{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Hadamard Multitask GP Regression\n", "\n", "## Introduction\n", "\n", "This notebook demonstrates how to perform \"Hadamard\" multitask regression. \n", "This differs from the [multitask gp regression example notebook](./Multitask_GP_Regression.ipynb) in one key way:\n", "\n", "- Here, we assume that we have observations for **one task per input**. For each input, we specify the task of the input that we observe. (The kernel that we learn is expressed as a Hadamard product of an input kernel and a task kernel)\n", "- In the other notebook, we assume that we observe all tasks per input. (The kernel in that notebook is the Kronecker product of an input kernel and a task kernel).\n", "\n", "Multitask regression, first introduced in [this paper](https://papers.nips.cc/paper/3189-multi-task-gaussian-process-prediction.pdf) learns similarities in the outputs simultaneously. It's useful when you are performing regression on multiple functions that share the same inputs, especially if they have similarities (such as being sinusodial).\n", "\n", "Given inputs $x$ and $x'$, and tasks $i$ and $j$, the covariance between two datapoints and two tasks is given by\n", "\n", "$$ k([x, i], [x', j]) = k_\\text{inputs}(x, x') * k_\\text{tasks}(i, j)\n", "$$\n", "\n", "where $k_\\text{inputs}$ is a standard kernel (e.g. RBF) that operates on the inputs.\n", "$k_\\text{task}$ is a special kernel - the `IndexKernel` - which is a lookup table containing inter-task covariance." ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "import math\n", "import torch\n", "import gpytorch\n", "from matplotlib import pyplot as plt\n", "\n", "%matplotlib inline\n", "%load_ext autoreload\n", "%autoreload 2" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Set up training data\n", "\n", "In the next cell, we set up the training data for this example. For each task we'll be using 50 random points on [0,1), which we evaluate the function on and add Gaussian noise to get the training labels. Note that different inputs are used for each task.\n", "\n", "We'll have two functions - a sine function (y1) and a cosine function (y2)." ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "TASK_NOISES = [math.sqrt(0.3), math.sqrt(0.1)]\n", "torch.manual_seed(1)\n", "train_x1 = torch.rand(20)\n", "train_x2 = torch.rand(20)\n", "\n", "train_i_task1 = torch.full((train_x1.shape[0],1), dtype=torch.long, fill_value=0)\n", "train_i_task2 = torch.full((train_x2.shape[0],1), dtype=torch.long, fill_value=1)\n", "\n", "train_f1 = torch.sin(train_x1 * (2 * math.pi))\n", "train_f2 = torch.cos(train_x2 * (2 * math.pi))\n", "\n", "train_noise1 = torch.randn(train_f1.size())\n", "train_noise2 = torch.randn(train_f2.size())\n", "\n", "full_train_x = torch.cat([train_x1, train_x2])\n", "full_train_i = torch.cat([train_i_task1, train_i_task2])\n", "full_train_f = torch.cat([train_f1, train_f2])\n", "full_train_noise = torch.cat([TASK_NOISES[0] * train_noise1, TASK_NOISES[1] * train_noise2])\n", "full_train_y = full_train_f + full_train_noise\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Set up a Hadamard multitask model\n", "\n", "The model should be somewhat similar to the `ExactGP` model in the [simple regression example](../01_Exact_GPs/Simple_GP_Regression.ipynb).\n", "\n", "The differences:\n", "\n", "1. The model takes two input: the inputs (x) and indices. The indices indicate which task the observation is for.\n", "2. Rather than just using a RBFKernel, we're using that in conjunction with a IndexKernel.\n", "3. We don't use a ScaleKernel, since the IndexKernel will do some scaling for us. (This way we're not overparameterizing the kernel.)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "class MultitaskGPModel(gpytorch.models.ExactGP):\n", " def __init__(self, train_x, train_y, likelihood):\n", " super(MultitaskGPModel, self).__init__(train_x, train_y, likelihood)\n", " self.mean_module = gpytorch.means.ConstantMean()\n", " self.covar_module = gpytorch.kernels.RBFKernel()\n", " \n", " # We learn an IndexKernel for 2 tasks\n", " # (so we'll actually learn 2x2=4 tasks with correlations)\n", " self.task_covar_module = gpytorch.kernels.IndexKernel(num_tasks=2, rank=1)\n", "\n", " def forward(self,x,i):\n", " mean_x = self.mean_module(x)\n", " \n", " # Get input-input covariance\n", " covar_x = self.covar_module(x)\n", " # Get task-task covariance\n", " covar_i = self.task_covar_module(i)\n", " # Multiply the two together to get the covariance we want\n", " covar = covar_x.mul(covar_i)\n", " \n", " return gpytorch.distributions.MultivariateNormal(mean_x, covar)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Training the model\n", "\n", "In the next cell, we handle using Type-II MLE to train the hyperparameters of the Gaussian process.\n", "\n", "See the [simple regression example](../01_Exact_GPs/Simple_GP_Regression.ipynb) for more info on this step." ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "scrolled": false }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "c:\\Users\\tobyb\\phd\\gpytorch\\venv\\Lib\\site-packages\\linear_operator\\utils\\interpolation.py:71: UserWarning: torch.sparse.SparseTensor(indices, values, shape, *, device=) is deprecated. Please use torch.sparse_coo_tensor(indices, values, shape, dtype=, device=). (Triggered internally at C:\\actions-runner\\_work\\pytorch\\pytorch\\pytorch\\torch\\csrc\\utils\\tensor_new.cpp:646.)\n", " summing_matrix = cls(summing_matrix_indices, summing_matrix_values, size)\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Iter 25/100 - Loss: 1.012\n", "Iter 50/100 - Loss: 1.004\n", "Iter 75/100 - Loss: 1.003\n", "Iter 100/100 - Loss: 1.003\n" ] } ], "source": [ "# this is for running the notebook in our testing framework\n", "import os\n", "smoke_test = ('CI' in os.environ)\n", "training_iterations = 2 if smoke_test else 100\n", "\n", "# We define the training loop in a function, which will let us use\n", "# it again later for a different likelihood.\n", "def train_model(train_data, likelihood_cls: type[gpytorch.likelihoods.Likelihood]):\n", " likelihood = likelihood_cls(num_tasks=2)\n", " (train_x, train_i), train_y = train_data\n", " # Here we have two terms that we're passing in as train_inputs\n", " model = MultitaskGPModel((train_x, train_i), train_y, likelihood)\n", " # Find optimal model hyperparameters\n", " model.train()\n", " likelihood.train()\n", "\n", " # Use the adam optimizer\n", " optimizer = torch.optim.Adam(model.parameters(), lr=0.1) # Includes GaussianLikelihood parameters\n", "\n", " # \"Loss\" for GPs - the marginal log likelihood\n", " mll = gpytorch.mlls.ExactMarginalLogLikelihood(likelihood, model)\n", "\n", " for i in range(training_iterations):\n", " optimizer.zero_grad()\n", " output = model(train_x, train_i)\n", " loss = -mll(output, train_y, [train_i])\n", " loss.backward()\n", " if (i + 1) % 25 == 0:\n", " print(f'Iter {i+1}/{training_iterations} - Loss: {loss.item():.3f}')\n", " optimizer.step()\n", "\n", " \n", " # Set into eval mode\n", " model.eval()\n", " likelihood.eval()\n", "\n", " return model, likelihood\n", "\n", "model, likelihood = train_model(\n", " ((full_train_x, full_train_i), full_train_y), \n", " gpytorch.likelihoods.GaussianLikelihood\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Make predictions with the model" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Initialize plots\n", "f, (y1_ax, y2_ax) = plt.subplots(1, 2, figsize=(8, 3))\n", "\n", "# Test points every 0.02 in [0,1]\n", "test_x = torch.linspace(0, 1, 51)\n", "test_i_task1 = torch.full((test_x.shape[0],1), dtype=torch.long, fill_value=0)\n", "test_i_task2 = torch.full((test_x.shape[0],1), dtype=torch.long, fill_value=1)\n", "\n", "# Make predictions - one task at a time\n", "# We control the task we cae about using the indices\n", "\n", "# The gpytorch.settings.fast_pred_var flag activates LOVE (for fast variances)\n", "# See https://arxiv.org/abs/1803.06058\n", "with torch.no_grad(), gpytorch.settings.fast_pred_var():\n", " observed_pred_y1 = likelihood(model(test_x, test_i_task1), [test_i_task1])\n", " observed_pred_y2 = likelihood(model(test_x, test_i_task2), [test_i_task2])\n", "\n", "\n", "# Define plotting function\n", "def ax_plot(ax, train_y, train_x, rand_var, title):\n", " # Get lower and upper confidence bounds\n", " lower, upper = rand_var.confidence_region()\n", " # Plot training data as black stars\n", " ax.plot(train_x.detach().numpy(), train_y.detach().numpy(), 'k*')\n", " # Predictive mean as blue line\n", " ax.plot(test_x.detach().numpy(), rand_var.mean.detach().numpy(), 'b')\n", " # Shade in confidence \n", " ax.fill_between(test_x.detach().numpy(), lower.detach().numpy(), upper.detach().numpy(), alpha=0.5)\n", " ax.set_ylim([-3, 3])\n", " ax.legend(['Observed Data', 'Mean', 'Confidence'])\n", " ax.set_title(title)\n", "\n", "# Plot both tasks\n", "train_y1 = train_f1 + TASK_NOISES[0] * train_noise1\n", "train_y2 = train_f2 + TASK_NOISES[1] * train_noise2\n", "ax_plot(y1_ax, train_y1, train_x1, observed_pred_y1, fr'Task 1 ($\\sigma_y^2 = {TASK_NOISES[0]}$)')\n", "ax_plot(y2_ax, train_y2, train_x2, observed_pred_y2, fr'Task 2 ($\\sigma_y^2 = {TASK_NOISES[1]}$)')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Task-specific Noise\n", "\n", "In this notebook so far, we assumed that each task had the same noise. However, \n", "this may be too strong an assumption. In this section, we use the `HadamardGaussianLikelihood`\n", "to learn independent noises for each task." ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iter 25/100 - Loss: 0.942\n", "Iter 50/100 - Loss: 0.935\n", "Iter 75/100 - Loss: 0.933\n", "Iter 100/100 - Loss: 0.933\n" ] } ], "source": [ "model_hd, likelihood_hd = train_model(\n", " ((full_train_x, full_train_i), full_train_y), \n", " gpytorch.likelihoods.HadamardGaussianLikelihood\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "First, we compare the models when the tasks have significantly different noises. \n", "Note that the training loss achieved using the `HadamardGaussianLikelihood` is lower\n", "than using `GaussianLikelihood`, which learns the same noise across all tasks. \n", "We also see that the predictive distribution for task 2 is much tighter." ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "with torch.no_grad(), gpytorch.settings.fast_pred_var():\n", " observed_pred_y1_hd = likelihood_hd(model_hd(test_x, test_i_task1), [test_i_task1])\n", " observed_pred_y2_hd = likelihood_hd(model_hd(test_x, test_i_task2), [test_i_task2])\n", "\n", "train_y1 = train_f1 + TASK_NOISES[0] * train_noise1\n", "train_y2 = train_f2 + TASK_NOISES[1] * train_noise2\n", "\n", "fig = plt.figure(figsize=(8, 7))\n", "subfigs = fig.subfigures(2, 1)\n", "subfigs[0].suptitle('Shared noise')\n", "subfigs[1].suptitle('Task-specific noise')\n", "\n", "for row, (subfig, pred_y1, pred_y2) in enumerate(zip(subfigs, (observed_pred_y1, observed_pred_y1_hd), (observed_pred_y2, observed_pred_y2_hd))):\n", " y1_ax, y2_ax = subfig.subplots(1, 2)\n", " ax_plot(y1_ax, train_y1, train_x1, pred_y1, fr'Task 1 ($\\sigma_y^2 = {TASK_NOISES[0]**2:.2f}$)')\n", " ax_plot(y2_ax, train_y2, train_x2, pred_y2, fr'Task 2 ($\\sigma_y^2 = {TASK_NOISES[1]**2:.2f}$)')\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Failure case of task-specific noise\n", "\n", "The downside to this approach is that, since each task has its own noise, learning\n", "the noise parameter requires more data. We demonstrate this failure case below,\n", "where each task has the same noise. In the low-data regime, learning a single\n", "noise parameter gives accurate results, however the task-specific noises are not accurate,\n", "with task 1 overestimating the noise, and task 2 underestimating.\n", "\n", "This can be mitigated by setting the `noise_prior` argument of the likelihood." ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Iter 25/100 - Loss: 1.190\n", "Iter 50/100 - Loss: 1.189\n", "Iter 75/100 - Loss: 1.188\n", "Iter 100/100 - Loss: 1.188\n", "likelihood.noise=tensor([0.3621], grad_fn=)\n", "Iter 25/100 - Loss: 1.150\n", "Iter 50/100 - Loss: 1.137\n", "Iter 75/100 - Loss: 1.136\n", "Iter 100/100 - Loss: 1.136\n", "likelihood.noise=tensor([0.5248, 0.1473], grad_fn=)\n" ] }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAp4AAAK6CAYAAACQS1yzAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjEsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvc2/+5QAAAAlwSFlzAAAPYQAAD2EBqD+naQABAABJREFUeJzsnQWcFOX/xz/X3c0F3V0ioQiCoBioKCoqWKBioSigf1sssEVARfiBDYKoKCkh3d11dxzX3Tn/1+fBPfc6uL2t75vXsLczszPP7O09+5lv2miapkEQBEEQBEEQDIytoU8gCIIgCIIgCESEpyAIgiAIgtAoiPAUBEEQBEEQGgURnoIgCIIgCEKjIMJTEARBEARBaBREeAqCIAiCIAiNgghPQRAEQRAEoVEQ4SkIgiAIgiA0CiI8BUEQBEEQhEZBhKcgCCaPjY0NnnjiCZgD48aNQ7NmzRr9vOfPn1fv04IFCxr93IIgCLVFhKcgCEbj0KFDGDVqFJo2bQpnZ2eEhoZi6NCh+Oyzz4w9NEEQBMEA2BvioIIgCDWxdetWDBo0CBEREXjkkUcQHByM6OhobN++HZ988gmefPJJYw/RrKB4z83NhYODg7GHIgiCUCUiPAVBMArTp0+Hl5cXdu3aBW9v7zLbEhISGn082dnZcHNzg7lCNzutxoIgCKaMuNoFQTAKZ86cQceOHSuIThIYGFjpa3799Vd06tQJTk5O6rUrV64ssz0yMhKPP/442rZtCxcXF/j5+eGOO+5Q8Y/6MA6SQm3jxo1qf54vLCysdPtff/2Fq666SglRDw8PjBgxAkeOHKlyPBR8fFy2bFmtr59xoDfeeCM2b96MK664Qh2jRYsWWLhwYYV9z549q67D19cXrq6uuPLKK7FixYoaYzzj4uLwwAMPqGvjexYSEoJbbrmlwvtR2+sVBEG4XER4CoJgNNfwnj17cPjw4VrtT4FGkXjXXXfh/fffR15eHm6//XYkJyeX7kPrKV343OfTTz/Fo48+inXr1uGaa65BTk5OhWPyeEePHsUrr7yCqVOnqnWLFi1Swsvd3R3vvfceXn75ZbXPgAEDygi21atXq/NT7L3zzjsYOXKkEnm7d++u9Xtw+vRpFePKuNYPPvgAPj4+KjlJX/TFx8ejX79+WLVqlRovLcW89ptvvrlGocvxcR+O64svvsBTTz2FzMxMREVFle5T2+sVBEFoEDRBEAQjsHr1as3Ozk4tffv21V544QVt1apVWkFBQYV9OVU5Ojpqp0+fLl134MABtf6zzz4rXZeTk1Phtdu2bVP7LVy4sHTd/Pnz1boBAwZoRUVFpeszMzM1b29v7ZFHHilzjLi4OM3Ly6vM+m7dumkhISFaWlpamWvicZs2bVrj9XMf7rtp06bSdQkJCZqTk5P23HPPla575pln1H7//PNPmXE2b95ca9asmVZcXKzWnTt3Tu3HayOpqanq+YwZM6ocQ12uVxAEoSEQi6cgCEaBVr5t27Ypy92BAweUFXPYsGEqs/23336rsP+QIUPQsmXL0uddunSBp6enckProHtdR2FhobKGtmrVSrnz9+7dW+GYTGqys7Mrfb5mzRqkpaXh7rvvRlJSUunCffr06YP169er/WJjY7F//36MHTtWxanqX1OHDh1q/R5wX7q4dQQEBKgwAf1r+vPPP5UrnhZIHbROjh8/XlkkaZ2sDL4Xjo6O2LBhA1JTUyvdp7bXKwiC0FCI8BQsmvz8fDz44IMqc5oihbFxFDuCadC7d28sXbpUCaOdO3di2rRpyhVM93N5QcXfYXnomtYXVczqpts8PDxcxTT6+/srMUdxlZ6eXuH1zZs3L/P81KlT6nHw4MHqdfoLXeu6pCfGkpLWrVtXOCaFY22pzTXxXJUds3379mXGUh5eP13njN8MCgrC1VdfrcQ94z7rer2CdSPzqNCQSFa7YNEUFRWpJA7GBzLB4ueff8ZNN92kLEW0GgmmAS1zFKFc2rRpo2ISFy9ejFdffbV0H33LpD6XPPGXYAmm+fPn45lnnkHfvn2VNZIxmIz5LCkpqfBafQsp0e3DuEeWdyqPvX3DTpm1uabLge8DP+9MgmKMKOM3GY/6999/o3v37o1+vYJ5IvOo0JCIxVOokddee019edMFZwxopWnXrl2lwqEmmKVLCxjv1G1tbZUAocg5ceIErIk5c+ao94CWC1OnV69epe7surJkyRLl/maiji5phy5qWjxrg86Vzyx3uvbLL0xS0iVG6VsM9WnozxbPVdkxjx8/XmYs1V3Tc889pyyYTOQqKChQ749uW22uV7h8ZB41f8xpHjVlRHiaEZy0arMwpssUyMrKUhar4cOHqzIw9Wnnl5GRodyFU6ZMURPe5UKhkJKSouL+DAknJo65SZMmyqrGeDnG09UGZjSzdA5L67B0Dt3FdJP+/vvv9T4XM6UpOObOnQtTgfGDlVn2GNNYV5e1vgWx/DHZBam4uLhWr2eMKV2Jb7/9tooRLU9iYqJ6ZFmibt264X//+18ZFz7f96piLuvLDTfcoMIQ9F2brDn65ZdfKitUVTGlzOJn9rs+FJosl6T74qzt9VoS5jaPslID28WyfBgFIIXPnXfeiZMnT9b6GDKPWu48ao6IH8WMoDtMH9b74x9G+fW62C9jwzv7N954Q02UXbt2rddE/s033yg3D5MfLhfG/917770qjlA/IcQQcIKi9Y2uTsYBUnBTQFBs6SeJVAZj9hjnSMsdJ0EKiF9++UUl4XDCY1JJXc/FGpE83ocffqjc0fxiNTYcB6/t1ltvVZYYTugshfTTTz8pQUV3e11hXUz+PfD3S0FGsbZ27VpVz7M2UITNnj0b9913H3r06KEsO4x3ZPkh1s3s378/Pv/8c7UvXdYsQ8T3mPFv/CKmyKVA4E1XQ8EyTz/88AOuv/56VQ6JN3EUvOfOnVOfi6qEBIXJtddeq0QK3wu6zVlaieWZeF11vV5LwdzmUQrGLVu2KBHFhDrG6PJ3wt8Xu3yxfmxNyDxqufOoWdIgufGCUZg4caIql2JoXn31VXWexMTEOr0uLy9Pi42NVT/v2rWrTKmX2tKlSxft3nvv1S4XlugZMWKEds8992glJSWaIdmxY0eFMja5ublay5YtVdmg+sCSP127dtXatm1b73Pt3r1b7btu3TrNFPjrr7+0Bx98UGvXrp3m7u6uyiW1atVKe/LJJ7X4+Pgy+3Lc/LxXVpJo7Nixpc9ZQuiBBx7Q/P391TGHDRumHT9+vMJ+unJK/FxWxvr169VrWVLI2dlZvZ/jxo1T76E+v/zyi9a+fXtVAqlDhw7a0qVL1XlqW06Jn8nyDBw4UC36nDlzRhs1apQqfcTxXHHFFdoff/xRZp/y5ZSSkpLUe8b3183NTV1Lnz59tJ9//rne12uJmPo8umXLFi0/P7/MupMnT6rP3JgxY2p1DJlHLXceNUdEeFrohHn+/Hntscce09q0aaO+SHx9fdUXF7+c9MnIyNCefvpp9SXIL/6AgABtyJAh2p49e6qdMHl8/lF27NhR1fyrifoIz7Nnz6rXLFiwoMp95s6dq3Xv3l1zcXFR++ovrHNIWOdw9OjR2o033qgVFhZqhub5559XtSnT09PLrH/77bfVuKKioup1XI4/KCjoss7Fz8FTTz1Vr/MLgiViTvOoPj169FBLTcg8WhaZR42PuNotFP0OLsxCZPYhXWpMFmAMGmNeCDu70L3AGCK641j3kJmLx44dU66cqlodsvwKXX50UTF2xhBw/KSqcUyaNAkff/wxrrvuOuWWvXDhAj766CMVq0aXa8+ePdV+EyZMUIkqzOqtKUuXr62s7E5l8Porc3Pu27dPZWbTjakPazES1n9kuZ+aYBwf3VocD+tasizO6NGjL+tcfC/pthMEwXznURqNGDLBsI6akHlU5lGTw9jKVzDMnXptO7jQtVaZC1Mf/Tv1Y8eOaU2aNNF69+6tpaSk1Hqs9bF4/t///Z96DburlIfdXriN1gh9Xn/9dbV+586dpRYFPqe1gu5G3aLfLaa8y7H8HX9VS3mrhw5aLwYPHlxh/ZEjR9Tr5syZU6vrnzBhQum5bG1tlaWl/Hte13ONHz9eWTUEQTC/eVTHokWL1LHmzZtX474yj8o8amqIxdNCKd/BhVmN+h1cmExA+HzHjh24ePGiCsCuDpZi4Z0ij8O7xvJ3hw0NrQa8s66sThzvyHmnPGPGjDLrBw4cWJpYwZqQLDVTl5qITIKqbdZkZXUPCe+uWby7PAxM122vDQxyZ0kg/m5YN4+Z2UzAuZxzsTg51zHQXmetEQTBfOZRltGaOHGiqlPLRJeakHlU5lFTQ4SnhcI/Cmbdsph2TExMmUlD3wXC2m6cvOhGoEuFWXz333+/KkFRHhYMZgcUulqMWTSY2Zmc1G655RZVXkQf3YRSX1HMCYX1Cy/3y6qyOm+60jbli5ZXBTO9uRD+TugK4++AX3C6bMq6nkv3OZBsTEEwv3mUGe2spMBscrr2q2pAUBtkHpV51FhIHU8LhaUepk+frkqp8C6PxaM5ybCsjH4BYW5nX2iWgeGdOu98GTfEO/Hy3H777Sou6bvvvmuUa+BYOTmyJIY+jLNiuZrKyojs2bPnskqhcMLl5F6bparakKzxWFnxc926miwiVcG7dsac6dfvq+u52IqRd+i1nbQFwZoxpXmUQpcltdgMYeXKlbWeR2QeLYvMo8ZHLJ4Win4HF/27t8o6uPCP7vHHH1cLezMzcJqTLSc5fTiZ0mXD/ViE+p577jHoNejuUlmvkPXrdOgmUHbOKH8XyjaLnPDrW9iYgfiDBg2q1b4cF+tNloeFxVn7jW45fYsB77B12+uDzt2jb2mp67k4ZlOpTygIpo6pzKM8J610FEusS1tV04DKkHm0LDKPGh8RnhZKbTq48Gfe8eoXAWbrPN7dVeZ2oFuB3VJ0RXnpJmIxXkPBGCaye/fuMhMmC9ITTsDPPvts6XpmZjLu6ttvv633ORsiNol31DNnzlTv1eTJk9U6vp9017Ebhn52JGOEWKybGa26rFZ+afH3oA/jy1jomnfY+l86dTkX4fszZsyYWr8fgmDNmMI8yuMzJpTNEJYvX146L9YWmUf/Q+ZR00CEp4VSmw4unPhYIoR/dJwoOAFyH7oh9O/w9WHZC05II0eOVO4ltjdkSZCqYIcNWgcY2E3YrozlOnRurOo6XzA+im4gjomdYXTwGnj+X3/9Vf3xs7sKS5ewu8vDDz9cOiGwkwwnHt7F6gLA2RXniy++wMaNGw0Wm8SJil1G2NmDkx+tBuw0Q9fWvHnzyuzLVoi0DLC1KHs568qWcMxs7xYaGqrcUXTLMamAvxf9uLC6nIvuM74njOkSBME85tHnnntOlQGixZN/v+UFIbsIVYfMozKPmhzGTqsXDFMGpDYdXNgNg4Vz2cnBw8NDlcfgz1988UWZY1VW+JhlRthdhcfevn17lWPk+epaRkOfDz/8UJ2jfFkTXh87q/j4+KgOHix+XFlpkWbNmqnyJ4RFj1kIevPmzZqhYdeLyZMna8HBwWp8LJuycuXKKsuO8D3W8cMPP6ji0yxybG9vr66Rz5cvX35Z55oyZYoWERFh8I4jgmBOmPo8yu3VlSOqDTKPyjxqStjwP2OLX0GoCsbh8I6dWaMPPfRQnV/Pu1haEh577DF8/fXXWLp0qbIuWBt0GzGOin2/n376aWMPRxCERkTm0YZB5tGGQbLaBZOGLq4XXnhBBeTrZ5HWll69eqmuFJww3nzzTbz11luwRhir5ODgoDqsCIJgXcg82jDIPNowiMVTsGjWrVun7k4Zr8T4JWapCoIgCLVH5lGhIRHhKVg0TGxidinv+Dl51qUMiSAIgiDzqNCwSFa7YNGwlR0nzH79+slkKQiCUA9kHhUaErF4ChYN6+u1bt0aW7ZsqbR9nSAIglA9Mo8KZpNcNHv2bFWwlp0AuLCQbWUtxATBULz88ssqI1MmS8FckXlUMDYyjwpmIzxZVPfdd99VBVfZNYHlGFh09ciRI4Y8rSBg//79Kh7p4MGDePvtt409HEGoNzKPCsZC5lHBIlztvr6+qqRDfWqJCYIgCDKPCoJgvjRachH7zS5evBjZ2dl17jUrCIIgyDwqCIL5Y3DheejQITVB5uXlqd6oy5YtqzIrjsVpuehgoVv2RGVPWRsbG0MPVRAEK4ROH/bbZtYue2ibIjKPCoJgMfOooXtyso/tqVOntN27d2tTp05VPW+PHDlS6b66XrayyCKLLI29REdHa6aKzKOyyCILLGQebfQYzyFDhqBly5aYO3dujXfq7C8bERGB6Oholc0pCILQ0GRkZCA8PFwVyWYihTkg86ggCOY6jzZ6AXm6ffQnRX2cnJzUUh5dGRFBEARDYU5uaJlHBUEw13nUoMJz2rRpuP7669XdNn3/33//PTZs2IBVq1YZ8rSCIAgWg8yjgiBYEgYVngkJCbj//vsRGxurTK8sgszJcujQoYY8rSAIgsUg86ggCJaEQYXnvHnzDHl4QRAEi0fmUUEQLIlGj/EUBEEQBGPVQS0sLDT2MAQLxMHBAXZ2dsYehlkgwlMQBEGwaFi8JS4uTmXcCoKh8Pb2RnBwsFklKhoDEZ6CIAiCRaMTnYGBgXB1dRVhIDT4jU1OTo6KxyYhISHGHpJJI8JTEARBsGj3uk50snuTIBgCFxcX9Ujxyc+auN2rxjT7wwmCIAhCA6CL6aSlUxAMie4zJnHE1SPCUxAEQbB4xL0uGBr5jNUOEZ6CIAiCYMY0a9YMH3/8MSwFS7seoSwiPAVBEATBBImOjsaDDz6IJk2awNHREU2bNsXTTz+N5ORkWDOvvfaasi5ysbe3h7+/P66++molVqtqJVsV7ALG40jFg8ZDhKcgCIIg1ILdu3dj8ODB6tHQnD17Fr169cKpU6fwww8/4PTp05gzZw7WrVuHvn37IiUlBcZM2CopKYEx6dixo+rmFRUVhfXr1+OOO+7AO++8g379+qnWsoLpIsJTEARBEGrBwoULlchZtGiRwc81ceJEZeVcvXo1Bg4ciIiICFx//fVYu3YtYmJi8NJLL5XZn2Lr7rvvhpubG0JDQzFr1qwy5X5oJeQxnJyclAX1qaeeKt1OK+HkyZPV6/j6Pn36KEugjgULFqgalb/99hs6dOigjvH111/D2dm5gqWQFlmKcx2bN2/GVVddpbK+w8PD1Xmzs7NLtzML/KabblLbmzdvju+++65W7w8tnayZyWvp3LkznnzySWzcuBGHDx/Ge++9V7off1cU8B4eHmr/e+65p7Ts0fnz5zFo0CD1s4+Pj7J8jhs3Tj1fuXIlBgwYoK6b1RBuvPFGnDlzplZjE6pHhKcgCIIgVEFkZCT27NmDvXv34qefflLrfvzxR/Wc67m9oaE1c9WqVXj88cdLy/TooHgaM2aMGgsFpY4ZM2aga9eu2LdvH6ZOnaoE4Jo1a9S2X375BR999BHmzp2rLKi//vqrEms6nnjiCWzbtk1d18GDB5X1cPjw4WpfHaxTSUFHwXnkyBE1BooyHlvfEspxcRuhUONxbr/9dnVcbqMQ5fl0UOgxpICCfsmSJfjiiy9KhWFdadeunRLnS5cuLV3HDPM333wTBw4cUNdNsakTlxTCuvGfOHFCWVA/+eQT9Zzi+Nlnn1XWbVqZbW1tceuttxrd0msRaCZMeno6/6rUoyAIgiGw9HnG0q+vJnJzc7WjR4+qx/rA90632NjYlHnULQ3N9u3b1XGXLVtW6fYPP/xQbY+Pj1fPmzZtqg0fPrzMPqNHj9auv/569fMHH3ygtWnTRisoKKhwrMjISM3Ozk6LiYkps/7aa6/Vpk2bpn6eP3++Ot/+/fvL7PP0009rgwcPLn2+atUqzcnJSUtNTVXPH3roIW38+PFlXvPPP/9otra26vdx4sQJddydO3eWbj927Jha99FHH1X5/rz66qta165dK902ZcoUzcXFpcrX7tq1Sx0/MzNTPV+/fr16rhtzVSQmJqr9Dh06ZLDPmrXMM2LxFARBEIQq+Pbbb5Vbl+gsjLpHrud2Q6Fv0awJxn2Wf37s2DH1My2Yubm5aNGiBR555BEsW7YMRUVFatuhQ4eUpbJNmzZwd3cvXei21nct0+3fpUuXMuegZZMu+YsXL6rndJOPGDFCWUIJrYx00+sfd9iwYcpqeO7cOTU+voc9e/YsY7XUvb6+75l+WSNapenKZ5gB3e0MWyCMDa0OWnsZusD3zNPTU2Xa1+Z1Qs1I5yJBEARBqAKKq/bt25cRRzp27NiBHj16NPg5W7VqpcQThRndu+XhesYkBgQE1Op4dCnTlcz4ULrf6cKna57iMisrS3XZoUAr322HQlEHXf7l61T27t0bLVu2VC76xx57TAlaCk0dPPaECRPKxJPqoBA8efIkGhq+N4wV1bnLKXS5UBTz/aJw5POCgoJqj0OxyioCX331lYojpVju1KlTja8TakaEpyAIgiDUAsb5UYDoHg0Fk1mGDh2q4h0nTZpUJs6Tfecpou6///4yQnD79u1ljsHnFMw6eAyKKS5MXKJlkdbO7t27K4sn4yqZBFQfYc7xhIWFqfeFFk8dFOVHjx5VQroyOAZaXil6KWIJBXJ9SxsdP35cJQVNmzat9DlLT7377rtKfJPyFQloySV8D3TwNRwHRafuPWFsqtAwiKtdEARBEKqBvbeZ1EOrJ0sa8ZHPud5QfP755yrbnNa5TZs2qQQciioKUmafT58+vcz+W7Zswfvvv6+siMxoX7x4sUowIrRCzps3T2V8s0wTwwMoRGnRo4ud4pFClkk5dIHv3LlTlSZasWJFjePka5loxfGMGjVKZbzrmDJlCrZu3aqSifbv36/c18uXLy9NLmrbtq1KPqJVlNZjCtCHH364QkJVZVCwUoTTzU8B/dlnnyk3erdu3fD888+XWlUpLLmN182sfCYa6cP3gAL+jz/+QGJiorLS0ppM8f/ll1+qMlZ///23SjQSGgjNhLH2oHhBEAyPpc8zln59NdFQCR95eXlaSUmJ+pmPfG5ozp8/r40dO1YLCgrSHBwctPDwcO3JJ5/UkpKSyuzH5KLXX39du+OOOzRXV1ctODhY++STT0q3M0mpT58+mqenp+bm5qZdeeWV2tq1a0u3M+nolVde0Zo1a6bOExISot16663awYMHS5OLvLy8qhznFVdcoT5jf//9d4VtTBwaOnSo5u7urs7dpUsXbfr06aXbY2NjtREjRqikpIiICG3hwoXqempKLtIldjExytfXVxswYIB6Tfnfy/fff6+ui8fv27ev9ttvv6nX7du3r3SfN954Q71nTBrj+03WrFmjtW/fXr2OY96wYUO1CV9EkotQq3nGhv/BRMnIyICXlxfS09NVcK8gCEJDY+nzjKVfX03k5eUpKx7j/lh3UhAMhTV/1jLqMM+Iq10QBEEQBEFoFER4CoIgCIIgCI2CCE9BEARBEAShURDhKQiCIFSgqFhaAwqC0PBIHU9BEAShAov3XEBKdgFcHe3g5mQPdyd7eDjbw9/dSS2+bo6wsy1bUFwQBKEmRHgKgiAIlVJQVKKWtJzCCtsoOn3cHBHo4YQIX1e1UKAKgiBUh8wSgiAIQp0pLtGQlJmvlqMXM9Q6f3dHRPi5oem/QtRWLKKCIJRDhKcgCILQICRlFahlb2SqctG3CfZA+2BPBHtZV01DQRCqRoSnIAiC0ODkFBRjf1SaWnxcHdA+xBOdw7zg6ihfO4JgzcgMIAiCIBiU1JxCbD2TjJ3nUpQA7dHURyUnCYJgfUg5JUEQBKFRKCrRcCgmHQu3ncfy/TG4kJpj7CGZNOPGjYONjQ0effTRCtsmTpyotnEfQTAnRHhaGCUlGvIKi5GeW4jEzHy1JGXlIzkrX5VGScspQE5BkdpPEATBGGgacDYxG4t3X8DSvReQkJln7CGZLOHh4fjxxx+Rm5tbpif4999/j4iICKOOTRDqg7jazSyLNDWnAKnZBcjIK0JmXiEy84qQlV+ErLwi5BcVo7C4doLSxgZwsreDi4MtXBzt4OnsAG9XR+X+8nFzgI+rIxzs5L5EEATDEpmcg6iUKLQL9kS/Vn5qLhL+o0ePHjhz5gyWLl2KMWPGqHX8maKzefPmpfuVlJTgvffew5dffom4uDi0adMGL7/8MkaNGqW2FxcXY/z48fj777/Vdr7+8ccfx9NPP116DFpP09LSMGDAAHzwwQcoKCjAXXfdhY8//hgODvJ7ERoGEZ4mCq2Wsel5iM/IQ3JWAZKz81UtPYrPhrI48BxcGH91EXkVhClFaKi3C5p4uyDUx0W+EARBMAicj47FZuBUfCa6RXjjiua+6sbYkOfLMYKX39X10txaVx588EHMnz+/VHh+8803eOCBB7Bhw4bSfd555x18++23mDNnDlq3bo1Nmzbh3nvvRUBAAAYOHKiEaVhYGBYvXgw/Pz9s3bpVCdGQkBDceeedpcdZv369WsfH06dPY/To0ejWrRseeeSRhnkTBKtHhKeJQDc4450oNmPTcpGWW6gmR2PBcyvBm1WAgxfS1TpPFwdVn691kDvCfaRGnyAIDR8Duvt8Kk7EZeKatgFoFehhkPNQdLq7o9HJygLc3Or+OgrIadOmITIyUj3fsmWLcr/rhGd+fj7efvttrF27Fn379lXrWrRogc2bN2Pu3LlKeNJi+frrr5cek9bSbdu24eeffy4jPH18fPD555/Dzs4O7dq1w4gRI7Bu3ToRnkKDIcLTSNBNHp2Si6iUHCU46TI3dTJyC1ViABe651sFuIsIFQShweF8+PuBWLQMzMSgtgHwsHJvC62WFIALFiyApmnqZ39//9LttEzm5ORg6NChZV5HV3n37t1Ln8+aNUtZS6OiolTMKLfTmqlPx44dlejUQevnoUOHDHp9gnVhUOFJ0z9jUY4fPw4XFxf069dPxaC0bdsW1gaTeWIz8nAuMRvnkrJUkWVzJreguFSEsn9zlzBvdA71UoJUEISGw5rn0TMJWYhOyUG/ln7oFu6tsrgbyuVN62Njw/PWF7rbn3jiiVIBqU/WvxezYsUKhIaGltnm5OSkHmkhnTx5sordpFXUw8MDM2bMwI4dO8rsXz6Wk+853fSCYBbCc+PGjarkQ+/evVFUVIQXX3wR1113HY4ePQq3+vgbzAwm+zBw/mxiFs4n5yixZqnWiS2nk7DjbDLahXiie4Q3/N0vTXaCIFwe1j6Pslf8hhOJOJ2QhWGdghsk1pz61dzeuuHDhysLJYXgsGHDymzr0KGDEpi0ZNKtXhl0z/OmhQlFOpi0JAgWJTxXrlxZ5jndBIGBgdizZw+uvvpqWCLZ+UU4k5ilFrrSGyoZyFzisw7HpKulmb8r+rfyR6CHtMoThMvBGufRyriQmotvt0dicLtAlQFvbdD9fezYsdKf9aH1ktbMSZMmKesks9LT09OV2PT09MTYsWNVwtHChQuxatUqFd+5aNEi7Nq1q0xmvCBYXIwn/xCIr68vLAnWyDyTmK0sm3EZeUZNCjIVziflIDKZJVI80LelP7xcrDtGSxAaCkudR2tDfmEJ/joUp0KWBrULhLODdYX2UERWxZtvvqliQRmacfbsWXh7e6tSTLSQkwkTJmDfvn0qS51W07vvvltZP//6669GvAJBAGw0Rio3ArwLu/nmm1WNMGbaVQYz87joyMjIUMVzOdFW9wfX2NCKeTEtF2eTLolNljkSqsbe1gZdwr3Rp7mv1X1RCKYP5xkvLy+Tm2eMPY/+sDMKcemmW9idseXXdw5RJd+qg8XWz507pyx7zs7igREMhzV/1jLqMI82msWTMUqHDx+ucrIkvFPTL/dgauWOIpOz/81Cz1VxR0LtXfB7I1NVnT6WSLFGN5kgNATmPo82dGz5kt0XMKC1H3o2tT7rryCYK41i8WQm3vLly1VB2+riSUzF4llYXKIKt3OJS89HbHquWZQ7MhdaBLjh2vZBcHeSal6C8TEXi2djz6OmbvHUp1WgO4Z2CKrUo2LNViihcbHmz1qGqVg8qWmffPJJLFu2TBW6rSmImVl5utIPl8P2s8lwcbCDm5M93JwuPTra2arYyxJNQ7GmQSsBCopLkJFXqOpTsgUlH1m4PSWrQO0nGAb2aI5JO4+BbQLQsYmXsYcjCCaNseZRc4IZ70lZ+RjRJUQSGgXBxLE3tFvo+++/V3fpzLpjf1hCVcx6dIZi17kU5d4VTDtJYPWReJyKz8KwjsFS/1MQTGweNTcYa//TzmgMbh8oN7SCYMLYGvLgs2fPVmbXa665RnU/0C0//fSTIU8rmBHnkrLx3Y5Is3HpCUJjI/No7aHBgTe0608kqKYdgiCYHgZ3tQtCTTB+9ufd0cr13jXc29jDEcycouIS2NsZ9J66UZF5tO7sj0pDUmY+buzSBNLMVxBMC8uZnQWzhiWq/j6egJWH41RylyDUh93nUxCf+V9ijWC9sPrI9zujVJ1lQRBMBxGegknBkks/7opGZp7URhXqxpGL6dh8OsnYw7A4ok8ewhfP368ezQ0mjP51OFZuZgXBhBDhKZgcdJH9tCtaZakKQm1gi9q1RxOka5gB2LVmOU4f2IHda5fDHCkq1pBbUIzs/EIJWxAEE0CEp2DScZ8XUnOMPRTBxIlJy8Vfh2KlBFoDEhkZibPHDuHCqSPYv3GFWrdvw5/qefTJw0iJj4G5kVtYgvTcQvmcVAIF+fjx41UbVrbT3L9/v0pme+aZZ6p9XbNmzfDxxx832jgFw8PmOPzeNWSIilTwFhoNuup+/2oGbnrkeYS36VyrkkvL9sZgeKdgtA7yaJQxCuZFYmY+lu+PQWGxiImGhIKiPFlpyfhw4m2lzz9cfQLmRn5RCVKzC+Dt6gA7W1t8tOZko5170tA29Xody2dNnz4dK1asQExMDAIDA9GtWzclCq+99toGGdvKlSuxYMECVSe2RYsW8Pf3x9KlS+Hg4NAgxxdMl+SsfHXzzsoybJqTkn3p5mxwu0D4uRumHrAIT8EoLrvaCE9deZQVh2IxML8I3SN8DD5GwXzIyi/Cr/ti1A2K0LB8++23GDt2HIqLK3Zss7Wzw92T34W5wjmFLZC9XBxh6pw/fx79+/eHt7c3ZsyYgc6dO6OwsBCrVq1S9V2PHz/eIOc5c+aMKtHVr1+/0nW0fgqWR1JWvkq8o1UzJjUXOQXFjT4GcbULBoUuObrmLsdlR8/YhhOJ2HU+pRFGLJhLFYQ/DlxU4lNoeMaMGYM3vvm10m3PfLoYPa+9GeYMS3ym5hSo0lumzOOPP65c3zt37sTtt9+ONm3aoGPHjnj22Wexfft2tU9UVBRuueUWuLu7q1aFd955J+Lj40uP8dprrykL6aJFi5Qlm40H7rrrLmRmZqrt48aNU52xeByeS2ftLu9qT0hIwE033aSaFrB71nfffVdhvGlpaXj44YcREBCgxjJ48GAcOHCg1mMhJSUleP/999GqVSvVgSsiIkJZfHVER0era6QYpzjmtVOgC5WTV1iMk/GZWH0kDl9tOotF2yKx/niCat5iDNFJxOIpGJS37hvcYC67zacuZSz3biZ34tYOJ85YaTrQKFCMMAZQ92hJ0PVOt6KjPW0wplXxMyUlRbnAKbrc3NwqbKfwokjTic6NGzeiqKhIWUJHjx6t3Ob6Fs1ff/0Vf/zxB1JTU5Vwe/fdd9WxP/nkE7Rs2RJffvkldu3aBTu7yrvIUaBevHgR69evVy74p556SolRfe644w4lTP/66y8lKufOnavCAU6ePFlqQa1uLGTatGn46quv8NFHH2HAgAGIjY0ttezS2jts2DD07dsX//zzD+zt7fHWW29h+PDhOHjwIBwdTd+K3RjQos+Ey3OJ2WqeNLW4ZhGegkEZM2UGfpg5FSXFxQ3isqP45NdDLxGfVsvBC2k4FJNu7GFYPJ4+/vDw8Yd3QAj6DB+FHSuXIC0xFu7efrAkGB9copXAycEWNiYkPk+fPq2Efrt27arcZ926dTh06BDOnTuH8PBwtW7hwoXKKkoR2bt3b7WOApUxnGy5Su677z71Woo9CkSup+AMDg6u9DwUjhSTtLzqjjlv3jy0b9++dJ/Nmzer7RSjtFSSmTNnKpG5ZMkSlbxU01ho+aQQ/vzzzzF27Fi1naKYApSwWxdf//XXX6sbITJ//nwlwim0r7vuOlgjmqYhLiMPZxKyleCk8DRlRHgKBoUuuaCIlmUsnPouu7DWHet8zH/+tXyK+LQ+LqblqrALwfD4BYXg5UXrYefgoL7k+44YjeLCQthboFWJoRssueTsYAfbfwWNsamNdfnYsWNKcOpEJ+nQoYMSYtymE4l0a+uEHmE8Z3lrZU3noXWxZ8+epesoiHkeHXSpZ2Vlwc+v7I1Jbm6usnLqqG4sPE9+fn6VSVM8BwW5/utJXl5emXNYA5qm4WJ6nnKjn0nIUpVgzAURnoJZuuxEfFofjOdccTBWiQShcdAXmfy7tUTRqYNTUm5hMZztbVXGu7Fp3bq1es8bIoGofHY6j0vLYUNC0UkRqe/i16EvUKsbC930NZ2D4rey+FLGlVqT2Dwdn2W2Me4iPAWDQ9ecIVx2FJ8OdrbS390KoNhccVCSiQQDozEZowSO9lBzizFhTCTjGWfNmqXiKcvHeTKRh65uJttw0Vk9jx49qrbR8tlQ0LrJ+NE9e/aUWlFPnDihzqOjR48eqvQTLaOVleOqrdim+KTrnUlK5eE56G5nSSkmL1kDmqapOE0lNs3MslkVIjwFg+MdEGwwl936EwlwdbSTOp8WzpbTSbiYJslEQuMV0WZChpORk44oOllO6YorrsAbb7yBLl26KAG4Zs0azJ49W4lMllhiFQIWcuc2ZsIPHDgQvXr1arBxtG3bViXwTJgwQZ2X4pIZ7/oWyiFDhqikn5EjR6qsdGbgMxmJ9UdvvfXWWo3H2dkZU6ZMwQsvvKAShXjtiYmJOHLkCB566CF1nSwrxYQqvh9hYWGq2QFrjvI1fG5JMZsn47NwKj7TIsSmPsb3JwhWAUWmLhi8IV12dI+tPByH6BTpcGSpRCXnYG9UqrGHIVgZulabxszkZzH3vXv3YtCgQXjuuefQqVMnDB06VFkEKQA5ly5fvhw+Pj64+uqrlfjja2gVbGiYxNOkSRMlam+77TaVLETLow6O5c8//1TjeOCBB5TwZKkkCsOgoKBan+fll19W1/rKK68oiy4z9HUxoK6urti0aZMqscQxcDsFKWM8zd0CqtzoKoY9AfM2n8OPO6OxNzLV4kQnsdFMuD5GRkaGyrhLT0+v04fqs3WnVJFgwXpgRuodPcMR4GGYTguCceAX/7fbI+vkYr+zdzhCvauPFWuIecZcqO/1/bAzSnUzMXdcbIvR3acQoRFN4eBY9/nB1gbwcnGAo33lZYYEQQcFMCsMsM4prbe1CSGKTsnB2aQsnE3MNimRyc5FdQljq8s8I652wSJg9xp2saHo4JeEYBmsPhoncZ2CCRSbL4SbkwY3R7tSz40g1PdmOjIlWwnNc0nZKqzD2hDhKVgMFCjL9l7A6N4RcHEU64S5sz86TU3OgmAKZOcXobC4BF7ODrClGVQQakFJiYbYjDxEJmUjMiVH9UM3XT9z4yDCU7AoaJn4/eBF3N4jDHby5WDW/YT/OSn1OgXTgtap5OwCeLs4wEElHglCWSgqKS4T4i71Qr+Ynqs8csJ/iPAULA7+sa87Fo/rOlbehUMwbdg/+6/DcRKnLZgkzHZPySmAu5O9qqghrnfrhSkyjNPkXEVreHZOATLzC7HpUCxyS8TrVhUiPAWL5MjFDPi5O6JnUykwb25sPZOMpMx8Yw9DEGoM7aEF1NPF3iQKzguGd5kXlZQokcmKB5ceS1j6tRR1syz3yzUiwlOwWFhg3sfVES0C3I09FKGWxKbnSukkoYH51yJpgMC6guJLrndPZwfVblMwf3GpLJj/WjJ1CwVnrT4+pTuJFbw6RHgKFgvnALpsR/cOh7+7lFkydWg9WHM03uoD74WGpaDEBsWahoL8PDg41Vzipq7w85qeW4j8ohJ4ONubTK93ofIwCZ2YVCKznMC83KmHnzH1WSuRz0B1iPAULBq6wpbvv4i7rwiHK/vgCSbL9rMpSM4qMPYwBAujGDa4mGMLh6RLyWqOFJ8GEIeFBUB2jo0queQk1k+jxl0WVSIsKToNFjb+741NclKi+qzxMydUjXwTCxZPRm4hVhyMVZnuUgbFNGGh8j2R4mIXDEN0HjulFaCwOB52jWCRtLezhbODrVg/DYxOXCpRWYLSn40RZ8lzU3Re+qwJ1SHCU7AKLqTm4p/TSRjYJsDYQxEqdbHHXfrCEASDYIPoPCdczNPgaMvPmeE/a44OtugZ4YM2QR6S+d4AZOcXIimrQHlFWG6NsbX5hcUmEk9pcymkwyTGYvqI8BSsBva9DfJ0Qrtgy2uLaM7sOJeivlAEwdBQGOQ2Uvxdbj7w96k0HInPw1Wt/RHu69oo57UUkrPycTEtDzFpOcpwUHk7SQlpMEdEeApWxdqj8fBzc5Ke7iYCCy3vPi8udsGyP+NL9lxAiwA3XNU6AL5u4oqtjNTsAkSn5iAq5ZLQZGtJwTIR4SlYFYXFGv44yGSjCCl/YgLxWWuPxYuLXbAK2P71fFIOOoV64soWfnBzsu6vXwpLiszI5Gz1WLlFU7BErPuTL1glaTmFWHk4Drd0ayKxV0ZkX3QqEjKkULxgPfAm6+CFdBy9mIEOTTzRs6kPvF0dratneXI2IpOlZ7k1I8JTsErOJWVj29lk9Gvpb+yhWCWse7jtTLKxhyEIRoHlfihAD8Wko3WgB3o180GQZ8PXGDU2mXmFSmSe/9eqKT3LBSLCU7Badp5LQai3C5r6uRl7KFbH+uMJKuxBEKwZWvxOxmeqJdTHBR2beCoh6mhvni042a+c8ZnR/7rQJWlQqAwRngKsvbPRmD4R8HB2MPZwrIYTcZnK4iwIwn/EpOaqZcOJRLQMcEeHEE+E+7qYdDgQC7MnZOYhKvlSUlBsep5aJwjVIcJTgLUHuP95KBajeobDTorLG5y8wmJsPJlg7GEIgkl3WzsWm6EWtuCkR6apnysifF2NnhDJscWm5yImLVeVOopLzxXPhVBnRHgKVg8n0M1SXL5R+OdUErLzpUyKINQGZnofjklXCw2fwZ7OiPBzVfGg/u5O8HJxMOhNYmJmvrJoXnrMR2p2oVShEExbeG7atAkzZszAnj17EBsbi2XLlmHkyJGGPKUg1Lu4fKi3M1oFehh7KBbLhdQcHLmYbuxhmB0yjwqEeo+ubC46nBxslQBlXWJPZ3u4OtrD3YmPdqpck4Odreqlo98qmNnl+UUlyC8qRl5hiRKYOQXFSMstUO2FWfWDyX9cJwhmJzyzs7PRtWtXPPjgg7jtttsMeSpBuGxWH41Xk7i1lDdpTBj39ffxBCmfUg9kHhWqglniutjQmqDF1AY20PhP/g4FSxWe119/vVoEwVwm8T8OxuKu3uGwtzPPrFJTZff5FNVjWag7Mo8KDQHFJkWnIBgb+XYVBD0Yy7TxZKKxh2FRpOcUYtf5FGMPQxAEQTABTCq5KD8/Xy06MjIyjDoewTphYecwH1e0DZZ4z4bg7xPxkvnaiMg8KgiCKWNSFs933nkHXl5epUt4eLixhyRYKewhnpotruHLhYWx2Z9aaDxkHhUEwZQxKYvntGnT8Oyzz5a5U5dJUzBWvbo/Dl2K92RmqFB3mDW78YSELZjjPPr998Cafe7IKnCCjS1gZ6fBxlaDg6MGL/+iS4tfEexM6htEEARzwKSmDScnJ7UIgimQlJmvuogM7RBk7KGYJVtPJyMrv8jYw7A6GmIefftt4MgR32r3sbHR4OFTDO/AQoS2zEd4mzxEtMtDcEQBbI1b51wQBGsVnllZWTh9+nTp83PnzmH//v3w9fVFRESEIU8tCA0CCzeH+bigfYinsYdiVsRn5OHAhTRjD8MiMMY8Onw44B6cjZy8EpSU2KCkmPUfbVCQZ4OMZHukJdmjuNAWGSn2aok67oJtKy691tGpBGGt89CmRw469s1CkxYFqpSPIAiCwYXn7t27MWjQoNLnOvfP2LFjsWDBAvkNCGYB608GejjBz12s8bWBBaoZIyu1As13Hp05E/hhZzLi9IqV61NSAmSn2yE9yR5JsQ6IPumMqBPOuHDKCfk5djh72FUtKxf6wyewUAnQjldmo1XXHHHPC4KVY9Ap4JprroEm3z6CBcR7rlDxnhFwtJd4z5rYF52KhIz/sqoFy5tHbW2h3Oxcwlrno9vVWWo9LaMJFxxx7ogLju5ww8m9rkhNcMDm5T5q8fApQu/rMnDl9enwb1Jo7MsQBMEIyL2nINQCFj//+3g8hncKMfZQTBq22tt+Vmp2WiuM7QxuWqCWvjekK9f8qX2uOLzdHUe2uiEz1R5//+Srltbds9U+nfplwd5wLccFQTAxxHwjNArRJw/hi+fvV4/myrHYTBy6IL3Gq2P98QRlIRYE4uisoWPfbIyeFI9XfziLca9cRNte2Sox6dQ+Nyyc3gTTxzbHxqXeyM+VQFDBfLGE77jGQoSnUAF69fglkJ1hi8ICmwaJ1du1ZjlOH9iB3WuXw5zZcCIBCRmVx71ZOyfiMnEuKdvYwxBMFMZ2dhmQhQlvx+Cl/53D0HuS4elbhPQkByyfE4i37m+OtT/4IjdbvpYE88NSvuMaA3G1Wym5WbaIPeeEmLNOiD3niLREB5UskJVmh6x0OxTm/zf50zph76jB0UmDm2cxgprlI6RZAUKa56NJi3z4BRdWWj4lJT4G2empsLGxwf6Nl1Je9234E72H3qpi1ty8fOAbFApzoqhEU/Ged18RAWcHqRmjI6+wGBtPJhh7GIKZ4BtchOvHJWPoPSnYtcYTf//sg+RYR/w531/9fPWtabhmVAqcXU0rtlUQLP07rjGw0Uwtal0PFj5m54309HR4eta+nM1n604pgSD8l4FKkXl6vwvOHHRFzBknFfDfUDi7FqNDn2x0vToT7XrnqCLT5Nnr2tb42g9Xn4A50irQHTd1bWLsYZgMa47Gq9JTpsCdvcMR6u1i8HnGXKjv9f2wM6rKrPaGprgY2LfeA+t+8kV85KXqEUxEGn5/Mq4Yng47uccTTBBL/o4b3C4QXcO9DTLPiMXTQkmJs8fRnW44td8VZw+6Ijuj4sztE1SIJs3zleXSL6QQ7t7FcPcqvvToza4kGgoLbFGYb/PvYov0ZHtlIY09T0upE+IiHZGXY4e96z3V4uRSgg5XZqHb1Zm4+/mZ+OnDKSjht0o5bO3scPfkd2GunE7Iwq7zKejdrPoi29bAhdQcHLloGqJTME8oLHsNyUSPwZk4uNkdf37jj6SLjlj8SRD++dUbNz2SqG5qpR6oYEqMmTIDP8ycapHfcYZEhKcFWTVjTjvh8FZ3lUEae7ZszUkKwuadclUdvWbt85TYdHGvOQnE3qEELm7/PQ9uVoC2Pf/rvc3yKZHHnXHwHw8c+Mdduez3rfdUi3+TJzH0nmuwalEf2jTKHPeZTxcjrHVHmHtnHtb3bOqn9wZZGYXFJVh7VGp2Cg1XpomlmTr1zcLWP7yx+ls/xEU64av/C0ObHtm47YkEBIZJGSbBNOh57c0IimiJDyfeVmGbJXzHGQoRnmYuNs8fccH+Te44uNlDdRTRwb7KzTvkKisBxSbb2RmicDNjO5t3zFPLTeMTVRHpA5vcsXuNl7JYrFrUE8BxNuEDsAg2NsUmV5OwvpRoGv46HKfiPb1crLMezJbTSUjNESEgNCwsr8Q4z95DM7Dme1/8s9wbJ/e6YcaEphhyVwoGj04tDekRBFOAcZ78btM9ClUjwtMcxeZRZ+zf6KGsjGxXp2/VbNcrW3UJaX9FNtw8SxrdWkFrKhfGZm353Rt//+SFnMxWAL6Bm+cHcHF/Ffm5P8Dd2w+WQG5BMf44eBGje4XD3s7W6lzs+6OlLaZgOOiVuXl8EvrdmI6lnwfi+G43rFrkr8J6Rj0Zj9bdc409RMHK4XeZh48/vANC0Gf4KOxYuQRpibEW8x1nCER4mgG8eYo+6YR9GzyU4GT5ER3ObsXo3C8LXa/OQpvuOSr7vDpYY+z3r2bgpkeeR3ibzgYbs5OLhsF3pqL/TWnYvNwdG5cGICvNB9kZn6LHoDfg4JREGQ1LgF161h5LwPBOwbAWWKuTCUVyYy80Buxy9Mj0GOVNWTY7EIkXHDF7Sjh6XpuBkY8lNPpNtiDo8A4IxsuL1sPOwUFZO/uOGI3iwkLYOzoae2gmiwhPE4Vf6EziodCk4GSpEf0scnb76DawdmKzqlpjhhSe+gL02rsyMeCWLKz53g/rF/tg73pvnDrgjjuejkenvpZR9/FYbAaCvZzRrQ5ZgObuYk8TF7vQiDCxiHNe2145+GuBH7b85o096zxVW85RT8Wjc/9so95sC9aLvsik+BTRWT0iPE1MbMadd8T+TR44sNFD9TzW4ehUolzo3a/JRLtedRObplBrjAL0xoeSlHX2h5lBSIh2wjevhqLH4AzcNjEBrh7mb7HYdDIRfm6OCPd1hSUTnZKDAxfExS4YBxe3Etw2MRE9r83Ejx8EqfJL818PRY9BGbh1YkXrZ2PfbAuCUD0iPE1AbF4866RKiBz8xx3xUU5lMsqZHNT9mgxVJ5PirT68dd/gCuuy0pLLZOI1Vq2xpu3z8NzsKKxc6IcNS3yw929PlRX/8BsxCIowbwtacYmGPw7G4q7e4fBxs8w73vyiYqwWF7tgAjRtl4dnZ0Vh1SKdJ8UTpw64Kk9KkxYnjX6zLQhC5YjwNAIsQXT+qAsObXFXS0r8fzGbdhSbvXJUHcyOV2bD2a3E4mqNMRv1poeT0KV/Fha9HYLki4745OkI3P9SrLp2c+/g8+v+GIvtbPTPySRk5Jr3DYJgOXAuqcyTAqwE8DSAbKPfbAuCUBYRno0E+54zI/PYTjec2OOK7PT/3noHpxJVG7PLgEx07JutXEnWUGuM1s9nPovCgjdCcPawK776v1DcMiERV41MM+tC0Yx9/P3ARdzWIwx2tmZ8IeU4GZ+JQybSnUgQKvWk/O+SJ0XTHgIwkLfdAHaW2VcKewuCcbE64dlYgeZFhVA1Ldk56MRuN+VO1kr+EyEu7sXKoskkIZZAcnTWrLLWGLskPfreBSz5NAg7V3nh19mBqhvS7U8kGKTuaGNxITUXa4/FY1hHy8h0T80uUFnsgmCqKE/KI0lo3ycbP7wfjNQElnHbAuCNf+sIX/L4SGFvQTAuZvzVXj8MFWhOoXnhtDPOHHDB6QOuOHfYBQX5Zes6slsQ62u2752NZh1yG1VYmXKtMRaLHv1sPIKbFuD3r/yx/U9v1Uv+gVcvwtHJ+OK4vhy9mAFfN0ezb6tZVFyCPw7FqhJKgmDqtOqSi8lzIrHobScc3x3+r/AcDuBeAOeMPTxBsHqsQng2dFY3DYVpifbKihl51EU9XjjlhKLCskLT3asILbvmqpJH7XpnwyewCMbC1GuN0bV+zahUBIQVqLhPWom/eaUJHnz9YqNZgw1VdsjN0R4dmnjCXFl/IhFJmfnGHoYg1Knw/J2TzuH98W8hP3cmtJJ+sLE9ACeXZxvsZlvKNAlC/bAK4Xk5Wd2FBTZIvOCgMs9jzjjh4hln9ZiTWTFxxM2zGM075qJVtxy07paj+pqbUqyiOdQaY/jB+Okx+PKlUJzc54Z5rzbBQ2YsPnmTQhe1o70NWgV6wByttoclrlMwQ3iz/cZPTyI9JQHfv2+Hc4c9kJf9FVYuTMetjyfUu0qIDinTJAj1wyqEZ01Z3aOffQ/JsfZIjnNQGdYJFxyQEO2olpQ4B2haRfVoa6shpEW+ag/ZtH2uCm5ndw1TEprmSovOuRj/9gV89VIYTu1zw9cvh+KhN2Iu+4vCmD3d/zwUh1u62aKpnxvMhaSsfPx9XOI6BfOFN9d+wcV4fMYFrPnOT/V9Zyz5+SMuuPfFWIS1qpsl3xRqIguCuWOjmUKGSRVkZGTAy8sL6enp8PSsvavys3WnUFSilfY2z8m0xal9jPn5AABjfsL+XSLg6XclslJdUaKX+FMetqVs0jwfTVrmI/TfJahpgQpmFwzHuSPOyvKZn2OHll1y8PCb5is+iYOdDW7tEYZQbxeYQ1mon3dHIzmrAObGnb3D6/Qe13eeMRfqe30/7IxCXHoeLInTB13w3bvBqu0wS9exrFtdqmg8e13bGveRMk2CJTC4XSC61qETX13mGYuyeGZnA5MmAVsPhSAjxR4ZKXbITLVHSTFnFWY4XlvhNRnJ/xVr9w0qgm9IIQLDChAYXoDAiAIEhReozGuxZDY+zTvmYcI7MfhyWijOHHTF16+EYsL0mDp1bTIlCos1/LovBqN6hiHI0xmmSmFxCX7bf9EsRacg1Cbx6KcPgnF4m7uqonFqnyvumhxXq37vplYTWRDMEYuyeBYVAQxbrOyKXD0KkJdzHI7OKWjSwg3JF7chP/cE7pr8CCLaecLTtwi2ZXODBBMh8pgz5k4LRV6OHboNzMC90+LM+nfl4miH23qEItDD9MRnSYmG3w9exNnEyvtemwNi8SyLWDwrwu+Irb97YfncAJUU6uVfiHunxqFll9waX3vh1JFKayI/O2uplGkSLIbBYvGsHfb2wHvvAXtj4+HmXQRPvyJ4+BTDw6dIlS4qKrCFnUPIvzUsh6G4cPC/CTbGyzYXaobxsyyt9OVLYdi/0RPeAUW4eXwSzJXcgmIs3n0BN3Vpggg/0+nrzntQtsM0Z9EpCLWBHqz+N6ejWcc8LJoegoQLjvjihTBcNyYZQ+9Jga2d+dVEFgRzwYztRpXz/PNAvxEZqgNQeJt8JVJ09TIpMjlJmHJWt1A5rbvn4q7n4tTPG5b4YtOy2t+JmSKsicnWmsfjMmAqbDqVhGOxpjMeQTA0jNefNCsSva9LVw0+Vi3yVwKU5fJqqokc1roTRj31unrkc1OoiSwI5oBFWTwFy6bntZnqC2HFNwFYPidA3VR0GZAFc6W4RMPKw3HIyitCLyMXmd95LgV7I1ONOgZBMAZMWLx7cryqt8wOamcPuWLmo00x+tk4dO6fbXY1kQXB1LE4i6dg2QwenYp+N6apElffvhusMt/NGXro/jmVhPUnEoziruM5WeSeiyBY+43ts19EIrxNnqrTPP/1UCz5NBAF+RUzS8V7Jgj1R4SnYFZwrr9tYgI69s1SMbvzXglVNVjNnf1RaVi6NwaZeYWNWjJp+f6LytopCAIQEFqIJz+KwqA7Lv1NbP3DGx8/EYHYcyIsBctF0wAWauBNVm62LbLSbZEQZ4PMTMOcz/y/sQWrg4H/902LxazJ4Yg+6Yz5rzfBUx9Hm213Ix1RKTlYtD0S17QJNHiLzeSsfPx+4CJScxpP6AqCOWDvANz0SJJyvX8/IxhxkU746IkI3PhwEgbckmbWFTUEy6GoEMjJsEP2vwut9Hk5tsjNskVeNpdLzwvybJCfa4v8PP586XlRgQ0K821RWHjpZy6VNcr58MNLJSobGhGegllCkclM9w8nRuDiWWf89GGQKrNk7vVW8wtLsOpIHE4nZmFI+0C4Ojb8n+jphCx1DiY4CYJQOW175aianz9+EIRjOy/V/Dy6ww13Px8HL7+KdTwFoSHIy7ZVuQxpSfbqMZM1yVP5aKfqk2em2iE7naKyFqUXLgM7Ow3FqgZ6wyPCUzBbmFw09v9iMXtKGPZt8FRVDK4ZZRkJMmcSshCblosBrf3RLtgTdraXPwGk5RQot/rR2IxKa90KglAWluN7+M2Lqubnb18F4OReN8wY3wx3PBOPrleZb2KjYDwK8m2QfNHhUovuWAfVllv3SKFZF0FpY6PB1bMYbh4l6tHZrQQubsVwcS9RPzu5XFocnXWPGhydSuDgpKmFjXPYgZFNWezsLy0sS2lrr2FIh7rV8awLIjwFs4YFn295NBHLZgXi96/90aRFPtr0yIElkFNQjNVH4rH1dDI6h3mhS5hXvSygKdkUnMk4EZel+sYLglD3mp+tuuXgu3dDcOG0M/73ZhP0GpKOWx9PVF/yglAeWibjzjsiLsoJiRcckBDtiMQYR6Ql2Ffq1tbHxb0YXv5FyrjipeqRF8HDtxiefPQphrtPEdw8i+HiVlKrmrOmhghPwewZcHMaLpxywq7VXlg4PQSTPo+EX4jlNAXIyi/CtjPJ2HUuBW2CPdA51Av+7k5wtLettkh9UlY+DsWk42R8plg4BeEyCYooxFOfRGH1t35Y96Mvdq/1wsl9brhzUjw6XCFNF6zZgkmBGXPGWSWhxZ53Us+z06uWV85uxfAPKYRvcCH8Qi4t/NknsEh10XJ2tewJW4SnYBEWiVFPJagkgOgTlpNsVJ6iEg1HL2aohbg72cPb1QE+ro7wdHFAVn6h6q9OCyetpYIgNHzi0Q0PJKP9Fdn4cWawsmB9/X+hqgD9yEfF+mnp5OfaIOaME6KOOyuhGXPaSVkyS0psKnWD+4YUIrhpAQLDChAQVoDA8AJVOcHdu9js8xFMXnjOmjULM2bMQFxcHLp27YrPPvsMV1xxRWOcWrASGKfywCv/JRstmx2I0ZPiYcnQEsrlQmrN/aUF80fmUdOhecc8PDc7En/9zw+blvoob8uJPW6485l4dOgj1k9LoKQEiI90xPmjLog64ayWuEhH1eGqPO5eRQhtla9CvYKb5SOk2SWRaWnGD7MRnj/99BOeffZZzJkzB3369MHHH3+MYcOG4cSJEwgMDDT06QUrgvEw906LxdypYdjxlxdadc1Bz8EGKkQmCI2IzKOmB0XFLROSVPe0Uuvny6HoenUmRj6WIJnvZmjNpMg8f9RZPUYed1Ylicrj6VeEiDZ5qtEAxWZoqzx4+lq3BbOu2GgGbpfCSbJ37974/PPP1fOSkhKEh4fjySefxNSpU6t9bUZGBry8vJCeng5Pz9rXNfxs3SnllhSsk5UL/VQcFrP42Ic5MExqVVoTd/YOR6i3S633r+8805gYYx79YWcU4tLzLnvs1gBrI3LeofWTblcn12LcMC4Z/W9KM8vkD2sgO8MW54644OyhS8uFU84VXObMBm/aLg9N2+cqoRneNh/e/paTP1Adg9vVLau9LvOMQS2eBQUF2LNnD6ZNm1a6ztbWFkOGDMG2bdsMeWrBirluTDLOHHLBmQOuWPhWCJ7+NFq54gXBHJF51DysnzePT1JtNxd/HIioEy5Y9kUgdq3xxB1Px6tSb4JxYUceCszT+11x+oALLp51qpBd7hNUiOYdctG8Yy6adcxTbnM7uXFocAwqPJOSklBcXIygoKAy6/n8+PHjFfbPz89Xi76CFoS6QgvDvVNj8cGjTVW85/I5ASr5SBDMEZlHzYfQlvkqsXH7X174Y56/sqJ9/GQEeg3NwPXjkq3GWmYKFBbY4PwRZ5zc54pT+1wRfcq5QnxmYHg+WnTORcvOueqRWeUNQfTJQ/j9qxm46ZHnEd6mc4Mc05Iwqaz2d955B6+//rqxhyFYAIyvGjMlDl++FKr6LbfskoPu10jBZ8HykXnU+De+/W5MR+f+WVg+NwB7//ZUyUcHNnrgmjtSVR94JxfxwBgiGYhWzJN7XdVy9rALigrKlpzzb1Kg6rG27pqLll1zVGymIdi1ZjlOH9iB3WuXi/BsbOHp7+8POzs7xMeXzS7m8+Dg4Ar705XEAHr9O3XGMQlCfVveXXtXCtb+4IefPw5S7i7/JhLvKZgXMo+aJyz0fe/UONXf/be5ASphhbHn2//0wvXjkpQVVNy4l0d6sh1O7nHDiT2XxGZWudqZnr5FqqFIa4rN7jkqAdVQpMTHIDs9FTY2Nti/cYVat2/Dn+g99FYwlcbNywe+QaEGO785YVDh6ejoiJ49e2LdunUYOXJkaVA8nz/xxBMV9ndyclKLIDQUw+5PvhQ8ftgV374TjCc/ioadSdn5BaF6ZB41b5q1z1PzzoF/3PHH1/5IiXPETx8GqyL0Q+5JQc/BGTIn1aFY+7lDLji+xxUndrup2s3lk4FYzYRik0tQREGjZZu/dd/gCuuy0pLx4cTbSp9/uPpE4wzGxDH4x5133mPHjkWvXr1UzTmWAcnOzsYDDzxg6FMLgrIojJkah5mPNlUB/6sW+akC0IJlwhod587aILQHLAqZR82P8nF+3a7OQqcrs/HHvFxs/i0CSRe9VRmm1d/6YsjdKeg1JEMVqBfK/j2zCxAtmqyTyqRRffc5i7SHtclH2x7ZaNszR2WfG+s9HDNlBn6YORUlxRXd97Z2drh78rtGGZdVCs/Ro0cjMTERr7zyiip83K1bN6xcubJCoLwgGCoomwHjLOz8v7eaKCsD74RbdZWi65ZGVrotFn8ShJf3OeHgAaBlS1gMMo+aH5XF+dk7aigpmYmS4qVo3vErJMbcoSygP38UjDXf+aH/zWnoMzwdbp4lVt3jnG7zS+5zN2SklHOf+xWhXS8KzWzlPnf3Mo33que1NyMoomUZC6eOZz5djLDWHY0yLlOkUQz8dAdV5hIShMYKyu56dRb6XJ+uCst/914wJs+JtOrJ3dI4usMNP30YhMxUe9jba2CVIUsSnkTmUdOnqji/tj36IycrAy4env+uz0ZizJN44FU3HNzcFHvWtkNqghP++DpA1QPtfk0mBtycZhVlmPJybHD20KUYzVP7XRF7rqz73MGpRGWdt+mRjXa9chDUtPHc5/WFv3/GdeoehbJIZIlgdBorKHvkowkq3jPxgqNKNhr3cqzJT2BCzd1GWC5r+1+XCh0HNc3H/xZqGHa1s7GHJlghVcX5ff3Ko5Wu/2zSLf8+c8adk85hy+/eiDntrLLguUS0zUXv6zJUdyQmK1lMh6AjLsptfvqAq2pFWVJcdiIObZmHNj1zlFWT7UnNpQ6zu7cfPHz84R0Qgj7DR2HHyiVIS4xV64VG7Fx0OUjnIuvg2eva1rhPQwVlXzjlhE+ejkBxkQ3ueCYefW9Ib5DjCo3PuSPO+P79YCTHOqpYr6tvS1Xxu2P6h1lc56LLQToXNR571v1WZZxfVeji/+iq5bdx5DFnbP7NGwc2eah5itjYamjdNQfdrslUZZrMyVtD1zmv6fwxF5w56ILokxWFpl9Igco814VBuXubr8guKiiAnYNDqbWzuLAQ9o6OMDcGm2vnIkEwtaDssNb5GPFgEn77MgC/zg5Ai065KvNRMB/45bxpqTd+/ypAtbjzCSzE3c/HSdyuYNJxfpznvnvv+Wrj/+iBadYhD806xOGWCYnYvZaueQ8l1k7uc1PLkk81NOuQW1oiKKJtnskkJdGayVqaMWecLvU7P+asbgzLww5BLTvnoGWXXJWF7hdiOYX19UUmxac5ik5DI8JTsLqgbFrGju++FLjOEktPfxKtgv4F84gHYykaWoNI90EZqiuVi5v5WIAE66CqOL/axv/RtT5IFZxPRdJFB+zf6K4+9zFnnFVMJJdViy6VEOINNDO6Q5oXoEmLfPgGF8K2bO30BqWowAZJsQ5IvOCAhGhHJTQ5rqQYhwptKOmN4M190w55aNExVzXz8A22HKEp1B0RnoLVBWVzQqaFbOaEZmqy/Ot/frjpkSRYK+bS3i0+yhHzXw9BQrQT7Ow13DwhUSVgSJyuYA5xfgGhzeod/8fGF0PuTlULRShbQDIR59R+F2Sn2+P4bje16KAYDWmej4DQQpUF7uVXVPro6lmsYibtHTR1w81HisPCAlsU5tmgIN8Whfk2yMuxVRnlXDJT7NRjWqI9kmIckZpgX0Fg6hdtD22VpyyxrGEa0T5PbgyFMojwFKwyKJstNe+cFIf5r4di/WJftO2VjTbdrdNVaw7t3Wjt+fGDYBTk2cLLvxBj/y9WuSQFwdTwDgjGy4vWl8b59R0xujTOr6r1dYEi1L9JOvqOSFdtIuMiHXF6v6uyOsaedVLP+XcSeYyu7trHOtcVJ9diBIYVwj+0AE2a5yO0Vb7qVW8pSVCC4RDhKZj8ZG0oOvfPxpU3pGH7n974YYZ1lVgyl/ZuNHqzwPaqRf7qOePB7nsxVr7cBLOM82vo+D96b5rQvd78vzh1hsrTKnnxrCNS4h2QkWyvlvR/l9xMWxQV2qCosKIvnklMjk6aKmHk5FKi/s48fItUT3NaMk/vX4ZT+xeiz/CuuHPSU1bnbTAX75CpI8JTsOqgbAbwnznoqkosLf44CGOtpMSSObR3KyywwU8fBGHv+ksZkteMSsGIh5Kkv7UgVAP/PhhTWVPSJG/qiosYr2mrfqbYZOvO8vOf/k3qP79O5Roc2X4MMaevNamb1MbAHLxD5oAIT8GqcXLRcO/UWFVi6eBmD+xanY0rhmXA0jH19m5ZaXaY/3oTnDviAls7DaOeiseV11v+70UQGgsKTGbD2zuUmP1NqiExF++QOSHCU7B62B3k+nFJWDEvAEtnBaJ5p1zkZe+1aJeKKbd3i49ywNcvh6oyLM5uxRj3ykWrjb8VBGNj6jephsbahbchMGDBBUEwHwaNSkXLrjkqKP/bd0KwY9UfpS4VS4d38vqPxoSdpT59JkKJThaVfvqTKBGdgmDkm1TejFYG13O7pQtvCuzK4HpuF+qGCE9BUBMIMOLB/XByLVDFmnet7l7qUrlw6giiTx5WLhdLrCQQ1roTRj31unrkc2O1dzu8zQ1zpoUiN8tOFch++tMoBEUUGmUsgiCY9k1qY2HtwtsQiKtdEP7l06f7s6M7gGUozH8KwG/ISltnsS4VY1QSqIodKz3x88dB0Eps0PHKLNz3UqzKrhUEwfhID/LGqzNtDYjwFIQKsUxzADwKYBGALgCSLDaWydjt3Th3//2zj4qvJVcMS8cdz8RL5rogmBCmdJNqDER4NywiPAWhQsLNGABXAWCCzXwANxk94