tensorboard-visualization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b65dfbdc-bb80-4bd4-8248-dd06483222cd",
   "metadata": {},
   "source": [
    "# The Measurements and Visualizations of PorkCNN with TensorBoard\n",
    "\n",
    "author: davidycliao(David Yen-Chieh Liao)                                                                      \n",
    "email: davidycliao@gmail.com                                                             \n",
    "date: 9-July-2021  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e8005d5a-f938-4c26-86d5-592408152d08",
   "metadata": {},
   "source": [
    "-------------------------\n",
    "\n",
    "### Stage 1: Libaries & Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "6ca92ba3-7480-418f-91e0-d2beea2d10af",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:06.695144Z",
     "iopub.status.busy": "2021-07-19T07:15:06.694924Z",
     "iopub.status.idle": "2021-07-19T07:15:10.699521Z",
     "shell.execute_reply": "2021-07-19T07:15:10.698921Z",
     "shell.execute_reply.started": "2021-07-19T07:15:06.695120Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# built-in library\n",
    "import math\n",
    "import re\n",
    "import collections\n",
    "import zipfile\n",
    "import random\n",
    "from itertools import chain\n",
    "\n",
    "# ML & Deep Learning/ NLP toolkit\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import jieba\n",
    "from sklearn.model_selection import train_test_split\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras import layers\n",
    "import tensorflow_datasets as tfds\n",
    "from tensorflow.keras.callbacks import EarlyStopping,TensorBoard\n",
    "\n",
    "# Visualization\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8bd77e05-355a-422f-87db-2cf77cba3bdd",
   "metadata": {},
   "source": [
    "-------------------------\n",
    "\n",
    "### Stage 2: Data Preprocessing (Training Data: Introduction of Bills and Legislation from 6th Session to 7th Session, 2004-2012)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b57b9c9-443e-4e26-bda2-15f6877845d5",
   "metadata": {},
   "source": [
    "#### (1) Read file "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e1973cfb-73d4-4717-a67a-a4e8a0c9edb0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:11.200686Z",
     "iopub.status.busy": "2021-07-19T07:15:11.200529Z",
     "iopub.status.idle": "2021-07-19T07:15:11.236315Z",
     "shell.execute_reply": "2021-07-19T07:15:11.235668Z",
     "shell.execute_reply.started": "2021-07-19T07:15:11.200667Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# read file\n",
    "df = pd.read_csv('data/Pork Bill - 2021-05-20.csv',encoding='utf-8')\n",
    "\n",
    "# combine abstract of bill and title \n",
    "df['text'] =  df['Title'] + df['Content'].fillna(df['Title'])\n",
    "\n",
    "# drop conten without having any characters\n",
    "# view na's row: df[df['text'].isnull()==True]\n",
    "data = df[['text', 'pork_bill']].dropna(subset=['text'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "f6160785-9522-4404-8f77-6df941869cde",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:11.238378Z",
     "iopub.status.busy": "2021-07-19T07:15:11.238165Z",
     "iopub.status.idle": "2021-07-19T07:15:11.242959Z",
     "shell.execute_reply": "2021-07-19T07:15:11.242384Z",
     "shell.execute_reply.started": "2021-07-19T07:15:11.238354Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Pork Legislation 2510 \n",
      " None-Pork Legislation 4733\n"
     ]
    }
   ],
   "source": [
    "print(\" Pork Legislation\", data['pork_bill'].value_counts()[1],'\\n', \n",
    "      \"None-Pork Legislation\", data['pork_bill'].value_counts()[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "11f1b3ab-7f1e-40b0-995a-e66ae58da102",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:11.244202Z",
     "iopub.status.busy": "2021-07-19T07:15:11.243924Z",
     "iopub.status.idle": "2021-07-19T07:15:11.380395Z",
     "shell.execute_reply": "2021-07-19T07:15:11.379860Z",
     "shell.execute_reply.started": "2021-07-19T07:15:11.244179Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlwAAAFlCAYAAAA+gTZIAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAAARZ0lEQVR4nO3dfYxmZ1nH8d9lFxRFKdANIW11G6kxxWjVtYDGlxQtBQ2tCtoGZSWNjUlVNL6BfwiiJKDBKiom1Ta2DbFUNLYaDGlaUHyhsLVQLJWwoKRtkC5sAVGBtF7+MWfJuO50Ztu9Zmann0+ymXPucz/Pc59k8+Q753mZ6u4AADDni7Z6AQAAO53gAgAYJrgAAIYJLgCAYYILAGCY4AIAGLZrqxfwUE455ZTes2fPVi8DAGBdt91228e7e/fRjm3r4NqzZ0/279+/1csAAFhXVX1krWNeUgQAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGLZrqxewHVx0zZu3egnwqHbdi1+w1UsAGOUKFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwzYcXFV1UlXdXlV/teyfUVW3VtWBqnpTVT12Gf/iZf/AcnzPqvt4+TL+gap6znE/GwCAbehYrnC9NMldq/Zfm+Ty7n5akvuTXLKMX5Lk/mX88mVequqsJBcleXqS85O8oapOemTLBwDY/jYUXFV1WpLvTfJHy34lOTfJm5cpVye5cNm+YNnPcvzZy/wLklzX3Z/r7n9NciDJOcfhHAAAtrWNXuH67SS/mOR/lv0nJ/lkdz+w7N+T5NRl+9QkdyfJcvxTy/wvjB/lNl9QVZdW1f6q2n/w4MGNnwkAwDa1bnBV1fclua+7b9uE9aS7r+juvd29d/fu3ZvxkAAAo3ZtYM63JXl+VT0vyZck+Yokv5Pk5KratVzFOi3Jvcv8e5OcnuSeqtqV5AlJPrFq/LDVtwEA2LHWvcLV3S/v7tO6e09W3vR+S3e/KMnbkrxgmbYvyQ3L9o3Lfpbjt3R3L+MXLZ9iPCPJmUneddzOBABgm9rIFa61/FKS66rq15PcnuTKZfzKJNdW1YEkh7ISaenuO6vq+iTvT/JAksu6+8FH8PgAACeEYwqu7n57krcv2x/OUT5l2N2fTfLCNW7/6iSvPtZFAgCcyHzTPADAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAw9YNrqr6kqp6V1W9t6rurKpfXcbPqKpbq+pAVb2pqh67jH/xsn9gOb5n1X29fBn/QFU9Z+ysAAC2kY1c4fpcknO7+xuSnJ3k/Kp6ZpLXJrm8u5+W5P4klyzzL0ly/zJ++TIvVXVWkouSPD3J+UneUFUnHcdzAQDYltYNrl7xmWX3Mcu/TnJukjcv41cnuXDZvmDZz3L82VVVy/h13f257v7XJAeSnHM8TgIAYDvb0Hu4quqkqnpPkvuS3JTkQ0k+2d0PLFPuSXLqsn1qkruTZDn+qSRPXj1+lNsAAOxYGwqu7n6wu89OclpWrkp97dSCqurSqtpfVfsPHjw49TAAAJvmmD6l2N2fTPK2JM9KcnJV7VoOnZbk3mX73iSnJ8ly/AlJPrF6/Ci3Wf0YV3T33u7eu3v37mNZHgDAtrSRTynurqqTl+3HJfmeJHdlJbxesEzbl+SGZfvGZT/L8Vu6u5fxi5ZPMZ6R5Mwk7zpO5wEAsG3tWn9Knprk6uUThV+U5Pru/quqen+S66rq15PcnuTKZf6VSa6tqgNJDmXlk4np7jur6vok70/yQJLLuvvB43s6AADbz7rB1d13JPnGo4x/OEf5lGF3fzbJC9e4r1cnefWxLxMA4MTlm+YBAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGDYrq1eAMCjwUXXvHmrlwCPate9+AVb+viucAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAwTXAAAwwQXAMAwwQUAMExwAQAME1wAAMMEFwDAMMEFADBMcAEADBNcAADDBBcAwDDBBQAwTHABAAxbN7iq6vSqeltVvb+q7qyqly7jT6qqm6rqg8vPJy7jVVWvr6oDVXVHVX3Tqvvat8z/YFXtmzstAIDtYyNXuB5I8nPdfVaSZya5rKrOSvKyJDd395lJbl72k+S5Sc5c/l2a5A+SlUBL8ookz0hyTpJXHI40AICdbN3g6u6Pdvc/Ldv/keSuJKcmuSDJ1cu0q5NcuGxfkOSaXvHOJCdX1VOTPCfJTd19qLvvT3JTkvOP58kAAGxHx/Qerqrak+Qbk9ya5Cnd/dHl0L8necqyfWqSu1fd7J5lbK1xAIAdbcPBVVWPT/JnSX6muz+9+lh3d5I+Hguqqkuran9V7T948ODxuEsAgC21oeCqqsdkJbbe2N1/vgx/bHmpMMvP+5bxe5Ocvurmpy1ja43/H919RXfv7e69u3fvPpZzAQDYljbyKcVKcmWSu7r7t1YdujHJ4U8a7ktyw6rxFy+fVnxmkk8tLz2+Ncl5VfXE5c3y5y1jAAA72q4NzPm2JD+a5H1V9Z5l7JeTvCbJ9VV1SZKPJPmh5dhbkjwvyYEk/5XkJUnS3Yeq6teSvHuZ96ruPnQ8TgIAYDtbN7i6+++S1BqHn32U+Z3ksjXu66okVx3LAgEATnS+aR4AYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBg2LrBVVVXVdV9VfXPq8aeVFU3VdUHl59PXMarql5fVQeq6o6q+qZVt9m3zP9gVe2bOR0AgO1nI1e4/jjJ+UeMvSzJzd19ZpKbl/0keW6SM5d/lyb5g2Ql0JK8IskzkpyT5BWHIw0AYKdbN7i6+2+THDpi+IIkVy/bVye5cNX4Nb3inUlOrqqnJnlOkpu6+1B335/kpvz/iAMA2JEe7nu4ntLdH122/z3JU5btU5PcvWrePcvYWuP/T1VdWlX7q2r/wYMHH+byAAC2j0f8pvnu7iR9HNZy+P6u6O693b139+7dx+tuAQC2zMMNro8tLxVm+XnfMn5vktNXzTttGVtrHABgx3u4wXVjksOfNNyX5IZV4y9ePq34zCSfWl56fGuS86rqicub5c9bxgAAdrxd602oqj9J8l1JTqmqe7LyacPXJLm+qi5J8pEkP7RMf0uS5yU5kOS/krwkSbr7UFX9WpJ3L/Ne1d1HvhEfAGBHWje4uvviNQ49+yhzO8lla9zPVUmuOqbVAQDsAL5pHgBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYYILAGCY4AIAGCa4AACGCS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIYJLgCAYZseXFV1flV9oKoOVNXLNvvxAQA226YGV1WdlOT3kzw3yVlJLq6qszZzDQAAm22zr3Cdk+RAd3+4uz+f5LokF2zyGgAANtVmB9epSe5etX/PMgYAsGPt2uoFHKmqLk1y6bL7mar6wFauhxPCKUk+vtWL4OF7076tXgGsy/PMCW6Tnme+aq0Dmx1c9yY5fdX+acvYF3T3FUmu2MxFcWKrqv3dvXer1wHsXJ5neKQ2+yXFdyc5s6rOqKrHJrkoyY2bvAYAgE21qVe4uvuBqvrJJG9NclKSq7r7zs1cAwDAZtv093B191uSvGWzH5cdzUvQwDTPMzwi1d1bvQYAgB3Nn/YBABgmuNh0VdVV9bpV+z9fVa9c5zYXrvVXCarqlVV1b1W9p6r+uaqef4zr+beqOuVYbgPsDFX14Krnjj+tqi89htv+WFX93uT62DkEF1vhc0l+4Bgj58Ks/DmotVze3WcneWGSq6pqQ/+3lz83BTx6/Xd3n93dX5fk80l+YiM3qqpt9z2WbG+Ci63wQFbegPqzRx6oqj1VdUtV3VFVN1fVV1bVtyZ5fpLfXH4T/eq17ri771ru/5Squriq3rf85vraVY/xmap6XVW9N8mzVo0/rqr+uqp+/DieK3DieEeSp1XVk6rqL5bnoXdW1dcnX7iafm1V/X2Sa1ffsKq+t6r+0dVy1iK42Cq/n+RFVfWEI8Z/N8nV3f31Sd6Y5PXd/Q9Z+b62X1h+E/3QWndaVc9I8j9JHpPktUnOTXJ2km+pqguXaV+W5Nbu/obu/rtl7PFJ/jLJn3T3Hx6PEwROHMsVq+cmeV+SX01y+/I89MtJrlk19awk393dF6+67fcneVmS53W3b6PnqFwSZUt096er6pokP53kv1cdelaSH1i2r03yGxu8y5+tqh9J8h9JfjjJ3iRv7+6DSVJVb0zyHUn+IsmDSf7siNvfkOQ3uvuNx342wAnscVX1nmX7HUmuTHJrkh9Mku6+paqeXFVfscy5sbtXP2edm5Xnm/O6+9ObtGZOQK5wsZV+O8klWbni9Ehdvlz9+vbufsc6cz/b3Q8eMfb3Sc6vqjoOawFOHIffw3V2d/9Ud39+nfn/ecT+h5J8eZKvmVkeO4XgYst096Ek12clug77h6z8yackeVFWfuNMVq5cffkx3P27knxnVZ2yvDH+4iR/8xDzfyXJ/Vl5qRN4dHtHVp5/UlXfleTjD3H16iNZuRp2TVU9fVNWxwlJcLHVXpdk9ZtMfyrJS6rqjiQ/muSly/h1SX6hqm5/qDfNH9bdH83KeyreluS9SW7r7hvWudlLs/LywkZfxgR2plcm+ebleeg1SfY91OTu/pesBNqfbuT5iUcn3zQPADDMFS4AgGGCCwBgmOACABgmuAAAhgkuAIBhggsAYJjgAgAYJrgAAIb9L+qlZn6yko4wAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "plt.figure(figsize=(10, 6))\n",
    "pork_plot = pd.DataFrame({\"num\":data['pork_bill'].value_counts(), \"type\":['Not Pork','Pork']})\n",
    "plt.bar(pork_plot['type'], pork_plot['num'], color='#57A1AB')\n",
    "#plt.show()\n",
    "plt.savefig('images/full_plot.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5e6d677d-e200-40e6-9252-130eb02fc682",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:01:03.001989Z",
     "iopub.status.busy": "2021-05-16T17:01:03.001801Z",
     "iopub.status.idle": "2021-05-16T17:01:03.004600Z",
     "shell.execute_reply": "2021-05-16T17:01:03.003789Z",
     "shell.execute_reply.started": "2021-05-16T17:01:03.001969Z"
    }
   },
   "source": [
    "#### (2) Tokenization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "120e8b54-7f75-4ca6-a0d0-9872b1522e8e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:11.381385Z",
     "iopub.status.busy": "2021-07-19T07:15:11.381255Z",
     "iopub.status.idle": "2021-07-19T07:15:11.387973Z",
     "shell.execute_reply": "2021-07-19T07:15:11.387191Z",
     "shell.execute_reply.started": "2021-07-19T07:15:11.381368Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import collections\n",
    "import numpy as np\n",
    "import jieba\n",
    "from itertools import chain\n",
    "\n",
    "\n",
    "def jieba_cut(filename):\n",
    "    \"\"\"\n",
    "    cut Chinese and remove stop words\n",
    "    Reference: https://www.cnblogs.com/Luv-GEM/p/10836454.html\n",
    "    Stopwords: https://www.kaggle.com/rikdifos/english-and-chinese-stopwords?select=cn_stopwords.txt\n",
    "    \"\"\"\n",
    "    stop_list = [i.strip() for i in open('cn_stopwords.txt','r',encoding='utf-8')]  \n",
    "    news_cut = []\n",
    "    news_list = []\n",
    "    for line in filename:    \n",
    "        if line:\n",
    "            news_cut = list(jieba.cut(''.join(line),cut_all=False,HMM=True))  \n",
    "            news_list.append([word.strip() for word in news_cut if word not in stop_list and len(word.strip())>0]) \n",
    "    news_list = list(chain.from_iterable(news_list))  \n",
    "    return news_list\n",
    "\n",
    "def clearPucts(context):\n",
    "    \"\"\"\n",
    "    remove punctuation\n",
    "    ref: https://chenyuzuoo.github.io/posts/28001/\n",
    "    \"\"\"\n",
    "    context = re.sub(\"[\\s+\\.\\!\\/_,$%^*(+\\\"\\']+|[+——！，。？、~@#￥%……&*（）]+\", \"\", context)\n",
    "    context = re.sub(\"[【】╮╯▽╰╭★→「」]+\",\"\", context)\n",
    "    context = re.sub(\"！，❤。～《》：（）【】「」？”“；：、\",\"\",context)\n",
    "    context = re.sub(\"\\s\",\"\",context)\n",
    "    return context\n",
    "\n",
    "def seg_char(sent):\n",
    "    \"\"\"\n",
    "    cut Chinese and remove stop words\n",
    "    ref: https://blog.csdn.net/renyuanfang/article/details/86487367\n",
    "    \"\"\"\n",
    "    # split\n",
    "    pattern_char_1 = re.compile(r'([\\W])')\n",
    "    parts = pattern_char_1.split(sent)\n",
    "    parts = [p for p in parts if len(p.strip())>0]\n",
    "    # cut sentence\n",
    "    pattern = re.compile(r'([\\u4e00-\\u9fa5])')\n",
    "    chars = pattern.split(sent)\n",
    "    chars = [w for w in chars if len(w.strip())>0]\n",
    "    chars = ' '.join(chars)\n",
    "    return chars"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "25f4277c-001d-47a4-81d0-f5ac5ec5e5d3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:11.389416Z",
     "iopub.status.busy": "2021-07-19T07:15:11.389270Z",
     "iopub.status.idle": "2021-07-19T07:15:13.437772Z",
     "shell.execute_reply": "2021-07-19T07:15:13.437193Z",
     "shell.execute_reply.started": "2021-07-19T07:15:11.389399Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "data_clean = [seg_char(text) for text in [clearPucts(text) for text in data.text]]\n",
    "\n",
    "tokenizer = tfds.deprecated.text.SubwordTextEncoder.build_from_corpus(\n",
    "    data_clean, target_vocab_size=2**18)\n",
    "\n",
    "data_inputs = [tokenizer.encode(sentence) for sentence in data_clean]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "089e65af-e49c-4d60-962e-a0a826809aa5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:01:47.611294Z",
     "iopub.status.busy": "2021-05-16T17:01:47.611020Z",
     "iopub.status.idle": "2021-05-16T17:01:47.614208Z",
     "shell.execute_reply": "2021-05-16T17:01:47.613487Z",
     "shell.execute_reply.started": "2021-05-16T17:01:47.611276Z"
    }
   },
   "source": [
    "#### (3) Padding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b801d231-972c-47e3-83c4-7375c2e11a18",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:13.440283Z",
     "iopub.status.busy": "2021-07-19T07:15:13.440126Z",
     "iopub.status.idle": "2021-07-19T07:15:13.505199Z",
     "shell.execute_reply": "2021-07-19T07:15:13.504698Z",
     "shell.execute_reply.started": "2021-07-19T07:15:13.440264Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Maximun length:785 \n",
      "Input:7243\n"
     ]
    }
   ],
   "source": [
    "MAX_LEN = max([len(sentence) for sentence in data_clean])\n",
    "data_inputs = tf.keras.preprocessing.sequence.pad_sequences(data_inputs,\n",
    "                                                            value=0,\n",
    "                                                            padding=\"post\",\n",
    "                                                            maxlen=MAX_LEN)\n",
    "\n",
    "print('Maximun length:{} \\nInput:{}'.format(MAX_LEN, data_inputs.shape[0]))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f69d90cf-ac3c-4e08-b46b-457d9708d832",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:02:28.107037Z",
     "iopub.status.busy": "2021-05-16T17:02:28.106843Z",
     "iopub.status.idle": "2021-05-16T17:02:28.109696Z",
     "shell.execute_reply": "2021-05-16T17:02:28.108954Z",
     "shell.execute_reply.started": "2021-05-16T17:02:28.107017Z"
    }
   },
   "source": [
    "#### (4) Spliting Training / Testing Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "42503355-eea3-4602-ae5a-ea18be84b8d1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:13.507023Z",
     "iopub.status.busy": "2021-07-19T07:15:13.506840Z",
     "iopub.status.idle": "2021-07-19T07:15:13.509743Z",
     "shell.execute_reply": "2021-07-19T07:15:13.509032Z",
     "shell.execute_reply.started": "2021-07-19T07:15:13.507002Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "data_labels = data.pork_bill.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "19235117-f63a-4e06-a102-6efe98c37dbf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:13.510605Z",
     "iopub.status.busy": "2021-07-19T07:15:13.510469Z",
     "iopub.status.idle": "2021-07-19T07:15:13.525158Z",
     "shell.execute_reply": "2021-07-19T07:15:13.524136Z",
     "shell.execute_reply.started": "2021-07-19T07:15:13.510587Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of X Train: (4852, 785) \n",
      "Shape of X Test : (2391, 785) \n",
      "Shape of Y Trian: (4852,) \n",
      "Shape of Y Test : (2391,)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "train_inputs, test_inputs, train_labels, test_labels = train_test_split(\n",
    "    data_inputs, data_labels, test_size=0.33, random_state=42)\n",
    "\n",
    "print(\"Shape of X Train:\", train_inputs.shape, '\\n'\n",
    "      \"Shape of X Test :\", test_inputs.shape,'\\n'\n",
    "      \"Shape of Y Trian:\", train_labels.shape , '\\n'\n",
    "      \"Shape of Y Test :\", test_labels.shape )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "725f7feb-b93b-4d0b-bcbf-562240a79f8b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:04:13.375644Z",
     "iopub.status.busy": "2021-05-16T17:04:13.375440Z",
     "iopub.status.idle": "2021-05-16T17:04:13.378111Z",
     "shell.execute_reply": "2021-05-16T17:04:13.377464Z",
     "shell.execute_reply.started": "2021-05-16T17:04:13.375623Z"
    },
    "tags": []
   },
   "source": [
    "-------------------------\n",
    "\n",
    "\n",
    "### Stage 3: Model and Building"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a0e74bf-eaf4-4043-8d23-88291d8ea344",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:13:18.807725Z",
     "iopub.status.busy": "2021-05-16T17:13:18.807498Z",
     "iopub.status.idle": "2021-05-16T17:13:18.810574Z",
     "shell.execute_reply": "2021-05-16T17:13:18.809625Z",
     "shell.execute_reply.started": "2021-05-16T17:13:18.807701Z"
    }
   },
   "source": [
    "#### (1) Using the Subclassing API to Build Dynamic Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "1f18f9b1-5f4d-4d0a-baf4-2eebed7f34f2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:13.527447Z",
     "iopub.status.busy": "2021-07-19T07:15:13.527137Z",
     "iopub.status.idle": "2021-07-19T07:15:13.537636Z",
     "shell.execute_reply": "2021-07-19T07:15:13.536910Z",
     "shell.execute_reply.started": "2021-07-19T07:15:13.527410Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "class DCNN(tf.keras.Model):\n",
    "    def __init__(self,\n",
    "                 vocab_size,\n",
    "                 emb_dim=128,\n",
    "                 nb_filters=100,\n",
    "                 # units: Positive integer, dimensionality of the output space.\n",
    "                 FFN_units=512,\n",
    "                 nb_classes=2,\n",
    "                 dropout_rate=0.1,\n",
    "                 training=False,\n",
    "                 name=\"PorkCNN\"):\n",
    "        super(DCNN, self).__init__(name=name)\n",
    "        self.embedding = layers.Embedding(vocab_size, emb_dim)\n",
    "        self.bigram = layers.Conv1D(filters=nb_filters, kernel_size=2, strides = 1, padding=\"valid\", activation=\"relu\")\n",
    "        self.bigram2 = layers.Conv1D(filters=nb_filters, kernel_size=2, strides = 2, padding=\"valid\", activation=\"relu\")\n",
