paperSource/Automatic_subj_9_training-only4inputs_v1.1-sliced-noVal-seed11.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "ccc04654",
   "metadata": {},
   "source": [
    "## Var name setting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "d887cfd2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input file:  inputFeatures9_sorted.mat\n"
     ]
    }
   ],
   "source": [
    "subjectID='9'\n",
    "noEpochs=1000\n",
    "noLSTMUnits=100\n",
    "\n",
    "# Change to true if input file 's features are sorted by time (False by default) \n",
    "orderedInput=True\n",
    "\n",
    "# Change at least one of these 2 variable to true for reading inputFeatures (Can be both)\n",
    "fourInputFeats=True\n",
    "onlyVel=False\n",
    "\n",
    "HRInputFeats=False\n",
    "assert onlyVel==True or fourInputFeats == True or HRInputFeats==True, \"Both condition are false!! Need at least one of these 2 variables to true\"\n",
    "\n",
    "# Change number of steps according to Matlab file for testing input data\n",
    "test_noTimeSteps=50 \n",
    "\n",
    "# Change to 1 if want to export to CSV O2 precict against test (0 by default)\n",
    "exportCSV=1\n",
    "\n",
    "# Chabge to 1 if want to save the training model (0 by default)\n",
    "savedModel=1\n",
    "\n",
    "# Set seed value for uniform (NOT RECOMMEND TO CHANGE, by default 12321)\n",
    "seed_value = 11\n",
    "\n",
    "# Getting input file name\n",
    "if orderedInput:\n",
    "    orderedPrefix = 'order-'\n",
    "    sortedPrefix = '_sored'\n",
    "    fileName='inputFeatures'+subjectID+'_sorted.mat'\n",
    "else:    \n",
    "    # Unordered input features\n",
    "    orderedPrefix = ''\n",
    "    sortedPrefix = ''\n",
    "    fileName='inputFeatures'+subjectID+'.mat'\n",
    "\n",
    "## For testinng purpose\n",
    "#fileName=\"inputFeatures29_sorted.mat\"\n",
    "#fileName='inputFeatures9_sorted_99.mat'\n",
    "print(\"Input file: \",fileName)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d64db4a1",
   "metadata": {},
   "source": [
    "## Include lib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "6f273d91",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import load_model\n",
    "import scipy.io as sio \n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os\n",
    "import random\n",
    "\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from tensorflow.keras.models import Model\n",
    "from keras.models import Sequential\n",
    "from keras.optimizers import Adam\n",
    "from keras.layers import LSTM\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Input, Bidirectional\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.python.client import device_lib\n",
    "from datetime import datetime as dt\n",
    "from keras import backend as K\n",
    "from keras.callbacks import EarlyStopping\n",
    "\n",
    "#from comet_ml import Experiment"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d232b155",
   "metadata": {},
   "source": [
    "## Set seed for training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "d846f5ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ['PYTHONHASHSEED']=str(seed_value)\n",
    "random.seed(seed_value)\n",
    "np.random.seed(seed_value)\n",
    "tf.random.set_seed(seed_value)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48fdef95",
   "metadata": {},
   "source": [
    "## Defilne parser function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "e95b0608",
   "metadata": {},
   "outputs": [],
   "source": [
    "list_test=[]\n",
    "\n",
    "\"\"\"\n",
    "The function take number of feature according to data \n",
    "and return a list normalized input features data\n",
    "\"\"\"\n",
    "def parse_input(noFeatures):\n",
    "    global list_test\n",
    "    print(\"Number of features: \",noFeatures)\n",
    "    if noFeatures==5:\n",
    "        print(\"Keep all data\")\n",
    "        for i in range(len(test)):\n",
    "            temp_trainData=np.transpose(np.array(test[i]))\n",
    "            list_test.append(temp_trainData)\n",
    "        for j in range(len(test2)):  \n",
    "            temp_testData=np.transpose(np.array(test2[j]))\n",
    "            list_test.append(temp_testData)\n",
    "    elif noFeatures==4:\n",
    "        print(\"Remove HR data\")\n",
    "        for i in range(len(test)):\n",
    "            temp_trainData=np.transpose(np.array(test[i]))\n",
    "            list_test.append(temp_trainData[:,:4])\n",
    "        for j in range(len(test2)):  \n",
    "            temp_testData=np.transpose(np.array(test2[j]))\n",
    "            list_test.append(temp_testData[:,:4])\n",
    "    elif noFeatures==1 and onlyVel==1:\n",
    "        print(\"Only vel feature\")\n",
    "        for i in range(len(test)):\n",
    "            temp_trainData=np.transpose(np.array(test[i]))\n",
    "            list_test.append(temp_trainData[:,[1]])\n",
    "        for j in range(len(test2)):  \n",
    "            temp_testData=np.transpose(np.array(test2[j]))\n",
    "            list_test.append(temp_testData[:,[1]])\n",
    "    elif noFeatures==1 and onlyVel==0:\n",
    "        print(\"Only HR data\")\n",
    "        for i in range(len(test)):\n",
    "            temp_trainData=np.transpose(np.array(test[i]))\n",
    "            list_test.append(temp_trainData[:,[-1]])\n",
    "        for j in range(len(test2)):  \n",
    "            temp_testData=np.transpose(np.array(test2[j]))\n",
    "            list_test.append(temp_testData[:,[-11]])\n",
    "    else:\n",
    "        # Code will runinto this part\n",
    "        print(\"Unknown number of features\")\n",
    "        \n",
    "    print(\"---------\")\n",
    "    print(np.shape(list_test))    \n",
    "    list_test=np.vstack(list_test)\n",
    "    print(np.shape(list_test))\n",
    "    scaler = StandardScaler()\n",
    "    scaler=scaler.fit(list_test)\n",
    "    scaled_list_test=scaler.transform(list_test)\n",
    "    \n",
    "    return scaled_list_test"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a9410f08",
   "metadata": {},
   "source": [
    "## Read input from Matlab file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "9be7b2e8",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5, 50)\n",
      "(5, 50)\n",
      "The time step now should be  50\n",
      "4 input features included: speed, speedChange, stepDuration, vertOscillation_dist_amp\n"
     ]
    }
   ],
   "source": [
    "gpu_devices = tf.config.experimental.list_physical_devices('GPU')\n",
    "for device in gpu_devices:\n",
    "    tf.config.experimental.set_memory_growth(device, True)\n",
    "\n",
    "i=sio.loadmat(fileName)\n",
    "\n",
    "TTest=i['TTest']\n",
    "TTrain=i['TTrain']\n",
    "XTrain=i['XTrain']\n",
    "XTest=i['XTest']\n",
    "\n",
    "XTrain=XTrain.reshape(-1)\n",
    "TTrain=TTrain.reshape(-1)\n",
    "XTest=XTest.reshape(-1)\n",
    "TTest=TTest.reshape(-1)\n",
    "\n",
    "test=np.transpose(XTrain).reshape(-1)\n",
    "test2=np.transpose(XTest).reshape(-1)\n",
    "\n",
    "print(np.shape(test[0]))\n",
    "print(np.shape(test2[0]))\n",
    "\n",
    "# Get number of features and number of number of time steps\n",
    "#noFeatures = np.shape(test[0])[0]\n",
    "noTimeSteps = np.shape(test[0])[1]\n",
    "#print(noFeatures)\n",
    "\n",
    "# Validate the input data's number of feature and number of time step\n",
    "assert noTimeSteps == test_noTimeSteps, \"The number of time steps does not match our configuration!! Check either of them\"\n",
    "print(\"The time step now should be \",noTimeSteps)\n",
    "\n",
    "if fourInputFeats and HRInputFeats:\n",
    "    #assert noFeatures==5, \"The number of input features is not 5! Check setting again\"\n",
    "    print(\"5 input features included: speed, speedChange, stepDuration, vertOscillation_dist_amp, heart rate\")\n",
    "    confPrefix=\"withHR_\"\n",
    "    noFeatures=5\n",
    "elif fourInputFeats:\n",
    "    #assert noFeatures==4, \"The number of input features is not 4! Check setting again\"\n",
    "    print(\"4 input features included: speed, speedChange, stepDuration, vertOscillation_dist_amp\")\n",
    "    confPrefix=\"NoHR_\"\n",
    "    noFeatures=4\n",
    "elif HRInputFeats:\n",
    "    print(\"1 input feture included: heart rate\")\n",
    "    confPrefix=\"onlyHR_\"\n",
    "elif onlyVel:\n",
    "    #assert noFeatures==1, \"The number of input features is not 1! Check setting again\"\n",
    "    print(\"1 input feture included: speed\")\n",
    "    confPrefix=\"onlyVel_\"\n",
    "    noFeatures=1\n",
    "else:\n",
    "    ## Both conditions are false, code will not run into this section\n",
    "    print(\"Both condition are false!! Need at least \")\n",
    "    \n",
    "# Automatically setting name of output files based on configuration    \n",
    "modelSaveName=confPrefix+'Subj-'+subjectID+'-'+str(noEpochs)+'pochs-'+orderedPrefix+str(noLSTMUnits)+'-LSTM_u.h5'\n",
    "figPredictName=confPrefix+'Subj'+subjectID+'-'+str(noEpochs)+'pochs-'+orderedPrefix+str(noLSTMUnits)+'-LSTM_u.jpg'\n",
    "figRMSEName=confPrefix+'RMSE_Subj'+subjectID+'-'+str(noEpochs)+'pochs-'+orderedPrefix+str(noLSTMUnits)+'-LSTM_u.jpg'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4c78ae0",
   "metadata": {},
   "source": [
    "## Call parser function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "3d2a27fe",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of features:  4\n",
      "Remove HR data\n",
      "---------\n",
      "(1344, 50, 4)\n",
      "(67200, 4)\n"
     ]
    }
   ],
   "source": [
    "scaled_list_test = parse_input(noFeatures)\n",
    "\n",
    "# Return back input data with scaled value\n",
    "XT_s_stack=list()\n",
    "i=0\n",
    "timeSteps=noTimeSteps\n",
    "while i+timeSteps <= (len(scaled_list_test)):\n",
    "    XT_s_stack.append(scaled_list_test[i:i+timeSteps])\n",
    "    i=i+timeSteps\n",
    "\n",
    "XT_s_stack=np.stack(XT_s_stack)\n",
    "cutPoint=np.shape(TTest.reshape(-1))[0]\n",
    "\n",
    "\n",
    "XTrain_s=XT_s_stack[:-cutPoint,:,:]\n",
    "XTest_s=XT_s_stack[-cutPoint:,:,:]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55305abc",
   "metadata": {},
   "source": [
    "## Configure the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "d6ba2216",
   "metadata": {},
   "outputs": [],
   "source": [
    "es = EarlyStopping(monitor='val_loss',mode='min', verbose=1,patience=20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "91546b25",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " lstm (LSTM)                 (None, 100)               42000     \n",
      "                                                                 \n",
      " dense (Dense)               (None, 1)                 101       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 42,101\n",
      "Trainable params: 42,101\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "length = noTimeSteps\n",
    "n_features = noFeatures\n",
    "model = Sequential()\n",
    "model.add(LSTM(noLSTMUnits,input_shape=(length,n_features)))\n",
    "model.add(Dense(1))\n",
    "opt=Adam(learning_rate=0.005)\n",
    "model.compile(optimizer=opt, loss='mean_squared_logarithmic_error')\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "66f31e4d",
   "metadata": {},
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "9b3c4d1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "#print(np.shape(XTrain_s))\n",
    "#print(np.isnan(XTrain_s))\n",
    "#print(XTrain_s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "d6fcf1ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "# print(np.shape(TTrain))\n",
    "# TTrain[np.isnan(TTrain)]=0\n",
    "# print(TTrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "9a62b3fc",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "34/34 - 5s - loss: 3.5084 - 5s/epoch - 150ms/step\n",
      "Epoch 2/1000\n",
      "34/34 - 1s - loss: 1.9469 - 1s/epoch - 33ms/step\n",
      "Epoch 3/1000\n",
      "34/34 - 1s - loss: 0.3968 - 1s/epoch - 34ms/step\n",
      "Epoch 4/1000\n",
      "34/34 - 1s - loss: 0.2216 - 1s/epoch - 36ms/step\n",
      "Epoch 5/1000\n",
      "34/34 - 1s - loss: 0.2209 - 1s/epoch - 37ms/step\n",
      "Epoch 6/1000\n",
      "34/34 - 1s - loss: 0.2205 - 1s/epoch - 36ms/step\n",
      "Epoch 7/1000\n",
      "34/34 - 1s - loss: 0.2199 - 1s/epoch - 34ms/step\n",
      "Epoch 8/1000\n",
      "34/34 - 1s - loss: 0.2188 - 1s/epoch - 33ms/step\n",
      "Epoch 9/1000\n",
      "34/34 - 1s - loss: 0.2177 - 1s/epoch - 34ms/step\n",
      "Epoch 10/1000\n",
      "34/34 - 1s - loss: 0.2142 - 1s/epoch - 37ms/step\n",
      "Epoch 11/1000\n",
      "34/34 - 1s - loss: 0.2081 - 1s/epoch - 36ms/step\n",
      "Epoch 12/1000\n",
      "34/34 - 1s - loss: 0.1960 - 1s/epoch - 33ms/step\n",
      "Epoch 13/1000\n",
      "34/34 - 1s - loss: 0.1812 - 1s/epoch - 39ms/step\n",
      "Epoch 14/1000\n",
      "34/34 - 1s - loss: 0.1563 - 1s/epoch - 37ms/step\n",
      "Epoch 15/1000\n",
      "34/34 - 1s - loss: 0.1412 - 1s/epoch - 35ms/step\n",
      "Epoch 16/1000\n",
      "34/34 - 1s - loss: 0.1134 - 1s/epoch - 35ms/step\n",
      "Epoch 17/1000\n",
      "34/34 - 1s - loss: 0.1221 - 1s/epoch - 40ms/step\n",
      "Epoch 18/1000\n",
      "34/34 - 1s - loss: 0.1200 - 1s/epoch - 43ms/step\n",
      "Epoch 19/1000\n",
      "34/34 - 1s - loss: 0.0878 - 1s/epoch - 36ms/step\n",
      "Epoch 20/1000\n",
      "34/34 - 1s - loss: 0.0848 - 1s/epoch - 33ms/step\n",
      "Epoch 21/1000\n",
      "34/34 - 1s - loss: 0.0821 - 1s/epoch - 42ms/step\n",
      "Epoch 22/1000\n",
      "34/34 - 1s - loss: 0.0809 - 1s/epoch - 41ms/step\n",
      "Epoch 23/1000\n",
      "34/34 - 1s - loss: 0.0806 - 1s/epoch - 34ms/step\n",
      "Epoch 24/1000\n",
      "34/34 - 1s - loss: 0.0789 - 1s/epoch - 32ms/step\n",
      "Epoch 25/1000\n",
      "34/34 - 1s - loss: 0.0782 - 1s/epoch - 33ms/step\n",
      "Epoch 26/1000\n",
      "34/34 - 1s - loss: 0.0768 - 1s/epoch - 36ms/step\n",
      "Epoch 27/1000\n",
      "34/34 - 1s - loss: 0.0766 - 1s/epoch - 35ms/step\n",
      "Epoch 28/1000\n",
      "34/34 - 1s - loss: 0.0755 - 1s/epoch - 35ms/step\n",
      "Epoch 29/1000\n",
      "34/34 - 1s - loss: 0.0752 - 1s/epoch - 35ms/step\n",
      "Epoch 30/1000\n",
      "34/34 - 1s - loss: 0.0860 - 1s/epoch - 35ms/step\n",
      "Epoch 31/1000\n",
      "34/34 - 1s - loss: 0.0992 - 1s/epoch - 37ms/step\n",
      "Epoch 32/1000\n",
      "34/34 - 1s - loss: 0.0833 - 1s/epoch - 37ms/step\n",
      "Epoch 33/1000\n",
      "34/34 - 1s - loss: 0.0800 - 1s/epoch - 34ms/step\n",
      "Epoch 34/1000\n",
      "34/34 - 1s - loss: 0.0816 - 1s/epoch - 35ms/step\n",
      "Epoch 35/1000\n",
      "34/34 - 1s - loss: 0.0852 - 1s/epoch - 34ms/step\n",
      "Epoch 36/1000\n",
      "34/34 - 1s - loss: 0.0785 - 1s/epoch - 34ms/step\n",
      "Epoch 37/1000\n",
      "34/34 - 1s - loss: 0.0783 - 1s/epoch - 34ms/step\n",
      "Epoch 38/1000\n",
      "34/34 - 1s - loss: 0.0768 - 1s/epoch - 32ms/step\n",
      "Epoch 39/1000\n",
      "34/34 - 1s - loss: 0.0754 - 1s/epoch - 33ms/step\n",
      "Epoch 40/1000\n",
      "34/34 - 1s - loss: 0.0754 - 1s/epoch - 34ms/step\n",
      "Epoch 41/1000\n",
      "34/34 - 1s - loss: 0.0746 - 1s/epoch - 33ms/step\n",
      "Epoch 42/1000\n",
      "34/34 - 1s - loss: 0.0753 - 1s/epoch - 40ms/step\n",
      "Epoch 43/1000\n",
      "34/34 - 1s - loss: 0.0759 - 1s/epoch - 40ms/step\n",
      "Epoch 44/1000\n",
      "34/34 - 1s - loss: 0.0765 - 1s/epoch - 37ms/step\n",
      "Epoch 45/1000\n",
      "34/34 - 1s - loss: 0.0733 - 1s/epoch - 42ms/step\n",
      "Epoch 46/1000\n",
      "34/34 - 2s - loss: 0.0729 - 2s/epoch - 44ms/step\n",
      "Epoch 47/1000\n",
      "34/34 - 1s - loss: 0.0738 - 1s/epoch - 37ms/step\n",
      "Epoch 48/1000\n",
      "34/34 - 1s - loss: 0.0763 - 1s/epoch - 34ms/step\n",
      "Epoch 49/1000\n",
      "34/34 - 1s - loss: 0.0738 - 1s/epoch - 33ms/step\n",
      "Epoch 50/1000\n",
      "34/34 - 1s - loss: 0.0730 - 1s/epoch - 34ms/step\n",
      "Epoch 51/1000\n",
      "34/34 - 1s - loss: 0.0717 - 1s/epoch - 34ms/step\n",
      "Epoch 52/1000\n",
      "34/34 - 1s - loss: 0.0730 - 1s/epoch - 34ms/step\n",
      "Epoch 53/1000\n",
      "34/34 - 1s - loss: 0.0722 - 1s/epoch - 34ms/step\n",
      "Epoch 54/1000\n",
      "34/34 - 1s - loss: 0.0694 - 1s/epoch - 33ms/step\n",
      "Epoch 55/1000\n",
      "34/34 - 1s - loss: 0.0702 - 1s/epoch - 34ms/step\n",
      "Epoch 56/1000\n",
      "34/34 - 1s - loss: 0.0717 - 1s/epoch - 34ms/step\n",
      "Epoch 57/1000\n",
      "34/34 - 1s - loss: 0.0709 - 1s/epoch - 34ms/step\n",
      "Epoch 58/1000\n",
      "34/34 - 1s - loss: 0.0717 - 1s/epoch - 33ms/step\n",
      "Epoch 59/1000\n",
      "34/34 - 1s - loss: 0.0696 - 1s/epoch - 35ms/step\n",
      "Epoch 60/1000\n",
      "34/34 - 1s - loss: 0.0680 - 1s/epoch - 33ms/step\n",
      "Epoch 61/1000\n",
      "34/34 - 1s - loss: 0.1175 - 1s/epoch - 33ms/step\n",
      "Epoch 62/1000\n",
      "34/34 - 1s - loss: 0.5549 - 1s/epoch - 33ms/step\n",
      "Epoch 63/1000\n",
      "34/34 - 1s - loss: 0.1890 - 1s/epoch - 33ms/step\n",
      "Epoch 64/1000\n",
      "34/34 - 1s - loss: 0.1615 - 1s/epoch - 35ms/step\n",
      "Epoch 65/1000\n",
      "34/34 - 1s - loss: 0.1068 - 1s/epoch - 35ms/step\n",
      "Epoch 66/1000\n",
      "34/34 - 1s - loss: 0.0939 - 1s/epoch - 36ms/step\n",
      "Epoch 67/1000\n",
      "34/34 - 1s - loss: 0.1062 - 1s/epoch - 35ms/step\n",
      "Epoch 68/1000\n",
      "34/34 - 1s - loss: 0.1368 - 1s/epoch - 35ms/step\n",
      "Epoch 69/1000\n",
      "34/34 - 1s - loss: 0.1417 - 1s/epoch - 39ms/step\n",
      "Epoch 70/1000\n",
      "34/34 - 2s - loss: 0.1069 - 2s/epoch - 45ms/step\n",
      "Epoch 71/1000\n",
      "34/34 - 1s - loss: 0.0959 - 1s/epoch - 41ms/step\n",
      "Epoch 72/1000\n",
      "34/34 - 1s - loss: 0.0892 - 1s/epoch - 42ms/step\n",
      "Epoch 73/1000\n",
      "34/34 - 1s - loss: 0.0855 - 1s/epoch - 39ms/step\n",
      "Epoch 74/1000\n",
      "34/34 - 1s - loss: 0.0938 - 1s/epoch - 37ms/step\n",
      "Epoch 75/1000\n",
      "34/34 - 1s - loss: 0.0909 - 1s/epoch - 39ms/step\n",
      "Epoch 76/1000\n",
      "34/34 - 1s - loss: 0.0825 - 1s/epoch - 34ms/step\n",
      "Epoch 77/1000\n",
      "34/34 - 1s - loss: 0.0796 - 1s/epoch - 35ms/step\n",
      "Epoch 78/1000\n",
      "34/34 - 1s - loss: 0.0784 - 1s/epoch - 35ms/step\n",
      "Epoch 79/1000\n",
      "34/34 - 1s - loss: 0.0777 - 1s/epoch - 37ms/step\n",
      "Epoch 80/1000\n",
      "34/34 - 1s - loss: 0.0822 - 1s/epoch - 36ms/step\n",
      "Epoch 81/1000\n",
      "34/34 - 1s - loss: 0.0793 - 1s/epoch - 36ms/step\n",
      "Epoch 82/1000\n",
      "34/34 - 1s - loss: 0.0812 - 1s/epoch - 35ms/step\n",
      "Epoch 83/1000\n",
      "34/34 - 1s - loss: 0.0847 - 1s/epoch - 36ms/step\n",
      "Epoch 84/1000\n",
      "34/34 - 1s - loss: 0.0851 - 1s/epoch - 36ms/step\n",
      "Epoch 85/1000\n",
      "34/34 - 1s - loss: 0.1023 - 1s/epoch - 37ms/step\n",
      "Epoch 86/1000\n",
      "34/34 - 1s - loss: 0.1318 - 1s/epoch - 36ms/step\n",
      "Epoch 87/1000\n",
      "34/34 - 1s - loss: 0.0947 - 1s/epoch - 38ms/step\n",
      "Epoch 88/1000\n",
      "34/34 - 1s - loss: 0.0931 - 1s/epoch - 36ms/step\n",
      "Epoch 89/1000\n",
      "34/34 - 1s - loss: 0.0894 - 1s/epoch - 41ms/step\n",
      "Epoch 90/1000\n",
      "34/34 - 1s - loss: 0.0881 - 1s/epoch - 42ms/step\n",
      "Epoch 91/1000\n",
      "34/34 - 1s - loss: 0.0847 - 1s/epoch - 36ms/step\n",
      "Epoch 92/1000\n",
      "34/34 - 1s - loss: 0.0830 - 1s/epoch - 37ms/step\n",
      "Epoch 93/1000\n",
      "34/34 - 1s - loss: 0.0828 - 1s/epoch - 35ms/step\n",
      "Epoch 94/1000\n",
      "34/34 - 1s - loss: 0.0801 - 1s/epoch - 37ms/step\n",
      "Epoch 95/1000\n",
      "34/34 - 1s - loss: 0.0789 - 1s/epoch - 39ms/step\n",
      "Epoch 96/1000\n",
      "34/34 - 1s - loss: 0.0806 - 1s/epoch - 37ms/step\n",
      "Epoch 97/1000\n",
      "34/34 - 1s - loss: 0.0824 - 1s/epoch - 39ms/step\n",
      "Epoch 98/1000\n",
      "34/34 - 1s - loss: 0.0780 - 1s/epoch - 34ms/step\n",
