Res_ODDL_0809.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\framework\\dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorboard\\compat\\tensorflow_stub\\dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "import statsmodels.api as sm      # 时序分解\n",
    "from sklearn.preprocessing import MinMaxScaler # 数据归一化\n",
    "from numpy import array\n",
    "import numpy as np\n",
    "import pandas as pd \n",
    "import matplotlib.pyplot as plt\n",
    "# 支持中文\n",
    "plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签\n",
    "plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号\n",
    "\n",
    "import tensorflow as tf\n",
    "\n",
    "import tensorflow.keras as keras\n",
    "from tensorflow.keras import Input, Model\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, SimpleRNN, LSTM, Bidirectional\n",
    "\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from math import sqrt,exp,log\n",
    "from sklearn import metrics,tree # 计算指标\n",
    "import random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "RATE = 0.7     # 训练集，测试集比例(7:3)\n",
    "N_timestamp=24   # 每24个数据（一天）预测下一个（一小时）\n",
    "X_dim = 10        # 六个维度特征：PM2.5 PM10 CO2....\n",
    "Epoch_WL = 10    #弱学习器迭代次数\n",
    "Epoch_SL = 50    #强学习器迭代次数\n",
    "Batch_Size = 64     # 批训练大小  \n",
    "\n",
    "Loss = 'mean_absolute_error'\n",
    "\n",
    "optimizer = keras.optimizers.Adam(0.001)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From C:\\Users\\Shiina\\AppData\\Local\\Temp\\ipykernel_5380\\3152071552.py:5: The name tf.set_random_seed is deprecated. Please use tf.compat.v1.set_random_seed instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# 设置随机数种子，使结果可复现\n",
    "my_seed = 23\n",
    "np.random.seed(my_seed)\n",
    "random.seed(my_seed)\n",
    "tf.set_random_seed(my_seed)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          No  year  month  day  hour  PM2.5  PM10   SO2   NO2     CO    O3  \\\n",
      "0          1  2013      3    1     0    4.0   4.0   4.0   7.0  300.0  77.0   \n",
      "1          2  2013      3    1     1    8.0   8.0   4.0   7.0  300.0  77.0   \n",
      "2          3  2013      3    1     2    7.0   7.0   5.0  10.0  300.0  73.0   \n",
      "3          4  2013      3    1     3    6.0   6.0  11.0  11.0  300.0  72.0   \n",
      "4          5  2013      3    1     4    3.0   3.0  12.0  12.0  300.0  72.0   \n",
      "...      ...   ...    ...  ...   ...    ...   ...   ...   ...    ...   ...   \n",
      "35059  35060  2017      2   28    19   12.0  29.0   5.0  35.0  400.0  95.0   \n",
      "35060  35061  2017      2   28    20   13.0  37.0   7.0  45.0  500.0  81.0   \n",
      "35061  35062  2017      2   28    21   16.0  37.0  10.0  66.0  700.0  58.0   \n",
      "35062  35063  2017      2   28    22   21.0  44.0  12.0  87.0  700.0  35.0   \n",
      "35063  35064  2017      2   28    23   19.0  31.0  10.0  79.0  600.0  42.0   \n",
      "\n",
      "       TEMP    PRES  DEWP  RAIN   wd  WSPM       station  \n",
      "0      -0.7  1023.0 -18.8   0.0  NNW   4.4  Aotizhongxin  \n",
      "1      -1.1  1023.2 -18.2   0.0    N   4.7  Aotizhongxin  \n",
      "2      -1.1  1023.5 -18.2   0.0  NNW   5.6  Aotizhongxin  \n",
      "3      -1.4  1024.5 -19.4   0.0   NW   3.1  Aotizhongxin  \n",
      "4      -2.0  1025.2 -19.5   0.0    N   2.0  Aotizhongxin  \n",
      "...     ...     ...   ...   ...  ...   ...           ...  \n",
      "35059  12.5  1013.5 -16.2   0.0   NW   2.4  Aotizhongxin  \n",
      "35060  11.6  1013.6 -15.1   0.0  WNW   0.9  Aotizhongxin  \n",
      "35061  10.8  1014.2 -13.3   0.0   NW   1.1  Aotizhongxin  \n",
      "35062  10.5  1014.4 -12.9   0.0  NNW   1.2  Aotizhongxin  \n",
      "35063   8.6  1014.1 -15.9   0.0  NNE   1.3  Aotizhongxin  \n",
      "\n",
      "[35064 rows x 18 columns]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 21600x1440 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD2CAYAAAA6eVf+AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAABTBklEQVR4nO2dd3hUxfrHP7ObTSeQ0HsN0hGIFKUEC3ZRFAUvIlhQQa7KzysqFtSL9V4bVhRBEBUUUSxI8RKIQqRpAqGFEiCEhJDet5z5/bGbsAkpm2Q3u5vM53nyZHd2ztnvnD3nPXPeeecdIaVEoVAoFI0DnbsFKBQKhaL+UEZfoVAoGhHK6CsUCkUjQhl9hUKhaEQoo69QKBSNCB93C6iKFi1ayC5durhbhkKhUHgVu3fvPielbFnRZx5t9Lt06cKuXbvcLUOhUCi8CiHEico+U+4dhUKhaEQoo69QKBSNCGX0FQqFohHh0T79ijCZTCQlJVFUVORuKfWCv78/HTp0wGAwuFuKQqFoAHid0U9KSqJJkyZ06dIFIYS75bgUKSXp6ekkJSXRtWtXd8tRKBQNAK9z7xQVFdG8efMGb/ABhBA0b9680TzVKBQK1+OQ0RdChAkhrhJCtHC1IEdoDAa/hMbUVoVC4XqqNfpCiFDgJ2AosFkI0VIIcVIIEWX762+rt1gIsV0I8Yzdtg6VKRQKhcKKyaLxQdQRYk9luWT/jvT0BwBzpJQLgPXAPcBXUspI299eIcQEQC+lHAF0E0KEO1rmkla5mGnTpvH0008DMH/+fObPn+9eQQqFokGw73Q2N7//B6//eohf41Nc8h3VGn0p5RYpZYwQYjTW3n4hcIMQYoet1+4DRAKrbJtsAEbWoKwMQogZQohdQohdaWlptW2Xy/nkk0+Ur12hUDiFIpOF1389yPj3/yA1p5iPpgxm7jW9XPJdDkXvCKtj+Q4gE/gLuFJKeUYIsQy4DggCTtuqZwCDa1BWBinlImARQERERJXLer3wYzz7k3McaYLD9GkXwvM39q22Xr9+/VixYgUAxcXFTJ48meTkZDp06MCSJUvw9fV1qi6FQtEw2ZmYwdxv4zh2Lp+JQzrwzPV9aBrouhBthwZypZVZQBzQTkp5xvbRLiAcyAMCbGXBtv06WuaVzJo1i48//hiw9vr79evHli1bCA8P57PPPnOzOoVC4enkFZt57od9TPxoO0aLxvJ7h/LGxIEuNfjgQE9fCDEXOCOlXAY0Az4SQhwE9gE3Ay8DaVhdNTHAQOAQkORgWa1xpEfuKtq0aUOvXr2Iiori9ttvZ9iwYQAMHz6cdevWuU2XQqHwfKIOnWXemn0kZxcy/bIuPD7uIoL86mfalCPfsghYJYS4D6uhHw2sAASwVkq5SQgRAkQLIdoB1wLDAelgmdfy2GOPMXjwYOLj42nXrh1XXnklMTEx9O3rvpuRQqHwXDLzjbz0836+23OaHq2C+fbBSxnSObReNVRr9KWUmcBV5YoHlKuTI4SItNV7XUqZDeBombcyaNAgxowZw4gRI4iPj2f06NF07NixNLJHoVAowDq7/pe9KTy/dh9ZBSb+eXkPZl3eAz8ffb1rcdrzhO3msKo2Zd7G0qVLS19HRUW5TYdCofB8UnOKePb7fWzYn0r/9k1Zfu8wercNcZser8u9o1AoFN6AlJJVu07x758PYDRrPHVtL+4d2RUfvXvjV5TRVygUCidzMr2AJ7+LY9vRdIZ1DePVWwfQtUWQu2UByugrFAqF07BokiV/HOe/Gw6j1wkW3NKPyZd0QqfznBxayugrFAqFEzicmssT38bx96ksLu/VigW39KNt04DqN6xnlNFXKBSKOmA0a3wYdZT3NifQxN/AO5Mu5qaB7Tw2Q64y+gqFQlFL/j6Vxdxv4ziUmsv4i9vx3A19aB7s525ZVaKMvkKhUNSQQqOFNzceYvHvx2nVxJ/Fd0dwRe/W7pblEN5t9Nc9CSl7nbvPNv3h2lerrPL888/Tu3dvJk2axPz58+nZsyffffcdZ8+epX///rz//vvk5eVx2223kZ+fT48ePViyZAkAkZGRXHLJJcTFxbF+/XrnalcoFC5n29FzPLl6LyczCrhzWCeevLYXIf7es4a11yY8cydTp07lyy+/BGD9+vWcOHGCfv36sXXrVs6cOUNcXBxnzpxh9uzZbNq0icTERFJTUwGIiYlhxIgRyuArFF5GTpGJp76L485P/kQn4Kv7h/PyLf29yuCDt/f0q+mRu4ru3buTm5tLVFQU/fr14+TJk2zbto2oqCiysrI4ffo0vXv35tNPP2XJkiVkZGRQWFgIWFMyT5gwwS26FQpF7di4P5Vnvt9LWm4xD4zuxqNX9iTAt/5TKDgD7zb6bmTSpEncc889fP755+zevZuhQ4cyffp0fvrpJzp16sTixYu57bbbuP322xkzZkzpdsHBwW5UrVAoasK5vGLmr43np7gz9GrThE+mRjCgQzN3y6oTyr1TS2677TaEEIwcOZL777+fdevWMXr0aD766CM6duzIVVddxSuvvMLll18OwOnTp6vZo0Kh8BSklHz/12muenMLG+JT+b+rerL24ZFeb/BB9fRrRXx8PNOnT+fpp59GCEFQUBCrVpXNITd69Gj27dt3wbYqQZtC4dkkZxUyb81eNh9KY1CnZrx+6wDCWzdxtyynoYx+Lejbty87duxwtwyFQuFENE2yYsdJXv3lAJqE52/sw9QRXdB7UAoFZ6CMvkKhaPQcS8vjydV72ZGYwcgeLXhlQn86hgW6W5ZLUEZfoVA0WswWjU+ij/PWpsP4++h4/bYBTBzSwWNTKDgDZfQVCkWjJD45m7mr49h3Oodr+rbhxfF9aRXi725ZLkcZfYVC0agoMllY+L8EPtpyjNBAXz78x2Cu7d/W3bLqDYdCNoUQYUKIq4QQLVwtqCETFRVFYmJi6ftp06aVea9QKFzLrsQMrns3mvc3H+WWQe3ZNGd0ozL44IDRF0KEAj8BQ4HNQoiWQojFQojtQohn7OrVuqyxUN7oKxSK+iG/2Mz8tfFM/Hg7xSaNZfcM5T8TB9Is0Nfd0uodR9w7A4A5UsoY2w3gckAvpRwhhPhMCBEO9K9tmZQyobbiX9vxGgczDtZ28wrpFdaLuUPnVvr5kCFDaNWqFb6+vqSkpDB+/Hi2b99OVlYWQ4YM4e2332batGl069aNjRs3YrFY+O2335g5cyabN2/m+++/p2/fvqxYsQKAZcuWlakXEOB5iy4oFN7MlsNpPP3dXpKzC7l7RBf+dfVFBPk1Xs92tT19KeUWm8EfjbW3fzVQMhNpAzASiKxDWRmEEDOEELuEELvS0tJq3iIXU1BQwDfffENcXBxffvklCQkJTJo0iejoaLKzs/n1118ByMvLIzo6ml69evHXX3+xZMkSpk2bxttvv11q8Cuqp1AonENWgZH/WxXL3Z/twN+g49sHRzD/pr6N2uCDgwO5whq/dAeQCUigJKdABjAYCKpDWRmklIuARQARERGyKl1V9chdRevWrQkODqZz587o9XpSU1MZNmwYAMOGDePAgQMA3H333QB06tQJo9FY6f4cradQKBznl71neO6HfWQVmHh4bA8evrwH/gbvTJDmbBwayJVWZgFxwKVAiQ8i2LaPvDqUeTVBQUHExMQA1rTJffv2LS0vT0BAAAUFBYA1t0dl9RQKRe04m1PEA8t3MXPFHto09eeHhy/j8asvUgbfDkcGcucKIaba3jYDXuW8W2YgkAjsrkOZV1NcXMzXX3/NyJEjadasGePGjau07q233sqrr77K8OHDOXr0aD2qVCgaNlJKVu08xZVvbiHqUBpPXtuL72deRt92Td0tzeMQJT3OSitYB29XAX7APuApYCvwG3AtMByryye6NmVSyuzKvjsiIkLu2rWrTNmBAwfo3bt3DZvp3TTGNisUjnIqo4CnvtvL70fOMbRrGK9O6E+3lo07hbkQYreUMqKiz6r16UspM4Gryu0w0lb2eonRrkuZQqFQ1BSLJvl8WyJvrD+EXif49839uHNoJ3QNLEGas6nVMLbtRrDKWWUKhUJRExJSc3lidRx/ncxi7EUtWXBLf9o1U+HOjtC4Y5cUCoVXYTRrfLTlKO/97whBfnremXQxNw1s16ATpDkbZfQVCoVXEHsqi7mr4ziYksuNA9sx/8Y+NA/2c7csr0MZfYVC4dEUGi28tekwn0Yfo2UTPz6ZGsFVfVq7W5bXoox+HZk/fz6RkZFERka6W4pC0eDYfjSdJ7+L40R6AZOHduKp63oR4m9wtyyvRhl9hULhceQUmXjll4N8teMknZsH8uX9w7i0u0ry6wy82uinvPwyxQecm3DNr3cv2jz9dJV1MjMzmThxIhaLBSklERERXHPNNeTn59OjRw+WLFnC888/T+/evZk0aRLz58+nV69eTJo0yalaFYqGyKb9qTzz/T7O5hYxY3Q3HruyJwG+akats/D6NAjuYNGiRdxwww1s3rwZg8HAmTNnmD17Nps2bSIxMZHU1FSmTp3Kl19+CcD69esZP368m1UrFJ5Nel4x//zqL+5btotmgQbWzLyMp6/rrQy+k/Hqnn51PXJXcfz4ce644w4AIiIiMBgMfPrppyxZsoSMjAwKCwvp3r07ubm5REVF0a9fP5UyWaGoBCkla2OTmb82nrxiM3Ou6smDY7rj66P6pK5AHdVa0KlTJ+Lj4wH4+++/Wbx4MbfddhtfffVVmQRqkyZN4p577mHq1KmV7UqhaNQkZxVy7+e7eOTrv+ncPIif/zmKf14Rrgy+C1FHthbMmDGD1atXExkZSU5ODldddRWvvPIKl19+OQCnT1uzR992220IIRg58oJlAxSKRo2mSb6IOcG4t7ay/Wg6z97Qh9UPXUrP1k3cLa3B49XuHXfRokULNm/eXKbsueeeK/M+Pj6e6dOn8/TTT6vZggqFHcfP5TN3dRw7jmdwWY/mvHLLADo1D3S3rEaDMvouom/fvuzYscPdMhQKj8Fs0fj09+O8tfEwvj46Xr91ABMjOqhOUT2jjL5CoXA5+5NzmLs6jr2nsxnXpzUv3dyP1iH+7pbVKFFGX6FQuIxis4X3/neED6OO0izQwAf/GMy1/dqo3r0bUUZfoVC4hN0nMnji2ziOpuVz6+AOPHN9b0KDfN0tq9GjjL5CoXAq+cVm3lh/iM+3J9KuaQCf3zOUMT1buluWwoYy+rVAJVlTKCpm6+E0nvpuL8nZhUwd3pl/XdOLYD9lZjwJ9WsoFIo6k1Vg5N8/H+Db3Ul0axnENw+MIKJLmLtlKSrAq41+9KrDnDuV59R9tugYzKjbe1Zbb+PGjTz33HPk5OTwww8/8OSTT5KcnEyHDh1YsmQJL7/8cunTwNKlSwGYNm0akZGRXHLJJcTFxbF+/Xqnalco3MG6vWd49od4MguMzBrbndmXh+NvUPlyPJVqZ+QKIZoKIdYJITYIIdYIIXyFECeFEFG2v/62eouFENuFEM/YbetQmTdy5MgRtm7dyoQJE/j888/p168fW7ZsITw8nM8++6zS7WJiYhgxYoQy+Aqv52xuEQ99sZuHVuyhdYgfax++jH9d3UsZfA/HkZ7+P4A3pZQbhRAfAk8CX0kp55ZUEEJMAPRSyhFCiM+EEOFAf0fKpJQJtRXvSI/cVZTk0+nUqRMzZ85k7dq1AAwfPpx169YRGhpaWrewsLA04Vq/fv2YMGFC/QtWKJyElJJvdyfx0k/7KTJrzL2mF/eP6oqPXmV18Qaq/ZWklB9IKTfa3rYEzMANQogdtl67DxAJrLLV2QCMrEFZGYQQM4QQu4QQu9LS0mrTpnrBPrHaG2+8QUxMDGDtyfft2xdfX19K9P/666+ldYODg+tXqELhRE5lFDD1sx3869s4erUJYd0jo3gosrsy+F6Ewz59IcQIIBTYCCyRUp4RQiwDrgOCgNO2qhnA4BqUlUFKuQhYBBARESFr2B634OPjQ3x8PKNHj6Zjx448/fTTHD58mJkzZ/Lbb7/RvHlzd0tUKOqERZMs257IG+sPIYCXxvflH8M6o9OpSVbehkNGXwgRBiwEbgVSpJTFto92AeFAHlCSMD4Y6xOEo2Vex/z580tfT5s2DYAHHnigTJ1+/fqxdevWC7aNiopyoTKFwvkcOZvLE9/GsedkFpEXtWTBLf1p30ytD+GtODKQ6wt8AzwlpTwBLBdCDBRC6IGbgVhgN+ddNQOBxBqUKRQKD8Rk0Vj4WwLXvfM7x8/l89YdA1ky7RJl8L0cR3r692J1w8wTQswDNgPLAQGslVJuEkKEANFCiHbAtcBwQDpYplAoPIy4pCye+DaOgym53DCgLfNv6kuLYD93y1I4gWqNvpTyQ+DDcsUvlKuTI4SIBK4CXpdSZgM4WqZQKDyDIpOFtzYe5pPoY7Rs4seiu4Ywrm8bd8tSOBGnTc6SUmZyPjKnRmUKhcL9xBxL58nVcSSmFzB5aEeevLY3TQMM7palcDJePSNXoVDUndwiE6+uO8iKP0/SKSyQL+8bxqU9WrhblsJFKKNfC+bPn8/KlStp2bIlBoMBvV5Ps2bNWLVqFZMmTcLf35933nmHSZMmYbFYCAoKYuXKlfj6XphW1mw2061bN7p16wbAwoUL6d+/f303SdFI+d/BVOat2UdqThH3j+rKnKsuIsBXzahtyHhlyKQnMG/ePLZu3cr06dNp3rw5sbGxAKX/V6xYwZw5c9iwYQNt2rQpM0HLnri4OCZPnkxUVBRRUVHK4CvqhfS8Yh75+i/uWbqLEH8D3828jHnX91EGvxHg1T39zUsXcfbEMafus1XnboydNsPh+pmZmaxcuZK+ffuSnp6OwWD1gc6cObO0TlpaGq1atapw+5iYGH766Sc2b95M//79+fjjj/Hx8eqfReHBSClZG5vMCz/uJ7fIxKNXhjMzsge+Pqr/11hQv3QtWbBgAaNHjyYmJobx48czcOBAVq5cycCBA8vU2759O5mZmQwfXnF06iWXXMKmTZvYsWMHJpOJX375pT7kKxohZ7ILue/zXTzy9d90DAvkp9mjePTKnsrgNzK8uktZkx65s5k3bx5TpkwBrLNyBwwYwNKlS5k8eXKpiycjI4PZs2ezevXqSvczYMAA/Pys8c8REREkJNQ6/5xCUSGaJvlq50le+eUgZk3jmet7M/2yruhVCoVGibrFO4nBgwezc+dOBg+2phMyGo1MnDiRV155hc6dO1e63V133UVsbCwWi4Xvv//+gicFhaIuJJ7L585PY5i3Zh8DOjRlw6NjuG9UN2XwGzHK6DuJLl260LNnz1IDv3jxYvbs2cOCBQuIjIxk5cqV7N+/n2eeKbuMwHPPPcddd93FxRdfzIgRI7jyyivdIV/RwDBbNBZtPcrVb28lPjmH127tz4r7htGpeaC7pSncjJDScxNZRkREyF27dpUpO3DgAL1793aTIvfQGNusqD0HzuQwd3UccUnZXNWnNf++uR+tQ/zdLUtRjwghdkspIyr6zKt9+gqF4jzFZgvv/+8IH0QdpVmggffvHMx1/dsghHLlKM6jjL5C0QDYfSKTuavjOHI2jwmD2vPsDX0IDbpwMqBCoYy+QuHF5Beb+c+GQyzdlkjbEH+WTL+EsRdVPCdEoQBl9BUKryU6IY2nvttLUmYhU0d05olrehHspy5pRdWoM0Sh8DKyC0ws+GU/q3Yl0a1FEKseGMHQrmHulqXwEpTRVyi8iF/3pfDsD/vIyDcyM7I7/7wiHH+DypejcBxl9GuBI1k2ly5dSmpqKrfddhvR0dGl2957773s37+f66+//oKYfYWiMs7mFjF/bTy/7E2hT9sQlky7hH7tm7pblsIL8Wqjn/XjUYzJ+U7dp2+7IJrd2L3aeiVpGL744gt+/vln9uzZA1izbA4bNozMzEzuvvtu8vPP6/vuu++wWCxs376de+65h4SEBMLDw52qX9GwkFKyes9pXvppP4UmC/+6+iJmjO6GQa/mVSpqhzpz6khJlk1fX98yWTb1ej0rV64kJCSktG5UVBS33347AOPGjeP33393i2aFd3Aqo4Cpn+3g8W9iCW8VzC//HMWssT2UwVfUCa/u6TvSI3cVCxYsYNGiRXTs2JHx48eXLpRSkjvH3tiXkJ+fT/v27QEICwsrfTpQKOzRNMmy7Ym8vv4QAnhxfF+mDOuMTuXLUTiBarsMQoimQoh1QogNQog1QghfIcRiIcR2IcQzdvVqXeaNlCyismLFCpo2bcrgwYNZunRpacK1iggODqawsBCAvLw8NE2rL7kKL+HI2Vwmfryd+T/u55IuYax/bDRTR3RRBl/hNBx5TvwH8KaUchyQAkwC9FLKEUA3IUS4EGJCbctc06z6p3yWzYoYMmRIqUsnNjaWLl261JM6hadjsmi8v/kI173zO0fT8njz9oEsnX4JHUJVgjSFc6nWvSOl/MDubUtgCvC27f0GYCQwCFhVy7IyCeSFEDOAGQCdOnWqSVvcSvksmxVx8803M2rUKJKTk1m3bh0xMTH1qFDhqexNyuaJ1XEcOJPD9QPaMv/GvrRs4uduWYoGisM+fSHECCAUSARO24ozgMFAUB3KyiClXAQsAmuWTYdbUo/Mnz+/zPulS5cCcOjQoTLvwTp4W0JISAhRUVFs3LiRJ554gqZNVchdY6bIZOHtTQl8En2M5kG+fHzXEK7u28bdshQNHIeMvhAiDFgI3ArMAQJsHwVjdRHl1aGsUREaGloawaNovPx5LJ0nv9vL8XP53BHRkaev703TAIO7ZSkaAdUafSGEL/AN8JSU8oQQYjdWt0wMMBA4BCTVoUyhaDTkFpl47deDfBFzko5hAay4bxiX9WjhblmKRoQjPf17sbph5gkh5gFLgLuEEO2Aa4HhgASia1mmUDQKNh88y9Nr9pKaU8R9I7syZ1xPAn29Ompa4YU4MpD7IfChfZkQYi1wFfC6lDLbVhZZ2zKFoiGTkW/kxR/j+f7vZMJbBfPBQ5cyqFOou2UpGim16mZIKTM5H4VT5zJvIz8/nylTppCRkUGnTp344IMPmDp1KmlpaXTv3p3FixeTn5/PpEmTsFgspRO3fH3VohaNCSklP8WdYf7aeHKKTDxyRTgzx3bHz0clSFO4j0Y3kOoMli9fzogRI9iyZQt+fn68/fbbhIeH8/vvv1NcXMyqVatYsWIFc+bMYcOGDbRp04Zff/3V3bIV9UhKdhH3L9vN7K/+okNoAD/OHsljV/VUBl/hdrzaobhu3TpSUlKcus82bdpw7bXXVlmnffv2fP7559xyyy18+umn3Hzzzdx3330AjBw5kp07d/LWW2+V1k9LS6NVK7WaUWNASsnXO0/x8s8HMGkaz1zfm+mXdUWvZtQqPASvNvru4sYbb6SwsJAJEyYwduxYsrOzCQoKAiAwMJCcnJzSutu3byczM5Phw9WYdUMn8Vw+T323l+3H0hnRrTmv3tqfzs2D3C1LoSiDVxv96nrkriIhIYFrrrmGW2+9lSlTphAVFcWcOXMAq7+/JNlaRkYGs2fPZvXq1W7RqagfLJrks9+P89+NhzDodLwyoT+TLumIEKp3r/A8lE+/Fnz66aesWbMGvV5Pv379mD9/funM2+joaIYOHYrRaGTixIm88sorVaZmUHg3B1NymPDBHyz45QAje7Rk45wxTB7aSRl8hceijH4teOSRR1i6dCmRkZHs2LGDxx57jGPHjnHppZcSEBDAxIkTWbx4MXv27GHBggVERkaycuVKd8tWOJFis4U3Nx7mhnd/JymzkIWTB/HJ1CG0aervbmkKRZUIKT0yvQ1gzb2za9euMmUHDhygd+/eblLkHhpjmz2ZPSczmfttHAln87hlUHuevaEPYUEqHFfhOQghdkspIyr6zKt9+gpFfVJgNPPfDYf57I/jtAnxZ8m0SxjbS0VlKbwLZfQVCgf448g5nvwujlMZhdw1vDNPXHMRTfxVgjSF96GMvkJRBdmFJl7++QArd52ia4sgVs4YzrBuzd0tS6GoNcroKxSVsD4+hWe/30d6vpGHIrvzyBXh+BvUjFqFd6OMvkJRjrTcYuavjefnvWfo0zaEz6ZdQr/2asEbRcNAGf1aMG3aNNq1a8fLL7/M/PnzycvL4+jRoyrhmpcjpeS7Pad58af9FBot/Ovqi5gxuhsGvYpsVjQc1NlcSz755BOKiooAWLhwoUq45uUkZRZw95Kd/N83sfRoFcwvj4xi1tgeyuArGhxe3dM/fPglcvMOOHWfTYJ707Pns9XW69evHytWrACs6SBGjx4NqIRr3oamSb748wSvrTuIBF64qS93De+MTiVIUzRQVDemlsyaNYuPP/4YgNzcXJVwzQs5mpbH7R9v57kf4hnSJYwNj43m7ku7KIOvaNB4dU/fkR65q2jTpg29evUiKiqK0NBQ8vLyAJVwzRswWTQWbT3GO78lEGDQ89+JA5kwuL3Kl6NoFHi10Xc3jz32GIMHD8bPz49u3bpx4403Eh0dzS233KISrnko+05n88S3cew/k8N1/dvwwk39aNnEz92yFIp6Q7l36sCgQYMYM2YMDz/8sEq45uEUmSy89utBxr//B2l5xXw0ZQgf/GOIMviKRodDCdeEEK2Bb6WUo4QQPsAx2x/AbCnlXiHEYqAP8LOU8t+27RwqqwyVcM1KY2yzM9lxPIMnV8dx7Fw+t0d0YN51fWgaqFIoKBoudUq4JoQIBT4HSpYAGgB8JaWca1dnAqCXUo4QQnwmhAgH+jtSJqVMqGP7FIoKySs289q6gyyPOUGH0AC+uHcYI8NbuFuWQuFWHPHpW4A7gB9s74cDNwghxgJ7gQeASGCV7fMNwEhgkINlZYy+EGIGMAOgU6dONW2PQgHA5kNnmffdXs7kFHHPZV15/OqeBPqqISyFolqfvpQyR0qZbVe0E7hSSjkUMADXYX0KOG37PANoXYOy8t+3SEoZIaWMaNmyZc1bpGjUZOYbmbPyb6Yv2UmQnw+rH7qU527sowy+QmGjNldCnJSy2PZ6FxAO5AEBtrJgrDcTR8sUijojpeTnvWd4/od4sgtN/POKcGaN7Y6fj0qQplDYUxuju1wIMVAIoQduBmKB3VhdNQADgcQalCkUdSI1p4gZy3fz8Jd/0T40gB9nj2TOVT2VwVcoKqA2Pf0XgS8BAayVUm4SQoQA0UKIdsC1WP3+0sEyr+TZZ5/lt99+o3Xr1ixfvpy77rqrTMI1Hx/lTnA1UkpW7jzFgl8OYDRrPH1dL+65rCs+Kl+OQlEpDl8dUspI2/99UsoBUsr+Usp5trIcrIO5McBYKWW2o2VObEu9sW3bNqKjo/njjz8YN24c77zzzgUJ1xSu5UR6Pv/49E+e/G4vfdqGsP7R0cwY3V0ZfIWiGpzWHZVSZnI+MqdGZbXl2YQk9uUVOmNXpfQLDuCl8A5V1lm/fj3XXXcdQgiuvvpqwsPD+eEHa3BTScK1O++806m6FFYsmmTJH8f5z4ZDGHQ6Xr6lP5Mu6ajy5SgUDqJ8ELUgNTWViAjrvIdu3boRGRlZacI1hfM4lJLLE6vjiD2VxRW9WvHvW/rRtmlA9RsqFIpSvNroV9cjdxUhISGlCdZ27NjB//73Px599FGgbMI1hXMwmjXe33yED6KO0MTfwLuTB3HjgLYqQZpCUQuUA7QWXHbZZWzcuBGALVu28NJLLxEVFQVAdHQ0Q4cOdaO6hsXfp7K4YWE07/yWwPX927JpzhhuGthOGXyFopZ4dU/fXdx0001s2rSJSy+9lBYtWrB8+XKmTZvGpZdeSnh4OBMnTnS3RK+nwGjmzQ2H+eyP47QO8eezaRFc3uuCuXwKhaKGKKNfC4QQLFy4sEzZmjVr3KSm4bHtyDme/G4vJzMK+MewTjx5bS+a+KsEaQqFM1BGX+ExZBeaeOWXA3y98xRdWwTx9YzhDO/W3N2yFIoGhTL6Co9gQ3wKz3y/j3N5xTwwphuPXdkTf4OaUatQOBtl9BVu5VxeMfPXxvNT3Bl6tWnCp3dHMKBDM3fLUigaLMroK9yClJI1f53mxZ/2U1Bs4fFxPXlgTHcMakatQuFSlNF3AhkZGezevZtBgwbRooVapKM6TmcVMm/NXqIOpTG4UzNev20APVo1cbcshaJRoLpVNSQvL49bbrmFkSNHcvfdd5OWlsYNN9zAjh07GDt2LGlpae6W6LFommT59kTGvbmFHcczmH9jH7558FJl8BWKekT19GvIwoULCQ8PZ82aNUyaNIkPPviAN998k+HDh5OZmcmePXu4+uqr3S3T4zialseTq+PYmZjJqPAWvHxLfzqGBbpblkLR6PBqo//Cj/HsT3Zunps+7UJ4/sa+lX7+559/ct999wHW5GpHjx5l+PDhbN26lR07dvDcc885VY+3Y7JofBJ9jLc3JRBg0POfiQO5dXB7NaNWoXATXm303UFubu4FydWklKxcuZLQ0FAMBjWJqIR9p7OZuzqO+OQcru3XhhfG96VVE393y1IoGjVebfSr6pG7CvtkayXJ1YQQvP/++zz77LOsXbuWO+64o951eRJFJgvv/pbAx1uPERroy4f/GMy1/du6W5ZCoUAN5NaYYcOGlUmu1qxZM5YtWwZAVlYWzZo1c584D2BnYgbXvRvNB1FHmTCoPb/NGaMMvkLhQSijX0Mefvhhjh07xqWXXkpAQAAzZ85k+fLljB49GovFwrhx49wt0S3kFZt57od9TPxoO0azxrJ7hvLGxIE0DVTuLoXCk/Bq9447CA4OviC5Wkma5cZK1KGzzFuzj+TsQqZf1oXHx11EkJ86tRQKT0RdmYpak5lv5KWf9/PdntP0aBXMtw9eypDOoe6WpVAoqsAh944QorUQItru/WIhxHYhxDPOKFN4F1JKfo47w1VvbWHt38n88/Ie/PzPkcrgKxReQLVGXwgRCnwOBNneTwD0UsoRQDchRHhdylzVMIVrSM0p4oHlu5n15R7aNg1g7cMjmTPuIvx8VEbMhsbPaVnsycm/oPxIQREZJrMbFHkX2SYzh/KL3C3jAhxx71iAO4AfbO8jgVW21xuAkcCgOpQl2H+ZEGIGMAOgU6dONWmLwoVIKVm16xT//vkARrPGU9f24t6RXfFRCdIaLE8eTuLaFk0ZHBJUWtZm898AtPb1Ifayfm5S5h3csCeBhIJiUsZezMITqfyemcfKi7u7W1b1Rl9KmQPYz6AMAk7bXmcAg+tYVv77FgGLACIiImRNGqNwDSfTC3hqTRx/HElnaNcwXrt1AF1bBFW/ocJr0aQkzWgm36JV+HmqUfX0qyOhoLj09akiI/vyCt2o5jy1GcjNAwJsr4OxuojqUuZ1vPfee3z77bf8+eefDBs2jMzMTIxGI61bW9dwffvtt3n00UcZOnQor7/+OsOHD+eaa64hMTGR2NhYgoODadasGV999RXBwcFubk3lWDTJ0m2J/Gf9IfQ6wYJb+jH5kk7odCqFQkOn0GbsV6dm8n6fzhXWefdEKv/srNYtdhRPuWpqY3R3Y3XLAAwEEutY5nU8/PDDREVF0b59e6KiorjllluYN28eUVFRREVFcfHFFwMQGxuLpmns27evdNuFCxcSHR3NsGHDWLFihZtaUD2HU3O59cNtvPTTfkZ0b87GOaP5x7DOyuArSvn89Dl3S/AaJOApl05tevrfA9FCiHbAtcBwrG2qbVntWfckpOyt0y4uoE1/uPZVp+zKbDZz5MiRCscmMjMzPXLMwmjW+DDqKO9tTqCJv4F3Jl3MTQPbqQRpjYzozLwLyjRZ1tuqTgnHWZ2aSUElrrL6xuGevpQy0vY/B+tgbgwwVkqZXZcyp7XEzSxYsIDIyEgiIyOxWCwAtGvXjk2bNjFkyJDSerNnz2bUqFEUFBQwefLkOn+vWZMUOelkij2VxY0Lf+etTYe5rn9bNj42mvEXq4yYruac0UybzX+TVGR0t5RSUo2mC8raRcWWeS88xmHh+ZQYfE+IeqrV5CwpZSbno3DqXFZrnNQjdwbz5s1jypQpZcoGDRrE0qVLmTx5MtnZ1vvbwoULGTlyZEW7qBV3xh1la2YeKWMvrvU+Co0W3tx4iMW/H6dVE38+nRrBlX2Ur7a+uGLnQQAitu+v0+/oTByJoFAmv+bkmi2EGdw7J9YrB1K9hcGDBxMbG0u/fq4LbdtawWN4Tdh29BxXv72VT6KPM2loJzbMGa0Mfj1j9tIYNWX0vROVhsFJLFiwgE8//RSAhx56CLAa/QEDBnhkjv2cIhOv/HKQr3acpEvzQL66fzgjujd3t6xGiXSoX12/ONTTV1bfK1FGvw4cOXIEgPnz5zN//vwyn5Xk1N+5cycAkZGR9SmtSjbtT2Xe93tJyy3mgdHdePTKngT4qhm17sJbfePeqVqhjH4D51SRkTWpmczu1Ir0fCMv/LifH2OT6dWmCZ9MjWBAh2bulqjwQKSsvq/vrTerxo4y+g2cu+OOsT+/iMC0Yt7/5SD5xRb+76qePDCmO74+akinPGeKjfyVU8B1LZvV23d6onvHEZTJdwxPCdUsQRn9Bk6OyRo++uLP+xnSsgmv3zqA8NZN3KzK8zhRWIxFwqTYo5wsMpIcORCdclornEBK8fnw1z+y8ugc4OdGNcroN1g0TbJix0mSswshwIfZl/dgzmXd0XvKtEAPY1jMAbd9tye6SU7YzRk4WVhMswrCDNU90TGu2X2o9HW+2f29fvV834AosmjMOXiSv85kM2lRDM9+v6805fFtQzpWaPA1KZl/5DQnC4sv+Kwxo3mnx8VpfHwqrfT10JgDfH0m/YI6RwrUOeMIOXaGXvMAV54y+rVg/vz59O7dm9GjR3PFFVcwbtw4Bg0axIgRI5g4cSImk4lp06YxaNCg0lm6KSkpxMfHM3bsWIYNG8Z7773ndF1fJqfz5ZkMrj14nIMpObx+2wBaN6n6UXJfXiEfnUrjgfgTTtfjzcTmFrhbQr1j0iTvnzxLsXZhb/S5I8luUOR9fJGczpniymdWe0JnQhn9WjJv3jy2bt3K9OnTad68OQsXLmT79u0EBwezadMmwDr7tiQJW5s2bXjppZd47bXX2LZtG