    "        self.trigram = layers.Conv1D(filters=nb_filters,kernel_size=3,strides = 1, padding=\"valid\",activation=\"relu\")\n",
    "        self.trigram2 = layers.Conv1D(filters=nb_filters,kernel_size=3,strides = 2, padding=\"valid\",activation=\"relu\")\n",
    "        self.fourgram = layers.Conv1D(filters=nb_filters,kernel_size=4,strides = 2, padding=\"valid\",activation=\"relu\")\n",
    "        self.fivegram = layers.Conv1D(filters=nb_filters,kernel_size=5,strides = 2, padding=\"valid\",activation=\"relu\")\n",
    "        self.pool = layers.GlobalMaxPool1D()\n",
    "        self.dense_1 = layers.Dense(units=FFN_units, activation=\"relu\")\n",
    "        self.dropout = layers.Dropout(rate=dropout_rate)\n",
    "        self.last_dense = layers.Dense(units=1, activation=\"sigmoid\")    \n",
    "    def call(self, inputs, training):\n",
    "        x = self.embedding(inputs)\n",
    "        x_1 = self.bigram(x)\n",
    "        x_1 = self.pool(x_1)\n",
    "        x_1_1 = self.bigram2(x)\n",
    "        x_1_1 = self.pool(x_1_1)        \n",
    "        x_2 = self.trigram(x)\n",
    "        x_2 = self.pool(x_2)\n",
    "        x_2_1 = self.trigram2(x)\n",
    "        x_2_1 = self.pool(x_2_1)     \n",
    "        x_3 = self.fourgram(x)\n",
    "        x_3 = self.pool(x_3)\n",
    "        x_4 = self.fourgram(x)\n",
    "        x_4 = self.pool(x_4) \n",
    "        x_5 = self.fivegram(x)\n",
    "        x_5 = self.pool(x_5)      \n",
    "        merged = tf.concat([x_1,x_1_1, x_2,x_2_1, x_3, x_4, x_5], axis=-1) \n",
    "        merged = self.dense_1(merged)\n",
    "        merged = self.dropout(merged, training)\n",
    "        output = self.last_dense(merged)        \n",
    "        return output"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "bfc02e93-3a0e-4c1c-b488-b2db98b53824",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:42.908094Z",
     "iopub.status.busy": "2021-07-19T07:15:42.907820Z",
     "iopub.status.idle": "2021-07-19T07:15:42.929346Z",
     "shell.execute_reply": "2021-07-19T07:15:42.928723Z",
     "shell.execute_reply.started": "2021-07-19T07:15:42.908068Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "VOCAB_SIZE = tokenizer.vocab_size #tokenizer.vocab_size  # 5000 tokenizer.vocab_size\n",
    "EMB_DIM = 200\n",
    "NB_FILTERS = 100\n",
    "FFN_UNITS = 256\n",
    "NB_CLASSES = 2 #len(set(train_labels))\n",
    "DROPOUT_RATE = 0.25\n",
    "BATCH_SIZE = 230\n",
    "NB_EPOCHS = 1000\n",
    "\n",
    "Dcnn = DCNN(vocab_size=VOCAB_SIZE,\n",
    "            emb_dim=EMB_DIM,\n",
    "            nb_filters=NB_FILTERS,\n",
    "            FFN_units=FFN_UNITS,\n",
    "            nb_classes=NB_CLASSES,\n",
    "            dropout_rate=DROPOUT_RATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0ec8f51-6ed8-49fa-9d5e-c6929184bdf1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T18:43:59.279424Z",
     "iopub.status.busy": "2021-05-16T18:43:59.279202Z",
     "iopub.status.idle": "2021-05-16T18:43:59.282718Z",
     "shell.execute_reply": "2021-05-16T18:43:59.281823Z",
     "shell.execute_reply.started": "2021-05-16T18:43:59.279399Z"
    }
   },
   "source": [
    "#### (2) Compile and Summary of the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2cce0414-05eb-4f79-9c27-bca9c91a2903",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:44.485117Z",
     "iopub.status.busy": "2021-07-19T07:15:44.484859Z",
     "iopub.status.idle": "2021-07-19T07:15:44.558572Z",
     "shell.execute_reply": "2021-07-19T07:15:44.557871Z",
     "shell.execute_reply.started": "2021-07-19T07:15:44.485089Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"PorkCNN\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding (Embedding)        multiple                  586600    \n",
      "_________________________________________________________________\n",
      "conv1d (Conv1D)              multiple                  40100     \n",
      "_________________________________________________________________\n",
      "conv1d_1 (Conv1D)            multiple                  40100     \n",
      "_________________________________________________________________\n",
      "conv1d_2 (Conv1D)            multiple                  60100     \n",
      "_________________________________________________________________\n",
      "conv1d_3 (Conv1D)            multiple                  60100     \n",
      "_________________________________________________________________\n",
      "conv1d_4 (Conv1D)            multiple                  80100     \n",
      "_________________________________________________________________\n",
      "conv1d_5 (Conv1D)            multiple                  100100    \n",
      "_________________________________________________________________\n",
      "global_max_pooling1d (Global multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                multiple                  179456    \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            multiple                  0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              multiple                  257       \n",
      "=================================================================\n",
      "Total params: 1,146,913\n",
      "Trainable params: 1,146,913\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "Dcnn.compile(loss=\"binary_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n",
    "Dcnn.build(input_shape = (train_inputs.shape[1], EMB_DIM)) # (train_inputs.shape[1] , EMB_DIM)  (785 , EMB_DIM)\n",
    "Dcnn.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "73d70667-275f-40c0-bd71-6b47e9722798",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-05-16T17:47:13.442664Z",
     "iopub.status.busy": "2021-05-16T17:47:13.442381Z",
     "iopub.status.idle": "2021-05-16T17:47:13.448371Z",
     "shell.execute_reply": "2021-05-16T17:47:13.446970Z",
     "shell.execute_reply.started": "2021-05-16T17:47:13.442620Z"
    }
   },
   "source": [
    "-------------------------\n",
    "\n",
    "### Stage 4: Train the Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4351aebf-d66c-4d7e-8463-9b8a8cc45748",
   "metadata": {},
   "source": [
    "#### (1) Set up the Check-point Path"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2a42cc9c-52a3-48f7-be81-93fdc7caa44d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:46.949067Z",
     "iopub.status.busy": "2021-07-19T07:15:46.948876Z",
     "iopub.status.idle": "2021-07-19T07:15:46.952579Z",
     "shell.execute_reply": "2021-07-19T07:15:46.951980Z",
     "shell.execute_reply.started": "2021-07-19T07:15:46.949047Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "checkpoint_path = \"checkpoint_recode_tensorboard/ckpt_manager\"\n",
    "\n",
    "ckpt = tf.train.Checkpoint(Dcnn=Dcnn)\n",
    "\n",
    "ckpt_manager = tf.train.CheckpointManager(ckpt, checkpoint_path, max_to_keep=10)\n",
    "\n",
    "if ckpt_manager.latest_checkpoint:\n",
    "    ckpt.restore(ckpt_manager.latest_checkpoint)\n",
    "    print(\"Checkpoint Located!!\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3cdfd46a-e4ab-4ade-92b9-d7480829adc5",
   "metadata": {},
   "source": [
    "#### (2) Define the TensorBoard Callback\n",
    "reference: https://www.tensorflow.org/tensorboard/get_started"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d19eaa00-aab8-4b36-ab5b-356260f0012c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:47.748189Z",
     "iopub.status.busy": "2021-07-19T07:15:47.747972Z",
     "iopub.status.idle": "2021-07-19T07:15:47.751863Z",
     "shell.execute_reply": "2021-07-19T07:15:47.751214Z",
     "shell.execute_reply.started": "2021-07-19T07:15:47.748167Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import datetime\n",
    "log_dir = \"logs/fit/\" + datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\")\n",
    "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=log_dir, \n",
    "                                                      write_images=True,\n",
    "                                                      histogram_freq=1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6efdbe9a-f152-4c7b-a2b0-39243c31eeb4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-08T21:39:05.952891Z",
     "iopub.status.busy": "2021-07-08T21:39:05.952657Z",
     "iopub.status.idle": "2021-07-08T21:39:05.955902Z",
     "shell.execute_reply": "2021-07-08T21:39:05.954866Z",
     "shell.execute_reply.started": "2021-07-08T21:39:05.952872Z"
    }
   },
   "source": [
    "#### (3) Training the Model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "baf0cade-2c45-43b2-9782-6a8c6e570153",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T07:15:51.424569Z",
     "iopub.status.busy": "2021-07-19T07:15:51.424383Z",
     "iopub.status.idle": "2021-07-19T20:40:56.071733Z",
     "shell.execute_reply": "2021-07-19T20:40:56.069831Z",
     "shell.execute_reply.started": "2021-07-19T07:15:51.424550Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n",
      "22/22 [==============================] - 39s 2s/step - loss: 0.6029 - accuracy: 0.6641 - val_loss: 0.5091 - val_accuracy: 0.7227\n",
      "Epoch 2/1000\n",
      "22/22 [==============================] - 37s 2s/step - loss: 0.3850 - accuracy: 0.8333 - val_loss: 0.2724 - val_accuracy: 0.8929\n",
      "Epoch 3/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.1823 - accuracy: 0.9349 - val_loss: 0.1759 - val_accuracy: 0.9264\n",
      "Epoch 4/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0897 - accuracy: 0.9722 - val_loss: 0.1630 - val_accuracy: 0.9402\n",
      "Epoch 5/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0605 - accuracy: 0.9802 - val_loss: 0.1646 - val_accuracy: 0.9456\n",
      "Epoch 6/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0483 - accuracy: 0.9854 - val_loss: 0.1665 - val_accuracy: 0.9419\n",
      "Epoch 7/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0476 - accuracy: 0.9847 - val_loss: 0.1669 - val_accuracy: 0.9435\n",
      "Epoch 8/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0471 - accuracy: 0.9856 - val_loss: 0.1693 - val_accuracy: 0.9431\n",
      "Epoch 9/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0462 - accuracy: 0.9860 - val_loss: 0.2037 - val_accuracy: 0.9339\n",
      "Epoch 10/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0465 - accuracy: 0.9852 - val_loss: 0.1798 - val_accuracy: 0.9448\n",
      "Epoch 11/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0400 - accuracy: 0.9868 - val_loss: 0.1767 - val_accuracy: 0.9440\n",
      "Epoch 12/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0389 - accuracy: 0.9878 - val_loss: 0.1983 - val_accuracy: 0.9414\n",
      "Epoch 13/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0435 - accuracy: 0.9864 - val_loss: 0.2000 - val_accuracy: 0.9444\n",
      "Epoch 14/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0389 - accuracy: 0.9858 - val_loss: 0.1899 - val_accuracy: 0.9410\n",
      "Epoch 15/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0372 - accuracy: 0.9858 - val_loss: 0.1982 - val_accuracy: 0.9373\n",
      "Epoch 16/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0368 - accuracy: 0.9862 - val_loss: 0.1864 - val_accuracy: 0.9431\n",
      "Epoch 17/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0343 - accuracy: 0.9864 - val_loss: 0.1863 - val_accuracy: 0.9419\n",
      "Epoch 18/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0317 - accuracy: 0.9864 - val_loss: 0.1904 - val_accuracy: 0.9427\n",
      "Epoch 19/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0295 - accuracy: 0.9862 - val_loss: 0.1892 - val_accuracy: 0.9452\n",
      "Epoch 20/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0395 - accuracy: 0.9868 - val_loss: 0.1971 - val_accuracy: 0.9410\n",
      "Epoch 21/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0417 - accuracy: 0.9864 - val_loss: 0.1917 - val_accuracy: 0.9410\n",
      "Epoch 22/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0463 - accuracy: 0.9862 - val_loss: 0.1899 - val_accuracy: 0.9410\n",
      "Epoch 23/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0360 - accuracy: 0.9878 - val_loss: 0.1939 - val_accuracy: 0.9440\n",
      "Epoch 24/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0313 - accuracy: 0.9870 - val_loss: 0.1959 - val_accuracy: 0.9410\n",
      "Epoch 25/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0375 - accuracy: 0.9856 - val_loss: 0.1998 - val_accuracy: 0.9435\n",
      "Epoch 26/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0326 - accuracy: 0.9864 - val_loss: 0.1987 - val_accuracy: 0.9414\n",
      "Epoch 27/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0318 - accuracy: 0.9862 - val_loss: 0.2027 - val_accuracy: 0.9410\n",
      "Epoch 28/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0305 - accuracy: 0.9860 - val_loss: 0.2056 - val_accuracy: 0.9381\n",
      "Epoch 29/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0272 - accuracy: 0.9876 - val_loss: 0.2051 - val_accuracy: 0.9423\n",
      "Epoch 30/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0248 - accuracy: 0.9876 - val_loss: 0.2101 - val_accuracy: 0.9419\n",
      "Epoch 31/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0253 - accuracy: 0.9864 - val_loss: 0.2160 - val_accuracy: 0.9381\n",
      "Epoch 32/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0270 - accuracy: 0.9862 - val_loss: 0.2238 - val_accuracy: 0.9373\n",
      "Epoch 33/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0264 - accuracy: 0.9854 - val_loss: 0.2250 - val_accuracy: 0.9385\n",
      "Epoch 34/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0247 - accuracy: 0.9874 - val_loss: 0.2296 - val_accuracy: 0.9427\n",
      "Epoch 35/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0249 - accuracy: 0.9874 - val_loss: 0.2304 - val_accuracy: 0.9431\n",
      "Epoch 36/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0251 - accuracy: 0.9870 - val_loss: 0.2472 - val_accuracy: 0.9318\n",
      "Epoch 37/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0252 - accuracy: 0.9866 - val_loss: 0.2269 - val_accuracy: 0.9373\n",
      "Epoch 38/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0223 - accuracy: 0.9878 - val_loss: 0.2401 - val_accuracy: 0.9368\n",
      "Epoch 39/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0207 - accuracy: 0.9887 - val_loss: 0.2479 - val_accuracy: 0.9389\n",
      "Epoch 40/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0211 - accuracy: 0.9880 - val_loss: 0.2534 - val_accuracy: 0.9419\n",
      "Epoch 41/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0209 - accuracy: 0.9870 - val_loss: 0.2558 - val_accuracy: 0.9410\n",
      "Epoch 42/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0201 - accuracy: 0.9868 - val_loss: 0.2640 - val_accuracy: 0.9398\n",
      "Epoch 43/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0211 - accuracy: 0.9870 - val_loss: 0.2616 - val_accuracy: 0.9385\n",
      "Epoch 44/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0213 - accuracy: 0.9870 - val_loss: 0.2505 - val_accuracy: 0.9394\n",
      "Epoch 45/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0201 - accuracy: 0.9885 - val_loss: 0.2678 - val_accuracy: 0.9423\n",
      "Epoch 46/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0195 - accuracy: 0.9883 - val_loss: 0.2754 - val_accuracy: 0.9381\n",
      "Epoch 47/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0202 - accuracy: 0.9883 - val_loss: 0.2748 - val_accuracy: 0.9402\n",
      "Epoch 48/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0207 - accuracy: 0.9880 - val_loss: 0.2716 - val_accuracy: 0.9394\n",
      "Epoch 49/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0195 - accuracy: 0.9872 - val_loss: 0.2784 - val_accuracy: 0.9373\n",
      "Epoch 50/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0200 - accuracy: 0.9872 - val_loss: 0.2739 - val_accuracy: 0.9406\n",
      "Epoch 51/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0189 - accuracy: 0.9878 - val_loss: 0.2804 - val_accuracy: 0.9373\n",
      "Epoch 52/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0195 - accuracy: 0.9885 - val_loss: 0.2855 - val_accuracy: 0.9414\n",
      "Epoch 53/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0193 - accuracy: 0.9868 - val_loss: 0.2910 - val_accuracy: 0.9402\n",
      "Epoch 54/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0191 - accuracy: 0.9880 - val_loss: 0.2903 - val_accuracy: 0.9398\n",
      "Epoch 55/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0185 - accuracy: 0.9887 - val_loss: 0.2941 - val_accuracy: 0.9410\n",
      "Epoch 56/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0187 - accuracy: 0.9876 - val_loss: 0.2940 - val_accuracy: 0.9414\n",
      "Epoch 57/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0194 - accuracy: 0.9880 - val_loss: 0.2941 - val_accuracy: 0.9419\n",
      "Epoch 58/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0189 - accuracy: 0.9876 - val_loss: 0.2976 - val_accuracy: 0.9402\n",
      "Epoch 59/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0200 - accuracy: 0.9876 - val_loss: 0.2801 - val_accuracy: 0.9414\n",
      "Epoch 60/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0185 - accuracy: 0.9891 - val_loss: 0.2886 - val_accuracy: 0.9423\n",
      "Epoch 61/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0185 - accuracy: 0.9885 - val_loss: 0.2936 - val_accuracy: 0.9410\n",
      "Epoch 62/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0188 - accuracy: 0.9887 - val_loss: 0.3072 - val_accuracy: 0.9394\n",
      "Epoch 63/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0186 - accuracy: 0.9885 - val_loss: 0.3065 - val_accuracy: 0.9419\n",
      "Epoch 64/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 0.3101 - val_accuracy: 0.9419\n",
      "Epoch 65/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0183 - accuracy: 0.9899 - val_loss: 0.3180 - val_accuracy: 0.9410\n",
      "Epoch 66/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0180 - accuracy: 0.9895 - val_loss: 0.3171 - val_accuracy: 0.9419\n",
      "Epoch 67/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0188 - accuracy: 0.9889 - val_loss: 0.3140 - val_accuracy: 0.9423\n",
      "Epoch 68/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0182 - accuracy: 0.9887 - val_loss: 0.3165 - val_accuracy: 0.9414\n",
      "Epoch 69/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0185 - accuracy: 0.9872 - val_loss: 0.3194 - val_accuracy: 0.9402\n",
      "Epoch 70/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9880 - val_loss: 0.3263 - val_accuracy: 0.9414\n",
