Skip to content

Instantly share code, notes, and snippets.

@mtsokol
Last active May 30, 2021 17:15
Show Gist options
  • Save mtsokol/fe00f78bb0b925f986d241622c2ec019 to your computer and use it in GitHub Desktop.
Save mtsokol/fe00f78bb0b925f986d241622c2ec019 to your computer and use it in GitHub Desktop.
Display the source blob
Display the rendered blob
Raw
{
"cells": [
{
"cell_type": "markdown",
"id": "bcc8b28e-8773-478e-9be6-87f9df0f2d77",
"metadata": {},
"source": [
"# Example: Utilizing Predictive and Deterministic with MCMC and SVI\n",
"\n",
"In this short tutorial we'll see how to use `deterministic` statements inside a model and inspect its samples with `Predictive` class. Additionally a `GammaPoisson` distribution will be discussed as it'll be used within our model.\n",
"\n",
"Check out other tutorials that use `Predictive` and `Deterministic`:\n",
"\n",
"- [Example: analyzing baseball stats with MCMC](http://pyro.ai/examples/baseball.html)\n",
"- [Bayesian Regression - Inference Algorithms (Part 2)](http://pyro.ai/examples/bayesian_regression_ii.html)"
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "35e41be6-291e-452d-b934-7e9f3a7532b1",
"metadata": {},
"outputs": [],
"source": [
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import pandas as pd\n",
"from sklearn.datasets import make_regression\n",
"import pyro.distributions as dist\n",
"from pyro.infer import MCMC, NUTS, Predictive\n",
"from pyro.infer.mcmc.util import summary\n",
"from pyro.distributions import constraints\n",
"import pyro\n",
"import torch\n",
"\n",
"pyro.set_rng_seed(101)\n",
"\n",
"%matplotlib inline\n",
"%config InlineBackend.figure_format='retina'"
]
},
{
"cell_type": "markdown",
"id": "49144812-c5d0-49d8-97fe-b425dda440f2",
"metadata": {},
"source": [
"## Data generation\n",
"\n",
"Let's generate our data with `sklearn.datasets.make_regression` method where we can determine the number of features, bias and noise power. Also we'll transform the target variable and make it a `torch` tensor."
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "19fdf2d3-51af-4909-bdc7-876726633a30",
"metadata": {},
"outputs": [],
"source": [
"X, y = make_regression(n_features=1, bias=150., noise=5., random_state=108)\n",
"\n",
"X_ = torch.tensor(X, dtype=torch.float)\n",
"y_ = torch.tensor((y**3)/100000. + 10., dtype=torch.float)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "f9314d98-59d3-451a-9909-3b1936aef7a7",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"image/png": {
"height": 248,
"width": 375
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(X_, y_);"
]
},
{
"cell_type": "markdown",
"id": "407399e5-31df-4fa8-9628-ddb0fd04b038",
"metadata": {},
"source": [
"## Model definition\n",
"\n",
"In our model we first sample coefficient from a normal distribution with zero mean and sampled standard deviation. We use `to_event(1)` to move the expanded dimension from `batch_size` to `event_size` as we want to sample from a multivariate normal distribution. `deterministic` part is used to register a `name` whose `value` is fully determined by arguments passed to it. Here we use `softplus` to be sure that the resulting `rate` isn't negative. Then we use vectorized version of `plate` to record `counts` from passed dataset as they were sampled from `GammaPoisson` distribution. \n",
"\n",
"For now this model might be a little obscure but later we will dive into sampled data to better grasp it's internals."
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "2f987086-f6a7-494c-a37d-5e6ba57e4867",
"metadata": {},
"outputs": [],
"source": [
"def model(features, counts):\n",
" N, P = features.shape\n",
" scale = pyro.sample(\"scale\", dist.LogNormal(0, 1))\n",
" coef = pyro.sample(\"coef\", dist.Normal(0, scale).expand([P]).to_event(1))\n",
" rate = pyro.deterministic(\"rate\", torch.nn.functional.softplus(coef @ features.T))\n",
" concentration = pyro.sample(\"concentration\", dist.LogNormal(0, 1))\n",
" with pyro.plate(\"bins\", N):\n",
" return pyro.sample(\"counts\", dist.GammaPoisson(concentration, rate, validate_args=False), obs=counts)"
]
},
{
"cell_type": "markdown",
"id": "1985507f-746e-42e4-b059-c14a1dbf05f8",
"metadata": {},
"source": [
"## Inference\n",
"\n",
"Inference will be done with MCMC algorithm. IMPORTANT! Please note that only `scale` and `coef` variables are returned in samples dict. `deterministic` parts are available via `Predictive`, similarly as observed samples."
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "e295c1a8-eb4e-4f66-8fe7-ca7d4ae1d2de",
"metadata": {},
"outputs": [],
"source": [
"nuts_kernel = NUTS(model)\n",
"mcmc = MCMC(nuts_kernel, num_samples=500)"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "41598ceb-09ec-4149-a733-222d0694d2b6",
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Sample: 100%|██████████| 1000/1000 [00:22, 43.82it/s, step size=6.35e-01, acc. prob=0.900]"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"CPU times: user 22.3 s, sys: 325 ms, total: 22.6 s\n",
"Wall time: 22.8 s\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"%%time\n",
"\n",
"mcmc.run(X_, y_);"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "134eeb75-3367-4177-9794-11319e1d696e",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"dict_keys(['scale', 'coef', 'concentration'])\n"
]
}
],
"source": [
"samples = mcmc.get_samples()\n",
"print(samples.keys())"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "e5b9d5c4-aed3-438a-88d1-4d50f249c985",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"dict_keys(['counts', 'rate'])\n"
]
}
],
"source": [
"predictive = Predictive(model, samples)(X_, None)\n",
"print(predictive.keys())"
]
},
{
"cell_type": "markdown",
"id": "772fa2a1-9336-4931-9f71-e427c96bf51f",
"metadata": {},
"source": [
"After sampling let's see how well our model fits the data. We compute sampled `counts` mean and standard deviation and plot it against the original data."
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "3b6f72cb-8344-4abc-ba6e-b38894377ba9",
"metadata": {},
"outputs": [],
"source": [
"def prepare_counts_df(predictive):\n",
" counts = predictive['counts'].numpy()\n",
" counts_mean = counts.mean(axis=0)\n",
" counts_std = counts.std(axis=0)\n",
" \n",
" counts_df = pd.DataFrame({\n",
" \"feat\": X_.squeeze(),\n",
" \"mean\": counts_mean,\n",
" \"high\": counts_mean + counts_std,\n",
" \"low\": counts_mean - counts_std,\n",
" })\n",
"\n",
" return counts_df.sort_values(by=['feat'])"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "a884c08f-e4ff-4acf-95ab-35d9a0d08646",
"metadata": {},
"outputs": [],
"source": [
"counts_df = prepare_counts_df(predictive)"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "ca56d0f9-11bd-48c0-8d4a-0162dac03071",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"image/png": {
"height": 248,
"width": 375
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(X_, y_, c='r')\n",
"plt.plot(counts_df['feat'], counts_df['mean'])\n",
"plt.fill_between(counts_df['feat'], counts_df['high'], counts_df['low'], alpha=0.5);"
]
},
{
"cell_type": "markdown",
"id": "1a572f9a-db20-4fa4-af80-a5ac416d4bbe",
"metadata": {},
"source": [
"But where do these values (and uncertainty) come from? Let's find out!"
]
},
{
"cell_type": "markdown",
"id": "e351f4ac-bb56-4cea-9dad-8db0bd433838",
"metadata": {},
"source": [
"## Inspecting deterministic part\n",
"\n",
"Now let's move to the essence of this tutorial. `GammaPoisson` distribution used here and parameterized with (`concentration`, `rate`) arguments is basically an alternative parametrization of `NegativeBinomial` distribution. \n",
"\n",
"`NegativeBinomial` answers a question: How many successes will we record before seeing `r` failures (overall) if each trial wins with probability `p`? \n",
"\n",
"The reparametrization occurs as follows:\n",
"\n",
"- `concentration = r`\n",
"- `rate = 1 / (p + 1)`\n",
"\n",
"First we check sampled mean of `concentration` and `coef` variables..."
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "eae492b2-3df4-462d-94cd-e60e939b0b49",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Concentration mean: 28.858654022216797\n",
"Concentration std: 0.8479085564613342\n"
]
}
],
"source": [
"print('Concentration mean: ', samples['concentration'].mean().item())\n",
"print('Concentration std: ', samples['concentration'].std().item())"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "aee98f2b-0201-4de3-abee-f5598757ac43",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Coef mean: -1.2455012798309326\n",
"Coef std: 0.037528540939092636\n"
]
}
],
"source": [
"print('Coef mean: ', samples['coef'].mean().item())\n",
"print('Coef std: ', samples['coef'].std().item())"
]
},
{
"cell_type": "markdown",
"id": "c6fd0556-a81e-4b6f-81eb-dd0877b8aafa",
