Inception-ResNet-4Piece-PCA-Model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "fd1efdf1",
   "metadata": {},
   "source": [
    "### Importing Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "0938de48",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import pickle\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from sklearn.decomposition import PCA\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import load_img, img_to_array\n",
    "\n",
    "from tensorflow.keras.layers import Input, Flatten, Dense, Dropout, GlobalAvgPool2D, BatchNormalization\n",
    "from tensorflow.keras.regularizers import l2\n",
    "\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "\n",
    "from sklearn.metrics import precision_recall_fscore_support as score"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c954b162",
   "metadata": {},
   "source": [
    "## Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68302673",
   "metadata": {},
   "source": [
    "A detailed discussion and visualization of the data can be seen in [here](Data-Overview.ipynb)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "75b06a70",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "labels = pd.read_csv(\"./Data/train_labels.csv\")\n",
    "images = ['./Data/train/train/'+str(name)+'.tif' for name in labels['id']]\n",
    "labels['images'] = images\n",
    "labels = labels[['id', 'images', 'label']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "fd2fe082",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>images</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>./Data/train/train/0.tif</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>./Data/train/train/1.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>./Data/train/train/2.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>./Data/train/train/3.tif</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>./Data/train/train/4.tif</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15995</th>\n",
       "      <td>15995</td>\n",
       "      <td>./Data/train/train/15995.tif</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15996</th>\n",
       "      <td>15996</td>\n",
       "      <td>./Data/train/train/15996.tif</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15997</th>\n",
       "      <td>15997</td>\n",
       "      <td>./Data/train/train/15997.tif</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15998</th>\n",
       "      <td>15998</td>\n",
       "      <td>./Data/train/train/15998.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15999</th>\n",
       "      <td>15999</td>\n",
       "      <td>./Data/train/train/15999.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>16000 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          id                        images  label\n",
       "0          0      ./Data/train/train/0.tif      1\n",
       "1          1      ./Data/train/train/1.tif     13\n",
       "2          2      ./Data/train/train/2.tif     13\n",
       "3          3      ./Data/train/train/3.tif     14\n",
       "4          4      ./Data/train/train/4.tif      6\n",
       "...      ...                           ...    ...\n",
       "15995  15995  ./Data/train/train/15995.tif      2\n",
       "15996  15996  ./Data/train/train/15996.tif     15\n",
       "15997  15997  ./Data/train/train/15997.tif      3\n",
       "15998  15998  ./Data/train/train/15998.tif      9\n",
       "15999  15999  ./Data/train/train/15999.tif      9\n",
       "\n",
       "[16000 rows x 3 columns]"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "276e2059",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "train_labels = pd.read_csv('./Data/train_labels_1.csv')\n",
    "test_labels = pd.read_csv('./Data/test_labels_1.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "2aea9ea9",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>images</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>./Data/train/train/0.tif</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>./Data/train/train/1.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>./Data/train/train/2.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>./Data/train/train/3.tif</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>./Data/train/train/4.tif</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15195</th>\n",
       "      <td>15995</td>\n",
       "      <td>./Data/train/train/15995.tif</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15196</th>\n",
       "      <td>15996</td>\n",
       "      <td>./Data/train/train/15996.tif</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15197</th>\n",
       "      <td>15997</td>\n",
       "      <td>./Data/train/train/15997.tif</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15198</th>\n",
       "      <td>15998</td>\n",
       "      <td>./Data/train/train/15998.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15199</th>\n",
       "      <td>15999</td>\n",
       "      <td>./Data/train/train/15999.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>15200 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          id                        images  label\n",
       "0          0      ./Data/train/train/0.tif      1\n",
       "1          1      ./Data/train/train/1.tif     13\n",
       "2          2      ./Data/train/train/2.tif     13\n",
       "3          3      ./Data/train/train/3.tif     14\n",
       "4          4      ./Data/train/train/4.tif      6\n",
       "...      ...                           ...    ...\n",
       "15195  15995  ./Data/train/train/15995.tif      2\n",
       "15196  15996  ./Data/train/train/15996.tif     15\n",
       "15197  15997  ./Data/train/train/15997.tif      3\n",
       "15198  15998  ./Data/train/train/15998.tif      9\n",
       "15199  15999  ./Data/train/train/15999.tif      9\n",
       "\n",
       "[15200 rows x 3 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "b9ea19c9",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>images</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>18</td>\n",
       "      <td>./Data/train/train/18.tif</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>25</td>\n",
       "      <td>./Data/train/train/25.tif</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>./Data/train/train/33.tif</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41</td>\n",
       "      <td>./Data/train/train/41.tif</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>65</td>\n",
       "      <td>./Data/train/train/65.tif</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>795</th>\n",
       "      <td>15816</td>\n",
       "      <td>./Data/train/train/15816.tif</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>796</th>\n",
       "      <td>15824</td>\n",
       "      <td>./Data/train/train/15824.tif</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>797</th>\n",
       "      <td>15832</td>\n",
       "      <td>./Data/train/train/15832.tif</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>798</th>\n",
       "      <td>15863</td>\n",
       "      <td>./Data/train/train/15863.tif</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>799</th>\n",
       "      <td>15994</td>\n",
       "      <td>./Data/train/train/15994.tif</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>800 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        id                        images  label\n",
       "0       18     ./Data/train/train/18.tif     11\n",
       "1       25     ./Data/train/train/25.tif      8\n",
       "2       33     ./Data/train/train/33.tif      6\n",
       "3       41     ./Data/train/train/41.tif      1\n",
       "4       65     ./Data/train/train/65.tif     14\n",
       "..     ...                           ...    ...\n",
       "795  15816  ./Data/train/train/15816.tif      2\n",
       "796  15824  ./Data/train/train/15824.tif      2\n",
       "797  15832  ./Data/train/train/15832.tif     11\n",
       "798  15863  ./Data/train/train/15863.tif     15\n",
       "799  15994  ./Data/train/train/15994.tif      6\n",
       "\n",
       "[800 rows x 3 columns]"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "29190fd4",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "class_labels = list(set(labels['label']))\n",
    "class_names = [\n",
    "    'letter', 'form', 'email', 'handwritten', 'advertisement', 'scientific report', 'scientific publication',\n",
    "    'specification', 'file folder', 'news article', 'budget', 'invoice', 'presentation', 'questionnaire', 'resume',\n",
    "    'memo'\n",
    "]\n",
    "label_names = pd.DataFrame({\n",
    "    'labels': class_labels,\n",
    "    'names': class_names\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2e251c1",
   "metadata": {},
   "source": [
    "## Data Preparation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "37a36e2b",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16000, 5, 1536)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data = np.load('./Data/Inception-ResNet-4P-train-precompute/whole_train_data_precomp.npy')\n",
    "fine_tune_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "568ffcb1",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16000,)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels = labels['label'].to_numpy()\n",
    "fine_tune_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "937e1e86",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200, 5, 1536)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_train = fine_tune_data[list(train_labels['id'])]\n",
    "fine_tune_data_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "af08c830",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200,)"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels_train = train_labels['label'].to_numpy()\n",
    "fine_tune_labels_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "c00f0b32",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(800, 5, 1536)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_test = fine_tune_data[list(test_labels['id'])]\n",
    "fine_tune_data_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "16f4bb93",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(800,)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels_test = test_labels['label'].to_numpy()\n",
    "fine_tune_labels_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "772c2480",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "N_COMPS = 840\n",
    "pcas = [PCA(N_COMPS) for _ in range(5)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "f0f2b450",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "for i in range(5):\n",
    "    pcas[i].fit(fine_tune_data_train[:,i,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "210b64f9",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fine_tune_data_train_red = np.zeros((fine_tune_data_train.shape[0], 5, N_COMPS))\n",
    "for i in range(5):\n",
    "    fine_tune_data_train_red[:,i,:] = pcas[i].transform(fine_tune_data_train[:,i,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "fb7bb046",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200, 5, 840)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_train_red.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "2c0e4617",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fine_tune_data_test_red = np.zeros((fine_tune_data_test.shape[0], 5, N_COMPS))\n",
    "for i in range(5):\n",
    "    fine_tune_data_test_red[:,i,:] = pcas[i].transform(fine_tune_data_test[:,i,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "b9e1dec6",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(800, 5, 840)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_test_red.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dff1f776",
   "metadata": {},
   "source": [
    "## Fine-Tuning Models"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7380947f",
   "metadata": {},
   "source": [
    "The fine-tuning part of our model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "96a2bdfe",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "def ftmodel():\n",
    "    inputs = Input(shape=(5, N_COMPS))\n",
    "    x = Flatten()(inputs)\n",
    "    x = BatchNormalization()(x)\n",
    "\n",
    "    top_dropout_rate = 0.5\n",
    "    # x = Dropout(top_dropout_rate, name='top_dropout_1')(x)\n",
    "    # x = Dense(1024, name='top_dense_1', kernel_initializer=\"he_normal\", kernel_regularizer=l2(5e-5), activation='relu')(x)\n",
    "    # x = BatchNormalization()(x)\n",
    "    x = Dropout(top_dropout_rate, name='top_dropout_2')(x)\n",
    "    x = Dense(512, name='top_dense_2', kernel_initializer=\"he_normal\", kernel_regularizer=l2(5e-5), activation='relu')(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Dropout(top_dropout_rate, name='top_dropout_3')(x)\n",
    "    outputs = Dense(16, activation='softmax', name='pred', kernel_regularizer=l2(5e-5))(x)\n",
    "\n",
    "    model1 = tf.keras.Model(inputs, outputs, name='Inception-ResNet-FineTune')\n",
    "    return model1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "c6574006",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"Inception-ResNet-FineTune\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_2 (InputLayer)        [(None, 5, 840)]          0         \n",
      "                                                                 \n",
      " flatten_1 (Flatten)         (None, 4200)              0         \n",
      "                                                                 \n",
      " batch_normalization_2 (Batc  (None, 4200)             16800     \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " top_dropout_2 (Dropout)     (None, 4200)              0         \n",
      "                                                                 \n",
      " top_dense_2 (Dense)         (None, 512)               2150912   \n",
      "                                                                 \n",
      " batch_normalization_3 (Batc  (None, 512)              2048      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " top_dropout_3 (Dropout)     (None, 512)               0         \n",
      "                                                                 \n",
      " pred (Dense)                (None, 16)                8208      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 2,177,968\n",
      "Trainable params: 2,168,544\n",
      "Non-trainable params: 9,424\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model1 = ftmodel()\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e56505e2",
   "metadata": {},
   "source": [
    "### Model Training"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "70b7e206",
   "metadata": {},
   "source": [
    "The fine-tuning will be done on the Inception-ResNet's $1536D$ representation vectors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "96524578",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "model1.compile(loss = 'sparse_categorical_crossentropy', metrics = ['accuracy'], optimizer = Adam(learning_rate = 0.0003))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "20df8897",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "EPOCHS = 1000\n",
    "checkpoint_filepaths = ['./Models/Inception-ResNet-4P-PCA/auto-Inception-ResNet-FT-model-weight']\n",
    "model_checkpoint_callback = [tf.keras.callbacks.ModelCheckpoint(\n",
    "    filepath=checkpoint_filepath,\n",
    "    save_weights_only=True,\n",
    "    monitor = \"loss\",\n",
    "    mode='min',\n",
    "    save_best_only=True) for checkpoint_filepath in checkpoint_filepaths]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "03354278",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/1000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-08 19:13:34.016500: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 2s 50ms/step - loss: 3.7089 - accuracy: 0.1201 - val_loss: 2.2748 - val_accuracy: 0.4725\n",
      "Epoch 2/1000\n",
      " 1/19 [>.............................] - ETA: 0s - loss: 2.9189 - accuracy: 0.2425"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-08 19:13:35.214834: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 25ms/step - loss: 2.6680 - accuracy: 0.3036 - val_loss: 1.9582 - val_accuracy: 0.5450\n",
      "Epoch 3/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 2.0875 - accuracy: 0.4376 - val_loss: 1.7752 - val_accuracy: 0.5750\n",
      "Epoch 4/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 1.7277 - accuracy: 0.5288 - val_loss: 1.6584 - val_accuracy: 0.5888\n",
      "Epoch 5/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 1.4988 - accuracy: 0.5874 - val_loss: 1.5730 - val_accuracy: 0.6075\n",
      "Epoch 6/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 1.3450 - accuracy: 0.6269 - val_loss: 1.5067 - val_accuracy: 0.6237\n",
      "Epoch 7/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 1.2324 - accuracy: 0.6575 - val_loss: 1.4493 - val_accuracy: 0.6375\n",
      "Epoch 8/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 1.1376 - accuracy: 0.6850 - val_loss: 1.3986 - val_accuracy: 0.6475\n",
      "Epoch 9/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 1.0563 - accuracy: 0.7068 - val_loss: 1.3535 - val_accuracy: 0.6538\n",
      "Epoch 10/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.9924 - accuracy: 0.7196 - val_loss: 1.3090 - val_accuracy: 0.6612\n",
      "Epoch 11/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.9473 - accuracy: 0.7291 - val_loss: 1.2668 - val_accuracy: 0.6662\n",
      "Epoch 12/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.8828 - accuracy: 0.7501 - val_loss: 1.2289 - val_accuracy: 0.6675\n",
      "Epoch 13/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.8361 - accuracy: 0.7609 - val_loss: 1.1908 - val_accuracy: 0.6787\n",
      "Epoch 14/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.8105 - accuracy: 0.7697 - val_loss: 1.1540 - val_accuracy: 0.6850\n",
      "Epoch 15/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.7755 - accuracy: 0.7799 - val_loss: 1.1165 - val_accuracy: 0.6975\n",
      "Epoch 16/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.7423 - accuracy: 0.7896 - val_loss: 1.0852 - val_accuracy: 0.7050\n",
      "Epoch 17/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.7169 - accuracy: 0.7972 - val_loss: 1.0544 - val_accuracy: 0.7088\n",
      "Epoch 18/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.6895 - accuracy: 0.8004 - val_loss: 1.0230 - val_accuracy: 0.7212\n",
      "Epoch 19/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.6538 - accuracy: 0.8147 - val_loss: 0.9980 - val_accuracy: 0.7312\n",
      "Epoch 20/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.6511 - accuracy: 0.8165 - val_loss: 0.9728 - val_accuracy: 0.7400\n",
      "Epoch 21/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 0.6331 - accuracy: 0.8210 - val_loss: 0.9516 - val_accuracy: 0.7462\n",
      "Epoch 22/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.6043 - accuracy: 0.8269 - val_loss: 0.9326 - val_accuracy: 0.7487\n",
      "Epoch 23/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.5831 - accuracy: 0.8405 - val_loss: 0.9117 - val_accuracy: 0.7562\n",
      "Epoch 24/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.5719 - accuracy: 0.8372 - val_loss: 0.8954 - val_accuracy: 0.7575\n",
      "Epoch 25/1000\n",
      "19/19 [==============================] - 0s 18ms/step - loss: 0.5608 - accuracy: 0.8379 - val_loss: 0.8865 - val_accuracy: 0.7537\n",
      "Epoch 26/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.5484 - accuracy: 0.8439 - val_loss: 0.8764 - val_accuracy: 0.7600\n",
      "Epoch 27/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.5254 - accuracy: 0.8520 - val_loss: 0.8676 - val_accuracy: 0.7675\n",
      "Epoch 28/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 0.5203 - accuracy: 0.8520 - val_loss: 0.8571 - val_accuracy: 0.7713\n",
      "Epoch 29/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.5030 - accuracy: 0.8565 - val_loss: 0.8460 - val_accuracy: 0.7812\n",
      "Epoch 30/1000\n",
      "19/19 [==============================] - 0s 24ms/step - loss: 0.4889 - accuracy: 0.8641 - val_loss: 0.8392 - val_accuracy: 0.7812\n",
      "Epoch 31/1000\n",
      "19/19 [==============================] - 0s 22ms/step - loss: 0.4803 - accuracy: 0.8651 - val_loss: 0.8385 - val_accuracy: 0.7825\n",
      "Epoch 32/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 0.4714 - accuracy: 0.8681 - val_loss: 0.8350 - val_accuracy: 0.7812\n",
      "Epoch 33/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.4475 - accuracy: 0.8758 - val_loss: 0.8349 - val_accuracy: 0.7837\n",
      "Epoch 34/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.4509 - accuracy: 0.8775 - val_loss: 0.8361 - val_accuracy: 0.7825\n",
      "Epoch 35/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 0.4369 - accuracy: 0.8784 - val_loss: 0.8366 - val_accuracy: 0.7837\n",
      "Epoch 36/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.4365 - accuracy: 0.8786 - val_loss: 0.8383 - val_accuracy: 0.7862\n",
      "Epoch 37/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.4296 - accuracy: 0.8809 - val_loss: 0.8416 - val_accuracy: 0.7800\n",
      "Epoch 38/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.4170 - accuracy: 0.8840 - val_loss: 0.8403 - val_accuracy: 0.7837\n",
      "Epoch 39/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.4164 - accuracy: 0.8838 - val_loss: 0.8394 - val_accuracy: 0.7862\n",
      "Epoch 40/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.4011 - accuracy: 0.8909 - val_loss: 0.8428 - val_accuracy: 0.7925\n",
      "Epoch 41/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3991 - accuracy: 0.8916 - val_loss: 0.8421 - val_accuracy: 0.7900\n",