      "Epoch 99/1000\n",
      "34/34 - 1s - loss: 0.0750 - 1s/epoch - 34ms/step\n",
      "Epoch 100/1000\n",
      "34/34 - 1s - loss: 0.0707 - 1s/epoch - 34ms/step\n",
      "Epoch 101/1000\n",
      "34/34 - 1s - loss: 0.0756 - 1s/epoch - 33ms/step\n",
      "Epoch 102/1000\n",
      "34/34 - 1s - loss: 0.0782 - 1s/epoch - 34ms/step\n",
      "Epoch 103/1000\n",
      "34/34 - 1s - loss: 0.0695 - 1s/epoch - 34ms/step\n",
      "Epoch 104/1000\n",
      "34/34 - 1s - loss: 0.0717 - 1s/epoch - 34ms/step\n",
      "Epoch 105/1000\n",
      "34/34 - 1s - loss: 0.0656 - 1s/epoch - 34ms/step\n",
      "Epoch 106/1000\n",
      "34/34 - 1s - loss: 0.0683 - 1s/epoch - 34ms/step\n",
      "Epoch 107/1000\n",
      "34/34 - 1s - loss: 0.0652 - 1s/epoch - 34ms/step\n",
      "Epoch 108/1000\n",
      "34/34 - 1s - loss: 0.0662 - 1s/epoch - 34ms/step\n",
      "Epoch 109/1000\n",
      "34/34 - 1s - loss: 0.0630 - 1s/epoch - 35ms/step\n",
      "Epoch 110/1000\n",
      "34/34 - 1s - loss: 0.0710 - 1s/epoch - 34ms/step\n",
      "Epoch 111/1000\n",
      "34/34 - 1s - loss: 0.0710 - 1s/epoch - 35ms/step\n",
      "Epoch 112/1000\n",
      "34/34 - 1s - loss: 0.0654 - 1s/epoch - 34ms/step\n",
      "Epoch 113/1000\n",
      "34/34 - 1s - loss: 0.0672 - 1s/epoch - 33ms/step\n",
      "Epoch 114/1000\n",
      "34/34 - 1s - loss: 0.0665 - 1s/epoch - 34ms/step\n",
      "Epoch 115/1000\n",
      "34/34 - 1s - loss: 0.0616 - 1s/epoch - 33ms/step\n",
      "Epoch 116/1000\n",
      "34/34 - 1s - loss: 0.0588 - 1s/epoch - 34ms/step\n",
      "Epoch 117/1000\n",
      "34/34 - 1s - loss: 0.0577 - 1s/epoch - 33ms/step\n",
      "Epoch 118/1000\n",
      "34/34 - 1s - loss: 0.0583 - 1s/epoch - 35ms/step\n",
      "Epoch 119/1000\n",
      "34/34 - 1s - loss: 0.0564 - 1s/epoch - 35ms/step\n",
      "Epoch 120/1000\n",
      "34/34 - 1s - loss: 0.0530 - 1s/epoch - 35ms/step\n",
      "Epoch 121/1000\n",
      "34/34 - 1s - loss: 0.0578 - 1s/epoch - 35ms/step\n",
      "Epoch 122/1000\n",
      "34/34 - 1s - loss: 0.0556 - 1s/epoch - 35ms/step\n",
      "Epoch 123/1000\n",
      "34/34 - 1s - loss: 0.0571 - 1s/epoch - 35ms/step\n",
      "Epoch 124/1000\n",
      "34/34 - 1s - loss: 0.0532 - 1s/epoch - 35ms/step\n",
      "Epoch 125/1000\n",
      "34/34 - 1s - loss: 0.0692 - 1s/epoch - 36ms/step\n",
      "Epoch 126/1000\n",
      "34/34 - 1s - loss: 0.0651 - 1s/epoch - 36ms/step\n",
      "Epoch 127/1000\n",
      "34/34 - 1s - loss: 0.0706 - 1s/epoch - 34ms/step\n",
      "Epoch 128/1000\n",
      "34/34 - 1s - loss: 0.0592 - 1s/epoch - 36ms/step\n",
      "Epoch 129/1000\n",
      "34/34 - 1s - loss: 0.0578 - 1s/epoch - 34ms/step\n",
      "Epoch 130/1000\n",
      "34/34 - 1s - loss: 0.0568 - 1s/epoch - 33ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 131/1000\n",
      "34/34 - 1s - loss: 0.0524 - 1s/epoch - 35ms/step\n",
      "Epoch 132/1000\n",
      "34/34 - 1s - loss: 0.0496 - 1s/epoch - 34ms/step\n",
      "Epoch 133/1000\n",
      "34/34 - 1s - loss: 0.0486 - 1s/epoch - 35ms/step\n",
      "Epoch 134/1000\n",
      "34/34 - 1s - loss: 0.0491 - 1s/epoch - 34ms/step\n",
      "Epoch 135/1000\n",
      "34/34 - 1s - loss: 0.0517 - 1s/epoch - 34ms/step\n",
      "Epoch 136/1000\n",
      "34/34 - 1s - loss: 0.0561 - 1s/epoch - 35ms/step\n",
      "Epoch 137/1000\n",
      "34/34 - 1s - loss: 0.0474 - 1s/epoch - 35ms/step\n",
      "Epoch 138/1000\n",
      "34/34 - 1s - loss: 0.0448 - 1s/epoch - 35ms/step\n",
      "Epoch 139/1000\n",
      "34/34 - 1s - loss: 0.0462 - 1s/epoch - 35ms/step\n",
      "Epoch 140/1000\n",
      "34/34 - 1s - loss: 0.0466 - 1s/epoch - 33ms/step\n",
      "Epoch 141/1000\n",
      "34/34 - 1s - loss: 0.0591 - 1s/epoch - 33ms/step\n",
      "Epoch 142/1000\n",
      "34/34 - 1s - loss: 0.0624 - 1s/epoch - 34ms/step\n",
      "Epoch 143/1000\n",
      "34/34 - 1s - loss: 0.0576 - 1s/epoch - 35ms/step\n",
      "Epoch 144/1000\n",
      "34/34 - 1s - loss: 0.0501 - 1s/epoch - 34ms/step\n",
      "Epoch 145/1000\n",
      "34/34 - 1s - loss: 0.0528 - 1s/epoch - 36ms/step\n",
      "Epoch 146/1000\n",
      "34/34 - 1s - loss: 0.0482 - 1s/epoch - 36ms/step\n",
      "Epoch 147/1000\n",
      "34/34 - 1s - loss: 0.0467 - 1s/epoch - 35ms/step\n",
      "Epoch 148/1000\n",
      "34/34 - 1s - loss: 0.0435 - 1s/epoch - 34ms/step\n",
      "Epoch 149/1000\n",
      "34/34 - 1s - loss: 0.0406 - 1s/epoch - 34ms/step\n",
      "Epoch 150/1000\n",
      "34/34 - 1s - loss: 0.0418 - 1s/epoch - 36ms/step\n",
      "Epoch 151/1000\n",
      "34/34 - 1s - loss: 0.0572 - 1s/epoch - 34ms/step\n",
      "Epoch 152/1000\n",
      "34/34 - 1s - loss: 0.0612 - 1s/epoch - 35ms/step\n",
      "Epoch 153/1000\n",
      "34/34 - 1s - loss: 0.0457 - 1s/epoch - 35ms/step\n",
      "Epoch 154/1000\n",
      "34/34 - 1s - loss: 0.0408 - 1s/epoch - 33ms/step\n",
      "Epoch 155/1000\n",
      "34/34 - 1s - loss: 0.0399 - 1s/epoch - 35ms/step\n",
      "Epoch 156/1000\n",
      "34/34 - 1s - loss: 0.0373 - 1s/epoch - 33ms/step\n",
      "Epoch 157/1000\n",
      "34/34 - 1s - loss: 0.0372 - 1s/epoch - 34ms/step\n",
      "Epoch 158/1000\n",
      "34/34 - 1s - loss: 0.0369 - 1s/epoch - 34ms/step\n",
      "Epoch 159/1000\n",
      "34/34 - 1s - loss: 0.0375 - 1s/epoch - 36ms/step\n",
      "Epoch 160/1000\n",
      "34/34 - 1s - loss: 0.0376 - 1s/epoch - 35ms/step\n",
      "Epoch 161/1000\n",
      "34/34 - 1s - loss: 0.0356 - 1s/epoch - 34ms/step\n",
      "Epoch 162/1000\n",
      "34/34 - 1s - loss: 0.0378 - 1s/epoch - 35ms/step\n",
      "Epoch 163/1000\n",
      "34/34 - 1s - loss: 0.0375 - 1s/epoch - 35ms/step\n",
      "Epoch 164/1000\n",
      "34/34 - 1s - loss: 0.0517 - 1s/epoch - 35ms/step\n",
      "Epoch 165/1000\n",
      "34/34 - 1s - loss: 0.0493 - 1s/epoch - 35ms/step\n",
      "Epoch 166/1000\n",
      "34/34 - 1s - loss: 0.0443 - 1s/epoch - 35ms/step\n",
      "Epoch 167/1000\n",
      "34/34 - 1s - loss: 0.0365 - 1s/epoch - 34ms/step\n",
      "Epoch 168/1000\n",
      "34/34 - 1s - loss: 0.0366 - 1s/epoch - 33ms/step\n",
      "Epoch 169/1000\n",
      "34/34 - 1s - loss: 0.0352 - 1s/epoch - 35ms/step\n",
      "Epoch 170/1000\n",
      "34/34 - 1s - loss: 0.0357 - 1s/epoch - 34ms/step\n",
      "Epoch 171/1000\n",
      "34/34 - 1s - loss: 0.0342 - 1s/epoch - 34ms/step\n",
      "Epoch 172/1000\n",
      "34/34 - 1s - loss: 0.0355 - 1s/epoch - 35ms/step\n",
      "Epoch 173/1000\n",
      "34/34 - 1s - loss: 0.0344 - 1s/epoch - 35ms/step\n",
      "Epoch 174/1000\n",
      "34/34 - 1s - loss: 0.0318 - 1s/epoch - 35ms/step\n",
      "Epoch 175/1000\n",
      "34/34 - 1s - loss: 0.0315 - 1s/epoch - 35ms/step\n",
      "Epoch 176/1000\n",
      "34/34 - 1s - loss: 0.0335 - 1s/epoch - 34ms/step\n",
      "Epoch 177/1000\n",
      "34/34 - 1s - loss: 0.0385 - 1s/epoch - 35ms/step\n",
      "Epoch 178/1000\n",
      "34/34 - 1s - loss: 0.0326 - 1s/epoch - 35ms/step\n",
      "Epoch 179/1000\n",
      "34/34 - 1s - loss: 0.0308 - 1s/epoch - 35ms/step\n",
      "Epoch 180/1000\n",
      "34/34 - 1s - loss: 0.0303 - 1s/epoch - 35ms/step\n",
      "Epoch 181/1000\n",
      "34/34 - 1s - loss: 0.0318 - 1s/epoch - 33ms/step\n",
      "Epoch 182/1000\n",
      "34/34 - 1s - loss: 0.0303 - 1s/epoch - 33ms/step\n",
      "Epoch 183/1000\n",
      "34/34 - 1s - loss: 0.0291 - 1s/epoch - 34ms/step\n",
      "Epoch 184/1000\n",
      "34/34 - 1s - loss: 0.0283 - 1s/epoch - 34ms/step\n",
      "Epoch 185/1000\n",
      "34/34 - 1s - loss: 0.0362 - 1s/epoch - 34ms/step\n",
      "Epoch 186/1000\n",
      "34/34 - 1s - loss: 0.0358 - 1s/epoch - 34ms/step\n",
      "Epoch 187/1000\n",
      "34/34 - 1s - loss: 0.0382 - 1s/epoch - 35ms/step\n",
      "Epoch 188/1000\n",
      "34/34 - 1s - loss: 0.0339 - 1s/epoch - 35ms/step\n",
      "Epoch 189/1000\n",
      "34/34 - 1s - loss: 0.0311 - 1s/epoch - 35ms/step\n",
      "Epoch 190/1000\n",
      "34/34 - 1s - loss: 0.0271 - 1s/epoch - 34ms/step\n",
      "Epoch 191/1000\n",
      "34/34 - 1s - loss: 0.0303 - 1s/epoch - 35ms/step\n",
      "Epoch 192/1000\n",
      "34/34 - 1s - loss: 0.0589 - 1s/epoch - 34ms/step\n",
      "Epoch 193/1000\n",
      "34/34 - 1s - loss: 0.0830 - 1s/epoch - 35ms/step\n",
      "Epoch 194/1000\n",
      "34/34 - 1s - loss: 0.0622 - 1s/epoch - 35ms/step\n",
      "Epoch 195/1000\n",
      "34/34 - 1s - loss: 0.0584 - 1s/epoch - 33ms/step\n",
      "Epoch 196/1000\n",
      "34/34 - 1s - loss: 0.0540 - 1s/epoch - 35ms/step\n",
      "Epoch 197/1000\n",
      "34/34 - 1s - loss: 0.0543 - 1s/epoch - 34ms/step\n",
      "Epoch 198/1000\n",
      "34/34 - 1s - loss: 0.0490 - 1s/epoch - 34ms/step\n",
      "Epoch 199/1000\n",
      "34/34 - 1s - loss: 0.0504 - 1s/epoch - 35ms/step\n",
      "Epoch 200/1000\n",
      "34/34 - 1s - loss: 0.0565 - 1s/epoch - 34ms/step\n",
      "Epoch 201/1000\n",
      "34/34 - 1s - loss: 0.0588 - 1s/epoch - 35ms/step\n",
      "Epoch 202/1000\n",
      "34/34 - 1s - loss: 0.0596 - 1s/epoch - 35ms/step\n",
      "Epoch 203/1000\n",
      "34/34 - 1s - loss: 0.0617 - 1s/epoch - 35ms/step\n",
      "Epoch 204/1000\n",
      "34/34 - 1s - loss: 0.0562 - 1s/epoch - 35ms/step\n",
      "Epoch 205/1000\n",
      "34/34 - 1s - loss: 0.0517 - 1s/epoch - 35ms/step\n",
      "Epoch 206/1000\n",
      "34/34 - 1s - loss: 0.0501 - 1s/epoch - 38ms/step\n",
      "Epoch 207/1000\n",
      "34/34 - 1s - loss: 0.0493 - 1s/epoch - 36ms/step\n",
      "Epoch 208/1000\n",
      "34/34 - 1s - loss: 0.0452 - 1s/epoch - 34ms/step\n",
      "Epoch 209/1000\n",
      "34/34 - 1s - loss: 0.0424 - 1s/epoch - 34ms/step\n",
      "Epoch 210/1000\n",
      "34/34 - 1s - loss: 0.0451 - 1s/epoch - 34ms/step\n",
      "Epoch 211/1000\n",
      "34/34 - 1s - loss: 0.0432 - 1s/epoch - 34ms/step\n",
      "Epoch 212/1000\n",
      "34/34 - 1s - loss: 0.0466 - 1s/epoch - 33ms/step\n",
      "Epoch 213/1000\n",
      "34/34 - 1s - loss: 0.0464 - 1s/epoch - 37ms/step\n",
      "Epoch 214/1000\n",
      "34/34 - 1s - loss: 0.0436 - 1s/epoch - 35ms/step\n",
      "Epoch 215/1000\n",
      "34/34 - 1s - loss: 0.0386 - 1s/epoch - 35ms/step\n",
      "Epoch 216/1000\n",
      "34/34 - 1s - loss: 0.0430 - 1s/epoch - 35ms/step\n",
      "Epoch 217/1000\n",
      "34/34 - 1s - loss: 0.0349 - 1s/epoch - 34ms/step\n",
      "Epoch 218/1000\n",
      "34/34 - 1s - loss: 0.0406 - 1s/epoch - 38ms/step\n",
      "Epoch 219/1000\n",
      "34/34 - 1s - loss: 0.0376 - 1s/epoch - 35ms/step\n",
      "Epoch 220/1000\n",
      "34/34 - 1s - loss: 0.0300 - 1s/epoch - 35ms/step\n",
      "Epoch 221/1000\n",
      "34/34 - 1s - loss: 0.0351 - 1s/epoch - 36ms/step\n",
      "Epoch 222/1000\n",
      "34/34 - 1s - loss: 0.0279 - 1s/epoch - 34ms/step\n",
      "Epoch 223/1000\n",
      "34/34 - 1s - loss: 0.0299 - 1s/epoch - 34ms/step\n",
      "Epoch 224/1000\n",
      "34/34 - 1s - loss: 0.0236 - 1s/epoch - 35ms/step\n",
      "Epoch 225/1000\n",
      "34/34 - 1s - loss: 0.0231 - 1s/epoch - 34ms/step\n",
      "Epoch 226/1000\n",
      "34/34 - 1s - loss: 0.0202 - 1s/epoch - 35ms/step\n",
      "Epoch 227/1000\n",
      "34/34 - 1s - loss: 0.0204 - 1s/epoch - 35ms/step\n",
      "Epoch 228/1000\n",
      "34/34 - 1s - loss: 0.0238 - 1s/epoch - 34ms/step\n",
      "Epoch 229/1000\n",
      "34/34 - 1s - loss: 0.0182 - 1s/epoch - 35ms/step\n",
      "Epoch 230/1000\n",
      "34/34 - 1s - loss: 0.0207 - 1s/epoch - 34ms/step\n",
      "Epoch 231/1000\n",
      "34/34 - 1s - loss: 0.0212 - 1s/epoch - 35ms/step\n",
      "Epoch 232/1000\n",
      "34/34 - 1s - loss: 0.0169 - 1s/epoch - 35ms/step\n",
      "Epoch 233/1000\n",
      "34/34 - 1s - loss: 0.0199 - 1s/epoch - 35ms/step\n",
      "Epoch 234/1000\n",
      "34/34 - 1s - loss: 0.0172 - 1s/epoch - 35ms/step\n",
      "Epoch 235/1000\n",
      "34/34 - 1s - loss: 0.0152 - 1s/epoch - 34ms/step\n",
      "Epoch 236/1000\n",
      "34/34 - 1s - loss: 0.0247 - 1s/epoch - 34ms/step\n",
      "Epoch 237/1000\n",
      "34/34 - 1s - loss: 0.0170 - 1s/epoch - 34ms/step\n",
      "Epoch 238/1000\n",
      "34/34 - 1s - loss: 0.0166 - 1s/epoch - 34ms/step\n",
      "Epoch 239/1000\n",
      "34/34 - 1s - loss: 0.0153 - 1s/epoch - 33ms/step\n",
      "Epoch 240/1000\n",
      "34/34 - 1s - loss: 0.0146 - 1s/epoch - 35ms/step\n",
      "Epoch 241/1000\n",
      "34/34 - 1s - loss: 0.0142 - 1s/epoch - 35ms/step\n",
      "Epoch 242/1000\n",
      "34/34 - 1s - loss: 0.0128 - 1s/epoch - 35ms/step\n",
      "Epoch 243/1000\n",
      "34/34 - 1s - loss: 0.0125 - 1s/epoch - 35ms/step\n",
      "Epoch 244/1000\n",
      "34/34 - 1s - loss: 0.0130 - 1s/epoch - 34ms/step\n",
      "Epoch 245/1000\n",
      "34/34 - 1s - loss: 0.0128 - 1s/epoch - 36ms/step\n",
      "Epoch 246/1000\n",
      "34/34 - 1s - loss: 0.0131 - 1s/epoch - 35ms/step\n",
      "Epoch 247/1000\n",
      "34/34 - 1s - loss: 0.0129 - 1s/epoch - 37ms/step\n",
      "Epoch 248/1000\n",
      "34/34 - 1s - loss: 0.0153 - 1s/epoch - 36ms/step\n",
      "Epoch 249/1000\n",
      "34/34 - 1s - loss: 0.0145 - 1s/epoch - 33ms/step\n",
      "Epoch 250/1000\n",
      "34/34 - 1s - loss: 0.0138 - 1s/epoch - 34ms/step\n",
      "Epoch 251/1000\n",
      "34/34 - 1s - loss: 0.0162 - 1s/epoch - 35ms/step\n",
      "Epoch 252/1000\n",
      "34/34 - 1s - loss: 0.0161 - 1s/epoch - 36ms/step\n",
      "Epoch 253/1000\n",
      "34/34 - 1s - loss: 0.0164 - 1s/epoch - 35ms/step\n",
      "Epoch 254/1000\n",
      "34/34 - 1s - loss: 0.0174 - 1s/epoch - 35ms/step\n",
      "Epoch 255/1000\n",
      "34/34 - 1s - loss: 0.0169 - 1s/epoch - 35ms/step\n",
      "Epoch 256/1000\n",
      "34/34 - 1s - loss: 0.0149 - 1s/epoch - 35ms/step\n",
      "Epoch 257/1000\n",
      "34/34 - 1s - loss: 0.0126 - 1s/epoch - 35ms/step\n",
      "Epoch 258/1000\n",
      "34/34 - 1s - loss: 0.0148 - 1s/epoch - 35ms/step\n",
      "Epoch 259/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34/34 - 1s - loss: 0.0184 - 1s/epoch - 35ms/step\n",
      "Epoch 260/1000\n",
      "34/34 - 1s - loss: 0.0208 - 1s/epoch - 35ms/step\n",
      "Epoch 261/1000\n",
      "34/34 - 1s - loss: 0.0234 - 1s/epoch - 35ms/step\n",
      "Epoch 262/1000\n",
      "34/34 - 1s - loss: 0.0208 - 1s/epoch - 35ms/step\n",
      "Epoch 263/1000\n",
      "34/34 - 1s - loss: 0.0150 - 1s/epoch - 34ms/step\n",
      "Epoch 264/1000\n",
      "34/34 - 1s - loss: 0.0152 - 1s/epoch - 34ms/step\n",
      "Epoch 265/1000\n",
      "34/34 - 1s - loss: 0.0120 - 1s/epoch - 34ms/step\n",
      "Epoch 266/1000\n",
      "34/34 - 1s - loss: 0.0121 - 1s/epoch - 33ms/step\n",
      "Epoch 267/1000\n",
      "34/34 - 1s - loss: 0.0111 - 1s/epoch - 36ms/step\n",
      "Epoch 268/1000\n",
      "34/34 - 1s - loss: 0.0110 - 1s/epoch - 35ms/step\n",
      "Epoch 269/1000\n",
      "34/34 - 1s - loss: 0.0153 - 1s/epoch - 35ms/step\n",
      "Epoch 270/1000\n",
      "34/34 - 1s - loss: 0.0118 - 1s/epoch - 36ms/step\n",
      "Epoch 271/1000\n",
      "34/34 - 1s - loss: 0.0127 - 1s/epoch - 35ms/step\n",
      "Epoch 272/1000\n",
      "34/34 - 1s - loss: 0.0110 - 1s/epoch - 36ms/step\n",
      "Epoch 273/1000\n",
      "34/34 - 1s - loss: 0.0124 - 1s/epoch - 35ms/step\n",
      "Epoch 274/1000\n",
      "34/34 - 1s - loss: 0.0092 - 1s/epoch - 35ms/step\n",
      "Epoch 275/1000\n",
      "34/34 - 1s - loss: 0.0117 - 1s/epoch - 36ms/step\n",
      "Epoch 276/1000\n",
      "34/34 - 1s - loss: 0.0139 - 1s/epoch - 33ms/step\n",
      "Epoch 277/1000\n",
      "34/34 - 1s - loss: 0.0112 - 1s/epoch - 35ms/step\n",
      "Epoch 278/1000\n",
      "34/34 - 1s - loss: 0.0100 - 1s/epoch - 38ms/step\n",
      "Epoch 279/1000\n",
      "34/34 - 1s - loss: 0.0114 - 1s/epoch - 36ms/step\n",
      "Epoch 280/1000\n",
      "34/34 - 1s - loss: 0.0092 - 1s/epoch - 36ms/step\n",
      "Epoch 281/1000\n",
      "34/34 - 1s - loss: 0.0160 - 1s/epoch - 37ms/step\n",
      "Epoch 282/1000\n",
      "34/34 - 1s - loss: 0.0147 - 1s/epoch - 35ms/step\n",
      "Epoch 283/1000\n",
      "34/34 - 1s - loss: 0.0169 - 1s/epoch - 35ms/step\n",
      "Epoch 284/1000\n",
      "34/34 - 1s - loss: 0.0146 - 1s/epoch - 34ms/step\n",
      "Epoch 285/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 35ms/step\n",
      "Epoch 286/1000\n",
      "34/34 - 1s - loss: 0.0097 - 1s/epoch - 36ms/step\n",
      "Epoch 287/1000\n",
      "34/34 - 1s - loss: 0.0095 - 1s/epoch - 35ms/step\n",
      "Epoch 288/1000\n",
      "34/34 - 1s - loss: 0.0120 - 1s/epoch - 35ms/step\n",
      "Epoch 289/1000\n",
      "34/34 - 1s - loss: 0.0103 - 1s/epoch - 34ms/step\n",
      "Epoch 290/1000\n",
      "34/34 - 1s - loss: 0.0087 - 1s/epoch - 34ms/step\n",
      "Epoch 291/1000\n",
      "34/34 - 1s - loss: 0.0079 - 1s/epoch - 35ms/step\n",
      "Epoch 292/1000\n",
      "34/34 - 1s - loss: 0.0079 - 1s/epoch - 34ms/step\n",
      "Epoch 293/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 33ms/step\n",
      "Epoch 294/1000\n",
      "34/34 - 1s - loss: 0.0446 - 1s/epoch - 36ms/step\n",
      "Epoch 295/1000\n",
      "34/34 - 1s - loss: 0.0462 - 1s/epoch - 35ms/step\n",
      "Epoch 296/1000\n",
      "34/34 - 1s - loss: 0.0243 - 1s/epoch - 35ms/step\n",
      "Epoch 297/1000\n",
      "34/34 - 1s - loss: 0.0184 - 1s/epoch - 36ms/step\n",
      "Epoch 298/1000\n",
      "34/34 - 1s - loss: 0.0159 - 1s/epoch - 35ms/step\n",
      "Epoch 299/1000\n",
      "34/34 - 1s - loss: 0.0178 - 1s/epoch - 35ms/step\n",
      "Epoch 300/1000\n",
      "34/34 - 1s - loss: 0.0147 - 1s/epoch - 36ms/step\n",
      "Epoch 301/1000\n",
      "34/34 - 1s - loss: 0.0141 - 1s/epoch - 34ms/step\n",
      "Epoch 302/1000\n",
      "34/34 - 1s - loss: 0.0123 - 1s/epoch - 35ms/step\n",
      "Epoch 303/1000\n",
      "34/34 - 1s - loss: 0.0118 - 1s/epoch - 34ms/step\n",
      "Epoch 304/1000\n",
      "34/34 - 1s - loss: 0.0131 - 1s/epoch - 34ms/step\n",
      "Epoch 305/1000\n",
      "34/34 - 1s - loss: 0.0436 - 1s/epoch - 35ms/step\n",
      "Epoch 306/1000\n",
      "34/34 - 1s - loss: 0.0527 - 1s/epoch - 34ms/step\n",
      "Epoch 307/1000\n",
      "34/34 - 1s - loss: 0.0477 - 1s/epoch - 35ms/step\n",
      "Epoch 308/1000\n",
      "34/34 - 1s - loss: 0.0428 - 1s/epoch - 35ms/step\n",
      "Epoch 309/1000\n",
      "34/34 - 1s - loss: 0.0457 - 1s/epoch - 35ms/step\n",
      "Epoch 310/1000\n",
      "34/34 - 1s - loss: 0.0361 - 1s/epoch - 36ms/step\n",
      "Epoch 311/1000\n",
      "34/34 - 1s - loss: 0.0402 - 1s/epoch - 35ms/step\n",
      "Epoch 312/1000\n",
      "34/34 - 1s - loss: 0.0413 - 1s/epoch - 35ms/step\n",
      "Epoch 313/1000\n",
      "34/34 - 1s - loss: 0.0584 - 1s/epoch - 35ms/step\n",
      "Epoch 314/1000\n",
      "34/34 - 1s - loss: 0.0440 - 1s/epoch - 36ms/step\n",
      "Epoch 315/1000\n",
      "34/34 - 1s - loss: 0.0427 - 1s/epoch - 35ms/step\n",
      "Epoch 316/1000\n",
      "34/34 - 1s - loss: 0.0272 - 1s/epoch - 35ms/step\n",
      "Epoch 317/1000\n",
      "34/34 - 1s - loss: 0.0268 - 1s/epoch - 34ms/step\n",
      "Epoch 318/1000\n",
      "34/34 - 1s - loss: 0.0189 - 1s/epoch - 34ms/step\n",
      "Epoch 319/1000\n",
      "34/34 - 1s - loss: 0.0136 - 1s/epoch - 35ms/step\n",
      "Epoch 320/1000\n",
      "34/34 - 1s - loss: 0.0147 - 1s/epoch - 34ms/step\n",
      "Epoch 321/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 36ms/step\n",
      "Epoch 322/1000\n",
      "34/34 - 1s - loss: 0.0114 - 1s/epoch - 36ms/step\n",
      "Epoch 323/1000\n",
      "34/34 - 1s - loss: 0.0099 - 1s/epoch - 35ms/step\n",
      "Epoch 324/1000\n",
      "34/34 - 1s - loss: 0.0085 - 1s/epoch - 36ms/step\n",
      "Epoch 325/1000\n",
      "34/34 - 1s - loss: 0.0108 - 1s/epoch - 35ms/step\n",
      "Epoch 326/1000\n",
      "34/34 - 1s - loss: 0.0135 - 1s/epoch - 36ms/step\n",
      "Epoch 327/1000\n",
      "34/34 - 1s - loss: 0.0100 - 1s/epoch - 35ms/step\n",
      "Epoch 328/1000\n",
      "34/34 - 1s - loss: 0.0081 - 1s/epoch - 35ms/step\n",
      "Epoch 329/1000\n",
      "34/34 - 1s - loss: 0.0089 - 1s/epoch - 35ms/step\n",
      "Epoch 330/1000\n",
      "34/34 - 1s - loss: 0.0082 - 1s/epoch - 35ms/step\n",
      "Epoch 331/1000\n",
      "34/34 - 1s - loss: 0.0090 - 1s/epoch - 34ms/step\n",
      "Epoch 332/1000\n",
      "34/34 - 1s - loss: 0.0087 - 1s/epoch - 35ms/step\n",