8uWLStN1+AM4pOz+Wjr0dL3m+aM4faIjqWjbZWdayXlZgeiNRQNm+XJ53jpaDIfnDzrbileSZrRxOOHTjEl7lildTzhKvNqn/5rO17jYMZBp+6zV1gv5g6d63D9zMxMVq5cyaxZs5BSkpeXh6+vb4V127dvzyeffEJ4eDg7duxwlmQAbnrvD3y6h5S+bxXi79B2jc0tm2Uyc/nOQ3zWrysXhwRWWq+xHRegNHd+nguiTVKLTYzbdYhvLu5BzyDHzk1vo6TjlG6svDNXPmmdO1A9/VqyYMECRo8eTUxMDOPHj2f27Nl06dKF1q1bc/nllwPW5GqRkZFMnDgRgFdffZWuXbsSERHBkiVL6qxhV2JG6eubL27PtNFdK61bnRHbm1fI6pSMamp5PzFZ+SQXm3jrRIq7pTRocsyWMm6in9OySDWaWZyUVsVWDRO93cXn7wErzXl1T78mPXJnY59gbdq0aSxcuJDff/8dPz+/0qyU5ZOrxcXF8fTTT/Pggw8yatQoxowZQ7du3Wr83fnFZt5Yf4jPtyfCuPYA/Pf2gbyTmFrpNo70L2YdOMmtbcJqrMeTkVLy7omzTGobRmu/GqTDaIxdfSfSM3ovg5oEsi6iJwC/pFkTDv6ZfeGauw2dYL2ebLO19+/nAdFz7r/tNCAeeOABFi9eXKmvftasWZw4cYKwsDDat2+PyXRh+trq2HI4jYhVO/ioiYkpIzpzb9vmBNl6DxWdT+4/xdxLfF4hrxw/wwPxiQAUS2vvs8hS9W2wPsMovXVyVnX8ZTcY/keWNTHgQQ9cKLyx4dU9fU8jNDSUyy+/nNWrVwNW907Tpk0BeOGFF3jttdeYOHEiBoOBMWPGcNFFFzm876wCIy/9dIDVe5IwXtkOgLnX9eaNxBSHTEZFhuWvnAIKnDiY7ImUOBhK/NUG21NYdR2uij4+kFdIc4MPrWryxOCF1PZ2dzC/kFAfnwqfqErOvsbSCTlbXLMOnZSSP7LyuKxZsMvXr1BGvxaUT662dOnS0tfvvvsuALfffnuF29Z0ANekaRQYLVz55layCow8PLYH7xuKKNQkgrLx1OXDwX7PzOV4YcXhYxYpuXb34Sq/+7/HU/DTCR724nVQt2bkAtb2wnnjk24yY9Q0fHUVP+xWdN2N3XmIAJ2O42MGuEKq26nr80bkjkPoBZyOvPiCzy5tFsy2rDxGhzbM2eALT6Ryym5C27asilOe24/j7s7Op2eQP0189Cw5fY6nE07zSd8u3NiqmUu1Nmj3jialQ4mjPBWTRWN/XhEZZjNtmvrxw8OX8fjVdk8H5QzTqXIrL92993jp67dPpPJ3zvnHbUcOyxuJKfz72JlaafcUSvSbbA0uuTHG5Rby+KFTlW5X4nv9MjmdLbYbB0BhBTHsdaW+Z+SmFJt4/sjp0hvhhXpqT2Ves2FNgwC4xPa/obHg2BmWJZ+fwGbfabB/yi55ZdIk1+9JIDx6L88nnGa37do8XQ+rpzVYo69Jyd7cQs7U8DHLE5BSkpFv5HCq1dhIneD7mZfRt53VVVRos1zljcUXFcyaLOGblEyuqaBnH9JIkq5VZIw2pedUWt9PWI/LnEOnuCP2aIUTlpxFffv0Hzt4ko9PpbG9kt6oKyhpob6x+HfssJ+Ru9S2mLzJ7ob7cVIaq1Mz601Pg73iS3oxnrAmZU0wmi0cP5dPUmYB/j7nc9z7VBDq5YzrZ1QDfdwuz7HCYu7fl+hw/fJmOMfccMY+jJVMC115xhqym++COP2S+HRdA/PqP3rgJIO3xZcpSzGaKp3hfsA2kH2kwH0D2jU2+kIIHyHESSFElO2vvxBisRBiuxDiGbt6DpW5ipLTupogDY9BSsm53GIOp+ZRaLTQvlkA3VpW/Sj88anzMydr6sZqbANrAD+mZTlct3zvuzJD2ZA4asu/VNGi6HWl5JbpARGLTuXrlAySa+FN+F8lT5n1cZbVpqc/APhKShkppYwEwgG9lHIE0E0IES6EmOBImbMaURHedIkWmSwcTcsnObuQID8fwls3oXmwX7Wj+K8er3yCUXVPOCVGzd5f3RiwPy+q8qWXP39cafQL6znfurNbcqrIyHS78aOKKO3pq9ScgHvtU22id4YDNwghxgJ7gWJgle2zDcBIYJCDZQm1k109rhy/nTZtGn///Tc+Pj488MAD/PHHH8TGxpaGZ5bk2Jk+fTpffvklt956KwkJCWzevJni4mJ0Oh3du3dn8WefkVFg5mxuMXoBHcMCaRZgqFXIVvnm9vl9n0Pb5XrYAg+ewub0XE7aRT4VudCnX1jPTxElBji7EpdVlqlmrqxLtu+vtk5JfH6D9SfXkMp+8T05rk/0VxujvxO4Ukp5RgixDLgc+Nj2WQYwGAgCTjtQdgFCiBnADIBOnTrVQp4VV19G7733Hr1792bgwIEMHTq0zOxbo9FISkoKR44cITExkeTkZDp37gzAN998Q4cOHZh2zz0s+/ZHLhk5lmYBvrRt5o+hDlO0a9rew410kox9BsSSW2tWBU9Fzx45Xea9K3vjESGB7KqHi72EGNus2OcSTnN9BSuENfWp23rJFbkaN9ueKBuae6e2VNYprYkLsrbUxujHSSlLEmnvAm4BAmzvg7HezPMcLLsAKeUiYBFARERElbYs5eWXKT5QccI1CxIfm0P/RA2MqV/vXrR5+mmH6jZv3pzrr7+e1atXM2fOnNJyX19fNE3jyJEjtG3blqSkJIYMGQJYFzdJzirgdGo6vn7+dGkeREiAY5N90owmzhmdMzB95a6qY/S9jYjt8czq1Jrp7VtUWW9/3oU3O0d8stftcdlDKU193DNdprLw0715hXXab1W3x4YwkJtnttAjei/L+lee66o63OneqU3XcrkQYqAQQg/cDMzC6qoBGAgkArsdLHMZ9fXE3Lx5c7Kysi5IrhYYGMi5c+cIDw9nz549dOnSBYsmuWnCrVw/7nL69O7NneOvdtjgA4zZcZCxOw9V+Jn3X0p1I6nIxFOHk6qtt9Iuqdw5Ww+/sRy7PLOFdLtOQ2UGuPx8j5qy8VzlobANwaW/3BaPP7WacYyqcGfqjdp0MV4EvsR6rawFvgeihRDtgGux+vylg2V1oqoe+f68Qsw2y9+pifUBwxXTmzMyMggLC7sguVrHjh1JSkriiiuu4IMPPoCgMIpMFt5e9DkRfXoQ7F/zqfwZVfha1QCZ9+LMn86kSf59LJlHOrcmrNwShz2i95Z574yY+YpSBU/bV7UxlFJilmDwUl/PT05wwVQ2U74+qHFPX0q5T0o5QErZX0o5T0qZA0QCMcBYKWW2o2XOakRF2Pf09+YVuiTRU1ZWFuvWrStNpWxPly5d6Ny5M207dCasRSvyjBKDXke3FsG1Mvj1xadJafUeTeIqClzcjjWpmR63zOQv57L4+FQazyWcrrauMzoKD+2v2YprFil57XgKHbfEeu155oz1Br6rx8lY5XGKM1FKmcn5yJwalbmKMlOfJRidHM4ze/Zs/Pz8eO211/j1118vSK7WoWNHCs0CQ1hb2nfsSPdWwfj66NB5eO/mmYTTpBSbeKZ7O3dLqTMVDdBWRG2N30P7T9DC4MO+kf1qtX0JzpylWvJ0W1mKBWd/7w9ns2pU3yJhWbJ1VmqexUKAB+SXrymesBBKXWiwCddc2YewT7AGMGnSpNLXUkqyC020DL+Y0RJaNfFjx/Y/0AlBVFRUme2klBRp0qETf38dB9dKuDP2KF8O7F5lnfdOnuWaFk2d8n31gf2chJv3JPD9YOsUEEfPgbrYvnPlbizT9h7jiuYh3NWu6gFle4aEBLG+Ej94mtGEJnF4LYDSSXcO3Mj0tpb/93hK6cQsV6NJWaWb0htw9XjhqD8PED2st8v232CNvqvHSaSUmKTEIETpBWY0ayRnFZJTZCLQ14cOoQH4GyoPf0s1mkktNtEzyJ8AvQ6LlOg4f8HaP/5eXskAbk35X0YuzyRUP+D5gt1C2CZNeoz/1SIlBRaNJnZhhfYDhzF2i3S4skdW2QzoX8/l8Ou5HFoaDCQXG7mnQ8tq91XVke3/h3WKf8rYix3TZfvvSP+55Cd9I7H+VhErtrOY3jrJ2VJH43KoGldzUpFr84V537NVHcg1W9iXW1BtHpV8s6Va10Cq0cyBvCJOFBmRUpKeV0xCai55xWbaNg2ge8sg/A16pJSkFpswa5Jcs4Vcu+/Ot+WyTy627mNfbiFJduGDJhcZrU+TztWo/tS9lS/0XB6jpvHf4ykUuchf+8ShU4RH7y3jvigfCfHf4ykUWjSX3vfL7/t0kZH+f5yfEDdt33GedsCvXtG+6kJNDGl9Z/eEsjcYb7L5MVl5/Gpb/auuqV3G7HDuut41peH29CugSNOwSDheUEwHf1+a+1bc/CMF1kfdZnbRD6cKi8kwWRhoW0w703ZTyDZZOJaXT36xmWA/H9qHBuBn1wvNMVtIKTZRrGlk2h5rB5ZbkDvPrJXGTGcYzXT0r3hh9frE3juwuQapGj5NOscbiSn4CMEjXZyfh39VinUAzCIleiFYl5ZVelxLKFlYJrSKp6y60i4qtvR1m81/u+x7akrJDdCRYQp3P7x504phN/91BLA+cXm7T79R9fTt3SVJDsQiHysoKp25WuKHLLJopYNlJRSZLHQIDaRri6AyBh/O92bsN8k1W9ibW1Bpj0FKSXxeYemNxRE8JYokz/b0Ut1Tyu7sfLpsiSWtmuRexwqKSbF7+imxU/kWjTab/2b6vkReOJp8wXb/SUzhi+TKU03XBU8y8uU5796xHqnFSWmM+vNAhXXrw+afM5ornf1d1SlytthEly2xZdaAqG/+yimgU1TsBatgnfbCdO32NCqjX75HWB25Zo1Ci1ZmoPBQfhGH8osw2d1AerZuQliQL0IIcswWYnMKyCvvQrK7wlKLrYNz9jchs90FUKxJzJqsUQ4UZw+O7XBgAWsp5QU3z5IbWYbJXGUO+un7jlOkSWKyqv6eS/88wMV2qWtLerAJDoTgni52LBbalcYvuchYxhWlScmM+ER2V3F8zZosky6iIp5POM3as1m8We7mVvJVJW2al3CahIJi9lTwffXh3rl42z5GV+LOiLcFJ+zNLWDcrkOlnak0o4kB2+Ip0iSfJKVVuG198PGpsxhtyxg2JBqV0a+KqlITnyo3keKpB+/n9lGXMvXKsTw+dQpoFv7++2+6du1aauzzLRqJiYncfM3Vpdvde/01fLvkMwCefWgGP6xYXvqZ/UzJ6gZ6KsLowoRglbEqJZOI7fv50+6iMNku3M9On+PB+BPszS3g3n3HL8gfftbW3rwartFbYqbMDjxh+zgYilkSQugKBm/fz6t2q4+lGk2sPZvFPVVMYHrpaDKDtu2vMuXGx0lpzIhP5PXjKTx+6BTfp2byW3qOXfRO2XO6ojQSPvXQ1a/qd7pr73E0Kblq12Hicgv5yJYq/KMyS4DW3ZWSZ7ZwsrC4xvsqyaxq/6Ri8tbRZzsatdHPMVsotGgUWzTicmvmTnnyjf+ybNNmAoOD2LRpE+vXrycpKYkDh6w5bS64nmznylcff+CSRSpcNehrzxfJ6WUGuHdkW439mrNZSClZkZzOB3Y5/tedy+as0czPadmkGc0sTkq7IH48oJI1aqHiWPOiGsShO/r080ema3ty9mMijvSuS1b0cvRJBeDB/Sf4R9yxMtE7y6pxb+mFKF04xV2U+Mrh/AIj9jjjrF6WnM7QmAP0/n0f22rwW5fM7XnrRKpdmes7V4Waxuenz7lsqVevHsiNXnWYc6cq/hEd6UHG2f7763QUaRr7haB1xyY0v66jQ98vpaQgLx9fX1/W/bqe2++fwYb1v3LngzOtGsyW0iiWkjCv0BYt+TNq8wX7quvKTI70fOtCQn4Rjx86xeOHTnEmciBCiNI4+KWnz2EQ8EkFUUG/2QzYuydS2ZyRS6Bex+S2zekS4EtioZHugX4cyi8iOjOXO9qElQnFtM+T82B8Ih/17VL6fmLsUae1zdVR4wL48WwWGpJhTYOrrGs/XvDM4fPRP2N3HGTz0F7Vfpf94OjcanIR7c0r5JGDJ6vdpyuxdyMKrOfZhyfPdxyqMrH/S8/hstBg/KroOACklwRdmC1M+PsIn/btwn3xiRwe1Z+QKjKKVuSedCSNtDOYeziJzgG+RIaFOH3fXm30K6WGd2NNswACs5Rkmy00d2CbV//1f+RkZTL6musY2q87Z9LPMXvqNN5+/lmr0TcVII3nT+g82zqZU2Y+zFcff0iTZs1qpLE6TEd+A7o4dZ/2jLLzy74Ts4FZfmlYCtoDoUDFBh+sbh6AjOxUIJCCE39CSgaJhdZZrFfZZfu8JvV/NNGdHyTLKGoBtAHg+7NZvFj0NVC94asRe5ZjKQwH/J27XztEQQb3x1v918P0+UAQ0lRI+q4vMSM4U9wKKjjr0vIyAD/A1gvesxyjFMRYAoFuFX/XyRigA5xLqHCfnozIOMqo8ukJMhNhz9YLRn33WAK4M78n9/mmkSX1DNIXcK9vOpqEw5ofvfTnAxvezxlYZtt3D+wHAtmx6weuNFwYmfaFMYxFxpYc1i48J+pzYtn6xCNEhlWYgb5OeLXRH3V7z4o/KMwk1uTn0u9e8MoLbN0Zh8HPj5/XryMrPZ1X/vV/JOzbh7G4GMghKDcJP63sI3rvgRfzxQfvkZOVRcTIUU7TY/rzE+i3wGn7q4pXi1qTevQPNJ9kaD3OoW1iLdYw1eRDUXD8Yxiz5YI6csOzUHz+UdrcaQp0vb/0/YAcJxt8gLUPY7lkBQR2cP6+bYisExBi1f6nxbYEZlEOfWWfKrfT8tMhwC4dxtqHeTr8cb5od2PlG/39FVz0L+TJ7dD2hrpKr19O7YA215QtO70HDrwAQI4+iBCLtSOVETYM+r/O0ZRjbA4bxremMK6Oeoy1LcfyYveZvH74PwRYitGjQe+yRl/LPg1NwplS2I2UX8fwTqcpDM+KZViONSHd4xWcm+7gtyznzMIvj1cb/UrxCwGTi0MY/Zpw2/R7mX7tVZwbOZK5r/2HK24az7wZ97Jn2x90vvpKdM2LMOovjLm/84GHmDNlMuOn3OU0Of8csMC1uSfKsaT9BCaE+kJmzbIFvt/pTnIHT4dzF/4+s6/4jv92CqK7v/WR23SmAFJcvNjLo/uwxGeC0XUH72Ro7wtn9AS1qNYndzKgXP6jR/exbX8mFFeuVVz1ApzM40tvM/iA7DMeMsqeF2tbXc7HV0/k91wTE4/ksLJHU/4uMLHiXJH1N+syEnKsT4ePXPk9YT46yCrmiZ6PV/o9WqteUGjrsT+2n1f2WAeOUwbbZk/vcV/EkD0nRdVrZNeWhjmQq3PdpJwSTEBIaChDR0fyv40bS3vtl4wew7ZNG0HowcevwljkyOtvoEOX2i/AUBG5bkhYqBkCq69UAcsqMPgAMflmLjuQzZmANrT5K4P/utrgAzTriObi8yWzggkZZ2sxCNPmr3SOVWHwAQgIrfF+PYVvMyo+L97I9OGBE9Ye/h1HsnkluYCTtpt0quX8b2fR+/JjVvWdveP2x7Bp+7Kv7d83UISrRoidQUREhNy1a1eZsgMHDtC7d/XJiGJdPKnDTyfK5BGpiCY+OnLNdbfGqUcTmJrleb/TqNBgol0Q+TKzY6syUUCuJGXsxQzZFu/1E24UMKxpEH86ML/EnpSxF3v0ZDtHcy6VRwixW0oZUdFnDbKnXx83suoMPuAUg+/JuMLgA/Vm8BUNi5oafICdtdjG22mQPv18Uz560xkQJfc0gX3EryyNlrYv0yFKA950ts/Kbne+TCAQZbYuS/lt7OP2K6tbUfy2BAl6SxbNUt6y1ZW2SAZpUyApdeZLCUKPFH5IYUBIE2CxRjMJPaCzlRchpBlrgJ8PCANS6KztFjpAjyx5LS2242JtNyX1ELZtbOXoQAjrcZRFCFmirQS7IyB8S2YP2bRr1vZI+/bI0s9L2yk1mwZsrzWE1KxtLP0cED4g7aMszus4/6tZ/09Y648pv4jQ8nqlBR+zNTmYya8nUhjwLYo//x7rcURakLoAmxYjUgTaXheBMKCJAM6fA5rtuPlY2yE1ED5IYXVRCGmyizwr+Y3Pvy45H8q8L923TXu57e3bKoUf589t7I6f7U9qiJLzBUrbZT2vSn4/gdAKbL+DHqkLtG6nFdj2o0Pq/K2/MRIhjbbv9gX01uOC3noeavlIXaBVlxC2c7J+mfIrNLW99i0qu7KY0a+33fGw2I6P9VjJEtMpDCDN1t9eF8T5a9FMif0QmJHCF4Ox4gl5ZkMnQCKFP0gjAg2pC6Ao6DLgYie3uIEa/ZIT0nrh2xtVYTu5bAah5KYgrT8Mpaa85AeznsRlKXeRld4IqOA/Fb8XFdQtnT8vylYUNvOuC7R96/l2lBrbEkW2tgmtAIHFevHZfmKBZj05MSN1TdCEj3VbabK2vcSIaiZKjIFAKzX+wmaERYlBLjEGJQa71OhqNuNSotFmZKRWauitNyMJQmdrj65cfVs7y9xUbO+lZjvkOiR6qwEVOptOve3CNJ+/KEsPZcl+yv4QHYObcsSYi0WzP0+slBh9iQ9Cs/MVSw1BobWdOl3pBS51QQjNOg5hfV2ITss6r1v4ABo6rQjrRe5nO/Yl56Nvud+3RI+9dmG7uZb9vLSzoqNM/dL9IdFp+dbfSxhsn1Vw/Ep+C0DIQiR6zl8LAoGGpm+OFDqENFuNvTCgGUJtZRpCFoMsxnoDCLZ1BIoR0oTUBVs7ErIYTd8UnVZou2FYW1F6Q6wTzppqbDPuOrtOECVjCBbrTUqardem8EVo+bZjaO38lOkkycoDHiz6MOs1phUihMAiAmwdACc1oxwN0uj7+gRg8Wnj2u/QidJp2q5G0+eR3eqJevmuxsY7Hu7TVSicTYP06deHL72+DL5CgVlDl14/WVTbVpNv+bGD9RBRpXApDbKnX1V2xyqREkwa+Opsbk6bYbdIa/JxHYyPGMSyXzZx0/BBfBu1nacevp8mIU3JycykTceO/PuDRXz88susX/sdoWHN8TEY+PdHn7LwlRcoyivgjeVfMHf6VHz9/Xnpw0Vlv7uOC1VHpJvZ1dzxn7RnjoXDIQ6GK2q2Y2Ow9RNKwhBt7/Un8jAczMYUHoIhIQdLK39MfZrhH5WCsV8oumwjPqfykQJMfZshJGDS0EL98P07HRngg6W1Pxh06E/lI3JMmAaEoTXzRZg08BHIANecroOaBPJX7vlor7mHMuiQ689r/f1J8rO2T+SZkD468K95eKfIMeK78xym/qForQKgyIIu14QhLgNhC9009WqK9Nfj+/f51BOWtgHoz1Q8QcfcMQhdthGdLUbd1Ksplg5BiBwjMtQPpESXVoQux4SlTQAy2GB9n1mMuXsTRLYJGXbhBMYf1+Ww48ogZtot4XnXcSPLu1rnm9yakcBb9K/xMagRmkSXXozvnnSKIttYz4nkAgzxWaXnV+lx6ByMzwlrQIH98TJ3DEKYNKSPwCfp/G9r7t4ELcSA718ZaCEG67loKnE9gsg2IoN8kH56ZIjB/YsOuAC3GH0hxGKgD/CzlPLfzt5/gNQwZBdiwILFR480n/d6Wr1t5/3xGjr0UsMXMxo6zOgQtnPk/DbWugJoEdqME7u2k5ebQ+revzhx8CD33DeDBx+cyVNz/4+Tf2zFz1LMP2fN4tZbbuW7NT/w3YfvYjCa2LcvluCsHBJi4xg4aBCBWXlIQCdK/OWg2dTphMQsdeiQ+BYU0+fHOMwhBsizYEKPpakv6ATNZDbnskOwBOrx8z/C0DP+JJ7sQIfgM+QRSF5RAF2bnORQZk9b2zV0aOilho8hl25Gf3q2PE7CuS62+5rEIE3oseBvNmGQZox6Ay1FFhlaCGapo5nMo5nIJ0cGIiUEUkx70jhtbEmrPZn4CjMJ+e0Rx0Hgiy7GQL7OBAj0aPjtNBFMMbkEUYxAQwdZEJyaT5H0RY9GrgjCtMMHCzo0YRvOlTo0ocM63CjQhL50XNPXYkYCJuGD1An89SaKMWCROoI6muntt5/t2cMxnrP95lgQ7XyIy0pl0umjZJ4JpHVgGpnFTflZbyS9KAxtow/NdXo0BDo0/KSZEC2PfF0AAg1fYR378JUWhE5DJ8BHmimyGBASgkURAbKYLiKFTEsTkne2wA8jgWYjfhYjRXpfgkxFWHR6xB4oMvsSIIz46Mxomh8FuToKdT4Y0PDDhIlAioWJpkJiPGxCEIBFWMdgdH9rFMb6Y0aglYxfCB2BFFFw0A+j0BMm8skgmOBDRRRbDBThR77BH50mQZN0b36UeYE6irf7EpbTB4NmRgjY6ZtDm/1N0SGZhB8dfz5oGynQ7EZgrH962wCyTlo/F8I6DoAAnUUjVxdIgDSi18zodRKT1KND0oJccqUfAdKEphMICzQ3ZJG1PgSdJinEgA7oFHeWHBlAEb74aSaM+30RUiOMXLQEnXV/UkK8dQQCaUFoRkw624rAe0DaxkQseTpMqQb0UuKr+XCRMZlMAQX40cF4lqiQARTp/BEW6ziSSfiQ5xOASegJ0IwEiEIKQ5og8oyYpR6L0KP56NGZNSQCU4gfFFgQFolm0GEwabSV2eRqvvhYzIQYCsiSQWTpgvCxXZf24SMF4c2hliGbVVHvRl8IMQHQSylHCCE+E0KESykvzPvqAJuXLuLsiQuX8jMVFdU6bDO0bXsibril0s87tWvD7m1bGTpkCDu3b6VN61Zs+vVnbrpyDG8teBGArZt+xcdixGAuIC/zDE0MOjI0E34+PhSfPYmfjw6DZqKJKdshTZmWQq4//T3Yr753qvL6Yyoou5RtlW+QBP3Z7ZCWqgi3EzWgKoH1jU1KV8rldT8JG2OsLye6WEJLoJKkIZ7D+QzQXEKM+3S4kVDbnwTGFGyqfgPHVsSsFYFJbWHGtU7frzt6+pHAKtvrDcBIoNToCyFmADMAOnXqVKsvELZHtXKldq8rvyEI9AhR+XKFHTp0ImbnLsaOHsPmrVu4Z+p0AgMCuHfmbC4bPoIXn3kWhA/vfvgRK1Z+Q7u27Xj1xZd45sUX6Nu7D2t/2UDf3n0AHUL4lom90EqiJmxPFdIWTSSEnmB9W4QECzr8MGFGT5AsoghfzOjRaxqaLXJDh8QHM0YMCCnRWSTSR2Cxf8YR9s8vIOT5o+IjLNaeGZJCfNFLDX9hoghfdFjwE2Ys6DDZQjsl4C+MFFj88NOb0KNRKP2s+9MFoenCwHyGknA2CfgKHUbRBGE5f+PTo6HZngYspWs/lQ15FbawxTIP3QI795hEj0agLCZPBuCLGbOwahSaLA2c0qOhExoGKQALxdIHKaxPECU9rXyy8JVB+GJAAEYsFIs8QosD0XR6TAYDmq2ykBIpSr0EgMBHM2PA+oRj0euQJoHUC4QOjOixoMMs9NanLiyYhR4hJJoUoAu2RgnJwvON1IWCloXUtwAtzxqFoxWBNOIjLFikDh9hQZMlEVMlLbH+N0gjUljbqRMaFqlHSA2dlLbz4XyIaEmZXmpYhI4yU8vtXgrt/BO0RKDZ3CHWSLKyYdI6oSEsIG0BLXo0LJo1MswHC2adHjN6pLRGremltZds9vFBJzXMOh0+UsMsdWg6nfVhRtq+F4FeaNbtS1w1Qo/QNUHT7OeTyNLfVyc1/I1Ginz9ED6tEMIXgzkLsyUNk8zGT9+ytK0l16lAIjTN+v0IdJpmfTKxPXIKOzetKNfxFART6N8aYTyFyXIGH2HNuBqMD2b0BGDEaDFg0lvP/ZbNDLgCdxj9IM7fHzOAMmnkpJSLgEVgnZFb1Y7GTpvhCn1V0nfgAD769BMWvv8eb7z9JtPvvYdJkydz78wHmDJlChu3RRMc2pTn5s9nypQppdsFNAliwIABfP3110yePJnY2Fhad+vs0HeeKy7kgS8/cVWTFApFI8Id0Tt5QIDtdbCbNNSaLl260KpVK/r27YvFYmHTpk2sWbMGvV5Pv379KCqqPLph8ODB7Ny5k8GDnZ8uVaFQKBzBHQZ3N1aXDsBAINENGmpN586d6d69O4GBgXTo0IH//Oc/LF26lMjISHbs2MFdd1WeObNLly707NmTzp0d6+ErFAqFs6n3hGtCiBAgGvgNuBYYLqWscESzLgnXGhKNsc0KhaL2eFTCNSllDtbB3BhgbGUGX6FQKBTOxy1x+lLKTM5H8NRme2uETiPAk1NfKxQK78OrBlEB/P39SU9PbxTGUEpJeno6/v6uW79VoVA0LrwuDUOHDh1ISkoiLc0zljRzNf7+/nTo4Lr1WxUKRePC64y+wWCga9eu7pahUCgUXolHL5cohEgDTtRy8xbAOSfKcTXepNebtIJ36VVaXYc36a2r1s5SypYVfeDRRr8uCCF2VRay5Il4k15v0grepVdpdR3epNeVWr1uIFehUCgUtUcZfYVCoWhENGSjv6j6Kh6FN+n1Jq3gXXqVVtfhTXpdprXB+vQVCoVCcSENuaevUCgUinIoo69QKLwWIUSYEOIqIUQLd2upDk/R2iCNvhBisRBiuxDiGQ/Q4iOEOCmEiLL99a9In6NlLtbaWggRXVNN7tBur7WiY+wpWoUQTYUQ64QQG4QQa4QQvp56XCvR6pHH1bb/UOAnYCiwWQjR0oOPbUVa3XJsG5zRF3Zr8ALdhBDhbpY0APhKShkppYwEwimnryLN9d0O20n5OdaVzSo8jnUpc6VWyh1jKeVeT9EK/AN4U0o5DkgBJtVWlxu0PonnHlew/u5zpJQLgPXA5bXVVg96y2u9Bzcd2wZn9Kl4DV53Mhy4QQixQwixGLiSC/VFOljmSizAHUCO7b2jmhwtc6XWMsdYCOHjKVqllB9IKTfa3rYEptRBV31rNeOhx9Wmd4uUMkYIMRprD/rqOmhzqd4KtBbipmPbEI1++TV4W7tRC8BO4Eop5VDAgHXhmPL6KtJcr+2QUuaUW9vAUU31rr0CreWP8XWeorUEIcQIIBQ4VQdd9a11I55/XAXWDkAm1mXMPfbYltP6F246tg3R6HvaGrxxUsoztte7sObUKK+vIs3uboejmjxBe/ljHO5JWoUQYcBCrI/0Hn1cy2n16OMKIK3MAuKAS+ugzeV6y2lt565j626D6Ao8bQ3e5UKIgUIIPXAzMIsL9VWk2d3tcFSTJ2gvf4xjPUWrEMIX+AZ4Skp5oo666lurxx5Xm965QoiptrfNgFfroM3Vx7a81o/cdWy9LrWyA3wPRAsh2mFbg9e9cngR+BIQwFoq1icdLKtP6qKzvrWXOcZSyk3CthazB2i9FxgMzBNCzAOWAHd56HEtr3UzsBzPPK5gnbW6SghxH7AP6zm71UOPbXmto4EVuOPYSikb3B9Wf+TtQBt3a3FUn6Nl3qLT3do9Was6rp6prbEcW5WGQaFQKBoRDdGnr1AoFIpKUEZfoVAoGhHK6CsUCkUjQhl9hUKhaEQoo69QKBSNiP8HDDNW0rvdjx0AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 导入数据 共一列数据, 默认索引\n",
    "#data = pd.read_csv('data_fillna.csv', usecols=['0'], squeeze=True)\n",
    "data = pd.read_csv('PRSA_Data_Aotizhongxin_20130301-20170228.csv')\n",
    "print(data)\n",
    "plt.figure(figsize=(300, 20))\n",
    "data.plot() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "       PM2.5  PM10   SO2   NO2     CO    O3  TEMP    PRES  DEWP  RAIN\n",
      "0        4.0   4.0   4.0   7.0  300.0  77.0  -0.7  1023.0 -18.8   0.0\n",
      "1        8.0   8.0   4.0   7.0  300.0  77.0  -1.1  1023.2 -18.2   0.0\n",
      "2        7.0   7.0   5.0  10.0  300.0  73.0  -1.1  1023.5 -18.2   0.0\n",
      "3        6.0   6.0  11.0  11.0  300.0  72.0  -1.4  1024.5 -19.4   0.0\n",
      "4        3.0   3.0  12.0  12.0  300.0  72.0  -2.0  1025.2 -19.5   0.0\n",
      "...      ...   ...   ...   ...    ...   ...   ...     ...   ...   ...\n",
      "35059   12.0  29.0   5.0  35.0  400.0  95.0  12.5  1013.5 -16.2   0.0\n",
      "35060   13.0  37.0   7.0  45.0  500.0  81.0  11.6  1013.6 -15.1   0.0\n",
      "35061   16.0  37.0  10.0  66.0  700.0  58.0  10.8  1014.2 -13.3   0.0\n",
      "35062   21.0  44.0  12.0  87.0  700.0  35.0  10.5  1014.4 -12.9   0.0\n",
      "35063   19.0  31.0  10.0  79.0  600.0  42.0   8.6  1014.1 -15.9   0.0\n",
      "\n",
      "[35064 rows x 10 columns]\n",
      "[[False False False ... False False False]\n",
      " [False False False ... False False False]\n",
      " [False False False ... False False False]\n",
      " ...\n",
      " [False False False ... False False False]\n",
      " [False False False ... False False False]\n",
      " [False False False ... False False False]]\n",
      "0\n",
      "[[   4.     4.     4.  ... 1023.   -18.8    0. ]\n",
      " [   8.     8.     4.  ... 1023.2  -18.2    0. ]\n",
      " [   7.     7.     5.  ... 1023.5  -18.2    0. ]\n",
      " ...\n",
      " [  16.    37.    10.  ... 1014.2  -13.3    0. ]\n",
      " [  21.    44.    12.  ... 1014.4  -12.9    0. ]\n",
      " [  19.    31.    10.  ... 1014.1  -15.9    0. ]]\n"
     ]
    }
   ],
   "source": [
    "#引用数据\n",
    "len(data)\n",
    "df = pd.DataFrame(data, columns=[\"PM2.5\", \"PM10\", \"SO2\", \"NO2\", \"CO\", \"O3\", \"TEMP\", \"PRES\", \"DEWP\", \"RAIN\"])\n",
    "print(df)\n",
    "data = np.array(df)\n",
    "\n",
    "# 空值处理\n",
    "for i in range(len(data)):\n",
    "    for j in range(len(data[i])):\n",
    "        if np.isnan(data[i, j]):\n",
    "            data[i, j] = data[i - 1, j]\n",
    "\n",
    "print(np.isnan(data))\n",
    "print(np.isnan(data).sum())\n",
    "print(data)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[array([[   4. ,    4. ,    4. , ..., 1023. ,  -18.8,    0. ],\n",
      "       [   8. ,    8. ,    4. , ..., 1023.2,  -18.2,    0. ],\n",
      "       [   7. ,    7. ,    5. , ..., 1023.5,  -18.2,    0. ],\n",
      "       ...,\n",
      "       [  20. ,   49. ,    9. , ...,  995.5,   19.2,    0. ],\n",
      "       [  22. ,   70. ,   10. , ...,  995.5,   19.1,    0. ],\n",
      "       [  22. ,   70. ,    8. , ...,  995.6,   20.1,    0. ]]), array([[  17. ,   40. ,    8. , ...,  995.4,   19.4,    0. ],\n",
      "       [  18. ,   52. ,    5. , ...,  995.3,   19.1,    0. ],\n",
      "       [  21. ,   49. ,    6. , ...,  996.1,   18.4,    0. ],\n",
      "       ...,\n",
      "       [  11. ,    7. ,   14. , ..., 1030.4,  -17.7,    0. ],\n",
      "       [   9. ,    6. ,   12. , ..., 1030.2,  -17.8,    0. ],\n",
      "       [  11. ,   10. ,    9. , ..., 1029.8,  -17.2,    0. ]]), array([[  11. ,   10. ,    9. , ..., 1029.4,  -16.7,    0. ],\n",
      "       [  10. ,    6. ,   16. , ..., 1029.2,  -16.9,    0. ],\n",
      "       [  12. ,    8. ,   19. , ..., 1029. ,  -17.7,    0. ],\n",
      "       ...,\n",
      "       [  43. ,  102. ,   14. , ...,  994.8,   11.6,    0. ],\n",
      "       [  42. ,  113. ,   13. , ...,  995. ,   12.5,    0. ],\n",
      "       [  48. ,  134. ,   11. , ...,  995.7,   12.6,    0. ]]), array([[  61. ,  165. ,   10. , ...,  995.9,   12.1,    0. ],\n",
      "       [  85. ,  188. ,    9. , ...,  996. ,   12.3,    0. ],\n",
      "       [  80. ,  169. ,   13. , ...,  995.8,   12.1,    0. ],\n",
      "       ...,\n",
      "       [  20. ,   20. ,    3. , ..., 1021.3,    5.2,    0. ],\n",
      "       [  16. ,   16. ,    2. , ..., 1021.5,    5.8,    0. ],\n",
      "       [  18. ,   29. ,    2. , ..., 1021.6,    7.5,    0. ]]), array([[  17. ,   36. ,    2. , ..., 1021.8,    5.6,    0. ],\n",
      "       [  19. ,   52. ,    2. , ..., 1021.4,    6.3,    0. ],\n",
      "       [  24. ,   56. ,    2. , ..., 1020.3,    4.3,    0. ],\n",
      "       ...,\n",
      "       [  37. ,   90. ,   11. , ..., 1018.5,  -13.2,    0. ],\n",
      "       [  44. ,  121. ,   14. , ..., 1018.6,  -10.6,    0. ],\n",
      "       [  53. ,  132. ,   15. , ..., 1018.8,  -12.7,    0. ]]), array([[  38. ,   93. ,   12. , ..., 1018.5,  -14.2,    0. ],\n",
      "       [  29. ,   84. ,   12. , ..., 1017.8,  -14. ,    0. ],\n",
      "       [  31. ,   58. ,   10. , ..., 1016.7,  -15.1,    0. ],\n",
      "       ...,\n",
      "       [  75. ,   93. ,    6. , ...,  997.6,   20.3,    0. ],\n",
      "       [  92. ,  113. ,    7. , ...,  997.4,   20.7,    0. ],\n",
      "       [ 111. ,  111. ,    8. , ...,  996.9,   21.1,    0. ]]), array([[ 127. ,  127. ,    7. , ...,  996.8,   20.6,    0. ],\n",
