age_estimation_regression.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "kT-P9uyJK7GT"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "# os.environ['CUDA_VISIBLE_DEVICES'] = \"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4fUpOXPUK7GY"
   },
   "outputs": [],
   "source": [
    "from NN import *\n",
    "from utils import get_tf_session\n",
    "from data_utils import make_batch, single_np_datapoint_generator\n",
    "\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 0,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "B1vLGApaK7Ge"
   },
   "outputs": [],
   "source": [
    "train_dataset_files = [\n",
    " './data/tfrecords/0_cropped.tfrecord',\n",
    " './data/tfrecords/1_cropped.tfrecord',\n",
    " './data/tfrecords/2_cropped.tfrecord',\n",
    " './data/tfrecords/3_cropped.tfrecord',\n",
    " './data/tfrecords/4_cropped.tfrecord',\n",
    " './data/tfrecords/5_cropped.tfrecord',\n",
    " './data/tfrecords/6_cropped.tfrecord',\n",
    " './data/tfrecords/7_cropped.tfrecord',\n",
    " './data/tfrecords/8_cropped.tfrecord',\n",
    "]\n",
    "\n",
    "validation_dataset_files = ['./data/tfrecords/9_cropped.tfrecord']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 88
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1467,
     "status": "ok",
     "timestamp": 1543102965462,
     "user": {
      "displayName": "Aarushi Bisht",
      "photoUrl": "https://lh6.googleusercontent.com/-2XghgRfShmY/AAAAAAAAAAI/AAAAAAAAS-8/XZz-3V9YukY/s64/photo.jpg",
      "userId": "00253849553632547740"
     },
     "user_tz": 300
    },
    "id": "UiiR3kfLK7Gh",
    "outputId": "92fdb9d0-c684-46db-8cb8-ea18ff85c63f",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/sparse_ops.py:1165: sparse_to_dense (from tensorflow.python.ops.sparse_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Create a `tf.sparse.SparseTensor` and use `tf.sparse.to_dense` instead.\n"
     ]
    }
   ],
   "source": [
    "X_batch, Y_batch = make_batch(train_dataset_files, 200, raw_values=True)\n",
    "no_of_outputs = 1\n",
    "age_dense = get_regression_NN(X_batch, 60, [lambda i: 10 + 3*i], 4, no_of_outputs)\n",
    "loss = tf.losses.mean_squared_error(tf.cast(tf.reshape(Y_batch[:, 1],[-1,1]), tf.int32), age_dense)\n",
    "train_op = get_training_op(loss)\n",
    "X = tf.placeholder(dtype=tf.float32, shape=(None, None, None, 3))\n",
    "age_pred = get_regression_NN(X, 60, [lambda i: 10 + 3*i], 4, no_of_outputs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 2011,
     "status": "ok",
     "timestamp": 1543102969579,
     "user": {
      "displayName": "Aarushi Bisht",
      "photoUrl": "https://lh6.googleusercontent.com/-2XghgRfShmY/AAAAAAAAAAI/AAAAAAAAS-8/XZz-3V9YukY/s64/photo.jpg",
      "userId": "00253849553632547740"
     },
     "user_tz": 300
    },
    "id": "4_Zakn1nK7Gl",
    "outputId": "13402d09-02ce-46d6-a536-f9f32eb5d444"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensorflow is able to detect gpu...\n"
     ]
    }
   ],
   "source": [
    "sess = get_tf_session()\n",
    "if tf.test.is_gpu_available():\n",
    "    print(\"Tensorflow is able to detect gpu...\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 442
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 385781,
     "status": "ok",
     "timestamp": 1543103387288,
     "user": {
      "displayName": "Aarushi Bisht",
      "photoUrl": "https://lh6.googleusercontent.com/-2XghgRfShmY/AAAAAAAAAAI/AAAAAAAAS-8/XZz-3V9YukY/s64/photo.jpg",
      "userId": "00253849553632547740"
     },
     "user_tz": 300
    },
    "id": "c4CQmSMXK7Go",
    "outputId": "49de2bf6-6cf2-4efc-f3a8-8136a221de6f",
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Batch No. 200 Avg loss: 331.8318941116333\n",
      "Batch No. 400 Avg loss: 157.04911346435546\n",
      "Batch No. 600 Avg loss: 154.50487873077392\n",
      "Batch No. 800 Avg loss: 151.47941005706787\n",
      "Batch No. 1000 Avg loss: 148.69893268585204\n",
      "Batch No. 1200 Avg loss: 146.11204273223876\n",
      "Batch No. 1400 Avg loss: 142.8570775604248\n",
      "Batch No. 1600 Avg loss: 140.3071291732788\n",
      "Batch No. 1800 Avg loss: 138.14426025390625\n",
      "Batch No. 2000 Avg loss: 136.46332008361816\n",
      "Batch No. 2200 Avg loss: 135.83071865081786\n",
      "Batch No. 2400 Avg loss: 134.92899894714355\n",
      "Batch No. 2600 Avg loss: 133.8459800720215\n",
      "Batch No. 2800 Avg loss: 132.29557598114013\n",
      "Batch No. 3000 Avg loss: 131.5482594680786\n",
      "Batch No. 3200 Avg loss: 130.68667163848878\n",
      "Batch No. 3400 Avg loss: 129.84105667114258\n",
      "Batch No. 3600 Avg loss: 129.20860927581788\n",
      "Batch No. 3800 Avg loss: 128.50024024963378\n",
      "Batch No. 4000 Avg loss: 127.55597728729248\n",
      "Batch No. 4200 Avg loss: 126.89839797973633\n",
      "Batch No. 4400 Avg loss: 126.2782810974121\n",
      "Batch No. 4600 Avg loss: 125.43907047271729\n",
      "Batch No. 4800 Avg loss: 124.7928237915039\n",
      "training took total 385.2217128276825 seconds\n"
     ]
    }
   ],
   "source": [
    "avg_loss = 0\n",
    "print_int = 200\n",
    "st = time.time()\n",
    "# while datetime.datetime.now().time() < datetime.time(hour=8, minute=30):\n",
    "for i in range(5000):\n",
    "    _, loss_value= sess.run([train_op, loss])\n",
    "    avg_loss += loss_value\n",
    "    if i>0 and i%print_int == 0:\n",
    "        print(\"Batch No.\", i, \"Avg loss:\", avg_loss/print_int)\n",
    "        avg_loss = 0\n",
    "print(\"training took total\", time.time() - st, \"seconds\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 4893,
     "status": "ok",
     "timestamp": 1543107928589,
     "user": {
      "displayName": "Aarushi Bisht",
      "photoUrl": "https://lh6.googleusercontent.com/-2XghgRfShmY/AAAAAAAAAAI/AAAAAAAAS-8/XZz-3V9YukY/s64/photo.jpg",
      "userId": "00253849553632547740"
     },
     "user_tz": 300
    },
    "id": "wqkHMFHwK7Gt",
    "outputId": "ee31d3e6-f683-4abc-d44c-c4810f83ff3e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Mean absolute Error 8.946773120425815\n"
     ]
    }
   ],
   "source": [
