBUFDQJ2NyuVzo6uqKJeqGJ4SQf6fg4GBs3boVMTExSE9Px+TJk1FcXIyAgAAAwNixY8UG10VGRiIhIQEPHz5Eeno61q5di927d+Prr7+W6r5K2xUPVA9Yu3r1KkJDQzFs2DDk5+fD1NQUfn5+CA0NhaGhYYPiHzp0CLNnz8bIkSNRXFwMKysrrFixAkB1iz45ORkLFixAly5dwDAMWrdujeHDh7Px1gghhKBxl5QdPnw4Xrx4gcWLFyM3NxeOjo6Ij48XDajLzs6Gisr/GmfFxcWYMmUK/v77b2hqaqJt27b49ddfpa4XOIysRo+Rf4SDqeRdOpK6k5/Nekx54GDUitV4Oe/yWY0HAOoqqqzH5KrW/uinvh4V5rIaDwByu1uxHtP0zANW46nU87HdP02Pp81qPDUO+7+TOQV3WIu11kK61m5dZmX/ymo8WaB+XEIIIUSBKHVXPCGEEMWmjF3SVLETQghRWMq4Hzt1xRNCCCEKhFrshBBCFFZjjopvLFSxE0IIUVjK+IyduuIJIYQQBUItdkIIIQpLqIRtdqrY5dy9gr9ZjznQzIXVeP/Nu8FqPAAw0tRhPeatV49YjffhilJsEQrZf2LI9p89fZYXQAHYX0wGAN7++g2r8XS//pnVeACgKoMFiV6/K2I13r990LkyPmOnrnhCCCFEgVCLnRBCiMJSvo54qtgJIYQoMOqKJ4QQQohcoxY7IYQQhaWMS8pSxU4IIURhKeN0t3p1xT958gTjx49Hs2bNoKGhgZYtW2LGjBl49eoV2+WTSG5uLqZNmwZLS0twuVyYm5ujX79+SExM/MfLwuFwEBcX94/flxBCCAHqUbE/fPgQLi4uyMjIwL59+/DgwQNERUUhMTERHh4eyM/Pl0U5PykrKwsdO3bEH3/8gdWrV+PWrVuIj49H9+7dMXXq1H+0LJIqLy9v7CIQQohSYFhO8kDqin3q1KnQ0NDAqVOn0K1bN1hYWKB37944ffo0nj59igULFgAABAIBlixZgpEjR0JbWxvNmzfHpk2bxGIVFBRg4sSJMDY2hq6uLnr06IG0tDTR+bCwMDg6OmL37t0QCATQ09PDiBEj8PbtW1GeKVOmgMPh4PLlyxg8eDBsbGzQrl07BAcH488//xTly87OxoABA8Dn86Grq4thw4YhLy9PdN7f3x9+fn5i5Zs5cya8vLxEr728vDB9+nTMnTsXhoaGMDU1RVhYmOi8QCAAAAwcOBAcDkf0+v372LZtG1q1agUej4ddu3bByMgIZWVlYvf08/PDmDFjJP55EEII+TQhy0keSFWx5+fn4+TJk5gyZQo0NTXFzpmammL06NE4cOAAGKb6e83q1avRoUMH3LhxA/Pnz8eMGTOQkJAgumbo0KF4/vw5fv/9d1y7dg3Ozs7o2bOnWKs/MzMTcXFxOHbsGI4dO4bk5GSsWLFCVJ74+HhMnToV2to1V7vS19cHUL1a14ABA5Cfn4/k5GQkJCTg4cOHGD58uDRvHwAQExMDbW1tXLp0CatWrUJERIToPV25cgUAsHPnTuTk5IheA8CDBw9w6NAhxMbGIjU1FUOHDkVVVRWOHj0qyvP8+XMcP34c48ePl7pchBBCCCDl4LmMjAwwDANbW9taz9va2uL169d48eIFAMDT0xPz588HANjY2CAlJQXr16+Hj48Pzp8/j8uXL+P58+fgcrkAgDVr1iAuLg4HDx7EpEmTAFRXytHR0dDRqV5CdMyYMUhMTMSyZcvw4MEDMAyDtm3bfrbciYmJuHXrFh49egRzc3MAwK5du9CuXTtcuXIFrq6uEn8GDg4OCA0NBQBYW1tj48aNSExMhI+PD4yNjQFUf6EwNTUVu668vBy7du0S5QGAUaNGYefOnRg6dCgA4Ndff4WFhYVYLwEhhJD6o8FzEnrfIq+Lh4dHjdfp6ekAgLS0NBQVFcHIyAh8Pl+UHj16hMzMTNE1AoFAVKkDgJmZGZ4/fy5VOdLT02Fubi6q1AHAzs4O+vr6ovJIysHBQez1h+X5nJYtW4pV6gAQGBiIU6dO4enTpwCA6Oho+Pv7g8OpfX5GWVkZ3rx5I5Yk/QwIIUQZKeMzdqla7FZWVuBwOEhPT8fAgQNrnE9PT4eBgUGNCqw2RUVFMDMzQ1JSUo1z77vQAUBdXV3sHIfDEW2EYW1tDQ6Hg7t370rzNmqloqJSo5KsqKioke9z5fmc2h4VODk5oUOHDti1axe+/PJL/PXXXzh+/PgnY0RGRiI8PFy83Ko6UFPTq/P+hBBClINULXYjIyP4+Phg8+bNePfundi53Nxc7NmzB8OHDxe1OD8cvPb+9ftufGdnZ+Tm5kJNTQ1WVlZiqUmTJhKVx9DQEL6+vti0aROKi4trnC8oKABQ/YjgyZMnePLkiejcnTt3UFBQADs7OwCAsbExcnJyxK5PTU2VqBwfUldXR1VVlcT5J06ciOjoaOzcuRPe3t5ivQofCwkJQWFhoVhSVdWVuoyEEKIsaPCcBDZu3IiysjL4+vri7NmzePLkCeLj4+Hj44PmzZtj2bJlorwpKSlYtWoV7t+/j02bNuG3337DjBkzAADe3t7w8PCAn58fTp06haysLFy4cAELFizA1atXJS7Ppk2bUFVVBTc3Nxw6dAgZGRlIT0/Hjz/+KHoU4O3tDXt7e4wePRrXr1/H5cuXMXbsWHTr1g0uLtVblPbo0QNXr17Frl27kJGRgdDQUNy+fVvajwcCgQCJiYnIzc3F69ev68w/atQo/P3339i6dWudg+a4XC50dXXF0qe67QkhhFQ/Y2czyQOpK3Zra2tcvXoVlpaWGDZsGFq3bo1Jkyahe/fuuHjxIgwNDUV5Z82ahatXr8LJyQlLly7FunXr4OvrC6C6C/vEiRPo2rUrAgICYGNjgxEjRuDx48cwMTGRuDyWlpa4fv06unfvjlmzZqF9+/bw8fFBYmIitmzZIrrXkSNHYGBggK5du8Lb2xuWlpY4cOCAKI6vry8WLVqEuXPnwtXVFW/fvsXYsWOl/Xiwdu1aJCQkwNzcHE5OTnXm19PTw+DBg8Hn82tMtyOEEEKkxWFkNPpKIBBg5syZmDlzpizCK5SePXuiXbt2+PHHH6W+lsv7dNd9ffU3cWY13n/zbrAaDwCMNHXqziSl58UFrMZTUWF/jyVJxnNIi+0/APq8muNJGqqgtOajtoZ6++s3rMbT/fpnVuMBgKqKKusxK4WSPyqUhCz6DCvKn7IW6zvBCNZiAcD6rP2sxpMFWiu+Eb1+/RpJSUlISkrC5s2bG7s4hBCicOTluTibqGJvRE5OTnj9+jVWrlyJNm3aNHZxCCGEKACZVexZWVmyCq0w6DMihBDZYuRkwBubqMVOCCFEYSljVzz7o3sIIYQQ0mioxU4IIURhycvcczZRxU4IIURhKV+1ThW73FPhsP80he155xY6TVmNBwAPC3PqziQle0MBq/GM1PmsxgOAW28esx6T7RZNpRRLKktKFnOl9cb8wmq8a83YXf8BANxy01iPyfYaEO8qy1mNRxqOKnZCCCEKi7riCSGEEAVCo+IJIYQQwppNmzZBIBCAx+PB3d0dly9f/mTerVu3okuXLjAwMICBgQG8vb0/m/9TqGInhBCisBiW/5PGgQMHEBwcjNDQUFy/fh0dOnSAr68vnj9/Xmv+pKQkjBw5EmfOnMHFixdhbm6OL7/8Ek+fSrd2PlXshBBCFFZj7se+bt06BAYGIiAgAHZ2doiKioKWlhZ27NhRa/49e/ZgypQpcHR0RNu2bbFt2zYIhUIkJiZKdV+q2AkhhBCWlZeX49q1a/D29hYdU1FRgbe3Ny5evChRjJKSElRUVIhthy4Jqtg/8OLFC0yePBkWFhbgcrkwNTWFr68vUlJSAFTv6x4XF9e4hSSEECIxtrviy8rK8ObNG7FUVlZW474vX75EVVUVTExMxI6bmJggNzdXorLPmzcPzZo1E/tyIAmq2D8wePBg3LhxAzExMbh//z6OHj0KLy8vvHr1itX7VFRUsBqPEEJI7djuio+MjISenp5YioyMZL3cK1aswP79+3H48GHweDyprqWK/f8VFBTg3LlzWLlyJbp3746WLVvCzc0NISEh6N+/PwQCAQBg4MCB4HA4otcAsGXLFrRu3RoaGhpo06YNdu/eLRabw+Fgy5Yt6N+/P7S1tbFs2TIAwJEjR+Ds7AwejwdLS0uEh4ejsrLyn3rLhBBCpBQSEoLCwkKxFBISUiNfkyZNoKqqiry8PLHjeXl5MDU1/ew91qxZgxUrVuDUqVNwcHCQuoxUsf8/Pp8PPp+PuLi4WrtVrly5AgDYuXMncnJyRK8PHz6MGTNmYNasWbh9+za++eYbBAQE4MyZM2LXh4WFYeDAgbh16xbGjx+Pc+fOYezYsZgxYwbu3LmDn3/+GdHR0aJKnxBCSMMJGYbVxOVyoaurK5a4XG6N+2poaKBjx45iA9/eD4Tz8PD4ZHlXrVqFJUuWID4+Hi4uLvV6z1Sx/z81NTVER0cjJiYG+vr68PT0xPfff4+bN28CAIyNjQEA+vr6MDU1Fb1es2YN/P39MWXKFNjY2CA4OBiDBg3CmjVrxOKPGjUKAQEBsLS0hIWFBcLDwzF//nyMGzcOlpaW8PHxwZIlS/Dzzz//s2+cEEIUGMNykkZwcDC2bt2KmJgYpKenY/LkySguLkZAQAAAYOzYsWKt/ZUrV2LRokXYsWMHBAIBcnNzkZubi6KiIqnuSxX7BwYPHoxnz57h6NGj6NWrF5KSkuDs7Izo6OhPXpOeng5PT0+xY56enkhPTxc79vE3r7S0NERERIh6Cvh8PgIDA5GTk4OSkpJa71XboA2GUb7lEgkhRB4MHz4ca9asweLFi+Ho6IjU1FTEx8eLBtRlZ2cjJ+d/+15s2bIF5eXlGDJkCMzMzETp44ZiXWhJ2Y/weDz4+PjAx8cHixYtwsSJExEaGgp/f/8GxdXW1hZ7XVRUhPDwcAwaNKjWMtQmMjIS4eHhYsdUVXWhrq7foLIRQoiiauy14oOCghAUFFTruaSkJLHXWVlZrNyTWux1sLOzQ3FxMQBAXV0dVR/tXGVrayuaDvdeSkoK7OzsPhvX2dkZ9+7dg5WVVY2kolL7j6W2QRtqanoNeHeEEKLYGnPlucZCLfb/9+rVKwwdOhTjx4+Hg4MDdHR0cPXqVaxatQoDBgwAAAgEAiQmJsLT0xNcLhcGBgaYM2cOhg0bBicnJ3h7e+O///0vYmNjcfr06c/eb/Hixejbty8sLCwwZMgQqKioIC0tDbdv38bSpUtrvYbL5dYYpMHhyGJDS0IIIfKKWuz/j8/nw93dHevXr0fXrl3Rvn17LFq0CIGBgdi4cSMAYO3atUhISIC5uTmcnJwAAH5+fvjhhx+wZs0atGvXDj///DN27twJLy+vz97P19cXx44dw6lTp+Dq6opOnTph/fr1aNmypazfKiGEKI3GXFK2sXAYGn0l1zQ12f8iwPavhIVOU1bjAcDDwpy6M0nJ3lDAajwjdT6r8QDg1pvHrMdk+xlk5UePq9hQVP6O9Zhs93ZdNXNiNR4AuOWmsR5Tj6vFarx3leWsxgOAN8UPWYs1tOUA1mIBwG+Pj7AaTxaoxU4IIYQoEHrGTgghRGHJy4A3NlHFTgghRGHJy3NxNlFXPCGEEKJAqMVOCCFEYSnj+HCq2AkhhCisxl55rjFQVzwhhBCiQKjFTgghRGEp4+A5qtjlnCyeH1UK2V1gJLckn9V4ACCLhXQLKqTbGvGfjgcAhWW17/z3b8L274+sGPLYXUDI+dl1VuMBwNvfQ1mPafBV7UtWKyplnO5GXfGEEEKIAqEWOyGEEIWljIPnqGInhBCisJRxuht1xRNCCCEKhFrshBBCFBaNiieEEEIUCI2KVyIvXrzA5MmTYWFhAS6XC1NTU/j6+iIlJQVA9V7NcXFx/1h5/un7EUIIUUxK22IfPHgwysvLERMTA0tLS+Tl5SExMRGvXr2SOEZ5eTk0NDRkWEpCCCENoYyj4pWyxV5QUIBz585h5cqV6N69O1q2bAk3NzeEhISgf//+EAgEAICBAweCw+GIXoeFhcHR0RHbtm1Dq1atwOPxRPEmTpwIY2Nj6OrqokePHkhLSxO755EjR+Ds7AwejwdLS0uEh4ejsrISAD55P0IIIQ3DMAyrSR4oZcXO5/PB5/MRFxeHsrKyGuevXLkCANi5cydycnJErwHgwYMHOHToEGJjY5GamgoAGDp0KJ4/f47ff/8d165dg7OzM3r27In8/OoV186dO4exY8dixowZuHPnDn7++WdER0dj2bJldd6PEEIIkYZSVuxqamqIjo5GTEwM9PX14enpie+//x43b94EABgbGwMA9PX1YWpqKnoNVHe/79q1C05OTnBwcMD58