"metadata": {},
"source": [
"...and do reparametrization (again please note that we get it from `predictive`!)."
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "00a2edd1-feea-4c8f-b752-5b41a40062e0",
"metadata": {},
"outputs": [],
"source": [
"rates = predictive['rate'].squeeze()\n",
"rates_reparam = 1. / (rates + 1.) # here's reparametrization"
]
},
{
"cell_type": "markdown",
"id": "d8b889b8-845f-4929-9e6d-5694cffad010",
"metadata": {},
"source": [
"Now we plot reparametrized `rate`:"
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "0f569654-eeca-47e1-9286-ceff6792712f",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAABfUAAAIPCAYAAADacDRxAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAABYlAAAWJQFJUiTwAACUdElEQVR4nOz9f3xcdZ33/z9fsWj6I5QOkZRtKIxNQvnRqkAauCI/anZr/bGXkK7tdV1LVvkg4gLbqnhtW0EFFFu8VjQFcVnsopb9rmG3Fa/VC6gbK7oDpFnQbYGWZEpKaKGBIfyYUgLWvL9/zAyGNJOZOXPm9+N+u+V2yDnn/T7vhDQ553Ve79fbnHMCAAAAAAAAAADFr6rQAwAAAAAAAAAAAOkhqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAAAAQIkgqA8AAAAAAICKZWbXmZkzsx8UeiwAkI4phR4AAFQ6MztG0uckyTl3XSHHAgAAAKAyleJziZl9StJJku5xzv2uoIMBgDwiqA8AhXeMpK/G//u6wg0DAAAAQAU7RqX3XPIpSedL2ivpd4UcCADkE+V3AAAAAAAAAAAoEQT1AQAAAAAAAAAoEQT1ASAFM9sbXzTpAjObY2a3mdlTZvaGmf1uzHn1ZvZFM7vPzPrN7JCZvWpmvzWz6+M1Ksf3/StJA2M+d+M+rpugzUlmdouZPRm/RtTMHjGz1WY23cPX96n4tX4V//x/mtmD8bG/YGY/MbNTxpx/fPz6e81sxMzCZrbGzN6R4jp/bmY/NbMDZvammT1vZv9mZh+apM15ZtZpZj1m9uyYdveZ2V9M0u4Hie+fmb3DzD5nZv8V/34Nm9nPzOysTL9XAAAAQKGU+3NJvM/3mtmP4l/rG/E+n4p/LZ8zs2nx8z5lZk6x0juSdOe48e6doO+Tzeyf488Tr5vZbjP7qpm9y8tYAaCQqKkPAOlrkvQvkmolHZL0+3HHvyNpWfy/35R0ULG6lO+Lf/ylmV3gnNs3ps2wpEi8T0kaGtfnwbGfmFm7pH+SVB3fdUjSuySdEf/4SzP7M+fc+H7SYmY3SfpbSYclvR4f14WSzjWz/ybJSfqlpHpJUcX+jsyTtE7SCZKunKDPoyTdKekvx+x+VdK7JX1M0sfM7JvOudXj2s2Q9MCYXdH4mN4t6UOSPmRm/+Ccu3ySL2mKpJ/Hz/+9pDckzZL0UUltZvZB59xDk39XAAAAgKJSls8lZvYRSfdIOiq+6w1Jo5KC8Y8PSbpP0m7FnguGJAXi578a35fwwri+z5N0r6Rp8V2vxvu8Lt7vr9IdJwAUAzL1ASB935L0nKRW59x059wMSWOzxXdJWqnYTfZU59yxit3kXiCpV7Hg9+1jO3TOtUtqHvP57HEff5c4ZmbNkn6sWKD6Rkn1zrnpkqZK+m+S/lPSAkk/8vj1vU/S5yV9TtJM59zRkhZKelLSsZLWS7pL0jOS3hc/frSka+Pt/9rMTp+g328qFtAPS1ouaYZzbma87RWKBev/1sz+57h2o5L+VdJFko51zh0dbzdL0lWKPVh8xsw+McnXdKVi398V8evWSHqvpMcU+3/TmfrbAgAAABSVcn0uuVWxAP3PJJ3snKuO3//PlHSepDskjcTH1+Wcmy3pwXjbVePG+9bXYmazFHsJMk3So4o9y8yUNEPSJxV7Prgiw7ECQEGZc67QYwCAohafunmipJclzfeSBW9mAcUySmolvcc5t3fMsZMUn+rqnLNJ+vgPSa2SPuucu32C4wHFgtXHS2p2zv1nmmP7lGKZ9JJ0nXPu+nHHz5X06/inL8XH//K4c7olfVDSV51zN4zZ36jYS4GIpDOdc89McP3/IemfJT3unJvopUCycXco9qDwK+fc4nHHfqDYDbokneuc+49xx89U7GFDkk50zg2me10AAACgEMr5ucTMjtMfZwfMTvdri5cNOl/SJc65HyQ558uSbpD0omLft8i44xdL2hT/9IfOuU+lc20AKCQy9QEgfT/yWtbGOTesWBaJKZa9khEzm6fYjfPLkjZOco1745/+mYdhvinp5gn2hxTPiJH0vfEB/bju+HZ8UP6vFPuauyYK6Mf9q2JTa08zs+MzGO+/xbdnT1LP/zfjA/qS5Jx7RFJiunHaLxIAAACAIlCOzyUHFZupK8VeBvgpMYvhjvEB/bh/kvS0z9cEgJyipj4ApC9l7XUzWyTps4rdINdLmmiBqD/xcO3EDfcMSfvMkibOzIhvT/Bwjb3Ouej4nc65UTOLKPb1PJakbeKhYta4/YlxfzJFmZxE3cwTFJtKLEkysymKZdx/QrFpsQFJ7xzXtjp+3Ylu0HsnueZ+xb6m8WMGAAAAilnZPZc45w6Z2QOSFku638xuUawMz07n3B88jFOSZGbvlHRa/NMHJjrHOefM7NeSOrxeBwDyjaA+AKTvhckOmtkXFasfn7iz/YNi5WrejH8+U7EA9EQ31KkkslWmSKpL4/xpqU85wnOTHPtDinMSx48atz8x7pr4RypvjTu+UO79ensG0euK/X9IZPEkvhfTNXFQ/4iXFGMkZh+MHzMAAABQzMr1ueTTigXyT5H0tfjHwXjA/Z8l/dg5dziD/qRYUlBiVu+zk5y3P8N+AaCgKL8DAOlLmiFiZqdJukmxG+dbFcsGeZdzLpBYrEmxMjPSH2+uM5H4ff1fzjlL4+NTHq6RC4lxfz7Ncf9qTNsvKxbQjyiWrV/nnJvmnDsu/v2cM+ZcL99TAAAAoBSV5XOJc+4pSQslXSTpHxRb8HeGpI8oVvO+J574AwAVj6A+APhjmWK/U+93zv2Nc+6JCaaJppPJkkyivI2XsjqFlBj3XA9tE+V6/sY59yPn3PPjjmfz/QQAAADKUUk/lzjnDjvn7nHOXe6cO1WxmQH/W7FZtmdI+mqGXQ7rjy9BJis35KUUEQAUDEF9APBHfXz724kOmtl0SWcnaTs65rxk2TKJupkBM2vxNMLCSIx7qYe2k35PJf2phz4BAACAclZWzyXOuQPOub+T9J34rvPHnZIY84Tjdc69Kenx+KfnTXRO/Gud8BgAFCuC+gDgj1fi2wVJjl+j5DXlXx3z38dMdIJzbrekh+OfftPMktaBN7OpZvau5EPNqx9JcpJOMbPLJzvRzMYvWJv0exqfdnuNLyMEAAAAykdJPpeY2VGTvEiQYmtrSdL4/hJjPmaStv8S315mZoEJjv8PSSelGiMAFBOC+gDgj1/Etx81s7VmNk2SzOzdZvZ/JK2V9OJEDZ1zL+uPizZdMsk1Vkp6Q7Eskm4z+4CZVcWv8w4zW2BmX5H0lP64gFVBOeeekPTt+Ke3mdk6M0tkD8nMasxsiZndpT/ebCckvqc3m9n5iZt8M2uW1C3p2BwPHwAAACg1pfpccpqkx8zsc2bWNObe/ygzWybpC/Hz7h/XLpGF325mM5P0/V1Jz0uqlXS/mS0c0/fFku7QH1+GAEBJIKgPAD5wzm2VtCX+6TckHTSzYcVqTn5R0kZJP5uki+/Ht98ys4Nmtjf+8bkx1+hVbNGoVySdK+k3kg6ZWUSxzJUdkq6XNFux7Phi8beSvqfY35w1kp4xs1fM7GXFvpb7Jf2lpHeMa3etYovkniDpV4p9rQclbVcs8+h/5WPwAAAAQKko8eeSUxVLCHpS0utm9qJitfT/VdJMSf8p6evj2myS9KakD0iKmNn++Hj/Y8x4X5K0PD62syT9V/xZJBpvv0Ox5xUAKBkE9QHAPysUC1rvkvR7xeo6hiR90jn36RRtb5C0WrEbSpN0YvzjmLEnOefuldSk2M3so4plyByj2LTTByWtl3Smc+5pP74gPzjn/uCcu0KxG+27JD2t2LTZakmDkv6vpKsk/cW4dk9JWhRv87xiQf+XJf2TpOb4AwsAAACAtyvF55Jdij0P/L1i6wG8LOloxV4c/Iekv5HU6pwbWyIoUQ7ozyTdFz93dny89ePOe0DS+yV1SXpBseeRvZKuk/TB+PgBoGSYc8WUzAkAAAAAAAAAAJIhUx8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBJBUB8AAAAAAAAAgBIxpdADyCczG5B0tKS9BR4KAAAAStNJkl51zgULPZBCMLN6STdIWirpWEnPSbpH0vXOuZcy6Ccg6SuSLpR0vKQXJd0n6SvOuX1+XtvM/kLSZZLOlDRD0vOSfitpnXPu4XTHnKRvni8AAADg1Uny+Gxhzjn/h1OkzOzFqVOnBk455ZRCDwUAAAAlaNeuXXr99deHnXPHFnos+WZm8yQ9KOk4ST+VtFvSIkmLJT0pqdU592Ia/Rwb76dJ0i8l9UqaL+njigXcz3HOPZXttc1siqQfSvpfkvol/ULSK5JmSzpH0q3Oue9m+n0Ydw2eLwAAAOBJNs8WFZWpL2nvKaecEnjkkUcKPQ4AAACUoDPPPFOPPvro3kKPo0BuUyyovtI5d0tip5ndLOnzkm6U9Nk0+vmGYgH9m51zV4/pZ6Wkzvh1lvpw7esVC+jfqNgMgNGxB83sqDTGmgrPFwAAAPAkm2cLauoDAAAAmFQ8U36JYmVmxme3f1XSa5I6zGx6in5mSOqIn3/duMO3Snpa0ofM7D3ZXNvMZkv6oqSHnXPXjg/oS5Jz7veTjRUAAAAoVgT1AQAAAKSyOL7dOj5A7pyLSgpJmibp7BT9nC1pqqRQvN3YfkYl3T/uel6v/ReS3inpx2Y21cz+wszWmNmVZvbeFGMEAAAAilqlld8BAAAAkLmT49u+JMf7Fcumb5LUnWU/iveTzbWb49tpitXfnzu2gZltlvRXzrlDk4x17PnJ6uvMT6c9AAAA4Ccy9QEAAACkMjO+fSXJ8cT+Y3LQj5c2x8W3X1OsbM8ZkmYols3/n5KWKVanHwAAACg5ZOoDAAAAKDeJ5KVhSX/unHs1/nmPmf13xbL+O8zsGufc/lSdOefOnGh/PIP/DD8GDAAAAKSLTH0AAAAAqSSy4WcmOZ7Y/3IO+vHSJvHf3WMC+pIk59xzknoUexY6a9LRAgAAAEWIoD4AAACAVJ6Mb5uSHG+Mb5PVvc+mn2zavJykzUvx7dQkxwEAAICiRVAfAAAAQCrb4tslZva2Zwgzq5HUKumQpIdT9POwpNcltcbbje2nSrEFb8dez+u1/z2+PT3JOE6LbwdSjBcAAAAoOgT1AQAAAEzKObdH0lZJJ0m6ctzh6yVNl7TJOfdaYqeZzTez+eP6OShpU/z868b1c1W8//udc09lc21Jv5H0O0kfMLOLxjYws8sknSIprNiiuQAAAEBJYaFcAAAAAOm4QtKDkjaYWZukXZJaJC1WrPTNNePO3xXf2rj9X5J0gaQvmNn7JG1XLMj+cUnP68jAfcbXds45M/ukpAckbTazf4ufd5qkD0t6TdInnXN/SP/LBwAAAIoDQX0AAACUlL6hqELhiA6OHNaM6ilqbahVU11N6obIinNuj5mdJekGSUslfUTSc5I6JV3vnHtpsvZj+nnRzM6R9FVJF0o6V9KLku6U9BXn3D4/ru2c22FmZ8SvsyTeJiLpnyR9zTn35Pg2AAAA5Yz76PJBUB8AAAAlIRSOqLO7X9sHho84tigY0Kq2RrU21BZgZJXDOfeMpEvSPHd8hv7YY8OSVsU/fL/2mDYDkj6VSRsAAIBSMzZYv+/l17XvpUM69MZhvfz67zX9XVP0jirT0Ctv6MCrI0e0TXYfzQuA4kZQHwAAAEWvq3dQa7fs1Kib+Pj2gWF1bOzR+vaFWt58Qn4HBwAAAOTZ1scP6B9+/ZSeHIoqOnLYcz/j76MnS6Q5MTBNVyyepxXNc7MZOnxAUB8AAABFLRSOTBrQTxh10potOzRn1lQy9gEAAFCWbn9gj27dFs4qkD9e4j66//moNv7HQNL77qeHD2n15p36+wee0mfPf48OvfkHMvkLhKA+AAAAikKyKb6d3f0pA/oJo07a0N1PUB8AAABl5+q7f6fNj+7PSd+jTrrjNwNpnTsQeU2rN+88Yj8lMfOHoD4AAAAKarIpvgvmzNTO/a9k1F/PwLD6hqJkCgEAAKCkbH38gLY8uk/RkcOqqZ6i9jPqteS02ZJiGfq5Cuj7hZKY+UNQHwAAAAWTqlZ+pgH9hFA4QlAfAAAAJSFZSZ37Hh9STfUUXbW4QbduCxdodJmhJGZ++BbUN7N6STdIWirpWEnPSbpH0vXOuZcy6GeZpL+R9H5J75T0lKS7JH3LOfemX+MFAABAYSTK7Dy+/1VtfnSf0qysk5GDPtYYBQAAAHIlVUmd6Mhhrbt3dx5HlD1KYuaeL0F9M5sn6UFJx0n6qaTdkhZJWiVpqZm1OudeTKOfb0haK+mgpM2ShiWdK+kbktrM7MPOud/7MWYAAADkx1tB/Gdf1faBYQ0OH8r5NWdUMyEVAAAAxalvKKqf/Haftj4+pD0vvFbo4eQEJTFzy6+nndsUC+ivdM7dkthpZjdL+rykGyV9drIOzOwMxQL6L0s60zn3VHy/xfv/rGIZ/Df7NGYAAADk0GS18nONrCAAAAAUm67eQX1ra5+ej75R6KHkBSUxcyfroH48S3+JpL2Svjvu8FclfUZSh5ld7Zyb7NXThfHt9xMBfUlyzjkz+5JiQf0rRVAfAACg6KWqlZ9LLcEADw8AAAAoGqFwRNfe85gGIuWZlZ9MqpKYiRm9B0cOa0b1FLU21HIfnyY/MvUXx7dbnXOjYw8456JmFlIs6H+2pO5J+pkd3z41/oBz7iUze0nSe8ws6Jwb8GHcAAAAyIFQOFKwgH6VSSvbGvN/YQAAAGACXb2DWrNlp1wB7o0LLVlJzMlm9C4KBrSqrZGZtylU+dDHyfFtX5Lj/fFtU4p+IvFtcPwBMztG0qxx10vKzB6Z6EPS/FRtAQAAkJ3O7v6CBfTXty/kAQAAAABFoat3UGs2F39Av8qky84Nqsr87Xei+/Ku3kF1bOxJWqJz+8CwOjb26O7eZ/wdTJnxI1N/Znz7SpLjif3HpOjn54rV1L/MzG5zzu2V3qqpf+OY82ZN0BYAAABFoG8omtMa+gvmHK2d+189Yn9LMKCVZPQAAACgwPqGotr00NP65e7ntf/l1ws9nJTG3kc3HDfDt1kFE5XETHdG76iT1mzZoTmzpnJ/n4RfC+VmzTkXMrONki6VtMPMNksalnSupIWSdiuWaT+avJe3+jpzov3xbP0zfBs0AAAA3iYUjqQ+yaOWYEBdl59D7U0AAAAUnclKyhST445+l85reLdOm3P0EffRK5rnqn7WNH35nsf0VBb1/5OVxMxkRu+okzZ09xPUT8KPoH4iE39mkuOJ/S+n0ddlkrbHt8slOUkPS7pA0rWKBfWf9zhOAAAAZClVQD3VYlhejX0waKqrIYgPAACAotHVO5iXNaVqqqfoqsUNunVbWNEJ7rvfOaVK82fPUE31UTr0xmG9/PrvNf1dU/TuGe/S/ONrdNH761PeR7c21OqXX7xAXb2Dum3bHj09fCijMSYriellRm/PwLD6hqLc+0/Aj6D+k/Ftspr5idcyyWruv8U55yT9Q/zjbcxsgWJZ+o96GCMAAACykO5iVskWw8oGtfIBAABQrNItKeOHqxY36PLz5+ny8+dp6+MHtOXRfYqOHFZN9RS1n1GvJafN9u1aK5rnakXz3COSeqa98x3a8uh+9UzwXDBZSUyvM3pD4QhB/Qn48dS1Lb5dYmZVzrm3yuOYWY2kVkmHFMu498TMLpA0V9K/OeeS1e4HAABADqTKPEosZpWLwDu18gEAAFDMMikpk41PnFmvy8+f99bnS06b7WsQP5mJZslOFOxPVRLT64zeXM0ELnVZB/Wdc3vMbKukJZKulHTLmMPXS5ou6Xbn3FuFmMxsfrzt7rF9mdnRzrlXx+07UdL3Jb2pWAkeAAAA5Emmi1lturRFi4IBz7VETdKyM+t12p8cWeMTAAAAKCZeSspkKlFyZ2xAvxhkWhLT64zeidqxxpZ/C+VeIelBSRvMrE3SLkktkhYrVnbnmnHn74pvbdz+jfEg/qOKLZIblPTfJR0lqcM5t8On8QIAACANXhazWtXWqI6NPRlnLCXK7CxvPiHzgQIAAAB55rWkTDqaTzxGl503Ly/Z+Pngdebt2HbplgStBFV+dOKc2yPpLEk/UCyYf7WkeZI6JZ3tnHsxza5+Jun3kj4h6