      "Epoch 71/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 0.3339 - val_accuracy: 0.9414\n",
      "Epoch 72/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0181 - accuracy: 0.9876 - val_loss: 0.3371 - val_accuracy: 0.9410\n",
      "Epoch 73/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0181 - accuracy: 0.9878 - val_loss: 0.3348 - val_accuracy: 0.9406\n",
      "Epoch 74/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.3407 - val_accuracy: 0.9414\n",
      "Epoch 75/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0181 - accuracy: 0.9878 - val_loss: 0.3433 - val_accuracy: 0.9419\n",
      "Epoch 76/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.3486 - val_accuracy: 0.9419\n",
      "Epoch 77/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.3462 - val_accuracy: 0.9394\n",
      "Epoch 78/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9895 - val_loss: 0.3305 - val_accuracy: 0.9373\n",
      "Epoch 79/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0180 - accuracy: 0.9887 - val_loss: 0.3337 - val_accuracy: 0.9389\n",
      "Epoch 80/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9889 - val_loss: 0.3418 - val_accuracy: 0.9394\n",
      "Epoch 81/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0184 - accuracy: 0.9887 - val_loss: 0.3480 - val_accuracy: 0.9394\n",
      "Epoch 82/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.3501 - val_accuracy: 0.9414\n",
      "Epoch 83/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.3513 - val_accuracy: 0.9402\n",
      "Epoch 84/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.3551 - val_accuracy: 0.9406\n",
      "Epoch 85/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.3535 - val_accuracy: 0.9414\n",
      "Epoch 86/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 0.3558 - val_accuracy: 0.9402\n",
      "Epoch 87/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.3623 - val_accuracy: 0.9402\n",
      "Epoch 88/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0179 - accuracy: 0.9874 - val_loss: 0.3668 - val_accuracy: 0.9414\n",
      "Epoch 89/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0181 - accuracy: 0.9889 - val_loss: 0.3593 - val_accuracy: 0.9402\n",
      "Epoch 90/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 0.3570 - val_accuracy: 0.9398\n",
      "Epoch 91/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.3618 - val_accuracy: 0.9402\n",
      "Epoch 92/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0181 - accuracy: 0.9895 - val_loss: 0.3689 - val_accuracy: 0.9398\n",
      "Epoch 93/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9895 - val_loss: 0.3577 - val_accuracy: 0.9402\n",
      "Epoch 94/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0183 - accuracy: 0.9893 - val_loss: 0.3550 - val_accuracy: 0.9385\n",
      "Epoch 95/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0188 - accuracy: 0.9872 - val_loss: 0.3501 - val_accuracy: 0.9348\n",
      "Epoch 96/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0183 - accuracy: 0.9893 - val_loss: 0.3483 - val_accuracy: 0.9368\n",
      "Epoch 97/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.3534 - val_accuracy: 0.9406\n",
      "Epoch 98/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9883 - val_loss: 0.3647 - val_accuracy: 0.9398\n",
      "Epoch 99/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9895 - val_loss: 0.3696 - val_accuracy: 0.9402\n",
      "Epoch 100/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 0.3647 - val_accuracy: 0.9368\n",
      "Epoch 101/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0180 - accuracy: 0.9893 - val_loss: 0.3706 - val_accuracy: 0.9394\n",
      "Epoch 102/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.3698 - val_accuracy: 0.9368\n",
      "Epoch 103/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.3738 - val_accuracy: 0.9398\n",
      "Epoch 104/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.3726 - val_accuracy: 0.9398\n",
      "Epoch 105/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.3800 - val_accuracy: 0.9389\n",
      "Epoch 106/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.3795 - val_accuracy: 0.9402\n",
      "Epoch 107/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.3878 - val_accuracy: 0.9406\n",
      "Epoch 108/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.3889 - val_accuracy: 0.9406\n",
      "Epoch 109/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 0.3780 - val_accuracy: 0.9402\n",
      "Epoch 110/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9887 - val_loss: 0.3720 - val_accuracy: 0.9373\n",
      "Epoch 111/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9885 - val_loss: 0.3610 - val_accuracy: 0.9402\n",
      "Epoch 112/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0179 - accuracy: 0.9891 - val_loss: 0.3642 - val_accuracy: 0.9410\n",
      "Epoch 113/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9893 - val_loss: 0.3714 - val_accuracy: 0.9368\n",
      "Epoch 114/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 0.3807 - val_accuracy: 0.9419\n",
      "Epoch 115/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.3849 - val_accuracy: 0.9419\n",
      "Epoch 116/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.3879 - val_accuracy: 0.9406\n",
      "Epoch 117/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 0.3902 - val_accuracy: 0.9423\n",
      "Epoch 118/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.3898 - val_accuracy: 0.9427\n",
      "Epoch 119/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 0.3967 - val_accuracy: 0.9419\n",
      "Epoch 120/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.3991 - val_accuracy: 0.9423\n",
      "Epoch 121/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9880 - val_loss: 0.3917 - val_accuracy: 0.9406\n",
      "Epoch 122/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.3971 - val_accuracy: 0.9410\n",
      "Epoch 123/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9893 - val_loss: 0.3757 - val_accuracy: 0.9435\n",
      "Epoch 124/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.3609 - val_accuracy: 0.9435\n",
      "Epoch 125/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.3665 - val_accuracy: 0.9402\n",
      "Epoch 126/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9891 - val_loss: 0.3715 - val_accuracy: 0.9389\n",
      "Epoch 127/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9876 - val_loss: 0.3769 - val_accuracy: 0.9410\n",
      "Epoch 128/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9891 - val_loss: 0.3615 - val_accuracy: 0.9410\n",
      "Epoch 129/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.3675 - val_accuracy: 0.9419\n",
      "Epoch 130/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.3838 - val_accuracy: 0.9398\n",
      "Epoch 131/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.3939 - val_accuracy: 0.9381\n",
      "Epoch 132/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9885 - val_loss: 0.3996 - val_accuracy: 0.9394\n",
      "Epoch 133/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9880 - val_loss: 0.4004 - val_accuracy: 0.9402\n",
      "Epoch 134/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9887 - val_loss: 0.3965 - val_accuracy: 0.9398\n",
      "Epoch 135/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.3920 - val_accuracy: 0.9381\n",
      "Epoch 136/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9883 - val_loss: 0.3960 - val_accuracy: 0.9406\n",
      "Epoch 137/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9889 - val_loss: 0.3974 - val_accuracy: 0.9402\n",
      "Epoch 138/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9878 - val_loss: 0.3788 - val_accuracy: 0.9398\n",
      "Epoch 139/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 0.3868 - val_accuracy: 0.9402\n",
      "Epoch 140/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9889 - val_loss: 0.3966 - val_accuracy: 0.9398\n",
      "Epoch 141/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.4031 - val_accuracy: 0.9398\n",
      "Epoch 142/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9901 - val_loss: 0.4024 - val_accuracy: 0.9406\n",
      "Epoch 143/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.3972 - val_accuracy: 0.9402\n",
      "Epoch 144/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.3962 - val_accuracy: 0.9402\n",
      "Epoch 145/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.4005 - val_accuracy: 0.9402\n",
      "Epoch 146/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 0.4065 - val_accuracy: 0.9402\n",
      "Epoch 147/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.4112 - val_accuracy: 0.9389\n",
      "Epoch 148/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.4173 - val_accuracy: 0.9381\n",
      "Epoch 149/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9895 - val_loss: 0.4134 - val_accuracy: 0.9406\n",
      "Epoch 150/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0184 - accuracy: 0.9889 - val_loss: 0.3967 - val_accuracy: 0.9360\n",
      "Epoch 151/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.4037 - val_accuracy: 0.9381\n",
      "Epoch 152/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0191 - accuracy: 0.9893 - val_loss: 0.3814 - val_accuracy: 0.9381\n",
      "Epoch 153/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0180 - accuracy: 0.9895 - val_loss: 0.3965 - val_accuracy: 0.9402\n",
      "Epoch 154/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.4062 - val_accuracy: 0.9377\n",
      "Epoch 155/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0212 - accuracy: 0.9891 - val_loss: 0.4364 - val_accuracy: 0.9251\n",
      "Epoch 156/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0184 - accuracy: 0.9878 - val_loss: 0.3915 - val_accuracy: 0.9356\n",
      "Epoch 157/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0192 - accuracy: 0.9880 - val_loss: 0.3968 - val_accuracy: 0.9322\n",
      "Epoch 158/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9893 - val_loss: 0.4154 - val_accuracy: 0.9339\n",
      "Epoch 159/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0203 - accuracy: 0.9897 - val_loss: 0.4237 - val_accuracy: 0.9260\n",
      "Epoch 160/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0199 - accuracy: 0.9878 - val_loss: 0.4879 - val_accuracy: 0.9251\n",
      "Epoch 161/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0200 - accuracy: 0.9876 - val_loss: 0.5240 - val_accuracy: 0.9205\n",
      "Epoch 162/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0225 - accuracy: 0.9878 - val_loss: 0.4085 - val_accuracy: 0.9414\n",
      "Epoch 163/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0198 - accuracy: 0.9885 - val_loss: 0.4627 - val_accuracy: 0.9318\n",
      "Epoch 164/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0193 - accuracy: 0.9889 - val_loss: 0.4412 - val_accuracy: 0.9377\n",
      "Epoch 165/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0184 - accuracy: 0.9897 - val_loss: 0.5249 - val_accuracy: 0.9335\n",
      "Epoch 166/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0185 - accuracy: 0.9891 - val_loss: 0.4874 - val_accuracy: 0.9385\n",
      "Epoch 167/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9899 - val_loss: 0.5055 - val_accuracy: 0.9385\n",
      "Epoch 168/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.5169 - val_accuracy: 0.9381\n",
      "Epoch 169/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0182 - accuracy: 0.9889 - val_loss: 0.5093 - val_accuracy: 0.9373\n",
      "Epoch 170/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.4962 - val_accuracy: 0.9402\n",
      "Epoch 171/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9880 - val_loss: 0.5035 - val_accuracy: 0.9394\n",
      "Epoch 172/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5143 - val_accuracy: 0.9394\n",
      "Epoch 173/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 0.5237 - val_accuracy: 0.9394\n",
      "Epoch 174/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.5239 - val_accuracy: 0.9385\n",
      "Epoch 175/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9885 - val_loss: 0.5270 - val_accuracy: 0.9385\n",
      "Epoch 176/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.5128 - val_accuracy: 0.9377\n",
      "Epoch 177/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5177 - val_accuracy: 0.9381\n",
      "Epoch 178/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9876 - val_loss: 0.5193 - val_accuracy: 0.9368\n",
      "Epoch 179/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9901 - val_loss: 0.5181 - val_accuracy: 0.9364\n",
      "Epoch 180/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5198 - val_accuracy: 0.9373\n",
      "Epoch 181/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.5259 - val_accuracy: 0.9364\n",
      "Epoch 182/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 0.5349 - val_accuracy: 0.9348\n",
      "Epoch 183/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.5346 - val_accuracy: 0.9360\n",
      "Epoch 184/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.5337 - val_accuracy: 0.9373\n",
      "Epoch 185/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.5415 - val_accuracy: 0.9373\n",
      "Epoch 186/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9876 - val_loss: 0.5422 - val_accuracy: 0.9356\n",
      "Epoch 187/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.5468 - val_accuracy: 0.9352\n",
      "Epoch 188/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.5383 - val_accuracy: 0.9364\n",
      "Epoch 189/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0171 - accuracy: 0.9897 - val_loss: 0.5281 - val_accuracy: 0.9364\n",
      "Epoch 190/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.5357 - val_accuracy: 0.9364\n",
      "Epoch 191/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.5353 - val_accuracy: 0.9360\n",
      "Epoch 192/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.5437 - val_accuracy: 0.9364\n",
      "Epoch 193/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.5467 - val_accuracy: 0.9364\n",
      "Epoch 194/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5476 - val_accuracy: 0.9360\n",
      "Epoch 195/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.5564 - val_accuracy: 0.9352\n",
      "Epoch 196/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.5428 - val_accuracy: 0.9360\n",
      "Epoch 197/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 0.5488 - val_accuracy: 0.9356\n",
      "Epoch 198/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9903 - val_loss: 0.5524 - val_accuracy: 0.9360\n",
      "Epoch 199/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.5541 - val_accuracy: 0.9360\n",
      "Epoch 200/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 0.5613 - val_accuracy: 0.9352\n",
      "Epoch 201/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9901 - val_loss: 0.5623 - val_accuracy: 0.9352\n",
      "Epoch 202/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5581 - val_accuracy: 0.9368\n",
      "Epoch 203/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.5626 - val_accuracy: 0.9360\n",
      "Epoch 204/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.5605 - val_accuracy: 0.9339\n",
      "Epoch 205/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9883 - val_loss: 0.5702 - val_accuracy: 0.9352\n",
      "Epoch 206/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.5703 - val_accuracy: 0.9352\n",
      "Epoch 207/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.5756 - val_accuracy: 0.9352\n",
      "Epoch 208/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 0.5770 - val_accuracy: 0.9348\n",
      "Epoch 209/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5788 - val_accuracy: 0.9348\n",
      "Epoch 210/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.5642 - val_accuracy: 0.9356\n",
      "Epoch 211/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.5594 - val_accuracy: 0.9352\n",
      "Epoch 212/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.5574 - val_accuracy: 0.9368\n",
      "Epoch 213/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.5599 - val_accuracy: 0.9364\n",
      "Epoch 214/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.5627 - val_accuracy: 0.9377\n",
      "Epoch 215/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5720 - val_accuracy: 0.9373\n",
      "Epoch 216/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.5779 - val_accuracy: 0.9373\n",
      "Epoch 217/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.5823 - val_accuracy: 0.9364\n",
      "Epoch 218/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.5756 - val_accuracy: 0.9373\n",
      "Epoch 219/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.5751 - val_accuracy: 0.9352\n",
      "Epoch 220/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.5808 - val_accuracy: 0.9356\n",
      "Epoch 221/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.5800 - val_accuracy: 0.9360\n",
      "Epoch 222/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.5847 - val_accuracy: 0.9327\n",
      "Epoch 223/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5756 - val_accuracy: 0.9352\n",
      "Epoch 224/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9885 - val_loss: 0.5765 - val_accuracy: 0.9360\n",
      "Epoch 225/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5816 - val_accuracy: 0.9356\n",
      "Epoch 226/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.5817 - val_accuracy: 0.9360\n",
      "Epoch 227/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.5607 - val_accuracy: 0.9373\n",
      "Epoch 228/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.5536 - val_accuracy: 0.9385\n",
      "Epoch 229/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.5570 - val_accuracy: 0.9364\n",
      "Epoch 230/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.5546 - val_accuracy: 0.9373\n",
      "Epoch 231/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5650 - val_accuracy: 0.9343\n",
      "Epoch 232/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5675 - val_accuracy: 0.9348\n",
      "Epoch 233/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.5705 - val_accuracy: 0.9348\n",
      "Epoch 234/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.5796 - val_accuracy: 0.9348\n",
      "Epoch 235/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5587 - val_accuracy: 0.9385\n",
      "Epoch 236/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.5537 - val_accuracy: 0.9360\n",
      "Epoch 237/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5609 - val_accuracy: 0.9348\n",