cYSYfHr378MwMalPur54NEpGPFgklWUsxIEc8PYN6nGxNqFd0MjwlMQKsBklnvoAAZwI4AnjT0gi6OkGPjpoyDsWu2lnt88PgHXjEoz9rCEmsrv2NqgqERTlmpBsCasWXg3NCI8BaFSl4oGv5D12L/xegAzkJV+2NhDsxiKCoHv3gvBgU0eqlPKXc/FoffQTGMPSyjH1W0CQOOzs4MdnB1s4WxvB1vbS+boouISFBZrKCwpQUFRCdJzC5GWU4DU7EKk8jGnANn50l1KEISKiPAUhCpcKoANCvKycHSHO379ojOe+TwSzq5i6rkcCvJtsPCtEPWe2tlrqv1llwFZxh6WUAmh3lX3+ba3s4W9HeCCS8G4/u5OFfahEL2QmosLqTnqMTOvyKDjFQTBPBDhKQjl0Heh3DU5Dh881hQJ/7bUvHdanMQg1pP8XBvMeyUUpw+4wt6xBA+8ehHte+cYe1iCgfB2dVRLp1CvUiF6OiELJ+OzEJ+RZ+zhCYJgJKSOpyBUg7tXCe5/KVa1bdy3wRNb/7j0JSrUjdwsW8ydFqZEp5NLCca/HSOi08qgCO3VzBf39InAA/2boX8rfwR4VLSUCoJg2YjwFIQaaN4xDzc+lKR+/nVOAKJPypdlXchKt8UXL4Th/FEXuHgU49H3LqBVF+lGZO0i9Irmvrj3yqZKiHZs4gkHO3ElCII1IK52oVLsbG2UNcLXzRH+7o7wdbv0s7uTPUo0DcX/Zrby59zCYsSl5yEmLRexablIzbG8bjMDb09V7RwPb3PH/94KwXNfRMHFvcTYwzJ5MlPtMHtKGOLOO8HdqwgT3r2A0JYFxh6WYEIEeTrjuo7BKpnpyMV0HLyQjjQLnEMEQbiECE+hFFocmvq5oVWgO5r7u6ls1sqwgw30N7k52avkAl0sV25BMaJScrA/OhUX0ywjlotxnYz3/HBiBFLiHPHDzGAVoyjxnlWTlmivRGfiBUd4+hbhsfcvIChCRKdQOZxvejb1RY8IH5xJzMLOc6kSC1oLok8ewu9fzcBNjzyP8DadjT0cQagREZ5WDoVTiwB3dAjxUKLTwe7yoy9cHO3QNthDLRfTcrEnMlV9kZh77T9XjxKM/b9YfDopHIe3umPjLz64ZlSqsYdlkqTE2Sv3OkW6T2Chcq8HhIoVS6gZ1khsFeihlvNJ2dh5PgUxqRKaURW71izH6QM7sHvtchGeglkgwtNKcbS3RYcmnuge7q3irQxFE28XtTCjdff5VBy+mG7WAjS8TT5GTkjEL58H4fev/RHSLB9te0mSjD6JMQ6Y/UIY0hId4BdSoCydvkFSSkeoO8383dTCkkzbz6YgOkX+1khKfAyy01OVSN+/cYVat2/Dn+g99FbVQ9zNywe+QaHGHqYgVIoITyvDw9ke3SO80bGJV5WudENAcTukQ5Cygq4+Go+MXPO1fvW7KR1RJ51V153/TQ/BM59GITDcfK+nIYk954g508KQmWKPwPB8PPpeDLz9RXQKl0eYjytG9XRFZHI2Np9OQkJGPqyZt+4bXGFdVloyPpx4W+nzD1efaORRCULtkKx2K8HNyQ7XtA3AuH7NVBxVY4pOfcJ9XXHvlRHo/G88qLmGJ9zxVAKadchFXrYd5r0aipxM+VNitv+syeFKdIa0yMfEmRdEdAoNCsOB7rkiAjd0DoG3K5s8WCdjpsyArV3lczjXc7sgmCrybWnhUGAOaO2PB/o3R/cIH9VxxNg42dsp6+et3UOVBdYcsXfUVHKRd0ChSp5Z9HYIiq24QyAz/hnTmZNph4h2uXj8/Wh4+FjxGyIYDLqX6TkZ27cZBrULNNpNtDHpee3NeObTxZVu43puFwRTxfgqRDBYhnqfFr54cEAz9G7m2yBJQw0NY7fG9GmKJt7OMEcorB58/SIcnUpwYo8bfv8yANbIid2umPtiKPJz7NCyaw4effcC3Dyl1JRgWNg3vlu4typG3y3CG7ZWWmKCQlz/URBMHdNTI8JlwbmHxZjH9W+Ofi39lXXRlGEG/G09wtAiwA3mSFirfNz9fJz6edMyH2z/yxPWxMHN7vj61SYozLdFu97ZeOStGOlnLzQqtHgOahuoQnia+bvCWnD39oOHjz/CWnfCqKdeV498zvWCYMqYp59TqJQIX1dc1cYfgR7mZUGkNfamLk3w9/EEHIpJh7nR9eosXHdvMlZ/64clnwTBw7sYHftmw9Jh+9BfPg+EVmKDrldlYszUWNhbb9idYGT83J1wa/cwnE3MwoYTiUg34wTG2uAdEIyXF62HnYODsnb2HTEaxYWFsHc0XJUSQWgIxOJpAbCz0Mjuobi9Z5jZiU59txnjPvu2NM+7dQrPntdmoKTERmW6nz7oAkuF5bBWLvTDkk+DlOi88vo03PuiiE7BNGBd4vv7NlWhRva2lu1+psjUd7WL6BTMARGeZp6pPqR9kIqTZKchS+DKFn7qmswtXMnWFrjruTh0vDILRQW2mPdKE4vs6c4EqsUfByrrrk5w3/FMAqpIsBUEo8AkSoYa3dfXcuZGQbAURHiaceLQuH7N0TnMS1kLLQleE/s2mxt29sD9/xerEmyYaPPlS6GIj7IcM2BBvg0WvNEE2//yho2thlFPxWP4/clmd5MgWA+sH0xv0E1dQ8y2goYgWBoGE57Tp09Hv3794OrqCm9vb0Odxqqws7VB13Cv0sQhdh+yVNiv+YrmvjA3HBw1PPR6DMLb5CE73R5zp4YhJd78v/Ayku3wxfNhOLLNHfYOl1qH9rvR/OJxzRGZSy8ftt+k9ZNZ8HKjJAjGxWDKpaCgAHfccQcee+wxQ53CamCZEGaqj+3XDIPbBcHdyfyFTG3o38rfLAvNM6v7kekXENQ0H2lJDpg9JQzJseb7O4s64YQPn2iKqOMucHEvxoR3YtBlQJaxh2U1yFzaMLDCB+t+3tU7AgEelhcGIwiwduH5+uuvY9KkSejcubOhTmHx8M6chZJ5p35dx2B4uViO27a2XNs+EK2D3GFuuHuVKIHmG1yA5IuO+HRSBGLONPyXXfTJQ/ji+fvVoyHYs84Dnz8bjoxkewRF5OOZz6LQskuuQc4lVI7MpQ1LsJez6n50dRt/FbYkCELjYlK+2vz8fGRkZJRZrBFOhnQJsb0lW8P5ullvpiIzNYd3DFatNs0Ntot88qNo1T6SbSQ/fy4Mp/Y1bLb7rjXLcfrADuxeu7xBj1tSDPz+lT++ey8ERYW26NAnC09/Eo2AUMsuUWMJyDxaM4yLZ+vg+65spsrQCYJgpcLznXfegZeXV+kSHh4Oa8tS79fSDw9f1UK5hBgYL1zKUGVygJ+7+b0fXn7FeOKD6P8Sjv4vFPs3Xp4FNyU+BtEnD+PCqSPYv3GFWrdvw5/qOddz++XGc371f6FYv/hSjO21dyXjwdcuwtlNuhGZA9Y+j9YFL1cHVYbuuo5BVtl6UxBMXnhOnTpVWaCqW44fP17vwUybNg3p6emlS3R0NKwhfpNde0Z0CcGD/ZujTws/mQCriM9ikXlzTKhycSvB+Okxqsh6caGt6uu+aZm3qodZH966bzA+euJ2fDjxNmSlpah1WWnJ6jnXc/vldCKaMaGZagFq71iCe6fFYsSDybCVj2SDYsi51Brn0culYxMvjO3XFO2CPYw9FEGweOqU8fDcc89h3Lhx1e7TokWLeg/GyclJLdaAv4cTOoR4qonOzUqShS4XHzdHDO8UjN8PXKy3aDNmtvt9L8bCfXYxtvzmjV9nB+LcYReMejq+zn3Nx0yZgR9mTkUJi2qWw9bODndPfrfO48vLsVFj2rnqUjJXaMs8jJkSh+BmBXU+lmDcudSa5tGGxNXRHtd3DlFx9eyilplXZOwhCYJFUifFExAQoBahfnGbjFNs6ueGZn6u4kavJy0D3HFFM1/sOHfJ0mdO0Gp428QE+AUX4o95/jjwjwfOHXXB3ZPj0LZnTq2P0/PamxEU0VJZOMvzzKeLEda6Y53GdfawM36YEYzkWHZB0TDozlQMvy8Z9o5mpu7NCJlLTbvzUaiPC7acTsLBC+lmd5MrCKaOwUxtUVFRSElJUY/FxcXYv3+/Wt+qVSu4u5tflnJdcXW0Q5CnMwI9nBDm44om3s4qVlG4fNhWMyEzH+eSss2yUsE1o1LRsksOvns3BAkXHDF3WhiuvjUVIx5KUpbRuh3PBpqmlT7WheRYB6z4xg/7N3qq5z5Bhbjn+TjJWjcxrH0uNVZoD0vXtQnywLpjCUjJFsu/IJi88HzllVfwv//9r/R59+7d1eP69etxzTXXwBKgiGBNTU8XB1XqyNvFQbnQKTY9nK2v9FGjZrp3CsYPO6OQlmOeWdbhbfLx7BeR+P2rAGz53Rublvng5F5XjHgwCR2uzK6xyLW7tx88fPzhHRCCPsNHYcfKJUhLjFXrayI7wxZrv/fD5t+8UVzEeEINVwzLwM0TElU8qmBaWMNcaqrQaDCmTwR2nkvB7shUFJeI+VMQLhcbra5mkkaEZUCYlckAeU/PS1aZ2vDZulMouswJwsnBVolKxv24O9mpR8ZiXlpnpx7Zgk2smMYjMTMfP+2KQmGxyX6Ea8XRHW748YMgZKXZl8ZXDh2Tgk79slQP+KooKiiAnYNDqbWzuLAQ9o5Vh3BkpdlhxypP/P2TL3KzLmULtemejZvGJyK0peVYdO7sHY5QbxeDzzPmgqVfX2PON2uPxSMuPc/YQxEEgzO4XSC6hnsbZJ6x2qwWe1sb+Lo7wtvFUVkrlcXS1QGezg6qrJEIStOH3UfY052uMHOmQ59svPDVeWxY4ovNy70Rc8ZZ9UQPbpaPIXenoOOVWXByqSiu9UUmxWdlorOoEDi20w27Vnvh6E43lBRfMqXy2Dc9koh2vXKkhaAg1HK+uat3OPZFp2HbmWQUFIl3QBDqg1UIT5Ys4qQR5MnFGYGeTvB3c1JFhAXzpkuYNyKTc3A6wbxbOLLT0Y0PJWHQHSnYtNQH//zqjbjzTvj2nRDY2Wto1iEXbbrnoE2PHIS1yYNdFeWN8nNtEBfphNhzjrhwyhkH/nFXPeN1hLfNQ/+b0tDr2gwpkSQIdYQ3eD0ifFSS4/rjCWYZZy4IxsZihSfd4cwgb+7vhqZ+rlIb04IZ2iEI8Rl5FlH+hKWVrh+XrBKQNv3qrcobpcY74MxBV7X89T/AyaUErp7FcHQugaOzph7t7DQkXXRASlxFq6enbxF6XpuB3kMzpDySIDQA9JCN7B6Kk/GZ2HAiAdn5FUubCYJgRcJzdO9wZeHk3alg+fCmYljHYPyy94LFlD5xcS/BsHtTcN2YFCUoT+1zxYm9rji931XFZ+bnVh0K4uFbhJCm+UpktumZg7Y9s6u0kAqCUH+Y9c6Wmyy9dChGSi9ZM0wy9nVzgJvKC7mUE8KFRjA7WxvleWWIn60tEzqBomINhcUlKmRDPRaXqBuYrPwiZOcXISuvCJn/PpZY2AfLIoVnoKezsYcgNDKskdqrqS92nTe/+p7VwQmK/dEDQtPR78Z01UM9McYRedm2KMizQUEeH21RWGgD36BCBDctgLu3WF8EoTFvfK9tH4T2IZ5YdzwBSZn5xh6SYOA52c/dCU28nOHv7qQq2fi7O6oSXIaguERDWk4BUnMKSx+Ts/KRnF1gtnHGFik8Beut7xmVkqPc7pYK4zKDIsRdLgimRhNvF4y5IgL7L0jykaVBYcnSWuG+Lgj1doWLY+O5kOxsbZTQ5aIPK5mk5xYiKSsfSVkFquoCF64zdUR4ChYD/0Bv6ByM73ZEyaQvCEKjQzcqk4/aBnngn1OJOBabaewhCfX8LmFJtpaB7mgZ4GaSdbltbGxUB0QurQL/W59XWKwEKJusXBKjeUjJLjQpd70IT8Gi4B/hVa39zb7EkiAI5gtj+4Z3CkGnUC+sP5Eo7nczgPGXzfzd0CrQXSUlm2tCsrODnQo946KjqLik1CqakJlXaiU1loHGIoQn28gVFpq+eVloHNr4O+O8ryMupl1u60cbFJTYoBiSpCYIQj07H10RgSMXM7D1TBJyCiT+2tTiNfk7ahfsoQSnuYrNmmBd8mAvZ7UAXhVc9YmZBUjOzletYVMbwTpq1sKTb1xcXBzS0tKMPRTBxGjpoiHYtuiys0yLNQ0Xc2wRnccyRSJABUGou/u9c5gX2gS7Y9e5VOyLSr3sznrC5eHn7ogOIZ5oG+xhkm50Y7rqmcyUmlMABwM20TFr4akTnYGBgXB1dZXySUIZ8gqKkHU5FgZNQ0F+HhySEgEUIDqvbHC3IAhCbWHW84DW/ugc6oUtZ5JUDVATCruzeGjNbBvsjg4hXv9a/oSq4luZrW9I7M3Zva4TnX5+fsYejmCCOHNuySlA/mXEsTg4XZqgCovjcTFPE7e7IAiXhZerA27oHIJezXxU9vvZROl+ZChoi2Kd1Y5NvFSSkLTCNg3MVnjqYjpp6RSEqqAbpSA7/7IsC45OzrCzsYGjrYbcEhGegiBcPoEezrilW6iKRd96JhnRKTnGHpJFdZbq0MRTLZ5W6ko3ZcxWeOoQ97pQk9vAw8kBGXmXkXxW+hkTv5ggCA1f/3NUzzAlPNkAIzJZBGh9cLS3VQlCjN0M83ERbWDCmL3wFISaYLHf/KLiy3K5C4IgGBJdCZyEjDzsjkzFqfgsk6q9aIpQW4b7uKquURSdFJ+C6SO/JROnWbNm+Pjjj2EpGOt66HKXG2BBEMyh5TNjQMf1a4au4V4ipiohyNMZV7fxx4MDmuP2nmHKpS7vk/kgvykjER0djQcffBBNmjSBo6MjmjZtiqeffhrJycmwZl577TXlIuFib28Pf39/XH311Uqs5ufXrQjzhg0b1HGYhKZzuQuCIJhLEtLgdkF45KoWGNI+SIkta29b2a+lHx7o3wz39IlAz6a+Er9ppojw/Jfdu3dj8ODB6tHQnD17Fr169cKpU6fwww8/4PTp05gzZw7WrVuHvn37IiUlBcasFlBSYlyXdMeOHREbG4uoqCisX78ed9xxB9555x3069cPmZmZl+Vyl7tiQRDMCc5ZrANKscWF5ZgstdC5PvRQsW0lLZu0/t7Xtxn6tPBTdScF80a+hf9l4cKFSuQsWrTI4OeaOHGisnKuXr0aAwcOREREBK6//nqsXbsWMTExeOmll8rsT7F19913w83NDaGhoZg1a1aZIvq0EvIYTk5OyoL61FNPlW6nlXDy5MnqdXx9nz59lCVQx4IFC+Dt7Y3ffvsNHTp0UMf4+uuv4ezsXKEwPy2yFOc6Nm/ejKuuugouLi4IDw9X583O/q80SEJCAm666Sa1vXnz5vjuu+9q9f7Q0hkcHKyupXPnznjyySexceNGHD58GO+9917pfvxdUcB7eHio/e+55x51TnL+/HkMGjRI/ezj46Msn+PGjYOnsz3Wr12Nm4cNRpuIYLRvFop777wN58+erdXYBEEQjAWtnkM6BGH81S0wsnuocjE7OVjO1ziNA22CPHBdxyBMuLol7uwdriybPm4iNi0Jy/nE1oPIyEjs2bMHe/fuxU8//aTW/fjjj+o513N7Q0Nr5qpVq/D4448rQaYPxdOYMWPUWCgodcyYMQNdu3bFvn37MHXqVCUA16xZo7b98ssv+OijjzB37lxlQf3111+VWNPxxBNPYNu2beq6Dh48qKyHw4cPV/vqyMnJUYKOgvPIkSNqDBSjPLa+JZTj4jZy5swZdZzbb79dHZfbKER5Ph0UegwpoKBfsmQJvvjii1JhWFfatWunxPnSpUvLlNR68803ceDAAXXdFJs8J6EQ1o3/xIkTyoL6ySefwM7WFlphHiZMfAqr1m/B4t/+hK2tLR64d7TRLb2CIAi1gaFD7Cc+rGOwEmgUod3CveFrZgKN1txm/q7KqjnmyghMuLoFRnQJUXU3KUIFy8Sqs9qZ6KJDV3ohMTERPXv2LF2vLwAbAgo+HrN9+/aVbuf61NRUNQ4Wxyf9+/dXgpO0adMGW7ZsUWJz6NChyh1NwTpkyBA4ODgoy+cVV1yh9uW2+fPnq0daDwmtnytXrlTr33777VIBR1FIcavjrrvuwvfff4+HHnpIPWcYAC2gFJqErm+K0GeeeUY9b926NT799FNlwZ09e7Y6519//YWdO3eid+/eap958+ZVed21FZ+0EutgjKyOFi1aqPPzXFlZWXB3d4evr6/axveRQlrH3aPvRGp2AQqKLwnNj2bNQccW4Thx/Bjad+hY7/EJgiAYS4RyIZl5hYhKyUF0Si5i03NVP25TSI7nV6yfmyOCvVwQ4uWsrLeM25SyR9aHVQvPb7/9VlnIiorY0/vSX6buke5euqENRV0ELeM+yz/XZYbTgsmfKbxogbzhhhuUe5vjP3TokLJUUqzqQ/e7frcnuv27dOlSZh+KyiuvvBIXL15UopVu8hEjRpQKOFoZaenUd5/zmmg1PHfuHE6ePKnGoC/iKRz1BWBd4fH1JylapRlmwLFQrOsslhS9DBuoTvy//PLL2LZ9B1JSkktfF3MhWoSnIAhmDSt40GLIheQVFiMxMx8JmfnqMSW7QIlRrjeUFdPdyV65xyk0ff995HND9v8WzAerFp4UV7TA6YsjHTt27ECPHj0a/JytWrVS4unYsWO49dZbK2znesYkBgQE1Op4dCnTlcz4ULrf6cKna54xkbT82dnZKYHGR31oEdRBl3/5u05aDlu2bKlc9I899hiWLVtWRojz2BMmTCgTT6qDVlcKz4aG7w1jRQljSYcNG6YWil++XxScfF5QUFDtcSjMWUXgizlz4ekXqITnNVf2RGENrxMEQTA3mIikqxGqD2sbZ+QWqeYa2flFqs5xfmFJac3jwn89Qvrwe8LRzhZO9rZKYPKRPejdnOyU4PVwtreKxCfh8rBq4akP4/woQHSPhoKWRrrI6dqeNGlSmTjPuLg4JaLuv//+MkJw+/btZY7B5/ouax6DYooLE5doWaS1s3v37sriybhKJgHVR5hzPGFhYep9ocVTB0X50aNHlZCuDI6BlmSKXp2rnQK5fMJSbTl+/LgKEZg2bVrpc5aeevfdd5X4JuUrEtCSS/ge6OBrOI6vvvoKAwYMQGpOITZv/qdeYxIEQTBXKBgDPLg4GXsogpVh9XZvxv8xRpJWT5Y04iOf6+IrDcHnn3+u3N20zm3atEkl4FBUUZAy+3z69Oll9mdM5/vvv6+siMxoX7x4sUowIrRCMnaSGd8s08TwAQpRWvToYqd4pJBlUg5d4Iy5ZHzmihUrahwnX8tEK45n1KhRKuNdx5QpU7B161aVTLR//37lvl6+fHlpclHbtm2V659WUVqPKUAffvjhCglVlUHBShFONz8F9GeffaZiR7t164bnn3++1KpKYcltvG5m5TPRSB++BxTwf/zxh4qZpZWW1mSK/y+//FIlSO3Ztgmvvjillr85QRAEQRAuB6sXnrTmMRua4kgnkvic6w0FE3FonWNc5p133qlc2uPHj1flf5iBrkuK0fHcc8+p/WnBfOutt/Dhhx8q0UoYM0nrHROQGKdJl/vvv/9eGsPJJCIKTx6DYnDkyJHYtWuXEm41QWsmE5UYy6nLZtfBc9GdTzFMayrH9sorr5QmMenOzecUjbfddpu6xtoIembWh4SEqDFec801+Pnnn5Wl859//ikNEaBrnaKbIpzxnLR8zpw5s8xxKOJff/11lZgVFBSkRDEttwwfoBDu1KkTJj/3HN597/0axyQIgiAIwuVjozV02nYDkpGRAS8vL6Snp8PT07PMtry8PGXBY8wfa04KQn3hnwBd7pXFNJHCgnzEREViX6oDckskfsnUYe0/Fp5uiHnGErD06xMEwbzmGau3eAoC3fEsLC9FPQRBEATBsIjwFARm2dnZwt1Zcu0EQRAEwZCI8BSEf3FxsJM6c4IgCIJgQORbVhD0Xe4u4nIXBEEQBEMhwlMQ9LC3tVVFkAVBEARBaHhEeApCOVwc7VVHDkEQBEEQGhb5dhWESvB0doCtmfrco08ewhfP368eBUEQBMEqhCeLsD/00EOqzia71bBI+quvvlpjH21BMAVsbVliyQHmyK41y3H6wA7sXrvc2EMRLhOZRwVBsDQMFszGXtrseT537lzVAYctHR955BFkZ2dX6DAjCKaIk4MdXIpKUGgG3/Ep8THITk9VCVL7N15qh7pvw5/oPfRWVSDfzcsHvkGhxh6mUEdkHhUEwdJo1M5FM2bMwOzZs1Vv7dognYsEY1OiaYhPyURU5HmT7lz07HVta9znw9UnYA1YeueihpxHBUEQLLpzEQdUvg+5Pvn5+Wrw+oslMm7cOGWZevTRRytsmzhxotrGfQTjY2tjc6mwvInHe46ZMgO2dpWLYq7ndsEykHlUEARzptGE5+nTp/HZZ59hwoQJVe7zzjvvKMWsW8LDw2Gp8Np+/PFH5ObmlrHifv/994iIiDDq2ISysKi8s71pWjp19Lz2Zjzw2qxKt909+V0EhrdQ7njBvJF5VBAEqxOeU6dOVRa56hbGJekTExOD4cOH44477lDxSVUxbdo0dTevW6Kjo2Gp9OjRQ30hLF26tHQdf6bo7N69e+k6xnfxi0SXXNC1a1csWbKkdHtxcXGZ5IO2bdvik08+KXMuWk9HjhypYsJCQkLg5+enLKuFhYWNdLXmj6O9LZr6u8GUmfdyRQs6+e695/HRE7fjrfsGN/qYhMqReVQQBGulzslFzz33XI1u4BYtWpT+fPHiRQwaNAj9+vXDl19+We3rnJyc1FJfGK2akwOj4OrKzjd1e82DDz6I+fPnY8yYMer5N998gwceeAAbNmwo3Yei89tvv8WcOXPQunVrbNq0Cffeey8CAgIwcOBAJUzDwsKwePFiJSi3bt2K8ePHK4F55513lh5n/fr1ah0faTUZPXo0unXrVu0XmFCW/i39EJeVgLQc0xTstz72EpbNnl6lu52WT8E0MOV5VBAEwWyTi3iHzsmyZ8+eSjzZVRGD1lDJRdnZgLs7jEJWFuBWS4MYv3DS0tLw1VdfKavniROXkj7atWunrBMPP/wwvL29VSYrY7nWrl2Lvn37lr6e23NycpRbvjKeeOIJxMXFlVpGeT6K2TNnzpT+DihKbW1tlbtfqB79z1pGIfDTzmgUlTRaTl6dOH90Pz59ZnSF9c/OWoqw1h1hDVhacpEh51FBEISGoC7zjL0hJ8trrrkGTZs2VS7exMTE0m3BwcGGOq1ZQavliBEjsGDBAlXyhj/7+/uXbqdlkgJz6NChZV7HGn767vhZs2Ypa2lUVJSKGeV2WjP16dixY5kvLFo/Dx2SAuN1JdDDGQPbBmDdsQSYIvYOl2qP0lXLz5TuUTBPZB4VBMHSMJjwXLNmjRJOXOgK1sdQX4R0d9PyaAx47vpAdzstlDoBqU/WvxezYsUKhIaWrcGoc6XRYjl58mR88MEHyirq4eGhyq3s2LGjzP4O/woSHRQkdNMLdadLmDcupuXiWGwmTA13bz94+PjDOyAEfYaPwo6VS5CWGKvWWzoujnZo6usKT1YhsBCMMY8KgiAYEoPN0HTvNnZJIMZY1tbdbSowWYAWSgrBYcOGldnWoUMHJTBpyWQ8Z2Vs2bJFxX09/vjjpevoUhcMy7Xtg5CSXYj4jDyYEt4BwXh50XrYOTioz1TfEaNRXFgIe0dHWBr8ew/ydEZTP1c093dDsKezumZLwhjzqCAIgiGxHNOAmUL397Fjx0p/1ofWS1ozJ02apKyTAwYMUPETFJuMoRg7dqxKOFq4cCFWrVqlYhAXLVqEXbt2qZ8Fw5ZYuqVbE/ywMwqZeUUwJfRFJoWYJYlOBzsbhPu6omWAuxKbbk4yhQmCIJgTMmubANUF4r755psqFpTZ7exUwqQjlmJ68cUX1XbW89u3b5/KUqfIuPvuu5X186+//mrEK7BOKHpGdg/FT7uiUVAkYQuGws3JDi383dEiwA0Rvq6wt2vUvheCIAiCubbMrCvSMlMwBWr6rEUmZ+PXfRdVe02hYfB1c1RWTYrNEC/DutAtPevb0q9PEATjYxJZ7YJgLTT1c8O17QOx5mi8sYdi9vGaFJutAt2V8BQEQRAsDxGegtAAdAr1UoXld51PMfZQzAZbGxuE+rgoodkywA0ezmUrLwiCIAiWhwhPQWgg+rfyQ2FJCfZHpRl7KCadHBTh56aEJq2bzg51K4YuCIIgmDciPAWhgWAc4qC2gbC3tcHu86nGHo7JQHHZ3N9VWTYZlsCKAIIgCIJ1IsJTEBqYq1oHwM7WBjvOWq/b3d3JHi0D3dAqwANhPi6wtbWs+pqCIAhC/RDhKQgGoF9Lf2XZ23wqCdaWiU7BaYnF3AVBEITLR4SnIBiI3s18leVz08lEWGKlJV0m+qXkIMlEFwRBEGpGhKcgGJAeET7wdnHA6qPxyC0ohrlDIU3Xua7GpmSiC4IgCHVBhKcgGJgWAe64p48TVh6KQ0xaLswNR3tb1Q9d16ZSMtEFQRCE+iLppRYMm1KNHz8evr6+Kt5u//79uOaaa/DMM89U+7pmzZrh448/brRxWgOezg4Y1TNMud/NIfTRxdEOHZp44uZuTTD+6ha4sUsTtA/xFNEpCIIgXBYWafH8aM3JRj3fpKFt6vyauLg4TJ8+HStWrEBMTAwCAwPRrVs3JQqvvfbaBhnXypUrsWDBAmzYsAEtWrSAv78/li5dCgcHcY8aA2Z2D2jtr1zV7HKUlV8EU8LTxaG0vmaot2SiC4IgCA2PRQpPU+f8+fPo378/vL29MWPGDHTu3BmFhYVYtWoVJk6ciOPHjzfIec6cOYOQkBD069evdB2tn4JxaebvhnH9m+HghTTsOp9qtNhPWl4DPZyV+5yCM9CzYh96QRAEQWhIxNVuBB5//HHl+t65cyduv/12tGnTBh07dsSzzz6L7du3q32ioqJwyy23wN3dHZ6enrjzzjsRH/9fL/DXXntNWUgXLVqkXONeXl646667kJmZqbaPGzcOTz75pDoOz8V9SHlXe0JCAm666Sa4uLigefPm+O677yqMNy0tDQ8//DACAgLUWAYPHowDBw7UeiykpKQE77//Plq1agUnJydEREQoi6+O6OhodY0U4xTHvHYKdEuFpZZ6NvXFA/2boW9LPxVH2RjwPC0D3TG0QxAevqoF7ukToc4volMQBEFoDER4NjIpKSnKBU7LppubW4XtFF4UaRRe3Hfjxo1Ys2YNzp49i9GjR1ewaP7666/4448/1MJ93333XbXtk08+wRtvvIGwsDDExsZi165dlY6HApWib/369ViyZAm++OILJUb1ueOOO9S6v/76C3v27EGPHj1UOADHV5uxkGnTpqnnL7/8Mo4ePYrvv/8eQUFBahutvcOGDYOHhwf++ecfbNmyRQnu4cOHo6CgAJaMk70drmzhh4cGNFdu+CberH/ZcMdnF6VwX1f0a+mH0b3D8ejAlri5axPVW55F3gVBEAShMZFvnkbm9OnTKumnXbt2Ve6zbt06HDp0COfOnUN4eLhat3DhQmUVpYDs3bu3WkeByhhOCjZy3333qdfSkkirI9fb2dkhODi40vOcPHlSiUlaXnXHnDdvHtq3b1+6z+bNm9V2Ck9aKsnMmTOVyKRQZfJSTWOh5ZNC+PPPP8fYsWPV9pYtW2LAgAHq559++km9/uuvvy4tOj5//nwlwhmfet1118HSYdIOE4+45BQU4VxStlqiU3KRV1hc6z7ofu5OCHB3gr+HE/zdHVUhd3tpUSkIgiCYCCI8GxmKzpo4duyYEpw60Uk6dOighBi36UQi