      "Epoch 333/1000\n",
      "34/34 - 1s - loss: 0.0074 - 1s/epoch - 34ms/step\n",
      "Epoch 334/1000\n",
      "34/34 - 1s - loss: 0.0064 - 1s/epoch - 35ms/step\n",
      "Epoch 335/1000\n",
      "34/34 - 1s - loss: 0.0053 - 1s/epoch - 35ms/step\n",
      "Epoch 336/1000\n",
      "34/34 - 1s - loss: 0.0057 - 1s/epoch - 37ms/step\n",
      "Epoch 337/1000\n",
      "34/34 - 1s - loss: 0.0059 - 1s/epoch - 36ms/step\n",
      "Epoch 338/1000\n",
      "34/34 - 1s - loss: 0.0060 - 1s/epoch - 35ms/step\n",
      "Epoch 339/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 35ms/step\n",
      "Epoch 340/1000\n",
      "34/34 - 1s - loss: 0.0054 - 1s/epoch - 35ms/step\n",
      "Epoch 341/1000\n",
      "34/34 - 1s - loss: 0.0096 - 1s/epoch - 35ms/step\n",
      "Epoch 342/1000\n",
      "34/34 - 1s - loss: 0.0061 - 1s/epoch - 35ms/step\n",
      "Epoch 343/1000\n",
      "34/34 - 1s - loss: 0.0056 - 1s/epoch - 34ms/step\n",
      "Epoch 344/1000\n",
      "34/34 - 1s - loss: 0.0050 - 1s/epoch - 34ms/step\n",
      "Epoch 345/1000\n",
      "34/34 - 1s - loss: 0.0056 - 1s/epoch - 34ms/step\n",
      "Epoch 346/1000\n",
      "34/34 - 1s - loss: 0.0059 - 1s/epoch - 34ms/step\n",
      "Epoch 347/1000\n",
      "34/34 - 1s - loss: 0.0050 - 1s/epoch - 35ms/step\n",
      "Epoch 348/1000\n",
      "34/34 - 1s - loss: 0.0068 - 1s/epoch - 37ms/step\n",
      "Epoch 349/1000\n",
      "34/34 - 1s - loss: 0.0091 - 1s/epoch - 35ms/step\n",
      "Epoch 350/1000\n",
      "34/34 - 1s - loss: 0.0072 - 1s/epoch - 35ms/step\n",
      "Epoch 351/1000\n",
      "34/34 - 1s - loss: 0.0048 - 1s/epoch - 40ms/step\n",
      "Epoch 352/1000\n",
      "34/34 - 1s - loss: 0.0047 - 1s/epoch - 39ms/step\n",
      "Epoch 353/1000\n",
      "34/34 - 1s - loss: 0.0050 - 1s/epoch - 38ms/step\n",
      "Epoch 354/1000\n",
      "34/34 - 1s - loss: 0.0075 - 1s/epoch - 37ms/step\n",
      "Epoch 355/1000\n",
      "34/34 - 1s - loss: 0.0076 - 1s/epoch - 39ms/step\n",
      "Epoch 356/1000\n",
      "34/34 - 1s - loss: 0.0073 - 1s/epoch - 43ms/step\n",
      "Epoch 357/1000\n",
      "34/34 - 1s - loss: 0.0054 - 1s/epoch - 37ms/step\n",
      "Epoch 358/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 43ms/step\n",
      "Epoch 359/1000\n",
      "34/34 - 1s - loss: 0.0043 - 1s/epoch - 37ms/step\n",
      "Epoch 360/1000\n",
      "34/34 - 1s - loss: 0.0071 - 1s/epoch - 38ms/step\n",
      "Epoch 361/1000\n",
      "34/34 - 1s - loss: 0.0086 - 1s/epoch - 38ms/step\n",
      "Epoch 362/1000\n",
      "34/34 - 1s - loss: 0.0054 - 1s/epoch - 38ms/step\n",
      "Epoch 363/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 41ms/step\n",
      "Epoch 364/1000\n",
      "34/34 - 1s - loss: 0.0044 - 1s/epoch - 39ms/step\n",
      "Epoch 365/1000\n",
      "34/34 - 1s - loss: 0.0045 - 1s/epoch - 37ms/step\n",
      "Epoch 366/1000\n",
      "34/34 - 1s - loss: 0.0043 - 1s/epoch - 39ms/step\n",
      "Epoch 367/1000\n",
      "34/34 - 1s - loss: 0.0049 - 1s/epoch - 37ms/step\n",
      "Epoch 368/1000\n",
      "34/34 - 1s - loss: 0.0046 - 1s/epoch - 37ms/step\n",
      "Epoch 369/1000\n",
      "34/34 - 1s - loss: 0.0049 - 1s/epoch - 41ms/step\n",
      "Epoch 370/1000\n",
      "34/34 - 2s - loss: 0.0046 - 2s/epoch - 45ms/step\n",
      "Epoch 371/1000\n",
      "34/34 - 1s - loss: 0.0060 - 1s/epoch - 36ms/step\n",
      "Epoch 372/1000\n",
      "34/34 - 2s - loss: 0.0036 - 2s/epoch - 54ms/step\n",
      "Epoch 373/1000\n",
      "34/34 - 2s - loss: 0.0036 - 2s/epoch - 62ms/step\n",
      "Epoch 374/1000\n",
      "34/34 - 1s - loss: 0.0027 - 1s/epoch - 42ms/step\n",
      "Epoch 375/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 39ms/step\n",
      "Epoch 376/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 42ms/step\n",
      "Epoch 377/1000\n",
      "34/34 - 1s - loss: 0.0022 - 1s/epoch - 40ms/step\n",
      "Epoch 378/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 40ms/step\n",
      "Epoch 379/1000\n",
      "34/34 - 1s - loss: 0.0132 - 1s/epoch - 41ms/step\n",
      "Epoch 380/1000\n",
      "34/34 - 2s - loss: 0.0053 - 2s/epoch - 45ms/step\n",
      "Epoch 381/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 42ms/step\n",
      "Epoch 382/1000\n",
      "34/34 - 1s - loss: 0.0055 - 1s/epoch - 38ms/step\n",
      "Epoch 383/1000\n",
      "34/34 - 1s - loss: 0.0063 - 1s/epoch - 41ms/step\n",
      "Epoch 384/1000\n",
      "34/34 - 1s - loss: 0.0047 - 1s/epoch - 40ms/step\n",
      "Epoch 385/1000\n",
      "34/34 - 1s - loss: 0.0048 - 1s/epoch - 39ms/step\n",
      "Epoch 386/1000\n",
      "34/34 - 1s - loss: 0.0029 - 1s/epoch - 41ms/step\n",
      "Epoch 387/1000\n",
      "34/34 - 1s - loss: 0.0026 - 1s/epoch - 41ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 388/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 37ms/step\n",
      "Epoch 389/1000\n",
      "34/34 - 2s - loss: 0.0020 - 2s/epoch - 45ms/step\n",
      "Epoch 390/1000\n",
      "34/34 - 2s - loss: 0.0018 - 2s/epoch - 52ms/step\n",
      "Epoch 391/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 41ms/step\n",
      "Epoch 392/1000\n",
      "34/34 - 2s - loss: 0.0022 - 2s/epoch - 49ms/step\n",
      "Epoch 393/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 44ms/step\n",
      "Epoch 394/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 44ms/step\n",
      "Epoch 395/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 43ms/step\n",
      "Epoch 396/1000\n",
      "34/34 - 2s - loss: 0.0015 - 2s/epoch - 46ms/step\n",
      "Epoch 397/1000\n",
      "34/34 - 2s - loss: 0.0013 - 2s/epoch - 45ms/step\n",
      "Epoch 398/1000\n",
      "34/34 - 1s - loss: 0.0014 - 1s/epoch - 39ms/step\n",
      "Epoch 399/1000\n",
      "34/34 - 1s - loss: 0.0111 - 1s/epoch - 39ms/step\n",
      "Epoch 400/1000\n",
      "34/34 - 1s - loss: 0.0075 - 1s/epoch - 39ms/step\n",
      "Epoch 401/1000\n",
      "34/34 - 1s - loss: 0.0060 - 1s/epoch - 37ms/step\n",
      "Epoch 402/1000\n",
      "34/34 - 1s - loss: 0.0072 - 1s/epoch - 39ms/step\n",
      "Epoch 403/1000\n",
      "34/34 - 1s - loss: 0.0093 - 1s/epoch - 39ms/step\n",
      "Epoch 404/1000\n",
      "34/34 - 1s - loss: 0.0193 - 1s/epoch - 39ms/step\n",
      "Epoch 405/1000\n",
      "34/34 - 1s - loss: 0.0098 - 1s/epoch - 39ms/step\n",
      "Epoch 406/1000\n",
      "34/34 - 1s - loss: 0.0061 - 1s/epoch - 40ms/step\n",
      "Epoch 407/1000\n",
      "34/34 - 1s - loss: 0.0048 - 1s/epoch - 39ms/step\n",
      "Epoch 408/1000\n",
      "34/34 - 1s - loss: 0.0042 - 1s/epoch - 39ms/step\n",
      "Epoch 409/1000\n",
      "34/34 - 1s - loss: 0.0036 - 1s/epoch - 39ms/step\n",
      "Epoch 410/1000\n",
      "34/34 - 1s - loss: 0.0040 - 1s/epoch - 40ms/step\n",
      "Epoch 411/1000\n",
      "34/34 - 1s - loss: 0.0033 - 1s/epoch - 40ms/step\n",
      "Epoch 412/1000\n",
      "34/34 - 1s - loss: 0.0037 - 1s/epoch - 42ms/step\n",
      "Epoch 413/1000\n",
      "34/34 - 1s - loss: 0.0037 - 1s/epoch - 38ms/step\n",
      "Epoch 414/1000\n",
      "34/34 - 1s - loss: 0.0044 - 1s/epoch - 40ms/step\n",
      "Epoch 415/1000\n",
      "34/34 - 2s - loss: 0.0032 - 2s/epoch - 47ms/step\n",
      "Epoch 416/1000\n",
      "34/34 - 2s - loss: 0.0027 - 2s/epoch - 47ms/step\n",
      "Epoch 417/1000\n",
      "34/34 - 2s - loss: 0.0026 - 2s/epoch - 51ms/step\n",
      "Epoch 418/1000\n",
      "34/34 - 2s - loss: 0.0026 - 2s/epoch - 46ms/step\n",
      "Epoch 419/1000\n",
      "34/34 - 1s - loss: 0.0030 - 1s/epoch - 43ms/step\n",
      "Epoch 420/1000\n",
      "34/34 - 1s - loss: 0.0029 - 1s/epoch - 40ms/step\n",
      "Epoch 421/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 41ms/step\n",
      "Epoch 422/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 40ms/step\n",
      "Epoch 423/1000\n",
      "34/34 - 1s - loss: 0.0029 - 1s/epoch - 38ms/step\n",
      "Epoch 424/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 41ms/step\n",
      "Epoch 425/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 39ms/step\n",
      "Epoch 426/1000\n",
      "34/34 - 1s - loss: 0.0030 - 1s/epoch - 41ms/step\n",
      "Epoch 427/1000\n",
      "34/34 - 1s - loss: 0.0045 - 1s/epoch - 39ms/step\n",
      "Epoch 428/1000\n",
      "34/34 - 1s - loss: 0.0027 - 1s/epoch - 39ms/step\n",
      "Epoch 429/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 39ms/step\n",
      "Epoch 430/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 39ms/step\n",
      "Epoch 431/1000\n",
      "34/34 - 1s - loss: 0.0049 - 1s/epoch - 42ms/step\n",
      "Epoch 432/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 40ms/step\n",
      "Epoch 433/1000\n",
      "34/34 - 1s - loss: 0.0026 - 1s/epoch - 38ms/step\n",
      "Epoch 434/1000\n",
      "34/34 - 2s - loss: 0.0021 - 2s/epoch - 45ms/step\n",
      "Epoch 435/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 40ms/step\n",
      "Epoch 436/1000\n",
      "34/34 - 1s - loss: 0.0012 - 1s/epoch - 39ms/step\n",
      "Epoch 437/1000\n",
      "34/34 - 1s - loss: 0.0017 - 1s/epoch - 36ms/step\n",
      "Epoch 438/1000\n",
      "34/34 - 1s - loss: 0.0013 - 1s/epoch - 42ms/step\n",
      "Epoch 439/1000\n",
      "34/34 - 1s - loss: 0.0016 - 1s/epoch - 41ms/step\n",
      "Epoch 440/1000\n",
      "34/34 - 1s - loss: 0.0014 - 1s/epoch - 39ms/step\n",
      "Epoch 441/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 40ms/step\n",
      "Epoch 442/1000\n",
      "34/34 - 1s - loss: 0.0017 - 1s/epoch - 39ms/step\n",
      "Epoch 443/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 39ms/step\n",
      "Epoch 444/1000\n",
      "34/34 - 1s - loss: 0.0036 - 1s/epoch - 42ms/step\n",
      "Epoch 445/1000\n",
      "34/34 - 1s - loss: 0.0022 - 1s/epoch - 43ms/step\n",
      "Epoch 446/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 44ms/step\n",
      "Epoch 447/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 42ms/step\n",
      "Epoch 448/1000\n",
      "34/34 - 1s - loss: 0.0030 - 1s/epoch - 42ms/step\n",
      "Epoch 449/1000\n",
      "34/34 - 1s - loss: 0.0021 - 1s/epoch - 39ms/step\n",
      "Epoch 450/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 42ms/step\n",
      "Epoch 451/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 40ms/step\n",
      "Epoch 452/1000\n",
      "34/34 - 1s - loss: 0.0019 - 1s/epoch - 39ms/step\n",
      "Epoch 453/1000\n",
      "34/34 - 1s - loss: 9.8868e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 454/1000\n",
      "34/34 - 1s - loss: 8.3011e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 455/1000\n",
      "34/34 - 1s - loss: 8.6245e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 456/1000\n",
      "34/34 - 1s - loss: 7.6828e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 457/1000\n",
      "34/34 - 1s - loss: 7.3204e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 458/1000\n",
      "34/34 - 1s - loss: 8.5516e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 459/1000\n",
      "34/34 - 1s - loss: 6.7020e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 460/1000\n",
      "34/34 - 1s - loss: 8.7381e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 461/1000\n",
      "34/34 - 1s - loss: 9.9172e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 462/1000\n",
      "34/34 - 1s - loss: 8.3314e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 463/1000\n",
      "34/34 - 1s - loss: 7.2031e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 464/1000\n",
      "34/34 - 1s - loss: 0.0010 - 1s/epoch - 39ms/step\n",
      "Epoch 465/1000\n",
      "34/34 - 1s - loss: 8.4166e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 466/1000\n",
      "34/34 - 1s - loss: 6.4270e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 467/1000\n",
      "34/34 - 1s - loss: 5.5567e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 468/1000\n",
      "34/34 - 1s - loss: 5.8235e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 469/1000\n",
      "34/34 - 1s - loss: 6.0848e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 470/1000\n",
      "34/34 - 1s - loss: 5.2314e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 471/1000\n",
      "34/34 - 1s - loss: 4.5415e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 472/1000\n",
      "34/34 - 1s - loss: 4.8418e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 473/1000\n",
      "34/34 - 1s - loss: 5.1250e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 474/1000\n",
      "34/34 - 1s - loss: 6.1281e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 475/1000\n",
      "34/34 - 1s - loss: 4.3445e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 476/1000\n",
      "34/34 - 1s - loss: 4.3929e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 477/1000\n",
      "34/34 - 1s - loss: 4.7537e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 478/1000\n",
      "34/34 - 1s - loss: 4.1952e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 479/1000\n",
      "34/34 - 1s - loss: 3.8301e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 480/1000\n",
      "34/34 - 1s - loss: 3.6883e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 481/1000\n",
      "34/34 - 1s - loss: 3.7372e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 482/1000\n",
      "34/34 - 1s - loss: 3.8635e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 483/1000\n",
      "34/34 - 1s - loss: 3.6784e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 484/1000\n",
      "34/34 - 2s - loss: 4.8958e-04 - 2s/epoch - 55ms/step\n",
      "Epoch 485/1000\n",
      "34/34 - 2s - loss: 0.0036 - 2s/epoch - 44ms/step\n",
      "Epoch 486/1000\n",
      "34/34 - 1s - loss: 0.0175 - 1s/epoch - 40ms/step\n",
      "Epoch 487/1000\n",
      "34/34 - 2s - loss: 0.0190 - 2s/epoch - 46ms/step\n",
      "Epoch 488/1000\n",
      "34/34 - 1s - loss: 0.0111 - 1s/epoch - 40ms/step\n",
      "Epoch 489/1000\n",
      "34/34 - 1s - loss: 0.0143 - 1s/epoch - 37ms/step\n",
      "Epoch 490/1000\n",
      "34/34 - 1s - loss: 0.0132 - 1s/epoch - 44ms/step\n",
      "Epoch 491/1000\n",
      "34/34 - 1s - loss: 0.0074 - 1s/epoch - 43ms/step\n",
      "Epoch 492/1000\n",
      "34/34 - 1s - loss: 0.0056 - 1s/epoch - 39ms/step\n",
      "Epoch 493/1000\n",
      "34/34 - 1s - loss: 0.0115 - 1s/epoch - 40ms/step\n",
      "Epoch 494/1000\n",
      "34/34 - 1s - loss: 0.0353 - 1s/epoch - 41ms/step\n",
      "Epoch 495/1000\n",
      "34/34 - 1s - loss: 0.0265 - 1s/epoch - 39ms/step\n",
      "Epoch 496/1000\n",
      "34/34 - 1s - loss: 0.0197 - 1s/epoch - 37ms/step\n",
      "Epoch 497/1000\n",
      "34/34 - 1s - loss: 0.0143 - 1s/epoch - 41ms/step\n",
      "Epoch 498/1000\n",
      "34/34 - 1s - loss: 0.0140 - 1s/epoch - 40ms/step\n",
      "Epoch 499/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 41ms/step\n",
      "Epoch 500/1000\n",
      "34/34 - 1s - loss: 0.0117 - 1s/epoch - 39ms/step\n",
      "Epoch 501/1000\n",
      "34/34 - 1s - loss: 0.0109 - 1s/epoch - 40ms/step\n",
      "Epoch 502/1000\n",
      "34/34 - 1s - loss: 0.0148 - 1s/epoch - 38ms/step\n",
      "Epoch 503/1000\n",
      "34/34 - 1s - loss: 0.0458 - 1s/epoch - 38ms/step\n",
      "Epoch 504/1000\n",
      "34/34 - 1s - loss: 0.0314 - 1s/epoch - 39ms/step\n",
      "Epoch 505/1000\n",
      "34/34 - 1s - loss: 0.0210 - 1s/epoch - 39ms/step\n",
      "Epoch 506/1000\n",
      "34/34 - 1s - loss: 0.0191 - 1s/epoch - 38ms/step\n",
      "Epoch 507/1000\n",
      "34/34 - 1s - loss: 0.0138 - 1s/epoch - 39ms/step\n",
      "Epoch 508/1000\n",
      "34/34 - 1s - loss: 0.0172 - 1s/epoch - 37ms/step\n",
      "Epoch 509/1000\n",
      "34/34 - 1s - loss: 0.0118 - 1s/epoch - 39ms/step\n",
      "Epoch 510/1000\n",
      "34/34 - 1s - loss: 0.0097 - 1s/epoch - 38ms/step\n",
      "Epoch 511/1000\n",
      "34/34 - 1s - loss: 0.0114 - 1s/epoch - 39ms/step\n",
      "Epoch 512/1000\n",
      "34/34 - 1s - loss: 0.0116 - 1s/epoch - 38ms/step\n",
      "Epoch 513/1000\n",
      "34/34 - 1s - loss: 0.0065 - 1s/epoch - 38ms/step\n",
      "Epoch 514/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 41ms/step\n",
      "Epoch 515/1000\n",
      "34/34 - 1s - loss: 0.0068 - 1s/epoch - 37ms/step\n",
      "Epoch 516/1000\n",
      "34/34 - 1s - loss: 0.0076 - 1s/epoch - 40ms/step\n",
      "Epoch 517/1000\n",
      "34/34 - 1s - loss: 0.0127 - 1s/epoch - 40ms/step\n",
      "Epoch 518/1000\n",
      "34/34 - 1s - loss: 0.0063 - 1s/epoch - 37ms/step\n",
      "Epoch 519/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 40ms/step\n",
      "Epoch 520/1000\n",
      "34/34 - 1s - loss: 0.0041 - 1s/epoch - 39ms/step\n",
      "Epoch 521/1000\n",
      "34/34 - 1s - loss: 0.0036 - 1s/epoch - 38ms/step\n",
      "Epoch 522/1000\n",
      "34/34 - 1s - loss: 0.0033 - 1s/epoch - 36ms/step\n",
      "Epoch 523/1000\n",
      "34/34 - 1s - loss: 0.0028 - 1s/epoch - 36ms/step\n",
      "Epoch 524/1000\n",
      "34/34 - 1s - loss: 0.0030 - 1s/epoch - 42ms/step\n",
      "Epoch 525/1000\n",
      "34/34 - 1s - loss: 0.0037 - 1s/epoch - 38ms/step\n",
      "Epoch 526/1000\n",
      "34/34 - 1s - loss: 0.0035 - 1s/epoch - 38ms/step\n",
      "Epoch 527/1000\n",
      "34/34 - 1s - loss: 0.0033 - 1s/epoch - 37ms/step\n",
      "Epoch 528/1000\n",
      "34/34 - 1s - loss: 0.0038 - 1s/epoch - 40ms/step\n",
      "Epoch 529/1000\n",
      "34/34 - 1s - loss: 0.0035 - 1s/epoch - 38ms/step\n",
      "Epoch 530/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 36ms/step\n",
      "Epoch 531/1000\n",
      "34/34 - 1s - loss: 0.0038 - 1s/epoch - 37ms/step\n",
      "Epoch 532/1000\n",
      "34/34 - 1s - loss: 0.0028 - 1s/epoch - 37ms/step\n",
      "Epoch 533/1000\n",
      "34/34 - 1s - loss: 0.0028 - 1s/epoch - 36ms/step\n",
      "Epoch 534/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 37ms/step\n",
      "Epoch 535/1000\n",
      "34/34 - 1s - loss: 0.0022 - 1s/epoch - 36ms/step\n",
      "Epoch 536/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 38ms/step\n",
      "Epoch 537/1000\n",
      "34/34 - 2s - loss: 0.0020 - 2s/epoch - 49ms/step\n",
      "Epoch 538/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 39ms/step\n",
      "Epoch 539/1000\n",
      "34/34 - 1s - loss: 0.0022 - 1s/epoch - 40ms/step\n",
      "Epoch 540/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 42ms/step\n",
      "Epoch 541/1000\n",
      "34/34 - 1s - loss: 0.0029 - 1s/epoch - 41ms/step\n",
      "Epoch 542/1000\n",
      "34/34 - 1s - loss: 0.0026 - 1s/epoch - 38ms/step\n",
      "Epoch 543/1000\n",
      "34/34 - 1s - loss: 0.0040 - 1s/epoch - 40ms/step\n",
      "Epoch 544/1000\n",
      "34/34 - 1s - loss: 0.0027 - 1s/epoch - 39ms/step\n",
      "Epoch 545/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 38ms/step\n",
      "Epoch 546/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 42ms/step\n",
      "Epoch 547/1000\n",
      "34/34 - 1s - loss: 0.0036 - 1s/epoch - 41ms/step\n",
      "Epoch 548/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 41ms/step\n",
      "Epoch 549/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 40ms/step\n",
      "Epoch 550/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 39ms/step\n",
      "Epoch 551/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 39ms/step\n",
      "Epoch 552/1000\n",
      "34/34 - 1s - loss: 0.0021 - 1s/epoch - 42ms/step\n",
      "Epoch 553/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 40ms/step\n",
      "Epoch 554/1000\n",
      "34/34 - 1s - loss: 0.0037 - 1s/epoch - 37ms/step\n",
      "Epoch 555/1000\n",
      "34/34 - 1s - loss: 0.0079 - 1s/epoch - 38ms/step\n",
      "Epoch 556/1000\n",
      "34/34 - 1s - loss: 0.0038 - 1s/epoch - 40ms/step\n",
      "Epoch 557/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 37ms/step\n",
      "Epoch 558/1000\n",
      "34/34 - 1s - loss: 0.0014 - 1s/epoch - 38ms/step\n",
      "Epoch 559/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 38ms/step\n",
      "Epoch 560/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 42ms/step\n",
      "Epoch 561/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 39ms/step\n",
      "Epoch 562/1000\n",
      "34/34 - 1s - loss: 0.0011 - 1s/epoch - 39ms/step\n",