YuSPiDpXyW91znX5cdYAQAAkB6vi1m9u+ZdWte+QFXj0zcm0RIMaNOlLQT0AQAAUDJyURrGJH1z2UL9y1+3lk1AX4pl9i8KBjJq0xIMvJWB39U7qI6NPUmfT7YPDOvi7/doQ3d/1mMtBb6tZOace0bSJWmeO+EjnnPuh5J+6NeYAAAA4F02i1ld0hpU/axp2tDdP+EiWiceO03NJwZ02hzK7AAAAKA0eS0pk8wps2t07cdOLdts80xm9FaZtLKtUVL6JUGdpJt/0aef73xOXynj76PkY1AfAAAA5SXbxaxaG2rV2lBLzUsAAACUJT+CxjOnHqXzm2p11Qcby/4eubWhVuvaF6QM0CfKcia+v5kuRvzkgag6NvaUdWlPgvoAAAAVItPgul+LWWW6iBYAAABQChIlZbwslmuS1i9boBXNc/0fWBFb0Tx30hm9LcGAVo6pje91MeJRJ63ZskNzZk0ty4x9gvoAAABlzuuCUn4sZgUgN5gBAwBAccikpExCIhO9XLPIU8lkRm82ixGPOmlDd39ZPp8Q1AcAAChjXb2Dk05v3T4wnHRqqpfMo7GLWQHwX1fvoL67bY8Ghw8dcWyyl3QAACA30i0pkzA+E72SpTOjN9vFiHsGhvXV//uYTjp2elklQRDUBwAAKCNjs12ej47orp5BuRQPF5NNTfW6mBUAf4XCEV17z2MaiLyW9JzJXtIVM2YdAABKXaqSMvWzpmrxye9Wxzkn8TcuQ34sRvzDB59+67/LJQmCoD4AAEAZCIUjWnfvLj22/1VP7ZNNTfW6mBUA/3T1DmrNlp0pX9BJpVU/1mtpMAAAilEmJWWQPr/vBUo1CWI8gvoAAAAl7safP6E7fjOQdT89A8PqG4oe8dCR6WJWAPwTCke0Ns2AfkIp1I/NpjQYAADFLJ2SMkhfNosRJ1NKSRDJENQHAAAoYX4F9BNC4ciEDyFkHgGF0dndn9HCewnJXtIVg8SLilRfVzk8cAMAgOx5WYw4lVJIgpgMQX0AAIASFQpHfA3oS6kXoiLzCMifvqFoVllpyV7SFVomLypK/YEbAABkL9PFiNNVzEkQqRDUBwAAKFHr793te59+LEQFwB+hcCSr9qle0hWClxcVpfzADQAofsxELQ2JkqBf/9kT2nUg6lu/xZoEkQpPbQAAACWobyiqnftf8b1fsmGB4pFtUL4YX9J5fVFRqg/cAIDixYLtpae1oVb3fu48dXb36Tu/6JcfSft7X3xNd4YGSu6lTvHd5QEAAOAtyTKHss3gnUhLMFASN7BApcg2KF+MgQivLyqKcdYBAKB0sWB7aVvV1qSzTgxoQ3e/erJcQPeHDz59xL5SeKlDUB8AAKAIpcocmhuY5uv1qkxa2dboa58AspPNg2SxvqTz+qIi17MOKL0AAJWDBdvLQ2tDrVobat/6G7438pp++NCRAXovSuGlDkF9AACAItI3FNWtv+zXv/3Xc0mnk24fGFZvlhkpY1WZtL59IQ8rQJFpqqvRomAg4xr0xfySzuvvmVz9fqL0AgBUnq/97AkWbC8jTXU1b72I33Ug87V7kin2lzpVhR4AAAAAYoGlj93yGy359q/1fycJ6Cf4UT9SimXzbrq0pWgzUIBKt6qtUVWW/vmm4n5Jl3hRkYlczTro6h1Ux8aepA//iSy9u3uf8f3aAIDC6Ozu0+4MF1lNLNiO4pfpfVMqiZc6xYhMfQAAgAL79I969e9PPJ+Xa3WcfaKOq3kX5SWAEtHaUKt17QvSKhMQrJ2ur194etEG9BNWtTWqY2NPWlmSuZp1QOkFAKg8oXBE3/mFtwAtC7aXhkzum9KVeKlTbP//CeoDAAAUQN9QVN+8b7d+uft53244U7ns3KCu+eip+bkYAN+saJ6r+lnTki4Gd+Kx03TFBfO0onluAUaXuXQfuHNZGqyzu5/SCwBQYTq7+z3PdmXB9tKR6r7Ji2J8qUNQHwAAII8mq9/slSl1OR4C+kBpG78YXKkv6JrqgbslGNDKHNWz7xvKvN5usWbpAQDS4+V3/1i5XrAd/kp23+R1Md1ifKnDTyQAAECedPUO+joVNGHZGfV65qVDEwbGFtbP1Oql88kwBcrE2MXgSl2hXlSEwhHP7crlew8Alcbr7/4E7qVL0/j7pjtDA576KcaXOsU3IgAAgDLU2d2nb3us4ZnKaXOO1t8tf2/ZZPACqCz5flHhNduuGLP0AADpyeZ3eK4WbEf+eX05U4wvdQjqAwAA5NDtD+zRrdvCiuYwGJS4ySynDF4AyBWv2XbFmKUHAEiP19/hptws2I7CaKqr0aJgIKNSTMX6Uqeq0AMAAAAoV1ff/Tutu3d3TgP6xXqTCQDFqpyy9AAA6fH6O/zzf9bE7/8ys6qtUVWW3rlVVrwvdQjqAwAA5MDtD+zR5kf35/QaxXyTCQDFKpGllwleoAJAafPyu3/+7BrutctQa0Ot1rUvSBnYrzJpffvCon2pQ1AfAAAgB27dFs5p/8V+kwkAxaxcsvQAAOnL9Hf/lz92am4HhIJZ0TxXmy5tUUuSFz0twYA2Xdqi5c0n5Hlk6aMoIAAAgM+2Pn4g5yV3VrY1EtAHAI8SWXprt+zUqEt+Hi9QAaB88LsfY7U21Kq1oVZ9Q1GFwhEdHDmsGdVT1NpQWxKz8wjqAwAAeNQ3FNVPfrtfu597VZI0//gaXfT+em15dF9OrnfOewK6/uOnl8RNJgAUuxXNc1U/a5o2dPerZ4IF8/x+gVqqQQMAKCf5/t2P4tdUV1OSf48J6gMAAGQoFI7oaz97QrsPRN+2f9uTL+h7v3pK0975Dt+vedm5QV3zUaYAA4Cf8pGlFwpH1Nndr+0TBI8WBQNaRfAIAPKq1DO0AYmgPgAAQEa6ege1ZvNOTTJjV4fe/INv15s17Sjd+r/OIOADADmUqyy9rt7BScs8bB8YVsfGHq1vX1jUdXsBoByVaoY2IBHUBwAASEvfUFSbHnpamx5+Oi/Xe9eUKnWcfaKuZYEuAChJoXAkZd1mSRp10potOzRn1lRe4AIAgLQQ1AcAAJhEKBzR+nt3a+f+V3J+raOrp6iprkafOe89WnLa7JxfDwCQO53d/SkD+gmjTtrQ3U9QHwA8oIwOKhFBfQAAgCRu/PkTuuM3A3m5FjXzAaB89A1FJ6yhP5megWH1DUUJRAFAmlizBJWsqtADAAAAKEZX3PVI3gL6nziznoA+AJSRUDiS13YAUGm6egfVsbEn6QvUxJold/c+k+eRAflBpj4AAMA4N/78Cf2/xw741t87p1TpzcOjR+yvqZ6iqxY36PLz5/l2LQBA4R0cOZxVO0pJAEByrFkCENQHAAB4S2wx3L3a9PCgr/2u/fB8zTlmqrY8uk/RkcOqqZ6i9jPqqZsPAGVqRrW3R+3noyNafvtDlJIAgEmwZglAUB8AAGDSepx+SGRYEsRHqTOzekk3SFoq6VhJz0m6R9L1zrmXMugnIOkrki6UdLykFyXdJ+krzrl9flzbzCZ73O9xzp2d7niBTHkNHt3VMyiX5Cc3UUpifftCLW8+IYvRAUDpYs0SIIagPgAAqGhdvYNas3mn0kz2yVhLMMADBMqCmc2T9KCk4yT9VNJuSYskrZK01MxanXMvptHPsfF+miT9UtKPJc2XdImkj5rZOc65p3y69tOSfjDB/glfHAB+aaqr0aJgIOPAU7KAfgKlJABUumzWLOGeHOWEoD4AAKhYtz+wR+vu3Z2z/k3SyrbGnPUP5NltigXVVzrnbknsNLObJX1e0o2SPptGP99QLKB/s3Pu6jH9rJTUGb/OUp+uvdc5d10aYwJ8t6qtUR0be9IuEZGuUSd96Sc7dcdfnUWACkDFyXbNEqBcVBV6AAAAAIXQ1TuY84D+TcsWkkmJshDPlF8iaa+k7447/FVJr0nqMLPpKfqZIakjfv514w7fqlhm/YfM7D1+XxvIt9aGWq1rX6Aqm/y8FIcn9PSLh7Tk27/W8tsf8py1CgClyOuaJV7bAcWKoD4AAKg4oXBEa7bszFn/82fX6K5Pt1DzGOVkcXy71Tk3OvaAcy4qKSRpmqRUderPljRVUijebmw/o5LuH3e9bK99jJn9f2b2JTO70syoo4+8WtE8V5subVFLMDDh8ZZgQBeffaLn/hN19u/ufcZzHwBQSrwmzJBog3LDayoAAFBxOrv7U9YtztR762fqvzUcq4veX085BJSjk+PbviTH+xXLpm+S1J1lP4r348e13ytp49gdZvZfkjqcc2m/2TOzR5Icmp9uH6hcrQ21am2oVd9QVKFwRAdHDmtG9ZS3FlG/pbs/dSeToM4+gEriZc0S1rhCOSKoDwAAKkrfUDTjhQtTuezcoK756Km+9gkUmZnx7StJjif2H5ODfrxe+2ZJmxV7GTCiWAB+taS/kPRLM3ufc25/ivECvmmqq5kwqORHSYhRJ23o7ieoD6AsjX8puuyMOfrPvcNprVlSZaxxhfJEUB8AAFQUv2sPf+T02QT0gSI0dhHeuP+U9Akz+1dJyyR9UbFFdtPp68yJ9scz+M/IZpyAX4H4noFh9Q1FyUYFUDZC4Yg6u/snTMgJ1k7X3shrmiyuX2XS+nbWuEJ5oqY+AACoKAdHDvvW12XnBnXbxRPG+oByk8iGn5nkeGL/yznox69rJ/x9fHtemucDOZUoJeEHFs0FUC66egfVsbEn6QzbgchrksWC+xNpCQa06VLWuEL5IlMfAABUFD/KHCysn6nVS+eT9YNK8mR825TkeGJee7K699n049e1E16IbyeOAgAFsKqtUR0be9IqJTEZP19cA0ChhMIRrd2yM+XvROekp198TTctW6BDb/7hiDVLgHJGUB8AAFSUbALxHWefqI5zTuQhAZVoW3y7xMyqnHOjiQNmViOpVdIhSQ+n6OdhSa9LajWzGudcdEw/VYoteDv2en5eO+Hs+PapNM8Hcq61oVbr2hekFcSajB8vrgGg0Dq7+9P+XTjqpC2P7lfX5efkdlBAkaH8DgAAqCheyxx8c9lCfe3C0wnooyI55/ZI2irpJElXjjt8vWJZ75ucc68ldprZfDObP66fg5I2xc+/blw/V8X7v98599SYNl6uvdDMjhr/dZjZQkk3xj+9K9nXCxTCiua52nRpi1qyKMXDDDIApa5vKJq05E4yiTVFgErCa3wAAFBxMi1zsPbD86nHCUhXSHpQ0gYza5O0S1KLpMWKlb65Ztz5u+JbG7f/S5IukPQFM3ufpO2STpH0cUnP68jAvZdrf0HSn5vZbyQ9I+kNSfMlLZX0Dkl3SPrn9L5sIH9aG2rV2lCrvqGoPv3D/9Tg8KG027YEA7x4BlDyvK4NEgpH+B2IikKmPgAAqDiJMgdV40ON45jFMvQvP39efgYGFLF4xvxZkn6gWED9aknzJHVKOts592Ka/bwo6RxJGyQ1xPtpkXSnpDPj18n22vdIekDS6ZI+KWmlpDMl3Svp4865zzjnsqxeDuROU11NWn+nEqpMWtnWmPpEAChyjz/7qqd2rCmCSkOmPgAAKDl9Q1GFwpGsFsNa0TxX9bOmaUN3v3ommOLbEgxoZVsjpQyAMZxzz0i6JM1zk4YjnXPDklbFP3Jx7XsUC+wDJSvdOvtVJq1vX8jfKwAlr6t3UJsf2eepLWuKoNL49hNvZvWSblBsSuuxkp5T7Eb6eufcSxn08wFJ/1vSeyXNVmwK7mOSNjjn7vNrvAAAoPSEwhF1dvdPWGdzUTCgVRkG4ceWOcj2JQEAAH7z8wU0f+sAFLNQOKK1W3bK6zQ6Xmyi0vgS1DezeYrVuDxO0k8l7Za0SLHMm6Vm1prOdFwz+2tJt0l6TdJPJO2TVC+pXdKHzexa59yNk3QBAADKVGd3n77zi/6kN/rbB4bVsbFH69sXZlz/vqmuhsAGAKAoZfsC2u8X4gCQC53d/WmvdzUea4qgEvmVqX+bYgH9lc65WxI7zexmSZ+XdKOkz07WgZkdJWmdpBHFamk+OebYNyT9VtI1ZvZ3zrk3fBo3AAAocqFwRDf87Ak9eSCa8txRJ63ZskNzZk0lQAEAKCteXkB39Q5OWr4nmxfiAOCXvqHohC8e08GaIqhUWS+UG8/SXyJpr6Tvjjv8VcWy7jvMbHqKrgKSZkrqGxvQlyTn3C5JfZKmSpqR7ZgBAEBp6OodVMfGnrQC+gmjTtrQ3Z/DUQEAUPwSpSxSZb6OOmn1lh0KhSP5GRgAjOP194+JNUVQubIO6ktaHN9udc6Njj3gnItKCkmaJunsFP08L+kFSU1m9rZXbGbWJKlR0u/SLOPzyEQfkuan9yUBAIBCSzcYMZGegWH1DaX/IgAAgHKTSSkL56TP3vUIgX0ABXFw5LCndsvOqGeWESqWH0H9k+PbviTHE6lyTZN14pxzkq6Mj+kRM/uhma0zsx9JekTS45I+4cN4AQBACcimrqbkPeMHAIBS56WURXTksC7+fo/u7n0mR6MCgInNqPZWHfy0OUf7PBKgdPhRU39mfPtKkuOJ/cek6sg59y9m9qykf5b0V2MODUm6U9JT6QzIOXfmRPvj2fpnpNMHAAAonGzqaiZ4zfgBAKDUeX2x7cTaNADyz+vvG35PoZL5kanvGzO7WNK/S/qNpFMUK9tziqRuSbdK+nHhRgcAAPLFjyx7rxk/AACUumxebLM2DYB8a6qr0aJgIKM2LcFAxouHA+XEj6B+IhN/ZpLjif0vT9ZJvG7+PypWZqfDObfbOfe6c263pA7FSvB8wswuyHbAAACguPmRZU/mDgCgUmX7Ypu1aQDk26q2RlVZeudWmbSyrTH1iUAZ8yOo/2R8m6xmfuJfWbKa+wlLJB0l6YEJFtwdlfTr+KcTltYBAADlI9tgBJk7AIBK5seLbdamAZBPrQ21Wte+IGVgv8qk9e0LSeBBxfMjqL8tvl1iZm/rz8xqJLVKOiTp4RT9vCu+fXeS44n9b3oZJAAAKB3Z3KSTuQMAqHReSlmMx9o0APJtRfNcbbq0RS1Jfn+1BAPadGmLljefkOeRAcUn62Kzzrk9ZrZVsUz7KyXdMubw9ZKmS7rdOfdaYqeZzY+33T3m3N/Et39hZn/nnNsx5vz3SfoLxdbt+WW2YwYAAIXTNxRVKBzRwZHDmlE9Ra0NtUdk1SeCEZkulkvmDgAAMavaGtWxsUejzlt71qYBUAitDbVqbahN65kBqGR+/ZW+QtKDkjaYWZukXZJaJC1WrOzONePO3xXfvjWpxjm33czulHSJpF4z+4mkpyWdJOlCSe+U9B3n3OM+jRkAAORRKBxRZ3f/hIH6RcGAVrU1vi0Yn2kw4pTZNbr2Y6cS0AcAQH8sZbFmy045D4F9/p4CKKSmuhqC+MAkfAnqx7P1z5J0g6Slkj4i6TlJnZKud869lGZXlypWO/9Tkj4kqUbSq5L+Q9Idzrkf+zFeAACQX129g1q7ZWfSAP32gWF1bOzR+vaFb02nTQQjJmsnxTIEPv9nTZTcAQBgnBXNc1U/a5o+e9cjimZQTiexNg2ZsgAAFCff5tM5555RLMs+nXMnXPbCOeck/SD+AQAAykAoHEkZmJekUSet2bJDc2ZNfSs7MBGM2NDdr54JMvxbggGtHJfhDwAA/qi1oVZ/f/GZuvj7PUonYb/KpA/OP07Lb38o7dl1AJAuXhYC/qBIHgAAyKnO7v60S+iMOmlDd//bAgXU1QQAIDutDbVavyz17Lcqk9rfX6+b7tud0ew6AEgl01KcACZHUB8AAPguEYDf++JrGS922zMwrL6h6ISL5xLEBwDAm3Rmv31w/nGTBvQTJppdBwDJeCnFCWByBPUBAIBvJsvAybQfAvgAAPgr1ey35bc/lNXsOgAYL5tSnACSI6gPAAB8kSoDJxMHM1jMDwAAZGai2W99Q1HfZtcBQEK2pTgBTKyq0AMAAAClr6t3UGs2+xPQl6QZ1eQdAACQT6FwJK/tAJS/bF4WApgcQX0AAOBZKBzR8tsf0urNO+VTPF+SyM4BACDPvM6S+3XfCwTgAEyIl4VA7pAGBwAAPPGz3M5YLcEA0/gBAMgzr7Pktj35grY9+YIWBQNa1dbIi3kAb/H6spBSnEBqZOoDAICMpbvgVaaqTFrZ1uhvpwAAIKVsg/HbB4bVsbFHd/c+49OIAJQ6ry8LKcUJpEZQHwAAZCyTBa/SVWXS+vaFZPgBAFAATXU1WhQMZNXHqJPWbNlB6QwAkry/LOR5AEiNoD4AAMiIlwWvUmkJBrTp0hYtbz7B134BAED6VrU1qsqy62PUSRu6+/0ZEICS5uVlIaU4gfQwnwUAAGTEr+y7T55zok6qna7Whlpu3AEAKAKtDbVa174g6xJ7PQPD6huK8vcdqGB9Q1GFwhHNnTVNvQPDSudXCqU4gfQR1AcAABnxY+GqlmBA13/8dB9GAwAA/LSiea7qZ03Thu5+9WQxMy8UjhDUBypQKBxRZ3d/xjN7KcUJZIagPgAASCqRYXNw5LBmVE9Ra0Nt1gtXkYEDAEBxa22oVWtDrfqGolr3/3Zp25MvZNzHr/teYDYeUGG6egc9zfRpCQa0sq2RgD6QAYL6AADgCJNl2Jw+52jP/ZKBAwBA6Wiqq9F5Te/2FNTf9uQL2vbkC1oUDGgVwTqg7IXCkbQD+iZp2Zn1Ou1PjublH+ARC+UCAIC36eodVMfGnqRTZh/b/6qnflkMFwCA0pNtMH77wLA6Nvbo7t5nfBoRgGLU2d2fdoa+k/TM8CFd0hokoA94RFAfAAC8JZMMm3SZpJuWLVDX5eeQpQcAQIlpqqvRomAgqz5GnbR6yw6FwhGfRgWgmPQNRTOuoZ9YUBuANwT1AQDAWzLJsElHlUk3LVuoFc1z/esUAADk1aq2RlVZdn04J332rkcI7ANlyOu/a34fAN4R1AcAAOobiuqm+3ZlnGEjSQvrZ064n3I7AACUh9aGWq1rX5B1YD86clh/+X1K8QDl5uDI4by2A8BCuQAAVLTJFsRN10Xvn6O/+8R7FQpHdHDksGZUT2HBKwAAysyK5rmqnzVNG7r71ZPFfYMk/e3mHXrp0Ju6/Px5Po0OQCHNqPYWXvTaDgBBfQAAKlZX76Av9fMPjhxWU10NQXwAAMpca0OtWhtq1TcU1br/t0vbnnzBc1/r7t2tWdPeyYw+oAx4XTeL9bYA7yi/AwBABfJzQVwybAAAqCxNdTU6r+ndWffD4rlAefCyoHZLMEBSEJAFgvoAAFQgPxfEJcMGAIDK48fff+ekDd39PowGQKFlsqB2lUkr2xpzOyCgzBHUBwCgwvQNRbOqoT8WGTYAAFQmL5m5E+kZGFbfUNSHEQEopHQX1K4yaX37QhKDgCwR1AcAoML4Nc2dDBsAACpbJpm5k6EED1AeVjTP1aZLW9SS5IVfSzCgTZe2sJYG4AOK4AIAUGEOjhzOug8ybAAAQCIzd82WnXJZlPXz494EQHEYu6B2KBzRwZHDmlE9Ra0NtczwBXxEUB8AgAqT7cK2LcGAVrY1EtAHAABa0TxX9bOm6bN3PaKox+B8tvcmAIpPU10NQXwgh/jLCQBAhfEajP/rC+bpovfP4eYcAAC8TWtDrf7+4jP1l9/v8dweAACkj5r6AABUGC8L27UEA1q9dD4BfQAAMKHWhlrdtGxBxu1ODEzT/Y8d0J2hARbMBQAgTWTqAwBQRtKtXbmqrVEdG3s0mkb9WxbEBQAA6VjRPFcvH/q91t27O+02Tw8f0rd+0ffW54uCAa2izB8AAJMiqA8AQBkIhSPq7O7X9oHhI45N9HCcWNhu7Zadkwb2WRAXAABk4vLz5+mYaUd5Xjx3+8CwOjb2aH37Qi1vPsH/AQIAUAYovwMAQInr6h1Ux8aeCQP60h8fju/ufeZt+1c0z9WmS1vUkqQUT0swoE2XtvBADQAAMrKiea7umuQeI5VRJ63ZskOhcMTnkQEAUB7I1AcAoISFwpGU2fbSHx+O58yaekTGfmtDbdplewAAANIx0T3Gvz6yT08PH0qr/aiTNnT3M1sQAIAJENQHAKBE9Q1F9aU0AvoJkz0cN9XVEMQHAAC+S9xj9A1F31Y7Px09A8PqG4pyjwIAwDgE9QEAKDGT1c9PhYdjAABQCF5L6YTCEe5bAAAYh6A+AAAlpKt3MK1yO5Ph4RgAAOTbwZHDeW0HAEA5Y6FcAABKRLr181Ph4RgAAOTbjGpvOYVe2wEAUM4I6gMAUCI6u/uzDuhLPBwDAID887rgLQvlAgBwJIL6AAAUub6hqG66b5enGvoT4eEYAADkW1NdjRYFAxm1aQkGKBkIAMAESNUDAKBIZbMgbjI8HAMAgEJZ1daojo09ac08rDJpZVtj7gcFAEAJIlMfAIAi1NU7qI6NPb4G9Hk4BpAtM6s3s380s2fN7A0z22tm3zGzWRn2E4i32xvv59l4v/W5uraZXWtmLv7xp5mMF4A/Whtqta59gaps8vOqTFrfvpDZhUCB9Q1FdWdoQLd09+vO0ID6hqKFHhKAODL1AQAoMn4tiDsWD8cAsmVm8yQ9KOk4ST+VtFvSIkmrJC01s1bn3Itp9HNsvJ8mSb+U9GNJ8yVdIumjZnaOc+4pP69tZmdI+oqkg5JmZPJ1A/DXiua5qp81TRu6+9UzQfJCSzCglW2N3LMABTTZjOFFwYBW8W8UKDiC+gAAFJG+oajvAX0ejgH45DbFguornXO3JHaa2c2SPi/pRkmfTaOfbygW0L/ZOXf1mH5WSuqMX2epX9c2s2pJmyT1StojqSONMQLIodaGWrU21KpvKKpQOKKDI4c1o3qKWhtqKRMIFFhX7+CkzyPbB4bVsbFH69sXannzCfkdHIC3ENQHAKAI5KJ+/omBabrjk2fxcAwga/FM+SWS9kr67rjDX5X0GUkdZna1c+61SfqZoVhQ/TVJ1407fKukL0j6kJm9J5Gt78O110kKSnqvpGsm/UIB5FVTXQ33KUARSXfG8KiT1mzZoTmzppI4BBQINfUBACiwXNXP/0b7Ah6UAfhlcXy71Tk3OvaAcy4qKSRpmqSzU/RztqSpkkLxdmP7GZV0/7jrZXVtM/ugYiV61jrn+lOMDQCAitbZ3Z/2jOFRJ23o5k8rUCgE9QEAKCDq5wMoESfHt31Jjiee6pty0I+na5vZTEk/kPQbSRtSjGtSZvbIRB+KrQUAAEDJ6xuKZpxk1DMwzOK5QIFQfgcAgALKJBsmHdTPB5AjM+PbV5IcT+w/Jgf9eL32LZICki5wzvn4mxYAgPITCkc8t2N2MJB/BPUBACgQL9kwyfz1BfN00fvncEMNAJLMbJlitfuvTNTmz4Zz7swk13lE0hnZ9g8AQKEdHDmc13YAskNQHwCAAvGaDTNeSzCg1UupAAEgpxLZ8DOTHE/sfzkH/WTUxswCkv5eUrek76UYDwAAkDSj2luI0Gs7ANnxraa+mdWb2T+a2bNm9oaZ7TWz75jZrDTbX2BmLo2PE/waMwAAheRHVkuVSSvbGn0YDQBM6sn4NlnN/MQvomR177PpJ9M2cyXVSmqTNDr2WULSJ+Pn/CK+73MpxgsAQEXwWr6Tsp9AYfjyOs3M5kl6UNJxkn4qabekRZJWSVpqZq3OuRdTdLNX0vVJji2Q1C7pMefcM36MGQCAQss2q4UFcQHk0bb4domZVTnnRhMHzKxGUqukQ5IeTtHPw5Jel9RqZjXOubdW1zOzKklLxl3Py7VflLQxyfXPU+wlwL2SnpX0WIrxAgBQEZrqarQoGMioPGhLMED5T6BA/Jojc5tiAf2VzrlbEjvN7GZJn5d0o6TPTtaBc26vpOsmOmZm/xz/zzt8GCsAAHnXNxRVKBzRwZHDmlE9Ra0NtVkF41kQF0A+Oef2mNlWxYLuVyq2CG3C9ZKmS7rdOfdaYqeZzY+33T2mn4NmtknSZxS79796TD9XSTpJ0v1j6+Bneu14EtCnJ/o6zOwHigX1b3bO/Xv63wEAAMpb31BUJ9fNUO/AsNJZXZ4Zw0BhZR3Uj2fpL1Es0/674w5/VbEb9g4zu3rsTX4G/ddKukixjJ4fZTdaAADyKxSOqLO7f8KMl0XBgObPrtHuA9EJWk7sxGOn6Y6/OouMGACFcIVis3M3mFmbpF2SWiQtVqz0zTXjzt8V39q4/V+SdIGkL5jZ+yRtl3SKpI9Lel6xwH221wYAAGmY7HklGWYMA4XnR6b+4vh269ipsJLknIuaWUixoP/Zii1WlalPSnqXpB85517OZqAAAORTV++g1m7ZqdEkqS7bB4ZlikW70s2G+cZFCwjoAyiIeMb8WZJukLRU0kckPSepU9L1zrmX0uznRTM7R7EEoAslnatYyZw7JX3FObcvV9cGAAB/lOp5ZSLMGAaKgx9B/ZPj22SLYvUrFtRvkreg/mXx7e3pNjCzR5Icmu/h+gAAZCwUjqR1g+wUC+qbSW6Sc8mGAVAM4qVtLknz3PEZ+mOPDSu2/taqXFx7kj4+JelT2fQBAEA5SPd5RYo9r1x89onqOOdEEoyAIuFHUH9mfPtKkuOJ/cdk2rGZna/YS4PHnHMPZj40AAAKo7O7P+2MFydpfl2NZk49Sj0TTHslGwYAAACAnzJ9XukbihLQB4qIXwvl5spn4tt/yKSRc+7MifbHM/jPyHZQAABMpG8oqp/8dp8eefolbR/IrBLE7gNRbf38eZJ0xIK63DwDAAAA8EvfUDSjGvqS1DMwTGAfKCJ+BPUTmfgzkxxP7H85k07NLCBpmWIL5G7yNDIAAPIgFI7ohp89oSczWPA2WT+XtAa5UQYAAACQM6FwxHM7nlWA4uBHUP/J+LYpyfHG+DZZzf1kEgvk/pAFcgEAxaqrd1CrN+/0pa+DI4d96QcAAAAAkvH63MHzClA8/Ajqb4tvl5hZlXNuNHHAzGoktUo6JOnhDPtNLJCbUekdAADy5fYH9mjdvbt9629GdbFXxQMAAABQ6rw+d/C8AhSPqmw7cM7tkbRV0kmSrhx3+HpJ0yVtcs69lthpZvPNbH6yPs3sXEmniAVyAQBFqqt30NeAviQWwgUAAACQc16fO3heAYqHX6/YrpD0oKQNZtYmaZekFkmLFSu7c82483fFt5akP08L5AIAkA+hcMS3kjsJLcEA9SkBAACy1DcUVSgc0cGRw5pRPUWtDbXcYwHjNNXVaFEwkNFiuTyvAMXFl6C+c26PmZ0l6QZJSyV9RNJzkjolXe+ceyndvsxslqS/EAvkAgCKUFfvoL7y08d97bPKpJVtjalPBAAAwIRC4Yg6u/snDFIuCga0qq2RLGNgjFVtjerY2KNRl/pcnleA4uNbMSzn3DOSLknz3GQZ+oq/AJjq17gAAPBDKBzRtfc8poHIa6lPzkCVSevbF/KQCQAA4FFX76DWbtmZNDi5fWBYHRt7tL59oZY3n5DfwQFFqrWhVuvaF0z6b0fieQUoVqxwAQBACl29g1qzeafSSGLJSEswoJVkjQEAAHgWCkdSBiUladRJa7bs0JxZU7n3QkUbX6JqXfsCbXl0v3ommOXC8wpQvAjqAwAwiVA4ojVb/A3o//UF79FF76+nJiUAAECWOrv70yofIsUC+xu6+wlQoiKlKlF107IFOvTmH1iPAigRBPUBAEgiFI7os3c9IudjRL8lGNDqpaf41yEAAECF6huKZrTQpyT1DAyrbyhKsBIVJZ0SVf+5d1jr2xfqktZgfgcHwJOqQg8AAIBi1NU7qL/8fo+iI4d969NYYAoAAMA3oXAkr+2AUpRpiSr+fQClgaA+AADj3P7AHq3evNPXPk3STSwwBQAA4JuDHpMvvv+bAQKXqBheSlQBKH4E9QEAGKOrd1Dr7t3ta59zZ03VXZ9u0fLmE3ztFwAAoJLNqPZWUXj/y6+rY2OP7u59xucRAcUlmxJVAIobNfUBAIhLLIrrpy/8WRMldwAAAHIgmxmQiVIjc2ZNZSYlylY2JapYdwIobmTqAwAQ19nd79uiuDXVU/RPn24hoA8AAJAjTXU1WhQMeG5PqRGUO68lqry2A5A/BPUBAJC3qanJmEl/f/GZZH0BAADk2Kq2RlWZ9/aUGkE581qiyms7APlDUB8AUNH6hqK6MzSgdf9vly/9mbEgLgAAQL60NtRqXfuCrAL7LJqLcuX1mYRnGaD48eoNAFCRQuGI1t+7Wzv3v+Jbn++pna6vXXg6N8EAAAB5tKJ5rupnTdPqzTu076XXM25PqRGUq0SJqkxmJLcEA9TTB0oAmfoAgIpz48+f0F9+v8e3gP67plTppmUL9MsvXkBAHwAAoABaG2p16QeCntoe+v0ffB4NUDwyKVFVZWJNMKBEENQHAFSU5X//oO74zYCvff7jp5q1onmur30CAAAgM16TK773qz1afvtDlOFBWXp3zbv0sYXHK1Vcv8qk9ZQRBUoG5XcAABWjY2OPtu99ydc+v7mMG18AAIBi4KXUSML2gWF1bOzR+vaFWt58Qg5GB+RXKBxRZ3d/Wv8eWoIBrWxr5LkGKCEE9QEAFeH2B/boN/3+Zl+t/fB8HvoAAACKyKq2RnVs7NGoy7ztqJNWb9mhObOmEtxESevqHdTaLTsn/Xdgkv78vcfrqg82UkMfKEGU3wEAVIQN3f2+9WUWy9C//Px5vvUJAACA7LU21Gpd+4K0a4iP55z05Xse83dQQB6FwpGUAX1JcpJ+tuM5vRB9Iy/jAuAvgvoAgLK39fEDeu1NfxZAawkGdNelLWToAwAAFKkVzXO16dIWtQQDnto/FXlNXb2DPo8KyI/O7v60Z6qMOn+TnwDkD+V3AABlb8uj+7Jqv/jkd+u8pnertaGWqakAAAAloLWhVq0Ntbrpvl363q+eyrj9bdv2aEXz3ByMDMidvqFoxmtK9AwMq28oynMOUGII6gMAyl505HBW7dd+5BRucgEAAErQtKO8hT2eHj5EoBMlJxT2toZYKBzhZx0oMZTfAQCUvZpq7++wF9bP5AYXAACgRM3I4j7w1l+GfRwJkHsHPSYzeW0HoHAI6gMAyl77GfWe265eOt/HkQAAACCfWhtqPbf9t/961nPmM1AIXl9iZfPyC0BhENQHAJS9JafN9pStf9m5waweBAEAAFBYTXU1OjEwzVNbJxYRRWnx+uzCMw9QegjqAwBKWt9QVDfdt1uX3Lldl9y5XTfdt0t9Q9EjzrtqcUNG/S46aZau+eipfg0TAAAABXLF4nme2yYWEQVKQVNdjRYFAxm1aQkGKDcKlCCC+gCAkhQKR7T0O7/Wkm//Wt/71R5te/IFbXvyBX3vV09pybd/rQ9959dvmy59+fnztOyMOWn1fV5jre7+7H/L1dABAACQRyua5ypYO91z+00P7fVvMECOrWprVJWld26VSSvbGnM7IAA5QVAfAFByunoH9Zff79HuA8mzpp48ENXF3+/R3b3PvLXvW8vfp7Ufnp+0FM/0d75Daz88Xz+6tMX3MQMAAKBwvn7h6Z7b3vXw4NvuKYFi1tpQq3XtC1IG9qtMWt++kNI7QIliJQwAQEkJhSNavXlnWuc6Sas379CcWVPfulm9/Px5uvz8edr6+AFteXSfoiOHVVM9Re1n1GvJabNzOHIAAAAUSmtDrf78vcfr3/7ruYzbTnRPCRSzFc1zVT9rmjZ096tnYPiI4y3BgFa2NfLzDJQwgvoAgJJyw8+eyOj8xAJn429Yl5w2myA+AABABfmbDzZ6CupLsXvKa+95TNu+eIGvYwJypbWhVq0NteobiioUjujgyGHNqJ6i1oZaaugDZYCgPgCgZGx9/ICenKTkTjKJBc64eQUAAKhciUVEt0+QuZyOgchr6uod1IrmuT6PDMidproanoOAMkRNfQBA0QuFI1p++0P6zKZHsuoDAAAAlS2TRUQnctuv9vg3GAAAPCKoDwAoal29g7p4Y4/njKqEgyOHfRoRAAAASlViEVGvcf2nXzykvqHMZ44CAOAngvoAgKJ1+wN7tHrzTjmXfV8zqqk4BwAAgNgioheffaLn9swABQAUGkF9AEBR+so9j2ndvbt962/8QrkAAACoXB3neA/qMwMUAFBopC0CAIrK7Q/s0Xf+vU+v/37Utz4XzDmaxaEAAADwlqa6Gs0NTNPg8KGM2zIDFABQaGTqAwCKxtV3/07r7t3ta0BfktZ8+BRf+wMAAEDpu3LxPE/tmAEKACg0gvoAgKJw+wN7tPnR/b73e9m5QR68AAAAcIQVzXMVrJ2eUZuWYIAZoACAgiOoDwAoCrduC/ve52XnBnXNR0/1vV8AAACUh69feLrM0jvXJJ0QmKY7QwPqG4rmdFwAAEyGQnAAgILb+vgBRX1ccGxh/UytXjqfDH0AAABMqrWhVuvbF2jtlp0adZOf6yT96yP79K+PxD5fFAxoVVsj95wAgLwjqA8AKKi+oahu/kWfL32deOw03fFXZzElGgAAAGlb0TxX9bOmaUN3v3oGhtNut31gWB0be7S+faGWN5+QwxECAPB2BPUBAAURCkfU2d2v7Rk8OE3GTPrGRQsI6AMAACBjrQ21am2oVd9QVKFwRI/vf1WbH92nFMn7GnXSmi07NGfWVDL2kTeJn9ODI4c1o3qKWhtqeQ4CKgxBfQBA3nX1DqY1xTkTN7Uv5EEKAAAAWWmqq1FTXY2W3/5QyoB+wqiTNnT3cy+KnJssMYpyUEBlYaFcAEBedXb3afVmfwP6az88nynPAAAA8EXfUDTj2aQ9A8Msnouc6uodVMfGnqQ/m4lyUHf3PpPnkQEoBIL6AIC8uP2BPVpw3f369i/6fevTJH1z2UJdfv483/oEAABAZQuFI57abXroaZ9HAsSEwpG0ZjonykF5/RkGUDoovwMAyLmr7/6dNj+639c+5wamah0ldwAAAOCzgyOHPbXbtvt5n0cCxHR296c905lyUEBlIFMfAJBTtz+wx/eA/pJT6/Trv/0gN6oAAADw3Yxqb/mP+15+nRI88B3loABMhKA+ACCnbt0W9q2vmuopWvvh+fqHvzrLtz4BAACAsbJJHKHsCfzm9WeKn0WgvBHUBwDkzNbHDyjqcfryWPNn1+gfOs7Uzus+RP18ACggM6s3s380s2fN7A0z22tm3zGzWRn2E4i32xvv59l4v/V+XNvMjo4f+038/BEze97MtpvZ58xsupevH0BlaKqrUf0xUz219Vq6B0jG688UP4tAeSOoDwDwXd9QVHeGBvR/7t/tS38b/uf7teS02b70BQDwxszmSXpE0iWStkv6tqSnJK2S9JCZHZtmP8dKeijebk+8n+3xfh8xs/f4cO2ApM9I+oOkn0u6WdK/SKpJXM/Mjk73awdQeRbPf7endl5L9wDJeP2Z4mcRKG/8CwcA+CYUjqizuz/jmo+TaQkG1FRX41t/AADPbpN0nKSVzrlbEjvN7GZJn5d0o6TPptHPNyQ1SbrZOXf1mH5WSuqMX2dpltd+RtJM59zvx1/czO6S9Jfx87+ZxngBVKCOc07SpocHM27Hmk/wm9efKX4WgfLmW6a+X1Nx432dYWb/PzPbF+9ryMweMLO/8mu8AAB/dfUOqmNjj68B/SqTVrY1+tYfAMCbeKb8Ekl7JX133OGvSnpNUkeqsjZmNkNSR/z868YdvlXS05I+NDZb38u1nXN/mCigH/cv8S1/YAAk1VRXo0XBQEZtSEZBLvCzCGAivgT1/ZqKG+/rKkm9it24d0v6lqSfSHqHpI/4MV4AgL9C4YjWbtmpUedfn1UmrW9fSIYJABSHxfHtVufc6NgDzrmopJCkaZLOTtHP2ZKmSgrF243tZ1TS/eOu5+e1E/48vt2R5vkAKtSqtkZVWXrnkoyCXOJnEcB4fpXf8WUqrpktkbRB0i8k/cX4G30zO8qn8QIAfNTZ3e9rQL8lGNDKtkYC+gBQPE6Ob/uSHO9XLCmnSbHEnGz6UbyfrK9tZlMkXRv/NCDpXEnvk7RN0h2TjPNtzOyRJIfmp9sHgNLT2lCrde0LUiavkIyCXONnEcB4WQf105gO+xnFpsNe7Zx7LUV3/0fS65L+1/iAviRNMoUWAFAgfUNRX0vuXHZuUNd89FTf+gMA+GJmfPtKkuOJ/cfkoJ9srj1FsWeSsTZJusI5N5J0lAAQt6J5rupnTdOG7n71THDPSzIK8oWfRQBj+ZGpP+l0WDMLKRb0P1uTZO2Y2emSFkq6R9KwmS2WdKYkJ+l3kraN7x8AUHihcMS3vj5xZj0BfQCAb+KBezMzk/Qnkv5U0jpJ/2lmS51ze9Ps58yJ9scz+M/wabgAilRrQ61aG2rVNxRVKBzRwZHDmlE9Ra0NtdQtR86N/7n72oWnSxI/i0CF8yOo79dU3Ob49nlJv5J03rjjO82s3TkXTjUgpscCQP4cHDmcdR811VN01eIGXX7+PB9GBADIgUQ2/MwkxxP7X85BP1lf2znnJO2X9EMze1LSQ4otzPuxyYcLAH/UVFdD4BR5EwpH1NndP+Gs6EXBgFaRlQ9UND+C+n5NxT0uvr1UsRvuj0r6D0l1kr4i6WJJPzezBc65Nz2PFgCQlfGZIod+/wdP/QRrp+nkuhq1n1GvJafN9nmUAACfPRnfNiU5nliRL1miTzb9+HVtSZJz7mEze1nSBemcDwBAvnX1Dk5aP3/7wLA6NvZofftCLW8+Ib+DA1AU/Foo1w9V8e07JP0P59xD8c9fNbO/UizL/ixJyyT982QdMT0WAPw3WaaIF7d3nEWmEwCUjm3x7RIzqxpbFtPMaiS1Sjok6eEU/Tys2BparWZWM3YdLTOrUmyG79jr+XntsW2OlnTEGl4AABRaKBzRmi075SZZEFeSRp20ZssOzZk1lYx9oAJVpT4lJb+m4iaOHxgT0Jf01nTZn8Y/XZTh+AAAWerqHVTHxh7fAvotwQABfQAoIc65PZK2SjpJ0pXjDl8vabqkTc651xI7zWy+mb2t/KVz7qBiC9VOl3TduH6uivd/v3PuqSyvvcDMqsd/HWb2TsXK7lRJ+vkkXzIAeNI3FNWdoQHd0t2vO0MD6hvi/SEyc+09j6UM6CeMOmlDd39uBwSgKPmRqe/3VNyXkxx/Kb6dmt6wAAB+CIUjk079zFSVSSvbGlOfCAAoNldIelDSBjNrk7RLUoukxYrd618z7vxd8a2N2/8lxUrffMHM3idpu6RTJH1csfW1xgfuvVz7UkmXmFlI0tOKPWP8iWIzAWYr9uzxxbS+agBIA/XP4Yeu3kENRF5LfeIYPQPD6huKkjQFVBg/MvXfNh127AEPU3Ffk3SSmU2f4Pjp8e1AFmMFAGSos7vf14D++vaFPNAAQAmKZ8yfJekHigXUr5Y0T1KnpLOdcy+m2c+Lks6RtEFSQ7yfFkl3Sjozfp1sr/0v8Y8TJf0PSf9b0p9L2hP/7zOcc8+l95UDwORSzWpN1D+/u/eZPI8Mpea72474E5iWUDji80gAFLusM/Wdc3vMbKtiWS9XSrplzOHEdNjbx0/FjbfdPaafQ2a2UdJKSV83sy/Ey+7IzBZI+pSkw5L+NdsxAwBS6xuK6ie/3edryZ2VZCgBQElzzj0j6ZI0zx2foT/22LCkVfGPXFw7JCmUbt8A4FW6s1qpf45U+oaiGhw+5KntwZHDPo8GQLHza6Fcv6biflnSeZI+J+mc+HTZOkntkqolfW6izB0AgH9C4YjW3btLj+1/1XMff33Be3RcTbUOjhzWjOopam2oZTooAAAAyk4ms1oT9c8J6mMi2WTbz6j2K7wHoFT48q8+nq1/lqQbJC2V9BFJzyk2HfZ659xLk7Uf08+rZnaupLWSPqHYYlmvS/oPSX/nnNvqx3gBABO78edP6I7fZF/lbNpRU3RJa9CHEQEAAADFqW8omvGsVuqfI5mH9qRVxW5CvCgCKo9vr/J8nIp7ULHM/vHZ/QCAHPIroC+RKQIAAIDy5zWzOhSOENTH24TCEf3iiSFPbU88dho/T0AF8mOhXABAiQuFI74F9CUyRQAAAFD+vNYxf/xZ72UuUZ46u/uVZhWnI1xxwTxfxwKgNBDUBwBo/b27U5+UppZggEwRAAAAlD2vs1M3P7JPd/c+4/NoUKq8lHFKCNZO14rmuT6PCEApIKgPABWubyiqnftf8aWvKpNWtjX60hcAAABQzLzOTnWS1mzZkdXCqCgf2fwcfP3C030cCYBSQlAfACpU31BUd4YGtO7/7fKlvyqT1rcvpPQOAAAAKkJTXY0WBQOe2o46aUN3v88jQinyWsbpQ6fW8ewFVDBWMgSAChMKR7T+3t2+ZedLsZI7K9sauakEAABARVnV1qiOjT0a9VAQvWdgWH1DUUpXVjivZZzOnneszyMBUEoI6gNABbnx50/4uiDuX18wTxe9fw4PIgAAAKhIrQ21Wte+QGs27/S00GkoHOFeusJ5TYwioQqobJTfAYAKcfXdv/M1oL9gztFavXQ+DyEAAACoaCua52rZGfWe2notvYLy0VRXo9PnHJ1Rm5ZggOcwoMIR1AeAMhcKR7T0O7/W5kf3+9rvmg+f4mt/AAAAQKk6LcOgbILX0isoD6FwRMtvf0iP7X817TZVJq1sa8zhqACUAv56AEAZ6+od1NotOz3V+JzMZecGme4JAAAAxFFCBZny8qxWZdL69oX83AAgUx8AylUoHMlZQP+aj57qb6cAAABACWuqq9GiYCCjNpRQqVxentVaggFturRFy5tPyN3AAJQMMvUBoEx1dvf7GtBfWD9Tq5fOJysEAAAAmMCqtkZ1bOxJ6x6cEiqVLdNntYX1M9V1+Tm5GxCAkkNQHwDKUN9QVNsHhrPuZ/HJ79Z5Te9Wa0MtWUQAAADAJFobarWufUHKDGxKqFQ2L89qO/a9or6hKM9kAN5CUB8AylAoHPGln7UfOYUbRwAAACBNK5rnqn7WNG3o7lfPBIHblmBAK9saCehXsE0PPe2pXSgc4dkMwFsI6gNAmegbiioUjujgyGE9OvhS1v0trJ/JTSMAAACQodaGWrU21L7t/nxG9RRmv0KS9Mvdz3tqd3DksM8jAVDKCOoDQIkLhSNad+8uPbb/VV/7Xb10vq/9AQAAAJWkqa6GID7epm8oqv0vv+6p7YxqQngA/ojfCABQwm78+RO64zcDvvd72blBpgQDAAAAgI+yKZPK8xmAsQjqA0CJuvru32nzo/t97/eyc4O65qOn+t4vAAAAgD+iPE/l8VpCp37WVH42ALwNQX0AKDGhcESrN+/Qvpe8TdtMJnjsNH39ogVkgAAAAAA5FApH1Nndr+0TLKS7KBjQKhbSLVteS+gsPvndPo8EQKkjqA8AJaSrd1BrNu+U87FPk/T5P2vSyrZGH3sFAAAAMF5X76DWbtmp0SQ39NsHhtWxsUfr2xdqefMJ+R0ccm7aO9/hqV3HOSf5OxAAJY+gPgCUiFA44ntAvyUY0EoygQAAAICcC4Ujkwb0E0adtGbLDs2ZNZX79DKSeKGTqZZggNI7AI5AUB8ASkRnd3/WAf3G46brv793DjU7AQAAgDzr7O5PGdBPGHXShu5+gvplIt0XOuNVmZhRDWBCBPUBoAT0DUUnrLmZqfpZ0/Q33BQCAAAAeeXlfr5nYFh9Q1ESccpAJi90EqpMWt++kBc7ACZUVegBAABSC4UjvvQz//ijfekHAAAAQPq83s/79RyAwvGaoLWufQHrKgBIiqA+AJSAgyOHfennovfP8aUfAAAAAOnzej/v13MACsfri5lDb/7B55EAKCcE9QGgBMyozr5a2imza5i6CwAAABSA1/t5P54DUFgP7XnRUzte6ACYDEF9ACgBftRRvPZjp/owEgAAAACZ8no/Tz310hYKR/SLJ4Y8teWFDoDJENQHgBLQVFejRcGA5/bfXMYCSwAAAECheLmfbwkGmGlb4jq7+5Xh+rhv4fkNwGQI6gNAEegbiurO0IBu6e7XnaEB9Q1FjzhnVVujLMN+T5g1Vf/06RYWWAIAAAAKbFVbo6rSvKGvMmllW2NuB4Sc8rpArsQLHQCpMZcHAAooFI5o3b279Nj+V484tigY0Kq2xrcyNFobarV+2QKt2bwzrWyPL/xZEw8CAAAAQJFobajVuvYFWrtlp0YnuaGvMml9OzNtS53XBXJNvNABkBpBfQAokBt//oTu+M1A0uPbB4bVsbFH69sXvpVpv6J5rupnTdPXfvaEdh84Mptfii2Ie+3HTuUhAAAAACgyifv5Dd396pkgi7slGNDKMYk9KF1eF7pdcmod//8BpERQHwAK4Oq7f6fNj+5Ped6ok9Zs2aE5s6a+LWP/vs+dp76hqH7y233a/VwsuD//+KN10fvnME0TAAAAKGKtDbVqbahV31BUoXBEB0cOa0b1FLU21HIvX0a8LnR79rxjfR4JgHJEUB8A8igUjmj15h3a99LrabcZddKG7v4jsjWa6mq0eukpfg8RAAAAQB401dUQxC9jXrPtydIHkA4WygWAPOnqHdTF3+/JKKCf0DMwPOHiuQAAAACA4tNUV6NFwUBGbVggF0C6COoDQB6EwpG0F7idrA8AAAAAQGlY1daoKkvv3CpjgVwA6SOoDwB50Nndn1VAX/K+0BIAAAAAIP9aG2q1rn1BysB+lUnr2xdSegdA2qipDwA51jcU1faB4az78brQEgAAAACgMFY0z1X9rGna0N2vngmeC1uCAa1saySgDyAjRIgAIMf8KpvDTR4AAAAAlJ7Whlq1NtSqbyiqUDiigyOHNaN6ilobaqmhD8ATgvoAkANjb9YeHXwp6/4WzDmamz0AAAAAKGFNdTU81wHwBUF9APBRKBxRZ3e/L+V2xlrz4VN87Q8AAAAAAACliaA+APikq3dQa7fs1Gi2K+KOc9m5QUrvAAAAABWKki0AgPEI6gOAD0LhSM4C+td89FR/OwUAAABQ9CabBbwoGNAqFlcFgIpVVegBAEA56Ozu9z2gv/bD8wnoAwAAABWoq3dQHRt7kpb13D4wrI6NPbq795k8jwwAUAwI6gNAlvqGor7X0P/msoW6/Px5vvYJAAAAoPilOwt41ElrtuxQKBzJz8AAAEWD8jsAkCU/b6JPmV2jaz92KtNoAQAAgAqVySzgUSdt6O7n+aFIsP4BgHwhqA8AWTo4cthTu0UnzdL0d8V+Dc8//mhd9P453PABAAAAFczLLOCegWH1DUV5ligg1j8AkG8E9QEgAxNlXsyo9var9MMLjtclrUGfRwgAAACgVHmdBRwKRwjqF0hX7+Ck5ZIS6x+sb1+o5c0n5HdwAMoWQX0ASMNkmRcL5sz01CeZGgAAAADG8joL2Gs7ZCfT9Q/mzJrKcyAAX7BQLgCk0NU7qI6NPUmnwe7c/0rGfbYEA2TSAAAAAHgbr7OAvbZDdrysfwAAfiCoDwCTSDfzIhNVJq1sa/SvQwAAAABlwWsWN9nf+ZfN+gcAkC3fgvpmVm9m/2hmz5rZG2a218y+Y2azMujjV2bmJvmo9mu8AJCOTDIv0lFl0vr2hdx0AwAAADhCU12NFgUDGbVhFnBhZLP+AQBky5f5WWY2T9KDko6T9FNJuyUtkrRK0lIza3XOvZhBl9cn2U+ROAB54yXzQpIWzDlaO/e/esT+lmBAK9saCegDAAAASGpVW6M6NvaklVzELODC2fvia3ltBwBj+VV07TbFAvornXO3JHaa2c2SPi/pRkmfTbcz59x1Po0LADzzmkHRfka9vrW8VqFwRAdHDmtG9RS1NtSSPQMAKHlmVi/pBklLJR0r6TlJ90i63jn3Ugb9BCR9RdKFko6X9KKk+yR9xTm3L9trm9kcSe2SPiLplPg1Dkp6VNL3nHNb0h0rAORba0Ot1rUvSFkGlFnAhfWLx4c8tRt6ZcTnkQCoRFkH9eNZ+ksk7ZX03XGHvyrpM5I6zOxq5xyvIwEUrb6h6NsC8V4zKA6OHFZTXQ1BfABAWfFrdq6ZHRvvp0nSLyX9WNJ8SZdI+qiZneOceyrLa/+NpNWSBiRtk3RA0omKBfr/1My+7Zz7gqdvBADkwYrmuaqfNU0buvvVM8HsYWYBF9btD+zRsx6D83VHU1kaQPb8yNRfHN9udc6Njj3gnIuaWUixoP/ZkrrT6dDMVkgKSnpT0i5Jv3TOveHDWAHgCKFwRJ3d/Z5K7UxkRrVfk6AAACgqfs3O/YZiAf2bnXNXj+lnpaTO+HWWZnnt7ZIucM49MLYTMztF0sOSPm9m/+SceySN8QJAQbQ21Kq1ofaI5CNmARferdvCntueVDvdx5EAqFR+RJ5Ojm/7khzvVyyo36Q0g/qKZeuM9byZXemc+9d0GptZspvz+WleH0CF6Ood1JrNO+XjWrhkywAAyo5fs3PNbIakDkmvSbpu3OFbJX1B0ofM7D2JbH0v105WXsc5t8vMuiRdJukCSQT1ARQ9ZgEXl62PH1B0xPuSjzwvAvBDlQ99zIxvX0lyPLH/mDT6+qmkP5dUL2mqYkH4dfG2XWY2PmMHADy7/YE9Wu1zQL8lGOCGGwBQjiadnSspJGmaYrNzJ3O2Yvf5oXi7sf2MSrp/3PX8vHbC7+PbtCMyZvbIRB8iaQgAKs6WRydc+iUtPC8C8IsfQX3fOOe+7Zz7mXNuv3NuxDn3pHPuS5KuVmys69Ls58yJPhSrvQkA6uod1Lp7/f2VUGXSyrZGX/sEAKBIpDM7V4rNzvW7H7+uLTM7WtIySU7S1lTnAwAwXjZZ+jwvAvCLH+V3Epn4M5McT+x/OYtrfF/StyW9z8xqxmf1AEAmQuGI1mzZ6WufVSatb1/IVEoAQLnya3aul358ubaZmWLPFXWSbnPO7Zrs/LHiCUIT9fmIpDPS7QcAUPpqPK6hdvqfHM3zIgDf+JGp/2R8mywzJvEaMllmTUrOuRFJiUA+K4oAyEpnd7+cjzV3WoIBbbq0RcubT/CvUwAA4LdvSfqEpN8oVrsfAICMtZ9R76kdWfoA/ORHpv62+HaJmVWNrXNpZjWSWiUdkvSw1wuY2cmSZikW2I9kMVYAFa5vKKrtA8Oe23/ynBN1Uu10HRw5rBnVU9TaUEtNRABAJfBrdq6XfrK+tpl9U9LnJf1a0kedc2+kGCcAABNactps1VRPyagMT031FC05bXYORwWg0mQd1HfO7TGzrZKWSLpS0i1jDl+vWGb97c651xI7zWx+vO3uMfuCkl5xzr0t2mZm75Z0Z/zTHzvnvBcvA1DxQuHs3gueVDtdl7QGfRoNAAAlw6/ZuV76yeraZvZtSZ9TLBnpY865QynGCAAlp28oqlA4QvJRnly1uCGjNdquWtyQw9EAqER+ZOpL0hWSHpS0wczaJO2S1CJpsWI319eMOz9Rv9LG7Dtf0t+b2X9IekrSsKS5kj6iWPbNf0r6W5/GC6BCjL+53fvia6kbTYIaiACACuXX7NyHJb0uqXX8WllmVqVYotDY63m+dryG/q2KPav8QtLHnXOvp/PFAkCpCIUj6uzun3A28qJgQKvaGnmGyYHLz5+nvqGoNj+6P+W5nzizXpefPy8PowJQSXwJ6sez9c+SdIOkpYoF4p+T1CnpeufcS2l084ikH0s6U9L7JR2tWLmdnZLuVizb/00/xgug/HX1Duq72/ZocNi/ZLyWYIBsFwBARfJrdq5z7qCZbZL0GUnXSbp6TD9XSTpJ0v3OuaeyvLZJ+gdJn5Z0r6T2+DpdAFA2unoHtWbzTiVbLmz7wLAu/n6Pblq2kPW/cuBby9+nproa3botPGEpnprqKbpqcQMBfQA54Vemvpxzz0i6JM1zbYJ9OyV9yq/xAKhMoXBE197zmAYi2WXkj2diYSMAQMXzY3auJH1J0gWSvmBm75O0XdIpkj4u6XnFAvfZXvsrigX0X5f0O0lrYnH+t/mdc+6e5F8uABSvUDgyaUA/wUlavXmH5syaSsZ+Dlx+/jxdfv48bX38gLY8uk/RkcOqqZ6i9jPqqaEPIKd8C+oDQKF19Q5qzZadcqnubD24adlCboIBABXNp9m5cs69aGbnSPqqpAslnSvpRcXW0fqKc26fD9dOLIAzVdLaJEP5oaR70hkzABSbG372RMqAfoKT9LWfPaH7PndeLodU0ZacNpsgPoC8IqgPoCyEwhGtzVFAf+2H5zNdFQAAZT87d8yxYUmr4h+5uPanxCxgAGWqbyiqJw9EU584xu4DUfUNRSknCgBloqrQAwCAbPUNRfWlLTs16nNA30z65rKF1EAEAAAAUDR+8tsjJjTltB0AoPiQqQ+gZHX1Duq2bXv0tI+L4Sa0BANa2dZIyR0AAAAARWX3c5ll6WfbDgBQfAjqAyg5fi+G+8lzTtRJtdN1cOSwZlRPUWtDLdNSAQAAAJSV5155vdBDAAD4hKA+gJLS1Tuo1Zt3+trnSbXTdUlrMPWJAAAAAFBg848/WtuefCHjdk8eOKhQOMJsZAAoA9TUB1Aybvz5E74H9CVxUwsAAACgZFz0/jme2jlJG7