BTC-RNN-Deep-Learning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Recurrent Neural Network - LSTM -  BTC"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Importing Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "plt.style.use(\"ggplot\")\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM, Dense, Dropout"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"BTC-USD.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Setting the datetime index as the date, only selecting the 'Close' column, then only the last 1000 closing prices.\n",
    "df = df.set_index(\"Date\")[['Close']].tail(1000)\n",
    "df = df.set_index(pd.to_datetime(df.index))\n",
    "\n",
    "# Normalizing/Scaling the Data\n",
    "scaler = MinMaxScaler()\n",
    "df = pd.DataFrame(scaler.fit_transform(df), columns=df.columns, index=df.index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>Close</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Date</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2017-03-08</th>\n",
       "      <td>0.011448</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-09</th>\n",
       "      <td>0.013522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-10</th>\n",
       "      <td>0.009655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-11</th>\n",
       "      <td>0.012840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2017-03-12</th>\n",
       "      <td>0.015294</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Close\n",
       "Date                \n",
       "2017-03-08  0.011448\n",
       "2017-03-09  0.013522\n",
       "2017-03-10  0.009655\n",
       "2017-03-11  0.012840\n",
       "2017-03-12  0.015294"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Checking the newly formatted DataFrame\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting the Closing Prices\n",
    "df.plot(figsize=(14,8))\n",
    "plt.title(\"BTC Closing Prices\")\n",
    "plt.ylabel(\"Price (Normalized)\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Functions to prepare the data for LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_sequence(seq, n_steps_in, n_steps_out):\n",
    "    \"\"\"\n",
    "    Splits the univariate time sequence\n",
    "    \"\"\"\n",
    "    X, y = [], []\n",
    "    \n",
    "    for i in range(len(seq)):\n",
    "        end = i + n_steps_in\n",
    "        out_end = end + n_steps_out\n",
    "        \n",
    "        if out_end > len(seq):\n",
    "            break\n",
    "        \n",
    "        seq_x, seq_y = seq[i:end], seq[end:out_end]\n",
    "        \n",
    "        X.append(seq_x)\n",
    "        y.append(seq_y)\n",
    "    \n",
    "    return np.array(X), np.array(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "def visualize_training_results(results):\n",
    "    \"\"\"\n",
    "    Plots the loss and accuracy for the training and testing data\n",
    "    \"\"\"\n",
    "    history = results.history\n",
    "    plt.figure(figsize=(12,4))\n",
    "    plt.plot(history['val_loss'])\n",
    "    plt.plot(history['loss'])\n",
    "    plt.legend(['val_loss', 'loss'])\n",
    "    plt.title('Loss')\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.show()\n",
    "    \n",
    "    plt.figure(figsize=(12,4))\n",
    "    plt.plot(history['val_accuracy'])\n",
    "    plt.plot(history['accuracy'])\n",
    "    plt.legend(['val_accuracy', 'accuracy'])\n",
    "    plt.title('Accuracy')\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Accuracy')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "def layer_maker(n_layers, n_nodes, activation, drop=None, d_rate=.5):\n",
    "    \"\"\"\n",
    "    Creates a specified number of hidden layers for an RNN\n",
    "    Optional: Adds regularization option - the dropout layer to prevent potential overfitting (if necessary)\n",
    "    \"\"\"\n",
    "    \n",
    "    # Creating the specified number of hidden layers with the specified number of nodes\n",
    "    for x in range(1,n_layers+1):\n",
    "        model.add(LSTM(n_nodes, activation=activation, return_sequences=True))\n",
    "\n",
    "        # Adds a Dropout layer after every Nth hidden layer (the 'drop' variable)\n",
    "        try:\n",
    "            if x % drop == 0:\n",
    "                model.add(Dropout(d_rate))\n",
    "        except:\n",
    "            pass"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Splitting the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# How many periods looking back to learn\n",
    "n_per_in  = 30\n",
    "\n",
    "# How many periods to predict\n",
    "n_per_out = 10\n",
    "\n",
    "# Features (in this case it's 1 because there is only one feature: price)\n",
    "n_features = 1\n",
    "\n",
    "# Splitting the data into appropriate sequences\n",
    "X, y = split_sequence(list(df.Close), n_per_in, n_per_out)\n",
    "\n",
    "# Reshaping the X variable from 2D to 3D\n",
    "X = X.reshape((X.shape[0], X.shape[1], n_features))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Modeling - LSTM (RNN)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating the Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_13\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_68 (LSTM)               (None, 30, 30)            3840      \n",
      "_________________________________________________________________\n",
      "lstm_69 (LSTM)               (None, 30, 12)            2064      \n",
      "_________________________________________________________________\n",
      "lstm_70 (LSTM)               (None, 30, 12)            1200      \n",
      "_________________________________________________________________\n",
      "lstm_71 (LSTM)               (None, 30, 12)            1200      \n",
      "_________________________________________________________________\n",
      "lstm_72 (LSTM)               (None, 30, 12)            1200      \n",
      "_________________________________________________________________\n",
      "lstm_73 (LSTM)               (None, 30, 12)            1200      \n",
      "_________________________________________________________________\n",
      "lstm_74 (LSTM)               (None, 30, 12)            1200      \n",
      "_________________________________________________________________\n",
      "lstm_75 (LSTM)               (None, 10)                920       \n",
      "_________________________________________________________________\n",
      "dense_13 (Dense)             (None, 10)                110       \n",
      "=================================================================\n",
      "Total params: 12,934\n",
      "Trainable params: 12,934\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Instatiating the model\n",
    "model = Sequential()\n",
    "\n",
    "# Activation\n",
    "activ = \"softsign\"\n",
    "\n",
    "# Input layer\n",
    "model.add(LSTM(30, activation=activ, return_sequences=True, input_shape=(n_per_in, n_features)))\n",
    "\n",
    "# Hidden layers\n",
    "layer_maker(n_layers=6, n_nodes=12, activation=activ)\n",
    "\n",
    "# Final Hidden layer\n",
    "model.add(LSTM(10, activation=activ))\n",
    "\n",
    "# Output layer\n",
    "model.add(Dense(n_per_out))\n",
    "\n",
    "# Model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Compiling the data with selected specifications\n",
    "model.compile(optimizer='adam', loss='mse', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fitting and Training the RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "collapsed": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 864 samples, validate on 97 samples\n",
      "Epoch 1/800\n",
      "864/864 [==============================] - 6s 7ms/step - loss: 0.0907 - accuracy: 0.0671 - val_loss: 0.0835 - val_accuracy: 0.0515\n",
      "Epoch 2/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 0.0474 - accuracy: 0.0660 - val_loss: 0.0283 - val_accuracy: 0.0515\n",
      "Epoch 3/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0353 - accuracy: 0.0764 - val_loss: 0.0188 - val_accuracy: 0.0722\n",
      "Epoch 4/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0293 - accuracy: 0.0752 - val_loss: 0.0072 - val_accuracy: 0.0515\n",
      "Epoch 5/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0163 - accuracy: 0.1123 - val_loss: 0.0030 - val_accuracy: 0.0722\n",
      "Epoch 6/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0142 - accuracy: 0.0833 - val_loss: 0.0029 - val_accuracy: 0.0515\n",
      "Epoch 7/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0125 - accuracy: 0.0914 - val_loss: 0.0028 - val_accuracy: 0.0619\n",
      "Epoch 8/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0115 - accuracy: 0.0903 - val_loss: 0.0038 - val_accuracy: 0.0515\n",
      "Epoch 9/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0112 - accuracy: 0.0984 - val_loss: 0.0032 - val_accuracy: 0.0722\n",
      "Epoch 10/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0107 - accuracy: 0.0984 - val_loss: 0.0033 - val_accuracy: 0.0722\n",
      "Epoch 11/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0101 - accuracy: 0.0868 - val_loss: 0.0040 - val_accuracy: 0.0722\n",
      "Epoch 12/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0093 - accuracy: 0.1065 - val_loss: 0.0026 - val_accuracy: 0.0722\n",
      "Epoch 13/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0087 - accuracy: 0.0972 - val_loss: 0.0025 - val_accuracy: 0.0722\n",
      "Epoch 14/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0075 - accuracy: 0.0914 - val_loss: 0.0030 - val_accuracy: 0.0722\n",
      "Epoch 15/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 0.0075 - accuracy: 0.1134 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 16/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 0.0069 - accuracy: 0.1007 - val_loss: 0.0030 - val_accuracy: 0.0722\n",
      "Epoch 17/800\n",
      "864/864 [==============================] - 4s 4ms/step - loss: 0.0061 - accuracy: 0.0891 - val_loss: 0.0032 - val_accuracy: 0.0825\n",
      "Epoch 18/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0059 - accuracy: 0.0914 - val_loss: 0.0033 - val_accuracy: 0.0722\n",
      "Epoch 19/800\n",
      "864/864 [==============================] - 4s 5ms/step - loss: 0.0058 - accuracy: 0.1053 - val_loss: 0.0026 - val_accuracy: 0.0515\n",
      "Epoch 20/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 0.0053 - accuracy: 0.0926 - val_loss: 0.0028 - val_accuracy: 0.0619\n",
      "Epoch 21/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0055 - accuracy: 0.0972 - val_loss: 0.0055 - val_accuracy: 0.0515\n",
      "Epoch 22/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0057 - accuracy: 0.0961 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 23/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0050 - accuracy: 0.0903 - val_loss: 0.0029 - val_accuracy: 0.0722\n",
      "Epoch 24/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0050 - accuracy: 0.0914 - val_loss: 0.0029 - val_accuracy: 0.0722\n",
      "Epoch 25/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0048 - accuracy: 0.0961 - val_loss: 0.0032 - val_accuracy: 0.0722\n",
      "Epoch 26/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0047 - accuracy: 0.1053 - val_loss: 0.0030 - val_accuracy: 0.0619\n",
      "Epoch 27/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0051 - accuracy: 0.1030 - val_loss: 0.0042 - val_accuracy: 0.0722\n",
      "Epoch 28/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0049 - accuracy: 0.1169 - val_loss: 0.0028 - val_accuracy: 0.0619\n",
      "Epoch 29/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0047 - accuracy: 0.1100 - val_loss: 0.0047 - val_accuracy: 0.0825\n",
      "Epoch 30/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0048 - accuracy: 0.1539 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 31/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0046 - accuracy: 0.1076 - val_loss: 0.0031 - val_accuracy: 0.0722\n",
      "Epoch 32/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0045 - accuracy: 0.1331 - val_loss: 0.0028 - val_accuracy: 0.0619\n",
      "Epoch 33/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0043 - accuracy: 0.1366 - val_loss: 0.0029 - val_accuracy: 0.0515\n",
      "Epoch 34/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0042 - accuracy: 0.1192 - val_loss: 0.0027 - val_accuracy: 0.0619\n",
