Bose_Hubbard.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "prerequisites:  \n",
    "tenpy: https://tenpy.github.io/INSTALL.html  \n",
    "tensorflow: https://www.tensorflow.org/install  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:09:14.363094Z",
     "start_time": "2020-01-23T21:09:10.012337Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:526: 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",
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:527: 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",
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:528: 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",
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:529: 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",
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:530: 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",
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:535: 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": [
    "from AD_tools import *\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "from matplotlib.ticker import MaxNLocator\n",
    "from matplotlib.colors import BoundaryNorm\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential,Model\n",
    "from tensorflow.keras.layers import Input,Dense,Conv2D,Conv1D,MaxPooling2D,MaxPooling1D,UpSampling2D,UpSampling1D,Flatten, Activation, BatchNormalization, GlobalAveragePooling2D,add"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:09:14.394227Z",
     "start_time": "2020-01-23T21:09:14.387107Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'1.13.1'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:09:14.428593Z",
     "start_time": "2020-01-23T21:09:14.414414Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'0.5.0'"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tenpy\n",
    "tenpy.__version__"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Create data\n",
    "Preferably run on a cluster / external session"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:09:15.677648Z",
     "start_time": "2020-01-23T21:09:15.673730Z"
    }
   },
   "outputs": [],
   "source": [
    "Vs = np.linspace(0,5,100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:16:27.860993Z",
     "start_time": "2020-01-23T21:11:46.084687Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initial sweeps...\n",
      "E = -6.0694700912326\n",
      "0:00:34.475932\n",
      "Initial sweeps...\n",
      "E = -5.7045767476760\n",
      "0:00:37.621923\n",
      "Initial sweeps...\n",
      "E = -5.3401870730464\n",
      "0:00:37.851126\n",
      "Initial sweeps...\n",
      "E = -4.9763110539437\n",
      "0:00:37.205504\n",
      "Initial sweeps...\n",
      "E = -4.6129590787029\n",
      "0:00:37.554460\n",
      "Initial sweeps...\n",
      "E = -4.2501419539892\n",
      "0:00:37.405192\n",
      "Initial sweeps...\n",
      "E = -3.8878709262272\n",
      "0:00:37.673483\n",
      "Initial sweeps...\n",
      "E = -3.5261577021729\n",
      "0:00:37.840000\n",
      "Initial sweeps...\n",
      "E = -3.1650144742022\n",
      "0:00:38.088353\n",
      "Initial sweeps...\n",
      "E = -2.8044539413091\n",
      "0:00:37.839484\n",
      "Initial sweeps...\n",
      "E = -2.4444893382511\n",
      "0:00:37.959309\n",
      "Initial sweeps...\n",
      "E = -2.0851344637237\n",
      "0:00:38.315691\n",
      "Initial sweeps...\n",
      "E = -1.7264037111817\n",
      "0:00:38.000469\n",
      "Initial sweeps...\n",
      "E = -1.3683121028578\n",
      "0:00:38.411420\n",
      "Initial sweeps...\n",
      "E = -1.0108753225437\n",
      "0:00:38.430523\n",
      "Initial sweeps...\n",
      "E = -0.6541097568645\n",
      "0:00:38.827047\n",
      "Initial sweeps...\n",
      "E = -0.2980325352883\n",
      "0:00:39.492909\n",
      "Initial sweeps...\n",
      "E = 0.0573384251874\n",
      "0:00:38.910516\n",
      "Initial sweeps...\n",
      "E = 0.4119843720527\n",
      "0:00:39.855499\n",
      "Initial sweeps...\n",
      "E = 0.7658856693623\n",
      "0:00:37.762056\n",
      "Initial sweeps...\n",
      "E = 1.1190217260989\n",
      "0:00:35.179285\n",
      "Initial sweeps...\n",
      "E = 1.4713709553556\n",
      "0:00:35.293258\n",
      "Initial sweeps...\n",
      "E = 1.8229107016486\n",
      "0:00:35.593302\n",
      "Initial sweeps...\n",
      "E = 2.1736171662276\n",
      "0:00:34.995346\n",
      "Initial sweeps...\n",
      "E = 2.5234653316776\n",
      "0:00:35.004444\n",
      "Initial sweeps...\n",
      "E = 2.8724288778859\n",
      "0:00:35.015555\n",
      "Initial sweeps...\n",
      "E = 3.2204800904312\n",
      "0:00:35.154190\n",
      "Initial sweeps...\n",
      "E = 3.5675897606866\n",
      "0:00:34.746841\n",
      "Initial sweeps...\n",
      "E = 3.9137270789963\n",
      "0:00:35.704218\n",
      "Initial sweeps...\n",
      "E = 4.2588595033188\n",
      "0:00:35.513564\n",
      "Initial sweeps...\n",
      "E = 4.6029526503319\n",
      "0:00:34.755173\n",
      "Initial sweeps...\n",
      "E = 4.9459701347927\n",
      "0:00:34.784427\n",
      "Initial sweeps...\n",
      "E = 5.2878734163018\n",
      "0:00:34.565569\n",
      "Initial sweeps...\n",
      "E = 5.6286216250363\n",
      "0:00:35.539459\n",
      "Initial sweeps...\n",
      "E = 5.9681713691372\n",
      "0:00:35.169126\n",
      "Initial sweeps...\n",
      "E = 6.3064765262177\n",
      "0:00:35.392163\n",
      "Initial sweeps...\n",
      "E = 6.6434879790931\n",
      "0:00:35.981783\n",
      "Initial sweeps...\n",
      "E = 6.9791534068839\n",
      "0:00:36.287465\n",
      "Initial sweeps...\n",
      "E = 7.3134169454724\n",
      "0:00:35.904424\n",
      "Initial sweeps...\n",
      "E = 7.6462188706989\n",
      "0:00:35.553029\n",
      "Initial sweeps...\n",
      "E = 7.9774952272915\n",
      "0:00:35.598280\n",
      "Initial sweeps...\n",
      "E = 8.3071774100441\n",
      "0:00:35.315713\n",
      "Initial sweeps...\n",
      "E = 8.6351916891429\n",
      "0:00:35.623402\n",
      "Initial sweeps...\n",
      "E = 8.9614586697557\n",
      "0:00:35.429244\n",
      "Initial sweeps...\n",
      "E = 9.2858926737046\n",
      "0:00:35.423358\n",
      "Initial sweeps...\n",
      "E = 9.6084010280864\n",
      "0:00:35.211367\n",
      "Initial sweeps...\n",
      "E = 9.9288832417793\n",
      "0:00:45.870729\n",
      "Initial sweeps...\n",
      "E = 10.2472300457704\n",
      "0:00:45.543036\n",
      "Initial sweeps...\n",
      "E = 10.5633222656905\n",
      "0:00:46.919503\n",
      "Initial sweeps...\n",
      "E = 10.8770294853768\n",
      "0:00:45.364529\n",
      "Initial sweeps...\n",
      "E = 11.1882084454388\n",
      "0:00:46.208655\n",
      "Initial sweeps...\n",
      "E = 11.4967011006182\n",
      "0:00:45.587084\n",
      "Initial sweeps...\n",
      "E = 11.8023318090155\n",
      "0:00:45.840069\n",
      "Initial sweeps...\n",
      "E = 12.1049056780509\n",
      "0:00:46.714058\n",
      "Initial sweeps...\n",
      "E = 12.4042056044397\n",
      "0:00:47.029159\n",
      "Initial sweeps...\n",
      "E = 12.6999778810123\n",
      "0:00:46.847452\n",
      "Initial sweeps...\n",
      "E = 12.9919399282032\n",
      "0:00:46.833720\n",
      "Initial sweeps...\n",
      "E = 13.2797607468738\n",
      "0:00:58.369606\n",
      "Initial sweeps...\n",
      "E = 13.5630521802472\n",
      "0:01:01.791726\n",
