ResNet-VGG-Inception-RoI-Model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "fd1efdf1",
   "metadata": {},
   "source": [
    "### Importing Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "0938de48",
   "metadata": {},
   "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.applications import EfficientNetV2L\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": 2,
   "id": "75b06a70",
   "metadata": {},
   "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": 3,
   "id": "fd2fe082",
   "metadata": {},
   "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": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "276e2059",
   "metadata": {},
   "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": 5,
   "id": "2aea9ea9",
   "metadata": {},
   "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": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b9ea19c9",
   "metadata": {},
   "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": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "29190fd4",
   "metadata": {},
   "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": "code",
   "execution_count": 8,
   "id": "f42b4e3c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16000, 100, 6)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roi_info = np.load('./Data/ROI-Extractions/train/train_roi_viz.npy')\n",
    "roi_info.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "60edccbe",
   "metadata": {},
   "outputs": [],
   "source": [
    "roi_count = [np.sum([roi[-1]==1 for roi in doc]) for doc in roi_info] # RoIs per document Image"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2e251c1",
   "metadata": {},
   "source": [
    "## Data Preparation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "85b2ccd1",
   "metadata": {},
   "source": [
    "We have already generated the Extracted Feature Vectors from the ResNet Model now we train using that, so we load those feature vectors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "37a36e2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_resnet = np.load('./Data/ResNet-kP-train-precompute/whole_train_data_precomp.npy')\n",
    "# fine_tune_data_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "4cc9fe34",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_vgg = np.load('./Data/VGG16-kP-train-precompute/whole_train_data_precomp.npy')\n",
    "# fine_tune_data_vgg.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "40e99544",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_inception_resnet = np.load('./Data/Inception-ResNet-kP-train-precompute/whole_train_data_precomp.npy')\n",
    "# fine_tune_data_inception_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "6e4e8b18",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data = np.concatenate([fine_tune_data_resnet, fine_tune_data_vgg, fine_tune_data_inception_resnet], axis=-1)\n",
    "# del fine_tune_data_resnet\n",
    "# del fine_tune_data_vgg\n",
    "# del fine_tune_data_inception_resnet\n",
    "# fine_tune_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c9821c99",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_data_precomp.npy', fine_tune_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "112978b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data = np.load('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_data_precomp.npy')\n",
    "# fine_tune_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "937e1e86",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_train = fine_tune_data[list(train_labels['id'])]\n",
    "# fine_tune_data_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0b93d55a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# roi_count_train = [roi_count[idx] for idx in list(train_labels['id'])]\n",
    "# fine_tune_data_train_no_pads = np.concatenate([fine_tune_data_train[i,:roi_count_train[i],:] for i in range(fine_tune_data_train.shape[0])])\n",
    "# fine_tune_data_train_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "559fcf15",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_train_data_precomp_no_pads.npy', fine_tune_data_train_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "dad95a7c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(242726, 4096)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roi_count_train = [roi_count[idx] for idx in list(train_labels['id'])]\n",
    "fine_tune_data_train_no_pads = np.load('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_train_data_precomp_no_pads.npy')\n",
    "fine_tune_data_train_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c00f0b32",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_test = fine_tune_data[list(test_labels['id'])]\n",
    "# fine_tune_data_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "23b86df5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# roi_count_test = [roi_count[idx] for idx in list(test_labels['id'])]\n",
    "# fine_tune_data_test_no_pads = np.concatenate([fine_tune_data_test[i,:roi_count_test[i],:] for i in range(fine_tune_data_test.shape[0])])\n",
    "# fine_tune_data_test_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "ac3d6656",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_test_data_precomp_no_pads.npy', fine_tune_data_test_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "922907fc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13154, 4096)"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roi_count_test = [roi_count[idx] for idx in list(test_labels['id'])]\n",
    "fine_tune_data_test_no_pads = np.load('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_test_data_precomp_no_pads.npy')\n",
    "fine_tune_data_test_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "568ffcb1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(16000,)"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels = labels['label'].to_numpy()\n",
    "fine_tune_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "3ba13d30",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_no_pads = np.concatenate([fine_tune_data[i,:roi_count[i],:] for i in range(fine_tune_data.shape[0])])\n",
    "# fine_tune_data_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "cbe27d98",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_data_precomp_no_pads.npy', fine_tune_data_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "a80c05c9",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fine_tune_data_roi = np.load('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_data_precomp_no_pads.npy')\n",
    "# fine_tune_data_roi.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "af08c830",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200,)"
      ]
     },
     "execution_count": 52,
     "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": 53,
   "id": "f7bad4fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200,)"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels_train_no_pads = [[label]*roi_count_train[i] for i, label in enumerate(list(fine_tune_labels_train))]\n",
    "fine_tune_labels_train_no_pads = np.array([label for label_set in fine_tune_labels_train_no_pads for label in label_set])\n",
    "fine_tune_labels_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "16f4bb93",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(800,)"
      ]
     },
     "execution_count": 54,
     "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": 55,
   "id": "edffbbcd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13154,)"
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_labels_test_no_pads = [[label]*roi_count_test[i] for i, label in enumerate(list(fine_tune_labels_test))]\n",
    "fine_tune_labels_test_no_pads = np.array([label for label_set in fine_tune_labels_test_no_pads for label in label_set])\n",
    "fine_tune_labels_test_no_pads.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": 56,
   "id": "96a2bdfe",
   "metadata": {},
   "outputs": [],
   "source": [
    "def ftmodel():\n",
    "    inputs = Input(shape=(4096,))\n",
    "    x = Flatten()(inputs)\n",
    "    x = BatchNormalization()(x)\n",
    "\n",
    "    top_dropout_rate = 0.5\n",
    "    x = Dropout(top_dropout_rate)(x)\n",
    "    x = Dense(2048,  kernel_initializer=\"he_normal\", kernel_regularizer=l2(5e-5), activation='relu')(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Dropout(top_dropout_rate)(x)\n",
    "    x = Dense(1024, kernel_initializer=\"he_normal\", kernel_regularizer=l2(5e-5), activation='relu')(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Dropout(top_dropout_rate)(x)\n",
    "    x = Dense(512, kernel_initializer=\"he_normal\", kernel_regularizer=l2(5e-5), activation='relu')(x)\n",
    "    x = BatchNormalization()(x)\n",
    "    x = Dropout(top_dropout_rate)(x)\n",
    "    outputs = Dense(16, activation='softmax', name='pred', kernel_regularizer=l2(5e-5))(x)\n",
    "\n",
    "    model1 = tf.keras.Model(inputs, outputs, name='RestNet-VGG-Inception-ResNet-RoI-FineTune')\n",
    "    return model1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "c6574006",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"RestNet-VGG-Inception-ResNet-RoI-FineTune\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_2 (InputLayer)        [(None, 4096)]            0         \n",
      "                                                                 \n",
      " flatten_1 (Flatten)         (None, 4096)              0         \n",
      "                                                                 \n",
      " batch_normalization_4 (Batc  (None, 4096)             16384     \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_4 (Dropout)         (None, 4096)              0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 2048)              8390656   \n",
      "                                                                 \n",
      " batch_normalization_5 (Batc  (None, 2048)             8192      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_5 (Dropout)         (None, 2048)              0         \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 1024)              2098176   \n",
      "                                                                 \n",
      " batch_normalization_6 (Batc  (None, 1024)             4096      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_6 (Dropout)         (None, 1024)              0         \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 512)               524800    \n",
      "                                                                 \n",
      " batch_normalization_7 (Batc  (None, 512)              2048      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_7 (Dropout)         (None, 512)               0         \n",
      "                                                                 \n",
      " pred (Dense)                (None, 16)                8208      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 11,052,560\n",
      "Trainable params: 11,037,200\n",
      "Non-trainable params: 15,360\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model1 = ftmodel()\n",
    "model1.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "04f47ff8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# tf.keras.utils.plot_model(model1, show_shapes=True, to_file='./Images/ResNet-VGG-Inception-ResNet-RoI-Model.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e56505e2",
   "metadata": {},
   "source": [
    "### Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "96524578",
   "metadata": {},
   "outputs": [],
   "source": [
    "model1.compile(loss = 'sparse_categorical_crossentropy', metrics = ['accuracy'], optimizer = Adam(learning_rate = 0.0003))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "20df8897",
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 2000\n",
    "checkpoint_filepaths = ['./Models/ResNet-VGG-Inception-ResNet-RoI/auto-ResNet-VGG-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": 60,
   "id": "03354278",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/2000\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-10 22:24:02.604401: W tensorflow/core/platform/profile_utils/cpu_utils.cc:128] Failed to get CPU frequency: 0 Hz\n",
      "2022-10-10 22:24:03.067231: 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": [
      "242/242 [==============================] - ETA: 0s - loss: 3.2974 - accuracy: 0.1500"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-10 22:24:08.967862: 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": [
      "242/242 [==============================] - 7s 24ms/step - loss: 3.2974 - accuracy: 0.1500 - val_loss: 2.8768 - val_accuracy: 0.2039\n",
      "Epoch 2/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.8779 - accuracy: 0.2004 - val_loss: 2.7386 - val_accuracy: 0.2454\n",
      "Epoch 3/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.7600 - accuracy: 0.2286 - val_loss: 2.6420 - val_accuracy: 0.2619\n",
      "Epoch 4/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.6930 - accuracy: 0.2467 - val_loss: 2.5913 - val_accuracy: 0.2850\n",
      "Epoch 5/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.6408 - accuracy: 0.2614 - val_loss: 2.5455 - val_accuracy: 0.2959\n",
      "Epoch 6/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.6004 - accuracy: 0.2713 - val_loss: 2.5096 - val_accuracy: 0.2978\n",
      "Epoch 7/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.5633 - accuracy: 0.2803 - val_loss: 2.4751 - val_accuracy: 0.3104\n",
      "Epoch 8/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.5311 - accuracy: 0.2884 - val_loss: 2.4447 - val_accuracy: 0.3182\n",
      "Epoch 9/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 2.5051 - accuracy: 0.2941 - val_loss: 2.4280 - val_accuracy: 0.3183\n",
      "Epoch 10/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.4771 - accuracy: 0.2987 - val_loss: 2.4031 - val_accuracy: 0.3273\n",
      "Epoch 11/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.4527 - accuracy: 0.3045 - val_loss: 2.3759 - val_accuracy: 0.3334\n",
      "Epoch 12/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.4341 - accuracy: 0.3076 - val_loss: 2.3654 - val_accuracy: 0.3324\n",
      "Epoch 13/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.4120 - accuracy: 0.3122 - val_loss: 2.3380 - val_accuracy: 0.3354\n",
      "Epoch 14/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.3894 - accuracy: 0.3166 - val_loss: 2.3193 - val_accuracy: 0.3419\n",
      "Epoch 15/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.3699 - accuracy: 0.3211 - val_loss: 2.3046 - val_accuracy: 0.3452\n",
      "Epoch 16/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.3508 - accuracy: 0.3245 - val_loss: 2.2857 - val_accuracy: 0.3496\n",
      "Epoch 17/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.3320 - accuracy: 0.3292 - val_loss: 2.2771 - val_accuracy: 0.3514\n",
      "Epoch 18/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.3167 - accuracy: 0.3321 - val_loss: 2.2623 - val_accuracy: 0.3489\n",
      "Epoch 19/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.3006 - accuracy: 0.3356 - val_loss: 2.2472 - val_accuracy: 0.3572\n",
      "Epoch 20/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.2844 - accuracy: 0.3399 - val_loss: 2.2347 - val_accuracy: 0.3561\n",
      "Epoch 21/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 2.2720 - accuracy: 0.3419 - val_loss: 2.2229 - val_accuracy: 0.3591\n",
      "Epoch 22/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.2596 - accuracy: 0.3456 - val_loss: 2.2246 - val_accuracy: 0.3534\n",
      "Epoch 23/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.2462 - accuracy: 0.3493 - val_loss: 2.2023 - val_accuracy: 0.3653\n",
      "Epoch 24/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 2.2338 - accuracy: 0.3519 - val_loss: 2.2088 - val_accuracy: 0.3653\n",
      "Epoch 25/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.2250 - accuracy: 0.3552 - val_loss: 2.1932 - val_accuracy: 0.3701\n",
      "Epoch 26/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.2150 - accuracy: 0.3582 - val_loss: 2.1948 - val_accuracy: 0.3662\n",
      "Epoch 27/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.2044 - accuracy: 0.3605 - val_loss: 2.1815 - val_accuracy: 0.3749\n",
      "Epoch 28/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1941 - accuracy: 0.3642 - val_loss: 2.1738 - val_accuracy: 0.3737\n",
      "Epoch 29/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1861 - accuracy: 0.3666 - val_loss: 2.1730 - val_accuracy: 0.3730\n",
      "Epoch 30/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1768 - accuracy: 0.3689 - val_loss: 2.1744 - val_accuracy: 0.3736\n",
      "Epoch 31/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1700 - accuracy: 0.3713 - val_loss: 2.1632 - val_accuracy: 0.3796\n",
      "Epoch 32/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1611 - accuracy: 0.3732 - val_loss: 2.1639 - val_accuracy: 0.3804\n",
      "Epoch 33/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1536 - accuracy: 0.3759 - val_loss: 2.1614 - val_accuracy: 0.3836\n",
      "Epoch 34/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1472 - accuracy: 0.3788 - val_loss: 2.1585 - val_accuracy: 0.3812\n",
      "Epoch 35/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1421 - accuracy: 0.3816 - val_loss: 2.1564 - val_accuracy: 0.3794\n",
      "Epoch 36/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1344 - accuracy: 0.3850 - val_loss: 2.1583 - val_accuracy: 0.3816\n",
      "Epoch 37/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1308 - accuracy: 0.3844 - val_loss: 2.1579 - val_accuracy: 0.3819\n",
      "Epoch 38/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1228 - accuracy: 0.3876 - val_loss: 2.1531 - val_accuracy: 0.3846\n",
      "Epoch 39/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1149 - accuracy: 0.3914 - val_loss: 2.1478 - val_accuracy: 0.3841\n",
      "Epoch 40/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1132 - accuracy: 0.3922 - val_loss: 2.1559 - val_accuracy: 0.3800\n",
      "Epoch 41/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1081 - accuracy: 0.3933 - val_loss: 2.1517 - val_accuracy: 0.3896\n",
      "Epoch 42/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1030 - accuracy: 0.3957 - val_loss: 2.1461 - val_accuracy: 0.3873\n",
      "Epoch 43/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.1006 - accuracy: 0.3969 - val_loss: 2.1487 - val_accuracy: 0.3873\n",
      "Epoch 44/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.0942 - accuracy: 0.3989 - val_loss: 2.1547 - val_accuracy: 0.3871\n",
      "Epoch 45/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.0929 - accuracy: 0.3992 - val_loss: 2.1479 - val_accuracy: 0.3901\n",
      "Epoch 46/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.0857 - accuracy: 0.4039 - val_loss: 2.1542 - val_accuracy: 0.3897\n",
      "Epoch 47/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.0824 - accuracy: 0.4040 - val_loss: 2.1505 - val_accuracy: 0.3879\n",
      "Epoch 48/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0803 - accuracy: 0.4052 - val_loss: 2.1516 - val_accuracy: 0.3862\n",
      "Epoch 49/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0748 - accuracy: 0.4074 - val_loss: 2.1487 - val_accuracy: 0.3895\n",
      "Epoch 50/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0710 - accuracy: 0.4078 - val_loss: 2.1529 - val_accuracy: 0.3893\n",
      "Epoch 51/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0668 - accuracy: 0.4103 - val_loss: 2.1496 - val_accuracy: 0.3911\n",
      "Epoch 52/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0633 - accuracy: 0.4112 - val_loss: 2.1509 - val_accuracy: 0.3924\n",
      "Epoch 53/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.0633 - accuracy: 0.4135 - val_loss: 2.1449 - val_accuracy: 0.3960\n",
      "Epoch 54/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 2.0583 - accuracy: 0.4143 - val_loss: 2.1498 - val_accuracy: 0.3955\n",
      "Epoch 55/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0563 - accuracy: 0.4159 - val_loss: 2.1520 - val_accuracy: 0.3942\n",
      "Epoch 56/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0529 - accuracy: 0.4176 - val_loss: 2.1604 - val_accuracy: 0.3917\n",
      "Epoch 57/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0535 - accuracy: 0.4170 - val_loss: 2.1562 - val_accuracy: 0.3944\n",
      "Epoch 58/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 2.0486 - accuracy: 0.4190 - val_loss: 2.1569 - val_accuracy: 0.3932\n",
      "Epoch 59/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0474 - accuracy: 0.4191 - val_loss: 2.1527 - val_accuracy: 0.3939\n",
      "Epoch 60/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0423 - accuracy: 0.4222 - val_loss: 2.1542 - val_accuracy: 0.3961\n",
      "Epoch 61/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0411 - accuracy: 0.4228 - val_loss: 2.1559 - val_accuracy: 0.3977\n",
      "Epoch 62/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0372 - accuracy: 0.4248 - val_loss: 2.1544 - val_accuracy: 0.3960\n",
      "Epoch 63/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0355 - accuracy: 0.4250 - val_loss: 2.1520 - val_accuracy: 0.3984\n",
      "Epoch 64/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0338 - accuracy: 0.4250 - val_loss: 2.1587 - val_accuracy: 0.3988\n",
      "Epoch 65/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.0338 - accuracy: 0.4265 - val_loss: 2.1593 - val_accuracy: 0.3979\n",
      "Epoch 66/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0306 - accuracy: 0.4278 - val_loss: 2.1604 - val_accuracy: 0.3968\n",
      "Epoch 67/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0290 - accuracy: 0.4282 - val_loss: 2.1646 - val_accuracy: 0.3949\n",
      "Epoch 68/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0252 - accuracy: 0.4300 - val_loss: 2.1551 - val_accuracy: 0.3965\n",
      "Epoch 69/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0228 - accuracy: 0.4317 - val_loss: 2.1629 - val_accuracy: 0.3976\n",
      "Epoch 70/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 2.0236 - accuracy: 0.4301 - val_loss: 2.1671 - val_accuracy: 0.3990\n",
      "Epoch 71/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0216 - accuracy: 0.4313 - val_loss: 2.1657 - val_accuracy: 0.3972\n",
      "Epoch 72/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0188 - accuracy: 0.4328 - val_loss: 2.1704 - val_accuracy: 0.4003\n",
      "Epoch 73/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0165 - accuracy: 0.4340 - val_loss: 2.1603 - val_accuracy: 0.3984\n",
      "Epoch 74/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0127 - accuracy: 0.4365 - val_loss: 2.1709 - val_accuracy: 0.3986\n",
      "Epoch 75/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 2.0160 - accuracy: 0.4358 - val_loss: 2.1691 - val_accuracy: 0.3977\n",
      "Epoch 76/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 2.0107 - accuracy: 0.4371 - val_loss: 2.1746 - val_accuracy: 0.4019\n",
      "Epoch 77/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0101 - accuracy: 0.4368 - val_loss: 2.1699 - val_accuracy: 0.4066\n",
      "Epoch 78/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0090 - accuracy: 0.4380 - val_loss: 2.1687 - val_accuracy: 0.4000\n",
      "Epoch 79/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0085 - accuracy: 0.4389 - val_loss: 2.1759 - val_accuracy: 0.4017\n",
      "Epoch 80/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0071 - accuracy: 0.4397 - val_loss: 2.1709 - val_accuracy: 0.3996\n",
      "Epoch 81/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0055 - accuracy: 0.4397 - val_loss: 2.1677 - val_accuracy: 0.4055\n",
      "Epoch 82/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0034 - accuracy: 0.4409 - val_loss: 2.1726 - val_accuracy: 0.3980\n",
      "Epoch 83/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0024 - accuracy: 0.4420 - val_loss: 2.1683 - val_accuracy: 0.4054\n",
      "Epoch 84/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 2.0011 - accuracy: 0.4424 - val_loss: 2.1708 - val_accuracy: 0.4047\n",
      "Epoch 85/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9990 - accuracy: 0.4435 - val_loss: 2.1726 - val_accuracy: 0.4051\n",
      "Epoch 86/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9967 - accuracy: 0.4451 - val_loss: 2.1798 - val_accuracy: 0.4012\n",
      "Epoch 87/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.9974 - accuracy: 0.4453 - val_loss: 2.1770 - val_accuracy: 0.4055\n",
      "Epoch 88/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9967 - accuracy: 0.4454 - val_loss: 2.1731 - val_accuracy: 0.4040\n",
      "Epoch 89/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9960 - accuracy: 0.4451 - val_loss: 2.1779 - val_accuracy: 0.4050\n",
      "Epoch 90/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9931 - accuracy: 0.4470 - val_loss: 2.1788 - val_accuracy: 0.4038\n",
      "Epoch 91/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9924 - accuracy: 0.4470 - val_loss: 2.1826 - val_accuracy: 0.4047\n",
      "Epoch 92/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9907 - accuracy: 0.4474 - val_loss: 2.1796 - val_accuracy: 0.4038\n",
      "Epoch 93/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9905 - accuracy: 0.4480 - val_loss: 2.1741 - val_accuracy: 0.4057\n",
      "Epoch 94/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9891 - accuracy: 0.4485 - val_loss: 2.1851 - val_accuracy: 0.4037\n",
      "Epoch 95/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9861 - accuracy: 0.4504 - val_loss: 2.1824 - val_accuracy: 0.4047\n",
      "Epoch 96/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9875 - accuracy: 0.4499 - val_loss: 2.1819 - val_accuracy: 0.4040\n",
      "Epoch 97/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9883 - accuracy: 0.4493 - val_loss: 2.1903 - val_accuracy: 0.4018\n",
      "Epoch 98/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9824 - accuracy: 0.4520 - val_loss: 2.1825 - val_accuracy: 0.4022\n",
      "Epoch 99/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9822 - accuracy: 0.4526 - val_loss: 2.1901 - val_accuracy: 0.4080\n",
      "Epoch 100/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9831 - accuracy: 0.4522 - val_loss: 2.1878 - val_accuracy: 0.3992\n",
      "Epoch 101/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9801 - accuracy: 0.4533 - val_loss: 2.1905 - val_accuracy: 0.4039\n",
      "Epoch 102/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9793 - accuracy: 0.4532 - val_loss: 2.1843 - val_accuracy: 0.4063\n",
      "Epoch 103/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.9816 - accuracy: 0.4530 - val_loss: 2.1851 - val_accuracy: 0.4045\n",
      "Epoch 104/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9775 - accuracy: 0.4543 - val_loss: 2.1873 - val_accuracy: 0.4079\n",
      "Epoch 105/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9766 - accuracy: 0.4548 - val_loss: 2.1884 - val_accuracy: 0.4031\n",
      "Epoch 106/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9751 - accuracy: 0.4559 - val_loss: 2.1879 - val_accuracy: 0.4044\n",
      "Epoch 107/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9742 - accuracy: 0.4559 - val_loss: 2.1908 - val_accuracy: 0.4028\n",
      "Epoch 108/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9750 - accuracy: 0.4555 - val_loss: 2.1865 - val_accuracy: 0.4056\n",
      "Epoch 109/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9730 - accuracy: 0.4568 - val_loss: 2.1870 - val_accuracy: 0.4048\n",
      "Epoch 110/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9711 - accuracy: 0.4574 - val_loss: 2.1978 - val_accuracy: 0.4040\n",
      "Epoch 111/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9722 - accuracy: 0.4586 - val_loss: 2.1882 - val_accuracy: 0.4041\n",
      "Epoch 112/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9678 - accuracy: 0.4586 - val_loss: 2.1914 - val_accuracy: 0.4030\n",
      "Epoch 113/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9705 - accuracy: 0.4578 - val_loss: 2.1888 - val_accuracy: 0.4089\n",
      "Epoch 114/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9692 - accuracy: 0.4591 - val_loss: 2.1907 - val_accuracy: 0.4087\n",
      "Epoch 115/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9661 - accuracy: 0.4591 - val_loss: 2.1936 - val_accuracy: 0.4088\n",
      "Epoch 116/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9691 - accuracy: 0.4597 - val_loss: 2.1973 - val_accuracy: 0.4046\n",
      "Epoch 117/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9656 - accuracy: 0.4604 - val_loss: 2.1885 - val_accuracy: 0.4078\n",
      "Epoch 118/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9664 - accuracy: 0.4608 - val_loss: 2.1911 - val_accuracy: 0.4068\n",
      "Epoch 119/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9636 - accuracy: 0.4621 - val_loss: 2.1972 - val_accuracy: 0.4045\n",
      "Epoch 120/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9637 - accuracy: 0.4614 - val_loss: 2.1878 - val_accuracy: 0.4086\n",
      "Epoch 121/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9637 - accuracy: 0.4613 - val_loss: 2.1886 - val_accuracy: 0.4077\n",
      "Epoch 122/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9633 - accuracy: 0.4618 - val_loss: 2.1924 - val_accuracy: 0.4079\n",
      "Epoch 123/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9609 - accuracy: 0.4622 - val_loss: 2.1975 - val_accuracy: 0.4057\n",
      "Epoch 124/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9616 - accuracy: 0.4621 - val_loss: 2.1972 - val_accuracy: 0.4091\n",
      "Epoch 125/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9618 - accuracy: 0.4626 - val_loss: 2.1976 - val_accuracy: 0.4082\n",
      "Epoch 126/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9597 - accuracy: 0.4632 - val_loss: 2.2005 - val_accuracy: 0.4071\n",
      "Epoch 127/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9591 - accuracy: 0.4635 - val_loss: 2.2021 - val_accuracy: 0.4076\n",
      "Epoch 128/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9590 - accuracy: 0.4638 - val_loss: 2.1999 - val_accuracy: 0.4082\n",
      "Epoch 129/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9562 - accuracy: 0.4653 - val_loss: 2.2029 - val_accuracy: 0.4077\n",
      "Epoch 130/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9559 - accuracy: 0.4658 - val_loss: 2.1948 - val_accuracy: 0.4098\n",
      "Epoch 131/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9562 - accuracy: 0.4650 - val_loss: 2.2031 - val_accuracy: 0.4083\n",
      "Epoch 132/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9554 - accuracy: 0.4666 - val_loss: 2.2059 - val_accuracy: 0.4051\n",
      "Epoch 133/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9550 - accuracy: 0.4659 - val_loss: 2.1997 - val_accuracy: 0.4088\n",
      "Epoch 134/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9545 - accuracy: 0.4662 - val_loss: 2.2079 - val_accuracy: 0.4056\n",
      "Epoch 135/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9534 - accuracy: 0.4674 - val_loss: 2.2048 - val_accuracy: 0.4066\n",
      "Epoch 136/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9511 - accuracy: 0.4679 - val_loss: 2.2020 - val_accuracy: 0.4095\n",
      "Epoch 137/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9494 - accuracy: 0.4688 - val_loss: 2.2031 - val_accuracy: 0.4073\n",
      "Epoch 138/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9516 - accuracy: 0.4679 - val_loss: 2.2059 - val_accuracy: 0.4077\n",
      "Epoch 139/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9511 - accuracy: 0.4677 - val_loss: 2.2045 - val_accuracy: 0.4055\n",
      "Epoch 140/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9517 - accuracy: 0.4676 - val_loss: 2.1910 - val_accuracy: 0.4092\n",
      "Epoch 141/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9501 - accuracy: 0.4686 - val_loss: 2.2017 - val_accuracy: 0.4117\n",
      "Epoch 142/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9485 - accuracy: 0.4692 - val_loss: 2.2030 - val_accuracy: 0.4111\n",
      "Epoch 143/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9454 - accuracy: 0.4702 - val_loss: 2.2090 - val_accuracy: 0.4111\n",
      "Epoch 144/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9491 - accuracy: 0.4700 - val_loss: 2.2099 - val_accuracy: 0.4085\n",
      "Epoch 145/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9476 - accuracy: 0.4710 - val_loss: 2.2039 - val_accuracy: 0.4097\n",
      "Epoch 146/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9467 - accuracy: 0.4705 - val_loss: 2.2112 - val_accuracy: 0.4135\n",
      "Epoch 147/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9455 - accuracy: 0.4710 - val_loss: 2.2087 - val_accuracy: 0.4094\n",
      "Epoch 148/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9443 - accuracy: 0.4706 - val_loss: 2.2078 - val_accuracy: 0.4060\n",
      "Epoch 149/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9444 - accuracy: 0.4707 - val_loss: 2.2097 - val_accuracy: 0.4098\n",
      "Epoch 150/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9453 - accuracy: 0.4698 - val_loss: 2.2077 - val_accuracy: 0.4057\n",
      "Epoch 151/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9437 - accuracy: 0.4729 - val_loss: 2.2085 - val_accuracy: 0.4081\n",
      "Epoch 152/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9431 - accuracy: 0.4716 - val_loss: 2.2103 - val_accuracy: 0.4070\n",
      "Epoch 153/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9424 - accuracy: 0.4721 - val_loss: 2.2141 - val_accuracy: 0.4054\n",
      "Epoch 154/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9428 - accuracy: 0.4723 - val_loss: 2.2060 - val_accuracy: 0.4106\n",
      "Epoch 155/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9418 - accuracy: 0.4729 - val_loss: 2.2068 - val_accuracy: 0.4077\n",
      "Epoch 156/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9417 - accuracy: 0.4717 - val_loss: 2.2152 - val_accuracy: 0.4095\n",
      "Epoch 157/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9370 - accuracy: 0.4748 - val_loss: 2.2106 - val_accuracy: 0.4092\n",
      "Epoch 158/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9363 - accuracy: 0.4735 - val_loss: 2.2105 - val_accuracy: 0.4114\n",
      "Epoch 159/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9410 - accuracy: 0.4735 - val_loss: 2.2150 - val_accuracy: 0.4088\n",
      "Epoch 160/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9382 - accuracy: 0.4735 - val_loss: 2.2063 - val_accuracy: 0.4127\n",
      "Epoch 161/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9373 - accuracy: 0.4747 - val_loss: 2.2107 - val_accuracy: 0.4098\n",
      "Epoch 162/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9375 - accuracy: 0.4749 - val_loss: 2.2158 - val_accuracy: 0.4094\n",
      "Epoch 163/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9388 - accuracy: 0.4742 - val_loss: 2.2142 - val_accuracy: 0.4081\n",
      "Epoch 164/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9385 - accuracy: 0.4745 - val_loss: 2.2192 - val_accuracy: 0.4082\n",
      "Epoch 165/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9344 - accuracy: 0.4750 - val_loss: 2.2207 - val_accuracy: 0.4064\n",
      "Epoch 166/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9345 - accuracy: 0.4764 - val_loss: 2.2144 - val_accuracy: 0.4105\n",
      "Epoch 167/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9326 - accuracy: 0.4769 - val_loss: 2.2149 - val_accuracy: 0.4117\n",
      "Epoch 168/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9326 - accuracy: 0.4767 - val_loss: 2.2215 - val_accuracy: 0.4086\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 169/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9335 - accuracy: 0.4766 - val_loss: 2.2255 - val_accuracy: 0.4076\n",
      "Epoch 170/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9348 - accuracy: 0.4753 - val_loss: 2.2119 - val_accuracy: 0.4101\n",
      "Epoch 171/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9307 - accuracy: 0.4784 - val_loss: 2.2173 - val_accuracy: 0.4083\n",
      "Epoch 172/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9322 - accuracy: 0.4772 - val_loss: 2.2181 - val_accuracy: 0.4079\n",
      "Epoch 173/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9323 - accuracy: 0.4771 - val_loss: 2.2231 - val_accuracy: 0.4081\n",
      "Epoch 174/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9338 - accuracy: 0.4751 - val_loss: 2.2217 - val_accuracy: 0.4096\n",
      "Epoch 175/2000\n",
      "242/242 [==============================] - 7s 28ms/step - loss: 1.9288 - accuracy: 0.4795 - val_loss: 2.2196 - val_accuracy: 0.4109\n",
      "Epoch 176/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9322 - accuracy: 0.4777 - val_loss: 2.2235 - val_accuracy: 0.4113\n",
      "Epoch 177/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9312 - accuracy: 0.4782 - val_loss: 2.2235 - val_accuracy: 0.4098\n",
      "Epoch 178/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9311 - accuracy: 0.4786 - val_loss: 2.2179 - val_accuracy: 0.4074\n",
      "Epoch 179/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9295 - accuracy: 0.4799 - val_loss: 2.2163 - val_accuracy: 0.4113\n",
      "Epoch 180/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9307 - accuracy: 0.4782 - val_loss: 2.2213 - val_accuracy: 0.4083\n",
      "Epoch 181/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9281 - accuracy: 0.4795 - val_loss: 2.2142 - val_accuracy: 0.4117\n",
      "Epoch 182/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9253 - accuracy: 0.4804 - val_loss: 2.2272 - val_accuracy: 0.4088\n",
      "Epoch 183/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9266 - accuracy: 0.4785 - val_loss: 2.2186 - val_accuracy: 0.4111\n",
      "Epoch 184/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9258 - accuracy: 0.4802 - val_loss: 2.2234 - val_accuracy: 0.4096\n",
      "Epoch 185/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9274 - accuracy: 0.4798 - val_loss: 2.2204 - val_accuracy: 0.4073\n",
      "Epoch 186/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9251 - accuracy: 0.4803 - val_loss: 2.2189 - val_accuracy: 0.4123\n",
      "Epoch 187/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9263 - accuracy: 0.4813 - val_loss: 2.2105 - val_accuracy: 0.4123\n",
      "Epoch 188/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9251 - accuracy: 0.4800 - val_loss: 2.2244 - val_accuracy: 0.4091\n",
      "Epoch 189/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9237 - accuracy: 0.4801 - val_loss: 2.2223 - val_accuracy: 0.4140\n",
      "Epoch 190/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9237 - accuracy: 0.4811 - val_loss: 2.2196 - val_accuracy: 0.4120\n",
      "Epoch 191/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9260 - accuracy: 0.4807 - val_loss: 2.2200 - val_accuracy: 0.4122\n",
      "Epoch 192/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9258 - accuracy: 0.4810 - val_loss: 2.2295 - val_accuracy: 0.4070\n",
      "Epoch 193/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.9234 - accuracy: 0.4801 - val_loss: 2.2210 - val_accuracy: 0.4117\n",
      "Epoch 194/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9227 - accuracy: 0.4829 - val_loss: 2.2181 - val_accuracy: 0.4110\n",
      "Epoch 195/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9214 - accuracy: 0.4831 - val_loss: 2.2249 - val_accuracy: 0.4093\n",
      "Epoch 196/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9225 - accuracy: 0.4817 - val_loss: 2.2274 - val_accuracy: 0.4100\n",
      "Epoch 197/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9220 - accuracy: 0.4822 - val_loss: 2.2188 - val_accuracy: 0.4077\n",
      "Epoch 198/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9224 - accuracy: 0.4817 - val_loss: 2.2154 - val_accuracy: 0.4125\n",
      "Epoch 199/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9220 - accuracy: 0.4830 - val_loss: 2.2269 - val_accuracy: 0.4138\n",