3do6oUcYz1neWlnTeezt7dGzZ8/SdRTEPI8OutSzsrLg5+dX5rW5ubnKyqmjurHwPPn5+VUmTfEcFOT6ryd5eXllzmEtuDrao2MTL7WQgqISZOcXqWSk7IIiFRNKV72zg62ymDrZ28LJwQ6ezvbSLUgQBEEwaUR4NjKtW7dW4qAhEojKZ6fzuLQcNiQUnRSRtDyWR1+gVjcWxo/WdA6K38riSxlXau0wLtPR3hE+0hlIEARBMHPEB9fIMHGG8YyzZs1CdnZ2pYk8dHUz7pKLDsZFchstnw0FrZtFRUUqblPHiRMn1Hl0MJ6TpZ9oGWVikP7C8ky1FdsUn3S9VwbPcerUKVVSqvw5GDIgCIIgCIJlIMLTCFB0FhcX44orrsAvv/yiRBfd0Z9++in69u2LIUOGqBJLY8aMwd69e1WM5f3334+BAweiV69eDTaOtm3bqgSeCRMmYMeOHUqAMntd30LJsXBMI0eOxOrVq1Wm+datW/HSSy9h9+7dtTqPs7MzpkyZghdeeEHFqtJ9zux9xpMSXidFLBOqmFzE2FZaWJ966ilcuHChwa5XEARBEATjIsLTCLCYOwXloEGD8Nxzz6FTp04YOnSosgjOnj1buamXL18OHx8fXH311Ur88TVMwmlomMTTpEkTJWpvu+02lSxEy6MOjuXPP/9U43jggQdU6SeWSoqMjCzNSq8NzGbntb7yyivKossMfV0MqKurKzZt2qRKLHEM3P7QQw+pGE+WbxIEQRAEwTKw0WqT7WIkMjIylKs1PT29ggChKKFljLUnaVETBEMhnzXLprp5xhKw9OsTBMG85hmxeAqCIAiCIAiNgghPQRAEQRAEoVEQ4SkIgiAIgiA0CiI8BUEQBEEQhEZBhKcgCIIgCILQKJi98GzoTj2CUB75jAmCIAiClbfMdHR0hK2tLS5evKjaKvK59KkWGhJWGisoKEBiYqL6rPEzJgiCIAiCFQpPCgHWVYyNjVXiUxAMBQvcs7g9P3OCIAiCIFih8CS0QFEQsN84W1AKQkNjZ2en+tSLNV0QBEEQrFx4EgoCBwcHtQiCIAiCIAimi0F9hzfffLOySLLNYEhICO677z5xiwuCINQBmUcFQbAkDCo8Bw0ahJ9//hknTpzAL7/8gjNnzmDUqFGGPKUgCIJFIfOoIAiWhI3G1N1G4rfffsPIkSORn59fK9d4XZrOC4Ig1Adzm2dkHhUEwdSoyzzTaDGeKSkp+O6779CvX78qJ0tOpFx08AJ0FyQIgmAIdPNLI96D1xuZRwVBMPt5VDMwL7zwgubq6sqRaFdeeaWWlJRU5b6vvvqq2k8WWWSRpbGX6OhozVSReVQWWWSBhcyjdXa1T506Fe+99161+xw7dgzt2rVTPyclJam79MjISLz++uvKFPvHH39UWp6m/J06O8bwtX5+frUuZ0PVHR4ejujoaLN3K1nStVja9ci1WM61cArMzMxEkyZNGq1Wq8yjjYclXYulXY9ci3XOo3UWnuzikpycXO0+LVq0qLTLy4ULF9TFbN26FX379oUhsKR4Jku6Fku7HrkW08RcrkXm0cbDkq7F0q5HrsU6r6XOMZ5sT8nlcnpe69+NC4IgWBsyjwqCYK0YLLlox44d2LVrFwYMGAAfHx9VAuTll19Gy5YtDXaXLgiCYEnIPCoIgqVha8j+1kuXLsW1116Ltm3b4qGHHkKXLl2wceNGODk5Geq06tivvvqqQc/RWFjStVja9ci1mCaWdC1E5tHLx5KuxdKuR67FOq+lUet4CoIgCIIgCNZL46RwCoIgCIIgCFaPCE9BEARBEAShURDhKQiCIAiCIDQKIjwFQRAEQRCERsEsheesWbPQrFkzODs7o0+fPti5c2e1+y9evFh1AOH+nTt3xp9//glzvJavvvoKV111lSqrwmXIkCE1Xrsp/150/Pjjj6qjysiRI2FK1PV60tLSMHHiRISEhKhswDZt2pjMZ62u1/Lxxx+rLGoXFxdVrHzSpEnIy8uDsdm0aRNuuukm1R2Dn5lff/21xtds2LABPXr0UL+TVq1aYcGCBY0yVlNH5lHTnEctbS6VeVTm0QpoZsaPP/6oOTo6at9884125MgR7ZFHHtG8vb21+Pj4SvffsmWLZmdnp73//vva0aNHtf/7v//THBwctEOHDmnmdi333HOPNmvWLG3fvn3asWPHtHHjxmleXl7ahQsXNHO7Fh3nzp3TQkNDtauuukq75ZZbNFOhrteTn5+v9erVS7vhhhu0zZs3q+vasGGDtn//fs3cruW7777TnJyc1COvY9WqVVpISIg2adIkzdj8+eef2ksvvaQtXbpU9QVetmxZtfufPXtW9Th/9tln1d//Z599puaDlStXataMzKOmOY9a2lwq86jMo5VhdsLziiuu0CZOnFj6vLi4WGvSpIn2zjvvVLr/nXfeqY0YMaLMuj59+mgTJkzQzO1aylNUVKR5eHho//vf/zRzvBaOv1+/ftrXX3+tjR071mQmy/pcz+zZs7UWLVpoBQUFmqlR12vhvoMHDy6zjhNO//79NVOiNhPmCy+8oHXs2LHMutGjR2vDhg3TrBmZR01zHrW0uVTmUZlHK8OsXO0FBQXYs2ePco3oYDN6Pt+2bVulr+F6/f3JsGHDqtzflK+lPDk5OSgsLISvry/M8VreeOMNBAYGqqLYpkR9rue3335TnWToIgoKCkKnTp3w9ttvo7i4GOZ2Lf369VOv0bmRzp49q1xdN9xwA8wNU/37NyYyj5rmPGppc6nMozKPNnrLTEOQlJSkPoD8QOrD58ePH6/0NXFxcZXuz/Xmdi3lmTJliorRKP+BMIdr2bx5M+bNm4f9+/fD1KjP9XBS+fvvvzFmzBg1uZw+fRqPP/64+kJjBwhzupZ77rlHvY5tGnlDXFRUhEcffRQvvvgizI2q/v4zMjKQm5urYq+sDZlHTXMetbS5VOZRmUerwqwsnsJ/vPvuuyqQfNmyZSrQ2ZzIzMzEfffdp4L8/f39YQmUlJQoi8OXX36Jnj17YvTo0XjppZcwZ84cmBsMIqeV4YsvvsDevXtVy8YVK1bgzTffNPbQBKFBMed51BLnUplHrQOzsnjyD8vOzg7x8fFl1vN5cHBwpa/h+rrsb8rXomPmzJlqwly7dq3q22xs6notZ86cwfnz51VWnf6EQ+zt7XHixAm0bNkS5vS7YQamg4ODep2O9u3bqztFumkcHR1hLtfy8ssvqy+zhx9+WD1nBnN2djbGjx+vvgToYjIXqvr79/T0tEprJ5F51DTnUUubS2UelXm0KsznygH1oeNd0Lp168r8kfE540Iqg+v19ydr1qypcn9Tvhby/vvvqzumlStXolevXjAF6notLMly6NAh5RrSLTfffDMGDRqkfmbZCXP73fTv31+5hXSTPjl58qSaSI01Wdb3WhjzVn5S1H0RXIpFNx9M9e/fmMg8aprzqKXNpTKPyjxaJZqZwZIGLFGwYMECldY/fvx4VdIgLi5Obb/vvvu0qVOnlikDYm9vr82cOVOVznj11VdNqgxIXa7l3XffVeUclixZosXGxpYumZmZmrldS3lMKROzPtcTFRWlMmOfeOIJ7cSJE9off/yhBQYGam+99ZZmbtfCvxFeyw8//KDKaKxevVpr2bKlymw2NvysswwOF05fH374ofo5MjJSbed18HrKlwF5/vnn1d8/y+hIOSWZR011HrW0uVTmUZlHK8PshCdhDamIiAg1ebDEwfbt20u3DRw4UP3h6fPzzz9rbdq0UfuzJMCKFSs0c7yWpk2bqg9J+YUfcHP8vZjqZFnf69m6dasqMcPJiSVBpk+frsqcmNu1FBYWaq+99pqaJJ2dnbXw8HDt8ccf11JTUzVjs379+kr/BnTj5yOvp/xrunXrpq6dv5f58+cbafSmhcyjpjmPWtpcKvOozKPlseF/VdtDBUEQBEEQBKFhMKsYT0EQBEEQBMF8EeEpCIIgCIIgNAoiPAVBEARBEIRGQYSnIAiCIAiC0CiI8BQEQRAEQRAaBRGegiAIgiAIQqMgwlMQBEEQBEFoFER4CoIgCIIgCI2CCE9BEARBEAShURDhKQiCIAiCIDQKIjwFQRAEQRCERkGEpyAIgiAIgtAoiPAUBEEQBEEQGgURnoIgCIIgCEKjIMJTEARBEARBaBREeAqCIAiCIAiNgghPQRAEQRAEoVEQ4SkIgiAIgiA0CiI8BUEQBEEQhEZBhKcgCIIgCILQKIjwFARBEARBEBoFEZ6CIAiCIAhCoyDCUxAEQRAEQWgURHgKgiAIgiAIjYIIT0EQjM758+dhY2ODmTNnwtK45ppr1KJPfHw8Ro0aBT8/P3XdH3/8MTZs2KB+5qMpMm7cODRr1szYwxAEwcyxN/YABEEwTSiCasP69esrCCuheiZNmoRVq1bh1VdfRXBwMHr16oW4uDhjD0sQBMHgiPAUBKFSFi1aVOb5woULsWbNmgrr27dv38gjMy9Wr15dYd3ff/+NW265BZMnTy5d16ZNG+Tm5sLR0RGmyFdffYWSkhJjD0MQBDNHhKcgCJVy7733lnm+fft2JTzLrxeqpzIhmZCQAG9v7zLrbG1t4ezsDFPFwcHB2EMQBMECkBhPQRDqzfz58zF48GAEBgbCyckJHTp0wOzZsyvst3v3bgwbNgz+/v5wcXFB8+bN8eCDD1Z7bE3TMH78eCXcli5dWu2+mZmZeOaZZ1QMIsfB8QwdOhR79+4t3YfhAJ06dcKePXvQr1+/0nHMmTOnwvHy8/OVG7xVq1bqeOHh4XjhhRfU+vJ8++23uOKKK+Dq6gofHx9cffXVZayc+jGeCxYsUCEMvLZZs2apn3UhDVXFeO7YsQM33HCDOrabmxu6dOmCTz75pNr3Q3eeLVu24Nlnn0VAQIB67a233orExMQK+3/xxRfo2LGjutYmTZpg4sSJSEtLqzHG88cff0TPnj3h4eEBT09PdO7cucLYeBz+bvge8vh8T9977z2xngqClSIWT0EQ6g1FJgXLzTffDHt7e/z+++94/PHHlaigeNFZ96677jolfqZOnaosfUwmqk5MFhcXK2H6008/YdmyZRgxYkS143j00UexZMkSPPHEE0r8JicnY/PmzTh27Bh69OhRul9qaqoScXfeeSfuvvtu/Pzzz3jssceUuNUJYY6d18PXU/gylODQoUP46KOPcPLkSfz666+lx3v99dfx2muvKSH7xhtvqONQKNKVzmsuD0UpQxXuu+8+JYzvv//+aq+LFuYbb7wRISEhePrpp1U8KK/pjz/+UM9r4sknn1SClSKa7zmTmPge8X3VwfHzOoYMGaLeixMnTqjf665du5RwrcrSybHxPbz22muVkCQcG1+jG1tOTg4GDhyImJgYTJgwAREREdi6dSumTZuG2NhYNR5BEKwMTRAEoRZMnDhRKz9l5OTkVNhv2LBhWosWLUqfL1u2TL1u165dVR773Llzap8ZM2ZohYWF2ujRozUXFxdt1apVtRqbl5eXGl91DBw4UJ3jgw8+KF2Xn5+vdevWTQsMDNQKCgrUukWLFmm2trbaP//8U+b1c+bMUa/fsmWLen7q1Cm136233qoVFxeX2bekpKTMebnow+OUH+/69evVej6SoqIirXnz5lrTpk211NTUKo9fGfPnz1fHGjJkSJl9J02apNnZ2WlpaWnqeUJCgubo6Khdd911Za7h888/V6//5ptvSteNHTtWjUXH008/rXl6eqpxVsWbb76pubm5aSdPniyzfurUqWocUVFR1V6HIAiWh7jaBUGoN3RX60hPT0dSUpKycJ09e1Y9J7pYRlrpCgsLqz1eQUEB7rjjDrXvn3/+WanVsDJ4DloaL168WO1+tMrS8qaDFko+p1WWLniyePFiZeVs166duh7dwpACXRY/oeWT1tFXXnlFxWfWpyJAdezbtw/nzp1Tbury8aC1PT4ttvr7XnXVVcqaHBkZqZ6vXbtWvec8h/41PPLII8p1vmLFiiqPzTFlZ2cry2dV8L3kOWl11X8vaV3lODZt2lSr6xAEwXIQ4SlYNIzJowuVLj5+kV555ZXYtm2bsYdlMdCtShHB+EEKEbrTX3zxRbVNJzwpRG+//XblzmWMJ7O5GRtaWbzkO++8owQd3eblSzRRqLDkkP5C0UTef/99HD58WMURMt6S7mOK3/IwfpFj1YfZ5ISuaHLq1CkcOXJEXYv+otuPIpWcOXNGiTW69g0Bj08Yl1pf+LnXhwJQF3JAdAK0bdu2ZfajIG/RokXp9spgSAXfk+uvvx5hYWHq72zlypVl9uF7yXXl30t+ZvTfS8G0kXlUaEgkxlOwaIqKilRCBOP1+OXImL6bbrpJiQx3d3djD8+soTBifB8tgx9++KESfRQstFQyHlKXPEKLG4Uks+IZA8r6lfwS++CDD9Q6/d8DE5AoVCgkKTz1s7yjo6NVMlBlNUQZs0nLGuNBmdgzY8YMFXfIOFIKo7rAcTNJhtdUGbxOc8HOzq7S9Ze8/ZcHE7j279+vfp9//fWXWnhDwbjV//3vf6XvJWNZmZhVGToxL5g2Mo8KDYqxff2C6fPqq6+qeK/ExESjnP+9997T2rZtWyGOrr6EhIRou3fv1qyJ2bNna+Hh4VpeXl6DxXh+9NFH6nlkZGSZ/V588UW1nnGbVfHdd9+pfb766qsKMZ5r1qzRnJyctBtvvFHFe+rIzc1V2/SXlJSUSo8fHx+vhYaGav379y9dxzhLe3t7LSsrq8J7w3Nv27ZNPb/hhhvUa2uKo+RY+bp9+/ZVu199YzwZE8vnfJ/rii7Gs3xcbflzfP/99+r5n3/+WWY/xr4ybvb222+vMsazPPz7nDBhgjoe419Jhw4dtL59+9Z5/JaIzKPmT0PMo4LEeJoVutIrNS2m0nIvKytLZdMOHz4cvr6+amws81IXMjIylOVqypQpFeLo6gNdfykpKaqki6FdUxwzXbuMg+zTp0+1sXD60M3LOEe6Olmih+5pZkPTWljfc7EUDt3Sc+fORUNb0/StZ3Sv0+qlD9265S1s3bp1Kx17eeiGZZkeWj6Z/a2znNL6yW36C13HdMHr3Pr61ji+H+WPT8uN/nuge0/o/mVZIELrKbOwWTC9PCzwzrhGMnLkSPWZZDZ7+dJADWFRZDY+LbzM/C5f2qghjk/4HtJK/emnn5Y55rx589R7Wl01AVYO0IfvBUs9Ed37zveSLllaRcvDa+Lvo7Ext3mU1QVYiYDVIxgmQnc331dWWKgtMo+a7jxqjYir3Ywwt04yTCLglzInyq5du9ZrIv/mm2/UlxPLtlwuFA0sfs5SLl5eXjAknKDoXmbSRuvWrZXgZhkfuoYHDBhQ7WsZV8e6lGPHjlWTIEvS/PLLL6rEDyc8JozU9VwUbTwe3ccssdMQyS9M/KFoocuNCTq80aBYo+hjqRwddLuyTiRrSLZs2VJdG/djrBjHWRkUdTq3LferbqLn8ej+Y+9zfs7o+mPSDL+w6c7Xh+8nv4DpIqSbl2WF6C7+8ssvS8sGUezSlcgSTXwP+/fvr8Tt8ePH1XqKKLa45JfuSy+9hDfffFO5+W+77TZVp5Ln5XkYr3o5UCCwrBHfXwr1Bx54QJVV4jj4pVqZmKsrFNz8e2D8LW8Q+RljOSX+vnr37l1ts4CHH35YiQ8mXfH95+f2s88+U2PVzUHPP/88fvvtN1USip9TinsKd5an4meWvwcKgsbE3OZRfl4ZS00RRWHP2ObPP/9c3ZgwVKU2McAyj5ruPGqVGNvkKjRseRtTchHRHREbG1vGbUgXYF3o0qWLdu+992qXC0vljBgxQrvnnntqdKFeLjt27Ch1G+u7iVu2bFlvtyNL1nTt2lW5yup7LrrFuO+6desa7PP222+/qd+Rs7Oz1qxZM+XOYwkefVf73r17tbvvvluLiIhQLnSWLqIbXd9Np+9q1+eLL75Q6ydPnlzluOgWfv7559X74+Hhocr38Ge+Vh+6uzt27KjOy/eGY6brmKWDKvu88Fq4P8fs4+Oj9ezZU3v99de19PT0Mvvyert37166H8/DMIDLdbXr2Lx5szZ06NDSa+P7/dlnn2kN4WrXwfegXbt2moODgxYUFKQ99thjFUo4lXe1L1myRJVh4u+TJZn4+6WrXfc3ryMzM1ObNm2a1qpVK7Wfv7+/1q9fP23mzJmlJayMianPoyzfxc+4PixPxc/bmDFjanUMmUdNZx4VNE2EpxlT3YR5/vx59eXRpk0b9QXr6+urjRo1qkLcXUZGhqrHxy8UfikEBASo2n979uypdsLk8flHyS/muLi4GsdaH+F59uxZ9ZoFCxZUuc/cuXPVlz5rPnJf/YU1EAljmlgXsnzMoKGgCGKNwvIC5e2331bjqm/tQo6fouByzsXPwVNPPaVZIzrhKQjmOo/q06NHD7XUhMyjZZF51PiIq91CobuPHULuuusu5QajS4tuO2YAHz16VMW81KXjS/lsZrrXGLdJF5WhXGUcP6lqHJMmTVLxb3T50g154cIFlU3NWpF07eli9ugGpuuXrknWcawOvrZ8vGBV8Pori5di/UW6ceki1odlfghdu7XJjKZLkm4tjofuSmYNjx49+rLOxfeSbjtBEMx3HqXRKD4+XsV91oTMozKPmhzGVr6CYe7UK+sow6xd7r9w4cI6dXzRv1M/duyY1qRJE613795VZhQ3lMXz//7v/9Rr6Korz6ZNm9Q2WiP0oSuU63fu3FlqUeBzWivoptQtfH1l6FyRtVmqytqm9WLw4MEV1h85ckS9jh1waoMuQ5gLO+TQ0lL+Pa/rucaPH6+sGtaIWDwFc59HdbC7Fo81b968GveVeVTmUVNDLJ5W0FGGd5/MamQyBIt87927VyVQlO/4wgDs6mCBbt4p8ji8ayx/d9jQ0GrAO+vK6sTxjpx3yqzXqA+LlRNmfDI5omnTpnXKAGZySm2zJtk3uzJ4d80kk/LoalJye21gkDsTZvi7YVILE1x0BdPrey5mgXMdA+111hpBEMxnHmVy2cSJE9G3b1+V6FITMo/KPGpqiPC0UPhHwaxaZgazNEz5kjc6WKibkxfdCHSpMIuPmcQsQVEeZtcGBQUpV4sxiwYzO5OTGjvglO9Co5tQ6iuKOaHouqpczpdVZWWC8vLySrfXBhZm50L4O6ErjL8DfsHpsinrei7d58AaszFNpTyOYD6Y2jzKjHaWuGI2OV37VTUIqA0yj8o8aiykjqeFwlIP06dPV/XeeJfHbi6cZPz8/MrUHOR2thZkGRTeqfPOl3FDvBMvD9seMi7pu+++a5Rr4Fg5ObIkhj6Ms2LpnsrKiOj6bde3FAon3PJtGataeOdcGSx5o19OSIduXU0WkargXTtjzvTr99X1XKypyTv02k7agmDNmNI8SqHLLlysf8oat7WdR2QeLYvMo8ZHLJ4WCu+GeQeuX8eQd2/lC1Hr/ujYd5kLeyczcJqTbflWg5xM6bLhfh4eHrjnnnsMeg26u9Rz586VFqYmugmUNSTL34UuXrxYTfj1LWzMQPxBgwbVal+Oi23kysM6hqz9RrecvsWAd9i67fVB5+7Rt7TU9Vwcs6nUJxQEU8dU5lGek1Y6iiXWqGUCU22RebQsMo8aHxGeFgpdMOVjcng3rn93yZ95x6tfBLiqji86twILbeuK8tJNxGK8hoIxTGT37t1lJkwWpCecgJ999tnS9czMZNzVt99+W+9zNkRsEu+oZ86cqd6ryZMnq3V8P+muYzcM/exIxghFRUWpjFZdViu/tPh70IfxZSx0zTts/S+dupyL8P0ZM2ZMrd8PQbBmTGEe5fEZE8oOUMuXLy+dF2uLzKP/IfOoaSDC00JhGQx24uBkyD8wTlqcYOh2qU/HFx0se8EJiZ1l6F76888/VUmQqmCHDVoHGNhN2K6M5Tp0bqzqOl8wPopuII7pwQcfLF3Pa+D5f/31V/XHz84yLF3yww8/qG4qugmBXVU48fAuVhcAzk417MqyceNGg8UmcaJilxF29uDkR6sBu/fQtcVWhPrs3LlTWQbYWvS1114rLVvCMbO9W2hoqHJH0S3HpAL+XvTjwupyLrrP+J4wpksQBPOYR5977jlVBogWT/79lheE1XWXIjKPyjxqchg7rV4wTBkQdh154IEHVJcQd3d3bdiwYdrx48dVgWN2IKlLx5fKCh+zzAjL0/DY27dvr3KMPF9dy2jo8+GHH6pzlC9rwusbN26c6hTDDh4sflxZaRF202H5E8KixywEzU4whoZdL9htJzg4WI2PZVNWrlxZZdkRvsc6fvjhB1V8mkWO7e3t1TXy+fLlyy/rXFOmTFHdZQzdcUQQzAlTn0e5vbpyRLVB5lGZR00JG/5nbPErCFXBOBzesTNr9KGHHqrz63kXS0vCY489hq+//hpLly5V1gVrg24jxlFNnToVTz/9tLGHIwhCIyLzaMMg82jDIFntgklDF9cLL7ygAvL1s0hrS69evVRXCk4Yb775Jt566y1YI4xVcnBwUB1WBEGwLmQebRhkHm0YxOIpWDTr1q1Td6eMV2L8ErNUBUEQhNoj86jQkIjwFCwaJjYxu5R3/Jw861KGRBAEQZB5VGhYJKtdsGjYyo4TZr9+/WSyFARBqAcyjwoNiVg8BYuG9fVat26NLVu2VNq+ThAEQagemUcFs0kumj17tipYy04AXFjItrIWYoJgKF5++WWVkSmTpWCuyDwqGBuZRwWzEZ4sqvvuu++qgqvsmsByDCy6euTIEUOeVhCwf/9+FY908OBBvP3228YejiDUG5lHBWMh86hgEa52X19fVdKhPrXEBEEQBJlHBUEwXxotuYj9ZhcvXozs7Ow695oVBEEQZB4VBMH8MbjwPHTokJog8/LyVG/UZcuWVZkVx+K0XHSw0C17orKnrI2NjaGHKgiCFUKnD/ttM2uXPbRNEZlHBUGwmHnU0D052cf21KlT2u7du7WpU6eqnrdHjhypdF9dL1tZZJFFlsZeoqOjNVNF5lFZZJEFFjKPNnqM55AhQ9CyZUvMnTu3xjt19peNiIhAdHS0yuYUBEFoaDIyMhAeHq6KZDORwhyQeVQQBHOdRxu9gDzdPvqToj5OTk5qKY+ujIggCIKhMCc3tMyjgiCY6zxqUOE5bdo0XH/99epum77/77//Hhs2bMCqVasMeVpBEASLQeZRQRAsCYMKz4SEBNx///2IjY1VplcWQeZkOXToUEOeVhAEwWKQeVQQBEvCoMJz3rx5hjy8IAiCxSPzqCAIlkSjx3gKgiAIgrHqoBYWFhp7GIIF4uDgADs7O2MPwywQ4SkIgiBYNCzeEhcXpzJuBcFQeHt7Izg42KwSFY2BCE9BEATBotGJzsDAQLi6uoowEBr8xiYnJ0fFY5OQkBBjD8mkEeEpCIIgWLR7XSc62b1JEAyBi4uLeqT45GdN3O5VY5r94QRBEAShAdDFdNLSKQiGRPcZkzji6hHhKQiCIFg84l4XDI18xmqHCE9BEARBMGOaNWuGjz/+GJaCpV2PUBYRnoIgCIJggkRHR+PBBx9EkyZN4OjoiKZNm+Lpp59GcnIyrJnXXntNWRe52Nvbw9/fH1dffbUSq1W1kq0KdgHjcaTiQeMhwlMQBEEQasHu3bsxePBg9Whozp49i169euHUqVP44YcfcPr0acyZMwfr1q1D3759kZKSAmMmbJWUlMCYdOzYUXXzioqKwvr163HHHXfgnXfeQb9+/VRrWcF0EeEpCIIgCLVg4cKFSuQsWrTI4OeaOHGisnKuXr0aAwcOREREBK6//nqsXbsWMTExeOmll8rsT7F19913w83NDaGhoZg1a1aZcj+0EvIYTk5OyoL61FNPlW6nlXDy5MnqdXx9nz59lCVQx4IFC1SNyt9++w0dOnRQx/j666/h7OxcwVJIiyzFuY7NmzfjqquuUlnf4eHh6rzZ2dml25kFftNNN6ntzZs3x3fffVer94eWTtbM5LV07twZTz75JDZu3IjDhw/jvffeK92PvysKeA8PD7X/PffcU1r26Pz58xg0aJD62cfHR1k+x40bp56vXLkSAwYMUNfNagg33ngjzpw5U6uxCdUjwlMQBEEQqiAyMhJ79uzB3r178dNPP6l1P/74o3rO9dze0NCauWrVKjz++OOlZXp0UDyNGTNGjYWCUseMGTPQtWtX7Nu3D1OnTlUCcM2aNWrbL7/8go8++ghz585VFtRff/1ViTUdTzzxBLZt26au6+DBg8p6OHz4cLWvDtappKCj4Dxy5IgaA0UZj61vCeW4uI1QqPE4t99+uzout1GI8nw6KPQYUkBBv2TJEnzxxRelwrCutGvXTonzpUuXlq5jhvmbb76JAwcOqOum2NSJSwph3fhPnDihLKiffPKJek5x/OyzzyrrNq3Mtra2uPXWW41u6bUINBMmPT2df1XqURAEwRBY+jxj6ddXE7m5udrRo0fVY33ge6dbbGxsyjzqloZm+/bt6rjLli2rdPuHH36otsfHx6vnTZs21YYPH15mn9GjR2vXX3+9+vmDDz7Q2rRpoxUUFFQ4VmRkpGZnZ6fFxMSUWX/ttddq06ZNUz/Pnz9fnW///v1l9nn66ae1wYMHlz5ftWqV5uTkpKWmpqrnDz30kDZ+/Pgyr/nnn380W1tb9fs4ceKEOu7OnTtLtx87dkyt++ijj6p8f1599VWta9eulW6bMmWK5uLiUuVrd+3apY6fmZmpnq9fv1491425KhITE9V+hw4dMthnzVrmGbF4CoIgCEIVfPvtt8qtS3QWRt0j13O7odC3aNYE4z7LPz927Jj6mRbM3NxctGjRAo888giWLVuGoqIite3QoUPKUtmmTRu4u7uXLnRb67uW6fbv0qVLmXPQskmX/MWLF9VzuslHjBihLKGEVka66fWPO2zYMGU1PHfunBof38OePXuWsVrqXl/f90y/rBGt0nTlM8yA7naGLRDGhlYHrb0MXeB75unpqTLta/M6oWakc5EgCIIgVAHFVfv27cuIIx07duxAjx49GvycrVq1UuKJwozu3fJwPWMSAwICanU8upTpSmZ8KN3vdOHTNU9xmZWVpbrsUKCV77ZDoaiDLv/ydSp79+6Nli1bKhf9Y489pgQthaYOHnvChAll4kl1UAiePHkSDQ3fG8aK6tzlFLpcKIr5flE48nlBQUG1x6FYZRWBr776SsWRUix36tSpxtcJNSPCUxAEQRBqAeP8KEB0j4aCySxDhw5V8Y6TJk0qE+fJvvMUUffff38ZIbh9+/Yyx+BzCmYdPAbFFBcmLtGySGtn9+7dlcWTcZVMAqqPMOd4wsLC1PtCi6cOivKjR48qIV0ZHAMtrxS9FLGEArm+pY2OHz+ukoKmTZtW+pylp959910lvkn5igS05BK+Bzr4Go6DolP3njA2VWgYxNUuCIIgCNXA3ttM6qHVkyWN+MjnXG8oPv/8c5VtTuvcpk2bVAIORRUFKbPPp0+fXmb/LVu24P3331dWRGa0L168WCUYEVoh582bpzK+WaaJ4QEUorTo0cVO8Ughy6QcusB37typShOtWLGixnHytUy04nhGjRqlMt51TJkyBVu3blXJRPv371fu6+XLl5cmF7Vt21YlH9EqSusxBejDDz9cIaGqMihYKcLp5qeA/uyzz5QbvVu3bnj++edLraoUltzG62ZWPhON9OF7QAH/xx9/IDExUVlpaU2m+P/yyy9VGau///5bJRoJDYRmwlh7ULwgCIbH0ucZS7++mmiohI+8vDytpKRE/cxHPjc058+f18aOHasFBQVpDg4OWnh4uPbkk09qSUlJZfZjctHrr7+u3XHHHZqrq6sWHBysffLJJ6XbmaTUp08fzdPTU3Nzc9OuvPJKbe3ataXbmXT0yiuvaM2aNVPnCQkJ0W699Vbt4MGDpclFXl5eVY7ziiuuUJ+xv//+u8I2Jg4NHTpUc3d3V+fu0qWLNn369NLtsbGx2ogRI1RSUkREhLZw4UJ1PTUlF+kSu5gY5evrqw0YMEC9pvzv5fvvv1fXxeP37dtX++2339Tr9u3bV7rPG2+8od4zJo3x/SZr1qzR2rdvr17HMW/YsKHahC8iyUWo1Txjw/9gomRkZMDLywvp6ekquFcoS0FRCXILi5HPpagE/E3qPC98tLWxgbODHVwd7dSjIAjWN89Y+vXVRF5enrLiMe6PdScFwVBY82ctow7zjMR4mjDFJRqSs/ORllOI1OwCpOUWIj2nEOm5hUpwcnttsbe1gYujHdyc7OHj6ogAD0f4uTnB38MJ7k7yMRAEoSz7olIR5uOKAI//XKeCIAiXiygOE6GkRENSdj4SMvIRn5GHuIw8JGcV1ElcVkdRiYbMvCK1xKXn4Vjsf9toEQ3xdkGEr6tafN0uBVsLgmC9HI/LxIYTiQj1dkHnMC+0CfKAnW3ZrGZBEIS6IsLTSDDCITEzH9GpubiQmoOYtFzkFxqnI0JOQTHOJGSphXg42ysB2jLQHc383OTLRhCsGM5NXDadTETnUC90j/BR3hNBEIT6IMKzEcnKL8L5pGycT85GdEou8gr/K99gStAqeuRihloYG9omyB1tgz2U5aN8HTdBEKwD3qDuOJeCfdFp6BLmhZ5NfeDqKF8hgiDUDZk1DOw+v5iei/NJOTiXnI2kzHyYGxTHBy+kq8XTxQEdQjzRNdxLvnAEwYqTGnefT8WB6DR0CvVCr2a+EicuCEKtkdnCAEItMjkH55KycC4px2StmvUhI7cQ288mY/f5FLQL8UT3CG/4u0vigSBYI4XFGvZFpSnPCK2fvZr6wN5OSkMLglA9IjwbgLScApxJzMa5pGxcTMttsIQgU4WJSodj0tXS1M8VvZv5ItzX1djDEgTBSBbQbWeSlQC9qrW/SkISBEGoChGe9YDCkgLzbFI2ziVmITWnENYKrbtcKED7t/JHkKd11S4TBOE/j8iKg7HY75OGa9oGINBD5gJBECoiwrOWpGQXIDI5W4ksZnjyLl/4D74vUSlRaBXojn4t/aUkkyBYKTGpufhhRzR6N/dBn+Z+UhVDEIQyiPCsotRRUlaBsmpyodBkprdQPeycdCqeZZmy0SnUU1lApWOSIFgfJZqGHWdTcDYxG9d1DBLrpyAIpVh9JDgzz5Oz8nEiLhNbTifh130xmL3xDL7dHom/jyeoIsoiOuv+pcMs+P9tPY8jF9ONPRxBEIwEaxX/uDMaO84mq7lWqBvjxo1TJeweffTRCtsmTpyotnEfQTAnLNLiWVhcouIw/3vUkF9UrOpoZrF7z7+PbD3JVpRMlhEMU/dv9ZF4HL2YgcHtAuEnGfCCYHVwDt56JlnFxN/QOQReLg7GHpJZER4ejh9//BEfffQRXFxcSnuCf//994iIiDD28AShzlikxXPOhjOYveEMvv7nHOZvOa+sl4t3X8Bfh+Lwz6kk7I9Kw+mELHU3LqLT8FxIzcV3O6Kw9XSSxWf8C4JQOWzV+/2OKNVEQ6g9PXr0UOJz6dKlpev4M0Vn9+7dS9eVlJTgnXfeQfPmzZVA7dq1K5YsWVK6vbi4GA899FDp9rZt2+KTTz4pcy5aT0eOHImZM2ciJCQEfn5+yrJaWGi9CbRCw2ORFk/B9KDgZNcTFtIf3jFYrJ+CYIWwrvGv+2NwZQs/9Gnua7ROaIxHz8lp/PO6ugL1ueQHH3wQ8+fPx5gxY9Tzb775Bg888AA2bNhQug9F57fffos5c+agdevW2LRpE+69914EBARg4MCBSpiGhYVh8eLFSlBu3boV48ePVwLzzjvvLD3O+vXr1To+nj59GqNHj0a3bt3wyCOPNMybIFg9IjyFRiUhI19ZPfq39kf3cG9pwSkIVgZFH+t+xmfkYVjHYKMkIFJ0urs3+mmRlQW4udX9dRSQ06ZNQ2RkpHq+ZcsW5X7XCc/8/Hy8/fbbWLt2Lfr27avWtWjRAps3b8bcuXOV8HRwcMDrr79eekxaPrdt24aff/65jPD08fHB559/Djs7O7Rr1w4jRozAunXrRHgKDYYIT6HRYXjDxhOJOPdvxquHs8R8CYK1wYx33oSO7B4q5ddqgFZLCsAFCxaoqiv82d/fv3Q7LZM5OTkYOnRomdcVFBSUccfPmjVLWUujoqKQm5urttOaqU/Hjh2V6NRB6+ehQ4cMen2CdWFQ4UnTP2NRjh8/rmJK+vXrh/fee0/FlghCVEoOvt0eheGdgtHcvx5mAEGwAix5HmWC50+7onFztyYI9b6UONNYLm9aHxsbnre+0N3+xBNPlApIfbL+vZgVK1YgNDS0zDYnp0thTbSQTp48GR988IGyinp4eGDGjBnYsWNHmf1pGdWHXim66QXBLITnxo0bVWBy7969UVRUhBdffBHXXXcdjh49Crf6+BsEi4z5Wr4/RhWavrKF8WK+BMFUsfR5lHPA0j0X1A1o60Zqt8lpxtzeuuHDhysLJefIYcOGldnWoUMHJTBpyaRbvTLonudNy+OPP1667syZ/2/vPMCjqNo2/GR7eu+F3nsXRFAQwYJiAVQ+7IINf7Bh+bArKHwWELBhwy4qKL1I7y0QIARIIL33upts9r/eEzckkEAStszsvrfXcbOzk50zZPfMM2+Nt/q8GcamwnPt2rX1npObICgoCAcPHsSwYcOseWhGZjFfexJqYr7o4sNF5xnGudZRCr9ZFZOOayqq0K+Vr72nI0nI/R0bG1v7c13IeknWzBkzZgjr5NChQ1FYWCjEppeXF+6//36RcPTdd99h3bp1Ir5z6dKl2L9/v/iZYRw2xpO+CISfn58tD8vIhLM5paLs0tieoQjinu8M41TrKN2AbjuVjeKKSgzvGMjejwYgEdkYb731logFpdCMhIQE+Pj4iFJMZCEnpk6disOHD4ssdfq3veeee4T1c82aNTY8A4YBXEwUqWwD6C7s1ltvRUFBgci0awjKzKNhpqioSNQvo4X2Ul+4C1mw6TTX55QxKoULRncPQUcbud0Y54bWGW9v72avM/bAluvoT/uSRO1Ne9AtzAujugZbRHxSsfWzZ88Ky55Oxze0jPVw5s9ac9ZRmxWQpxilY8eOiQDnxqA7NZq4edBiyTgfdNOwOiYd+87m2XsqDCMpnGUdPZ5WhHXHM7jNJsM4IDYRnpSJt3LlSlGQlgrYNgbVKSO1bB7Jycm2mB4jQcgOv/NMDtYfz+BuRwzjhOtobHox1rL4ZBiHw6oxnuTFnzZtGv78809R6PZyQcyUlWcu/cAwZstHUUUVbukZyklHjFPizOtoXEYxqk0m3Ng9FEoFx3wyjCOgsLZbiFp4/fjjjyLrLiMjQwwqXMswTSU5r0zU+qOafwzjbDj7Ono6s0RkvLPng2EcA6sKz8WLFwtXz7XXXiu6H5jHL7/8Ys3DMg5IXqkBvx1IRm7J+aQJhnEGeB0F4rNKsOZYOrvdGcYBsLqrnWEsRXFFFX47mIJxvcMR4u1cGYOM88Lr6HnL50ZlpsWy3RmGsQ82y2pnGEtQbjDi90MpSMots/dUGIaxQ8z31lPZ9p4GwzBXAAtPRnYYqqpFm80zWcX2ngrDODzJp2Kw6Pn7xKMUOJxUgF3xOfaeBsMwLYSFJyPfFntHMxCbXmTvqTCMQ7N/wwqcObIXBzaugFTYm5CHg4n59p4GwzAtgIUnI1uozAoVmT6RxuKTYSxJYmIiEmJjkHL6OKK3rhLbDm9ZLZ4nnzqGvMxUe09RtNc8nlbTPpS58jjiKVOmiDasFD8bHR0tktmmT59+yd9r3bo1PvroI5vNk3EMbNqrnXFuyFX39xdzMfbR5xHZsYdF3pPyLtafyBA/dw2TdrtDhpELJCgupKQgFx88eUft8w/Wx8HebDyRBQ+tCq383Vv0+x9uOAVbMWNUxxb9HpXOeuedd7Bq1SqkpqYiKCgIvXv3FqJw5MiRFpnb2rVr8c0334g6sW3btkVAQAD++OMPqNVqi7w/w9SFLZ6M7F12ZvHJ1g+mRF+FgjKDvache6huqFLZsF1CoVRi0sy5kIrXY+XRdGQV2aenvLU5d+4c+vXrh3/++Qdz585FTEyMEInXXXedqO9qKeLj40WJriFDhiAkJAQqlUpYP6luLMNYGhaejFUhlxy55qztsiPxueFEJo6lsvh09qSzUoPR3lORPZMmTcKbXy1v8LXp839Dv5G3Qkp/9+XRqQ7ZYOKJJ54Qru99+/bhzjvvRMeOHdGtWzc888wz2LNnj9gnKSkJt912Gzw8PODl5YUJEyYgMzOz9j1ef/11YSFdunSpsGR7e3vj7rvvRnFxTXLmAw88IDpj0fvQsczW7gtd7VlZWRg7dixcXV1F96wffvjhovkWFBTgkUceQWBgoJjLiBEjcOTIkSbPhaiursb777+P9u3biw5cUVFRwuJrhlrA0jn6+PgIcUznTgKdsRxUL7e4wnrfJxaejFV5e/IIfPjUncJFV1KQV89lR9vpdUtB4nNjbCZbPp10oaQC41lF3GDA0phrZkq5dmap3ojlh1NRUek4Nx15eXnCukmWTXf3i0MJSHiRSCPhRftu3boVGzZsQEJCAiZOnHiRRXP58uVYuXKlGLTvnDlzxGsff/wx3nzzTURERCA9PR379+9vcD4kUEn0bd68GcuWLcOiRYuEGK3L+PHjxbY1a9bg4MGD6Nu3rwgHoPk1ZS7ESy+9JJ7PmjULJ06cEB27goODxWuVlZUYPXq0sMRu374dO3fuFIJ7zJgxMBjY02EJEnNL8f3eRCRkl8JacIwnY1XIJffTvBdRbTQ26LK757nzC47FxOeJLGiUCnQIZjeRs7DlVJZVF0pnxMs3AJ6+AfAJDMWgMXdh79plKMhOh4ePP6Ta3eyv6DTc0TccKqX8bSpnzpwRST+dO3dudJ9NmzYJ9/vZs2cRGRkptn333XfCKkoCcsCAAWIbCVSK4TS7zidPnix+lyyJZHWk7UqlUrjZG+LUqVNCTJLl1fyeS5YsQZcuXWr32bFjh3idhCdZKol58+YJkUlClZKXLjcXsnySEP7kk09w//33i9fbtWuHoUOHip+pWxf9/pdffll7I/T1118LEU7xqTfccMMV/7s7K/mlBmw7nW2TdZSFJ2NVyCUXHNWuXlJCXZddRIduVon7WnMsQ1x82gS0LOmAkQ8HE/NwJJmt3JbGPzgUs5ZuhlKtFhf5wTdPhLGyEiqNBlIltaAc645n4qYeIZK20FqqY1VsbKwQnGbRSXTt2lUIMXrNLBLJrV03XpPiOS+0Vl7uOBT3SfGmZkgQ03HMkEu9pKQE/v71b0zKy8uFldPMpeZCx9Hr9Y0mTdExSJBfGHtaUVFR7xhM80JV9iTkIjq5AEYbtaRl4cnYDLoQ0GJqfrQm9AVaeSQN4/qEI9LPzarHYuzH6cxibD/NxcStRV2RSd9bKYtOM6cyi+HrpsaQ9gGQMx06dBD/5idPnrzi97owO53elyyHloREJ4lIsjxeSF2Beqm5UPzo5Y5B4reh+FKKK2WaR0J2Cf45mSXaUdsS+fsjGMlDrjly2UV06I67nn5DPNJza7vsqMj8X0fSkF5YbtXjMPYhs6hC1HHlVubMhew9myf7+r6UOEPxjAsXLkRpaWmDiTzk6qa4SxpmKC6SXiPLp6Ug62ZVVZWI2zQTFxcnjmOG4jmp9BNZRikxqO6g8kxNFdskPsn13hB0jNOnT4uSUhceg0IGmKZRqq/CqqPpWBGdZnPRSbDwZKyOT2CIcNlNX/Abhtxyt3ik57TdFm6EPw+nIqvYMcutOHPZJIrnqzSy6mQahhINU/LLIGdIdBqNRgwcOBC///67EF3kjp4/fz4GDx6M66+/Hj169BBVCA4dOiRiLO+77z4MHz4c/fv3t9g8OnXqJBJ4pk6dir179woBStnrdS2UNBea07hx47B+/XqRab5r1y688sorOHDgQJOOo9PpMHPmTLzwwgsiVpXc55S9T/GkBJ0niVhKqKLkIoptJQvr008/jZSUFIudr6NiMpkQk1KIb3efE54Be8HCk7EJ5KKrmx1rS5edvrJaZLxyfUfHoMpYjb+PpAnxyTCXCrf5+0g6CmX8vadi7iQoqW7ns88+i+7du2PUqFHCIrh48WKxlq5YsQK+vr4YNmyYEH/0O5SEY2koiScsLEyI2jvuuEMkC5Hl0QzNZfXq1WIeDz74oCj9RKWSqAuWOSu9KVA2O53rq6++Kiy6lKFvjgF1c3PDtm3bRIklmgO9/vDDD4sYTyrfxDROUUUl/jiUKm7I6JpoT1xM1g62uwKKioqE+bywsLBZH6oFm04LNyvD1MXbVY2JAyLhruXQZjmz9lg6YtMbv1ufMCAS4T6XjhWzxDojF1p6fj/tS0JGofw9BSHuCvTyMQhBRhY1hrEWJIDJCkt1TqX0WTuWWigy1psjOEd0DkKvyPOxuZZcZ9jiyTgNVGD6Twer9edsHDiXd0nRyTANfe/LDEarJzQyjBRjOVdEp4rmKva2ctaFhSfjVGQX64Wblty1jPwyMHec4Qx2pmVud7oIM4wzVfxYuse6heBbCgtPxulIyS/H6mMZotsNIw9yS/SiNisbrWzDiRNAYa5jXR4qqqpZfDIOj6GqGuuPZ2Dl0XSUS7R9MAe7MU5JfFYJNp3MwqiuTQ96Z+wDhUaQlZoWVMY2TJ1KnWgi4B9qQKsuFWjdtRytu1QgtK0eSiVkCyWkqRQu0KplfBIM0whUOnDtsQwUlFmvz7olYOHJOC0UcO2hVWFwO2m2AGRqyn9Qrc58iS+kjgRZlfV6ylI2ITddI8ahf2qSBbRuRvQYUoo+1xahY98yKFXyjPn0VbhA7QBtNRmGIO/dvnN52JuQJzr3SR0ZLhsMYzmoVZinToXu4Vx8WIrsTsiVZIySI0NVz/btA5b8k4L9e11wLlaHc7GuSIzVoaJUiQMbvcRw965C72El6DuiCK27Vojfkyb/TuzfCzL9nyxCfu4aKBWSnTQjQ+yRwFZYXol1xzJEu1i5wMKTcXo2xWbBTaNE20APe0+FqcOZrBLsO5tn72k4LW4eJnTqX45O/WuKsFNXQxKfh7d4InqLJ0oKVdj5t48YEe0rMPKePPS4ugQKiRkSDdUuMJpMMOgroNbWlLghq1BhuQG+bufrCzPMlVJWVtZgW1BrEZdRjE0npZWx3hRYeDJOD12EVsek465+kQjxlk7tNWdPJuJ2mNKCBGWbbhVi3PZYNs4cdsOhzZ44ss0TKWd0+PatMARF6jFyYr6wgkrFDW+EC9LKFFDnZIvnGhKfLi6g4A2jQQ9PV+n3n2ekb+kk0UmF7qkvvdLKgdCGqmpsjsuSbVtYiSwNDGNfqPUi1Tub0D8Svu58IbInnEwkfei6SpZQGmOnZGP7cl/sWOGDrGQtfpoXgrXf+WPMfbnod32RJCygyRX0nTag0pgJ5QUWTq1aAa2Kk42YK4dEZ0iIdVtBZxRWYM2xdMknEF0KFp4M8y9UZHp5dKrobuSm4a+GPeBkIvnh4V2NG+/PxXV35WPXKm9s/d0X+VlqIUB3r/bGndOyEN5Ob+dZuiC5Qou0ChM0CjKjnzelkw4d1iEIbQLd7TpDRt6Qe92als5qmSUQXQq+ujJMHeguckV0Gu7qF8FZr3ZgT0IeJxPJFJ17NUZMyMfQ2wqwfbkPNvzgj3MnXPHBk1EYckshbrw/B26e1XZ3u5dXXxzTufFUHu7yckOod9NbrTKMrSgsq8Ta4+lIK5B/C1uCr6wM04Arg2I+ucC87TsT7T2ba+9pMFeIRmsScZ4vLjmH3sOLYap2wc6/fDD7odY4uMlTknG7VdUmEd5RVMGWdkZ6Zf++35voMKKTYOHJMA1AVjcK3mZsQ36pAWs5mcih8Amswn2vpOOx95IRHKVHaaEKP7wXiu/nhKC8RHqXnlK9UXg79FXS7PbCOBel+ipxM0R91h0t3l16336GkQhHUwq5nI8NoEV15dE02ZUEYZpGxz7leO7TRIy+LwcKhQmHN3th3mOtEH9Uem7tnGI9Vh1lbwdjX2LTi0SfdSop54iw8GSYS7ArPkcsAoz1oDv6nBKDvafBWBEqrTT6P3l46sNk+IcZRPLRoucjsPprfxgl1j49MbcMG2Mz7T0NxkmtnH8dSRNtL6XaZ90SsPBkmEtArl8SRkm5NYWBGcuy/1weTmUW23sajI2gfu/PLkrEwNGFMJlcsPEnf3zybCQKc6VVzuh4WpHoasYwtuJ4WiG+252IeAe1ctaFhSfDXAYjJR4cTUNWseMEd0uBczml2Hkmx97TYGyMzs2Eu5/NxP3/TYOrhxGJsa748MlWoiuSlNgdnyvbAt2MfMgu1uPX/clYfzxT1DB2Blh4MkwT4xBXHE4TfXEZyyQTrT6WzslETkyvYSWY8UkSQlrpUZSnwifPRWDfOi9ICXK5J+ext4OxPBWVRpHA+uPeJFn1WbcELDwZpomU6Kuw/HCq09yVWgvKGiYLMicTMQFhlXj64yR0H1ICY6UCP/8vBH8uDoRRIl8x9nYw1miScVy41c8hOqlA9sXgJSc8t23bhrFjxyIsLAwuLi5Yvny5NQ/HMFYnr9QgWmtWGlk0tbwzUSZyOZmoyTj6Okqu9wdeTcPoyTVhF9v/9MXnL4ejvFQadhG6QaIbzoIy/swyV7b2UTz70j2Jwq1O5bucFat+s0tLS9GrVy8sXLjQmodhGJtChXzXHMvgkist7EzkDMHzlsQZ1lHq5z56cp4QoBpdNU4fdsfCZyNQJJGkIxIJfxxKFVnHDNNc4rNL8MPeJFGqK5dvuq3bMvPGG28Ug2EcDRJPm05mYVTXYHtPRTacySrmzkQtwJnW0Z5DS+AXnIwv/huOtAQd5k+PwpTZKQiKsH9sNcV3/3E4FeP7RUCnloYgZqQLhWmczirG4aQC0Q2POY80fBkMI9NWZttPZ9t7GrIgq6hC1KZzwnAmpplEdNDj6Y+SERBmQF6mGgtmREom450KzP8VncahNswlcwGoIsJXO85iTUwGi06pC0+9Xo+ioqJ6g2GkzIFz+TiYyN2NLrcQU1HkSiOrTlvgCOuof2glpn2YjMiOFaLV5uIXInBinzukAGUgr45JFxYthiHos3A2p1R8LkhwUg1YWvcYGQjP2bNnw9vbu3ZERkbae0oMc1m2ncoR1k/mYsgyRMlYxRW8CNsKR1lHPX2NeGJuMjr1L4VBr8BXr4UhepsHpEBCdo3I4Dhv54X+9tRYhBqMfL4tQSSgxWUU8w2J3ITnSy+9hMLCwtqRnJxs7ykxTJPYFJvlsH11rySLk9zrWUV6e0/FqXCkdVTrasIjb6ai73VFqDa6YOm7oTi4yRNSgL7vVIuWxafzQMll1EJ53fEMfLE9Ab8fShFGBy6xJ6Hkouai1WrFYBi5QbXY1sSk49beYWjlLw2XoL3ZeSaXxbgdcLR1lPq83/tCBlQaE/at88aP74fAWOWCgaPtH0JwOrMEq5GOm7qHQqFwsfd0GAtDgjK9sEI0EUjKK0NOiZ7j1KUuPEtKSnDmzJna52fPnkV0dDT8/PwQFRVlzUMzjM2pomLTR9JwW+9wRPq5wZkhKwD1YWeuHF5HAYUSmDAjEyq1CbtW+ohC81WVLhhyS6EkxOcaZODG7iEsPmUMeWgKyiqRVliO9IIK8Uh1m1loykx4HjhwANddd13t82eeeUY83n///fjmm2+seWiGsQuUQEOJNOP6hCPcxxXOWrOOQg8Yy8Dr6Plan3dOy4JSbRJF5pfNDxaWz2vGFdh7aqIwOMHiU37WTMo6zygqR0ahnl3mjiA8r732WnEXwTDO1tedAs3v6BuOUG/nEp8p+WVYfTTdKdvAWQteR8/j4gKMeyxbWD43/+qHPxcFCYvUsNulIT6rqqtxU49QqJWSSp9weuj7Q9ZLEpppBeXiMb+MrZn2QlIxngzjSOLzz8OpuLNvBIK9pFGD0NpQP2uy9lLIAcNYU3ze8nAOlEoTNv7kj+WLg4QQlYLbnbLd/ziUIsJtuMi8fYVmbqlBxGam5JeLwdZM6cC3ZYxNSD4Vg0XP3ycenQXq8Uxt9jKLHL+AMPWxJisvnTPD2EJ83vhALq4bXxNHTG73feu8IJWWur8eSEZxhf27LTkTZYYqHE8rFGWuqLzR0t2J2BKXLRIcbSE6nfEa11LY4ulEkFuhokyBKoOL+FmM6pqfta7VcPWoFgu6Ndi/YQXOHNmLAxtXILJjDzgLtOAtO5iCW3uFOWzCEZUYqeljzRYFxsaWz0dyUFXlImI+f/kgGEqVCf1G1sRb2hPqx/3L/mTc3icc/h6OU2FAil4WsjKfyylFRlGFXV3nznqNawksPB2MsmIFMhI1yDinRWaSBgU5KhTnqVCUp0RRngpVhsaN3Ep1NTx9jPD0M8LLtwp+IZWic0hkRz0CIwwiuL855GWmorQwHy4uLojeukpsO7xlNQaMul24Qty9feEXHA5nifm8qWco2gVKowC2ZUVniuhjzTD2ivk0VrqIbPcf54aI5KPew+xfxouaJvx6IAVje4UiwtcxbzrtJTapkgDF1FIWuj3ha1zLYOEpY/TlLkiMdcXZ4zrxmH5Og8IcdZN+18XFJBZtF0XNz1WVChgrFSjIpnHxe2jdjIhor0erLhXoMaQEUZ0rLmsdfXvyiIu2lRTk4oMn76h9/sH6ODgDFPe48kg6bugWjC6h0nAJXinkSiRLJwXtM4y9oHXojqeyRHklqvP5/exQqNRp6D64VBIeD/qODOsYiN6RPvaejmzJLzUgNqMIpzKKkW9nsVkXvsa1DBaeMqKizAVnjrjh9GE3nD3mirQELaqrL1Z/PoGVCGltQEgrPXyDq4T10su/Cl5+Rnj4VEGjqxGddSH3e3FBjVW0OF+JolwVMpM1SDmlQ8oZLfRlSsQfdRPjn1/84BNQiR5DS9BzaAnadCsXdfYuZNLMufhp3ouoNl7sglUolbjnuTlwJijTmzpe6KuqZX8RKiyrxLJDKShiSycjAcgbM2F6piivdHCTF757OxSPvp2KDn3K7T010UJx88ksEes9snMQVJzx3mTRTpbNE+mFIm5WivA1rmWw8JQwFK9C4vLkfjecPOiOc8ddxcJaF9/gSrTpWi7EX1g7vRCcru7NT/CgriC+QVViXAh9p7KSNEg+pcPJA26I3eeBghw1ti/3FcPTrwpDbinA1bcUwsPn/Bew38hbERzVrt7dn5np839DRIducMa/KV2EyFo4tH2AcNHIDbJwknud+68zUoJufu9+LgP6cgWO7fLAktfC8fh7KcJLIwVOpBWJ2M9beoXCS9c0z5SzQe5pykCnBhRUD5jqIksZvsa1DBaeEsNYBWFVjNnlgWO73C9ynfuHGdCpbxna9SxDm24V8Am0/sVfqQRC2xjEoDZ1lQYXnDrohqM7PHBsj4eIIV33XQA2/eyH/tcXYfgd+QiOqm8JI4FFi4r50dk5cC5fXITGdA+RVdmV7GK9EJ1lBk4kYqQHrVX3vZyOL2eF4dRhd3z+33A8OS8ZYW2kEQ5CVs+f9iaJ7z231q2fkU7CnASnlFzpzYGvcU2HhacEMFS4IHa/O2J2euDEXndUlJ4XIhptNdr3LkPnAaXo1K8MgeH2/1KqNSZ0G1wqhrEqE0e3e2LL777CIrpntY8YXQeViHInHj7+8PQNgE9gKAaNuQt71y5DQXa62O7snM0pxc/7knBr73D4uWsgdShzdPWxdC6ZxEga8t48+HoaPn0xQsS+f/ZSBJ76X7Ik1k6Cbtqoxm/PCG8MbR8Ijcp5Xe9UZzMmtVCUPKKQBDnC17jm42KSsDQvKiqCt7c3CgsL4eXV9ISMBZtOS76ItRCb+9xxZLsnTuxxh0F/fvGhOEwSdZTE06FPmRB6loJqjP39xVyMffR5i5Z8oE8RxZ2SAD2+2x0mk4tIWho4pgg33JsGnyBF7Z2gsbISKo30hZatoAsPtdprK9GMd/qb7Tubh90JubLo9DFhQGSz2pW2dJ2RCy09v5/2JYl2gnKu8LHw+UikJ2hFSNK0D5Jt4iFqDj5uaozqGuxUWe/6KqOwbpLgJK+PI1BlMECpVjvUNW5E5yD0akYuQnPWGbZ42hCD3gUn97sjeuvFYpMWxl7XFKPH1SVo1bmiwWQdKdcao1DFtj3KxchOVWPtt/44vMULe9d4I3qrB0bdm4dh4wqg0kD2X0hrlFuijj8DW/thUFt/KCXU65kuEuuOZyI+y/7laRimObh5VuOx2Sn48Klg5Gd64JNnAzDjkyy4e0nHYk/lgKjOb58oX1zdzt+hE4+oDNLR5ELEZRaLNc+RqHtNI/HJ17hLw8LTylRVAnEHa8Tmsd3uIjvcjB+JzWHFYlCtTGvlmdi61hi5tCa/nIGhtxWIdnbkgl/5ZSB2r/LGXU9niZABpj5kSdx7Ng/xOaW4oWuwJNps5pbosfJoOpdLYmSLp68R7XvPwoENzyAvIxJfvKLGY++lQOdmktR3/1Bivri5G9ohAB2DPeEoVBqrEZdRLKybcraeM5aFhaeVEoROR7sJsUlJQuXFynqljnoPLxbDmmJTCrXGKPnp/+Yn4eAmT6xaEojcdI2Itxp6W77otUxlnZj65BTr8fO+ZPRv7Yur7GT9rK42ITqlALvjcx3OMsE4B3Vvtk/u/w7AWsBlB5Li/LF4pjcmzTyBoIgwSAlqwrDqaDoOeefjmo6BzQoXkWISIiUKUe1NjglnLoSFp4WgkkPxR11rxOYOT5QWnRebVG6o97AasUmlPZrbAUjOtcboXAeMKhb1PlcuCcTOv3ywY4UvTh1yw6SZGUJ8MxfX+6SYyoTsEozoEmzTC1BKfhk2x2ULAcwwcuXim+08wHQjgH+QHBeEOQ9tw7w1YVYLaboS0gsr8Ov+ZLQL8hAl1+SQeEiUG4xCaMamFyGriNcPpnFYeF6hZfNMtBuObPcQls3SwvP/nB7eVeh5TYkQm227N1xg3ZlqjWldTbjzqSyR7f7L/0KQlazFx/8XhRsm5WLkPXmiDApTn5wSg7gARfm5YWAbP6v2ei/RV2HH6WzEptu/zzXDWOdmez+A2wCsBnAXfptfKIrOS7WULrne6eazTYA7ekX4oJW/m+Tq/pIrPTG3FCfSi0XVC7lmpjO2xemE55VmdVOCEHUOotJHVKS4rI4b3d3LiO5Xl6DP8GK061UmSTFl71pjXQaU4fnPz2HZ/GAc2eaJtd8FiLCE+/9L5Se4NmRDJOWViRHu6yoSkFoHWK7+X06JvjbDlN3qjKPQ+M32Pxj7aDRWLhkoEh9pzaawH6lCS3RCdqkYZPmkEkxdw7ygVSntmnBIpeCoqxCJTqkXeWekh9MJz5ZkdVNZDqqvSULz5AF3GCrqlD7yrhKtI3tdUyKKuisl+i8qpVpjlFV63yvpODS4RAhQKpj/4bQoPPR6GsLbsYumMVLzy/FnfioCPbVoG+guLCEhXrpmW0GoFd2pzGIcTyvigH/G4bnwZrtD7zSM/79M/PphiGj/6+ZpxIgJ+ZA6lOS3JS4bu+JzhfejbUDNGuCute5Fh/7Nsor1IgyHboCT88rZsslcERKVSfbN6qa7zIxzGlHUnQQntaqs2xOd+pR3G1KKnkOLRfkgKVo2L8QnMASzlm6urTU2+OaJdq01Rlqp38hiITS/ej0MOWkaLJgeKVre9R7OpXsuF7hPY29CHlw1SrTyc0OUvxs8tWro1AphDdGKR4UoVk0XrPwyg3ikkVZQzlYKxuG51M32VTcWCW8VVdugQeKTtkmhJvLlIM8EueFp0Doa5KmruQn11sHfQ3PF7TipnS+tE+QNofaVqQXlnCDEWBSnEJ5Nyep++/fTOHPEDXEHqRe5Owqy6395Q1rrRUH37kNKENHBNtnozlBrjHrLT1+QhKXvhoqyU9+9E4bU+FzceH+uJAP/pRjQfzKjWIwLoc+oHAq+M4w9brbJyklJoJt/9cNvHwfD1aNaeK7sXRO5OdD3m9pw0jBDN53+7hoEeGiFNZQaVGiUCvGoVipAly59VbVwmYvHymrRspLEZm6pgUNuGKujct5Ac6qTOBhwuQH+Ifdj1vgQmOpYNVWaanToXYYuA0tFXKJ/qDTarTlqoedH307Fqq8CsPk3P2z62R+ZiVr85+V0aLSsnFoKi07mSugU4olgL62wjpNrlRJJSKRkl+hlI04ud7NN8Z1k+aR4z+/nhMDVPQ0d+5ZJpiZyS6C/UVpBhRgMI0VUzhJo7h3QGYue/xTANQCGAhhAzQoBE5CbXrNfUKReLDqdB5Shfc8yrjNpQ8i6OfbRHIS11eOXD4JxbLcHPn85HA+/mQZXd3lc5BjGkegb5dtondecUj0yC/XIKCKBUy7bJgPkFRj/dCYqShSifTGF/Tz+XoooeyelmsgM40g4pPCsNLggOV6DpDgdkuN04jEzqSPd316wZxq6DKxC7+FqdOhTDp8AafXxdUYo7pP6KS95NQwJMW5Y+GwEprybCi8/znhnGCmgULiIuEIaPeAttpGrl6ojUDtECv+Q200v1RQuL1Xg1CF3fP7fcDw5LxlhbQySqonMMI6Ci8keNXWs0HSeyMwExo4FDkVXw1h5cZV2F8UZuHkcRZdBCqSe/hzFBfvxzMLfRSwQIy1S47XC4lmcr4J/qAFTZ6ciIIzDHZjLM2FAZLOK7jd3nZEbtjw/csmfzSnBsdQiUXJHTujLXURntXMnXOHpW4WnPkgW7X8vJOX08QZrIj+z8A+b1ERmGFswonMQekX6WGWdcSiLZ0AAcOIEhOik+myRHSsQ1bmi5rFTBVzdK6BUd/u3rMaHds3qZi4NZbtP+zAZn70ULlptLpgRiSnvpiC8nTxdegzjDFCL1/ZBnmKQFXRXfA7O5TQeMyklqMnFI2+lYtHzEUhL0OHTmRFCfPoGVUmyJjLDWJPqaqCqClBZQSU6lPCkskbLlwPb08/CO7iygcxz6WV1M41DFs5pHyTjs1cikJ6gxaIXIvH4nBRRVYBhGGkT7KXD7X0iRP1Hqj1JdWjlkOhI3pWFz0YiK0VTKz49fY2SrInMMI1B90JUg7ykQCm6KhYXKMXPZUVKEVZSXqJEeYlC/FxRqoShwkXUKKcmOfRYqVfg44+Bp5+GxXEoV7uZBZtOo4oL3DoM9OX44r/hwgXm6mkUwf8R7Vl8Mg3DrnZpnh+1VNwSl4X8MumHzORnqfDJs5HIz1QjtI0eT8xNFo0vzFQZDLVlmugSyt4zxtZUGVyQl6lCTroGuelqFGSpUJCjQmH2v4+5qgZDDpvDu+8CL73UtH2d1tXOOCZUX2/KO6n47OVwJMa6CivE4++T253FJ8PIBWr1+h/fVtiTkIeDifmolq7NQ7jXH5uTIsRn+lktvnglHFPnpNZW2JBiTWTG8TCZqGqCEhmJGmQm0dAiM1GDnDQ1CnNUMJkuX1Bc526Ep48R7t5G0ZaaHt08jOK66upuhM6jGjq3amhdq6HWmqDVVUOjq8aI7v4Y0rUmedDSsPBkZIHOvVpkt1PMZ9JJVyx+gcRnMsd8MoyMUCkVGNohAB1DPLDxRFa9wudSgxKLHpudgkXPRyIpzlWIT4oz17lJVzAz8qXaCGSnapByRotUGvFapMXrRJODxiCB6B9WCf+QSvgGV8I3sAreAZWiMox3QBW8fI1QaVr2efULMEGrhVVg4cnIBrI2UPzVZy+GiwvBpzMjhfgMa8vik2HkBJViuntAJA4n52PXmVzJhkaFtjFg6pwULJ5Zk+1OIT/kfaFEJIa5EkoKlDgXqxOfq8RYHZJP6URs5YW4uJjgF1KJ4CgDglsZxGNguEHkQJAFU5ZdFO09AYZpifj89MUI8UU1B/8HRUo/boxhmPr1QPu18kOknxtWH02XbOwnxZOT5ZPE59ljbkJ8Uqc1Fp9McyjIVuHMEVfRmjs+xhW5aReHZ2i01QhrpxdhZOHtK8QjiU1H6+DHwpORHRSbMnV2ihCdKWd0ovbepcqeMAwjbevnvYNa4Z+TmYhNL4YUieyor/W2UGOLJa+Gi9JL3N2OaQzKKD91yA1xh9wRf8QVOQ0IzeBWerTuUoFWXcrFY1CkQTQ0cHRYeDLy7e/+TqoI/s9O0YjYTxKfHt7cXpNh5IZGpcCY7qGI8HUTme/UH15qtOpc8W+ceYSwWi15LQwPvZ7Glk+mtu5lymktTu53x8kD7kg8qYOp+rwf3EVhEqUA2/eqacndumuFMKI4Iyw8GdlCtfUo83TB9EhkJWvxxX8j8Ph7yRz8zzAypXu4N0K9dVh5NF2S/d9JLDz6Tgo+fzkCpw+7i4SjR95O5TXHSaH23Kej3XBspzuO7/EQnfYutGh27leGDn3K0KZ7eW1VBGeHhScja8i9TmVOPnkmEslxOnz9RhgefSutxZl8DMPYF38PLSYOiMTaYxmSbLvZtnuFCPUh0ZlwzE2E/JAllLwwjONDrVVJZMbs9BDWTX35+YQgrZsRHfuUoXP/MnQeUMrhX41wZdVFm8jChQvRunVr6HQ6DBo0CPv27bPFYRkngbL8yApBdcjICvH9nBBRmoJhHAlnWkd1aiVu6x2Gfq18IUXadKsQtYTdPI2iwgaVd6MsZcYxoa4+R7Z54Nu3QvHqhHb4fnYojmzzFKKTyhddPbYAU99NwVu/xePB19Ix+OZCFp32FJ6//PILnnnmGbz22ms4dOgQevXqhdGjRyMrK8vah2aciKhOejz4eiqU6moc3eGJ5Z8GiuK7DOMIOOM6SoXZh3UMxOhuIVApXCSZcPTEPGqnWYXUeB0WPheBolwWn46C0QjE7nfD97NDhNj89u0wHNnuKVpJBoQZcP09uZi+IBGzvj+LO6dloVP/MqjU9p61PLB6y0y6Mx8wYAA++eQT8by6uhqRkZGYNm0aXnzxxUv+LrfMZJpL9DYPLH0nVHR0GPtoNq4bn2/vKTE2xhFbZtpjHZUSGYUV+OtIKkr10nNlZCWrRamlwhw1/MMMmPpuqqixyMgPUkNUuP3gRi8c2uxZL2bTL7gSvYYXo8/wYoS318uyfmZzGNE5CL0ifZq8v2RaZhoMBhw8eBAv1Wn2qVAocP3112P37t3WPDTjpPQeVoKCrGz89XkQ/v4iED4BVehznTRLtDBMU+B1FAjx1mHigCgsP5wquaQjqiH81P+Shfik2ozz/y9SJByRF4aRByWFCiE29633Fi1Szbh7V6HPtcXoN6IYUZ0rHF5s2gqrCs+cnBwYjUYEBwfX207PT548edH+er1ejLoKmmGay/A7C5Cfrcb2P33x47xgePpXoX3PcntPi2FaBK+jNXi7qkXS0V9H0pCaL63vs39oFZ7+MBlfzApH6hkdFj0XiftmpaPrQOklRzE1UB5A3CE37FvrjWO7PWCsqlGVKnU1ug0uRf+RRSJBSNkClZR8KgZ/fzEXYx99HpEde1h+8jLHJslFTWX27NnCVGse5EpimOZCd6W3TclGj6HFMFYq8PXrYchIvLh4L8M4Io68jlLS0R19wtEx2BNSw8vfiCfnJaNTv1IY9Ap89WoY9qyRZ2iDo3cQWrfUD2/f1wZfvBIh4jZJdEZ2rMCd0zLx+s8JuP+/6UJ8tkR0Evs3rMCZI3txYOMKS0/fIbCq8AwICIBSqURmZma97fQ8JCTkov3JlUTxAeaRnJxszekxDgx1f5g0MwOtu5ajvEQpSp9w4D8jR3gdrY9KqcBNPUIkmfFO9Typo9GAUYWornbBrx+GYO13/pzoKAHrZuw+N3z1WhjemtwG65YGoCBbLaoSXDMuH88uPocZnyTh6rGFLS6LlZeZiuRTx5By+jiit64S2w5vWS2e03Z6nbGBq12j0aBfv37YtGkTxo0bVxsUT8+feuqpi/bXarViMIwloP62D72RigUzokR3oy9fDRcWCe40wsgJXkcbz3h316qw/XS2pIQdWcnufi4TPoFV2PCjP9Z/74/MJA3ufjaD1x4bQyWu9q71wu5VPsjLPJ9y3rZHmSh51HNoCdQWqvn89uQRDRw/Fx88eUft8w/Wx1nkWHLH6q52KgHyxRdf4Ntvv0VsbCwef/xxlJaW4sEHH7T2oRlGtNB89O1UESSeclqHH+aEco1PRnbwOtowZPUc1TUYCollfdB0ug/ZgsCId6FQVouaj/OnRyEnTS3i/xY9f594ZCwP3YSci9Xhh/dC8MakNlj1VaAQna4eRlxzez5e+OIcnvpfikgYspToJCbNnAuFsmGvGm2n1xkbdS6aOHEisrOz8eqrryIjIwO9e/fG2rVrLwqUZxhrBWVTaRPqqbzohQgRRP73lwG4bWqORd6bYWwBr6ON0y3MG1qVEmti0iVVRo/i/LJTlqLnNUacPf6SyJb+cFoU2nZfVhv/x4knlsOgd8HhzZ7Y+ZcPUs7oardHdKjA0FsL0Ht4MTQ6630++o28FcFR7epZOM1Mn/8bIjp0s9qx5YZNWmaSO6ghlxDDXCoo25KLMnUauee5TNFxYuvvfkKMUjwPw8gFXkcbp32QB8b1CRcZ74Yq+7WupDi+0sJ8EQpgjvNLiPkYYx9pjfU/jERuehiO736GahXg0OYlGDDqdlApbXdvX/gFh9tt3nImN12FXSt9sHetN8qKlbWZ6VQG6epbC+xS1or+/vR3NT8y9eFe7YzdaWixpqBsSy/Kfa8rRm6aGmu+DcAfC4NEQeAuA8sscAYMw9ibSD83jO8XgT8Pp6LMYJ94msbi/H6adx9F6wKgBgCPAngPpYXD8cGTD1GamNiP4/+aTnU1cOqQm7BuntjrLhqGELSmDxlbgIGjC0WYla3x8PGHp28AfAJDMWjMXdi7dhkKstPFdsaGnYuuBO5c5Bw8c0Ony+5jqUWZPu0/zwvG/g3eorf7tA+TENZWWgWpmSvDETsXXQmOfn4XQgXmfz+YghK97XtlH9z0F36a9yKqqd9io0wB8DHlwAPIhotiCu59/mrhqmUuTXmJAvvWe2HX3z7ITj1fIo9KWJE7vcvAUlHRxJ5UGQxQqtW11k5jZSVUGvmV8xthxc5FkqrjyTgntgzKpqD/8dMz0a5XGfTlCnw5i8ssMYwj4eeuwYT+kfBytX3jbBKPFM/XEOfXsc9pT2rwCyAQpuo/EX/0EejLpZUgJSXSEjT49aMgvHFPW6z4NEiITp1bTSmkF5ecxdTZqaLupr1FJ0Eik0QnQY9yFJ3Whl3tjN2xdVC2Sg08MCtNZJlSmaUlr4XjiblcZolhHAVvNzUm9I8Qls/8Mvv0TW8szq/m+QkAVwF4C3B5DnvW+CD+qBvGz8jkLmv/UmlwwZFtHiJ+89yJ8x6MkNZ6Yd3sN7KI12yZwhZPRlLUvVO0Ju5e/5ZZ8jIi+ZQOP74fIuKGnBEu78I4Ip46Ncb3j0SAp9YucX4RHbrjrqffEI/0PDC8db3tkR07wtP3fUx+KQY+AZXCiketNr+fE+LUXpjsVDX++jxAWDd/fD9UiE6F0oRew4rxxLxkPP9ZIobcUsiiU8awxZORBPYIyqbM9gdfT8XimRGI2emJVUsqMfZR5yuzZK1KAgxjb6jAvDnhKKOwwibH9AkMwaylm2vj/AbfPLE2zq+x7Z36JWL11wHYvcobh/7xwvE97hhzXy6G3laARqKQ7IK1epAbKlxwdIeHyEwny68Z36BKXHVTIQaNLhQtSRnHgIUnIwkutVhbk7bdK3D3M5n44b1QbP7NDwHhlRh8k+OXWbJVJQGGkUR/977hWH44FWkFthGfddetunF+jW2nNo13PZ2FQWMKsWxBMJLjdCKWkYTY2Eey0XlAmYhPd6SbVIo+SDmtxb513jj4jycqSmsUtovChE79ynD1LdJIFrKF8HY2WHgykqGxRdna9BtZLDqKUP/e3+fXlFmihc+R4fZujDNBBeZv7xOBFdGpSMmXbgxlZEc9/u/jJCE4V38VgIxzWnzx3whEdqzAqEm56HZVqc0FqKVvUqnuJll1SWxmJZ8Pg6B1l8ogDbihCL5Btq9I0BTYO2QZWHgyDIAb/pOHnDQNDm7ywrdvheKpD5MR1sZxyyxRhm1jZV+oksA9z82xy7wYxlpoVApRZP7vI2lIzJXujaVCAeF16Tm0GJt+9hOlgygO/avXwhHWlgRoHnpcXSL2k8tNamGuEsd2euDQZi+cPX4+UUilqUb3wSW46sYitO9dZrNzag7sHbI8XMeTYf6lyuCCT18KR0KMGzx9y+AfOgG3P/Ggw97Zppw+3mAlgWcW/iHr9m5ju0eifSjX8XSW82suVcZqrIpJR0J2KeRASYESW373FcXSqQSc2To44IZCDBhVBL+QKrvVJjXfpDZUgzQrRY1juzwQs9MDibHnv48uLiZ06F2GviOL0fPqEujcpZ3Vacs6085Sx5MtngzzLyqNCQ++loYF06OQleKG4vw3sXftYocVno7W3q3aCGz8yQ8fbtEi+jDg52fvGTFSRKVU4JaeYVgdk44zWSWQOh4+RtzycA6uG5+HbX/4YvsKH+Rl1oQG0SARRyKUrKDWyPRuarm7ilIF4o+64lS0G04ddENmUv1qAlGdy9F7WAl6X1sMnwBputIbgr1DloeFJ8Nc4FK55dF4fP16X5hMfbF37XgMvOEEXBTVDudScaT2bvlZKvwwJwQJx2oyYn/4AZg2zd6zYqSKUuGCm3uEYu3xDMRlFEMOUAm4Gx/Ixci784QVkTr4nD7sjtPRbmIoVSa07lqOjn3LxIjsUGHxxJzzN6cUaBqC+JgQxOz0F8dPOqlDdfX5AFQqgdS+V5kQxFTcXU5i0551pp0BFp4M02As0wAAW2CsugEfPb0IwJMO51KxVyUBS0NFpn/9KBjlJUrRBvX9D6swbaq8zoGxPQqFC8Z0C4HCBYhNl4f4JDQ6k0iIpJGXocL+DV44sNELuekaUYqIxppvAFcPI1p1qUBwlB7BUQYxgiINQsA2FTLyFWarkJPeFjq3J6HR9YCb1whkJfui2hiAFZ/W398/zCAssB371IhfytZ3JBzFO2RvWHgyTIMulf20BcDvAJ6Ai0sS7n0hAI6GvSoJWAJqMbicSs6s8RbPozqV4z8vZeDO20LsPTVGRuJztBCfLjieVgS5QfGdoyfn/ZscqcapQ25ikAWSbsRO7ncXoy7UapLiKmm4ikejCDOq1CtEt6DKChdUGhSiL3phrgqmWivmJ6goA4ryUFv2iMRsRPsKtOtZjg59yuAXLE+rpjN5h6QAC0+GadSlspxCywF8BJNpDkymdADysYw4eu/m794JFeVYKFmB3I+jJ+dCySsa00zohmtU12Dhfj+aIs8avlRiKTC8EoHhhbh6bKGwVKac1iH1jBaZyRpkJWmQmaRBQbYaFWVKMZDdtPdWqqvhG1glEpp8Q6oQ3q4CEe31CGurF9ZXZ8BRvENSgZdphrmkS2U+TCaK63weP/8vRLTY7DJQuqVYHB3yblF3l+WLA1FVqYCXfxUmzUxHh97Src3IyOP7PrJLsLCARicVQO5Qt6NWnSvEqEtFmQuK8lQiEYgKtpNVs7xUAWOVC9RaE9SaamjoUWsSYSvUOcjD1yjJMke2Rs7eIanBwpNhLuNSyc+aizbd7kPMzmB8+1YYHn8/RcROMbaFLpC/fRSM6K2e4nmXgSW45/kMeHg7VhwZYz+u6xQElcIFB87lwxHRuZmgc6u09zQYJ4eFJ8M0waXioijEl696IO6AO76cFY5pHyYhKJIXcFuRfEorXOuUQEHZslReZtgd+WyJYSzONR0Chdt9b8K/wYwMw1gUXrYZ5gLIhUKis65LhWIHH5iVhshOFSgtUuKzlyJQkMP3bbZwrW/70wfzp0cJ0ekbXIlpHyTj2rtYdDLWY0i7AFzd3vGSCRlGCvDSzTBNhIozP/pWKgLDDcjPUuOzl8JFVxHGOlD82TdvhmL54iARg9bj6mI8uyiRwxwYmzCwjR+GdQy09zQYxuFg4ckwzewiMnV2ikhqyUzUCvFZVsxfI0uTFKfF/56IQsxOT1EYe9zjWXjg1XSHqwvISJt+rXxF68B/HSAMw1gAvmIyTAtq5z3+XjI8fKqQGq/D5y+HiyxRxnKu9QUzopCXoYFfiEHE0w67vYAv/oxdoH7VVG6Jan0yDHPl8NWSYVpAcFQlHnsvBW6eRiTFueKLWWGioDnTcshy/PUbYRe41pMQ1Ulv76kxTk63MG/c1CNEJB1JjeRTMVj0/H3ikWHkAAtPhmkhYW0MeGxOiuj8cfaYG756LRwGvfQuTHLgXKwO/3uiFY7t8hAFq82udVcPdq0z0qBDsCfG9gqDWimt7/j+DStw5sheHNi4wt5TYZgmwcKTYa6AiA56THk3VRRbpjZ1X78eBkOFtC5MUqa6Gtj8my8+eSYS+Zlq+Ica8PSHyexaZyRJmwB33NY7HBqVfS+deZmpSD51DCmnjyN66yqx7fCW1eI5bafXGUaqcD0YhrlCWnepwCNvpeLzV8IRd9BdxHw+8laa6IXMNA5VBPhpXjBi93mI572HF2P89EzRP5phpEqknxvu7BuB5dGpKDcY7TKHtyePuGhbSUFunXa/wAfr42w8K4ZpGmzxZBgL0K5nuch217kZkXDMDYtnRqCkkL9ejRF30A1zH2slRKdKU43x/5eJyS+ns+hkZEGItw4T+kfCU2cf282kmXOhoL6YDUDb6XXGvlAHLD93DaL83MTNSoSvK8J9XRHmo0OQl9buVnN7whZPhrEQbbtX4Im5Kfjs5XAkn9Jh4XORIgbU298+VhEpUlUJrP46AFuW+Ynnwa30mPxSOsLaGuw9NYZpFiQqJgyIxPLDqcgtse3nt9/IWxEc1a6ehdPM9Pm/IaJDN5vOx9lx1SjR2t8NYT6u8HHVwNtNDS+dqrYRSWOU6KuQX2pAfplBfIaS88ts/lmyByw8GcbCMZ9P/S9FWDypzifFLpL49A+tgrOTlazG97NDkXJGJ55fPbYAY6dkQ6M12XtqDNMivHRqjO8XiRXRqUgvtE9jAxI3JpOp9pGxPqQnQ7x0aOXvLuJ+g720lxWZDeGhVYlBFlEzRRWVSMotw7ncUiTllUFf6XheIBaeDGNhgqMMmPZBEj59MUK0eaSalA+/mYrIjs5ZFoiuhbtWeuPvzwNh0Cvg7mXExGcz0H1wqb2nxjAWsXbd0TcCq2LScC6nzGbH9fDxh6dvAHwCQzFozF3Yu3YZCrLTxXbGen/rrqFe6BnhDR83jdVuZrqHe4thrDbhbE4pTqQX4Wx2Kaod5MaChSfDWAGycD71QbLo6Z5xTosFz0Tinucy0OfaEjgT+Vkq/PK/YJw67C6ed+hTintfyODwA8ahoHi9W3uFY8OJTMSmF9nkmD6BIZi1dDOUarWwtg2+eSKMlZVQaawjiJyZcB9X9IjwRocgD6iUtovNVCpc0D7IQ4wyQxVi04uFCM0plrcRg4Unw1gJEldUGmjp7BCRRLP03TBkJuXihv/kQuHgceV0Y75vnRdWfBqIijIl1Npq3PxQDobeVuDw5844JyQSxnQPgZerCnsT8mxyzLoik8Qni07L0jrADVe19Ueot6u9pwI3jUq0cKWRnFeGQ0n5whoqRyMoC0+GsSJUUunhN9KwcklNQs367/2RmajBPc9nQKOT4YrRBApyVFj2cRBO7K0pk9S6aznufi4DQRGV9p4aw1idIe0C4O2qxqbYLOEqZeSHlARnQ1BMKA1KTCIBSlb2SqN8PmssPJ0cchEFeGjEQqlTK+GqVoo7K1eNQtxBVxqrYag6PygLL7NYj7wSg8PEm1gbhRK4dUqOiP1cNj8YR7Z7IiddLcoHOZIYMxqBHct9sPa7AOjLFaID0Y335+LaO/PFvwHDOFOLTYrV+/tomkMmhzgqVPpoSHvpCs4L8XXXYGSXYHGzczgpH9EpBbL4vFlNeL7zzjtYtWoVoqOjodFoUFBQYK1DMU1E4eKCUG+dqEFHdcSCPHXwdauJD2ouJEizi/XILKpAakG5CHyu4rv7SzJoTBECwyvx9ZuhSD2jwwePt8KtU7Mx+OZC2XfpoZaXZOVMS9DVWjknTM9ESGvHLw1ibXgtlSdkkZrYPxLLo9NQVO44N5iOiL+HBtd0CBQZ6nJNehrSPgB9W/kiOrkAh5MKUFFpdD7haTAYMH78eAwePBhLliyx1mGYy0ClGlr5u6F1gLu4myOrpiVQKxWiZhmNPlG+4kN+OrMEsRlFSCsol2XciS1o26Mczy5Kwk9zQ0SLTbKAntjrjonPZMLTV7oLRWOUFilEXc49q71hMrnA1dOIWx7OwaAxhRzLaSF4LZUv/h5a3DswCqti0kVcHiMt3LVK4VLvHuYNhcJF/qFd6prz6RPlgyPJhcINb6/uWnYRnm+88YZ4/Oabb6x1CKYRqJtGx2BPdArxRLBXjQXKFh94yvqjUVheieOphTicXCDc80x9fAKrMHVOCrb/6YOVXwWIWMi5U3RCfHZrZomh5FMx+PuLuRj76POI7NgDtoL60W/70wf//OqHitKam5kBowox9tEcePhIb6GTM7yWOkC5pT7h2Ho6G9FJbK2WSlchsg72b+0Lrcrx4oC0KiUGtvFDr0hv8Zk7JDELqKRiPPV6vRhmiopsU5bCESDh1zHYQwhOas3VEve5paB4UTL7947ywb6zeYhJKWQ3/AWQNXD4nQXo0LcMP8wJRfpZLZa8Fo5uV5XglkeyERzVNNfc/g0rcObIXhzYuMImwtNYBexd6y2SpIryapaP0DZ63P5kFtr3LLf68ZnLw+uo9CBr2nWdghDkqcU/sVm8HtqRdkEeGN4hUHQXcnS0KiUGtfVHr0gf4X4/nJwviRhQSQnP2bNn197dM5eHtGWkr5uwMrYL9BDlPKQEJSld2ylI3Fnuic8VNcg4Iak+YW0MmLEgCWu+9cfW331xfI8HYve5Y/AthRg9OQce3hcvEnmZqSgtzBc3F9FbV4lth7esxoBRt4vOJe7evvALDrd4q8vorZ7Y8IM/slNrSrb4BVdizP056DuimN3qEoLXUWknHfm7a7HyaBqKK7ibma3jOId3DBTdhpzRMDW4XY0L/lBivt29kS6mZvTYevHFF/Hee+9dcp/Y2Fh07ty59jm5h6ZPn96kgPiG7tQjIyNRWFgILy+vpk4TCzaddug7Sorb7BrmJeJS5HTXRslI645niEem4ZaSf38ZiOO7a8oQ6dyNuP6ePAy5pQA6t/Of52du6HTZ9/pgfZxF5lScr8TuVd7YudIHxf9aOD28qzBqUh4G31QIlUZ63zPqn00Fn5sKrTPe3t7NXmeuBGuupZZaRxnrQXF3609kICGbu3dZG61aIeIee0f4OEQcp6U+fwcS83A0pbBRATqic5CwlFpjHW2WxfPZZ5/FAw88cMl92rZti5ai1WrFYBrPkuz1r3VTjl+gQE8t7hkYhd3xueJDz8bP+gRFVoqan6ejXfHX54Ei833ll4HCythvZBEG31yA8HYGTJo5Fz/NexHVVL/oAhRKJe55bs4VzYP+LsmntNi10geH/vFEVWWNOdPLvwpDby3A0Nvy6wlhpvlYcy3ldVQecZ+39goTsXc7z+RwvU8rVXHpEeGFwW0DxL83cx7696AsfipGv/9cPmJSCmxaB7RZwjMwMFAMxrZ1Nsm62TPcW2RIyh0KBxjaIQBtAt2x7liGSERi6tOhdzlmfJKEgxu9sOkXX2Ql14hAGlSmaPDNkzB1dmcsfmHsRb87ff5viOjQrdnHrDYCZ4+74uhOD8Ts8EBB9nlLemSnCgy7PR+9rimGSj4GdknDaylDoTJ04aeY/NUx6Sgo47XQkkYacquTsYO5dDgc/Tv1b+WLg4n5iElt3AIqixjPpKQk5OXliUej0Shq0BHt27eHh0eNK5FpHCrqTmbuziFeQnw6GuQK/c9VrbAlLgvH0zj54UIoZnLADUXoP6oIZ464CtEZs9MD5064igGXmwEcBLAFwD8AtpOzo8nvX16qQGq8FqlnaOhw8oAbSgrOLwcaXTW6DS7BNbcVoFWXCtnXGZUzvJY6NlR55N5BUdgSl40TvBZeEVSXmgwb7YM87T0VWeGuVWEYCdDWvjiUWIAjKQXSifFsDuRG+vbbby/avnnzZlx77bVWjb2Sa4wnuQbaBbmjV4SPuGNzFqjjwrZTOZx4dBmK8pQio/zgJk9hBa2PES4uWQht4wHfIBfhFjfXBdWXKVBRroC+zEX0Tc9OUSM3/eKezlSDs/tVJegxtAQd+5ZBo5Xn30MOMZ62XEulfn7Meaj39qbYTE48aiZumprsbfIMyjEMTWpQ6SWKA6XOSE2lOeuM1YSnJXAW4UnJQt3CvdAj3BueOrXTLrjkbuK6n02jKFeJuEManD3miTNH3JCT1vQFwoxvUCXC2+kR3r4CrbtWoH2vMiglVeeiZTia8LxSHP38HA19lVHEfVLih3SvztJArXRB3yhf9HPQepxywmrJRYxlkXuykCWhVmUTB0RiBbeXaxJe/kYMGFUuhjn7vCBbhaJclaivWSgelcJlr3Wrhs61WjxqXavhG1wjON29WOQzjNQgATWic7CoybzxRCbyOfazwQLw3cO9MaCNnzDcMPKC/2J2yCbrElpj3fRrhhnbGQjwoKz3SPx9JA1pBRX2no6sILd6jWudS1UxjCMQ4euGyYNbIzo5H3vP5kmi8LdUBCfFIjqrd9ARYOFpIyhzkQq9tw/0gErpeMlClsyyu7NvBP4+moZzOdzbmGEY54WqgPRr5Yeuod7YczZXdIFzxtJL5FKn4vssOB0DFp5WhFwAnUM9RaH35gTpOjskzMf2DMOqmHQusMwwjNNDnjJquUlF0HecyUF8dolTxH+6a5Ui2bZnhA/X4nQgWHha4Q61baC7uDtr5efm9LGbVyI+b/lXfMZnldh7OgzDMHaHDBhje4Uhp0SPA+fycSqz2CEtoFR/k5KGOoV4Sq4VNHPlsPC0EKHeOnQO9UKnYE++M7MQtODc3CMUa46l43Qmi0+GYRhzPPyY7iEY0t4fh5MKcMxGhb+t3dqSrp/UMCXUu+lVKRj5wcLzCvBxU4s7si4hXuxKt6L4vKl7KNa6ZCAuo9je02EYhpEMXjq16DwzqI2fWB9j04uQXiifxExqTBHl5ybEJuc/OA8sPFsQc9Ih2FPcmYU1o1Yg03IoXGFMtxDhUjrDbneGYZh66NRK0emORn6pAbEZRTiZXizJlsTUia+Vv5soodc2wIM9hE6IQwhPaiNXWXn+C+aqMKIKlot70agViPKt+aKQC8Act1lRIZ87S0fguvY+MFbqkWGzO3oXGKpdqCeQjY7HMAxzZZD3bUi7AAxu64/sEj2S88qQmFuGtIJyVBpNdvFaUcwmhaO19ncXFV7YsuncyFp4UtOljIwMFBTU7yvaw6