Part II - MLP with Keras.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "distinguished-foster",
   "metadata": {},
   "source": [
    "# Assignment - Part II"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "engaged-transport",
   "metadata": {},
   "source": [
    "### Imports and global settings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "continental-beginning",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.utils import plot_model\n",
    "from keras.callbacks import ModelCheckpoint\n",
    "import matplotlib.pyplot as plt\n",
    "import time\n",
    "plt.style.use('seaborn-whitegrid')\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "increased-friendship",
   "metadata": {},
   "source": [
    "## Approximation of Mackey Glass time series function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "graphic-discussion",
   "metadata": {},
   "outputs": [],
   "source": [
    "def data_split(training_size, inputs, targets):\n",
    "    inputs_training = inputs[0:training_size, :]\n",
    "    targets_training = targets[0:training_size, :]\n",
    "    inputs_testing = inputs[training_size:1000, :]\n",
    "    targets_testing = targets[training_size:1000,:]\n",
    "    inputs_evaluation = inputs[1000:1200, :]\n",
    "    targets_evaluation = targets[1000:1200, :]\n",
    "    print(\"\\n Training samples: \" + str(inputs_training.shape[0] ), \n",
    "            \"\\n Testing samples: \" + str(inputs_testing.shape[0]), \n",
    "            \"\\n Evaluation samples: \" + str(inputs_evaluation.shape[0]))\n",
    "\n",
    "    return inputs_training, targets_training, \\\n",
    "        inputs_testing, targets_testing, \\\n",
    "        inputs_evaluation, targets_evaluation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "brutal-rachel",
   "metadata": {},
   "outputs": [],
   "source": [
    "def Euler_Approximation(t, beta, gamma, tau, n):\n",
    "    x = np.zeros(t)\n",
    "    x[0] = 1.5\n",
    "    for time in range(t-1):\n",
    "        if time-tau < 0:\n",
    "            x[time+1] = 0\n",
    "        else:\n",
    "            x[time+1] = x[time]+((beta*x[time-tau])/(1+(x[time-tau]**n)))-gamma*x[time]\n",
    "    return x"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "figured-tulsa",
   "metadata": {},
   "source": [
    "## Data generation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "stable-andorra",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAESCAYAAADZt+8IAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAABvNklEQVR4nO2dd2AU1fbHv7O76SEkIaGXBGkriCKIiBQFpOqz8mxg+fl89oYPBcHybA99Tx+K3WcFGyACilQpgigoIlKWGkIJhPSQtim78/tjd2an3OmzyW64H/4gO+XeO7ecOXPuuecyLMuyoFAoFErE42jqAlAoFApFH1RgUygUSpRABTaFQqFECVRgUygUSpRABTaFQqFECVRgUygUSpRABXYzpmfPnnjwwQdlx2fMmIGePXuaTnfOnDmYMWOGlaKp8tNPP+Hmm2/GpZdeissuuwzXXXcdvv/+e/78tGnT8NZbb4Ut/7feegtjx47F2LFjce6552LIkCH8719//RV33HFH2PKeO3currjiCowbNw4jR47E448/jsrKSsPpjB07FkVFRWEoIaUpcTV1ASjhZd++faisrERycjIAoL6+Hjt37mziUimzYcMGzJgxA7Nnz8aAAQMAAL/++ivuu+8+pKSkYMiQIWEvw7333ot7770XADB58mRcd911uPLKK/nzF1xwQVjy/fHHH/HZZ5/hiy++QFpaGmpra/Hoo4/i5ZdfxrPPPqsrDZ/PB6fTiRUrVoSljJSmhWrYzZxBgwZhzZo1/O9Nmzahb9++omsWLVqEcePGYfTo0bj55puRl5cHAPD7/Xj11Vf5c++//74s/fz8fIwYMQK///47WJbFm2++iXHjxmHs2LF4/vnnUV9fj88++wx33XUXf4/f78fgwYPh8Xhk6b322mt4+OGHeWENBATk2rVricL6zz//xLXXXouxY8diwoQJ2Lx5MwCgoaEBM2bMwJgxYzBq1Cjcf//9qKysVDyuly1btuCyyy4DAMyePRvPPPMM7r77bgwZMgRTp07F+vXrcfXVV2PIkCF8vdfV1eH555/H2LFjMX78eLz55psgrVc7ePAgOnfujLS0NABAXFwcZs2ahccffxwAUFFRgX/84x8YM2YMLr/8csyfP5+/t2fPnnjzzTcxZswY+Hw+9OzZE/n5+QCA+fPnY/z48Rg7diweffRR/nm3bt2Kq6++GuPGjcO4ceOwfPly3fVAaRqowG7mjB07Ft999x3/e9myZRgzZgz/u6SkBE899RQ++OADrFq1Cp07d+bNDStWrMDWrVuxdOlSLFq0CJ999hl27NjB3+v1enHvvffikUcewfnnn49Vq1Zh5cqVWLBgAb777jscO3YM8+fPx9ixY/HLL7+gtLQUALBt2zakpKTA7XaLylpVVYU9e/Zg2LBhsufgvhCkPPnkk7j11luxYsUK3HnnnXj66acBBF5Mx48fx4oVK7B69Wp0794d27dvVzxuBqfTiXXr1uHFF1/Et99+ixUrVmDdunX45ptvcP/99+O9994DAMybNw+HDx/Gt99+i4ULF2LNmjXYsGGDLL2LL74YmzdvxuOPP44NGzbwX0ZJSUkAAi+I2NhYrFixAp9++inefPNN7N+/X1SeNWvWwOl08sd27tyJOXPm4JNPPsGKFSuQlJSEN998EwDw0ksvYfr06Vi+fDneffdd0YudEplQgd3MGThwIA4cOICSkhJ4vV5s374dF110EX8+PT0d27ZtQ/v27QEAAwYMwLFjxwAA69atw9ixYxETE4Pk5GR8//33Iu38iSeewIgRI3DFFVcAAH744QdcccUVSE5OhsvlwsSJE7F69Wq0atUKAwYMwMqVKwEAq1atwvjx42VlraysBMuySE9P54/dfffdGDt2LEaMGIF77rlHds/ChQsxYcIEAED//v35smdkZODQoUNYvXo1ampq8NBDD2Ho0KGKx83Sr18/pKenIy0tDZmZmRg+fDgAoHv37igsLOTr5dprr0VMTAwSExNx5ZVXYvXq1bK0evbsiYULF8Lv92PatGkYNGgQ7rvvPpw4cYJP54YbbgDDMEhPT8fo0aNFQvaSSy6Rpbl27VqMHDkSmZmZAIAbb7yRvyczMxOLFy/GoUOH0LlzZ7zyyium64HSOFAbdjPH6XTisssuw/Lly9GqVStcfPHFcLlCzd7Q0ID//ve/2LRpExoaGlBZWYns7GwAAe07JSWFvzYxMZH/e9WqVairqxMJ/+LiYmzevBlff/01gIA9tVWrVgCACRMmYNGiRbjhhhvwww8/4L333sOpU6dw6623AgD69u2L559/Hg6HAwUFBfwL5J133gEALFmyBAsXLpQ935IlS/DZZ5+huroafr+fNzX06dMH//znP/HJJ5/g8ccfx8iRI/HUU08pHhc+pxE47Zera+630+mEz+fj6+Xll1/G66+/DiBgIjn33HOJ6fXq1Qv//ve/wbIsdu/ejddeew2PPPIIvvrqKxQXF2PKlCl8+9XW1mLcuHH8vampqbL0iouLsXLlSvzyyy8AAJZlUVdXBwCYNWsW3njjDdx+++2Ij4/HP/7xD4wePdpUPVAaByqwzwDGjx+P1157Denp6bj++utF57777jts2LABn332GdLT07FgwQIsXboUQED75swYAFBUVIT4+HgAwNlnn41p06bh9ttvx+DBg9GhQwdew7zllltkZbjsssvwz3/+E+vWrUNSUhK6desGALLJsf79+2PVqlW47bbbNJ/r+PHjeOaZZzB//nycffbZOHLkiEjgjBw5EiNHjsTp06cxc+ZMfPjhh3j44YcVj4eLzMxM3HbbbRg5cqTqddu2bUOHDh3Qtm1bMAyDPn36YOrUqXybZWZm4o033kCvXr0M5X3VVVdh+vTpsnOpqamYOXMmZs6ciZ9//hn33XcfhgwZInoxUyILahI5A+jXrx8KCgqwf/9+DBw4UHSuvLwcHTt2RHp6OoqLi7F06VJUVVUBAC699FIsW7YMtbW1qK6uxk033YR9+/YBADp27Ai3241bb70VTzzxBFiWxciRI7FkyRL+/i+//BKLFy8GALRo0QLDhg3Ds88+K9IKpUyZMgXvvvsu1q9fzx/7888/8fbbb6NLly6iaysrK5GYmIizzjoL9fX1mDdvHoCALXzhwoW8rTYlJQVdunQBy7KKx8PJyJEjsWDBAtTX14NlWbz11lvYuHGj7LolS5bgySefREVFBYDAF8qyZct4r5QRI0bgiy++AMuyaGhowIsvvog9e/ao5j1ixAisXr0axcXFAIA1a9bg/fffR319PSZNmoSCggIAAXOMy+UCwzB2PjrFZqiGfQbAMAxGjhyJmpoaOBzid/Tll1+O7777DqNGjUKnTp0wdepU3HPPPXjllVfwyCOPYO/evfjLX/4Cv9+PiRMnon///rwnBgD8/e9/xw8//IB58+Zh0qRJOHDgAK699lr4/X506dIFL774In/thAkTFO3XHOeffz7eeOMNvPHGG3j22WfhcDjQsmVL3HLLLbKvg549e2LYsGEYM2YM0tLSMGPGDOzcuRN33HEH3nnnHUybNg2jR4+G0+lEdnY2/vWvf4FlWeLxcHLTTTfh+PHjuOKKK+Dz+dC3b1/eFCRkxowZePXVV3HttdfC6XSCYRicd955mDVrFgDg4Ycfxj//+U+MHTsWfr8fw4cPR48ePVTz7t27N+655x7ccsstaGhoQHp6Ol544QXExMTguuuuw2233QaWZeFyufDkk08iISEhLHVAsQeGxsOmNBZ//vknnn32WaItmkKhaENNIpRGwefz4a233sKkSZOauigUStRCBTYl7OzZswcjR45EWlqaaMUghUIxBjWJUCgUSpRANWwKhUKJEqjAplAolCghbG5927ZtC1fSFAqF0qzp378/8XhY/bCVMo0UPB6PLABRNNJcngOgzxKpNJdniYbnUFN2qUmEQqFQogQqsCkUCiVKoAKbQqFQogQqsCkUCiVKoAKbQqFQogQqsCkUCiVKoAKbQqFQogQqsCkRy/Xv/ozCitqmLgaFEjFQgU2JWLYcLsGek6ebuhgUSsRABTYloqEbVlEoIajApkQ0dItBCiUEFdiUiIahOjaFwkMFNuWMwnPyNBp8/qYuBoViCiqwKRGN3SaRca9txKLf8+xNtJFo8Plx2lvf1MWgNCFUYFMimnAYRLwNvjCkGn5mrzmAvs+saupiUJoQKrApkQ01YfMcK61u6iJQmhgqsCmUKIG+uyiWBLbX68WoUaOwaNEiu8pDoYgIh5cIFXyUaMWSwH777bfRsmVLu8pCocgIhx82a3+SjQJDndLPeEwL7EOHDuHQoUO45JJLbCwOhdL0VNU2NHURKEG89T589evRpi5GxGB6E96XX34ZTz75JL755hvFazwej9nkGwWv1xvxZdRDc3kOQP4sR48cQcvaAlvzOJWfD4+nhnhuT4EXjy4/geW3drWcj93tUl5eDqBpxlVT9bFfj1fjqR/y0Te5yvC94z7JwYIbuyA51skfi/axYkpgL168GAMGDEDHjh1Vr4v03YkjdQflCm89iivrkJWRpOv6SH0OM4ifJQdZWVlwZ6fbmEMO2rZtC7c7i3j2iC8fwAlb6tPudkn9sxZAZZO0dVP1sXxHAYB8k3nnIKNDNrIF4ygaxorarummBPb69etx/PhxrF69Gvn5+YiNjUXbtm0xePBg04WkhJjxzS4s3XECubMmGL63uLIWrZLjwlCqpqGxzbYRbSaO5LJRGgVTAnv27Nn833PmzEGHDh2osLaRCgur2fo/vwYrHh6KXm1TbCxR0xEWGRXRUpkihLaUGOqHHWZ2Hi/HocJKQ/dY9WKo9DafSTMqWymUEKYnHTkeeOABO8rRbLnijU1o3SIOW2eMarQ8qZDTgFV+JdKqiyyoK6MYqmE3AkY1ZhV5cgZCBywHDTVLoQK7ETAqgKm8DhEWBStKtbYoLbYlzsBHVoUK7EbBmAhmLavYzaebN58noVCsQwV2I3Ammjh+ySnGnB8ONHUxDENtps2L5taaVGA3AmegvMab6w7ildX7LacTDgEarYM4WsttBfr+FEMFNiUs2CVoo2m8vrnuIHafKA9b+lR4UajAbgSs26SN0ZwGdqOvdLRw7/ajpThaTDcZsBPOM+bH/YXw+c/Eb1UxVGA3AmeiW18zemfohmWBBipUwsItH27FlsPFTV2MJocK7EagOQhgo9ilGTe277GVcvtZFn6Vxv5m+3FU1zWfVaiNzZk4jqRQgd0INLpJpFFzI2O1DI1dZ3bAAmjwKZf7ka92YMWufNPpn+kLZ6KwS9hO1Avs9348hB3Hypq6GKoYNok0A78Sq5OOTTU4rWnY0LSzWjGZmClbvc9vOr9IQPjMzWFcWCXqBfaL3+/FW+sPNnUx1DkD+5llDZv/X1x59T4/sqYts5R2uCYyWZYlCuQf9xdi/6kKAIC/kW3c3Wcsx8GCikbNkxI+ol5gRwNnyqRjSVUd7yVhVSgq2YIjWWNkWcDnl5fvlg+34pGv/gAA+BqhcU9761FQ4eV/l1SZD9ebNW0Z8srIu/M0BsJuZKTqOJNalA4lRajAbgSi0R5rhns/24Zh/14X/GWPSSSaqo4Fq2kSsaJh630J3vXpNgx84QfD9ymRX+7VvqgRUJvQlRLqP/Z3oKve/AknmuglRgV2I9DYGnZTLa+uEMThtloEJXtlJAtwv1/bRm3Nl1hfpQq1a/132QPLsiiurLUvPR3X/JZbAm+9j3hfOLrLH8fK8OfxsjCkrE2zENiRPIjPJBwCKW1USBRX1uJYSWjRCa8hGUhjwW/HLC+usOKJoUfDVnEisQ37s9Cf4qo9p9D/+TW25SzUqpVKcd07P+PTn3NFx3iTSJjqe1fe6SZZJNUsBHak09gvlKZy/nJYyPj2j3/F0JfX8b+16oz0qTt14Z84Xtp0Kw39KgtnuOI29qRjuKiqbcCeE6dlx0ur6mzNR9TMKlUnndoIVbN2ffv8LC54wdhL5o11BzH5wy2G7rEDUwK7pqYGDz30ECZNmoRrrrkGP/zwg/ZNZzBG3ZGi1n1JqGEbFN7lNeKJMa4OpIKZ/9RVqKI31x1EbYOPfJIrm9orzcrbTuDW98XWozhYLDcNNMako5RwWMhmr9mP8a9vtD9hCWIN24ANG/o17LoGPworjJtxmkIxMiWw165diz59+mDevHmYM2cOXn75ZbvLZYhIF2+NrmEzwK0fbpXZMsONUMM2alqQ1pHw96Lfj2PzoaLgcfXZ//m/HUdOYZWhvO1CaBKZvmgnPtleIrvGqMlmxa6TmL7oTwBWBK/9osVbT/bW0SrjH8fKDIXdFSnYKlUnzdeISa0pXqJmMSWwJ0yYgDvvvBMAkJ+fjzZt2thaqOZGU7j1bdhfiKcW77aekAGEY8aocJFqT36BYJ4yfwdmfLMreDx4vaSSjAjCcH3B6Fk4U1xZh62H5YJciXm/HMUXW48B0Ba7jy3cgYnvbG5SDUbrRf3+jzmGwu4K29nMuNBzj89H/pqLRCxtwjtx4kQUFRXhvffeI573eDxWktdNZUWFqby8Xm+jlNHv9xvKp7o6YIfVe4/0OQ4fzgUArNid32htAADemoCrk8fjwemKCv5vPdTVBUwi3LNU1QU0uNzcXABAbW0tPB4PKmoD5g7P3r1wCVT6OoER83BODlAWp5hX/sl8eDxkW/fxPGN1L6S6uhqnCv3weAJl8flC7e71Br52PvzpMD786TCW39pVV5pVVVV8eUpLS1XLtnr3SZTW+NAhJUZ03ZEjuUisNr8kPjf3CLqmiPNVKsvJk+rtXn76tOp5KUePhb6Wjh07Bg9DftkVFBTA4wnZz70NgTY4lJMDf2ls6DhhzJfVBPrU7j0eOA1MxNTV1TXq+AIsCuwFCxZg9+7dmDJlCpYsWQKHQ6ywu91uS4XTRw6SkluYysvj8TRCGXPAMIyhfBJ+LAPg1X2P+DlykJ2dBSAPQGO1QYCkDaVAYS3cbjda/l4NoEp3/jExJwE0ID4+Hm63O2jTzkWXLlkATiAmNhZutxtl1XUAjqBnz16IdYX6W4U3cD0AZHftCne7FIWcctCmbVt0zG6PFvExsrOnHAUA8jFjXQk+vO0CpCbGypNQIG5tCVLT0oLPnAOHw8H/HRsXByAkUPTWS/Lm0wBq4Ha7kba3AUCF4r0uVx4AH+JiYwHU83lnZ2XB3TlN93OIyUFWVhfEV58S5Zu6byeA07Ky7Ko6BqBQsYwpBvvFcfYUgFMAgA4dO6JDdivEOh2Ij3GKytimdWu43WfxRwJBtnKRnZ0t6gukMR/wMz+Cnr16Icapx+iQAwCIDfZJu9m2bZviOVMmkZ07d+LEiRMAgN69e8Pv9/Nv3DMJz0n5LDmJKPjSknGspBob9hcauocRufVZtWGTPWlDtknpknX9lfzZlqM455lVxHPcM/x+tAyHCit1pxkoU3jDq2qZmbjTTdndtNYAGN6QWuLW1+/Z1Xjwi+2q97y57iBvQrvmrc2aeTQEV6daHae/5ZZgzH9/tJaIBqYE9vbt2/HJJ58AAIqKilBVVYW0NLNv8Ohl3Gsbceq09sReYw8gO6K6PblkF279cKvBfJV+GEdpEpK3bUvOG1mynmfR9e+LrUeJ7mssy4bVbc/uaH0sy+LLrUdtTVOrhEaForA62eAcwdESefsJ3xMfb87FN9sDX5g19eoeQ0AowqKRlZQkNh4owr5T4Y3bYkpg33DDDSgqKsJNN92Eu+++G08//bTMHNK4NJ1OoUejMjyZEQEauZm+a2XhjJb7Hu+mpXB/XYM9MUb0xK6YvmgnLxCEROoGBkpab53Pj2mLdppLU6GFtb4CjE/4yh2x1bT4Bp9xFz27NGyrAl8PpmzYsbGxeOWVV+wuS1SiRxg3uobdRCtnlPI9WFCJrhlJcKhM6EjrSCbAWfL/HEINW3PRDeHYyfIaVNf5LLWVnxWvdFRLa+3eUxjRS7931S85xZruZ1z9S+uuMbuDpsC2qGED6s/z9e/HjWWA0EvWqvdQY2xhRlc6WkRPBzSuYDe9lmamBMLBKtSCRr26Aav2nDKUFu++x5VHomlLtRkjNmwSN72/BSNf2WD4vtWC5zKiYeeVGfORv+G9X7BhH3lO4c11B/GflfsMpRctCJuZ+5P0Mc9p/NV12iYQKSGTiOFbRTSGP3ezENjROKkX6ZjxSVUziUiD88jzC/09+YMtyrZVhQURVrWb2mD5RCYRleu5R73z0994c4xUw1ZDy3vM52fBsqzoxafUJq/9cABvrAtfTHhStkpKhZad3WgriVY68hq2ypeaiW7AfZ1ZNWk0RtiBZiGwKZEBY2FpulAAbDxQhG//DHghScNkKi2cMTTYCJeS7KLCJNftLVC0kwtvVRLY0uI5gzct+SMPRYTodj1mLserkgUmirZbDUGmeJtCXXK8u+EQgIAGf8vCI+REdOYVytOYUBNefd/nvwOwFrOGBG8SsSBvd58oxx6dXmNWaNYC2+dn8bqBZbArd+fj6SW7DOURDu0+Wr8YGIW/9SB9ZqncYxGYWOS0IbUqsmJSUhI4t3/8K1buFi8+kUaE47wYrpizSTMf7mvkoS//wCebcwEALyzbg//7+FcAgXR2S4Ir6X0JSp9eU+tVqK5/Ld8LANhyuASFVfpMDXaH9iUKeEIeVrJtsGGl44TXN+Gng+Hf1b1ZCGylal69J1+mpfx0sAjl1eQdOD766TA++VmfJhHKW99EkLE0mx4zfdfOsRpy3wsJxTGzf8Sdn/7G/xYSnhencqIMQi8VP6/9s2jw+7Ezr1xWJmk/EdbVkj8CXxPf/XkSa/cWyPIh3SMqp8Ueo3W3w8aGNe6HLT+241iZYsxtUvLHCG6AQjgvEbssGs8s3Y3dJ8rtSUxCsxDYStw973fZsZv/twVvrCNr3eEY9Ga6utk3vV2xEPblV2DTwSLD94ls2AYHubTkJHvg4aIq7M2vIN5gVWhxxRVqozJNVRZgSO6KqNeGLVwCTfIrJpZRoTdpNbtZk0goX/1o+mEL/j5UWEmsL5Zl+R1dlExdRvqn1vwJVwajNmylPn6osBJFlfaGmeVo1gJbCaWKNjPkI8l8YVdZnllqLmiUnSYRbsad9xKRrnhUWAFpFj3vF6nA5LL0syzm/3oMBwsqRV4ioh2/JeUjaa1az6Bku5XeprcutHzbeQw0ppH39MhXNhD92VftOYXBs9YGyqZQOOXj8hNayoPPBhu2uAzhc6VsFgJb2kjeeh9Olhvfc82Mhqp1B9dZDhZUIrdIX9hPs/1G6gJnFiVtVWtVp2hgGO6xkklEyfyelgmEVTmnnlMAkvYqTUc67oWrLl8OutWJ/LBVyqFXsCm5StpBeL4otezl4kwDMT/ECLeaU9J6jWjD0hedt94n2pMxNOloT4WwYMO2FqJZCGwpL37vwUX/Wqt4XqlhwuGVw7XbqFc36JqMsoJtHY6QTIW3Hhe+qL5Rhb3xsCUTepLrpQPW6rPzJhEDexuIl8sHfuh361PO6PejpcH0JXZvjTSNe+YE/2/Er0St6tlxrAz/WLCD/610uZGxKo3A96/vPbwGDwhNIvrTBAImuie+ka8UDWjY4ZHYzVJgk9yk9MANkO+CLmVK7Morx6X/Wc/f4633KQ5U4SAK97JlJfOB4XQIt+tZmGKnVhEyiZC1HxbA5kNFyJq2jP+tF+JnM+k6lVRF/tEI1ZneNlYT2IoBi4K3LN95UlxOWd2I68z+5eLKWO0Df+aJJ+uUlSsjGra4UCVBpwPO48jKSseNB+SLmVg2fKuNm6XAVtLWtODG2vsbD6te91tuCQ4HzRssgF5PrsCrq8krzYRvWq1G3HywCMdKqk1rPLbZ4AgdV60OJ3+wBd56n+azVtc1KM6eyzVo9TJ+seUonloSsrVbt2EHCiwqtl6TiMB8o+iHLfmtx5eYlZSHEzz3fPa77DoS64MrI7cfLSOnL/mKsQMjk4567jdiq1a6Xtpu3M/vgy8+v0kNGwj504vKAJbasNXQqmcl++ei3WWY9vWfguP6Wky8+izw/6ECBfu0gZa76X9bMHXhDu0LNRDHXzBhlzd4y8YDRSipqhMtGSY99py1BzHhdbJZSLYQRrI2XVqmV1bvx8GCQPjTr7cdh3YvEJSN5Mer707Rr5C5JvSK061h61z9IQ5ZS0apvW4P+nTPXExeW6C3xohfHwo3a4dXVc9V/lIkX3e8tEb0Jc3la0ZL5trMzEpFUjsGNGxqElFEa0JK6fPp272n8eWvx0L3Bf83ZisM3LVid75I+PPXaqQlpabeb+oD9fejpVi6I2DKIS3ntYpWMg6G0dSwaxX2ASSlL43LoJb/owt2WH9OPV4ijPjvkP91SBDpHfRmfJv1xsPWC6vwMrSCVTkltf0qxeeYs/YgbvtIX/hftZfLqt358ElnuA1A1LCpSUQfXIxi+f6A5OuVJrLM9l+h8Ocw6jhRqyN+r5B6nx+7C7x4cvEufrJGKDTMxEcg3aHH15fUSfcIVusZ6cR+v7gttDQzKzbsAc+v5jfuFZ7R/HITpMf9vVNig1VCz8YmMi8VnSJZd5MLvhCMYtW3Wy8+lTjnwo2AueIYybasug5/n7stpGGbKDNpSzFqEtHBnhOn0e+51QC0NezQQJNcp/Gi5VZIChtDU5AZbDotJ38pWw+X4B/LT8AlkAB6Q3wqYabjMghpjZ6Tp3G6JuCaNf71jaJrlOCyPFhcK/qtF7Xr6xr8aCAM/Dk/HMArq/YpLnLQdA8Mng9o2FrXii8gxi4htFa4Br4oXxs1bK7HKSpJOl78QtTmuuNc+sQXl2fWtGXIr6gXeAQF/jBiw5a2o5I/fbhMIpb2dIwkAnv6BTDT/6pqG1SDt/j9LM59dhVyZ00Q9SrNSRSd7cbt7lJT70NLAyOoZUJgX0KvIKyksOOZ0rAJt2gvwWf4Zx332kbiNarxsIOZfvFnKQCBl4jYlK15v/AejsGzfsDQ7pn47/XnidJ6Y91B1EoCOqn6Tgv+/mZ7Hj9YWbCGvwDMmET0tIGZ9OyU10JXRyfJt91gbs99t0fxnHhfRy59QpkER09U1IMJ6qmcOcOIH7b0Ek7DZhhxX6UmEQFjZ/+IFbvEgXhIE4EcSkJL2JA7jpWp5in8/DWiYSuVUQq3f+Kp07XYcVzfZ7U47dDfJBv26j2nUFUrX6RAgvhIekwiGumqdeLQpF3gf6m3hZnFMBxFlXWi/Te5tFwaE3/y+B+h67cfLQtN5rHGhZ6e8WxWkOrtk1Y0a0XvjeD/imNOZuZR/62GUMNW7VsKZX17QyAkrRE/bOlzcUqIWCZEqEnk1VdfxfXXX49rrrkGy5cvt6tMmuzNr8CPAv9H+RtNpw1bcFxN++v91Ar8mlsiu0cP4f6k5coj1NhIg+XOT3/D178fR05hJV5Ypqy1iBLVlX/oWrUX0sNf/aFuHhJoZsIiKPlhy8uhXk6SRusiGJKlQjqwS3sApSX7Px4oEq3OC6UF9Jhhflz8uL8QP0iCQSlhJciYXfZmgLzn5jNLd2PFrpPE663kTDKJkL8OyeePlQRWOwo17InvbFZd/CQ945SYV7hrIk7D/vXXX+HxePDVV1/hww8/xL/+9S87y6WJsOKVtpPi8LMsdh4v50NXkj6zhUtkjxRXiaKBVdX5eA3cz4qXnRr5xAtHG3IDRDj54VOZdPxme56mnznZvkpGGIdBq5Pq8WTjTM1K8w5KCNuhsNKLy+eIzTLC+uHKGeNUL9CBU5U479nV/NdcXhk53MFeFVNaHfdAkgew1WwMc32L98MG0H3G90Q7P3def5qB/4Xt9/HmXHz0U67oPEeDj+UXPxlF2E//+e0e/HyIHN5U64XkE0xw/5pbqro/qLRfuoK+rML6L6+pR7hUNdMCu1+/fpg9ezYAoEWLFqivr4ffgnuMXjbzUbpCFbfxQJHqTiEsC/yw95QsdKXwwv/7+Df+77Lqejwt0aY4GVhV68OP+wXavd0jT8Db6wMB5Cd/sAXTFTZL5bIXCiQ1tz49fsKxOjWXQF5cOVhNu6w+k0jgL6kNW7fLBgDPyQrsyhMLUdLLgjjDL0iH05qX7pAHKBJiZgWrnVotYG2Si2UDK1mVnkPNHVOWVvB/mUkr+L9U4El3NTfyGNI63HJYQWBLfkvrSrprupoSJm020nZlOYVVYdOwTU86ulwuuFyB2xcuXIjhw4fLdk73eDzWSkfg71/kAgBKS8tE6R85cgQAsGv3HhSXiQfrvn37UFIUOlZSUgyPx8M3EKmcTG2V6PipgsDefS8v2YqVu8v444dyckT3SdMSvsR8fp+hOnlpxV5c0qYOGw8UIS3eiVvcoeaqqfcjzsXgcFHgS6CuNqT9HTkSci/cu28fEmMC7XIqPx8FlQ3EcgqprQmF++SuK6oi38dpkAf2H0B5eZnq85QUBwbUgx9vxD0XZojO+XyBgctvUBAcGMt/C8Qzb/Cre8/kHg1tKVZQUCAra12tl//N+90S0jwqSKewKPBirqyo4I+R6q2giCwo/IL0a+vEnijHjh2Dxx+6z+PxoKFefY6hrjb01ScrB8uivj645LpeHu+dVO6ymkD59u0LrNLdu3cv0czAtbEwjbKyMmK6eccrg2nuR1JsKK3qqmp4PB5UV1eL7vt3MGgW9/uk4GtFa6ys21eIL9eGVn3OXnMAt5+fLrvu0KFDqCuKBQDU19ehvFzcFvnB/nLoUGAs7927DwkxZF3WK9G+vTU1wXKKJfmR3FwkVInn2ezAspfImjVrMH/+fHz00Ueyc26322ryMuJijqOyrg6pqanB9AOVnJWVBeAEvjvqwI58cVS5bt17oE3pUQClAID09FZwu91gmCOCcooFb7s2GYLy5yAjIxNAKVqktgJQxl+XlZUNIKSBcfeUVdfhvGdXIynWCSAwMJxOp6xO/H4WBwoqZfnz6Z/VA0AOXDEu0b1Z05bhmSvORt+sdgBOoEVSEoDAgG7XoSOAwAume48eSImPAZCDtm3boqaoGkC5atvErC8F4BU9TyD64VHZfQFTUi66de+OtMP7AFQqppuZkQGgDKsPVeH124aKzjGOIwB8Mg/WL/4sAwA4GAe4eiTRrn1HAAE7aWZmoK2E7ZeclMi3c0DDYpEYHwdUioVkp86dAQQGWqtWGQBKkZKSAqBKUB/itmqRkgqgAlL+PBUSDHGxsQBCgrRjp05w92zNp+V2u+F05ak+Y2xcHJ+GuB1y4HAwiI2NAaoaEBMTA0D8XNz1mw8WYUBWOmJdDhRW1AI4gm1lCQCAXr16STwvcohpAEDq3p0ATsv6wz5vHoACdO/Rg/dgAnKQkJgAt9uNhB/LAHhl9cils7PqGIAiwTHyuOCYtUkcFzuzdSaAEtGxrl27olvrFgByEBMTi5YtEyDspylp6QDK0CU4lnv07InkOLJo5Po7R4vkJLjdbjiYXAAhYZ6dnQ13p1TVsiuxbds2xXOWJh03btyIt956C//73/+CnTr8cJODX/56DOMF7mPcJ8ihQrnA8LMs0SNA7aNUat/kvvCk9i2lL1vOt1frU7XrE99jzOwfFc+7n1qheO7kaS//WegSlFfRxY1hdK3qqifYMpUWRIT2WNR2VePjdahcpuzRo85rGlvBiTcIDvxN7BOC/DmXP2F9kEwZ9TrqNGDXFGYE7D8lF/Jq7D+l/DLUy03/24LlwQlA7tP/pRV7LafLwZsTpDZ7vaYtg9TrGI/CYzPX5Musy/UNYpPIMpXgb1KrEW9WkyRq976TfLpmb6yoqMCsWbPw3nvvIS0tzc4yqSJcCkrymybVk59lifZKtc7jcjiQNW0ZfpN4h3gbxBqQsr3L3p5JKn6Mw0H0EhEGB5IKGM5OqebiJ/RN9klWHEoR2v207HZcWqTLuGIqzfdo2YlLqtR3+HCIZvHlE7UkuEVMwpc0qRgNOiIZFkvKV+fzY/R/lV/URhHtlEOQWsKJPZ+fRV2DHwNfUA+Xq4RwZaf8XPB/yRWaUxCsvG/oWeZfp7ISUm/e3BZhbwZ3nn/8a/J8ESBXKDh5JHdPjLBJx++//x7l5eV45JFHMHnyZEyePBknTqiHJbUDpUHGHSUKDZY8YNUakpt4yy0O2Ny4hvp8y1HRdUbmm6w0IakDxDgdOgaC+DcnNHs/vRK1DeTPb6FGyW08qxTZjYtUp2d11xvBAaG2yo8kmLu0SlSduZelRdKyBItbSC+50HUhuPoRCoUGgjZNOqZdRvtnq7VemNzehiwrn+zjjgOB1bPf7lAey2Nnb5SNA2kayn7a+p9bKY6IED1hf7XWZXD9feXuU9ppSZraIVg4IyTiJh2vv/56XH/99XaWRReKApt/08nP+1mxyYBDrfNIP5eVBthVb/6kmIadkDqAy8loaiEiiwj07YgiFFCcZqi04lC4lZfRz8CVu/NRU+fDVf068MfyTssnzDqkJuBIsb59D5Xw++XPQKrT2z/6lf+bi1XBfTJz6UjRIzSk2C6vddT90JfXBfIGMO+XI7Lzp731SIh1YurCHar1vU/FlKNk+eCDYykJcpaLR6O+nkAL0jiVjvPFf4hfRnUN+vORmt5cDkb3gjQ7iLqVjmY0bD9LdjlTXYYsuVyvJl1YUYudx8sxZf6OYB72jEzSU8c4GcW4KBzSTq+1IwrLsiiqCH2+l1eLBTbHrrxyeOt9st3N9cA9y5Sv/sDDX/0hSt/sHg/CQUlKosEviIKoMw/uxSVMm6hh6/gsl2LHXhb/WbkPx0sDglXNrVWeN8t7ZwjhdhQy22W3HSnlA5A9/90eXP1WSJnR/BIkldMmL2HOBKQEac5GiQMF4peVw8Gg99MrRYGogAjUsJsKJYHN7aKsJLC1liFL4Totd5fet/1jC3dg3b6Qn7bws86KryzpXpfAhq2E8DzD6LMFx8U4+E/m/6zaj26tk+FulxJMjwXA4PI5m/DQyO64eVBnPh/dA11hdZieZ9AD6XpxQKygvVSjOUhBgUhCxKgftoPR97mvxRvrDiJJwZtBFfutMQCA7cGtzYBAKIQKgeapNGnNcdYT3+Pdyf1FLx4zdUS65T8r92HjAeVd1o2YtBIk8UuU5ErE2bCbClL8WSDkz6m0mapw6TnLAo8v/BOna9VWNAX+57LTOyalg1coKLi0DhdVGfYQIBHQsEOTfiSkx7U07LyyGnRITRAde3nlPsECmRC1DX6RANNbR8QWVLnXaMAg0vUNfrkNW2tQheJdq2vYRjQ0ABjRq43Cp7txhHtR2iUkzG4ZJuxb0slAJVOJkD+Pl4leolIPELNU1jaovlSNmLQSYsUCW0keRdzS9KZCa6cO8vZW8or96jd57Goha/eKJyDMmjZIHeUvczYZ9hAgdYAYp0OguZDvU5p0VLrneGkNOqaJBbaDYRRtg+LYEfrqiOTepyYkjJoPlDRsPUKDlK8wf9ILzw4bdta0ZUG/aHOIvUQ08taoAbPKv7Cvy15iOhL1+cV9wtQKUsKxhFj1rxAjL9yTZeI1HspzarqTNETUCWwt00ZNnXz2W9GtT4Vfc0tFv81+RZL6qZmOyNngF/x2jN8ayWVQYDNgRHmT/L9P19QHF9oI7xNoSJJ8hG59Rge6yO6qahPRTkvp/rk/5wIICmzJZ7nm5rQE+zzpM13J20YJ4W41ViHN3Wj1L63up1S019ao+7qLY9hI0tRIO3BOfJKro8RYeRhVO9ErsI+XVmNrrnhRjvKcGjWJANDWsIX2Yw4zAluKlikh3HADcurCPzE/+HUgNIkoIZ90DHXOoyXVoh1hAtfLO+GBgsqQP7owPzZk0zViwzaqfVjZ1ZvbiDagYXMvl2A5NO4VLgriIPlcHykKeVR0zUzSLJPUW8duhHsdklBrJ24Sk8Qaj7rbm55nUrvC52dFgq7BzyLO5UB2hnad8ukTMtD6Oq7X6SUy5KV1smNUw9ZAI8AaEZIQ0kvIhm19gHElMLUlk+DvRMHEhzD4EgnhUYaR7+Ax/vWNKDjtFVzPEic4PSflNvd3f8wRbatm9LnsnHRU9vsN0CDSsHUVj2jDJvUD4eTaVed1kJ2XIgx2bxWuCqvrfDga9LPW6utq7fTmukOKUQl35pWrzr2ohiXVoWJLb/f7tQOK2YGelapKaHmt2U3EC+zK2gb4/aEQjC5SeCxN1Bv+1dX7tVOwUSGymhYX7yGg1Yon0uR5KWvYHG9vOMT/7WfJ/tTcZ780nyve2MTnY2XSUU0L0vOyFAoZoinDz2LxdknUPQ1hQApsr2Vu0KMYMGAMKwAXd2ul+1qliTAOtUf4TmXBDKAulNXqRs9+qdLQxQ1+Fg7GYAQ/QQ4T+rZD344tNe8xOmkshGrYAiq89ejz9MqQpuNnieEMtfCzkgkuyWB5XSUOhd22qP2nKmTbUulBOPHHBd4/WlLN+38qDYRtR8S2eNInPRerGAjUDenlRpobAEIhSFkYMYmQV4cpYfT95mNDgparM5+fxfajZeJyaKRD8jHXWqikx300YMPWvEyEUjAibvNgIXqWayuh5Uqn9kKysvs4EOhLwhDIPosadryLtEmZHK3QAkp9H1DzEjkDbdhcRYVm61mclZlsOB0/a3xCTIpem6NaQzEMYz5+hGCQc9ruq6v3Y/aa4NeBQvEe+vIPkclD6zn8Qa1GinRhgBSW1W8SYST/A/b6YXPySugZ0uBnEeMSP5jWmOI2EhZqw1oeIbo0bBN+2KTdcYBAEDTDWBgMareqadh6zFFf/34cSwSrEH1+7fg0aiTEOnRpBWovuG1HSmUB2IRN/L9Nh4n3nZEmEU4T5XxffSyL5HiXorahROB2axJb7RNWGENY7dNeGqBojAHhzQjKIHzjl3IrEVWeT7gruB4PAtJLx6tgEgnB6q5i6a7V6ukabzmh7ZlLt6iyFvN+Ice/UGJnXrmsbFqLLPRohAzIbpJqHCZo0mZ5cgl5qzM9qLpfSvqW8OW15+RpLN6epzqOkiTeIA1+v6aTgax8guRjgi+5XcF2VELNJEKawNXVxmekhi2JlrYrrxzbckuNC2yJhm2mMj9TCHYDkPeW04NaTAYpDMPwzyCOphf4X238c4NEzTvhwS+2Y9Hvx/FLTjFRKeFeEoqTm0F5fVFX/bZW3a3AkrV+JYTblqkJCKmJRIkurRL5v7XMWbrKyeiLRCekoCL0lRSvEFy/MSBVZ7cnvsfavadQL3kmqXmIC0OgRKJkXPv8rKY9Xo0YpwM7jpWhtFoen0ZIg09Zkycd1vMSOSM17Oo6scD+2ye/4bcjpYY/kyq8DSIxYyjuhY68YkW7N4enqZQ0bD2TOWINkXzl0h0nMGX+DqzacwoFp2vxxZ2DROe1BFXAhs0Sg2xJEWr8ejDq5cMJbDuWfwPiNtXaLks6mK84tz3xOqM2bE6ri3M50Dk9UePq8HFaGtcbgT6141g5qiVBkGob/LJ9FqW+8EJiJHUXMIkY1LAFf+vtM0WVtYpaM+m4npfIGTnpyAkmTli0SYkHoO+TRMiN7/+iGi7SKuY8V4whfGSvIDQmJ5ykk4tChFqxXs1OugSXy1NtgU7Aw0R/2+ifdDQ2cFmBScQOhF8lWotkpKVsmxInu8bvZ/Hsdxo710vghI+RyV07kO7vqRR3nAVwmrBz/M+HimTXAcC41zbKrpU+lo/zEjGirwoqR2/8oHofi97tyRuwkLqdXk+gcBDZAju4xx0nsI3ERJZyotyrfZFJ9GiVVgnYPQN/CzdR0COAuUumLdqpK9ANC+lGXQKBrXIPC3M7bWhFu2MVXA2V4L1E/NYnmwGpwFYvq/TFQposrFLxOlBCKCT0PtKALmmG85HWl1SbVOtuJO27hWTVrCe46cjefLk5UGquM+MlYkbDBpTNmqTs9SR7RmrYB4JbInGCmrNpCytD7+ehaNssm8rHYTQSoBmEy5mFK7P0fPZXCjQfLW8PAHxsYiG6vEQUXALJ14e05ncEfuBKaSulSxpoQv9pK6skOYSauqaGLSlmDEFgG7VfA2Lho/fLIT7G+pJueZhh5bwrCXGhSRslKCGtlme/24P80+YVLbVxKX0upccifdlZXTVthYgMr1pSVYfzn1vN/+YGCafdCOtQ92d1GD8jlVyu7IZ7hHrRTuzaDzbpgy3830qr2KRIP+lCXiIqk46svgkZ7nruygKNoEdK8cwBICM5TvZMIYHN2hJTWeinq2XDltab1C4LmPtS5DRdlpvdDROynVMk55UE9pI/8oibHuhZlMYh7Vu7JWETjKLWF10ORuSiqfRcpBT0um6GA0uSZv/+/Rg1ahTmzZtnV3kABHYcF8J1cM5+Jt5QVR9m7Zl6bKeNo2GHVscJBUg4XkTCNGeMD+xmrcckUtvg1+0xI54E1r5eqYpJLyzOwmJGk9XKQ9skIv5NepnXmljYwtuwg/J6qY45GTu+LqQvSqW2srojEGDPnIMwCbVxKc2KlPNVb/7Et/cnm3P543Eu7S+XiHPrq66uxnPPPYeLLrrIzvIAkDecdJDoXXAhTlNwj4F+8eAX2zWvIX322g2DULmFg9XKqjYlhBvqcvZ5fkWlyqRjbYNPV2cOXB/KQ0+4T6G2dNnZbURllRLyw5afP7dTqq7yCRGanbSWMUtlRAxhfsOUhi2YdDxcVKWrX9rxMjdiErGKUspml6Y7VZwBpN5SpHf7H8fKUBz0aHp6ach3XTohTyLi3PpiY2Px/vvvo3Xr1naWB4C8MqUCm3vrX3Vee2KnvP/SbrJjHsIO63YhnHQMV0MxDAx/Cmckx2JgdrqlfElaipKnQF2DH3E6fYSNPAoriQXTKik2dI6QkNCtT3j+xoGdkCG4Vy969sHkkJpESPVXZzAcKxAS2OG2n1ZLJkSlmmI4g1YqfREZeUfo1bD1ZkJKQm/4gXBg2obtcrngcqnf7vF4TKW9r0g80XD4iHjRSl1dQGCUlZ/m/xZSUqK8HRAAlJYqu8CZod4bKm9lZaWtaXPU1Xqxd598Hz41fA0+eGuMf6qePl2B3NzDAICCAnFIzX379mHRbvnKsZzDh1FeWYWqeH0bkno8e9HQELi2tLRM9doaby38As32dHno+vp6eX5l5YHy7dizH1XVoecvLyvD6SrjG6YK08g/pR5i9OTJk6LfRYUFsmuqaoxvVFBXW4u3/tIRFbU+PL7ypPYNAKqqrK+OZP1iAZ6XJzbFmB3jJEhugQDg9eqbdwGAwqLQ2C8s0N4FnaO6RpwH91z5+fK6rqvTbr9DBw+iLNH+KcKwTjq63W5T953OKQZwAi0TYtDg86Nt+w4AQh0/Li4WQD1SUlLgKvMBEDd0x3ZtgB1liumnpaUBsE/jTmmRBBQEhHZScjIA/R1MLwkJCejeoweAI7rvccW4kJyUBMDYTHtyixbomt0VQB46dWgPIDQIYlt1REIKA6BMdE99Qgb2FubhkrM7Ah7tuu3ZqydiY04C8CE1NRVq7RETE4t41gd4A8Ijs1U6gIBbmNPlBCAWKsktWgCowr1LjyNWYK5KTUtDJWpgtH1iYuMBBBSDjMzWAEoUr+3QoT2AUEz2Du3bQVh/AMA6nJD2WS0SExMwfvC5OFpcDegU2AmJiTDa9lJcLhdQG1KK0lu3gfD5AmM8x1IeWiQkJICrfy3S01sBCLywOxLqXom4+FAbA0B1YlsAOWjfXtz/ASAxQXwtiR7du6N1cN2IUbZt26Z4LiLd+oQmkIRYlywKGPeZRlpE0DUjCcO6Z6qmb8dkjBDhxFI4P1iN2g8ZGF9kxN/L2bAln39XvfkTqmvln/Scp4buSUdWf101+MUbUAhtk6SvaKW9BVmWlX126/lq1oqHLUSaHGlVnJXNCxohPLQIaf08uXhX4xbAIMK6NWI+kjYrF4qXlIKuMRWJXiLhgvNIqG3wweVgZOEPuXYguZgte3CoZqfWu8OEXkiuW3YjnHTUfQ9jboAL/bBJnb6qTq4dcrt367Zhs+IXrxo+icAWTuSxLItbL+oiuZ6cjt8vF7h6JowbRDZsDYEtdYsj1H9jbV5kz/xg0/kcm0HYVkYWtEnrau4vgS9ZknBW8wC56rxAKIKIW+m4a9cuTJ48Gd988w0+/fRTTJ48GWVlZbYUiqt0b70fTgcj00gcKgM9IdapWVkVterBYIwi7BjhGosMo70dmNJ9ViAJtCrCAgmSN4QaOUUhW7+WYPH5WZGmL15EIq9zJS1YuGkwh3TpNQmhVq4kbO+95CwAcgFNeuGZ3dAZ0O/nDij3xfM7p+rPr4nktVlXWbGGrV+8KfphE4qhVrJ+ndMU77MD0wK7T58+mDt3LtauXYtVq1Zh7ty5QVukdYSVF+NklEOCsuY6fzlhCa0VGmPhjJltpRgwJgdcaG89soYtN4lwX0Hpifq8MCa8vol4PImgoTf4WZGgEg5mUvsrmRz8LGQLaWJ1tJ1Ph0mEG6BSjYxUf1Y0bEPNGcxHKqBfvq6v/vwEGaYmxihfaDMJsU64HAwW3TvY0H0iDduGt42S+eOSnmSzK5dlxLn1hRNhhyZp2Fp7AWq93XYel3s5WEFoEgmbWx/MmUTM2LCFJhFSp68mmER8fhb9OqcaGtQkYdsijrCUm5Vq2ML44/J0leKl+Fmxhv2P0T10fYE0+LQ1bCPR3qz4MtuhuRlZN8CV/6ZzU1GmEabUTtqkxKNB8mWlB2FcGiN934iGzQJIV3AP5RSLiFs4E06EA9nlcMg0bK4qWJCD+2juJKLgPqSHPh3kUb0aR8M2vg9gYNLRZH7B/0kaImnSsYGPXWzmk119oqjB5xcd1zJBKS1MkcbHDsRV165Tnw4bNlci6UAlm0Q0s1REr2303E6piiY0QxEVg/83ZoRAAEgMLk5xMOQnfvOm84n3+Uxq2Eaf7/oBnYjHubo9wzRs8cSBNJobv0KOsJINCJ/BHwBG9mojO9YoS9Nhzj5u5k0vzIckcEhBfnzB3UGsKhakqpROOgoFDmn7N6VAVSzLilapuZwOXQNVaBLRChIkLb9ZDXvKZT0U8tG8FQBQXFmLX3NLRWUzmgagb1VfOODK7HQwxH4/oW874n1C5c5pYF5FqU2UzGsXdm2F/7s4W3acN4lEmg07nAi/aF1OuYbNTzoqDh7jebZvqc9n8q8XyN+sRrcx0oJkYwzYsFl0SE3QnQ7DmLNhs2zoNUhySyMtz673BT5fjWQX2uQ2dIwosFmxwBYWidQHlCLq+VigRVzIZBPjJAsD2X2iSUd1rVVaA6Tn0ZLXbVPicePAzsSb9Pbt46XKvuaJsfqXXyTHN57dWojaFx7HbzNHyY6dEAQCU1KkBmbJV/8qNYnafANRWVToB3YRmQJbEoRc7iUS+F/RJGIizzY6BTZJYOoVinpfCiSE8bCNYGXXaUD/y0iqBeuB69TC5yK9ILz1fsXBJ4x7wqEUoKnB50dKQkhY6d14QmhiUbZhB/6XB38yrmGX19QrPq8dgiA5zoXbBmfpurapTCLc4yuZRIBApEYpvwk28lDqj3P/NlB2TOn5jAYQ48fbmaRhiwawg+AlItCwzUw62o3eQTRldE99CSo5xbCAQ99G0IFyMSb9sBFqA1KnL1PYJsrBMIZMMNylPwl2JVGaDlCK0EgaaF6FGMxVdT7cclEW/zvG5dDlZVRZ24DO6YkYmJ2ueD0/2aRSbm5CVivHmnqf/HOe09xs6Nt60+iYltDoY4lDaBIxi9ILOYZwXKldiS9XwpchB9d/w2Uljch42FK3PtlKx+D/QsEC4hWNg9743FZs3ZwPMRN8Peh975vVyLjPPZKGTg64FJgYNJIbd60wnrXSF4HSIgguDrcQJRt2dW0D+nRoyf+OUbCPComPccBb78fNF3YGwwCFGrG71SYdQ9qqduspa9jGUXuJqHFBVjpyi+3brd0I3ONb2YRXSdiTvhqVWsTovqC8aexM8hIRu/U5UEXwSgC4gcqie+tk0fHG17B1XmehXNzXREBr1peQ8DKSvU8tLw69Gk5Iw9adDfE5lLJTWgQRiOQnPqZkw5ZOluqZdEwJ2nA5F0mluM8OBQ1YKHC459Xzla1U73ZonnrvFL5YpDui60Vpr0QlQn7MwUlcCxLKDi8Rs5slh0sERajAFtuwPxYEDweEg4IV+Qy/O7m/6DqpILfCNed3UDwn8gu3wdZHmsxgEVhhaEyDDa2OJNn7lPMXmER0SuCADdv8UngOpfyUBp+flbtVkmzY53ZKxUVntRKnqcOLIDm45J4z96zaQ44ARzJd3tg3lRh/W4+GTfpsF6bvdDDEyTM9mGmjkhryS/Dt9eLt3XJnTRD95jbAEHJ2O2UhLn3xWXkxGblXaV5BzYZNuifcymJECmxWIrCl9GjdgngfNyHI3TFjgrlogSTUBJddjTTvjgsBkIW+n2Vx+ZxNyC2uNiS0rb5A9D6bNECTlJsv7Cw7Rkpbjw1bCMuyMsFLeuZv7hmMp6/oLToW49C2YQvTVhv/wkkyjlv6pfMxVoTo0dqUJns5IWHEVCCPb6LvXhahsdQqgeze9/LKvRp5y/N65i+9CVeKcQi/JEwOMLOxRHq0CSl6wo0LlBjlDrn6htOlGIhQgS3s0KRVWRdkp+P9WwaAZQMuZlI3v3DYj9QaXzhI1caiVpS2Id0zAAAtE+SuVMJ7jUw6mhHYKfEhIaPX3unzqU86Wv3wEHYDYRZ+Vt/yclKx9AxobuJqy+ES1brgBip3yS/TRxKusQ7XDRyMtaiTRjXX2/uTtfkkDRdBUjZqWfNfKsH/hROHpHGhhtZLLbNFHFY8PBT3XHKW6MXdTePLnLuSu0U4Xs44DdvvZ0VbYJE6FoPQQhJvg5/fHZrrwNwdhjwWNM6rD9YQx0uUNwwgCWzp4+18ZjR6ET4Zhcuj9b7FA3VkbFBPHtQFs67tK3j56buP07ANmWwIFyu1mVrKetzzSOm6dEw6ckK9pKpOtS6kgiZccTc4wWLGXfORUaHFOA+N6q4jr9DfSvklChbWkELrkupdbVxKX3zJcS7sOFYGQDl+hxIk2dEzI2QadDIMerVNweNjexneY1QEI/8zXF6QESewaxv82CbwpSSZRDh3NT/Lot7n5zUsqZAxJjzUr1bTSIQzyTlFyrPqZIEtTrdFfAyx3KJ7dWvYjOFJk87piUiOc4Vefjrz+vr344qTw4rlIzyIUnaq3jcGIwWG0mQ0Rxb3hdfg86sLGsmCCfLLyFQxRaQEtUyHw7hfvjD/FB0LYvQknyww+ex7fpzsPDdshFER9Wj3nKeP8D6jfVnrRS4sh7Au9dZrSEEMpcPVsZ6vPjNEnMCWLoUl2fI4ge2t94kqhqtntUFjFrVORgo3SoLkIqRXU9p3qoL/+65hXXFNP+VJUCF6Ox/3uSmzdxp47ZXX1KFvx1SRTU8NkjeHUjWrVZPR0K4cSsuehXAKQ51P7o0ihNN8uXJaXbBEyAAA+K9JIxYNK3ZVLUWGK4/W/Zsev5Sf+FYte/DcaYKvv9HInKSl6dzj3DiwE24d3CWUtgWdWFpFubMm6Arba4aIE9hSSI3LBP/V1AfMIdLKDmnY9g0aNXsYKbaGFAdDnnE2o4k9OronnrribM08g5Z9zesA4NkrAxNBnCZv1CQSuJZBelIs/nfrAF3Xnzot92lWys+qSYSEnok7TntvEMyVkOA17OD/tgtsUV6BaHEWvtxtQ0tb5k7HOByGXmYkJcjoF4XLweDcji1Fx+KC7fmva/ri78POMpX2tHG9RPeIqyC8RuwoENgKb0kG2HGsTBTb2mxg+Ll3yJeqSlELJFOhI/ofw8iX2AeOE67V0ei6XkaM/o7I1bM00p0RTS6sQbDCoGEzDNAqWTl+d2KsE92DHkl9O6YqXvfdA0N4Wy5XEnJVWKgfQUfZPG0E5t91kZUkbGNnXrnqed7bwxnqsWrl4OqNZP4wqgU7GAZL7h8iOjZtOPnrT5gfKZ82KYGvg6ljemLwWRnB6wKQTCLhIuIF9uI/8vi/3745FFJRWC+cUOIrUDIBpEXn9ETNa5RsUrOvPw/3XtINgLrfN4NA8CH5cZPCRlKcEb1aE/PU28V5gS0L7KS/fFZ8ZqXlkMKI/hZfc07HVFN5Ox0MFt0zGH8d0JF4fs+zY3Fd/8C5l649RxY1kkO4epIvo4mRK93qTIl2LRPQKjlOf8uYbBYrO+NwcO2pV8Pm48sQeq5CmHNFSOVPjVcy4Wi4dwa/4vp3SZOdEz5OmOW1eYE9e/Zs3HDDDbjmmmuwc+dOO8skQrjMeHC3wJuNUXAfk7aP3srTM5lx1/Cz8OXfBwXzDx2/ql8HZGUEBL6a0GAUTCJK12rBdfrN00YE7lG4Tu+g44ou1bCNyB0lDfuszCQb/MGVC3Jep1QcenG84TSdDgatkuPQSmVREe9e5nSgnvTGDTL+nHZ47qo+qsNeqy7PJwgDNfQG5gpptsbEidAP2yyh+mP4PqvHJEIM4WFQw04k+MAroTU0uSKTJieF/SJcS9I5TAnsX375Bbt27cKXX36JWbNmYdasWXaXiwjJ++P2i7NkgfAZ0oUKZLVKRGYL7VWAyXEuDOoaWCknFUyk7bRWPjxMlgY36fjC1X20C6aBdKGGkoucbg07mKB0laDW4GotqDulF9aaKcN1lkLFS4Rw7Lr+HbFmirye9WJUcCjtZAMEJm0nD+qi+oJ89LIeeGdSf8XzHL3akheGSXlaxzyGEDOyxC75Iwy9S/pY7dMhBQ+M6BYSjIQ0jHiJPHdlb5EHixbCdlOJ94TOreRf41//fpz/OyI17C1btmDkyMDCgB49eqCgoAA1Ncrxd+1CKJy4iqmp8/GV3baleKWjHtZPvdRQwwJywcR1MuHxnpJBxyBkwxbmZ3ZAyAU1wdYP/TbskDeEWChxvrXLHhwiuwcQDyxFH2qGQYKGNwEHSea/fG1fYj1lZyShm8KqVyN56bWmqGnYHGpXtE9NwNg+bZXvDd68gvCyJ9G7fUvcPfws7QvNYoMzsXAxm3RiVkh2RjIeFUSzJAtN/QUyWnQ/X/dDFa/JnTVBEl45SpamFxYWIj09tPIpPT0dRUVFKnfYA/9ph1CQoZKqOv48vzSdCV0XDqReCaGJJuX8brmoC3HS0cwWZ6S8lbxNhMevVnEF5DTs+qCG3bt9CubdcSHapybgx6mXond7uZ0W0P9CeHQ0eQeVqWPEIWeJz65QH1ZtrNwLVs1bRJgFaeMGI2i1q1l/cr35mhkPvdqmWPIpFrqycpPDwnHCvSw5hUGthNyO5HowGsea60u92qZYek+FW2CbCq8aEyN2umdZlvjW9Hg85kqlwP79+wAAeXnHUR4b6ERVVZWormkQ5VdUFfh99OgRzTS5e2qqlVcoCq8DgOv7pODLP8tQVe+Hx+NBSTD/Om8N8XoAmNCFxbd7Ay+1E3mhlZx+1i+751Sl2AdVOnno8Xh4be/gwQMAgMqKCkipra1FdZWfv+fqs5z4Zjv5+U4cPwYAKCwp48vRCoDHUxi4v4B8X0NDyEOmsqKC2ObcsS6pMThSJn42f1WJ+DfB7JB/8iTKy0N1e6ogEICpoLAQHo/YQ0dpopVUrsM5h1BdEIPi4hLCHYF7amp9aJvsgsfjQf4p5es4jh6v5o95vV54PB68Oq49piw/gWNHj8LTUITLuiVj9cFKWTp5eXmKZfXW1BCPlwjKPvOSNvj3xgLUSr4Eqqur0SrBiaT6MsX0SaQ6anB993hc160LvF6v5vWkdHNyDvPnfA2Bts85dEh23enT5fB4PPx48PnEYxoARrStx8Abu+CWBUc1n+Fk/il4PPIyc20ixecLjZNTJXLPl/r6etl9paVlsuuOHTsOD0vuJ3ZgSmBnZmaiuLiY/11SUoKMjAzZdW632eBLOcSj7l69AOSiU8eOwRVf+UhNSUGVvwZALZ/fqdNeAEeR1aULgJOqOXH3JKwvBaAc61j4LE+6gaxOR/Dk4l1wu93BGMlHkZycBBTWCq4PPYe7Vy+4ewHXXlyJQ4VVAAISMBA21CfKI6WsBsAx/t6UhBjExzh4v2W32x30WDiMHj16ADiKFi1aABC/dOLj4pHSIgFANdxuN9LKA/VCIrtLFwD5SEhKVmk3ebu4XE6+/C1atBDcK3j24LE1PXriv6v34y1BhLfe3boAmwr5306n3HTSoX17HKwsBBAQcm3btAFQgsyMTLjdwiXWOYhxOgieLtK+GChbj+7d0T41Aa3zDgAoU7znl/MCcw4tD3tUrwOAk8wpAPlwu93weDxwu92oSSoFcAJdunSBu1sG3ne7seSPPDz05R+idNq17wCggFiHj4ztA7dbbk5pdWQvX6bOnTrC6SwCfOIFSa3TUrD04QtwsKAS+CGfmL6UNVOGo2tGEv/lpUfIk9Lt1KULgBNwu91IXlUElNejR/du4Pq308HA72OR2jIVbrcbTsdRAA1wuVwAfKJxxKXvOVc4B0R+hszWreF2d5VdEx8fT+zfDsdRAH643W5Ufi+PyBgTEyO7r+XuegBiRalzp05wEzy2jLBt2zbFc6a+dYYNG4YffvgBALB792506tQJ8fHmt78yitCGTfZj5s6FewpAXAathRjxMU6RaaFdy3ik6gxoI/UDls64Ky/AYSS/yTgcDK7p10HVbEKCZYF/B/eg1KruGKdDZv9PSxT7QZOsAg4HWWsmHTOyhiZUh/quJ70I9MBI/geAK8/rILPrK5l4Lu7WStH2LUzTwTAys9w7k87HKxPPM1ZgBNxY7dirVGgG5EwiYjc48TG7RqxRa5kwVOp9QTddU0SiDbtPnz7o1asXrr76ajz99NOYNm2a3eXiGX9OoKM6GKHtKySxicuLdVba1zdlmS4XKYucIvlnLn+9qJMG+P7BocQlrCT7m3TsSDu4kn3SQciXfB2DV68/D6N7K0+KkTBq75MOpPQkscBOjCEEEAKj215tZNUjd6lewdSgZ9KROCdBTl8aT1kpJrNe4SOdswCArIwktAxTICo9+EUCO1DhTkKndEjealblnp6d6YUIr762f0dkqCyo4iDtKB+RXiIAMHXqVHzzzTdYuHAhevbUuVehCd66OeAGFZhlRvDvkIByMAxuvKCTSDMUvrU3TL1EMW2ScBBy1/CuePQy8mSZEK6RSqvk8Q+kZeLKBShPUkonJxlGfq1UCDAMsOOp0fyKLGlepHuEGJ1Xeiu4iMnqxF9qgnhgJAja5O/DAp+0SsUm76mnf8g4JV8pWqi59amVScmRRyawFZJXq2Jh0UkathCSgAkXnCKSIviC5CYWSftz8mOW+21C8gldJoVD6LmrtN1ojQp4APjHmJ4yr5KI9MNubNKTYtGmRRxx+x2GAW4Y2Bn/vf482X0MgNRE7TelEud2TMUDI7vLdtGQ5SNRd1+8+hzCNcJyBZ9DofbNBolqmRgjGtyMZACrJWG0o13aM2Snk0aqM0Kc5KWZGBNK44ngbiVK/uSkhRTSZeovXStvCw4nL0D0lXXSIO2ViKRyKnlpCIXK1DE9TQkNcUYEl1NBnh1SE4hxusNJnw4t8eczowGENGxl8501gSd8dGFdTtbTbhpVT1JMkuNc6NVWHAo5YjXsxmTlw8Pw7QNDRLZpNc1OpIk3ghmby8KoX6+SECa6JClpmdxpBVs+SZu5/eIs3WVRgjMniP2wte8TCtmUeJcs2ltmknwenJFmFCQ7I0l2TLpMnLRsnE/XoIbdu31L3cvHRflI7LQcQqHi97MYd0474gtG7wo/l4NRNJ1xtG2pb65JqUpIdf7k5Wdj9z/HKKbFhXJ1Edz6lPIx8/IXbSRi1CRi4V0pXEgXkX7YjU1mi8DyYaEiq+ZmKRRRRn1ISTE5tJBqUGoToeLryTQEH46bzAPUt8gKpBk4LxOgBJMIKS2ju1NLd/nh8zNArCsgrIV27D5tEvDOpPNF1zGM3Af64Avj8Jdz24uO5bw4Hvddqn/CiNNGSeW+1GCwfA6SoFBqb9EyZwQ0tusvkG+lpmoSEaTqdCjv+KNE10y5AAbUBCkZ0lZoUrJaBfIS+2FzbSBuCzOCTzh9QaqzgdnKe2BaCa8q/CI+I7cIU0K02a1KBQs7RHyMU9OkAQSEozRqn+7JHomgJu3wLSx7SCNXt2FPHNCJv15Layedltq+HXz55FcbHSAhgc0qdtH7dQpPRvK3NLaHg2FkHhoup0MmnBwOuQ1XbQBxLylSfXx0u3YERxJqJhEpQ7tnYHiPwIvBjs2bY5wOvo25STOtZr2uf0fsl2w88PU9g9ExTTsgGgfxJUW47onxbvzx1GXineS5/22Qc8K2NrrZgdb1aqetKC1GiUiBPXOCWzO+h1oFOwnuQ1p0zUzG0O76tSpGKmVC/5F3ySHcyzBAcrxcM+ncKlGmDSgJd6UA8kvvvxjvTOovEYbK9WI02p2ey7X34OO+DiRHJW3LQB6USgkjz8F7iQgKMKRbBr6+x3joUjVC9S4u29w7LsQn/xd4MagpIGrCQpikU/DCilWxFwtxMozMU4kUkS6UoXp6asS6HEhNjBXN3YTmPwLEu8Rhajn07BAvfFbDXiKy68096Blpw/7b0K4YfbbyriUs1BskJjgSG8F8HciHt10ra2ykgeNgGHx8+0DESybeUuJjZPGOSc+SO2sCkuJceOiiDDw0MrCAhKuWvh1T0TEtUSxYJeWUlsUIRBOMpJSd0hPQVxJAXojSpsnSwcMwjH6BbeA5uGcW7kvYuVUi+nfRFg4AOQ63WqgBtaKpyhdd5r9Av+fDmerY8eT5q/rw4WMbE6KXSPCPhfdchLWPDpffo0NSWbJhG7q66YhIga0LVvK/ABevYTeWyA4SzI68D6VccjoYICM5Dp00Pj+VwslyjO2Rgu5tyEGQiCaR4DHh7tBGFOwMocmCVRZEyXExWHo/OWiUEOnXh7RJ0xJjdC9a0TOwv3sgUCauHq7u1wHLHwq4ZxkZ5+Rr9duw1e8SntNXKKeDweSLuuCvAzryfVDteSYN6qIaWpaEHSNKTYnomJaIrpnJcnOXjrEsvMbAvD0Ae0xSmpnYQMQKbKX6++uAjji3Y0tVDdvKzid638wiv2rJ/1qf5VKNXA96H0ma5O0XZ/O7ZUsndoZ0y+Dt+8ZWtbGEvyDrrKTH0wp2xUiu2TZzFC7s2sq0SSQrI5FffMXRI/hy457Z5XTAze9Ub23kkv2wdZjoVPodyTMjlHgoUZeTwd3Dz8LL152L9qKoco3PWZnKm3kQJx1V0kqOc+HcTqmaeSq59ZlB2lZ6x2o4t4YDIlhgK/HydeeidUq8updIsNJI0fG03LIeG9tL5J2hB2nYSL12VCNtq7/DiH+f07ElHhoVMJeEPj85DV8+cPQgDuKu7V6pmI7i8dAZTgOUxulWQvocibEufvEVh5WdcbTEgJFJRyFK/Xl4j0y8QPDrJyFUVPiFTbru1I+ejUMAoFe7FoqT/VwSdwzJxgc69v/c+cxoPDZGe3Eeo6Fha9+vfE5tU91JgzrjxoEBBwE7dl1SI+oENoeeN6hUYPfpkIK7NOIHu9ul8N4ZeuGaSLpztuZ9Oi9koD9GhppXBO/FEvyjJWEVmlFYAF2C7lrCFD6/80IMym6lfm+wvm6/OFtcRkLTCjVstXrT8+JxOhhFYWLIJKLzfvm3GCktcsYJMU5+wQkJYYrCZfktgr7Ptn3qE1BzmdXjnTO0ewYGBCcTSSt3Q3/rc1cUdmGjNuz/uzgbfxuSrXhe7Vmfv+ocTB3TK1iGM1Rgm1l5JEW6lLh76xaqnd8sXBvxk2gahiyjgyiQvjkNW3wucDIpLjDJdvclgaXfn/3tQnRM0/8JLS0+yatg8FkZmmYWLp27h58lWjBCqp6aeh/hqByrCo7VT+luhH09rUw6GvEPthJPe+4dA3Vt5CGdb1C8Tkc/FLm66jCnaSHsb6R2VEvzqSvOxowJyjv4xLg0zJzB/6mGrYCeTx6phh0f47A8oElwAnoSZ27RzMO4UNBvw9a+cPJFXbD8oaGIC7pQXdwtw9AErdJ2Sh0s2E17Bpf4MmCQnZEk02iq67R3pgesTzRb1Uh7tiWZArTttIrmIQPlsTJ3M7R7JlIILqZqcHZy6UvlwZHdcetFWYr3kSZh7VhwEicwW5gxiaihpehJv17Dhal42JGAno7cIGm1OJczLJ4jXJITzmmHt9fLg7NLMSMU9H5q6dF64lxOwSSbcVjR34Ff25+8jOhTrpqOQj20T03A/hfEizmEmzGrYbV1jTSN0Q2O1fqeYqQ+jbSFSboIQsWIhq5HyHH5vXB1HxwqqMLhosOydpyiETBNcVWu8LeJluTMQID1LyUpmgKb8/wKswocwRq2eoXrsmFLwmHGGdSw1Tq7qmDUSNf425/R/YnYGJ6MpJ2j05JidZmbRMJetKQ3+L9K+RN1RJuz+vzhsPlKF4eQM1Y4bKBApM9xI89jRMhxC1ysIMxPKwaKHoRfCNJHiXM5cHY75TUBWsRp+bVzGnaYTSIRq2Fr9R2usdUuI2rYNjlKCm2VUhslN0BbxLmQ2SIOOUVVonvNxC2ww63IrgkRKyFVrQjExFgXquvUbdlW29dI2+i9Uk+JlEKfar3chc9rxSQC6HseUv2abVLRC9ti32QYsYYt7aO7/jkGTobBvn17TaV/gcZKS2oSsQGhDfu5K3tjTO+2iiFNjXJBVrrMVimN6fvzEyNRWlWHoS+vE11nRmjp1rBVzsU4Hdj+5GXGM1fBikJK+iRWKv/bN5+PlokxuOn9LappWh4vYdGwxf9LWTNlmGLsDq2XYytBoH3SpKOxhUCNu95PPaiVMW69KAuxLgcu7ZmJdfsKZS86K84GPdok48GR3VWv4cprxy49akStwNZlEhHGxA1OglR4lTcZMItU6+AGZnKcCzUEjdDMsNCrPWlp0WlJ5uODc7CKP6wkpM64c9rxf6v5xFpFa2CbetnyL3Jy23RrTV6lCmhX0U0DO2P02W2QnhRr2YbtbpeCw5KvQSl2KpCioEmScx/fPhDlNfrGakZyLJ75S28AgaBdo/+7ARd3k+8xawSjj8mvwYhUDXvr1q146KGH8OKLL+LSSy+1s0wAdJhENOagPr79Alx0ltwPOBx+kjKTiKZbn7FRzzDA/SO64/2Nh3VcbChpcwht2BYktkjD1rMaEMDi+y7G2SoTplaad9mDQ9ApXX+UupYJMbJ5ErUymSmbVldxOBi0TlGOcW2kq31w6wWaipBwD07L8wWCv6UmkSy11Z3SdCRFXvXIcAul0k5fjYi0YR89ehQfffQRzj//fO2Lw8Qot3JwKAC4pCc5rnU4XoDSJEVb1tmUn3bkOy7vxpDYIYxqnEoCXk+8DQA4T2OJshVbqHCDZC0cDLDioWGm89JLYxoptL5cNk8bgaRYF859dpXoeI82ysvQ1RDbsE0l0Sjos+0HCLdJxNS3ZWZmJt544w20aKH8KRduzG4sGh4NO6QdfnHnIAzqGtLsWyXF8iE0OcJpKgxzfwEgdeuzGYvlb6xxzzAM2raM17WDi5Uu19h2ZTXapybw405YqhG91JUnJcRfWJGLnjbg2tjqxK8WpjTshITwB5ax8qnd2PAuaWBkZhiGYfgg9RxGfUSNdIFw73gB2OfjGg5h1FiampFxKQ26pZcLs9NVwwzrIYLkvQy7NOxwP6IR75lwf+FqCuwFCxZgwYIFomMPPPAAhg4dqnBHCI/HY7pgl2c50LNFG800ysvLDeVTL7A5er1e1Xvz8vLgiSnXTJPreDk5OfCXak/q5Z2oABCqn9raWtFvKQ0NDfw5ByO/TvgcdXXqadmB3+/n02dZ1lBerR1VSHAxqGlgRekcLg6Uu6621lLZi6rEKyLtrofS0lIAgZeoVtpcuxQGy5R7+DAc5frDmT4zLBVAjaVnOJx7GK4KYyFUSUjHyokTJ1BSWgfAfB0fO34cHmcZAKC4qNh0OsLxoYXWmOfwNYT6UW1tneY9XHCyQwcPoCg+fLvTawrsiRMnYuLEiaYSd7vdpu4DADcAbQthDlJSUgzlE9gbMDB5Fx8fr3JvDjp06AC3u73CeSmHcdZZXfnQnWrsrj4OoJDPO255AYB6hbLkwOVyBc/lgGEY2XUej4c/Fr+qCECdpbpXJgcAwDAOvjwsjLWz2w3cMQbImrYMjMPB3+s/UQ4gD/Fxam2izanTXgBHcefQbCTFueB2q6+6M0ravl0ATsPhkLeDFK5d0soDZerataulFabGyUFWVhbcHVMtpyTsY0AO2rdvj3LmNIByk+2Vgw4dOsLtbgcgB5mZGXC7tSPykdIJjQ9txM+hjCsmD6gJeHjFx8Vq3uOt9wHIhbtnT9PmWo5t27Ypl8tSylGI2TCidmLFFKD1Kd4YNmwhVuqIeK9NNuxhPTINbflmFCP9yIqXiFXC1YfbtYyH5+RpS2moufUZSye8RhE9Ptxc24Z7abopgb1+/Xp88MEHyMnJwe7duzF37lx8+OGHdpctLOjtGG/dfL6pHdT1YLR7cZ3hm3sH694cIZzYNb8gTIezAVoIOMclJErPbs7vkoq5vxwx5OfblBNqqRa1PRL7nh+LOJcTa/cWWEpHNBcSYW4i0t3o9V4fkW59l1xyCS655BKbi9I46O0X4wULNfSi+0VvUt7165ymeU1juPWFQ6HhhHdijDUVRa8/t1mu7tcRV/cztg8i5+rVGBPCQvY/Py4si4zibIgjAoiHQWN/GRpBj+cHr2GfqfGww0Wj7/NIwKiGamSgh/Ppts0cZWt6QsHPrQi1qqFEQPPKSA360Evjs4ebcK4ItQNx8C/zDRduLxF9Grb+a60Q2S0aZegVxFINtXN6oulg8PJrw9dhuO267BogwnQqvPriXWvBSP6PBLgl43Y9Y6RgOWxL5FpERJCW/EuJ+KXpFPNIA9O8cdP5ujeZ1SLcHf/GgZ35UJMMY595JEXnSk4t+BdWhAmAVY8MQ3fCbjRnMkIFx4piGu6m1mMS4S6hwZ/CxHNX9gZQ0yR5SzXxhFinYnhNo4Tbhvava0JbeTGwqG0Lbu7fJQ0HXhiHg/v3WUlRtIgpktDj7nmmwb3sP/vbhZohB1TTsac4iugyiTDK+4TayRlrEvnLuR1sT1OvthntS9M5LG/HJRlqduy32ZQudGcal53dBkO7m4+Kx42Di7tlIEnHfpJNRTj2gTVL5NZSGGnXMt42jdYMht36DF0bPZIqkpdNU7S5sGsrXNhVHhFTL3Zv4xUuwj2RaISoF9hmmvzn6SNtL4chwthRp4/vhX35FWFLX0jkdOMQ0k0kKJGLbZPXYZb7VjaXtpuoF9jRSDj7V7/Oabr8te0gIs0OvEkkEgtHERJJkQilcN1n28xRoq3HmprIMc6cQRjtp5EqfKzvn2g/1IYdHdw4sDOG9Qhf6ACrnN8lDV0zk9AqOS6i/Nmphm0j+icd9Yuqi7q2Mh0gPtIJS3hVyf+UyETobWSVcPSjOTf0i8gAz1EvsFu3sB46srHR2glbyBd/HxS+glglAqVipH6NUKKLcPtTmyWqBfZvM0chJYLsS7pXOoa5HI2F5ZVutpRCDK9hR+Z4o1AsEdUCOyM5+rRrILInW4xgeRPWMFRDqExUYp8pNI/RpI+oFtjRyvUXdELHtMhxFTJLJPp8hztaH4XSlETO9OcZRIv4GIztYzx8a6QRiUIxQkOJUCi2QAU2pVlCJx8pzREqsG2kmZimdROJIpHKaUpzxpQN2+fzYebMmThy5Ajq6urw2GOPYeDAgXaXjRLhWNFi7QzNKkqXLk2nNGNMCexvv/0WcXFx+Pzzz3Hw4EE89thjWLRokd1lo0Q4kSgU6UrHM5Az6MvWlMAeP348xowZAwBIS0tDVVWVrYWiNH8sx9JWSTfwP5XYZwIDs9ORkRzb1MVoNEwJ7NjYUAV98sknuPzyy20rECWKiECZyJlpqIZ9ZjD/rouaugiNiqbAXrBgARYsWCA69sADD2Do0KH47LPPsGvXLrzzzjvEez0ejz2lDBNer9fWMh4+fBjO042/mMfu59DLY0MyUFPvt5S39F6rz8LFWG6qthDSVO0SDprLs0T7c2gK7IkTJ2LixImy4wsWLMDq1avxzjvviDRuIW6323oJw4jH47GxjDnIzs6Gu0NLm9LTj73PoR8rWXZIOwlvvV9WbqvPElhFerjJ2kJIU7VLOGguzxINz7Ft2zbFc6ZMIseOHcPnn3+Ozz77DPHx8aYLRjlz+fb+IWFamk5NIpTmiymBvWDBApw+fRp33XUXf+yDDz5Q1LTPFM40P2wrpCae2X2FQjGDKYE9ZcoUTJkyxe6yRD1tUqIzGFVzhL48Kc0RGvzJJhpji3sKhXJmQ5emUygUSpRABTaFQqFECVRgUygUSpRABTaFQqFECVRgUygUSpRABTaFQqFECVRgUygUSpRABTaFQqFECVRgUygUSpRABTalWUKXplOaI1RgUygUSpRABTaFQqFECVRgUygUSpRABTaFQqFECVRgU5olbFj2ZKdQmhYqsCkUCiVKMLWBQXFxMR577DHU1dXB6/XiiSeeQL9+/ewuG4VCoVAEmNKwFy9ejKuuugpz587F1KlTMWfOHLvLRaFQKBQJpjTsO+64g/87Pz8fbdq0sa1AFAqFQiFjek/HwsJC3HXXXaipqcGnn35qZ5koFAqFQoBhWfVFvAsWLMCCBQtExx544AEMHToUALB+/Xp8/PHH+Pjjj0XXbNu2DYmJifaW1ma8Xi/i4+ObuhiWaS7PAdjzLOM+ycHsCe3RM6Np64S2S+QRDc9RXV2N/v37k0+yJvjll1/Y0tJS/veFF14ou+a3334zk3SjsmfPnqYugi00l+dgWXuepcvj37F/HC21XhiL0HaJPKLhOdRkp6lJx7Vr12Lp0qUAgH379qFt27am3yYUSjho1zKytSgKxQymbNh33303pk2bhlWrVqG+vh7PPPOMzcWiUMyTO2tCUxeBQgkLpgR2Wloa3n33XbvLQqFQKBQV6EpHCoVCiRKowKZQKJQogQpsCoVCiRKowKZQKJQogQpsCoVCiRKowKZQKJQogQpsCoVCiRI0Y4mYZdu2beFIlkKhUJo9SrFEwiawKRQKhWIv1CRCoVAoUQIV2BQKhRIlNHuB7fV6MWrUKCxatAjFxcW444478Ne//hUPPvgg6urqAACrV6/G9ddfj6uuugoLFy5s4hKT+fbbb3HNNdfg6quvxvr166P2WaqqqnDfffdh8uTJ+Otf/4oNGzbg8OHDmDRpEq699lo888wz4Kx0n3/+OW644QZceeWV2LBhQxOXPMT+/fsxatQozJs3DwAMtYXP58NTTz2FG264ATfccAOOHTvWZM8ByJ/l1KlT+L//+z9MmjQJt9xyC06dOgUgOp+FY+PGjejZsyf/OxqeRZHGifDadLz66qvsNddcw3799dfsY489xi5btoxlWZadNWsWu2DBAraiooIdNWoUe/r0aba6upodM2YMW1lZ2cSlFlNZWcleffXVrNfrZfPz89kZM2ZE7bPMnTuX/fe//82yLMvm5+ezo0ePZm+++Wb2jz/+YFmWZR944AF28+bN7JEjR9i//OUvbF1dHVtYWMiOGzeO9fv9TVl0lmVZtqqqip00aRI7c+ZMdu7cuSzLsobaYuHChexTTz3FsizLrl27lp06dWpEPcu0adP4Z5k3bx47a9asqH0WlmVZr9fLTpo0ib344otZlmWj4lnUaNYa9qFDh3Do0CFccsklAICtW7dixIgRAICRI0di06ZN2LlzJ/r06YMWLVogISEB559/Pn777bcmLLWcTZs2Yfjw4YiLi0ObNm3w/PPPR+2zpKWlobi4GABQVlaGtLQ0HD16FOeeey4AYMSIEdi0aRO2bt2KoUOHIiYmBhkZGcjMzEROTk5TFh0AEBsbi/fffx+tW7fmjxlpiy1btmDkyJEAgCFDhmDr1q1N8hwA+VlmzpyJ0aNHAwi0VWVlZdQ+CwC88847uPnmmxEbGwsAUfEsajRrgf3yyy9j2rRp/O+qqip+e6D09HQUFRWhsLAQ6enp/DWtWrVCUVFRo5dVjZMnT6KkpAR/+9vfcNNNN+Hnn3+O2mcZP348Tp48iTFjxuDWW2/F1KlT0bJlS/48V2bSsxQWFjZFkUW4XC7ZFlNG2kJ4PCYmBj6fDz6fr/EeQADpWZKSkuByueDz+fD5559jwoQJUfsshw8fxoEDBzB27Fj+WDQ8ixqmN+GNdBYvXowBAwagY8eO/LGYmBj+b5ZlwTCM6JjweCRRV1eH4uJivPvuuzh27Bhuu+02OJ1O/nw0PcuSJUvQoUMHfPzxx9i7dy/uv/9+JCQk8Oej6Vk4jPQr6fFIxOfz4bHHHsPAgQMxaNAgLF++XHQ+Wp7lpZdewsyZM0XHorldgGYssNevX4/jx49j9erVyM/PR2xsLOLi4lBTU4OEhAQUFRWhdevWyMzM5D/RAaCoqAiDBg1qwpLLyczMxHnnnQen04msrCwkJyfD4XBE5bNs374dw4YNAwD06tULXq8XXq+XPy98lv3798uORyJJSUm620J4vK6uDjExMaKXbyQwffp0dOzYEQ8++CAAROWznDp1CocOHcKUKVMAAAUFBZg0aRIefvjhqHsWIc3WJDJ79mwsXLgQ8+fPx8SJE3Hvvffi0ksvxQ8//AAgMFM8fPhw9O3bF/v27UNFRQWqqqqwY8cODBgwoIlLL2bw4MH45ZdfwLIsiouLUVVVFbXP0rlzZ+zatQtAYFAlJSWhT58+2L59O4DQs1x88cXYtGkT6uvrcerUKZSVlSE7O7spi67I0KFDdbfFsGHD+GvXr1+PwYMHN2XRZSxduhQOhwOPPPIIfywan6VNmzZYvXo15s+fj/nz56N169aYN29eVD6LkDNipeOcOXPQoUMHDBkyBI8++iiqq6uRnZ2NWbNmweVyYfny5Xj77bfhcDjwt7/9DZdffnlTF1nGV199he+++453izvnnHOi8lmqqqowbdo0lJaWor6+Hg8//DAyMzMxffp0+Hw+DBw4kJ93+PTTT/H111/D4XDgsccew0UXXdTEpQd27dqFl156CXl5eXC5XGjTpg3+85//4B//+IeutvD5fHjiiSdw4MABJCQk4JVXXmmyTaxJz1JcXIy4uDgkJycDAM466yw888wzUfksc+bMQWpqKoDAZPbatWsBIOKfRY0zQmBTKBRKc6DZmkQoFAqluUEFNoVCoUQJVGBTKBRKlEAFNoVCoUQJVGBTKBRKlNBsF85QoptZs2Zh9+7dKCwsRE1NDTp37oyWLVuib9++uOCCC9CvXz/LecyZMwfffvst2rRpA7/fj/j4eLz44oto06YN8fpFixahRYsWuOyyy4jnT5w4gaKiIvTt29dy2SgUEtStjxLRLFq0CAcOHMDjjz9ue9pz5sxBWloaJk2aBCAQzmDr1q148cUXTaW3aNEiVFdX8+lRKHZDNWxKVDFt2jSMGTMG8fHxmDdvHpxOJ3bv3o0HH3wQq1atwv79+/Hcc89h0KBBWLVqFT766CM4HA707dtXU+ifc845fHzk77//Hh999BFcLhd69+6NmTNn8gJ++PDhmD59Ojp16gSPx4NzzjkHjzzyCN544w24XC60a9eOj/xGodgJFdiUqMThcGDv3r1Yvnw5Vq9ejddeew3Lli3DqlWrsHTpUvTt2xfvvPMOvvzyS8TGxuLBBx/E9u3bVU0p69atQ58+fVBVVYVXX30VixcvRnJyMu6++278/PPP/HUMw2DXrl3473//i7S0NIwYMQJTp07F1VdfjbS0NCqsKWGDCmxK1NKzZ0/ExsYiIyMD3bp142NnV1ZWIjc3F3l5ebjjjjsAABUVFcjLy5MJ7E8//RQrV66E3+9HVlYWpk+fjtzcXGRnZ/PLs/v374+9e/eK7uvSpQsyMzMBABkZGaioqGiEJ6ac6VCBTYlaXC4X8W+Os88+Gx999JFqGrfccovM5swwDIRTO6TQrtJIbnQqiNIYULc+SrMkOzsbOTk5fMjM119/nd+fUIusrCzk5uaisrISLMtiy5Yt6NOnj+Z9DMOgoaHBUrkpFDWohk1pliQkJGDGjBm48847ERMTg969e+uOp52YmIhHH30Ut99+O5xOJy644AIMGDBAZMcm0a9fP0yfPh2tWrXCFVdcYcdjUCgiqFsfhUKhRAnUJEKhUChRAhXYFAqFEiVQgU2hUChRAhXYFAqFEiVQgU2hUChRAhXYFAqFEiVQgU2hUChRAhXYFAqFEiX8PzPIfX/ITEhAAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAW4AAAESCAYAAADdQj81AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAACFU0lEQVR4nO19aYAkVZXul5kRuVfWXtVL9Qa9FbStCCICInazNCA6gEsP4DY6b9zGARVZZXDjNc6gMKCjgzM+BR0GsAdQZN9GVMBBxO6met+ru/Yl18hYMt6PiBsZGRlbZt6szKqO70935XLzxnbuud/5zjk+WZZlePDgwYOHWQN/oyfgwYMHDx4qg2e4PXjw4GGWwTPcHjx48DDL4BluDx48eJhl8Ay3Bw8ePMwyeIbbgwcPHmYZPMN9jGDVqlX44he/WPb6jTfeiFWrVlU97l133YUbb7yxlqnZ4ne/+x2uuOIKvPe978W5556LD37wg/jNb36jvX/dddfhBz/4Qd1+/wc/+AE2bNiADRs24K1vfSvOPPNM7e8//vGP+NSnPlW337733ntx8cUX44ILLsD69etx7bXXIp1OVzzOhg0bMDY2VocZemgUmEZPwMPMYceOHUin04jH4wAAQRCwZcuWBs/KGi+++CJuvPFG3HHHHTjllFMAAH/84x/x+c9/HolEAmeeeWbd5/C5z30On/vc5wAAH/3oR/HBD34QH/jAB7T33/GOd9Tld//nf/4HP//5z/Gf//mfaG9vRz6fx5e//GV85zvfwTe+8Q1XY0iShEAggCeeeKIuc/TQOHge9zGE0047Dc8884z290svvYS1a9eWfGbz5s244IILcN555+GKK67A4OAgAKBQKOC73/2u9t4999xTNv7Q0BDWrVuHP/3pT5BlGd///vdxwQUXYMOGDfjWt74FQRDw85//HH/3d3+nfadQKOD000/HwMBA2Xh33nknrrrqKs1oA4qhfO6550yN9l/+8hdcdtll2LBhAy666CL8/ve/BwCIoogbb7wR559/Ps455xx84QtfQDqdtnzdLV555RWce+65AIA77rgDt9xyCz7zmc/gzDPPxDXXXIMXXngBl1xyCc4880ztvPM8j29961vYsGEDLrzwQnz/+9+HWQ7c7t27sXjxYrS3twMAQqEQNm3ahGuvvRYAkEql8JWvfAXnn38+3ve+9+GBBx7Qvrtq1Sp8//vfx/nnnw9JkrBq1SoMDQ0BAB544AFceOGF2LBhA7785S9rx/vqq6/ikksuwQUXXIALLrgAjz/+uOvz4GHm4RnuYwgbNmzAr3/9a+3vxx57DOeff77298TEBG6++Wb8+7//O5566iksXrxYoyGeeOIJvPrqq3j00UexefNm/PznP8cbb7yhfZfjOHzuc5/D1Vdfjbe//e146qmn8OSTT+LBBx/Er3/9axw6dAgPPPAANmzYgJdffhmTk5MAgNdeew2JRAL9/f0lc81kMnjzzTdx1llnlR0H2TEY8bWvfQ0f//jH8cQTT+Bv//Zv8Y//+I8AlAXq8OHDeOKJJ/D0009jxYoVeP311y1frwaBQADPP/88br31VvzqV7/CE088geeffx7//d//jS984Qv4t3/7NwDAfffdh3379uFXv/oVHnroITzzzDN48cUXy8Y744wz8Pvf/x7XXnstXnzxRW2nFIvFACgLRTAYxBNPPIGf/exn+P73v4+dO3eWzOeZZ55BIBDQXtuyZQvuuusu/PSnP8UTTzyBWCyG73//+wCA2267Dddffz0ef/xx/OhHPypZ4D00HzzDfQzh1FNPxa5duzAxMQGO4/D666/jXe96l/Z+R0cHXnvtNSxYsAAAcMopp+DQoUMAgOeffx4bNmwAy7KIx+P4zW9+U+Kt33DDDVi3bh0uvvhiAMCzzz6Liy++GPF4HAzD4EMf+hCefvppdHZ24pRTTsGTTz4JAHjqqadw4YUXls01nU5DlmV0dHRor33mM5/Bhg0bsG7dOnz2s58t+85DDz2Eiy66CABw8skna3Pv6urCnj178PTTTyOXy+Ef/uEf8O53v9vy9Wpx0kknoaOjA+3t7eju7sZ73vMeAMCKFSswOjqqnZfLLrsMLMsiGo3iAx/4AJ5++umysVatWoWHHnoIhUIB1113HU477TR8/vOfx5EjR7RxNm7cCJ/Ph46ODpx33nklxvbss88uG/O5557D+vXr0d3dDQD467/+a+073d3dePjhh7Fnzx4sXrwYt99+e9XnwUP94XHcxxACgQDOPfdcPP744+js7MQZZ5wBhineAqIo4nvf+x5eeukliKKIdDqNZcuWAVC88UQioX02Go1q/3/qqafA83zJIjA+Po7f//73+OUvfwlA4Vs7OzsBABdddBE2b96MjRs34tlnn8W//du/YXh4GB//+McBAGvXrsW3vvUt+P1+jIyMaAvJD3/4QwDAI488goceeqjs+B555BH8/Oc/RzabRaFQ0CiINWvW4Otf/zp++tOf4tprr8X69etx8803W76uP85KQLxhcq7J34FAAJIkaeflO9/5Dv7lX/4FgEKdvPWtbzUdb/Xq1finf/onyLKMbdu24c4778TVV1+N//qv/8L4+Di+9KUvadcvn8/jggsu0L7b1tZWNt74+DiefPJJvPzyywAAWZbB8zwAYNOmTbj77rvxyU9+EuFwGF/5yldw3nnnVXUePNQfnuE+xnDhhRfizjvvREdHBz7ykY+UvPfrX/8aL774In7+85+jo6MDDz74IB599FEAijdO6A0AGBsbQzgcBgCccMIJuO666/DJT34Sp59+OhYuXKh5nB/72MfK5nDuuefi61//Op5//nnEYjEsX74cAMqCaCeffDKeeuopfOITn3A8rsOHD+OWW27BAw88gBNOOAEHDhwoMTzr16/H+vXrkUwmcdNNN+E//uM/cNVVV1m+Xi90d3fjE5/4BNavX2/7uddeew0LFy7EvHnz4PP5sGbNGlxzzTXaNevu7sbdd9+N1atXV/Tbf/VXf4Xrr7++7L22tjbcdNNNuOmmm/CHP/wBn//853HmmWeWLNAemgceVXKM4aSTTsLIyAh27tyJU089teS96elp9PX1oaOjA+Pj43j00UeRyWQAAO9973vx2GOPIZ/PI5vN4vLLL8eOHTsAAH19fejv78fHP/5x3HDDDZBlGevXr8cjjzyiff/+++/Hww8/DABoaWnBWWedhW984xslXqIRX/rSl/CjH/0IL7zwgvbaX/7yF/zrv/4rlixZUvLZdDqNaDSK448/HoIg4L777gOgcOUPPfSQxuUmEgksWbIEsixbvl5PrF+/Hg8++CAEQYAsy/jBD36A3/72t2Wfe+SRR/C1r30NqVQKgLJjeeyxxzQVy7p16/Cf//mfkGUZoiji1ltvxZtvvmn72+vWrcPTTz+N8fFxAMAzzzyDe+65B4Ig4Morr8TIyAgAhaZhGAY+n4/moXugCM/jPsbg8/mwfv165HI5+P2l6/b73vc+/PrXv8Y555yDRYsW4ZprrsFnP/tZ3H777bj66quxfft2vP/970ehUMCHPvQhnHzyyZpyAwD+z//5P3j22Wdx33334corr8SuXbtw2WWXoVAoYMmSJbj11lu1z1500UWW/DbB29/+dtx99924++678Y1vfAN+vx+tra342Mc+VrZbWLVqFc466yycf/75aG9vx4033ogtW7bgU5/6FH74wx/iuuuuw3nnnYdAIIBly5bh//7f/wtZlk1frycuv/xyHD58GBdffDEkScLatWs1ikiPG2+8Ed/97ndx2WWXIRAIwOfz4W1vexs2bdoEALjqqqvw9a9/HRs2bEChUMB73vMerFy50va3TzzxRHz2s5/Fxz72MYiiiI6ODnz7298Gy7L44Ac/iE984hOQZRkMw+BrX/saIpFIXc6Bh9rh8+pxe2gE/vKXv+Ab3/iGKVftwYMHe3hUiYcZhyRJ+MEPfoArr7yy0VPx4GFWwjPcHmYUb775JtavX4/29vaSDEQPHjy4h0eVePDgwcMsg+dxe/DgwcMsgyvDvXPnTpxzzjmaxMoMt99+Oz760Y9Sm5gHDx48eDCHoxwwm83im9/8ZklWnBG7d+/GH//4R7Asa/r+a6+9Vv0MPXjw4OEYxsknn1z2mqPhDgaDuOeee0yrwRHcdttt+NKXvoS77rqroh9vFgwMDJQVOZqt8I6l+TBXjgPwjmWmYeX0OhpuhmFK6lkYsXnzZrzzne/U6klYwaxsZ7OA47imnl8l8I6l+TBXjgPwjqVZUFPm5NTUFB599FH8+Mc/1ur9WqGZV7bZsPK6hXcszYe5chyAdywzDSuPuyZVycsvv4zR0VFcfvnl+MIXvoBt27aVpDV78ODBgwf6qMnjJr33AKU62/XXX48bbriBysQ8ePDgwYM5HA331q1bcdttt2FwcBAMw+DJJ5/EunXr0NfXp7Vt8uDBgwcPMwdHw71mzRrce++9jgP19fW5+pwHDx48eKgNXuakBw8ePMwyeIbbw6yHVJBxxY9fhiAVGj0VD7McKU7AH/aMN3oajvAMd424/9WDODyZbfQ0jmlsHZzG73aPY8dQqtFT8TDLce/LB/DX97zc6Gk4wjPcNUCWZVy3eQse+fORRk/lmMbOYcVgD07lGjwTD7MdHK80dW723ducN9wPvz6Idbe/UJexR1N5AEBOvdgeGoNDE8qOJ8WJDZ6Jh9mOPWNKj9Rmv5fmvOF+ctsQ9o5mcHSavjd2UDUYUzme+tge3GNwikM0GEAyJzR6Kh5mCE9sHQIn0HeY9oykAaDp76U5b7h3DKfQEQtiz0iG+tiHVG47k/c87kbiyFQOq+a1NL2X5IEePnPfa3hmYJjqmFJBxr6xDI7rjiHJeYa7YeAECYOTOZy1oqsuAcSD4zms7I3XzWA8v2MEUsFrUOSEo9M5rOiJI8PTvw5Hp3N4eW/zqwyOJRCjOpGhu9PdM5rG/NYwelpCSDe5EzCnDffRaQ49iRAWtkcwls5TH//gRBb98xPI5Olf5APjGXzyJ3/E9qEk9bHnEmRZxpFpDku7YuBF+gGlf//tPmz8t+ZXGRxLODiuOGG06Yytg9NYs7AVYTYATmzuXfScNtyjqTx6WsJoiwQxmaW/9TlEDHcdPL396s05kqK/4MwljGd4xEMM2iJB5OvwsI2qC36ztWbdcni6LhzvbAAJRicpe8XbjiRx4oJWhBg/8oKnKmkYRlIcelpCaIuymMzSDyAenMhiZW+8Lhf5iCptG016htsOR6ZymN8aRojxg6vDdSBeXa7JjOTFd7+EHzy/u9HTaAgOTmQRDzGYpuyM7RhKYfX8Fs/jbjRGknnVcAepX2ROkDCR5bG4I1YXT29c9fSydfDm5xLG0nl0t4QQYv11uQ5jaWXBr8eOrVqQ4+Sl5toFzBQOTGRxwoIE0pQpyu1DKaye1+J53I3GSCqPnkQY7XXwuA9P5rCwLYJoMIB8HbjVsTQPNuADV4ex5xImMgI6YkGEmEBdHrbxdB5Bxt9UWn2SP9Dsyod64ZC606VJFSU5ATlexLxEWPG4m2yHZcScNtyjqTy64wpVMkXZYzoylcOCNmWLXh/DnUdfe7SpDEYzYjLDoyMarMt1kGUZYxkefW2RpnqQiZpiqg7032zA/vEMVva2UL3eI0kO81rD8Pl8CLP1ccZoYk4b7pEUh+6EQpXQ9riHkxx6E2GE2ADydXiox9M8+tojTc+1OUGUCnhuO129rR4TWR7tsaD6sNE9V6m8CNbvQ1uUrQsNUy3GVcOdPcYW9cOTWdz+1A4MJ/NY2dtCdTEdms6jNxEGgLrFS2hiThtuRVUSQkuYoc6HjaSUC10vj3s8k1cM9yx/OF/eO4G/+X//q9UToY2JNK9SJfSvw3iaR1dLSN06N8+DPJnh0d0SaqpdwEzg31/ah7ue243LT11MnaIkjhgALzjZaExmlYc6GPBDlkFV5zuc5NDbEgLj96EgyxApF6UZT/NY2BZpOjVDpSAG+80jSUxleSy97jGq0rqJLI/2aFAJTlI2ruPpPDpVb76ZKKuJDI8FbZGm387Txq7hNH76N6filvefSJ2HHk4pOR8AvOBkozGVFdAWCcLn8yEapPvwkRXa5/MhxATAUzTcUkHGdE7A/NZIU3l61WA4xQFQKvdtV8uuHp6kVzdmMsMXg5OUvaSxdB6d8RAiTeaBjWd4LGgNz/p7o1IMTimCAADUd1gkHgYAoSa73maYs4abEyTIAMKscojRIEM1UWY8zaNTu9B0V+jJLI9EhEU8zMx6j3t4msOahQmMpvJaoa8R1ZjTwESWR0eMrQtVMpbm0RVXpIbNZCQn0sTjnt33RiUoFGQc0Rlu2h73dE5Aa4RVxvY87sZB8bZZ+Hw+AEA0FKCqiU5yxQtN22hMZXm0RdlZIUtywnAyj7csbMNwksNkRlH2JHP0rsNkRqVK6sVxx4NNdx0msqrhbnLjQhNjmTxiIQaRYAAA/WcumROQIM+z53E3DtM5AW1RVvs7FmSoVvFL5kS0hJVey0GGbvJHihPREmaVLXoTGYxqMJzisLavFcNJDlNqFiIt/bFUkJHkRLRG2Lqoe8YzKsfNNNd1mNCokpmZkyAVMHC0sTVzxlI8elpC2t+0F9NkTjR43M1zvc0wZw33VJZHWySo/R0NBqhSJUlOt0IzdCPc6byIlhCDMOuf9VTJVFbA6nktGE7mNd3xNKXiQNM5AS1hBkzAr0i4aHvcGR4d8RAiQX/TGe75Mxic/MUrB3HBnb+dkd+ywrTOIwYUj5uXCtQC3XqqJOTpuBuHqZyAVr3HHWKQpeRxi1IBebGAmH7bRnHbmuZExEOM6nE39w1kB1mWkcwJOL4njtF0HpNZAb2JEDX98YSafAMAjN8HmbK6R9n5MKrH3TzXYTydn1GPmyy09aiCWckcWnWG2+fzgQ3Qo0uMHHczLdRmmLOGezpbeqFpetwp1bAS/jxEmyrJi4iHmaaToVUKTiiACfiQCCvBw4PjGcxL0DM4k2ryDQBN3UPTU8poO5/moUoEqYAsL6ErHoJUBxmqGUimJu3615UgmROQCLMlr4Up8tx6wx1k/HUpEUwTc9ZwT+V4tEVKOW5aRlChSRjtb+pUibow1Ktw0kwhyQloUR+2npYQdgynMI+ijG1CDUwSsAEf1QcuzZEFtHkoqynVIfH7fQhTvu+sQAx2IzsMGT1uANTiGrxYgCAVEFV30ME6JdXRxNw13NnS4GQ0FKCWPZniRLSEdHwbZQObzitb9HoVTpoppDgBCTWA29OiGOwFFOt+KBru4nUIUtbTp/PKAtpMmZMpTtCC4sp9N3OGm+xYH33jCK7f/Je6/64eehUXQZiSTJPEq4o76IDncTcK0zkBrTpvjGbhGEU6pPe46XLcKU5QPO5ZsPLbYTonah53l6oIWNQepbbITeioEkANWFE8X9p1qEMdlGqhLOqEi50ZCmc8o8giieNz+1M78J+vHqr77+oxbXjmAFBLujJ687PhuXNluHfu3IlzzjkH9913X9l7r776KjZu3IiNGzfi2muvRaHQHAc8lRNKqBKaN3mSK+XbaGdOkoczxPghFOhFzmng0z/9X+wZTbv6bEqnvAkGlFstEWGpLXKTBqokqCoNaECWZWR4CbEQowarmuO+JjQaMJMet1KpkgQnGb+v7r9phBlVEqQUnDQqVmjeR/WCo+HOZrP45je/iXe9612m73/ta1/DnXfeifvvvx8cx+HFF1+kPslqMG2gSmje5Emu6PUA9fC4FW6VduScBp4ZGMbjW466+mySK2rdv7h+Oe7665OU7S0l73VcTXcnCAboedycUAAbUM5/M3ncJHANzJzHneJELGgLa4ZbaEADB1PDTcnAmnrcTRLTsIKj4Q4Gg7jnnnvQ09Nj+v6DDz6I3t5eAEB7ezvSaXfeWL0xleNLLgZNiU8ZVVIPjpt4VU20bUupiTOiy87zKd3OZElnDBe/dQFVad2kTg4I0FUDpPKC5tk2Uwp0mtPdG2z91Q+FgoycIKE7HkJaldOS3yy4vA9owMrjFigcf9Iw9pzwuBmGQTgctnw/kUgAAEZGRvCHP/wBZ555Jr3Z1QBSYIogzNIL9CU5sYwqoa4qIQGoOhRPqgSC7gYmzSjcqguSOVELThLQlNZNZAV0xOtDlegpiXrU+q4WaZ3HzQbqb2CygoQIG0BLmEVave7E855JpU09PW5Fali8T8nOrZkoSiMY5484Y3x8HJ/5zGdw4403or293fQzAwMDNH7KNSbSHIYP70VmRJH4jI+kMDyeM50Hx3EVze/AkTH0xhjtO+npSRzK+DAwQKex71gyg9HBgxjIDcMvS3hz+06Mx1nnL6LyY7HDK4cyuOW5YTz+8eMAALvGleM7NDTm6jf2D04gzPhKPjs0wmEymXb1fadjGZ5MY/LoIQxkhgAAYj6HXXv3I56rvXHDrvE8GFnEwMAABid5TKezVZ9Xmtdk76FJ5EUZAwMDEPM57KZ0vFYYz4oI+YHM9ASSeQnZjigyvIjWsB9/3jqAjigVE+KIiRSHkcP7IU4Uf4/nstiz7wC6xLGqxiTXZdfBSYjqOdVj65sDDeHz3aDms55Op/HpT38a//AP/4CzzjrL8nP9/f21/pRriFIBeWkfTn7LifCrJ36/eBRbJo+YzmNgYKCi+TFb3sDypR3o718EAFgwuAtSoYD+/lVU5i88chRv6V+JRR1RxCPD6FtyHJb3xF19t9JjscOLI3sADOO4FSsRYgIY3zUGYBCBcMzVb7A7tuD43hb09y/VXpMS0/D9Oe3q+07HkhYO4pS1/Zon1vaHFOYv6EP/anNarxJM7hlDV1sO/f39iIxlgN9NVH1eaV6TyP7t6I0w6O9fjtY/pDBvYR/6V9V+vFbYM5pGa2wMxy1egIGjScgBP6JBBu2xIOYvXobjut3dl7VAlmVkhH04ee0JCLMB7fX2/81g3oKF6O+fX9W45LqE9r6JhfEQ+vuP194LMQdw3PKViIVmZmGywmuvvWb6es1ywE2bNuGjH/0ozj777FqHogaFymA0ow0ofCCtoFiZqoRydJ/ouAH6BawqAek0PzxdbE4bZv3IuExk0tMNBLSyTPOihLwolW5xKcYDjOqNplGV5AWN46YZjLVCNq8oa2LBALJ5EVmhMOPa9pwgwa/2gtQjSImiNOZ8KGM3d/ak43KydetW3HbbbRgcHATDMHjyySexbt069PX14cwzz8TDDz+MAwcO4L//+78BAO973/vwkY98pO4Tt4NSFjVY8hrNCLw+sQQgPerojC3LMtJ5UVvpG1nwZjytZszli7UqOqJB1xF3IqfTg9YDP5kR0BYNakkTAGWOO6/juBscZ9BDH/8IMr6SGERdfi8vIhZk1Hr2EjK8X00Om7kAnhm/DdBbuKZyQpm9aCZRgBkcDfeaNWtw7733Wr6/detWqhOigSnT9FiKcsCcaKhURs+45gQJwYAfrKp7bmQbpVSeBKMUo5XlJbTHgq6PNcdLWhoxAS0FzkSGR2fM8LBR9EAzuiBgc3ncIuKhoja+7h43LyIWCmgdpLJCQKnI6K9NMjdhkHLawdJwMz4q0kSlkujs8rjnZOakscAUoBhXWg9fUpd2DCipt1S36AZvvlHSpExeBBvwaSqCDC+iowLDneXFMsNNy+M21ikBaMsBdbseRims3wwqg5SOwgky/pnxuEMMYiGlSFuGLyAeZmva3eweSeHt33zadV12s+cZIAtX7U6AQpWY3EtSc+yyzDAnDfdUji/jrMIUtdYpEzkgNRpGp+EmYzcqGSCTF9HTEtZSnXO8hPZo0PV5zPISImzppi5IScI2omvuqo3N0HmQgVK9dMDvA+On491Vg4PjWfzwxT3KvHTxD3YGPO5MXkIsyCDCKkXaMkJBo0qq3QluO6I0Zdih9iB1wlSunIMG6FFjZuPTlvjSxtw03FmhbOtDq2CTLMslhX6Usevocc9QWrPpXPIiehOhosedlxSP2+V5zJpQJWxA8RJr9V6PTnOYlyjNL6C1KAClHDegxkgaxHNvfv0wNj2+HVJBLvO4631vKFRJ0ePO8gUkwkxNv01iJ4cmsq4+b0xJJ6Cxw5Jl2dSjb/YKgXPScBsLTAH0uNVUXkSEDYAJFE8dzQQNo8GodVEoFOSqy9lmeQm9iaLHneVF1eOuwHCHSg13wO+D3+eDVGPW3XCSQ6/RcFOkSpREF4NyqEE8N0l8OjKVK3EaggF/3XcByv0YQETjuAtaHZ1qqQTSCWkk5S7vwZjZSKAkINV2/FleQsBfrlihGS+pB+ak4TbzuGlxq6b8OcXgVYozGu7aFoXrN2/Be/7p+aq+q3jcYS04meEldMRY1/PJ8SKiwfL4NxuonXYYmuYwr7WOhptTDBZBrXTY73aPYf9YpqrvjhJZZpIryZyciQBaJi8iGmK0nq0ZXpEDBmugSiazAua3hjHpsjGDdXCy9uO3omFmopxALZiThtu06Dol/bC+1yQBTblYihPKgpO1eHov7R5z7dnoIcsysryE7pYQ0qocMJsXXatKZFmpcRExeDIAnVTtoWS54WYDfuSpUiX0tPpX/PgVXGeoYb17JKV5n3ZI5gRE2AAOTWa1Tj9AkXaqJ4ikM6LuKjN8QasVX+01zORFLOqIuu6oU085oJl0mIztUSUzDOVilBekEQtyzVt0804c9DzudN4Y+KztBmoJOyo+TUGq4yUirFZcKKNSJW446rxYABvwI2CSMhys0eDIsox9oxks64yVvE6zHrfeswXo5AEYz8U53/0f/PNTOxy/l+RELO+JY/dIuizhqN6Ko4xKlfj9yoIxrtZYr8XjzvIS+toimHSxaAH2HnetC5fZ7pyM7XncMwyzC+3z+ah4xkmTQAlNj9uYbVjrTiHEVHeJCdceDwWQ5YuqkliIcVVqNpMvlwIS1OopjqbzYAK+kiYKQB04bj1VQiFIbOY0jCSdd0OpnIAVPXHsGcmUlBOuxOOs1mHRK4NioQDGsxJaI2xNMtWsIGFhe8S1x50ylFEmoOFxj2d4tMdMqBJPDjjzsFqhaXjGVh43rcCV0dNrVOZkVuWnIyyjdWXPqCoRN7sARVFi7u2zjA+CWP3OZ+9oxrRGRq2evB7KAlraiKNWWabedhLv3c21TXIijuuOYc9oumQHxbo0noJUwPE3/AZ/OTxV8Zw5QUJEXYAjwQDGMmJRVVLl+eB4CQvbIpjMutNxG1VcBDR2HMPTHOYlIqZjN0spXzPMXcNtEnCg4Rlb8ec0ddxlHncNNxAxFpV6XDlBQpj1IxosetxZ1Yt2EzDNCeVSQIJaOe69oxks64qVvU5fVUJHlknOvb4j+6gad3DTB5UTJCztimG3wXC7VT6QjkWv7J2oaN7kt8Pqri0WZDCek5CIsMo9ULXHLWJBW6Uet1mQu/brPZTkML+1vGx1s9fknnOGW5blGfe4afazTBtu0lq3bJrR5YsGQpAKjkXwOaGAMBtQDXfR444F3SVfmGm4CWr1jPeNpXFct4XhpvCwSQUZnCAhqgus1sJxc+ruIquTZQ4nOUTYgKaRt4ISJBaxpCMGWUbJos66rFUyllIM5F6dqiXJCVh63WMli4np3NX7AIB2PVsjtXPc3S1KfoAbPb+xOTcBjes9NM2h18xwB5q7YfCcM9w5QQLj92uRdz1oBJiGk3n0GjP2VA+SRkp0uY67tsQhslDptdzXPPgGzr/jfxy+JyHMFPW7gEqfhAKuvM8sL2pbbCNq5bj3j2extNPEcFPS3mZUmshYXbLaxTmnXgO9dz2W5rGkM+rYlIKXCmD8fsxvU4xLaas2d8ZlLJ1HkPFjLF3k04emOQDAzmH7jlWcIOkMt3JfJlSOu9rzwfGK1x7w+1yNoc8W1YPGDuvodM7U425k4psbzDnDPWVR1wAgHndthlvRD5dyYn4/vd6QZcFJ1g+uhnGzvJK6rff2fr9nHLtGnB/YEKvUXs7yktoRRDGObjTN2bwNxx2orardwfEslnRGy16nle1mVo42wgaq7vjCiUq2od67TudFzG8NI8PbG26OLyDM+rUWbZ2xotPgVlUxls5jVW9LCTUxnFQMt5McUS/pDGqUSaAmo5lVdzMtYcZx4SLVMuMmhptGksyQSQYu4MkBZxxWNAlARy5mxYnRSntPGnTcETZQdeYjoGxLO+PBEgPhpmckJxQQ0VElOV4JUvl8PlcGMqsLahmhcJPV7U5kWcaBiQyWmHnclDhupbBS6dxjIcaR1rACJ8ro0SUyAcU6ME5jZoVS719PEbEBdx7rWJrHit54ieEm2ZjTOfsAoUKVKGaC3DeKlrx6tVNWvZfiIcaR488Jkta02Qi2RjmgVJAxms6XZeACnhxwxmEVmARQU9IAwdB0eao1GZtGMSijjlvxeKszGIWCDF4qoD0WLDH+hACwo3byorJFVqgSERleqcsMKJ6O0wOT40XErDjuGh64yayAEBMo84gBepUUzSSfMRdGxgqcWEB7lEVBljVjkFaTmcSCbBtvIAsmALz5jfNx6dv7tPcq8bhX9rZojTGAoqrFqUJfXpAQMkmiqtYJkgoyRKmAEON3tRga1T161FqbZpqTNL7eCBoJezuGUvj8L/5U0xhWaGxfnjrAjiqpVeIzns7D50NZA1yAXmnXZE4oMUrRUMB1x5lP//SPOC4mgnTJyqk8dSxYpEqUIlnKw8JLBdNYAEC4TT+iKkWgpD6rGXsugmK2csAaOO5pixRlgF5AyWzXFg8FNCVIpcgJMqJBBrGQsggHmaDG25ItudXuJMuXc8wEbjn9FCdgcUcUWV6CICmJUYR+S+YcqBpR0jzu9791AaS8Uhiq2mtIqBefz4d4yJkqId2szFCrVzyek9DTYt4IncbufNdIylEEUC3mnMdtVZAGqH2F/tPBKaztay3pukJAY4UWpQIyauCGIBZkkHXp6T0zMIJn9xa5a6LsiOiUIVlegt+vZFRyvPW5IGoCJuAH4/djMitoHrcbg5Hl7aiS6jluOyqM1vY2yZX/Rq0ed9RADWTUILTTfcPZyCrdqiqyauJUW5TVshU59X6wi1WIUgFiQUZQpSk+eHIfbjq7V/ttt/VmprMCrv6vPyPFCSVB65awC6pEt3AZUausdCIrlgkNCGjcS0emcpjfWq4Rp4E5Z7incuXdLAhqrRD4PztHcdaKbtP3ai1iRWSMLWGmJDVaL8ezA1nZ9dXSiCIgGgwgJ5QajGgwgKxg/dDo1QSRYABj6bxmQNx4W1leLJHT6VFLVTdHw02BKpnOlvYUBaAaXefrIMsyrnnwjRI+mRNlRAwLaJpTyqU6JVgpmYs26hwXsQKygCcirOZhc4IEnw+2AVdOLGjesdlvuzVszwwM479fH8T/7BwroX6Uc+pA1eg8fiNq9YrtPe7aJb47htJY2VufZspzznDbPtg1RqFfPzSJU5Z2mL5Xi8e9fSiJZdf/BqPpfFlXF30CjB1I8DGdLx4feWCjwYAWGCMdTZyCnpxQ0BIvogbD7So4aeNxBwN+CFVeB6vazGRedDxu0YQqcRecPDSRw4OvHcYLO0a01/Kqx61fhInsM+yQdZsTrD3OSjzuaDCAFp3Hz4lKUwxbw23z25V4u28eTSLg92HPaFprzAAA8TCDtANVkhcL9sdfw/WezEp187hHkhxe2DGCdx7XWfUYdpibhtsqOFkDD50XJewaTuPEBQnzsWvQW285PA0AeHXfRBl/Gwsxrjju6ZyA7pYQknlJ875J9mKEZbQtMXlwwgZ528HxbElHEk4sBqUiQYXfjeq6izttk0ldEzPUynHXa2F+80gSrx2YNP0Nt1TJgQklyWXvaDHZJS/JiLABRNjiIpzWqBL7LFRbqsTl8ZLyunpDmeMLaIuytrtEfdakEaEKAsz7xzI4e2U3DoxntcYMADmn9vd2XpQs6+3UTJXkRHS1WBjuGsf+4Yt78f63LTDN8KWBOWe4bYOTNQSvprKKp2e1+odZf9UdUoim9re7xso87hDjV7hGh5toOiegMxZEmPFrSoEsL6rKkKJ+Pa2jSvT85sf+45WSpJzSxIsARpJ5LUjkRs1glznpNuPPDLYxjBq9pL++52Vc9q+/V7360kUnHnL2DgFoSS4jKU57LS/KmrSS7HIyquTQSUaqBCfNH1O3HjepMaOnJojHbcdxO3rcLs/1kWkO7zq+EwfGM9pcAKg7AGc5olUAPchUv3MDgKwgW1bPrJVW3TI4hXNP6K36+06Yc4bbVsddg8dtNy5Qm8ed5ET0tITw0q6xMo/b5/MpkkAHqSGhEFrDfo1fJV3W9ckjxGBEDNy50ZvU63ejQQbDKU6r0ObmoXXiZutxHUJMbfW4ybmfyvLmVIkLymo8zaPPUPkuL8oIsQEtmQkoZgM61bnJ21Elrj1uoptmNRUHJ0hoj7LOdJnlNXS/+B6dzuHUZR04PJlDNl+UlbpZDPOikghmhlq94pxQsJca1rAoDCU5LKhTYBKYg4bbSVVSL8Ndi8ed4gScdlwncoJkmsUVDQaQddhSkuNOhANFwy0okrwwG0BOVZBkeHccd16VEpLfH07mtea5bjzunGDe/QaoreVW1kYfThaUaksPiOqcXt03UabVj7vI8gMUg7ykM4pxveGWlCCfvnwAiTU4BcHyYsGaKnBxHUitE5KpqHHcgoQ2E45772haK0pFJKFmcKsqyanJW6vmtWA8k0dKJyuNh1mkHOinvGBz/GonpWold1mhUJZoRVDL7k2WZbU0hnngkwbmnOFWWhGVd7QAatv+mLUsKxm7Fo87J+I0NYhhVq7UDb9KFpbWUEAzGiRAGAkWPW5ClUSCTMlDS+598hBwYilVMpriSrqLO9fjts+crJYqydtsnQN+HwI1dGOfzPJ413GdSHJK9To9WiMskjnBcVFQmgRES9py5VVViT7QnMlLCsftcE/yYsE0QQRw5xXmxYIi6Qz4SzxcTlCSgoze/rrbX8RV9/9Z+4ydt+/GCTo6ncO81jBCTACJMIvDE1mdxx2oKTjp8/kUJ6BQfQ0Zs0QuoDZVSV4swAdY3v80MOcMd72CV9M5wTIRAKiNhknnRcxrDeHbl6zBhW+ZV/Z+IsI6piaT0pctIT+S6mdzKl0RYYt8djZPWlH5SzzuNCciyPh1W+kiVRJhGYylea15riuPm5csvZmaDLdYsNw6A9VXHpQKSqu1NQuV4HOPIWjFBvwIswHHBTTLi+hrjxg8blnzuLOCEjzOqpmoYYcFn5cKCAas6QqnzEt9rCEWYpDmSz1uM5pGa5zhoGhxc56HksVaID2JMPaOZTSPO+aCfrILTpJ5VPtM50TZMoBey7hWZWhpYk4Z7kJByQq0MrC1NCVwpkqqrzyYFxVa4op3LjGlF9qjrGMxIGVLHUAs6EeSKz54Gsdt2KLreW9RKkCSZXREizVN9IEpchMWu4v7nDluQUSUtW6kUC03Wa8HmWT0XXXOSvzmi+821S63RlitxocVsnkJva1h5NQsRWXOaolcllE7pSvn1u/3OS74dgsV8TjtzmUmX9TTl6hKBGs5IFkHOIueoYD7xXdERxn0JkLYN5bRPG43qhK74CRQm4HN8gWN/jOiltpDxgqf9cCcMtwp9SZlTArSALVVE3MOTlZ/oXkHL7IjGnQsOk+21LFg0eMmVEk4WBqcjIeU17RsSrVaW0zXpkzPbxKlS0sFqpIcLyEctN7iV9sBhyxQVqg2CYc0iYiFGJxgIflsizrvfLK8QoG0RopZinlR1jWlkLQkKMC5CQcvFrTMRTM4XQtOkBAuUXEYqZLidwkNxAZ82net7ku3qpLhJKdppXtaQtg7mtG8XDfaeKeFmg1U7wRkxYKlx12T4ebMqxnShCvDvXPnTpxzzjm47777yt57/fXXsXHjRlxyySX4wQ9+QH2ClcCsOJAetdIZZn3vCMJs9d118qL1dhgA2qJBR0+PVwv3xNki3ZHjRY0q0Qy31pKsuPvIqny0UvSHpEIXDWSH2pMvUYGqJG/Hj9ZQZMouWAVUT4fZJQwRtEVdeNxqVcTWSFHBQeSARMmT4vSG2zk4acVxA6rhcvg+CTLrOe68SXCSxAYkLc5hrypxomkAlATpehPhks5IbopM5QUHaoypzgmQCjIESbYvJ1Ct2CAvWAbmacHRcGezWXzzm9/Eu971LtP3r7vuOnzve9/DL3/5Szz//PM4ePAg9Um6xVTWugARQB7q6ukMuxtI8Zzqw9t2xFjHjtjEM4sFA5qOW0+VFHXcikcY1GV6Kv0l1QxL4nHrgpPdKt9LHkDFq3Xu8m6XOFE9x+18HapZnLO8ZEntELRFgpjK2V8HIndr0cUl8lJBC07meFGVZBY9bv2Cb/S+XVFDNudSb/jjYUZTceQESUnA0cU5CJ1G/tUri4xwGxgcTnHa/dOj3j/dceXveNA56G4XnASIJLDyZzqdFxFifKaUGFBb+QSnOdOAo+EOBoO455570NPTU/beoUOH0Nraivnz58Pv9+Pss8/GSy+9VJeJuoEjnVFL5qSDpxdy4XGTdlEk4YaAtzFygOJxuzHcbMBnQpUwJTK0jF6Gpi405HPRIKPJDvO64ORbF7UBKBpwJ4+7UJAhFKy3+LVkvLmiShyu8eY/HcZH//2VktdyNvXDCVrdeNykLkiY0a6DPgGnjCrRtdPjxQJWf+0JbDsyrY1npyoBnOuV6L9fqipRE3B096zSPcqneeF2ckDlt51jHSPJYhnk49Va4vPUevYKNSc5lhe2p0qqe6ZzvHVWKFDcuVUjLXWit2jAcXSGYRAOm+sRR0ZG0NFRrN3R2dmJsbExerOrEE6Gu5bMSV6yf4DceNx/OaQ8kH/cP1Hyel6UbMdujwYxmbE3GIJUQFALTparSvRyQGPGXk6QEDN43MoDoxiy+a0R7N90kfZbTsoNRQnht/RmiP62GjhRB248pd9sGcJvd42VPJROBgIA2lyoe0j1u9YIqwWJ85KsZrAqEsxUicddvCcPTSolU7cfLZYecFrUgw49SfXHVarjLiARYSAWZI0ayQkSOmJBLfbhZpF0uo56quStfW349JnLsLK3BQDUypP2zSCc5hCqoEqhHopxNb8/ydx8Pp+rpiNmYzvdS7WiJiKGZUuNpCzLlg/rwMBALT/lCtv3JiHn85a/dXSYw1QyXfY+x3GO8xubnMbokIiBgZTp++MjaYyMZ2zH+dNu5buvvrkfx7NFryqbF3Bg7x5Mhc1v0NRYDofHpmzHHh2fxFiQw4KojNEp5RhHJ6cxNgwcEseQzvEYGBjA+HQao0cOY2KKx+h4DgMDA9g+mEVB4CDkROw5IGAglFTntBtjJgHGkaE0xifLz6M237wE1m99zUeG0piw+T6B2XVJZzkc2r8P0oT5Ai3yHHbt2YtQ2jprbXhCOfe//9NWdESVR2DPYBZCPmc7p3xqCkMjEgYGrI13MpvHkQP7UODS2LH3EAaC0+CEAg7t34ORpIjxqRR27D0IWf2tqYkpTHMSBgYKeP2IYri37DmIE6LKvTIxncTwUR8GAlOmv1cQBezYtQfCuHnNjb0HM+BzWeW3OAlTGQ5vvvkmOEHCvt07EQz48MbWNxFh/dg7kUc0UMCoWMC2N9/E4NAE2iMBDAwU6Qz9NfHJBby5fQc6o+ZmRJZlDE3nMHlkP7hR5T667Hgf9uzaoX0mzACvbxlAW8T83h+dmMLosIiBgaTp+yKfd7zeZjg4xYPx2dslxg9s2TaAiM2uwwz7D6WQy9jfS7WiJsPd09OD8fFx7e+xsTFTSgUA+kl1/zri+eHdWBISLH+Lj08hsCVT9v7AwIDj/EKvpHHcksXo7zevP3CwMIQ/jR22Hef54d0IMuPwRRIln5PkA3jLCastI9xoS+I/3njdduzw6zksXTwP8fwYeCjH6P+fKaw6finWLGyF8PAg+vv7If16CGtWL4fvSBI7U0Po7+/HfvEouo/KWNAWQaItgv7+4yAW9mPtif2m3u1heRivDB+0nM9IkkM0NGT5/qA8jJeHrL9PYHZdCr5BnLh6pbbdNqLtpWnMX7gY/Su6LMfN/mZY2Yl096F/cTsA4FBhCB2DBds5rUofwh/3T9h+RigcxNoTV+N3o3sQDTPo718OXtqHtSf2IzGWgfxGGi0d3egrKNdo0cQ+SGPK/3fzRwAMwR9p036D/e00li9bgv7l5seTeHYcCxcvRb9KZxmxRziCzrGj6O/vR16UwD10CMetWAWWOYATTzgB0eBhLDluOTrjIWQPTKI9kcFQJolly1citnM7FvfE0d+/VBtPf00ioaNYsux4LOoo7/8JKElrIeYQ3r72RMvzlYgexYIly0xb0QFA8OUUjl+6GP2rzZ+7tt85X28zFI5MI8iM2F7LSPAQlh2/Au0x84Q+K7yePIhuboqKzXvttddMX6/Jn583bx5EUcSRI0cgSRKef/55nHXWWbUMWRPccdzVBSeduEY3Ou6JNI+VvXGMGaR9Ttv/jlgQEw5UCeHV4kF/Sc3lCBtAmPEjJ0hq41VJV8C/yHFHg4zWKYdw1KzFVtJNQMwugMjWGPiphToAimnp+uxGJyoMUDlupx6Nav1ohSpRMi3zkoww49eCk0qChnKf6tvpTWV5+H0oUVo4zcuJttJnmpJ/k5yg8bthNqB1wyH3C6mp4kQfObWKG05x6LYom0oQcwhQ5oWCZYAUqD7QzYsFsH5rqgSovkQG7xBApwFHj3vr1q247bbbMDg4CIZh8OSTT2LdunXo6+vDueeeixtuuAGf+9zn4PP58P73vx/z58+v64TtMJ0VsKTDuoxiLZmTTobbjZohw0tY0hHDmK4FligpARDG5iZqjbCYzvG2VJQyPx+iPj9SnKBm5ymBMtLFhpcKWlnNEsOtBuaioQCmc4ImLbTjqO0laJItL1lLBxwnVYmba5zmRLy1r61EG+8moNQWYTFtE5wsqBKzYMCPRITBwYkMBEmG36dwpqREb4oT0KUqK/Tt9KayAha0RbTsRu14HYJzdsdL4g0E8TCDsRSvqR70OvKsPibCS7Yp725+eyorlFW7NCKuk6Cawc31rs64Wjsm2tguAt0//u1evG1RW0mdfuM5rwccDfeaNWtw7733Wr7/jne8Aw8//DDNOVUNu5KugKL8qNZwOz1AYda5WXCOF7GoI4odw7rgk9r30cpIkrH9Pp9WNMoMgpoaHZB9iLBKkFHfbSTM+pHNS0p9ZzZQ0txYkbAFEGUDODrF2dYDAZxrMXMutNbVBJRktdmuU0KK3YNMUtsXtkdKlDpOux5A1dPbyAHJLsXn8yERVrrN5HQ1rVvV4GYyJ2o1afQLaIZXioyVeNxugrF2htvw/XiIwWg6r7svijvFnJqsE1FL/jouGg7leVOc4Jj67aTlrlfCFS+5NNw2uzdZlvGtxwZw4VvmlRpuF/dSrZhTmZOTWR7tMXsdt9PqPJLkcMam58rUA04PtptmwRlewuKOqFazGVC2gm4ucquDokF/syTU5I8kJ6BFLVsZCQYwnskjoqZa6w2cXg6Yc/PAOsj53Hy/Go9bLMjw+3yWmbGAsyEj3eq74qES+snNw+aUgEO8bQAaVcIJkqZeUBZgpVY3MWillJWIXhPDbb97qexatIQZjKXyGv0Q0rXc43Qed9aFx+1E0+gpISvo+3CagRMcdOxV7qIVqsT+MyHGvh3hpHovGO8Jz3BXiMksjw6bQEKIdb7Ifzo4hcGpHN44NFXyOqEPLMdmnDnuHC9hQVsYyZyg6xHpTjrkZDT0XGgirHw2ky+m3kZYpRlCkVstlQNGdanxjoWcHLLVnDz2Sorwl4zrQmblxLuSXo+KxLKUKrGbM6DWKrGpEMiLBbC6xXM6JyDLSwgxRc+uPRrEwYmsdh30iVCZvITullBJ/Q4nh8Gp76KVx0302WHd75N6LUQ+WquGOsXZF2YDFC13rR531Ry3K4/beuxxrWlGvuR1pznTwBwz3PacmhuP+7CqpT0wnil53ckYhVlnHXeGV0uqsgGNx3RK6SVw5XEHinrdoWQOETagNR6OBhkcmS56emHWJHOSDYDTglK1eHlOhr86jjsvFNupWY7tsCik1cWsI8ZiIltZcDLMBuADLK+zIBXABsjiqSTgpDkRMd25aI2wODSR03ncRfoukxfRkwiVdIWptVaJ0XC3hBmMpvKaJ61QfKULOEnY4oSC7fl20nEnXXjcTiWL3TgRVXncknNw0qm20USGx4qeeFlCnRvarVbMGcMtyzKmsrxtynuIUVJ07bKhJjI8GL8Po+lSLtM5Acc5czKnqjdawvo6FpKrQEZrJGhvuCUlOAko3t7gFFfy0CQiDAYnc1otF73ByOalYqF/QXLFUTsHJ52oEmeOO8OX/oYbj9vJmyd1QhJhFimueD7zLrPd7ApN6Y0sScBJcgJiOi18ezQIXiqgLaL3uAnHLaKnJVwSrHMqQOamVol+EY6HGIyl9Ya7NDgZ1pUBVlK3HWIVNr+ddMFxx0PFrkCm87dJuwdq2705edxOmdaTWQFLOmPgxUJJQ283TkCtmDOGO5UXEWICtp6iz+cD67e/GONpHit7WzBq2P44eT5uPG6t27auvKbbbVWrg6KB1xWqSoQZHJnKlVQoa42wODyZteBWlRrdYZXbdCe5qz7bzQ3H/frBSXzwP/fjyFTOMK79LevkgZGWYfFwqafnlpdMhK0Nt6B7YFvCSuOF6ZyAuN5wqzEYUrfDeB16WkIl83JaUCqpVQKoqhI9VcIGtLR3TlUXhVlFPpoXatt52ZVYJnAVnHTyuB3upd/tHsO+sdIdtCuqxGFRyOSV4+tJhDCSLNoLXiwg1OiU99mCqYx9gSkCJw40lRewvCdeEkAE3BQ3UjxuO28+yyttm5Q2WKSOhTuqJBFhtFR2MxiDk4OTuRJvJxFmcXgyp1X4K1GVaPpdvZqg+qCUU10XN4b7f/dPAgBe2l0soeA0L8CZOsio5VtbwmxJ9xW3acpRXYlcI/RKhSDjBxvwY2iaK2m1RtRDpFZJmC0WPsvkRXTEgmrlOmVn6CQtcyUHLOG4WXVO6u/rAnBaiQRt5+VQq8ThXKc40bZaJwDEgtbNKWRZdl64XHjcV/z4FXzhF38qec2VjtvBCSAqnI5YqEyhVG8d95wx3BMOgUkCvW7WDFlV+VGpx+2mbVa2BqokFrTfUvIl3h6DwalcCVXSGmHV14rFjYinx5EWZ6y74CTrcEPT0FoPTuXQHglgl0466SYeEGTsGwbn1aqH8VBpD0lesq8XQxC26dUpiHLJGMouJ1ficZ+5vAvHdRdzDYKBQInHHQ8xWjEqYrT8NgbG0eM2LKItYQZD05y2G9NTJSQ4SY7RMTDoFE/gBG2BsIKdx03j+IkjZdwluZEDGis3GkESlhJhRqtLAygJOA0vMjVbMJnlLXtN6uHkcedMJHui+nk7GRqgeC9WF7qg6ocjatNWUl5T4TCdqRJjV3YjSjzuMIsjUwaPW6NKVG5VDdTKcnFe5DecPWb74KIjVeKiA85wksNJ8yPYOZw2jOukKnFoBaYaA/010L/uBKJxNh1bF5wElF3SocksojrD/denLsZzXz67OF+d0onsyEgCjBuu1KkphVG11BJWkoCKHr/O4xaKC3heLDh63EGH6+imvKldFxxX1JiLYDSgxK70u2EqHre6Q0mo/Uj1Y3sct0tMZnh0uKBKFI/b2gDmBAnzWsMlKzRJknFCyIbn5tRVOOD3oSVkoErcbtFt+vMJOg+iLcri6DSnBcAAxfsryMUuNn6/T+seolcTuFGVkAQaqyL6TsfkJjiZzovo7w5h94jecDtTJU4ZrGQ3EWL8KBRkbaF1+7DpKy0aofCmOsMdLve4jdArnTJ5CTFShleQXM3JqR610fsjBjuu7byKC5Fex+3G42YdgpNuKAO7LjjKzq02amwyI2BhWwQBn08rBQHQkQOShU7xuEvthWe4XWIyK7gqBqOvDWGGLK9oaUnNDsB9koydllvPobaEGR1V4m5speSqtTevT/6Y16pUSpuvK8SUiCgPareuCS7pvpLTqQmUoJS94fX5FKNvVUSfcwhqMX4fJF05UTOk8yIWtwUxmsprOx4318Fpe0uCuD6fD/FwMd3a7cNmt/MRDHx0a4R0NbeJjaiyTKkggxP1CTCi62Cs3SJoNJ4k1Z7svMKsXwtOaq3u1KbIBVm2NW5OGbCkl6od7BoGO92HgLNXTCjUea1hHE0WA93uqBL73RspEUCyZLWxXe7easHcMdwZ3rEuAuDMcefUbaRea+32obaTDwmSrHlj+sCY+6AYY8mtGutfL2pXDPeCtmKpy/mqMV+oey2knguiJiDzS+dFZy7ZZovq5Gn5fD5HTynNiUiEAmiPsRhW4w1uJHthh4bQ+rnpGwu4pkpsPG7BcJ90t4SQyovotChZChTPY5ZX+qX6/T6NjnHjcTsGJ3VqI6DYxWi+pmop1XGHWSURayqr1DOxK8XgVCyMcxGTiIesg5NuPHan459UJcLz2yI4OlXUW+cFyR1V4tDPMxIMqJnKpdJSN/RnLZgzhnsiy7v0uJ23P8YOJm6Na9jG4xYLBTCBoqKgSJU4b/8BqAEr8xvcuLAs64rhHUvbcaau1GX/vAQA4ERdI1xyLgjHDSge5VSWd5wTa+Pp5QX77iIA2eI7qD9YPxa2RTRJYF6051wBshjZFS0qGuh4iEEqX6Ss3FIlnI3Hrffi+tQFtDtmHaAjBfuTnIioSmOQ4KSbxcSpNovR+C/qUOa0tEspxaoPTpJgW4QNYCon1Mwvu2lOYRecdNq5Ac7GdTKjeNxd8WBJ3MpVrRKHhL1cSXDSfdIUDcwZwz2V5dHumuO22/6IulVUp/yo1eMW9R63jipxsR0E7LfoRr7O5/Phwc+cjp6WIlXSGmWxf9NFJQFcwm8STwtARQ+tlcfsxuNgAz5bfjSVFxFhfVjQFsHgZE47TqcH2cnj1ie0GPX0bqkSK487b+C4367W+p7fYn9fhhg/JjO8JhsscszO5UEdpZkGpyMaZLDjWxuwvEfpQqN3NvTBycms4BhYdNo12TWMJiC7niwv4qaHtxi6EtUenJzKCmiLsGgzJLDlXQQnncpAE4pRCU5WnhNQC+aM4Z7I8OhwrSqx5opJV+yWkp6B7lbQMGNdIVCpHFekSjRViUsaxk4OaNyiu0WIUTrCs2rQFCAet+DscTtRJTVws7Isax53T0tY85TcbJ31ZUrN5yaVeNzpfGWUVdiWKimVA56+vAvbv7lBO7dWCDJ+TGR4rfIjaXHm1uOupFYJgJJrW6Iq4UmQ2o/pLO8iS9U+a9OpQBSgXAOxIOPJbUO47+WDODKtozNc7LCclC05QUI0xJTV+qGRgFMMTrIlHrdbwUEtmDOGW+nwXhvHzanBFL9fLctZIQ9tVyFQlIo1t1t0CTicC1oBcKBKqlzhQ4wfU1leo0kAncftwkBaPTBudih2STg5Qfl+wO9DZzyIMbX8gBvNe8iVx60cbzzMGgy3C1mmnY7bJFnGTbfvEOPHZJZHVPO4/Yoc0CXH7dj/02YMfc0ajeNmXHrcrrx9+zF8Ph96E2H8dpeSaHVAl+Ho5vvBQMChxLCyeLRF2ZKSvIrhth3asYCXpuOOMCVyQM9wV4CJjLsEHDtVCYmqA6XG1W1aup2qRF+AyNi01a2O2zI4WSWnFmT8mM4JpYZb47irDwrxOlrI8rdt+MM0JyKulqPtige1KmxOdTsA1RDZUGF5XTqyPgmnElWJ3XVwOm4zhJgAJjJ8SSVHLRHKDcdrxzM7eL1GjzuiBienc86GWwlOWu+a3Fa+7E2E8Ds1Q1bfYYiGqoRoyVsjBo/bTZEpxt4JKHLcbHkCjme4nSHLshY9doKdx01uXAAlonq3Hq1eWmWEPnClNxhuAjiAwk1ayQEVfrbyKHaIUWgRslgBqsftgiqx4zdFm7Znbr6fzouIh5Tf74yFSqkSRx13wPIaAOUcd4qyqqQaw02oEtJzNKK1DnOh43YIzrnxuDlRUlLsodAfpL6IU4Eo28VbKoD122c9EvQkwhhO5nHC/ISBh3bWcbuha8KMH23RUo6bRgccheP2a3VptLE9Hbc7ZHgJbMDveltqlRLNCcVgkJEqcavjtloUxIKsZV7q5YBOxeoJlGw2uy268wNSPl9lix42UiU18pui5OxxO1ElZE6d8SDGM4QqcanjtvW4SznuTL6ya2zHcVf7wIYYP8YzPOLBosfNCZIrj9XJcDktSKRZQDavdEby+XyaxjvhUJLVrhOSU4VJPTrVnfLJS9pLvGK3HredV8yp91J7lC2pJ0Ijc5Jcb30dITddmmigpi7vzQK3Gm7APnNSn8RC6hYDFQQnWevAmN7j1qdbk9oZruYtFkz7TlbNcbN+xePW0Q+xkFJ3wY2iwMrT46UCGH/1Bkd/PF3xEMYJx+0i2BW2WeDI2Hod98FMVptzrcoh5YGtbgEdT+fR165I9KLBAIaSAvKC85bbSd6qdxjMQFruZQUl3R4oGmySZWkF+wC1+4a5l5y0EGlORHdLqEQa6GY36qqNHutHWyRYRpU4XSvHcysp5zbCBiBIstI/Vn3PqTxGrZgTHrdTyzI97AyOWChoCoBEpBgp5iXn6DZgHxjTJ+CEGL9a+cw5S5Eg4PeB8ZtH0Kv39AJI5kqpErI9jtewTRZdaGTtSoLqPZZOVX9b5EydU97tyuvmdQkpcSNV4lKrb73wV0eVRIOKk0CoEtKJyJ3HbW+49EFxMxBngxRAI68BcJX8YikJdWg8osdJi9vx3Y+8rYyGclPrJBiwz4Qm2Zut0dKyyG6pEjsnQJBksH6lx2gsGEAmL7kOcteKOWG4JyrwuO3oDEHS0xmlWmt3D7W10RClgvYA+Xw+Tb/KCe48bm3uJsay2q1ZiPFjyhCcbHHpbdmlOwsuqBI7OaB+IYoGGQT8PmRcFL8Ciinvdu3FtCqKIUbrNuP2HIZYPzibxbmaBTQaDGAkldd4/SgpMkVBDigWCg6GOwBOLXtA7gOyoyvYlCjWfttWy1/ZuTDmKrhqnME4cdyK8W8JMcgKkrbQ0JADKkl1RfozyQkzouEG5ojhduo1qYfd9kcqyBrvVVp61W2RKTseunTLSsavRDpkxd/WKgcMs+Uet1Ngys5gCFIxS9Ty+7aqlNLj6YwHMZbKu+K4mYAffp91eV39+SbNFAoFGULBpeG2odqqVZXEQgxGknlNSaNvHeao7LBZQAsFGQUZtjpy0uWd1Ckn+If1K3DFO5c4/La10XST9WgE4fYrGcOpXgqJWykSX0YLULpLwAnYG25J1s4tUYrNRGASmCOGeyJj32tSD7vglagzOHo5oFsdt9023ajxJckflXnc5ls3YzlRtwipmm29x00WQDf8prWqRHY0gnbfN3qaHbEQxjO8awWOXaEp/aJAdj1u6j4T2GVm8lKxWXAliAaV7X4spMucdNGsF7BfAAVV3WNXb4QYywxfTLkHgKvPXYn++QnL75HftryGVRgwY66CG6/dacfB6ZyuVr1SjELKu149RZ7nmQhMAnPEcI+n81pk2glBm8xJoVDc4ifCTMUp7/o2UEboa5UAysKQ5AQl6cflltLKaFTrcQcD/jI5IDHcpIqc5XcdPW43VIm7h74zpnRkd1u8x864munpSQacG9h53IJOI14JCLcd1+SAurKqDsdrt4NU+G37+QT8Po1jj1ZYGMmO7hIl2dGjNULh9ovH4iYY7VRkKq+rKZ6IlCrFalWV6M8v6WpVSVC2FswRw82jq8Xe0BA4etw6qqQ0c9Kd1NDamJU+RKRORiVBHCsNuttEByMiQSXxQ+9xn7ykHX996iJ3VelsDYaL4KRNyrz+9ztiQUwQw+1652NXekCZW0y/63F5DRwXBaZyVUmrWjedtPlyW14XcAoSy46UFaDci8PJfAlV4gZOAWo3v61H1FDAy11wUrkPrWIa+jESOr21Igt10wHH/PhkWYZYkHVKsSL16XncLjGeqdTjdhucrKxynJ2qRJRkBBm9x62kW1ficVtx6NVuz0iiRUy3Re6Kh/B/L13rPBcbj9lN7RS7GhNGaqozFlSpEjrXgVxjQpVU6nG77YDjFqRGeq9aFIxQBm4bCdhTJc7ziYcYDCc51+eAwI7u0u9e3SISDCArGKgSh+tNGoJYef56KlKvt3YTnLQz3GJB4bf1PUQJVVLvdHdgjhju0Yo8brskmeLFjLABiAVFTF+Jp2cnFdN73CR7shKP2+pGqpYqIQqSVoeGrmawC0wJkr2SASAtt9xy3EFMZPKuts6A/a5K0G3hY2oxp2ze3biAA8ftItXfDD3qvduTCGm/wQkFV0ksdioaqeC88wGKHnfMIa5h9tvWuy7ne8AIYx0YtwFOu92f/tklDQ9IYw43xb94K+rTsKskO2hPVVIBxtN5dMXcGW67OtB6zkrJICO8lTtpk90KbVRaEH7Vqa+fm/Gr1Q8TTrUaw22b8i7JjkE6OzWEkePuUD1u15SVQ6yBPLB+ld8dy+Rde5uM34eCLGsPvx7VVmk8/fguPPz5MzTPsMTjdqGiCbPm3ZHcLKCAsoCPpLgSyswN7KgSo4rKDYx1YNw6THa7Dv1iT3Iz3KrEbHfnht2M5og1k477jjvuwMaNG3HppZdiy5YtJe/dd999+PCHP4yNGzfiW9/6liXXVC/IsoyxdB6dcZeqEpsau2bGVbkY7ro2h2yaBRuVFnEtOFmJx21exKraVZ7UdqnO43auU2H7fbvMSxM54ESGV2uYOHuFYRuPWzIYlHiIwVgq75rj9vl8lnp6s+qAbhBk/Hjbojbtb1Lz260R0Dfm0EN0aTxbwgwGJ3Max+4Wdk2j3dSrMaI8AcddoM9O3aLPKWiNsJjOudda6xtqG2GMH5Q6Yk1guF9++WVs3boV999/PzZt2oRNmzZp76XTafz4xz/GL37xC9x///3Ys2cP/vznP9dzvmXI8hJ88Lne5tnJp8SCXGJwWkJKwEFfCtQO9o0UDItCiMFURoBUkCvguM3Hz1fp6S3tigEAjuuOVfxdxRuxUBQUnINi9t5aabutjlgIExkembyoSebsYOdxCwaDEg8zGE2797gB69IG1eq4y8Zn3HVZJ9Ani+lhVDLZfX88w5c0l3YDJ1WJk6LFiKih5rybRgxkHtbPdHHXQbpauZUqkpwA0aQ3qmhBfVYSL6kFjrN/5ZVXsH79egDAypUrMTIyglxO6UjCsixYlkU6nYYoisjlcmhra6vrhI0YT/OuvW3AKXBl5nELriPFttxqwagqYXE0ySEeYmx1tuXj0wtOdsaC+OZfrcHijmjF37VaAAtqE2CnLbpdcNPoEXVEgxhP82pbOecF2l45JJc9cGMp3vXiqYxv43FT4Df9fh9CjFJy143HrU8WK52P7LjzAYq1SdxU19TDafGtVFUSZpVdmNak2y1VYqG3NiYgETlgJc+L1X0uFEobKSset6DW6G4Cjnt0dBQdHR3a3x0dHRgbU2rnhkIhfP7zn8f555+Pc889F29729uwbNmy+s3WbH7pPDodNMd62CYsGFK1iSTQPcdtQ5UYBP/xEIOjUzlXW//S8ekZDJ/Ph4+etsT1wqGHFcftJukDsA9ulskB40EMTuUQYgKOASXAutCUUcIFqMXE0vmKtrdWDTN4yb4reiXQyuu69ritqBJ3ZVUBVPQcAfZSPOPu1Q0UGqpYGtltP1are5Fca3IvEjlgJVprq12u0cmLq7tzfemAesLRarBs6Sqsr06XTqfxox/9CI8//jji8Tg++clP4s0338QJJ5xQNs7AwAClKZfijYMZhGXe9fiD0zxSWa7k8xyn/D14dArJnKS9J+cz2LnvIKaSKQwPHsZAYdx27OG0gHQubzqXoZFxtEcCGBhQPKPJkRwOTWQwv4V1PfdsagoHBzMYGMiWvD4yNolOXxYDAzntWOqNsZEkRsaVYx1Ji+iJK7dSTijA73O+3hNj0xhOCqafGxkbR0xKg4sFMTAwoBkGHwqujo1LJ7HvII+BULLkdbEgw+8Dtm/frr1W4HM4mEpjYcL9dYAkYGDHLnCjpYYunc3h0IH98E+X7gCruSasT8Z4isORQwcQyQzZflbmc9i+5wB6DffnrlEOIm9+P5Ygo3Sd4ccPYyB91PajxmMJ+ICtbw6U7bAOD04jlXT/XBIE/cAb27ajLRxAMsPh8IF9KEza7wQkgceO3XuAqXDJ65xQgB/Fe3FihMPwZBLbd+2BLAqurotfLmDbwA50GRo+H5rmURBF7ftjYxxGp9I4MCgjw0t1fwYdDXd3dzfGx4s3xMTEBLq6lO7he/bswZIlSzSP/O1vfzu2bdtmarj7+/tpzbkEf04dxNJ5QdfjxyeywAvjJZ8fGBhAf38/OoZ2g+VE9PevBgD07ZQQb48hEBSw4vil6F/SYTUsAKAzxaHw1IjpXBK7tmFBRxT9/cqOJNydAffEUXS3xV3Pff7+7UhEGPT3Ly95PfoGhyWLutHf36cdS72xLXsYh7kx9C1bjgtueQovX78e81rDai3vw45z+FPyAKaRNP1cdCuPJX0dCIfTuvf3gWUYV8fWs11EZ3cL+vuXlrye4yWwgQMlYyzYyuP1Q1OY39Pl+ry1PjuOBYuXol8XUAQA36+GsHrFci12QFDNNWmJDmMkk8HqFcuxvCdu+9mF2wQkOtvQ37+45PV0ZALxbTnH3162XEJb11Gc+rY+x3kZjyXIHMBxy1eWxZj+ML4PmUC28uOOHMWipcehrz0K2T+IE1avwPzWiO13Ei9MYuGiJehfWvp8TucEBHX3ItORgvC/01i4aAkSsTTC4bDj/KLho1iy7Hgs7iylE31DScQiU9r3g6NpSK9Mo6W9E60+oL9/ZUXHbYXXXnvN9HXH/cJZZ52FZ599FgCwbds2LFq0COGwsrItWLAAe/fuBc8r9ZIHBgawdOlSKhN2i7GUe0UJYK8qESXjNpqtSGsdsin5aYyyk6SLHpf6c2V868zJmdCO6hFklIYUO4ZSAICtg9MA3FUGBFzUKjE5Hquem0ZYaa3NElLiYQZj6XzFHLdZcJIWxw1A4/LdcLxWVIlbOWCYDeCyk52NthmsaYrKddzKXIqBX9K423EOFvSnVJAR0D1zhOOuRLJnVffGGCtpUevYkz6U9Yajx71mzRqsXr0al1xyCQKBAL797W9j8+bNaGlpwbnnnotPfOITuPzyy8EwDE466SS84x3vqPuk9RjP8BUF10IBm+BkoVCWlj6ayrvObrTNtDIYNMJtVyKeDLF+JHPlxmumCtvoEQz4IIgFrTvNhPqv6DJbr5LqgADwncvWuk4nt8puNON8W8JKL8IWh24vJeNbcdyUVCUAtIffDffeFmUxmTXnuGnNxwpWyS/V6LgBouVWxssL7mJLVnpro/KDFJmqRD4btIkrGZ28dF5QOO72JjDcAHDNNdeU/L1q1Srt/5dffjkuv/xyurOqAGPpPE5a3Ob687aSPUlGmC29GHtHM6oMzflU6YM1ZV1qTLyfvzljGS58yzz3c2cCyIv5std5FwkvtEEelim1HdRYRpmXILoLiNn3rCxXpXz4HYtcz83K4zY+yEBx59PhsrokYKOnr1LHbYZwsJiM44SueAgHxifLXncrB6wFQYsgs3H36hZEy00ajbg5n1bqFmMgWmlgAqQ4sQLDbbEoFOSSQHmYVaSRKc65yTINzPrWZWPpvGMlOz2IWN/MuBqVH4mIUr83m3cnQ/P7fWD9ftNOLWbez80Xl8cC7GCd8i5VVZWuFgQDSq1i4ulNqh632/oY9mVhVYNTZS5XiPGbUweFcmPSrd47lUjhrBZ/mlQJoUjcGm7SULl0PpVrqSuFlZbbuHt1C9LTU5Bk+H0+V167VXE3fb1sQFGtJCIKNVZrDXxjchNpjjKW5meEKpn1Ke/jab4iw+1XW4CZ32ylNzp5IDK8iFglZT9NV//avR8rCmCm6iPoEQ0padZTWQHzW8NayrVxC2kFNuCz3PlIhdoMjuU1MNEWL1GDTgvb7QNgeli1LxMoygGJp+lGqtnVEsJYmi97vVqvtxJYLcCCSymiEaReidva64CTNLV0jESYxWgq7/p5sarHYpYZGg8xVZUOqAaz33BnKkvAAYhg34QDNVyMrngIhydzYAN+13xdiLXqUlM732iVPNSI4CTpkD6V5dHXHkFWbfLqNmPOPk25usAWgdIA19yYGLXFK3tb8MGT+9A/z75pgB5m7cukgoyCLLvSmbtBVwX3dGcsaOpx61tr1QuWNIXLBdyISFChodzWXgccgpOG69ESYTGc5FyXsLUa2+w+bwkzGElVloVbLWa14RalApI5991vCKxaEhm3P90tIQyp2Y2ux7ZIwqmmdkP52HRbl9WCaDCAbF7EVFbAwrYI0nmdx+2yMJBVurTk0JncCYphNb8GRi8w4Pfhnz/0VlfdbwjMPG5Sp6SaZCYzfGHdCjx19VmuPtvdEsJ4mi9LhHFTF71W2HHA1eyait1/Cgi7vKdZixLBZg5AIsxgKMm5oj4BaxWa2c6yLaoEuiuta14NZrXhnsjyaIuyFXs5VimyguFGN6tX7QTrbXrtfKNVRqBbCR5NkPrDk1keC9oiyKget75sqh2cKsvV4rlale5VOM/az5MZx00zMAkoxnhlb4urz4bZAIKMX2v8QVCtJK8SWKaEV+lxhwlVIjjXIi/OwdwRk0xqgiciLI5Oc66pT0uP22Rh6lbrqVdC3VaLWW24x1I8Ol2Wc9UjxFpdjFKPjHhPcgVRMrsuNTVTJXb1uGfYcEeDjMZx97VHNY21284ndsFJySSIWAnCFl6SUWVQLcxqxvBidf0maaEzXk6XVCvJqwQsY14h0E37OjOQyohuapETWHn9Zg5Aa4TF0DRX0l/TcWyLhcl4n5OcjEqp22owqw33aDqvybkqgZXHbeUVVxJssOpSUy3nVzK2heGupFM8LQQZPwI+H4aSHPraI0irHrfbBcq+E1GhZo/brGlzNcX9zaD0Fi0dv1oPkxY6Y0ohLj1o3HNOsIpVVEvTkObFFQUnLSSJZg5AQu085V5sYCUtLffmiaftloapBbNaDjic5LSuIZXAyuM2e/ie+dJZWvU0V2Nbqkpq936sMjN5l82MaSMWCmAyK2Bhe0Qrx+k26cMuOEm2uOY5qM6w8rhpeaBmHrcgyg25BgSJCFsmgayWZ64EVpSXUKjufETYAJI5NZHFLZ3B+EvKwRKIJg5AIqI2Zg7X5nGb0VAfP30Jzlje6WrcWjGrDfdoKo+elrDzBw2wVpWU3+jLe9zxjATWXWpq36ZbLTiNUJXo0RULaR632yAsy1hXByTJDdUabkuPm0KAGDBP8KFBhdUC0sNUD5ryRCuwFnXZzZKd3CCsUiXZCqrsBdUSuEaY9b0kTlhXPASIKVdju80MjQYZrO1rczXnWjGrqZLhJIfeajxuxjyYUU0NYbOxzTW+9eO4GxGcJL8LKJruLK9ku/EuvdqgTesyseDcQccObutL1DK+UU/fiDiDHqSQvx5SoTbKyQ1CFr1Da9JxCxKyLmuvA9Zev2TiFS9sU/T6bgOIdjkBjaTGZrXhHklW6XHbKD9q11rbJH/U+BCZpVoT/XC91QNmIFQHG/Aj4FcSakSp4E5Vwlg3Cxal0uJAlcKyyBSlh81sfJpZk9WAtM7SY0aCkxa1Sqrd3RCOm6uQKrEOTpYe/4kLE4iHGCzrctf1KWjlBMwADWWHWU2VjKSq5LjtshtrNq7184pDJi2ziKdHSz9cCfT+coRVvO5KOO68Ta0SN8bfCopHTE9bbD6+4TpQCnxWC7O+k2LBXXW9WmClKql2dxMJ+pHjJWR5scIkGfMu98bFo6cljK1fP9/1fKzljvWnoewwqw33cDKP3qo97voEr6wScGhRJUbPohHJNwTf+/DbIKlJH1GVm+RdysCIHNCqZkzNOm6La0Cj6JKZ0sBtB/p6oSXMYHysVA4oSjKYUJ1rlVhoqKs912HVAchW0LvR2uOunSqyTtarfwEvO8xawy3LshKcpMhx1zO70U0DXScEA36Iak9HckPSTvyoBBetna/9XynHKbrm/gJ+H/w+n5olaTDcFDIn60WFAebNghu5gAKqx92Q4KR51qJoEhh0g5aQEmStpAWY4hWbxzRqdpaskvUaTJXMWo57KisgzPqrKqFox3HXejHsKsfVEnADlISgsIHnbrSihEApDlSAIMmuFxI2YPHQ15iqTRoRk6az2riUMgnNPG5jn8yZRkuYQbpBmZOCCU1RbUwnEWGQ5ARkeck9VWIh2XPTtLrasWnlBFSLxj/xVWIkldeanFYKqzKQNIJXtlSJy0YAtuMbFoZGe3oE0WAAWV5UZHFua0xYPPRmXngl8Pl8Wm10PWgF60w97gYvoKRZrR40cgecYKXHrzamozQ7EJETJERcqkqs6t4IFCpy2i4KDVQRNf6JrxLVSgGBeifJ1M+bB1DucTdYhkYQCTLIClJFXWCsSmYKVdZyNo5tNK60JFyWHHcDr0MsFChr7TYTXiFrU/a0GqNJauBk8yKiLnfT1nJACjtom3iYJwesAiNVJt8A9oXXqahKDBy3LMv06mQ0qccdYf3geAmC5D793qxeSaEgQ5aBWu2NWQlcWqqSsFmRqQZfh2iQKcseNFa7rAesy7pWd66ZgB8RNoDRdL4iqsRMoVRrsTIytuXC5FEllWM4yVXUaFcPa20mBeWHSa0S0oqLhmTPjONuZMYeATEcldBNZg+9oAaIaz1XYZO66DQCxIC5nr6S2hr1gEJVGdLwKTkLdrBvYlDdbyfCDI5Oce5VJRZJQDR2WHZyQI8qqQKjNXDc1gGH2h9sM6qElgwNKOdXG20wCCLBQMVUidlDb1b8vhqEmEBZTW5a1IFpWdeGe9xWVEm9OW7rnpPVGrZEhMXBiazrOvv2fSHr07zEy5ysEiOp6j1uq4pfNJQf5oa7vPNK9eMHmpQqCYDjJaVxsWtVSfkDpyTf0CoEVZ/gJKHa9KqVRl+HWIjR2scRVJt2XgksPW6XddnNkAizEAsyOmIuDbcNXVMvj7vRmZONf+KrxHAyj17aHjeFrbRZrRKaYn1j9mS+wYkfBGSrXknqt5m3Vmu6O4FZ0wmlBkrtY/t8vjIvT9HTN+46hBg/REkpOUBAq6iWHViLmjO1tE0juRluGzhbKz9q33FUUo97JjFrDXdtHrdNESIKFfxMPW5KfJiZx90MVEmYDSAriFoLLzcweyiq7Q5uhFnaO81gXdigWmm0x+3z+ZQ4g25OEgWqwAmsXW37Kp8lcv+0uCynbNmwmIojVt9krmoxKzMnZVlWCkxVKQe0TJGtsSodYL4o0Eh3JyjnuJvDcEeDAYwkuQrlgOUNCWj1STTzuGlKuIzcZ6MTcACSvSpppUsFlwW/avpNy3Z61T9Lnzn7eJywIOH68/VMkqlnXaNaMCsN92RWQIjxV91pwqwnoVSQ4QMqahprObaZDI1acLJ07o329AiqoUoU5Ud5tUM6yg8Tj5viw2ZUrTTDAhoLBrTen8DMyAFjodLf1H67BkXLyt4W1/02AdW4WgYnaSTgNEefVz0a/8RXgSNTOSxQ6+pWAzOPu9oeeUaYacRFih630aPnm0RVEmZJkSn3D2zIogUYnbR0s50PPUNmVK00wwJq1HLToAoq/U1A0eLTUge5gb5gmR5UapVYtS6jkJVZCxr/xFeBo9Mc5rdWF5gEzBUHtZYSLRnbpDsKPU+vtLtLM2zRAeUBzvGS0kbN5cNi1KQD9FKJzWpm05RwGT1uvoLEo3rBqOUWKaiknBALMsgYZIi0tPhuQQqWiWW1aSgUdlOdPOOiIFDKhK4Wrn75jjvuwMaNG3HppZdiy5YtJe8NDQ3hyiuvxIc+9CHcfPPNdZmkEUenc5hfo8dt3FrVWkqUwKxWCc1ARrnH3WyqEtl1rRIzqkSgxHGbpbzTSIEujm/0uN0vWPVCNFRqRGkE251/M4Bs3iyWMLPnwky2R4PjtlwUml3H/fLLL2Pr1q24//77sWnTJmzatKnk/e9973v4+7//ezz44IPw+/0YHBys22QJjkxxWFCzx12fmy1kmrFH09NrTo+bUCWVqErMtqHUOG6zLjUUqQPjrq0pqBJWCU4SzIQc0MzjbkTlPDM9Oc3GGeZxqyb2uF955RWsX78eALBy5UqMjIwgl8tp72/btg3vfOc7AQC33HILFi5cWKepFnF0Oof5rdV73GbFjWhxVqZUiUhRhmZQlTS6Kh1BVFU0VKIqMauyJxQKVCRsYTM6jGImYZhtvvK6UUOgkEbmoBPCagNrSeeRNsLjNiufQK/EQf12htXC8eyOjo6io6ND+7ujowNjY2MAgGQyiXA4jBtvvBEbN27E7bffXr+Z6nB0msP8tlo87nJVCb1GshaJH9Q8PYMMTWiO4GQ0qOi4FQPm7liNuwdAbTdFJS29POWdZiahcYHOC42v0ki6EBHQklbaQdOP67xumvJXtzD1iikt1KSBsR5Sg4OTjno6li0VwetbTfE8j7179+LOO+9Eb28v/u7v/g7PPfcc1q1bVzbOwMAApSkDB0eTyI0dwQA/WtX3x7MisnlemxPHcTi8cxdkSah5nlJBRl4o4M0339TO097DWeRzWSrnYHIsiaExThtreGwCCTmDgQEOgHIsNM+1W0xkRaSzPEKMDwf27YMw7pw8MT0xhcmchIGB4gO352gOeS6HgYGBmo5lanwKoxkRAwNFT3A6mcTRwcMY8E9WNaYe+Wwa+w4ewgA7pYydzuLo4YMYyI+UfXamrgmXmsY+Po2BNmVHnMvz2L93D9LD9FS/ZscS9Mt4Y9t2dEaV3zmaEiAXxJm9Dwsitu/chXRrMU1+ajqJoaOydo2McHtdfLKEN3fsQko3diaXx8H9+yBPukvLpw3HK9rd3Y3x8XHt74mJCXR1dQEA2tvb0dfXp9Ejp59+Ovbs2WNquPv7+6lMuFCQMZHbjzPefmJV3W8AYCrLQ/rVEW1OAwMDWLKgD9HIFJV5MoH9WL5ytbZ1PlgYQtsRmcrYO7hB7M2MaGNF/5LHkkVd6O/vA6AcC61zXQlSnAD+kUGE/AxOWL3CFZW1eHI/8iOpkvmOMaNo3Sugv7+/pmNZPH0AmaPJku+Hf5/EsqVL0L+yu6ox9eh5U0BHdxv6+xcDAAJPjmLl8uPRb5I4MlPXpG9wJ6SCjP7+VcoL/sPoX70SXfHqEtXMYHYsrbFhCLFehDqjOK47juBoGtHwxIzeh7EnRrBo6TKsnlc8/+GXU1i6eBH6+3tNv+P2urQ+NYYFi5aif2Gr9pr/V0exasVyLOl01y2+Wrz22mumrzvuI8466yw8++yzABQ+e9GiRQiHFZoiEAhgwYIFOHToEADgjTfewLJly2jN2RRjmTxawkzVRhuw0XFT2lYa6RIaxW4IyjInm4BbBYrbyUrKzJqWXqVQQ5nMp4w/p5hJaMzMbIbgZMyo454BOSCgUDSf+ukfse72F4u/O8M0gnn5hPp1PJqJ5CY7OHrca9aswerVq3HJJZcgEAjg29/+NjZv3oyWlhace+65uP7663HzzTcjl8thxYoVWiCzXjg6VRu/DSjSobxY2mGcvmSvAJL7VUvBnbKxDUkrzVKrhAn4wfj9SOaECjInTVLeKdWQNjPcNB82Y2ZmM2RORg1dcGZCDggAHbEg9Go5WvGiSmAuB6RXPsEYi6ml+iENuCK/rrnmmpK/V61apf1/yZIl+MlPfkJ3VjY4Op3DvET1ihJAMTJEm0mMBM20dGOghGqNDEOEu1nkgICS/jyZFRBzWYrArCEBLT19JOgvkcYBdFKgCcr09E2w8zEm4MxU9mI8VLzeeVFtpjHD58JsF02rxIFZcJKmM1YNmuOJrwBHpjgsqNHjBsqNK80ssxBbWtqV5pbV6KU2S8o7AG3H4tZYmG5BKWZOmj1sNItMlXjcQhMk4AQZZHTJMAKFjk5uQOIZ3S0hTGWFhnij1nJAOveS+e6tieWAzYZa65QQGDmxenrcVOtxm3jczWK4K/XuzAty0fSS6iP5BMw97hDbaMMdQE5QqBKiq54Jj/umi/rxxj+eh45oEBMZnur97hbmNYJoUiVm8RLP43aNQ5NZLGqP1jyO8cGjVWSqOLaRKqHncZd1wGlgAX89KjUSZinvtIxrJKh05CkZm2IPRvPr0GjDXfS4Z9Kw+P0+tEZYtMdYTGZ4tddlM1AldJwx652h53G7xqGJHBZ11MHjpri9M3bBoRllN67+ebHxnh5BpQ+rWdCHNFauFaa8JCX+HCithSJIBfh8voZynkAxexVojGFpjwYxkeUhVJA9SwtmzRRoJeAY71NZltW6N00enGwWyLKMQxO0PO6AgSqh214sX0bD1MebbwZPj+CEBYmS1GcnhE2yG8UCndZlZoa7XjsfTpAQbgK6Sl83pBH1QtpjQUxmeHS3hGa+VkkdU96N9xJJd5+p6odmmFWGezonQIb7XnR2IJJAArq1mo1j09UPl3V5bxKP+46PvK0iw21WwU+kdK7CplRJfWINnFCoKa+AFiI6VYnYALpC4bgFtEWDjVGVmMRLaNBFYdZfWi63wenuwCyjShSaJEplpSv3iukZV2MCDs3aDc3scbMBf0UGzLxmNp3CSOb1JSgGJw0edzMEiGM6HTetRKZK0BphMZ0TZqRlmhFswA9eKm+kQKvxtJ4qUVQzjb3ejb/bKoASmKyd3waIx23Y/lA0riXlNSmOTRYAwufxYgGhJvD2qoFVdUAa8QA24IcMlPCeNCu66ZsF58Xm8LjDqkpHKsgNKfREDHcjsgoVR8wkGE1JVVJavMvzuCvCoYksFnXUzm8DisdUFpykdDEiwVI6g2YCDqCWLFXnnm8ij7tSEENT0NErNDNYjdmTVGWZRo+7CQy33+/TdhqNCE4miMc9Q/pxPaIsU55wRSkYHTbkZUgzUC7XCbPqiafvcRuCk5QuRlkLKcodoUOqQZIKMiSZ7qIwk/D7fcrupE7G1egp0ZSp6eus5EUJ4SaJM0SDDLJ5sSHByUSEQZITVFXJzP62QhPVp8RBWXCSoqy0WjTH3eYShOOmgRBbGsygWas5YlLsJ0ixvRjZphN+u5HR7VoRDzElxf9pGtdI0A+OL15jmhIufesyTigg3ATt4wBDC7kGUCXJnECt80wlUAKzhk48lAyskdLzqJIKcWiSIlVi9LgpcoLGFZoX6a7QZJveDBXpakUsxCCt79pC0VPUXweivaUVsNN73JzQTB53ABleVNP7G8Nx06YG3SBmSPcHlJ0uvUqTXnCyKhQKMgYnc+ijRJUYPW5aiR8A8XpKO4LQNLBERpcXpaZoFFwLykuR0gts6Q03be2t0eNulutQ9Lhnnq4ghjvfgPo5xmcOIMFJGn1kDcFJTw7oHkNJDokIi6jLynNOKFeVUPT0DBw37Qg/kSc1U52SahE3etyUGysXMwnpPmz6ZsHNxHHHQspCqOzyZnZO8RCDvFhAmhNnPFhLjpuA5BP4qQQnjVTJzFNBRjTH3eYC+8YyOK6LXrcJM1UJLU8vWqYqoWu4SWrzXDDcMWODW4oPhV7dQ8v7IggxfnCiBFmWmyYBB1A9z7xIfZfnBj6fD4kwg5FUfsbPRyQYQKZOTSSM6qRG7GaMmDVP/d6xDI7rpme4yzInKXpkRlUJT5nzI9vCucBxRw0eN9UgsW6LSysZg4AJ+BHw+SBIsspxN4vhZnRUyczfG60RFsNJbsZ3IDFVTUNAN55RynHTLGFRLWbNU79vNINlND1uQ3VAiWLAIcIaeFvKRXfIw8lLs9/jjhuCSjSpkoieKqFUcEgPUmuFa6KyA2RRb5RXmIiwisc9w5x/mQSXogNgrKlDk1atFs1xt7nAvrE0lnXFqY1n5KFpJizoq7QB9aFKsrykFO+f5YY7ZpAD0qRKwkFdcLIO2lvCczejHJBvgBwQ0HvcjTDcxlgJxSC3YVFoxLnVY9Y89fsoUyXRIFNmXGly3EZVCV2qhEGWF5Froi16tYiHGaTy9TlXem5SqkPHEtIwOC80k8fNIMNLEBqUUZuIsMjyEiLBGaZKQkwJx027TR0vFTN8PVWJS/BiAUemOSrlXAnMtla0CuNEyjxuuis0mTsnSIjMcsPdFmExneW1v2kmb+gNt0AxM5aANAzOi83lcec0qqQxHjcA131HaSHE+CFKBa02Dc1CV36/ryQmJniqEnc4OJHFgtYwVVqgzCum2PwzGmSQNUShqc49pCRZ5AQJkWBzGIxq0RELYjIraH8LFLPSIsHS4CRtXjKoxkmaKjipep5Kw96Z9wqJ4W6PBWf0d30+X0lOgESprjuB8V7yVCUusG+MbmASKAb4CKgWmWLrq+NWIugSOKEw6z3u9lgQk3qPm3Kzgxxf9MBoKwHIzqqpMidZRQ7YKI67pyUEgE7N/Eqh3+nSzm5UMjPVkrkN7vAOzBrDTTcwCShea3khKHoFiASpoCUB0E4BJoHVHN88nl61IA1mCWgW5FLkgMXmubQ97rjq3TaTjpsUWxKkxnDchM5sj86sxw0Qnlt3vSk+czF1lwvQLQ9cLWaJ4aYbmATKqRKa9UR8Pp/qdSvj85QfotgcCk6SBrMENFPeleSe+mROAmpdkLwIrokyJzUddwP6PgLAe1f34PmvnN2Q31aSj4p9QGmWXtWrn8QGLYp6NMfd5oC9lDXcQHmNDJ42D62XotVJDjgXgpMdMaXBLAFNHbd+e1uPwkCkQJbSAac5rkOpjnvmH++A30f9WXWLWIhBKq/ES2hVBtTGDjJIq4sC36D4gR6zwnDvHkljRQ9dqsSo4+ZFuproEr5NLFDtwUcezlwDZFe0EWED8MGHFKc+cBQj9vrKgzQTMgjiISVbT7kOzWK4maKOu8HGZaaRCLNIcYQqoVMZkEBfmkEpp9zY6930T/1YOg+xIKNbDXrQAqEbCARJprr9ieqyJ2lz3OTh5MTZ73H7fD70JkIYSeUB0NVxx3WcJ60Sn3oQ7XA6L6Il1Bx9t/XVARu9nZ9pJMKMZrh5ke7zrHcChCZYFJv+yu4cTmFVbwv1ZgFhtawrCSDSrvtBPHpZlpWtFcVtOgms5vjmCYrVgt5EGMNJDgDdOhAlHHcdVBaxYADpvIh0XkQ83CSGO9RYqqSRSKiNHAD61Ke+4UczNOh29et33HEHNm7ciEsvvRRbtmwx/cztt9+Oj370o1QnBwC7htNY0UuXJgEUT0/f2oq24SZp76TON43ykvqxMyq3OucMN8WIvb5kLO32cUAxYJXJS4g1jcet1H45Fg13i87jph2c1Qcnm2E34/jrL7/8MrZu3Yr7778fmzZtwqZNm8o+s3v3bvzxj3+sywR3DKewsrelLmNHdXQJbeVHPYNEJF0/NweCkwDQmwhhOKmnSuhx3CXBSdoed4hBmhOR5cUZzxS0QkQtiJQXGl96dKaRCLNIqrES2sZVcQL0wckmN9yvvPIK1q9fDwBYuXIlRkZGkMvlSj5z22234Utf+lJdJrirroa7KB/iKa/QxJsXRPpBsVgwgKwgIZMXmyYoVgvKqRJ6xb44QUKhINeH4w4yGE0rtadpj10tAmoD5umcMOsLkFWKRIQpoUpoGteYTj5M21ZUA0c3YXR0FKtXr9b+7ujowNjYGBYtWgQA2Lx5M975zndiwYIFtuMMDAxUPDlZljFwZBpIHsXAwEjF33dCQBaxbccuLIjKyOZ5HNy3B+lhOp6TmEtj9wEBXYUJ+FGo6vjtEGF82DeSxNjRwxgQRrXXOY6j/lv1hphKY/fhDAYGBpDnRezdvQujoQCVYwkxPvxpy5s4cCgNLpOnem4mR7LYPzKNcMD+/p7paxIKAEcnpjF8FBgIJamO3cz31/RYGkdG0xgYGMCBQylk0zlq1yU5kcLgiDLeyNg4YlIaAwMcralXDEcrxbKlqauyLGuBwqmpKTz66KP48Y9/jKGhIdtx+vv7K57ccJJDODiId520puLvukH78xOY17cY4ewwCvDjxP5VaKOU8dW3R0ZLIoTFS+cjHh6t6vjt0JMYxt6xDN7avwJLdbrZgYEB6r9Vb6TCE3j6wHb09/dDlPdj7Yn9CLMBKseSiAyib+nx6MgOobeQpnpuhJYpjL44ggWtEdtxZ/qaJKJHwSOAZUsWo7+/l+rYzXx/jTNjeOHwbvT39+PPqYPo5qeoXZf94lFsmxpEf38/Ylt5LO5rR3//YlpTt8Rrr71m+rqjv9/d3Y3x8XHt74mJCXR1dQFQ+O/R0VFcfvnl+MIXvoBt27bh1ltvpTRlYMdQCit66kOTAETiU5/gZIuqKeXqVAiqM64sMInIzNeEoA2F4+YgyzLylCP2RMaVF+h3C2qPBsEJhaZRlBBEWQajqTxiodlPo1WCREQXnKRdHyhUbPjRiLZwRjj++llnnYVnn30WALBt2zYsWrQI4XAYALBhwwY89thjeOCBB3D33XfjxBNPxA033EBtcjuHU1g1r36GuyWsBJcA+hc6rka4c3UqQERqQbQ0mdGoBr2JMEaSec1o01TgEBlXPboFkUJK8SZRlBBEQwFM54SmCZjOFFp0wUnaPHQsFNAUSnwTKHYcr+yaNWuwevVqXHLJJQgEAvj2t7+NzZs3o6WlBeeee25dJxcJBnDm8q66jU+i0AVGhiTTLRxDpEn16o6iys8bfgPRQJgNIMT6MZLMU29IQNLe83VISycGO9pkAWIyn2aRKM4UEmFjcJJm5qRex934Djiuruw111xT8veqVavKPtPX14d7772XzqxUXPHOJVTHM4II9oWIkmVFM8lHyeIS6lYzu689Qn3MRqI7HsLhySx146pRJWIB7TG6hozcL72JMNVxawWJ0xxrVAmhJ2VZhiDKCFEu7Fai424wVXJsLckGJMIMkpwAoSBTvxDxEFvXAkRfe98JuOqcFdTHbRS6WkI4PJmjTivF1XKcebE+jZVfuWE9EuHmijOQmtjHmscdZPxgAj5FhkuZGtMnc82azMm5CsXjFiFKoH4hilRJfTzugN9HTQHTDOiOh3B4KkfduJIAdF6szwLamwg3nZaeOCHNUj9lJkEKTQm0ddxqXRpZlpvC4z62DXeYxXSuTh53uFjyM3yMJUJUg654EIcnstRT+Elwsl4edzNiw4nzsKInTr2+z2xAi8pz0w4gBhk/Aj4f8mKhKYKTx8adbIFERKVKJPqGu4Vw3E1U8rOZ0RUPYd94hrpCI6Yz3I32kmYKJy1ux9Nfek+jp9EQJCKs+kwXEKSdsawqSxTFilcdsGFIhNXgJOWSrmTsFCeCE+dGBb96o7slhL2jGbRQ5ov1Ou5jxeM+ltEWYTGVFeqSlt4WDSo79DpISyvFMX0nK6uzCL4gU5ehhRg/pIKMFCd4VIkLdMVDmM4JSEToetzxUABpTlQ4bm8BnfNojwYxmRXqUlSslSwKHlXSWLSqcsCcUKCerODz+dASVjLYwh5V4oiehKKEoK3QaI0oXtKxxHEfy2iLBjGZ4anX41bGZjGd49Xdm9cBp2EgCTicKNdFOtUSZjGSytclAWeugdRbmd9KVxPdFi1unT3DPffREWMxmeXr0iyZ0DDNELc6pu/kMKvQGcm8VJfst+6WEPaN0Q+4zUUkwixOXtKOU5d1UB1X2TrzqsftLaBzHW0aVVJAkHJ7MTJ2tgnq4B/TFsXn86ErHsLRlIBYkH72W28ihNcOTGo1LTzY45efPZ36mG1RFlM5AS1h5phRlRzLaFepknxdgpMsRlN5+ICG30vH/J3ckwjj8LSAaB3Sg3talMVgLiXKzDYoASUembx4zKWAH4toV6mSLE+/nVxbhMXR6VzDaRLAM9yYlwhh7wSvVdujCVLDot3zuBuGMBsA4/djOJlH6xwogevBHu3RIKayQl3aybVFgzg6xTVFUbFj3nD3JsI4nBTQEaNvuLvVmhGLO6PUx/bgHsTzajQv6aH+aI8GMZHlkalD3Ko1yuLwZLYpasA0fgYNxjxVxdAVp2+437d2PsKs3wuKNRixUABjaRyTKeDHGhQVEY9okKkLVXJkmsPJS9qpjlsNjnmPu39eAgCwog4NicNsAO9ba9+L00P9IUpyo6fgYYYQZgMIs0ojCdoed2dM2UE3A/V5zHvc717Rha++uwfLOmPOH/YwK/HtS9Zg53Cq0dPwMEOIBZXKnLRLTcxvU3bnzVDG95g33EzAj/ceF6faLstDc+HsVT04e1VPo6fhYYZQL6kekRd2qbGrRuKYN9wePHiYW/jkGUuxdTBZl7Ffvn492mOex+3BgwcPVPHJM5bVbex5lEsyVItjPjjpwYMHD7MNnuH24MGDh1kGz3B78ODBwyyDZ7g9ePDgYZbBM9wePHjwMMvgGW4PHjx4mGXwDLcHDx48zDJ4htuDBw8eZhl8sizXvQLPa6+9Vu+f8ODBg4c5iZNPPrnstRkx3B48ePDggR48qsSDBw8eZhk8w+3BgwcPswzHlOHmOA7nnHMONm/ejPHxcXzqU5/Chz/8YXzxi18Ez/MAgKeffhof+chH8Fd/9Vd46KGHGjxjc/zqV7/CpZdeiksuuQQvvPDCrD2WTCaDz3/+8/joRz+KD3/4w3jxxRexb98+XHnllbjssstwyy23gDB5v/jFL7Bx40Z84AMfwIsvvtjgmRexc+dOnHPOObjvvvsAoKJrIUkSbr75ZmzcuBEbN27EoUOHGnYcQPmxDA8P42/+5m9w5ZVX4mMf+xiGh4cBNP+xGI+D4Le//S1WrVql/d3sx2EL+RjCd7/7XfnSSy+Vf/nLX8pf/epX5ccee0yWZVnetGmT/OCDD8qpVEo+55xz5GQyKWezWfn888+X0+l0g2ddinQ6LV9yySUyx3Hy0NCQfOONN87aY7n33nvlf/qnf5JlWZaHhobk8847T77iiivkP//5z7Isy/Lf//3fy7///e/lAwcOyO9///tlnufl0dFR+YILLpALhUIjpy7LsixnMhn5yiuvlG+66Sb53nvvlWVZruhaPPTQQ/LNN98sy7IsP/fcc/I111zTVMdy3XXXacdy3333yZs2bWr6YzE7DlmWZY7j5CuvvFI+44wzZFmWm/44nHDMeNx79uzBnj17cPbZZwMAXn31Vaxbtw4AsH79erz00kvYsmUL1qxZg5aWFkQiEbz97W/H//7v/zZw1uV46aWX8J73vAehUAi9vb341re+NWuPpb29HePj4wCAqakptLe34+DBg3jrW98KAFi3bh1eeuklvPrqq3j3u98NlmXR1dWF7u5u7N27t5FTBwAEg0Hcc8896OkpNmmo5Fq88sorWL9+PQDgzDPPxKuvvtqQ4wDMj+Wmm27CeeedB0C5Vul0uumPxew4AOCHP/whrrjiCgSDSm/ZZj8OJxwzhvs73/kOrrvuOu3vTCaDcFiprdvR0YGxsTGMjo6io6ND+0xnZyfGxsZmfK52OHr0KCYmJvDpT38al19+Of7whz/M2mO58MILcfToUZx//vn4+Mc/jmuuuQatra3a+2TOZscyOjraiCmXgGEY7bwTVHIt9K+zLAtJkiBJ0swdgA5mxxKLxcAwDCRJwi9+8QtcdNFFTX8sZsexb98+7Nq1Cxs2bNBea/bjcMIx0Ujh4YcfximnnIK+vj7tNZYtdrGQZRk+n6/kNf3rzQSe5zE+Po4f/ehHOHToED7xiU8gECj21ptNx/LII49g4cKF+H//7/9h+/bt+MIXvoBIJKK9P5uOhaCS+8r4ejNCkiR89atfxamnnorTTjsNjz/+eMn7s+FYbrvtNtx0000lr83mawIcI4b7hRdewOHDh/H0009jaGgIwWAQoVAIuVwOkUgEY2Nj6OnpQXd3t7Z1B4CxsTGcdtppDZx5Obq7u/G2t70NgUAAS5cuRTweh9/vn5XH8vrrr+Oss84CAKxevRocx4HjOO19/bHs3Lmz7PVmRCwWc30t9K/zPA+WZUsW4WbA9ddfj76+Pnzxi18EgFl3LMPDw9izZw++9KUvAQBGRkZw5ZVX4qqrrppVx2HEMUGV3HHHHXjooYfwwAMP4EMf+hA+97nP4b3vfS+effZZAEp0+T3veQ/Wrl2LHTt2IJVKIZPJ4I033sApp5zS4NmX4vTTT8fLL78MWZYxPj6OTCYza49l8eLF2Lp1KwDlAYvFYlizZg1ef/11AMVjOeOMM/DSSy9BEAQMDw9jamoKy5bVrz1VLXj3u9/t+lqcddZZ2mdfeOEFnH766Y2cehkeffRR+P1+XH311dprs+1Yent78fTTT+OBBx7AAw88gJ6eHtx3332z7jiMOOYyJ++66y4sXLgQZ555Jr785S8jm81i2bJl2LRpExiGweOPP45//dd/hd/vx6c//Wm8733va/SUy/Bf//Vf+PWvf63J6d7ylrfMymPJZDK47rrrMDk5CUEQcNVVV6G7uxvXX389JEnCqaeeqsUlfvazn+GXv/wl/H4/vvrVr+Jd73pXg2cPbN26FbfddhsGBwfBMAx6e3vxz//8z/jKV77i6lpIkoQbbrgBu3btQiQSwe2334558+Y1zbGMj48jFAohHo8DAI4//njccsstTX0sZsdx1113oa2tDYAS8H7uuecAoKmPwwnHnOH24MGDh9mOY4Iq8eDBg4e5BM9we/DgwcMsg2e4PXjw4GGWwTPcHjx48DDL4BluDx48eJhlOCYScDzMXmzatAnbtm3D6OgocrkcFi9ejNbWVqxduxbveMc7cNJJJ9X8G3fddRd+9atfobe3F4VCAeFwGLfeeit6e3tNP79582a0tLTg3HPPNX3/yJEjGBsbw9q1a2uemwcPZvDkgB5mBTZv3oxdu3bh2muvpT72XXfdhfb2dlx55ZUAlBIJr776Km699daqxtu8eTOy2aw2ngcPtOF53B5mJa677jqcf/75CIfDuO+++xAIBLBt2zZ88YtfxFNPPYWdO3fim9/8Jk477TQ89dRT+MlPfgK/34+1a9c6Gv+3vOUtWn3m3/zmN/jJT34ChmFw4okn4qabbtIM/Xve8x5cf/31WLRoEQYGBvCWt7wFV199Ne6++24wDIP58+drleY8eKAJz3B7mNXw+/3Yvn07Hn/8cTz99NO488478dhjj+Gpp57Co48+irVr1+KHP/wh7r//fgSDQXzxi1/E66+/bkuxPP/881izZg0ymQy++93v4uGHH0Y8HsdnPvMZ/OEPf9A+5/P5sHXrVnzve99De3s71q1bh2uuuQaXXHIJ2tvbPaPtoW7wDLeHWY9Vq1YhGAyiq6sLy5cv1+p2p9Np7N+/H4ODg/jUpz4FAEilUhgcHCwz3D/72c/w5JNPolAoYOnSpbj++uuxf/9+LFu2TEv5Pvnkk7F9+/aS7y1ZsgTd3d0AgK6uLqRSqRk4Yg/HOjzD7WHWg2EY0/8TnHDCCfjJT35iO8bHPvaxMk7a5/NBHwIyKydrrBznhYw8zAQ8OaCHOY1ly5Zh7969WqnOf/mXf9F6Jzph6dKl2L9/P9LpNGRZxiuvvII1a9Y4fs/n80EUxZrm7cGDHTyP28OcRiQSwY033oi//du/BcuyOPHEE13X8o5Go/jyl7+MT37ykwgEAnjHO96BU045pYTnNsNJJ52E66+/Hp2dnbj44otpHIYHDyXw5IAePHjwMMvgUSUePHjwMMvgGW4PHjx4mGXwDLcHDx48zDJ4htuDBw8eZhk8w+3BgwcPswye4fbgwYOHWQbPcHvw4MHDLINnuD148OBhluH/A9MgP8WbEl7GAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def generate_data(Noise):\n",
    "    \"\"\" Generate Mackey-Glass time series.\n",
    "\n",
    "    Parameters \n",
    "    ----------\n",
    "    Noise : Boolean\n",
    "        A boolean that determines if gaussian noise should be added to the targets\n",
    "\n",
    "    Returns\n",
    "    -----------\n",
    "    inputs : array\n",
    "        Returns the inputs that can later be split and used for training\n",
    "    target : array\n",
    "        Returns the targets that can later be split and used for training\n",
    "\n",
    "    \"\"\"\n",
    "    inputs = np.zeros((time_to-time_from, 5))\n",
    "    targets = np.zeros((time_to-time_from, 1))\n",
    "    f = Euler_Approximation(time_to + 5, beta, gamma, tau, n)\n",
    "    \n",
    "    if Noise == True:\n",
    "        for time in range(time_from, time_to):\n",
    "            inputs[time-time_from][0] = f[time-20]  \n",
    "            inputs[time-time_from][1] = f[time-15]  \n",
    "            inputs[time-time_from][2] = f[time-10]\n",
    "            inputs[time-time_from][3] = f[time-5]\n",
    "            inputs[time-time_from][4] = f[time]\n",
    "            targets[time-time_from] = f[time+5] + np.random.normal(loc=0.05, scale=1)\n",
    "\n",
    "    else:\n",
    "        for time in range(time_from, time_to):\n",
    "            inputs[time-time_from][0] = f[time-20]  \n",
    "            inputs[time-time_from][1] = f[time-15]  \n",
    "            inputs[time-time_from][2] = f[time-10]\n",
    "            inputs[time-time_from][3] = f[time-5]\n",
    "            inputs[time-time_from][4] = f[time]\n",
    "            targets[time-time_from] = f[time+5] \n",
    "\n",
    "    plt.plot( range(time_from, time_to), targets, linewidth=1.0)\n",
    "    plt.xlabel(\"Time Point\")\n",
    "    plt.title(\"Mackey-Glass Time Series\")\n",
    "    plt.show()\n",
    "    return inputs, targets\n",
    "\n",
    "beta = 0.2\n",
    "gamma = 0.1\n",
    "n = 10\n",
    "tau = 25\n",
    "time_from = 301\n",
    "time_to = 1501\n",
    "t = time_to - time_from\n",
    "\n",
    "x = Euler_Approximation(t, beta, gamma,tau,n)\n",
    "inputs_noisy, targets_noisy = generate_data(Noise = True) \n",
    "inputs, targets = generate_data(Noise = False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "outer-robin",
   "metadata": {},
   "source": [
    "More chaos in the time series with Noise = True."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "rising-detail",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1200, 5)\n",
      "(1200, 1)\n"
     ]
    }
   ],
   "source": [
    "print(inputs.shape)\n",
    "print(targets.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "auburn-letter",
   "metadata": {},
   "outputs": [],
   "source": [
    "def show_curves(history):\n",
    "    # train_loss = history.history[\"loss\"]\n",
    "    # val_loss = history.history[\"val_loss\"]\n",
    "    # loss is MSE so loss and MSE are the same\n",
    "    mean_squared_error = history.history[\"mean_squared_error\"]\n",
    "    val_mean_squared_error = history.history[\"val_mean_squared_error\"]\n",
    "    \n",
    "    plt.plot(mean_squared_error, color=\"b\", label=\"MSE train\")\n",
    "    plt.plot(val_mean_squared_error, color=\"r\", label=\"MSE val\")\n",
    "    plt.title(\"MSE vs epochs\")\n",
    "    plt.legend(loc=\"upper right\")\n",
    "    plt.xlabel(\"Number of epochs\")\n",
    "    plt.ylabel(\"MSE\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "seven-object",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_uncomplied_model(input_dim, hidden_layer_1_dim, hidden_layer_2_dim):\n",
    "    # Input layer dim\n",
    "    input_shape = (input_dim,)\n",
    "    # Create model using Sequential\n",
    "    model = Sequential()\n",
    "    output_dim = 1\n",
    "    l2 = 10e-8\n",
    "    model.add(Dense(hidden_layer_1_dim, \n",
    "                    input_shape=input_shape, \n",
    "                    activation='sigmoid',\n",
    "                    kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=1),\n",
    "                    bias_regularizer=tf.keras.regularizers.l2(l2=l2),\n",
    "                    kernel_regularizer=tf.keras.regularizers.l2(l2=l2)))\n",
    "    \n",
    "    model.add(Dense(hidden_layer_2_dim, \n",
    "                    activation='sigmoid',\n",
    "                    kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=1),\n",
    "                    bias_regularizer=tf.keras.regularizers.l2(l2=l2),\n",
    "                    kernel_regularizer=tf.keras.regularizers.l2(l2=l2)))\n",
    "    \n",
    "    model.add(Dense(output_dim, \n",
    "                    activation='linear',\n",
    "                    kernel_initializer=tf.keras.initializers.RandomNormal(mean=0.0, stddev=1),\n",
    "                    bias_regularizer=tf.keras.regularizers.l2(l2=l2),\n",
    "                    kernel_regularizer=tf.keras.regularizers.l2(l2=l2)))\n",
    "    \n",
    "    return model\n",
    "\n",
    "def get_compiled_model(model, learning_rate):\n",
    "    # Configure the model and start training\n",
    "    # val_loss is almost mse_val, except the weight decay term, same with train\n",
    "    optimizer = tf.keras.optimizers.SGD(learning_rate=learning_rate, \n",
    "                                        momentum=0.9, \n",
    "                                        nesterov=True, \n",
    "                                        name=\"SGD\")\n",
    "    \n",
    "    model.compile(loss=\"mse\", optimizer=optimizer, metrics=['mean_squared_error'])\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "continuing-hacker",
   "metadata": {},
   "source": [
    "No noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "surprising-tattoo",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'inputs_training, targets_training,     inputs_testing, targets_testing,     inputs_evaluation, targets_evaluation = data_split(training_size=700, inputs=inputs, targets=targets)\\n\\ninput_dim = inputs_training.shape[1]\\nhidden_layer_1_dim = 8\\nhidden_layer_2_dim = 8\\nlearning_rate = 10e-3\\nbatch_size = 100\\nepochs = 10000\\n\\nprint(inputs_training.shape)'"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"\"\"inputs_training, targets_training, \\\n",
    "    inputs_testing, targets_testing, \\\n",
    "    inputs_evaluation, targets_evaluation = data_split(training_size=700, inputs=inputs, targets=targets)\n",
    "\n",
    "input_dim = inputs_training.shape[1]\n",
    "hidden_layer_1_dim = 8\n",
    "hidden_layer_2_dim = 8\n",
    "learning_rate = 10e-3\n",
    "batch_size = 100\n",
    "epochs = 10000\n",
    "\n",
    "print(inputs_training.shape)\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "another-coupon",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Training samples: 700 \n",
      " Testing samples: 300 \n",
      " Evaluation samples: 200\n",
      "(700, 5)\n"
     ]
    }
   ],
   "source": [
    "inputs_training, targets_training, \\\n",
    "    inputs_testing, targets_testing, \\\n",
    "    inputs_evaluation, targets_evaluation = data_split(training_size=700, inputs=inputs, targets=targets)\n",
    "\n",
    "input_dim = inputs_training.shape[1]\n",
    "hidden_layer_1_dim = 5\n",
    "hidden_layer_2_dim = 6\n",
    "learning_rate = 10e-3\n",
    "batch_size = 100\n",
    "epochs = 10000\n",
    "\n",
    "print(inputs_training.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "racial-excitement",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "Epoch 1/10000\n",
      "7/7 [==============================] - 1s 69ms/step - loss: 2.1143 - mean_squared_error: 2.1143 - val_loss: 0.3471 - val_mean_squared_error: 0.3471\n",
      "Epoch 2/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.4175 - mean_squared_error: 0.4175 - val_loss: 0.2062 - val_mean_squared_error: 0.2062\n",
      "Epoch 3/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.1285 - mean_squared_error: 0.1285 - val_loss: 0.0645 - val_mean_squared_error: 0.0645\n",
      "Epoch 4/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.1717 - mean_squared_error: 0.1717 - val_loss: 0.0427 - val_mean_squared_error: 0.0426\n",
      "Epoch 5/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.1120 - mean_squared_error: 0.1120 - val_loss: 0.0590 - val_mean_squared_error: 0.0590\n",
      "Epoch 6/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0986 - mean_squared_error: 0.0986 - val_loss: 0.0394 - val_mean_squared_error: 0.0394\n",
      "Epoch 7/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0997 - mean_squared_error: 0.0997 - val_loss: 0.0378 - val_mean_squared_error: 0.0378\n",
      "Epoch 8/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0966 - mean_squared_error: 0.0966 - val_loss: 0.0396 - val_mean_squared_error: 0.0396\n",
      "Epoch 9/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0909 - mean_squared_error: 0.0909 - val_loss: 0.0369 - val_mean_squared_error: 0.0369\n",
      "Epoch 10/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0882 - mean_squared_error: 0.0882 - val_loss: 0.0348 - val_mean_squared_error: 0.0348\n",
      "Epoch 11/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0857 - mean_squared_error: 0.0857 - val_loss: 0.0340 - val_mean_squared_error: 0.0340\n",
      "Epoch 12/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0826 - mean_squared_error: 0.0826 - val_loss: 0.0328 - val_mean_squared_error: 0.0328\n",
      "Epoch 13/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0799 - mean_squared_error: 0.0799 - val_loss: 0.0315 - val_mean_squared_error: 0.0315\n",
      "Epoch 14/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0775 - mean_squared_error: 0.0775 - val_loss: 0.0305 - val_mean_squared_error: 0.0304\n",
      "Epoch 15/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0751 - mean_squared_error: 0.0751 - val_loss: 0.0295 - val_mean_squared_error: 0.0295\n",
      "Epoch 16/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0729 - mean_squared_error: 0.0729 - val_loss: 0.0285 - val_mean_squared_error: 0.0285\n",
      "Epoch 17/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0707 - mean_squared_error: 0.0707 - val_loss: 0.0276 - val_mean_squared_error: 0.0276\n",
      "Epoch 18/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0687 - mean_squared_error: 0.0687 - val_loss: 0.0268 - val_mean_squared_error: 0.0268\n",
      "Epoch 19/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0668 - mean_squared_error: 0.0668 - val_loss: 0.0260 - val_mean_squared_error: 0.0260\n",
      "Epoch 20/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0650 - mean_squared_error: 0.0650 - val_loss: 0.0253 - val_mean_squared_error: 0.0253\n",
      "Epoch 21/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0633 - mean_squared_error: 0.0633 - val_loss: 0.0246 - val_mean_squared_error: 0.0246\n",
      "Epoch 22/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0617 - mean_squared_error: 0.0617 - val_loss: 0.0240 - val_mean_squared_error: 0.0239\n",
      "Epoch 23/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0602 - mean_squared_error: 0.0602 - val_loss: 0.0234 - val_mean_squared_error: 0.0234\n",
      "Epoch 24/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0588 - mean_squared_error: 0.0588 - val_loss: 0.0228 - val_mean_squared_error: 0.0228\n",
      "Epoch 25/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0575 - mean_squared_error: 0.0575 - val_loss: 0.0223 - val_mean_squared_error: 0.0223\n",
      "Epoch 26/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0562 - mean_squared_error: 0.0562 - val_loss: 0.0218 - val_mean_squared_error: 0.0218\n",
      "Epoch 27/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0551 - mean_squared_error: 0.0551 - val_loss: 0.0214 - val_mean_squared_error: 0.0214\n",
      "Epoch 28/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0540 - mean_squared_error: 0.0539 - val_loss: 0.0210 - val_mean_squared_error: 0.0210\n",
      "Epoch 29/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0529 - mean_squared_error: 0.0529 - val_loss: 0.0206 - val_mean_squared_error: 0.0206\n",
      "Epoch 30/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0519 - mean_squared_error: 0.0519 - val_loss: 0.0203 - val_mean_squared_error: 0.0202\n",
      "Epoch 31/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0510 - mean_squared_error: 0.0510 - val_loss: 0.0199 - val_mean_squared_error: 0.0199\n",
      "Epoch 32/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0502 - mean_squared_error: 0.0502 - val_loss: 0.0196 - val_mean_squared_error: 0.0196\n",