    "gen = single_np_datapoint_generator(validation_dataset_files, raw_values=True)\n",
    "total = 0\n",
    "absolute_error = 0\n",
    "actual_pred_pair = []\n",
    "for img, attr in group_iterable_into_list(gen, 20, 2):\n",
    "    pred = sess.run([age_pred], feed_dict={X: img})[0]\n",
    "    age_value = np.array(attr)[:, 1]\n",
    "    pred = np.round(pred).reshape(1,pred.shape[0])[0]\n",
    "    absolute_error += np.sum(np.absolute(pred-age_value))\n",
    "    total += pred.shape[0]\n",
    "    for i in range(len(pred)):\n",
    "      actual_pred_pair.append((age_value[i], pred[i]))\n",
    "print(\" Mean absolute Error\", absolute_error/total)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 347
    },
    "colab_type": "code",
    "executionInfo": {
     "elapsed": 1506,
     "status": "ok",
     "timestamp": 1543109881834,
     "user": {
      "displayName": "Aarushi Bisht",
      "photoUrl": "https://lh6.googleusercontent.com/-2XghgRfShmY/AAAAAAAAAAI/AAAAAAAAS-8/XZz-3V9YukY/s64/photo.jpg",
      "userId": "00253849553632547740"
     },
     "user_tz": 300
    },
    "id": "8dQLb9hVzail",
    "outputId": "0f70877b-ad09-437c-a673-aae935f666de"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAe0AAAFKCAYAAAAwrQetAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzsnXdgFGX+/9+zu0l209smECD0KiKo\niKKAiOVQv2e5r4qc5c52/mzcVz05PRuodzYQC+d56llOURQBUZReFKX3DgmQRsqmJ1uSLfP7YzNb\nZmdmZ3Z3toTPiz/YTHmeZ+pnnk9lWJZlQRAEQRBE3KOJ9QAIgiAIgpAHCW2CIAiCSBBIaBMEQRBE\ngkBCmyAIgiASBBLaBEEQBJEgkNAmCIIgiARBF+sBSGEytUW8zZycVDQ1WSLebqyh40os6LgSCzqu\nxCLRj8tozBBdd8bNtHU6bayHoAp0XIkFHVdiQceVWHTX4wJUnGmbzWbMnDkTLS0tsNvtePDBB2E0\nGvH8888DAIYOHYpZs2ap1T1BEARBdDtUE9pLlixB//798dhjj6G2thZ33nknjEYjnnrqKYwaNQqP\nPfYYNm7ciEmTJqk1BIIgCILoVqimHs/JyUFzczMAoLW1FdnZ2aiqqsKoUaMAAJMnT8bmzZvV6p4g\nCIIguh2qzbSvueYaLF68GFdccQVaW1vx7rvvYvbs2Z71eXl5MJlMkm3k5KSqYpuQMvInMnRciQUd\nV2JBx5VYdNfjUk1of/vttygqKsKHH36II0eO4MEHH0RGhvckyqlToob3n9GYoYpXeqyh40os6LgS\nCzquxCLRj0vqg0M1ob1r1y5ccsklAIBhw4aho6MDDofDs762thYFBQVqdU8QBEEQ3Q7VbNp9+/bF\n3r17AQBVVVVIS0vDwIEDsWPHDgDAqlWrMGHCBLW6JwiCIIhuh2pC+5ZbbkFVVRVuu+02PPbYY3j+\n+efx1FNPYe7cuZg2bRqKi4sxfvx4tbpXndWrV2DSpHEeZzsxSkqOo7y8THH7//u//wOLJXGTAxAE\nQRCRRzX1eFpaGt58882A5QsWLFCry6iyevVK9OrVGxs2rMH11/+v6HYbN67DsGEjUFzcN4qjIwiC\nILojcZ3GNF5pbW3B4cMH8eSTz2LBgk89QvvYsSOYM+cVaDQMRo48B7/5zTX49tvF2LhxHXJycvDs\ns0/i008XIjU1Fe+8Mw8DBgzEpEmTMWvW07BarbDZbPi///sLRowYKdjvqlU/YtGihdBqNejXbyBm\nzvwb2tvb8fTTT8DlcuD88y/Ed98txddfL8Pevbvx3nvzodPpUFBQiJkzn0ZSUlI0TxNBEAQRYRJa\naH+1rgTbj9Qp2kerZeB0inuujx1WgJsvGyTZxrp1azB+/CUYN+4ivPLKizCZ6mA0FmDevNfxl788\nhUGDBuOFF55FWloaxo27CJdeOkVUEDc0NODaa6/HxImXYufO7fj880/w0kuvCW5rtVoxZ87byMjI\nwIMP3ovS0hLs3r0D/foNwEsvzcJ77/3H45U/b95rePPNd5GZmYV//vNNrF+/BldeOVXmWSIIgiCC\n0dTWgT3HTbjwrB4wpERHnCa00I4Va9asxJ133g2tVovJk6dg7dpVmDbtNpSXl2HQoMEAgGeemR2k\nFTe5uXn45JMP8MUX/4XdboderxfdNjMzE08++RgAoKzsJFpamnHq1CmMGXMeAOCSSyZiwYJP0djY\ngMrKCjz11F8AADabDVlZ2eEcMkEQBMHjh81lWLurElqtBhPPKYpKnwkttG++bFDQWTGfcOP36upq\ncejQAbzzzjwwDAObzYaMjHRMm3YbNBppvz6GYTy/ufC3r75agPz8AjzzzAs4cuQQ3nlnnuC+drsd\nc+e+io8/XoC8vHw88cSfu9aw0GgYv/Z1uiTk5xvxzjv/Dvk4CYIgCGlazB0AgLMH5EWtzzOuyle4\nrFmzEjfccBM++eQLfPzxAnzxxTdobW1FVVUl+vXrj4MHDwAA/vGP2Th16iQYhoHT6QQApKamoaGh\nHk6nEwcP7gcAtLQ0o1ev3gCAjRvX+8Wy+2KxmKHVapGXl4/a2hocOXIYDocDRUW9ceTIYQDAli2/\nAnDPyAHg5MkTAIBFi75ESclxlc4IQRDEmYnZ5n5fZ6RGz18ooWfasWDNmpV4+mlvdTKGYTB16rVY\ns2YlZsx4HK+//g8AwFlnnY1+/frjnHPGYN6815Camorf/e5mzJz5fygu7ov+/QcAAH7zm2vw4ovP\nYf36Nfjd727GmjWrsHz5soB+s7KyMXbsONxzzx0YNGgwpk+/HW+9NRdvv/0ennrqMdx+++0455zz\nPLP9v/71Wfz977OQlOSedf/2tzdG4ewQBEGcORwua0JykgY6bfTmvwwrJ59ojFAjDV2ip7fjU1NT\njbKyU7j22iuxfv0v+PDD9/DGG/NjPayI0d2uFwcdV2JBx5VYROO4Glps+Mu7v0KrYfD+E5Mj2nZM\n0pgS0SEtLR0LF36Ozz77D+x2J/7858djPSSCIIhuT6ulEwBw3lBjVPsloZ3gZGRkYO7cd7rtFzNB\nEEQ8YrbZAQC98tOi2i85ohEEQRCEQixdTmhphugmrSKhTRAEQRAKMVvdM+1UfXQV1iS0CYIgCEIh\nXLhXmp5m2gRBEAQR13DqcZppxznV1adxxRUT8dBD9+Ghh+7Dfff9ARs3rg+prW++WYgPP3wPx48f\nxYcfvie63aZNG2G322W1eeJECR566L6QxkMQBEHIY//JBgBAepRn2uQ9HgLFxX09KUJbW1vwxz/+\nHhdeeBFSUsTzhksxePBQDB48VHT9l19+jnPPHUtVugiCIOKEmgYLACAzLTmq/ZLQDpPMzCzk5eWj\noaEBH330PnS6JLS2NmP27Jfx6qsv4fTpKjgcDtxzz/0477yx2LFjG956aw5yc/OQl5ePoqJe2LVr\nBxYv/govvvgqVqxYjkWLFoJhGEyb9nvY7XYcOnQAjz/+CN58810sW7YEa9asAMNoMGHCpbj11ttQ\nV1eLhx66B4AGgwYNCRijw+HASy89D5OpDlarFXfddR8uvngCtm/f2jWWfBQX90V2djbuvvtPeO+9\n+di3bw9cLiduvPFmXHHFb6J/YgmCIOKczLTkqFX34khoob245HvsrtuvaB+thoHTJZ4EbkzB2bhx\n0LWy26uuPo3W1hYUFBQCcOf9njnzb1ixYjny8vLx5JPPorm5GTNm3I9PPvkS7733Dp555gUMHjwE\njz/+CIqKennasljM+PjjD/DJJ1+gs9OOl156Di+/PBcffPAvvP76WzCZ6rBhw1r8858fAgD+3/+7\nG5MnX47Fixfi6quvxtVX34jPPvsYJSXH/MbY1taKCy64EFOnXouqqko888xfcfHFE/Duu2/jmWdm\nY+DAwXjwwXsxduw47N27G7W1NZg//310dnbirrtuw8SJl4asRSAIguhuOF0uOF0sivJSo953Qgvt\nWFFeXuaxGycnJ+Ppp2dBp3OfyhEjzgIAHDiwD3v37sa+fXsAAB0dHbDb7aiursbgwe7Z8OjR56Kj\no8PT7qlTJ1Fc3A8pKXqkpOjx8stz/fo9fPggKisr8PDDfwLgFvI1Nadx6tRJ3HDDbwEAY8ac7ykc\nwpGRkYnDhw9i2bLFYBgNWltbAAC1tdUYMmQYAODCC8fD6XRi//69OHhwv+f4WNaF+vp6T1ETgiCI\nMx27wwUASE7SRr3vhBbaNw66VtGsGIhMTlpfmzYfnS7J8/8dd9wVoFr2Ld/JT/uu0WjBsi7RfnW6\nJFx00cV44om/+S3//PNPPO0K7b969Qq0trZi/vwP0NrainvuuT1gG66sZ1JSEq699jrcfvsfRcdB\nEARxJsMJ7aQoFgrhIO9xlRgxYiQ2bdoIAGhqasR777mLeOTnG1Fefgosy2L37p1++/Tt2w/l5WWw\nWCzo6OjAn//8AFiWBcNo4HQ6MXTocOzatRM2mw0sy2LevNfR0WFDcXFfHDjgLgm6a9eOgLE0Nzej\nZ88iaDQabNy4zuOJnpubh7KyU3A6ndi+fatn3L/88jNcLhc6OjrwxhuvqnaOCIIgEhGP0E6KvghN\n6Jl2PHPZZZdj167tuP/+u+B0OnHXXW518333PYCnn56JHj16euzgHAaDAXfffT/+/OcHAAC33DId\nDMNgzJhz8cADd+Ptt/+Nm2++FQ8+eC80Go3H1nzTTbdi9uy/YfnyHzFw4OCAsVx66WX4618fxaFD\nB3DNNb9FQUEBPvrofdx77wP429/+gp49i9C3bz9otVqcffY5GDPmPPzpT38EwOKGG25S/VwRBEEk\nEp2celwXfaFNpTm7CaEc17ZtW9CnTzF69izCq6++hNGjz8OVV8aXpzhdr8SCjiuxoOMKjS0Ha/Dv\n7w5hyrm98fsrAyN2woVKcxKCsCyLp556HKmpacjJycXkyVNiPSSCIIi4Z/WOCgBAdkZ0Y7QBEtpn\nNOPGXYRx4y6K9TAIgiASik67Wz1+1QXFUe+bHNEIgiAIQgFmmx3GbD105D1OEARBEPFLU1sHmts7\no17di4OENkEQBEHIZNX2cgCAPjn6iVUAEtoEQRAEIZtWcycAYPoVkfcalwMJbYIgCIKQibmrjnZe\nZmzqMZDQJgiCIAiZ7CttgFbDkHqcIAiCIOKZVotbNe50sZ56DdFGtTjtr7/+GsuWLfP8feDAAXzx\nxRd4/vnnAQBDhw7FrFmz1OqeIAiCICKKtcOtGh85IDdmY1BNaN9000246SZ33upt27bhxx9/xEsv\nvYSnnnoKo0aNwmOPPYaNGzdi0qRJag2BIAiCICIGVyjEmG2I2Riioh6fP38+7r33XlRVVWHUqFEA\ngMmTJ2Pz5s3R6J4gCIIgwsYew0IhHKqnMd23bx969uwJrVaLzMxMz/K8vDyYTCbJfXNyUqHTRd7Y\nL5WMPZGh40os6LgSCzquxEKN46pt7QAAZGcaYnbeVBfaixYtwg033BCwXE5xsaYmS8THQ1VtEgs6\nrsSCjiuxoONShqmhHQBg73Soet6kPghUn+Nv3boVY8aMQW5uLpqbmz3La2trUVBQoHb3BEEQBBER\n7F2FQpJiqB5Xtefa2lqkpaUhOTkZSUlJGDBgAHbs2AEAWLVqFSZMmKBm9wRBEAQRMarqzQC6sU3b\nZDIhN9frGv/UU0/h2WefhcvlwjnnnIPx48er2T1BEARBRIwTp1sBAIaU2FW1VrXnkSNH4oMPPvD8\nPWjQICxYsEDNLgmCIAhCFRxOt3p8zGBjzMZAGdEIgiAIQgZmmwM6rQbJSd1UPU4QBEEQiU5zewd+\n2V8NU7MVaXpdzFKYAiS0CYIgCEKStTsrsXxzGQBgYK/MIFurCwltgiAIgpCAq6F977UjcPbAvJiO\nhWzaBEEQBCGBxeYtFJJuSIrpWEhoEwRBEIQInXYndh5zp9xO1cdeOU1CmyAIgiBE2Fva4Pmt1cRe\nZMZ+BARBEAQRp3D27FunDI7xSNyQ0CYIgiAIEcxWOwCgyJgW45G4ib2CniAIgiDiDBfLYv2uKhw4\n2QgASIsDezZAQpsgCIIgAjhR1YrPVx8DAGgYBrkZ+hiPyA0JbYIgCILg0Wpx27IvO7cXLj+/DzLT\nkmM8Ijdk0yYIgiAIHpwtu1+PTPTITY3xaLzQTJsgCIIgurB2OLBhdxUOlzUBiB9bNkd8jYYgCIIg\nYsj2I3X4ekOp529jtiGGowmEhDZBEARBdNHWZcu+efIgjByQi97G9BiPyB+yaRMEQRBEF2arO8/4\n4D5ZcSewARLaBEEQBAHA7TG+Yls5ACBNH9vCIGKQ0CYIgiAIAD/vPe35nRUnIV58SGgTBEEQBIA2\nizvM64HrR8KQEp8uXyS0CYIgCALe2OziHhkxHok48fkpQRAEQRBRorbRgs0Ha3CiuhVA/MVm+xK/\nIyMIgiCIKPDdr6fw64EaAG6BHa+qcYCENkEQBHGGw+UZf+yW0SjKT4OGYWI8InFIaBMEQRBnNGar\nAzotgxH9csDEscAGSGgTBEEQZyhNbR34ee9pmJqtSNUnxb3ABkhoEwRBEGco63ZVYvnmMgDAoF5Z\nMR6NPEhoEwRBEGckrWa3LftPvz0LIwfkxng08qA4bYIgCOKMxGJz5xkf0S8nbtOW8qGZNkEQBNGt\nabV0YuPuKtidLr/lZbVtAIDUOI7L5pM4IyUIgiCIEPh572ks+fmk4Lq8zBRoNYmjdFZVaC9btgwf\nfPABdDodHnnkEQwdOhRPPPEEnE4njEYjXnvtNSQnx2dSdoIgCKJ7wOUU/8PUYeiRm+q3rpD3d7yj\nmtBuamrC/Pnz8c0338BiseDtt9/GypUrMX36dEydOhVz587FokWLMH36dLWGQBAEQRAw29xCe1hx\nNgpyEktI81FNJ7B582ZcdNFFSE9PR0FBAV544QVs3boVU6ZMAQBMnjwZmzdvVqt7giAI4gymvLYN\ni38qxTcbS1Fa5c4pnpogzmZSqDbTrqyshM1mw/3334/W1lY8/PDDsFqtHnV4Xl4eTCaTZBs5OanQ\n6bQRH5vRGL8VXMKBjiuxoONKLOi4EovvNpdh55E6z98ZqUko7p0DrSb+E6hIoapNu7m5Ge+88w5O\nnz6NO+64AyzLetb5/hajqckS8TEZjRkwmdoi3m6soeNKLOi4Egs6rsTCaMxAfbMVSToNHp82GgCQ\nn2VAY0N7jEcmD6kPKdWEdl5eHsaMGQOdTofi4mKkpaVBq9XCZrNBr9ejtrYWBQUFanVPEARBnMFY\nbHak6XUY3Ds71kOJKKrZtC+55BJs2bIFLpcLTU1NsFgsGD9+PFauXAkAWLVqFSZMmKBW9wRBEMQZ\nSIfdiUXrjqPF3JkwCVOUoNpMu7CwEFdddRVuvvlmAMDTTz+Ns88+GzNnzsTChQtRVFSE66+/Xq3u\nCYIgiDOQfaUN+GT5IQCAMdsQ49FEHlVt2tOmTcO0adP8ln300UdqdkkQBEGcwXD5xG+Y0B9Xji2O\n8WgiT+KkgSEIgiCIIFi6YrL7F2UiJTny0UexhtKYEgRBEAmPy8VizY4K7CttAIBuac8GSGgTBEEQ\n3YAT1a34cl0JAECrYZCbqY/xiNSBhDZBEASR8LR12bKnnNcb064aBq3LFWSPxIRs2gRBEETCY+6q\njV1ckI4eeWkxHo160EybIAiCSEhazZ1Yv7sKdocLp2q6T35xKUhoEwRBEAnJpv3V+HaTf53sgpzu\nF5vtCwltgiAIIiHhYrL/ePUw9MxLQ7ohKaBedneDhDZBEASRkFi67NhD+mSjMMHrZMuFhDZBEASR\nMLRb7Vi7sxKdDidKqloAdN+YbCFIaBMEQRAJw68Havzs2OmGJKSmnDmi7Mw5UoIgCCLhabO47dh3\n/GYo+hjTkZ+lh0bDxHhU0YOENkEQBJEwcPHYg3tloZcxPcajiT4ktAmCIIi4ZfPBGlSa2j1/H6to\nBgCkGc4cO7YvJLQJgiCIuKSj04kPvjsElrfckKJDOgltgiAIgogf2q12sABGDsjFdRf39yzPy9JD\npz0zs3CT0CYIgiDiEnNXbezC7FQM7JUV49HEByS0CYIgiLjhaHmTpyZ2U1sHACDNQKKKg84EQRAE\nETd8vvq4n+MZABizu3c+cSWQ0CYIgiDihjZrJ3IyUvDADSMBAMk6LXobu2+pTaWQ0CYIgiDiBovN\ngd7GNAwsIhu2ECS0iW7Hjto9MBry0DezT6yHQhBnLJuqtmBIzkAUpBolt+uwO7FqWzmsnU6wLAu7\nwxXXucR/Pb0d/bOK0TOtMCb9k9AmuhVOlxMfHVwAAJh/2asxHg1BnJmUtVbgi6OLAQR/DveXNmDJ\nz4lRE7vOUo/Pj3wNIHbvFxLaRLfCFZCGgSCIaGO2W2Rvy+USv3HiAIzolwuGAfoUxGd60g5nZ6yH\nQEKb6GawJLQJItYwkF/Ao70rl3i/HhkYUJSp1pAigoaJfWESEtoEQRBE1LDYHFi9owIdnU4AwNEE\nyiWuYWKfhY2ENkEQBBFZJCakO4/W+dXDBgCthkFepl7lQYUPCW2CiDCkHCeI2COlHm+zulOT3jpl\nMAb1dod1ZaUlIzMtOSpjCwcNSGgTRIQhsU0Q8QyXT7x/z0z07xnfNmw+cWDSJqFNEARBqMf2I3U4\nebrV8/ehU40AEjOfeDz4uSbeWSMICeLgmSKIMx5f9fj73x2Ew+n/ZCbrNMhOT4n2sMKGjYM3jGpC\ne+vWrZgxYwYGDx4MABgyZAjuuecePPHEE3A6nTAajXjttdeQnBz/dgwicWDj4VOYIAgPDieLIb2z\ncMuUwZ5l2ekpMKQk4pwx9u8XVc/aBRdcgLfeesvz95NPPonp06dj6tSpmDt3LhYtWoTp06erOQTi\njCP2DxVBEP7kZxsSzn4tRDy8XaLqCrd161ZMmTIFADB58mRs3rw5mt0TZwDx8FARxJnI7uMmLFx3\nHAvXHceGPVWe5drCU9DqO2I4ssgRD5o8VWfaJSUluP/++9HS0oKHHnoIVqvVow7Py8uDyWSS3D8n\nJxU6nTbi4zIaMyLeZjxAxwVYOr23dLyfj3gfX6jQcSUWkTquj9/a5ElJqsloQMpw9/LkvkdQqquH\n0TgxIv3IRY3r1ZHirfMdq/tBNaHdr18/PPTQQ5g6dSoqKipwxx13wOl0etbL+WJpapKfv1YuRmMG\nTKa2iLcba+i43FjsVs/veD4fdL0SCzouaVwuFm2WTvQtzMAdvxmKCnMZFpZv96xvctRH9fypdb0a\nzWbPbzWPR+qDQDX1eGFhIa6++mowDIPi4mLk5+ejpaUFNpsNAFBbW4uCggK1uifOWGKvviKIMw1L\nhzt/eG5mCvr3zETPvNQYj0gdurV6fNmyZTCZTLj77rthMpnQ0NCAG2+8EStXrsR1112HVatWYcKE\nCWp1T5yhxP6RIojuA8uy2FD5C87OH4F8Qy4Ad/3rlVvLPYIaAGyd7t/eOtiBWUh+Pb0NfTP7oFd6\nT9XHrRbdOuTrsssuw+OPP461a9fCbrfj+eefx/DhwzFz5kwsXLgQRUVFuP7669XqnjhDiYeHiiC6\nCwcbjmDR8WX48dQavDrheQDu+tdLebnDOQpzxetgf35kEQCqcx8uqgnt9PR0/Otf/wpY/tFHH6nV\nJUHQVJsgIghXF9u3PjaXO/zGiQMwckCuZ7lWo0EvYxoAgImHfJ8q0K3V4wRBEET3w9KVO7xvjwz0\n65H4sddKiL3IjnKcNkGoDanHCbUoaT6JXXX7Yj2MkGnvNGNN+UZ0Ojtlbb9hTxW2Hqr1/P3l2uP4\ncu1x7DrmDtVN1Ud+zre//hAONxyTvf22ml042VIesPyw6XjY12pL9Q6Ut1b6LWPh8vxeV/5TWO2H\nCs20CYIgZPDGrncBAOcmqE32syNfYX/9YZjtFlw3cKrktq2WTny64ii0eY1IHuhetmp7hWe9VsMg\nP0vcfi1VmlOKf+37GIA8u7eLdeGTQ18Kbv/curkAQr9WFrsV/z38VWDbPnOCb0q+x6V9Lol6jW0S\n2kS3gmbaBCFMjbkOANBoawq6bbvFWz6Ty2323B/GetZnpiUjK8b1r9W0L7tYl+By/vvF4XIgWRvd\n80BCm+hWxIGfCEHEJUoeDa7mdVZaCqq6duzbQ34GsGj4oblU/EDXyDyATpedhDZBhAdJ7e7I1uqd\nKEwzol9mcayHkvBwqmuLzYHPVhxGU7M1YJv6FncSrJQkLSDPBB5Rvj+xClf3v9yjeq5oO42KtiqM\nL/LO9uXMtPfXH4JOo8Pw3CGeZZurdyBJo8P5haNF99tUtVVwOX+mbXfagSTBTVWDhDbRrSD1ePfD\n7rTj08MLAVCMbyTZdcyEhaulnb6y01OAxlBaD2+q/eOpNeid3hOjC84GALy8fR4AYETeEGSnZAGQ\n96zzbeSdTjs+67JViwnt8rZKfHviR8F1/A8Fu8sedAyRhoQ2QRBxjZN1Bt+IUEyb1T2FvnXKYAzp\nkx2wPjlJg9OOEqwPSWiH//Hc2tkesMzmsAGc0A7BFuZkHUG3sdptouv4PdpdwduLNCS0iW5FPCQ/\nICILXVF1MFvdAqdfzwxRe3V1XWgz5kg8hkIzaYfLKbk+GC4ZA5Nul2baBEHw2Gs6gBRtCoblDo71\nUFTnZEsZTNYGXNDjXNFtlHyIRcP2vb5iE87KG4aC1Pyw2mm0NWFX3T5c1mcCNIwGLR1t2FazE5f2\nuQRJGvFXs4t1YUPFJpxjHIk8Qy4q2qpQ0XYaFxSehx+3lKHd6p391eMkdEhBNorQztgBBjhV04Y3\nq39EbS0LQA+n1oY15Rtxae+LoZPol+Ngw1GwrAsj84eHdfzBEBTaPjPlUD7Qfb3CD9QfBsMwOCtv\nWND9HC4HNlT+glx9jt9yu5OENkGERXewaf97/6cAzgz77es75wNw2xfF4l3lXtPOKNm+Fx1fhmWl\nP+KNS18Kq503d72HelsjMpMzcEGPc/HBgf/iRMspaBgNphSL157eZzqIb0q+x+ryjfjHJc/g5e1v\nAgBSrD2x5Gf/nOCGC9YAAKzbfoOUUQ5o9EB1gxmm/PVATyC55hosrfgKFe2V0DJaTO5zid/+QvHW\n/9z7IQDpcxyJ51BIKIc/0/YK7Xf3uVNq849DqN1Np7diSclyyfFEC1lR4U1NTdi/fz8AwOUSjl8j\nCIJQA7GYWT7RtH13RkAtWm9zG4vbumy31eYaAEBzR4vkfu12d03n1k7/es6tFrct9pqL+uK5P4z1\ni6t+7g9jkZuRAgA4e0CeZ/kHT12OGot4v7HMIS4kPJ2+QjvMmbYSWjpahduLwSQhqND+/vvvccst\nt+DJJ58EALzwwgv4+uuvVR8YQYQCmbS7H3LskG4S8+IzvF/BZpBiGglLh1ug9Tamo28Pfzt13x4Z\n0Gnd7afqtZ7lOZl6yX5DFdmR8C0RnGmzkZtpi3csvz02xI+AcAgqtD/66CN8++23yMlx6/JnzpyJ\nr776SvWBEURoJOaLWwgX68Ka8o1osAbPYBXPlDSfxI6a3ZLb+L6ga8y12FDxi2eZb75n6TZCH6Mv\n22p24URLWWQak8H6yl9QZ6n3CkiJ41i7sxJbD9V5/l6wxhuytetYPQAgLUhOcF9h9+3hVV5nKsF+\nQxbbomv459bm6MDqsg1o7zTuqumnAAAgAElEQVQHjLPOUo91FT97llW1n/b8lv8x554pry7bIEtD\nIvQxoBE5Dx3OTqwu2+DRlkSDoDbtjIwMGAzeHLN6vR5JSVGOJicImXQfkQ3sqtuHJSXLsaHiF7x4\n8VOxHk7IcDm7z+8xRtb2L2ydAwDom9kH/bOKFag0IzO7E8tnrRaNtib8Y/s8JDHu17HYDLKprQOf\nrz4GbV6TJx/4mh2VMFzg/n3ydCsYpKAgx/2+Dpipdqm6fZd/vm+J53dEZ9oS6+bsnO93blecWovV\n5RtQ0uxvj2dZFq9sfws2pzcEq9rsLWAi92MOQJe/QBnqLPWy9/FDxExwpPEYfq3ejhRtCib2vii0\nthUSVGjn5ORgyZIl6OjowMGDB/HDDz8gNzc32G4EESO6j9jmvt6bOppjPBL1ERIYFoc7U5dcVauS\nmZeScUSDTmcnkpKkhXZ7Vx3rAUVZ4GpPPf/HsXjl4AoAwBO3jkFBRjZyM/WS7YgdoaDQDtmmLf88\ncrnQfQUyNx5fgQ0AWsar2leigufyrnN+BFIItSp2Fjq6KqZF874Jqh6fNWsW9u/fD7PZjKeffhod\nHR148cUXozE2glAM2bTjF6kZs9Bl4+yFcoWxWh7L0YLz1BYbgrlLaGen6T3Ligu9dutexnSPwAbk\nF73wWRETOBt9gCOhwHh8r4+S68314ZCVDCV4u+lJaQAAZ9c5jmalr6Az7czMTDz77LPRGAsRI2wO\nG36q2oyLi8YhLSk11sMJi+4Q8hVJXKwL6yp+xmjjSOQb8oLv0EV7pxm/Vm/DxF7jodelRGQsLMuK\nT1kEJBV3LV0y1aByhPse0wGkaJP9clH7tREH909FXRsWlHtt1Ta0oYE5ieTmQQAAfbIOEEjaxb/3\nA4Q25yMgcp6Enh25HzHbanYh35CLAVn9fLuSBTeb549X6Fr4jtF3bHUWE/bXH8ZlfSYI9sF53DsE\nbNptne3YXL0dl/a+BMnaJFS1VQuN0u8v74eGo+vv6HnZBxXakyZNClCRaLVa9O/fHzNnzsTgwd0/\nAUR35/sTq7C+chPK26pwz8jbYj2cMIn9SzdShFqT2Jf99YewpGQ5VpdtwCsTnpO93xdHv8Ee0wG0\ndbbjd4P/J+xxAO6XshZawXVCV417icv10JVj43w/SAx8PGTUO17ZjENllZ6/U0ZthEZvRWeTDUAR\ncjL0wkKbN3SlM21BbYeM50morrWSj2fuPuePV+ha+Altn9+v75wPs92CHmmFGJYzSLQvobSjHx1c\ngKNNJXC4HLi6/xWCecf5T6JHaLu4mbbwfa0GQYX273//e7S3t+Oqq66CVqvFqlWrkJycjIEDB+L5\n55/H559/Ho1xEirC2XlMoTppxBGxf+VGkAh8vHN2cW6mIReTtQEAUG8NKfG0IE7WJVEQKfDKcTNn\n2erxBLZpA/4faeNH9sCVY/sAgMdmfeX4AkwquhDVjhOASaiFIDNtbisF50nOls4wE4xwk0K+elzo\nI8x36L73hdluAQC0dLRIakuEhHZNly29ySYcH+90OQMmrh51e9eYxbzL1SCoIv6XX37Bo48+irPO\nOgvDhg3DI488gh07duCKK66ARhM9PT5ByCIOZkqRIhIz7VDPhkZk9hPeWJTZtF0em7Zc9Xj4Y41F\n3K0Hxvt/YY4BxYUZfjbr7PRkFOSkylZvi33siH+YhKYeF0pqo8jeDE5oy5lpu/z+4tPpskveB0I2\nbe48iWm47S57wLPonWm729PGk027ubkZx44dw5AhbhvQyZMncfr0aVRVVaG9PXqxaUR0ae5owbaa\nXZjcZ4InD3K1uRZHG0swqff4kLxK2+1m/Fq1DRN7XwS9Th98hxBQU2TXWkw42HAEk3tfEnD8nU47\nNlRuwrge5yMrRbj4QiLBdL2E5NqTjzWVYl3FTzDoUzC195WCebn5L2VfWJaF0+X0i8mVY9N2uBxY\nX7EJY3uMiYwjmkp30I4jdThWwYsC4GlUbR3OLsHNIlUfqJPwng+Z3vQi501U6PMWH20swc66vUH7\ncfCEdknzSWyr2SVrjIB3pt3p9C/cLZYRbXXZBpxfOFrwOOxOO9aV/yTal1CucK+QZ7ClekfgPgKC\nnrNhc8fOxJPQfvTRR/GnP/0JFosFGo0GGo0Gd955J44cOYIHHnggGmMkYsB7+z5GeVsVdBqdx7nj\nxa742f5Zxeib2Udxm18eXYLddfvQ0tmKm4ZcF9HxRoOXts6Fk3WiV1pPDM31t5utr/gZy06swIH6\nI3j0vP8XoxEKEZoQ4mYOclWpb+5+z/N7T/UhvDEpMMJEeibMYlvNLiwt/SFge6kx/FS1GUtLf8Du\nuv34w1nTZI1VCrUUNR//eASWDv+XPxdfzdFpd4FJdv/ukRfoEOo1F8gTxoq9x3nL39rzb5Ht/OGr\nx7m4fLmICTyhUe427cdu035sq9mFu0f+PmD9saZSHGo8KtqXUFUu7mPS6XLiv4cDE4e59+HPtN1f\nXK6uY48r7/FJkyZh/fr1qK6uxtatW7FkyRJ8+umn2LRpUzTGR0SFwMfjdJedp7WjLWCd1SFeb1aK\n+i6beSTtpIGoN9fm1IBc/LAvTV15m2stdQHrQiUSVrJQhRDj8Y5Vri7mz5g4goV88XNpszJs2k02\n9+y1xlIbtzZth9MFS4cD/Xtm4s7fDPUs52zVHBmpyWh3dGDs8AKc1S8wFwYbRGjzUWrTDvXIwy2a\nIWYPlrqe1eZawfuipVM4RziH0KzZ1fVci117u9MeoDrnxuyxaceT0N6zZw8WL16MH374AS6XCy+8\n8AKuvPLKaIyNiCUqTDmUqlxDIRoWbTVdTvxfVBER2yHtpfFkz4rctQo20w5c0hWiJHG/uFiv924w\ntbEcYaeG9zg3w87NSPGzUeOg/3Y6rQZwAIZkMQ97ac1DsJAvbq3YeQr1gyXcQi1ipjap8TAMI7je\nFmRCIbQPd57EfEjcgl4s5IsT2nEQ8vX+++9jyZIlsFqtuO666/DNN99gxowZuOaaa6I2OLXYXrMb\n2SlZGJwzIGp9OlwOrKv4GWMLxyBHnx21fuUReMNxt7aU7brTaceGik0Y11Pajnuo4Sg6XXbPja5m\nWE2sQnYiNUPzbUfue8DFurC2/CecWzAKeQb/GZrvqNaW/4TL+kyQ5Y/gUf+JCLpaiwkHuuJi5fo3\nSM2YhVYJeY8fqD8MF+vC2fkjsL7iZ08WLS2jkbz29dZG7KzdE3SMQoJvbflPGOEYKLrPqu0VMDVb\nwYJFNXMQWWxPNDGVMLKDkII02LqEdmqQnOAcBxuOYq/pAM4xjvQfm8BM+9fT2zy/V5atx/8MuMqT\n+EP0I0XsPAV5dqraq9ErvWfA8rCFtoiwrGw7LbgccB/bylPrApY32JTn6ee0SWLP8PKTq1GUVui3\njBPScRXyNW/ePAwaNAjPPvssLrzwQgCxLdMWST4+9AWA6NYr/vX0Nnxb+iO21+zG38Y9GrV+Q0Xy\nK7frIVtX8RO+O7EShxqP4s/n3i+6/fyu+rsDsvoCiKxHckyQeA4i4fGtlB21e7C09Af8XLUZs8c/\n6bfO9zouLvkePdMKMSJvKL+JAIJ5j8/ZMR9mhwWFqUaMzB8ua5wuyZe7kOdyoPc4VwN5xpg/4ZuS\n7z3LGYaRvK/m7pyPls5AU0/gGP3HwdnZl5YKb1/fYsWXa48DADTpTUgZsdWz7pT1EDr2e5N9CNmp\nfeHunZbOVvx7/6ee9xMD96xSyDHv8yOLPL83VW1BW2c77jv7jq5jCT9O25e/b3sD70x+JWB5uOpx\nMbkiZZsGIMtJTg7BPrj31R/Evnp/tUhbV3GTuJppb9iwAUuWLMFzzz0Hl8uFG264AXZ7+DVkz1Sa\nu+qx1kTQ5hkNpIQQZ8fl5wwWb6tLPa6i0I5ZnG2EZvihqMe5Wr9yZhmWrnjWYHBaEbHZsdnhbkdJ\nXnSlucE54SR0v/D9CjTQSF57OQIbCFTFc/HqYrRZ3O/Ei87qgYHDs/BNuVdoawxmzLrL7W2m1TDo\nGURoi8EwDFiW9VY9kziPdRZvALfY+Ra/DsGvj6BHNysnNag4sfjQDRdOWDtiEPIl2pPRaMR9992H\nlStX4u9//zvKy8tRVVWF+++/Hxs3bozaACNNrGZ58Z1eM0S7Z9ftI/fYNJ50hfF8LoIj9IphpVYq\nIJT7RNJZjN+czBmBRqb/gc3RIas9IJTc4+JCSihDVWTitP37CmYjtdjcL+3CXAPysgLTvfYpSEef\ngnQU5acF1VSKrfdkDONm2lJmBp/f4veFvJAv4W2EwrDCO++JKLS588AJbyZ4ypOIIaunsWPH4uWX\nX8bPP/+MSy+9FPPnz1d7XKqhRGBsq9mF400nItp/NDPnhEqn0+7jnAH8UrUVZa0VAdsxCoXw8Wb3\nuSxtCayvbLFbsPLUOljsgZ7ZSghm0z7ZUo4Xts7xm5HIoaKtyuev0NTjFrs16DG2CHjrB8PXprjP\ndBBlrRX45fRWHGk8jhWn1vLGF8hPlZtR4WM/PN50AgcaDgMIFLQnWk5hc/UOz3EGE2q+SAnVn6s2\nByxbcWodmjtaZDkuNnU0+5VdPFB/GPtMByX28LKpagvKWitQWtOA13/5xLP881XHsKHyF9H9Pl61\nD6u2u5+LNH2S4HNQazHBxbqwumwDTJYGOF1OrCpbj9PtNQHbNopoShiPUyCLww3HsLR0ueiY/Apq\niJzv0pZTYntLjg/wd2JjWRZWhw0/nFztXR/Ch1O4NvFYwN2TXCSNNoqJxuR5RnSRnp6OadOmYdq0\n8OMhY4WSmyqSdXU9D1MC+AWs9UlOYHZYseJooMMH4OthrHx2+NGhL/zqK3917Ftsr92NRlsTbh32\nO8XtcQQbyes73wEAzNrymqLr+vL2Nz2/Q72C35R8hy3VO1BnrcejRXcLbvPBgf8q7sf3nn5v/ycS\nWwZSZ6nHwmPumsrc+Zi3+1+CbQPAnJ3/BAAkaZJgd9nRKRD3KmecfL47sRIXF43zW9bS2Yp3936E\nawcIRasEnh3OVwXw2r6DXeO2znZ8cXQxACDfPgzNSd6Ps7V7TsFwvvi+W1pXo/OE+x7ukZsKGxuY\nBvj1He/gtuE3Y2npD1hdvgG/G/Q/+Lb0R3xbGpjfWgxP9S+weGfvB5Lb+qr3Az92pJ8OFsDWmp2S\n4+NX2fq29EccaTruWRZKSlODSomWlKJEy8X/QIu7mXao2Gw2XH755Vi8eDGqq6tx++23Y/r06Zgx\nYwY6O4VjOdXGFabTRKhwN0T8i2x/O6VTopSd92USvlqSi2/m7OShEyvVe/B+G7ri0+sl7KS+Pg9y\n1YbKZjf+bXaIxFR72xY+Lq9aUP4dHWzG3OEMVLVXm4XjryP1HPmmtTSz/vfeE78/W3LfvB6dmH3X\nBXjl/otwVv9cwXFaHFaPNsJstyjOAQ/4z7SD4rOJ0m9pFmxArLzQNt72WdTyNFahVEnTacQz0kcT\nJeeLfy3iwqYdCd59911kZWUBAN566y1Mnz4dCxYsQN++fbFo0aIge6tDrOypXqEdj2KbPybfsKPA\n8XKLFL1MgsDZ37Rhhk7E2ndA6uoGq5esqDEflKgXlUaAiHl8e8wnCtoL1eYsJAjUiGTh22aNudIz\nQAfbid4F6TBmGwCICyydxntPh5KEQ+Mz0w6Gn/pa6bPAAsFuOl+VOws24KM+lGscD5XVlML/AI2r\nNKahUlpaipKSElx66aUAgK1bt2LWrFkAgMmTJ+M///kPpk+frlb3ogi94CrbTqOk5SQu7X2x6H5c\nvOY5xpGCeZX5rK/YhCE