      "Epoch 200/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9208 - accuracy: 0.4830 - val_loss: 2.2267 - val_accuracy: 0.4096\n",
      "Epoch 201/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9223 - accuracy: 0.4820 - val_loss: 2.2225 - val_accuracy: 0.4117\n",
      "Epoch 202/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9187 - accuracy: 0.4848 - val_loss: 2.2288 - val_accuracy: 0.4095\n",
      "Epoch 203/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9192 - accuracy: 0.4846 - val_loss: 2.2359 - val_accuracy: 0.4096\n",
      "Epoch 204/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9199 - accuracy: 0.4830 - val_loss: 2.2282 - val_accuracy: 0.4134\n",
      "Epoch 205/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9214 - accuracy: 0.4835 - val_loss: 2.2258 - val_accuracy: 0.4090\n",
      "Epoch 206/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9186 - accuracy: 0.4826 - val_loss: 2.2282 - val_accuracy: 0.4122\n",
      "Epoch 207/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9178 - accuracy: 0.4841 - val_loss: 2.2193 - val_accuracy: 0.4117\n",
      "Epoch 208/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9155 - accuracy: 0.4841 - val_loss: 2.2275 - val_accuracy: 0.4118\n",
      "Epoch 209/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9186 - accuracy: 0.4834 - val_loss: 2.2211 - val_accuracy: 0.4117\n",
      "Epoch 210/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9167 - accuracy: 0.4855 - val_loss: 2.2336 - val_accuracy: 0.4101\n",
      "Epoch 211/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9149 - accuracy: 0.4855 - val_loss: 2.2253 - val_accuracy: 0.4131\n",
      "Epoch 212/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9179 - accuracy: 0.4843 - val_loss: 2.2323 - val_accuracy: 0.4121\n",
      "Epoch 213/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9123 - accuracy: 0.4861 - val_loss: 2.2220 - val_accuracy: 0.4111\n",
      "Epoch 214/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9175 - accuracy: 0.4845 - val_loss: 2.2189 - val_accuracy: 0.4107\n",
      "Epoch 215/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9138 - accuracy: 0.4848 - val_loss: 2.2293 - val_accuracy: 0.4123\n",
      "Epoch 216/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9142 - accuracy: 0.4858 - val_loss: 2.2243 - val_accuracy: 0.4154\n",
      "Epoch 217/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9118 - accuracy: 0.4867 - val_loss: 2.2267 - val_accuracy: 0.4149\n",
      "Epoch 218/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9139 - accuracy: 0.4853 - val_loss: 2.2248 - val_accuracy: 0.4143\n",
      "Epoch 219/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9137 - accuracy: 0.4861 - val_loss: 2.2223 - val_accuracy: 0.4101\n",
      "Epoch 220/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9132 - accuracy: 0.4859 - val_loss: 2.2258 - val_accuracy: 0.4098\n",
      "Epoch 221/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9136 - accuracy: 0.4870 - val_loss: 2.2168 - val_accuracy: 0.4124\n",
      "Epoch 222/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9134 - accuracy: 0.4873 - val_loss: 2.2194 - val_accuracy: 0.4151\n",
      "Epoch 223/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9124 - accuracy: 0.4867 - val_loss: 2.2219 - val_accuracy: 0.4146\n",
      "Epoch 224/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9111 - accuracy: 0.4874 - val_loss: 2.2286 - val_accuracy: 0.4108\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 225/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9111 - accuracy: 0.4869 - val_loss: 2.2297 - val_accuracy: 0.4129\n",
      "Epoch 226/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9095 - accuracy: 0.4876 - val_loss: 2.2353 - val_accuracy: 0.4119\n",
      "Epoch 227/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9088 - accuracy: 0.4876 - val_loss: 2.2317 - val_accuracy: 0.4134\n",
      "Epoch 228/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.9116 - accuracy: 0.4864 - val_loss: 2.2256 - val_accuracy: 0.4135\n",
      "Epoch 229/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9087 - accuracy: 0.4884 - val_loss: 2.2191 - val_accuracy: 0.4120\n",
      "Epoch 230/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.9093 - accuracy: 0.4871 - val_loss: 2.2301 - val_accuracy: 0.4101\n",
      "Epoch 231/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9090 - accuracy: 0.4891 - val_loss: 2.2245 - val_accuracy: 0.4117\n",
      "Epoch 232/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9070 - accuracy: 0.4888 - val_loss: 2.2299 - val_accuracy: 0.4127\n",
      "Epoch 233/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9110 - accuracy: 0.4874 - val_loss: 2.2287 - val_accuracy: 0.4105\n",
      "Epoch 234/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9078 - accuracy: 0.4881 - val_loss: 2.2205 - val_accuracy: 0.4162\n",
      "Epoch 235/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9078 - accuracy: 0.4880 - val_loss: 2.2266 - val_accuracy: 0.4117\n",
      "Epoch 236/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9086 - accuracy: 0.4871 - val_loss: 2.2247 - val_accuracy: 0.4094\n",
      "Epoch 237/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9075 - accuracy: 0.4876 - val_loss: 2.2299 - val_accuracy: 0.4115\n",
      "Epoch 238/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9083 - accuracy: 0.4884 - val_loss: 2.2350 - val_accuracy: 0.4111\n",
      "Epoch 239/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9077 - accuracy: 0.4890 - val_loss: 2.2333 - val_accuracy: 0.4133\n",
      "Epoch 240/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9065 - accuracy: 0.4897 - val_loss: 2.2325 - val_accuracy: 0.4096\n",
      "Epoch 241/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9059 - accuracy: 0.4896 - val_loss: 2.2321 - val_accuracy: 0.4107\n",
      "Epoch 242/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9075 - accuracy: 0.4887 - val_loss: 2.2318 - val_accuracy: 0.4096\n",
      "Epoch 243/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9049 - accuracy: 0.4893 - val_loss: 2.2310 - val_accuracy: 0.4101\n",
      "Epoch 244/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9058 - accuracy: 0.4890 - val_loss: 2.2303 - val_accuracy: 0.4098\n",
      "Epoch 245/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9053 - accuracy: 0.4906 - val_loss: 2.2331 - val_accuracy: 0.4095\n",
      "Epoch 246/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9072 - accuracy: 0.4895 - val_loss: 2.2280 - val_accuracy: 0.4104\n",
      "Epoch 247/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9046 - accuracy: 0.4910 - val_loss: 2.2322 - val_accuracy: 0.4115\n",
      "Epoch 248/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9045 - accuracy: 0.4898 - val_loss: 2.2272 - val_accuracy: 0.4129\n",
      "Epoch 249/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9063 - accuracy: 0.4899 - val_loss: 2.2251 - val_accuracy: 0.4130\n",
      "Epoch 250/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9023 - accuracy: 0.4907 - val_loss: 2.2380 - val_accuracy: 0.4098\n",
      "Epoch 251/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.9013 - accuracy: 0.4914 - val_loss: 2.2306 - val_accuracy: 0.4088\n",
      "Epoch 252/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9027 - accuracy: 0.4907 - val_loss: 2.2326 - val_accuracy: 0.4115\n",
      "Epoch 253/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9024 - accuracy: 0.4908 - val_loss: 2.2331 - val_accuracy: 0.4123\n",
      "Epoch 254/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9019 - accuracy: 0.4918 - val_loss: 2.2378 - val_accuracy: 0.4085\n",
      "Epoch 255/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9024 - accuracy: 0.4909 - val_loss: 2.2326 - val_accuracy: 0.4089\n",
      "Epoch 256/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9015 - accuracy: 0.4916 - val_loss: 2.2378 - val_accuracy: 0.4104\n",
      "Epoch 257/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9022 - accuracy: 0.4910 - val_loss: 2.2357 - val_accuracy: 0.4123\n",
      "Epoch 258/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.9007 - accuracy: 0.4913 - val_loss: 2.2375 - val_accuracy: 0.4113\n",
      "Epoch 259/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8991 - accuracy: 0.4925 - val_loss: 2.2331 - val_accuracy: 0.4098\n",
      "Epoch 260/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9018 - accuracy: 0.4916 - val_loss: 2.2347 - val_accuracy: 0.4124\n",
      "Epoch 261/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9009 - accuracy: 0.4928 - val_loss: 2.2401 - val_accuracy: 0.4137\n",
      "Epoch 262/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9006 - accuracy: 0.4920 - val_loss: 2.2326 - val_accuracy: 0.4107\n",
      "Epoch 263/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8989 - accuracy: 0.4928 - val_loss: 2.2393 - val_accuracy: 0.4128\n",
      "Epoch 264/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.9009 - accuracy: 0.4929 - val_loss: 2.2317 - val_accuracy: 0.4087\n",
      "Epoch 265/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8988 - accuracy: 0.4928 - val_loss: 2.2445 - val_accuracy: 0.4123\n",
      "Epoch 266/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8993 - accuracy: 0.4930 - val_loss: 2.2351 - val_accuracy: 0.4138\n",
      "Epoch 267/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.9023 - accuracy: 0.4910 - val_loss: 2.2351 - val_accuracy: 0.4114\n",
      "Epoch 268/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8978 - accuracy: 0.4933 - val_loss: 2.2339 - val_accuracy: 0.4112\n",
      "Epoch 269/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8969 - accuracy: 0.4936 - val_loss: 2.2333 - val_accuracy: 0.4119\n",
      "Epoch 270/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8965 - accuracy: 0.4932 - val_loss: 2.2335 - val_accuracy: 0.4106\n",
      "Epoch 271/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8970 - accuracy: 0.4934 - val_loss: 2.2368 - val_accuracy: 0.4111\n",
      "Epoch 272/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8956 - accuracy: 0.4943 - val_loss: 2.2443 - val_accuracy: 0.4075\n",
      "Epoch 273/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8986 - accuracy: 0.4926 - val_loss: 2.2286 - val_accuracy: 0.4098\n",
      "Epoch 274/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8991 - accuracy: 0.4924 - val_loss: 2.2346 - val_accuracy: 0.4138\n",
      "Epoch 275/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8969 - accuracy: 0.4937 - val_loss: 2.2352 - val_accuracy: 0.4153\n",
      "Epoch 276/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8956 - accuracy: 0.4937 - val_loss: 2.2328 - val_accuracy: 0.4111\n",
      "Epoch 277/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8949 - accuracy: 0.4942 - val_loss: 2.2307 - val_accuracy: 0.4118\n",
      "Epoch 278/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8970 - accuracy: 0.4939 - val_loss: 2.2322 - val_accuracy: 0.4094\n",
      "Epoch 279/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8955 - accuracy: 0.4941 - val_loss: 2.2389 - val_accuracy: 0.4085\n",
      "Epoch 280/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8954 - accuracy: 0.4929 - val_loss: 2.2391 - val_accuracy: 0.4092\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 281/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8980 - accuracy: 0.4942 - val_loss: 2.2357 - val_accuracy: 0.4111\n",
      "Epoch 282/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8967 - accuracy: 0.4953 - val_loss: 2.2402 - val_accuracy: 0.4114\n",
      "Epoch 283/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8949 - accuracy: 0.4933 - val_loss: 2.2387 - val_accuracy: 0.4088\n",
      "Epoch 284/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8968 - accuracy: 0.4944 - val_loss: 2.2342 - val_accuracy: 0.4092\n",
      "Epoch 285/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8930 - accuracy: 0.4951 - val_loss: 2.2390 - val_accuracy: 0.4136\n",
      "Epoch 286/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8923 - accuracy: 0.4955 - val_loss: 2.2349 - val_accuracy: 0.4093\n",
      "Epoch 287/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8903 - accuracy: 0.4949 - val_loss: 2.2449 - val_accuracy: 0.4101\n",
      "Epoch 288/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8930 - accuracy: 0.4946 - val_loss: 2.2481 - val_accuracy: 0.4101\n",
      "Epoch 289/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8926 - accuracy: 0.4952 - val_loss: 2.2343 - val_accuracy: 0.4130\n",
      "Epoch 290/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8922 - accuracy: 0.4956 - val_loss: 2.2368 - val_accuracy: 0.4132\n",
      "Epoch 291/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8926 - accuracy: 0.4951 - val_loss: 2.2454 - val_accuracy: 0.4103\n",
      "Epoch 292/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8916 - accuracy: 0.4966 - val_loss: 2.2408 - val_accuracy: 0.4104\n",
      "Epoch 293/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8929 - accuracy: 0.4964 - val_loss: 2.2396 - val_accuracy: 0.4099\n",
      "Epoch 294/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8929 - accuracy: 0.4952 - val_loss: 2.2361 - val_accuracy: 0.4125\n",
      "Epoch 295/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8927 - accuracy: 0.4954 - val_loss: 2.2397 - val_accuracy: 0.4118\n",
      "Epoch 296/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8921 - accuracy: 0.4959 - val_loss: 2.2478 - val_accuracy: 0.4082\n",
      "Epoch 297/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8913 - accuracy: 0.4966 - val_loss: 2.2353 - val_accuracy: 0.4125\n",
      "Epoch 298/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8890 - accuracy: 0.4974 - val_loss: 2.2468 - val_accuracy: 0.4072\n",
      "Epoch 299/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8883 - accuracy: 0.4974 - val_loss: 2.2454 - val_accuracy: 0.4120\n",
      "Epoch 300/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8922 - accuracy: 0.4957 - val_loss: 2.2418 - val_accuracy: 0.4143\n",
      "Epoch 301/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8928 - accuracy: 0.4954 - val_loss: 2.2427 - val_accuracy: 0.4107\n",
      "Epoch 302/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8891 - accuracy: 0.4972 - val_loss: 2.2428 - val_accuracy: 0.4098\n",
      "Epoch 303/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8917 - accuracy: 0.4950 - val_loss: 2.2396 - val_accuracy: 0.4137\n",
      "Epoch 304/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8928 - accuracy: 0.4951 - val_loss: 2.2320 - val_accuracy: 0.4165\n",
      "Epoch 305/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8876 - accuracy: 0.4969 - val_loss: 2.2377 - val_accuracy: 0.4113\n",
      "Epoch 306/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8911 - accuracy: 0.4966 - val_loss: 2.2307 - val_accuracy: 0.4136\n",
      "Epoch 307/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8905 - accuracy: 0.4952 - val_loss: 2.2392 - val_accuracy: 0.4147\n",
      "Epoch 308/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8900 - accuracy: 0.4979 - val_loss: 2.2414 - val_accuracy: 0.4133\n",
      "Epoch 309/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8878 - accuracy: 0.4970 - val_loss: 2.2400 - val_accuracy: 0.4148\n",
      "Epoch 310/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8889 - accuracy: 0.4965 - val_loss: 2.2402 - val_accuracy: 0.4120\n",
      "Epoch 311/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8868 - accuracy: 0.4982 - val_loss: 2.2430 - val_accuracy: 0.4142\n",
      "Epoch 312/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8905 - accuracy: 0.4967 - val_loss: 2.2357 - val_accuracy: 0.4130\n",
      "Epoch 313/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8855 - accuracy: 0.4981 - val_loss: 2.2361 - val_accuracy: 0.4152\n",
      "Epoch 314/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8869 - accuracy: 0.4983 - val_loss: 2.2364 - val_accuracy: 0.4144\n",
      "Epoch 315/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8867 - accuracy: 0.4985 - val_loss: 2.2380 - val_accuracy: 0.4098\n",
      "Epoch 316/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8853 - accuracy: 0.4972 - val_loss: 2.2384 - val_accuracy: 0.4129\n",
      "Epoch 317/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8885 - accuracy: 0.4975 - val_loss: 2.2382 - val_accuracy: 0.4134\n",
      "Epoch 318/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8849 - accuracy: 0.4992 - val_loss: 2.2395 - val_accuracy: 0.4107\n",
      "Epoch 319/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8868 - accuracy: 0.4980 - val_loss: 2.2409 - val_accuracy: 0.4138\n",
      "Epoch 320/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8857 - accuracy: 0.4987 - val_loss: 2.2440 - val_accuracy: 0.4150\n",
      "Epoch 321/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8864 - accuracy: 0.4981 - val_loss: 2.2437 - val_accuracy: 0.4130\n",
      "Epoch 322/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8865 - accuracy: 0.4984 - val_loss: 2.2398 - val_accuracy: 0.4155\n",
      "Epoch 323/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8856 - accuracy: 0.4984 - val_loss: 2.2479 - val_accuracy: 0.4127\n",
      "Epoch 324/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8835 - accuracy: 0.4982 - val_loss: 2.2413 - val_accuracy: 0.4158\n",
      "Epoch 325/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8847 - accuracy: 0.4981 - val_loss: 2.2436 - val_accuracy: 0.4125\n",
      "Epoch 326/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8838 - accuracy: 0.4994 - val_loss: 2.2435 - val_accuracy: 0.4149\n",
      "Epoch 327/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8880 - accuracy: 0.4976 - val_loss: 2.2383 - val_accuracy: 0.4149\n",
      "Epoch 328/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8824 - accuracy: 0.4989 - val_loss: 2.2472 - val_accuracy: 0.4127\n",
      "Epoch 329/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8870 - accuracy: 0.4988 - val_loss: 2.2457 - val_accuracy: 0.4094\n",
      "Epoch 330/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8851 - accuracy: 0.4984 - val_loss: 2.2437 - val_accuracy: 0.4133\n",
      "Epoch 331/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8815 - accuracy: 0.4998 - val_loss: 2.2425 - val_accuracy: 0.4122\n",
      "Epoch 332/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8835 - accuracy: 0.4983 - val_loss: 2.2382 - val_accuracy: 0.4128\n",
      "Epoch 333/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8835 - accuracy: 0.4991 - val_loss: 2.2448 - val_accuracy: 0.4092\n",
      "Epoch 334/2000\n",
      "242/242 [==============================] - 6s 27ms/step - loss: 1.8812 - accuracy: 0.5007 - val_loss: 2.2385 - val_accuracy: 0.4155\n",
      "Epoch 335/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8842 - accuracy: 0.4991 - val_loss: 2.2345 - val_accuracy: 0.4160\n",
      "Epoch 336/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8835 - accuracy: 0.4992 - val_loss: 2.2430 - val_accuracy: 0.4127\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 337/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8804 - accuracy: 0.5003 - val_loss: 2.2506 - val_accuracy: 0.4101\n",
      "Epoch 338/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8820 - accuracy: 0.4999 - val_loss: 2.2482 - val_accuracy: 0.4093\n",
      "Epoch 339/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8796 - accuracy: 0.4999 - val_loss: 2.2405 - val_accuracy: 0.4151\n",
      "Epoch 340/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8844 - accuracy: 0.4993 - val_loss: 2.2491 - val_accuracy: 0.4088\n",
      "Epoch 341/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8802 - accuracy: 0.5002 - val_loss: 2.2534 - val_accuracy: 0.4118\n",
      "Epoch 342/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8840 - accuracy: 0.4988 - val_loss: 2.2448 - val_accuracy: 0.4134\n",
      "Epoch 343/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8820 - accuracy: 0.4997 - val_loss: 2.2432 - val_accuracy: 0.4135\n",
      "Epoch 344/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8817 - accuracy: 0.4994 - val_loss: 2.2516 - val_accuracy: 0.4092\n",
      "Epoch 345/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8778 - accuracy: 0.5009 - val_loss: 2.2439 - val_accuracy: 0.4104\n",
      "Epoch 346/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8839 - accuracy: 0.4988 - val_loss: 2.2407 - val_accuracy: 0.4182\n",
      "Epoch 347/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8825 - accuracy: 0.5006 - val_loss: 2.2383 - val_accuracy: 0.4120\n",
      "Epoch 348/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8798 - accuracy: 0.5004 - val_loss: 2.2498 - val_accuracy: 0.4115\n",
      "Epoch 349/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8792 - accuracy: 0.5002 - val_loss: 2.2448 - val_accuracy: 0.4117\n",
      "Epoch 350/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8812 - accuracy: 0.4989 - val_loss: 2.2361 - val_accuracy: 0.4125\n",
      "Epoch 351/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8829 - accuracy: 0.5005 - val_loss: 2.2471 - val_accuracy: 0.4132\n",
      "Epoch 352/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8792 - accuracy: 0.5006 - val_loss: 2.2358 - val_accuracy: 0.4173\n",
      "Epoch 353/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8802 - accuracy: 0.4999 - val_loss: 2.2414 - val_accuracy: 0.4164\n",
      "Epoch 354/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8814 - accuracy: 0.5006 - val_loss: 2.2400 - val_accuracy: 0.4133\n",
      "Epoch 355/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8795 - accuracy: 0.5008 - val_loss: 2.2370 - val_accuracy: 0.4131\n",
      "Epoch 356/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8806 - accuracy: 0.5008 - val_loss: 2.2405 - val_accuracy: 0.4177\n",
      "Epoch 357/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8825 - accuracy: 0.5000 - val_loss: 2.2482 - val_accuracy: 0.4120\n",
      "Epoch 358/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8808 - accuracy: 0.5005 - val_loss: 2.2373 - val_accuracy: 0.4139\n",
      "Epoch 359/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8799 - accuracy: 0.5011 - val_loss: 2.2411 - val_accuracy: 0.4121\n",
      "Epoch 360/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8775 - accuracy: 0.5008 - val_loss: 2.2382 - val_accuracy: 0.4165\n",
      "Epoch 361/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8795 - accuracy: 0.5010 - val_loss: 2.2397 - val_accuracy: 0.4139\n",
      "Epoch 362/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8799 - accuracy: 0.5015 - val_loss: 2.2430 - val_accuracy: 0.4135\n",
      "Epoch 363/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8778 - accuracy: 0.5022 - val_loss: 2.2462 - val_accuracy: 0.4142\n",
      "Epoch 364/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8764 - accuracy: 0.5007 - val_loss: 2.2500 - val_accuracy: 0.4183\n",
      "Epoch 365/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8773 - accuracy: 0.5025 - val_loss: 2.2447 - val_accuracy: 0.4123\n",
      "Epoch 366/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8766 - accuracy: 0.5015 - val_loss: 2.2474 - val_accuracy: 0.4122\n",
      "Epoch 367/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8805 - accuracy: 0.5004 - val_loss: 2.2418 - val_accuracy: 0.4136\n",
      "Epoch 368/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8735 - accuracy: 0.5015 - val_loss: 2.2464 - val_accuracy: 0.4130\n",
      "Epoch 369/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8779 - accuracy: 0.5016 - val_loss: 2.2452 - val_accuracy: 0.4113\n",
      "Epoch 370/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8756 - accuracy: 0.5029 - val_loss: 2.2388 - val_accuracy: 0.4146\n",
      "Epoch 371/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8779 - accuracy: 0.5015 - val_loss: 2.2460 - val_accuracy: 0.4142\n",
      "Epoch 372/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8767 - accuracy: 0.5015 - val_loss: 2.2323 - val_accuracy: 0.4169\n",
      "Epoch 373/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8780 - accuracy: 0.5017 - val_loss: 2.2443 - val_accuracy: 0.4103\n",
      "Epoch 374/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8789 - accuracy: 0.5007 - val_loss: 2.2399 - val_accuracy: 0.4172\n",
      "Epoch 375/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8737 - accuracy: 0.5031 - val_loss: 2.2377 - val_accuracy: 0.4123\n",
      "Epoch 376/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8743 - accuracy: 0.5027 - val_loss: 2.2476 - val_accuracy: 0.4109\n",
      "Epoch 377/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8751 - accuracy: 0.5025 - val_loss: 2.2456 - val_accuracy: 0.4139\n",
      "Epoch 378/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8763 - accuracy: 0.5022 - val_loss: 2.2451 - val_accuracy: 0.4169\n",
      "Epoch 379/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8743 - accuracy: 0.5034 - val_loss: 2.2407 - val_accuracy: 0.4157\n",
      "Epoch 380/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8753 - accuracy: 0.5028 - val_loss: 2.2461 - val_accuracy: 0.4137\n",
      "Epoch 381/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8755 - accuracy: 0.5017 - val_loss: 2.2394 - val_accuracy: 0.4098\n",
      "Epoch 382/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8744 - accuracy: 0.5024 - val_loss: 2.2394 - val_accuracy: 0.4126\n",
      "Epoch 383/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8740 - accuracy: 0.5031 - val_loss: 2.2386 - val_accuracy: 0.4130\n",
      "Epoch 384/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8757 - accuracy: 0.5017 - val_loss: 2.2430 - val_accuracy: 0.4139\n",
      "Epoch 385/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8736 - accuracy: 0.5024 - val_loss: 2.2426 - val_accuracy: 0.4151\n",
      "Epoch 386/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8737 - accuracy: 0.5031 - val_loss: 2.2466 - val_accuracy: 0.4131\n",
      "Epoch 387/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8731 - accuracy: 0.5027 - val_loss: 2.2425 - val_accuracy: 0.4138\n",
      "Epoch 388/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8744 - accuracy: 0.5022 - val_loss: 2.2379 - val_accuracy: 0.4143\n",
      "Epoch 389/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8744 - accuracy: 0.5027 - val_loss: 2.2547 - val_accuracy: 0.4133\n",
      "Epoch 390/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8718 - accuracy: 0.5039 - val_loss: 2.2393 - val_accuracy: 0.4142\n",
      "Epoch 391/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8747 - accuracy: 0.5031 - val_loss: 2.2441 - val_accuracy: 0.4157\n",
      "Epoch 392/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8733 - accuracy: 0.5029 - val_loss: 2.2439 - val_accuracy: 0.4146\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 393/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8741 - accuracy: 0.5046 - val_loss: 2.2444 - val_accuracy: 0.4104\n",
      "Epoch 394/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8747 - accuracy: 0.5037 - val_loss: 2.2364 - val_accuracy: 0.4153\n",
      "Epoch 395/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8719 - accuracy: 0.5039 - val_loss: 2.2433 - val_accuracy: 0.4136\n",
      "Epoch 396/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8745 - accuracy: 0.5027 - val_loss: 2.2390 - val_accuracy: 0.4145\n",
      "Epoch 397/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8744 - accuracy: 0.5028 - val_loss: 2.2409 - val_accuracy: 0.4120\n",
      "Epoch 398/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8736 - accuracy: 0.5036 - val_loss: 2.2394 - val_accuracy: 0.4162\n",
      "Epoch 399/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8730 - accuracy: 0.5034 - val_loss: 2.2375 - val_accuracy: 0.4137\n",
      "Epoch 400/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8704 - accuracy: 0.5045 - val_loss: 2.2457 - val_accuracy: 0.4114\n",
      "Epoch 401/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8744 - accuracy: 0.5040 - val_loss: 2.2491 - val_accuracy: 0.4135\n",
      "Epoch 402/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8708 - accuracy: 0.5041 - val_loss: 2.2486 - val_accuracy: 0.4117\n",
      "Epoch 403/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8735 - accuracy: 0.5034 - val_loss: 2.2456 - val_accuracy: 0.4139\n",
      "Epoch 404/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8678 - accuracy: 0.5038 - val_loss: 2.2410 - val_accuracy: 0.4149\n",
      "Epoch 405/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8709 - accuracy: 0.5048 - val_loss: 2.2475 - val_accuracy: 0.4151\n",
      "Epoch 406/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8704 - accuracy: 0.5036 - val_loss: 2.2459 - val_accuracy: 0.4152\n",
      "Epoch 407/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8710 - accuracy: 0.5053 - val_loss: 2.2493 - val_accuracy: 0.4130\n",
      "Epoch 408/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8715 - accuracy: 0.5036 - val_loss: 2.2428 - val_accuracy: 0.4167\n",
      "Epoch 409/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8697 - accuracy: 0.5051 - val_loss: 2.2398 - val_accuracy: 0.4151\n",
      "Epoch 410/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8721 - accuracy: 0.5027 - val_loss: 2.2492 - val_accuracy: 0.4150\n",
      "Epoch 411/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8708 - accuracy: 0.5047 - val_loss: 2.2452 - val_accuracy: 0.4172\n",
      "Epoch 412/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8699 - accuracy: 0.5044 - val_loss: 2.2448 - val_accuracy: 0.4149\n",
      "Epoch 413/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8700 - accuracy: 0.5044 - val_loss: 2.2469 - val_accuracy: 0.4166\n",
      "Epoch 414/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8713 - accuracy: 0.5035 - val_loss: 2.2382 - val_accuracy: 0.4174\n",
      "Epoch 415/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8691 - accuracy: 0.5040 - val_loss: 2.2460 - val_accuracy: 0.4160\n",
      "Epoch 416/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8713 - accuracy: 0.5043 - val_loss: 2.2403 - val_accuracy: 0.4171\n",
      "Epoch 417/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8677 - accuracy: 0.5058 - val_loss: 2.2491 - val_accuracy: 0.4146\n",
      "Epoch 418/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8709 - accuracy: 0.5056 - val_loss: 2.2424 - val_accuracy: 0.4160\n",
      "Epoch 419/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8687 - accuracy: 0.5052 - val_loss: 2.2483 - val_accuracy: 0.4174\n",
      "Epoch 420/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8677 - accuracy: 0.5046 - val_loss: 2.2480 - val_accuracy: 0.4166\n",
      "Epoch 421/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8673 - accuracy: 0.5056 - val_loss: 2.2513 - val_accuracy: 0.4157\n",
      "Epoch 422/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8703 - accuracy: 0.5034 - val_loss: 2.2539 - val_accuracy: 0.4131\n",
      "Epoch 423/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8707 - accuracy: 0.5045 - val_loss: 2.2454 - val_accuracy: 0.4146\n",
      "Epoch 424/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8715 - accuracy: 0.5043 - val_loss: 2.2455 - val_accuracy: 0.4144\n",
      "Epoch 425/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8655 - accuracy: 0.5055 - val_loss: 2.2414 - val_accuracy: 0.4146\n",
      "Epoch 426/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8699 - accuracy: 0.5045 - val_loss: 2.2443 - val_accuracy: 0.4154\n",
      "Epoch 427/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8696 - accuracy: 0.5050 - val_loss: 2.2405 - val_accuracy: 0.4162\n",
      "Epoch 428/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8659 - accuracy: 0.5052 - val_loss: 2.2490 - val_accuracy: 0.4109\n",
      "Epoch 429/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8662 - accuracy: 0.5066 - val_loss: 2.2409 - val_accuracy: 0.4155\n",
      "Epoch 430/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8674 - accuracy: 0.5065 - val_loss: 2.2521 - val_accuracy: 0.4124\n",
      "Epoch 431/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8698 - accuracy: 0.5046 - val_loss: 2.2441 - val_accuracy: 0.4175\n",
      "Epoch 432/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8673 - accuracy: 0.5051 - val_loss: 2.2528 - val_accuracy: 0.4147\n",
      "Epoch 433/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8667 - accuracy: 0.5072 - val_loss: 2.2439 - val_accuracy: 0.4101\n",
      "Epoch 434/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8689 - accuracy: 0.5050 - val_loss: 2.2445 - val_accuracy: 0.4161\n",
      "Epoch 435/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8670 - accuracy: 0.5067 - val_loss: 2.2465 - val_accuracy: 0.4126\n",
      "Epoch 436/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8659 - accuracy: 0.5059 - val_loss: 2.2494 - val_accuracy: 0.4123\n",
      "Epoch 437/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8691 - accuracy: 0.5050 - val_loss: 2.2472 - val_accuracy: 0.4164\n",
      "Epoch 438/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8669 - accuracy: 0.5053 - val_loss: 2.2439 - val_accuracy: 0.4171\n",
      "Epoch 439/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8660 - accuracy: 0.5069 - val_loss: 2.2496 - val_accuracy: 0.4153\n",
      "Epoch 440/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8680 - accuracy: 0.5050 - val_loss: 2.2449 - val_accuracy: 0.4167\n",
      "Epoch 441/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8672 - accuracy: 0.5064 - val_loss: 2.2484 - val_accuracy: 0.4135\n",
      "Epoch 442/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8667 - accuracy: 0.5069 - val_loss: 2.2527 - val_accuracy: 0.4121\n",
      "Epoch 443/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8647 - accuracy: 0.5068 - val_loss: 2.2451 - val_accuracy: 0.4144\n",
      "Epoch 444/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8682 - accuracy: 0.5057 - val_loss: 2.2495 - val_accuracy: 0.4135\n",
      "Epoch 445/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8658 - accuracy: 0.5061 - val_loss: 2.2476 - val_accuracy: 0.4158\n",
      "Epoch 446/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8623 - accuracy: 0.5067 - val_loss: 2.2554 - val_accuracy: 0.4123\n",
      "Epoch 447/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8646 - accuracy: 0.5066 - val_loss: 2.2467 - val_accuracy: 0.4172\n",
      "Epoch 448/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8639 - accuracy: 0.5066 - val_loss: 2.2531 - val_accuracy: 0.4120\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 449/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8677 - accuracy: 0.5053 - val_loss: 2.2545 - val_accuracy: 0.4146\n",
      "Epoch 450/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8615 - accuracy: 0.5072 - val_loss: 2.2542 - val_accuracy: 0.4107\n",
      "Epoch 451/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8670 - accuracy: 0.5068 - val_loss: 2.2512 - val_accuracy: 0.4150\n",
      "Epoch 452/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8640 - accuracy: 0.5071 - val_loss: 2.2456 - val_accuracy: 0.4149\n",
      "Epoch 453/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8651 - accuracy: 0.5064 - val_loss: 2.2465 - val_accuracy: 0.4167\n",
      "Epoch 454/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8684 - accuracy: 0.5053 - val_loss: 2.2437 - val_accuracy: 0.4155\n",
      "Epoch 455/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8646 - accuracy: 0.5062 - val_loss: 2.2443 - val_accuracy: 0.4168\n",
      "Epoch 456/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8657 - accuracy: 0.5065 - val_loss: 2.2507 - val_accuracy: 0.4111\n",
      "Epoch 457/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8661 - accuracy: 0.5058 - val_loss: 2.2500 - val_accuracy: 0.4163\n",
      "Epoch 458/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8630 - accuracy: 0.5082 - val_loss: 2.2489 - val_accuracy: 0.4147\n",
      "Epoch 459/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8631 - accuracy: 0.5070 - val_loss: 2.2513 - val_accuracy: 0.4121\n",
      "Epoch 460/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8639 - accuracy: 0.5074 - val_loss: 2.2480 - val_accuracy: 0.4153\n",
      "Epoch 461/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8637 - accuracy: 0.5085 - val_loss: 2.2521 - val_accuracy: 0.4123\n",
      "Epoch 462/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8602 - accuracy: 0.5082 - val_loss: 2.2542 - val_accuracy: 0.4154\n",
      "Epoch 463/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8653 - accuracy: 0.5063 - val_loss: 2.2509 - val_accuracy: 0.4158\n",
      "Epoch 464/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8639 - accuracy: 0.5076 - val_loss: 2.2481 - val_accuracy: 0.4122\n",
      "Epoch 465/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8619 - accuracy: 0.5070 - val_loss: 2.2530 - val_accuracy: 0.4136\n",
      "Epoch 466/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8626 - accuracy: 0.5080 - val_loss: 2.2507 - val_accuracy: 0.4113\n",
      "Epoch 467/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8625 - accuracy: 0.5075 - val_loss: 2.2460 - val_accuracy: 0.4140\n",
      "Epoch 468/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8615 - accuracy: 0.5069 - val_loss: 2.2495 - val_accuracy: 0.4126\n",
      "Epoch 469/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8618 - accuracy: 0.5079 - val_loss: 2.2457 - val_accuracy: 0.4121\n",
      "Epoch 470/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8631 - accuracy: 0.5076 - val_loss: 2.2561 - val_accuracy: 0.4113\n",
      "Epoch 471/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8638 - accuracy: 0.5076 - val_loss: 2.2461 - val_accuracy: 0.4121\n",
      "Epoch 472/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8592 - accuracy: 0.5084 - val_loss: 2.2561 - val_accuracy: 0.4115\n",
      "Epoch 473/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8623 - accuracy: 0.5060 - val_loss: 2.2526 - val_accuracy: 0.4122\n",
      "Epoch 474/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8604 - accuracy: 0.5088 - val_loss: 2.2555 - val_accuracy: 0.4085\n",
      "Epoch 475/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8620 - accuracy: 0.5085 - val_loss: 2.2500 - val_accuracy: 0.4124\n",
      "Epoch 476/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8613 - accuracy: 0.5076 - val_loss: 2.2514 - val_accuracy: 0.4144\n",
      "Epoch 477/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8597 - accuracy: 0.5076 - val_loss: 2.2557 - val_accuracy: 0.4127\n",
      "Epoch 478/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8592 - accuracy: 0.5094 - val_loss: 2.2465 - val_accuracy: 0.4187\n",
      "Epoch 479/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8644 - accuracy: 0.5078 - val_loss: 2.2578 - val_accuracy: 0.4122\n",
      "Epoch 480/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8645 - accuracy: 0.5083 - val_loss: 2.2521 - val_accuracy: 0.4114\n",
      "Epoch 481/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8622 - accuracy: 0.5075 - val_loss: 2.2512 - val_accuracy: 0.4104\n",
      "Epoch 482/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8590 - accuracy: 0.5092 - val_loss: 2.2555 - val_accuracy: 0.4130\n",
      "Epoch 483/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8589 - accuracy: 0.5084 - val_loss: 2.2438 - val_accuracy: 0.4135\n",
      "Epoch 484/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8598 - accuracy: 0.5085 - val_loss: 2.2494 - val_accuracy: 0.4120\n",
      "Epoch 485/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8627 - accuracy: 0.5077 - val_loss: 2.2510 - val_accuracy: 0.4117\n",
      "Epoch 486/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8610 - accuracy: 0.5090 - val_loss: 2.2474 - val_accuracy: 0.4138\n",
      "Epoch 487/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8609 - accuracy: 0.5084 - val_loss: 2.2504 - val_accuracy: 0.4118\n",