      "Epoch 35/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0042 - accuracy: 0.1192 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 36/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0045 - accuracy: 0.1400 - val_loss: 0.0036 - val_accuracy: 0.0309\n",
      "Epoch 37/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0041 - accuracy: 0.1331 - val_loss: 0.0032 - val_accuracy: 0.0515\n",
      "Epoch 38/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0040 - accuracy: 0.1227 - val_loss: 0.0028 - val_accuracy: 0.0412\n",
      "Epoch 39/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0039 - accuracy: 0.1019 - val_loss: 0.0026 - val_accuracy: 0.0412\n",
      "Epoch 40/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0039 - accuracy: 0.1273 - val_loss: 0.0034 - val_accuracy: 0.0309\n",
      "Epoch 41/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0039 - accuracy: 0.1181 - val_loss: 0.0026 - val_accuracy: 0.0515\n",
      "Epoch 42/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0037 - accuracy: 0.1296 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 43/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0043 - accuracy: 0.1088 - val_loss: 0.0034 - val_accuracy: 0.0619\n",
      "Epoch 44/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0041 - accuracy: 0.1435 - val_loss: 0.0026 - val_accuracy: 0.1443\n",
      "Epoch 45/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0038 - accuracy: 0.1273 - val_loss: 0.0031 - val_accuracy: 0.0722\n",
      "Epoch 46/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0038 - accuracy: 0.1215 - val_loss: 0.0027 - val_accuracy: 0.0515\n",
      "Epoch 47/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0035 - accuracy: 0.1250 - val_loss: 0.0028 - val_accuracy: 0.0515\n",
      "Epoch 48/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0033 - accuracy: 0.1331 - val_loss: 0.0028 - val_accuracy: 0.0515\n",
      "Epoch 49/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0032 - accuracy: 0.1285 - val_loss: 0.0027 - val_accuracy: 0.0309\n",
      "Epoch 50/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0034 - accuracy: 0.1435 - val_loss: 0.0028 - val_accuracy: 0.0412\n",
      "Epoch 51/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0032 - accuracy: 0.1181 - val_loss: 0.0029 - val_accuracy: 0.0412\n",
      "Epoch 52/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0033 - accuracy: 0.1424 - val_loss: 0.0029 - val_accuracy: 0.0309\n",
      "Epoch 53/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0035 - accuracy: 0.1296 - val_loss: 0.0029 - val_accuracy: 0.0722\n",
      "Epoch 54/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0030 - accuracy: 0.1516 - val_loss: 0.0027 - val_accuracy: 0.0619\n",
      "Epoch 55/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0029 - accuracy: 0.1296 - val_loss: 0.0028 - val_accuracy: 0.0619\n",
      "Epoch 56/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0028 - accuracy: 0.1354 - val_loss: 0.0028 - val_accuracy: 0.0412\n",
      "Epoch 57/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0028 - accuracy: 0.1343 - val_loss: 0.0045 - val_accuracy: 0.0825\n",
      "Epoch 58/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0029 - accuracy: 0.1412 - val_loss: 0.0029 - val_accuracy: 0.0722\n",
      "Epoch 59/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0027 - accuracy: 0.1574 - val_loss: 0.0028 - val_accuracy: 0.0825\n",
      "Epoch 60/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0026 - accuracy: 0.1296 - val_loss: 0.0029 - val_accuracy: 0.0619\n",
      "Epoch 61/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0027 - accuracy: 0.1574 - val_loss: 0.0031 - val_accuracy: 0.1443\n",
      "Epoch 62/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0025 - accuracy: 0.1678 - val_loss: 0.0033 - val_accuracy: 0.0412\n",
      "Epoch 63/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0027 - accuracy: 0.1400 - val_loss: 0.0029 - val_accuracy: 0.0928\n",
      "Epoch 64/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0026 - accuracy: 0.1562 - val_loss: 0.0034 - val_accuracy: 0.1031\n",
      "Epoch 65/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0025 - accuracy: 0.1493 - val_loss: 0.0028 - val_accuracy: 0.1237\n",
      "Epoch 66/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0025 - accuracy: 0.1262 - val_loss: 0.0039 - val_accuracy: 0.0515\n",
      "Epoch 67/800\n",
      "864/864 [==============================] - 4s 5ms/step - loss: 0.0024 - accuracy: 0.1632 - val_loss: 0.0030 - val_accuracy: 0.2062\n",
      "Epoch 68/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0024 - accuracy: 0.1447 - val_loss: 0.0025 - val_accuracy: 0.1753\n",
      "Epoch 69/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1655 - val_loss: 0.0027 - val_accuracy: 0.0515\n",
      "Epoch 70/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0023 - accuracy: 0.1493 - val_loss: 0.0033 - val_accuracy: 0.0825\n",
      "Epoch 71/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0022 - accuracy: 0.1701 - val_loss: 0.0036 - val_accuracy: 0.1753\n",
      "Epoch 72/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1736 - val_loss: 0.0032 - val_accuracy: 0.1031\n",
      "Epoch 73/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1493 - val_loss: 0.0035 - val_accuracy: 0.1340\n",
      "Epoch 74/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0021 - accuracy: 0.1574 - val_loss: 0.0038 - val_accuracy: 0.0206\n",
      "Epoch 75/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1644 - val_loss: 0.0030 - val_accuracy: 0.1237\n",
      "Epoch 76/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1620 - val_loss: 0.0032 - val_accuracy: 0.0619\n",
      "Epoch 77/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0022 - accuracy: 0.1458 - val_loss: 0.0031 - val_accuracy: 0.0619\n",
      "Epoch 78/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0022 - accuracy: 0.1748 - val_loss: 0.0033 - val_accuracy: 0.0619\n",
      "Epoch 79/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0020 - accuracy: 0.1644 - val_loss: 0.0032 - val_accuracy: 0.0515\n",
      "Epoch 80/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0021 - accuracy: 0.1655 - val_loss: 0.0030 - val_accuracy: 0.0515\n",
      "Epoch 81/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0021 - accuracy: 0.1539 - val_loss: 0.0035 - val_accuracy: 0.0515\n",
      "Epoch 82/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0019 - accuracy: 0.1632 - val_loss: 0.0040 - val_accuracy: 0.0928\n",
      "Epoch 83/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0021 - accuracy: 0.1470 - val_loss: 0.0042 - val_accuracy: 0.1031\n",
      "Epoch 84/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0023 - accuracy: 0.1701 - val_loss: 0.0035 - val_accuracy: 0.0515\n",
      "Epoch 85/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0022 - accuracy: 0.1782 - val_loss: 0.0038 - val_accuracy: 0.1546\n",
      "Epoch 86/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0019 - accuracy: 0.1597 - val_loss: 0.0039 - val_accuracy: 0.2268\n",
      "Epoch 87/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0020 - accuracy: 0.1736 - val_loss: 0.0041 - val_accuracy: 0.0619\n",
      "Epoch 88/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0022 - accuracy: 0.1678 - val_loss: 0.0037 - val_accuracy: 0.0722\n",
      "Epoch 89/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0022 - accuracy: 0.1470 - val_loss: 0.0039 - val_accuracy: 0.0722\n",
      "Epoch 90/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0019 - accuracy: 0.1667 - val_loss: 0.0035 - val_accuracy: 0.1340\n",
      "Epoch 91/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0019 - accuracy: 0.1678 - val_loss: 0.0029 - val_accuracy: 0.0206\n",
      "Epoch 92/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0019 - accuracy: 0.1493 - val_loss: 0.0037 - val_accuracy: 0.0722\n",
      "Epoch 93/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1748 - val_loss: 0.0033 - val_accuracy: 0.0722\n",
      "Epoch 94/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1632 - val_loss: 0.0044 - val_accuracy: 0.2577\n",
      "Epoch 95/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1655 - val_loss: 0.0037 - val_accuracy: 0.1959\n",
      "Epoch 96/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1713 - val_loss: 0.0031 - val_accuracy: 0.1134\n",
      "Epoch 97/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1632 - val_loss: 0.0033 - val_accuracy: 0.0619\n",
      "Epoch 98/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0017 - accuracy: 0.1470 - val_loss: 0.0035 - val_accuracy: 0.1649\n",
      "Epoch 99/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1863 - val_loss: 0.0034 - val_accuracy: 0.0825\n",
      "Epoch 100/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0017 - accuracy: 0.1528 - val_loss: 0.0032 - val_accuracy: 0.0825\n",
      "Epoch 101/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0017 - accuracy: 0.1817 - val_loss: 0.0031 - val_accuracy: 0.1649\n",
      "Epoch 102/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0020 - accuracy: 0.1644 - val_loss: 0.0039 - val_accuracy: 0.1753\n",
      "Epoch 103/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1794 - val_loss: 0.0037 - val_accuracy: 0.0619\n",
      "Epoch 104/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1551 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 105/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1771 - val_loss: 0.0044 - val_accuracy: 0.2062\n",
      "Epoch 106/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0016 - accuracy: 0.1933 - val_loss: 0.0041 - val_accuracy: 0.1031\n",
      "Epoch 107/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1632 - val_loss: 0.0037 - val_accuracy: 0.2371\n",
      "Epoch 108/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1968 - val_loss: 0.0033 - val_accuracy: 0.1443\n",
      "Epoch 109/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0017 - accuracy: 0.1701 - val_loss: 0.0034 - val_accuracy: 0.0619\n",
      "Epoch 110/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1551 - val_loss: 0.0042 - val_accuracy: 0.2784\n",
      "Epoch 111/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0017 - accuracy: 0.2060 - val_loss: 0.0035 - val_accuracy: 0.0825\n",
      "Epoch 112/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1574 - val_loss: 0.0042 - val_accuracy: 0.1856\n",
      "Epoch 113/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1690 - val_loss: 0.0037 - val_accuracy: 0.1856\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 114/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.2106 - val_loss: 0.0034 - val_accuracy: 0.0309\n",
      "Epoch 115/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0016 - accuracy: 0.1736 - val_loss: 0.0034 - val_accuracy: 0.2680\n",
      "Epoch 116/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1806 - val_loss: 0.0038 - val_accuracy: 0.1237\n",
      "Epoch 117/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1968 - val_loss: 0.0033 - val_accuracy: 0.1340\n",
      "Epoch 118/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.1910 - val_loss: 0.0039 - val_accuracy: 0.2680\n",
      "Epoch 119/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.1968 - val_loss: 0.0042 - val_accuracy: 0.2474\n",
      "Epoch 120/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.1736 - val_loss: 0.0039 - val_accuracy: 0.1340\n",
      "Epoch 121/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.2002 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 122/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1933 - val_loss: 0.0040 - val_accuracy: 0.1959\n",
      "Epoch 123/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.1817 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 124/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1690 - val_loss: 0.0040 - val_accuracy: 0.2680\n",
      "Epoch 125/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.2072 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 126/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1921 - val_loss: 0.0041 - val_accuracy: 0.1753\n",
      "Epoch 127/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1748 - val_loss: 0.0036 - val_accuracy: 0.2680\n",
      "Epoch 128/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1898 - val_loss: 0.0033 - val_accuracy: 0.1340\n",
      "Epoch 129/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.2106 - val_loss: 0.0035 - val_accuracy: 0.1237\n",