5Az1xjWLq8VqLCfvrD2FKn4khv4gONRxFik2DgfrgZVKbO1qw6Ph3GJBZ\njMuKJ6KstQL76w/5bSPXg5lzRAu1BnGjrQkH6g+jf1Zfj1otlC/Wn6u2oG9GbzR2NKO6XZ4nezgI\nXSeh9862ml3ITslCjbkO/bL6eK59actJNFlbwH0zc9sNyRGPBeazpXoHjIZ8DMzup+hFua1mF5I0\nOrAsi1HGsxBMQ+D7kWuyBGoI1pRvhFUgQ5xYW0cbS9DW2YZDjccC8s/vEIijZsGG5WB24nQrNh+o\nCXjTfbbKHU5Uwxzy5ADvdPgLoOUnV0m23W43Y8WpdZjSZwKStEk4JeD3AQA6jbdzodzWweDuN7MM\nz3//Z9f7u6pdqEY0b185HwV+6vHAcC8hbUnwfuMlBFSBepx33HER8hUur7zyCp555hksXboUAGC1\nWpGc7E6sm5eXB5MpuCNQTk4qdLrIBq232Lxp7oxGd0KQB9fNAwCMHzgavTJ7BOxjNGZge9XeLrvU\nenx041zJPk41VWLR8WUAgK9ucefhNVS6j13DMJ5+AeDRRU+jw9mJ4UUDMLrniJCOaf66D/36kuLV\nlW+hrLkSu+v2YdKQsXh13duedTqdBkZjBlJOetVVBoHCBVlZqTAaM5BW7faWZVmX3zHJ5aVtcz3e\nxz3SjQCAVENKQFtSbTdbW/Bll02ST7AxhTJmAMjMNATsqz/pfpQ0Wo1n3SfrvvTb5uzCYZ7f8zZ/\niFmXPeq33Ve3vOv3QZCZYRAcK8uy+O+6rzz7JJfLf0b21x/yfKR9dcu7aNelCfbBwWi8y59Y8rxg\nm7/4JPiQaistIwkvblSWlxoAMjICvbIzMw0CWwayYnsFdhyuDcjzvW6X23ZtuMDrAOd0ufxqeGyt\n2Rm0/e9OrEBGuh7XD78KG9cFOq0la5OQm+09F5Xt4glD+HDnUNs1U5czHo3P9WqAN8Ts79vegDEt\nT3LfFL0OaWnJktvk5Xnvl/y8NEDjL3CrHVX8XQTxvT8MtdJ9RosUgXedGPwiKfl5GTBmRqdQkCpC\ne+nSpRg9ejT69BEuKiFXHdLUJC+mVAnadO9NZjL522+qTY1I7nDflL5jNJnacLrePcsw260B+/Gp\nbvbm1ua2tVi7vkBZ/345m2JVvQm9dMo9h30JNi4AKGuu9Pw+WePvIepwuGAytcFq89o5fX9ztLRY\nYNK0wdYVo+piWdG+pa61b7gQN8txdPq3ZTRmSB5Xk00813Cw8yHnfAnR2mqFKdl/X5vNfS5Yl3i7\nDrv33jvdVguTqc1vFmkytfmdr7Y2m986Dl9nH5OpDWZLaLngTaY2NLWZ/f7m43S6PMvNnfKfR6G2\nqhuU55xnWRbNLYH9trbKm93XN5uh0wbOgmbf7Zbi/9jvzf/dt0c6Ki3KM2pVNdaJ30ssYGkLLb+F\np00FiiyHz/Vqava3n7ucQXwKbA6YNdI+Dqb6Nr/fNrv/9vXN0rm/Pfv6nC+zObT7N9J02ELPQ9La\n1AFTCJEfYkhNKFQR2hs2bEBFRQU2bNiAmpoaJCcnIzU1FTabDXq9HrW1tSgoKFCj66DIVbWFYxsV\ntFsHCeSNRfy4qFrTz5lFKFOV+3/OjiN1ruQel7curTLNSjTTB0qh9G4RqyEsx+AdLMd0JImkj4Dc\nxC58wjk+s82ONH0S+KKotzE9YNskXWh2yWCXLFzVqRI/GL6jmNJ+gvXk276LZQPU40L1qoORiDZt\nPr5+C6r3pUaj8+bN8/x+++230atXL+zevRsrV67Eddddh1WrVmHChAkSLUQelmWxvuJnDO5ZHNL+\nSh47oWfU44jGMDBZGrC7bp9fVqdgL6Zacx32NxyWtH2vOLUOU4oneupf81l1ar3f3/xqXSxYrC7b\n4OcRKvQ4cU4Ycl5Gcl+4XMYqLe/m/+nUVug6U+BwOWFxWHFe4Tmede2dZqwu8z8mX3z9BWosdVha\n4h/fuvn0dhSmGTEgqx9+qdqKovSe6J8V/P7wfbVtqd6BqvZqbK7e7reNUMyt735NthZ8emihnx2U\nn6hkr+mA57fVYfOExvjaFb8/sRK/8voOlRWn1mJKn4l+y5o7WsCyrGJ/C7PdgrQk/wxgocbgC/lM\n7D0pT83cYrYiLyM9QGhvqtqCQw3+aTJD/UBhwXqqjfHpdNmxRqK+sxyUnHsxmzYQGSfcZp/ojlOt\n5Wiw+WtPhEpfCrGh8hePD1GoPjGRps4aGLInF/57S01Us2nzefjhhzFz5kwsXLgQRUVFuP7666PV\nNQB3zOc3Jd9jeMug4BtD4OtPwYMj9L3qG/L16o63AtIwBiuB+NrO+bA6rOiRWiCa6/m7Eyug02hx\nefGkgHUtHW349oR/7CVfSFS1Vwc4rAjPtN3L5MxylZZ29HVEY1kW72z92G+9r9D+7MjXAU50vvx9\n61w4WCd6pxfh7T3vB6z/7MjXAIDXJjyPBUe/AaA8Jl+o/i4AvLU7sBYx/+XLt1GurfB/uR9oOOL5\nvfzEKvzvkN8C8BcuP/KSpyjFt63vTqwU1HQcbz6hyEkOcH/MTCn2/wCwOkNTgwolCdncJu0kxtHp\ndKBnXhr4IvULAT+IUGf0LOvyxIULwSWqUd6u+2NJyeeS30ybJwyVpJwV44P93jwCQs+UUOlLIb4+\n9i2G5w5BYaoxbmbaQgmkxBicPcCTLAoIP+pFCaoL7Ycfftjz+6OPxG9stUnRup0dmq3iNhepG14Z\n4t7FDMMIeoEGK4HIqbL5Dx7/RdNsE45zdgh+AYeWa1hZEgJlcfG+N3+wUCbfDFhCcM4iwTxa5b5o\n+Ei9bIRyVgd7/Uqpj33ziqupDheqR9wcQuy8UM3xUF/Ocvfrk1aMCnO5p8AGADzx+9EYVFiAGTIy\nL4dahpSFWxMWaViwbpW1gogK/4xoyvsMdo/W26T9EpSox7nnUs1SvWoxrsd5MRPa0QsuizEp2hQw\nYNAq01nAT8mk1DYkpR4XtWnL64Ov0uYLHLEHQGjGK6dHqfqzcgSy4pm2j5opmKpNbqiITsRcEC5C\n50bqpRdMWWPQ6UWvie+u0Z6Z8O85OQh9WIQUwwtW9ks9Sec+S76x0IW5enedahmErqZlVYnT5a5z\nyDbtGKidFdm0u4YXLzNtJfDvyWjatM8Yoa1hNEhNMsAs167mcyM5WWfQx+Z0ew3WVfwMQFo9LoaU\nAPTNosRXafMFm6c2LMtiTflG1HRV4BJqX85DLfRArTy1Hi7WhQP1RwT28Efpi3p9xc9gWRY2hw2L\nj38vuW2wmTbHzlrpEn788/BT5a8ob3V72W+o+MUvLzfgttd+W/ojDtQrU3uWNp+SXP9z1RYE+5Sy\nOqz48dQaRf1KIuN9qSS/uKdZlkW73eyXDrOmSXk2MABYekQ4jS4fLjmKrxlij+lA13kNjpxYZiFY\nllUlI5YLLLbV7FKk1uae1zpLvaD6WootNTtCNmFwKNFarSxbBxfrinlCpFDgv9e6lXo8ntBr9bIS\nFAD+7zKHyxn0a/elbe7Y7cHZA4TbC/LVLDUj/frYt57fnby0k/wvW1fXi6uk+QSWlCzHkpLlmH/Z\nq8Lty8qKGLjRydYy7Krbh7K24DYgpcUA7C4HDjUehdluiZiDVbD4Vt8HsNHWhIXH3LkFnh33OL4+\n/m3A9j9Vbcap1nLF4xBSGfsileKSuwrfnViJjZW/Ku47HPj3nBxcrAtfHV3qV/P4SHkjkKu8fxsj\nTz3v8qRW9cLlS1ATFupkxGJZ1lP/QC6ctkDovpXDrrp9Ie3HIdcRDXB/UNWY6xJypj0wq7/f32pk\n6BPjjJlpA8G9nf3vHV81k/zZYtBShSJDkJqR1lu9diT+zcFPvci1Y+YJCKH2Q51pA27P7WDtSy2X\nwmy3iObEVuMB9x2jryq4wyU8BqkMYGo/vI0iXsqhEnAPCJzeUI6JBRtQSa3TqW7ef+7DNNqpgtWa\naYc2A+2aaZuVVbHjsIfwgeaL0pAvJ+tMyFK9RemBSbiixRkltIO/fFiBX8pDJYQfti6P6xBs2lKj\n5gtFz4ya117IQltkG7mxwkpt2p72Rc6HGk5YvmP01QyIvfyjqQrj4Eaidt/Cdl3lQtDFugLaCtXR\nSy6xKvHIQnlInBxC0XBwj02sVM5KhXaiqsdjyRkltDud/qqbH0+uwZpyr1vp7rr9nt++QsPFugQf\nympzLZ7b/IpfqMDKsnVoEvDg5tqzingy8x0bLHYrfjy5tkud7+17W80uz+8TLWX4qcpfVcqpCH0f\ngxMtp/BTZWC9YjmzVrFt+HbV5SdXo6q9GrO3vI7Djcc8y0Mp1ceyLNb6XBdfXD42+0jhO0bfZBFC\n6TmB0Io+RIKTLeXYY9offMMw2ChQP/rHU2sEQ9ikOFzWiKY2f/uorqBSZOvIwOUXUFIuNBJ0OjvR\n1tke8XbNIVQEszltKG+r9Is2UEJLkCpfwVAaieFiWcFa6oQ4Z5RNmx+68j2vIMDaip9w4+BrAfAz\n/wjPEF7cOgcA8OoOb/7uw43H/IQWB9ea2Nczv48lJcvxa/U21Frq/LyOfe2ec3bOF23Hd/xcDeTA\nMYU+0+b7BqwqW49VXYlO3tnzgSfeOZSZ8YmWU6KhJdzsLZJf575j9K1aJPYykRLaailmWXhrgUcS\nuVqko00litqtqm+HJsMOTWrwbROd3Sp9SLWHmEHule1vRXgk8lEqtEtbTqo0EvW4sIdEkfUocEYJ\nbWXwZtphvo6Vzgy5jFrujEPKyx+GM4v220Z2z9LjUYJUiBGnSQhV7S7cpq96PHi70vbLeCy9Ko5a\nqsnzhxtRYbaioSPyM9AzhVBm2rFGOB+EOHIdg+OF24ffjHE9zvNb9pu+l0V1DGeUelwJvvIsEnbU\noC/HEG3afLxq9tBn0X7bhKmGDsXOKJ3LvGumHUGh7TtGOTa5WKjH1fsUUEdo61M00MZHWviEJdEE\nGhCCI1oI5rNYkqozBJhKU7SBVejUhIS2KL5x2i7V50/Sr07/3teUb0S1Wbhu9P76w1hdtkGWsJXz\ngIVb6zaUGbHU2Lkax5GcafumC5VzTqScjhpsjahsk19+US576w9GvM115T8FRB9EimNNpbLj6Alh\njjQej/UQFKNUPc5P3RvvCD370Qz3Akhoi+KfEc2lKPe4EJEMa1hSstxjTxdiaekPONZcGrSdX08H\nj4MOd6YdWhYscTZW/qW8954AACAASURBVAqnyxnRmfYKn/zd/Dq5ofCP7fOCbxQHfFPyvZ9jYyQR\nKphCKIMfMpcIhJoSOFGIdjihECS0RWB5M+1wCZbyU0olHEppP7E4Z1+E8kzzCbcCj1p5siM50/ZF\nznjD+Y7J14eQXURF5NwDiQS/SIlSLi+ehOKM3gCAHmmFYbV1fuFoxftcN2AqzsobBkD96lcj84QL\nD4VDKKU5Y0lRWuzirUOFhLYYrO/P8B+eaNfLlhPPK2sWHeahR9qmza1XKyGDPOe80K9lenJgHedY\nEos67mqSqgvPXT1Zk4RUnQEAYAjTVskvTSqrf20y0pPSAKj3YcrBlXqNJEoyosUD2fosRdtHWxUu\nBHmPi+A/0w6ee1yMzae3o91uDprXmGVZWB02LDq2DAadXlaSDykq2qqCbiPnY0QofE0uJc0nsfDo\nUsX7+daRFoIF8JNAPHEkkPPSOdFSFnL73As5Xqiqb+9Wn+5iyYtk789oPC/mcD8LNaGcWMb7vLtU\ndtJSQwAlmvOc8mRFJLTjDq6GrV+1HJZFqBeLq9ksh+UnVmFLzY7AFSE8XKfNNYr3ESIcLcOCI990\nhaxFFruzEyvK5BWRUEqwPOXhkppkULV9pTRbzNDE1+Q/PMJ8p/oKsnA1bKFGGXBjUHumHQ/22Vij\nVGiH+1EYCbrRN3Zk4B5UfsiX2moRF1yizjuxv01CoyWEOsxyUNPZJWju+DDRRzk8JBhFxu6V/SRc\nQcR0/QMQ9lQ7FKHt27+UaUnJcRak5gu3EQeq3kjSN6OPrO1uHXqj57dQzgWDTuLDOg5OGQltHl4b\nX2Qd0YLBsqyo40mifhGrZS91qKg2TNYmq9Y2EP2YzmBoNN0r73O4gohhGHhldnjnJtSxaJjgQlsJ\n2SnZwv10s9d/RrI801OO3ns+hD6sUiTeAfHwLu5eVy0CuFgWq8s24HS7V738xq538ZNPOcSVp9ZF\nXCC5WFdClqgTY96uf6mWA3qNSF7ySKC+0Fa3faUkmrdvMCI60w5TaIdS+YuBt8ynVD5zJR8UYuPo\nbjNtnSZJ1na+zoqRSJQU7YInZNPmUdp8EktLfwhYfrz5hOf3shMrIl6aze0NLTLTTsCHy/d8RRp+\nkZRIorbNKkUXXzPtSMSlxxPhXj2GiZTIDl0gdIgUFQoVMaEdq8I3kSZPn4MGWxMm9b5IVkEdncZt\nx8435Cn+sBIKv1UrQZEYJLR5tMvM98svPhIuLgSWMgwHHaPtdi/kcDi/cDQYMNheu1tyO7XLO8bb\nTNsZJzPt4blDJCMVrh94teDHNB8mTEHkFmScfjwWjmgMzi0YFdGkNxoRZys1aoCrQWGq0VPBjc/D\no+/FkJyB6HB2wqDT47UJs2DQ6ZGenYQ/LHlUpEUGcye9CB2jxZdHFyscjVdo6zQ6OFwOWJ1WhW2E\nR2JctSgiV9Dxy3yGCyuhHg9F5RcNO3wika/PlRU3q6a9HACSZKrwokW8fNgFuzb6KGoomMj4oYX0\nDDKMuJANFbGPh+4w007RpkDDaDwx56lJ7tzgqcnSURop2mRoNVoo1c34bm3QuvtU23mVT+JftQgj\nd+axrPTHiPbrYtmI2rSpsLw/jE/8rRROVt2Z5y97hXPGx4p4iasN9mEq59FgwIRtSnL7qnBx2mFm\nAwzxAzCUDIhSiM2odZrEULRKXYdwayMInWrp97B3B0MSJ7TFqxKqAQltHnJnHpGeoUg5tnWHL+JY\nc1beMFnesmrPtA8ctcFlja8EK/EBgyv7Tg6rBQ2jCdsRLRJleDlC03YxEVdbx0I9Pr7nBZFrTEKG\nGg3C4WwA0D+zWHA5I/FXMHw/Ci/vMwkAcG7BKEVthAtJAx6xKhXHQjzkK5qqwe5K/6xiWR8/Smza\nw3OHKBrDpB6T8eLvr8TT48RsbWcuDANcN3AqnrvwLyJbBJ/1yhXaU/tNEV3nklHR738GXBW0DyB0\n/wgmwq9lrUa5evz3w/7X83vOxBdk9fPGpJcwZ+ILeGPSS7h12I2C28wYcx9evuRZXFwkT6jfPfI2\nz0x7tHEkZo59xLMuWZOEDIm0wI+e9wAKJIQ6ICayxe8132iLi3uNw+sTZ2F0wdmSfUSaxNCPRJFY\nCW13yJfwl3lynNlBExU5qlMlM23OC1UuvbJz0TOPZtlCcMI2O0V+LmiDzgCrw+sEpGU0spIHSiXP\ncPpU9BNXk8qbnYUitBmIC9lQEfceF79/c/U5nt9yJw3JWv/3VHpSWoBjLwMNMpLTZdvtkzQ6jwhN\n0iQhMznDsy5Y7nQNo3H7SvD8xMIxoVgc/o1JJmJRCRLaPNRIuymHHbV7BJefaClDQaoxyqPpnsgJ\n51Jy/StqzYACuR3pGVT3RMkL1V+o+nl+h9gD6zPTFrOlyh1hqBOAiM+0RQSk1HFEIt2u8PlzL5Ob\n54IB4+PQoFzYaoN9WCsU4FZHdD3FhaC3CI9Np7fGeggBdKekK7Ek3HAgPg0tyrxGfV8PZxcOk73f\nBT3OVdRPIsLNfoReoWKaJv5joWE0sj7MpO4Dl0+dARbucCO59M30T6M5NHew7H05BmUPiLitWWl+\n7azkTOgYZfO5oTmDAhf6XB+jIQ8AUJhWAAAYI6FS9i2qwzAaP9Hvb/4Ifq0nFF0YdBs+vv3xTQj9\nM/sqbi/SkNBOALpT1qoLe5wfsGxgVj88eu4DEe/r4qJxfn9HOknNWf0V1sb26f9vkx7GM+Mel7Xb\nHcNvCTg/fxxxq7K+FRBq20lheCNzL2P+NXrlkufw2sRZsvy4fVOQhoqfTZtl8bcLHsWI3KEB203q\nPd7v74dH34vHz3vQ8/drE2ahd3pPRX3/v1F/RI+0Alkfl0pMZmKzTSH7/+sTZ+G5i54QbWvm2Edw\ndv6IgOUPjb5HdJ9xPc7Dcxc+gdcnzvaot8X8Qa4ovhR/v/hpz9/ujzD31Xd71Su7wOf3GIPRRvEP\nBO4cJGuTBd9NL138N1xePMnzd6STaoUCCe0EgG9HSWSEwllYsKqkD+XH/kY617IhRZmvge8s0De2\nNBgMwyCNp66MZCIePqGGAoWjyWA8//vfH+nJaRLj8T8HWkYrS7Us5azmZJ1emzbcAi9JG3id+bNX\nvS7Fb1aWmmRQrCHTd90PckK+9ApqYYs6nAl0Y9AZJBMAGbQG6ARm7kJ9cOpxvS4FDMPIut+1jMbv\nI4NhmC7tB/c3/NbJIVWiX66FJI3O06+vWj9Jk6TIzyIakNBOACrbTsd6CBFEQGizrCrJQ/kvl9Kq\n1oi2H27KUyUzf75AVDV5TqiFLsI6H0HitAU+UvjL3N7j8roSE9z+9yLrWRYMoQ9CpZ9V3JjkqMdj\n4QAFdDlfKr4/lNzn/IAsxu93KOF40s+ZN5OO0FbxGMtOQjsBaLOLFw5INDIFQjQuLhqnSn51rc8D\n12bpxL7Shoi2r3TMUi+kYPBf5AOz3La1S3pJ2+x6pxfJ7sM7rtAI5xpyu4bThkZmchXf896Lp8J2\nsi6MyHP7G1zQ4zwA8hJ45BsCTSVKk7NwJTSDaSxG5A31U9mGitT9J3YetRqt7PsjNF2Qf+t+Wgde\nx+Nlho3xjzPXp8pXsNtFx2jjorKXL6oJbavVihkzZuC2227DTTfdhPXr16O6uhq33347pk+fjhkz\nZqCzs1Ot7mPCjDF/8vubb7f509l34r6z75Td3uDsAREZVzBmX/Sk4PLXJ86KeF9J2iTMmfgCLusz\nwbPsoqKxqjwYviFZ7VY7wIr3kaJRHguvdMz87RXNtHmPakGqEa9PnI1pQ27wLHt94mw8dcH/+W03\nc+wjuH/UH8IaJ+COkeUzMKs//jr2z5L7Kek13DY0Gnlx2u5KWu7tRuYN91vnYp24qOf5+PvFz+Cq\nrmQvUjPtFG0y5kycjVSBNKxC+z123gN+9Zw55kyc7Yk5DjbTfmDUXRhfNFZym3ARO2YlQoxrIxwN\njNsRzd0Of6Z9db/L5TYCwO3gNmfibEEthWikABN+lr1Io5rQXr9+PUaOHInPPvsM8+bNw8svv4y3\n3noL06dPx4IFC9C3b18sWrRIre5jgp5XKzmVd3MYdAbk+cQ/BoPztFQbodkvoI4KjmXdNi7OJqim\n+snXA9Zsc4CVENqhhLgonmkH/C1/fyE7p0Gn9xuDjtEG3HMaRqO4hrfQcbkEXuIp2mQYfGJ4w5pp\nh9AGf0QaRitzf18FuH8r3HFmpWR42pJOasmI2peFBIGW0Qo+V75tBLPLKz3PkYw+0YbyvCoYrtAz\nwo2f4W0g9zxwHw0sWIFrJdAGG3SLmKKa0L766qtx7733AgCqq6tRWFiIrVu3YsoUdzaiyZMnY/Pm\nzWp1HxP4Aoh/UylNmpCiiU5FqOjabboewK5zw8VrRjrfMgBsO1Tn+b3sl5OS24bSf7iObUpqsstJ\nRsEwjKBDkNIjExLQDpGc7HybY6gEdWITDPnl2bRl9u8uvym8rUtAFc4XeuFqhfhJefjtRTq5ihhS\nQk9UPS4zh7+bED4W+CYkxus9DoRn05b6eGF9+g782Iovsa3623ratGmoqanBv/71L/zxj39EcrJb\nEOXl5cFkEi63xpGTkwqdLrIVb9Sk0Jjt93dyiv/pzcvJQFFGIbBdXnsGg7qZ0IoyCnG6rRYFBZmC\n643GDMHl4ZCalgKjMQOZte7Zhot1wWjMQGeKvJKoSjhe0YbkLgvDgRON0BaIP3w6rfL7zGBQ9lGV\nkWnwO6e9e7htmMPyB+JIfanofkZjBlJsgcv4FBgz0d6pDdiuHtJZ2NKSDDDb3REKVw6aiMzMwJnj\nVUMn4NCvR3H98Kuw9PBK97ZDJyA/zzuO60dcic/2LpHsSwyDIUnwmLhlFyaPwtfHv/VbxxceyUk6\nZGUGr+SWmaHH786aiq8OfI8Jg87Dnvp9qDO7/R0uG3xhwDiSkv2FaFp6Chy2Ts8YxMadluV+/sf0\nPAu7qw8CAHJy0jBAXwT4lH2+ccRUvzaSrOIfcwadXvFzmZoqfJ+mp+mRlpwKc6e7aEyOPsvTdkey\n14/Gt78eBdm43DkeO2r3eN4f/G04nF0fQOmpwmPuldkDVa01AWPy3TY3Jw2Mxn2dDYZk5Od5tYLB\nzgO3fjI7Dhsrf8X1I64K2Ce1wn1uGAZI1bt/azRuT3erwwajMQPpzSkBbcYS1YX2l19+icOHD+Mv\nf/mL35eOHJVNU5O6FYim9puCovSe+PDAZxFpr63ZP9lGh82/fGdriw1ZLgfmTHwBj/30TND22i3K\nqsf0TCtEtVm6itQD59yFf+79DwDgr+f9GU7WCZOpTXBbseXhYDbbYDK1wbcwjsnUhiZzZIT2zLOf\nxCv7/wEAuOXSIVhSfgAA8I/7LsTupiR8V3ZYcD+GlZ7djMo/Cz3SCrCqbL1nWYfNO/s06PR45ZLn\n8MgGf/+AORNn47GfngUAtLZaPefUaMxAS6MNcya+gGRtEo41leLtPe8L9m0ytcHCq8YldG3q69th\n5VUcMpna0NwsHjL45qV/BwMGDtYJm6MDmcnp2G3aH7DdIP0QvD5xNgw6PSYYLwbgVuk2NjZ7thmX\nOw5X3DgRdy7+v4D9g9FhcwgeE7dMjwzMmTgbALCkZDk2nd4aoBFwOYG2NuFnhgHjmUG1t3VgUq+J\nGDfxAuhZPZ4e+zicrBNO1gWDTh8wjo4Ofy2Dub0DVrtbaLtYl+S450x8ASnaZDxUPRMA0NxkQWZm\nbtexMGDhgkFn8GujrVP4vffqhOeRok1W/FxaLMK+Q+1mG96/7lXU1bW4w9sYjaftRp/n0be/+vp2\n9EnqhzkTZ0Ov0+PBdU8EbMNh58oX27WC6/963p/hcDnwfxu9cdkWc6ffti3NVjhdbuHfYbOjocH7\nMSF1HozGDO+zxvT0jJe/j83qHqPLxcJq8/5++ZJnwbIsTKY2tLZ5nx813oli4xdDNaF94MAB5OXl\noWfPnhg+fDicTifS0tJgs9mg1+tRW1uLgoLo2GzF0DI6pMuosSyXYLYoTnUpN5ev0kpi2SlZQYW2\nb6ykVqOFVkkezgjAvWb5saBK7XQ6Rit4foqNXp+BXJ9ZV2FuKtKs4pqLYCaCFG1yoJrUZ8wp2hTB\nJBa+NjQhGyd3LwTrX04MNANGUM0vpVLk+tVC67kmYmp77t7xs7/y+uPKFSpHfmyyt39+nLa4gjwt\nKdWbB5vxby/YcxBOiU6xZ10qzlrseiVpkiJuytKJxKEHQ0mcuNi2GkYTkJ+Bf/v62rQRQnKVYGPw\nyaTj17LveRYymcQS1YwnO3bswH/+457R1dfXw2KxYPz48Vi50q1aW7VqFSZMmCDVhOowDBPReNeA\nG45v01aYfEJpPV557cfYPtP1APIfVqWvRacrcA/+y44fpy1ltw72MhR6cUeyZGowW52ca+u2aQcK\nHzUdlyJm01a0rfDW7prpwufJLzWmwnFGNI1wGI9fpJ9cyZCvCPbDd46Uhm/T1visYUJNISDRW/AG\n4y2NtGpCe9q0aWhsbMT06dNx33334dlnn8XDDz+MpUuXYvr06Whubsb111+vVveC/P/2zj06iirP\n49/qdzoP8uxAIgkxKMgQQAEhJCE8Aogoq7zNhll21h0dxcFVB1EzggIKyMMZ5szBQWHPwrgDRldx\nxwUOM6LRjXEQBXEGEWEWDRACBEPeSffdP7qrH9VV1VWd7nRX+veZg5NU3VTdW4/7q9/9vca54i69\nCWWKUOEDILzZSivb8IzOHKWqvSJnJVcf+5l87djprtzAPClmX/t8qOHD2UqyCwGoK67AHJxoRrA7\nBCUXhZ6uco5jSvIzCwX3yIwfBfwbn7+XefkDCVY5Tbu/1bNi1dOEL4A6BznvfvfMe7zn/Zb