      "Initial sweeps...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tenpy/algorithms/dmrg.py:439: UserWarning: final DMRG state not in canonical form within `norm_tol` = 1.00e-05\n",
      "  warnings.warn(msg.format(nt=norm_tol))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "E = 13.8414568496390\n",
      "0:01:02.377193\n",
      "Initial sweeps...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tenpy/algorithms/dmrg.py:439: UserWarning: final DMRG state not in canonical form within `norm_tol` = 1.00e-05\n",
      "  warnings.warn(msg.format(nt=norm_tol))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "E = 14.1141439818695\n",
      "0:04:30.892020\n",
      "Initial sweeps...\n",
      "E = 14.3803194854456\n",
      "0:01:46.884816\n",
      "Initial sweeps...\n",
      "E = 14.6396359668920\n",
      "0:01:15.853833\n",
      "Initial sweeps...\n",
      "E = 14.8916018687893\n",
      "0:01:00.005118\n",
      "Initial sweeps...\n",
      "E = 15.1355689532774\n",
      "0:00:44.346408\n",
      "Initial sweeps...\n",
      "E = 15.3708079693096\n",
      "0:00:43.586511\n",
      "Initial sweeps...\n",
      "E = 15.5964847893283\n",
      "0:00:44.375457\n",
      "Initial sweeps...\n",
      "E = 15.8116041591080\n",
      "0:00:43.293440\n",
      "Initial sweeps...\n",
      "E = 16.0149094895726\n",
      "0:00:43.148042\n",
      "Initial sweeps...\n",
      "E = 16.2046641902431\n",
      "0:00:43.923758\n",
      "Initial sweeps...\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/export/anaconda3/envs/QC/lib/python3.7/site-packages/tenpy/algorithms/dmrg.py:439: UserWarning: final DMRG state not in canonical form within `norm_tol` = 1.00e-05\n",
      "  warnings.warn(msg.format(nt=norm_tol))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "E = 16.3779765275218\n",
      "0:01:30.226278\n",
      "Initial sweeps...\n",
      "E = 16.5276995780174\n",
      "0:00:44.043351\n",
      "Initial sweeps...\n",
      "E = 16.6584050291836\n",
      "0:00:33.302620\n",
      "Initial sweeps...\n",
      "E = 16.7762915749048\n",
      "0:00:33.535399\n",
      "Initial sweeps...\n",
      "E = 16.8840396524024\n",
      "0:00:32.656205\n",
      "Initial sweeps...\n",
      "E = 16.9833615341109\n",
      "0:00:31.425643\n",
      "Initial sweeps...\n",
      "E = 17.0754849104977\n",
      "0:00:30.946087\n",
      "Initial sweeps...\n",
      "E = 17.1613458581636\n",
      "0:00:30.849293\n",
      "Initial sweeps...\n",
      "E = 17.2416866916748\n",
      "0:00:30.438407\n",
      "Initial sweeps...\n",
      "E = 17.3171125327822\n",
      "0:00:30.125962\n",
      "Initial sweeps...\n",
      "E = 17.3881269109902\n",
      "0:00:29.222887\n",
      "Initial sweeps...\n",
      "E = 17.4551555719907\n",
      "0:00:28.585969\n",
      "Initial sweeps...\n",
      "E = 17.5185631397365\n",
      "0:00:26.986609\n",
      "Initial sweeps...\n",
      "E = 17.5786652018354\n",
      "0:00:26.606510\n",
      "Initial sweeps...\n",
      "E = 17.6357373309060\n",
      "0:00:25.328577\n",
      "Initial sweeps...\n",
      "E = 17.6900219795173\n",
      "0:00:24.553684\n",
      "Initial sweeps...\n",
      "E = 17.7417338558647\n",
      "0:00:23.425376\n",
      "Initial sweeps...\n",
      "E = 17.7910641857573\n",
      "0:00:22.618228\n",
      "Initial sweeps...\n",
      "E = 17.8381841410737\n",
      "0:00:21.871099\n",
      "Initial sweeps...\n",
      "E = 17.8832476326119\n",
      "0:00:21.473610\n",
      "Initial sweeps...\n",
      "E = 17.9263936104192\n",
      "0:00:21.211063\n",
      "Initial sweeps...\n",
      "E = 17.9677479770330\n",
      "0:00:20.608504\n",
      "Initial sweeps...\n",
      "E = 18.0074251926463\n",
      "0:00:20.218021\n",
      "Initial sweeps...\n",
      "E = 18.0455296323216\n",
      "0:00:20.259730\n",
      "Initial sweeps...\n",
      "E = 18.0821567416232\n",
      "0:00:20.262705\n",
      "Initial sweeps...\n",
      "E = 18.1173940268716\n",
      "0:00:20.137177\n",
      "Initial sweeps...\n",
      "E = 18.1513219086040\n",
      "0:00:19.776233\n",
      "Initial sweeps...\n",
      "E = 18.1840144610205\n",
      "0:00:18.996105\n",
      "Initial sweeps...\n",
      "E = 18.2155400557695\n",
      "0:00:18.962705\n",
      "Initial sweeps...\n",
      "E = 18.2459619249455\n",
      "0:00:18.911395\n"
     ]
    }
   ],
   "source": [
    "datas = []\n",
    "for V in Vs:\n",
    "    datas.append(hubbard_dmrg(L=8,U=5.,V=V,chi_max=64,bc=\"infinite\"))\n",
    "datas = np.array(datas)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:16:41.955207Z",
     "start_time": "2020-01-23T21:16:41.680307Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f54ebbc5ba8>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebd2d8d0>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebd2deb8>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebb08dd8>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebb08908>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebb082e8>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebb08198>,\n",
       " <matplotlib.lines.Line2D at 0x7f54ebb08518>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ent = np.array([psi.entanglement_entropy() for psi in datas[:,0]])\n",
    "fig,ax = plt.subplots(figsize=(8,5))\n",
    "ax.plot(Vs,ent)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:07.149431Z",
     "start_time": "2020-01-23T21:17:07.137651Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(100, 64, 1)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_test = np.array([psi.entanglement_spectrum(by_charge=False)[int(psi.L/2)-1] for psi in datas[:,0]])\n",
    "x_test = np.sqrt(np.exp(-1*x_test)) # Go from entanglement_spectrum to singular values l_v^2 = exp(-s_v)\n",
    "x_test = x_test.reshape(*x_test.shape,1) # Need extra \"color\" channel (though trivial) for tensorflow later\n",
    "x_test.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Define CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:10.164563Z",
     "start_time": "2020-01-23T21:17:10.154149Z"
    }
   },
   "outputs": [],
   "source": [
    "def scc_cnn1D(loss,optimizer,activation0,activation,x_shape):\n",
    "    in_dim = x_shape\n",
    "    input_img = Input(shape=in_dim)\n",
    "   \n",
    "    x1 = Conv1D(64, 3, activation=activation0, padding='same')(input_img) # 100 100 64\n",
    "    x2 = MaxPooling1D(2, padding='same')(x1) # 50 50 64\n",
    "    x2 = Conv1D(64, 3, activation=activation0, padding='same')(x2) # 50 50 128\n",
    "    x3 = MaxPooling1D(2, padding='same')(x2) # 25 25 128\n",
    "    encoded = Conv1D(64, 3, activation=activation0, padding='same')(x3) # 25 25 128\n",
    "\n",
    "    y=UpSampling1D(2)(encoded) # 50 50 128\n",
    "    y=add([x2,y]) # 50 50 128\n",
    "    y=Conv1D(64, 3, activation=activation0, padding='same')(y) # 50 50 128\n",
    "    y=UpSampling1D(2)(y) # 100 100 128\n",
    "    y=add([x1,y]) # 100 100 128\n",
    "    decoded = Conv1D(x_shape[-1], 3, activation=activation, padding='same')(y) # 100 100 4\n",
    "\n",
    "    cnn = Model(input_img, decoded)\n",
    "\n",
    "    cnn.compile(loss=loss,optimizer=optimizer)#,metrics=['accuracy']) #adadelta\n",
    "    return cnn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:13.101167Z",
     "start_time": "2020-01-23T21:17:12.679784Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "WARNING:tensorflow:From /export/anaconda3/envs/QC/lib/python3.7/site-packages/tensorflow/python/keras/utils/losses_utils.py:170: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "input_1 (InputLayer)            (None, 48, 1)        0                                            \n",