      "Epoch 563/1000\n",
      "34/34 - 1s - loss: 0.0012 - 1s/epoch - 36ms/step\n",
      "Epoch 564/1000\n",
      "34/34 - 1s - loss: 0.0014 - 1s/epoch - 37ms/step\n",
      "Epoch 565/1000\n",
      "34/34 - 1s - loss: 0.0011 - 1s/epoch - 38ms/step\n",
      "Epoch 566/1000\n",
      "34/34 - 1s - loss: 9.9806e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 567/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 37ms/step\n",
      "Epoch 568/1000\n",
      "34/34 - 1s - loss: 0.0016 - 1s/epoch - 37ms/step\n",
      "Epoch 569/1000\n",
      "34/34 - 1s - loss: 0.0030 - 1s/epoch - 37ms/step\n",
      "Epoch 570/1000\n",
      "34/34 - 1s - loss: 0.0034 - 1s/epoch - 36ms/step\n",
      "Epoch 571/1000\n",
      "34/34 - 1s - loss: 0.0020 - 1s/epoch - 36ms/step\n",
      "Epoch 572/1000\n",
      "34/34 - 1s - loss: 0.0013 - 1s/epoch - 36ms/step\n",
      "Epoch 573/1000\n",
      "34/34 - 1s - loss: 9.3366e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 574/1000\n",
      "34/34 - 1s - loss: 7.1967e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 575/1000\n",
      "34/34 - 1s - loss: 6.1058e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 576/1000\n",
      "34/34 - 1s - loss: 5.4090e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 577/1000\n",
      "34/34 - 1s - loss: 5.6285e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 578/1000\n",
      "34/34 - 1s - loss: 4.7370e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 579/1000\n",
      "34/34 - 1s - loss: 4.5808e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 580/1000\n",
      "34/34 - 1s - loss: 4.1493e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 581/1000\n",
      "34/34 - 1s - loss: 4.6569e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 582/1000\n",
      "34/34 - 1s - loss: 5.1720e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 583/1000\n",
      "34/34 - 1s - loss: 4.6751e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 584/1000\n",
      "34/34 - 1s - loss: 6.2101e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 585/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 39ms/step\n",
      "Epoch 586/1000\n",
      "34/34 - 1s - loss: 0.0126 - 1s/epoch - 36ms/step\n",
      "Epoch 587/1000\n",
      "34/34 - 1s - loss: 0.0115 - 1s/epoch - 38ms/step\n",
      "Epoch 588/1000\n",
      "34/34 - 1s - loss: 0.0076 - 1s/epoch - 38ms/step\n",
      "Epoch 589/1000\n",
      "34/34 - 1s - loss: 0.0106 - 1s/epoch - 38ms/step\n",
      "Epoch 590/1000\n",
      "34/34 - 1s - loss: 0.0039 - 1s/epoch - 38ms/step\n",
      "Epoch 591/1000\n",
      "34/34 - 2s - loss: 0.0032 - 2s/epoch - 46ms/step\n",
      "Epoch 592/1000\n",
      "34/34 - 1s - loss: 0.0031 - 1s/epoch - 43ms/step\n",
      "Epoch 593/1000\n",
      "34/34 - 1s - loss: 0.0047 - 1s/epoch - 42ms/step\n",
      "Epoch 594/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 42ms/step\n",
      "Epoch 595/1000\n",
      "34/34 - 1s - loss: 0.0017 - 1s/epoch - 39ms/step\n",
      "Epoch 596/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 38ms/step\n",
      "Epoch 597/1000\n",
      "34/34 - 1s - loss: 0.0026 - 1s/epoch - 38ms/step\n",
      "Epoch 598/1000\n",
      "34/34 - 1s - loss: 0.0019 - 1s/epoch - 37ms/step\n",
      "Epoch 599/1000\n",
      "34/34 - 1s - loss: 0.0019 - 1s/epoch - 38ms/step\n",
      "Epoch 600/1000\n",
      "34/34 - 1s - loss: 0.0017 - 1s/epoch - 36ms/step\n",
      "Epoch 601/1000\n",
      "34/34 - 1s - loss: 0.0016 - 1s/epoch - 41ms/step\n",
      "Epoch 602/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 38ms/step\n",
      "Epoch 603/1000\n",
      "34/34 - 1s - loss: 0.0013 - 1s/epoch - 38ms/step\n",
      "Epoch 604/1000\n",
      "34/34 - 1s - loss: 8.0326e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 605/1000\n",
      "34/34 - 1s - loss: 5.6407e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 606/1000\n",
      "34/34 - 1s - loss: 5.2828e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 607/1000\n",
      "34/34 - 1s - loss: 4.5039e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 608/1000\n",
      "34/34 - 1s - loss: 4.2543e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 609/1000\n",
      "34/34 - 1s - loss: 4.2194e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 610/1000\n",
      "34/34 - 1s - loss: 3.7951e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 611/1000\n",
      "34/34 - 1s - loss: 3.7073e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 612/1000\n",
      "34/34 - 1s - loss: 3.7321e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 613/1000\n",
      "34/34 - 1s - loss: 3.3629e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 614/1000\n",
      "34/34 - 1s - loss: 3.1785e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 615/1000\n",
      "34/34 - 1s - loss: 3.3146e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 616/1000\n",
      "34/34 - 1s - loss: 3.3626e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 617/1000\n",
      "34/34 - 1s - loss: 3.4461e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 618/1000\n",
      "34/34 - 1s - loss: 3.3651e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 619/1000\n",
      "34/34 - 1s - loss: 3.2467e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 620/1000\n",
      "34/34 - 1s - loss: 3.2698e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 621/1000\n",
      "34/34 - 1s - loss: 2.9959e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 622/1000\n",
      "34/34 - 1s - loss: 2.8890e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 623/1000\n",
      "34/34 - 1s - loss: 2.8830e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 624/1000\n",
      "34/34 - 1s - loss: 2.8451e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 625/1000\n",
      "34/34 - 1s - loss: 2.7942e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 626/1000\n",
      "34/34 - 1s - loss: 2.6833e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 627/1000\n",
      "34/34 - 1s - loss: 2.6921e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 628/1000\n",
      "34/34 - 1s - loss: 2.6066e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 629/1000\n",
      "34/34 - 1s - loss: 2.3665e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 630/1000\n",
      "34/34 - 1s - loss: 2.2958e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 631/1000\n",
      "34/34 - 1s - loss: 2.7662e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 632/1000\n",
      "34/34 - 1s - loss: 2.6096e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 633/1000\n",
      "34/34 - 1s - loss: 2.3975e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 634/1000\n",
      "34/34 - 1s - loss: 2.3236e-04 - 1s/epoch - 42ms/step\n",
      "Epoch 635/1000\n",
      "34/34 - 1s - loss: 2.3607e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 636/1000\n",
      "34/34 - 1s - loss: 2.2899e-04 - 1s/epoch - 42ms/step\n",
      "Epoch 637/1000\n",
      "34/34 - 1s - loss: 2.1465e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 638/1000\n",
      "34/34 - 1s - loss: 2.4382e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 639/1000\n",
      "34/34 - 1s - loss: 2.0321e-04 - 1s/epoch - 40ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 640/1000\n",
      "34/34 - 1s - loss: 2.3368e-04 - 1s/epoch - 42ms/step\n",
      "Epoch 641/1000\n",
      "34/34 - 1s - loss: 6.9443e-04 - 1s/epoch - 44ms/step\n",
      "Epoch 642/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 43ms/step\n",
      "Epoch 643/1000\n",
      "34/34 - 1s - loss: 9.0693e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 644/1000\n",
      "34/34 - 1s - loss: 0.0010 - 1s/epoch - 36ms/step\n",
      "Epoch 645/1000\n",
      "34/34 - 1s - loss: 6.4405e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 646/1000\n",
      "34/34 - 1s - loss: 0.0113 - 1s/epoch - 37ms/step\n",
      "Epoch 647/1000\n",
      "34/34 - 1s - loss: 0.0089 - 1s/epoch - 37ms/step\n",
      "Epoch 648/1000\n",
      "34/34 - 1s - loss: 0.0100 - 1s/epoch - 38ms/step\n",
      "Epoch 649/1000\n",
      "34/34 - 1s - loss: 0.0110 - 1s/epoch - 42ms/step\n",
      "Epoch 650/1000\n",
      "34/34 - 1s - loss: 0.0131 - 1s/epoch - 43ms/step\n",
      "Epoch 651/1000\n",
      "34/34 - 1s - loss: 0.0074 - 1s/epoch - 37ms/step\n",
      "Epoch 652/1000\n",
      "34/34 - 1s - loss: 0.0058 - 1s/epoch - 37ms/step\n",
      "Epoch 653/1000\n",
      "34/34 - 1s - loss: 0.0021 - 1s/epoch - 37ms/step\n",
      "Epoch 654/1000\n",
      "34/34 - 1s - loss: 0.0033 - 1s/epoch - 35ms/step\n",
      "Epoch 655/1000\n",
      "34/34 - 1s - loss: 0.0034 - 1s/epoch - 36ms/step\n",
      "Epoch 656/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 35ms/step\n",
      "Epoch 657/1000\n",
      "34/34 - 1s - loss: 0.0012 - 1s/epoch - 34ms/step\n",
      "Epoch 658/1000\n",
      "34/34 - 1s - loss: 9.3824e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 659/1000\n",
      "34/34 - 1s - loss: 8.9882e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 660/1000\n",
      "34/34 - 1s - loss: 7.8685e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 661/1000\n",
      "34/34 - 1s - loss: 6.7046e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 662/1000\n",
      "34/34 - 1s - loss: 6.0592e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 663/1000\n",
      "34/34 - 1s - loss: 5.9356e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 664/1000\n",
      "34/34 - 1s - loss: 4.9085e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 665/1000\n",
      "34/34 - 1s - loss: 4.4140e-04 - 1s/epoch - 42ms/step\n",
      "Epoch 666/1000\n",
      "34/34 - 1s - loss: 4.0688e-04 - 1s/epoch - 42ms/step\n",
      "Epoch 667/1000\n",
      "34/34 - 1s - loss: 3.8374e-04 - 1s/epoch - 40ms/step\n",
      "Epoch 668/1000\n",
      "34/34 - 1s - loss: 4.7255e-04 - 1s/epoch - 41ms/step\n",
      "Epoch 669/1000\n",
      "34/34 - 1s - loss: 3.9981e-04 - 1s/epoch - 39ms/step\n",
      "Epoch 670/1000\n",
      "34/34 - 1s - loss: 3.1213e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 671/1000\n",
      "34/34 - 1s - loss: 3.0595e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 672/1000\n",
      "34/34 - 1s - loss: 2.7859e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 673/1000\n",
      "34/34 - 1s - loss: 2.9529e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 674/1000\n",
      "34/34 - 1s - loss: 3.9956e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 675/1000\n",
      "34/34 - 1s - loss: 3.5323e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 676/1000\n",
      "34/34 - 1s - loss: 3.1539e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 677/1000\n",
      "34/34 - 1s - loss: 2.7420e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 678/1000\n",
      "34/34 - 1s - loss: 2.5649e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 679/1000\n",
      "34/34 - 1s - loss: 2.5225e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 680/1000\n",
      "34/34 - 1s - loss: 2.4669e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 681/1000\n",
      "34/34 - 1s - loss: 2.1166e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 682/1000\n",
      "34/34 - 1s - loss: 2.2070e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 683/1000\n",
      "34/34 - 1s - loss: 2.2165e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 684/1000\n",
      "34/34 - 1s - loss: 2.0335e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 685/1000\n",
      "34/34 - 1s - loss: 2.3661e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 686/1000\n",
      "34/34 - 1s - loss: 2.4234e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 687/1000\n",
      "34/34 - 1s - loss: 2.3289e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 688/1000\n",
      "34/34 - 1s - loss: 2.2781e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 689/1000\n",
      "34/34 - 1s - loss: 2.0845e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 690/1000\n",
      "34/34 - 1s - loss: 1.9592e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 691/1000\n",
      "34/34 - 1s - loss: 1.9225e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 692/1000\n",
      "34/34 - 1s - loss: 1.9865e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 693/1000\n",
      "34/34 - 1s - loss: 1.7984e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 694/1000\n",
      "34/34 - 1s - loss: 1.7296e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 695/1000\n",
      "34/34 - 1s - loss: 2.2716e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 696/1000\n",
      "34/34 - 1s - loss: 1.9826e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 697/1000\n",
      "34/34 - 1s - loss: 2.1507e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 698/1000\n",
      "34/34 - 1s - loss: 2.0615e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 699/1000\n",
      "34/34 - 1s - loss: 1.7249e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 700/1000\n",
      "34/34 - 1s - loss: 1.8328e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 701/1000\n",
      "34/34 - 1s - loss: 1.9682e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 702/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 37ms/step\n",
      "Epoch 703/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 36ms/step\n",
      "Epoch 704/1000\n",
      "34/34 - 1s - loss: 0.0076 - 1s/epoch - 35ms/step\n",
      "Epoch 705/1000\n",
      "34/34 - 1s - loss: 0.0108 - 1s/epoch - 36ms/step\n",
      "Epoch 706/1000\n",
      "34/34 - 1s - loss: 0.0046 - 1s/epoch - 35ms/step\n",
      "Epoch 707/1000\n",
      "34/34 - 1s - loss: 0.0052 - 1s/epoch - 38ms/step\n",
      "Epoch 708/1000\n",
      "34/34 - 1s - loss: 0.0027 - 1s/epoch - 40ms/step\n",
      "Epoch 709/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 38ms/step\n",
      "Epoch 710/1000\n",
      "34/34 - 1s - loss: 0.0021 - 1s/epoch - 36ms/step\n",
      "Epoch 711/1000\n",
      "34/34 - 1s - loss: 0.0011 - 1s/epoch - 35ms/step\n",
      "Epoch 712/1000\n",
      "34/34 - 1s - loss: 6.6101e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 713/1000\n",
      "34/34 - 1s - loss: 8.7956e-04 - 1s/epoch - 38ms/step\n",
      "Epoch 714/1000\n",
      "34/34 - 1s - loss: 0.0022 - 1s/epoch - 37ms/step\n",
      "Epoch 715/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 37ms/step\n",
      "Epoch 716/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 37ms/step\n",
      "Epoch 717/1000\n",
      "34/34 - 1s - loss: 0.0019 - 1s/epoch - 36ms/step\n",
      "Epoch 718/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 35ms/step\n",
      "Epoch 719/1000\n",
      "34/34 - 1s - loss: 0.0025 - 1s/epoch - 36ms/step\n",
      "Epoch 720/1000\n",
      "34/34 - 1s - loss: 0.0015 - 1s/epoch - 35ms/step\n",
      "Epoch 721/1000\n",
      "34/34 - 1s - loss: 9.9362e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 722/1000\n",
      "34/34 - 1s - loss: 8.0244e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 723/1000\n",
      "34/34 - 1s - loss: 6.9789e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 724/1000\n",
      "34/34 - 1s - loss: 6.5065e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 725/1000\n",
      "34/34 - 1s - loss: 5.7158e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 726/1000\n",
      "34/34 - 1s - loss: 5.7682e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 727/1000\n",
      "34/34 - 1s - loss: 5.5517e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 728/1000\n",
      "34/34 - 1s - loss: 5.2349e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 729/1000\n",
      "34/34 - 1s - loss: 4.8879e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 730/1000\n",
      "34/34 - 1s - loss: 4.4552e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 731/1000\n",
      "34/34 - 1s - loss: 4.3481e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 732/1000\n",
      "34/34 - 1s - loss: 4.0167e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 733/1000\n",
      "34/34 - 1s - loss: 4.8659e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 734/1000\n",
      "34/34 - 1s - loss: 3.8937e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 735/1000\n",
      "34/34 - 1s - loss: 3.6878e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 736/1000\n",
      "34/34 - 1s - loss: 4.0762e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 737/1000\n",
      "34/34 - 1s - loss: 5.6366e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 738/1000\n",
      "34/34 - 1s - loss: 3.8878e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 739/1000\n",
      "34/34 - 1s - loss: 3.9235e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 740/1000\n",
      "34/34 - 1s - loss: 4.1296e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 741/1000\n",
      "34/34 - 1s - loss: 3.2213e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 742/1000\n",
      "34/34 - 1s - loss: 2.9072e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 743/1000\n",
      "34/34 - 1s - loss: 3.2318e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 744/1000\n",
      "34/34 - 1s - loss: 3.1168e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 745/1000\n",
      "34/34 - 1s - loss: 2.7953e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 746/1000\n",
      "34/34 - 1s - loss: 2.6346e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 747/1000\n",
      "34/34 - 1s - loss: 4.1103e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 748/1000\n",
      "34/34 - 1s - loss: 3.0094e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 749/1000\n",
      "34/34 - 1s - loss: 2.6321e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 750/1000\n",
      "34/34 - 1s - loss: 2.6027e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 751/1000\n",
      "34/34 - 1s - loss: 2.4741e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 752/1000\n",
      "34/34 - 1s - loss: 2.1992e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 753/1000\n",
      "34/34 - 1s - loss: 2.5625e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 754/1000\n",
      "34/34 - 1s - loss: 2.5819e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 755/1000\n",
      "34/34 - 1s - loss: 2.9385e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 756/1000\n",
      "34/34 - 1s - loss: 2.9506e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 757/1000\n",
      "34/34 - 1s - loss: 2.0908e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 758/1000\n",
      "34/34 - 1s - loss: 2.0083e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 759/1000\n",
      "34/34 - 1s - loss: 1.9962e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 760/1000\n",
      "34/34 - 1s - loss: 2.1170e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 761/1000\n",
      "34/34 - 1s - loss: 1.8906e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 762/1000\n",
      "34/34 - 1s - loss: 2.1246e-04 - 1s/epoch - 35ms/step\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 763/1000\n",
      "34/34 - 1s - loss: 3.0109e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 764/1000\n",
      "34/34 - 1s - loss: 2.2678e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 765/1000\n",
      "34/34 - 1s - loss: 2.0043e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 766/1000\n",
      "34/34 - 1s - loss: 1.8521e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 767/1000\n",
      "34/34 - 1s - loss: 1.7107e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 768/1000\n",
      "34/34 - 1s - loss: 1.6460e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 769/1000\n",
      "34/34 - 1s - loss: 1.6658e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 770/1000\n",
      "34/34 - 1s - loss: 1.7437e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 771/1000\n",
      "34/34 - 1s - loss: 1.5733e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 772/1000\n",
      "34/34 - 1s - loss: 1.6574e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 773/1000\n",
      "34/34 - 1s - loss: 1.6833e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 774/1000\n",