7sCm5\nodD9s9qa6T3RtINBciUuyOsbTP/5LGBDU5z504XZBZXAl98cnJwn207ug1nHecwanNd/Q43cNQq2\nxGq4CNvyuMViwaZNm/y279y5M1ynDMjiWxag9uJn7t/tzB5UJZ7+Vhsutl7y2x4KbYmHt8GYdEZ0\nOrqQl5SDs03nZP9Giean4zhsKV3r15dnxz3hY499QGVsr1L4VyMtzmnX48OR1Kx4xJucJRm9XzP+\nenkjHKOcUBHLm317/9tw8uo3aOq87uq774t9c3I+brWNwOeXjru3pcel4XKb06nppZJVSobj7FuA\n/XLPzjPjHnN/fIZC+1cltGV6PvvGOzA6cxRW1qzz2/cP+TNhi0vH9hO7nMcJQiD5+fhKFAxZW/QM\nHMyBN0694/o7dUIslEJbyceJQWfA5tI1zp85vZ+fhJB5N81G1Tf7JPd79798yFy8/vWbrr5IYzFY\n3PZ4AHix6JeqngsAuG/IXMwZfFfApfRNE5/HssNPA/D/6HRW+fL5LewI73e01X6IqYxowonB4bAH\ndUMkE/CLhit4UDOh8g86v4wcsMQcFC6hgoNJb/Q7nvD3cIRgCbF4xRk7VHzNCvNw88cSIvyCl7v+\nUl/73rH3Qk1Zr9P7xeZ7F3JQl089gKYts9/bNhgSoa3ChicnbE16k+QzadAZgrKl+iC8H5xOVBu1\n6C0+10Xtcmdol0eVvVdmvQlmvck3D7dEW++l/4Bn903eLduWzxkOSOdgD3QuJbZvn3dPxMTo1rSF\ngdohgH+vvHV54TcavzITSnNYT4i+xKq9iJ05VDt7ATI2tUCatgLBK8Q7G1AgFC2/yxzG28M2XN+0\nX5xuwKWT/h7czdwlQGEOkDhDnCLtRygw5JaORQseqIx2EOJnLpHpc6BvJKWaaCiyN4nFaUshd00D\nPUM6gb1SKVJJT6Tir3Uc52P+UK9pC0+k6s9DhtQ1UjOvRNtSrxC/rIHgfKO0Q2/UdsKY5DvIK3bR\nkoc8Oj4dIoSd2XFjv0EAAttdvBHTQi16s99NF6Yx7cmXmpJJ7SYFY5DzcPcWcmo0bd7upYRz9c34\n6PgFv39ffOUblsSX8BPDrDfhzrxpkvsHJmYDCJzsxhupibxwgDNd5LC0IbLOaGJmFuH5AmXWUsqE\nAWpTWKrVLEOjaYPz38/b34em3OTzvI1Q6R8ghrT3OOdzLjXjc7b3vX5SZSUB53K22HvOp0tVkxFR\nKXlJObL7R6R7rq3d4cDUgRMBqHtv5VBTJlSK6a6UsUMF19Y7uUpwqVWUwQB3Cc9puZN89nW5hLZR\nZY3xcBEdvYgQdmaHzZqOTRNX42p7I9Z+utmvTXbCADx220M+cdVigu+X45/we6SyE7KQaExwF/xQ\nW4wegFcyIPnH9RfFD6KrRUE1Ipnj6Dgd4P4SF283KmM4vmg44bNt6aj7sfT9J/3avly6Fo9+8Izz\nF8YBHMO4YTbcfZd4kYtO+zi8eNwZY/184Qr33z5f+BQSTPHY+vnvcLbpHPScHrPypiE7vr/bJurN\n8jGPoNvRjYY23wIhcs5cjDHcknoz/nb1lGcbGKblTkJh1lgkmhJQdcrfbsiHQbXZ213nkPGm7uEy\n65bStbCzblj08kuOP/nRP2LHV78P+jxqNG15D2TffZsmrobFYEZ7dwcsBjO+aTzj3qfmo5nHv8qX\nXvTDSpgpTm1ZU/6Yt6TejPuHL4bFYManF4+Ktt1Uulp0+4MjlqDT0SVb9jIQUh9IKV4FMB4e+S+w\nGuPw0pHfOPvjuuY8DmbHnJvuwp150xSHnsrx8qQXQiJI/yF/JmbkTvHrkw46wWsTnuVxgOHmlHy/\n6wUA3Q5n/LYwk12kiG2h7dKOvG03Qroddr+bKNbWqDNC+EDpOA5mvQnXXTkG1IZ8qSHOYEYX1CVj\nEeJcXnd2Vup6GEVSq0q2FbGBJVqNsKVIeaJ6tnsv+VkMZpj1Jvdky59PymbM23iFGo/8SgcTnVA5\njnMXchCb7Pk80mK+EaHWtE16I4DAWo3/5KJuolNn05a+psKz8u8R//+hTtep43SiZgwOvh/Mqm3a\nXppeIEEntYTKueaCcGPUGXyiJIT95R0+QyGw+fOFCrE++WjaXPgd0cT60OWSE7Q8HgV427OlHgfR\nyVjUbqbz02I55/qgT5tgCRTK46yGExi5x17vY2NUzo73xLOMiW036pVdA51IX5g7T7myY4jZx6Rw\ngPkJIOHkLjbZWwWONvLTSvAhX2rw15TVCalQeY8HItj3QeqcOoiHAjqTzvTcES3aqj2JEcheH9Y6\n7GGAk7inoTw+IP+G8Hn3DSEwA4SCmBbaY2yeOGhpTTuwBgXwX/L+3uPeW8Lpfah0OpFbHp8xaIr7\nZ36i4m26vM1pbH//2PGPjl8QPR6/nXWZ0FWXDwAYaRsm278B8ZkiW3kPc1epvzA4ZRUOGKsgxtnz\navNx7bwNbvINxYrPJU7oBEJu0g0AgAkD5OsND0kZLLpdaLOVagcEusbKHdGU9CMQkpp2iJbHxT4W\nerPWsrJEIJ7a3gXpnndtpiuHwbA0ddc00uiE5qYIfDiN638bAGDqwMgmA+OJDn0/AtyVNwO3eD3A\nPdW09ZzObzLQCVwnxCapqTkTMTAhG//+1/8UPT+D0q98hYJMpt3UnIl46/R/A/B8kf/j0HmYd9Ns\nWAxmUZsTAIwanI6vBdseGfYIUm9NQ7djNDhOBz2nh8UCJMXJJ2l4+vZ/87vm/ND5a8Ev/wWaeoXJ\nVDrsntzwfM7vbkc3HIzBYjDjmMBWL9Ra+N9MehNWjv8FALh9IpQsfcp6j3v93FObdKolJWCfirPG\nYeGQe0X3ZSc4czXzKw9yjkayIpsLtLIjbyN8eOS/yJ7TP3mRTvIae78/wWvaIvt6MfGKko9QBgar\n0YrNpWt87ttdN85AWc6kkC2L9xYcx3l8QbjQO+67r6nMM1GQPkzU1h0pYlZomwX2VintrYt5klbI\nlZF0OnEJlp446aeMD68y6YyKXvxAX9mKP0AVNuTH6oy19LVFCkkTqQqVmZyIFIv6LEqi+Yjdmjbf\nJ4XL44KxehfT4MfifS7hR5X/IrMrBavO6NPW+7rIXl2Z2+zd12BqewsJNMEY9UZZTVdpbunA2p+6\nMDufv5R6ViW2S9u0fdurzSXteT+1szwu9sEWLUJHDcIw1HCtbASagaPp2sWs0PavmCqvaRt0BnTa\n+XJ8IjWLOU70zkvnSXYmwncwpuzLP8CzGqhYSaD+COEntu8vNePQZ9/B4T3PCWSK1WIABAsSobX/\nBbk8Lhhre7e8o15Ab28mvUyqBKVambE3bGchUhADLY/L7Q7epi19PCnvcW+C1bRDkSI27ERXmuwe\no/Py1QmnKaK3U9X2hJi1aQtvktTkw9uC7hxU5t5WnDVOtK2ayYGfAJjrf4EIRtPOtPpXUQsk8EZm\nDAcApFucdrFDn32HD49dwEdfev55Y7+aieyMeLe3Np9TPF5FlqZAeIrYCx3R5K+bVNz8LIkY70AZ\n7JhEO29KbyiS3KfUe7x/fPir34VqkgpUdIL3rs/v5x/SJZcLPhhG20bK53eH73OklMkDnf4K4waM\nCb5zYSDOEOce09hMp901K6F/JLsUcryrfHFc6JOr8DHs3vN7tBOzmrYQscnnucIVbqeOabmT3MUt\nLAaz2+Z4//DFGJ42FIBzkt9SusZdH1Y+A5ZTM2dM2fQZeBHSt8WW0rUw6PR45P0Vsu2E/OvwxT7x\npM1tThX62SVjYLU4NcDnju4HAPxixHIkmRKQmmTB6CHPgYFBx+nQ7egOaSgI32P38niQqwo5STfI\n2qaEQkTYzj15yJyzMGssdp98Q2KvMkGZYIzHC0W/xNMfi8f8qkVMjoVMaMtMohw4GHUGbCldIxou\n01PHTO8xbC5dA7PehJoLR2T7yhhT7Yg2Iet2zPhRMa43iteljhTri591j+Wfhi3EfUPn9EpYWW8S\nbo/9vH457mdHK5DQdiH2cAgnbbHJ3ijIoextI3VO8hLL4/ySLxyy2oES7c6/NR/TK35mOYTxpC2u\nIvEDbQl+cbVZKf3c5/GOq1aXc1sJ/LVStzwuhpxtSniN/e264bNtCsck9LkINb1ZblDqeQg+Ttv/\n+vPPrOzqlssvutgfWAAAFX9JREFUJZixWwxmXEd0CW1nHXAnvRUH3tv0hk1ba9ctZoW28MUV096U\n2LDkAu4ZmOQRPEu+DEyRY0zPwmc87TzHOfDpOXzf0Czbvu5yC+LMeokJtncmfqGmrThOW6VwF7aP\n0/sa79V9QPkjLysEppWgziBxZJHuhjP21XPiQLtDa9MGIPsueWJytWO/jHX41RHAed+1ECsfbmLW\npi1ETHtT8oCIha3w8ZGplmRMyRGP7ZvpsqGMzBgeILexcieYAfFOe9Zo20jJNvyYWtu7sOfPp/Hx\nlxdl/zW3dSE7PcHnGINcuY6DSssaDH427Z6Ht4khXB4fkeEbU97zyV5ZyBfgCbPKS8rt4TkllsfD\nqGnzMeu8f4MUUvdRz+mDXq3xHlecweKT69v7Q7mn8HmqhTXcidDCgcOsvOkAgIIM+RwPsULsatpC\nR7QgEyeIVdh5oOCf0OXogklvQlHWONyaMcJvCWZa7iRMvGECzHoTLrRcdG9/qeQ5dz5r384EXtZO\nNCX4xWcK4fc1u5a9xwy1Yf6kfNljpyT6Lik/PvohdDvsQVUtCwaPps07pCj91gxO07Ya4rCm6Bm/\ne8Z67D2uHL1O7/ZLCA/hE9rzbp6Nu/PvCLjsKLVisrl0tfw1Vlj4Zc2EZ3zOwX/4qg35EiM/eZDm\nbKFaRMdxmJk3FVNySuhau4hZoS0S8+WHEuFgENE2OY7z0RSk4m7F7HBSbZVGwwZ6sPlUfC3tTgez\n1EQzMpLVxQU7Y6l7b5FGuKzpTq4SQGNSu5LGa34OJp6HnPXQpq5WU5f2S+g54bZpK5lgpa5joBzP\nssvjXuPyrxmg7LlRCgmR8MO/+3StPcSs0ParEBSkph2K1KTK7IvyffE+wpdnruDTv9Y7fxHkN/mP\n/zkFDhx+aHE61cRbtPMIeCd8UUKwy+NSmphnrg/SrhZFptResWkHIFibthxy43Jr2r3ohEf0ELJh\n+xFzNu25w+4EAAxzhWl5CM6mHQqhLes9zmfwU/Hwvl19Bh+fuIiPT1z02/e/J+rx8YmLOHH2KgAg\nOyPBr020wY+dz41+m22Eor/jzQyBbKs8/HK/WG1sALi9/60AgLKcibLHsegtPs/FtJxJAIAhqdI5\nvMNJZnyG37be9B6Xgl9JCFQPWg5hfeqhrjzpZTmlfm3vcD8/BUGfj+gdPH4z4vOreI2C2EA7alaI\nWFhwN4ozJvinyhSRiUqcv0LhzagsuYpymtu60C/ehGcWj4aDjUOHvQPrjq8DAGx4sNDdzmjUo1+8\ndpadynJKMTG7ULGTkkFnwJbStYpjxuNcdartTFxoD0sbgi2lawKef0PJSp97es/gO3FnXlkYQuGU\nkWCMx5bSNXj8w2fdqxXR4EHtzGug3m7vvYKyqvBJn8pV/eMzJe/RlJyJKM4eH7H7QCiH95sRU4pe\nLl3ba/400UjMCW0gtHHEoYgblNd6lDk/eR+jpa0bKUlmpLtt1Z7sZOkq7dfRgPfY1d47NXZhJfm2\nlZxfbEKJtKAQnj8KFG0AwdrtPc+DsySu78Qud60jfR8IZcj5zRjDnMMg2olJoa0UZcvj4dG0m9u6\n8E71WXTqHQAHnDnfJHu3/vPg1zC2JwMA2jq6cYM5dGlEI01vxWaKeu33Er0tQ5XlBiAIItqIOZu2\nFCad8wv8xn6D3NvktNsfuWzi/cz9enzuYalDAAAzcj31rD//pgF/Ovo97K5KHQ3XPMUuHK2Jfsf4\ny9/q3XZsBiDb5mur1nE6JBqj336thoGJ2QCA8f1DkxN6cLIzPzafrravEQ127FAwPP0WAJ5cB30d\nvnY7QQCkabvxjovl83XLaXg/G/HP6HJ0hyQsJyuhv58djo+jNup16IYdt96