      "__________________________________________________________________________________________________\n",
      "conv1d (Conv1D)                 (None, 48, 64)       256         input_1[0][0]                    \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling1d (MaxPooling1D)    (None, 24, 64)       0           conv1d[0][0]                     \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_1 (Conv1D)               (None, 24, 64)       12352       max_pooling1d[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "max_pooling1d_1 (MaxPooling1D)  (None, 12, 64)       0           conv1d_1[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_2 (Conv1D)               (None, 12, 64)       12352       max_pooling1d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "up_sampling1d (UpSampling1D)    (None, 24, 64)       0           conv1d_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "add (Add)                       (None, 24, 64)       0           conv1d_1[0][0]                   \n",
      "                                                                 up_sampling1d[0][0]              \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_3 (Conv1D)               (None, 24, 64)       12352       add[0][0]                        \n",
      "__________________________________________________________________________________________________\n",
      "up_sampling1d_1 (UpSampling1D)  (None, 48, 64)       0           conv1d_3[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "add_1 (Add)                     (None, 48, 64)       0           conv1d[0][0]                     \n",
      "                                                                 up_sampling1d_1[0][0]            \n",
      "__________________________________________________________________________________________________\n",
      "conv1d_4 (Conv1D)               (None, 48, 1)        193         add_1[0][0]                      \n",
      "==================================================================================================\n",
      "Total params: 37,505\n",
      "Trainable params: 37,505\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "cnn = scc_cnn1D(\"mse\",\"adam\",\"relu\",\"relu\",(48,1,))\n",
    "cnn.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Training\n",
    "pick data points from test set for training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:17.228217Z",
     "start_time": "2020-01-23T21:17:17.224375Z"
    }
   },
   "outputs": [],
   "source": [
    "Vs_train = Vs[np.where(Vs<2)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:17.987837Z",
     "start_time": "2020-01-23T21:17:17.972312Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(40, 64, 1)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train = x_test[np.where(Vs<2)]\n",
    "x_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:32.798654Z",
     "start_time": "2020-01-23T21:17:32.794419Z"
    }
   },
   "outputs": [],
   "source": [
    "def train_loss2(y_true,y_pred):\n",
    "    return tf.keras.backend.sqrt(tf.keras.backend.sum((y_true - y_pred)**2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-01-23T21:17:38.783156Z",
     "start_time": "2020-01-23T21:17:38.766985Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 40 samples, validate on 40 samples\n",
      "Epoch 1/40\n",
      "40/40 [==============================] - 1s 19ms/sample - loss: 4.4280 - val_loss: 3.7862\n",
      "Epoch 2/40\n",
      "40/40 [==============================] - 0s 2ms/sample - loss: 3.6686 - val_loss: 3.1398\n",
      "Epoch 3/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 3.0370 - val_loss: 2.5701\n",
      "Epoch 4/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 2.4771 - val_loss: 2.0722\n",
      "Epoch 5/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 2.0018 - val_loss: 1.6990\n",
      "Epoch 6/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 1.6550 - val_loss: 1.4503\n",
      "Epoch 7/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 1.4179 - val_loss: 1.2901\n",
      "Epoch 8/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 1.2580 - val_loss: 1.1077\n",
      "Epoch 9/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 1.0732 - val_loss: 0.9224\n",
      "Epoch 10/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.9013 - val_loss: 0.8230\n",
      "Epoch 11/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.8095 - val_loss: 0.7504\n",
      "Epoch 12/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.7266 - val_loss: 0.6220\n",
      "Epoch 13/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.5944 - val_loss: 0.5114\n",
      "Epoch 14/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.5047 - val_loss: 0.4753\n",
      "Epoch 15/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.4686 - val_loss: 0.4139\n",
      "Epoch 16/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.4137 - val_loss: 0.3798\n",
      "Epoch 17/40\n",
      "40/40 [==============================] - 0s 988us/sample - loss: 0.3699 - val_loss: 0.3322\n",
      "Epoch 18/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.3343 - val_loss: 0.2802\n",
      "Epoch 19/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.2718 - val_loss: 0.2865\n",
      "Epoch 20/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.2884 - val_loss: 0.2386\n",
      "Epoch 21/40\n",
      "40/40 [==============================] - 0s 992us/sample - loss: 0.2504 - val_loss: 0.2376\n",
      "Epoch 22/40\n",
      "40/40 [==============================] - 0s 982us/sample - loss: 0.2237 - val_loss: 0.1935\n",
      "Epoch 23/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.2112 - val_loss: 0.1893\n",
      "Epoch 24/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.1847 - val_loss: 0.1804\n",
      "Epoch 25/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.1843 - val_loss: 0.1429\n",
      "Epoch 26/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.1381 - val_loss: 0.1870\n",
      "Epoch 27/40\n",
      "40/40 [==============================] - 0s 999us/sample - loss: 0.1969 - val_loss: 0.1272\n",
      "Epoch 28/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.1815 - val_loss: 0.1672\n",
      "Epoch 29/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.2398 - val_loss: 0.4108\n",
      "Epoch 30/40\n",
      "40/40 [==============================] - 0s 998us/sample - loss: 0.3062 - val_loss: 0.3475\n",
      "Epoch 31/40\n",
      "40/40 [==============================] - 0s 1ms/sample - loss: 0.3125 - val_loss: 0.2112\n",
      "Epoch 32/40\n",
      "40/40 [==============================] - 0s 980us/sample - loss: 0.2145 - val_loss: 0.1194\n",
      "Epoch 33/40\n",
      "40/40 [==============================] - 0s 910us/sample - loss: 0.1300 - val_loss: 0.0993\n",
      "Epoch 34/40\n",
      "40/40 [==============================] - 0s 922us/sample - loss: 0.1032 - val_loss: 0.1300\n",
      "Epoch 35/40\n",
      "40/40 [==============================] - 0s 931us/sample - loss: 0.1185 - val_loss: 0.1076\n",
      "Epoch 36/40\n",
      "40/40 [==============================] - 0s 899us/sample - loss: 0.1150 - val_loss: 0.0961\n",
      "Epoch 37/40\n",
      "40/40 [==============================] - 0s 951us/sample - loss: 0.1162 - val_loss: 0.1110\n",
      "Epoch 38/40\n",