+dx+fJl/Pbbb3BxcYG1tTXWrFkDfX19HDx4EAAQHh6O+fPnY9y4cbC0tISPjw+WLFmCn3/+uc77EUIIqT8hGFaTPFDqZ+xfffUVzp07hz///BO///47Vq1ahW3btsHf3/+T17Vs2VKs4k1LS0NRURGMjIzE8r179w6ZmZmiPCkpKaIWOgBUVVWhtLQUJSUl0NLSkqjMZWVlNXoYGIYBhyOLldMJIUT+KeOoeKWt2AGAx+PBx8cHPj4+WLRoESZOnIjQ0NDPVuza2tpir4uKimBmZoakpKQaefX19UV5wsPDMWjQoFrLIKnIyEiEh4eLHVNV1YWamp7EMQghhCg2pa7YP2ZnZyeacqauro6qqrp3qXJ2dkZubi7U1NQ+OejN2dkZ9+7dg5WV1SfjSHK/kJAQBAcHix0zNm5XZxkJIURZCeVkwBublLJif/XqFYYOHYrx48fDwcEBOjo6uHr1KlatWoUBAwYAqB6pnpiYCE9PT3C5XBgYGNQay9vbGx4eHvDz88OqVatgY2ODZ8+e4fjx4xg4cCBcXFywePFi9O3bFxYWFhgyZAhUVFSQlpaG27dvY+nSpRLfj8vlgsvlih2jbnhCCPk05avWlXTwHJ/Ph7u7O9avX4+uXbuiffv2WLRoEQIDA7Fx40YAwNq1a5GQkABzc3M4OTl9MhaHw8GJEyfQtWtXBAQEwMbGBiNGjMDjx49hYmICAPD19cWxY8dw6tQpuLq6olOnTli/fj1atmwpiiPp/QghhJDP4TDyMjGP1IrHs2A9ZqWw7kcQ0tBS59adSUrvKmpOU2yoFjpNWI/Jtpzi141dhDqx/fsjK0aaOqzGe/XuLavxAODt76GsxzT4ainrMdlWWprNWizP5j1YiwUAKU//YDWeLChlVzwhhBDlIC9T1NiklF3xhBBCiKKiFjshhBCFpYxPm6liJ4QQorCoK54QQgghco1a7IQQQhQWLSlLCCGEKBBlfMZOXfGEEEKIjGzatAkCgQA8Hg/u7u64fPnyJ/P+9ddfGDx4MAQCATgcDjZs2FCve1KLXc4JGWFjF6FOJTJYTEYWXpWyu8CILBbmkQW2y1lU/o7VeIBs1vt+V1nOajyumjqr8QCgud9q1mMW/vUfVuNp2w5mNR7bGnPw3IEDBxAcHIyoqCi4u7tjw4YN8PX1xb1799C0adMa+UtKSmBpaYmhQ4fiu+++q/d9qcVOCCFEYTEMw2qSxrp16xAYGIiAgADY2dkhKioKWlpa2LFjR635XV1dsXr1aowYMaLGviDSoIqdEEIIkVBZWRnevHkjlsrKavZKlpeX49q1a/D29hYdU1FRgbe3Ny5evCjTMlLFTgghRGEJwbCaIiMjoaenJ5YiIyNr3Pfly5eoqqoSbQb2nomJCXJzc2X6nukZOyGEEIXF9nS3kJAQBAcHix1rSLe5LFDFTgghhEiIy+VKVJE3adIEqqqqyMvLEzuel5cHU1NTWRUPAHXFf1JSUhI4HA4KCgpkEl8gENR7KgMhhBDJCBmG1SQpDQ0NdOzYEYmJif8ri1CIxMREeHh4yOKtiihsxf7ixQtMnjwZFhYW4HK5MDU1ha+vL1JSUhq7aIQQQv4hDMv/SSM4OBhbt25FTEwM0tPTMXnyZBQXFyMgIAAAMHbsWISEhIjyl5eXIzU1FampqSgvL8fTp0+RmpqKBw8eSHVfhe2KHzx4MMrLyxETEwNLS0vk5eUhMTERr169atRylZeXQ0NDo1HLQAghRPaGDx+OFy9eYPHixcjNzYWjoyPi4+NFA+qys7OhovK/9vWzZ8/g5OQker1mzRqsWbMG3bp1Q1JSksT3VcgWe0FBAc6dO4eVK1eie/fuaNmyJdzc3BASEoL+/fsjKysLHA4HqampYtdwOJwaH15KSgocHBzA4/HQqVMn3L59W3QuLCwMjo6OYvk3bNgAgUAgeu3v7w8/Pz8sW7YMzZo1Q5s2bUTn3r59i5EjR0JbWxvNmzfHpk2b2PwYCCFE6TVWV/x7QUFBePz4McrKynDp0iW4u7uLziUlJSE6Olr0WiAQ1Dp3XppKHVDQip3P54PP5yMuLq7W+YXSmDNnDtauXYsrV67A2NgY/fr1Q0VFhVQxEhMTce/ePSQkJODYsWOi46tXr0aHDh1w48YNzJ8/HzNmzEBCQkKDyksIIeR/GrMrvrEoZMWupqaG6OhoxMTEQF9fH56envj+++9x8+ZNqWOFhobCx8cH9vb2iImJQV5eHg4fPixVDG1tbWzbtg3t2rVDu3btRMc9PT0xf/582NjYYNq0aRgyZAjWr18vdRkJIYSQ9xSyYgeqn7E/e/YMR48eRa9evZCUlARnZ2exbg9JfDh60dDQEG3atEF6erpUMezt7Wt9rv7xyEgPD4/Pxq5txSNl3LmIEEIk1dhd8Y1BYSt2AODxePDx8cGiRYtw4cIF+Pv7IzQ0VDRY4cNKUdrudaB6ecCPK9ba4mhra0sduza1rXgkrGJ34xJCCFEk1BWv4Ozs7FBcXAxjY2MAQE5OjujchwPpPvTnn3+K/v/169e4f/8+bG1tAQDGxsbIzc0Vq9w/Faeu2O9fv49dm5CQEBQWFoolFVUdie9HCCFE8SnkdLdXr15h6NChGD9+PBwcHKCjo4OrV69i1apVGDBgADQ1NdGpUyesWLECrVq1wvPnz7Fw4cJaY0VERMDIyAgmJiZYsGABmjRpAj8/PwCAl5cXXrx4gVWrVmHIkCGIj4/H77//Dl1dXYnKmZKSglWrVsHPzw8JCQn47bffcPz48U/mr23FIw6HI9mHQgghSkheus/ZpJAtdj6fD3d3d6xfvx5du3ZF+/btsWjRIgQGBmLjxo0AgB07dqCyshIdO3bEzJkzsXTp0lpjrVixAjNmzEDHjh2Rm5uL//73v6Ln5ba2tti8eTM2bdqEDh064PLly5g9e7bE5Zw1axauXr0KJycnLF26FOvWrYOvr2/DPwBCCCEAlLMrnsPQ6Cu5psFtwXpMZfyGCwBa6uxu5MB2PAAoKC1mPSbb5Swqf8dqPEA2v5Nsv+8qRshqPADgqqqzHjMndTer8bRtB7MaDwAqy5+yFsuyiVPdmaTw8OUNVuPJgkJ2xRNCCCEAwMjgC9e/HVXshBBCFJZQTrrP2aSQz9gJIYQQZUUtdkIIIQpLGYeRUcVOCCFEYVFXPCGEEELkGrXYCSGEKCzqiidyR12V/R9hWaX06+Z/jpqKKqvxAKBSWMV6TBvd5qzGu1v4N6vxAICvwWM9ZnlVJavxZDHn3ECTz3rMgndFrMZ7vwcFm9j+twiwP+/87e+hrMZjmzKuy0Fd8YQQQogCoRY7IYQQhSUvy8CyiSp2QgghCksZn7FTVzwhhBCiQKjFTgghRGHRPPZ/GYFAgA0bNoheczgcxMXFsX4fLy8vzJw585P3rc2HZcnKygKHw0FqairrZSOEEFJ/DMOwmuSBTCt2Dofz2RQWFsZafDU1NVhYWCA4OBhlZWVSxYmNjcWSJUvqXQ5zc3Pk5OSgffv29Y5BCCGEsEGmXfE5OTmi/z9w4AAWL16Me/fuiY7x+Q2fm7pz50706tULFRUVSEtLQ0BAALS1taWqqA0NDRtUBlVVVZiamjYoBiGEEPbRPHaWmZqaipKenh44HI7odXFxMUaPHg0TExPw+Xy4urri9OnTUt9DX18fpqamMDc3R9++fTFgwABcv35ddN7f3x9+fn5i18ycORNeXl6i1x93xX8sIyMDXbt2BY/Hg52dHRISEsTOf9wVn5SUBA6Hg8TERLi4uEBLSwudO3cW+1IDAEuXLkXTpk2ho6ODiRMnYv78+XB0dJT6MyCEEFI76or/BxUVFaFPnz5ITEzEjRs30KtXL/Tr1w/Z2dn1jnn//n388ccfcHd3Z62cQqEQgwYNgoaGBi5duoSoqCjMmzdPomsXLFiAtWvX4urVq1BTU8P48eNF5/bs2YNly5Zh5cqVuHbtGiwsLLBlyxbWyk0IIUQ5Ndqo+A4dOqBDhw6i10uWLMHhw4dx9OhRBAUFSRxn5MiRUFVVRWVlJcrKytC3b1+EhISwVs7Tp0/j7t27OHnyJJo1awYAWL58OXr37l3ntcuWLUO3bt0AAPPnz8dXX32F0tJS8Hg8/PTTT5gwYQICAgIAAIsXL8apU6dQVMTuMpeEEKLMaFT8P6ioqAizZ8+Gra0t9PX1wefzkZ6eLnWLff369UhNTUVaWhqOHTuG+/fvY8yYMayVMz09Hebm5qJKHQA8PDwkutbBwUH0/2ZmZgCA58+fAwDu3bsHNzc3sfwfv/5YWVkZ3rx5I5bkpWuIEEIagzJ2xTdai3327NlISEjAmjVrYGVlBU1NTQwZMgTl5eVSxTE1NYWVlRUAoE2bNnj79i1GjhyJpUuXwsrKCioqKjV+GBUV7G+sUBt1dXXR/3M4HADVXfv1FRkZifDwcLFjamp60FDXr3dMQgghiqXRWuwpKSnw9/fHwIEDYW9vD1NTU2RlZTU4rqpq9U5i7969AwAYGxuLjc4HINV8c1tbWzx58kQsxp9//tngcrZp0wZXrlwRO/bx64+FhISgsLBQLKmr6TW4LIQQoqiEDMNqkgeNVrFbW1sjNjZW1I0+atSoerVmCwoKkJubi2fPniE5ORkRERGwsbGBra0tAKBHjx64evUqdu3ahYyMDISGhuL27dsSx/f29oaNjQ3GjRuHtLQ0nDt3DgsWLJC6nB+bNm0atm/fjpiYGGRkZGDp0qW4efOmqGVfGy6XC11dXbH0ufyEEEKUT6NV7OvWrYOBgQE6d+6Mfv36wdfXF87OzlLHCQgIgJmZGVq0aIGRI0eiXbt2+P3336GmVv2UwdfXF4sWLcLcuXPh6uqKt2/fYuzYsRLHV1FRweHDh/Hu3Tu4ublh4sSJWLZsmdTl/Njo0aMREhKC2bNnw9nZGY8ePYK/vz94PPb32yaEEGXFsPyfPOAw8jIaQAn4+PjA1NQUu3fvlvgabS0B6+Uoq2R3DIKaiiqr8QCgUljFekxHI0tW490t/JvVeADAU1OvO5OUyqsqWY1XUiHdyo+SMNBs+GJWHyt4x+4MFBUV9ttJVQ0Yk/NPeft7KOsxNXtOYi+WZkvWYgHAu3ePWY0nC7QJTCMpKSlBVFQUfH19oaqqin379uH06dM1Fr8hhBBCpEEVeyPhcDg4ceIEli1bhtLSUrRp0waHDh2Ct7d3YxeNEEIUhjJ2SlPF3kg0NTXrtYQuIYQQycnLc3E2/au3bSWEEEKIdKhiJ4QQorAae+W5TZs2QSAQgMfjwd3dHZcvX/5s/t9++w1t27YFj8eDvb09Tpw4IfU9qWInhBCisBqzYj9w4ACCg4MRGhqK69evo0OHDvD19RUtLf6xCxcuYOTIkZgwYQJu3LgBPz8/+Pn5SbX2CkDT3eQeTXdjD013YwdNd2MPTXdrOHWN5qzFAoCK8qcS53V3d4erqys2btwIoHpJcXNzc0ybNg3z58+vkX/48OEoLi7GsWPHRMc6deoER0dHREVFSXxfarETQghRWAzLqbbNuMrKan6ZLS8vx7Vr18RmOqmoqMDb2xsXL16stawXL16sMTPK19f3k/k//aaJwistLWVCQ0OZ0tJSpYopD2WURUx5KKMsYspDGWURUx7KKKuYjSE0NLRGfR8aGloj39OnTxkAzIULF8SOz5kzh3Fzc6s1trq6OrN3716xY5s2bWKaNm0qVRmpYlcChYWFDACmsLBQqWLKQxllEVMeyiiLmPJQRlnElIcyyipmYygtLWUKCwvFUm1fVhqzYqd57IQQQoiEuFwuuFxunfmaNGkCVVVV5OXliR3Py8uDqalprdeYmppKlf9T6Bk7IYQQwjINDQ107NgRiYmJomNCoRCJiYnw8PCo9RoPDw+x/ACQkJDwyfyfQi12QgghRAaCg4Mxbtw4uLi4wM3NDRs2bEBxcTECAgIAAGPHjkXz5s0RGRkJAJgxYwa6deuGtWvX4quvvsL+/ftx9epV/PLLL1Ldlyp2JcDlchEaGipR95EixZSHMsoipjyUURYx5aGMsogpD2WUVcx/u+HDh+PFixdYvHgxcnNz4ejoiPj4eJiYmAAAsrOzxaZJdu7cGXv37sXChQvx/fffw9raGnFxcWjfvr1U96V57IQQQogCoWfshBBCiAKhip0QQghRIFSxE0IIIQqEKnZCCCFEgVDFTiSWnZ1d6+5GDMMgOzu7EUpECCHkY1SxE4m1atUKL168qHE8Pz8frVq1qlfMs2fPorKy5u5ilZWVOHv2bL1iEuVTXl6Oe/fu1fq7JK1x48bJ9HevtLS0wTEsLS3x6tWrGscLCgpgacnuLoX15eTkBGdn5xqpY8eO8PT0xLhx43DmzJnGLqZCooqdSIxhGHA4nBrHi4qKwOPx6hWze