      "       [ 117. ,  117. ,    7. , ...,  996.2,   19.9,    0. ],\n",
      "       [  85. ,   85. ,    7. , ...,  995.7,   19.9,    0. ],\n",
      "       ...,\n",
      "       [  54. ,   81. ,   23. , ..., 1030.6,  -11.1,    0. ],\n",
      "       [  30. ,   43. ,   15. , ..., 1030.1,  -11. ,    0. ],\n",
      "       [  23. ,   30. ,   13. , ..., 1029.7,  -10.7,    0. ]]), array([[  26. ,   36. ,   13. , ..., 1029.4,  -10.7,    0. ],\n",
      "       [  28. ,   39. ,   12. , ..., 1029.4,  -10.6,    0. ],\n",
      "       [  24. ,   34. ,   10. , ..., 1030. ,  -10.6,    0. ],\n",
      "       ...,\n",
      "       [  21. ,   21. ,    2. , ..., 1010. ,   -1.1,    0. ],\n",
      "       [   7. ,   16. ,    2. , ..., 1010.2,    0. ,    0. ],\n",
      "       [   9. ,   39. ,    2. , ..., 1009.7,   -1.2,    0. ]]), array([[   5. ,    5. ,    2. , ..., 1010. ,   -4.2,    0. ],\n",
      "       [  17. ,   17. ,    2. , ..., 1010.4,   -3.9,    0. ],\n",
      "       [  10. ,   16. ,    3. , ..., 1010.9,   -5.6,    0. ],\n",
      "       ...,\n",
      "       [  66. ,   66. ,    2. , ..., 1013.1,   13.2,    0. ],\n",
      "       [  61. ,   66. ,    2. , ..., 1013.1,   13. ,    0. ],\n",
      "       [  63. ,   75. ,    2. , ..., 1013.7,   13.3,    0. ]]), array([[  64. ,   76. ,    2. , ..., 1014.3,   13.9,    0. ],\n",
      "       [  73. ,   87. ,    2. , ..., 1014.9,   14.5,    0. ],\n",
      "       [  85. ,  108. ,    2. , ..., 1015.7,   14.8,    0. ],\n",
      "       ...,\n",
      "       [  16. ,   37. ,   10. , ..., 1014.2,  -13.3,    0. ],\n",
      "       [  21. ,   44. ,   12. , ..., 1014.4,  -12.9,    0. ],\n",
      "       [  19. ,   31. ,   10. , ..., 1014.1,  -15.9,    0. ]])]\n",
      "3506\n",
      "10\n"
     ]
    }
   ],
   "source": [
    "# 数据集分组\n",
    "datas = []\n",
    "\n",
    "#分组个数\n",
    "num_data_set = 10\n",
    "\n",
    "\n",
    "length = int(len(data)/num_data_set)\n",
    "for i in range(num_data_set - 1):\n",
    "    datas.append(data[i*length:(i+1)*length])\n",
    "datas.append(data[(num_data_set - 1)*length:])\n",
    "\n",
    "print(datas)\n",
    "print(np.shape(datas[0])[0])\n",
    "print(np.shape(datas[0])[1])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "data = datas[0]\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = data[:int(len(data)*RATE),:] \n",
    "test_set = data[int(len(data)*RATE):,:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def data_split(sequence, n_timestamp):   # \n",
    "    # 取前 n_timestamp 天的数据为 X；n_timestamp+1天数据为 Y。\n",
    "    X = []\n",
    "    y = []\n",
    "\n",
    "    for i in range(len(sequence)):\n",
    "        end_ix = i + n_timestamp\n",
    "\n",
    "        if end_ix > len(sequence) - 1:\n",
    "            break\n",
    "\n",
    "        seq_x, seq_y = sequence[i:end_ix], sequence[end_ix,0]\n",
    "        X.append(seq_x)\n",
    "        y.append(seq_y)\n",
    "    return array(X), array(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#数据归一化处理\n",
    "\n",
    "sc  = MinMaxScaler(feature_range=(0, 1))\n",
    "train_set_scaled = sc.fit_transform(train_set)\n",
    "test_set_scaled  = sc.transform(test_set) \n",
    "#print(train_set_scaled)\n",
    "#print(test_set_scaled)\n",
    "\n",
    "#训练集输入输出\n",
    "X_train, y_train = data_split(train_set_scaled,N_timestamp)\n",
    "X_train = X_train.reshape(X_train.shape[0], X_train.shape[1], X_dim)\n",
    "X_test, y_test = data_split(test_set_scaled,N_timestamp)\n",
    "X_test = X_test.reshape(X_test.shape[0], X_test.shape[1], X_dim)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2430\n"
     ]
    }
   ],
   "source": [
    "#sample \n",
    "print(X_train.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 画图 Loss, result\n",
    "def plot_loss(history):\n",
    "    plt.plot(history.history['loss']    , label='Training Loss')\n",
    "    plt.plot(history.history['val_loss'], label='Validation Loss')\n",
    "    plt.title('Training and Validation Loss')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "    \n",
    "def plot_load(model, X_test, y_test, scaler=None, figsize=(60,10)):\n",
    "    predicted_load = model.predict(X_test)                        # 测试集输入模型进行预测\n",
    "    if scaler:\n",
    "        predicted_load = scaler.inverse_transform(predicted_load)  # 对预测数据还原---从（0，1）反归一化到原始范围\n",
    "        real_load      = scaler.inverse_transform(y_test)# 对真实数据还原---从（0，1）反归一化到原始范围\n",
    "    else:\n",
    "        real_load = y_test\n",
    "\n",
    "    plt.figure(figsize=figsize)\n",
    "    # 画出真实数据和预测数据的对比曲线\n",
    "    plt.plot(real_load, color='red', label='Load')\n",
    "    plt.plot(predicted_load, color='blue', label='Predicted Load')\n",
    "    plt.title('R N N')\n",
    "    plt.xlabel('Time')\n",
    "    plt.ylabel('load')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "    MSE   = metrics.mean_squared_error(real_load, predicted_load)\n",
    "    RMSE  = metrics.mean_squared_error(real_load, predicted_load)**0.5\n",
    "    MAE   = metrics.mean_absolute_error(real_load, predicted_load)\n",
    "    R2    = metrics.r2_score(real_load, predicted_load)\n",
    "    EV    = metrics.explained_variance_score(real_load,predicted_load)   #新添加EV\n",
    "\n",
    "    print(\"EV\", EV)\n",
    "    print(\"MSE:\", MSE)\n",
    "    print(\"RMSE:\", RMSE)\n",
    "    print(\"MAE:\", MAE)\n",
    "    print(\"R2:\", R2)\n",
    "    \n",
    "    return real_load, predicted_load"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\ops\\init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "WARNING:tensorflow:From D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\ops\\init_ops.py:97: calling GlorotUniform.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "WARNING:tensorflow:From D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\ops\\init_ops.py:97: calling Orthogonal.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "WARNING:tensorflow:From D:\\adconda\\envs\\tensorflow37\\lib\\site-packages\\tensorflow\\python\\ops\\init_ops.py:97: calling Zeros.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Call initializer instance with the dtype argument instead of passing it to the constructor\n",
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "inputs (InputLayer)          [(None, 24, 10)]          0         \n",
      "_________________________________________________________________\n",
      "bidirectional (Bidirectional (None, 24, 100)           24400     \n",
      "_________________________________________________________________\n",
      "bidirectional_1 (Bidirection (None, 100)               60400     \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 50)                5050      \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1)                 51        \n",
      "=================================================================\n",
      "Total params: 89,901\n",
      "Trainable params: 89,901\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "# 构建神经网络模型LSTM\n",
    "# 输入层\n",
    "inputs = Input((N_timestamp, X_dim), name='inputs')\n",
    "\n",
    "# LSTM \n",
    "# 全连接层\n",
    "#LSTM_1 = LSTM(units=100, activation='relu', dropout=0.1)(inputs)\n",
    "#dense_2 = Dense(units=50, activation='relu')(LSTM_1)\n",
    "\n",
    "\n",
    "# 双向 LSTM\n",
    "Bi_LSTM_1 = Bidirectional(LSTM(units = 50, activation='relu',return_sequences= True, dropout = 0.1))(inputs)\n",
    "Bi_LSTM_2 = Bidirectional(LSTM(units = 50, activation='relu', dropout = 0.1))(Bi_LSTM_1)\n",
    "dense_1 = Dense(units=50, activation='relu')(Bi_LSTM_2)\n",
    "output = Dense(units=1, activation='relu')(dense_1)\n",
    "\n",
    "\n",
    "# 构造一个新模型\n",
    "model_1 = Model(inputs=inputs, outputs= output)\n",
    "print(model_1.summary())\n",
    "# # 编译\n",
    "model_1.compile(optimizer=optimizer, loss=Loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1944 samples, validate on 486 samples\n",
      "Epoch 1/10\n",
      "1944/1944 [==============================] - 1s 408us/sample - loss: 0.0258 - val_loss: 0.0229\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 2/10\n",
      "1944/1944 [==============================] - 1s 393us/sample - loss: 0.0254 - val_loss: 0.0228\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 3/10\n",
      "1944/1944 [==============================] - 1s 404us/sample - loss: 0.0244 - val_loss: 0.0230\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 4/10\n",
      "1944/1944 [==============================] - 1s 403us/sample - loss: 0.0245 - val_loss: 0.0259\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 5/10\n",
      "1944/1944 [==============================] - 1s 400us/sample - loss: 0.0239 - val_loss: 0.0219\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 6/10\n",
      "1944/1944 [==============================] - 1s 402us/sample - loss: 0.0241 - val_loss: 0.0227\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 7/10\n",
      "1944/1944 [==============================] - 1s 386us/sample - loss: 0.0244 - val_loss: 0.0251\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 8/10\n",
      "1944/1944 [==============================] - 1s 393us/sample - loss: 0.0227 - val_loss: 0.0210\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 9/10\n",
      "1944/1944 [==============================] - 1s 410us/sample - loss: 0.0239 - val_loss: 0.0212\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 10/10\n",