      "40/40 [==============================] - 0s 939us/sample - loss: 0.1406 - val_loss: 0.1625\n",
      "Epoch 39/40\n",
      "40/40 [==============================] - 0s 905us/sample - loss: 0.1710 - val_loss: 0.2302\n",
      "Epoch 40/40\n",
      "40/40 [==============================] - 0s 922us/sample - loss: 0.1893 - val_loss: 0.1717\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cnn0 = training(x_train, scc_cnn1D, name = \"\", provide_cnn = False,\n",
    "             load_prev = False, num_epochs = 40,  verbose_val = 1, batch_size = 24, shuffle = True, early = False,\n",
    "             loss = train_loss2, activation0 = 'relu', activation = 'tanh', optimizer = \"adam\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Expected Output:**\n",
    "\n",
    "The Output should look similar to this\n",
    "\n",
    "<img src=\"Images/Learning_curve.png\">"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "xi = np.array([psi.correlation_length() for psi in datas[:,0]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f54e831c438>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeAAAAEvCAYAAACdahL0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nOzdd1iUV/bA8e+dYQBBrNgrosaCCIpdsURN0RTXkmjqpu+uqbvZ5Jeym90ku9n0YvrGmKLGaJrJZrOaREWDsfcSFcFYEAEF6Uy5vz9ehjp0hpnB83keHpx33nnnYMwc7n3PvUdprRFCCCFE4zJ5OgAhhBDiQiQJWAghhPAAScBCCCGEB0gCFkIIITxAErAQQgjhAZKAhRBCCA/wa8w3Cw0N1T179mzMtxRCCCE8Ztu2bWla63aunmvUBNyzZ0+2bt3amG8phBBCeIxS6lhlz8kUtBBCCOEBkoCFEEIID5AELIQQQnhAo94DdsVqtXLixAny8/M9HYrwcYGBgXTt2hWLxeLpUIQQoloeT8AnTpwgJCSEnj17opTydDjCR2mtSU9P58SJE4SFhXk6HCGEqJbHp6Dz8/Np27atJF9RL0op2rZtKzMpQgif4fEEDEjyFQ1C/h0JIXyJVyRgX/Pll1+yf/9+t1x7woQJHlsrvWjRIk6dOtUg10pKSmLJkiUNci0hGsSGlyExruyxxDjjeA28tS6B+IS0MsfiE9J4a11CQ0UoLjCSgOvAnQnYkxorAdtstgZ5DyFqpcsQWH5zSRJOjDMedxlSo5dHdm3J/CU7ipNwfEIa85fsILJrS/fEK5o8n0rA7voN9OOPP2b48OFERUVx5513YrfbAWjevDmPPvoogwcPZuTIkaSkpBAfH8/KlSt58MEHiYqKIiEhgXfffZdhw4YxePBgZs6cSW5uLgA333wz99xzD6NHj6ZXr16sWLECAIfDwe9//3sGDhzI9OnTufzyy4ufK23VqlWMGjWKIUOGMHv2bLKzswFjR7FHHnmEUaNGERMTw/bt27nkkksIDw/nrbfeKn79c889x7Bhw4iMjOSvf/0rYCTG/v37c/vttzNw4ECmTp1KXl4eK1asYOvWrVx33XVERUWRl5dXJpaEhAQuvfRShg4dyrhx4zh48GCVP+PDDz/M+vXriYqK4qWXXmLRokXMnj2bK664gqlTp6K15sEHHyQiIoJBgwaxbNkyANauXUtsbCwzZsxgwIAB3HXXXTgcDt577z3uv//+4njeffddHnjggXr9dxcXmLBYmL3ISLo/Pg1LroGxDxjHnaoYEY8OD2XBvGh+v3g7D3+2m/lLdrBgXjSjw0MbJXzRBGmtG+1r6NChurz9+/dXOFaZn46k6ui/r9I/HUl1+bgu9u/fr6dPn64LCwu11lr/7ne/0x988IHWWmtAr1y5Umut9YMPPqiffPJJrbXWN910k16+fHnxNdLS0or//Oijj+pXX321+LxZs2Zpu92u9+3bp8PDw7XWWi9fvlxfdtll2m636+TkZN2qVavi640fP15v2bJFp6am6nHjxuns7GyttdbPPPOM/tvf/qa11rpHjx76jTfe0Fprfd999+lBgwbp8+fP6zNnzuh27dpprbX+3//+p2+//XbtcDi03W7X06ZN0+vWrdOJiYnabDbrHTt2aK21nj17tv7oo4/KvLcrkyZN0ocOHdJaa/3zzz/riRMnVvkzrlmzRk+bNq349e+//77u0qWLTk9P11prvWLFCj158mRts9n06dOndbdu3fSpU6f0mjVrdEBAgE5ISNA2m01PnjxZL1++XGdnZ+tevXoV/3caNWqU3r17t8v/nkJorbVe/5LWR9eVPM5O0/rTm7R+f5rWf22h9YIRWj/VUetdy4znj67T+l9hZV/jwrCnVuseD32jX/jfQffFLpoMYKuuJCd6fBlSede8vbHCsemRnbhhVE+iu7WmfUgAN763mQ4tAkg5X0Dv9s05ec4YrZ3NKeR3H28r89pld46q8v1++OEHtm3bxrBhwwDIy8ujffv2APj7+zN9+nQAhg4dyurVq11eY+/evTz22GNkZGSQnZ3NJZdcUvzc1VdfjclkYsCAAaSkpACwYcMGZs+ejclkomPHjkycOLHCNX/++Wf279/PmDFjACgsLGTUqJKf5corrwRg0KBBZGdnExISQkhICIGBgWRkZLBq1SpWrVpFdHQ0ANnZ2Rw+fJju3bsTFhZGVFRU8c+VlJRU5d9RdnY28fHxzJ49u/hYQUFBlT+jK1OmTKFNmzbFfwdz587FbDbToUMHxo8fz5YtW2jRogXDhw+nV69eAMydO5cNGzYwa9YsJk2axDfffEP//v2xWq0MGjSoyrjFBc455Tx7EVjz4PM7ID8DLEEQ+6Ax0nVY4fPbYcNLkHUa5nxQdkRcTnxCGufzrPiZFB9v+pWR4W1lBCzqzOsScHVaNrPQoUUAJzPy6dIqkJbN6rfpgtaam266iX/+858VnrNYLMWVtWazudJ7lzfffDNffvklgwcPZtGiRaxdu7b4uYCAgDLvVfp7dXFNmTKFpUuXunzeeV2TyVTmPUwmEzabDa01//d//8edd95Z5nVJSUllzjebzRWmm8tzOBy0atWKnTt3VhmLM+7KBAcH1+i88tXMzse33XYb//jHP+jXrx+//e1vq4xZCMJi4arXYfEssBUAJvAPhrmfGM+FxcKnN0LzjnBmP1iCISe17DUS4+Dkdhh7X/E930sGduTbvcksmBct09CiXrzuHvCyO0dV+LphVE8AmvmbuXdyH/KsDu6Z1Js8q4N7J/dhdkw3ANoE+1d4bXUuvvhiVqxYwZkzZwA4e/Ysx45V2rwCgJCQELKysoofZ2Vl0alTJ6xWK4sXL672PceOHctnn32Gw+EgJSWlTMJ2GjlyJD/99BNHjhwBIDc3l0OHDlV7badLLrmEhQsXFt83PnnyZPHPWNOfy6lFixaEhYWxfPlywEieu3btqtO1nGJjY1m2bBl2u53U1FTi4uIYPnw4AJs3byYxMRGHw8GyZcsYO3YsACNGjOD48eMsWbKEuXPnVvn+4gLkqso57vmi5Av0HFOSfMH4Pu5PkHEMImaCNRdW3lNpkdbuE5ksmBfNV7tOYbVrRvVqy4J50ew+kdk4P59ocrwuAVfF+RvognnRPDD1ouLfQMsXZtXGgAEDeOqpp5g6dSqRkZFMmTKF5OTkKl9z7bXX8txzzxEdHU1CQgJPPvkkI0aMYMqUKfTr16/a95w5cyZdu3YlIiKCO++8kxEjRtCyZdlKynbt2rFo0SLmzp1LZGQkI0eOLC58qompU6cyb948Ro0axaBBg5g1a1aVCRGMkfxdd93lsghr8eLFvPfeewwePJiBAwfy1VdfVXmtyMhI/Pz8GDx4MC+99FKF52fMmEFkZCSDBw9m0qRJPPvss3Ts2BGAUaNG8fDDDxMREUFYWBgzZswoft2cOXMYM2YMrVu3rulfhbhQOKec934GuWeNBHrmgDHqjf2zMcotLTEONrwI85bBrIVwwxegTPDJdfD5nSXT10UJ+67x4WVGujaHZnR4KHeND2+0H1E0MZXdHHbHV32LsN5ce6RCwdVPR1L1m2uP1Pga3iIrK0trbRRw9erVSycnJ3s4Iu9QvnirvGnTpunvv/++0uelCOsCUb7ASmutD/+g9TsTjQKrt8Zp/Y+uxpfzvPJFVq6ucXSd1q/FGNdYcavLtx7x9Pe6x0PfaKvN3sA/lGiK8KUirKq4+k1zdHioT95/mT59OhkZGRQWFvL4448Xj/6EaxkZGQwfPpzBgwdz8cUXezoc4WmlC6zCYuGnV+H7J0DboU04JO+CsPEQ+6eyU86zFxn3dMNiYex9rq+dkwbB7WDPCug1CaLnlXm6b8cQOrQMxM/sUxOIwgv5VAJuSlzd9xXGTmATJkyocLxVq1a1ugcumrjSa3o7RsLRNUYx1Yg7YOPrxpTz1vdcv66yKmfnPd85H0DrnrBgGHz1ewhsCf2nFZ8Te+YzXsi5DKvdgUWSsKgH+dcjhPB+5Qus8s8bFcsdIozk23OcUfG88XUjMU96tCRBly/MqszJ7SUj6lbdYcqTgIb/PABaFyfoH853Ic9qJz27sMF/THFhkRGwEML7OaecZy6EzF9h1eNQkFVSYLX1PTj4dZmiqQpTztUpPyU94g7IOQOb3oY1/zDeY/YiZqSHsXHFbqx2RwP/kOJCIwlYCOF9NrxsJF1n4uwxFnpPho+uBrRRrWxpBtcuLppWHmck6IiZZa9T1ZRzTUx6DFAQ96yR6MNi8c84CRhV0ELUh0xBCyG8T/nGCd89BLuXQUCI8bjn2Iprep2j3YaUGGeMfDsOgp9fh8Q47ltmbEgjI2BRX9UmYKVUoFJqs1Jql1Jqn1Lqb0XHn1BKnVRK7Sz6utz94QohLgjOhPrpTUbjhD0rYNA1YLYYI9GUfa5fU1llc104i7JmLwKTn1EZvfxmRpmM95YELOqrJiPgAmCS1nowEAVcqpQaWfTcS1rrqKKvb90WpZepqm3fhAkTuOiii1i5ciUADz30EJGRkdx4443F53z00Ue88sorxY/Xr1/PgAEDiIiIcG/gQviStENgtxrTv32mQML3dS+wqovSRVlR18G5JJj0GJHqKP5+Jrq0aua+9xYXhGoTcNFa4uyih5aiL8/c/KhnQ+2GUl3f3MWLF3PllVeSmZlJfHw8u3fvxm63s2fPHvLy8li0aBG///3vi88fN24c3357wfz+IkT1jq6F/zwI9nxju8gDX5dtHeiuKefSxt5X8n6DZoE5AM4cYEXgTGYP7UqrIH/3vbe4INToHrBSyqyU2gmcAVZrrTcVPTVfKbVbKbVQKeX+vQHr2VC7Mi+++CIRERFERETw8stGMk9KSiozIn3++ed54oknqu2bW5rJZKKwsBCtNXl5eVgsFp577jnuueceLJb6NZEQokkp/ct12hFYOg8UEDEbLn7c2C5yw4tlfwFv6CnnqjRrDf2mwZ7lZOXksOPXDDJyZRmSqJ8aVUFrre1AlFKqFfCFUioCeBMoWijHk8ALwC3lX6uUugO4A6B79+7Vv9n70yoeG3g1DL8dusRASCf4aIbxPSsZ2vWDjOPGeTnpRneT0n77nyrfbtu2bbz//vts2rQJrTUjRoxg/Pjxle41PGvWLBYsWMDzzz9PTExMldcOCQlh5syZREdHc/HFF9OyZUu2bNnCX/7ylypfJ8QFx/nL9ZWvweq/gMNmtA2MKmq6UdslRe4QcwsEh9J+s539yefZfSKT2L7tPBOLaBJqVQWttc4A1gKXaq1TtNZ2rbUDeBcYXslr3tFax2itY9q1a4B/rIGtjOSbedz4HtiqXpfbsGEDM2bMIDg4mObNm/Ob3/yG9evX1z/OIn/+85/ZuXMnL7zwAo8//jh///vf+fe//82cOXN46qmnGux9hPBpzgT71XzQDvALgLlLyybbxhzxuhI2Di5/jtdvNbZCtTmkCEvUT7UjYKVUO8Cqtc5QSjUDJgP/Ukp10lo72wbNAPY2SERVjVj9g2DCQ8Zvys7F9xMeKvmfNLhttSPe8nQlfWn9/PxwlPofLD8/v1bXLW/Hjh0A9O3bl3vvvZe4uDiuvfZaDh8+TJ8+fep1bSGahLBYGHZbmTW3XmXDy9A5mhbnCmhLJoU2XaZfsBC1VZMRcCdgjVJqN7AF4x7wN8CzSqk9RccnAve7MU5D6WUBDVQJGRsby5dffklubi45OTl88cUXjBs3jg4dOnDmzBnS09MpKCjgm2++KX5Ndb1uXXGOfq1WK3a7HTDuEefm5tY5diGalMQ42PJvGHW38cu1Oyuc66LLEPSnNxL29WxmmeNombKxQWpQxIWr2hGw1no3EO3i+A1uiagqpZcFQIPcFxoyZAg333xzcTP42267jeho48f9y1/+wogRIwgLCyvT59fZN7dZs2Zs3LiRZs2qXo7w5ZdfMmzYMDp37gxQ3KPX2Q9XiAue85fr0D7G3s7OX65L///uaWGx5M14H8uSWdzu9w0hm76DuR96T3zC56jKpmDdISYmRm/durXMsQMHDtC/f/9Gi8HdJkyYUKMCrfKSkpKYPn06e/c2zEz+haqp/Xu6YDi3nox7Hmz5cOsqr5zeTc8uIOnZ0Qw1HSFrxP2EXPaEp0MSXk4ptU1r7TIhyFaUDaxNmzbcfPPNxRtx1MT69eu54oorCA31vb7GQjQI55pba65R/QyeL7pyQSWtp7/6FYCQPR963zS58CnSjKGBff7557V+zbhx49izZ48bohHCxxTmQvMOno7CtcQ4Wn5zO5/Yx3Kd34+cufgV2nvbNLnwKTICFkJ4D2tOyQjY25zcTub0d4nrfCvTCp7m65x+7t+NSzRpMgIWQniPcX801vd7o7H30QZ4ua+d/n/5jukOU/3bHYoLmiRgIYT3GHJj9ed4mF/Wca4zf49ffnsg3NPhCB92wU9BZ2Rk8MYbb9TptZdffjkZGRl1fu/mzZtX+Xx9YhPC52gNKfsh96ynI6nUzuMZPPj6Up62LCQ494