diff --git a/semi-supervised/semi-supervised_learning_2.ipynb b/semi-supervised/semi-supervised_learning_2.ipynb new file mode 100644 index 0000000000..cb038affb6 --- /dev/null +++ b/semi-supervised/semi-supervised_learning_2.ipynb @@ -0,0 +1,694 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "In this notebook, we'll learn how to use GANs to do semi-supervised learning.\n", + "\n", + "In supervised learning, we have a training set of inputs $x$ and class labels $y$. We train a model that takes $x$ as input and gives $y$ as output.\n", + "\n", + "In semi-supervised learning, our goal is still to train a model that takes $x$ as input and generates $y$ as output. However, not all of our training examples have a label $y$. We need to develop an algorithm that is able to get better at classification by studying both labeled $(x, y)$ pairs and unlabeled $x$ examples.\n", + "\n", + "To do this for the SVHN dataset, we'll turn the GAN discriminator into an 11 class discriminator. It will recognize the 10 different classes of real SVHN digits, as well as an 11th class of fake images that come from the generator. The discriminator will get to train on real labeled images, real unlabeled images, and fake images. By drawing on three sources of data instead of just one, it will generalize to the test set much better than a traditional classifier trained on only one source of data." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "%matplotlib inline\n", + "\n", + "import pickle as pkl\n", + "import time\n", + "\n", + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "from scipy.io import loadmat\n", + "import tensorflow as tf\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "!mkdir data" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "from urllib.request import urlretrieve\n", + "from os.path import isfile, isdir\n", + "from tqdm import tqdm\n", + "\n", + "data_dir = 'data/'\n", + "\n", + "if not isdir(data_dir):\n", + " raise Exception(\"Data directory doesn't exist!\")\n", + "\n", + "class DLProgress(tqdm):\n", + " last_block = 0\n", + "\n", + " def hook(self, block_num=1, block_size=1, total_size=None):\n", + " self.total = total_size\n", + " self.update((block_num - self.last_block) * block_size)\n", + " self.last_block = block_num\n", + "\n", + "if not isfile(data_dir + \"train_32x32.mat\"):\n", + " with DLProgress(unit='B', unit_scale=True, miniters=1, desc='SVHN Training Set') as pbar:\n", + " urlretrieve(\n", + " 'http://ufldl.stanford.edu/housenumbers/train_32x32.mat',\n", + " data_dir + 'train_32x32.mat',\n", + " pbar.hook)\n", + "\n", + "if not isfile(data_dir + \"test_32x32.mat\"):\n", + " with DLProgress(unit='B', unit_scale=True, miniters=1, desc='SVHN Training Set') as pbar:\n", + " urlretrieve(\n", + " 'http://ufldl.stanford.edu/housenumbers/test_32x32.mat',\n", + " data_dir + 'test_32x32.mat',\n", + " pbar.hook)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "trainset = loadmat(data_dir + 'train_32x32.mat')\n", + "testset = loadmat(data_dir + 'test_32x32.mat')" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "idx = np.random.randint(0, trainset['X'].shape[3], size=36)\n", + "fig, axes = plt.subplots(6, 6, sharex=True, sharey=True, figsize=(5,5),)\n", + "for ii, ax in zip(idx, axes.flatten()):\n", + " ax.imshow(trainset['X'][:,:,:,ii], aspect='equal')\n", + " ax.xaxis.set_visible(False)\n", + " ax.yaxis.set_visible(False)\n", + "plt.subplots_adjust(wspace=0, hspace=0)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def scale(x, feature_range=(-1, 1)):\n", + " # scale to (0, 1)\n", + " x = ((x - x.min())/(255 - x.min()))\n", + " \n", + " # scale to feature_range\n", + " min, max = feature_range\n", + " x = x * (max - min) + min\n", + " return x" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "class Dataset:\n", + " def __init__(self, train, test, val_frac=0.5, shuffle=True, scale_func=None):\n", + " split_idx = int(len(test['y'])*(1 - val_frac))\n", + " self.test_x, self.valid_x = test['X'][:,:,:,:split_idx], test['X'][:,:,:,split_idx:]\n", + " self.test_y, self.valid_y = test['y'][:split_idx], test['y'][split_idx:]\n", + " self.train_x, self.train_y = train['X'], train['y']\n", + " # The SVHN dataset comes with lots of labels, but for the purpose of this exercise,\n", + " # we will pretend that there are only 1000.\n", + " # We use this mask to say which labels we will allow ourselves to use.\n", + " self.label_mask = np.zeros_like(self.train_y)\n", + " self.label_mask[0:1000] = 1\n", + " \n", + " self.train_x = np.rollaxis(self.train_x, 3)\n", + " self.valid_x = np.rollaxis(self.valid_x, 3)\n", + " self.test_x = np.rollaxis(self.test_x, 3)\n", + " \n", + " if scale_func is None:\n", + " self.scaler = scale\n", + " else:\n", + " self.scaler = scale_func\n", + " self.train_x = self.scaler(self.train_x)\n", + " self.valid_x = self.scaler(self.valid_x)\n", + " self.test_x = self.scaler(self.test_x)\n", + " self.shuffle = shuffle\n", + " \n", + " def batches(self, batch_size, which_set=\"train\"):\n", + " x_name = which_set + \"_x\"\n", + " y_name = which_set + \"_y\"\n", + " \n", + " num_examples = len(getattr(dataset, y_name))\n", + " if self.shuffle:\n", + " idx = np.arange(num_examples)\n", + " np.random.shuffle(idx)\n", + " setattr(dataset, x_name, getattr(dataset, x_name)[idx])\n", + " setattr(dataset, y_name, getattr(dataset, y_name)[idx])\n", + " if which_set == \"train\":\n", + " dataset.label_mask = dataset.label_mask[idx]\n", + " \n", + " dataset_x = getattr(dataset, x_name)\n", + " dataset_y = getattr(dataset, y_name)\n", + " for ii in range(0, num_examples, batch_size):\n", + " x = dataset_x[ii:ii+batch_size]\n", + " y = dataset_y[ii:ii+batch_size]\n", + " \n", + " if which_set == \"train\":\n", + " # When we use the data for training, we need to include\n", + " # the label mask, so we can pretend we don't have access\n", + " # to some of the labels, as an exercise of our semi-supervised\n", + " # learning ability\n", + " yield x, y, self.label_mask[ii:ii+batch_size]\n", + " else:\n", + " yield x, y" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_inputs(real_dim, z_dim):\n", + " inputs_real = tf.placeholder(tf.float32, (None, *real_dim), name='input_real')\n", + " inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')\n", + " y = tf.placeholder(tf.int32, (None), name='y')\n", + " label_mask = tf.placeholder(tf.int32, (None), name='label_mask')\n", + " \n", + " return inputs_real, inputs_z, y, label_mask" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def generator(z, output_dim, reuse=False, alpha=0.2, training=True, size_mult=128):\n", + " with tf.variable_scope('generator', reuse=reuse):\n", + " # First fully connected layer\n", + " x1 = tf.layers.dense(z, 4 * 4 * size_mult * 4)\n", + " # Reshape it to start the convolutional stack\n", + " x1 = tf.reshape(x1, (-1, 4, 4, size_mult * 4))\n", + " x1 = tf.layers.batch_normalization(x1, training=training)\n", + " x1 = tf.maximum(alpha * x1, x1)\n", + " \n", + " x2 = tf.layers.conv2d_transpose(x1, size_mult * 2, 5, strides=2, padding='same')\n", + " x2 = tf.layers.batch_normalization(x2, training=training)\n", + " x2 = tf.maximum(alpha * x2, x2)\n", + " \n", + " x3 = tf.layers.conv2d_transpose(x2, size_mult, 5, strides=2, padding='same')\n", + " x3 = tf.layers.batch_normalization(x3, training=training)\n", + " x3 = tf.maximum(alpha * x3, x3)\n", + " \n", + " # Output layer\n", + " logits = tf.layers.conv2d_transpose(x3, output_dim, 5, strides=2, padding='same')\n", + " \n", + " out = tf.tanh(logits)\n", + " \n", + " return out" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def discriminator(x, reuse=False, alpha=0.2, drop_rate=0., num_classes=10, size_mult=64):\n", + " with tf.variable_scope('discriminator', reuse=reuse):\n", + " x = tf.layers.dropout(x, rate=drop_rate/2.5)\n", + " \n", + " # Input layer is 32x32x3\n", + " x1 = tf.layers.conv2d(x, size_mult, 3, strides=2, padding='same')\n", + " relu1 = tf.maximum(alpha * x1, x1)\n", + " relu1 = tf.layers.dropout(relu1, rate=drop_rate)\n", + " \n", + " x2 = tf.layers.conv2d(relu1, size_mult, 3, strides=2, padding='same')\n", + " bn2 = tf.layers.batch_normalization(x2, training=True)\n", + " relu2 = tf.maximum(alpha * x2, x2)\n", + " \n", + " \n", + " x3 = tf.layers.conv2d(relu2, size_mult, 3, strides=2, padding='same')\n", + " bn3 = tf.layers.batch_normalization(x3, training=True)\n", + " relu3 = tf.maximum(alpha * bn3, bn3)\n", + " relu3 = tf.layers.dropout(relu3, rate=drop_rate)\n", + " \n", + " x4 = tf.layers.conv2d(relu3, 2 * size_mult, 3, strides=1, padding='same')\n", + " bn4 = tf.layers.batch_normalization(x4, training=True)\n", + " relu4 = tf.maximum(alpha * bn4, bn4)\n", + " \n", + " x5 = tf.layers.conv2d(relu4, 2 * size_mult, 3, strides=1, padding='same')\n", + " bn5 = tf.layers.batch_normalization(x5, training=True)\n", + " relu5 = tf.maximum(alpha * bn5, bn5)\n", + " \n", + " x6 = tf.layers.conv2d(relu5, 2 * size_mult, 3, strides=2, padding='same')\n", + " bn6 = tf.layers.batch_normalization(x6, training=True)\n", + " relu6 = tf.maximum(alpha * bn6, bn6)\n", + " relu6 = tf.layers.dropout(relu6, rate=drop_rate)\n", + " \n", + " x7 = tf.layers.conv2d(relu5, 2 * size_mult, 3, strides=1, padding='valid')\n", + " # Don't use bn on this layer, because bn would set the mean of each feature\n", + " # to the bn mu parameter.\n", + " # This layer is used for the feature matching loss, which only works if\n", + " # the means can be different when the discriminator is run on the data than\n", + " # when the discriminator is run on the generator samples.\n", + " relu7 = tf.maximum(alpha * x7, x7)\n", + " \n", + " # Flatten it by global average pooling\n", + " features = raise NotImplementedError()\n", + " \n", + " # Set class_logits to be the inputs to a softmax distribution over the different classes\n", + " raise NotImplementedError()\n", + " \n", + " \n", + " # Set gan_logits such that P(input is real | input) = sigmoid(gan_logits).\n", + " # Keep in mind that class_logits gives you the probability distribution over all the real\n", + " # classes and the fake class. You need to work out how to transform this multiclass softmax\n", + " # distribution into a binary real-vs-fake decision that can be described with a sigmoid.\n", + " # Numerical stability is very important.\n", + " # You'll probably need to use this numerical stability trick:\n", + " # log sum_i exp a_i = m + log sum_i exp(a_i - m).\n", + " # This is numerically stable when m = max_i a_i.\n", + " # (It helps to think about what goes wrong when...\n", + " # 1. One value of a_i is very large\n", + " # 2. All the values of a_i are very negative\n", + " # This trick and this value of m fix both those cases, but the naive implementation and\n", + " # other values of m encounter various problems)\n", + " \n", + " raise NotImplementedError()\n", + " \n", + " return out, class_logits, gan_logits, features" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_loss(input_real, input_z, output_dim, y, num_classes, label_mask, alpha=0.2, drop_rate=0.):\n", + " \"\"\"\n", + " Get the loss for the discriminator and generator\n", + " :param input_real: Images from the real dataset\n", + " :param input_z: Z input\n", + " :param output_dim: The number of channels in the output image\n", + " :param y: Integer class labels\n", + " :param num_classes: The number of classes\n", + " :param alpha: The slope of the left half of leaky ReLU activation\n", + " :param drop_rate: The probability of dropping a hidden unit\n", + " :return: A tuple of (discriminator loss, generator loss)\n", + " \"\"\"\n", + " \n", + " \n", + " # These numbers multiply the size of each layer of the generator and the discriminator,\n", + " # respectively. You can reduce them to run your code faster for debugging purposes.\n", + " g_size_mult = 32\n", + " d_size_mult = 64\n", + " \n", + " # Here we run the generator and the discriminator\n", + " g_model = generator(input_z, output_dim, alpha=alpha, size_mult=g_size_mult)\n", + " d_on_data = discriminator(input_real, alpha=alpha, drop_rate=drop_rate, size_mult=d_size_mult)\n", + " d_model_real, class_logits_on_data, gan_logits_on_data, data_features = d_on_data\n", + " d_on_samples = discriminator(g_model, reuse=True, alpha=alpha, drop_rate=drop_rate, size_mult=d_size_mult)\n", + " d_model_fake, class_logits_on_samples, gan_logits_on_samples, sample_features = d_on_samples\n", + " \n", + " \n", + " # Here we compute `d_loss`, the loss for the discriminator.\n", + " # This should combine two different losses:\n", + " # 1. The loss for the GAN problem, where we minimize the cross-entropy for the binary\n", + " # real-vs-fake classification problem.\n", + " # 2. The loss for the SVHN digit classification problem, where we minimize the cross-entropy\n", + " # for the multi-class softmax. For this one we use the labels. Don't forget to ignore\n", + " # use `label_mask` to ignore the examples that we are pretending are unlabeled for the\n", + " # semi-supervised learning problem.\n", + " raise NotImplementedError()\n", + " \n", + " # Here we set `g_loss` to the \"feature matching\" loss invented by Tim Salimans at OpenAI.\n", + " # This loss consists of minimizing the absolute difference between the expected features\n", + " # on the data and the expected features on the generated samples.\n", + " # This loss works better for semi-supervised learning than the tradition GAN losses.\n", + " raise NotImplementedError()\n", + "\n", + " pred_class = tf.cast(tf.argmax(class_logits_on_data, 1), tf.int32)\n", + " eq = tf.equal(tf.squeeze(y), pred_class)\n", + " correct = tf.reduce_sum(tf.to_float(eq))\n", + " masked_correct = tf.reduce_sum(label_mask * tf.to_float(eq))\n", + " \n", + " return d_loss, g_loss, correct, masked_correct, g_model" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_opt(d_loss, g_loss, learning_rate, beta1):\n", + " \"\"\"\n", + " Get optimization operations\n", + " :param d_loss: Discriminator loss Tensor\n", + " :param g_loss: Generator loss Tensor\n", + " :param learning_rate: Learning Rate Placeholder\n", + " :param beta1: The exponential decay rate for the 1st moment in the optimizer\n", + " :return: A tuple of (discriminator training operation, generator training operation)\n", + " \"\"\"\n", + " # Get weights and biases to update. Get them separately for the discriminator and the generator\n", + " raise NotImplementedError()\n", + "\n", + " # Minimize both players' costs simultaneously\n", + " raise NotImplementedError()\n", + " shrink_lr = tf.assign(learning_rate, learning_rate * 0.9)\n", + " \n", + " return d_train_opt, g_train_opt, shrink_lr" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "class GAN:\n", + " \"\"\"\n", + " A GAN model.\n", + " :param real_size: The shape of the real data.\n", + " :param z_size: The number of entries in the z code vector.\n", + " :param learnin_rate: The learning rate to use for Adam.\n", + " :param num_classes: The number of classes to recognize.\n", + " :param alpha: The slope of the left half of the leaky ReLU activation\n", + " :param beta1: The beta1 parameter for Adam.\n", + " \"\"\"\n", + " def __init__(self, real_size, z_size, learning_rate, num_classes=10, alpha=0.2, beta1=0.5):\n", + " tf.reset_default_graph()\n", + " \n", + " self.learning_rate = tf.Variable(learning_rate, trainable=False)\n", + " inputs = model_inputs(real_size, z_size)\n", + " self.input_real, self.input_z, self.y, self.label_mask = inputs\n", + " self.drop_rate = tf.placeholder_with_default(.5, (), \"drop_rate\")\n", + " \n", + " loss_results = model_loss(self.input_real, self.input_z,\n", + " real_size[2], self.y, num_classes,\n", + " label_mask=self.label_mask,\n", + " alpha=0.2,\n", + " drop_rate=self.drop_rate)\n", + " self.d_loss, self.g_loss, self.correct, self.masked_correct, self.samples = loss_results\n", + " \n", + " self.d_opt, self.g_opt, self.shrink_lr = model_opt(self.d_loss, self.g_loss, self.learning_rate, beta1)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def view_samples(epoch, samples, nrows, ncols, figsize=(5,5)):\n", + " fig, axes = plt.subplots(figsize=figsize, nrows=nrows, ncols=ncols, \n", + " sharey=True, sharex=True)\n", + " for ax, img in zip(axes.flatten(), samples[epoch]):\n", + " ax.axis('off')\n", + " img = ((img - img.min())*255 / (img.max() - img.min())).astype(np.uint8)\n", + " ax.set_adjustable('box-forced')\n", + " im = ax.imshow(img)\n", + " \n", + " plt.subplots_adjust(wspace=0, hspace=0)\n", + " return fig, axes" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def train(net, dataset, epochs, batch_size, figsize=(5,5)):\n", + " \n", + " saver = tf.train.Saver()\n", + " sample_z = np.random.normal(0, 1, size=(50, z_size))\n", + "\n", + " samples, train_accuracies, test_accuracies = [], [], []\n", + " steps = 0\n", + "\n", + " with tf.Session() as sess:\n", + " sess.run(tf.global_variables_initializer())\n", + " for e in range(epochs):\n", + " print(\"Epoch\",e)\n", + " \n", + " t1e = time.time()\n", + " num_examples = 0\n", + " num_correct = 0\n", + " for x, y, label_mask in dataset.batches(batch_size):\n", + " assert 'int' in str(y.dtype)\n", + " steps += 1\n", + " num_examples += label_mask.sum()\n", + "\n", + " # Sample random noise for G\n", + " batch_z = np.random.normal(0, 1, size=(batch_size, z_size))\n", + "\n", + " # Run optimizers\n", + " t1 = time.time()\n", + " _, _, correct = sess.run([net.d_opt, net.g_opt, net.masked_correct],\n", + " feed_dict={net.input_real: x, net.input_z: batch_z,\n", + " net.y : y, net.label_mask : label_mask})\n", + " t2 = time.time()\n", + " num_correct += correct\n", + "\n", + " sess.run([net.shrink_lr])\n", + " \n", + " \n", + " train_accuracy = num_correct / float(num_examples)\n", + " \n", + " print(\"\\t\\tClassifier train accuracy: \", train_accuracy)\n", + " \n", + " num_examples = 0\n", + " num_correct = 0\n", + " for x, y in dataset.batches(batch_size, which_set=\"test\"):\n", + " assert 'int' in str(y.dtype)\n", + " num_examples += x.shape[0]\n", + "\n", + " correct, = sess.run([net.correct], feed_dict={net.input_real: x,\n", + " net.y : y,\n", + " net.drop_rate: 0.})\n", + " num_correct += correct\n", + " \n", + " test_accuracy = num_correct / float(num_examples)\n", + " print(\"\\t\\tClassifier test accuracy\", test_accuracy)\n", + " print(\"\\t\\tStep time: \", t2 - t1)\n", + " t2e = time.time()\n", + " print(\"\\t\\tEpoch time: \", t2e - t1e)\n", + " \n", + " \n", + " gen_samples = sess.run(\n", + " net.samples,\n", + " feed_dict={net.input_z: sample_z})\n", + " samples.append(gen_samples)\n", + " _ = view_samples(-1, samples, 5, 10, figsize=figsize)\n", + " plt.show()\n", + " \n", + " \n", + " # Save history of accuracies to view after training\n", + " train_accuracies.append(train_accuracy)\n", + " test_accuracies.append(test_accuracy)\n", + " \n", + "\n", + " saver.save(sess, './checkpoints/generator.ckpt')\n", + "\n", + " with open('samples.pkl', 'wb') as f:\n", + " pkl.dump(samples, f)\n", + " \n", + " return train_accuracies, test_accuracies, samples" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "!mkdir checkpoints" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "real_size = (32,32,3)\n", + "z_size = 100\n", + "learning_rate = 0.0003\n", + "\n", + "net = GAN(real_size, z_size, learning_rate)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true, + "scrolled": false + }, + "outputs": [], + "source": [ + "dataset = Dataset(trainset, testset)\n", + "\n", + "batch_size = 128\n", + "epochs = 25\n", + "train_accuracies, test_accuracies, samples = train(net,\n", + " dataset,\n", + " epochs,\n", + " batch_size,\n", + " figsize=(10,5))" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "fig, ax = plt.subplots()\n", + "plt.plot(train_accuracies, label='Train', alpha=0.5)\n", + "plt.plot(test_accuracies, label='Test', alpha=0.5)\n", + "plt.title(\"Accuracy\")\n", + "plt.legend()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "When you run the fully implemented semi-supervised GAN, you should usually find that the test accuracy peaks at 69-71%. It should definitely stay above 68% fairly consistently throughout the last several epochs of training.\n", + "\n", + "This is a little bit better than a [NIPS 2014 paper](https://arxiv.org/pdf/1406.5298.pdf) that got 64% accuracy on 1000-label SVHN with variational methods. However, we still have lost something by not using all the labels. If you re-run with all the labels included, you should obtain over 80% accuracy using this architecture (and other architectures that take longer to run can do much better)." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "_ = view_samples(-1, samples, 5, 10, figsize=(10,5))" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "!mkdir images" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "for ii in range(len(samples)):\n", + " fig, ax = view_samples(ii, samples, 5, 10, figsize=(10,5))\n", + " fig.savefig('images/samples_{:03d}.png'.format(ii))\n", + " plt.close()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true + }, + "source": [ + "Congratulations! You now know how to train a semi-supervised GAN. This exercise is stripped down to make it run faster and to make it simpler to implement. In the original work by Tim Salimans at OpenAI, a GAN using [more tricks and more runtime](https://arxiv.org/pdf/1606.03498.pdf) reaches over 94% accuracy using only 1,000 labeled examples." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.6.1" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/semi-supervised/semi-supervised_learning_2_solution.ipynb b/semi-supervised/semi-supervised_learning_2_solution.ipynb new file mode 100644 index 0000000000..844ee1757b --- /dev/null +++ b/semi-supervised/semi-supervised_learning_2_solution.ipynb @@ -0,0 +1,783 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "In this notebook, we'll learn how to use GANs to do semi-supervised learning.