      "Epoch 130/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.2211 - val_loss: 0.0035 - val_accuracy: 0.0619\n",
      "Epoch 131/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1736 - val_loss: 0.0043 - val_accuracy: 0.2990\n",
      "Epoch 132/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1910 - val_loss: 0.0050 - val_accuracy: 0.2268\n",
      "Epoch 133/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1956 - val_loss: 0.0038 - val_accuracy: 0.2784\n",
      "Epoch 134/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.2257 - val_loss: 0.0033 - val_accuracy: 0.2062\n",
      "Epoch 135/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1875 - val_loss: 0.0042 - val_accuracy: 0.2577\n",
      "Epoch 136/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.1968 - val_loss: 0.0039 - val_accuracy: 0.1031\n",
      "Epoch 137/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.2037 - val_loss: 0.0043 - val_accuracy: 0.1959\n",
      "Epoch 138/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1840 - val_loss: 0.0082 - val_accuracy: 0.2577\n",
      "Epoch 139/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0016 - accuracy: 0.2049 - val_loss: 0.0043 - val_accuracy: 0.2268\n",
      "Epoch 140/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.1794 - val_loss: 0.0047 - val_accuracy: 0.2680\n",
      "Epoch 141/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.2106 - val_loss: 0.0046 - val_accuracy: 0.2577\n",
      "Epoch 142/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0013 - accuracy: 0.2164 - val_loss: 0.0038 - val_accuracy: 0.2268\n",
      "Epoch 143/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.1887 - val_loss: 0.0035 - val_accuracy: 0.1649\n",
      "Epoch 144/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0012 - accuracy: 0.2118 - val_loss: 0.0034 - val_accuracy: 0.2371\n",
      "Epoch 145/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2106 - val_loss: 0.0054 - val_accuracy: 0.2577\n",
      "Epoch 146/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.1944 - val_loss: 0.0036 - val_accuracy: 0.1959\n",
      "Epoch 147/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2141 - val_loss: 0.0061 - val_accuracy: 0.2474\n",
      "Epoch 148/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.1887 - val_loss: 0.0039 - val_accuracy: 0.1753\n",
      "Epoch 149/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.2002 - val_loss: 0.0030 - val_accuracy: 0.2165\n",
      "Epoch 150/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.2049 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 151/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2176 - val_loss: 0.0031 - val_accuracy: 0.2165\n",
      "Epoch 152/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.1863 - val_loss: 0.0031 - val_accuracy: 0.2577\n",
      "Epoch 153/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0014 - accuracy: 0.2106 - val_loss: 0.0036 - val_accuracy: 0.2577\n",
      "Epoch 154/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0018 - accuracy: 0.1782 - val_loss: 0.0028 - val_accuracy: 0.2474\n",
      "Epoch 155/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0017 - accuracy: 0.1921 - val_loss: 0.0027 - val_accuracy: 0.2165\n",
      "Epoch 156/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.1956 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 157/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0013 - accuracy: 0.2176 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 158/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1782 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 159/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0011 - accuracy: 0.2118 - val_loss: 0.0045 - val_accuracy: 0.3093\n",
      "Epoch 160/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0011 - accuracy: 0.2164 - val_loss: 0.0035 - val_accuracy: 0.2577\n",
      "Epoch 161/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1991 - val_loss: 0.0041 - val_accuracy: 0.2268\n",
      "Epoch 162/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1956 - val_loss: 0.0033 - val_accuracy: 0.2371\n",
      "Epoch 163/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2095 - val_loss: 0.0035 - val_accuracy: 0.2474\n",
      "Epoch 164/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.1979 - val_loss: 0.0039 - val_accuracy: 0.2474\n",
      "Epoch 165/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1979 - val_loss: 0.0035 - val_accuracy: 0.2062\n",
      "Epoch 166/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1944 - val_loss: 0.0030 - val_accuracy: 0.2474\n",
      "Epoch 167/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0013 - accuracy: 0.2072 - val_loss: 0.0051 - val_accuracy: 0.1340\n",
      "Epoch 168/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0014 - accuracy: 0.1736 - val_loss: 0.0028 - val_accuracy: 0.2268\n",
      "Epoch 169/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2153 - val_loss: 0.0029 - val_accuracy: 0.2784\n",
      "Epoch 170/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2164 - val_loss: 0.0036 - val_accuracy: 0.2577\n",
      "Epoch 171/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0011 - accuracy: 0.2188 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 172/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1829 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 173/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2245 - val_loss: 0.0030 - val_accuracy: 0.2680\n",
      "Epoch 174/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.1968 - val_loss: 0.0029 - val_accuracy: 0.3196\n",
      "Epoch 175/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2164 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 176/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0010 - accuracy: 0.2106 - val_loss: 0.0037 - val_accuracy: 0.2165\n",
      "Epoch 177/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.1991 - val_loss: 0.0030 - val_accuracy: 0.2165\n",
      "Epoch 178/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0011 - accuracy: 0.2083 - val_loss: 0.0033 - val_accuracy: 0.2165\n",
      "Epoch 179/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.2176 - val_loss: 0.0030 - val_accuracy: 0.2062\n",
      "Epoch 180/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.1933 - val_loss: 0.0030 - val_accuracy: 0.2062\n",
      "Epoch 181/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.7486e-04 - accuracy: 0.1852 - val_loss: 0.0031 - val_accuracy: 0.2577\n",
      "Epoch 182/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.9327e-04 - accuracy: 0.2176 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 183/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.9449e-04 - accuracy: 0.2222 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 184/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.3915e-04 - accuracy: 0.2153 - val_loss: 0.0031 - val_accuracy: 0.3196\n",
      "Epoch 185/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2141 - val_loss: 0.0055 - val_accuracy: 0.2474\n",
      "Epoch 186/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0015 - accuracy: 0.2025 - val_loss: 0.0023 - val_accuracy: 0.2474\n",
      "Epoch 187/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2222 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 188/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2153 - val_loss: 0.0025 - val_accuracy: 0.3196\n",
      "Epoch 189/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.6520e-04 - accuracy: 0.2338 - val_loss: 0.0031 - val_accuracy: 0.3196\n",
      "Epoch 190/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.8436e-04 - accuracy: 0.1863 - val_loss: 0.0035 - val_accuracy: 0.3299\n",
      "Epoch 191/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.8362e-04 - accuracy: 0.2106 - val_loss: 0.0029 - val_accuracy: 0.3093\n",
      "Epoch 192/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.5760e-04 - accuracy: 0.2442 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 193/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.1933 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 194/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.9285e-04 - accuracy: 0.2211 - val_loss: 0.0032 - val_accuracy: 0.2577\n",
      "Epoch 195/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.4669e-04 - accuracy: 0.2130 - val_loss: 0.0029 - val_accuracy: 0.2474\n",
      "Epoch 196/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.1921 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 197/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.4855e-04 - accuracy: 0.2326 - val_loss: 0.0028 - val_accuracy: 0.2577\n",
      "Epoch 198/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2396 - val_loss: 0.0045 - val_accuracy: 0.2165\n",
      "Epoch 199/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0011 - accuracy: 0.2280 - val_loss: 0.0037 - val_accuracy: 0.2784\n",
      "Epoch 200/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.9761e-04 - accuracy: 0.2095 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 201/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.7464e-04 - accuracy: 0.2072 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 202/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.5458e-04 - accuracy: 0.2095 - val_loss: 0.0035 - val_accuracy: 0.2062\n",
      "Epoch 203/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.3166e-04 - accuracy: 0.2303 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 204/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.9721e-04 - accuracy: 0.2338 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 205/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.8234e-04 - accuracy: 0.2326 - val_loss: 0.0035 - val_accuracy: 0.1856\n",
      "Epoch 206/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.0097e-04 - accuracy: 0.1910 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 207/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.8516e-04 - accuracy: 0.2396 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 208/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2488 - val_loss: 0.0024 - val_accuracy: 0.1753\n",
      "Epoch 209/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.1944 - val_loss: 0.0025 - val_accuracy: 0.3196\n",
      "Epoch 210/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.7196e-04 - accuracy: 0.1991 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 211/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.1276e-04 - accuracy: 0.2188 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 212/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.5389e-04 - accuracy: 0.2581 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 213/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.2216e-04 - accuracy: 0.2384 - val_loss: 0.0029 - val_accuracy: 0.2474\n",
      "Epoch 214/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2338 - val_loss: 0.0030 - val_accuracy: 0.2680\n",
      "Epoch 215/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.5951e-04 - accuracy: 0.2523 - val_loss: 0.0033 - val_accuracy: 0.2268\n",
      "Epoch 216/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.5604e-04 - accuracy: 0.2211 - val_loss: 0.0035 - val_accuracy: 0.2165\n",
      "Epoch 217/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.6408e-04 - accuracy: 0.2303 - val_loss: 0.0031 - val_accuracy: 0.2268\n",
      "Epoch 218/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.0955e-04 - accuracy: 0.2292 - val_loss: 0.0037 - val_accuracy: 0.2371\n",
      "Epoch 219/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.2585e-04 - accuracy: 0.2454 - val_loss: 0.0032 - val_accuracy: 0.2680\n",
      "Epoch 220/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.3806e-04 - accuracy: 0.2442 - val_loss: 0.0031 - val_accuracy: 0.3196\n",
      "Epoch 221/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.4507e-04 - accuracy: 0.2662 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 222/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.5512e-04 - accuracy: 0.2292 - val_loss: 0.0035 - val_accuracy: 0.3299\n",
      "Epoch 223/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.6927e-04 - accuracy: 0.2431 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 224/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.2996e-04 - accuracy: 0.2280 - val_loss: 0.0035 - val_accuracy: 0.2577\n",