      "Epoch 42/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3859 - accuracy: 0.8941 - val_loss: 0.8440 - val_accuracy: 0.7912\n",
      "Epoch 43/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3843 - accuracy: 0.8936 - val_loss: 0.8471 - val_accuracy: 0.7887\n",
      "Epoch 44/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3832 - accuracy: 0.8973 - val_loss: 0.8489 - val_accuracy: 0.7900\n",
      "Epoch 45/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3734 - accuracy: 0.8991 - val_loss: 0.8502 - val_accuracy: 0.7900\n",
      "Epoch 46/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3642 - accuracy: 0.9013 - val_loss: 0.8529 - val_accuracy: 0.7950\n",
      "Epoch 47/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3595 - accuracy: 0.9024 - val_loss: 0.8566 - val_accuracy: 0.7925\n",
      "Epoch 48/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3586 - accuracy: 0.9032 - val_loss: 0.8590 - val_accuracy: 0.7950\n",
      "Epoch 49/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3546 - accuracy: 0.9064 - val_loss: 0.8657 - val_accuracy: 0.7937\n",
      "Epoch 50/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3480 - accuracy: 0.9066 - val_loss: 0.8669 - val_accuracy: 0.8012\n",
      "Epoch 51/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3427 - accuracy: 0.9091 - val_loss: 0.8699 - val_accuracy: 0.7987\n",
      "Epoch 52/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3359 - accuracy: 0.9096 - val_loss: 0.8694 - val_accuracy: 0.7962\n",
      "Epoch 53/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3241 - accuracy: 0.9141 - val_loss: 0.8697 - val_accuracy: 0.7975\n",
      "Epoch 54/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3209 - accuracy: 0.9162 - val_loss: 0.8730 - val_accuracy: 0.7975\n",
      "Epoch 55/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.3228 - accuracy: 0.9137 - val_loss: 0.8794 - val_accuracy: 0.8000\n",
      "Epoch 56/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.3170 - accuracy: 0.9181 - val_loss: 0.8805 - val_accuracy: 0.7975\n",
      "Epoch 57/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.3200 - accuracy: 0.9168 - val_loss: 0.8767 - val_accuracy: 0.8012\n",
      "Epoch 58/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.3225 - accuracy: 0.9167 - val_loss: 0.8778 - val_accuracy: 0.8012\n",
      "Epoch 59/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 15ms/step - loss: 0.3092 - accuracy: 0.9189 - val_loss: 0.8812 - val_accuracy: 0.7962\n",
      "Epoch 60/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.2987 - accuracy: 0.9241 - val_loss: 0.8809 - val_accuracy: 0.7937\n",
      "Epoch 61/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2980 - accuracy: 0.9232 - val_loss: 0.8818 - val_accuracy: 0.7912\n",
      "Epoch 62/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2979 - accuracy: 0.9225 - val_loss: 0.8807 - val_accuracy: 0.7862\n",
      "Epoch 63/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2939 - accuracy: 0.9264 - val_loss: 0.8822 - val_accuracy: 0.7850\n",
      "Epoch 64/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2854 - accuracy: 0.9269 - val_loss: 0.8813 - val_accuracy: 0.7887\n",
      "Epoch 65/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2895 - accuracy: 0.9263 - val_loss: 0.8806 - val_accuracy: 0.7887\n",
      "Epoch 66/1000\n",
      "19/19 [==============================] - 0s 14ms/step - loss: 0.2843 - accuracy: 0.9264 - val_loss: 0.8799 - val_accuracy: 0.7825\n",
      "Epoch 67/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2848 - accuracy: 0.9288 - val_loss: 0.8801 - val_accuracy: 0.7837\n",
      "Epoch 68/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2788 - accuracy: 0.9296 - val_loss: 0.8791 - val_accuracy: 0.7875\n",
      "Epoch 69/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2732 - accuracy: 0.9299 - val_loss: 0.8803 - val_accuracy: 0.7900\n",
      "Epoch 70/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2702 - accuracy: 0.9338 - val_loss: 0.8840 - val_accuracy: 0.7900\n",
      "Epoch 71/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2668 - accuracy: 0.9325 - val_loss: 0.8851 - val_accuracy: 0.7912\n",
      "Epoch 72/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2655 - accuracy: 0.9347 - val_loss: 0.8843 - val_accuracy: 0.7925\n",
      "Epoch 73/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2530 - accuracy: 0.9380 - val_loss: 0.8859 - val_accuracy: 0.7862\n",
      "Epoch 74/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.2582 - accuracy: 0.9372 - val_loss: 0.8862 - val_accuracy: 0.7912\n",
      "Epoch 75/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2511 - accuracy: 0.9399 - val_loss: 0.8841 - val_accuracy: 0.7900\n",
      "Epoch 76/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2526 - accuracy: 0.9398 - val_loss: 0.8911 - val_accuracy: 0.7875\n",
      "Epoch 77/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2529 - accuracy: 0.9372 - val_loss: 0.8981 - val_accuracy: 0.7837\n",
      "Epoch 78/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2452 - accuracy: 0.9407 - val_loss: 0.8988 - val_accuracy: 0.7850\n",
      "Epoch 79/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2317 - accuracy: 0.9462 - val_loss: 0.9031 - val_accuracy: 0.7862\n",
      "Epoch 80/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2403 - accuracy: 0.9435 - val_loss: 0.9053 - val_accuracy: 0.7800\n",
      "Epoch 81/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2486 - accuracy: 0.9409 - val_loss: 0.9036 - val_accuracy: 0.7875\n",
      "Epoch 82/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2380 - accuracy: 0.9432 - val_loss: 0.9037 - val_accuracy: 0.7912\n",
      "Epoch 83/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2309 - accuracy: 0.9473 - val_loss: 0.9064 - val_accuracy: 0.7912\n",
      "Epoch 84/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2324 - accuracy: 0.9461 - val_loss: 0.9097 - val_accuracy: 0.7887\n",
      "Epoch 85/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2396 - accuracy: 0.9429 - val_loss: 0.9133 - val_accuracy: 0.7887\n",
      "Epoch 86/1000\n",
      "19/19 [==============================] - 0s 17ms/step - loss: 0.2273 - accuracy: 0.9466 - val_loss: 0.9134 - val_accuracy: 0.7900\n",
      "Epoch 87/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.2236 - accuracy: 0.9471 - val_loss: 0.9110 - val_accuracy: 0.7862\n",
      "Epoch 88/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.2237 - accuracy: 0.9475 - val_loss: 0.9130 - val_accuracy: 0.7912\n",
      "Epoch 89/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2214 - accuracy: 0.9525 - val_loss: 0.9104 - val_accuracy: 0.7900\n",
      "Epoch 90/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.2123 - accuracy: 0.9531 - val_loss: 0.9145 - val_accuracy: 0.7887\n",
      "Epoch 91/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2205 - accuracy: 0.9484 - val_loss: 0.9171 - val_accuracy: 0.7887\n",
      "Epoch 92/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2077 - accuracy: 0.9552 - val_loss: 0.9204 - val_accuracy: 0.7875\n",
      "Epoch 93/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2057 - accuracy: 0.9545 - val_loss: 0.9173 - val_accuracy: 0.7900\n",
      "Epoch 94/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2152 - accuracy: 0.9522 - val_loss: 0.9180 - val_accuracy: 0.7925\n",
      "Epoch 95/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2128 - accuracy: 0.9495 - val_loss: 0.9152 - val_accuracy: 0.7900\n",
      "Epoch 96/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2090 - accuracy: 0.9534 - val_loss: 0.9120 - val_accuracy: 0.7900\n",
      "Epoch 97/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.2065 - accuracy: 0.9556 - val_loss: 0.9109 - val_accuracy: 0.7912\n",
      "Epoch 98/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2055 - accuracy: 0.9567 - val_loss: 0.9127 - val_accuracy: 0.7875\n",
      "Epoch 99/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.2007 - accuracy: 0.9566 - val_loss: 0.9149 - val_accuracy: 0.7862\n",
      "Epoch 100/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.2002 - accuracy: 0.9547 - val_loss: 0.9134 - val_accuracy: 0.7850\n",
      "Epoch 101/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.2007 - accuracy: 0.9564 - val_loss: 0.9127 - val_accuracy: 0.7825\n",
      "Epoch 102/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1986 - accuracy: 0.9570 - val_loss: 0.9149 - val_accuracy: 0.7812\n",
      "Epoch 103/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1968 - accuracy: 0.9584 - val_loss: 0.9149 - val_accuracy: 0.7850\n",
      "Epoch 104/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1942 - accuracy: 0.9574 - val_loss: 0.9133 - val_accuracy: 0.7925\n",
      "Epoch 105/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1926 - accuracy: 0.9593 - val_loss: 0.9114 - val_accuracy: 0.7937\n",
      "Epoch 106/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1989 - accuracy: 0.9569 - val_loss: 0.9101 - val_accuracy: 0.7862\n",
      "Epoch 107/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1931 - accuracy: 0.9585 - val_loss: 0.9081 - val_accuracy: 0.7900\n",
      "Epoch 108/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1878 - accuracy: 0.9628 - val_loss: 0.9079 - val_accuracy: 0.7937\n",
      "Epoch 109/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1811 - accuracy: 0.9638 - val_loss: 0.9119 - val_accuracy: 0.7950\n",
      "Epoch 110/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1825 - accuracy: 0.9636 - val_loss: 0.9154 - val_accuracy: 0.7937\n",
      "Epoch 111/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1855 - accuracy: 0.9615 - val_loss: 0.9156 - val_accuracy: 0.7950\n",
      "Epoch 112/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1822 - accuracy: 0.9626 - val_loss: 0.9125 - val_accuracy: 0.7975\n",
      "Epoch 113/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1809 - accuracy: 0.9618 - val_loss: 0.9134 - val_accuracy: 0.7900\n",
      "Epoch 114/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1774 - accuracy: 0.9650 - val_loss: 0.9121 - val_accuracy: 0.7925\n",
      "Epoch 115/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1764 - accuracy: 0.9651 - val_loss: 0.9125 - val_accuracy: 0.7962\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 116/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1855 - accuracy: 0.9603 - val_loss: 0.9140 - val_accuracy: 0.7950\n",
      "Epoch 117/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1737 - accuracy: 0.9661 - val_loss: 0.9128 - val_accuracy: 0.7937\n",
      "Epoch 118/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1740 - accuracy: 0.9656 - val_loss: 0.9158 - val_accuracy: 0.7962\n",
      "Epoch 119/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1749 - accuracy: 0.9659 - val_loss: 0.9149 - val_accuracy: 0.8000\n",
      "Epoch 120/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1743 - accuracy: 0.9662 - val_loss: 0.9151 - val_accuracy: 0.7987\n",
      "Epoch 121/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1696 - accuracy: 0.9678 - val_loss: 0.9128 - val_accuracy: 0.7925\n",
      "Epoch 122/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1714 - accuracy: 0.9667 - val_loss: 0.9175 - val_accuracy: 0.7925\n",
      "Epoch 123/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1691 - accuracy: 0.9664 - val_loss: 0.9190 - val_accuracy: 0.7962\n",
      "Epoch 124/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1620 - accuracy: 0.9697 - val_loss: 0.9196 - val_accuracy: 0.7950\n",
      "Epoch 125/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1712 - accuracy: 0.9670 - val_loss: 0.9210 - val_accuracy: 0.7925\n",
      "Epoch 126/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1610 - accuracy: 0.9703 - val_loss: 0.9235 - val_accuracy: 0.7912\n",
      "Epoch 127/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1600 - accuracy: 0.9702 - val_loss: 0.9221 - val_accuracy: 0.7875\n",
      "Epoch 128/1000\n",
      "19/19 [==============================] - 0s 17ms/step - loss: 0.1600 - accuracy: 0.9687 - val_loss: 0.9169 - val_accuracy: 0.7887\n",
      "Epoch 129/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1614 - accuracy: 0.9693 - val_loss: 0.9163 - val_accuracy: 0.7925\n",
      "Epoch 130/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1573 - accuracy: 0.9705 - val_loss: 0.9150 - val_accuracy: 0.7912\n",
      "Epoch 131/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1583 - accuracy: 0.9723 - val_loss: 0.9132 - val_accuracy: 0.7925\n",
      "Epoch 132/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1608 - accuracy: 0.9717 - val_loss: 0.9138 - val_accuracy: 0.7975\n",
      "Epoch 133/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1626 - accuracy: 0.9724 - val_loss: 0.9085 - val_accuracy: 0.7925\n",
      "Epoch 134/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1579 - accuracy: 0.9724 - val_loss: 0.9114 - val_accuracy: 0.7850\n",
      "Epoch 135/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1563 - accuracy: 0.9712 - val_loss: 0.9108 - val_accuracy: 0.7875\n",
      "Epoch 136/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1541 - accuracy: 0.9718 - val_loss: 0.9099 - val_accuracy: 0.7900\n",
      "Epoch 137/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1552 - accuracy: 0.9710 - val_loss: 0.9108 - val_accuracy: 0.7950\n",
      "Epoch 138/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1538 - accuracy: 0.9726 - val_loss: 0.9092 - val_accuracy: 0.7900\n",
      "Epoch 139/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1555 - accuracy: 0.9721 - val_loss: 0.9084 - val_accuracy: 0.7900\n",
      "Epoch 140/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1527 - accuracy: 0.9726 - val_loss: 0.9065 - val_accuracy: 0.7875\n",
      "Epoch 141/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1479 - accuracy: 0.9732 - val_loss: 0.9033 - val_accuracy: 0.7875\n",
      "Epoch 142/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1506 - accuracy: 0.9739 - val_loss: 0.9058 - val_accuracy: 0.7837\n",
      "Epoch 143/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1495 - accuracy: 0.9746 - val_loss: 0.9077 - val_accuracy: 0.7837\n",
      "Epoch 144/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1494 - accuracy: 0.9747 - val_loss: 0.9107 - val_accuracy: 0.7850\n",
      "Epoch 145/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1462 - accuracy: 0.9753 - val_loss: 0.9098 - val_accuracy: 0.7850\n",
      "Epoch 146/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1549 - accuracy: 0.9712 - val_loss: 0.9067 - val_accuracy: 0.7900\n",
      "Epoch 147/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1430 - accuracy: 0.9768 - val_loss: 0.9116 - val_accuracy: 0.7900\n",
      "Epoch 148/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1428 - accuracy: 0.9755 - val_loss: 0.9115 - val_accuracy: 0.7912\n",
      "Epoch 149/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1435 - accuracy: 0.9757 - val_loss: 0.9079 - val_accuracy: 0.7962\n",
      "Epoch 150/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1400 - accuracy: 0.9768 - val_loss: 0.9081 - val_accuracy: 0.7912\n",
      "Epoch 151/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1447 - accuracy: 0.9736 - val_loss: 0.9072 - val_accuracy: 0.7925\n",
      "Epoch 152/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1421 - accuracy: 0.9754 - val_loss: 0.9122 - val_accuracy: 0.7962\n",
      "Epoch 153/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1437 - accuracy: 0.9763 - val_loss: 0.9152 - val_accuracy: 0.7950\n",
      "Epoch 154/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1393 - accuracy: 0.9776 - val_loss: 0.9162 - val_accuracy: 0.7950\n",
      "Epoch 155/1000\n",
      "19/19 [==============================] - 0s 17ms/step - loss: 0.1391 - accuracy: 0.9789 - val_loss: 0.9181 - val_accuracy: 0.7950\n",
      "Epoch 156/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1424 - accuracy: 0.9771 - val_loss: 0.9199 - val_accuracy: 0.8000\n",
      "Epoch 157/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1428 - accuracy: 0.9752 - val_loss: 0.9223 - val_accuracy: 0.7950\n",
      "Epoch 158/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1349 - accuracy: 0.9779 - val_loss: 0.9199 - val_accuracy: 0.7975\n",
      "Epoch 159/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1402 - accuracy: 0.9769 - val_loss: 0.9191 - val_accuracy: 0.7937\n",
      "Epoch 160/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1322 - accuracy: 0.9788 - val_loss: 0.9178 - val_accuracy: 0.7937\n",
      "Epoch 161/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1378 - accuracy: 0.9775 - val_loss: 0.9158 - val_accuracy: 0.7900\n",
      "Epoch 162/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1366 - accuracy: 0.9786 - val_loss: 0.9165 - val_accuracy: 0.7950\n",
      "Epoch 163/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1309 - accuracy: 0.9806 - val_loss: 0.9186 - val_accuracy: 0.7937\n",
      "Epoch 164/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1345 - accuracy: 0.9786 - val_loss: 0.9208 - val_accuracy: 0.7975\n",
      "Epoch 165/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1335 - accuracy: 0.9784 - val_loss: 0.9237 - val_accuracy: 0.7950\n",
      "Epoch 166/1000\n",
      "19/19 [==============================] - 0s 17ms/step - loss: 0.1304 - accuracy: 0.9793 - val_loss: 0.9257 - val_accuracy: 0.7937\n",
      "Epoch 167/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1348 - accuracy: 0.9775 - val_loss: 0.9235 - val_accuracy: 0.7912\n",
      "Epoch 168/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1334 - accuracy: 0.9787 - val_loss: 0.9244 - val_accuracy: 0.7937\n",
      "Epoch 169/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1347 - accuracy: 0.9786 - val_loss: 0.9244 - val_accuracy: 0.7925\n",
      "Epoch 170/1000\n",
      "19/19 [==============================] - 0s 17ms/step - loss: 0.1292 - accuracy: 0.9814 - val_loss: 0.9255 - val_accuracy: 0.7962\n",
      "Epoch 171/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1315 - accuracy: 0.9797 - val_loss: 0.9285 - val_accuracy: 0.7950\n",
      "Epoch 172/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1314 - accuracy: 0.9809 - val_loss: 0.9302 - val_accuracy: 0.7887\n",
      "Epoch 173/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1255 - accuracy: 0.9807 - val_loss: 0.9340 - val_accuracy: 0.7900\n",
      "Epoch 174/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1309 - accuracy: 0.9803 - val_loss: 0.9335 - val_accuracy: 0.7925\n",
      "Epoch 175/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1340 - accuracy: 0.9793 - val_loss: 0.9366 - val_accuracy: 0.7950\n",
      "Epoch 176/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1294 - accuracy: 0.9807 - val_loss: 0.9379 - val_accuracy: 0.7900\n",
      "Epoch 177/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1265 - accuracy: 0.9820 - val_loss: 0.9390 - val_accuracy: 0.7887\n",
      "Epoch 178/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1300 - accuracy: 0.9801 - val_loss: 0.9409 - val_accuracy: 0.7925\n",
      "Epoch 179/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1322 - accuracy: 0.9799 - val_loss: 0.9403 - val_accuracy: 0.7900\n",
      "Epoch 180/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1260 - accuracy: 0.9833 - val_loss: 0.9460 - val_accuracy: 0.7900\n",
      "Epoch 181/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1269 - accuracy: 0.9819 - val_loss: 0.9483 - val_accuracy: 0.7975\n",
      "Epoch 182/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1307 - accuracy: 0.9809 - val_loss: 0.9476 - val_accuracy: 0.7937\n",
      "Epoch 183/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1282 - accuracy: 0.9806 - val_loss: 0.9446 - val_accuracy: 0.7937\n",
      "Epoch 184/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1240 - accuracy: 0.9829 - val_loss: 0.9406 - val_accuracy: 0.7962\n",
      "Epoch 185/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1287 - accuracy: 0.9804 - val_loss: 0.9398 - val_accuracy: 0.7925\n",
      "Epoch 186/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1255 - accuracy: 0.9818 - val_loss: 0.9421 - val_accuracy: 0.7937\n",
      "Epoch 187/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1283 - accuracy: 0.9817 - val_loss: 0.9434 - val_accuracy: 0.7900\n",
      "Epoch 188/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1242 - accuracy: 0.9828 - val_loss: 0.9437 - val_accuracy: 0.7950\n",
      "Epoch 189/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1212 - accuracy: 0.9843 - val_loss: 0.9468 - val_accuracy: 0.7925\n",
      "Epoch 190/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1265 - accuracy: 0.9809 - val_loss: 0.9436 - val_accuracy: 0.7912\n",
      "Epoch 191/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1240 - accuracy: 0.9815 - val_loss: 0.9415 - val_accuracy: 0.7987\n",
      "Epoch 192/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1213 - accuracy: 0.9843 - val_loss: 0.9462 - val_accuracy: 0.7975\n",
      "Epoch 193/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1248 - accuracy: 0.9830 - val_loss: 0.9543 - val_accuracy: 0.7937\n",
      "Epoch 194/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1243 - accuracy: 0.9799 - val_loss: 0.9544 - val_accuracy: 0.7962\n",
      "Epoch 195/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1242 - accuracy: 0.9813 - val_loss: 0.9520 - val_accuracy: 0.7900\n",
      "Epoch 196/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1213 - accuracy: 0.9826 - val_loss: 0.9518 - val_accuracy: 0.7912\n",