\n", + "\n", + "In supervised learning, we have a training set of inputs $x$ and class labels $y$. We train a model that takes $x$ as input and gives $y$ as output.\n", + "\n", + "In semi-supervised learning, our goal is still to train a model that takes $x$ as input and generates $y$ as output. However, not all of our training examples have a label $y$. We need to develop an algorithm that is able to get better at classification by studying both labeled $(x, y)$ pairs and unlabeled $x$ examples.\n", + "\n", + "To do this for the SVHN dataset, we'll turn the GAN discriminator into an 11 class discriminator. It will recognize the 10 different classes of real SVHN digits, as well as an 11th class of fake images that come from the generator. The discriminator will get to train on real labeled images, real unlabeled images, and fake images. By drawing on three sources of data instead of just one, it will generalize to the test set much better than a traditional classifier trained on only one source of data." + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "%matplotlib inline\n", + "\n", + "import pickle as pkl\n", + "import time\n", + "\n", + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "from scipy.io import loadmat\n", + "import tensorflow as tf\n", + "\n", + "# There are two ways of solving this problem.\n", + "# One is to have the matmul at the last layer output all 11 classes.\n", + "# The other is to output just 10 classes, and use a constant value of 0 for\n", + "# the logit for the last class. This still works because the softmax only needs\n", + "# n independent logits to specify a probability distribution over n + 1 categories.\n", + "# We implemented both solutions here.\n", + "extra_class = 0" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "mkdir: cannot create directory ‘data’: File exists\r\n" + ] + } + ], + "source": [ + "!mkdir data" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "from urllib.request import urlretrieve\n", + "from os.path import isfile, isdir\n", + "from tqdm import tqdm\n", + "\n", + "data_dir = 'data/'\n", + "\n", + "if not isdir(data_dir):\n", + " raise Exception(\"Data directory doesn't exist!\")\n", + "\n", + "class DLProgress(tqdm):\n", + " last_block = 0\n", + "\n", + " def hook(self, block_num=1, block_size=1, total_size=None):\n", + " self.total = total_size\n", + " self.update((block_num - self.last_block) * block_size)\n", + " self.last_block = block_num\n", + "\n", + "if not isfile(data_dir + \"train_32x32.mat\"):\n", + " with DLProgress(unit='B', unit_scale=True, miniters=1, desc='SVHN Training Set') as pbar:\n", + " urlretrieve(\n", + " 'http://ufldl.stanford.edu/housenumbers/train_32x32.mat',\n", + " data_dir + 'train_32x32.mat',\n", + " pbar.hook)\n", + "\n", + "if not isfile(data_dir + \"test_32x32.mat\"):\n", + " with DLProgress(unit='B', unit_scale=True, miniters=1, desc='SVHN Training Set') as pbar:\n", + " urlretrieve(\n", + " 'http://ufldl.stanford.edu/housenumbers/test_32x32.mat',\n", + " data_dir + 'test_32x32.mat',\n", + " pbar.hook)" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "trainset = loadmat(data_dir + 'train_32x32.mat')\n", + "testset = loadmat(data_dir + 'test_32x32.mat')" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAASwAAAElCAYAAABect+9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzsvUuPbUuSJvSZu6/H3rEj4rzvzZuZN7Oysh79QE1PoEE9\n6R4yQTACJiD+AHSPkLoHiAGiBtBCQgghJCRmjGBY9AwJMaORgKLpzMrK132dR5wTEfu11vKHMXBz\nc9/n3qw857ZKrdPanlLefSJ2rIcvX+Zmn332GTEzzuM8zuM8PoRh/llfwHmcx3mcx7uOs8E6j/M4\njw9mnA3WeZzHeXww42ywzuM8zuODGWeDdR7ncR4fzDgbrPM4j/P4YMbZYJ3HeZzHBzPOBus8zuM8\nPphxNljncR7n8cGM9zJYRPTHf1EX8hc5iOjlP+tr+LbjQ732D/W6gQ/32j/U6wbe3bbQ+5TmrMYV\nP3n6JP+j/TMC6OSbrL+mfDH5p8yIMQEAYowIIZ4cxJj8PWutfiYiENWDEZ2eSc+Y2guqn5kZ290B\nw9DpT5kZnFL93PwpEZ3cy8nsyBfJmHodnO+1zCMz6/USGT0Wo8yRzAVYD87Nb8s1lP9ut1s8evTo\na9eAt6fhrefR3n9q5lzvG5yvXY9HsDrnRn+eYkSMCe06MSbvc/k5GTkenTz1mCK+8/HHJ9d9utK+\n+Tn+xl8xkMr8N/Pzta8xn8w5mIFvXDMM+g3X8LOf/xn6YQCAk3s3AAzl+yVDsNbCOgsgzwV0STTr\n7xvuo1wPgU7WTYr5fQghgFMEod6v+YZLbWeUAJDtsbrY6DVaa/N1ATD0tZf0rYuS4xCj+DGcAO/z\nNS2LR4ihzoUldM4BALq+g7PyNwxwSohyLzFGlFeTmVGeTmJGKPcbI/a3N/fMfP2brrAM99u+0I6n\nT5/iP/wP/k4+oU9yc4B1Ri+YOCEmjySTY42Btfk0IQTc3+8AAG/e3OLN61u5KcC5DuO4AgA8fHCF\ncezzZHQGts8Hc47gnAFZmQAYyD0jLhEcQ74GMBLnX3AM+B//53+If+lf/msISSbfe/hplmvy8EuU\nvyM45/QhA6eGqDyEcRzh5GEleTjee7mXqC9x3/ew5YVOCcZYfUliTGq8U0pIskiMMei6Ls+dtfjj\nf/i/4N/79/9dmVvooi7nyNcWv2ZQynnCsmC7zXO+297jeDzm8wcPDgkhenmGhPU4AgCGYUAM+dq2\nt/e4v71H8EGPvbq4AABcbjYY15t8PCZ4pHIF+Nmvf45/8Ed/BC4vSeK6cIH60jIhob4yRObEvhSj\nFBMwz7PMizt5GZny8fNpkr6XIeQXrMwVoT5PQ4xOXmwQAWz0Wfzr/9a/gR/95b8KALi7u0ec8xyN\n1mLssyEbVgOuHlzhwaMHeS6uNjByPT56JHk3IjFYpoUTgZgAk9eOhUWSNeu9x2G7BQC8fvkC8/YO\nFvlZr3qLVZ//hgwQZZ6XuGiI5KzB3TTib/9r/ybWQ36OD66vcXV5BQDoB6cGy5hscKk8Lw55cSEb\nLGfyPR73wPOv7gAAv/71V3j15gaR8/WuNz2ePssb6Sff/QiPri/lHiOmwxH3t/ndvr3bYl7yaSYf\nEOXl3c8eN3dyv9s7/O//03//U7zDeC+DdToS1GegvNAA2SmYwWIwIoxa0mVe8Pr1DQDg+fOXePPm\nTm4yL8Jxtc7HQ8LDh9nYkulg0cl5ym6O+t/qIIDKAo4BtuwSZYexQDFDBgksiwHESCnoEdkzkngh\nrQdR/g3kRV0M1DzPmKZJ/83M6Ptev1c+97bLXoIYzRgCprkaOd3ViDDI7q4vmrxMBqQvhaVTlypx\nqp4dgCAG5/5+i7u7O/l8h0Ve+pQiwIwUgs7DPOWVtRpGIOTj391ucffmFsuSr9U4i2vZjQw5JFlC\nkRmBitGVZUUmvx2Ql6R4J2SqN5IAREZM+fiJI1iOn1L1AGOMarCcc7BdBxPrnJWnxK2HTASyNnsX\n+eFULxlAMW3EQJTnwsww1uDBw/yiT/OEnc4ZIcayVhysZdjyBlEEiRtkcigg54k6B0DesEjmx8AA\nnJ9n11ss81EvE8bAFA+LqydmmPR+LNnqPTbRiROvr+vqZ0MW1T4z6C1PtKx5oG7MHAzCnH++u9/j\n9vU92Mg1WVJngcjCyj3FyAg+4HjMc3bcTwgpnzgywQ75XR6GDv0gm+D+iHcd72mwGLpdAKBy8QSA\nJNQg2S81LCJECf222x1ub+XludvBL/mCnevAnDAd9wCAu7sOxskDNxt0vRgiZwAYEBfPJALJ6HmM\nrkuDJC9AsWgxxhM39RvvjoHIUV8s50iNlLX25OUpBmqaJuz3e/13a+CccxqCReRwrHzvcJzUCCzL\nAuYa5vTi6dRQpPzOqmeYiGFk/l1nMQw9rNy/DwEH8aS29/f46qvnAID9fqee8MXmApvNBrPM+f2b\nO+zv8ufVuMaFbB7GGBhjsSz5eD1ZXdxkHbb7/DfT4jGuxUN+/DT/zCeQrAsiA5i6dooXyQzElBDV\nQ4rQjZ9PQyY1cpQ9zaBhPYAmfCINiUxe4K0l12MzImrYG8Rwz8sCawwei+d0e/sGx3vx8K3VQ1mT\njVEK+XnstzWqMHaAKYbCsEYYzuW5LNeazymffcCLFy/lnLcYbN5YAeTrjPpCwarl6eo5y2YAwMgz\nNs7CujoX6rkyA+yRxFtKKeqmzRzVkC2zxXzMa3R7d8D93Q5uyPe1uVyjPBBjLIx4TpECYkhYZDOe\nl6AGK5GFHcWwdqa+K+8BS52zhOdxHufxwYz38rAYDdD3NQCvABQRCZzdXWTLH0O2trvtDrvtIX+N\nofH1xWYNArA/5N16u9+BZIdyvcG4ziHSyvbZ7dZTRf1syMLKjpSQvma053lGFLwmhHCCHw0S86eU\nwxGjYYw5xYrkvyklLEsOn6ZpwvE4wcsO3Q8dBvmeMRbFD5/DghgiFvGQ9vs99ofstRznGRqhEqET\ntzlyDrvV+0oJroS9DW6VUgLHiCBX6OcFydcQp+BAxpgaUjoHYwjWdPI7B0C80gT1eCJTxteacK9c\nTwjVMwqc4CWsmqZJbqYmJ7gBmAHAS8jKKSGGqKEyt1ADUxNGogmtAJ8CUrnGFPWzsQbWyNphh7xi\ni0cC9XgRa+jvvVcP63A4FJRRz1uORwwNxzgxgveY5BmGmDBLxMDkdE1tNiNAghNaA2Ocrr2u62Ao\nv4LzcUKQtQFOSLFGMJYdFAysTmrOJ8itGaZ8rwy93xx6V8+srGBDGfIr9x9jnf8Qgy7042Sw2+Vn\neX+7x/3tHpvrtZ5fE2Oo80qJwYHhl3yQafKY5Nm4vodDicQAY/M5rXt3D+v9QkKuYQqBdTHpL7/h\neyEGTIdspO7vtjjs82drOzwSt/vhwwcAM56/yH/z/OY13DF/b5pHBMHDEgOWqgdpGnTWGgKihA+p\nZnZYXe6oOEWMCTFVg7W5khAssrw8NTzTMDAlDUH8suhLeThMOBwmDTMTA+NYzk1Icv4leMQQMZWQ\ncJ6x3eV7PEwTUsmYMsMNBRdKGWdqQp9OcbT250nwhAo813uNqFhjfdyEHOZ2nQDA1qI43CklxcDy\nKYweg5mbcMfqvuW6Tt8lxfMSEAtwyIx2jejzSYyUakgWm02IuT7fmBIO01HuA2BD9f5TfYtt5zTc\nBOWMLRf0kqDZZEpJEwnLsugGtCwLGKwwBsekobaBUWzUMBCWiH3Iz/B4XLA/zHLdBn2fr2G7crop\n2M7BWaubxtD1ug53+wMOAkIjRjAlfV6MVOeC64ZNnI16uR4AWEJALPgxZ2wRAGJYAAn7nDXorIGj\nrh4yldA7Yjrk7233jP02z/n93R53b+5xeZ2TLBebC02MheUIP4uRTPld2wsu9fLmDXZy3u/98PtY\nXQ5yTwwyOXnz6PEa/yvebZxDwvM4j/P4YMb7g+6NeS+ZwPxTyewQZ+tJNdNzFADucJiUe7VeX+Dq\nKlvrR4+vwMzYH3P6/fmbN+qJTLPXLCMjglCpAS2PxnDjNTX8ouJh/eoXnyGhZP8cLi6zdb+4uNDt\nqXcO3HcQjx3ERj2pu7t77Pf587IsmMR9Px4neB+UI7SkBTB514XtcXmZgWhjDXb397h7k+9xXiJM\nlz27q35UT8d1Lh8DyNm9o9fdNXGEU2/BgEvKnxMoxwL5nvk0/OIS/saoPk4IAZS4hjvUAtynnmXk\npOBuogqYd0MPV8Du4HX+Jy95bFM9qSTeIgDEGEAFiCYLkFWPcJo9vKyXFJprohyCAYBPHolZwxiy\nlQoyECnIneet8UjyD+W+AmIBmlPUhM3mYp2zuSWk9iFDD/IMTXllEiEtjEVA9+3uoOuDyCIO4n3M\nBr1QIcZxgOkGdROm6ajJkf3hAC+ZNevyWkdLr0nNTZQMZGJNyGQOHcOQqWG0Mfq8OckaKf9GAki8\nepMKqwEpQeGN42HGTqIj7yNYvCdAIpYS+SCCU+EXASlEpcUwN1lR6zBKQmk1dkiXJbHz53Dy3hrv\njWHFhgJQiG2cWFPEiQMSpWbSLIrHHkJCJxe/WfUYR+ErDQaJGat1vpn1aq0u5XRcEGWSLAFIURcQ\nGIqV5ayRXI8xgoVAQ43Xr9+ACrbRO/3cdR26vLZAfb4XTdeyweILVjUrh2nxAYuQS+bZYwlRDaNl\nwC01Q9IvJROTH/4kvwM5dL08vPWA9Uo4UCuLu23OpC5hxvHuoBsDADUcDFKDZUAn2B6BQFTDtrJO\n/RLAkmEdrENYBc0o5b/TWFspIiYloDFmKTFcoWqMg4aBcSFMEsZHCd1iZN0kOMcv+ed55cuzYiBm\nDAgAliVhnsSQREbFs6BGLsWAxYeaJTUl+wY426OXB5pJnQ3FoH0vmPUlI4Km/zebDWJK2N3ljSX6\npCl76yo1gDkhhqCZ7uSTGr2+c1jJNQy9w2qVN62ut5myIZM2zx7TTtbUcYGVzcggwpxsIFD8LoFg\nCl4HU7ODJWQkqsAWdPoki16x2cyhrNnPyAX3ioiyYe6nI7aHHKZGBMAWZhjgY9RnFjlVDDIkpJga\nKo6FK5lVsuhlXV2Oo2as2zX428Y/BQ8LDeJHNbhMlEE/KqQ+ox4LUgV5u65H7/JDzbwVRidxvyWr\nN4yQ1EBx4hZOkRcpf2Q0Pz/5mD9N0wIn57Ndpz+PDN0NgongmBTAZaaKrYSEJNykFIBCx4mREUJU\nrIrJIMbiVbB6XiklJM6/BwBjrXK0hnGFcVWMV1d3rpTw+ovXldmNBvthVoPFZMDUMK1NUuA1RCjI\nG7xHEEMxOwO/ePTiCQDUGPEeQ+GSFdC+XIM16ITgO65WkhgApuj12rwkNw7TDDlkTqsrVcXoNSEE\nwQ5lQlNSbJJsm/SoHDSQcORs5R51fSEad0rqzVhqqovB0EkConiXLfza9z384nH7Om8aiMCgRN6W\ncZ5f+EKe7juLrssew/pijeuNePDrAYWUOi0zjtOE6ZgNwuEw62eAYLvCZYxyneX5UsMQhz4LQxWL\nj9RsNl9b/ZAESzHqmctVNtm2GCAlVi9qmibMPnuNpjfoV5WuwUY2IQh1qazZ4AGwGtDoE5IRTG32\niBKZxN7BybxSg1X+tnHGsM7jPM7jgxnv7WEVa5gNfZv+L1kkwJgEKrVFqXpYsXHrh2GFrlcCAAh0\nUpJSsh4pJs3YpJRyxkfOlSJqypDrvkKoO1It7SDFf1w3oB/Xch0jbCkrorLVlGPWtLyhuiNbwxqC\nWLKwFHSrI9gajpGpZT5slISZ/87VsMvYigt1Ha6ur+u1UxvjG836kCGND1nwLHVy2aiXmGLUWsIU\nkobXISSkyPo9NAxs1/fohG0fOOV7KJ4OmcrodhZRvM4QPLyEzwXLOh4PcILN9b1FJ1kpmIotcmQg\nJiUDG0vqaTvjlCTJxiilo/MOPgYUsjKIQCVkSgleMn45DCIYlrIWa4DGK2hLsMogIoQQcXt7DwAY\nXI9BvND83HUaYCiXNAHAYDus5HtX15fYbPL6co4QJGw8eo9lOeAwVw9rnvPz6Ide4Q0gR43JVBz2\nhOrMdaUn/VQ9LHUDCTpH2Ukuazm/HeXRt2Vifgk4HrNXNfkZseC7qxFjCHByj8YZJXdnTzffY4ge\nHJN6aYfdEZMRj/t+i/u1ZCbnSTPeJfJ5l/FeBis7qaU8hmvYxrV2zRAkVCyTZlBfJaMP3FgHU2oM\nYwIZUk6KI4uSMOfEOrOcynFloRnW0IxACiLS2/EhgBAY3VhetA5WJsm5XkF3Ikkfl2dvjOIXzjkN\n4cg4NYgxhAwYywtnTWvYahHx7BcswSugic7qAuq6DuNYgNkezAXDyZSEYuQYRnGXfLoS3uTYoFAR\nYuRq2ACdy8wCr+Fl4kodoMYQWec0RA0xANY1mCSpQWoZ//M0YV4K/8rCUOaDFQPRGSC5QgkwIE2Q\nJLQ1VtaQ4hyd7WFtXuApx8P1gRoCGnys3G+KUflgoPwMyrNhON2c8rzVKomW00YgJC84kTNaasTW\nn/wtGVaD2lmD9Ua4V1crjEOpNfUZ2ARgHIM6YH0xyi10MIdFvndaytxsm1JryXUeGntUt5sCSUAZ\n/LFJRAEM1vfDgpoSnMUHzGLkD/sDDoJFLnGBkz1mWHWY5l43BjYm0/0h6E+pdOGIxXvshW959/oO\noRPqxvUWW6kL9veEAs1aWxyX3z7OIeF5nMd5fDDjPUNCOnFHdXdmo94WEcGYSiughu2cs1dlezgl\nKuaQTrIWZGr9XGLNqhBnUp4mQZgV7AaxsH3lTLoLCficUi12Jas7NxmnhbdEVFIy+rcKVHYdZKOA\nsZWY6meP4APKpp7LLSvLvISz03HGPC0aAjBbDceI6m7fdZ2C9gX41p2MSeclptSEJxncL4B6SlE9\niWEccXGZ6SPzsmAWgl/X9zCmkm+NMSdyPxeiyJBSQpt1To1qxXE64ihh4G6/0zlxLnuCKQQgFi+8\n0zDfcmXSa10iVQ9LQz9XZYYiM3yo83qClGffM39KrEoJDJEROknny/M0pB4lx/r7lBKssxgls8cw\nmg3rjGngg+zhUQH+ndFrCt5j4kJKnZUyYK3Bg+sr9c7ndUR3L8TM7R5enh8xw7EpPE+w5dPb1bpd\no1ljTbg0wgOcUjO/UechyHxp/eQ0a7HydJyUUGuJNJnRhwDjjIaRKaaGsJs0+wxKSCFoBcB+u0eS\nyMYfJsz7bHJ88jjuBYAP726G3o/WwIxYShYSqSQKu6pzQ2LUCq8mpaTZJzKm0aGC0h2ss0icTjJg\nNbFTX9JcU90w7A3UR6REQAml0qnGVT53B2skk9Sv0EnphOt6RJGXMcy5ZEBO56wF9fkEqzHByX3E\nWA1MjBEhBXBN9qiulG1KI1KSDEwoqfRFy3SWacayFI5Xlf0oDGste6Cq9lAoRjqZnCpeYViLXlcX\nK2zmKz3OUdz9UsmvlQAx6qLLIVI17kRGF34KbVnSQRUUlmXRjKHtmuxb8yCq3WOY8hZLBlJ/Z6vB\nckSaiUKM9RqEHV9KV/Jaqr8jqiFhfqnLM4glOsPp0uccmiJTUZyzuLjKcinTbtLNKFJdA0aMBSt/\nbFbO3na7gys4H5EaLGMNLi43cILddj0jyLM++CPmnbwQKS9rLpBLOp27OqFvaX8hG/aSmY4clVbi\niHQdJeRnXIxP8FGLlUOISn9YjSMSl3laYF4fwAVPjqHSbbjSehRwUOfBKFaVM7gCb4SghnF3P+Fd\nxzkkPI/zOI8PZrwf6E5V6iSh1lidbvYm/7ZIwLCp2TBnEefKCSraQj0NMFTDu8RJdway1MjYMJgb\nV5RNzSY2aoaM6mFVT8GhkIKs62BsKfrtYFwhpqZalAWgc10N2xjwwu/hBCW9pRAQg2+KfDNjHgB6\n18GpgF+UbEwRGaRav+YX9VS6gdArMHwa+mS+jHhvhuBsCedyiKXSSNbUY/BFDUkM4Layw6WQ+TNF\nQG6ZK38sVQmeUjxdQemgWbhldpinyv5Hkykm+VsNOVvAn6hqM3FOqlRV2SqDkhBV8dIvEZOETDHl\nEKmsC2OMenKcGhKvzSqqyuRqpjPzs8S7bIqnQ1hgrMVKeFTT0StPj3yCl0JmGAfbWwSuNZBJya8L\nqNS1hlT5Yr2F6QyuVSjR4UKA+pA26iXOk4cxTRF+I5lDxJqFzkmUEv7m/z/h/nHSuQCRPlMyBmyq\nllaMDO+Lh2rg5N1YrzpYJwkWnwm6Wr8Zm1CbWeeyZMKN8C37fsRwkT3Ki80FLsVzpeQh/FTM+x3e\ndbw3raGm2BuciEpeA5q6aL83iPrAuFrhUIiLYVa1S5geDCAIHuKjV6a6MdDYKFEmX6ovi0bmlmum\nhIhO08DIWESJuEMIiL5k4nq4UjpgIlwjD2zIwJQMzzhiPQgeFrze+zIPJwJ+AGEcKsuZqVbmO2sb\nZjJQ0iSZeiBp4cWrSNo36XapwbJWU+qGBK9ojFnJ5hgYxQBDCJgk9PRLRERS3GtZFj1f8FHF/Bbv\ngZNCa0IF7KDnsdZqFGCNEbZzbIxUxZmsIXjN6mVljRo5shaZx5iLxgFgmhfsBWfJ2WaDriv4otPz\nONdQUb6h5IPq0jkpgFeysKh1XKwzLeE1XquBZhNQghLTZay1EG9d5+ClisHPs5aTTYdZsUYXDFbr\nlT73btVjZaTwnlirO8ISMtGzJMNPJMPrfRmjZAYtu8mF/yUr2khgMzViF5nlX0X79JKAZFT6eBwc\nRgkB552Hs4TFF4JoQipZ6dSKCXCGBuRc/eCwvshzeXl9jSsRRrQckHy+j+P920/pN49zSHge53Ee\nH8x4f+Jo61VRdVmL9S4ZFavm3Gi2axgGHLaSXfNegXkQgVPOqAAQTXgJSSxXBUVRTWx3Y+hZqnRs\nasKPdpRQ0vughbQpRrArgKE5/btUpU6ctbrzLEuVUOmchSXAFu/OENay6459jyXle+q7Dp2zGiJa\nqiGdJapcoRQVvM5hYgVt2woGAldvQUPHsmOak/IW5yqHqgClgQM4ERbJFC3ea0bTe4+DaJMtPut4\nqVdL0HkY+kGJwOsGdHfGYok5Q1U4YK1ia/C+hvVUyJgldCYlt/ppwUFCzlxgjuY+uK6fAC3Hydy5\nmvk8kWY2pNeYTrzXuo5CCCAyUr6Vs8uLn2UugaKaatlgNKT8OSNeJQD4xeO4k+L4Q1Dvg2L2HoNc\nw2iBTnhJY+rgXONFWWitpDmRtarlQdSQnNVxZK6JKEuaxYRl9fxDCPBz0FDXmi6XdiFri2mB8jDA\nyzvZGUJvCHOR5DkesYjH6+cRxyhk3Ujw89JET1C5pPX1GmspeKYY0K/y37jVX5hEMhp3E+pfMyd9\n4YhSFvBTLXWrTOHV2OFOXthpOmoqdbN4xBSwF62s4BuCHjVdc/IZTjISZXHm7FCDkTSfy++V8d2w\n5/0SKs6EnJWqSg/Q+0KskrqGKoZlDcOajJcAOb3dS8OA1WqEldT0dpsJnK5p3FEUGrrOoiuChbYW\nK7cLEhCsptGAUpazSVKrWLC9NlQmve8QPLwsrBADEgizvIzzMoNFhjd6j0kKvVMqmKHcOyp7fxhG\njGX+OeI4V70q+JJJLZhlUIyMTDX4KSVpmtHiZ/Lih6Cp9yWEOi8mH1sxy0adg+D03q0tDVDKH56k\n207rMsvG5PL3S6WG6zrsS5Fv8BgFBx3RYxgHDCIu6UxLU7GYDoKL2g5UUpMUc0azwZVKeNcPHXqB\nToxldJ0rbBa4zmqdIQEKK7ToccaWoxQ217VfjJsxqMXylOkZpf5wDh47CUf3u7nUpefpamS6O1tD\nb78E7Lcir31hsJH135tOspPicJikjStiDJr1NdaoSGf4um/xG8e3MFhlJ2xAbgKS7DxGgPHg64tZ\n4vyHjx/g7vY1AOD2zS1uXrzRyUgp4PZ1DmaTz1hS/p1pPKosHVKURRlJX2BCrRAnbko9yosQUR2Q\nVPlRIXiAezmelA+VE6ao0iwEagyWaaofEsCxoSJY9LK4xqFTVUVnDJwFOtlFXWcxSgnK0DsVfOuc\nVfyjFMAWpnsp+s23EPXlo7c9Sm7aqYXKRg+LV3WB6CPYGAShdIQl6PylRv8+pQJO102i8JcGZ+DE\nw4gctICk8Inal6e0FSsTHdXwQkqpZMpT9ZwCUn3JTC3JKsXEZZ5SoXUgr8UC+Fpkz1WBZ2Z9YVoj\nly9JzmONYGqNdSyWI1ZssOs7DGOHYTXonM2ivHB3f4+9cIyMNYpjFpGAwuWyruKLTKjeEEWArN57\nlhOon/VZUKUPFPCdqJZ/nQhsUmug8wbnC862BBxK04hpVuZ5W/oWI8tzyv9eZo+tsNnHFQHiRfU2\n5s2lPHcDpW7MMao0vTW2zut7qDWcMazzOI/z+GDGt8ew8Ja0i/raBKLU1P+xksUuLzd4/PgxgLyj\n399nrZ1lWZA44ah1VYROdhtnTBZ5Q/busmdXztjszaaGPrlIuqHHo6SDy1202ZN0svNT4hr2xqRY\nB6MyHqyx+vMUQ+44ot2Eku4CnUGuuUIOIR0ZdFQKPh36Ijo3dFXAzxmkInvbhLPlApV4x1WbLIQI\nIqDvZGfkhOALXuex34lo4PGoBeAhEo6Ho0pWex+wWuVUvhtcZS7n9OuJDlTBnxa/ZJ0kAMHPjZ9U\nsJdKQM0eW/U00omH1XhizEoozpkt0aLqkuJUMeS2YMWLZK4ZyGStit0VueWTPgSNh3UydI1lAm3p\nYcmmYqi97TXczyHcoBjWfJxUO+1w3GMWz7UzveK7CXkeiyeVewNKJHFC7cg0gdMrrPjWyc/Kc9HQ\nz6LU2jKqN2yt1exuTAlL8DiKtM393Q4vb3L7vekQ4ZeSbXbo5e+nacayLFoLa5akXtn+4OAky72Y\ngMUTihq3G6zi0T559eqsrddM35DN/U3jvQ1WW91fBuv/1Q/lxUocddGuVgOefpRbQDEnfP7ZFwCA\n+/tdbkCq8TjD9SW1OtZqbiYB9SX8sg3VHbUMoeWfFE5K58xJv7ZOgOjeuYZOkvICV2A2KE+JW7Cf\nE6JgQTF6pBRqm6Rklf2cYgTJ31tkTlJ5OJ2tGFbf2cqpanCNqoBRQ5o26VH1uRKcM6qMkDjBC31h\nno7Y7zLjKRGmAAAgAElEQVTWsD8cVTnAzwH7u6PihimyYo2us+3NAsS1aNgwjtIabLutba/msNS2\nW8qDaoD2JWhbKZimz57JAX4phg6RFetirlwkOlG+iIgtVywFpFJC1c6fNRlbkuuJqer6M9ekBTmj\nnK4USrhYIIXKvB/7Hv1QSqiyPruGTI0hJ7KVQmAaoUUJ3QvWGFNTtN2aJwZa9ZGsTV9x1Rr+1/VQ\n1UmqIi9XxESKDuRZpIglRBxFNWK7O+C1qFMct0ENFrNwEwG8ev0ad9t7HOe8roI1GPZ5XlZrh14Y\n8dEmRHbopEKkH53u9CH62g4vMXypPbKnpvnPG+eQ8DzO4zw+mPHtFUdrggLAaVo1MsBUpIErINp3\nAy5Fxz00us9vXr/B7n5fuxJzUrbtanVRu6Ag7xi6q3AFFplTo/rZ7DjFo+mcpnVXQ4/VWHW22yaS\niSuwGNsGk2hoBQ2REmAYSjUcYlQN9bDojk75FyChFRg0ALyzDQ2ktlav4GrZXetORKjeB0yu9yue\nbN5ZhfC3eMySmp7mCX4W130KCKHKQXfdoDVubffk3FiTGskcq/STeZo1VDwus+o3GSlsjzEqoMoc\nawiAGu5m4Dh7VkBeFyHUcLFIEDlrlVLAzKBQay5jjBqux6bm0BiCswahJHpTrMkW7eCcU/FWqxiy\n9zeL50kxotfC9OqZc2LEVLXQLRNKJbKBqd/j5t1A9lSdrZ2KFiHobrcHzOLxWOpyobu2U2MYKk2H\nq2Y7U1L9+pwM8FKVUJ5dTWAsIdU2eFwSTPmLPiatKNgdZoBKJ6AJx0OGE3b7PQ77BUsqEQOBRqlU\nWRusu0wI7ayF7S2Gtbxr84BF9LAYtTsSxYSwFNEBvPP41kz3Nislaw4AkFD0kho+lhgsH70yai+v\nNlgJ3+Py4hIvvnyJV69yHD1PO11Afd+j66qML3Nzh1SF+nL/PP1a+b/KgG4yccMwaDv4YRwwHYvB\nKiFkeTEqhgXmkpwSqd/i/gv3p3DFDFBmI3GsIWUS5rcy+FlLmyyhwX/eTlvXQdSEQcRoiVlERtUo\nQkoaVsUYs9gdgMUn1b1PJredKv3zurFTydqcKZXbBoFsp7wkZlalC5BtIseqtkHI3Y299zpnppUn\npqa5AhMSs2YWp9lr5tLAqt6+MRaUihEIJcYpF1mL5Y1pOFkWbT/E2HR49sFXw+YsrIahOZsZRJfe\ncEQnWV9nOHPmkMPwsHiQ9CGwxqqeIPuoz8baqnBANoftZa2kxJikNdjd6zv9TDBAIsU/o48ITZhL\nGssCti8csyILXSs1mLnB89KJ1Du4GraYIuYpz//d3Rb7vZRA+Rn7/a18hwFymqklYtAh3/BmWiEi\ni066cYQzHVZy7IuwRlz2cr0VV+GYqh4WvXugdw4Jz+M8zuODGd9CIrl8qD9jVHCbmUEJGq4459R3\n8GFByWD03YiVyBSHKeL1y1sNC6yxmkGztgmXTNkVKsBYhjGkHgY3CLmSK6mGh4YqQO26DvFYtkYB\n6wtozqlVyaj1edaguA7O2ZNrtKYpSE1JP+daU4ZKwGSGJAAgLgtQ1Dib0M5Q5om3XKbiIcDWgt8y\n7+Vmo68u/rJ4LBIGeh9qI1Hr0I0Wrukwo9phvhZzG2OELV9Z0tDi2ypPw8aql2iMBVLIGaHCCmwS\nDpbRcKqy56Nt0/YTFkkMWOOw4rF8rWbbYkQMQWvZUjMPhkhJnwl5J9cayRA1y+WDR5L15kCArVUQ\nnGqWuzMG6AuJ1Cj7vHT/Vs5co52WQ7aSpU4qpdP1WUpcu83sjtje55Dr/n6rCRHLBieFC7aGMCkm\n9cBt57QOtggMvJ39LOFj56zWxSbm/C6VZ5dqMm2eF+z3OTTtB4dRdNGWxYNsB5L6TTMAKylqXl9c\n4vIqe1gXm0sYsojyLq7nIw6psOUtrKQPc6VGCyG82/jWGBY3ha3ZvZRFwoIzaRtr05QKmEpIRFQN\n8HmZMU9zdqUBrMZe2151XdNaKcVcRNu2j6+0cGXXJ6aaHpEH4cOiJRZLDIp3kTW1FVXKBkvxG1N1\nv7OMsWhokdH4O4aI6ViLnztrNGwwRHAScvXjAOoMSF6MRDO2u+xuhzgjxYwBXF1d1Uwsf90BLgbL\nwTWGLAOKqjgQGUspoZgbxYNQmd6dNeid1blMqWoixRjUoHddBwOrWmKeMuYBAEtgza6Bjabrc5lH\nwBwDSoUGmdoBh+JyQhT1YVH8Jiy1v2FiBqby/Dx6eVlK1+aqfIHaUadhwPuYmfJFUWJaZu3ok2kv\nQvFYAqyk3uc5ZLniItpnnWpg2Q6a6UwpYlkWzEINsMahkz6Tm4sNJi0zW/S+V+OIvusRC+VkCTiU\ndnaHGYZLIXnWry/F7Y5Q1R/A4FJVYWpRdBlZgq1mFAvZOaSE3hWKjUU/dLjcZOWEp8+AT3f5Prr+\nQjufd32HKKH6drtFZIIpBedrg8tHWeTwo4+f4uGTTFe6GMesmSZYYd/3On+d7Ro8OjUOwLtnCb91\nq3pueCLM1VtgIiQ2SksKIZ68d8UA+OBx3GYe1s3NK+z3Oy13ubq8wJXUHI3DoAYkcQTZWrfIjYwN\nQMrLSWrfK1A9DINiG2SoNotcFvSCZxnuYNDwsFLInDKIwRqKFnfFY7q+z00bSgsx61S6xtquckyM\nyQCy7CZLiDBSysIg3XkYFpDrnBSUrF5Kq0NeSzBy0wnV6F4W9VJ8jFjEE/GpbjKO5JrkXMmzStx4\n7zEkaTZABqZphknwCtTnsprqbRVaSlH1XGIEq5hhBrSBvIbKhhFClqspqhGMpnaUSNnzwUcQl16J\nnGtAFVyvxXZM1eOKIWCaZuxF/fI4H3W9GGtgTeNxCEXELwFgVhDfWFsxMdckB1JECIxJsJ+hr01U\nLjYbuKWTZ7hDJ/ezXq8x9iOCnGs6TqqKwZxLX4Bcq2mMbYQg0Wj5c63rbDhrKsmTgs5/W4OYOGky\ng5CN0eZK7ot69H02Pp98t5ZopQTshBLz8uUr3O12SPI+rK4HPHn2AADw5OljXDzIHtbKWcSlasRb\na9EV0UOQvhvMCUnuz1f0+beOM4Z1HudxHh/MeD8Pi3AShpSdzOA0NWms1d0vLEmtMplaVAqQyvXe\n3d3jOB3RC7Hz6voSl5fZXe17p14Os3SUVlpDTWETNzK+jaBYwXkuLy80pGtDSr/MSnFwlCkYNSSL\n6s31zqlKQQxBd6slJISY4IuGE0dMEt4cZ4+e5NhuhX64QNeJ/tS8YK+FvUft/nOMCUVocApB5rxm\nw1SityFlFoG9khn0Pqr++eKjMrC99zr/FLocFlGpOQxVUHBZ9HsdiyR1wY8ITQPXNhypWUqldTRF\nxZwMomZfqyRxilGIvnqLNeVsanG7IdPyA050n2Ks3mXk+vOQEkL0WhGQiaMVQ4wNTYLF6ylKESvx\npgnVSycD2CIZLJ2KCj7IkTTb5ZzTc440KB5rqMNxv2CSeb67vVcNLGeshrWlllHDZq61l9YY1cOf\n56CZxBK5xDiDuWRJSdesMVV1onR2KtnPzcap/hcHh2WW408Bt9IBO8WE43LELCF131ts1tkr26xX\n2lGq7zuwS4pJOmu1o7WfF81GgggFGJi+pl73m8d7N6GokjIGrBwlVotlkDsQF/UCwFXWL1U1xHma\nazv6+YiYPEwvzNmLEd0oLbWsqQtG348K8FeR/Ybin0hTzOVBP3r4QDk9fddVUD5EGC4hnAHDAKmE\nJ1WmhZkUC5qnBftdxkUO+xnHw6Lt1YEIa/N9WXfQ9K61Pcb+Ase+GIVUDUlY4KVZg/VRw8ayOKp4\nXl3E1ll9gZdlORFjywu/YoplUTORgu4h5NCnZeWjoawUALnv+xx+Fv6Nc4p+G2sw9PnFnn0FoMtI\nPmlRPFLKBdsAwFU4MKZ4wp0CV4ySDYG5lJlUbLF9/vl668/QFDVno9ZUBzTX1q6dPE+lbOiU45Yz\n8ZXBr4kbZANTriqEqEYqpag8OEtONyM/L/DBN+t+1moFat9El1vjqcEi1mYTBFL+FDgpBaRIJFmq\nBf9t+JRSUp4YjDzjoqjSEZBKi+5qeA0Ix32hsBBiCko/8cusgpvMrKqqrjOwncNKMOi+7/Vd22/3\neGVe5r8hi4Ow5m3/7qD7OSQ8j/M4jw9mvH+WsGwFTS1W7mxTu4kASZnETBa9KSJnFjvR0Hn55Wv8\n+lefAwAOuy02mzUeP34EAFhdbyAcQfjUaGNZautX0RmnHVo4JSDW0LFY9fLdhw+vUewzEde6NCKV\nS7ZwOXOhjkBClF3Jc1Qxs3maVS9qtz/iOEfMvoYnTEXw7YhApaZqAJkexmXXO/KsWlSJEyhMcjk1\nBA5FOkSF6xudb1NDpALAq/cLq/e3Wo3askopEc3f1ONR7XBMVEXcVqvsTcvfsSEFoftxRF9Cp67q\nuwNAWCYgVUoBgdTDYg6a9Mha6LHSLQBltBsAQT67ppaQpcNS7fJjwU2mt3g3RDlU0i48jYxzO1Lj\nIZDLWudWsqRE1cON0VetLSOt4qiSUtuOMuXhWK6s+uO05GoDX5IpFquhvhsl2UAme9JF8hpcw++E\n+t5leel8LCseettyz+SVgDKZJdKxIDhuaDGNUGWmGuRzzSmqfpqfZxx2W+wm0TOmCVfX+dqPx0uk\nmNeYsSMoVYmbrusUxlh8wG4rISYIXqWiayXLbxvfQg+rfCK9Mbyt1MmpZp/Y4HDIC3m3PeJG2Owv\nX7zEfptf+nFc4dGTx/joO7kwelyNJeMMNpWvQQRpai+Lq4WqEsMUTI2hecKipHhxcaGhQpKGEPku\nElgedrIJHVo+l5Huy5mfshVju9/vMEml+mF/xOEwY5pLyUdEiEe99yhvyAUTAAMnrXTJOrBMf0LS\nlu8+TbmfHzIVoEOTJWz6PbajGKuC97Rdqi8uLuDFIBui0/Cy6dnnmmxODB4X0oRhc7UBGQMrekfj\nsqjhuFhfaEjovcckWE1KCcdlCydNIOSHei6OSYudDecsJ1O9rha8UfzOQLNKBhI1lyjSVOOVEIXv\nB3lDE8oBjaG2x0hVOOByVOG+EdR4tDtk7idQMUTjHKrwYFUSZWa978xtK99hpMhwVhRpxzUGLYfK\nBhGAZEZjXa+NEUlNpp4S1dATEYCR3pVyL0ywsv6tdbr5GoEkCxyQFUoknI1AlCzhPC847KQoetrB\nL0fMc8ad3QTFoJf5qKEip6yqcXGRDdjjR49y6QRymRjLprrEoPPVr86dn8/jPM7jn8PxrYmjLWB5\nOhgpkQLczMB+nz2T58+f48WXGXRb/IKVsGifPX2MZx89wkZ29SwtU3SokhYMAxm0N0VrC6i6VCko\nuGqoNoksft/QWb3mECqTOYWkPDFnuiwDqy511F0pxhkH4fNst3uVd17mRZqQlnkh5T2Z2cNJtmQY\nRwzDgHGVQ8L1esZSricFwBddKVY33loLeI/iSkSOSLL70WKUNJgEXC7aUV3XYS1hoGl0y4a+1/kK\nRX5YfmmtqVy6FLGWHXK9WYGMxTBm78mHoB7WOKyVe+S9R++z5xB9xKs3X6HruhrGUbNe2MAVMF34\nfEm5b9yEXUblq01TTWDEuXdtZrqA5IBmZ4pyaltb11wETtKOZaWk03CMqUpjMxG4eP7ESLFWBGQm\nv9FjaMBhoR7PsOpy0kLWVN/36gkTgJjy8wvRIwSvTYuZklY/OGs1oUWohFKdB1BV3oWp3hYyPxLI\nYd+JPlwbHBmoVTDkAWTPaTU6PHx4ifFCnv3lgM0mr5G+syoaSpxy7wJ5t5GgazGkmnSb/KI1roWk\n/C7jn4Lp3moznUr0MmroaK3TC+qHEddCMAMSHjzMn58+eYiLzaqK8BvJ1gFIcVHmbSauxoZmRnWy\nuX4mMtXINal/04SSVSgowlFdDF3X6eL33oN9yTjVwuMUuZZHCBm0VBeYUI2Xc33NxBjKrF8SFYpl\nwiJueIhBm0HQRLqIjQW2L45vZbNK5rAKtZU2VcZI0bAxGmo422nn7bEfNITIBstrWGCIoOks4hyW\nAxhWI2AsujDItcbaSr4flewYGv31oojQNUz6nPGrKgxaSJ4qSTQ/xibbSUY3vpbRncmoRiGAfKBS\nGG30/FFaWZ1sro3UR9vyS/ttcgJQi+oTGvwoYxLyJwm56FrCLG77YjYtxMigVPlmTS/UtWJNzciR\nQdmkY0zgVJtXAOnUMBa8mJv7SsgLm6nGuqXEBwDHJtQuxymZfVvDSCSj2mzj2OH6Khse0xlcP1oh\niWBjNzqsr7Ihur7eYF3WrJTcFEWKi/UKg4gOMFndmJbQlDW9h4DfOSQ8j/M4jw9mfAsP6xvc6Pa3\nwoMpu9rQEy7FdQQ/wuPH2auyzuBSGiwOvUVIHkHA0t6utCAyA+LVWzKNjTWmkdKwVToZ/A3haooV\nCO0snAjtM/foTKfHc67TMgXnZhgqiYEBDx7kUgRjHSYpfQFnzlLJ0IWY1O0dhxHjqrb8MrYW5hqb\nwy0ACJy0YejhsNMsXNd1uH/+vALlbR0l1W7FmUB7Ki1cBhG0I48ZB50v7xcEX7siGwcNk4lI1VBN\n58AwtUSpd5kDhlw/p41dG28IXQ1pa9ei6iUUDSxAuE2c6Znl3yfCWXofVLWUpLNR2526yEozq7MF\n29kTz5+oJmzagxOR1vHVBEf5FuscUb1M8XQqNBJiAjRLWIF+avhPKTESQz1Ua2xdD2RUeghUQk6B\nAhB1TXA+aP4a16VOKQE2t4Rr77GUQ7XuFUnI2nptGolYaFH4OA54+iyf90G6RqKPCqcZpieY0sjX\naK9jGIJ68UD2crXcF0kv2BLBlE5AJWP/DoO+Kev0m4ZzK+5H6dxqqhQtoYaAIUa9Yf27svgNgIbt\n3KrwZlJks2g1vjPKKo8xgUAqdexcdV9bGdlh6PWFI2IcdjsQOc3QDcOAQdjtXdfp8bKb/E1m+HT8\nuXPW4gFvf629Z5xEJ9/w5TwPv/j5n+HZ06dfP00rQpZjk+YQ7Wby1gvbGJCCfQFCTWkqAIrrHmLu\nrlJezM5aXG6yCGOuhcsGves7TWUbQ7i9v8dut68GzZmGDd1nGWbkTYJMm2UmnZis2F9C91q4TfK/\nb34OJ1ZOZwMo8/MNWdZ2jgC8ev4ST6T3ANCw9U/+lNGS70Gt8WIt/o2hvqT5GG2asraLG3un65CE\nRlJtbZuRbwxP88wJBtv9FvbB6vSr9Qt6DGZuD452zto54uamy4qqCg+1WuFr4ze8Hm//fX0ujN0X\nr5j5G6r93xrv5WF1wxX+4F/4dwAA15c9LgSIJUDLQm5u7rHfLwoSAhHPvpMf/nptEFNe4DF4CFEW\nHPmkoUHkiChxOtkVXt9m4Pr29R6d6fH0SfbSnjwc8eBaFn7vYV1eJD/+/R/g448e5mt2Cf/Df/Nf\n46//i38bH330EQDgR7/7u/jR7/0YAPDJdz/GA2mf7XoL54xyct7GZTV1zqnubJQfgr74tqaxY4jN\nPOQFXb2CppwiBKRQsCSjfKXEEX/rb/4N/Of/2R/pMYr3YAw1fLEAclUrPEXWAlgiq3/kXI9ejPY0\nH7Hf36vXMnQ9LgS3AgE7KS96c7/D87stdtJY9aOH1/hb/8rfBAD85P/+E/zsp/9PnsfvfRfrTSmn\n6vGf/oP/Ard3B4xSsXD9YIPvf/oJAOD7v/N9PH4mnLvNiHG1hnNybmM1eWCMQ0w5dR7jpKC7sx0c\nG1WszWVg6is2hCOLBANfNtOQVAkXVNvCObJNgxHGf/x3/z7+k//o7+e5RKzGx7cof4LtkzLMQaRF\n1i/uJnz+KvON3rzegoTLZJCkgbA8Nw88vcpRxh/8zlM8k3XtuhERFSc1phfMMnsppXUcGWglgEOP\n//K//a/wr/69fzu3DwNAocHvKCGVZqlS/lU2GCZbxU0azC82nLaY4gnrf5knTMojLF6yeH0NPshM\niidTQ6uZlwW+qKHGgP/t7/13/wjvMN7LYKWU1CX8zifP8Ol3nwDIi/3+Nj+sn+LXmA43CiJvrlf4\n4Y++CwB49uwCw1B2eI+wFGo/QLaGHZETvCymaXb49a+zDMs/+X9/jfs3BxxFmfHeJHTCA7psdijX\nKAcUg2PcgPVlNkyPP36GZ59k4/Xw2QOsVo27jaS8G6LWua7ETG5UHcuo4U47X93Jv5lRuxWj6TEY\nKnGy61ytiZSV3dvSYLPWSxpDSIUflABjGFTcf0sgkowdA0fxgtJxwcOHOaz93qffR2cZn332WZ7n\nwxGp9BO0BoO469/5+DF204R9fv9w+/oNXrx4AQD4g7/0h/j5n/0EALC936sSaZn7cRy0Wcd8PCix\nd3QWY9EjMxbENXQ5HgNmyaxmYqbsamnR0HazHmG7VT4moNI0QN501uIBWtchguGFB5iCV6JxDAlx\nKQRdaI9I12UNMickWqIEkvA/UNOs1yQEWnCUa93uD/jqJq/Tm7sFN3dSMzp5OEjDkjQjpgSWUjCK\nwK28N59/foDlbLwePupyyF8yn7Z2lWYkFIWWrutQ0pba2GI6VFJ0SsphtMTg0oHI9khUgXZqkleR\nDKKpq75wEiMTQvRCigV6bhqTALClUQwMiGs4S4m0rjhHVZKQYYYXI7y8R1/C9zJYDGC1zrP10UcP\n8Omn+aVfjwNePc8r+qsvbxHCC037Xz9a4+HjvPN+8r3HuFiXYDcqQTIG0UDSTZJBEE9gciAhu7/5\n4g3iIak3sj/OGA/iUq9JDU+T84AteNJqjYuNaE49eKgUinFw6rUcjrkwtUzqOHToXX2xSjibPaWa\n8QLQhD5Wa/JSTJotcTZ3ZilaWdkrL9a0utdL9E0T00x/LUKFBtX7M8ZUgwUAMWpok2KVptnuDri5\neaPP8MXzVwCAV69u8KPf/T6++73vAQB+/qc/w3IsTHVCvyqdiiLmwxGpaDgR41effwkA+PT7n+I7\n3/sBAOCrLz7H+iK/mNfX4rE6aE3iahywEoLg5eYS11f5O8kQfCQcpHLgq+f3eP5VbrY7HwLWwoLu\ne6PX0Hcd+m5AmIvnwlivS7PeK1gU2WcgmQQr89Qbp4jUYfK4eZnljfZ3B8VgCtsfpTbPEAYqa4DR\nFeJvB2wPAbv7bDQ/e/4GX76R6oejxyLG1AaPWLS2IsPDAVJz6ChgK2HGF1/26OWlHwZGPxCcbMbU\nGEpwrSFNbDTdWozCyAxXmPSoHlHnOt1NjzHAwsGU4u9mM7ZUE+ggW6XK10XauRTLL5iWItszY1aj\nlA1+mU+HRtixQSr8sEKQKpZjqNUsv22cs4TncR7n8cGM9/KwiAjX11KycdnBuCL4toBMqVML8EtQ\nNz3GWLMaHWv1PvOiUINxJOJeBSSNWroBE2FjPg/8DIoRJbeVYlShuhQHzfSYVMmlJbxbX1xhfZE9\nvdVqDaslMvmMAHA4TtjtD4BgacZs0BdSmwEmEVs7HvY4HCQcmTzAUAJlN/Sa0bHW4UKIomYcMgen\nEZorOw+huso+hJM28dQiq+AqQkioWSUmsHWKVR13O/xKej5++eKlSqA8fvIMDx9lbO/mzQ6v/88/\nwV/5w4zl9f2Ae/HENuu1Zgzv39zj9vWtuu/oO7woTTeDx+OPMizwq1/+meKYhUCZSaoFs6tZwr7v\nMQwrPcYyT7gTnPLzz7/ET/+/X+Vzv96piicS4+42e0Rx8TDGohfv9fpqjR/88DsAgL/0V3+MzZXU\nHw45g9sVNQCy8ML7ev5qj1/86jkA4Gc/+RWO0my2c1kdJArfzZCr9ZYdwZRuOCbBx4StyAnf3044\nSvOGw/GIw5SvtYNHV5yjZEQ2W37QebCEb3t/wM3dHQDgyVOLYbUqNUMwxsBIOQ8zgYXMuSyLigGu\nhXfXxQSndbS1RtCl3DMQyLWEkSvR1RpTYRCGcuMSx8rJIgtnjeJq5AhesoQHZ3EjQn9LDHBkKreR\nWWsJHVPt/2hMVSj5i5JItobw8XeeAQAeP76E6+XFWoK+cNFP8MuimIzrHLpBiGPgzOoGEHxAX9K7\nNru52uk4RkAoDjwDURp38jKDfQC5cjzSWr02ELRNB5fSKqkf1nBCVzCuVyXMBCg9YLff4/b+Tg3J\naj3WjJUhbW91d3eP169z2LK93wOc1SSBXORcOvKM46hAcd87bcmUrzbfQZmXgm1N89RI5tS0MeSb\nRR/cGKPSJNZZMBvciWzJ81c3+PKrrwAAL1++xCwG6+5ui993fwgAuLy6xPZwwJ/+7BcAgL/yez/C\ni/Rcv7eWkJmYcH2xwV7wHnQOd2KsX968wiMxgH3f6ZzPcyG3ktb7eZ8QQsXvquHOWcHaoZhQGsDA\nELzIp3ifVKdpOWQp6yJrbOBwOJQQvUff5Wsfug7Rej0ewxRhUUwxYStY6M2bLW5f38qxElZEWuGA\nyNrnk5iVerOEBbfbPW63Uie7LFiEApOsRy8nvepWGCW83C9HHMKsmloWCVR01tMR+zmv693xCg8f\nNsmbFJWwGSMjRGlmGmZNPEA29ZETCpoEjohFAibW6pOu62GZVIKZGEhFa50JJGtxsAYXfdHyAtjP\n4IJhDX3e2ABcrzcoagX7xcMRoZd3sWNCV3TljdWNMHLCIjZjiOeQ8DzO4zz+ORzvR2voOnW9Hz3a\nIKbs9iYfEcXCx7iAEWEKubBzChqDaxttolTLLpxFjL42EohQ9cvjbsa9tNGe9jOC91it8g7KfZIe\ngRAiXwWkte297OTeJwRf0/6u3DkBh0PeNV7dvMHNzQ2cuP2bzYXuIqbhgy1LDnsBIC1ZcC/JDhM9\nsJQwKC049HleejeIhEjxrWrNXIgJy+L12Kq7XuRGlLuSVE99GAY4yeRN3iPCYCeg+Zu7Ow0rnzx6\niN19fk4vb17hyy9zVrAbfgdd3+PVXfYsttOClai8fvGrX6EvDTeGHp88e4LPX+Qa0J1fNNX92Rdf\n4G/89b8GAHj85KnCACWATcFqtmg5BhwO0ntw8lVBwmSg+/JBvufvfPJE9d53H+81C9a7EVayaykQ\n9v0e60AAACAASURBVNsJL77K2cp5nhDkmg7zorrrMWRPqdR2xuRxLFkzZ/D4o0yt+P34Y8x7kQSK\nET/9R59hmspaYgwKwFu992kJuN9PuJe/OywJHsXjCxi7/Pnp9QZX0rvw1c4i7Q4wXkKuFAApz4oU\nsJvzory983hwvWAoYLRJsKYkfBiJy7s2a2Sz3+dn/4PH17iQxR3mBV5E8pZpwb28U/fTEckliBgu\nkknaDYgZqlDSmQ4PLnLG9dnjB7i6uFBVYLIWO4FIbrY7vHyZYQIzeazWKzwUUvTDcYNrIYhvulo7\nGWLEXtbLzVFS0O8w3stgjUOPq8t8IT4smk6wtoPKS7kB7JxKR8yBYSj/jeURXl6q7d6jEwL8+rJH\n8B5r4Q6N4yW223wzv/z1V0pr2O0Y1mwQg/BnRtKXOoFh+sK0dip/W7qcTJPHNJV2V1HpGUxQ1vpu\nv8dut9OGq21bLEOEteBRjx4+xCDtsfy1B5FVekWWjq6M5cJ1Ad7i6gFNTWTFBiwZbVraF+kRJe8l\neO24MmuoHVOmVxbh/9kHVXwcHeF4L1nCNONwyDjJshzQ2QHzkp/Hy9c3+OhBfoHtl19VSd7EGHqD\ntbS6moNX7OP1zQ2Okt17+uwZ/vQnfypzPcldWUSFCoBQ5n8O2tXHrTqs1x2c8LW6rseVZBmPu1nD\nlsvLK1yM2aBytPjis5f4k/8rz8uvP/sCi7SSOh73ev6LzQiCQxJqRYhJcaHNZsAPfpiz3N//5GOw\ncKzi7PGT/+OP1cilEJBkYxh6VzN+zDgGYCdhqg8GJPFmCguOErYducNlX5r29ugPs3K3oq/qoTEm\nkYjJxePRJ8SSEbYJTCWcTkotSCmpXHKhl/3lH34PT64zn2veHTAJN+z29hb/5Od/BgC42+4wJcL4\nMPMjDyEqzAIySBJGPlw9wXcf5Wfxez/6Pp5ePUQnJiPC4Kt76foUPsdG1n+YIp5cXeGHn2TH5pOH\nj/CR1A+PzmKU88QIvBZRhJciYfMu47013QvYTBTRCdDeU1cLlJHB3+JUOddpt5N5Zrx8kbGfn//i\nl0jyQJ58/AjrdYdPnmU8JPSM27s8aZ99dYeXN/nGQhpgbFe7bcSgJLxEXDXDKWmFfZGeWxavXWHm\naVYg2wDKTWBkTEWF8IzTwmFnDK7KTrFeIz0tIGX2k4qNWgJjEgznOB+1NZVz9oTXVXi/+Ty1VXo/\nDNrFZxhXmZTaNIMs+NZhmuFKK6q+x7LMuN1nT4pdhx98Ny+YcH+Hr4Rr9fDqOvfgQ05LDzRqCdSb\nN7fYuOLBTLh/nefq0ZPHsIb13kMy2AplZb894pefZ6zsB59+il/+8hfyXPLL5Tpb9ZgMKYZnTVcF\nH41DPwxwVBQgDDYblue0YF4EvO4J61E8Us9wYwBc0b1aYAqvayC4UQzMhctKB0V3HYARY2Gcwapg\nmtzn/mXI7H0i0pZYPlaPY/ERc+GVMcMz4yBe0BI9nHjZlBzE8cKLmyPIFHE7AhzgS9F7TxDHGnEm\nDEIIRVjAMcCZPOddZ1FEDb33mESTKpOS83V3gkENljDKta/WA5IkHC4Hp97Wbv8z2GhgZDM5Rq8O\nhrUWK3ECLhxhJde0RsTAHiRGdUkML9hynCb0Ut52tb7A06sH+ORRTsZ88ugaD8Vbt0oHB6hz6G3e\ngK4EL32XccawzuM8zuODGe/lYQUf8OplDi+cW6NUUzATgtQ/hJClTopsxtD3SsHf72d8+UUOSf7x\nP/4M95KlefLxA/z4xx/jobiywXu8uc072ctXO9wfpMmjXQMWmAUvC8uEIeUYG2SqpEeKTV1iYYsP\n1Qvyi1ICBuTwAACuLy8RFp93NADr1ag1iYwawjHVBqKGIZSMchmEKDvt4pt6y6/VKNYjhrAoI9x1\nVhndbz+cLPmrpbiKDfb9gMMyVzliY7Be53kJ0Wt34NVmDVc6B6WE4Gf1Jv28YLvNrvl8PIBKLSZn\nGZVyH52tGlXz7PH5FxlH+vSTT/DsWc4g393eyncBK6FCiqnOBUMzhrRETOxhJffvuqFKE606LUMJ\nacFRQpW7N3vc3d3hKH0dQ5hhnezSjuFKp+beIHJEL/t6N/YqSx2mgEE8yothhEVeA4c+q4j28my6\n0WmD2JAWleAxADqb4Iwcz+8B6eU4B4N5kfXnJxDymh+GDqUNbn6CtTEHoepmIUWpfoj6vVr8bBQG\nCTFpdUEvP3tzd4ueChXBoJc5X489HkkJ2tMH17hfAo4lVx1mLLFUWvTYbERaqGN0JRKJEcHP2o3p\ndnfAy+cZt7p5+RJJQvzeOqz7EZfr/DwuV5foCmWi6XbOHqpw+0AoLu8y3stgzYvHlxLSPX66Ujq+\npR6Fjh6WCIPa6SakiPv77Doe9xZffpUf3s3NjKOkyi+vHfruCiQJ2d39ES9fZiDu9t7DR1lwqw4R\nAfuj8HFSwEMSAN44NAyHrw+uIm/WOqRQiQWjGKiPP36GYRxUqtbaKvrXlhVKrbEOycwDEFBUFxpr\nR3Yyb1uspmNNirXBJ+w3lC2X8xilOBBZZTFP/z97b85sWZKlC30+7eFMd4h7Y8q5qrqAR7dBG2YY\nT3sCGjqIT+dP8AMA7RkCKgI6ZohooL42w/p1d3VVdWVlxnznM+zJJwRfvnyfyKzqDKGFxK4LESdO\n3HvOHny7r+Ebhh7OOS7ixxBxoHD96ckZLs5TraIfepydUZ2qbuCjYCzT5Cz2ZGw7DiNDTqSUkFqi\nIgjLagGGkoy7Hrd36X5e3dzgy6++BAD87X16L3jPnDLvHadWQhbpXiEURFSw2WYrTMjS88FH5Mqw\niBqHfVocrq4OeP32Brc36XvG0cGQhPNiWbNm2P6wS3LVdM3qqmZMUFs3qBQh4pWBJiqTWkiEGBnC\nIpTijcH7wAt8oxTO10tcbNLxHW4fcLel+mBwyHz6ZVvB5EngSfOfZpL1Hm7KG6dGW2dyPhnw5t9T\ngjdRVUmea94HxhPmhemfXr9Dt0sp40lbY0Wa+21dMfn/2bOn0PsO1/u04MsY4KlcYkSEkVT/rCrU\ndeELu8nicEiffXt9h+v3qRFze3OLiZYSUzfwIbI7+Og9NB2b60Y4SksRFKomLVQ103r++fGYEj6O\nx/E4fjbj01JC59FRwTUKWSRhRUFqj30HhVJMvP2ww/ffpZW4Mgav36YIrR8iWmL3X1xc4MXz54xU\nv716wNvXKdx0o0S7yNw0hcHaFJojkXQzNCLxl3IrOhQ11Hzsk+VUpq0aeIqwbACDO188O8dqvcCW\nzCNjjNyqbyrDEIm5rpKgP3K6aO3E7joxhuJjN5fwRUqj2bA1/QQAwLsIrwtHcB5lRRTdJxlD1vZP\nnCzvOFUABLb7tBP+4sULfPHlFwCA3/zDP5RUuDJJolnkQrbHgX7HjhaxzR2vCG0kd06VNLAk03wY\nC6Ph1Zt3+PK//GsARRIXXjGpOXgBhFx0ryCoSBxCUlDIRd/ElaOUc/KIhLDfHya8fZd4kN9++z3e\nvHqDPXWSz86e4NnzRLBfr0+hZq3zYbAz/S4wt7MSGoHgBf3gIAURpH2EDwFb6mYnTwy6Ft5xd6Wq\nDEwUuCRO5HDawdOz0Y+OSf4nbYWWOsoeGtpH7IkwHTzTCtGuKlwS5/bsdIHFooXJBGwFhnooqdBS\nt9p7z3PSUET87Zu36AjGcrle4oyaJecnG+5kP312gaDvsaMOsRERiqL9CgErSj/XdYMVRUEiRtih\nx0CMgH5/QN+l124aMYVsCuyx3e1wTZG3dA5LagKMuwcEuq6VrrE+OaVzT8f4U8anqTVEQFA9pG4a\n5Md0mkZG31rXAzFyqHtzdQBkqnO4EPCOUkqlBC4u083+8vMnOFnV2N+n/3v1x9d4/yr9DmyNVZNO\nKIiAfrKFqGoqluhNBpxlQZmDBgAkC6bMII/FlFPM0jsFwEjBOCw4P3Ps+THpN8DHCB8Epz7WOnjK\naZImeUmtgFlaKVC6ZqpodgtRZD9+kNnO17dZGpo+a2YZpQYmE0/O4ouvEsH56voKIxFjtR0htEFD\nLfcxRuzvM+TB8ndl6RaGZ6gAQ1CD2kjWpb+6u8VAdYwnVMuyNtlzAan041zOmyXX1RwcAizXBIWM\nvAnttwfc3BI5+M0Wr1+nje/N67e4ur7huuH5xQWW61T/XJ+eoaX6nYCnul5OlTWQxRqVYoK+HSym\nIS3W9/dbhBBwoOsnhGCSfpwRzN1gIY3EhhbnL16e4wnZXnV9z93lSmhMtKjvhgkIlknvEgqajmfZ\ntDg7TfN8vWnQLmo2E46IvLkdqYE4z7AIQZvc9+/eY1imWqQfzuCndC3qSuOUyedLbA8dsrlWLYFV\n7iY2NU5J7WJZV1DZeLfrYKcJA21qfhxZornRGj1h7A79iHfvPyDQPLtpG6xps8PUQ9LCuGmXTHPa\n2DV+6nhMCR/H43gcP5vxyRLJBQhZiofOOwTaIZMsSQ9D5qnWGxz2aVvYdR2jyp8/X+PFi7Syvnix\nhEKPw0OKsK7evcVIgnFGadYqCiJiu/NMJm2URp2LwzMXECEFByOsYeUDAoMBw8wE9vj8Urcmp2Rh\nRjw+HhnMOU0O0+jYjLXrDtyta5qaw6SiqFk+jyU9lGZJD+8Do8DDTLL44xGAmYxvkt1dU4q97Bz6\nLqUFb6+u8K//i5Sq/ddPLvHb3yZw55t3bzHZiZH8EBETddAiYhHJDOmKMK8RDnXGjBkNRee63W/x\nR8J7/eqLL/j8igyvYsG8GASzGqQMyaAid5JkAdFa7/H99wnn9R/+9jXev0tlgt1+h67vuJsVYuKm\npWP33BlOduyGI28l9fzqc7YgZMSeMGx/+KdvoQSwppROKIkYc8fUcErpXURUAQv693qjAOpY7w8j\nSIIMw97hwaXPnmKH/dRBZ66jMFCk99VWCm1NzSUdoLVgKeUgADaegC7SRrGUM3JR/n67gyZs2KZq\nsCKZoBADKkpNtVZQiCysV6tEzgeAk+USZ4SL2tQNk5i9neCmMeWxdO/Xee6oClamDOtw/4D7uzt0\n1HFutcaTNXUMK4MVof7XK3CEJT9+CP/M+OQFi9MnUax9vBLsZtv1PYZhxDhQaz8EOBIw2/cjJHUq\nzp6c4eVnqWP17HKJSnq0TTqcJ2dLuI4WQythiCA6Wg8jDFxMn1cLhYZuqpiB2oCiEMEaVr6IigVf\nhM3EbD0SAGSMrBQRZxLCc9XhGIvC6jhO2G0P2Ge2+tBzSqlUcYrJogssADe7pkpqOFrwp374swtV\nseVSKOTYVM/ImeyqVvBjui5vP9zi3//dbwAAX758zoBIFSIcAiOyD909Il1XpUSp24DQ+3RMNng2\nw1y0Lcbc3XMe3377LQDgP/mLrwEAVaV48YkCTEsKIvL1C4ipzpcBwDFCU3pSyRr7bTq+16/f4/5u\nR9cl3Yx8bw7dnkUFN+sKhtKjzckCUoLvRwyOEexROFRUahACcDHrh+2hlMSTNf2fkvBZT6tqWA9r\nsg42WKa0CGFhBwJSOmD7QATxuy32RJsZfASkYdNcJYog5ugCtoRKXy4V2rAsDuemYkVWAQ07ZVHD\ngIogHJXJYGrP88ZUhm3lmsUChtJ/7z0mZ4thLwRqWvBOlg17LTR1hSqb61oPby2LZK4WOimEAFhB\nAHU6dxsCPlzfYaCAY1ASim7u+sUzLMj+6/zJE5yQa9anLFiPKeHjeByP42czPlkPy1RFQ8fkEEtH\n9m47DCP2fY/gsxeZYbt4oWusKDy8uLjAxVmi4tRaAn7As6epayD+6tc4O00doVff3XNBT0YBGRUc\n2cIHW9JDnYR86EhLUZ2xWS6iiHvHGQoU/FpIotpkXa2oWJ5GxZI+RoAJtggR3jp0OcKyEzvlRFF2\nj483kfjRv7JipHOOid8pqjn2fMyvwtwgNEcAFG22RuPAZhPAq+9SWvWbv/07XH+XtKaePX+G0y+/\nQE80ou7QM05KSTBYMsrUlcjRkQgShi66qVt0xA88DBN3kN8SUTrqUjYIwcNmHSc7Ysh4HBVgRFW8\n/2KEJDCkUTUuLxPF45tffsW6XotFjUO3w/19AjHvtnt8991rAEBdSaxJJ71daDStYR137ywG6tDJ\nqsV6mZo+EgG5d7NYNjTPqYEhAWadCcHcSKkldFCM2aukRlWnlNC2NW7oGjm/ZRll6yOkrCCzDrsS\nEHWaaw+Dw/eEPdSVQttOMDrjxASnfi6CNbAQwVSX7NRklEJLjYDVaoUNEdqX6yW73ByGCYd+wERZ\nUXCWlV1Pl0tsiIdaJQucdO2swzRavi6VVlhlCo+p0VNyc7c/4H2Y4Oi6SGkY8Fs1NXcEm7Zl7Nww\ndfip45MWLCnBbicxOkS2RpKI9FFBKFR1yzyoyStWVKh1i/VJen+zPkVDubwdHYb+gFMiSX75i5fM\nDRtGi7dv0glNTiBKxd9rrU2MdwBagVMiKcUPiMYxgt18gpuZSMywAwJpUWaRMiVKl0+UcDQAs3ay\nRmUM6uzgKwUqakebqiJTijlz8IfD+8BmEOm7cnHquNMpZyflnGOAqpSpi5eZ8HVV8WbS2xF3H9Li\n/3B3zWmxEwouJr4eAPT7gXWUggRkRvuvW1RtzYuoVpqtqaJQWC7SBN6PE7ZEZv2nP6bFw/oJhuom\n1aICIRkwuaL4UDcm8ZFzLUZq9LT4V1rhm68SXGFz9oThCVJG3Nze4vvvUs3s299/h7vbVDN5//4O\nL16kbufZ2Zo2D0rfvUNDNRQpNBPB7QS0y/T+V7/4IgF0TenQlUU3Fg6qSoamec+OMWKiDXGwkVOn\ns7M1Q0lud3sMNha/gQBk1YAQPbZU3725nXC+tGjqdM2FEvA+w3AkL7pKuOIcTZNr0TRYZOJ7bVBR\nmUVrxR3F7T6VMLJphvceNRHuF22DVZOR7oavXQghuXLn50aD68deC7S08NYaMDJizKR3j2KaKwWv\nHz5aWEc10yxB8hPGJ0dYOWKI0XMR29nIFureJZJrprRMvcdEeI/eW1ib0ewjXr9KRfY3f7yC0Af8\nq79MF+r8fIOL52lnuLs7wTWRn8PBQWuDJSmHauWTagQSrSHjkJQojrpMbI7Fjdc5n5xwQXWrWWIs\n4qwYrjGzXirRTSpy00JU18AG7FQyzazc27blJkWqBRVJGaAsYNa7WaG+AbFEqOgFpmXImeuIUmqG\n7UqPVD6NSkksKcqzvYHNKOnNKUcLoq4RfYQl0T8lFC/OdaPR0E548vQCi2aBq/epRuSdS3QdpIVy\nQa332mhMMd2/91mR1E5MSFdBYaRi8OgnRCLzSpUiDTUvwNJD4XzA+UX6zNVlzdAH7xzqxQoRCcFv\nxxFvX6Wobvsw4pbmy+XFgMVigaqhYnbVwtC19JNl5VptJDYUlVW1Sbi6LPAYy30SSvJ8EjGR5HPh\n3seAHUUVD9PE5P9nT1Y4zfbudxrX2w4dif5NDoxTEMbDEp3nYevR9YEbCSF61j/3VvD1F1IjULSW\nN+BVU2HFC1ZpSoXo0GXoxm6Lu90eh6w7L2Q6bwCLZZvmIBIkxdGGNnY9pm7k+VubwpIIElhRtLQ0\nEo2ROHRFqDObzQzjwM/rNBww0iKn8FF08WfGYw3rcTyOx/GzGZ/cJcxyKUKGEnFEgUh5tRsChoMD\n6rzCBkSqfyh4TERYffPuCnfXyX3F2ve4fFHhOYkDLk9bbE5THv70+RqkWoKb6wERGstl7uCMR7Wo\nHCEoKWY7IdV5YuSU0NmQu7M5iOEhBNhjLf2A4J+LR6ljeq2NwELVqGkXH11gVLJW6og3ltLN8j2F\nm1gcnbXWjFpWUgJxroc1P86jo0aMkeEYlVZoKS31qxYVffZhu8ee6gnNogWcQ6QddFEZhl0bXcMQ\nIXVywMVqDXOTouHbww5Nlr/REpXOkAEHEbNbCxF1RUzQEBoha/urUtsrvUK6FBDFlxGR5WAqXQjn\nXnmcnlYQPhnMjjuPwzadR9f1uLtJtaDdtsf5uYemdCf4iNyWEzFwxC2lQFVnwcl0HI4niOCURgnB\nkyBlF4prb2OIGOn/Ou+hZPr903WN8zNKi9cG+oPEG+rMjjvL560MWJY5ydjMnL194G6zhESTnX0E\nWA8rj+fnZ3hOstUny4Z5siJ47Cll3+732PYH9BT5LJsWJpvcGgOT65fOo6cobLfboe87lmpWRjDP\nUAkBnfm3boTxFpIiqWFyOFDG8uHdW1RE8leXp9BU09ysfjpw9JMXrPIwlSddCoFpoALeGFHBsP27\nVyM8wRKMARDSzXp4iLijkzK1x5mo4TI9JQT2xasbidpQ3UDaZBpK9TElFWNVhIjF1GLm4Z1RxTGi\nmEJ6x1iWo5AfOGI5xwDWzPY+ciFazOpZESmlyQalugJs1q12PonGAUAAjJ4RtFGOB2KOS5Jcq6kI\nusAa7/JPB8QptaW6mlZY1FlO0kMyhGOFOgsCGIVpd4Ciz65qzSqqkILZ+6/evMNms4Zusqhh8QcU\nokAGFq3BwabNaF5LzOmxNIIL21pJbhQE4eBCAFxOCR1sKPpVht2bPaLNJqjAZrlGq1PR/OF2wtvX\nWWN/h+1Dgj8M3QgJhVqnazGFiTdWQCHQQyZEhMl1FpXUKbikIAUvp/ONSigJFyQO1LS43x9we8jK\nrnf8eVp4nBNB+uy0hY8Wu11aUPfjgIm0saQw3Cqa/IBdf8Awkh6W1tA6lwIkp8+AYI8EQXPu4myN\ni3PSmVos0GazYwEWW3zY7tANA9OhpFFsoqJmm6z3nj0id9sdhqHj86/begbpCNDZGg8Bp02DqSF0\nPya22tvd3eFDblK4DhUtWNmj8qeMx5TwcTyOx/GzGZ8YYQlItplSiATUDL6ALWX0kNEhErq30hH1\nmtC2WiBbkIxDl/h9ANpNhaquWf0yigJcBDwXnZPlloWj4q2q5Gy3Dx91Bn/Yk8upVYi+IKNnYHYh\nQNsoaTC5gBx01E5DyVL8ztFlBkDGefRF/2dD4CKjEAm5PEe855BfKcmV/xgjJ1GpgSlYnTPOoj/E\nklMyGD8DTiEY1aykLB1RUzPQ1jmP7vaBIRiIjiO7ZrHASNHMbrrH26t3uCAtpclZqCFHVSv+zkoZ\n1p16+8fvQReKO4qpkUHRkojwIbvABEQESJqKUhXzTqUU//40WQZMKq1RrSrWnW/rBd8bOzkMlMZ4\nZyFVxIIMdiUc28JJodlgVYrSsOEObeZiy8DXWUoqNyDd98lbPBAZ+M27G1xTp/Lh0DP84f5uhy+e\npm74y6enWNcN1tQ0ut3vMZGsi3OGI2kbB+z3HfaHnLI1XMSPPvIzBCkS7ATgNuG6MVgTar2pND8f\nMQB7csC+ut8mPXY6F11V0CbDOGTxRohgTuA4dpimHtZS5B4DDIWRUiusKE09Xy4hngmcErH5oe9h\nbZZ+HtndZ7+7R7emyHdIc+unjE/sEkoolfNniZit5UPgySRFBMQESTiUl19e4Ff/8VcAgM2TNasI\n/Obvv8PrV+/p5CVhTfL3lPaxCIJF66RMyPEQMsajZs1vKUsd6Ai3RC+VnNFEII7yQDbVpX+zjhoC\n22+FEI5SRxYIjAEhANlROKeN5bPzQpawNLmlHSIYexVRoBRSy2J39tGiK+J8GRY4pvsUKIeUisN1\nIOFfgORLN9ImERGxWCxZUM3agbtp9WqNzFP2k8fd/T1a6gKFGNET436zntWHnENHelra57pfLKl3\nCLAE3bDewpEQnvYaURaKSUqJc11MM4QlWsG660IoIAiE7GsHxSlc9GGmqhHgpxGB1D1itNxC17qC\nzzfaR6ZWBSTi90TcmoDAGvum0jwvffAYrcMdpXfX91vsdqTWMASMdH77bYdImLNWaqyXLbSkzUTV\ngCBDiaBnxVSPcXSYhtzuL4lQiIGtxqAlskZ9XtjPVg3WBDUxUiGvdJOLuKcF8LbrsXcBbfa811Ux\njZGCU3ovA2LuTk4j7DSVZ08BOluAKcn1rNWiha5qnND02x867kY+3F9Dxux6naBR6eD/FODnh+Mx\nJXwcj+Nx/GzGJ+OwVJYMFoJTEe8CeoqcBAChPVabtGL/6tfP8Z/+VZI3WZ8vcXudduGbD9f4kADY\nCN5CIECqzLECd2PSDs3oPEiJmcFFhM4FeSUY+6WUKsXRfIwhcFQkBFin6mOAqZCCU6MwjKw/lX4w\n97RmP4+UHuZmjYw4AmzlNLT0xGY4LO7+RY4KJI5Br3H2cxDHqPc8pJSIsZC/Y4xc7IYIbOwaVMSe\n5Iu7roM2CoKchmTboCXS6+rkFLdk/2XDgPv7LTRFmmM3cJTovOXobblocUnMBbtIO+c4FjsvAOgp\nJen6kdMEUxvoGfA4QsNRFORsmFmdGWRLrGTzNaInie2Hux0zDew0QSJ3nSQma+GnvKsruNyBtBaB\ndMsSYjyXOlI5IkeDUkmeJN4FRs0no5OALZ3TwUbs6fO64YCaIlIEge19itbulntUsmLwrvcSUWTA\nsWZZohh1whUyaTtwxuIhWXsKNjCuLEeoX33+gp1pggsYyWXp/d0O93S9rg89Dv2Ec7rfQWrEnJ0o\nxUV9XWmsyen9ZXgGHwJEjsRqwx1SHyPuHkhee5qw2pyipnkxna0xUoR5t9LoSS24MQJL4mtK89Pj\npk9OCTXrT4EtsBACVNYZkgrOT9AkAHf+bIGTJzSpVwrjQLo7mworAusFuCO9KSV1Ngkhe+/ykGot\nsqQRlElEXQBQIs6Q0JK7g3nFGqYeljoUURZOQyyd6gS+VIpD3YDA0sVCFdnnMFt4pBYQUbCjMARK\n/SNGBoR6JMJsZfJ1Ai8q0YdCzI6uaLpny/eZE7TirujxwpX+nUnX8/fBZGUvC4XDU0qYF44oJAa6\nEMN2i+02TUCjFMZ+wNUhpT52GBkVDboaALBc1qhMghns7tPPehf54RII6OjhtpMv5wQJYxoG6Frr\n0R2KwJ0ieHwMgefBODkMcY/b67QQvH3zDg/0wFSVQUPASaUUtDLo6EGVM1BwcBZ25qJUjzX/qvzF\nIgAAIABJREFUPiCgdRHPY+aDlCxUGWICLQ8ZFNmPcKR2IaJlx2U3TLyx73Y96nrLtnIueFjSjZM6\nYCKIiYwCy7bFhtQojJ4Rx4Ei2WwLRSw32s7PThBokep6x7peV9sDbva0YTiBg4+oKePsXMCBuvz9\nMGFapP9YVhqXl2kTOj1ZQRqDMWuyKwmbu3/7A65IsM8OHu3KM/Ris6ohSMb8yYlBR44/4zRhuUrM\nlljN59OfH5+4YMVSGwme8+foc7UaEEahWjVYP0mFtNWTNWBy7SJgnNJkDvBoVxnrM8IYBZNzYkhu\nsU+j58I2kFDJebFRGpgcMdyVOUKVZ4xNbp+7MLFSqfUWPnu7zepCEVmQr9RdMo4oICDkifsxWl2A\nRQVjBEcVfdezioVSCogRWfhGQDKUwUjJkzrE489PcIWyA3HE9tGCxVHYRyOhsSkKRZFhWZ+c4HS9\nxgcywAwo9lPeeqZTSSEwjRMOO1KQjAGL86SyEbxnaIZSCqIiFDQJxWlpZiKDRS5FC8U8OS0Ngo+M\nCTp0Iz68TxCF/bZHWxEvsKq5fjJOA/rR4v4+Tf4PV/cYCd9X1wZr4qsuFi2kEDx/BAREbqhMI3YP\n6Zwe7neQJIdUmfpoo0mmvLk5ohAJWR5sxGQnBKKX+KkHqD6oBZhjOI4DAl2Eh+4Af2exo2tmo4Mh\nyIWAQSBVk0oIbFYt1rxglfqWQFlAbQwMu5GKMHRNg4ns9/wEDFmx1QIHWhyDTEKGI83z/TDhnpoH\nt7s9Nou02DTVitUVdOMhKwNJz+XBWhxont/te3ygDc6ODnrZQlMzZ6NqrKimpqo1Wp+e+W6ykKQ3\n78RPj7Aea1iP43E8jp/N+ERYQ4RWuRUsILI9u7dcdwgR0IsG6/MU7rXLllMs54ZiJTX2jGiGimja\nGi0xwY3WOFBIPRxGlphFTLWeyO3ogJySJEhC7rpFjiQEp3GBO1OTnTg9DDFwhy8ipUx5d52nhCH6\nWSRWIqJkX1/qTuNo0RGptOsOBcHeaMiZoN18p5w3LdkZlq4lhCjyyjH+yUgq4qiFyENAFqUIETil\nXywWCEJgyrWaquKo08NzijbsOxzu79HdpainNhoXFGF5b4tDdGW43tQ06b12UafuF313S/rgTd2g\nod3VmAqTnbAnTuPVhwf84Q8JFvHuzTVUTDv1ql1CEerdugHDNKLrSZb3YDnVPT89x+WzNPdOTlpo\nIyEJda6geM71Y4/b+xRdvvrje0xjOmatGsToOb2rVM2lBkgBn41PPYAQIJHhPA4qR89SsOS1qjRH\n/l3wGPcHHOgYpqghSBxwOvQQlJ6fX2xwsjYz9L1GzDzZOKsdB4+RgKuZpdBZAU9R4D5K7CkC613A\nlGW4ZQWpA0ee267HNYFtT9cPWDRZatskDS8g4SKsw0BmDQc74dCn5/L6foe3d/RcDxNQVfBUBxvC\nEpYBzUDM+loqmbECwOBLnfOfG5+s1qDIJy4GVyRRJouR7MGdn2CqCk2bQkkfIqYp/c7+YYvbq6Qc\ncNje80LUrGusN0tm9tvRcx1k+9CxJZcQGlp5OEm1KESmLASEIw2XvDAU7FVZsKydEHIu/lFK+INF\nYY4cyLQMRFZXGAYLZ0tNZrc7cDFYCHB6UjcVjNFcXI+zP30omKxUqKfv/EjIL8G/MnRDIS/WkSQn\n+FxDnGE1JEAwA+k9WypZ63C/OzAZWsoyGUSIGMnO6cPbt7h9+xa+z4Tic+zJWnx1ukbLi6Fk1dTc\nmFks6lk6H9HSg1BXmksLzjvYqci+HPYH3N2mVO3Duxv0u/RztV5guaJz1x4QFpGwV9potOQtefHs\nBE8uU9F9tdKo6sAwCeE8i9b56LGlc3z74T0XxkVI5OdcNDeVYd+AECLbkcWoYJTBIhehZ+q0URZz\nlHqhGCPmIBLRnSaVdRGeUtlp1+GU0qhnF6fYbBqez1EqiJxOQzDUJtWR6SVtdDfdgJHkl263Hd7T\nQnS177EdckkkwnuBjlZfjQOqXP/UignrBzuy43ddVdCmwkgL1vZwwNVdkvd58+ED3hHh3TmP0Nbw\nBHOYRECkJkCzqJjatx06DJRCd9NPX7AeU8LH8Tgex89mfGKEJbAkoa/aKMiRdnTnsMtt8KmDNCt4\nWj0P+x5nxKUKo8dhmyKnh7tr9HsKgZ99htW64fRruxtwfZ0+7+bqHh0ZaCZTz1hExaJDYHT8LKqa\nAdFyACAFOKqy1lIqV7pm+WeVlAyPaJqmtI2VYicUHzzG3B3qRgzdwCJtGVQJAKvVEivaoZZtAz3T\nmk/fl6EM4HN31nGkksY84ivdSTE/OeSAqoBUS2QYuWNaac3t732XoAVtQ0a0sZhkPtzf4d3rpGn1\n4c0HKBmxIPiDMZpT/MoYqEw4h2bUdf5bzcjL6RSoO+ZGJsHroGCtZ8K5qQwXek9PT5G9kv0U4Wzu\nOE6QdUBLEUndtFidkKvx8zMsV7nDFxCCZ8E8H8Hd3KAUS8g4RIy5g2wdpJB8XYSSzKzwQTA3VJJQ\n3mad0ty21tgdqMwQIiPdhTIsSDhahxAEKABC31sIOqdFpXF5kSLD07MWqhI4EEzCSQGTo1fFalqJ\ndJxhHxR6/PbdNSb6gvtdj+u7FBm/f9jjgSKsznpYF2ApnXTOcQllcAMeSOr5arfD2SYd06JpYHQF\nS1nK7tDh6jaVCd7f3OCByiAhRrjtA6eB23HEntaCVVuhpu7hME3YdzTfKCP5KeOTU8I659Wq1G2C\nm1jAa9VWGBGwv09pw93NEl8+TzWP0+UJWwidrpfIDtUvn1/gZLNifa3d/R431yktuL3boevz3Wqo\nxkR4lTBvOSsWy0v0AppYXDMqWlLTaDFSS3uaHExbfOukEtytXPgFwwiUMow/8sHDU+vYOw9rHZNE\nYyi4p9WiZW2iVifBvDhfso58DtPw1vMi+nFaS1c7/RkFT9KPMWVCCLY9jwjcJTRKoabFeLvrIILA\nDelcHXZ32FMn8O72CnekzuAGi8VqycRrXRnUuRZVNahkwQExPISO19qRF8uAgKFP1/Kwf0B3SNSN\npm0gpEZDJN0nT84QCcF+vrlAtARrsBHjlI5v8FtEZVmMb7k6wXKVFqzNyYoR+yEGTJMti0zQpdOr\nNBraTE4vzlHVZKgQIu5u/jFZgtH1o1Itpmni1FVXElornJ2m7z0/P8X722yt5jHRgqWNxkSLUpw8\nYizYqDAFrOm8v3r5BF99ntRVV6sa1k1syTa5iHZBm47RXBaQIqXX6f10zv/4/RtMtEDs9p7hCtvJ\nYZJFvyroCGsLYFBS6j4c9ti6tOne9AcsqK7c1BUEii9miAJ9XgBHj0OGnHiHcddhT6n36v4e76j+\n2ZgKRpfUs6OS0P4xJXwcj+Nx/P9xfFKEpbWCJsCR9ZbdMEyr8QVpWe36gNcf7rG9TQW5uw9LdF+R\n7k2r8ex5MtkUWsMTL/Hi2VMsmwW2d2mH+v7ba3xLbtE39z2EIPfZEGG9hw2zKIOlik2RBEHBhfFf\nMalMAsDQdUXiY99BU9TYNknFMzujoBZHEVr+dBUEu1QbqdHWdTF0heCwd71eMdfOI8l1MD9wLikj\nwHxJIQVzWxURhhmHFSMDHwUkg2IFpcNzjBZHXcCMJB2wWKRrqe8PCCGgJ0Do7u4OltI05QKWdG/0\nZoF60aIibE67WqOllE0ZPQOygkO8/J6bRsbMJRdm6pYdDuwyDQB13aAyJMu7rmDol843JzAqvR+9\nxzCk9OQwPsCGgVH6y8WKVWjrui4uOTEBUcVYIkxJ91objQUBl58+f4JwmX5GS4W//5v/BzZzDqVm\nwKqIsci8CA8Ej4ZKBp9dXmC/S/Pr+7c32FMZYxQOgTregnSyMmf28myNZ89ShPbZsxM8Jefn1bJB\n9B6RroPSBiozPybPlXYRwCBeQ0Xu728eYClT6abILkFjlAiUvutFcr2WGTGvBQJdl/3k0VFk97Dd\nQ8RUtF80NQQEm/9qUzEyf/IBju5555N57UAg4bsQcU9a/5UockQhRua1BvkvBBxVSmHMssjOQ2Yx\nr1WNNXVmntye4MP9DvuHdMC37zv8/reJ5Pzi5QbLTULOfn6y4hQiBIXrd1v88Z/Sz/3uH/+ILeXe\niLqAN0MHSIt2ScC2umUhvuWs3uStY+Bj1hiSQnINZPuwxTXpnJ89WWOdLZ1qlRal/ODpoq4gZ2Uh\npSTaLENrDGJoj0jJ+YE1leGaUV5MeAGUc3hoeb+uzcwzUZAGU+ksMp0nllSvrIF0rEe6WaLYlsUU\nlgNAXSl0fceWTlW8YOdt7RyW1E1cLFdQxkDT4nBxfoHFOqX10piygIq5FySlDaEobawWLU6IwS8g\nuUNc1zUqHRDJn0/AYUFUjVYbJvo64bBY0H1pGrioOL1LysX591Wi8SBtcCLIIsgoJUK2JQuO09DL\np6co0mSR2A/0MAlFwGgAURaCvYyQIrIE8fl6hV99/TKdk6mweJ/m164b4KkA1VYtFo3Gmubv04sN\nTs8JGFtLVEQzq5SCrhrI/CBLwR3jEAIjX6SU0JTiZ/Ly++2BRQqD0AgZkiEM10y9jxC6Ql2nxTIG\njymTnOsaoHS2cxM81bnupjERo6kep6SeSXRLIGvg1w2G6DBR2aaKAm7MC5aEsel3XHSAzJAX/OTx\nmBI+jsfxOH4245MirOAjdncFL5Q5c36IuLvJ1IoIZyt0u7RKv/luC+dSx2m76/D887TLVq3EQB21\n26sHfHh/h6t3qah6c/2AgSydEt6IiowKWKwMlmRdtN60MCZTQwpA76hLxhGSgCNgXnfocX+XupCH\n/YExLCKmb+INedaRwccRFqVMCj/cIBhGNX8dxJGJxzwiiSgRVmM+Co9n3U8pBEc0iAJZOSvpks34\njbJoT0GU89GILO8x9gfc3V5xEb5tDBylFQYrVKSP1LQL+FCwTItFi5qiS20q3s0jftgkyG4+QJJz\nyfy8EMGWXVM/QkbBXSUlInS2hZOKo2JLuln5yooIbqxEB7iYUzAgZH6k1lBRIs8fNzp4KmmMmDil\nETOLuJjxVCJHYhNGm8nYkc0zap2aThk7KCUQWSVX4Nllun59P3ITRZsaptLYUHdzvWjYCDVxFDPO\nT0Co0oEVIjImK0BCZkXbEJFJt8xFrVpkiHOERjwy0yDsW0w5C3NMEVlqJ8TSPPHRwFKXV9HcZRkg\nAaYspewjR/Hp+mnuAMmCJ1OaGxjRV8V96c8o6X48xJ9CTv/YaNplXG1O6cBmHzLnvtFikd/7EXGB\n9Ps/8s6fPpJSj0kdtfmHlrY5a1R99AXRdexuAqTuUfDlAguuvcgj3awQ4kyzanY04vgdqea/E47I\nyuU6SPq/wiXELIUTPLGK6oKQAtu7O5ycrPkylL5O+TMfTTGpPoZqFIRDxPHtjjhakY84jB/duNmc\nOv7W+b0vPzCNHpeXT/DjY/YJdLOy3n4Ms6P/SDpazJgBdMb8GfzVMwG6GCKc9wxnEUJwJ1krxdc8\nRhyBbv/xN/+BH8yi6khX5WgHO56LR8/EDOzLgFJSYDyGqfxwfAxe/ilPqABwenaGQzeW35g9CMfP\nzfHnzzfSpKU267aX1vNcsAQhFtC296GwQ2ZzPH9G/jylNNdqj1yfhMD7929izOJnf2Z8UoS12pzi\nv/nv/ns64HLwEpLhBUopaK2Z6PqxR+BcGYEnCQIpbZYbW6BIsrTUpYSRCrXKCpKiYKuCx5g120M4\nmjx/83/9b/jr/+rfIE+8cRhwIEH+cexREyTjZLPCetmwBnjfjeipYJhE+qh4a4AoM7TCYX2yZJG3\nw6FjTJbRhtv+lWlw2HUYyP5IyQZZdmK52UBSjWjyFoKCrKat8L//r/8L/u2//W/TG8IngUQkAnEu\n/AchEaLkAqtHKgin7wEXed1kiwknBCKKrbmUCohZiUNy/SQiADJCUyQAWXBdMs489qRk30Epgf/z\n//i/8e/+3f8Ewd6KAkUmaEb2DmnC90TNGTsLR0VuVTVYUDTdLJcwBBcRSgDCM8E4eM8Nl6ZZsJ/A\n2Fvc3j4wHUxrhc0mY51WMGRRFkNkiks/TPg