      "Epoch 488/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8595 - accuracy: 0.5077 - val_loss: 2.2521 - val_accuracy: 0.4105\n",
      "Epoch 489/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8606 - accuracy: 0.5080 - val_loss: 2.2550 - val_accuracy: 0.4120\n",
      "Epoch 490/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8597 - accuracy: 0.5085 - val_loss: 2.2506 - val_accuracy: 0.4121\n",
      "Epoch 491/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8600 - accuracy: 0.5092 - val_loss: 2.2527 - val_accuracy: 0.4092\n",
      "Epoch 492/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8609 - accuracy: 0.5077 - val_loss: 2.2486 - val_accuracy: 0.4142\n",
      "Epoch 493/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8615 - accuracy: 0.5078 - val_loss: 2.2575 - val_accuracy: 0.4110\n",
      "Epoch 494/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8591 - accuracy: 0.5080 - val_loss: 2.2499 - val_accuracy: 0.4146\n",
      "Epoch 495/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8572 - accuracy: 0.5095 - val_loss: 2.2447 - val_accuracy: 0.4165\n",
      "Epoch 496/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8562 - accuracy: 0.5094 - val_loss: 2.2561 - val_accuracy: 0.4139\n",
      "Epoch 497/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8575 - accuracy: 0.5096 - val_loss: 2.2490 - val_accuracy: 0.4130\n",
      "Epoch 498/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8620 - accuracy: 0.5074 - val_loss: 2.2488 - val_accuracy: 0.4114\n",
      "Epoch 499/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8577 - accuracy: 0.5102 - val_loss: 2.2449 - val_accuracy: 0.4135\n",
      "Epoch 500/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8566 - accuracy: 0.5099 - val_loss: 2.2460 - val_accuracy: 0.4180\n",
      "Epoch 501/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8588 - accuracy: 0.5091 - val_loss: 2.2530 - val_accuracy: 0.4139\n",
      "Epoch 502/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8569 - accuracy: 0.5095 - val_loss: 2.2541 - val_accuracy: 0.4137\n",
      "Epoch 503/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8590 - accuracy: 0.5086 - val_loss: 2.2544 - val_accuracy: 0.4127\n",
      "Epoch 504/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8566 - accuracy: 0.5103 - val_loss: 2.2493 - val_accuracy: 0.4129\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 505/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8548 - accuracy: 0.5109 - val_loss: 2.2556 - val_accuracy: 0.4154\n",
      "Epoch 506/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8571 - accuracy: 0.5092 - val_loss: 2.2546 - val_accuracy: 0.4167\n",
      "Epoch 507/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8557 - accuracy: 0.5098 - val_loss: 2.2552 - val_accuracy: 0.4158\n",
      "Epoch 508/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8585 - accuracy: 0.5085 - val_loss: 2.2596 - val_accuracy: 0.4130\n",
      "Epoch 509/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8549 - accuracy: 0.5103 - val_loss: 2.2554 - val_accuracy: 0.4127\n",
      "Epoch 510/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8579 - accuracy: 0.5093 - val_loss: 2.2523 - val_accuracy: 0.4130\n",
      "Epoch 511/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8567 - accuracy: 0.5102 - val_loss: 2.2515 - val_accuracy: 0.4123\n",
      "Epoch 512/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8584 - accuracy: 0.5099 - val_loss: 2.2523 - val_accuracy: 0.4117\n",
      "Epoch 513/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8576 - accuracy: 0.5089 - val_loss: 2.2587 - val_accuracy: 0.4139\n",
      "Epoch 514/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8560 - accuracy: 0.5096 - val_loss: 2.2539 - val_accuracy: 0.4152\n",
      "Epoch 515/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8576 - accuracy: 0.5097 - val_loss: 2.2566 - val_accuracy: 0.4127\n",
      "Epoch 516/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8559 - accuracy: 0.5104 - val_loss: 2.2526 - val_accuracy: 0.4165\n",
      "Epoch 517/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8570 - accuracy: 0.5097 - val_loss: 2.2498 - val_accuracy: 0.4138\n",
      "Epoch 518/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8538 - accuracy: 0.5111 - val_loss: 2.2538 - val_accuracy: 0.4134\n",
      "Epoch 519/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8561 - accuracy: 0.5105 - val_loss: 2.2479 - val_accuracy: 0.4156\n",
      "Epoch 520/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8581 - accuracy: 0.5095 - val_loss: 2.2534 - val_accuracy: 0.4152\n",
      "Epoch 521/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8568 - accuracy: 0.5103 - val_loss: 2.2554 - val_accuracy: 0.4135\n",
      "Epoch 522/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8563 - accuracy: 0.5100 - val_loss: 2.2537 - val_accuracy: 0.4130\n",
      "Epoch 523/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8549 - accuracy: 0.5102 - val_loss: 2.2611 - val_accuracy: 0.4105\n",
      "Epoch 524/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8542 - accuracy: 0.5104 - val_loss: 2.2508 - val_accuracy: 0.4168\n",
      "Epoch 525/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8574 - accuracy: 0.5090 - val_loss: 2.2534 - val_accuracy: 0.4146\n",
      "Epoch 526/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8554 - accuracy: 0.5117 - val_loss: 2.2487 - val_accuracy: 0.4144\n",
      "Epoch 527/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8539 - accuracy: 0.5110 - val_loss: 2.2507 - val_accuracy: 0.4155\n",
      "Epoch 528/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8543 - accuracy: 0.5102 - val_loss: 2.2449 - val_accuracy: 0.4190\n",
      "Epoch 529/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8527 - accuracy: 0.5105 - val_loss: 2.2605 - val_accuracy: 0.4120\n",
      "Epoch 530/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8552 - accuracy: 0.5106 - val_loss: 2.2534 - val_accuracy: 0.4156\n",
      "Epoch 531/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8543 - accuracy: 0.5098 - val_loss: 2.2556 - val_accuracy: 0.4146\n",
      "Epoch 532/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8528 - accuracy: 0.5106 - val_loss: 2.2534 - val_accuracy: 0.4154\n",
      "Epoch 533/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8555 - accuracy: 0.5096 - val_loss: 2.2494 - val_accuracy: 0.4166\n",
      "Epoch 534/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8536 - accuracy: 0.5106 - val_loss: 2.2517 - val_accuracy: 0.4173\n",
      "Epoch 535/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8562 - accuracy: 0.5101 - val_loss: 2.2522 - val_accuracy: 0.4139\n",
      "Epoch 536/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8565 - accuracy: 0.5093 - val_loss: 2.2597 - val_accuracy: 0.4147\n",
      "Epoch 537/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8542 - accuracy: 0.5102 - val_loss: 2.2573 - val_accuracy: 0.4134\n",
      "Epoch 538/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8510 - accuracy: 0.5109 - val_loss: 2.2534 - val_accuracy: 0.4140\n",
      "Epoch 539/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8541 - accuracy: 0.5110 - val_loss: 2.2608 - val_accuracy: 0.4119\n",
      "Epoch 540/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8564 - accuracy: 0.5091 - val_loss: 2.2511 - val_accuracy: 0.4184\n",
      "Epoch 541/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8530 - accuracy: 0.5113 - val_loss: 2.2615 - val_accuracy: 0.4079\n",
      "Epoch 542/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8571 - accuracy: 0.5093 - val_loss: 2.2549 - val_accuracy: 0.4151\n",
      "Epoch 543/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8548 - accuracy: 0.5105 - val_loss: 2.2531 - val_accuracy: 0.4112\n",
      "Epoch 544/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8547 - accuracy: 0.5106 - val_loss: 2.2526 - val_accuracy: 0.4149\n",
      "Epoch 545/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8552 - accuracy: 0.5102 - val_loss: 2.2517 - val_accuracy: 0.4117\n",
      "Epoch 546/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8494 - accuracy: 0.5120 - val_loss: 2.2602 - val_accuracy: 0.4121\n",
      "Epoch 547/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8524 - accuracy: 0.5113 - val_loss: 2.2529 - val_accuracy: 0.4137\n",
      "Epoch 548/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8561 - accuracy: 0.5102 - val_loss: 2.2538 - val_accuracy: 0.4136\n",
      "Epoch 549/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8539 - accuracy: 0.5103 - val_loss: 2.2569 - val_accuracy: 0.4138\n",
      "Epoch 550/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8513 - accuracy: 0.5111 - val_loss: 2.2535 - val_accuracy: 0.4117\n",
      "Epoch 551/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8539 - accuracy: 0.5108 - val_loss: 2.2563 - val_accuracy: 0.4153\n",
      "Epoch 552/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8512 - accuracy: 0.5112 - val_loss: 2.2500 - val_accuracy: 0.4158\n",
      "Epoch 553/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8547 - accuracy: 0.5118 - val_loss: 2.2598 - val_accuracy: 0.4132\n",
      "Epoch 554/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8549 - accuracy: 0.5111 - val_loss: 2.2485 - val_accuracy: 0.4139\n",
      "Epoch 555/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8534 - accuracy: 0.5103 - val_loss: 2.2525 - val_accuracy: 0.4152\n",
      "Epoch 556/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8533 - accuracy: 0.5118 - val_loss: 2.2563 - val_accuracy: 0.4122\n",
      "Epoch 557/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8541 - accuracy: 0.5121 - val_loss: 2.2562 - val_accuracy: 0.4113\n",
      "Epoch 558/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8517 - accuracy: 0.5118 - val_loss: 2.2478 - val_accuracy: 0.4122\n",
      "Epoch 559/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8524 - accuracy: 0.5117 - val_loss: 2.2524 - val_accuracy: 0.4116\n",
      "Epoch 560/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8527 - accuracy: 0.5113 - val_loss: 2.2562 - val_accuracy: 0.4128\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 561/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8499 - accuracy: 0.5131 - val_loss: 2.2622 - val_accuracy: 0.4079\n",
      "Epoch 562/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8516 - accuracy: 0.5116 - val_loss: 2.2588 - val_accuracy: 0.4127\n",
      "Epoch 563/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8502 - accuracy: 0.5121 - val_loss: 2.2495 - val_accuracy: 0.4145\n",
      "Epoch 564/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8528 - accuracy: 0.5119 - val_loss: 2.2614 - val_accuracy: 0.4113\n",
      "Epoch 565/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8547 - accuracy: 0.5118 - val_loss: 2.2548 - val_accuracy: 0.4127\n",
      "Epoch 566/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8517 - accuracy: 0.5115 - val_loss: 2.2571 - val_accuracy: 0.4151\n",
      "Epoch 567/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8517 - accuracy: 0.5120 - val_loss: 2.2546 - val_accuracy: 0.4156\n",
      "Epoch 568/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8497 - accuracy: 0.5125 - val_loss: 2.2577 - val_accuracy: 0.4165\n",
      "Epoch 569/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8503 - accuracy: 0.5118 - val_loss: 2.2549 - val_accuracy: 0.4165\n",
      "Epoch 570/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8510 - accuracy: 0.5117 - val_loss: 2.2506 - val_accuracy: 0.4155\n",
      "Epoch 571/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8533 - accuracy: 0.5113 - val_loss: 2.2532 - val_accuracy: 0.4168\n",
      "Epoch 572/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8506 - accuracy: 0.5113 - val_loss: 2.2591 - val_accuracy: 0.4127\n",
      "Epoch 573/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8510 - accuracy: 0.5115 - val_loss: 2.2476 - val_accuracy: 0.4133\n",
      "Epoch 574/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8494 - accuracy: 0.5131 - val_loss: 2.2475 - val_accuracy: 0.4149\n",
      "Epoch 575/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8505 - accuracy: 0.5123 - val_loss: 2.2432 - val_accuracy: 0.4136\n",
      "Epoch 576/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8530 - accuracy: 0.5108 - val_loss: 2.2530 - val_accuracy: 0.4134\n",
      "Epoch 577/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8486 - accuracy: 0.5125 - val_loss: 2.2536 - val_accuracy: 0.4133\n",
      "Epoch 578/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8489 - accuracy: 0.5130 - val_loss: 2.2475 - val_accuracy: 0.4155\n",
      "Epoch 579/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8505 - accuracy: 0.5130 - val_loss: 2.2542 - val_accuracy: 0.4170\n",
      "Epoch 580/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8482 - accuracy: 0.5130 - val_loss: 2.2529 - val_accuracy: 0.4134\n",
      "Epoch 581/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8465 - accuracy: 0.5133 - val_loss: 2.2586 - val_accuracy: 0.4124\n",
      "Epoch 582/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8478 - accuracy: 0.5134 - val_loss: 2.2599 - val_accuracy: 0.4171\n",
      "Epoch 583/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8483 - accuracy: 0.5136 - val_loss: 2.2510 - val_accuracy: 0.4177\n",
      "Epoch 584/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8493 - accuracy: 0.5129 - val_loss: 2.2585 - val_accuracy: 0.4149\n",
      "Epoch 585/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8523 - accuracy: 0.5119 - val_loss: 2.2517 - val_accuracy: 0.4182\n",
      "Epoch 586/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8482 - accuracy: 0.5129 - val_loss: 2.2518 - val_accuracy: 0.4154\n",
      "Epoch 587/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8530 - accuracy: 0.5114 - val_loss: 2.2554 - val_accuracy: 0.4120\n",
      "Epoch 588/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8477 - accuracy: 0.5144 - val_loss: 2.2489 - val_accuracy: 0.4149\n",
      "Epoch 589/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8483 - accuracy: 0.5143 - val_loss: 2.2557 - val_accuracy: 0.4133\n",
      "Epoch 590/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8495 - accuracy: 0.5133 - val_loss: 2.2564 - val_accuracy: 0.4135\n",
      "Epoch 591/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8489 - accuracy: 0.5126 - val_loss: 2.2565 - val_accuracy: 0.4123\n",
      "Epoch 592/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8489 - accuracy: 0.5122 - val_loss: 2.2533 - val_accuracy: 0.4153\n",
      "Epoch 593/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8462 - accuracy: 0.5144 - val_loss: 2.2579 - val_accuracy: 0.4131\n",
      "Epoch 594/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8474 - accuracy: 0.5138 - val_loss: 2.2496 - val_accuracy: 0.4141\n",
      "Epoch 595/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8504 - accuracy: 0.5124 - val_loss: 2.2594 - val_accuracy: 0.4161\n",
      "Epoch 596/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8501 - accuracy: 0.5120 - val_loss: 2.2564 - val_accuracy: 0.4123\n",
      "Epoch 597/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8488 - accuracy: 0.5129 - val_loss: 2.2521 - val_accuracy: 0.4138\n",
      "Epoch 598/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8492 - accuracy: 0.5127 - val_loss: 2.2521 - val_accuracy: 0.4163\n",
      "Epoch 599/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8493 - accuracy: 0.5130 - val_loss: 2.2520 - val_accuracy: 0.4133\n",
      "Epoch 600/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8486 - accuracy: 0.5124 - val_loss: 2.2543 - val_accuracy: 0.4119\n",
      "Epoch 601/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8467 - accuracy: 0.5137 - val_loss: 2.2519 - val_accuracy: 0.4135\n",
      "Epoch 602/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8474 - accuracy: 0.5131 - val_loss: 2.2485 - val_accuracy: 0.4155\n",
      "Epoch 603/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8471 - accuracy: 0.5131 - val_loss: 2.2545 - val_accuracy: 0.4123\n",
      "Epoch 604/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8450 - accuracy: 0.5148 - val_loss: 2.2490 - val_accuracy: 0.4161\n",
      "Epoch 605/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8461 - accuracy: 0.5140 - val_loss: 2.2531 - val_accuracy: 0.4114\n",
      "Epoch 606/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8444 - accuracy: 0.5153 - val_loss: 2.2537 - val_accuracy: 0.4124\n",
      "Epoch 607/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8451 - accuracy: 0.5136 - val_loss: 2.2480 - val_accuracy: 0.4148\n",
      "Epoch 608/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8482 - accuracy: 0.5125 - val_loss: 2.2491 - val_accuracy: 0.4155\n",
      "Epoch 609/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8455 - accuracy: 0.5146 - val_loss: 2.2496 - val_accuracy: 0.4142\n",
      "Epoch 610/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8493 - accuracy: 0.5124 - val_loss: 2.2517 - val_accuracy: 0.4113\n",
      "Epoch 611/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8505 - accuracy: 0.5131 - val_loss: 2.2581 - val_accuracy: 0.4105\n",
      "Epoch 612/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8469 - accuracy: 0.5131 - val_loss: 2.2634 - val_accuracy: 0.4130\n",
      "Epoch 613/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8481 - accuracy: 0.5130 - val_loss: 2.2597 - val_accuracy: 0.4125\n",
      "Epoch 614/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8469 - accuracy: 0.5145 - val_loss: 2.2621 - val_accuracy: 0.4130\n",
      "Epoch 615/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8453 - accuracy: 0.5135 - val_loss: 2.2516 - val_accuracy: 0.4163\n",
      "Epoch 616/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8480 - accuracy: 0.5138 - val_loss: 2.2562 - val_accuracy: 0.4127\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 617/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8488 - accuracy: 0.5134 - val_loss: 2.2548 - val_accuracy: 0.4114\n",
      "Epoch 618/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8470 - accuracy: 0.5138 - val_loss: 2.2569 - val_accuracy: 0.4132\n",
      "Epoch 619/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8483 - accuracy: 0.5139 - val_loss: 2.2549 - val_accuracy: 0.4139\n",
      "Epoch 620/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8463 - accuracy: 0.5142 - val_loss: 2.2589 - val_accuracy: 0.4136\n",
      "Epoch 621/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8469 - accuracy: 0.5130 - val_loss: 2.2521 - val_accuracy: 0.4158\n",
      "Epoch 622/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8471 - accuracy: 0.5142 - val_loss: 2.2588 - val_accuracy: 0.4134\n",
      "Epoch 623/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8457 - accuracy: 0.5148 - val_loss: 2.2579 - val_accuracy: 0.4093\n",
      "Epoch 624/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8486 - accuracy: 0.5130 - val_loss: 2.2553 - val_accuracy: 0.4133\n",
      "Epoch 625/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8449 - accuracy: 0.5150 - val_loss: 2.2617 - val_accuracy: 0.4130\n",
      "Epoch 626/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8501 - accuracy: 0.5122 - val_loss: 2.2510 - val_accuracy: 0.4123\n",
      "Epoch 627/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8449 - accuracy: 0.5149 - val_loss: 2.2558 - val_accuracy: 0.4119\n",
      "Epoch 628/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8452 - accuracy: 0.5147 - val_loss: 2.2511 - val_accuracy: 0.4149\n",
      "Epoch 629/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8459 - accuracy: 0.5145 - val_loss: 2.2483 - val_accuracy: 0.4148\n",
      "Epoch 630/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8451 - accuracy: 0.5134 - val_loss: 2.2536 - val_accuracy: 0.4145\n",
      "Epoch 631/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8443 - accuracy: 0.5150 - val_loss: 2.2528 - val_accuracy: 0.4122\n",
      "Epoch 632/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8475 - accuracy: 0.5130 - val_loss: 2.2646 - val_accuracy: 0.4149\n",
      "Epoch 633/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8445 - accuracy: 0.5154 - val_loss: 2.2603 - val_accuracy: 0.4139\n",
      "Epoch 634/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8439 - accuracy: 0.5147 - val_loss: 2.2582 - val_accuracy: 0.4142\n",
      "Epoch 635/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8473 - accuracy: 0.5142 - val_loss: 2.2522 - val_accuracy: 0.4155\n",
      "Epoch 636/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8463 - accuracy: 0.5139 - val_loss: 2.2485 - val_accuracy: 0.4153\n",
      "Epoch 637/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8445 - accuracy: 0.5151 - val_loss: 2.2519 - val_accuracy: 0.4145\n",
      "Epoch 638/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8443 - accuracy: 0.5142 - val_loss: 2.2520 - val_accuracy: 0.4147\n",
      "Epoch 639/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8458 - accuracy: 0.5134 - val_loss: 2.2507 - val_accuracy: 0.4157\n",
      "Epoch 640/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8434 - accuracy: 0.5153 - val_loss: 2.2569 - val_accuracy: 0.4129\n",
      "Epoch 641/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8450 - accuracy: 0.5139 - val_loss: 2.2549 - val_accuracy: 0.4129\n",
      "Epoch 642/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8440 - accuracy: 0.5143 - val_loss: 2.2592 - val_accuracy: 0.4138\n",
      "Epoch 643/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8476 - accuracy: 0.5138 - val_loss: 2.2556 - val_accuracy: 0.4128\n",
      "Epoch 644/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8445 - accuracy: 0.5139 - val_loss: 2.2474 - val_accuracy: 0.4106\n",
      "Epoch 645/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8446 - accuracy: 0.5148 - val_loss: 2.2542 - val_accuracy: 0.4127\n",
      "Epoch 646/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8426 - accuracy: 0.5146 - val_loss: 2.2663 - val_accuracy: 0.4136\n",
      "Epoch 647/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8442 - accuracy: 0.5139 - val_loss: 2.2620 - val_accuracy: 0.4111\n",
      "Epoch 648/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8428 - accuracy: 0.5147 - val_loss: 2.2572 - val_accuracy: 0.4142\n",
      "Epoch 649/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8430 - accuracy: 0.5148 - val_loss: 2.2563 - val_accuracy: 0.4129\n",
      "Epoch 650/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8425 - accuracy: 0.5153 - val_loss: 2.2614 - val_accuracy: 0.4130\n",
      "Epoch 651/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8450 - accuracy: 0.5149 - val_loss: 2.2569 - val_accuracy: 0.4126\n",
      "Epoch 652/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8404 - accuracy: 0.5154 - val_loss: 2.2514 - val_accuracy: 0.4149\n",
      "Epoch 653/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8448 - accuracy: 0.5149 - val_loss: 2.2548 - val_accuracy: 0.4133\n",
      "Epoch 654/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8425 - accuracy: 0.5151 - val_loss: 2.2576 - val_accuracy: 0.4121\n",
      "Epoch 655/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8465 - accuracy: 0.5144 - val_loss: 2.2511 - val_accuracy: 0.4154\n",
      "Epoch 656/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8413 - accuracy: 0.5150 - val_loss: 2.2551 - val_accuracy: 0.4128\n",
      "Epoch 657/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8419 - accuracy: 0.5153 - val_loss: 2.2541 - val_accuracy: 0.4131\n",
      "Epoch 658/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8405 - accuracy: 0.5160 - val_loss: 2.2551 - val_accuracy: 0.4167\n",
      "Epoch 659/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8459 - accuracy: 0.5144 - val_loss: 2.2530 - val_accuracy: 0.4147\n",
      "Epoch 660/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8419 - accuracy: 0.5153 - val_loss: 2.2519 - val_accuracy: 0.4147\n",
      "Epoch 661/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8432 - accuracy: 0.5160 - val_loss: 2.2556 - val_accuracy: 0.4123\n",
      "Epoch 662/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8413 - accuracy: 0.5154 - val_loss: 2.2494 - val_accuracy: 0.4196\n",
      "Epoch 663/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8436 - accuracy: 0.5142 - val_loss: 2.2475 - val_accuracy: 0.4152\n",
      "Epoch 664/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8438 - accuracy: 0.5146 - val_loss: 2.2526 - val_accuracy: 0.4108\n",
      "Epoch 665/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8430 - accuracy: 0.5154 - val_loss: 2.2505 - val_accuracy: 0.4184\n",
      "Epoch 666/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8449 - accuracy: 0.5140 - val_loss: 2.2561 - val_accuracy: 0.4133\n",
      "Epoch 667/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8407 - accuracy: 0.5160 - val_loss: 2.2522 - val_accuracy: 0.4142\n",
      "Epoch 668/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8433 - accuracy: 0.5153 - val_loss: 2.2585 - val_accuracy: 0.4133\n",
      "Epoch 669/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8424 - accuracy: 0.5154 - val_loss: 2.2532 - val_accuracy: 0.4176\n",
      "Epoch 670/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8396 - accuracy: 0.5167 - val_loss: 2.2553 - val_accuracy: 0.4154\n",
      "Epoch 671/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8450 - accuracy: 0.5149 - val_loss: 2.2518 - val_accuracy: 0.4121\n",
      "Epoch 672/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8434 - accuracy: 0.5149 - val_loss: 2.2592 - val_accuracy: 0.4139\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 673/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8392 - accuracy: 0.5166 - val_loss: 2.2526 - val_accuracy: 0.4155\n",
      "Epoch 674/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8432 - accuracy: 0.5150 - val_loss: 2.2461 - val_accuracy: 0.4145\n",
      "Epoch 675/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8436 - accuracy: 0.5149 - val_loss: 2.2556 - val_accuracy: 0.4133\n",
      "Epoch 676/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8406 - accuracy: 0.5152 - val_loss: 2.2555 - val_accuracy: 0.4156\n",
      "Epoch 677/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8407 - accuracy: 0.5160 - val_loss: 2.2597 - val_accuracy: 0.4136\n",
      "Epoch 678/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8422 - accuracy: 0.5150 - val_loss: 2.2564 - val_accuracy: 0.4123\n",
      "Epoch 679/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8398 - accuracy: 0.5162 - val_loss: 2.2501 - val_accuracy: 0.4131\n",
      "Epoch 680/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8413 - accuracy: 0.5155 - val_loss: 2.2537 - val_accuracy: 0.4114\n",
      "Epoch 681/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8432 - accuracy: 0.5158 - val_loss: 2.2526 - val_accuracy: 0.4130\n",
      "Epoch 682/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8423 - accuracy: 0.5163 - val_loss: 2.2601 - val_accuracy: 0.4130\n",
      "Epoch 683/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8449 - accuracy: 0.5147 - val_loss: 2.2434 - val_accuracy: 0.4111\n",
      "Epoch 684/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8406 - accuracy: 0.5168 - val_loss: 2.2506 - val_accuracy: 0.4152\n",
      "Epoch 685/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8403 - accuracy: 0.5161 - val_loss: 2.2583 - val_accuracy: 0.4126\n",
      "Epoch 686/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8433 - accuracy: 0.5150 - val_loss: 2.2569 - val_accuracy: 0.4136\n",
      "Epoch 687/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8418 - accuracy: 0.5148 - val_loss: 2.2503 - val_accuracy: 0.4181\n",
      "Epoch 688/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8424 - accuracy: 0.5142 - val_loss: 2.2479 - val_accuracy: 0.4113\n",
      "Epoch 689/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8394 - accuracy: 0.5169 - val_loss: 2.2567 - val_accuracy: 0.4139\n",
      "Epoch 690/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8403 - accuracy: 0.5161 - val_loss: 2.2502 - val_accuracy: 0.4161\n",
      "Epoch 691/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8379 - accuracy: 0.5176 - val_loss: 2.2513 - val_accuracy: 0.4165\n",
      "Epoch 692/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8411 - accuracy: 0.5158 - val_loss: 2.2548 - val_accuracy: 0.4123\n",
      "Epoch 693/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8404 - accuracy: 0.5158 - val_loss: 2.2442 - val_accuracy: 0.4153\n",
      "Epoch 694/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8372 - accuracy: 0.5161 - val_loss: 2.2471 - val_accuracy: 0.4154\n",
      "Epoch 695/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8406 - accuracy: 0.5171 - val_loss: 2.2607 - val_accuracy: 0.4163\n",
      "Epoch 696/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8388 - accuracy: 0.5163 - val_loss: 2.2476 - val_accuracy: 0.4161\n",
      "Epoch 697/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8386 - accuracy: 0.5185 - val_loss: 2.2531 - val_accuracy: 0.4139\n",
      "Epoch 698/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8423 - accuracy: 0.5163 - val_loss: 2.2519 - val_accuracy: 0.4099\n",
      "Epoch 699/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8388 - accuracy: 0.5158 - val_loss: 2.2509 - val_accuracy: 0.4144\n",
      "Epoch 700/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8404 - accuracy: 0.5160 - val_loss: 2.2550 - val_accuracy: 0.4140\n",
      "Epoch 701/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8414 - accuracy: 0.5158 - val_loss: 2.2523 - val_accuracy: 0.4127\n",
      "Epoch 702/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8404 - accuracy: 0.5164 - val_loss: 2.2583 - val_accuracy: 0.4135\n",
      "Epoch 703/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8422 - accuracy: 0.5152 - val_loss: 2.2515 - val_accuracy: 0.4119\n",
      "Epoch 704/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8410 - accuracy: 0.5159 - val_loss: 2.2574 - val_accuracy: 0.4154\n",
      "Epoch 705/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8425 - accuracy: 0.5161 - val_loss: 2.2533 - val_accuracy: 0.4136\n",
      "Epoch 706/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8382 - accuracy: 0.5176 - val_loss: 2.2596 - val_accuracy: 0.4089\n",
      "Epoch 707/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8405 - accuracy: 0.5162 - val_loss: 2.2512 - val_accuracy: 0.4116\n",
      "Epoch 708/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8395 - accuracy: 0.5167 - val_loss: 2.2550 - val_accuracy: 0.4145\n",
      "Epoch 709/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8365 - accuracy: 0.5175 - val_loss: 2.2586 - val_accuracy: 0.4124\n",
      "Epoch 710/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8420 - accuracy: 0.5154 - val_loss: 2.2455 - val_accuracy: 0.4129\n",
      "Epoch 711/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8370 - accuracy: 0.5173 - val_loss: 2.2607 - val_accuracy: 0.4149\n",
      "Epoch 712/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8397 - accuracy: 0.5166 - val_loss: 2.2571 - val_accuracy: 0.4158\n",
      "Epoch 713/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8382 - accuracy: 0.5173 - val_loss: 2.2474 - val_accuracy: 0.4139\n",
      "Epoch 714/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8372 - accuracy: 0.5164 - val_loss: 2.2586 - val_accuracy: 0.4136\n",
      "Epoch 715/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8394 - accuracy: 0.5166 - val_loss: 2.2512 - val_accuracy: 0.4133\n",
      "Epoch 716/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8367 - accuracy: 0.5177 - val_loss: 2.2481 - val_accuracy: 0.4168\n",
      "Epoch 717/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8379 - accuracy: 0.5178 - val_loss: 2.2496 - val_accuracy: 0.4166\n",
      "Epoch 718/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8404 - accuracy: 0.5161 - val_loss: 2.2558 - val_accuracy: 0.4131\n",
      "Epoch 719/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8407 - accuracy: 0.5160 - val_loss: 2.2559 - val_accuracy: 0.4128\n",
      "Epoch 720/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8404 - accuracy: 0.5161 - val_loss: 2.2521 - val_accuracy: 0.4156\n",
      "Epoch 721/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8404 - accuracy: 0.5166 - val_loss: 2.2508 - val_accuracy: 0.4137\n",
      "Epoch 722/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8390 - accuracy: 0.5174 - val_loss: 2.2536 - val_accuracy: 0.4158\n",
      "Epoch 723/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8375 - accuracy: 0.5164 - val_loss: 2.2608 - val_accuracy: 0.4104\n",
      "Epoch 724/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8378 - accuracy: 0.5163 - val_loss: 2.2549 - val_accuracy: 0.4168\n",
      "Epoch 725/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8381 - accuracy: 0.5173 - val_loss: 2.2496 - val_accuracy: 0.4158\n",
      "Epoch 726/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8393 - accuracy: 0.5174 - val_loss: 2.2540 - val_accuracy: 0.4172\n",
      "Epoch 727/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8380 - accuracy: 0.5172 - val_loss: 2.2573 - val_accuracy: 0.4142\n",
      "Epoch 728/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8358 - accuracy: 0.5176 - val_loss: 2.2586 - val_accuracy: 0.4140\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 729/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8368 - accuracy: 0.5169 - val_loss: 2.2517 - val_accuracy: 0.4165\n",
      "Epoch 730/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8399 - accuracy: 0.5171 - val_loss: 2.2578 - val_accuracy: 0.4150\n",
      "Epoch 731/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8371 - accuracy: 0.5168 - val_loss: 2.2567 - val_accuracy: 0.4144\n",
      "Epoch 732/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8361 - accuracy: 0.5177 - val_loss: 2.2526 - val_accuracy: 0.4145\n",
      "Epoch 733/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8356 - accuracy: 0.5184 - val_loss: 2.2564 - val_accuracy: 0.4155\n",
      "Epoch 734/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8337 - accuracy: 0.5188 - val_loss: 2.2587 - val_accuracy: 0.4131\n",
      "Epoch 735/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8363 - accuracy: 0.5181 - val_loss: 2.2543 - val_accuracy: 0.4107\n",
      "Epoch 736/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8376 - accuracy: 0.5178 - val_loss: 2.2465 - val_accuracy: 0.4155\n",
      "Epoch 737/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8367 - accuracy: 0.5177 - val_loss: 2.2527 - val_accuracy: 0.4150\n",
      "Epoch 738/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8373 - accuracy: 0.5176 - val_loss: 2.2520 - val_accuracy: 0.4151\n",
      "Epoch 739/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8344 - accuracy: 0.5189 - val_loss: 2.2523 - val_accuracy: 0.4165\n",
      "Epoch 740/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8386 - accuracy: 0.5179 - val_loss: 2.2498 - val_accuracy: 0.4160\n",
      "Epoch 741/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8386 - accuracy: 0.5170 - val_loss: 2.2558 - val_accuracy: 0.4124\n",
      "Epoch 742/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8375 - accuracy: 0.5176 - val_loss: 2.2559 - val_accuracy: 0.4107\n",
      "Epoch 743/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8352 - accuracy: 0.5189 - val_loss: 2.2562 - val_accuracy: 0.4139\n",
      "Epoch 744/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8366 - accuracy: 0.5177 - val_loss: 2.2590 - val_accuracy: 0.4143\n",
      "Epoch 745/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8369 - accuracy: 0.5179 - val_loss: 2.2457 - val_accuracy: 0.4161\n",
      "Epoch 746/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8369 - accuracy: 0.5177 - val_loss: 2.2539 - val_accuracy: 0.4148\n",
      "Epoch 747/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8378 - accuracy: 0.5174 - val_loss: 2.2518 - val_accuracy: 0.4135\n",
      "Epoch 748/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8370 - accuracy: 0.5171 - val_loss: 2.2507 - val_accuracy: 0.4168\n",
      "Epoch 749/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8363 - accuracy: 0.5183 - val_loss: 2.2530 - val_accuracy: 0.4118\n",
      "Epoch 750/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8345 - accuracy: 0.5177 - val_loss: 2.2587 - val_accuracy: 0.4154\n",
      "Epoch 751/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5177 - val_loss: 2.2503 - val_accuracy: 0.4159\n",
      "Epoch 752/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8343 - accuracy: 0.5184 - val_loss: 2.2543 - val_accuracy: 0.4117\n",
      "Epoch 753/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8374 - accuracy: 0.5171 - val_loss: 2.2527 - val_accuracy: 0.4135\n",
      "Epoch 754/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5184 - val_loss: 2.2519 - val_accuracy: 0.4150\n",
      "Epoch 755/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8343 - accuracy: 0.5204 - val_loss: 2.2620 - val_accuracy: 0.4103\n",
      "Epoch 756/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8370 - accuracy: 0.5173 - val_loss: 2.2576 - val_accuracy: 0.4135\n",
      "Epoch 757/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8368 - accuracy: 0.5183 - val_loss: 2.2540 - val_accuracy: 0.4122\n",
      "Epoch 758/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5187 - val_loss: 2.2670 - val_accuracy: 0.4111\n",
      "Epoch 759/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8393 - accuracy: 0.5175 - val_loss: 2.2592 - val_accuracy: 0.4114\n",
      "Epoch 760/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8370 - accuracy: 0.5178 - val_loss: 2.2622 - val_accuracy: 0.4141\n",
      "Epoch 761/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5186 - val_loss: 2.2622 - val_accuracy: 0.4151\n",
      "Epoch 762/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8356 - accuracy: 0.5179 - val_loss: 2.2657 - val_accuracy: 0.4138\n",
      "Epoch 763/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8357 - accuracy: 0.5189 - val_loss: 2.2578 - val_accuracy: 0.4151\n",
      "Epoch 764/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8396 - accuracy: 0.5159 - val_loss: 2.2582 - val_accuracy: 0.4124\n",
      "Epoch 765/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8337 - accuracy: 0.5179 - val_loss: 2.2625 - val_accuracy: 0.4115\n",
      "Epoch 766/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8362 - accuracy: 0.5180 - val_loss: 2.2586 - val_accuracy: 0.4129\n",
      "Epoch 767/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8324 - accuracy: 0.5195 - val_loss: 2.2520 - val_accuracy: 0.4152\n",
      "Epoch 768/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8380 - accuracy: 0.5176 - val_loss: 2.2585 - val_accuracy: 0.4135\n",
      "Epoch 769/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8346 - accuracy: 0.5178 - val_loss: 2.2548 - val_accuracy: 0.4156\n",
      "Epoch 770/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8355 - accuracy: 0.5179 - val_loss: 2.2513 - val_accuracy: 0.4158\n",
      "Epoch 771/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8365 - accuracy: 0.5188 - val_loss: 2.2651 - val_accuracy: 0.4128\n",
      "Epoch 772/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8340 - accuracy: 0.5181 - val_loss: 2.2614 - val_accuracy: 0.4131\n",
      "Epoch 773/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5180 - val_loss: 2.2606 - val_accuracy: 0.4164\n",
      "Epoch 774/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5194 - val_loss: 2.2542 - val_accuracy: 0.4168\n",
      "Epoch 775/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8365 - accuracy: 0.5173 - val_loss: 2.2609 - val_accuracy: 0.4139\n",
      "Epoch 776/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8369 - accuracy: 0.5181 - val_loss: 2.2620 - val_accuracy: 0.4158\n",
      "Epoch 777/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8358 - accuracy: 0.5179 - val_loss: 2.2636 - val_accuracy: 0.4107\n",
      "Epoch 778/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8329 - accuracy: 0.5194 - val_loss: 2.2534 - val_accuracy: 0.4168\n",
      "Epoch 779/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8339 - accuracy: 0.5183 - val_loss: 2.2554 - val_accuracy: 0.4139\n",
      "Epoch 780/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8351 - accuracy: 0.5187 - val_loss: 2.2639 - val_accuracy: 0.4110\n",
      "Epoch 781/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8379 - accuracy: 0.5171 - val_loss: 2.2557 - val_accuracy: 0.4155\n",
      "Epoch 782/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8339 - accuracy: 0.5188 - val_loss: 2.2635 - val_accuracy: 0.4104\n",
      "Epoch 783/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5191 - val_loss: 2.2571 - val_accuracy: 0.4144\n",
      "Epoch 784/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8363 - accuracy: 0.5179 - val_loss: 2.2568 - val_accuracy: 0.4101\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 785/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8327 - accuracy: 0.5196 - val_loss: 2.2495 - val_accuracy: 0.4129\n",
      "Epoch 786/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8341 - accuracy: 0.5180 - val_loss: 2.2552 - val_accuracy: 0.4156\n",
      "Epoch 787/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8339 - accuracy: 0.5186 - val_loss: 2.2585 - val_accuracy: 0.4111\n",
      "Epoch 788/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8341 - accuracy: 0.5177 - val_loss: 2.2670 - val_accuracy: 0.4121\n",
      "Epoch 789/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8325 - accuracy: 0.5193 - val_loss: 2.2596 - val_accuracy: 0.4136\n",
      "Epoch 790/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8319 - accuracy: 0.5197 - val_loss: 2.2573 - val_accuracy: 0.4108\n",
      "Epoch 791/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8325 - accuracy: 0.5189 - val_loss: 2.2547 - val_accuracy: 0.4158\n",
      "Epoch 792/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8312 - accuracy: 0.5192 - val_loss: 2.2673 - val_accuracy: 0.4117\n",
      "Epoch 793/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8329 - accuracy: 0.5186 - val_loss: 2.2600 - val_accuracy: 0.4154\n",
      "Epoch 794/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8313 - accuracy: 0.5204 - val_loss: 2.2631 - val_accuracy: 0.4125\n",
      "Epoch 795/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8347 - accuracy: 0.5182 - val_loss: 2.2616 - val_accuracy: 0.4146\n",
      "Epoch 796/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8350 - accuracy: 0.5181 - val_loss: 2.2634 - val_accuracy: 0.4168\n",
      "Epoch 797/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8349 - accuracy: 0.5191 - val_loss: 2.2593 - val_accuracy: 0.4156\n",
      "Epoch 798/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8305 - accuracy: 0.5198 - val_loss: 2.2648 - val_accuracy: 0.4127\n",
      "Epoch 799/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8344 - accuracy: 0.5184 - val_loss: 2.2583 - val_accuracy: 0.4114\n",
      "Epoch 800/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8318 - accuracy: 0.5191 - val_loss: 2.2612 - val_accuracy: 0.4140\n",
      "Epoch 801/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8336 - accuracy: 0.5179 - val_loss: 2.2663 - val_accuracy: 0.4117\n",
      "Epoch 802/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8352 - accuracy: 0.5182 - val_loss: 2.2573 - val_accuracy: 0.4152\n",
      "Epoch 803/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8352 - accuracy: 0.5185 - val_loss: 2.2625 - val_accuracy: 0.4158\n",
      "Epoch 804/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8320 - accuracy: 0.5204 - val_loss: 2.2671 - val_accuracy: 0.4123\n",
      "Epoch 805/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8344 - accuracy: 0.5180 - val_loss: 2.2504 - val_accuracy: 0.4139\n",
      "Epoch 806/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8338 - accuracy: 0.5185 - val_loss: 2.2572 - val_accuracy: 0.4142\n",
      "Epoch 807/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8342 - accuracy: 0.5175 - val_loss: 2.2515 - val_accuracy: 0.4138\n",
      "Epoch 808/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8314 - accuracy: 0.5196 - val_loss: 2.2535 - val_accuracy: 0.4109\n",
      "Epoch 809/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8338 - accuracy: 0.5191 - val_loss: 2.2558 - val_accuracy: 0.4151\n",
      "Epoch 810/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8338 - accuracy: 0.5197 - val_loss: 2.2536 - val_accuracy: 0.4130\n",
      "Epoch 811/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8298 - accuracy: 0.5198 - val_loss: 2.2591 - val_accuracy: 0.4138\n",
      "Epoch 812/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8322 - accuracy: 0.5196 - val_loss: 2.2609 - val_accuracy: 0.4133\n",
      "Epoch 813/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8334 - accuracy: 0.5189 - val_loss: 2.2600 - val_accuracy: 0.4141\n",
      "Epoch 814/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8359 - accuracy: 0.5193 - val_loss: 2.2548 - val_accuracy: 0.4133\n",
      "Epoch 815/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8302 - accuracy: 0.5196 - val_loss: 2.2599 - val_accuracy: 0.4130\n",
      "Epoch 816/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8334 - accuracy: 0.5182 - val_loss: 2.2575 - val_accuracy: 0.4155\n",
      "Epoch 817/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8338 - accuracy: 0.5187 - val_loss: 2.2581 - val_accuracy: 0.4122\n",
      "Epoch 818/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8295 - accuracy: 0.5198 - val_loss: 2.2634 - val_accuracy: 0.4123\n",
      "Epoch 819/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8346 - accuracy: 0.5196 - val_loss: 2.2564 - val_accuracy: 0.4175\n",
      "Epoch 820/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8338 - accuracy: 0.5181 - val_loss: 2.2556 - val_accuracy: 0.4129\n",
      "Epoch 821/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8319 - accuracy: 0.5206 - val_loss: 2.2535 - val_accuracy: 0.4136\n",
      "Epoch 822/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8304 - accuracy: 0.5199 - val_loss: 2.2586 - val_accuracy: 0.4129\n",
      "Epoch 823/2000\n",
      "242/242 [==============================] - 6s 26ms/step - loss: 1.8292 - accuracy: 0.5202 - val_loss: 2.2559 - val_accuracy: 0.4148\n",
      "Epoch 824/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8314 - accuracy: 0.5199 - val_loss: 2.2632 - val_accuracy: 0.4132\n",
      "Epoch 825/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8308 - accuracy: 0.5203 - val_loss: 2.2585 - val_accuracy: 0.4133\n",
      "Epoch 826/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8335 - accuracy: 0.5204 - val_loss: 2.2569 - val_accuracy: 0.4130\n",
      "Epoch 827/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8337 - accuracy: 0.5196 - val_loss: 2.2635 - val_accuracy: 0.4139\n",
      "Epoch 828/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8332 - accuracy: 0.5192 - val_loss: 2.2507 - val_accuracy: 0.4187\n",
      "Epoch 829/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8296 - accuracy: 0.5206 - val_loss: 2.2652 - val_accuracy: 0.4100\n",
      "Epoch 830/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8311 - accuracy: 0.5198 - val_loss: 2.2616 - val_accuracy: 0.4136\n",
      "Epoch 831/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8306 - accuracy: 0.5197 - val_loss: 2.2541 - val_accuracy: 0.4181\n",
      "Epoch 832/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8302 - accuracy: 0.5205 - val_loss: 2.2636 - val_accuracy: 0.4127\n",
      "Epoch 833/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8309 - accuracy: 0.5206 - val_loss: 2.2564 - val_accuracy: 0.4156\n",
      "Epoch 834/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8316 - accuracy: 0.5191 - val_loss: 2.2539 - val_accuracy: 0.4151\n",
      "Epoch 835/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8315 - accuracy: 0.5208 - val_loss: 2.2611 - val_accuracy: 0.4101\n",
      "Epoch 836/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8309 - accuracy: 0.5202 - val_loss: 2.2656 - val_accuracy: 0.4096\n",
      "Epoch 837/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8335 - accuracy: 0.5186 - val_loss: 2.2608 - val_accuracy: 0.4129\n",
      "Epoch 838/2000\n",
      "242/242 [==============================] - 6s 25ms/step - loss: 1.8291 - accuracy: 0.5212 - val_loss: 2.2595 - val_accuracy: 0.4120\n",
      "Epoch 839/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8286 - accuracy: 0.5212 - val_loss: 2.2576 - val_accuracy: 0.4136\n",
      "Epoch 840/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8325 - accuracy: 0.5189 - val_loss: 2.2634 - val_accuracy: 0.4094\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 841/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8324 - accuracy: 0.5206 - val_loss: 2.2550 - val_accuracy: 0.4153\n",
      "Epoch 842/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8318 - accuracy: 0.5207 - val_loss: 2.2525 - val_accuracy: 0.4149\n",
      "Epoch 843/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8344 - accuracy: 0.5186 - val_loss: 2.2607 - val_accuracy: 0.4130\n",
      "Epoch 844/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8289 - accuracy: 0.5209 - val_loss: 2.2565 - val_accuracy: 0.4151\n",
      "Epoch 845/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8289 - accuracy: 0.5201 - val_loss: 2.2685 - val_accuracy: 0.4095\n",
      "Epoch 846/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8308 - accuracy: 0.5188 - val_loss: 2.2578 - val_accuracy: 0.4115\n",
      "Epoch 847/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8291 - accuracy: 0.5210 - val_loss: 2.2551 - val_accuracy: 0.4142\n",
      "Epoch 848/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8289 - accuracy: 0.5203 - val_loss: 2.2546 - val_accuracy: 0.4138\n",
      "Epoch 849/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8292 - accuracy: 0.5198 - val_loss: 2.2610 - val_accuracy: 0.4135\n",
      "Epoch 850/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8325 - accuracy: 0.5199 - val_loss: 2.2586 - val_accuracy: 0.4142\n",
      "Epoch 851/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8289 - accuracy: 0.5201 - val_loss: 2.2539 - val_accuracy: 0.4165\n",
      "Epoch 852/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8291 - accuracy: 0.5197 - val_loss: 2.2579 - val_accuracy: 0.4111\n",
      "Epoch 853/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8315 - accuracy: 0.5195 - val_loss: 2.2524 - val_accuracy: 0.4139\n",
      "Epoch 854/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8303 - accuracy: 0.5201 - val_loss: 2.2579 - val_accuracy: 0.4124\n",
      "Epoch 855/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8304 - accuracy: 0.5211 - val_loss: 2.2624 - val_accuracy: 0.4141\n",
      "Epoch 856/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8296 - accuracy: 0.5209 - val_loss: 2.2592 - val_accuracy: 0.4148\n",
      "Epoch 857/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8305 - accuracy: 0.5201 - val_loss: 2.2567 - val_accuracy: 0.4092\n",
      "Epoch 858/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8318 - accuracy: 0.5196 - val_loss: 2.2557 - val_accuracy: 0.4142\n",
      "Epoch 859/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8289 - accuracy: 0.5205 - val_loss: 2.2645 - val_accuracy: 0.4107\n",
      "Epoch 860/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8284 - accuracy: 0.5204 - val_loss: 2.2578 - val_accuracy: 0.4105\n",
      "Epoch 861/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8303 - accuracy: 0.5200 - val_loss: 2.2570 - val_accuracy: 0.4155\n",
      "Epoch 862/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8291 - accuracy: 0.5203 - val_loss: 2.2563 - val_accuracy: 0.4165\n",
      "Epoch 863/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8305 - accuracy: 0.5200 - val_loss: 2.2607 - val_accuracy: 0.4111\n",
      "Epoch 864/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8294 - accuracy: 0.5203 - val_loss: 2.2645 - val_accuracy: 0.4120\n",
      "Epoch 865/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8307 - accuracy: 0.5199 - val_loss: 2.2558 - val_accuracy: 0.4145\n",
      "Epoch 866/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8298 - accuracy: 0.5204 - val_loss: 2.2513 - val_accuracy: 0.4154\n",
      "Epoch 867/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8264 - accuracy: 0.5210 - val_loss: 2.2609 - val_accuracy: 0.4138\n",
      "Epoch 868/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8303 - accuracy: 0.5204 - val_loss: 2.2644 - val_accuracy: 0.4154\n",
      "Epoch 869/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8267 - accuracy: 0.5218 - val_loss: 2.2695 - val_accuracy: 0.4133\n",
      "Epoch 870/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8313 - accuracy: 0.5206 - val_loss: 2.2631 - val_accuracy: 0.4126\n",
      "Epoch 871/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8288 - accuracy: 0.5202 - val_loss: 2.2620 - val_accuracy: 0.4176\n",
      "Epoch 872/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8321 - accuracy: 0.5197 - val_loss: 2.2599 - val_accuracy: 0.4131\n",
      "Epoch 873/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8293 - accuracy: 0.5193 - val_loss: 2.2611 - val_accuracy: 0.4117\n",
      "Epoch 874/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8274 - accuracy: 0.5212 - val_loss: 2.2551 - val_accuracy: 0.4140\n",
      "Epoch 875/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8259 - accuracy: 0.5217 - val_loss: 2.2587 - val_accuracy: 0.4146\n",
      "Epoch 876/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8287 - accuracy: 0.5211 - val_loss: 2.2661 - val_accuracy: 0.4113\n",
      "Epoch 877/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8280 - accuracy: 0.5210 - val_loss: 2.2639 - val_accuracy: 0.4136\n",
      "Epoch 878/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8320 - accuracy: 0.5191 - val_loss: 2.2617 - val_accuracy: 0.4146\n",
      "Epoch 879/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8280 - accuracy: 0.5213 - val_loss: 2.2735 - val_accuracy: 0.4131\n",
      "Epoch 880/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8293 - accuracy: 0.5198 - val_loss: 2.2642 - val_accuracy: 0.4113\n",
      "Epoch 881/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8295 - accuracy: 0.5201 - val_loss: 2.2571 - val_accuracy: 0.4123\n",
      "Epoch 882/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8282 - accuracy: 0.5208 - val_loss: 2.2594 - val_accuracy: 0.4139\n",
      "Epoch 883/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8326 - accuracy: 0.5208 - val_loss: 2.2630 - val_accuracy: 0.4130\n",
      "Epoch 884/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8280 - accuracy: 0.5216 - val_loss: 2.2650 - val_accuracy: 0.4130\n",
      "Epoch 885/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8271 - accuracy: 0.5222 - val_loss: 2.2639 - val_accuracy: 0.4123\n",
      "Epoch 886/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8258 - accuracy: 0.5219 - val_loss: 2.2623 - val_accuracy: 0.4104\n",
      "Epoch 887/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8307 - accuracy: 0.5194 - val_loss: 2.2571 - val_accuracy: 0.4098\n",
      "Epoch 888/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8311 - accuracy: 0.5199 - val_loss: 2.2666 - val_accuracy: 0.4103\n",
      "Epoch 889/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8299 - accuracy: 0.5202 - val_loss: 2.2675 - val_accuracy: 0.4129\n",
      "Epoch 890/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8299 - accuracy: 0.5202 - val_loss: 2.2708 - val_accuracy: 0.4079\n",
      "Epoch 891/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8267 - accuracy: 0.5211 - val_loss: 2.2558 - val_accuracy: 0.4120\n",
      "Epoch 892/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8273 - accuracy: 0.5209 - val_loss: 2.2593 - val_accuracy: 0.4136\n",
      "Epoch 893/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8295 - accuracy: 0.5203 - val_loss: 2.2679 - val_accuracy: 0.4127\n",
      "Epoch 894/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8255 - accuracy: 0.5225 - val_loss: 2.2594 - val_accuracy: 0.4137\n",
      "Epoch 895/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8286 - accuracy: 0.5211 - val_loss: 2.2625 - val_accuracy: 0.4122\n",
      "Epoch 896/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8297 - accuracy: 0.5210 - val_loss: 2.2569 - val_accuracy: 0.4136\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 897/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8254 - accuracy: 0.5226 - val_loss: 2.2660 - val_accuracy: 0.4109\n",
      "Epoch 898/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8271 - accuracy: 0.5216 - val_loss: 2.2623 - val_accuracy: 0.4151\n",
      "Epoch 899/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8277 - accuracy: 0.5205 - val_loss: 2.2624 - val_accuracy: 0.4098\n",
      "Epoch 900/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8279 - accuracy: 0.5216 - val_loss: 2.2632 - val_accuracy: 0.4107\n",
      "Epoch 901/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8283 - accuracy: 0.5216 - val_loss: 2.2608 - val_accuracy: 0.4120\n",
      "Epoch 902/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8281 - accuracy: 0.5220 - val_loss: 2.2659 - val_accuracy: 0.4117\n",
      "Epoch 903/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8289 - accuracy: 0.5211 - val_loss: 2.2636 - val_accuracy: 0.4145\n",
      "Epoch 904/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8265 - accuracy: 0.5211 - val_loss: 2.2594 - val_accuracy: 0.4130\n",
      "Epoch 905/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8272 - accuracy: 0.5213 - val_loss: 2.2635 - val_accuracy: 0.4141\n",
      "Epoch 906/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8291 - accuracy: 0.5214 - val_loss: 2.2623 - val_accuracy: 0.4114\n",
      "Epoch 907/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8263 - accuracy: 0.5212 - val_loss: 2.2676 - val_accuracy: 0.4114\n",
      "Epoch 908/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8281 - accuracy: 0.5213 - val_loss: 2.2724 - val_accuracy: 0.4112\n",
      "Epoch 909/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8311 - accuracy: 0.5205 - val_loss: 2.2668 - val_accuracy: 0.4096\n",
      "Epoch 910/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8268 - accuracy: 0.5216 - val_loss: 2.2609 - val_accuracy: 0.4092\n",
      "Epoch 911/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8285 - accuracy: 0.5208 - val_loss: 2.2604 - val_accuracy: 0.4125\n",
      "Epoch 912/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8255 - accuracy: 0.5219 - val_loss: 2.2617 - val_accuracy: 0.4150\n",
      "Epoch 913/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8262 - accuracy: 0.5213 - val_loss: 2.2696 - val_accuracy: 0.4115\n",
      "Epoch 914/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8253 - accuracy: 0.5231 - val_loss: 2.2658 - val_accuracy: 0.4129\n",
      "Epoch 915/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8280 - accuracy: 0.5211 - val_loss: 2.2619 - val_accuracy: 0.4151\n",
      "Epoch 916/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8260 - accuracy: 0.5219 - val_loss: 2.2676 - val_accuracy: 0.4096\n",
      "Epoch 917/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8277 - accuracy: 0.5217 - val_loss: 2.2642 - val_accuracy: 0.4142\n",
      "Epoch 918/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8244 - accuracy: 0.5215 - val_loss: 2.2621 - val_accuracy: 0.4112\n",
      "Epoch 919/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8245 - accuracy: 0.5227 - val_loss: 2.2613 - val_accuracy: 0.4113\n",
      "Epoch 920/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8291 - accuracy: 0.5210 - val_loss: 2.2596 - val_accuracy: 0.4109\n",
      "Epoch 921/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8246 - accuracy: 0.5231 - val_loss: 2.2583 - val_accuracy: 0.4162\n",
      "Epoch 922/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8248 - accuracy: 0.5233 - val_loss: 2.2673 - val_accuracy: 0.4113\n",
      "Epoch 923/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8264 - accuracy: 0.5217 - val_loss: 2.2578 - val_accuracy: 0.4132\n",
      "Epoch 924/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8274 - accuracy: 0.5214 - val_loss: 2.2642 - val_accuracy: 0.4116\n",
      "Epoch 925/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8254 - accuracy: 0.5228 - val_loss: 2.2668 - val_accuracy: 0.4117\n",
      "Epoch 926/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8256 - accuracy: 0.5219 - val_loss: 2.2585 - val_accuracy: 0.4122\n",
      "Epoch 927/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8284 - accuracy: 0.5204 - val_loss: 2.2605 - val_accuracy: 0.4124\n",
      "Epoch 928/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8260 - accuracy: 0.5213 - val_loss: 2.2578 - val_accuracy: 0.4117\n",
      "Epoch 929/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8275 - accuracy: 0.5209 - val_loss: 2.2603 - val_accuracy: 0.4114\n",
      "Epoch 930/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8273 - accuracy: 0.5213 - val_loss: 2.2597 - val_accuracy: 0.4142\n",
      "Epoch 931/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8277 - accuracy: 0.5214 - val_loss: 2.2687 - val_accuracy: 0.4111\n",
      "Epoch 932/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8263 - accuracy: 0.5223 - val_loss: 2.2645 - val_accuracy: 0.4120\n",
      "Epoch 933/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8227 - accuracy: 0.5226 - val_loss: 2.2697 - val_accuracy: 0.4111\n",
      "Epoch 934/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8272 - accuracy: 0.5207 - val_loss: 2.2697 - val_accuracy: 0.4101\n",
      "Epoch 935/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8221 - accuracy: 0.5220 - val_loss: 2.2706 - val_accuracy: 0.4107\n",
      "Epoch 936/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8279 - accuracy: 0.5215 - val_loss: 2.2668 - val_accuracy: 0.4111\n",
      "Epoch 937/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8260 - accuracy: 0.5220 - val_loss: 2.2654 - val_accuracy: 0.4117\n",
      "Epoch 938/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8263 - accuracy: 0.5201 - val_loss: 2.2705 - val_accuracy: 0.4082\n",
      "Epoch 939/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8224 - accuracy: 0.5228 - val_loss: 2.2627 - val_accuracy: 0.4127\n",
      "Epoch 940/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8280 - accuracy: 0.5210 - val_loss: 2.2619 - val_accuracy: 0.4099\n",
      "Epoch 941/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8262 - accuracy: 0.5213 - val_loss: 2.2536 - val_accuracy: 0.4137\n",
      "Epoch 942/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8253 - accuracy: 0.5223 - val_loss: 2.2621 - val_accuracy: 0.4142\n",
      "Epoch 943/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8249 - accuracy: 0.5230 - val_loss: 2.2616 - val_accuracy: 0.4115\n",
      "Epoch 944/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8255 - accuracy: 0.5223 - val_loss: 2.2586 - val_accuracy: 0.4091\n",
      "Epoch 945/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8266 - accuracy: 0.5213 - val_loss: 2.2564 - val_accuracy: 0.4127\n",
      "Epoch 946/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8254 - accuracy: 0.5227 - val_loss: 2.2668 - val_accuracy: 0.4096\n",
      "Epoch 947/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8234 - accuracy: 0.5224 - val_loss: 2.2645 - val_accuracy: 0.4127\n",
      "Epoch 948/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8249 - accuracy: 0.5225 - val_loss: 2.2607 - val_accuracy: 0.4135\n",
      "Epoch 949/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8268 - accuracy: 0.5208 - val_loss: 2.2569 - val_accuracy: 0.4152\n",
      "Epoch 950/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8268 - accuracy: 0.5232 - val_loss: 2.2664 - val_accuracy: 0.4139\n",
      "Epoch 951/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8231 - accuracy: 0.5224 - val_loss: 2.2655 - val_accuracy: 0.4114\n",
      "Epoch 952/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8284 - accuracy: 0.5217 - val_loss: 2.2606 - val_accuracy: 0.4104\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 953/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8250 - accuracy: 0.5216 - val_loss: 2.2519 - val_accuracy: 0.4135\n",
      "Epoch 954/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8252 - accuracy: 0.5216 - val_loss: 2.2627 - val_accuracy: 0.4120\n",
      "Epoch 955/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8285 - accuracy: 0.5207 - val_loss: 2.2573 - val_accuracy: 0.4155\n",
      "Epoch 956/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8239 - accuracy: 0.5224 - val_loss: 2.2695 - val_accuracy: 0.4129\n",
      "Epoch 957/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8219 - accuracy: 0.5234 - val_loss: 2.2595 - val_accuracy: 0.4133\n",
      "Epoch 958/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8241 - accuracy: 0.5222 - val_loss: 2.2695 - val_accuracy: 0.4133\n",
      "Epoch 959/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8234 - accuracy: 0.5227 - val_loss: 2.2526 - val_accuracy: 0.4162\n",
      "Epoch 960/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8226 - accuracy: 0.5228 - val_loss: 2.2586 - val_accuracy: 0.4152\n",
      "Epoch 961/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8233 - accuracy: 0.5223 - val_loss: 2.2611 - val_accuracy: 0.4116\n",
      "Epoch 962/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8279 - accuracy: 0.5214 - val_loss: 2.2660 - val_accuracy: 0.4125\n",
      "Epoch 963/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8260 - accuracy: 0.5228 - val_loss: 2.2596 - val_accuracy: 0.4143\n",
      "Epoch 964/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8254 - accuracy: 0.5212 - val_loss: 2.2613 - val_accuracy: 0.4139\n",
      "Epoch 965/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8238 - accuracy: 0.5240 - val_loss: 2.2546 - val_accuracy: 0.4163\n",
      "Epoch 966/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8233 - accuracy: 0.5231 - val_loss: 2.2609 - val_accuracy: 0.4136\n",
      "Epoch 967/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8247 - accuracy: 0.5229 - val_loss: 2.2541 - val_accuracy: 0.4097\n",
      "Epoch 968/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8226 - accuracy: 0.5235 - val_loss: 2.2603 - val_accuracy: 0.4107\n",
      "Epoch 969/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8230 - accuracy: 0.5222 - val_loss: 2.2629 - val_accuracy: 0.4111\n",
      "Epoch 970/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8227 - accuracy: 0.5236 - val_loss: 2.2603 - val_accuracy: 0.4107\n",
      "Epoch 971/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8224 - accuracy: 0.5233 - val_loss: 2.2665 - val_accuracy: 0.4112\n",
      "Epoch 972/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8256 - accuracy: 0.5223 - val_loss: 2.2637 - val_accuracy: 0.4123\n",
      "Epoch 973/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8238 - accuracy: 0.5229 - val_loss: 2.2695 - val_accuracy: 0.4147\n",
      "Epoch 974/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8227 - accuracy: 0.5230 - val_loss: 2.2659 - val_accuracy: 0.4156\n",
      "Epoch 975/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8250 - accuracy: 0.5220 - val_loss: 2.2685 - val_accuracy: 0.4109\n",
      "Epoch 976/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8228 - accuracy: 0.5233 - val_loss: 2.2604 - val_accuracy: 0.4124\n",
      "Epoch 977/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8255 - accuracy: 0.5221 - val_loss: 2.2674 - val_accuracy: 0.4117\n",
      "Epoch 978/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8244 - accuracy: 0.5226 - val_loss: 2.2614 - val_accuracy: 0.4114\n",
      "Epoch 979/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8232 - accuracy: 0.5226 - val_loss: 2.2594 - val_accuracy: 0.4100\n",
      "Epoch 980/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8253 - accuracy: 0.5231 - val_loss: 2.2573 - val_accuracy: 0.4112\n",
      "Epoch 981/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8222 - accuracy: 0.5234 - val_loss: 2.2579 - val_accuracy: 0.4121\n",
      "Epoch 982/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8256 - accuracy: 0.5214 - val_loss: 2.2573 - val_accuracy: 0.4108\n",
      "Epoch 983/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8229 - accuracy: 0.5238 - val_loss: 2.2523 - val_accuracy: 0.4104\n",
      "Epoch 984/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8214 - accuracy: 0.5236 - val_loss: 2.2606 - val_accuracy: 0.4114\n",
      "Epoch 985/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8235 - accuracy: 0.5219 - val_loss: 2.2588 - val_accuracy: 0.4104\n",
      "Epoch 986/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8248 - accuracy: 0.5237 - val_loss: 2.2609 - val_accuracy: 0.4085\n",
      "Epoch 987/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8254 - accuracy: 0.5210 - val_loss: 2.2604 - val_accuracy: 0.4144\n",
      "Epoch 988/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8213 - accuracy: 0.5235 - val_loss: 2.2588 - val_accuracy: 0.4131\n",
      "Epoch 989/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8208 - accuracy: 0.5235 - val_loss: 2.2611 - val_accuracy: 0.4139\n",
      "Epoch 990/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8241 - accuracy: 0.5230 - val_loss: 2.2565 - val_accuracy: 0.4115\n",
      "Epoch 991/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8218 - accuracy: 0.5229 - val_loss: 2.2562 - val_accuracy: 0.4132\n",
      "Epoch 992/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8227 - accuracy: 0.5229 - val_loss: 2.2605 - val_accuracy: 0.4119\n",
      "Epoch 993/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8210 - accuracy: 0.5234 - val_loss: 2.2591 - val_accuracy: 0.4118\n",
      "Epoch 994/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8233 - accuracy: 0.5230 - val_loss: 2.2621 - val_accuracy: 0.4121\n",
      "Epoch 995/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8201 - accuracy: 0.5249 - val_loss: 2.2684 - val_accuracy: 0.4085\n",
      "Epoch 996/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8220 - accuracy: 0.5231 - val_loss: 2.2661 - val_accuracy: 0.4117\n",
      "Epoch 997/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8230 - accuracy: 0.5229 - val_loss: 2.2644 - val_accuracy: 0.4118\n",
      "Epoch 998/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8253 - accuracy: 0.5223 - val_loss: 2.2677 - val_accuracy: 0.4129\n",
      "Epoch 999/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8245 - accuracy: 0.5229 - val_loss: 2.2614 - val_accuracy: 0.4127\n",
      "Epoch 1000/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8210 - accuracy: 0.5243 - val_loss: 2.2615 - val_accuracy: 0.4143\n",
      "Epoch 1001/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8204 - accuracy: 0.5240 - val_loss: 2.2652 - val_accuracy: 0.4137\n",
      "Epoch 1002/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8213 - accuracy: 0.5234 - val_loss: 2.2711 - val_accuracy: 0.4104\n",
      "Epoch 1003/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8246 - accuracy: 0.5234 - val_loss: 2.2631 - val_accuracy: 0.4146\n",
      "Epoch 1004/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8228 - accuracy: 0.5232 - val_loss: 2.2641 - val_accuracy: 0.4121\n",
      "Epoch 1005/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8215 - accuracy: 0.5232 - val_loss: 2.2612 - val_accuracy: 0.4134\n",
      "Epoch 1006/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8231 - accuracy: 0.5235 - val_loss: 2.2587 - val_accuracy: 0.4127\n",
      "Epoch 1007/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8210 - accuracy: 0.5236 - val_loss: 2.2609 - val_accuracy: 0.4122\n",
      "Epoch 1008/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8200 - accuracy: 0.5233 - val_loss: 2.2651 - val_accuracy: 0.4100\n",
      "Epoch 1009/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8232 - accuracy: 0.5226 - val_loss: 2.2610 - val_accuracy: 0.4130\n",
      "Epoch 1010/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8241 - accuracy: 0.5229 - val_loss: 2.2643 - val_accuracy: 0.4094\n",
      "Epoch 1011/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8221 - accuracy: 0.5240 - val_loss: 2.2586 - val_accuracy: 0.4142\n",
      "Epoch 1012/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8213 - accuracy: 0.5234 - val_loss: 2.2616 - val_accuracy: 0.4159\n",
      "Epoch 1013/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8236 - accuracy: 0.5230 - val_loss: 2.2639 - val_accuracy: 0.4109\n",
      "Epoch 1014/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8223 - accuracy: 0.5241 - val_loss: 2.2650 - val_accuracy: 0.4108\n",
      "Epoch 1015/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8232 - accuracy: 0.5239 - val_loss: 2.2643 - val_accuracy: 0.4095\n",
      "Epoch 1016/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8210 - accuracy: 0.5233 - val_loss: 2.2639 - val_accuracy: 0.4107\n",
      "Epoch 1017/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8198 - accuracy: 0.5243 - val_loss: 2.2621 - val_accuracy: 0.4127\n",
      "Epoch 1018/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8203 - accuracy: 0.5244 - val_loss: 2.2619 - val_accuracy: 0.4120\n",
      "Epoch 1019/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8215 - accuracy: 0.5244 - val_loss: 2.2604 - val_accuracy: 0.4131\n",
      "Epoch 1020/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8196 - accuracy: 0.5244 - val_loss: 2.2624 - val_accuracy: 0.4095\n",
      "Epoch 1021/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8238 - accuracy: 0.5229 - val_loss: 2.2592 - val_accuracy: 0.4142\n",
      "Epoch 1022/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8221 - accuracy: 0.5238 - val_loss: 2.2618 - val_accuracy: 0.4120\n",
      "Epoch 1023/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8215 - accuracy: 0.5242 - val_loss: 2.2601 - val_accuracy: 0.4169\n",
      "Epoch 1024/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8239 - accuracy: 0.5235 - val_loss: 2.2611 - val_accuracy: 0.4127\n",
      "Epoch 1025/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8231 - accuracy: 0.5240 - val_loss: 2.2520 - val_accuracy: 0.4114\n",
      "Epoch 1026/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8195 - accuracy: 0.5234 - val_loss: 2.2599 - val_accuracy: 0.4134\n",
      "Epoch 1027/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8226 - accuracy: 0.5236 - val_loss: 2.2522 - val_accuracy: 0.4140\n",
      "Epoch 1028/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8215 - accuracy: 0.5237 - val_loss: 2.2594 - val_accuracy: 0.4111\n",
      "Epoch 1029/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8196 - accuracy: 0.5236 - val_loss: 2.2664 - val_accuracy: 0.4114\n",
      "Epoch 1030/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8217 - accuracy: 0.5243 - val_loss: 2.2590 - val_accuracy: 0.4141\n",
      "Epoch 1031/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8211 - accuracy: 0.5238 - val_loss: 2.2561 - val_accuracy: 0.4140\n",
      "Epoch 1032/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8214 - accuracy: 0.5242 - val_loss: 2.2662 - val_accuracy: 0.4107\n",
      "Epoch 1033/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8245 - accuracy: 0.5224 - val_loss: 2.2645 - val_accuracy: 0.4128\n",
      "Epoch 1034/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8219 - accuracy: 0.5247 - val_loss: 2.2612 - val_accuracy: 0.4109\n",
      "Epoch 1035/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8228 - accuracy: 0.5230 - val_loss: 2.2645 - val_accuracy: 0.4117\n",
      "Epoch 1036/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8235 - accuracy: 0.5234 - val_loss: 2.2625 - val_accuracy: 0.4157\n",
      "Epoch 1037/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8211 - accuracy: 0.5245 - val_loss: 2.2651 - val_accuracy: 0.4122\n",
      "Epoch 1038/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8212 - accuracy: 0.5230 - val_loss: 2.2622 - val_accuracy: 0.4137\n",
      "Epoch 1039/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8225 - accuracy: 0.5242 - val_loss: 2.2553 - val_accuracy: 0.4144\n",
      "Epoch 1040/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8220 - accuracy: 0.5239 - val_loss: 2.2620 - val_accuracy: 0.4143\n",
      "Epoch 1041/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8176 - accuracy: 0.5240 - val_loss: 2.2587 - val_accuracy: 0.4116\n",
      "Epoch 1042/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8211 - accuracy: 0.5232 - val_loss: 2.2604 - val_accuracy: 0.4125\n",
      "Epoch 1043/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8217 - accuracy: 0.5241 - val_loss: 2.2549 - val_accuracy: 0.4147\n",
      "Epoch 1044/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8215 - accuracy: 0.5238 - val_loss: 2.2631 - val_accuracy: 0.4115\n",
      "Epoch 1045/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8180 - accuracy: 0.5249 - val_loss: 2.2635 - val_accuracy: 0.4132\n",
      "Epoch 1046/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8209 - accuracy: 0.5244 - val_loss: 2.2591 - val_accuracy: 0.4115\n",
      "Epoch 1047/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8194 - accuracy: 0.5245 - val_loss: 2.2621 - val_accuracy: 0.4156\n",
      "Epoch 1048/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8223 - accuracy: 0.5238 - val_loss: 2.2574 - val_accuracy: 0.4137\n",
      "Epoch 1049/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8251 - accuracy: 0.5208 - val_loss: 2.2589 - val_accuracy: 0.4095\n",
      "Epoch 1050/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8192 - accuracy: 0.5253 - val_loss: 2.2601 - val_accuracy: 0.4138\n",
      "Epoch 1051/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8196 - accuracy: 0.5252 - val_loss: 2.2593 - val_accuracy: 0.4162\n",
      "Epoch 1052/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8199 - accuracy: 0.5237 - val_loss: 2.2580 - val_accuracy: 0.4131\n",
      "Epoch 1053/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8189 - accuracy: 0.5232 - val_loss: 2.2713 - val_accuracy: 0.4127\n",
      "Epoch 1054/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8232 - accuracy: 0.5239 - val_loss: 2.2589 - val_accuracy: 0.4171\n",
      "Epoch 1055/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8205 - accuracy: 0.5235 - val_loss: 2.2622 - val_accuracy: 0.4147\n",
      "Epoch 1056/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8202 - accuracy: 0.5241 - val_loss: 2.2649 - val_accuracy: 0.4133\n",
      "Epoch 1057/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8168 - accuracy: 0.5256 - val_loss: 2.2722 - val_accuracy: 0.4121\n",
      "Epoch 1058/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8196 - accuracy: 0.5252 - val_loss: 2.2695 - val_accuracy: 0.4122\n",
      "Epoch 1059/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8202 - accuracy: 0.5252 - val_loss: 2.2610 - val_accuracy: 0.4131\n",
      "Epoch 1060/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8198 - accuracy: 0.5248 - val_loss: 2.2584 - val_accuracy: 0.4146\n",
      "Epoch 1061/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8186 - accuracy: 0.5252 - val_loss: 2.2630 - val_accuracy: 0.4138\n",
      "Epoch 1062/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8202 - accuracy: 0.5242 - val_loss: 2.2632 - val_accuracy: 0.4157\n",
      "Epoch 1063/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8210 - accuracy: 0.5244 - val_loss: 2.2674 - val_accuracy: 0.4124\n",
      "Epoch 1064/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8213 - accuracy: 0.5234 - val_loss: 2.2622 - val_accuracy: 0.4154\n",