      "Epoch 33/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0494 - mean_squared_error: 0.0494 - val_loss: 0.0193 - val_mean_squared_error: 0.0193\n",
      "Epoch 34/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0486 - mean_squared_error: 0.0486 - val_loss: 0.0191 - val_mean_squared_error: 0.0191\n",
      "Epoch 35/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0479 - mean_squared_error: 0.0479 - val_loss: 0.0188 - val_mean_squared_error: 0.0188\n",
      "Epoch 36/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0472 - mean_squared_error: 0.0472 - val_loss: 0.0186 - val_mean_squared_error: 0.0186\n",
      "Epoch 37/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0466 - mean_squared_error: 0.0466 - val_loss: 0.0184 - val_mean_squared_error: 0.0184\n",
      "Epoch 38/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0460 - mean_squared_error: 0.0460 - val_loss: 0.0182 - val_mean_squared_error: 0.0182\n",
      "Epoch 39/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0455 - mean_squared_error: 0.0455 - val_loss: 0.0180 - val_mean_squared_error: 0.0180\n",
      "Epoch 40/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0449 - mean_squared_error: 0.0449 - val_loss: 0.0178 - val_mean_squared_error: 0.0178\n",
      "Epoch 41/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0444 - mean_squared_error: 0.0444 - val_loss: 0.0177 - val_mean_squared_error: 0.0177\n",
      "Epoch 42/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0440 - mean_squared_error: 0.0440 - val_loss: 0.0175 - val_mean_squared_error: 0.0175\n",
      "Epoch 43/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0435 - mean_squared_error: 0.0435 - val_loss: 0.0174 - val_mean_squared_error: 0.0174\n",
      "Epoch 44/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0431 - mean_squared_error: 0.0431 - val_loss: 0.0172 - val_mean_squared_error: 0.0172\n",
      "Epoch 45/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0427 - mean_squared_error: 0.0427 - val_loss: 0.0171 - val_mean_squared_error: 0.0171\n",
      "Epoch 46/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0423 - mean_squared_error: 0.0423 - val_loss: 0.0170 - val_mean_squared_error: 0.0170\n",
      "Epoch 47/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0420 - mean_squared_error: 0.0420 - val_loss: 0.0169 - val_mean_squared_error: 0.0169\n",
      "Epoch 48/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0416 - mean_squared_error: 0.0416 - val_loss: 0.0168 - val_mean_squared_error: 0.0168\n",
      "Epoch 49/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0413 - mean_squared_error: 0.0413 - val_loss: 0.0167 - val_mean_squared_error: 0.0166\n",
      "Epoch 50/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0410 - mean_squared_error: 0.0410 - val_loss: 0.0166 - val_mean_squared_error: 0.0166\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 51/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0407 - mean_squared_error: 0.0407 - val_loss: 0.0165 - val_mean_squared_error: 0.0165\n",
      "Epoch 52/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0404 - mean_squared_error: 0.0404 - val_loss: 0.0164 - val_mean_squared_error: 0.0164\n",
      "Epoch 53/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0401 - mean_squared_error: 0.0401 - val_loss: 0.0163 - val_mean_squared_error: 0.0163\n",
      "Epoch 54/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0399 - mean_squared_error: 0.0399 - val_loss: 0.0162 - val_mean_squared_error: 0.0162\n",
      "Epoch 55/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0396 - mean_squared_error: 0.0396 - val_loss: 0.0161 - val_mean_squared_error: 0.0161\n",
      "Epoch 56/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0394 - mean_squared_error: 0.0394 - val_loss: 0.0161 - val_mean_squared_error: 0.0161\n",
      "Epoch 57/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0391 - mean_squared_error: 0.0391 - val_loss: 0.0160 - val_mean_squared_error: 0.0160\n",
      "Epoch 58/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0389 - mean_squared_error: 0.0389 - val_loss: 0.0159 - val_mean_squared_error: 0.0159\n",
      "Epoch 59/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0387 - mean_squared_error: 0.0387 - val_loss: 0.0159 - val_mean_squared_error: 0.0158\n",
      "Epoch 60/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0385 - mean_squared_error: 0.0385 - val_loss: 0.0158 - val_mean_squared_error: 0.0158\n",
      "Epoch 61/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0383 - mean_squared_error: 0.0383 - val_loss: 0.0157 - val_mean_squared_error: 0.0157\n",
      "Epoch 62/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0381 - mean_squared_error: 0.0381 - val_loss: 0.0157 - val_mean_squared_error: 0.0157\n",
      "Epoch 63/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0379 - mean_squared_error: 0.0379 - val_loss: 0.0156 - val_mean_squared_error: 0.0156\n",
      "Epoch 64/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0377 - mean_squared_error: 0.0377 - val_loss: 0.0156 - val_mean_squared_error: 0.0155\n",
      "Epoch 65/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0376 - mean_squared_error: 0.0376 - val_loss: 0.0155 - val_mean_squared_error: 0.0155\n",
      "Epoch 66/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0374 - mean_squared_error: 0.0374 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
      "Epoch 67/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0372 - mean_squared_error: 0.0372 - val_loss: 0.0154 - val_mean_squared_error: 0.0154\n",
      "Epoch 68/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0371 - mean_squared_error: 0.0371 - val_loss: 0.0153 - val_mean_squared_error: 0.0153\n",
      "Epoch 69/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0369 - mean_squared_error: 0.0369 - val_loss: 0.0153 - val_mean_squared_error: 0.0153\n",
      "Epoch 70/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0368 - mean_squared_error: 0.0368 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
      "Epoch 71/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0366 - mean_squared_error: 0.0366 - val_loss: 0.0152 - val_mean_squared_error: 0.0152\n",
      "Epoch 72/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0365 - mean_squared_error: 0.0365 - val_loss: 0.0152 - val_mean_squared_error: 0.0151\n",
      "Epoch 73/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0363 - mean_squared_error: 0.0363 - val_loss: 0.0151 - val_mean_squared_error: 0.0151\n",
      "Epoch 74/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0362 - mean_squared_error: 0.0362 - val_loss: 0.0151 - val_mean_squared_error: 0.0151\n",
      "Epoch 75/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0361 - mean_squared_error: 0.0361 - val_loss: 0.0150 - val_mean_squared_error: 0.0150\n",
      "Epoch 76/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0359 - mean_squared_error: 0.0359 - val_loss: 0.0150 - val_mean_squared_error: 0.0150\n",
      "Epoch 77/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0358 - mean_squared_error: 0.0358 - val_loss: 0.0149 - val_mean_squared_error: 0.0149\n",
      "Epoch 78/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0357 - mean_squared_error: 0.0357 - val_loss: 0.0149 - val_mean_squared_error: 0.0149\n",
      "Epoch 79/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0356 - mean_squared_error: 0.0356 - val_loss: 0.0149 - val_mean_squared_error: 0.0148\n",
      "Epoch 80/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0354 - mean_squared_error: 0.0354 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
      "Epoch 81/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0353 - mean_squared_error: 0.0353 - val_loss: 0.0148 - val_mean_squared_error: 0.0148\n",
      "Epoch 82/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0352 - mean_squared_error: 0.0352 - val_loss: 0.0147 - val_mean_squared_error: 0.0147\n",
      "Epoch 83/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0351 - mean_squared_error: 0.0351 - val_loss: 0.0147 - val_mean_squared_error: 0.0147\n",
      "Epoch 84/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0350 - mean_squared_error: 0.0350 - val_loss: 0.0147 - val_mean_squared_error: 0.0147\n",
      "Epoch 85/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0349 - mean_squared_error: 0.0349 - val_loss: 0.0146 - val_mean_squared_error: 0.0146\n",
      "Epoch 86/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0348 - mean_squared_error: 0.0348 - val_loss: 0.0146 - val_mean_squared_error: 0.0146\n",
      "Epoch 87/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0347 - mean_squared_error: 0.0347 - val_loss: 0.0145 - val_mean_squared_error: 0.0145\n",
      "Epoch 88/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0346 - mean_squared_error: 0.0346 - val_loss: 0.0145 - val_mean_squared_error: 0.0145\n",
      "Epoch 89/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0345 - mean_squared_error: 0.0345 - val_loss: 0.0145 - val_mean_squared_error: 0.0145\n",
      "Epoch 90/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0344 - mean_squared_error: 0.0344 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
      "Epoch 91/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0343 - mean_squared_error: 0.0343 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
      "Epoch 92/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0342 - mean_squared_error: 0.0342 - val_loss: 0.0144 - val_mean_squared_error: 0.0144\n",
      "Epoch 93/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0341 - mean_squared_error: 0.0341 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n",
      "Epoch 94/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0340 - mean_squared_error: 0.0340 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n",
      "Epoch 95/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.0339 - mean_squared_error: 0.0339 - val_loss: 0.0143 - val_mean_squared_error: 0.0143\n",
      "Epoch 96/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0338 - mean_squared_error: 0.0338 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
      "Epoch 97/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0337 - mean_squared_error: 0.0337 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
      "Epoch 98/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0336 - mean_squared_error: 0.0336 - val_loss: 0.0142 - val_mean_squared_error: 0.0142\n",
      "Epoch 99/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0335 - mean_squared_error: 0.0335 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
      "Epoch 100/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0334 - mean_squared_error: 0.0334 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
      "Epoch 101/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0333 - mean_squared_error: 0.0333 - val_loss: 0.0141 - val_mean_squared_error: 0.0141\n",
      "Epoch 102/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0332 - mean_squared_error: 0.0332 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
      "Epoch 103/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0331 - mean_squared_error: 0.0331 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
      "Epoch 104/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0331 - mean_squared_error: 0.0331 - val_loss: 0.0140 - val_mean_squared_error: 0.0140\n",
      "Epoch 105/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0330 - mean_squared_error: 0.0330 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
      "Epoch 106/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0329 - mean_squared_error: 0.0329 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
      "Epoch 107/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0328 - mean_squared_error: 0.0328 - val_loss: 0.0139 - val_mean_squared_error: 0.0139\n",
      "Epoch 108/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0327 - mean_squared_error: 0.0327 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
      "Epoch 109/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0326 - mean_squared_error: 0.0326 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
      "Epoch 110/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0326 - mean_squared_error: 0.0325 - val_loss: 0.0138 - val_mean_squared_error: 0.0138\n",
      "Epoch 111/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0325 - mean_squared_error: 0.0325 - val_loss: 0.0137 - val_mean_squared_error: 0.0137\n",
      "Epoch 112/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0324 - mean_squared_error: 0.0324 - val_loss: 0.0137 - val_mean_squared_error: 0.0137\n",
      "Epoch 113/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0323 - mean_squared_error: 0.0323 - val_loss: 0.0137 - val_mean_squared_error: 0.0137\n",
      "Epoch 114/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0322 - mean_squared_error: 0.0322 - val_loss: 0.0137 - val_mean_squared_error: 0.0136\n",
      "Epoch 115/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0321 - mean_squared_error: 0.0321 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
      "Epoch 116/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0321 - mean_squared_error: 0.0321 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
      "Epoch 117/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0320 - mean_squared_error: 0.0320 - val_loss: 0.0136 - val_mean_squared_error: 0.0136\n",
      "Epoch 118/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0319 - mean_squared_error: 0.0319 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
      "Epoch 119/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0318 - mean_squared_error: 0.0318 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
      "Epoch 120/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0318 - mean_squared_error: 0.0317 - val_loss: 0.0135 - val_mean_squared_error: 0.0135\n",
      "Epoch 121/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0317 - mean_squared_error: 0.0317 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
      "Epoch 122/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0316 - mean_squared_error: 0.0316 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
      "Epoch 123/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0315 - mean_squared_error: 0.0315 - val_loss: 0.0134 - val_mean_squared_error: 0.0134\n",
      "Epoch 124/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0314 - mean_squared_error: 0.0314 - val_loss: 0.0134 - val_mean_squared_error: 0.0133\n",
      "Epoch 125/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0314 - mean_squared_error: 0.0314 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
      "Epoch 126/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0313 - mean_squared_error: 0.0313 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
      "Epoch 127/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0312 - mean_squared_error: 0.0312 - val_loss: 0.0133 - val_mean_squared_error: 0.0133\n",
      "Epoch 128/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0311 - mean_squared_error: 0.0311 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
      "Epoch 129/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0311 - mean_squared_error: 0.0311 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
      "Epoch 130/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0310 - mean_squared_error: 0.0310 - val_loss: 0.0132 - val_mean_squared_error: 0.0132\n",
      "Epoch 131/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0309 - mean_squared_error: 0.0309 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
      "Epoch 132/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0308 - mean_squared_error: 0.0308 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
      "Epoch 133/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0308 - mean_squared_error: 0.0308 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
      "Epoch 134/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0307 - mean_squared_error: 0.0307 - val_loss: 0.0131 - val_mean_squared_error: 0.0131\n",
      "Epoch 135/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0306 - mean_squared_error: 0.0306 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
      "Epoch 136/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0305 - mean_squared_error: 0.0305 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
      "Epoch 137/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0305 - mean_squared_error: 0.0305 - val_loss: 0.0130 - val_mean_squared_error: 0.0130\n",
      "Epoch 138/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0304 - mean_squared_error: 0.0304 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
      "Epoch 139/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0303 - mean_squared_error: 0.0303 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
      "Epoch 140/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0302 - mean_squared_error: 0.0302 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
      "Epoch 141/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0302 - mean_squared_error: 0.0302 - val_loss: 0.0129 - val_mean_squared_error: 0.0129\n",
      "Epoch 142/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0301 - mean_squared_error: 0.0301 - val_loss: 0.0128 - val_mean_squared_error: 0.0128\n",
      "Epoch 143/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0300 - mean_squared_error: 0.0300 - val_loss: 0.0128 - val_mean_squared_error: 0.0128\n",
      "Epoch 144/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0299 - mean_squared_error: 0.0299 - val_loss: 0.0128 - val_mean_squared_error: 0.0128\n",
      "Epoch 145/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0299 - mean_squared_error: 0.0299 - val_loss: 0.0128 - val_mean_squared_error: 0.0127\n",
      "Epoch 146/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0298 - mean_squared_error: 0.0298 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
      "Epoch 147/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0297 - mean_squared_error: 0.0297 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
      "Epoch 148/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0297 - mean_squared_error: 0.0296 - val_loss: 0.0127 - val_mean_squared_error: 0.0127\n",
      "Epoch 149/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0296 - mean_squared_error: 0.0296 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
      "Epoch 150/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0295 - mean_squared_error: 0.0295 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
      "Epoch 151/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0294 - mean_squared_error: 0.0294 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
      "Epoch 152/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0294 - mean_squared_error: 0.0294 - val_loss: 0.0126 - val_mean_squared_error: 0.0126\n",
      "Epoch 153/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0293 - mean_squared_error: 0.0293 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
      "Epoch 154/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0292 - mean_squared_error: 0.0292 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
      "Epoch 155/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0291 - mean_squared_error: 0.0291 - val_loss: 0.0125 - val_mean_squared_error: 0.0125\n",
      "Epoch 156/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0291 - mean_squared_error: 0.0291 - val_loss: 0.0125 - val_mean_squared_error: 0.0124\n",
      "Epoch 157/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0290 - mean_squared_error: 0.0290 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
      "Epoch 158/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0289 - mean_squared_error: 0.0289 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
      "Epoch 159/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0288 - mean_squared_error: 0.0288 - val_loss: 0.0124 - val_mean_squared_error: 0.0124\n",
      "Epoch 160/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0288 - mean_squared_error: 0.0288 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
      "Epoch 161/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0287 - mean_squared_error: 0.0287 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
      "Epoch 162/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0286 - mean_squared_error: 0.0286 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
      "Epoch 163/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0286 - mean_squared_error: 0.0286 - val_loss: 0.0123 - val_mean_squared_error: 0.0123\n",
      "Epoch 164/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0285 - mean_squared_error: 0.0285 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
      "Epoch 165/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0284 - mean_squared_error: 0.0284 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
      "Epoch 166/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0283 - mean_squared_error: 0.0283 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
      "Epoch 167/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0283 - mean_squared_error: 0.0283 - val_loss: 0.0122 - val_mean_squared_error: 0.0122\n",
      "Epoch 168/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0282 - mean_squared_error: 0.0282 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
      "Epoch 169/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0281 - mean_squared_error: 0.0281 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
      "Epoch 170/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0280 - mean_squared_error: 0.0280 - val_loss: 0.0121 - val_mean_squared_error: 0.0121\n",
      "Epoch 171/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0280 - mean_squared_error: 0.0280 - val_loss: 0.0121 - val_mean_squared_error: 0.0120\n",
      "Epoch 172/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0279 - mean_squared_error: 0.0279 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
      "Epoch 173/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0278 - mean_squared_error: 0.0278 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
      "Epoch 174/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0278 - mean_squared_error: 0.0277 - val_loss: 0.0120 - val_mean_squared_error: 0.0120\n",
      "Epoch 175/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0277 - mean_squared_error: 0.0277 - val_loss: 0.0120 - val_mean_squared_error: 0.0119\n",
      "Epoch 176/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0276 - mean_squared_error: 0.0276 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
      "Epoch 177/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0275 - mean_squared_error: 0.0275 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
      "Epoch 178/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0275 - mean_squared_error: 0.0275 - val_loss: 0.0119 - val_mean_squared_error: 0.0119\n",
      "Epoch 179/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0274 - mean_squared_error: 0.0274 - val_loss: 0.0119 - val_mean_squared_error: 0.0118\n",
      "Epoch 180/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0273 - mean_squared_error: 0.0273 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
      "Epoch 181/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0272 - mean_squared_error: 0.0272 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
      "Epoch 182/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0272 - mean_squared_error: 0.0272 - val_loss: 0.0118 - val_mean_squared_error: 0.0118\n",
      "Epoch 183/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0271 - mean_squared_error: 0.0271 - val_loss: 0.0118 - val_mean_squared_error: 0.0117\n",
      "Epoch 184/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0270 - mean_squared_error: 0.0270 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
      "Epoch 185/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0269 - mean_squared_error: 0.0269 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
      "Epoch 186/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0269 - mean_squared_error: 0.0269 - val_loss: 0.0117 - val_mean_squared_error: 0.0117\n",
      "Epoch 187/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0268 - mean_squared_error: 0.0268 - val_loss: 0.0117 - val_mean_squared_error: 0.0116\n",
      "Epoch 188/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0267 - mean_squared_error: 0.0267 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
      "Epoch 189/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0267 - mean_squared_error: 0.0266 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
      "Epoch 190/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0266 - mean_squared_error: 0.0266 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
      "Epoch 191/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0265 - mean_squared_error: 0.0265 - val_loss: 0.0116 - val_mean_squared_error: 0.0116\n",
      "Epoch 192/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0264 - mean_squared_error: 0.0264 - val_loss: 0.0115 - val_mean_squared_error: 0.0115\n",
      "Epoch 193/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0264 - mean_squared_error: 0.0263 - val_loss: 0.0115 - val_mean_squared_error: 0.0115\n",
      "Epoch 194/10000\n",
      "7/7 [==============================] - 0s 11ms/step - loss: 0.0263 - mean_squared_error: 0.0263 - val_loss: 0.0115 - val_mean_squared_error: 0.0115\n",
      "Epoch 195/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0262 - mean_squared_error: 0.0262 - val_loss: 0.0115 - val_mean_squared_error: 0.0115\n",
      "Epoch 196/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0261 - mean_squared_error: 0.0261 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
      "Epoch 197/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0261 - mean_squared_error: 0.0260 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
      "Epoch 198/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0260 - mean_squared_error: 0.0260 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
      "Epoch 199/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0259 - mean_squared_error: 0.0259 - val_loss: 0.0114 - val_mean_squared_error: 0.0114\n",
      "Epoch 200/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0258 - mean_squared_error: 0.0258 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
      "Epoch 201/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0258 - mean_squared_error: 0.0257 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
      "Epoch 202/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0257 - mean_squared_error: 0.0257 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
      "Epoch 203/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0256 - mean_squared_error: 0.0256 - val_loss: 0.0113 - val_mean_squared_error: 0.0113\n",
      "Epoch 204/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0255 - mean_squared_error: 0.0255 - val_loss: 0.0113 - val_mean_squared_error: 0.0112\n",
      "Epoch 205/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0255 - mean_squared_error: 0.0254 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
      "Epoch 206/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0254 - mean_squared_error: 0.0254 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
      "Epoch 207/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0253 - mean_squared_error: 0.0253 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
      "Epoch 208/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0252 - mean_squared_error: 0.0252 - val_loss: 0.0112 - val_mean_squared_error: 0.0112\n",
      "Epoch 209/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0252 - mean_squared_error: 0.0251 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
      "Epoch 210/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0251 - mean_squared_error: 0.0251 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
      "Epoch 211/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0250 - mean_squared_error: 0.0250 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
      "Epoch 212/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0249 - mean_squared_error: 0.0249 - val_loss: 0.0111 - val_mean_squared_error: 0.0111\n",
      "Epoch 213/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0249 - mean_squared_error: 0.0248 - val_loss: 0.0111 - val_mean_squared_error: 0.0110\n",
      "Epoch 214/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0248 - mean_squared_error: 0.0248 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
      "Epoch 215/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0247 - mean_squared_error: 0.0247 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
      "Epoch 216/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0246 - mean_squared_error: 0.0246 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
      "Epoch 217/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0245 - mean_squared_error: 0.0245 - val_loss: 0.0110 - val_mean_squared_error: 0.0110\n",
      "Epoch 218/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0245 - mean_squared_error: 0.0245 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
      "Epoch 219/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0244 - mean_squared_error: 0.0244 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
      "Epoch 220/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0243 - mean_squared_error: 0.0243 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
      "Epoch 221/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0242 - mean_squared_error: 0.0242 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
      "Epoch 222/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0242 - mean_squared_error: 0.0242 - val_loss: 0.0109 - val_mean_squared_error: 0.0109\n",
      "Epoch 223/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0241 - mean_squared_error: 0.0241 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 224/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0240 - mean_squared_error: 0.0240 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 225/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0239 - mean_squared_error: 0.0239 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 226/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0239 - mean_squared_error: 0.0239 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 227/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0238 - mean_squared_error: 0.0238 - val_loss: 0.0108 - val_mean_squared_error: 0.0108\n",
      "Epoch 228/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0237 - mean_squared_error: 0.0237 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
      "Epoch 229/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0236 - mean_squared_error: 0.0236 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
      "Epoch 230/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0236 - mean_squared_error: 0.0235 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
      "Epoch 231/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0235 - mean_squared_error: 0.0235 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
      "Epoch 232/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0234 - mean_squared_error: 0.0234 - val_loss: 0.0107 - val_mean_squared_error: 0.0107\n",
      "Epoch 233/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0233 - mean_squared_error: 0.0233 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
      "Epoch 234/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0232 - mean_squared_error: 0.0232 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
      "Epoch 235/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0232 - mean_squared_error: 0.0232 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
      "Epoch 236/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0231 - mean_squared_error: 0.0231 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
      "Epoch 237/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0230 - mean_squared_error: 0.0230 - val_loss: 0.0106 - val_mean_squared_error: 0.0106\n",
      "Epoch 238/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0229 - mean_squared_error: 0.0229 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 239/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0229 - mean_squared_error: 0.0228 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 240/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0228 - mean_squared_error: 0.0228 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 241/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0227 - mean_squared_error: 0.0227 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 242/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0226 - mean_squared_error: 0.0226 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 243/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0225 - mean_squared_error: 0.0225 - val_loss: 0.0105 - val_mean_squared_error: 0.0105\n",
      "Epoch 244/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0225 - mean_squared_error: 0.0225 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 245/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0224 - mean_squared_error: 0.0224 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 246/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0223 - mean_squared_error: 0.0223 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 247/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0222 - mean_squared_error: 0.0222 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 248/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0222 - mean_squared_error: 0.0222 - val_loss: 0.0104 - val_mean_squared_error: 0.0104\n",
      "Epoch 249/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0221 - mean_squared_error: 0.0221 - val_loss: 0.0104 - val_mean_squared_error: 0.0103\n",
      "Epoch 250/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0220 - mean_squared_error: 0.0220 - val_loss: 0.0103 - val_mean_squared_error: 0.0103\n",
      "Epoch 251/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0219 - mean_squared_error: 0.0219 - val_loss: 0.0103 - val_mean_squared_error: 0.0103\n",
      "Epoch 252/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0218 - mean_squared_error: 0.0218 - val_loss: 0.0103 - val_mean_squared_error: 0.0103\n",
      "Epoch 253/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0218 - mean_squared_error: 0.0218 - val_loss: 0.0103 - val_mean_squared_error: 0.0103\n",
      "Epoch 254/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0217 - mean_squared_error: 0.0217 - val_loss: 0.0103 - val_mean_squared_error: 0.0103\n",
      "Epoch 255/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0216 - mean_squared_error: 0.0216 - val_loss: 0.0103 - val_mean_squared_error: 0.0102\n",
      "Epoch 256/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0215 - mean_squared_error: 0.0215 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 257/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0215 - mean_squared_error: 0.0215 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 258/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0214 - mean_squared_error: 0.0214 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 259/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0213 - mean_squared_error: 0.0213 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 260/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0212 - mean_squared_error: 0.0212 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 261/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0212 - mean_squared_error: 0.0211 - val_loss: 0.0102 - val_mean_squared_error: 0.0102\n",
      "Epoch 262/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0211 - mean_squared_error: 0.0211 - val_loss: 0.0102 - val_mean_squared_error: 0.0101\n",
      "Epoch 263/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0210 - mean_squared_error: 0.0210 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 264/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0209 - mean_squared_error: 0.0209 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 265/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0208 - mean_squared_error: 0.0208 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 266/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0208 - mean_squared_error: 0.0208 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 267/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0207 - mean_squared_error: 0.0207 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 268/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0206 - mean_squared_error: 0.0206 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 269/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0205 - mean_squared_error: 0.0205 - val_loss: 0.0101 - val_mean_squared_error: 0.0100\n",
      "Epoch 270/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0205 - mean_squared_error: 0.0205 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 271/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0204 - mean_squared_error: 0.0204 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 272/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0203 - mean_squared_error: 0.0203 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 273/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0202 - mean_squared_error: 0.0202 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 274/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0202 - mean_squared_error: 0.0201 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 275/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0201 - mean_squared_error: 0.0201 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 276/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0200 - mean_squared_error: 0.0200 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 277/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0199 - mean_squared_error: 0.0199 - val_loss: 0.0100 - val_mean_squared_error: 0.0099\n",
      "Epoch 278/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0198 - mean_squared_error: 0.0198 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 279/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0198 - mean_squared_error: 0.0198 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 280/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0197 - mean_squared_error: 0.0197 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 281/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0196 - mean_squared_error: 0.0196 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 282/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0195 - mean_squared_error: 0.0195 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 283/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0195 - mean_squared_error: 0.0195 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 284/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0194 - mean_squared_error: 0.0194 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 285/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0193 - mean_squared_error: 0.0193 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 286/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0192 - mean_squared_error: 0.0192 - val_loss: 0.0099 - val_mean_squared_error: 0.0098\n",
      "Epoch 287/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0192 - mean_squared_error: 0.0192 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 288/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.0191 - mean_squared_error: 0.0191 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 289/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0190 - mean_squared_error: 0.0190 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 290/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0189 - mean_squared_error: 0.0189 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 291/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0189 - mean_squared_error: 0.0189 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 292/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0188 - mean_squared_error: 0.0188 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 293/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0187 - mean_squared_error: 0.0187 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 294/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0186 - mean_squared_error: 0.0186 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 295/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0186 - mean_squared_error: 0.0186 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 296/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0185 - mean_squared_error: 0.0185 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 297/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0184 - mean_squared_error: 0.0184 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 298/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0183 - mean_squared_error: 0.0183 - val_loss: 0.0098 - val_mean_squared_error: 0.0097\n",
      "Epoch 299/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0183 - mean_squared_error: 0.0183 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 300/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0182 - mean_squared_error: 0.0182 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 301/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0181 - mean_squared_error: 0.0181 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 302/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0181 - mean_squared_error: 0.0180 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 303/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0180 - mean_squared_error: 0.0180 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 304/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0179 - mean_squared_error: 0.0179 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 305/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.0178 - mean_squared_error: 0.0178 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 306/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0178 - mean_squared_error: 0.0178 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 307/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0177 - mean_squared_error: 0.0177 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 308/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0176 - mean_squared_error: 0.0176 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 309/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0176 - mean_squared_error: 0.0175 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 310/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0175 - mean_squared_error: 0.0175 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 311/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0174 - mean_squared_error: 0.0174 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 312/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0173 - mean_squared_error: 0.0173 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 313/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0173 - mean_squared_error: 0.0173 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 314/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0172 - mean_squared_error: 0.0172 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 315/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0171 - mean_squared_error: 0.0171 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 316/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0171 - mean_squared_error: 0.0171 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 317/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0170 - mean_squared_error: 0.0170 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 318/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0169 - mean_squared_error: 0.0169 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 319/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0169 - mean_squared_error: 0.0168 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 320/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0168 - mean_squared_error: 0.0168 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 321/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0167 - mean_squared_error: 0.0167 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 322/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0167 - mean_squared_error: 0.0166 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 323/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0166 - mean_squared_error: 0.0166 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 324/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0165 - mean_squared_error: 0.0165 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 325/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0164 - mean_squared_error: 0.0164 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 326/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0164 - mean_squared_error: 0.0164 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 327/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0163 - mean_squared_error: 0.0163 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 328/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.0162 - mean_squared_error: 0.0162 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 329/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0162 - mean_squared_error: 0.0162 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 330/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0161 - mean_squared_error: 0.0161 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 331/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0161 - mean_squared_error: 0.0160 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 332/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0160 - mean_squared_error: 0.0160 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 333/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0159 - mean_squared_error: 0.0159 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 334/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0159 - mean_squared_error: 0.0159 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 335/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0158 - mean_squared_error: 0.0158 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 336/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0157 - mean_squared_error: 0.0157 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 337/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0157 - mean_squared_error: 0.0157 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 338/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0156 - mean_squared_error: 0.0156 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 339/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0155 - mean_squared_error: 0.0155 - val_loss: 0.0096 - val_mean_squared_error: 0.0096\n",
      "Epoch 340/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0155 - mean_squared_error: 0.0155 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 341/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0154 - mean_squared_error: 0.0154 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 342/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0154 - mean_squared_error: 0.0153 - val_loss: 0.0097 - val_mean_squared_error: 0.0096\n",
      "Epoch 343/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0153 - mean_squared_error: 0.0153 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 344/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0152 - mean_squared_error: 0.0152 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 345/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0152 - mean_squared_error: 0.0152 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 346/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0151 - mean_squared_error: 0.0151 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 347/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0151 - mean_squared_error: 0.0150 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 348/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0150 - mean_squared_error: 0.0150 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 349/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0149 - mean_squared_error: 0.0149 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 350/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0149 - mean_squared_error: 0.0149 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 351/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0148 - mean_squared_error: 0.0148 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 352/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0148 - mean_squared_error: 0.0148 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 353/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0147 - mean_squared_error: 0.0147 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 354/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0146 - mean_squared_error: 0.0146 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 355/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0146 - mean_squared_error: 0.0146 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 356/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0145 - mean_squared_error: 0.0145 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 357/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0145 - mean_squared_error: 0.0145 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 358/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0144 - mean_squared_error: 0.0144 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 359/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0144 - mean_squared_error: 0.0144 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 360/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0143 - mean_squared_error: 0.0143 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 361/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0143 - mean_squared_error: 0.0142 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 362/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0142 - mean_squared_error: 0.0142 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 363/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0141 - mean_squared_error: 0.0141 - val_loss: 0.0097 - val_mean_squared_error: 0.0097\n",
      "Epoch 364/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0141 - mean_squared_error: 0.0141 - val_loss: 0.0098 - val_mean_squared_error: 0.0097\n",
      "Epoch 365/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0140 - mean_squared_error: 0.0140 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 366/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0140 - mean_squared_error: 0.0140 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 367/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 368/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0139 - mean_squared_error: 0.0139 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 369/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0138 - mean_squared_error: 0.0138 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 370/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0138 - mean_squared_error: 0.0138 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 371/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0137 - mean_squared_error: 0.0137 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 372/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0137 - mean_squared_error: 0.0137 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 373/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0136 - mean_squared_error: 0.0136 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 374/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0136 - mean_squared_error: 0.0136 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 375/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 376/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0135 - mean_squared_error: 0.0135 - val_loss: 0.0098 - val_mean_squared_error: 0.0098\n",
      "Epoch 377/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0134 - mean_squared_error: 0.0134 - val_loss: 0.0099 - val_mean_squared_error: 0.0098\n",
      "Epoch 378/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0134 - mean_squared_error: 0.0134 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 379/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0133 - mean_squared_error: 0.0133 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 380/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.0133 - mean_squared_error: 0.0133 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 381/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0132 - mean_squared_error: 0.0132 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 382/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0132 - mean_squared_error: 0.0132 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 383/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.0131 - mean_squared_error: 0.0131 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 384/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0131 - mean_squared_error: 0.0131 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 385/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0131 - mean_squared_error: 0.0130 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 386/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0130 - mean_squared_error: 0.0130 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 387/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0130 - mean_squared_error: 0.0130 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 388/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0129 - mean_squared_error: 0.0129 - val_loss: 0.0099 - val_mean_squared_error: 0.0099\n",
      "Epoch 389/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0129 - mean_squared_error: 0.0129 - val_loss: 0.0100 - val_mean_squared_error: 0.0099\n",
      "Epoch 390/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0128 - mean_squared_error: 0.0128 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 391/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0128 - mean_squared_error: 0.0128 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 392/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0127 - mean_squared_error: 0.0127 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 393/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0127 - mean_squared_error: 0.0127 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 394/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0127 - mean_squared_error: 0.0127 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 395/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0126 - mean_squared_error: 0.0126 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 396/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0126 - mean_squared_error: 0.0126 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 397/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 398/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0125 - mean_squared_error: 0.0125 - val_loss: 0.0100 - val_mean_squared_error: 0.0100\n",
      "Epoch 399/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0125 - mean_squared_error: 0.0124 - val_loss: 0.0101 - val_mean_squared_error: 0.0100\n",
      "Epoch 400/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 401/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0124 - mean_squared_error: 0.0124 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 402/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0123 - mean_squared_error: 0.0123 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 403/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0123 - mean_squared_error: 0.0123 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 404/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0123 - mean_squared_error: 0.0122 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 405/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0122 - mean_squared_error: 0.0122 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 406/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.0122 - mean_squared_error: 0.0122 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n",
      "Epoch 407/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.0121 - mean_squared_error: 0.0121 - val_loss: 0.0101 - val_mean_squared_error: 0.0101\n"
     ]
    }
   ],
   "source": [
    "uncompiled_model = get_uncomplied_model(input_dim, hidden_layer_1_dim, hidden_layer_2_dim)\n",
    "model = get_compiled_model(uncompiled_model, learning_rate)\n",
    "\n",
    "model_saver = ModelCheckpoint(os.path.join(\"saved_models\", \"save_at_{epoch}.h5\"), period=100)\n",
    "early_stopper = tf.keras.callbacks.EarlyStopping(monitor='val_mean_squared_error', \n",
    "                                                 min_delta=10e-5, \n",
    "                                                 restore_best_weights=True, \n",
    "                                                 patience=100)\n",
    "callbacks = [model_saver, early_stopper]\n",
    "\n",
    "history = model.fit(inputs_training, targets_training,\n",
    "                    epochs=epochs, \n",
    "                    batch_size=batch_size, \n",
    "                    verbose=1, \n",
    "                    validation_data=(inputs_testing, targets_testing), \n",
    "                    shuffle=False,\n",
    "                    callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "tender-adolescent",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    " show_curves(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "assisted-monday",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 1ms/step - loss: 0.0175 - mean_squared_error: 0.0175\n",
      "Test MSE 0.0175\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results = model.evaluate(inputs_evaluation, targets_evaluation)\n",
    "print(f\"Test MSE {results[1] :.4f}\")\n",
    "pred_target_evaluation = model.predict(inputs_evaluation)\n",
    "plt.plot(pred_target_evaluation, color=\"b\", label=\"predicted\")\n",
    "plt.plot(targets_evaluation, color=\"r\", label=\"true\")\n",
    "plt.title(\"Mackey-Glass time series\")\n",
    "plt.legend(loc=\"upper right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "honest-starter",
   "metadata": {},
   "source": [
    "Noise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "serious-observation",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      " Training samples: 700 \n",
      " Testing samples: 300 \n",
      " Evaluation samples: 200\n",
      "(700, 5)\n"
     ]
    }
   ],
   "source": [
    "inputs_noisy_training, targets_noisy_training, \\\n",
    "    inputs_noisy_testing, targets_noisy_testing, \\\n",
    "    inputs_noisy_evaluation, targets_noisy_evaluation = \\\n",
    "        data_split(training_size=700, inputs=inputs_noisy, targets=targets_noisy)\n",
    "\n",
    "input_dim = inputs_noisy_training.shape[1]\n",
    "hidden_layer_1_dim = 5\n",
    "hidden_layer_2_dim = 3\n",
    "learning_rate = 10e-3\n",
    "batch_size = 100\n",
    "epochs = 10000\n",
    "\n",
    "print(inputs_noisy_training.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "honest-macedonia",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:`period` argument is deprecated. Please use `save_freq` to specify the frequency in number of batches seen.\n",
      "Epoch 1/10000\n",
      "7/7 [==============================] - 1s 27ms/step - loss: 2.0985 - mean_squared_error: 2.0985 - val_loss: 1.2403 - val_mean_squared_error: 1.2403\n",
      "Epoch 2/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0723 - mean_squared_error: 1.0723 - val_loss: 1.0878 - val_mean_squared_error: 1.0878\n",
      "Epoch 3/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.1342 - mean_squared_error: 1.1342 - val_loss: 1.1067 - val_mean_squared_error: 1.1067\n",
      "Epoch 4/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.1089 - mean_squared_error: 1.1089 - val_loss: 1.0379 - val_mean_squared_error: 1.0379\n",
      "Epoch 5/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0296 - mean_squared_error: 1.0296 - val_loss: 1.0494 - val_mean_squared_error: 1.0494\n",
      "Epoch 6/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0226 - mean_squared_error: 1.0226 - val_loss: 1.0466 - val_mean_squared_error: 1.0466\n",
      "Epoch 7/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0195 - mean_squared_error: 1.0195 - val_loss: 1.0383 - val_mean_squared_error: 1.0382\n",
      "Epoch 8/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0220 - mean_squared_error: 1.0220 - val_loss: 1.0369 - val_mean_squared_error: 1.0369\n",
      "Epoch 9/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0250 - mean_squared_error: 1.0250 - val_loss: 1.0371 - val_mean_squared_error: 1.0371\n",
      "Epoch 10/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0238 - mean_squared_error: 1.0238 - val_loss: 1.0377 - val_mean_squared_error: 1.0377\n",
      "Epoch 11/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0222 - mean_squared_error: 1.0222 - val_loss: 1.0381 - val_mean_squared_error: 1.0381\n",
      "Epoch 12/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0216 - mean_squared_error: 1.0216 - val_loss: 1.0379 - val_mean_squared_error: 1.0379\n",
      "Epoch 13/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0218 - mean_squared_error: 1.0218 - val_loss: 1.0376 - val_mean_squared_error: 1.0376\n",
      "Epoch 14/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0221 - mean_squared_error: 1.0221 - val_loss: 1.0375 - val_mean_squared_error: 1.0375\n",
      "Epoch 15/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0221 - mean_squared_error: 1.0221 - val_loss: 1.0375 - val_mean_squared_error: 1.0375\n",
      "Epoch 16/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0220 - mean_squared_error: 1.0220 - val_loss: 1.0376 - val_mean_squared_error: 1.0376\n",
      "Epoch 17/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0218 - mean_squared_error: 1.0218 - val_loss: 1.0375 - val_mean_squared_error: 1.0375\n",
      "Epoch 18/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0218 - mean_squared_error: 1.0218 - val_loss: 1.0375 - val_mean_squared_error: 1.0375\n",
      "Epoch 19/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0217 - mean_squared_error: 1.0217 - val_loss: 1.0374 - val_mean_squared_error: 1.0374\n",
      "Epoch 20/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0217 - mean_squared_error: 1.0217 - val_loss: 1.0374 - val_mean_squared_error: 1.0374\n",
      "Epoch 21/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0216 - mean_squared_error: 1.0216 - val_loss: 1.0374 - val_mean_squared_error: 1.0374\n",
      "Epoch 22/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0216 - mean_squared_error: 1.0216 - val_loss: 1.0374 - val_mean_squared_error: 1.0374\n",
      "Epoch 23/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0215 - mean_squared_error: 1.0215 - val_loss: 1.0373 - val_mean_squared_error: 1.0373\n",
      "Epoch 24/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0214 - mean_squared_error: 1.0214 - val_loss: 1.0373 - val_mean_squared_error: 1.0373\n",
      "Epoch 25/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0214 - mean_squared_error: 1.0214 - val_loss: 1.0373 - val_mean_squared_error: 1.0373\n",
      "Epoch 26/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0213 - mean_squared_error: 1.0213 - val_loss: 1.0372 - val_mean_squared_error: 1.0372\n",
      "Epoch 27/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0213 - mean_squared_error: 1.0213 - val_loss: 1.0372 - val_mean_squared_error: 1.0372\n",
      "Epoch 28/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0212 - mean_squared_error: 1.0212 - val_loss: 1.0372 - val_mean_squared_error: 1.0372\n",
      "Epoch 29/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0211 - mean_squared_error: 1.0211 - val_loss: 1.0372 - val_mean_squared_error: 1.0372\n",
      "Epoch 30/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0211 - mean_squared_error: 1.0211 - val_loss: 1.0371 - val_mean_squared_error: 1.0371\n",
      "Epoch 31/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0210 - mean_squared_error: 1.0210 - val_loss: 1.0371 - val_mean_squared_error: 1.0371\n",
      "Epoch 32/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0210 - mean_squared_error: 1.0210 - val_loss: 1.0371 - val_mean_squared_error: 1.0371\n",
      "Epoch 33/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0209 - mean_squared_error: 1.0209 - val_loss: 1.0370 - val_mean_squared_error: 1.0370\n",
      "Epoch 34/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0209 - mean_squared_error: 1.0209 - val_loss: 1.0370 - val_mean_squared_error: 1.0370\n",
      "Epoch 35/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0208 - mean_squared_error: 1.0208 - val_loss: 1.0370 - val_mean_squared_error: 1.0370\n",
      "Epoch 36/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0207 - mean_squared_error: 1.0207 - val_loss: 1.0370 - val_mean_squared_error: 1.0370\n",
      "Epoch 37/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0207 - mean_squared_error: 1.0207 - val_loss: 1.0369 - val_mean_squared_error: 1.0369\n",
      "Epoch 38/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0206 - mean_squared_error: 1.0206 - val_loss: 1.0369 - val_mean_squared_error: 1.0369\n",
      "Epoch 39/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0206 - mean_squared_error: 1.0206 - val_loss: 1.0369 - val_mean_squared_error: 1.0369\n",
      "Epoch 40/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0205 - mean_squared_error: 1.0205 - val_loss: 1.0368 - val_mean_squared_error: 1.0368\n",
      "Epoch 41/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0204 - mean_squared_error: 1.0204 - val_loss: 1.0368 - val_mean_squared_error: 1.0368\n",
      "Epoch 42/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0204 - mean_squared_error: 1.0204 - val_loss: 1.0368 - val_mean_squared_error: 1.0368\n",
      "Epoch 43/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0203 - mean_squared_error: 1.0203 - val_loss: 1.0367 - val_mean_squared_error: 1.0367\n",
      "Epoch 44/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0203 - mean_squared_error: 1.0203 - val_loss: 1.0367 - val_mean_squared_error: 1.0367\n",
      "Epoch 45/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0202 - mean_squared_error: 1.0202 - val_loss: 1.0367 - val_mean_squared_error: 1.0367\n",
      "Epoch 46/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0201 - mean_squared_error: 1.0201 - val_loss: 1.0367 - val_mean_squared_error: 1.0367\n",
      "Epoch 47/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0201 - mean_squared_error: 1.0201 - val_loss: 1.0366 - val_mean_squared_error: 1.0366\n",
      "Epoch 48/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0200 - mean_squared_error: 1.0200 - val_loss: 1.0366 - val_mean_squared_error: 1.0366\n",
      "Epoch 49/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0200 - mean_squared_error: 1.0199 - val_loss: 1.0366 - val_mean_squared_error: 1.0366\n",
      "Epoch 50/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0199 - mean_squared_error: 1.0199 - val_loss: 1.0365 - val_mean_squared_error: 1.0365\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 51/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 1.0198 - mean_squared_error: 1.0198 - val_loss: 1.0365 - val_mean_squared_error: 1.0365\n",
      "Epoch 52/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0198 - mean_squared_error: 1.0198 - val_loss: 1.0365 - val_mean_squared_error: 1.0365\n",
      "Epoch 53/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0197 - mean_squared_error: 1.0197 - val_loss: 1.0364 - val_mean_squared_error: 1.0364\n",
      "Epoch 54/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0196 - mean_squared_error: 1.0196 - val_loss: 1.0364 - val_mean_squared_error: 1.0364\n",
      "Epoch 55/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0196 - mean_squared_error: 1.0196 - val_loss: 1.0364 - val_mean_squared_error: 1.0364\n",
      "Epoch 56/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0195 - mean_squared_error: 1.0195 - val_loss: 1.0363 - val_mean_squared_error: 1.0363\n",
      "Epoch 57/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0194 - mean_squared_error: 1.0194 - val_loss: 1.0363 - val_mean_squared_error: 1.0363\n",
      "Epoch 58/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0194 - mean_squared_error: 1.0194 - val_loss: 1.0363 - val_mean_squared_error: 1.0363\n",
      "Epoch 59/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0193 - mean_squared_error: 1.0193 - val_loss: 1.0362 - val_mean_squared_error: 1.0362\n",
      "Epoch 60/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0192 - mean_squared_error: 1.0192 - val_loss: 1.0362 - val_mean_squared_error: 1.0362\n",
      "Epoch 61/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0191 - mean_squared_error: 1.0191 - val_loss: 1.0362 - val_mean_squared_error: 1.0362\n",
      "Epoch 62/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0191 - mean_squared_error: 1.0191 - val_loss: 1.0361 - val_mean_squared_error: 1.0361\n",
      "Epoch 63/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0190 - mean_squared_error: 1.0190 - val_loss: 1.0361 - val_mean_squared_error: 1.0361\n",
      "Epoch 64/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0189 - mean_squared_error: 1.0189 - val_loss: 1.0360 - val_mean_squared_error: 1.0360\n",
      "Epoch 65/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0189 - mean_squared_error: 1.0188 - val_loss: 1.0360 - val_mean_squared_error: 1.0360\n",
      "Epoch 66/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 1.0188 - mean_squared_error: 1.0188 - val_loss: 1.0360 - val_mean_squared_error: 1.0360\n",
      "Epoch 67/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0187 - mean_squared_error: 1.0187 - val_loss: 1.0359 - val_mean_squared_error: 1.0359\n",
      "Epoch 68/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0186 - mean_squared_error: 1.0186 - val_loss: 1.0359 - val_mean_squared_error: 1.0359\n",
      "Epoch 69/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0185 - mean_squared_error: 1.0185 - val_loss: 1.0359 - val_mean_squared_error: 1.0358\n",
      "Epoch 70/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0185 - mean_squared_error: 1.0185 - val_loss: 1.0358 - val_mean_squared_error: 1.0358\n",
      "Epoch 71/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0184 - mean_squared_error: 1.0184 - val_loss: 1.0358 - val_mean_squared_error: 1.0358\n",
      "Epoch 72/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0183 - mean_squared_error: 1.0183 - val_loss: 1.0357 - val_mean_squared_error: 1.0357\n",
      "Epoch 73/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0182 - mean_squared_error: 1.0182 - val_loss: 1.0357 - val_mean_squared_error: 1.0357\n",
      "Epoch 74/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0181 - mean_squared_error: 1.0181 - val_loss: 1.0356 - val_mean_squared_error: 1.0356\n",
      "Epoch 75/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0180 - mean_squared_error: 1.0180 - val_loss: 1.0356 - val_mean_squared_error: 1.0356\n",
      "Epoch 76/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0179 - mean_squared_error: 1.0179 - val_loss: 1.0356 - val_mean_squared_error: 1.0355\n",
      "Epoch 77/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0178 - mean_squared_error: 1.0178 - val_loss: 1.0355 - val_mean_squared_error: 1.0355\n",
      "Epoch 78/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0177 - mean_squared_error: 1.0177 - val_loss: 1.0355 - val_mean_squared_error: 1.0355\n",
      "Epoch 79/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0177 - mean_squared_error: 1.0176 - val_loss: 1.0354 - val_mean_squared_error: 1.0354\n",
      "Epoch 80/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0176 - mean_squared_error: 1.0176 - val_loss: 1.0354 - val_mean_squared_error: 1.0354\n",
      "Epoch 81/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0175 - mean_squared_error: 1.0175 - val_loss: 1.0353 - val_mean_squared_error: 1.0353\n",
      "Epoch 82/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0174 - mean_squared_error: 1.0173 - val_loss: 1.0353 - val_mean_squared_error: 1.0353\n",
      "Epoch 83/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0172 - mean_squared_error: 1.0172 - val_loss: 1.0352 - val_mean_squared_error: 1.0352\n",
      "Epoch 84/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0171 - mean_squared_error: 1.0171 - val_loss: 1.0352 - val_mean_squared_error: 1.0352\n",
      "Epoch 85/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0170 - mean_squared_error: 1.0170 - val_loss: 1.0351 - val_mean_squared_error: 1.0351\n",
      "Epoch 86/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0169 - mean_squared_error: 1.0169 - val_loss: 1.0350 - val_mean_squared_error: 1.0350\n",
      "Epoch 87/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0168 - mean_squared_error: 1.0168 - val_loss: 1.0350 - val_mean_squared_error: 1.0350\n",
      "Epoch 88/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0167 - mean_squared_error: 1.0167 - val_loss: 1.0349 - val_mean_squared_error: 1.0349\n",
      "Epoch 89/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0166 - mean_squared_error: 1.0166 - val_loss: 1.0349 - val_mean_squared_error: 1.0349\n",
      "Epoch 90/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0165 - mean_squared_error: 1.0164 - val_loss: 1.0348 - val_mean_squared_error: 1.0348\n",
      "Epoch 91/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0163 - mean_squared_error: 1.0163 - val_loss: 1.0348 - val_mean_squared_error: 1.0347\n",
      "Epoch 92/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0162 - mean_squared_error: 1.0162 - val_loss: 1.0347 - val_mean_squared_error: 1.0347\n",
      "Epoch 93/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 1.0161 - mean_squared_error: 1.0161 - val_loss: 1.0346 - val_mean_squared_error: 1.0346\n",
      "Epoch 94/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0159 - mean_squared_error: 1.0159 - val_loss: 1.0346 - val_mean_squared_error: 1.0346\n",
      "Epoch 95/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0158 - mean_squared_error: 1.0158 - val_loss: 1.0345 - val_mean_squared_error: 1.0345\n",
      "Epoch 96/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0157 - mean_squared_error: 1.0157 - val_loss: 1.0344 - val_mean_squared_error: 1.0344\n",
      "Epoch 97/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0155 - mean_squared_error: 1.0155 - val_loss: 1.0343 - val_mean_squared_error: 1.0343\n",
      "Epoch 98/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0154 - mean_squared_error: 1.0154 - val_loss: 1.0343 - val_mean_squared_error: 1.0343\n",
      "Epoch 99/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0152 - mean_squared_error: 1.0152 - val_loss: 1.0342 - val_mean_squared_error: 1.0342\n",
      "Epoch 100/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0151 - mean_squared_error: 1.0151 - val_loss: 1.0341 - val_mean_squared_error: 1.0341\n",
      "Epoch 101/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0149 - mean_squared_error: 1.0149 - val_loss: 1.0340 - val_mean_squared_error: 1.0340\n",
      "Epoch 102/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0147 - mean_squared_error: 1.0147 - val_loss: 1.0340 - val_mean_squared_error: 1.0340\n",
      "Epoch 103/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0146 - mean_squared_error: 1.0146 - val_loss: 1.0339 - val_mean_squared_error: 1.0339\n",
      "Epoch 104/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0144 - mean_squared_error: 1.0144 - val_loss: 1.0338 - val_mean_squared_error: 1.0338\n",
      "Epoch 105/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0142 - mean_squared_error: 1.0142 - val_loss: 1.0337 - val_mean_squared_error: 1.0337\n",
      "Epoch 106/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0140 - mean_squared_error: 1.0140 - val_loss: 1.0336 - val_mean_squared_error: 1.0336\n",
      "Epoch 107/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0138 - mean_squared_error: 1.0138 - val_loss: 1.0335 - val_mean_squared_error: 1.0335\n",
      "Epoch 108/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0136 - mean_squared_error: 1.0136 - val_loss: 1.0334 - val_mean_squared_error: 1.0334\n",
      "Epoch 109/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0134 - mean_squared_error: 1.0134 - val_loss: 1.0333 - val_mean_squared_error: 1.0333\n",
      "Epoch 110/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0132 - mean_squared_error: 1.0132 - val_loss: 1.0332 - val_mean_squared_error: 1.0332\n",
      "Epoch 111/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0130 - mean_squared_error: 1.0130 - val_loss: 1.0331 - val_mean_squared_error: 1.0331\n",
      "Epoch 112/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0128 - mean_squared_error: 1.0128 - val_loss: 1.0330 - val_mean_squared_error: 1.0330\n",
      "Epoch 113/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0126 - mean_squared_error: 1.0126 - val_loss: 1.0329 - val_mean_squared_error: 1.0329\n",
      "Epoch 114/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0123 - mean_squared_error: 1.0123 - val_loss: 1.0328 - val_mean_squared_error: 1.0328\n",
      "Epoch 115/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0121 - mean_squared_error: 1.0121 - val_loss: 1.0327 - val_mean_squared_error: 1.0327\n",
      "Epoch 116/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0119 - mean_squared_error: 1.0119 - val_loss: 1.0326 - val_mean_squared_error: 1.0326\n",
      "Epoch 117/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0116 - mean_squared_error: 1.0116 - val_loss: 1.0324 - val_mean_squared_error: 1.0324\n",
      "Epoch 118/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0113 - mean_squared_error: 1.0113 - val_loss: 1.0323 - val_mean_squared_error: 1.0323\n",
      "Epoch 119/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0111 - mean_squared_error: 1.0111 - val_loss: 1.0322 - val_mean_squared_error: 1.0322\n",
      "Epoch 120/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0108 - mean_squared_error: 1.0108 - val_loss: 1.0321 - val_mean_squared_error: 1.0320\n",
      "Epoch 121/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0105 - mean_squared_error: 1.0105 - val_loss: 1.0319 - val_mean_squared_error: 1.0319\n",
      "Epoch 122/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 1.0102 - mean_squared_error: 1.0102 - val_loss: 1.0318 - val_mean_squared_error: 1.0318\n",
      "Epoch 123/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0099 - mean_squared_error: 1.0099 - val_loss: 1.0316 - val_mean_squared_error: 1.0316\n",
      "Epoch 124/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0096 - mean_squared_error: 1.0096 - val_loss: 1.0315 - val_mean_squared_error: 1.0315\n",
      "Epoch 125/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0093 - mean_squared_error: 1.0093 - val_loss: 1.0313 - val_mean_squared_error: 1.0313\n",
      "Epoch 126/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0089 - mean_squared_error: 1.0089 - val_loss: 1.0312 - val_mean_squared_error: 1.0311\n",
      "Epoch 127/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0086 - mean_squared_error: 1.0086 - val_loss: 1.0310 - val_mean_squared_error: 1.0310\n",
      "Epoch 128/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0082 - mean_squared_error: 1.0082 - val_loss: 1.0308 - val_mean_squared_error: 1.0308\n",
      "Epoch 129/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0078 - mean_squared_error: 1.0078 - val_loss: 1.0306 - val_mean_squared_error: 1.0306\n",
      "Epoch 130/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 1.0074 - mean_squared_error: 1.0074 - val_loss: 1.0304 - val_mean_squared_error: 1.0304\n",
      "Epoch 131/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0070 - mean_squared_error: 1.0070 - val_loss: 1.0303 - val_mean_squared_error: 1.0303\n",
      "Epoch 132/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0066 - mean_squared_error: 1.0066 - val_loss: 1.0301 - val_mean_squared_error: 1.0301\n",
      "Epoch 133/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0062 - mean_squared_error: 1.0062 - val_loss: 1.0299 - val_mean_squared_error: 1.0299\n",
      "Epoch 134/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0058 - mean_squared_error: 1.0057 - val_loss: 1.0296 - val_mean_squared_error: 1.0296\n",
      "Epoch 135/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0053 - mean_squared_error: 1.0053 - val_loss: 1.0294 - val_mean_squared_error: 1.0294\n",
      "Epoch 136/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0048 - mean_squared_error: 1.0048 - val_loss: 1.0292 - val_mean_squared_error: 1.0292\n",
      "Epoch 137/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0043 - mean_squared_error: 1.0043 - val_loss: 1.0290 - val_mean_squared_error: 1.0290\n",
      "Epoch 138/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0038 - mean_squared_error: 1.0038 - val_loss: 1.0287 - val_mean_squared_error: 1.0287\n",
      "Epoch 139/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0033 - mean_squared_error: 1.0033 - val_loss: 1.0285 - val_mean_squared_error: 1.0285\n",
      "Epoch 140/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0028 - mean_squared_error: 1.0028 - val_loss: 1.0282 - val_mean_squared_error: 1.0282\n",
      "Epoch 141/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0022 - mean_squared_error: 1.0022 - val_loss: 1.0280 - val_mean_squared_error: 1.0280\n",
      "Epoch 142/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0016 - mean_squared_error: 1.0016 - val_loss: 1.0277 - val_mean_squared_error: 1.0277\n",
      "Epoch 143/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0010 - mean_squared_error: 1.0010 - val_loss: 1.0274 - val_mean_squared_error: 1.0274\n",
      "Epoch 144/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 1.0004 - mean_squared_error: 1.0004 - val_loss: 1.0271 - val_mean_squared_error: 1.0271\n",
      "Epoch 145/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9998 - mean_squared_error: 0.9998 - val_loss: 1.0269 - val_mean_squared_error: 1.0269\n",
      "Epoch 146/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9991 - mean_squared_error: 0.9991 - val_loss: 1.0266 - val_mean_squared_error: 1.0265\n",
      "Epoch 147/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9985 - mean_squared_error: 0.9985 - val_loss: 1.0262 - val_mean_squared_error: 1.0262\n",
      "Epoch 148/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9978 - mean_squared_error: 0.9978 - val_loss: 1.0259 - val_mean_squared_error: 1.0259\n",
      "Epoch 149/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9971 - mean_squared_error: 0.9971 - val_loss: 1.0256 - val_mean_squared_error: 1.0256\n",
      "Epoch 150/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9964 - mean_squared_error: 0.9964 - val_loss: 1.0253 - val_mean_squared_error: 1.0253\n",
      "Epoch 151/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9956 - mean_squared_error: 0.9956 - val_loss: 1.0249 - val_mean_squared_error: 1.0249\n",
      "Epoch 152/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9949 - mean_squared_error: 0.9949 - val_loss: 1.0246 - val_mean_squared_error: 1.0246\n",
      "Epoch 153/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9941 - mean_squared_error: 0.9941 - val_loss: 1.0242 - val_mean_squared_error: 1.0242\n",
      "Epoch 154/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9933 - mean_squared_error: 0.9933 - val_loss: 1.0239 - val_mean_squared_error: 1.0239\n",
      "Epoch 155/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9925 - mean_squared_error: 0.9925 - val_loss: 1.0235 - val_mean_squared_error: 1.0235\n",
      "Epoch 156/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9917 - mean_squared_error: 0.9917 - val_loss: 1.0231 - val_mean_squared_error: 1.0231\n",
      "Epoch 157/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9909 - mean_squared_error: 0.9909 - val_loss: 1.0227 - val_mean_squared_error: 1.0227\n",
      "Epoch 158/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9900 - mean_squared_error: 0.9900 - val_loss: 1.0224 - val_mean_squared_error: 1.0224\n",
      "Epoch 159/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9892 - mean_squared_error: 0.9892 - val_loss: 1.0220 - val_mean_squared_error: 1.0220\n",
      "Epoch 160/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9883 - mean_squared_error: 0.9883 - val_loss: 1.0216 - val_mean_squared_error: 1.0216\n",
      "Epoch 161/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9874 - mean_squared_error: 0.9874 - val_loss: 1.0212 - val_mean_squared_error: 1.0212\n",
      "Epoch 162/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9866 - mean_squared_error: 0.9866 - val_loss: 1.0208 - val_mean_squared_error: 1.0208\n",
      "Epoch 163/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9857 - mean_squared_error: 0.9857 - val_loss: 1.0204 - val_mean_squared_error: 1.0204\n",
      "Epoch 164/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9848 - mean_squared_error: 0.9848 - val_loss: 1.0200 - val_mean_squared_error: 1.0200\n",
      "Epoch 165/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9839 - mean_squared_error: 0.9839 - val_loss: 1.0197 - val_mean_squared_error: 1.0196\n",
      "Epoch 166/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9831 - mean_squared_error: 0.9831 - val_loss: 1.0193 - val_mean_squared_error: 1.0193\n",
      "Epoch 167/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9822 - mean_squared_error: 0.9822 - val_loss: 1.0189 - val_mean_squared_error: 1.0189\n",
      "Epoch 168/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9813 - mean_squared_error: 0.9813 - val_loss: 1.0185 - val_mean_squared_error: 1.0185\n",
      "Epoch 169/10000\n",
      "7/7 [==============================] - 0s 10ms/step - loss: 0.9805 - mean_squared_error: 0.9805 - val_loss: 1.0182 - val_mean_squared_error: 1.0181\n",
      "Epoch 170/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.9797 - mean_squared_error: 0.9797 - val_loss: 1.0178 - val_mean_squared_error: 1.0178\n",
      "Epoch 171/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9788 - mean_squared_error: 0.9788 - val_loss: 1.0174 - val_mean_squared_error: 1.0174\n",
      "Epoch 172/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9780 - mean_squared_error: 0.9780 - val_loss: 1.0171 - val_mean_squared_error: 1.0171\n",
      "Epoch 173/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9772 - mean_squared_error: 0.9772 - val_loss: 1.0168 - val_mean_squared_error: 1.0167\n",
      "Epoch 174/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9764 - mean_squared_error: 0.9764 - val_loss: 1.0164 - val_mean_squared_error: 1.0164\n",
      "Epoch 175/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9757 - mean_squared_error: 0.9756 - val_loss: 1.0161 - val_mean_squared_error: 1.0161\n",
      "Epoch 176/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9749 - mean_squared_error: 0.9749 - val_loss: 1.0158 - val_mean_squared_error: 1.0158\n",
      "Epoch 177/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9742 - mean_squared_error: 0.9742 - val_loss: 1.0155 - val_mean_squared_error: 1.0155\n",
      "Epoch 178/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9734 - mean_squared_error: 0.9734 - val_loss: 1.0152 - val_mean_squared_error: 1.0152\n",
      "Epoch 179/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9728 - mean_squared_error: 0.9727 - val_loss: 1.0149 - val_mean_squared_error: 1.0149\n",
      "Epoch 180/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9721 - mean_squared_error: 0.9721 - val_loss: 1.0147 - val_mean_squared_error: 1.0147\n",
      "Epoch 181/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9714 - mean_squared_error: 0.9714 - val_loss: 1.0144 - val_mean_squared_error: 1.0144\n",
      "Epoch 182/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9708 - mean_squared_error: 0.9708 - val_loss: 1.0142 - val_mean_squared_error: 1.0142\n",
      "Epoch 183/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9702 - mean_squared_error: 0.9702 - val_loss: 1.0140 - val_mean_squared_error: 1.0139\n",
      "Epoch 184/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9696 - mean_squared_error: 0.9696 - val_loss: 1.0137 - val_mean_squared_error: 1.0137\n",
      "Epoch 185/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9690 - mean_squared_error: 0.9690 - val_loss: 1.0135 - val_mean_squared_error: 1.0135\n",
      "Epoch 186/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9684 - mean_squared_error: 0.9684 - val_loss: 1.0133 - val_mean_squared_error: 1.0133\n",
      "Epoch 187/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9679 - mean_squared_error: 0.9679 - val_loss: 1.0131 - val_mean_squared_error: 1.0131\n",
      "Epoch 188/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9674 - mean_squared_error: 0.9674 - val_loss: 1.0130 - val_mean_squared_error: 1.0130\n",
      "Epoch 189/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9669 - mean_squared_error: 0.9669 - val_loss: 1.0128 - val_mean_squared_error: 1.0128\n",
      "Epoch 190/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9664 - mean_squared_error: 0.9664 - val_loss: 1.0127 - val_mean_squared_error: 1.0126\n",
      "Epoch 191/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9659 - mean_squared_error: 0.9659 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 192/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9655 - mean_squared_error: 0.9655 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 193/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9650 - mean_squared_error: 0.9650 - val_loss: 1.0123 - val_mean_squared_error: 1.0122\n",
      "Epoch 194/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9646 - mean_squared_error: 0.9646 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 195/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9642 - mean_squared_error: 0.9642 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 196/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9638 - mean_squared_error: 0.9638 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 197/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9635 - mean_squared_error: 0.9635 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 198/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9631 - mean_squared_error: 0.9631 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 199/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9628 - mean_squared_error: 0.9628 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n",
      "Epoch 200/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9625 - mean_squared_error: 0.9625 - val_loss: 1.0116 - val_mean_squared_error: 1.0116\n",
      "Epoch 201/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9622 - mean_squared_error: 0.9622 - val_loss: 1.0116 - val_mean_squared_error: 1.0116\n",
      "Epoch 202/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9619 - mean_squared_error: 0.9619 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 203/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9616 - mean_squared_error: 0.9616 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 204/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9613 - mean_squared_error: 0.9613 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 205/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9611 - mean_squared_error: 0.9611 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 206/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9608 - mean_squared_error: 0.9608 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 207/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9606 - mean_squared_error: 0.9606 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 208/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9604 - mean_squared_error: 0.9604 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 209/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9602 - mean_squared_error: 0.9601 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 210/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9599 - mean_squared_error: 0.9599 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 211/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9598 - mean_squared_error: 0.9598 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 212/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9596 - mean_squared_error: 0.9596 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 213/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9594 - mean_squared_error: 0.9594 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 214/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9592 - mean_squared_error: 0.9592 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 215/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9591 - mean_squared_error: 0.9591 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 216/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9589 - mean_squared_error: 0.9589 - val_loss: 1.0113 - val_mean_squared_error: 1.0113\n",
      "Epoch 217/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9588 - mean_squared_error: 0.9588 - val_loss: 1.0114 - val_mean_squared_error: 1.0113\n",
      "Epoch 218/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9586 - mean_squared_error: 0.9586 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 219/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9585 - mean_squared_error: 0.9585 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 220/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.9584 - mean_squared_error: 0.9584 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 221/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9583 - mean_squared_error: 0.9583 - val_loss: 1.0114 - val_mean_squared_error: 1.0114\n",
      "Epoch 222/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9582 - mean_squared_error: 0.9582 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 223/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9581 - mean_squared_error: 0.9580 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 224/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9579 - mean_squared_error: 0.9579 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 225/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9579 - mean_squared_error: 0.9578 - val_loss: 1.0115 - val_mean_squared_error: 1.0115\n",
      "Epoch 226/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9578 - mean_squared_error: 0.9578 - val_loss: 1.0116 - val_mean_squared_error: 1.0116\n",
      "Epoch 227/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9577 - mean_squared_error: 0.9577 - val_loss: 1.0116 - val_mean_squared_error: 1.0116\n",
      "Epoch 228/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9576 - mean_squared_error: 0.9576 - val_loss: 1.0116 - val_mean_squared_error: 1.0116\n",
      "Epoch 229/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.9575 - mean_squared_error: 0.9575 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n",
      "Epoch 230/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9574 - mean_squared_error: 0.9574 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n",
      "Epoch 231/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9574 - mean_squared_error: 0.9574 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n",
      "Epoch 232/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9573 - mean_squared_error: 0.9573 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n",