Uji8bncU6ni96HKs+\nX+VzjMp/vh2J3c486eCAVEG5zM2la7RRpUgF/cxJ2FK6JmRVsVIsySE9XjSjZQE+MDFLkW9BX2Hl\n+F/42O2J2IaEthdqBLAzFjt0k7twAmp1xVHrdBzgACwmz63KSPGvU52aZEE/h7S9OpQFPKKJUE/c\nkRMEvStEoyHkqyfEisAGxO32ROzSN2dyjfBt3Q84/EWdqFPQ2QtNqo4VDd7AhHbQsqZNELEMCe0I\nsr/2HD471SC5P85swLSBU/DH/9uPUbYC/KX+c/e+9Lg0XG670hvdJHqF8Jsu7rpxOt49cwAAMH7A\n6JAd16QzkiZIEL0ECe0Iwtut1z1YCJ3InB1vMSLObEDZoGJ02H3LAvJ2rkcPP90bXSXCTvg13zsG\nTcWUgc4CNqFcXt448fmQHYsgCHlIaEeQlvZuxJn1sAWInTbpTWi3d/hsIzsXEQzhsAXHcqILguht\nSGiHmMs/tOG///f/0NUd2Nuz4Yc2JFhC48xGNkqCIIi+DwntEFNz4iI+PHZecftbclICN4pRMq0Z\nqG+Vtvn3JZLNzsQ4Ba4YZIIgCDHCKrQ3bNiAzz77DN3d3XjggQdQUFCA5cuXw263IyMjAy+99BJM\npr4VutHc5gzVWjZvBG5QUIxDWKua8FA57nGkpllx7Wp74MYax2Iwq8qTThBEbBK2GeKTTz7BN998\ngz179qCxsRH33nsvCgsLUV5ejpkzZ2Lz5s2oqqpCeXl5uLoQEVrbnc5lA9LjkdYvcmkx+wI6TufK\nM9z3hTYQ3vrZBEH0DcImtMeOHYsRI5wlFJOSktDW1oba2lo899xzAIDJkydjx44dUS+0m9u68F8f\nnkF7Z7ei9l9/dw2AtupUEwRBENogbJJFr9fDanVm7qqqqsLEiRPx0UcfuZfD09LS0NAgb69MSbHC\nYAi9Z2pGhn/ubql2x/9yDu9/Xqfq+LaUOORkpzizmYWIFLvTw7wgc4hP/4tyxuDjc0eQlZiJhD5U\nJMQbpfdLa9C4tAWNS1v01XGFXR08dOgQqqqqsGPHDkyfPt29XYm3c2Nja8j7k5GRiIaG64raNjRc\nx4VLzrZLZg7F8LxURX+XEGfElSvNQfdRipdL18KgM/j0/778+Zifdy8SzPGKx6Ul1NwvLUHj0hY0\nLm2h9XHJfXCEVWhXV1dj27ZtePXVV5GYmAir1Yr29nZYLBbU19fDZrOF8/QhocWVAzwzJc6vCEdv\nI1ZHluO4mK8vSxAEESuELTvH9evXsWHDBrzyyitITnaGs0yYMAEHDjjTKB48eBAlJSXhOn1I2P7u\nV/js60sAAGuI4qkJgiAIIljCpmm/9957aGxsxKOPPuretm7dOlRWVmLPnj3IysrCPffcE67T9wg9\nM6HbYUfNV/UAnDnA0yKsZRMEQRBE2IT2woULsXDhQr/tO3fuDNcpQ8bga/PwxenLeP4nt8NqMcBq\nMfiUxiQIgiCISECSSITWdgfAdMhKjw+pBzhBEARB9AQS2l502x2oOvwt6hqaEWc2kMAmCIIgogoS\n2l6cOd+Eg3/5DgAwZGByhHtDEARBEL6Q0PaixVXf+u4Jg3B30aDIdoYgCIIgBFBBZi/4mOz0fhYY\n9HRpCIIgiOiCNG0ATS2d+K/qM/j7RWcGHYrJJgiCIKIREtoAjn7TgA++cNbA1us4ZKVbI9wjgiAI\ngvCHhDY8tux/vWsYRg5Oh5UqdBEEQRBRCBlu4bFl21LiSGATBEEQUUvMC227w4H9tecAAPFxZMsm\nCIIgopeYF9rn6j0lNFMSzBHsCUEQBEHIE/NCu9llz75rwiCYTfoI94YgCIIgpIl5od3S7hTaKQmm\nCPeEIAiCIOSJaaFdd7kFv9v3VwBkzyYIgiCin5gW2h+6YrMBICs9PoI9IQiCIIjAxLTQ5u3Zz/3k\ndtyQkRDh3hAEQRCEPDEttFtd9uy0JEuEe0IQBEEQgYnJTCJXm9rxdvVZnLnQBB3HIc5MXuMEQRBE\n9BOTQvvTv13CR19eAADk9k8Ex3ER7hFBEARBBCYmhTYf5vXInAIU5KdFuDcEQRAEoYyYtGl75xqn\nutkEQRCEVohJiVVz4iIAis0mCIIgtEXMCW2Hg6Gjyw4ASCChTRAEQWiImBParR3OpfHheam0NE4Q\nBEFoipiTWs2tnQCAfvGUa5wgCILQFjEntF/bdwIA2bMJgiAI7RFTQru5rQufuJzQBtoobSlBEASh\nLWJKaLe4co3fdnMGigoGRLg3BEEQBKGO2BLafHx2clyEe0IQBEEQ6okpoX3yXCMAwGqJyURwBEEQ\nhMaJKaF94UoLACAl0RzhnhAEQRCEesIqtE+dOoWysjLs3r0bAHDhwgUsXrwY5eXlWLZsGTo7O8N5\nej9aXcvjIwen9+p5CYIgCCIUhE1ot7a2YvXq1SgsLHRv+/Wvf43y8nK8/vrryM3NRVVVVbhOL0pL\nezc4DrCaaXmcIAiC0B5hE9omkwnbt2+HzWZzb6utrcXUqVMBAJMnT0ZNTU24Tu8HYwynvrsGq8UI\nnY5KcRIEQRDaI2wqp8FggMHge/i2tjaYTM5MZGlpaWhoaJA9RkqKFQaDPiT9ae/oRrzFgPzsfsjI\nSAzJMaMNGpe2oHFpCxqXtuir44rYOjFjLGCbxsbWkJ5z89IiZGb2Q0PD9ZAeNxrIyEikcWkIGpe2\noHFpC62PS+6Do1e9x61WK9rb2wEA9fX1PkvnvYHRoIeelsYJgiAIjdKrQnvChAk4cOAAAODgwYMo\nKSnpzdMTBEEQhKYJ2/L4iRMnsH79etTV1cFgMODAgQPYuHEjVqxYgT179iArKwv33HNPuE5PEARB\nEH2OsAnt4cOHY9euXX7bd+7cGa5TEgRBEESfJqYyohEEQRCEliGhTRAEQRAagYQ2QRAEQWgEEtoE\nQRAEoRFIaBMEQRCERiChTRAEQRAagYQ2QRAEQWgEEtoEQRAEoRE4pqRyB0EQBEEQEYc0bYIgCILQ\nCCS0CYIgCEIjkNAmCIIgCI1AQpsgCIIgNAIJbYIgCILQCCS0CYIgCEIjxIzQfuGFF7Bw4UIsWrQI\nx48fj3R3gmLDhg1YuHAh5s6di4MHD+LChQtYvHgxysvLsWzZMnR2dgIA9u3bh7lz52L+/Pl44403\nItxrZbS3t6OsrAxvvfVWnxnXvn37MHv2bMyZMweHDx/uE+NqaWnB0qVLsXjxYixatAjV1dU4efIk\nFi1ahEWLFmHlypXutq+++irmzZuH+fPn44MPPohgr6U5deoUysrKsHv3bgBQdY+6urrw+OOP4777\n7kNFRQW+++67iI1DiNi4lixZgoqKCixZsgQNDQ0AtD8unurqagwZMsT9u9bGpQoWA9TW1rKf/vSn\njDHGTp8+zRYsWBDhHqmnpqaG3X///Ywxxq5evcpKS0vZihUr2HvvvccYY2zTpk3s97//PWtpaWHT\np09nTU1NrK2tjc2aNYs1NjZGsuuK2Lx5M5szZw578803+8S4rl69yqZPn86uX7/O6uvrWWVlZZ8Y\n165du9jGjRsZY4xdvHiRzZgxg1VUVLBjx44xxhh77LHH2OHDh9m5c+fYvffeyzo6OtiVK1fYjBkz\nWHd3dyS77kdLSwurqKhglZWVbNeuXYwxpuoevfXWW2zVqlWMMcaqq6vZsmXLIjYWb8TGtXz5cvbH\nP/6RMcbY7t272fr16/vEuBhjrL29nVVUVLCioiJ3Oy2NSy0xoWnX1NSgrKwMAJCfn48ffvgBzc3N\nEe6VOsaOHYtf/epXAICkpCS0tbWhtrYWU6dOBQBMnjwZNTU1OHbsGAoKCpCYmAiLxYLbbrsNR48e\njWTXA/Ltt9/i9OnTmDRpEgD0iXHV1NSgsLAQCQkJsNlsWL16dZ8YV0pKCq5duwYAaGpqQnJyMurq\n6jBixAgAnnHV1taipKQEJpMJqampyM7OxunTpyPZdT9MJhO2b98Om83m3qbmHtXU1GDatGkAgAkT\nJkTNfRMb18qVKzFjxgwAnnvYF8YFANu2bUN5eTlMJhMAaG5caokJoX358mWkpKS4f09NTXUvD2kF\nvV4Pq9UKAKiqqsLEiRPR1tbmflDT0tLQ0NCAy5cvIzU11f13Whjr+vXrsWLFCvfvfWFc33//Pdrb\n2/Hggw+ivLwcNTU1fWJcs2bNwvnz5zFt2jRUVFRg+fLlSEpKcu/X0rgMBgMsFovPNjX3yHu7TqcD\nx3Hu5fRIIjYuq9UKvV4Pu92O119/HXfffXefGNfZs2dx8uRJzJw5071Na+NSiyHSHYgETMOZWw8d\nOoSqqirs2LED06dPd2+XGlO0j/Xtt9/GqFGjMHDgQNH9Wh0XAFy7dg2/+c1vcP78efz4xz/26bNW\nx/XOO+8gKysLr732Gk6ePImHH34YiYmJ7v1aHZcYascS7WO02+1Yvnw5xo8fj8LCQrz77rs++7U4\nrhdffBGVlZWybbQ4LjliQtO22Wy4fPmy+/dLly4hIyMjgj0Kjurqamzbtg3bt29HYmIirFYr2tvb\nAQD19fWw2WyiYxUuJ0UThw8fxp/+9CcsWLAAb7zxBn7729/2iXGlpaXh1ltvhcFgQE5ODuLj4xEf\nH6/5cR09ehTFxcUAgKFDh6KjowONjY3u/VLj4rdHO2qePZvN5l496OrqAmPMraVHI0899RRyc3Ox\ndOlSAOLzopbGVV9fjzNnzuCJJ57AggULcOnSJVRUVGh+XIGICaFdVFSEAwcOAAC++uor2Gw2JCQk\nRLhX6rh+/To2bNiAV155BcnJyQCcdhl+XAcPHkRJSQlGjhyJL7/8Ek1NTWhpacHRo0cxZsyYSHZd\nlpdffhlvvvkm9u7di/nz5+Ohhx7qE+MqLi7GJ598AofDgcbGRrS2tvaJceXm5uLYsWMAgLq6OsTH\nxyM/Px9HjhwB4BnX+PHjcfjwYXR2dqK+vh6XLl3C4MGDI9l1Rai5R0VFRdi/fz8A4P3338e4ceMi\n2XVZ9u3bB6PRiJ///OfubVofV2ZmJg4dOoS9e/di7969sNls2L17t+bHFYiYqfK1ceNGHDlyBBzH\nYeXKlRg6dGiku6SKPXv2YOvWrcjLy3NvW7duHSorK9HR0YGsrCy8+OKLMBqN2L9/P1577TVwHIeK\nigrMnj07gj1XztatW5GdnY3i4mI8+eSTmh/XH/7wB1RVVQEAfvazn6GgoEDz42ppacHTTz+NK1eu\noLu7G8uWLUNGRgaeffZZOBwOjBw5Ek899RQAYNeuXXj33XfBcRweffRRFBYWRrj3vpw4cQLr169H\nXV0dDAYDMjMzsXHjRqxYsULRPbLb7aisrMTf//53mEwmrFu3DgMGDIj0sETHdeXKFZjNZreykp+f\nj1WrVml+XFu3bnUrMVOmTMGf//xnANDUuNQSM0KbIAiCILROTCyPEwRBEERfgIQ2QRAEQWgEEtoE\nQRAEoRFIaBMEQRCERiChTRAEQRAagYQ2QRAEQWgEEtoEQRAEoRFIaBMEQRCERvh/+L2VAFXJbDsA\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff4637676d8>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "actual_pred_pair_sorted = sorted(actual_pred_pair, key=lambda x: x[0])\n",
    "plt.plot([x[0] for x in actual_pred_pair_sorted], label=\"Actual age\")\n",
    "plt.ylabel(\"Age\")\n",
    "plt.plot([x[1] for x in actual_pred_pair_sorted],label=\"Predicted age\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "age_estimation_regression.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}