      "Epoch 238/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.5690 - val_accuracy: 0.9339\n",
      "Epoch 239/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9883 - val_loss: 0.5614 - val_accuracy: 0.9352\n",
      "Epoch 240/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.5585 - val_accuracy: 0.9352\n",
      "Epoch 241/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.5620 - val_accuracy: 0.9352\n",
      "Epoch 242/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.5627 - val_accuracy: 0.9348\n",
      "Epoch 243/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 0.5485 - val_accuracy: 0.9314\n",
      "Epoch 244/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0172 - accuracy: 0.9891 - val_loss: 0.5484 - val_accuracy: 0.9322\n",
      "Epoch 245/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.5562 - val_accuracy: 0.9310\n",
      "Epoch 246/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0181 - accuracy: 0.9887 - val_loss: 0.5651 - val_accuracy: 0.9335\n",
      "Epoch 247/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.5650 - val_accuracy: 0.9335\n",
      "Epoch 248/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.5662 - val_accuracy: 0.9348\n",
      "Epoch 249/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.5714 - val_accuracy: 0.9343\n",
      "Epoch 250/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0180 - accuracy: 0.9887 - val_loss: 0.5695 - val_accuracy: 0.9339\n",
      "Epoch 251/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 0.5662 - val_accuracy: 0.9339\n",
      "Epoch 252/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.5700 - val_accuracy: 0.9348\n",
      "Epoch 253/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.5735 - val_accuracy: 0.9352\n",
      "Epoch 254/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.5678 - val_accuracy: 0.9352\n",
      "Epoch 255/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.5717 - val_accuracy: 0.9364\n",
      "Epoch 256/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.5794 - val_accuracy: 0.9348\n",
      "Epoch 257/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 0.5886 - val_accuracy: 0.9352\n",
      "Epoch 258/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0177 - accuracy: 0.9880 - val_loss: 0.5826 - val_accuracy: 0.9364\n",
      "Epoch 259/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.5877 - val_accuracy: 0.9360\n",
      "Epoch 260/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5896 - val_accuracy: 0.9356\n",
      "Epoch 261/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.5827 - val_accuracy: 0.9377\n",
      "Epoch 262/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5605 - val_accuracy: 0.9381\n",
      "Epoch 263/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.5677 - val_accuracy: 0.9389\n",
      "Epoch 264/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.5726 - val_accuracy: 0.9385\n",
      "Epoch 265/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.5787 - val_accuracy: 0.9373\n",
      "Epoch 266/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9903 - val_loss: 0.5859 - val_accuracy: 0.9368\n",
      "Epoch 267/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.5839 - val_accuracy: 0.9373\n",
      "Epoch 268/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0177 - accuracy: 0.9899 - val_loss: 0.5826 - val_accuracy: 0.9352\n",
      "Epoch 269/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.5874 - val_accuracy: 0.9352\n",
      "Epoch 270/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.5902 - val_accuracy: 0.9368\n",
      "Epoch 271/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.5707 - val_accuracy: 0.9360\n",
      "Epoch 272/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.5778 - val_accuracy: 0.9377\n",
      "Epoch 273/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.5149 - val_accuracy: 0.9423\n",
      "Epoch 274/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.5567 - val_accuracy: 0.9394\n",
      "Epoch 275/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.5747 - val_accuracy: 0.9385\n",
      "Epoch 276/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.5807 - val_accuracy: 0.9389\n",
      "Epoch 277/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.5860 - val_accuracy: 0.9381\n",
      "Epoch 278/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.5876 - val_accuracy: 0.9377\n",
      "Epoch 279/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0190 - accuracy: 0.9899 - val_loss: 0.5558 - val_accuracy: 0.9268\n",
      "Epoch 280/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0184 - accuracy: 0.9885 - val_loss: 0.5811 - val_accuracy: 0.9256\n",
      "Epoch 281/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0228 - accuracy: 0.9874 - val_loss: 0.4844 - val_accuracy: 0.9348\n",
      "Epoch 282/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0253 - accuracy: 0.9862 - val_loss: 0.5005 - val_accuracy: 0.9348\n",
      "Epoch 283/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0274 - accuracy: 0.9868 - val_loss: 0.5072 - val_accuracy: 0.9331\n",
      "Epoch 284/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0305 - accuracy: 0.9850 - val_loss: 0.4708 - val_accuracy: 0.9264\n",
      "Epoch 285/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0234 - accuracy: 0.9874 - val_loss: 0.5280 - val_accuracy: 0.9394\n",
      "Epoch 286/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0187 - accuracy: 0.9887 - val_loss: 0.5812 - val_accuracy: 0.9402\n",
      "Epoch 287/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0182 - accuracy: 0.9891 - val_loss: 0.6324 - val_accuracy: 0.9285\n",
      "Epoch 288/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 0.5677 - val_accuracy: 0.9360\n",
      "Epoch 289/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.5975 - val_accuracy: 0.9343\n",
      "Epoch 290/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0179 - accuracy: 0.9885 - val_loss: 0.6122 - val_accuracy: 0.9356\n",
      "Epoch 291/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.6229 - val_accuracy: 0.9335\n",
      "Epoch 292/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 0.6294 - val_accuracy: 0.9331\n",
      "Epoch 293/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.6361 - val_accuracy: 0.9343\n",
      "Epoch 294/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 0.6442 - val_accuracy: 0.9348\n",
      "Epoch 295/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9880 - val_loss: 0.6429 - val_accuracy: 0.9348\n",
      "Epoch 296/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.6468 - val_accuracy: 0.9352\n",
      "Epoch 297/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.6610 - val_accuracy: 0.9368\n",
      "Epoch 298/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 0.6682 - val_accuracy: 0.9348\n",
      "Epoch 299/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.6637 - val_accuracy: 0.9385\n",
      "Epoch 300/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.6658 - val_accuracy: 0.9377\n",
      "Epoch 301/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 0.6578 - val_accuracy: 0.9377\n",
      "Epoch 302/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.6576 - val_accuracy: 0.9356\n",
      "Epoch 303/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9899 - val_loss: 0.6711 - val_accuracy: 0.9364\n",
      "Epoch 304/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.6688 - val_accuracy: 0.9368\n",
      "Epoch 305/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0183 - accuracy: 0.9897 - val_loss: 0.6336 - val_accuracy: 0.9377\n",
      "Epoch 306/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.5769 - val_accuracy: 0.9394\n",
      "Epoch 307/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.5872 - val_accuracy: 0.9389\n",
      "Epoch 308/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.6067 - val_accuracy: 0.9381\n",
      "Epoch 309/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.6179 - val_accuracy: 0.9368\n",
      "Epoch 310/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.6275 - val_accuracy: 0.9368\n",
      "Epoch 311/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.6324 - val_accuracy: 0.9368\n",
      "Epoch 312/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9901 - val_loss: 0.6359 - val_accuracy: 0.9368\n",
      "Epoch 313/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.6404 - val_accuracy: 0.9368\n",
      "Epoch 314/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.6527 - val_accuracy: 0.9377\n",
      "Epoch 315/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.6542 - val_accuracy: 0.9373\n",
      "Epoch 316/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.6604 - val_accuracy: 0.9368\n",
      "Epoch 317/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.6648 - val_accuracy: 0.9381\n",
      "Epoch 318/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.6629 - val_accuracy: 0.9381\n",
      "Epoch 319/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.6643 - val_accuracy: 0.9377\n",
      "Epoch 320/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0169 - accuracy: 0.9891 - val_loss: 0.6724 - val_accuracy: 0.9373\n",
      "Epoch 321/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 0.6696 - val_accuracy: 0.9373\n",
      "Epoch 322/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.6674 - val_accuracy: 0.9381\n",
      "Epoch 323/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.6689 - val_accuracy: 0.9385\n",
      "Epoch 324/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.6649 - val_accuracy: 0.9377\n",
      "Epoch 325/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.6692 - val_accuracy: 0.9385\n",
      "Epoch 326/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.6738 - val_accuracy: 0.9373\n",
      "Epoch 327/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.6789 - val_accuracy: 0.9385\n",
      "Epoch 328/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 0.6846 - val_accuracy: 0.9381\n",
      "Epoch 329/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.6881 - val_accuracy: 0.9368\n",
      "Epoch 330/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.6928 - val_accuracy: 0.9368\n",
      "Epoch 331/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.6946 - val_accuracy: 0.9381\n",
      "Epoch 332/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.6989 - val_accuracy: 0.9377\n",
      "Epoch 333/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.7037 - val_accuracy: 0.9364\n",
      "Epoch 334/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 0.6905 - val_accuracy: 0.9364\n",
      "Epoch 335/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0173 - accuracy: 0.9903 - val_loss: 0.6787 - val_accuracy: 0.9389\n",
      "Epoch 336/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.6785 - val_accuracy: 0.9394\n",
      "Epoch 337/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.6767 - val_accuracy: 0.9385\n",
      "Epoch 338/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 0.6812 - val_accuracy: 0.9389\n",
      "Epoch 339/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 0.6847 - val_accuracy: 0.9389\n",
      "Epoch 340/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.6904 - val_accuracy: 0.9394\n",
      "Epoch 341/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.6939 - val_accuracy: 0.9394\n",
      "Epoch 342/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.7010 - val_accuracy: 0.9385\n",
      "Epoch 343/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0171 - accuracy: 0.9897 - val_loss: 0.7010 - val_accuracy: 0.9381\n",
      "Epoch 344/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0172 - accuracy: 0.9891 - val_loss: 0.7079 - val_accuracy: 0.9394\n",
      "Epoch 345/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 0.7089 - val_accuracy: 0.9385\n",
      "Epoch 346/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.7052 - val_accuracy: 0.9385\n",
      "Epoch 347/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.7096 - val_accuracy: 0.9385\n",
      "Epoch 348/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.7141 - val_accuracy: 0.9385\n",
      "Epoch 349/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.7192 - val_accuracy: 0.9398\n",
      "Epoch 350/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.7126 - val_accuracy: 0.9398\n",
      "Epoch 351/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0233 - accuracy: 0.9903 - val_loss: 0.6613 - val_accuracy: 0.9285\n",
      "Epoch 352/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.5451 - val_accuracy: 0.9356\n",
      "Epoch 353/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0183 - accuracy: 0.9887 - val_loss: 0.5789 - val_accuracy: 0.9385\n",
      "Epoch 354/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0189 - accuracy: 0.9880 - val_loss: 0.6665 - val_accuracy: 0.9318\n",
      "Epoch 355/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0192 - accuracy: 0.9887 - val_loss: 0.5471 - val_accuracy: 0.9318\n",
      "Epoch 356/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0182 - accuracy: 0.9891 - val_loss: 0.6243 - val_accuracy: 0.9431\n",
      "Epoch 357/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0188 - accuracy: 0.9878 - val_loss: 0.6926 - val_accuracy: 0.9373\n",
      "Epoch 358/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0188 - accuracy: 0.9887 - val_loss: 0.7663 - val_accuracy: 0.9289\n",
      "Epoch 359/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0220 - accuracy: 0.9876 - val_loss: 0.6584 - val_accuracy: 0.9364\n",
      "Epoch 360/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0276 - accuracy: 0.9866 - val_loss: 0.6351 - val_accuracy: 0.9247\n",
      "Epoch 361/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0206 - accuracy: 0.9887 - val_loss: 0.6519 - val_accuracy: 0.9276\n",
      "Epoch 362/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0193 - accuracy: 0.9880 - val_loss: 0.6149 - val_accuracy: 0.9335\n",
      "Epoch 363/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0192 - accuracy: 0.9891 - val_loss: 0.5852 - val_accuracy: 0.9360\n",
      "Epoch 364/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0185 - accuracy: 0.9883 - val_loss: 0.6128 - val_accuracy: 0.9364\n",
      "Epoch 365/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0182 - accuracy: 0.9893 - val_loss: 0.6514 - val_accuracy: 0.9360\n",
      "Epoch 366/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.6668 - val_accuracy: 0.9343\n",
      "Epoch 367/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.6891 - val_accuracy: 0.9364\n",
      "Epoch 368/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0170 - accuracy: 0.9893 - val_loss: 0.7005 - val_accuracy: 0.9364\n",
      "Epoch 369/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0176 - accuracy: 0.9883 - val_loss: 0.7085 - val_accuracy: 0.9364\n",
      "Epoch 370/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0179 - accuracy: 0.9901 - val_loss: 0.7141 - val_accuracy: 0.9356\n",
      "Epoch 371/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.7233 - val_accuracy: 0.9364\n",
      "Epoch 372/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0181 - accuracy: 0.9893 - val_loss: 0.7136 - val_accuracy: 0.9368\n",
      "Epoch 373/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0181 - accuracy: 0.9901 - val_loss: 0.6942 - val_accuracy: 0.9331\n",
      "Epoch 374/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.6964 - val_accuracy: 0.9335\n",
      "Epoch 375/1000\n",
      "22/22 [==============================] - 39s 2s/step - loss: 0.0171 - accuracy: 0.9883 - val_loss: 0.7107 - val_accuracy: 0.9331\n",
      "Epoch 376/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.7149 - val_accuracy: 0.9373\n",
      "Epoch 377/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9883 - val_loss: 0.7099 - val_accuracy: 0.9398\n",
      "Epoch 378/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7174 - val_accuracy: 0.9377\n",
      "Epoch 379/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.7259 - val_accuracy: 0.9352\n",
      "Epoch 380/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.7312 - val_accuracy: 0.9352\n",
      "Epoch 381/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.7309 - val_accuracy: 0.9402\n",
      "Epoch 382/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9901 - val_loss: 0.7371 - val_accuracy: 0.9360\n",
      "Epoch 383/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.7381 - val_accuracy: 0.9381\n",
      "Epoch 384/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.7427 - val_accuracy: 0.9356\n",
      "Epoch 385/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 0.7477 - val_accuracy: 0.9352\n",
      "Epoch 386/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.7450 - val_accuracy: 0.9364\n",
      "Epoch 387/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.7402 - val_accuracy: 0.9373\n",
      "Epoch 388/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9883 - val_loss: 0.7394 - val_accuracy: 0.9377\n",
      "Epoch 389/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.7448 - val_accuracy: 0.9373\n",
      "Epoch 390/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.7478 - val_accuracy: 0.9377\n",
      "Epoch 391/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.7480 - val_accuracy: 0.9377\n",
      "Epoch 392/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.7547 - val_accuracy: 0.9377\n",
      "Epoch 393/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7536 - val_accuracy: 0.9402\n",
      "Epoch 394/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7591 - val_accuracy: 0.9394\n",
      "Epoch 395/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.7584 - val_accuracy: 0.9352\n",
      "Epoch 396/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.7619 - val_accuracy: 0.9373\n",
      "Epoch 397/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.7648 - val_accuracy: 0.9364\n",
      "Epoch 398/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.7665 - val_accuracy: 0.9364\n",
      "Epoch 399/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7833 - val_accuracy: 0.9331\n",
      "Epoch 400/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9887 - val_loss: 0.7818 - val_accuracy: 0.9352\n",
      "Epoch 401/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9903 - val_loss: 0.7828 - val_accuracy: 0.9343\n",
      "Epoch 402/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 0.7859 - val_accuracy: 0.9352\n",