r7/R0MAKAgyNQHUPS6egd1031Pavi1N33vuyUYoNQOAAAAgJLRVFej+bNrtPtA5jXyewaG1TcU5RkIAEocmfoAilYoHNHiv/uVVm/emZOAfpVJK9safe8XAAAAAHLpyx871XPbTQ897eNIAACFQFAfQFHq6h3UxRt7fFsMdzyTtL59IaV3AAAAAJSc1oZaLTm1zlPbbbuf93k0AIB8I6gPoOiEwhGt3bJTzuWm/2DtdN316RYtbz4hNxcAAAAAgBw7Z96xntrte/l19Q1lXroHAFA8qKkPoOisu3eXRnMQ0D/x2Gm64oJ5WtE81//OAQAAACCPspl1HApHqKsPACWMoD6AohEKR7T+3t16bP+rvvZ7YmCa7vjkWdy0AgAAACgbTXU1qj9mqva9/HrGbQ+OHM7BiAAA+UJQH0DB9Q1Fdesv+/Vv//Wc/E7QrzLpG+0LCOgDAAAAKDuL579bmx4ezLjdjGrCQQBQyvgtDqBgQuGIOrv7tX1gOCf9m7EYLgAAAIDy1XHOSZ6C+jwjAUBpI6gPoCC6ege1dsvOnNTOl6T31E7X1y48nZtVAAAAAGWrqa5Gi4KBjBKlWoIBZjIDQImrKvQAAFSeUDiS04D+pz8Q1C+/eAEBfQAAAABlb1Vbo6osvXOrTFrZ1pjbAQEAco6gPoC86+zuz1lA/5vLFuraj52am84BAAAAoMi0NtRqXfuClIH9KsqTAkDZoPwOgLzqG4rmpIZ+sHa6vk65HQAAAAAVaEXzXNXPmqYN3f3qmeB5qyUY0Mq2Rp6XAKBMENQHkFehcMTX/o6d/k797dKTtaJ5rq/9AgAAAEApaW2oVWtDrfqGogqFIzo4clgzqqeotaGWGvoAUGYI6gPIq4Mjh33p5/iZ1frcnzYSzAcAAACAMZrqagjiZ4GXIgBKAUF9AHk1ozr7XzsL62fq/171AR9GAwAAAABAbFZ5Z3f/hOViFwUDWkX5IgBFhIVyAeRVtjdBVSatXjrfp9EAAAAAACpdV++gOjb2JF3/bfvAsDo29uju3mfyPDIAmBhBfQB51VRXo0XBgKe2VSatb19IdgQAAAAAwBehcERrt+zUqJv8vFEnrdmyw/d14gDAC4L6APJuVVujqiyzNi3BgDZd2qLlzSfkZlAAAAAAgIrT2d2fMqCfMOqkDd39uR0QAKSBmvoA8q61oVbr2hekzIYwSX/+3uN11QcbWZgIAAAAAOCrvqFo0pI7yfQMDKtvKMozKoCCIqgPoCBWNM9V/axp2tDdr54JbqJaggGtZCEiAAAAAECOeC2lEwpHCOoDKCiC+gAKprWhVq0NteobiioUjujgyGHNqJ6i1oZabpAAAAAAADn128GXPLe7pDXo82gAIH0E9QFkJBcB+Ka6GoL4AAAAAIC86hs6mNd2AOAXgvoA0hIKR9TZ3T9hvcFFwYBWUSoHAAAAAFBCqo+qyms7APALv4UApNTVO6iOjT1JFxDaPjCsjo09urv3mTyPDAAAAAAAb2YfXZ3XdgDgF4L6ACYVCke0dstOjbrJzxt10potOzwvNAQAAAAAQD61n1Gf13YA4BeC+gAm1dndnzKgnzDqpA3d/bkdEAAAAAAAPlhy2mzVVGdWmbqmeoqWnDY7RyMCgPRQUx/A24xdCPfQ7w8nLbmTTM/AsPqGoix8CwAAAAAoelctbtC6e3dndD4AFBpBfQCSYnXzb9u2R08PH8q6r1A4QlAfAAAAAFD0Lj9/nvqGotr86P6U537izHpdfv68PIwKACZHUB+ocKFwRNfe85gGIq/51ufBkcO+9QUAAAAAQC59a/n71FRXo1u3hRWd4Hm2pnqKrlrcQEAfQNEgqA9UsK7eQa3evNP3fmdkWJMQAAAAAIBCuvz8ebr8/Hna+vgBbXl0n6Ijh1VTPUXtZ9RTQx9A0SHyBlSo2x/Yk1HdwEy0NtTmpF8AAAAAAHJpyWmzCeIDKHoE9YEK1NU7mLOAfkswQD19AAAAAEBR6xuKKhSO6ODIYc2onqLWhlqeZQGUDIL6QIUJhSNas8X/kjuSVGXSyrbGnPQNAAAAAMidSglyh8IRdXb3a/vA8BHHFgUDWtXWyOxzAEWPoD5QIRI3aN//zYCc87//KpPWty/k5gcAAAAASkglBbm7ege1dstOjSZ5Jt4+MKyOjT1a375Qy5tPyO/gACADBPWBMjfZDZpfWoIBrSyjGz0AAAAAqARdvYNas3mnkuV9bR8Y1sXf79FNy0o/yB0KRyYN6CeMOmnNlh2aM2sqz7gAihZBfaCMpcpC8OqvL5inaUe9o6ynZAIAAABAOQuFI5MG9BOcpNWbSz/I3dndn/az8aiTNnT3l/TXC6C8EdQHykyizM7jz76qzY/sS3mDlqmWYECrl873uVcAAAAAQD7d8LMn0n5edJK+9rMndN/nzsvlkHKmbyia8ez1noFh9Q1FSWIDUJQI6gNlIh9ldlgIFwAAAABKX99QVE8eiGbUZveBaMkGuUPhiOd2pfj1Aih/VYUeAIDsdfUOqmNjT04D+sZCuAAAAABQFn7y2315bVdoB0cO57UdAOQamfpAiUt3sZ9svKd2ur524ekE9AEAAACgDOx+LrMs/WzbFdqMam/hL6/tACDX+O0ElLhMFvvx4rJzg7rmo6fm7gIAAAAAAOTQqyO/99SOxDYAxcq38jtmVm9m/2hmz5rZG2a218y+Y2azsujzPDP7g5k5M/u6X2MFyoWXxX7SZSZ9c9lCAvoAAAAAUGbmH3+0p3almLkeCkfU+e/9GbdrCQaopw+gaPny29jM5kl6UNJxkn4qabekRZJWSVpqZq3OuRcz7LNG0g8lHZI0w49xAqWsbyiqUDiigyOHNaN6ilobaj0v9pNKSzCglW2NZCUAAAAAQBm66P1z9L1f7cm43c/+6zmd2/CMljefkINR5YaX2e1VJq1sa8zNgADAB369Yr1NsYD+SufcLYmdZnazpM9LulHSZzPss1PSTEnr4u2BihQKR9TZ3T9hRv6cY6b6dp36WVN16QeCam2oJRsBAAAAAMpYU12N5s+u0e4DmdXId5JWb96hObOmlkQSmNfZ7Z/706aS+PoAVK6sy+/Es/SXSNor6bvjDn9V0muSOsxsegZ9flzSJZJWSno22zECpaqzu08Xf78n6U3I/pdf9+U6VSbdtGyhLmkNEtAHAAAAgArw5Y+dKvPQzkn6+s+e8Hs4OeF1dntNCZYZAlBZ/Kipvzi+3eqcGx17wDkXlRSSNE3S2el0ZmbHSbpD0j3Oubt8GB9QckLhiJZ+59f69i/6lcM1cCXFAvrr2xeShQAAAAAAFaS1oVbrly3w1HbXgaj6hjLL8i+EgyOH89oOAPLFj6D+yfFtX5LjidVImtLs7w7FxpVpuZ63mNkjE31Imu+1TyBfunoH1bGxJ+NpkF60BAPadGlLSdVDBAAAAAD4Y0XzXP339x7vqe1Pfrvf59H47993DXlqV4oLAgOoLH78lpoZ376S5Hhi/zGpOjKz/0/Sf5e0wjnn7TcvUMJC4YjWbtmZ8SI+6TJJy86o12lzjqZ2PgAAAABAB155w1O7R/a+5PNI/HX7A3v0X/uShaomx0x2AMXOj0x9X5jZSZK+I+lfnHN3Z9OXc+7MiT4k7fZhqEDOdHb35yygn6ib/3fL30vtfAAA4ImZ1ZvZP5rZs2b2hpntNbPvmNmsDPsJxNvtjffzbLzfer+ubWaXmtntZtZjZofMzJnZ1zP9mgGg3L10yFtQ32u7fLl1W9hTu5ZggOdlAEXPj0z9xGvPmUmOJ/a/nKKff5T0uqQrfBgTUHL6hqJJF8RNV/2sqdr30pGL57YEA1rZ1ki2AQAA8MzM5kl6UNJxkn6qWMLMIkmrJC01s1bn3Itp9HNsvJ8mSb+U9GPFymReIumjZnaOc+4pH679LcWeRV6S9KykeV6+bgAod7OmvVPSax7bFaetjx9Q1ENdfJO0sq3R/wEBgM/8COo/Gd8mq5mf+G2YrOZ+whmK3XS/YDbh+uvXmNk1kn7qnLsw00ECxS4UjmTdx6UfCKq1oVahcEQHRw5rRvUUyuwAAAC/3KZYUH2lc+6WxE4zu1nS5yXdqPTWxfqGYs8ONzvnrh7Tz0pJnfHrLPXh2v9D0i7n3NNm9ilJd6YxNgCoOGeeNEvbPZTSOfOkjCZp5dWWR/d5anfanxxNMhyAkuBHUH9bfLvEzKqcc6OJA2ZWI6lV0iFJD6fo50eSpk2wv1HSeZJ+J+kRSb/NdsBAMTroIYtgvEQAnyA+AADwUzxTfomkvZK+O+7wVyV9RlKHmV3tnEua7mlmMyR1KJYSet24w7dK+oKkD5nZexLZ+l6v7Zy7L4MvEQAq1kXvr9f3fvVU6hMnaFesvGTpS9LMqUf5PBIAyI2sa+o75/ZI2irpJElXjjt8vaTpkjaNvcE2s/lmNn9cPyudc58e/6E/ZtT8PL5v/I08UBZmVGf3jo26fwAAIIcWx7dbxybxSJJzLioppFiCztkp+jlb0lRJoXi7sf2MSrp/3PX8vDYAYAJNdTU6eXZmz5LzZxd3MlmNx+drr+0AIN/8Wij3CknPS9pgZveY2Toz+6ViU2H7JF0z7vxd8Q8AcdlM8asy6v4BAICcOjm+TVZSsz++TVaSM5t+/Lq2Z2b2yEQfiq0FAAAl7ysfO1UTFkKegEn68sdOzeVwstZ+hrdZBF7bAUC++RLUj2frnyXpB5JaJF2t2EJUnZLOTmfBLKAc9Q1FdWdoQLd09+vO0ID6hqJJz22qq9GiYCDja1SZtL59IXX/AABALs2Mb19Jcjyx/5gc9OPXtQEASbQ21Gr9sgUpA/sm6aZlxf/8ueS02Rln3ddUT9GS02bnaEQA4C/f5hU5556RdEma56b7AljOuR8o9rIAKBmhcESd3f3aPjB8xLFFwYBWtTVOeBO0qq1RHRt7NOrSu8782TX68sdOLfobKgAAgFLmnDtzov3xbP0z8jwcAMiJFc1zVT9rmjZ096tngmfZlmBAK5M8yxajqxY3aN29uzM6HwBKBcXCAJ919Q5q7ZadSQPz2weG1bGxR+vbF2p58wlvO9baUKt17QsmbS/FsiM+/2dNlNwBAAD5ksiGn5nkeGL/yznox69rAwBSaG2oVWtDrfqGogqFIzo4clgzqqeotaG2qGvoT+Ty8+epbyiqzY/uT3nuJ86s1+Xnz8vDqADAHwT1AZ/0DUW16aGnddfDTytVov2ok9Zs2aE5s6YekeVQbtkRAACgLDwZ3yarW5/INEhW9z6bfvy6NgAgTU11xb0Qbrq+tfx9aqqr0a3bwoqOHD7ieE31FF21uIGAPoCSQ1AfyNJkpXYmM+qkDd39Ewbnyyk7AgAAlIVt8e0SM6tyzo0mDphZjaRWSYckPZyin4clvS6p1cxqnHNvLThkZlWSloy7np/XBgBUoMvPn6fLz5+nrY8f0JZH9yk6clg11VPUfkY9NfQBlCyC+kAWUpXaSaVnYFh9Q9GkgfpyyY4AAAClzTm3x8y2KhZ0v1LSLWMOXy9puqTbnXOvJXaa2fx4291j+jloZpskfUbSdZKuHtPPVZJOknS/c+6pbK4NAMB4S06bTRAfQNkgqA94FApHsgroj+2HwD0AACgBV0h6UNIGM2uTtEtSi6TFipW+uWbc+bviWxu3/0uSLpD0BTN7n6Ttkk6R9HFJzysWuM/22jKzT0v6QPzTxOqHf25m9fH/3u2cWz/pVwwAAAAUIYL6gEed3f1ZB/Ql6eAEdf0AAACKTTxj/ixJN0haKukjkp6T1CnpeufcS2n286KZnSPpq5IulHSupBcl3SnpK865fT5d+wOSPjlu38L4hyQ9IImgPgAAAEoOQX3Ag76haMY19JOZUc0/QwAAUBqcc89IuiTNc8dn6I89NixpVfzD92vHz/+UpE+lez4AoPSxLh2ASkE0EZhEshuCUDji2zUmWigXAAAAAACkJxSOqLO7f8Lku0XBgFa1NfLsDaCsENQHJpDqhmBuYJov12kJBsgaAAAAAACUnGLJiu/qHZx0vbvtA8Pq2Nij9e0Ltbz5hPwODgByhKA+ME46NwS9PpTeqTJpZVtj1v0AAAAAAJAvxZQVHwpHJn1+Txh10potOzRn1lQy9gGUhapCDwAoJuneEGS7Pm6VSevbF3IzAQAAAAAoGV29g+rY2JN0jblEVvzdvc/kZTyd3f0pn98TRp20obs/twMCgDwhqA+MkckNgVctwYA2XdrCtD8AAAAAQMnINCvez7XoJtI3FE36ciGZnoFh9Q1FczQiAMgfyu8AcV5uCNJlki4+e646zjmJGvoAAAAAgJLjJSs+l7PTvb40CIUjPJcDKHkE9YE4rzcEpsnL8SRK7ZCZDwAAAAAoRdlkxecqgH5w5HBe2wFAMaH8DhDn9Q/7sjPq1RIMTHiMUjsAAAAAgFKXTVZ8rsyo9pan6rUdABQTfpMBcV7/sJ8252j93fL3qm8oqlA4ooMjhzWjeopaG2qZ0gcAAAAAKHnFmBXvtbRPLksCAUC+ENQH4rK9IWiqqyGIDwAAAAAoO8WYFd9UV6NFwUBGZYFaggGe2wGUBcrvAHGJG4JMcEMAAAAAACh3XpPgojmuX7/sjDmyNM+tMmllW2NOxwMA+UJQHxhjVVujqtK8I+CGAAAAAABQCbwkwUnSt3/Rl5O6+qFwRMtvf0irN++US+P8KpPWty+k9A6AskFQHxijtaFW69oXpAzsc0MAAAAAAKgkmSTBJThJ197zmK/j6Ood1MUbe9Iuu9MSDGjTpS1a3nyCr+MAgEKipj5KUi4XpV3RPFf1s6ZpQ3e/eia4SWgJBrSyrZGAPgAAAACgYrQ21GrVnzbq27/oz6jdQOQ1dfUOakXz3KzHEApHtCbN7HxJumnZAl+uCwDFhqA+SkooHFFnd/+Eb+QXBQNa5VOwvbWhVq0NtTl9eQAAAAAAQCk5uvooT+1u+9UeX4LrX/vZE2kH9CXpB6G9BPUBlCWC+igZXb2DWrtlp0aT/AXfPjCsjo09Wt++0LdpdU11NQTxAQAAAACQdNDjwrdPv3hIfUPRrJ6v+4ai2n0gmlGbXQeiWV8XAIoRNfVREkLhyKQB/YRRJ63ZsiMnC/EAAAAAAFDJZlR7zw3N9jn9J7/dn9d2AFDMCOqjJHR296cM6CeMOmlDd2Y1/gAAAAAAwOSyKXfrNcs/Yfdzr+a1HQAUM4L6KHp9Q9G0V7VP6BkYVt9QZtPyAAAAAABAck11NZobmOapbTZZ/gCAtyOoj6LndYoeJXgAAAAAAPDXlYvneWqXTZa/JM0/3ltdfK/tAKCYEdRH0fM6RS/bqX0AAAAAAODtVjTPVbB2ekZtWoKBrBervej99XltBwDFjKA+ip7XKXpM7QMAAAAAwH9fv/B0maV3bpVJK9sas75mU12NTp6d2YuB+bNrsn6ZAADFiKA+ip7XKXrZTu0DAAAAAABHam2o1fr2BapKEdivMml9+0Lfns+/8rFTlea7BJmkL3/sVF+uCwDFhqA+il5TXY0WBQMZtfFjah8AAAAAAJjYiua52nRpi1qSPK+3BAPadGmLljef4Ns1WxtqtX7ZgpSBfZN00zL/XiYAQLGhPglKwqq2RnVs7NGoS32uX1P7AAAAAABAcq0NtWptqFXfUFShcEQHRw5rRvUUtTbU5izRbkXzXNXPmqYN3f3qGRg+4nhLMKCVbY0E9AGUNYL6KAmtDbVa175Aa7fsnDSw7/fUPgAAAAAAMLmmuvzWri/EywQAKCYE9VEyeBsPAAAAAAAS8v0yAQCKBUF9lBTexgMAAAAAAACoZAT14Umhg+q8jQcAAAAAAABQiQjqIyOhcESd3f3aPkH5m0XBgFZR/gYAAAAAAAAAcoagPtLW1Ts46UK12weG1bGxR+vbF2p58wn5HRwAAAAAACg6fsz0L3S1AAAoNgT1kZZQODJpQD9h1ElrtuzQnFlTydgHAAAAAKBC+THTn2oBADCxqkIPAKWhs7s/ZUA/YdRJG7r7czsgAAAAAABQlLp6B9WxsWfCYLz0x5n+d/c+k9M+AKBcEdRHSn1D0aR/RJPpGRhW31A0RyMCAAAAAADFKNOZ/qFwJCd9AEA5I6iPlLz+ceSPKgAAAAAAlcWPmf5UCwCAyRHUR0oHRw7ntR0AAAAAACg9fsz0p1oAAKRGUB8pzaj2tp6y13YAAAAAAKD0eJ2xv+mhp7Pug2oBACoJQX2k5HUleVagBwAAAACgcnidsb9t9/NZ90G1AACVhKA+Umqqq9GiYCCjNi3BgJrqanI0IgAAAAAAUGy8ztjf9/Lrb5XPoVoAAKRGUB9pWdXWqCpL79wqk1a2NeZ2QAAAAAAAoKhkM2M/UT6HagEAkBpBfaSltaFW69oXpAzsV5m0vn0hf0wBAAAAAKgwTXU1qj9mqqe2ifI5VAsAgNQI6iNtK5rnatOlLWpJ8se1JRjQpktbtLz5hDyPDAAAAAAAFIPF89/tqd3Y8jlUCwCAyflWcMzM6iXdIGmppGMlPSfpHknXO+deSrOP/y1psaRTJdVKGpX0tKRfSLrZObfPr/HCm9aGWrU21KpvKKpQOKKDI4c1o3qKWhtqeSsOAAAAAECF6zjnJG16eDDjdmNn/CeqBazdslOjLnkbqgUAqFS+BPXNbJ6kByUdJ+mnknZLWiRplaSlZtbqnHsxja4ul3RQ0gOShiQdJf3/27v7aDnq8oDj3yco7xEIIFgDJkBC0AIqYCI5FTAVqR6BqoDtMRYqVSoKpXhaq61CW4o9p/KqFK1IWmgLFK1YjwqWEMDwHq3