      "1944/1944 [==============================] - 1s 411us/sample - loss: 0.0223 - val_loss: 0.0255\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
     ]
    }
   ],
   "source": [
    "history_1 = model_1.fit(X_train,\n",
    "                        y_train,\n",
    "                        epochs = Epoch_WL, batch_size = Batch_Size, validation_split = 0.2)#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "#模型保存\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 1944 samples, validate on 486 samples\n",
      "Epoch 1/10\n",
      "1944/1944 [==============================] - 1s 389us/sample - loss: 0.0236 - val_loss: 0.0247\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 2/10\n",
      "1944/1944 [==============================] - 1s 384us/sample - loss: 0.0235 - val_loss: 0.0214\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 3/10\n",
      "1944/1944 [==============================] - 1s 378us/sample - loss: 0.0225 - val_loss: 0.0219\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 4/10\n",
      "1944/1944 [==============================] - 1s 384us/sample - loss: 0.0221 - val_loss: 0.0209\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 5/10\n",
      "1944/1944 [==============================] - 1s 385us/sample - loss: 0.0233 - val_loss: 0.0242\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 6/10\n",
      "1944/1944 [==============================] - 1s 383us/sample - loss: 0.0226 - val_loss: 0.0200\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 7/10\n",
      "1944/1944 [==============================] - 1s 378us/sample - loss: 0.0223 - val_loss: 0.0204\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 8/10\n",
      "1944/1944 [==============================] - 1s 385us/sample - loss: 0.0232 - val_loss: 0.0233\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 9/10\n",
      "1944/1944 [==============================] - 1s 383us/sample - loss: 0.0220 - val_loss: 0.0201\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n",
      "Epoch 10/10\n",
      "1944/1944 [==============================] - 1s 376us/sample - loss: 0.0219 - val_loss: 0.0239\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\n"
     ]
    }
   ],
   "source": [
    "history_2 = model_1.fit(X_train,\n",
    "                        y_train,\n",
    "                        epochs = Epoch_WL, batch_size = Batch_Size, validation_split = 0.2)#"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXgAAAEFCAYAAADpIfy5AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8qNh9FAAAACXBIWXMAAAsTAAALEwEAmpwYAAA1S0lEQVR4nO3dd3xUVdrA8d+TRkISCJCQQggBqQIJaOgtoAgo4IqorK4F17aW1X1fXcuq6zbd4u6rrhVXcVcF14qCghRBkCpg6F0ChBRCAqRA+nn/uENPmYSZ3MnM8/188mFy59xzHybw5MyZc58jxhiUUkp5Hz+7A1BKKeUemuCVUspLaYJXSikvpQleKaW8lCZ4pZTyUprglVLKS2mC91Ii8rCIZIvICRE55ng88gL7e9hd7RuLiCwRkdRanp8pIo+c8f3dIvKVE/3eJiLvVHP8BhF5sZbz3hGR2+rqv5bza+3/Avq9TUQ+cHW/qnEF2B2Acg9jzPPA846ks8oY87oL+nNbew+yCJgI/M3x/VDHsQYxxvwX+O+FBiUiDwHvGGOOuqN/5Z10BK/U2RYBQ0VEHN8P4wISvAs9BETYHINqYjTB+yDHNMVkEZklIt+ccfweEdkvIgdF5NFzznlGRJ454/tURz9/F5HDIrJUREIa0l5EpohIhoisF5EPROTtOuI/L846+n9aRLJEZC7Qora+jTF7gWNATxFpD4QBG+p6fWqJ9bypGxG5Q0QyRWQF0O6M4z8Rkd2O5/7pOHa/iGQD7YHvHVNtzevo/woR2S4iB0Tkibpen/oSkUccr8MOERnnOOYnItMcr/N+EbmmtuOqcWiC913PAm8DJ/8jBgM3AwOBLsDDIhJWRx+DgL1ALFbiHNvA9i8ClwFvABXGmNtr6qCOOM/rX0T6A7cDFwNPA8l1xAjWiH2Y42uxMcY08PWpLv52wF8csd7o6O+ku4HrgAQgVUQuNsa8bIyJAQ4A/YwxMcaY47X03wb4D3A90Au4QUSudDxd359Xdf1fDtyK9TpeA0wXkWigDzAe6ACMAUY7TqnpuGoEmuB919vGmC+MMccAjDElwC1YSexDoDUQWUcfOcArxphyrFFuywa2LwGCHF+1/pusI87q+h8MfGmMOWKM+R7YVEeMAN9wOsEvcuK69dEP6zORfcaYdM6e/rndcc2ZWL9E2jag/8FAmjFmo+Nn+w5wMsHX9+dVnXHAe47Xcyuw2hHzHqAK67OLrlhTStRyXDUCTfC+a9WZ34jIRcBSIB/4X6wRY132mtPV6pypWldT+7XAJ8BU4Pe1dVBHnNX1L+dcq8qJOE8m+BGOxw19far9K1QXj4hEYL0OfsCfOOfnU0/mnMcnv6/vz8vp/h2/TC4GlmG9M1mA9US1x1Xj0ASvTuoLpGNN23QD4p04p75J4rz2IpIAdAIuNsZcYozZfgFxVhfPGmCciLQUkUuApDqDNCYbax4+3Biz04nr1sd6YKCIxDv+7pc5jncGKoBXgRDg0nPOOwx0EEvrWvpfAfQRkV4i0gJrOmXuyb9aA2M+01zgJhGJEJHuwADgOxG5DOu1mQU8DvR3xFrtcRfEoZygCV6dtNDxZw4wBWuutmsjXPcA1r/DTBFJF5E5IlJb8qxXnMaY5cAHwE7gBWCrk3EtwjF6b8h1a4lnH/Bb4HvgcyDN8dQGx1cW1mcFm87p/2ms6ZbDWHPZNfWfh5XUPwa2AB8ZY+pcx1+Da0Wk6IyvXxpjFgLvAhuBL4DbjTE5wLdAIXAQa7T+a8e7hZqOq0Yg+lorO4nIRGCiMeYOEfEHXgJ2GmNcfvOOUr5GR/DKbj8AXUQkC9iHNV3zob0hKeUddASvlFJeSkfwSinlpTTBK6WUl/KYYmORkZEmMTHR7jCUUqpJWbdu3WFjTFR1z3lMgk9MTGTt2rV2h6GUUk2KiOyr6TmdolFKKS+lCV4ppbyUJnillPJSHjMHr5RqHOXl5WRkZFBSUmJ3KKoegoODiY+PJzAw0OlzNMEr5WMyMjIIDw8nMTERrfvVNBhjyMvLIyMjg44dOzp9nk7RKOVjSkpKaNOmjSb3JkREaNOmTb3fdWmCV8oHaXJvehryM3MqwYvIWyKyUkSedLaNiPzCsQfkEhFJE5E36h2dE3YfKuQPc7ZSVuHMPg5KKeU76kzwIjIJ8DfGDAI6iUgXZ9oYY14zxqQaY1Kx6kC/6eLYAdiff5y3vtvLsl257uheKeViL7/8MqmpqYSEhJCamspnn31W7z4eeughl7aryzPPPMN7773nkr4akzMj+FROl2+dDwytTxvHJsPRxpjzblMVkbtEZK2IrM3NbViCHto5ipYhgczekNmg85VSjev+++9nyZIltGvXjiVLlnDNNdfUu48XXnjBpe28lTOraEKxdmMBaz/KS+rZ5j7gteo6NsZMA6YBpKSkNKhucVCAH2N7xjBnYyYl5ZUEB/o3pBulfNLvZm9ha2aBS/u8OK4Fv53Qs97npaam0q9fPzZu3MjXX39NUVERkydPpri4mM6dOzN9+vSz2i5ZsgSwRtfl5eUsW7aMgoIC5s2bR0xMjFPtWrZsyaRJk8jPz+eiiy6iV69ePPHEE07FW1paym233UZmZibx8fFMnz6dyspKrrvuOgoKCmjTpg0fffQR5eXl5x0LCGicBYzOjOCLsPaIBAir4Zxq24iIHzASWHJBUdZhQnIcxWWVLN5+yJ2XUUq50apVqxg0aBBff/01AFlZWTzwwAMsXLiQ9PR0cnJyajx39+7dLF26lEmTJvHNN9843W779u3Ex8fz3XffsXv3bqeTO8Cbb75Jr169+Pbbb+nSpQtvv/02W7duxc/Pj6VLlzJ16lSKioqqPdZYnPk1sg5rymUVkAzsqEebYcBqd+/BOLBTa9qEBjFnYxbjese681JKeZWGjLTdpVevXkyaNOnU94GBgfzrX/9i+vTp5Ofnc+LEiRrPveWWWwBISEigrKzM6Xbt2rVj3bp1DB8+nAcffLBe8W7duvVUvAMHDmTu3Lncfffd9OrViyuuuIIuXbowduxYLrnkkvOONRZnRvCzgJtF5B/A9cAWEfljHW2+dBwfAyx1Tag1C/D348resSzankNxaYW7L6eUcoOwsLCzvn/rrbeYPHkyM2fOJDQ0tNZz63q+pnbz5s3jqaeeYuXKldx00031irdnz56sWrUKsN599OzZkw0bNjBkyBDmz5/PkSNHWLZsWbXHGkudCd4YU4D1IeoqYKQxZoMx5sk62hxzHH/CGPOpq4OuzvikWErKq1i4rea3cUqppmP06NE899xzjBo1CoCDBw/WcUb99e3blwceeIBRo0YxZcoUNm/eXGPbp59+mpSUFFJSUnj55Ze544472LJlC8OHD2fXrl3cdtttJCYm8tJLLzF48GCys7NJSUmp9lhj8Zg9WVNSUsyF1IOvqjIM+vMikuIjePOWxnsBlWpqtm3bRo8ePewOwyO8+eabzJw5k8DAQAIDA3n44YdJTU21O6waVfezE5F1xphqk57X1KLx8xOu6h3He6v2cexEOS1DnC/Io5TyTXfeeSd33nmn3WG4jVeVKhifHEtZZRULtuo0jVJKeVWC79s+gnYRIXrTk1JK4WUJXkQYnxzL8t2HOVJc81IppZTyBV6V4AEmJMVRUWWYtyXb7lCUUspWXpfge8a1oGNkqE7TKOWhBgwYwK5duwD44osvmDp1aq3tq1vVUlMRsWeeeeZUaYLqpKWlkZaW5nR/9eVpRcm8LsGLCOOTYln1Yx6HCnVLMqU8zdixY1mwYAEAixYtYsyYMfXuo6FFxGpK8N5alMxrlkmeaUJyHP/8ZjdzN2Vz6+BEu8NRynPNfQyyN7m2z5jeMO7PNT49ZswYnn/+ee69914WL17M008/XWthseqcWUTsyJEjXHfddVRWVmKMITU1tdr+Hn/88VOlid99910WLVpUbX/VFRF79tlnayxoVhc7i5J53QgeoGt0OF2jw5izUadplPI0AwYMIC0tjYyMDJo3b06bNm3qVVjsXNOmTWP8+PEsXrz41IbU1fX33HPP8dhjj/HYY4+dldzPVV0RMXC+oJkz/TVWUTKvHMGD9WHr3xfsJOvYCWJbhtR9glK+qJaRtrv4+/tz6aWX8pe//IUrrrgCqF9hsXPt3buXG264AeBUGYAL6a+6ImKtWrVyuqCZM/01VlEyrxzBA4xPjgPgy41ZNkeilDrX2LFjef31108lsfoUFjtXQkICW7ZsATg1v15TfyEhIRw/fhyAmsq0VFdEDJwvaOZMf41VlMxrE3zHyFB6tWvBbE3wSnmcMWPGEBYWxoABA4ALKyx211138cknn5CamkpBQUGt/Y0ePZpPP/2UIUOG1JhAqysiVh+eVJTMa4qNVef1b/fw57nbWfrISBLaNHdp30o1VVpsrOmqb7Exrx3BA1zl2Pxjzib9sFUp5Xu8OsG3b92cvgkRzN6g0zRKnclT3rkr5zXkZ+bVCR5gfFIc27IK2JPbePsgKuXJgoODycvL0yTfhBhjyMvLIzg4uF7nee0yyZOu6h3LH7/cypwNWTx4eRe7w1HKdvHx8WRkZJCbm2t3KKoegoODiY+Pr9c5Xp/gY1oG0y+xNbM3ZvLLyzojInaHpJStAgMD6dixo91hqEbg9VM0YJUu2H2oiB05hXaHopRSjcYnEvy4XjH4CVphUinlU3wiwUeGNWPwRZHM2ZilHywppXyGTyR4gAnJsezLO86mg8fsDkUppRqFzyT4MT1jCPAT5mjpAqWUj/CZBB/RPIjhXaP4cmMWVVU6TaOU8n4+k+ABxifFcvDoCX44cMTuUJRSyu18KsGPvjiaoAA/LV2glPIJPpXgw4MDGdktii83ZVGp0zRKKS/nUwkerNo0uYWlrNmbb3coSinlVj6X4C/r0ZaQQH9m636tSikv53MJvnlQAJf1aMu8zdlUVFbZHY5SSrmNzyV4sGrT5BeXsWJPnt2hKKWU2/hkgh/RNYrwZgFam0Yp5dV8MsEHB/ozumc0X2/JprSi0u5wlFLKLXwywQNMSIqjoKSCZTsP2x2KUkq5hc8m+CGdI2kZEsgcXU2jlPJSTiV4EXlLRFaKyJP1bSMir4rIhAsN1NWCAvwY1yuGBVtzKCnXaRqllPepM8GLyCTA3xgzCOgkIudtbFpTGxEZBsQYY2a7OG6XGJ8UR3FZJYu3H7I7FKWUcjlnRvCpwIeOx/OBoc60EZFA4E0gXUSurq5jEblLRNaKyFo7NgAe2Kk1kWFBWkJYKeWVnEnwocBBx+N8INrJNrcAW4G/Av1F5IFzTzLGTDPGpBhjUqKiouob+wUL8PdjXK9YFm3Pobi0otGvr5RS7uRMgi8CQhyPw2o4p7o2fYFpxphs4D1g5IWF6h4TkuMoKa9i4bYcu0NRSimXcibBr+P0tEwykO5km91AJ8exFGBfQ4N0p5QOrYhpEawlhJVSXifAiTazgGUiEgeMA6aIyB+NMU/W0mYgUAW8LSJTgEBgsisDdxU/P+HK3rG8t2ofx06U0zIk0O6QlFLKJeocwRtjCrA+RF0FjDTGbDgnuVfX5pgxptAYc50xZrgxZpAx5uC5fXuKCcmxlFVWsWCrTtMopbyHU+vgjTFHjDEfOubTG9zGU/VpH0F8qxCtTaOU8io+eyfrmUSE8UlxLN99mPziMrvDUUopl9AE7zA+KZaKKsO8zU3uDYhSSlVLE7xDz7gWdIoM1do0SimvoQnewZqmiWXVj3kcKiyxOxyllLpgmuDPMD45jioDczfpNI1SqunTBH+GrtHhdIsO12kapZRX0AR/jvFJsXyffoTMoyfsDkUppS6IJvhzjE+OA+CrTVq6QCnVtGmCP0fHyFB6tWuhNz0ppZo8TfDVmJAUx4aMY+zPO253KEop1WCa4KtxVVIsAHM26SheKdV0aYKvRnyr5vRNiNASwkqpJk0TfA0mJMWxLauA3YeK7A5FKaUaRBN8Da5KikUEXROvlGqyNMHXILpFMP0TWzN7QybGGLvDUUqpetMEX4vxyXHsyS1me3ah3aEopVS9aYKvxbheMfj7iU7TKKWaJE3wtYgMa8bgi9owZ2OWTtMopZocTfB1GJ8Uy76842w6eMzuUJRSql40wddhTM8YAv2FORt1TbxSqmnRBF+HiOZBDOsSxZwNmVRV6TSNUqrp0ATvhAnJsWQeK+GHA0fsDkUppZymCd4Jl/eIJijAT0sXKKWaFE3wTggPDmRUt7Z8uSmLSp2mUUo1EZrgnTQ+OZbcwlLW7M23OxSllHKKJngnjerelpBAf2brTU9KqSZCE7yTmgcFcPnF0czbnE15ZZXd4SilVJ00wdfD+KRY8ovLWLEnz+5QlFKqTprg62FE1yjCmwUwR/drVUo1AZrg6yE40J/RPaP5eks2pRWVdoejlFK10gRfTxOS4ygoqWDZzsN2h6KUUrXSBF9PQztHEtE8UEsIK6U8nib4egr092NszxgWbM2hpFynaZRSnksTfANMSI6juKySxdsP2R2KUkrVSBN8Awzo2JrIsCC96Ukp5dGcSvAi8paIrBSRJ51tIyIBIrJfRJY4vnq7Kmi7Bfj7cWXvWL7Zfoji0gq7w1FKqWrVmeBFZBLgb4wZBHQSkS5OtkkCZhpjUh1fm1wdvJ3GJ8VRUl7Fwm05doeilFLVcmYEnwp86Hg8HxjqZJuBwHgRWeMY3Qece5KI3CUia0VkbW5ubn1jt1VKh1bEtAjWEsJKKY/lTIIPBQ46HucD0U62+R643BjTHwgErjz3JGPMNGNMijEmJSoqqr6x28rPT7gqKZalO3M5dqLc7nCUUuo8ziT4IiDE8TishnOqa7PRGHNyeLsWOG9qp6kbnxRLWWUV87dk2x2KUkqdx5kEv47T0zLJQLqTbd4VkWQR8Qd+Amy4kEA9UZ/2EcS3CtENuZVSHum8efFqzAKWiUgcMA6YIiJ/NMY8WUubgcBGYAYgwBfGmIWuDNwTiAjjk+J4c9mP5BeX0To0yO6QlFLqlDpH8MaYAqwPUVcBI40xG85J7tW1OWaM2WyMSTLG9DbG/Mb1oXuGCcmxVFYZ5m3WaRqllGdxah28MeaIMeZDY0yNWcyZNt7o4tgWdIoM1do0SimPo3eyXiARYXxyHKt+zONQYYnd4Sil1Cma4F1gQlIsVQbmbvKpNy9KKQ+nCd4FukSH0y06nE/WZ1BYomvilVKeQRO8i/xsYAIbM44x6LlveOaLLaQfLrY7JKWUj3NmmaRyws2DEukdH8H05Xt5b9U+/r0ynVHd2jJ1SEeGdG6DiNgdolLKx4gxxu4YAEhJSTFr1661OwyXyCko4f1V+3h/9X7yisvoGh3GbYM7ck3fdoQE+dsdnlLKi4jIOmNMSrXPaYJ3n5LySmZvyGT68nS2ZhUQ0TyQn/ZP4OaBHYiLCKm7A6WUqoMmeJsZY1izN5/py9OZvzUbEWFsrximDk7k0g6tdPpGKdVgtSV4nYNvBCLCgE5tGNCpDQfyj/Puqn18sGY/X27Mone7lkwdksj4pDiCAvQzb6WU6+gI3ibFpRV8+sNB3lm+lz25xUSFN+OmAQncNKADUeHN7A5PKdVE6BSNB6uqMizbfZjpy/eyZEcuQf5+jE+O5fYhHenVrqXd4SmlPJxO0bjLsQw4sMb6ylgD4g/XvgmtEp3uws9PGNE1ihFdo9iTW8S/V6Tz8boMPl1/kH6JrZg6pCNXXBxNgL9O3yil6kdH8M6qKIWsjVYiP5nUCx0FxgJCoN0lkLMFAkPg5lnQtnuDL3XsRDkfrT3AOyvSyThygnYRIdw8qANT+rUnormWJFZKnaZTNA1RkHU6mWd8D5lpUFlqPdcyAdr3t77i+0FMb/APtBL8u9dAZTnc/CnE9b2gECqrDAu35TB9+V5W/ZhPSKA/11zSjqmDE+kSHX7hf0elVJOnCb4uleWQvclK5AdWw4Hv4dh+6zn/ZhDXx0rk7QdYST08pua+8vbAf34CJ47Ajf+FxCEuCXFbVgHTl+9lVlomZRVVDOsSydQhiaR2bYufny6zVMpXaYI/V1Hu2VMtmT9AxQnrufC4M0bn/SE2CQLquarl2EF49ydwdD/c8B50Ge2y0POKSpm5Zj/vrtpHTkEpHSNDuXVQByantCesmX6kopSv8e0EX1kBh7ZaI/OM762EfmSv9ZxfoJXA2w9wjND7Q8t411y3+LA1XXNom/XBa89rXNOvQ3llFXM3ZzN9+V5+2H+U8GYBXJfSntsGJ5LQprlLr6WU8ly+leCP559O5AdWw8H1UO6o7Bja9vTovP0AiE22PhR1lxNHYcYN1ruFCS/CJbe45TJpB44yfflevtyYRaUxjO0Zw0OXd6VbjM7TK+XtvDvBHzsIu+afnj/P220dF3+I6eUYnfeH9v0gogM0dlmAsuPw35/BnkUw5lkYdJ/bLpVTUMK7K/fx7xXpFJVVMDE5jl9d3pXEyFC3XVMpZS/vTvBbZsFHt0LzNqcTefsB1gqWIA9JbBWl8MkdsO0LGPEYpD7m1l80R4+X8cbSH3lneTpllVVcd2k8D1zWhXZa4Ewpr+PdCb60EIoOQetOjT86r4/KCpj9S0h7Hwbea43m3RzvocISXl28hxmrrRVBNw5I4N6RF9E2PNit11VKNR7vTvBNSVUVfP04rH4d+v4MJrwEfu6vD3/w6Ale/mYXH67NINBfuHVwIvcMv4hWoXrTlFJNnSZ4T2IMLHkOvv0LXHw1TPoXBDROok0/XMwLC3fy+YZMwoIC+Pmwjvx8aEfCgwMb5fpKKderLcFrgZPGJgIjn4Ar/gRbP4cPfmp9ENsIEiNDeWFKX75+aDhDOkfywsJdDPvrYl7/dg8nyiobJQafVFEGaTOtz2KUakSa4O0y+H5rimb3InjvWig51miX7hodzus3X8rs+4fSp30Ef567nWF/Xcw7y/dSWqGJ3uVWvwaz7oE10+yORPkYnaKx2+ZP4NO7ILon/OxTCI1s9BC+T8/n+a93sHpvPu0iQvjlZZ259pJ4rWDpCieOwovJUHLUug/joY3uvfdC+RydovFkva6FKTMhdwdMHwcFmY0eQr/E1nxw10De+/kAIsOb8egnm7j8H9/yedpBqqo8YwDQZC1/wXp3Nu6vUHwI1v3b7oiUD9EE7wm6XgE/+8SqYPn2GMj/sdFDEBGGdolk1r2DefOWFIID/XnwgzTGvbiMeZuz8ZR3ek1KQRaseh16XwcD7oYOQ6yEX15id2TKR2iC9xSJQ+HWL6x1/W+Pg5yttoQhIoy+OJqvfjmMf/60L+WVVdzz3jqufmU53+7M1URfH9/+GaoqYNRvrO9H/BoKsyDtPXvjUj5DE7wnaXcJTJ1rPX7nSshYZ1sofn7ChOQ45v9qOH+dnEReURm3vr2GG95Yxeof82yLq8k4vBvWvwspt5/e4avjCOtu6+9esFbWKOVmmuA9TdsecPs8aNYC/jMR9i6zNZwAfz+uT2nP4odT+cPVPUnPK+aGaau4+a3VbDhw1NbYPNo3f4CAYBj+8OljIjDiUTh2ADbMtC825TM0wXui1h2tJN8y3lpCuWOe3RERFODHzYMS+faRkfzmyh5sPniMq19Zzp3/Wcv27AK7w/MsB9fD1lnWUtiwtmc/1/kyq07Ssr9bG80o5Uaa4D1Vizi47StrRP/fm2DTx3ZHBEBIkD93Du/EskdH8T+ju7JqTx7jXlzGAzN/4MfcIrvD8wwLn7GK3w26//znTo7ij+6DTR81emjKt2iC92ShbeDW2VZ1zE/ugLXT7Y7olLBmAfzysi4se3QkvxhxEQu35jD6/5by6483kHGkce7M9Uh7voG938LwRyC4RfVtuo619vFd+jxU6Y1lyn00wXu64BZw08fQ+XKY8xAsf9HuiM4S0TyIX4/tztJfj+TWQYnMSstk5PNLePrzzb6X6KuqYOHvrE3ZU26vuZ0IDP815O+BzZ82XnzK5ziV4EXkLRFZKSJP1reNiESLyA8XGqhPC2oOU2ZY2/4teBoW/cEqWuZBosKb8fSEi1nycCqTL23PjNX7GfqXxUx+bQXvrkwnr8gH6rBsnQVZaVatobr28e0+HtpeDMuet34xKOUGdSZ4EZkE+BtjBgGdRKRLPds8D+i92RcqIAiufcva9m/Z8zD3UY9MDHERITw3qTdLHknlkTHdKCgp56nPt9D/2UXc+vYaPlmXQWGJF364WFlurZxpezEkXV93ez8/GPa/kLvd2ghGKTcIcKJNKvCh4/F8YCiwy5k2IjIKKAayq+tYRO4C7gJISEioR9g+ys/fKlDWrAWsfBlKC2Diy+DvzI+xccW3as59Iztz38jObM8u4PO0TL5Iy+R/P9pAs8/8uKxHWyYmx5HarS3Bge6vie926/9j3YH80/86X+O/5zWw5M+w9G/QY6KV9JVyIWcyQyhw0PE4H7jEmTYiEgQ8BVwDzKquY2PMNGAaWMXGnI7al4nAFX+E4Jaw+E/Wna+T3657SsBG3WNa0H1sC349phvr9x/li7SDzNmYxVebsglvFsCYXjFc3SeOQZ3aNM0CZ2XFVn3/hEHQdYzz5/n5W+vkP7sbds6F7le5L0blk5xJ8EWcnmIJo/ppneraPAa8aow5Kp68lV5TJGLd9t4sHOY9BjNugCnve84etDUQES7t0IpLO7TiqfEXs2JPHp+nZTJvczYfr8sgMiyI8UlxTEiO45KECJrMv5vVr0NRDlz/n/pvw9hrsjWK//Yv0O1Kz952UjU5zgyX1mFNuQAkA+lOtrkcuE9ElgB9RORfFxKoqsbAX8DVr1jL8t69xipN20QE+PsxvGsUf78+mbVPXs5rN11Cv8TWzFizn2tfW8Hwvy3mr/O2syO70O5Qa3c8H757EbqOg4SB9T/fP8Cai8/aALsWuD4+5dPqrAcvIi2AZcAiYBwwBbjOGPNkLW0GGmOOnfH8EmNMam3X8dl68K6wZZa1Tr5td/jZZxAWZXdEDVZQUs78LTl8sSGT5bsPU1ll6BYdzsQ+cUxMjqN96+Z2h3i2+U/CipfhFysg+uKG9VFZDi9dYt31esdCHcWrerngPVlFpBUwGlhqjKnpA9M629RGE/wF2r0QPvgZtGwHt3xulTlo4g4XlfLVpiw+T8tk3b4jAFySEMHE5DiuSoojKtzmzx2OZViJude1cM1rF9bX2rdhzq/g5s/golGuiU/5BN1021fsWwkzrrd2DJr0JnQaYXdELnMg/zizN1orcbZnF+InMKRzJBOS4xjbK4YWdmwc/vl9sPFDeGAdRFzgKrCKUnipr9XP1Lk6ildO0wTvS3K2wEe3weFd1gqNEY955DLKC7Ezp5Av0jL5fMNBDuSfICjAj5Hdori6TztGdW+kZZe5O+DVgTDgHhj7nGv6XD0N5j4Ct86BjsNc06fyeprgfU1ZMXz1CKS9by3du/ZfXjFlcy5jDGkHjvJ5WiZzNmZxuKiUsGYBXNEzmonJcQztHOm+ZZcf3AQ/fgsPbrBqBrlCeYm1f2tkF7htjmv6VF5PE7yv2vBf+PJ/wC8AfvKqV6+zrqwyrNyTxxcbDjJ3czaFJRW0CQ3iyt6xDOjUmu4xLUhs09w1Cf/A9/DW5TDyN9ZyVVda+Qp8/QRMnQcdBrm2b+WVNMH7srw98PFUaxle/7vhij949E1RrlBaUcmSHbl8sSGThVtzKK2wSjo0C/CjS3SYdeNVTDg9Yq0/24TV4/UwBt4ZD4d3wi9/gGZhrg2+7Di8mGRVm7z5M9f2rbxSbQneuyZn1fnaXAQ/XwALfgurX4P9K2HydIjsbHdkbtMswJ8xPWMY0zOGkvJKdh8qYnt2ITuyC9ieXciSHbl8vC7jVPvIsGb0iA2ne0y4lfxjw+ncNoxmAdXM5e9eCPu+gyufd31yB6uw3KD7YeFvIWMtxFf7/1Ypp+gI3pds/wo+v9faD3T8/0HyDXZHZJvcwlJ2ZBey3ZH0t2cXsDOniDLHaN/fT7goKpRup0b74XSPDiP2gyuQsiK473urAJw7lBbBC70hvh/c9GHd7ZVP0xG8snS/EmKXWzdFfXYX/LgErvybe0aiHi4qvBlR4c0Y2iXy1LGKyirS84qthJ9lJf31+44we0MmABP9lvNS0GZejHiUvC93nhrtd4sOJ7SZC/8rNQuDQffCN3+EzDSI6+O6vpVP0RG8L6qssGqfLP0btOkM10235nxVtQpKytl1MI+uH19GoQnhly1eYHtOMUWlFafadGjTnO4x4XSLaUGPmHC6x7YgoXVz/P0auJ695Jg1ik8cZtUZqqeqKkN5VRUVlYbyyirKKw0VVVUEB/jTKtRN7zyULfRDVlW9vUvhkzvhxBEY8yfod4feYFOTk2vUb/oEulyOMYaMIycco31rmmdbdgHph4upcvyXCgn0p2tMOD1iwrkoKgw/P6GisuqshFt+KgGfTMbW9xVVVYzNfYeJR//D421fZ49/ouPck8+bM845fX6Z47nKqur/X4vA8C5R3DQggVHd2zbN6p3qLJrgVc2KcmHWL2D3AmuXoatfhpBWdkflWUqL4KU+ENXd2iO3ll+CJeWV7MopYlt2walpnu3ZheQXl53XNsBPCPAXAv39CPT3I8Dv5GPrzwg5zr8LpvJDUD9eav3EqecC/P0I8vcjwF8I8PMjKMD6M8BfTh2vrs8Afz8OHjnBR+sOkFNQSkyLYG7o154p/dsT21L35GmqNMGr2lVVwapXYOEzEB5r7RyVMMDuqDzHt3+1au//fCG071fv040xFJywpnMCHck40F+cK4e86Pew7B9w32qI6lbva1enorKKRdsP8f7q/SzblYsAo7pHc9OABIZ3jWr4tJKyhSZ45ZyMddaa+WMZMOo3MORXustQ8WF4sY9V16cBc+EXfv08ay6++1Vw7Zsu7/5A/nFmrtnPh2sPcLiojHYRIfy0f3uuT2lP2xbBLr+ecj1N8Mp5Jcdg9oOw5TPolArXTIPwaLujss+8x60NPe5d5bIRdL3Nf8raovH+tdZ9DW5QVlHFgq05zFizj+W78wjwE0ZfHM2NAxIYclEkfjqq91ia4FX9GAPr/21t7N0sHK55AzpfZndUje/ofvjnpZB0g/XZhF2KDsELSdBrklVyws32Hi5m5pr9fLT2AEeOl9OhTXN+2j+B6y6Nr99dv6pRaIJXDZOz1Zqyyd0OQ39l1V7xt6Esr10+uwc2f2qVJGjZzt5Y5j0Oq9+wShO37tgolywpr+TrLdm8v2o/a9LzCfQXxvaK5cb+CQzs1LrpbKno5TTBq4YrO27t+7r+39adlde+Ba062B2V++VshdcGw+D7rU3O7VaQZVWaTJ4CE19q9Mvvyilkxpr9fLIug4KSCjpFhXJj/wQmXxpPRHNdV28nTfDqwm3+BGY/ZC0RnPhPuPhquyNyrxlTYN8KeDANmre2OxrLlw/DunesdxQR7W0J4URZJV9uymLG6n2s33+UoAA/xveO5cYBCVzaoZWO6m2gCV65Rv5e+Ph2yFwPKT+3bo4K9ML10/tWwvSxcNnT1obYnuJYhrWi59Jb4aq/2x0N27IKmLF6P5/9cJCi0gq6RYdz44AErrmknT07bPkoTfDKdSrK4Jvfw4p/QnQvqzJlVFe7o3IdY+DtsXAk3RopB3nYJt+zH4S0GdZGIy3i7I4GgOLSCmZvyGTGmv1szDhGSKA/E5JjuXFAB5LjW+qo3s00wSvX2zkfZt0D5Ses0rl9bvSOMgc75sLMKVa1zZTb7Y7mfEfSrY2++98F4/5sdzTn2ZRxjBlr9vF5WibHyyrpGdeCGwckcHWfdoS5siCbOkUTvHKPgiz49E5IXwa9r4fx/7CWVTZVVZXw+lBrA+z7VnvuiqFZ98Hmj+HBjR57j0JhSTmz0jJ5f9U+tmcXEhrkz9V923Fj/wR6tWtpd3heRRO8cp+qSlj2d1jyHLRKtKZsmmp527SZ1ruSydOtNeeeKm8PvJwCg+7zjBU+tTDG8MOBo8xYvZ/ZGzIpragiuX0EP+kTR692LekWE67z9RdIE7xyv3RHnfniXGtbwAH3NK0pm4pS+GeKtWLmzsWeX6Lhkzth+xx4aBOERtbd3gMcO17Opz9kMGP1fnYdKjp1vF1EiKPUslVmuXtMOB0jQwnUSpdO0QSvGsfxfJh1L+ycC13HwdWvQGgbu6NyzspX4evH4eZZcNFIu6OpW+4OeGWAdQPa5b+1O5p6McaQXVBy1sYq27MK2ZNbRIWjzHGQvx+d24ZZ2yjGhp/aRzcqvJl+aHsOTfCq8Rhj1W6Z/xQEhULqY9aSSndtb+cKJQVWOeCY3nDL53ZH47yPboNdC+GhjZ6zVv8ClFVUsSe3iB2O2vrbswrZkV1IdkHJqTatQ4NOjfZ7xLSgW0w4XaPDCQmqZv9cH6EJXjW+nK3w9RPw42JofZE1bdPtSs+ctln8rLXD1Z2Lod0ldkfjvJwt1t22Ix6FkU/YHY3bHCkuO2vT9G3ZhezMLuREeSVg/ZPq2CbUmuJxbKPYPSac9q2a+0SRNE3wyh7GwK4FMP83cHintf3cmD9BbLLdkZ1WdMi6eajLaLj+33ZHU38f3AR7l8GvNkGw76xOqaoy7M8/fnrTdMdUz77845xMac2D/B1J30r8Jx97W2kFTfDKXpXl1i32i5+1tgfscxOMehJaxNodGXz1CHz/Fty3BiI72x1N/WVtgDeGw8gnYcQjdkdju+NlFezMKTq1jeLJXwBHj5efahPbMvjUaL996xBaBAfSMuTsrxYhgU1m4xNN8MoznDgKy56HVa+DfxAMfQgG3W/f3aL5e+HlftD3ZzDhBXticIUZN8CB1daKmqZ8H4KbGGM4VFjKNkfS35FdyLasAvbkFlFeWXP+C28WQItTCT/gvF8CJ38RVPe4MVcAaYJXniX/R1jwW9j2BYTHWatAel/f+EsTP7kDts2xShJ4wruJhspYB/8aBZc/Y62qUU4pr6wiv7iMYyfKra/j5RSUlJ/+3vFVcOL8YyXlVbX23TzIv/pfAqfeLQTQsvnp4/GtmhPdwB20NMErz7RvhfVBbOYPENcXxjwLHQY3zrWzNsIbwxzLDJ9pnGu607uTrOmahzZaq5eUW5VWVFJwoqLWXwLnHjv5fXFZ5Xn93T2iE4+P69GgWGpL8FocQtmnw2C44xvY9CEs/B1MHwc9JsLo30HrTu699qLfQ3AEDHnIvddpLCMehbevgLXTrRr2yq2aBfgTFe5PVHj9d7gqr6w6lewLSqxfEu0i3LP/rSZ4ZS8/P2sTix4TrX1Hv/s/2DkPBtwNwx6GkAjXXzP9O9i9AEb/3j392yFhAHQcDitegn4/984yzl4i0N+PNmHNGmX7Q70XWHmGoOYw4tfwwHprPn7Fy/BSX1jzJlRWuO46xljz/y3aWRUZvcmIR6EoB9b/x+5IlIdwa4IXkdYiMlpEmkaxDGW/FrHwk1fg7m8huid89bB1M8/O+eCKz4u2z4GDa607bL1tlJs4FBIGw3cvWLV1lM9zKsGLyFsislJEnnS2jYi0AuYA/YHFIhLlkoiVb4hNhltnw5QZUFUBM66Dd6+x7t5sqMoKWPQHiOwKyTe6LlZPMuIRKMyEH96zOxLlAepM8CIyCfA3xgwCOolIFyfbJAH/Y4z5E/A10ITuAVceQQS6XwX3roKxf7ZW27w+1NrVqOhQ/fvbMBMO74BRT4G/l3781GmktTn6dy9YN5gpn+bMCD4V+NDxeD4w1Jk2xphvjTGrRGQ41ih+5YWFqnxWQBAM/IW1Xr3/3dbo9KVLYNk/oLyk7vPB2nlqyXPQLgV6THBvvHYSgeG/hmP7YcMHdkejbOZMgg8FDjoe5wPVbSFTbRux6nreABwBzhtOiMhdIrJWRNbm5ubWM3Tlc5q3trapu3e1tWJk0e+sO1E3fVz3/PyaN6HgoLXm3RMLnrlSl9EQ28e6a9iVH1CrJseZBF8EnPw0KqyGc6ptYyz3ARuBieeeZIyZZoxJMcakREXpFL1yUmRn+OkMuOULCGkJn/wc3hoNB76vvv2Jo9auU50vh47DGjVUW4hYK5KOpFtb+ymf5UyCX8fpaZlkIN2ZNiLyqIjc4jgWARxtcJRKVafTCLjrW2tjkaP74a3L4ePbrcdnWvESlByFy562JUxbdLsSonvD0r9Z2yoqn+RMgp8F3Cwi/wCuB7aIyLkbQZ7b5ktgmuPYUsAfa25eKdfy87eKhT2w3pp73v6VtfXewt9ZG3kUZlu7NfWa7Fllit1NBIY/DHm7YctndkejbOJULRrHksfRwFJjTHZD29RGa9EolziWYZUh2PhfCI2ylkQeWA33f+/+8geepqoKXhsECPxihefvM6sapLZaNE79xI0xR4wxH9aWuJ1po5TbtYyHSdPgzm+gTWfYtxwunep7yR2shD78EcjdBttn2x2NsoFWk1TeyxjIWAuxSRDg/rofHqmqEl7pDwEhcM8y719B