      "Epoch 403/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.7916 - val_accuracy: 0.9335\n",
      "Epoch 404/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.7937 - val_accuracy: 0.9348\n",
      "Epoch 405/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7923 - val_accuracy: 0.9364\n",
      "Epoch 406/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9905 - val_loss: 0.7913 - val_accuracy: 0.9360\n",
      "Epoch 407/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0178 - accuracy: 0.9899 - val_loss: 0.7895 - val_accuracy: 0.9343\n",
      "Epoch 408/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9883 - val_loss: 0.8013 - val_accuracy: 0.9352\n",
      "Epoch 409/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.7978 - val_accuracy: 0.9368\n",
      "Epoch 410/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.7936 - val_accuracy: 0.9373\n",
      "Epoch 411/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 0.7989 - val_accuracy: 0.9368\n",
      "Epoch 412/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.7998 - val_accuracy: 0.9368\n",
      "Epoch 413/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 0.8032 - val_accuracy: 0.9368\n",
      "Epoch 414/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.8078 - val_accuracy: 0.9368\n",
      "Epoch 415/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.8072 - val_accuracy: 0.9356\n",
      "Epoch 416/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 0.8114 - val_accuracy: 0.9352\n",
      "Epoch 417/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.8101 - val_accuracy: 0.9348\n",
      "Epoch 418/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9880 - val_loss: 0.8108 - val_accuracy: 0.9343\n",
      "Epoch 419/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.8163 - val_accuracy: 0.9348\n",
      "Epoch 420/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.8199 - val_accuracy: 0.9348\n",
      "Epoch 421/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.8209 - val_accuracy: 0.9364\n",
      "Epoch 422/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0172 - accuracy: 0.9907 - val_loss: 0.8238 - val_accuracy: 0.9352\n",
      "Epoch 423/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 0.8186 - val_accuracy: 0.9377\n",
      "Epoch 424/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.8235 - val_accuracy: 0.9364\n",
      "Epoch 425/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9885 - val_loss: 0.8242 - val_accuracy: 0.9368\n",
      "Epoch 426/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.8281 - val_accuracy: 0.9360\n",
      "Epoch 427/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.8316 - val_accuracy: 0.9364\n",
      "Epoch 428/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.8383 - val_accuracy: 0.9360\n",
      "Epoch 429/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.8363 - val_accuracy: 0.9364\n",
      "Epoch 430/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.8380 - val_accuracy: 0.9348\n",
      "Epoch 431/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.8398 - val_accuracy: 0.9364\n",
      "Epoch 432/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.8349 - val_accuracy: 0.9364\n",
      "Epoch 433/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9901 - val_loss: 0.8309 - val_accuracy: 0.9352\n",
      "Epoch 434/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9901 - val_loss: 0.8277 - val_accuracy: 0.9360\n",
      "Epoch 435/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.8266 - val_accuracy: 0.9377\n",
      "Epoch 436/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.8294 - val_accuracy: 0.9373\n",
      "Epoch 437/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.8336 - val_accuracy: 0.9348\n",
      "Epoch 438/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.8324 - val_accuracy: 0.9385\n",
      "Epoch 439/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.8307 - val_accuracy: 0.9389\n",
      "Epoch 440/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9907 - val_loss: 0.8314 - val_accuracy: 0.9377\n",
      "Epoch 441/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.8337 - val_accuracy: 0.9373\n",
      "Epoch 442/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9874 - val_loss: 0.8370 - val_accuracy: 0.9381\n",
      "Epoch 443/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9909 - val_loss: 0.8438 - val_accuracy: 0.9348\n",
      "Epoch 444/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 0.8466 - val_accuracy: 0.9356\n",
      "Epoch 445/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.8516 - val_accuracy: 0.9356\n",
      "Epoch 446/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.8527 - val_accuracy: 0.9352\n",
      "Epoch 447/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.8462 - val_accuracy: 0.9356\n",
      "Epoch 448/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9901 - val_loss: 0.8560 - val_accuracy: 0.9343\n",
      "Epoch 449/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 0.8572 - val_accuracy: 0.9356\n",
      "Epoch 450/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.8580 - val_accuracy: 0.9352\n",
      "Epoch 451/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.8524 - val_accuracy: 0.9356\n",
      "Epoch 452/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.8435 - val_accuracy: 0.9368\n",
      "Epoch 453/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0171 - accuracy: 0.9885 - val_loss: 0.8402 - val_accuracy: 0.9381\n",
      "Epoch 454/1000\n",
      "22/22 [==============================] - 72s 3s/step - loss: 0.0177 - accuracy: 0.9903 - val_loss: 0.8370 - val_accuracy: 0.9381\n",
      "Epoch 455/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.8349 - val_accuracy: 0.9381\n",
      "Epoch 456/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0178 - accuracy: 0.9901 - val_loss: 0.8433 - val_accuracy: 0.9373\n",
      "Epoch 457/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.8408 - val_accuracy: 0.9364\n",
      "Epoch 458/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.8412 - val_accuracy: 0.9356\n",
      "Epoch 459/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 0.8486 - val_accuracy: 0.9364\n",
      "Epoch 460/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0178 - accuracy: 0.9899 - val_loss: 0.8552 - val_accuracy: 0.9352\n",
      "Epoch 461/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.8614 - val_accuracy: 0.9348\n",
      "Epoch 462/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.8591 - val_accuracy: 0.9377\n",
      "Epoch 463/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.8524 - val_accuracy: 0.9368\n",
      "Epoch 464/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0172 - accuracy: 0.9889 - val_loss: 0.8511 - val_accuracy: 0.9368\n",
      "Epoch 465/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.8609 - val_accuracy: 0.9364\n",
      "Epoch 466/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 0.8618 - val_accuracy: 0.9373\n",
      "Epoch 467/1000\n",
      "22/22 [==============================] - 76s 3s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.8649 - val_accuracy: 0.9368\n",
      "Epoch 468/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 0.8664 - val_accuracy: 0.9377\n",
      "Epoch 469/1000\n",
      "22/22 [==============================] - 89s 4s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.8615 - val_accuracy: 0.9360\n",
      "Epoch 470/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0175 - accuracy: 0.9885 - val_loss: 0.8721 - val_accuracy: 0.9356\n",
      "Epoch 471/1000\n",
      "22/22 [==============================] - 66s 3s/step - loss: 0.0177 - accuracy: 0.9899 - val_loss: 0.8741 - val_accuracy: 0.9360\n",
      "Epoch 472/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.8716 - val_accuracy: 0.9364\n",
      "Epoch 473/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.8671 - val_accuracy: 0.9368\n",
      "Epoch 474/1000\n",
      "22/22 [==============================] - 76s 3s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 0.8721 - val_accuracy: 0.9373\n",
      "Epoch 475/1000\n",
      "22/22 [==============================] - 80s 4s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.8758 - val_accuracy: 0.9360\n",
      "Epoch 476/1000\n",
      "22/22 [==============================] - 68s 3s/step - loss: 0.0182 - accuracy: 0.9901 - val_loss: 0.8450 - val_accuracy: 0.9318\n",
      "Epoch 477/1000\n",
      "22/22 [==============================] - 98s 5s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.7947 - val_accuracy: 0.9302\n",
      "Epoch 478/1000\n",
      "22/22 [==============================] - 90s 4s/step - loss: 0.0177 - accuracy: 0.9885 - val_loss: 0.7877 - val_accuracy: 0.9297\n",
      "Epoch 479/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 0.7362 - val_accuracy: 0.9368\n",
      "Epoch 480/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.7060 - val_accuracy: 0.9364\n",
      "Epoch 481/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9883 - val_loss: 0.7206 - val_accuracy: 0.9368\n",
      "Epoch 482/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.7353 - val_accuracy: 0.9352\n",
      "Epoch 483/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9885 - val_loss: 0.7430 - val_accuracy: 0.9364\n",
      "Epoch 484/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9905 - val_loss: 0.7503 - val_accuracy: 0.9364\n",
      "Epoch 485/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.7598 - val_accuracy: 0.9360\n",
      "Epoch 486/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 0.7685 - val_accuracy: 0.9356\n",
      "Epoch 487/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.7776 - val_accuracy: 0.9360\n",
      "Epoch 488/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.7860 - val_accuracy: 0.9360\n",
      "Epoch 489/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 0.7936 - val_accuracy: 0.9364\n",
      "Epoch 490/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 0.7956 - val_accuracy: 0.9360\n",
      "Epoch 491/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.7993 - val_accuracy: 0.9360\n",
      "Epoch 492/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 0.8094 - val_accuracy: 0.9364\n",
      "Epoch 493/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 0.8190 - val_accuracy: 0.9364\n",
      "Epoch 494/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9885 - val_loss: 0.8389 - val_accuracy: 0.9356\n",
      "Epoch 495/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.8206 - val_accuracy: 0.9377\n",
      "Epoch 496/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9885 - val_loss: 0.8261 - val_accuracy: 0.9360\n",
      "Epoch 497/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.8277 - val_accuracy: 0.9360\n",
      "Epoch 498/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 0.8254 - val_accuracy: 0.9373\n",
      "Epoch 499/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 0.8287 - val_accuracy: 0.9352\n",
      "Epoch 500/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.8281 - val_accuracy: 0.9352\n",
      "Epoch 501/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.8497 - val_accuracy: 0.9348\n",
      "Epoch 502/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.8504 - val_accuracy: 0.9348\n",
      "Epoch 503/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.8483 - val_accuracy: 0.9352\n",
      "Epoch 504/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.8471 - val_accuracy: 0.9348\n",
      "Epoch 505/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.8492 - val_accuracy: 0.9348\n",
      "Epoch 506/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9901 - val_loss: 0.8496 - val_accuracy: 0.9339\n",
      "Epoch 507/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.8510 - val_accuracy: 0.9339\n",
      "Epoch 508/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.8578 - val_accuracy: 0.9335\n",
      "Epoch 509/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.8550 - val_accuracy: 0.9339\n",
      "Epoch 510/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0172 - accuracy: 0.9891 - val_loss: 0.8513 - val_accuracy: 0.9343\n",
      "Epoch 511/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.8505 - val_accuracy: 0.9339\n",
      "Epoch 512/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.8468 - val_accuracy: 0.9343\n",
      "Epoch 513/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0172 - accuracy: 0.9891 - val_loss: 0.8510 - val_accuracy: 0.9335\n",
      "Epoch 514/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.8554 - val_accuracy: 0.9343\n",
      "Epoch 515/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.8669 - val_accuracy: 0.9343\n",
      "Epoch 516/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.8634 - val_accuracy: 0.9331\n",
      "Epoch 517/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 0.8652 - val_accuracy: 0.9331\n",
      "Epoch 518/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0184 - accuracy: 0.9893 - val_loss: 0.7846 - val_accuracy: 0.9218\n",
      "Epoch 519/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0222 - accuracy: 0.9889 - val_loss: 0.6210 - val_accuracy: 0.9360\n",
      "Epoch 520/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0184 - accuracy: 0.9889 - val_loss: 0.6785 - val_accuracy: 0.9373\n",
      "Epoch 521/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0181 - accuracy: 0.9883 - val_loss: 0.6895 - val_accuracy: 0.9343\n",
      "Epoch 522/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.6678 - val_accuracy: 0.9360\n",
      "Epoch 523/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.7017 - val_accuracy: 0.9352\n",
      "Epoch 524/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 0.7313 - val_accuracy: 0.9364\n",
      "Epoch 525/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0180 - accuracy: 0.9893 - val_loss: 0.7172 - val_accuracy: 0.9356\n",
      "Epoch 526/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.7321 - val_accuracy: 0.9377\n",
      "Epoch 527/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.7452 - val_accuracy: 0.9373\n",
      "Epoch 528/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 0.7674 - val_accuracy: 0.9381\n",
      "Epoch 529/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.7801 - val_accuracy: 0.9364\n",
      "Epoch 530/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 0.7775 - val_accuracy: 0.9364\n",
      "Epoch 531/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0178 - accuracy: 0.9880 - val_loss: 0.7793 - val_accuracy: 0.9368\n",
      "Epoch 532/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.7843 - val_accuracy: 0.9364\n",
      "Epoch 533/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0181 - accuracy: 0.9885 - val_loss: 0.7893 - val_accuracy: 0.9377\n",
      "Epoch 534/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.7931 - val_accuracy: 0.9377\n",
      "Epoch 535/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0177 - accuracy: 0.9901 - val_loss: 0.8094 - val_accuracy: 0.9373\n",
      "Epoch 536/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.8563 - val_accuracy: 0.9327\n",
      "Epoch 537/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0174 - accuracy: 0.9885 - val_loss: 0.7944 - val_accuracy: 0.9373\n",
      "Epoch 538/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 0.6891 - val_accuracy: 0.9377\n",
      "Epoch 539/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0198 - accuracy: 0.9885 - val_loss: 0.6732 - val_accuracy: 0.9368\n",
      "Epoch 540/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0206 - accuracy: 0.9889 - val_loss: 0.7736 - val_accuracy: 0.9189\n",
      "Epoch 541/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0217 - accuracy: 0.9885 - val_loss: 0.7818 - val_accuracy: 0.9377\n",
      "Epoch 542/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0198 - accuracy: 0.9883 - val_loss: 0.6946 - val_accuracy: 0.9385\n",
      "Epoch 543/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0232 - accuracy: 0.9874 - val_loss: 0.7530 - val_accuracy: 0.9293\n",
      "Epoch 544/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0200 - accuracy: 0.9895 - val_loss: 0.7328 - val_accuracy: 0.9318\n",
      "Epoch 545/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0185 - accuracy: 0.9889 - val_loss: 0.8641 - val_accuracy: 0.9306\n",
      "Epoch 546/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0187 - accuracy: 0.9891 - val_loss: 0.7227 - val_accuracy: 0.9310\n",
      "Epoch 547/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0180 - accuracy: 0.9889 - val_loss: 0.8339 - val_accuracy: 0.9302\n",
      "Epoch 548/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0180 - accuracy: 0.9897 - val_loss: 0.8704 - val_accuracy: 0.9281\n",
      "Epoch 549/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.8758 - val_accuracy: 0.9289\n",
      "Epoch 550/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0180 - accuracy: 0.9893 - val_loss: 0.8938 - val_accuracy: 0.9276\n",
      "Epoch 551/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.9013 - val_accuracy: 0.9276\n",
      "Epoch 552/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0177 - accuracy: 0.9899 - val_loss: 0.9068 - val_accuracy: 0.9272\n",
      "Epoch 553/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 0.9136 - val_accuracy: 0.9285\n",
      "Epoch 554/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.9123 - val_accuracy: 0.9302\n",
      "Epoch 555/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 0.9117 - val_accuracy: 0.9302\n",
      "Epoch 556/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.9125 - val_accuracy: 0.9297\n",
      "Epoch 557/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 0.9238 - val_accuracy: 0.9306\n",
      "Epoch 558/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0180 - accuracy: 0.9885 - val_loss: 0.9347 - val_accuracy: 0.9297\n",
      "Epoch 559/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.9351 - val_accuracy: 0.9293\n",
      "Epoch 560/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.9405 - val_accuracy: 0.9289\n",
      "Epoch 561/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 0.9444 - val_accuracy: 0.9293\n",
      "Epoch 562/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 0.9451 - val_accuracy: 0.9285\n",
      "Epoch 563/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0173 - accuracy: 0.9887 - val_loss: 0.9494 - val_accuracy: 0.9293\n",
      "Epoch 564/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0179 - accuracy: 0.9883 - val_loss: 0.9441 - val_accuracy: 0.9289\n",