gkZfwEkBRKEQgAQJinv4xc+2yuTd3787s3Z3s93POnMmd2fnt7/fcye4+z539DW8ATgI+GBEHZuYP6uizqpm9w1SL+JIkSZIk6SVGps+55YGVHR8z2vQ5R+23M9O32Zxzrl7OzaO0NXfmNE5YMMuCvqShVNeV+udRFPRPyMxzRzZGxBkUBfnTgOM6aOc3M3NN+8aI+CPgS2U776ilx5IkSZIkSardiQtmsfCCm9d7lf2I9U2f42wBkjS6ykX98ir9g4EVwBfadn8G+BCwMCJOzsxn1tfWaAX90mUURX0nSZMkSZIkSRpgdU+f42wBkvRSdVypf1C5vioz17buyMxVEbGUoug/D7i6y+d4V7m+vZMHR8SyMXbN6fL5JUmSJEmS1CGnz5Gk3qmjqL97ub5njP3LKYr6s+mwqB8RxwLTgS2BPYHfprhh7icq9VSSJEmSJEmTwulzJKk36ijqb1Wunxpj/8j2rSfQ5rHA3JafbwV+PzPv7eTgzNxntO3lFfxvnEA/JEmSJEmSVIHT50hSveq6UW6tMnMeQERsS1GEPw1YFhFHZuaVfe1cBf5lWpIkSZIkSZJURR1F/ZEr8bcaY//I9icn2nBmPgF8NyJuBe4CLoqI12TmcxPuZZ/c8+gqLrrxQRbf9Rg/fXLdbr9p5jROdA45SZIkSZI0ZLz4UZK6U0dR/+5yPXuM/bPK9Vhz7o8rM5+MiBuBw4HXAbd129ZkWXrv45x99XJuGeVmMK1ueWAlCy+4mc++ey+O3G+nSeqdJEmSJElSf6yvZuLFj5I0vik1tHFNuT44Il7SXkRMBeYDzwI3VXyeV5frFyu203OX3voQCy+4edyC/oi1CZ/42u0svffxHvdMkiRJkiSpf8armYxc/HjZrQ9Pcs8kqTkqF/Uz8z7gKmAGcHzb7lOBLYCLMvOZkY0RMSci5rQ+MCJ2jogdRnuOiPgwsB/wMHBH1T730tJ7H+cvvnYHa3Nix61NOOfq5b3plCRJkiRJUp91WjPx4kdJWr86rtQH+AjwGHBORHw9Ik6PiMXASRTT7nyq7fF3lkurNwI/iYjrIuIrZRtfiojbgfOB1cDCzPxVTX3uibOvXj7hgv6Imx9YyT2Prqq3Q5IkSVJNImJ6+Vn9kYh4PiJWRMRZEbHNBNuZVh63omznkbLd6XU+d0S8NiIui4jHImJNRNwdEadGxGYT6a8kqR4TqZl48aMkja2Won55tf6+wCJgLnAysCtwNjCvvOHteL5fPn4T4J3Ax4HfAxL4HPDazLy2jv72yj2Prup4yp2x+FdoSZIkDaKI2BVYBhwD3AKcCdwPnAjcGBHbdtjOtsCN5XH3le3cUra7LCJ2qeO5I2IucCvFfbn+myLXeBr4NPDdiNikw6FLkmrQTc3Eix8laXR13CgXgMx8mOJDdiePjVG2PURRyG+sOgryq9cM/C0DJEmSNJzOA14JnJCZ545sjIgzKL6hexpwXAft/B0wGzgjM09uaecEisL7ecAhVZ47IjYCLgQ2Bw7LzG+U26cAlwHvKY/7bCcDlyRV123NZOm9jzN7h6k190aSmq2u6XdEPQX5LTet7e8skiRJUi3KK+UPBlYAX2jb/RngGWBhRGwxTjtbAgvLx5/StvvzwIPA21uv1u/yuQ8A9gCuGynoA2TmWuDPyh+Pi4h1LjaSJPVGtzUTL36UpHVZ1K9RHQX5+bttV0NPJEmSpFodVK6vKgvjv5aZq4ClFFfFzxunnXnAZsDS8rjWdtYCV7Y9X7fP/dZy/Z32DmTm/RT3/XoNsM5UP5Kk3ui2ZuLFj5K0Lov6NapakJ87c5pfKZMkSdIg2r1c3zPG/pE7Gc7uQTuTdcyYImLZaAswp5PjJUnd10y8+FGS1mVRv0azd5jKm2ZO6+rYKQEnLJhVc48kSZKkWmxVrp8aY//I9q170M5kHSNJ6qFuaiZe/ChJo7OoX7MTF8xiygRn5pwS8Nl37+VfnyVJkqQBlJn7jLYAd/W7b5LUJBOpmXjxoySNzaJ+zebvth2nv3vPjt+k5s6cxkUfnMuR++3U245JkiRJ3Ru5sn2rMfaPbH+yB+1M1jGSpB7rtGbixY+StH7ebaQHjtpvZ6ZvsznnXL2cmx9Yuc7+6dtsxkG7b8/CN8/wa2SSJElqgrvL9Vhz0I9cSjnWHPZV2pmsYyRJk2C8msncmdM4YcEsC/qStB4W9Xtk/m7bMX+37bjn0VUsvfdxVq95kS03fRnzd9vOQr4kSZKa5ppyfXBETMnMtSM7ImIqMB94FrhpnHZuAp4D5kfE1Mxc1dLOFODgtufr9rkXA58CDgFOb+1AROxCUex/ELh/nP5KknrAmokkVWNRv8dm7zDVNyRJkiQ1WmbeFxFXURTdjwfObdl9KrAF8MXMfGZkY0TMKY+9q6Wd1RFxEfAh4BTg5JZ2PgrMAK7MzPtbjpnwcwPXAncCb4mIQzPzG2WfpgB/Xz7m/MzMCYZCklQjayaS1B2L+pIkSZI68RHgBuCciFhAUTSfCxxEMY3Np9oef2e5bp85+ZPAgcCfRsTrgVuAPYDDgMcoCveVnjszfxURx1BcsX95RFwOPAQsAPYFlgJndj50SZIkaXB4o1xJkiRJ48rM+ygK4osoCuonA7sCZwPzMvOJDtt5AngzcA6wW9nOXOBCYJ/yeSo/d2beDOwHXEFxlf9JFDfI/WvgbZn5fGcjlyRJkgaLV+pLkiRJ6khmPgwc0+Fj26/Qb923EjixXGp/7pZjfgwcMZFjJEmSpEHnlfqSJEmSJEmSJDWERX1JkiRJkiRJkhrCor4kSZIkSZIkSQ1hUV+SJEmSJEmSpIawqC9JkiRJkiRJUkNY1JckSZIkSZIkqSEs6kuSJEmSJEmS1BCRmf3uw6SJiCc222yzaXvssUe/uyJJkqQGuvPOO3nuuedWZua2/e6L+s/8QpIkSd2qklsMW1H/AeAVwIoam51Tru+qsc1hYvyqMX7VGL9qjF81xq8a41eN8eveDODpzJzZ746o/8wvBpLxq8b4dc/YVWP8qjF+1Ri/aoxf92bQZW4xVEX9XoiIZQCZuU+/+9JExq8a41eN8avG+FVj/KoxftUYP2lw+f+zGuNXjfHrnrGrxvhVY/yqMX7VGL/+cE59SZIkSZIkSZIawqK+JEmSJEmSJEkNYVFfkiRJkiRJkqSGsKgvSZIkSZIkSVJDWNSXJEmSJEmSJKkhIjP73QdJkiRJkiRJktQBr9SXJEmSJEmSJKkhLOpLkiRJkiRJktQQFvUlSZIkSZIkSWoIi/qSJEmSJEmSJDWERX1JkiRJkiRJkhrCor4kSZIkSZIkSQ1hUV+SJEmSJEmSpIawqN+hiJgVEX8eEYsj4uGIeCEiHo2IKyLioC7b3D8ivhURKyPiuYi4PSL+JCI2qrv//RYRL4+IEyPiwoj4nzJ+GRHHdtHWjPLYsZZLejGGfqozfi1tDs35N6KuMY9z/t3Uq/73WkRMj4ivRMQjEfF8RKyIiLMiYpsJtjOtPG5F2c4jZbvTe9X3QVBH/CJiyTjn16a9HEO/RMR7I+LciLg+Ip4ux3pxl23Vch43SV3xK2M11rn38170XRpm5hfVmF90z9yiHuYW4zO/qMb8ojvmFtWYWzTHy/rdgQb5G+Ao4MfAt4CVwO7AocChEXFiZp7TaWMRcRjwVWANcGnZ3ruAM4H5wBG19r7/tgDOKv/9KPBzYKeKbf4Q+Poo239Usd1BVGv8hvD868WYHwQWjbL9J933sn8iYlfgBuCVwBXAXcCbgBOBQyJifmY+0UE725btzAYWA5cAc4BjgHdGxJsz8/7ejKJ/6opfi1PH2P5ipY4Orr8E9gZWU/wfmtNNIz34PTRFLfErPcX/v9+0Wl2hTUmjM7+oxvyie+YWFZlbjM/8ohrzi0rMLaoxt2iKzHTpYAGOBt4wyvYDgBeA54FXddjWK4DHymP2bdm+KcULRgLv6/eYa47fxsDvjMQIOKUc57FdtDWjPHZRv8fV0PgN4/lX65jLxy/p97hqjtGV5bg+1rb9jHL7+R2288Xy8Z9r235Cuf07/R7rgMdvSfHW3P8xTXL8DgJmAQEcWMbs4n79Hpq21Bi/FcCKfo/HxWVYFvOLyvEzvxiM2A3juWdu0dm4zC8GI35Dl1+YWwxM/Mwterw4/U6HMnNRZv5glO3XUrxIbgzs32Fz7wW2By7JzNta2lpD8RcxgD+u1OEBk5kvZOa3M/Nn/e5LE9Ucv6E7/xjOMXesvALhYIo33S+07f4M8AywMCK2GKedLYGF5eNPadv9eYorkN4eEbtU7/XgqCt+wywzr8nM5Vl++uvGMP8e6oifpMlnflGN+UX3zC0qG8YxT4j5RTXD/Lm2DuYW1ZhbNIfT79Tjl+W6068tvbVcf2eUfdcBzwL7R8Qmmfl81c5twH4jIj4MbAs8AdyYmbf3uU9NMIznXy/GvHVE/CGwI8VXypZlZlPnvByZt/eqzFzbuiMzV0XEUooPNPOAq9fTzjxgs7KdVW3trI2IK4EPlc+3IX1Ftq74/VpEHAXMpLhS805g8Qb0/7FXav89DKlNIuL9wM4UycrtwHWZ+av+dksaOuYX/WF+MXHDeO6ZW4zP/KIa84v+M7eoh7lFD1nUrygiXgMsoHjjvq7Dw3Yv1/e078jMFyPiAeB1wC4UL7Ya3dvK5dciYgnwB5n5UF961AzDeP71Ysx7Axe0boiIHwILM/OOCn3thzHjU1pO8YFlNuv/wNJJO5TtbEjqil+r9hvyPRYRx2fm5V30b1j04vcwjHYELmrb9kBEHFNePSypx8wv+sr8YuKG8dwztxif+UU15hf9Z25RD3OLHnL6nQoiYhPgX4FNgFMy8xcdHrpVuX5qjP0j27fuvncbtGcpbiy2D7BNuRwAXEMx39fVG+JXoGo0jOdf3WM+g+IGWNsDU4H9gMspPowvjohXd9fNvqkrPsN4bkG9476C4iZr0ymuSpoDnF4ee2lEHNJ1Lzd8w3r+1elCikLijhQ3UdyTYh7bGcC3I2Lv/nVNGg7mF31jftG9YTz3zC3GZ35RjflF/w3ruVcnc4seG6qifkSsiIicwHLxetraiOKvTfMp7nb/D5M1jn6pM35VZOZjmfnpzPx+Zj5ZLtdR/JX0ZmA34NhePHcVgxK/phqk+GXmyZl5Q2Y+npmrM/O2zDwC+CqwHfDxXj23NmyZeWZmfjMzf5qZazLz7sz8JHAyxXv26X3uojZgmXlqZi7OzEcz89nM/FFmHkdRbNiMdeeylYae+UU1g/L5ron5xaDErqkGKX7mFuol8wv1i7lF7w3b9Dv3AWsm8PhHRttYfuC+GDgCuAx4/wRvIDHyF72txtg/sv3JCbQ5GWqJX6+UX3X8MjAXeAtw9mQ+fwcGJX7DeP5N1pjPB95Dcf41SV3xaeq5VdVkjPvLwJnA6yNiavucogKG9/ybDOdTJH5Ne22TJoP5RTWD8vl4VAOeXwxK7Ibx3DO3GJ/5RTXmF/03rOfeZDC3qMlQFfUzc0HVNiLi5RRfiT0C+DfgA13c4OFuYF+KubeWtbX/Moqbl7zIgN3opY74TYL/LdcD9/XYAYrfMJ5/kzXmgT3/xnF3uR5rLspZ5Xqs+QTrbqdpej7uzFwTEasopgPYAvBD97qG9fybDE19bZN6zvyimgH6fLw+A/kaOECxG8Zzz9xifOYX1Zhf9N+wnnuTocmvbQNlqKbfqSoiNgb+g+ID979Q3LSmmzs2Ly7Xo81d9hZgc+AG70TelXnleqA+NA6YYTz/JmvMTT3/rinXB0fES94XImIqxTQAzwI3jdPOTcBzwPzyuNZ2plB8hb31+TYUdcVvTBGxO8UH7lXA4922s4Hr+e9hiDX1tU0aeOYXjeBr4PoN47lnbjE+84tqzC/6z9yid5r82jZQLOp3KIqbVv0ncBjFXemPycy14xyzVUTMiYhXte26nOJF830RsW/L4zcF/rb88R9r63xDjRW/iHhj+4tquX0BcFL549DPGen59xITHnNEbF7Gb+e27XuVV9TRvh04rfyxUedfZt4HXEVxw5rj23afSvEX9Isy85mRjWVs5rS1s5piLuAtWHd+vI+W7V+ZmRvUm3dd8YuImRExrb39iNie4iZDAJdk5os1dr9xIuLlZfx2bd3eze9hGI0Vv4jYI0a5CWREzAA+X/7YqNc2adCZX0w+84vuee69hLnFOMwvqjG/mDzmFtWYW/RXTGyqxuEVERcCR1O8eZ8HjBa4JZm5pOWYoyleKP85M49ua+9wig8Da4BLgJXAocDu5fYjJziP5sCLiE9Q3Gkd4PXA3sANwPJy2/cy88stjz+aUeIXEUsovup0A/CTcvNewFvLf/9VZo58mNpg1BW/ct/hDN/5dzgTGHNEHEjx1/lrM/PAlu2LgHcB1wMPA89T/F4OATYC/gn4cNPiV74J3wC8ErgCuJNi/tiDKL5SuH9mPtHy+ATIzGhrZ9uyndkUVzHdAuxBUbB4rGznvl6PZ7LVEb/y/+z5wPcorlpYCewMvINizsbbgLdl5pM9H9AkK/9/Hl7+uCPwdooYXF9uezwzP14+dgbwAPBgZs5oa2dCv4cNRR3xi4hTKOa2vA54kOKqrV2BdwKbAt8CfjczX+jlWKRhYn5RnflF98wtqjG3GJ/5RTXmF90zt6jG3KJBMtOlgwVYQvFBe33LKW3HHF1uXzRGm/MpTuRfUHyl7A6KK0E26vd4+xTDRW2PHzV+wAeBbwIrgNUUH3weAi4Ffqvf4xz0+A3r+TfRMQMHlvFb0rb9cOBrwL3A08ALwM+A/wIO7fcYK8ZnJ4pk7WfluB4EzgK2GeWxWbyFjNrONIobyT3YEp+vANP7PcZBjh+wJ7CoPC+fAH5J8cH7euBjwMb9HmMPY3fKOK9vK1oeO6N9W7e/hw1lqSN+wAHAvwN3Udzw65cU811+F/gA5YUgLi4u9S2YX0xGDBe1PX7U+DGE+UVdsRvWc2+iY2YIc4tyfOYXfYwfQ5pfYG7R9/hhbjEpi1fqS5IkSZIkSZLUEM6pL0mSJEmSJElSQ1jUlyRJkiRJkiSpISzqS5IkSZIkSZLUEBb1JUmSJEmSJElqCIv6kiRJkiRJkiQ1hEV9SZIkSZIkSZIawqK+JEmSJEmSJEkNYVFfkiRJkiRJkqSGsKgvSZIkSZIkSVJDWNSXJEmSJEmSJKkhLOpLkiRJkiRJktQQFvUlSZIkSZIkSWoIi/qSJEmSJEmSJDWERX1JkiRJkiRJkhrCor4kSZIkSZIkSQ1hUV+SJEmSJEmSpIawqC9JkiRJkiRJUkP8HxuKcHzjgMGkAAAAAElFTkSuQmCC\n",
"text/plain": [
"<Figure size 936x288 with 2 Axes>"
]
},
"metadata": {
"image/png": {
"height": 263,
"width": 762
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fig, (ax1, ax2) = plt.subplots(1, 2)\n",
"fig.set_size_inches(13, 4)\n",
"\n",
"ax1.scatter(X_, rates_reparam.mean(axis=0))\n",
"ax1.set_title('rate mean')\n",
"ax2.scatter(X_, rates_reparam.std(axis=0))\n",
"ax2.set_title('rate std');"
]
},
{
"cell_type": "markdown",
"id": "b5480f4e-cfdc-43f1-8e24-a0edf1268d9b",
"metadata": {},
"source": [
"We see that the probability of success rises with `x`. This means that it will take more and more trials before we observe those 28 failures imposed by `concentration` parameter.\n",
"\n",
"Intuitively if we want to record 28 failures where each failure occurs with probability 0.5 then it should also take 28 successes. Let's check if our model follows this logic:"
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "6c70bdcd-e72e-44e8-879f-6f0e69d30c42",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 936x288 with 2 Axes>"
]
},
"metadata": {
"image/png": {
"height": 263,
"width": 766
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fig, (ax1, ax2) = plt.subplots(1, 2)\n",
"fig.set_size_inches(13, 4)\n",
"\n",
"ax1.scatter(X_, y_, c='r')\n",
"ax1.plot(counts_df['feat'], counts_df['mean'])\n",
"ax1.fill_between(counts_df['feat'], counts_df['high'], counts_df['low'], alpha=0.5)\n",
"ax1.axhline(samples['concentration'].mean().item(), c='g', linestyle='dashed')\n",
"ax1.axvline(-0.46, c='g', linestyle='dashed')\n",
"ax1.set_title('fitted model')\n",
"ax2.scatter(X_, rates_reparam.mean(axis=0))\n",
"ax2.axhline(0.5, c='g', linestyle='dashed')\n",
"ax2.axvline(-0.46, c='g', linestyle='dashed')\n",
"ax2.set_title('rate mean');"
]
},
{
"cell_type": "markdown",
"id": "a3a7b0a7-90bb-490b-8cb9-971b9b2a7338",
"metadata": {},
"source": [
"It indeed does. Red lines show that 28 successes and rate 0.5 are located with the same `x` argument. "
]
},
{
"cell_type": "markdown",
"id": "78066438-2aaa-4706-981f-d77eba71b9d2",
"metadata": {},
"source": [
"## SVI approach\n",
"\n",
"`Predictive` class can also be used with the SVI method. In the next section we will use it with `AutoGuide`'s guide and manually designed one."
]
},
{
"cell_type": "code",
"execution_count": 17,
"id": "3ea5009f-c82e-4bfb-aab4-2c950eb6f753",
"metadata": {},
"outputs": [],
"source": [
"from pyro.infer import SVI, Trace_ELBO\n",
"from pyro.optim import Adam\n",
"from pyro.infer.autoguide import AutoNormal"
]
},
{
"cell_type": "markdown",
"id": "f1dc46d1-cd9b-4b9d-b463-6f50ccfa1cbc",
"metadata": {},
"source": [
"### Manually defined guide\n",
"\n",
"First we define our guide with all `sample` sites that are present in the model and parametrize them with learnable parameters. Then we perform gradient descent with `Adam` optimizer."
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "9a4339a2-c8f3-44e0-a189-82893c27b557",
"metadata": {},
"outputs": [],
"source": [
"def guide(features, counts):\n",
" N, P = features.shape\n",
" \n",
" scale_param = pyro.param(\"scale_param\", torch.tensor(0.1), constraint=constraints.positive)\n",
" loc_param = pyro.param(\"loc_param\", torch.tensor(0.0))\n",
" scale = pyro.sample(\"scale\", dist.Delta(scale_param))\n",
" coef = pyro.sample(\"coef\", dist.Normal(loc_param, scale).expand([P]).to_event(1))\n",
" \n",
" concentration_param = pyro.param(\"concentration_param\", torch.tensor(0.1), constraint=constraints.positive)\n",
" concentration = pyro.sample(\"concentration\", dist.Delta(concentration_param))"
]
},
{
"cell_type": "code",
"execution_count": 19,
"id": "c173a362-acc5-4018-afbe-ebde64ffe273",
"metadata": {},
"outputs": [],
"source": [
"pyro.clear_param_store()\n",
"\n",
"adam_params = {\"lr\": 0.005, \"betas\": (0.90, 0.999)}\n",
"optimizer = Adam(adam_params)\n",
"\n",
"svi = SVI(model, guide, optimizer, loss=Trace_ELBO())"
]
},
{
"cell_type": "code",
"execution_count": 20,
"id": "d453e4af-0763-4a46-a8c6-1827d5a4f439",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Loss: 4083.343687057495\n",
"Loss: 467.4017288684845\n",
"Loss: 394.44257640838623\n",
"Loss: 408.9837645292282\n",
"Loss: 409.8578680753708\n",
"Loss: 394.2356821298599\n",
"CPU times: user 23.6 s, sys: 226 ms, total: 23.8 s\n",
"Wall time: 24.5 s\n"
]
}
],
"source": [
"%%time\n",
"\n",
"n_steps = 5001\n",
"\n",
"for step in range(n_steps):\n",
" loss = svi.step(X_, y_)\n",
" if step % 1000 == 0:\n",
" print('Loss: ', loss)"
]
},
{
"cell_type": "markdown",
"id": "0fed7b0c-ca4f-4c66-9268-858827f2aee0",
"metadata": {},
"source": [
"`Pyro`s parameter store is comprised of learned parameters that will be used in `Predictive` stage. Instead of providing samples we pass `guide` parameter to construct predictive distribution."
]
},
{
"cell_type": "code",
"execution_count": 21,
"id": "94acd52c-8e14-4f5e-93a8-1f568bbd6bc0",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[('scale_param', tensor(0.1934, grad_fn=<AddBackward0>)),\n",
" ('loc_param', tensor(-1.1939, requires_grad=True)),\n",
" ('concentration_param', tensor(28.2309, grad_fn=<AddBackward0>))]"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"list(pyro.get_param_store().items())"
]
},
{
"cell_type": "code",
"execution_count": 22,
"id": "006799eb-1f80-4981-b8ad-c1643497c384",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"dict_keys(['scale', 'coef', 'concentration', 'counts', 'rate'])\n"
]
}
],
"source": [
"predictive_svi = Predictive(model, guide=guide, num_samples=500)(X_, None)\n",
"print(predictive_svi.keys())"
]
},
{
"cell_type": "code",
"execution_count": 23,
"id": "d208a674-22e3-42ae-b59e-986a6fd81434",
"metadata": {},
"outputs": [],
"source": [
"counts_df = prepare_counts_df(predictive_svi)"
]
},
{
"cell_type": "code",
"execution_count": 24,
"id": "3f6b8e7d-714d-44c6-b8b2-140b161ed9e0",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"image/png": {
"height": 248,
"width": 375
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(X_, y_, c='r')\n",
"plt.plot(counts_df['feat'], counts_df['mean'])\n",
"plt.fill_between(counts_df['feat'], counts_df['high'], counts_df['low'], alpha=0.5);"
]
},
{
"cell_type": "markdown",
"id": "f11cf5f8-d3df-42f4-a191-8c257476c3df",
"metadata": {},
"source": [
"### AutoGuide\n",
"\n",
"Another approach for conducting SVI is to rely on automatic guide generation. Here we use `AutoNormal` that underneath uses a normal distribution with a diagonal covariance matrix."
]
},
{
"cell_type": "code",
"execution_count": 25,
"id": "fdfb6fd3-2217-4376-95b4-04ff1c47fbf6",
"metadata": {},
"outputs": [],
"source": [
"pyro.clear_param_store()\n",
"\n",
"adam_params = {\"lr\": 0.005, \"betas\": (0.90, 0.999)}\n",
"optimizer = Adam(adam_params)\n",
"\n",
"auto_guide = AutoNormal(model)\n",
"\n",
"svi = SVI(model, auto_guide, optimizer, loss=Trace_ELBO())"
]
},
{
"cell_type": "code",
"execution_count": 26,
"id": "5fe54144-3da1-4d55-b40e-e417b2ae0b46",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Loss: 4005.407080989331\n",
"Loss: 379.12515234947205\n",
"Loss: 376.1868230998516\n",
"Loss: 376.45190131664276\n",
"CPU times: user 18.8 s, sys: 47.8 ms, total: 18.9 s\n",
"Wall time: 18.9 s\n"
]
}
],
"source": [
"%%time\n",
"\n",
"n_steps = 3001\n",
"\n",
"for step in range(n_steps):\n",
" loss = svi.step(X_, y_)\n",
" if step % 1000 == 0:\n",
" print('Loss: ', loss)"
]
},
{
"cell_type": "code",
"execution_count": 27,
"id": "3eccb4d8-e08f-4f03-842a-3e53ecdbd7b8",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'scale': tensor(1.0996, grad_fn=<ExpandBackward>),\n",
" 'coef': tensor([-1.2571], grad_fn=<ExpandBackward>),\n",
" 'concentration': tensor(28.2732, grad_fn=<ExpandBackward>)}"
]
},
"execution_count": 27,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"auto_guide(X_, y_)"
]
},
{
"cell_type": "markdown",
"id": "13bb4ff7-060a-4079-8c81-1cd41c511622",
"metadata": {},
"source": [
"As we check `PARAM_STORE` we see that each `sample` site is approximated with a normal distribution. "
]
},
{
"cell_type": "code",
"execution_count": 28,
"id": "dc8f6f2c-4edd-496b-91ce-df0ba9886a20",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[('AutoNormal.locs.scale',\n",
" Parameter containing:\n",
" tensor(0.3109, requires_grad=True)),\n",
" ('AutoNormal.scales.scale', tensor(0.5159, grad_fn=<AddBackward0>)),\n",
" ('AutoNormal.locs.coef',\n",
" Parameter containing:\n",
" tensor([-1.2521], requires_grad=True)),\n",
" ('AutoNormal.scales.coef', tensor([0.0397], grad_fn=<AddBackward0>)),\n",
" ('AutoNormal.locs.concentration',\n",
" Parameter containing:\n",
" tensor(3.3570, requires_grad=True)),\n",
" ('AutoNormal.scales.concentration', tensor(0.0291, grad_fn=<AddBackward0>))]"
]
},
"execution_count": 28,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"list(pyro.get_param_store().items())"
]
},
{
"cell_type": "markdown",
"id": "211c5936-4d5f-459c-8495-6cb925a66187",
"metadata": {},
"source": [
"Finally we again construct a predictive distribution and plot `counts`. For all three methods we managed to get similar results for our parameters. "
]
},
{
"cell_type": "code",
"execution_count": 29,
"id": "bcbdc495-9248-4556-98a6-4281d094f56d",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"dict_keys(['scale', 'coef', 'concentration', 'counts', 'rate'])\n"
]
}
],
"source": [
"predictive_svi = Predictive(model, guide=auto_guide, num_samples=500)(X_, None)\n",
"print(predictive_svi.keys())"
]
},
{
"cell_type": "code",
"execution_count": 30,
"id": "1f663b61-a6c6-4d30-bff1-875633b56ada",
"metadata": {},
"outputs": [],
"source": [
"counts_df = prepare_counts_df(predictive_svi)"
]
},
{
"cell_type": "code",
"execution_count": 31,
"id": "e3beaa96-a387-4c3c-849f-e6dc2bc6a0d8",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
]
},
"metadata": {
"image/png": {
"height": 248,
"width": 375
},
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"plt.scatter(X_, y_, c='r')\n",
"plt.plot(counts_df['feat'], counts_df['mean'])\n",
"plt.fill_between(counts_df['feat'], counts_df['high'], counts_df['low'], alpha=0.5);"
]
}
],
"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": "3.8.9"
}
},
"nbformat": 4,
"nbformat_minor": 5
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment