deep_learning_lstm_transformer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "tags": [
     "input"
    ]
   },
   "outputs": [],
   "source": [
    "splitPartCount = 5\n",
    "splitSeed = 27\n",
    "hoursPerWindow = 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Preproccess"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## read data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1213"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from utils.class_patient import Patients\n",
    "\n",
    "\n",
    "patients = Patients.loadPatients()\n",
    "len(patients)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## remove missing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# fill measures whose null represent false value\n",
    "\n",
    "from constants import NULLABLE_MEASURES\n",
    "\n",
    "\n",
    "nullableMeasures = NULLABLE_MEASURES\n",
    "\n",
    "for measureName in nullableMeasures:\n",
    "    patients.fillMissingMeasureValue(measureName, 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pco2 917\n",
      "ph 954\n",
      "po2 917\n",
      "albumin 406\n",
      "hba1c 326\n",
      "lymphocyte 446\n",
      "height 415\n",
      "urine-ketone 294\n",
      "crp 19\n"
     ]
    }
   ],
   "source": [
    "# remove measures with less than 80% of data\n",
    "\n",
    "measures = patients.getMeasures()\n",
    "\n",
    "for measure, count in measures.items():\n",
    "    if count < len(patients) * 80 / 100:\n",
    "        patients.removeMeasures([measure])\n",
    "        print(measure, count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1206"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# remove patients with less than 80% of data\n",
    "\n",
    "patients.removePatientByMissingFeatures()\n",
    "len(patients)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "82"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# # remove patients with positive tag in first 12 hours\n",
    "\n",
    "from pandas import Timedelta\n",
    "\n",
    "\n",
    "patients.removePatientAkiEarly(Timedelta(hours=12))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total  1124\n",
      "AKI  392\n",
      "Ratio  0.3487544483985765\n"
     ]
    }
   ],
   "source": [
    "print(\"Total \", len(patients))\n",
    "print(\"AKI \", sum([1 for p in patients if p.akdPositive]))\n",
    "print(\"Ratio \", sum([1 for p in patients if p.akdPositive]) / len(patients))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## split patients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "225"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "splitedPatients = patients.split(splitPartCount, splitSeed)\n",
    "\n",
    "len(splitedPatients[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "splitedPatients = patients.split(splitPartCount, splitSeed)\n",
    "\n",
    "\n",
    "def trainTest():\n",
    "    for i in range(splitedPatients.__len__()):\n",
    "        testPatients = splitedPatients[i]\n",
    "\n",
    "        trainPatientsList = splitedPatients[:i] + splitedPatients[i + 1 :]\n",
    "        trainPatients = Patients(patients=[])\n",
    "        for trainPatientsElem in trainPatientsList:\n",
    "            trainPatients += trainPatientsElem\n",
    "\n",
    "        yield trainPatients, testPatients\n",
    "\n",
    "\n",
    "def trainValTest():\n",
    "    for i in range(splitedPatients.__len__()):\n",
    "        testPatients = splitedPatients[i]\n",
    "\n",
    "        trainPatientsList = splitedPatients[:i] + splitedPatients[i + 1 :]\n",
    "        trainPatients = Patients(patients=[])\n",
    "        for trainPatientsElem in trainPatientsList:\n",
    "            trainPatients += trainPatientsElem\n",
    "\n",
    "        *trainPatients, valPatients = trainPatients.split(5, 27)\n",
    "        tmpPatients = Patients(patients=[])\n",
    "        for trainPatientsElem in trainPatients:\n",
    "            tmpPatients += trainPatientsElem\n",
    "        trainPatients = tmpPatients\n",
    "\n",
    "        yield trainPatients, valPatients, testPatients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "899 225\n",
      "899 225\n",
      "899 225\n",
      "899 225\n",
      "900 224\n"
     ]
    }
   ],
   "source": [
    "for trainPatients, testPatients in trainTest():\n",
    "    print(len(trainPatients.patientList), len(testPatients.patientList))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Transformer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seperate static and dynamic"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (74) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional_6\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional_6\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)        </span>┃<span style=\"font-weight: bold\"> Output Shape      </span>┃<span style=\"font-weight: bold\">    Param # </span>┃<span style=\"font-weight: bold\"> Connected to      </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ input_layer_13      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)        │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ input_layer_12      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)    │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lambda_12 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Lambda</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)     │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ input_layer_13[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lstm_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">LSTM</span>)       │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)    │     <span style=\"color: #00af00; text-decoration-color: #00af00\">19,200</span> │ input_layer_12[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lambda_13 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Lambda</span>)  │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)    │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ lambda_12[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ concatenate_6       │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">74</span>)    │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ lstm_6[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],     │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Concatenate</span>)       │                   │            │ lambda_13[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ transformer_encode… │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">24</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">74</span>)    │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ concatenate_6[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]… │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">TransformerEncode…</span> │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ global_average_poo… │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">74</span>)        │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ transformer_enco… │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">GlobalAveragePool…</span> │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │         <span style=\"color: #00af00; text-decoration-color: #00af00\">75</span> │ global_average_p… │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)       \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape     \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m   Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to     \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ input_layer_13      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m)        │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ input_layer_12      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m, \u001b[38;5;34m10\u001b[0m)    │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lambda_12 (\u001b[38;5;33mLambda\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m, \u001b[38;5;34m10\u001b[0m)     │          \u001b[38;5;34m0\u001b[0m │ input_layer_13[\u001b[38;5;34m0\u001b[0m… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lstm_6 (\u001b[38;5;33mLSTM\u001b[0m)       │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m, \u001b[38;5;34m64\u001b[0m)    │     \u001b[38;5;34m19,200\u001b[0m │ input_layer_12[\u001b[38;5;34m0\u001b[0m… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ lambda_13 (\u001b[38;5;33mLambda\u001b[0m)  │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m, \u001b[38;5;34m10\u001b[0m)    │          \u001b[38;5;34m0\u001b[0m │ lambda_12[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ concatenate_6       │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m, \u001b[38;5;34m74\u001b[0m)    │          \u001b[38;5;34m0\u001b[0m │ lstm_6[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],     │\n",
       "│ (\u001b[38;5;33mConcatenate\u001b[0m)       │                   │            │ lambda_13[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ transformer_encode… │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m24\u001b[0m, \u001b[38;5;34m74\u001b[0m)    │          \u001b[38;5;34m0\u001b[0m │ concatenate_6[\u001b[38;5;34m0\u001b[0m]… │\n",
       "│ (\u001b[38;5;33mTransformerEncode…\u001b[0m │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ global_average_poo… │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m74\u001b[0m)        │          \u001b[38;5;34m0\u001b[0m │ transformer_enco… │\n",
       "│ (\u001b[38;5;33mGlobalAveragePool…\u001b[0m │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense_6 (\u001b[38;5;33mDense\u001b[0m)     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │         \u001b[38;5;34m75\u001b[0m │ global_average_p… │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">19,275</span> (75.29 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m19,275\u001b[0m (75.29 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">19,275</span> (75.29 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m19,275\u001b[0m (75.29 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from keras.models import Sequential, Model\n",
    "from keras.layers import (\n",
    "    LSTM,\n",
    "    Dense,\n",
    "    Dropout,\n",
    "    Input,\n",
    "    Concatenate,\n",
    "    Masking,\n",
    "    Conv1D,\n",
    "    MaxPooling1D,\n",
    "    BatchNormalization,\n",
    ")\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "from tensorflow.keras.regularizers import l2\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from official.nlp.modeling.models import TransformerEncoder\n",
    "\n",
    "\n",
    "@keras.saving.register_keras_serializable()\n",
    "class TransformerEncoderWrapper(keras.layers.Layer):\n",
    "    def __init__(\n",
    "        self,\n",
    "        dropout_rate=0.1,\n",
    "        attention_dropout_rate=0.1,\n",
    "        **kwargs\n",
    "    ):\n",
    "        super(TransformerEncoderWrapper, self).__init__(**kwargs)\n",
    "        self.encoder = TransformerEncoder(\n",
    "            dropout_rate=dropout_rate,\n",
    "            attention_dropout_rate=attention_dropout_rate\n",
    "        )\n",
    "\n",
    "    def call(self, inputs, **kwargs):\n",
    "        return self.encoder(inputs, **kwargs)\n",
    "\n",
    "\n",
    "def createModel2(timeSteps, timeFeatures, staticFeatures):\n",
    "    # time series layers\n",
    "    seriesInputLayer = Input(shape=(timeSteps, timeFeatures))\n",
    "    seriesLayer = seriesInputLayer\n",
    "    # seriesLayer = LSTM(\n",
    "    #     64,\n",
    "    #     return_sequences=True,\n",
    "    # )(seriesInputLayer)\n",
    "    \n",
    "\n",
    "    # static layers\n",
    "    staticInputLayer = Input(shape=(staticFeatures,))\n",
    "    staticLayer = keras.layers.Lambda(lambda x: tf.expand_dims(x, axis=1))(staticInputLayer)\n",
    "    staticLayer = keras.layers.Lambda(lambda x: tf.repeat(x, timeSteps, axis=1))(staticLayer)\n",
    "\n",
    "    # combine layers\n",
    "    combined = Concatenate(axis=2)([seriesLayer, staticLayer])\n",
    "    combined = TransformerEncoderWrapper(\n",
    "        dropout_rate=0.1,\n",
    "        attention_dropout_rate=0.1\n",
    "    )(combined)\n",
    "    combined = keras.layers.GlobalAveragePooling1D()(combined)\n",
    "    # combined = Dense(32, activation=\"relu\")(combined)\n",
    "    # combined = Dropout(0.2)(combined)\n",
    "    # combined = Dense(16, activation=\"relu\")(combined)\n",
    "    combined = Dense(1, activation=\"sigmoid\")(combined)\n",
    "\n",
    "    model = Model(inputs=[seriesInputLayer, staticInputLayer], outputs=combined)\n",
    "    model.compile(\n",
    "        optimizer=Adam(learning_rate=0.0001),\n",
    "        loss=\"binary_crossentropy\",\n",
    "        metrics=[\"AUC\", \"accuracy\", \"precision\", \"recall\"],\n",
    "    )\n",
    "\n",
    "    return model\n",
    "createModel2(24, 10, 10).summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (126) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 21ms/step - AUC: 0.5660 - accuracy: 0.4862 - loss: 0.8433 - precision: 0.3844 - recall: 0.6906 - val_AUC: 0.5812 - val_accuracy: 0.5251 - val_loss: 0.8278 - val_precision: 0.3905 - val_recall: 0.6613\n",
      "Epoch 2/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.6070 - accuracy: 0.5240 - loss: 0.7896 - precision: 0.4117 - recall: 0.7534 - val_AUC: 0.5986 - val_accuracy: 0.5475 - val_loss: 0.7968 - val_precision: 0.4040 - val_recall: 0.6452\n",
      "Epoch 3/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.6034 - accuracy: 0.5531 - loss: 0.7826 - precision: 0.4207 - recall: 0.6960 - val_AUC: 0.6172 - val_accuracy: 0.5587 - val_loss: 0.7695 - val_precision: 0.4124 - val_recall: 0.6452\n",
      "Epoch 4/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6175 - accuracy: 0.5485 - loss: 0.7562 - precision: 0.4102 - recall: 0.6890 - val_AUC: 0.6341 - val_accuracy: 0.5698 - val_loss: 0.7457 - val_precision: 0.4211 - val_recall: 0.6452\n",
      "Epoch 5/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6429 - accuracy: 0.5648 - loss: 0.7302 - precision: 0.4388 - recall: 0.6994 - val_AUC: 0.6485 - val_accuracy: 0.5922 - val_loss: 0.7257 - val_precision: 0.4396 - val_recall: 0.6452\n",
      "Epoch 6/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6682 - accuracy: 0.6033 - loss: 0.7016 - precision: 0.4747 - recall: 0.7300 - val_AUC: 0.6610 - val_accuracy: 0.5978 - val_loss: 0.7096 - val_precision: 0.4444 - val_recall: 0.6452\n",
      "Epoch 7/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.6848 - accuracy: 0.6173 - loss: 0.6742 - precision: 0.4618 - recall: 0.6798 - val_AUC: 0.6743 - val_accuracy: 0.6034 - val_loss: 0.6940 - val_precision: 0.4494 - val_recall: 0.6452\n",
      "Epoch 8/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6693 - accuracy: 0.6277 - loss: 0.7015 - precision: 0.5040 - recall: 0.6751 - val_AUC: 0.6838 - val_accuracy: 0.6201 - val_loss: 0.6824 - val_precision: 0.4659 - val_recall: 0.6613\n",
      "Epoch 9/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7041 - accuracy: 0.6273 - loss: 0.6523 - precision: 0.4757 - recall: 0.6917 - val_AUC: 0.6938 - val_accuracy: 0.6201 - val_loss: 0.6700 - val_precision: 0.4659 - val_recall: 0.6613\n",
      "Epoch 10/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7166 - accuracy: 0.6428 - loss: 0.6500 - precision: 0.5093 - recall: 0.7039 - val_AUC: 0.7011 - val_accuracy: 0.6425 - val_loss: 0.6601 - val_precision: 0.4881 - val_recall: 0.6613\n",
      "Epoch 11/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7171 - accuracy: 0.6454 - loss: 0.6421 - precision: 0.5077 - recall: 0.7446 - val_AUC: 0.7087 - val_accuracy: 0.6369 - val_loss: 0.6509 - val_precision: 0.4819 - val_recall: 0.6452\n",
      "Epoch 12/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7236 - accuracy: 0.6612 - loss: 0.6361 - precision: 0.5113 - recall: 0.6760 - val_AUC: 0.7162 - val_accuracy: 0.6425 - val_loss: 0.6431 - val_precision: 0.4881 - val_recall: 0.6613\n",
      "Epoch 13/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7362 - accuracy: 0.6539 - loss: 0.6243 - precision: 0.5099 - recall: 0.7219 - val_AUC: 0.7221 - val_accuracy: 0.6536 - val_loss: 0.6359 - val_precision: 0.5000 - val_recall: 0.6774\n",
      "Epoch 14/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7460 - accuracy: 0.6715 - loss: 0.6122 - precision: 0.5289 - recall: 0.7046 - val_AUC: 0.7297 - val_accuracy: 0.6704 - val_loss: 0.6301 - val_precision: 0.5176 - val_recall: 0.7097\n",
      "Epoch 15/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7583 - accuracy: 0.6870 - loss: 0.5928 - precision: 0.5313 - recall: 0.7163 - val_AUC: 0.7368 - val_accuracy: 0.6704 - val_loss: 0.6251 - val_precision: 0.5176 - val_recall: 0.7097\n",
      "Epoch 16/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7472 - accuracy: 0.6773 - loss: 0.6102 - precision: 0.5388 - recall: 0.7051 - val_AUC: 0.7413 - val_accuracy: 0.6760 - val_loss: 0.6205 - val_precision: 0.5238 - val_recall: 0.7097\n",
      "Epoch 17/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7247 - accuracy: 0.6577 - loss: 0.6314 - precision: 0.5103 - recall: 0.6733 - val_AUC: 0.7457 - val_accuracy: 0.6816 - val_loss: 0.6155 - val_precision: 0.5301 - val_recall: 0.7097\n",
      "Epoch 18/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7399 - accuracy: 0.6805 - loss: 0.6182 - precision: 0.5391 - recall: 0.7001 - val_AUC: 0.7497 - val_accuracy: 0.6872 - val_loss: 0.6123 - val_precision: 0.5366 - val_recall: 0.7097\n",
      "Epoch 19/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7412 - accuracy: 0.6773 - loss: 0.6134 - precision: 0.5307 - recall: 0.6916 - val_AUC: 0.7530 - val_accuracy: 0.6872 - val_loss: 0.6083 - val_precision: 0.5366 - val_recall: 0.7097\n",
      "Epoch 20/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7990 - accuracy: 0.7397 - loss: 0.5532 - precision: 0.6076 - recall: 0.7937 - val_AUC: 0.7571 - val_accuracy: 0.6872 - val_loss: 0.6040 - val_precision: 0.5375 - val_recall: 0.6935\n",
      "Epoch 21/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7806 - accuracy: 0.7038 - loss: 0.5610 - precision: 0.5436 - recall: 0.7317 - val_AUC: 0.7587 - val_accuracy: 0.6816 - val_loss: 0.6007 - val_precision: 0.5309 - val_recall: 0.6935\n",
      "Epoch 22/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7858 - accuracy: 0.7189 - loss: 0.5563 - precision: 0.5582 - recall: 0.7114 - val_AUC: 0.7608 - val_accuracy: 0.6816 - val_loss: 0.5979 - val_precision: 0.5309 - val_recall: 0.6935\n",
      "Epoch 23/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7731 - accuracy: 0.7224 - loss: 0.5731 - precision: 0.5714 - recall: 0.7425 - val_AUC: 0.7641 - val_accuracy: 0.6983 - val_loss: 0.5963 - val_precision: 0.5500 - val_recall: 0.7097\n",
      "Epoch 24/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7823 - accuracy: 0.7284 - loss: 0.5549 - precision: 0.5551 - recall: 0.7321 - val_AUC: 0.7655 - val_accuracy: 0.6983 - val_loss: 0.5938 - val_precision: 0.5500 - val_recall: 0.7097\n",
      "Epoch 25/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8034 - accuracy: 0.7331 - loss: 0.5420 - precision: 0.5966 - recall: 0.7772 - val_AUC: 0.7669 - val_accuracy: 0.7039 - val_loss: 0.5919 - val_precision: 0.5570 - val_recall: 0.7097\n",
      "Epoch 26/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7994 - accuracy: 0.7343 - loss: 0.5531 - precision: 0.6204 - recall: 0.7633 - val_AUC: 0.7703 - val_accuracy: 0.6872 - val_loss: 0.5909 - val_precision: 0.5375 - val_recall: 0.6935\n",
      "Epoch 27/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7860 - accuracy: 0.7160 - loss: 0.5676 - precision: 0.5915 - recall: 0.7401 - val_AUC: 0.7728 - val_accuracy: 0.6983 - val_loss: 0.5884 - val_precision: 0.5513 - val_recall: 0.6935\n",
      "Epoch 28/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7794 - accuracy: 0.7145 - loss: 0.5682 - precision: 0.5719 - recall: 0.7255 - val_AUC: 0.7758 - val_accuracy: 0.6927 - val_loss: 0.5866 - val_precision: 0.5455 - val_recall: 0.6774\n",
      "Epoch 29/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7979 - accuracy: 0.7352 - loss: 0.5530 - precision: 0.6007 - recall: 0.7525 - val_AUC: 0.7775 - val_accuracy: 0.6983 - val_loss: 0.5846 - val_precision: 0.5513 - val_recall: 0.6935\n",
      "Epoch 30/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8299 - accuracy: 0.7623 - loss: 0.5128 - precision: 0.6354 - recall: 0.7884 - val_AUC: 0.7794 - val_accuracy: 0.6983 - val_loss: 0.5844 - val_precision: 0.5513 - val_recall: 0.6935\n",
      "Epoch 31/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8168 - accuracy: 0.7500 - loss: 0.5316 - precision: 0.6332 - recall: 0.7407 - val_AUC: 0.7810 - val_accuracy: 0.6927 - val_loss: 0.5825 - val_precision: 0.5443 - val_recall: 0.6935\n",
      "Epoch 32/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8036 - accuracy: 0.7374 - loss: 0.5492 - precision: 0.6263 - recall: 0.7288 - val_AUC: 0.7826 - val_accuracy: 0.6983 - val_loss: 0.5819 - val_precision: 0.5500 - val_recall: 0.7097\n",
      "Epoch 33/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8115 - accuracy: 0.7485 - loss: 0.5298 - precision: 0.6032 - recall: 0.7494 - val_AUC: 0.7845 - val_accuracy: 0.6983 - val_loss: 0.5796 - val_precision: 0.5500 - val_recall: 0.7097\n",
      "Epoch 34/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8121 - accuracy: 0.7455 - loss: 0.5268 - precision: 0.5991 - recall: 0.7464 - val_AUC: 0.7867 - val_accuracy: 0.7039 - val_loss: 0.5786 - val_precision: 0.5570 - val_recall: 0.7097\n",
      "Epoch 35/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8107 - accuracy: 0.7573 - loss: 0.5315 - precision: 0.6190 - recall: 0.7481 - val_AUC: 0.7875 - val_accuracy: 0.7039 - val_loss: 0.5783 - val_precision: 0.5570 - val_recall: 0.7097\n",
      "Epoch 36/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8411 - accuracy: 0.7684 - loss: 0.4988 - precision: 0.6410 - recall: 0.7913 - val_AUC: 0.7896 - val_accuracy: 0.7039 - val_loss: 0.5775 - val_precision: 0.5570 - val_recall: 0.7097\n",
      "Epoch 37/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8290 - accuracy: 0.7621 - loss: 0.5189 - precision: 0.6442 - recall: 0.7819 - val_AUC: 0.7909 - val_accuracy: 0.6872 - val_loss: 0.5764 - val_precision: 0.5395 - val_recall: 0.6613\n",
      "Epoch 38/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8421 - accuracy: 0.7761 - loss: 0.4980 - precision: 0.6719 - recall: 0.7686 - val_AUC: 0.7917 - val_accuracy: 0.6872 - val_loss: 0.5753 - val_precision: 0.5395 - val_recall: 0.6613\n",
      "Epoch 39/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8309 - accuracy: 0.7825 - loss: 0.5018 - precision: 0.6486 - recall: 0.7797 - val_AUC: 0.7930 - val_accuracy: 0.6872 - val_loss: 0.5739 - val_precision: 0.5395 - val_recall: 0.6613\n",
      "Epoch 40/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8588 - accuracy: 0.7778 - loss: 0.4713 - precision: 0.6460 - recall: 0.8121 - val_AUC: 0.7945 - val_accuracy: 0.6927 - val_loss: 0.5729 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 41/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8440 - accuracy: 0.7768 - loss: 0.4902 - precision: 0.6567 - recall: 0.7795 - val_AUC: 0.7952 - val_accuracy: 0.6927 - val_loss: 0.5726 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 42/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8512 - accuracy: 0.7844 - loss: 0.4849 - precision: 0.6734 - recall: 0.8004 - val_AUC: 0.7957 - val_accuracy: 0.6927 - val_loss: 0.5708 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 43/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8495 - accuracy: 0.7914 - loss: 0.4848 - precision: 0.6811 - recall: 0.7814 - val_AUC: 0.7966 - val_accuracy: 0.6927 - val_loss: 0.5712 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 44/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8555 - accuracy: 0.7787 - loss: 0.4696 - precision: 0.6242 - recall: 0.8219 - val_AUC: 0.7954 - val_accuracy: 0.6927 - val_loss: 0.5686 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 45/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8567 - accuracy: 0.7874 - loss: 0.4729 - precision: 0.6651 - recall: 0.8026 - val_AUC: 0.7958 - val_accuracy: 0.6983 - val_loss: 0.5693 - val_precision: 0.5541 - val_recall: 0.6613\n",
      "Epoch 46/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8344 - accuracy: 0.7789 - loss: 0.4948 - precision: 0.6386 - recall: 0.7703 - val_AUC: 0.7960 - val_accuracy: 0.6983 - val_loss: 0.5690 - val_precision: 0.5541 - val_recall: 0.6613\n",
      "Epoch 47/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8466 - accuracy: 0.7896 - loss: 0.4794 - precision: 0.6430 - recall: 0.8108 - val_AUC: 0.7981 - val_accuracy: 0.6983 - val_loss: 0.5685 - val_precision: 0.5541 - val_recall: 0.6613\n",
      "Epoch 48/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8580 - accuracy: 0.7931 - loss: 0.4796 - precision: 0.6796 - recall: 0.8233 - val_AUC: 0.8001 - val_accuracy: 0.7039 - val_loss: 0.5674 - val_precision: 0.5600 - val_recall: 0.6774\n",
      "Epoch 49/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8609 - accuracy: 0.7836 - loss: 0.4638 - precision: 0.6451 - recall: 0.8053 - val_AUC: 0.8002 - val_accuracy: 0.7151 - val_loss: 0.5667 - val_precision: 0.5714 - val_recall: 0.7097\n",
      "Epoch 50/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8376 - accuracy: 0.7840 - loss: 0.4955 - precision: 0.6522 - recall: 0.7865 - val_AUC: 0.8008 - val_accuracy: 0.7151 - val_loss: 0.5659 - val_precision: 0.5714 - val_recall: 0.7097\n",
      "Epoch 51/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8546 - accuracy: 0.7961 - loss: 0.4700 - precision: 0.6538 - recall: 0.8158 - val_AUC: 0.8014 - val_accuracy: 0.7207 - val_loss: 0.5652 - val_precision: 0.5789 - val_recall: 0.7097\n",
      "Epoch 52/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8679 - accuracy: 0.7997 - loss: 0.4515 - precision: 0.6599 - recall: 0.7971 - val_AUC: 0.8024 - val_accuracy: 0.7151 - val_loss: 0.5663 - val_precision: 0.5714 - val_recall: 0.7097\n",
      "Epoch 53/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8422 - accuracy: 0.7835 - loss: 0.4853 - precision: 0.6426 - recall: 0.7852 - val_AUC: 0.8027 - val_accuracy: 0.7151 - val_loss: 0.5664 - val_precision: 0.5714 - val_recall: 0.7097\n",
      "Epoch 54/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8625 - accuracy: 0.7894 - loss: 0.4674 - precision: 0.6674 - recall: 0.8232 - val_AUC: 0.8027 - val_accuracy: 0.7263 - val_loss: 0.5644 - val_precision: 0.5844 - val_recall: 0.7258\n",
      "Epoch 55/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8589 - accuracy: 0.7942 - loss: 0.4651 - precision: 0.6614 - recall: 0.8158 - val_AUC: 0.8033 - val_accuracy: 0.7207 - val_loss: 0.5641 - val_precision: 0.5769 - val_recall: 0.7258\n",
      "Epoch 56/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8610 - accuracy: 0.7903 - loss: 0.4660 - precision: 0.6686 - recall: 0.7970 - val_AUC: 0.8026 - val_accuracy: 0.7263 - val_loss: 0.5637 - val_precision: 0.5844 - val_recall: 0.7258\n",
      "Epoch 57/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8745 - accuracy: 0.7925 - loss: 0.4511 - precision: 0.6717 - recall: 0.8357 - val_AUC: 0.8036 - val_accuracy: 0.7374 - val_loss: 0.5641 - val_precision: 0.5949 - val_recall: 0.7581\n",
      "Epoch 58/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8733 - accuracy: 0.7951 - loss: 0.4476 - precision: 0.6678 - recall: 0.8235 - val_AUC: 0.8026 - val_accuracy: 0.7430 - val_loss: 0.5635 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 59/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8868 - accuracy: 0.8024 - loss: 0.4225 - precision: 0.6458 - recall: 0.8319 - val_AUC: 0.8037 - val_accuracy: 0.7374 - val_loss: 0.5607 - val_precision: 0.5949 - val_recall: 0.7581\n",
      "Epoch 60/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8780 - accuracy: 0.7901 - loss: 0.4416 - precision: 0.6465 - recall: 0.8420 - val_AUC: 0.8035 - val_accuracy: 0.7430 - val_loss: 0.5623 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 61/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8530 - accuracy: 0.7682 - loss: 0.4771 - precision: 0.6410 - recall: 0.7879 - val_AUC: 0.8042 - val_accuracy: 0.7430 - val_loss: 0.5635 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 62/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8509 - accuracy: 0.7887 - loss: 0.4829 - precision: 0.6590 - recall: 0.8051 - val_AUC: 0.8058 - val_accuracy: 0.7430 - val_loss: 0.5651 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 63/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8716 - accuracy: 0.7876 - loss: 0.4522 - precision: 0.6575 - recall: 0.8305 - val_AUC: 0.8066 - val_accuracy: 0.7430 - val_loss: 0.5636 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 64/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8798 - accuracy: 0.8003 - loss: 0.4440 - precision: 0.6906 - recall: 0.7768 - val_AUC: 0.8067 - val_accuracy: 0.7430 - val_loss: 0.5636 - val_precision: 0.6000 - val_recall: 0.7742\n",
      "Epoch 65/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8760 - accuracy: 0.7929 - loss: 0.4466 - precision: 0.6634 - recall: 0.8296 - val_AUC: 0.8075 - val_accuracy: 0.7486 - val_loss: 0.5613 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 66/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8581 - accuracy: 0.7919 - loss: 0.4777 - precision: 0.6873 - recall: 0.7862 - val_AUC: 0.8076 - val_accuracy: 0.7486 - val_loss: 0.5615 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 67/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8873 - accuracy: 0.8283 - loss: 0.4244 - precision: 0.7181 - recall: 0.8380 - val_AUC: 0.8085 - val_accuracy: 0.7542 - val_loss: 0.5611 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 68/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8761 - accuracy: 0.8128 - loss: 0.4471 - precision: 0.7022 - recall: 0.8307 - val_AUC: 0.8093 - val_accuracy: 0.7542 - val_loss: 0.5609 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 69/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8849 - accuracy: 0.8087 - loss: 0.4284 - precision: 0.6758 - recall: 0.8180 - val_AUC: 0.8087 - val_accuracy: 0.7542 - val_loss: 0.5601 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 70/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8903 - accuracy: 0.8184 - loss: 0.4295 - precision: 0.7118 - recall: 0.8366 - val_AUC: 0.8086 - val_accuracy: 0.7542 - val_loss: 0.5610 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 71/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8913 - accuracy: 0.8214 - loss: 0.4222 - precision: 0.7006 - recall: 0.8429 - val_AUC: 0.8080 - val_accuracy: 0.7598 - val_loss: 0.5602 - val_precision: 0.6267 - val_recall: 0.7581\n",
      "Epoch 72/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8934 - accuracy: 0.8204 - loss: 0.4223 - precision: 0.7222 - recall: 0.8444 - val_AUC: 0.8080 - val_accuracy: 0.7598 - val_loss: 0.5616 - val_precision: 0.6267 - val_recall: 0.7581\n",
      "Epoch 73/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8928 - accuracy: 0.8268 - loss: 0.4149 - precision: 0.6886 - recall: 0.8487 - val_AUC: 0.8073 - val_accuracy: 0.7598 - val_loss: 0.5597 - val_precision: 0.6267 - val_recall: 0.7581\n",
      "Epoch 74/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8772 - accuracy: 0.7904 - loss: 0.4416 - precision: 0.6640 - recall: 0.7994 - val_AUC: 0.8085 - val_accuracy: 0.7598 - val_loss: 0.5598 - val_precision: 0.6267 - val_recall: 0.7581\n",
      "Epoch 75/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8821 - accuracy: 0.8082 - loss: 0.4337 - precision: 0.6815 - recall: 0.8131 - val_AUC: 0.8081 - val_accuracy: 0.7654 - val_loss: 0.5595 - val_precision: 0.6351 - val_recall: 0.7581\n",
      "Epoch 76/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8845 - accuracy: 0.8232 - loss: 0.4279 - precision: 0.7030 - recall: 0.8236 - val_AUC: 0.8078 - val_accuracy: 0.7654 - val_loss: 0.5624 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 77/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8804 - accuracy: 0.8257 - loss: 0.4344 - precision: 0.7149 - recall: 0.8205 - val_AUC: 0.8085 - val_accuracy: 0.7654 - val_loss: 0.5623 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 78/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8941 - accuracy: 0.8344 - loss: 0.4209 - precision: 0.7511 - recall: 0.8193 - val_AUC: 0.8078 - val_accuracy: 0.7654 - val_loss: 0.5639 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 79/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8976 - accuracy: 0.8141 - loss: 0.4088 - precision: 0.6812 - recall: 0.8475 - val_AUC: 0.8089 - val_accuracy: 0.7709 - val_loss: 0.5624 - val_precision: 0.6400 - val_recall: 0.7742\n",
      "Epoch 80/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8900 - accuracy: 0.8189 - loss: 0.4170 - precision: 0.6896 - recall: 0.8577 - val_AUC: 0.8083 - val_accuracy: 0.7709 - val_loss: 0.5624 - val_precision: 0.6400 - val_recall: 0.7742\n",
      "Epoch 81/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9026 - accuracy: 0.8315 - loss: 0.4059 - precision: 0.7218 - recall: 0.8558 - val_AUC: 0.8082 - val_accuracy: 0.7654 - val_loss: 0.5630 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 82/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8962 - accuracy: 0.8437 - loss: 0.4093 - precision: 0.7408 - recall: 0.8535 - val_AUC: 0.8078 - val_accuracy: 0.7598 - val_loss: 0.5656 - val_precision: 0.6234 - val_recall: 0.7742\n",
      "Epoch 83/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9082 - accuracy: 0.8438 - loss: 0.4007 - precision: 0.7423 - recall: 0.8613 - val_AUC: 0.8066 - val_accuracy: 0.7654 - val_loss: 0.5655 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 84/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9056 - accuracy: 0.8339 - loss: 0.3938 - precision: 0.7095 - recall: 0.8424 - val_AUC: 0.8068 - val_accuracy: 0.7598 - val_loss: 0.5646 - val_precision: 0.6234 - val_recall: 0.7742\n",
      "Epoch 85/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8946 - accuracy: 0.8263 - loss: 0.4150 - precision: 0.7169 - recall: 0.8226 - val_AUC: 0.8065 - val_accuracy: 0.7598 - val_loss: 0.5649 - val_precision: 0.6234 - val_recall: 0.7742\n",
      "Epoch 86/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9065 - accuracy: 0.8461 - loss: 0.3895 - precision: 0.7103 - recall: 0.8738 - val_AUC: 0.8053 - val_accuracy: 0.7542 - val_loss: 0.5657 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 87/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9099 - accuracy: 0.8511 - loss: 0.3864 - precision: 0.7523 - recall: 0.8532 - val_AUC: 0.8071 - val_accuracy: 0.7542 - val_loss: 0.5686 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 88/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9092 - accuracy: 0.8420 - loss: 0.4000 - precision: 0.7603 - recall: 0.8603 - val_AUC: 0.8074 - val_accuracy: 0.7486 - val_loss: 0.5686 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 89/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8970 - accuracy: 0.8425 - loss: 0.4105 - precision: 0.7542 - recall: 0.8257 - val_AUC: 0.8062 - val_accuracy: 0.7486 - val_loss: 0.5676 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 90/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9174 - accuracy: 0.8489 - loss: 0.3686 - precision: 0.7292 - recall: 0.8600 - val_AUC: 0.8083 - val_accuracy: 0.7486 - val_loss: 0.5684 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 91/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9000 - accuracy: 0.8406 - loss: 0.4004 - precision: 0.7335 - recall: 0.8561 - val_AUC: 0.8055 - val_accuracy: 0.7542 - val_loss: 0.5702 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 92/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8931 - accuracy: 0.8256 - loss: 0.4176 - precision: 0.7117 - recall: 0.8569 - val_AUC: 0.8059 - val_accuracy: 0.7542 - val_loss: 0.5706 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 93/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8984 - accuracy: 0.8374 - loss: 0.4124 - precision: 0.7480 - recall: 0.8333 - val_AUC: 0.8048 - val_accuracy: 0.7542 - val_loss: 0.5710 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 94/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9219 - accuracy: 0.8532 - loss: 0.3626 - precision: 0.7347 - recall: 0.8648 - val_AUC: 0.8058 - val_accuracy: 0.7542 - val_loss: 0.5707 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 95/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9194 - accuracy: 0.8580 - loss: 0.3684 - precision: 0.7529 - recall: 0.8730 - val_AUC: 0.8053 - val_accuracy: 0.7542 - val_loss: 0.5723 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 96/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9177 - accuracy: 0.8517 - loss: 0.3771 - precision: 0.7575 - recall: 0.8480 - val_AUC: 0.8047 - val_accuracy: 0.7542 - val_loss: 0.5745 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 97/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9172 - accuracy: 0.8463 - loss: 0.3701 - precision: 0.7325 - recall: 0.8855 - val_AUC: 0.8040 - val_accuracy: 0.7542 - val_loss: 0.5731 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 98/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9244 - accuracy: 0.8640 - loss: 0.3660 - precision: 0.7770 - recall: 0.8758 - val_AUC: 0.8043 - val_accuracy: 0.7486 - val_loss: 0.5760 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 99/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9211 - accuracy: 0.8631 - loss: 0.3559 - precision: 0.7544 - recall: 0.8784 - val_AUC: 0.8034 - val_accuracy: 0.7542 - val_loss: 0.5742 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 100/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9236 - accuracy: 0.8499 - loss: 0.3617 - precision: 0.7442 - recall: 0.8746 - val_AUC: 0.8042 - val_accuracy: 0.7542 - val_loss: 0.5785 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 101/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9117 - accuracy: 0.8349 - loss: 0.3851 - precision: 0.7334 - recall: 0.8429 - val_AUC: 0.8036 - val_accuracy: 0.7486 - val_loss: 0.5798 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 102/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9300 - accuracy: 0.8689 - loss: 0.3441 - precision: 0.7686 - recall: 0.8829 - val_AUC: 0.8024 - val_accuracy: 0.7486 - val_loss: 0.5783 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 103/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9336 - accuracy: 0.8641 - loss: 0.3355 - precision: 0.7434 - recall: 0.9163 - val_AUC: 0.8029 - val_accuracy: 0.7486 - val_loss: 0.5783 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 104/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9319 - accuracy: 0.8553 - loss: 0.3484 - precision: 0.7324 - recall: 0.8975 - val_AUC: 0.8029 - val_accuracy: 0.7542 - val_loss: 0.5809 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 105/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9330 - accuracy: 0.8656 - loss: 0.3400 - precision: 0.7825 - recall: 0.8595 - val_AUC: 0.8008 - val_accuracy: 0.7430 - val_loss: 0.5837 - val_precision: 0.6026 - val_recall: 0.7581\n",
      "Epoch 106/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9319 - accuracy: 0.8792 - loss: 0.3405 - precision: 0.7969 - recall: 0.8792 - val_AUC: 0.8018 - val_accuracy: 0.7486 - val_loss: 0.5831 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 107/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9367 - accuracy: 0.8690 - loss: 0.3333 - precision: 0.7661 - recall: 0.8998 - val_AUC: 0.8009 - val_accuracy: 0.7486 - val_loss: 0.5819 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 108/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9308 - accuracy: 0.8604 - loss: 0.3477 - precision: 0.7468 - recall: 0.8964 - val_AUC: 0.8007 - val_accuracy: 0.7542 - val_loss: 0.5830 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 109/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9324 - accuracy: 0.8656 - loss: 0.3381 - precision: 0.7633 - recall: 0.8778 - val_AUC: 0.7998 - val_accuracy: 0.7542 - val_loss: 0.5865 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 110/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9275 - accuracy: 0.8633 - loss: 0.3497 - precision: 0.7668 - recall: 0.8594 - val_AUC: 0.7993 - val_accuracy: 0.7486 - val_loss: 0.5880 - val_precision: 0.6104 - val_recall: 0.7581\n",
      "Epoch 111/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9245 - accuracy: 0.8623 - loss: 0.3566 - precision: 0.7725 - recall: 0.8447 - val_AUC: 0.7998 - val_accuracy: 0.7542 - val_loss: 0.5874 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 112/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9276 - accuracy: 0.8595 - loss: 0.3512 - precision: 0.7651 - recall: 0.8786 - val_AUC: 0.7992 - val_accuracy: 0.7542 - val_loss: 0.5876 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 113/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9391 - accuracy: 0.8664 - loss: 0.3274 - precision: 0.7736 - recall: 0.8658 - val_AUC: 0.7991 - val_accuracy: 0.7542 - val_loss: 0.5907 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 114/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9517 - accuracy: 0.8803 - loss: 0.3048 - precision: 0.7806 - recall: 0.8970 - val_AUC: 0.7982 - val_accuracy: 0.7542 - val_loss: 0.5937 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 115/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9422 - accuracy: 0.8632 - loss: 0.3240 - precision: 0.7522 - recall: 0.8801 - val_AUC: 0.7965 - val_accuracy: 0.7542 - val_loss: 0.5940 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 116/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9266 - accuracy: 0.8522 - loss: 0.3397 - precision: 0.7312 - recall: 0.8393 - val_AUC: 0.7971 - val_accuracy: 0.7542 - val_loss: 0.5939 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 117/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9368 - accuracy: 0.8748 - loss: 0.3305 - precision: 0.7824 - recall: 0.8884 - val_AUC: 0.7953 - val_accuracy: 0.7542 - val_loss: 0.5965 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 118/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9515 - accuracy: 0.8872 - loss: 0.3040 - precision: 0.7983 - recall: 0.9109 - val_AUC: 0.7959 - val_accuracy: 0.7486 - val_loss: 0.5985 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 119/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9369 - accuracy: 0.8713 - loss: 0.3346 - precision: 0.7893 - recall: 0.8870 - val_AUC: 0.7964 - val_accuracy: 0.7486 - val_loss: 0.5975 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 120/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9270 - accuracy: 0.8619 - loss: 0.3407 - precision: 0.7549 - recall: 0.8646 - val_AUC: 0.7947 - val_accuracy: 0.7542 - val_loss: 0.5983 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 121/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9433 - accuracy: 0.8849 - loss: 0.3130 - precision: 0.8071 - recall: 0.8880 - val_AUC: 0.7945 - val_accuracy: 0.7486 - val_loss: 0.6020 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 122/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9482 - accuracy: 0.8793 - loss: 0.3030 - precision: 0.7689 - recall: 0.9032 - val_AUC: 0.7947 - val_accuracy: 0.7486 - val_loss: 0.6007 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 123/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9344 - accuracy: 0.8779 - loss: 0.3323 - precision: 0.8047 - recall: 0.8730 - val_AUC: 0.7943 - val_accuracy: 0.7486 - val_loss: 0.6037 - val_precision: 0.6076 - val_recall: 0.7742\n",
      "Epoch 124/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9417 - accuracy: 0.8753 - loss: 0.3224 - precision: 0.7886 - recall: 0.8805 - val_AUC: 0.7938 - val_accuracy: 0.7542 - val_loss: 0.6031 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 125/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9391 - accuracy: 0.8907 - loss: 0.3279 - precision: 0.8529 - recall: 0.8589 - val_AUC: 0.7932 - val_accuracy: 0.7542 - val_loss: 0.6060 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "\u001b[1m8/8\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - AUC: 0.7831 - accuracy: 0.6989 - loss: 0.6412 - precision: 0.6907 - recall: 0.3425 \n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 21ms/step - AUC: 0.6886 - accuracy: 0.6520 - loss: 0.7344 - precision: 0.5309 - recall: 0.3526 - val_AUC: 0.6929 - val_accuracy: 0.6816 - val_loss: 0.6190 - val_precision: 0.5581 - val_recall: 0.3871\n",
      "Epoch 2/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.6847 - accuracy: 0.6727 - loss: 0.7188 - precision: 0.5436 - recall: 0.3825 - val_AUC: 0.7120 - val_accuracy: 0.6816 - val_loss: 0.6018 - val_precision: 0.5556 - val_recall: 0.4032\n",
      "Epoch 3/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7252 - accuracy: 0.6986 - loss: 0.6207 - precision: 0.5236 - recall: 0.4952 - val_AUC: 0.7247 - val_accuracy: 0.6872 - val_loss: 0.5893 - val_precision: 0.5625 - val_recall: 0.4355\n",
      "Epoch 4/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7437 - accuracy: 0.7176 - loss: 0.6520 - precision: 0.6347 - recall: 0.5438 - val_AUC: 0.7359 - val_accuracy: 0.6927 - val_loss: 0.5786 - val_precision: 0.5660 - val_recall: 0.4839\n",
      "Epoch 5/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7220 - accuracy: 0.6957 - loss: 0.6067 - precision: 0.4921 - recall: 0.5669 - val_AUC: 0.7466 - val_accuracy: 0.6983 - val_loss: 0.5701 - val_precision: 0.5714 - val_recall: 0.5161\n",
      "Epoch 6/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7440 - accuracy: 0.6983 - loss: 0.6262 - precision: 0.5723 - recall: 0.5808 - val_AUC: 0.7559 - val_accuracy: 0.7039 - val_loss: 0.5637 - val_precision: 0.5738 - val_recall: 0.5645\n",
      "Epoch 7/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7517 - accuracy: 0.7001 - loss: 0.6038 - precision: 0.5625 - recall: 0.6262 - val_AUC: 0.7631 - val_accuracy: 0.7207 - val_loss: 0.5578 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 8/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7437 - accuracy: 0.6856 - loss: 0.6293 - precision: 0.5747 - recall: 0.6118 - val_AUC: 0.7711 - val_accuracy: 0.7151 - val_loss: 0.5533 - val_precision: 0.5902 - val_recall: 0.5806\n",
      "Epoch 9/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7609 - accuracy: 0.7041 - loss: 0.5957 - precision: 0.5868 - recall: 0.6821 - val_AUC: 0.7783 - val_accuracy: 0.7263 - val_loss: 0.5492 - val_precision: 0.6032 - val_recall: 0.6129\n",
      "Epoch 10/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7825 - accuracy: 0.7295 - loss: 0.5710 - precision: 0.5902 - recall: 0.7214 - val_AUC: 0.7823 - val_accuracy: 0.7598 - val_loss: 0.5460 - val_precision: 0.6377 - val_recall: 0.7097\n",
      "Epoch 11/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7832 - accuracy: 0.7269 - loss: 0.5680 - precision: 0.5935 - recall: 0.7254 - val_AUC: 0.7874 - val_accuracy: 0.7542 - val_loss: 0.5426 - val_precision: 0.6286 - val_recall: 0.7097\n",
      "Epoch 12/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7924 - accuracy: 0.7243 - loss: 0.5460 - precision: 0.5808 - recall: 0.7607 - val_AUC: 0.7909 - val_accuracy: 0.7486 - val_loss: 0.5395 - val_precision: 0.6164 - val_recall: 0.7258\n",
      "Epoch 13/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7750 - accuracy: 0.7099 - loss: 0.5749 - precision: 0.5718 - recall: 0.7440 - val_AUC: 0.7949 - val_accuracy: 0.7486 - val_loss: 0.5379 - val_precision: 0.6164 - val_recall: 0.7258\n",
      "Epoch 14/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7947 - accuracy: 0.7190 - loss: 0.5464 - precision: 0.5604 - recall: 0.7699 - val_AUC: 0.7972 - val_accuracy: 0.7542 - val_loss: 0.5352 - val_precision: 0.6216 - val_recall: 0.7419\n",
      "Epoch 15/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7846 - accuracy: 0.7096 - loss: 0.5566 - precision: 0.5480 - recall: 0.7260 - val_AUC: 0.8004 - val_accuracy: 0.7654 - val_loss: 0.5328 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 16/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7954 - accuracy: 0.7329 - loss: 0.5500 - precision: 0.5873 - recall: 0.7341 - val_AUC: 0.8033 - val_accuracy: 0.7654 - val_loss: 0.5305 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 17/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7938 - accuracy: 0.7233 - loss: 0.5314 - precision: 0.5418 - recall: 0.7679 - val_AUC: 0.8065 - val_accuracy: 0.7654 - val_loss: 0.5284 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 18/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8142 - accuracy: 0.7311 - loss: 0.5397 - precision: 0.6182 - recall: 0.7448 - val_AUC: 0.8075 - val_accuracy: 0.7709 - val_loss: 0.5268 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 19/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7983 - accuracy: 0.7269 - loss: 0.5446 - precision: 0.5790 - recall: 0.7745 - val_AUC: 0.8100 - val_accuracy: 0.7709 - val_loss: 0.5247 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 20/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8266 - accuracy: 0.7231 - loss: 0.5164 - precision: 0.5741 - recall: 0.7947 - val_AUC: 0.8114 - val_accuracy: 0.7709 - val_loss: 0.5221 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 21/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8264 - accuracy: 0.7346 - loss: 0.5097 - precision: 0.5788 - recall: 0.7696 - val_AUC: 0.8127 - val_accuracy: 0.7709 - val_loss: 0.5211 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 22/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7951 - accuracy: 0.7275 - loss: 0.5490 - precision: 0.5846 - recall: 0.7779 - val_AUC: 0.8144 - val_accuracy: 0.7709 - val_loss: 0.5194 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 23/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8183 - accuracy: 0.7136 - loss: 0.5333 - precision: 0.5848 - recall: 0.8009 - val_AUC: 0.8158 - val_accuracy: 0.7654 - val_loss: 0.5174 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 24/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7925 - accuracy: 0.6917 - loss: 0.5542 - precision: 0.5502 - recall: 0.7126 - val_AUC: 0.8175 - val_accuracy: 0.7654 - val_loss: 0.5155 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 25/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8177 - accuracy: 0.7231 - loss: 0.5276 - precision: 0.5876 - recall: 0.7764 - val_AUC: 0.8179 - val_accuracy: 0.7654 - val_loss: 0.5150 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 26/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7920 - accuracy: 0.7115 - loss: 0.5646 - precision: 0.5699 - recall: 0.7421 - val_AUC: 0.8182 - val_accuracy: 0.7654 - val_loss: 0.5135 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 27/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8362 - accuracy: 0.7387 - loss: 0.4992 - precision: 0.5858 - recall: 0.8013 - val_AUC: 0.8198 - val_accuracy: 0.7654 - val_loss: 0.5118 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 28/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7940 - accuracy: 0.7200 - loss: 0.5574 - precision: 0.5929 - recall: 0.7880 - val_AUC: 0.8213 - val_accuracy: 0.7654 - val_loss: 0.5109 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 29/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8186 - accuracy: 0.7248 - loss: 0.5212 - precision: 0.5701 - recall: 0.7639 - val_AUC: 0.8224 - val_accuracy: 0.7654 - val_loss: 0.5092 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 30/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8313 - accuracy: 0.7507 - loss: 0.4994 - precision: 0.5938 - recall: 0.7748 - val_AUC: 0.8237 - val_accuracy: 0.7709 - val_loss: 0.5079 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 31/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8366 - accuracy: 0.7499 - loss: 0.4958 - precision: 0.5862 - recall: 0.8283 - val_AUC: 0.8251 - val_accuracy: 0.7765 - val_loss: 0.5062 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 32/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8313 - accuracy: 0.7395 - loss: 0.5118 - precision: 0.6006 - recall: 0.7910 - val_AUC: 0.8264 - val_accuracy: 0.7821 - val_loss: 0.5056 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 33/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8308 - accuracy: 0.7460 - loss: 0.5158 - precision: 0.6219 - recall: 0.8051 - val_AUC: 0.8276 - val_accuracy: 0.7821 - val_loss: 0.5043 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 34/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8253 - accuracy: 0.7532 - loss: 0.5124 - precision: 0.5993 - recall: 0.8031 - val_AUC: 0.8285 - val_accuracy: 0.7765 - val_loss: 0.5027 - val_precision: 0.6447 - val_recall: 0.7903\n",
      "Epoch 35/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8416 - accuracy: 0.7549 - loss: 0.4894 - precision: 0.5935 - recall: 0.8079 - val_AUC: 0.8295 - val_accuracy: 0.7821 - val_loss: 0.5016 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 36/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8316 - accuracy: 0.7540 - loss: 0.5035 - precision: 0.6054 - recall: 0.7887 - val_AUC: 0.8297 - val_accuracy: 0.7821 - val_loss: 0.5016 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 37/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8576 - accuracy: 0.7747 - loss: 0.4770 - precision: 0.6447 - recall: 0.8099 - val_AUC: 0.8307 - val_accuracy: 0.7877 - val_loss: 0.5016 - val_precision: 0.6538 - val_recall: 0.8226\n",
      "Epoch 38/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8562 - accuracy: 0.7761 - loss: 0.4766 - precision: 0.6267 - recall: 0.8347 - val_AUC: 0.8312 - val_accuracy: 0.7821 - val_loss: 0.4998 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 39/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8221 - accuracy: 0.7577 - loss: 0.5045 - precision: 0.5833 - recall: 0.7737 - val_AUC: 0.8313 - val_accuracy: 0.7877 - val_loss: 0.4985 - val_precision: 0.6579 - val_recall: 0.8065\n",
      "Epoch 40/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8284 - accuracy: 0.7613 - loss: 0.5065 - precision: 0.6130 - recall: 0.7679 - val_AUC: 0.8321 - val_accuracy: 0.7877 - val_loss: 0.4978 - val_precision: 0.6579 - val_recall: 0.8065\n",
      "Epoch 41/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8442 - accuracy: 0.7528 - loss: 0.4989 - precision: 0.6232 - recall: 0.8158 - val_AUC: 0.8329 - val_accuracy: 0.7877 - val_loss: 0.4973 - val_precision: 0.6579 - val_recall: 0.8065\n",
      "Epoch 42/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8505 - accuracy: 0.7594 - loss: 0.4964 - precision: 0.6450 - recall: 0.8125 - val_AUC: 0.8333 - val_accuracy: 0.7821 - val_loss: 0.4964 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 43/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8571 - accuracy: 0.7654 - loss: 0.4799 - precision: 0.6224 - recall: 0.8271 - val_AUC: 0.8336 - val_accuracy: 0.7821 - val_loss: 0.4954 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 44/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8694 - accuracy: 0.7834 - loss: 0.4539 - precision: 0.6316 - recall: 0.8446 - val_AUC: 0.8344 - val_accuracy: 0.7821 - val_loss: 0.4944 - val_precision: 0.6494 - val_recall: 0.8065\n",
      "Epoch 45/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8418 - accuracy: 0.7643 - loss: 0.4905 - precision: 0.6195 - recall: 0.7798 - val_AUC: 0.8346 - val_accuracy: 0.7765 - val_loss: 0.4942 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 46/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8526 - accuracy: 0.7746 - loss: 0.4833 - precision: 0.6555 - recall: 0.8002 - val_AUC: 0.8364 - val_accuracy: 0.7765 - val_loss: 0.4941 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 47/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8630 - accuracy: 0.7729 - loss: 0.4654 - precision: 0.6099 - recall: 0.8417 - val_AUC: 0.8362 - val_accuracy: 0.7821 - val_loss: 0.4929 - val_precision: 0.6456 - val_recall: 0.8226\n",
      "Epoch 48/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8639 - accuracy: 0.7865 - loss: 0.4584 - precision: 0.6341 - recall: 0.8379 - val_AUC: 0.8371 - val_accuracy: 0.7765 - val_loss: 0.4920 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 49/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8526 - accuracy: 0.7731 - loss: 0.4802 - precision: 0.6373 - recall: 0.8311 - val_AUC: 0.8368 - val_accuracy: 0.7765 - val_loss: 0.4917 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 50/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8582 - accuracy: 0.7848 - loss: 0.4728 - precision: 0.6638 - recall: 0.8026 - val_AUC: 0.8372 - val_accuracy: 0.7709 - val_loss: 0.4916 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 51/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8575 - accuracy: 0.7685 - loss: 0.4708 - precision: 0.6249 - recall: 0.8155 - val_AUC: 0.8368 - val_accuracy: 0.7765 - val_loss: 0.4914 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 52/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8746 - accuracy: 0.8110 - loss: 0.4510 - precision: 0.6958 - recall: 0.8115 - val_AUC: 0.8376 - val_accuracy: 0.7765 - val_loss: 0.4907 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 53/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8653 - accuracy: 0.7903 - loss: 0.4657 - precision: 0.6735 - recall: 0.8077 - val_AUC: 0.8372 - val_accuracy: 0.7709 - val_loss: 0.4902 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 54/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8638 - accuracy: 0.7959 - loss: 0.4726 - precision: 0.6942 - recall: 0.8134 - val_AUC: 0.8368 - val_accuracy: 0.7765 - val_loss: 0.4901 - val_precision: 0.6410 - val_recall: 0.8065\n",
      "Epoch 55/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8686 - accuracy: 0.7917 - loss: 0.4547 - precision: 0.6406 - recall: 0.8251 - val_AUC: 0.8364 - val_accuracy: 0.7709 - val_loss: 0.4894 - val_precision: 0.6364 - val_recall: 0.7903\n",
      "Epoch 56/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8752 - accuracy: 0.7912 - loss: 0.4522 - precision: 0.6722 - recall: 0.8183 - val_AUC: 0.8355 - val_accuracy: 0.7598 - val_loss: 0.4900 - val_precision: 0.6203 - val_recall: 0.7903\n",
      "Epoch 57/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8765 - accuracy: 0.8078 - loss: 0.4420 - precision: 0.6888 - recall: 0.8161 - val_AUC: 0.8368 - val_accuracy: 0.7598 - val_loss: 0.4891 - val_precision: 0.6203 - val_recall: 0.7903\n",
      "Epoch 58/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8821 - accuracy: 0.8028 - loss: 0.4388 - precision: 0.6922 - recall: 0.8106 - val_AUC: 0.8353 - val_accuracy: 0.7598 - val_loss: 0.4888 - val_precision: 0.6234 - val_recall: 0.7742\n",
      "Epoch 59/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8747 - accuracy: 0.7967 - loss: 0.4506 - precision: 0.6669 - recall: 0.8182 - val_AUC: 0.8357 - val_accuracy: 0.7598 - val_loss: 0.4879 - val_precision: 0.6234 - val_recall: 0.7742\n",
      "Epoch 60/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8555 - accuracy: 0.7683 - loss: 0.4739 - precision: 0.6392 - recall: 0.7942 - val_AUC: 0.8364 - val_accuracy: 0.7654 - val_loss: 0.4878 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 61/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8631 - accuracy: 0.7788 - loss: 0.4647 - precision: 0.6631 - recall: 0.7981 - val_AUC: 0.8363 - val_accuracy: 0.7654 - val_loss: 0.4878 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 62/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8707 - accuracy: 0.7881 - loss: 0.4537 - precision: 0.6642 - recall: 0.7973 - val_AUC: 0.8353 - val_accuracy: 0.7598 - val_loss: 0.4877 - val_precision: 0.6203 - val_recall: 0.7903\n",
      "Epoch 63/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8882 - accuracy: 0.8056 - loss: 0.4257 - precision: 0.6760 - recall: 0.8435 - val_AUC: 0.8354 - val_accuracy: 0.7542 - val_loss: 0.4871 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 64/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8724 - accuracy: 0.7860 - loss: 0.4464 - precision: 0.6412 - recall: 0.8182 - val_AUC: 0.8362 - val_accuracy: 0.7542 - val_loss: 0.4861 - val_precision: 0.6154 - val_recall: 0.7742\n",
      "Epoch 65/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8931 - accuracy: 0.8075 - loss: 0.4244 - precision: 0.6919 - recall: 0.8429 - val_AUC: 0.8364 - val_accuracy: 0.7654 - val_loss: 0.4873 - val_precision: 0.6282 - val_recall: 0.7903\n",
      "Epoch 66/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8726 - accuracy: 0.7979 - loss: 0.4447 - precision: 0.6544 - recall: 0.8353 - val_AUC: 0.8358 - val_accuracy: 0.7542 - val_loss: 0.4872 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 67/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8733 - accuracy: 0.8110 - loss: 0.4460 - precision: 0.6855 - recall: 0.8478 - val_AUC: 0.8357 - val_accuracy: 0.7542 - val_loss: 0.4860 - val_precision: 0.6184 - val_recall: 0.7581\n",
      "Epoch 68/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8721 - accuracy: 0.7921 - loss: 0.4409 - precision: 0.6377 - recall: 0.8309 - val_AUC: 0.8359 - val_accuracy: 0.7598 - val_loss: 0.4862 - val_precision: 0.6267 - val_recall: 0.7581\n",
      "Epoch 69/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8766 - accuracy: 0.7981 - loss: 0.4410 - precision: 0.6717 - recall: 0.8195 - val_AUC: 0.8362 - val_accuracy: 0.7654 - val_loss: 0.4868 - val_precision: 0.6316 - val_recall: 0.7742\n",
      "Epoch 70/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8789 - accuracy: 0.8064 - loss: 0.4403 - precision: 0.6903 - recall: 0.8031 - val_AUC: 0.8364 - val_accuracy: 0.7486 - val_loss: 0.4863 - val_precision: 0.6133 - val_recall: 0.7419\n",
      "Epoch 71/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8917 - accuracy: 0.8211 - loss: 0.4195 - precision: 0.6888 - recall: 0.8376 - val_AUC: 0.8362 - val_accuracy: 0.7709 - val_loss: 0.4854 - val_precision: 0.6438 - val_recall: 0.7581\n",
      "Epoch 72/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8673 - accuracy: 0.7927 - loss: 0.4585 - precision: 0.6796 - recall: 0.7866 - val_AUC: 0.8367 - val_accuracy: 0.7654 - val_loss: 0.4859 - val_precision: 0.6389 - val_recall: 0.7419\n",
      "Epoch 73/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8981 - accuracy: 0.8159 - loss: 0.4129 - precision: 0.7056 - recall: 0.8199 - val_AUC: 0.8376 - val_accuracy: 0.7654 - val_loss: 0.4860 - val_precision: 0.6389 - val_recall: 0.7419\n",
      "Epoch 74/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8875 - accuracy: 0.8198 - loss: 0.4251 - precision: 0.7051 - recall: 0.8320 - val_AUC: 0.8368 - val_accuracy: 0.7598 - val_loss: 0.4854 - val_precision: 0.6338 - val_recall: 0.7258\n",
      "Epoch 75/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8841 - accuracy: 0.7968 - loss: 0.4326 - precision: 0.6825 - recall: 0.7996 - val_AUC: 0.8367 - val_accuracy: 0.7598 - val_loss: 0.4857 - val_precision: 0.6338 - val_recall: 0.7258\n",
      "Epoch 76/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8816 - accuracy: 0.7960 - loss: 0.4325 - precision: 0.6551 - recall: 0.8293 - val_AUC: 0.8371 - val_accuracy: 0.7598 - val_loss: 0.4847 - val_precision: 0.6338 - val_recall: 0.7258\n",
      "Epoch 77/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9131 - accuracy: 0.8349 - loss: 0.3830 - precision: 0.6908 - recall: 0.8667 - val_AUC: 0.8373 - val_accuracy: 0.7598 - val_loss: 0.4848 - val_precision: 0.6338 - val_recall: 0.7258\n",
      "Epoch 78/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8679 - accuracy: 0.7954 - loss: 0.4464 - precision: 0.6550 - recall: 0.8003 - val_AUC: 0.8377 - val_accuracy: 0.7598 - val_loss: 0.4855 - val_precision: 0.6338 - val_recall: 0.7258\n",
      "Epoch 79/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8976 - accuracy: 0.8167 - loss: 0.4069 - precision: 0.6829 - recall: 0.8420 - val_AUC: 0.8363 - val_accuracy: 0.7654 - val_loss: 0.4847 - val_precision: 0.6429 - val_recall: 0.7258\n",
      "Epoch 80/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9116 - accuracy: 0.8428 - loss: 0.3874 - precision: 0.7055 - recall: 0.8992 - val_AUC: 0.8380 - val_accuracy: 0.7654 - val_loss: 0.4857 - val_precision: 0.6389 - val_recall: 0.7419\n",
      "Epoch 81/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9000 - accuracy: 0.8314 - loss: 0.4038 - precision: 0.7065 - recall: 0.8670 - val_AUC: 0.8370 - val_accuracy: 0.7709 - val_loss: 0.4857 - val_precision: 0.6479 - val_recall: 0.7419\n",
      "Epoch 82/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8849 - accuracy: 0.8112 - loss: 0.4370 - precision: 0.7125 - recall: 0.8490 - val_AUC: 0.8364 - val_accuracy: 0.7709 - val_loss: 0.4866 - val_precision: 0.6479 - val_recall: 0.7419\n",
      "Epoch 83/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8970 - accuracy: 0.8211 - loss: 0.4228 - precision: 0.7346 - recall: 0.8205 - val_AUC: 0.8371 - val_accuracy: 0.7654 - val_loss: 0.4863 - val_precision: 0.6429 - val_recall: 0.7258\n",
      "Epoch 84/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9056 - accuracy: 0.8266 - loss: 0.4033 - precision: 0.7094 - recall: 0.8638 - val_AUC: 0.8378 - val_accuracy: 0.7709 - val_loss: 0.4861 - val_precision: 0.6522 - val_recall: 0.7258\n",
      "Epoch 85/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9027 - accuracy: 0.8355 - loss: 0.3950 - precision: 0.7190 - recall: 0.8553 - val_AUC: 0.8362 - val_accuracy: 0.7709 - val_loss: 0.4858 - val_precision: 0.6479 - val_recall: 0.7419\n",
      "Epoch 86/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8727 - accuracy: 0.7981 - loss: 0.4526 - precision: 0.6866 - recall: 0.8477 - val_AUC: 0.8368 - val_accuracy: 0.7821 - val_loss: 0.4861 - val_precision: 0.6667 - val_recall: 0.7419\n",
      "Epoch 87/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9059 - accuracy: 0.8335 - loss: 0.3962 - precision: 0.7288 - recall: 0.8364 - val_AUC: 0.8368 - val_accuracy: 0.7765 - val_loss: 0.4863 - val_precision: 0.6571 - val_recall: 0.7419\n",
      "Epoch 88/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9123 - accuracy: 0.8482 - loss: 0.3913 - precision: 0.7595 - recall: 0.8538 - val_AUC: 0.8373 - val_accuracy: 0.7709 - val_loss: 0.4882 - val_precision: 0.6479 - val_recall: 0.7419\n",
      "Epoch 89/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8848 - accuracy: 0.8060 - loss: 0.4347 - precision: 0.7054 - recall: 0.8269 - val_AUC: 0.8366 - val_accuracy: 0.7765 - val_loss: 0.4883 - val_precision: 0.6618 - val_recall: 0.7258\n",
      "Epoch 90/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9128 - accuracy: 0.8452 - loss: 0.3842 - precision: 0.7473 - recall: 0.8606 - val_AUC: 0.8362 - val_accuracy: 0.7821 - val_loss: 0.4879 - val_precision: 0.6667 - val_recall: 0.7419\n",
      "Epoch 91/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9069 - accuracy: 0.8482 - loss: 0.3865 - precision: 0.7354 - recall: 0.8373 - val_AUC: 0.8351 - val_accuracy: 0.7765 - val_loss: 0.4876 - val_precision: 0.6618 - val_recall: 0.7258\n",
      "Epoch 92/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8884 - accuracy: 0.8232 - loss: 0.4224 - precision: 0.7241 - recall: 0.8066 - val_AUC: 0.8364 - val_accuracy: 0.7709 - val_loss: 0.4886 - val_precision: 0.6567 - val_recall: 0.7097\n",
      "Epoch 93/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8976 - accuracy: 0.8296 - loss: 0.4141 - precision: 0.7378 - recall: 0.8270 - val_AUC: 0.8351 - val_accuracy: 0.7709 - val_loss: 0.4891 - val_precision: 0.6567 - val_recall: 0.7097\n",
      "Epoch 94/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9042 - accuracy: 0.8482 - loss: 0.4021 - precision: 0.7504 - recall: 0.8767 - val_AUC: 0.8344 - val_accuracy: 0.7709 - val_loss: 0.4893 - val_precision: 0.6567 - val_recall: 0.7097\n",
      "Epoch 95/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9092 - accuracy: 0.8267 - loss: 0.3866 - precision: 0.6957 - recall: 0.8560 - val_AUC: 0.8343 - val_accuracy: 0.7709 - val_loss: 0.4892 - val_precision: 0.6567 - val_recall: 0.7097\n",
      "Epoch 96/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9078 - accuracy: 0.8376 - loss: 0.3935 - precision: 0.7275 - recall: 0.8643 - val_AUC: 0.8343 - val_accuracy: 0.7765 - val_loss: 0.4912 - val_precision: 0.6618 - val_recall: 0.7258\n",
      "Epoch 97/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9039 - accuracy: 0.8276 - loss: 0.4012 - precision: 0.7248 - recall: 0.8472 - val_AUC: 0.8340 - val_accuracy: 0.7821 - val_loss: 0.4921 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 98/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9105 - accuracy: 0.8455 - loss: 0.3910 - precision: 0.7423 - recall: 0.8792 - val_AUC: 0.8340 - val_accuracy: 0.7821 - val_loss: 0.4930 - val_precision: 0.6716 - val_recall: 0.7258\n",
      "Epoch 99/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9232 - accuracy: 0.8438 - loss: 0.3654 - precision: 0.7316 - recall: 0.8676 - val_AUC: 0.8344 - val_accuracy: 0.7877 - val_loss: 0.4929 - val_precision: 0.6765 - val_recall: 0.7419\n",
      "Epoch 100/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9177 - accuracy: 0.8446 - loss: 0.3764 - precision: 0.7418 - recall: 0.8617 - val_AUC: 0.8331 - val_accuracy: 0.7821 - val_loss: 0.4931 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 101/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9260 - accuracy: 0.8458 - loss: 0.3661 - precision: 0.7387 - recall: 0.8846 - val_AUC: 0.8332 - val_accuracy: 0.7877 - val_loss: 0.4938 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 102/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9223 - accuracy: 0.8481 - loss: 0.3651 - precision: 0.7485 - recall: 0.8727 - val_AUC: 0.8325 - val_accuracy: 0.7821 - val_loss: 0.4946 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 103/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9307 - accuracy: 0.8731 - loss: 0.3473 - precision: 0.7940 - recall: 0.8772 - val_AUC: 0.8311 - val_accuracy: 0.7821 - val_loss: 0.4971 - val_precision: 0.6716 - val_recall: 0.7258\n",
      "Epoch 104/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9015 - accuracy: 0.8314 - loss: 0.4011 - precision: 0.7306 - recall: 0.8524 - val_AUC: 0.8297 - val_accuracy: 0.7765 - val_loss: 0.4983 - val_precision: 0.6667 - val_recall: 0.7097\n",
      "Epoch 105/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9334 - accuracy: 0.8653 - loss: 0.3436 - precision: 0.7756 - recall: 0.8502 - val_AUC: 0.8311 - val_accuracy: 0.7877 - val_loss: 0.4972 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 106/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9188 - accuracy: 0.8528 - loss: 0.3837 - precision: 0.7929 - recall: 0.8213 - val_AUC: 0.8295 - val_accuracy: 0.7765 - val_loss: 0.4992 - val_precision: 0.6667 - val_recall: 0.7097\n",
      "Epoch 107/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9169 - accuracy: 0.8353 - loss: 0.3695 - precision: 0.7155 - recall: 0.8244 - val_AUC: 0.8287 - val_accuracy: 0.7821 - val_loss: 0.4993 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 108/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8994 - accuracy: 0.8194 - loss: 0.4129 - precision: 0.7435 - recall: 0.8035 - val_AUC: 0.8272 - val_accuracy: 0.7821 - val_loss: 0.5001 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 109/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9108 - accuracy: 0.8339 - loss: 0.3849 - precision: 0.7429 - recall: 0.8323 - val_AUC: 0.8262 - val_accuracy: 0.7877 - val_loss: 0.5026 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 110/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9338 - accuracy: 0.8681 - loss: 0.3425 - precision: 0.7785 - recall: 0.8836 - val_AUC: 0.8255 - val_accuracy: 0.7821 - val_loss: 0.5031 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 111/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9369 - accuracy: 0.8652 - loss: 0.3399 - precision: 0.7856 - recall: 0.8533 - val_AUC: 0.8258 - val_accuracy: 0.7877 - val_loss: 0.5045 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 112/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9145 - accuracy: 0.8310 - loss: 0.3764 - precision: 0.7350 - recall: 0.8073 - val_AUC: 0.8242 - val_accuracy: 0.7877 - val_loss: 0.5064 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 113/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9241 - accuracy: 0.8506 - loss: 0.3568 - precision: 0.7373 - recall: 0.8595 - val_AUC: 0.8248 - val_accuracy: 0.7821 - val_loss: 0.5062 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 114/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9386 - accuracy: 0.8667 - loss: 0.3326 - precision: 0.7741 - recall: 0.8762 - val_AUC: 0.8245 - val_accuracy: 0.7877 - val_loss: 0.5077 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 115/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9323 - accuracy: 0.8566 - loss: 0.3388 - precision: 0.7352 - recall: 0.8635 - val_AUC: 0.8246 - val_accuracy: 0.7877 - val_loss: 0.5086 - val_precision: 0.6818 - val_recall: 0.7258\n",
      "Epoch 116/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9386 - accuracy: 0.8641 - loss: 0.3288 - precision: 0.7686 - recall: 0.8845 - val_AUC: 0.8227 - val_accuracy: 0.7821 - val_loss: 0.5104 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 117/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9340 - accuracy: 0.8635 - loss: 0.3382 - precision: 0.7596 - recall: 0.8681 - val_AUC: 0.8235 - val_accuracy: 0.7821 - val_loss: 0.5115 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 118/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9349 - accuracy: 0.8542 - loss: 0.3318 - precision: 0.7267 - recall: 0.8829 - val_AUC: 0.8227 - val_accuracy: 0.7821 - val_loss: 0.5136 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 119/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9327 - accuracy: 0.8609 - loss: 0.3416 - precision: 0.7717 - recall: 0.8691 - val_AUC: 0.8206 - val_accuracy: 0.7709 - val_loss: 0.5156 - val_precision: 0.6615 - val_recall: 0.6935\n",
      "Epoch 120/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9336 - accuracy: 0.8683 - loss: 0.3359 - precision: 0.7716 - recall: 0.8812 - val_AUC: 0.8210 - val_accuracy: 0.7821 - val_loss: 0.5158 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 121/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9323 - accuracy: 0.8567 - loss: 0.3443 - precision: 0.7710 - recall: 0.8709 - val_AUC: 0.8209 - val_accuracy: 0.7765 - val_loss: 0.5172 - val_precision: 0.6667 - val_recall: 0.7097\n",
      "Epoch 122/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9285 - accuracy: 0.8647 - loss: 0.3465 - precision: 0.7762 - recall: 0.8922 - val_AUC: 0.8195 - val_accuracy: 0.7765 - val_loss: 0.5176 - val_precision: 0.6667 - val_recall: 0.7097\n",
      "Epoch 123/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9419 - accuracy: 0.8746 - loss: 0.3244 - precision: 0.7738 - recall: 0.9154 - val_AUC: 0.8198 - val_accuracy: 0.7821 - val_loss: 0.5186 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 124/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9319 - accuracy: 0.8640 - loss: 0.3354 - precision: 0.7675 - recall: 0.8711 - val_AUC: 0.8186 - val_accuracy: 0.7821 - val_loss: 0.5203 - val_precision: 0.6769 - val_recall: 0.7097\n",
      "Epoch 125/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9430 - accuracy: 0.8669 - loss: 0.3196 - precision: 0.7644 - recall: 0.8696 - val_AUC: 0.8188 - val_accuracy: 0.7877 - val_loss: 0.5230 - val_precision: 0.6875 - val_recall: 0.7097\n",
      "Epoch 126/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9512 - accuracy: 0.8865 - loss: 0.2998 - precision: 0.8062 - recall: 0.8894 - val_AUC: 0.8187 - val_accuracy: 0.7877 - val_loss: 0.5242 - val_precision: 0.6875 - val_recall: 0.7097\n",
      "\u001b[1m8/8\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - AUC: 0.7876 - accuracy: 0.7335 - loss: 0.5529 - precision: 0.6106 - recall: 0.7027 \n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 21ms/step - AUC: 0.5486 - accuracy: 0.3767 - loss: 0.9286 - precision: 0.3616 - recall: 0.9590 - val_AUC: 0.5695 - val_accuracy: 0.5196 - val_loss: 0.7576 - val_precision: 0.3723 - val_recall: 0.5645\n",
      "Epoch 2/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.5986 - accuracy: 0.3848 - loss: 0.8612 - precision: 0.3672 - recall: 0.9534 - val_AUC: 0.5929 - val_accuracy: 0.5587 - val_loss: 0.7252 - val_precision: 0.4023 - val_recall: 0.5645\n",
      "Epoch 3/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6397 - accuracy: 0.3812 - loss: 0.8127 - precision: 0.3445 - recall: 0.9190 - val_AUC: 0.6210 - val_accuracy: 0.5810 - val_loss: 0.6974 - val_precision: 0.4198 - val_recall: 0.5484\n",
      "Epoch 4/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6509 - accuracy: 0.4090 - loss: 0.7905 - precision: 0.3526 - recall: 0.9120 - val_AUC: 0.6419 - val_accuracy: 0.6145 - val_loss: 0.6741 - val_precision: 0.4533 - val_recall: 0.5484\n",
      "Epoch 5/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6590 - accuracy: 0.4759 - loss: 0.7399 - precision: 0.3921 - recall: 0.8976 - val_AUC: 0.6559 - val_accuracy: 0.6089 - val_loss: 0.6555 - val_precision: 0.4444 - val_recall: 0.5161\n",
      "Epoch 6/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6782 - accuracy: 0.5331 - loss: 0.7136 - precision: 0.4233 - recall: 0.8724 - val_AUC: 0.6707 - val_accuracy: 0.5978 - val_loss: 0.6399 - val_precision: 0.4265 - val_recall: 0.4677\n",
      "Epoch 7/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7211 - accuracy: 0.5904 - loss: 0.6616 - precision: 0.4549 - recall: 0.8398 - val_AUC: 0.6816 - val_accuracy: 0.6257 - val_loss: 0.6267 - val_precision: 0.4627 - val_recall: 0.5000\n",
      "Epoch 8/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7091 - accuracy: 0.5938 - loss: 0.6707 - precision: 0.4649 - recall: 0.8455 - val_AUC: 0.6896 - val_accuracy: 0.6257 - val_loss: 0.6153 - val_precision: 0.4603 - val_recall: 0.4677\n",
      "Epoch 9/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7460 - accuracy: 0.6247 - loss: 0.6240 - precision: 0.4918 - recall: 0.8209 - val_AUC: 0.6991 - val_accuracy: 0.6313 - val_loss: 0.6062 - val_precision: 0.4677 - val_recall: 0.4677\n",
      "Epoch 10/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7560 - accuracy: 0.6395 - loss: 0.6118 - precision: 0.5086 - recall: 0.8190 - val_AUC: 0.7089 - val_accuracy: 0.6480 - val_loss: 0.5989 - val_precision: 0.4918 - val_recall: 0.4839\n",
      "Epoch 11/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7299 - accuracy: 0.6494 - loss: 0.6229 - precision: 0.4853 - recall: 0.7346 - val_AUC: 0.7155 - val_accuracy: 0.6704 - val_loss: 0.5925 - val_precision: 0.5254 - val_recall: 0.5000\n",
      "Epoch 12/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7825 - accuracy: 0.6860 - loss: 0.5760 - precision: 0.5263 - recall: 0.7817 - val_AUC: 0.7215 - val_accuracy: 0.6704 - val_loss: 0.5873 - val_precision: 0.5254 - val_recall: 0.5000\n",
      "Epoch 13/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7480 - accuracy: 0.6504 - loss: 0.5985 - precision: 0.4697 - recall: 0.6883 - val_AUC: 0.7271 - val_accuracy: 0.6872 - val_loss: 0.5829 - val_precision: 0.5536 - val_recall: 0.5000\n",
      "Epoch 14/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7788 - accuracy: 0.6894 - loss: 0.5774 - precision: 0.5538 - recall: 0.7750 - val_AUC: 0.7350 - val_accuracy: 0.6927 - val_loss: 0.5785 - val_precision: 0.5636 - val_recall: 0.5000\n",
      "Epoch 15/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7933 - accuracy: 0.7094 - loss: 0.5595 - precision: 0.5569 - recall: 0.7763 - val_AUC: 0.7386 - val_accuracy: 0.6983 - val_loss: 0.5750 - val_precision: 0.5769 - val_recall: 0.4839\n",
      "Epoch 16/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8249 - accuracy: 0.7565 - loss: 0.5174 - precision: 0.6167 - recall: 0.7631 - val_AUC: 0.7439 - val_accuracy: 0.7095 - val_loss: 0.5722 - val_precision: 0.5962 - val_recall: 0.5000\n",
      "Epoch 17/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7956 - accuracy: 0.7249 - loss: 0.5466 - precision: 0.5655 - recall: 0.7342 - val_AUC: 0.7466 - val_accuracy: 0.7207 - val_loss: 0.5703 - val_precision: 0.6154 - val_recall: 0.5161\n",
      "Epoch 18/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8185 - accuracy: 0.7489 - loss: 0.5252 - precision: 0.5958 - recall: 0.7908 - val_AUC: 0.7482 - val_accuracy: 0.7207 - val_loss: 0.5684 - val_precision: 0.6200 - val_recall: 0.5000\n",
      "Epoch 19/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8050 - accuracy: 0.7224 - loss: 0.5433 - precision: 0.5783 - recall: 0.7671 - val_AUC: 0.7519 - val_accuracy: 0.7263 - val_loss: 0.5658 - val_precision: 0.6275 - val_recall: 0.5161\n",
      "Epoch 20/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8330 - accuracy: 0.7623 - loss: 0.5103 - precision: 0.6502 - recall: 0.7665 - val_AUC: 0.7539 - val_accuracy: 0.7263 - val_loss: 0.5646 - val_precision: 0.6275 - val_recall: 0.5161\n",
      "Epoch 21/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8385 - accuracy: 0.7580 - loss: 0.5000 - precision: 0.6054 - recall: 0.8062 - val_AUC: 0.7551 - val_accuracy: 0.7318 - val_loss: 0.5635 - val_precision: 0.6400 - val_recall: 0.5161\n",
      "Epoch 22/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8336 - accuracy: 0.7572 - loss: 0.5069 - precision: 0.6188 - recall: 0.7654 - val_AUC: 0.7574 - val_accuracy: 0.7263 - val_loss: 0.5626 - val_precision: 0.6275 - val_recall: 0.5161\n",
      "Epoch 23/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8260 - accuracy: 0.7548 - loss: 0.5144 - precision: 0.6178 - recall: 0.7421 - val_AUC: 0.7585 - val_accuracy: 0.7151 - val_loss: 0.5613 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 24/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8029 - accuracy: 0.7324 - loss: 0.5417 - precision: 0.5895 - recall: 0.7360 - val_AUC: 0.7588 - val_accuracy: 0.7151 - val_loss: 0.5610 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 25/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8065 - accuracy: 0.7335 - loss: 0.5441 - precision: 0.6047 - recall: 0.7833 - val_AUC: 0.7619 - val_accuracy: 0.7151 - val_loss: 0.5605 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 26/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8453 - accuracy: 0.7822 - loss: 0.4933 - precision: 0.6642 - recall: 0.7853 - val_AUC: 0.7633 - val_accuracy: 0.7151 - val_loss: 0.5601 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 27/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8272 - accuracy: 0.7571 - loss: 0.5127 - precision: 0.6188 - recall: 0.7862 - val_AUC: 0.7649 - val_accuracy: 0.7151 - val_loss: 0.5598 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 28/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8283 - accuracy: 0.7538 - loss: 0.5122 - precision: 0.6221 - recall: 0.7413 - val_AUC: 0.7653 - val_accuracy: 0.7151 - val_loss: 0.5603 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 29/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8476 - accuracy: 0.7822 - loss: 0.4838 - precision: 0.6482 - recall: 0.7982 - val_AUC: 0.7653 - val_accuracy: 0.7151 - val_loss: 0.5594 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 30/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8360 - accuracy: 0.7756 - loss: 0.5079 - precision: 0.6669 - recall: 0.7737 - val_AUC: 0.7661 - val_accuracy: 0.7151 - val_loss: 0.5589 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 31/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8205 - accuracy: 0.7607 - loss: 0.5218 - precision: 0.6302 - recall: 0.7994 - val_AUC: 0.7674 - val_accuracy: 0.7151 - val_loss: 0.5592 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 32/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8362 - accuracy: 0.7652 - loss: 0.5018 - precision: 0.6407 - recall: 0.8010 - val_AUC: 0.7667 - val_accuracy: 0.7151 - val_loss: 0.5595 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 33/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8282 - accuracy: 0.7568 - loss: 0.5042 - precision: 0.6133 - recall: 0.7717 - val_AUC: 0.7666 - val_accuracy: 0.7151 - val_loss: 0.5592 - val_precision: 0.6078 - val_recall: 0.5000\n",
      "Epoch 34/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8447 - accuracy: 0.7791 - loss: 0.4944 - precision: 0.6686 - recall: 0.8068 - val_AUC: 0.7680 - val_accuracy: 0.7095 - val_loss: 0.5590 - val_precision: 0.5962 - val_recall: 0.5000\n",
      "Epoch 35/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8607 - accuracy: 0.7970 - loss: 0.4649 - precision: 0.6721 - recall: 0.8015 - val_AUC: 0.7677 - val_accuracy: 0.7151 - val_loss: 0.5598 - val_precision: 0.6038 - val_recall: 0.5161\n",
      "Epoch 36/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8471 - accuracy: 0.7750 - loss: 0.4817 - precision: 0.6212 - recall: 0.8174 - val_AUC: 0.7677 - val_accuracy: 0.7151 - val_loss: 0.5601 - val_precision: 0.6038 - val_recall: 0.5161\n",
      "Epoch 37/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8375 - accuracy: 0.7677 - loss: 0.4986 - precision: 0.6384 - recall: 0.7663 - val_AUC: 0.7674 - val_accuracy: 0.7151 - val_loss: 0.5600 - val_precision: 0.6038 - val_recall: 0.5161\n",
      "Epoch 38/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8288 - accuracy: 0.7651 - loss: 0.5097 - precision: 0.6385 - recall: 0.7793 - val_AUC: 0.7679 - val_accuracy: 0.7151 - val_loss: 0.5599 - val_precision: 0.6038 - val_recall: 0.5161\n",
      "Epoch 39/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8633 - accuracy: 0.7916 - loss: 0.4629 - precision: 0.6540 - recall: 0.8343 - val_AUC: 0.7676 - val_accuracy: 0.7151 - val_loss: 0.5608 - val_precision: 0.6038 - val_recall: 0.5161\n",
      "Epoch 40/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8626 - accuracy: 0.7826 - loss: 0.4640 - precision: 0.6450 - recall: 0.8260 - val_AUC: 0.7691 - val_accuracy: 0.7318 - val_loss: 0.5602 - val_precision: 0.6250 - val_recall: 0.5645\n",
      "Epoch 41/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8485 - accuracy: 0.7722 - loss: 0.4800 - precision: 0.6223 - recall: 0.8006 - val_AUC: 0.7685 - val_accuracy: 0.7318 - val_loss: 0.5611 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 42/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8409 - accuracy: 0.7812 - loss: 0.4956 - precision: 0.6566 - recall: 0.7999 - val_AUC: 0.7688 - val_accuracy: 0.7318 - val_loss: 0.5623 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 43/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8636 - accuracy: 0.7700 - loss: 0.4588 - precision: 0.6151 - recall: 0.7914 - val_AUC: 0.7688 - val_accuracy: 0.7318 - val_loss: 0.5614 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 44/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8515 - accuracy: 0.7819 - loss: 0.4747 - precision: 0.6467 - recall: 0.7944 - val_AUC: 0.7699 - val_accuracy: 0.7318 - val_loss: 0.5619 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 45/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8660 - accuracy: 0.7907 - loss: 0.4545 - precision: 0.6478 - recall: 0.8444 - val_AUC: 0.7703 - val_accuracy: 0.7318 - val_loss: 0.5616 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 46/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8629 - accuracy: 0.7897 - loss: 0.4629 - precision: 0.6725 - recall: 0.8218 - val_AUC: 0.7708 - val_accuracy: 0.7318 - val_loss: 0.5628 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 47/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8646 - accuracy: 0.7927 - loss: 0.4599 - precision: 0.6683 - recall: 0.8303 - val_AUC: 0.7714 - val_accuracy: 0.7318 - val_loss: 0.5631 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 48/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8536 - accuracy: 0.7726 - loss: 0.4771 - precision: 0.6442 - recall: 0.8097 - val_AUC: 0.7721 - val_accuracy: 0.7374 - val_loss: 0.5633 - val_precision: 0.6316 - val_recall: 0.5806\n",
      "Epoch 49/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8363 - accuracy: 0.7527 - loss: 0.4977 - precision: 0.6071 - recall: 0.8191 - val_AUC: 0.7712 - val_accuracy: 0.7374 - val_loss: 0.5642 - val_precision: 0.6316 - val_recall: 0.5806\n",
      "Epoch 50/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8420 - accuracy: 0.7811 - loss: 0.4924 - precision: 0.6721 - recall: 0.8068 - val_AUC: 0.7720 - val_accuracy: 0.7318 - val_loss: 0.5646 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 51/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8604 - accuracy: 0.7973 - loss: 0.4659 - precision: 0.6780 - recall: 0.8110 - val_AUC: 0.7727 - val_accuracy: 0.7263 - val_loss: 0.5655 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 52/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8517 - accuracy: 0.7870 - loss: 0.4746 - precision: 0.6581 - recall: 0.7908 - val_AUC: 0.7734 - val_accuracy: 0.7318 - val_loss: 0.5652 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 53/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8799 - accuracy: 0.8067 - loss: 0.4372 - precision: 0.6826 - recall: 0.8396 - val_AUC: 0.7743 - val_accuracy: 0.7263 - val_loss: 0.5661 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 54/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8778 - accuracy: 0.8023 - loss: 0.4364 - precision: 0.6729 - recall: 0.8580 - val_AUC: 0.7744 - val_accuracy: 0.7263 - val_loss: 0.5674 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 55/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8825 - accuracy: 0.8007 - loss: 0.4269 - precision: 0.6542 - recall: 0.8441 - val_AUC: 0.7747 - val_accuracy: 0.7318 - val_loss: 0.5677 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 56/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8735 - accuracy: 0.7877 - loss: 0.4410 - precision: 0.6392 - recall: 0.8193 - val_AUC: 0.7755 - val_accuracy: 0.7374 - val_loss: 0.5682 - val_precision: 0.6316 - val_recall: 0.5806\n",
      "Epoch 57/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8894 - accuracy: 0.8136 - loss: 0.4170 - precision: 0.6832 - recall: 0.8283 - val_AUC: 0.7747 - val_accuracy: 0.7318 - val_loss: 0.5686 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 58/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8666 - accuracy: 0.7700 - loss: 0.4594 - precision: 0.6563 - recall: 0.8191 - val_AUC: 0.7750 - val_accuracy: 0.7318 - val_loss: 0.5684 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 59/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8697 - accuracy: 0.7853 - loss: 0.4531 - precision: 0.6569 - recall: 0.8106 - val_AUC: 0.7743 - val_accuracy: 0.7263 - val_loss: 0.5698 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 60/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8706 - accuracy: 0.7826 - loss: 0.4499 - precision: 0.6601 - recall: 0.7883 - val_AUC: 0.7732 - val_accuracy: 0.7318 - val_loss: 0.5712 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 61/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8745 - accuracy: 0.7901 - loss: 0.4429 - precision: 0.6669 - recall: 0.7946 - val_AUC: 0.7744 - val_accuracy: 0.7207 - val_loss: 0.5716 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 62/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8847 - accuracy: 0.8059 - loss: 0.4229 - precision: 0.6559 - recall: 0.8649 - val_AUC: 0.7745 - val_accuracy: 0.7207 - val_loss: 0.5726 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 63/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8807 - accuracy: 0.7994 - loss: 0.4304 - precision: 0.6646 - recall: 0.8354 - val_AUC: 0.7732 - val_accuracy: 0.7207 - val_loss: 0.5724 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 64/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8884 - accuracy: 0.8070 - loss: 0.4354 - precision: 0.7300 - recall: 0.8105 - val_AUC: 0.7741 - val_accuracy: 0.7263 - val_loss: 0.5734 - val_precision: 0.6066 - val_recall: 0.5968\n",
      "Epoch 65/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8787 - accuracy: 0.7875 - loss: 0.4416 - precision: 0.6790 - recall: 0.8127 - val_AUC: 0.7746 - val_accuracy: 0.7263 - val_loss: 0.5734 - val_precision: 0.6066 - val_recall: 0.5968\n",
      "Epoch 66/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.8728 - accuracy: 0.7790 - loss: 0.4477 - precision: 0.6576 - recall: 0.8253 - val_AUC: 0.7744 - val_accuracy: 0.7318 - val_loss: 0.5745 - val_precision: 0.6167 - val_recall: 0.5968\n",
      "Epoch 67/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9027 - accuracy: 0.8252 - loss: 0.4071 - precision: 0.7355 - recall: 0.8674 - val_AUC: 0.7731 - val_accuracy: 0.7207 - val_loss: 0.5767 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 68/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8951 - accuracy: 0.8094 - loss: 0.4129 - precision: 0.6891 - recall: 0.8515 - val_AUC: 0.7726 - val_accuracy: 0.7151 - val_loss: 0.5780 - val_precision: 0.5902 - val_recall: 0.5806\n",
      "Epoch 69/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8869 - accuracy: 0.8020 - loss: 0.4262 - precision: 0.6762 - recall: 0.8283 - val_AUC: 0.7741 - val_accuracy: 0.7318 - val_loss: 0.5774 - val_precision: 0.6207 - val_recall: 0.5806\n",
      "Epoch 70/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8957 - accuracy: 0.8091 - loss: 0.4098 - precision: 0.6887 - recall: 0.8211 - val_AUC: 0.7737 - val_accuracy: 0.7374 - val_loss: 0.5774 - val_precision: 0.6271 - val_recall: 0.5968\n",
      "Epoch 71/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8752 - accuracy: 0.7804 - loss: 0.4334 - precision: 0.6208 - recall: 0.8045 - val_AUC: 0.7741 - val_accuracy: 0.7374 - val_loss: 0.5778 - val_precision: 0.6271 - val_recall: 0.5968\n",
      "Epoch 72/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9123 - accuracy: 0.8182 - loss: 0.3765 - precision: 0.6598 - recall: 0.8687 - val_AUC: 0.7727 - val_accuracy: 0.7318 - val_loss: 0.5804 - val_precision: 0.6167 - val_recall: 0.5968\n",
      "Epoch 73/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9055 - accuracy: 0.8274 - loss: 0.3979 - precision: 0.7294 - recall: 0.8311 - val_AUC: 0.7743 - val_accuracy: 0.7318 - val_loss: 0.5803 - val_precision: 0.6167 - val_recall: 0.5968\n",
      "Epoch 74/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8988 - accuracy: 0.8244 - loss: 0.3961 - precision: 0.6778 - recall: 0.8466 - val_AUC: 0.7712 - val_accuracy: 0.7207 - val_loss: 0.5824 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 75/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9044 - accuracy: 0.8233 - loss: 0.4006 - precision: 0.7165 - recall: 0.8607 - val_AUC: 0.7716 - val_accuracy: 0.7207 - val_loss: 0.5811 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 76/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9109 - accuracy: 0.8315 - loss: 0.3859 - precision: 0.7129 - recall: 0.8554 - val_AUC: 0.7711 - val_accuracy: 0.7207 - val_loss: 0.5835 - val_precision: 0.6000 - val_recall: 0.5806\n",
      "Epoch 77/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9083 - accuracy: 0.8286 - loss: 0.3876 - precision: 0.7118 - recall: 0.8652 - val_AUC: 0.7714 - val_accuracy: 0.7263 - val_loss: 0.5841 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 78/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9143 - accuracy: 0.8484 - loss: 0.3753 - precision: 0.7462 - recall: 0.8466 - val_AUC: 0.7703 - val_accuracy: 0.7263 - val_loss: 0.5865 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 79/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9040 - accuracy: 0.8289 - loss: 0.3975 - precision: 0.7185 - recall: 0.8587 - val_AUC: 0.7705 - val_accuracy: 0.7263 - val_loss: 0.5874 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "Epoch 80/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8859 - accuracy: 0.8009 - loss: 0.4237 - precision: 0.6721 - recall: 0.8227 - val_AUC: 0.7714 - val_accuracy: 0.7263 - val_loss: 0.5876 - val_precision: 0.6102 - val_recall: 0.5806\n",
      "\u001b[1m8/8\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - AUC: 0.7816 - accuracy: 0.7541 - loss: 0.5614 - precision: 0.7415 - recall: 0.4887 \n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (127) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 21ms/step - AUC: 0.6643 - accuracy: 0.6057 - loss: 0.7147 - precision: 0.4741 - recall: 0.7291 - val_AUC: 0.6564 - val_accuracy: 0.6201 - val_loss: 0.6953 - val_precision: 0.4651 - val_recall: 0.6452\n",
      "Epoch 2/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6770 - accuracy: 0.6080 - loss: 0.7140 - precision: 0.4626 - recall: 0.7654 - val_AUC: 0.6749 - val_accuracy: 0.6369 - val_loss: 0.6753 - val_precision: 0.4824 - val_recall: 0.6613\n",
      "Epoch 3/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7142 - accuracy: 0.6410 - loss: 0.6525 - precision: 0.4865 - recall: 0.7441 - val_AUC: 0.6881 - val_accuracy: 0.6313 - val_loss: 0.6604 - val_precision: 0.4762 - val_recall: 0.6452\n",
      "Epoch 4/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7307 - accuracy: 0.6363 - loss: 0.6227 - precision: 0.4540 - recall: 0.7187 - val_AUC: 0.6992 - val_accuracy: 0.6369 - val_loss: 0.6471 - val_precision: 0.4819 - val_recall: 0.6452\n",
      "Epoch 5/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7340 - accuracy: 0.6491 - loss: 0.6212 - precision: 0.5043 - recall: 0.7252 - val_AUC: 0.7091 - val_accuracy: 0.6480 - val_loss: 0.6362 - val_precision: 0.4940 - val_recall: 0.6613\n",
      "Epoch 6/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7409 - accuracy: 0.6599 - loss: 0.6179 - precision: 0.5151 - recall: 0.7591 - val_AUC: 0.7169 - val_accuracy: 0.6480 - val_loss: 0.6272 - val_precision: 0.4938 - val_recall: 0.6452\n",
      "Epoch 7/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7835 - accuracy: 0.7067 - loss: 0.5707 - precision: 0.5677 - recall: 0.7904 - val_AUC: 0.7214 - val_accuracy: 0.6592 - val_loss: 0.6191 - val_precision: 0.5063 - val_recall: 0.6452\n",
      "Epoch 8/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7555 - accuracy: 0.6764 - loss: 0.6040 - precision: 0.5363 - recall: 0.7127 - val_AUC: 0.7286 - val_accuracy: 0.6480 - val_loss: 0.6125 - val_precision: 0.4937 - val_recall: 0.6290\n",
      "Epoch 9/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7446 - accuracy: 0.6760 - loss: 0.6187 - precision: 0.5364 - recall: 0.7334 - val_AUC: 0.7329 - val_accuracy: 0.6536 - val_loss: 0.6060 - val_precision: 0.5000 - val_recall: 0.6290\n",
      "Epoch 10/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7726 - accuracy: 0.6816 - loss: 0.5761 - precision: 0.5252 - recall: 0.7186 - val_AUC: 0.7378 - val_accuracy: 0.6704 - val_loss: 0.6010 - val_precision: 0.5195 - val_recall: 0.6452\n",
      "Epoch 11/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7953 - accuracy: 0.7147 - loss: 0.5509 - precision: 0.5506 - recall: 0.7359 - val_AUC: 0.7410 - val_accuracy: 0.6927 - val_loss: 0.5960 - val_precision: 0.5467 - val_recall: 0.6613\n",
      "Epoch 12/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8148 - accuracy: 0.7171 - loss: 0.5258 - precision: 0.5710 - recall: 0.7782 - val_AUC: 0.7448 - val_accuracy: 0.6983 - val_loss: 0.5922 - val_precision: 0.5541 - val_recall: 0.6613\n",
      "Epoch 13/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7927 - accuracy: 0.7147 - loss: 0.5458 - precision: 0.5474 - recall: 0.7489 - val_AUC: 0.7483 - val_accuracy: 0.7151 - val_loss: 0.5890 - val_precision: 0.5733 - val_recall: 0.6935\n",
      "Epoch 14/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7717 - accuracy: 0.7069 - loss: 0.5842 - precision: 0.5582 - recall: 0.7424 - val_AUC: 0.7508 - val_accuracy: 0.7151 - val_loss: 0.5853 - val_precision: 0.5733 - val_recall: 0.6935\n",
      "Epoch 15/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8048 - accuracy: 0.7197 - loss: 0.5327 - precision: 0.5425 - recall: 0.7666 - val_AUC: 0.7528 - val_accuracy: 0.7207 - val_loss: 0.5823 - val_precision: 0.5811 - val_recall: 0.6935\n",
      "Epoch 16/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8142 - accuracy: 0.7326 - loss: 0.5315 - precision: 0.5908 - recall: 0.7809 - val_AUC: 0.7542 - val_accuracy: 0.7151 - val_loss: 0.5796 - val_precision: 0.5733 - val_recall: 0.6935\n",
      "Epoch 17/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7995 - accuracy: 0.7055 - loss: 0.5526 - precision: 0.5689 - recall: 0.7333 - val_AUC: 0.7557 - val_accuracy: 0.7151 - val_loss: 0.5770 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 18/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8028 - accuracy: 0.7274 - loss: 0.5411 - precision: 0.5814 - recall: 0.7435 - val_AUC: 0.7583 - val_accuracy: 0.7207 - val_loss: 0.5741 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 19/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8148 - accuracy: 0.7309 - loss: 0.5275 - precision: 0.5841 - recall: 0.7334 - val_AUC: 0.7594 - val_accuracy: 0.7151 - val_loss: 0.5725 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 20/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8157 - accuracy: 0.7264 - loss: 0.5272 - precision: 0.5859 - recall: 0.7247 - val_AUC: 0.7608 - val_accuracy: 0.7151 - val_loss: 0.5702 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 21/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8332 - accuracy: 0.7453 - loss: 0.5020 - precision: 0.6044 - recall: 0.7404 - val_AUC: 0.7621 - val_accuracy: 0.7151 - val_loss: 0.5676 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 22/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8244 - accuracy: 0.7268 - loss: 0.5182 - precision: 0.5967 - recall: 0.7154 - val_AUC: 0.7628 - val_accuracy: 0.7151 - val_loss: 0.5670 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 23/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8322 - accuracy: 0.7498 - loss: 0.5076 - precision: 0.6123 - recall: 0.7536 - val_AUC: 0.7632 - val_accuracy: 0.7151 - val_loss: 0.5654 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 24/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8522 - accuracy: 0.7720 - loss: 0.4724 - precision: 0.6158 - recall: 0.7824 - val_AUC: 0.7643 - val_accuracy: 0.7151 - val_loss: 0.5630 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 25/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8359 - accuracy: 0.7483 - loss: 0.5029 - precision: 0.6317 - recall: 0.7637 - val_AUC: 0.7670 - val_accuracy: 0.7151 - val_loss: 0.5628 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 26/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8495 - accuracy: 0.7667 - loss: 0.4792 - precision: 0.6282 - recall: 0.7896 - val_AUC: 0.7680 - val_accuracy: 0.7263 - val_loss: 0.5616 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 27/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8329 - accuracy: 0.7471 - loss: 0.4935 - precision: 0.5739 - recall: 0.7587 - val_AUC: 0.7697 - val_accuracy: 0.7263 - val_loss: 0.5591 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 28/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8345 - accuracy: 0.7647 - loss: 0.4971 - precision: 0.6315 - recall: 0.7284 - val_AUC: 0.7714 - val_accuracy: 0.7263 - val_loss: 0.5585 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 29/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8451 - accuracy: 0.7646 - loss: 0.4973 - precision: 0.6523 - recall: 0.7891 - val_AUC: 0.7722 - val_accuracy: 0.7263 - val_loss: 0.5577 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 30/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8515 - accuracy: 0.7708 - loss: 0.4761 - precision: 0.6314 - recall: 0.7814 - val_AUC: 0.7740 - val_accuracy: 0.7263 - val_loss: 0.5566 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 31/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8347 - accuracy: 0.7717 - loss: 0.5077 - precision: 0.6570 - recall: 0.7520 - val_AUC: 0.7745 - val_accuracy: 0.7318 - val_loss: 0.5557 - val_precision: 0.6000 - val_recall: 0.6774\n",
      "Epoch 32/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8485 - accuracy: 0.7567 - loss: 0.4875 - precision: 0.6382 - recall: 0.7676 - val_AUC: 0.7760 - val_accuracy: 0.7374 - val_loss: 0.5549 - val_precision: 0.6056 - val_recall: 0.6935\n",
      "Epoch 33/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8726 - accuracy: 0.7848 - loss: 0.4420 - precision: 0.6355 - recall: 0.8098 - val_AUC: 0.7781 - val_accuracy: 0.7430 - val_loss: 0.5538 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 34/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8579 - accuracy: 0.7919 - loss: 0.4867 - precision: 0.7006 - recall: 0.7961 - val_AUC: 0.7785 - val_accuracy: 0.7430 - val_loss: 0.5537 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 35/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8350 - accuracy: 0.7570 - loss: 0.5061 - precision: 0.6276 - recall: 0.7800 - val_AUC: 0.7796 - val_accuracy: 0.7486 - val_loss: 0.5523 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 36/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8432 - accuracy: 0.7606 - loss: 0.4878 - precision: 0.6202 - recall: 0.7572 - val_AUC: 0.7803 - val_accuracy: 0.7486 - val_loss: 0.5520 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 37/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8412 - accuracy: 0.7751 - loss: 0.4849 - precision: 0.6219 - recall: 0.7728 - val_AUC: 0.7801 - val_accuracy: 0.7486 - val_loss: 0.5511 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 38/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8637 - accuracy: 0.7761 - loss: 0.4619 - precision: 0.6371 - recall: 0.8013 - val_AUC: 0.7809 - val_accuracy: 0.7486 - val_loss: 0.5502 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 39/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8592 - accuracy: 0.7861 - loss: 0.4731 - precision: 0.6664 - recall: 0.8038 - val_AUC: 0.7810 - val_accuracy: 0.7486 - val_loss: 0.5488 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 40/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8810 - accuracy: 0.7963 - loss: 0.4365 - precision: 0.6740 - recall: 0.8111 - val_AUC: 0.7812 - val_accuracy: 0.7542 - val_loss: 0.5479 - val_precision: 0.6286 - val_recall: 0.7097\n",
      "Epoch 41/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8814 - accuracy: 0.7968 - loss: 0.4405 - precision: 0.6829 - recall: 0.8299 - val_AUC: 0.7823 - val_accuracy: 0.7430 - val_loss: 0.5487 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 42/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8508 - accuracy: 0.7747 - loss: 0.4774 - precision: 0.6326 - recall: 0.8040 - val_AUC: 0.7825 - val_accuracy: 0.7486 - val_loss: 0.5479 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 43/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8731 - accuracy: 0.8035 - loss: 0.4472 - precision: 0.6758 - recall: 0.8403 - val_AUC: 0.7815 - val_accuracy: 0.7486 - val_loss: 0.5480 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 44/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8717 - accuracy: 0.7869 - loss: 0.4521 - precision: 0.6718 - recall: 0.8106 - val_AUC: 0.7830 - val_accuracy: 0.7430 - val_loss: 0.5470 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 45/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8729 - accuracy: 0.7969 - loss: 0.4441 - precision: 0.6653 - recall: 0.8317 - val_AUC: 0.7839 - val_accuracy: 0.7374 - val_loss: 0.5470 - val_precision: 0.6027 - val_recall: 0.7097\n",
      "Epoch 46/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8662 - accuracy: 0.7857 - loss: 0.4535 - precision: 0.6334 - recall: 0.8272 - val_AUC: 0.7832 - val_accuracy: 0.7430 - val_loss: 0.5459 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 47/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8746 - accuracy: 0.7969 - loss: 0.4484 - precision: 0.6941 - recall: 0.7957 - val_AUC: 0.7843 - val_accuracy: 0.7374 - val_loss: 0.5465 - val_precision: 0.6027 - val_recall: 0.7097\n",
      "Epoch 48/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8806 - accuracy: 0.8144 - loss: 0.4411 - precision: 0.7095 - recall: 0.8231 - val_AUC: 0.7846 - val_accuracy: 0.7486 - val_loss: 0.5451 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 49/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8666 - accuracy: 0.7957 - loss: 0.4519 - precision: 0.6566 - recall: 0.8354 - val_AUC: 0.7847 - val_accuracy: 0.7486 - val_loss: 0.5444 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 50/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8893 - accuracy: 0.7949 - loss: 0.4223 - precision: 0.6642 - recall: 0.8422 - val_AUC: 0.7843 - val_accuracy: 0.7430 - val_loss: 0.5436 - val_precision: 0.6143 - val_recall: 0.6935\n",
      "Epoch 51/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8793 - accuracy: 0.8037 - loss: 0.4456 - precision: 0.6894 - recall: 0.8410 - val_AUC: 0.7852 - val_accuracy: 0.7486 - val_loss: 0.5442 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 52/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8785 - accuracy: 0.8065 - loss: 0.4362 - precision: 0.6751 - recall: 0.8240 - val_AUC: 0.7866 - val_accuracy: 0.7486 - val_loss: 0.5433 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 53/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8666 - accuracy: 0.7858 - loss: 0.4559 - precision: 0.6573 - recall: 0.7885 - val_AUC: 0.7874 - val_accuracy: 0.7486 - val_loss: 0.5435 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 54/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8637 - accuracy: 0.7779 - loss: 0.4541 - precision: 0.6371 - recall: 0.8062 - val_AUC: 0.7883 - val_accuracy: 0.7430 - val_loss: 0.5427 - val_precision: 0.6111 - val_recall: 0.7097\n",
      "Epoch 55/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8924 - accuracy: 0.8223 - loss: 0.4236 - precision: 0.7184 - recall: 0.8599 - val_AUC: 0.7878 - val_accuracy: 0.7486 - val_loss: 0.5443 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 56/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8824 - accuracy: 0.8185 - loss: 0.4360 - precision: 0.7033 - recall: 0.8304 - val_AUC: 0.7880 - val_accuracy: 0.7486 - val_loss: 0.5428 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 57/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8482 - accuracy: 0.7591 - loss: 0.4859 - precision: 0.6339 - recall: 0.7950 - val_AUC: 0.7885 - val_accuracy: 0.7430 - val_loss: 0.5411 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 58/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8869 - accuracy: 0.7981 - loss: 0.4276 - precision: 0.6763 - recall: 0.8378 - val_AUC: 0.7893 - val_accuracy: 0.7486 - val_loss: 0.5410 - val_precision: 0.6232 - val_recall: 0.6935\n",
      "Epoch 59/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8781 - accuracy: 0.7914 - loss: 0.4414 - precision: 0.6593 - recall: 0.8487 - val_AUC: 0.7892 - val_accuracy: 0.7486 - val_loss: 0.5418 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 60/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8800 - accuracy: 0.8033 - loss: 0.4350 - precision: 0.6850 - recall: 0.8505 - val_AUC: 0.7901 - val_accuracy: 0.7486 - val_loss: 0.5423 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 61/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8496 - accuracy: 0.7555 - loss: 0.4811 - precision: 0.6141 - recall: 0.8353 - val_AUC: 0.7894 - val_accuracy: 0.7486 - val_loss: 0.5411 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 62/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8875 - accuracy: 0.7904 - loss: 0.4265 - precision: 0.6597 - recall: 0.8623 - val_AUC: 0.7911 - val_accuracy: 0.7542 - val_loss: 0.5413 - val_precision: 0.6286 - val_recall: 0.7097\n",
      "Epoch 63/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8798 - accuracy: 0.8047 - loss: 0.4461 - precision: 0.7057 - recall: 0.8343 - val_AUC: 0.7919 - val_accuracy: 0.7486 - val_loss: 0.5408 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 64/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8794 - accuracy: 0.7850 - loss: 0.4358 - precision: 0.6430 - recall: 0.8299 - val_AUC: 0.7916 - val_accuracy: 0.7654 - val_loss: 0.5399 - val_precision: 0.6471 - val_recall: 0.7097\n",
      "Epoch 65/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8878 - accuracy: 0.8092 - loss: 0.4337 - precision: 0.7048 - recall: 0.8166 - val_AUC: 0.7911 - val_accuracy: 0.7598 - val_loss: 0.5403 - val_precision: 0.6377 - val_recall: 0.7097\n",
      "Epoch 66/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9056 - accuracy: 0.8273 - loss: 0.4040 - precision: 0.7420 - recall: 0.8424 - val_AUC: 0.7908 - val_accuracy: 0.7486 - val_loss: 0.5399 - val_precision: 0.6232 - val_recall: 0.6935\n",
      "Epoch 67/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8863 - accuracy: 0.8160 - loss: 0.4262 - precision: 0.7025 - recall: 0.8150 - val_AUC: 0.7909 - val_accuracy: 0.7430 - val_loss: 0.5391 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 68/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8987 - accuracy: 0.8325 - loss: 0.4087 - precision: 0.7272 - recall: 0.8395 - val_AUC: 0.7911 - val_accuracy: 0.7430 - val_loss: 0.5387 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 69/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8769 - accuracy: 0.7971 - loss: 0.4379 - precision: 0.6900 - recall: 0.7978 - val_AUC: 0.7925 - val_accuracy: 0.7486 - val_loss: 0.5392 - val_precision: 0.6232 - val_recall: 0.6935\n",
      "Epoch 70/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8860 - accuracy: 0.8117 - loss: 0.4208 - precision: 0.6795 - recall: 0.8155 - val_AUC: 0.7923 - val_accuracy: 0.7430 - val_loss: 0.5384 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 71/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8867 - accuracy: 0.8108 - loss: 0.4338 - precision: 0.7141 - recall: 0.8298 - val_AUC: 0.7923 - val_accuracy: 0.7486 - val_loss: 0.5390 - val_precision: 0.6197 - val_recall: 0.7097\n",
      "Epoch 72/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9090 - accuracy: 0.8205 - loss: 0.3947 - precision: 0.7120 - recall: 0.8637 - val_AUC: 0.7938 - val_accuracy: 0.7430 - val_loss: 0.5380 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 73/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8939 - accuracy: 0.8114 - loss: 0.4197 - precision: 0.7049 - recall: 0.8291 - val_AUC: 0.7936 - val_accuracy: 0.7318 - val_loss: 0.5384 - val_precision: 0.6000 - val_recall: 0.6774\n",
      "Epoch 74/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8837 - accuracy: 0.7971 - loss: 0.4312 - precision: 0.6818 - recall: 0.8019 - val_AUC: 0.7933 - val_accuracy: 0.7430 - val_loss: 0.5388 - val_precision: 0.6176 - val_recall: 0.6774\n",
      "Epoch 75/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8960 - accuracy: 0.8150 - loss: 0.4088 - precision: 0.6972 - recall: 0.8119 - val_AUC: 0.7922 - val_accuracy: 0.7374 - val_loss: 0.5384 - val_precision: 0.6056 - val_recall: 0.6935\n",
      "Epoch 76/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9050 - accuracy: 0.8210 - loss: 0.3979 - precision: 0.7173 - recall: 0.8414 - val_AUC: 0.7917 - val_accuracy: 0.7374 - val_loss: 0.5379 - val_precision: 0.6087 - val_recall: 0.6774\n",
      "Epoch 77/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9207 - accuracy: 0.8362 - loss: 0.3786 - precision: 0.7366 - recall: 0.8681 - val_AUC: 0.7930 - val_accuracy: 0.7374 - val_loss: 0.5393 - val_precision: 0.6087 - val_recall: 0.6774\n",
      "Epoch 78/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9180 - accuracy: 0.8422 - loss: 0.3729 - precision: 0.7299 - recall: 0.8501 - val_AUC: 0.7931 - val_accuracy: 0.7430 - val_loss: 0.5394 - val_precision: 0.6143 - val_recall: 0.6935\n",
      "Epoch 79/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9110 - accuracy: 0.8199 - loss: 0.3932 - precision: 0.7203 - recall: 0.8446 - val_AUC: 0.7927 - val_accuracy: 0.7263 - val_loss: 0.5390 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 80/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8973 - accuracy: 0.8091 - loss: 0.4150 - precision: 0.7132 - recall: 0.8136 - val_AUC: 0.7919 - val_accuracy: 0.7263 - val_loss: 0.5395 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 81/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8848 - accuracy: 0.7979 - loss: 0.4279 - precision: 0.6762 - recall: 0.8077 - val_AUC: 0.7923 - val_accuracy: 0.7207 - val_loss: 0.5392 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 82/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8947 - accuracy: 0.8086 - loss: 0.4145 - precision: 0.6809 - recall: 0.8263 - val_AUC: 0.7929 - val_accuracy: 0.7151 - val_loss: 0.5390 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 83/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9080 - accuracy: 0.8301 - loss: 0.3896 - precision: 0.7150 - recall: 0.8553 - val_AUC: 0.7929 - val_accuracy: 0.7207 - val_loss: 0.5398 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 84/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9006 - accuracy: 0.8281 - loss: 0.3991 - precision: 0.7192 - recall: 0.8171 - val_AUC: 0.7926 - val_accuracy: 0.7263 - val_loss: 0.5396 - val_precision: 0.5890 - val_recall: 0.6935\n",
      "Epoch 85/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9184 - accuracy: 0.8407 - loss: 0.3718 - precision: 0.7239 - recall: 0.8412 - val_AUC: 0.7916 - val_accuracy: 0.7263 - val_loss: 0.5389 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 86/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9153 - accuracy: 0.8411 - loss: 0.3695 - precision: 0.7234 - recall: 0.8547 - val_AUC: 0.7928 - val_accuracy: 0.7263 - val_loss: 0.5383 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 87/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9079 - accuracy: 0.8323 - loss: 0.3921 - precision: 0.7403 - recall: 0.8152 - val_AUC: 0.7920 - val_accuracy: 0.7207 - val_loss: 0.5403 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 88/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9076 - accuracy: 0.8253 - loss: 0.3972 - precision: 0.7380 - recall: 0.8289 - val_AUC: 0.7926 - val_accuracy: 0.7207 - val_loss: 0.5414 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 89/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9068 - accuracy: 0.8220 - loss: 0.3940 - precision: 0.7060 - recall: 0.8437 - val_AUC: 0.7920 - val_accuracy: 0.7263 - val_loss: 0.5393 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 90/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9000 - accuracy: 0.8160 - loss: 0.3994 - precision: 0.6915 - recall: 0.8165 - val_AUC: 0.7931 - val_accuracy: 0.7207 - val_loss: 0.5389 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 91/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9128 - accuracy: 0.8377 - loss: 0.3902 - precision: 0.7487 - recall: 0.8414 - val_AUC: 0.7915 - val_accuracy: 0.7151 - val_loss: 0.5417 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 92/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8988 - accuracy: 0.8277 - loss: 0.4035 - precision: 0.7241 - recall: 0.8137 - val_AUC: 0.7930 - val_accuracy: 0.7207 - val_loss: 0.5407 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 93/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9229 - accuracy: 0.8418 - loss: 0.3631 - precision: 0.7385 - recall: 0.8520 - val_AUC: 0.7926 - val_accuracy: 0.7207 - val_loss: 0.5398 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 94/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9226 - accuracy: 0.8306 - loss: 0.3659 - precision: 0.7179 - recall: 0.8587 - val_AUC: 0.7933 - val_accuracy: 0.7207 - val_loss: 0.5403 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 95/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9078 - accuracy: 0.8287 - loss: 0.3886 - precision: 0.7184 - recall: 0.8439 - val_AUC: 0.7931 - val_accuracy: 0.7151 - val_loss: 0.5419 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 96/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9125 - accuracy: 0.8287 - loss: 0.3827 - precision: 0.7284 - recall: 0.8316 - val_AUC: 0.7935 - val_accuracy: 0.7151 - val_loss: 0.5413 - val_precision: 0.5753 - val_recall: 0.6774\n",
      "Epoch 97/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9179 - accuracy: 0.8321 - loss: 0.3747 - precision: 0.7263 - recall: 0.8330 - val_AUC: 0.7926 - val_accuracy: 0.7095 - val_loss: 0.5418 - val_precision: 0.5694 - val_recall: 0.6613\n",
      "Epoch 98/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9228 - accuracy: 0.8397 - loss: 0.3604 - precision: 0.7164 - recall: 0.8717 - val_AUC: 0.7940 - val_accuracy: 0.7207 - val_loss: 0.5414 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 99/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9247 - accuracy: 0.8468 - loss: 0.3605 - precision: 0.7348 - recall: 0.8578 - val_AUC: 0.7947 - val_accuracy: 0.7207 - val_loss: 0.5417 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 100/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9096 - accuracy: 0.8193 - loss: 0.3885 - precision: 0.7154 - recall: 0.8370 - val_AUC: 0.7940 - val_accuracy: 0.7207 - val_loss: 0.5424 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 101/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9285 - accuracy: 0.8496 - loss: 0.3438 - precision: 0.7163 - recall: 0.8584 - val_AUC: 0.7934 - val_accuracy: 0.7151 - val_loss: 0.5415 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 102/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9193 - accuracy: 0.8533 - loss: 0.3675 - precision: 0.7639 - recall: 0.8382 - val_AUC: 0.7945 - val_accuracy: 0.7207 - val_loss: 0.5425 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 103/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9227 - accuracy: 0.8442 - loss: 0.3642 - precision: 0.7377 - recall: 0.8779 - val_AUC: 0.7934 - val_accuracy: 0.7151 - val_loss: 0.5431 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 104/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9197 - accuracy: 0.8421 - loss: 0.3663 - precision: 0.7378 - recall: 0.8560 - val_AUC: 0.7928 - val_accuracy: 0.7207 - val_loss: 0.5434 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 105/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9187 - accuracy: 0.8550 - loss: 0.3694 - precision: 0.7474 - recall: 0.8689 - val_AUC: 0.7943 - val_accuracy: 0.7151 - val_loss: 0.5437 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 106/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9247 - accuracy: 0.8490 - loss: 0.3688 - precision: 0.7811 - recall: 0.8374 - val_AUC: 0.7931 - val_accuracy: 0.7151 - val_loss: 0.5453 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 107/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9174 - accuracy: 0.8412 - loss: 0.3739 - precision: 0.7390 - recall: 0.8623 - val_AUC: 0.7934 - val_accuracy: 0.7151 - val_loss: 0.5446 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 108/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9262 - accuracy: 0.8725 - loss: 0.3464 - precision: 0.7761 - recall: 0.8727 - val_AUC: 0.7931 - val_accuracy: 0.7207 - val_loss: 0.5456 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 109/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9235 - accuracy: 0.8338 - loss: 0.3631 - precision: 0.7118 - recall: 0.8609 - val_AUC: 0.7914 - val_accuracy: 0.7207 - val_loss: 0.5458 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 110/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9145 - accuracy: 0.8406 - loss: 0.3756 - precision: 0.7396 - recall: 0.8160 - val_AUC: 0.7938 - val_accuracy: 0.7207 - val_loss: 0.5466 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 111/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9391 - accuracy: 0.8650 - loss: 0.3351 - precision: 0.7844 - recall: 0.8751 - val_AUC: 0.7939 - val_accuracy: 0.7151 - val_loss: 0.5475 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 112/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9228 - accuracy: 0.8482 - loss: 0.3431 - precision: 0.6963 - recall: 0.8444 - val_AUC: 0.7925 - val_accuracy: 0.7207 - val_loss: 0.5467 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 113/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9294 - accuracy: 0.8464 - loss: 0.3471 - precision: 0.7367 - recall: 0.8393 - val_AUC: 0.7935 - val_accuracy: 0.7207 - val_loss: 0.5481 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 114/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9349 - accuracy: 0.8616 - loss: 0.3387 - precision: 0.7782 - recall: 0.8693 - val_AUC: 0.7938 - val_accuracy: 0.7151 - val_loss: 0.5497 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 115/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9231 - accuracy: 0.8477 - loss: 0.3626 - precision: 0.7557 - recall: 0.8450 - val_AUC: 0.7938 - val_accuracy: 0.7207 - val_loss: 0.5494 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 116/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9286 - accuracy: 0.8625 - loss: 0.3516 - precision: 0.7616 - recall: 0.8779 - val_AUC: 0.7939 - val_accuracy: 0.7207 - val_loss: 0.5494 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 117/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9449 - accuracy: 0.8731 - loss: 0.3181 - precision: 0.7778 - recall: 0.8754 - val_AUC: 0.7926 - val_accuracy: 0.7151 - val_loss: 0.5495 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 118/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9372 - accuracy: 0.8814 - loss: 0.3341 - precision: 0.8047 - recall: 0.8646 - val_AUC: 0.7934 - val_accuracy: 0.7207 - val_loss: 0.5503 - val_precision: 0.5857 - val_recall: 0.6613\n",
      "Epoch 119/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9289 - accuracy: 0.8345 - loss: 0.3496 - precision: 0.7243 - recall: 0.8435 - val_AUC: 0.7912 - val_accuracy: 0.7207 - val_loss: 0.5523 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 120/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9529 - accuracy: 0.8805 - loss: 0.3036 - precision: 0.7934 - recall: 0.8805 - val_AUC: 0.7916 - val_accuracy: 0.7263 - val_loss: 0.5512 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 121/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9433 - accuracy: 0.8726 - loss: 0.3186 - precision: 0.7785 - recall: 0.8747 - val_AUC: 0.7914 - val_accuracy: 0.7151 - val_loss: 0.5531 - val_precision: 0.5775 - val_recall: 0.6613\n",
      "Epoch 122/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9381 - accuracy: 0.8614 - loss: 0.3264 - precision: 0.7410 - recall: 0.8797 - val_AUC: 0.7927 - val_accuracy: 0.7263 - val_loss: 0.5508 - val_precision: 0.5915 - val_recall: 0.6774\n",
      "Epoch 123/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9367 - accuracy: 0.8735 - loss: 0.3295 - precision: 0.8046 - recall: 0.8787 - val_AUC: 0.7920 - val_accuracy: 0.7207 - val_loss: 0.5529 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 124/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9340 - accuracy: 0.8721 - loss: 0.3372 - precision: 0.7789 - recall: 0.8743 - val_AUC: 0.7917 - val_accuracy: 0.7207 - val_loss: 0.5529 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 125/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9380 - accuracy: 0.8638 - loss: 0.3276 - precision: 0.7687 - recall: 0.8835 - val_AUC: 0.7912 - val_accuracy: 0.7207 - val_loss: 0.5540 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "Epoch 126/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9348 - accuracy: 0.8684 - loss: 0.3381 - precision: 0.7926 - recall: 0.8797 - val_AUC: 0.7925 - val_accuracy: 0.7207 - val_loss: 0.5538 - val_precision: 0.5833 - val_recall: 0.6774\n",
      "\u001b[1m8/8\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - AUC: 0.8356 - accuracy: 0.7741 - loss: 0.4920 - precision: 0.6795 - recall: 0.7141 \n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n",
      "WARNING:absl:The input size (125) is not a multiple of the number of attention heads (8)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 20ms/step - AUC: 0.6123 - accuracy: 0.5812 - loss: 0.7171 - precision: 0.4361 - recall: 0.6044 - val_AUC: 0.5435 - val_accuracy: 0.5000 - val_loss: 0.8452 - val_precision: 0.3608 - val_recall: 0.5556\n",
      "Epoch 2/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6451 - accuracy: 0.6062 - loss: 0.6880 - precision: 0.4670 - recall: 0.6997 - val_AUC: 0.5621 - val_accuracy: 0.5278 - val_loss: 0.8218 - val_precision: 0.3804 - val_recall: 0.5556\n",
      "Epoch 3/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6448 - accuracy: 0.6063 - loss: 0.6907 - precision: 0.4555 - recall: 0.6773 - val_AUC: 0.5807 - val_accuracy: 0.5444 - val_loss: 0.8013 - val_precision: 0.4000 - val_recall: 0.6032\n",
      "Epoch 4/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.6742 - accuracy: 0.6366 - loss: 0.6651 - precision: 0.4876 - recall: 0.6931 - val_AUC: 0.5927 - val_accuracy: 0.5611 - val_loss: 0.7836 - val_precision: 0.4149 - val_recall: 0.6190\n",
      "Epoch 5/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7183 - accuracy: 0.6736 - loss: 0.6246 - precision: 0.5192 - recall: 0.7275 - val_AUC: 0.6055 - val_accuracy: 0.5500 - val_loss: 0.7682 - val_precision: 0.4062 - val_recall: 0.6190\n",
      "Epoch 6/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.6980 - accuracy: 0.6719 - loss: 0.6521 - precision: 0.5327 - recall: 0.7259 - val_AUC: 0.6161 - val_accuracy: 0.5611 - val_loss: 0.7554 - val_precision: 0.4149 - val_recall: 0.6190\n",
      "Epoch 7/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7304 - accuracy: 0.7014 - loss: 0.6265 - precision: 0.5592 - recall: 0.7391 - val_AUC: 0.6277 - val_accuracy: 0.5778 - val_loss: 0.7428 - val_precision: 0.4301 - val_recall: 0.6349\n",
      "Epoch 8/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7154 - accuracy: 0.6829 - loss: 0.6349 - precision: 0.5329 - recall: 0.7138 - val_AUC: 0.6391 - val_accuracy: 0.6000 - val_loss: 0.7335 - val_precision: 0.4505 - val_recall: 0.6508\n",
      "Epoch 9/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7316 - accuracy: 0.6911 - loss: 0.6122 - precision: 0.5312 - recall: 0.7012 - val_AUC: 0.6499 - val_accuracy: 0.6167 - val_loss: 0.7230 - val_precision: 0.4674 - val_recall: 0.6825\n",
      "Epoch 10/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7538 - accuracy: 0.7020 - loss: 0.5992 - precision: 0.5695 - recall: 0.7101 - val_AUC: 0.6580 - val_accuracy: 0.6222 - val_loss: 0.7160 - val_precision: 0.4725 - val_recall: 0.6825\n",
      "Epoch 11/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7791 - accuracy: 0.7235 - loss: 0.5816 - precision: 0.6158 - recall: 0.7491 - val_AUC: 0.6640 - val_accuracy: 0.6278 - val_loss: 0.7089 - val_precision: 0.4783 - val_recall: 0.6984\n",
      "Epoch 12/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7613 - accuracy: 0.7201 - loss: 0.5915 - precision: 0.5879 - recall: 0.6879 - val_AUC: 0.6713 - val_accuracy: 0.6278 - val_loss: 0.7017 - val_precision: 0.4783 - val_recall: 0.6984\n",
      "Epoch 13/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7578 - accuracy: 0.6897 - loss: 0.5902 - precision: 0.5241 - recall: 0.7387 - val_AUC: 0.6766 - val_accuracy: 0.6389 - val_loss: 0.6958 - val_precision: 0.4889 - val_recall: 0.6984\n",
      "Epoch 14/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7697 - accuracy: 0.6945 - loss: 0.5735 - precision: 0.5226 - recall: 0.7172 - val_AUC: 0.6810 - val_accuracy: 0.6556 - val_loss: 0.6901 - val_precision: 0.5057 - val_recall: 0.6984\n",
      "Epoch 15/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8119 - accuracy: 0.7279 - loss: 0.5375 - precision: 0.5926 - recall: 0.7377 - val_AUC: 0.6853 - val_accuracy: 0.6611 - val_loss: 0.6869 - val_precision: 0.5116 - val_recall: 0.6984\n",
      "Epoch 16/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.7599 - accuracy: 0.6983 - loss: 0.5912 - precision: 0.5533 - recall: 0.7168 - val_AUC: 0.6887 - val_accuracy: 0.6611 - val_loss: 0.6821 - val_precision: 0.5116 - val_recall: 0.6984\n",
      "Epoch 17/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7855 - accuracy: 0.7157 - loss: 0.5678 - precision: 0.5914 - recall: 0.7223 - val_AUC: 0.6913 - val_accuracy: 0.6611 - val_loss: 0.6791 - val_precision: 0.5116 - val_recall: 0.6984\n",
      "Epoch 18/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7739 - accuracy: 0.7151 - loss: 0.5671 - precision: 0.5329 - recall: 0.7297 - val_AUC: 0.6956 - val_accuracy: 0.6611 - val_loss: 0.6743 - val_precision: 0.5119 - val_recall: 0.6825\n",
      "Epoch 19/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.7910 - accuracy: 0.7289 - loss: 0.5554 - precision: 0.5830 - recall: 0.7638 - val_AUC: 0.6998 - val_accuracy: 0.6667 - val_loss: 0.6703 - val_precision: 0.5176 - val_recall: 0.6984\n",
      "Epoch 20/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8263 - accuracy: 0.7537 - loss: 0.5112 - precision: 0.6011 - recall: 0.7830 - val_AUC: 0.7029 - val_accuracy: 0.6556 - val_loss: 0.6680 - val_precision: 0.5060 - val_recall: 0.6667\n",
      "Epoch 21/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8044 - accuracy: 0.7617 - loss: 0.5344 - precision: 0.6219 - recall: 0.7546 - val_AUC: 0.7058 - val_accuracy: 0.6500 - val_loss: 0.6644 - val_precision: 0.5000 - val_recall: 0.6508\n",
      "Epoch 22/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8072 - accuracy: 0.7465 - loss: 0.5398 - precision: 0.6251 - recall: 0.7817 - val_AUC: 0.7067 - val_accuracy: 0.6500 - val_loss: 0.6611 - val_precision: 0.5000 - val_recall: 0.6508\n",
      "Epoch 23/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8148 - accuracy: 0.7358 - loss: 0.5250 - precision: 0.5752 - recall: 0.7826 - val_AUC: 0.7080 - val_accuracy: 0.6556 - val_loss: 0.6584 - val_precision: 0.5063 - val_recall: 0.6349\n",
      "Epoch 24/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8278 - accuracy: 0.7812 - loss: 0.5138 - precision: 0.6581 - recall: 0.7898 - val_AUC: 0.7099 - val_accuracy: 0.6611 - val_loss: 0.6561 - val_precision: 0.5125 - val_recall: 0.6508\n",
      "Epoch 25/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8092 - accuracy: 0.7431 - loss: 0.5325 - precision: 0.5901 - recall: 0.7746 - val_AUC: 0.7114 - val_accuracy: 0.6611 - val_loss: 0.6535 - val_precision: 0.5125 - val_recall: 0.6508\n",
      "Epoch 26/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8221 - accuracy: 0.7501 - loss: 0.5238 - precision: 0.6353 - recall: 0.7823 - val_AUC: 0.7143 - val_accuracy: 0.6611 - val_loss: 0.6522 - val_precision: 0.5125 - val_recall: 0.6508\n",
      "Epoch 27/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8115 - accuracy: 0.7530 - loss: 0.5315 - precision: 0.6175 - recall: 0.7472 - val_AUC: 0.7150 - val_accuracy: 0.6611 - val_loss: 0.6489 - val_precision: 0.5125 - val_recall: 0.6508\n",
      "Epoch 28/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8178 - accuracy: 0.7515 - loss: 0.5194 - precision: 0.6034 - recall: 0.7511 - val_AUC: 0.7167 - val_accuracy: 0.6667 - val_loss: 0.6470 - val_precision: 0.5185 - val_recall: 0.6667\n",
      "Epoch 29/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8327 - accuracy: 0.7744 - loss: 0.5102 - precision: 0.6568 - recall: 0.7772 - val_AUC: 0.7194 - val_accuracy: 0.6722 - val_loss: 0.6450 - val_precision: 0.5250 - val_recall: 0.6667\n",
      "Epoch 30/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8286 - accuracy: 0.7521 - loss: 0.5082 - precision: 0.5908 - recall: 0.7697 - val_AUC: 0.7203 - val_accuracy: 0.6722 - val_loss: 0.6429 - val_precision: 0.5250 - val_recall: 0.6667\n",
      "Epoch 31/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8157 - accuracy: 0.7381 - loss: 0.5414 - precision: 0.6339 - recall: 0.7316 - val_AUC: 0.7213 - val_accuracy: 0.6722 - val_loss: 0.6416 - val_precision: 0.5250 - val_recall: 0.6667\n",
      "Epoch 32/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8483 - accuracy: 0.7615 - loss: 0.4884 - precision: 0.6231 - recall: 0.7832 - val_AUC: 0.7230 - val_accuracy: 0.6722 - val_loss: 0.6389 - val_precision: 0.5250 - val_recall: 0.6667\n",
      "Epoch 33/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8153 - accuracy: 0.7383 - loss: 0.5337 - precision: 0.6090 - recall: 0.7486 - val_AUC: 0.7251 - val_accuracy: 0.6778 - val_loss: 0.6385 - val_precision: 0.5316 - val_recall: 0.6667\n",
      "Epoch 34/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8234 - accuracy: 0.7607 - loss: 0.5157 - precision: 0.6376 - recall: 0.7622 - val_AUC: 0.7271 - val_accuracy: 0.6778 - val_loss: 0.6358 - val_precision: 0.5316 - val_recall: 0.6667\n",
      "Epoch 35/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8406 - accuracy: 0.7765 - loss: 0.4913 - precision: 0.6343 - recall: 0.7958 - val_AUC: 0.7270 - val_accuracy: 0.6778 - val_loss: 0.6346 - val_precision: 0.5316 - val_recall: 0.6667\n",
      "Epoch 36/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8632 - accuracy: 0.7840 - loss: 0.4746 - precision: 0.6841 - recall: 0.8015 - val_AUC: 0.7298 - val_accuracy: 0.6778 - val_loss: 0.6332 - val_precision: 0.5316 - val_recall: 0.6667\n",
      "Epoch 37/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8156 - accuracy: 0.7498 - loss: 0.5296 - precision: 0.6294 - recall: 0.7462 - val_AUC: 0.7300 - val_accuracy: 0.6833 - val_loss: 0.6311 - val_precision: 0.5385 - val_recall: 0.6667\n",
      "Epoch 38/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8384 - accuracy: 0.7617 - loss: 0.5031 - precision: 0.6352 - recall: 0.7802 - val_AUC: 0.7318 - val_accuracy: 0.6833 - val_loss: 0.6300 - val_precision: 0.5385 - val_recall: 0.6667\n",
      "Epoch 39/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8421 - accuracy: 0.7866 - loss: 0.4961 - precision: 0.6746 - recall: 0.8013 - val_AUC: 0.7336 - val_accuracy: 0.6778 - val_loss: 0.6281 - val_precision: 0.5316 - val_recall: 0.6667\n",
      "Epoch 40/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8583 - accuracy: 0.7769 - loss: 0.4661 - precision: 0.6145 - recall: 0.8093 - val_AUC: 0.7351 - val_accuracy: 0.6833 - val_loss: 0.6258 - val_precision: 0.5385 - val_recall: 0.6667\n",
      "Epoch 41/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8440 - accuracy: 0.7536 - loss: 0.5043 - precision: 0.6358 - recall: 0.7934 - val_AUC: 0.7366 - val_accuracy: 0.6778 - val_loss: 0.6253 - val_precision: 0.5325 - val_recall: 0.6508\n",
      "Epoch 42/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8488 - accuracy: 0.7634 - loss: 0.4848 - precision: 0.6161 - recall: 0.8009 - val_AUC: 0.7368 - val_accuracy: 0.6778 - val_loss: 0.6240 - val_precision: 0.5325 - val_recall: 0.6508\n",
      "Epoch 43/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8200 - accuracy: 0.7440 - loss: 0.5231 - precision: 0.6066 - recall: 0.7297 - val_AUC: 0.7380 - val_accuracy: 0.6778 - val_loss: 0.6228 - val_precision: 0.5325 - val_recall: 0.6508\n",
      "Epoch 44/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8390 - accuracy: 0.7745 - loss: 0.4939 - precision: 0.6550 - recall: 0.7452 - val_AUC: 0.7392 - val_accuracy: 0.6833 - val_loss: 0.6214 - val_precision: 0.5395 - val_recall: 0.6508\n",
      "Epoch 45/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8577 - accuracy: 0.7776 - loss: 0.4835 - precision: 0.6715 - recall: 0.7900 - val_AUC: 0.7399 - val_accuracy: 0.6833 - val_loss: 0.6209 - val_precision: 0.5395 - val_recall: 0.6508\n",
      "Epoch 46/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8571 - accuracy: 0.7708 - loss: 0.4708 - precision: 0.6414 - recall: 0.7618 - val_AUC: 0.7411 - val_accuracy: 0.6833 - val_loss: 0.6194 - val_precision: 0.5395 - val_recall: 0.6508\n",
      "Epoch 47/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8542 - accuracy: 0.7749 - loss: 0.4810 - precision: 0.6669 - recall: 0.7515 - val_AUC: 0.7425 - val_accuracy: 0.6889 - val_loss: 0.6181 - val_precision: 0.5467 - val_recall: 0.6508\n",
      "Epoch 48/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8688 - accuracy: 0.7755 - loss: 0.4543 - precision: 0.6156 - recall: 0.8183 - val_AUC: 0.7437 - val_accuracy: 0.6944 - val_loss: 0.6147 - val_precision: 0.5541 - val_recall: 0.6508\n",
      "Epoch 49/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8582 - accuracy: 0.7809 - loss: 0.4712 - precision: 0.6426 - recall: 0.7772 - val_AUC: 0.7435 - val_accuracy: 0.6889 - val_loss: 0.6159 - val_precision: 0.5467 - val_recall: 0.6508\n",
      "Epoch 50/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8409 - accuracy: 0.7537 - loss: 0.4944 - precision: 0.6223 - recall: 0.7509 - val_AUC: 0.7444 - val_accuracy: 0.6944 - val_loss: 0.6148 - val_precision: 0.5541 - val_recall: 0.6508\n",
      "Epoch 51/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8697 - accuracy: 0.7835 - loss: 0.4621 - precision: 0.6682 - recall: 0.7871 - val_AUC: 0.7446 - val_accuracy: 0.6944 - val_loss: 0.6145 - val_precision: 0.5541 - val_recall: 0.6508\n",
      "Epoch 52/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8600 - accuracy: 0.7751 - loss: 0.4718 - precision: 0.6628 - recall: 0.7590 - val_AUC: 0.7450 - val_accuracy: 0.6944 - val_loss: 0.6126 - val_precision: 0.5541 - val_recall: 0.6508\n",
      "Epoch 53/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8624 - accuracy: 0.7850 - loss: 0.4685 - precision: 0.6668 - recall: 0.7847 - val_AUC: 0.7460 - val_accuracy: 0.7000 - val_loss: 0.6117 - val_precision: 0.5616 - val_recall: 0.6508\n",
      "Epoch 54/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8632 - accuracy: 0.7781 - loss: 0.4650 - precision: 0.6540 - recall: 0.7912 - val_AUC: 0.7460 - val_accuracy: 0.6944 - val_loss: 0.6111 - val_precision: 0.5541 - val_recall: 0.6508\n",
      "Epoch 55/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8492 - accuracy: 0.7748 - loss: 0.4757 - precision: 0.6307 - recall: 0.7441 - val_AUC: 0.7468 - val_accuracy: 0.6889 - val_loss: 0.6096 - val_precision: 0.5479 - val_recall: 0.6349\n",
      "Epoch 56/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8651 - accuracy: 0.7890 - loss: 0.4627 - precision: 0.6722 - recall: 0.7907 - val_AUC: 0.7482 - val_accuracy: 0.6944 - val_loss: 0.6086 - val_precision: 0.5556 - val_recall: 0.6349\n",
      "Epoch 57/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8585 - accuracy: 0.7728 - loss: 0.4679 - precision: 0.6416 - recall: 0.7408 - val_AUC: 0.7492 - val_accuracy: 0.6944 - val_loss: 0.6082 - val_precision: 0.5556 - val_recall: 0.6349\n",
      "Epoch 58/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8812 - accuracy: 0.8091 - loss: 0.4423 - precision: 0.7021 - recall: 0.7988 - val_AUC: 0.7503 - val_accuracy: 0.7000 - val_loss: 0.6075 - val_precision: 0.5634 - val_recall: 0.6349\n",
      "Epoch 59/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8378 - accuracy: 0.7522 - loss: 0.4988 - precision: 0.6187 - recall: 0.7466 - val_AUC: 0.7510 - val_accuracy: 0.7000 - val_loss: 0.6064 - val_precision: 0.5634 - val_recall: 0.6349\n",
      "Epoch 60/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8596 - accuracy: 0.7909 - loss: 0.4577 - precision: 0.6485 - recall: 0.7880 - val_AUC: 0.7513 - val_accuracy: 0.7056 - val_loss: 0.6045 - val_precision: 0.5714 - val_recall: 0.6349\n",
      "Epoch 61/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8733 - accuracy: 0.7845 - loss: 0.4551 - precision: 0.6690 - recall: 0.8017 - val_AUC: 0.7514 - val_accuracy: 0.7056 - val_loss: 0.6046 - val_precision: 0.5714 - val_recall: 0.6349\n",
      "Epoch 62/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8789 - accuracy: 0.8124 - loss: 0.4391 - precision: 0.6987 - recall: 0.7986 - val_AUC: 0.7517 - val_accuracy: 0.7056 - val_loss: 0.6044 - val_precision: 0.5714 - val_recall: 0.6349\n",
      "Epoch 63/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8789 - accuracy: 0.8089 - loss: 0.4404 - precision: 0.6873 - recall: 0.8136 - val_AUC: 0.7532 - val_accuracy: 0.7000 - val_loss: 0.6029 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 64/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8778 - accuracy: 0.7996 - loss: 0.4453 - precision: 0.6834 - recall: 0.8156 - val_AUC: 0.7525 - val_accuracy: 0.7000 - val_loss: 0.6030 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 65/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8778 - accuracy: 0.8102 - loss: 0.4384 - precision: 0.6906 - recall: 0.7987 - val_AUC: 0.7540 - val_accuracy: 0.7000 - val_loss: 0.6023 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 66/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8637 - accuracy: 0.8027 - loss: 0.4601 - precision: 0.6657 - recall: 0.8165 - val_AUC: 0.7550 - val_accuracy: 0.7000 - val_loss: 0.6009 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 67/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8664 - accuracy: 0.7940 - loss: 0.4549 - precision: 0.6759 - recall: 0.7754 - val_AUC: 0.7556 - val_accuracy: 0.7000 - val_loss: 0.6004 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 68/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8736 - accuracy: 0.8090 - loss: 0.4514 - precision: 0.7128 - recall: 0.7862 - val_AUC: 0.7563 - val_accuracy: 0.7000 - val_loss: 0.5995 - val_precision: 0.5652 - val_recall: 0.6190\n",
      "Epoch 69/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8919 - accuracy: 0.8256 - loss: 0.4166 - precision: 0.7085 - recall: 0.8433 - val_AUC: 0.7566 - val_accuracy: 0.7056 - val_loss: 0.5986 - val_precision: 0.5735 - val_recall: 0.6190\n",
      "Epoch 70/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8752 - accuracy: 0.8095 - loss: 0.4457 - precision: 0.7043 - recall: 0.7879 - val_AUC: 0.7557 - val_accuracy: 0.7056 - val_loss: 0.5995 - val_precision: 0.5735 - val_recall: 0.6190\n",
      "Epoch 71/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8761 - accuracy: 0.8157 - loss: 0.4498 - precision: 0.7117 - recall: 0.8101 - val_AUC: 0.7561 - val_accuracy: 0.7111 - val_loss: 0.5986 - val_precision: 0.5821 - val_recall: 0.6190\n",
      "Epoch 72/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8686 - accuracy: 0.8087 - loss: 0.4607 - precision: 0.6982 - recall: 0.8154 - val_AUC: 0.7559 - val_accuracy: 0.7111 - val_loss: 0.5977 - val_precision: 0.5821 - val_recall: 0.6190\n",
      "Epoch 73/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8718 - accuracy: 0.8061 - loss: 0.4514 - precision: 0.6931 - recall: 0.8119 - val_AUC: 0.7564 - val_accuracy: 0.7167 - val_loss: 0.5970 - val_precision: 0.5909 - val_recall: 0.6190\n",
      "Epoch 74/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9039 - accuracy: 0.8458 - loss: 0.4017 - precision: 0.7518 - recall: 0.8336 - val_AUC: 0.7572 - val_accuracy: 0.7167 - val_loss: 0.5967 - val_precision: 0.5909 - val_recall: 0.6190\n",
      "Epoch 75/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8886 - accuracy: 0.8261 - loss: 0.4329 - precision: 0.7295 - recall: 0.8516 - val_AUC: 0.7586 - val_accuracy: 0.7167 - val_loss: 0.5957 - val_precision: 0.5909 - val_recall: 0.6190\n",
      "Epoch 76/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8753 - accuracy: 0.8142 - loss: 0.4503 - precision: 0.7255 - recall: 0.7992 - val_AUC: 0.7585 - val_accuracy: 0.7111 - val_loss: 0.5960 - val_precision: 0.5821 - val_recall: 0.6190\n",
      "Epoch 77/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8854 - accuracy: 0.8180 - loss: 0.4317 - precision: 0.7104 - recall: 0.8193 - val_AUC: 0.7608 - val_accuracy: 0.7222 - val_loss: 0.5939 - val_precision: 0.6000 - val_recall: 0.6190\n",
      "Epoch 78/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8765 - accuracy: 0.7917 - loss: 0.4439 - precision: 0.6631 - recall: 0.8182 - val_AUC: 0.7618 - val_accuracy: 0.7167 - val_loss: 0.5940 - val_precision: 0.5938 - val_recall: 0.6032\n",
      "Epoch 79/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8895 - accuracy: 0.8207 - loss: 0.4236 - precision: 0.7106 - recall: 0.8321 - val_AUC: 0.7618 - val_accuracy: 0.7111 - val_loss: 0.5939 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 80/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8983 - accuracy: 0.8279 - loss: 0.4043 - precision: 0.7009 - recall: 0.8486 - val_AUC: 0.7608 - val_accuracy: 0.7167 - val_loss: 0.5930 - val_precision: 0.5938 - val_recall: 0.6032\n",
      "Epoch 81/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8957 - accuracy: 0.8125 - loss: 0.4088 - precision: 0.6960 - recall: 0.8095 - val_AUC: 0.7613 - val_accuracy: 0.7111 - val_loss: 0.5938 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 82/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8841 - accuracy: 0.8180 - loss: 0.4279 - precision: 0.6998 - recall: 0.8155 - val_AUC: 0.7619 - val_accuracy: 0.7111 - val_loss: 0.5926 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 83/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8827 - accuracy: 0.8049 - loss: 0.4319 - precision: 0.6769 - recall: 0.8189 - val_AUC: 0.7622 - val_accuracy: 0.7111 - val_loss: 0.5925 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 84/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9023 - accuracy: 0.8287 - loss: 0.4013 - precision: 0.7269 - recall: 0.8340 - val_AUC: 0.7627 - val_accuracy: 0.7111 - val_loss: 0.5922 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 85/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8962 - accuracy: 0.8246 - loss: 0.4110 - precision: 0.7006 - recall: 0.8606 - val_AUC: 0.7635 - val_accuracy: 0.7111 - val_loss: 0.5916 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 86/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.8897 - accuracy: 0.8067 - loss: 0.4195 - precision: 0.6780 - recall: 0.7827 - val_AUC: 0.7637 - val_accuracy: 0.7111 - val_loss: 0.5913 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 87/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8666 - accuracy: 0.8078 - loss: 0.4540 - precision: 0.6887 - recall: 0.8104 - val_AUC: 0.7637 - val_accuracy: 0.7111 - val_loss: 0.5920 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 88/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9142 - accuracy: 0.8425 - loss: 0.3769 - precision: 0.7248 - recall: 0.8665 - val_AUC: 0.7639 - val_accuracy: 0.7111 - val_loss: 0.5908 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 89/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9029 - accuracy: 0.8317 - loss: 0.3938 - precision: 0.6976 - recall: 0.8706 - val_AUC: 0.7646 - val_accuracy: 0.7111 - val_loss: 0.5900 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 90/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9172 - accuracy: 0.8469 - loss: 0.3701 - precision: 0.7422 - recall: 0.8519 - val_AUC: 0.7654 - val_accuracy: 0.7167 - val_loss: 0.5900 - val_precision: 0.5909 - val_recall: 0.6190\n",
      "Epoch 91/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8902 - accuracy: 0.8117 - loss: 0.4176 - precision: 0.6844 - recall: 0.8410 - val_AUC: 0.7642 - val_accuracy: 0.7111 - val_loss: 0.5903 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 92/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8832 - accuracy: 0.8120 - loss: 0.4375 - precision: 0.7264 - recall: 0.8049 - val_AUC: 0.7648 - val_accuracy: 0.7111 - val_loss: 0.5918 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 93/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9085 - accuracy: 0.8332 - loss: 0.3913 - precision: 0.7387 - recall: 0.8731 - val_AUC: 0.7659 - val_accuracy: 0.7111 - val_loss: 0.5904 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 94/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8837 - accuracy: 0.8019 - loss: 0.4344 - precision: 0.7014 - recall: 0.7933 - val_AUC: 0.7654 - val_accuracy: 0.7056 - val_loss: 0.5896 - val_precision: 0.5781 - val_recall: 0.5873\n",
      "Epoch 95/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9050 - accuracy: 0.8371 - loss: 0.3913 - precision: 0.7371 - recall: 0.8491 - val_AUC: 0.7645 - val_accuracy: 0.7056 - val_loss: 0.5902 - val_precision: 0.5781 - val_recall: 0.5873\n",
      "Epoch 96/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8918 - accuracy: 0.8115 - loss: 0.4260 - precision: 0.7262 - recall: 0.7989 - val_AUC: 0.7652 - val_accuracy: 0.7111 - val_loss: 0.5901 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 97/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9034 - accuracy: 0.8358 - loss: 0.3876 - precision: 0.7149 - recall: 0.8197 - val_AUC: 0.7659 - val_accuracy: 0.7167 - val_loss: 0.5889 - val_precision: 0.5938 - val_recall: 0.6032\n",
      "Epoch 98/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9012 - accuracy: 0.8301 - loss: 0.4100 - precision: 0.7448 - recall: 0.8367 - val_AUC: 0.7658 - val_accuracy: 0.7111 - val_loss: 0.5891 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 99/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9073 - accuracy: 0.8420 - loss: 0.3890 - precision: 0.7490 - recall: 0.8419 - val_AUC: 0.7671 - val_accuracy: 0.7111 - val_loss: 0.5894 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 100/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9173 - accuracy: 0.8484 - loss: 0.3723 - precision: 0.7650 - recall: 0.8598 - val_AUC: 0.7671 - val_accuracy: 0.7167 - val_loss: 0.5892 - val_precision: 0.5909 - val_recall: 0.6190\n",
      "Epoch 101/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9005 - accuracy: 0.8452 - loss: 0.3984 - precision: 0.7502 - recall: 0.8266 - val_AUC: 0.7674 - val_accuracy: 0.7111 - val_loss: 0.5887 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 102/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9045 - accuracy: 0.8339 - loss: 0.3909 - precision: 0.7211 - recall: 0.8553 - val_AUC: 0.7652 - val_accuracy: 0.7111 - val_loss: 0.5884 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 103/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9128 - accuracy: 0.8477 - loss: 0.3779 - precision: 0.7331 - recall: 0.8574 - val_AUC: 0.7654 - val_accuracy: 0.7167 - val_loss: 0.5891 - val_precision: 0.5938 - val_recall: 0.6032\n",
      "Epoch 104/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9196 - accuracy: 0.8612 - loss: 0.3649 - precision: 0.7529 - recall: 0.8977 - val_AUC: 0.7682 - val_accuracy: 0.7111 - val_loss: 0.5901 - val_precision: 0.5846 - val_recall: 0.6032\n",
      "Epoch 105/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.8885 - accuracy: 0.8245 - loss: 0.4250 - precision: 0.7186 - recall: 0.8235 - val_AUC: 0.7671 - val_accuracy: 0.7278 - val_loss: 0.5884 - val_precision: 0.6094 - val_recall: 0.6190\n",
      "Epoch 106/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9083 - accuracy: 0.8482 - loss: 0.3829 - precision: 0.7399 - recall: 0.8597 - val_AUC: 0.7668 - val_accuracy: 0.7278 - val_loss: 0.5882 - val_precision: 0.6094 - val_recall: 0.6190\n",
      "Epoch 107/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9111 - accuracy: 0.8420 - loss: 0.3868 - precision: 0.7561 - recall: 0.8584 - val_AUC: 0.7669 - val_accuracy: 0.7222 - val_loss: 0.5908 - val_precision: 0.6000 - val_recall: 0.6190\n",
      "Epoch 108/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9173 - accuracy: 0.8484 - loss: 0.3697 - precision: 0.7436 - recall: 0.8737 - val_AUC: 0.7652 - val_accuracy: 0.7222 - val_loss: 0.5898 - val_precision: 0.6000 - val_recall: 0.6190\n",
      "Epoch 109/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9277 - accuracy: 0.8594 - loss: 0.3489 - precision: 0.7429 - recall: 0.8712 - val_AUC: 0.7681 - val_accuracy: 0.7278 - val_loss: 0.5888 - val_precision: 0.6061 - val_recall: 0.6349\n",
      "Epoch 110/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9273 - accuracy: 0.8603 - loss: 0.3504 - precision: 0.7664 - recall: 0.8948 - val_AUC: 0.7669 - val_accuracy: 0.7278 - val_loss: 0.5903 - val_precision: 0.6061 - val_recall: 0.6349\n",
      "Epoch 111/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9187 - accuracy: 0.8484 - loss: 0.3610 - precision: 0.7255 - recall: 0.8693 - val_AUC: 0.7665 - val_accuracy: 0.7278 - val_loss: 0.5904 - val_precision: 0.6094 - val_recall: 0.6190\n",
      "Epoch 112/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9239 - accuracy: 0.8707 - loss: 0.3535 - precision: 0.7869 - recall: 0.8580 - val_AUC: 0.7673 - val_accuracy: 0.7333 - val_loss: 0.5902 - val_precision: 0.6154 - val_recall: 0.6349\n",
      "Epoch 113/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9146 - accuracy: 0.8508 - loss: 0.3739 - precision: 0.7376 - recall: 0.8704 - val_AUC: 0.7683 - val_accuracy: 0.7389 - val_loss: 0.5892 - val_precision: 0.6250 - val_recall: 0.6349\n",
      "Epoch 114/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9219 - accuracy: 0.8531 - loss: 0.3547 - precision: 0.7336 - recall: 0.8621 - val_AUC: 0.7673 - val_accuracy: 0.7222 - val_loss: 0.5893 - val_precision: 0.6000 - val_recall: 0.6190\n",
      "Epoch 115/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9243 - accuracy: 0.8699 - loss: 0.3549 - precision: 0.7750 - recall: 0.8963 - val_AUC: 0.7686 - val_accuracy: 0.7333 - val_loss: 0.5921 - val_precision: 0.6154 - val_recall: 0.6349\n",
      "Epoch 116/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9252 - accuracy: 0.8717 - loss: 0.3585 - precision: 0.7825 - recall: 0.8727 - val_AUC: 0.7691 - val_accuracy: 0.7389 - val_loss: 0.5908 - val_precision: 0.6250 - val_recall: 0.6349\n",
      "Epoch 117/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9333 - accuracy: 0.8658 - loss: 0.3335 - precision: 0.7407 - recall: 0.9157 - val_AUC: 0.7671 - val_accuracy: 0.7333 - val_loss: 0.5900 - val_precision: 0.6190 - val_recall: 0.6190\n",
      "Epoch 118/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9094 - accuracy: 0.8404 - loss: 0.3778 - precision: 0.7194 - recall: 0.8633 - val_AUC: 0.7688 - val_accuracy: 0.7389 - val_loss: 0.5894 - val_precision: 0.6290 - val_recall: 0.6190\n",
      "Epoch 119/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9250 - accuracy: 0.8499 - loss: 0.3540 - precision: 0.7567 - recall: 0.8759 - val_AUC: 0.7690 - val_accuracy: 0.7222 - val_loss: 0.5895 - val_precision: 0.6000 - val_recall: 0.6190\n",
      "Epoch 120/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9350 - accuracy: 0.8656 - loss: 0.3315 - precision: 0.7514 - recall: 0.8851 - val_AUC: 0.7692 - val_accuracy: 0.7500 - val_loss: 0.5898 - val_precision: 0.6452 - val_recall: 0.6349\n",
      "Epoch 121/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9282 - accuracy: 0.8648 - loss: 0.3503 - precision: 0.7523 - recall: 0.9063 - val_AUC: 0.7695 - val_accuracy: 0.7333 - val_loss: 0.5910 - val_precision: 0.6190 - val_recall: 0.6190\n",
      "Epoch 122/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9327 - accuracy: 0.8628 - loss: 0.3340 - precision: 0.7523 - recall: 0.8890 - val_AUC: 0.7692 - val_accuracy: 0.7500 - val_loss: 0.5925 - val_precision: 0.6452 - val_recall: 0.6349\n",
      "Epoch 123/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9280 - accuracy: 0.8599 - loss: 0.3534 - precision: 0.7854 - recall: 0.8636 - val_AUC: 0.7687 - val_accuracy: 0.7444 - val_loss: 0.5939 - val_precision: 0.6308 - val_recall: 0.6508\n",
      "Epoch 124/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9301 - accuracy: 0.8704 - loss: 0.3448 - precision: 0.7963 - recall: 0.8605 - val_AUC: 0.7665 - val_accuracy: 0.7333 - val_loss: 0.5936 - val_precision: 0.6190 - val_recall: 0.6190\n",
      "Epoch 125/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9369 - accuracy: 0.8745 - loss: 0.3237 - precision: 0.7569 - recall: 0.9304 - val_AUC: 0.7690 - val_accuracy: 0.7444 - val_loss: 0.5925 - val_precision: 0.6393 - val_recall: 0.6190\n",
      "Epoch 126/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9293 - accuracy: 0.8774 - loss: 0.3429 - precision: 0.8008 - recall: 0.8924 - val_AUC: 0.7694 - val_accuracy: 0.7500 - val_loss: 0.5930 - val_precision: 0.6452 - val_recall: 0.6349\n",
      "Epoch 127/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9285 - accuracy: 0.8764 - loss: 0.3426 - precision: 0.7783 - recall: 0.8887 - val_AUC: 0.7698 - val_accuracy: 0.7444 - val_loss: 0.5923 - val_precision: 0.6393 - val_recall: 0.6190\n",
      "Epoch 128/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9438 - accuracy: 0.8905 - loss: 0.3085 - precision: 0.8059 - recall: 0.9082 - val_AUC: 0.7685 - val_accuracy: 0.7389 - val_loss: 0.5940 - val_precision: 0.6333 - val_recall: 0.6032\n",
      "Epoch 129/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9377 - accuracy: 0.8833 - loss: 0.3204 - precision: 0.7975 - recall: 0.8978 - val_AUC: 0.7693 - val_accuracy: 0.7500 - val_loss: 0.5924 - val_precision: 0.6552 - val_recall: 0.6032\n",
      "Epoch 130/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9226 - accuracy: 0.8614 - loss: 0.3579 - precision: 0.7713 - recall: 0.8589 - val_AUC: 0.7678 - val_accuracy: 0.7500 - val_loss: 0.5941 - val_precision: 0.6552 - val_recall: 0.6032\n",
      "Epoch 131/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9339 - accuracy: 0.8777 - loss: 0.3371 - precision: 0.8064 - recall: 0.8654 - val_AUC: 0.7704 - val_accuracy: 0.7556 - val_loss: 0.5925 - val_precision: 0.6667 - val_recall: 0.6032\n",
      "Epoch 132/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9367 - accuracy: 0.8682 - loss: 0.3238 - precision: 0.7826 - recall: 0.8657 - val_AUC: 0.7693 - val_accuracy: 0.7611 - val_loss: 0.5959 - val_precision: 0.6724 - val_recall: 0.6190\n",
      "Epoch 133/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9317 - accuracy: 0.8756 - loss: 0.3276 - precision: 0.7644 - recall: 0.8931 - val_AUC: 0.7708 - val_accuracy: 0.7611 - val_loss: 0.5925 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 134/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9368 - accuracy: 0.8782 - loss: 0.3265 - precision: 0.7898 - recall: 0.8964 - val_AUC: 0.7706 - val_accuracy: 0.7611 - val_loss: 0.5947 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 135/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9508 - accuracy: 0.8859 - loss: 0.2991 - precision: 0.8123 - recall: 0.8954 - val_AUC: 0.7705 - val_accuracy: 0.7611 - val_loss: 0.5939 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 136/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9352 - accuracy: 0.8567 - loss: 0.3364 - precision: 0.7820 - recall: 0.8516 - val_AUC: 0.7711 - val_accuracy: 0.7611 - val_loss: 0.5953 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 137/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9322 - accuracy: 0.8698 - loss: 0.3280 - precision: 0.7571 - recall: 0.8936 - val_AUC: 0.7701 - val_accuracy: 0.7611 - val_loss: 0.5959 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 138/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9398 - accuracy: 0.8836 - loss: 0.3196 - precision: 0.8091 - recall: 0.8683 - val_AUC: 0.7724 - val_accuracy: 0.7611 - val_loss: 0.5950 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 139/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9321 - accuracy: 0.8696 - loss: 0.3292 - precision: 0.7693 - recall: 0.8768 - val_AUC: 0.7719 - val_accuracy: 0.7611 - val_loss: 0.5938 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 140/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9505 - accuracy: 0.8931 - loss: 0.2926 - precision: 0.8161 - recall: 0.8899 - val_AUC: 0.7745 - val_accuracy: 0.7611 - val_loss: 0.5949 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 141/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9440 - accuracy: 0.8776 - loss: 0.3001 - precision: 0.7594 - recall: 0.9184 - val_AUC: 0.7730 - val_accuracy: 0.7611 - val_loss: 0.5961 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 142/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9466 - accuracy: 0.8925 - loss: 0.3040 - precision: 0.8148 - recall: 0.8993 - val_AUC: 0.7726 - val_accuracy: 0.7611 - val_loss: 0.5979 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 143/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9532 - accuracy: 0.8992 - loss: 0.2851 - precision: 0.8148 - recall: 0.9068 - val_AUC: 0.7735 - val_accuracy: 0.7611 - val_loss: 0.5965 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 144/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9631 - accuracy: 0.9038 - loss: 0.2658 - precision: 0.8352 - recall: 0.9205 - val_AUC: 0.7731 - val_accuracy: 0.7611 - val_loss: 0.5973 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 145/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9486 - accuracy: 0.8814 - loss: 0.2893 - precision: 0.7733 - recall: 0.8982 - val_AUC: 0.7719 - val_accuracy: 0.7611 - val_loss: 0.5977 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 146/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9662 - accuracy: 0.9218 - loss: 0.2519 - precision: 0.8578 - recall: 0.9345 - val_AUC: 0.7724 - val_accuracy: 0.7611 - val_loss: 0.6007 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 147/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9474 - accuracy: 0.9046 - loss: 0.2926 - precision: 0.8449 - recall: 0.8951 - val_AUC: 0.7751 - val_accuracy: 0.7611 - val_loss: 0.5982 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 148/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9511 - accuracy: 0.8973 - loss: 0.2818 - precision: 0.7994 - recall: 0.9318 - val_AUC: 0.7745 - val_accuracy: 0.7611 - val_loss: 0.5995 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 149/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9476 - accuracy: 0.8982 - loss: 0.2967 - precision: 0.8207 - recall: 0.9116 - val_AUC: 0.7757 - val_accuracy: 0.7611 - val_loss: 0.5991 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 150/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9564 - accuracy: 0.9002 - loss: 0.2732 - precision: 0.8253 - recall: 0.9144 - val_AUC: 0.7719 - val_accuracy: 0.7500 - val_loss: 0.6012 - val_precision: 0.6607 - val_recall: 0.5873\n",
      "Epoch 151/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 7ms/step - AUC: 0.9555 - accuracy: 0.8980 - loss: 0.2760 - precision: 0.8170 - recall: 0.9235 - val_AUC: 0.7751 - val_accuracy: 0.7611 - val_loss: 0.6006 - val_precision: 0.6786 - val_recall: 0.6032\n",
      "Epoch 152/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 6ms/step - AUC: 0.9568 - accuracy: 0.9097 - loss: 0.2703 - precision: 0.8323 - recall: 0.9002 - val_AUC: 0.7747 - val_accuracy: 0.7556 - val_loss: 0.6001 - val_precision: 0.6727 - val_recall: 0.5873\n",
      "Epoch 153/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9556 - accuracy: 0.9164 - loss: 0.2689 - precision: 0.8439 - recall: 0.9376 - val_AUC: 0.7775 - val_accuracy: 0.7500 - val_loss: 0.6000 - val_precision: 0.6607 - val_recall: 0.5873\n",
      "Epoch 154/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9632 - accuracy: 0.9171 - loss: 0.2566 - precision: 0.8440 - recall: 0.9325 - val_AUC: 0.7734 - val_accuracy: 0.7500 - val_loss: 0.6020 - val_precision: 0.6607 - val_recall: 0.5873\n",
      "Epoch 155/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9446 - accuracy: 0.8911 - loss: 0.2974 - precision: 0.7916 - recall: 0.9170 - val_AUC: 0.7752 - val_accuracy: 0.7556 - val_loss: 0.6018 - val_precision: 0.6727 - val_recall: 0.5873\n",
      "Epoch 156/1000\n",
      "\u001b[1m23/23\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 5ms/step - AUC: 0.9452 - accuracy: 0.8955 - loss: 0.2934 - precision: 0.8069 - recall: 0.9151 - val_AUC: 0.7759 - val_accuracy: 0.7556 - val_loss: 0.6008 - val_precision: 0.6667 - val_recall: 0.6032\n",
      "\u001b[1m7/7\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - AUC: 0.7951 - accuracy: 0.7241 - loss: 0.5725 - precision: 0.6196 - recall: 0.7177 \n"
     ]
    }
   ],
   "source": [
    "from utils.prepare_data import normalizeData, patientsToNumpy\n",
    "from constants import CATEGORICAL_MEASURES\n",
    "import numpy as np\n",
    "from tensorflow.keras.callbacks import EarlyStopping\n",
    "\n",
    "\n",
    "loses = []\n",
    "aucs = []\n",
    "accuracies = []\n",
    "precisions = []\n",
    "recals = []\n",
    "\n",
    "train_loss_list = []\n",
    "val_loss_list = []\n",
    "models = []\n",
    "\n",
    "for i, (trainPatients, valPatients, testPatients) in enumerate(trainValTest()):\n",
    "    npTrainX, dataNormalizer = patientsToNumpy(\n",
    "        trainPatients,\n",
    "        hoursPerWindow,\n",
    "        timeSeriesOnly=True,\n",
    "        fromHour=0,\n",
    "        toHour=12,\n",
    "    )\n",
    "\n",
    "    npTestX, *_ = patientsToNumpy(\n",
    "        testPatients,\n",
    "        hoursPerWindow,\n",
    "        timeSeriesOnly=True,\n",
    "        fromHour=0,\n",
    "        toHour=12,\n",
    "        dataNormalizer=dataNormalizer,\n",
    "        isTrainPatients=False,\n",
    "    )\n",
    "\n",
    "    npValX, *_ = patientsToNumpy(\n",
    "        valPatients,\n",
    "        hoursPerWindow,\n",
    "        timeSeriesOnly=True,\n",
    "        fromHour=0,\n",
    "        toHour=12,\n",
    "        dataNormalizer=dataNormalizer,\n",
    "        isTrainPatients=False,\n",
    "    )\n",
    "\n",
    "    npTrainX = np.nan_to_num(npTrainX, nan=0)\n",
    "    npTestX = np.nan_to_num(npTestX, nan=0)\n",
    "    npValX = np.nan_to_num(npValX, nan=0)\n",
    "\n",
    "    ################### Static ###################\n",
    "    staticTrainX = trainPatients.getMeasuresBetween(measureTypes=\"static\")\n",
    "    staticTestX = testPatients.getMeasuresBetween(measureTypes=\"static\")\n",
    "    staticValX = valPatients.getMeasuresBetween(measureTypes=\"static\")\n",
    "\n",
    "    staticTrainX = staticTrainX.drop(columns=[\"subject_id\", \"hadm_id\", \"stay_id\", \"akd\"])\n",
    "    staticTestX = staticTestX.drop(columns=[\"subject_id\", \"hadm_id\", \"stay_id\", \"akd\"])\n",
    "    staticValX = staticValX.drop(columns=[\"subject_id\", \"hadm_id\", \"stay_id\", \"akd\"])\n",
    "\n",
    "    staticTrainX, staticTestX, staticValX = normalizeData(\n",
    "        staticTrainX, staticTestX, staticValX\n",
    "    )\n",
    "\n",
    "    staticLen = len(staticTrainX.columns)\n",
    "\n",
    "    staticTrainX = staticTrainX.to_numpy().astype(np.float32)\n",
    "    staticTestX = staticTestX.to_numpy().astype(np.float32)\n",
    "    staticValX = staticValX.to_numpy().astype(np.float32) # type: ignore\n",
    "\n",
    "    staticTrainX = np.nan_to_num(staticTrainX, nan=0)\n",
    "    staticTestX = np.nan_to_num(staticTestX, nan=0)\n",
    "    staticValX = np.nan_to_num(staticValX, nan=0)\n",
    "\n",
    "    ################### labels ###################\n",
    "    trainY = [p.akdPositive for p in trainPatients]\n",
    "    testY = [p.akdPositive for p in testPatients]\n",
    "    valY = [p.akdPositive for p in valPatients]\n",
    "\n",
    "    neg, pos = np.bincount(trainY)\n",
    "    weight0 = (1 / neg) * (len(trainY)) / 2.0\n",
    "    weight1 = (1 / pos) * (len(trainY)) / 2.0\n",
    "    weight = {0: weight0, 1: weight1}\n",
    "\n",
    "    early_stopping = EarlyStopping(\n",
    "        monitor=\"val_loss\", patience=50, restore_best_weights=True\n",
    "    )\n",
    "\n",
    "    model = createModel2(npTrainX.shape[1], npTrainX.shape[2], staticLen)\n",
    "\n",
    "    history = model.fit(\n",
    "            [npTrainX, staticTrainX],\n",
    "            np.array(trainY),\n",
    "            epochs=1000,\n",
    "            batch_size=32,\n",
    "            validation_data=([npValX, staticValX], np.array(valY)),\n",
    "            class_weight=weight,\n",
    "            callbacks=[early_stopping],\n",
    "        )\n",
    "\n",
    "    loss, auc, accuracy, precison, recal = model.evaluate([npTestX, staticTestX], np.array(testY))\n",
    "\n",
    "    # model.save(f\"result/lstm_split_{i}.h5\")\n",
    "\n",
    "    models.append(model)\n",
    "    loses.append(loss)\n",
    "    aucs.append(auc)\n",
    "    accuracies.append(accuracy)\n",
    "    precisions.append(precison)\n",
    "    recals.append(recal)\n",
    "\n",
    "    train_loss_list.append(history.history[\"loss\"])\n",
    "    val_loss_list.append(history.history[\"val_loss\"])\n",
    "\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loses: [0.5932092070579529, 0.530018150806427, 0.5766720175743103, 0.5491611361503601, 0.6139375567436218] 0.5725996136665344 0.03004233199595232\n",
      "AUCs: [0.7870661616325378, 0.7984039187431335, 0.7621813416481018, 0.7957777380943298, 0.7770020961761475] 0.7840862512588501 0.013270845118083097\n",
      "Accuracies: [0.7244444489479065, 0.746666669845581, 0.7599999904632568, 0.7377777695655823, 0.7142857313156128] 0.7366349220275878 0.016099456184412407\n",
      "Precisions: [0.699999988079071, 0.6265060305595398, 0.7358490824699402, 0.6136363744735718, 0.5686274766921997] 0.6489237904548645 0.06059709571404675\n",
      "Recals: [0.3589743673801422, 0.6666666865348816, 0.49367088079452515, 0.6835442781448364, 0.7435897588729858] 0.5892891943454742 0.14203361766810968\n"
     ]
    }
   ],
   "source": [
    "print(\"Loses:\", loses, np.mean(loses), np.std(loses))\n",
    "print(\"AUCs:\", aucs, np.mean(aucs), np.std(aucs))\n",
    "print(\"Accuracies:\", accuracies, np.mean(accuracies), np.std(accuracies))\n",
    "print(\"Precisions:\", precisions, np.mean(precisions), np.std(precisions))\n",
    "print(\"Recals:\", recals, np.mean(recals), np.std(recals))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+kAAAIjCAYAAAB/OVoZAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8hTgPZAAAACXBIWXMAAA9hAAAPYQGoP6dpAAEAAElEQVR4nOzdd3iUVdrH8e+k90pCCmmEkITekSawIm1pitjFIKLIsqyyNEWKsioiKKIrKi8QRBQFXEUEERHUBRYQBAsQCCWFAKGEhCSkz/vHkJGYQhISksDvc11zkXnmPOfcM4Fd7zn3OcdgNBqNiIiIiIiIiEiNs6jpAERERERERETEREm6iIiIiIiISC2hJF1ERERERESkllCSLiIiIiIiIlJLKEkXERERERERqSWUpIuIiIiIiIjUEkrSRURERERERGoJJekiIiIiIiIitYSSdBEREREREZFaQkm6iIjUGSdOnMBgMBAdHW2+NnPmTAwGQ7nuNxgMzJw5s0pj6tGjBz169KjSPm82u3fvpnPnzjg6OmIwGNi3b19Nh1TlevToQbNmzWo6DACCg4OJioq6Zrvo6GgMBgMnTpyo9phERKT8lKSLiEi1GDRoEA4ODly6dKnUNg899BA2NjacP3/+BkZWcQcOHGDmzJm1KpnZunUrBoOB1atX13QoZcrNzWXYsGFcuHCBN954g+XLlxMUFFRt4xV+LqU9Vq5cWW1jX4/g4OBSY87Kyqrp8ERE5AayqukARETk5vTQQw/x5Zdf8p///Ifhw4cXez0zM5MvvviCvn374unpWelxnn/+eaZMmXI9oV7TgQMHeOGFF+jRowfBwcFFXvvmm2+qdey67ujRo8TFxbFo0SIef/zxGzbuuHHjaN++fbHrnTp1umExVFSrVq345z//Wey6jY1NDUQjIiI1RUm6iIhUi0GDBuHs7MxHH31UYpL+xRdfkJGRwUMPPXRd41hZWWFlVXP/d6YEqmzJyckAuLm5VVmfGRkZODo6ltmmW7du3HPPPVU25o3g7+/Pww8/XNNhiIhIDVO5u4iIVAt7e3vuvvtuNm/ebE7UrvbRRx/h7OzMoEGDuHDhAhMmTKB58+Y4OTnh4uJCv3792L9//zXHKWlNenZ2Ns888wxeXl7mMRITE4vdGxcXx5gxYwgPD8fe3h5PT0+GDRtWpKw9OjqaYcOGAdCzZ09zCfLWrVuBktekJycnM3LkSOrXr4+dnR0tW7Zk2bJlRdoUrq+fO3cu77//PqGhodja2tK+fXt27959zfddXseOHWPYsGF4eHjg4ODAbbfdxldffVWs3VtvvUXTpk1xcHDA3d2ddu3a8dFHH5lfv3TpEk8//TTBwcHY2tri7e3NnXfeyd69e0sdOyoqiu7duwMwbNgwDAZDkc/qu+++o1u3bjg6OuLm5sbgwYM5ePBgkT4Kf78HDhzgwQcfxN3dna5du17np2KydOlS/vKXv+Dt7Y2trS1NmjRh4cKFJbbdsGED3bt3x9nZGRcXF9q3b1/k8yl04MABevbsiYODA/7+/syZM6dKYgXTlxP//Oc/CQgIwNbWlvDwcObOnYvRaLzmvb///jt/+ctfsLe3p0GDBvzrX/+ioKCgymITEZGqo5l0ERGpNg899BDLli3j008/ZezYsebrFy5cYOPGjTzwwAPY29vz+++/8/nnnzNs2DBCQkI4c+YM7733Ht27d+fAgQP4+flVaNzHH3+cDz/8kAcffJDOnTvz3Xff8de//rVYu927d7N9+3buv/9+GjRowIkTJ1i4cCE9evTgwIEDODg4cPvttzNu3DgWLFjAc889R2RkJID5zz+7fPkyPXr0IDY2lrFjxxISEsKqVauIiori4sWL/OMf/yjS/qOPPuLSpUs8+eSTGAwG5syZw913382xY8ewtrau0Pv+szNnztC5c2cyMzMZN24cnp6eLFu2jEGDBrF69WruuusuABYtWsS4ceO45557+Mc//kFWVha//PILO3fu5MEHHwRg9OjRrF69mrFjx9KkSRPOnz/Pf//7Xw4ePEibNm1KHP/JJ5/E39+fl19+2Vx+Xr9+fQC+/fZb+vXrR8OGDZk5cyaXL1/mrbfeokuXLuzdu7fYsoJhw4YRFhbGyy+/XK6k9NKlS5w7d67YdU9PT/OXOgsXLqRp06YMGjQIKysrvvzyS8aMGUNBQQF/+9vfzPdER0fz2GOP0bRpU5599lnc3Nz4+eef+frrr82fD0BKSgp9+/bl7rvv5t5772X16tVMnjyZ5s2b069fv2vGnJubWyxmBwcHHBwcMBqNDBo0iC1btjBy5EhatWrFxo0bmThxIidPnuSNN94otd/Tp0/Ts2dP8vLymDJlCo6Ojrz//vvY29tfMyYREakBRhERkWqSl5dn9PX1NXbq1KnI9XfffdcIGDdu3Gg0Go3GrKwsY35+fpE2x48fN9ra2hpffPHFItcA49KlS83XZsyYYbz6/8727dtnBIxjxowp0t+DDz5oBIwzZswwX8vMzCwW844dO4yA8YMPPjBfW7VqlREwbtmypVj77t27G7t3725+Pn/+fCNg/PDDD83XcnJyjJ06dTI6OTkZ09LSirwXT09P44ULF8xtv/jiCyNg/PLLL4uNdbUtW7YYAeOqVatKbfP0008bAeOPP/5ovnbp0iVjSEiIMTg42PyZDx482Ni0adMyx3N1dTX+7W9/K7NNReJs1aqV0dvb23j+/Hnztf379xstLCyMw4cPN18r/P0+8MADFRqvtMepU6fMbUv6/ffp08fYsGFD8/OLFy8anZ2djR07djRevny5SNuCggLzz927dy/29yY7O9vo4+NjHDp06DXjDgoKKjHewr+vn3/+uREw/utf/ypy3z333GM0GAzG2NjYIn09+uij5ueFfw927txpvpacnGx0dXU1Asbjx49fMz4REblxVO4uIiLVxtLSkvvvv58dO3YUKSH/6KOPqF+/PnfccQcAtra2WFiY/i8pPz+f8+fP4+TkRHh4eJnl1CVZv349YNo47GpPP/10sbZXzyTm5uZy/vx5GjVqhJubW4XHvXp8Hx8fHnjgAfM1a2trxo0bR3p6Ot9//32R9vfddx/u7u7m5926dQNMZerXa/369XTo0KFIebiTkxNPPPEEJ06c4MCBA4BpvXhiYmKZZfZubm7s3LmTpKSk647r1KlT7Nu3j6ioKDw8PMzXW7RowZ133mn+HV5t9OjRFRpj+vTpbNq0qdjj6vGu/v2npqZy7tw5unfvzrFjx0hNTQVg06ZNXLp0iSlTpmBnZ1dkjD8vs3ByciqyptzGxoYOHTqU+3fZsWPHYvEW7uewfv16LC0ti/29/uc//4nRaGTDhg2l9rt+/Xpuu+02OnToYL7m5eV13ftBiIhI9VCSLiIi1aowEShcv5uYmMiPP/7I/fffj6WlJQAFBQW88cYbhIWFYWtrS7169fDy8uKXX34xJ0vlFRcXh4WFBaGhoUWuh4eHF2t7+fJlpk+fbl7jWzjuxYsXKzzu1eOHhYWZv3QoVFgeHxcXV+R6YGBgkeeFCXtKSkqlxv9zLCW97z/HMnnyZJycnOjQoQNhYWH87W9/Y9u2bUXumTNnDr/99hsBAQF06NCBmTNnVvqLhMJxS4vt3LlzZGRkFLkeEhJSoTGaN29Or169ij2u3uhv27Zt9OrVy7wm3svLi+eeew7A/Ps/evQoQLnOQG/QoEGxxN3d3b3cv8t69eoVi7dhw4aA6TPz8/PD2dm5yD2l/b26WuHfyT8r6fMXEZGapyRdRESqVdu2bYmIiODjjz8G4OOPP8ZoNBaZxXv55ZcZP348t99+Ox9++CEbN25k06ZNNG3atFo3t/r73//OSy+9xL333sunn37KN998w6ZNm/D09Lxhm2oVflHxZ8ZyrLuuKpGRkcTExLBy5Uq6du3KmjVr6Nq1KzNmzDC3uffeezl27BhvvfUWfn5+vPbaazRt2rTMGdyqVNXrp48ePcodd9zBuXPneP311/nqq6/YtGkTzzzzDEClfv+14XcpIiJ1nzaOExGRavfQQw8xbdo0fvnlFz766CPCwsKKnGG9evVqevbsyeLFi4vcd/HiRerVq1ehsYKCgigoKODo0aNFZgpjYmKKtV29ejWPPvoo8+bNM1/Lysri4sWLRdr9eXb0WuP/8ssvFBQUFJlNP3TokPn1GyUoKKjE911SLI6Ojtx3333cd9995OTkcPfdd/PSSy/x7LPPmsu8fX19GTNmDGPGjCE5OZk2bdrw0ksvlWtTtD/HBSX/Tg4dOkS9evWuecTa9fryyy/Jzs5m7dq1RaoZtmzZUqRdYUXGb7/9RqNGjao1prIEBQXx7bffcunSpSKz6eX5exUUFMSRI0eKXS/p8xcRkZqnmXQREal2hbPm06dPZ9++fcXWwlpaWhabbVy1ahUnT56s8FiFCeOCBQuKXJ8/f36xtiWN+9Zbb5Gfn1/kWmHC+OfkvST9+/fn9OnTfPLJJ+ZreXl5vPXWWzg5OZmPJLsR+vfvz65du9ixY4f5WkZGBu+//z7BwcE0adIEgPPnzxe5z8bGhiZNmmA0GsnNzSU/P79Y+b+3tzd+fn5kZ2dXOC5fX19atWrFsmXLinymv/32G9988w39+/evcJ8VVTjrffXvPzU1laVLlxZp17t3b5ydnXnllVfIysoq8tqNnCHv378/+fn5vP3220Wuv/HGGxgMhjK/KOnfvz//+9//2LVrl/na2bNnWbFiRbXFKyIilaeZdBERqXYhISF07tyZL774AqBYkj5gwABefPFFRowYQefOnfn1119ZsWKFeT1uRbRq1YoHHniAd955h9TUVDp37szmzZuJjY0t1nbAgAEsX74cV1dXmjRpwo4dO/j222/x9PQs1qelpSWvvvoqqamp2Nrams/X/rMnnniC9957j6ioKPbs2UNwcDCrV69m27ZtzJ8/v9ia4uu1Zs0a82zq1R599FGmTJnCxx9/TL9+/Rg3bhweHh4sW7aM48ePs2bNGvNMf+/evfHx8aFLly7Ur1+fgwcP8vbbb/PXv/4VZ2dnLl68SIMGDbjnnnto2bIlTk5OfPvtt+zevbtIFUJFvPbaa/Tr149OnToxcuRI8xFsrq6uzJw583o+EgB+/PHHYkk1mDana9GiBb1798bGxoaBAwfy5JNPkp6ezqJFi/D29ubUqVPm9i4uLrzxxhs8/vjjtG/f3nxW+/79+8nMzGTZsmXXHWt5DBw4kJ49ezJ16lROnDhBy5Yt+eabb/jiiy94+umni+3BcLVJkyaxfPly+vbtyz/+8Q/zEWyFVR8iIlLL1Ni+8iIickv597//bQSMHTp0KPZaVlaW8Z///KfR19fXaG9vb+zSpYtxx44dxY43K88RbEaj0Xj58mXjuHHjjJ6enkZHR0fjwIEDjQkJCcWOYEtJSTGOGDHCWK9ePaOTk5OxT58+xkOHDhU7wspoNBoXLVpkbNiwodHS0rLIcWx/jtFoNBrPnDlj7tfGxsbYvHnzIjFf/V5ee+21Yp/Hn+MsybWOGis8du3o0aPGe+65x+jm5ma0s7MzdujQwbhu3boifb333nvG22+/3ejp6Wm0tbU1hoaGGidOnGhMTU01Go2mo8QmTpxobNmypdHZ2dno6OhobNmypfGdd94pM8ar4yzpqLhvv/3W2KVLF6O9vb3RxcXFOHDgQOOBAweKtCn8/Z49e/aaY5Xnc7n6c127dq2xRYsWRjs7O2NwcLDx1VdfNS5ZsqTEY8nWrl1r7Ny5sznWDh06GD/++GPz6927dy/xGLtHH33UGBQUdM24g4KCjH/961/LbHPp0iXjM888Y/Tz8zNaW1sbw8LCjK+99lqRo+AK+/rz399ffvnF2L17d6OdnZ3R39/fOGvWLOPixYt1BJuISC1kMBq1m4mIiIiIiIhIbaA16SIiIiIiIiK1hJJ0ERERERERkVpCSbqIiIiIiIhILaEkXURERERERKSWUJIuIiIiIiIiUksoSRcRERERERGpJaxqOoAbraCggKSkJJydnTEYDDUdjoiIiIiIiNzkjEYjly5dws/PDwuLsufKb7kkPSkpiYCAgJoOQ0RERERERG4xCQkJNGjQoMw2t1yS7uzsDJg+HBcXlxqORkRERERERG52aWlpBAQEmPPRstxySXphibuLi4uSdBEREREREblhyrPkWhvHiYiIiIiIiNQSStJFREREREREagkl6SIiIiIiIiK1xC23Jl1ERERERK6P0WgkLy+P/Pz8mg5FpNawtrbG0tLyuvtRki4iIiIiIuWWk5PDqVOnyMzMrOlQRGoVg8FAgwYNcHJyuq5+lKSLiIiIiEi5FBQUcPz4cSwtLfHz88PGxqZcu1WL3OyMRiNnz54lMTGRsLCw65pRV5IuIiIiIiLlkpOTQ0FBAQEBATg4ONR0OCK1ipeXFydOnCA3N/e6knRtHCciIiIiIhViYaE0QuTPqqqqRP+6RERERERERGoJJekiIiIiIiIitYSSdBERERERkWvo0aMHTz/9dJltgoODmT9//g2JpyKio6Nxc3MzP585cyatWrUq856oqCiGDBly3WNXVT+3EiXpIiIiIiJy04uKisJgMBR7xMbG3rAYfv/9d4YOHUpwcDAGg+GaCf2aNWuwtLTk5MmTJb4eFhbG+PHjKxzHhAkT2Lx5c4XvK8uJEycwGAzs27evyPU333yT6OjoKh2rJAaDgc8//7zax7kRlKSLiIiIiMgtoW/fvpw6darIIyQk5IaNn5mZScOGDZk9ezY+Pj7XbD9o0CA8PT1ZtmxZsdd++OEHYmNjGTlyZIXjcHJywtPTs8L3VYarq2uRWXy5NiXpIiIiIiJSaUajkcycvBp5GI3GCsVqa2uLj49PkUfhUVnff/89HTp0wNbWFl9fX6ZMmUJeXl6pfSUnJzNw4EDs7e0JCQlhxYoV1xy/ffv2vPbaa9x///3Y2tpes721tTWPPPJIiTPRS5YsoWPHjjRt2pTXX3+d5s2b4+joSEBAAGPGjCE9Pb3Ufv9c7p6fn8/48eNxc3PD09OTSZMmFftsv/76a7p27WpuM2DAAI4ePWp+vfDLjtatW2MwGOjRowdQvNw9OzubcePG4e3tjZ2dHV27dmX37t3m17du3YrBYGDz5s20a9cOBwcHOnfuTExMzDU/r9IUFBTw4osv0qBBA2xtbWnVqhVff/21+fWcnBzGjh2Lr68vdnZ2BAUF8corrwCmv98zZ84kMDAQW1tb/Pz8GDduXKVjKQ+dky4iIiIiIpV2OTefJtM31sjYB17sg4PN9ac0J0+epH///kRFRfHBBx9w6NAhRo0ahZ2dHTNnzizxnqioKJKSktiyZQvW1taMGzeO5OTk647lz0aOHMnrr7/ODz/8wO233w5Aeno6q1ev5o033gBMR+ItWLCAkJAQjh07xpgxY5g0aRLvvPNOucaYN28e0dHRLFmyhMjISObNm8d//vMf/vKXv5jbZGRkMH78eFq0aEF6ejrTp0/nrrvuYt++fVhYWLBr1y46dOjAt99+S9OmTbGxsSlxrEmTJrFmzRqWLVtGUFAQc+bMoU+fPsTGxuLh4WFuN3XqVObNm4eXlxejR4/mscceY9u2bZX6DN98803mzZvHe++9R+vWrVmyZAmDBg3i999/JywsjAULFrB27Vo+/fRTAgMDSUhIICEhATAtOXjjjTdYuXIlTZs25fTp0+zfv79ScZSXknQREREREbklrFu3DicnJ/Pzfv36sWrVKt555x0CAgJ4++23MRgMREREkJSUxOTJk5k+fXqxc+EPHz7Mhg0b2LVrF+3btwdg8eLFREZGVnnMTZo04bbbbmPJkiXmJP3TTz/FaDRy//33AxTZ0C44OJh//etfjB49utxJ+vz583n22We5++67AXj33XfZuLHoFy9Dhw4t8nzJkiV4eXlx4MABmjVrhpeXFwCenp6llvJnZGSwcOFCoqOj6devHwCLFi1i06ZNLF68mIkTJ5rbvvTSS3Tv3h2AKVOm8Ne//pWsrCzs7OzK9Z6uNnfuXCZPnmz+vF599VW2bNnC/Pnz+fe//018fDxhYWF07doVg8FAUFCQ+d74+Hh8fHzo1asX1tbWBAYG0qFDhwrHUBFK0muphAuZ/HoylQbu9rRo4FbT4YiIiIiIlMje2pIDL/apsbEromfPnixcuND83NHREYCDBw/SqVMnDAaD+bUuXbqQnp5OYmIigYGBRfo5ePAgVlZWtG3b1nwtIiKi2tZeP/bYYzzzzDO89dZbODs7s2TJEoYNG4azszMA3377La+88gqHDh0iLS2NvLw8srKyyMzMxMHBocy+U1NTOXXqFB07djRfs7Kyol27dkVK3o8cOcL06dPZuXMn586do6CgADAlsc2aNSvX+zh69Ci5ubl06dLFfM3a2poOHTpw8ODBIm1btGhh/tnX1xcwLTH48+/iWtLS0khKSioyJph+v4Uz4lFRUdx5552Eh4fTt29fBgwYQO/evQEYNmwY8+fPp2HDhvTt25f+/fszcOBArKyqL5XWmvRaavn/4hizYi+f7S15J0cRERERkdrAYDDgYGNVI4+rk+rycHR0pFGjRuZHYfJX2xXOAH/66accOXKEbdu2mTeMO3HiBAMGDKBFixasWbOGPXv28O9//xswrbWuKgMHDuTChQssWrSInTt3snPnziof42rW1tbmnwt/z4VfDFS1Nm3acPz4cWbNmsXly5e59957ueeeewAICAggJiaGd955B3t7e8aMGcPtt99Obm5utcQCStJrrYb1TN/qHT1b+oYPIiIiIiJy/SIjI9mxY0eRmeNt27bh7OxMgwYNirWPiIggLy+PPXv2mK/FxMRw8eLFaonP2dmZYcOGsWTJEpYuXUrjxo3p1q0bAHv27KGgoIB58+Zx22230bhxY5KSksrdt6urK76+vuakGyj23s6fP09MTAzPP/88d9xxB5GRkaSkpBTpp3ANen5+fqljhYaGYmNjU2RteW5uLrt376ZJkybljrkiXFxc8PPzK7aefdu2bUXGdHFx4b777mPRokV88sknrFmzhgsXLgBgb2/PwIEDWbBgAVu3bmXHjh38+uuv1RIvqNy91mroZVorc+xsRg1HIiIiIiJycxszZgzz58/n73//O2PHjiUmJoYZM2Ywfvz4YuvRAXNZ9JNPPsnChQuxsrLi6aefxt7evsxxcnJyOHDggPnnkydPsm/fPpycnGjUqFGZ944cOZJu3bpx8OBBJk+ebL7eqFEjcnNzeeuttxg4cCDbtm3j3XffrdD7/8c//sHs2bMJCwsjIiKC119/vcgXDu7u7nh6evL+++/j6+tLfHw8U6ZMKdKHt7c39vb2fP311zRo0AA7OztcXV2LtHF0dOSpp55i4sSJeHh4EBgYyJw5c8jMzKzUUXJ/dvz48WLntIeFhTFx4kRmzJhBaGgorVq1YunSpezbt8+8I//rr7+Or68vrVu3xsLCglWrVuHj44ObmxvR0dHk5+fTsWNHHBwc+PDDD7G3ty+ybr2qaSa9lmroZZpJT0q9TFZu6d9GiYiIiIjI9fH392f9+vXs2rWLli1bMnr0aEaOHMnzzz9f6j1Lly7Fz8+P7t27c/fdd/PEE0/g7e1d5jhJSUm0bt2a1q1bc+rUKebOnUvr1q15/PHHrxlj165dCQ8PJy0tjeHDh5uvt2zZktdff51XX32VZs2asWLFCvPxYeX1z3/+k0ceeYRHH32UTp064ezszF133WV+3cLCgpUrV7Jnzx6aNWvGM888w2uvvVakDysrKxYsWMB7772Hn58fgwcPLnGs2bNnM3ToUB555BHatGlDbGwsGzduxN3dvUIxl2T8+PHmz7fw8fPPPzNu3DjGjx/PP//5T5o3b87XX3/N2rVrCQsLA0yVCnPmzKFdu3a0b9+eEydOsH79eiwsLHBzc2PRokV06dKFFi1a8O233/Lll19W6znzBmNFDxes49LS0nB1dSU1NRUXF5eaDqdURqORli98Q1pWHhv+0Y1I39obq4iIiIjcGrKysjh+/DghISGV2mVb5GZW1r+PiuShmkmvpQwGg0reRUREREREbjFK0muxwpL3Y9o8TkRERERE5JagJL0WC70yk378nGbSRUREREREbgVK0muxkMJj2JSki4iIiIiI3BKUpNdiV5e732L7+4mIiIiIiNySlKTXYsGejhgMcCkrj3PpOTUdjoiIiIiIiFQzJem1mJ21Jf5u9oA2jxMREREREbkVKEmv5czHsGlduoiIiIiIyE1PSXot17CejmETERERERG5VShJr+VCr2wep2PYRERERERqTo8ePXj66afLbBMcHMz8+fNvSDwVMXPmTFq1amV+HhUVxZAhQ8q8pzzvtzyqqp9biZL0Ws5c7n5WSbqIiIiISGVFRUVhMBiKPWJjY29YDIsWLaJbt264u7vj7u5Or1692LVrV6nt582bh7u7O1lZWcVey8zMxMXFhQULFlQ4jjfffJPo6OgK31eWrVu3YjAYuHjxYpHrn332GbNmzarSsf7sxIkTGAwG9u3bV63j3ChK0mu5wrPS4y9kkptfUMPRiIiIiIjUXX379uXUqVNFHiEhITds/K1bt/LAAw+wZcsWduzYQUBAAL179+bkyZMltn/kkUfIyMjgs88+K/ba6tWrycnJ4eGHH65wHK6urri5uVX4vsrw8PDA2dn5hox1s1CSXsv5uNhhb21JXoGR+AuZNR2OiIiIiEhRRiPkZNTMw2isUKi2trb4+PgUeVhaWgLw/fff06FDB2xtbfH19WXKlCnk5eWV2ldycjIDBw7E3t6ekJAQVqxYcc3xV6xYwZgxY2jVqhURERH83//9HwUFBWzevLnE9t7e3gwcOJAlS5YUe23JkiUMGTIEDw8PJk+eTOPGjXFwcKBhw4ZMmzaN3NzcUuP4c7l7RkYGw4cPx8nJCV9fX+bNm1fsnuXLl9OuXTucnZ3x8fHhwQcfJDk5GTDNZPfs2RMAd3d3DAYDUVFRQPFy95SUFIYPH467uzsODg7069ePI0eOmF+Pjo7Gzc2NjRs3EhkZiZOTk/nLlcrKzs5m3LhxeHt7Y2dnR9euXdm9e3eRmB566CG8vLywt7cnLCyMpUuXApCTk8PYsWPx9fXFzs6OoKAgXnnllUrHUh5W1dq7XDcLCwMh9Rw5cCqNY2czCL1S/i4iIiIiUivkZsLLfjUz9nNJYON43d2cPHmS/v37ExUVxQcffMChQ4cYNWoUdnZ2zJw5s8R7oqKiSEpKYsuWLVhbWzNu3Dhz0lpemZmZ5Obm4uHhUWqbkSNHMmDAAOLi4ggKCgLg2LFj/PDDD2zcuBEAZ2dnoqOj8fPz49dff2XUqFE4OzszadKkcsUxceJEvv/+e7744gu8vb157rnn2Lt3b5F17Lm5ucyaNYvw8HCSk5MZP348UVFRrF+/noCAANasWcPQoUOJiYnBxcUFe3v7EseKioriyJEjrF27FhcXFyZPnkz//v05cOAA1tbW5s9l7ty5LF++HAsLCx5++GEmTJhQri9CSjJp0iTWrFnDsmXLCAoKYs6cOfTp04fY2Fg8PDyYNm0aBw4cYMOGDdSrV4/Y2FguX74MwIIFC1i7di2ffvopgYGBJCQkkJCQUKk4yktJeh3Q0KswSU8H6td0OCIiIiIiddK6detwcvpj0qtfv36sWrWKd955h4CAAN5++20MBgMREREkJSUxefJkpk+fjoVF0QLkw4cPs2HDBnbt2kX79u0BWLx4MZGRkRWKZ/Lkyfj5+dGrV69S2/Tp0wc/Pz+WLl1q/sIgOjqagIAA7rjjDgCef/55c/vg4GAmTJjAypUry5Wkp6ens3jxYj788ENzf8uWLaNBgwZF2j322GPmnxs2bMiCBQto37496enpODk5mb9o8Pb2LrWUvjA537ZtG507dwZM1QUBAQF8/vnnDBs2DDB9IfDuu+8SGhoKwNixY3nxxRev+V5KkpGRwcKFC4mOjqZfv36AaW+ATZs2sXjxYiZOnEh8fDytW7emXbt2gOkzLBQfH09YWBhdu3bFYDCYvyipTkrS6wBtHiciIiIitZa1g2lGu6bGroCePXuycOFC83NHR9Ms/MGDB+nUqRMGg8H8WpcuXUhPTycxMZHAwMAi/Rw8eBArKyvatm1rvhYREVGhdd6zZ89m5cqVbN26FTs7u1LbWVpa8uijjxIdHc2MGTMwGo0sW7aMESNGmL88+OSTT1iwYAFHjx4lPT2dvLw8XFxcyhXH0aNHycnJoWPHjuZrHh4ehIeHF2m3Z88eZs6cyf79+0lJSaGgwLRfVnx8PE2aNCnXWIWf29VjeXp6Eh4ezsGDB83XHBwczAk6gK+vb4WrFK5+f7m5uXTp0sV8zdramg4dOpjHfOqppxg6dCh79+6ld+/eDBkyxPwlQlRUFHfeeSfh4eH07duXAQMG0Lt370rFUl5ak14H6Bg2EREREam1DAZTyXlNPK5KqsvD0dGRRo0amR++vr7V9KGUbe7cucyePZtvvvmGFi1aXLP9Y489Rnx8PN999x2bN28mISGBESNGALBjxw4eeugh+vfvz7p16/j555+ZOnUqOTk5VRZvRkYGffr0wcXFhRUrVrB7927+85//AFTpOIUKy94LGQwGjBXcf6Ai+vXrR1xcHM888wxJSUnccccdTJgwAYA2bdpw/PhxZs2axeXLl7n33nu55557qi0WUJJeJzSsd2Um/Vx6DUciIiIiInLziYyMZMeOHUUSwW3btuHs7Fys7BtMs+Z5eXns2bPHfC0mJqbY8WMlmTNnDrNmzeLrr782l1dfS2hoKN27d2fJkiUsXbqUXr16mcuut2/fTlBQEFOnTqVdu3aEhYURFxdXrn4L+7a2tmbnzp3maykpKRw+fNj8/NChQ5w/f57Zs2fTrVs3IiIiis1s29jYAJCfn1/qWJGRkeTl5RUZ6/z588TExJR7Nr6iQkNDsbGxYdu2beZrubm57N69u8iYXl5ePProo3z44YfMnz+f999/3/yai4sL9913H4sWLeKTTz5hzZo1XLhwoVriBZW71wkhV2bSz6XnkHo5F1d762vcISIiIiIi5TVmzBjmz5/P3//+d8aOHUtMTAwzZsxg/PjxxdajA+bS5yeffJKFCxdiZWXF008/XepmaYVeffVVpk+fzkcffURwcDCnT58GwMnJqcha+ZKMHDmSUaNGARQ54zwsLIz4+HhWrlxJ+/bt+eqrr8yz3OXh5OTEyJEjmThxIp6ennh7ezN16tQi7zswMBAbGxveeustRo8ezW+//Vbs7POgoCAMBgPr1q2jf//+2NvbF3tPYWFhDB48mFGjRvHee+/h7OzMlClT8Pf3Z/DgweWOuTQxMTHFrjVt2pSnnnqKiRMn4uHhQWBgIHPmzCEzM5ORI0cCMH36dNq2bUvTpk3Jzs5m3bp15v0FXn/9dXx9fWndujUWFhasWrUKHx+faj3CTjPpdYCTrRX1XWwBrmweJyIiIiIiVcXf35/169eza9cuWrZsyejRoxk5cmSRDdn+bOnSpfj5+dG9e3fuvvtunnjiCby9vcscZ+HCheTk5HDPPffg6+trfsydO/eaMQ4dOhRbW1scHByKHJ82aNAgnnnmGcaOHUurVq3Yvn0706ZNK/d7B3jttdfo1q0bAwcOpFevXnTt2rXIensvLy+io6NZtWoVTZo0Yfbs2cVi9vf354UXXmDKlCnUr1+fsWPHljjW0qVLadu2LQMGDKBTp04YjUbWr19frMS9Mu6//35at25d5HHmzBlmz57N0KFDeeSRR2jTpg2xsbFs3LgRd3d3wFQF8Oyzz9KiRQtuv/12LC0tWblyJWDaOX/OnDm0a9eO9u3bc+LECdavX1/ilzdVxWCszuL+WigtLQ1XV1dSU1PLvZlCbXD/+zv437ELzBvWkqFti5fciIiIiIhUt6ysLI4fP05ISEiZm52J3IrK+vdRkTxUM+l1hHmHd61LFxERERERuWkpSa8jGtYzrUvXMWwiIiIiIiI3LyXpdUTolZl0HcMmIiIiIiJy81KSXkc0vOqs9IKCW2obARERERERkVuGkvQ6ooG7AzaWFmTnFXDy4uWaDkdERERERESqgZL0OsLSwkCQpwMAx1TyLiIiIiIiclNSkl6HFJa866x0ERERERGRm5OS9DokpN6VY9i0w7uIiIiIiMhNSUl6HWKeSddZ6SIiIiIiIjclJel1SGjhDu+aSRcRERERuaF69OjB008/XWab4OBg5s+ff0PiqYiZM2fSqlUr8/OoqCiGDBlS5j3leb/lUVX93EqUpNchDa+UuyelZpGZk1fD0YiIiIiI1B1RUVEYDIZij9jY2BsWw2effUa7du1wc3PD0dGRVq1asXz58lLbz5s3D3d3d7Kysoq9lpmZiYuLCwsWLKhwHG+++SbR0dEVvq8sW7duxWAwcPHixSLXP/vsM2bNmlWlY/3ZiRMnMBgM7Nu3r1rHuVGUpNch7o42uDtYA6bz0kVEREREpPz69u3LqVOnijxCQkJu2PgeHh5MnTqVHTt28MsvvzBixAhGjBjBxo0bS2z/yCOPkJGRwWeffVbstdWrV5OTk8PDDz9c4ThcXV1xc3Or8H2V4eHhgbOz8w0Z62ahJL2OaeilzeNEREREpPYwGo1k5mbWyMNoNFYoVltbW3x8fIo8LC0tAfj+++/p0KEDtra2+Pr6MmXKFPLySq9eTU5OZuDAgdjb2xMSEsKKFSuuOX6PHj246667iIyMJDQ0lH/84x+0aNGC//73vyW29/b2ZuDAgSxZsqTYa0uWLGHIkCF4eHgwefJkGjdujIODAw0bNmTatGnk5uaWGsefy90zMjIYPnw4Tk5O+Pr6Mm/evGL3LF++nHbt2uHs7IyPjw8PPvggycnJgGkmu2fPngC4u7tjMBiIiooyv+ery91TUlIYPnw47u7uODg40K9fP44cOWJ+PTo6Gjc3NzZu3EhkZCROTk7mL1cqKzs7m3HjxuHt7Y2dnR1du3Zl9+7dRWJ66KGH8PLywt7enrCwMJYuXQpATk4OY8eOxdfXFzs7O4KCgnjllVcqHUt5WFVr71LlGtZzZE9cipJ0EREREakVLuddpuNHHWtk7J0P7sTB2uG6+zl58iT9+/cnKiqKDz74gEOHDjFq1Cjs7OyYOXNmifdERUWRlJTEli1bsLa2Zty4ceaktTyMRiPfffcdMTExvPrqq6W2GzlyJAMGDCAuLo6goCAAjh07xg8//GCegXd2diY6Oho/Pz9+/fVXRo0ahbOzM5MmTSpXLBMnTuT777/niy++wNvbm+eee469e/cWWceem5vLrFmzCA8PJzk5mfHjxxMVFcX69esJCAhgzZo1DB06lJiYGFxcXLC3ty9xrKioKI4cOcLatWtxcXFh8uTJ9O/fnwMHDmBtbaoazszMZO7cuSxfvhwLCwsefvhhJkyYUK4vQkoyadIk1qxZw7JlywgKCmLOnDn06dOH2NhYPDw8mDZtGgcOHGDDhg3Uq1eP2NhYLl++DMCCBQtYu3Ytn376KYGBgSQkJJCQkFCpOMpLSXodY55J1w7vIiIiIiIVsm7dOpycnMzP+/Xrx6pVq3jnnXcICAjg7bffxmAwEBERQVJSEpMnT2b69OlYWBQtQD58+DAbNmxg165dtG/fHoDFixcTGRl5zRhSU1Px9/cnOzsbS0tL3nnnHe68885S2/fp0wc/Pz+WLl1q/sIgOjqagIAA7rjjDgCef/55c/vg4GAmTJjAypUry5Wkp6ens3jxYj788ENzf8uWLaNBgwZF2j322GPmnxs2bMiCBQto37496enpODk54eHhAZhm/0srpS9Mzrdt20bnzp0BWLFiBQEBAXz++ecMGzYMMH0h8O677xIaGgrA2LFjefHFF6/5XkqSkZHBwoULiY6Opl+/fgAsWrSITZs2sXjxYiZOnEh8fDytW7emXbt2gOkzLBQfH09YWBhdu3bFYDCYvyipTkrS65jCY9iOnlWSLiIiIiI1z97Knp0P7qyxsSuiZ8+eLFy40Pzc0dH039YHDx6kU6dOGAwG82tdunQhPT2dxMREAgMDi/Rz8OBBrKysaNu2rflaREREudZ5Ozs7s2/fPtLT09m8eTPjx4+nYcOG9OjRo8T2lpaWPProo0RHRzNjxgyMRiPLli1jxIgR5i8PPvnkExYsWMDRo0dJT08nLy8PFxeXcn0mR48eJScnh44d/6iG8PDwIDw8vEi7PXv2MHPmTPbv309KSgoFBQWAKYlt0qRJucYq/NyuHsvT05Pw8HAOHjxovubg4GBO0AF8fX0rVKXw5/eXm5tLly5dzNesra3p0KGDecynnnqKoUOHsnfvXnr37s2QIUPMXyJERUVx5513Eh4eTt++fRkwYAC9e/euVCzlVeNr0v/9738THByMnZ0dHTt2ZNeuXaW2zc3N5cUXXyQ0NBQ7OztatmzJ119/fQOjrXnh9U2bLhw5k05+QcXW4IiIiIiIVDWDwYCDtUONPK5OqsvD0dGRRo0amR++vr7V9KmUzsLCgkaNGtGqVSv++c9/cs8991xzjfNjjz1GfHw83333HZs3byYhIYERI0YAsGPHDh566CH69+/PunXr+Pnnn5k6dSo5OTlVFnNGRgZ9+vTBxcWFFStWsHv3bv7zn/8AVOk4hQrL3gsZDIYK7z9QEf369SMuLo5nnnmGpKQk7rjjDiZMmABAmzZtOH78OLNmzeLy5cvce++93HPPPdUWC9Rwkv7JJ58wfvx4ZsyYwd69e2nZsiV9+vQp9VuS559/nvfee4+33nqLAwcOMHr0aO666y5+/vnnGxx5zQn0cMDe2pLsvAJOnNe6dBERERGR6xUZGcmOHTuKJILbtm3D2dm5WNk3mGbN8/Ly2LNnj/laTExMsePHyqOgoIDs7Owy24SGhtK9e3eWLFnC0qVL6dWrl7nsevv27QQFBTF16lTatWtHWFgYcXFx5R4/NDQUa2trdu78oxoiJSWFw4cPm58fOnSI8+fPM3v2bLp160ZERESxnM3GxgaA/Pz8UseKjIwkLy+vyFjnz58nJiam3LPxFRUaGoqNjQ3btm0zX8vNzWX37t1FxvTy8uLRRx/lww8/ZP78+bz//vvm11xcXLjvvvtYtGgRn3zyCWvWrOHChQvVEi/UcLn766+/zqhRo8zfAr377rt89dVXLFmyhClTphRrv3z5cqZOnUr//v0BU1nCt99+y7x58/jwww9vaOw1xcLCQLiPM/sSLnLo1CVCvZyufZOIiIiIiJRqzJgxzJ8/n7///e+MHTuWmJgYZsyYwfjx44utRwfMpc9PPvkkCxcuxMrKiqeffrrUzdIKvfLKK7Rr147Q0FCys7NZv349y5cvL1KCX5qRI0cyatQogCJnnIeFhREfH8/KlStp3749X331lXmWuzycnJwYOXIkEydOxNPTE29vb6ZOnVrkfQcGBmJjY8Nbb73F6NGj+e2334qdfR4UFITBYGDdunX0798fe3v7Iuv/C2MdPHgwo0aN4r333sPZ2ZkpU6bg7+/P4MGDyx1zaWJiYopda9q0KU899RQTJ07Ew8ODwMBA5syZQ2ZmJiNHjgRg+vTptG3blqZNm5Kdnc26devM+wu8/vrr+Pr60rp1aywsLFi1ahU+Pj7VeoRdjc2k5+TksGfPHnr16vVHMBYW9OrVix07dpR4T3Z2NnZ2dkWu2dvbl3pkQeE9aWlpRR51XaSvqeT90Om6/15ERERERGqav78/69evZ9euXbRs2ZLRo0czcuTIIhuy/dnSpUvx8/Oje/fu3H333TzxxBN4e3uXOU5GRgZjxoyhadOmdOnShTVr1vDhhx/y+OOPXzPGoUOHYmtri4ODQ5Hj0wYNGsQzzzzD2LFjadWqFdu3b2fatGnlfu8Ar732Gt26dWPgwIH06tWLrl27Fllv7+XlRXR0NKtWraJJkybMnj2buXPnFunD39+fF154gSlTplC/fn3Gjh1b4lhLly6lbdu2DBgwgE6dOmE0Glm/fn2xEvfKuP/++2ndunWRx5kzZ5g9ezZDhw7lkUceoU2bNsTGxrJx40bc3d0BUxXAs88+S4sWLbj99tuxtLRk5cqVgGkPgTlz5tCuXTvat2/PiRMnWL9+fYlf3lQVg7E6i/vLkJSUhL+/P9u3b6dTp07m65MmTeL7778vUgJR6MEHH2T//v18/vnnhIaGsnnzZgYPHkx+fn6pJSIzZ87khRdeKHY9NTW13Jsp1DbLtp9gxtrf6RVZn/97tF1NhyMiIiIit4isrCyOHz9OSEhIsckzkVtdWf8+0tLScHV1LVceWuMbx1XEm2++SVhYGBEREdjY2DB27NgiuxqW5NlnnyU1NdX8qO4z7W6ECB/NpIuIiIiIiNyMaixJr1evHpaWlpw5c6bI9TNnzuDj41PiPV5eXnz++edkZGQQFxfHoUOHcHJyomHDhqWOY2tri4uLS5FHXRfhY3oPiSmXScvKreFoREREREREpKrUWJJuY2ND27Zt2bx5s/laQUEBmzdvLlL+XhI7Ozv8/f3Jy8tjzZo1VbLJQF3i6mCNn6upfOLw6Us1HI2IiIiIiIhUlRotdx8/fjyLFi1i2bJlHDx4kKeeeoqMjAzzbu/Dhw/n2WefNbffuXMnn332GceOHePHH3+kb9++FBQUMGnSpJp6CzUmwtc0m35QSbqIiIiIiMhNo0aPYLvvvvs4e/Ys06dP5/Tp07Rq1Yqvv/6a+vXrAxAfH19kvXlWVhbPP/88x44dw8nJif79+7N8+fJq3f6+torwcea7Q8kcPKV16SIiIiIiIjeLGk3SAcaOHVvq9vxbt24t8rx79+4cOHDgBkRV+xXOpB9Ski4iIiIiInLTqFO7u8sfIq/s8B5z+hIFBTVyip6IiIiIiIhUMSXpdVRIPUdsLC3IyMknMeVyTYcjIiIiIiIiVUBJeh1lZWlBWH0nAA7qvHQREREREZGbgpL0OqzwvPRDp7TDu4iIiIhIderRowdPP/10mW2Cg4OZP3/+DYmnImbOnEmrVq3Mz6OiohgyZEiZ95Tn/ZZHVfVzK1GSXktdXLOGEw8+xIVly0ptE+lrWpd+SDPpIiIiIiJlioqKwmAwFHvExsbWSDwrV67EYDCUmSzPmzcPd3d3srKyir2WmZmJi4sLCxYsqPDYb775JtHR0RW+ryxbt27FYDBw8eLFItc/++wzZs2aVaVj/dmJEycwGAzs27evWse5UZSk11J5585zee9esg4cLLWNeSZdZ6WLiIiIiFxT3759OXXqVJFHSEjIDY/jxIkTTJgwgW7dupXZ7pFHHiEjI4PPPvus2GurV68mJyeHhx9+uMLju7q63rBjrD08PHB2dr4hY90slKTXUtYN/AHISUwstU3ElZn0E+czyMzJuyFxiYiIiIhczWg0UpCZWSMPo7FipxzZ2tri4+NT5GFpaQnA999/T4cOHbC1tcXX15cpU6aQl1f6f2MnJyczcOBA7O3tCQkJYcWKFeWKIT8/n4ceeogXXniBhg0bltnW29ubgQMHsmTJkmKvLVmyhCFDhuDh4cHkyZNp3LgxDg4ONGzYkGnTppGbm1tqv38ud8/IyGD48OE4OTnh6+vLvHnzit2zfPly2rVrh7OzMz4+Pjz44IMkJycDpi8devbsCYC7uzsGg4GoqCigeLl7SkoKw4cPx93dHQcHB/r168eRI0fMr0dHR+Pm5sbGjRuJjIzEycnJ/OVKZWVnZzNu3Di8vb2xs7Oja9eu7N69u0hMDz30EF5eXtjb2xMWFsbSpUsByMnJYezYsfj6+mJnZ0dQUBCvvPJKpWMpjxo/J11KZhMQAEBuGUl6PSdb6jnZci49m8Nn0mkV4HaDohMRERERMTFevkxMm7Y1Mnb43j0YHByuu5+TJ0/Sv39/oqKi+OCDDzh06BCjRo3Czs6OmTNnlnhPVFQUSUlJbNmyBWtra8aNG2dOWsvy4osv4u3tzciRI/nxxx+v2X7kyJEMGDCAuLg4goKCADh27Bg//PADGzduBMDZ2Zno6Gj8/Pz49ddfGTVqFM7OzkyaNKlc73/ixIl8//33fPHFF3h7e/Pcc8+xd+/eIuvYc3NzmTVrFuHh4SQnJzN+/HiioqJYv349AQEBrFmzhqFDhxITE4OLiwv29vYljhUVFcWRI0dYu3YtLi4uTJ48mf79+3PgwAGsra0BUyn/3LlzWb58ORYWFjz88MNMmDCh3F+E/NmkSZNYs2YNy5YtIygoiDlz5tCnTx9iY2Px8PBg2rRpHDhwgA0bNlCvXj1iY2O5fNl0gtaCBQtYu3Ytn376KYGBgSQkJJCQkFCpOMpLSXotZd2gAQB5Z85QkJ2Nha1tie0ifZ358Ug2h06lKUkXERERESnDunXrcHJyMj/v168fq1at4p133iEgIIC3334bg8FAREQESUlJTJ48menTp2NhUbQA+fDhw2zYsIFdu3bRvn17ABYvXkxkZGSZ4//3v/9l8eLFFVo73adPH/z8/Fi6dKn5C4Po6GgCAgK44447AHj++efN7YODg5kwYQIrV64sV5Kenp7O4sWL+fDDD839LVu2jAZX8pFCjz32mPnnhg0bsmDBAtq3b096ejpOTk54eHgAptn/0krpC5Pzbdu20blzZwBWrFhBQEAAn3/+OcOGDQNMXwi8++67hIaGAjB27FhefPHFa76XkmRkZLBw4UKio6Pp168fAIsWLWLTpk0sXryYiRMnEh8fT+vWrWnXrh1g+gwLxcfHExYWRteuXTEYDOYvSqqTkvRaytLdHYODA8bMTHJPJmHbsOS1MhE+zvx45JzWpYuIiIhIjTDY2xO+d0+NjV0RPXv2ZOHChebnjo6OABw8eJBOnTphMBjMr3Xp0oX09HQSExMJDAws0s/BgwexsrKibds/KggiIiLKXOd96dIlHnnkERYtWkS9evXKHbOlpSWPPvoo0dHRzJgxA6PRyLJlyxgxYoT5y4NPPvmEBQsWcPToUdLT08nLy8PFxaVc/R89epScnBw6duxovubh4UF4eHiRdnv27GHmzJns37+flJQUCgoKAFMS26RJk3KNVfi5XT2Wp6cn4eHhHDz4x15cDg4O5gQdwNfXt1xVCqW9v9zcXLp06WK+Zm1tTYcOHcxjPvXUUwwdOpS9e/fSu3dvhgwZYv4SISoqijvvvJPw8HD69u3LgAED6N27d6ViKS8l6bWUwWDApkEDsg8fJjcxoYwk3fSP7+Ap7fAuIiIiIjeewWCokpLzG8HR0ZFGjRrVyNhHjx7lxIkTDBw40HytMNG1srIiJiamSGJ6tccee4xXXnmF7777joKCAhISEhgxYgQAO3bsMK9x79OnD66urqxcubLEdeWVlZGRQZ8+fejTpw8rVqzAy8uL+Ph4+vTpQ05OTpWNU6iw7L2QwWCo8P4DFdGvXz/i4uJYv349mzZt4o477uBvf/sbc+fOpU2bNhw/fpwNGzbw7bffcu+999KrVy9Wr15dbfFo47harLDkvTybxx06fala/+KKiIiIiNysIiMj2bFjR5H/nt62bRvOzs7Fyr7BNGuel5fHnj1/VBDExMQUO37sz/f8+uuv7Nu3z/wYNGgQPXv2ZN++fQRc2ZOqJKGhoXTv3p0lS5awdOlSevXqZS673r59O0FBQUydOpV27doRFhZGXFxcud97aGgo1tbW7Ny503wtJSWFw4cPm58fOnSI8+fPM3v2bLp160ZERESxmW0bGxvAtDFeaSIjI8nLyysy1vnz54mJiSn3bHxFhYaGYmNjw7Zt28zXcnNz2b17d5Exvby8ePTRR/nwww+ZP38+77//vvk1FxcX7rvvPhYtWsQnn3zCmjVruHDhQrXEC5pJr9VsAkz/g5CbeLLUNo28nbC0MJB6OZfTaVn4ulas5EdERERE5FY3ZswY5s+fz9///nfGjh1LTEwMM2bMYPz48cXWowPm0ucnn3yShQsXYmVlxdNPP13qZmkAdnZ2NGvWrMi1wvL4P18vyciRIxk1ahRAkTPOw8LCiI+PZ+XKlbRv356vvvqK//znP+V41yZOTk6MHDmSiRMn4unpibe3N1OnTi3yvgMDA7GxseGtt95i9OjR/Pbbb8XOPg8KCsJgMLBu3Tr69++Pvb19kfX/hbEOHjyYUaNG8d577+Hs7MyUKVPw9/dn8ODB5Y65NDExMcWuNW3alKeeeoqJEyfi4eFBYGAgc+bMITMzk5EjRwIwffp02rZtS9OmTcnOzmbdunXm/QVef/11fH19ad26NRYWFqxatQofH59qPcJOM+m1mLX/lSS9jN0Dba0sCfUyraU5dErr0kVEREREKsrf35/169eza9cuWrZsyejRoxk5cmSRDdn+bOnSpfj5+dG9e3fuvvtunnjiCby9vastxqFDh2Jra4uDg0OR49MGDRrEM888w9ixY2nVqhXbt29n2rRpFer7tddeo1u3bgwcOJBevXrRtWvXIuvtvby8iI6OZtWqVTRp0oTZs2czd+7cIn34+/vzwgsvMGXKFOrXr8/YsWNLHGvp0qW0bduWAQMG0KlTJ4xGI+vXry9W4l4Z999/P61bty7yOHPmDLNnz2bo0KE88sgjtGnThtjYWDZu3Ii7uztgqgJ49tlnadGiBbfffjuWlpasXLkSMO2cP2fOHNq1a0f79u05ceIE69evL/HLm6piMN5iNdJpaWm4urqSmppa7s0UasqlLVtIfGoMtk0iafjZZ6W2G/fxz6zdn8SkvuGM6VEza2xERERE5OaXlZXF8ePHCQkJwc7OrqbDEalVyvr3UZE8VDPptZhNg8KZ9NLXpMNV69I1ky4iIiIiIlKnKUmvxQo3jiu4dIn81NRS20Ve2eH90Gnt8C4iIiIiIlKXKUmvxSzs7bG8coZiThmz6YUz6UfPZpCdV/puiiIiIiIiIlK7KUmv5cwl72Ucw+bjYoervTX5BUZik9NvVGgiIiIiIiJSxZSk13LW5iS99B3eDQYDET5aly4iIiIiIlLXKUmv5awb+AOQU8ZMOkCkr9ali4iIiIiI1HVK0ms5m4AAAHITT5bZzjyTfloz6SIiIiIiInWVkvRaztq/8Bi20svdASKuzKQfVLm7iIiIiIhInaUkvZazCbiSpCclYcwvfef2xvWdsDDAufRsktOyblR4IiIiIiIiUoWUpNdyVj4+YGWFMTeXvOTkUts52FjRyNsJgF9Pln6muoiIiIiIVFyPHj14+umny2wTHBzM/Pnzb0g8FTFz5kxatWplfh4VFcWQIUPKvKc877c8qqqfW4mS9FrOYGmJtZ8fUPYxbADN/F0B+CVRSbqIiIiIyNWioqIwGAzFHrGxsTcshujo6GLj29nZldp+3rx5uLu7k5VVvFI2MzMTFxcXFixYUOE43nzzTaKjoyt8X1m2bt2KwWDg4sWLRa5/9tlnzJo1q0rH+rMTJ05gMBjYt29ftY5zoyhJrwNsCnd4Tyg7SW9xJUnXTLqIiIiISHF9+/bl1KlTRR4hISE3NAYXF5ci48fFxZXa9pFHHiEjI4PPPvus2GurV68mJyeHhx9+uMIxuLq64ubmVuH7KsPDwwNnZ+cbMtbNQkl6HWDdoHCH97KT9OYN3ABTkm40Gqs7LBERERERjEYjudn5NfKo6H/z2tra4uPjU+RhaWkJwPfff0+HDh2wtbXF19eXKVOmkJeXV2pfycnJDBw4EHt7e0JCQlixYkW5YjAYDEXGr1+/fqltvb29GThwIEuWLCn22pIlSxgyZAgeHh5MnjyZxo0b4+DgQMOGDZk2bRq5ubml9vvncveMjAyGDx+Ok5MTvr6+zJs3r9g9y5cvp127djg7O+Pj48ODDz5I8pXluCdOnKBnz54AuLu7YzAYiIqKAoqXu6ekpDB8+HDc3d1xcHCgX79+HDlyxPx6dHQ0bm5ubNy4kcjISJycnMxfrlRWdnY248aNw9vbGzs7O7p27cru3buLxPTQQw/h5eWFvb09YWFhLF26FICcnBzGjh2Lr68vdnZ2BAUF8corr1Q6lvKwqtbepUpYNzBtHpeTWPYO7018XbAwwNlL2ZxJy8bHtfTSGRERERGRqpCXU8D7//i+RsZ+4s3uWNtaXnc/J0+epH///kRFRfHBBx9w6NAhRo0ahZ2dHTNnzizxnqioKJKSktiyZQvW1taMGzfOnLSWJT09naCgIAoKCmjTpg0vv/wyTZs2LbX9yJEjGTBgAHFxcQQFBQFw7NgxfvjhBzZu3AiAs7Mz0dHR+Pn58euvvzJq1CicnZ2ZNGlSud7/xIkT+f777/niiy/w9vbmueeeY+/evUXWsefm5jJr1izCw8NJTk5m/PjxREVFsX79egICAlizZg1Dhw4lJiYGFxcX7O3tS/3cjhw5wtq1a3FxcWHy5Mn079+fAwcOYG1tDZhK+efOncvy5cuxsLDg4YcfZsKECeX+IuTPJk2axJo1a1i2bBlBQUHMmTOHPn36EBsbi4eHB9OmTePAgQNs2LCBevXqERsby+XLlwFYsGABa9eu5dNPPyUwMJCEhAQSrnHy1vVSkl4HmHd4v8ZZ6fY2ljSu78yh05f49WSqknQRERERkausW7cOJycn8/N+/fqxatUq3nnnHQICAnj77bcxGAxERESQlJTE5MmTmT59OhYWRQuQDx8+zIYNG9i1axft27cHYPHixURGRpY5fnh4OEuWLKFFixakpqYyd+5cOnfuzO+//06DKxNzf9anTx/8/PxYunSp+QuD6OhoAgICuOOOOwB4/vnnze2Dg4OZMGECK1euLFeSnp6ezuLFi/nwww/N/S1btqxYPI899pj554YNG7JgwQLat29Peno6Tk5OeHh4AKbZ/9JK6QuT823bttG5c2cAVqxYQUBAAJ9//jnDhg0DTF8IvPvuu4SGhgIwduxYXnzxxWu+l5JkZGSwcOFCoqOj6devHwCLFi1i06ZNLF68mIkTJxIfH0/r1q1p164dYPoMC8XHxxMWFkbXrl0xGAzmL0qqk5L0WqrAWEBSehLWFta4NChM0ssudwfT5nGHTl/i18SL3Nmk9NIZEREREZGqYGVjwRNvdq+xsSuiZ8+eLFy40Pzc0dERgIMHD9KpUycMBoP5tS5dupCenk5iYiKBgYFF+jl48CBWVla0bdvWfC0iIuKa67w7depEp06dzM87d+5MZGQk7733Xqmbq1laWvLoo48SHR3NjBkzMBqNLFu2jBEjRpi/PPjkk09YsGABR48eJT09nby8PFxcXMr1mRw9epScnBw6duxovubh4UF4eHiRdnv27GHmzJns37+flJQUCgoKAFMS26RJk3KNVfi5XT2Wp6cn4eHhHDx40HzNwcHBnKAD+Pr6lqtKobT3l5ubS5cuXczXrK2t6dChg3nMp556iqFDh7J371569+7NkCFDzF8iREVFceeddxIeHk7fvn0ZMGAAvXv3rlQs5aU16bXUa7tfo99n/VhxaIW53D0vOZmCEnZ2vFqLBto8TkRERERuHIPBgLWtZY08rk6qy8PR0ZFGjRqZH76+vtX0qZSPtbU1rVu3vuYO84899hjx8fF89913bN68mYSEBEaMGAHAjh07eOihh+jfvz/r1q3j559/ZurUqeTk5FRZnBkZGfTp0wcXFxdWrFjB7t27+c9//gNQpeMUKix7L2QwGKp1z61+/foRFxfHM888Q1JSEnfccQcTJkwAoE2bNhw/fpxZs2Zx+fJl7r33Xu65555qiwWUpNdawS7BAMSmxGLp5obFlW/5cpOSyryv2VU7vGvzOBERERGRa4uMjGTHjh1F/vt527ZtODs7l1iGHhERQV5eHnv27DFfi4mJKXb82LXk5+fz66+/XvPLgtDQULp3786SJUtYunQpvXr1Mpddb9++naCgIKZOnUq7du0ICwsrc8f4kvq2trZm586d5mspKSkcPnzY/PzQoUOcP3+e2bNn061bNyIiIorNbNvY2JjfU2kiIyPJy8srMtb58+eJiYkp92x8RYWGhmJjY8O2bdvM13Jzc9m9e3eRMb28vHj00Uf58MMPmT9/Pu+//775NRcXF+677z4WLVrEJ598wpo1a7hw4UK1xAsqd6+1Qt1M5R1HLx41fTvZoAHZMTHkJiRg27Bhqfc18XXB0sLAufQcTqVm4edW8oYNIiIiIiJiMmbMGObPn8/f//53xo4dS0xMDDNmzGD8+PHF1qMD5tLnJ598koULF2JlZcXTTz9d6mZphV588UVuu+02GjVqxMWLF3nttdeIi4vj8ccfv2aMI0eOZNSoUQBFzjgPCwsjPj6elStX0r59e7766ivzLHd5ODk5MXLkSCZOnIinpyfe3t5MnTq1yPsODAzExsaGt956i9GjR/Pbb78VK88PCgrCYDCwbt06+vfvj729fZH1/4WxDh48mFGjRvHee+/h7OzMlClT8Pf3Z/DgweWOuTQxMTHFrjVt2pSnnnqKiRMn4uHhQWBgIHPmzCEzM5ORI0cCMH36dNq2bUvTpk3Jzs5m3bp15v0FXn/9dXx9fWndujUWFhasWrUKHx+faj3CTjPptVQjt0YAJGUkkZmbiXVA4Q7vZa9Lt7M2bR4HKnkXERERESkPf39/1q9fz65du2jZsiWjR49m5MiRRTZk+7OlS5fi5+dH9+7dufvuu3niiSfw9vYuc5yUlBRGjRpFZGQk/fv3Jy0tje3bt5drFnno0KHY2tri4OBQ5Pi0QYMG8cwzzzB27FhatWrF9u3bmTZtWrnfO8Brr71Gt27dGDhwIL169aJr165F1tt7eXkRHR3NqlWraNKkCbNnz2bu3LlF+vD39+eFF15gypQp1K9fn7Fjx5Y41tKlS2nbti0DBgygU6dOGI1G1q9fX6zEvTLuv/9+WrduXeRx5swZZs+ezdChQ3nkkUdo06YNsbGxbNy4EXd3d8BUBfDss8/SokULbr/9diwtLVm5ciVg2jl/zpw5tGvXjvbt23PixAnWr19f4pc3VcVgvMVqotPS0nB1dSU1NbXcmynUlB6f9OB81nk+6v8R3v/3FReWLcMjKor6UyaXed+k1fv59KdExvZsxIQ+4WW2FREREREpr6ysLI4fP05ISAh2djpJSORqZf37qEgeqpn0WqxwNj32YizWAQEA5J689g7vzRu4AZpJFxERERERqWuUpNdiV69Lt27gD0BOQjmSdG0eJyIiIiIiUicpSa/FCpP02NRYbApn0hMTr5l4R/g4Y2Vh4EJGDkmpZR/ZJiIiIiIiIrWHkvRarLDc/ejFo1j7m2bSC9LTKUgtu4zdztqScJ8rm8clXqzWGEVERERERKTqKEmvxQpn0k9nnCbTIg8rLy+g4iXvIiIiIiJVSUsqRYqrqn8XStJrMVdbV7ztTcc4HE09inUD0zFs5ds8zpSk/5KoJF1EREREqkbhMVmZmZk1HIlI7ZOTkwOApaXldfVjVRXBSPUJdQsl+XIyRy8epWNAAy7//DM5CQnXvO/Pm8cZDIbqDlVEREREbnKWlpa4ubmRnJwMgIODg/47UwQoKCjg7NmzODg4YGV1fWm2kvRaLtQtlB2ndhB7MZZuhTPpiSeveV+4jzPWlgYuZuaSmHKZAA+H6g5VRERERG4BPj4+AOZEXURMLCwsCAwMvO4vrpSk13JFN4/rA0BuOWbSba1Mm8f9djKNX0+mKkkXERERkSphMBjw9fXF29ub3Nzcmg5HpNawsbHBwuL6V5QrSa/lzMewpcRiHTwSgJxyrEkHaO7vZk7S+zf3rbYYRUREROTWY2lped1rb0WkOG0cV8sVJunJl5PJru8GQG7SKYz5+de8t8WVzeN+1eZxIiIiIiIidYKS9FrO2caZ+g71AYizuQTW1pCbS96ZM9e898+bx4mIiIiIiEjtpiS9Dihclx576RjWfqay9ZzEa5e8N67vjI2lBamXc0m4cLlaYxQREREREZHrpyS9DigseT968Sg2DQIAyE24dpJuY2VBhK8zAL+cvFht8YmIiIiIiEjVUJJeB5hn0i/GYl14DFu5N4/7o+RdREREREREajcl6XVAkZn0AFOSnlOOmXS4KknX5nEiIiIiIiK1npL0OqAwST93+Ry59T2A8p2VDtC8gTaPExERERERqSuUpNcBjtaO+Dn6AZDkZkq0c+Ljy3Vv4/rO2FhZcCkrj7jzmdUWo4iIiIiIiFw/Jel1ROFs+jFnU6Kdf+EC+anXLmG3trSgqZ8LAD8npFRfgCIiIiIiInLdlKTXEYWbxx3OTsDK2xuAnLi4ct3bNtAdgJ9OKEkXERERERGpzZSk1xFFNo8LDgYg58SJct3bLlhJuoiIiIiISF2gJL2OuPoYtoom6W2DTJvNHU6+ROrl3OoIT0RERERERKqAkvQ6IsQ1BIALWRfIa3Cl3L2cSbqXsy3Bng4YjbA3XrPpIiIiIiIitZWS9DrCwdoBfyd/AM7WswYgu5xJOvwxm75HJe8iIiIiIiK1lpL0OqSw5P2ESzYAuSfiyn32uXldetyF6glORERERERErpuS9DqkcPO4g3YXwMKCgsxM8s6eLde97YJMSfq+hIvk5hdUW4wiIiIiIiJSeUrS65DCmfQj6cexbtAAKP+69FAvJ1ztrcnKLeD3pLTqClFERERERESug5L0OqTIMWxBQUD5k3QLCwNtgwqPYlPJu4iIiIiISG2kJL0OCXENwYCBi9kXKQjwASDnRFy57y9cl74nTpvHiYiIiIiI1EZK0usQeyt7ApwDALjgZQeUfyYdoN2VHd5/iksp94ZzIiIiIiIicuMoSa9jCkveE93zgYol6S0auGJtaeDspWwSLlyujvBERERERETkOihJr2MKN4+LcUoHICchAWNeXrnutbO2pJm/K6Cj2ERERERERGojJel1TOFM+q+GJAy2tpCbS25SUrnvLzyK7SetSxcREREREal1lKTXMWHuYQAcSY2t8A7vAG0L16Vrh3cREREREZFaR0l6HRPiGoK1hTXpuenk+3sDFU3STTPph8+kk5qZWx0hioiIiIiISCUpSa9jrC2szSXvKd72QMWOYfNytiWkniMAe+NV8i4iIiIiIlKbKEmvgxq7NwbgpLvpGLWKzKTDH7Pp2jxORERERESkdlGSXgdFeEQAcMgxFah4km7ePO6EZtJFRERERERqEyXpdVC4ezgAP9meAiD31CkKsrLKfX+7YFOSvj/xIjl5BVUfoIiIiIiIiFSKkvQ6KNzDlKQfLjiFwdkZjEZy4uPLfX/Dek64OViTlVvA70mp1RWmiIiIiIiIVJCS9DrI1daV+g71wWAgv0HFd3i3sDDQNtA0m75H56WLiIiIiIjUGkrS66jC2fSL3g5AxXZ4B2gbrHXpIiIiIiIitY2S9DqqcF16ZXd4bxfkAcBPcSkYjcYqjU1EREREREQqR0l6HWVel+54Cah4kt6igSs2lhacS88m/kJmVYcnIiIiIiIilaAkvY4qnEnfZ5cMVDxJt7O2pJm/CwC7VfIuIiIiIiJSKyhJr6MCnAOwt7InzjUHgPwLF8hPrdhO7e2DTSXvu46fr/L4REREREREpOKUpNdRlhaWhLmHkWVrIM/DNCOeE1exzeNua+gJwM7jF6o8PhEREREREak4Jel1WGHJe1p9R6DiSXq7YHcsDBB3PpOki5erPD4RERERERGpGCXpdVhhkp5kOk2NnOMnKnS/s501zf1dAdipkncREREREZEapyS9Divc4f2IUwZQ8c3jADpeKXn/31GVvIuIiIiIiNQ0Jel1WJh7GACHXdKByiXptzU0bR6nmXQREREREZGaV+NJ+r///W+Cg4Oxs7OjY8eO7Nq1q8z28+fPJzw8HHt7ewICAnjmmWfIysq6QdHWLo7WjgQ6B3LKwwCYknSj0VihPtoFe2BhgBPnMzmVqnXpIiIiIiIiNalGk/RPPvmE8ePHM2PGDPbu3UvLli3p06cPycnJJbb/6KOPmDJlCjNmzODgwYMsXryYTz75hOeee+4GR157hHuEc8YNjBYGCjIzyTt7tkL3u9hZ06xwXfoxlbyLiIiIiIjUpBpN0l9//XVGjRrFiBEjaNKkCe+++y4ODg4sWbKkxPbbt2+nS5cuPPjggwQHB9O7d28eeOCBa86+38wauzcm39JAuqcDUMl16SGmkvf/HVPJu4iIiIiISE2qsSQ9JyeHPXv20KtXrz+CsbCgV69e7Nixo8R7OnfuzJ49e8xJ+bFjx1i/fj39+/cvdZzs7GzS0tKKPG4mhTu8nzLl2ZVcl67z0kVERERERGoDq5oa+Ny5c+Tn51O/fv0i1+vXr8+hQ4dKvOfBBx/k3LlzdO3aFaPRSF5eHqNHjy6z3P2VV17hhRdeqNLYa5PCHd6PuWTRGMg5UbGz0uGPdenHz2VwOjULH1e7Ko5SREREREREyqPGN46riK1bt/Lyyy/zzjvvsHfvXj777DO++uorZs2aVeo9zz77LKmpqeZHQkLCDYy4+vk6+uJs48xJd9OGcZWZSXe1t6apn85LFxERERERqWk1NpNer149LC0tOXPmTJHrZ86cwcfHp8R7pk2bxiOPPMLjjz8OQPPmzcnIyOCJJ55g6tSpWFgU/87B1tYWW1vbqn8DtYTBYCDcPZwkD9MSgMok6WBal/7ryVT+d+wCg1v5V2GEIiIiIiIiUl41NpNuY2ND27Zt2bx5s/laQUEBmzdvplOnTiXek5mZWSwRt7S0BKjw0WM3k3CP8D+OYUtIwJiXV+E+zOvStXmciIiIiIhIjanRcvfx48ezaNEili1bxsGDB3nqqafIyMhgxIgRAAwfPpxnn33W3H7gwIEsXLiQlStXcvz4cTZt2sS0adMYOHCgOVm/FYW7h3PeBXJtLCA3l5z4ipf0tw/xwGCAY+cyOJN2a547LyIiIiIiUtNqrNwd4L777uPs2bNMnz6d06dP06pVK77++mvzZnLx8fFFZs6ff/55DAYDzz//PCdPnsTLy4uBAwfy0ksv1dRbqBUaezTGaDCQWM9ASFIB2UeOYNswpEJ9mNalu/DbyTT+d+y8St5FRERERERqgMF4i9WJp6Wl4erqSmpqKi4uLjUdTpXIzs+m44qOPLEuh56/GKk3dixeY/9W4X5mrTvA4v8e58GOgbx8V/NqiFREREREROTWU5E8tE7t7i4ls7W0JcQ1hIR6pnXp2UeOVKqfwnXp/9O6dBERERERkRqhJP0m0di9MQlepp8rm6R3CL6yLv1sBslaly4iIiIiInLDKUm/SYR7hBPvdWWH97g4CnJyKtyHq4M1TXxNpRf/O36hSuMTERERERGRa1OSfpMIdw8nxQky7SwgP5+c48cr1U/HEB3FJiIiIiIiUlOUpNdW+blwbCucP1qu5uEe4WAwEFfPtA9g9uHKrkv3ALQuXUREREREpCYoSa+t1j0DHwyGPUvL1byefT3qO9S//nXpV85LP3o2g+RLWpcuIiIiIiJyIylJr60a9TL9eXAdlPOUvGb1ml33Du9uDjZE+pjWpe/SunQREREREZEbSkl6bdXoDrC0hZTjcPZQuW5pVq/Zdc+kA3S8UvK+46hK3kVERERERG4kJem1la0zNOxh+vnQunLd0rxecxKu7PCem5hIQUZGpYbuEloPgO8Pn8VYzll8ERERERERuX5K0muziL+a/jz0VbmaN/FsQrqDBRcdTc+zjx2r1LCdG3liY2lBYspljp5Nr1QfIiIiIiIiUnFK0muz8H6AAZJ+htTEazZ3tnEm2DXYfF56ZXd4d7CxMpe8bzl0tlJ9iIiIiIiISMUpSa/NnLwhoKPp55gN5bqleb3mJJiq1a9rXXrPcG8AtsQkV7oPERERERERqRgl6bWdueS9fOvSm9VrRqLX9e3wDvCXCFOSvuv4BS5l5Va6HxERERERESk/Jem1XWGSfuK/cDnlms2beTYj/jqPYQMIrudISD1H8gqMbIs9V+l+REREREREpPyUpNd2nqHgFQkFeXBk0zWbh3uEc7q+FQB5ycnkp6ZWeuge4abz3LQuXURERERE5MZQkl4XVKDk3cbShiCfSM66mJ5nx8ZWetjCkvctMck6ik1EREREROQGUJJeFxQm6Ue+hdysazZvVq+Z+bz06yl57xDigb21JcmXsvk9Ka3S/YiIiIiIiEj5KEmvC/xag7Mf5GbA8e+v2dyUpJt+ruwxbAC2VpZ0aWTaKn6rdnkXERERERGpdkrS6wKDoUIl783rNTdvHpd15PB1Dd0z4sq69BitSxcREREREaluStLrisIkPWYDFOSX2TTYNZjzvg4AXD4cc13ryQvPS/85PoWUjJxK9yMiIiIiIiLXpiS9rgjuCraukHEWEneX2dTCYIFb46YUAKReIv9c5Y9Q83OzJ8LHmQIj/HBEs+kiIiIiIiLVSUl6XWFpDY37mH4uR8l7hF9LTrubfr6ezeMAelyZTd9ySOvSRUREREREqpOS9LqksOT94Dq4Rgl783rN/9jh/TqOYYM/jmL7/vBZ8gt0FJuIiIiIiEh1UZJelzS6AyxtIeU4nD1UZtOrd3jPiCm77bW0CXTD2c6KlMxc9iVcvK6+REREREREpHRK0usSW2do2MP08zVK3us71OeivwsAaYd+u65hrSwtuL2xKePXUWwiIiIiIiLVR0l6XXN1yXsZDAYDjo0jATAejbuuHd7hj13ev9O6dBERERERkWqjJL2uifgrGCzg1D64cLzMpgFNOpBnAZZZOeQlJV3XsD3CTTPpvyelcSYt67r6EhERERERkZIpSa9rHOtByO2mnw98XmbTpvVbkuRp+jnrOnd4r+dkS8sGrgB8H6Oj2ERERERERKqDkvS6qOldpj9//0/Zzeo1JaGeaYf31Otclw5/HMWmkncREREREZHqoSS9LooYCAZLOLUfzh8ttZmrrStpAabD0s/9vve6hy08iu3HI2fJys2/7v5ERERERESkKCXpdZGjJzTsbvr5GiXvNo1CAcg5cn1npQM093elvostGTn57Dh6/rr7ExERERERkaKUpNdV5Sx5r9+0HQD2iecx5l/f7LeFhYHeTXwA2Pj76evqS0RERERERIpTkl5XRQwACys4/SucK32WPKxJV7KtwCqvgOy4uOsetnfT+gBsOnCG/ILrO9ZNREREREREilKSXlc5eEDDHqafD5Q+mx7h1YQEb9Pmcad/3n7dw97W0BMXOyvOZ+SwJy7luvsTERERERGRPyhJr8vMJe+fl9rEzsqO8yEeAJze/cN1D2ltacEdkabZ9G9U8i4iIiIiIlKllKTXZRF/BQtrOPMbnD1cajPb5s0AyP31QJUM2+dKyfvGA6cxGlXyLiIiIiIiUlWUpNdl9u4Q2tP0cxm7vAd0/AsArnEXMObkXPewtzf2wtbKgoQLlzl46tJ19yciIiIiIiImStLrunLs8t6iTV8u2YN1npGT+65/XbqDjRXdwrwA+OaASt5FRERERESqipL0ui68v6nkPfkAJB8qsYmLrQtnAp0BOLr96yoZ1lzy/vuZKulPRERERERElKTXegXXOubM3g0a3WH6uYySd2PTMAAy9v1cJXH1iqyPhQEOnkoj4UJmlfQpIiIiIiJyq1OSXkvt/y6BD6ZuZ/+3CdduXI6S9/rtugLgeCSpKsLD3dGGDld2jd+oXd5FRERERESqhJL0Wqog38il81kkHSnHWeTh/cDSBs4eguSDJTZp0m0wAN7n80g8WXKbiurT1AeAb1TyLiIiIiIiUiWUpNdS/o3dAEiKTb12ybudKzTqZfq5lNl0Zy8/LtSzBeDAj2urJMbeV5L03XEXOJeeXSV9ioiIiIiI3MqUpNdS9Ro4YW1nSc7lPM6fTL/2DU2GmP787TMo5ezy7PBAAC7s+V+VxOjvZk9zf1eMRvj2gGbTRURERERErpeS9FrKwtIC31BXAJIOX7z2DeH9wMoOzh+BU/tKbOLauj0AVoeOYywlka+o3k0Kd3nXunQREREREZHrpSS9FvMLcwMg6cjFaze2c4GIAaaff15RYpPQLn0BCEjMJj4tvgoihD7NTCXv22LPk56dVyV9ioiIiIiI3KqUpNdi/o3dAVOSXq6Z71YPmv78dRXkZhV72aVpS/IsDbhchv0/V8156WHeToTUcyQnv4CtMclV0qeIiIiIiMitSkl6LeYV6IyVtQVZGblcOJVx7Rsa9gAXf8i6CIc3FHvZwsaGjBBvAJJ2f18lMRoMBno3LSx517p0ERERERGR61GpJD0hIYHExETz8127dvH000/z/vvvV1lgApZWFvhcWZd+qjwl7xaW0PIB08+llLzbt2gBQMFvMVW4Lt1U8r7lUDLZeflV0qeIiIiIiMitqFJJ+oMPPsiWLVsAOH36NHfeeSe7du1i6tSpvPjii1Ua4K2ucF36yfIk6fBHyfvRzZB2qtjLDTr2NP0Zn8nxtONVECG0DnDDx8WO9Ow8Nh9UybuIiIiIiEhlVSpJ/+233+jQoQMAn376Kc2aNWP79u2sWLGC6OjoqozvlmfePO5wOdele4ZCYCcwFsAvK4u97Ny6LQDByfBT/I4qidHCwsDdbfwB+PSnhCrpU0RERERE5FZUqSQ9NzcXW1tbAL799lsGDRoEQEREBKdOFZ+9lcqrH+KCpZUFmWk5pCZfLt9NhbPp+z4qdma6dUAAOc52WOfDsZ++q7I4h7ULAOCHw2c5nVp80zoRERERERG5tkol6U2bNuXdd9/lxx9/ZNOmTfTtazraKykpCU9PzyoN8FZnZW1J/RAXoJxHsQE0vQusHeDcYUj8qchLBoMBy6bhAGT98kuVrUsPqedI+2B3Cozw2c+J175BREREREREiqlUkv7qq6/y3nvv0aNHDx544AFatmwJwNq1a81l8FJ1/liXnlK+G2ydIdJU3cC+4hvI1Wvb2dRvXAZHLx6tihABGNbWNJu+6qfEKkv+RUREREREbiWVStJ79OjBuXPnOHfuHEuWLDFff+KJJ3j33XerLDgxMa9LL+9MOkDrh0x//vYZ5BYtk3e6si49LMnIrtO7qiBCk/4tfHGwseT4uQz2xJXzCwURERERERExq1SSfvnyZbKzs3F3dwcgLi6O+fPnExMTg7e3d5UGKODT0BULCwPpF7JJO1fOdelBXcEtELJT4dBXRV6yb9Hc1O9F+CV2W5XF6WRrRf/mvoBpNl1EREREREQqplJJ+uDBg/nggw8AuHjxIh07dmTevHkMGTKEhQsXVmmAAta2lngFOQMVmE23sICWVzaQ+/nDIi9ZurhQEOgHwMWfd1NgLKiqUBnWtgEA635JIjMnr8r6FRERERERuRVUKknfu3cv3bp1A2D16tXUr1+fuLg4PvjgAxYsWFClAYqJf2M3oIIl760eMP15bCukFp3Zdr1S8u4Xl86RlCPXH+AVHUI8CPZ0ICMnn/W/nq6yfkVERERERG4FlUrSMzMzcXY2zex+88033H333VhYWHDbbbcRFxdXpQGKiV+YaWnByYok6e7BENwNMML+j4u85NCyFQBhSfC/U/+rkhjBtHv8PVdm03VmuoiIiIiISMVUKklv1KgRn3/+OQkJCWzcuJHevXsDkJycjIuLS5UGKCa+oa4YDJB29jLpKdnlv7GUM9PtW5h25G90ysj3CVurLlDg7jYNMBhg1/ELxJ3PqNK+RUREREREbmaVStKnT5/OhAkTCA4OpkOHDnTq1Akwzaq3bt26SgMUExt7K+oFXFmXHluBndObDAYbJ7hwDOL/mDG3C28MtjY4ZUHSwT2kZqdWWax+bvZ0C/MCYPUebSAnIiIiIiJSXpVK0u+55x7i4+P56aef2Lhxo/n6HXfcwRtvvFFlwUlRfoXr0g9fLP9NNo7QZIjp5/0fmS8brK2xb9IUgNCT+fyQ+EPVBHlF4QZyq/ckkl+gM9NFRERERETKo1JJOoCPjw+tW7cmKSmJxETTbGmHDh2IiIiosuCkKL9GbkAFN4+DPzaQ++0/kJNpvmzfqhUA4YlGtiRsuf4Ar3Jnk/q42FlxKjWLbbHnqrRvERERERGRm1WlkvSCggJefPFFXF1dCQoKIigoCDc3N2bNmkVBQdUd5yVF+YW5gQFSTmeSmZZT/hsDO4NbEORcKnJmukP7dgBEJhjZdnIbOfkV6PMa7KwtGdLaH4BVKnkXEREREREpl0ol6VOnTuXtt99m9uzZ/Pzzz/z888+8/PLLvPXWW0ybNq2qY5Qr7Byt8fRzAio4m25hAS2vzKbvW2G+7NCmDQANzoNVagY7T+2sqlABGNY2AICNv58mNTO3SvsWERERERG5GVUqSV+2bBn/93//x1NPPUWLFi1o0aIFY8aMYdGiRURHR1dxiHI1vzA3AJJiL1bsxpb3m/48thVSTwJg6eaGbePGgGk2fWvC1qoI0ayZvwsRPs7k5BWweq9m00VERERERK6lUkn6hQsXSlx7HhERwYULF647KCldYZJ+MqYCO7wDeIRAUFfACL+sNF92aPdHyfvWhK0UGKtuuYLBYODh24IA+GDHCQq0gZyIiIiIiEiZKpWkt2zZkrfffrvY9bfffpsWLVpcd1BSugYR7hgMcCEpg0sXsip2c+EGcledmV64Lr1pgoHky8kcOH+gKsPl7jb+ONtZEXc+k62Hk6u0bxERERERkZtNpZL0OXPmsGTJEpo0acLIkSMZOXIkTZo0ITo6mrlz51Z1jHIVO0dr6oe4ABD/+/mK3dxkMFg7wPlYSPwJAPu2bQEITC7AIavqd3l3sLHi/vamtelLt52o0r5FRERERERuNpVK0rt3787hw4e56667uHjxIhcvXuTuu+/m999/Z/ny5VUdo/xJYFNPAOIPVHBpga0zRA4y/XxlAzlrb29sgoIwGKvnKDaA4Z2CMRjgxyPniE2+VOX9i4iIiIiI3CwqfU66n58fL730EmvWrGHNmjX861//IiUlhcWLF1dlfFKCwiQ98eAF8vMruIa81YOmP3/7DHJN5fL2V5W8H0k5QuKlqt3kLcDDgV6R9QFYtj2uSvsWERERERG5mVQ6SZea4x3ojL2zNTlZ+Zw+mlqxm4O7gWsAZKdCjOnM9MLN49qdtgeoltn0EZ2DAVizN5HUyzqOTUREREREpCRK0usgg4WBgCYeQCXWpVtY/HEc276PAXBo1x4An4RMbHOq/ig2gE6hnjSu70RmTj6rfkqo8v5FRERERERuBkrS66jAJqaS97jfK3HkXcsru7wf3Qxpp7D298PK1xeL/ALCkozsObOH1OwKztBfg8FgIKpzCAAf7IgjX8exiYiIiIiIFGNVkcZ33313ma9fvHjxemKRCghs6gEGOJ+YTsbFbBzdbMt/s2coBNwGCf+DXz7B0PVpHNq1I+3LL+mS7M5vwan8kPgDA0MHVmnMd7X259WvDxF/IZMth5Lp1aR+lfYvIiIiIiJS11VoJt3V1bXMR1BQEMOHD6+uWOUq9k42eAeZjmKLq2jJO/xxZvr+j8FoNK9Lb33KlOxXR8m7vY0l93cwHccWvf1ElfcvIiIiIiJS11VoJn3p0qXVFYdUQlBTD5JPpBH/+3madPGr2M1N74INk+HsIUjai8OVHd7dY89ilWfkvyf/S05+DjaWNlUa8yO3BbHoh2P8N/YcR85cIqy+c5X2LyIiIiIiUpdpTXodVngUW8LBFAoqehSbnStEXiln3/V/2ISEYOnhgSEnl7YX3MjMy2TX6V1VHDE0cHegdxMfQLPpIiIiIiIif6YkvQ7zDnbBztGanMt5nD6eVvEOOj5l+vPXTzGkJZlL3ntf9Afgu/jvqirUIqK6BAPw2d6TpGbqODYREREREZFCStLrMIurj2L7rRLr0hu0haCuUJAHOxeak/SIeNOs/Ldx35JbUPVJdMcQDyJ8nLmcm8/K3fFV3r+IiIiIiEhdpSS9jgtqakrSK7V5HECXcaY/f4rGoUUEALYHjlPP2p2U7BT+l/S/qgizCIPBwGNdTMexRW8/QW5FS/VFRERERERuUkrS67iAK+eln0tIJyM1u+IdNLoTvCIh5xK2aT9i4exMQUYGwyw6APDV8a+qMlyzQa38qOdky6nULL765VS1jCEiIiIiIlLXKEmv4xxcbPAKNO2QnnDgQsU7sLCAzn8HwLDrPRzatAbg9nOm5P+7+O/IzM2smmCvYmdtSVTnIAAW/XgMo9FY5WOIiIiIiIjUNUrSbwJBzUwJdaVL3psPA2dfSD+Ng7/pVD7Xgydp4NSAy3mX+T7x+6oKtYiHOgZhb23J70lp7DhaydhFRERERERuIkrSbwLmo9gOXKj4UWwAVjZw2xgAHHJMa9Av//QT/YP7AfDVseopeXd3tGFYuwYAvP/jsWoZQ0REREREpC5Rkn4TqB/sjK2DFdmZeZw5calynbSNAlsX7CxiMdjakJ+aSj+aAbDt5DYuZl2ssnivNrJrCAYDbI05y+EzlYxdRERERETkJqEk/SZgYWlBQOSVo9gqW/Ju5wLtRmCwAAcf018Lj5jTRHpEkmfM45u4b6oq3CKCPB3p29QHgP/TbLqIiIiIiNzilKTfJApL3iudpAN0fAosrLF3TgYgc+cu+of0B6qv5B3g8W4NAfj85ySS07KqbRwREREREZHarlYk6f/+978JDg7Gzs6Ojh07smvXrlLb9ujRA4PBUOzx17/+9QZGXPsEXjkvPTnuEpcuVDLRdfGFFvfh5GM6yi1j2zb6NOiFAQN7k/dyKr16jkprG+RO2yB3cvILWLbjRLWMISIiIiIiUhfUeJL+ySefMH78eGbMmMHevXtp2bIlffr0ITk5ucT2n332GadOnTI/fvvtNywtLRk2bNgNjrx2cXS1xS/MDYDDu05XvqPOf8fOIxdL23wK0tNxOZRIO592AGw4saEKIi3ZqCuz6R/+L57MnLxqG0dERERERKQ2q/Ek/fXXX2fUqFGMGDGCJk2a8O677+Lg4MCSJUtKbO/h4YGPj4/5sWnTJhwcHG75JB0gvKNpbXfMzjOVP3fcOwJDRF+c/Eyz6elbtt6Qkvc7m9QnyNOB1Mu5rPopsdrGERERERERqc1qNEnPyclhz5499OrVy3zNwsKCXr16sWPHjnL1sXjxYu6//34cHR1LfD07O5u0tLQij5tVaBsvLK0sSDmVwbmE9Mp31OUfOPmZSubTt2zmzqA7sbKw4nDKYY6kHKmiaIuytDDweNcQAP7vv8fIL6jklwwiIiIiIiJ1WI0m6efOnSM/P5/69esXuV6/fn1On752yfauXbv47bffePzxx0tt88orr+Dq6mp+BAQEXHfctZWtgzXBLeoBELPzOkreAzvh2LYpWBjJiU/E7uQFuvp3BWDD8eoreb+nbQDuDtYkXLjMxt+vI34REREREZE6qsbL3a/H4sWLad68OR06dCi1zbPPPktqaqr5kZCQcAMjvPHCbzOVvB/efYaC/ILKdWIwYNlzPI5eV0rev93IXxuaNuZbf3x95Uvpr8HexpJHbgsCYOHWo9U2joiIiIiISG1Vo0l6vXr1sLS05MyZM0WunzlzBh8fnzLvzcjIYOXKlYwcObLMdra2tri4uBR53MwCm3hg52jN5bQcEg6lVL6j8P44hZk+q/SvVtG9QXccrBw4mX6S/Wf3V1G0xT3aORgHG0t+PZnKpgNnrn2DiIiIiIjITaRGk3QbGxvatm3L5s2bzdcKCgrYvHkznTp1KvPeVatWkZ2dzcMPP1zdYdYpllYWhLXzBuDw9ZS8W1jgNPQxADIPJ2GTmsEdgXcA1buBnKeTLVGdgwF4fdNhCrQ2XUREREREbiE1Xu4+fvx4Fi1axLJlyzh48CBPPfUUGRkZjBgxAoDhw4fz7LPPFrtv8eLFDBkyBE9Pzxsdcq3X+ErJ+7F9Z8nJqvxxZjZ3jsbGzQhGSP/kdfo3NO3yvv74ejJzM6sk1pI8cXtDnG2tOHT6Eht+09p0ERERERG5ddR4kn7fffcxd+5cpk+fTqtWrdi3bx9ff/21eTO5+Ph4Tp06VeSemJgY/vvf/16z1P1WVT/YBVdve/JyCji272zlO7Kyxfm21gCkf/0lnep3JMA5gLScNP4T+58qirY4NwcbHruy0/sb3x7WTu8iIiIiInLLMBhvsd250tLScHV1JTU19aZen777q+Ps+vI4AZHuDPpH60r3k7nje+JGjMbCpoDGK19nlWUGs/43C19HX766+yusLayrMOo/pGXl0u3VLaRezuWN+1pyV+sG1TKOiIiIiIhIdatIHlrjM+lSPRp3MJW8Jx5KIeNidqX7sW/fBUsHGwpyLLi86jUGhQ7Cw86DUxmn+ObEN1UVbjEudtY8cXtDAN789gh5ld2pXkREREREpA5Rkn6TcvWyxzfUFaMRDu+q/C7pBisrHLvfDsCln49il7SfByMeBGDpb0ur9Zi0qM7BeDjacOJ8Jp/tPVlt44iIiIiIiNQWStJvYo07mmbTY65nl3fA+c5+AKQn2cK2N7k/4n7sreyJSYlhR9KO646zNI62VozufmU2ffMRcvI0my4iIiIiIjc3Jek3sUZtvbGwMnD+ZDrnEtMr3Y9j165gaUlOmjU5P32Na9oZhoYNBWDJ70uqKtwSPXJbMF7Otpy8eJlPf0qo1rFERERERERqmpL0m5idozXBzeoB13dmuqWLCw7t2gGQnmQH299keJPhWBos2XlqJ7+f/71K4i2JvY0lY3qEAvD2d7Fk5eZX21giIiIiIiI1TUn6Ta5xR9NRdod3nabgOo4yc+rRA4BLSXawfyW+OVn0CzGVwS/9bel1x1mWBzoE4utqx+m0LD7eFV+tY4mIiIiIiNQkJek3ueBm9bB1sCIjNYeEAxcq3Y9zzx4AZJ61JT87H7bOJqppFACb4jaRkFZ9peh21paM/UsjwDSbfjEzp9rGEhERERERqUlK0m9yltYWRNzmC8Av31U+kbYJDsYmOBgKIOOULfzyCeF5BXTx70KBsYBlB5ZVUcQlG9Y2gFAvR85n5PDSVwerdSwREREREZGaoiT9FtC8ZwMwQPyBC1w4lVHpfgpL3tMzGgFG2PISjzV9DIDPYz/nQlblZ+qvxcbKgleHtsBggFV7EvnxyNlqG0tERERERKSmKEm/Bbh62RPSwrSB3C9bEivdj1PPngCkH8/BmG+Ag1/SvsCKZp7NyM7P5uNDH1dJvKVpF+zBo52CAXj2s1/JyM6r1vFERERERERuNCXpt4iWdwQAELPjFFkZuZXqw6FdW6y8vMhPS+eSzV8AMGx5iRHNRgDw8aGPSc+p/FFv5TGxTzj+bvYkplxm7jcx1TqWiIiIiIjIjaYk/RbhF+aGZwMn8nILOPDfpEr1YbC0xGXQQABSE9zAwgpiv+UOHAh2CSY1O5VFvy6qwqiLc7S14uW7mwMQvf0Ee+NTqnU8ERERERGRG0lJ+i3CYDDQ8i+m2fRftyaSn19QqX5cBw8GIH3HT+Q1vhcAy+9eYkK7fwKw/MByEi5V307vAN0be3F3G3+MRpi8+hey83R2uoiIiIiI3ByUpN9Cwtp7Y+9sTXpKNsd+rtzGa3aNG2PbJBJyc0lLbw6WthC/nduzcunk24ncglxe/+n1Ko68uGl/bUI9JxuOJKfz7y1Hq308ERERERGRG0FJ+i3EytqSprf7A9d3HJvbldn01G++h/aPA2D47l9Maj8RC4MF38Z/y+7Tu68/4DK4O9owc1BTAN7ZEsuh02nVOp6IiIiIiMiNoCT9FtPsdn8sLA2cPpbGmeOVS2xdBgwAS0uyfvmF7AZ3g7UjnNpHo1MHGdZ4GABzds8hv6B6y9D/2tyXO5vUJ6/AyOTVv5BXyRJ+ERERERGR2kJJ+i3G0dWWsPb1Adhfydl0K09PnLp1AyB103/htqdML2x5ib+1GI2zjTOHLhzii6NfVEnMpTEYDPxrSDOc7azYn5jKaxu127uIiIiIiNRtStJvQYUbyB3dk0x6Snal+nAdcqXkfe1ajLf9Dezc4Owh3A98yVMtTUn7m3vfrPYj2eq72PHq0BYAvPfDMTb8eqpaxxMREREREalOStJvQV6Bzvg2cqWgwMhv3ydWqg+nnj2xcHYm79QpMn89Aj2mmF74bhb3B/Yl2CWYC1kXqv1INoD+zX154vaGAExYtZ/Y5Or9YkBERERERKS6KEm/RbW8wzSb/vuPSeTlVHztuIWtLS79+gGQ+sUXpg3k6jWGzPNY//cNJrSbANyYI9kAJvUJ57aGHmTk5DP6wz2kZ+dV+5giIiIiIiJVTUn6LSqkpRfOnnZkZeRyaEflSsRdhwwB4NLGjRRk50KfV0wv7HqP2+186ezX+YYdyWZlacFbD7ShvostscnpTF79C0ajsdrHFRERERERqUpK0m9RFhYGWvUyzabvXn+C3ErMptu3boV1YCAFmZlc+vZbCOsFYX2gIA/DN1OZ2O6PI9l2ndpV1W+hGC9nW955qC3Wlga++vUUi/97vNrHFBERERERqUpK0m9hTbv64+xpR2ZqTqXOTTcYDLgOHgRA6udXdnLv8xJYWMGRb2h07jj3hd8HwEs7XyI3P7fKYi9N2yB3pg1oAsArGw7xv2Pnq31MERERERGRqqIk/RZmaW1Bx0GmDdf2fh1HVnrFk2jXQaYkPWPHDnLPnIF6YdBxtOnFjc8xtsVoPOw8OJZ6jA8OfFBlsZflkduCuKu1P/kFRsZ+tJczaVk3ZFwREREREZHrpST9Fte4fX08GziRk5XPnq9PVPh+m4AA7Nu1BaORtC+/NF28fSI41INzh3HZ/4l5E7n3fnmPU+nVf0SawWDg5buaE+HjzLn0HP711cFqH1NERERERKQqKEm/xRksDHS6KxSAX7YmculCxWedXQebzky/+Pnnps3a7N3gL8+bXtz6CgPqd6Jt/bZczrvMq7tfrarQy2RvY8m8e1tiYYAv9yex46jK3kVEREREpPZTki4ENvHAv7EbBXlGdn15rML3u/Tti8HWlpzYo2Tt32+62GY41G8OWakYtr7M1I5TsTRYsjl+Mz8k/lDF76BkTf1cebBjIAAz1/5OXn7BDRlXRERERESkspSkCwaDgU53NQIg5n+nOX8yvUL3Wzo749K/PwApH39sumhhCf1mm37es5Sw7GweafIIAK/sfIWsvBuzTvyfd4bj5mBNzJlLLP9f3A0ZU0REREREpLKUpAsA9UNcCG3thdEI//ui4rPp7g8+AEDa+g3kpaSYLgZ3hSaDwVgAGybxVIvReDt4k5ieyJLfllRl+KXH5WjDxD7hALy+6TDn0rNvyLgiIiIiIiKVoSRdzDoObojBwsCJX86RFHuxQvfaN2+OXbNmGHNzSV2z5o8Xer8EVvYQtw2HmA1Maj8JgMW/LiY+Lb4Koy/d/e0DaebvwqWsPOZ8feiGjCkiIiIiIlIZStLFzN3HkcguvgDs+OyoaRO4itz/gGk2PWXlJxjz800X3QLg9n+afv7meXr7dKKzX2dyCnJ4eefLFR6jMiwtDLwwqCkAn/6UyM/xKdU+poiIiIiISGUoSZciOvw1BCtrC04fS+X4/nMVutelfz8sXF3JTUwk47///eOFTn8H9xC4dArDj3N5ruNzWFtYsy1pG2uPrq3id1CytkEe3N3GH4AZa3+noKD6vxwQERERERGpKCXpUoSjmy0t7ggAYMd/jpKfV/4d0S3s7XG76y4AUj76+I8XrO2g35Wj13a8Q1B2No83fxyAGdtn8M2Jb6om+GuY0i8CJ1srfklM5dOfEm7ImCIiInVZ4qVELmRdqOkwRERuKUrSpZg2fYKwd7bm4plMft2aWKF73e+/D4D0H34gJ/Gqexv3gcZ9oSAXNkziyeZPMCh0EPnGfCb/MJlv476tyrdQIm9nO57uFcb/s3ff4ZFcZb7Hv51zt9TKWRpN0uScHHDE4IANJto4Ab7LEnbBXGAxywK7cGHXBC9gliVHR8DYGGcbx7En5zzKOXfOVXX/qB5pZGk0mqCRZub9PE891equrq6aHqn7V+ec9wD81zMHCMbSk/6aQgghxHQQS8fY2r2VNzvfJK0c//Nvb/9evvDyF7jm0Wv4zZ7fnIEjFEIIcYR5qg9ATD82h5k1N9Ty99/tZ9Pfmpi9qhin1zqh51qrq3FdcAHR118n8NBDFH7uc8MPvuNbUP8iNPwd08Gn+Pd1/46qqTzR8ASff/nzfOeS73B55eWTdFa629ZV8+CmVg73RPjMQ9v48c3LcVhNk/qaQgghxJkUS8fYP7Cfvf172du/lz39e2gMNqKhD/XyWr1cWXUl76h5ByuLVmIy6p+DmqaxvmM9v9rzKzZ0bhjaX1v4xC7YCyGEODUG7UxU7ppGQqEQPp+PYDCI1+ud6sOZtlRV44/f3kxvS5h5F5Vy6c1zJ/zc8Asv0PbJT2HKzWXmS3/HaLMNP/jiN+CVe8BXCZ/cgGK2cfdrd/Nk45OYjWa+f8n3uaTiktN/QkfZ1DTAzT/fQCqjsrQyh1/cthK/a2IXIYQQQogzLaWkMGDAbDRjMBhGPBZLx9g3sG8okO/t3zsikB+t0FGIoin0J/qH7st35PP2qrdTm1PLwwce5sDgAQBMBhPvqHkHt8+/nbn+iX8HEEIIMbYTyaES0sUxdRwO8Oh3toIB3n/3SgoqPBN6npbJcPjKt5Pp7KT0v/4T37veNfxgKgb3rYJgK7zti3Dp3WTUDHe/ejdPNT2F2Wjmvy/9by4uv3iSzkq3qWmAj/1mM8F4mhkFLn5zxyoq/M5JfU0hhBDiRMQzce7dci8PH3iYjJYBwGayYTVZsRqtmIwmemO9xwzk8/LmMS9/HvPz5jMvbx75jnwUVWFL9xaebHyS55qfI5QKjXiew+zgxlk3csu8Wyh1l56R8xRCiPOBhPRxSEg/Mc/+fDeHNvdQOiuHG+5aOuoK/rH0/eQn9N773zgWL6b6oQdHPrj3MXj4VjDZ4JMbwF9DRs3wxVe+yLPNz2IxWvjO277DZZWXTcIZDTvUHea2X26kI5igwGPjV7evZEGZb1JfUwghhJiIPf17+NKrX6Ix2HjcbQud2UCeNzKQH09aSfNG5xs81fgU9YF6rqy6kvfPeT8+m3wWCiHE6SYhfRwS0k9MeCDB/V99k0xa5ao7FzBzeeGEnpfp7eXQZZdDOk31n/6IY/784Qc1DX53AzS8BDMuhVseBYOBtJrmi698keeanwPglnm38Jlln8Fqmryu6F3BBLf/aiP7u8K4bWZ+8uHlXDjr+F9shBBCiMmQUTP8Ytcv+MmOn5DRMuQ78vn6uq+zpHAJKSVFSkmRVJKklBRpNU2xq3hCgVwIIcTUkpA+DgnpJ27jE41seqIRt9/GzV9bg3mChdba7/ocoSefJOd976XkP/5j5IP99fA/6yCTgOt+AMtvAyCtpvnOpu9w//77Aajz13HP2+6hylt1Ws/paMF4mn/43WbebBjAYjLwnfct5volZZP2ekIIIc5tHZEO/lr/V8KpMDn2HHw2Hz6rjxxb9rbNh9fqxWF2jOih1hJq4e7X7mZH7w4Arqy6kn9b82/k2HOm6EyEEEKcLhLSxyEh/cSlUwr3f/VNIoNJVl1Xw8praib0vNjmzTR/+BYMdjuzXn4Jk+8t3efW/xCe/VeweeETb4JvOBi/1PoSX3n9KwSSAZxmJ/+65l+5rva603hWIyUzCnc9vIO/7ezEaIDvvn8x715aPmmvJ4QQYupomkYgGaAr2qUvsa6h22k1TW1OLXNy5zAndw5lnjKMhuPPWJtRM7za9iqPHHyE19pfG3Oc+FuZDWa8Ni8eqwev1cvhwGHimThui5u7V9/NtTOunfAwMyGEENObhPRxSEg/OYc2d/Psz/dgthi56etr8Pjtx32Opmk0Xn8DyYMHyf/0pyj45CdHbqAq8MuroG0TzLwSbn4Ejvoy0h3t5l9e/Rc2d28G4NoZ1/Kva/4Vl8V1Ws9t6HBUjS//ZRcPbGzFaIDvvX8JNyyVFnUhhDjbqZrKocFDbO7ezOauzWzp3sJgcnBCz3VZXMzKmcWs3FkUOArIteeSa8/Fb/eTY8vBYrTwVNNT/Ongn+iOdQ89b3XJaubkziGYDBJMBQkmgwSSAYLJIKFkaKgQ3FutLF7JNy74hhRtE0KIc4yE9HFISD85mqbx6He30nk4yMzlhbz9Y/MndHU/9NRTtH/2LowuF7XPP4c5N3fkBr0H4CcXgZKEG/4Hltw04mFFVfjZrp/xPzv+B1VTKXOX8fV1X2d1yerTeXpDVFXj7kd38eAmPah//wNLpOu7EEJMMU3Thub6Xlm8kmJX8XGf0x3t5vmW59nQuYGtPVsJJoOjtvHb/RS7iilxlVDsKqbYWYzJaOLg4EEODBygPlBPSk1N+DhzbDncMPMG3jv7veMO09I0jXgmTigVIpwKD62tJitrStZMqOVeCCHE2UVC+jgkpJ+83pYwj3xrE5oGb/vQbBa87fjdwTVVpfG97yW5dx/+O+6g6ItfGL3Ra9+H578Gdh98YgN4S0Ztsq1nG1985Yt0RjsBuHHWjdy14i681tP/HkpQF0KIqaeoCtt7t/N88/M83/I8XdGuoceWFCzhquqruLLqSopcRUP3DyQGeL75eZ5sfJKt3VtHdDl3mB0sLVzKiqIVrChewby8edhMtnGPIaNmaAo2cWDwAA3BBgYSAwwmBhlMDOq3k4NEUhEWFyzm/XPez5VVV05qsVMhhBBnLwnp45CQfmq2PdvC+j8fxmgycMNdyyipPf40LZFXX6X1zv+DwWql9tlnsBS/pQVEycAvroSOrTD7nfChB0Z0ex/aTyrCvVvv5aEDDwFQ4Cjgy2u+zOWVl5+WczuaBHUhhDjzNE1jS/cWnmp8ihdbX6Qv3jf0mMPsoMZXw77+fSPC97LCZawtXcv2nu282fkmiqaMeOxtFW9jRdEK6vLqsBgtk3LMMm5cCCHE8UhIH4eE9FOjaRrP/nwPh7f04PRaef+XV+Lyjd8SoWkaLbfcSmzz5rErvQN074X/vRjUNLznZ7Do/cfc35buLXxt/ddoCjUB8Paqt/Ol1V867VPQqKrGl/68i4c2S1AXQojJ1Bfv4/H6x3n00KNDf9sBPBYPl1RcwhVVV7CudB12s32oG/szTc+wrWfbqH3Ny5vHO6vfyVXVV1HiHt0zSwghhJgKEtLHISH91KUSGf70X1sY6IhSPMPHDXctxWQef/xcbOs2mm+6CUwmZjzxV2w1Y1SIf/ke+Ps3wJGrd3v3FI3eJiupJPnJjp/wq92/QtEUPBYPH1n4EW6uuxmH2XGqpzhEVTX+5c87eXhzGwYDfPXaedx+wcSq2wshxPkskUnwRMMT/LX+r1iMFso95VR6K6nwVFDhqaDMXcaO3h386eCfeKn1paFCak6zk6uqr+Lt1W9ndfFqLKZjt353Rbt4rvk5tnRvoc5fxztq3jGpU3YKIYQQJ0tC+jgkpJ8ege4Yj3x7M6l4hgVvK+NtH5pz3Oe0fvwfibz0Ep53voPy739/9AZKGn52GXTthLrr4AO/P+4+9/Xv46vrv8q+gX0A5Dvy+YdF/8CNs24c94vdiVBVja88tps/bGgB4M6LavjSO+swGqV7oxBCvFVfvI8H9j/AIwcemXAFdYBF+Yu4cfaNvKP6HTgtzkk8QiGEEOLMk5A+Dgnpp0/Tzj7+9uOdAFx2ax1168bvVpg4cIDGG94Nmkb1n/6IY/780Rt17YKfXgJqBt73G5h/w3GPQ1EVnmx8kvu230d7pB2Acnc5n1z6Sa6uufq0VMnVNI0fv1TPPc8cAOCaRSV8932LsVtMp7xvIYSYTrqiXbzZ+SZpNc2akjVUeCom9LwDAwf47d7f8lTjU6TVNAClrlJuqruJXHsuLaEWWsOttIXbaAm3EEgG8Fq9XFd7He+Z9R5m586ezNMSQgghppSE9HFISD+9Nj7RyKYnGjGZjbzn88sorBr/37T9/36e0BNP4LroIip/9tOxN3rxG/DKPeAqgE9uBKd/QseSVtI8cvARfrrzp/Qn+gGYlTuLu5bfxYVlF57QeR3Lo9va+MIfd5JWNFbV+PnpLcvJcUolXyHE9KSoCs+1PMfjhx/HbXFT46uhJqeGGm8N1b5qbCYbsXSMzd2beaPjDdZ3rKch2DBiH9Xeai4ou4ALSi9gZfFK7GY7sXSMPf172NW3i129u9jZt5OeWM/Qc5YULOGWebdwWeVlmI3mMY8tkopgM9smpZibEEIIMd1ISB+HhPTTS1M1nvyfnTTt6sftt/GBL6/C7jr2F65USwv1V18DmQxVv/stzpUrR2+USepF5Hr3w6IPwnv+94SOKZaO8Yd9f+BXu39FOB0GYF3pOu5afhdz/Mfvln88rx/u4+O/20I4mWFmoZtf37GS8lzpmimEmD4UVeHppqf56c6fjgrdRxgwUOIqoTfeO9TyDWA0GFmQtwCz0czO3p1DY8UBbCYbZe4ymkJNqJo6Yn9mg5krqq7glnm3sKhg0eScmBBCCHGWkpA+Dgnpp18ylubhb20m1Btn9uoirrxjjG7sR+n82tcIPPgQjqVLqbr/D2NPXdO6SZ+WDQ1uegRmv/2EjyuYDPKznT/j/v33k1bTGDBww8wb+OSST46YV/dk7O8KcfsvN9EVSlDgsfHTW5aztDL3lPYphDg/nc4pvDJqhqcan+KnO386VCXdY/Vw09ybcFlcNAYbaQg20BBsIJwKDz2v1FXKurJ1rCtdx6riVfhs+vSa4VSYjZ0bea3jNV5rf23EXOVFziIWFSxiYf5CFuYvZF7ePBlLLoQQQhyDhPRxSEifHF0NQf58zxY0Dd7xfxZQu6zwmNumu3uov+oqtESC8vt+hOfyY8xz/syX4Y0fgbcMPvEm2E/u/WoNt/KDrT/g6aanAX2u3dvm38Yd8+84pS+UncE4d/xqE/u7wljNRr717oXcuLz8pPcnhDg/aJpGY7CRV9tf5dW2V9nSs4UqTxUfnPtBrqu9DpfFdUL7i6VjbO/ZzubuzTzd9DSt4VYAfDYft827jQ/N/RBuq3vUMQwkBmgONZPnyKPSU3ncCwWaplEfqKcj2sGc3DmnfLFTCCGEOJ9ISB+HhPTJ88Zf6tn6dDN2l4UP/tuqcedP7/nu9+j/2c8wl5Yw4/G/YnKP8aU0FYP/WQeDjbDiI3DtGBXhT8CO3h18d/N3h+bVLXQU8tkVn+WammtOuhUrkszw2Ye289zebgA+emENX3rnXMymUy9WJ4Q4u6WUFOFUmEg6QiQVoSfWw+sdr/Na+2tDRS7fym1xc/3M6/nAnA9Q4xs93aOiKvTGezkwcIAt3VvY3L2Zvf17UTRlaJtcWy63zb+ND8794AkHfiGEEEJMDgnp45CQPnmUjMoj395Mf1uE6kX5XP2PC48ZftVolIZ3XU+6vZ3cm26i+N++MvZOG1+F31yr377tCai56JSOUdM0Xmh5ge9u/i5tkTZAL3D0L6v/hfl543fTPxZV1bj3+YP84MXDAFw0K58ffmipFJQT4iwXTUfZ2r0Vu9nO4oLFWE3H/p1OKSlebX+VJxueZGvPVkLJECk1dcztLUYLK4tXclHZRawsXsnm7s08uP/BoS7qoNfSWJC/gK5oFx2RDjqjnXRHu0eMET+i1FXKiuIVrChawVXVV0m3cyGEEGKakZA+Dgnpk6uvLcIj396EmtG47Na51K0rPea20fXrafnIRwGo+v3vcK5YMfaGf/0MbPkV5NbAP64H66l/+UwqSX6393f8dOdPiWfiGDDwnlnv4dNLP02eI++k9vnkrk4+9/AO4mmFqjwnP7t1BbOLPKd8rEKI00PTNJJKEpvJNuYFRFVT2T+wn/Ud63mt/TV29OwYCsQOs4PlRctZW7KWtaVrmZkzEw2NLd1b+FvD33i2+dkRY7yP5jQ7cVvd+Gw+lhYs5aLyi1hVvGpUkFY1lTc73uSB/Q/wctvLaIz98Ww2mCn3lLOsaBkrilawvGg5pe5j/60VQgghxNSTkD4OCemTb+szzbzxaD0Wu4kP/usqvPmOY27b8eUvE/zTn7FWV1Pzl0cx2u2jN0qE4MdrINQOaz8FV33ztB1rd7Sb72/9Pn9r+BsAHouHNaVrsJvsWE1WrCYrNpMNq8lKubuct1W8Db/92FPC7esMcedvN9M2GMdlNXHvB5dy5TwZtynEyeiOdrOzbyd7+/dS6Czk8srLKXQeu94F6C3aG7s2srtvN33xvhFLf7yfhJLAZDDhsXrwWD14rV68Vi82k42dfTsZSAyM2F+Fp4JYOjY0reMRBY4CDAbDiGnHCh2FvLPmnVxRdQUFzgLcFjduixuT0XTC594WbuPPh/7MYHKQUlcpJe4SSl2llLpLKXAUnNQ+hRBCCDF1JKSPQ0L65FNVjUe/s5WuhiBls3O4/jNLMRjH7vauhEI0XHMtmd5e8u78GIWf+9zYOz34LNz/PsAAdzwFVWtP6zFv69nGtzZ8i30D+8bdzmgwsqxwGVdUXcHllZdT7Coetc1ANMUn/rCFNxsGMBjg/759Dp+4pPa0VW8W4lyUVtPs6dvDjt4d7Ojdwc7enXTHukdsY8DA4oLFXFl1JVdUXTHUejyYGOTV9ld5qfUlXm9/nVgmdtLH4TQ7WVWyigtK9XnBK7wVaJrGwcGDvNn5Jus71rOlewtJJQnoF/aurL6Sa2quYXnRcgnPQgghhBiThPRxSEg/M4K9MR78xiYySYUL3zeLxZdXHHPb8Asv0PbJT4HJRPVDD+FYcIyx4X/5JGz/PeRUwT++DrbT25VcURVeaXuFrlgXKSVFUkmSVJKklBTxTJxdfbvY2793xHMW5i/k0opLWV60nPn587GZ9GJ5aUXl3/+6l9+92QzAuxaX8l/vXYTdIl/ghYDhCudvdL7BGx1vsKlr06hwbTKYmJ07m3l58zgcOMyO3h0jHp+fp//Obe/dPmLO7kJHIWtK11DiKiHfkT9i8dl8xNIxwqkw4XSYUDJEKBUimo5Sm1PLkoIlWEyWcY89qSTZ1rONtJJmdcnqcceqCyGEEEKAhPRxSUg/c3a/0s7L9x/AZDFy3acWUzbn2POIt332s4Sfehrb3LnUPPIwBssYX5ITIfjJBRBogaW3wPU/msSjH1t7pJ0Xml/ghZYX2NazbcSYUYvRwoL8BSwpXMKywmUsKVjCE9uDfO3xPWRUjYVlPn5663JKfMfu/i/E2eLQ4CGeaHiCZ5qeIZqOUuoupcxdRpm7bOi21+olmo4OVTePpPWlI9LBm51vjugqDpBjy2Fp4VIWFyxmUcEi5ufNHzFuuzvazfMtz/N88/Ns7dk6IpjPyZ3DJRWXcGnFpdTl1WE0yAwLQgghhJg+JKSPQ0L6maNpGn+7byfNu/sxGg1c/KHZzL+obMxtM/39NFx9DUowSMFn/pn8j3987J02vQ6/vgbQ4IP3w9xrJu8EjqMv3seLLS/yRscbbO3ZOmosqwEDy4uWM9dzMQ/8PYdAxEqBx8b/3rKcZZXHvmAhxFSKZ+JYjBbMRvOox7qj3TzV+BRPNDzBgcEDp/xaVqOVZUXLWFu6lrUla5njnzPhcN0X7+Pl1pdRNIULyy6UwmlCCCGEmNYkpI9DQvqZlUkpvPi7/RzapI8tXXhpORe+dybGMeYRDz7+OB1f+CIGi4WaR/+MbebMsXf67Fdg/Q/AmQ+feBPcBZN5ChOiaRqt4Va29mxle892tvZspTHYOPS4yWDCnKwj0DsfY3wBH1k3l2sXlTC/1Ctj1cWUSatpDg4cHDEO/MjUhDaTDafZidPixGVxYTKY2D+wf6j3iNlo5uKyi7m29loqPZV0RDpoj7QPLR2RDiLpiF48zeoesc6157KscBlLC5diN49RLFIIIYQQ4hwjIX0cEtLPPE3T2PJ0MxseawCgfG4uV925ALvLMmq71o9/nOjLr2CfP5+qB+7HaB1jrGcmCT+9FHr2wJyr9Rb1aRh0OyOdPNX0FE81PsX+gf1D92uqGU1xg2bAbDTisllx2yw4LGbK3GWsLlnNquJV47YqJpUkDYEGAskAiwsWy5zI5wFFVWgKNWEz2ShwFgzVP5goVVNpDjWzt38ve/v3srtvN3v795JQEie0n2WFy7hmxjVcVX0VPpvvhJ4rhBBCCHG+kpA+DgnpU6dhey/P/WovmaSCr9DBNZ9YRG6xa8Q26a4uGq+/ASUYJPfmmyn+yr+OvbOu3fCzS0FJwbt+BMtuOQNncPIaAg082fgkTzU+RUu4ZULP8dl8rCxayaqSVRQ4CjgUOMShQX1pCbcMjcc1G80sL1zOhWUXcmHZhdTmSCX5c4GqqRwcPMjGzo1s6trElu4thNPD83B7rB4KHAUUOArIc+ThsXqwGC1YTBYsRgtWoxWLyUJ/vJ+9/XvZN7CPaDo66nW8Vi+LChaxqGARiwsWM88/D4PBQDQdHVpi6RhxJc5c/1zK3GMPWRFCCCGEEMcmIX0cEtKnVl9bmL/9eCeRgSRWh5m3f3Q+VQvyRmwTfukl2j7+jwCU3ft9vO94x9g7e+1eeP6rYHXDx18Df80kH/2p0zSNxlAjsXSMWCrNluYB1tf3sqVlgGQmjdHeSWlxGwnToeNOI+Wz+XCanXRGO0fcX+IqYV3pOmp8NRS7ivXFWUy+I1+mhzrKgYEDPHLwETqjndT561iYv5D5+fPJd+SfsWNQVIX+RD/d0W66Y8NLc7CZLT1bCCaDI7Z3mB0oqkJKTZ3U69lMNub45zDPP4/5+fNZVLCIam+1FFkTQgghhJhkEtLHISF96sVCKZ7+6S46DwfBAKuurWHFO6tHzKXe893v0v+zn2N0uaj585+wVlWN3pGqwK+vhZb1ULEGbn8CjjN10nQVTyk8vLmV/3hir14JvtzNXdc6ORzazoauDYRTYWbmzGRWzixm5epLgUMfi98caua19td4rf01NnVtOmaAMxvMFDgLMBvNpNU0GTVDWk2TVtKk1TRui5sqbxWV3kqqvdVUeauo8lZR4ak4qe70kVSExmAjDcEGGoINNAWbMBqM+O1+/A4/ubZc/A4/efY8MmqGzmgnHZGOEet4Jk61t5ranFpqc2qZmTOT2pxa8ux5J9VbIK2kea75OR488CDberaNuU2Jq4QF+QuYnzef2pxaZvhmUOouHbOQ2ngUVWFrz1ZebHmR1nArsUyMeDpOLBPTb2fiRFIRFE055j6cZifLipaxqnjV0BAIk8FEKBWiL95Hb7yX3lgvffE+4pk4KSWlv6dqeui2y+Kizl/H/Pz5zPDNOOHzEEIIIYQQp05C+jgkpE8PSlrltUcOsfuVdgCqFuZxxe3zhsapa5kMzbffTnzzFmx1dVQ/+ABG2xhjcAeb4H8ugFQEFt8E198HxrO3VfDNhn7+8fdbGIylKfLa+PmtK1lYPvFxv/FMnE1dm9jcvZnOSCdd0S66Yl30xnrHDYPH47a4KXAWUOgopMBZQIGzgBxbDkklORw803r4DKfCNIWaRk2vdTp5rV5ybDm4LC5cFhduixunxYnb4sZj9eCxevDavPra6sVpdvJq+6v86eCf6E/0A/pFi8sqL2NZ0TL29e9jd99uGoINI6bVO8JitFDlraLGV0O1t5pKbyXl7nLKPeUUOguHWqLTapqNnRt5rvk5/t7691EV/8diMpj0f1tnIUXOIoqcRRS7illSuIR5efOwGM/OC09CCCGEEGKYhPRxSEifXvat7+TlBw6gpFW8BQ7e+Q8LyS93A5Du7qbx3e9BGRgg5wMfoOTrXxt7JweehgdvAk2BtZ+Ct39jWhaSm6jm/igf/c1mDvdEsFuMfO/9S7h6Yckp7TOjZuiL99Ed60bTNCwmC2aDeWj8ssVoIZAM0BRqojnYTEu4Rb8dah7V5fpE5DvymeGbQY2vhhpfDSaDiYHEwNDSH+9nIDGAwWCg1FVKqVtfSlwllLpLsZvsNAQbqA/UczhwmPpAPa3h1jGD9EQVOAp47+z38t7Z76XQWTjisUgqwr6Bfezq28W+/n00BhtpCjWRVJLH3J/VaKXUXUqRq4i9/XsJp4bHjXutXi6puISlhUtxW9w4zA6cFqe+NjtxW93k2fNkGIIQQgghxDlOQvo4JKRPP70tYZ76312E+xOYLUYu+fBc5qwuBiDy2uu03nknaBql99yD77prx97J9gfgL9m51S//Klx01xk6+skRSqT59P3bePlgLwD/dNlMPnHpTOyWMx/mIqnIULfqnngPfbE+euI9BJNB7CY7TotzaKquIyG03F1Oja9mUqp/JzIJ2sJtRNKRoSWaig7/nIoQSoX0JRkinA4TSoao8lbxvjnv4/LKy0+odVrVVDqjnTQEGmgMNtIYaqQ93E5ruJXOaOeoHgp+u5/LKy/niqorWFm8UlrChRBCCCGEhPTxSEifnhKRNM/9cg8te/XuwXPXlbDu3bU4PFZ6f/AD+n78PxicTmr++Ai2GTPG3skb98Ezd+u3r/tvWH77mTn4SZJRVP7fk/v55ev6fOtFXhufvHQmH1hZgc0sLa/TQUbN0BXtoi3SRmekkwpPBUsLl0rLuBBCCHGOG4im6A4l6A0n6YscWVL0hZOU5Nj5/FVzp/oQxTQjIX0cEtKnL1XV2PREI5ufbALA5jSz6roa5l9QQtudHyO2YQO2WTOpeuBBTG7X2Dt5/uvw2vfAYIT3/RrmXX/Gjn+y/GVbO//19H46gvp81iU+O5+8dCbvX1GB1Xz2jr8XQgghhDhbJNIKGxoHePlAL68c6uVwT+SY29aVeHnqny86g0cnzgYS0schIX3666wP8sqDB+hr1f/45ZW5WHdVIcnP3Uamtxf35ZdT/sMfYBirQJymwV//Gbb+BkxWuPmPMONtZ/gMTr9kRuHhzW3c9+JhukJ6WC/LcfDxS2q5al4RhV77FB+hEEIIIcS5ZSCa4i/b2nn5YC9vNvSTzKhDjxkM4HdayXfbyPdk19ml0u/kmkWnVk9InHskpI9DQvrZQVU19r7WwZuP1ZOMZgComWml9MG7sUV6yfuHf6Dws585xpMVeOR22Pe4Pof6rY9D+fIzduyTKZFWeHBjCz9+qZ6e8HAxs1mFbtbV5rG2Np81M/zkOK1TeJRCCCGEEGe3zU0DfOIPW0d83yrx2Xnb7AIunl3ABbX5+JxSd0ZMnIT0cUhIP7skImk2PN7A7lfbQQOLWWPWzt9Q3LOJ0u98B9+114z9xEwS/vA+aHwZbD647TEoXXpmD34SJdIK929o4Y9b2tjXFeLo32KDAeaXerlxWTnvW1GB2ybzYgshhBBCTISmafx6fRPf/Ns+MqrGjAIXN62q5OLZBcwqdGM4i2cQElNLQvo4JKSfnXpbwrz8wAG6G0MAFHVvYk7zo8z8zc9xLFww9pOSEfjDe6HlDbDnwO1PQPHCM3fQZ8hgNMWGxn7W1+vL0WOkPDYz719Zwe3rqqnwO6fwKIUQQgghprdoMsOX/ryLx3d0AHDtohL+88ZFuKTBQ5wGEtLHISH97KUqKluebmbT3xrRVLAlBljY/VeW/eYeLIWFYz8pGYbfvRvaNoEzD257AormndkDP8N6Qgme2dPFr9c3Ud8bBcBogCvqivjIhTWsrvHLVWAhhBDiPJVRVJ7b283Cch/luefWBfxYKsOze7p5YX8PZqOBPJeVPLeNPLeVfLc+brzAY6PAbcNsGlnbqKE3wsd/v4WD3RHMRgN3X13HHRdUy3cmcdpISB+HhPSzX1dDkOd+sZtQfxI0lRmJHVzx409gcTnGfkIiCL+9Hjq2gasAbv8bFMw5swc9BVRV45VDvfzy9SZeyc63DpDjtDC7yMPcYg9zivX1rCIPXruMqxJCCCHOZcF4mk/dv5VXD/WR47Twu4+sZmG5b6oP65SkFZXXDvXxl+3tPLunm3haOe5zjAbId9so8dkp8trJ99h4fHsHkWSGQo+N+25exspq/xk4enE+kZA+Dgnp54ZUIsMrv9rGgR1hAHzGEG/79MVU1B3jD2p8EH7zLujaCe4iPajnzzqDRzy1DveE+dXrTfx5a/sxP7wWV+TwxXfMYV1t/hk+OiGEEEJMtsa+KB/9zSYasr3sQB8W96s7VrJiigKppmmEEhl9jvFwkmA8TTSVIZpUiGXX0WSGlKJiNBgwGw2YjloGoime3t1FfzQ1tM+qPCfvWlyKx26mP5KiL5KiP5rM3k7SG06SUceOP6tq/PzopqUUemTWHHH6SUgfh4T0c8veB1/jtecGSFvcAFTU5bL23TMpqPSM3jg2AL+5Drp3g6cE7ngS/DPO8BFPrURaob43woGuMAe6wuzvCnOwO0xndg52gCvqCvmXd9Yxs9A9hUcqhBBCiNNl/eE+/vEPWwnG05T47Nz7gSV897mDbGwcwGEx8fPbVnDBzNN/kT6cSNM2GKd1IKavB/V1TyhBXyRFbyRJ6qhpzU5WnsvKdYtLuX5JKUsqcsbtoq6qGn3RJN3BJJ3BON2hBJ3BBCU+Ox9cVYnFNMYUv0KcBhLSxyEh/dzT9dBjbPj9VtpLLkAz6oU9Zq4oZPV1M8gpestYq2gf/Ppa6N0H/lr42PPglO5MPeEE9714mN9vaEFRNUxGAzetquQzV8wiz22b6sMTQgghxEn63ZvNfO3xPSiqxpKKHH5663IKPXbiKYX/87vNvHqoD6vZyE8+vIzL5hZNaJ8ZRWVrS4AX9nWzvTVAIqOSTCukFJVURl/iKYVwMjOh/XlsZvI9NnwOC26bGZfNhMtqxpldW81GVE0jo2qo6vDaaDRw8ewCLpyZL+FaTHsS0schIf3cFHr2Werv/hb1FVfRXbgSDAaMRgN1F5ay5voZ2F1HjbcOd8PPr4BgC1RdALf8BcwyrzhAfW+Ebz25n+f3dQPgtpn5yAXVLK/2U1fsocBjkwIqQgghxFkgo6j8+xN7+e0bzQDcsKSUb9+4CLvFNLRNMqPwqfu38dzebiwmA//9waVcvbBkzP0FY2leOtjDi/t7eOlAL8F4ekLHkeu0UJ7rpMLvoDzXSXmug+LsOPCCbCG3o49JiHOVhPRxSEg/d0Vef522T32akNFP05Jb6LVUAOD0WbnsljqqFuQNb9y9F355FSRDsPgmuOHH+gTjAoA36vv55pN72d0eGnG/32VlTpGHuSUeZhV68DrM+pVuqwmXTV+77WYK3BLmhRBCiBMViKV45VAfVX4niytyTno/0WSGT/xhKy9nC8d+/qo5fOKS2jE/m9OKyl0P7+CvOzowGuBDqypRNY1QPEMwnh5a2gNxlKPGcuc4LVw6p5ALZ+bjc1iwmo3YzEas2cVmNlHss+OW6cuEACSkj0tC+rkttnUbrR//OGooRGTR5Ryc/SECfUkA5l1UygU3zsRqz35YHH4e/vB+0BS47Ctw8f+dwiOfflRV4687O3h2Tzf7u0I09kU5Rp2VURaW+bjz4hlcvaB41BQnQgghhBgWTyk8v6+bx7a38/LBXtKK/mH7vuXl3H11HbmuE+vt1xNO8JFfb2J3ewi7xci9H1jKOxYUj/scRdX40p938vDmtnG3m1Pk4bK6Qi6fW8jSylxMRrkgL8RESUgfh4T0c19i/35aPvoxlP5+TDUz6Xzfv7N74yAA3nw7l982j9JZOfrGm34Of/ucfvu9v4IF75magz4LJNIKh7oj7O8Ksb8rTGNflGgyQyylEE1liCX1dSSZ4chflbIcBx+9sIYPrKzA9ZYr6T3hBNtaAmxvDWAzG3n30jKq8lxTcGZCCCHEmZXKqKyv7+Px7R08s6eLaGp45pUZ+S4a+6Noml4Q7SvXzuP6JaUT6qFW3xvhtl9upG0wjt9l5Re3rWBpZe6EjklVNR7e3MrB7gg+hwWfw4zPacnetlCa46DEd4zpboUQxyUhfRwS0s8PqaYmmj/yETIdnZjy8zHd/X1eey1JZCAJBlhyRSWrrqvBYjXB03fDm/eByaZPzVaxcqoP/6zWH0nyuzeb+e0bzQxkp0Tx2s18eE0VeW4b21oG2dYSoD0QH/XctTPy+OCqCq6aXyzj04QQQkyajKLS1B+lO5RkaWUOTuupdcne1RZkT0eQ6nwXc4o8Y7Z+D0ZT/P1ADy/s6+Hlg71EjiqqVp7r4PolpdywpIxZRR62NA/wpT/v4mB3BICLZuXzzRsWUpnnHLXfI7Y0D/DR32wmEEtTlefkN3esojpfLn4LMV1ISB+HhPTzR7qri9aP/yPJ/fsx2O3kf/M/2dFfwf71nQDYnGbq1pWw4KISfC98DA4+Ba4C+NgLkFs1xUd/9kukFf60tY2fv9pIY1901OMGg95tbklFDp3BBK8c6h1qgfc5LLx7aRnXLymlOs9FjtMiY9yFEEKclLSi8kZ9v94TrFOffvRwb2Ro6q9Kv5N73ruI1TPyjrOn0Vr6Y/znM/v5287OEfcXeGzMKfIwu8hDrtPCq4f62Nw8MGLYWL7bxtULi7l+SSnLKnNHfc6lMio/e7WB/37hEKmMit1i5JY1Vcwu8gwVYCvx2TGbjDy9u4t/fnAbyYzK4oocfnHbCvJldhYhphUJ6eOQkH5+USJR2j93F9GXXwGDgcL/+38JLr+W1/94mFBfdm5wA1TV+ViY+BGV0Ucx+Kvhlj+fd3OoTxZF1Xh+XzcPbmzBZDSwtDKXpRU5LCz34bEPV91vD8R5ZHMrj2xuG9XKbjUbKfLaKPLYKfLZKctx8N7l5cwu8pzp0xFCCHEWCcbT3PGrjWxtCYx6zGk1YTUbCcTSGAxwx7oaPn/VHBzW4/fkCsRS/PDFw/z2jSbSiobBACur/XQG47QOjO4pdsTcYg9Xzivi8roiFpX5ME5gTHdjX5S7/7yLNxr6Rz1mNECx105nKIGmweVzC/nhTUtPuWeAEOL0O6tC+n333cc999xDV1cXixcv5oc//CGrVq065vaBQIAvf/nL/PnPf2ZgYICqqiruvfderr766gm9noT084+WydD9//4fg/c/AEDOBz5A4Ze/TOuBILteaqNlz8DQtj5LLwvsf2Wufwf2W38DZcum6rDPW4qq8drhPh7a1MKbDQNDXebfymCA6xaV8s9XzKK2wH2Gj1IIIcR0F4iluOUXG9nVHsRjM3Px7ALmFHuYU+yhrthLea6DaCrDN/+2jwc3tQL6ePB73reY5VVjj+NOZhR+u76ZH754iFBC765+0ax8vvTOOuaV6t8ro8kMh3oiHOwKc6A7THcowcpqP5fXFVKee+zu6uPRNI0ndnayvr6PtsE4bYNx2gfjpBR1aJubVlfy7++aLwVbhZimzpqQ/tBDD3Hrrbfyk5/8hNWrV3PvvffyyCOPcODAAQoLC0dtn0qluOCCCygsLOTuu++mrKyM5uZmcnJyWLx48YReU0L6+UnTNAZ/+1u6v/2foGm4LriAsnu/j8njIdAdY/cr7exb30kqrn/gmg0JZjvXs/DGt5G/7sopPvrzWzKj0BNK0h1K0B1K0hVKsLGxn2f26HO5Gw1ww9Iy/vnyWccsPKdp2oS7y2uaxhsN/eztCHHJnAJmFkprvRBCnG36I0k+/IuN7OsM4XdZ+f1HVw+F6LH8/UAP//KnnXSHkhgNcOdFM7hkTiHN/VGa+mM090dp7IvS3B8jntaLvM0p8nD3NXW8bXbBmTqtEVRVozeSpG0wjtVkZEGZV4aGCTGNnTUhffXq1axcuZIf/ehHAKiqSkVFBZ/+9Kf5l3/5l1Hb/+QnP+Gee+5h//79WCyWUY9PhIT081v4hRdo/7+fR4vHMZeWUPqNb+Batw6AdFLh4MYudr3YQn/ncFe1kuIkC69dxoylBZjk6vS0sbs9yL3PH+L5fXpYNxkN3LisjJp8N92hBD1hPdTrt5OU+uy8e2k571lWRoV/dEtGRlF5ancXP32lgV3twaH7V1bn8qFVlVy9sESK2QkhxFmgJ5zg5p9t4FBPhHy3jfvvXD2h4VHBWJqvP7GHP29tH3e7Iq+Nz105hxuXl8sUZEKICTsrQnoqlcLpdPLHP/6RG264Yej+2267jUAgwGOPPTbqOVdffTV+vx+n08ljjz1GQUEBN910E1/84hcxmcb+8pxMJkkmk0M/h0IhKioqJKSfx+K799D+mc+QbtPnAs35wAco/PznMbn1VlhN0+g80MeuB/5GfXcZGvr/LbPNhL/ERX6Zi7xyN3mlbvLK3NjdJ3fBSJweO1oDfO+5g7x8sHfCz1k7I4/3Li/nnQv1eWMf2dzGz19rGBpHaLcYWVyew+bmQZRslR+v3cx7lpXzwVUVzCnySGuFEEIcRVU1Xtzfw+83NJNMq9y4vJxrF03s4mZvOEnLQJSWgRitA3FaBmK0DMRoH4xjMRnwOix47Ra8DnN2baEi18HiihzmFnuxmocvoHcFE9z0szdp6ItS5LVx/51rTnhI1LN7urjnmQOkFZWqPBfVeU59ne+kOs9Fpd8pXcqFECfsrAjpHR0dlJWVsX79etauXTt0/xe+8AVefvllNmzYMOo5c+fOpampiZtvvplPfOITHD58mE984hP80z/9E1/96lfHfJ2vfe1rfP3rXx91v4T085sajdLz3e8xeP/9AFhKSyn5f9/EtWbNURupRJ74FntebmFv/Epiqn/MfXn8dqoW5lGzKJ+y2bmYLPLBPRW2NA/whzdbwABFXjtFHhtFXjuFXjv5bitbmgf545Y21tcPF95xWU1YskWDAHKdFm5dW82ta/Xp4rqCCR7Z3MqDm1pHFLNzWU1DX9iq8lxU+fX1wnIfbpsU6xFCnD+iyQx/3NLGr15vpKk/NuIxn8PCjcvKuXlN5YigHEtleKO+n1cO9vLKob4xZwCZKKvZyIJSL4srcphf6uMHLxyiZSBGWY6D++9cfcxhUEIIcaadsyF99uzZJBIJGhsbh1rOv/e973HPPffQ2dk5anuQlnQxvuibb9J595dJd3QAkHvThyj83Ocwuo76UN/wv6hPfomgUkT/zE/R738n/e0R+tsjwxXisyx2E5Xz8qhZnE/VgjzsLmlln27aBmM8urWdP25tozn7hbLS7+RjF9XwvuUVY1b1VVWNVw/38eDGFp7f101aGfvPps1s5PK6Qq5dVMplcwule7wQYtpIpBV+90YzTf1Rbl1bzZziU6u30R6I89v1TTywsWWogJrHbuamVZV4HRbu39Ay4uLm2hl5rJ7hZ2PjAJubBkcUPDMaoMTnoNLvpMJ/ZK1PMaZqEIqnCSXShOIZwok0gViagz0RdrQGCMbTo46t0u/k/jtXn3SRNiGEmAxnRUg/me7ub3vb27BYLDz//PND9z311FNcffXVJJNJrFbrcV9XxqSLt1IiUXq+cw+BBx8CwFxSQtEXPo/nHe8Y7tK89bfw+Kf12++8B1b/HwBS8QwdhwM07eyjcWcfseBwJXKD0UDZ7BxqlxUyY0kBTu/x/3+KM0fTNLa2BIilMqydkTfhrovJjELbYFwvJtSnFxNqHohxqDsyqrX9ynlFXLuolLpSL4PRFANvWVKKSmG2xb/Ia6fYa6fQa5NwL4Q4bRRV4y/b2vnuswfoCA5fWH7H/GI+ddlMFpT5Jryvpr4oz+zp4pk9XWxrDXDkG2R1npOPXFjDjcvKcWV7EymqxiuHevnDmy28uL97xPzgAOW5Dt42u4CLZxewrjZvxJScE6VpGk39Mba3DrKjNci21gA2s5H//uASSnyOE96fEEJMprMipINeOG7VqlX88Ic/BPTCcZWVlXzqU58as3Dc3Xffzf33309DQwNGo/6F+r//+7/5z//8TzqyLaHHIyFdHEv0jTfo/NevkG7XC8Y4V62i6Mtfxj5ntr7By/fA378BGOB9v4b5N4x4vqZq9LSEadzRS9POPvrbh7vvGQxQMnM4sLtzbWfmpMQZo2kaeztD/HVHJ0/s7KBt8Njz5B5PrtNCTb6L2gI3tYVufV0w+eMgNU0jkVaJpxXiaYXE0KKSzChYTEYcFhNOqwmH1YTTYsaRnWdYiMnUH0mytzNEXyTJiir/mMUfz7Qjv/MvH+wlGE+TyqjDi6KvCzw2FpT6mF/mZVah54z/rrxysJdvPbWffZ0hAEp9duaVenl+X8/QNpfNLeRTl81kWeXIKccUVSMQS9EeiPP8vh6e3dPF/q7wiG3WzsjjoxfWcNncwnHn++4IxHlwUysNvRFWVOVy8ewCavJdUttDCHFeOWtC+kMPPcRtt93G//7v/7Jq1SruvfdeHn74Yfbv309RURG33norZWVlfOtb3wKgtbWV+fPnc9ttt/HpT3+aQ4cO8ZGPfIR/+qd/4stf/vKEXlNCuhiPmkjQ/4tf0P/Tn6Elk2A0kvuhD1Hw6U9h8vngb3fB5l+CyQq3PArVFx5zX8HeGPVbe6nf2kNP88gvNr4Ch158rsxNfpmbvHI33jw7BqkSe07QNI3trQH+uqOTJ3d10h9Nkuu04ndZyXNbyXVayXNZMZuM9ISPTC+XoCuYIJlRj7lfs9GAz2HBbTfjsppx28z6bZuZPJeVYp+dEp/eIl/s01vnLSYjA9EUfZEk/RF9rS8p+iNJ+qP6ui+Soj+aJJE+9usfS67TwtxiL3NL9LmH55Z4mFXoGXPogBDHMxhN8Xp9H3s7QuzrDLG3M0R3KDlim5mFbi6dU8ClcwpZUe0/Y+H3yO/207u7eGp3Fy0DseM/KctqMjK3xMP8Uh8rq3N5x4JinNbxa1homsYb9f28dLCXXKeVmiN1MPKco56raRrhZIbecJKOQJyfvtLAq4f6AL0b+icvncnt66qxW0wc7A7z478f5vEdHUMt3EsrczAbDUM9fQLxNG/9hmgyGlgzw8875hdz5bxiin32CZ+/EEKc786akA7wox/9iHvuuYeuri6WLFnCD37wA1avXg3AJZdcQnV1Nb/+9a+Htn/jjTf47Gc/y/bt2ykrK+OjH/3ouNXd30pCupiIdHs73f91D+FnngHAlJNDwT//EznveTeGRz8K+58Amw8+8hQUzT/u/kL9cRq29VK/tZeuhuCY21hsJgqrPZTPyaVsjp/Cao9M+XaOmOg87ZqmEYpnaA/EaeiLUN8Tpb43Qn1vhIbe6NDcvCfCYGDUF+2JsJqN2M1G7BYTdosJm9lIRtWIpTLEUgrxlELmrf1Xj2I0QIHHhjnb68loBKPBgDH775BWVDKKRlpRs4uGommU5zqYXehhdpGbWUUeZhd5qMl3EU8pHOwJc7A7zMGuMAe7IxzqCRNLKeQ4LPicVnKdFnKcFnwOa3Zt0R/LLl6HhVyXlUKPDct59rs1GE3hsZunfUXqvx/o4a6HtjMYGznO2GCAmjwXPqeFnW3BoVkXQB9acuGsfC6vK+LyuYXkuU+tp9KRFmT9ApYeWPujSRp69a7enUd1GbeZjbxtdgGVfidWs3F4MRmxmIy0B+LsaguyuyNIODtu+wi3zcx1i0v5wMoKFpf7RvyNCMbT/GlLG7/f0ExD79hF1Qo9NqrynKQUjb5wkt5IktRbLvJZTAZuXVvNpy6dSa5r9JCrxr4o//PSYf68tf2Yv885Tgsrq/1cNb+YK+oKyXHK0C0hhDgZZ1VIP9MkpIsTEX3zTbq/+U2Shw4DYKmoIP/jd+Ib+BmG9g3gKYWPPQe+8gnvMx5O0dceob8tki1AF2WgI4ryli9XZpuJ0pk5lM/JpXxuLvnlbmlpP4+pqkZ3OEEwniaazBBJKkQSGaLJDOFkhr5Iku5ggs5ggq5Qgs5gfKhV3GAAv9NKvttGnnvkOt9tJc81/HOuy4rDYprQ3L+pjEo8pdA6GGNfZ4j9XWH2d4XY1xlmIJo67vMnymQ0jAhlp8qYnQGgLMdBWa6D0hwHZTmOoZoAhR59RoDpGGgTaYVtLQH6o0nmlXipznON2c1Y0zQOdkd4ancnT+/WuynbzEbmlnhZUOplfqmP+aVe5hR7pkUNhIyi8r3nDvLjl+oBqMl3sWZGHvNKvcwr8TK32DM01jkYT/PaoT7+fqCHlw700hcZbmU3GmBFlZ+3zy/iynlFIyp7pzIq3aEE7YE4HYE4XaHEyN4lYT2MD0RTo8ZPH81lNXHp3EKuXljCJXMKjtsaDvr70ToQZ3dHkJ1tQZ7e3TmiEvrcYg8fWFnB/FIff9zSyuM7OoZ+f11WE+9cWEJGUWns1+tgBGKji6Ud4bGZyffYWFqZw2cun01l3vGHBrQNxlh/uB+XzYzfZR1acp2Wafl7IIQ4/ygZlcYdfVjsJsrn5GI6C4faSUgfh4R0caK0TIbBhx6i739+gtKndx201lRTMKcLj68BQ8EcuO1x8BSf9Guoispgd4zOw0Ha9g/SfmCQRHTklzCby0zZ7NxsS3suucVOGc8njknTNH2crKLid57ZwKlpGr2RJN3BJBoaqgaqpqFp+mMaetd9i0lvdTxyG6CpP8rB7ojeWt4T5lB3hEhSb4Esy3Ewq8jN7CIPswr1tc9hIRBPE4ilCMb1qs+BWJrBWIpQPE0wuwSOrGOpY1bnP5rBAHkuKwUeO0VeG0UeO0VDQwn0Yn/luU58juMXu4okM2xrGWRfZ4gcp5WyHP3CQInPftyAnMwo7GgN8kZ9P2809LG1JTCitdRjN7OwzMfCch+Ly3Mo8Nh4cX8Pz+zuomEC01qZjAbmFHlYVpXD0opcllXlUp137L8tqqoRSekXh6JJRV+n9NupjEp5roPaQvcJTUXYFUzwTw9sY2PTAAC3rq3iy9fUYTMf/+KBqmrs6Qjxwv5untvbzZ6O0IjH5xR5cNlMdAQSdIcTJ9SrJMdpwe+yku+y4XdZKfTauGhWARfNyj/lCxuapvFmwwAPb27lyV2dYw5zmVvs4eY1Vbx7admof89ALEVTvz6XuN1sJN9jo8Bto8AjhSeFEOee1v0DvPLAQQLd+sVNm9NMzeJ8apcVUlHnP2sCu4T0cUhIFydLjcUY+MMf6P/5L1CDepd1Wx4UzO/HPTsPw033Q+nS0/JamqrR3xEZCuzthwKkEyO7Ojt9Voqq9f/DqqKhZNShNUBRjZfqhfmUzso5a/54CfFWmqbRHUrisplOqvrzW6mqRl8kSXsgri+DeqtqeyBOTzhJT0jvNjzRlvtir505xR7mFnuYk11ynVa2twb0qaaaB9jbETpmy2y+20qJz4HNbETVNBRNP0ZV01BUjab+6Kg6AQUeG6U+O/u7wuPWMLCajVw8K5+r5hdzeV0RgViKPR0hdncE2dsRYnd7cFS3ctBrDCytzKXIax/q9j0YTTEYSzEYS0/o36bEZ6e2wM3MQr34YVV2Sq3SHPuI8P3ywV4++9B2BqIp3DYz375xIdcuKj3u/o+lbTDGc3u7eXZPNxubBkYdq9VkpDTHTlmugyKvnQK3Te9R4tF7lBy5neu0nrEhEcFYmsd2tPPgxlaa+qO8fV4RH15TxfKqXLkQK4Q4r0UGk7z+x0Mc3qIXu3R4LGAwEA8N99azOszULMqndnkhlXV+TJbp+51XQvo4JKSLU6WEwwz8+jcM/PrXqFG9pcqRn6RwWRLnP/wI5r/7tL+mqqj0tIRpPzBI2/5BOuuDKBMs8GWxm6io81O9MI+qBfkyFZwQx6GoGgPRFD3hRDa4J+gKJunKFvg7svRFJt6lvzzXwaJyH5Gkol8UGIxPuMZAvtvK6hl5rJ2Rx9raPGZkq2KnFZVD3RF2tgXY2R5kZ1uAzkCCNTPyuGpBMZfNLRy3RVvTNDqCCba3BNjWMsjWlkF2t4dGzF99LGajAZdNL17otJpw2cyYjAaa+2Mjup+/lcEARR475bkOcpwWXtjfg6bBvBIvP755GdX5rmM+90QFYileO9yH0WAYGtKQ57KOW4VcCCHE6aUoKvFQimggRTSYJBrILsEk6YSCt8BBbrGL3GInuSUubA4ziqKy84U2Nv6tkUxSwWCABZeUs/q6Gix2M131AQ5v6aV+W8+I6Y9v/voacoqmfvaPY5GQPg4J6eJ0yQwO0v/znzP4+z/oleABd2mCgo+8D/tN39YrZU3Wa6cVuhtCDHRGMRgNmMwGjCYjJrMRo8lAJq3Qtm+Qpt39I642Arj9Nrx5Drx5djz5Drz5drx5Djx5dlw5NvkCK8QEhRJpDnaFOdAd5kBXmP1d+jqUSFNX7GVldS4rqv2sqM4dNWfzkeEI+vjoBBlFxWg0YDIYhorsmYwGir12Zha6z1iLajKjsK8zzNbmQYLx9IjZCHKz45R9Dsu4XaoDsRT1vREO9+hLfW+U1oEYbce4MPHhNZX86zXzpJu2EEKchdJJhcEuvb5SeCBBJJAkFkgSDaaIBpLEwik4gbTp9Foxmg1EBvTv1sUzvFz8oTkUVHhGbaupGp0NQeq39BDoiXHdp5ecprOaHBLSxyEhXZxu6e5u+u67j8Af/4jep1XDuyifgm//EuuM2VN6bJqq0dsapmlnH827+0dNBfdWRpMBj9+ONz8b4PPsuHPt2F0W7C4LNpcZu8uC1WGWMC/EGDRNI61oMnf8GDRN76HQOhinbTBG+2CceaVeLppVMNWHJoQQYgIigwk6Dwf1wscdUQY6IoT6E8cN4UajAafPiivHhstnw+Wz4sq1YbaYCPbGGeyKMtgVIxoY7olld1tY955a5q4pOWcKJ0tIH4eEdDFZko2N9H39c4Te3KffYQTfNe8g7x8/jW3GjKk9uKx4JEWwJ06oL06oL0G4P06oP0GoL05kIIk60QraBrA7LfgKHeQUOfWl0ElusRNfgQOzzI8thBBCCHHW0jSNYE+cjsMBOg4F6DwcINSXGHNbh8eCv9SFL9+hB/Eji09fO9yWCQXtVDzDYHeMWDBJycwc7K5Tr0UznUhIH4eEdDHZ4s/dT++3v060PduSZjDgueJy8j72MRyLF0/twY1DVVQigSTh/gShvgSh/jjhPr3bUjKWJhnNkIilRxWwG8UAeaUuimf4KK71UTzDh6/AIQWQhBBCCCGmMU3V6KwPcHBTD43be4m9ZcikwWigoMJNQZUXf4mLvFIX/lIXDo/UO5oICenjkJAuzoiBRuL3XEffxgiR9uGxqM6VK8n7P3fiuvDCsza0KopKMpohFkoR6I4NLz36OhnLjHqOw2OheIaPgkoP/hL9D7q3wIFJ5t8VQgghhJgymqbR1xrh4KZuDm/uJjI43OXcZDZSVOOlZKaP0lk5FM/wYbVPfIpNMZKE9HFISBdnTLANfnMdycYW+usLCdabQNFboW2zZ+O//Xa8116D0XruXH3UNI1YKEV3Q4jOhiBd9UF6WkKomdF/ZowmAzlFTvylLnIKnXj8dtx+G+5cO+5cm3wICCGEEEKcIk3TaDswSF9LhHRKIZNUSCeVodv9HdGh+ccBrHYTM5YWMGtlEaWzcjBLUc/TRkL6OCSkizMq3AW/eRf0HSBNEQPqDQw+8QJaTP9jaCrIx3/zzeR84AOYc3On+GAnh5JW6W0N01kfZKA9wkBnlIGuGJnk+N3mbU4zbr8dT64Nt18P7h6/fSjEOzxWzFbjWdsjQQghhBBisqiqRsO2XrY83URfa2TcbU0WI9UL85m9sojKBX4J5pNEQvo4JKSLMy7SC7+9Hnr2gDMP5d1/IPDyHgZ+93sy3d0AGOx2fO++Af+HP4yttnaKD3jyaapGeCChB/bOKKG+BJGBhD51x2CSVHx0l/mxmMxG7C4zdrdlqAK9w2vF5bMdVUXUitM78aIlQgghhBBnKyWjcmBDF9uebRlqITdb9RBuc5ox20xYbCYsVn3t8FipnOfH6pAejJNNQvo4JKSLKREbgN/dAJ07wJ4DH/4TWuEiQk8/Tf+vfk1y376hTe2LFuG74Xp8V1+NKSdnqo54SqXiGcKDCSIDSSKDw+F9KMgHkmN2oR+XQW+dtzst2FwWPdy7sred5ux9w1PNOdwWXD6bVKoXQgghxLQS6osTD6dJJTOkEwqphL6OhVPsX985NK7c5jSz8NJyFl1ajsN97gyvPFtJSB+HhHQxZeIB+P2N0L4ZjBZ4+zdg9T+gAbENGxn47W+JvPzy0Lh1g8WC+5JL8L37BtwXXYTBcm5NQ3EqNE0jnVRIRNMkIml9nb0dC6aIhlLEgkmigRSxUJJ4OH3Sr2VzmkdMJ+LOsVFcqxdQsUiAF0IIIcQZEA+nOLixm31vdNLfNn73dafPypIrKpl/UanU+JlGJKSPQ0K6mFLJMDz6cdj/hP7z3Gvh+h+BQx+PnunrI/jEEwT/8hjJ/fuHnmby+/Fecw2+G67HPm+ejMM+QUpGJREdnkYueSTURzP67VjmqPv07eKRFJmUesx9msxGSmb6qJjnp3Ken7wyt7wvQgghhDgmVdWIDCSys+LECfbos+NkUio5hQ5yS1x6Ud0SFx6/HU3TaN4zwP71nTTt6kNV9NhmNBlw+qxY7WYsNhNWhxmrzYTFbqJ4ho85a4plXPk0JCF9HBLSxZTTNNj4U3j2X0FJQU4lvPfXUL58xGaJ/fsJPvoXgk88gdLfP3S/bdYsfDdcj/fa67AUFZ7hgz9/aJpGKqEQHUwSDSSJBPR1sC9O276BEVOUADi9Vspm51AyM4eSmT78pW6MMgZeCCGEOK+FBxIc3tzD4S3d9LVHJjxcz2QxYrYYR0xtW1DpoW5dCbNWFmF3SQ/Ls42E9HFISBfTRsc2eOR2GGwCoxmu+Dqs/SS8pTVWy2SIvPYawcceI/LCi2iplP6A0Yhr3Tp811+P54rLMToco15CTA5N0wh0x2jZO0DrvgHaDwyOanW3Zq9ml8z04fTZSI5orc+QjKVRFU3vQp87PPWcO1fvVm93WjBZZB55IYQQ4mwTDSY5vKWHw5t76GoIjnjMZDbiLXCQU+jAV+gkp9CB2Woi0B1jsCvGYFeUQE9sKMw7PBZmryqmbl0JeWXuqTgdcZpISB+HhHQxrSSC8Pg/wd6/6D/Pfifc8GNw+sfcXAkGCT39DMG//IX4tm1D9xtdLjzvuArf9dfjXLECg1HC3ZmkpFW6GoJ0HA7QWR+kqyFIOjH+FHMTYTQbsDnMWO1mvSubw0x+hZuy2bmUzsrBJpVYhRBCiJOiqRqBnhg9zWF6m8Nomkb1wnxK5+RgMo39PUpVNdr2D3DgzS5a9g6gaRpmsxGTxYgpuwboa4vAkYRlgNKZOcxaUUjl/Dzcfvtxe9qpikqoL0EimqagynPM4xFnFwnp45CQLqYdTYPNv4Cn7wYlCZ5SuPHnUH3BuE9LNTURfPyvBB97jHR7+9D9lrIyfNe/C88VV2Crq5Nx0lNAVTX62yN0Hg7SVR8gGc8MV5LPVpa3OS0YjBAdTBEJZKvXD+rV7GOh1PCH+zEYDHq3t9LZuZTNzsGda0dVVFRFyy767WQ8QyyYInakmF52nUmp+Aod5Ba7yC12ZhcXdrdF/s8IIYQ456TiGdr2D9LVEKSnJURvc5jUGBfUbU4z1YvymbGkgMp5fsxWE/3tEQ682cXBjV1Eg6kJvV5RjZdZK4qYubwQV47tdJ+OOAtJSB+HhHQxbXXuhD9+BPoPgcEIb/siXPx5MI5f+ENTVeJbthB47DHCTz2NGo0OPWYuKMB18UW4L7oY1wXrMHk8k30W4jTQVI1UUiEVzwwtyXiGeDhNV0OQ9gODBHvjk/LaVrsJg8mApuoXGzRVQ1M0NE3DV+ikeIaX4hk+imt9+ItdMve8EEKIaSvQHaNpVx9Nu/rpPBwYKrx2hNliJL/CQ2G1h0xKpXFH74gZYcw2Ex6/ncHO4e9WNpeZ2SuKmLWyCJvLgpJRUdLZJaMveWVuvPkyDFGMJCF9HBLSxbSWjMBTX4Dtf9B/rroQ3vNT8JVN6OlqPE74xRcJPfkU0TfeQIvFhh80m3EuXYpzzWqcy5bhWLQIo8s1CSchzoTIYIL2gwHaDw7ScShAKp7BaDJiNBkwmfW10WTAYjPh9Npw+aw4fdah2yazkUDPkfFvMQLdUUL9ieO24B/N6jBTXOPFk2cfqqUwFNkNvOVnwzHuP7IyYDCCr8BBXrmHvDKXTBsjhBCCaCBJPJImr3RiF4b72yPsW99J086+URe0fYUOyuf6KazyUFjlxV/ixHhUV3JV1eiqD1C/rZeGbb1DRWKNJgPVC/OZs6aYqgV5mMzS/VycOAnp45CQLs4KOx6Cv90FqQg4/Po0bXOvOaFdqKkU8c2bibz8CpFXXyXV0DByA5MJ+5w5OJYtw7lsKY4VK7AUSrX481kmpRAeSKBpYDQaMBj14Gw0GtA0/YtPV4M+5r67MTTuFHWng7fAQX65m/xyN1aHGSWjomY0FGV4bbGZyC3Su+r7Ch0S7IUQ4iyXSSl0HA7QuneAlr0DDHTordgev51Zq4qYvaqIvFL3qOfUb+1h9ysdIwq1GU0GSmflUL0wn6oFeeQUOSd8HJqm0dsSJtgbp3xuLg639fScoDhvSUgfh4R0cdbor4c/3gGdO/Sfqy+Cy74ClatPanep1lair71GbPMWYtu2kunoHLWNbdZMXOsuwHXBOpwrVmB0TvzDTJxfVEWlvz1KV0OQRFTvGjj0aZK9MfTh8pZPmaGPnZErlIzKYGeM/rbwhMf8vZU710ZOkZO8cjczlxVSVOOVMfZCCDGNKRmV3pYwHYcCtB3Qe4cp6aMuAhv0bulHXxjOr3Aze2UxJTN9HN7aw/43OklG9anKjEYDNUvymbWyiIo6v1y8FdOGhPRxSEgXZ5VMEl78Bmz4iT6nOsCst8OlX4bSJae063RXF/GtW4lt3UZs6xaS+/YflbLAYLHgWLYM17p1uNatwz6vDoNp/PHxQpwu8XCKvrYIfW0RBtojZDKqXjnXZMBoNmIyGTGaDSSjaQa7YwS6YyPGER7hK3QwZ3Uxs1cV4ysYPT4wEU0z0BEh0B3HV+igeIZPujEKIcQkSsUzdDUE6awP0nEoQHdTaGQoB1w5Nirn+amY56dirh+z1Ujjzj4ObuymZU//qLHloLe0z7uolLp1Jbh8UqhNTD8S0schIV2clQKt8Mp/wbY/gJatRFr3Lj2sF849LS+RGRwk9uabRNevJ/L666Na2k0+H861a3GtW4tr3QVYyyc2Tl6IMyURTWfnmY3SdmCQhm29I1peSmp9VC/OJxZKMdARpb89QuwtLfYWm4myOblDXw59BQ4MBgPJeIa+1jC9LcNLLJzSLxyMWAxYHWb8JS7yK9zkl3vwl7hkznshxLSlKio9LWHa9g3SfnAQg9FA8QwfJbU+imq8J90SraoawZ4Y/e3639sjS6gvMWpbu9tCSa2P0lk5VM7LI7fEecxeUIlImsNbezi4sYue5jAVdX7mX1RK5fy8405tJsRUkpA+Dgnp4qzWXw8vfRt2PQJoehX45XfAZf96zLnVT4amaaSamoi+vp7o+vXENmwYUTUewFJViefSy/BccTmOpUullV1MO6lEhsbtvRzY0EXb/kGO9WnnybPjK3DQ3x4Z1RrvzbdjMBhOqZq+0Wggt8SJv9SNyWzQp8jLVs1XVX0xGAwYjYahgn/6YkRTNTIphXRKJZNSyKRU0ikFTdUwmY2YLfq8vEfWJosRq92sLw5T9rYJq8OMK8eGx2+XafaEOE+kkwqDXVE0Vf+6AGAw6LVGVEWjuzFE674B2g/qxUfHYjBAXrmbkhk+8is9OL1WnF4rDo8Vh9uC2ap/9sdCqRFBvL89ykBndFQL+RGePDulM3MomemjZGYOucXHDuVCnCskpI9DQro4J3Tvhb9/E/Y/of/syNWD+vI7jjtl28nQMhniO3cRXa+H9viOHaAMzy1q8vtxX3oJniuuwLV2LUa7/bQfgxCnIhpIcnBTN52HA3j8dvylLvLK3PhLXFgdeiuRpmr0tUVo2dtP694BOuuDI7pUevx2Cio9FFS6ya/w4M1zoChqtqCdipLWUDIq8Uia/vYIfW1h+lojJGNjf/mdKmaLEbffjifPjsdvx+m1YnOasbss2FwW7E4zNpcFd65NxnIKcZZQ0ir9HRF6mkJ0N4fpaQox2Bk95sXJt7I6zJTPyaV8bi4AnfVBuuqDhAdGt3ofzWIzYTIbh2qTvJXZasRf4iKv3E1e2ZHFJUXYxHlJQvo4JKSLc0rTa/DkF6Bnj/5z0UK4+r+gat2kvqwSiejd4l94gfBLL6MGhyupGpxOHPPnY59Xh62uDnvdPGy1MzCY5cu+OLukEhk664MYDQYKKj3Y3ZYT3oemaUQGk/S1RYbm2TWa9Mr5R1rODUYDaHqXU0XR9JZ2RUVVNIwmA2aLCbPViMVmwmzVbxuNBpSMRiatoKRVMtk5ejMplVQik10U0nH9djKWIRJIjurePx6j2UBlnZ/aZYVUL8rH7jrx8z8dNE0jGc0QDSZJpxTyytxYrNJzR5x9VEWltyVC+8FB2g4M0tMUwum1Zi/+eSio8JBf4cbm1H/X0imFYM+RaTL1JRZKDf3OZ1KKvk6rJKPpMcdpOzwWzBbTUMFOTdWGgntuiYuKulzK5/opqPSM2VU8MpjMjh8PMNgVIx5OEQ+niYdTI1/PAL58hx7GS7OhvNSNt8AhXdCFyJKQPg4J6eKco2Rg8y/h79+ARDYsL3gvXPVN8BRP+str6TSxLVsIP/c84RdeINPVNWobg9WKbfZsrNXVmAsLMRcWYCkszN4uxFxcjNEqV9WFmGxKWiU8mCAykCA8kCA8kCQRTpGIZUhG0ySi6aHbR/cAMBoNlM/NpXZZIaWzcoiHU0QGk4QHEkPrWCiF0QhGU7brvcmAyWzEaDbicFtw+qy4cmy4vDacOVZcPhuaqhENJokGU8SCSaKB7DqYIhrULypEQ0nUjDbiWPIr3JTU5lBcq4+bdeVIkSgxtZLxDPFQikw6OywlqWSHqijEginaDwboODhIKqEcd1/efDuqog3N0T1RNpeZoiovhdXeoXnAJ+t3Q9M0UvEM8XCadEohp9CJxSYXz4QYj4T0cUhIF+esaD+8+O+w5TeABjYfvP0/YNmt+qCyM0DTNJIHD5HYs4fEvn0k9u0luW//qPHso5hM2GbOxF5Xh33ePOzz52GbMxeT23VGjlsIMdpAR5T6bT3Ub+2hv/04v8NngN1lwWAyEA+N7g3g8dvxFTpw59hw5dqyazvuHBvefPtQy+TJ0jSNTFrVWyGzLZFDa03D4bZgNJ1acUBN0/d9qvuZbJqqER5MEOyO48yx4i9xnbGxxKlEhshAkshgAk1jqN6CxT5cf+FE//2UjEqoL663VPfECfbECPTECPbGMRoN2N1W7C4LDrcFu8uC3W1BUVSig/pxRAaTRAJJ0hMI36B3Ky+bnUPZnFxKan3EQqlsUcoIvS3hUd3LbS4zuUVOcrKLO9eO2WrEbDEN1aMwW41YHWY8fruM6xZiGpOQPg4J6eKc17EdnvgMdGzTf66+CK77b8irnZLD0VSVdGsriX37SXd0kOnpGV56e0n39KDFxyjKZTBgra7GuWI5zhUrcK5YgaVMKsoLMRUGu6LUb+ulfmsPA51RXD69CJ0716aPb8+14cxOeTQ0Rl/RUDN6V9x4OJ1tIc+2kgeSQy31Do8Fp9eGK8eK02fD5dVb3J0+vbXd6bPi8towWYxomkZ4IEFXvT59U2d9kIH2yHHH3drdFnIKneQUOvAVOckpdGJzmEmnFNLJ4SWTUkjGMySy3XnjkfRQ914lM3YBLNBb9915dnz5drz5DrwFDnz5DowmfWaAVFwfcpCMZ0hl18lY9v6j7tNUDYvNhM2l1wiwuyzYnHpAzCt3U1jlIa/MPe40gaqqERlIYDAacOfY9OEU42wb6IrR2xIi0BMfMbziSAA0GA2E+uIMdsYY6Iwy2B0jkxwOpO5cG5Xz86ian0f53NyhGg9HZNIK4f4Eob4E8XAK41E9LExmAyaT/hqpxBj/TrE0kUBqKAwfq7jZ0Y7MsmA8espGsxGDARRFQ0nrdSSOrMfqIn6yLDZTdljKyOEpVruZ4hk+yufmkl8xdrfyIxKRNH1tYUxmIznFThm7LcQ5REL6OCSki/OCktHnVn/xG5CJg9kOl94Naz4Jpuk1NlzTNDJdXXrL+569JPbuJbFv35jd5s2lJXpgX74C+9w5WGtrMbndU3DUQohTlUkrGAyGU56XPhXP0NsS1oNcIKm3cAaSQy2cY7W8n81MZiP5FW4Kq7wUVHpQ0spQC3CwN06wLz40PMBsMeIrdOAr1C9M+AodmEwGelsi9LSE6G2NjAjcE2U0GfDmOwgPJEZU7zYaDRTX+vDm2Qn1Jwj1xYkEknAav2nanPpMBSazkVR8uP7CsaqIT4TFZsJX6NAv5BTp/06+AicAiYh+sSYRTZOI6IvBZMCTa8OVY8ft13tuuHPt0t1bCDEuCenjkJAuzisDjfDXf4bGl/WfS5bANd+F8hVTelgTkRkYIL59B7Etm4lt3kxi954RFeWPMBcXY6utxTazFuvMmdhnzcI2axZGl3SVF0LoXaSDPXG9C3NPjEC3fjuTUoZaO4daQG16t2mH26JPMeWx4HDra7vbohf6MxgwGMBg1G8fGVcf6osT7E1k13FCfXE0TQ+VNocZa3Ztc+pT5NmcZqwOMzanZeh+k9lIMp4mEcmQiKWH6gREgyn6WsJ0N4UmNFuA0WwAVW8pPx6z1UhBpQd/iQtNI1uULFuQLKWiKiruXDv+Ehf+Ehe5JU68BQ5MJiOZlEL7wQAte/pp3tNPsGfsqQotNhPeAgcurxVV1WdBUDJ6gUQlow33IBjj38rps+k9NnLt4844oGRU0gmFVDKDqmjZHh0aiqKv1ey0hSazYXjaQrPebdzmMks3cSHEpJOQPg4J6eK8o2mw/Q/wzN3DheVqLoYLPwszLj1j49VPlRqNEt+xg9jmzcS3byd56DCZ3t5jbm+prMQ+Zza2WbOxzZmDfV4dlvJy+SImhDhraZpGqC9Od1OInqYwfW1hLDbzUCuwr8Chj83PtYOmEepP6BcoumNDY60zaZWCCg8FVR4KK73kFDtPW/XtYG+Mlj0DJGMZvAV2fPlOvAV2vZ6A/O0VQpznJKSPQ0K6OG+Fu+GFr8POh0DNtsSULNHDet11kzK/+mRTgkGS9Q0k6w+TOnyY5KHDJA8dOmZ4N/n9OBYuxL54EY5Fi3EsWohJ/g4IIYQQQohJJiF9HBLSxXkv0Apv/EivAp/Jdk3MmwkXfQ4WfRCM07uy8ERkBgZIHjxI8sABEgcPktyvr0mnR21rra3FuWolrtVrcK5ehTk3dwqOWAghhBBCnMskpI9DQroQWdE+vbjcxp8Od4MvWQLv/E+oXDOlhzYZ1FSK5L59xHfsJL5TX9ItLaO2s82Zg3P1KpwrV2KfMwdLWRkG09nXy0AIIYQQQkwfEtLHISFdiLdIhGDTz+HV70EqrN+34L1w5dfBVz61xzbJMoODxLduJfrmBmIbNpA8eHDUNgabDWtNDbbaWqy1M7DNqMVSVoq5qAhzXp4EeCGEEEIIcVwS0schIV2IY4j0wAv/Dtt+D2hgdujj1dd9GqzOqT66MyIzMEBs40aiGzYQ37adVGMjWjJ57CeYzZgLCrAUFWEuKsJSWoqlrAxLeRnW8nIspaUYnefHv50QQgghhDg2CenjkJAuxHF0bIenvwQt6/WfvWVw0V2w9BYw26b00M40TVFId3SQPHyYVH09yfoGUg0NpLu7yfT0gHr8eXlNeXlYKyv1KeJqa7HVzsRWOwNzSYlUOxZCCCGEOE9ISB+HhHQhJkDTYM+j8Ny/QbBVv89bpresL70FLPapPb5pQMtkyPT3k+nq0kN7Vxfp9g5S7W2k2ztIt7WhhsPHfL7R6cRSXYW5oABzfj7mvHx9XZCPKS8Pc24uppwcTDk5GCyWM3hmQgghhBDidJOQPg4J6UKcgHQCtv4WXvs+hDv0+zwlelhfdpuE9eNQQiHSbW2kmppIHq4nWV+vTxfX3DJmpfljMbrdemD3+7HX1enV6FetwlxQMIlHL4QQQoipomkasU2bcK5cKT3vzhES0schIV2Ik5BOwLbf6WE91K7f5y6CpR+GxTdB/sypPb6zjJZOk2ptJdXSgtLfT6a3j0zfkaUXpa8fJRBACQb1Xg3HMDx93GqsNTMwulwYXU5Mbre0vgshhBDThJpKEXnxRQxmM+5LLsFgNo+7ffLwYbq/9W2ir79O2fe+i/fqq8/QkYrJJCF9HBLShTgFmaReWO7V70Gobfj+itWw+EMw/93gyJmywzvXaIqCEgrpgX0wQKanh/i2bUQ3biS5f/+4Ad5gtWJ0uzHn52Ovq8M+fx72ujpsdXWY3O4zeBZCCCHEuSPd3U3oiSeIbdyEff48PFddhW327DFbuzMDAww++CCD9z+A0tcHgKWykrw7P0bO9ddjsFpHbK8EAvT+6D4GH3gAFAWDxULBZz9L3kfuOCPnJiaXhPRxSEgX4jTIpGD/E7DjATj8PGjZAmpmO8y9BpbcDDMuAaNMTzZZlECA2ObNRDduJLZpM5meHtRoFC2ROO5zLVWV2GpnYrBZMZjM+jRyZpN+22LB5M/NjpUv0NeFBZj9fmmdF0IIcV5So1HCzz9P8LHHib7xxqiL5NbqajxXXYX3qrdjq6sjVV/PwG9+S/Dxx4dmiTEXF6MlEiiBwNDPeR/5CDnvey8Gi4XBBx+i74c/1HvRAe4rLqfoC1/AWll5Rs9VTB4J6eOQkC7EaRbugp0Pw/b7oXff8P3eMr11fclNkFc7dcd3ntEyGdRoFDUaRYlESLe1k9i3l8TefST27iXT2XlyOzYYsJSVYZs1a3iZPQtrTQ3Gt7QECCGEEGcDLZ0mvns3sY2bSHd26BerzSY46gJ2ur2d8PMvoMViQ89zLFuG+5JLiO/YQfTVV9FSqaHHzAUFZHp7h362L1iA/47b8b797WjpNIMPP8zAL3+lzxIDmPx+TDk5pBoaALDNmkXR3V/CtXbtGfpXEGeKhPRxSEgXYpJoGnRs08P6rkcgERh+rOoCvXV93rvA5pmyQxSQGRwksXcv6dZWtIwCSgYto6Ap+m01mUTpHyDT26uPkc+uUZSxd2gyYa2owDpjBtaaamwzZmCtqcFaU4M5N/fMnpwQQohzWqa3l9jWbcS3biG2ZSup1lbcF19M7oc+hGPpkuMWWNMyGRJ79xLdsIHYho3Etm4dEb7HY6msxHf9u/C9611YKyqG7lciUSIvv0T4mWeJvPKK3qPNYMBzxeX4b78dx7Jlo45LTaUI/vlR+n/2M9Lteq0fU04OBf/8T+S8733HHbMuzk4S0schIV2IMyCdgANP6uPX618Esn9mjBaoXAOzroSZV0JhHUjF0mlPU1WU/n6SjY0kDx3KLodJHjqEGgod83lGl0uvSu/zYcrxYfT5MPl8mP1+rDUz9Lnja2ow2mWWACGEOJdoqRRKOIzR4zlubys1FiPV2ka6rRUlEEBLpdDS6aG1mkqR6eomtnUL6eaWY+7HVldH7oc+iO/aazE6nfpxqCrJQ4eIvfkm0Tc3ENu0CTUSGfE8k8+Hc9UqbLNmoWkqKCqakoHsBWyj3Yb78stxLDn+RQA1FiO2bRvWysoRQf6Y/07pNKGnniLd0UHuhz6Eyec77nPE2UtC+jgkpAtxhgXb9bHrOx6A/sMjH/OWwczLYfY7YeYVYJZu02cTTdPI9PSQamgg2dBAqrGJVGMjycYGMh0T7FZvNGKpKMdWOxNrTTUGiwUtnYZMBi2d0b+oZTKY8/zZFnq9xV5a6YUQYmppmkb01VcJP/ccmb5+lIEBMoODKIODqOHw0HZGnw9zXh7m/HzM+fmY8vJQQ0FSLa2k2lpRevsm/qIGA7bZs3EuX4Zj2XLMhQUEH3uM0BN/Gxr7bXS78V57DUowSGzDRpSBgRG7MHq9OFeuxLV6lR7OZ8/GYDSeln8TIcYjIX0cEtKFmEL99XqhuUPPQdOrkDmqyJkjV68Ov+iDULFKWtjPcmosRrqrGzUURAlml4C+zvT26nPGHz6Mmi2Qc6JMOTlYZ8zAUl6GOS8fc54fkz8Pc36evi7Ix5yXJ10GhRDiNNM0jcjfX6Lvxz8msXv3admn0efDWlGh/922WjFYLMNriwWjz4tzyRIcS5aM2dqsBAIEHv0Lgw8+MKq13eBw4Fy+HNfaNThXr8FeN1cfby7EGSYhfRwS0oWYJtJxaHodDj0Lex+DSNfwY7nVsOgDsOC9kD9LAvs5StM0vRv94cMkD9eTam4GTdO/lJnN2S9nZjCZh1vsmxpPqJXeXFCAubgIS2ER5uJizAUFGG1WOPIaZsvQl0BLeTm22hkYHY7JPXEhhJhm1FiMxL59JPbsAQxYZ8zANqMGc3HxUCuzpqqEX3iBvh//D8l9eqFYg8NBzo03Yps1S58ZxO/HlJurL14vaiSSrW/Sp6/7elH6+zG6PVgrK7BUVGKtKD9t3bw1VSW6/g3CzzyDuagI15rVOBYtGjXVmRBTQUL6OCSkCzENqQo0vqxXid/7OKSjw495y6D6Iqi+EGougpwqCe3nOTUWI9XcTKqxkXRnF5n+fpT+fjLZ5cjtYxa7G4/BgLWyEtvsWdhmzcY2exbmvLwRjx9ZG0wmDE4nxiOLw4HBbj/umEUhhDgZyfp6lEAA+/z5p1TLQ8tkSOw/QGLXTuK7dpPYvZvk4cOgqqO2NTgcelHQmhn6BdUDBwAwOp3k3nwz/jtux+z3n/SxCHE+kZA+DgnpQkxzqSjsfxJ2PggNL4OaHvm4r0IP7Utv1qvGSyASY9AURQ/t3d2ku7rIdPeQ6e4i09unFyPKZLKLPv5dTSRJNTWNGrt4wgwGjG43luIizCUlWIpLsJQUD98uL8dSXCTd8IUQE6IEAgSf+BvBRx/NtnIDFguO+fNxrliOY9lynMuWYsrJOeY+1FSKxK5dxDZtJrZ5M/GtW1HHqGhuLizEvmABBpOJZGMDqeYWSI/8DDa6XOTe8mH8t90mtUGEOEES0schIV2Is0gqBq0boOk1fQx7+xZQM8OPlyyBtZ+C+TeAyTJVRynOIZm+PpIHD5I4eHCoiv1QAaS3fFxq6TRqPI4ai+lT7kyU2YylpARLeRnW8gos5eVDlYiHW+r1lcnj0bctLcVcNDrca5qGMjBAuqODdHsHSiCAtaoS28yZmPLzpVVfiEmiqSqZzk69V09zM6mmZjIDAxiMBjCawGTEYDCCyYjRZsc2ezb2hQuw1dYedzy0lskQff11An9+lMiLL+rFNAEsFkw+H0rf6EJrlspKjHa7/jfCkh3KYzajpVIk9u4dKqp2hNHjwbFoEfaFC3AsXIh9wUIsRYWjjiPd1kayoYFkfT1GqxXfDTdIBXIhTpKE9HFISBfiLJaK6qF972Ow48HhwnOeUlj9D7D8dnDkTOURivOUpiio8QRaPIYSCund8Ls6SXd0ku7qIt3ZQaajk3RHx/AX7hNlNGIuKsJSWorRbifdmd3fMS4QmHw+bLNmYZ01E9uMWgwWM2oigZZMoSWTqEn9NpoGRuNwuDAaMBiNaBkFNRpFjcWGl2gULZ3GaLdjdDowOPRu/kaHA6PbjW3mTOwL5utBRHoLiCmW7ukh8tJLRF78O6nGRuwLFuBctQrnqpVYq6vHvIilqSrpjk69Z01/34iil0eWdGcH6ZZWtFTqhI/J4HRin1eHY8FC7PPqUJNJMp1dpLu7yHR16+uOzhEt3bZ5deTc8G68112LKSeHdFsbsc1biG3ZTHzzFlJNTcd9XZPfj3PFCn1ZuUKvaC7F04Q4oySkj0NCuhDniGg/bP4lbPwpRHv0+ywu8M/ItmJY9HnZj9z2z9DnZq+5CKyuqT12cd7SVJVMTw/p1lZSbe2k29pIt7WhpVMMfRwf+VTWND0QdHSQ6ewcN9ybCwv18O7zkm5qJtXSMqrl/0wy2GzY5s7BMX8+9vnzca5eg7W8bMqOR0xPmqKQ2L+f5KFDmAsKsFZVYykpPm54VJNJtHRaD9lGo36RyWAAg4FkQyORv79I+MW/k9i585j7MBcU4Fy1CsfixSiBQZKNjaQaGkk1NY1qdT4miwVrRYU+J3ZVFeZCvSVaUxV9ru3sWo2ESezZS3zvXrQxupmPxZSbi+9d1+F797uxz5077raZ/n5SDQ1DU1Zq6bQ+hWVG73lmn1eHtaZGetYIMcUkpI9DQroQ55hMEnb9Ed64D3r2HH97k00vQjfrSpj1dsirnfxjFOIUaaqqV0fu7CDd0YEaj+vd4MvKMJeUYHxL5WI1kdCr4R8+TPLQoaGWNoPNjsFmxWizYbDaMNhsYDSAqoGqoKlq9rYKJhNGlxOj0zVibbBY0RLxbFf/OGo8hhaPkxkcJLlvP4m9e1Gj0VHnYJszB/dll+K57DLs8+dPu3mJNU0j09mJEgxirazE6JKLeaeblsmQ2LeP2MZNxDZuJLZ164j5tAEMViuWigqsVVVYKypQU0mU/gG9KGRfH5n+ftRIZMKvaV+0CM9ll2KfN4/4zl3ENm4kvn37uK3gBosFS1UllsIiTDk5mHJ8GH0+TD4fJl+OfkGhugpLSckJtUZrikKqoSFbrG0XiYMHMbpcer2K4iLMRcV6DYuiIqwVFRgsMoxLiHOJhPRxSEgX4hyladC5HWID+rh1NQNKWl9nktCxFQ4+C8GR86fiKYX8mXpL+9BSC/4asMhUXEKcKE1VSTU3k9izl8SePcR37iS+bduIytHmwkLcl16KrbYWTVFAOVLIT0HLpDEYjXrFfJdLX47cdnswFxZizs8bNxzpXYj1oQZaRsnOuWzBYLFm1xaUgQESBw6QPHiQ5MFDJA8dGhEYzUVFWGfUYKuZgXXGDKzV1VhKS7AUFY0Z4DVNI93eTnL/fhL79f1iMmItL8dSVoalrBxLeRmW0lIwGEg3N+utt41NpBobSTU2kunv14Og34/Zn4spJxeT34/J60EJBMj09upLT3Y9MIDJn4tt5kxsM2dl17XYamtHHKOmaaDoF2HUcHjM4RhKf7ZoYrZFGgP6mGqjEaPLhcnrxej1YPJ4Mfm8GD1eTB73UWsPRrcHo8uJlkiQamnRz62pKTtmu4nkwYOjLuAY3W7sdXVk+vtJt7ae/HCQLIPNhmvtWtyXXYr7kkuwFBaO2kZNJolv30Fs40YSe/dizs/X3+OaamwzZmApK5Ou4EKI005C+jgkpAtxHtM06DsIh57T52dvXj+6evwRRjNUrYM5V8Ocd+pztwshTkpmcJDoK68QfvHvRF99dczK0ifEZMJcUIC5qBBLUTGmPD/KwKBeQK+zc8zCWhNiNmNyu1ECgXE3M3o8Qy2f5rw8Uu1tJPcfmHgLr9E45nRXp5PB4YBMRr8IMsmvNYLJNO70h0avF+fy5frY8JUrsdfNHQrEmqKQ7uwk1dRMqqWZdGsbBrsNc14+5vw8zHl5mPL0tcFu189L09BUDTQVVBWDwzGqZ4kQQkwHEtLHISFdCDEkGYaefdBfDwMN2aUe+hsgGRy5beE8PazPfieULAKzbWqOWYiznJpKEduwgcjfX0IJDILJrM85bzGDyYTBbAFVQY3GUGNRvXhd9rYSDJHp65tQ6DQ4nXp3ZKtVn3YvnR6xNnrc2GfNxjb7qKWmGoPVihIIHNXK3UCyoZFUcxOZru5xg7jBYsE6cyb2OXOwzZkDBkgfqT3Q3k6qvX1oTLLR5cJaU5Ntra/BWlODubBQL042GEAZHCAzMIAyGEANhzD6fPqFifx8fV1QgDkvj0xPjz6s4XB9dn34uBcpzAUF2WkBi7GUlGAuKcZcUKAXDFRVvS6CpgGaXhQxEkUNh1BCYZRQEDUURgmFUMNhlEgENRRCiUQgMzz7htHnw1ZdjbW6Cmt1tb7MmIFt5kxppRZCnJckpI9DQroQ4rg0TQ/sB5+GA0/pLe7aUS1DBqPesp4/B/JnQcEcyJ8NRQvA6pyywxbifKBlMmT6+8l0d5Pu6iLT3UOmvw+zP0/vjl5ami2i55uUQllKJEKmuzv7+t1kenuxlBRjmzMX24yacccRa5qGMjgIijKpU+RlBgdRo1E9DJtM+kUQkwnMZoxWK4ZJaGnWNA0tkUAJhTFYLTKHthBCvIWE9HFISBdCnLDYABx+Hg48CfUvQiI49nZmO1RdADOv0AvT5c0cnvdaCCGEmAxatru/UXooCDGdSUgfh4R0IcQp0TSIdEPvAX18+5GlZ59+/9FyKvVp38qWgT0H7F6w+4YXm08fmyqEEEKcCE2Dzh2w51HY+xcItELtpbDgRph7jf4ZI4SYViSkj0NCuhBiUmiaHtwPP6e3ujevB+XYU/wAYHFCzcXDLe9SnE4IIc4dShraNkHnTjBbweoGqyu7uPXPAIMhW/RO0ddHFqNZn2HEbNe3s9jB7IDu3Xoo3/MoDDaN/bomq36BeMF7YPY7wOY+k2cthDgGCenjkJAuhDgjUlFofFUP7INNkAzp3eSPLOkxqlvnzdLD+szLoXyV3vIuhBDi7BFogcMv6H/7G1/R//ZPFrND/8yY/269Nsr+v8HuP0Hv/pHbrPlHuOKrk3ccQogJkZA+DgnpQohpIZPSv0gdfk7/Qtfy5sjidBj0ivIVK6F8pR7a82ZK93ghhJhMg82w6Wew/QG9lTu3Bvw12fUM/bbJCpEeiPbow5wiPfrSvVsf/nQ0h1+fzhMgFdEv4KaikIzoP4M+ltxg0ouSGk366yoZyMQhHYdMYnh/Q8H8Bph11dit5N179bC++08w2AiX/xtc9LlJ+ecSQkychPRxSEgXQkxLiSA0vKTP4d74st4a81b2HL2afE4V5FYdta4Eb5lMCyeEmL5aN0H9C5AIQSo8HFKTET2Mmmx69+4jXbstDrC4wFuqB2P/DD0ovzWUZpIQbINgq77GAOUr9J5JE72oqWnQ9Bps+IleIFQ7hXnlDSb9wurMK2DmZVCy5NQLuqkqKEk9sB/595kITYPO7eApBU/RqR2DEOKUSUgfh4R0IcRZIdylj2Vs26R/ue3Ypn+RHY/dB65CcBeCq0Bf23PAZNHHN5oseguQ0ayvj3whtjqzX/wc+lhJXyWYzGfkNIUQ57iO7fD3b8KhZ0/P/lyFemhXFT2Yv7Vg5xH2HKhYlV1W61Nkqpnhlux0TF8PNsKmX0L3ruHnzrgEVn8cPCX6dJyDjTDQqA9dGmjQX9ud/VvrLsr+zS3UL5hWXwiOnNNzrkKIc4qE9HFISBdCnJWUNPTs1b8oBpr1LpmBFv12oGVkd8hTZbLp4xuL5g8vhfPBkasHfOlyL8Tpp2n6crp/v1JRPVj2H9aXRFAPsE6//jvtyK7Ndgi2HBVGs+tQmz4ThbcUvCV6q6y3RA+w/hn63wqra/Tr9uyDv/8/2Pe4/rPBBPOuh5wKsHr0FnGrW19bnHqLeDquh+cj61RUD+IDjfo5xAfGPkezQ9+vr0L/W9i+9fgXNcfax+IP6uG8cO6JPVcIISZAQvo4JKQLIc45mgaJwPC4yGgPRHr1FqZkSA/4aia7TutrJZX9Ihwf+cU4EZzAl1vDcMu80ax/2c+bpXfFz5sJ+bP12w4/xPoh1gfRXoj26+tUFFz5+pd+T7H+pd+Zd/aGf1WBrl16kajGV/RQUbVOHy9ac7HeU0Gcn9IJPVz2HYT+Q9B3WL8d7dO7Lyup4d9HJQUY9N8Nd7HePdldrLfSOvMgGdZ/f2J9+u9SrE///TKYstXCncPVwy1O/W9Cfz2E2if/PHMqoaBOD+wFc/Tfg50PA5p+TgvfB5f8C+TVntrrxAPDrdpGkx7Kcyr1fx+DYXg7Ja3/TrZuhNYN+jrUpj9mcQ7/G1ndYPPA3Kth6S363zIhhJgkEtLHISFdCCHGoap663z3Hr3lvnu3fru/Hv0L9yQxWvRuo1an3hXfZNXH2Jssesu+zZNt+cu2+h257SnWu77aPKf3eDRN/3K/82E97Nhz9OEEjuzanqNfEGl8WR/LmgiMvR+zXQ/qs96uF3tyFWaHHJhGhopzVSqqD90wWfX3zeo6t8472qeH7mA7hDsg1HnUujM7RnoafM1y+PULaHm1eqCNByA+qLdMxwchNqC3QPvK9akgh4qlVev3JUJvOb/s0n9Yv3BwLHXvgkvvhsK6M3Si40hnx72frRcDhRBnPQnp45CQLoQQJ+FIV1RV0Vvlj17CndB3SF/6s+tAs158yWDUQ4GrYHhtdeot/eFskBnvS/5EuQr1AOKfoS+uguEQEuuH2KC+zsSheBFUXaC3dudUjgyNoQ7Y8QBsv18PIBNl9UD1BXog91VkiwA+q7eqj8kwfDHiSK2AEbct+kUKTdNbXDOpo9YpvRWwfLk+1rZiNRQv1J9zJmWSeittsD27bhv981svXpisemA80tX6yDmijVwbjvHvY7YPjwP2lAy3Nrvy9aAZatf/Tx0JkeFO/f9BMqL3KkmGhxe04dbUo1uibZ5sN/CjLgY5/fprDzToszL0HtTXx+p+fTSbN9vLJNvbJH/WcKHHEedn02d4iGQrhoe7spXDu/X/uzavfp5Dv0v5+lrTjhpnHR2+bXUfFcwnsYU42p/9NzlqcebBBf8MpUsn73WFEOIsIyF9HBLShRDiDMgk9aBgzzl+y1UmNRxGMgn9uUo62x04rd+XCGUD98BR4TsbymL9J3+c3nI9rJcs1oN1/QvDlZ0tLn2ao6L5w/PbxwPZ2wG90F71hVDzNr2C81uL7WmaPi730DNw8Bm9Zf5UqkaPx+yAsuVQukQ/rrfSND2YJgLZcwgMn4uazm6UvVhx5KKFwTT6woHJqm8f6pj4xRWLS7+wMPQ65xKDPhY6pyp70aA4O4wjO2Y7t0q/oHAu9R4QQghxUiSkj0NCuhBCnGPiAb2F8+glNpBtCc0DZ3bt8OvdzNs2QfN6vWK+mhm9v8q1sORmPaCfzm70SmbscciZbIA90ko+9HgSMIDZOtzSas6uY33Z8bbZMbfH6m4/2cx2vVXYV6Zf8PCVZX8uH77f7htu7T26e3V8QD9PDNkQaxgOs5o2sn7CUB2F2HAr81BLc4/eAm0wDrewe0uHw7IzT38fbd7sOrsYDJCKjZy7OpVtcR86xqOONRXTQ3fB3OGx13mzpOaAEEKICZGQPg4J6UIIIQA9lLVtgqbXoXOH3mK+5GbInznVR3ZiVFUfZtC6QW+5V5XsA2/5eLe6s2Pqc4bXdp/e7Rqy3c2P3m/mLUE5NRyqvSV6KHf6p76VWFX0HgE2r0wdKIQQYtqSkD4OCelCCCGEEEIIIc6kE8mhUuJSCCGEEEIIIYSYJiSkCyGEEEIIIYQQ04SEdCGEEEIIIYQQYpqQkC6EEEIIIYQQQkwTEtKFEEIIIYQQQohpQkK6EEIIIYQQQggxTUhIF0IIIYQQQgghpgkJ6UIIIYQQQgghxDQhIV0IIYQQQgghhJgmJKQLIYQQQgghhBDThIR0IYQQQgghhBBimpCQLoQQQgghhBBCTBMS0oUQQgghhBBCiGliWoT0++67j+rqaux2O6tXr2bjxo3H3PbXv/41BoNhxGK328/g0QohhBBCCCGEEJNjykP6Qw89xF133cVXv/pVtm7dyuLFi7nqqqvo6ek55nO8Xi+dnZ1DS3Nz8xk8YiGEEEIIIYQQYnJMeUj/3ve+x5133skdd9zBvHnz+MlPfoLT6eSXv/zlMZ9jMBgoLi4eWoqKis7gEQshhBBCCCGEEJNjSkN6KpViy5YtXHHFFUP3GY1GrrjiCt54441jPi8SiVBVVUVFRQXXX389e/bsOea2yWSSUCg0YhFCCCGEEEIIIaajKQ3pfX19KIoyqiW8qKiIrq6uMZ8zZ84cfvnLX/LYY4/x+9//HlVVWbduHW1tbWNu/61vfQufzze0VFRUnPbzEEIIIYQQQgghTgfzVB/AiVq7di1r164d+nndunXU1dXxv//7v/zHf/zHqO2/9KUvcddddw39HAwGqayslBZ1IYQQQgghhBBnxJH8qWnacbed0pCen5+PyWSiu7t7xP3d3d0UFxdPaB8Wi4WlS5dy+PDhMR+32WzYbLahn4/840iLuhBCCCGEEEKIMykcDuPz+cbdZkpDutVqZfny5bzwwgvccMMNAKiqygsvvMCnPvWpCe1DURR27drF1VdfPaHtS0tLaW1txePxYDAYTvbQT5tQKERFRQWtra14vd6pPhxxBsh7fn6R9/v8Iu/3+Ufe8/OLvN/nF3m/zz+T+Z5rmkY4HKa0tPS42055d/e77rqL2267jRUrVrBq1SruvfdeotEod9xxBwC33norZWVlfOtb3wLg3//931mzZg0zZ84kEAhwzz330NzczMc+9rEJvZ7RaKS8vHzSzudkeb1e+eU/z8h7fn6R9/v8Iu/3+Ufe8/OLvN/nF3m/zz+T9Z4frwX9iCkP6R/4wAfo7e3l3/7t3+jq6mLJkiU8/fTTQ8XkWlpaMBqH69sNDg5y55130tXVRW5uLsuXL2f9+vXMmzdvqk5BCCGEEEIIIYQ4LQzaREaui0kTCoXw+XwEg0G5QneekPf8/CLv9/lF3u/zj7zn5xd5v88v8n6ff6bLez6lU7AJvbDdV7/61RHF7cS5Td7z84u83+cXeb/PP/Ken1/k/T6/yPt9/pku77m0pAshhBBCCCGEENOEtKQLIYQQQgghhBDThIR0IYQQQgghhBBimpCQLoQQQgghhBBCTBMS0oUQQgghhBBCiGlCQvoUu++++6iursZut7N69Wo2btw41YckToNvfetbrFy5Eo/HQ2FhITfccAMHDhwYsU0ikeCTn/wkeXl5uN1ubrzxRrq7u6foiMXp9O1vfxuDwcBnPvOZofvk/T73tLe38+EPf5i8vDwcDgcLFy5k8+bNQ49rmsa//du/UVJSgsPh4IorruDQoUNTeMTiZCmKwle+8hVqampwOBzU1tbyH//xHxxde1fe77PXK6+8wnXXXUdpaSkGg4G//OUvIx6fyHs7MDDAzTffjNf7/9u7/5io6z8O4M+Dg+OHISBy/HAoLiaIyhDSEFsrnEAuUiknu9FpbQwFRS3CKNJWZlipaXWmS/sDkqJJIZMcAlE4fsVvAtEtQguJTBEE+RH3/v7R+ux7YkoJfO7O52O7jXu/3354fXjujs+Lz30+OsDR0RHPP/88bty4MYl7Qf/GnTIfHh5GSkoK5s+fD3t7e3h4eODZZ59FR0eHwTaYuem422v8/8XHx0OhUGD//v0G45OdN5t0GX3++efYtm0bduzYgZqaGgQEBCA8PBxdXV1yl0b3qKSkBAkJCSgvL0dBQQGGh4exfPly9PX1SWu2bt2KkydPIjs7GyUlJejo6MDq1atlrJrGQ1VVFT7++GMsWLDAYJx5m5dr164hNDQUVlZWyM/PR3NzM9577z04OTlJa/bs2YMDBw7g0KFDqKiogL29PcLDwzEwMCBj5fRfpKenQ6fT4YMPPkBLSwvS09OxZ88eHDx4UFrDvE1XX18fAgIC8OGHH952fizZajQa/PjjjygoKEBeXh6+++47xMXFTdYu0L90p8z7+/tRU1ODtLQ01NTU4MSJE2htbUVUVJTBOmZuOu72Gv9bTk4OysvL4eHhMWpu0vMWJJtFixaJhIQE6fnIyIjw8PAQu3fvlrEqmghdXV0CgCgpKRFCCNHd3S2srKxEdna2tKalpUUAEGVlZXKVSfeot7dX+Pj4iIKCAvHoo4+KpKQkIQTzNkcpKSli6dKl/ziv1+uFm5ubeOedd6Sx7u5uoVKpxPHjxyejRBpHK1asEM8995zB2OrVq4VGoxFCMG9zAkDk5ORIz8eSbXNzswAgqqqqpDX5+flCoVCIX3/9ddJqp//m1sxvp7KyUgAQ7e3tQghmbsr+Ke9ffvlFeHp6iqamJjFz5kyxb98+aU6OvHkmYNQESAAACmJJREFUXSZDQ0Oorq7GsmXLpDELCwssW7YMZWVlMlZGE+H69esAAGdnZwBAdXU1hoeHDfL39fWFl5cX8zdhCQkJWLFihUGuAPM2R7m5uQgODsYzzzwDV1dXBAYG4siRI9J8W1sbOjs7DTKfOnUqFi9ezMxN0JIlS1BYWIjz588DAOrr61FaWorIyEgAzNucjSXbsrIyODo6Ijg4WFqzbNkyWFhYoKKiYtJrpvF3/fp1KBQKODo6AmDm5kav1yM2NhbJycnw9/cfNS9H3soJ2Srd1ZUrVzAyMgK1Wm0wrlarce7cOZmqoomg1+uxZcsWhIaGYt68eQCAzs5OWFtbS2/2f1Or1ejs7JShSrpXWVlZqKmpQVVV1ag55m1+fvrpJ+h0Omzbtg2pqamoqqrC5s2bYW1tDa1WK+V6u/d4Zm56tm/fjp6eHvj6+sLS0hIjIyPYtWsXNBoNADBvMzaWbDs7O+Hq6mowr1Qq4ezszPzNwMDAAFJSUhATEwMHBwcAzNzcpKenQ6lUYvPmzbedlyNvNulEEywhIQFNTU0oLS2VuxSaIJcuXUJSUhIKCgpgY2Mjdzk0CfR6PYKDg/HWW28BAAIDA9HU1IRDhw5Bq9XKXB2Nty+++AKZmZn47LPP4O/vj7q6OmzZsgUeHh7Mm8iMDQ8PY82aNRBCQKfTyV0OTYDq6mq8//77qKmpgUKhkLscCT/uLhMXFxdYWlqOurvzb7/9Bjc3N5mqovGWmJiIvLw8FBcXY8aMGdK4m5sbhoaG0N3dbbCe+Zum6upqdHV1YeHChVAqlVAqlSgpKcGBAwegVCqhVquZt5lxd3fH3LlzDcb8/Pxw8eJFAJBy5Xu8eUhOTsb27duxdu1azJ8/H7Gxsdi6dSt2794NgHmbs7Fk6+bmNuqmv3/++SeuXr3K/E3Y3w16e3s7CgoKpLPoADM3J99//z26urrg5eUlHcO1t7fjhRdewKxZswDIkzebdJlYW1sjKCgIhYWF0pher0dhYSFCQkJkrIzGgxACiYmJyMnJQVFREby9vQ3mg4KCYGVlZZB/a2srLl68yPxNUFhYGBobG1FXVyc9goODodFopK+Zt3kJDQ0d9d8qnj9/HjNnzgQAeHt7w83NzSDznp4eVFRUMHMT1N/fDwsLw0MmS0tL6PV6AMzbnI0l25CQEHR3d6O6ulpaU1RUBL1ej8WLF096zXTv/m7QL1y4gDNnzmDatGkG88zcfMTGxqKhocHgGM7DwwPJyck4ffo0AJnynpDb0dGYZGVlCZVKJT799FPR3Nws4uLihKOjo+js7JS7NLpHGzZsEFOnThXffvutuHz5svTo7++X1sTHxwsvLy9RVFQkfvjhBxESEiJCQkJkrJrG0//f3V0I5m1uKisrhVKpFLt27RIXLlwQmZmZws7OTmRkZEhr3n77beHo6Ci+/vpr0dDQIJ566inh7e0tbt68KWPl9F9otVrh6ekp8vLyRFtbmzhx4oRwcXERL730krSGeZuu3t5eUVtbK2prawUAsXfvXlFbWyvdyXss2UZERIjAwEBRUVEhSktLhY+Pj4iJiZFrl+gu7pT50NCQiIqKEjNmzBB1dXUGx3GDg4PSNpi56bjba/xWt97dXYjJz5tNuswOHjwovLy8hLW1tVi0aJEoLy+XuyQaBwBu+zh27Ji05ubNm2Ljxo3CyclJ2NnZiVWrVonLly/LVzSNq1ubdOZtfk6ePCnmzZsnVCqV8PX1FYcPHzaY1+v1Ii0tTajVaqFSqURYWJhobW2VqVq6Fz09PSIpKUl4eXkJGxsbMXv2bPHKK68YHLAzb9NVXFx829/ZWq1WCDG2bP/44w8RExMjpkyZIhwcHMT69etFb2+vDHtDY3GnzNva2v7xOK64uFjaBjM3HXd7jd/qdk36ZOetEEKIiTlHT0RERERERET/Bq9JJyIiIiIiIjISbNKJiIiIiIiIjASbdCIiIiIiIiIjwSadiIiIiIiIyEiwSSciIiIiIiIyEmzSiYiIiIiIiIwEm3QiIiIiIiIiI8EmnYiIiIiIiMhIsEknIiKicadQKPDVV1/JXQYREZHJYZNORERkZtatWweFQjHqERERIXdpREREdBdKuQsgIiKi8RcREYFjx44ZjKlUKpmqISIiorHimXQiIiIzpFKp4ObmZvBwcnIC8NdH0XU6HSIjI2Fra4vZs2fjyy+/NPj3jY2NePzxx2Fra4tp06YhLi4ON27cMFhz9OhR+Pv7Q6VSwd3dHYmJiQbzV65cwapVq2BnZwcfHx/k5uZKc9euXYNGo8H06dNha2sLHx+fUX9UICIiuh+xSSciIroPpaWlITo6GvX19dBoNFi7di1aWloAAH19fQgPD4eTkxOqqqqQnZ2NM2fOGDThOp0OCQkJiIuLQ2NjI3Jzc/Hggw8afI/XX38da9asQUNDA5544gloNBpcvXpV+v7Nzc3Iz89HS0sLdDodXFxcJu8HQEREZKQUQgghdxFEREQ0ftatW4eMjAzY2NgYjKempiI1NRUKhQLx8fHQ6XTS3MMPP4yFCxfio48+wpEjR5CSkoJLly7B3t4eAHDq1Ck8+eST6OjogFqthqenJ9avX48333zztjUoFAq8+uqreOONNwD81fhPmTIF+fn5iIiIQFRUFFxcXHD06NEJ+ikQERGZJl6TTkREZIYee+wxgyYcAJydnaWvQ0JCDOZCQkJQV1cHAGhpaUFAQIDUoANAaGgo9Ho9WltboVAo0NHRgbCwsDvWsGDBAulre3t7ODg4oKurCwCwYcMGREdHo6amBsuXL8fKlSuxZMmS/7SvRERE5oRNOhERkRmyt7cf9fHz8WJrazumdVZWVgbPFQoF9Ho9ACAyMhLt7e04deoUCgoKEBYWhoSEBLz77rvjXi8REZEp4TXpRERE96Hy8vJRz/38/AAAfn5+qK+vR19fnzR/9uxZWFhYYM6cOXjggQcwa9YsFBYW3lMN06dPh1arRUZGBvbv34/Dhw/f0/aIiIjMAc+kExERmaHBwUF0dnYajCmVSunmbNnZ2QgODsbSpUuRmZmJyspKfPLJJwAAjUaDHTt2QKvVYufOnfj999+xadMmxMbGQq1WAwB27tyJ+Ph4uLq6IjIyEr29vTh79iw2bdo0pvpee+01BAUFwd/fH4ODg8jLy5P+SEBERHQ/Y5NORERkhr755hu4u7sbjM2ZMwfnzp0D8Ned17OysrBx40a4u7vj+PHjmDt3LgDAzs4Op0+fRlJSEh566CHY2dkhOjoae/fulbal1WoxMDCAffv24cUXX4SLiwuefvrpMddnbW2Nl19+GT///DNsbW3xyCOPICsraxz2nIiIyLTx7u5ERET3GYVCgZycHKxcuVLuUoiIiOgWvCadiIiIiIiIyEiwSSciIiIiIiIyErwmnYiI6D7DK92IiIiMF8+kExERERERERkJNulERERERERERoJNOhEREREREZGRYJNOREREREREZCTYpBMREREREREZCTbpREREREREREaCTToRERERERGRkWCTTkRERERERGQk/gd/QjuugYaV7AAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "for i, train_loss in enumerate(train_loss_list):\n",
    "    plt.plot(\n",
    "        range(1, len(train_loss) + 1), train_loss, label=f\"Fold {i+1} Training Loss\"\n",
    "    )\n",
    "plt.title(\"Training Loss for Each Fold\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.savefig(\"result/transformer_split_train_loss.svg\")\n",
    "plt.show()\n",
    "\n",
    "# Plot the validation loss for each fold\n",
    "plt.figure(figsize=(12, 6))\n",
    "for i, val_loss in enumerate(val_loss_list):\n",
    "    plt.plot(range(1, len(val_loss) + 1), val_loss, label=f\"Fold {i+1} Validation Loss\")\n",
    "plt.title(\"Validation Loss for Each Fold\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend()\n",
    "plt.savefig(\"result/transformer_split_val_loss.svg\")\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.11.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}