5IMueASvVJMkAu37+W5yB+szimEPQ84m2DHX7mhUI9MEr5S3630dtEq0Nhb3kHfsqnFoglfK2/kHwLD/haw02L3Q7mhUI9IEr5QvSJoCLRPgmz9YOz9V1b7lnHKjqkqrdEZJARTnQUGWdTOeG3hpxSWl1FkCguCyp+DTO+GN4dZuVolDrZIPicOgbQ/f+gC2tBAOrrfuk6gsg6pyqzhbZZnjq+KMx+WO58vOaHNO+6pz2tf2vKnml6ubto7UBK+Ur0i63krqe5dB+lLrz+1zrOdCo6znEodBxxHQ5iLvSfjGWDd8HVgDGd9bX4e2Vp9ozyR+4B/k+AoEv8DTj0/9ecbjoFBHm8Azzgs4uw//oOrbxLjnJjxdJqmULzuyD9KXWcl+71KrljxAeKwj2Q+zRvmtEm0Ns15KCuDgOmuJbIYjqZ84Yj3XrCXEXwrx/a0VVhGJ1rub8xJvoLUCqQnQTbeVUtVr1cH66vsza6Sb/6OV6PcuhR8XWxuigzV/fzLZJw6Dlu3sjfuks0bna6yCc4e2Ao6Ba1R36D7eqpHfvj9EdvOpO3p1BK+Uqp4xkLvDSvbpS63Nyk+OhFt3Op3sOw6HsLaNE9Op0fn3p6dcSo5azzVrCfEpjmTez6r97y2bqteithG8JnillHOqqiBns2NKZynsWwGlBdZzUd1PT+kkDrNq97vienm7T0+zVDc6Pzkyj+9v1R3yodH5SZrglVKuV1kB2RscUzrLYP9KKD8OCET3skb2HYdBh8EQ3LLu/koKrA3RM9bWPDpv399K6u0u9YnRuTM0wSul3K+iDDLXn16ls381VJZaq1Fikx1TOsMhYSAENj89Oj+ZzA9twxqdizU6b9/PSuY+PDp3hiZ4pVTjKy+xEvfJKZ2MtdZ6cr8ACAyF0mNWu+CW1nz5ydF5fIpzI34F6CoapZQdAoMdK2+GwcgnoKzYqs2/d6n1Ye3JpN6mi47O3UQTvFKqcQSFwkWjrC/VKPTXplJKeSlN8Eop5aU0wSullJfSBK+UUl5KE7xSSnkpTfBKKeWlNMErpZSX0gSvlFJeymNKFYhILrCvgadHAoddGE5Tp6/H2fT1OE1fi7N5w+vRwRgTVd0THpPgL4SIrK2pFoMv0tfjbPp6nKavxdm8/fXQKRqllPJSmuCVUspLeUuCn2Z3AB5GX4+z6etxmr4WZ/Pq18Mr5uCVUkqdz1tG8Eoppc6hCV4ppbxUk0/wIvKWiKwUkSftjsVuItJSROaKyHwR+UxEguyOyW4iEi0iP9gdh6cQkVdFZILdcdhNRFqJyFcislZE3rA7Hndp0gleRCYB/saYQUAnEelid0w2uwn4hzHmCiAbGGtzPJ7geSDE7iA8gYgMA2KMMbPtjsUD3Ay871gDHy4iXrkWvkkneCAV+NDxeD4w1L5Q7GeMedUYs8DxbRRwyM547CYio4BirF92Pk1EAoE3gXQRudrueDxAHtBLRCKA9sABe8Nxj6ae4EOBg47H+UC0jbF4DBEZBLQyxqyyOxa7OKanngIeszsWD3ELsBX4K9BfRB6wOR67fQd0AH4JbMPKH16nqSf4Ik6//Q6j6f99LpiItAb+Cdxudyw2ewx41Rhz1O5APERfYJoxJht4Dxhpczx2+y1wjzHm98B2YKrN8bhFU0+I6zg9LZMMpNsXiv0co9aPgMeNMQ0t3OYtLgfuE5ElQB8R+ZfN8dhtN9DJ8TiFhhf28xatgN4i4g8MALzyhqAmfaOTiLQAlgGLgHHAQGPMMXujso+I/AJ4FtjgOPSaMea/NobkEURkiTEm1e447CQi4cDbWNOYgcBkY8zB2s/yXiLSH5iONU2zErjGGFNkb1Su16QTPFjLnYDRwFLH20+llFJ4QYJXSilVvaY+B6+UUqoGmuCVUspLaYJXSikvpQleKaW8lCZ4pZTyUv8PE+Ph7ImHHbsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_loss(history_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "#保存模型\n",
    "model_save_path = \"model_file_path.h5\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'History' object has no attribute 'shape'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[1;32m~\\AppData\\Local\\Temp\\ipykernel_5380\\11448768.py\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mhistory_1\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mAttributeError\u001b[0m: 'History' object has no attribute 'shape'"
     ]
    }
   ],
   "source": [
    "history_1.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#算法一\n",
    "#predicted表示n个弱学习器对数据集第n+1的预测结果\n",
    "#y_test表示第n+1个数据集的实际值\n",
    "def algorithm_1(predicted,y_test): \n",
    "    #样本大小\n",
    "    sample_size = y_test.shape[0]\n",
    "    #初始化新数据集中的样本权重\n",
    "    w = np.zeros([sample_size])\n",
    "    w[:] = 1 / sample_size\n",
    "    #初始化a_k\n",
    "    a_k_all = []\n",
    "    #初始化a_k_final\n",
    "    a_k_final = []\n",
    "    \n",
    "    #对每一个弱学习器使用新数据集作为输入来输出预测\n",
    "    for i in range(len(predicted)):\n",
    "        E_k_max = 0\n",
    "        E_k = np.zeros(sample_size)\n",
    "        e_k = np.zeros(sample_size)\n",
    "        \n",
    "        \n",
    "        for j in range(sample_size):\n",
    "        #计算最大误差率数据集\n",
    "            E_k[j] = abs(y_test[j,0] - predicted[i][j,0])\n",
    "            \n",
    "        E_k_max = np.max(E_k)\n",
    "         #计算相关误差   \n",
    "        for j in range(sample_size):\n",
    "            e_k[j] = 1 - exp(abs(y_test[j,0] - predicted[i][j,0]))\n",
    "            \n",
    "        #计算回归误差\n",
    "        regression_error_rate = np.sum(w * e_k)\n",
    "        #计算标准误差率\n",
    "        beta_k = regression_error_rate / ( 1 - regression_error_rate)\n",
    "        \n",
    "        #计算弱学习器比率\n",
    "        a_k = 1/2 * log(1 / beta_k)\n",
    "        a_k_all.append(a_k)\n",
    "        #计算每个弱学习器的最终比率\n",
    "    for a_k in a_k_all:\n",
    "        a_k_final.append(a_k / np.sum(a_k_all))\n",
    "        \n",
    "    return a_k_final\n",
    "\n",
    "#算法一使用时须在外部调用保存的模型，使用函数返回的a_k_final列表生成强学习器\n",
    "\n",
    "    \n",
    "       "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# -*- coding: utf-8 -*-\n",
    "\"\"\"\n",
    "@Copyright (C) 2022 mewhaku . All Rights Reserved\n",
    "@Time ： 2022/8/8 11:19\n",
    "@Author ： mewhaku\n",
    "@File ：vfkmm.py\n",
    "@IDE ：PyCharm\n",
    "\"\"\"\n",
    "import numpy as np\n",
    "from sklearn import tree\n",
    "\n",
    "\n",
    "# 算法二\n",
    "# H 测试样本分类结果\n",
    "# train_old 原训练样本 np数组\n",
    "# train_new 辅助训练样本\n",
    "# label_old 原训练样本标签\n",
    "# label_new 辅助训练样本标签\n",
    "# 使用老数据集训练的弱学习器weak_learner\n",
    "# S_Test  测试样本\n",
    "# N 迭代次数\n",
    "def algorithm_2(train_old, train_new, label_old, label_new, S_test, weak_learner, N):\n",
    "    train_data = np.concatenate((train_new, train_old), axis=0)\n",
    "    train_label = np.concatenate((label_new, label_old), axis=0)\n",
    "\n",
    "    # 赋值训练集的行数\n",
    "    row_new = train_new.shape[0]\n",
    "    row_old = train_old.shape[0]\n",
    "    row_T = S_test.shape[0]\n",
    "\n",
    "    test_data = np.concatenate((train_data, S_test), axis=0)\n",
    "\n",
    "    # 初始化权重\n",
    "    #缺少veryfastKMM\n",
    "    #合并数据集train_old U train_new\n",
    "    beta = train_new.shape[0]\n",
    "    max_beta = train_old.shape[0]\n",
    "    # beta 和 max_beta由veryfastKMM得到\n",
    "    weight_1 = beta\n",
    "    weight_2 = np.max(beta)\n",
    "    weights = np.concatenate((weight_1, weight_2), axis=0)\n",
    "\n",
    "    # 设置rho值 #\n",
    "    rho = 1 / (1 + np.sqrt(2 * np.log(row_new / N)))\n",
    "\n",
    "    # 存储每次迭代的标签和bata值\n",
    "    bata_T = np.zeros([1, N])\n",
    "    result_label = np.ones([row_new + row_old + row_T, N])\n",
    "\n",
    "    predict = np.zeros([row_T])\n",
    "\n",
    "    print('params initial finished.')\n",
    "    train_data = np.asarray(train_data, order='C')\n",
    "    train_label = np.asarray(train_label, order='C')\n",
    "    test_data = np.asarray(test_data, order='C')\n",
    "\n",
    "    for i in range(N):\n",
    "        P = calculate_P(weights, train_label)\n",
    "\n",
    "        result_label[:, i] = train_classify(train_data, train_label,\n",
    "                                            test_data, P)\n",
    "        print('result,', result_label[:, i], row_new, row_old, i, result_label.shape)\n",
    "\n",
    "        error_rate = calculate_error_rate(label_old, result_label[row_new:row_new + row_old, i],\n",
    "                                          weights[row_new:row_new + row_old, :])\n",
    "        print('Error rate:', error_rate)\n",
    "        \n",
    "        if error_rate > 0.5:\n",
    "            error_rate = 0.5\n",
    "        if error_rate == 0:\n",
    "            N = i\n",
    "            break  # 防止过拟合\n",
    "            # error_rate = 0.001\n",
    "        \n",
    "        #计算bete_t\n",
    "        bata_T[0, i] = error_rate / (1 - error_rate)\n",
    "        \n",
    "        #权重更新\n",
    "        # 调整源域样本权重\n",
    "        for j in range(row_old):\n",
    "            weights[row_new + j] = weights[row_new + j] * np.power(bata_T[0, i],\n",
    "                                                               (-np.abs(result_label[row_new + j, i] - label_old[j])))\n",
    "\n",
    "        # 调整辅域样本权重\n",
    "        for j in range(row_new):\n",
    "            weights[j] = weights[j] * np.power(rho, np.abs(result_label[j, i] - label_new[j]))\n",
    "    # print bata_T\n",
    "    for i in range(row_T):\n",
    "        # 跳过训练数据的标签\n",
    "        left = np.sum(\n",
    "            result_label[row_new + row_old + i, int(np.ceil(N / 2)):N] * np.log(1 / bata_T[0, int(np.ceil(N / 2)):N]))\n",
    "        right = 0.5 * np.sum(np.log(1 / bata_T[0, int(np.ceil(N / 2)):N]))\n",
    "\n",
    "        if left >= right:\n",
    "            predict[i] = 1\n",
    "        else:\n",
    "            predict[i] = 0\n",
    "            # print left, right, predict[i]\n",
    "\n",
    "    return predict\n",
    "\n",
    "#计算p_t\n",
    "def calculate_P(weights, label):\n",
    "    total = np.sum(weights)\n",
    "    return np.asarray(weights / total, order='C')\n",
    "\n",
    "#训练分类器\n",
    "def train_classify(trans_data, trans_label, test_data, P):\n",
    "    clf = tree.DecisionTreeClassifier(criterion=\"gini\", max_features=\"log2\", splitter=\"random\")\n",
    "    clf.fit(trans_data, trans_label, sample_weight=P[:, 0])\n",
    "    return clf.predict(test_data)\n",
    "\n",
    "#输入用旧数据集训练的弱学习器\n",
    "def calculate_error_rate(row_new, y_test, predicted, weight):\n",
    "    E_k_max = 0\n",
    "    E_k = np.zeros(sample_size)\n",
    "    e_k = np.zeros(sample_size)\n",
    "    for j in range(row_new):\n",
    "        E_k[j] = abs(y_test[j,0] - predicted[j,0])\n",
    "            \n",
    "    E_k_max = np.max(E_k)\n",
    "    \n",
    "    for j in range(row_new):\n",
    "        e_i[j] = 1 - exp(abs(y_test[j,0] - predicted[j,0]))\n",
    "    regression_error_rate = np.sum(weight * e_i)\n",
    "    return regression_error_rate\n",
    "\n",
    "\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (tf37)",
   "language": "python",
   "name": "tensorflow37"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}