      "34/34 - 1s - loss: 1.7065e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 775/1000\n",
      "34/34 - 1s - loss: 1.7992e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 776/1000\n",
      "34/34 - 1s - loss: 1.8943e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 777/1000\n",
      "34/34 - 1s - loss: 1.8321e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 778/1000\n",
      "34/34 - 1s - loss: 1.6867e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 779/1000\n",
      "34/34 - 1s - loss: 1.7825e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 780/1000\n",
      "34/34 - 1s - loss: 1.5245e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 781/1000\n",
      "34/34 - 1s - loss: 1.5686e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 782/1000\n",
      "34/34 - 1s - loss: 1.8528e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 783/1000\n",
      "34/34 - 1s - loss: 2.2373e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 784/1000\n",
      "34/34 - 1s - loss: 1.9718e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 785/1000\n",
      "34/34 - 1s - loss: 1.6267e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 786/1000\n",
      "34/34 - 1s - loss: 1.6013e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 787/1000\n",
      "34/34 - 1s - loss: 1.5806e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 788/1000\n",
      "34/34 - 1s - loss: 1.5515e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 789/1000\n",
      "34/34 - 1s - loss: 1.2897e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 790/1000\n",
      "34/34 - 1s - loss: 1.5202e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 791/1000\n",
      "34/34 - 1s - loss: 1.2666e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 792/1000\n",
      "34/34 - 1s - loss: 1.2756e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 793/1000\n",
      "34/34 - 1s - loss: 1.3756e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 794/1000\n",
      "34/34 - 1s - loss: 1.3476e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 795/1000\n",
      "34/34 - 1s - loss: 1.2312e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 796/1000\n",
      "34/34 - 1s - loss: 1.3768e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 797/1000\n",
      "34/34 - 1s - loss: 1.4887e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 798/1000\n",
      "34/34 - 1s - loss: 1.3293e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 799/1000\n",
      "34/34 - 1s - loss: 1.2526e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 800/1000\n",
      "34/34 - 1s - loss: 1.3374e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 801/1000\n",
      "34/34 - 1s - loss: 1.3054e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 802/1000\n",
      "34/34 - 1s - loss: 1.4284e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 803/1000\n",
      "34/34 - 1s - loss: 1.5212e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 804/1000\n",
      "34/34 - 1s - loss: 0.0013 - 1s/epoch - 37ms/step\n",
      "Epoch 805/1000\n",
      "34/34 - 1s - loss: 0.0044 - 1s/epoch - 36ms/step\n",
      "Epoch 806/1000\n",
      "34/34 - 1s - loss: 0.0135 - 1s/epoch - 37ms/step\n",
      "Epoch 807/1000\n",
      "34/34 - 1s - loss: 0.0091 - 1s/epoch - 36ms/step\n",
      "Epoch 808/1000\n",
      "34/34 - 1s - loss: 0.0046 - 1s/epoch - 36ms/step\n",
      "Epoch 809/1000\n",
      "34/34 - 1s - loss: 0.0026 - 1s/epoch - 35ms/step\n",
      "Epoch 810/1000\n",
      "34/34 - 1s - loss: 0.0019 - 1s/epoch - 34ms/step\n",
      "Epoch 811/1000\n",
      "34/34 - 1s - loss: 0.0032 - 1s/epoch - 34ms/step\n",
      "Epoch 812/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 34ms/step\n",
      "Epoch 813/1000\n",
      "34/34 - 1s - loss: 0.0014 - 1s/epoch - 36ms/step\n",
      "Epoch 814/1000\n",
      "34/34 - 1s - loss: 9.3985e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 815/1000\n",
      "34/34 - 1s - loss: 8.0095e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 816/1000\n",
      "34/34 - 1s - loss: 7.4421e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 817/1000\n",
      "34/34 - 1s - loss: 5.6055e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 818/1000\n",
      "34/34 - 1s - loss: 5.3514e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 819/1000\n",
      "34/34 - 1s - loss: 4.8216e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 820/1000\n",
      "34/34 - 1s - loss: 4.8870e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 821/1000\n",
      "34/34 - 1s - loss: 4.3252e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 822/1000\n",
      "34/34 - 1s - loss: 4.1351e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 823/1000\n",
      "34/34 - 1s - loss: 3.7905e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 824/1000\n",
      "34/34 - 1s - loss: 3.6958e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 825/1000\n",
      "34/34 - 1s - loss: 3.5229e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 826/1000\n",
      "34/34 - 1s - loss: 3.5586e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 827/1000\n",
      "34/34 - 1s - loss: 3.3732e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 828/1000\n",
      "34/34 - 1s - loss: 3.2890e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 829/1000\n",
      "34/34 - 1s - loss: 3.2296e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 830/1000\n",
      "34/34 - 1s - loss: 3.1009e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 831/1000\n",
      "34/34 - 1s - loss: 2.7789e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 832/1000\n",
      "34/34 - 1s - loss: 2.5506e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 833/1000\n",
      "34/34 - 1s - loss: 2.6085e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 834/1000\n",
      "34/34 - 1s - loss: 2.6472e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 835/1000\n",
      "34/34 - 1s - loss: 2.3132e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 836/1000\n",
      "34/34 - 1s - loss: 2.4134e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 837/1000\n",
      "34/34 - 1s - loss: 2.1813e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 838/1000\n",
      "34/34 - 1s - loss: 2.2868e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 839/1000\n",
      "34/34 - 1s - loss: 2.1100e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 840/1000\n",
      "34/34 - 1s - loss: 2.1089e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 841/1000\n",
      "34/34 - 1s - loss: 2.0835e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 842/1000\n",
      "34/34 - 1s - loss: 2.0699e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 843/1000\n",
      "34/34 - 1s - loss: 1.8187e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 844/1000\n",
      "34/34 - 1s - loss: 2.1149e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 845/1000\n",
      "34/34 - 1s - loss: 1.8413e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 846/1000\n",
      "34/34 - 1s - loss: 1.7451e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 847/1000\n",
      "34/34 - 1s - loss: 1.6647e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 848/1000\n",
      "34/34 - 1s - loss: 1.6593e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 849/1000\n",
      "34/34 - 1s - loss: 1.7037e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 850/1000\n",
      "34/34 - 1s - loss: 1.7299e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 851/1000\n",
      "34/34 - 1s - loss: 1.7691e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 852/1000\n",
      "34/34 - 1s - loss: 1.7434e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 853/1000\n",
      "34/34 - 1s - loss: 1.5668e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 854/1000\n",
      "34/34 - 1s - loss: 1.4314e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 855/1000\n",
      "34/34 - 1s - loss: 1.3595e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 856/1000\n",
      "34/34 - 1s - loss: 1.4091e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 857/1000\n",
      "34/34 - 1s - loss: 1.4230e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 858/1000\n",
      "34/34 - 1s - loss: 1.4822e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 859/1000\n",
      "34/34 - 1s - loss: 1.5064e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 860/1000\n",
      "34/34 - 1s - loss: 1.3935e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 861/1000\n",
      "34/34 - 1s - loss: 1.5968e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 862/1000\n",
      "34/34 - 1s - loss: 1.3782e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 863/1000\n",
      "34/34 - 1s - loss: 1.6976e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 864/1000\n",
      "34/34 - 1s - loss: 1.8972e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 865/1000\n",
      "34/34 - 1s - loss: 1.5306e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 866/1000\n",
      "34/34 - 1s - loss: 1.2880e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 867/1000\n",
      "34/34 - 1s - loss: 1.3665e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 868/1000\n",
      "34/34 - 1s - loss: 1.3558e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 869/1000\n",
      "34/34 - 1s - loss: 1.3332e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 870/1000\n",
      "34/34 - 1s - loss: 1.4159e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 871/1000\n",
      "34/34 - 1s - loss: 1.5437e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 872/1000\n",
      "34/34 - 1s - loss: 1.4164e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 873/1000\n",
      "34/34 - 1s - loss: 1.4514e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 874/1000\n",
      "34/34 - 1s - loss: 1.5911e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 875/1000\n",
      "34/34 - 1s - loss: 1.3221e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 876/1000\n",
      "34/34 - 1s - loss: 1.4848e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 877/1000\n",
      "34/34 - 1s - loss: 1.5216e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 878/1000\n",
      "34/34 - 1s - loss: 1.5516e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 879/1000\n",
      "34/34 - 1s - loss: 1.4951e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 880/1000\n",
      "34/34 - 1s - loss: 1.4126e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 881/1000\n",
      "34/34 - 1s - loss: 1.1746e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 882/1000\n",
      "34/34 - 1s - loss: 1.2469e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 883/1000\n",
      "34/34 - 1s - loss: 1.2636e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 884/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34/34 - 1s - loss: 1.1265e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 885/1000\n",
      "34/34 - 1s - loss: 1.2040e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 886/1000\n",
      "34/34 - 1s - loss: 1.1921e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 887/1000\n",
      "34/34 - 1s - loss: 1.2845e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 888/1000\n",
      "34/34 - 1s - loss: 1.3036e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 889/1000\n",
      "34/34 - 1s - loss: 1.3215e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 890/1000\n",
      "34/34 - 1s - loss: 1.2351e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 891/1000\n",
      "34/34 - 1s - loss: 1.0923e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 892/1000\n",
      "34/34 - 1s - loss: 1.1352e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 893/1000\n",
      "34/34 - 1s - loss: 1.0530e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 894/1000\n",
      "34/34 - 1s - loss: 1.0400e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 895/1000\n",
      "34/34 - 1s - loss: 1.2588e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 896/1000\n",
      "34/34 - 1s - loss: 1.2694e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 897/1000\n",
      "34/34 - 1s - loss: 1.0613e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 898/1000\n",
      "34/34 - 1s - loss: 1.2705e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 899/1000\n",
      "34/34 - 1s - loss: 1.2721e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 900/1000\n",
      "34/34 - 1s - loss: 1.3100e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 901/1000\n",
      "34/34 - 1s - loss: 1.9332e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 902/1000\n",
      "34/34 - 1s - loss: 3.9071e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 903/1000\n",
      "34/34 - 1s - loss: 2.1143e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 904/1000\n",
      "34/34 - 1s - loss: 4.3978e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 905/1000\n",
      "34/34 - 1s - loss: 6.1571e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 906/1000\n",
      "34/34 - 1s - loss: 0.0016 - 1s/epoch - 34ms/step\n",
      "Epoch 907/1000\n",
      "34/34 - 1s - loss: 0.0071 - 1s/epoch - 35ms/step\n",
      "Epoch 908/1000\n",
      "34/34 - 1s - loss: 0.0088 - 1s/epoch - 37ms/step\n",
      "Epoch 909/1000\n",
      "34/34 - 1s - loss: 0.0068 - 1s/epoch - 35ms/step\n",
      "Epoch 910/1000\n",
      "34/34 - 1s - loss: 0.0123 - 1s/epoch - 35ms/step\n",
      "Epoch 911/1000\n",
      "34/34 - 1s - loss: 0.0078 - 1s/epoch - 36ms/step\n",
      "Epoch 912/1000\n",
      "34/34 - 1s - loss: 0.0071 - 1s/epoch - 35ms/step\n",
      "Epoch 913/1000\n",
      "34/34 - 1s - loss: 0.0067 - 1s/epoch - 37ms/step\n",
      "Epoch 914/1000\n",
      "34/34 - 1s - loss: 0.0037 - 1s/epoch - 35ms/step\n",
      "Epoch 915/1000\n",
      "34/34 - 1s - loss: 0.0024 - 1s/epoch - 35ms/step\n",
      "Epoch 916/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 35ms/step\n",
      "Epoch 917/1000\n",
      "34/34 - 1s - loss: 0.0023 - 1s/epoch - 34ms/step\n",
      "Epoch 918/1000\n",
      "34/34 - 1s - loss: 0.0018 - 1s/epoch - 35ms/step\n",
      "Epoch 919/1000\n",
      "34/34 - 1s - loss: 0.0013 - 1s/epoch - 36ms/step\n",
      "Epoch 920/1000\n",
      "34/34 - 1s - loss: 9.5175e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 921/1000\n",
      "34/34 - 1s - loss: 0.0010 - 1s/epoch - 36ms/step\n",
      "Epoch 922/1000\n",
      "34/34 - 1s - loss: 5.2363e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 923/1000\n",
      "34/34 - 1s - loss: 4.7631e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 924/1000\n",
      "34/34 - 1s - loss: 4.3673e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 925/1000\n",
      "34/34 - 1s - loss: 4.0527e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 926/1000\n",
      "34/34 - 1s - loss: 4.3306e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 927/1000\n",
      "34/34 - 1s - loss: 3.5292e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 928/1000\n",
      "34/34 - 1s - loss: 3.2076e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 929/1000\n",
      "34/34 - 1s - loss: 2.9626e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 930/1000\n",
      "34/34 - 1s - loss: 2.8594e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 931/1000\n",
      "34/34 - 1s - loss: 2.9863e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 932/1000\n",
      "34/34 - 1s - loss: 2.7923e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 933/1000\n",
      "34/34 - 1s - loss: 2.6749e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 934/1000\n",
      "34/34 - 1s - loss: 2.3406e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 935/1000\n",
      "34/34 - 1s - loss: 2.2958e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 936/1000\n",
      "34/34 - 1s - loss: 2.2178e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 937/1000\n",
      "34/34 - 1s - loss: 2.2909e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 938/1000\n",
      "34/34 - 1s - loss: 2.1116e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 939/1000\n",
      "34/34 - 1s - loss: 2.0735e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 940/1000\n",
      "34/34 - 1s - loss: 2.1359e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 941/1000\n",
      "34/34 - 1s - loss: 2.1129e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 942/1000\n",
      "34/34 - 1s - loss: 2.0658e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 943/1000\n",
      "34/34 - 1s - loss: 1.9456e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 944/1000\n",
      "34/34 - 1s - loss: 1.8643e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 945/1000\n",
      "34/34 - 1s - loss: 1.8706e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 946/1000\n",
      "34/34 - 1s - loss: 2.0755e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 947/1000\n",
      "34/34 - 1s - loss: 1.9149e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 948/1000\n",
      "34/34 - 1s - loss: 1.5623e-04 - 1s/epoch - 37ms/step\n",
      "Epoch 949/1000\n",
      "34/34 - 1s - loss: 1.6659e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 950/1000\n",
      "34/34 - 1s - loss: 1.5374e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 951/1000\n",
      "34/34 - 1s - loss: 1.4276e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 952/1000\n",
      "34/34 - 1s - loss: 1.5730e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 953/1000\n",
      "34/34 - 1s - loss: 1.4785e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 954/1000\n",
      "34/34 - 1s - loss: 1.4644e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 955/1000\n",
      "34/34 - 1s - loss: 1.6151e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 956/1000\n",
      "34/34 - 1s - loss: 1.7286e-04 - 1s/epoch - 35ms/step\n",
      "Epoch 957/1000\n",
      "34/34 - 1s - loss: 1.5741e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 958/1000\n",
      "34/34 - 1s - loss: 1.8000e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 959/1000\n",
      "34/34 - 1s - loss: 2.1901e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 960/1000\n",
      "34/34 - 1s - loss: 1.9123e-04 - 1s/epoch - 34ms/step\n",
      "Epoch 961/1000\n",
      "34/34 - 1s - loss: 2.0412e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 962/1000\n",
      "34/34 - 1s - loss: 4.3073e-04 - 1s/epoch - 36ms/step\n",
      "Epoch 963/1000\n",
      "34/34 - 1s - loss: 6.0017e-04 - 1s/epoch - 31ms/step\n",
      "Epoch 964/1000\n",
      "34/34 - 1s - loss: 7.2014e-04 - 984ms/epoch - 29ms/step\n",
      "Epoch 965/1000\n",
      "34/34 - 1s - loss: 0.0013 - 978ms/epoch - 29ms/step\n",
      "Epoch 966/1000\n",
      "34/34 - 1s - loss: 0.0088 - 894ms/epoch - 26ms/step\n",
      "Epoch 967/1000\n",
      "34/34 - 1s - loss: 0.0019 - 922ms/epoch - 27ms/step\n",
      "Epoch 968/1000\n",
      "34/34 - 1s - loss: 0.0012 - 943ms/epoch - 28ms/step\n",
      "Epoch 969/1000\n",
      "34/34 - 1s - loss: 0.0029 - 867ms/epoch - 25ms/step\n",
      "Epoch 970/1000\n",
      "34/34 - 1s - loss: 0.0010 - 821ms/epoch - 24ms/step\n",
      "Epoch 971/1000\n",
      "34/34 - 1s - loss: 6.0194e-04 - 813ms/epoch - 24ms/step\n",
      "Epoch 972/1000\n",
      "34/34 - 1s - loss: 6.1367e-04 - 837ms/epoch - 25ms/step\n",
      "Epoch 973/1000\n",
      "34/34 - 1s - loss: 4.1598e-04 - 812ms/epoch - 24ms/step\n",
      "Epoch 974/1000\n",
      "34/34 - 1s - loss: 3.6924e-04 - 827ms/epoch - 24ms/step\n",
      "Epoch 975/1000\n",
      "34/34 - 1s - loss: 3.6893e-04 - 830ms/epoch - 24ms/step\n",
      "Epoch 976/1000\n",
      "34/34 - 1s - loss: 3.3032e-04 - 822ms/epoch - 24ms/step\n",
      "Epoch 977/1000\n",
      "34/34 - 1s - loss: 3.0156e-04 - 857ms/epoch - 25ms/step\n",
      "Epoch 978/1000\n",
      "34/34 - 1s - loss: 3.0807e-04 - 876ms/epoch - 26ms/step\n",
      "Epoch 979/1000\n",
      "34/34 - 1s - loss: 2.6898e-04 - 930ms/epoch - 27ms/step\n",
      "Epoch 980/1000\n",
      "34/34 - 1s - loss: 2.6897e-04 - 900ms/epoch - 26ms/step\n",