      "Epoch 565/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0180 - accuracy: 0.9883 - val_loss: 0.9468 - val_accuracy: 0.9310\n",
      "Epoch 566/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 0.9594 - val_accuracy: 0.9297\n",
      "Epoch 567/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.9610 - val_accuracy: 0.9306\n",
      "Epoch 568/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 0.9655 - val_accuracy: 0.9306\n",
      "Epoch 569/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 0.9681 - val_accuracy: 0.9310\n",
      "Epoch 570/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 0.9708 - val_accuracy: 0.9310\n",
      "Epoch 571/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.9700 - val_accuracy: 0.9306\n",
      "Epoch 572/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.9696 - val_accuracy: 0.9297\n",
      "Epoch 573/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 0.9749 - val_accuracy: 0.9306\n",
      "Epoch 574/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 0.9793 - val_accuracy: 0.9306\n",
      "Epoch 575/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0182 - accuracy: 0.9899 - val_loss: 0.9788 - val_accuracy: 0.9310\n",
      "Epoch 576/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.9816 - val_accuracy: 0.9310\n",
      "Epoch 577/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.9885 - val_accuracy: 0.9318\n",
      "Epoch 578/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.9988 - val_accuracy: 0.9314\n",
      "Epoch 579/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 1.0022 - val_accuracy: 0.9297\n",
      "Epoch 580/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.0026 - val_accuracy: 0.9297\n",
      "Epoch 581/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.0014 - val_accuracy: 0.9314\n",
      "Epoch 582/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.0074 - val_accuracy: 0.9306\n",
      "Epoch 583/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.0066 - val_accuracy: 0.9297\n",
      "Epoch 584/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.0084 - val_accuracy: 0.9297\n",
      "Epoch 585/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.0080 - val_accuracy: 0.9306\n",
      "Epoch 586/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 1.0130 - val_accuracy: 0.9314\n",
      "Epoch 587/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.0172 - val_accuracy: 0.9310\n",
      "Epoch 588/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 1.0157 - val_accuracy: 0.9310\n",
      "Epoch 589/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 1.0145 - val_accuracy: 0.9314\n",
      "Epoch 590/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0175 - accuracy: 0.9885 - val_loss: 1.0132 - val_accuracy: 0.9322\n",
      "Epoch 591/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0179 - accuracy: 0.9899 - val_loss: 1.0124 - val_accuracy: 0.9314\n",
      "Epoch 592/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.0089 - val_accuracy: 0.9306\n",
      "Epoch 593/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.0129 - val_accuracy: 0.9314\n",
      "Epoch 594/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 1.0216 - val_accuracy: 0.9297\n",
      "Epoch 595/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 1.0244 - val_accuracy: 0.9306\n",
      "Epoch 596/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.0314 - val_accuracy: 0.9302\n",
      "Epoch 597/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.0305 - val_accuracy: 0.9289\n",
      "Epoch 598/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.0280 - val_accuracy: 0.9293\n",
      "Epoch 599/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 1.0250 - val_accuracy: 0.9297\n",
      "Epoch 600/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9880 - val_loss: 1.0261 - val_accuracy: 0.9293\n",
      "Epoch 601/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.0294 - val_accuracy: 0.9289\n",
      "Epoch 602/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 1.0333 - val_accuracy: 0.9276\n",
      "Epoch 603/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0331 - val_accuracy: 0.9302\n",
      "Epoch 604/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0407 - val_accuracy: 0.9281\n",
      "Epoch 605/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0388 - val_accuracy: 0.9314\n",
      "Epoch 606/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0425 - val_accuracy: 0.9285\n",
      "Epoch 607/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.0429 - val_accuracy: 0.9293\n",
      "Epoch 608/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 1.0421 - val_accuracy: 0.9302\n",
      "Epoch 609/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.0418 - val_accuracy: 0.9297\n",
      "Epoch 610/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.0381 - val_accuracy: 0.9297\n",
      "Epoch 611/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.0493 - val_accuracy: 0.9297\n",
      "Epoch 612/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0178 - accuracy: 0.9889 - val_loss: 1.0523 - val_accuracy: 0.9293\n",
      "Epoch 613/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 1.0511 - val_accuracy: 0.9306\n",
      "Epoch 614/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.0573 - val_accuracy: 0.9297\n",
      "Epoch 615/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.0628 - val_accuracy: 0.9297\n",
      "Epoch 616/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.0643 - val_accuracy: 0.9302\n",
      "Epoch 617/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.0732 - val_accuracy: 0.9293\n",
      "Epoch 618/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.0740 - val_accuracy: 0.9297\n",
      "Epoch 619/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 1.0715 - val_accuracy: 0.9297\n",
      "Epoch 620/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.0733 - val_accuracy: 0.9297\n",
      "Epoch 621/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0175 - accuracy: 0.9901 - val_loss: 1.0640 - val_accuracy: 0.9318\n",
      "Epoch 622/1000\n",
      "22/22 [==============================] - 77s 4s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 1.0653 - val_accuracy: 0.9297\n",
      "Epoch 623/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.0724 - val_accuracy: 0.9310\n",
      "Epoch 624/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 1.0759 - val_accuracy: 0.9302\n",
      "Epoch 625/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 1.0751 - val_accuracy: 0.9306\n",
      "Epoch 626/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 1.0822 - val_accuracy: 0.9302\n",
      "Epoch 627/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.0767 - val_accuracy: 0.9306\n",
      "Epoch 628/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.0805 - val_accuracy: 0.9306\n",
      "Epoch 629/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0181 - accuracy: 0.9891 - val_loss: 1.0471 - val_accuracy: 0.9239\n",
      "Epoch 630/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 0.9603 - val_accuracy: 0.9251\n",
      "Epoch 631/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.9634 - val_accuracy: 0.9302\n",
      "Epoch 632/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 0.9712 - val_accuracy: 0.9310\n",
      "Epoch 633/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 0.9861 - val_accuracy: 0.9318\n",
      "Epoch 634/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0190 - accuracy: 0.9895 - val_loss: 0.9092 - val_accuracy: 0.9239\n",
      "Epoch 635/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 0.9133 - val_accuracy: 0.9285\n",
      "Epoch 636/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0180 - accuracy: 0.9885 - val_loss: 0.9352 - val_accuracy: 0.9310\n",
      "Epoch 637/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 0.9451 - val_accuracy: 0.9297\n",
      "Epoch 638/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.9460 - val_accuracy: 0.9297\n",
      "Epoch 639/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 0.9516 - val_accuracy: 0.9293\n",
      "Epoch 640/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.9626 - val_accuracy: 0.9289\n",
      "Epoch 641/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.9614 - val_accuracy: 0.9289\n",
      "Epoch 642/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.9735 - val_accuracy: 0.9293\n",
      "Epoch 643/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 0.9776 - val_accuracy: 0.9310\n",
      "Epoch 644/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.9782 - val_accuracy: 0.9306\n",
      "Epoch 645/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 0.9828 - val_accuracy: 0.9293\n",
      "Epoch 646/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0175 - accuracy: 0.9880 - val_loss: 0.9909 - val_accuracy: 0.9285\n",
      "Epoch 647/1000\n",
      "22/22 [==============================] - 75s 3s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 0.9915 - val_accuracy: 0.9293\n",
      "Epoch 648/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0173 - accuracy: 0.9899 - val_loss: 0.9929 - val_accuracy: 0.9322\n",
      "Epoch 649/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 0.9919 - val_accuracy: 0.9335\n",
      "Epoch 650/1000\n",
      "22/22 [==============================] - 67s 3s/step - loss: 0.0230 - accuracy: 0.9901 - val_loss: 0.9290 - val_accuracy: 0.9210\n",
      "Epoch 651/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0175 - accuracy: 0.9901 - val_loss: 0.8966 - val_accuracy: 0.9335\n",
      "Epoch 652/1000\n",
      "22/22 [==============================] - 87s 4s/step - loss: 0.0200 - accuracy: 0.9887 - val_loss: 0.6594 - val_accuracy: 0.9368\n",
      "Epoch 653/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0195 - accuracy: 0.9883 - val_loss: 0.6798 - val_accuracy: 0.9368\n",
      "Epoch 654/1000\n",
      "22/22 [==============================] - 66s 3s/step - loss: 0.0187 - accuracy: 0.9891 - val_loss: 0.7346 - val_accuracy: 0.9364\n",
      "Epoch 655/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0181 - accuracy: 0.9887 - val_loss: 0.8826 - val_accuracy: 0.9348\n",
      "Epoch 656/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0189 - accuracy: 0.9893 - val_loss: 0.7525 - val_accuracy: 0.9377\n",
      "Epoch 657/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0188 - accuracy: 0.9893 - val_loss: 0.7111 - val_accuracy: 0.9352\n",
      "Epoch 658/1000\n",
      "22/22 [==============================] - 67s 3s/step - loss: 0.0185 - accuracy: 0.9887 - val_loss: 1.0389 - val_accuracy: 0.9331\n",
      "Epoch 659/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0188 - accuracy: 0.9889 - val_loss: 1.1710 - val_accuracy: 0.9272\n",
      "Epoch 660/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0199 - accuracy: 0.9889 - val_loss: 1.0382 - val_accuracy: 0.9339\n",
      "Epoch 661/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0185 - accuracy: 0.9891 - val_loss: 1.0620 - val_accuracy: 0.9373\n",
      "Epoch 662/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0221 - accuracy: 0.9883 - val_loss: 0.9972 - val_accuracy: 0.9293\n",
      "Epoch 663/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0284 - val_accuracy: 0.9335\n",
      "Epoch 664/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0198 - accuracy: 0.9889 - val_loss: 1.0437 - val_accuracy: 0.9343\n",
      "Epoch 665/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0272 - accuracy: 0.9866 - val_loss: 0.9869 - val_accuracy: 0.9310\n",
      "Epoch 666/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0194 - accuracy: 0.9891 - val_loss: 0.9530 - val_accuracy: 0.9285\n",
      "Epoch 667/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0182 - accuracy: 0.9887 - val_loss: 0.8781 - val_accuracy: 0.9322\n",
      "Epoch 668/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0186 - accuracy: 0.9878 - val_loss: 0.8755 - val_accuracy: 0.9293\n",
      "Epoch 669/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 0.9005 - val_accuracy: 0.9285\n",
      "Epoch 670/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 0.9225 - val_accuracy: 0.9306\n",
      "Epoch 671/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 0.9383 - val_accuracy: 0.9306\n",
      "Epoch 672/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.9514 - val_accuracy: 0.9306\n",
      "Epoch 673/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0180 - accuracy: 0.9895 - val_loss: 0.9591 - val_accuracy: 0.9306\n",
      "Epoch 674/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 0.9631 - val_accuracy: 0.9310\n",
      "Epoch 675/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 0.9712 - val_accuracy: 0.9310\n",
      "Epoch 676/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0178 - accuracy: 0.9899 - val_loss: 0.9801 - val_accuracy: 0.9297\n",
      "Epoch 677/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 0.9843 - val_accuracy: 0.9302\n",
      "Epoch 678/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 0.9875 - val_accuracy: 0.9306\n",
      "Epoch 679/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 0.9907 - val_accuracy: 0.9310\n",
      "Epoch 680/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0173 - accuracy: 0.9887 - val_loss: 1.0046 - val_accuracy: 0.9314\n",
      "Epoch 681/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 1.0084 - val_accuracy: 0.9302\n",
      "Epoch 682/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.0117 - val_accuracy: 0.9302\n",
      "Epoch 683/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.0168 - val_accuracy: 0.9302\n",
      "Epoch 684/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.0193 - val_accuracy: 0.9302\n",
      "Epoch 685/1000\n",
      "22/22 [==============================] - 74s 3s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.0204 - val_accuracy: 0.9306\n",
      "Epoch 686/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.0300 - val_accuracy: 0.9302\n",
      "Epoch 687/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.0366 - val_accuracy: 0.9302\n",
      "Epoch 688/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0182 - accuracy: 0.9889 - val_loss: 1.0375 - val_accuracy: 0.9310\n",
      "Epoch 689/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 1.0450 - val_accuracy: 0.9306\n",
      "Epoch 690/1000\n",
      "22/22 [==============================] - 74s 3s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 1.0542 - val_accuracy: 0.9302\n",
      "Epoch 691/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.0545 - val_accuracy: 0.9302\n",
      "Epoch 692/1000\n",
      "22/22 [==============================] - 70s 3s/step - loss: 0.0170 - accuracy: 0.9891 - val_loss: 1.0587 - val_accuracy: 0.9302\n",
      "Epoch 693/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.0618 - val_accuracy: 0.9302\n",
      "Epoch 694/1000\n",
      "22/22 [==============================] - 68s 3s/step - loss: 0.0175 - accuracy: 0.9883 - val_loss: 1.0596 - val_accuracy: 0.9310\n",
      "Epoch 695/1000\n",
      "22/22 [==============================] - 72s 3s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 1.0666 - val_accuracy: 0.9302\n",
      "Epoch 696/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 1.0692 - val_accuracy: 0.9293\n",
      "Epoch 697/1000\n",
      "22/22 [==============================] - 63s 3s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 1.0724 - val_accuracy: 0.9302\n",
      "Epoch 698/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.0757 - val_accuracy: 0.9297\n",
      "Epoch 699/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0175 - accuracy: 0.9887 - val_loss: 1.0807 - val_accuracy: 0.9293\n",
      "Epoch 700/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0170 - accuracy: 0.9899 - val_loss: 1.0796 - val_accuracy: 0.9293\n",
      "Epoch 701/1000\n",
      "22/22 [==============================] - 68s 3s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.0793 - val_accuracy: 0.9293\n",
      "Epoch 702/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 1.0859 - val_accuracy: 0.9297\n",
      "Epoch 703/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 1.0907 - val_accuracy: 0.9297\n",
      "Epoch 704/1000\n",
      "22/22 [==============================] - 81s 4s/step - loss: 0.0175 - accuracy: 0.9878 - val_loss: 1.0928 - val_accuracy: 0.9310\n",
      "Epoch 705/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9899 - val_loss: 1.0957 - val_accuracy: 0.9302\n",
      "Epoch 706/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0172 - accuracy: 0.9899 - val_loss: 1.0992 - val_accuracy: 0.9314\n",
      "Epoch 707/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0178 - accuracy: 0.9903 - val_loss: 1.0995 - val_accuracy: 0.9297\n",
      "Epoch 708/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 1.1010 - val_accuracy: 0.9310\n",
      "Epoch 709/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 1.1066 - val_accuracy: 0.9297\n",
      "Epoch 710/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0180 - accuracy: 0.9889 - val_loss: 1.1000 - val_accuracy: 0.9297\n",
      "Epoch 711/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 1.1014 - val_accuracy: 0.9297\n",
      "Epoch 712/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0170 - accuracy: 0.9889 - val_loss: 1.1046 - val_accuracy: 0.9314\n",
      "Epoch 713/1000\n",
      "22/22 [==============================] - 70s 3s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 1.1065 - val_accuracy: 0.9314\n",
      "Epoch 714/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.1129 - val_accuracy: 0.9302\n",
      "Epoch 715/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.1130 - val_accuracy: 0.9306\n",
      "Epoch 716/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.1168 - val_accuracy: 0.9297\n",
      "Epoch 717/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.1179 - val_accuracy: 0.9297\n",
      "Epoch 718/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0182 - accuracy: 0.9897 - val_loss: 1.1161 - val_accuracy: 0.9306\n",
      "Epoch 719/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 1.1198 - val_accuracy: 0.9310\n",
      "Epoch 720/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.1247 - val_accuracy: 0.9306\n",
      "Epoch 721/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.1261 - val_accuracy: 0.9322\n",
      "Epoch 722/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.1295 - val_accuracy: 0.9306\n",
      "Epoch 723/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 1.1323 - val_accuracy: 0.9310\n",