3//qvcLpJdTqoglqXwnDRWUgNITRE4QeURTdG1ASN\n0UowQdy5CQjluL2bkNMUKQtTwLkkcMgE/liMij8OMHgxEAJff/01+qkFL+TB8+cbpXguBu/hqBEA\nJD3/bO21XC6xXGR2RsMQH2UMlJb8/ZN1jDnc7fas7nE4dBj6kR/0drHAkkwmNicn2BBebr1esZmM\nMRr/8//4P/x7/ITxaSlhiOg6KoLOzEghZoqZZAzBzHytjzpgeX4H4Dh6EWCjR+ccu4GEWRG7qSrU\nRiLOdLPy82B9xGBz18IV6Ru6eV//+tfFiNRH3BGt4NXr17i+Th3JQ3ePtmqwIIWA6ItkzWqzZDmc\n1apGlFkp0sNUCh1RV4b7B9xtyQV6iKhEnrgGw2HCNOToa2Ac0Pr8GS6fJgyOg8VIriq5cZDHfFEP\ns13SOYd+sDiQLMtkPYNrlYjFsTqUJkQIoRhkpDs3U7iQqOqSKrZtzeDH+QYkhYTOSp9SzBxcKvxw\nxD/xOp1NXoiFKmYiyiR3FoBkfDLFRkra8HI6FOFpIQ7dBEFpb78fcHe7ZekZY2ogZlWGgIYZLhEd\n8eJ2pFqQt8zgA/NkJTxLCRsjIZWY3QPPJQclBWSWb25b7oylVNbD2nR/h2HPr6MH348fo9aVZku5\nfj+mjXZxecoRvveW095gHZc+IPyR52bTNFgsaC6ul1hRF7hpW2g2BTbQWvOds9axvplWku9fUiwt\nhZLFcsWbxMnpCU7PUoZ2crLBapUXrH+hLmGIEcOQ+XjhI9mOvGD55GaTNZ10YOS4nGlKRaDYGAGJ\nWJo1ra0vZGoIaJ3DdYUYFB+1noXbkw9c95qsZeR4vsJ/+Z//JROLlTK4IWBks1rDZy3yN+9wf3/A\nKXVx2qbBJXW2Xn7xAi8/T3pPp6ctpkC64cEjRo9bAse5oHBzmxaOu9s7JrdKGPhJ8GQUwuLpkvhv\nT1/gl7/+Jv2cTosWAAiTJ2a+ToIlfYIIrBnmQ8DQj9g9pGPq9xODcLVQHO5rXTTWfYiYbGAgYJKI\np1Z+LQCS8a1rBSlDcpMBpRaZ2CvLRmV9YJ2ngTa1o2Vplk5AzNO+CBIEon87zi2F9EDmrsLB+bzY\nCjjvMVKUEGwA6erBuwGOkN6Hhw63Nw/YEruiqReQIBt2s8A00XnoSKkUsL0/QKBokU3UpgcAh8B1\nHwGNEIr7uXeFD7poW6xIHO/s9IzZDkoDCA6TTfdpv7/HbptqqfvdAQN104RMm2VekJM9XO7w+aOF\nihcyeu/i8qxow08jJuL4Dl2PkdHoSRc/w2+Wy3K888inbprCaqgqmNrwdztbNK+UKnO0OKen67Jc\nr7A5yVHtBudnmZh+wjzbT1mwHruEj+NxPI6fzfikCCtGYLKZ3uJnEVZZ95IZRITOUZARcKEoYc5N\nQecRVpxFWJMt6UqIQIaSxKCB4KGo3hNlUeMcrcdIqhCTK+oJeVs/PT1hAwJT19BEZZgmi+0u7XhX\nVw94d3uPqUuf/8XnKzx5+gwA8OU3X+HzL1Patjlp0Y85dUi+gbkZYcwS794lsN2r729hiRphR4/o\nNfIe0S40WgI7Xj57ihefvaRjkwiKeG0qVZPZeSbEcq1FQODIKyI4z9ywYCPgssphBUURpJyJLiZ+\nhmQZlKTpRse2VFiS01Fba1SmpPUQESKDlkLkNH4cHXa7tIPvtnNumPjo7+MRkeSCJjJhdcEicK1r\ngsl1qqmHn0ixNERMzqdaCVI0kjuLQzdhoNR4d7/H/c0Ddg8k8rgYoak+WdUtGkoD22VVZFMOAyAE\nqiojXsHlBR/Ka2c9wjjy9Zur8TZNg7ZNAN3lcoHNJtVtlAaiHwGkY1i0hoUf+67HQNFQXdfArKit\nlOZCpA8WgtLzGMEmHznyuXh6xmng0PfoD5T5IEBkv8eQfAizIe5qtcRymY53sazQEl9YGwGVyy8q\nQKtYRCe1ZLVfIfysVhkQvOMa5XJRYbmk+tiiQUvijk1jOLIyn+BL+Gl6WLHYWnnPhitciwJyVyAW\nHSkELjSFKDiFC6EQjwO13vOCNU5lwfIBjK6XIkCLiGlGss2SMePkMFK6OjmHedkeANzUl+K9qrAi\nNcfnzy9Zyubdqyu8+fY9Hkg+5evqczx9mRapl58/xemTFCq7MGFP0rOmqrFolrggKyfIBk+eJhiH\nrL7H4SF99tgJeGu5Db46WWFJNbHFqkazIARxLTARCDKQEElepJSc1208MsAvUscsazOJ6KFEJiWX\nlrOUYP5hiBFKgEG5WimuWy1XBqcnaQJXSkMJxccQ4GfHACa57/cdtg+pdpe5gLkrxmP+Mpa/Y0Tp\nUinJ36V1kSS2zjJouesn7A4dnMs1Iw1DFf9psNhTCnjYHxCCx5oK95VpWA12+7CDOCWJXjGho/uZ\nF0Huw4qiDAshuLgvgFS3Ycfu8gDWdcUlEussun5Pn+mBOPFmHqJHldOy1RLLLEM9BcRYFoGkzzWT\nQ+LCu8DHTlEXTzYMBekOCiZvUN6xZlYMEXVVMfh5vW55wWraihdrISIibSSpc1mYJUqr0oAQLaeE\n3lmMQ8/iAnVt0NC8qivFnMd53etYEuDPj8eU8HE8jsfxsxmfbFWfC6JRSu46zGOZ1IkCb6Eyggu7\nYqaWCHHc+QiRJbHhfeT2uwuRuVxWeVgdYFkFESziP02OxcWss0eYIAC4eveezQmaSqKmSOfi7ATd\ny8Rv/PKL5/jD777Fe6JVVK3E2UWKnNq1gYsp8nl39Rbfv3qb3m83+OzzL3B6knbrxWqNZy9TGtks\nGnRD+ixnDbwFGuqMnJyvcPk80ZTqhUIg4KO1AZ5eK0rRuOMZ5xireNRYUCpJ/gCAaEphXMoCUtK1\nhKaum3IBQgQIupZVI7BYkUrFwrAqhxEKIkowOyVGvqjeA2OfO3IWI0E2chRJh0YvAre6s4lCGrmb\nSc0JWQrbUqoiuxM9Bkp5t9sdbu7uOdpfNC3WbYIhhOAQQna8AZbrBdYLiqSi4qZH3w9oqTMWIWGn\nLEyZ0ErcdQwFupHwcjldJWqMLhI3GRpgjGZFke6wxziQKUewUDJgRSlSVWnU9DurzRrbbSoljOMD\nIgpvNMR4hG3KNB0p5EypI13Ts5M1+wgYlaJtALDjCDtkc4+kKFtRaaVtG+aY1nXFrjkheI5ivZsw\nwbOufl01HGE1teLu6zS2OOzrmRKqh6PjQYyMHkh4x3xCPz3C+mSbr0wMDsGXCVjEFiEgKS3MN1nM\nwGJydmgRIpNKQyTwHy2GiJwuxlhe5xuX61MugruJk/WpnY/UHlYzRDkA/OF3f8SC5Ehqo/HkItWc\n2rbB+XlalC4vT3B6tsLdNkEe1qctTs+JVLtQGEgn6O7uAa/fJI+y9dri9PySsSbtaomLy9RN3Jye\nwpjEjwyTQLte4uXnaTH7+pef4/Ov00K5OWsgcpsrBk4FvPNABEtRQwR2gQ4yFIFVKaDWLSqqbTgb\nOSWUUvBDlgCu6fUwTJCHEYauX73Q2JwQr21p0FBdQbiYNg0O3wUvLtY7DCRTPOwdLHXdXPixvvxs\njswRiAxYzamNKptiBPP4XCxYqd3ugNvbB7biOj85xWax4vOtF5SaNRqNabGiBWscAvbkHOOsRT90\ndGgV8iKsVGrdO1YmDSzVI0QpaXjvEREhZU6LioNTjAHjlFLMfnCIjFPwaNuKTWllW/N1ds5xSulc\nUsRlV6WZbDfLNSM9Jx8DfNfrBS9YAgGeXnf1AYf8eSFxMPKCYSqDhmpLptL8vrOByw7WWoxjKPZx\nS4+moe6nMmiom9g2NRaLhjFpdprQE+bRbk4LmkBITu/EJ4DXP23BkgINydwK4aBkrjPFgsmivDpj\nT4zRqKvMOldcRIkxMujThQiJQntQDiyhGwG2OBKJD8BRho+xeNjN2s8hzFyg6Yb+0+9foaFi4tn5\nGefpq01k2snmdIXN2RqLa3pw25aBc0Iq9h6EMHCO2t5Tcn7OD09Va2xII2m1WkHShPTSol01+Ozz\n5wCA/+hf/QJf/+Jz+t4lsgtu9IFJuUlUbYaWj0WhQUJwK1spBbOooImmFP3MSVmhUCOCTrAQJDE7\nZSdWAVi0hiWMjZS53IFoI3xMDY50HdJVTZ8XMdEiMo0OgfSVdKbrJKxEvmjl92KJmmMQiAHw1CQI\nLrKwXrAzWkcUDFfY7zrc396x7XyjKuBpPr6IKrv3GIPNaoUlLWbbXY9+IujHMKAfs8uSY+wXF7dn\nEtrZCllIxfPceQfvi+Z/2gzyJhvgfMYrTgTgTDW5RZukhoFUnM9ze5qm2UKRNNHyvJwj36WQHHHP\nvSjzhU6Aggw7kQUPqQqYOzpP/pe5NlrqUeneZQDyLDhwDs5ZpgvVpkKgBoYSipsHRkoYpTlgsMHz\nopk08Ok+/XgP5p8djzWsx/E4HsfPZnya4qgQqIkgiyghSYLD+hl/SBACmjtygh1stSo6TQGAzzlN\nSIGDD7MWey7PiNLZEkpQN5HSkxjgMqo3BMyCPAba5Uhjtxux3acw/dBPsHmnjBGKzmm1WWO5WkKK\nklploKxzgJIpzzdmCa2yWaeBFBKB2/KBd2utijxNBFC1GmcXKfr6/PMXOD9Pr+tawo6zHZ2ioG57\nSOkxgSqViCw9E4JAvn3Z1CBS+0kawZFYnLG5YxBcBxqmHhEjqgWhnZcS7JEbBfvv+RgRYQsLTZQ0\n3MaIQOlqdJJT13z90meVmlvZH+fba4q88vzxPvBnSik4UpdSc71ICgXMZGOkKMwKH13yrgTQtgaL\nVYUFdWD7MSIKigj9AEEprA+Wgc5SSe5cAoCEgqLzEUqxlHCIqb7D6rBSMPA2wnFKGYKDoUjJmBbt\noi4duboudUiAU7kQM7UsR1XzbuBxxadkU+lFdzhw3WkaptJdDHTdAEiliXOZa6QiZ37OwYL8AAAg\nAElEQVSIH5cXMwIgBCCwmlvS6yLp6ADBkWdwDpjX/Wa0PTHLFhAxq8v9C9awsr5R0POLBXhuwaY2\npaCbJ5BwGun3SwtTADODz8QllCrjTop78dwAJqOiA6VPCIEnkBexFKSFYP313JIWaoGB6DP94BGo\nFiR0BZ2Pw2g4FzB09FD3AdNAoa2VUIT+NrpGTZK6VUVhfcxW7iOmIdcsFHPypHIwRmFB6PbNyZL5\nZpK58gROoF/fb3tChTM8fuYCFBlFDykSZUflIm25zjIWJr+bPHpCofvg0CwU1idUAK4VQwhCKMTU\nqAKCj6UFLQTrlUcIdnORM26jG+dtmD8V+8+7IiVdnKsKSAhOn7SpsKJdZr3cYLNes0z1erlERal3\n1IE3xbrVqBrJuvKQDh7U3AgjBLkxOWfZkbtpKqSCm+BrW3jQgh1zDWmlCVajUDxPx3HgxUtJiYrw\nTuvNCqvlAhVtkCEG9ASn2O12LNstiebERPkZgTj1PIpix5xoD6RmAiPdrZ05bIuyYImApBum+HO4\nfxPKrfkYlTIfKa2nZ9571uiKIUDEeKzpNrO2mzH0j6AtP3U8poSP43E8jp/N+GRYg5ppM5UAP5ZK\nfywM8/x3Ti/S7uDLh80lfhHIGDW1tnPRXcgIIfMuEhBFYHQ1YuB0CQhFawpFijW7imjRoDskvt92\nt+edp21q7Anp7qyHiArTkD7//q7DzVXS6Lp8/gRVnd5v2gqfERRCKAmtJJSo8lHg5iYBTz98uGWo\nhQsWQXqOKr0v3ZMYXXHn0RUjvQ/DlHalnDqLmfJFjAyOFYIC8Dn3LIf7zsMRgrrfWbgpfX9dazRt\ngyrrOSnFqZ4XE0euMaYdntVNEXiDl5p4hwBUHSF8ToPoECE4UElvlM5tnL0VMYuEZ85CUgnuPpm6\n5YbB5fmAaRjhCSB5ulqzjpg2GoJ8K02deKiB0kAXJ7iQXzsIei2c4/nbVBqAYDR5gjJQ00hI3uK1\nSYDSIhmsOBUNwXFnfLlY4oQgL6enRV4aAA6HA27Jz+/6w3tYW1RYRzux6auKsshOO8eFb4gfRljz\noZTizqOesRWCp04+N3P4UYSIs2K8kFxSUVImv8XczDkWqzmKlsQsDUzvZYhIkXUKPnD3+s8JS308\nPnHBEowBQQjc/Yg+cOs25eSRQ8kgI4JX/FrkB0HiaNKKOE8jPS9sQoRZ9qABIdn8EyKSiWiKknPH\nRUrBHZbcau52A4TInRDNdACBiIGIr8Nhgp8kuwC/e3uDP/zhFQDg4vkGT56nmtNyvcBXberweWdh\nTANJB+V6h3evkhPN1YdbTFPO8z08AkaSBr69foCldEDIgBOS3VhtWiidUe8N5l1CzB4YMctactEl\n1+20NJAEBxgGi4GoMtNgYaiz01QNalNzty64GexARk7jQwzQKLW4AAZ3w9QSC7LU8s6jbrLUD0Ds\npGP9KL6RsrQdCdd11MKP5d5nloNRxdx0vVrifHOKYcxUFgNDNajlsnRLhUyWoDbLLHvHyP4oChdC\nIDLNKV0OwRuIVKog3SHyD6Q6HwJ8zt99efAWi5pSy6RK8OQ84e1ONmsoJZk603UH3BNp/vb6mjeM\nqqrhfcBENa0oZ5isEKHyZn4EDymbPy8WupRwlJphtpBweiKWBWuel8013PLiqKUCVOm+a634niVp\nIkrjJTEDRAlSSn3SF/UI7wvd6xPGY0r4OB7H4/jZjE/kEkaO9+fKmiF4FimLIVDkU34lzjhMx+Hf\n7HUUKJh5D4FsmjkHx81lctMuyWmlBGROA2c4sJwS3t3v8PRFAosuFgvuKg39iJ66h9vbPe5uH1i+\n9+Fhh9u7lBJutwc8eZZ+XwEzMnBFOwydv4sIhJqudMWo734c0R8sbj6kdPHbP7xjL8e2lpDfkOnD\n6qTw1bLnXyYvQzL2KplYlFBbhLKzqijgiQEw7CYc7lOEoaSEqbMLjcbUC0yZXCwCJJkl1IuS3iTe\noCqRdQwMItVaQhM63lSKcVI+Ar/97jj1+7h6W4QsZULcZ9mcODFSPcIgErsgBMVpbm00lm0FbzP5\nuezclVmhbnL3b0Q39ix745znwrPRNTTJXPsYGCkupUIinJemTX4dAkpX2idAcz7uOCPcS9litc5m\nsBtsVtlowiAi4LBPpYD722ts7+/pskYYlTF/Ekpp6NyBFBJZx0zNlHOTWm5J1dMPK846ggsFAOt8\nScf4HLKwoIcn7l/UpcslhSjpp07SMjmqqkyNijwLpVTMOjCjJ1lswdcsZkD35LgT6pxjECyzB37C\n+OQaVg7vwqyLEWMoHYOQLmLJdcHkW6AghTFr2+YG6LHKef6/Y1KviBJx1lNj9UsJvtCKbjhQLKes\ns3j6LBm6npyumKkuYunKbe86bO972EwL8n5WmxPQ1N6WMWLMdaHOQkvJ4FgjDc6JVPv08gz3RLe4\nubvHftvj7ZuUAsADFVlGXTzZ4MklIR+j4vOLJCJVtNLBC5YUmmsDSqQ0PNJxW+9Yk2rsLBzZXKEq\nkBDXW/Rjh55SVB8tqpbIzxuNJrsqGwWpFV/nEF0huyvBNmbGaISG3pdlU/kpQwhkaXjE6HnBgvAl\nNQ2O00ijBUyleL7YccJETtzpoSL0uJ8wjhZDVsywnjtsxhhOj0c/cbqUN7pMxE/t/zQCIlvMWR8h\nWKwu1/nyzwV+GI1WXI8FHKQQ8FMuQexhKa01WiPbek02QEgFrWedU/qIueIJMFswc/0yMeTp/B13\nHsdp4sXLOQcRy2ZuJ8sLiTalZodYFiwBApjSedVNy4a6SmomO49DhJI9BxnBBw5SJmv5PjnnUEWi\ncH1Cm/AxJXwcj+Nx/GzGp0dYc15fLqx/hKlI2KscCZQml4wlVURASecgEESclWQjZKY8xFJdFjFQ\nNIfyZbw4i/I9cob9oL8/+/IpfvnrrwAAX37xknWr+8OI/UNa9d++ucXD3YGdXqqqwpq0jJq25t3U\n2glvXydS89tX77HZLPA5EZ4vn17iiy8S/eYX33yGh12KsN6+v0LfDXj3lvzq7vaoiULivvGYphz+\nK+7Y5BQBs7C/SLuk2BVI0CApFAJFI9M4YiSZFGddKbwaXcjEPsJODhPZjU/eFkt1r+HpeOSqQdNS\nIwUobT2kOeAzeVkK7vIqMyvg/kjRPXHycgE4foQRmvefih5Oyv7pmETip+YCuvVzoGbpUlrr0XUD\nq4laa7kJ0y6K+W10E0fxUUTCCJVCMReNY1KYzacjpYLWJcIsmU3ESKa0XVehokhJK4GmqrAgrazV\naoWH+zQ/drsDyy8ngwccZRJHKTXPhx+WVwQKVi/x//K5T7MU0AEucJdwHAaMRIRWAglkCZC8TDj6\n/GIoIWcFfcP81+zlyUcYSxnIez9LUR0TxHN3/6eMTxPwQ1mkEk+sgP9YKxy5LpEnpyp670Ic1bDK\nEpU6Fpr+rSHg6LWPsZS2QgR8mF3E0pkREkfZ5keZPf6zv/41fvmrzwAAF09OECnnvr3d4s2rtIj8\n/vff4/3VLUyV6jxnZ6d4+UVytL68vGAA4P3tFr/7zR8AAL/9ze/w+cvn2NAC+MXnn+Gbb75MP7fb\n4p7qFd+9eov7ux73d6mG1cseLXEGn18+mXXoSrs3s+nz2QSEcv1FgJzBBKTykNRlddpBaaqjtQot\nifGZxkCZ3B4HqqZG1afvOHQddzCnIULTQ+GUgzd2pqagOC1ND3JRjGCrrKymIQoHTgBF93tOgg+U\nNmS+ndKlShkUYsj1PMH6/fDJ6SiTrKOMqGgRgJAgADb6zuLhYYf3tykNN7rC2Wly8jnZnLHSx3WY\n0GUWxD4pJeQ6Tzr+vDAGrvUIpY5gA1VlOG12zrKzdeLspePRJv181sBqm5bT18k6xJzKhsBac8AP\nIQtHKeFHGzNQalTjOKLr0nFMw5hQ6ACCdXCT5dS+q3XhLcLDhwx2LuUcIdNipUJZMoosumSrLilF\nmge5yy9TFzh9dmCoho+O694+/EstWBFw2QEkOFhfiplzLzplFJOGq6rk4lLOL3aYRWhJKiWwvkzk\nhzFhX/KklUeMefaQQy6JzU6c18L04q/++hs8eUIW6xq4vkrFzj/8/hX+37/9DQDgzZsPsNZhtSGX\n42cnePkyKS+8ePGUmwfbhx0+vEv4md/97jv0+wHffJ0WKe8d1uu0sP3FX3yJD9dpMfyH35zhsBt5\nx/MxwFM+H8OsxZy0AujYKYIsUhi8KEgpuRCecD8FylE3xfUGUUORDbzUhUwcAtB6oOpyzWbCNKSH\n1o4O2VNlkAGVkWjzglXpzAWGxwzaosr1F3EuLzMPvXMtLoCrEXl3mZF7c9MhqRRQgR8lao4iPai5\ncZKiRXK9mSwmm37n0I0Yeo9AiHhpBOoqO1HXsFV6f3ev0XMUkK99Xh2Lw8888hJREFK8HGveWKZx\nhMsNgRhgaJNoFw1q07CShqnqwhwJ8ch1J87R9rPxsWsOH48o/z7GQM6j3ZLe+OAAVg/WGPscVHjE\nmLXaRWGjSHFk+2WthXNZFWIm7BkcEP3RQpdHquXljaAsXuFHzvNPjcca1uN4HI/jZzM+XdOdoqBu\nnFifyFvHZOd2UaNtG6za7E4iwY4oKPZTqdaQxcE8Jusw5fzWB95djNYQ3PHTtMLnULJoWsc4M90U\nktvUOWx99vyUu0q39w/43e+/AwD8zd/8Lf7u7/4eAHC/u4MyAnWbfme9rrFc5u8OGF3R25oIurDb\ndriuHvCwI3lb60DoALStxsVFkjY5P1vhda0Qs5aZjXz5Q4gsTxNiSfXybsRgx1mUIZXg9nX0ibmW\nuQeyMqiotgKp+Br4EEo6J5LHXC6T1bVm/8LgLQLxFL2V8J1DrGinrcrxhBkZVsbSTfPTbB/k3b6Q\nmoUoNbFk4gnkkNhoxTUaBcHzJZElcveUuKeZLxpLecE5j4nmVdL412jrdA9Wi+K51zYNKgqmFm2L\nvh7494FSJhKyFDHmtdl0TpElomMIzPnsh4GJ5FWl4XIkHCnumPtJZkAoYrG2Cxk5/sPoaF6jirOU\ncH5M+T1jDIsKxikg0px1SkDKAH6O3AQ75ZTOM5Tk/2PvTXplO650sS+63WR3mtuxFSlSVFHVqKpe\nY3vgod+vMGw8eGB4bgMG7IH/yfPMI/8Ae2LAEzcPsI0qVZWoJ5KixEbkbU6XzW6i8yBWrIg891K6\nx0YNWDgBSMybJ3Pn3rFjr1jNt75Pqmq9SUE9jhS6G83uk9aan8NpGhCiZ8+KMbd8jHw9kXtX4z9a\nSIiaf6qEhMH7gqIlpKsyudWkcqlTgouPxaT4zsNbV/isKooGJXURsiSXNB6FgcVgcRwoSrIW3KEu\ncPmcwsDffIO//dsUBn762RcYSNzg/NEpYohYZpmvXiKKdPPGeYuQy+qNgcwJUqhEHpjPXQg2Pl3b\n4I0nKaR89OgMpvkCA6O4NZeMp9kyl3jKPxTyPYiSh6uT0wKxRFVBIkYFV7XaZ9R/ROQwJyAyAWBC\neleLW0tkLYDZ+hLSmDYJWuQ9J8biyqPglxBLqjxzZB2FJPEYg1eb5CgAQeXyptOgTpp0D3PIichf\nigjw3vHRlDLQKt2P6AVsFpQNgNYN1gRf2KzWWPcJctI3S3gSWFivNhgOWYRiT/cnG4jC/BFFeeAg\nBD2ktswztfrMzqKTRYC0bXODeQcoU0RbJ8ubdEDJ70YcG548f+C/fv8o4isJMb9Y0I2zAT4z8k4K\nTgrOBcfoKnHXCCdJh9FXXRYy5RHzJjmMurRoScW5vXGck1AGh5Lg+yZVgWFABOQG8buAYO5Dwvtx\nP+7HD2bcGdbArcZCMvd2VEAgb8FDwEfAZrVckdhEgWwdy87FzImJUyQxaiJ7EqUBM3sVSmhACARO\n3vqy40UwXUZNnZH/OxwmfPttSoD/3S9+hV998jmAxKH+8Z9+DAB4/OQhbq4v8c1XX6bz1TOcT1WW\nyW4hiA9LNQoNhbzKdHAQDAmw3kL7jLJX2KwSLOJkvYKWEoH62owQTK9rJ4+RJKeCr136fA3krsfC\nAVWxf0FEDQnNYXSsQiQhAn9fJDg8v5YqQBJ4tTMKS6qCToNl5ZvgJJyX35sYrQAIjKSeCRiJcOxh\nvaoMT41sHB4kSu3Sh1aSyqLGSCCEAJWVfIJgcCjkhIl+fxgsnBPos4ejOkhKKHsvGaCqdcOo7WEY\nX4LlcDFBCp6HBD0oYsIhBi5IKanQEV/8crXGkuTc2q6HlAojcbxvd3uMYwa8lgQ++Z0vJdjz73Ig\nISoPiObTeccuizaaIRTRejiqArtpgreGONUSWDYXbKQShd5ciApCke4FswRby/dZCMGh9DRbeD8j\nM+gK4SuqJxx5WNma3IVe5s45rCIhj9IFG6vXQpKgRMbngKWGUg6mYA8i57MiQgX1SRw6OSRUJVch\nZVq0GQsmYkHRxxILewd4jo4YB4DDPi3q337xDV5QePjGW+/gn//7/wwA8JOP3sVXX/4WF5e/p+96\nZJkjKQqfdte3LGgRRMRhmrHNeCbnsVaZRjomeXQA/WIJrQyAkY5d1HzHccZIAgHeF6xLhlFknikp\nJC/GUIXXChreATOFlXa2DF8wrWGFX6FEadaVkVpi0gGbRqLL56ojJqI7dl5hCkAGsesqsJMQbDS8\nC0VleMrX+IdGhRtCxGxzxTSWW5beoM+HI+soUAjogMiq3nbyrAA+TokD30gShxAts2pI2ULpLAbh\n0dC1N03id+fnSpTWHJo2+kkBKRVE5gaOEibTTS/XODtLG9Xm9ATdokKE+4AdKTBfb3cYKEyTuqgq\nZxwTkwPWsxYjb+yQ4BAtPxPWWt7sW9OgoU0xWAt7SGGqbTXcrHO0jabRDKFpmqYQAwiAjUqqyRcq\nLngOgWOkLgKkjhLnZxariAgltBWhdEHIUHKwd+BLvg8J78f9uB8/mHHnpLvPicHkNwMAhCoqOYkV\ns1QAg48IXLEDhzQ1DW3MCPb8QxUlrJSqNGBKlRKSHEoGpnYNvvBDpd9OFl4Ri+n19R7Pnibs1DzN\nOCN64o9++g5+9qfvAQDe/+AtCDliuUq7rXUzBhbWVFhQtanvA6uMKC1gvWORTwcNQd5N9DNfU6N6\n9O0KUgx8zTInmlvNO3UIHiEj2LXBrf21vBSFhhYhYtpb3FySaOswQWcJp1UPeUJz2UkEVn72ULEw\nnYaqa0DLBsQeDBsE5hBgydNpoiyI8whO3HrrMe0zqvoVqjnppF/xVgoJcziFGLnJucb9pAR+obyJ\nEUWs10euUvnZYk8ezGE4QAuJlphiNQrtiRcKbVcol5k3SutcDqS/lfQElGTRDgHCJalSBMkq0KvN\nGienudF+zUUj6x2iB4u2HoaBIxHdGMYh5mCjpNkLg2eNIheynFsOx6OPuWZzJAajtSr0S4TZy8+Y\naTSapojA5tey8uAymDZfrzGGaaV9iBAZ6SoiUmRCBbXoqpAvoDRpx1IXu4PbdPccVgUjOLqpTOEq\nUstGbrINiYAMSCFE3c5TUWuREnQBmPKxq5BQCEXHLuowzCUdyvvCgxdPpl12s8dqlUraf/0v/pJ5\ntd986xHWJ7np9AbztGUK5qvLLa6eE1J4LyFJPsrPMwzlfs4edBgmxUSCu2ECKV/B6Igxg0O9xKJZ\nIbhn6XxlwPo0ncNHP3sfb/8otfaothDpMYUtl8ci5xSkKIbbzQ7zMGHeU0g2TNBz4fvKnFIL1UM2\npeIFSH7ovQ8cIljrOCQUQqCRpe0HFXIbQJFzso7zSJm08DjHIip7VVe/QNdTckOlEVkzzxUQi0ae\n9ykkoeqntQEg6uMgZtxcJ4aN690NRAQmmpcrY7BYptzS5vwUG+Iga40+MoxAMYCIZV06V/JUpm3R\nL1Zo+wLKzUoxD588ZOWjti2MHRDA1fUFbm5St4N3np+bmttd5OcpQzpuTVmetBhCYXJ+BeOBkqKS\nuysGHyCCAqZ71mgp/dD3HRtypWSpRMeQyADo99qu5cb3bHSBlNtKEJa8uRQ5QB8Kuh0I3PCegeWv\nM+6cw+KNMFTJYAHkGQ0xwHnBuCwvK9EIAUjGmhSpe+/TZ+ry/ZH5zaVzIUEJMzofwdge7yJPXBSx\niLdSifbxw3No2uneeueQcCQA1pse600yHFoLIESmZpnmARfPUq/X9eUeD07TsTbLDm+9kXbQn/zk\nDeyGEQ8fJWPYtgGqzQUDwbmpebKYxpmltNabDX784bsAgI8+/gBvvpsM1nLdI4bMux5oXkuinffX\nilM7ONKfy6KoUkPRrY3Ww5LHEdcdNOVwPBwQQ5VE9Ywzm7yDo22x1Qq6kRAqb/G+Kp1Lbp3xLjI2\nbRqKwcpeYJRFEzEZh9rTFtyT53zJwRhjOG/oURLwgIDSmjey2U6cCwoRuLxKBuvq6gLBWWzNVT5b\nZuk4fXCOh48fAgCePH6EgsD2R/lVH4Gsvmp9RO7v08qgb3ssl8nr7hYtP+gnpyfwOfcpC9RgGEY8\nf/Yc1zfp/FzwvCazzmEeouJCR92DmyctvSj5RFGMT6Y+kkIwYWbts0XcOp4oRQatNVoqQKhG8eYR\nkRLu3GbUGJ7/CFsMPkGfMuTJB8eFHu8ts6p659hI34Vx9D6HdT/ux/34wYw7eljFLU+S9FXeKjcr\nhwDnAcuhS2SRRUjBHpbzRaLLhZBQ2Dl3c8vi1pStEoUuNufL0vlUDasQ8JlumSx/3yuckJz2yusC\nm2gUx5V2jnCzx3xI37252uOb3yXl5u/efoq3nyQvaL3s8f77b6Vzj3+O7WGHR28kgOjJRqPJDcaj\nx9VVcv+vr6+xP+zQEM3v228/xo8/SP2HP3r/bZycrujcLUJIXpnJYTYDJgt1sUBAdnpCjEDF3WVE\nCU8SVz5NPwQCcWZNPsDNFoEUfuaDYy/FiQBF59n0Gm0rYbLEvaq8D1/yNhGCgbUxFpmvUiECV4hT\nobcqx8cS2nsfuRQXo4AiAGaE4MXatQbRL7DrKFz3Dpn+GhAYpoxaH2HnCS0rYlsW65UK3FfYmEKD\nvDlJ96FI0gvmxoohQpInqIRKDgv9TUuFniAEKeyhtW09DvuUW7y8uMDTp09xOBA4FZK9FBUFXIWg\nvN0HWI+j8DWWzwOJnJIfNylYJj74UJ6VSM9Lrix6z6FuFAKS+huNNkU1nEDMeTElObRI8+9ZZ2F2\nHpMt/FohBPbyZjvjQP2qTd+hdT3fi9cddzdYPrNB1hOnyyIOgKgwMgqSu/SjEJw89M4X0YNAbIgV\nTKICFFfslACqmyyFZm1EAQGRIf4xvlQy9X5EJEOgpWSaESkB+PxwK2h0CFP69/ZywO+Jn/13n32F\nNx6lEOLt957g8eNEBrhYfYzDtEO/TLmMB+crDnW/e36Jb75KBu+br3+PFxfPOYTqFxqPHyeu7zff\nfIjTsyWd5w6HCv2c/ls4zovwRGACtmgStY6oV2SGISjN7RkiSgyUGN+OE6bJIszZYE2YCacDLdGS\nSna/0ugWmlqsACFclW8pcmCJiSO9dqxKVjaUEgTmMLAOf2Rhr/UldLEhcpIbQrK+pVIJ0rA5Tfgm\n0zWQZJSU1FjQvTh/sMY8jbzhxRgItgCcbE5wfp7C+uWq56XXtEm/LnO6xygQbMm75OG8wziNXCxJ\n2Dxilx1nLkaE4LDbp03r2bOnuL6+5HloTItcNXLeMq4RMs3bH+0JFnVqPn2367qSKvCuYheNRxKD\nzpeCiXUetmo7K3lmkeidkDaMtMnTXAYBm6Ekc8BE62iaHebZY66srwjFmA3U/N+OE9pu5mO/7rgP\nCe/H/bgfP5hx5yoh9+4BvAUEWZqVBVKCNFdZnJCQLntYki22D56reoGSilX0wK+lqDGpMjU5V/xa\nVQ2KPYyYcBX0B9oRrOVQUyuUyqMXcJm/KQrAahhSdR4PM779ffKwPv31Z1hRct7D4hEl3btFC9UU\njndvHb77NiV5//4Xv8Knn3wGAPj6q68wj3to6mtDtNAqdwNUKGQ3IRAsIIcfHBJDsosu4lxK+QKI\njYSjasu0n1kktDcSkcK+vdthooLCMCe6WuZI8pZL4G1v0K6ILnlp0LcKkuJPHwU37Ipqx9VGY7NO\nFTglNfAUR0R9qPoWaxFOiCSVFTJ6BII9y9l7CJvR6JoT1FoqxFbh5Dz1BS6c5d8xpsGjR+neWDtj\nmEauJpqm4aJH1zQVdKbs24G4oLIDKKVk5ebUfZmG90kE1boi4+aoiyHKIn472xH7XSrc7K5vELxH\nR0lto1VVda/pjl8vCV0n5nPFWElVPPOafNB7Tuc45+F94N+xNmDKYNvRQpsMgi7pBCEFrb3j5DoA\n7IcDdsT/NQwznIscSSSldvK8beA+02GYYEwK3XOx5nXGnQ0W5yFCVempWACShBAqFHtFRlYRuvnK\n9QSSUeIGanVcUSqy8yK5sxy4h4J0h+dw6Qi3Quf46S9/xWXd1DaU0fMNBOVcfJD49vdXHPd777Aj\nAr7ffvUVDHGeT37CG89TSLg+WcBj5nmRQuNbYiP99Sdf4Osvk0zYsNuiNRKKHvxxuMKzb1ML0G8/\nXzLJnvcHrgTmUI7R7SK19KSplKUJWQrIRQs55qqh5zAmzjOGm4yRUZCESYoxJp41mvPONEwouFw1\naCm8Ma1KeSsOqxRXcAVVVYHUPL1epfNdrFvg18TmcFSlKptaHR56FHEFHyJ3LKTOCmIHCLIUi5VE\nIwuuaHaScUjGKJjcJmYEtAGzohpj0LWZq76Bpuuwc2lCzkYjcz0Zo2B0hiUEDoNiTBhAS3i/3c7D\n2vQANq3mvIxzZVMwSkF3HUxmFanqY0qr75Vuv81pVQIoUX2Uqu++iMN4V4yUtb60jzkP5ytlaeEx\nSdrUzAyIkc5JonqQgFKjTuwS9GCP44hhoBa2aU7HZryl4JSLdRGSQkd1mCCQcnu5Qvw64z4kvB/3\n4378YMadPCwBMAgxVFiaIy8IAkqWzP+x2GMBugkRmX8JKvUfCpX7B8F/UzJySCfeRMsAACAASURB\nVMj/zaybIrLLqgSgmCM+cqJfiLSD/1//+7/lMDMJmuZQSEKR3JMQDcYpwvtUyVgsCjfTcNjhm6++\nBgBM8x6//W3mVWow+5HPTQmJcZ/FUi+xv74EABhhsVlKZugcd8/xu9/8u3Q+fgvTZa4nB0kVr0a3\nx3gr6bip3ADUmwh4mTi8NhtCziuNKfOYewefqT5URN7MtNSA0KzoorVgD2uxaKBNxj/F1GjNfW6o\nwn9RvGKjkQt1oQptbjNi3h75jmX+eu8jU/W4qYS9QTmmMGlihGpb9rhklcRPeLHSx9oYDU9YvITu\nps9Fm2Frib8pFzOyzlNm/1Qo8nGiNExHAsKyNzNZzHNaN8vYYUGdEMu+x5pkvqL3cPOMkZDu4zCw\nlyLFLZXsyst6NWso6D4ce2PDMDI4c55mDAdSdxonTozPNkU9OSqfEBAyTY4cuVoplTw6fqJIzzjK\nwNXTaZowEUDae3/UzBxRUjUBnkHZ3g+Y53yfXr9MeCeDJaUgKe90sZFDNc0GSyDBGHIVTslK3r46\n+QiBTGQmSMY4w/61kkz8JSTABPciJPApc+0Uo6eVQGQtvcgS20IIzADm3YE78kNwsEO6kdNsoUSm\nc+7Rtgv87KeJ+/1H7z5gcKyvuud98DiQuMT+2mE8HFjHb7VYshFdGIE3HyY09aPzVL3JxkdIQIbk\nRl88/YrL7UpLDlG5uZdR5uWGB6FYFENBQrYKhsLcRitYAjFa51KOEUBUkSl5jTaJjliV1o18DUYL\nnr/ZWUQPzt7EGLkBOMY6jI/MjSVvPUQvjxIeptxWQbdHgcK5FjwbaCcsk+UFEdFIMDtcjLEiFfSl\nqgpAy2JwUriXab0FGwv6I30nJggGJ07LIxvjrXpWxc3mfSHwizGw8vODB2fcYSFiwLDf4+nTlBcd\nxwMcASmlVEVe6zas53us/avosvb7PYOBk8FKa+xwGJkZwloPV1EyIwY4WlcOAyab+djrACy1vrGB\njoE3E2stc8CltrqiFC5EZECz82Dg6DwH7ox4FUr/+8Z9SHg/7sf9+MGMu4WEog4Jaw+rhIRA8qjy\n55QoCsUpiMzll8KKeRRugLyq/JuI5fsivoTLqquJzPSB4uFlS79sWmxIsktqycnTtEtQ4rhp0bYd\nWgIASq2hqQ8sxIibbdqtdvs9V4eGwwGHnWZBhLOzE3RdxvAU0JwQEd45WGLy9MHDMO2LZFc7xdKZ\nYVUDKNUjKcCeQIyCqXc9AoSQCQQLYKE7xEXp84o8z654p0pBRMkV19RUS5Vd70rbVPQIAuwN12GB\niLUwRJnHvC6OGUer0CbGEjZEyeynaV5Kv551jkNoFwVXwoTWEEazJxtiSe5774+KNEKKSqU7sqKP\nd6H0ngpZvBWRamyGYmcpJHt23oeSqCbgZBFfEYwF65oG63UCoJ6dnmGzXtPvBBzaBiOBJ7c315go\nPEQMiLHci3Q+3xMK5hFj9exQYts69rCsddyI7pznEM7HeNTLGyviAGEr1fXK88nNz0XppoSE1he2\nYCFSw3XpLT4+31otPnOJfZ8H+aohXreECgCr9TqeP3jIl1AdpnJPUxXjKI49Pmv+HOpX3xO/15Jh\nEPlYdY4E1ev6e+Vow/6AbrjF0ZQPeXQZuWT7feM15qp204++Kl5691XAwONgCdjGiLMHD14++Evf\nq669/v8KehBCqKpGFs4VYKFUinmQmqZ04gtRDEU+qXhUKqcHwnvKeSSwqrMjWPWHT+fl83/dtRr/\nwL/qTE56zv/A2qx/T5SZqlejdQHvvf8e8qFelUPKn2edR+eONPdqbqvq54GIAjWoqnip+Tw/zPFo\nzqWURUm6ujfELsifu7h4geVqU713nH+q64u3V7J4xavjEV9r+d8+/tFt/57vRwBf/u63MWaL/QfG\nnTys8wcP8V/9t/8d/SvweUgpmSUyC2Eyk4D1LGMdQpEDExVtR8pTlQsrXeUJO2PaTHchkKTMC56k\nfmA4JxGPrfb/8G/+Df5jkrqiPyOD4o8StjRfLlsxKTAQKV276qCIdsPHsjillKhD8HGeCm1H23Ne\nZxgGeB95nkKMXFmIoiJr04UPH0riv//FJ/gv/5v/mo7eVM9iTKKfqFpBRJlbRfAFJTtuW9nvJnz7\nXSoC/O53X+Pp8+eYiM/+5GSNt95JrUfv//htPHiQME6t0bDjiEhYuuAi7JCu/eLiEt89TRCOZxfP\ncXqekPuPHj/C//o//Y/4y3/2p6h95VA/MLlanr1mXRmPnH+EYG+hFm2NMQAyshckAIjMlR/KwZP4\nquB7oIQoxaCjbgfJoqoiAv/z//Jv8b/9n/8HAGCeLRuYpmkYAS9FSiJnErv9fscsDME5tMRksF6t\n0DWZwE/ADhaX36Xm50/+/jN8/usEbfnq6++wvUke/PXVAYfDiNUqeWZvvP0mfvyTZEA/+vhD/OjH\nqS1ssWmgSaoswOFf/av/CP/Jv/4vELIwIwK3gvWrnje1KJI/m9ujMk4qz3luWFeVsfXWATFygjyE\n0jgvquc/3auCVxOyooqCZPhPEt0A3c+I//w/+0//b7zGuM9h3Y/7cT9+MOPuvYQ2Z/YL93NqbK1c\n4Co+9j4wCK8OQVJJPCONKTWTcxLBV15kIfriMitTxwqulIlIslBIv/3HIl2mwK3fQ90tRru6LGjy\nXLZ3PmCSpb8vV7gAILaK8Rexrfim5giLUHHflznzsTSCw0XumROSSlSsfqyqsLVUpYRKybzj0Kcc\nIwMxrYs4EPXLzXbC9npi9eS2A5zPeapbVDBCIAuLhhBK/m6YcKBQ2zrHuauM0E/3oMph1Wvkdh4y\n1tec86Set1QlBMCqQBWg8dYoLfT5BGQVipS5qG+2EMXLLrz4Jd/GrL5V/TOvk1zRjBLc/Ny3DYul\ndpVwqh8tti/2+PI33wIAPv93X+KLz78BAHz33SX2BEHYjxbWegiT4BBzFIiUHzNtx3oCXd+gaXLP\nbKq+XV5cc+VRS2C1ScdoFw0M5TiDUHRv6H5FzdcbfETM1b/gCqTG2hSBVBYj+19GK+bGEhApB0pR\nkFSCITJKaPawUq6SPL47xJp3M1ghwlEST+lK+48wKfw5oEK6l7BtmmY2XkApZ3JYdRTigF6XQEJr\nRSyL9H0hOAmd6M6zwSqNnq8a4sgsvexk5txSEIChxWFWLRRha2LQ7HYHEY+YGetEqGwNn4ezI+wU\n4FRmuyj89rPzpbtdSDammhZV4R2TbAxDBD98QipACYZ/IAq22N5bOFJCHgeLkeSshv2Mw2EuuBgX\nkdH/x4IIqaDieQNymIjj63AYMRPDQ9t0rPmX//t93bsx8G2DEFQCrwwQ4+ViLAl0VeUyY6REe85l\nohQF6nwTn0M2xIUFXgZRmC4DX/oR51Y+9tHvVH+dvWclb4fInQmLfsG8YwjA9nkKFb/64mt88dlv\n8fmvfgcA+Lu/+RVevEjwGOvA+pvSKGw2SyzP0zyuTlc4fyOF26uHK7RUUBHCc7dDLsw8e3bBCuVd\nq5EFwLtRQ5ue5zxtSkXvM2TGXF+wW9Owh6Priz5xd3UNPQ9aM0urMUWyLjXBe3ZYpAA36afnleZZ\nSMZT/oFH9aVxHxLej/txP34w404eVgiBd9e2axmEmHiqirdVUyRb6zFRY+80TexJoGqGlTkxl91F\n75H9+LpJ0xgNpRVXSYxUpRH6yPZGPtarnc3ixQBl5wwg54TeT03F5M72LUxGLAsHP2d1XCQV3Yzu\nDr4g9nUDT+A4bwRmW+hiY1SMxh+dx2wLN5ih3qumDoGQigJcSInFXxAUkucQPYa6n8xWja0jhgow\n61wpgkihGJohUQofuWIVmErE43BI1z4cRi7596sFln0q5bdVgaOe8wIUZlqlqixagYhp521MgYto\npVhPwHkPOBwDP7kqWqqYiJKKW5zhP+aEYhBzhcgn9HihYgklBSFiSVvElFzPxzPaoGtJUitIdhNv\nXmzxyS9SA/zf/80v8cVnX+LrL1NI+N3vn6PvU2L9/NEZ1qQz8OiNRzh7dIr1efrbg0dneOudxwCA\nk/MOqks/qnRB2udg9fr6BplxWIlCNSOl5GdSRip0VNCDDAI9HCZcE2Pr9voKloCn0Qc02mBJCkBt\n22KxSvdbGY3c9qqUSv2mKgOkj9XKGZYjAp9z/EPh0K1xN4MVA0ZyEbXRVf2njIgUkmW64tnaYrDG\nCXPmOI+RH7BcSeBQwDteZN57zgO1DRHkk7acMqVCJ1D4wwVKWPGqoES+2oqlaxSlghiV4EbrueLA\nDlLC0ocCApQCX4tqm3JOyvCCmmTEGB1bw+ADcsunjYlQL11vKKIdLj3NWbgDEYwaBlDyTAhQotDk\nBlGahu0cuQVinA6s2ed9etgya0VjDJqqAhYzxS09mDkVYKeZNRTHcaZQEmhMx1JZmVUhGYF8T0SV\nq6gkumQ8CglTfiv9rdGa+b6PObMohOZ0VMUiEGMJD4//j6AvonpNpxMLl5qg90sLTsEoOe/QZOaG\nmNhGcuNua0zSzQSghcLNRXrQf/UPn+Nv/p9fAgA+/eQLvPjuAttdmr/16QO898EHAIC33nkLj95M\nDfXvfvgO1udLdMt07NVmiabP98YhyrTpCNMAc9ZzzCriM6toJ+64HD4qpjcXihTE81x4ME/azfUW\nly8S28j2+hI+h4TOQyuFkcLFtm0xzTkcVWgoPFSLdP9YW1LEihtMVTvVMSTjdcd9SHg/7sf9+MGM\nO1cJc0g3WwtjM+2Ghw7Fl3EucDJ2mmb2sMZpLiGhKBxSKiaeq5yEdr4gl1OoM+evJOminLhW+siF\nyh6WfJVbdTRqO/2yO5oT6B6A1Nmb04yPcoigglrqdfOlvw5hZg+rbSVGSs6PzuLgfPHEpOTQZY4e\nNvfCVch+mxPdmaI3OhTqLlFkm6KEiJVARagapr1n78g7i6Ii4qBkZLWUtjGsUKNk8bCcCwizh6P7\naa0ttLtzQLDFrc9eXQnVIntVQcQy70K8JJ6Z/xkEmN755OysoNRRxDqV2kOpCdNUJNNYXDRUSfYo\ngBhQ5eqrIQuQMqYkdJrX7Onlz1XycXDlc4jQsjCuNspAUAg/Hiw+//S3AIBf/M0n+OyzRDH0/PkN\nAIUnb7wJAPjpn36Mn/7ZnwAA3nznCU4fpiT76cMWwgguiioDLqgktebsMVquUGePxjmHkEO9EOAz\nmHX2aKlrQwkNIcs1e+swkFDJ9uoGN0TrPQ8zTH5GVZqtHCE55+DoGVVGoSMmVy0VpJacFokI7IUL\nXULquwDW63E3PqxYFs08W0id3MMQI7QmIxLTIs/Vk3GaOIycZssUy0pJ1msTUlGrSIYrVK0WLsAS\nV