SnwxE+ThJwFUnOuWFGZb799ltatarfVphVkQQsLijWPHhzFGxb5OlIKpVXaCcl32w8sMomOqJ+fCoBL9y7kM3Jm8sc25y8mYV7F9b5mg8//DAJCQlERUXx4IMPsnbtWiZOnMi8efMYNGgQAFdffTVDhw5l4MCBvPPOO8Wv7dmzJ2lpaSQlJdG/f39uv/12Bg4cyNSpU112SUpMTGTUqFEMGzaMxx9/vPh4dnY2F198MUOGDGHQoEF89dVXLmOr7DwhmgRnQvMP9mwcVbDaHeTqAAAmh4d4OBrh87TWjfY1dOhQXd7+/fsrHKvMplOb9Lil4/SmU5tcPq6LxMREPXDgwOLHa9as0UFBQfro0aPFx9LT07XWWufm5uqBAwfqtLQ0rbXWPXr00KmpqToxMVGbzWa9Y8cOrbXWs2fP1h999FGF97riiiv0Bx98oLXWesGCBTo4OFhrrbXVatWZmZlaa61TU1N1eHi4djgcFWKr7DxRojb/noSXOZuk9V9baL3tQ09HUqnV+07rKQ+/acS55zNPhyN8ALBVV5ITfWoEPLzTcJ4f/zx/WvcnFuxYwJ/W/Ynnxz/P8E4uGzHV/X2GDycsLKz48auvvsrgwYMZOXIkx48f5/DhwxVeExYWRlRUFABDhw4lKSmpwjk//fQTc+ca7dVuuKFkJ0+tNY888giRkZFMnjyZkydPkpKSUuH1NT1PCJ/kKyNgjBHwsZQ0D0cjfJ3PVUEP7zScORfN4e3db3Nn5J0NnnwBgoNLPgDWrl3L999/z8aNGwkKCmLChAkuOyMFBAQU/9lsNrucggZQSlU4tnjxYlJTU9m2bRsWi4WePXu6fI+anieETyr0/gTcsWUgwyMjmLT7ecakDeJJTwckfJpPjYDBuOf76S+fcmfknXz6y6cV7gnXVnWdjTIzM2ndujVBQUEcPHiQn3/+uc7vNWbMGD755BPASKal36N9+/ZYLBbWrFnDsWPHXMZW2XlCNAmtusOVr0GHCE9HUqno7q15ce4wCluFk0Ogp8MRPs6nEvDm5M3F087zo+cXT0fXJwm3bduWMWPGEBERwYMPPljh+UsvvRSbzUZkZCSPP/44I0eOrPN7vfLKK7z++usMGzaMzMzM4uPXXXcdW7duJSYmhsWLFxd3XiofW2XnCdEkNG9nrANu2cXTkVRrnv1remTv8HQYwsf5VDekhXsXEtE2osy08+bkzexN38stEbc0aKzCN0k3JB+WnQoZx4wRsMU7R5cf/XyMp/+zn53mG4lreTVT73/X0yEJL1dVNySfugfsKskO7zTcLfeBhRCN7NB3sHI+3LfHmI72QgVWO/lWB/nmQCwO13UeQtSUT01BCyGaMGcVtMV7i7AK7Q4AgpqHMKyzd47She/wqRGwEKIJK8wxvvt7aTckwGozbtlZApvjby70cDTC13nFCLgx70OLpkv+Hfk4ay6gwM97R5ZWuwOzSZHl8Cf17DlPhyN8nMdHwIGBgaSnp9O2bVuXa2SFqAmtNenp6QQGeu+Ht6hGYa6xBtiLPwciu7bk+hHdeTjlcc4VmFjq6YCET/N4Au7atSsnTpwgNTXV06EIHxcYGEjXrl09HYaoq6h50GOUp6Oo0tSBHZk6sCO3LMojq0A2wRH14/EEbLFYymz7KIS4QHWMML68mM3uwKQUI/LiUDkngHGeDkn4MI8nYCGEAODkNtBA16GejqRSj3yxh7hDabwcuImwgu3AS54OSfgwScBCCO/ww9+NSujbvvd0JJWy2jX+fiZspkACtExBi/qptgpaKRWolNqslNqllNqnlPpb0fE2SqnVSqnDRd9buz9cIUST5SzC8mKFNgcWsyI6vAst/KyeDkf4uJosQyoAJmmtBwNRwKVKqZHAw8APWus+wA9Fj4UQom6suV69CQcYG3FYzCaCm7fAZMsHh8PTIQkfVm0CLuopnF300FL0pYGrgA+Kjn8AXO2WCIUQF4bCHK/ehAOMdcABfiaSsooSr022oxR1V6N7wEopM7AN6A28rrXepJTqoLVOBtBaJyul2rsxTiFEU1eYAxbvTsCXR3Qi32ZnSeo0PnVEstPL4xXerUYJWGttB6KUUq2AL5RSNV4roJS6A7gDoHt379xgXQjhBWYvgqA2no6iSnOGdQPgmf8eJNfh59WbhgjvV6utKLXWGcBa4FIgRSnVCaDo+5lKXvOO1jpGax3Trl27eoYrhGiyeo6B9t7dSjIzz0puoY1O+Ud4TL0H5095OiThw2pSBd2uaOSLUqoZMBk4CKwEbio67SbgK3cFKYRo4uxW2LMC0hM8HUmV5r37M/cs3Ukrayo3mldjz5QELOquJiPgTsAapdRuYAuwWmv9DfAMMEUpdRiYUvRYCCFqL/88fHYrHPHeNcBgLEPy91NoSzMA7PnZ1bxCiMpVew9Ya70biHZxPB242B1BCSEuMNaiVoReXtRkLVqGNDWqF+wCi0M24xB15xXtCIUQF7jCXOO71y9D0ljMJoKCQwBQ1lwPRyR8mSRgIYTnFY+AvXsjjgKbMQI+mG7HgZm8fFkHLOpOErAQwvN8ZAR81/heTB3Qgd1ZLeiV/xFne//G0yEJHyYJWAjheZ0Gw20/QKcoT0dSpdvG9WJiv/b4+RkfnVabbEUp6k4SsBDC8wJbQNcY47sXO342l/P5VvzMJp7xe4eAQys9HZLwYZKAhRCed+YgbP/Q2I7SS2mtGffsGt5bn4i/WXGleSOW0zs8HZbwYZKAhRCelxgHK+8Gq/cWNRXajelmfz8TfiYTefiXFI8JUQc12gtaCCHcygfWAVvtGgCLWTH+onaYW7VCWWwejkr4MknAQgjPK8wBFBTtMOWNnAVXFrMJi9kE/sFGD2Mh6kgSsBDC8wpzjdGvF3cXspaagj5+NhdbfjChNj9CPByX8F1yD1gI4XnWHK9fAxwc4McTVwwgpkcb0rILmJj6R7YO/ZenwxI+TEbAQgjPm/gYjJrv6SiqFBzgx81jwgDYezITKBkVC1EXMgIWQnhe83YQ2sfTUVQp32rnl9NZZBfYsJhN3Gr+lou2/tXTYQkfJglYCOF5e1bAvi88HUWVjpzJ5pKX44g/koafWdHf9CvtTq/zdFjCh8kUtBDC8za/A36BMHCGpyOpVOl1wBaTiVwdgNku7QhF3ckIWAjheYW5xrIeL+ZchuRvNtGtTTNuiO1PgPQDFvUgCVgI4XnWHK/ehANKbcThZ0IphbIEgS0PHFKIJepGErAQwvMKc71+GZKz4tliNpFXaOeLw1bymvcAe4GHIxO+ShKwEMLzCnPA4t1T0Bd1DOG5WZF0bxOEQ2vuPzqUD4d94dW7dwnvJkVYQgjPu2cHmL3746hzq2bMjukGQIHNDoDNoT0ZkvBxMgIWQnhe83bQrLWno6hSalYB246do8Bmx2IyEaMOcuWOO+DsUU+HJnyUJGAhhGcV5sCPT8HJ7Z6OpEo/Hkxh5pvxpGcXYjIpWpry6HZ+G+Se83RowkdJAhZCeFbeOYh7DlL2ejqSKhUWtyM0PjbNAUX3rKUnsKgjScBCCM8q9P5ewFB2HTDAO7fGGk8USktCUTeSgIUQnuVMwN6+EUepnbCAkqpt6Qks6qjaBKyU6qaUWqOUOqCU2qeUurfo+BNKqZNKqZ1FX5e7P1whRJPjTGBePgIutDnXARs9i59dm8zZkH5e/4uD8F41qfu3AX/UWm9XSoUA25RSq4uee0lr/bz7whNCNHnOKVwvT2SXDepIr3bNMZuMBPzVUU1Kr3d4oe9gD0cmfFW1CVhrnQwkF/05Syl1AOji7sCEEBeI3pPhkVNgDvB0JFXq3T6E3u1Dih9bzEr6AYt6qdU9YKVUTyAa2FR0aL5SardSaqFSyuUiPqXUHUqprUqprampqfUKVgjRBJlMxujXyzfiOJySRXxCWvFjP5Pi/qS7YPO7HoxK+LIaJ2ClVHPgM+A+rfV54E0gHIjCGCG/4Op1Wut3tNYxWuuYdu3aNUDIQogmJXE9fPdISTGWl/pw4zH+sLhkrbLFz0znwiQ4l+SxmIRvq1ECVkpZMJLvYq315wBa6xSttV1r7QDeBYa7L0whRJN1chv8/Lqno6iW1e4oXgMM0KFFAIWmQKmCFnVWkypoBbwHHNBav1jqeKdSp80AvHsVvRDCOzkTmJ93NzUotDtKliABi347nJDmLcCa58GohC+ryU2XMcANwB6l1M6iY48Ac5VSUYAGkoA73RKhEKJpKyzqBWzy7m0JCm2O4k04ivkHef3UufBeNamC3gAoF0992/DhCCEuOM4E7OXKT0G/sOoXJuq+DAnt48GohC/z7rJDIUTT57B6/RpggHsv7ktuoa348d6TmazjTlZePNaDUQlfJglYCOFZV70O2vv76g7o3KLMYz+zqXh3LCHqwrtvugghLgzK1V0u7xJ/JI3NiWeLH1vMitty/w0fXuXBqIQvkwQshPCstf+Cjd6/DOm5Vb/w2o+Hix9bzCaCHech7YgHoxK+TBKwEMKzDnxtbMbh5az2slXQnVo2wy+guawDFnUmCVgI4VnWHA5nOMps8wgQn5DGW7QqkBYAACAASURBVOsSPBRURVabLlMF/fBl/ZgyOEwSsKgzScBCCM8qzKVFi5bMX7KDuEOpZOVbiU9IY/6SHUR2benp6IpZ7Q4sfuU+Mi1BYMsHhxRjidqTKmghRKN7a10CkV1bMjo8FKy5tG/TmkkB7blx4WZCAv2wmE0smBdtPO8lCu2O4l7AAB9uTCJzrx9395tuLKUyeXc3J+F9ZAQshGh0kV2NEW98Qho2czOW7j7Pim0nAMjKtzFveDevSr4Ab10/lLsnlWy6cfJcHgtSIuDaxeAnyVfUnoyAhRCNbnR4KAvmRTN/yQ7aBL9LwplsLhnYgc2JZ7lhZA8+3vQro3uHelUSjuhSdjrcz6ywObx//bLwXjICFkJ4xOjwUK4f0Z0jZ7K5fFBHtiSd4/XrhvDA1IuKk3P5wixPWrHtBLuOZxQ/tphNxLId/WwvOHPAg5EJXyUJWAjhEfEJaXwdv5tVnd4m6+BafjehF+1DApj26noUigXzotl9ItPTYRZ75Is9/Hfv6eLHFrMJDajcdGnIIOpEpqCFEI3OWeXcwZZJ33PreGz81Vy79iihzQPYd+o8pzLymDm0q9dMQWuti9YBlxRhdWwRSJd2bSETScCiTmQELIRodLtPZPKv30Rithu9dPt27ciCedEcP2s8Ts8p8GR4FdgdGq0psw545tCuPD1nuPFAegKLOpAELIRodHeND6dL62YEUZRo/YMYHR7K3ZN6E+BnIi270LMBlmO1G8VWFdcBF3VxssoIWNSeJGAhhEckZ+YRpIoScFEiU0oR2jyAtGzvGgE7ux6VHgH/b99prvn4EPn9ZkKLLp4KTfgwScBCCI84lZmPHRO2Ft0gsKTV39jeofRs6139gZsH+rH6/lhmRJck2qx8G5vOmDkzZQF0H+nB6ES9bXgZEuPKHkuMM467kRRhCSE8IqZHa3IvmYMa9zCYSoqb/jUr0oNRuWY2Kfp0CClzzLkrllW2ofQ9G16GLkMgLNZ43GUIfHIdDPwNXPmKkXyX3wyzF7k1DBkBCyE8on+nFtw5Phyzyft7AWfmWnlvQyIJqdnFx4zpaE3Y271hzT88F5yoXvkRrjPhrry37Hn7PocfnypJvs4E7SaSgIUQHrH/1HkyNi2GD68Ga37x8X+vP8qof/6A1t6zy1Rqdj5PfrOf/afOFx/zMymg6JcHWYbkXapLuHYbOOyweym8HAmLZxlbio64E+Keg5hb3Z58QRKwEMJDfrd4Gxs3b4Kja8DsX+a55Mx8sgtsHoqsogIXRVihIQGM6d0WhyVIWhJ6WnUJ17ll6L7P4e3x8PEMo3LdVgAZx6BLjPH81oUQ+2fY+l7Fe8JuIAlYCNHo9IaX6X5+G20sVqOln8lUXPTStrmRjL1pKZJzGZK/X8l0+ZDurVl820j8AoKhUBJwo6o24dpLRrivDIaPr4Y5H8GIuyB5J3QeAhP+D5q1MRLu6T3G62cvgkmPGt+X3+z2JCwJWAjR6M63GcTLppfpZDtpJGBn0UuXIbQNNjoLpXvRUiSrveIIuJglSNYBN7YuQ8omSF1qhLvkWvio1Aj3XBJ0igJbnjGyjf0zpB+Bja/DnA+MhBvxm7LXD4s1kvDJ7W79MSQBCyEa3a8tYphvvYcuZzeCvbBM0YtXjoCLpqD9SyXgfacyGffsjxzrMh3CL/ZUaE2fqyVCAL0mwdK58MZI+PAqGP+QMcI99F/oFGkk2matSxLu53eUjHDLJ9wrXjHuAZdOuGGxMPY+t/5o1SZgpVQ3pdQapdQBpdQ+pdS9RcfbKKVWK6UOF31v7dZIhRBNxqnMPDY6BpLfPgoKzpcpeunYIpBpkZ2KE7E3GNqzNfEPT2Jwt1bFx+wOzfGzeRzuewfE/NaD0TUx1U0v7/nMKNzbuxwKs41OVO37g39QyQj3bCJsegvmfOhVCbe8moyAbcAftdb9gZHAH5RSA4CHgR+01n2AH4oeCyFEtaK6tWLx5EKCMhMqFL20bR7A6/OGMKxnGw9HWSLAz0znVs0ItJiLjzmno602m+wFXR9VJdy0I7D/a6PKfNcS+PFp+O+D0DUGhv4WAltB7IOQeQJW/9XrRrjVqTYBa62Ttdbbi/6cBRwAugBXAR8UnfYBcLW7ghRCNC0d0jczZvsfUXMWVVr04vCiZve/nM7itR8OczanZFrcuRFH5KYH4G33L1lpMmqyJteWD9s/gAVDYcs7ENIBuo+CuGeN2ZJJj8GBlXDNR8affSThllere8BKqZ5ANLAJ6KC1TgYjSQPtK3nNHUqprUqprampqfWLVgjRJBzfu4HMNpHw/d+MAppyRS9XLdjA3Z/s8GyQpRxIPs8Lqw+RkVs6ARsfnzZToFRBV6W6hJuTZtQB7FpsjHCX3wxh46FNmPH88DtgxtuQsrdktmTvZ2U3yvCRhFtejROwUqo58Blwn9b6fHXnO2mt39Fax2itY9q1a1eXGIUQTcx9x8dTmLwPQjqCKlraU+oDM8DP7FVV0IVFVdD+pbohhe56k/lhp/BvFlKyDrgR9g/2OeUrlsFYIrRrCTzXG1b81hjx+jUrGeGOuQfyM42Eu+uTikuEDnxd8X18IOGWV6MErJSyYCTfxVrrz4sOpyilOhU93wk4454QhRBNjf+5I7Szn4HwSS6fb9vcn3QvqoIudFEFHRw2nD9l/pNOlmwjAZdaSnVBKz/iDYuFUb+Hj2fBfx8y/o5G/h7sVshJNUa7V70OZouRcDe9VTbhemiJUGOothmDUkoB7wEHtNYvlnpqJXAT8EzR96/cEqEQokmxOzQD8rYanz69XS/fadvcn02J3pOAXa4DdiaCJXOMEVwj7R/sdSprbNB7CnQfAbuWwqmi2wmb3jKSbNfhxhKhYbcZx07tMKaQw2Ih5wzs/bzk+le8AhEzjYTrfI+w2Cbx91yTEfAY4AZgklJqZ9HX5RiJd4pS6jAwpeixEEJU6UxWPmPVLs4H