/fuyM/Pr3G8sLAQ3bt3r1fM0NBQPH78uF7X/lMEAgEiIiKop6OBSkpKMGHCBGhpaaFdu3aiz3PatGlYsWJFvWIWFhbC29sb1tbWWL58OZ4+lXybzk8RCoVYsmQJmjdvDj6fj4cPHwIAFi1ahO3bt0sdLysrC1VVNbcuLisra1B5ExMT0bdvX7Ru3RqtW7dG3759cfr06XrF6tWrFx4+fAhtbW10794d3bt3B5/PR2ZmJlxdXZGTkwNvb28cOXKk3uUltaMFahRYaWkpfvrpJ5w5cwbPnz+H8KO9na9fvy5RnODgYAAAh8PBokWLoKWlJTpXVVWFS5cuwdHRsV5l/NSXhVevXkFbW7teMY8cOYJly5ahW7dumDBhAgYPHiz1ohhOTk61lqs2kn6OH5o5cyaio6MRERGB7t27Y8KECRg4cGCDFu8oLi7GihUrkJiYWOvP+31lIimGYXDw4MFP/v7ExsbWq5yJiYmfLOOOHTukihUSEoK0tDQkJSWhV69eouPe3t4ICwurdc/rusTFxeHFixfYvXs3YmJiEBoaCm9vb0yYMAEDBgyAurq61DGXLl2KmJgYrFq1CoGBgaLj7du3x4YNGzBhwgSJ4hw9elT0/ydPnoSenp7odVVVFRITEyEQCKQuHwBs3rwZM2bMwJAhQzBjxgwAwJ9//ok+ffpg/fr1mDp1qlTxXr58iVmzZmHRokVix5cuXYrHjx/j1KlTCA0NxZIlSzBgwIB6lZnUjhaoUWCjR4/GqVOnMGTIEJiYmNSoqEJDQyWK877lnJycDA8PD2hoaIjOaWhoQCAQYPbs2bC2tpa4bIMGDQJQXQn36tVLrEKrqqrCzZs30aZNG8THx0sc80M3btzAzp07sW/fPlRWVmLEiBEYP348XF1dJbo+PDxc4ntJ+jnW5vr164iOjsa+fftQVVWFUaNGYfz48XB2dpY61siRI5GcnIwxY8bAzMysxs/7/R9rSc2YMQM///wzunfvXuvvz86dO6UuY3h4OCIiIuDi4lJrGQ8fPixVvJYtW+LAgQPo1KkTdHR0kJaWBktLSzx48ADOzs548+aN1GX82PXr17Fz505s27YNfD4fX3/9NaZMmSLV77uVlRV+/vln9OzZU6ycd+/ehYeHB16/fi1RnPerlHE4nBrjXdTV1SEQCLB27Vr07dtX8jf4/1q0aIH58+cjKChI7PimTZvq1XOhp6eHa9euwcrKSuz4gwcP0LFjRxQWFuLu3btwdXXF27dvpS4v+Qyp9oIjckVXV5c5f/48a/H8/f1Z23LR39+f8ff3ZzgcDjN8+HDRa39/f2bSpEnM8uXLmRcvXjT4PuXl5cyhQ4eYvn37Murq6oy9vT2zYcMGpqCggIV3wZ7y8nJmw4YNDJfLZVRUVJgOHTow27dvZ4RCocQx9PT0WP15GxgYMMePH2ctHsMwjKmpKbNr1y7W4mlqajKZmZkMwzAMn88X/X9qaiqjq6vb4PjPnj1jVqxYwbRp04bR1tZmxo4dy/Ts2ZNRU1Nj1q1bJ3EcHo/HZGVl1SjnX3/9xWhra0tdLoFAwMq/jw9pa2szGRkZNY7fv3+/XmVs2rQpExMTU+N4TEyMaBvSv/76i2nSpIn0hSWfRV3xCqx58+bQ0dFhLd7HLbQ3b97gjz/+QNu2bdG2bdt6xXrf2q9vt3tdGIZBRUUFysvLwTAMDAwMsHHjRixatAhbt27F8OHDZXJfSVVUVODw4cPYuXMnEhIS0KlTJ0yYMAF///03vv/+e5w+fRp79+6VKJaBgQEMDQ1ZK5uenh7rA7HKy8vRuXNn1uK5uLjg+PHjmDZtGgCIegC2bdsm9Y5Y71VUVODo0aPYuXMnTp06BQcHB8ycOROjRo2Crq4ugOqehfHjx+O7776TKKadnR3OnTuHli1bih0/ePAgnJycpC7jo0ePRP9fWlpa7zEuH+rfvz8OHz6MOXPmiB0/cuRIvXoApk2bhm+//RbXrl0T9ZRduXIF27Ztw/fffw+g+nFCfR/jkc9o7G8WRHZOnDjB9OrVS9RSaKihQ4cyP/30E8MwDFNSUsJYW1sz6urqjJqaGnPw4MF6xSwpKWGKi4tFr7Oyspj169czJ0+ebFBZr169ykydOpUxNDRkzMzMmHnz5om1Rn788UdRq6EulZWVzOrVqxlXV1fGxMSEMTAwEEv1ce3aNSYoKIgxMjJijI2NmVmzZjHp6elieW7dusXweDyJY+7evZsZMmSI2OfZENHR0cyIESOYkpISVuIxDMPMnTuXiYiIYC3euXPnGD6fz3z77bcMj8djZsyYwfj4+DDa2trM1atX6xXTyMiIMTAwYKZMmcLcuHGj1jyvX79mBAKBxDHj4uIYPT09ZsWKFYyWlhazevVqZuLEiYyGhgZz6tQpqctYVVXFREREMM2aNWNUVVVFPQALFy5ktm3bJnGcH374QZSWLFnC6OnpMX369GGWLFnCLFmyhPnqq68YfX19ZsmSJVKXkWEY5tdff2U6deok+rfSqVMnZs+ePaLzJSUlzLt37+oVm3waVewK7Pnz54yXlxejoqLC8Pn8BldIJiYmTGpqKsMwDLNnzx7GysqKKS4uZjZv3sw4OjrWq4w+Pj7Mli1bGIap/mPZtGlTpkWLFgyPx2M2b95cr5jt27dn1NTUmD59+jCHDx9mKisra+R58eIFw+FwJIq3aNEixszMjFmzZg3D4/GYJUuWMBMmTGCMjIyYH374oV5lVFFRYXx9fZn//Oc/THl5ea15ioqKGH9/f4ljOjo6Mjo6Ogyfz2fat2/PODk5iSVplZSUML6+vqzFYxiGmT59OqOvr8907dqVCQoKYr777juxVB8PHjxgJk6cyLi6ujK2trbM6NGjmZs3b9YrFsMwzK5du2RS2Zw9e5bx9vZmjI2NGU1NTcbT07PeX2DDw8MZS0tL5tdffxV7HLF//36mU6dOEscRCAQSpVatWtWrnKRx0OA5Bebt7Y3s7GxMmDCh1sFP48aNkyqepqYm7t+/D3Nzc4wdOxbNmjXDihUrkJ2dDTs7OxQVFUldxiZNmiA5ORnt2rXDtm3b8NNPP+HGjRs4dOgQFi9ejPT0dKljLlmyBOPHj0fz5s2lvrY2rVu3xo8//oivvvoKOjo6SE1NFR37888/Je4q/9Djx49rdMs2VF0D/qQd5Dds2DCcOXOmwYMvP/S5KYwcDgd//PGH1DHZVFFRAU1NTaSmpkq9VeanVFZWYvny5Rg/fjxatGjBSky2BuP9E8rLy2udAWFhYdFIJVICjf3NgsiOpqamqIXNBmtra+bAgQNMUVERY2xszCQmJjIMUz1QycjIqN5lfPz4McMw1V39YWFhDMMwTHZ2NqOpqclOwRtIS0tLVEZTU1Pm2rVrDMMwTGZmZr0HaLVq1Yp5+fJljeOvX7/+17SOtLS0mHPnzjV2MT7r+PHjTHx8fI3j8fHxzIkTJ+oVs1WrVqz+u2GY6oFpjx49Yi0e24PxysvLGUtLS+bOnTuslfH+/fvMF198waioqIglDofDqKiosHYfUhMNnlNgbdu2xbt371iLN3PmTIwePRp8Ph8tW7aEl5cXgOrV4+zt7esV08rKCnFxcRg4cCBOnjwpGoz0/Plz0UCl+vj7779x9OhRZGdno7y8XOzcunXrpIrVokUL5OTkwMLCAq1bt8apU6fg7OyMK1eu1HveuawWGCkoKMDBgweRmZmJOXPmwNDQENevX4eJiYnUPRjm5uYN+hn8E+bPn1/rQjQMw2D+/Pno3bu31DEXLFiA77//Hrt372ZtMGLPnj2RnJxc7znmH2N7MJ66ujorK+J9yN/fH2pqajh27FitUxuJ7FDFrsBWrFiBWbNmYdmyZbC3t6+xsIa0f7SnTJkCd3d3ZGdnw8fHRzSn1tLSEkuXLq1XGRcvXoxRo0bhu+++Q48ePUQjmU+dOlWvP1BA9QIo/fv3F3VNtm/fHllZWWAYpl7zwwcOHIjExES4u7tj2rRp+Prrr7F9+3ZkZ2dLPCr6PVkuMHLz5k14e3tDT08PWVlZCAwMhKGhIWJjY5GdnY1du3ZJFW/t2rWYO3cuoqKiGlQhDRo0CNHR0dDV1RWtX/Ap0i56k5GRATs7uxrH27ZtiwcPHkgV672NGzfiwYMHaNasGVq2bFljxkZ9FiTq3bs35s+fj1u3bqFjx441Yvbv31+qeIsXL8a4cePw9OlTCIVCxMbG4t69e9i1axeOHTsmdfkAYOrUqVi5ciW2bdsGNbWGVw2pqam4du2a1DNmSMPRM3YF9uFiFh9i/n+1t9pajI0hNzcXOTk56NChg6jMly9fhq6ubr3+KLi5uaF3794IDw8XPX9s2rQpRo8ejV69emHy5MkNKu/Fixdx8eJFWFtbo1+/flJdK8sFRry9veHs7IxVq1aJPXe9cOECRo0ahaysLKniGRgYoKSkBJWVldDS0qrxxbC2pYBrExAQgB9//BE6OjoICAj4bF5pF70xNTXF3r170aNHD7Hjp0+fxqhRo/D8+XOp4gFAWFjYZ1uX9Rlb8P7nXpv6/ls8d+4cIiIikJaWhqKiIjg7O2Px4sX48ssvpY4F/O8LLJ/Ph729fY0vH9J+6XJ1dcX69evxxRdf1Ks8pP6oYldgycnJnz3frVs3qWOy2cX9oQcPHiAzMxNdu3aFpqbmJ5ealcSHA9wMDAxw/vx5tGvXDmlpaRgwYIDUFZwstGrVCleuXEGTJk1Yi6mnp4fr16+jdevWYhX748eP0aZNG6m7WmNiYj57XtrBl7LwzTff4OLFizh8+DBat24NoPp3afDgwXB1dcW2bdsauYTyg+0vXX/88QcWLlyI5cuXs9JjSCRHXfEKrD4V9+ew3cUNVK8J/370NYfDQUZGBiwtLTFhwgQYGBhg7dq1UsfU1tYWfekwMzNDZmYm2rVrB6B6/WpJHD16FL1794a6urpY93ltpO1GBcQXGGELl8utdQnV+/fvw9jYWOp4/4aKuy6rVq1Cr1690LZtW9GI87///htdunTBmjVr6hXT0tISV65cgZGRkdjxgoICODs7S73mvqy8H0/x8OFDzJ49u0HjKYD6LRH8Od7e3gCqxxd86N/WY6iIqGJXcAUFBdi+fbto2li7du0wfvx4sWe7kgoJCcHs2bNFXdyHDh0S6+Kuj++++w7q6urIzs6Gra2t6Pjw4cMRHBxcr4q9U6dOOH/+PGxtbdGnTx/MmjULt27dQmxsLDp16iRRDD8/P+Tm5qJp06bw8/P7ZD5p/kD9+OOPmDRpEng8Hn788cfP5p0+fbpEMT/Uv39/RERE4D//+Y+obNnZ2Zg3bx4GDx4sdTyg+rl/XFyc2O9P//79oaqqWq94QPUAr//85z+19vpI+/xaT08PFy5cQEJCAtLS0qCpqQkHBwd07dq13uX73MDGv//+u95xk5OTsWbNGtFnaWdnhzlz5qBLly5Sx/p4PMXEiRMbNJ5CFmhL1kbUSKPxyT/gypUrjKGhIdO8eXNm4MCBzMCBA5kWLVowRkZGoilb0uDz+cyDBw8YhmEYfX195vbt2wzDVE93a9myZb3K+OGiNx9O28nMzKzXtJ3316alpTEMU73IyzfffMPY29szgwYNYm0VvvoQCASiKW6yWAykoKCA8fb2ZvT19RlVVVXG3NycUVdXZ7p27coUFRVJHS8jI4OxtrZmtLS0RIvSaGlpMW3atBH9Hkjrhx9+YPh8PhMUFMRoaGgw33zzDePt7c3o6ekx33//fb1isuXIkSPMkSNHGA6Hw+zatUv0+siRI0xsbCwzdepUxsbGpl6xd+/ezaipqTHDhg0TrfQ2bNgwRl1dXWwlNkn17NmTmTNnDsMw4v9uUlJS6v1v8f3v3qcSkR/0jF2BdenSBVZWVti6datolGtlZSUmTpyIhw8f4uzZs1LFMzU1xZkzZ2Braws7OzusWLEC/fv3R1paGjw9Peu1QI2Ojg6uX78Oa2trsefCV69eha+vL169eiV1TDZVVFSgV69eiIqKkmo3r8Z0/vx53Lx5UzSg6n2XqLT69OkDhmGwZ88e0bSvV69e4euvv4aKigqOHz8udcy2bdsiNDQUI0eOFPt5L168GPn5+di4cWOdMWTV8yHLgY22traYNGlSjVkU69atw9atW6VeiInt8RQA8MMPP4i9rqiowI0bNxAfH485c+ZItAXuzZs30b59e6ioqODmzZufzevg4CB1GYmEGvd7BZElHo9XY/1xhqlexKI+i78MGDCA+eWXXxiGYZhZs2YxVlZWzNKlSxlnZ2emZ8+e9Spj7969mYULFzIMU93yePjwIVNVVcUMHTqUGTx4cL1iLlq0iPnjjz9YWxa0SZMmzP3791mJxTAMU1hYyFRVVdU4XlVVxdrueWzQ0tKqdWnW1NTUevemaGpqinpNjI2NRb019+/fZwwNDSWKIeueD1nsnKahoVHrzmkZGRkMl8uVOp6xsTFz/fp1hmHEW+ynTp1iWrRo0bDCfmTjxo0SL23M4XCYvLw80f+/X5Dm40QL1MgWPWNXYLq6usjOzq4xZezJkyf12vVt3bp1olZ5eHg4ioqKcODAAVhbW9d7RPyqVavQs2dPXL16FeXl5Zg7dy7++usv5OfnIyUlpV4xL168iHXr1qGyshKurq7o1q0bvLy84OnpCU1NTanjvZ+3XttCKNI6fPgw5s2bh9TUVGhpaYmde/fuHVxdXbFmzRqpp9G9l5iYiMTExFqX8NyxY4dUsbhcbq37ZBcVFUFDQ6Ne5TM1NUV+fj5atmwJCwsL/Pnnn+jQoQMePXpUo5X8KampqaIxIrIYhCiLmObm5khMTKyxN/np06dhbm4udTxZjKf4lN69eyMkJESiwXWPHj0SDdSUxedIJEMVuwIbPnw4JkyYgDVr1oi2ykxJScGcOXMwcuRIqeN9uIWntrY2oqKiGlzG9u3b4/79+9i4cSN0dHRQVFSEQYMGYerUqTAzM6tXzISEBFRWVuLSpUs4e/YskpOT8eOPP6KsrAyurq44f/68VPEqKyuxY8cOnD59utbFRaT5UrNlyxbMnTu3RqUOVH+m8+bNw8aNG+tVsYeHhyMiIgIuLi6srPTVt29fTJo0Cdu3b4ebmxsA4NKlS/j222/rNRMAAHr06IGjR4/CyckJAQEB+O6773Dw4EFcvXq1zsVr3jM0NEROTg6aNm2KHj16IDY2Fvr6+vUqz6ew+QUJAGbNmoXp06cjNTVV7N9idHR0jS5wSaxduxZDhgxB06ZN8e7dO3Tr1g25ubnw8PDAsmXLpI73OQcPHpR4Bb4PV8Jjey8EIjl6xq7AysvLMWfOHERFRaGyshJA9bPCyZMnY8WKFfVeDvXatWtio6Tru0LcP/H8+v79+zhz5gxOnz6NuLg46OnpSTzl7b3PbVwCSDf6t1mzZjh79myNltt7Dx48QNeuXfHs2TOpyghUT+1btWoVxowZI/W1tSkoKMC4cePw3//+VzQHubKyEv3790d0dHS9ZlYIhUIIhULRmI/9+/fjwoULsLa2xjfffCNRT4Cenh7+/PNP2NraQkVFBXl5efWazvcpdX1BOnz4cL3iHj58GGvXrhX927G1tcWcOXMwYMCAepc1JSVFbIGa+o6nAAAnJyex98owDHJzc/HixQts3rwZkyZNqjNGXVNDP1TfL4ekblSxK4GSkhJkZmYCqN6prLbWoiSeP3+OESNGICkpSdRCKigoQPfu3bF///56/XE1NjYW/WFnyy+//IKkpCQkJyejrKwMXbp0gZeXF7y8vODg4NCoa1Zramrixo0bn1xRLz09Hc7OzvVa49/IyAiXL18WLdTCloyMDKSnp4PD4cDW1vaTX0okkZ2dDXNz81pXQ3zy5IlEO34NHjwYKSkpsLW1RXJyMjp37vzJLwT12S2O7S9IbJPFDnRAzd0BVVRUYGxsDC8vL4lXgPx4hb2PByJ++HOneewy1HiP98k/6cmTJ8yTJ08aFGPYsGGMi4uL2A5Qf/31F+Pi4sKMGDGiXjFnzpzJzJs3r0Hl+hiHw2GaNm3KrFy5knn79m2D4wUEBDBv3rypcbyoqIgJCAiQKlbbtm2Z3bt3f/L8rl27mDZt2khdRoZhmLlz5zIRERH1urYuQqGQEQqFDY6joqIiGlz1oZcvX0o8oKqkpITZsmULM3v2bIbD4TCTJk1iZs6cWWuqD0NDw3pP5/sUtnfzk8UOdGxLSEhgnJ2dmfj4eKawsJApLCxk4uPjGRcXF+bUqVONXTyFRi12BSYUCrF06VKsXbtWNOhNR0cHs2bNwoIFCz67fnVt9PT0cPr0abi6uoodv3z5Mr788ksUFBRIXcZp06Zh165dsLa2bvDz6/fi4uJw9uxZJCUlIT09HU5OTqIW+xdffCF1j4Wqqqrome6HXr58CVNTU9FjDkksWLAAv/76Ky5fvgwTExOxc7m5uXB3d8fXX38t8XPS4OBg0f8LhULExMTAwcEBDg4ONZbwrM9nuWvXLqxevRoZGRkAABsbG8yZM6ferdlPdZ0/fvwYdnZ2KC4urjPGmzdvRMuRdu/eHYcPH2b1Gfu8efPA5/OxaNEi1mKqqKiIFjz6UF5eHiwsLFBWViZVvO3btyM2NpbVHeiA6lb04cOHxRbRGTBgQL02hWnfvj2ioqJqrBV/7tw5TJo0SeopfkRyNHhOgS1YsEA0mtvT0xNA9RznsLAwlJaWSj3IRigU1qgsgOrn9h8PMJLU7du3RcvR3r9/X+xcfbvM/fz8RKvFFRYW4ty5c/jtt9/Qt29fqKioSDzH982bN2AYBgzD4O3bt+DxeKJzVVVVOHHiRI0/1HWZP38+jhw5Amtra3z99ddo06YNAODu3bvYs2cPzM3NJZov/N6NGzfEXjs6OgKo/lw/VJ/Pct26dVi0aBGCgoLEfn++/fZbvHz5Uqqd7d5/AeFwOFi0aJHYl6uqqipcunRJVPa6GBgYiL5oyeKxSmlpKX755RecPn26wV+QZLWbnyx2oPvrr7/Qr18/5OXliX4vV65cCWNjY/z3v/+Vuts/MzOz1i9c71fLI7JDLXYF1qxZM0RFRdUYpHLkyBFMmTJF6n2/BwwYgIKCAuzbtw/NmjUDADx9+hSjR4+GgYFBvQcVycKrV6+QnJyMpKQkJCUl4a+//oKBgQG6dOkicTlVVFQ+W3FwOByEh4djwYIFUpWtsLAQISEhOHDgAF6/fg0A0NfXx4gRI7Bs2TIYGBhIFU9WWrVqhfDwcIwdO1bseExMDMLCwqSazvR+AGJycjI8PDzEnolraGhAIBBg9uzZEo21+HDwnKqqKnJzc1kdPPe5wZIcDkeq5/ayWvRGFjvQeXh4wNjYGDExMaLfwdevX8Pf3x8vXrzAhQsXpIrXtWtX8Hg87N69W9Q7lZeXh7Fjx6K0tLTOTapI/VHFrsB4PB5u3rwJGxsbseP37t2Do6Oj1AO0njx5gv79++Ovv/4Szb198uQJ2rdvj6NHj4o24aiv9+twNzSOvb090tPTYWBggK5du8LLywvdunWTeqWr5ORkMAyDHj164NChQ2JdnhoaGmjZsqXoC059MAyDly9fgmEYGBsbN7j1WVhYiKqqqhpds/n5+VBTU5N6Ny0ej4fbt2/XGCyXkZEBe3v7eq1uFhAQgB9++KFBO3vJevCcLMhiNz+2aWpq4urVq6INk967ffs2XF1dpf578eDBAwwcOBD3798X+3thbW2NuLi4Bg3CJJ9HXfEKrEOHDti4cWONZTc3btyIDh06SB3P3Nwc169fx+nTp3H37l0A1VN2GjLFhu1xAADw7bffolu3bg0eMfx+d7xHjx7BwsKC9W7fqqoqpKWlITMzE6NGjYKOjg6ePXsGXV1d8Pl8qeONGDEC/fr1w5QpU8SO/+c//8HRo0dx4sQJqeJZWVnhP//5D77//nux4+8XJaoPNnYQ+/XXXxETE4PMzEwkJyejXbt29Z7pIWsXL17Eq1evxHo3du3ahdDQUBQXF8PPzw8//fST1FNPZbEDnY2NDfLy8mpU7M+fP69XJWxlZYWbN28iISGhxt+LxpyZohQaZ8we+SckJSUx2trajK2tLTN+/Hhm/PjxjK2tLcPn85mzZ89KHCcxMZGxtbWtdbnTgoICxs7OTqp4H5o/fz5jbGzMbN68mUlLS2PS0tKYTZs2McbGxo2+Kch7v//+O3Pu3DnR640bNzIdOnRgRo4cyeTn59crZlZWFtO2bVtGS0uLUVVVFS0JOn36dOabb76pV0wDAwOxGQvvpaenS7xc64cOHjzIqKqqMr6+vkxERAQTERHB+Pr6MmpqakxsbGy9ylhUVMQsXLiQ8fDwYFq3bt3gjUa8vLyY/Px85sWLF6wtA+vl5cV07979k0kavr6+zIoVK0Svb968yaipqTETJ05k1q5dy5iamjKhoaFSl/HDpVs/lJuby6irq0sc5/1o9cLCQub48eNMu3btmN9++000i+a3335j7O3tmePHj0tdRtJ4qMWuwLp164b79+9j06ZNom/MgwYNwpQpU6TqQt6wYQMCAwNr7T7V09PDN998g3Xr1tVr+8mYmBhs27ZNbByAg4MDmjdvjilTptRrFa2qqipER0d/cuUwabtn58yZg5UrVwIAbt26heDgYMyaNQtnzpxBcHBwvVqhM2bMgIuLC9LS0sRaXQMHDkRgYKDU8YDqbUVrG6FfUVFRr3nxgwcPxqVLl7B+/XrExcUBqG5xXb58ud6LEk2cOBHJyckYM2ZMg1fHKygogK2tLaytrUVjFQwMDDBixAgsXbq03iPlPx7EV1FRgdTUVNy+fVvqPerT0tKwdOlS0ev9+/fD3d0dW7duBVDdCxYaGoqwsDCJ4kkyGK9Vq1YSl09fX7/GojTDhg0THWP+/0ltv379JJp3XtemPB+qz9bERDL0jF1BsbmqW8uWLREfHy+2X/qH7t69iy+//BLZ2dlSx2Z7HAAABAUFITo6Gl999VWtlcf69eulisfn83H79m0IBAKEhYXh9u3bOHjwIK5fv44+ffogNzdX6jIaGRnhwoULaNOmjdjuXFlZWbCzs0NJSYnUMbt374727dvjp59+Ejs+depU3Lx5E+fOnZM65qccPHgQQ4YMkfo6fX19HD9+XDTKvr7y8/Ph4eEhGrz5/nfzzp072Lt3L8zNzXHhwgVWByKGhYWhqKgIa9askfgaHo+HjIwM0TPmL774Ar179xYNuMzKyoK9vX2ta/LXhu3BeNIMYHv/aOpzJP1SweFw6vW4gEioUfsLiEyxtSsZl8utdWeq9zIyMhgej1ev2G5ubsy0adNqHA8KCmLc3d3rFdPIyIjVrkMDAwPmr7/+YhiGYTw9PZmff/6ZYRiGefToUb12yWOY6v3s38f8cHeuc+fOMU2bNq1XzPPnzzM8Ho/p0qULExYWxoSFhTFdunRheDye1I9KKioq/q+9O49q+sz6AP5NUEEUEEEdZAmbW1lUpFYdN7AK7mPHWhGtIGOrAqI4DNjqa4vVQ8cq1qPWjktFrWAHta5VCypI0cqihqpQZEcFN4KIGoQ+7x+8+b0kYUlCQiC9n3M4B/KDhwtobn7Pci/LyspiOTk5Uo//+OOPzNXVlXXp0kWlGG1tbRtdLlBWSEgIc3Z2ZmVlZXLXHj58yFxcXFQuUNOU3NxcZmpqqtTX2NjYsKSkJMYYY2KxmHXt2pUlJCRw14VCodJjMqaZDnREd1Bi12Hqqupmb2/Pjh8/3uT1o0ePqtwiU137ABqysLCQS0itMX36dG6duXPnzqy0tJQxxtj58+dZv379VBpzzpw5bPHixYyx/29XW1VVxTw9PRVukdmYGzdusHnz5rG33nqLDRs2jPn7+yv94i4rK4sJBALG5/MZn89ns2bNYmVlZWzs2LGsZ8+eLDw8XOUqhgcPHmSzZ89m1dXVKn29hEAgYOfOnWvy+k8//cQEAkGrvoesAwcOMAsLC6W+ZsmSJWzkyJEsOTmZhYaGMjMzMyYWi7nrhw4dYu7u7gqPl5qayk6dOiX1WExMDLO1tWW9evViixcvZq9fv1YqRomkpKRm30jHQVPxOkxdVd2Cg4Nx+fJlpKWlSRVpAepbjQ4fPhweHh5Kra/l5+fDzs4OPB4PDx48wM6dO6WaYyi7D6ChzZs3Iz8/H9u3b1fL7tvi4mIsW7YMJSUlWL58OQICAgAAK1euRF1dnVI/t0RpaSm8vLzAGENubi7c3d2Rm5sLc3NzJCcnK134Rp2mTp0KsViMFStWIDY2FrGxsRgwYAACAgIQGBioUutbiaFDhyIvLw+MMdja2soVf1G0sIq+vj7y8vKaPBpZWloKR0dHlY7kyXaZY4zh4cOHSE9Px9q1a5U6I/7kyRO89957SElJQffu3RETE4NZs2Zx1ydMmIARI0YovJfE29sbHh4eCA8PB1C/58PNzQ1+fn4YNGgQNm3ahI8//ljhNfuGGjuBomxt99DQUKxfvx7dunWTqorYGFVbPZOWUWLXYeoqtFFeXg43Nzfo6ekhKChIqlrajh07UFdXh8zMTLkSqc2RLdP6wQcfYNu2bUqN0ZDsk/HFixfRs2dPODk5ySWPY8eOqfQ91K22thZxcXEQCoVcdy5fX1+VE2dTpW+fPn2K3r17K9x0o3fv3rhw4QKGDBmCyspKmJqaIiYmRi1NUWQbjchSNGlaWlriyJEjcuVKJa5cuYIPPvhApS55/v7+Uh9LmqF4enpi0qRJSo8H1NcY6N69O/T09KQef/bsGbp3765wf3sLCwucOnUK7u7uAOqrSyYlJXGtiP/73/9i3bp1uHPnjkoxNvTmzRvcuHEDa9euxYYNGzBhwoQWx2hY4ledhX6IciixE4UUFRVh6dKlOH/+PLdph8fjwcvLCzt27FBqJy4gXzvb2NgYN2/elOr5rgzZJ+PmKLuLvaVNgYp0JGsLTdUjf/DgARwcHBTeiCg7jpGRETIzMzXWWlcVixYtQl5eHn7++We5pCgWi+Hl5QV7e3uVeqe3Z+rejKeIpKQkhIaGIiMjQ21jEs2i425EIQKBAGfPnkVFRQXu3bsHxhj69euntl3HrX19qY7CJ02xtbVtdkpf0TthTfWqliwF8Hg87NmzR6q4TV1dHZKTkxVuuykZR1IbnzEGHo+HV69e4fnz51Kf15rqca0l6Zfer18/BAYGYuDAgWCM4e7du9i5cyfEYjEOHjzYqu+RkZHBLQ85OTmpfMRPnfr06YOCggJYW1ujpqYGmZmZUrMgVVVVjfZzaO33zMnJUeuYRLMosesY2Snp5qgyJW1qairX3U0VPB5PLlm212pUso1WJFOUW7ZsUeqcvaQxjURjR5YkvwNlelVLju8xxrBr1y6p6V5JHfZdu3YpPB5jTOr4IWNMKqlJkr0q/bRbqr+v6JhWVla4evUqli1bhtWrV0vNIk2cOBHbt2/n7mqV9ejRI8ydOxeXL1/mzsKLRCJ4eHggLi5OrXXplTVlyhRERETgyy+/xI8//ghDQ0Op+hFCoRAODg4qjS0UCqU+luwtiIqKUrhBT0MeHh7N/q1pKl5zKLHrmIYFKxhjOH78OExMTLg1uYyMDIhEIqVeAGgCYwx+fn5cKc3Xr19jyZIlchv8VHnxMXTo0EafUHg8HgwMDODo6Ag/P79m1wAbaqz8rru7O/r27YtNmzYp/LtsWCgnISEB4eHh2LhxI0aOHAmgvvzomjVrsHHjRoXGk5CUK/Xw8MCxY8daPYty6dKlVn19c2Qb8EheJMXExLS4/i7Lzs4OP/30EyoqKri2so6Ojq1uYxocHIyqqircvn1b6nz8woULsXz5csTGxrZq/NZYv3493nvvPYwbN47bjNdwKWLfvn0q7wMYMmRIoy82R4wYodKShjoL/RDl0Bq7DgsPD8ezZ8+k7uLq6uqwbNkyGBsbY9OmTVqLTdE1cVWm2FevXo1vvvkGLi4uGD58OAAgLS0NQqEQfn5+uHPnDhITE3Hs2DHMnDlT6fEl7t27h8GDByvUQ1wW9aqWdvjwYRw5cgQnTpzQdigwMTFBQkKC3MzU9evXMWnSJIhEIu0E1oC6NuM1VFRUJPWxZNOg7EmY1lKl0A9RDiV2HdarVy+kpKRwu9glcnJyMGrUKDx9+lRLkWnW4sWLYWNjg7Vr10o9/sUXX6CoqAi7d+/GunXrcObMGaSnp7c4nuzasmSK8rPPPkN2djZu3rypdIxdu3ZFWlqaXKMaoVCId955R+GNbpo8XiT7c0vweDzo6+urlDyakp+fD1dXV64RkDYZGRnhypUrcnecN27cwLhx45r8vXRUkkY1DavVqaNRTVPu3buH4cOH49mzZ2oZj8ijqXgdVltbi+zsbLnEnp2dLVc/XZf88MMPje7gnTt3LoYNG4bdu3fDx8dH4UQnW08bqE/u1tbWiIuLUynGt99+G6GhoXK9qsPCwrhZBkXcuHEDb9684d5viiq7pBv7uRuysrKCn58f1q1bp1IXPolXr15h27ZtsLS0VHkMdfL09ERISAhiY2O5Wgr379/HypUrFTry1dFERkZi/PjxXGLPyspCQECA1Nn4vn37qnQ2vjFXr15V+ywAkUaJXYf5+/sjICAAeXl5XLL49ddfERUVpdTxsI7GwMAAqampcq0mU1NTuSeUP/74Q+EnF9k1Z8kUpaOjIzp1Uu2/0L59+zBr1izY2Ng02qtaUZcuXUJ0dDRWrlzZ5Np4VVUVvL29lY5x//79+PTTT+Hn58f9+7l+/TpiYmKwZs0aPH78GF999RX09fXlWrs2xdTUVK7pSFVVFQwNDXHo0CGlY9SE7du3Y8aMGbC1tZX62zg7O7ebGNXp5s2bWL9+PfdxaxvVSLRU6IdoDiV2HfbVV1/hL3/5CzZv3oyHDx8CqC9wERYWhlWrVmk5Os0JDg7GkiVLkJGRwa2TpqWlYc+ePVwCOn/+vMI7fRVpfqEsdfaq/uSTT2BmZoYPP/xQ7lp1dTUmT56s0rJLTEwMNm/ejDlz5nCPTZ8+HS4uLvj222+RmJgIGxsbbNiwQeHEvnXrVqmPJS+S3nnnHbU2bGkNa2trZGZmIiEhQe5vo4sqKiqkCkMlJSVh8uTJ3Mdvv/02SkpKlB634UZeoP5vPWDAAERGRqq8wY8ohtbY/yQk64LaPHvclr7//nts376dO387YMAABAcHY968eQDqp38lu+Qbo6kz55oQHx+PBQsW4MiRI1KxvHjxAt7e3nj06BEuX76sdInerl27QigUyhWmyc3NxeDBg/Hy5UsUFBTAyclJpW50pH0QCAQ4ePAgxo4di5qaGvTo0QOnTp3ilh2ysrIwbtw4WhPvQOiOXcfV1tbi8uXLyMvL45LagwcPYGxsLFXIRNf4+vrC19e3yestlW1t6cy5sjW0gfpCMh999BEMDAxarC+vTK/q2bNnQyQSwcfHB2fOnMH48eO5O/Xy8nIkJSWpVHff2toae/fuRVRUlNTje/fu5aaonz59qvSdtkgkwt69e6WKvyxatEjuDq+tXbx4EUFBQbh27ZrcC+DKykqMGjUKu3btkjo3rgs0eTa+ofz8fLx69QqDBg1q1Z4MogANN5khWlRYWMgGDhzIDA0NmZ6eHtcadPny5ezjjz/WcnQdx88//8zc3NzYuXPnWGVlJausrGTnzp1j7u7u7MKFCwqPY2try548ecK939Sbqp3yvvzyS2ZsbMwuXbrExowZw+zt7VXuwsYYYydOnGBdunRhrq6uLCAggAUEBLDBgwczfX19rsPYzp072cqVKxUeMy0tjfXs2ZNZWlqyWbNmsVmzZjErKytmZmbGMjIyVI5VHaZPn862bNnS5PWvv/6a/e1vf2vDiNrG48eP2ZgxYxiPx2NGRkbs2LFjUtc9PT3ZJ598ovB4NTU17H/+53/YtGnT2BdffMFqa2vZ3LlzuW6BgwYNYgUFBWr+KUhDlNh12MyZM9n8+fOZWCyW6vl96dIl5ujoqOXo1MvU1JTrT92jRw9mamra5JuynJyc2JUrV+QeT05OZgMHDmx17OoUHh7O+Hw+s7e3Z8XFxa0eLz8/n4WHh3NJOCIiolVPyqNHj2Z+fn7szZs33GNv3rxhCxcuZGPGjGl1vK1hY2PTbK/4u3fvMmtr6zaMqG2JRCJWW1sr9/jTp0+lWs22JDQ0lPXq1Yv94x//YPb29mzGjBlswIABLC4ujv3www/MxcWFzZs3T52hExk0Fa/Drly5gtTUVLnzxra2trh//76WotKM6OhoGBkZce+rszxtXl4eV1q0IRMTExQWFqrle9TV1SErKwsCgUDpqW3Z3cedO3eGubk5QkJCpB5XpYqfnZ2d3FR8a6Snp2P37t1Spwk6deqEf/3rX1x1RG0pLy9vts56p06d8Pjx4zaMqG01tRSibCW/+Ph47N+/H1OmTMHvv/+OgQMH4syZM9yGvN69eze7TEZajxK7Dvvjjz8aXf8tLS3lkqCuWLhwIZ4/fw6xWKz2crnqOnPe0IoVK+Di4oKAgADU1dVh7NixuHr1KgwNDXH69GmMHz9e4bFkn5B9fHxUiqkxIpEI169fx6NHj+RqHzS2C78lxsbGKC4ulmtKU1JSovV/k5aWlvjtt9/kjklKCIVCWFhYtHFUHc+DBw+4Msz9+/eHvr6+1O+0f//+KCsr01Z4fw7anjIgmjNnzhy2ePFixhhj3bt3Z/n5+ayqqop5enoyPz8/LUenfjwej1vHa+5NWbm5uczZ2Zl16dKFOTg4MAcHB9alSxfm5OTEcnNzVYrV0tKSpaWlMcYYO378OOvbty/Lyclha9asYaNGjVJpTHU7efIkMzIyYjwej5mYmLAePXpwb6osaTDGWHBwMLOysmJxcXGsuLiYFRcXs9jYWGZlZcVCQkLU+wMoKSgoiDk7O7NXr17JXXv58iVzdnZmwcHBWoisY+HxeKy8vJz7uOEyIGOMlZWVqfT/kCiOjrvpsNLSUnh5eYExhtzcXLi7uyM3NxdmZma4cuWKXN/uji4pKYl7nzGGKVOmYM+ePXIVzVQ5l84YU8uZcwkDAwPcu3cPVlZW+Oijj2BoaIitW7eioKAAgwcPbhdlS/v3748pU6Zg48aNMDQ0VMuYNTU1CAsLw65du1BbWwvGGLp06YKlS5ciKipKbWVLVVFeXg43Nzfo6ekhKCiIq9iYnZ2NHTt2oK6uDpmZmVJnvok8Pp+PmJgYbibJx8cHW7du5X5vIpEI/v7+KnUHJIqhxK7jamtrERcXB6FQiBcvXsDNzQ2+vr4tHvfSBUZGRrh16xbs7e21HYocgUCA3bt3Y8KECbCzs8M333yDqVOn4vbt2xg9ejQqKiq0HSK6deuGrKwsjfz+Xr58iby8PACAg4OD2l44tFZRURGWLl2K8+fPS7WC9fLywo4dO2BnZ6flCNs/RY6yqdr2lyiG1th12NOnT2FmZob58+ejpKQEu3fvRk5ODtLT03XuLK6mJSYmIjExsdG1ZlVaWvr7+2POnDmwsLAAj8fjqpr9+uuvcuvP2uLl5YX09HS1JPZFixYp9Hmq/C7VSSAQ4OzZs6ioqMC9e/fAGEO/fv3aTVW8jkCX+1B0FJTYdVBWVhamT5/O1R6Pi4uDt7c3qqurwefzER0djfj4eLkiLKRxn3/+OSIjI+Hu7s4l4tb67LPP4OzsjJKSErz//vvcFLSenh4iIiJaPb46TJ06FWFhYbhz5w5cXFzkdowrU3Fv//79EAgEGDp0qFy/7/bI1NRUrm0rIR0FTcXroMmTJ6NTp06IiIjAwYMHcfr0aXh5eXFNHYKDg5GRkYFr165pOVLNMjIyglAobPX0qYWFBf79739jwYIFaoqsY2huSlXZqdTAwEDExsZCIBDA398f8+fPV/oYVVuprq5GVFRUkzM0+fn5Woqs/etIpZh1GSV2HWRubo6LFy9y/a2NjY2RlpaGYcOGAajfDDRixAiIRCLtBqpmssfcTp06BU9PT3Tr1k3qcWXPc5uZmeH69etqKavZkLqn99s7sViMY8eOYd++fUhNTcXUqVMREBCASZMmqbXuQGv5+PggKSkJCxYsaHSGRrY+APl/ipaKpTV2zaLEroP4fD7Kysq4Xe+ym8jKy8vRt29fnfuPpWgr2u+++06pccPDw9G9e3e1tppsaXr/+PHjavte7VFRURH279+PAwcOoLa2Frdv3243vQt69OiBM2fO4K9//au2QyFEJbTGrqNkE0V7uiPSFGUTtqJev36N//znP0hISICrq6vcWvOWLVuUHnPXrl3Yv39/u5ve11SjGll8Pp9rrNPeXmCampq222UCQhRBd+w6iM/nY/LkydyGLNkpabFYjHPnzrW7J9T2ysPDo9nrly5dUnpMTU3vt5adnR3S09NhZmbW7N4EHo+n9Fpzw6n4lJQUTJs2Df7+/vD29m5X3b4OHTqEEydOICYmpt0cw+soWnox2FBrXhiS5lFi10GampIm6qOJ6f32bNmyZYiLi4O1tTUWLVoEX19fmJubazusRg0dOhR5eXlgjMHW1lZuhiYzM1NLkbV/im5UVeWFIVEcJXZCmqBIzXkej4ejR48qPXZISAgOHDgAV1dXtU3vq1tKSgpGjx6tlrH4fD5sbGwwdOjQZpeFVGlUo26ff/55s9fXrVvXRpEQohpaYyekCU11u1IHoVCIIUOGAAB+++03jX2f1vD09ISlpSV8fHzg6+sLJycnlcf68MMPO8w+D0rcpKOjO3ZCSKOePHmCuLg4xMbG4urVq3B1dYWvry98fHxgZWWl7fA0SiQSIT4+Hnl5eQgLC0PPnj25OvGyvQdI00pLS3Hy5EkUFxejpqZG6lp7mJXSVZTYCWlDmpze16SCggIcPnwYsbGxyM7OxtixY3Hx4kVth6URQqEQ7777LkxMTFBYWIicnBzY29tjzZo1KC4uxoEDB7QdYoeQmJiIGTNmwN7eHtnZ2XB2dkZhYSEYY3Bzc9PZfz/tQfvZikrIn4CJiUmLb8bGxtoOU46dnR0iIiIQFRUFFxcXqU56uiY0NBR+fn7Izc2FgYEB9/iUKVOQnJysxcg6ltWrV+Of//wnsrKyYGBggKNHj6KkpATjxo3D+++/r+3wdBrdsRNCmvXLL7/g+++/R3x8PF6/fo2ZM2fC19cX3t7e2g5NI0xMTJCZmQkHBwep4k5FRUUYMGAAXr9+re0QOwQjIyPcvHkTDg4OMDU1RUpKCpycnHDr1i3MnDkThYWF2g5RZ9HmOUJIo1avXo24uDg8ePAAEydOxNdff42ZM2fq/NlufX19PH/+XO7x33//Hb169dJCRB1Tt27duHV1CwsL5OXlcRswnzx5os3QdB5NxRNCGpWcnIywsDDcv38fp0+fho+Pj84ndaC+OUlkZCTevHkDoH7PQ3FxMcLDw/H3v/9dy9G1f5GRkaiursaIESOQkpICoH4ZY9WqVdiwYQMWLVqEESNGaDlK3UZT8YQQ0kBlZSVmz56N9PR0VFVVoW/fvigrK8PIkSNx9uxZuaZCRJqenh4ePnyIFy9e4MWLF3B1dUV1dTVWrVqF1NRU9OvXD1u2bIFAINB2qDqLEjshhHPy5ElMnjwZnTt3brEFp6633UxJSYFQKMSLFy/g5uaGd999V9shdQiyTahI26PETgjhNHxSVmc/dvLnwefzUV5eTvsRtIg2zxFCOA37wsv2iJcoKSlBZGRkW4WkFYmJiYiOjsbdu3cBAIMGDcKKFSvorl1B/fv3b7HS4LNnz9oomj8fumMnhCjl1q1bcHNz09k79p07dyIkJASzZ8/GyJEjAQDXrl1DfHw8oqOjERgYqOUI2zc+n4+tW7e2WJJ54cKFbRTRnw8ldkKIUnQ9sVtZWSEiIgJBQUFSj+/YsQMbN27E/fv3tRRZx0Br7NpHx90IIaQBkUjUaPGdSZMmobKyUgsRdSwdpdmPLqPETgghDcyYMQPHjx+Xe/zEiROYNm2aFiLqWGgSWPto8xwhREpLjWpEIlHbBKIlb731FjZs2IDLly9LrbH/8ssvWLVqFbZt28Z97vLly7UVZrvV1KZL0nZojZ0QIsXf31+hz/vuu+80HIl22NnZKfR5PB4P+fn5Go6GEOVRYieEkEZI6pmbm5trORJClENr7IQQ8n9EIhECAwNhbm6OPn36oE+fPjA3N0dQUJDOL0EQ3UF37IQQgvqCKSNHjsT9+/fh6+uLQYMGAQDu3LmDw4cPw9raGqmpqTA1NdVypIQ0jxI7IYQAWLFiBRITE5GQkIA+ffpIXSsrK8OkSZMwYcIEREdHaylCQhRDiZ0QQgDY2tri22+/hZeXV6PXz507hyVLlqCwsLBtAyNESbTGTgghAB4+fAgnJ6cmrzs7O6OsrKwNIyJENZTYCSEE9bvfm7sbLygoQM+ePdsuIEJURImdEEIAeHl54dNPP0VNTY3cNbFYjLVr1zZaapaQ9obW2AkhBEBpaSnc3d2hr6+PwMBADBw4EIwx3L17Fzt37oRYLEZ6ejqsra21HSohzaLETggh/6egoADLli3DhQsXuJrnPB4PEydOxPbt2+Ho6KjlCAlpGSV2QgiRUVFRgdzcXACAo6Mjra2TDoUSOyGEEKJDaPMcIYQQokMosRNCCCE6hBI7IYQQokMosRNCCCE6hBI7IYQQokMosRNCCCE6hBI7IYQQokMosRNCCCE65H8BofQSV+i00HAAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 500x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the trained network on the test set\n",
    "\n",
    "train_test.perform_test(model = model_data_augmentation, model_path = task2_model_path, test_loader = test_loader, dataset_test = dataset_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I will now investigate what happens using also:\n",
    "\n",
    "* Adam optimizer\n",
    "* dropout "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH 1:\n",
      "LOSS: train: 2.28084769397974, validation: 2.3108415603637695; accuracy validation set: 19.11111111111111%\n",
      "\n",
      "EPOCH 2:\n",
      "LOSS: train: 1.7741989016532898, validation: 1.8608949184417725; accuracy validation set: 35.55555555555556%\n",
      "\n",
      "EPOCH 3:\n",
      "LOSS: train: 1.49086654484272, validation: 1.539387822151184; accuracy validation set: 44.44444444444444%\n",
      "\n",
      "EPOCH 4:\n",
      "LOSS: train: 1.2809935905039311, validation: 2.058729410171509; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 5:\n",
      "LOSS: train: 1.1190727047622204, validation: 1.7346161603927612; accuracy validation set: 39.55555555555556%\n",
      "\n",
      "EPOCH 6:\n",
      "LOSS: train: 0.9818670429289341, validation: 2.5807106494903564; accuracy validation set: 29.333333333333332%\n",
      "\n",
      "EPOCH 7:\n",
      "LOSS: train: 0.8968664184212685, validation: 1.393500566482544; accuracy validation set: 49.333333333333336%\n",
      "\n",
      "EPOCH 8:\n",
      "LOSS: train: 0.8149099528789521, validation: 2.2077295780181885; accuracy validation set: 44.44444444444444%\n",
      "\n",
      "EPOCH 9:\n",
      "LOSS: train: 0.7659089729189873, validation: 1.488778829574585; accuracy validation set: 49.333333333333336%\n",
      "\n",
      "EPOCH 10:\n",
      "LOSS: train: 0.6875857532024383, validation: 1.202567458152771; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 11:\n",
      "LOSS: train: 0.6104370255023241, validation: 1.2940789461135864; accuracy validation set: 57.77777777777778%\n",
      "\n",
      "EPOCH 12:\n",
      "LOSS: train: 0.5592426490038633, validation: 2.844183921813965; accuracy validation set: 30.666666666666668%\n",
      "\n",
      "EPOCH 13:\n",
      "LOSS: train: 0.5068889329209924, validation: 1.346366047859192; accuracy validation set: 52.888888888888886%\n",
      "\n",
      "EPOCH 14:\n",
      "LOSS: train: 0.4739034574478865, validation: 1.6310070753097534; accuracy validation set: 43.55555555555556%\n",
      "\n",
      "EPOCH 15:\n",
      "LOSS: train: 0.4430043075233698, validation: 1.310176968574524; accuracy validation set: 55.55555555555556%\n",
      "\n",
      "EPOCH 16:\n",
      "LOSS: train: 0.4039236966520548, validation: 1.5839035511016846; accuracy validation set: 52.888888888888886%\n",
      "\n",
      "EPOCH 17:\n",
      "LOSS: train: 0.34708257112652063, validation: 1.0790120363235474; accuracy validation set: 63.55555555555556%\n",
      "\n",
      "EPOCH 18:\n",
      "LOSS: train: 0.3178729586303234, validation: 1.2688888311386108; accuracy validation set: 53.333333333333336%\n",
      "\n",
      "EPOCH 19:\n",
      "LOSS: train: 0.31373314522206786, validation: 1.696549654006958; accuracy validation set: 47.111111111111114%\n",
      "\n",
      "EPOCH 20:\n",
      "LOSS: train: 0.27718300437554716, validation: 1.5954402685165405; accuracy validation set: 53.333333333333336%\n",
      "\n",
      "EPOCH 21:\n",
      "LOSS: train: 0.2433685871772468, validation: 1.3470773696899414; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 22:\n",
      "LOSS: train: 0.24330491647124292, validation: 1.3632090091705322; accuracy validation set: 58.22222222222222%\n",
      "\n",
      "EPOCH 23:\n",
      "LOSS: train: 0.22592963092029095, validation: 1.3985025882720947; accuracy validation set: 54.666666666666664%\n",
      "\n",
      "EPOCH 24:\n",
      "LOSS: train: 0.22727432427927852, validation: 1.278116226196289; accuracy validation set: 60.0%\n",
      "\n",
      "EPOCH 25:\n",
      "LOSS: train: 0.18450991744175554, validation: 1.2677972316741943; accuracy validation set: 60.888888888888886%\n",
      "\n",
      "EPOCH 26:\n",
      "LOSS: train: 0.19256655061617495, validation: 1.2711374759674072; accuracy validation set: 61.333333333333336%\n",
      "\n",
      "EPOCH 27:\n",
      "LOSS: train: 0.17659549936652183, validation: 1.3288421630859375; accuracy validation set: 58.22222222222222%\n",
      "\n",
      "EPOCH 28:\n",
      "LOSS: train: 0.16868602712638675, validation: 1.8492947816848755; accuracy validation set: 50.666666666666664%\n",
      "\n",
      "EPOCH 29:\n",
      "LOSS: train: 0.16520714410580695, validation: 1.4492486715316772; accuracy validation set: 54.22222222222222%\n",
      "\n",
      "EPOCH 30:\n",