49UEib40jEeS5AhAF5lqa72/RNyu0bIISCqah8N3RSDJXNztRN84w7TmLr9qUF5tiMbga5zmCci\nzBsnWB94woUWheQQsXB8KYEgc2hXICB87tkYyLIoBCTJwRW4R07CODthohDOTiNEzNxMQGMkh1w1\nkwGi58ZYDWpxyVXHacaw3dOxLb8/jzO211uexzRnirm+vXOce2zbjq99mmZMfsKayu/r1RLnBEB9\n9913eS6V0iyXdX19iZubazx9mhS6Xzx7xmGqd4AhRecYBZy3hVpZKsamSK0SYwVAEvQUWuUQltdi\nBaupdkKBZCQyrMa7CEf53a8+/xqf/vLXAIDPP/0Nnj9P4FohBN5652389GfJSH3884/xNmkDrM9b\nZDUw3aZ77yJRViMiK5CH6Bm+I5SBpsb9nMMN1WYfKp2D4ANX8lLIrgr0yDoMu/Qsb693GIiMTwuB\nfpmM6KJrIGLkZ3kcB26s3t606LuU2zJGo+1bmDy3kNXzCshKKYef3T/qYJRxZw+LF9rsAJHlxXHL\nYHlOzo/ThJl2ZOddhXSXUKqUVZWSvAWqivgtxMCwCO1U6gWrOeG+Z9QP+Usfrf4UYk2jVtgf87ei\nt/mD7B0O44D9Pj20290Ozlv+rnMObUuLay04eWttQuKrPONSwNviVaHyJPiu6LTIc1tGDJIhGamU\nTx4WVEJ01w8dS7fPGEkEwY4Tc+C3RmJuZNr1kARnc5k5BocY8vYuYYyBpXvtJovdjo43z5xEHYcB\nlxdJqHZPv/fwycOyc+/2pSUlRIAWdFQBRgl0PRms0w3WpxueowMxDBid8mQAcH7+EJvNafEIhcTF\nRcq77G+GIkNGxQuGqUjJD3aMAp6NU2BUxe3mkIjj9ZDvcxJSVcjakfNhwDXBF7749Lf4PVERbW8u\nmA//7ffexp/92c/xsz//GQDgyY8eYnFKD7Yp2EUXEsbOZwiQCCUvGgocSFTwkIzVU7roeCqluDCm\njURObSkhEGIRJJknh2Gf7tOwG+CIFbRfLXGyTvfiwfkZgIjLyzTPSbsyPePDOGCgnshp7qEbDR2q\njSFjB2XJYxJih+bynq3hftyP+/FPcNxd+Tl3aFuLEksHtuQxkign5a3myXLHeAiRdwelNBpy9xtt\nIKXkvEtE1b3vLBO3FVmpV9UnAdQB4i3QYo63bw8FwSGhECJVUDK6WqKg0b2Hze7w4YDDNnkRh+0W\ns7dlF/YRkcrOdir9jd4JKGWQ94h5sjBUVQtKwI2ZG9xDEoo+6lRBy8KWMUoOS3wosAYFDSUCUFWz\nsleaqnpEXztOoOIXpIhojIKhXrOuLco4Sgr2cJ0LkCBvCil36bJCyjhTbor4qihsmSm39Pab7+CK\nREOlvmRAYvJuMpxComkbbNapd/HBg4c4PU2v97sDLqhiJaXBhnrrHj9+iMWyw8OHqdSfQp90Xa7i\nWYpCUHdAWRcsbx8lr2Upqop1zrmhjLwWbfDsoWmhkqI5hc77qx2efpNC1C+//AIXz0nhOzi8827i\nV/vn//Kv8Oc//ws8IhCo7AQookMUkfUO0jkpCOKXDz5wGOijh0KWBpOQudKX4SVNi7bJ/G4Nr51F\n10NRPlZApjxnxeowZwDs7NhbM6bFmub87Cx5WDkVcnkpMWdE/TgWgVQfIGRBvkulilclC3+YiBXD\nyz9aSAgUSggZOG8jhCiagDHC2QDLSVrHrieEYO5uYxo0FLQ3pkm5pCxK4TW8oKRvRPXQe7yEZK7O\n7ZWkEDzkLfVcSjhWbJoFhVv/N2OgKjhEiKwJKCGgYym5iyihsnseZFZxQowp2ZmbbGWji5iDjwDp\n/gkvSiOwJuPMlDLHdC1Mtx1jWoB5YXiAUlVHHN8C4DxS33eQupSZtSw0QEkkIr8GXHDMKhBqFgFV\nmlmByCFMhgScnz+AbnJZXeACyfhMo+NksBBA07RYEDr+dHOKDYUhF8+vONSLQWK/Hejcl1itVliu\nkmE79xEjYcP2uxnzlDYTBw8pZMnNVaPuoIAskBhUm1ce1V5UUcAoRCHhKCy6vrrEN1+nNptvvvka\n+zGlDB49Psef/fxjAMBf/NXHeOPdc0af++DhchgvSmoiugilS37KhpnXqxKCW32id6W4tU3n0Xcd\nF51iRecSo2CGWiFUIiPMeC2oBHOg687Yt0XXYb1K92KxWMLamcPNpNVQiixMBkgyfKYyWLkYE2Pk\njSo5FQWX97rjPiS8H/fjfvxgxh2T7gURq7W6tQuRK02o6IyMTmFedv8V7w5aGrb4Wum0u1d9XMy5\n5EpYFbxPZGMh/1btPYhjD+v2ub/Cu6KTqq4jHoUCCcRYu+mU5JaFoKzROqmg0DGttSzWqZRBCFks\nMiVBmzYljlVjMEwZrqEgMzxZSn6tCWVvQ/Ewc+JfotAbp3mp+bIje3YIAjLvzlJDElWPUBpqnBhI\nGqNnr8c5x9WynGcuclSp+gcAfd9ypRI6MDo+h5NN02K1XNO8OIxD+q153pbdOSBBXPJ8Ni3P7X6/\nx44qknby2OvkYZ1sznB6coYTElJ9cGZgh3S8w+iwP+Te1UOCxVTudukLjBWBHI7WwNEQheRQi5qw\nMP13ovtxtb/Gdy+SaO719hItyYm9/8E7+PhnCRz66M0zSBMRiJJGSEBXFaRAuQnnLaRMgr8AoE0J\n0esuARc9Jqro/Z7CUSll0TlQkufVWc/khUJV3SWgwg5HPoa9mFRNzM9hOCYwJCpkIHlRuTCkNRXQ\nKiBvDsNjFQUgljm/C4HfHWENkTEVsT1uXi5RFRkeLguX0E0JRXkcQKsGRmVktAIQmF0SsbSCBH/L\n9YwF+V6HSLG2d4ivNF5sDo+AWtUH5e0JDMVgycitDMYYbnRNC7cYMGstemqSBiQwpmONc4DUApKM\nkG4MVCZvi/HopjG3mJRH2DdnHecKWtMy4V4MnvJWGSPjMWUJ+rFAK+zsEek8PXWc5NTeNM7Y75JB\naDuN2GVJdgE4z6wMw3hghgelFfoFMQcsWuaGmilMctamzgQAi2WPJXGb74YD3JaMa5DpPArlZ2FU\ngGZBDjtbzMTU+eL5FR4+2GNJxrBte6xOkhbh6e6Aq6sEr3BzgIgWmcc9olSPAyKyVZdBol4HArce\nphySBxS2Ep0Mw35Iv/X04imeXlDeSkScPUio9SdvvoHTsxS6Khkxj3sOvaXRzJLrfYSmWFG3XUqt\n5FDKO8bcKaWQOf61EJjHZLA+/zTBKEIIvH6M0lyJ9z7wOtLEfCGqDTjzfLVNC8ckAoFl7hSFdjmH\nBQCacmJt0zH8xBiTGBuY4x1Fji7Ewtghaz68+yrh/bgf9+Of4LhzSFiAaCWhh1jxJ1U9fek7pX9L\nCMl9Zo0xXMEwRiEEh0Cg0hTeZbzGsauUesXoH6JQr0SB0qBrHYtGctLxe67ptjuagKMVUDO/lKIw\nhCoJQd6NUQJKGigK4xYCvNvEGBHSBgy7P6RzRFF+iblhWpRihrWlStj36fpZ6iqExH2FnPAGfx8x\ngSaBBGK01D84jx4zNT/bObA0FqQkDBFVLWePA2HnllMPk5PlSsIFi2HKHtbI4NV20WK9SV7T5mTF\nPFlZxmoYBmYhPT8/4x1+tzvgxiTPwI0OzgVeI+l80jmdnpzjZkOKwgeHHSXdnz27xNnZBVab5GE9\n7M6x2ZDas/MYxuzhBRz2V4ktE0ig0Rwrh4oJs65eVRQodBML3Y9UFQgVQAVwrl36rl8y19OzZ8/x\nd7/4uzRfn2i4MMGQ8tTjJ49xfp48sc3ZKRSFkQgCQkaSLAMgNdPHAJ4rk/N+wMV3CZR69SwxyfoQ\nYCqPhf3W+rpiWlf5n8ZoXrNN0zCf1mEccXmVjjtNA3lchMGzljsSmsaUNIhWtGbJO0QJCetzqKnP\n//FCQtS5Kg+fCd+0KiVKKQrlMY5DtRosakzDF6m1grNgEN4xWO+4ihBiyTMJIbm6EWLJZ03zzNLZ\n1r1CBjuUEEygAhaCSq/ZTZcFba5EYGDmHD2mjPAOAVraVOJGohrOEk8xRgyEVpZuxGQ9BIVN0ZXM\nivce05RbmSa2krYNlEKiG64k2gwKlIJDMwlB+UESnggRGe86TQEjURrPLsK0+YFryOhlNeCIafQ0\nfx5LAo4KbRCiw0QP/WRnZEa7rm+wOcliCxsOVw+mdDAsFgkQ2nSaDef19Q4vLhPcYRg9EGONPGAj\nIIVk/qW+6zEesuH12F7vsaX81nqzQku5ub5fYHOSKlvb7Q1mu8dsBzpeBWGJ1boSVfayJi5Mfyo5\nIRlZLVrTfWOV8Nkx3GPcT5iH9KCPO4svf5Oqh9vdDSY7sGbhTz76AD/5KOW33nv/A5wl24Vu0UM1\nughwzBbB53xnYNW3m8sbvHiefmd7neZiniYGagKikg6rNm2RcqC5yyGh0wkK0TU4HNI1Hg4DLIXh\nV6rAe9J/HQzrWHZc5e3aHlpWqAERb4XX5fVdDFUe9yHh/bgf9+MHM+5ILxMZ92FaxWC7VDXLlI0C\nUvjSk3ULFZZ7tFApg8QQjiqLL1UZKxKjWDX2HjfcBgaojqPlsCC/l0bePSv2SkROrCdJLgluvxCC\nOYRm7wCbvjMMA/b7FOsd7ASlFNMnnz5Yg3ljIMHEkzr104F2yuTFUTO1mzGS2MV4GEBOD2xnEXzg\nhLfWCprChhADIl1bBt4yRsYndlEAuLkecHlF4VcQWKxIikoaWBewox6yKDx7bE3foqcdU2oNV+Nn\nNHg3Pjlb4cGDFIotFi2mMacFwOfBCWDToM/Uvl1XKHXFoaQWkFINOTzuuoZD0xgFY8GkDJisw4Eo\nUcZhZnyZbjRWWWLr/Az7/TXGQwpjRA1OFuVMj2lnwO+lL0X2rGumVHiBMFpYmr9xOzEWatpZlnMb\ntoFplF+8eIbBHtgbfH65xfMXqZ3n4sUeT54kEZCTkxO0fYuJEur77Q0ceYlaaTTkwR62A778Inlv\nlxdUTbUe3uQ0SElsS1mavQWqUiNS1VDws1y8TetnTGMRtEjVxOLhZxGQruvQ92mhd41OaQTKTwQE\nyKqIdNSCRuMPFPdfGnczWN7jcMgy7i20yiFdwyV2AQEtFGxWARYSAvmGRyZ8c/MMSw9pUArOWmYE\n8NaXheUrQGSIECEwqte7wi9lrcdIRmoa55K3ySq2AH9WCsGvYxXaSwgglBBRREBm0jMPzJHK8sOA\nA/XLHcYhIao1haPuCXQGxEmN3JLnZUj/IzR4DAImPzDBI5DM+TweoLOaiZ3hvcfNdVrUbdeUvIHK\neGcQ457gh26eHfb7NLfX1ztcXhKQMgrMFMb3LsI6h2tSavFxhiWD1S8XWG3SQy+VhJ8to8IhBRQh\n4rveoF9Qpag13HCd7/HhcOCcR4Rno+y9LymDEOC94/xWTdi4WPQcUhpjOE/qfMAwTHy/rXXcYKu0\nRkdGfbVaoG0blh+TsoThdcohAuWmRwCowpUjPqfqc0HAjjNm6sEbtwc2WPurAwYKD00TMNHDe7Ub\n4YLHjs778uYa25s0P5eXN3jrSWJxOD87R9No7GhT3F1dwbk0d13bcKOxsx7Pnqb+zWdPr+gcUy00\n3SrJxkJrXarjMnHq59ySdZ7zj8nQ5I0BRexTpLRPhiUpVcEiZIFPKJUUKrp4TwAAIABJREFU3PMz\nKyCODRZ/p1LN/sci8Kubn4WQbGGNNnzCASFJohPjoFaW85wigr0qax0vOClTGTQzPMzjBDvlBewh\nVe1hFbl35zy3ZEyT5eON04gpy3LXOSzmkK6s/BEeQgAhlHxGBDcrby+uMFIyfLITlpRbETFgu99B\n0VxMhwN7IGePTxGp9eXF9TVm79ConOQewZ5c9Jgpqd1KiQdUBjdKwTmH588vAADL5SLBDACo9ZIT\npVJLeBv5Wg/DhC3l8G62ey5NuwA4agg+jBN8CDhkIwLLCPv9fmTPSyuJ6KciOqpE7ltGkB6eb27J\nl+SFfHOzxdOn6b3V2DNs4nA4wBL+y3mHaZ4xzPl+WawUCSeYBdbr5On1fcfrzXuZ7ndu/5ot51aU\nBHu7ptVQRkHQAyOU4tYk+Lpr4o/lUkrRiA8egeAd7JxboCzmKaPt94gy3Zv1ZoGevMlus4FqgZk2\np+urZxhpk3763XM4gp/sbga0rYFzZACvbzCMyXi1bYM+03sGgW02mGTwqzRRMiSZndeYkrjPTDyh\nMNHuDmmNDPPIFDtt28KsiIVBawgILmBYOxdNRjtz65bzgdrE6By+h0aofv9eqv5+3I/78U9y3FGq\nXnDuwRjD/UKN0cXDkhHwgqXO7ey46lXDFax1EKKoMdt5xkRexn5/4NeQAUaU04woLqT3vuJVGnEg\nfp7DfsDhQOGCy/2OEWyfb6NKY8m9hBgrnviAifJHAZ6rKq0QnEeQooXwDpoqJmmnIYCpbtG26bea\npjnK14gQmR5FiICWS+yKObkbreC9Z7S3FMBMvPKI8YjczocAS6VB6yx8Ta5G4M4IhZ5CLKkTR7qm\nEruHw2JFjbKLJUNOpNIIcJAELFRas4cFIdjzci5w7il7vU+fv8Awp3uy2Sz477vdllMLzjoI4TCT\nd+xdLJXkpmGw6XK1QEtzPI4Wdh4xkXc4Wcv5LWUkq7QYQ83ilcfB6YmakYmI/tLr4yohIDnfo5Rm\nDrdUPlSFmMgXKfi2a3F2TqK0P/0IGwKRqr5Bv24wDCkM/OXf/wI3L5L3HGeLLTXUhxDRd6WJOQCw\nvG4cQN5p4pqiE9BUvdSaK/ZCBBaLNU1pAhdI4OuRRHS3+xtsdyntcBgOnNPdrJZY03pb9D3ggd02\nh7AX3Ec5Hg44kIfWHxbo+r7kxHQlXluRIcYQqkb0189i3clgSSk5cVqr4BpdZH6iRGJezJznLhRy\nsZBOO72OFYuDxzxNvADHYSyd4EaiaSuXvMZIhXhkADMKd55nLg4cwxrqhGtpaq6dVikEQpUGzF3y\nSgoYgmFIoRhDJSxgp6Fcv/NM6maMgc6ocGXSQsjEeFLB0CITWjIntghVU6pOLUqO8nHeFgPtQ93Y\nmlDM2WBFCSxW6T49enKONZX5AyQU3bOkC1eKFlARi2Vy/8/O1lgs831WcFZBk4FWVnEL0Ha3g6HO\nhbAR8FlOjMLx3TBC5AepbRDmzObgjjibYgj8IAyHgbF1QghmeV2fLHH+KCX4n377HDc3l7i4SOf7\nZPcI9ixhsuTsqXiSQuhHjwon17Qb+UHyEchNKGkjzQrTqZUsh5hCyJIXkJr59hEDrI8Yac15GXFy\nnkL55WqD9z/8CADwF3/913jj3XfSXG4aqE5xSuOnf/JjfPr3vwIAfP6rX+PFi0u+7vXpBmsy1lpL\nhmZ47xhKYWeHOSR4iN2l9xJJIarB1poxgAICPhSSze1+i2tSrR7HEQsSmlisV8SDlebSzY7zk1dX\nAhM5BXu9w57SDqvNCUzTcAO1kKW1KYSqsIbIe8RdDNZ9SHg/7sf9+MGMu3lYSmFFVr/rOg4JtW6Y\nDwtRQMrAPWCAKP1MrjRPJuqL0liZ+uRypcKX0EmoUkpVmvoWXw4xj/sKI8BVilePzNwpXvpU4ZUC\nCs9XazQjkY0qitB27wElS9lfycLyKCQzf8IHwAcIop5p24bnD9ZDkycyTZbnIVWnBLLgpRTmyHMt\nkI0IFywi4SGaTuLh4+SNPHryAIHCnNl6ZoY8jAOmaeJQqmkNluT+r9c9FrmXUAvIqNHkooKUOJAY\n63b7jCuQ770tkEnIB6IrfvDwDZwSsPT0bMPwAtO2aJu0i182N9gfRszZK9vuGUTb9C1z+hgjsV4n\nr+/6SuPmZsaWuLZ22x3s/CDNkSyUN03b4NHjRzhQAeLp+B1cDqciuHolQoCoVYjrDT+C2VEhNRds\npBLQrYGkkNq0Bt2aqpPLDd77SfKqfvzxu3j4RroXQQJBlcLTP/sXf4EHm+SVeetg6f31coUf/+RD\nfPjhjwEAy2XL1eVxGrAlb+jZ0+f47LPP0xxQ6K0gq+pb5NRMUc0GQhDwLlWTAWAcJoxZwdvOkCI/\n4z3W61M6hwWmw4gblZL/3kcuGOhp5tfOOYIyEL2MrJ/R4k2FSlHpLn7TnQyWUpJbMbpKl1BpXQi7\nhILRkeWPpJT8Nzd7OJtzHo5VoL07+hnqBKdQQhUJoUbrJIvNZdICT1CySAtprdFk6t/X0JXNxZPw\nCusWOX8hGT8TjWKDlUUmuBWhbmSOEWBYRYAMkQ2l0hJtpj7WmudynmdYl9V2Q2qjIALD4ALTKseA\n0i0v0y/koSr2CKUbfsjGcYLeEkxDeSjpMducY1Nom5z7AQxzvSuIoKBy+FqhmOdphqMWoG+/e4bc\n95ON4PmDh1hRaHmyWWNDFb9p3OCaqHYhFUK8hKVr3B8OuLhKfzOLhueyaQ1awvo0rULTKG62HseB\nUwD9ooeQueHXQQhRtY1UkuzOg34SCkXQgze6iiOt5IQkN7QIIWC6ljnPu3WP5Y6qm12DxYaM11qD\neoQxugnBBWYpaYzGQ2rNeXj+EPs302ZyenqKD37yAT760xRWto0CfQWzm7Gn8PmL3/wOL25SGPni\nJuXChCrnHkKp6nsfWZ9AQGL2ExwRGYQKZiKVKs+RNCzd1jQ9/Aye28SpRQ3/PmLOYhfeQ0BwflWK\nQpRY+REJCyZu5wv/+LgPCe/H/bgfP5hx56R7plVpGlNCn8rDUjIJNCpV2Eiz6+20Y3zVPBdErXTJ\nO8p0KSE2iDFzLjXoshRV26IxJZRSUqWqJJ1P11IPXxDM1lhwYzVQMNyC177awscYGUGNxqClgkMw\niqk/gNQIOlNyfeXXHF54W5g1dRSQEPCZSnaeoLuUKDZaMiVNsK7qq0sJYHbxXeB+NWePcURKSaZy\nkbJgopRuuCgqZYCzVM10Ci4UIJ/WgqugWgkWDDVGwckW1qfzm6YGA993xR7zxfUVq3/nuR3HEYou\nZtm3WFIhwBjN1a3JBVjnsCf6ZOcKONlaC0P3dLlaYiIvqlt0kEYx2HG722JPYd9y0yV+MiRArlYG\nbZt/11Q4QItMGyu0ZtqTwptWgI95X/feJ5VvAFGmQkS7IIWkvuP1K2SE99nzdsWrhoOSgMnHdoAi\nXNx0mFgyTRuDdtFjuc4VXHDKQRiFluZxuVmw15mb7xPPVQ7/BN+feXbc7wchMcAh0m9rYbDs09+c\n9zDkEsaK/ypSw0kG7xrdMsW3Eoppp+Y5g3hzaVCgPGyxoO2rhuP4+g7WXQ2W4NKyaTS7fbV8vJZJ\nwcWzuwdu4ZlloeSNoiL28hoxNFz+1Eaxgek6wxWrvm/QdrcMFpP5laqDVh6uDfx+Og+B+EeqETHG\n1Iohy3c4XyZLbioazRTEEmkh56qmUorNnx0nWHoQ4TyU91wlSeEhGVulsabQQkPAUazSUljNBixE\n5hma55mrPIlMMR5dK5f2dQljnFPQJlcwFRqn+Ry00XwPjS6bR9c1EGiYZjgGzw+jFMBAlaJpGLl5\nPVdJr15cYJ+l0YKDMW8AAFbrBU7PcqO7gPUeluiTh2nCxXXKTZ2P5zglBZ3FasHA4MWiR2M058p2\n+z22O/qO3TAtM4SEUoblwZQqc+F9YBWZl0KS6p8JBV8euAoqDyEK5bTUAjORNR62I168SCwKL54/\nxWpJoWyn080kQxH2Frtn6byvnl7icJMMNZ6kFAer6JQzSfeWQrt+WXioGISsyik671mxys6ec1aS\ncqs5N9y1LRZk1KfJstK4tZ7BvuPgcNjtcU1cY+Mwl84RqaGk4deIkpXfpZBHGpQl1K5Cwj8K3C3j\nzjishhe8LFZdVR5MjoUp2BRSltaISuBBiIgsDilkhNYpLk7HLjmFrmuwyCRxnUHTlnxPki9Px2tj\nAJDpagLrhTIDzmuUToWMKY+VH3wJzAS1ENJj4ZNRURBMM6Ip/9RQWbzRpggGTDNm8hbiZCFshKFe\nnVYoCLqpRggsyZPQUZSks0yw5ZxTUEpxiTjG5JnQ1dG9L4WGyKyqkdvGtBIwdP+apkEIVe5Cyoo6\npVyfUWkzykY8Rs/iBUYaGE3YN3HgB6DRDb4FcHHxjHN78zwV5oa2Y0bOru/RdT2ETInky6srTNRv\n+eStRzihFqGmbbngs16vsFz1DJ8YDgfsdulBmq1FL1q6ZxoxuOMeulg8hiKIIgsOi0cpxzCxZECR\nXIseMQYW8VCN4oTysxdPOdmyWvVo6J49eHwGrSXklP79m0++wt/97acAgE//4dcYaA09eesJIiSy\n5rBXgR0WGZOsGQAs+gXOTlNSvCfcXAieN65pmrjIMgwD+j4VOpqmgZIKHX2n73oMDbWGzY4xkJeX\nl4xsl1JjGmcctkSOGDzDfJbLBZa04fbdIkF4uDklllY33DZY2VDd57Dux/24H/8Ex92R7k2uDMqC\nYBVFC8OHABFEFap57lkKYebKjg8zPFXDQrCI8Bx2pHxM+p3FsmNAY9OmMJRbj2KlVgzFoMDgUcq5\nrygShhB4l5KyxNKSGkJz2OUQEekie6U5BGt1UQUxSmOzXMHF5D082JzBk4e0HyZM1Fx8uNwhjjM6\nKucrH9Fk4KIUiJRPCdqj1elYSibUcks7WeouyBCHInXvg09HqvnnY/HEChATHCq2jQGiZ89YQhRZ\nclkVnOk7MRLfVNvDr/KuKxit32vN9Ndt0+JXSPc+i6pqrbDfE2uEtdxv2fUt+uUCXUcIfrnFSKHL\n7mYP+4j44fuG0xF5R7+6pN3elUZob20p9wqFGF0Bx8YqxBex2tdFjtIKzrIKY3xmHggWWmZV6Qil\nJUNBVqsV35vdbsB4SPL0fg7YXSXv5dGjB2iNRkinik8/+QK//IfPAABfff0tVg8SxKFt21TtZO+3\nKFPXFMjL1QoPHz6i1yd04hF50fvgOM83zRO/1iY9Rz03iS+Z23+2MzesX11NGKl7JNEhCwQ6RtOW\n+3F2dorT08zasYRpDLNaxJz8emnUua3XH3c2WPmm1AYLKMk5RA+EoirrvC8GwM3c0OncDJeNV3SA\nCBw6GqPQEgle3zdYLAozaWqnyG59nbcqvOohgAn9XzUnQhSSvpqtIZJcFod0ojAPeBugKCkqgmD+\n7WWzwOnqHJ5Y1YxqOOkeJodAYYuaPYyTMNRArWbP+S2tNf+mgeQcAryDlBJLQh73bcfSaGnRZuwQ\n6fxVrJmMJAuRWFQBxABNoXZjFBANs30IgNlgjVKc2/LOwSFjwpIByx0OfdtCkzFcNB2/3+S/L3sc\nMu5rGLkJexhH9DlxvFxgtR6xopCibTrcEL7q6XfP8PjhQzqmYsO72ayxWa9xs97ysfcUEu5udlhT\nGGmMTAWRqWjm6YpAMpfbZ2s5VD7KU4FysDRnMaK0PBFdS15zi8USp6dndD8NE+r94vmv8ZvPvgUA\nnJ2eojUt7JTWwLffPGUNxYfvvI2f/+WfAwD+7K/+Aufnp4WoEgV6Iyooj5Qa52Sw3vvxhzSfPduH\npjGFRUWAZfSACKUVOioirf26tMkEzwZ65/aY7cjf0drAELSnbVtuTD87P8MJNez3yw7ayKoFp6zL\nGnl1HAXeI93vx/24H/8Ex908LCk4VFNScPUvIrL1jiEiugRyBBLdRE4OW2/hfOHdyYRxIfpEmZ4b\nNY1ikGbbNUw5LKWA845DIaDk7WRVAUoeFujcXh5ZjQYgxH1xsQj4SUlu4bkX0U4zS2RF5zlh3jU9\nHp5IRO6JlZyoP2wPmDNB3uShfITOIqs+svcVmsgVE42ygzrnIKVAT7/btQaNzgRqFeKfwuQqIiyK\nJKLIn9XAPSlSOMegSBSpLSUFctUihVgVXUgIMNnf14Ypg+HjUXEAAM7OzjEcCs3PNGX5rZk3WNNo\nrFZL7l+TQmCi6t/liytckvLzql+g7YoU1WLRMdXwOA4YM9ndYQtvE+q9M11iD+Km7LrTopTo0zwW\njzyiaAEIIQoIulYFFxqIBTbTdUs8epSqoG+//SN8HZNX9fWXT3F1lSqGT5/uIIRkwKa1Hk/eTOrV\nH/35z/Dzf++vAADvffAu2l6xO+G9Q6Q+QK0U38+AyJz5P/2TJNb65htvsLdUw5D6vuPwP0YPpXSB\nJajyXBuj0NB62y33DMuJMULWifq+x5o49Tcna1ZPMk3y/Fma71YxI1br7f/LuHsOi9x9KcFhw1Hn\ntfdJw4zl5X1RDfZFaCJEx3E5RCQcFlWZ2oZDwrZtmXUSiPDBVc3JsapKFiGMEHCMrv2ea8mH4Api\nEAnBW1XbMuVbsJ7l5G1bFHBFjOhUA0Hnfr29KqwR11sMV8S0MCXWiZboSHUsVcIwOWZiFEJA5Pab\n2aYGYJ3zbXTCdGGCSRIFlJBlEUSUY/jA1SvvHEJuBg8RIopSzUGBgHjrecOR0h3LolUN0zFGcLQZ\nIx870nW1i55L8OO8x5byWTc3W5yeUdjWaWgl0ZHx6ZoOhogh/RRxSWX/k+UaZ2cpBFk0LZ48fICb\nK2rNub5BpJLasN8xi0CMLZy13MLknDtq+ZKvQFrXSHaa5gqdXTT/cgZGqVwpW+HsQTKUH3/8Myji\nw5Kix56I/ba7CYfDBE2VxR99+Dbe/8l7AIA/+fnHePKjxPDQLFWqCuabE0quMf04VZeVhlpSHolC\n57ffebOwfYbA16hUU+EmJYxRVQN6wfBpXfLH+9WKDY+dE4K9JxhF3/aMQUvwCoIaKYUY/BEnfu02\ncLKilk+7QyrrPiS8H/fjfvxgxp1Vc8qokt+oaWQDQqgojutGZAFO1CslCoocKZGbsTm16GbbtYyU\nD8GjwC0BVAnz5K7nZlYwOO9V1rsGkR7hs0hCrLSRSQZBWh/hZsIUjVORjpICQitImz63vbjBMCav\n6nCzY6+s9akqk5uIdS0S6j33VQohwOTHNofZ2UO1sMRKKXWEyQ2tISXFma4jhCOKWn/kOWUpNQIW\nzlmZ+hgUeTxfx0KvtbRVvs/BB/as8pyuN2s0mc73asegw+uTG4xjCmW6oYEvMCMYrdnDGg4Tnn6X\nxEk3yx4dgVA36yVWiwVOiLv9atkVeuzhAEs4IokNovWFvda5o3VZq7nkK/5DoUoMkYVoo6Cwkemc\nO4hHycM62Zyh7dL6ffTkBfbbdD4XF1sc9hN3TPz0Tz7E+x++CwD40ftv4eyMqsMmPyccOOPYSwGf\ndw7ps3rRyenmWHg4N9+jPB9aqxQh5aq8LN0oi9jz7zRNWz3XERKKWV+bpoGm+6GNYuruHPXkdFHq\naa2jguwjlcJQfAkD9/3jjgarZPxBqPD0shaUSHmgWBupfJNVhKKHTEeNWuyhaRq+kYvVEksCGWad\nM/4dlIcxOfCMUC3d9zVSOZ9u9Z6SkoGlEaEARQMt3qKnwZUv70YO4fzoMFFt2sUArSUczcV+u2fl\n42gDGgrbemOwMC0aehiV0RCZSypMXJlJKrjpWFpIxBi4soroWSU3CMeS50orYr/IlLUl9KlprYMv\nG4ki8rlc6q6/H5xHzbVVb04Aqoc+cH4ousLXn797+uAEi98XwsCZlKmnaeK8pvMeUmq0VP3sm5ab\n3Q/bLS4v0zGvr09xfpZQ7yerFfq2xZLCyEXX8QYyHYaUdwOw6FrciAhLla7g/dG81LqEOc3wqjYR\nbo3yoYpJRDoePWxGa8glPfSPVjihMv9uW8gkb64PmGfL3GJvvf0WHj5Ohnu1VryZZ9buso4F30Og\n8KWh0qZctDkfJbkyGKMuuaRQ1NCFTFTmOTdZB29KKebE11rxSQhIKFnCyGSQ8po4phWPCCVVUwm+\npE4SmT+GvMO+Dqg7j/uQ8H7cj/vxgxn/P0LC1x+vaB9CEtAoFBRSVrQWWjPTpJSiVCBv1/wi2OQK\nIdj7AGIFnMzfqWxzlKgRpaz6C8IcFuVXGNpRRi+KgogrohGztVBKJSwZgOgCNIWjrem4obhXBo02\njAOS0MxzpYKFjkXAQYvcBJ5aSUJNYUO7Z4iOaWi01AgxFOUT66qdNRy9ztemtT4KjWs2SG8dJ26L\ngi/HmyUMrL4DJwpmzeeKlkCm/NIG3NLjg2V81totsVr23NqxWvXYkBLPcC0xEpXK9fMLXC7Jo2pS\nf+mamnkfnJ1wUtp7j2FIIfn19VVSqmbgsiv9qyIc+emctqD3Xk15IpjeTFAahL1/ESrxB+DsPHmD\n5+en3BtqbYCPkYsgfd+i73PFtfqZWLwsmvJKhqz6HREqLxF0/a56T6bePpqXrCAVgofwx56NYoZQ\nwf2mid6YXiMVKXKkFyrGWx8iF7lyYag03Bd8pK8a+fN1pmt6fQ9L3OXDTdPEBYVtQhRJ+iMV1yqU\nvT1ilSY5/o44/vftkzxqPxLVT1RhjC/Kz0cPEgRM28BOUwGz+VAqX1VJWxCkoJx6KLYLJcz1ocTd\n+dv8nSiOrpENav5zVVZnAyyAmha4GFMPP3ksVaXeWyVbeCZuzVusPviH7u4f4iE6Vt6uMeGvPma9\nmQgIqM0awm9vndUrFivyw0hVzVh2NSE1X1sMgp9KIdSxsa3XcDUv6daWdRVDrNZFvpp89mUNuHmH\nMwo/cfTpupp4h9LWre8ff5OuwXv4vDEFf1SeFCibsRCiurfFqoUYEOUCFxc7Xg4vP9plI44o91VX\nnRu1TkCMhV0hxEgkm7Qxh3C0/upcstKy5L5eCsPTJ4+VciLmeY6xNMB+77iTh7XoO/yH/8G/xP/L\n3rv0ypJk6ULfWmbm7hF777PPIzO7qrtuP253DxgBV4I/gMR/YMQQJMQAMeIfIOYwZsyA+ycYIIQu\n0r1CIAFNd3VXZVVmnueOHeHu9lgMzNYyi1MnK/O06JJS2laqPHHixMPD3N1sPb4HULEwimj2LhhW\nxX5wa6vnUqwwkKTUCxKA42C5PE8BYZpMmdQ5N97hZvM1BYcpOEPEx7xjbXSP88OjGY6u5wsem3ED\ngfHNwwf80d0NYvN2uzw+mDEDnAOKFsKr0kTUYILOyKlNsOyI7WI/7alatgMIBZhKl/+Q4pHUdGNx\n4KZH7xmgDOSGcr7sCcmMIgjciNtzWJBFb/QT/uH//g7/4asv7Bzsqdf91BCB2UGon+u6i7cLZlBi\nLaUXml1jBlh98SPF1lElkpkHxDSpLD2kmWsCaGyG+p3ee/zyqy/w7/zpG9hNRxmEVuAvvUGQM5Bz\nQCpNwDBPgG+Qh8MLwNci8BoZFOrz8+EWHILdPHvcDZLBzttN4bwD+WA1sXiJV9eFSL+zS7vgvAv4\n3/7nf4n/8j//j/HxuG48fLTsDH8X+miR1/taGMB1w0cFAE4f3uPtd78FADw+vEXOm9WYvAtYWvMi\nBN8jxhStvrnHM/7Xv5lwenxneLpR1VcEnfHgPQTdAuzlF/f4xZ9UDNmz+xtD1ceUKt4MVfzxN7/5\nDt/+tvogrpcNAXrPB6PIHe4Yz18ekZtyxeP793hsBhclxy5kcJz7RoKCv/2bv/lXvzPhnxif7Uu4\nN85RlZfpQMFRYXDk48WUbNKq9HH9LMcerItcDEgpDYteL/aJZJBqEC0egglTmyiCWJekvqAXJhWE\nqgV14dRBa+xB0orfxIjNQ3FLhKmw4ZtEggH29sxYd8U2EdR9JXGGUIBX1xYfEC2dSvC5pTc7w0lB\niapTdTMYQGQwmlrDNME15953D+/qztYmLYSApV1A5J0t/gAsha7H2i+0nDOSqU8MCxEz2Aebvxyj\nSYJUhY0mpVMEqSSL+Zx33f1YuvpovUcUB9c+RwTC/bGdJhlTAwJxAZM2GsgkiEjKVSTrLHQglJQR\nTe55tZQkzF26m8jBh8Hw07NxXERSN0cQ9KCWvn+T/53O4vdEoXXtG1MNHv9heA8s6PEDWHpfq0aM\nuZoXQW4LeoAHN1qYC9yvc2QAgldf3Nn8Xx4vWNcuTtNxew4FhPvWWfzZz7/As4Zx2/eId69PdrQv\nX6mxxhE///lXdou9/u4t9gbwLbIbRSuVShOz4vzoNCT5qgxkTbtPSf1+z/jsGpZCEVJMWNHIwKF3\nT5hdA5LqZCcDke57tCiF4Azj4LyD830BFGILub1ns8GG1N1TdxDiHmbzkKJegUEt63Jw7SAlcxXY\nBoBM5oBTCNi52E4kQuZys+8Ja1KBwK55RVRQigdcW6SYIaQaRLsB8gQBUnaLxFx7L1AttnQljynZ\nglUdejvvsUBM/8h7b3CFmHMN8w1cyFbrYs/wQ9qs7xEACcV4iwXSdYvY2XHmklBKhtM6H/ebU51m\nPh4dlFugOsQZ2TaUnLOdF9ei86K7dQnIKpgnBdkkpjsqX3LEvkbTNn84PZoN1v3zlzZHzlc/ADLw\nJBs4OeVoXbkSkzEwVI5Gb+7vH4IxWqqPB1Bknw37eymib7Pj6cBnMoR5mAJigt1DKW61q4c656qH\nX4UKVWrJA4j48o/uTZ895R1bg62w9IWu8imKyf08f3FvC8yb12/w619W2eVlnjA32aP5iwl3z5/h\ni0ZMfzxfzMm7lK7hXtNUNmlxQedfylC/q8u4Mit+PKzhqUv4NJ7G0/jJjM9XHG27V0oj2BGD1G5d\nSWPLszM6dogcQwkkRM4K0EaZaDWxArKOlwvOFAbmWVO1MXpqEZbzxptyDSMDANuluaQUgNpzDIZa\njhcScJsGRx6gbMoLGcDeTBY2EIrT7yWwaG3LIWVGtGNPtqPEuEKPQL/tAAAgAElEQVRya1LkWtg1\n1dKS4Nr3CJMVfVOMlh5hbxGU0ZyiRS8ZYi4rMUZkEghrxMPwVHfGIr087Kdu4CGtwaARzJq7guQ8\nOZs/oYp8MxoWXXfSTN3nyp68/TkUbYXIUtga8fQeXX1yeK/hnnKNzFCjDa1XZtmw7SseHms9atui\n7ehV8qapRkyTGcDWrxLrejnqBeC8JVMl4CYeiB+o/9YGjlw9M47eCBArS5RUUCAImm4zWUrnOZg8\nD5NrEkkqM54BqOKEQ0yaKThTf02Dhbx2eMfOuWRANMMoBBe65HcIE3KLnPZzxHru3dwPbc7vXt7g\nuMy4eVaju/vnR1yaC9J6Wa2hVUpBytnMgAVi2cN13U+s8fXjE8J/FJewy3CUcd0YXiPS25wggmvc\npCX4Tix2wcBwKRV47wa51yqVC1TCtN4AUwhwzvc6QullKyLGoRUmD9PcpYTbggUqYK7PBUdWuM+c\nu7SLVOLm3iY/ScamXftCEKmf5diBDShHKEM6KiKDG5AgNnS8pBXkBKw8uZTMlmtL3d3aoaZq9Usr\nuFN8x4KsuujKjr1dmHuKKFQsy2V2mLXr6IbW9BzAqjsOQskCWVuxOnvY4h86WLcIQNRrhTzoSAl+\nfyhPIHsNSed6XouUVfDj2IErWdPRbOawqVSHZwCIhbBuEY9DSqLa7wDDsRr8TtWCLQ3cunaTpNx1\nrva4DenNRxZO7ZfoL77+bfLRqzQNErswU3M+AurmmUux63yhybrDzgVMkyqCLvAuIFL/XikKtF3N\nyDaKs7mq3U+u90Rqt3We4Fu6VQiI2nwBcPvsBnethsVM2AfZ6zDV740pYt0UqiLIZcXUnMyfPz/i\nw9tm/XZejbuaU0YpGaHd89PMg1M4d5uvIghBd6lPTPn3jKeU8Gk8jafxkxmfGWHBOnbselThfYAL\nnXLCIgja3SF0uMIY9nJXyy9FME2TmQUAgqV1Iy/ramHzEibMPhifCzlbWOmp0xQKkX22UZpoArdi\nto8C0vSOSifFS0Qkqe8HwIXhmv42b4TMLQVmts5bKYKYCKySHr7z69h5KM/HO4ZwN2aVVigHgECw\nEJoloUMBasF2ObZ2/hJ6WM0E1yKOsl9qzNJqquw95tY+ng9L94z0vkcBpba8p9TazLEgKeduT0Nj\nq7L02aKHTu+BZIzbY3f7URAiDNLyux01TQdqq78XqrscTooFm4Jhc4EGy1mqlj9bFOksqmKQfWdu\n0U3O/VxptLWvF+QWtazrithgKmpi+8kxFMzr3+kqvbCmbWe9tHmuUcr5fEbcN+QGeBXJ5vQ0T8GK\n6cfDLfbtAimqFpog0JSvAGjRkDiQFd3VD2HGdtY0sRhXlNhBQx1iql1+VV4BBtWO3pkVwNxwagQn\nBtD1gU3T3Xk3nj7kkiEa7hu1q2LFTOkj9/j0ny4lBBlyth641o86rMG56+4KEfrFRL3FOaZ2QDM/\nDV1+1iAyA3TBhwnM3rp6tR3aix+dcNnbz5oOZTCKwhVCz6sdsy3CmWKFECbVrAKk1QZyKYgtlZkC\nweVGXN4jCm/Y0GogcbJaAbmEWReLvKDMBX7Rdm9AWx8wud61FMBSRSoVOBma1E44HHBskrxCBY+X\nWl+QCyFBAJXnWSYst+3iv7tBMJ6Z72qmUtcbu1G3iPWksjgnOy+lqJ2YriLZwnoaCG8jyJA+QWat\ntSk932MNq55DGlr/MmxkalO17QUptnkN1UhVYTFFOqyDiKz2s21b1T5T8GMWM1x9fDwhqUFGjFfO\n3XV6PlGfovHm4pYUKnJ++EXcS3op7mbqsG9n5Jywr/VTHAG+bb5T6AtIWG5wuN2tC5vzat3dUjJK\nc4EWZNvMK1yArjBzQHd+5gGcy44RgkcYZIv01mFmI0WXJF3LLtZKJpSNMgccmoZZWM69m9uAulZS\nQLHTziCrBmTJIIUnfQZ4/fNgDUTmWlt5jG0nda5jXdoBq7dZdXH+3a8Z3Z2JqjOMFUFzx+kQUb8Y\nvQN513b2ejzWYh83O+rQB8N9uA5aJe/ghgKwtWClgJLrkAdOoFmljz3YNQlcbJiPDXdFhC17ZG0Z\nJ4esNk4C07LyjrGnjFlFHubZGhOlBJAyrmmv/0fdCZkZy20VSpuWCcfn9XHMEeeiHoeuRm/HxqS/\nPeCmiavNt0ewXpjU5ZeZuDYP9nZu/GYo8LgloFzauSBQpq7EMCDWayCrN7N8MprSIR89P4rn1cd9\nweq1TaBkfd51p/GpWr0paD3m3G8k191ruEirgekNmKw4XUqxulXKuWK0ACzzoR3vp+tz6r3IULiJ\nwkLygMSvFm8AcHp4j3fvqivz5BwOhwOknbf9IrhphW9AeuNgcjg8u+uwi5KwnptFPDbkFmGVUlBa\nMb4GLnMVw2yu51NwpjZShI2o7gKuHNSJxKg93rPd1yIZe1uIYq4LlqoyTPNsoosheCOfF6kROJHq\na3ksrZ61pWS1PSlDJPc98JhPz//TeBpP42n8RMZn1rAY0zSkbfr8QPHgJt1ruxzTEKL2qMo5fxVh\nedcBfqDulzdyFu3vKlHBPV3Ueo8O+ZizVArQYASOGQNCwSIvkSo1rF25Ih3xTclhck1FMpZu+uqq\nK0uAur4AqaWY616s1gUfqwRL0/SXWbHyABUHyj08V8wiuYr+f/WzKqPrl8lArduerZsYJdXUUT0j\njwfMzYA0LJPtrHHdrGsqWRB86GnIYTaYybauln4h1ra8DIqXCq5mYiM0EzrY13FPzyzuHeRrmHlQ\n+6zpoX5+KcUUalPqbj3MAd430Oy8gJlRlNcmYmmoGyRXmAk5iancSs4GGSHiK90ohULc3Nzi4/Ep\nSk7VgJPr5w38G3Fq/n3n01kvu1bTS13yCM6iC0aN/AEglFBrxCoXdHwO1yKWlDakdhFt+8UMNmoX\ncQYgHQY0+e5jOaLJpXpVjl6TCmL2fvS+TIYEiKlmKNpRZ3KYZjVEgV2XKSXkNIEa3W123iA/kdjg\nMnHfkfbPB45+JtKdTAe6JgYfF1Ir3obRKQNE3Gsbjjtkf6RAEDWcTlcVMMIkAGoVaWqLYRdT6Ckh\nUy8MF6KrBU+/T+sk9RZRFQIyHllticduW14cPNUFeo07IioCmHePIvX5TBnF5abvBRCnoYBOgLbO\nWSCecG6YKqyE3ObI+Y5zYji4ZgSKHMGO4ZrN2c7J5uL0eMF5ayatUuDDjEMTPbx58Ry+qR8UETy+\nr4jwy5sP2JubCwnh2f0zhJZiuuNk9ce079by37YL9j2aFlj1ih6K2m3IwFPUq5yldJ0pXN/4OrrO\neLE/r/T2VXUD2ZQvvCMUFGvvCzp0IMWIMClBtzpiK4QFBHDq9CGT7i6wsoVe3/34Pjpm+2stJts9\nIB2OcTlvODct/7hnk1GuNaOuHRfLjm2vrzvKser5ozWxiPvikB3irNSZDet+avPzAdv+vk25poli\nLj8+DLr/0ilTKO0MWnrWN3tV8QDq+dUFJu0JkgrKuHEN86obhuRc5aobtIJLgW/nkAVWNyzIiKtu\n5j8+0XtKCZ/G03gaP5nxmUV3GEH5StoCveCtmqB+SAvG6Mu4RPKRUD2RRRkpxa6ESQLvGoLdORD3\nNKSgc5MgMN4qsdU/9dNBcBa2MufOoaMCbt0KyQwvbGoEJBlLA9HJgZAarKHwjtSQ0Yk8krjOrKcI\nbjmAF8Hcwl0fGDvvNmeyMdDC/NWtQCM/c2agdHR8kffYSInW3XZ+L9lI0ewYPjgD63HwSPq6NeJ0\nanblbx+R3tZOYCBCcgvSse26i1hzY7pZENrO784TOEZwi36d9J0fZdDLokEH9nu2QRojcu6FexAs\nOmbuzjQDPxYENtCyJ8KWk8mxpBitY7jvO3zjb5J37fprX8rDY3TNpvr5WmjuZYrfOX4aixOtQzg0\nh1R+en1czZeTyZsR7mFmgCPWprAQ44r1XCPeeHsLcj26Y/bWCPB3C2JoBe59A120Qy3wyhvUzl7J\nOLTvmxYP12ofksi4u8HXYjwZAr2XTrwPBhWSIrUBg0qOLznbuWUi+HZ8YRQ/EAfkgtgcoUgipEXr\nNdrSqLGgbKp2+0O8zT4+b8ESWMu+/mH9527nU2P/Xt8isnrS8A6MGuKu1Tf0VVeCYO66mzgSKFPO\nhusoUqyDUw/u+q4R7ooDhQheEfJSDENFFECFwNJuVgFIU9tbIHK9aPK+o6n9wkXGHrvELucFQdvl\nWI3IPPlScUMt7dpAaDAblN2BW+Eqc1enIM8okhGbHLOfZ4Mh5NI1vRx5BNdrD1myif5t+4at1Tni\nnk2XHlQxR6lRMpAKoBZikwcHfRzgfDBBRBaAy6jPpfPccWVEn1ixBrQ8uKfA+qc3YcIAH1pnNhOE\ntSvtbCEryIj7jnWtC3HMEVAql/d2zsCqLuHsuxRDNnapUzGOO1Isvyd97cj9Ig3u0d6XY8alUYXW\ny8nqZnOYTN6GSBpUo5PjL1x/Q9p3c1Jm9k3hRBcRD3fQrttsNdfLeUOYau1UKTox7jg2sYApsBnl\n5iFtq+RuGqgxYl113wjj7Rfb9ZL2XDNITfeZTd99nibsbZEUbqmfYvrio2nsUy622Tnmvnn/eFTD\nU0r4NJ7G0/jpjM/Tw0JfDWkI8AVWv6v/MkiiyABKA6jvdsMurCRN87VDJ6Z637uJapSg0VwuuXd6\nhogDw3EanoujHWSmDnJE9kMnagFcvOIMSlKQJIHb65yf4fWz9g2eCLlohAAjSYtL1v0iZkzwhk8p\njrG1LlcAGsEVKDh1batLqemSygxPjMdzK+Lvybp/5AjeTwiG9h5OyIh9YoLzvUkhxMia1ucMKopQ\n72YDxIrLab9XxITveMTiYWiAmGej/gct8h5zsN7drSKjLb0IDqzH6AmOhqhZO1ESEfNu2mx7jCDf\n0N+rM2FIFxy8BLDXyAJIRQGObAyMPKD31VPyk0NkKLrXD9QC+vn8iIcP1fR1Wy8d7OznQbOqOkp1\nM+EIUlHJ7RFhayl5oCr9o00iTya250AIzcw0TAt8axLQ1vBWWSzSnubJ7Oh9cFBJ7sw1He5GLB0R\nO5LHPTko8lQiQIXBVlbp3f9lnrC3KFJYgBQNSLqtO5K5GJEdG7MH8+9eoz80PlvAT9clln4BEbF1\nJka4Qx1dGbOUrkcdQui1LWXnS0eV0xCikgFK81VNrAzpYRkJ1yKWomm3y5fFOlGZo7k9E3wveLEg\nFUHW1JLEUPWUArwK4REhkupftWyKdFXZkEmxC2KpHrkAR51CUjLj3PJKRoJvxpuO7vH8rl5kZXvd\n3HvbQuQd9gZIjFvWBiTmEDCHyYTrvHMoBh+Bddeyh51xgaBwMZ0lkmLdP3IBpDeCn0A+QmL77aXL\nCYNgYFGFs+hjQOEhurgVSNRFPRvJWLhShkzhQrpu/XSYEFqXltwRjWWCvEeAC8KkoGBvqbZzhMtW\nu2ipJBxu7nBz1zqhzsO3muRcCLNq0K8XSFNpWM8/QOjWKggRaBC0fDx9MPUCyQWsDAfqEBiGtNxT\nu3MRpRUbU9wt3ScOTR9OkfyDUxQL3NRqUVPXCKuWcQWHZekGqSxYjkqZW7AcKmTjm9evq7591vPo\n7T5hJiwKj+F+/ZZcVWJdA4SXEu1+DctkG1DOAsnJupYk6OyKkuzYPbGtE9dyyb9/PKWET+NpPI2f\nzPjMorsYAI0CWcH12iy1RUAaWZDrRFBwb+VhSNe0mD5wttwgsUoWLXzMU/yoMDqkgaPwfX1zgbd6\n84SB4ARu0RIjIQtbN0Www7UPDWVBaQRUoYI0eJYXEohXHFXHncigdU8utO5l+82x66E7LNZVnaeC\n57cvAQC+HPCrb75GaNbmILZQu8RiBFMHj8Aek9FTGNJSiCpR0iIJ2bHmBhwtgE8bQtslZya4oFwQ\nASkxNkyA2wDXcTZaDCgYsjwa0nxtIAwRuRR0tdPSo3B2UlMDVV+VAhkcXFSrnjzMYDaIoOSpF7yb\nwW49N2Td05gzFoFpYk3TbNQy75xF5KUUK0dop+9To+qi94heSrYU8nw+IbbOajWD1V+YBlnwVvhX\nd6Kc7dqL24rLJbTfg0qEbhHk4MtRU7amC7ccZxxvGrc0RwAPmAeV0BURXl87z5gXpcs1k1XDYcEy\npBAYYdL7rJhKKTMhpYgUdZ572YYcmbLpfr5g8mI8RYIHF9WAGxyc0LMobfD8mPF5KWHpOtMeAWMk\nVyySlqsODNDFzNh1RDwG4iwRXdlRiYgxwae5AxrJXjsknfLRn218nG6C2OAGwrG3r7PvaQ0RwDDg\nKHsyyy4kmPY7SMBGRq1JVWoSyVNkuNhqXTSbi3EVCeyCdIkLCisCu3Qn6cCYGjzh5avnGHH+MSWr\nqZUoQLL8BChkmkQkBEfaVu8pSIZgV7BlqVrxRvrlQext7Pg1ccHcdwMDSJIUOENSdx1vQz0Im/KF\noEsIC7ouEiNAEKy1XdABvkTc5aHLWNcMwMzGV825YFfCc4y2iDI7sPfGzlgOB7Ci/vNsC1ZOyVQN\norS80wYBw0ZqC6MIYoy4NAJ6TrtmpWAWqxUKktl8cQiV82oLRTIliRhXuFXLBdWUdmnX2Hxchlps\nN48Iy4SpmZ6u6wzgofF6Fb4gZp+2HG7ByhRwvpKZleHBA6zEA8tBzWWAnFXFYkOMu903XG/Gdg4J\nKW72usl3hL1z3E1jCszFHJDqigTgMzLCzy26iy1MTJ3VLUJXFysN1AsAtpJfL2RDgbzeb1dUidGC\naIysxnXpikzLHa07Rli266cJqtpY5GI3EjEQpNVrpC5qXt8jC6zewMUWvESVmFvnwcEV6YYBxFAZ\nD8FsKF6iXOsMxrvZQa7l+SlhogqZYDkgQXfFuvjrOd4vO9AK45IJ0m6EkgQldZsojt7cq4kqORWo\n5OnYCKuSBH6eTM1CPprPEdFN3Bsd0swOADR604B2t86LUqfCQI72V9G1riqFHJgm5HYpRmEI6WUZ\nDB9VCsz3sP5MtoJyEQE36EYl6+rBO3g/Y2ko8dubo6nOXta1e0QyWy1oCqE2lK7AZH1j7RsB4XH9\ngPXcoBXbBmpYtTksHSYwL8DwPTntg378bgvg47vXODc6D7sA8s6K61989TMsh+aEfjiitJMb5gPu\nXlZHJb/cAP/mu4pOV8T8NOH2rkEPlltIm0sfZuwxGVNhhGsQE6a5LTaejPy8xg1Jcq9hE6y2zARz\n0EreQWSw2RvlkqRDW5xzJknVZaV+eDzVsJ7G03gaP5nx2a45Yp0y111gqdiOSc7DMZvWdxl0kqoP\nXtv6PyKOjnCF+toOQvw+P7hqVGoH0fmDwMBRa3+OSkYy95WaXfOLQyUhi4fuCFJSlxnxqacq4MFl\negchI2iLljZkq2cJStNZD57giFFUP547UDbwjAM3NL8QuKV6rlrrWIS1rdEciCrKvAFo94i47dgb\nApqG7pEnh0PbneVGuu5WEhyOB8wtpQg+2PEUdKG2ijzvWmOFLZCtKdLgFNM5gPW5PZGRZ1HILM5K\nKdal5cJgMEQlq4ktVSDy0BeKAI8Nsb/vBfO8mBSMD77WHgGAVvtsRwzHzpyLluVQeW4AHs9nE+tL\nKVrkrp3Wa/wC7DlNo3KublCKFqdSrpD89oiccQlVE0ws4ihGwi9pMwhB4Q2SGLmRnE8Pk8FUfHDw\nDcDs/ITlqJZfLXKWDmJelgVOoyU3IbZOb5hCtUYzg8leqimODBYyzRPWtYGgt3rtdWs/mC9kcA5L\n01zbtxU5b5aiOxSwgbt7GYhdwNQiyNGi7ofGZ5OfdTLqRd05FLooOVVu0ExAitVWmHs+izJY/oja\nn/e6yZXywngEFYbcnx9WqWsn5B7GA0AhhlMvwuItnR2/QFAqZqb036ifU9jZRcNwlvOj1Yic1/pb\nt8TiwSqKqEGAW+E+BLKaR07A3mDvC0+YghbPBQQ2Tah9T8bO56G6FfeI7bLBK/xh8phaGuLnCbdL\nbWdP9wGHhoyWUjWNlkaY5qkrQeRchvSuXmQ6za4yXW2arUQoMtRmWppzyf1zCpB33Txg1xG5DMoZ\nZMh6b2kg4DrRuhDOjy39iqXaydn14gH1riTXTygcyHlbhIIPZjOXUjY9rH3b7Xz6ww+kJ+2jU8xg\n6VS1aZq6cgVx34+lpzEEbnZ09oQ9JjRNrTY3pSSUJjZ4ev/G6nI+eBycatbPttDYQjuombDrDApi\n1inCFGbk/IB1V/9MdN1/qpg+fb+WOlLmlkJqnVOwNTT7CLHJKdfGDKl6Bhn+i1yBatkVGty2f9BS\nrY+nlPBpPI2n8ZMZn01+1iGlGBFyTNt037eapZBpb9fXtujIwdq7WqDT7ocqkOpj4399FKbzWMQf\nCv0iHbBohqxOtLuPIs7gCsjZWtWCDKFkOxSJh5CSvRloHEMU6akABGBBKapEOaRT8CDp5pVgHnSb\nCMGrGmSNkgDUrmKLlJIKKQ2pqHblHPfiPiQjbhHbuXEOpwDf3E28dwgtJTpMN5BbTXdrO1rhC8KE\n3KK8mIvNuZRSd8yOTenzj26+ytLPj57j1+/O/dzJEI2Rs9ZQ4QJZM7hFqPPBYWlyOkCxAvweMy5N\nyyuEuaY7rVhbwOhKN2wa6sQO87xgbqkjc9fAklJMT2u/rBB17mZuEI2Pusw2NJ2rrjyHQ/ts5Apo\nRYuwNJUt0rXiwCAZOnICK2k45y2lrJFul6I+P37onb9lNo1+LAQyrqVe5x7TpBESD1FMT+vneUYp\ngnVVXmp1lNY58kGvncnmNZcaUV3aNQYknD7UDul6XrFrtBVLZVGY2XFvmnEhpKiNLzGWRb6+rX/v\n+GzysxEWS7FViamnehWD1cmt9cnh7e0px92XUFv9KtnKzF2+FRhgDFKpInrCqd/AzZK4f6GmgnqB\nlWSLV3ZDvatqPbdj59rxpB4eZ3WBRuoKAxGG4xBJADIkq4wzLGWSskPV2wQVN6THXjKMUByQDIfl\nQQYbQFVqt7oJCVnJztFYR8uIewadGqXITfChXhjeH4xCQd5bTF0kowgMnhG436RUxNr/JQuQpdsO\nCnWZZemifbVKOOLygMu5d4gAGrq9pWunS0WNm6Z4Jixtk3BuRrH08GI3omOHnLORurMAqzL/QXZD\nz8uMaQ4mbFgNKnqd1K4jkDED8iAK+PuGc3UjkNyEG0UQqc5/iqnv2EKD4oS0eqleb73WGPxkWlO5\n7ACSlSZS3hBTq99tF4NgcFgQnG6IWqvydi8WKaDhnnKlL1hZiqV0pfS0jNlfLViKW4MAj48XvP7u\nTfu5PWWVPLJMqoWeYvrmw2zI+Zx2rGtTp8jRSNzlB+Z6HE8p4dN4Gk/jJzM+u0s4alhpqledNuwV\nlWelzSMIRAuiuN65xsfMnejqXBfI59GCBGhFS3XCyUPn7jphJEudGs6ItursDKAU39MT9l3+Rgqo\nMIxtTBkFDRmOaFHjaLrB2cMVxq7NzyG6Exa0xkyNYFg6r8o5+KZ/FPdiSqJHFnXrMtxMV+MU43zl\nEk1eVxpuLTb+1pkv9t6UHbwaZoTJot3UOqBa4M/B9XOYxQCqiKXG7AZKFTtPTNR31nYWgB45psyD\nnAs6+LQedD0X7GontWjRHVBXbvZz1VcC4DhhOTS9qEJIOeOynttvISsgp5wQmqHucljgg+v4PsZg\nUTZce0RX6MXazOopob4f6Omcc7WrStY4EotSvE+wKIpdV2ZtxXhN0/x8wO1NA7VOi73ucf2AtEVz\niJ6CN6L2GKUTqOMijazujKlRRMxshZ03oxMXAogY54Yh22PGjXI2iYz7GGbuYF0C0nk14xMfyO7R\nnK8ByIWcOVTBB4RmOcfJ4dyYFnAEMW29fyKkOxFwc6wh8BQ66NB7bzdClgwW6VroRBbq1muiPt73\n3U6QiMA57idomob2cr/IdJHQC8g5B3WRE7EIutYPRMnT7SIfOjKOe6o4NDQBERBCBw1yX6S4cLcj\nKgT1g0P2gAi4ifu5UZ6CGFNLbyoYUcDWQSSbCxBha+lm9gCHli7Fqtag0Iqcd9M9KjlZrYbbPOt9\ntT1u9njdMtyxpYrLrOtChWI4Qmk0DJmCtbMpEUrr6OW9IO8ZqgjoSoeBQKQ/BqxGqM8V8tZ5BErf\nd2iw8hKpKPf29z2LAfi9cJcXDgkhqOxwxB537Jq2gq37Ry5gaQvW8XiEc2yA2oguvU3SbxPHzs5L\nlQTueu0pJXMRJyJzoQFXq3WvKHpic/Xe9rXrRjlnbtvIBcze/AeJg1mwzcsR5jpEhJQLUhP6C2Ar\n9JQkBqDNqZi/pjlIh8kUHCrcSM0qg4nxCwjeTQYTOT9ueP78ts1lNv/Ow5Fr7bcdA+3ZkOrhZjZV\nh71sRocq7R7e2/lYc8ZB74cpIGrqHwJEa7Cf0SX8vAWLGaHlx2EQ1mPqNykJV0twrQs5BsHCraGN\nK9b6RLuZjQ4CsfdUakj7kdJ2MsWCFTF+2ah4SY6M6a5/EgWY+CDoage1yEkYkAlqWglkON0pS7CF\nowyC/sQeSNmK+CxDFMXoiGeaQNTF+ZgYvr1nAsyCPtnqCbgiV8qOgmIGEMUNkiMc4IVMhC6DEHXB\nQYKkGonE0xmprSrsBbfHGe5ZhTUclwW+zeW+rchbWyS3jLxnUNJIoheUMXjdSemGtKpVL3mIorl0\nAT+CFevrJxCKNUy6VnwuMJXReZqx6s2AvWKolGnBweqQx5sjnjcDjuf3z7DMi53fMkRKYz0V6BGV\nmqtqnWjf914zYsak6OwpNLR244D6AFaFi9UbxAFMXZxwz/DzjKM8a9+ZbWMmP3cbLp/hw44UGydy\nB1gbU27HOtWFjNwO17w8TWDETbYICglyg8Ssa8TlsS7cHz6ckbIYW+Pd2wfc39dA5NWrGzPqmOeA\nm2P9rHTekWMyhkjeuhlJWvcrQ5CCritf0Avwjjtkoqq69Fr0jx1PNayn8TSexk9mfFaExdRNLksR\n2/mZuetLyeCegtow051Vhv8CffctpQE2O7f/6jXWTSz5yr4WnIUAACAASURBVFGnXJkxdgeTIgMT\nXFHTYIhBD9xAyi3NFbdGR6MrSGWq9shRl3cZdKQKC+A7WA9CyGbSKkOXss2LBV9iUZmjYpLNvpB1\n6KgWLAypfVhmPJhyBTdCde2uMgDZO6LYxNoKYVvrzrqWhF2BqzNjWSb4lsYcpoP91rxfkPe2u2+l\nAj5bMc7hGl5iWuiFoEwzsdx8ta4ciWAeHWmszkd1XjUSdQRpKVx2O4pXPSdn5rB+CVhS6YBDnqxu\nczjc4FYJv/OM4EKXn5aMPWqtSzDNNTo5HlMnQrfj3bamc/X4iL09ZscGY/DhDoJO7gZgEJHleDPw\nBbO5Owv5er2pKEZO6DlHaKavAGjG8XgP387h6eFkJY5SyFQvcipWw9LrL8bevQ+BkVsY9fDhEb/6\n1Xf18z5s2DcyQPK77z7g1YuaEn7x4tbOX8kJakaXywaWbPdQToL1otfIZhLVBJVE1yicm1YXwE7g\nWkSZc7/vyP0T1rAGuWf7YcJ9kSrQAmRPuXpje6RvDBZJ7e/FIPyuWxKN35OkiplpHYi563cXsZC/\nlGQF6a68yfbQU7YwNA2F2IJUay5lqE81ag0cBmmRjNwULotLADm4VmNwiZG1UOkI08hIJzGzhMAT\nuNVksqRa7EeFLiiUgud64l1Tbzgcj5b6iITeOReHIn0zydKVXdeUsarluauEWAA43k443t1iUjI0\nM9LWbcmzihYkBnI3LChwvWWPfqPUtN46DwCAm7lvKMSE49zfZzAY9rVAq0Vlz3Dq1Sfdysu5CW7R\nhdtjIddrPoV7Sni8xaQ1o1LTc7vKChkim9jj7r46eS/LAXvShSwCAqwNVX95PCE1+ITzBG4L/jwx\nJswgUufmDldwwVuJRCR1CA0TAok1kdbL2fTOz+nRsExMhCkwgj+24+tI9DAfjBTtp9BhMu1HPjye\n8arc13+n3hyKe8R3DZKwXgQOk3kgbJcNH95VK7jXryeri6UUQa6ei5tbj8Uv8GYVFi1VLvsOX9SM\nhCElWOyRckZsk+5CGKh5AzyGf/yC9ZQSPo2n8TR+MuOzIizHDjdtRy5FMCDibKiAn8EfhjRofHG1\nTFLQZ5VONuAhelLIoO46TLVAN0ZvGonlnAcUs5ggWxq4S9T0rMqw6xLBxOnQxNTQdvVCVOViUEXy\nFMBKtJpA2yFMWB8zVPCVXTEE+3EOCIOMDTnBTeOqHcIN3r2tu1oWAjebr+yoAQdrSA/A0MZuPuLu\neZUTOfMZ2/vmuJIE1bmoTXQIcK3wunhATWMRCHP7/uPNjNv7Zzje1VQgbRGv2/G8f/0BqRVoPSZM\nxwW5wQb282aRHHPn/bEfmQr1DL58fuyijBAEk9bpaQCRg3BA0dTde0DFAz1Z0wLUBQYFaGmG7tad\nVD/Ps5mhTtOCEGZrzWcki+ycdwbILSVgUk4ePEQEl1OVO877BlINsViwNnAuZMPNs+e40055OFjk\nKYBF0pUP2TWgkNnO076tOH2o5rzr5Qy9NyY/A+Vgki3OTQhNlM/Ps2UVIXh4u0bq+bqsq5mVsmez\n1ZrmYPZf6bJBSrZohYVwfqiNmW+/+Q63dw0JME/44198CQA4eIfFeayn+rq3332LtUWHJW6goij/\n0KS3FVbT3c+FYKDeTMk6TGXsNP3A+MwuIeGw1BB4jxEaoLHrNSGUcqXvDeDa5stwO95IwoKKPGdL\nn7rxBDB4D9K1OGDViP/E4hV7KzoPpgMGDxAMuX9HWRdhSGY4TW85ISs9JidzPw7ed+JyAErYB5pS\nt4w/HEJv/YIxuQk3TVd79gfc3tSazBaj1aOOx9lwWBMFiADqxHXZs7WNC4eajgIQyWBmzEryPR5w\neFa/J3rB1BZA8YBv6WWYGHAOW9M7ituO86WRbc8bpHUJXWZM5BupGKCJlL9dxeRU68lzp1OxADjh\n/u7QNeNxbbGllBtiD+IJuf1dnEcxJTzqm6L0NBJU2QhawyLypozpfbBOlHMezN3OSkS683BvSlcr\nMP0LNzXOoiT1bp8W04bUdPi3FQhhRZrb5uIPVqvJ6KlwQTdeEClVbLCpROz7hqy6/py7eqsAKGwb\nQ4wrOgbDwQ0bs+noK+tk8ECACELD9z17doM//dOfAwC+/uW3ePv6wXCU08TYmlfAd9/stpB/8bOX\neNb8BW6XGb4Ab37b6mCnE9AWr4JeK6JmCqpd4FQ6BQcDRowZhs8k/vGJ3lNK+DSextP4yYzPRrob\nsZTICp3suqwtUHeyjrnpll8AutIn9R1ZGi5Kw3pG5wWWUnqUpAhkBWHnCiAEqkTItnW5EH1ekeGQ\nvtPa8UOLxrq1ZTDH3tVo/6uvy2Y/hTKbdtFhiZiWgB11J5JSnYkBAN533Scm3NwcTZsKRfDqVcXj\nLEdGwwji2d0Ngu8yNgBVZ2kAubARVtd1x3lXJ9aC2U0Ic9N9enaH2+e18Mo3M7KCUilXMCEAKRFA\nwWMr9J5PZzy0XTbBgUyL3iOzM3XTDLEUPQtMx8qxQ3FaRK3PzX7QIyXAZFCZr6SJwA6lzXkCIWtz\nAzA8Uy4RfTId4Dp+pxakNeruyPR67sWuv1SyFcOZO5tCXAeuOs9V7kj1pzybymjOHbQsW8QaNiyr\nRljJlF2J2EoEBHQ5bq7HrETrHJNdn5UUr5FTRMoFaHJIWTJ8UfbFQKYe5GP0O453t2Cv91Vngtzd\nTHA/ew4AiA8nXD58MLJ2SYzUUrq8ZtNYv3t1D7RoNYKRUsG54fset4I9akS9WDOi5Iaqr0cLsRi7\nToBGU8TSAdYfiRr8vvGZmu7FNISI2QCqktkUChR2YCx+5t4Z/L56Fpqrr0EN5MpWz8CHAqAAsWhI\nveNyqXWcx8vZ2s95j4M4W7LvsE4rZevUlHrQ9VgBSAkmQVsoAV67VLnafaPWPLTDyaEJFEpffL1d\nUGSp4hQYh0Mwy624buDWgZkmj9DesywerOBOYuQiOLVU7bLteHeu879tO1LR38bwnlEmFfDzQHs8\nH49AuwCjJOyxLeq7IMYVW6NnPJzOuLQbiW8O4NR+g/i6oSS9gd1AzrY9C/Bu6Py1tCsXS8NBAjep\nkB3BgntuKUTLeWhU2igZpXEZMgQ5az2wIA+NRQGZ1hOTs2sPEXCFDDCcc7aNLKXYU0oMIpNtP9Q6\nZDUo1tqnWGeypIy0J+wNMrKHzUoazntbAKv8cIOEsEfx05X+k23sAutsF+SqSabXAXsDYjrHmIJq\ns/dOqXbcbm5uTJSRXDHTEscMr2KSLmOZgE27yilBdCVwwONjvUa+/s0bPDaFDM8MlwkfXlfvxbfv\nd6AoOHsyt+4pFxRHtjbEsvYufztXdc6p5+SfQc15SgmfxtN4Gj+Z8dkmFD2E9UafEaSOwyowJxlA\nd5g6mHmIojqeCsC1hRLQeXuDLIkW1jUsTykZxmNPEbsCDnPqj1uEVffQFqa7rpxYNZcU3MhAmSAq\n0OPJ3GmrYUDjaAnBtnQn4DJVvlf9kbYbgsm6YdN0AHMwDqIPE7YWBW57sYI5HMGpxlGpOkQfHtU8\ntcCrtpObjaLkyGMJE5bWwQ3HG6B9XpT+SwVk0SPIAzybhDNPBxzuWyeKghVxHQUEOPhhZ7wyTGXt\njqLL2JDgl//vWzzG2YrmRICnOn/sHYKmnMxI4hBbupelp6PsJiM/E6grw5SqyKkpIjtSvjQoBOvq\nlVQgUcztOZcMJV0DCc3opTZ8DOBbSc3afSsSrQziHOPYuLQxpaqmqVe3FOPZ1bRoiAWsJlIL8N73\nSMyaRuglEWpYLZVdIp4sTfOTg5+6SIDhsMzcwcM1U96UV0Szhdvx8OY9AODD6YR1X7v9HpFFryIO\nLQjHm29PeP1dNcbIuVTqVou000p2TcyBcfe8gnWfLwFrTnh/et/OFXV8ZSqG/UpScXbA5+Gw6Id0\nf8bx7NmtfPXlq/rDvifv1BTOGgPUQYyfYvbX98jVe65gEui1J2pSvdbR+SjDNBfo0tHxgOB8OuHx\n4cEiUAW3wv7Um1H/+7tz0l+lR6XPXztdV9Ds+KpP/KjfnYLxA4ZQGSAOOD6rPoUpFVsUbo5HzLMS\nq6U5lahC4SBxPEoXD4+LCEqRWitpx1oMQZ0Gva9qr64/o0jXJxtBvHrb6Zzc3d/h8byjy2J8NBVX\nU/zxfNPvPCT0j9In6NN/+eh7vv/67ufw+k1EBc+fP+//OJzPj33N7bPoe+6H7/323/8Cqh/6A+8d\nNnwA27aipP0aUmSf1wndpdSN30QaP7q6P3V8VWXi+367HXFL9Ya5vTrG75+N02UTubYq+uT4rAjr\nqy9f4b/5r/+r9uVusEIi20VSFmwxITdsxTRNlYkO4OGyIatMyaB7nWLGHrORd2moGwTvzUg0TB4u\nOMMnse/yF1SArfX/Hx8vSLEj3f/7/+6/xeXNt6ZGsEIsuolbh2d4+OqK5/QSLiOCxqgEDmzEzywF\n+1DMTSmZqGDgfmf2i6jPp2681/fuNbXpw8r4D/6j/wIA8PbNCTe3Fa7w7/97/wJ/9ec/a8eTkLcz\n1lOtL+TLCaWZeiIn7O1xjBu2vdastm3D6bzj7Xsl0k64nOrrHl5/C2wqGHfGejlBWSPnteCihfp9\nNfUIL8VI40UiXvyzV9jSfY/IiSxqVuPc+nOlFoeH36xRlZB00UShXgtpm5YZ7A4wGkG/SYjQCug2\nuf18gOw8ldLrRRDCt9/9Df7T/+Q/s/OhETMTDXUmAkCmSsPcFUtAsOivUO7NAmgZti/sWnui0oX+\nAAEPEkttl7bjGalp9g4R/I//8n/AX/7iJSavZrHdDIZQrK67bxdcLhdczuqpmNHrSJ1mJtI9KFOq\nG5dSl2T47c6x1W0dhytYU8nRzkeW3Cl8ZSTiFfxP//r/+lf4EeOzu4T/mKG4juPhtmloAtuesZ6V\nr7XhfF7NPVakGE/x9vYGN9woCocDpnm2yMIHbyc5xYzLpYL9Lue9R3htMfqLP/tn+Pmf1Bs8Bsbf\nf/0bAMDf/j9/i7UVGeuHuavo6eoq1G4PGN7oN9XQM6oBgQDGnxt3HtIbqSdoPya6lbzDffj7+hkP\nJ3z3bb3o/pfHX+OXL2uX8fz+HU7v3yNeqppj2TfEtuCc17Px2lJMSJoqxx1bKliTHs/Ub7KYTcqZ\nKcNxseg1gw0UmAcLQHZiv1sr8WlwUa7Kru03Sf/t1ArZ3d0I1hCpDWFNMwmuzXnM1eh1nD5dDIt0\n96XfcU6SvmAxcV84pNjz9TwD60WbGxuODXsYHGNvuCk4wjwvnas6qIWUmBFHqk/p3XRuTQb9TXrX\nxpQM7KyKEVMD/x6PR8PzjfNacjZMlz53f39vEjjHw2ILlkhCbPfX+RwAx5D2j1cKqeiyyilJNewF\nUGRHGrIgIkE2Oh7Ze0ptoVkzIUuwzUDGCJV6RpRHGMEPjH/yBasqkNQDevbsGTZdYN494rzW/Pi7\nN+/x7u0HnM/thpOCpSF7f/4nX8Evqu/DmKYZsxJQvUdS54/1jPVSH58eo7nG6In+iz//Y/zVv/WX\nAACaAnwjFP/q736NC6/tWAtq8KyE6UHvSWBiaA6EP/2TP66f+1d/AXeY8PU3vwUA/O//x/9p7i5F\nyG4aNE6lWTyha79fDxqiMEJJEdu7r+vnPW5VIwvAh282nL/5BwDA17/6Nd5898Zs1l1jDgBo7kZD\nuK7fL/XfkiqKSbZOJxPDGYK6IOYEUV2kacasO7gKHQLY9ovBSvRXpRih5L2qv6+rW+8co9TFK5sj\nMEPR85W7qtGSs52/oLQord34OUPv/IJiC2f9CvlowbJZ7trqxMNiWrXU1YL+8fFsXMJlCohtMXG+\nvZs6SFU/P6Zo8s0pJSubEVGVWxogP2ZjlntKL6U05xztzI7RGyEmjZQ2I2nrgpX2iPenip5/K4K5\nLXohTDjc1mO9vX8GmgjLXfs3F+xcpghsjVP6eN7w2IIKWTdwyT1Y8Fx1vwBs64rYFts9biByPeIi\nb4t6mDyC3ssl27HTVano94+nLuHTeBpP4ycz/iApoe4O881NldwFUOSMU6uZvHnzHt+9fovHxxph\nkQhu71vXYX2Bu8HvrUL7tV7BSC3HPl8iTq2bdllTRS2iG1vMhwlz665Q8CbeX8oAemUCSQ+J0d0H\nq0Jl26Bn7/Bnv/gFAOBf/Lv/Nvjg8fyXtUj7t3//9zg9tN8B3+sBaGnHQEfqjASxtBmtNqKvESnY\nm3C/5ASvFIyympDgvl+QcuxYON9rZlKkU1Oo/x7HDsEzDr4XDjWadFxBggBwOl1QHHDbOIevvnyJ\nuwZ+naeDfd4WN7x9W9UAvv5tjTZz3u0Cq7Pcu4hWyG7RljnOfBQRFUtBcpcpcjxOJQil801H6WNX\nQcFkOKDBCw89aClUrr5TROxafHj/gNSilDhNlt74JYAcLBJlZqsRretqDj9p7zVSHXaEhB7BF+ly\ny+W6GJGWxbqOGYKL+imeL5a6anS2Xlbs7btPDw+4ua3n6vb2Fn7RTmNAWBZwu34ce1OxkAIU4/R6\nU2dZDh6gQZivxA78ZOqPhUEUoG3bQuj3k3Tp9Cqr3D7L/fi46Q+zYLVBzFZEjSni3GpHD6czTuez\nWQg5xziYY0i2dKFIRTGnpDIywNpC7/NlxflSH8ddMPsBCQeA2JtscxEPUQkZCQB6tw3kulsxBvef\nIhY2z96bvvi8TOB5MtmWui4NKO5hIaprV+9eahpTb5KhZzVc0UTAMvf6wh41petNipIjmIqRYqs0\nbmtaTA73zytn8XA49lDdOYCpt98FNmdhmvD1r2uN7+2H15iWA776WSXBfvnyBVLDA2zbyRbk29tb\n3N7VRXzNEYAg573pnAEYwI/MYsDCkgtSTHCKxs+pN1/QC+0YLnw15i1aAC/JPs+xM+1+bZf7QeM8\nDfpcpgw7sDEk18UrtlJD3Dc4PYZYOljaE+YiVi8rkqxWuG871gbIzbkgsF53FWLRO6vjIj6ojEm9\nVzCkjqYJl5OlUpfL2rW62ls/fDhhbWBq74KBVCuav3EiY8b5smLd2sZKzqA8cSs4n+tvv6zRmi3O\nO/jJwxmkAmDVIHIO1N7PnsDSeZ5ARwdIARBVlECwK1DZ2y//wfGUEj6Np/E0fjLjH+2aU+FCfWXs\n3mJa5O3Rw/heVUuUAmtLhxCqML+oREjA3W2NCrQzCDRYwxA+liJmwrrvGZcWYe1rgiy6E6jUyDys\n+gHU5FyqYUUHpgIM1ZoXEtvNmTtb0jNjaYX/aZkgzncNezCcmlKCLUWpHaEhFWrdKKDRM6hHOjZj\npWK6bu7UtGADoZtf6nAM0OgsNBaeCT368J0LmrYN+56sbR2CR3hZ+Ydx263l7Yjx1atX+OplxYJ9\nePcGX/+2NgFOjxeLel68fIE//bM/BwD87Bd/jHe/+RXO6wn+RsG2GUWj2gHHFlPBHnt0EmM0SEyN\nQU1Wo/MPoXOk/80WjATvTAcePLUWe1OCGNRrqzpST5U1qqt+mxg6vV01IVE0fXZHdd6NDlfEPA33\nNRplpxB3WaOckHO+uie6JI2YlI6m69oVdQPMJ8ZoTYC075YKqpLn+9PF6FeT80bLmnIeuqKEkgRx\nU6xeRm60s33LOD+237F1+tI0TzguC6Zm8hvzjn3voF67/k02Su/zgtxA1gnFoqpqpKomKv+EKaEt\nUlcLVl+grl4DXPMKvYMSPJ2fcPustuW/TMDxeG+L3vFmwb3WsL64xYuW0tze3cB73/E4RYY6TsbZ\neHa93z652pJ2oS4sACCZe9omfJ2kSTaMlQAmyO8cX+GALNVRrphy2aQbvdaf3TFBVyaeQxRM45zR\ncFM1Jde7uzpPuXxAbjVASt112bl6cSvSGlQsxYlbwrvXFXV8ensy0GdKETl1cuzt7REvX9QF6/x4\nwsP72sENLuD+9hbS6jNf/8Ov8KHVdwoxYgv3f/v6NQ539Tz987/+K7z7za+qLIqK5Qy/V9BR61XH\nrJPV9303DF1BN+gkMEidsttC3rFtZegmBpOaybl2EpXbSHBdXbaIXQNbSoaGN8hFu65CCEgNDuAI\nmFop4HizIEy+19Wk82zfv/+AfW3dxMkjtdfs66V2TrX753xfDKXAt+7rzc0N5uNs+l9AMRek8+kR\nW6tbiXQitC525z32FAwFqoeqCWj7B6RYsLWueoliNcttz1gvCpWQKkOEuoguywI/N0jLWob6bIcr\nkDCEBqwZwerMdS2APVannRB6+vhD4/+3Gta1K+Cn8UU5Jcv52TNuWhTl3AEvX9UICgDu7m5w36D+\n7CKObZdeDvUEb20ni1sxCMHDwwXnJjpH4qFszh5V9Z0WQ3tfxpBGCoBitJSq6T78KsWdEHUsIF8v\n0CJkrjK1P6A1mIYBstd216DKCe+R66BrCCmC22YLlQ87crtY07abW/RxDjgzIavu9yA/7V33xcu5\ndDeXUqvWFOr3vvzy3gqsv/3uWxNd8yRgz1hbVfZxj9UpqH6I3WRxT3j/ri5yx1ALvT44gzI4rsVx\n/cEm+46MIl3SupSMqIJ5ZYjQaXDlkSpxbZ573OlbzNzBnK5CnjRSKbn0pkcqBgte1wu2rBQuwZHE\nAt4Qgp0y58VECtlXzbYuIFmQ200vuWt/OVBXTwgeMtawBkaCYzK1B+8d3NQdnLMUi4TrhqMLda8J\nk762sGmNhfkA13Tr3bSYqGDJwH7JODUByBwLWuCDfc+IUfFzVIncqNc4c88kKoi2Ya1y32Sp9kP6\nRkPefBL21N3XiyQLUJz8+GXoqYb1NJ7G0/jJjD9ol/ByuWDfdcd0OKjh5cHD+cXcYW5vF9w2pcP5\nQPC+r97r5RGxtcBlANellGxncNQ7LLpDLMuM0LR9SiEktWQuPYeu2OeeLjIVBSk3QKK2YckiC5S6\nT28tVUu5QKWWeZAWEdT6FRkZuKODhQoUXVjjsDGlJBwa4TnlDQ9an6EOMA0+YJSlHiNIgCyqVehq\nPbZaX7hpzsPPXzzD+dyYAtup0ybIgV0HfTrHSFkpOALnFSZAPeVt52RZZjAPIergPWicxlJR9Ap8\njCldSV3bXDAbrER/q1JrnGOrbfrgjIJVG20dJpFbRxJArZu1sCLGiDRwIsX143OutelR3Y6npRtA\nVOkk2GdblCGD4zLIpIPcMmOeukFD3jdI6/K54K0+5GdX06Q253vcqxsOKnzEVFwHehJ1nox1w+E9\nuF3zHCYzehXJSEkQV52LDGrE9LE27VwFagNVenpZFos8zzjbvSelyyAXqfVEk/UOPRJDIaTmTp5L\nMeJ8+Z6M7FPjD7pghWlqLe8Ga7Bc3GNZGFOb3OAnLLNaQkW7WZgAR24wR4WlXyV2R1yhZBepMuPD\nwIwncha+o/QLvxYMyRamajs/Tma/ebR2IFKQpZ50oLZrBxjXdao81NYr7qmMLxxe1utjEIFrNKXD\n4WAteuc8iuKmvK8YF1vouiOJjBkveoolImDHuL9vjr9S8PZtRUinmFDaSp1zxL6ecf/lHwEAnt8/\nx9e//U2bZ0JsVBUXGC9fVSyautYsywzVx6/c0V4qNyJu48Z1t202IQwKI/HddRpMswbT4rpzhKnN\nUfDOaiJKh7JNLSejf8U9mlFpGvBeysBSTJX33pQg3ORMWz1MdZMQ1bcHwSuH1PX3TJPH4ahNk4bs\n00bROpmahwtukK92IEcG51m3zXiyl3U1bKH3oZvpOlVBYMxtYbq5e4bbJpV9PB5NBWM7b1UA8qxo\nfODmpmmVIYG5fpcPAcuhbmiHw4Jpckii6WKfsyFLrek+AWIu0+OlLdDuhsig9v/j16unlPBpPI2n\n8dMZf6AIq66LDw+PeNfAom/fvMPbN7XbFCPg3WQGF2/f3OLFq1p0v3u+4HBsxeVDQB46gzkJtqYc\ncDqdDWF+PN7artq1m7p8CgFWfKbSVR+r50GXAMbQ/XDERouur++RSiFBNHArLCQnh6EzWKM1g4Uw\ngE+RPkWPEA3iIHbczo/OPZ0LR40I3KOErmKRqRf6BUMTohTc3N7g+YsaFT18OOFNU5MEHKa22+eY\n8PbNW7x8XmENX7x8gctaIQ9biuD2uhdfvsCrL+prNLV0geFS07Yaos1cOgyBm1yypXQSjFhOjrow\nKTmLloGqM6bpZrVAb+/3zjq4zJVUrVEQUfcIBHqay6nrYWn0Z76W6N5+zrvBaKMSfPt1NsB0vIdX\nOMAUMM0KqyjIezIV0AKpUSgANzFYjYk9A9R1vPYUsZ5VcSNaR5AxuFO1uY2pYF7q4+VwwF3r3N7c\nHhEb7OeyrhAhg/zkJIAxP3oKHZYFd89rh/r22Q3IVWK3DkuBqZ8LauUYzZZKgnUnq7y51liSNbGu\nG3a/f/xBU8LDcsS6awfuAY+n2qV4//aMuAtCQ9E+f36DL061xf7zeI/nzZV2Cvd1ctRxRToSN8Zi\nNYmcexqgjj1jvUFK13Sq3br6qC5PpVNZAKhagAwSIY56bUzaQnjtDKK0hHSlI187gFqI0W8Y/g26\nqIwLkWDfm9IcC7IBhjqFiLmmS6VE+2z5BLG6EtHr4zAtePHqlbnenE5nmBEqYClzAfD63QPmX1Xs\n1c+++hJ//Zd/0T6vgBv+x08T3r2pC975fV2w9jVaqsbEpj/uOkwHRBmEoT1P0qRbFJ/W63lW98oC\nSglFz5PnoSvcO4bUCNe2SUzeUmXHHnFXyldqBGogSyWlTybnzNa9m+epyxm1eTJX6eE3HG+WDiFw\nsPeUwrismy0cRbJhDJ3zttDWNG/Qr8rFLLUgbN21lFI/EE1NyXcnI+71LOc9dmpCkAnYY7fiylRw\nUX13qqUbAJhvZixHXVDdVc045zRgL/uFpRZm5o5Vug6+wlH0PVcN9B85nlLCp/E0nsZPZvwjIixF\ndPfF/fcukMO2Pk0LUOoq/+HNA/7ub6rO09/98mucH1fc3NTi5B//yVfYY9WummZY5HV//xzLEhBM\nkD9bAX7y3UCzqmQme1yPN0AR7QWErB6F1F1VCtdd268FgAAAHcFJREFUTX0Jx/RMhOwHM8G2BWng\nRpObBVuEkKQY8JSk/WcIsEb+oIwFeH3UjmNr2lbsXE9v6kG1eXAIwRtavH+D/Qh7qKXO+TDj/sW9\n7W7Hwy38VzUlXy8bHh5qlLTlhJgJv/7NNwCqCOAffVEt3g+HYMai3z3u+E0jPR+PRyyLx7buEFNF\nBSgM3S1jENToS9r5KiV33FjhHpFiSO8FV6DJXDrWrMCPWnQIjgxf5oMbSgQObHivISKvAt3GexQp\ndh4O82TRFhNXpkU7B+wc5oZ5Km6yRk6S1F2GSsT5fMb22BoV3hmmbUKP2n3wlZKq0UxKJqMEjCYn\nFT0PoBf/2dl94H2AD0322nuLyLeYsKfczTqcwxq10O7g2zk73Byx3HTObEq7CTLGmAYBwXE1aKmv\nAnEzrqrqo1uVSpNfoah/YHzmgkW/52/Ds1crmDPYwL7GaxcPDd1FEPcNp1Y3eP9+wfv3dfF6+/Zg\nrfftiwwfyFqmIrAO2nSYDHhaEd+6UGgXx4NbezpXRSUADdCsNwW3vNwmsCsYgAcE+wAVSFLagtXT\nFTuRpQM4ierlwqNkrKWBXS2gajMpcrTWyl6//hYAMC83FnIzs4nEOefqYmafW4yCIzwoa+YE35DL\nL18ccP/8gLzV1z3HfXWQRtW5Op0qCPRXv/4NPpzOZvP17bev8aF1Ez33TlYUxt66bmFesKCKKg4N\n06Fd3uEd3jn40M1FCcP00zCXKPYP7AJM1x21TmLagQ7gWF/nCSjM9vuZCdJudu8K0JQWYmJwo6bU\nuhBsYdr3aIBXzw7O0OKCktFT9AKI2Y71VNbBdenpuGLbkmllueSNoH0sCyZ1dw4OqQikQW+kFFuY\nAB6KlV2sUGlBTL46RwOY5wXzPCxYqtcVd2wxm3CeODagp/MMbr89LL1LOC0B2BNyW+T3feuS3BAo\nWldK6x4qPhmdA0NEzZKt1Qz1/ne9FPFD4yklfBpP42n8ZMbnp4S6k4ExRoRkWKYaIJAVGjtBebtc\nLAV5cX+Pv/7Lfw4AeHZ7j2+++RYPjaPmg8ND6/h9+81r0/R5+eGEEGaTcnEhWXduj5uFqyVnkwLW\nzdp7B8pKBXDdTYQMGgIvBE8ezgipg4xuyTXdAEDsrbCLUuBEANVzqu6iAFA9/TRk52LpBtCAm6Zp\n4nq4xdSBhW2yFK9WENBRML2Dycw1nbDmwQCLKWXodBJetK7RqxfPsa1nfPvb1wCAhw+Pdtz3d/d4\n0bqHUgri3/3DldyxcsNqMbXNzxAa6e8qJavJCriQ6S/R/9fetS3JcSPXkwmgqi8zJCVqtftgr//B\n/x9+cNhPdvgbvI7wSityqOFcu6sApB+AzERTXomUwrGeiMoIiXPp7qmurgIy81ySycru1iRnQyUh\nMI0ZiKCkLBm87qcYUUrGqmapKHYt5gzLlnIvB8f9OxgCGSBdlsRL6UhAK62IZGhc+9SckrMhdECT\nuEhxdExRwtacVusUF8Cva0EYMuHz+eyWQPUKkwEYEVyc5BsGHzMGG9+LmQ1pVt5X5IS5Z1iH3R77\nTvxkcuH7mgvWUtyPiny6EyU236xpP2E+tGNKM2OtZLMwl3WxASYijpACtQ+50JmKToBl+Aj7Jlvr\nWWhy59pfii9csAYWIrkPd1us9AIUMAdL63P2sVwkwLGT6P7wh4Svv/4GAPDHPz7hz9/9Bd9/33og\n9w/3eOpePY+Pz7i/bzD63d0jjlevcOwseIpslrXnfLLptVLF7IL1OCIH9xsTNzxruKBSFAaBKNA9\nk/StuzcWk6vTWZp1MuuFWwTcU3YWNs1hG7ZQDXUk+BBMqQ7tXgwSlfZz1UCGeY+Y28V0ej67FisE\npK5t079l9svFvcLjPGG/P+qP8f6H7/Hdd91sb4VZ+d7f3lnf5vp4wHE34/FRtXaXI9w05BPdJdAW\nNJuGDPqEITg2FKpTDALbYsYRXkIwTFvHIljXYC+3lupaRAFKUdIrt2nNemMMjg+jfpMC24JF5tbR\n+21C1hZooux+LoVaea73AIdh8SnmhU4UoMqHWgtijNj3DbjkgnnXe7BTsn4Wc0SAAIs+D3a9TSEY\nhSME8mtAdYjslJTjfu+DLFCMWJ1L7W4J3hbRUpljROrI5e64c5Swl+fjgGIrhwf7amKCrKuV8iFG\no98IOZVE4J7yYfo/W7A8LttsHvKTn7s5/zRPiF3Iu9+z9WNevbnGvEvYH1rmdPPhA24+NvfKUs7Q\nO6HkirVUF1YyMM3uKKrq9lyy+35Yk5VNIIq1eh+giDfGuS1dYnR0d4Ek4sFgcZhwUgVUCKSNtVra\nf0BfcUZ+j9/TTZ5ymZG0b9xRVUOvC44JYeq9n1JRNIskHi7MzjnTQ+jvHQAkVzzcdcrBesLt7Y+o\ny+CG0B93Op3xrjfQd3//d7i+2uHx6a6fT6d6NDXOwBUxvo3zhGwju3hH/h0HndOni61LcFJimw8Y\nojtklLWAaXGaA2D9t1qqjyGrbRqlwxl8sWBhbPoOJn9AdSpJre5dL+TmkT2roCHjGhdAUk0Xje+V\nEBPbppFSMjna7rCzvqqgNfSX1Rvclk1HAqkcKpBleDpZqo2y75lomqwSqKX4fkGMZlSpYACcExUC\n0k7Z7QfM/etSzsjFhdGlFHs+cTCOorq6qo1UInJHCya7B7nCVh+On78MbT2sLbbY4sXEF2dYjmZd\n5v8+Y6/vquQ/V1g3RLF0FuJJ0CHMwDevQV3kPO8iwtR+efdwa2LWSnUoSIAQk03QORwPhrKs58X+\nvg5xDsy2I5darN8VyItA6rvrhb267myoRjpkaZpGoLGEa3XfpgaD96dzNauTJnYeUMcL7R/j0+wE\ngO1ga1XiaMWulxBcgdzJhKXgYqcfSzMe5j9CCPd3Df2rD81qxaxiqtv1cmAfJVUz0jyYGlZ41jBQ\nVkbxtYmyQxpIg8NxgJ3caaiy90MUfk8pWIaVEhkinIP717e3y9CZiFIZg5chBDKYy7l/v5CTGDlE\nQD2zamt1aN81gKAOiKW6PQ2o0Qs0e2UedY/cfOc/+Twa5WCybHiaElKnHTBHy/rXJaNUt6gZaS+l\nFrOGDuylmP693Txh3++JebcD9fNyOi3mM38+L8g5o5qti5fK826PV68aaXt/PBpJezk/4/R0MuOB\nOlBBIMU0hlXaNafk01qrfdbTlLDv1CVOEafctcT8+XnTrzfwa0wi/4VWBiKXrYp22ACAnBfEqYtK\np4ipl4fTtMP1qyPmLlNgZpxy4x4JZZMs5LxiPZ9wPquLpRuXxTgwfEO0RaeIQ68uuPWx4ixOY5C+\nQI0Tau3GqC4gaM9he2dF3NtbpDq/BHUoO2S4OdEvfP2m+N8hh+ilNbds0MG8S4D2Z06rfxZS+41z\nsdIO72H48dB7U0mPxsj+16/iFBAKW39Gah2GOgyLFNMnCz0uhMeljjPtyMfagy8WrFEa0sTuClIE\nb9SHAEnJTO1iIfNzquT9Hu4jw5SNzhwGH3eXmNThmFU4rjKgBhRob6r642qT7OjNGII7J1QBdF9e\nc8Zy1vqcm+d8UN+rCWnq/Vh25ULJDSjR0i4GP1/mq9bPjwwLJgBMuxn7o5pW+jWRa8WiI/HWglrJ\nxsbXQaY0xYh955Pt0s6G2tZKWNdijXZgHIZbUUg/v3bm9b4sNTtYMkVcX7VyOO4npKf22o99RsDn\nxFYSbrHFFi8mfkXT3VN3DboAjtsOYKO/B+pYCN4ClSH74EiY44zr2iD3c1lxWhsySEEMcp2mBAJZ\nOQdiKxfn3Yx538qlZUmGQKx9skeFkzZl7H7XCg56TB2V++S99Ddpb1oGrKyiohKhKDBAFVnf75hh\nQVE/fWUy1PHivA7saXR2vfTfzXPCPDWEaTmviE99+GU+YZ6TAQQAmRVLldVeu6Ja9sZgUA2mkWx0\nFH/c/qrpNznNeLq9Aamgu1CzmNZD7h3yIhlzR4Df/u4tzg+33a/ppyUjMxlhkhAakupppTWSKwmq\nduAHK9YYE6QSVvWqJ4GSBdrnomUfgXjIsGgs08iyTcIAFDCDwTgeWyaQT64eEIFljBzqhW6Rhwb8\n+LnW02qWNjXXjqZq1hhtwo8wW5bdEmGxrGqakiF/IThbH0QKbppgPKRgDqcUPYtKU0S02Qg7pFhR\nsp+PMfvV8xQ4OUpLfeqOV6mDJ74PtQU6INF7MSFOSFHtcKJ9nULEotZP4tnuL8VvEj/bfSVyceuN\nJvvjWPJaxabFnpaC0O2N590JIczOtg1sPuZCzmje72dMc3B3hRBw7CaAr69f4dWrvuAtZ7O/VbSn\nVJjAtZaMXHVEkffFWGrr4+h8PzCUxbNm96VKh9mlG7WiIOC5oyctI+6pPWDco2bxK05lEHG0DfBx\nTxdi7C4B6a9NVWwidohs+XHVwwzKaHckFUyowwuGvlBIFtSqKTzQr0AAwPH6iG9+1ygnzAEPd8/d\nJx8gST4sIQaoV9y8P+Ltt20U2Fdfv8H3D7eIKYwtH5/8DBqunUZ70THqy3kd1q5gC14ILpcJMYGo\n2s1OVGxjoMEKmzg1yN3Gyff3icueE8jeui2sk5ZqhZHVXPGTdsflIszeWmgQMIDGtVIP9sjNA17Z\n8lT9+YGD0Ws4EEIk6xOnFM0HTUp1blopiN37Si+GSmRoY4gzmHRS9YRoLiLce6p6nof3JI7qMfvm\nrXxKMQB8MOYT71ObRbi6N6BtGkDbaNLgeRejMv5HSdnPx1YSbrHFFi8mfpu9jDWUq8OEkIvyaSyJ\nHh7ubarseakm0syFQLSY9e5aijs+pojlrCWgoJQMor4yixgKt9slHDrJbbeb7OdLz+iKVNS+ki95\nwaokU/amKqjvKIoeCZmhPwLMqiZM0VjJTAGoZBlNHQpKkE/k0f1Mhr81knA1xkyV+gSdrNlNAaYu\n9g4x2PGACAjOVi7FLWuZHbGtw99szPiA0FP3NCfs+/n79vdvcdVHi91++BGnp7OhRV+9+RpvXjcU\naZqTp/gpQMcN3d3+2N+MeFOaL4XH+nUtgpoLJCsvDobQrcSgXhI2/WF7SKSW6Vg5K8HKqUCwTCJw\nAAcva6j9QTtnhhLCvcUAtoG97WE8TFES8EWJP4IzbFCFiLhb6OmEpSNr03ECCXzAawxIChTFiKy6\nU0jLKPvfitEniOdSDLENoUKiCqT76SsZa9YpQU46HsmdIo1w6tcsMJtg2hv8PBLCtVmiFULxyUeA\n87hEz4uqNXK58J4j8hJ6JE5/bnz5gqUf2E+QwJ95it4wlfHcewI3t3dQKUhK9wghGsS7288ufQnJ\nCJKJQrtBlKoQYZa11+VoEp7wnqyHpjPaKorNuzuvZyszwcN60ekFNNYxakNMAQwVG0f3MYqpTwt2\nuoIRQikOCza1m4r8Q7KknPwDx8CuFzQo/flR6QvVSgOUat22CjHHBqBNHh7HT7mXly8g33z1DY7T\nFWJfADkGhC5VSVMwb6vv/vw98lrtQosxmTSqjblqr/z4+ISPD41c+vH+Dm/fXAEEo5qkFDF3F4CU\nkt8UtSBzhIjD5Yr+BRaYVKAy1GBfOuFcBlhy7KPasYYJMSQEVscInyIuUm1hlFycSNyRzdNJR3vF\nwdPspxe99hdLyfbZ1kEMX0o2VUUr8X1zCiHYxkcpWI+Vub1u7b3CFBKmfm+sS7bNnCk4Ih3s8K3H\nezqdrAzPS7FRXlIbTcKJn4y598imGN2tBM7uZAptI9D3KEM7gdj8A0Rh7v4SJRfrlUkd3NvFLbOV\ngf858eW0Bv3wKv4Kyxnd/L+v3qXifFaZTMXamdWPd8+4u2uN9ZwLOEZzR/z67RvMff5bmgNev9ZR\n680EX90Npxhs97w67vD1V63v9eOPV27X0mt4CuIawVrbfDigLRAGGZfG063eELYdQbyeb/0j7VPV\nC25MHQlmxPCmfR9tpdovJpj8Y0jKIDRkZe0/VQQ8P5+dtd0dL4HW9w7w3ggTDxmd9xdFgNR32evD\nEdf7K+PVPK3POD0pT+eEx49t8Tk9n0BMNrr9w4db3D882Gvb36nFhmQWFLx903hklrdccJb860I/\nt736iSHiAUho/TdfiMfOhg+uiJwQiS2bkwofLV+rcc1KKW3RAszp4qG/x8PuyhZa+d92aaPKFPsg\nSymDe0exBYul9acss+OhLxcYPPQdc63GGUtxwpxaT+0JJ+vLigRM5A6oAFBzxrm7k95+aH1EAFjP\nZ+sT1lpBzGpW0fSDStOpFbnPiFzPC1hdSUpxjmX/DOw6HXp72s+22Zq1eCYMd+1t7qNfnmFtPawt\nttjixcQXZVhtffVa1UI+JVvCYBeRgmJmeq4xOp0W3N0+9K/PXdXenn887K1kmKdkaXNKofWwrNYv\nhq4FFmOBX13tbJdQi9728kqaHPJZqcMSLyOHsdEcrFKrGASErmOTjIqCtZvnVWSDmsOAvAFtJ/Kk\nwYmWMvzf+mj6E/Jhlo/PzxeCZ0X8Gc3zSieARSLrFVaBU0wIVvrcvHuPm/zOyodzzT5MtMJ23DYy\nTSA6DFQKup3TxXETAdV8vMZsu79OcDJmTMFGvkOAuFaEXjpyEQSFomi0A/bpzmvOKNmtdyFsWXcz\n0tNyvdEOFG0r8J1/zdk8uHLOPnEJglwybm6ai0W5FlDP8KcpGqNeQ+cLyFgG1WLlZsnZrn90yoCX\nhX7fyJgLW6egX7shIamjgQiWRbPB1XR42p9aTic8P3fr8Y8fcZjVt72YQwkgiByQupqkciN1A230\n2NIVFKfTEyDt+ctacF4LsrU72MeNgcF2Mbdy0y3dxvLB+3etxeJff278Rk938X+HwQ2j6BfkbOV5\nnuxiTxyHxUDaSwwQrzZO4wCtSikQcRGn1NWg4FqKQbiHXcTSod+kVIQQoIV+EJhYmWsBmZ1JAw+K\nupQSmzyCyKUyzXVBpQgZVQqK0iRQjBUOqsPi2oZbKHeFAKx6sRKs6Tv+H+iWKv1cnvMZtUsgEgcz\n+g90RkHBrveW7uudaZJCGCxypFgZ9JAfQdU3kEpD+Qay3giFduz6eQhFiJUhBLLhD2ybgN6Qu92M\n3b77g8/pwpnAb1qGCBlIIAKcSUs1Qb+PUNYVZ/VgR4WKbNv3jGj9t4R59gnKF4olKT6qvrjtzrgZ\nBQgqkQ1E2c+zLbq1Ds4eVVCFrGQSiMH3VKqVgbVm2wDP5xNkmkDd/iWjXlj16PWVYmjUFC0l2aei\nH49X9px1XW3wijmnsGA59b7teoWP51baSxGcuzQnn1bUdbV2R6li1/zDA+H9+2DvUc/l8/MTHp+e\nXZBdqjX0WxO/f/bMiHGyhGM5PY1N4oEdL34NXK5pPxtbSbjFFlu8mPgNGZbAc1eBU2AJbR10pEGz\njP1uQu2IwZtX1zj1dGuaJsTAeHXUSdCTMXuZBDT4Q1cpBrwVWlFZU2/PlOYUDIHUB//w5xub2vuX\n779H6XQHlmWwDSeAgtm5VBG45kyMsEdTMFb4FAOQz0BHIxOx7YxNXyZ2ihrc699H9SNSsh0GtBAK\nSzN2RqEQrL3UbSWMWkwHkARw/zinMPl5GUS+FeyoGyKIA6YLRKg3q2OwjIOosZK5ZwWUDuCeyYHJ\nTAZzLdbQT0yQ8zscjwejRzQyZC8VGZ6phJY9TEXRyoiQ2vk7nxYsHVVe1hWlTypmZoTkqCgnMivg\n3TzbINc2+WbIVBBA0PaCOOQvhElpIIFx9/GEXdKpOYSiGjyqhsY1wia5dQ85IFUJlm1K8ev/fH5G\nlWoWxLVWZD22MoAjiI0ew56Fqjf71atrm5T1cPeI5allWOp59u2339rEn2ZP7tSFXdcIHo9HzGln\nHlprzXh87pkaijHz758fcc7tMaenZzw+nbBmzeQi7B4fmuYVFVRhKHyakls/x2Soey4V516iKij3\nOfErxM/tXxUK9+8GIa6Wg2OXS9NFwTy3x11fH8yV8Hy9ACAcDo2WME/JxKMixWgNkIyC4kh3cXSy\n1gLpj2OIcWeUd/PP//RvNnjinE/4+PARQPfu1n5b06dYP6CCzRurlMVusoenJ9zctOefloqb97d4\nfmgfOBO8vEEdpEg/RZjschRHAkdGkEAAcUPAsuRuIoculdF+YusuGr8nTbZQVha7noQDYmznmIQR\nOdpND/ESqX2ULlsJYcLUFQVxvwf1C7AGssW95OJ0jFzxfH6H3W7GlIbyUakMg1CaECAE7LjdTLkI\nOPYFP/qk5bwW9zWnRsOIfThJitH+zpSScZtibIJin5btrYZad4bkHQ57MwBkBNx9vMGiBpCnky1E\nHNlKszZCjF0GQz6jsG18igYzuPtXIXBngLsUrKoLhm5oaKUecwAn7UMKevWPCcBhaFloSb4uGXgH\nvHnzxpQIx/3BXjNQMD7Ymzetx6UzEZdlwbsP79vjEiF1j7ndfjfMYYyoFIA+F2E5O62GfB/uPevV\neqNXx4O5RxwOB6PECLHRLNYvoDVsJeEWW2zxYuLXO47SkG1d/JwutFQt8+op8fJkjcWYKg7HLn6c\nmrZMUQsgo3TUTUDGdCfuTU/B8NpK1hPjVrEUxL5zKVHwv/70HRbNwHYCCt2tk3nY7Qltiq36b3km\nwBwhXUj9w7v3+Nd/+ff+/OYP/+H2Yz/2imlWlGwkyv313iKJaYjbrmVZbMud5tin/awZz33SdaGI\nh05uPOXmnxQ6InT1+rXrwzi0nREAwoSYWobF1Ca26PvLA8Fx9KSSIsjEkL6zPp0FOZ/7O3X+TwzJ\nJvCsy4oZbaCoGaGSI1mAO7ZyAIS9Yc3sqFGMAVGRrezIrITGNZsumut6HGQIp1B7jioCUnWxcnsD\n/c0fJi+B5VJnl2uxTCLCLWQ4RXAMJhS+RMkJMamT6B61s8gbYTY4mXaa+tDUloFqeShWsfTMDu4x\nn5Ijrykl41mpdczt/Z0z2uFASOI0MNulzThQJH6/w7d/+H3/XTGEr7HtFYWPiHHC1K+fnMV0qTyc\nr1oK1nIyi/LdNA3ocLJre11Xm9p+On2+lvA3SnPI/rlctHhgB1fr9zyffoQ2AWpdwIrkTYRShjHu\nWFCNlcueMkpDdEaSmpZSLcvuz6FitbsuWJUKOHqpZtOFR+cGCiBE7wUlRlZHCalY+gX+9HTCf97+\ndz8Gaja/Wp7M7h2V13UQlioSqqUW2VAMCLxnAsZiEqXmiaR0jfsCPDyu/XfAqbbnnEqDzqv6ekcx\nJKlSgHSmt/CE0ntRNQQQyGHq6CVJrQMC1wG0j0sreR9Od0bhEACpDzl4dRWBPvyh5IgZDRX2ke++\nXMv4LfVNztBU98tgJiRlN1Z3txBuv9NzzkN7oimRRspKge9IYmxyIgL3fgoFL6dR24KgRnORJz82\nbr74ABCn1BYbPbw8jGoLjIna4w7HoyHDMTTW/K6X19NuMnXHOMSjeddXVD2fNZtXGTFj6oLnEBwF\nVevmH368sRmFt3cP9nUKwZnla0YIbKjyNM94/abJrXLNtjEQy6CYaBSOqffB5l1AUA/6wI7G1oJS\n9oZmN0+y8bMfWjg/Nc77xdhKwi222OLFxJc33S27GX4olytfxcg/qtDM5+nhDqHvDjmvpr9jChD2\ntLrZDmtj0ZuZRTIuNEhloP0TwcXYxayYayfy1HC246hEGEXIbEKsZuqvHCRhgignCNmnttShHiYC\nBbd+bk3un9MaXJYOQJu6oxKLlCZw18xxaeX10tHUpQQ8Pnf5Cx3wXJtn01Nl5LLa9OMUI84d3cyB\nUFl38QhmteRt5S+Ri7it7OBoHzBTK0nK/U0/tevAySLT6dXCViJQFCADOkHp01AWVXtw/0dRkkCI\n2qQWF8iqfTWA5j0WHFhpHLl+zqsDPsz9bYzyVxqul4G75Lqr5iL69dtv+t9l5KrTZlYbytDKGxe9\nZwxTleCSm+PVwUrPEBggNiIwGH6tyGgrRF3iYqm/l6LUKgQAYGEDJUIHKiq7FlMiWzM8wyfsVDRb\nGxVb12UBuhSr1NWqEpC49Xa32jbfMArQuz4EtvFpzG2CEQ3NfpfpwHzjlrXguRNU1/pz98tl0Jek\nZUT0DsCfPvsJ/3/iHwH8x9/6IH5lvNRjf6nHDbzcY3+pxw0A/yAiv/ulB33RgrXFFlts8beMrYe1\nxRZbvJjYFqwtttjixcS2YG2xxRYvJrYFa4sttngxsS1YW2yxxYuJbcHaYostXkxsC9YWW2zxYmJb\nsLbYYosXE9uCtcUWW7yY+B8z4nOOt1L25AAAAABJRU5ErkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "idx = np.random.randint(0, trainset['X'].shape[3], size=36)\n", + "fig, axes = plt.subplots(6, 6, sharex=True, sharey=True, figsize=(5,5),)\n", + "for ii, ax in zip(idx, axes.flatten()):\n", + " ax.imshow(trainset['X'][:,:,:,ii], aspect='equal')\n", + " ax.xaxis.set_visible(False)\n", + " ax.yaxis.set_visible(False)\n", + "plt.subplots_adjust(wspace=0, hspace=0)" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def scale(x, feature_range=(-1, 1)):\n", + " # scale to (0, 1)\n", + " x = ((x - x.min())/(255 - x.min()))\n", + " \n", + " # scale to feature_range\n", + " min, max = feature_range\n", + " x = x * (max - min) + min\n", + " return x" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "class Dataset:\n", + " def __init__(self, train, test, val_frac=0.5, shuffle=True, scale_func=None):\n", + " split_idx = int(len(test['y'])*(1 - val_frac))\n", + " self.test_x, self.valid_x = test['X'][:,:,:,:split_idx], test['X'][:,:,:,split_idx:]\n", + " self.test_y, self.valid_y = test['y'][:split_idx], test['y'][split_idx:]\n", + " self.train_x, self.train_y = train['X'], train['y']\n", + " # The SVHN dataset comes with lots of labels, but for the purpose of this exercise,\n", + " # we will pretend that there are only 1000.\n", + " # We use this mask to say which labels we will allow ourselves to use.\n", + " self.label_mask = np.zeros_like(self.train_y)\n", + " self.label_mask[0:1000] = 1\n", + " \n", + " self.train_x = np.rollaxis(self.train_x, 3)\n", + " self.valid_x = np.rollaxis(self.valid_x, 3)\n", + " self.test_x = np.rollaxis(self.test_x, 3)\n", + " \n", + " if scale_func is None:\n", + " self.scaler = scale\n", + " else:\n", + " self.scaler = scale_func\n", + " self.train_x = self.scaler(self.train_x)\n", + " self.valid_x = self.scaler(self.valid_x)\n", + " self.test_x = self.scaler(self.test_x)\n", + " self.shuffle = shuffle\n", + " \n", + " def batches(self, batch_size, which_set=\"train\"):\n", + " x_name = which_set + \"_x\"\n", + " y_name = which_set + \"_y\"\n", + " \n", + " num_examples = len(getattr(dataset, y_name))\n", + " if self.shuffle:\n", + " idx = np.arange(num_examples)\n", + " np.random.shuffle(idx)\n", + " setattr(dataset, x_name, getattr(dataset, x_name)[idx])\n", + " setattr(dataset, y_name, getattr(dataset, y_name)[idx])\n", + " if which_set == \"train\":\n", + " dataset.label_mask = dataset.label_mask[idx]\n", + " \n", + " dataset_x = getattr(dataset, x_name)\n", + " dataset_y = getattr(dataset, y_name)\n", + " for ii in range(0, num_examples, batch_size):\n", + " x = dataset_x[ii:ii+batch_size]\n", + " y = dataset_y[ii:ii+batch_size]\n", + " \n", + " if which_set == \"train\":\n", + " # When we use the data for training, we need to include\n", + " # the label mask, so we can pretend we don't have access\n", + " # to some of the labels, as an exercise of our semi-supervised\n", + " # learning ability\n", + " yield x, y, self.label_mask[ii:ii+batch_size]\n", + " else:\n", + " yield x, y" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_inputs(real_dim, z_dim):\n", + " inputs_real = tf.placeholder(tf.float32, (None, *real_dim), name='input_real')\n", + " inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')\n", + " y = tf.placeholder(tf.int32, (None), name='y')\n", + " label_mask = tf.placeholder(tf.int32, (None), name='label_mask')\n", + " \n", + " return inputs_real, inputs_z, y, label_mask" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def generator(z, output_dim, reuse=False, alpha=0.2, training=True, size_mult=128):\n", + " with tf.variable_scope('generator', reuse=reuse):\n", + " # First fully connected layer\n", + " x1 = tf.layers.dense(z, 4 * 4 * size_mult * 4)\n", + " # Reshape it to start the convolutional stack\n", + " x1 = tf.reshape(x1, (-1, 4, 4, size_mult * 4))\n", + " x1 = tf.layers.batch_normalization(x1, training=training)\n", + " x1 = tf.maximum(alpha * x1, x1)\n", + " \n", + " x2 = tf.layers.conv2d_transpose(x1, size_mult * 2, 5, strides=2, padding='same')\n", + " x2 = tf.layers.batch_normalization(x2, training=training)\n", + " x2 = tf.maximum(alpha * x2, x2)\n", + " \n", + " x3 = tf.layers.conv2d_transpose(x2, size_mult, 5, strides=2, padding='same')\n", + " x3 = tf.layers.batch_normalization(x3, training=training)\n", + " x3 = tf.maximum(alpha * x3, x3)\n", + " \n", + " # Output layer\n", + " logits = tf.layers.conv2d_transpose(x3, output_dim, 5, strides=2, padding='same')\n", + " \n", + " out = tf.tanh(logits)\n", + " \n", + " return out" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def discriminator(x, reuse=False, alpha=0.2, drop_rate=0., num_classes=10, size_mult=64):\n", + " with tf.variable_scope('discriminator', reuse=reuse):\n", + " x = tf.layers.dropout(x, rate=drop_rate/2.5)\n", + " \n", + " # Input layer is 32x32x3\n", + " x1 = tf.layers.conv2d(x, size_mult, 3, strides=2, padding='same')\n", + " relu1 = tf.maximum(alpha * x1, x1)\n", + " relu1 = tf.layers.dropout(relu1, rate=drop_rate)\n", + " \n", + " x2 = tf.layers.conv2d(relu1, size_mult, 3, strides=2, padding='same')\n", + " bn2 = tf.layers.batch_normalization(x2, training=True)\n", + " relu2 = tf.maximum(alpha * x2, x2)\n", + " \n", + " \n", + " x3 = tf.layers.conv2d(relu2, size_mult, 3, strides=2, padding='same')\n", + " bn3 = tf.layers.batch_normalization(x3, training=True)\n", + " relu3 = tf.maximum(alpha * bn3, bn3)\n", + " relu3 = tf.layers.dropout(relu3, rate=drop_rate)\n", + " \n", + " x4 = tf.layers.conv2d(relu3, 2 * size_mult, 3, strides=1, padding='same')\n", + " bn4 = tf.layers.batch_normalization(x4, training=True)\n", + " relu4 = tf.maximum(alpha * bn4, bn4)\n", + " \n", + " x5 = tf.layers.conv2d(relu4, 2 * size_mult, 3, strides=1, padding='same')\n", + " bn5 = tf.layers.batch_normalization(x5, training=True)\n", + " relu5 = tf.maximum(alpha * bn5, bn5)\n", + " \n", + " x6 = tf.layers.conv2d(relu5, 2 * size_mult, 3, strides=2, padding='same')\n", + " bn6 = tf.layers.batch_normalization(x6, training=True)\n", + " relu6 = tf.maximum(alpha * bn6, bn6)\n", + " relu6 = tf.layers.dropout(relu6, rate=drop_rate)\n", + " \n", + " x7 = tf.layers.conv2d(relu5, 2 * size_mult, 3, strides=1, padding='valid')\n", + " # Don't use bn on this layer, because bn would set the mean of each feature\n", + " # to the bn mu parameter.\n", + " # This layer is used for the feature matching loss, which only works if\n", + " # the means can be different when the discriminator is run on the data than\n", + " # when the discriminator is run on the generator samples.\n", + " relu7 = tf.maximum(alpha * x7, x7)\n", + " \n", + " # Flatten it by global average pooling\n", + " features = tf.reduce_mean(relu7, (1, 2))\n", + " \n", + " # Set class_logits to be the inputs to a softmax distribution over the different classes\n", + " class_logits = tf.layers.dense(features, num_classes + extra_class)\n", + " \n", + " \n", + " # Set gan_logits such that P(input is real | input) = sigmoid(gan_logits).\n", + " # Keep in mind that class_logits gives you the probability distribution over all the real\n", + " # classes and the fake class. You need to work out how to transform this multiclass softmax\n", + " # distribution into a binary real-vs-fake decision that can be described with a sigmoid.\n", + " # Numerical stability is very important.\n", + " # You'll probably need to use this numerical stability trick:\n", + " # log sum_i exp a_i = m + log sum_i exp(a_i - m).\n", + " # This is numerically stable when m = max_i a_i.\n", + " # (It helps to think about what goes wrong when...\n", + " # 1. One value of a_i is very large\n", + " # 2. All the values of a_i are very negative\n", + " # This trick and this value of m fix both those cases, but the naive implementation and\n", + " # other values of m encounter various problems)\n", + " \n", + " if extra_class:\n", + " real_class_logits, fake_class_logits = tf.split(class_logits, [num_classes, 1], 1)\n", + " assert fake_class_logits.get_shape()[1] == 1, fake_class_logits.get_shape()\n", + " fake_class_logits = tf.squeeze(fake_class_logits)\n", + " else:\n", + " real_class_logits = class_logits\n", + " fake_class_logits = 0.\n", + " \n", + " mx = tf.reduce_max(real_class_logits, 1, keep_dims=True)\n", + " stable_real_class_logits = real_class_logits - mx\n", + "\n", + " gan_logits = tf.log(tf.reduce_sum(tf.exp(stable_real_class_logits), 1)) + tf.squeeze(mx) - fake_class_logits\n", + " \n", + " out = tf.nn.softmax(class_logits)\n", + " \n", + " return out, class_logits, gan_logits, features" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_loss(input_real, input_z, output_dim, y, num_classes, label_mask, alpha=0.2, drop_rate=0.):\n", + " \"\"\"\n", + " Get the loss for the discriminator and generator\n", + " :param input_real: Images from the real dataset\n", + " :param input_z: Z input\n", + " :param output_dim: The number of channels in the output image\n", + " :param y: Integer class labels\n", + " :param num_classes: The number of classes\n", + " :param alpha: The slope of the left half of leaky ReLU activation\n", + " :param drop_rate: The probability of dropping a hidden unit\n", + " :return: A tuple of (discriminator loss, generator loss)\n", + " \"\"\"\n", + " \n", + " \n", + " # These numbers multiply the size of each layer of the generator and the discriminator,\n", + " # respectively. You can reduce them to run your code faster for debugging purposes.\n", + " g_size_mult = 32\n", + " d_size_mult = 64\n", + " \n", + " # Here we run the generator and the discriminator\n", + " g_model = generator(input_z, output_dim, alpha=alpha, size_mult=g_size_mult)\n", + " d_on_data = discriminator(input_real, alpha=alpha, drop_rate=drop_rate, size_mult=d_size_mult)\n", + " d_model_real, class_logits_on_data, gan_logits_on_data, data_features = d_on_data\n", + " d_on_samples = discriminator(g_model, reuse=True, alpha=alpha, drop_rate=drop_rate, size_mult=d_size_mult)\n", + " d_model_fake, class_logits_on_samples, gan_logits_on_samples, sample_features = d_on_samples\n", + " \n", + " \n", + " # Here we compute `d_loss`, the loss for the discriminator.\n", + " # This should combine two different losses:\n", + " # 1. The loss for the GAN problem, where we minimize the cross-entropy for the binary\n", + " # real-vs-fake classification problem.\n", + " # 2. The loss for the SVHN digit classification problem, where we minimize the cross-entropy\n", + " # for the multi-class softmax. For this one we use the labels. Don't forget to ignore\n", + " # use `label_mask` to ignore the examples that we are pretending are unlabeled for the\n", + " # semi-supervised learning problem.\n", + " d_loss_real = tf.reduce_mean(\n", + " tf.nn.sigmoid_cross_entropy_with_logits(logits=gan_logits_on_data,\n", + " labels=tf.ones_like(gan_logits_on_data)))\n", + " d_loss_fake = tf.reduce_mean(\n", + " tf.nn.sigmoid_cross_entropy_with_logits(logits=gan_logits_on_samples,\n", + " labels=tf.zeros_like(gan_logits_on_samples)))\n", + " y = tf.squeeze(y)\n", + " class_cross_entropy = tf.nn.softmax_cross_entropy_with_logits(logits=class_logits_on_data,\n", + " labels=tf.one_hot(y, num_classes + extra_class,\n", + " dtype=tf.float32))\n", + " class_cross_entropy = tf.squeeze(class_cross_entropy)\n", + " label_mask = tf.squeeze(tf.to_float(label_mask))\n", + " d_loss_class = tf.reduce_sum(label_mask * class_cross_entropy) / tf.maximum(1., tf.reduce_sum(label_mask))\n", + " d_loss = d_loss_class + d_loss_real + d_loss_fake\n", + " \n", + " # Here we set `g_loss` to the \"feature matching\" loss invented by Tim Salimans at OpenAI.\n", + " # This loss consists of minimizing the absolute difference between the expected features\n", + " # on the data and the expected features on the generated samples.\n", + " # This loss works better for semi-supervised learning than the tradition GAN losses.\n", + " data_moments = tf.reduce_mean(data_features, axis=0)\n", + " sample_moments = tf.reduce_mean(sample_features, axis=0)\n", + " g_loss = tf.reduce_mean(tf.abs(data_moments - sample_moments))\n", + "\n", + " pred_class = tf.cast(tf.argmax(class_logits_on_data, 1), tf.int32)\n", + " eq = tf.equal(tf.squeeze(y), pred_class)\n", + " correct = tf.reduce_sum(tf.to_float(eq))\n", + " masked_correct = tf.reduce_sum(label_mask * tf.to_float(eq))\n", + " \n", + " return d_loss, g_loss, correct, masked_correct, g_model" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def model_opt(d_loss, g_loss, learning_rate, beta1):\n", + " \"\"\"\n", + " Get optimization operations\n", + " :param d_loss: Discriminator loss Tensor\n", + " :param g_loss: Generator loss Tensor\n", + " :param learning_rate: Learning Rate Placeholder\n", + " :param beta1: The exponential decay rate for the 1st moment in the optimizer\n", + " :return: A tuple of (discriminator training operation, generator training operation)\n", + " \"\"\"\n", + " # Get weights and biases to update. Get them separately for the discriminator and the generator\n", + " t_vars = tf.trainable_variables()\n", + " d_vars = [var for var in t_vars if var.name.startswith('discriminator')]\n", + " g_vars = [var for var in t_vars if var.name.startswith('generator')]\n", + " for t in t_vars:\n", + " assert t in d_vars or t in g_vars\n", + "\n", + " # Minimize both players' costs simultaneously\n", + " d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)\n", + " g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)\n", + " shrink_lr = tf.assign(learning_rate, learning_rate * 0.9)\n", + " \n", + " return d_train_opt, g_train_opt, shrink_lr" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "class GAN:\n", + " \"\"\"\n", + " A GAN model.\n", + " :param real_size: The shape of the real data.\n", + " :param z_size: The number of entries in the z code vector.\n", + " :param learnin_rate: The learning rate to use for Adam.\n", + " :param num_classes: The number of classes to recognize.\n", + " :param alpha: The slope of the left half of the leaky ReLU activation\n", + " :param beta1: The beta1 parameter for Adam.