      "Epoch 197/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1214 - accuracy: 0.9823 - val_loss: 0.9534 - val_accuracy: 0.7900\n",
      "Epoch 198/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1219 - accuracy: 0.9831 - val_loss: 0.9495 - val_accuracy: 0.7925\n",
      "Epoch 199/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1227 - accuracy: 0.9830 - val_loss: 0.9466 - val_accuracy: 0.7950\n",
      "Epoch 200/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1188 - accuracy: 0.9842 - val_loss: 0.9482 - val_accuracy: 0.7975\n",
      "Epoch 201/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1215 - accuracy: 0.9839 - val_loss: 0.9465 - val_accuracy: 0.7925\n",
      "Epoch 202/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1170 - accuracy: 0.9864 - val_loss: 0.9482 - val_accuracy: 0.7937\n",
      "Epoch 203/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1256 - accuracy: 0.9818 - val_loss: 0.9491 - val_accuracy: 0.7937\n",
      "Epoch 204/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1234 - accuracy: 0.9818 - val_loss: 0.9421 - val_accuracy: 0.7975\n",
      "Epoch 205/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1195 - accuracy: 0.9835 - val_loss: 0.9382 - val_accuracy: 0.7975\n",
      "Epoch 206/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1209 - accuracy: 0.9839 - val_loss: 0.9376 - val_accuracy: 0.7962\n",
      "Epoch 207/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1168 - accuracy: 0.9855 - val_loss: 0.9352 - val_accuracy: 0.7912\n",
      "Epoch 208/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1158 - accuracy: 0.9853 - val_loss: 0.9338 - val_accuracy: 0.7925\n",
      "Epoch 209/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1148 - accuracy: 0.9854 - val_loss: 0.9383 - val_accuracy: 0.7912\n",
      "Epoch 210/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1168 - accuracy: 0.9844 - val_loss: 0.9367 - val_accuracy: 0.7887\n",
      "Epoch 211/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1188 - accuracy: 0.9837 - val_loss: 0.9269 - val_accuracy: 0.7937\n",
      "Epoch 212/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1184 - accuracy: 0.9849 - val_loss: 0.9245 - val_accuracy: 0.7987\n",
      "Epoch 213/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1191 - accuracy: 0.9843 - val_loss: 0.9291 - val_accuracy: 0.8012\n",
      "Epoch 214/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1188 - accuracy: 0.9835 - val_loss: 0.9292 - val_accuracy: 0.7950\n",
      "Epoch 215/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1134 - accuracy: 0.9852 - val_loss: 0.9287 - val_accuracy: 0.7925\n",
      "Epoch 216/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1159 - accuracy: 0.9849 - val_loss: 0.9283 - val_accuracy: 0.7962\n",
      "Epoch 217/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1182 - accuracy: 0.9842 - val_loss: 0.9296 - val_accuracy: 0.7950\n",
      "Epoch 218/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1189 - accuracy: 0.9841 - val_loss: 0.9324 - val_accuracy: 0.7975\n",
      "Epoch 219/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1127 - accuracy: 0.9865 - val_loss: 0.9337 - val_accuracy: 0.7925\n",
      "Epoch 220/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1120 - accuracy: 0.9857 - val_loss: 0.9325 - val_accuracy: 0.7912\n",
      "Epoch 221/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1158 - accuracy: 0.9845 - val_loss: 0.9297 - val_accuracy: 0.7950\n",
      "Epoch 222/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1102 - accuracy: 0.9866 - val_loss: 0.9319 - val_accuracy: 0.7975\n",
      "Epoch 223/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1099 - accuracy: 0.9875 - val_loss: 0.9378 - val_accuracy: 0.7975\n",
      "Epoch 224/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1188 - accuracy: 0.9840 - val_loss: 0.9389 - val_accuracy: 0.7887\n",
      "Epoch 225/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1118 - accuracy: 0.9861 - val_loss: 0.9410 - val_accuracy: 0.7912\n",
      "Epoch 226/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1146 - accuracy: 0.9862 - val_loss: 0.9371 - val_accuracy: 0.7912\n",
      "Epoch 227/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1165 - accuracy: 0.9847 - val_loss: 0.9337 - val_accuracy: 0.7962\n",
      "Epoch 228/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1097 - accuracy: 0.9856 - val_loss: 0.9341 - val_accuracy: 0.7962\n",
      "Epoch 229/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1139 - accuracy: 0.9857 - val_loss: 0.9336 - val_accuracy: 0.8025\n",
      "Epoch 230/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1131 - accuracy: 0.9854 - val_loss: 0.9381 - val_accuracy: 0.7925\n",
      "Epoch 231/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1126 - accuracy: 0.9865 - val_loss: 0.9429 - val_accuracy: 0.7925\n",
      "Epoch 232/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1114 - accuracy: 0.9857 - val_loss: 0.9420 - val_accuracy: 0.7937\n",
      "Epoch 233/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1142 - accuracy: 0.9856 - val_loss: 0.9472 - val_accuracy: 0.7925\n",
      "Epoch 234/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1161 - accuracy: 0.9852 - val_loss: 0.9465 - val_accuracy: 0.7912\n",
      "Epoch 235/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1155 - accuracy: 0.9845 - val_loss: 0.9490 - val_accuracy: 0.7925\n",
      "Epoch 236/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1138 - accuracy: 0.9853 - val_loss: 0.9453 - val_accuracy: 0.7925\n",
      "Epoch 237/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1069 - accuracy: 0.9874 - val_loss: 0.9448 - val_accuracy: 0.7950\n",
      "Epoch 238/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1097 - accuracy: 0.9865 - val_loss: 0.9456 - val_accuracy: 0.7925\n",
      "Epoch 239/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1081 - accuracy: 0.9877 - val_loss: 0.9495 - val_accuracy: 0.7925\n",
      "Epoch 240/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1103 - accuracy: 0.9874 - val_loss: 0.9519 - val_accuracy: 0.7937\n",
      "Epoch 241/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1115 - accuracy: 0.9862 - val_loss: 0.9472 - val_accuracy: 0.7937\n",
      "Epoch 242/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1087 - accuracy: 0.9876 - val_loss: 0.9462 - val_accuracy: 0.7962\n",
      "Epoch 243/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1116 - accuracy: 0.9873 - val_loss: 0.9452 - val_accuracy: 0.7925\n",
      "Epoch 244/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1112 - accuracy: 0.9859 - val_loss: 0.9461 - val_accuracy: 0.7950\n",
      "Epoch 245/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1051 - accuracy: 0.9886 - val_loss: 0.9554 - val_accuracy: 0.7875\n",
      "Epoch 246/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1107 - accuracy: 0.9872 - val_loss: 0.9576 - val_accuracy: 0.7887\n",
      "Epoch 247/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1103 - accuracy: 0.9864 - val_loss: 0.9610 - val_accuracy: 0.7900\n",
      "Epoch 248/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1092 - accuracy: 0.9876 - val_loss: 0.9652 - val_accuracy: 0.7825\n",
      "Epoch 249/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1065 - accuracy: 0.9884 - val_loss: 0.9666 - val_accuracy: 0.7875\n",
      "Epoch 250/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1102 - accuracy: 0.9874 - val_loss: 0.9670 - val_accuracy: 0.7875\n",
      "Epoch 251/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1141 - accuracy: 0.9857 - val_loss: 0.9663 - val_accuracy: 0.7812\n",
      "Epoch 252/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1093 - accuracy: 0.9866 - val_loss: 0.9670 - val_accuracy: 0.7788\n",
      "Epoch 253/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1092 - accuracy: 0.9877 - val_loss: 0.9674 - val_accuracy: 0.7825\n",
      "Epoch 254/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1113 - accuracy: 0.9862 - val_loss: 0.9632 - val_accuracy: 0.7850\n",
      "Epoch 255/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1124 - accuracy: 0.9865 - val_loss: 0.9648 - val_accuracy: 0.7887\n",
      "Epoch 256/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1101 - accuracy: 0.9887 - val_loss: 0.9631 - val_accuracy: 0.7887\n",
      "Epoch 257/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1104 - accuracy: 0.9876 - val_loss: 0.9613 - val_accuracy: 0.7912\n",
      "Epoch 258/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1087 - accuracy: 0.9871 - val_loss: 0.9588 - val_accuracy: 0.7875\n",
      "Epoch 259/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1073 - accuracy: 0.9871 - val_loss: 0.9576 - val_accuracy: 0.7887\n",
      "Epoch 260/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1074 - accuracy: 0.9886 - val_loss: 0.9591 - val_accuracy: 0.7862\n",
      "Epoch 261/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1103 - accuracy: 0.9866 - val_loss: 0.9556 - val_accuracy: 0.7887\n",
      "Epoch 262/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1101 - accuracy: 0.9884 - val_loss: 0.9557 - val_accuracy: 0.7887\n",
      "Epoch 263/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1077 - accuracy: 0.9876 - val_loss: 0.9550 - val_accuracy: 0.7925\n",
      "Epoch 264/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1085 - accuracy: 0.9874 - val_loss: 0.9532 - val_accuracy: 0.7937\n",
      "Epoch 265/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1086 - accuracy: 0.9871 - val_loss: 0.9559 - val_accuracy: 0.7925\n",
      "Epoch 266/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1097 - accuracy: 0.9867 - val_loss: 0.9585 - val_accuracy: 0.7887\n",
      "Epoch 267/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1107 - accuracy: 0.9874 - val_loss: 0.9592 - val_accuracy: 0.7912\n",
      "Epoch 268/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1102 - accuracy: 0.9866 - val_loss: 0.9620 - val_accuracy: 0.7850\n",
      "Epoch 269/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1122 - accuracy: 0.9868 - val_loss: 0.9678 - val_accuracy: 0.7862\n",
      "Epoch 270/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1076 - accuracy: 0.9898 - val_loss: 0.9681 - val_accuracy: 0.7887\n",
      "Epoch 271/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1091 - accuracy: 0.9868 - val_loss: 0.9698 - val_accuracy: 0.7862\n",
      "Epoch 272/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1075 - accuracy: 0.9876 - val_loss: 0.9728 - val_accuracy: 0.7912\n",
      "Epoch 273/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1050 - accuracy: 0.9882 - val_loss: 0.9753 - val_accuracy: 0.7875\n",
      "Epoch 274/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1063 - accuracy: 0.9887 - val_loss: 0.9752 - val_accuracy: 0.7850\n",
      "Epoch 275/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1083 - accuracy: 0.9873 - val_loss: 0.9747 - val_accuracy: 0.7887\n",
      "Epoch 276/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1070 - accuracy: 0.9880 - val_loss: 0.9771 - val_accuracy: 0.7900\n",
      "Epoch 277/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1069 - accuracy: 0.9882 - val_loss: 0.9807 - val_accuracy: 0.7925\n",
      "Epoch 278/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1081 - accuracy: 0.9882 - val_loss: 0.9777 - val_accuracy: 0.7912\n",
      "Epoch 279/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1064 - accuracy: 0.9888 - val_loss: 0.9709 - val_accuracy: 0.7925\n",
      "Epoch 280/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1096 - accuracy: 0.9878 - val_loss: 0.9741 - val_accuracy: 0.7950\n",
      "Epoch 281/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1092 - accuracy: 0.9872 - val_loss: 0.9766 - val_accuracy: 0.7925\n",
      "Epoch 282/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.1032 - accuracy: 0.9897 - val_loss: 0.9763 - val_accuracy: 0.7925\n",
      "Epoch 283/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1077 - accuracy: 0.9878 - val_loss: 0.9753 - val_accuracy: 0.7925\n",
      "Epoch 284/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1063 - accuracy: 0.9891 - val_loss: 0.9738 - val_accuracy: 0.7900\n",
      "Epoch 285/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1111 - accuracy: 0.9869 - val_loss: 0.9684 - val_accuracy: 0.7912\n",
      "Epoch 286/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1076 - accuracy: 0.9880 - val_loss: 0.9641 - val_accuracy: 0.7900\n",
      "Epoch 287/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1090 - accuracy: 0.9882 - val_loss: 0.9651 - val_accuracy: 0.7925\n",
      "Epoch 288/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1072 - accuracy: 0.9886 - val_loss: 0.9651 - val_accuracy: 0.7875\n",
      "Epoch 289/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1072 - accuracy: 0.9880 - val_loss: 0.9676 - val_accuracy: 0.7912\n",
      "Epoch 290/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1082 - accuracy: 0.9882 - val_loss: 0.9644 - val_accuracy: 0.7912\n",
      "Epoch 291/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1108 - accuracy: 0.9871 - val_loss: 0.9691 - val_accuracy: 0.7912\n",
      "Epoch 292/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1090 - accuracy: 0.9878 - val_loss: 0.9677 - val_accuracy: 0.7925\n",
      "Epoch 293/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1071 - accuracy: 0.9876 - val_loss: 0.9613 - val_accuracy: 0.7950\n",
      "Epoch 294/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1079 - accuracy: 0.9882 - val_loss: 0.9546 - val_accuracy: 0.7975\n",
      "Epoch 295/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1064 - accuracy: 0.9884 - val_loss: 0.9507 - val_accuracy: 0.8012\n",
      "Epoch 296/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1063 - accuracy: 0.9889 - val_loss: 0.9523 - val_accuracy: 0.8000\n",
      "Epoch 297/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1033 - accuracy: 0.9897 - val_loss: 0.9463 - val_accuracy: 0.7975\n",
      "Epoch 298/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.1030 - accuracy: 0.9894 - val_loss: 0.9478 - val_accuracy: 0.7962\n",
      "Epoch 299/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1052 - accuracy: 0.9897 - val_loss: 0.9507 - val_accuracy: 0.7987\n",
      "Epoch 300/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0992 - accuracy: 0.9907 - val_loss: 0.9535 - val_accuracy: 0.7975\n",
      "Epoch 301/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1053 - accuracy: 0.9889 - val_loss: 0.9544 - val_accuracy: 0.7950\n",
      "Epoch 302/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1044 - accuracy: 0.9886 - val_loss: 0.9505 - val_accuracy: 0.8012\n",
      "Epoch 303/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1036 - accuracy: 0.9888 - val_loss: 0.9503 - val_accuracy: 0.8012\n",
      "Epoch 304/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1066 - accuracy: 0.9878 - val_loss: 0.9509 - val_accuracy: 0.8025\n",
      "Epoch 305/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1006 - accuracy: 0.9904 - val_loss: 0.9535 - val_accuracy: 0.7987\n",
      "Epoch 306/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1039 - accuracy: 0.9897 - val_loss: 0.9575 - val_accuracy: 0.8025\n",
      "Epoch 307/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9900 - val_loss: 0.9561 - val_accuracy: 0.7987\n",
      "Epoch 308/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1085 - accuracy: 0.9874 - val_loss: 0.9574 - val_accuracy: 0.8012\n",
      "Epoch 309/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 0.9905 - val_loss: 0.9543 - val_accuracy: 0.8037\n",
      "Epoch 310/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1054 - accuracy: 0.9883 - val_loss: 0.9534 - val_accuracy: 0.8000\n",
      "Epoch 311/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1051 - accuracy: 0.9895 - val_loss: 0.9550 - val_accuracy: 0.7962\n",
      "Epoch 312/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1057 - accuracy: 0.9887 - val_loss: 0.9479 - val_accuracy: 0.7950\n",
      "Epoch 313/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1033 - accuracy: 0.9888 - val_loss: 0.9447 - val_accuracy: 0.7950\n",
      "Epoch 314/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1050 - accuracy: 0.9893 - val_loss: 0.9491 - val_accuracy: 0.7975\n",
      "Epoch 315/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1032 - accuracy: 0.9890 - val_loss: 0.9542 - val_accuracy: 0.7962\n",
      "Epoch 316/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1033 - accuracy: 0.9898 - val_loss: 0.9593 - val_accuracy: 0.8062\n",
      "Epoch 317/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1078 - accuracy: 0.9881 - val_loss: 0.9624 - val_accuracy: 0.8012\n",
      "Epoch 318/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1039 - accuracy: 0.9896 - val_loss: 0.9629 - val_accuracy: 0.8000\n",
      "Epoch 319/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1048 - accuracy: 0.9897 - val_loss: 0.9630 - val_accuracy: 0.8050\n",
      "Epoch 320/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1049 - accuracy: 0.9886 - val_loss: 0.9617 - val_accuracy: 0.8050\n",
      "Epoch 321/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1023 - accuracy: 0.9897 - val_loss: 0.9591 - val_accuracy: 0.8025\n",
      "Epoch 322/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1020 - accuracy: 0.9894 - val_loss: 0.9641 - val_accuracy: 0.8012\n",
      "Epoch 323/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0979 - accuracy: 0.9911 - val_loss: 0.9652 - val_accuracy: 0.7962\n",
      "Epoch 324/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1046 - accuracy: 0.9890 - val_loss: 0.9647 - val_accuracy: 0.7975\n",
      "Epoch 325/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1050 - accuracy: 0.9894 - val_loss: 0.9612 - val_accuracy: 0.7962\n",
      "Epoch 326/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1037 - accuracy: 0.9897 - val_loss: 0.9529 - val_accuracy: 0.7975\n",
      "Epoch 327/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 0.9906 - val_loss: 0.9490 - val_accuracy: 0.7975\n",
      "Epoch 328/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1007 - accuracy: 0.9904 - val_loss: 0.9492 - val_accuracy: 0.8000\n",
      "Epoch 329/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1023 - accuracy: 0.9900 - val_loss: 0.9490 - val_accuracy: 0.8012\n",
      "Epoch 330/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1084 - accuracy: 0.9878 - val_loss: 0.9532 - val_accuracy: 0.7975\n",
      "Epoch 331/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1043 - accuracy: 0.9893 - val_loss: 0.9531 - val_accuracy: 0.7962\n",
      "Epoch 332/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1026 - accuracy: 0.9893 - val_loss: 0.9510 - val_accuracy: 0.7975\n",
      "Epoch 333/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1017 - accuracy: 0.9903 - val_loss: 0.9525 - val_accuracy: 0.7950\n",
      "Epoch 334/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1035 - accuracy: 0.9894 - val_loss: 0.9503 - val_accuracy: 0.7962\n",
      "Epoch 335/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1034 - accuracy: 0.9891 - val_loss: 0.9533 - val_accuracy: 0.7975\n",
      "Epoch 336/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0997 - accuracy: 0.9904 - val_loss: 0.9565 - val_accuracy: 0.7950\n",
      "Epoch 337/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1025 - accuracy: 0.9897 - val_loss: 0.9589 - val_accuracy: 0.7925\n",
      "Epoch 338/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1038 - accuracy: 0.9896 - val_loss: 0.9670 - val_accuracy: 0.7925\n",
      "Epoch 339/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1056 - accuracy: 0.9895 - val_loss: 0.9654 - val_accuracy: 0.7925\n",
      "Epoch 340/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1040 - accuracy: 0.9886 - val_loss: 0.9604 - val_accuracy: 0.7925\n",
      "Epoch 341/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1000 - accuracy: 0.9907 - val_loss: 0.9619 - val_accuracy: 0.7887\n",
      "Epoch 342/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9908 - val_loss: 0.9630 - val_accuracy: 0.7900\n",
      "Epoch 343/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1009 - accuracy: 0.9909 - val_loss: 0.9681 - val_accuracy: 0.7937\n",
      "Epoch 344/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9915 - val_loss: 0.9691 - val_accuracy: 0.7887\n",
      "Epoch 345/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1025 - accuracy: 0.9903 - val_loss: 0.9787 - val_accuracy: 0.7875\n",
      "Epoch 346/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9915 - val_loss: 0.9741 - val_accuracy: 0.7837\n",
      "Epoch 347/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1048 - accuracy: 0.9893 - val_loss: 0.9697 - val_accuracy: 0.7837\n",
      "Epoch 348/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9914 - val_loss: 0.9716 - val_accuracy: 0.7862\n",
      "Epoch 349/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9913 - val_loss: 0.9712 - val_accuracy: 0.7875\n",
      "Epoch 350/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0984 - accuracy: 0.9924 - val_loss: 0.9767 - val_accuracy: 0.7875\n",
      "Epoch 351/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1011 - accuracy: 0.9912 - val_loss: 0.9758 - val_accuracy: 0.7925\n",
      "Epoch 352/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1022 - accuracy: 0.9903 - val_loss: 0.9734 - val_accuracy: 0.7912\n",
      "Epoch 353/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.1021 - accuracy: 0.9902 - val_loss: 0.9716 - val_accuracy: 0.7937\n",
      "Epoch 354/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1032 - accuracy: 0.9900 - val_loss: 0.9716 - val_accuracy: 0.7925\n",