      "Epoch 1065/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8220 - accuracy: 0.5233 - val_loss: 2.2614 - val_accuracy: 0.4140\n",
      "Epoch 1066/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8217 - accuracy: 0.5244 - val_loss: 2.2686 - val_accuracy: 0.4134\n",
      "Epoch 1067/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8194 - accuracy: 0.5245 - val_loss: 2.2604 - val_accuracy: 0.4124\n",
      "Epoch 1068/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8169 - accuracy: 0.5261 - val_loss: 2.2637 - val_accuracy: 0.4139\n",
      "Epoch 1069/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8184 - accuracy: 0.5255 - val_loss: 2.2612 - val_accuracy: 0.4138\n",
      "Epoch 1070/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8189 - accuracy: 0.5255 - val_loss: 2.2645 - val_accuracy: 0.4151\n",
      "Epoch 1071/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8161 - accuracy: 0.5258 - val_loss: 2.2562 - val_accuracy: 0.4164\n",
      "Epoch 1072/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8197 - accuracy: 0.5238 - val_loss: 2.2634 - val_accuracy: 0.4136\n",
      "Epoch 1073/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8218 - accuracy: 0.5242 - val_loss: 2.2589 - val_accuracy: 0.4137\n",
      "Epoch 1074/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8185 - accuracy: 0.5254 - val_loss: 2.2640 - val_accuracy: 0.4118\n",
      "Epoch 1075/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8182 - accuracy: 0.5257 - val_loss: 2.2669 - val_accuracy: 0.4111\n",
      "Epoch 1076/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8166 - accuracy: 0.5259 - val_loss: 2.2564 - val_accuracy: 0.4160\n",
      "Epoch 1077/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8185 - accuracy: 0.5247 - val_loss: 2.2640 - val_accuracy: 0.4127\n",
      "Epoch 1078/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8227 - accuracy: 0.5235 - val_loss: 2.2621 - val_accuracy: 0.4130\n",
      "Epoch 1079/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8195 - accuracy: 0.5241 - val_loss: 2.2590 - val_accuracy: 0.4137\n",
      "Epoch 1080/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8184 - accuracy: 0.5250 - val_loss: 2.2529 - val_accuracy: 0.4123\n",
      "Epoch 1081/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8187 - accuracy: 0.5256 - val_loss: 2.2598 - val_accuracy: 0.4131\n",
      "Epoch 1082/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8192 - accuracy: 0.5246 - val_loss: 2.2591 - val_accuracy: 0.4107\n",
      "Epoch 1083/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8197 - accuracy: 0.5240 - val_loss: 2.2597 - val_accuracy: 0.4140\n",
      "Epoch 1084/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8184 - accuracy: 0.5244 - val_loss: 2.2560 - val_accuracy: 0.4162\n",
      "Epoch 1085/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8179 - accuracy: 0.5246 - val_loss: 2.2556 - val_accuracy: 0.4117\n",
      "Epoch 1086/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8191 - accuracy: 0.5246 - val_loss: 2.2576 - val_accuracy: 0.4114\n",
      "Epoch 1087/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8160 - accuracy: 0.5249 - val_loss: 2.2636 - val_accuracy: 0.4106\n",
      "Epoch 1088/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8190 - accuracy: 0.5251 - val_loss: 2.2646 - val_accuracy: 0.4120\n",
      "Epoch 1089/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8175 - accuracy: 0.5256 - val_loss: 2.2593 - val_accuracy: 0.4138\n",
      "Epoch 1090/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8212 - accuracy: 0.5248 - val_loss: 2.2563 - val_accuracy: 0.4123\n",
      "Epoch 1091/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8195 - accuracy: 0.5241 - val_loss: 2.2553 - val_accuracy: 0.4120\n",
      "Epoch 1092/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8207 - accuracy: 0.5241 - val_loss: 2.2613 - val_accuracy: 0.4128\n",
      "Epoch 1093/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8176 - accuracy: 0.5256 - val_loss: 2.2590 - val_accuracy: 0.4103\n",
      "Epoch 1094/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8180 - accuracy: 0.5248 - val_loss: 2.2590 - val_accuracy: 0.4133\n",
      "Epoch 1095/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8193 - accuracy: 0.5256 - val_loss: 2.2501 - val_accuracy: 0.4152\n",
      "Epoch 1096/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8201 - accuracy: 0.5250 - val_loss: 2.2581 - val_accuracy: 0.4119\n",
      "Epoch 1097/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8193 - accuracy: 0.5244 - val_loss: 2.2583 - val_accuracy: 0.4106\n",
      "Epoch 1098/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8195 - accuracy: 0.5246 - val_loss: 2.2549 - val_accuracy: 0.4176\n",
      "Epoch 1099/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8184 - accuracy: 0.5242 - val_loss: 2.2610 - val_accuracy: 0.4132\n",
      "Epoch 1100/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8200 - accuracy: 0.5250 - val_loss: 2.2627 - val_accuracy: 0.4144\n",
      "Epoch 1101/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8200 - accuracy: 0.5246 - val_loss: 2.2609 - val_accuracy: 0.4107\n",
      "Epoch 1102/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8180 - accuracy: 0.5250 - val_loss: 2.2677 - val_accuracy: 0.4152\n",
      "Epoch 1103/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8180 - accuracy: 0.5253 - val_loss: 2.2547 - val_accuracy: 0.4144\n",
      "Epoch 1104/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8168 - accuracy: 0.5254 - val_loss: 2.2614 - val_accuracy: 0.4145\n",
      "Epoch 1105/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8177 - accuracy: 0.5250 - val_loss: 2.2584 - val_accuracy: 0.4131\n",
      "Epoch 1106/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8194 - accuracy: 0.5237 - val_loss: 2.2666 - val_accuracy: 0.4117\n",
      "Epoch 1107/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8167 - accuracy: 0.5252 - val_loss: 2.2627 - val_accuracy: 0.4135\n",
      "Epoch 1108/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8164 - accuracy: 0.5254 - val_loss: 2.2629 - val_accuracy: 0.4117\n",
      "Epoch 1109/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8160 - accuracy: 0.5260 - val_loss: 2.2627 - val_accuracy: 0.4117\n",
      "Epoch 1110/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8225 - accuracy: 0.5240 - val_loss: 2.2668 - val_accuracy: 0.4133\n",
      "Epoch 1111/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8186 - accuracy: 0.5250 - val_loss: 2.2636 - val_accuracy: 0.4103\n",
      "Epoch 1112/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8215 - accuracy: 0.5246 - val_loss: 2.2688 - val_accuracy: 0.4112\n",
      "Epoch 1113/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8196 - accuracy: 0.5246 - val_loss: 2.2631 - val_accuracy: 0.4123\n",
      "Epoch 1114/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8153 - accuracy: 0.5261 - val_loss: 2.2718 - val_accuracy: 0.4116\n",
      "Epoch 1115/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8203 - accuracy: 0.5234 - val_loss: 2.2676 - val_accuracy: 0.4124\n",
      "Epoch 1116/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8160 - accuracy: 0.5257 - val_loss: 2.2724 - val_accuracy: 0.4119\n",
      "Epoch 1117/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8171 - accuracy: 0.5256 - val_loss: 2.2688 - val_accuracy: 0.4138\n",
      "Epoch 1118/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8159 - accuracy: 0.5261 - val_loss: 2.2656 - val_accuracy: 0.4135\n",
      "Epoch 1119/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8188 - accuracy: 0.5258 - val_loss: 2.2659 - val_accuracy: 0.4104\n",
      "Epoch 1120/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8212 - accuracy: 0.5238 - val_loss: 2.2634 - val_accuracy: 0.4094\n",
      "Epoch 1121/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8186 - accuracy: 0.5247 - val_loss: 2.2629 - val_accuracy: 0.4122\n",
      "Epoch 1122/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8170 - accuracy: 0.5244 - val_loss: 2.2626 - val_accuracy: 0.4120\n",
      "Epoch 1123/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8202 - accuracy: 0.5244 - val_loss: 2.2631 - val_accuracy: 0.4121\n",
      "Epoch 1124/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8165 - accuracy: 0.5260 - val_loss: 2.2600 - val_accuracy: 0.4153\n",
      "Epoch 1125/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8162 - accuracy: 0.5263 - val_loss: 2.2610 - val_accuracy: 0.4126\n",
      "Epoch 1126/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8181 - accuracy: 0.5253 - val_loss: 2.2622 - val_accuracy: 0.4104\n",
      "Epoch 1127/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8158 - accuracy: 0.5262 - val_loss: 2.2639 - val_accuracy: 0.4130\n",
      "Epoch 1128/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8134 - accuracy: 0.5267 - val_loss: 2.2723 - val_accuracy: 0.4098\n",
      "Epoch 1129/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8140 - accuracy: 0.5256 - val_loss: 2.2650 - val_accuracy: 0.4141\n",
      "Epoch 1130/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8191 - accuracy: 0.5251 - val_loss: 2.2577 - val_accuracy: 0.4131\n",
      "Epoch 1131/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8189 - accuracy: 0.5255 - val_loss: 2.2581 - val_accuracy: 0.4171\n",
      "Epoch 1132/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8171 - accuracy: 0.5259 - val_loss: 2.2652 - val_accuracy: 0.4168\n",
      "Epoch 1133/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8185 - accuracy: 0.5251 - val_loss: 2.2612 - val_accuracy: 0.4133\n",
      "Epoch 1134/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8185 - accuracy: 0.5245 - val_loss: 2.2602 - val_accuracy: 0.4152\n",
      "Epoch 1135/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8139 - accuracy: 0.5275 - val_loss: 2.2604 - val_accuracy: 0.4138\n",
      "Epoch 1136/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8152 - accuracy: 0.5264 - val_loss: 2.2684 - val_accuracy: 0.4114\n",
      "Epoch 1137/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8191 - accuracy: 0.5251 - val_loss: 2.2601 - val_accuracy: 0.4145\n",
      "Epoch 1138/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8190 - accuracy: 0.5249 - val_loss: 2.2602 - val_accuracy: 0.4118\n",
      "Epoch 1139/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8191 - accuracy: 0.5249 - val_loss: 2.2621 - val_accuracy: 0.4157\n",
      "Epoch 1140/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8203 - accuracy: 0.5247 - val_loss: 2.2664 - val_accuracy: 0.4158\n",
      "Epoch 1141/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8176 - accuracy: 0.5246 - val_loss: 2.2655 - val_accuracy: 0.4105\n",
      "Epoch 1142/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8145 - accuracy: 0.5274 - val_loss: 2.2685 - val_accuracy: 0.4141\n",
      "Epoch 1143/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8131 - accuracy: 0.5267 - val_loss: 2.2583 - val_accuracy: 0.4140\n",
      "Epoch 1144/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8181 - accuracy: 0.5263 - val_loss: 2.2670 - val_accuracy: 0.4119\n",
      "Epoch 1145/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8144 - accuracy: 0.5264 - val_loss: 2.2672 - val_accuracy: 0.4107\n",
      "Epoch 1146/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8152 - accuracy: 0.5267 - val_loss: 2.2661 - val_accuracy: 0.4123\n",
      "Epoch 1147/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8176 - accuracy: 0.5266 - val_loss: 2.2638 - val_accuracy: 0.4120\n",
      "Epoch 1148/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8200 - accuracy: 0.5242 - val_loss: 2.2622 - val_accuracy: 0.4134\n",
      "Epoch 1149/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8146 - accuracy: 0.5258 - val_loss: 2.2713 - val_accuracy: 0.4111\n",
      "Epoch 1150/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8174 - accuracy: 0.5252 - val_loss: 2.2660 - val_accuracy: 0.4130\n",
      "Epoch 1151/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8148 - accuracy: 0.5264 - val_loss: 2.2662 - val_accuracy: 0.4133\n",
      "Epoch 1152/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8201 - accuracy: 0.5233 - val_loss: 2.2646 - val_accuracy: 0.4117\n",
      "Epoch 1153/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8149 - accuracy: 0.5260 - val_loss: 2.2656 - val_accuracy: 0.4115\n",
      "Epoch 1154/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8185 - accuracy: 0.5253 - val_loss: 2.2627 - val_accuracy: 0.4135\n",
      "Epoch 1155/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8156 - accuracy: 0.5273 - val_loss: 2.2648 - val_accuracy: 0.4127\n",
      "Epoch 1156/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8194 - accuracy: 0.5248 - val_loss: 2.2665 - val_accuracy: 0.4088\n",
      "Epoch 1157/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8136 - accuracy: 0.5252 - val_loss: 2.2659 - val_accuracy: 0.4101\n",
      "Epoch 1158/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8151 - accuracy: 0.5267 - val_loss: 2.2605 - val_accuracy: 0.4120\n",
      "Epoch 1159/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8197 - accuracy: 0.5242 - val_loss: 2.2578 - val_accuracy: 0.4139\n",
      "Epoch 1160/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8182 - accuracy: 0.5252 - val_loss: 2.2560 - val_accuracy: 0.4149\n",
      "Epoch 1161/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8155 - accuracy: 0.5260 - val_loss: 2.2519 - val_accuracy: 0.4137\n",
      "Epoch 1162/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8163 - accuracy: 0.5258 - val_loss: 2.2612 - val_accuracy: 0.4107\n",
      "Epoch 1163/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8130 - accuracy: 0.5260 - val_loss: 2.2591 - val_accuracy: 0.4141\n",
      "Epoch 1164/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8183 - accuracy: 0.5265 - val_loss: 2.2637 - val_accuracy: 0.4123\n",
      "Epoch 1165/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8139 - accuracy: 0.5257 - val_loss: 2.2628 - val_accuracy: 0.4133\n",
      "Epoch 1166/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8168 - accuracy: 0.5252 - val_loss: 2.2625 - val_accuracy: 0.4142\n",
      "Epoch 1167/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8150 - accuracy: 0.5256 - val_loss: 2.2622 - val_accuracy: 0.4133\n",
      "Epoch 1168/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8136 - accuracy: 0.5266 - val_loss: 2.2642 - val_accuracy: 0.4128\n",
      "Epoch 1169/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8164 - accuracy: 0.5256 - val_loss: 2.2683 - val_accuracy: 0.4073\n",
      "Epoch 1170/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8158 - accuracy: 0.5265 - val_loss: 2.2696 - val_accuracy: 0.4112\n",
      "Epoch 1171/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8134 - accuracy: 0.5264 - val_loss: 2.2609 - val_accuracy: 0.4108\n",
      "Epoch 1172/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8170 - accuracy: 0.5261 - val_loss: 2.2591 - val_accuracy: 0.4135\n",
      "Epoch 1173/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8127 - accuracy: 0.5261 - val_loss: 2.2628 - val_accuracy: 0.4146\n",
      "Epoch 1174/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8156 - accuracy: 0.5267 - val_loss: 2.2640 - val_accuracy: 0.4113\n",
      "Epoch 1175/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8126 - accuracy: 0.5285 - val_loss: 2.2665 - val_accuracy: 0.4127\n",
      "Epoch 1176/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8142 - accuracy: 0.5263 - val_loss: 2.2744 - val_accuracy: 0.4127\n",
      "Epoch 1177/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8142 - accuracy: 0.5256 - val_loss: 2.2675 - val_accuracy: 0.4120\n",
      "Epoch 1178/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8133 - accuracy: 0.5258 - val_loss: 2.2662 - val_accuracy: 0.4101\n",
      "Epoch 1179/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8155 - accuracy: 0.5268 - val_loss: 2.2611 - val_accuracy: 0.4130\n",
      "Epoch 1180/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8162 - accuracy: 0.5265 - val_loss: 2.2660 - val_accuracy: 0.4133\n",
      "Epoch 1181/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8138 - accuracy: 0.5266 - val_loss: 2.2599 - val_accuracy: 0.4120\n",
      "Epoch 1182/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8144 - accuracy: 0.5258 - val_loss: 2.2645 - val_accuracy: 0.4085\n",
      "Epoch 1183/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8183 - accuracy: 0.5248 - val_loss: 2.2590 - val_accuracy: 0.4156\n",
      "Epoch 1184/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8155 - accuracy: 0.5262 - val_loss: 2.2620 - val_accuracy: 0.4138\n",
      "Epoch 1185/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8159 - accuracy: 0.5258 - val_loss: 2.2643 - val_accuracy: 0.4117\n",
      "Epoch 1186/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8150 - accuracy: 0.5260 - val_loss: 2.2545 - val_accuracy: 0.4137\n",
      "Epoch 1187/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8165 - accuracy: 0.5259 - val_loss: 2.2611 - val_accuracy: 0.4128\n",
      "Epoch 1188/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8174 - accuracy: 0.5256 - val_loss: 2.2610 - val_accuracy: 0.4070\n",
      "Epoch 1189/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8143 - accuracy: 0.5263 - val_loss: 2.2569 - val_accuracy: 0.4142\n",
      "Epoch 1190/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8154 - accuracy: 0.5264 - val_loss: 2.2600 - val_accuracy: 0.4114\n",
      "Epoch 1191/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8150 - accuracy: 0.5275 - val_loss: 2.2566 - val_accuracy: 0.4136\n",
      "Epoch 1192/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8200 - accuracy: 0.5246 - val_loss: 2.2651 - val_accuracy: 0.4160\n",
      "Epoch 1193/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8176 - accuracy: 0.5259 - val_loss: 2.2563 - val_accuracy: 0.4167\n",
      "Epoch 1194/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8145 - accuracy: 0.5267 - val_loss: 2.2607 - val_accuracy: 0.4127\n",
      "Epoch 1195/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8133 - accuracy: 0.5263 - val_loss: 2.2645 - val_accuracy: 0.4143\n",
      "Epoch 1196/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8134 - accuracy: 0.5274 - val_loss: 2.2680 - val_accuracy: 0.4114\n",
      "Epoch 1197/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8153 - accuracy: 0.5255 - val_loss: 2.2579 - val_accuracy: 0.4144\n",
      "Epoch 1198/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8143 - accuracy: 0.5278 - val_loss: 2.2604 - val_accuracy: 0.4144\n",
      "Epoch 1199/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8172 - accuracy: 0.5253 - val_loss: 2.2592 - val_accuracy: 0.4130\n",
      "Epoch 1200/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8168 - accuracy: 0.5246 - val_loss: 2.2596 - val_accuracy: 0.4147\n",
      "Epoch 1201/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8161 - accuracy: 0.5256 - val_loss: 2.2649 - val_accuracy: 0.4141\n",
      "Epoch 1202/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8186 - accuracy: 0.5251 - val_loss: 2.2585 - val_accuracy: 0.4167\n",
      "Epoch 1203/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8137 - accuracy: 0.5263 - val_loss: 2.2586 - val_accuracy: 0.4145\n",
      "Epoch 1204/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8152 - accuracy: 0.5260 - val_loss: 2.2559 - val_accuracy: 0.4160\n",
      "Epoch 1205/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8151 - accuracy: 0.5260 - val_loss: 2.2632 - val_accuracy: 0.4138\n",
      "Epoch 1206/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8134 - accuracy: 0.5273 - val_loss: 2.2522 - val_accuracy: 0.4149\n",
      "Epoch 1207/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8130 - accuracy: 0.5268 - val_loss: 2.2559 - val_accuracy: 0.4159\n",
      "Epoch 1208/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8162 - accuracy: 0.5269 - val_loss: 2.2570 - val_accuracy: 0.4129\n",
      "Epoch 1209/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8134 - accuracy: 0.5259 - val_loss: 2.2560 - val_accuracy: 0.4133\n",
      "Epoch 1210/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8157 - accuracy: 0.5265 - val_loss: 2.2580 - val_accuracy: 0.4145\n",
      "Epoch 1211/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8147 - accuracy: 0.5278 - val_loss: 2.2561 - val_accuracy: 0.4159\n",
      "Epoch 1212/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8135 - accuracy: 0.5266 - val_loss: 2.2567 - val_accuracy: 0.4153\n",
      "Epoch 1213/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8141 - accuracy: 0.5266 - val_loss: 2.2617 - val_accuracy: 0.4150\n",
      "Epoch 1214/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8131 - accuracy: 0.5278 - val_loss: 2.2565 - val_accuracy: 0.4155\n",
      "Epoch 1215/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8161 - accuracy: 0.5263 - val_loss: 2.2609 - val_accuracy: 0.4127\n",
      "Epoch 1216/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8157 - accuracy: 0.5260 - val_loss: 2.2531 - val_accuracy: 0.4162\n",
      "Epoch 1217/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8144 - accuracy: 0.5269 - val_loss: 2.2616 - val_accuracy: 0.4150\n",
      "Epoch 1218/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8180 - accuracy: 0.5247 - val_loss: 2.2555 - val_accuracy: 0.4165\n",
      "Epoch 1219/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8137 - accuracy: 0.5278 - val_loss: 2.2627 - val_accuracy: 0.4121\n",
      "Epoch 1220/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8147 - accuracy: 0.5266 - val_loss: 2.2513 - val_accuracy: 0.4180\n",
      "Epoch 1221/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8104 - accuracy: 0.5283 - val_loss: 2.2583 - val_accuracy: 0.4130\n",
      "Epoch 1222/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8119 - accuracy: 0.5276 - val_loss: 2.2578 - val_accuracy: 0.4142\n",
      "Epoch 1223/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8121 - accuracy: 0.5271 - val_loss: 2.2608 - val_accuracy: 0.4139\n",
      "Epoch 1224/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8135 - accuracy: 0.5262 - val_loss: 2.2583 - val_accuracy: 0.4143\n",
      "Epoch 1225/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8146 - accuracy: 0.5277 - val_loss: 2.2572 - val_accuracy: 0.4123\n",
      "Epoch 1226/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8146 - accuracy: 0.5276 - val_loss: 2.2589 - val_accuracy: 0.4125\n",
      "Epoch 1227/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8136 - accuracy: 0.5269 - val_loss: 2.2536 - val_accuracy: 0.4146\n",
      "Epoch 1228/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8153 - accuracy: 0.5263 - val_loss: 2.2586 - val_accuracy: 0.4139\n",
      "Epoch 1229/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8155 - accuracy: 0.5267 - val_loss: 2.2561 - val_accuracy: 0.4127\n",
      "Epoch 1230/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8134 - accuracy: 0.5274 - val_loss: 2.2550 - val_accuracy: 0.4143\n",
      "Epoch 1231/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8141 - accuracy: 0.5259 - val_loss: 2.2561 - val_accuracy: 0.4117\n",
      "Epoch 1232/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8129 - accuracy: 0.5282 - val_loss: 2.2509 - val_accuracy: 0.4136\n",
      "Epoch 1233/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8104 - accuracy: 0.5284 - val_loss: 2.2571 - val_accuracy: 0.4137\n",
      "Epoch 1234/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8164 - accuracy: 0.5261 - val_loss: 2.2569 - val_accuracy: 0.4133\n",
      "Epoch 1235/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8128 - accuracy: 0.5271 - val_loss: 2.2621 - val_accuracy: 0.4145\n",
      "Epoch 1236/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8164 - accuracy: 0.5266 - val_loss: 2.2585 - val_accuracy: 0.4114\n",
      "Epoch 1237/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8121 - accuracy: 0.5267 - val_loss: 2.2606 - val_accuracy: 0.4115\n",
      "Epoch 1238/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8143 - accuracy: 0.5262 - val_loss: 2.2615 - val_accuracy: 0.4115\n",
      "Epoch 1239/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8127 - accuracy: 0.5275 - val_loss: 2.2576 - val_accuracy: 0.4109\n",
      "Epoch 1240/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8127 - accuracy: 0.5278 - val_loss: 2.2628 - val_accuracy: 0.4131\n",
      "Epoch 1241/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8115 - accuracy: 0.5268 - val_loss: 2.2609 - val_accuracy: 0.4139\n",
      "Epoch 1242/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8125 - accuracy: 0.5277 - val_loss: 2.2691 - val_accuracy: 0.4121\n",
      "Epoch 1243/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8105 - accuracy: 0.5273 - val_loss: 2.2597 - val_accuracy: 0.4117\n",
      "Epoch 1244/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8157 - accuracy: 0.5261 - val_loss: 2.2621 - val_accuracy: 0.4149\n",
      "Epoch 1245/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8110 - accuracy: 0.5278 - val_loss: 2.2542 - val_accuracy: 0.4180\n",
      "Epoch 1246/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8130 - accuracy: 0.5282 - val_loss: 2.2609 - val_accuracy: 0.4154\n",
      "Epoch 1247/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8139 - accuracy: 0.5260 - val_loss: 2.2668 - val_accuracy: 0.4110\n",
      "Epoch 1248/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8108 - accuracy: 0.5284 - val_loss: 2.2608 - val_accuracy: 0.4140\n",
      "Epoch 1249/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8168 - accuracy: 0.5264 - val_loss: 2.2559 - val_accuracy: 0.4174\n",
      "Epoch 1250/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8124 - accuracy: 0.5267 - val_loss: 2.2593 - val_accuracy: 0.4136\n",
      "Epoch 1251/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8124 - accuracy: 0.5272 - val_loss: 2.2556 - val_accuracy: 0.4154\n",
      "Epoch 1252/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8145 - accuracy: 0.5256 - val_loss: 2.2560 - val_accuracy: 0.4143\n",
      "Epoch 1253/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8158 - accuracy: 0.5254 - val_loss: 2.2502 - val_accuracy: 0.4162\n",
      "Epoch 1254/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8119 - accuracy: 0.5275 - val_loss: 2.2518 - val_accuracy: 0.4166\n",
      "Epoch 1255/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8126 - accuracy: 0.5284 - val_loss: 2.2533 - val_accuracy: 0.4209\n",
      "Epoch 1256/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8117 - accuracy: 0.5274 - val_loss: 2.2603 - val_accuracy: 0.4149\n",
      "Epoch 1257/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8137 - accuracy: 0.5264 - val_loss: 2.2497 - val_accuracy: 0.4143\n",
      "Epoch 1258/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8119 - accuracy: 0.5280 - val_loss: 2.2593 - val_accuracy: 0.4095\n",
      "Epoch 1259/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8106 - accuracy: 0.5286 - val_loss: 2.2549 - val_accuracy: 0.4139\n",
      "Epoch 1260/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8140 - accuracy: 0.5267 - val_loss: 2.2562 - val_accuracy: 0.4147\n",
      "Epoch 1261/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8137 - accuracy: 0.5265 - val_loss: 2.2537 - val_accuracy: 0.4145\n",
      "Epoch 1262/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8140 - accuracy: 0.5265 - val_loss: 2.2566 - val_accuracy: 0.4152\n",
      "Epoch 1263/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8102 - accuracy: 0.5269 - val_loss: 2.2596 - val_accuracy: 0.4111\n",
      "Epoch 1264/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8119 - accuracy: 0.5277 - val_loss: 2.2647 - val_accuracy: 0.4120\n",
      "Epoch 1265/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8115 - accuracy: 0.5275 - val_loss: 2.2616 - val_accuracy: 0.4105\n",
      "Epoch 1266/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8118 - accuracy: 0.5277 - val_loss: 2.2570 - val_accuracy: 0.4171\n",
      "Epoch 1267/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8126 - accuracy: 0.5278 - val_loss: 2.2583 - val_accuracy: 0.4117\n",
      "Epoch 1268/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8127 - accuracy: 0.5280 - val_loss: 2.2547 - val_accuracy: 0.4149\n",
      "Epoch 1269/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8122 - accuracy: 0.5276 - val_loss: 2.2648 - val_accuracy: 0.4086\n",
      "Epoch 1270/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8138 - accuracy: 0.5278 - val_loss: 2.2584 - val_accuracy: 0.4128\n",
      "Epoch 1271/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8114 - accuracy: 0.5272 - val_loss: 2.2583 - val_accuracy: 0.4137\n",
      "Epoch 1272/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8099 - accuracy: 0.5290 - val_loss: 2.2613 - val_accuracy: 0.4136\n",
      "Epoch 1273/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8130 - accuracy: 0.5275 - val_loss: 2.2629 - val_accuracy: 0.4095\n",
      "Epoch 1274/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8109 - accuracy: 0.5272 - val_loss: 2.2653 - val_accuracy: 0.4104\n",
      "Epoch 1275/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8122 - accuracy: 0.5282 - val_loss: 2.2599 - val_accuracy: 0.4155\n",
      "Epoch 1276/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8127 - accuracy: 0.5279 - val_loss: 2.2667 - val_accuracy: 0.4121\n",
      "Epoch 1277/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8112 - accuracy: 0.5286 - val_loss: 2.2676 - val_accuracy: 0.4123\n",
      "Epoch 1278/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8093 - accuracy: 0.5289 - val_loss: 2.2654 - val_accuracy: 0.4085\n",
      "Epoch 1279/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8112 - accuracy: 0.5273 - val_loss: 2.2598 - val_accuracy: 0.4142\n",
      "Epoch 1280/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8117 - accuracy: 0.5268 - val_loss: 2.2697 - val_accuracy: 0.4090\n",
      "Epoch 1281/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8142 - accuracy: 0.5268 - val_loss: 2.2622 - val_accuracy: 0.4114\n",
      "Epoch 1282/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8105 - accuracy: 0.5279 - val_loss: 2.2607 - val_accuracy: 0.4148\n",
      "Epoch 1283/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8141 - accuracy: 0.5270 - val_loss: 2.2617 - val_accuracy: 0.4122\n",
      "Epoch 1284/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8135 - accuracy: 0.5267 - val_loss: 2.2621 - val_accuracy: 0.4114\n",
      "Epoch 1285/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8124 - accuracy: 0.5278 - val_loss: 2.2554 - val_accuracy: 0.4138\n",
      "Epoch 1286/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8127 - accuracy: 0.5275 - val_loss: 2.2520 - val_accuracy: 0.4138\n",
      "Epoch 1287/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8114 - accuracy: 0.5280 - val_loss: 2.2599 - val_accuracy: 0.4138\n",
      "Epoch 1288/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8099 - accuracy: 0.5285 - val_loss: 2.2586 - val_accuracy: 0.4122\n",
      "Epoch 1289/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8130 - accuracy: 0.5281 - val_loss: 2.2627 - val_accuracy: 0.4094\n",
      "Epoch 1290/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8125 - accuracy: 0.5276 - val_loss: 2.2684 - val_accuracy: 0.4133\n",
      "Epoch 1291/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8103 - accuracy: 0.5275 - val_loss: 2.2655 - val_accuracy: 0.4103\n",
      "Epoch 1292/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8135 - accuracy: 0.5269 - val_loss: 2.2639 - val_accuracy: 0.4126\n",
      "Epoch 1293/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8115 - accuracy: 0.5274 - val_loss: 2.2571 - val_accuracy: 0.4100\n",
      "Epoch 1294/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8115 - accuracy: 0.5278 - val_loss: 2.2696 - val_accuracy: 0.4090\n",
      "Epoch 1295/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8119 - accuracy: 0.5276 - val_loss: 2.2603 - val_accuracy: 0.4104\n",
      "Epoch 1296/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8129 - accuracy: 0.5266 - val_loss: 2.2619 - val_accuracy: 0.4161\n",
      "Epoch 1297/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8114 - accuracy: 0.5278 - val_loss: 2.2629 - val_accuracy: 0.4114\n",
      "Epoch 1298/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8060 - accuracy: 0.5287 - val_loss: 2.2662 - val_accuracy: 0.4110\n",
      "Epoch 1299/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8131 - accuracy: 0.5270 - val_loss: 2.2625 - val_accuracy: 0.4093\n",
      "Epoch 1300/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8070 - accuracy: 0.5308 - val_loss: 2.2621 - val_accuracy: 0.4117\n",
      "Epoch 1301/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8106 - accuracy: 0.5278 - val_loss: 2.2628 - val_accuracy: 0.4158\n",
      "Epoch 1302/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8128 - accuracy: 0.5269 - val_loss: 2.2643 - val_accuracy: 0.4130\n",
      "Epoch 1303/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8120 - accuracy: 0.5277 - val_loss: 2.2643 - val_accuracy: 0.4142\n",
      "Epoch 1304/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8096 - accuracy: 0.5279 - val_loss: 2.2691 - val_accuracy: 0.4129\n",
      "Epoch 1305/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8129 - accuracy: 0.5275 - val_loss: 2.2658 - val_accuracy: 0.4097\n",
      "Epoch 1306/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8114 - accuracy: 0.5283 - val_loss: 2.2679 - val_accuracy: 0.4122\n",
      "Epoch 1307/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8111 - accuracy: 0.5279 - val_loss: 2.2688 - val_accuracy: 0.4146\n",
      "Epoch 1308/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8109 - accuracy: 0.5280 - val_loss: 2.2647 - val_accuracy: 0.4153\n",
      "Epoch 1309/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8117 - accuracy: 0.5280 - val_loss: 2.2637 - val_accuracy: 0.4115\n",
      "Epoch 1310/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8096 - accuracy: 0.5283 - val_loss: 2.2668 - val_accuracy: 0.4141\n",
      "Epoch 1311/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8105 - accuracy: 0.5276 - val_loss: 2.2695 - val_accuracy: 0.4127\n",
      "Epoch 1312/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8082 - accuracy: 0.5291 - val_loss: 2.2725 - val_accuracy: 0.4122\n",
      "Epoch 1313/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8128 - accuracy: 0.5269 - val_loss: 2.2662 - val_accuracy: 0.4149\n",
      "Epoch 1314/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8116 - accuracy: 0.5277 - val_loss: 2.2673 - val_accuracy: 0.4125\n",
      "Epoch 1315/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8084 - accuracy: 0.5291 - val_loss: 2.2633 - val_accuracy: 0.4101\n",
      "Epoch 1316/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8103 - accuracy: 0.5293 - val_loss: 2.2664 - val_accuracy: 0.4116\n",
      "Epoch 1317/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8113 - accuracy: 0.5269 - val_loss: 2.2673 - val_accuracy: 0.4114\n",
      "Epoch 1318/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8060 - accuracy: 0.5299 - val_loss: 2.2665 - val_accuracy: 0.4083\n",
      "Epoch 1319/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8099 - accuracy: 0.5289 - val_loss: 2.2686 - val_accuracy: 0.4091\n",
      "Epoch 1320/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8124 - accuracy: 0.5277 - val_loss: 2.2613 - val_accuracy: 0.4132\n",
      "Epoch 1321/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8110 - accuracy: 0.5283 - val_loss: 2.2619 - val_accuracy: 0.4102\n",
      "Epoch 1322/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8111 - accuracy: 0.5266 - val_loss: 2.2659 - val_accuracy: 0.4117\n",
      "Epoch 1323/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8106 - accuracy: 0.5281 - val_loss: 2.2696 - val_accuracy: 0.4077\n",
      "Epoch 1324/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8124 - accuracy: 0.5278 - val_loss: 2.2611 - val_accuracy: 0.4116\n",
      "Epoch 1325/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8097 - accuracy: 0.5287 - val_loss: 2.2684 - val_accuracy: 0.4108\n",
      "Epoch 1326/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8113 - accuracy: 0.5275 - val_loss: 2.2653 - val_accuracy: 0.4153\n",
      "Epoch 1327/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8109 - accuracy: 0.5274 - val_loss: 2.2545 - val_accuracy: 0.4159\n",
      "Epoch 1328/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8089 - accuracy: 0.5281 - val_loss: 2.2600 - val_accuracy: 0.4130\n",
      "Epoch 1329/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8104 - accuracy: 0.5280 - val_loss: 2.2681 - val_accuracy: 0.4090\n",
      "Epoch 1330/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8083 - accuracy: 0.5290 - val_loss: 2.2674 - val_accuracy: 0.4134\n",
      "Epoch 1331/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8101 - accuracy: 0.5293 - val_loss: 2.2612 - val_accuracy: 0.4116\n",
      "Epoch 1332/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8118 - accuracy: 0.5284 - val_loss: 2.2629 - val_accuracy: 0.4092\n",
      "Epoch 1333/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8103 - accuracy: 0.5275 - val_loss: 2.2692 - val_accuracy: 0.4086\n",
      "Epoch 1334/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8097 - accuracy: 0.5290 - val_loss: 2.2587 - val_accuracy: 0.4098\n",
      "Epoch 1335/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8084 - accuracy: 0.5289 - val_loss: 2.2585 - val_accuracy: 0.4133\n",
      "Epoch 1336/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8077 - accuracy: 0.5291 - val_loss: 2.2682 - val_accuracy: 0.4093\n",
      "Epoch 1337/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8102 - accuracy: 0.5302 - val_loss: 2.2654 - val_accuracy: 0.4122\n",
      "Epoch 1338/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5294 - val_loss: 2.2688 - val_accuracy: 0.4110\n",
      "Epoch 1339/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8115 - accuracy: 0.5281 - val_loss: 2.2674 - val_accuracy: 0.4092\n",
      "Epoch 1340/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8082 - accuracy: 0.5284 - val_loss: 2.2655 - val_accuracy: 0.4112\n",
      "Epoch 1341/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8097 - accuracy: 0.5290 - val_loss: 2.2625 - val_accuracy: 0.4133\n",
      "Epoch 1342/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8103 - accuracy: 0.5277 - val_loss: 2.2599 - val_accuracy: 0.4147\n",
      "Epoch 1343/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8106 - accuracy: 0.5289 - val_loss: 2.2619 - val_accuracy: 0.4144\n",
      "Epoch 1344/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8095 - accuracy: 0.5282 - val_loss: 2.2672 - val_accuracy: 0.4125\n",
      "Epoch 1345/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8078 - accuracy: 0.5292 - val_loss: 2.2637 - val_accuracy: 0.4150\n",
      "Epoch 1346/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8084 - accuracy: 0.5290 - val_loss: 2.2679 - val_accuracy: 0.4134\n",
      "Epoch 1347/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8095 - accuracy: 0.5277 - val_loss: 2.2664 - val_accuracy: 0.4127\n",
      "Epoch 1348/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8086 - accuracy: 0.5290 - val_loss: 2.2703 - val_accuracy: 0.4128\n",
      "Epoch 1349/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8086 - accuracy: 0.5295 - val_loss: 2.2643 - val_accuracy: 0.4111\n",
      "Epoch 1350/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8100 - accuracy: 0.5286 - val_loss: 2.2713 - val_accuracy: 0.4093\n",
      "Epoch 1351/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8071 - accuracy: 0.5296 - val_loss: 2.2700 - val_accuracy: 0.4082\n",
      "Epoch 1352/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8087 - accuracy: 0.5281 - val_loss: 2.2659 - val_accuracy: 0.4126\n",
      "Epoch 1353/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5304 - val_loss: 2.2673 - val_accuracy: 0.4110\n",
      "Epoch 1354/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8069 - accuracy: 0.5291 - val_loss: 2.2616 - val_accuracy: 0.4120\n",
      "Epoch 1355/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8054 - accuracy: 0.5291 - val_loss: 2.2693 - val_accuracy: 0.4109\n",
      "Epoch 1356/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8095 - accuracy: 0.5279 - val_loss: 2.2660 - val_accuracy: 0.4126\n",
      "Epoch 1357/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8098 - accuracy: 0.5287 - val_loss: 2.2604 - val_accuracy: 0.4131\n",
      "Epoch 1358/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8097 - accuracy: 0.5288 - val_loss: 2.2661 - val_accuracy: 0.4131\n",
      "Epoch 1359/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8079 - accuracy: 0.5298 - val_loss: 2.2678 - val_accuracy: 0.4133\n",
      "Epoch 1360/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8078 - accuracy: 0.5295 - val_loss: 2.2651 - val_accuracy: 0.4119\n",
      "Epoch 1361/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8098 - accuracy: 0.5292 - val_loss: 2.2697 - val_accuracy: 0.4101\n",
      "Epoch 1362/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8088 - accuracy: 0.5300 - val_loss: 2.2600 - val_accuracy: 0.4139\n",
      "Epoch 1363/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8097 - accuracy: 0.5285 - val_loss: 2.2671 - val_accuracy: 0.4127\n",
      "Epoch 1364/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8125 - accuracy: 0.5282 - val_loss: 2.2569 - val_accuracy: 0.4140\n",
      "Epoch 1365/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5291 - val_loss: 2.2639 - val_accuracy: 0.4116\n",
      "Epoch 1366/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8075 - accuracy: 0.5297 - val_loss: 2.2625 - val_accuracy: 0.4107\n",
      "Epoch 1367/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8060 - accuracy: 0.5302 - val_loss: 2.2664 - val_accuracy: 0.4119\n",
      "Epoch 1368/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8077 - accuracy: 0.5295 - val_loss: 2.2619 - val_accuracy: 0.4136\n",
      "Epoch 1369/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8081 - accuracy: 0.5291 - val_loss: 2.2670 - val_accuracy: 0.4109\n",
      "Epoch 1370/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8073 - accuracy: 0.5294 - val_loss: 2.2680 - val_accuracy: 0.4115\n",
      "Epoch 1371/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8097 - accuracy: 0.5294 - val_loss: 2.2606 - val_accuracy: 0.4106\n",
      "Epoch 1372/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8051 - accuracy: 0.5298 - val_loss: 2.2691 - val_accuracy: 0.4127\n",
      "Epoch 1373/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8072 - accuracy: 0.5295 - val_loss: 2.2563 - val_accuracy: 0.4142\n",
      "Epoch 1374/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8076 - accuracy: 0.5280 - val_loss: 2.2710 - val_accuracy: 0.4100\n",
      "Epoch 1375/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8043 - accuracy: 0.5293 - val_loss: 2.2626 - val_accuracy: 0.4139\n",
      "Epoch 1376/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8121 - accuracy: 0.5283 - val_loss: 2.2673 - val_accuracy: 0.4138\n",
      "Epoch 1377/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8087 - accuracy: 0.5284 - val_loss: 2.2663 - val_accuracy: 0.4126\n",
      "Epoch 1378/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8110 - accuracy: 0.5288 - val_loss: 2.2630 - val_accuracy: 0.4123\n",
      "Epoch 1379/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8084 - accuracy: 0.5278 - val_loss: 2.2680 - val_accuracy: 0.4137\n",
      "Epoch 1380/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8073 - accuracy: 0.5290 - val_loss: 2.2695 - val_accuracy: 0.4119\n",
      "Epoch 1381/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8099 - accuracy: 0.5292 - val_loss: 2.2683 - val_accuracy: 0.4092\n",
      "Epoch 1382/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8076 - accuracy: 0.5299 - val_loss: 2.2651 - val_accuracy: 0.4093\n",
      "Epoch 1383/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8082 - accuracy: 0.5285 - val_loss: 2.2752 - val_accuracy: 0.4114\n",
      "Epoch 1384/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8124 - accuracy: 0.5279 - val_loss: 2.2672 - val_accuracy: 0.4123\n",
      "Epoch 1385/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8069 - accuracy: 0.5293 - val_loss: 2.2581 - val_accuracy: 0.4140\n",
      "Epoch 1386/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8065 - accuracy: 0.5301 - val_loss: 2.2640 - val_accuracy: 0.4128\n",
      "Epoch 1387/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8075 - accuracy: 0.5291 - val_loss: 2.2719 - val_accuracy: 0.4104\n",
      "Epoch 1388/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8068 - accuracy: 0.5290 - val_loss: 2.2615 - val_accuracy: 0.4139\n",
      "Epoch 1389/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8064 - accuracy: 0.5295 - val_loss: 2.2717 - val_accuracy: 0.4116\n",
      "Epoch 1390/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8094 - accuracy: 0.5288 - val_loss: 2.2557 - val_accuracy: 0.4113\n",
      "Epoch 1391/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8104 - accuracy: 0.5283 - val_loss: 2.2689 - val_accuracy: 0.4104\n",
      "Epoch 1392/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8117 - accuracy: 0.5279 - val_loss: 2.2577 - val_accuracy: 0.4112\n",
      "Epoch 1393/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8066 - accuracy: 0.5295 - val_loss: 2.2679 - val_accuracy: 0.4121\n",
      "Epoch 1394/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8116 - accuracy: 0.5278 - val_loss: 2.2645 - val_accuracy: 0.4104\n",
      "Epoch 1395/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8081 - accuracy: 0.5293 - val_loss: 2.2639 - val_accuracy: 0.4099\n",
      "Epoch 1396/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8071 - accuracy: 0.5302 - val_loss: 2.2676 - val_accuracy: 0.4144\n",
      "Epoch 1397/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8076 - accuracy: 0.5295 - val_loss: 2.2595 - val_accuracy: 0.4142\n",
      "Epoch 1398/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8069 - accuracy: 0.5302 - val_loss: 2.2635 - val_accuracy: 0.4155\n",
      "Epoch 1399/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8086 - accuracy: 0.5286 - val_loss: 2.2648 - val_accuracy: 0.4123\n",
      "Epoch 1400/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8093 - accuracy: 0.5296 - val_loss: 2.2596 - val_accuracy: 0.4124\n",
      "Epoch 1401/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8109 - accuracy: 0.5285 - val_loss: 2.2628 - val_accuracy: 0.4132\n",
      "Epoch 1402/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8075 - accuracy: 0.5292 - val_loss: 2.2592 - val_accuracy: 0.4127\n",
      "Epoch 1403/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8075 - accuracy: 0.5300 - val_loss: 2.2633 - val_accuracy: 0.4117\n",
      "Epoch 1404/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8061 - accuracy: 0.5292 - val_loss: 2.2643 - val_accuracy: 0.4125\n",
      "Epoch 1405/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8088 - accuracy: 0.5294 - val_loss: 2.2631 - val_accuracy: 0.4127\n",
      "Epoch 1406/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5305 - val_loss: 2.2607 - val_accuracy: 0.4112\n",
      "Epoch 1407/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8096 - accuracy: 0.5282 - val_loss: 2.2605 - val_accuracy: 0.4109\n",
      "Epoch 1408/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8096 - accuracy: 0.5295 - val_loss: 2.2603 - val_accuracy: 0.4117\n",
      "Epoch 1409/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8101 - accuracy: 0.5289 - val_loss: 2.2615 - val_accuracy: 0.4155\n",
      "Epoch 1410/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8106 - accuracy: 0.5277 - val_loss: 2.2646 - val_accuracy: 0.4114\n",
      "Epoch 1411/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8084 - accuracy: 0.5299 - val_loss: 2.2559 - val_accuracy: 0.4138\n",
      "Epoch 1412/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8089 - accuracy: 0.5285 - val_loss: 2.2510 - val_accuracy: 0.4137\n",
      "Epoch 1413/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8087 - accuracy: 0.5287 - val_loss: 2.2577 - val_accuracy: 0.4132\n",
      "Epoch 1414/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8099 - accuracy: 0.5280 - val_loss: 2.2584 - val_accuracy: 0.4142\n",
      "Epoch 1415/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8061 - accuracy: 0.5298 - val_loss: 2.2574 - val_accuracy: 0.4133\n",
      "Epoch 1416/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8084 - accuracy: 0.5285 - val_loss: 2.2590 - val_accuracy: 0.4131\n",
      "Epoch 1417/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8032 - accuracy: 0.5308 - val_loss: 2.2638 - val_accuracy: 0.4150\n",
      "Epoch 1418/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8100 - accuracy: 0.5291 - val_loss: 2.2599 - val_accuracy: 0.4105\n",
      "Epoch 1419/2000\n",
      "242/242 [==============================] - 5s 23ms/step - loss: 1.8076 - accuracy: 0.5288 - val_loss: 2.2650 - val_accuracy: 0.4135\n",
      "Epoch 1420/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5301 - val_loss: 2.2543 - val_accuracy: 0.4149\n",
      "Epoch 1421/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8082 - accuracy: 0.5287 - val_loss: 2.2640 - val_accuracy: 0.4095\n",
      "Epoch 1422/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8080 - accuracy: 0.5289 - val_loss: 2.2636 - val_accuracy: 0.4123\n",
      "Epoch 1423/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8086 - accuracy: 0.5293 - val_loss: 2.2579 - val_accuracy: 0.4117\n",
      "Epoch 1424/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8076 - accuracy: 0.5292 - val_loss: 2.2664 - val_accuracy: 0.4095\n",
      "Epoch 1425/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8070 - accuracy: 0.5293 - val_loss: 2.2640 - val_accuracy: 0.4107\n",
      "Epoch 1426/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8032 - accuracy: 0.5299 - val_loss: 2.2605 - val_accuracy: 0.4107\n",
      "Epoch 1427/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8055 - accuracy: 0.5300 - val_loss: 2.2621 - val_accuracy: 0.4125\n",
      "Epoch 1428/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8056 - accuracy: 0.5305 - val_loss: 2.2668 - val_accuracy: 0.4075\n",
      "Epoch 1429/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5286 - val_loss: 2.2571 - val_accuracy: 0.4110\n",
      "Epoch 1430/2000\n",
      "242/242 [==============================] - 7s 27ms/step - loss: 1.8071 - accuracy: 0.5296 - val_loss: 2.2592 - val_accuracy: 0.4130\n",
      "Epoch 1431/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8070 - accuracy: 0.5300 - val_loss: 2.2540 - val_accuracy: 0.4123\n",
      "Epoch 1432/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8081 - accuracy: 0.5294 - val_loss: 2.2586 - val_accuracy: 0.4149\n",
      "Epoch 1433/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8069 - accuracy: 0.5302 - val_loss: 2.2598 - val_accuracy: 0.4171\n",
      "Epoch 1434/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8065 - accuracy: 0.5300 - val_loss: 2.2575 - val_accuracy: 0.4137\n",
      "Epoch 1435/2000\n",
      "242/242 [==============================] - 6s 24ms/step - loss: 1.8076 - accuracy: 0.5290 - val_loss: 2.2597 - val_accuracy: 0.4127\n",
      "Epoch 1436/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8086 - accuracy: 0.5297 - val_loss: 2.2617 - val_accuracy: 0.4130\n",
      "Epoch 1437/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8051 - accuracy: 0.5299 - val_loss: 2.2553 - val_accuracy: 0.4138\n",
      "Epoch 1438/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8059 - accuracy: 0.5294 - val_loss: 2.2611 - val_accuracy: 0.4155\n",
      "Epoch 1439/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8099 - accuracy: 0.5296 - val_loss: 2.2629 - val_accuracy: 0.4126\n",
      "Epoch 1440/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8072 - accuracy: 0.5293 - val_loss: 2.2627 - val_accuracy: 0.4133\n",
      "Epoch 1441/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8099 - accuracy: 0.5286 - val_loss: 2.2557 - val_accuracy: 0.4146\n",
      "Epoch 1442/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8061 - accuracy: 0.5305 - val_loss: 2.2640 - val_accuracy: 0.4116\n",
      "Epoch 1443/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8069 - accuracy: 0.5295 - val_loss: 2.2608 - val_accuracy: 0.4147\n",
      "Epoch 1444/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8067 - accuracy: 0.5295 - val_loss: 2.2629 - val_accuracy: 0.4110\n",
      "Epoch 1445/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8047 - accuracy: 0.5310 - val_loss: 2.2663 - val_accuracy: 0.4133\n",
      "Epoch 1446/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8076 - accuracy: 0.5296 - val_loss: 2.2708 - val_accuracy: 0.4075\n",
      "Epoch 1447/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8065 - accuracy: 0.5298 - val_loss: 2.2644 - val_accuracy: 0.4118\n",
      "Epoch 1448/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8071 - accuracy: 0.5292 - val_loss: 2.2628 - val_accuracy: 0.4102\n",
      "Epoch 1449/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8072 - accuracy: 0.5297 - val_loss: 2.2667 - val_accuracy: 0.4158\n",
      "Epoch 1450/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8071 - accuracy: 0.5290 - val_loss: 2.2648 - val_accuracy: 0.4117\n",
      "Epoch 1451/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8033 - accuracy: 0.5296 - val_loss: 2.2629 - val_accuracy: 0.4116\n",
      "Epoch 1452/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8058 - accuracy: 0.5304 - val_loss: 2.2600 - val_accuracy: 0.4142\n",
      "Epoch 1453/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5299 - val_loss: 2.2666 - val_accuracy: 0.4089\n",
      "Epoch 1454/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8056 - accuracy: 0.5299 - val_loss: 2.2663 - val_accuracy: 0.4089\n",
      "Epoch 1455/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8076 - accuracy: 0.5296 - val_loss: 2.2586 - val_accuracy: 0.4133\n",
      "Epoch 1456/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8051 - accuracy: 0.5300 - val_loss: 2.2664 - val_accuracy: 0.4095\n",
      "Epoch 1457/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8060 - accuracy: 0.5297 - val_loss: 2.2659 - val_accuracy: 0.4120\n",
      "Epoch 1458/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8078 - accuracy: 0.5303 - val_loss: 2.2661 - val_accuracy: 0.4106\n",
      "Epoch 1459/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8078 - accuracy: 0.5299 - val_loss: 2.2593 - val_accuracy: 0.4123\n",
      "Epoch 1460/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8061 - accuracy: 0.5296 - val_loss: 2.2607 - val_accuracy: 0.4120\n",
      "Epoch 1461/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8061 - accuracy: 0.5305 - val_loss: 2.2707 - val_accuracy: 0.4086\n",
      "Epoch 1462/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8070 - accuracy: 0.5295 - val_loss: 2.2586 - val_accuracy: 0.4139\n",
      "Epoch 1463/2000\n",
      "242/242 [==============================] - 6s 23ms/step - loss: 1.8067 - accuracy: 0.5298 - val_loss: 2.2595 - val_accuracy: 0.4149\n",
      "Epoch 1464/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8069 - accuracy: 0.5303 - val_loss: 2.2623 - val_accuracy: 0.4115\n",
      "Epoch 1465/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8059 - accuracy: 0.5301 - val_loss: 2.2635 - val_accuracy: 0.4113\n",
      "Epoch 1466/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8059 - accuracy: 0.5300 - val_loss: 2.2597 - val_accuracy: 0.4126\n",
      "Epoch 1467/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8053 - accuracy: 0.5301 - val_loss: 2.2612 - val_accuracy: 0.4107\n",
      "Epoch 1468/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8083 - accuracy: 0.5307 - val_loss: 2.2590 - val_accuracy: 0.4127\n",
      "Epoch 1469/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8075 - accuracy: 0.5304 - val_loss: 2.2559 - val_accuracy: 0.4130\n",
      "Epoch 1470/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8108 - accuracy: 0.5291 - val_loss: 2.2600 - val_accuracy: 0.4126\n",
      "Epoch 1471/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8058 - accuracy: 0.5309 - val_loss: 2.2578 - val_accuracy: 0.4139\n",
      "Epoch 1472/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8097 - accuracy: 0.5288 - val_loss: 2.2600 - val_accuracy: 0.4131\n",
      "Epoch 1473/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8070 - accuracy: 0.5292 - val_loss: 2.2593 - val_accuracy: 0.4084\n",
      "Epoch 1474/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8079 - accuracy: 0.5287 - val_loss: 2.2532 - val_accuracy: 0.4117\n",
      "Epoch 1475/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8067 - accuracy: 0.5293 - val_loss: 2.2617 - val_accuracy: 0.4095\n",
      "Epoch 1476/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8040 - accuracy: 0.5302 - val_loss: 2.2573 - val_accuracy: 0.4129\n",
      "Epoch 1477/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8055 - accuracy: 0.5291 - val_loss: 2.2629 - val_accuracy: 0.4132\n",
      "Epoch 1478/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8052 - accuracy: 0.5306 - val_loss: 2.2561 - val_accuracy: 0.4126\n",
      "Epoch 1479/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8049 - accuracy: 0.5314 - val_loss: 2.2631 - val_accuracy: 0.4094\n",
      "Epoch 1480/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8045 - accuracy: 0.5315 - val_loss: 2.2613 - val_accuracy: 0.4127\n",
      "Epoch 1481/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8068 - accuracy: 0.5292 - val_loss: 2.2605 - val_accuracy: 0.4128\n",
      "Epoch 1482/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8058 - accuracy: 0.5308 - val_loss: 2.2537 - val_accuracy: 0.4135\n",
      "Epoch 1483/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8026 - accuracy: 0.5319 - val_loss: 2.2652 - val_accuracy: 0.4123\n",
      "Epoch 1484/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8074 - accuracy: 0.5297 - val_loss: 2.2668 - val_accuracy: 0.4069\n",
      "Epoch 1485/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8062 - accuracy: 0.5303 - val_loss: 2.2637 - val_accuracy: 0.4118\n",
      "Epoch 1486/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8044 - accuracy: 0.5305 - val_loss: 2.2650 - val_accuracy: 0.4097\n",
      "Epoch 1487/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8040 - accuracy: 0.5308 - val_loss: 2.2654 - val_accuracy: 0.4134\n",
      "Epoch 1488/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8039 - accuracy: 0.5304 - val_loss: 2.2596 - val_accuracy: 0.4119\n",
      "Epoch 1489/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8041 - accuracy: 0.5306 - val_loss: 2.2617 - val_accuracy: 0.4098\n",
      "Epoch 1490/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8052 - accuracy: 0.5300 - val_loss: 2.2578 - val_accuracy: 0.4139\n",
      "Epoch 1491/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8040 - accuracy: 0.5304 - val_loss: 2.2571 - val_accuracy: 0.4158\n",
      "Epoch 1492/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8064 - accuracy: 0.5287 - val_loss: 2.2610 - val_accuracy: 0.4117\n",
      "Epoch 1493/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8068 - accuracy: 0.5301 - val_loss: 2.2545 - val_accuracy: 0.4133\n",
      "Epoch 1494/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8038 - accuracy: 0.5307 - val_loss: 2.2607 - val_accuracy: 0.4114\n",
      "Epoch 1495/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8039 - accuracy: 0.5307 - val_loss: 2.2640 - val_accuracy: 0.4098\n",
      "Epoch 1496/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8051 - accuracy: 0.5309 - val_loss: 2.2600 - val_accuracy: 0.4140\n",
      "Epoch 1497/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8076 - accuracy: 0.5298 - val_loss: 2.2615 - val_accuracy: 0.4128\n",
      "Epoch 1498/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8064 - accuracy: 0.5293 - val_loss: 2.2647 - val_accuracy: 0.4092\n",
      "Epoch 1499/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8034 - accuracy: 0.5317 - val_loss: 2.2618 - val_accuracy: 0.4098\n",
      "Epoch 1500/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8085 - accuracy: 0.5301 - val_loss: 2.2689 - val_accuracy: 0.4113\n",
      "Epoch 1501/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8087 - accuracy: 0.5300 - val_loss: 2.2626 - val_accuracy: 0.4100\n",
      "Epoch 1502/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8046 - accuracy: 0.5306 - val_loss: 2.2596 - val_accuracy: 0.4107\n",
      "Epoch 1503/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8049 - accuracy: 0.5305 - val_loss: 2.2635 - val_accuracy: 0.4104\n",
      "Epoch 1504/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8037 - accuracy: 0.5302 - val_loss: 2.2592 - val_accuracy: 0.4128\n",
      "Epoch 1505/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8044 - accuracy: 0.5308 - val_loss: 2.2605 - val_accuracy: 0.4117\n",
      "Epoch 1506/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8058 - accuracy: 0.5296 - val_loss: 2.2633 - val_accuracy: 0.4101\n",
      "Epoch 1507/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8066 - accuracy: 0.5297 - val_loss: 2.2624 - val_accuracy: 0.4094\n",
      "Epoch 1508/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5305 - val_loss: 2.2611 - val_accuracy: 0.4095\n",
      "Epoch 1509/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8057 - accuracy: 0.5302 - val_loss: 2.2589 - val_accuracy: 0.4107\n",
      "Epoch 1510/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8039 - accuracy: 0.5295 - val_loss: 2.2615 - val_accuracy: 0.4120\n",
      "Epoch 1511/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8021 - accuracy: 0.5318 - val_loss: 2.2596 - val_accuracy: 0.4147\n",
      "Epoch 1512/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5311 - val_loss: 2.2608 - val_accuracy: 0.4112\n",
      "Epoch 1513/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8083 - accuracy: 0.5303 - val_loss: 2.2580 - val_accuracy: 0.4127\n",
      "Epoch 1514/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8024 - accuracy: 0.5311 - val_loss: 2.2610 - val_accuracy: 0.4119\n",
      "Epoch 1515/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8071 - accuracy: 0.5293 - val_loss: 2.2643 - val_accuracy: 0.4118\n",
      "Epoch 1516/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8055 - accuracy: 0.5310 - val_loss: 2.2634 - val_accuracy: 0.4107\n",
      "Epoch 1517/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8037 - accuracy: 0.5310 - val_loss: 2.2662 - val_accuracy: 0.4111\n",
      "Epoch 1518/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8059 - accuracy: 0.5303 - val_loss: 2.2597 - val_accuracy: 0.4155\n",
      "Epoch 1519/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8063 - accuracy: 0.5293 - val_loss: 2.2622 - val_accuracy: 0.4111\n",
      "Epoch 1520/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8013 - accuracy: 0.5319 - val_loss: 2.2596 - val_accuracy: 0.4097\n",
      "Epoch 1521/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8084 - accuracy: 0.5293 - val_loss: 2.2588 - val_accuracy: 0.4107\n",
      "Epoch 1522/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8042 - accuracy: 0.5301 - val_loss: 2.2504 - val_accuracy: 0.4095\n",
      "Epoch 1523/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8034 - accuracy: 0.5314 - val_loss: 2.2547 - val_accuracy: 0.4100\n",
      "Epoch 1524/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8036 - accuracy: 0.5313 - val_loss: 2.2607 - val_accuracy: 0.4093\n",
      "Epoch 1525/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8040 - accuracy: 0.5309 - val_loss: 2.2605 - val_accuracy: 0.4135\n",
      "Epoch 1526/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8010 - accuracy: 0.5308 - val_loss: 2.2541 - val_accuracy: 0.4127\n",
      "Epoch 1527/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8026 - accuracy: 0.5315 - val_loss: 2.2504 - val_accuracy: 0.4144\n",
      "Epoch 1528/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8043 - accuracy: 0.5314 - val_loss: 2.2561 - val_accuracy: 0.4103\n",
      "Epoch 1529/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8051 - accuracy: 0.5295 - val_loss: 2.2609 - val_accuracy: 0.4114\n",
      "Epoch 1530/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8042 - accuracy: 0.5298 - val_loss: 2.2604 - val_accuracy: 0.4134\n",
      "Epoch 1531/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8046 - accuracy: 0.5316 - val_loss: 2.2605 - val_accuracy: 0.4101\n",
      "Epoch 1532/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8008 - accuracy: 0.5319 - val_loss: 2.2580 - val_accuracy: 0.4115\n",
      "Epoch 1533/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8041 - accuracy: 0.5313 - val_loss: 2.2578 - val_accuracy: 0.4130\n",
      "Epoch 1534/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8050 - accuracy: 0.5310 - val_loss: 2.2617 - val_accuracy: 0.4147\n",
      "Epoch 1535/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8026 - accuracy: 0.5312 - val_loss: 2.2708 - val_accuracy: 0.4120\n",
      "Epoch 1536/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8028 - accuracy: 0.5320 - val_loss: 2.2566 - val_accuracy: 0.4083\n",
      "Epoch 1537/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8052 - accuracy: 0.5297 - val_loss: 2.2604 - val_accuracy: 0.4104\n",
      "Epoch 1538/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8013 - accuracy: 0.5311 - val_loss: 2.2693 - val_accuracy: 0.4120\n",
      "Epoch 1539/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8029 - accuracy: 0.5312 - val_loss: 2.2572 - val_accuracy: 0.4136\n",
      "Epoch 1540/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8028 - accuracy: 0.5306 - val_loss: 2.2634 - val_accuracy: 0.4114\n",
      "Epoch 1541/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8019 - accuracy: 0.5323 - val_loss: 2.2623 - val_accuracy: 0.4146\n",
      "Epoch 1542/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5305 - val_loss: 2.2656 - val_accuracy: 0.4103\n",
      "Epoch 1543/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8033 - accuracy: 0.5307 - val_loss: 2.2604 - val_accuracy: 0.4076\n",
      "Epoch 1544/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5303 - val_loss: 2.2615 - val_accuracy: 0.4105\n",
      "Epoch 1545/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8040 - accuracy: 0.5316 - val_loss: 2.2595 - val_accuracy: 0.4146\n",
      "Epoch 1546/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8056 - accuracy: 0.5309 - val_loss: 2.2559 - val_accuracy: 0.4149\n",
      "Epoch 1547/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8049 - accuracy: 0.5308 - val_loss: 2.2505 - val_accuracy: 0.4153\n",
      "Epoch 1548/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8038 - accuracy: 0.5307 - val_loss: 2.2604 - val_accuracy: 0.4145\n",
      "Epoch 1549/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8046 - accuracy: 0.5304 - val_loss: 2.2667 - val_accuracy: 0.4082\n",
      "Epoch 1550/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8033 - accuracy: 0.5300 - val_loss: 2.2584 - val_accuracy: 0.4127\n",
      "Epoch 1551/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8087 - accuracy: 0.5290 - val_loss: 2.2587 - val_accuracy: 0.4107\n",
      "Epoch 1552/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8058 - accuracy: 0.5296 - val_loss: 2.2652 - val_accuracy: 0.4098\n",
      "Epoch 1553/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8029 - accuracy: 0.5304 - val_loss: 2.2621 - val_accuracy: 0.4114\n",
      "Epoch 1554/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8047 - accuracy: 0.5312 - val_loss: 2.2604 - val_accuracy: 0.4084\n",
      "Epoch 1555/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8035 - accuracy: 0.5308 - val_loss: 2.2642 - val_accuracy: 0.4119\n",
      "Epoch 1556/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8042 - accuracy: 0.5301 - val_loss: 2.2609 - val_accuracy: 0.4143\n",
      "Epoch 1557/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8045 - accuracy: 0.5297 - val_loss: 2.2669 - val_accuracy: 0.4117\n",
      "Epoch 1558/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8032 - accuracy: 0.5309 - val_loss: 2.2695 - val_accuracy: 0.4092\n",
      "Epoch 1559/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8049 - accuracy: 0.5305 - val_loss: 2.2670 - val_accuracy: 0.4147\n",
      "Epoch 1560/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8030 - accuracy: 0.5305 - val_loss: 2.2639 - val_accuracy: 0.4130\n",
      "Epoch 1561/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8055 - accuracy: 0.5312 - val_loss: 2.2647 - val_accuracy: 0.4111\n",
      "Epoch 1562/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8049 - accuracy: 0.5312 - val_loss: 2.2664 - val_accuracy: 0.4141\n",
      "Epoch 1563/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5315 - val_loss: 2.2699 - val_accuracy: 0.4149\n",
      "Epoch 1564/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8050 - accuracy: 0.5300 - val_loss: 2.2694 - val_accuracy: 0.4141\n",
      "Epoch 1565/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8012 - accuracy: 0.5325 - val_loss: 2.2638 - val_accuracy: 0.4146\n",
      "Epoch 1566/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8062 - accuracy: 0.5300 - val_loss: 2.2662 - val_accuracy: 0.4123\n",
      "Epoch 1567/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8007 - accuracy: 0.5312 - val_loss: 2.2562 - val_accuracy: 0.4150\n",
      "Epoch 1568/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8067 - accuracy: 0.5303 - val_loss: 2.2658 - val_accuracy: 0.4133\n",
      "Epoch 1569/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5304 - val_loss: 2.2610 - val_accuracy: 0.4127\n",
      "Epoch 1570/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8083 - accuracy: 0.5288 - val_loss: 2.2665 - val_accuracy: 0.4106\n",
      "Epoch 1571/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8050 - accuracy: 0.5305 - val_loss: 2.2656 - val_accuracy: 0.4120\n",
      "Epoch 1572/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8026 - accuracy: 0.5320 - val_loss: 2.2670 - val_accuracy: 0.4121\n",
      "Epoch 1573/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8060 - accuracy: 0.5311 - val_loss: 2.2652 - val_accuracy: 0.4125\n",
      "Epoch 1574/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8053 - accuracy: 0.5313 - val_loss: 2.2635 - val_accuracy: 0.4106\n",
      "Epoch 1575/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8031 - accuracy: 0.5305 - val_loss: 2.2679 - val_accuracy: 0.4117\n",
      "Epoch 1576/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8078 - accuracy: 0.5292 - val_loss: 2.2619 - val_accuracy: 0.4139\n",
      "Epoch 1577/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8011 - accuracy: 0.5320 - val_loss: 2.2662 - val_accuracy: 0.4102\n",
      "Epoch 1578/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8063 - accuracy: 0.5305 - val_loss: 2.2650 - val_accuracy: 0.4114\n",
      "Epoch 1579/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.8002 - accuracy: 0.5329 - val_loss: 2.2656 - val_accuracy: 0.4104\n",
      "Epoch 1580/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8031 - accuracy: 0.5310 - val_loss: 2.2626 - val_accuracy: 0.4146\n",
      "Epoch 1581/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8075 - accuracy: 0.5304 - val_loss: 2.2594 - val_accuracy: 0.4127\n",
      "Epoch 1582/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8040 - accuracy: 0.5304 - val_loss: 2.2671 - val_accuracy: 0.4112\n",
      "Epoch 1583/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8050 - accuracy: 0.5305 - val_loss: 2.2672 - val_accuracy: 0.4105\n",
      "Epoch 1584/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8060 - accuracy: 0.5305 - val_loss: 2.2635 - val_accuracy: 0.4098\n",
      "Epoch 1585/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8037 - accuracy: 0.5308 - val_loss: 2.2693 - val_accuracy: 0.4117\n",
      "Epoch 1586/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8018 - accuracy: 0.5321 - val_loss: 2.2606 - val_accuracy: 0.4134\n",
      "Epoch 1587/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8042 - accuracy: 0.5308 - val_loss: 2.2611 - val_accuracy: 0.4144\n",
      "Epoch 1588/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5332 - val_loss: 2.2561 - val_accuracy: 0.4142\n",
      "Epoch 1589/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8030 - accuracy: 0.5317 - val_loss: 2.2593 - val_accuracy: 0.4136\n",
      "Epoch 1590/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8013 - accuracy: 0.5324 - val_loss: 2.2638 - val_accuracy: 0.4123\n",
      "Epoch 1591/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8010 - accuracy: 0.5337 - val_loss: 2.2613 - val_accuracy: 0.4134\n",
      "Epoch 1592/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7995 - accuracy: 0.5324 - val_loss: 2.2603 - val_accuracy: 0.4149\n",
      "Epoch 1593/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8013 - accuracy: 0.5325 - val_loss: 2.2612 - val_accuracy: 0.4130\n",
      "Epoch 1594/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8020 - accuracy: 0.5312 - val_loss: 2.2624 - val_accuracy: 0.4142\n",
      "Epoch 1595/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8011 - accuracy: 0.5327 - val_loss: 2.2676 - val_accuracy: 0.4106\n",
      "Epoch 1596/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8008 - accuracy: 0.5322 - val_loss: 2.2568 - val_accuracy: 0.4119\n",
      "Epoch 1597/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8032 - accuracy: 0.5306 - val_loss: 2.2638 - val_accuracy: 0.4103\n",
      "Epoch 1598/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8024 - accuracy: 0.5317 - val_loss: 2.2593 - val_accuracy: 0.4119\n",
      "Epoch 1599/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8051 - accuracy: 0.5304 - val_loss: 2.2698 - val_accuracy: 0.4093\n",
      "Epoch 1600/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8045 - accuracy: 0.5308 - val_loss: 2.2606 - val_accuracy: 0.4127\n",
      "Epoch 1601/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5310 - val_loss: 2.2639 - val_accuracy: 0.4123\n",
      "Epoch 1602/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8047 - accuracy: 0.5305 - val_loss: 2.2649 - val_accuracy: 0.4127\n",
      "Epoch 1603/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5315 - val_loss: 2.2672 - val_accuracy: 0.4128\n",
      "Epoch 1604/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8024 - accuracy: 0.5312 - val_loss: 2.2673 - val_accuracy: 0.4157\n",
      "Epoch 1605/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8016 - accuracy: 0.5325 - val_loss: 2.2629 - val_accuracy: 0.4137\n",
      "Epoch 1606/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8018 - accuracy: 0.5313 - val_loss: 2.2648 - val_accuracy: 0.4161\n",
      "Epoch 1607/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8019 - accuracy: 0.5319 - val_loss: 2.2597 - val_accuracy: 0.4146\n",
      "Epoch 1608/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8034 - accuracy: 0.5311 - val_loss: 2.2591 - val_accuracy: 0.4127\n",
      "Epoch 1609/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8009 - accuracy: 0.5310 - val_loss: 2.2612 - val_accuracy: 0.4111\n",
      "Epoch 1610/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8022 - accuracy: 0.5318 - val_loss: 2.2617 - val_accuracy: 0.4117\n",
      "Epoch 1611/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8056 - accuracy: 0.5304 - val_loss: 2.2653 - val_accuracy: 0.4116\n",
      "Epoch 1612/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8034 - accuracy: 0.5315 - val_loss: 2.2635 - val_accuracy: 0.4123\n",
      "Epoch 1613/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8024 - accuracy: 0.5307 - val_loss: 2.2620 - val_accuracy: 0.4108\n",
      "Epoch 1614/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8035 - accuracy: 0.5308 - val_loss: 2.2624 - val_accuracy: 0.4142\n",
      "Epoch 1615/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8025 - accuracy: 0.5315 - val_loss: 2.2658 - val_accuracy: 0.4130\n",
      "Epoch 1616/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8027 - accuracy: 0.5316 - val_loss: 2.2624 - val_accuracy: 0.4154\n",
      "Epoch 1617/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8007 - accuracy: 0.5312 - val_loss: 2.2674 - val_accuracy: 0.4121\n",
      "Epoch 1618/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8016 - accuracy: 0.5320 - val_loss: 2.2663 - val_accuracy: 0.4108\n",
      "Epoch 1619/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8035 - accuracy: 0.5298 - val_loss: 2.2661 - val_accuracy: 0.4120\n",
      "Epoch 1620/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8024 - accuracy: 0.5308 - val_loss: 2.2695 - val_accuracy: 0.4122\n",
      "Epoch 1621/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8004 - accuracy: 0.5326 - val_loss: 2.2628 - val_accuracy: 0.4127\n",
      "Epoch 1622/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8045 - accuracy: 0.5312 - val_loss: 2.2657 - val_accuracy: 0.4092\n",
      "Epoch 1623/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5325 - val_loss: 2.2698 - val_accuracy: 0.4117\n",
      "Epoch 1624/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8018 - accuracy: 0.5311 - val_loss: 2.2661 - val_accuracy: 0.4102\n",
      "Epoch 1625/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8048 - accuracy: 0.5307 - val_loss: 2.2675 - val_accuracy: 0.4081\n",
      "Epoch 1626/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8061 - accuracy: 0.5299 - val_loss: 2.2678 - val_accuracy: 0.4112\n",
      "Epoch 1627/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8002 - accuracy: 0.5318 - val_loss: 2.2623 - val_accuracy: 0.4093\n",
      "Epoch 1628/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8001 - accuracy: 0.5326 - val_loss: 2.2677 - val_accuracy: 0.4082\n",
      "Epoch 1629/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8034 - accuracy: 0.5306 - val_loss: 2.2590 - val_accuracy: 0.4084\n",
      "Epoch 1630/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8007 - accuracy: 0.5322 - val_loss: 2.2639 - val_accuracy: 0.4107\n",
      "Epoch 1631/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8014 - accuracy: 0.5324 - val_loss: 2.2610 - val_accuracy: 0.4165\n",
      "Epoch 1632/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8052 - accuracy: 0.5312 - val_loss: 2.2623 - val_accuracy: 0.4133\n",
      "Epoch 1633/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8019 - accuracy: 0.5311 - val_loss: 2.2534 - val_accuracy: 0.4110\n",
      "Epoch 1634/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8033 - accuracy: 0.5316 - val_loss: 2.2561 - val_accuracy: 0.4136\n",
      "Epoch 1635/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8042 - accuracy: 0.5292 - val_loss: 2.2622 - val_accuracy: 0.4150\n",
      "Epoch 1636/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8025 - accuracy: 0.5322 - val_loss: 2.2684 - val_accuracy: 0.4092\n",
      "Epoch 1637/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8038 - accuracy: 0.5308 - val_loss: 2.2619 - val_accuracy: 0.4095\n",
      "Epoch 1638/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8037 - accuracy: 0.5317 - val_loss: 2.2622 - val_accuracy: 0.4095\n",
      "Epoch 1639/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8019 - accuracy: 0.5312 - val_loss: 2.2613 - val_accuracy: 0.4101\n",
      "Epoch 1640/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8026 - accuracy: 0.5317 - val_loss: 2.2550 - val_accuracy: 0.4114\n",