      "Epoch 225/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 3s 3ms/step - loss: 7.8978e-04 - accuracy: 0.2477 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 226/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.6018e-04 - accuracy: 0.2546 - val_loss: 0.0035 - val_accuracy: 0.2474\n",
      "Epoch 227/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 0.0011 - accuracy: 0.2674 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 228/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0012 - accuracy: 0.2396 - val_loss: 0.0031 - val_accuracy: 0.3402\n",
      "Epoch 229/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2593 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 230/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.9190e-04 - accuracy: 0.2431 - val_loss: 0.0029 - val_accuracy: 0.2680\n",
      "Epoch 231/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.5123e-04 - accuracy: 0.2465 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 232/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0010 - accuracy: 0.2627 - val_loss: 0.0032 - val_accuracy: 0.2268\n",
      "Epoch 233/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.6919e-04 - accuracy: 0.2396 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 234/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.2840e-04 - accuracy: 0.2697 - val_loss: 0.0034 - val_accuracy: 0.3196\n",
      "Epoch 235/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.0177e-04 - accuracy: 0.2697 - val_loss: 0.0033 - val_accuracy: 0.3299\n",
      "Epoch 236/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8744e-04 - accuracy: 0.2558 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 237/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.5329e-04 - accuracy: 0.2662 - val_loss: 0.0030 - val_accuracy: 0.3402\n",
      "Epoch 238/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.2290e-04 - accuracy: 0.2546 - val_loss: 0.0032 - val_accuracy: 0.3299\n",
      "Epoch 239/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.6151e-04 - accuracy: 0.2396 - val_loss: 0.0029 - val_accuracy: 0.2784\n",
      "Epoch 240/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.6304e-04 - accuracy: 0.2454 - val_loss: 0.0029 - val_accuracy: 0.3196\n",
      "Epoch 241/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.6011e-04 - accuracy: 0.2755 - val_loss: 0.0030 - val_accuracy: 0.2577\n",
      "Epoch 242/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.1014e-04 - accuracy: 0.2627 - val_loss: 0.0030 - val_accuracy: 0.2680\n",
      "Epoch 243/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.2561e-04 - accuracy: 0.2593 - val_loss: 0.0027 - val_accuracy: 0.2577\n",
      "Epoch 244/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8158e-04 - accuracy: 0.2882 - val_loss: 0.0035 - val_accuracy: 0.2268\n",
      "Epoch 245/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6924e-04 - accuracy: 0.2546 - val_loss: 0.0029 - val_accuracy: 0.3505\n",
      "Epoch 246/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8329e-04 - accuracy: 0.2697 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 247/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.9679e-04 - accuracy: 0.2674 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 248/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7207e-04 - accuracy: 0.2755 - val_loss: 0.0030 - val_accuracy: 0.3402\n",
      "Epoch 249/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7947e-04 - accuracy: 0.2512 - val_loss: 0.0034 - val_accuracy: 0.3505\n",
      "Epoch 250/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.4722e-04 - accuracy: 0.2917 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 251/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.6212e-04 - accuracy: 0.2743 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 252/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.4834e-04 - accuracy: 0.2697 - val_loss: 0.0036 - val_accuracy: 0.3299\n",
      "Epoch 253/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.9709e-04 - accuracy: 0.2338 - val_loss: 0.0025 - val_accuracy: 0.3299\n",
      "Epoch 254/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.7469e-04 - accuracy: 0.2604 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 255/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.0000e-04 - accuracy: 0.2454 - val_loss: 0.0028 - val_accuracy: 0.2268\n",
      "Epoch 256/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.1304e-04 - accuracy: 0.2535 - val_loss: 0.0030 - val_accuracy: 0.2990\n",
      "Epoch 257/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.9657e-04 - accuracy: 0.2731 - val_loss: 0.0029 - val_accuracy: 0.2990\n",
      "Epoch 258/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.5629e-04 - accuracy: 0.2697 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 259/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.3828e-04 - accuracy: 0.2558 - val_loss: 0.0032 - val_accuracy: 0.2268\n",
      "Epoch 260/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.0237e-04 - accuracy: 0.2766 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 261/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.0266e-04 - accuracy: 0.2569 - val_loss: 0.0031 - val_accuracy: 0.2474\n",
      "Epoch 262/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.7052e-04 - accuracy: 0.2859 - val_loss: 0.0028 - val_accuracy: 0.2784\n",
      "Epoch 263/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.8839e-04 - accuracy: 0.2604 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 264/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6002e-04 - accuracy: 0.2755 - val_loss: 0.0036 - val_accuracy: 0.3402\n",
      "Epoch 265/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.3447e-04 - accuracy: 0.2894 - val_loss: 0.0032 - val_accuracy: 0.3299\n",
      "Epoch 266/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.1526e-04 - accuracy: 0.2535 - val_loss: 0.0034 - val_accuracy: 0.3402\n",
      "Epoch 267/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9813e-04 - accuracy: 0.2778 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 268/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.2816e-04 - accuracy: 0.2650 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 269/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.0866e-04 - accuracy: 0.2778 - val_loss: 0.0036 - val_accuracy: 0.3402\n",
      "Epoch 270/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9509e-04 - accuracy: 0.2836 - val_loss: 0.0037 - val_accuracy: 0.2474\n",
      "Epoch 271/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.8982e-04 - accuracy: 0.2766 - val_loss: 0.0032 - val_accuracy: 0.3505\n",
      "Epoch 272/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.1776e-04 - accuracy: 0.2697 - val_loss: 0.0037 - val_accuracy: 0.2680\n",
      "Epoch 273/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.4064e-04 - accuracy: 0.2650 - val_loss: 0.0031 - val_accuracy: 0.2577\n",
      "Epoch 274/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.2141e-04 - accuracy: 0.2836 - val_loss: 0.0035 - val_accuracy: 0.2680\n",
      "Epoch 275/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7154e-04 - accuracy: 0.2616 - val_loss: 0.0032 - val_accuracy: 0.2165\n",
      "Epoch 276/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6681e-04 - accuracy: 0.2639 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 277/800\n",
      "864/864 [==============================] - 4s 5ms/step - loss: 7.6642e-04 - accuracy: 0.2894 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 278/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.7358e-04 - accuracy: 0.2708 - val_loss: 0.0030 - val_accuracy: 0.3196\n",
      "Epoch 279/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9462e-04 - accuracy: 0.2801 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 280/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 7.2977e-04 - accuracy: 0.2720 - val_loss: 0.0036 - val_accuracy: 0.2268\n",
      "Epoch 281/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.3227e-04 - accuracy: 0.2477 - val_loss: 0.0034 - val_accuracy: 0.2474\n",
      "Epoch 282/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.9619e-04 - accuracy: 0.2894 - val_loss: 0.0034 - val_accuracy: 0.2371\n",
      "Epoch 283/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.3995e-04 - accuracy: 0.2755 - val_loss: 0.0027 - val_accuracy: 0.2887\n",
      "Epoch 284/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.2801e-04 - accuracy: 0.2928 - val_loss: 0.0024 - val_accuracy: 0.2680\n",
      "Epoch 285/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6229e-04 - accuracy: 0.2859 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 286/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9833e-04 - accuracy: 0.2697 - val_loss: 0.0031 - val_accuracy: 0.2990\n",
      "Epoch 287/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7640e-04 - accuracy: 0.2500 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 288/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9276e-04 - accuracy: 0.2963 - val_loss: 0.0030 - val_accuracy: 0.2165\n",
      "Epoch 289/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7630e-04 - accuracy: 0.2755 - val_loss: 0.0027 - val_accuracy: 0.2887\n",
      "Epoch 290/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.4498e-04 - accuracy: 0.2523 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 291/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.9946e-04 - accuracy: 0.2731 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 292/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 6.9294e-04 - accuracy: 0.2870 - val_loss: 0.0031 - val_accuracy: 0.2887\n",
      "Epoch 293/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.0103e-04 - accuracy: 0.2847 - val_loss: 0.0030 - val_accuracy: 0.2784\n",
      "Epoch 294/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.9339e-04 - accuracy: 0.2685 - val_loss: 0.0040 - val_accuracy: 0.2680\n",
      "Epoch 295/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.1000e-04 - accuracy: 0.2685 - val_loss: 0.0034 - val_accuracy: 0.2474\n",
      "Epoch 296/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.2282e-04 - accuracy: 0.2778 - val_loss: 0.0031 - val_accuracy: 0.3402\n",
      "Epoch 297/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.7065e-04 - accuracy: 0.2755 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 298/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.8986e-04 - accuracy: 0.2662 - val_loss: 0.0029 - val_accuracy: 0.3299\n",
      "Epoch 299/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.5274e-04 - accuracy: 0.2789 - val_loss: 0.0032 - val_accuracy: 0.2474\n",
      "Epoch 300/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.2300e-04 - accuracy: 0.2755 - val_loss: 0.0033 - val_accuracy: 0.1856\n",
      "Epoch 301/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6559e-04 - accuracy: 0.2743 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 302/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.6679e-04 - accuracy: 0.2986 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 303/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.2281e-04 - accuracy: 0.2512 - val_loss: 0.0030 - val_accuracy: 0.2268\n",
      "Epoch 304/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8784e-04 - accuracy: 0.2836 - val_loss: 0.0030 - val_accuracy: 0.2784\n",
      "Epoch 305/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0015 - accuracy: 0.2488 - val_loss: 0.0025 - val_accuracy: 0.3093\n",
      "Epoch 306/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 0.0013 - accuracy: 0.2569 - val_loss: 0.0021 - val_accuracy: 0.3093\n",
      "Epoch 307/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 8.8395e-04 - accuracy: 0.2720 - val_loss: 0.0025 - val_accuracy: 0.3093\n",
      "Epoch 308/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.7154e-04 - accuracy: 0.2812 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 309/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.0191e-04 - accuracy: 0.2847 - val_loss: 0.0031 - val_accuracy: 0.2474\n",
      "Epoch 310/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.1565e-04 - accuracy: 0.2812 - val_loss: 0.0030 - val_accuracy: 0.3402\n",
      "Epoch 311/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.8558e-04 - accuracy: 0.2882 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 312/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4891e-04 - accuracy: 0.2882 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 313/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4257e-04 - accuracy: 0.2650 - val_loss: 0.0033 - val_accuracy: 0.3505\n",