      "Epoch 355/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.0995 - accuracy: 0.9916 - val_loss: 0.9693 - val_accuracy: 0.7912\n",
      "Epoch 356/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9908 - val_loss: 0.9682 - val_accuracy: 0.7887\n",
      "Epoch 357/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1009 - accuracy: 0.9913 - val_loss: 0.9700 - val_accuracy: 0.7912\n",
      "Epoch 358/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1026 - accuracy: 0.9904 - val_loss: 0.9693 - val_accuracy: 0.7862\n",
      "Epoch 359/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9914 - val_loss: 0.9718 - val_accuracy: 0.7887\n",
      "Epoch 360/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1030 - accuracy: 0.9900 - val_loss: 0.9741 - val_accuracy: 0.7887\n",
      "Epoch 361/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1031 - accuracy: 0.9899 - val_loss: 0.9758 - val_accuracy: 0.7950\n",
      "Epoch 362/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1007 - accuracy: 0.9905 - val_loss: 0.9787 - val_accuracy: 0.7900\n",
      "Epoch 363/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1028 - accuracy: 0.9895 - val_loss: 0.9822 - val_accuracy: 0.7937\n",
      "Epoch 364/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1033 - accuracy: 0.9904 - val_loss: 0.9851 - val_accuracy: 0.7912\n",
      "Epoch 365/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1043 - accuracy: 0.9895 - val_loss: 0.9797 - val_accuracy: 0.7950\n",
      "Epoch 366/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9913 - val_loss: 0.9782 - val_accuracy: 0.7950\n",
      "Epoch 367/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9901 - val_loss: 0.9752 - val_accuracy: 0.7950\n",
      "Epoch 368/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0963 - accuracy: 0.9917 - val_loss: 0.9791 - val_accuracy: 0.7937\n",
      "Epoch 369/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1000 - accuracy: 0.9912 - val_loss: 0.9777 - val_accuracy: 0.7975\n",
      "Epoch 370/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1020 - accuracy: 0.9895 - val_loss: 0.9807 - val_accuracy: 0.7962\n",
      "Epoch 371/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1021 - accuracy: 0.9907 - val_loss: 0.9753 - val_accuracy: 0.7937\n",
      "Epoch 372/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1000 - accuracy: 0.9906 - val_loss: 0.9749 - val_accuracy: 0.7925\n",
      "Epoch 373/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1013 - accuracy: 0.9904 - val_loss: 0.9781 - val_accuracy: 0.7887\n",
      "Epoch 374/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1011 - accuracy: 0.9903 - val_loss: 0.9849 - val_accuracy: 0.7850\n",
      "Epoch 375/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1005 - accuracy: 0.9911 - val_loss: 0.9870 - val_accuracy: 0.7862\n",
      "Epoch 376/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1021 - accuracy: 0.9897 - val_loss: 0.9849 - val_accuracy: 0.7862\n",
      "Epoch 377/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9910 - val_loss: 0.9810 - val_accuracy: 0.7862\n",
      "Epoch 378/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0989 - accuracy: 0.9910 - val_loss: 0.9862 - val_accuracy: 0.7887\n",
      "Epoch 379/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1014 - accuracy: 0.9907 - val_loss: 0.9895 - val_accuracy: 0.7912\n",
      "Epoch 380/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9924 - val_loss: 0.9886 - val_accuracy: 0.7850\n",
      "Epoch 381/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1014 - accuracy: 0.9905 - val_loss: 0.9857 - val_accuracy: 0.7862\n",
      "Epoch 382/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1013 - accuracy: 0.9907 - val_loss: 0.9916 - val_accuracy: 0.7825\n",
      "Epoch 383/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1041 - accuracy: 0.9901 - val_loss: 0.9912 - val_accuracy: 0.7887\n",
      "Epoch 384/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1036 - accuracy: 0.9903 - val_loss: 0.9977 - val_accuracy: 0.7875\n",
      "Epoch 385/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1039 - accuracy: 0.9897 - val_loss: 0.9946 - val_accuracy: 0.7875\n",
      "Epoch 386/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1024 - accuracy: 0.9903 - val_loss: 0.9943 - val_accuracy: 0.7887\n",
      "Epoch 387/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1037 - accuracy: 0.9904 - val_loss: 0.9965 - val_accuracy: 0.7862\n",
      "Epoch 388/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1007 - accuracy: 0.9912 - val_loss: 0.9949 - val_accuracy: 0.7837\n",
      "Epoch 389/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1009 - accuracy: 0.9906 - val_loss: 0.9981 - val_accuracy: 0.7862\n",
      "Epoch 390/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1012 - accuracy: 0.9903 - val_loss: 0.9932 - val_accuracy: 0.7862\n",
      "Epoch 391/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1010 - accuracy: 0.9897 - val_loss: 0.9925 - val_accuracy: 0.7875\n",
      "Epoch 392/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1019 - accuracy: 0.9903 - val_loss: 0.9898 - val_accuracy: 0.7925\n",
      "Epoch 393/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.0956 - accuracy: 0.9922 - val_loss: 0.9913 - val_accuracy: 0.7925\n",
      "Epoch 394/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0967 - accuracy: 0.9924 - val_loss: 0.9912 - val_accuracy: 0.7950\n",
      "Epoch 395/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1006 - accuracy: 0.9918 - val_loss: 0.9934 - val_accuracy: 0.7962\n",
      "Epoch 396/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1054 - accuracy: 0.9899 - val_loss: 0.9898 - val_accuracy: 0.7950\n",
      "Epoch 397/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1010 - accuracy: 0.9911 - val_loss: 0.9941 - val_accuracy: 0.7912\n",
      "Epoch 398/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1020 - accuracy: 0.9909 - val_loss: 0.9951 - val_accuracy: 0.7887\n",
      "Epoch 399/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0997 - accuracy: 0.9914 - val_loss: 0.9904 - val_accuracy: 0.7900\n",
      "Epoch 400/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1028 - accuracy: 0.9901 - val_loss: 0.9844 - val_accuracy: 0.7900\n",
      "Epoch 401/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0993 - accuracy: 0.9910 - val_loss: 0.9838 - val_accuracy: 0.7925\n",
      "Epoch 402/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1004 - accuracy: 0.9911 - val_loss: 0.9835 - val_accuracy: 0.7937\n",
      "Epoch 403/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0986 - accuracy: 0.9921 - val_loss: 0.9875 - val_accuracy: 0.7912\n",
      "Epoch 404/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1000 - accuracy: 0.9911 - val_loss: 0.9928 - val_accuracy: 0.7925\n",
      "Epoch 405/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9907 - val_loss: 0.9975 - val_accuracy: 0.7925\n",
      "Epoch 406/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1024 - accuracy: 0.9899 - val_loss: 0.9990 - val_accuracy: 0.7950\n",
      "Epoch 407/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9914 - val_loss: 0.9957 - val_accuracy: 0.7962\n",
      "Epoch 408/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1008 - accuracy: 0.9911 - val_loss: 0.9965 - val_accuracy: 0.7962\n",
      "Epoch 409/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9912 - val_loss: 0.9987 - val_accuracy: 0.8000\n",
      "Epoch 410/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1013 - accuracy: 0.9907 - val_loss: 0.9963 - val_accuracy: 0.7962\n",
      "Epoch 411/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1009 - accuracy: 0.9918 - val_loss: 0.9934 - val_accuracy: 0.7950\n",
      "Epoch 412/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1002 - accuracy: 0.9904 - val_loss: 0.9963 - val_accuracy: 0.7937\n",
      "Epoch 413/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9918 - val_loss: 0.9984 - val_accuracy: 0.7950\n",
      "Epoch 414/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1020 - accuracy: 0.9897 - val_loss: 0.9976 - val_accuracy: 0.7962\n",
      "Epoch 415/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1026 - accuracy: 0.9900 - val_loss: 1.0044 - val_accuracy: 0.7962\n",
      "Epoch 416/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9911 - val_loss: 1.0062 - val_accuracy: 0.7962\n",
      "Epoch 417/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1022 - accuracy: 0.9903 - val_loss: 1.0028 - val_accuracy: 0.7925\n",
      "Epoch 418/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0967 - accuracy: 0.9920 - val_loss: 1.0068 - val_accuracy: 0.7937\n",
      "Epoch 419/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9907 - val_loss: 1.0057 - val_accuracy: 0.7937\n",
      "Epoch 420/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1040 - accuracy: 0.9892 - val_loss: 1.0045 - val_accuracy: 0.7937\n",
      "Epoch 421/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0994 - accuracy: 0.9919 - val_loss: 1.0016 - val_accuracy: 0.7887\n",
      "Epoch 422/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9922 - val_loss: 0.9983 - val_accuracy: 0.7937\n",
      "Epoch 423/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0988 - accuracy: 0.9912 - val_loss: 0.9944 - val_accuracy: 0.7962\n",
      "Epoch 424/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1053 - accuracy: 0.9895 - val_loss: 0.9914 - val_accuracy: 0.7987\n",
      "Epoch 425/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9914 - val_loss: 0.9907 - val_accuracy: 0.8000\n",
      "Epoch 426/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9928 - val_loss: 0.9915 - val_accuracy: 0.8012\n",
      "Epoch 427/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1033 - accuracy: 0.9903 - val_loss: 0.9944 - val_accuracy: 0.8025\n",
      "Epoch 428/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1012 - accuracy: 0.9913 - val_loss: 0.9959 - val_accuracy: 0.8012\n",
      "Epoch 429/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9912 - val_loss: 1.0007 - val_accuracy: 0.8000\n",
      "Epoch 430/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1020 - accuracy: 0.9898 - val_loss: 0.9994 - val_accuracy: 0.8000\n",
      "Epoch 431/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9916 - val_loss: 0.9928 - val_accuracy: 0.8012\n",
      "Epoch 432/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1007 - accuracy: 0.9916 - val_loss: 0.9848 - val_accuracy: 0.8037\n",
      "Epoch 433/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9926 - val_loss: 0.9869 - val_accuracy: 0.8012\n",
      "Epoch 434/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9918 - val_loss: 0.9934 - val_accuracy: 0.8000\n",
      "Epoch 435/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9911 - val_loss: 0.9952 - val_accuracy: 0.8000\n",
      "Epoch 436/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9924 - val_loss: 0.9921 - val_accuracy: 0.8012\n",
      "Epoch 437/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1023 - accuracy: 0.9903 - val_loss: 0.9937 - val_accuracy: 0.8025\n",
      "Epoch 438/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0965 - accuracy: 0.9924 - val_loss: 0.9892 - val_accuracy: 0.8025\n",
      "Epoch 439/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9910 - val_loss: 0.9884 - val_accuracy: 0.7987\n",
      "Epoch 440/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1040 - accuracy: 0.9901 - val_loss: 0.9913 - val_accuracy: 0.7975\n",
      "Epoch 441/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1019 - accuracy: 0.9901 - val_loss: 0.9915 - val_accuracy: 0.8000\n",
      "Epoch 442/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0984 - accuracy: 0.9920 - val_loss: 0.9969 - val_accuracy: 0.7962\n",
      "Epoch 443/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9913 - val_loss: 0.9978 - val_accuracy: 0.7925\n",
      "Epoch 444/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1033 - accuracy: 0.9903 - val_loss: 0.9990 - val_accuracy: 0.7975\n",
      "Epoch 445/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1011 - accuracy: 0.9902 - val_loss: 0.9947 - val_accuracy: 0.7950\n",
      "Epoch 446/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0977 - accuracy: 0.9923 - val_loss: 0.9911 - val_accuracy: 0.7937\n",
      "Epoch 447/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.0993 - accuracy: 0.9912 - val_loss: 0.9879 - val_accuracy: 0.7950\n",
      "Epoch 448/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1006 - accuracy: 0.9918 - val_loss: 0.9902 - val_accuracy: 0.7962\n",
      "Epoch 449/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.0989 - accuracy: 0.9909 - val_loss: 0.9907 - val_accuracy: 0.7975\n",
      "Epoch 450/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0994 - accuracy: 0.9914 - val_loss: 0.9898 - val_accuracy: 0.7950\n",
      "Epoch 451/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0974 - accuracy: 0.9930 - val_loss: 0.9880 - val_accuracy: 0.7912\n",
      "Epoch 452/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1000 - accuracy: 0.9909 - val_loss: 0.9862 - val_accuracy: 0.7962\n",
      "Epoch 453/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0969 - accuracy: 0.9926 - val_loss: 0.9835 - val_accuracy: 0.7975\n",
      "Epoch 454/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9920 - val_loss: 0.9789 - val_accuracy: 0.7950\n",
      "Epoch 455/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0999 - accuracy: 0.9922 - val_loss: 0.9832 - val_accuracy: 0.8000\n",
      "Epoch 456/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1005 - accuracy: 0.9907 - val_loss: 0.9890 - val_accuracy: 0.7962\n",
      "Epoch 457/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9924 - val_loss: 0.9905 - val_accuracy: 0.7975\n",
      "Epoch 458/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9911 - val_loss: 0.9903 - val_accuracy: 0.7950\n",
      "Epoch 459/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9924 - val_loss: 0.9893 - val_accuracy: 0.7975\n",
      "Epoch 460/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1009 - accuracy: 0.9913 - val_loss: 0.9881 - val_accuracy: 0.7975\n",
      "Epoch 461/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9928 - val_loss: 0.9875 - val_accuracy: 0.8025\n",
      "Epoch 462/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9927 - val_loss: 0.9930 - val_accuracy: 0.8012\n",
      "Epoch 463/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9922 - val_loss: 0.9992 - val_accuracy: 0.8000\n",
      "Epoch 464/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0950 - accuracy: 0.9932 - val_loss: 1.0020 - val_accuracy: 0.7975\n",
      "Epoch 465/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1012 - accuracy: 0.9913 - val_loss: 1.0024 - val_accuracy: 0.8037\n",
      "Epoch 466/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9916 - val_loss: 0.9952 - val_accuracy: 0.8062\n",
      "Epoch 467/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1013 - accuracy: 0.9922 - val_loss: 0.9896 - val_accuracy: 0.8062\n",
      "Epoch 468/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9921 - val_loss: 0.9799 - val_accuracy: 0.8050\n",
      "Epoch 469/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9923 - val_loss: 0.9816 - val_accuracy: 0.8012\n",
      "Epoch 470/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9922 - val_loss: 0.9778 - val_accuracy: 0.8025\n",
      "Epoch 471/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9916 - val_loss: 0.9760 - val_accuracy: 0.8050\n",
      "Epoch 472/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0957 - accuracy: 0.9928 - val_loss: 0.9792 - val_accuracy: 0.8025\n",
      "Epoch 473/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0937 - accuracy: 0.9935 - val_loss: 0.9802 - val_accuracy: 0.8050\n",
      "Epoch 474/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1003 - accuracy: 0.9924 - val_loss: 0.9774 - val_accuracy: 0.8050\n",
      "Epoch 475/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9920 - val_loss: 0.9810 - val_accuracy: 0.8025\n",
      "Epoch 476/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9919 - val_loss: 0.9803 - val_accuracy: 0.8025\n",
      "Epoch 477/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 0.9913 - val_loss: 0.9766 - val_accuracy: 0.7987\n",
      "Epoch 478/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9918 - val_loss: 0.9721 - val_accuracy: 0.7987\n",
      "Epoch 479/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0949 - accuracy: 0.9937 - val_loss: 0.9760 - val_accuracy: 0.8012\n",
      "Epoch 480/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9924 - val_loss: 0.9764 - val_accuracy: 0.8037\n",
      "Epoch 481/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9917 - val_loss: 0.9740 - val_accuracy: 0.8075\n",
      "Epoch 482/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0996 - accuracy: 0.9920 - val_loss: 0.9745 - val_accuracy: 0.8062\n",
      "Epoch 483/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9915 - val_loss: 0.9765 - val_accuracy: 0.8037\n",
      "Epoch 484/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9923 - val_loss: 0.9723 - val_accuracy: 0.8075\n",
      "Epoch 485/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9916 - val_loss: 0.9663 - val_accuracy: 0.8075\n",
      "Epoch 486/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0963 - accuracy: 0.9926 - val_loss: 0.9650 - val_accuracy: 0.8087\n",
      "Epoch 487/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9930 - val_loss: 0.9750 - val_accuracy: 0.8037\n",
      "Epoch 488/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9927 - val_loss: 0.9830 - val_accuracy: 0.8062\n",
      "Epoch 489/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1003 - accuracy: 0.9913 - val_loss: 0.9859 - val_accuracy: 0.8075\n",
      "Epoch 490/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1000 - accuracy: 0.9914 - val_loss: 0.9846 - val_accuracy: 0.8062\n",
      "Epoch 491/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9916 - val_loss: 0.9800 - val_accuracy: 0.8062\n",
      "Epoch 492/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9931 - val_loss: 0.9829 - val_accuracy: 0.8000\n",
      "Epoch 493/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9918 - val_loss: 0.9825 - val_accuracy: 0.8025\n",
      "Epoch 494/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9917 - val_loss: 0.9774 - val_accuracy: 0.8075\n",
      "Epoch 495/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9914 - val_loss: 0.9773 - val_accuracy: 0.8075\n",
      "Epoch 496/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9934 - val_loss: 0.9879 - val_accuracy: 0.8062\n",
      "Epoch 497/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9918 - val_loss: 0.9870 - val_accuracy: 0.8000\n",
      "Epoch 498/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0968 - accuracy: 0.9919 - val_loss: 0.9913 - val_accuracy: 0.8012\n",
      "Epoch 499/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9918 - val_loss: 0.9976 - val_accuracy: 0.8000\n",
      "Epoch 500/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9920 - val_loss: 0.9966 - val_accuracy: 0.8075\n",
      "Epoch 501/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9913 - val_loss: 0.9990 - val_accuracy: 0.8012\n",
      "Epoch 502/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9926 - val_loss: 1.0032 - val_accuracy: 0.8000\n",
      "Epoch 503/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9920 - val_loss: 1.0057 - val_accuracy: 0.8000\n",
      "Epoch 504/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1006 - accuracy: 0.9920 - val_loss: 1.0045 - val_accuracy: 0.7975\n",
      "Epoch 505/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9926 - val_loss: 1.0025 - val_accuracy: 0.7987\n",
      "Epoch 506/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9918 - val_loss: 0.9944 - val_accuracy: 0.7962\n",
      "Epoch 507/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9921 - val_loss: 0.9878 - val_accuracy: 0.7987\n",
      "Epoch 508/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1011 - accuracy: 0.9911 - val_loss: 0.9934 - val_accuracy: 0.7950\n",
      "Epoch 509/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1016 - accuracy: 0.9911 - val_loss: 0.9989 - val_accuracy: 0.7937\n",
      "Epoch 510/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9917 - val_loss: 1.0033 - val_accuracy: 0.7925\n",
      "Epoch 511/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0954 - accuracy: 0.9930 - val_loss: 1.0018 - val_accuracy: 0.7900\n",
      "Epoch 512/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0981 - accuracy: 0.9926 - val_loss: 0.9978 - val_accuracy: 0.7950\n",
      "Epoch 513/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0975 - accuracy: 0.9924 - val_loss: 0.9965 - val_accuracy: 0.7962\n",
      "Epoch 514/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1005 - accuracy: 0.9912 - val_loss: 1.0007 - val_accuracy: 0.7962\n",
      "Epoch 515/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9911 - val_loss: 1.0072 - val_accuracy: 0.8000\n",
      "Epoch 516/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9920 - val_loss: 1.0069 - val_accuracy: 0.8000\n",
      "Epoch 517/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1003 - accuracy: 0.9905 - val_loss: 1.0094 - val_accuracy: 0.7962\n",
      "Epoch 518/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1005 - accuracy: 0.9913 - val_loss: 1.0070 - val_accuracy: 0.7987\n",
      "Epoch 519/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0992 - accuracy: 0.9919 - val_loss: 1.0028 - val_accuracy: 0.7975\n",
      "Epoch 520/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0967 - accuracy: 0.9934 - val_loss: 1.0020 - val_accuracy: 0.7900\n",
      "Epoch 521/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1001 - accuracy: 0.9922 - val_loss: 1.0021 - val_accuracy: 0.7925\n",
      "Epoch 522/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9915 - val_loss: 0.9990 - val_accuracy: 0.8012\n",
      "Epoch 523/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9924 - val_loss: 0.9957 - val_accuracy: 0.8012\n",
      "Epoch 524/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9912 - val_loss: 0.9924 - val_accuracy: 0.8025\n",
      "Epoch 525/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1053 - accuracy: 0.9907 - val_loss: 0.9962 - val_accuracy: 0.7962\n",
      "Epoch 526/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9924 - val_loss: 0.9948 - val_accuracy: 0.8025\n",
      "Epoch 527/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9924 - val_loss: 0.9922 - val_accuracy: 0.8012\n",
      "Epoch 528/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0983 - accuracy: 0.9927 - val_loss: 0.9955 - val_accuracy: 0.8075\n",
      "Epoch 529/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1012 - accuracy: 0.9915 - val_loss: 0.9921 - val_accuracy: 0.8062\n",
      "Epoch 530/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1011 - accuracy: 0.9916 - val_loss: 0.9911 - val_accuracy: 0.8012\n",
      "Epoch 531/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9920 - val_loss: 0.9921 - val_accuracy: 0.8050\n",
      "Epoch 532/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0977 - accuracy: 0.9933 - val_loss: 0.9996 - val_accuracy: 0.7987\n",
      "Epoch 533/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9909 - val_loss: 0.9967 - val_accuracy: 0.7912\n",
      "Epoch 534/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0986 - accuracy: 0.9923 - val_loss: 0.9904 - val_accuracy: 0.7962\n",
      "Epoch 535/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0982 - accuracy: 0.9928 - val_loss: 0.9859 - val_accuracy: 0.7962\n",
      "Epoch 536/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9913 - val_loss: 0.9851 - val_accuracy: 0.7962\n",
      "Epoch 537/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9925 - val_loss: 0.9915 - val_accuracy: 0.7987\n",
      "Epoch 538/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.0990 - accuracy: 0.9926 - val_loss: 0.9929 - val_accuracy: 0.7975\n",
      "Epoch 539/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9924 - val_loss: 0.9917 - val_accuracy: 0.7975\n",
      "Epoch 540/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1014 - accuracy: 0.9924 - val_loss: 0.9900 - val_accuracy: 0.8025\n",
      "Epoch 541/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1051 - accuracy: 0.9908 - val_loss: 0.9853 - val_accuracy: 0.7962\n",
      "Epoch 542/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9929 - val_loss: 0.9844 - val_accuracy: 0.8025\n",
      "Epoch 543/1000\n",
      "19/19 [==============================] - 0s 13ms/step - loss: 0.1008 - accuracy: 0.9918 - val_loss: 0.9872 - val_accuracy: 0.7975\n",
      "Epoch 544/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0975 - accuracy: 0.9919 - val_loss: 0.9899 - val_accuracy: 0.8037\n",
      "Epoch 545/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0978 - accuracy: 0.9922 - val_loss: 0.9889 - val_accuracy: 0.8025\n",
      "Epoch 546/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1022 - accuracy: 0.9914 - val_loss: 0.9873 - val_accuracy: 0.8012\n",
      "Epoch 547/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9926 - val_loss: 0.9860 - val_accuracy: 0.8012\n",
      "Epoch 548/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1003 - accuracy: 0.9922 - val_loss: 0.9823 - val_accuracy: 0.8037\n",
      "Epoch 549/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0986 - accuracy: 0.9921 - val_loss: 0.9810 - val_accuracy: 0.8012\n",
      "Epoch 550/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9916 - val_loss: 0.9833 - val_accuracy: 0.7987\n",
      "Epoch 551/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1004 - accuracy: 0.9920 - val_loss: 0.9810 - val_accuracy: 0.7962\n",
      "Epoch 552/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0993 - accuracy: 0.9922 - val_loss: 0.9774 - val_accuracy: 0.7962\n",
      "Epoch 553/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1001 - accuracy: 0.9918 - val_loss: 0.9785 - val_accuracy: 0.8025\n",
      "Epoch 554/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0972 - accuracy: 0.9926 - val_loss: 0.9820 - val_accuracy: 0.7975\n",
      "Epoch 555/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9917 - val_loss: 0.9799 - val_accuracy: 0.8037\n",
      "Epoch 556/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9917 - val_loss: 0.9788 - val_accuracy: 0.8050\n",
      "Epoch 557/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9917 - val_loss: 0.9763 - val_accuracy: 0.8025\n",
      "Epoch 558/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1028 - accuracy: 0.9913 - val_loss: 0.9707 - val_accuracy: 0.8037\n",
      "Epoch 559/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0971 - accuracy: 0.9929 - val_loss: 0.9728 - val_accuracy: 0.8037\n",
      "Epoch 560/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1004 - accuracy: 0.9912 - val_loss: 0.9790 - val_accuracy: 0.8012\n",
      "Epoch 561/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1017 - accuracy: 0.9907 - val_loss: 0.9772 - val_accuracy: 0.7962\n",
      "Epoch 562/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0996 - accuracy: 0.9919 - val_loss: 0.9786 - val_accuracy: 0.8050\n",
      "Epoch 563/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9923 - val_loss: 0.9799 - val_accuracy: 0.8012\n",
      "Epoch 564/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1023 - accuracy: 0.9906 - val_loss: 0.9817 - val_accuracy: 0.8012\n",
      "Epoch 565/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9926 - val_loss: 0.9848 - val_accuracy: 0.8050\n",
      "Epoch 566/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1005 - accuracy: 0.9921 - val_loss: 0.9917 - val_accuracy: 0.8025\n",
      "Epoch 567/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1018 - accuracy: 0.9909 - val_loss: 0.9967 - val_accuracy: 0.8000\n",
      "Epoch 568/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0970 - accuracy: 0.9932 - val_loss: 0.9964 - val_accuracy: 0.7987\n",
      "Epoch 569/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9922 - val_loss: 0.9935 - val_accuracy: 0.7925\n",
      "Epoch 570/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1006 - accuracy: 0.9922 - val_loss: 0.9933 - val_accuracy: 0.7975\n",
      "Epoch 571/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0988 - accuracy: 0.9922 - val_loss: 0.9896 - val_accuracy: 0.7975\n",
      "Epoch 572/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9922 - val_loss: 0.9870 - val_accuracy: 0.8012\n",
      "Epoch 573/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0965 - accuracy: 0.9925 - val_loss: 0.9878 - val_accuracy: 0.8000\n",
      "Epoch 574/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0965 - accuracy: 0.9936 - val_loss: 0.9957 - val_accuracy: 0.7975\n",
      "Epoch 575/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9928 - val_loss: 0.9984 - val_accuracy: 0.7962\n",
      "Epoch 576/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9926 - val_loss: 0.9935 - val_accuracy: 0.7987\n",
      "Epoch 577/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9925 - val_loss: 0.9896 - val_accuracy: 0.8012\n",
      "Epoch 578/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1003 - accuracy: 0.9920 - val_loss: 0.9907 - val_accuracy: 0.8025\n",
      "Epoch 579/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9930 - val_loss: 0.9963 - val_accuracy: 0.8062\n",
      "Epoch 580/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9928 - val_loss: 1.0001 - val_accuracy: 0.8012\n",
      "Epoch 581/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9928 - val_loss: 1.0043 - val_accuracy: 0.7950\n",
      "Epoch 582/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1016 - accuracy: 0.9917 - val_loss: 1.0056 - val_accuracy: 0.7962\n",
      "Epoch 583/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9922 - val_loss: 1.0070 - val_accuracy: 0.7975\n",
      "Epoch 584/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9932 - val_loss: 1.0075 - val_accuracy: 0.7950\n",
      "Epoch 585/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0951 - accuracy: 0.9934 - val_loss: 1.0107 - val_accuracy: 0.7937\n",
      "Epoch 586/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0994 - accuracy: 0.9928 - val_loss: 1.0079 - val_accuracy: 0.7950\n",
      "Epoch 587/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9926 - val_loss: 1.0097 - val_accuracy: 0.7962\n",
      "Epoch 588/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1004 - accuracy: 0.9919 - val_loss: 1.0108 - val_accuracy: 0.7937\n",
      "Epoch 589/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9920 - val_loss: 1.0161 - val_accuracy: 0.7950\n",
      "Epoch 590/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9926 - val_loss: 1.0208 - val_accuracy: 0.7912\n",
      "Epoch 591/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1002 - accuracy: 0.9916 - val_loss: 1.0137 - val_accuracy: 0.7875\n",
      "Epoch 592/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1003 - accuracy: 0.9925 - val_loss: 1.0077 - val_accuracy: 0.7912\n",
      "Epoch 593/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9918 - val_loss: 1.0108 - val_accuracy: 0.7950\n",
      "Epoch 594/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0979 - accuracy: 0.9929 - val_loss: 1.0086 - val_accuracy: 0.7950\n",
      "Epoch 595/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9926 - val_loss: 1.0079 - val_accuracy: 0.7950\n",
      "Epoch 596/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 0.9913 - val_loss: 1.0055 - val_accuracy: 0.7962\n",
      "Epoch 597/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9915 - val_loss: 1.0001 - val_accuracy: 0.7912\n",
      "Epoch 598/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1006 - accuracy: 0.9924 - val_loss: 0.9949 - val_accuracy: 0.7962\n",
      "Epoch 599/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1028 - accuracy: 0.9914 - val_loss: 0.9904 - val_accuracy: 0.7987\n",
      "Epoch 600/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1038 - accuracy: 0.9918 - val_loss: 0.9943 - val_accuracy: 0.8000\n",
      "Epoch 601/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1016 - accuracy: 0.9916 - val_loss: 0.9995 - val_accuracy: 0.7950\n",
      "Epoch 602/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1011 - accuracy: 0.9913 - val_loss: 0.9963 - val_accuracy: 0.7962\n",
      "Epoch 603/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9924 - val_loss: 0.9989 - val_accuracy: 0.7937\n",
      "Epoch 604/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9922 - val_loss: 0.9982 - val_accuracy: 0.7925\n",
      "Epoch 605/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9914 - val_loss: 0.9948 - val_accuracy: 0.7925\n",
      "Epoch 606/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0999 - accuracy: 0.9918 - val_loss: 0.9915 - val_accuracy: 0.7950\n",
      "Epoch 607/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9918 - val_loss: 0.9909 - val_accuracy: 0.7950\n",
      "Epoch 608/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1027 - accuracy: 0.9904 - val_loss: 0.9894 - val_accuracy: 0.7937\n",
      "Epoch 609/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9929 - val_loss: 0.9885 - val_accuracy: 0.7950\n",
      "Epoch 610/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1007 - accuracy: 0.9915 - val_loss: 0.9853 - val_accuracy: 0.7950\n",
      "Epoch 611/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0974 - accuracy: 0.9924 - val_loss: 0.9870 - val_accuracy: 0.7925\n",
      "Epoch 612/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0966 - accuracy: 0.9932 - val_loss: 0.9870 - val_accuracy: 0.7925\n",
      "Epoch 613/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9932 - val_loss: 0.9831 - val_accuracy: 0.7937\n",
      "Epoch 614/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9929 - val_loss: 0.9862 - val_accuracy: 0.7937\n",
      "Epoch 615/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1009 - accuracy: 0.9916 - val_loss: 0.9897 - val_accuracy: 0.7900\n",
      "Epoch 616/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1000 - accuracy: 0.9927 - val_loss: 0.9873 - val_accuracy: 0.7900\n",
      "Epoch 617/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9928 - val_loss: 0.9903 - val_accuracy: 0.7900\n",
      "Epoch 618/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9928 - val_loss: 0.9926 - val_accuracy: 0.7887\n",
      "Epoch 619/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9924 - val_loss: 0.9931 - val_accuracy: 0.7912\n",
      "Epoch 620/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9926 - val_loss: 0.9922 - val_accuracy: 0.7937\n",
      "Epoch 621/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0976 - accuracy: 0.9924 - val_loss: 0.9984 - val_accuracy: 0.7900\n",
      "Epoch 622/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9922 - val_loss: 1.0044 - val_accuracy: 0.7937\n",
      "Epoch 623/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9933 - val_loss: 1.0066 - val_accuracy: 0.7912\n",
      "Epoch 624/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0991 - accuracy: 0.9929 - val_loss: 1.0056 - val_accuracy: 0.7925\n",
      "Epoch 625/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9928 - val_loss: 1.0046 - val_accuracy: 0.7937\n",
      "Epoch 626/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9930 - val_loss: 1.0032 - val_accuracy: 0.7925\n",
      "Epoch 627/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9939 - val_loss: 1.0085 - val_accuracy: 0.7925\n",
      "Epoch 628/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0949 - accuracy: 0.9942 - val_loss: 1.0104 - val_accuracy: 0.7925\n",
      "Epoch 629/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9924 - val_loss: 1.0127 - val_accuracy: 0.7900\n",
      "Epoch 630/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9926 - val_loss: 1.0129 - val_accuracy: 0.7887\n",
      "Epoch 631/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9926 - val_loss: 1.0129 - val_accuracy: 0.7912\n",
      "Epoch 632/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1027 - accuracy: 0.9918 - val_loss: 1.0178 - val_accuracy: 0.7862\n",
      "Epoch 633/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9941 - val_loss: 1.0236 - val_accuracy: 0.7850\n",
      "Epoch 634/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9936 - val_loss: 1.0316 - val_accuracy: 0.7875\n",
      "Epoch 635/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0991 - accuracy: 0.9924 - val_loss: 1.0309 - val_accuracy: 0.7875\n",
      "Epoch 636/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0986 - accuracy: 0.9925 - val_loss: 1.0280 - val_accuracy: 0.7912\n",
      "Epoch 637/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9928 - val_loss: 1.0318 - val_accuracy: 0.7862\n",
      "Epoch 638/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1028 - accuracy: 0.9915 - val_loss: 1.0327 - val_accuracy: 0.7887\n",
      "Epoch 639/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9920 - val_loss: 1.0281 - val_accuracy: 0.7900\n",
      "Epoch 640/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9928 - val_loss: 1.0265 - val_accuracy: 0.7962\n",
      "Epoch 641/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9932 - val_loss: 1.0295 - val_accuracy: 0.7962\n",
      "Epoch 642/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1012 - accuracy: 0.9921 - val_loss: 1.0245 - val_accuracy: 0.7962\n",
      "Epoch 643/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9929 - val_loss: 1.0248 - val_accuracy: 0.7962\n",
      "Epoch 644/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1024 - accuracy: 0.9924 - val_loss: 1.0228 - val_accuracy: 0.7975\n",
      "Epoch 645/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9925 - val_loss: 1.0189 - val_accuracy: 0.7987\n",
      "Epoch 646/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1005 - accuracy: 0.9923 - val_loss: 1.0257 - val_accuracy: 0.7950\n",
      "Epoch 647/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9928 - val_loss: 1.0349 - val_accuracy: 0.7950\n",
      "Epoch 648/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0994 - accuracy: 0.9924 - val_loss: 1.0415 - val_accuracy: 0.7925\n",
      "Epoch 649/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9932 - val_loss: 1.0478 - val_accuracy: 0.7950\n",
      "Epoch 650/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0952 - accuracy: 0.9932 - val_loss: 1.0418 - val_accuracy: 0.7937\n",
      "Epoch 651/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0977 - accuracy: 0.9928 - val_loss: 1.0382 - val_accuracy: 0.7987\n",
      "Epoch 652/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9936 - val_loss: 1.0380 - val_accuracy: 0.7987\n",
      "Epoch 653/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0962 - accuracy: 0.9938 - val_loss: 1.0402 - val_accuracy: 0.8012\n",
      "Epoch 654/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1008 - accuracy: 0.9918 - val_loss: 1.0394 - val_accuracy: 0.8025\n",
      "Epoch 655/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0967 - accuracy: 0.9933 - val_loss: 1.0327 - val_accuracy: 0.8012\n",
      "Epoch 656/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9924 - val_loss: 1.0298 - val_accuracy: 0.7975\n",
      "Epoch 657/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1028 - accuracy: 0.9916 - val_loss: 1.0314 - val_accuracy: 0.7962\n",
      "Epoch 658/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0962 - accuracy: 0.9933 - val_loss: 1.0402 - val_accuracy: 0.7962\n",
      "Epoch 659/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1008 - accuracy: 0.9917 - val_loss: 1.0459 - val_accuracy: 0.7912\n",
      "Epoch 660/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9926 - val_loss: 1.0445 - val_accuracy: 0.7900\n",
      "Epoch 661/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1011 - accuracy: 0.9917 - val_loss: 1.0522 - val_accuracy: 0.7937\n",
      "Epoch 662/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1033 - accuracy: 0.9922 - val_loss: 1.0531 - val_accuracy: 0.7925\n",
      "Epoch 663/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1004 - accuracy: 0.9919 - val_loss: 1.0539 - val_accuracy: 0.7875\n",
      "Epoch 664/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0994 - accuracy: 0.9920 - val_loss: 1.0528 - val_accuracy: 0.7850\n",
      "Epoch 665/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9930 - val_loss: 1.0480 - val_accuracy: 0.7900\n",
      "Epoch 666/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0991 - accuracy: 0.9928 - val_loss: 1.0478 - val_accuracy: 0.7975\n",
      "Epoch 667/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9926 - val_loss: 1.0492 - val_accuracy: 0.7962\n",
      "Epoch 668/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9922 - val_loss: 1.0474 - val_accuracy: 0.7937\n",
      "Epoch 669/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9930 - val_loss: 1.0475 - val_accuracy: 0.7937\n",
      "Epoch 670/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9926 - val_loss: 1.0465 - val_accuracy: 0.7937\n",
      "Epoch 671/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1018 - accuracy: 0.9918 - val_loss: 1.0447 - val_accuracy: 0.7950\n",
      "Epoch 672/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9928 - val_loss: 1.0414 - val_accuracy: 0.7950\n",
      "Epoch 673/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1025 - accuracy: 0.9911 - val_loss: 1.0303 - val_accuracy: 0.7962\n",
      "Epoch 674/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9924 - val_loss: 1.0281 - val_accuracy: 0.7937\n",
      "Epoch 675/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9932 - val_loss: 1.0220 - val_accuracy: 0.7925\n",
      "Epoch 676/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1001 - accuracy: 0.9922 - val_loss: 1.0189 - val_accuracy: 0.7887\n",
      "Epoch 677/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1009 - accuracy: 0.9923 - val_loss: 1.0160 - val_accuracy: 0.7912\n",
      "Epoch 678/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0969 - accuracy: 0.9932 - val_loss: 1.0118 - val_accuracy: 0.7950\n",
      "Epoch 679/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0996 - accuracy: 0.9923 - val_loss: 1.0121 - val_accuracy: 0.7925\n",
      "Epoch 680/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1026 - accuracy: 0.9922 - val_loss: 1.0087 - val_accuracy: 0.7950\n",
      "Epoch 681/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9927 - val_loss: 1.0044 - val_accuracy: 0.7962\n",
      "Epoch 682/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9930 - val_loss: 1.0064 - val_accuracy: 0.7975\n",
      "Epoch 683/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1025 - accuracy: 0.9916 - val_loss: 1.0020 - val_accuracy: 0.7975\n",
      "Epoch 684/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0977 - accuracy: 0.9938 - val_loss: 1.0039 - val_accuracy: 0.7975\n",
      "Epoch 685/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0978 - accuracy: 0.9932 - val_loss: 1.0079 - val_accuracy: 0.7962\n",
      "Epoch 686/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1019 - accuracy: 0.9926 - val_loss: 1.0130 - val_accuracy: 0.7912\n",
      "Epoch 687/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9934 - val_loss: 1.0142 - val_accuracy: 0.7950\n",
      "Epoch 688/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9941 - val_loss: 1.0081 - val_accuracy: 0.7937\n",
      "Epoch 689/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1007 - accuracy: 0.9928 - val_loss: 1.0062 - val_accuracy: 0.7962\n",
      "Epoch 690/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0969 - accuracy: 0.9933 - val_loss: 1.0007 - val_accuracy: 0.7975\n",
      "Epoch 691/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9926 - val_loss: 0.9977 - val_accuracy: 0.7975\n",
      "Epoch 692/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0995 - accuracy: 0.9930 - val_loss: 0.9990 - val_accuracy: 0.7975\n",
      "Epoch 693/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0966 - accuracy: 0.9931 - val_loss: 1.0001 - val_accuracy: 0.7925\n",
      "Epoch 694/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1016 - accuracy: 0.9916 - val_loss: 1.0076 - val_accuracy: 0.7962\n",
      "Epoch 695/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9936 - val_loss: 1.0080 - val_accuracy: 0.7962\n",
      "Epoch 696/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9930 - val_loss: 1.0056 - val_accuracy: 0.7925\n",
      "Epoch 697/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9926 - val_loss: 0.9998 - val_accuracy: 0.7887\n",
      "Epoch 698/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9930 - val_loss: 0.9975 - val_accuracy: 0.7887\n",
      "Epoch 699/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9926 - val_loss: 0.9972 - val_accuracy: 0.7912\n",
      "Epoch 700/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0996 - accuracy: 0.9923 - val_loss: 0.9971 - val_accuracy: 0.7925\n",
      "Epoch 701/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9925 - val_loss: 1.0026 - val_accuracy: 0.7937\n",
      "Epoch 702/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0950 - accuracy: 0.9947 - val_loss: 1.0064 - val_accuracy: 0.7900\n",
      "Epoch 703/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9931 - val_loss: 1.0051 - val_accuracy: 0.7862\n",
      "Epoch 704/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9930 - val_loss: 1.0101 - val_accuracy: 0.7912\n",
      "Epoch 705/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0948 - accuracy: 0.9941 - val_loss: 1.0165 - val_accuracy: 0.7900\n",
      "Epoch 706/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0994 - accuracy: 0.9922 - val_loss: 1.0172 - val_accuracy: 0.7900\n",
      "Epoch 707/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1015 - accuracy: 0.9914 - val_loss: 1.0204 - val_accuracy: 0.7925\n",
      "Epoch 708/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9930 - val_loss: 1.0200 - val_accuracy: 0.7950\n",
      "Epoch 709/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9943 - val_loss: 1.0216 - val_accuracy: 0.7937\n",
      "Epoch 710/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9926 - val_loss: 1.0180 - val_accuracy: 0.7912\n",
      "Epoch 711/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0955 - accuracy: 0.9938 - val_loss: 1.0154 - val_accuracy: 0.7925\n",
      "Epoch 712/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0976 - accuracy: 0.9923 - val_loss: 1.0140 - val_accuracy: 0.7937\n",
      "Epoch 713/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9926 - val_loss: 1.0079 - val_accuracy: 0.7937\n",
      "Epoch 714/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0988 - accuracy: 0.9933 - val_loss: 1.0106 - val_accuracy: 0.7962\n",
      "Epoch 715/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9939 - val_loss: 1.0106 - val_accuracy: 0.7950\n",
      "Epoch 716/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9936 - val_loss: 1.0177 - val_accuracy: 0.7912\n",
      "Epoch 717/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9930 - val_loss: 1.0196 - val_accuracy: 0.7937\n",
      "Epoch 718/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0977 - accuracy: 0.9933 - val_loss: 1.0172 - val_accuracy: 0.7962\n",
      "Epoch 719/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9930 - val_loss: 1.0201 - val_accuracy: 0.7950\n",
      "Epoch 720/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0951 - accuracy: 0.9948 - val_loss: 1.0203 - val_accuracy: 0.7912\n",
      "Epoch 721/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9932 - val_loss: 1.0229 - val_accuracy: 0.7825\n",
      "Epoch 722/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0976 - accuracy: 0.9930 - val_loss: 1.0250 - val_accuracy: 0.7875\n",
      "Epoch 723/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9944 - val_loss: 1.0214 - val_accuracy: 0.7887\n",
      "Epoch 724/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9932 - val_loss: 1.0190 - val_accuracy: 0.7937\n",
      "Epoch 725/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9924 - val_loss: 1.0178 - val_accuracy: 0.7950\n",
      "Epoch 726/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1008 - accuracy: 0.9925 - val_loss: 1.0222 - val_accuracy: 0.7950\n",
      "Epoch 727/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9933 - val_loss: 1.0250 - val_accuracy: 0.7962\n",
      "Epoch 728/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9936 - val_loss: 1.0227 - val_accuracy: 0.7975\n",
      "Epoch 729/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9936 - val_loss: 1.0263 - val_accuracy: 0.7987\n",
      "Epoch 730/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0971 - accuracy: 0.9936 - val_loss: 1.0242 - val_accuracy: 0.7950\n",
      "Epoch 731/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0963 - accuracy: 0.9949 - val_loss: 1.0272 - val_accuracy: 0.7987\n",
      "Epoch 732/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9927 - val_loss: 1.0286 - val_accuracy: 0.7937\n",
      "Epoch 733/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0959 - accuracy: 0.9934 - val_loss: 1.0218 - val_accuracy: 0.7937\n",
      "Epoch 734/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9928 - val_loss: 1.0216 - val_accuracy: 0.7912\n",
      "Epoch 735/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9930 - val_loss: 1.0263 - val_accuracy: 0.7912\n",
      "Epoch 736/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9941 - val_loss: 1.0317 - val_accuracy: 0.7950\n",
      "Epoch 737/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9936 - val_loss: 1.0315 - val_accuracy: 0.7962\n",
      "Epoch 738/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9931 - val_loss: 1.0283 - val_accuracy: 0.8000\n",
      "Epoch 739/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0969 - accuracy: 0.9929 - val_loss: 1.0271 - val_accuracy: 0.7950\n",
      "Epoch 740/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9936 - val_loss: 1.0349 - val_accuracy: 0.7937\n",
      "Epoch 741/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0981 - accuracy: 0.9923 - val_loss: 1.0444 - val_accuracy: 0.7912\n",
      "Epoch 742/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9929 - val_loss: 1.0414 - val_accuracy: 0.7912\n",
      "Epoch 743/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0993 - accuracy: 0.9941 - val_loss: 1.0468 - val_accuracy: 0.7900\n",
      "Epoch 744/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9927 - val_loss: 1.0464 - val_accuracy: 0.7900\n",
      "Epoch 745/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1008 - accuracy: 0.9920 - val_loss: 1.0454 - val_accuracy: 0.7912\n",
      "Epoch 746/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9928 - val_loss: 1.0414 - val_accuracy: 0.7900\n",
      "Epoch 747/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9924 - val_loss: 1.0352 - val_accuracy: 0.7862\n",
      "Epoch 748/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0963 - accuracy: 0.9940 - val_loss: 1.0355 - val_accuracy: 0.7925\n",
      "Epoch 749/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9932 - val_loss: 1.0328 - val_accuracy: 0.7912\n",
      "Epoch 750/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9928 - val_loss: 1.0302 - val_accuracy: 0.7925\n",
      "Epoch 751/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9933 - val_loss: 1.0326 - val_accuracy: 0.7887\n",
      "Epoch 752/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9920 - val_loss: 1.0382 - val_accuracy: 0.7887\n",
      "Epoch 753/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9924 - val_loss: 1.0367 - val_accuracy: 0.7887\n",
      "Epoch 754/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1000 - accuracy: 0.9922 - val_loss: 1.0353 - val_accuracy: 0.7912\n",
      "Epoch 755/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0954 - accuracy: 0.9938 - val_loss: 1.0384 - val_accuracy: 0.7925\n",
      "Epoch 756/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9932 - val_loss: 1.0387 - val_accuracy: 0.7975\n",
      "Epoch 757/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0995 - accuracy: 0.9930 - val_loss: 1.0355 - val_accuracy: 0.7925\n",
      "Epoch 758/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9930 - val_loss: 1.0370 - val_accuracy: 0.7937\n",
      "Epoch 759/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9925 - val_loss: 1.0460 - val_accuracy: 0.7875\n",
      "Epoch 760/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9928 - val_loss: 1.0453 - val_accuracy: 0.7850\n",
      "Epoch 761/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9928 - val_loss: 1.0498 - val_accuracy: 0.7837\n",
      "Epoch 762/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0971 - accuracy: 0.9929 - val_loss: 1.0521 - val_accuracy: 0.7875\n",
      "Epoch 763/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0954 - accuracy: 0.9936 - val_loss: 1.0494 - val_accuracy: 0.7837\n",
      "Epoch 764/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9930 - val_loss: 1.0478 - val_accuracy: 0.7875\n",
      "Epoch 765/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0971 - accuracy: 0.9939 - val_loss: 1.0486 - val_accuracy: 0.7875\n",
      "Epoch 766/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9926 - val_loss: 1.0458 - val_accuracy: 0.7825\n",
      "Epoch 767/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9934 - val_loss: 1.0466 - val_accuracy: 0.7875\n",
      "Epoch 768/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0944 - accuracy: 0.9942 - val_loss: 1.0478 - val_accuracy: 0.7862\n",
      "Epoch 769/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9930 - val_loss: 1.0553 - val_accuracy: 0.7825\n",
      "Epoch 770/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9932 - val_loss: 1.0565 - val_accuracy: 0.7825\n",
      "Epoch 771/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9926 - val_loss: 1.0571 - val_accuracy: 0.7825\n",
      "Epoch 772/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0989 - accuracy: 0.9932 - val_loss: 1.0546 - val_accuracy: 0.7800\n",
      "Epoch 773/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9932 - val_loss: 1.0594 - val_accuracy: 0.7825\n",
      "Epoch 774/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9939 - val_loss: 1.0638 - val_accuracy: 0.7812\n",
      "Epoch 775/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9932 - val_loss: 1.0651 - val_accuracy: 0.7775\n",
      "Epoch 776/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9932 - val_loss: 1.0713 - val_accuracy: 0.7812\n",
      "Epoch 777/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9936 - val_loss: 1.0669 - val_accuracy: 0.7788\n",
      "Epoch 778/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9924 - val_loss: 1.0642 - val_accuracy: 0.7800\n",
      "Epoch 779/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1000 - accuracy: 0.9924 - val_loss: 1.0628 - val_accuracy: 0.7800\n",
      "Epoch 780/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9929 - val_loss: 1.0657 - val_accuracy: 0.7788\n",
      "Epoch 781/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9935 - val_loss: 1.0672 - val_accuracy: 0.7812\n",
      "Epoch 782/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9926 - val_loss: 1.0630 - val_accuracy: 0.7812\n",
      "Epoch 783/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0978 - accuracy: 0.9941 - val_loss: 1.0680 - val_accuracy: 0.7788\n",
      "Epoch 784/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9924 - val_loss: 1.0698 - val_accuracy: 0.7825\n",
      "Epoch 785/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9934 - val_loss: 1.0649 - val_accuracy: 0.7862\n",
      "Epoch 786/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0969 - accuracy: 0.9936 - val_loss: 1.0624 - val_accuracy: 0.7862\n",
      "Epoch 787/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1010 - accuracy: 0.9924 - val_loss: 1.0587 - val_accuracy: 0.7825\n",
      "Epoch 788/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9936 - val_loss: 1.0493 - val_accuracy: 0.7875\n",
      "Epoch 789/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9935 - val_loss: 1.0461 - val_accuracy: 0.7862\n",
      "Epoch 790/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9920 - val_loss: 1.0451 - val_accuracy: 0.7837\n",
      "Epoch 791/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0947 - accuracy: 0.9948 - val_loss: 1.0419 - val_accuracy: 0.7850\n",
      "Epoch 792/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1025 - accuracy: 0.9929 - val_loss: 1.0370 - val_accuracy: 0.7862\n",
      "Epoch 793/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9927 - val_loss: 1.0379 - val_accuracy: 0.7825\n",
      "Epoch 794/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9926 - val_loss: 1.0378 - val_accuracy: 0.7875\n",
      "Epoch 795/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9928 - val_loss: 1.0359 - val_accuracy: 0.7862\n",
      "Epoch 796/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9939 - val_loss: 1.0414 - val_accuracy: 0.7887\n",
      "Epoch 797/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9928 - val_loss: 1.0429 - val_accuracy: 0.7862\n",
      "Epoch 798/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0982 - accuracy: 0.9931 - val_loss: 1.0420 - val_accuracy: 0.7825\n",
      "Epoch 799/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0954 - accuracy: 0.9941 - val_loss: 1.0438 - val_accuracy: 0.7825\n",
      "Epoch 800/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1015 - accuracy: 0.9923 - val_loss: 1.0490 - val_accuracy: 0.7812\n",
      "Epoch 801/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9929 - val_loss: 1.0438 - val_accuracy: 0.7850\n",
      "Epoch 802/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0993 - accuracy: 0.9924 - val_loss: 1.0478 - val_accuracy: 0.7812\n",
      "Epoch 803/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9937 - val_loss: 1.0497 - val_accuracy: 0.7875\n",
      "Epoch 804/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0964 - accuracy: 0.9932 - val_loss: 1.0505 - val_accuracy: 0.7875\n",
      "Epoch 805/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0943 - accuracy: 0.9950 - val_loss: 1.0512 - val_accuracy: 0.7862\n",
      "Epoch 806/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0962 - accuracy: 0.9936 - val_loss: 1.0493 - val_accuracy: 0.7862\n",
      "Epoch 807/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0950 - accuracy: 0.9943 - val_loss: 1.0459 - val_accuracy: 0.7837\n",
      "Epoch 808/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9937 - val_loss: 1.0440 - val_accuracy: 0.7875\n",
      "Epoch 809/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0947 - accuracy: 0.9941 - val_loss: 1.0449 - val_accuracy: 0.7887\n",
      "Epoch 810/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0990 - accuracy: 0.9930 - val_loss: 1.0489 - val_accuracy: 0.7912\n",
      "Epoch 811/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9935 - val_loss: 1.0517 - val_accuracy: 0.7887\n",
      "Epoch 812/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9932 - val_loss: 1.0521 - val_accuracy: 0.7875\n",
      "Epoch 813/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9939 - val_loss: 1.0520 - val_accuracy: 0.7837\n",
      "Epoch 814/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9921 - val_loss: 1.0539 - val_accuracy: 0.7850\n",
      "Epoch 815/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0991 - accuracy: 0.9930 - val_loss: 1.0567 - val_accuracy: 0.7825\n",
      "Epoch 816/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9938 - val_loss: 1.0588 - val_accuracy: 0.7850\n",
      "Epoch 817/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9927 - val_loss: 1.0580 - val_accuracy: 0.7837\n",
      "Epoch 818/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9932 - val_loss: 1.0593 - val_accuracy: 0.7812\n",
      "Epoch 819/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9930 - val_loss: 1.0591 - val_accuracy: 0.7812\n",
      "Epoch 820/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1014 - accuracy: 0.9920 - val_loss: 1.0566 - val_accuracy: 0.7875\n",
      "Epoch 821/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0996 - accuracy: 0.9935 - val_loss: 1.0590 - val_accuracy: 0.7862\n",
      "Epoch 822/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9945 - val_loss: 1.0572 - val_accuracy: 0.7887\n",
      "Epoch 823/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9932 - val_loss: 1.0598 - val_accuracy: 0.7925\n",
      "Epoch 824/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0971 - accuracy: 0.9931 - val_loss: 1.0602 - val_accuracy: 0.7925\n",
      "Epoch 825/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9930 - val_loss: 1.0630 - val_accuracy: 0.7937\n",
      "Epoch 826/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9927 - val_loss: 1.0665 - val_accuracy: 0.7900\n",
      "Epoch 827/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9934 - val_loss: 1.0638 - val_accuracy: 0.7900\n",
      "Epoch 828/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0952 - accuracy: 0.9932 - val_loss: 1.0661 - val_accuracy: 0.7900\n",
      "Epoch 829/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9915 - val_loss: 1.0651 - val_accuracy: 0.7912\n",
      "Epoch 830/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0979 - accuracy: 0.9934 - val_loss: 1.0650 - val_accuracy: 0.7937\n",
      "Epoch 831/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9933 - val_loss: 1.0653 - val_accuracy: 0.7950\n",
      "Epoch 832/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0969 - accuracy: 0.9939 - val_loss: 1.0648 - val_accuracy: 0.7937\n",
      "Epoch 833/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0981 - accuracy: 0.9936 - val_loss: 1.0636 - val_accuracy: 0.7925\n",
      "Epoch 834/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0977 - accuracy: 0.9939 - val_loss: 1.0571 - val_accuracy: 0.7950\n",
      "Epoch 835/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0966 - accuracy: 0.9933 - val_loss: 1.0545 - val_accuracy: 0.7962\n",
      "Epoch 836/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0965 - accuracy: 0.9939 - val_loss: 1.0556 - val_accuracy: 0.7925\n",
      "Epoch 837/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0962 - accuracy: 0.9940 - val_loss: 1.0548 - val_accuracy: 0.7925\n",
      "Epoch 838/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0963 - accuracy: 0.9936 - val_loss: 1.0577 - val_accuracy: 0.7937\n",
      "Epoch 839/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9934 - val_loss: 1.0576 - val_accuracy: 0.7862\n",
      "Epoch 840/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0943 - accuracy: 0.9949 - val_loss: 1.0534 - val_accuracy: 0.7912\n",
      "Epoch 841/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9927 - val_loss: 1.0540 - val_accuracy: 0.7937\n",
      "Epoch 842/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9936 - val_loss: 1.0578 - val_accuracy: 0.7937\n",
      "Epoch 843/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9936 - val_loss: 1.0598 - val_accuracy: 0.7950\n",
      "Epoch 844/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0996 - accuracy: 0.9929 - val_loss: 1.0615 - val_accuracy: 0.7987\n",
      "Epoch 845/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0966 - accuracy: 0.9938 - val_loss: 1.0642 - val_accuracy: 0.8012\n",
      "Epoch 846/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0997 - accuracy: 0.9932 - val_loss: 1.0637 - val_accuracy: 0.7987\n",
      "Epoch 847/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0958 - accuracy: 0.9942 - val_loss: 1.0628 - val_accuracy: 0.7937\n",
      "Epoch 848/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0966 - accuracy: 0.9935 - val_loss: 1.0639 - val_accuracy: 0.7950\n",
      "Epoch 849/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9926 - val_loss: 1.0616 - val_accuracy: 0.7950\n",
      "Epoch 850/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9929 - val_loss: 1.0625 - val_accuracy: 0.7950\n",
      "Epoch 851/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1003 - accuracy: 0.9922 - val_loss: 1.0580 - val_accuracy: 0.7887\n",
      "Epoch 852/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0962 - accuracy: 0.9940 - val_loss: 1.0598 - val_accuracy: 0.7900\n",
      "Epoch 853/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0963 - accuracy: 0.9934 - val_loss: 1.0607 - val_accuracy: 0.7912\n",
      "Epoch 854/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9932 - val_loss: 1.0538 - val_accuracy: 0.7887\n",
      "Epoch 855/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0965 - accuracy: 0.9932 - val_loss: 1.0480 - val_accuracy: 0.7925\n",
      "Epoch 856/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0956 - accuracy: 0.9936 - val_loss: 1.0504 - val_accuracy: 0.7950\n",
      "Epoch 857/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0963 - accuracy: 0.9941 - val_loss: 1.0492 - val_accuracy: 0.7937\n",
      "Epoch 858/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9938 - val_loss: 1.0504 - val_accuracy: 0.7925\n",
      "Epoch 859/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1006 - accuracy: 0.9929 - val_loss: 1.0559 - val_accuracy: 0.7900\n",
      "Epoch 860/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9944 - val_loss: 1.0591 - val_accuracy: 0.7912\n",
      "Epoch 861/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9934 - val_loss: 1.0655 - val_accuracy: 0.7875\n",
      "Epoch 862/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9934 - val_loss: 1.0680 - val_accuracy: 0.7887\n",
      "Epoch 863/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9937 - val_loss: 1.0722 - val_accuracy: 0.7862\n",
      "Epoch 864/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9927 - val_loss: 1.0746 - val_accuracy: 0.7850\n",
      "Epoch 865/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0999 - accuracy: 0.9928 - val_loss: 1.0710 - val_accuracy: 0.7850\n",
      "Epoch 866/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9933 - val_loss: 1.0717 - val_accuracy: 0.7825\n",
      "Epoch 867/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0944 - accuracy: 0.9941 - val_loss: 1.0729 - val_accuracy: 0.7812\n",
      "Epoch 868/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0962 - accuracy: 0.9932 - val_loss: 1.0696 - val_accuracy: 0.7788\n",
      "Epoch 869/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0968 - accuracy: 0.9934 - val_loss: 1.0671 - val_accuracy: 0.7825\n",
      "Epoch 870/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0966 - accuracy: 0.9930 - val_loss: 1.0609 - val_accuracy: 0.7850\n",
      "Epoch 871/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0959 - accuracy: 0.9934 - val_loss: 1.0567 - val_accuracy: 0.7925\n",
      "Epoch 872/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0992 - accuracy: 0.9927 - val_loss: 1.0516 - val_accuracy: 0.7900\n",
      "Epoch 873/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0956 - accuracy: 0.9936 - val_loss: 1.0542 - val_accuracy: 0.7887\n",
      "Epoch 874/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1029 - accuracy: 0.9924 - val_loss: 1.0583 - val_accuracy: 0.7950\n",
      "Epoch 875/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9928 - val_loss: 1.0593 - val_accuracy: 0.7925\n",
      "Epoch 876/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0979 - accuracy: 0.9932 - val_loss: 1.0576 - val_accuracy: 0.7900\n",
      "Epoch 877/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0974 - accuracy: 0.9939 - val_loss: 1.0580 - val_accuracy: 0.7887\n",
      "Epoch 878/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0970 - accuracy: 0.9939 - val_loss: 1.0589 - val_accuracy: 0.7875\n",
      "Epoch 879/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0948 - accuracy: 0.9943 - val_loss: 1.0589 - val_accuracy: 0.7900\n",
      "Epoch 880/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9937 - val_loss: 1.0600 - val_accuracy: 0.7900\n",
      "Epoch 881/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0959 - accuracy: 0.9942 - val_loss: 1.0616 - val_accuracy: 0.7900\n",
      "Epoch 882/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0970 - accuracy: 0.9934 - val_loss: 1.0604 - val_accuracy: 0.7925\n",
      "Epoch 883/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9930 - val_loss: 1.0631 - val_accuracy: 0.7900\n",
      "Epoch 884/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0961 - accuracy: 0.9934 - val_loss: 1.0586 - val_accuracy: 0.7925\n",
      "Epoch 885/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0972 - accuracy: 0.9941 - val_loss: 1.0558 - val_accuracy: 0.7900\n",
      "Epoch 886/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0941 - accuracy: 0.9939 - val_loss: 1.0580 - val_accuracy: 0.7887\n",
      "Epoch 887/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0987 - accuracy: 0.9936 - val_loss: 1.0612 - val_accuracy: 0.7937\n",
      "Epoch 888/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1007 - accuracy: 0.9921 - val_loss: 1.0599 - val_accuracy: 0.7925\n",
      "Epoch 889/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9934 - val_loss: 1.0585 - val_accuracy: 0.7962\n",
      "Epoch 890/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9938 - val_loss: 1.0599 - val_accuracy: 0.7925\n",
      "Epoch 891/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9943 - val_loss: 1.0659 - val_accuracy: 0.7900\n",
      "Epoch 892/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0998 - accuracy: 0.9928 - val_loss: 1.0783 - val_accuracy: 0.7887\n",
      "Epoch 893/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0970 - accuracy: 0.9935 - val_loss: 1.0816 - val_accuracy: 0.7925\n",
      "Epoch 894/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9934 - val_loss: 1.0845 - val_accuracy: 0.7887\n",
      "Epoch 895/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9943 - val_loss: 1.0850 - val_accuracy: 0.7912\n",
      "Epoch 896/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0977 - accuracy: 0.9925 - val_loss: 1.0790 - val_accuracy: 0.7912\n",
      "Epoch 897/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9932 - val_loss: 1.0761 - val_accuracy: 0.8000\n",
      "Epoch 898/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0975 - accuracy: 0.9932 - val_loss: 1.0762 - val_accuracy: 0.7950\n",
      "Epoch 899/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0977 - accuracy: 0.9934 - val_loss: 1.0774 - val_accuracy: 0.7950\n",
      "Epoch 900/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9929 - val_loss: 1.0810 - val_accuracy: 0.7937\n",
      "Epoch 901/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0998 - accuracy: 0.9928 - val_loss: 1.0809 - val_accuracy: 0.7925\n",
      "Epoch 902/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9925 - val_loss: 1.0803 - val_accuracy: 0.7912\n",
      "Epoch 903/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0981 - accuracy: 0.9938 - val_loss: 1.0821 - val_accuracy: 0.7950\n",
      "Epoch 904/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9924 - val_loss: 1.0859 - val_accuracy: 0.7937\n",
      "Epoch 905/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0988 - accuracy: 0.9935 - val_loss: 1.0875 - val_accuracy: 0.7937\n",
      "Epoch 906/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0974 - accuracy: 0.9941 - val_loss: 1.0848 - val_accuracy: 0.7925\n",
      "Epoch 907/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1001 - accuracy: 0.9930 - val_loss: 1.0813 - val_accuracy: 0.7925\n",
      "Epoch 908/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9946 - val_loss: 1.0792 - val_accuracy: 0.7937\n",
      "Epoch 909/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9941 - val_loss: 1.0745 - val_accuracy: 0.7912\n",
      "Epoch 910/1000\n",
      "19/19 [==============================] - 0s 16ms/step - loss: 0.0936 - accuracy: 0.9947 - val_loss: 1.0719 - val_accuracy: 0.7887\n",
      "Epoch 911/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0983 - accuracy: 0.9925 - val_loss: 1.0744 - val_accuracy: 0.7887\n",
      "Epoch 912/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9925 - val_loss: 1.0720 - val_accuracy: 0.7912\n",
      "Epoch 913/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0988 - accuracy: 0.9932 - val_loss: 1.0670 - val_accuracy: 0.7937\n",
      "Epoch 914/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0966 - accuracy: 0.9940 - val_loss: 1.0690 - val_accuracy: 0.7912\n",
      "Epoch 915/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0985 - accuracy: 0.9930 - val_loss: 1.0695 - val_accuracy: 0.7937\n",
      "Epoch 916/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0994 - accuracy: 0.9926 - val_loss: 1.0691 - val_accuracy: 0.7925\n",
      "Epoch 917/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9945 - val_loss: 1.0708 - val_accuracy: 0.7950\n",
      "Epoch 918/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9936 - val_loss: 1.0708 - val_accuracy: 0.7987\n",
      "Epoch 919/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0945 - accuracy: 0.9941 - val_loss: 1.0731 - val_accuracy: 0.8037\n",
      "Epoch 920/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0960 - accuracy: 0.9941 - val_loss: 1.0815 - val_accuracy: 0.8025\n",
      "Epoch 921/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0983 - accuracy: 0.9930 - val_loss: 1.0766 - val_accuracy: 0.8037\n",
      "Epoch 922/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0939 - accuracy: 0.9943 - val_loss: 1.0772 - val_accuracy: 0.8037\n",
      "Epoch 923/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9933 - val_loss: 1.0802 - val_accuracy: 0.8012\n",
      "Epoch 924/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0958 - accuracy: 0.9941 - val_loss: 1.0816 - val_accuracy: 0.7975\n",
      "Epoch 925/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0940 - accuracy: 0.9941 - val_loss: 1.0857 - val_accuracy: 0.7937\n",
      "Epoch 926/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0978 - accuracy: 0.9935 - val_loss: 1.0872 - val_accuracy: 0.7950\n",
      "Epoch 927/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9936 - val_loss: 1.0870 - val_accuracy: 0.7925\n",
      "Epoch 928/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9924 - val_loss: 1.0880 - val_accuracy: 0.7925\n",
      "Epoch 929/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0975 - accuracy: 0.9937 - val_loss: 1.0910 - val_accuracy: 0.7900\n",
      "Epoch 930/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1011 - accuracy: 0.9922 - val_loss: 1.0951 - val_accuracy: 0.7950\n",
      "Epoch 931/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9936 - val_loss: 1.0938 - val_accuracy: 0.7912\n",
      "Epoch 932/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0994 - accuracy: 0.9929 - val_loss: 1.0928 - val_accuracy: 0.7925\n",
      "Epoch 933/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0963 - accuracy: 0.9939 - val_loss: 1.0938 - val_accuracy: 0.7925\n",
      "Epoch 934/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0972 - accuracy: 0.9936 - val_loss: 1.0971 - val_accuracy: 0.7912\n",
      "Epoch 935/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0974 - accuracy: 0.9933 - val_loss: 1.0942 - val_accuracy: 0.7850\n",
      "Epoch 936/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9939 - val_loss: 1.0925 - val_accuracy: 0.7850\n",
      "Epoch 937/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0961 - accuracy: 0.9942 - val_loss: 1.0895 - val_accuracy: 0.7850\n",
      "Epoch 938/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1002 - accuracy: 0.9926 - val_loss: 1.0881 - val_accuracy: 0.7887\n",
      "Epoch 939/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0989 - accuracy: 0.9932 - val_loss: 1.0830 - val_accuracy: 0.7912\n",
      "Epoch 940/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0958 - accuracy: 0.9936 - val_loss: 1.0848 - val_accuracy: 0.7862\n",
      "Epoch 941/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9938 - val_loss: 1.0920 - val_accuracy: 0.7887\n",
      "Epoch 942/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1023 - accuracy: 0.9932 - val_loss: 1.0919 - val_accuracy: 0.7900\n",
      "Epoch 943/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0986 - accuracy: 0.9935 - val_loss: 1.0963 - val_accuracy: 0.7887\n",
      "Epoch 944/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9936 - val_loss: 1.0996 - val_accuracy: 0.7912\n",
      "Epoch 945/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0965 - accuracy: 0.9934 - val_loss: 1.1041 - val_accuracy: 0.7912\n",
      "Epoch 946/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0957 - accuracy: 0.9945 - val_loss: 1.1049 - val_accuracy: 0.7937\n",
      "Epoch 947/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0985 - accuracy: 0.9931 - val_loss: 1.1069 - val_accuracy: 0.7937\n",
      "Epoch 948/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0961 - accuracy: 0.9941 - val_loss: 1.0933 - val_accuracy: 0.7950\n",
      "Epoch 949/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0965 - accuracy: 0.9930 - val_loss: 1.0884 - val_accuracy: 0.7950\n",
      "Epoch 950/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9946 - val_loss: 1.0910 - val_accuracy: 0.7925\n",
      "Epoch 951/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1004 - accuracy: 0.9928 - val_loss: 1.0848 - val_accuracy: 0.7962\n",
      "Epoch 952/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0973 - accuracy: 0.9941 - val_loss: 1.0839 - val_accuracy: 0.7925\n",
      "Epoch 953/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0987 - accuracy: 0.9938 - val_loss: 1.0811 - val_accuracy: 0.7937\n",
      "Epoch 954/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0993 - accuracy: 0.9930 - val_loss: 1.0767 - val_accuracy: 0.7962\n",
      "Epoch 955/1000\n",
      "19/19 [==============================] - 0s 12ms/step - loss: 0.0948 - accuracy: 0.9947 - val_loss: 1.0785 - val_accuracy: 0.7950\n",
      "Epoch 956/1000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0954 - accuracy: 0.9941 - val_loss: 1.0782 - val_accuracy: 0.7937\n",
      "Epoch 957/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0989 - accuracy: 0.9938 - val_loss: 1.0847 - val_accuracy: 0.7937\n",
      "Epoch 958/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1004 - accuracy: 0.9932 - val_loss: 1.0881 - val_accuracy: 0.7912\n",
      "Epoch 959/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9926 - val_loss: 1.0972 - val_accuracy: 0.7900\n",
      "Epoch 960/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0968 - accuracy: 0.9939 - val_loss: 1.1073 - val_accuracy: 0.7950\n",
      "Epoch 961/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0975 - accuracy: 0.9936 - val_loss: 1.1014 - val_accuracy: 0.7962\n",
      "Epoch 962/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0946 - accuracy: 0.9945 - val_loss: 1.0927 - val_accuracy: 0.7987\n",
      "Epoch 963/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0990 - accuracy: 0.9932 - val_loss: 1.0893 - val_accuracy: 0.7950\n",
      "Epoch 964/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0980 - accuracy: 0.9940 - val_loss: 1.0909 - val_accuracy: 0.7950\n",
      "Epoch 965/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0976 - accuracy: 0.9936 - val_loss: 1.0888 - val_accuracy: 0.7925\n",
      "Epoch 966/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0983 - accuracy: 0.9940 - val_loss: 1.0898 - val_accuracy: 0.7962\n",
      "Epoch 967/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0967 - accuracy: 0.9938 - val_loss: 1.0894 - val_accuracy: 0.8000\n",
      "Epoch 968/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0984 - accuracy: 0.9933 - val_loss: 1.0914 - val_accuracy: 0.8012\n",
      "Epoch 969/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0991 - accuracy: 0.9936 - val_loss: 1.0918 - val_accuracy: 0.7962\n",
      "Epoch 970/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.1015 - accuracy: 0.9932 - val_loss: 1.0953 - val_accuracy: 0.7950\n",
      "Epoch 971/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9938 - val_loss: 1.0994 - val_accuracy: 0.7987\n",
      "Epoch 972/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0984 - accuracy: 0.9935 - val_loss: 1.0992 - val_accuracy: 0.7987\n",
      "Epoch 973/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0973 - accuracy: 0.9936 - val_loss: 1.0926 - val_accuracy: 0.7962\n",
      "Epoch 974/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0976 - accuracy: 0.9935 - val_loss: 1.0877 - val_accuracy: 0.7912\n",
      "Epoch 975/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0966 - accuracy: 0.9938 - val_loss: 1.0835 - val_accuracy: 0.7900\n",
      "Epoch 976/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0962 - accuracy: 0.9934 - val_loss: 1.0814 - val_accuracy: 0.7950\n",
      "Epoch 977/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0958 - accuracy: 0.9941 - val_loss: 1.0838 - val_accuracy: 0.7962\n",
      "Epoch 978/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0976 - accuracy: 0.9930 - val_loss: 1.0866 - val_accuracy: 0.8037\n",
      "Epoch 979/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1005 - accuracy: 0.9930 - val_loss: 1.0937 - val_accuracy: 0.7987\n",
      "Epoch 980/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0964 - accuracy: 0.9936 - val_loss: 1.1065 - val_accuracy: 0.7975\n",
      "Epoch 981/1000\n",
      "19/19 [==============================] - 0s 15ms/step - loss: 0.0929 - accuracy: 0.9949 - val_loss: 1.1125 - val_accuracy: 0.7937\n",
      "Epoch 982/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0991 - accuracy: 0.9931 - val_loss: 1.1081 - val_accuracy: 0.7925\n",
      "Epoch 983/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0982 - accuracy: 0.9936 - val_loss: 1.1119 - val_accuracy: 0.7937\n",
      "Epoch 984/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1004 - accuracy: 0.9934 - val_loss: 1.1142 - val_accuracy: 0.7950\n",
      "Epoch 985/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0964 - accuracy: 0.9942 - val_loss: 1.1173 - val_accuracy: 0.7950\n",
      "Epoch 986/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0960 - accuracy: 0.9951 - val_loss: 1.1158 - val_accuracy: 0.7912\n",
      "Epoch 987/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0958 - accuracy: 0.9942 - val_loss: 1.1152 - val_accuracy: 0.7912\n",
      "Epoch 988/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0989 - accuracy: 0.9928 - val_loss: 1.1152 - val_accuracy: 0.7912\n",
      "Epoch 989/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.1005 - accuracy: 0.9934 - val_loss: 1.1144 - val_accuracy: 0.7850\n",
      "Epoch 990/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0946 - accuracy: 0.9943 - val_loss: 1.1172 - val_accuracy: 0.7850\n",
      "Epoch 991/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0957 - accuracy: 0.9947 - val_loss: 1.1136 - val_accuracy: 0.7875\n",
      "Epoch 992/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0935 - accuracy: 0.9951 - val_loss: 1.1123 - val_accuracy: 0.7887\n",
      "Epoch 993/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0960 - accuracy: 0.9939 - val_loss: 1.1070 - val_accuracy: 0.7862\n",
      "Epoch 994/1000\n",
      "19/19 [==============================] - 0s 10ms/step - loss: 0.0969 - accuracy: 0.9932 - val_loss: 1.1100 - val_accuracy: 0.7887\n",
      "Epoch 995/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0971 - accuracy: 0.9939 - val_loss: 1.1113 - val_accuracy: 0.7900\n",
      "Epoch 996/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0983 - accuracy: 0.9945 - val_loss: 1.1045 - val_accuracy: 0.7900\n",
      "Epoch 997/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0981 - accuracy: 0.9928 - val_loss: 1.1022 - val_accuracy: 0.7912\n",
      "Epoch 998/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0980 - accuracy: 0.9934 - val_loss: 1.1028 - val_accuracy: 0.7937\n",
      "Epoch 999/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0972 - accuracy: 0.9931 - val_loss: 1.1031 - val_accuracy: 0.7875\n",
      "Epoch 1000/1000\n",
      "19/19 [==============================] - 0s 11ms/step - loss: 0.0993 - accuracy: 0.9936 - val_loss: 1.1053 - val_accuracy: 0.7925\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x371587610>"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history = model1.fit(fine_tune_data_train_red, fine_tune_labels_train, epochs=EPOCHS, batch_size=800, validation_data=(fine_tune_data_test_red, fine_tune_labels_test), validation_batch_size=800, callbacks=[model_checkpoint_callback])\n",
    "model1.save_weights(\"./Models/Inception-ResNet-4P-PCA/resnet-4p-model-weights\")\n",
    "model1.load_weights('./Models/Inception-ResNet-4P-PCA/resnet-4p-model-weights')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "a8380dbb",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x36a2dcdc0>"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1.load_weights('./Models/Inception-ResNet-4P-PCA/auto-Inception-ResNet-FT-model-weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "9116a6a6",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "def plot_train_loss(history, model_name):  \n",
    "    plt.figure(figsize=(8,6))\n",
    "    plt.plot(history.history['loss'])\n",
    "    plt.plot(history.history['accuracy'])\n",
    "    plt.plot(history.history['val_loss'])\n",
    "    plt.plot(history.history['val_accuracy'])\n",
    "    plt.title(model_name+' loss')\n",
    "    plt.ylabel('loss')\n",
    "    plt.xlabel('epochs')\n",
    "    plt.legend(['train loss', 'train accuracy', 'val loss', 'val accuracy'], loc='upper left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "c55b5870",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_train_loss(history, 'Fine-Tuned Inception-ResNet-4Piece Model')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03390fa6",
   "metadata": {},
   "source": [
    "### Model Evaluation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48701107",
   "metadata": {},
   "source": [
    "#### On Train Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0594069",
   "metadata": {},
   "source": [