9YDWPV2e0zY4gHO5hdjs3tFpyOpiCjqnwMbQjws42naCcWDITU0iKVSrqnu6eeeMY1rDvs/gv3+G5N3QYwwEtIBxfzQS7me3VL4Jho/ez62LakfAWusNFO84XoGsPhdC1MqpjHw+s8fSYVAXBlRyzrCebbhlTBhWu8bPXMlJjeim0T25JqY7Qf4lwZhNir55O+iUvrGkOCh8YtNLwlW17pv+krEO15pnVC0n/GBMx89dApvfMe7VDpoFR74vGeHmpl8wI9zqqMbsOBITE6O3bt3aaO8nhPA+OQU2DiSfp0/7EFoGWTwdTp3ZE9aR8eF1rB30HDOvvNLY2nDTW01vGrp0b9ywWOPxJ9cZlcvKbGz5qMzQugecPWr8MhI2znhNzK2wcQFMfBRGzy97zZPbm+Sotjyl1DatdYyr52QrSiFEowpO2UZMyLlqk29eoZ18q72RoqrajwdTeP5/v5Q5Zk7ewd3We0hqMcRYFrPmaWNNqq8XB7kqohpzr9Gy7+1Y+PSmktGsNQf6TTceO6uWyxdRzVsGG16seM0LIPlWRxKwEKJRZX9xL+mf3FXlOQmp2fT/y3f8b9/pRoqqahsT0nn/p8SyB8fex1bTIGOJUr/pxigw41ffSiyu9lk2+cGSayBhHST9ZDQ3WP1Xo7FB8i7oe4lx3sltRsJNjKt6n+UmVLXc0CQBCyHcz/lBn5VC83MH+PL8RVWumQ0t6ojkLQ0ZCm0OLH4VPy5/E92FgZ1bQlAbI9Hs/6qkM48vKL9GNzHOGK1OfBSW3wiLLjeaG3SIgIAQGPsAHPxP1cuELqAiqvqSBCyEcD/nB/2mt4zHQW2qXDPbopkffiblNZtxFNq1y1aEz8yM5MrBnY0HA6827oGe3tPI0dWCq+nlsQ8YS6k+vNr4at/fuF87/A7jnIEzIesUXLsEJv9VEm4DkgQshHA/5zTkxtfJw59rzi+qslhJKeVVm3FY7Y4ym3C41O8KYxo64cfGCaomqtsG8qdX4Ie/GVXMR9dAcDujIjkxrqSz0KH/Gknax7d99EbVLkMSQogG0X00WploRh4/d76BkdVUCrcNDiA9xztGwFa7o7j5QmkTn1/L8J5t+NesSAhuC3dvhdZhHoiwEs6Zh9K/7DjsRuP65u0h/jUw+YMyQdQ8Ywq9MLfsa5wVzZ0iL7hlQu4mCVgI0Th+jcfhF8gH+bHMS/kcEq+u8kP8ptE9yrT/86SXr4nC5qI7k92hKbCVqtRu06sRo3Kh/Jrd0lPM3UfD0bWg7cYoN+5ZYy1v4jqYvdQ4d+AMowBr4qNlr+EsopKk26BkCloI4X7Fa0k/YOofF2H7zcKKG/SXc82w7lwV1aXRQqyKUsrlPWA/s8LqTMwbXjYqh7+aDz+9ahxr7OYM5YuqDvynaIo539gkI7gdjLgLEtYUTS9/V3Z6OSzWWDbksJW9rkwxu4WMgIUQ7pe4AUI6YfYPomvrIGg9CQIWVTmqyi20cTozn17tmjdurC68tyERu8PBHbHhZY77m01Yixo1FCe/kM5wfDN0jiqZynUXVyPewXONxgdj7jXu4wa3h9w04/jez2DnkpJ1vDK97FEyAhZCuN36ZCBlL9tO5fPG2iM4HJp4xwDesl9R6Ws+3HiMSS+sI7fQVuk5jeX7/Sl8v79iwzc/syqZmnZO1Z47Cmm/GFO5Db0rVmVFVZ/fARvfgFeijJ2nlMmYYu49GWx5cN1yuOJlWaPrZSQBCyHcLurcKg7Rg7cONuPtdUf5OTGd+Ut2ENm1ZaWvaRvsD+AVldBWuwOLX8UirOmRnZl4Uak+52GxxhQvGHsi1zexVVfFDMYGGbuXwf/+DzKPG5uCWJoZU8wHvpYKZi8mCVgI4V7pCYSk7cB/yLWsPXQGP5Ni/pIdLJgXzejw0EpfFtrcuRmH5yuhC+0Ol/eA7xofzg2jepYcSIyDbYtg3J+MJUkbXoL882Wfr+qecHUJNyfd2IN550ew8m5j+njqk9B9lPF8xEz4dWNJpyHZBtKrSQIWQrjX7mWAoueEmwn29yM9p5DrR3SvMvmC0RMYvGMEXGhznYDtDk2h8x5w6aYFFz8Ok58w9kfe/qHRfD5hjTEtbSpVerPhZYhfUJKUXY1wtTZ66j5/Eay42eg9HNDCuG7MrcbGGWmHXI94ZYrZq0kCFkI0vNIjudC+MOIuduzYwjWFnzMirA0fb/qV+IS0Ki/RtmgE7A1rgf3MqkwrQqdbFm1h9lvxxoOT28ve8x1zD0x9ymjS8O5E+GiGsUZ47T9hz2dGZbLJD1Y9ZtyzzThu7LVsL4RdS4xCquU3G6NZbYfs09BzHFz1hnG+ND7weVIFLYRoeKU3gBg0i73nA+ix+nfs0fdwy9gwQgL9qp2GDm3uzxNXDGBI99aNGror39w9zuVxi1lhtRcVYblKcqPnQ14GrH8OOkZCThoUZhsN6UM6g73ASNKr/wKrHy/1QhMcWW0kWbPFGPEOu81IuMm7SqqYc86U7a0ra3Z9iiRgIUTDcyaCZdfDkJsI3/IhCVPe5P0RlwMQ4Gdmwbxodp/IrDQBB/iZuXmMF+0q5YLFbMJqd1R+QmIcbFtoJNKt78Gs9421uKseh4TvjeOj5xvbQB75HiLnQJ9L4L9/NqaXN71lfFWWcC/gZvZNgSRgIYR7dB5iTLPGv0qz2D8TMabskqPR4aHV3gdOTMvBZnfQp0OIOyOt1v99voehPVoza2jXMsf9zCaXO2QBFRvZO9fcjn0AkneUJOXAlnBqR8mU8i/fScK9QEgCFkK4xw9/N6ZYo66Hre/x/qmu+IWPL1s1XI0/frqTIH8/Pr5thPvirIFvdp8iwM9UIQEbU9CVjIDL3xN2bgu55mnjXm1YrJF8Vz1mTEOPni8J9wIjCVgI0fAS1sKWdyG0H1z9Oo6EdVz14XWsVP+AGiTgt9YlENm1JW2bB3D8bC4A8Qlp7D6RyV3jw6t5dcOz2h34u+gHfHG/DvRpX8no3NU9YYetJPk6H099qmTrR0m4FxRJwEKIhrdrKWgHTPw/AI61jOER6z08bE6s0csju7Zk/pIdRHVrSVp2IfEJacVFW55gtWuX7QinRXaq3YXKJ2VXSVoS7gVDliEJIRpeUFto1d3YlQnYf+o8Gx0DMY+r2XKY0eGhLJgXTXxCOmnZBcxfXP3GHe5id2jsDu1yHXC+1c65HM+vUxa+SRKwEKLhXfoPuDMOzMYk24Hk85hNit7ta95YYXR4KCN7tQXgqujOHkm+ADaHg/YhAYQEVpwwfGHVL4x+5kcPRCWaApmCFkI0rPzzENgCmpWs3zWbFCPC2tSqv298Qho7fs3g8kGd+HLHSaYM6OCRJBzgZ2bzo5NdPmcxm7A5qliGJEQVqk3ASqmFwHTgjNY6oujYE8DtQGrRaY9orb91V5BCCC/nbIvXphcsGAZXvAohHYxiorH3cf+UvrW6nPOe75vXD2F0eGiZe8CeGgm74mc2YbVrtNYoVbFZgxBVqckU9CLgUhfHX9JaRxV9SfIV4kLm3Plq1ePGXsVKGY+7DKnT5XafyCxOtpuOppfZuKOxnc0p5NZFW1h/OLXCcxaTkXQrXQssRBWqTcBa6zjgbCPEIoTwVWGxcMVrsO8LaNvH2MmpaA3sxoR0Jr2wloOnz1d7Gae7xocXj3Qf/2ovb65NYHR4qEeWIOUU2Pjh4BlSzlfck9pStDSpyt2whKhEfYqw5iuldiulFiqlPL9ZqxDCs47/DGijGX3MrcVLafadyuRoak5xe8HaCgsNJjEtuwEDrZ2Com5HFnPFKebhYW148JKLMJtk+lnUXl0T8JtAOBAFJAMvVHaiUuoOpdRWpdTW1NSKUzhCiCagIAu2/BvMASVbLBZ15NmffJ72IQF1TsA9Q4P59Wwudg9N8zpHt67WAQ/p3po/TOxNgF/Ni8uEcKpTAtZap2it7VprB/AuMLyKc9/RWsdorWPatWtX1ziFEN7s1A4j+c5422iLN3uRcQ84MY4DyVkM6NyizpfuFRqM1a45eS6vwcKtjeIE7GInrJwCG8fP5soUtKiTOiVgpVTp7V9mAHsbJhwhhM+x5hvVztd8CBEzjGNF3ZBsx7dx5EwW/TvVPQH3bBsMQGJ6TkNEW2smpQgLDaZ5QMVFI//Zk8y4Z9dwOjPfA5EJX1eTZUhLgQlAqFLqBPBXYIJSKgrQQBJwpxtjFEJ4G+eyo7BYWHELWAKNxxteLtleMSyW7A4juTrlAKOKNtSoi0FdW7Jy/pjK91x2s4guLVnzpwkun3NOS8sIWNRFtQlYaz3XxeH33BCLEMJXOJcdjf8z/PIfiL6+pPUeJc0URoeH8tzswUDdmykE+fsR2bVVw8bfQPzMsgxJ1J1sRSmEqL2wWKO5/P8eBUswHPy2TOs9ZzOF1ftT0FoXb6QR2bVlnd7ux4MpLN96vAF/gJrbmnSWa97eSGJaxSlwi4yART3IVpRCiLo5f8poo+ewwag/lOng42ymcON7m+nQIpA8q71eu1h9vv0ku09kMjumW0NFX2OpWQVsSjxLgc1e4Tnn0iSrXUbAovZkBCyEqD2HA9b8A0x+MO7BMsuOnGJ6tEEBJzPyuH5E93ptIdkrNJgT53IptDX+SLPQ7lwHXPHjsm+HEJ64YgCdWwU2dliiCZAELISovWMboDDbWHZ08WNllh05fbgxCatDM21QJz7e9CvxCWl1frueocE4NPx6NrfeodeWc3Trah1w19ZB3DwmjPYhkoBF7UkCFkLUTnoCnNgCcz6AQbOMY0XLjji5HTAKrp5f9QsmBc/MHMSCedHMX7Kjzkm4Z6ixFCnJxX1Ydyu0VT4Cziu0cyD5PFn51sYOSzQBkoCFENXb8LIxui3Igg+ugJR9JcedwmKLlyDtPpFJh5BAhvVsQ0igpfiecF2bKfQqSsDHPDACbtnMwsDOLQi0VPy4PJSSxWWvrGdzomyXL2pPirCEENVzLjvqPsoovuo2ssyyo/LuGh/ORR1CipfpgFGYVdf7wK2C/Nny6GRCm/vX6fX1MS2yE9MiO7l8zk+KsEQ9SAIWQlQvLBbGPgCrHoXOQ2DdM2WWHbkysV/7Bg2hXUjd9pJ2J+d9YZtDliGJ2pMpaCFE9XLSIP41aNYGTm0v0+3Ile/2JrP3ZMP27l29P4UnVu5r0GvWxMc/H+OK1zbgcLHZhp+sAxb1IAlYCFG9vAywBIG2V+h2VJ7DoXnki728tyGxQUM4kHyeRfFJ5BVWXI/rTsmZeexPPo/JRctBP5NMQYu6kwQshCjLWXBV2qHvIPs0XPNxhW5H5e09lcnZnEJi+9Z93a8rYcWFWI1bCW21a5e9gAHaNvfn2VmRDOvZplFjEk2DJGAhRFnOgqvEOGNZ0eLZ8ONTMPHRkmnncsuOSos7ZPT9HtenYduPOhNwYmrjJuBCm8PlGmAw9qmeE9OtODYhakOKsIQQZTmT66c3gXZAwXmY8zH0n1bxPBf3geMOpRHRpQWhzRu2aMq5Frix2xJa7Q6XvYAB7A7NzuPn6NyqGZ1aNmvUuITvkxGwEKKiHmOgWSvIz4DIayom30rkW42NKWIbePQL0DzAjy6tmpFTYGvwa7vy1roE4hPS6NE2iJgexhRzfEIab61LKD6n0OZg5psb+XLHqUaJSTQtkoCFEBWtuAXOHoXeU+DwqkoLrpycySrQYmbLY5O5c3x4hWRVH87rb3hoIg9e0g+omAwbmrOjU0SXlrx1w1CXHZ2K2xFKFbSoA0nAQlzoyhdd7fsS9n8J7frB9SuqLLhyciYrZxLedyqzXu0HK7v+xqPpAPVub1gTzt275i/ZwYurfmH+kh0VOjqVVEFLAha1J/eAhbjQOYuunBtrHF0D/s3h0meM50sXXFWy9nd0eCgL5kZz08LNTLyoPVuPnatX+0GX158XzZ0fbaNFoIXcQhuvXzekwa5f1fu2CPTj1R+PcM+k3hXeTymFxaywulgjLER1ZAQsxIXOmWCX3QAfzoADX8PcpRA+sew5Rfs8V8ZkUljtmlX7U+rdftCV0eGhTOjbjpMZeUzq197tyReMkfaxs7kM6BRSaUcni9kkU9CiTiQBCyEg9CJAw9EfYdA1Ve5yVZnXfjiMAn43Prze7QddiU9IY/0R45rf7jnd4Nd39X5/WLwdreGyiE6VdnR6+ZooZg7t6tZYRNMkCViIC11eBiycCvmZEH097FlWbdFVeav3neanhHQu7t+ehy7rV+/2g+U57/m+cd0QwkKDGdC5RYNe35XdJzJ59PL+AHRrE1RpR6epAzvSr2MLt8Uhmi5JwEJcaEoXXRXmwvuXwbkk6DMVrnq9RkVX5S3behyA+yb3Bah3+8Hydp/ILL6nHN29FUlpOSyY23DXd+Wu8eG0LVrL3K2NscZ3dHgod40PL3PepqPpHEg+77Y4RNMlRVhCXGhKF13ZC+HMAfBrBqPvNp6vQdFVeb8dE0bX1kFEdCmpSq5P+8HySie9cX1Cycy1MqhrS0b3du994BbN/LgsoiM92la+09Ufl+9ieM82vHhNlFtjEU2P0rrxqvdiYmL01q1bG+39hBCVSIwzknDMrbDlXZjzYZ3u+wqY9PxaBnZpyWtzoz0divBCSqltWusYV8/JFLQQFxpbIWx9H8ImQNyzMOz2eiXf7/encCojr+Hiq6ECm/u7ItVkfa+fWWG1SRW0qL1qE7BSaqFS6oxSam+pY22UUquVUoeLvrd2b5hCiDorfc/XVgCf3gD7PoeDX1fbWrAyzp2psgts3PPJDl5afcjtO1OV9ucVu5jxerzb3+fq13/i7qU7qjzHYjZhc0gCFrVXkxHwIuDScsceBn7QWvcBfih6LITwRs57vodXwyfzjNaCABf/tdrWgpVx7kz10upD5BbaGdilpdt3piqtU8tmHDx9nqx8q1vf59ezubQOslR5jp/ZRKH0AxZ1UG0C1lrHAWfLHb4K+KDozx8AVzdwXEKIhhIWCzPehqXXwpHvwewPU5+G0fNLnq+ktWBlnFXOi+KTaBNs4dUfDjfozlfVGdKjNQ4Nu467rwo6M9dKVr6Nbq2DqjzviSsG8OdLLnJbHKLpqus94A5a62SAou/tGy4kIUS9lN/bGYyk27yD8ecx95UkX6ca7HTlit2hOZtjdcvOV1WJ6tYKpWD7r+fc9h7Hz+UCJUuQKhPdvXWZ6m8hasrtRVhKqTuUUluVUltTU1Pd/XZCCOeUc2IcnN4LOxbDsuuhIKvO93xdWbnzFCYFv5/gnp2vqtKymYW+7UPYdsx9CfjXs84EXPUIeGvSWdYfls82UXt1XQecopTqpLVOVkp1As5UdqLW+h3gHTCWIdXx/YQQNeWcUv7kOrDmg7YZzRWuXWw8FzaubPOFOohPSGPV/hQW3jSMCf3aM7ZPqMtuQe5027iw4naA7tC9TRC3jAmjezUJ+M21CaRk5TPODT2QRdNW1wS8ErgJeKbo+1cNFpEQon60hpT9xogXDd1GGsVWzmRbh402SrPZHazen1Im2Zbe+aqxEvDsmG5uvX5El5Y1mlo2liHJ2ELUXk2WIS0