      "Epoch 233/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9572 - mean_squared_error: 0.9572 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 234/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9572 - mean_squared_error: 0.9572 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 235/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9571 - mean_squared_error: 0.9571 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 236/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9570 - mean_squared_error: 0.9570 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 237/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9570 - mean_squared_error: 0.9570 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 238/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9569 - mean_squared_error: 0.9569 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 239/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.9569 - mean_squared_error: 0.9569 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 240/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9568 - mean_squared_error: 0.9568 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 241/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9568 - mean_squared_error: 0.9568 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 242/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9567 - mean_squared_error: 0.9567 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 243/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9567 - mean_squared_error: 0.9567 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 244/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.9566 - mean_squared_error: 0.9566 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 245/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9566 - mean_squared_error: 0.9566 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 246/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9566 - mean_squared_error: 0.9565 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 247/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9565 - mean_squared_error: 0.9565 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 248/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.9565 - mean_squared_error: 0.9565 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 249/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9564 - mean_squared_error: 0.9564 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 250/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9564 - mean_squared_error: 0.9564 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 251/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9564 - mean_squared_error: 0.9564 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 252/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9563 - mean_squared_error: 0.9563 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 253/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9563 - mean_squared_error: 0.9563 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 254/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9563 - mean_squared_error: 0.9563 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 255/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9562 - mean_squared_error: 0.9562 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 256/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9562 - mean_squared_error: 0.9562 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 257/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9562 - mean_squared_error: 0.9562 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 258/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9561 - mean_squared_error: 0.9561 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 259/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9561 - mean_squared_error: 0.9561 - val_loss: 1.0125 - val_mean_squared_error: 1.0124\n",
      "Epoch 260/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9561 - mean_squared_error: 0.9561 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 261/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9560 - mean_squared_error: 0.9560 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 262/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9560 - mean_squared_error: 0.9560 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 263/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9560 - mean_squared_error: 0.9560 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 264/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9560 - mean_squared_error: 0.9560 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 265/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9559 - mean_squared_error: 0.9559 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 266/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9559 - mean_squared_error: 0.9559 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 267/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9559 - mean_squared_error: 0.9559 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 268/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9558 - mean_squared_error: 0.9558 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 269/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9558 - mean_squared_error: 0.9558 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 270/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9558 - mean_squared_error: 0.9558 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 271/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9558 - mean_squared_error: 0.9558 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 272/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9557 - mean_squared_error: 0.9557 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 273/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9557 - mean_squared_error: 0.9557 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 274/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9557 - mean_squared_error: 0.9557 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 275/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9557 - mean_squared_error: 0.9557 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 276/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9556 - mean_squared_error: 0.9556 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 277/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9556 - mean_squared_error: 0.9556 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 278/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9556 - mean_squared_error: 0.9556 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 279/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9556 - mean_squared_error: 0.9556 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 280/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9555 - mean_squared_error: 0.9555 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 281/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9555 - mean_squared_error: 0.9555 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 282/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9555 - mean_squared_error: 0.9555 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 283/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9555 - mean_squared_error: 0.9555 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 284/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9554 - mean_squared_error: 0.9554 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 285/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9554 - mean_squared_error: 0.9554 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 286/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9554 - mean_squared_error: 0.9554 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 287/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9554 - mean_squared_error: 0.9554 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 288/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9553 - mean_squared_error: 0.9553 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 289/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.9553 - mean_squared_error: 0.9553 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 290/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9553 - mean_squared_error: 0.9553 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 291/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9553 - mean_squared_error: 0.9553 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 292/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9552 - mean_squared_error: 0.9552 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 293/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9552 - mean_squared_error: 0.9552 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 294/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9552 - mean_squared_error: 0.9552 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 295/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9552 - mean_squared_error: 0.9552 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 296/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9551 - mean_squared_error: 0.9551 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 297/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9551 - mean_squared_error: 0.9551 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 298/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9551 - mean_squared_error: 0.9551 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 299/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9551 - mean_squared_error: 0.9551 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 300/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9550 - mean_squared_error: 0.9550 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 301/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9550 - mean_squared_error: 0.9550 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 302/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9550 - mean_squared_error: 0.9550 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 303/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9550 - mean_squared_error: 0.9550 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 304/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9549 - mean_squared_error: 0.9549 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 305/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9549 - mean_squared_error: 0.9549 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 306/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9549 - mean_squared_error: 0.9549 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 307/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9549 - mean_squared_error: 0.9549 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 308/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9548 - mean_squared_error: 0.9548 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 309/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9548 - mean_squared_error: 0.9548 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 310/10000\n",
      "7/7 [==============================] - 0s 8ms/step - loss: 0.9548 - mean_squared_error: 0.9548 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 311/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9548 - mean_squared_error: 0.9548 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 312/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9547 - mean_squared_error: 0.9547 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 313/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9547 - mean_squared_error: 0.9547 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 314/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9547 - mean_squared_error: 0.9547 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 315/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9547 - mean_squared_error: 0.9547 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 316/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9546 - mean_squared_error: 0.9546 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 317/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9546 - mean_squared_error: 0.9546 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 318/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9546 - mean_squared_error: 0.9546 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 319/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9546 - mean_squared_error: 0.9546 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 320/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9545 - mean_squared_error: 0.9545 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 321/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9545 - mean_squared_error: 0.9545 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 322/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9545 - mean_squared_error: 0.9545 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 323/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9545 - mean_squared_error: 0.9545 - val_loss: 1.0127 - val_mean_squared_error: 1.0127\n",
      "Epoch 324/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9544 - mean_squared_error: 0.9544 - val_loss: 1.0127 - val_mean_squared_error: 1.0126\n",
      "Epoch 325/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9544 - mean_squared_error: 0.9544 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 326/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9544 - mean_squared_error: 0.9544 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 327/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9544 - mean_squared_error: 0.9544 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 328/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9543 - mean_squared_error: 0.9543 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 329/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9543 - mean_squared_error: 0.9543 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 330/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9543 - mean_squared_error: 0.9543 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 331/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9543 - mean_squared_error: 0.9543 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 332/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9542 - mean_squared_error: 0.9542 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 333/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9542 - mean_squared_error: 0.9542 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 334/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9542 - mean_squared_error: 0.9542 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 335/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9542 - mean_squared_error: 0.9542 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 336/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9541 - mean_squared_error: 0.9541 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 337/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9541 - mean_squared_error: 0.9541 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 338/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9541 - mean_squared_error: 0.9541 - val_loss: 1.0126 - val_mean_squared_error: 1.0126\n",
      "Epoch 339/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9541 - mean_squared_error: 0.9541 - val_loss: 1.0126 - val_mean_squared_error: 1.0125\n",
      "Epoch 340/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9540 - mean_squared_error: 0.9540 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 341/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9540 - mean_squared_error: 0.9540 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 342/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9540 - mean_squared_error: 0.9540 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 343/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9540 - mean_squared_error: 0.9540 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 344/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9539 - mean_squared_error: 0.9539 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 345/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9539 - mean_squared_error: 0.9539 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 346/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9539 - mean_squared_error: 0.9539 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 347/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9539 - mean_squared_error: 0.9539 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 348/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9538 - mean_squared_error: 0.9538 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 349/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9538 - mean_squared_error: 0.9538 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 350/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9538 - mean_squared_error: 0.9538 - val_loss: 1.0125 - val_mean_squared_error: 1.0125\n",
      "Epoch 351/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9538 - mean_squared_error: 0.9538 - val_loss: 1.0125 - val_mean_squared_error: 1.0124\n",
      "Epoch 352/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9537 - mean_squared_error: 0.9537 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 353/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9537 - mean_squared_error: 0.9537 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 354/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9537 - mean_squared_error: 0.9537 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 355/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9537 - mean_squared_error: 0.9537 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 356/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9536 - mean_squared_error: 0.9536 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 357/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9536 - mean_squared_error: 0.9536 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 358/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9536 - mean_squared_error: 0.9536 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 359/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9536 - mean_squared_error: 0.9536 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 360/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9535 - mean_squared_error: 0.9535 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 361/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9535 - mean_squared_error: 0.9535 - val_loss: 1.0124 - val_mean_squared_error: 1.0124\n",
      "Epoch 362/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9535 - mean_squared_error: 0.9535 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 363/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9535 - mean_squared_error: 0.9535 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 364/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9534 - mean_squared_error: 0.9534 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 365/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9534 - mean_squared_error: 0.9534 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 366/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9534 - mean_squared_error: 0.9534 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 367/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9534 - mean_squared_error: 0.9534 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 368/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9533 - mean_squared_error: 0.9533 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 369/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9533 - mean_squared_error: 0.9533 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 370/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9533 - mean_squared_error: 0.9533 - val_loss: 1.0123 - val_mean_squared_error: 1.0123\n",
      "Epoch 371/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9533 - mean_squared_error: 0.9533 - val_loss: 1.0123 - val_mean_squared_error: 1.0122\n",
      "Epoch 372/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9532 - mean_squared_error: 0.9532 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 373/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.9532 - mean_squared_error: 0.9532 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 374/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9532 - mean_squared_error: 0.9532 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 375/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9532 - mean_squared_error: 0.9532 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 376/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9531 - mean_squared_error: 0.9531 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 377/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9531 - mean_squared_error: 0.9531 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 378/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9531 - mean_squared_error: 0.9531 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 379/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9531 - mean_squared_error: 0.9531 - val_loss: 1.0122 - val_mean_squared_error: 1.0122\n",
      "Epoch 380/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9530 - mean_squared_error: 0.9530 - val_loss: 1.0122 - val_mean_squared_error: 1.0121\n",
      "Epoch 381/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9530 - mean_squared_error: 0.9530 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 382/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9530 - mean_squared_error: 0.9530 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 383/10000\n",
      "7/7 [==============================] - 0s 5ms/step - loss: 0.9530 - mean_squared_error: 0.9530 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 384/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9529 - mean_squared_error: 0.9529 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 385/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9529 - mean_squared_error: 0.9529 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 386/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9529 - mean_squared_error: 0.9529 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 387/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9529 - mean_squared_error: 0.9529 - val_loss: 1.0121 - val_mean_squared_error: 1.0121\n",
      "Epoch 388/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9528 - mean_squared_error: 0.9528 - val_loss: 1.0121 - val_mean_squared_error: 1.0120\n",
      "Epoch 389/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9528 - mean_squared_error: 0.9528 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 390/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9528 - mean_squared_error: 0.9528 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 391/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9528 - mean_squared_error: 0.9528 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 392/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9527 - mean_squared_error: 0.9527 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 393/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9527 - mean_squared_error: 0.9527 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 394/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9527 - mean_squared_error: 0.9527 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 395/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9527 - mean_squared_error: 0.9527 - val_loss: 1.0120 - val_mean_squared_error: 1.0120\n",
      "Epoch 396/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9526 - mean_squared_error: 0.9526 - val_loss: 1.0120 - val_mean_squared_error: 1.0119\n",
      "Epoch 397/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9526 - mean_squared_error: 0.9526 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 398/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9526 - mean_squared_error: 0.9526 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 399/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9526 - mean_squared_error: 0.9526 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 400/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9525 - mean_squared_error: 0.9525 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 401/10000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9525 - mean_squared_error: 0.9525 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 402/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9525 - mean_squared_error: 0.9525 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 403/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9525 - mean_squared_error: 0.9525 - val_loss: 1.0119 - val_mean_squared_error: 1.0119\n",
      "Epoch 404/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9524 - mean_squared_error: 0.9524 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 405/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9524 - mean_squared_error: 0.9524 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 406/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9524 - mean_squared_error: 0.9524 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 407/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9524 - mean_squared_error: 0.9524 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 408/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9523 - mean_squared_error: 0.9523 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 409/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9523 - mean_squared_error: 0.9523 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 410/10000\n",
      "7/7 [==============================] - 0s 7ms/step - loss: 0.9523 - mean_squared_error: 0.9523 - val_loss: 1.0118 - val_mean_squared_error: 1.0118\n",
      "Epoch 411/10000\n",
      "7/7 [==============================] - 0s 6ms/step - loss: 0.9523 - mean_squared_error: 0.9523 - val_loss: 1.0117 - val_mean_squared_error: 1.0117\n"
     ]
    }
   ],
   "source": [
    "uncompiled_model = get_uncomplied_model(input_dim, hidden_layer_1_dim, hidden_layer_2_dim)\n",
    "model = get_compiled_model(uncompiled_model, learning_rate)\n",
    "\n",
    "model_saver = ModelCheckpoint(os.path.join(\"saved_models\", \"save_at_{epoch}.h5\"), period=100)\n",
    "early_stopper = tf.keras.callbacks.EarlyStopping(monitor='val_mean_squared_error', \n",
    "                                                 min_delta=10e-6, \n",
    "                                                 restore_best_weights=True, \n",
    "                                                 patience=200)\n",
    "callbacks = [model_saver, early_stopper]\n",
    "\n",
    "history_noisy = model.fit(inputs_noisy_training, targets_noisy_training,\n",
    "                    epochs=epochs, \n",
    "                    batch_size=batch_size, \n",
    "                    verbose=1, \n",
    "                    validation_data=(inputs_noisy_testing, targets_noisy_testing), \n",
    "                    shuffle=False,\n",
    "                    callbacks=callbacks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "beautiful-living",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    " show_curves(history_noisy)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "approved-tournament",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 1ms/step - loss: 1.1062 - mean_squared_error: 1.1062\n",
      "Test MSE 1.1062\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAEECAYAAAAMOA6OAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAByDElEQVR4nO1dd5gURfp+J27Ouyyw5LwCKoqo/FQQUBDwzsSZwHjcnSIYzkNQMJ0BjBjOwOmhZxZzABVFPEwgIEpYMiywsGzOMzuzO/37o+abru7pnumZndmd3a33efaZ2Znp7q+rq95++62vqkySJEkQEBAQEIh5mNs6AAEBAQEBYxCELSAgINBOIAhbQEBAoJ1AELaAgIBAO4EgbAEBAYF2AkHYAgICAu0EgrDbAQYPHow777xT8dm6deswY8aMoNteffXV2LZtW4tjmDdvHoYNG4aqqirF5xs2bMDgwYPxwQcfhLXfw4cP47jjjmtxfHooLS3F/PnzMXbsWEycOBHnnHMOFi1ahIaGBgCsHM8555yoHR8AXC4XPvroIwDAsWPHMHXq1KgeryWYO3cuVq9e3dZhCOhAEHY7wS+//ILt27eHvN2rr76KoUOHRiSG7OxsfPXVV4rPVqxYgW7dukVk/5FGfX09pk+fjm7dumHVqlX48ssv8c4772D37t1YuHBhq8Wxfft2H2Hn5ubis88+a7VjhwKPx4NHHnkE48aNa+tQBHQgCLud4LbbbsNDDz2k+Z3H48GTTz6JiRMnYuLEiZg/f75PQY4bNw4bNmxAU1MT7rrrLkycOBETJkzATTfdhLq6Olx88cX44osvfPtavXo1/vjHP2oeZ8yYMQqyaW5uxtq1a3HSSSf5PissLMSVV16JSZMm4dxzz1X8ft26dbjgggswceJEzJw5E5WVlX7HuP322/HPf/4TALBp0yZcfPHFmDRpEqZPn479+/ejuroaJ5xwAsrKynzbLFq0CA8++KDfvj744ANkZ2djzpw5sNlsAIDMzEw899xzWLx4sd/vGxsbcfvtt2PixIkYN26c4jcrV67E1KlTcd555+H888/HunXrAn5OKCsrw0033YTNmzfjiiuuUDxRLF++HHPmzME//vEPjBkzBtdeey02btyIP/3pTxg9ejTeeOMNAIAkSfjXv/6F8847D5MmTcIDDzwAt9vtF/+xY8dw9dVXY/LkyZgwYQKefPLJoNvPmDEDjz32GCZOnIhNmzZhxowZ+Pjjj3XLP9BxBKIPQdjtBOeddx4kSVKQK2HlypX43//+h48++ggrV65ETU0Nli1bpvjN999/j8OHD+OLL77AqlWrMHDgQPz666+YOnWqglRXrVqFKVOmaMZw/PHH4/DhwygpKQHACHj48OGw2+2+3yxevBhnnHEGvvjiCzz44IO466674Ha70dDQgFtvvRWLFy/Gl19+iT59+mDJkiWK/S9duhQ1NTW488474XA4MHv2bNxxxx344osvcNlll+H2229HWloaTj/9dKxYsSJozL/88gvOOussv8/j4uJgtVr9Pn/rrbdQWVmJlStX4sMPP8QHH3yADRs2AADuu+8+vPjii1i5ciXuvfden22g9zkhOzsbt912G0488US8+eabiu8sFgvWrl2LWbNmYdWqVdi7dy+WLl2Kt956C4sXL8Zzzz0HAPjqq6/w5ZdfYvny5fjss89w6NAhvPvuu37xv/LKKzjllFOwYsUKfPrpp75rFWz7goICrFy5EiNHjvR9plf+gY4jEH0Iwm5HuPPOO/HYY4+hsbFR8fmaNWswdepUJCQkwGw24w9/+AN++OEHxW+ys7Oxd+9erFq1Cg6HAzfffDPOPPNMTJ48GWvXrkVtbS2am5vx7bff4rzzztM8vslkwsSJE31k+fnnn2Py5MmK3zzzzDOYOXMmAOCkk05CY2MjSktL8dtvvyE7OxuDBw8GAPzjH//AXXfdpTiHFStW4IknnoDFYsHmzZuRlpaGUaNGAQCmTp2KwsJCFBcXY+rUqfj8888BMLKRJAknnniiX7x1dXXIzMz0/f/+++9j0qRJmDRpEk477TQcPXpU8ftrrrkGzz//PMxmM9LS0jBw4EAcPnwYAJCTk4N33nkHhw8fxsknn4z58+cH/NwoBgwYgD59+sBut6N3794444wzYLFYMGjQIJSVlcHj8eCbb77B+eefj+TkZFitVkybNg2rVq3y21eXLl3w/fffY8OGDbDb7Xj88cfRpUuXoNuPGTMGZrOSCgKVv95xBKIPf5khELMYOnQoTjnlFCxbtgwjRozwfV5RUYH09HTf/2lpaaioqFBsO2zYMNx333149dVXcccdd2D8+PG4++67kZubi+OPPx5fffUV8vLy0KNHD/Ts2VM3hqlTp+Kee+7BlVdeiZ9++gn33HMPvv76a9/33333HV588UVUVVXBZDJBkiR4PB6Ul5cjLS3N9ztelXs8Htx1113o27cvkpKSfOdUVFSESZMmKbYpLy/HuHHjsHDhQhw6dAirVq3y/Wbu3Ln4/fffATDvPiMjA8eOHfNtf/HFF+Piiy8GABx33HFobm5WnNvevXuxePFiFBYWwmQyobi4GBdddBEA4MUXX8Szzz6LSy65BLm5uVi4cCFGjhyp+7lR0PkCTHHT/0SgVHY//vgj3n//fQDMisrKyvLb19VXXw2Px4P77rsPJSUlmDFjBm666aag2/PXhRCo/PWOIxB9CMJuZ7j11ltx0UUXoUePHr7PMjMzFX5wZWUlsrOz/bYdP348xo8fj5qaGixYsAD/+c9/cMstt2DKlClYsWIF+vbt66eY1Rg6dChqa2vxwQcf4JRTTlEQr8vlwi233IIlS5Zg3LhxcLvdGD58OAAgKytLEaPD4UB1dbXv/zfffBPz5s3Dq6++imuuuQbZ2dno168fPvzwQ804zj77bKxcuRKrVq3Cww8/DAB45JFHFL8ZPXo0XnvtNcyaNQsmkyngeQHM3hg2bBheeOEFmM1mXH755b7vunfvjoceeggejweff/45/v73v+O7777T/TySyMnJwZgxY3DVVVcF/J3ZbMb111+P66+/HgcPHsTMmTNx8sknG96eR7Dy1zrO6aefHtJ5CYQOYYm0M3Tp0gVXXnklnnnmGd9nZ599NlasWAGn04nm5mZ8+OGHGDNmjGK79957D//6178AAKmpqejduzdoosZJkyZh06ZN+OKLL3TtEB5TpkzBs88+60fuTqcTjY2NOOGEE+DxePDSSy/BbrejoaEBJ554IsrLy30K+LnnnsOzzz4LgBFN79698fDDD+P555/Hvn37fB2LmzZtAgAcOnQId9xxhy/mqVOn4q233oLT6cSwYcM04zz//PNhMpnw4IMPwul0AgCqq6uxYMECpKenIyUlRfH7mpoaDBs2DGazGd9++y0KCwvR0NCA8vJyXHPNNairq4PZbMbQoUMhSZLu52pYrVbU1dVpfmcE48ePx8cff4z6+noAwNtvv+3LOuGxcOFCnxWWl5eHrKwsSJJkeHsegcpf7zgC0YdQ2O0Q1113HZYvX+77f9KkSdi5c6cvu+O0007zU1MTJkzAvHnzcO6558JisaBv374+ZZqeno5Ro0ahurraUIrelClT8MYbb2D06NGKz1NTUzFz5kxccMEFSE9Px+zZs3HOOedg9uzZ+PDDD7FkyRIsXLgQTqcTvXr1wuLFi33ZLADQp08fzJo1C3fccQfefvttPP3003jwwQdRV1cHm82GW265xaeUzzjjDNTX1ytUsBp2ux2vvvoqnnzySUyePBk2mw2SJOGMM87Ahx9+6GcF3HDDDXjooYfw9NNPY9KkSZg1axaee+45DB06FGeccQYuvvhiWCwWxMXF4eGHH0ZWVpbm52qcfPLJeOyxxzBmzBi8/vrrQctXjQkTJmD37t24+OKL4fF40Lt3b82MoUsvvRT33nuvL8tm/PjxPtVrZHse8fHxuuUf6DgC0YVJzIctADA7YMCAAbjyyivbOhTDmDJlCpYsWYKBAwe2dSgCAq0CYYkIoLCwEGvWrNHNv45FrFixAtnZ2YKsBToVhCXSyfH000/j/fffx8KFC5GcnNzW4RjCddddh2PHjvnylAUEOguEJSIgICDQTiAsEQEBAYF2AkHYAgICAu0EUfOwN27cGK1dCwgICHRonHzyyZqfR7XTUe+gwVBQUID8/PwIRxMZxGpsIq7QEKtxAbEbm4grNIQbVyCxKywRAQEBgXYCQdgCAgIC7QSCsAUEBATaCQRhCwgICLQTtIiwnU4nJkyYEPYCrAICAgICxtEiwn7++ec1Jz8XEBAQEIg8wibsvXv3Yu/evRg7dmwEwxEQEBAQ0EPYhP3II49g3rx5kYwldvHYY4B3qSiBCKKhAVAt0yUgIKCPsAbOfPTRRxg5cqRimSotFBQUhBWU0+kMe9tooPuaNUhatw67CwpiLjZCe4xrwPjxKL/2WlROn97KUcVueQGxG5uIKzREI66wCHvNmjU4fPgwVq1aheLiYtjtdnTt2tVvBZJwRx/F3Mglux1wu5Gfnx97sXnR7uKSJODoUXR1u9G1DeKO1fICYjc2EVdoiMZIx7AIe8mSJb73zzzzDPLy8vzIukOhsRHwrgkoECGQFeJ2t20cAgLtCCIP2wgaG9mfx9PWkXQcEFELwhYQMIwWT/40e/bsSMQR22hsZK9CZUcOgrAFBEKGUNhGQITtcLRtHB0JgrAFBEKGIGwjEIQdeQjCFhAIGYKwjUBYIpGHIGwBgZAhCNsIhMKOPIiom5raNg4BgXYEQdhGIAg78nC52KtQ2AIChiEI2wgEYUcewhIREAgZgrCNQBB25CEIW0AgZAjCNgLR6Rh5CMIWEAgZgrCDweORO8aEwo4cBGELCIQMQdjBQOoaEIQdSQjCFhAIGYKwg8EIYf/6K1BZ2TrxdBSItD4BgZAhCDsYjBD22LHAM8+0SjgdBkJhCwiEDEHYwcATtl6nY20tUFXVKuF0GAjCFhAIGYKwg4EnaS2F7fGwyfhpIIiAMYiBMwICIUMQdjAEs0RoIn7+dwLBIRS2gEDIEIQdDIKwowNB2AICIUMQdjAYJWxhiYQGkSUiIBAyBGEHQ7BOR6Gww4NQ2AICIUMQdjAISyQ6EIQtIBAyBGEHAxFxfLwg7EhCELaAFq67Dnj99baOImYhCDsYiIjT0jqGh11RwdIQ2xo8YcdCPAKxgQ8/BFavbusoYhaCsIOBCDs9vf0r7OJioGtX4Ntv2zoSpbKmMhQQaGwEGhraOoqYhSDsYOAJu713OpaVMaI8cqStI1E+kXR2W8TjaesIYgeNjUB9vfZ3+/cDffsChw+3bkwxBEHYwdCRFDal0MWCouVJujOn9h04ACQkAJs3t3UkbY+mJnbz0lPYO3ey8tqzp1XDiiUIwg4Go4TdHjxsIslYI+zOrLALCljd2b+/rSNpe9ATLCnsl18GvvlG/l50VMMazkYOhwPz5s1DeXk5GhoaMGvWLIwfPz7SscUGgnU6kjoUCjs0CMJmOHaMvYrVjOQ2RAr7n/8ETj0VIG4Rg63CI+zVq1dj2LBhmDlzJoqKinDdddd1fMJOTW3/lohQ2LGHkhL2KghbbkOksOvqtOtJJ64vYRH2lClTfO+Li4uRm5sbsYBiDo2NgN0OJCay9+oOovZkiQiFHXsQCluGWmHX1mp3Tnfi+hIWYROmTZuGsrIyLF26VPP7goKCsPbrdDrD3jbS6HL0KNLtdpTX1qILgMaaGkVscbt3ox8Aj9OJnW0Ys5EyS9q7F70AFB85gspWilUvrm5lZUj3vt+7YwdcrUxYsVLHuu/ejTQAxwoLUeGNJ1ZiUyPacdn37kV/AM21tdj122/Id7lQV1GBQ95jphUWojuAogMHUMPF0ZnKq0WEvXz5cmzbtg233XYbPv74Y5jNyj7M/Pz8sPZbUFAQ9rYRR1ISkJCALr17AwASAAzmY/OqAbPbjfwhQwCTqQ2CNFhm+/YBALrm5KBrK5WvblyJib63/Xv1Alr5esdMHfPeqHLT0pDrjSdmYlMh6nF5FbbF6UR+z54AgGSbTT5mTg4AIK9LF+RxcXS08tq4caPud2FliWzZsgVHvLm8Q4cOhcfjQWVHXdPQ6QTi4ljqFQCz2qvm7YVQbJEjR4CePVmqUmuhtTxsSQKeegqoqQkeC8CsmsWLO+cIN+Fhy6C21dwsr5EaS5aIJAFPPw1UV7fN8REmYf/666949dVXAQBlZWWor69HRkZGRAOLGTQ2KgjbpG5Y4RL23r1sAEBrEnZredi7dgG33AJ8/rn+b9QNcdEi4K23ohtXLEJ42DJ4MUTlEkuEvW0bcPPNwGeftc3xESZhX3bZZSgrK8MVV1yBv/3tb7jnnnv87JAOAyLs+HgAQRR2KJkilHHSmpWvtRQ2HSfQDUzd6ehw6I9w66jweIDSUvZeELayDOjJg69Dbb2sHF2rNswIC8vDttvtePzxxyMdS2xCrbAjRdhUOVuz8kVLYR84wB4X+/ZV7j/QufHfuVyBhyR3VFRWymUlCDv2FXZZGXttw4ywFnU6dgp0JMKOlsKeNYvt+6uvlPsPRtgmEyP6ujr2WWcjbCIlQBA2oGw/Wgq7rQfOCMJuB1ARtjlSHjbtpzUvfrQUdmWlMj+djhOoYbndLFOkvl7unCTi7iwgUgIEYQOxr7DLy9lrGxJ2BzWeI4hgnY48KXVWD7uxUVkORhV2UhJ7T4TdWRW23S4IG1CWQSwStlDY7QDR6nTsSB62ywVYLPL/RgmbcrFra9lrZyNsUtg9ewrCBmJfYccAYQuFHQyheNjqCzlrFvDMM9r77UgedmOj9oIERiwRoPMS9rFjgNkM5OUJwgaMe9jRbDM//aQ/1a0g7HYAlcI2qS9WIIX98cfAY49pL4HVFpZItBR2uJaI9ybYaS2RkhI2ei8xURA2EBsK+5ZbgAULtL8ThN0OQIRtZe6RSU12gQi7rg44eBDYtMl/v1oKu7k5ujmera2wAzUsl8tfYTc0sM7L++8H3nknsjHGIo4dA3JzmRgQhK0sg6oq9urx+NenaBK2w6Hf+W2003HVKqCoKLJxeSEIOxgaG1mD8hK2H9kFImxSjB9+6L9fLcJeuBAYO7ZF4QZEaylsem/UwyaFLUmswbz4onaZdTSUlABdugjCJuiJFfWAmWgSttutfy1IYQc7/gUXADoT4rUUHZuwf/tNvlOHC1LY3k41k9qX1fOwXS6ZuD74wH+/ZInw2xQWArt3tyzeQIiWwubPld+/UQ+bn3Okvp5ds84whaYgbCUaG4GUFP8J1IjIWyMPm0bdquF0yso7kMKWJPakqGWDRgAdm7DPPJNN1tIShGuJkLru25ctA1VYqNxOS2E3Ncn2QDQQTYUdqiWi1ekIsIl1Gho6B2ELS0QJamuU7klobYWtRdhkh/DxaIHqvs0W2bi86LiE3dzMiIAfnBAqJIldHI6wDVsiRNiDB7NXtdLXImy3Wx6mHQ1EQ2E3N7O/cDodtQibVnTv6IRdX8/+hMKW4XSysuCm3gUgE2RrzCWiZ4mQHcLHoQX6zm6PbFxedAzC/v57/0Im0gs0xWcwUOHzlkiohJ2ezl7Vj3FaWSL0m2ip7GgobPXjKr//UNP6ALmzpq0Ju6KC9SccPBid/ZOQEApbRrgKW5KAZcsiM1K2pQqbYhMKWwdlZcBZZwFvv638nBpASwibyCgujuXLmkyBCZu/kGrCVlcyPYUNRI+wo6GwqYxaorD5axQrhL1lC/Ddd8D69dHZPxE2KWyXy3/5uc4GImyqF0R6QQjbvncvcN11wKeftjyGYAqbrlWg7QFB2Logg189qTgRSUvIjydsgNki4SpsI4TdnhV2KITd3MyumTe3XZOw23plbKpP1dWygotknjjlGZPCBtrHQs7RhFphZ2Wx1yCEba2oYG8i8ZSip7CJsLt3F4TdIlDDVlf2SFgiVAE4wvZT2HpziQSzRLQmf2qPCpviJxIGgqf18ZXaZlOWTawobKo3NTXAjh1MwX3ySeT2zytsql+d3RZRe9iZmew1CGFbqH8oEjf5piZWl9X1jwi7a1dB2C0CFZCasCNtiQDMxw7VEklLU8ZJ6Ggettb+9RoQlRMRNo9Y6XTkFTb5l5EkVFLYZIlEev/tEWqF3RaETftWX4uyMtaWk5IEYbcI0VTY3gV2fXd8qzVwHnZLLZFACluSgEsuaZlPR/uPpFfKn7NawRtV2ADrIwBiT2FXV8sZPpGMqaQESE1lZC0Im0HtYRslbFr/saXXh39KVNsiZWXMorHbjS3MIQhbB8EIuyVqlSoCrVdptfqTHZFTQkJ4lohRhb1/P/D++8A334R0Cgq0tsIOhbCpccaiwibCNjKHxNatwJw5wW+KlIMNCA+bQAteB/OwVW0pYgqbr3Pqm2dDAxvUY7eLtL4WQY+wqcDr68MnKA3C1lXYeoQdiiUSSGH//DN7bcnIzWhmiQDGLRGesCm/PS2NWU5tPYUmgfewtVbw1sPKlWyGRtpGDzTKERAKm0DTQISqsOnm2tI6w2+vVthNTayuBiNsobCDIJjCBsJX2USOpJItFv20vsRE5YWknFA9wg5VYa9bx17V2TChIBoKW6vT1KjCttvlip2QoMy/jQRhr1sH/OMf4W2rpbCNxETlEYx8S0r8FbYg7JZZItFU2G63IOyIIBqETR1CWpaIHmFrKez4eLnDMhQPW8t3jwRht5bCDjVLBGBlFWnC/vRTNr1tOOer5WEbUdhUHlqpYTyOHRMKW42WdjoKhd0OEMwSAULreNy9G+jWjU1kXlnJJqJJTWXfaaX1NTez38TH+xN2crJ84fi7v9stkwh/8fUUdmMj8Ouv7H2sKexIdTomJLDyUv+mJWiJvdJShR2IsJuaWOaJIGwlyMNWK2z1aFp1HnY0FLYg7ChBLw2HJ5JQCLuoiPUU79rFGmpampzBoKewrVZW0dSEnZQke7R6j1tGPOzNm1klSU7ueB52NBV2Swg7FIX9xhvACSewemNEYZeVsd8KS0QJ8rBjQWFrrd1qswnCbjEibYlQZ2FpKVPYZIcA+h62xeJ/IYmw6cIZIWw9hU3Do8ePj22F3ZIskWh42LSPcFYIoXLmOx31YvrtN+D339n5G1HY/KAZIHTCfughYMMGY79tTyBLZPhwoFcvNtMlEHjyp6YmmKm9tIbCttnaL2E/8cQTuPTSS3HRRRdh5cqVkYopNBgh7FAUNuVel5SwhkodjoB+lojFoq+wtSwRvjIYUdhHjrD9HHecPFQ6HARS2I8+CuzZE/o+W2qJ0BNINAibyjwcwqY643Cwm3eg/fDzJFN5BCLflhL23Xf7z53T3kEzPsbFsbmBCguNDU2vrISJ2kM0FbbRTsdYTev75ZdfUFBQgHfeeQf/+c9/8PDDD0cyLuOItIdNhF1ayh6FeYWtlYfd1BSYsANZImazMYXtcDBCS0tjv6EYQ4Wewq6rA+bOBZYvD32fWh58MMLWGumotkT4QQzhIlyFTXPTpKSw/2kuc73zIcJubDSmsNXZR6EQNk1nG6xTs71BPaqYf+9ysWuiRdj8LHotVdh6ooq+I8JublbygCQBt97KlgKMVYU9YsQILFmyBACQkpICt9sNT1vMNhZpS0StsNWWiJ7CDscSSUkxprCdTkbY1MDDtUX0FLb6kTMUBFLYoXjYvMLWGx0aKsIlbIeDnUPPnux/uh7BFLZRwqbrRymfoRA2nVNHW7CY6hGVBSCrVJdLWWf5esXPU91aWSLq3xYVAUuWACtWRJ2wrWFvaLXC6lWP7733HsaMGQOzWcn/BQUFYe3b6XQa3ja1sBB5AJw1NdjPbZN9+DByvO9L9+5FmcH9Ze7fj1wAjkOHYC0tRX1+Po56t+3tdsNjNiti61pWhhQADY2NiKutxT7vd/0qK9HYpQuK9uxBPoDSo0d9McQXFKAvAHdCAkwOB3Z7Px/idsMEwF1ZiT3cMboVFyPRYkFpXR3yAOzdtAkujrQzX34ZcV27ItgZ9m9ogB2As75eUVaW0lIMAlB25AhKCwrQ9d57UXPuuWgYPTpoeWUdOgTvgz0O7NkDR0oKco4dQzaAJodD81om79uHngD2Hz6MHJcLyQDKGxpgdruRAcCVmgp7VRV2bNkCST2ZfQjoXl6ONAB7CwrgUt08AtUxa2kpBgKoy8gAl7eC6tJSHNHYpuexY0gGsGfbNuR668ORvXtRrbP/zN27kQtgZ3ExPA0NMDU2YgiAkoMHUV5QEDA2c10dBgOoKS5GUZjtK1yE0i5DBdXBoxUVqOKOMcRsRvmRIyjbsgVDAEgmEySXCzu9v0nevBne2yqqysp8bTUcJOzejT7e98cKC1HB7atfQwOcDgeclZXs2m3ZAo9XYCSuW4feAMqKiuAymdAdwJ7CQjizsiJeXmETNuHrr7/Gu+++i2XLlvl9l5+fH9Y+CwoKjG+7cSMAIF59vJQUn0eaExeHHKP78z4GJ9TUALW1SO/bF+m0bUoK6l0u5XFSUwG7Hak5OcCuXfJ3bjfiunVD6tChgMmEnPR0OQavh2nLzASOHmXbSJJPOdgaGpTHsNuB1FTkHXccAKB/VhbAf//uu6jr0QPJt98e+Ny8a+XF22zK/XsrXnZqKrKHDAGWL0dGr17A9dcHLy9SiQD69OjB4vIqZKskIT4+3v9a7twJAOg7eLDvt1l5eT51as/LAw4exJD+/ZV9CKHCS/b9e/ZUlheC1DGv8EjOzwd++MH3cVpiItK0tvFaNwN69vQ9xnfPyEB3vf17VeTgkSPZ05l3+y6pqeiSnx84Nq+iTLVakRpm+woXIbXLUOG9Vt369EE3Vd3PTk1Fdv/+AACTV+T44vjpJ/ZqtSI9KUluq+GA+ioA5KamIpffl9mMuMxMpHmfugb37St77GvXAgCyU1KAHCYTB+Tno6CmJqzy2ujlNC20qNNx7dq1eO655/DSSy8hlXKVWxuBLJG4OEao4VgiR46wR9SWdjoC7MahZYmkpvpPyGQysUds3l4iD1vPEnE4ELd/f/Bz0/OweUukqUleSNQIomGJeCt9RC2RbduAm2825otT+ZIlQjBiiRhJ6yN/3LuKEUwmVn9CsUT468Pd7Nst1FMZE8hqpPNOTGTnS3WMLJHc3NYb6Qgo6wItnM2vbRprHnZtbS0WLVqEpUuXIoP3eVsbVEBahB0fzxpGOJ2OtF+Vh62Zhx3Iwwb853zW8rDVx9u+HVi1Sv49dToCmoRtKy4O7mvqedg8sRnxYHmE0+molSXCdzpmZwfe3ij48/rsM7Ygc7A5PgC5vqgJO5ROx0DkW12teDIBEHyZsPJyZdogT9hffcXqTUty9NsaWh42oE3YgPx/eTk8NhsTM63pYbc3wl6xYgWqq6tx6623YsaMGZgxYwaO0ExrrYlAWSKksMMhbIIqS0Q3D5tX2G63PNDFu51mZeAJm86DBgvMng1ceqn8+/h4uZHzDVOS5P3t2hX43Iwo7FAJO5IjHaOlsOl6qOPVQ6gKm57gQul0VD+RBiLsxkagf3/glVe0CXvHDnbT4B7p2x20skQAQ4TdnJHhL4rCQTDCpoEzgD5hRzmtL2wP+9JLL8WlRChtiWhZIgSVJaKrsHnCJqUbzBJJSWHWh8cjf0+E7fXFfIScnq5tibjdsn2yYwcwYoT+uRnJEqFzCMcS0RrpqGVB6I10HD0auOgiYOBA5e/CBZ+HbSQ/mqClsJOTjSlso5ZIKAq7qoptc+SINmFT/W7PIyXDJeyyMjSnp8OmFkXhINhIRxo4A8jXweMB9u6VzyFWFXbMIBhhh2uJEIwobBqaTvmiWoStZYmQynK75e/peJRv63bLHnZiIrs58ITNE8OOHYHPLdoKWz35k9axAH2FfdxxbM5vKrdIWiJahH3vvcCbb/pvR+WbkyM/onfpoq2wm5rkfRpV2DU1oRE2xcPvvyMR9ujRwLPPsvfhKOz09NZR2FqWyKFDct3i4xSErQOehHhyoPXhQrVE6uvZ5E+EUPKwAXbB1IStZ4kQYfMXmhQ2Hw952CYTa+i8JcJXLG/2hS6CKWyjhMMjkCUC+JcXfzw1YRO0RoeGg2CE/fLLwFtv+W9H9SUlRb5GeoTN9xtES2FTPHw94Qmbvm+PhC1JLNPj/ffZ/6F62F6F7dfG9LBxIzBjhvYCE/wxgnU60m/50cGksC0WX0ZWpNFxCBtQkkdLFHbv3vL/oVgiALvQoVgigLbC5uMhD5viCUdh85kERiyRUAibYtPYvyZh61kiBK3BRmps2sQWCwgELcLm60hdHVBc7L9dTQ1rtDabTKxZWdrxkB1C+45Gp2NHVtjqPhwthc2fNxE2/T5Uhb1qFfD669odtLR9SopxhU3+dV6eTNhRUtdAZyDscDzsnBy5YqjT+vQImxSiHmFrWSLUKel2+ytsqhj19bIlArCGzhO2t+G6c3NZp6PeaFP+cyOWiFEP2+WSz1PrhhCMsPm5RAhGCPuBB4ApU4D//lf/N4E8bEliZHv0qP92fKdgWhp7n5CgrbD1CDtUhR0orY9X2Py8GuqRse1xuLr6nEOxRDweoKKCdToaVdjUNrV+S5+lpirLktIItTodd+5kN9t+/QRhGwJf8Dxh85YI/6gaDA0NrFJ06cIaKV+BtNL6aC4RUogOR3BLhGLjH6+IXPLy2HHPP1+OR03YGpaIMz+fvT90SPu8+ONH2sNWe85GFbZ6xRmCEcJ2OllDuvZa/ZnrAlgiJso5P3bM/ybHe8ypqeymrbf4Kk/YRp5QXC4WQ0sVNn+M9myJqNtlKIRdXQ00N4emsI0QdkqKsiypPmsp7A0bWEc/zYcvCDsIgilsusBGFSNP2Gp7oiUKW+1hx8criYm+z8pi2QAzZ7L/a2tZ5Qhiibjz8tj/eqldfDmpCaqllgg9KRi1RKqr2bWxWMK3RBobWV+DxwNs2aL9mwCWiJnqAy0moI6PFPbgwexPb1rNUBU2XbtQ0vq0FDYg1+n2bImoYw7Fw/Zet5A8bKOErTWjppqw3W7miY8aJWeJud1RS+kDOgNh0x07VIWdlydPME8INNKRCFtLYWtZImrCpu9pEABtW1HBXvUsEW/FaiYrhScQHlTpTKbIDpwJR2Fv3y4PFQ9XYbtc8gAbvQFDfB62SmGb+Ru42seurZUJ9emn2aAbIwrbSKejeuInQqBpO3mFzcfQXgnb4wHOPhv44gu5vKijPxTC9o5ybIq0wk5NVZYl7VdN2Nu2set86qkyYbtcQmEHRCDCjo+XK0CohP344/7+qNb0qlqETQ0pkCWSkKDM6eTv4vy2pP6CWCJNwQibyik+PrIeNk/YWml9Wo1o61Zg2DD2PlyF7XLJfn+wc9awRAISdl2dctATWTfBFDbN8kfvtUBqWU3Y6qcwrW3aq8JevRp47jn5/7o6YM0atjAHxTx/PvDvf/vPHRMtha11LfUUth5h08IipLApTkHYAaBH2DTSMRSF7fGwC5WYyFa7IFIhBFpxhp8ikxoSVS49S0TLw6aLTdvSXAlE2OnprIHSjSNUhR0XZ4ywm5qMKZYgnY5+Cru6Gjh8GBg6lP2v1emoNYe41nFTU9lETcEUtpaHzRO2uuORJ2yCngLmy5vv3A7mR4eisImw1R42nUOse9ivvMI6iQkUp8Mhv+/TB/jzn/235YkQ0CfsSCtsLcLms5pcLrYwdlYW63DkLRFB2AEQSUuEKo/elJ5Wq3+lCKSwecIOZonoKWyyROiG0KMH62yjlDZS2OS3t1Rh8+VkxBYJ1RLZto290s2wRw85E4NgVGHTCtvBblIaHraFJ/lACpuPKZglwhN2qJZIIIVN22gpbI9HJqFYJWynU1ke9N7p1J9DhGDAEololojJxOqUEUtk/XqmrmnyLkHYBhDMEgmFsNVEq4bVCpNRS4QfxqqXJaLnYfMxqC2Rq65iZDdzJvsuWgqbziUYtCwRnrDVx9q6lb0SYV9+OXDgQHidjnY7I9YwFLZhS4QQTGHHxclK12yOLGEHUthqSyYWoSZsLYWtzg4hqAmb2gEpbIsFHppKORIK22Zjx3A45GkVtDodq6pYX8yoUXKcgrANQGtACr0PVWEbIWwjCptPwwP8GyNlIQRS2HqWSFwc8NprbE7txx83Tti8wpYk5RwfeoRtxMemG6PFoj2SUl1eW7cyMuzVi/1vsfj7lkYVtt0eWGEH8LB9lojVqrREPB523lqE7fFoL69mszHfkxR2Wlp0Oh21rg8/KCyWFXZjo2zjUZwtUdiU3ZOZyRRupBS2zSa3EfWgHp6wCwvZ+QwYwP4XCluFigrg3HOZGuOhpbBpmDpP2EYqczDC1pte1Wr197D5fajv/mVlzPsKpLDj4phSUytsADjxRDYx0aFDMmEnJbHKZERhU9yElloicXFKu4hX2OqGsXUr86/NAapeqJZIKApbndbXt69SYdPnWpaIVkykxnmFnZamJCgeeml96pu62w1MmwZs3qxU2OosESO+eSSwfTssRqam1QJP0IBcr8JR2GpLhDKF+DZWUqI/gMyowubj1CJsmpmUFjHgs0REWh+A335jw0pfe035uRZh8zN/haKw6WIGUtiS5D9qUMsS4fehvvuXl/sTtlphm0xsH0TYagWSlsYassMh5zMnJxtT2BQ3oaWWiBZhewlZ08OmDkc9hKKw9c6ZH4qvZYnQtR44UEnYtC8tha0VE0/YvMIGtOtcTY0yQ4igJuwjR4D33mMphYEUdmsR9nnnIWvp0vC2VRM1T+D0PhwPm9oRILexmhrWgam3qnwgwqYpVNVrbPJiSk3YdMMgnmloEAobgJzK9tlnys+bmvxT9/hKEEpanwFLBICS7IwQNt8YqVJlZys7MNQKG2DqUUthA3I+Nm+/BCLsaChsSZKJkz/HpiZfTAoPu6aGqZ/BgwPvV4uw1Q2Mt0S0FDZ/o+CJjvewTSbWw3/0KDvvpiZ9wlZPq0nQImyyeLTKT2tYOsDOhZ+OluIoLAzsYbeWJVJTE77CVuem869606oSQlXYVVVsv5s3+++Ln0lTL60vmMK2WJgY0SPs+npB2ABkwl6/ng0nJjQ1yZ1eLVXYRiwROiZ/fPXQ9IYGfw+btqGsj2AKm+IIRtj8zSFchU3HVi8NFszD5m8CaoXtLXeFwqZRmOoBSWqoCXvLFnaN+QUa+E5HrXPmj6s3cCY5mQ3YqK5mSvv661umsNU51qEQtvqciVgOHPDPw6Z6yCtskym6hO12wxxup6aewuYtkUAKu7lZ3jaYwqb9aS2Z53IFXlxDj7DVbdNul/s91IRNfRpRQvshbP7uvmKF/D4ahE37U4MumHq+Z4tFtiWoR1zPEiECDuZhUxxqoiVEQ2EDoWUd8OXMK+zmZjlevmEQYdOKMnpQk1dBAXtPKYEeDyuXQAqbP66eh52cDHTtyj4/dIhNRB8JhR0JwqY4tm2TVTcpbLJU6uvlY2Zmtj/CNtrpCPiPHlYTts2mXH1J3dfF74O2V4PvdOTj5BU2xdTcLAsGPk5B2F5UVTEVkZcHfPqp/LkWYfN37QhnifiOSSDCBuR0oECWiB5h6ylsQkstESMeNhA+YasVtvc4CkuEMl5IlehBTV5UZuQ18zcfvbQ+dfYQ/a9W2OPGsdTCU09loiCYwg5E2BSHFmGXlADXXAP8/LO+JcLvn+IgJZeSopyrIjFRqbBzcqKb1ud2K1MhQ4FWZyO9Gul0BOSypbpeWamcnoDaDZWblsI2Sth6lgjVS3rNzpbnvRYKW4WqKuYNXnghU9ikuFvBEpEkxhmllaxSFBc1y3xnlLC5+XsB6Ctso4Sdmhodhc13YgVroEYIW8sSCaawqTz1CJviDZTWxzdI/nu+0zE5mXVQvfkmywsPRNhGskQIWoT9zTfAq6+yDJ+bbvKPV09hE3JyZEuEJ2yyS3JyoqewPR5AkmAKd/9GFLYRwqYBKoB8I+MVNiCXW3m5/7TKRghba0Y+LYXNH5uP3+kUhA1AJuzrrmMXmZZ2crtlYgtE2CGk9ZU1JOK//wX+9jfg//6PHTY7G7j7fkYkJw5vQnw8cMIJQEVZMw4XW9i15QlbPdTaqMJWWyIELYXd3Mz211YKmydOtSUSyMMOQtj7D5jgsdjw4//cuOce4NevWZl5jhz1P25ysnIODwLfIPmGy1sitIAEwGZmbKnCJhBh83WOiPXNN4E//MH/pPU8bEJOjuzRqhW21crijxZhe2OKmCWiVtg2m36aJ0/Y3NDw/T+xm/cNC7Jw8smDcfs8RqZP/pOrv2pbxKjCVq94pEfY/JMif/1FWh9YY0pPZ3PPjhjBlncC5FQcfhHcUCyRigpfYyrcwQi715BEXH018O677BrNmAE8+SRw1bXsgj3yYBNuv531VzU2NGPlVxZ06QIcqYxH2WEnpFAtET5LREthm83KzwGZFIqLI+9hE5GFa4nwmTtqwo6P1+wjOHKETTcxZAhL3HA02/Djd27cfz+w5TtWZt+8XozHHweaGlQKG/B/GuCPq5H65rNECER4ZNsY7HSU6upQ40lGabXcSL/9NR0AsPpzBz75hC2OU1/sjYG/SWjtX22JEOgmV1/vT9gpKawOxCJhezz+M0Cq0/r0/GtAQdgeqw033crazKF1RQCAtMFdcemllThzHPt8z2a53A58q7JFjBK2ei4brU5HQJ+wo6iwrcF/EiMghQ2w3vybbmKpO7R0D0/Y6s4wk0mfsC++GI7UXFxlfxvHv9eAO2HGTbfZcellwEknqZZmW2YFlgFXXdkMeFcRk3Kbce5JFvzcHSh9JQEHv3VgjNmBA3sS0b/eyye8JVJWJg/4UA99BbQVNq3nyIMnbMprNqKwjVgiNMFUSzodKa1PPWCI9/28H913H7B0KQvj7LOBG28E4u6y4earmvD3Z4HGc8qBb4BetqM493Zg9X8a8Tmg7PSpr1eSoZ7CDkTYAJuYCgjY6VhezpYg/OwTDz6qq8eSl5MxAHG4wvvT599Kw9kAnn7EgY+9n92LWtwDIH9kEvr0k6fZHjSIveZZbTDxcesRNnmkvCWSkhJ4Pu2Wggg7HA9bK6+fV9yUx68HLznu/b0emQ02LHvdhmcBnNrtEHAUWLQsFwWOEuR/ZwW+Bp56sA64hW36zG37cUoX4LLLvPtqLYUtCBuMsAcNYu8nTGCvW7cGJ2x+YhYN1O84iKKSYnweB9x2egPMWxLxyKM6C2hqdDqampvRu58FL/8LaN6egOxqBxJ2NOCzbxPxWE/gL38B7nRZkdrcLJvhWVksLv4xmEhMS2Gr7RBA6ZPyCru+nqka9SNmKB52UhLzkaPhYXN2yEcfAX/9KyuSa68F7rhDHumLf9oAuAETEF/HFPbgtGK8/wqw+HoW767COAzqr+psUp+v+ju1h00gwj50SLlGJ8HbSB972I27vmJFNjfnVZgh4awZvTG45BDwJfvp0rfTgMuA5x93YMFZ7B7Qe0kNnD+lYOhwM/buBf73P2XxXhVvw6sAFs5zo3FANv62vw59vZaeqaaGLahB56JW2Kmp0SVs7zU0haOw+Zi0LBF+vVINVDYmIgNA+d5KZMfZULDTAvQB4sqYwkZuLrM+vG3J6pSv9am5B3Dp5azazZ4NJWEHysNWK2x1p2MbEnb7s0QAmaBosEMgwqZXFWFLEvDoo4CjuBq9pf34daMHp5/QAJNehgignYfNdTpakhKQl1ILm+TGldcn4Oyz2TEefYpdwB/WuCGVlft3lATzsAMRNv89EZAW0YbiYdvtsh8fCIEIW8/DzsmBJAEPPsj6j/PygF9/ZVMh+8ga0LaRiotx0QUeLH+DxXv/Ijv2HPWWkdrzpW0tFpmwLRb/tD4CT9jJyYqngMZG4KVX2XX59ksXrr0WKHjzVyyquQGYMAFj/3M1uvWRG2x6b3ZtuqU7MHIkcMEFwIgBtYjPTsF777Hzratjh/rmG+Bf/wL+72y2/x/XuPDoo7lY8V49yhuTsa2OPcq9/Q270Xmqa7Utkfj40LNEVq9mBw8GUth857je/r76SvmZFmHzn9XW6hL2gQPAjLv7AgBOTNiJtCwbevU2yXUjLU2u++oskZQUXHzyflx4ITBnDrBsGUJX2FpziQDahM351l98Y8PPP2ueUovRIsLetWsXJkyYgNdffz1S8eiDt0T4PEktwlYn46sIW5KYmps7V0K6qRpxUiMGJxexCxqIsPXS+ujz+HjfwJje+Yl4/31gzx5g1GhWAc4Z14RfvihHQWkWXngB+OFnCySTSZklQjcFQI5Fq0Lzc1FQpSVLIFDWhJbCdrtlRc4rOKOErdXpqOVhey2RuXOBBQuAK68EfvoJGD5cY99qwqYbQkUFenVlhJ2QZsf8B5O1z5mOy/vltJKIJGl3OgIyYXuxZw8wejSw9FXWIF99yYUXXgCGfP0sTAkJwFtvyfWPoNXpSMTqhcnEZpYdN45ZQH+Zxfb/zRdu/PDDLlwyqQ5xmUlAr97wwIQfd7LJvQq31+FIuR2ehAgo7FdeYXfOYFDPX6KHe+8F7r5b+VkghQ2wdq1hiRw7BowdC/xay+7idkeNf1od5dDzn1E9GDIElsL9eOst4Jxz2OSW238JQtjUH6Zu5wayREpq5HMo2GMzNHFgOAibsBsaGvDPf/4Tp59+eiTj0YbbzciUCJvvSAxDYc+fz5TvzTMdsErekt271z8dT41AQ9MBRpw0ktG7n759gfMvZNu9stSNbnHl2FORhRtuAM44A3BJNvzrKTfeecONZrMVh4s4OyZchR1o5J+ewqZta2tZeRlR2PxgBoOWyKZDOXjsMUZSr70WwL4kwqbhxmSHFRf7ngjufcgOh5mVUflBHYXNE3Zami+VzNTcrK2wjx3zbfPDDyw9+8AB4JEnGCFkp3r3+/vvwMkn+490A7SHptfU+E/4pD5fb9wZGc3omlSHlK7JGHbJEJi75mLJC+wap1tqsXOfDV/+LxFNtSoP2+XSn/RIC3xaXSDw5KY30RbArhM/VJ6OQdBS2FVVfoLE6WRPJSUlwMdfJ7HHMEA5XTGgJGy1wh48GCgsRFwc628YOhR466UwFXaQTsevvgL+cIl8/WfdYsMZZ/jvPhIIm7Dtdjv+/e9/owt5a9EETX6jpbApdzIYYXsrydNPA4sXs5S9J+/jKpcRwg5iiSAhQc4PV2eJAPjThW70TCjH1KuzUFgIfP45YLLb0LeHG3WVTXB5rOjTB5g+nYVjyMPmjxWIsIN52LStxyNbIsE8bOrMo6liA6X1NTYCtbX4YG0OLr6YXQd1P6oCtL/KSvZIRB2rNO8HgLy+djzyHIv7gXl1iqUu/eaeoDgbG7VT94iwJQlITsbq1cD48UxE/fILMPZcLoujuZmNQOQfDYLlYasUtub58nFTuuDChcDatTDHs+OnW+sweLgdhyoSUX2kAa5yjrAB40vhAbFB2JWVfnftu+5i44tefx0YORLyzVqtsPkpDtQKu2dPxhtuN1JS2Fi7NCuLXUpMNJYlYqDT8eWXgfPOAzK7yedgT45eWl/YnY5WqxVWdaqZCgUFBWHt2+l0+ra1HjsGk8OBAQCKGhpQU1AASBLyAZQePox0hwP19fWwezyQSkvhGTcOJqcTyQB2FhbCU1mJvgBcZWX477OHcMstPTB+fB1mzTqMfb/uR3/vMct+/hnJ+/ahKTsbh3TiTj56FD0B7N+9G04viQ5pakJ5ZSVKCwrQ1elEhjfz43B5OWq9+0kvK0M3ALu3b8eAigqUQ0J9fQH69gUs8RacNqIYktUK23ILpv+pHO++m4H33gOWndWEywHUSxIOqmPyeDDEZIJJklBSVwen04nC8nL0BnBg61Y4VI0go6gIXQEcqahAdwD79uxBo7eS9/NmHtAWdW43rCYT3KWlOBzgGmbs2oWuAHYVF6O70wlLXR0OFBRgkMuF6ro6ZJhMaPZey9odZRgFANlpmD9/B3btknT3CwB9m5vhqqhA6fr16A+gNDcXOQCKNm1Cc3Y2egE4cPQo7FlMbVcersN559Xh+ecPMUdnzx70BuC0WkH6rcFqRbzDgX2//ooBAI7U1KCazq+pCfne35U6rDj//Gb06uXGK68cRGNjM/YcPMjq4IEDcKxahQEOB45kZfm2z6qpAUmXggMHMMRsRvWGDXDcdx+qpk1D39JSuHv00C3PhCNH0AfAwT174ExLQ0NpKSSbDQeLWOdackkJeoJ1/KVkuTBuqgtZH5UBJcAvrp7oWV2NrgB2/vYbPFojKTXQs6ICSY2N2BGkncbt2oV+3vf7tm5FoxbZARjkFSu7CgqQ8dprsB05gppJk9DX+31FURGOFRSgZ2kp6FbZVF6OxpwcX/3evj0eS5b0wbRpVcjPL0ZBAdA1OxsZAJweD/YXFGCgyQQrgAq7HccKCuB0OnHI2zZri4uRAqDYbGZ186ef0Ozt6J569hE4PouHqzEOrqMlKFOdd//6ejgaGnBs/34MAnD00CFUFRQg/dAh1n4PHECTw4EeTidSAPz7Qzf++pCEM86ox8N3HAKmsv0cq6xEhTeucDlQD1HNEsmnlbFDREFBAdv2p59YnteLLwIA8oYORR7tMy4OOSkpgMmE9OxsdnffvFmxQO3g449nSjEtDU6HHXPn9sTIkcAnn6QgMTEfaJDTvbJ37AB27gSuvVY/7oMHAQB9e/aUV/1ubkZ2bi6y8/PllZ8B9Bg0SP5Nz54AgIFZWUBTE7KHDGG/B4D4eGTSihnxdrzyShYeegi4+Wbgw/e64XIAtuRM7ZhSUoCaGnTp1Qvl8fHo7VWhfbKz5WMTvAscdO/Lmk+/3r2Vv8nMZDPDAUjOzAQkCfEWS+Br6FWvg0aO9KUC5ufnA5KEzJwcwGaD1WRCfn4+/jFnM0YB+PP87uhz8hD9fRKSkxEfH49U71NVztixwLPPIs9k8j0K9xk0yFe2f5tej9P/m4wnnsjHK68AJu+1iqeFHQAkdusGbNqEAd6nwu6DBqE7f37eRQg2781GTjcL1qyxoHt3r7rz3qDzuCHg3SdOlLfv0YO92u3IP+44ICEB6R9/jPSPP0a3GTMAlwvxeXn65el9WunVrRvq4+OR6PEAXbrIv6d0QwApmZlIGZYHfASUWLpi4nt3YM31r6ErgMEpKcxruv9+/9x9NSwWdqMaPDjw/OTck0K/3Fz/ugUolirLHzIE2LgR2LsXWdde6/tJZkICMvPzFXFZa2thzWT1u6mJjXnIyQFefDEDGfTUc+qpwPLliE9JYeXhvRaZxx2HzPx8FBQUoGc/dktJAQCbDV2PPx4AMCgri2uHVjSmJMFZa8HWjckYNyRf+ZRnMsGenY00bzvqlp2Nbvn5vsymgfn5TNV7vet/LB6OM84w4csvk5FQN9S3m9y8POR64wqHAzdu3Kj7XWxniWzZwl6/+IK98iuTxMf7e9hE1vQY7FWZTsTh9/WNyM1lj0a+p2R6hk5OBr7/nr0/7zz9eNQeNvmFvCVC0LBE/IbT0neUJeL9XffubNDOzDlsH2vWJ2jOZeN79I6khw0YzxKprZVnKlRbIt7JsExuN778Etj8NRvl2OeUIMPSCbQ/yhDp3Zt5y5yHzQ+cOW14Pe67jy10f/fd0LZEqLxonyqLoimVEUS9KRmrVrHroCgTePe7dSvzc447Tv6et98A5m/SsUtKwrdE1PunWLz7Trx/PvoPT8STz3uv63vvAQ8/zDz2YFBbiHowYonU1MgrGdXXs7ZVW6vf6UhM6fH46uSzzzKef+op2aECwGZSBAJ3OvIedlyc3MZoIJQ39riMJCSk2nBglwsPP6xxnlZr0IEzZTV2OBCPPscl4tNPvc1PpPVBHlq6di175QmbfGmesAF2N/zwQ5bLYzajshLYXBAHm6cRK1aoZvYkv+3EE9lrXp5OyoIXag+bSE+LsNXTqwLyXBhahE3n4YXJBJxzASOj6sYE/N//yfcvH0IhbKMeNmDcw6aONFqiiS8XL2F7Gptx003AsK7ehhNsHhGCmrAzM9kTzNGjSsImUqyrw8KFbEzVAw8AH7xjgLC5c66oAPaUM5Y4Y1KyjyMU8QDs2Fu2sOGYfIcmlSsR+9q18iRlZWVy56AetEY6qq8JH4t30qrk2/6C1auB3oMZ6e3++gD7DU+s69YBP/7of0wi00gQNt+BUFPDxJMeYTudyj6Y+HgcPMgyh847D/jTn1T7DtXDjo+X2xhda4o9KQkpmTYM6uvGXXex+5viPIOk9W3dCrz/XTaO2Hrjy69M8mmor0+UEDZhb926FTNmzMCHH36I//73v5gxYwaqODsiIvA+ovvukvxtl9KY1IR92mlsKbGnnkJdHTBlClDliMPQgY3+T3JUyU46ib1Onhy4J0zdGRGIsNXTqwJMaQE+ewKApsJW72PCVNYYzzpL1e4o6yAUhU0VKxhhG0nr4zMf+NGcRNhWK3ZstWLPHuCvFzKFHXSmPoKasLOyWAMtLlamE9LiEfX1MJmYe3bFFcDbr7PGJqnT+gB5ThPvOVdWMqIobmT1K7u3apQjHQtgMW3Z4n9jVyvsnj3ZxFIAUFTEtjOYJQKAkUsghX3KKWxekvh4pKQA8+9jdeTw9wfk7Qnz5wNz5/ofU710lx74TnY9wubbfm2tPDEZ1UW+PjmdirYsxcXhppuYOH/uOY0m2K8fs2xCUdhUz1QKG0lJMNntGD3SjdGjmQXz+efe74N0Ou7aZ8W55wJPZtwP+5pVSvEX64Q9bNgwvPbaa1i9ejW++uorvPbaa0hXL6baUqh9ALXCVlsiAPO7wLjkD39g6x0cPzIOqTaNSkmEfcop7HXy5MDxBCNsPj1JyxIh8uEbro7CBuBTcBndE/Djj0ycnnMOG2wBIHSFbbPJsQYibKNpfTxh0wRXkuQj7CazDbu2W3DppcCQrFLW6PibVSDwhG21suOkpTEyIBVK15wbkm+xsEnxJoxh1+jb9aq0PkAxX8iRI0ysbt4MDD41Q96fVjx0zrt3+xM2NVi+4dLTBE31adQSkaTglogKcWms7uUnMpHz2TscsVZVadeJcCwRvacutcKm/3mxxVsiHGHvPxKPTz9ltjvd4xSw29kXgdL6QlDYsNlg8bjx8ccs+eiPf2Q3CinISMezz7WhuRl4/+s09BzdUxmj1g0lCog5S2TnTuD88/vhssuA+u0HIFHltFqVJEgKm0/rA4DTTsOhQ0yN/u9/bGxA977+Ix0ByJXqT38C3nlHexY1HmoP26jCpgtIOdpqwqYlkHQUNhIS0Ls3O59+/dhTw6efwp+w4+KUI/t40A1Bi7Dd7vAsEd6XJUvE6+tLZgsqamyIM7nx+ONgDScjI3DnFg+esGllbFqsgLdEAL9FDKxWYOY1rLFt3ScTtjvJW17eDrw3v8zC8OFsIZtPPgG6HReAsOlY+/ezsuvfX/m92hKh/djtwL597H+DlojJO7hHYbnw+9WaDc4rFnJdhwAAy1+pkz3a2lrtaxktS6S8XCZnIuz0dF2F/c33cRgxgnW062LRIuCWW9h7mh+It9fUCjshQV6xae9e4KWXFIQNtxvZ2cC337IH8lmzAHeDG8XlNlZHzWagqQm1tcD/VjPCtsZbsXZtgCVJtepAhBFzhN2tG3DiiQ78tKYRCVVH8YVrHACg3p6OrdtMvvmSFJ2O3tn6JJMJy7aegmHD2DX6/HOW0+z7rRrV1XKj+tOfgpNJuB42VSYibL7h2u36CluVh921K7BmDRN3F10E7ClTEbbJpD8BlJ7Cbm5mJBtOp6PaEnG7ffvdvtuKWqcVQwfXsXEPlZWqnqQg4Amb1BKdmxZhq87Z1MRIZtrV8o3zxnmsvDYu34d6JOLKWWno358NFZ84EXJ8WoRtsbDy9abZ+XnxaksEkEmFCNugJeKbZCkEha0eqHT2KfW4806Wxi21BmHzlgiX0eKznwIo7IqGeCxdGiSpZdo070UCK6vsbKXA4SdsoifdrCx2w3j+ee9Qx+0KwgZYU/zsM9bhaZGa8O9XbcjPB1yw4Z3X3ejaFfjuGzc8MGHDJrPPTtcEXaPOpLBTU4F//vMoDqw9BDMkmM9nyY1HGtIxfDjrMJ41CyiuikNlUQMgSdi+24pnq2fggdRHcd0tqRgxgj3i0vXVnfyppkZ79Q89hOthqxU2T9iBPGxqsNy+srKYJXLaacAHX6sIm7YJRWET+anJgTxH3x1SA2pLpKnJt99PP7fAHGdD355eoogEYespbK1VZ7wNstsAWaWeeBaLtZtjH6qTumLlShN++knu0wpI2DRZFy2+qh4wpqeusrNDtkR805gG6nRUQzVa8KqL630dsK6yGjblrxpGLZFAHvaePSzdlVfYhw7J77UIW6WwTxgVxwbIGIXN5r8uKM/2fKYOKWw6bnKyMqMJTKfNusEDCzyYMMmGgQMBt2RFU2MTrrwSuG5GE8w2a/D+8s5I2ART4QEAwMTbhgL9+6P3iRl48UXWuP77X2DL7njs/Z0R0+tvW3Hbu6dh7ai/49NP2Rw0iidWPcLWW19PD3qWCD+XCKBcGQNQEnZCgrJyBfKwk5NZTu1VVyk+Tk1lmY7ZA1jsz7ycKI9IDlVhE/klJMhPGKSwaVV0PdA8FnQeHGFX1FjQJc8Gc7O3sVdVhUfY/MrYBi0RAP5ziVgsmDWXve/edBBp+ZmYNEk5dUtAwqbjhaKwARa71o1aa98A4HK1SGETzI56LF0K3LewCXHNDriqGqAYwyFJxjsdAynsq69mCsooYTc3Ay4XGhPlunD2eQHmw9bCBRewiWh48CSpVtj0hAPICltdr73nePpZNnzyCZCUZsOV09xYuhTIy20yRsJ0XTojYfsyRPr0AebMgf3a6fjLX9jqYNXVwOixccjvxSrPjbOtqKpiY/qnTtVwNiJF2EYtkcREZVc3b4moG20ghQ0wT0eREMyQlATMuInF/sQLCbjhhp5sMfnkZP/hwRRzIIVtsyk7zojoAg1F5lPVvJ2O77zJ9jv+HAuS0mzy0PRwFfaxY3LnUlISi7e+Xrmog9ZNSp2Hza8+1NyMJq2V24MRts0mE5MeYavJlP9dqJZIKB62evqC+nqYzcDdt7FyiZMaMXJEMxYu9FaPpibl4r6BEKjTsaKCqeyqKrluqQnbZmPnQvNfA/hoTbrvJ3GpqptcMNxyCzBvnvIzPYVNhH355SyGrl39FLbiHPmOQz5LJNggJP64nZKwDxxgjTIvj+VUz5nj+8psBpKy4pHkYZWxRx9rwClAIq6wgxG2uvHwClvdaAMp7CCwXTAV0s234I5ne2H9+kQMHw4cdaRB0iLsYArbbleqRCJi9bp4hOZmRpKcJeJxN2HOLLbfCRNZWp+CsEPJIrLZWAMvK5MJm4i0okJJWlrrOqoJ225XqNAmredbOg6fJ8+DjpmY6D/nTCBLhGDUEomAwvbdaLnr96epDXjgATaN7dKnOVVtkLAlq9X/Bt7QwCyRqip2c7JY/Ak7Pl7uE/Gq+R9/S0STxa4dezjgSZLKKiuL8UhdHfMQCwqA224zRtj8sn6CsA2gsJAN99U7eX6V6mCFGRfHCEY952F1dWDVo4bepDBaCpsHnxKmbrR2u7wadqgXundvmJY8ib/NsuDdd/ejTx/gp+1p2LepGitXquznYArbblcqbCqXmho2RcCJJyrVFRGk93e/bWf2x/B8ViYWu3ekIym5cCwRepwmNUyKs7JSSVpaOePq2fri4pSErTVp2dixwJdfymmeWjEB2oN/9CwR/reBCJsGH7lcMGkRtsUiPzoa8LB914cj7GX/asD69WyA5l23y4RddcwYYTenpvoTNq1hun8/uyGnpioJu6LCR9iSw4GnFrHrdNLoBFiSE7RjDwc8B/CWCLXRfv1YbnxCgtzRr3GOCoXNj3QUhB0EBw6w4ch6iI+XK6URwgb8lUS4nY56aX1UUdSEzcenbrTp6ezGEYbC5jFwoAs//QQMHZ2GeFc1Jk9mvPPWW942ZkRh6xH2+vXAb7/JHW6AjwiKalNx0UXAex+z2N99zVvG3pGOaGpiDdrtDp2w6Y7DWyKAP2FrZbSoPWwVYbu1LBGTieV46Q2eomMGIuxwFTbgIwlNhR3oGID/GAANhY2GBpxyCktl++BNuS3cNLMRf/wjy18nu10Bb1l6tDp3KdYtW1hbSklR3tglCYiLQ21TAkweD154lD39Tf9zPEz8fPUthZbC5su+Xz/lb40o7HAtkc6U1ufDkSPyhDpaiIvTXqVFC3pTT0bTw+bBx6dW9BkZrJWEo7A1wht8Shq6J1Xj5ZfZ6V1xBUto2PBjI6rqrThWFgZhU8qWwwGnE9iwAVj2FGt4t92bglWrgHHnstgzk7xl7J2TwdTUJE85G6olQlArbLUlojV5v9utXOqL97ChY4kEQ0sIOzFR1cOpATVhqxcr1hqcw29LN5p+/XQJG2A/O/MUubymjHfi11+Ba65hdWXUKJYT/fbbwI4dQLNTR2FLkkzOJSWsLXH1W/Keb1ldPB56krWNBTeyumBJTpDbSzQVNoEfkROIsGk//G8odTgYOrXCLi5WjmRSg7/I4Shst5upstb0sAF/lZWZyZ4UHI4WKWwf0tJgqq3Fddd4sGMHy9ueMQPIPLoNXx8chP8bw2Jd8ngzbrkFeOMVVinXb7ajvokRwW874vDNL6zhffJ6DdZ8zDrapk11ICmJKfcXH2OEffE1qThwADj7HG/sjUqFbWpulgk/VIVNUHvY6jmUSWHzHhA9xhK5GbFEjMaktW0wSySYuqb962WJ8PvWImyTiZ1fSgo7ZgDCBqC4wV1+USMKC9mD1B13sOL8979ZP11+PnDnXFZHth3KxsEdDbjzTuCRR4DnlzQqynxPeToO17DzrLemotrD3h+tike/YaxNXHmeV8KTr82fV0sQSGF36+afZhtNhR1Fwo4AQ0Qepvp6VuG0HlsJLSVs6u1viYdtVGEHskRoqHZJSeCJp4wiLY01otpaWNLSMGYMMGZYOfDiXthmzcSCLAtwP1Ba3IyXXgJ+rnfhSgB3P2jHI7DjeAD3L7ZjHVJxGMCnb9bgTCsrqyG9GnDnVSzM8Z5a4HLgT39OBbK4cyQi4Gbr8ynscAlbS2HzN3Mqd5dLvtb0xBJJwg6ksPXUL5GGkXrm9VZ9edjqehRIYQPykOykJDn9UI+w+bbQ2AiTid2Iyb6nSQm3bgW6vOwGvgMqpHRk1R7Go4+yJpCBBtzAHX71xjT0QjV6AKgxpcOWIAENVRhyQjyGz04AroVcF+Lj5esRbYXN2yGA8SyRUDsdWyGtLyYJ20rj/wMpbP6u3BLCbg0PO5AlQoRdURExhQ1Aafds2AAA6HnRKbhmACPsB+9vxgPXATWfuYA/AE89Z0f3p+zATuCBxXa4Tk8FzgKefagWcRurgfeBf97pAGj22fdrlOdD59io4WG3xBJJSZHLkwi7uhro1Uv+LZW7wyFfa3qM5QmbvsvIgBQOSbSk09Gowna7YW5u1rZQgim4+HjWZvg0Rz5jSEdha2WJ2GzAiBHsD/ubgO+AURPMSP+pHo2H2ebOvQ7geHmbK25IR3xxFfAh0G1IGmsf2wFbCqemqS4kJERWYeul9QHahK2Th62psGOo0zG2CTuaCpsqciQ9bOrUC0TY6obLq85IXGiesAnr17NH5pNPlhtyczNMJiAtgVXcwcPtQCYjt/wT44D/Y4/jcY1KD9sHKj8+DxtQKmyrlVkiLVHY/E1bb+QfNXyHQ74pqKfKpJXdzWbNvHZDCEdhk8oLxRLxePz960DHIKSmsjRYfiCRAUvEUFqfyQSPd79mszezMV6Zk52clwY0es8zPV0mRd7+qOAskUgqbJOJ1TlueTpkZ7PP1OPJtRQ2temWKOzO6mFbacKYSCtsvpK2RGHrETagVA5a8ekpbPXvwoUWYf/yC1uUNC3N+MAZs1kehEP74ht8jY7CVlsitIgu0HLC1htIwq/x2dDAOqy1PGzyecMl7EAK22RiXtGQIf7bqDrjdEGWSEOD9uCdYFkIr7zCxqLrETbfYciTtNPJJkcaO1Z7v96bn0c9IRi9p5sRf56UMQIoCVtLYUeCsAH5+tD+kpLYsOfZs/1/F6olIjod9WFpDYVNJBKKhx1MYQNsbu3jj1duF6zTUet34YIIm85PkpjCHjVKGWuwLBGAlQ1P2OpFZQF/hc1bIklJMNfWyo00nJsjXwe0FgwAlAp70SI2xa6Wh02/pVW4Q0UghQ2wVV7+/Gf/zwcODJyiSvCShMnhCDxjoB5hn3YaGxUTjsL+5Rfgu+/kRTZ4EGHHx7O6T3WG9kc3qfR0uT7w74Mp7EhYIoBcZ/j9nXWWf70zkocdo2l9sWmJlJb6T5+oRihL8mil9f38M7sI6mkyA8FkgmQ2s8f8r76S98dfzO++89/OKGFHQ2EfPsyGd1NvkhHCprINRNg1NawR8ooEkInAagUGD4a1qorNH52aGjytjYeWwuatJj2FfeQIO2enU5uwn39ee03CUGIKNSXwq6+MkRKfJRKOwiYkJcnEWlvL6kR1dcBORx+xb94MTJqk3B8RNpV/fT2LgerD4MGM8NUKmwb60HSngLLTMdoKO9DvjChssg9jyMOOSYVtLS9n/lOgQmqpwv7oI/YIGOKiC5LFwir/eecBL7zAPgxGRIEskbQ0OX82Gh42TXVJNyZqRDRbVDCFXVsrq3W1JcLffLQUNq15+OOPodkhgDZhk3nKxwgoFTbFSHNYqAl72jRg2LDQYiHQvkLNMMnI8O/X0Ns/WSJaHrZRQuAXsqitlZ9SAilsIqfNm/3351WYEpUzqXfa34gR7LV79+AKm7dEWkNha4EGZannhKfvaF+hKuzOOvmTlZ8/Qg8tIWxaIf2CC8IIzsrmt/B45JF/wQibJkQH/BW2xSKTbDQUNlVEqkyhWiIlJfJv1JYIf/PR6nQkwj5wIHzCVttiRGR6CpvI5NgxpYcdicdUmo9Ei0wjAX7gTDiWCIGfuKu2lj3F2WzahE3z7AQibN7Dpv0C8v4mTgQ2bWKd2sE8bN4SaUuFDShVdiTS+oYOZXU+Eu1YBzFJ2Jby8sD+NdCytL6PP2avwVaY0YBkscgqgTpHjTzqU0XQ8szJFonEnZlWMCfC5jsVAX/C5gldyxLh54VQK2z1yjmAUmH37IlmUpahLh+npbABbcLWUtglJdoKuyXo0oU9qQRa97MlIEtEz8MOxRIBZMKm1EgtSyQ11d8SUUNN2LQffkQmqWw9hU3Xn5b9ay0PWwtUt/jUvmAK20jbvPJKYNu26NUPxChhR11hf/opUwM9e+pvowezWVYJNDmREcKmGLXSu4iwI3FnNplkzxLwr4ihKmw6R8Dfw9ayRHiFbTLBRVZMqAqbrpkeYfONUkthl5ezczbaiI3ggQdY1kG0ECxLJFyFrUXYdJ3S0th7Uti7dmkvBmGz6VsifFaUnsLOyGDDKJ1OOQupa1f2u0h10rVEYasH0oWjsFsBsUfYksQ87GCE3ZK0vgMHgBNOCC88q1UmbF5NBgM/EESNSCpsILKEzUNtiahXzgGUhA2gMVzCnjiRZXyQciMQkQVT2BSTycSGJoebGcIjKSl0/zoU8JZIIA87FMKmG6seYZPCrqtjOeOSxCZy4uFVmH6WiNaIzOOOYxlJI0cqCRsAHnyQdWhS277xRpZZY3Sdz2AIVWHzhE3ZMfRkH87AmVZA7BF2bS3MTmdwS6QlCru+Xn+S+iCQLBb/Kc0iRdiRqhRahB2Oh62OVT29arBOR3CEHaolkprKFJm6bANZIpSHrY5pyxb/XNxYhM0GOJ36lojRTi3aVq2w1XnYtLIQWSJnnMG++/VX5f68hKXrYfMKOzsbWLeOjS5UE7bFwp5uN26UP+dHrLYU/CApI79TE3ZqqnzzEQrbII4dY6+hKOxQ0/po9eRwYLEoFxylz4KBVnbXUkekPltDYZtM7C/QwBnewyZYLEqFXVenLEO1JeL93zVgAPs/VIWth0Cdjg6HkpTonLOyopobGzHY7fJ1i4SHTVkiegqbhuuTwj7uOPbb7duV+wvU6Wi365MZEbb6aZif9jSSUC/VpwcqPzVh85wTzsCZVkDsEbb60UQP4Spsl4tdiDAJW7JY/BemNaqw9QbptIbC5iscDeEFWHnQ5PjqtDE+3i5dlIStfkqhRkC/8ZaJkwg72PU0ilAUdow0MsOw2eQO7Uh42OXl7DrrdTrGx7NrXl0t/27IECgXf4S/h813OgZKV1Qr7GgjHIVN56Im7HDS+loBYRP2kiVLcNlll+Giiy7CFrXn1RKQCg42MiwUwqY5JBobZXXQEoVt5DM1bDb9+SRa08MG/AmbCOCaa4Bly2RPkSfsbt3kyi1JTJHxpELvaci6t0ya8vJYR90VV0Tm3Og4Wp2Oegq7vcBmk+t/JBQ2iZ+UFPaZWmETYVO2ExH2jh3K/QXysNXTMPCg+tNahG1UYVO9+PhjNgiqtDS4wm7PhP3zzz9j69atePvtt7Fo0SIsWrQochGNHYv9773nP2GLGqF0OtLvI0DYktaxjFoiwQg7GgpbndYHKAnb4ZAreP/+jLQJ1OBMJqXCdjpZHroWYZNdxJfJ2WdHLndZS2Gbzez/2lrlY26MNDLDUK9TqUa3bqyuBDsvLcIOZIkQYScns1GgRUXKWf4CediBFPaQIcA99wCTJweON1IIVWH/9hs7h61bAyvs9t7puG7dOowfPx4AMGjQIJSUlMChXqIpXFitcNKAi0AIRWED7CLyaV/hErZWj3akLJFIKuzaWkaqwRR2UZH+ZEgULyk0usZaZRiIsCMJLcIGWH1Qdwa3R4VN0FLYf/sbsyuC5fkmJLDfHD3K/g9kidD8IPQ7GrbPq2yal4UGIhklbIsFuPde/UWNI41QFTbd0LZsYQKnoyrs0tJSZHJzYGRmZqKM7tKthVAVNq1HR5UtzCwRTSIycvy0NP05KKhDLpIK27uIQVDCPnRIPx+dCDs9XdngKWdXS2GTso8WYWul9QGMpGjSMEJHI2y73VhaIQ3hD0TYvMLmj0kTOfE+Nr98HW+tBCPs1kaoCpsI+/vv2auWwpakmOp0DIshbKrgJUmCSeOuX6DuvDAIp9MZfFtJwhCTCSZJwq59+9CsuXqojH5xcWgsKkLFtm3oA6CwrAwNYcTX26uwJVqvEMDuffvQxDcGDdgWLoRkt6NJ45i2+noMAHC4uhq1ESiz9IYGdAOwe8MGpBUVoQuAHfv2+TqNBgKoKSvDsYICDDxwALX9+qFYK66SEgwA4IyPh8PpREptLXYXFCBu9270A3C4qkoR72C7Hc3l5bAB2HvgAFwmk7FrGQIyamvRFcDRigpUcfvtb7Gg6fBhJAJoTkyEpaEBVfX1OKpz7EjHFQnkVFeD8if2lZSgsQXxDYyPh2XTJsBsxh6TCZkOBzLq67HTu89e5eUwSRKc9fUg6XWgrAyObt0wxGpF+fffo9Q7w2Pf2lq4U1LgdDrhtttRX1SEowUF6F1WBsliwcE2Lke6lj2cTqQAKNi/P6D4STxyBL0BNBUVwQqg6dtvYQVw0OVCvfdcsquqkAOgYOtWDHa5UFldjZIQzzMadSwsws7JyUE5p2YqKiqQrZGqkx/mrGgFBQXGtvXaHIOGDg0+TWpODuIkCaneOHsfd1xYs7Y5vDcrU69ewL59AICBQ4YwfzEQAh0rPx9YswY9Ro8O+06uKDOvShrYpYtPvQ8ZPlzet92OzNRUZPbtC5SXI+P445GhFZ/3iSC+SxfE5+UBbjc7hldF9xg8WHleKSkwe59g+g8eDPTvb/xaGoU3r7tb797oxu83NRV277EtvXsDBQVIz85Gus6xIx5XJMBZU/2GD/dfKSUUpKayJ45rr8XACROAtWsBhwP5gwczBW61AsnJSOTqbZ/hw9nUwAMHIru0FNlUPlYr4jMyEB8fD1taGtJtNlauJhOQnd3m5ei7lhkZgNmM/GBL7XnFndVbj2n+/V6jRsn12Vsu+QMGAB4PsnJzkRXieYZbxzZSnroGwrJEzjrrLHzzzTcAgG3btqFnz56Ib62eYB50TCNWAk0V2lIPmx71+VWYI/H4P2ZM5B67+BXP1atBA7IlQjP56Vki/MT0/OT1erZScrI8J0Vre9i8JdKjB3uNkcdYwwhmiYSCpCS2v7vvZv+TdUF58nqWCOCf2qe2RIx62K0NGusQDHpZNmoPG2B2SAx1OoYVxbBhwzBkyBBceOGFsFgsePDBByMdlzGEStiHD0eOsPPyZOKLFjmFCyoXyjm3WpUdVRQ3TeykR9g00Cc9nRFiUxP70/Kw1f+3tofNdzoKwmZLnsfFycKCiJVIlu90VB9z8GDgk0+Yf2sytR/CttmMpRDy5UxetXr+fX6CKI+nfRM2APzjH/+IZBzhge6mraiwfUSUlsYewcrKYo+wicwaG5WNjUCEffAg+z/Q8ODevdkfNUyHQyZsdRnyaYttobBpju/2Stjec5JMJpgC5TcbwZ13Kv/nCRvQVth0/VJTWf1wOlm58nUoMVG5ZFxL44wkjCpsvl6ccAIbKp+To+QR9cjdGKlLsTfSMRTEx7M7o5HJY9SEHe5cInQhU1PldLxYJWyXi/3pETYpbCI4Lfz4I7BggdwwGxoCWyL8MaKBHj3YNVenIvLKip4YYkQVGYb3OnkoLS+SoBsdT9g0cAZgbYjKkB85CiizJHiF7XDElsKeMAG4+OLgv+Pbw2mnsVf1VBj0G0pljZG6FBtRhAtaCdsIUlKYv0oea7jKgB8FGKuETY2QLJFAhJ2TE/gxkjqT+Rnx2tISGTiQ5Y6rO3n569leFbY3XikaJKhW2PzQdIC1D7pJ8Nc6I0Pp4RJhS1LsWSJXXcX+gkGtsOPi/AlbZ26ctkb7V9hGCzI1lVWyY8dYJQtzSsd2pbDJElHbB7wlYnROcCOWSGsQNqCdkcPfdNorYXuvk6c1CFttifDXjh/qD2h72G43q0OxRNhGwdeLnBzW4U9rnqp/IxR2BBGKwqbMiaNHWzRM2tfp2B4I24jCHjjQ2D55S6SujjVq9XlrTbfaWuAVdq9ebIBJ376tG0NLQZZINAmb7AyyRIicecLmFTagTdhE/O2dsLOygC+/9P8N1V9B2BEELYdlBBEibIUlQiMUY42wQ7FExo0ztk9eYevNJ95aClsLRDwmE7s2RUWxd12CIZqEzZMwTVnAK2z+ZhuIsGnEpNZc2O0FasIO9Bsqgxh5WutclgjACLsFKVMKS2TYMKbmoriGW1gworArK1knrFFLRO1ha9302pKwKb7ERHY91KmM7QGtZYnQjIC8h62lsJ1OeWg272FLkpxC2R4VNm8RGiXsGFHY7Zuww7FEiotbprB5S2TmTHlR0ViCkbQ+IxkiPNSWSKwq7PZIIITWUthGCdvhkOec4S0RQF7rsz2WN98euDmRFBCWSBSQkBC6JdLcHDkPO1YVnJG0PhrlGGxIPcGIJdIaedh6IJKJ1DSubYHW8LAbGuTMBz1LhO90VE8e1pEIOzVVnz9iVGHHRhTh4rbbgEsuMfZbfq6RSCnsWIURS4Qaot7UqmqEYonQMmStCd4Saa8gSyQavjB//YiwjXQ6qgmbypdm52yPHrbZzP4CTfsaowNn2jdhn3gi+zOCCBG21B4I22SSVy/RS+sjGFXYaktEq7JTo2+Lzj4iHqGwtaFF2ME6HZ1Oea5s3sMGZMJurzdImy0wYceowm7fHnYo4CtkSwibJnE3MgS2LWG36ytsynRJTtZfBUeNULJE2qJydwSFHc2BM7RmJz9SNSkpuIetZ4kUFbFXo/Un1hCMsNUedoxkHMXGbaM1YLOxiuhwtChLxDFihNwRE8uIi5MJW32+VPmM2iGAvyUSyMMWCjs8RDNLBJDrfyQI2zvXdoumgG1LGFXYtIJSSyfjihA6j8IGZBujBY26ZsoU4O23IxRQFGG3B84SAUIjbIuF7ZMskUAedlsQdgdS2FHxsAH/HOqkJDaWIDFROcjISKfj778zso71J009jB8PjB2r/z0pbJqyN0Ys0M6jsAFW6MeOtW8VZhSBLBEiVKP+NYHmxA5miQiFHR6ysgCrFW71vBaRgpbCTk1lKZ7p6fLv7HbWD+Jw+HvYdEN0ucJaACRmsHx54O+pzRBhp6VFNx6D6HyEDbTvRm0UZInopfUBoSlsgDX4igo2cCLWCLsjKOxu3YADB1BP05dGGomJ/oQN+Ocim0ysPJ1OfYUNtG/CDgY1YceIwhaWSEeFEYUdKmEnJ8tzaMeaJdIRFDbAFsaIVkok/4QEBC4rUuOdlbDVlkiMKGxB2B0VRjzsUC2R004DfviBvddS2Far9qRQrYGOoLCjDbUlEqis9Aib34ZWWO+I4BW23W5sJZtWQOck7Bjp8Y0q+CwRvTzsUBX25MlyhoxeGSYnt21aX2e4GYcL6nQ0orDj47UJm1/ooDMo7IqKmLFDAOFhd1xEwxKZOJE1WI9HvwyTk9tmyL4g7ODgFXZcXOAnIfKw1Z2OgJxdEiM2QVRAbcbjianz7JwKuzM06mhYIpmZwOjR7H0ghd0Wlki3bsDTTxufqqAzgjodGxqCtwE9SwRg23ZkdQ0ob1BCYbcROhNhx8WxKVT1CDslJTxraMoU4Pvv9bdNSZEbeWvCZAJmz27947Yn8J2OLSHsv/0N6N8/enHGAvjzjSGFLQi7o4IsEa20vnHjwvfx//IX1oiHD9f+PjlZXlVbILbAp/UFawPx8cy/1SLs+fOjF2OsgFfYgrDbCBdeyCphnz5tHUn0YbczD9Lj8Sfsa65hf+EgMxNYuFD/+wEDwl4vUyDKCEdha3nYnQF8mxGWSBuhWzdgwYK2jqJ1EBcnZwO05tSQTz3FBtYIxB4SE5lirqkxRthaA2c6C2gK1o7S6bh+/Xqcfvrp+PbbbyMZj0CkYLfLq5ur0/qiCZutdY8nYByUSVNWFjxfPZCH3VnALwcYIwiLsA8ePIhly5bhpJNOinQ8ApGC3d42ClsgdsETdks6HTsL6Jzbu8LOycnBs88+i5T2OhduZ4Dd7r8en0DnBqnqigpjnY48YXc2DxuISYUd1lVIaI/LAnU28NNeCsIWAGSFHWjgE/9bobDZawwp7KCEvXz5cixXTUU4e/ZsnHnmmUF3XlBQEFZQTqcz7G2jjViNTR1Xdk0Ncrzvj5SWorqNYm4v5RVLiFZsyeXl6Ol9X+50oiTAMbJqa9EFwLF9+5ALYNf+/XAmJMRkmUWrvAaCEeTB6mrUh7H/aMQVlLCnTZuGadOmhbXz/DBHQxUUFIS9bbQRq7H5xcUNO+/euze6t1HM7aa8YghRi+3wYd/brF69kBXoGL17AwByvU9qg4YNQ0FRUUyWWdTKy/tE0mvYsLBGdoYb18aNG3W/EwmzHRXCEhFQg88MMWKJAMDevWy7zthfRR52DFkiYRH2mjVrMGPGDKxduxZPPPEErrvuukjHJdBS8Kl1Is1OAJBJGDDW6QgAu3cDPXu2zYRebQ0SOu2903Hs2LEYG2g9NIG2B0/SQmELAErCNpKHDQC7dgGjRkUvplhGDHY6Ckuko0JYIgJqhGOJ1NYyhd0ZQZZIDNlBgrA7KoTCFlAjFEuE/21nJWybre2mC9aBIOyOCkHYAmqEo7CBzkvYVmtM2SGAIOyOC2GJCKgRTqcj0HkJ22aLqQ5HoLPN1teZIBS2gBoWizxPulDYwRGD8+YLwu6oEGl9AlpISBCEbRRPPinPxxMjEITdUSEsEQEtJCSwFYGMpvWlpsacLdBqGDKkrSPwg/CwOyqEJSKgBSJqowq7R4/oxiMQEgRhd1QIwhbQQkICG7XIdypqgb7vrHZIjEIQdkeFIGwBLSQmMnUdbKh5XBz7jSDsmIIg7I4K4WELaCEhwVj2g8kETJsGTJ0a/ZgEDEN0OnZUCIUtoAVS2EbwzjvRjUUgZAjC7qgQhC2ghW7d5MWZBdodBGF3VPCWiMjDFiAsWcLysAXaJQRhd1TwJB1Dk9cItDE6a051B4HodOyoIMK22Trn5PMCAh0QgrA7KqxWwGwW/rWAQAeCIOyODLtdELaAQAeCIOyODEHYAgIdCqLTsSMjLk5e5khAQKDdQ7Tmjgy7XRC2gEAHgmjNHRl2u0jpExDoQBCE3ZFBE/gICAh0CAjC7sgQIxwFBDoUwiLs5uZmLFiwAIWFhXC5XJg7dy5GjRoV6dgEWgq7HZCkto5CQEAgQgiLsD/99FPExcXhzTffxJ49ezB37lx88MEHkY5NoKWw2wGPp62jEBAQiBDCIuzJkydj4sSJAICMjAzU19dHNCiBCKFXr5hbRFRAQCB8hEXYds4bffXVVzFVTHIem3jllbaOQEBAIIIwSVJgk3P58uVYvny54rPZs2fjzDPPxBtvvIFvvvkGL7zwgoLEAWDjxo1IDLYysw6cTifig60510aI1dhEXKEhVuMCYjc2EVdoCDeuhoYGnHzyydpfSmHi3Xffla6++mrJ4XBofr9hw4Zwdy1t37497G2jjViNTcQVGmI1LkmK3dhEXKEh3LgCcWdYlsihQ4fw5ptv4o033ojJO5uAgIBAR0RYhL18+XLU1NTgr3/9q++zl19+2c8WERAQEBCIHMIi7Ntuuw233XZbpGMREBAQEAgAMb2qgICAQDuBIGwBAQGBdgJB2AICAgLtBEHzsMPFxo0bo7FbAQEBgQ4PvTzsqBG2gICAgEBkISwRAQEBgXYCQdgCAgIC7QQxt4DBkiVL8PPPP8PlcuG+++7D8OHD2zSeJ554AuvWrYPb7cbMmTPx3XffYdu2bUhPTwcAXH/99Rg7dmyrxrRu3TrcfPPNGDhwIABg0KBBuPHGGzF37lzU1taia9eueOyxx1p9INPy5cvxySef+P7funUrhg0bhoaGBt+8MnfccQeGDRvWajHt2rULN954I6655hpMnz4d5eXlmuW0atUqvPTSS2hsbMT06dNxySWXtGpcx44dw/z58+FyuWA2m/Hoo48iNzcXQ4cOxUknneTb7pVXXoElisu+qeOaN2+eZn1v6/KaM2cOKisrAQBVVVU48cQT8de//hXnn3++r35lZGTg6aefjmpcan4YNWpUdOtXWIPdo4SffvpJuv766yVJkqSdO3dKV1xxRZvGs379eunPf/6zJEmSVFlZKZ155pnSHXfcIa1evbpN4/r555+l2bNnKz6bO3eu9Pnnn0uSJEmLFi2Sli9f3hah+fDLL79Id999tzR9+nRp586dbRJDfX29NH36dGnBggXSa6+9JkmSdjnV1tZKEyZMkGpqaqSGhgZp4sSJUl1dXavGNW/ePF9cr7/+urRo0SJJkiRp1KhRUYvDSFxa9T0WyovH/Pnzpc2bN0uHDh2SLrzwwqjFoYYWP0S7fsWUJbJu3TqMHz8eAFONJSUlcDgcbRbPiBEjsGTJEgBASkoK3G43PDG6IMD69esxbtw4AMD48ePx/ffft2k8zzzzDG688cY2jcFut+Pf//43unTp4vtMq5y2bNmCYcOGISUlBQkJCTjppJOwYcOGVo1rwYIFOPfccwEwZVhXVxe144cSlxZiobwI+/btQ2VlJU444YSoHV8PWvzw888/R7V+xRRhl5aWIjMz0/d/ZmYmysrK2iweq9WKpKQkAMB7772HMWPGwGw24/XXX8f06dNxyy23oKKiok1i27NnD/785z/j8ssvxw8//ID6+nrfRFxtXW6///47cnNzkZubC4DZXFdccQUWLlwIp9PZanFYrVa/ycm0ykld77KysqJaflpxJSUlwWq1orm5GW+++SamTJkCAHC5XLj55ptx2WWX4T//+U/UYtKLC4BffY+F8iL897//xVVXXeX7v6ysDDfeeCMuu+wyhT0XrbjU/OBwOKJav2LKw7bZbIr/JUmCKQZW/f7666/x7rvvYtmyZdi2bRuSk5MxfPhwvPzyy3j66adx7733tmo8ffr0wQ033IApU6agqKgIV111FSQuO7Oty+3dd9/F5MmTAQBXXXUVBgwYgL59++L+++/Ha6+9hpkzZ7ZZbHwdo3KKlXrX3NzsWx/1tNNOAwDMnTsXU6dOhc1mw/Tp0zFy5Egcf/zxrRbTH//4R7/6fuqppyp+01bl5XA48NNPP+Huu+8GAKSnp2P27Nn44x//iIaGBkybNg2nnnqqTzhECzw/rF271vd5NOpXTCnsnJwclJeX+/6vqKhAdnZ2G0YErF27Fs899xxeeuklpKam4vTTT/d1hI4dOxZ79uxp9Zhyc3Nx/vnnw2w2o2fPnsjOzkZDQ4PPPiorKwv6WBtN/PLLLzj99NMBAOeccw769u0LABgzZkyblBePpKQkv3JS17u2Kr/58+ejR48emDNnju+zyy+/HCkpKYiPj8epp56K3bt3t2pMWvU9Vspr06ZNOOmkk2A2MxpLTk7GtGnTYLfbkZ6ejqFDh2L//v1RjUHND9GuXzFF2GeddRa++eYbAMC2bdvQs2fPNp1vu7a2FosWLcLSpUuRkZEBALj55puxY8cOAGw0J2VqtCY+//xzPPPMMwDYTa28vByXXHKJr+xWrVqFMWPGtHpcAFBcXAy73Y64uDhIkoQZM2agtLQUAGtgbVFePM4880y/cjr++OOxc+dO1NbWor6+Hr/99htGjhzZqnF98sknMJvNuPXWW32fFRYW4oYbbkBzczOam5uxefPmVi8/rfoeC+UFAL/99hsGDRrk+/+XX37BggULALDVXnbu3OkTC9GAFj9Eu37FlCUybNgwDBkyBBdeeCEsFgsefPDBNo1nxYoVqK6uVjSiOXPmYMGCBUhISEBSUhIeeuihVo9r7NixWLFiBS677DJIkoR77rkH+fn5+Pvf/45ly5ahb9++PkuitVFSUuJTDyaTCdOnT8df//pXJCYmIjc3t1Wv6datW7F48WIUFRXBarXiyy+/xGOPPYbbb79dUU5WqxVz5szBlVdeCbPZjFmzZkVVKGjFVV5ejri4OMyYMQMA0L9/f9x7770YNGgQpk2bBpvNhrPPPjuqdohWXFr13W63t3l5PfPMMygtLVUQ34gRI/Dxxx9j2rRpMJlMmDlzZlTtEC1+WLRoEebNmxe1+iWGpgsICAi0E8SUJSIgICAgoA9B2AICAgLtBIKwBQQEBNoJBGELCAgItBMIwhYQEBBoJxCELSAgINBOIAhbQEBAoJ1AELaAgIBAO8H/A0BwgwFziQN4AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "results_noisy = model.evaluate(inputs_noisy_evaluation, targets_noisy_evaluation)\n",
    "print(f\"Test MSE {results_noisy[1] :.4f}\")\n",
    "pred_target_noisy_evaluation = model.predict(inputs_noisy_evaluation)\n",
    "plt.plot(pred_target_noisy_evaluation, color=\"b\", label=\"predicted\")\n",
    "plt.plot(targets_noisy_evaluation, color=\"r\", label=\"true\")\n",
    "plt.title(\"Noisy Mackey-Glass time series\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "single-tradition",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "selective-police",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "ann_lab_1",
   "language": "python",
   "name": "ann_lab_1"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}