      "Epoch 724/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0182 - accuracy: 0.9901 - val_loss: 1.1346 - val_accuracy: 0.9306\n",
      "Epoch 725/1000\n",
      "22/22 [==============================] - 67s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.1369 - val_accuracy: 0.9302\n",
      "Epoch 726/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1397 - val_accuracy: 0.9297\n",
      "Epoch 727/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0171 - accuracy: 0.9887 - val_loss: 1.1399 - val_accuracy: 0.9310\n",
      "Epoch 728/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0177 - accuracy: 0.9889 - val_loss: 1.1430 - val_accuracy: 0.9297\n",
      "Epoch 729/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0174 - accuracy: 0.9899 - val_loss: 1.1444 - val_accuracy: 0.9310\n",
      "Epoch 730/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0171 - accuracy: 0.9880 - val_loss: 1.1439 - val_accuracy: 0.9297\n",
      "Epoch 731/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.1459 - val_accuracy: 0.9293\n",
      "Epoch 732/1000\n",
      "22/22 [==============================] - 73s 3s/step - loss: 0.0183 - accuracy: 0.9887 - val_loss: 1.1494 - val_accuracy: 0.9314\n",
      "Epoch 733/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9907 - val_loss: 1.1538 - val_accuracy: 0.9302\n",
      "Epoch 734/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.1583 - val_accuracy: 0.9306\n",
      "Epoch 735/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0171 - accuracy: 0.9891 - val_loss: 1.1561 - val_accuracy: 0.9310\n",
      "Epoch 736/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1565 - val_accuracy: 0.9310\n",
      "Epoch 737/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.1565 - val_accuracy: 0.9306\n",
      "Epoch 738/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.1581 - val_accuracy: 0.9306\n",
      "Epoch 739/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0180 - accuracy: 0.9893 - val_loss: 1.1625 - val_accuracy: 0.9310\n",
      "Epoch 740/1000\n",
      "22/22 [==============================] - 79s 4s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1647 - val_accuracy: 0.9322\n",
      "Epoch 741/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 1.1639 - val_accuracy: 0.9318\n",
      "Epoch 742/1000\n",
      "22/22 [==============================] - 75s 3s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1651 - val_accuracy: 0.9310\n",
      "Epoch 743/1000\n",
      "22/22 [==============================] - 63s 3s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 1.1719 - val_accuracy: 0.9310\n",
      "Epoch 744/1000\n",
      "22/22 [==============================] - 72s 3s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.1705 - val_accuracy: 0.9310\n",
      "Epoch 745/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 1.1654 - val_accuracy: 0.9310\n",
      "Epoch 746/1000\n",
      "22/22 [==============================] - 77s 4s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 1.1721 - val_accuracy: 0.9310\n",
      "Epoch 747/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 1.1859 - val_accuracy: 0.9297\n",
      "Epoch 748/1000\n",
      "22/22 [==============================] - 79s 4s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.1840 - val_accuracy: 0.9306\n",
      "Epoch 749/1000\n",
      "22/22 [==============================] - 77s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.1880 - val_accuracy: 0.9314\n",
      "Epoch 750/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 1.1891 - val_accuracy: 0.9310\n",
      "Epoch 751/1000\n",
      "22/22 [==============================] - 68s 3s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 1.1872 - val_accuracy: 0.9306\n",
      "Epoch 752/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 1.1938 - val_accuracy: 0.9310\n",
      "Epoch 753/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 1.1926 - val_accuracy: 0.9306\n",
      "Epoch 754/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 1.1971 - val_accuracy: 0.9310\n",
      "Epoch 755/1000\n",
      "22/22 [==============================] - 72s 3s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.2092 - val_accuracy: 0.9293\n",
      "Epoch 756/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0171 - accuracy: 0.9897 - val_loss: 1.2047 - val_accuracy: 0.9289\n",
      "Epoch 757/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.2080 - val_accuracy: 0.9289\n",
      "Epoch 758/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2083 - val_accuracy: 0.9302\n",
      "Epoch 759/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2072 - val_accuracy: 0.9289\n",
      "Epoch 760/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0173 - accuracy: 0.9887 - val_loss: 1.2067 - val_accuracy: 0.9289\n",
      "Epoch 761/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.2106 - val_accuracy: 0.9302\n",
      "Epoch 762/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9880 - val_loss: 1.2107 - val_accuracy: 0.9302\n",
      "Epoch 763/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0175 - accuracy: 0.9883 - val_loss: 1.2108 - val_accuracy: 0.9293\n",
      "Epoch 764/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 1.2116 - val_accuracy: 0.9297\n",
      "Epoch 765/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2133 - val_accuracy: 0.9306\n",
      "Epoch 766/1000\n",
      "22/22 [==============================] - 84s 4s/step - loss: 0.0170 - accuracy: 0.9891 - val_loss: 1.2177 - val_accuracy: 0.9302\n",
      "Epoch 767/1000\n",
      "22/22 [==============================] - 68s 3s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.2188 - val_accuracy: 0.9302\n",
      "Epoch 768/1000\n",
      "22/22 [==============================] - 63s 3s/step - loss: 0.0172 - accuracy: 0.9891 - val_loss: 1.2218 - val_accuracy: 0.9293\n",
      "Epoch 769/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0172 - accuracy: 0.9883 - val_loss: 1.2179 - val_accuracy: 0.9297\n",
      "Epoch 770/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0172 - accuracy: 0.9899 - val_loss: 1.2267 - val_accuracy: 0.9297\n",
      "Epoch 771/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.2259 - val_accuracy: 0.9285\n",
      "Epoch 772/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.2283 - val_accuracy: 0.9276\n",
      "Epoch 773/1000\n",
      "22/22 [==============================] - 63s 3s/step - loss: 0.0175 - accuracy: 0.9897 - val_loss: 1.2244 - val_accuracy: 0.9285\n",
      "Epoch 774/1000\n",
      "22/22 [==============================] - 75s 3s/step - loss: 0.0175 - accuracy: 0.9903 - val_loss: 1.2226 - val_accuracy: 0.9302\n",
      "Epoch 775/1000\n",
      "22/22 [==============================] - 70s 3s/step - loss: 0.0170 - accuracy: 0.9901 - val_loss: 1.2214 - val_accuracy: 0.9302\n",
      "Epoch 776/1000\n",
      "22/22 [==============================] - 76s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2176 - val_accuracy: 0.9314\n",
      "Epoch 777/1000\n",
      "22/22 [==============================] - 93s 4s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2257 - val_accuracy: 0.9310\n",
      "Epoch 778/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 1.2206 - val_accuracy: 0.9318\n",
      "Epoch 779/1000\n",
      "22/22 [==============================] - 84s 4s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.2221 - val_accuracy: 0.9306\n",
      "Epoch 780/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0175 - accuracy: 0.9901 - val_loss: 1.2259 - val_accuracy: 0.9293\n",
      "Epoch 781/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 1.2281 - val_accuracy: 0.9306\n",
      "Epoch 782/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 1.2296 - val_accuracy: 0.9318\n",
      "Epoch 783/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.2305 - val_accuracy: 0.9297\n",
      "Epoch 784/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 1.2289 - val_accuracy: 0.9289\n",
      "Epoch 785/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2297 - val_accuracy: 0.9293\n",
      "Epoch 786/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0174 - accuracy: 0.9889 - val_loss: 1.2299 - val_accuracy: 0.9289\n",
      "Epoch 787/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.2317 - val_accuracy: 0.9289\n",
      "Epoch 788/1000\n",
      "22/22 [==============================] - 67s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2332 - val_accuracy: 0.9297\n",
      "Epoch 789/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 1.2388 - val_accuracy: 0.9285\n",
      "Epoch 790/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.2444 - val_accuracy: 0.9285\n",
      "Epoch 791/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0176 - accuracy: 0.9878 - val_loss: 1.2429 - val_accuracy: 0.9289\n",
      "Epoch 792/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 1.2458 - val_accuracy: 0.9285\n",
      "Epoch 793/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2255 - val_accuracy: 0.9281\n",
      "Epoch 794/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 1.2314 - val_accuracy: 0.9276\n",
      "Epoch 795/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.2347 - val_accuracy: 0.9285\n",
      "Epoch 796/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.2415 - val_accuracy: 0.9281\n",
      "Epoch 797/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.2388 - val_accuracy: 0.9276\n",
      "Epoch 798/1000\n",
      "22/22 [==============================] - 55s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2399 - val_accuracy: 0.9276\n",
      "Epoch 799/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2505 - val_accuracy: 0.9276\n",
      "Epoch 800/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.2602 - val_accuracy: 0.9272\n",
      "Epoch 801/1000\n",
      "22/22 [==============================] - 62s 3s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.2734 - val_accuracy: 0.9276\n",
      "Epoch 802/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2729 - val_accuracy: 0.9285\n",
      "Epoch 803/1000\n",
      "22/22 [==============================] - 58s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2802 - val_accuracy: 0.9281\n",
      "Epoch 804/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0178 - accuracy: 0.9887 - val_loss: 1.2788 - val_accuracy: 0.9285\n",
      "Epoch 805/1000\n",
      "22/22 [==============================] - 64s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2905 - val_accuracy: 0.9281\n",
      "Epoch 806/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0171 - accuracy: 0.9895 - val_loss: 1.2970 - val_accuracy: 0.9281\n",
      "Epoch 807/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.2968 - val_accuracy: 0.9281\n",
      "Epoch 808/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0175 - accuracy: 0.9883 - val_loss: 1.2989 - val_accuracy: 0.9297\n",
      "Epoch 809/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.3032 - val_accuracy: 0.9285\n",
      "Epoch 810/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2725 - val_accuracy: 0.9281\n",
      "Epoch 811/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.2153 - val_accuracy: 0.9281\n",
      "Epoch 812/1000\n",
      "22/22 [==============================] - 55s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2108 - val_accuracy: 0.9285\n",
      "Epoch 813/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0171 - accuracy: 0.9891 - val_loss: 1.2273 - val_accuracy: 0.9276\n",
      "Epoch 814/1000\n",
      "22/22 [==============================] - 65s 3s/step - loss: 0.0174 - accuracy: 0.9901 - val_loss: 1.1500 - val_accuracy: 0.9302\n",
      "Epoch 815/1000\n",
      "22/22 [==============================] - 61s 3s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.1520 - val_accuracy: 0.9310\n",
      "Epoch 816/1000\n",
      "22/22 [==============================] - 63s 3s/step - loss: 0.0177 - accuracy: 0.9887 - val_loss: 1.1646 - val_accuracy: 0.9302\n",
      "Epoch 817/1000\n",
      "22/22 [==============================] - 59s 3s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.1731 - val_accuracy: 0.9297\n",
      "Epoch 818/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 1.1731 - val_accuracy: 0.9297\n",
      "Epoch 819/1000\n",
      "22/22 [==============================] - 77s 4s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.1872 - val_accuracy: 0.9297\n",
      "Epoch 820/1000\n",
      "22/22 [==============================] - 56s 3s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.1985 - val_accuracy: 0.9297\n",
      "Epoch 821/1000\n",
      "22/22 [==============================] - 60s 3s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.2079 - val_accuracy: 0.9285\n",
      "Epoch 822/1000\n",
      "22/22 [==============================] - 111s 5s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2205 - val_accuracy: 0.9289\n",
      "Epoch 823/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.2254 - val_accuracy: 0.9297\n",
      "Epoch 824/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2305 - val_accuracy: 0.9302\n",
      "Epoch 825/1000\n",
      "22/22 [==============================] - 69s 3s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.2393 - val_accuracy: 0.9302\n",
      "Epoch 826/1000\n",
      "22/22 [==============================] - 76s 3s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.2500 - val_accuracy: 0.9297\n",
      "Epoch 827/1000\n",
      "22/22 [==============================] - 74s 3s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2523 - val_accuracy: 0.9297\n",
      "Epoch 828/1000\n",
      "22/22 [==============================] - 53s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2568 - val_accuracy: 0.9297\n",
      "Epoch 829/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2525 - val_accuracy: 0.9297\n",
      "Epoch 830/1000\n",
      "22/22 [==============================] - 86s 4s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2549 - val_accuracy: 0.9293\n",
      "Epoch 831/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2550 - val_accuracy: 0.9293\n",
      "Epoch 832/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.1823 - val_accuracy: 0.9264\n",
      "Epoch 833/1000\n",
      "22/22 [==============================] - 51s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 1.1045 - val_accuracy: 0.9281\n",
      "Epoch 834/1000\n",
      "22/22 [==============================] - 52s 2s/step - loss: 0.0177 - accuracy: 0.9891 - val_loss: 1.1094 - val_accuracy: 0.9276\n",
      "Epoch 835/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 1.1558 - val_accuracy: 0.9268\n",
      "Epoch 836/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.1939 - val_accuracy: 0.9281\n",
      "Epoch 837/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1971 - val_accuracy: 0.9302\n",
      "Epoch 838/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0174 - accuracy: 0.9897 - val_loss: 1.2064 - val_accuracy: 0.9302\n",
      "Epoch 839/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0189 - accuracy: 0.9880 - val_loss: 1.0408 - val_accuracy: 0.9235\n",
      "Epoch 840/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0178 - accuracy: 0.9885 - val_loss: 1.1071 - val_accuracy: 0.9264\n",
      "Epoch 841/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1315 - val_accuracy: 0.9272\n",
      "Epoch 842/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0176 - accuracy: 0.9899 - val_loss: 1.1538 - val_accuracy: 0.9281\n",
      "Epoch 843/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0174 - accuracy: 0.9893 - val_loss: 1.1628 - val_accuracy: 0.9272\n",
      "Epoch 844/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1751 - val_accuracy: 0.9264\n",
      "Epoch 845/1000\n",
      "22/22 [==============================] - 49s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.1787 - val_accuracy: 0.9264\n",
      "Epoch 846/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.1870 - val_accuracy: 0.9272\n",
      "Epoch 847/1000\n",
      "22/22 [==============================] - 54s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.1732 - val_accuracy: 0.9297\n",
      "Epoch 848/1000\n",
      "22/22 [==============================] - 85s 4s/step - loss: 0.0181 - accuracy: 0.9895 - val_loss: 1.1732 - val_accuracy: 0.9281\n",
      "Epoch 849/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9897 - val_loss: 1.1881 - val_accuracy: 0.9268\n",
      "Epoch 850/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2150 - val_accuracy: 0.9268\n",
      "Epoch 851/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2247 - val_accuracy: 0.9264\n",
      "Epoch 852/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2334 - val_accuracy: 0.9260\n",
      "Epoch 853/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.2424 - val_accuracy: 0.9268\n",
      "Epoch 854/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 1.2490 - val_accuracy: 0.9260\n",
      "Epoch 855/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0180 - accuracy: 0.9891 - val_loss: 1.2505 - val_accuracy: 0.9276\n",
      "Epoch 856/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2541 - val_accuracy: 0.9281\n",
      "Epoch 857/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.2609 - val_accuracy: 0.9251\n",
      "Epoch 858/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2681 - val_accuracy: 0.9251\n",
      "Epoch 859/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0174 - accuracy: 0.9891 - val_loss: 1.2782 - val_accuracy: 0.9260\n",
      "Epoch 860/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.2756 - val_accuracy: 0.9268\n",
      "Epoch 861/1000\n",
      "22/22 [==============================] - 48s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.2829 - val_accuracy: 0.9268\n",
      "Epoch 862/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0181 - accuracy: 0.9893 - val_loss: 1.2978 - val_accuracy: 0.9268\n",
      "Epoch 863/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.3025 - val_accuracy: 0.9268\n",
      "Epoch 864/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.3620 - val_accuracy: 0.9247\n",
      "Epoch 865/1000\n",
      "22/22 [==============================] - 46s 2s/step - loss: 0.0190 - accuracy: 0.9885 - val_loss: 0.9764 - val_accuracy: 0.9318\n",
      "Epoch 866/1000\n",
      "22/22 [==============================] - 47s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.0543 - val_accuracy: 0.9327\n",
      "Epoch 867/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0178 - accuracy: 0.9893 - val_loss: 1.0908 - val_accuracy: 0.9314\n",