      "LOSS: train: 0.1364018780644983, validation: 1.5867174863815308; accuracy validation set: 51.111111111111114%\n",
      "\n",
      "EPOCH 31:\n",
      "LOSS: train: 0.15219310349784793, validation: 1.613331913948059; accuracy validation set: 53.77777777777778%\n",
      "\n",
      "EPOCH 32:\n",
      "LOSS: train: 0.14807872432284058, validation: 2.137676954269409; accuracy validation set: 48.888888888888886%\n",
      "\n",
      "EPOCH 33:\n",
      "LOSS: train: 0.1458157072775066, validation: 1.5925248861312866; accuracy validation set: 54.666666666666664%\n",
      "\n",
      "EPOCH 34:\n",
      "LOSS: train: 0.14076434187591075, validation: 1.5024787187576294; accuracy validation set: 61.333333333333336%\n",
      "\n",
      "EPOCH 35:\n",
      "LOSS: train: 0.11626591673120856, validation: 1.4012484550476074; accuracy validation set: 59.55555555555556%\n",
      "\n",
      "EPOCH 36:\n",
      "LOSS: train: 0.13554597729817033, validation: 1.5273078680038452; accuracy validation set: 58.22222222222222%\n",
      "\n",
      "EPOCH 37:\n",
      "LOSS: train: 0.14423925697337836, validation: 1.452329158782959; accuracy validation set: 57.333333333333336%\n",
      "\n",
      "EPOCH 38:\n",
      "LOSS: train: 0.10769187523983419, validation: 1.4971725940704346; accuracy validation set: 56.0%\n",
      "\n",
      "EPOCH 39:\n",
      "LOSS: train: 0.09843675484880805, validation: 1.4914593696594238; accuracy validation set: 56.888888888888886%\n",
      "\n",
      "EPOCH 40:\n",
      "LOSS: train: 0.0932410730747506, validation: 1.4751276969909668; accuracy validation set: 56.0%\n",
      "\n",
      "EPOCH 41:\n",
      "LOSS: train: 0.09470800545532257, validation: 2.1554970741271973; accuracy validation set: 56.44444444444444%\n",
      "\n",
      "EPOCH 42:\n",
      "LOSS: train: 0.08269263219553977, validation: 1.7414891719818115; accuracy validation set: 58.666666666666664%\n",
      "\n",
      "EPOCH 43:\n",
      "LOSS: train: 0.08073800107231363, validation: 1.8453227281570435; accuracy validation set: 56.44444444444444%\n",
      "\n",
      "EPOCH 44:\n",
      "LOSS: train: 0.08897935606073588, validation: 2.186727285385132; accuracy validation set: 47.55555555555556%\n",
      "\n",
      "EPOCH 45:\n",
      "LOSS: train: 0.10516968239098787, validation: 1.6038130521774292; accuracy validation set: 57.333333333333336%\n",
      "\n",
      "EPOCH 46:\n",
      "LOSS: train: 0.09082092663738876, validation: 2.1784250736236572; accuracy validation set: 47.55555555555556%\n",
      "\n",
      "EPOCH 47:\n",
      "LOSS: train: 0.08301683524623513, validation: 1.5306215286254883; accuracy validation set: 59.55555555555556%\n",
      "\n",
      "EPOCH 48:\n",
      "LOSS: train: 0.08792851890902967, validation: 1.7590131759643555; accuracy validation set: 54.22222222222222%\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Instantiate a model\n",
    "\n",
    "model_data_augmentation_1 = CNN.CNN3()\n",
    "\n",
    "\n",
    "# Set parameters for the training\n",
    "\n",
    "learning_rate = 0.0005\n",
    "loss_function = nn.CrossEntropyLoss()\n",
    "optimizer = optim.Adam(model_data_augmentation_1.parameters(), lr=learning_rate)\n",
    "\n",
    "# Perform the training\n",
    "task2_2_model_path, task2_2_loss_train, task2_2_loss_val, task2_2_accuracies_train, task2_2_accuracies_val = train_test.train_model(model = model_data_augmentation_1, EPOCHS = 48, train_loader = data_augmentation_train_loader, val_loader = val_loader, optimizer = optimizer, loss_function = loss_function)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the test images: 56 %\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the trained network on the test set\n",
    "\n",
    "train_test.perform_test(model = model_data_augmentation_1, model_path = task2_2_model_path, test_loader = test_loader, dataset_test = dataset_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 3 <a class=\"anchor\" id=\"task3\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this section the transfer learning based on a pre-trained network (here, AlexNet in particular) is used."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Alexnet (1) <a class=\"anchor\" id=\"alex1\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The weights of all the layers but the last fully connected layer will be frozen and the weights of the last layer will be fine-tuned."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the dataset; remark that AlexNet requires 3 channels as input (so, RGB)\n",
    "\n",
    "train_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/train'\n",
    "test_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/test'\n",
    "\n",
    "train_ALEXNET = ImageFolder(root=train_path, transform=train_test.transform_ALEXNET)\n",
    "dataset_test_ALEXNET = ImageFolder(root=test_path, transform=train_test.transform_ALEXNET)\n",
    "\n",
    "\n",
    "# Split the provided training set in 85% for training set and 15% for validation set\n",
    "\n",
    "train_size_ALEXNET = int(0.85 * len(train_ALEXNET))\n",
    "val_size_ALEXNET = len(train_ALEXNET) - train_size_ALEXNET\n",
    "\n",
    "dataset_train_ALEXNET, dataset_val_ALEXNET = random_split(train_ALEXNET, [train_size_ALEXNET, val_size_ALEXNET])\n",
    "\n",
    "# Data augmentation\n",
    "\n",
    "flipped_dataset_train_ALEXNET = train_test.HorizontallyFlippedDataset(dataset_train_ALEXNET)\n",
    "data_augmentation_dataset_train_ALEXNET = torch.utils.data.ConcatDataset([dataset_train_ALEXNET, flipped_dataset_train_ALEXNET])\n",
    "\n",
    "\n",
    "# Create the loaders\n",
    "\n",
    "batch_size = 32\n",
    "data_augmentation_train_loader_ALEXNET = DataLoader(data_augmentation_dataset_train_ALEXNET, batch_size, shuffle=True) \n",
    "val_loader_ALEXNET = DataLoader(dataset_val_ALEXNET, batch_size, shuffle=False)\n",
    "test_loader_ALEXNET = DataLoader(dataset_test_ALEXNET, batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "EPOCH 1:\n",
      "LOSS: train: 1.103673218563199, validation: 0.6218058466911316; accuracy validation set: 74.22222222222223%\n",
      "\n",
      "EPOCH 2:\n",
      "LOSS: train: 0.6198100719600916, validation: 0.6079599857330322; accuracy validation set: 75.55555555555556%\n",
      "\n",
      "EPOCH 3:\n",
      "LOSS: train: 0.5317423537373542, validation: 0.564347505569458; accuracy validation set: 75.11111111111111%\n",