      "Epoch 314/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.5440e-04 - accuracy: 0.2824 - val_loss: 0.0034 - val_accuracy: 0.3402\n",
      "Epoch 315/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6806e-04 - accuracy: 0.2859 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 316/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7053e-04 - accuracy: 0.2685 - val_loss: 0.0032 - val_accuracy: 0.3299\n",
      "Epoch 317/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.3526e-04 - accuracy: 0.2766 - val_loss: 0.0033 - val_accuracy: 0.2474\n",
      "Epoch 318/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4852e-04 - accuracy: 0.2951 - val_loss: 0.0034 - val_accuracy: 0.3299\n",
      "Epoch 319/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6101e-04 - accuracy: 0.2650 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 320/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6190e-04 - accuracy: 0.2824 - val_loss: 0.0032 - val_accuracy: 0.2474\n",
      "Epoch 321/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7773e-04 - accuracy: 0.2778 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 322/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4948e-04 - accuracy: 0.2708 - val_loss: 0.0029 - val_accuracy: 0.2887\n",
      "Epoch 323/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.3685e-04 - accuracy: 0.2685 - val_loss: 0.0031 - val_accuracy: 0.2577\n",
      "Epoch 324/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2373e-04 - accuracy: 0.3032 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 325/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4448e-04 - accuracy: 0.2894 - val_loss: 0.0030 - val_accuracy: 0.3402\n",
      "Epoch 326/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6057e-04 - accuracy: 0.2743 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 327/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6801e-04 - accuracy: 0.2685 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 328/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7283e-04 - accuracy: 0.2812 - val_loss: 0.0028 - val_accuracy: 0.2990\n",
      "Epoch 329/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.9200e-04 - accuracy: 0.2789 - val_loss: 0.0030 - val_accuracy: 0.2165\n",
      "Epoch 330/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 6.3513e-04 - accuracy: 0.2731 - val_loss: 0.0033 - val_accuracy: 0.2062\n",
      "Epoch 331/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7794e-04 - accuracy: 0.2720 - val_loss: 0.0029 - val_accuracy: 0.2577\n",
      "Epoch 332/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6520e-04 - accuracy: 0.2905 - val_loss: 0.0030 - val_accuracy: 0.2268\n",
      "Epoch 333/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.5718e-04 - accuracy: 0.2639 - val_loss: 0.0033 - val_accuracy: 0.3402\n",
      "Epoch 334/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2015e-04 - accuracy: 0.2847 - val_loss: 0.0033 - val_accuracy: 0.2371\n",
      "Epoch 335/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 3s 3ms/step - loss: 6.0720e-04 - accuracy: 0.2882 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 336/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.5680e-04 - accuracy: 0.2755 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 337/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.4309e-04 - accuracy: 0.2801 - val_loss: 0.0029 - val_accuracy: 0.2680\n",
      "Epoch 338/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.5283e-04 - accuracy: 0.2720 - val_loss: 0.0030 - val_accuracy: 0.2887\n",
      "Epoch 339/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.0114e-04 - accuracy: 0.2662 - val_loss: 0.0030 - val_accuracy: 0.2577\n",
      "Epoch 340/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8362e-04 - accuracy: 0.2697 - val_loss: 0.0051 - val_accuracy: 0.2577\n",
      "Epoch 341/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0011 - accuracy: 0.2500 - val_loss: 0.0028 - val_accuracy: 0.2268\n",
      "Epoch 342/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.2489e-04 - accuracy: 0.2512 - val_loss: 0.0027 - val_accuracy: 0.1546\n",
      "Epoch 343/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8011e-04 - accuracy: 0.2639 - val_loss: 0.0031 - val_accuracy: 0.2268\n",
      "Epoch 344/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.3043e-04 - accuracy: 0.2743 - val_loss: 0.0028 - val_accuracy: 0.2165\n",
      "Epoch 345/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6131e-04 - accuracy: 0.2859 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 346/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.3565e-04 - accuracy: 0.2812 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 347/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.1887e-04 - accuracy: 0.2951 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 348/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.5540e-04 - accuracy: 0.2755 - val_loss: 0.0031 - val_accuracy: 0.3196\n",
      "Epoch 349/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.1370e-04 - accuracy: 0.2824 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 350/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.1015e-04 - accuracy: 0.2917 - val_loss: 0.0034 - val_accuracy: 0.3505\n",
      "Epoch 351/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.0792e-04 - accuracy: 0.2789 - val_loss: 0.0032 - val_accuracy: 0.2474\n",
      "Epoch 352/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.9545e-04 - accuracy: 0.2766 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 353/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.9852e-04 - accuracy: 0.2778 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 354/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 5.8650e-04 - accuracy: 0.2847 - val_loss: 0.0030 - val_accuracy: 0.2784\n",
      "Epoch 355/800\n",
      "864/864 [==============================] - 4s 4ms/step - loss: 5.8695e-04 - accuracy: 0.2778 - val_loss: 0.0031 - val_accuracy: 0.3093\n",
      "Epoch 356/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 6.0901e-04 - accuracy: 0.2801 - val_loss: 0.0031 - val_accuracy: 0.1753\n",
      "Epoch 357/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.3003e-04 - accuracy: 0.2523 - val_loss: 0.0025 - val_accuracy: 0.2268\n",
      "Epoch 358/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.3596e-04 - accuracy: 0.2731 - val_loss: 0.0030 - val_accuracy: 0.3093\n",
      "Epoch 359/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.1649e-04 - accuracy: 0.2674 - val_loss: 0.0030 - val_accuracy: 0.2784\n",
      "Epoch 360/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.0977e-04 - accuracy: 0.2998 - val_loss: 0.0036 - val_accuracy: 0.3608\n",
      "Epoch 361/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.1170e-04 - accuracy: 0.2801 - val_loss: 0.0032 - val_accuracy: 0.2165\n",
      "Epoch 362/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.2162e-04 - accuracy: 0.2801 - val_loss: 0.0029 - val_accuracy: 0.2990\n",
      "Epoch 363/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7885e-04 - accuracy: 0.2859 - val_loss: 0.0031 - val_accuracy: 0.3093\n",
      "Epoch 364/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0632e-04 - accuracy: 0.2766 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 365/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0683e-04 - accuracy: 0.2859 - val_loss: 0.0030 - val_accuracy: 0.2990\n",
      "Epoch 366/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4015e-04 - accuracy: 0.2870 - val_loss: 0.0040 - val_accuracy: 0.2371\n",
      "Epoch 367/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.4164e-04 - accuracy: 0.2847 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 368/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.2237e-04 - accuracy: 0.2743 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 369/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7744e-04 - accuracy: 0.2674 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 370/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2543e-04 - accuracy: 0.2766 - val_loss: 0.0031 - val_accuracy: 0.2990\n",
      "Epoch 371/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2620e-04 - accuracy: 0.2824 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 372/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.9841e-04 - accuracy: 0.2824 - val_loss: 0.0035 - val_accuracy: 0.2474\n",
      "Epoch 373/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.3867e-04 - accuracy: 0.2755 - val_loss: 0.0034 - val_accuracy: 0.2474\n",
      "Epoch 374/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.9167e-04 - accuracy: 0.2801 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 375/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.8128e-04 - accuracy: 0.2847 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 376/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.8949e-04 - accuracy: 0.2593 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 377/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9887e-04 - accuracy: 0.2743 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 378/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.1864e-04 - accuracy: 0.2593 - val_loss: 0.0028 - val_accuracy: 0.2784\n",
      "Epoch 379/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.5849e-04 - accuracy: 0.2812 - val_loss: 0.0030 - val_accuracy: 0.2784\n",
      "Epoch 380/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0248e-04 - accuracy: 0.2604 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 381/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.8092e-04 - accuracy: 0.2836 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 382/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7129e-04 - accuracy: 0.2928 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 383/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7003e-04 - accuracy: 0.2720 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 384/800\n",
      "864/864 [==============================] - 4s 4ms/step - loss: 6.0684e-04 - accuracy: 0.2789 - val_loss: 0.0034 - val_accuracy: 0.2474\n",
      "Epoch 385/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.2431e-04 - accuracy: 0.2558 - val_loss: 0.0036 - val_accuracy: 0.3299\n",
      "Epoch 386/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.9479e-04 - accuracy: 0.2905 - val_loss: 0.0036 - val_accuracy: 0.2165\n",
      "Epoch 387/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0901e-04 - accuracy: 0.2986 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 388/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7874e-04 - accuracy: 0.2708 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 389/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7724e-04 - accuracy: 0.2963 - val_loss: 0.0032 - val_accuracy: 0.2680\n",
      "Epoch 390/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 3s 3ms/step - loss: 6.7145e-04 - accuracy: 0.2917 - val_loss: 0.0027 - val_accuracy: 0.2165\n",
      "Epoch 391/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.4311e-04 - accuracy: 0.2731 - val_loss: 0.0044 - val_accuracy: 0.2474\n",
      "Epoch 392/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 9.4822e-04 - accuracy: 0.2812 - val_loss: 0.0031 - val_accuracy: 0.2887\n",
      "Epoch 393/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2901e-04 - accuracy: 0.2905 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 394/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.9385e-04 - accuracy: 0.2998 - val_loss: 0.0031 - val_accuracy: 0.2887\n",
      "Epoch 395/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6821e-04 - accuracy: 0.2951 - val_loss: 0.0030 - val_accuracy: 0.3299\n",
      "Epoch 396/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6364e-04 - accuracy: 0.2812 - val_loss: 0.0033 - val_accuracy: 0.3299\n",
      "Epoch 397/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7238e-04 - accuracy: 0.2650 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 398/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7043e-04 - accuracy: 0.2801 - val_loss: 0.0032 - val_accuracy: 0.2680\n",
      "Epoch 399/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0252e-04 - accuracy: 0.2685 - val_loss: 0.0032 - val_accuracy: 0.2371\n",