      "Epoch 1641/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7996 - accuracy: 0.5319 - val_loss: 2.2609 - val_accuracy: 0.4114\n",
      "Epoch 1642/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7989 - accuracy: 0.5330 - val_loss: 2.2672 - val_accuracy: 0.4111\n",
      "Epoch 1643/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5310 - val_loss: 2.2636 - val_accuracy: 0.4088\n",
      "Epoch 1644/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7985 - accuracy: 0.5332 - val_loss: 2.2591 - val_accuracy: 0.4120\n",
      "Epoch 1645/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8015 - accuracy: 0.5325 - val_loss: 2.2718 - val_accuracy: 0.4095\n",
      "Epoch 1646/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8018 - accuracy: 0.5307 - val_loss: 2.2608 - val_accuracy: 0.4116\n",
      "Epoch 1647/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8017 - accuracy: 0.5313 - val_loss: 2.2578 - val_accuracy: 0.4146\n",
      "Epoch 1648/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8013 - accuracy: 0.5326 - val_loss: 2.2660 - val_accuracy: 0.4096\n",
      "Epoch 1649/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8022 - accuracy: 0.5324 - val_loss: 2.2637 - val_accuracy: 0.4152\n",
      "Epoch 1650/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8030 - accuracy: 0.5314 - val_loss: 2.2678 - val_accuracy: 0.4130\n",
      "Epoch 1651/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5313 - val_loss: 2.2590 - val_accuracy: 0.4125\n",
      "Epoch 1652/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8001 - accuracy: 0.5325 - val_loss: 2.2583 - val_accuracy: 0.4149\n",
      "Epoch 1653/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5317 - val_loss: 2.2598 - val_accuracy: 0.4123\n",
      "Epoch 1654/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8009 - accuracy: 0.5326 - val_loss: 2.2596 - val_accuracy: 0.4111\n",
      "Epoch 1655/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7988 - accuracy: 0.5322 - val_loss: 2.2598 - val_accuracy: 0.4113\n",
      "Epoch 1656/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8009 - accuracy: 0.5314 - val_loss: 2.2651 - val_accuracy: 0.4096\n",
      "Epoch 1657/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8015 - accuracy: 0.5307 - val_loss: 2.2510 - val_accuracy: 0.4161\n",
      "Epoch 1658/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8015 - accuracy: 0.5317 - val_loss: 2.2649 - val_accuracy: 0.4109\n",
      "Epoch 1659/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8006 - accuracy: 0.5321 - val_loss: 2.2631 - val_accuracy: 0.4131\n",
      "Epoch 1660/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8035 - accuracy: 0.5323 - val_loss: 2.2565 - val_accuracy: 0.4148\n",
      "Epoch 1661/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8044 - accuracy: 0.5301 - val_loss: 2.2607 - val_accuracy: 0.4104\n",
      "Epoch 1662/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8029 - accuracy: 0.5321 - val_loss: 2.2614 - val_accuracy: 0.4144\n",
      "Epoch 1663/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8016 - accuracy: 0.5316 - val_loss: 2.2637 - val_accuracy: 0.4152\n",
      "Epoch 1664/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8028 - accuracy: 0.5308 - val_loss: 2.2633 - val_accuracy: 0.4137\n",
      "Epoch 1665/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8034 - accuracy: 0.5304 - val_loss: 2.2583 - val_accuracy: 0.4135\n",
      "Epoch 1666/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7994 - accuracy: 0.5316 - val_loss: 2.2670 - val_accuracy: 0.4109\n",
      "Epoch 1667/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8009 - accuracy: 0.5328 - val_loss: 2.2620 - val_accuracy: 0.4139\n",
      "Epoch 1668/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8023 - accuracy: 0.5313 - val_loss: 2.2566 - val_accuracy: 0.4136\n",
      "Epoch 1669/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7975 - accuracy: 0.5336 - val_loss: 2.2655 - val_accuracy: 0.4120\n",
      "Epoch 1670/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8032 - accuracy: 0.5312 - val_loss: 2.2622 - val_accuracy: 0.4126\n",
      "Epoch 1671/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7994 - accuracy: 0.5321 - val_loss: 2.2632 - val_accuracy: 0.4151\n",
      "Epoch 1672/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7989 - accuracy: 0.5326 - val_loss: 2.2614 - val_accuracy: 0.4102\n",
      "Epoch 1673/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8016 - accuracy: 0.5318 - val_loss: 2.2621 - val_accuracy: 0.4123\n",
      "Epoch 1674/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8010 - accuracy: 0.5320 - val_loss: 2.2613 - val_accuracy: 0.4108\n",
      "Epoch 1675/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7977 - accuracy: 0.5336 - val_loss: 2.2602 - val_accuracy: 0.4136\n",
      "Epoch 1676/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8019 - accuracy: 0.5324 - val_loss: 2.2642 - val_accuracy: 0.4118\n",
      "Epoch 1677/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8044 - accuracy: 0.5299 - val_loss: 2.2588 - val_accuracy: 0.4117\n",
      "Epoch 1678/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7989 - accuracy: 0.5326 - val_loss: 2.2623 - val_accuracy: 0.4081\n",
      "Epoch 1679/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8016 - accuracy: 0.5318 - val_loss: 2.2584 - val_accuracy: 0.4142\n",
      "Epoch 1680/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8016 - accuracy: 0.5318 - val_loss: 2.2602 - val_accuracy: 0.4112\n",
      "Epoch 1681/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8027 - accuracy: 0.5321 - val_loss: 2.2663 - val_accuracy: 0.4119\n",
      "Epoch 1682/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7983 - accuracy: 0.5324 - val_loss: 2.2650 - val_accuracy: 0.4123\n",
      "Epoch 1683/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8013 - accuracy: 0.5314 - val_loss: 2.2614 - val_accuracy: 0.4130\n",
      "Epoch 1684/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8011 - accuracy: 0.5325 - val_loss: 2.2591 - val_accuracy: 0.4114\n",
      "Epoch 1685/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7999 - accuracy: 0.5323 - val_loss: 2.2574 - val_accuracy: 0.4114\n",
      "Epoch 1686/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7999 - accuracy: 0.5325 - val_loss: 2.2639 - val_accuracy: 0.4104\n",
      "Epoch 1687/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8004 - accuracy: 0.5327 - val_loss: 2.2639 - val_accuracy: 0.4101\n",
      "Epoch 1688/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8013 - accuracy: 0.5314 - val_loss: 2.2692 - val_accuracy: 0.4105\n",
      "Epoch 1689/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8018 - accuracy: 0.5321 - val_loss: 2.2622 - val_accuracy: 0.4091\n",
      "Epoch 1690/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8030 - accuracy: 0.5309 - val_loss: 2.2639 - val_accuracy: 0.4122\n",
      "Epoch 1691/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7981 - accuracy: 0.5335 - val_loss: 2.2657 - val_accuracy: 0.4107\n",
      "Epoch 1692/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7989 - accuracy: 0.5328 - val_loss: 2.2674 - val_accuracy: 0.4142\n",
      "Epoch 1693/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5319 - val_loss: 2.2616 - val_accuracy: 0.4137\n",
      "Epoch 1694/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8003 - accuracy: 0.5323 - val_loss: 2.2620 - val_accuracy: 0.4118\n",
      "Epoch 1695/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8010 - accuracy: 0.5313 - val_loss: 2.2652 - val_accuracy: 0.4104\n",
      "Epoch 1696/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5314 - val_loss: 2.2621 - val_accuracy: 0.4153\n",
      "Epoch 1697/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8027 - accuracy: 0.5314 - val_loss: 2.2591 - val_accuracy: 0.4107\n",
      "Epoch 1698/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8013 - accuracy: 0.5315 - val_loss: 2.2615 - val_accuracy: 0.4143\n",
      "Epoch 1699/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8014 - accuracy: 0.5321 - val_loss: 2.2694 - val_accuracy: 0.4089\n",
      "Epoch 1700/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8010 - accuracy: 0.5320 - val_loss: 2.2730 - val_accuracy: 0.4111\n",
      "Epoch 1701/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8021 - accuracy: 0.5320 - val_loss: 2.2660 - val_accuracy: 0.4154\n",
      "Epoch 1702/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8033 - accuracy: 0.5318 - val_loss: 2.2613 - val_accuracy: 0.4150\n",
      "Epoch 1703/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8001 - accuracy: 0.5316 - val_loss: 2.2663 - val_accuracy: 0.4098\n",
      "Epoch 1704/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8015 - accuracy: 0.5312 - val_loss: 2.2633 - val_accuracy: 0.4106\n",
      "Epoch 1705/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8006 - accuracy: 0.5326 - val_loss: 2.2641 - val_accuracy: 0.4136\n",
      "Epoch 1706/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5317 - val_loss: 2.2638 - val_accuracy: 0.4122\n",
      "Epoch 1707/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7953 - accuracy: 0.5349 - val_loss: 2.2688 - val_accuracy: 0.4095\n",
      "Epoch 1708/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8053 - accuracy: 0.5293 - val_loss: 2.2609 - val_accuracy: 0.4083\n",
      "Epoch 1709/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7993 - accuracy: 0.5330 - val_loss: 2.2586 - val_accuracy: 0.4139\n",
      "Epoch 1710/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7989 - accuracy: 0.5323 - val_loss: 2.2628 - val_accuracy: 0.4129\n",
      "Epoch 1711/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7991 - accuracy: 0.5339 - val_loss: 2.2658 - val_accuracy: 0.4125\n",
      "Epoch 1712/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7988 - accuracy: 0.5320 - val_loss: 2.2579 - val_accuracy: 0.4137\n",
      "Epoch 1713/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7988 - accuracy: 0.5329 - val_loss: 2.2600 - val_accuracy: 0.4119\n",
      "Epoch 1714/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8031 - accuracy: 0.5309 - val_loss: 2.2564 - val_accuracy: 0.4143\n",
      "Epoch 1715/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8008 - accuracy: 0.5326 - val_loss: 2.2640 - val_accuracy: 0.4138\n",
      "Epoch 1716/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8001 - accuracy: 0.5328 - val_loss: 2.2578 - val_accuracy: 0.4137\n",
      "Epoch 1717/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8012 - accuracy: 0.5325 - val_loss: 2.2637 - val_accuracy: 0.4106\n",
      "Epoch 1718/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5327 - val_loss: 2.2616 - val_accuracy: 0.4107\n",
      "Epoch 1719/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7990 - accuracy: 0.5324 - val_loss: 2.2529 - val_accuracy: 0.4146\n",
      "Epoch 1720/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8008 - accuracy: 0.5325 - val_loss: 2.2609 - val_accuracy: 0.4135\n",
      "Epoch 1721/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7996 - accuracy: 0.5330 - val_loss: 2.2600 - val_accuracy: 0.4143\n",
      "Epoch 1722/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8008 - accuracy: 0.5323 - val_loss: 2.2572 - val_accuracy: 0.4111\n",
      "Epoch 1723/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7994 - accuracy: 0.5324 - val_loss: 2.2657 - val_accuracy: 0.4114\n",
      "Epoch 1724/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7983 - accuracy: 0.5328 - val_loss: 2.2604 - val_accuracy: 0.4134\n",
      "Epoch 1725/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8001 - accuracy: 0.5328 - val_loss: 2.2618 - val_accuracy: 0.4119\n",
      "Epoch 1726/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8016 - accuracy: 0.5320 - val_loss: 2.2564 - val_accuracy: 0.4158\n",
      "Epoch 1727/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8026 - accuracy: 0.5316 - val_loss: 2.2646 - val_accuracy: 0.4146\n",
      "Epoch 1728/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8013 - accuracy: 0.5324 - val_loss: 2.2651 - val_accuracy: 0.4117\n",
      "Epoch 1729/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7979 - accuracy: 0.5341 - val_loss: 2.2594 - val_accuracy: 0.4136\n",
      "Epoch 1730/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8008 - accuracy: 0.5316 - val_loss: 2.2646 - val_accuracy: 0.4118\n",
      "Epoch 1731/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8011 - accuracy: 0.5318 - val_loss: 2.2613 - val_accuracy: 0.4135\n",
      "Epoch 1732/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7995 - accuracy: 0.5328 - val_loss: 2.2490 - val_accuracy: 0.4142\n",
      "Epoch 1733/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5323 - val_loss: 2.2562 - val_accuracy: 0.4133\n",
      "Epoch 1734/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7976 - accuracy: 0.5336 - val_loss: 2.2542 - val_accuracy: 0.4113\n",
      "Epoch 1735/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7989 - accuracy: 0.5331 - val_loss: 2.2592 - val_accuracy: 0.4146\n",
      "Epoch 1736/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8006 - accuracy: 0.5322 - val_loss: 2.2632 - val_accuracy: 0.4112\n",
      "Epoch 1737/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7989 - accuracy: 0.5327 - val_loss: 2.2585 - val_accuracy: 0.4104\n",
      "Epoch 1738/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7987 - accuracy: 0.5332 - val_loss: 2.2633 - val_accuracy: 0.4095\n",
      "Epoch 1739/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7979 - accuracy: 0.5325 - val_loss: 2.2642 - val_accuracy: 0.4178\n",
      "Epoch 1740/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8046 - accuracy: 0.5307 - val_loss: 2.2561 - val_accuracy: 0.4135\n",
      "Epoch 1741/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7989 - accuracy: 0.5328 - val_loss: 2.2657 - val_accuracy: 0.4155\n",
      "Epoch 1742/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7974 - accuracy: 0.5329 - val_loss: 2.2588 - val_accuracy: 0.4142\n",
      "Epoch 1743/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7998 - accuracy: 0.5333 - val_loss: 2.2626 - val_accuracy: 0.4150\n",
      "Epoch 1744/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7958 - accuracy: 0.5340 - val_loss: 2.2725 - val_accuracy: 0.4101\n",
      "Epoch 1745/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7971 - accuracy: 0.5334 - val_loss: 2.2566 - val_accuracy: 0.4146\n",
      "Epoch 1746/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7998 - accuracy: 0.5328 - val_loss: 2.2674 - val_accuracy: 0.4145\n",
      "Epoch 1747/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8003 - accuracy: 0.5329 - val_loss: 2.2646 - val_accuracy: 0.4109\n",
      "Epoch 1748/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7988 - accuracy: 0.5325 - val_loss: 2.2591 - val_accuracy: 0.4130\n",
      "Epoch 1749/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8020 - accuracy: 0.5326 - val_loss: 2.2516 - val_accuracy: 0.4152\n",
      "Epoch 1750/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5332 - val_loss: 2.2586 - val_accuracy: 0.4118\n",
      "Epoch 1751/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7995 - accuracy: 0.5329 - val_loss: 2.2698 - val_accuracy: 0.4127\n",
      "Epoch 1752/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7980 - accuracy: 0.5341 - val_loss: 2.2612 - val_accuracy: 0.4150\n",
      "Epoch 1753/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7994 - accuracy: 0.5321 - val_loss: 2.2716 - val_accuracy: 0.4111\n",
      "Epoch 1754/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7976 - accuracy: 0.5329 - val_loss: 2.2625 - val_accuracy: 0.4117\n",
      "Epoch 1755/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8015 - accuracy: 0.5322 - val_loss: 2.2630 - val_accuracy: 0.4145\n",
      "Epoch 1756/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5335 - val_loss: 2.2634 - val_accuracy: 0.4113\n",
      "Epoch 1757/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7998 - accuracy: 0.5326 - val_loss: 2.2612 - val_accuracy: 0.4120\n",
      "Epoch 1758/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7987 - accuracy: 0.5333 - val_loss: 2.2595 - val_accuracy: 0.4128\n",
      "Epoch 1759/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7977 - accuracy: 0.5337 - val_loss: 2.2548 - val_accuracy: 0.4157\n",
      "Epoch 1760/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8008 - accuracy: 0.5323 - val_loss: 2.2542 - val_accuracy: 0.4122\n",
      "Epoch 1761/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8013 - accuracy: 0.5314 - val_loss: 2.2675 - val_accuracy: 0.4133\n",
      "Epoch 1762/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8000 - accuracy: 0.5328 - val_loss: 2.2628 - val_accuracy: 0.4099\n",
      "Epoch 1763/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5328 - val_loss: 2.2632 - val_accuracy: 0.4117\n",
      "Epoch 1764/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8001 - accuracy: 0.5331 - val_loss: 2.2607 - val_accuracy: 0.4130\n",
      "Epoch 1765/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7955 - accuracy: 0.5340 - val_loss: 2.2588 - val_accuracy: 0.4141\n",
      "Epoch 1766/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8014 - accuracy: 0.5331 - val_loss: 2.2583 - val_accuracy: 0.4098\n",
      "Epoch 1767/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7993 - accuracy: 0.5337 - val_loss: 2.2493 - val_accuracy: 0.4148\n",
      "Epoch 1768/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8015 - accuracy: 0.5321 - val_loss: 2.2505 - val_accuracy: 0.4152\n",
      "Epoch 1769/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7981 - accuracy: 0.5332 - val_loss: 2.2641 - val_accuracy: 0.4130\n",
      "Epoch 1770/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7989 - accuracy: 0.5339 - val_loss: 2.2579 - val_accuracy: 0.4115\n",
      "Epoch 1771/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8018 - accuracy: 0.5310 - val_loss: 2.2542 - val_accuracy: 0.4135\n",
      "Epoch 1772/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8015 - accuracy: 0.5324 - val_loss: 2.2574 - val_accuracy: 0.4114\n",
      "Epoch 1773/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8029 - accuracy: 0.5322 - val_loss: 2.2556 - val_accuracy: 0.4113\n",
      "Epoch 1774/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8014 - accuracy: 0.5321 - val_loss: 2.2548 - val_accuracy: 0.4145\n",
      "Epoch 1775/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7996 - accuracy: 0.5331 - val_loss: 2.2606 - val_accuracy: 0.4113\n",
      "Epoch 1776/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8004 - accuracy: 0.5330 - val_loss: 2.2650 - val_accuracy: 0.4126\n",
      "Epoch 1777/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7978 - accuracy: 0.5335 - val_loss: 2.2645 - val_accuracy: 0.4142\n",
      "Epoch 1778/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7993 - accuracy: 0.5334 - val_loss: 2.2648 - val_accuracy: 0.4127\n",
      "Epoch 1779/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8003 - accuracy: 0.5327 - val_loss: 2.2636 - val_accuracy: 0.4098\n",
      "Epoch 1780/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7996 - accuracy: 0.5327 - val_loss: 2.2595 - val_accuracy: 0.4154\n",
      "Epoch 1781/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5321 - val_loss: 2.2562 - val_accuracy: 0.4156\n",
      "Epoch 1782/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5322 - val_loss: 2.2581 - val_accuracy: 0.4129\n",
      "Epoch 1783/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5337 - val_loss: 2.2534 - val_accuracy: 0.4139\n",
      "Epoch 1784/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8001 - accuracy: 0.5321 - val_loss: 2.2684 - val_accuracy: 0.4114\n",
      "Epoch 1785/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7956 - accuracy: 0.5339 - val_loss: 2.2597 - val_accuracy: 0.4104\n",
      "Epoch 1786/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5331 - val_loss: 2.2630 - val_accuracy: 0.4114\n",
      "Epoch 1787/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8010 - accuracy: 0.5326 - val_loss: 2.2582 - val_accuracy: 0.4101\n",
      "Epoch 1788/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7963 - accuracy: 0.5341 - val_loss: 2.2634 - val_accuracy: 0.4161\n",
      "Epoch 1789/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7968 - accuracy: 0.5327 - val_loss: 2.2622 - val_accuracy: 0.4139\n",
      "Epoch 1790/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7968 - accuracy: 0.5336 - val_loss: 2.2604 - val_accuracy: 0.4145\n",
      "Epoch 1791/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8004 - accuracy: 0.5321 - val_loss: 2.2593 - val_accuracy: 0.4107\n",
      "Epoch 1792/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7997 - accuracy: 0.5323 - val_loss: 2.2528 - val_accuracy: 0.4132\n",
      "Epoch 1793/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7960 - accuracy: 0.5330 - val_loss: 2.2642 - val_accuracy: 0.4130\n",
      "Epoch 1794/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5332 - val_loss: 2.2504 - val_accuracy: 0.4146\n",
      "Epoch 1795/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5320 - val_loss: 2.2554 - val_accuracy: 0.4162\n",
      "Epoch 1796/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8000 - accuracy: 0.5327 - val_loss: 2.2498 - val_accuracy: 0.4145\n",
      "Epoch 1797/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8000 - accuracy: 0.5328 - val_loss: 2.2533 - val_accuracy: 0.4116\n",
      "Epoch 1798/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7950 - accuracy: 0.5341 - val_loss: 2.2615 - val_accuracy: 0.4146\n",
      "Epoch 1799/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7949 - accuracy: 0.5337 - val_loss: 2.2610 - val_accuracy: 0.4111\n",
      "Epoch 1800/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8003 - accuracy: 0.5319 - val_loss: 2.2622 - val_accuracy: 0.4121\n",
      "Epoch 1801/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7958 - accuracy: 0.5338 - val_loss: 2.2662 - val_accuracy: 0.4148\n",
      "Epoch 1802/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7981 - accuracy: 0.5329 - val_loss: 2.2567 - val_accuracy: 0.4150\n",
      "Epoch 1803/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7995 - accuracy: 0.5319 - val_loss: 2.2617 - val_accuracy: 0.4095\n",
      "Epoch 1804/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8004 - accuracy: 0.5328 - val_loss: 2.2562 - val_accuracy: 0.4116\n",
      "Epoch 1805/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7971 - accuracy: 0.5336 - val_loss: 2.2613 - val_accuracy: 0.4111\n",
      "Epoch 1806/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7996 - accuracy: 0.5322 - val_loss: 2.2604 - val_accuracy: 0.4133\n",
      "Epoch 1807/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7976 - accuracy: 0.5331 - val_loss: 2.2621 - val_accuracy: 0.4152\n",
      "Epoch 1808/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5324 - val_loss: 2.2655 - val_accuracy: 0.4120\n",
      "Epoch 1809/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7977 - accuracy: 0.5330 - val_loss: 2.2586 - val_accuracy: 0.4098\n",
      "Epoch 1810/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7958 - accuracy: 0.5346 - val_loss: 2.2702 - val_accuracy: 0.4076\n",
      "Epoch 1811/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7982 - accuracy: 0.5332 - val_loss: 2.2705 - val_accuracy: 0.4096\n",
      "Epoch 1812/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7986 - accuracy: 0.5329 - val_loss: 2.2677 - val_accuracy: 0.4076\n",
      "Epoch 1813/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8004 - accuracy: 0.5325 - val_loss: 2.2616 - val_accuracy: 0.4103\n",
      "Epoch 1814/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8025 - accuracy: 0.5317 - val_loss: 2.2637 - val_accuracy: 0.4125\n",
      "Epoch 1815/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7989 - accuracy: 0.5330 - val_loss: 2.2650 - val_accuracy: 0.4100\n",
      "Epoch 1816/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7995 - accuracy: 0.5325 - val_loss: 2.2584 - val_accuracy: 0.4150\n",
      "Epoch 1817/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7981 - accuracy: 0.5330 - val_loss: 2.2693 - val_accuracy: 0.4121\n",
      "Epoch 1818/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8001 - accuracy: 0.5320 - val_loss: 2.2562 - val_accuracy: 0.4141\n",
      "Epoch 1819/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7983 - accuracy: 0.5328 - val_loss: 2.2601 - val_accuracy: 0.4123\n",
      "Epoch 1820/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7988 - accuracy: 0.5318 - val_loss: 2.2640 - val_accuracy: 0.4114\n",
      "Epoch 1821/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7991 - accuracy: 0.5326 - val_loss: 2.2602 - val_accuracy: 0.4126\n",
      "Epoch 1822/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8003 - accuracy: 0.5325 - val_loss: 2.2652 - val_accuracy: 0.4098\n",
      "Epoch 1823/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7960 - accuracy: 0.5329 - val_loss: 2.2628 - val_accuracy: 0.4136\n",
      "Epoch 1824/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7985 - accuracy: 0.5326 - val_loss: 2.2634 - val_accuracy: 0.4108\n",
      "Epoch 1825/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7988 - accuracy: 0.5318 - val_loss: 2.2584 - val_accuracy: 0.4149\n",
      "Epoch 1826/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8002 - accuracy: 0.5318 - val_loss: 2.2580 - val_accuracy: 0.4141\n",
      "Epoch 1827/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7975 - accuracy: 0.5344 - val_loss: 2.2546 - val_accuracy: 0.4108\n",
      "Epoch 1828/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7988 - accuracy: 0.5332 - val_loss: 2.2612 - val_accuracy: 0.4129\n",
      "Epoch 1829/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7983 - accuracy: 0.5333 - val_loss: 2.2535 - val_accuracy: 0.4127\n",
      "Epoch 1830/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7974 - accuracy: 0.5336 - val_loss: 2.2639 - val_accuracy: 0.4107\n",
      "Epoch 1831/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7975 - accuracy: 0.5342 - val_loss: 2.2630 - val_accuracy: 0.4130\n",
      "Epoch 1832/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7951 - accuracy: 0.5340 - val_loss: 2.2635 - val_accuracy: 0.4115\n",
      "Epoch 1833/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8005 - accuracy: 0.5322 - val_loss: 2.2643 - val_accuracy: 0.4151\n",
      "Epoch 1834/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8003 - accuracy: 0.5321 - val_loss: 2.2603 - val_accuracy: 0.4133\n",
      "Epoch 1835/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7987 - accuracy: 0.5324 - val_loss: 2.2602 - val_accuracy: 0.4117\n",
      "Epoch 1836/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5336 - val_loss: 2.2601 - val_accuracy: 0.4130\n",
      "Epoch 1837/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7965 - accuracy: 0.5339 - val_loss: 2.2585 - val_accuracy: 0.4141\n",
      "Epoch 1838/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7976 - accuracy: 0.5332 - val_loss: 2.2614 - val_accuracy: 0.4127\n",
      "Epoch 1839/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7983 - accuracy: 0.5328 - val_loss: 2.2579 - val_accuracy: 0.4134\n",
      "Epoch 1840/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7992 - accuracy: 0.5326 - val_loss: 2.2656 - val_accuracy: 0.4136\n",
      "Epoch 1841/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7970 - accuracy: 0.5331 - val_loss: 2.2606 - val_accuracy: 0.4123\n",
      "Epoch 1842/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7993 - accuracy: 0.5331 - val_loss: 2.2641 - val_accuracy: 0.4102\n",
      "Epoch 1843/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8004 - accuracy: 0.5324 - val_loss: 2.2650 - val_accuracy: 0.4120\n",
      "Epoch 1844/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7973 - accuracy: 0.5336 - val_loss: 2.2577 - val_accuracy: 0.4159\n",
      "Epoch 1845/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8010 - accuracy: 0.5332 - val_loss: 2.2587 - val_accuracy: 0.4129\n",
      "Epoch 1846/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7928 - accuracy: 0.5338 - val_loss: 2.2609 - val_accuracy: 0.4100\n",
      "Epoch 1847/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5331 - val_loss: 2.2582 - val_accuracy: 0.4136\n",
      "Epoch 1848/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7971 - accuracy: 0.5338 - val_loss: 2.2646 - val_accuracy: 0.4114\n",
      "Epoch 1849/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7964 - accuracy: 0.5334 - val_loss: 2.2592 - val_accuracy: 0.4123\n",
      "Epoch 1850/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7973 - accuracy: 0.5335 - val_loss: 2.2577 - val_accuracy: 0.4132\n",
      "Epoch 1851/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8014 - accuracy: 0.5317 - val_loss: 2.2555 - val_accuracy: 0.4135\n",
      "Epoch 1852/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7975 - accuracy: 0.5336 - val_loss: 2.2565 - val_accuracy: 0.4151\n",
      "Epoch 1853/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7969 - accuracy: 0.5338 - val_loss: 2.2580 - val_accuracy: 0.4111\n",
      "Epoch 1854/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7978 - accuracy: 0.5341 - val_loss: 2.2637 - val_accuracy: 0.4083\n",
      "Epoch 1855/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7955 - accuracy: 0.5337 - val_loss: 2.2604 - val_accuracy: 0.4092\n",
      "Epoch 1856/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7972 - accuracy: 0.5335 - val_loss: 2.2606 - val_accuracy: 0.4149\n",
      "Epoch 1857/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7952 - accuracy: 0.5337 - val_loss: 2.2602 - val_accuracy: 0.4132\n",
      "Epoch 1858/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7988 - accuracy: 0.5337 - val_loss: 2.2635 - val_accuracy: 0.4109\n",
      "Epoch 1859/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7973 - accuracy: 0.5333 - val_loss: 2.2596 - val_accuracy: 0.4106\n",
      "Epoch 1860/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7951 - accuracy: 0.5342 - val_loss: 2.2602 - val_accuracy: 0.4128\n",
      "Epoch 1861/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7984 - accuracy: 0.5346 - val_loss: 2.2566 - val_accuracy: 0.4130\n",
      "Epoch 1862/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7963 - accuracy: 0.5340 - val_loss: 2.2726 - val_accuracy: 0.4103\n",
      "Epoch 1863/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7971 - accuracy: 0.5336 - val_loss: 2.2612 - val_accuracy: 0.4141\n",
      "Epoch 1864/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7963 - accuracy: 0.5343 - val_loss: 2.2613 - val_accuracy: 0.4133\n",
      "Epoch 1865/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7981 - accuracy: 0.5331 - val_loss: 2.2602 - val_accuracy: 0.4146\n",
      "Epoch 1866/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7995 - accuracy: 0.5336 - val_loss: 2.2617 - val_accuracy: 0.4116\n",
      "Epoch 1867/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7952 - accuracy: 0.5347 - val_loss: 2.2665 - val_accuracy: 0.4145\n",
      "Epoch 1868/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7986 - accuracy: 0.5327 - val_loss: 2.2718 - val_accuracy: 0.4101\n",
      "Epoch 1869/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7932 - accuracy: 0.5349 - val_loss: 2.2664 - val_accuracy: 0.4128\n",
      "Epoch 1870/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7954 - accuracy: 0.5342 - val_loss: 2.2628 - val_accuracy: 0.4132\n",
      "Epoch 1871/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5323 - val_loss: 2.2629 - val_accuracy: 0.4154\n",
      "Epoch 1872/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7972 - accuracy: 0.5336 - val_loss: 2.2566 - val_accuracy: 0.4162\n",
      "Epoch 1873/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5331 - val_loss: 2.2603 - val_accuracy: 0.4101\n",
      "Epoch 1874/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7990 - accuracy: 0.5326 - val_loss: 2.2614 - val_accuracy: 0.4159\n",
      "Epoch 1875/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7964 - accuracy: 0.5352 - val_loss: 2.2571 - val_accuracy: 0.4148\n",
      "Epoch 1876/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7972 - accuracy: 0.5335 - val_loss: 2.2622 - val_accuracy: 0.4139\n",
      "Epoch 1877/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.8002 - accuracy: 0.5338 - val_loss: 2.2599 - val_accuracy: 0.4133\n",
      "Epoch 1878/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7934 - accuracy: 0.5345 - val_loss: 2.2631 - val_accuracy: 0.4109\n",
      "Epoch 1879/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7989 - accuracy: 0.5333 - val_loss: 2.2621 - val_accuracy: 0.4128\n",
      "Epoch 1880/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5336 - val_loss: 2.2671 - val_accuracy: 0.4071\n",
      "Epoch 1881/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7980 - accuracy: 0.5330 - val_loss: 2.2605 - val_accuracy: 0.4111\n",
      "Epoch 1882/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7956 - accuracy: 0.5343 - val_loss: 2.2606 - val_accuracy: 0.4123\n",
      "Epoch 1883/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7955 - accuracy: 0.5347 - val_loss: 2.2640 - val_accuracy: 0.4157\n",
      "Epoch 1884/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7950 - accuracy: 0.5341 - val_loss: 2.2660 - val_accuracy: 0.4115\n",
      "Epoch 1885/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7982 - accuracy: 0.5329 - val_loss: 2.2626 - val_accuracy: 0.4084\n",
      "Epoch 1886/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7945 - accuracy: 0.5346 - val_loss: 2.2665 - val_accuracy: 0.4116\n",
      "Epoch 1887/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5345 - val_loss: 2.2668 - val_accuracy: 0.4092\n",
      "Epoch 1888/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7970 - accuracy: 0.5342 - val_loss: 2.2665 - val_accuracy: 0.4130\n",
      "Epoch 1889/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7971 - accuracy: 0.5340 - val_loss: 2.2530 - val_accuracy: 0.4136\n",
      "Epoch 1890/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7970 - accuracy: 0.5340 - val_loss: 2.2588 - val_accuracy: 0.4115\n",
      "Epoch 1891/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7932 - accuracy: 0.5352 - val_loss: 2.2568 - val_accuracy: 0.4111\n",
      "Epoch 1892/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5319 - val_loss: 2.2560 - val_accuracy: 0.4155\n",
      "Epoch 1893/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5340 - val_loss: 2.2614 - val_accuracy: 0.4117\n",
      "Epoch 1894/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7954 - accuracy: 0.5339 - val_loss: 2.2630 - val_accuracy: 0.4083\n",
      "Epoch 1895/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7999 - accuracy: 0.5320 - val_loss: 2.2629 - val_accuracy: 0.4125\n",
      "Epoch 1896/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7994 - accuracy: 0.5333 - val_loss: 2.2610 - val_accuracy: 0.4131\n",
      "Epoch 1897/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7947 - accuracy: 0.5350 - val_loss: 2.2630 - val_accuracy: 0.4148\n",
      "Epoch 1898/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7977 - accuracy: 0.5337 - val_loss: 2.2593 - val_accuracy: 0.4127\n",
      "Epoch 1899/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.8000 - accuracy: 0.5327 - val_loss: 2.2570 - val_accuracy: 0.4121\n",
      "Epoch 1900/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7968 - accuracy: 0.5347 - val_loss: 2.2582 - val_accuracy: 0.4132\n",
      "Epoch 1901/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5337 - val_loss: 2.2604 - val_accuracy: 0.4130\n",
      "Epoch 1902/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5332 - val_loss: 2.2629 - val_accuracy: 0.4118\n",
      "Epoch 1903/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7985 - accuracy: 0.5332 - val_loss: 2.2574 - val_accuracy: 0.4121\n",
      "Epoch 1904/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7962 - accuracy: 0.5343 - val_loss: 2.2609 - val_accuracy: 0.4124\n",
      "Epoch 1905/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7962 - accuracy: 0.5346 - val_loss: 2.2628 - val_accuracy: 0.4127\n",
      "Epoch 1906/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7959 - accuracy: 0.5347 - val_loss: 2.2687 - val_accuracy: 0.4097\n",
      "Epoch 1907/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7998 - accuracy: 0.5343 - val_loss: 2.2636 - val_accuracy: 0.4146\n",
      "Epoch 1908/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7981 - accuracy: 0.5327 - val_loss: 2.2620 - val_accuracy: 0.4130\n",
      "Epoch 1909/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7953 - accuracy: 0.5335 - val_loss: 2.2626 - val_accuracy: 0.4145\n",
      "Epoch 1910/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5338 - val_loss: 2.2579 - val_accuracy: 0.4134\n",
      "Epoch 1911/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5334 - val_loss: 2.2614 - val_accuracy: 0.4098\n",
      "Epoch 1912/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5336 - val_loss: 2.2654 - val_accuracy: 0.4115\n",
      "Epoch 1913/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5340 - val_loss: 2.2669 - val_accuracy: 0.4139\n",
      "Epoch 1914/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7986 - accuracy: 0.5323 - val_loss: 2.2679 - val_accuracy: 0.4133\n",
      "Epoch 1915/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7952 - accuracy: 0.5360 - val_loss: 2.2656 - val_accuracy: 0.4142\n",
      "Epoch 1916/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5339 - val_loss: 2.2604 - val_accuracy: 0.4141\n",
      "Epoch 1917/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7974 - accuracy: 0.5340 - val_loss: 2.2593 - val_accuracy: 0.4148\n",
      "Epoch 1918/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5337 - val_loss: 2.2633 - val_accuracy: 0.4125\n",
      "Epoch 1919/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7974 - accuracy: 0.5324 - val_loss: 2.2622 - val_accuracy: 0.4131\n",
      "Epoch 1920/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7968 - accuracy: 0.5343 - val_loss: 2.2590 - val_accuracy: 0.4135\n",
      "Epoch 1921/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7968 - accuracy: 0.5349 - val_loss: 2.2610 - val_accuracy: 0.4119\n",
      "Epoch 1922/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5341 - val_loss: 2.2599 - val_accuracy: 0.4149\n",
      "Epoch 1923/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7957 - accuracy: 0.5340 - val_loss: 2.2571 - val_accuracy: 0.4130\n",
      "Epoch 1924/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5331 - val_loss: 2.2546 - val_accuracy: 0.4143\n",
      "Epoch 1925/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7966 - accuracy: 0.5338 - val_loss: 2.2573 - val_accuracy: 0.4122\n",
      "Epoch 1926/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7966 - accuracy: 0.5341 - val_loss: 2.2585 - val_accuracy: 0.4126\n",
      "Epoch 1927/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7965 - accuracy: 0.5337 - val_loss: 2.2592 - val_accuracy: 0.4139\n",
      "Epoch 1928/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7993 - accuracy: 0.5319 - val_loss: 2.2590 - val_accuracy: 0.4136\n",
      "Epoch 1929/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7953 - accuracy: 0.5339 - val_loss: 2.2612 - val_accuracy: 0.4101\n",
      "Epoch 1930/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5338 - val_loss: 2.2655 - val_accuracy: 0.4123\n",
      "Epoch 1931/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7975 - accuracy: 0.5338 - val_loss: 2.2588 - val_accuracy: 0.4168\n",
      "Epoch 1932/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7967 - accuracy: 0.5336 - val_loss: 2.2619 - val_accuracy: 0.4095\n",