\n", + " \"\"\"\n", + " def __init__(self, real_size, z_size, learning_rate, num_classes=10, alpha=0.2, beta1=0.5):\n", + " tf.reset_default_graph()\n", + " \n", + " self.learning_rate = tf.Variable(learning_rate, trainable=False)\n", + " self.input_real, self.input_z, self.y, self.label_mask = model_inputs(real_size, z_size)\n", + " self.drop_rate = tf.placeholder_with_default(.5, (), \"drop_rate\")\n", + " \n", + " loss_results = model_loss(self.input_real, self.input_z,\n", + " real_size[2], self.y, num_classes, label_mask=self.label_mask,\n", + " alpha=0.2,\n", + " drop_rate=self.drop_rate)\n", + " self.d_loss, self.g_loss, self.correct, self.masked_correct, self.samples = loss_results\n", + " \n", + " self.d_opt, self.g_opt, self.shrink_lr = model_opt(self.d_loss, self.g_loss, self.learning_rate, beta1)" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def view_samples(epoch, samples, nrows, ncols, figsize=(5,5)):\n", + " fig, axes = plt.subplots(figsize=figsize, nrows=nrows, ncols=ncols, \n", + " sharey=True, sharex=True)\n", + " for ax, img in zip(axes.flatten(), samples[epoch]):\n", + " ax.axis('off')\n", + " img = ((img - img.min())*255 / (img.max() - img.min())).astype(np.uint8)\n", + " ax.set_adjustable('box-forced')\n", + " im = ax.imshow(img)\n", + " \n", + " plt.subplots_adjust(wspace=0, hspace=0)\n", + " return fig, axes" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "def train(net, dataset, epochs, batch_size, figsize=(5,5)):\n", + " \n", + " saver = tf.train.Saver()\n", + " sample_z = np.random.normal(0, 1, size=(50, z_size))\n", + "\n", + " samples, train_accuracies, test_accuracies = [], [], []\n", + " steps = 0\n", + "\n", + " with tf.Session() as sess:\n", + " sess.run(tf.global_variables_initializer())\n", + " for e in range(epochs):\n", + " print(\"Epoch\",e)\n", + " \n", + " t1e = time.time()\n", + " num_examples = 0\n", + " num_correct = 0\n", + " for x, y, label_mask in dataset.batches(batch_size):\n", + " assert 'int' in str(y.dtype)\n", + " steps += 1\n", + " num_examples += label_mask.sum()\n", + "\n", + " # Sample random noise for G\n", + " batch_z = np.random.normal(0, 1, size=(batch_size, z_size))\n", + "\n", + " # Run optimizers\n", + " t1 = time.time()\n", + " _, _, correct = sess.run([net.d_opt, net.g_opt, net.masked_correct],\n", + " feed_dict={net.input_real: x, net.input_z: batch_z,\n", + " net.y : y, net.label_mask : label_mask})\n", + " t2 = time.time()\n", + " num_correct += correct\n", + "\n", + " sess.run([net.shrink_lr])\n", + " \n", + " \n", + " train_accuracy = num_correct / float(num_examples)\n", + " \n", + " print(\"\\t\\tClassifier train accuracy: \", train_accuracy)\n", + " \n", + " num_examples = 0\n", + " num_correct = 0\n", + " for x, y in dataset.batches(batch_size, which_set=\"test\"):\n", + " assert 'int' in str(y.dtype)\n", + " num_examples += x.shape[0]\n", + "\n", + " correct, = sess.run([net.correct], feed_dict={net.input_real: x,\n", + " net.y : y,\n", + " net.drop_rate: 0.})\n", + " num_correct += correct\n", + " \n", + " test_accuracy = num_correct / float(num_examples)\n", + " print(\"\\t\\tClassifier test accuracy\", test_accuracy)\n", + " print(\"\\t\\tStep time: \", t2 - t1)\n", + " t2e = time.time()\n", + " print(\"\\t\\tEpoch time: \", t2e - t1e)\n", + " \n", + " \n", + " gen_samples = sess.run(\n", + " net.samples,\n", + " feed_dict={net.input_z: sample_z})\n", + " samples.append(gen_samples)\n", + " _ = view_samples(-1, samples, 5, 10, figsize=figsize)\n", + " plt.show()\n", + " \n", + " \n", + " # Save history of accuracies to view after training\n", + " train_accuracies.append(train_accuracy)\n", + " test_accuracies.append(test_accuracy)\n", + " \n", + "\n", + " saver.save(sess, './checkpoints/generator.ckpt')\n", + "\n", + " with open('samples.pkl', 'wb') as f:\n", + " pkl.dump(samples, f)\n", + " \n", + " return train_accuracies, test_accuracies, samples" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "mkdir: cannot create directory ‘checkpoints’: File exists\r\n" + ] + } + ], + "source": [ + "!mkdir checkpoints" + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "real_size = (32,32,3)\n", + "z_size = 100\n", + "learning_rate = 0.0003\n", + "\n", + "net = GAN(real_size, z_size, learning_rate)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "scrolled": false + }, + "outputs": [], + "source": [ + "dataset = Dataset(trainset, testset)\n", + "\n", + "batch_size = 128\n", + "epochs = 25\n", + "train_accuracies, test_accuracies, samples = train(net, dataset, epochs, batch_size, figsize=(10,5))" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 19, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAEICAYAAACktLTqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzt3Xt0nHd95/H3d24aXa2rr5Isx/FNiXNxRFJCCIEkkJCe\nhLYEEjZbCKReziGFbcq2odsT0nD2NLSFlkO823rBNJRLGqCA23Uw0IVtICSxnNiO73Zky5Ys2dKM\nbtZtbt/94xlZE1myRtaM5vZ9naMz81w083009md+83t+83tEVTHGGJNfXJkuwBhjTOpZuBtjTB6y\ncDfGmDxk4W6MMXnIwt0YY/KQhbsxxuQhC3djjMlDFu4m54jIL0WkT0SKMl2LMdnKwt3kFBFpAt4J\nKHDvAj6vZ6Gey5hUsHA3ueb3gZeBfwQ+OrFSRIpF5Esi0i4iAyLyKxEpjm+7RUReEpF+ETktIh+L\nr/+liDyS8BgfE5FfJSyriHxKRI4Bx+LrvhJ/jEER2S0i70zY3y0ifyYib4rIUHx7g4hsEZEvJR6E\niGwXkT9Kxx/IGLBwN7nn94Fvx3/eJyJL4uv/BrgBuBmoBv4EiInISuAF4KtAHXAdsGcOz/cB4Cag\nOb68K/4Y1cB3gO+JiD++7THgQeD9QAXwcWAEeBZ4UERcACJSC9wR/31j0sLC3eQMEbkFWAk8r6q7\ngTeBj8RD8+PAZ1S1U1WjqvqSqo4DHwF+rqrfVdWwqgZUdS7h/peqGlTVUQBV/Vb8MSKq+iWgCFgX\n3/cR4M9V9Yg69sb3fRUYAG6P7/cA8EtVPTvPP4kxM7JwN7nko8BPVbU3vvyd+LpawI8T9lM1zLA+\nWacTF0TksyJyKN710w8sij//bM/1LPBQ/P5DwD/NoyZjZmUniUxOiPeffwhwi0h3fHURUAksA8aA\n1cDeKb96GrhxhocdBkoSlpdOs8+FaVPj/et/gtMCP6CqMRHpAyThuVYD+6d5nG8B+0XkWmAD8KMZ\najImJazlbnLFB4AoTt/3dfGfDcCLOP3w24Avi8jy+InNt8eHSn4buENEPiQiHhGpEZHr4o+5B/hd\nESkRkSuBT8xSQzkQAXoAj4g8gdO3PuFrwBdEZI04rhGRGgBV7cDpr/8n4AcT3TzGpIuFu8kVHwW+\noaqnVLV74gd4BvhPwOPAGzgBGgS+CLhU9RTOCc4/jq/fA1wbf8y/BULAWZxuk2/PUsNO4CfAUaAd\n59NCYrfNl4HngZ8Cg8DXgeKE7c8CG7EuGbMAxC7WYczCEJFbcbpnVqr9xzNpZi13YxaAiHiBzwBf\ns2A3C8HC3Zg0E5ENQD/Oid+/y3A5pkBYt4wxxuQha7kbY0weytg499raWm1qasrU0xtjTE7avXt3\nr6rWzbZfxsK9qamJ1tbWTD29McbkJBFpT2Y/65Yxxpg8ZOFujDF5yMLdGGPyUFZNHBYOh+no6GBs\nbCzTpSwYv99PfX09Xq8306UYY/JIVoV7R0cH5eXlNDU1ISKz/0KOU1UCgQAdHR2sWrUq0+UYY/JI\nVnXLjI2NUVNTUxDBDiAi1NTUFNQnFWPMwsiqcAcKJtgnFNrxGmMWRlZ1yxhjMktVicaUqCqqEI0p\nMVViMYips/4tyzFn+pIijwu/z02x143XnZo2YyymjEdijIajjIajxGKKyyW4BNwiiAju+LKzXuLr\nce67xKl1Sv0aI+E4lJjGlyeONX7cOnG88d+dON6YcuF3o6oI8ZpcU2oSpyaX6+J6q0t9lPvTe57N\nwj1BIBDg9tudy1x2d3fjdrupq3O+CPbqq6/i8/lmfYyHH36Yxx9/nHXr1s26rzGZEorE6B8JERwJ\nERwO0Tccpm8kRP9IiHB0fvNNed2C3+umOB72fm/CbXyd2yWMhaOMxYN7NOTcjocnw3wsHCVfp766\nfcNirqmvTOtzWLgnqKmpYc8e59rJTz75JGVlZXz2s599yz6qzju6yzV96+Qb3/hG2us0C0NVCUf1\nQtBMBNBoOMpYKMpYJMpoyAmjcDS2IDV53S6KPC58nsRbN0UJyxPrfB4XMVX6h8MER0L0DceDfCTE\n0FjkwmOKQIXfS3Wpj4bqknj4Ol2G7oTWp2uaVqnb5XQrjkdiFwX1xN9sYDTMWNjZPh23SyjxOeHv\n97qpLSui2Oe68KYw8YbgEpm1Zf2WbQkt/cR6Jd7KvtCyTlxO8pPB5O87xxBL/JQzTR0XWvvx5cqS\n2RuK82XhnoTjx49z7733cv311/P666/zs5/9jL/4i7/gtddeY3R0lA9/+MM88cQTANxyyy0888wz\nXH311dTW1vLJT36SF154gZKSEn784x+zePHiDB9NflONf5RPCOJwNEYkqs5tTIlElUjsresm9onE\nYoSiyng8mCKx6ZuOIkyGj9eN3+tCSO/5E0XjLe4I45EY45EYoUjybyo+j4vqUh/1VSVUlThhXlXq\no7LYiydFXSmXEotp/A3R+btO/P28bsn5c0/ueOBnk6wN918eOUfP0HhKH7OuvIjb1l1euB4+fJhv\nfvObtLS0APD0009TXV1NJBLh3e9+Nx/84Adpbm5+y+8MDAzwrne9i6effprHHnuMbdu28fjjj8/7\nOApV4Pw4ZwfHp29Jx5fHwjFiSXyW97oFj9uFxyV43S7cLnHWuVxOq7G86K3dCr7JIC/2OS3lbAik\niTezUDTGeNi5DUVijEeiF4K/qsQJ8VKfO6M1u1xCic9DiS9rYyev2F85SatXr74Q7ADf/e53+frX\nv04kEuHMmTMcPHjwonAvLi7m7rvvBuCGG27gxRdfXNCa88FIKMKR7iEOdQ1xdnByyKhLhGKf60L4\nVpf6KK6c/Gif+HHe53HhcQtel3PrceV+S3GCiFw4ZvyZrsZkk6wN98ttYadLaWnphfvHjh3jK1/5\nCq+++iqVlZU89NBD045VTzwB63a7iUQiF+1jLhaJxjjRO8zBrkFO9o4QU2VxRRHvWlfHqprSrGo5\nG5Otsjbcs9ng4CDl5eVUVFTQ1dXFzp07ueuuuzJdVk5TVboGxjjUNcjRs+cZC0cpK/KwaWUl65dW\nUFdelOkSjckpSYW7iNwFfAVw41zg9+kp21cC24A6IAg8pKodKa41a2zatInm5mbWr1/PypUrecc7\n3pHpknLWwGiYQ12DHOoapH8kjNctXLm4jA3LKmioKsGVZSepjMkVs15DVUTcwFHgTqAD2AU8qKoH\nE/b5HvBvqvqsiLwHeFhV//OlHrelpUWnXqzj0KFDbNiw4bIOJJcV2nH3DYc4ERjm+LnzdPaNAlBf\nVUzz8gquXFxGkced4QqNyV4isltVW2bbL5mW+43AcVVtiz/wc8B9wMGEfZqBx+L3fwH8aG7lmnwW\nicbo6BvlRGCYk73D9I+EAagp83Hz6hrWL6tgUbHNimlMKiUT7iuA0wnLHcBNU/bZC/wuTtfN7wDl\nIlKjqoHEnURkM7AZoLGx8XJrNjlgcCzMyd5hTvQOczo4QjiqeFxCQ3UJmxqraKopZVGJBbox6ZKq\nE6qfBZ4RkY8B/wF0Ahd9HU1VtwJbwemWSdFzmywQjSln+kc5GW+d954PAVBR7OWq5Ytoqi2lvqo4\nZfOOGGMuLZlw7wQaEpbr4+suUNUzOC13RKQM+D1V7U9VkSZ7DY9H2Hu6n32dA4yGorhEWFFVzK1r\nK2iqKaW61GdDFo3JgGTCfRewRkRW4YT6A8BHEncQkVogqKox4HM4I2dMHusZGue1U30c6R4ipsoV\ndWVsWFpOY02JnRA1JgvMGu6qGhGRR4GdOEMht6nqARF5CmhV1e3AbcBfiojidMt8Ko01mwxRVU4G\nRnitvY9TwRG8bmHjikVc11BJVWn6J0IyxiQvqT53Vd0B7Jiy7omE+98Hvp/a0hZeKqb8Bdi2bRvv\nf//7Wbp0adpqXUjhaIzDXUO8dqqP4HCIsiIPt6ypZeOKRc7X3o0xWce+oZogmSl/k7Ft2zY2bdqU\n8+E+PB5hb0c/+zqc/vTFFUXcdfVS1i4pz7oZ8Iwxb2XhnqRnn32WLVu2EAqFuPnmm3nmmWeIxWI8\n/PDD7NmzB1Vl8+bNLFmyhD179vDhD3+Y4uLiObX4s0Xg/Di72/s4HO9PX1VbyqbGKuqriu3kqDE5\nInvD/djP4fzZ1D5m2RJYc8ecf23//v388Ic/5KWXXsLj8bB582aee+45Vq9eTW9vL2+88QYA/f39\nVFZW8tWvfpVnnnmG6667LrX1p1ng/DivnAhy9OwQHpdw9YoKrm+osv50Y3JQ9oZ7Fvn5z3/Orl27\nLkz5Ozo6SkNDA+973/s4cuQIn/70p7nnnnt473vfm+FKL0/g/Divnghy5OwQXreLtzVVs6mximKf\n9acbk6uyN9wvo4WdLqrKxz/+cb7whS9ctG3fvn288MILbNmyhR/84Ads3bo1AxVenuBwiFfaAhdC\nvWVlNTestFA3Jh9kb7hnkTvuuIMPfvCDfOYzn6G2tpZAIMDw8DDFxcX4/X7uv/9+1qxZwyOPPAJA\neXk5Q0NDGa56Zn3DIV45EeBwt9P9csPKKm5YWWVXyDEmj9j/5iRs3LiRz3/+89xxxx3EYjG8Xi9/\n//d/j9vt5hOf+ASqiojwxS9+EYCHH36YRx55JOtOqDqhHuRw96CFujF5btYpf9PFpvydlO7j7h+J\nh3rXEG4XXFNfSUuThboxuSiVU/6aHHa4e5Cd+8/iEriusZKWlVWUFtnLbky+s//leezc4Bg/P3iW\nZZV+7tm4zELdmAKSdfOvZqqbKFPSdbyjoSj/uq8Lv9fNb19jwW5MocmqcPf7/QQCgYIJeFUlEAjg\n9/tT+rixmLLjjS5GxiP89jXLrW/dmAKUVf/r6+vr6ejooKenJ9OlLBi/3099fX1KH/PXb/ZyKjjC\nnc1LWLootW8cxpjckFXh7vV6WbVqVabLyGlHzw7RerKPaxsWcfWKRZkuxxiTIVnVLWPmp2donJ8e\n6GZ5pZ93rV2c6XKMMRmUVLiLyF0ickREjovI49NsbxSRX4jI6yKyT0Ten/pSzaWMhaP8274zFHnc\n3HPNcpuS15gCN2u4i4gb2ALcDTQDD4pI85Td/hx4XlWvx7kM3/9MdaFmZrGY8pP93QyNRbjnmmWU\n2cgYYwpeMi33G4HjqtqmqiHgOeC+KfsoUBG/vwg4k7oSzWxebgtwoneY29bVsbyyONPlGGOyQDJN\nvBXA6YTlDuCmKfs8CfxURP4QKAWyZ0rHPHf83HleORHkquUVbLQTqMaYuFSdUH0Q+EdVrQfeD/yT\niFz02CKyWURaRaS1kIY7pktwOMTOA90sXeTnPesX21WSjDEXJBPunUBDwnJ9fF2iTwDPA6jqbwA/\nUDv1gVR1q6q2qGrLxIWnzeUZj0T5171n8LiEe65ZhsdtA5+MMZOSSYRdwBoRWSUiPpwTptun7HMK\nuB1ARDbghLs1zdNEVdl54Cz9I2Hev3EZFX5vpksyxmSZWcNdVSPAo8BO4BDOqJgDIvKUiNwb3+2P\ngT8Qkb3Ad4GPaaHMIZABr54I8ua589y6tpaG6pJMl2OMyUJJjZlT1R3Ajinrnki4fxB4R2pLM9M5\n0TvMb9oCbFhWznUNlZkuxxiTpayjNoeMhqK8sL+LuvIibt+wxE6gGmNmZOGeQ1rbg4QiMe66aile\nO4FqjLkES4gcMTweYe/pftYvLaemrCjT5RhjspyFe47YdTJINAa/dUVNpksxxuQAC/ccMDQW5o2O\nATYsK6eyxJfpcowxOcDCPQfsOhlEgZus1W6MSZKFe5YbGA2zv3OQq5ZXsKjYvqxkjEmOhXuWe/VE\nEAFuXFWd6VKMMTnEwj2L9Y+EOHhmkKvrF1FuUwwYY+bAwj2LvdwWxO2CG5us1W6MmRsL9ywVHA5x\nuHuQa+orKbUrKxlj5sjCPUu93BbA63bR0lSV6VKMMTnIwj0L9QyNc6R7iOsaKinxWavdGDN3Fu5Z\n6OW2AD6PixtWWqvdGHN5LNyzzLnBMY6fO8+mxir8XnemyzHG5CgL9yzzm7YAfq+b6xttrnZjzOVL\nKtxF5C4ROSIix0Xk8Wm2/62I7In/HBWR/tSXmv+6BkZp6xnmhpXWajfGzM+sZ+tExA1sAe4EOoBd\nIrI9fvUlAFT1jxL2/0Pg+jTUmvd+82aAEp/brrBkjJm3ZFruNwLHVbVNVUPAc8B9l9j/QZzrqJo5\n6OgboT0wQktTFT6P9ZYZY+YnmXF2K4DTCcsdwE3T7SgiK4FVwP+dYftmYDNAY2PjnArNZ6rKb94M\nUFrk5pp6a7VnNVWIjENkFMJjEB6ByBhEQ6AxZ7tq/H4MSLh/YXt8vccPReVQVAH+Rc59dwqmmYjF\nnLrCI06tLo/zuC43uLzOssvjLBf6pRpVYbQPBjpg8AyMD0JxNZTWQmmdc+vJzYvjpHoQ9QPA91U1\nOt1GVd0KbAVoaWnRFD93zjodHKWjb5Tb1tXZ5fMWWjQMoWEnCEPDk/fDo85PZCzhfjzQNXZ5zyUu\nJ0xFAHGeeypvMfgrnMAvqnACf2LZV+qEdWgYwsMQGkm4P6V2TeK/l0g88D2Toe/2OOFWvQqqVkFx\nFjQ2QsMw1O0c08Tfw+O/vDemSAiGzsBApxPmg53O3wvA43PeZPvbIRqZ/B1/xWTQlySEfireiNMo\nmXDvBBoSluvj66bzAPCp+RZVSFSV37T1Uu73sHHFokyXk51iMRjqgsBxGOsHccdbnW5wuaYsT7M+\nEoLQ+ckAD4/Eg/H89AELTsh5ip2w9RY7/5kn7nuKwesHb4kTMt5icPvi4e1KCPH4fSQh1BNEI05L\ncXzIuR1LvN8P/aecML8Ut9cJfV8pFFfBogbwlTjL3lInsGJR5zhjkbf+XFgXhVh4ct3gGeg54jx+\nSbUT8tWroLIx/a3Y8KgT5EPdzms+1A1jA9MctweKFiW8+ZVPfgoqqnDWuX1Oq3wwHuQDHTDcM/nG\nV1IDtWugYjlU1Duvschka34k4Ow/3APDvdB30vlbgbOff5ET9L6y6V93ccX/PU7z72JRg/N8aZRM\nuO8C1ojIKpxQfwD4yNSdRGQ9UAX8JqUV5rn2wAhn+se4fcNiPNZqnxQeg74TTqAH25wwnvgPpQoa\ndf6jTdzGpv2wOEkkHs4lzn/GimVO+CUGoa8kvr10YVplbo8TniWXmBguMu4E/tiA88bk8cfDvGQy\nvFNN1Qm24AnnNejeC527nVCqWD7Zqi9f5ryJXq7I+MVBPto3ub24ynm+FTdA+RLn08XEm1/im2Gw\nzfnbTP204vI4b1jg/J3Kl8PKm6FihfO43uLp6xKZfF1q10yuj8Xiod/71tAf7Ly46222f49r35f5\ncFfViIg8CuwE3MA2VT0gIk8Braq6Pb7rA8Bzqsl8HjTgtNpfejNARbGXq5YXeKtdFUaCTpgHjjut\nLI05LeTqK6DmSud2pv+QE33ZiYE/cespclrb8wmiTPEUOT9pDoK3EIn3OddCw9ucTxiDnU7QB0/A\nyV/BiReduqpWQmWT00JN/ARw4Wfqp4aocxsahtHgZCD7K6B8KSy71rktXzrzaz2dWDQe/EOTbwCh\n807rvGKF050y39ff5YLSGuenbt3s++uU8y2x6OR9j39+tSRBMpXFLS0t2trampHnzhZv9pxn+54z\n3Nm8hKsLsUsmGoGBUxB40wn00fjXI8rqoHq1E+gVK3IzlPNZaMTpl55o2Y8NXrzPRJfExMnbCyd0\n48sefzzElzm3vtKFP44cJSK7VbVltv1sVqoMUVVebgtQWeKleVlFpstJPVWnb3v8/FtbUomtq7F+\nJ+BdHqhqgoYbnUD3F+AbXS7xlcDiDc6PqvPawltPzNobcsZZuGfI6eAo5wbHubN5CS5XDg9Hi0Wd\nLpS+k05/5PjQZIhP7XcUcVpoReWTJ+qqmpyP9lk+8sDMYOI8iMk6Fu4Z0toepLTIzfql5ZkuZe7G\nBpyTWIE3nY/nkZDzMXxirHbFCigqmxzK5yubvLUWnTELwsI9A84NjtEeGOGWNbW5MUJmom882Ob0\nsw73Ouv9FbC42ekfr1qZs1/2MCYfWbhnwO72PnweV3aPax8JOkEebIP+k/G+cbcz1nnZtc7IlZIa\n+4ajMVnKwn2BDYyEOXr2PNc3VmZ25kdVp298tN/pKx+L304sT3xrr7gKlsbDvLIxPeOqjTEpZ+G+\nwF473YcICzdfeyzqfNNxNPjW8J4YqTJBxOkjL65yxvCWLna+rHKpL9gYY7KWhfsCGg1FOdA5wLql\n5ZT7F2B0yFA3HNkBQ2edZZfHmSukuCo+b0jV5LJ/kdPtYozJCxbuC2hvRz/hqKb/2qjRCLT/Ck69\n4nzLr/leZy6LonLrIzemQFi4L5BwNMae0/1cUVdKbVkaR5UMdMDhHc7cIEs3wpW3z+1r3MaYvGDh\nvkAOnBlkNBSlpSlNfdiREJz4f84ET0XlcM2HoGZ1ep7LGJP1LNwXQCymvNbex/JKP8sXpWHCoGAb\nHPmJ8+WiFTfAFe+yMefGFDgL9wVw7Nx5BkbD3Lq2Dklln3d4DN78d+ja54xquf4hqGyY/feMMXnP\nwj3NVJXW9iBVJV5W16Vw5rueo3BspzNDX+NvQdM7nfnBjTEGC/e0S5wgLCWt9tAwHPspnDsMZYth\n4/3OlKnGGJPAwj3NUjZBWDTinCw99ZJz8YNVtzotdhubboyZRlLhLiJ3AV/BuRLT11T16Wn2+RDw\nJKDAXlW96FJ8hSYlE4TFYnB2P5x80bkoQvUVzvDGhbwyjzEm58wa7iLiBrYAdwIdwC4R2a6qBxP2\nWQN8DniHqvaJyOJ0FZxL5jVBmKozpW7bL5xZGMuXwvp7nPnPjTFmFsm03G8EjqtqG4CIPAfcBxxM\n2OcPgC2q2gegqudSXWiumdcEYQMd0PZL6D/tjIK56gNQt96+XWqMSVoy4b4COJ2w3AHcNGWftQAi\n8mucrpsnVfUnUx9IRDYDmwEaGxsvp96ccVkThA33OqHee8y5YtHa9znT61q/ujFmjlJ1QtUDrAFu\nA+qB/xCRjaran7iTqm4FtoJzgewUPXfWmfMEYWODzhXlu/c5l5tbdSvUv82m1zXGXLZkwr0TSPxm\nTH18XaIO4BVVDQMnROQoTtjvSkmVOWbP6SQnCAuPwanfQEcroLCiBVa+3a4Eb4yZt2TCfRewRkRW\n4YT6A8DUkTA/Ah4EviEitTjdNG2pLDRXhKMx9nYkOUHYgX9x5lpf3Ayr3ulMvWuMMSkwa7irakRE\nHgV24vSnb1PVAyLyFNCqqtvj294rIgeBKPDfVDWQzsKz1cQEYbO22ofOQl87rH63M17dGGNSKKk+\nd1XdAeyYsu6JhPsKPBb/KVgTE4QtW+RnReUs0+x2tjrTBSy7bmGKM8YUlMv8Zo2ZzsQEYS1N1Zee\naiA0AmcPwpKN4E3DLJHGmIJn4Z4ic5ogrGsvxCLO9LzGGJMGFu4p0tHnTBA2a6s9FoMzr0HVSiir\nW7gCjTEFxcI9RY6fO4/XLbNPEBY45oxrt1a7MSaNLNxT5GRgmIbqktknCOvcDf4KqFmzMIUZYwqS\nhXsKDIyE6R8J01hdcukdz/c4wx+XbwKX/emNMeljCZMC7cFhAFbWzHIitXM3uDzOfDHGGJNGFu4p\n0B4YodzvoarkEvPIhEfh7BuwpBl8s7TwjTFmnizc5ykWU04FR2iqKb30KJnuN5yrKa1oWbjijDEF\ny8J9nroGxwhFYqysuURrPBZzumQW1UP5koUrzhhTsCzc56k9MIwINFzqZGqwDUb7od5a7caYhWHh\nPk+nAiMsrfBf+mpLna1QVA61axeuMGNMQbNwn4excJTuwTEaL9UlMxyA4AlYfr1dUckYs2As3Ofh\ndHAEVWi61BDIzt1OqC+32R+NMQvHwn0eTgZGKPK6WFoxw8yOkXHn0nmLN9jVlYwxCyqpcBeRu0Tk\niIgcF5HHp9n+MRHpEZE98Z9HUl9qdlFV2gPDNFSV4HLNMASy+w2Ihm0eGWPMgpv1Yh0i4ga2AHfi\nXCt1l4hsV9WDU3b9Z1V9NA01ZqW+kTBDYxFuXDVDf7uq0yVTsdz5McaYBZRMy/1G4LiqtqlqCHgO\nuC+9ZWW/9kB8yoHqGbpbgm0wErRWuzEmI5IJ9xXA6YTljvi6qX5PRPaJyPdFpGG6BxKRzSLSKiKt\nPT09l1Fu9jgVHKGqxMuimaYc6HzN6WdfvGFhCzPGGFJ3QvVfgSZVvQb4GfDsdDup6lZVbVHVlrq6\n3L1QRSQa43RwZOaJwkaCEHzTGSFjwx+NMRmQTLh3Aokt8fr4ugtUNaCq4/HFrwF53RfRNTBGOKoz\nj28/8xogzth2Y4zJgGTCfRewRkRWiYgPeADYnriDiCxLWLwXOJS6ErNPe2AElwj1VcUXb4yEoGsf\n1K1zvpVqjDEZMOtoGVWNiMijwE7ADWxT1QMi8hTQqqrbgU+LyL1ABAgCH0tjzRnXHhxmWaWfIs80\nXS5n9zvj220eGWNMBs0a7gCqugPYMWXdEwn3Pwd8LrWlZaeRUIRzg+O848raizdODH8sXwIV051z\nNsaYhWHfUJ2j9sAIwPRT/Pa3w3CvM2f7peZ2N8aYNLNwn6P2wAjFPjeLy4su3tjRCt5iWNy88IUZ\nY0wCC/c5UFVOBYdprC65+KpLo/0QOO4Mf3Qn1dtljDFpY+E+B73nQwyPR2mcemGOsUF443sgbhv+\naIzJCtbEnINTwfiUA4n97cO9sPc5iI7DNfeDf1GGqjPGmEkW7nNwsneE2jIf5f74lAP9p2H/950W\n+3UP2fVRjTFZw7plkhSOxjjTP0rjxJQDPUedFru3BDb9vgW7MSarWMs9SZ19o0RiysrqEjjzOhzd\nCeVLYeP9diEOY0zWsXBPUntwBI9A/cBuOPVrqFkNzR8Ajy/TpRljzEUs3JN0qneIG0Kv4jl1GpZu\nhHV324yPxpisZeGehKHhEapP7qCp5Bysfw+sepd9A9UYk9XshOpswqMMvPItqkfbKW6+C664zYLd\nGJP1LNwvZWwAXv8WI72n6Vh2J5Vr3p7piowxJinWLTOT8z2w75+JRcbZVXEnNfVrL55ywBhjspS1\n3KcTHoXVdAXgAAANjUlEQVQ93wKUwJr7OedeSlPtDFddMsaYLGThPp3eYxAeg6t+hxNjZQAXzydj\njDFZLKlwF5G7ROSIiBwXkccvsd/viYiKSG5fhqj3KPgroGIFJwPDLK4oosRnPVjGmNwxa7iLiBvY\nAtwNNAMPishFE5aLSDnwGeCVVBe5oCLjEDwBtesYj8bo6h9jZbV9A9UYk1uSabnfCBxX1TZVDQHP\nAfdNs98XgC8CYymsb+EF3oRYBOrW0tE3Skx1+qsuGWNMFksm3FcApxOWO+LrLhCRTUCDqv6fSz2Q\niGwWkVYRae3p6ZlzsQui57AzV0xFPacCI/g8LpYt8me6KmOMmZN5n1AVERfwZeCPZ9tXVbeqaouq\nttTV1c33qVMvGobgm1C7Flwu2gPD1FcV43HbeWdjTG5JJrU6gYaE5fr4ugnlwNXAL0XkJPBbwPac\nPKkabINoBOrWMTASpm8kbKNkjDE5KZlw3wWsEZFVIuIDHgC2T2xU1QFVrVXVJlVtAl4G7lXV1rRU\nnE49R8Drh8pG2i9cdclOphpjcs+s4a6qEeBRYCdwCHheVQ+IyFMicm+6C1ww0QgEjsW7ZNy0B0Yo\n93uoKvFmujJjjJmzpAZvq+oOYMeUdU/MsO9t8y8rA/rbIRKC2nXEYsqp4AjrlpTblAPGmJxkZwon\n9Bx2LrxR1cSp4AihSIymWuuSMcbkJgt3gFjMmXKg5kpwezjcPUiR10WTjW83xuQoC3dwumTCo1C3\nnlAkxvFz51m7uNyGQBpjcpalFzhzybg9UH0Fb/acJxxV1i8rz3RVxhhz2SzcVZ0hkNWrwe3lcPcg\nFcVeVlQWZ7oyY4y5bBbuAx0QGoa6dQyPR2gPjLB+qY2SMcbkNgv3niPgckPNlRw9O4QqrF9qXTLG\nmNxW2OGuCr1HoGoVeIo43D3E4ooiasqKMl2ZMcbMS2GH+1AXjA1C3TqCwyG6B8ZYv7Qi01UZY8y8\nFXa49xwBcUHtGg53DyIC66xLxhiTBwo33CdGyVStRD1+DncN0VhdQlmRXU7PGJP7Cjfch3tgtA9q\n19I1MMbAaNi6ZIwxeaNww73nMIhA7VoOdw/idQurF9tcMsaY/FDA4X4EFtUT9ZZypPs8q+vKKPK4\nM12VMcakRGGG+3AAhnuhbj0nA8OMhaOsX2ZdMsaY/JFUuIvIXSJyRESOi8jj02z/pIi8ISJ7RORX\nItKc+lJTqPeIc1u7lsNdQ5T43HY5PWNMXpk13EXEDWwB7gaagQenCe/vqOpGVb0O+CucC2Znr57D\nULGcMXcpbT3nWbu0HLfLphswxuSPZFruNwLHVbVNVUPAc8B9iTuo6mDCYimgqSsxxUb7YOgs1K3j\n+LnzRGJq0w0YY/JOMoO6VwCnE5Y7gJum7iQinwIeA3zAe6Z7IBHZDGwGaGxsnGutqdFz1LmtW8fh\ng0NUlnhZWuHPTC3GGJMmKTuhqqpbVHU18KfAn8+wz1ZVbVHVlrq6ulQ99dz0HoGyxQxJGR19I6xf\nWmEzQBpj8k4y4d4JNCQs18fXzeQ54APzKSptxgZhoBPq1nOk25kBcoNdlMMYk4eSCfddwBoRWSUi\nPuABYHviDiKyJmHxHuBY6kpMod54WXXrONQ9xLJFfipLfJmtyRhj0mDWPndVjYjIo8BOwA1sU9UD\nIvIU0Kqq24FHReQOIAz0AR9NZ9GXrecwlNbSEyundyjIu9cvznRFxhiTFknNkqWqO4AdU9Y9kXD/\nMymuK/VCwzBwGhrfzpHuIVwirF1SlumqjDEmLQrnG6q9R0EVrVvH4e5BmmpLKPHZDJDGmPxUOOHe\ncxSKK+kIlTM0FrEZII0xea0wwj08Cn0nnbHtZ8/j87i4os5mgDTG5K/CCPfAcdAYkeo1HD07xJWL\ny/C6C+PQjTGFqTASrucIFJVzYnwRoUiMDdYlY4zJc/kf7mMDTst9yVUcOnuesiIP9VXFma7KGGPS\nKv/D/czrAIwtvpaTvcOsW1qOy2aANMbkufwO92gEzuyBmis5OuAiGlPW23QDxpgCkN/h3nPIGSmz\n4gYOdw1RW+ajrqwo01UZY0za5Xe4d+6GkhoGilbQ2T/K+mU2A6QxpjDkb7gPnoHBLqfVfnYIgHV2\nUQ5jTIHI33Dv3A0eH2M1G9jXMUB9VTEVfm+mqzLGmAWRn+EeGoZzh2DJRl48McRIKMqtazN0cRBj\njMmA/Az3rr0Qi9JRsoH9nQPcsLKKJXYpPWNMAcm/cI/FoPM1IotWsvNkhOpSHzddUZ3pqowxZkHl\nX7j3HoXxIVqjVzA0FuaO5iU2j4wxpuAklXoicpeIHBGR4yLy+DTbHxORgyKyT0T+XURWpr7UJHXu\npl9LeLm/kmsbKllRaVMNGGMKz6zhLiJuYAtwN9AMPCgizVN2ex1oUdVrgO8Df5XqQpNyvodoXzsv\njTZQXlzEO1bXZqQMY4zJtGRa7jcCx1W1TVVDwHPAfYk7qOovVHUkvvgyUJ/aMpPUuZvT/SHedF/J\nnRuW4PNYd4wxpjAlk34rgNMJyx3xdTP5BPDCdBtEZLOItIpIa09PT/JVJiM8xlD76+wLr2B94xIa\na0pS+/jGGJNDUtq0FZGHgBbgr6fbrqpbVbVFVVvq6lI77jzatY+2s/0MVl/DO9dYd4wxprAlc4Xo\nTqAhYbk+vu4tROQO4L8D71LV8dSUlyRVTr3xK85KLTdfdxV+r3tBn94YY7JNMi33XcAaEVklIj7g\nAWB74g4icj3wD8C9qnou9WVeWuD0Yc52n6G46W1cUVe20E9vjDFZZ9ZwV9UI8CiwEzgEPK+qB0Tk\nKRG5N77bXwNlwPdEZI+IbJ/h4VIuFlMO7/4l6iulpeXmhXpaY4zJasl0y6CqO4AdU9Y9kXD/jhTX\nlbS9x07gCrZRf/3tFPt9mSrDGGOySk6PFQwOh2h/49dUlfmob7ZWuzHGTMjZcFdV/n1/B0uGD7Ny\n3SbEX5HpkowxJmvkbLjv7RggdGY/q6s8FDe9LdPlGGNMVsnJcB8YDfPrYz006zHqljbAoobZf8kY\nYwpIzoW7qvLzg2cpH++muXwUqW8Buy6qMca8Rc6F+4Ezg5wKjnBr6SmK/CWw5KpMl2SMMVkn58K9\nutTHtYvdrIydhmXXgtuui2qMMVMlNc49myyvLGZ5Xyf0Acuvz3Q5xhiTlXKu5U4sCmf2QPVqKLHL\n5xljzHRyL9x7DkNoGFZsynQlxhiTtXIv3N0+qF0D1VdkuhJjjMlaOdfnTu0a58cYY8yMcq/lbowx\nZlYW7sYYk4cs3I0xJg8lFe4icpeIHBGR4yLy+DTbbxWR10QkIiIfTH2Zxhhj5mLWcBcRN7AFuBto\nBh4UkeYpu50CPgZ8J9UFGmOMmbtkRsvcCBxX1TYAEXkOuA84OLGDqp6Mb4uloUZjjDFzlEy3zArg\ndMJyR3zdnInIZhFpFZHWnp6ey3kIY4wxSVjQE6qqulVVW1S1pa6ubiGf2hhjCkoy3TKdQOLVMOrj\n6+Zl9+7dvSLSfpm/Xgv0zreGHFbIx1/Ixw6Fffx27I6VyfxCMuG+C1gjIqtwQv0B4COXVV4CVb3s\npruItKpqy3xryFWFfPyFfOxQ2Mdvxz63Y5+1W0ZVI8CjwE7gEPC8qh4QkadE5N74E79NRDqA+4F/\nEJEDcy/fGGNMqiQ1t4yq7gB2TFn3RML9XTjdNcYYY7JArn5DdWumC8iwQj7+Qj52KOzjt2OfA1HV\ndBRijDEmg3K15W6MMeYSLNyNMSYP5Vy4zzaJWT4TkZMi8oaI7BGR1kzXk24isk1EzonI/oR11SLy\nMxE5Fr+tymSN6TLDsT8pIp3x13+PiLw/kzWmi4g0iMgvROSgiBwQkc/E1xfKaz/T8c/p9c+pPvf4\nJGZHgTtxpkHYBTyoqgcv+Yt5QkROAi2qWhBf5BCRW4HzwDdV9er4ur8Cgqr6dPzNvUpV/zSTdabD\nDMf+JHBeVf8mk7Wlm4gsA5ap6msiUg7sBj6AMzlhIbz2Mx3/h5jD659rLfcLk5ipagiYmMTM5CFV\n/Q8gOGX1fcCz8fvP4vyjzzszHHtBUNUuVX0tfn8I5/s1Kyic136m45+TXAv3lE1ilqMU+KmI7BaR\nzZkuJkOWqGpX/H43sCSTxWTAoyKyL95tk5fdEolEpAm4HniFAnztpxw/zOH1z7VwL3S3qOomnLn1\nPxX/6F6w1OlTzJ1+xfn7X8Bq4DqgC/hSZstJLxEpA34A/FdVHUzcVgiv/TTHP6fXP9fCPS2TmOUK\nVe2M354DfojTTVVozsb7JCf6Js9luJ4Fo6pnVTWqqjHgf5PHr7+IeHGC7duq+i/x1QXz2k93/HN9\n/XMt3C9MYiYiPpxJzLZnuKYFISKl8ZMriEgp8F5g/6V/Ky9tBz4av/9R4McZrGVBTQRb3O+Qp6+/\niAjwdeCQqn45YVNBvPYzHf9cX/+cGi0DEB/+83eAG9imqv8jwyUtCBG5Aqe1Ds6cQN/J92MXke8C\nt+FMd3oW+DzwI+B5oBFoBz6kqnl34nGGY78N5yO5AieB/5LQB503ROQW4EXgDWDi6m5/htPvXAiv\n/UzH/yBzeP1zLtyNMcbMLte6ZYwxxiTBwt0YY/KQhbsxxuQhC3djjMlDFu7GGJOHLNyNMSYPWbgb\nY0we+v9xluUBL4jnagAAAABJRU5ErkJggg==\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots()\n", + "plt.plot(train_accuracies, label='Train', alpha=0.5)\n", + "plt.plot(test_accuracies, label='Test', alpha=0.5)\n", + "plt.title(\"Accuracy\")\n", + "plt.legend()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "When you run the fully implemented semi-supervised GAN, you should usually find that the test accuracy peaks a little above 71%. It should definitely stay above 70% fairly consistently throughout the last several epochs of training.\n", + "\n", + "This is a little bit better than a [NIPS 2014 paper](https://arxiv.org/pdf/1406.5298.pdf) that got 64% accuracy on 1000-label SVHN with variational methods. However, we still have lost something by not using all the labels. If you re-run with all the labels included, you should obtain over 80% accuracy using this architecture (and other architectures that take longer to run can do much better)." + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlIAAAEyCAYAAAAxwz60AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nS6hCA6VhuhwAAAAASUVORK5CYII=\n", + "text/plain": [ + "" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "_ = view_samples(-1, samples, 5, 10, figsize=(10,5))" + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "mkdir: cannot create directory ‘images’: File exists\r\n" + ] + } + ], + "source": [ + "!mkdir images" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [ + "for ii in range(len(samples)):\n", + " fig, ax = view_samples(ii, samples, 5, 10, figsize=(10,5))\n", + " fig.savefig('images/samples_{:03d}.png'.format(ii))\n", + " plt.close()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true + }, + "source": [ + "Congratulations! You now know how to train a semi-supervised GAN. This exercise is stripped down to make it run faster and to make it simpler to implement. In the original work by Tim Salimans at OpenAI, a GAN using [more tricks and more runtime](https://arxiv.org/pdf/1606.03498.pdf) reaches over 94% accuracy using only 1,000 labeled examples." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "collapsed": true + }, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.6.1" + } + }, + "nbformat": 4, + "nbformat_minor": 2 +}