      "Epoch 868/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0170 - accuracy: 0.9899 - val_loss: 1.1235 - val_accuracy: 0.9318\n",
      "Epoch 869/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0184 - accuracy: 0.9897 - val_loss: 0.9750 - val_accuracy: 0.9314\n",
      "Epoch 870/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.0455 - val_accuracy: 0.9297\n",
      "Epoch 871/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.0688 - val_accuracy: 0.9293\n",
      "Epoch 872/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0175 - accuracy: 0.9899 - val_loss: 1.0660 - val_accuracy: 0.9318\n",
      "Epoch 873/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 1.0685 - val_accuracy: 0.9306\n",
      "Epoch 874/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0173 - accuracy: 0.9891 - val_loss: 1.0640 - val_accuracy: 0.9302\n",
      "Epoch 875/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.0756 - val_accuracy: 0.9285\n",
      "Epoch 876/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9885 - val_loss: 1.0898 - val_accuracy: 0.9281\n",
      "Epoch 877/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.0859 - val_accuracy: 0.9306\n",
      "Epoch 878/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.1010 - val_accuracy: 0.9314\n",
      "Epoch 879/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.1115 - val_accuracy: 0.9314\n",
      "Epoch 880/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0171 - accuracy: 0.9893 - val_loss: 1.1137 - val_accuracy: 0.9310\n",
      "Epoch 881/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.1142 - val_accuracy: 0.9310\n",
      "Epoch 882/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.1219 - val_accuracy: 0.9310\n",
      "Epoch 883/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1246 - val_accuracy: 0.9314\n",
      "Epoch 884/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1262 - val_accuracy: 0.9314\n",
      "Epoch 885/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0171 - accuracy: 0.9895 - val_loss: 1.1275 - val_accuracy: 0.9310\n",
      "Epoch 886/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1383 - val_accuracy: 0.9314\n",
      "Epoch 887/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.1414 - val_accuracy: 0.9314\n",
      "Epoch 888/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.1455 - val_accuracy: 0.9314\n",
      "Epoch 889/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.1564 - val_accuracy: 0.9310\n",
      "Epoch 890/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.1617 - val_accuracy: 0.9310\n",
      "Epoch 891/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.1551 - val_accuracy: 0.9310\n",
      "Epoch 892/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1557 - val_accuracy: 0.9310\n",
      "Epoch 893/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.1557 - val_accuracy: 0.9310\n",
      "Epoch 894/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 1.1584 - val_accuracy: 0.9306\n",
      "Epoch 895/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 1.1493 - val_accuracy: 0.9314\n",
      "Epoch 896/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9895 - val_loss: 1.1492 - val_accuracy: 0.9310\n",
      "Epoch 897/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.1683 - val_accuracy: 0.9310\n",
      "Epoch 898/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.1777 - val_accuracy: 0.9306\n",
      "Epoch 899/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.1844 - val_accuracy: 0.9302\n",
      "Epoch 900/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.1772 - val_accuracy: 0.9314\n",
      "Epoch 901/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1833 - val_accuracy: 0.9314\n",
      "Epoch 902/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.1826 - val_accuracy: 0.9318\n",
      "Epoch 903/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0171 - accuracy: 0.9895 - val_loss: 1.1921 - val_accuracy: 0.9318\n",
      "Epoch 904/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.1876 - val_accuracy: 0.9318\n",
      "Epoch 905/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0171 - accuracy: 0.9893 - val_loss: 1.2046 - val_accuracy: 0.9310\n",
      "Epoch 906/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.2414 - val_accuracy: 0.9306\n",
      "Epoch 907/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.2487 - val_accuracy: 0.9310\n",
      "Epoch 908/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2377 - val_accuracy: 0.9310\n",
      "Epoch 909/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0179 - accuracy: 0.9893 - val_loss: 1.2421 - val_accuracy: 0.9310\n",
      "Epoch 910/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2337 - val_accuracy: 0.9310\n",
      "Epoch 911/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0169 - accuracy: 0.9897 - val_loss: 1.2288 - val_accuracy: 0.9310\n",
      "Epoch 912/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.2272 - val_accuracy: 0.9310\n",
      "Epoch 913/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.2311 - val_accuracy: 0.9310\n",
      "Epoch 914/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.2340 - val_accuracy: 0.9310\n",
      "Epoch 915/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.2450 - val_accuracy: 0.9297\n",
      "Epoch 916/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0181 - accuracy: 0.9895 - val_loss: 1.2200 - val_accuracy: 0.9314\n",
      "Epoch 917/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1355 - val_accuracy: 0.9310\n",
      "Epoch 918/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9895 - val_loss: 1.1474 - val_accuracy: 0.9306\n",
      "Epoch 919/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1671 - val_accuracy: 0.9310\n",
      "Epoch 920/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1822 - val_accuracy: 0.9314\n",
      "Epoch 921/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1879 - val_accuracy: 0.9310\n",
      "Epoch 922/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.1791 - val_accuracy: 0.9314\n",
      "Epoch 923/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.1807 - val_accuracy: 0.9306\n",
      "Epoch 924/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.1856 - val_accuracy: 0.9306\n",
      "Epoch 925/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1880 - val_accuracy: 0.9306\n",
      "Epoch 926/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.1994 - val_accuracy: 0.9297\n",
      "Epoch 927/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1957 - val_accuracy: 0.9306\n",
      "Epoch 928/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.1850 - val_accuracy: 0.9302\n",
      "Epoch 929/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.1978 - val_accuracy: 0.9306\n",
      "Epoch 930/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2021 - val_accuracy: 0.9306\n",
      "Epoch 931/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2076 - val_accuracy: 0.9306\n",
      "Epoch 932/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2093 - val_accuracy: 0.9310\n",
      "Epoch 933/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2230 - val_accuracy: 0.9306\n",
      "Epoch 934/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2261 - val_accuracy: 0.9306\n",
      "Epoch 935/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2257 - val_accuracy: 0.9310\n",
      "Epoch 936/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2348 - val_accuracy: 0.9310\n",
      "Epoch 937/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.2537 - val_accuracy: 0.9310\n",
      "Epoch 938/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.2666 - val_accuracy: 0.9310\n",
      "Epoch 939/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9895 - val_loss: 1.2714 - val_accuracy: 0.9306\n",
      "Epoch 940/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.2647 - val_accuracy: 0.9310\n",
      "Epoch 941/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2571 - val_accuracy: 0.9297\n",
      "Epoch 942/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9895 - val_loss: 1.2767 - val_accuracy: 0.9302\n",
      "Epoch 943/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2865 - val_accuracy: 0.9302\n",
      "Epoch 944/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.2787 - val_accuracy: 0.9314\n",
      "Epoch 945/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.2837 - val_accuracy: 0.9314\n",
      "Epoch 946/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.2836 - val_accuracy: 0.9322\n",
      "Epoch 947/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0189 - accuracy: 0.9883 - val_loss: 1.2704 - val_accuracy: 0.9310\n",
      "Epoch 948/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 1.3144 - val_accuracy: 0.9272\n",
      "Epoch 949/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0178 - accuracy: 0.9891 - val_loss: 1.2892 - val_accuracy: 0.9218\n",
      "Epoch 950/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0185 - accuracy: 0.9895 - val_loss: 1.2854 - val_accuracy: 0.9306\n",
      "Epoch 951/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0186 - accuracy: 0.9887 - val_loss: 1.1515 - val_accuracy: 0.9318\n",
      "Epoch 952/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0207 - accuracy: 0.9880 - val_loss: 1.2454 - val_accuracy: 0.9302\n",
      "Epoch 953/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0242 - accuracy: 0.9872 - val_loss: 0.9543 - val_accuracy: 0.9306\n",
      "Epoch 954/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0193 - accuracy: 0.9891 - val_loss: 0.9840 - val_accuracy: 0.9322\n",
      "Epoch 955/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0228 - accuracy: 0.9870 - val_loss: 1.2518 - val_accuracy: 0.9293\n",
      "Epoch 956/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.3936 - val_accuracy: 0.9289\n",
      "Epoch 957/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0193 - accuracy: 0.9887 - val_loss: 1.5558 - val_accuracy: 0.9197\n",
      "Epoch 958/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0210 - accuracy: 0.9885 - val_loss: 1.4109 - val_accuracy: 0.9335\n",
      "Epoch 959/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0200 - accuracy: 0.9891 - val_loss: 1.4020 - val_accuracy: 0.9289\n",
      "Epoch 960/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.4492 - val_accuracy: 0.9302\n",
      "Epoch 961/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0181 - accuracy: 0.9893 - val_loss: 1.4023 - val_accuracy: 0.9293\n",
      "Epoch 962/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0171 - accuracy: 0.9893 - val_loss: 1.3846 - val_accuracy: 0.9297\n",
      "Epoch 963/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.4036 - val_accuracy: 0.9297\n",
      "Epoch 964/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0173 - accuracy: 0.9893 - val_loss: 1.4118 - val_accuracy: 0.9281\n",
      "Epoch 965/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9897 - val_loss: 1.4354 - val_accuracy: 0.9281\n",
      "Epoch 966/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0175 - accuracy: 0.9889 - val_loss: 1.4604 - val_accuracy: 0.9289\n",
      "Epoch 967/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.4420 - val_accuracy: 0.9314\n",
      "Epoch 968/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0173 - accuracy: 0.9895 - val_loss: 1.4653 - val_accuracy: 0.9314\n",
      "Epoch 969/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0179 - accuracy: 0.9889 - val_loss: 1.4904 - val_accuracy: 0.9243\n",
      "Epoch 970/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.4526 - val_accuracy: 0.9289\n",
      "Epoch 971/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0202 - accuracy: 0.9889 - val_loss: 1.5084 - val_accuracy: 0.9281\n",
      "Epoch 972/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0196 - accuracy: 0.9887 - val_loss: 1.5321 - val_accuracy: 0.9327\n",
      "Epoch 973/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0196 - accuracy: 0.9893 - val_loss: 1.5135 - val_accuracy: 0.9276\n",
      "Epoch 974/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0184 - accuracy: 0.9891 - val_loss: 1.4532 - val_accuracy: 0.9302\n",
      "Epoch 975/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9895 - val_loss: 1.4448 - val_accuracy: 0.9314\n",
      "Epoch 976/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0193 - accuracy: 0.9895 - val_loss: 1.4243 - val_accuracy: 0.9281\n",
      "Epoch 977/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9880 - val_loss: 1.3107 - val_accuracy: 0.9310\n",
      "Epoch 978/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9887 - val_loss: 1.3250 - val_accuracy: 0.9310\n",
      "Epoch 979/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0174 - accuracy: 0.9887 - val_loss: 1.3314 - val_accuracy: 0.9310\n",
      "Epoch 980/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9891 - val_loss: 1.3447 - val_accuracy: 0.9322\n",
      "Epoch 981/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0176 - accuracy: 0.9891 - val_loss: 1.3545 - val_accuracy: 0.9314\n",
      "Epoch 982/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9895 - val_loss: 1.3612 - val_accuracy: 0.9310\n",
      "Epoch 983/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9901 - val_loss: 1.3674 - val_accuracy: 0.9306\n",
      "Epoch 984/1000\n",
      "22/22 [==============================] - 41s 2s/step - loss: 0.0177 - accuracy: 0.9897 - val_loss: 1.3700 - val_accuracy: 0.9310\n",
      "Epoch 985/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.3731 - val_accuracy: 0.9314\n",
      "Epoch 986/1000\n",
      "22/22 [==============================] - 40s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.3788 - val_accuracy: 0.9302\n",
      "Epoch 987/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0176 - accuracy: 0.9889 - val_loss: 1.3830 - val_accuracy: 0.9310\n",
      "Epoch 988/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.3849 - val_accuracy: 0.9314\n",
      "Epoch 989/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0176 - accuracy: 0.9893 - val_loss: 1.3897 - val_accuracy: 0.9310\n",
      "Epoch 990/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.3962 - val_accuracy: 0.9310\n",
      "Epoch 991/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.3968 - val_accuracy: 0.9314\n",
      "Epoch 992/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.3969 - val_accuracy: 0.9314\n",
      "Epoch 993/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0168 - accuracy: 0.9895 - val_loss: 1.4015 - val_accuracy: 0.9314\n",
      "Epoch 994/1000\n",
      "22/22 [==============================] - 45s 2s/step - loss: 0.0175 - accuracy: 0.9893 - val_loss: 1.4046 - val_accuracy: 0.9314\n",
      "Epoch 995/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0170 - accuracy: 0.9893 - val_loss: 1.4096 - val_accuracy: 0.9314\n",
      "Epoch 996/1000\n",
      "22/22 [==============================] - 44s 2s/step - loss: 0.0173 - accuracy: 0.9897 - val_loss: 1.4155 - val_accuracy: 0.9314\n",
      "Epoch 997/1000\n",
      "22/22 [==============================] - 42s 2s/step - loss: 0.0172 - accuracy: 0.9893 - val_loss: 1.4202 - val_accuracy: 0.9314\n",
      "Epoch 998/1000\n",
      "22/22 [==============================] - 50s 2s/step - loss: 0.0172 - accuracy: 0.9897 - val_loss: 1.4196 - val_accuracy: 0.9314\n",
      "Epoch 999/1000\n",
      "22/22 [==============================] - 57s 3s/step - loss: 0.0177 - accuracy: 0.9893 - val_loss: 1.4201 - val_accuracy: 0.9314\n",
      "Epoch 1000/1000\n",
      "22/22 [==============================] - 43s 2s/step - loss: 0.0175 - accuracy: 0.9895 - val_loss: 1.4238 - val_accuracy: 0.9310\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'checkpoint_recode_tensorboard/ckpt_manager/ckpt-1'"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Dcnn.fit(train_inputs, train_labels,\n",
    "         vavlidation_data=(test_inputs, test_labels),\n",
    "         batch_size=BATCH_SIZE,\n",
    "         epochs=NB_EPOCHS,\n",
    "         callbacks=[tensorboard_callback])\n",
    "ckpt_manager.save()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ae6fbb4-43be-400f-ab15-9b3175560113",
   "metadata": {},
   "source": [
    "### Stage 5: Tensorboard"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cf4e2d89-7c7a-4173-842d-a784315a9ac7",
   "metadata": {},
   "source": [
    "#### (1) Call Tensorboard from Local"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "954b4e59-802a-4038-b17d-9285d97c31c8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2021-07-19T20:42:27.513233Z",
     "iopub.status.busy": "2021-07-19T20:42:27.512969Z",
     "iopub.status.idle": "2021-07-19T20:42:34.095049Z",
     "shell.execute_reply": "2021-07-19T20:42:34.094135Z",
     "shell.execute_reply.started": "2021-07-19T20:42:27.513203Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "      <iframe id=\"tensorboard-frame-6ccddc1a8e225bb\" width=\"100%\" height=\"800\" frameborder=\"0\">\n",
       "      </iframe>\n",
       "      <script>\n",
       "        (function() {\n",
       "          const frame = document.getElementById(\"tensorboard-frame-6ccddc1a8e225bb\");\n",
       "          const url = new URL(\"/\", window.location);\n",
       "          const port = 6006;\n",
       "          if (port) {\n",
       "            url.port = port;\n",
       "          }\n",
       "          frame.src = url;\n",
       "        })();\n",
       "      </script>\n",
       "    "
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%load_ext tensorboard\n",
    "\n",
    "%tensorboard --logdir logs/fit"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9d58ba1-90f8-459a-a160-a2cc297abd46",
   "metadata": {},
   "source": [
    "#### (2) Upload Pre-run Data to Tensorboard.dev\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "684949a0-a128-40df-8c6d-4329fc52ff33",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "tensorboard dev upload --logdir \\\n",
    "    'logs/fit'"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d35cb76-7423-46d5-b982-9abf29c27d5c",
   "metadata": {},
   "source": [
    "#### (2) Re name\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d1d14684-bc0e-4e93-bfc5-68af3dca871d",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "tensorboard dev update-metadata --experiment_id xxxxxxxxxxxxxxx \\\n",
    "--name \"PorkCNN 🐖🐖🐖\" --description \"A Small Project for Pork Barrel Legislation Classification Using Convolutional Neural Networks\"\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}