      "Epoch 400/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 7.6429e-04 - accuracy: 0.2755 - val_loss: 0.0029 - val_accuracy: 0.2887\n",
      "Epoch 401/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 6.2534e-04 - accuracy: 0.2778 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 402/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.5664e-04 - accuracy: 0.2963 - val_loss: 0.0031 - val_accuracy: 0.3402\n",
      "Epoch 403/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.6343e-04 - accuracy: 0.2708 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 404/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.8454e-04 - accuracy: 0.2708 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 405/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 5.5613e-04 - accuracy: 0.3032 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 406/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.6922e-04 - accuracy: 0.2731 - val_loss: 0.0031 - val_accuracy: 0.2680\n",
      "Epoch 407/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7342e-04 - accuracy: 0.2951 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 408/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4636e-04 - accuracy: 0.2870 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 409/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.8300e-04 - accuracy: 0.2951 - val_loss: 0.0032 - val_accuracy: 0.3402\n",
      "Epoch 410/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2095e-04 - accuracy: 0.2755 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 411/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.9032e-04 - accuracy: 0.2940 - val_loss: 0.0031 - val_accuracy: 0.3299\n",
      "Epoch 412/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.4954e-04 - accuracy: 0.2847 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 413/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6074e-04 - accuracy: 0.2870 - val_loss: 0.0031 - val_accuracy: 0.2680\n",
      "Epoch 414/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7459e-04 - accuracy: 0.2766 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 415/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.8296e-04 - accuracy: 0.3009 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 416/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.8018e-04 - accuracy: 0.2766 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 417/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5796e-04 - accuracy: 0.2917 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 418/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6592e-04 - accuracy: 0.2778 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 419/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6030e-04 - accuracy: 0.2512 - val_loss: 0.0044 - val_accuracy: 0.3505\n",
      "Epoch 420/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.5035e-04 - accuracy: 0.2674 - val_loss: 0.0029 - val_accuracy: 0.2887\n",
      "Epoch 421/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2083e-04 - accuracy: 0.2685 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 422/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2604 - val_loss: 0.0023 - val_accuracy: 0.3093\n",
      "Epoch 423/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7193e-04 - accuracy: 0.2708 - val_loss: 0.0029 - val_accuracy: 0.2371\n",
      "Epoch 424/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0576e-04 - accuracy: 0.2928 - val_loss: 0.0029 - val_accuracy: 0.2887\n",
      "Epoch 425/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7945e-04 - accuracy: 0.2894 - val_loss: 0.0029 - val_accuracy: 0.2990\n",
      "Epoch 426/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4691e-04 - accuracy: 0.2812 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 427/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4234e-04 - accuracy: 0.2847 - val_loss: 0.0032 - val_accuracy: 0.2680\n",
      "Epoch 428/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.4950e-04 - accuracy: 0.2720 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 429/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.8086e-04 - accuracy: 0.2963 - val_loss: 0.0031 - val_accuracy: 0.2680\n",
      "Epoch 430/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.5390e-04 - accuracy: 0.2639 - val_loss: 0.0031 - val_accuracy: 0.2990\n",
      "Epoch 431/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7242e-04 - accuracy: 0.3021 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 432/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3488e-04 - accuracy: 0.2569 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 433/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3651e-04 - accuracy: 0.2928 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 434/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4506e-04 - accuracy: 0.2963 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 435/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7285e-04 - accuracy: 0.2870 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 436/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4056e-04 - accuracy: 0.2616 - val_loss: 0.0030 - val_accuracy: 0.2887\n",
      "Epoch 437/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5896e-04 - accuracy: 0.2697 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 438/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3001e-04 - accuracy: 0.2824 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 439/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6919e-04 - accuracy: 0.2812 - val_loss: 0.0031 - val_accuracy: 0.2887\n",
      "Epoch 440/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4788e-04 - accuracy: 0.3067 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 441/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2984e-04 - accuracy: 0.2836 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 442/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4164e-04 - accuracy: 0.2639 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 443/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3823e-04 - accuracy: 0.2998 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 444/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5637e-04 - accuracy: 0.2685 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 445/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5690e-04 - accuracy: 0.2894 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 446/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6304e-04 - accuracy: 0.2789 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 447/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3126e-04 - accuracy: 0.2812 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 448/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2968e-04 - accuracy: 0.2917 - val_loss: 0.0031 - val_accuracy: 0.2990\n",
      "Epoch 449/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2533e-04 - accuracy: 0.2870 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 450/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.4695e-04 - accuracy: 0.2546 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 451/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7358e-04 - accuracy: 0.2963 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 452/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7234e-04 - accuracy: 0.2558 - val_loss: 0.0033 - val_accuracy: 0.3402\n",
      "Epoch 453/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.9633e-04 - accuracy: 0.2789 - val_loss: 0.0031 - val_accuracy: 0.2990\n",
      "Epoch 454/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.2979e-04 - accuracy: 0.2951 - val_loss: 0.0042 - val_accuracy: 0.3299\n",
      "Epoch 455/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.7592e-04 - accuracy: 0.2743 - val_loss: 0.0031 - val_accuracy: 0.2784\n",
      "Epoch 456/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6516e-04 - accuracy: 0.2731 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 457/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3211e-04 - accuracy: 0.2778 - val_loss: 0.0033 - val_accuracy: 0.2062\n",
      "Epoch 458/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3760e-04 - accuracy: 0.3090 - val_loss: 0.0035 - val_accuracy: 0.2268\n",
      "Epoch 459/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6077e-04 - accuracy: 0.2720 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 460/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0678e-04 - accuracy: 0.2975 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 461/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1640e-04 - accuracy: 0.2801 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 462/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2741e-04 - accuracy: 0.2778 - val_loss: 0.0034 - val_accuracy: 0.3196\n",
      "Epoch 463/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3316e-04 - accuracy: 0.2986 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 464/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2141e-04 - accuracy: 0.3021 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 465/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2551e-04 - accuracy: 0.2894 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 466/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3375e-04 - accuracy: 0.2894 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 467/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.6735e-04 - accuracy: 0.3009 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 468/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1861e-04 - accuracy: 0.2859 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 469/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3050e-04 - accuracy: 0.3032 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 470/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1252e-04 - accuracy: 0.2824 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 471/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1480e-04 - accuracy: 0.2917 - val_loss: 0.0033 - val_accuracy: 0.2784\n",
      "Epoch 472/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1150e-04 - accuracy: 0.2766 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 473/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1965e-04 - accuracy: 0.2593 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 474/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0907e-04 - accuracy: 0.2882 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 475/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0727e-04 - accuracy: 0.2905 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 476/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9903e-04 - accuracy: 0.2951 - val_loss: 0.0034 - val_accuracy: 0.3505\n",
      "Epoch 477/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0764e-04 - accuracy: 0.3009 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 478/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0838e-04 - accuracy: 0.3079 - val_loss: 0.0035 - val_accuracy: 0.2577\n",
      "Epoch 479/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2361e-04 - accuracy: 0.2477 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 480/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4488e-04 - accuracy: 0.2963 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 481/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.6932e-04 - accuracy: 0.2523 - val_loss: 0.0037 - val_accuracy: 0.2784\n",
      "Epoch 482/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0021 - accuracy: 0.2535 - val_loss: 0.0078 - val_accuracy: 0.3608\n",
      "Epoch 483/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0027 - accuracy: 0.2083 - val_loss: 0.0028 - val_accuracy: 0.2474\n",
      "Epoch 484/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 0.0012 - accuracy: 0.2280 - val_loss: 0.0030 - val_accuracy: 0.3093\n",
      "Epoch 485/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 8.1927e-04 - accuracy: 0.2697 - val_loss: 0.0028 - val_accuracy: 0.2887\n",
      "Epoch 486/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4913e-04 - accuracy: 0.2512 - val_loss: 0.0028 - val_accuracy: 0.3093\n",
      "Epoch 487/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.6950e-04 - accuracy: 0.2720 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 488/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.4703e-04 - accuracy: 0.2870 - val_loss: 0.0033 - val_accuracy: 0.3299\n",