      "Epoch 1933/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7954 - accuracy: 0.5330 - val_loss: 2.2624 - val_accuracy: 0.4133\n",
      "Epoch 1934/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7947 - accuracy: 0.5343 - val_loss: 2.2606 - val_accuracy: 0.4123\n",
      "Epoch 1935/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7979 - accuracy: 0.5333 - val_loss: 2.2600 - val_accuracy: 0.4107\n",
      "Epoch 1936/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7942 - accuracy: 0.5345 - val_loss: 2.2571 - val_accuracy: 0.4114\n",
      "Epoch 1937/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7957 - accuracy: 0.5334 - val_loss: 2.2625 - val_accuracy: 0.4102\n",
      "Epoch 1938/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7976 - accuracy: 0.5328 - val_loss: 2.2600 - val_accuracy: 0.4110\n",
      "Epoch 1939/2000\n",
      "242/242 [==============================] - 5s 20ms/step - loss: 1.7957 - accuracy: 0.5342 - val_loss: 2.2694 - val_accuracy: 0.4103\n",
      "Epoch 1940/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5338 - val_loss: 2.2656 - val_accuracy: 0.4113\n",
      "Epoch 1941/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7963 - accuracy: 0.5343 - val_loss: 2.2568 - val_accuracy: 0.4133\n",
      "Epoch 1942/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7936 - accuracy: 0.5359 - val_loss: 2.2556 - val_accuracy: 0.4121\n",
      "Epoch 1943/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7953 - accuracy: 0.5346 - val_loss: 2.2571 - val_accuracy: 0.4161\n",
      "Epoch 1944/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7971 - accuracy: 0.5329 - val_loss: 2.2595 - val_accuracy: 0.4175\n",
      "Epoch 1945/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5334 - val_loss: 2.2562 - val_accuracy: 0.4127\n",
      "Epoch 1946/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7982 - accuracy: 0.5341 - val_loss: 2.2628 - val_accuracy: 0.4138\n",
      "Epoch 1947/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7962 - accuracy: 0.5343 - val_loss: 2.2595 - val_accuracy: 0.4136\n",
      "Epoch 1948/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5325 - val_loss: 2.2613 - val_accuracy: 0.4124\n",
      "Epoch 1949/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7965 - accuracy: 0.5335 - val_loss: 2.2614 - val_accuracy: 0.4140\n",
      "Epoch 1950/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5331 - val_loss: 2.2657 - val_accuracy: 0.4134\n",
      "Epoch 1951/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7965 - accuracy: 0.5341 - val_loss: 2.2645 - val_accuracy: 0.4130\n",
      "Epoch 1952/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7997 - accuracy: 0.5330 - val_loss: 2.2631 - val_accuracy: 0.4115\n",
      "Epoch 1953/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7944 - accuracy: 0.5343 - val_loss: 2.2556 - val_accuracy: 0.4138\n",
      "Epoch 1954/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7957 - accuracy: 0.5339 - val_loss: 2.2667 - val_accuracy: 0.4111\n",
      "Epoch 1955/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7980 - accuracy: 0.5342 - val_loss: 2.2631 - val_accuracy: 0.4120\n",
      "Epoch 1956/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7969 - accuracy: 0.5338 - val_loss: 2.2620 - val_accuracy: 0.4108\n",
      "Epoch 1957/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7984 - accuracy: 0.5335 - val_loss: 2.2609 - val_accuracy: 0.4119\n",
      "Epoch 1958/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7932 - accuracy: 0.5354 - val_loss: 2.2629 - val_accuracy: 0.4143\n",
      "Epoch 1959/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7944 - accuracy: 0.5344 - val_loss: 2.2663 - val_accuracy: 0.4156\n",
      "Epoch 1960/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7973 - accuracy: 0.5327 - val_loss: 2.2663 - val_accuracy: 0.4104\n",
      "Epoch 1961/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7977 - accuracy: 0.5338 - val_loss: 2.2602 - val_accuracy: 0.4088\n",
      "Epoch 1962/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7949 - accuracy: 0.5337 - val_loss: 2.2643 - val_accuracy: 0.4107\n",
      "Epoch 1963/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7963 - accuracy: 0.5339 - val_loss: 2.2643 - val_accuracy: 0.4139\n",
      "Epoch 1964/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7929 - accuracy: 0.5347 - val_loss: 2.2697 - val_accuracy: 0.4107\n",
      "Epoch 1965/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7958 - accuracy: 0.5342 - val_loss: 2.2634 - val_accuracy: 0.4133\n",
      "Epoch 1966/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7962 - accuracy: 0.5351 - val_loss: 2.2558 - val_accuracy: 0.4148\n",
      "Epoch 1967/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7959 - accuracy: 0.5344 - val_loss: 2.2674 - val_accuracy: 0.4131\n",
      "Epoch 1968/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7933 - accuracy: 0.5344 - val_loss: 2.2585 - val_accuracy: 0.4127\n",
      "Epoch 1969/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7985 - accuracy: 0.5321 - val_loss: 2.2564 - val_accuracy: 0.4133\n",
      "Epoch 1970/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7977 - accuracy: 0.5329 - val_loss: 2.2568 - val_accuracy: 0.4149\n",
      "Epoch 1971/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7978 - accuracy: 0.5338 - val_loss: 2.2539 - val_accuracy: 0.4135\n",
      "Epoch 1972/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7944 - accuracy: 0.5335 - val_loss: 2.2569 - val_accuracy: 0.4141\n",
      "Epoch 1973/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7947 - accuracy: 0.5344 - val_loss: 2.2641 - val_accuracy: 0.4123\n",
      "Epoch 1974/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7970 - accuracy: 0.5339 - val_loss: 2.2545 - val_accuracy: 0.4193\n",
      "Epoch 1975/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7950 - accuracy: 0.5338 - val_loss: 2.2588 - val_accuracy: 0.4163\n",
      "Epoch 1976/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7964 - accuracy: 0.5342 - val_loss: 2.2648 - val_accuracy: 0.4143\n",
      "Epoch 1977/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5337 - val_loss: 2.2645 - val_accuracy: 0.4130\n",
      "Epoch 1978/2000\n",
      "242/242 [==============================] - 5s 22ms/step - loss: 1.7920 - accuracy: 0.5355 - val_loss: 2.2586 - val_accuracy: 0.4186\n",
      "Epoch 1979/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7950 - accuracy: 0.5352 - val_loss: 2.2585 - val_accuracy: 0.4165\n",
      "Epoch 1980/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7954 - accuracy: 0.5341 - val_loss: 2.2538 - val_accuracy: 0.4153\n",
      "Epoch 1981/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7938 - accuracy: 0.5348 - val_loss: 2.2573 - val_accuracy: 0.4144\n",
      "Epoch 1982/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7972 - accuracy: 0.5329 - val_loss: 2.2589 - val_accuracy: 0.4136\n",
      "Epoch 1983/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7941 - accuracy: 0.5343 - val_loss: 2.2567 - val_accuracy: 0.4119\n",
      "Epoch 1984/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7962 - accuracy: 0.5332 - val_loss: 2.2545 - val_accuracy: 0.4142\n",
      "Epoch 1985/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7953 - accuracy: 0.5348 - val_loss: 2.2590 - val_accuracy: 0.4128\n",
      "Epoch 1986/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7975 - accuracy: 0.5332 - val_loss: 2.2574 - val_accuracy: 0.4130\n",
      "Epoch 1987/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7953 - accuracy: 0.5350 - val_loss: 2.2585 - val_accuracy: 0.4132\n",
      "Epoch 1988/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7947 - accuracy: 0.5350 - val_loss: 2.2588 - val_accuracy: 0.4148\n",
      "Epoch 1989/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7936 - accuracy: 0.5355 - val_loss: 2.2599 - val_accuracy: 0.4107\n",
      "Epoch 1990/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7983 - accuracy: 0.5328 - val_loss: 2.2644 - val_accuracy: 0.4098\n",
      "Epoch 1991/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7974 - accuracy: 0.5337 - val_loss: 2.2593 - val_accuracy: 0.4133\n",
      "Epoch 1992/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7931 - accuracy: 0.5352 - val_loss: 2.2604 - val_accuracy: 0.4155\n",
      "Epoch 1993/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7946 - accuracy: 0.5335 - val_loss: 2.2577 - val_accuracy: 0.4109\n",
      "Epoch 1994/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7938 - accuracy: 0.5345 - val_loss: 2.2542 - val_accuracy: 0.4138\n",
      "Epoch 1995/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7940 - accuracy: 0.5346 - val_loss: 2.2602 - val_accuracy: 0.4133\n",
      "Epoch 1996/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7955 - accuracy: 0.5345 - val_loss: 2.2617 - val_accuracy: 0.4107\n",
      "Epoch 1997/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7958 - accuracy: 0.5338 - val_loss: 2.2677 - val_accuracy: 0.4099\n",
      "Epoch 1998/2000\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7947 - accuracy: 0.5338 - val_loss: 2.2640 - val_accuracy: 0.4120\n",
      "Epoch 1999/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7920 - accuracy: 0.5358 - val_loss: 2.2681 - val_accuracy: 0.4104\n",
      "Epoch 2000/2000\n",
      "242/242 [==============================] - 5s 21ms/step - loss: 1.7944 - accuracy: 0.5352 - val_loss: 2.2645 - val_accuracy: 0.4072\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x4c9546c10>"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history = model1.fit(fine_tune_data_train_no_pads, fine_tune_labels_train_no_pads, epochs=EPOCHS, batch_size=1003, validation_data=(fine_tune_data_test_no_pads, fine_tune_labels_test_no_pads), validation_batch_size=1000, callbacks=[model_checkpoint_callback])\n",
    "model1.save_weights(\"./Models/ResNet-VGG-Inception-ResNet-RoI/resnet-vgg-inception-resnet-roi-model-weights\")\n",
    "model1.load_weights('./Models/ResNet-VGG-Inception-ResNet-RoI/resnet-vgg-inception-resnet-roi-model-weights')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "b1f883d7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.training.tracking.util.CheckpointLoadStatus at 0x4c76d0100>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1.load_weights('./Models/ResNet-VGG-Inception-ResNet-RoI/auto-ResNet-VGG-Inception-ResNet-FT-model-weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "9116a6a6",
   "metadata": {},
   "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": 64,
   "id": "c55b5870",
   "metadata": {},
   "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, 'ResNet-VGG-Inception-ResNet-RoI 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": 65,
   "id": "cf7ded38",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  33/7586 [..............................] - ETA: 24s "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-11 01:27:16.156232: 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": [
      "7586/7586 [==============================] - 23s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "train_preds_no_pads = model1.predict(fine_tune_data_train_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "f2a28269",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pred_labels_no_pads = np.argmax(train_preds_no_pads, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "768775b2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 1, 1, ..., 9, 9, 9])"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_pred_labels_no_pads"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "636c565d",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pred_labels = list()\n",
    "k = 0\n",
    "for i in range(len(roi_count_train)):\n",
    "    train_pred_labels.append(list(train_pred_labels_no_pads[k:(k+roi_count_train[i])]))\n",
    "    k += roi_count_train[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "b92d8192",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_pred_labels = np.array([pd.value_counts(ens_pred).index[0] if len(ens_pred) > 0 else -1 for ens_pred in train_pred_labels])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "eff52a3d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "14356"
      ]
     },
     "execution_count": 109,
     "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": 110,
   "id": "8f594a33",
   "metadata": {},
   "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>11</th>\n",
       "      <td>11</td>\n",
       "      <td>./Data/train/train/11.tif</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>./Data/train/train/15.tif</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>./Data/train/train/16.tif</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>22</td>\n",
       "      <td>./Data/train/train/22.tif</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>30</td>\n",
       "      <td>./Data/train/train/30.tif</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15151</th>\n",
       "      <td>15950</td>\n",
       "      <td>./Data/train/train/15950.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15159</th>\n",
       "      <td>15958</td>\n",
       "      <td>./Data/train/train/15958.tif</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15166</th>\n",
       "      <td>15965</td>\n",
       "      <td>./Data/train/train/15965.tif</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15177</th>\n",
       "      <td>15976</td>\n",
       "      <td>./Data/train/train/15976.tif</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15182</th>\n",
       "      <td>15981</td>\n",
       "      <td>./Data/train/train/15981.tif</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>844 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "          id                        images  label\n",
       "11        11     ./Data/train/train/11.tif      1\n",
       "15        15     ./Data/train/train/15.tif      8\n",
       "16        16     ./Data/train/train/16.tif     10\n",
       "21        22     ./Data/train/train/22.tif      8\n",
       "28        30     ./Data/train/train/30.tif      8\n",
       "...      ...                           ...    ...\n",
       "15151  15950  ./Data/train/train/15950.tif     13\n",
       "15159  15958  ./Data/train/train/15958.tif     11\n",
       "15166  15965  ./Data/train/train/15965.tif      5\n",
       "15177  15976  ./Data/train/train/15976.tif      2\n",
       "15182  15981  ./Data/train/train/15981.tif      5\n",
       "\n",
       "[844 rows x 3 columns]"
      ]
     },
     "execution_count": 110,
     "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": 111,
   "id": "2ea8f949",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/rishideychowdhury/Desktop/Document-Classification/env/lib/python3.8/site-packages/sklearn/metrics/_classification.py:1334: UndefinedMetricWarning: Recall and F-score are ill-defined and being set to 0.0 in labels with no true samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "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>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.749214</td>\n",
       "      <td>0.974438</td>\n",
       "      <td>0.847111</td>\n",
       "      <td>978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.967462</td>\n",
       "      <td>0.964324</td>\n",
       "      <td>0.965891</td>\n",
       "      <td>925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.972281</td>\n",
       "      <td>0.985946</td>\n",
       "      <td>0.979066</td>\n",
       "      <td>925</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.906931</td>\n",
       "      <td>0.967265</td>\n",
       "      <td>0.936127</td>\n",
       "      <td>947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.985539</td>\n",
       "      <td>0.952688</td>\n",
       "      <td>0.968835</td>\n",
       "      <td>930</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.973893</td>\n",
       "      <td>0.891892</td>\n",
       "      <td>0.931091</td>\n",
       "      <td>962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.942693</td>\n",
       "      <td>0.988978</td>\n",
       "      <td>0.965281</td>\n",
       "      <td>998</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.984683</td>\n",
       "      <td>0.985761</td>\n",
       "      <td>0.985222</td>\n",
       "      <td>913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.987179</td>\n",
       "      <td>0.766169</td>\n",
       "      <td>0.862745</td>\n",
       "      <td>1005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.969664</td>\n",
       "      <td>0.939140</td>\n",
       "      <td>0.954158</td>\n",
       "      <td>953</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.971302</td>\n",
       "      <td>0.960699</td>\n",
       "      <td>0.965971</td>\n",
       "      <td>916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.985027</td>\n",
       "      <td>0.964398</td>\n",
       "      <td>0.974603</td>\n",
       "      <td>955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.974972</td>\n",
       "      <td>0.877175</td>\n",
       "      <td>0.923491</td>\n",
       "      <td>977</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.987885</td>\n",
       "      <td>0.951220</td>\n",
       "      <td>0.969206</td>\n",
       "      <td>943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.920973</td>\n",
       "      <td>0.979526</td>\n",
       "      <td>0.949347</td>\n",
       "      <td>928</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.938962</td>\n",
       "      <td>0.976720</td>\n",
       "      <td>0.957469</td>\n",
       "      <td>945</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    precision    recall    fscore  support\n",
       "0    0.000000  0.000000  0.000000        0\n",
       "1    0.749214  0.974438  0.847111      978\n",
       "2    0.967462  0.964324  0.965891      925\n",
       "3    0.972281  0.985946  0.979066      925\n",
       "4    0.906931  0.967265  0.936127      947\n",
       "5    0.985539  0.952688  0.968835      930\n",
       "6    0.973893  0.891892  0.931091      962\n",
       "7    0.942693  0.988978  0.965281      998\n",
       "8    0.984683  0.985761  0.985222      913\n",
       "9    0.987179  0.766169  0.862745     1005\n",
       "10   0.969664  0.939140  0.954158      953\n",
       "11   0.971302  0.960699  0.965971      916\n",
       "12   0.985027  0.964398  0.974603      955\n",
       "13   0.974972  0.877175  0.923491      977\n",
       "14   0.987885  0.951220  0.969206      943\n",
       "15   0.920973  0.979526  0.949347      928\n",
       "16   0.938962  0.976720  0.957469      945"
      ]
     },
     "execution_count": 111,
     "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": 112,
   "id": "0d1e8f85",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "412/412 [==============================] - 1s 3ms/step\n"
     ]
    }
   ],
   "source": [
    "test_preds = model1.predict(fine_tune_data_test_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "49233509",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_pred_labels_no_pads = np.argmax(test_preds, axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "8f5cba39",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([11,  3,  6, ...,  6,  6,  6])"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_pred_labels_no_pads"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "e765dc12",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_pred_labels = list()\n",
    "k = 0\n",
    "for i in range(len(roi_count_test)):\n",
    "    test_pred_labels.append(list(test_pred_labels_no_pads[k:(k+roi_count_test[i])]))\n",
    "    k += roi_count_test[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "id": "552e84e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_pred_labels = np.array([pd.value_counts(ens_pred).index[0] if len(ens_pred) > 0 else -1 for ens_pred in test_pred_labels])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "id": "ee078822",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "526"
      ]
     },
     "execution_count": 118,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sum(test_pred_labels == test_labels['label'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "c3db9b5c",
   "metadata": {},
   "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>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>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>785</th>\n",
       "      <td>15621</td>\n",
       "      <td>./Data/train/train/15621.tif</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>786</th>\n",
       "      <td>15660</td>\n",
       "      <td>./Data/train/train/15660.tif</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>787</th>\n",
       "      <td>15668</td>\n",
       "      <td>./Data/train/train/15668.tif</td>\n",
       "      <td>11</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>793</th>\n",
       "      <td>15762</td>\n",
       "      <td>./Data/train/train/15762.tif</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>274 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        id                        images  label\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",
       "7      100    ./Data/train/train/100.tif     10\n",
       "..     ...                           ...    ...\n",
       "785  15621  ./Data/train/train/15621.tif     15\n",
       "786  15660  ./Data/train/train/15660.tif      4\n",
       "787  15668  ./Data/train/train/15668.tif     11\n",
       "788  15687  ./Data/train/train/15687.tif     12\n",
       "793  15762  ./Data/train/train/15762.tif     13\n",
       "\n",
       "[274 rows x 3 columns]"
      ]
     },
     "execution_count": 119,
     "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": 120,
   "id": "b538b721",
   "metadata": {},
   "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>0.383838</td>\n",
       "      <td>0.703704</td>\n",
       "      <td>0.496732</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.550000</td>\n",
       "      <td>0.750000</td>\n",
       "      <td>0.634615</td>\n",
       "      <td>44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.948718</td>\n",
       "      <td>0.860465</td>\n",
       "      <td>0.902439</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.578125</td>\n",
       "      <td>0.804348</td>\n",
       "      <td>0.672727</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.766667</td>\n",
       "      <td>0.638889</td>\n",
       "      <td>0.696970</td>\n",
       "      <td>36</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.708333</td>\n",
       "      <td>0.340000</td>\n",
       "      <td>0.459459</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.627119</td>\n",
       "      <td>0.804348</td>\n",
       "      <td>0.704762</td>\n",
       "      <td>46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.816327</td>\n",
       "      <td>0.769231</td>\n",
       "      <td>0.792079</td>\n",
       "      <td>52</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.846154</td>\n",
       "      <td>0.407407</td>\n",
       "      <td>0.550000</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.607843</td>\n",
       "      <td>0.659574</td>\n",
       "      <td>0.632653</td>\n",
       "      <td>47</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.794872</td>\n",
       "      <td>0.534483</td>\n",
       "      <td>0.639175</td>\n",
       "      <td>58</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.680851</td>\n",
       "      <td>0.571429</td>\n",
       "      <td>0.621359</td>\n",
       "      <td>56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.677419</td>\n",
       "      <td>0.488372</td>\n",
       "      <td>0.567568</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.728814</td>\n",
       "      <td>0.671875</td>\n",
       "      <td>0.699187</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.733333</td>\n",
       "      <td>0.859375</td>\n",
       "      <td>0.791367</td>\n",
       "      <td>64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.604167</td>\n",
       "      <td>0.674419</td>\n",
       "      <td>0.637363</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    precision    recall    fscore  support\n",
       "0    0.383838  0.703704  0.496732       54\n",
       "1    0.550000  0.750000  0.634615       44\n",
       "2    0.948718  0.860465  0.902439       43\n",
       "3    0.578125  0.804348  0.672727       46\n",
       "4    0.766667  0.638889  0.696970       36\n",
       "5    0.708333  0.340000  0.459459       50\n",
       "6    0.627119  0.804348  0.704762       46\n",
       "7    0.816327  0.769231  0.792079       52\n",
       "8    0.846154  0.407407  0.550000       54\n",
       "9    0.607843  0.659574  0.632653       47\n",
       "10   0.794872  0.534483  0.639175       58\n",
       "11   0.680851  0.571429  0.621359       56\n",
       "12   0.677419  0.488372  0.567568       43\n",
       "13   0.728814  0.671875  0.699187       64\n",
       "14   0.733333  0.859375  0.791367       64\n",
       "15   0.604167  0.674419  0.637363       43"
      ]
     },
     "execution_count": 120,
     "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": 121,
   "id": "ed4db76c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6561534764075567"
      ]
     },
     "execution_count": 121,
     "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": 122,
   "id": "4b085c0c",
   "metadata": {},
   "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": 122,
     "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": 123,
   "id": "077cf5fe",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 2048)"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_resnet_validation = np.load('./Data/ResNet-kP-validation-precompute/whole_validation_data_precomp.npy')\n",
    "fine_tune_data_resnet_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "id": "f4a09a28",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 512)"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_vgg_validation = np.load('./Data/VGG16-kP-validation-precompute/whole_validation_data_precomp.npy')\n",
    "fine_tune_data_vgg_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "id": "d5825c95",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 1536)"
      ]
     },
     "execution_count": 125,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_inception_resnet_validation = np.load('./Data/Inception-ResNet-kP-validation-precompute/whole_validation_data_precomp.npy')\n",
    "fine_tune_data_inception_resnet_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 270,
   "id": "53635253",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 2048)"
      ]
     },
     "execution_count": 270,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation_resnet = np.load('./Data/ResNet-4P-validation-precompute/whole_validation_data_precomp.npy')\n",
    "fine_tune_data_validation_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "id": "3b5a3b1c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 4096)"
      ]
     },
     "execution_count": 126,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation = np.concatenate([fine_tune_data_resnet_validation, fine_tune_data_vgg_validation, fine_tune_data_inception_resnet_validation], axis=-1)\n",
    "del fine_tune_data_resnet_validation\n",
    "del fine_tune_data_vgg_validation\n",
    "del fine_tune_data_inception_resnet_validation\n",
    "fine_tune_data_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "89b8fbee",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('./Data/ResNet-VGG-Inception-kP-validation-precompute/whole_validation_data_precomp.npy', fine_tune_data_validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "id": "8a7ed761",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 4096)"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation = np.load('./Data/ResNet-VGG-Inception-kP-validation-precompute/whole_validation_data_precomp.npy')\n",
    "fine_tune_data_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "id": "faf9ee38",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 6)"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "roi_info_validation = np.load('./Data/ROI-Extractions/validation/validation_roi_viz.npy')\n",
    "roi_info_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "id": "fa5495ed",
   "metadata": {},
   "outputs": [],
   "source": [
    "roi_count_validation = [np.sum([roi[-1]==1 for roi in doc]) for doc in roi_info_validation] # RoIs per document Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "id": "53764811",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14540, 4096)"
      ]
     },
     "execution_count": 131,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation_no_pads = np.concatenate([fine_tune_data_validation[i,:roi_count_validation[i],:] for i in range(fine_tune_data_validation.shape[0])])\n",
    "fine_tune_data_validation_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "id": "fb220306",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('./Data/ResNet-VGG-Inception-kP-validation-precompute/whole_validation_data_precomp_no_pads.npy', fine_tune_data_validation_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "id": "e41d68f4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14540, 4096)"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_data_validation_no_pads = np.load('./Data/ResNet-VGG-Inception-kP-validation-precompute/whole_validation_data_precomp_no_pads.npy')\n",
    "fine_tune_data_validation_no_pads.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2f51f41f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "id": "8e78b6f1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 512)"
      ]
     },
     "execution_count": 271,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation_vgg = np.load('./Data/VGG16-4P-validation-precompute/whole_validation_data_precomp.npy')\n",
    "# fine_tune_data_validation_vgg.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "id": "88e39e91",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 1536)"
      ]
     },
     "execution_count": 272,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation_inception_resnet = np.load('./Data/Inception-ResNet-4P-validation-precompute/whole_validation_data_precomp.npy')\n",
    "# fine_tune_data_validation_inception_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "id": "13b10856",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 1300)"
      ]
     },
     "execution_count": 273,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation_red_resnet = np.zeros((fine_tune_data_validation_resnet.shape[0], 5, N_COMPS_1))\n",
    "# for i in range(5):\n",
    "#     fine_tune_data_validation_red_resnet[:,i,:] = pcas_resnet[i].transform(fine_tune_data_validation_resnet[:,i,:])\n",
    "# fine_tune_data_validation_resnet = fine_tune_data_validation_red_resnet\n",
    "# del fine_tune_data_validation_red_resnet\n",
    "# fine_tune_data_validation_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "id": "b0a31a0b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 300)"
      ]
     },
     "execution_count": 274,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation_red_vgg = np.zeros((fine_tune_data_validation_vgg.shape[0], 5, N_COMPS_2))\n",
    "# for i in range(5):\n",
    "#     fine_tune_data_validation_red_vgg[:,i,:] = pcas_vgg[i].transform(fine_tune_data_validation_vgg[:,i,:])\n",
    "# fine_tune_data_validation_vgg = fine_tune_data_validation_red_vgg\n",
    "# del fine_tune_data_validation_red_vgg\n",
    "# fine_tune_data_validation_vgg.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "id": "5e12894e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 900)"
      ]
     },
     "execution_count": 275,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation_red_inception_resnet = np.zeros((fine_tune_data_validation_inception_resnet.shape[0], 5, N_COMPS_3))\n",
    "# for i in range(5):\n",
    "#     fine_tune_data_validation_red_inception_resnet[:,i,:] = pcas_inception_resnet[i].transform(fine_tune_data_validation_inception_resnet[:,i,:])\n",
    "# fine_tune_data_validation_inception_resnet = fine_tune_data_validation_red_inception_resnet\n",
    "# del fine_tune_data_validation_red_inception_resnet\n",
    "# fine_tune_data_validation_inception_resnet.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "id": "586ac8df",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 5, 2500)"
      ]
     },
     "execution_count": 276,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# fine_tune_data_validation = np.concatenate([fine_tune_data_validation_resnet, fine_tune_data_validation_vgg, fine_tune_data_validation_inception_resnet], axis=-1)\n",
    "# fine_tune_data_validation.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "id": "db7d1256",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "29/29 [==============================] - 0s 3ms/step\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-09 15:57:30.272902: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    }
   ],
   "source": [
    "# valid_preds = model1.predict(fine_tune_data_validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "id": "18b17d2e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# valid_pred_labels = np.argmax(valid_preds, axis=-1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ee154efd",
   "metadata": {},
   "source": [
    "## Extract the Feature Vectors for the ROIs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "id": "8c34e761",
   "metadata": {},
   "outputs": [],
   "source": [
    "model2 = tf.keras.Model(model1.input, model1.get_layer(index=len(model1.layers)-4).output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "id": "1a59afad",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_2 (InputLayer)        [(None, 4096)]            0         \n",
      "                                                                 \n",
      " flatten_1 (Flatten)         (None, 4096)              0         \n",
      "                                                                 \n",
      " batch_normalization_4 (Batc  (None, 4096)             16384     \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_4 (Dropout)         (None, 4096)              0         \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 2048)              8390656   \n",
      "                                                                 \n",
      " batch_normalization_5 (Batc  (None, 2048)             8192      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_5 (Dropout)         (None, 2048)              0         \n",
      "                                                                 \n",
      " dense_4 (Dense)             (None, 1024)              2098176   \n",
      "                                                                 \n",
      " batch_normalization_6 (Batc  (None, 1024)             4096      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " dropout_6 (Dropout)         (None, 1024)              0         \n",
      "                                                                 \n",
      " dense_5 (Dense)             (None, 512)               524800    \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 11,042,304\n",
      "Trainable params: 11,027,968\n",
      "Non-trainable params: 14,336\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model2.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "id": "443de618",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "  69/7586 [..............................] - ETA: 17s"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-10-11 02:29:21.061007: 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": [
      "7586/7586 [==============================] - 14s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "train_embed = model2.predict(fine_tune_data_train_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "id": "aa0d556c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(242726, 512)"
      ]
     },
     "execution_count": 141,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "6759d7a2",
   "metadata": {},
   "outputs": [],
   "source": [
    "fine_tune_train_embed = list()\n",
    "k = 0\n",
    "for i in range(len(roi_count_train)):\n",
    "    fine_tune_train_embed.append(np.concatenate([train_embed[k:k+roi_count_train[i]], np.zeros((100-roi_count_train[i], 512))])[np.newaxis,:])\n",
    "    k += roi_count_train[i]\n",
    "fine_tune_train_embed = np.concatenate(fine_tune_train_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "74070d9d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(15200, 100, 512)"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_train_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "854b2364",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_train_embed_precomp.npy', fine_tune_train_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "f7877aa9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "412/412 [==============================] - 1s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "test_embed = model2.predict(fine_tune_data_test_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "c2702d14",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13154, 512)"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "1ab98959",
   "metadata": {},
   "outputs": [],
   "source": [
    "fine_tune_test_embed = list()\n",
    "k = 0\n",
    "for i in range(len(roi_count_test)):\n",
    "    fine_tune_test_embed.append(np.concatenate([test_embed[k:k+roi_count_test[i]], np.zeros((100-roi_count_test[i], 512))])[np.newaxis,:])\n",
    "    k += roi_count_test[i]\n",
    "fine_tune_test_embed = np.concatenate(fine_tune_test_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "210df0c6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(800, 100, 512)"
      ]
     },
     "execution_count": 161,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_test_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "a130ce85",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('./Data/ResNet-VGG-Inception-kP-train-precompute/whole_train_test_embed_precomp.npy', fine_tune_test_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "354884fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "455/455 [==============================] - 1s 2ms/step\n"
     ]
    }
   ],
   "source": [
    "validation_embed = model2.predict(fine_tune_data_validation_no_pads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "id": "965cc8c3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(14540, 512)"
      ]
     },
     "execution_count": 166,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "validation_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "id": "ece4ad36",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "900"
      ]
     },
     "execution_count": 168,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(roi_count_validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "id": "ffa5ce99",
   "metadata": {},
   "outputs": [],
   "source": [
    "fine_tune_validation_embed = list()\n",
    "k = 0\n",
    "for i in range(len(roi_count_validation)):\n",
    "    fine_tune_validation_embed.append(np.concatenate([validation_embed[k:k+roi_count_validation[i]], np.zeros((100-roi_count_validation[i], 512))])[np.newaxis,:])\n",
    "    k += roi_count_validation[i]\n",
    "fine_tune_validation_embed = np.concatenate(fine_tune_validation_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "id": "02affb5c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(900, 100, 512)"
      ]
     },
     "execution_count": 170,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fine_tune_validation_embed.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "id": "726bf2ba",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.save('./Data/ResNet-VGG-Inception-kP-validation-precompute/whole_validation_embed_precomp.npy', fine_tune_validation_embed)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "690174cb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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
}