    "Now, after the model is being trained, we will look at the performance of the model on the training set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "cf7ded38",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "475/475 [==============================] - 1s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "train_preds = model1.predict(fine_tune_data_train_red)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "f2a28269",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "train_pred_labels = np.argmax(train_preds, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "768775b2",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 1, 13, 13, ...,  3,  9,  9])"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_pred_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "eff52a3d",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "15197"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(train_pred_labels == train_labels['label'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3fd403fa",
   "metadata": {},
   "source": [
    "Looking at the misclassified images to get a better idea about what features the model is unable to capture."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "8f594a33",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>images</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>9198</th>\n",
       "      <td>9682</td>\n",
       "      <td>./Data/train/train/9682.tif</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12255</th>\n",
       "      <td>12896</td>\n",
       "      <td>./Data/train/train/12896.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14569</th>\n",
       "      <td>15333</td>\n",
       "      <td>./Data/train/train/15333.tif</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          id                        images  label\n",
       "9198    9682   ./Data/train/train/9682.tif     12\n",
       "12255  12896  ./Data/train/train/12896.tif      9\n",
       "14569  15333  ./Data/train/train/15333.tif      7"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mis_train_data = train_labels[train_pred_labels != train_labels['label']]\n",
    "mis_train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "2ea8f949",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>fscore</th>\n",
       "      <th>support</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.997959</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.998979</td>\n",
       "      <td>978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.998926</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.999463</td>\n",
       "      <td>930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.998905</td>\n",
       "      <td>0.999452</td>\n",
       "      <td>913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.998951</td>\n",
       "      <td>0.999475</td>\n",
       "      <td>953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.998976</td>\n",
       "      <td>0.999488</td>\n",
       "      <td>977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>945</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    precision    recall    fscore  support\n",
       "0    0.997959  1.000000  0.998979      978\n",
       "1    1.000000  1.000000  1.000000      925\n",
       "2    1.000000  1.000000  1.000000      925\n",
       "3    1.000000  1.000000  1.000000      947\n",
       "4    0.998926  1.000000  0.999463      930\n",
       "5    1.000000  1.000000  1.000000      962\n",
       "6    1.000000  1.000000  1.000000      998\n",
       "7    1.000000  0.998905  0.999452      913\n",
       "8    1.000000  1.000000  1.000000     1005\n",
       "9    1.000000  0.998951  0.999475      953\n",
       "10   1.000000  1.000000  1.000000      916\n",
       "11   1.000000  1.000000  1.000000      955\n",
       "12   1.000000  0.998976  0.999488      977\n",
       "13   1.000000  1.000000  1.000000      943\n",
       "14   1.000000  1.000000  1.000000      928\n",
       "15   1.000000  1.000000  1.000000      945"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, fscore, support = score(train_labels['label'], train_pred_labels)\n",
    "pd.DataFrame({\n",
    "    'precision': precision,\n",
    "    'recall': recall,\n",
    "    'fscore': fscore,\n",
    "    'support': support\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "659ae313",
   "metadata": {},
   "source": [
    "#### On Train-Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "0d1e8f85",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "25/25 [==============================] - 0s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "test_preds = model1.predict(fine_tune_data_test_red)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "49233509",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "test_pred_labels = np.argmax(test_preds, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "8f5cba39",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([11,  8,  0,  7, 10,  6, 13, 12, 14,  1,  6,  0,  7, 14, 13, 11,  0,\n",
       "        9,  6, 12,  8, 14,  3, 12,  6,  0,  3, 14,  3,  3, 10,  1,  9,  2,\n",
       "        6,  8,  4, 13, 11,  9,  6, 12, 14,  2,  5, 14,  5,  0, 13,  4,  0,\n",
       "       10, 12,  0, 11, 14, 15,  9,  9,  2, 15, 13, 11, 11,  6, 10,  8,  0,\n",
       "        6,  3,  4,  4, 12,  5,  8,  6,  7,  0, 12, 11, 11, 15, 12, 12,  9,\n",
       "        8, 13, 13, 13, 13, 10, 11,  4, 10,  1, 10, 11,  8, 13, 10,  3, 11,\n",
       "       12,  9,  7, 14,  8,  2,  5, 15,  7, 13,  9,  4, 12, 12, 12, 13, 11,\n",
       "        6, 13,  1,  0, 12,  1,  5, 10, 12, 15, 11,  3, 11,  6,  1,  6,  9,\n",
       "       14, 10,  7, 14, 10,  8, 11, 12,  8,  5,  8,  9,  2, 13,  0,  5,  4,\n",
       "       14,  4,  9,  9, 14,  7,  7,  8,  4,  9,  5, 13,  8,  2,  8,  8,  8,\n",
       "       13,  6,  3, 11,  7, 12, 10, 14, 14,  8,  6, 10, 13, 11,  6, 13,  4,\n",
       "       14, 11,  7, 15,  0,  3,  0, 10, 13, 13, 10, 11, 11, 10,  1,  9,  1,\n",
       "        9,  5, 13, 14, 12,  1,  4,  8,  7,  6, 13,  5,  4, 12,  8,  0,  9,\n",
       "       13, 15,  1,  7,  7,  2,  2,  9,  1,  7,  0,  3, 12,  3,  5,  2, 12,\n",
       "        4,  0, 12, 15, 12,  8, 13, 10, 15,  0, 11,  0, 11, 13, 15, 15,  8,\n",
       "        1, 10, 13,  5, 11,  9, 14,  6, 13, 12, 15, 11, 14, 11, 11, 14,  3,\n",
       "        0, 11,  4,  3,  1,  4,  6,  2,  7,  8, 10, 10,  8, 12,  7,  8, 13,\n",
       "        6,  8, 15,  0,  2,  3, 12, 10,  0,  8, 13,  0,  7, 15, 10, 10,  5,\n",
       "        3, 13, 11, 10,  9,  2, 10,  3, 15,  6,  0,  4, 14, 11, 13,  2, 15,\n",
       "        9, 15, 15, 11,  0, 15,  3, 11,  0, 15, 15, 11, 10, 13, 13, 15, 15,\n",
       "       13,  0,  3,  3,  2,  8,  2, 15, 11,  3,  3,  2,  5, 14,  4, 12,  3,\n",
       "        8, 12,  1,  3, 11,  5, 10,  3,  7, 15, 15,  3,  2, 12, 15,  3,  6,\n",
       "        0, 14, 14,  0,  5,  8, 14,  6, 12, 13,  6,  9,  8,  7, 10,  1,  2,\n",
       "       11, 14,  2, 14, 15, 15,  0,  1,  2, 10,  1,  0, 14,  6,  9,  0,  2,\n",
       "        1,  3, 14,  8, 13, 11,  6, 14, 13,  1,  4,  7, 13, 14,  5, 11,  8,\n",
       "        0,  0, 11,  0, 10,  1,  7,  2,  5,  6, 15,  9, 11,  0,  3,  0,  8,\n",
       "       12,  0,  3,  5,  5,  9,  6, 14,  7,  8,  3,  2, 13, 13,  1, 15, 11,\n",
       "        9,  8, 10, 14, 12, 13, 15,  1,  9,  9, 15, 12,  7, 14,  9,  0, 11,\n",
       "        5, 10,  4,  8,  9,  3,  7,  8,  6, 10,  1, 10,  7,  5,  0, 13, 10,\n",
       "        2,  3,  7, 10, 13,  7,  8,  5,  9,  1,  6,  7, 15,  0,  8,  7, 15,\n",
       "       12,  1,  0, 12,  5,  1,  1,  7,  9,  1, 14,  3,  5, 14,  6, 14,  7,\n",
       "        1, 14,  6, 14,  9,  7, 13, 14,  1,  5,  2, 13,  7,  2,  0,  8,  8,\n",
       "       15,  8,  6,  5, 14,  3,  6, 11,  8, 14,  6,  9,  5,  9, 11,  2, 13,\n",
       "        3,  1, 11, 10, 15,  2, 10,  0, 14,  9,  8,  1,  9, 13, 10, 10, 12,\n",
       "        9, 11,  6, 10, 12,  8, 10, 14, 14,  9,  0, 15,  8,  0,  8,  6, 13,\n",
       "       12,  5,  0, 15, 14,  5,  3,  2, 12,  1, 15, 11, 10,  3, 13,  1,  1,\n",
       "        1, 13, 14,  4,  9,  1, 11, 15, 14, 13,  4,  8,  5, 12, 13,  4, 15,\n",
       "       13,  0,  1, 14, 13,  7,  6, 12,  8,  7,  5, 15,  5,  9,  9,  7, 14,\n",
       "        1, 13,  6,  6,  3,  7,  7,  5,  5, 12,  6,  8,  0,  4,  7,  7,  0,\n",
       "       12, 14, 11,  0,  3, 11,  2,  3,  5,  4, 13,  1, 15,  7, 14, 15,  6,\n",
       "        1, 12, 10,  7,  2,  2,  8,  8, 11,  6, 15,  2, 10, 13,  7,  6,  3,\n",
       "       13,  3,  2, 14,  2,  8, 12, 15,  8, 13, 13, 14,  4, 12,  2,  5, 14,\n",
       "       12, 13, 13,  4, 12,  0, 11, 15,  8, 13, 13,  8,  9,  9,  2,  8,  0,\n",
       "        1,  2, 14, 15,  0,  3,  9, 11,  0,  8,  6,  7,  9, 11, 13, 11,  3,\n",
       "        7,  7, 13,  8, 12,  2, 14, 15,  7, 10, 13, 14, 12,  6, 14,  0,  0,\n",
       "        5,  3, 12, 13,  0,  7,  2,  8,  7, 13,  2,  9,  8, 14, 12, 14, 12,\n",
       "       13,  1, 12, 15,  4, 11, 15,  0,  9,  9,  5, 13, 10,  2,  2, 12, 15,\n",
       "        6])"
      ]
     },
     "execution_count": 105,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_pred_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "ee078822",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "635"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(test_pred_labels == test_labels['label'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "c3db9b5c",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>images</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>33</td>\n",
       "      <td>./Data/train/train/33.tif</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41</td>\n",
       "      <td>./Data/train/train/41.tif</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>65</td>\n",
       "      <td>./Data/train/train/65.tif</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>67</td>\n",
       "      <td>./Data/train/train/67.tif</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>100</td>\n",
       "      <td>./Data/train/train/100.tif</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>782</th>\n",
       "      <td>15582</td>\n",
       "      <td>./Data/train/train/15582.tif</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>783</th>\n",
       "      <td>15589</td>\n",
       "      <td>./Data/train/train/15589.tif</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>784</th>\n",
       "      <td>15596</td>\n",
       "      <td>./Data/train/train/15596.tif</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>788</th>\n",
       "      <td>15687</td>\n",
       "      <td>./Data/train/train/15687.tif</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>797</th>\n",
       "      <td>15832</td>\n",
       "      <td>./Data/train/train/15832.tif</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>165 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        id                        images  label\n",
       "2       33     ./Data/train/train/33.tif      6\n",
       "3       41     ./Data/train/train/41.tif      1\n",
       "4       65     ./Data/train/train/65.tif     14\n",
       "5       67     ./Data/train/train/67.tif      9\n",
       "7      100    ./Data/train/train/100.tif     10\n",
       "..     ...                           ...    ...\n",
       "782  15582  ./Data/train/train/15582.tif      3\n",
       "783  15589  ./Data/train/train/15589.tif      5\n",
       "784  15596  ./Data/train/train/15596.tif      8\n",
       "788  15687  ./Data/train/train/15687.tif     12\n",
       "797  15832  ./Data/train/train/15832.tif     11\n",
       "\n",
       "[165 rows x 3 columns]"
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mis_train_data = test_labels[test_pred_labels != test_labels['label']]\n",
    "mis_train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "b538b721",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>fscore</th>\n",
       "      <th>support</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.706897</td>\n",
       "      <td>0.759259</td>\n",
       "      <td>0.732143</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.674419</td>\n",
       "      <td>0.659091</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.906977</td>\n",
       "      <td>0.906977</td>\n",
       "      <td>0.906977</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.933333</td>\n",
       "      <td>0.913043</td>\n",
       "      <td>0.923077</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.925926</td>\n",
       "      <td>0.694444</td>\n",
       "      <td>0.793651</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.820513</td>\n",
       "      <td>0.640000</td>\n",
       "      <td>0.719101</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.847826</td>\n",
       "      <td>0.847826</td>\n",
       "      <td>0.847826</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.857143</td>\n",
       "      <td>0.807692</td>\n",
       "      <td>0.831683</td>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.868852</td>\n",
       "      <td>0.981481</td>\n",
       "      <td>0.921739</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.770833</td>\n",
       "      <td>0.787234</td>\n",
       "      <td>0.778947</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.808511</td>\n",
       "      <td>0.655172</td>\n",
       "      <td>0.723810</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.781818</td>\n",
       "      <td>0.767857</td>\n",
       "      <td>0.774775</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.618182</td>\n",
       "      <td>0.790698</td>\n",
       "      <td>0.693878</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.732394</td>\n",
       "      <td>0.812500</td>\n",
       "      <td>0.770370</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.900000</td>\n",
       "      <td>0.843750</td>\n",
       "      <td>0.870968</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.660377</td>\n",
       "      <td>0.813953</td>\n",
       "      <td>0.729167</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    precision    recall    fscore  support\n",
       "0    0.706897  0.759259  0.732143       54\n",
       "1    0.674419  0.659091  0.666667       44\n",
       "2    0.906977  0.906977  0.906977       43\n",
       "3    0.933333  0.913043  0.923077       46\n",
       "4    0.925926  0.694444  0.793651       36\n",
       "5    0.820513  0.640000  0.719101       50\n",
       "6    0.847826  0.847826  0.847826       46\n",
       "7    0.857143  0.807692  0.831683       52\n",
       "8    0.868852  0.981481  0.921739       54\n",
       "9    0.770833  0.787234  0.778947       47\n",
       "10   0.808511  0.655172  0.723810       58\n",
       "11   0.781818  0.767857  0.774775       56\n",
       "12   0.618182  0.790698  0.693878       43\n",
       "13   0.732394  0.812500  0.770370       64\n",
       "14   0.900000  0.843750  0.870968       64\n",
       "15   0.660377  0.813953  0.729167       43"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "precision, recall, fscore, support = score(test_labels['label'], test_pred_labels)\n",
    "pd.DataFrame({\n",
    "    'precision': precision,\n",
    "    'recall': recall,\n",
    "    'fscore': fscore,\n",
    "    'support': support\n",
    "})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "ed4db76c",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7927985931890755"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.mean(fscore)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "067c5096",
   "metadata": {},
   "source": [
    "#### On Validation Data"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3e665591",
   "metadata": {},
   "source": [
    "Generating Predictions for the Validation Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "4b085c0c",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>images</th>\n",
       "      <th>id</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>850</th>\n",
       "      <td>./Data/validation/validation/17801.tif</td>\n",
       "      <td>17801</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <td>./Data/validation/validation/17802.tif</td>\n",
       "      <td>17802</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>852</th>\n",
       "      <td>./Data/validation/validation/17803.tif</td>\n",
       "      <td>17803</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>798</th>\n",
       "      <td>./Data/validation/validation/17804.tif</td>\n",
       "      <td>17804</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>750</th>\n",
       "      <td>./Data/validation/validation/17805.tif</td>\n",
       "      <td>17805</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>246</th>\n",
       "      <td>./Data/validation/validation/18696.tif</td>\n",
       "      <td>18696</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>211</th>\n",
       "      <td>./Data/validation/validation/18697.tif</td>\n",
       "      <td>18697</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>760</th>\n",
       "      <td>./Data/validation/validation/18698.tif</td>\n",
       "      <td>18698</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>787</th>\n",
       "      <td>./Data/validation/validation/18699.tif</td>\n",
       "      <td>18699</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>891</th>\n",
       "      <td>./Data/validation/validation/18700.tif</td>\n",
       "      <td>18700</td>\n",
       "      <td>-1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>900 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                     images     id  label\n",
       "850  ./Data/validation/validation/17801.tif  17801     -1\n",
       "898  ./Data/validation/validation/17802.tif  17802     -1\n",
       "852  ./Data/validation/validation/17803.tif  17803     -1\n",
       "798  ./Data/validation/validation/17804.tif  17804     -1\n",
       "750  ./Data/validation/validation/17805.tif  17805     -1\n",
       "..                                      ...    ...    ...\n",
       "246  ./Data/validation/validation/18696.tif  18696     -1\n",
       "211  ./Data/validation/validation/18697.tif  18697     -1\n",
       "760  ./Data/validation/validation/18698.tif  18698     -1\n",
       "787  ./Data/validation/validation/18699.tif  18699     -1\n",
       "891  ./Data/validation/validation/18700.tif  18700     -1\n",
       "\n",
       "[900 rows x 3 columns]"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validation_data = pd.DataFrame({'images':['./Data/validation/validation/'+name for name in os.listdir('./Data/validation/validation/')]})\n",
    "validation_data['id'] = [name.split('.')[0] for name in os.listdir('./Data/validation/validation/')]\n",
    "validation_data['label'] = -1 # Simply added to prevent re-writing code\n",
    "validation_data.sort_values(by=['id'], inplace=True)\n",
    "validation_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "095dff17",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fine_tune_data_validation = np.load('./Data/Inception-ResNet-4P-validation-precompute/whole_validation_data_precomp.npy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "8685b228",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 1536)"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "69bbf5fc",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "fine_tune_data_validation_red = np.zeros((fine_tune_data_validation.shape[0], 5, N_COMPS))\n",
    "for i in range(5):\n",
    "    fine_tune_data_validation_red[:,i,:] = pcas[i].transform(fine_tune_data_validation[:,i,:])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "db7d1256",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "valid_preds = model1.predict(fine_tune_data_validation_red)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "18b17d2e",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "valid_pred_labels = np.argmax(valid_preds, axis=-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bece594",
   "metadata": {},
   "source": [
    "### Submission"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "304c47de",
   "metadata": {},
   "source": [
    "For Kaggle Submission"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "e7ca9d19",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>17801</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>17802</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>17803</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>17804</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>17805</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>895</th>\n",
       "      <td>18696</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>896</th>\n",
       "      <td>18697</td>\n",
       "      <td>12</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>897</th>\n",
       "      <td>18698</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>898</th>\n",
       "      <td>18699</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>899</th>\n",
       "      <td>18700</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>900 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        id  label\n",
       "0    17801      9\n",
       "1    17802      6\n",
       "2    17803      4\n",
       "3    17804      8\n",
       "4    17805      3\n",
       "..     ...    ...\n",
       "895  18696      7\n",
       "896  18697     12\n",
       "897  18698      8\n",
       "898  18699      3\n",
       "899  18700      4\n",
       "\n",
       "[900 rows x 2 columns]"
      ]
     },
     "execution_count": 116,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "submission_data = pd.DataFrame({'id':list(range(17801, 18701))})\n",
    "submission_data['label'] = valid_pred_labels\n",
    "submission_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "6eea7d9c",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "submission_data.to_csv('./Submissions/submission-inception-resnet-4p-PCA.csv', index=False)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}