      "Epoch 981/1000\n",
      "34/34 - 1s - loss: 2.2968e-04 - 870ms/epoch - 26ms/step\n",
      "Epoch 982/1000\n",
      "34/34 - 1s - loss: 2.1283e-04 - 811ms/epoch - 24ms/step\n",
      "Epoch 983/1000\n",
      "34/34 - 1s - loss: 2.0168e-04 - 832ms/epoch - 24ms/step\n",
      "Epoch 984/1000\n",
      "34/34 - 1s - loss: 1.9444e-04 - 815ms/epoch - 24ms/step\n",
      "Epoch 985/1000\n",
      "34/34 - 1s - loss: 1.8690e-04 - 813ms/epoch - 24ms/step\n",
      "Epoch 986/1000\n",
      "34/34 - 1s - loss: 1.7815e-04 - 844ms/epoch - 25ms/step\n",
      "Epoch 987/1000\n",
      "34/34 - 1s - loss: 1.9131e-04 - 817ms/epoch - 24ms/step\n",
      "Epoch 988/1000\n",
      "34/34 - 1s - loss: 1.9280e-04 - 832ms/epoch - 24ms/step\n",
      "Epoch 989/1000\n",
      "34/34 - 1s - loss: 1.8084e-04 - 812ms/epoch - 24ms/step\n",
      "Epoch 990/1000\n",
      "34/34 - 1s - loss: 1.5760e-04 - 840ms/epoch - 25ms/step\n",
      "Epoch 991/1000\n",
      "34/34 - 1s - loss: 1.6613e-04 - 836ms/epoch - 25ms/step\n",
      "Epoch 992/1000\n",
      "34/34 - 1s - loss: 1.7559e-04 - 805ms/epoch - 24ms/step\n",
      "Epoch 993/1000\n",
      "34/34 - 1s - loss: 1.5671e-04 - 825ms/epoch - 24ms/step\n",
      "Epoch 994/1000\n",
      "34/34 - 1s - loss: 1.4825e-04 - 831ms/epoch - 24ms/step\n",
      "Epoch 995/1000\n",
      "34/34 - 1s - loss: 1.3717e-04 - 820ms/epoch - 24ms/step\n",
      "Epoch 996/1000\n",
      "34/34 - 1s - loss: 1.2856e-04 - 819ms/epoch - 24ms/step\n",
      "Epoch 997/1000\n",
      "34/34 - 1s - loss: 1.3484e-04 - 853ms/epoch - 25ms/step\n",
      "Epoch 998/1000\n",
      "34/34 - 1s - loss: 1.2594e-04 - 1s/epoch - 31ms/step\n",
      "Epoch 999/1000\n",
      "34/34 - 1s - loss: 1.2094e-04 - 990ms/epoch - 29ms/step\n",
      "Epoch 1000/1000\n",
      "34/34 - 1s - loss: 1.6355e-04 - 1s/epoch - 30ms/step\n"
     ]
    }
   ],
   "source": [
    "#TTrain[np.isnan(TTrain)]=0\n",
    "with tf.device('/gpu:0'):\n",
    "    start = dt.now()\n",
    "    hist=model.fit(XTrain_s,np.array(TTrain),epochs=noEpochs,verbose=2)\n",
    "    trainTime=(dt.now() - start)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c494eda6",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import matplotlib.pyplot as plt\n",
    "# from matplotlib.pyplot import figure\n",
    "# plt.figure(figsize=(35,15))\n",
    "# plt.plot(hist.history['loss'])\n",
    "# plt.plot(hist.history['val_loss'])\n",
    "# plt.title('model train vs validation loss')\n",
    "# plt.ylabel('loss')\n",
    "# plt.xlabel('epochs')\n",
    "# plt.legend(['train', 'validation'], loc='upper right',prop={'size': 10})\n",
    "# fig = plt.gcf()\n",
    "# plt.yscale(\"log\")\n",
    "# fig.set_size_inches(20.5, 10.5)\n",
    "# plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a852559c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train time:  0:20:38.225338 s\n"
     ]
    }
   ],
   "source": [
    "print(\"Train time: \",trainTime,\"s\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e24d20bf",
   "metadata": {},
   "source": [
    "## Inference"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "5df4584d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "9/9 [==============================] - 0s 9ms/step\n"
     ]
    }
   ],
   "source": [
    "with tf.device('/cpu:0'):\n",
    "    startInf = dt.now()\n",
    "    Y_hat=model.predict(XTest_s)\n",
    "    inferenceTime=(dt.now() - startInf)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "eda1f3f9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Inference time:  0:00:00.484836 s\n"
     ]
    }
   ],
   "source": [
    "print(\"Inference time: \",inferenceTime,\"s\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "0edfc195",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1075, 50, 4)\n",
      "(1075,)\n"
     ]
    }
   ],
   "source": [
    "print(XTrain_s.shape)\n",
    "print(TTrain.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d28eda96",
   "metadata": {},
   "source": [
    "## Plot prediction "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "e0d91afc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1476x756 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.pyplot import figure\n",
    "plt.figure(figsize=(15,7))\n",
    "plt.plot(TTest,color='blue',label='Actual',alpha=0.5,linewidth=6.0)\n",
    "plt.plot(Y_hat,color='red',label='Predicted')\n",
    "plt.xlabel('Steps',fontsize=25,backgroundcolor='w')\n",
    "plt.ylabel(\"Oxygen uptake\\n\"+\"$(mlxmin^{-1}xkg^{-1})$\",fontsize=25,backgroundcolor='w')\n",
    "plt.tight_layout()\n",
    "\n",
    "plt.yticks(fontsize=25)\n",
    "plt.xticks(fontsize=25)\n",
    "\n",
    "plt.gca().spines['right'].set_visible(False)\n",
    "plt.gca().spines['left'].set_linewidth(2)\n",
    "plt.gca().spines['top'].set_visible(False)\n",
    "plt.gca().spines['bottom'].set_linewidth(2)\n",
    "\n",
    "plt.ylim(0)\n",
    "plt.xlim(0)\n",
    "\n",
    "plt.legend(prop={'size': 25},loc='upper left')\n",
    "fig = plt.gcf()\n",
    "fig.set_size_inches(20.5, 10.5)\n",
    "plt.savefig(figPredictName)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f8ef8bc",
   "metadata": {},
   "source": [
    "## Save model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "95627e67",
   "metadata": {},
   "outputs": [],
   "source": [
    "if savedModel:\n",
    "    model.save(modelSaveName)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "4e33285e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(269, 1)\n",
      "(269,)\n",
      "9/9 [==============================] - 0s 8ms/step - loss: 0.0065\n"
     ]
    }
   ],
   "source": [
    "print(np.shape(Y_hat))\n",
    "print(np.shape(TTest))\n",
    "#Y_hat=np.expand_dims(Y_hat,axis=0)\n",
    "#TTest=np.expand_dims(TTest,axis=0)\n",
    "#print(np.shape(XTrain_s))\n",
    "#print(np.shape(TTrain))\n",
    "results = model.evaluate(XTest_s,TTest)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "7441ce70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABLAAAAJ4CAYAAAB1ShosAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACaT0lEQVR4nO3dd5hcZfnG8fvdTe+QTg0ECKFDIAhBitKkSAeRIlJUflQRFBQRCwqKFMFCsSBFelMREAQiNSTUUBJaKEmAFNLbbvb9/fHM65ydPdM2M+fM7nw/15VrslPPzp56n+d9jvPeCwAAAAAAAKhVDWlPAAAAAAAAAFAIARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAACXLO/cE598PIzyc75z5xzi1yzg10zo1zzr2V+fnAFCe13Zxz05xzu6c9HfVoVb77pP5uzrldnXMfVftzCny+d85tkPl/q+Wxip95nHPuySq874XOuZsq/b4AANQiAiwAACokEwAsdc4tdM7Nc8497Zz7lnPuf9tb7/23vPc/zTy/q6TLJO3pve/jvZ8j6SeSrs78fG8qv0iKCL+QpOjyWIhz7nHn3IlJTBMAAIhHgAUAQGXt773vK2ldSRdL+p6kP+Z57lBJPSS9Frlv3ZyfS+ac69Ke16F6+JtUF98vAAD1gwALAIAq8N7P997fL+kISV9zzm0mSc65vzjnfuac20jSlMzT5znn/uOce0fS+pL+nhlC2N05198590fn3Ezn3PTMaxsz73Wcc+4p59zlzrm5ki7MvOZS59wHmaGJf3DO9cw8f1fn3EfOue845z7NvOfXwzQ753o6537tnHvfOTffOfdk5LWfy1SUzXPOveyc27XIV7Cdc+5159xnzrk/O+d6RD5nP+fcS5EqtS0y998oaZ3I7/9d59wNzrnvZB5fMzP86/8yP2/gnJvrnHOF3jfz2BrOubucc7Occ+85506PPHahc+5259xfM9Vzrznnts33iznn9nTOTcl8R79zzj0RqnPy/E36Z957Vua7PT9U5eUOAXPOjcj8jl0yPz/unPtp5j0XOuceds4Nijz/mMx7znHO/SBnOhucc+c6597JPH67c271Ul4b8zvv65x70Tm3wDn3oXPuwphp/lpmvpsdfb/MfPWXzLzwuqTtinyWd86d7px7N/Nev4p8X2XN85nXnONsXp/hnDs+57P+4pz7WeTnAzLz0ILM97a3c+4iSZ+XdHVmvrw689yNnXP/zsyDU5xzh0feZ6Bz7v7M+0yQNLLA7/ugc+7UnPteds4dnPn/lZnvfIFzbpJz7vN53qfN0EwXqWgsNj8AAFDrCLAAAKgi7/0ESR/JDoCj90+VtGnmxwHe+y9470dK+kBWxdXHe79c0g2SmiVtIGlrSXtKig5l2l7Su5KGSLpI0iWSNpK0VeY1a0q6IPL8YZL6Z+4/QdJvnXOrZR67VNIYSTtKWl3SdyW1OOfWlPRPST/L3H+2pLucc4ML/OpHSdpLduC+kaTzJck5t42kP0n6pqSBkq6RdL9zrrv3/pic3/+Xkp6QtGvmPXfJ/K67ZH7eWdJ/vfe+0Ptmwo+/S3o583t/UdKZzrm9ItP7ZUm3Shog6X5JV8f9Upnw6E5J52U+Z0rm+4rK/ZtcJfvO189M+7GSvq7SfTXz/CGSusm+fznnNpH0e0nHSFojMz1rRV53uqQDM5+5hqTPJP22xNfmWpyZ7gGS9pV0smvbo20nSaNk3+8FzrnRmft/JJsPRsrmia+V8DsfJGlbSdtIOkBSNHgqeZ53zu0t+772kLShpLzDU51zYyX9VdI5md9zZ0nTvPc/kPRfSadm5stTnXO9Jf1b0i2Z6ThS0u+cc2GZ/q2kZZKGZ6a9VXCW45bM68N0bCKrxPxn5q7nM7/b6pnn3uEigXAZ8s4PAAB0BARYAABU3wzZwWdZnHNDJX1J0pne+8Xe+08lXS7pK9H39t5f5b1vlh0wnyTp2977ud77hZJ+nvP8Jkk/8d43ee8fkLRI0qhMyHO8pDO899O99yu9909nQrSjJT3gvX/Ae9/ivf+3pImS9ikw+Vd77z/03s+VhQzhAP0kSdd475/LfMYNkpZL+lye93lC0ucz07ezpF9KGpd5bJfM48XedztJg733P/Her/Devyvpupzv5cnM77dS0o2StswzPftIes17f3fmO/+NpI9znhP9m6yQVeGd571f6L2fJunXsuCoVH/23k/13i+VdLsszJCkQyX9w3s/PvN3+qGklsjrvinpB977jzKPXyjp0Ex1V7HXtuK9f9x7/2rm7/+KpL8pGyQGP/beL/XevywLC8N3eLikizLz5Iey76yYSzLP/0DSFYoEPCpvnj9c9v1N9t4vznwH+Zwg6U/e+39nfs/p3vs38zx3P1m49WfvfbP3/gVJd8m+30ZJh0i6ILPcTpYF0fncI2kr59y6mZ+PknR35u8i7/1N3vs5mc/5taTusqCwXIXmBwAAah4bLAAAqm9NSXPb8bp1JXWVNNPZKDnJTj59GHlO9P+DJfWSNCnyfCepMfKcOZkD/2CJpD6SBsn6cb2TZzoOc87tH7mvq6THCkx7dLrel1V8hPf6mnPutMjj3SKPt+K9f8c5t0gW2nxe0k8lneCcGyULUEIYUuh9V0pawzk3L/JYo6yqJoiGUEsk9XDOdcn5rpR5v//9bpnqr9wr6kV/90GZ6Xg/ct/7snmiVLnT1ifPtCx2zs2JPHddSfc456LB1EpZ77Vir23FObe9rKfbZpnfp7ukO9oznWr9XeSTb/7JfazYPL+GpEklfvbakh4oYdok+263z5mnusjCz8GZ/5f0O3vvFzrn/ikL3S7J3H4jPO5sCO2Jst/FS+onm6/KVWh+mN6O9wMAIFEEWAAAVJFzbjtZWPFkO17+oayKaFBMkBL4yP9nS1oqaVPvfbkHpLNl1SwjZdUzudNxo/f+pDLeb+3I/9eRVaGF97rIe39Rntf5mPuekFUMdfPeT3fOPSEbzraapJeKva9zbgdJ73nvNyxj+vOZqchQO2epSe7Qu9y/SZMsPHg9c986ygYGi2UBTDCszGkJw/TknOslGwoYfCjpeO/9U7kvdM4Ve22uW2TDKr/kvV/mnLtCpYcoM2XzQ7g4wTolvCb3+TMij5Uzz4fPDgp99ofK36sqd778UNIT3vs9cp+YqcBqznxuqOAq9jv/TdKPnHPjJfVUJhzO9Lv6nmxY5mve+xbn3GeykC5Xq3kpMx3RYb555wcAADoChhACAFAFzrl+zrn9ZH2VbvLev1rue3jvZ0p6WNKvM+/X4Jwb6ZzLHboVnt8iGxp3uXNuSGY61szp9ZTvs1pkPaQuc9bwvNE5t4NzrrukmyTt75zbK3N/j0zD6EI9k05xzq2VaRL9fUm3Ze6/TtK3nHPbO9PbWYPwvpnHP5H1iop6QtKpksZnfn5c0mmyYX8rS3jfCZIWOOe+56yheKNzbrNMuFiuf0ra3Dl3YGbo1SkqEDplpu92SRc55/pmhomdJftOJQvgdnbOreOc6y/rrVWqOyXt55zbyTnXTdJP1Hrf7g+Zz11Xkpxzg51zB5T42lx9Jc3NhFdjZX25SnW7pPOcc6tl5pnTir1A0jmZ568t6Qxl559WSpjnb5d0nHNuk0xI96MCn/lHSV93zn0xs6yt6ZzbOPNY7nz5D0kbOWuE3zXzbzvn3OjM3/xuWYP5XpmeVsX6fj0gCzl/Ium2zO8l2ffeLGmWpC7OuQtkFVhxpsoqB/d1znWV9Z3rHnm80PwAAEDNI8ACAKCy/u6cWyirdviBpMtUXsPuXMfKhmy9Lmu6fKesMXQ+35P0tqRnnXMLJD2i0vvlnC3pVVnT6Lmy4UwNmb5FB8iCqFmy3+0cFd6PuEUWvr2b+fczSfLeT5T1LLo68/u8Lem4yOt+Iel8Z1cSPDtz3xOyA/kQYD0pqzQJPxd830ygsL9sGOJ7sqqd62WN1cvivZ8t6TBZL645kjaR9QNbXuBlp8mqY97NTPstsrBQmX5it0l6RTbU7R9lTMtrsgDtFlml0WeyCwYEV8oa0j+cmSeflTVAL+W1uf5P0k8y73OBLBgq1Y9lQ+jek80TN5bwmvtk38dLstDwjwWem3ee997/S9ZD6z+Z5/wn35t4u+DC12V95ubL5rvQl+pKWb+oz5xzv8n02tpTNtxvhmz45CXKBkanyoZQfizpL5L+XOiXzfSkulvWZP6WyEMPSfqXLJx6X1Yl+WGbN7D3mC/7O10vq/BbrBLnB0lydoXF2CscAgBQC5z3cZX6AAAAKMZZc/mPJB3lvS/UEwwlcs55SRt6799Oe1oAAEDtoAILAACgDJmhlAMywyu/L+tH9GzKkwUAANCpEWABAACUZwfZ1Rpny4YmHui9X5ruJAEAAHRuDCEEAAAAAABATaMCCwAAAAAAADWNAAsAAAAAAAA1rUvaE1AJgwYN8iNGjEh7MgAAAAAAANBOkyZNmu29Hxz3WKcIsEaMGKGJEyemPRkAAAAAAABoJ+fc+/keYwghAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYAAAAAAABqGgEWAAAAAAAAahoBFgAAAAAAAGoaARYARL39tnTMMVJTU9pTAgAAAADIIMACgKgnnpBuukn64IO0pwQAAAAAkEGABQBRzc12SwUWAAAAANQMAiwAiAoB1ooV6U4HAAAAAOB/CLAAIIoKLAAAAACoOQRYABBFgAUAAAAANYcACwCiQnDFEEIAAAAAqBkEWAAQRQUWAAAAANQcAiwAiKKJOwAAAADUHAIsAIiiAgsAAAAAag4BFgBEUYEFAAAAADWHAAsAoqjAAgAAAICaQ4AFAFEhuCLAAgAAAICaQYAFAFEMIQQAAACAmkOABQBRDCEEAAAAgJpDgAUAUVRgAQAAAEDNIcACgCgqsAAAAACg5hBgAUAUFVgAAAAAUHMIsAAgigosAAAAAKg5BFgAEBWCKwIsAAAAAKgZBFgAEMUQQgAAAACoOQRYABDFEEIAAAAAqDkEWAAQRQUWAAAAANQcAiwAiKICCwAAAABqDgEWAEQRYAEAAABAzSHAAoCoEFwxhBAAAAAAagYBFgBEUYEFAAAAADWHAAsAomjiDgAAAAA1hwALAKKowAIAAACAmkOABQBRVGABAAAAQM0hwAKAKCqwAAAAAKDmEGABQBQBFgAAAADUHAIsAIgKwRVDCAEAQBrmzpW++lXps8/SnhIAqCkEWAAQRQUWAABI06RJ0t/+Jj3zTNpTAgA1hQALAKJo4g4AANIUTqJ9+mm60wEANYYACwCiqMACAABpCvsin3yS7nQAQI0hwAKAKCqwAABAmqjAAoBYBFgAEEUFFgAASBMVWAAQiwALAKJCcEWABQAA0kAFFgDEIsACgKClxf5JDCEEAADpCBVYBFgA0AoBFgAEK1dm/08FFgAASANDCAEgFgEWAARhh7GxkQosAACQjnASbdasbGU4AIAACwD+JwRYvXpZNRY7jQAAIGlhf2TlSmnu3HSnBQBqCAEWAARhh7FnT7tlGCEAAEhadP+DPlgA8D8EWAAQRCuwJAIsAACQvLA/ItEHCwAiCLAAIAiBFRVYAAAgLVRgAUAsAiwACHIrsGjkDgAAkhatwCLAAoD/6ZL2BABAzaAHFgAASFtTk+Sc1NDAEEIAiKjZAMs5N03SQkkrJTV777dNd4oAdHpUYAEAgLQ1N0vdukmrrUYFFgBE1GyAlbGb93522hMBoE7QxB0AAKStqUnq0kUaMoQKLACIoAcWAAQMIQQAAGlrbpa6drUAiwosAPifWg6wvKSHnXOTnHPfSHtiANSBEFgxhBAAAKQlVGANHUoFFgBE1PIQwnHe+xnOuSGS/u2ce9N7Pz48mAm1viFJ66yzTlrTCKAzoQILAACkjQosAIhVsxVY3vsZmdtPJd0jaWzO49d677f13m87ePDgNCYRQGdDE3cAAJC2aAXW4sX2DwBQmwGWc663c65v+L+kPSVNTneqAHR6VGABAIC0RSuwJKqwACCjJgMsSUMlPemce1nSBEn/9N4/mPI0AejsqMACAABpi1ZgSQRYAJBRkz2wvPfvStoy7ekAUGdyAywqsAAAQNJyK7Bo5A4Akmq3AgsAkscQQgAAkLZQgcUQQgBohQALAIIQWDGEEAAApKW5mQALAGIQYAFAQAUWAABIW1OTDSHs0UPq148hhACQQYAFAAFN3AEAQNpCBZZkjdypwAIASQRYAJBFBRYAAEhbaOIu2TBCKrAAQBIBFgBkUYEFAADSFpq4SxZgUYEFAJIIsAAgKzfAogILAAAkLVqBNXQoFVgAkEGABQBBCKwYQggAANKSW4E1Z072JBsA1DECLAAIcntgMYQQAAAkLbcCy3sLsQCgzhFgAUAQAqzu3aWGBiqwAABA8nIrsCSGEQKACLAAICsEWF262JlPKrAAAEDSciuwJBq5A4AIsAAgKxpgdetGBRYAAEgeFVgAEIsACwCCEGB17UoFFgAASEe0AisEWFRgAQABFgD8T6i4amigAgsAAKQjWoE1YICFWVRgAQABFgD8T3Oz7TA6ZzuLBFgAACBp0Qos56wKiwosACDAAoD/CQGWZBVYDCEEAABJi1ZgSdbInQALAAiwAOB/ogEWFVgAACAN0QosySqwGEIIAARYAPA/uQEWFVgAACBpVGABQCwCLAAIomc8aeIOAACS1tJi/+IqsLxPb7oAoAYQYAFAwBBCAACQpuZmu41WYA0ZIi1fLi1cmM40AUCNIMACgCBask8TdwAAkLS4AGvwYLudPTv56QGAGkKABQABFVgAACBNYd8jOoSwb1+7XbQo+ekBgBpCgAUAAU3cAQBAmuIqsPr0sVuGEAKocwRYABBEAyyauAMAgKTFVWCFAIsKLAB1jgALAAIqsAAAQJriKrAYQggAkgiwACCruTl7xpMKLAAAkLQQYMVVYDGEEECdI8ACgCB6FUKauAMAgKSFfQ8qsACgDQIsAAhye2AxhBAAACSJCiwAyIsAC1nNzdKrr6Y9FUB6cntgUYEFAACSFFeB1b271NhIBRaAukeAhaw77pC22kr65JP4x2+5Rfre9xKdJCBRNHEHAABpiqvAcs6GERJgAahzBFjI+vhjqaVFmjUr/vH77pNuuCHZaQKSlDuEkAosAACQpLgKLMmGETKEEECdI8BCVjirs2BB/OPz5+d/DOgMolchrMUKrKuukj7/+bSnAgAAVEtcBZZkARYVWADqHAEWshYvttt8IdWCBdLSpVSloPPKrcBqbpa8T3eaol58UZowIe2pAAAA1ZKvAoshhABAgIWIYhVY4X7Kl9FZNTW17oElZc+E1oIFC6wqbNmytKcEAABUQ6EKLPbBAdQ5AixklRpgMYwQnVVuE3eptoYRhh1XlkEAADonKrAAIC8CLGQRYKHe5Q4hlGpryGxY9ubPT3c6AABAdVCBBQB5EWAhq1CA5T0BFjq/Wq/AIsACAKBzK3QVQiqwANQ5AixkFQqwFi/ONrPm4BmdVa1XYDGEEACAzi1fBRZDCAGAAAsRYaMYV54cPWDm4BmdVXNzdocx3NZSgEUFFpCu666TRo1KeyoAdGaFKrAWL5ZaWpKfJgCoEQRYyCpUgUWAhXoQvQphqMCqlSGE3lOBBaRt4kRp6lSuBAqgegr1wJIsxAKAOkWAhSwCLNS7uB5YtVKBtWRJ9qwrFVhAOj75xG5ZBgFUS6GrEEoMIwRQ1wiwkEWAhXpXy03co8sdB89AOkKANW9eqpMBoBMLFVhxQwglrkQIoK4RYMG0tGRLkgmwUK9quYk7yyCQPiqwAFRb2O+Ia+IuUYEFoK4RYMEsWZL9f6EAyzl23NF51XIFVvSMK8sgkA4CLADVRgUWAORFgAUTzuZ07Vo4wBo+nOoPdE4tLdYoPQRXtVyBxcEzkLxFi7Ine1gGAVRLvgqsEGBRgQWgjhFgwYThgyGg8r714+Hgec01CbDQOeWe8ay1Ju5huevWjWUQ9eWDD6TLLmu7XUpaqL6S6IEFoHryVWAxhBAACLCQETaGa6xhlSjRIYWSHTD37CkNHMjBMzqn3Kv+1NoQwmiITPUH6slVV0nf+U7rACkN0c9nGQRQLcUqsBhCCKCOEWDBRAMsqW1ItWCB1K+f/SPAQmeUe8az1oYQhh3Wtdbi4Bn15Zln7JYAC0A9aG62nrMNOYdpDCEEAAIsZJQaYPXvT4CFzinfEMJaq8Baay2WQdSPFSukiRPt/59+mu60MIQQQBKam9tWX0kEWAAgAiwEYWM4fLjdUoGFelPrFVgLFtgO7ZAhVH+gfrz0krR8uf2/Viqw1liDZRBA9TQ1te1/Jdk+QPfuDCEEUNcIsGDKGUK4eHH2YB/oLGq9AmvhQmvg2r+//X/lyrSnCKi+MHxQqo0KrNVXlwYNIsACUD35KrAk2w+gAgtAHSPAgiknwJI4+4POJwRYYaexFiuwwjBeiR1Y1Ienn7Zhs9261UYF1tChtgwSYAGolnwVWJINI2QfHEAdI8CCKTfAYhghOpt8VyGs1QCLA2jUg2eekXbc0YbO1lKARQ8sANVSqAKrTx9OYAGoawRYMIsW2RVPhg61nwmwUG86whBClkHUk+nTpQ8/tABr6NDaGEI4dKg0YAABMoDqKVSBxRBCAHWOAAtm0SI7qxN3cOw9ARY6v47QxD30wJI4gEbnF/pf7bCDBUfVqsC6//7SgmqGEAJIQrEKLIYQAqhjBFgwIcDq3t3+RQOqZcvsID46fIkAC51NrVdg5YbIHECjs3vmGalHD2mrrWwIYTUqsKZMkQ44QLrjjsLPW7rUDhqjAZb3lZ8eACjWA4sKLAB1jAALJgRYkh0gR8/uhLCKCix0ZrkBVmOjDautlQqsMISQEBn14plnpDFjrBoyDCGsdGg0a5bdvvNO4eeF6q8whHDlSrsiLwBUWrGrEFKBBaCOEWDBRAOsvn1bHxzHBVhUf6Czyb0KoXP2/1qqwGIIIerF8uXSpEk2fFCyCqwVKyo/34f3mzat8POiARbLIIBqogILAPIiwIJZvLh1BVaxAIvqD3Q2uRVYkgVYtVCBFao9CJFRL154wQKrHXe0n8MFRirdBytcTfD99ws/jwALQFKKVWARYAGoYwRYMLlDCPMFWL17W2UKARY6mxBURQOsbt1qI8AKO6thGWxsZBlE5xZt4C5ZBZZU+T5Yq1KBFcIvoN7MmiU99ljaU9F5FavAWr68NvZNACAFBFgwpQZYDQ1thxgCnUG+CqxaGEIYlre+fS1A7teP6g90bs88I40YIQ0bZj9XuwLrww+t0jGf8LlDhlgPLIllEPXryiulPffMbjdRWcWuQihRhQWgbhFgwZQaYMU9DnQGcQFWrVRgxS2DHDyjM3vmmWz1lVT9CqymJmnmzPzP++QTq7zq0YMhhMAHH9g2k33B6ihUgdW3r90SYAGoUwRYMOUEWP37s9OCzqeWK7DCFYdYBlEPPvxQmj492/9KkgYNsurDalVgSYWHEX7ySbYKjCGEqHfTp9sty0B1NDcXHkIocSVCAHWLAAtm0SLrrSPlD7DCTjvVH+iMOkIFVjjz2r8/yyDK4710yinSU0+lPSXFTZxot2PHZu/r0kUaOLDyAdb8+dllvlAj97gAi2UQ9WrGDLslwKqOpiaGEAJAHgRYsAqTpqbWFVjLl9s/yQ6eu3WTunfPPk71BzqbEGBFdxpr5SqEDOPFqpo/X/rd76Qbbkh7Sor76CO7XW+91vcPHVr5IYTz5kkbb2z/L7UCq1cvC70IsFCvqMCqrkIVWOFEFhVYAOoUARayZ3GiAZaU3TguWJC9LzzOwTM6m7irENbyEEIOnlGOcMD56qvpTkcpZsywZW/gwNb3Dx1anQqs4cOtx1apFVjOsQyifi1cmN0mEWBVBxVYAJAXARbyB1ghpCLAQj3oSEMIGcaLcoUhP6++KrW0pDstxcyYYaFSQ84uypAh1anAGjDArniYrwJr+XJ7XgiwJAuwOHhHPQrrEolloFpKqcAiwAJQpwiwQIAFSLXdxD2uBxbLIMoRDjoXLy48VK4WzJghrbFG2/urVYHVv7+07rr5v5cQmuUGWITIqEcEWNVXSgUWQwgB1CkCLJQfYPXvb69ZuTK5aQSqrdYrsHr2zO7Q9u9vwdqyZelOFzqO6EHnK6+kNx2lCBVYuYYMsYO2pUsr91nRCqwPPoivTguhWTTAGjCAAAv1KQxHlgiwqqWUqxBSgQWgThFgoX0VWNHXAZ1BLVdgLVyYrb6SsssgB9Ao1YwZ1nxcqv0+WIUqsKTKDSNcscLCsFCBtXx5fIVXXIDFEELUqxCGd+vGMlAtzc35K7DCFcPZBwc6D++liy+W3nor7SnpEAiwYENKpPIDLA6e0ZnEXYWwliqwcqsgw/1AKWbMsKv6jRxZ2wHW0qV2UBwXYA0ZYreVCrDCNqx/f6vAkuIbuecLsNgGoh5Nn27bo+HDCbCqpakpfwVWQ4OFWAwhBDqPd96RzjtP+tWv0p6SDoEAC+2vwOLgGZ1JvgqsWg6wOIBGqaZPt1Bo881rewjhzJl2W6gCq1J9sMLyM2CAVWBJ8X2wGEIIZIUKyQEDCLCqpVAFlmT761RgAZ3HhAl2+49/WDUWCiLAAgEWIGWDKoYQojMKB51bbGEl6pXsI1VJYXhSEhVY4eA7DCGU8gdYffpkh2CG1yxcWPtXdAQqbfp0ac01CbCqqVAFlmT7A1RgoTNqaZGuvFKaMyftKUlWCLBmzpRefDHdaekACLDQNsDq1ctKlBcsyDaKJsBCZ1frTdwZQoj2ammxnaJQgdXSIr3+euvnrFghXX99+vN7KQFWNSqw+vaVBg7MP4QwWn0l2TLofdtl8Nvfls4+uzLTt6pOPFH65S/Tngp0NlRgVV/SFVgrVkinny5NnVq590RlvfSSNGqUNGtW2lNSXZMmSWeeKd1yS9pTkqwJE6TRoyXnrAoLBRFgIbsRDI0hnbOd+QULsmd4CLDQ2dVyE3f60GFVzJ5t83eowJLa9sG69VbppJOkBx5IfvqiCgVYvXrZgVulAqxoBZZkVVj5KrDiAiyp7TL497+n/x1KViXzxz9Kv/sdwxFQOS0ttoxSgVVdpVRgVTLAuuMO6aqrpMsvr9x7orIef9wCxlCp01lNmmS3776b7nQkqalJeuEFaZ99pM99TvrnP9OeoppHgAXbCPbo0Xpj2a+fhVchpKL6A51dc7OFtw2R1WKtVGDlDiGkBxbKEQ2FRo6UevZs2wfr9tvtthL9sVaulHbbTTr//PJfO2OG1L27tNpq8Y8PHVr5Ju4DBtjtiBGlV2CF10SXwZUrpQ8+sBAs7dDonnvs9v33uaoRKmfOHNsmUoFVPd7buqRYBVYlhxD+5jd2e/fd9tmoPW+/bbdvvFHa82+8UbrssupNT7VMnGi377yT7nQk6dVX7SrIY8dK++1nIWWlTtR1UgRYsAArDB8M+vWzgCouwKICC51Rc3PbM561XoHFMohSTJ9ut2uuKTU2Sptu2roC67PPpIcftv9X4gqFN91kZ4sffLD8186YYVc3cy7+8aFDq1+BlRs+ffJJdvhiEF4TPYCfPt0O7pcutaq3OM3NyYRbd96ZDd3C3xZYVdF1yYABFqKE6mVURlw1eK5KDiF87jk7YN5tNzs58N//VuZ9UVnlBliXXUaA1VGEqroQYEm1UcldwwiwYBvBMHwwKBRghbCL6g90JnE9J2rhKoQrVtiZmegy2LWrVdGwDKIUucPycq9EeN99Np+PGLHqFVhLlmQrr15/vfwm56G/Tj5DhlS2Asu57LI1YoSFT9EeI01NVnVSyhDC997L/j9uKOLKlRaSXX11JaY+v08+kcaPl775Tau4I8BCpUTXJaEKkRMplRUCrEIVWJUcQnjVVfZ+t9xi+xV33FGZ90VllRNgLV0qTZ5sgfOSJdWdrkpaulR67TU70fbuu/VzkZQJE6TBg23/YPPNpbXXpg9WEQRYKL8Cq6Eh2yML6GhCs9KPPmp9f1zPiVoYQhiGCUSHEEp2AM0yiFKEg85hw+x2iy0sBAqVTHfcYeHN0Uev+hUKr7jClq1jj7X3iYY6pQjN5vOpdAVW377ZYcPhSoTRYYQhzColwIr27Igbivjhh/a3qMQB4rRp0kEHSbvs0nYddc89VuV16KHSnntKjz1WWiXpjBnpD31EbcutwJIYRlhpcVdEzlWpIYQff2zDx48/3rYP++zDMMJa1NSUPSnyxhvF19OvvJINQjtSL6kw3bvtZhcQ+/jjtKcoGRMmWPWVc/Zvv/3sxNPy5aW/x+9/L+20U92EfgRYKBxghZ3zaIAVfRzoaCZNsjOOuVUJhYYQpnlQFxciS3YATQUWSjFjhlUuhTP6m29ut6++mh0+eNhhFmy1tJQ+RCHXJ59Iv/iFdOCB0skn232vvVb+tBarwJo9uzIHWPPnZw/CJQvxpNbVUyEsK6UHVrEKrDAk4pln2n/w2dQkXXKJtMkm0r/+ZZVW11zT+jl33ilttJG02WYWYC1aJD37bOH3/c9/pLXWslsgnxkz7ABr2LD0Aqz5822IcmdVSgVWGEK4qvsm11xj65RTTrGfDzvMQoOnnlq190Vlvf++bfO22caWt2InccIwPClbudURhAbuRxxht/UwjHDBAqtWHzs2e9+++9ryPX586e/z5pu2T9dQH9FOffyWKKzcCqzo40BH8/rrdptbZRIXYHXrZrdpno0stAwSYKEUuaFQ9EqE995r8/7hh7cOttrjxz+2s6YhYJFsGEOpFi2y+b1YBZb3+XtMlWPevGwllRRfgZUvwIrrgfXee9I669iyGVeBFQ4kmpvbdwD+7rvS1ltL554r7b23Vct94QvShRdaEClZxdjjj1v1lXN2JruxsfAwQu9t2Kf32QMIIM706dkwPC7ETcJVV9l8X4l1QC0qpQKrb1/bL1m2rP2fs2KFVW3ss4+04YZ237772kWd7ryz/e9bCc8/XzeVJCUJ247QH6nYSaaJE7NV+x0pBJo40YbS7bqr/dyRpr29Jk2ybW80wPrCF2w4bznDCOfMkQYNqvz01SgCLBBgob6UE2CFM6BpNnJnCGHnNH++9KMfWdVPtUOD3ABr8GALZF591YaPjBghjRkjbbCBXQGwPQHWG29I114rfetbVv3Tr5+FOeVUYM2cabfFKrCkyvTByq3AGjDAlqto9VS4gl9ugNW9u/3LrcBaf/38VzN8+217Ta9e7etLdc01dhn1+++3YT5rr21NeufOlS66yJ5z3312YHvoofZz//52We5Cn/fQQ1YVJklTppQ/Xagf0XVJWhVYkyfbAV8lrphai0pt4i6tWh+sO+6wgP6001q/75e+JN11V3oB0muv2cH8vfem8/m1KARY++9vt6UEWJ//vLT66h2rAmviRGnbbe1kUkNDfIC1cKFdYbHU6sMHH5RuvbWy01lJoYH7dttl7+vZU/riF6W//73033P2bGngwMpPX40iwEL+AGvRIjur29BgO9xRHDyjowoBVm5jy0IVWGn2wWIIYeeydKn0q19Z0PGTn1jQUe2rzUyfbj1rorbYwq429cgjVn3lnM3/m2zSvgDroovsYiAXXJC9b9NNywuwcpvNxwlBUiX6YOVWYEkWPoUAa8oUq0waM8b+Xrlyl8H33pPWWy97NcNcb79tTdV32UX697/Ln94pU6xSIhzESNKWW1r/mt/8xt7/jjtsWrfaKvucPfe0A4M5c9q+p/f2N1t3XWmHHWwYApDP9OnpB1hhHq3EFVNrUdjfKNbEXVq1AOuqq+xkw557tr7/0ENtXRxC7aSFEL2z/n3b4+23bfu6zTb2ty8UYC1ebPu5225rJ6U6SoC1ZIlN95gxNu+vs058/64//9l6bD75ZGnv+4MfSEceKd18c2Wnt1ImTLC/U274tN9+tk9R6j7U7NlUYKETmTTJVmaFLF4cH2BJthHr16/tJc2pwEJHVSjAirsKoVSbARZDCDue5cstWPjud+0M88SJFki8+GL1PrOpyaqVckOhzTe3HdswfDDYYov2VTZMnCjtvrtVdwWbbmo72qGioJhSAqxqVmBJFuS8/76d5T3oIAux77rLhuHlGjAge/C+bJlN/3rrZSuwcs+cvv227ajusYcdpH3wQXnTO2WKNGpU2/t/+lObzpNPth5WYfhgsOeeNi2PPtr2tf/8pw3X+eEPbZ6gAguFzJiRDcPTCLBaWrLzKBVY7e+l99Zb0nPP2ZVKc3vm7LefVYqmdTXC0EuQdVFW2HY0NEgbb1w4wHr5ZVtOOlqA9fLLVj287bb288iR8RVYob/X3/9e/D1XrrR9/i5dpK9/vX0njqotNHDPdcABtpwfdVRp69g5c6jAQifx9tu2UJxxRuHn5avAkuxqUrkHzuFxDp7R0SxcmD1oZAghkvbCCzYE7A9/sAbcY8ZYT6NqBliffGLhRW4oFPpgrbeendUNNt/cmviW01+mudnOlIY+KsFmm9myU2ofixBgDR+e/zlJVWAdd5z9rW6/PdsbK1e0AitUXIUKrAULWu90trTY97DBBtmKh3J2ppub7fVxAdbw4dJ551k1XXNzdvhgsO22FjbkDiMM1Vfrr29ntEeNsp3gztpbCKtmxYrWYXg4uZlkgPXBB9m+T501wCqlAmtVhxDec4/dHnJI28f69ZP22iu9YYTRCliYEGBJ0ujRhQOsEPCEtgAffFDe1ezSElopjBljt/kCrPC8UgKsd9+19cUvf2nB38EH235YrZgxw46z4wKsYcNsOX3jDQuzil0dmgosdBq/+51tfP7yl/wbgpaWwhVYhQIsDp7R0USHx+RWYDU1dawhhP36WbjF5a47jqefttsDDsjet/XWtsM+d251PjNfVVNo2B6GD+beX87wjQ8+sGVko41a37/ppnZbagn8jBnW+yE3VIoaMMAO7Fa1Asv7/BVYixZZj6lf/cqaoOcTDbBC1UCowJJa98GaOdN2QDfYwIZprrFGeX2w3nvPvuO4AEuSzjrLemKtu272DHbQpYv103j44dZVYffdZ+HpBRfYdxremwNHxAmXtA8VWA0Nth1KMsAK2/DttrP1Smfc/pVSgbWqQwjvvtuCgnzh/GGH2f5/Gld7DOvSqVPz9/9J88rQSVu50r6TaIA1Y0b+IoKJE237ssYa9pqWlvgh7bVm4kQ7QRXWLyNHWigTPdZcvNjWAcOH222x6rJwEZmddrKThquvbhctiBuamIbnn7fbuABLsqr2G2+0dg9f/Wr+avbly21dQICFmtPSYmdXjznGenIUs3ix9Kc/2VCFHj2sWXCccBDfngBr4UKuEoKOJQwf7NKl41RghY137jIaDvLbO4QAyXv6aat2GTYse9/WW9vtSy9V5zPzBVhbbWVnJc86q/X97Qmwpk6129wKrNGjLRwr9UqEoUF07pD1KOdsGOGqVmAtWmTbr9ywbL317ParX5XOPLPwewwYEB9ghYPC6EFDOJO8wQb2O+yxh23TSz0AD6FSbkgY9OxpAdXf/x7//e25p/Thh7bT/8EHNjzovPPsb3bUUfacjTdu/VlA1PTpdhtdl0SH0SYhVJ4cfrhtwzvjVcrKqcBqz/b/o49s+GBc9VVw0EEWiH/726UPAa+UsC5dtCi7/Yq6+24b4njGGXbV1c7uo49sHzQaYEn5+xVOnJitYgqv6QjDCMN0h+1X6DsZXcbD8Mjvfc9+LnaVvrDvsckmFow9+KCtN6K9OtM0YYIdd0R7VuY64gjrcXnvvdYmIC68Df0tGUKImvHZZ7agrbee7fDefLN08cXFV9o335y9ytW3vy3ddlv8AVI4e5MvwFqyJH+AFX090BG8/rrtFG6wQcdp4r5woS2fuX0qwoE3lZAdg/cWYO24Y+v7Q4BVrWGE4QAgt4l7Q4N0zjnZnlLBsGG2E1ROgBWu1JcbYPXubduuUiuwZs4s3P8qGDp01QOsEDzlVmDtvbc1N77uusJBmmTLYDh4f+89O1k0bFg2wIpWYIUDiHBAscceVnVX6t89hIT5KrAkC6BCAJlrjz3sNlzh6fDDbfouuyy73hsxwtZ5BFiIE7cuSTrAevNNWz+FysjOOIyw2lchDFf3O/jg/M/p3Vu64gr7fq++uvzPaC/vLfgPQ9zj1kX//rc977e/tSqdn/wkuWORlSuTXz/mbjtCgBU3jHDhQltGon2kpNoPehcvtt8nWj0cpj1aLRWGDx56qIVSxYYRTp5sQVjv3vbz6NF2Vd6wz1Jps2eXdzLyuedsXu/Zs/DzTj3VQrvrr48fAhkCLCqwUDNOPtmu7rTJJnYZ0KefthVoGL8ex3vbAd96aztY+s53pNVWi6/cKhZg5f4/4OAZHdHrr9sBYL9+HasCq1CITC+6jmHaNBuCkxtgDRlioU1ckNHSYqHOqpg+3RqQR5urF+KchSDlBlh9+2b7U0WVcyXCUIFVzJAhqz6EMBx051Zg9expO4u5V96NkzuEcMQICwYHDbLX5wZYXbpYVYNkQwOk0vtgTZliB+7tPcO63nrSKafYQetVV9nQhXnzrGFz0NhoIWS1r0S4bJlNz0EHxVdYoDbVQgXWm29aULvJJrasdcYAq5yrELanAuvuu+37KxSGS7Z87r23nUSPLqctLVa9OWZM5XsrzZplJxf33tt+jguLXnxRGjfOtit77mkn6seMKf1E489/bhevaI9TT7X5r5zh36sqN8Baf3070RAXYL34oh0DhiBo8GCbV2q9Auull7KN54O48G3SJNvPWGMNuxrv+PGF94EnT7ZenFEjRmSr/CrtkktsuGIpxwzeW9XZdtuV9t5HHGG3cRd/CX0rqcBCTfjwQ+nOO22Ix7/+ZTPv9tvbDuZtt+V/3fjxttCeeqodjAwYYMntP/8pPfVU6+e2N8AK90UDrP/8p3BjQSBtr79uO249e5ZWgVUrVyEkRO74Qv+r3ABLsibqcQHW9ddbtcyq9GuYMcP6ReRW8BWy+ea2DSl1iPjUqbZdiqtY2mwzOwgptkPnfekB1tChFuytSh+UfBVY5ejf39YjTU22QxyGHzpnf7foEMK337bHwzpm6FBpyy1LPxDKdwXCclx9tfXTOPVUO1AIFaZRo0ZVv8LgpZfsu7n3Xlsf//nP9dXTpqOaMcO2idGz/GkFWD172nDacoL2jqKaFVizZ0tPPFG4+ipwzsLuFSvsRLhkgdVRR9lIkBdeyLZlqJQQLIwbZ1UzueuilSsttNx6a1tX3Xmn9Ne/2jbokUeKv39Li1VsXXpp+dN2xx12AZYuXWz4YlL7hW+/bdW9YdvYpYttb+OOt6IN3CX7G3aEKxHmNnCXbL930KC2AVYYZrj//rasPPRQ/HsuX27zRVyANWuWVX1V2pw59r6lVGG9847th+T2rMwntJ6Iqz6nAgs15be/tZ26U0/N3uecBVmPP55/CMXVV1ujuiOPzN536qk283//+613FMPGL5RXBtErnpVS/bFgga1MDjuMvlioTUuX2s7RJptYdURcBVbuGc9aGUKYewVCKRtgUYHVMTz9tB105O5MSbYz/uabbUPV226zee+GG9r/uaWGQlGbb247YaWepXzrrbbDB4NNN7Vlq1jJ/sKF9pmlTOt221k1yDe/2f4mzvkqsMoRwq/58y1kDAGWZDvJuRVY4Qx6sMcedlKplB3pSgRYpRg1ynasq7nOmzDBbh97zIZPHH+8VVyEnXDUpunT2/aoSzLAmjvXKi9Dr7bNN4+vwLr5ZruIUUcVAqxCFVg9eljFZLkB1v332z56KQGWZOusc8+1ESB33y196Uv2/299yx6vdO/GEPqvv74FlLkB1pQptu8Wht5Ldky0+urSTTcVf/8PP7Rg49lny6usf+896cQTrYjg9ttte33VVaW/flW8/bZ9H9GTUPmuRDhxolX5RquhkwiwbrzRKo/yzY8XX2x9nPKZONFOtOVu/9dfPxtgLVligWkIuT73Oas4yjeMcOpUW5Zyh9XHXWSlUsLvn1ssEicMBYyGdoUMHmzr3nAxjSgqsFAzliyRrr3WSnhzrxJy+OG2Abrrrrav+/BDG1544omtx9T27m1DCMePt39Bvgqs6AFz3A5+bgXWLbfYNL/2mp2lAGrNlCkW3oYAq5SrEDKEMOv55630Hu3z9NO2w9XY2Paxrbe2dXq0mmDuXDtTLlmA1d4TA+0JsEL/kVKqG1assIOOfM3FS70SYRiiMnx48c/8v/+z7dl119n2cNmy4q/JVakKLMl2hOfNax1gRSuwvI8PsPbc09Y7xSoHFiywndYkAqyNN7ad/mpepen5522e3HVXOxl39dV2+41vUIlVy2bMaNtLL8kAK4QZIcDaYgubT6PD6FautGqh88/vuCdTQ3hcqALLOdtvL3cI4d132wF8oabRub73PRvOdcgh0pNPWlhx9dV2jPHyy+V9fjHhpMmIEfEBVqhUjgZY3brZduCee4p/H+FEyrJl2WqlYlaskL7yFfvOb71VOvBAC/IuvDA+TKi0uG3H6NE27+du+yZObFvRs8EG9r1Wsxn/n/9soc1557V97B//sPvPOit/dW+08XzUyJHZbVFo4B6e19hoVxR84IH43y00cI+rwJKqc2XGcgKsSZNs3g37SMV06WIVVnHFKwRYqBk33mgN3M84o+1jm21mK6/bb2/72DXX2AJ+8sltH/va12wFXEqA1aVLtgdIKUMIr7vOdiY22UT68Y8756WN0bGFUvdyhhDWQgVWrQwhvOIK6Qc/qF6z8c5s4UKrFIgbPijFN3J/4AFbj555pgUk7b2ceXsCrLBDVUqA9e67ts3JV4G18cZ25rjYlQjzXS0xjnPST38qXXmlHZDts0/5y0ElKrDCa8NBXLhqkmQ7yXPn2t9+1iy7zT0I2XlnaZ11rIdLoW1m2OlPqgIr+pnVMGFCtu9HQ4P15vrxj+1vGbdfU65PP03u6mTjx0uXX57MZ6UtVGBFDRhgy14S+3yhN1s0wJJah+OPPWYHeJ991nH7Y5VSgSXZfns5FVgLFljPvYMPLn6BiqiePe2E+qhRtl06+mgLD7bYojoB1qBB9ruNGmUhQ7TP1osv2hUIwzwQHH20VWaFBvX5RCuBo8dChfzgB7bO+uMfbb3unO0PLVtmo1qqqaXFKpDiAqyWlta/z/z59nNuELTBBjZPxfVOqoQlSyywGTDAgs3o9zpnjnTSSdn97rjv6+mnbdneYYe2j40cadPd1BQ/zHD//W07+8wzbV87ebLt0+eeXAsnmqodYBU7GfPCC1YdFjeUP5+hQ+ND0zlzbJnp3r309+rgCLBqkfdWarn11laSmcs5O9swfnzrxorvvWfDDvffP5swR/XpYyuyaMlvvgBLyh40FwuwXnjB/n3jG9bs8Y036rsK62c/s55lqC2vv55tUpxvCCEVWPmFM0rXXZfM53UmEybYzma+AGvdde1CG9EA6957rRrpoossKPnLX8r/3GXLbOcut2qimD59LIwpJcDKdwXCoEcP2+6UWoFVTth2+uk2bOS//5WOPbb010mVrcAKf7fcCizJwscwBCL3IKR7d+vF8vLLhZerECblq3KrpBBgVauR+7x5NrRj7NjW9599tt13yimrfoXJAw4of35oryuvtKt51sMVmfNVYEnJnEh54w072AvLWRgaFA2qbrklO/qgvaF/2kqpwJJspEQ5890DD9i+TKnDB6O+8AVbJ4SLT0jWw+/llytbNTltWvb4ZdSobPVq8OKL9nfPDfd23NFeV2wY4Vtv2TZp441tu1HIypXZflknn2wVaMFGG9kV3v/8Z7uSXLXMnGn7qnEBltR6GGEYkhZXgSVVbxjh+PE2X/3pT7bfcPzx2RPEp51m1UE33yx997t2kiL0A5Vs/j32WNteRtvlBCNH2t/h/fctwBoypPU6aK+9bF6IG0Y4ebLNQ7kB0dChNg9Uo5F7WB4//rhwQOa9/T7bbFPe++e7AvPs2XXV/0oiwKpNjzxiB9tnnpn/LMnhh9sCEIYRLl5sZa2SXRY7n7DBCSoRYF13ne0wHHWU9cDadNP6rcKaMUP64Q/jr/iIdL3+uh1kd+tW3Qqsgw+25bNSO/T5emD17m2BXBIB1vTptgPRq5ftiFSj+WVn9vTTti7ffvv4x52zIR0hCFm2THrwQTsQ79XLhi/ceWf581R7QqGg1CsRFguwJNsmVLICK+qoo+zse6nDQYJ582z57tGjvNdFhYP3cFIotweWZMtNOHAIV1WKOvRQG0r3gx9Y2BhnyhSrVIp7faUNGGA7ydWqwAp/p9wAq0sXOxhcuNCGiOY7KPbennf22fHP+egj629TzSGQUa+9Zvs61TyIrQULF9q/uAosqTLDCL237fSVV0onnND2SqNvvmnBQRiGve66tm0MAdayZbZPfMQRdiDdUQOsciqwyhlCePfd1gs3rtKlPbbc0irdPvywMu8ntb4YRm41qPe2jYwOHwycs+3AI48UvnLvW29ZoLPLLjYcMt9xykcfWWj3ox/Z+8YdV51/vn2fp59evaHPuVcgDEaNst85GmD95z92G1eBJbVuhl5JDz9sJ2P23ttCrHfese3ZHXdIf/ubfYdbbWVDCIcNsyArfF/nnGPr6htuiD/WjF6JMNrAPejXz/6W+QKsuJ6jcRdZqZRFi7LhYjSoyzVtmi07pfa/CoYNy1+BRYCF1F1xhe1AhktmxtlkE1swb7/dVgQnnGAHG3/7W+Gd3K22shVB2Oi1N8AKB9QzZ9oB7WGH2Y5MQ4OtrN58s/CVEjure+6x2xdeqP2rftSbcAVCqXoVWNOm2Txwxx22k7iq84D3+SuwnLP7kzjzHcqzf/xj+7x6rrBsj6efthCnULXP1lvbOry5WXr0UQsJDzjAHjvuOJtfy/3eVzXAmjq17XKS6623rIFuod4Lm21my0KhXlUzZ9p2KC6sLWaNNeysZDk9b+bPX7XqKylbgfXSS/Ze0fcLFVjTptnv3tAQXxntnB2wz5tnFcxxpkyxg7qkhgdsvHH1AqzQwD3uykubbGIVD/mGEs6ebScIjj9e+vWvs0NKou6/324LHcRWyvLl2XV8KT1Pat199+VfRsO6JF8F1qoEWE1NFgKsvbatJ8880w6EcxuxhysQBg0NrYP2Bx6w7dNXv2qh8PjxHbMPVqkVWOUOIXziCatYKeeKtIWEPlqVGkbY0mKBfwiwQsVpWBd98IEd9McFWJIFTS0t1qcqn3DBkZ13tnklbpjpffdZODdpkl3h8MYb40909O1rIdaECcUrjNsrX4DVs6dtT954w/ZPv/1tG/3xxS+2DTKGD7fnV+uY5N//ttFCPXtamHTKKbZNO+kkGyp+7rn2vN69bf3+1FP2Hf/rX3ZVx+98x/4eccKw/MmTWzdwjzroIFs3RLcHixdbMBYXYEn23VUrwPrc52zeKLRNCNVy7a3Ayg1MZ8+uq/5XEgFW7Zk61TbC3/pW8Z3VI46wMwhnnWVh0c9/bgl4IVtuabdhpb14sW3M4lbOhQKsxkbbeP71rxaGnXRS9rFDDrGdip/8JP/ZjZUrbdrT2LlYscKGaFbjwP/uu7MHi/UY4FWL96s2r4QDjRBg9expO4nRxo9xVyEMP5dagRUOnq691s6SjB1rYUR7LV1qy0rcMijZ/ZWqwGppyb9Bf/ppW0ecfrqd+VvVYYQrVljwneTl16vl2WctePjqV+3vPWiQ/T/MMy0tFgDmGz4YbL21HTy++aYNH+zbV9ptN3ts++3twK3cYYSrEmBttZVNe7GDk6lTC1dfSXZQ2tJSOBRpT6+uYNgwW37zVTDFmTdv1fpfSdnXL1rUuvpKsh3N7t2zFVjrrJN/m77FFjZE5fe/j696mzo1mf5XwahR1RtC+PzzdmCaLzz8zndsOfq//7N+KX/7m30n//yn7Vc88ID1P+vWzdYhuUIPnHBVy2qaMiW7j9PRA6zXX7cq/muuiX98+nS7rUYF1nXX2RXdttvOtp3vvWdX6Pzzn7Pf7/LldkCa2/soXInQe5tXhg619eauu1rYUUoVaa0ptQKrb9/SK7A++8wq2kptGF2KMISzUgHWzJm2bxCC/r59bX4L2424Bu5Ro0bZPJRvGGG4OMWGG0qf/7zdlzuM8MUXLRAZMcL+f8wxhaf5oIPsNuz7Vdrbb9t8sPbabR8bPdrCs512ssKH006z9WQu56ywoRoB1syZFi7tsUf2vosvthM4y5ZZZVU0iP361226v/tdK7zYbDNbn+czfLjte95zj60L4gKso46yk9J/+EP2vtDzNukAa/Fi2y//3OcKV2BNmmTfS+4VEosZNsxGj+QG11RgIXX33GMrq3CJ2kIOP9xur7jCKqC+973ir8k9Y7JokQVRcUMVCwVY4f5PPrGV0bhx2ftDFdaUKdb0MM7559sG5Npri09zpV16qTXHr3Tj1dmz7QzXccfZwSoBVvstWmRnvc46y8q4Bw60MzHt7TPy1lu28YtWYEmtq0sqMYTw3nttB/Gkk7JX2tprr/bv3ISd03xVKf37ZwMs721nPXfYRSk+/tiuqrPeevEHYk8/bTuG3brZFU6ffrr9ZxynTbNl/+ijbfhUNa+MUwm/+Y2tN6+4ovXBwvTp9jvssIP1qXr2WTuY++IX7UDquOMstHnjDfsblRJgSbZjc//91pg8BB7O2fs9+WTrpq2vv57tPxFnVQKsML3FDszDGe1CwkFToWGEqxpgSeVV3VSiAiu6bcwNsBoaLLQKAVbuGfRcP/mJTU/ucJSWlnQCrDlzslc2qqQJE9oOH4zq0sVOjI0YIf3qVxYGb7GFtN9+toP+/PO2/7DvvlZpET1JNn++DRsL81EpvbS8t/2Us8+2g9W99rLA+LDD7HNuvDH/fBvWgTvtZCF1R26bEIYXxR0AS9kG0JUOsBYtsnn/85+3E4AnnWR/+xNOsM8MJ4Deece+39wAa4st7LNff92GER1xhM1Du+xij3fEYYTVqMDKvYJjJfTta8FIpQKs0JMoui4dNap1gNXQkG3eH+foo626JQQYUR9+aN/thhtaIDRiRNtG7pdeat/ro48W365JtjyMHVvdAGu99eLnhdGj7TubOtWGzv7mN/lPkmywQXUCrHAF3T33zN7Xp49dTOG//80Opwu6dJEuucT2G2bPzl/dFjQ02L5/2A+JC7D695eOPNL634X94XxXIAzWW88+v5K9C73PHlOPG2f74/kKJV54wfaLym1hMHSo3eZu26jAqh3Oub2dc1Occ287585Ne3oS893v2oo37JAXstFGtpBsuaWVW5dyVZE117ThHqFnR1jY4pQSYEl2QJv72QcdZGfATj3Vykuj7rzTEvquXW0HNe4AdunS1s3mi1m2rLQD4WnTrMzWOTvTWEow8fjjFnjl9kzKdf/9toN1yCHWs+bVV+M3oh3VTTfZ37PaVTMvvWQbqWOPtTMqS5bYwcr771tZcnuEv0PYmIYmr9G/aVPTqg0hnDPHdoZCL7r117eDmrXXzn9Wu5iw8cu3DPbvbwHJiSdKa61lO3bjxpVXlfXgg7YOGT/edn5uuaX140uX2sY2BBrHHmvfS75wupD77rOgZsoUq7B49NFseXmtaWmxAPWMM+xv++1v29/yu9+1atdRo2xddv759n2/+671grjtNukXv7Dv8bTTsjtexQKsUaNsZ+b3v7cQMgwfDI45xnbmfvMbq1bYdlvbAdplF2nwYDuhccMNdqY9mDHD/qarrVb+7z98eOsdxzhLl9pBQbHm4httZPNMKJuPU4kAq5xLmleiAqtrVxsWIbW+AmEQ+myUEmCtvrptmx5/3C47Hnz0kX3PSQZY4SC30sMIp0+3v3O4AmE+o0ZZkLt4sVXX3HyzrUOffz578PrVr9rf+7HHsq/7179sPX7CCfZzKYHm5Mm2/rz6ajvg+uwzW9++/LLtpxx7rJ0ljwbHwWuvWTX6179u4XaxPm+1LFQjPPFE/EHdgw/aeiZ3Ps4XYN1+u60niw1BvvxyOxi75JLW+5EHHmjLRNjO5F6BMAjzw09+YlVaRx5pP6+zjh2kPvFE4c+PE6rN0lJOBVa5AVal1yNbblnevnohYR6MC7BC/6tRo7InIOMccYQtk3FVWLn9Gnfe2fZ7wgmDDz6w7fdJJ5V3cuPLX7YeeOVsf0pVaNtx9NFWffTii8Ub82+wgYXAlR718vDDtl4Io3uCESPyr+f328/2ja65JltUUUhoizN4sO3nxjn5ZNuf/+tf7efJk20/P/fEUnT6pMpWYS1bZt9vnz62v9fSEt8bMTRwL7f/lRS/r9PUZMcKdVaBJe99zf2T1CjpHUnrS+om6WVJm+R7/pgxY3zdWrTI+6am8l6z227ejx1r///KV7zfaKP45516qveS9/Pnxz8+dqz33bp5P2tW/OOffeb95pt736eP9xMn2n2TJ3vfu7f3n/uc97feau//t7+1fe1hh9ljzz1X/PeZPNn7tde23+OZZwo/98ADve/Vy/vf/c7e/7bbCj+/pcX7rbe25+68s/cLFuR/7r77ej9ihL1mxgzvnfP+gguKT3+ta2nx/sc/tu9Ast/x2Wer8zm//a333bt7v8Ya3j/0kPfNzdnHDzjA+379vJ89u/z3/tGP7O+xZIn9/Oc/2+/y3nvZ5wwb5v1JJ7V+3aef2vOuvrr4Z9xwgz13woTW9592mvc9e3q/dGn50z1xor3nfffFP37oofZ4//62zPz85943Nnp/8MH2fRbS3Oz9d75jr99sM1uODjnE+6FDW3/v48fbc+6/v/XnDhzo/bJlpf0eTU3en3WWvc+YMd6/847df8opdt8tt7R+/pQp3v/pT+Wv2ypl6dLsOuj00+37ePZZ7w8/3PuGBrv/wAOzv0eulhbvzznHnjdokP0r9vfw3tapkvddu3o/b17bx7/0pexyuPXW3l9xhff33GPz7fDhdv9qq3n/hz/YNH/1q96vv377v4djj/V+8OD80/7KK/nX4bn23tv7ddeNf6+WFu979LD5sT2mTrXp+OtfS3/N6NE2H6+qNdawz/7tb9s+duKJtr2TvL/00uLv1dRk69ftt89+Tw8/bK9/7LFVn9ZSvf22feYf/1jZ973nHnvfYtvpUixZYtuDr389e99XvmLz6wsv2OfceWfx97n2WnvuW2+1fWz5cu8ffdQe/9Of2j5+4IHeb7yxbUfyzQO5fvc7W25fe620dUJSwrZA8v7ee1s/tmSJzcff/Gbb1332mb3mssta37/JJnb/+uvbtjzOp59637evfY9xTj/d9jFnz/b+oovs/RYujP98yfv11mv9nX79696vvrr3K1cW+s2zXn3V1lOS9zfeWNprquGqq2waPv208PPOOsv+LqU491zvu3TxfsWKVZ++qJ/8xPatcv8u7X0vqfW+0uWXZ7+LtdaybVoxe+/t/ciRbe+/+mp7r+nT7efrrrOf33zTfj7rLNt/ev/98qY7bAevvbb4cxcutPVKKVpa7Pjp9NPLm544f/iDTeMHH5T2/FmzvN9jD+//8Y/C0zdsmK13q+mMM2za99678PO228626y0t3u+5p+1r5vPss/aef/975aYzHC9cdZUdNzc02LFHrg8+KP24ItdLL7Xdts2cWfr2p4ORNNHnyX5qtQJrrKS3vffveu9XSLpV0gFFXlOfevcuXmaca6utrDpo5crCFViDB9vZ+3yPH3SQdN55+VPfAQPsrN3AgTY8adIke02fPlbuethhdjbt4otbD5n417+sWbFzVvVQ6Ooe48db+X5zs6Xf48ZZ34zly9s+94EHbIjXBRdI3/iGJfC//W3+95ZsuMOLL9rZwKeesnHe0QqHYMECqzQ7+GCb7uHDrTLittuqd3WSSnvqKasYOOusbCl3c7MNZ/3Rj+xMdDhbtdNOVj1XztmcQhVsy5fb/HDKKTZk8KWXrCQ5XG1IsmFaCxfa/FKu11+3ColQeRVXgbWqTdzvvdcqHHPPquy9t52FLnbJ5jjFhhBeeaUNK5s1y854n3eencm+++7CQ2S9tzNWv/613U6YYNU8hx9uZ8OffDL73DCOP3rlopNOsqqku++2KoRrrrEhbhdcED82f++97So+p55q81moVrnsMpuXTjjB/uavvmpn0EePtkbNX/lK6Q30o+bOte+kPRYssHnvjjtsOMEVV9h8uP32tjy/+65VEt1zT3zVjWTrgEsuscqO2bPtbFwpFbKhoeduu8VXB118sX3HL79s03DGGbZuuvZaqxp47jk7E/qtb1kPhkmT2l/VJNn6dNas/EMPSrkCYXDkkVZFGS4IEDVvnq2/O1oFlpR9j7gzveuum+3DVMoVBLt0sVYAzz2XHfpUrcqJQkaMsOHCla7AmjDBfsdSzroX07OnbW/vusvmnRUrbBv/5S9nG42XMj88+6zto8T9fbp1s0ry1VaLn29fe83Wm+uua/NuseG2L75oladnnmmvW3ttq96KaySdtGnTbH3Wr1/bYYQPPWTz8SGHtH1d2DZFK7CammxY0z772Lpzr72sUiS3Iu7nP7f3/fnP46fphBPs73rTTVZpvPbabfdHBwywaivJqvKi69lddrFtQbHKuI8/tn3CLbe0+WGttWx/I60G8KVWYK2+un1/0SvR5TNlilXhFHvPcm25ZbZ9wap6771sz6MgrPeeesqqUfP1v4raYw+rNgpD6IO33rJjpuHD7efQOHz8eKuivu462wcK81OpNtvM1pnFhhEuW2brvtNPL+19Z82y/ali1bulCO9R6jDC226zY5qDDsr2Fcw1ebItO9H+V9UQ1s3FKpZOPtmWhf/+N/8VCINqVGBFL4rWr59V7sb1wQrN5itVgRWG+lOBlf4/SYdKuj7y8zGSrs73/LquwGqPv/zF0to33vB+112tsijO3LmlVUAVM2WKVSA4Z2eA/vvf7GOhEuaBB+znxYvtDPTGG2erpG69Nf5977jDzs5tvLH306ZZ4n3CCfaazTf3/l//ylZwLF1qZ2RGjcqe/fjlL+25r7ySf9q/9jU7A7JggZ097tbN+y239P6TT1o/75Zb7L2efDJ73+9/b/e99FLp31WaDj7YKoW6dLEzBwcfbFVlkp29C2c25861xyTvt93WvscpUwq/9x13WIVFvgqJ+++397vwwsJnS4891t7nww/L+9023dT7/fdv+3mhMtB77wcMsGqpqEWL7HmXXFL4/Zcsscq+//u/to8tWmTzzVlnlTfN3lvlVe50FtPS4v1BB9mZxOiyFvX979v7fv/7bae1Z8/Wv8eXv9y2SnPlSltOw5lvySofJDtLeuedNh2TJ9sZ+G7dbFmPM3Om92uuadUUki1v55zj/c9+Zj/vv3/p1WstLfY5oeplm23sdxw/vvSz8N/8ps3/pVQVFdPc7P3555dePXPNNat+Jq2lxdZHoSLrsMPa/16vvZa/+sR77y++2Bes0o2aP9+W3VNPbfvYiy/GV+KVo3fv8paxXr3aX/EVtcMO2e1prhtvzC4fr75a2vstXWpntXff3X4+9VRbJpKu1tl0U1v2K2n33W2ZrJRQnXbnnVblEypFm5tt/Xf++cXfY5NNbDtXyN572z5F1NKltp4IVdaHHWYVhoUcfrit515+2ao1DjvMqme7dPH+vPOyFcLttXy5rcPbY8wY7/fay6pw11yz9fx29NFWyZSveqdfv9ZVIq+/nq1iWrrUvqOuXbPL//vv2z5bt27eH3984enadlv77seMsYqQOPvtZ5/32mut7w+Vcb/5Tf73f/ddq1rt2tX7M8+0aq+bb7bX3XNP4WmrlrBvWqyqaeZM+7uMHVu8WnmTTayKvdKmTbNp/f3vy3vdihVt12m77eb9jju2vu+dd+z9jzjCbh95pPh7P/ecjx1hsc8+tv8etLRYxfnRR3v/q1+Vv68VdfrpNn8XWv5+/Wv7jE03Le09n3/eF6zAL0f4O5VSJea997vs4v2GG9pomS5dvL/99rbPCb9PqVVd7fWvf9nn3H134ectXmz78XvtZc//1a/yP7elxfZ1K7EPEIRKvDvusJ//7/+swjQ6osF773/4Q9t2LF5c/mc0N9trf/jD7H2PP176stHBqEAFVuphVexESYfFBFhX5TznG5ImSpq4zjrrVP5b68xCCeKtt9rOwT77VP8zJ0ywoRZ/+EPr+5cvt+F/IUQ791ybtscftwV1q628X2ed1jt2LS1WVuycbexyh5T94x/Zg7ehQ6389OST7ed//zv7vNmzbYMTVxbvvfdz5tjj3/pW9r4HH7T7Ro1qPXzo0EPtoCN6kPzpp7YTfd55pX1HL7xggdmmm9rOQJLDpz7+2DZSZ59t4dC559pOnXNWDpurpcXKr7faKnuAtvHG3v/iF213ui67zN5Hyn9AEUq5i5Vuv/ee7WjmDvWLM3Wq91deacOunPP+e9/LPvbII/Z548dn7+vTp+0B8IoV9ryf/rTwZ4VA7OGH4x/ffXfbgSympcXmu7BjFw6Ap04t/tqoefO832ADWw6mTWv9WCjJP+mk+IPi6DDClhYLn487ru3z7r/flo2//MWGHLW0eP/UU7aDKFk43qePLRdPP114ep97zg5SLrig9fIcQuw99yy+sZ8/34YXSLbz9dOfer/TTrYMSvGl3Lmeesqe256wsRJmzLCD3PYMk801f76FgE891f73WLnSDpDyHWQef7zNK6U69FDvhwxpu247+WRbrssNpqNGjixteIn32eX6Jz9p/+cFYchRXMgaht9K5e2shoOp556zg/Y0TtIdfLBt5ypl5UoLa6Lb01XV3Gzz38EH28FCr17ZfYXhw+2EViFh+Fmx9fuFF9o2JBrUhtA1HCRfcYX9nG8enjrVDjyi2yHvbX3/9a/bazfc0PZ92uu442x/6qOPyn/twIG2L/SnP7U+8bZsmQVUhYKmddaxk0vBnXfae0yalL3vrbfs79Gliy3rG25o7QKKHfyGE4ENDW1PMAV33JF/P27ECJs/4qxcafue/fq1Dr+ammw44tixqx4cX3ll/pM3+fz85/nXKbn+9jd77i9+kf85TU32nX/3u+VNRylaWiw0yPf9x3ngAdu/zN03HjGi7Tq8udnmk1697PecM6f4+69YYc/PPVmy4YZth40feqiddFtzTQvQ2ivsU+YOvw0++8y2pc7ZPkkpf9uwP/Lgg+2frqC52QLj3PVPnJkzsy1Q5s+3/aiGBgt2o/bay/b7q62pyfvrry/tmCgMN5Qs+Cpk9Gjb362Up59u/bk33WQ/v/hi6+fts4+17WivoUNbHwOF9W1HKZYoQ0cMsHaQ9FDk5/MknZfv+VRglWn5ctuYnXuurXwOPzyZz823IxB2/K691nZuogfLjz1mj110kf28dKk9LlnfhHxnLJcts7T+kENspS3F/57HHWdn7uP6zVx2WfxKYfx42/gOHmzjqEP1zcknt32PPfe0CpR8v/vy5bbztdNO9lm9e2dDoc02y987otJCJUW0imDx4tKCk2nT7AznLrvYewwaZBVLCxbYGU3JdiAHDIj/jqKfX8pB3mmn2Q5AXNXXrFnWa2bjjbMbsY02srNj0Z36uB2DHj3a7uC1tNjzivUyO/54OzjL19vg0kuLB3RTp2Z7HPXv7/3nP5+dL2bOLPz5cV56yX4nyUKls8/OVjUdfHDbs0LBbbdlQ+QpU+z/111X+uc2Ndky3bevHXivSijhvfXhcc7+prvtZt/L5z5nBx5f+YqdQfvlLy3AaGiwUCL6u82bZztaQ4YU7v+xYoUtc2uvXZl+Hp3FfvvlDzLCPFqqu+6y+Sm6XvvwQ1tHl3MAFGfcuNIPQGbN8kUrM0p11FF2ABQn9LpYc83y3nPBAtvGHHighQOlBnOV9P3vV7Znzptv+oLVfO11xhk2/wwb1rqX0tZbF6+sChVc0RNbcUJ1V/QMdzg4mTzZfg7VEvkqxk86yQ7E863LH3nE9hUk2161pxorVMVuvXV567CFC7MhSOinEva5/vEP+zlUycfZYovW1T2hL1Lc9vz99y1Y6NHD+x/8oPi0zZtnlRLtrUz92tcsnIurwA0VJHEBUzh5siq955qaLBzbZZfyXhd6jubbRke1tFgI061b/irPt96qzrIX7LKLbZOLWbnSwuIwGqNfv2xf2aYm26+Lmyc23dSmv5xihS9+0falg6ambKVj1G9+k91X/Oc/S3//XCtW2H5bvqD3vPP8/6reS630CidAiq2fSjV6tJ1YLOa3v/WtqoYXLrTXSfa93nyzBXI9e+YPldPyxhvZv2exfc8vfak6FcFh5ENcb8RQ9RcN/Mu1xRatq6ND5X57TlzUuI4YYHWR9K6k9ZRt4r5pvucTYLXDllvawrvWWq2boKZh0SI7MyHZjkZuU/gDD7Rg5/nns02Of/Sj0ocEzZ1r5a+ffdb2sdAk+8orW9/f0mLBxw47xL/nm2/azmao4MpXvvnHP9pjN99sperLl9t0P/GE99/4hh2kSHa277LLbBpbWixRDzuzW29tlXKjR9sGfN99Sxuy470FecXOILa0WLXO5z9f2nsW8uyz2fLd7t3t9swzbUdso43yh6XnnGPfZSk+/tjmh003tYOC88+3v9+RR2bDyh13tAaJ+Zpsh7Pn0SECjY1th9R5b2FvoSq65mYL7QodZE6enA1pcy1caGFyt24W+px3ngV9O+5ov2c5zdJzvfmmnc3dbbfsd7PbboXP/oVhhKeckh3imzs0oxQLF1auivD22y2g2Gknm/499rD5deTIbEi39tr5h0yGCrlCQ0JCiFqJcv3OJHwvcQ2Fhw0rPgQoaulSO2iJnqQ47TQ7sIheUKE9DjnE1pGlCE3Kb7hh1T7Tezs4zFfl1txsv1u5B7De2zYu7Ij/+MerMoXt89e/2mcPGWLVozvvbEPevv99+96efba8kCRUk5Y6lLJUYbiQZNWgwT77FK9cC0FL3AmsqFCp9bOfZe877zz724aTFqHqI+6Abvp0W/8Wqz5bvDhbQbDppjbUsFShgfCXvmRB/n77lRaAeJ/dPoUhvGPG2PrWe1tWC52c8d7mjeg8fsQRtk9TSCn7JsExx9j0Pfpoac+PCtuw3FYRr71m+yhf/nL8dCxdageae+7Z+v7PPiu9tUY4Ubb11uVN8w9/aK8r9fv59FM7oTpmTHzgHELIVanGLeT0021fpdA++fz5ti8vWegfLo7wu9/Z4+FgP+5kWWhZUc4QyHDhnrBs5wvxwoiUTTYp/ZginyOPtL9D7nL30Ue2T/XVr2ano5QLZIShYe2Z7+OEE6mFwmjvLazaeOPW89/ixba+DCF52O+qZBP0StltNzumLLb8nHyyPa9SwkVKQsVVS4uNPIo2uZ8+Pf6Ysxx77pm9EJv32QtctOdCUTWuwwVYNs3aR9JU2dUIf1DouQRY7XDssVZiH9f3Jw0XXhi/cfHeVvhdu9rjffpUvi/B9ttbuBLdeIWNa6GrWn36qZ11kvL3h5g7N1v6LNkGtU8f/79qq6OOsrM+cTuay5bZUJJdd7Wd0kMPted36WIH8sV6XUyebOFHGON/440W/uQKVW7lXMGrmKeftj5M0eGHO+5oZ2/iHH98eVUKN9xg1TJDhmSvCte/v83LpRwghWqAUBJdqNKqWK+ccJas0BUtW1osLM4tV37nHbtfsjPFuWfnV65sf3iVa9Ei2yEqpcotDCM8/nhbR6zqjl01hWGXhSpFmppsfbfffvGPv/OO7WAedFB1prEje/JJmz9zh0YsWOD/V7VRjuOOsxBr6VIbMtm9e/GhXqU45RQ7IVCKYlf3rKSxY0sbtpFr9uxsL7d8VT3VNH++HQSedJIdQO68sw3BCUNyw3av1IOr006z36fUUKVU4QRMQ0Prk1/HH28HD4Xss0/p/Whye2V9+ctth4Xvtlv8Gf2zz7bpe/vt0j7rwQctHO7WzU5APPqoncB78838J68eeMD/r3I2VFCUevWyEHCEod6hR8vHH9sydcwxhV//5S+37i20+eb517Xt8cordtBW6CrQ+cT1wVqxwsKeQYPi94mCX/zCXjtpUra34NChvlWVRSEhhI67Il4h555r+7zlCMOI4obDhuCiEkPT44QTtfkq9l980faxGxutOrulxf6NGWPLUEuL9//5T/4TwaF66cILS5+mMKQvDOcKy0fu36252fpsViKICcM5c4PCk06yv+c779i+VJ8+pR13heOQVRlWHLV8ua3DN944//7Sxx+37bEUtXKlTddRR9mxUy1Wq7/3Xut+xPlccol9v9F16rJltr7LHfZXiriWH8cfb/d9+9v23uFkainTl88xx7Tut1jO1Ug7mA4ZYJXzjwCrHcLwOKn0Hk3VtGKFraTzJeY/+5ntIIVy/UoKDdi33NJWLqEke+DA4on2kiV2VvXyy/M/58MPbQjC9dfbBvjkky04aW+z1dtvtw3MF76Qf5jBhx9aMDJ0qJ0VGjQoG6BdfHHr5371qxb+rGoD2WL237/1Tm7Ul79sZbHt0dxsTfXLOfsQhvZcf7393NRkP8f1xBkwoPCBwHe+YwcaxariTjzRDtzDjsPKlXbWul+/VduYVUMYRtijh4WnncG559pyEy6hHbS0WNVgnz6rPtyxM1q61Obvs89uff+kSTaPRC/nXIowHOuuu2ynrrGx9AP7Qn76U3vfUgLfSh8YFNLU1P4A+Dvfsemspd4WK1ZYkHLvvXbw2aWLDWEoZOVKO1htTyVaKW67rW2Pu+9/3+atfN99S4sFcKWGp8cfb/sEYR9l5Mi2F0g4/3z7zOhB3dy5tm458sjSPif49FPbLob9tPCvTx97z1yhT1cIec46y55/4ol20q9QUBMCrxkz7OdwifnQrqFY0HvssdkDqqYmW19Uo99Se40YYd/N6NF24H344fZ7hWbL+cybZ9vnPfawf5JVww8fbifkilV4hBOcgweXN71nn20nVMp10EGt9zGCk06yfcBqCduC3O+zpcXmre7dLUzOXd+G6rj//CcbgsVVzYfnlXPCYeHC1lX1V15p71FoOVhVn31m68PoCYs33rD9jug+5I475r94VlTukLRKCAHKFVfEPx6Gzha6uFVnEfZzo79ruHBStGqqVKFfX1iPem/7T6ee6v9XiXnssbYuWpXg75xzbJkK659jjy1veG0HQoCFtsIOvJTtdVCvWlosUBo50r6P7bfPNjSvVX/9q60Ev/Sltgdsc+dadVLfvtYY3nvbiZ840YI5KXtVvTlzbEV4yinVn+bjjsvfK6ac/jWVkNsDZ+lS+/nnP2/73MGDCw/92HZbCxOLCWdIw85IOGgIIVotCcMI853R7YimTo3/G197beEdOtjymTuc+tZb7XsrZ5iT93aAO2SIVZb27GmVh5Vw/fU2PcUuBOF9thdXLQVDcRYssHV90lcgLNW8edkm9mGoeK5PP80+59JLk5u20Nsmbuir99n1Qan9/cJ6YupUq2J1rm1FyIMP+lZVJC0t1tOnvfNaS4tVFD/+uB14hoqgu+5q+9x9921dEdbcbKFbGDouWVuCuNeec449L4R9YVh8CMyKnRw6/XQ7CeZ9tm9idDhn2l580f5W++9vQYpkVemlCBcW6tvXKsqbm7PzQr6G3d7bflhDg1XedO9e3vSecYYFUeUK+xi5J8R23jk7JLQali5t279qzhyr5A7DWuOWwyVLLBQ++OBs1V9cZdDs2RYClHvSd7vtsq0xTj3V/obVXpd+8Yu2bujRw/517WqfG/39Tz7Zlpdi0xKuvlfsIjjlaGmxasYBA9q2a/He9sNHjardbU4lheHn99+fve/II7Mnb4sNLc8VLr4SVyl6//02r0ultzrIJ1RUhunbd9/K9vKqIYUCrAahPm25Zfb/ffqkNx21wDnpq1+V3nhDuvZaafp0u/+b30x3ugo55hjpmmukf/1LGjtW+ulPpRdekJYulQ44QJoyRbrnHmnrre35DQ3SmDHS3/4mfeUr0ve+J116qXTTTdLy5dJJJ1V/mgcOlObMiX9szhx7PCm9etnt0qV229xst126tH1u165SU1P8+7S0SK+91np5yueLX5QaG6UHH5SmTZO++11pzz2l448ve/KrrndvaZ997P/jxqU7LZWy4YbSzjtLf/qTHc5J0ssvS6edJu2xh3TqqelOXy0bN06aODG7vEjSW2/Z7QYblPdeXbpIhx8uPf64rXu+//3KTOOwYXY7c2bx586fb7f9+1fms6ulb19b1zuX9pTE699f+vvfpdNPl664QtphB7t97z17fPx4aautpMcek/7wB+mss5KbtuHD7fbjj+Mff/ZZu/3c50p7vx12yL7uzTdtHbLppq2f87nP2d/qnnukiy6SNtrIbg84oLRtRC7npM02k3bZRdp/f+k737H9tUceaf0876Xnn5e22y57X2Oj9Mc/2rz+1FO2vV+yRLr++rafM22atO66tp8QXvulL9n/99tP6tGj8HQOGCAtWJDdHkrSJpuU//tWy1ZbST/6kXT//bZ/N2uW9Oc/l/bac8+VLr7Y/uannmrfzde/Lo0aZY+FfYdc//mPfR977GHrueXLS5/e5ub4fZFivvAF+xs+9FDr+99806a3Wnr0kDbeWHrySel3v5P23tuWv/vuk371K+kf/5AGD277up49pRNPlO6919YVa61l+1u5Bg6UrrrK9kvK8fnPSxMm2Hf/1lu2rar2uvRXv5LOOcfWiaedJp15pq0jo7//llvacvn++4Xfq6XFbhsqeKjunHT55dLChdIFF7R+7NNPpSeekA47rHa3OZW03np2G7ZXS5bYOmLMGGnZMunOO8t7v0WL7DYcX0Ttv7/tbx5yyKrv8w8dardh2zZ7drLHTzWCAKteDRxoGwuJACvo2tWCnLfekqZOLf/ALGknnSTdeKPtPPzoR7bSHThQ+u9/pb/+1QKTXF262GuOOMI2shdcYDu97dm5LtegQXYAvGRJ28eSDrDCDnmYlrATGrfz1K1b/gDrvffsd8o9kIkzYIAd4PzrX7bT1tAgXXdd7e4onHaaBRfbb5/2lFTOCSdIb79tO8sLF9qO2uqrW5Db2Jj21NWunXayZWDiRPt50iTpN7+RRo+O31kr5sgjs7cbbVSZaQwBVr7AImrePLsdMKAyn13PunSRrrzSguGlS6Vvf1taf31bJ+62m+1fPPecnRBKcl1XLNB89lkLCEePLu39Ro+25z/zTDakyV3v9+8vbb659NvfSuefb/tYN9xgJ44qoWtXaddd2wZYH35oB59jx7Z9TY8e0o47Wvi111627IYAP3j/fWnEiNb37buv3R56aPHpGjDA3nPBAun11+2+Ur/XNAwaVHpA1L+/nfBbY43sfV26SL/4hQVDf/lL/Oseekjq10/afXf7ecGC0qevqSl+X6SY1Vaz7fXDD2fvmzfP5o2NNy7//cqx1VYWfpxyivTOO7b/MHGidPbZhQOYb33Lbp94IhsoVMpOO1l4NXGi7ddvuGFl3z/O1ltLl1xi/375S/u3yy6tn7PVVnb70kuF32vlSrut9L7JJptIJ59sJ8GfeCK7Prj7bgvNDjussp9XqwYNsv2XadPs5wcekBYvtr/dRhvZsVI5Fi2y98v391pzTQvFzj57lSb7f9u2Tz6x29mz7XepMwRY9SyEFgRYrfXoUfkNabUcfbQdHHz8sYVWBx9sochXvpL/NV262AH7YYfZWaATT0xmWkNAlVuF5b00d26yK+CGBvs7l1qBtWJF/PtMnmy3m21W2ufuvbdVyj36qJ0RX2ed8qY7SbvsYmdU2xNQ1KpDD7WDij/+UfrGN2xH+29/k4YMSXvKatuOO9rtk09aZcFuu9l8ce+97Xu/HXawnedLL63YJJYVYIUKrL59K/f59e7rX5defdUC4l//2v4eJ55oB49JnCDJVWx+ePZZC3xKPThsbLRw4NlnLcDq2jX+JNfFF1tF9DvvWOXZscdapUml7L67HYxHqzcmTLDbaAVWnDFjLMyYMaP1/dOmtQ2wDj3UKkcOOqj4NIUgeN48C7DWXbfz71ceeKCtx370o7Yn5by3EOkLX8ju95QTYLW3Akuyqu7nn7d9Ksmq8aXqVmBJFthecYX9/adOtXV7Kcv9iBFWnRL+X0k77WS3//mPzeNJBFil2GwzC/Nffrnw86oVYEnSj39s8+auu0ojR1rF2HXXWXCz+eaV/7xa5JzNcyHAuvVWq27adVdbbz/xRPaxUixalMx6L7cCK+kCgBpBgFXPwlmAzr6jUQ+GDLGhJjfdVFog1aWLdPPNdpbwhBOqP31S/gBr/nzbUCe9Au7ZM7vjGSqs4nYaC1VglTtcYu+97Xb33ZMZtonWevWyqp+bbrKdlZ/+tO3ZUbQ1cKBVVFx/vQ0tWmcdG5bU3uop5yxADCFDJYQQstQKrH79qLqrhpEjbajgo49aSJlWSFgowFqyxA4eSx0+GHzuc9Irr1hAMGpUfJXMl75kB/Prr1/+NJciVPQ8+mj2vueft2nZYovCr91mG7udNCl735IlFmqtu27r5zY22vDBUoYv5QZYtTR8sFqcs+qaGTOsAjEqBIx77WXrGSmZCizJAqyWFgttJKsSk6ofYG28sXTGGbadKLfSMgzfr/SJ48GDbbpuusm+k1oJsHr3tm1nmgHW6qvbCdhrrrFA7frr7eTqEUfU7qiAaggB1sKF0j//aSf2Gxulo46yx2+6qfT3SirAilZgNTfbepcKLNSVEGCFDSzqS9eutrOT1EFcvgBr9uzWjyelV6/KVGCtu27pB2ljxtgOw4031tdOQi058UQ7Q7733tbDBKXZaSfp3Xelbbe1IZhrrpn2FLXWtavtxJVagVXr/a+wavr0sX9xQwgnTbKDw/YEWCtXWmVVKcPGq2GTTewAJjqM8Pnnrdqle/fCr91qK9vuvPBC9r5QybUq1S8hwJozxwKTegiwJFsnfvnL0s9/bsFmEIbw7blndj2TVAXW2LG2Tx+mYcoUe69qBaqV8MUvWtXmccdV/r132skqwqTaCbAkW16LDSGsRg+sqCFD7ETS/ffbsvvEE9J551Xns2pVCLDuv9/6Xh1xRPb+XXaxffXcIdf5LF6cTIA1cKAdt33ySbbSkgALdeXLX7am5aE5KVBNYQWbG2CFn9MIsHJ7YJXbxH3y5NKHD0rVqTxBebbd1nbUbr+9ejuGndG3v20N1//9bzt7W4uGDy+9Aov+V53fsGHx80No4F5uf78QeMU1cE+Kc1aF9eijNh0tLRbIFRs+KFnlx8YbVy/AevFFOwhM67tJw+9+Z4HRvvtmh2Y+9JANL11//fZVYDU3t78Cq0sXC4QeftjmjzfftKrI9r5fEpyzqs3cKsBK+Pzns/+vpQBrq62sh2oYzh6nmhVYuXr2tIvcVHK4c0ew3nrSZ5/ZsfBaa2XbJUg2jHDq1OwQ7WKSqsBqaLDw8eOP0ysAqAHsvdezbt1sGBPDKJCEsIINK9wgrQArOoSwUIDVrVt8BVZTk53dLCfAQm3YeWf6H5Vr9Gi7olot90QbNqz0qxBSgdX5FQqwRo6MvzJaIQMHZg+C01zv7767DfubPNkOsBYsKC3AkqwKODqEMPR4WZXwIARYTz9tt/VSgSVZJeo//mEHwfvvb7ePPWbVV1L7hxC2twJLss9+/30byjhlSvWHD9ay0Aerf//aqlIJ/cGilXu5kgyw6lUI7sePt6sjR09qHnqo9cr9619Le69Fi8q/UmZ7DR1qFVjh+KmW5u2EEGABSEao2qilCqzcIYRxZynzVWC9/bYFW/V0thmoZfkCi1xUYNWH4cPjA81nny1/+GAQXpfmej9cYfiRR0pv4B5ss41VCoXlZNo028YNH97+6ckNsGr5CoTVsPXW0m232ZCwnXayoUS5AVahSptcq1KBJWU/+4EHbD+l2lcgrGXrrWdXkNxww9pq2xACrEJ9sAiwqi9aeZp78at+/exiDbfemr+NSFRSFVhSdl+HCiwAqLJu3azqpVYCrHIqsOICrHKvQAigusJOXbGeFVRg1Ye4QHP6dAtwxo5t33sedZS0xx5WwZWWtdayUOKRR6z/Ve/epYdGY8bYbRhG+P77dlGGVTlIDiHNJ5/YtNVjX9V995V+8xtrYt+li12pVUqnAmv99W0I43XX2YF3PVdgOSddfrn0wx+mPSWtrbGGVc0UCrCq3QML2QBr/fWtvUSuY46xPlMPPlj8vZIMsKjA0iqsIQGgTIMGxQdYDQ3JV0T06pU9uCl0FcJ8Tdxfe82mu57PbgK1ZNgwaflyC6gKrU+owKoPw4bZvLB0aba3y3PP2W25/a+Cvfayf2n74helv/zFtmFjxpQeQIWL97zwgrTPPlaBtSr9ryT77H79LKSp54rkU06x+W3+/Gxw1aOH7UMk1QMr2HNP688l1XeAJdnQsFrjXPFG7lRgVd/AgVald+KJ8RV6u+9u+wp33219owtJugLrk0+kWbPsZyqwAKCKBg6M74G1+urJn2WqRAXWyJH11/QSqFXh4giFhhF6TwVWvYhebjyYMMHCgTCEp6PafXcbqvbCC6UPH5QsWNloo2wfrGnTKtM8OwTC9dT/Ks73vy9dckn2Z+ey4V6pVrUCS8oOI5Q4yVarttrK9iPD/mcuAqzqc856xeW7InW3btJ++0l//3v+v1OQdAXWihU2RLhnz9ruTVolBFgAkjNwYNsKrNmz0zl7ENcDq5wKrHKvQAigukIfn0KN3D/6yA4Mym3gjY4nbn6YMMEOHHv0SGWSKmbXXbMnfcoJsCTrg/XCC3bFwI8/XvUKLIkAq5ByA6xKVGDttpsFHwMH1mV1Roew5Za2DE6dGv84AVYyGhsL90c76CAbRjh+fP7nrFxpf8skK7AkGwlSp8s3ARaA5MQFWHPmpLMCLrUCK66J+7JlduaDAAuoHaVUYN17r93uvXfVJwcpy50fVq6UJk5sf/+rWjJgQDa4KjfAGjNG+uCDbB8sAqzqSqMCq18/C7FCzzPUnmKN3OmBVRv22stOeNxzT/7nLF5st0lWYEnWc68O+19J9MACkKRBg+KHEK6zTvLTUmoFVrdubSuwpkyxg6F67vcB1JpSAqy77rKDbIbVdH6588OUKdLChZ0jwJKkI4+0bdh665X3um22sdu777ZbAqzq6tcv2asQBnfeWfyCFkjP6NEWTr3+evzjVGDVht697YTXvffahRriqrUWLbLbpCuwFi6kAgsAqm7gQDsTGa1oqqUKrLidxrgKLK5ACNSeAQMscM4XYH3yifTf/0qHHproZCElgwfbwUYYQjhhgt12lgDrjDOseqPQ8Jc4uQFWJXpgrbmmBWlcHKGtNCqwJOvzx9+jdnXtKnXvbhX9cQiwasdBB1n7gYkT4x9POsAKFVhS3VZgEWABSE4IqubOzd6XVoDVq5ftIDQ1ld/E/bXXbOdjww2rP50ASuOcnZnMF2Dde68NyzjkkEQnCynp0kUaMiQ7P0yYkG1iXs8GDLDLxr/3nn1Ha6yx6u950UXSY4+t+vt0Rv37l98DqxIBFmpfY2M2qMpFgFU79tvP/g4h9M8VAqzevZOZntVWy55wJ8ACgCoLQVXog7V0qf1LYwUcrtqxZEk2oCq1ifvkyXYQ1K1bdacRQHkKBVh33SVtsIG0+ebJThPSE50fJkywflH0lMn2Rlp77cqEJautVplKrs6oPRVYlRhCiNrX0JA/wKIHVu1YfXW7cEa+PlhJV2A1NNjJGYkhhABQdSGoCn2wQpCV1hBCyQKscpu4cwVCoDYNGxZ/FcK5c61C5JBDyh9yhY4rBFhLl9pwu84yfHBVhWGEhE7V156rEFKBVR8aG7NBVS4qsGrLQQdZH8U33mj7WNIBlpQdRkgFFgBUWW4FVgiy0hpCKNmBTSlN3EMT1sWLbegFARZQe4YPj6/Auv9+W87pf1VfQqD50kv29yfAMqECqxIN3FFYv37W5yi3kjsfKrDqB0MIO44DD7TbuCqsNAKs0MidCiwAqLLcAKsjVGDtvLOV626zjfT889krxnAFQqD2DBtmwXhu1eSdd1q1CZeVry/Dh1vz/meftZ+33z7d6akV22xj27UNNkh7Sjq/fv3sttQqLCqw6gcBVsex5pq2/YgLsBYvtlsqsBJDgAUgObUUYMVVYMWd9dxjD2n8eHvOuHHS+efb/VRgAbVn2DC7dPysWdn7FiyQ/v1v6eCDGT5Yb4YNszDzoYektdayQAu2zR0/Xjr11LSnpPMrN8CiAqt+FAqwwtBCtlm146CD7EqE06e3vp8KrMQRYAFITq9eUo8eHasHliTtuKP04ovSvvtKDz9sv8P66ycznQBKF3bqosMI//EPG77D1QfrT5gf/vMfhg/mGjfOrpCH6grfMRVYyFWsAquhgQCrloTege+80/r+NAKscDJm8ODkPrOGsIYEkBznLKyqtQqsQlchDFZf3S6he/31FnpR1g3UnhBYRBu533WX7eztsEM604T0hJ38piYCLKSDCizkUyzAYj+ztoRhe5980vr+RYvs+CHJK5MffbQ0YIC0zjrJfWYNIcACkKzcAKtv32RX+kE5FViBc9JJJ1V3ugC0X24F1uWXW/D87W9zOfJ6FOYHiQAL6aAHFvIhwOpYCgVYffokWy232mrSMcck93k1hr05AMmKBlizZ6c3frvUqxAC6DhCYDFjhvTd70pnnWVDB3/+83SnC+kI84NzNPBHOsoJsLy3/REqsOpDsR5YnHSpLYMG2d8kLsDq3TudaapTHK0BSNagQdIrr9j/58xJP8AqpwILQG3r0cPK6i+91A4YTzlFuvJKzmTXq759rdp2vfWyQQKQpDDfzZ9f/LkhzGBfpD5QgdWxNDbaMUy+CiwkhmgXQLJyhxCmFWCFIYRUYAGdy7BhFl5ddJF01VUcBNQz56QttpD22ivtKUG9KqcCq9AVkdH5EGB1PEOHEmDVAI7WACRr4EBp7lwrj54zRxo5Mp3piKvAYqcR6Ph+/GMLLg47LO0pQS0YP56hOEhPz552cqyUAKuUC8qg82hoIMDqaOICrMWLCbASxhoSQLIGDrTwav78dCuwune3g9wlS7IHN+wsAB3f4YenPQWoJWlcJAQInLMqLCqwkKux0faH49ADqzYNHSq9/Xbr+xYtan3BEFQdSwaAZA0aZLeffCLNm5degOWcnRldutTOejY2JnsFEQAA0PmVGmBRgVVfGELY8YQKLO+z9zGEMHEEWACSFQKrt96y2xBopaFnz+wQQnYYAQBApVGBhTgEWB3P0KF24nvRoux9BFiJI8ACkKwQYE2d2vrnNPTqlW3iToAFAAAqrV+/0q5CSAVWfSHA6niGDrXbaB8sAqzEEWABSFYIrKZMaf1zGqjAAgAA1UQFFuIUCrDogVWbcgMs7wmwUsCSASBZYchgrVVgscMIAAAqrX9/emChLSqwOp7cAGv5cvtbEWAligALQLL697eNci0EWFRgAQCAaiq3Aov9kfpAgNXx5AZYoRdW797pTE+dIsACkCznpNVXl2bOtJ9roQKrqYkdRgAAUHnlXoWQivD6QIDV8QwebLchwFq82G6pwEoUARaA5IVhhF27prvS79WLCiwAAFA9/fplT5YVQgVWfaEHVsfTtaudeM+twCLAShRLBoDkhaqrQYOsIistDCEEAADV1K+f3RarwqICq75QgdUxDR1KgJUyAiwAyQsBVprDB6XWTdwJsAAAQKWVGmBRgVVfGhoIsDoiAqzUEWABSF6tBFhUYAEAgGrq399uqcBCVGOjDRWMQ4BVuwiwUkeABSB5oQdW2gFWtAKLHUYAAFBpVGAhDj2wOiYCrNSxZABIXq1VYHEVQgAAUA30wEIcemB1TEOHWnC1ZAkBVkoIsAAkr1YCrF69JO/tMrgEWAAAoNJCgDV/fuHnUYFVXwiwOqahQ+32k0/s+EEiwEoYARaA5NVKgNWzp90uWMAOIwAAqDwqsBCHAKtjigZYoQKrV6/0pqcOEWABSF7ogRVu0xI2OAsXEmABAIDKowcW4tADq2PKDbB69iRsTBhLBoDkjR4tbb65NHZsutNBBRYAAKimXr3sAJcKLERRgdUx5QZYDB9MHEdsAJI3cKD0yitpT0XrCix2GAEAQKU5Z1VYVGAhigCrYxoyxG4JsFJDBRaA+hUCrOXL2WEEAADVUU6AxQm1+kCA1TF17y4NGCB9/DEBVkoIsADUrzCEUCLAAgAA1dGvX/GrEIYhhOyP1Ad6YHVcQ4dSgZUi1pAA6lf0qiHsMAIAgGqgAgu5GhqowOqoQoC1ciUBVgqIdgHULyqwAABAtZUSYFGBVV8aG63SKg4BVm0LAdbixQRYKSDAAlC/qMACAADV1r8/FVhojR5YHRdDCFPFERuA+kUFFgAAqDYqsJCLHlgd19Ch2eW5d+90p6UOsWQAqF/RCizOeAIAgGoopwcWlTf1gQqsjmvoULtdsIAKrBQQYAGoX1RgAQCAauvXT1qyJFtlFaepyfZFnEtuupAeAqyOKwRYEgFWCgiwANSvbt2yJdoEWAAAoBr69bPbhQvzP6e5mX2RehICqrhG7gwhrG0EWKliyQBQv5zLVmGx0wgAAKohBFiFhhE2NdHOoJ6EACuuCosKrNpGgJUqAiwA9S30wSLAAgAA1dC/v90WCrCowKovBFgdFwFWqgiwANQ3AiwAAFBNVGAhFwFWx9Wzp9S3r/2fACtxBFgA6lsYQshOIwAAqIZSAiwqsOpLoQCLHli1L1RhEWAljiUDQH2jAgsAAFRTCLDmz8//HCqw6ksIqKjA6pgIsFJDgAWgvtHEHQAAVBMVWMhV6CqEBFi1LwRYvXunOx11iAALQH2jAgsAAFQTPbCQix5YHRsVWKkhwAJQ36jAAgAA1dS7tw0ZowILAT2wOjYCrNSwlgRQ36jAAgAA1eScVWHNm5f/OVRg1RcqsDq2Qw+1nnYDB6Y9JXWHaBdAfaMCCwAAVNsWW0h//7sFVXGowKovBFgd26abSpddZuE0EkWABaC+hQosznoCAIBq+d73pPffl26+Of5xKrDqS74Ay3sbQkiABcQiwAJQ36jAAgAA1falL0lbby39/OfxVTdUYNWXQgGWRA8sIA+WDAD1jR5YAACg2pyTfvAD6a23pDvuaPs4FVj1JV+AFX6mAguIRYAFoL4RYAEAgCQcdJA0erR00UU2TCyKCqz6QoAFtAsBFoD6xhBCAACQhIYGq8KaPNkaukc1N1OBVU8IsIB2IcACUN+owAIAAEk54ghp5EjpZz/L9juSbAgh+yL1I/S4yq3ECz/TAwuIxZIBoL6FCizOegIAgGrr0kU691xp4kTp4Yez91OBVV+owALahQALQH2jAgsAACTp2GOl/v2le+7J3kcFVn0hwALahQALQH2jBxYAAEhSt27SDjtITz6ZvY8KrPpCgAW0CwEWgPq2zTbSnntKm2+e9pQAAIB6MW6c9Npr0mef2c9UYNWXfAEWPbCAglgyANS3IUOkhx6SBg1Ke0oAAEC9GDfObp95xm6bmwmw6gkVWEC7EGABAAAAQJLGjrXAKgwjbGpiCGE9IcAC2oUACwAAAACS1Lu3tPXW0lNP2c9UYNUXAiygXQiwAAAAACBp48ZJEyZIK1ZQgVVv6IEFtAtLBgAAAAAkbdw4adky6YUXqMCqN1RgAe1CgAUAAAAASQuN3J96igqsekOABbQLARYAAAAAJG34cGn99aXx4+1nKrDqRxgiGIYMBgRYQEEEWAAAAACQhnHjsgEWFVj1gx5YQLuwZAAAAABAGsaNk+bNs/9TgVU/GEIItAsBFgAAAACkYaedsv+nAqt+EGAB7UKABQAAAABpGD1aGjDA/k8FVv0gwALahQALAAAAANLQ0CDtuKP9nwqs+kEPLKBdCi4ZzrnuSU0IAAAAANSdcePslgqs+kEFFtAuxaLdZyTJOXdjAtMCAAAAAPUl9MHq1i3d6UByCLCAdikW83dzzn1N0o7OuYNzH/Te312dyQIAAACAOjBunPTLX0r77JP2lCApBFhAuxQLsL4l6ShJAyTtn/OYl0SABQAAAADt1dgonXNO2lOBJNEDC2iXYgHWcO/9yc65F7331yYyRQAAAAAAdFZUYAHtUizaPS9z+61qTwgAAAAAAJ1eCKhCxVVAgAUUVKwCa45z7jFJ6znn7s990Hv/5epMFgAAAAAAnVAYIkgFFlCWYgHWvpK2kXSjpF9Xf3IAAAAAAOjE6IEFtEvBAMt7v0LSs865Hb33sxKaJgAAAAAAOid6YAHtUjDAcs5d4b0/U9KfnHM+9/FqDCF0zl0o6SRJITD7vvf+gUp/DgAAAAAAiSPAAtql2BDCGzO3l1Z7QnJc7r1P+jMBAAAAAKguAiygXYoNIZyUuX3COTc483+GEgIAAAAA0B75mrjTAwsoqOCS4cyFzrnZkt6UNNU5N8s5d0GVp+tU59wrzrk/OedWq/JnAQAAAACQnMZGKrCAMhWLds+UNE7Sdt77gd771SRtL2mcc+7b7f1Q59wjzrnJMf8OkPR7SSMlbSVppvJc/dA59w3n3ETn3MRZsygKAwAAAAB0EARYQNmK9cA6VtIe3vvZ4Q7v/bvOuaMlPSzp8vZ8qPd+91Ke55y7TtI/8rzHtZKulaRtt922TYN5AAAAAABqEgEWULZiFVhdo+FVkOmD1bUaE+ScGx758SBJk6vxOQAAAAAApCIuwKIHFlBQsQqsFe18bFX80jm3lSQvaZqkb1bpcwAAAAAASF5jYzawCqjAAgoqFmBt6ZxbEHO/k9SjCtMj7/0x1XhfAAAAAABqQkMDQwiBMhUMsLz3LDkAAAAAAFQSPbCAsjG4FgAAAACAJNEDCygbSwYAAAAAAEmiAgsoGwEWAAAAAABJIsACykaABQAAAABAkgiwgLIRYAEAAAAAkCR6YAFlY8kAAAAAACBJVGABZSPAAgAAAAAgSYUCLCqwgFgsGQAAAAAAJKmxMTtkMFi50sIr59KZJqDGEWABAAAAAJCkfD2wqL4C8mLpAAAAAAAgSQ0N8UMI6X8F5EWABQAAAABAkvL1wCLAAvIiwAIAAAAAIEkEWEDZCLAAAAAAAEgSPbCAsrF0AAAAAACQJCqwgLIRYAEAAAAAkCQCLKBsBFgAAAAAACSJAAsoGwEWAAAAAABJogcWUDaWDgAAAAAAkkQFFlA2AiwAAAAAAJLU2GgVV1EEWEBBBFgAAAAAACSJCiygbARYAAAAAAAkqaGBHlhAmVg6AAAAAABIEhVYQNkIsAAAAAAASBIBFlA2AiwAAAAAAJJEgAWUjQALAAAAAIAkxQVY9MACCmLpAAAAAAAgSVRgAWUjwAIAAAAAIEkEWEDZCLAAAAAAAEgSARZQNgIsAAAAAACSRA8soGwsHQAAAAAAJKmx0QKrKCqwgIIIsAAAAAAASBJDCIGyEWABAAAAAJCkhgYCLKBMBFgAAAAAACSJHlhA2Vg6AAAAAABIEkMIgbIRYAEAAAAAkCQCLKBsBFgAAAAAACSJAAsoGwEWAAAAAABJCkFVS0v2PnpgAQWxdAAAAAAAkKQQYEWrsKjAAgoiwAIAAAAAIEkEWEDZCLAAAAAAAEgSARZQNgIsAAAAAACSRA8soGwsHQAAAAAAJIkKLKBsBFgAAAAAACQpVFoRYAElI8ACAAAAACBJVGABZSPAAgAAAAAgSXEBFj2wgIJYOgAAAAAASBIVWEDZCLAAAAAAAEgSARZQNgIsAAAAAACSRIAFlI0ACwAAAACAJNEDCygbSwcAAAAAAEmiAgsoGwEWAAAAAABJCkFVS0v2PgIsoCACLAAAAAAAkkQFFlA2AiwAAAAAAJJEDyygbCwdAAAAAAAkKQRVIcAKQwmpwALyIsACAAAAACBJuRVY4ZYAC8iLAAsAAAAAgCQRYAFlI8ACAAAAACBJuQFWGEJIDywgL5YOAAAAAACSRAUWUDYCLAAAAAAAkkSABZSNAAsAAAAAgCQRYAFlI8ACAAAAACBJ9MACysbSAQAAAABAkkKAFYIrKrCAogiwAAAAAABIEkMIgbIRYAEAAAAAkCQCLKBsBFgAAAAAACQp9LqiBxZQMpYOAAAAAACSRAUWUDYCLAAAAAAAkkSABZSNAAsAAAAAgCQRYAFlI8ACAAAAACBJuQEWPbCAolg6AAAAAABIEhVYQNkIsAAAAAAASBIBFlA2AiwAAAAAAJKUL8BiCCGQF0sHAAAAAABJCgFW6H0VbqnAAvIiwAIAAAAAIEkMIQTKRoAFAAAAAECSCLCAshFgAQAAAACQpNDrih5YQMlYOgAAAAAASFJuBRY9sICiCLAAAAAAAEgSQwiBshFgAQAAAACQJAIsoGwEWAAAAAAAJClfgEUPLCAvlg4AAAAAAJKU28SdHlhAUQRYAAAAAAAkrbGRIYRAGQiwAAAAAABIGgEWUBYCLAAAAAAAktbYmB06SA8soCiWDgAAAAAAkhatwKIHFlAUARYAAAAAAEljCCFQFgIsAAAAAACS1tBAgAWUgQALAAAAAICkxVVg0QMLyIulAwAAAACApNEDCygLARYAAAAAAEmjBxZQFgIsAAAAAACSRoAFlIUACwAAAACApNEDCygLSwcAAAAAAEmjBxZQFgIsAAAAAACSxhBCoCwEWAAAAAAAJK2xMVt5RYAFFEWABQAAAABA0uiBBZSFpQMAAAAAgKTRAwsoSyoBlnPuMOfca865FufctjmPneece9s5N8U5t1ca0wcAAAAAQFXRAwsoS5eUPneypIMlXRO90zm3iaSvSNpU0hqSHnHObeS9X5n8JAIAAAAAUCUNDQRYQBlSqcDy3r/hvZ8S89ABkm713i/33r8n6W1JY5OdOgAAAAAAqoweWEBZam3pWFPSh5GfP8rcBwAAAABA55HbA8s5+wcgVtWGEDrnHpE0LOahH3jv78v3spj7fJ73/4akb0jSOuus065pBAAAAAAgFbkVWAwfBAqqWoDlvd+9HS/7SNLakZ/XkjQjz/tfK+laSdp2221jQy4AAAAAAGoSARZQllobQni/pK8457o759aTtKGkCSlPEwAAAAAAlZUbYNH/CigolSXEOXeQc+4jSTtI+qdz7iFJ8t6/Jul2Sa9LelDSKVyBEAAAAADQ6TQ2Wu8ryW6pwAIKqtoQwkK89/dIuifPYxdJuijZKQIAAAAAIEEMIQTKQo0iAAAAAABJI8ACykKABQAAAABA0uiBBZSFJQQAAAAAgKRFAyx6YAFFEWABAAAAAJC0hgaGEAJlIMACAAAAACBp9MACykKABQAAAABA0uiBBZSFJQQAAAAAgKTRAwsoCwEWAAAAAABJYwghUBYCLAAAAAAAkkaABZSFAAsAAAAAgKQ1NtrQQYkeWEAJWEIAAAAAAEgaPbCAshBgAQAAAACQNIYQAmUhwAIAAAAAIGkEWEBZCLAAAAAAAEhaboBFDyygIJYQAAAAAACS1tBADyygDARYAAAAAAAkjSGEQFkIsAAAAAAASBoBFlAWAiwAAAAAAJJGDyygLCwhAAAAAAAkLVRctbTQAwsoAQEWAAAAAABJC4HVypUMIQRKQIAFAAAAAEDSohVYBFhAUQRYAAAAAAAkLbcCix5YQEEsIQAAAAAAJC0aYNEDCyiKAAsAAAAAgKTRAwsoCwEWAAAAAABJI8ACykKABQAAAABA0kLPK3pgASVhCQEAAAAAIGn0wALKQoAFAAAAAEDSGEIIlIUACwAAAACApBFgAWUhwAIAAAAAIGm5ARY9sICCWEIAAAAAAEgaPbCAshBgAQAAAACQtBBYtbQwhBAoAQEWAAAAAABJowcWUBYCLAAAAAAAkkYPLKAsLCEAAAAAACSNHlhAWQiwAAAAAABIGkMIgbIQYAEAAAAAkLQwZJAACygJARYAAAAAAEmjBxZQFpYQAAAAAACSRg8soCwEWAAAAAAAJI0eWEBZCLAAAAAAAEhaCKyam1v/DCAWARYAAAAAAEkLgdWKFXZLDyygIJYQAAAAAACSFgKspqbWPwOIRYAFAAAAAEDSCLCAshBgAQAAAACQtNwhhARYQEEEWAAAAAAAJI0eWEBZWEIAAAAAAEgaFVhAWQiwAAAAAABIWqi4ogcWUBICLAAAAAAAkkYFFlAWAiwAAAAAAJJGDyygLCwhAAAAAAAkjQosoCwEWAAAAAAAJC0EVvTAAkpCgAUAAAAAQNKowALKQoAFAAAAAEDSciuw6IEFFMQSAgAAAABA0qjAAspCgAUAAAAAQNLogQWUhQALAAAAAICkUYEFlIUACwAAAACApOUGWPTAAgpiCQEAAAAAIGkhsKICCygJARYAAAAAAEmjBxZQFgIsAAAAAACSRg8soCwEWAAAAAAAJI0eWEBZWEIAAAAAAEgaPbCAshBgAQAAAACQhsZGemABJSLAAgAAAAAgDQRYQMkIsAAAAAAASENjIz2wgBKxhAAAAAAAkIZogEUFFlAQARYAAAAAAGlgCCFQMgIsAAAAAADSQAUWUDICLAAAAAAA0tDQQA8soEQsIQAAAAAApIEKLKBkBFgAAAAAAKSBHlhAyQiwAAAAAABIAxVYQMkIsAAAAAAASEM0wKIHFlAQSwgAAAAAAGlobJS8z/4fQF4EWAAAAAAApCEaWhFgAQURYAEAAAAAkAYCLKBkBFgAAAAAAKQhGlrRAwsoiCUEAAAAAIA0UIEFlIwACwAAAACANBBgASUjwAIAAAAAIA3RYYMEWEBBBFgAAAAAAKSBHlhAyVhCAAAAAABIA0MIgZIRYAEAAAAAkAYCLKBkBFgAAAAAAKSBIYRAyVhCAAAAAABIAxVYQMkIsAAAAAAASAMBFlAyAiwAAAAAANIQQivn7B+AvAiwAAAAAABIQwiw6H8FFMVSAgAAAABAGkKAxfBBoCgCLAAAAAAA0kCABZSMAAsAAAAAgDSEoYMEWEBRBFgAAAAAAKSBHlhAyVhKAAAAAABIA0MIgZIRYAEAAAAAkAYCLKBkBFgAAAAAAKSBAAsoGQEWAAAAAABpoAcWUDKWEgAAAAAA0kAFFlCyVAIs59xhzrnXnHMtzrltI/ePcM4tdc69lPn3hzSmDwAAAACAqiPAAkrWJaXPnSzpYEnXxDz2jvd+q2QnBwAAAACAhBFgASVLJcDy3r8hSc65ND4eAAAAAID00QMLKFktLiXrOededM494Zz7fNoTAwAAAABAVVCBBZSsahVYzrlHJA2LeegH3vv78rxspqR1vPdznHNjJN3rnNvUe78g5v2/IekbkrTOOutUarIBAAAAAEgGARZQsqoFWN773dvxmuWSlmf+P8k5946kjSRNjHnutZKulaRtt93Wr9rUAgAAAACQsDB0kAALKKqmhhA65wY75xoz/19f0oaS3k13qgAAAAAAqAJ6YAElS2Upcc4d5Jz7SNIOkv7pnHso89DOkl5xzr0s6U5J3/Lez01jGgEAAAAAqCqGEAIlS+sqhPdIuifm/rsk3ZX8FAEAAAAAkDACLKBk1CkCAAAAAJAGAiygZARYAAAAAACkgR5YQMlYSgAAAAAASAMVWEDJCLAAAAAAAEgDARZQMgIsAAAAAADSQIAFlIwACwAAAACANNADCygZSwkAAAAAAGmgAgsoGQEWAAAAAABpCJVXBFhAUQRYAAAAAACkgQosoGQEWAAAAAAApIEeWEDJWEoAAAAAAEgDFVhAyQiwAAAAAABIAwEWUDICLAAAAAAA0kCABZSMAAsAAAAAgDTQAwsoGUsJAAAAAABpoAILKBkBFgAAAAAAaSDAAkpGgAUAAAAAQBoIsICSEWABAAAAAJAGemABJWMpAQAAAAAgDSG4ogILKIoACwAAAACANDCEECgZARYAAAAAAGkgwAJKRoAFAAAAAEAa6IEFlIylBAAAAACANFCBBZSMAAsAAAAAgDQQYAElI8ACAAAAACANBFhAyQiwAAAAAABIAz2wgJKxlAAAAAAAkAYqsICSEWABAAAAAJAGAiygZARYAAAAAACkgQALKBkBFgAAAAAAaQi9r+iBBRTFUgIAAAAAQBqowAJKRoAFAAAAAEAaCLCAkhFgAQAAAACQBgIsoGQEWAAAAAAApCEEV/TAAopiKQEAAAAAIA1du9ptly7pTgfQARBgAQAAAACQhiFDpMsvlw46KO0pAWoeMS8AAAAAAGlwTjrzzLSnAugQqMACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNI8ACAAAAAABATSPAAgAAAAAAQE0jwAIAAAAAAEBNc977tKdhlTnnZkl6P+3pqKBBkmanPRFAylgOAJYDQGI5ACSWA0BiOagX63rvB8c90CkCrM7GOTfRe79t2tMBpInlAGA5ACSWA0BiOQAklgMwhBAAAAAAAAA1jgALAAAAAAAANY0AqzZdm/YEADWA5QBgOQAklgNAYjkAJJaDukcPLAAAAAAAANQ0KrAAAAAAAABQ0wiwaoxzbm/n3BTn3NvOuXPTnh4gKc65ac65V51zLznnJmbuW90592/n3FuZ29XSnk6gkpxzf3LOfeqcmxy5L+9875w7L7N9mOKc2yudqQYqK89ycKFzbnpmm/CSc26fyGMsB+h0nHNrO+cec8694Zx7zTl3RuZ+tgmoGwWWA7YJkMQQwprinGuUNFXSHpI+kvS8pCO996+nOmFAApxz0yRt672fHbnvl5Lmeu8vzgS6q3nvv5fWNAKV5pzbWdIiSX/13m+WuS92vnfObSLpb5LGSlpD0iOSNvLer0xp8oGKyLMcXChpkff+0pznshygU3LODZc03Hv/gnOur6RJkg6UdJzYJqBOFFgODhfbBIgKrFozVtLb3vt3vfcrJN0q6YCUpwlI0wGSbsj8/wbZBgzoNLz34yXNzbk733x/gKRbvffLvffvSXpbtt0AOrQ8y0E+LAfolLz3M733L2T+v1DSG5LWFNsE1JECy0E+LAd1hgCrtqwp6cPIzx+p8AILdCZe0sPOuUnOuW9k7hvqvZ8p2QZN0pDUpg5ITr75nm0E6s2pzrlXMkMMw7AplgN0es65EZK2lvSc2CagTuUsBxLbBIgAq9a4mPsY44l6Mc57v42kL0k6JTOkBEAW2wjUk99LGilpK0kzJf06cz/LATo151wfSXdJOtN7v6DQU2PuY1lApxCzHLBNgCQCrFrzkaS1Iz+vJWlGStMCJMp7PyNz+6mke2Tlv59kxsKHMfGfpjeFQGLyzfdsI1A3vPefeO9Xeu9bJF2n7JAQlgN0Ws65rrKD9pu993dn7mabgLoStxywTUBAgFVbnpe0oXNuPedcN0lfkXR/ytMEVJ1zrnemUaOcc70l7Slpsmz+/1rmaV+TdF86UwgkKt98f7+krzjnujvn1pO0oaQJKUwfUHXhgD3jINk2QWI5QCflnHOS/ijpDe/9ZZGH2CagbuRbDtgmIOiS9gQgy3vf7Jw7VdJDkhol/cl7/1rKkwUkYaike2ybpS6SbvHeP+ice17S7c65EyR9IOmwFKcRqDjn3N8k7SppkHPuI0k/knSxYuZ77/1rzrnbJb0uqVnSKVxlB51BnuVgV+fcVrKhINMkfVNiOUCnNk7SMZJedc69lLnv+2KbgPqSbzk4km0CJMl5zxBRAAAAAAAA1C6GEAIAAAAAAKCmEWABAAAAAACgphFgAQAAAAAAoKYRYAEAAAAAAKCmEWABAAAAAACgphFgAQAAVIlzzjvnfh35+Wzn3IUVeu+/OOcOrcR7AQAA1DoCLAAAgOpZLulg59ygtCckyjnXmPY0AAAAlIMACwAAoHqaJV0r6du5D+RWUDnnFmVud3XOPeGcu905N9U5d7Fz7ijn3ATn3KvOuZGRt9ndOfffzPP2y7y+0Tn3K+fc8865V5xz34y872POuVskvVrNXxoAAKDSuqQ9AQAAAJ3cbyW94pz7ZRmv2VLSaElzJb0r6Xrv/Vjn3BmSTpN0ZuZ5IyTtImmkpMeccxtIOlbSfO/9ds657pKecs49nHn+WEmbee/fW8XfCQAAIFEEWAAAAFXkvV/gnPurpNMlLS3xZc9772dKknPuHUkhgHpV0m6R593uvW+R9JZz7l1JG0vaU9IWkequ/pI2lLRC0gTCKwAA0BERYAEAAFTfFZJekPTnyH3NyrRzcM45Sd0ijy2P/L8l8nOLWu+/+ZzP8ZKcpNO89w9FH3DO7SppcXsmHgAAIG30wAIAAKgy7/1cSbdLOiFy9zRJYzL/P0BS13a89WHOuYZMX6z1JU2R9JCkk51zXSXJObeRc653e6cdAACgFhBgAQAAJOPXkqJXI7xO0i7OuQmStlf7qqOmSHpC0r8kfct7v0zS9ZJel/SCc26ypGtE1T0AAOjgnPe5lecAAAAAAABA7aACCwAAAAAAADWNAAsAAAAAAAA1jQALAAAAAAAANY0ACwAAAAAAADWNAAsAAAAAAAA1jQALAAAAAAAANY0ACwAAAAAAADWNAAsAAAAAAAA17f8BUDxd3paUeN4AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1476x756 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#Y_hat=np.transpose(Y_hat)\n",
    "TTest=TTest.reshape(len(TTest),1)\n",
    "#print(Y_hat)\n",
    "#diff=np.concatenate((Y_hat,TTest),axis=1)\n",
    "diff=np.subtract(Y_hat,TTest)\n",
    "#print(diff)\n",
    "plt.figure(figsize=(20.5,15.5))\n",
    "plt.plot(diff,color='red')\n",
    "plt.xlabel('Number')\n",
    "plt.ylabel('Diff')\n",
    "plt.title('Difference between grounded and predicted value.')\n",
    "fig = plt.gcf()\n",
    "fig.set_size_inches(20.5, 10.5)\n",
    "#fig.savefig(figPredictName+'L_'+str(results)+'.jpg', dpi=150)\n",
    "plt.show()\n",
    "#model.save('1st-stage-subj-1.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "bb05838c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE Error:  1.4942832558198134\n"
     ]
    }
   ],
   "source": [
    "rmse=np.sqrt(np.square(np.subtract(TTest,Y_hat)).mean())\n",
    "print(\"RMSE Error: \",rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "5d4a8537",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 6.02222919  5.8028      0.21942919  5.9125146 ]\n",
      " [ 6.43925953  5.8028      0.63645953  6.12102976]\n",
      " [ 6.50778103  5.8028      0.70498103  6.15529051]\n",
      " ...\n",
      " [38.206604   37.526       0.680604   37.866302  ]\n",
      " [36.2293663  37.526      -1.2966337  36.87768315]\n",
      " [36.03376389 34.956       1.07776389 35.49488194]]\n"
     ]
    }
   ],
   "source": [
    "res=np.concatenate((Y_hat,TTest,Y_hat-TTest,(Y_hat+TTest)/2),axis=1)\n",
    "print(res)\n",
    "\n",
    "if exportCSV:\n",
    "    if HRInputFeats and fourInputFeats:\n",
    "        pd.DataFrame(res).to_csv('Subj-'+subjectID+'-predict-wHR.csv')\n",
    "    elif HRInputFeats:        \n",
    "        pd.DataFrame(res).to_csv('Subj-'+subjectID+'-predict-oHR.csv')\n",
    "    elif fourInputFeats:\n",
    "        pd.DataFrame(res).to_csv('Subj-'+subjectID+'-predict-noHR.csv')\n",
    "    elif onlyVel:\n",
    "        print(\"Store to oVel\")\n",
    "        pd.DataFrame(res).to_csv('Subj-'+subjectID+'-predict-oVel.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}