      "Epoch 489/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.3056e-04 - accuracy: 0.2627 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 490/800\n",
      "864/864 [==============================] - 4s 4ms/step - loss: 5.4225e-04 - accuracy: 0.2801 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 491/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2734e-04 - accuracy: 0.2859 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 492/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1573e-04 - accuracy: 0.2812 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 493/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2277e-04 - accuracy: 0.2685 - val_loss: 0.0032 - val_accuracy: 0.2990\n",
      "Epoch 494/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1386e-04 - accuracy: 0.2986 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 495/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.0416e-04 - accuracy: 0.2685 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 496/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9129e-04 - accuracy: 0.2917 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 497/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1278e-04 - accuracy: 0.2778 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 498/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1239e-04 - accuracy: 0.2685 - val_loss: 0.0032 - val_accuracy: 0.2887\n",
      "Epoch 499/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9323e-04 - accuracy: 0.2685 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 500/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1982e-04 - accuracy: 0.2975 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 501/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1167e-04 - accuracy: 0.3125 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 502/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8696e-04 - accuracy: 0.2581 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 503/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.8783e-04 - accuracy: 0.2847 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 504/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9872e-04 - accuracy: 0.2986 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 505/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.0440e-04 - accuracy: 0.2720 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 506/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8086e-04 - accuracy: 0.2882 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 507/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8212e-04 - accuracy: 0.2650 - val_loss: 0.0034 - val_accuracy: 0.3402\n",
      "Epoch 508/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1087e-04 - accuracy: 0.2836 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 509/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 5.0652e-04 - accuracy: 0.2674 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 510/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.3666e-04 - accuracy: 0.2812 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 511/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0538e-04 - accuracy: 0.2743 - val_loss: 0.0037 - val_accuracy: 0.2680\n",
      "Epoch 512/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8198e-04 - accuracy: 0.2743 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 513/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8860e-04 - accuracy: 0.2708 - val_loss: 0.0035 - val_accuracy: 0.3196\n",
      "Epoch 514/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8892e-04 - accuracy: 0.2616 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 515/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9997e-04 - accuracy: 0.3009 - val_loss: 0.0033 - val_accuracy: 0.2887\n",
      "Epoch 516/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3224e-04 - accuracy: 0.2697 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 517/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3430e-04 - accuracy: 0.2743 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 518/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1762e-04 - accuracy: 0.2755 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 519/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9066e-04 - accuracy: 0.2917 - val_loss: 0.0035 - val_accuracy: 0.3402\n",
      "Epoch 520/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8486e-04 - accuracy: 0.2731 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 521/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9206e-04 - accuracy: 0.2975 - val_loss: 0.0034 - val_accuracy: 0.2680\n",
      "Epoch 522/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.9453e-04 - accuracy: 0.2650 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 523/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9087e-04 - accuracy: 0.3009 - val_loss: 0.0037 - val_accuracy: 0.3196\n",
      "Epoch 524/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.0940e-04 - accuracy: 0.3009 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 525/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9564e-04 - accuracy: 0.2940 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 526/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7602e-04 - accuracy: 0.2639 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 527/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.9214e-04 - accuracy: 0.2940 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 528/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7887e-04 - accuracy: 0.2743 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 529/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.9581e-04 - accuracy: 0.2894 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 530/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1528e-04 - accuracy: 0.2859 - val_loss: 0.0033 - val_accuracy: 0.2577\n",
      "Epoch 531/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2649e-04 - accuracy: 0.2720 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 532/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0633e-04 - accuracy: 0.2951 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 533/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3624e-04 - accuracy: 0.2639 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 534/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0965e-04 - accuracy: 0.2755 - val_loss: 0.0040 - val_accuracy: 0.3402\n",
      "Epoch 535/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0022e-04 - accuracy: 0.2535 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 536/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.2652e-04 - accuracy: 0.2870 - val_loss: 0.0037 - val_accuracy: 0.2784\n",
      "Epoch 537/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7325e-04 - accuracy: 0.2650 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 538/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.1529e-04 - accuracy: 0.2905 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 539/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7959e-04 - accuracy: 0.2477 - val_loss: 0.0035 - val_accuracy: 0.2474\n",
      "Epoch 540/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7354e-04 - accuracy: 0.2986 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 541/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7564e-04 - accuracy: 0.2708 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 542/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8089e-04 - accuracy: 0.2998 - val_loss: 0.0037 - val_accuracy: 0.2680\n",
      "Epoch 543/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7990e-04 - accuracy: 0.2824 - val_loss: 0.0034 - val_accuracy: 0.2784\n",
      "Epoch 544/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6289e-04 - accuracy: 0.3102 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 545/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8358e-04 - accuracy: 0.2720 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 546/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2747e-04 - accuracy: 0.2755 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 547/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3839e-04 - accuracy: 0.2882 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 548/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9070e-04 - accuracy: 0.2859 - val_loss: 0.0037 - val_accuracy: 0.3196\n",
      "Epoch 549/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6913e-04 - accuracy: 0.2812 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 550/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6854e-04 - accuracy: 0.2940 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 551/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6744e-04 - accuracy: 0.3032 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 552/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8193e-04 - accuracy: 0.2859 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 553/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.7170e-04 - accuracy: 0.2708 - val_loss: 0.0035 - val_accuracy: 0.2371\n",
      "Epoch 554/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 7.8486e-04 - accuracy: 0.2789 - val_loss: 0.0031 - val_accuracy: 0.2577\n",
      "Epoch 555/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 6.2833e-04 - accuracy: 0.2674 - val_loss: 0.0035 - val_accuracy: 0.3608\n",
      "Epoch 556/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5313e-04 - accuracy: 0.2593 - val_loss: 0.0035 - val_accuracy: 0.3608\n",
      "Epoch 557/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7656e-04 - accuracy: 0.2847 - val_loss: 0.0032 - val_accuracy: 0.2784\n",
      "Epoch 558/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6196e-04 - accuracy: 0.2662 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 559/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6357e-04 - accuracy: 0.2928 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 560/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7299e-04 - accuracy: 0.2789 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 561/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6938e-04 - accuracy: 0.3009 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 562/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7315e-04 - accuracy: 0.2905 - val_loss: 0.0039 - val_accuracy: 0.3196\n",
      "Epoch 563/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0107e-04 - accuracy: 0.2755 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 564/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5928e-04 - accuracy: 0.2882 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 565/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.5514e-04 - accuracy: 0.2685 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 566/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5384e-04 - accuracy: 0.3148 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 567/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.5518e-04 - accuracy: 0.2905 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 568/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5466e-04 - accuracy: 0.2593 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 569/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.6443e-04 - accuracy: 0.2836 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 570/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.6805e-04 - accuracy: 0.2743 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 571/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9146e-04 - accuracy: 0.2998 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 572/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7735e-04 - accuracy: 0.3067 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 573/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7686e-04 - accuracy: 0.2789 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 574/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7355e-04 - accuracy: 0.2963 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 575/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0679e-04 - accuracy: 0.2940 - val_loss: 0.0036 - val_accuracy: 0.2680\n",
      "Epoch 576/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.7619e-04 - accuracy: 0.2639 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 577/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.6769e-04 - accuracy: 0.2674 - val_loss: 0.0035 - val_accuracy: 0.3196\n",