FNgIXKaVOKKVuxUi8U5RSh4EpRY+FEJ5QfpnRyvnw3UOgTDDmXkg/XPE1dbznC/D59pO8/1MSzSzmMsddbdPoTlprktJySEzLccv1T2fmk2+tfq2xn9mEVZYhiTqoSRX0XK11J621RWvdVWv9ntY6XWt9sda6T9H38lXSQojGUvqeLwqSNhjHJ/8Npvy9TsuMynOu+8232nnp+0NEdWtFXqG90db9uopnY0I6V7/xE2+uPQLQ4OuQ5737M/cv21ntef5mU4027BCiPNkJSwhfFxYLsQ/CpzdC3HOQfcZYZjTGxd7OdeRc9/vkN/tJzsxnemQn5i9tvHW/LuNZuoOwtsFsO3auuFtSQ8XjcGhOnMurtgALwM+ksMk6YFEHshe0EL5mw8vGqDcsFhwOiH8Fvv+bscwo+7RR6Tzp0QZ/21X7TnPnR9vo47LvuwAAIABJREFU1iaI7AJboxZcuRKfkMati7aSZ7XTOsjC69cNabB4TmfmM/KfP/DkVQO5YVTPKs89lp5DVr5NliIJl2QvaCGaEueU84H/wCdz4fsnwGQGa06DLjMqL8jfD38/E7+ezW30db+ujA4P5fJBHQEY2attg8bjXAPctQYj4B5tgyX5ijqRBCyErwmLNbaR/PR6OLQKzIFgCYJrl9R5a8maMJkgyN/MPZN6N/q6X1fiE9JYczCVZhYTPxw406DxHHeuAa5mFyyAXcczWF7UD1mI2pB+wEJ4u9JTzk5BoRAQAvmZ0H0ExP6pwZYZlbf/1Hm+3HmSFdtOFE/zjgxv2+jrfktz3vNdcF00AX4mzmYXNmg8kV1b8ti0/nRtXfkuWG+tSyCya0viDqWxcEMis2O6EZ+Qxu4TmY1aDS58lyRgIbydc8p56tNw7CeI+A18eZfxnHPKubyw2DonX2diGR0eSm6hjbuXbufkuTymRXby6Lrf0nafyKyQbIMD/Rosnt7tQ+jdPqTKc5yFaeP7hmJ1OIg/ksb8pcYvAULUhBRhCeFtyo94HXZjV6tD/wW/QFBm456vc2erxLh672xVWvHocl40X+86xdLNxwkJ9OPtG4Z6/L5vZZZvPU7K+XzmT+rTINfbcyKT0BB/OrWseh9odxaCiaZBirCE8CWl1/WeOQhvjDSSb0hnsOVD15iS5AsNssyoNOfo9vYPtrJ083ECLSavTr4A246d4821CTXaOKMm7vp4G89+90u1540OD2VoD6Md+qyhXb3670h4H0nAQngbZ0JdfjN8eCWkHYZBc8BeYEw5p+x1/Zo67mzlSrfWQeQWGsns1rFhXp9YLh/UiZxCO+sP178Qq9Dm4FRmzdYAxyeksfN4BgDLthz3eGGa8C2SgIXwtPLtA3/9GQqyoUMEZKdA/ysh4QcjKbuhytm5y1Vp3+1LxuJn4nfje7F0s/cnllHhbWnZzMJ/9yTX+1qnMvLQGrpVUYAFJVP1r86N4vsHxvP6dUOYv2SH1/9dCe8hRVhCeJpzyvmKV42p5h0fg8kfLIHGiHfjApjYcM0UynMWEy2YF03rIH8+/jmJJZuO88i0ftw+Lpxxfdt5tOK5JixmE1MHdOC7facpsNkJ8DNX/6JKlPQBrnoEXL4QrHf75h4tTBO+RxKwEI2tfJFV99HQfQwsuw5QRck3oGL7wE6RZZNwAyRfKLnn+7uPt2O1O8grtPPQZUbyLf28tyeW6YM7k5SeQ2pWAV1rsH63PGf19/GzeYCRgKtaVlT62Ld7kknNKuCm0T29+u9IeBeZghaisZVpngCsuAUOroTgdoCGHqOMTTXcVGTlit2hySu0kVtoZ+6I7hUSTmN3OqqLA8nnuX9K3zLJtzYNGpwzAc0D/VgwL5rE1Jwa7y+9en8Kb3uoMYXwXZKAhXCn8vd3ncJijeYJPz4Nietg8FzQjkYpsip/z/eD+CRufG8zVrvmplE9+G7vaZ+8j+lMoKv3ncZqd9S6QYNzpP/Eyn0cOp3FPZ/UfNq9d/vmnMrMJyvfWt8fQ1xAJAEL4U7lR7v7voTFs2DfVxDcHuKehb6XwOFVbiuyKs+ZqJxJ9utdp9DAny7py9+uimDBvGifLCYaHR7KvZP7cPtH27h/2c463bceHR5KVLdWvPrjkVrtd92nfXMAElLd05tYNE2SgIVoSOVHvGGxMPYBWDIH3rsElt8Ediv0uxxyzhgj3gNfG+e4acq5/Ih3dHgo1w7rxm/f38KLq35h36lMHrmsH3+Y2Kf4eec9X18zd1h3gv3NfLM7uU4NI77YfoIfD55hXJ/QWu133beDsWvW4ZSsWscsLlySgIVoSOVHvIlxsOFFaN0Tjv8M7frD1W/BrxthzofGiHfeMuOcCom7YaacS494tdb8678HeWNtAoEWE6/+eITbx/XiDh+85+vK1mNncRRt7rcoPqlWo/j4hDQe/dKY/n/66kG1mgno1iaIAD8TpzLy6xS3uDBJAhaiPqoa8S68zNhCcuwDkJUCQ24yRr2/xpfdNtLNRVbOEe0fFm/nkpfjeHNdAt1aN0OhvKazUUNw3vN9bV40zQP8iOjSslZT6btPZNK9TRAXdQihe9ugWs0EmE2KnX+Zyr2TG2YrTPH/7d15XJVV4sfxz2ETRFEERRRxuai4YSqhYm6kprmUOVq5NKU/M8vRcqzJfs2MNb+WSSfDLLcsS7SysqbcTdwSRQERQUVWWQVxQUBZ7r3n98cDV0BcEOR69bxfL15xeHiee3gwvvcszzkPBhXAilITlVu80T9D8L9AX6wFrVNLrXU74WsYs0QL2pO/XX+duzjJCiDrciH5hXpOZ+XTqXlDCooMfD65J3OHdbTYMd/Kyp7LHdLJjWf9WhGWfJEPn+p2213pT/u2Ii47n6Gd3Uxfq05PgIPdnT97rDyYVAArSnXcsMX7NHw5HH58HvRFYGUDvV6AS2fu6vhuVcp3ORcU6QlJyOEfv8SAgKcf9iD5/BVmDm5X5c5GluylgTrTz/TiAB2/zx3IsC7NbztAo9K1n39IuQCujpD4HKZ/E8aVYv0dna88eNRCHIpSHWUt3vFroE1/CFkCBwKh02iI+h5cvODKeW18t+0AbevAu7iIBlTcPhC0QJ3Sx5PnVh/G1lpgZSWwEoKvp/rhr3PliYdaMmv9Ubq0qHjO/bSARNOG9UyfSykRQtzynIEdmhLx9lAa2t/Zn8XcqyXsPJFFQnYB3W7z0SflwaZawIpyM1W1ePu+AkHjYKEOdv4D2g6C+N+1Gc2XM6D/PLO1eM/nFzFrfQSBu+LRGyVXS4x0cXeqsJvR/dLivZWlu+MY+9kBlpVbIONWC3M0qm+LldWtw7oq7d20R5HistVMaOX2qABWlPIqB27LntpEql/naF3L302CXf8CQ7HW0m3VG5L2XHuG9y7PaIaqHyuaOagdU9ccoc8Hu9gUlUm3Fk40crBldoAX8VU8m2qps5yro6enM8czclmxN5EiveGmC3P8EZfDhBUHSb1w5Y5fr7WLIzZWgrjs/JpUW3mAqABWlPIqT6oqKQSjAWI2wr6FcHqb1p1s3wgGvA5nj9/1Md7Kgevj0YgZa8OZvzGKY6mXWPBrDMv2JDKiqzslBskQ72ak5xay7D6bZFVd/jpXXh/WkdyrJcxYG37ThTm2x5zleFpuha7r6rK1tqKtqyNxWSqAldtTowAWQiQLIY4LISKFEGG1VSlFqTNVTqp6DdaNh+X9Yf14MOq1CVX7FkKXpyA3DZ4OgoC366TFW3nlqqvFBkr0Rr4/ksoTnx3g65Bkpj3Shr2nzzE7wIsDCeeZOej+m2R1J14c0I4WjezZE3uOcT1bVhm+Ukp+P5lF//au2NvWbCZzD8/G2Nuqdo1ye2pjEtZgKeWD9dZasVyVdyIq62Lu8hSMCYT9iyH4XW1d5rNR4NYNekzWloysg60BoepJVTMHtWPamjAe7dSMTVHanrfO9W25eKWEUT7urP4j2dS666Nzue8nWd2ug4nnKSjWI4C1B88w2LuZ6T6U3Wcne1sycwt5bWiHm+5+dDs++lP3Wqy9cr9Tb9WU+9vNxnSlhAtJoC+E4z9oGyMc+ASaeIFdA+g3R3uMaPd7d3WM90ZdzG/+FEX4mQu8tDaMwN/jGd61OZuiMunSwol3nuiCENpCGr+fzFYt3iqUjfkum9yLFVN68dmknhV6Esp6Fr46kIQQ4GRvU63NGxSlpmoawBLYIYQIF0K8WBsVUpRaVXlMV0ptTPfYevh3G/httrY2s0evaxsjXD0Pz34LQ9/VHiMqrxbGeG82piulJDLlEoXFBr4/ksq4ZQfZFpPFw22cTV3MKReusGh7LEsn9mDusI6sft6XZXsSr5uYdb9PsrqVsoU5/HWuDOvSnEc7uRH4zENEpl4Crr1R2RaThU/LRrz1c3S1N2+oLOtyIWOW/sG26Mza+jGU+1hNA7iflLInMAJ4RQhxXR+cEOJFIUSYECLs3LlzNXw5RbmFqsZ0/edoY7rB78GPL4CNvRa6hZeg/TDtmd2smKo3RhgdCM+sqxi41WjxVrUqlbUVTFsTZlqbOePSVUoMRjZFZbJ452kW7YjFykrQ1tURgFE+7hwrDZO5wzoyyse9wvVUi7dq5RfmAPjPjljmb4yikYMteoORjRFpSCN092jEsbTcO9q8obLG9W2JybjMiYzLNa2+8gAQUsrauZAQC4B8KeWiG32Pr6+vDAtTc7WUWlJ5PBcgZKnWZfzkcii6DJHrtI0PrO3BUKiFbAM37Xse/h8IXa6d98w67TpJ+64ttHEHY7qVx29DEnKYsTacUT7ufPCUDyEJObwcFMEj7V3ZcSILO2tBfpEBB1srpj3SlqW7E5jSpzWPdmrG3A3HmNzbk1X7k5g7rD3T+19r0dZ0rPJBFBKfw3NfHqaejRXt3RoSmXoJe1srbK2teMG/DUGhKTVqAZf97t/+JZqObg1ZNrmX+j0pCCHCpZS+VR2740lYQghHwEpKmVf6+TDg3Tu9nqLc0s0mUD3+EZw5pI3Ptu4HPzynfY+wAo+H4dwp8J91feAWZEP0xmuvUc1JVZUDt6w7uSxwi/VGDAYjGyPSaeRgy4awNAZ2bMp/IzOoZ2NFfpGBwR2b8sRDLXl30wlmB3jxVUgyv0SmmxbPUJOqaoe/lyvvP9WNN36M4mTmZVP4Vr7PdxrCZWPK7VwdOZ2VZxqDXjqxx134aZT7QU1mQbsBP5cu8WYDrJdSbquVWikK3DxwR30MOXHa4hiRQdqHnaP2eJB1PW1iVfJ+6DZeW6XqmfXada7kVAzc0YHQdVzFwC23VGTlgF2+NwFrKzAYtS7OyoELYDRKfgxPI/hkNll5RaaXWr43kdkBXkx4uBX+Olf+vfUUMwZ48lVIMmFnLpqC4Fx+kWmmM1TsYlahWzMTfFsRkXKR7w6n0k/nwisBXlVOXruT+1x2/tQ1RygqMfLKugg+m9RT/c6UG7rjAJZSJgJqzr1Se27awl0IxaWrFEX/AMe+BUNpuNV30Val6jb+Wjdy9omqHxuqFLhawHbG/5FrgXuzgLW2gvc3n+KVQTqCT2WxPSaLq6UTpnKvlnAo8QKvD/fm/c0nyMoropenMwM6NOWrA0k817c1QaEpNHSwYdmeRJZO6lFl4H7wlA+ju7eoEASqxVs7QhJy2BGTZdqGsbKa3md/nSuPerux+XgmE3xbqd+ZclNqMwal7lQO2D8+0XYNMuq1SU3lAzfgfyHlkLYS1dG1cDQIrG1h0gaID4YDi8F7tBa6m1/TwjZsNTi31bqhS8dwf72sY8Te+diWboagBWxLDMaWvMT1XcZlAfvWSG/yCktIOldAicHITxHpFBQZ+CM+h1cCvFgaHG/6sVwcbWlgb8uW42eZHeBF+6YNaFDPlsl9tNbt6ey8Ct2c09aEMXdYe9MfZxW4daN8l3BtdDnf6DUOJp43BfzAjk3V71G5IfUcsHL33OwZXNDCd8fb2ozkxL2QFq59PWYjLO6qTZSSenBsClJPrMujRGfkwtFvYMAbFMUFU/Tzy/za4X0tsMevoWjnv9jU6FlTyLt2G8KMolls3b4ZwBSw1lZQpDeQlVuI3mDk54h0Zq2PYNmeROaP8Ob9LafotmAH//tLNIUlRor1Rn49lsHk3p78dWgHArybAdpKS59O7Eleod40fvvSunCWTrrxjOXVz/tiMHLd19VEnbur/GNJUPuzx8sH/NxhHZk/wpuZQREP3BKgyu1TLWDlzlWnRTsmUHsGF7SAzcvAEL8bo21DbHf/HwBGrNnS43OaXwjD98wqTjUfTaLbMAaf+DsOA95Ad3AZV3cGs6rdB0wPeJ7E+HhapG+lbPOaEGNnvpBv0DUzgQNxOXi7NyT1wlUOGjoTkt6ZX9aGcyT5Am+N9ObDrbG8t/lUhR9nU1QmswO8eHGgju0nsgg/c5HHuzVnYIdmfLj1JFP6XOtCjky9ZArcbw+n3HD8VrVu7x1VvcGpzd9F+YAv0ht4b8tJ2jdroMbulRuyvAC+1R/9ymWA3+ZoS4aMCbz9c5L2aeOEtbimr1lV975VcU9OrHiBJg1saT5pJQC/5jTnsT3PktB0KJ1nrOFAUi7+CYsJ0b1Gvy5nOLJ3Ez2KryKOBmF9ehvG/Gy29FimBWzcKox2jdlV3AkvxyK8CiI47jmRdaEpfOn4Awx4g7YHPqNF5i529VzMYwMnkJmaTOOk3ziYeJ5TP0SyNe1ppjTvg03YXpYkupN8/gprXvgzAM+tPozeWPERu+0xWhfx9P46gk+e42DieYZ0akbvdk1YGpzA5D6epoBNyikwBez+uBxTwDZ0sDF1UU/vr1OBq5iUD/h6NtZM6dOapbvj+XCcjxlrpdzLLC+Ay2+I7ukPVtZaN+YwrRVl6tYsKyftuzbrtds4LYAqf0/l8m9ztHOeWaeVqxtWdxBu1X2TcOibv+Pm3IC2zvVMZWllgzDq6fPcv64r3zQsH+GWZZL2ocveTlGGJPrAb3TtNQDnq8kYSorokLUZVgykz9loPiyZwKspX8KP+3k4/QgAxdb1sc4/S1aDLnwXmsQqxx+45DuHese+5qjBC7/8X9joNJEhZ35kTX1r/mKYx5VEf0YWhzHS+iDrQlOYdXAr8CSbRo/h+dOHeC48HYAVKR7YWbeiOCsfzyb1TUH3aCc3tsecZXjX5jzcxpklu+KZ1PtawMZm5ZkCNjTpgilgG9e3vWnAGozw1khvUxeyClzlRv7s34aV+xJZuS9BrRGtVMnixoCXp7Qkul+gFsJLuiN3vI1e2CB3vA0ftsawcwEHvV7j6u6FsOpRCBpHnE0HTtp11lZDCvoTV3cv5KDXaxTveg/WPkXxrveIdxtG8a734ZeXKYr8gSKDkX0HQyA9nKOx8cgdbxN+OgWuXOBAUi5yx9uEJF4Ao9FUPpCkjSVVt0zSPq4e/YGrkT+YxkxvdY6bcwNah3/AphhtdTFpZU3vuMWgPRaGEEZ6x32MY/E5SD5As7wTGEqK8creAcHv0evMF2w1+OKXvAzWjMYv6XNCDJ3odeYLCH4P3+SVnDJ40DtxCUULO2H8+kku6u0I7TCPljtfJnbpn+gf+wH1RQk2xmLIjOScsSGufSezVj8E0o/wvWEQMwyvk6e3JlA/lvp5iax2WMocw6tMTx3Ooisj+ZtVEJ/pxzD33Cg2GfpiZ2XF0E7N+CM+h+VOswnyfJ/HGmcAMKhDUy679eXVtEFM7deGxg62vDmiIw3stX1v84v0hCTkEJKQw5HkC9rOQPE5fPJ7HMsm9+SN4d7MHNSO9zefYuagdlWO0VYVsCum9DKNE740UMf0/roKrR01fqtU5cfwNAa0b8rPR9M5m1sIaOPEy/cmmLlmyr2i1lbCuh21sRJW2USH/3beQ6vjS0lx9GFXrjsjnBJpfiWOtAY+9M95k5+8d9MzeRWFto05V2xHPUpoYlWAjSwmos10xp0azMn607E3Xr9ZeUSb6Sw83Yz1du8jqHh/9K6dKcnNYNGVkbxtuw6B1nDVSytshBHZ0o+i7NMsvjKCN22/R9g7IQsvk2NsgKtVPrLdIIpTI/j0yjDm2f2EdGyGKMgmweiOANrYXsTG/xUKQ1Zwvtgad6tcjA7OWF29wHljAxrZ6rHrN4uSg8spKS7EnmKksMIaLTEKsOe4xyS6pH1LQ67fXHw7fXiMQ2ymHyM5UOFYkbRhu+jLGPazSfajvTxDI5FPc3GJRGNzomVbdE9/QN6htfRJW81vhj4cwod5Vuv5xjCU52x+Rz4yF/tDgawqHMw0m+1YWwk+cHqbb8625kuXIAaX/MGP7T/k9fDGfNZ6Hx5NGrL7ZCaFvWexISyNd33OkxC5H0Pf2QSFpjBzUDuW7Ulkcm9tRjFgaq2u2p9QobVatupU+e+ZvzGKTVGZpnLlx4zK/k2p1YqU2haSkMPMoAiK9AYCn+lBw9LNHmpz1rVy77vZSlgW1wL217nyTUAhDY5/w5fW43HMTyJNumJbcJZA/Vgc8pJY3zWMtsnfs0r8iYJiA2+UTGdOycvkGuvxlc142iZ/z/quYVwxCFboR3JRNuDTkie4JB35yWYkbZO/Z1JvT342+AOw2dCb2cWz2GnoiU3OCdbqh9B8+F8JMXQB4IixA6sNjxNp1GGVfpi1+iE0G/IqscaWiMJLxBvdCZPeJBndsE4MZq1+CI0GvkyqwQXr/ExSDS5kGp3JMDoTXNIF9n3EBn1/TstWnDE2xeZqDonG5oTKzuzQ94B9H/FdSX9+NAzkmNRhjZFDRm/+XfIMRxo8Sp+01US4jGZm8Rz+a+gLwEbjAJbYz6Sf9QkC9WMZZB3NjqYvcFk05DP9GC4LJ47o/sKjNscJ1I9lsG0Mhu6TcLSRBOrH0sz2Kg89+Sol51PpmPYDB1tOZZD1Mf5Rbz1/s56HYcB8vhRP4fzHuywzPIFhwHy2C3+sBJy9XMTsAC/mFU5lc5eFZJwIYXaAF29mBTD55MP4TXmHN0d0YuagdvzlYEPqB8xl7rCO1W6t+utcGeXjzigf9wqP+KgWrGIO/jpXlk3uiYOtNTHpuSp8letY3hhw0j66HpjDD10+5N3wxjh5uPJ2znJ+dHmJxen98XR3Z2z8Yva0mc17p3pT1L4fa8/+A4AV7u+y8LQbbb1bMCh+Mb+4zeSDlEdwbe7JrJzl/OT6EvPS+rOw10jGn3idK7YGAovHMtUuGA+fAbQ7mUBg8Vim2/9Ofese5NmmEVg8lhfsgjHqRqNL2G86bm/3EPk2l03Hc9o9QtPE1abjtvW7c9WmyHQ8tpf2aI5/xF8JLBrLdPvdxHWejuuplVod6gVj6DoIj5jlBBZqx6PaT6N13CF+b/ZnemZvJL/1Y3RP3cwhj2n4pG1AdPwfuiVFc7DlNIamf4udPoK/yHl4D3ic5QedmHtuDR+L5xADXmHVQWfmJn5iKi8/6MTc4x+ZynMPdefTLVNxNkjODF1J336jObnyEi3StzKmpzujhnUkNLsR/3dyIsM6NaL3sI6scviI57f+xKudL9B7WEcaOtjwl83FvDVy7m2Nr95qvLWq0Cxbjao8NSarmIu/zpUpfVqzJDieRzs1o1tLtdWhco31ggUL6uzFVq5cueDFF2u4a2HMz0R7TuSvRxoxvX9b8k/vR3oN4VTGRfwGjCDv9D6k1xAiknLwGzCCr0/B2LZ6rjRuz7yUvkzv35bs6D1IryGcSLuA34AR166Rrl3jcGQkwzjAFkNfLj6ygPDUPJ7MXsZSnkY+MpcjKZfpm7DYVA5PzWP8+eV8Vu64f8InpnJ4ah4TLqyocLxf4rXj69NcmZX9T9qe3c5s41xc+k3lSMplnjy3rMJrjDr7OZ9KrXzozCXGX1hBVq+/0mvye0Sm5dE/KZBj7V+h7/MfEJmWh3/CJ1r5hQ9IPB2Dw6XTOPQYz59HDsI6LYSvsrwY3MGZ8WPH37Ls4tGewxFHcWjZhS5j/wbAT/nduOTsg2tuDB7dA9h0sTVO7f3JcHoI3zZNOJJ8EV2HzhXKAzu6YjCCb5smDOnkxkOtGhOVlotvmyb4tmlCr9baf4HrygCtmtSvUFaUe1lIQg4fbotlXM+W/Dcyg92nsvlTLw9srC2u81G5Q++8807mggULVlZ1zOJawCHukyt05axymMvEcuOAlct9dC4MXas1oVZMub1zQrODeP7EHAJGjGNuf911LbvaLofoXNi4dj8CeH7KFPx1rrc8Z2uiI4vOTmaC1AMgjHpC27+GMFZd3tvxbTzcHsM1NwaAo62ep3lrOGqE3rdR9te5EvLnlexNy6VT6e9Ca4Fea4VWbpHeqgyqdarcvyqvvNW4vi1Ldycwasl+tr46ANvSEFZzEB5cFjcJ61aL41c1yWb+xijgWvfkrc6p6/Kd1BHU/7iKci+r/LcKYObaMLbGZNFX50LQtN6EJp1XY8P3uZtNwrK4AFYURbFkf/spiu+PpDLB14PfjmWqvZ7vc/fVLGhFURRL9u9xPkzw9WBDWBrDu7qxZFc8++O05/nLuq19PNRkrQeBxY0BK4qiWLKQhBx+P5nN7AAvvj54hrxCPc9/dYRRPu7siMmqsFNW2ferFvH9SbWAFUVR6kjlHZOWTe5Jg3o22FoJ/huZgY0VLN4Zx57YbECbGzJjbXiFFrFaTev+oQJYURSljlS1JeKcIV4IIRja2Y0rJUaK9AZe+z6Sj3fE8mtkRoXzVSDfX1QXtKIoSh2p3I0ckpDDsj2JrH7eV3vULz6HmesiGNSxGUuC42nWsB4FxVoX9ZBObuw7nY0oXe8dqLDUavlrqi5ry6BawIqiKGZyXYvYy5VZATq2RWtbZ14pNtCzlTNIyZbjmRQUGQjo2IxZ64/ynx2x/Hbs5i3k5XsTWLU/oUILWbWY7x2qBawoimImt2oR99G5MGNtOPVsrRnd3Y3NxzNp4+pIa5f6LAmOB8DNqR5//vIwPVo5E5V+CRura+0qaytMG5bA9S3m21mToKxeqlVd+1QAK4qi3CMqt4jLjPJx54OnfBjXy8O049e0R9qyPjSF1k3qU1hi4HDyBQBmDW7LrPVHGdKpGT9FpNOxeUP+s+M0f8SdJzTp/E0DOiQhx7Q+++juLa7bUQyqv5BR2XVVgF9PBbCiKMo9onJARaXlmrbSLG+Ujzt/H9WZRzs1MwXyy4N0rAtNobuHM1ZCsCQ4nlZNHCjWGynWG9l7WnvWeMFob2atP8rDbZzZEZOFawM7/r01lh/D0jhz/gofT+hOY0c7ZgZFEODdjO3RmQgrKwpLDEgprwvtW5VrGuB3Y3VDuDfeFKgxYEVRlHvUSwN1FcK3LJAr7/o1ysedN4Z7s2xyT+b+EMlXIcnMDvCioMjAs71b0djBjmmPtMXJ3obWLvWZ3NuT7TFZdG7hxEOezrg0sCM2Kx+DlHTzaIy/zpXOLZz4+Wg6V0qtOgt1AAAI7ElEQVSMFBTpmbomjHbzt/DZ7gTeGunN4p1x9PtwFx9ti6VLCycWbT/NuGUhfF56/NNd8cwMCueXo+kYjJLknALOnC8wBXTZfhTVLZe10jdFZRKSkHNH17hXxspVC1hRFMVCVKeFPLd0C9D3y200U77FPDvAi6DQFJ7s0YLwMxdN5dSLV0i9eIWTGZd5qkdLdpzIYqJfK2IyLnMg4TwT/TyZ3l9HSPwFdsdm49rAjtzCEgDCz1zkL4O9mN5fxy9HM9gafdZUp7d+jsbaCho52PHWSG8Wbj/NR1tj0RslDe1t+HDLKb45eIaCIgNvjfTm4x1xrA9NIe3iVTq3cGLh9tPsiM4iIaeAFVN6setkNlPXHKG7h7ajWoB3Uz75PY68q3qCQlOY4OvBJ7/HEX7mEn/E5TDe14NPd8WTd1XPpqhMDEZJfHY+Tva2nM8v4ov9Scx+1Auo+Lz23VSjABZCDAcCAWvgCynlh7VSq5v4MvpLurp0xc/dz1S2ETbopZ6pXadeVwZ4J+QdEPDPvv+8rXPquqzqaLl1BDiceZjo89GmsqLUlVsFcuU9tcvcKKDLJn0BpuuEJORcF9rOjrYcS7tkKj/XtzXL9iQyY4AnQaEpONW3If3SVSb2bsWmY5m8MtiLY2m5bDmeyeTeWoCHJV9ke0wWXVs60cGtIVFpl4jPLmB2gBbgm45lciwtlyb1bSko0lPP2oojpW8U/HWuLA2Op7DESGiSNva969Q5mjWsx5LgeGYHeLHzZDYFRQa2lb4J2BCWRsvGDqbj34el8o//xlS4L5/vSUBKCApNqZMNMu64C1oIYQ18BowAOgPPCiE611bFbqSrS1fm7Z3H4czDANgIGxaFLcJG2FRZPpx5mG3J29iWtO22z6nrsqqj5dbxcOZh5u2dR1eXriiKuVXusn5poI7p/XWmoK7chV05oP11rozycWeUj/sNW9UzB7Xj/c2nmDmo3U3LrwzW8f5YH5ZP6cXS3dp612WBvWp/AkeStTDNuFRIJ/eGXCgoqXA89eJVZgd4gRBM6uOJrY2V6XhIQg6zArxo4mjHrMFeONe35fXHOlBiMJq+Z2jnZjRysGVSb0+c7G2Y2q8N+UV60/EZA3SM6d4CgMe7NWfpxB6M6d6CJcHxTO7tWSe7U9WkBewHxEspEwGEEN8BTwAnaqNiN3xRdz8WDVzEvL3zmNBxAhtiNzDPdx5fHP+Cy8WXqywHDg4EqNY5dV1WdbTcOi4auMjUI6Mo97Lb2aO78vjyrVrVtyqXuVGru7rlqlrpjerf/Jw2rvVveI2yQO7h2ZjdsdfeJPTRudz1EK5JALcEUsuV09D2br/r/Nz9mNBxAiuiVjDDZwbPdXmOy8WXb1gu++NYnXPquqzqaLl1VOGr3M9uFdq3Ktc0wG/USi/7/E6uUeZmXfF1sU9zTQJYVPG16zYXFkK8CLwI4OnpWYOXu+Zw5mE2xG5ghs8MNsRuwMnO6aZlv+baH8jqnFPXZVVHy62jX3M/FcKKcgM1DfDbaaXX9psCf50rSyf2ICot9+62gqWUd/QB9AW2lyvPB+bf7JxevXrJmgrNCJX9v+0vQzNCpZRSfh39tey2ppv8OvrrKsuhGaGyz7o+ss+6Prd9Tl2XVR0tt46V/z0qiqKUB4TJG2RiTVrAR4D2Qoi2QDrwDDCxZm8Hbi36fHSFMTe91DPPdx56qa+y7Ofux/A2w0Fw2+fUdVnV0XLrWDYnIfp8tGoFK4pSLUIL6Ds8WYjHgU/QHkP6Ukr53s2+39fXV4aFhd3x6ymKoiiKJRFChEspfas6VqPngKWUW4AtNbmGoiiKojyI1FKUiqIoimIGKoAVRVEUxQxUACuKoiiKGagAVhRFURQzUAGsKIqiKGagAlhRFEVRzEAFsKIoiqKYQY0W4qj2iwlxDjhTi5d0BXJq8XoPInUPa4e6jzWn7mHNqXtYc7V9D1tLKZtWdaBOA7i2CSHCbrTCiHJ71D2sHeo+1py6hzWn7mHN1eU9VF3QiqIoimIGKoAVRVEUxQwsPYBXmrsC9wF1D2uHuo81p+5hzal7WHN1dg8tegxYURRFUSyVpbeAFUVRFMUiWWwACyGGCyFihRDxQog3zV0fSyOE+FIIkS2EiDZ3XSyVEKKVEGK3EOKkECJGCDHH3HWyNEIIeyHEYSHEsdJ7+I6562SphBDWQoijQohN5q6LpRJCJAshjgshIoUQd33zeovsghZCWAOngaFAGnAEeFZKecKsFbMgQogBQD7wjZSyq7nrY4mEEO6Au5QyQgjREAgHnlT/Dm+fEEIAjlLKfCGELfAHMEdKecjMVbM4Qoi5gC/gJKUcZe76WCIhRDLgK6Wsk2epLbUF7AfESykTpZTFwHfAE2auk0WRUu4DLpi7HpZMSpkppYwo/TwPOAm0NG+tLIvU5JcWbUs/LK9VYGZCCA9gJPCFueui3D5LDeCWQGq5chrqD59iRkKINkAPINS8NbE8pV2nkUA2sFNKqe5h9X0CvAEYzV0RCyeBHUKIcCHEi3f7xSw1gEUVX1PvmhWzEEI0AH4CXpVSXjZ3fSyNlNIgpXwI8AD8hBBqSKQahBCjgGwpZbi563If6Cel7AmMAF4pHaq7ayw1gNOAVuXKHkCGmeqiPMBKxy1/AtZJKTeauz6WTEp5CdgDDDdzVSxNP2BM6fjld0CAECLIvFWyTFLKjNL/ZgM/ow133jWWGsBHgPZCiLZCCDvgGeBXM9dJecCUTiBaDZyUUn5s7vpYIiFEUyFE49LPHYAhwCnz1sqySCnnSyk9pJRt0P4WBkspJ5u5WhZHCOFYOpkSIYQjMAy4q0+JWGQASyn1wCxgO9rElw1Syhjz1sqyCCG+BQ4CHYUQaUKIaeaukwXqB0xBa3FEln48bu5KWRh3YLcQIgrtjfVOKaV6jEYxBzfgDyHEMeAwsFlKue1uvqBFPoakKIqiKJbOIlvAiqIoimLpVAAriqIoihmoAFYURVEUM1ABrCiKoihmoAJYURRFUcxABbCiKIqimIEKYEVRFEUxAxXAiqIoimIG/w9xSprMhZfaEQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "out = 100*eval_loss(x_test,cnn0.predict(x_test),norm=norm2)\n",
    "\n",
    "fig,ax = plt.subplots(figsize=(8,5))\n",
    "ax.plot(Vs,np.max(out)*xi/xi.max(),\"x--\",label=\"entanglement entropy\") # normalized to loss output\n",
    "ax.plot(Vs,out,\"x--\",label=\"out [%]\")\n",
    "ax.plot(Vs_train,out[np.where(Vs<2)]/np.max(out),\"x\",label=\"train data\")\n",
    "ax.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Expected Output:**\n",
    "\n",
    "<img src=\"Images/Phase_Boundaries.png\">"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}