cSV6o7yQWgUiigUrpApTDBxaUEhoISJFq3hSlzGTq6mRChMtosoN5oMiGtTIHkCgqlYAHKMIw8cHFxQZCnTgwqyVRlrBb1kdMKy0VyUk6xHgXllRbNjieR6e2qFuI3xEsnxGaItw5qFpqMowhPs+gMCgqCm5ub+KIR9KVqKRqlCzQqpXAHUO90RnqQ6KTuRiQKrYrJBIO+AuqcbAAGJFdc2B+dYRhGHpCVMcrfTQwzlcZq5JUakF9mQKm+pj93qaFKPJZXGoUoNdF/4rHm1pvc+jq1Aq7iUSlc5T4kNt00QnRyJRfGYYUnudfNotPf37/ea2pD896L7sboi6RTUYtB/tLIAa1Oi/zSSlRbWXyqtTrxWGnMRFqFid3uDMM4DHTtC/bSicEwtkK2wtMc00mWzpa23jLfuSmpoTUjKygMgqoK0J26tQ2fGq0OShcXaBQmlFez8GQYhmEYpxOeZszu9aZArnNh2VQrOebEAaC/Ibl2CsqsnLlZ+xlzvELNDMMwDGNLZC88L4fLv4VpKR6FhEpzhCojfbQqJbx0JhRVVNl7KgzDMAzDXAaHNfuplS7w1KkQ4KmFt6tGJAzJUXS2bt0aH330ERwFa5yPq0YFN64FxzAMwzCSxyGFp7+7Bn7uWrhpVJLNTE9OTsZDDz2EsLAwaDQatGrVCv/3f/+H3NxcODOvv/66uEGgoVKpEBAQgGHDhgmxqtc3Xq3AQ6uCVlX/47xz+zaEeLui8IJyWwzDMAzD2AeHFJ4tKU574MABjBgxQjxam4SEBPTv3x+nT5/GTz/9hDNnzuDTTz/Fpk2bMHjwYOTl5cGe1QKqq+3b87dbt25IT09HUlISNm/ejPHjx2P27NkYMmQIiosbbptJQtXLVS3qrjIMwzAMI00cUni2hO+++06InKVLl1r9WE8++aSwcq5fvx7Dhw9HVFQUbrzxRmzcuBGpqal45ZVX6u1PYuuee+6Bu7s7wsPDsXDhwnpF9MlKSO+h1WqFBfXpp5+ufZ2shM8995z4Pfr9QYMGYcuWLbWvf/PNN/Dx8cFff/2Frl27ivf48ssvodPpLirMTxZZEudmduzYgWuuuQaurq6IjIwUxy0tLa19PSsrC2PHjhWvt2nTBj/88EOT/n3I0hkSEiLOpUePHpg2bRq2bt2KY8eO4b333qvdj/5WJOA9PT3F/v+ZNAmG4gJh5U5KTMSdt4wW+3VqFSosn08//qh4/s/G9bh19Ah0jApBl9bh+M+EO3AuIaFJc2MYhmEYpuU4tfBMTEzEwYMHcejQIfzyyy9i288//yye03Z63dKQNXPdunV44oknhCCrC4mnSZMmibmQoDQzd+5c9OrVC4cPH8aLL74oBOCGDRvEa7///js+/PBDfPbZZ8KCunz5ciHWzDz11FPYvXu3OK+jR48K6+GYMWPEvmbKysqEoCPBefz4cTEHEqP03nUtoTQveo2Ij48X73PnnXeK96XXSIjS8cw88MADIqSABP2yZcuwaNEiIUZbQufOnYU4/+OPP+qV1Hrrrbdw5MgRcd7nzp3Dww89KMosRUREYMnSn8R+Ow8exdFTZ/H2nHk151taiqlPPo11m3fit79WQ6FQ4MH/TLS7pZdhGIZhHB2Hz2q/XKKLGXPiUXZ2Nvr161e7va4AtAQk+Og9u3Tp0uDrtD0/P1/Mg4rjE1dffbUQnETHjh2xc+dOITZHjRol3NEkWK+//nqo1Wph+Rw4cKDYl177+uuvxSNZDwmyfq5du1Zsf/fdd2sFHIlCErdm7r77bvz44494+OGHxXMKAyALKAlNglzfJEKnT58unnfo0AHz588XFtzFixeLY65Zswb79u3DgAEDxD5Llixp9LybKj7JSmyGYmTNtG3bVhyfjqUvL4Ovhyt8fP3EawEBgfD28and95bbbq/3vh8u/BTd2kYi7mQsunTt1uL5MQzDMAxzaZza4vn9998Lt25dgWl+pO30urVojqCluM8Ln8fGxoqfyYJZXl4uhNejjz6KP//8E1VVNaWFYmJihKWSxKqHh0ftILc1WSzNkNu/Z8+e9Y5BopJc8mlpaeI5uclvvvlmYQklyMpIbvq67zt69GhhNTx79qyYH/0b1hXxJBzNv9/Sf7O6lQnIKk2ufBLb5G4n0UuQ6BWt3NQNf7wT4s/gsYfuw8CeXdA+IggDenQW21NTkls8N4ZhGIZhLo9TWzxJXJEFrq44MrN371707dvX4sds3769EE8kzG6/vb7ljaDtvr6+CAwMbNL7UWxlXFyciA8l9zu58Mk1T+KypKQESqVSCDR6rAsJRTPk8r+w1BRZDtu1aydc9I8//rgQtCQ0zdB7T506tV48qRkSgqdOnYKloX8bihUlKJaUhC4NEsX070WCk54bDDU9U0l8NsR9E+9ERGQU/jd/EYJDQ4VYvvaqfqj89/cYhmEYhrEOTi0860JxfiRAzI/WgvrKk4ucXNszZsyoF+eZkZEhRNR9991XTwju2bOn3nvQ87oua3oPsvzRoMQlsiyStbNPnz7C4klxlZQE1BJhTvOheEn6dyGLpxkS5SdOnBBCuiFoDmR5JdFrdrWTQL4wYampnDx5UoQIvPTSS7XPqfTUnDlzhPgmLqxIQIlShIvp/N8zLy8XZ06fwrz5C3HVkKFi297dO1s0J4ZhGIZhmodTu9oJiqOkGEmyelJJI3qk5+b4SmvwySefiGxzss5t27ZNJOCQqCJBStnn77zzTr39Kabz/fffF1ZEymj/7bffRIIRQVZIip2kjG8q00ThASREqS4oudhJPJKQpaQccoFTzCXFZ65ateqy86TfpUQrms9dd91VK+SImTNnYteuXSKZKDo6WsSurlixoja5qFOnTiL5iKyiZD0mAfrII49clFDVECRYSYSTm58E9IIFC4QbvXfv3nj++edrraoUIkCv0XlTVj4lGtWF/g1IwG//Zx1yc7NRWlICHx9f+Pn54/tvvsLZ+Hjs2LoFr708s4l/OYZhGIZhrgSnF55kzaNsaBJHZpFEz2m7taBEHLLOUVzmhAkThEt7ypQpuO6660QGup9fTVKMmWeffVbsTxbMt99+Gx988IEQrQTFTH7xxRciAYniNMnl/vfffwvLKkFJRCQ86T1IDI4bNw779+8Xwu1ykDWTEpUoa92czW6GjkXufBLDZE2lub366qu1SUzmY9NzEo133HGHOMemCHrKrA8NDRVzvPbaa/Hrr78KS+f27dtrQwTItU6im0Q4lYEiy+e8eTVZ62ZIxL/xxht45eWX0b1dK7z0/Axhuf30q+9wJPowrh3cD6++/AJefWv2ZefEMAzDMMyV42KydNq2BSkqKoK3tzcKCwvh5eVV77WKigphwaOYP6o5yTCXo1RfhRJ983u6Vxr0SE1KxOF8NcqruTWn1JkwIBLhPpe3rDdlnXEEHP38GIaR1zrj9BZPxnlw16oaTThiGIZhGMb6sPBknAovnYrbajIMwzCMnWDhyTgVlGzk46qGgrUnwzAMw9gcFp6M06FUKODtqrH3NBiGYRjG6WDhyTglGpVCuN0ZhmEYhrEdLDwZp8VVo4IrJxsxDMMwjM1g4ck4NZ46FTRKx/oaJJ+KwaLn7xOPDMMwDCMlrHbFpSLsDz/8sKizSd1qqEj6a6+9VttHm2Gkkmzk7aqG0oGyjfZvWIEzR/biwMYV9p4Kc4XwOorF4DwAABbeSURBVMowjKNhtSA36qVNPc8/++wz0QGHWjo++uijKC0tvajDDMPYE4WiRnzmlxog2W4KlyEvMxWlhflCSEdvrWmHenjLagwYdTuoR4S7ty/8gsPtPU2mmfA6yjCMo2HTzkVz587F4sWLRW/tpsCdixhbUlFpRGF5pSw7Fz1zQ6fL7vPB+jg4A47euciS6yjDMIxDdy6iCV3Yh7wuer1eTL7ucEQeeOABYZl67LHHLnrtySefFK/RPoxtoa5GHlp5ZrpPmjkXCmXDopi20+uMY8DrKMMwcsZmwvPMmTNYsGABpk6d2ug+s2fPForZPCIjI+Go0Ln9/PPPKC8vr2fF/fHHHxEVFWXXuTkzcm2r2W/krXjw9YUNvnbPc3MQFNlWuOMZecPrKMMwTic8X3zxRWGRu9SguKS6pKamYsyYMRg/fryIT2qMl156SdzNm0dycjIclb59+4oLwh9//FG7jX4m0dmnT5/abRTfRRcSc3JBr169sGzZstrXjUZjveSDTp064eOPP653LLKejhs3TsSEhYaGwt/fX1hWKysvdiszNW011TLMdF8y62ILOvHDe8/jw6fuxNuTR9h8TkzD8DrKMIyz0my/4rPPPntZN3Dbtm1rf05LS8N1112HIUOG4PPPP7/k72m1WjFaCkWrlpXBLri5UYZ0837noYcewtdff41JkyaJ51999RUefPBBbNmypXYfEp3ff/89Pv30U3To0AHbtm3Df/7zHwQGBmL48OFCmEZEROC3334TgnLXrl2YMmWKEJgTJkyofZ/NmzeLbfRIVpOJEyeid+/el7yAOXtbzbwyA4zV8kk3uv3xV/Dn4ncadbeT5ZORBlJeRxmGYSQlPEnw0GgKdIdOi2W/fv2EwFIorGtFItHp4QG7UFICuLs373dIQJJ1IjExUTzfuXOncL+bhSfFar377rvYuHEjBg8eXHsx2rFjh8hyJeGpVqvxxhtv1L4nWT53796NX3/9tZ7w9PX1xSeffAKlUonOnTvj5ptvxqZNm1h4XiLTncRnfpl8ytZcc/t9iOzUE/OnT7zotenzf0NEh252mRcjr3WUYRjGmlgtk4IWy2uvvRatWrUSLt7s7Oza10JCQqx1WFlBFx4SgN98840oeUM/BwQE1L5OlsmysjKMGjWq3u9RDb+67viFCxcKa2lSUpKIGaXXyZpZl27dugnRaYasnzExXGD8UqiUNT3ds/R6yAWVWl1rtaXPlPmRkSe8jjIM42hYTXhu2LBBCCca5Aqui7UuhOTuJsujPaBjtwRytz/11FO1ArIuJf+ezKpVqxAeXr8Go9mVRhbS5557Dv/73/+EVdTT01OUW9m7d2+9/ckyWhcSJOSmZy7f092TMt1lUl/ew8cfnr4B8AkMxaAxd2Hv2mUoyE4X2xn5YY91lGEYRpbCk+KXbF0SiGIsm+vutjeULEAWShKCo0ePrvda165dhcAkSya51RuC3PMU9/XEE0/UbouPj7f6vJ0JrVoJnUoeme4+gSGYtXQzlGq1+EwNvnkijJWVUGk09p4aI5N1lGEYxprIs2ihA0Hu79jY2Nqf60LWS7JmzpgxQ1gnhw4dKrJUSWxSgdb7779fJBx99913WLdunYjvXLp0Kfbv3y9+Zixr+ewV6YM9icWQOnVFJolPFp0MwzCMVGDhKQEuVeX/rbfeErGglN1OnUp8fHxEKaaXX35ZvE71/A4fPiyy1Elk3HPPPcL6uWbNGhuegXPQJ8oXFdVKRCcX2HsqDMMwDCNLbNoys7lwy0xGCtT9rFHow5pjGYjLkL7l01lx9JaZzcXRz49hGPsj2ZaZDCN3RCxutxC08m9hNhnDMAzDODEsPBmmmSgVLhjbKwyRfiw+GYZhGKY5sPBkmBZALTVv6x2GKBafDMMwDNNkWHgyzBWKz9YBLD4ZhmEYpimw8GSYK+xuNLZnGNoEyKyALMMwDMPYARaeDGMJ8dkrDG0DWXwyDMMwzKVg4ckwFko4uqVnGLqGcbkahmEYhmkMFp4MY0HxSaWWBrfjvugMwzAM0xAsPBnGwlzV1h9juocIIcowDMMwzHlYeDow1JRqypQp8PPzE4XPo6Ojce2112L69OmX/L3WrVvjo48+stk8HZEuoV64o284dGqlvafCMAzDMJLBIXu1f7jhlE2PN2NUx2b/TkZGBt555x2sWrUKqampCAoKQu/evYUoHDlypEXmtXbtWnzzzTfYsmUL2rZti4CAAPzxxx9Qq9UWeX/m0kT4uuHuAZH4+2gacksM9p4OwzAMw9gdhxSeUufcuXO4+uqr4ePjg7lz56JHjx6orKzEunXr8OSTT+LkyZMWOU58fDxCQ0MxZMiQ2m1k/WRsh6+7BvcMjMLWuGzEpBbaezoMwzAMY1fY1W4HnnjiCeH63rdvH+6880507NgR3bp1wzPPPIM9e/aIfZKSknDbbbfBw8MDXl5emDBhAjIzM2vf4/XXXxcW0qVLlwrXuLe3N+6++24UFxeL1x944AFMmzZNvA8di/YhLnS1Z2VlYezYsXB1dUWbNm3www8/XDTfgoICPPLIIwgMDBRzGTFiBI4cOdLkuRDV1dV4//330b59e2i1WkRFRQmLr5nk5GRxjiTGSRzTuZNAd5RC89d3DcbNPUOhVfNXjmEYhnFe+CpoY/Ly8oQLnCyb7u4X130k4UUijYQX7bt161Zs2LABCQkJmDhx4kUWzeXLl2PlypVi0L5z5swRr3388cd48803ERERgfT0dOzfv7/B+ZBAJdG3efNmLFu2DIsWLRJitC7jx48X29asWYODBw+ib9++IhyA5teUuRAvvfSSeD5r1iycOHECP/74I4KDg8VrZO0dPXo0PD09sX37duzcuVMI7jFjxsBgcBwXdcdgT0wa1AphPjp7T4VhGIZh7AK72m3MmTNnRNJP586dG91n06ZNiImJwdmzZxEZGSm2fffdd8IqSgJywIABYhsJVIrhJMFGTJ48WfwuWRLJ6kjblUolQkJCGjzOqVOnhJgky6v5PZcsWYIuXbrU7rNjxw7xOglPslQS8+bNEyKThColL11uLmT5JCH8ySef4P777xevt2vXDkOHDhU///LLL+L3v/zyS2GdJb7++mshwik+9YYbboCj4O2qxvh+kTiYlI99Z/NgqKq295QcAvrYhHjp0MrfHf7uGntPh2EYhmkEFp42hkTn5YiNjRWC0yw6ia5duwohRq+ZRSK5tc1Cj6B4zgutlZc7jkqlQr9+/Wq3kSCm45ghl3pJSQn8/evXpiwvLxdWTjOXmgsdR6/XN5o0RccgQV7394mKiop6x3AUFAoXDGjth84hnthxOgcnM86HJDBNx02jRCt/NyE2W/u7w1XDFQQYhmGkDgtPG9OhQwdh1bNEAtGF2en0vmQ5tCQkOklEkuXxQuoK1EvNheJHL3cMEr8NxZdSXKmj4qlT48YeoegR4Y3NcdnIKdbbe0qSRqVwQZiPqxCbUX5uCPTU1lrIGYZhGHnAwtPGUOIMxTMuXLgQTz/99EVxnpTIQ65uirukYbZ6UlwkvUaWT0tB1s2qqioRt2m2osbFxYnjmKF4Tir9RJZRc4JSS8Q2iU9yvVOS0oXQMcjdTiWlKHnJGcsuTRoYhRPpRTiYmI+8UseJa70SqAA/uc8jfF3Fv1Goj04kajEMwzDyhVdxO0Ci02g0YuDAgfj9999x+vRp4Y6eP38+Bg8ejOuvv16UWJo0aRIOHTokYizvu+8+DB8+HP3797fYPDp16iQSeKZOnYq9e/cKAUrCsK6FkuZCcxo3bhzWr18vMs137dqFV155BQcOHGjScXQ6HWbOnIkXXnhBxKqS+5yy9ymelKDzpBqjlFBFyUUU20oWVhLmKSkpcAbI/d493Bv3DW6FW3uHIdz30lZiR0SjUghr5qC2frizbwQev7YdJgyIxJD2AYjyd2PRyTAM4wCwxdMOUDF3EpSUePPss8+KrHNyKZO7efHixcJ9uGLFClEOadiwYVAoFEIgLliwwOJzoSQeEpskainL/O233xaZ52ZoLqtXrxZC88EHH0R2drZIVqJ5mbPSmwK9J1lNX331VaSlpQn3/WOPPSZec3Nzw7Zt24Q4veOOO0QyUnh4uIgJdTYLKP17twv0ECOjsAKHk/KRkFPqcElIChcX+HloEOSpRai3DqHergjw0LDrnGEYxsFxMTUl28VOFBUViezswsLCiwQIJZ6QZYxqT5JFjWGshb0/a5XGaiTmluJ0ZoksRShZMv3cNSLbPMhLh2AvLQI8tJKxYF5qnXEEHP38GIaR1zrDFk+GkTgk0NoHeYpRRSI0rwxpBeVIL6xAVlEFKo32v3ckQ6WHViXKRfm4aeDrpq4Rmx5aeOlUbMlkGIZhBCw8GUZGqJSKWlc8Yaw2IadEL0RofpkBReWVKCyvFI+WEqTkFtepFaJ8katGJQQmiUkPnUpk5nvqagSnVCyYDMMwjHRh4ckwMs/8Dhbu64tDAEr1VSivNArXvL6qWjzSMDYQXaN0cYFa5SLEo1qhED9rVUq4qpVCdLLFkmEYhrEELDwZxkFx16rEYBiGYRipwL4xhmEYhmEYxibIXnhaulMPw1wIf8YYhmEYxjLI1g+n0WhEfUuqCUk1MOk5x6ExloQqjRkMBlG7lD5r9BljGIZhGMYJhScJAaqrSMXXSXwyjLWgAvdRUVHiM8cwDMMwjBMKT4IsUCQIqN84taBkGEujVCpFxyW2pjMMwzCMkwtPggSBWq0Wg2EYhmEYhpEuVvUd3nrrrcIiSW0GqTf35MmT2S3OMAzTDHgdZRjGkbCq8Lzuuuvw66+/Ii4uDr///jvi4+Nx1113WfOQDMMwDgWvowzDOBIuJkrdtRF//fUXxo0bB71e3yTXeHOazjMMw7QEua0zvI4yDCM1mrPO2CzGMy8vDz/88AOGDBnS6GJJCykNM3QC5hNiGIaxBub1xYb34C2G11GGYWS/jpqszAsvvGByc3OjmZiuuuoqU05OTqP7vvbaa2I/Hjx48LD1SE5ONkkVXkd58OABB1lHm+1qf/HFF/Hee+9dcp/Y2Fh07txZ/JyTkyPu0hMTE/HGG28IU+zKlSsbLE9z4Z06dYyh3/X3929yORtS3ZGRkUhOTpa9W8mRzsXRzofPxXHOhZbA4uJihIWF2axWK6+jtsORzsXRzofPxTnX0WYLT+rikpube8l92rZt22CXl5SUFHEyu3btwuDBg2ENHCmeyZHOxdHOh89FmsjlXHgdtR2OdC6Odj58Ls55Ls2O8aT2lDSupOd13btxhmEYZ4PXUYZhnBWrJRft3bsX+/fvx9ChQ+Hr6ytKgMyaNQvt2rWz2l06wzCMI8HrKMMwjobCmv2t//jjD4wcORKdOnXCww8/jJ49e2Lr1q3QarXWOqx479dee82qx7AVjnQujnY+fC7SxJHOheB19MpxpHNxtPPhc3HOc7FpHU+GYRiGYRjGebFNCifDMAzDMAzj9LDwZBiGYRiGYWwCC0+GYRiGYRjGJrDwZBiGYRiGYWyCLIXnwoUL0bp1a+h0OgwaNAj79u275P6//fab6ABC+/fo0QOrV6+GHM/liy++wDXXXCPKqtC4/vrrL3vuUv67mPn5559FR5Vx48ZBSjT3fAoKCvDkk08iNDRUZAN27NhRMp+15p7LRx99JLKoXV1dRbHyGTNmoKKiAvZm27ZtGDt2rOiOQZ+Z5cuXX/Z3tmzZgr59+4q/Sfv27fHNN9/YZK5Sh9dRaa6jjraW8jrK6+hFmGTGzz//bNJoNKavvvrKdPz4cdOjjz5q8vHxMWVmZja4/86dO01KpdL0/vvvm06cOGH673//a1Kr1aaYmBiT3M7l3nvvNS1cuNB0+PBhU2xsrOmBBx4weXt7m1JSUkxyOxczZ8+eNYWHh5uuueYa02233WaSCs09H71eb+rfv7/ppptuMu3YsUOc15YtW0zR0dEmuZ3LDz/8YNJqteKRzmPdunWm0NBQ04wZM0z2ZvXq1aZXXnnF9Mcff4i+wH/++ecl909ISBA9zp955hnx/V+wYIFYD9auXWtyZngdleY66mhrKa+jvI42hOyE58CBA01PPvlk7XOj0WgKCwszzZ49u8H9J0yYYLr55pvrbRs0aJBp6tSpJrmdy4VUVVWZPD09Td9++61JjudC8x8yZIjpyy+/NN1///2SWSxbcj6LFy82tW3b1mQwGExSo7nnQvuOGDGi3jZacK6++mqTlGjKgvnCCy+YunXrVm/bxIkTTaNHjzY5M7yOSnMddbS1lNdRXkcbQlaudoPBgIMHDwrXiBlqRk/Pd+/e3eDv0Pa6+xOjR49udH8pn8uFlJWVobKyEn5+fpDjubz55psICgoSRbGlREvO56+//hKdZMhFFBwcjO7du+Pdd9+F0WiE3M5lyJAh4nfMbqSEhATh6rrpppsgN6T6/bcnvI5Kcx11tLWU11FeR23eMtMa5OTkiA8gfSDrQs9PnjzZ4O9kZGQ0uD9tl9u5XMjMmTNFjMaFHwg5nMuOHTuwZMkSREdHQ2q05HxoUfnnn38wadIksbicOXMGTzzxhLigUQcIOZ3LvffeK36P2jTSDXFVVRUee+wxvPzyy5AbjX3/i4qKUF5eLmKvnA1eR6W5jjraWsrrKK+jjSEriydznjlz5ohA8j///FMEOsuJ4uJiTJ48WQT5BwQEwBGorq4WFofPP/8c/fr1w8SJE/HKK6/g008/hdygIHKyMixatAiHDh0SLRtXrVqFt956y95TYxiLIud11BHXUl5HnQNZWTzpi6VUKpGZmVlvOz0PCQlp8Hdoe3P2l/K5mJk3b55YMDdu3Cj6Ntub5p5LfHw8zp07J7Lq6i44hEqlQlxcHNq1awc5/W0oA1OtVovfM9OlSxdxp0huGo1GA7mcy6xZs8TF7JFHHhHPKYO5tLQUU6ZMERcBcjHJhca+/15eXk5p7SR4HZXmOupoaymvo7yONoZ8zhwQHzq6C9q0aVO9Lxk9p7iQhqDtdfcnNmzY0Oj+Uj4X4v333xd3TGvXrkX//v0hBZp7LlSSJSYmRriGzOPWW2/FddddJ36mshNy+9tcffXVwi1kXvSJU6dOiYXUXotlS8+FYt4uXBTNF4KaWHT5INXvvz3hdVSa66ijraW8jvI62igmmUElDahEwTfffCPS+qdMmSJKGmRkZIjXJ0+ebHrxxRfrlQFRqVSmefPmidIZr732mqTKgDTnXObMmSPKOSxbtsyUnp5eO4qLi01yO5cLkVImZkvOJykpSWTGPvXUU6a4uDjTypUrTUFBQaa3337bJLdzoe8InctPP/0kymisX7/e1K5dO5HZbG/os05lcGjQ8vXBBx+InxMTE8XrdB50PheWAXn++efF95/K6HA5JV5HpbqOOtpayusor6MNITvhSVANqaioKLF4UImDPXv21L42fPhw8cWry6+//mrq2LGj2J9KAqxatcokx3Np1aqV+JBcOOgDLse/i1QXy5aez65du0SJGVqcqCTIO++8I8qcSIHmnEtlZaXp9ddfF4ukTqczRUZGmp544glTfn6+yd5s3ry5we+Aef70SOdz4e/07t1bnDv9Xb7++ms7zV5a8DoqzXXU0dZSXkd5Hb0QF/pf4/ZQhmEYhmEYhrEMsorxZBiGYRiGYeQLC0+GYRiGYRjGJrDwZBiGYRiGYWwCC0+GYRiGYRjGJrDwZBiGYRiGYWwCC0+GYRiGYRjGJrDwZBiGYRiGYWwCC0+GYRiGYRjGJrDwZBiGYRiGYWwCC0+GYRiGYRjGJrDwZBiGYRiGYWwCC0+GYRiGYRgGtuD/AeTvqAOmMnA/AAAAAElFTkSuQmCC", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig = plt.figure(figsize=(8, 7))\n", "subfigs = fig.subfigures(2, 1)\n", "subfigs[0].suptitle('Shared noise')\n", "subfigs[1].suptitle('Task-specific noise')\n", "\n", "# Reduce the size of the training set to show the effect of independent noises\n", "# in low data settings\n", "N_train = 10\n", "TASK_NOISE = TASK_NOISES[0]\n", "full_train_x = torch.cat([train_x1[:N_train], train_x2[:N_train]])\n", "full_train_i = torch.cat([train_i_task1[:N_train], train_i_task2[:N_train]])\n", "full_train_f = torch.cat([train_f1[:N_train], train_f2[:N_train]])\n", "full_train_noise = torch.cat([TASK_NOISE * train_noise1[:N_train], TASK_NOISE * train_noise2[:N_train]])\n", "full_train_y = full_train_f + full_train_noise\n", "\n", "likelihoods = (gpytorch.likelihoods.GaussianLikelihood, gpytorch.likelihoods.HadamardGaussianLikelihood)\n", "for row, (subfig, likelihood_cls) in enumerate(zip(subfigs, likelihoods)):\n", " model, likelihood = train_model(\n", " ((full_train_x, full_train_i), full_train_y), \n", " likelihood_cls\n", " )\n", "\n", " with torch.no_grad(), gpytorch.settings.fast_pred_var():\n", " observed_pred_y1 = likelihood(model(test_x, test_i_task1), [test_i_task1])\n", " observed_pred_y2 = likelihood(model(test_x, test_i_task2), [test_i_task2])\n", "\n", " y1_ax, y2_ax = subfig.subplots(1, 2)\n", " train_x1_sub = train_x1[:N_train]\n", " train_x2_sub = train_x2[:N_train]\n", " train_y1_sub = train_f1[:N_train] + TASK_NOISE * train_noise1[:N_train]\n", " train_y2_sub = train_f2[:N_train] + TASK_NOISE * train_noise2[:N_train]\n", " ax_plot(y1_ax, train_y1_sub, train_x1_sub, observed_pred_y1, fr'Task 1 ($\\sigma_y^2 = {TASK_NOISE**2:.2f}$)')\n", " ax_plot(y2_ax, train_y2_sub, train_x2_sub, observed_pred_y2, fr'Task 2 ($\\sigma_y^2 = {TASK_NOISE**2:.2f}$)')\n", "\n", " # Print the standard deviation, sigma_y which should be \n", " # close to TASK_NOISE\n", " print(f\"{likelihood.noise=}\")" ] } ], "metadata": { "@webio": { "lastCommId": null, "lastKernelId": null }, "anaconda-cloud": {}, "kernelspec": { "display_name": "venv", "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": "3.11.9" } }, "nbformat": 4, "nbformat_minor": 1 }