      "\n",
      "EPOCH 4:\n",
      "LOSS: train: 0.45119409933686255, validation: 0.5810096859931946; accuracy validation set: 75.11111111111111%\n",
      "\n",
      "EPOCH 5:\n",
      "LOSS: train: 0.4314974702894688, validation: 0.5394126772880554; accuracy validation set: 78.22222222222223%\n",
      "\n",
      "EPOCH 6:\n",
      "LOSS: train: 0.4029895424842834, validation: 0.5319639444351196; accuracy validation set: 76.44444444444444%\n",
      "\n",
      "EPOCH 7:\n",
      "LOSS: train: 0.3647210907191038, validation: 0.5312435626983643; accuracy validation set: 77.77777777777777%\n",
      "\n"
     ]
    }
   ],
   "source": [
    "alexnet = models.alexnet(weights=AlexNet_Weights.DEFAULT) # weights=AlexNet_Weights.DEFAULT instead of pretrained=True as suggested by a warning\n",
    "\n",
    "for param in alexnet.parameters(): # freeze all layers except the last fully connected layer\n",
    "    param.requires_grad = False\n",
    "\n",
    "\n",
    "in_features = alexnet.classifier[6].in_features # modify the last fully connected layer\n",
    "alexnet.classifier[6] = nn.Linear(in_features, 15)\n",
    "\n",
    "for param in alexnet.classifier[6].parameters(): # during training, only the weights and biases of the last layer will have their gradients updated\n",
    "    param.requires_grad = True\n",
    "\n",
    "loss_function = nn.CrossEntropyLoss()\n",
    "optimizer = optim.SGD(alexnet.parameters(), lr=0.001, momentum=0.9)\n",
    "\n",
    "\n",
    "task3_model_path, task3_loss_train, task3_loss_val, task3_accuracies_train, task3_accuracies_val = train_test.train_model(model = alexnet, EPOCHS = 7, train_loader = data_augmentation_train_loader_ALEXNET, val_loader = val_loader_ALEXNET, optimizer = optimizer, loss_function = loss_function)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the test images: 82 %\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_test.perform_test(model = alexnet, model_path = task3_model_path, test_loader = test_loader_ALEXNET, dataset_test = dataset_test_ALEXNET)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Alexnet (2) <a class=\"anchor\" id=\"alex2\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The pre-trained network will be employed as a feature extractor, accessing the activation of an intermediate layer (for instance, one of the fully connected layers); a multiclass linear SVM will be trained."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load the dataset; remark that AlexNet requires 3 channels as input (so, RGB)\n",
    "\n",
    "train_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/train'\n",
    "test_path = '/Users/Gaia/Desktop/CVPR-project/CVPR-project/data/test'\n",
    "\n",
    "dataset_train_ALEXNET_2 = ImageFolder(root=train_path, transform=train_test.transform_ALEXNET)\n",
    "dataset_test_ALEXNET_2 = ImageFolder(root=test_path, transform=train_test.transform_ALEXNET)\n",
    "\n",
    "\n",
    "# Data augmentation\n",
    "\n",
    "flipped_dataset_train_ALEXNET_2 = train_test.HorizontallyFlippedDataset(dataset_train_ALEXNET_2)\n",
    "data_augmentation_dataset_train_ALEXNET_2 = torch.utils.data.ConcatDataset([dataset_train_ALEXNET_2, flipped_dataset_train_ALEXNET_2])\n",
    "\n",
    "\n",
    "# Create the loaders\n",
    "\n",
    "batch_size = 32\n",
    "data_augmentation_train_loader_ALEXNET_2 = DataLoader(data_augmentation_dataset_train_ALEXNET_2, batch_size, shuffle=True) \n",
    "test_loader_ALEXNET_2 = DataLoader(dataset_test_ALEXNET_2, batch_size, shuffle=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Gaia\\AppData\\Roaming\\Python\\Python311\\site-packages\\sklearn\\svm\\_classes.py:31: FutureWarning: The default value of `dual` will change from `True` to `'auto'` in 1.5. Set the value of `dual` explicitly to suppress the warning.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test Accuracy: 79.53%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Gaia\\AppData\\Roaming\\Python\\Python311\\site-packages\\sklearn\\svm\\_base.py:1237: ConvergenceWarning: Liblinear failed to converge, increase the number of iterations.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "alexnet = models.alexnet(weights=AlexNet_Weights.DEFAULT)\n",
    "\n",
    "for param in alexnet.parameters(): # freezes the parameters of the pre-trained model; these parameters will not be updated during the training process\n",
    "    param.requires_grad = False\n",
    "\n",
    "train_features, train_labels = train_test.extract(alexnet, data_augmentation_train_loader_ALEXNET_2) # uses AlexNet model to extract features\n",
    "test_features, test_labels = train_test.extract(alexnet, test_loader_ALEXNET_2) # uses AlexNet model to extract features\n",
    "\n",
    "\n",
    "# Train a multiclass linear SVM\n",
    "\n",
    "SVM = svm.LinearSVC(C=1.0, max_iter=1000) # uses one-vs-the-rest; C has the default value\n",
    "SVM.fit(train_features, train_labels)\n",
    "\n",
    "\n",
    "# Evaluate the SVM on the test set\n",
    "\n",
    "y_pred = SVM.predict(test_features)\n",
    "accuracy = accuracy_score(test_labels, y_pred)\n",
    "print(f'Test Accuracy: {accuracy * 100:.2f}%')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References <a class=\"anchor\" id=\"ref\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "[1] Lazebnik, S., Schmid, C., and Ponce, J., Beyond bags of features: Spatial pyramid matching for recognizing natural scene categories, in 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06), volume 2, pages 2169-2178.\n",
    "\n",
    "[2] https://dingyan89.medium.com/calculating-parameters-of-convolutional-and-fully-connected-layers-with-keras-186590df36c6\n",
    "\n",
    "[3] Ioffe, S. and Szegedy, C., Batch normalization: Accelerating deep network training by reducing internal covariate shift, arXiv preprint arXiv:1502.03167, 2015\n",
    "\n",
    "[4] Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A., Going deeper with convolutions, in The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015\n",
    "\n",
    "[5] Krizhevsky, A., Sutskever, I., and Hinton, G. E., Imagenet classification with deep convolutional neural networks, in Advances in neural information processing systems, pages 1097–1105, 2012"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.11.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}