      "Epoch 578/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 5.6374e-04 - accuracy: 0.2639 - val_loss: 0.0033 - val_accuracy: 0.2990\n",
      "Epoch 579/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.8180e-04 - accuracy: 0.2847 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 580/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7345e-04 - accuracy: 0.2593 - val_loss: 0.0035 - val_accuracy: 0.2680\n",
      "Epoch 581/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5635e-04 - accuracy: 0.2882 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 582/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.6314e-04 - accuracy: 0.2627 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 583/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.6103e-04 - accuracy: 0.2940 - val_loss: 0.0035 - val_accuracy: 0.3196\n",
      "Epoch 584/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6912e-04 - accuracy: 0.2801 - val_loss: 0.0035 - val_accuracy: 0.2577\n",
      "Epoch 585/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.8071e-04 - accuracy: 0.2963 - val_loss: 0.0043 - val_accuracy: 0.3299\n",
      "Epoch 586/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7994e-04 - accuracy: 0.2870 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 587/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.7294e-04 - accuracy: 0.2928 - val_loss: 0.0038 - val_accuracy: 0.3196\n",
      "Epoch 588/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8410e-04 - accuracy: 0.2882 - val_loss: 0.0037 - val_accuracy: 0.2268\n",
      "Epoch 589/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.6574e-04 - accuracy: 0.2870 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 590/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6298e-04 - accuracy: 0.2812 - val_loss: 0.0039 - val_accuracy: 0.3196\n",
      "Epoch 591/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9415e-04 - accuracy: 0.2986 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 592/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5568e-04 - accuracy: 0.2928 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 593/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.8156e-04 - accuracy: 0.2986 - val_loss: 0.0037 - val_accuracy: 0.2680\n",
      "Epoch 594/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.7316e-04 - accuracy: 0.2975 - val_loss: 0.0033 - val_accuracy: 0.2680\n",
      "Epoch 595/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0656e-04 - accuracy: 0.2963 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 596/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9483e-04 - accuracy: 0.2789 - val_loss: 0.0039 - val_accuracy: 0.2165\n",
      "Epoch 597/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9024e-04 - accuracy: 0.2766 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 598/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7253e-04 - accuracy: 0.2674 - val_loss: 0.0037 - val_accuracy: 0.3196\n",
      "Epoch 599/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6047e-04 - accuracy: 0.2986 - val_loss: 0.0038 - val_accuracy: 0.3196\n",
      "Epoch 600/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9221e-04 - accuracy: 0.3067 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 601/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5692e-04 - accuracy: 0.2870 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 602/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5522e-04 - accuracy: 0.2951 - val_loss: 0.0036 - val_accuracy: 0.2577\n",
      "Epoch 603/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7727e-04 - accuracy: 0.2789 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 604/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1182e-04 - accuracy: 0.3079 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 605/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6544e-04 - accuracy: 0.2917 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 606/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4829e-04 - accuracy: 0.2836 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 607/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4081e-04 - accuracy: 0.3009 - val_loss: 0.0038 - val_accuracy: 0.2577\n",
      "Epoch 608/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4035e-04 - accuracy: 0.2847 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 609/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4528e-04 - accuracy: 0.2766 - val_loss: 0.0037 - val_accuracy: 0.3196\n",
      "Epoch 610/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5630e-04 - accuracy: 0.2859 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 611/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4383e-04 - accuracy: 0.2986 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 612/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4331e-04 - accuracy: 0.2616 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 613/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5045e-04 - accuracy: 0.2859 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 614/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.5378e-04 - accuracy: 0.2975 - val_loss: 0.0033 - val_accuracy: 0.3196\n",
      "Epoch 615/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.9361e-04 - accuracy: 0.2604 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 616/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.5471e-04 - accuracy: 0.3102 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 617/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.3559e-04 - accuracy: 0.2975 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 618/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4747e-04 - accuracy: 0.2998 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 619/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3420e-04 - accuracy: 0.2812 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 620/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6781e-04 - accuracy: 0.2963 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 621/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4505e-04 - accuracy: 0.2859 - val_loss: 0.0040 - val_accuracy: 0.3093\n",
      "Epoch 622/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5329e-04 - accuracy: 0.3032 - val_loss: 0.0039 - val_accuracy: 0.3299\n",
      "Epoch 623/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.9429e-04 - accuracy: 0.2940 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 624/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7912e-04 - accuracy: 0.2975 - val_loss: 0.0034 - val_accuracy: 0.3093\n",
      "Epoch 625/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4023e-04 - accuracy: 0.2720 - val_loss: 0.0036 - val_accuracy: 0.2577\n",
      "Epoch 626/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.3724e-04 - accuracy: 0.2986 - val_loss: 0.0038 - val_accuracy: 0.2784\n",
      "Epoch 627/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4473e-04 - accuracy: 0.2940 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 628/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.5230e-04 - accuracy: 0.2951 - val_loss: 0.0038 - val_accuracy: 0.2474\n",
      "Epoch 629/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4404e-04 - accuracy: 0.2604 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 630/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2625e-04 - accuracy: 0.2801 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 631/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.3100e-04 - accuracy: 0.3160 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 632/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2980e-04 - accuracy: 0.2697 - val_loss: 0.0037 - val_accuracy: 0.2062\n",
      "Epoch 633/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.4058e-04 - accuracy: 0.2778 - val_loss: 0.0036 - val_accuracy: 0.3402\n",
      "Epoch 634/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.6227e-04 - accuracy: 0.2593 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 635/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.4897e-04 - accuracy: 0.2928 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 636/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.6994e-04 - accuracy: 0.3032 - val_loss: 0.0038 - val_accuracy: 0.2784\n",
      "Epoch 637/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.4430e-04 - accuracy: 0.3125 - val_loss: 0.0037 - val_accuracy: 0.3402\n",
      "Epoch 638/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3081e-04 - accuracy: 0.2766 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 639/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2438e-04 - accuracy: 0.2581 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 640/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3874e-04 - accuracy: 0.2928 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 641/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2261e-04 - accuracy: 0.2882 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 642/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2411e-04 - accuracy: 0.2847 - val_loss: 0.0038 - val_accuracy: 0.2371\n",
      "Epoch 643/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2410e-04 - accuracy: 0.2708 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 644/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.1976e-04 - accuracy: 0.2870 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 645/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.9284e-04 - accuracy: 0.3113 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 646/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3771e-04 - accuracy: 0.2882 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 647/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.4617e-04 - accuracy: 0.3194 - val_loss: 0.0036 - val_accuracy: 0.2371\n",
      "Epoch 648/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2071e-04 - accuracy: 0.2940 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 649/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 5.2017e-04 - accuracy: 0.2917 - val_loss: 0.0034 - val_accuracy: 0.2990\n",
      "Epoch 650/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3481e-04 - accuracy: 0.2639 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 651/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2506e-04 - accuracy: 0.3021 - val_loss: 0.0038 - val_accuracy: 0.2474\n",
      "Epoch 652/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.4661e-04 - accuracy: 0.2870 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 653/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.9720e-04 - accuracy: 0.2963 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 654/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 4.9394e-04 - accuracy: 0.2882 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 655/800\n",
      "864/864 [==============================] - 4s 4ms/step - loss: 4.8723e-04 - accuracy: 0.2650 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 656/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3607e-04 - accuracy: 0.3079 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 657/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3466e-04 - accuracy: 0.2905 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 658/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.1875e-04 - accuracy: 0.2870 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 659/800\n",
      "864/864 [==============================] - 4s 5ms/step - loss: 4.1596e-04 - accuracy: 0.3113 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 660/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.2212e-04 - accuracy: 0.2697 - val_loss: 0.0037 - val_accuracy: 0.2062\n",
      "Epoch 661/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2727e-04 - accuracy: 0.3044 - val_loss: 0.0040 - val_accuracy: 0.2990\n",
      "Epoch 662/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1491e-04 - accuracy: 0.2847 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 663/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.0653e-04 - accuracy: 0.2894 - val_loss: 0.0048 - val_accuracy: 0.2268\n",
      "Epoch 664/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 9.8719e-04 - accuracy: 0.2801 - val_loss: 0.0032 - val_accuracy: 0.3093\n",
      "Epoch 665/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3990e-04 - accuracy: 0.2801 - val_loss: 0.0036 - val_accuracy: 0.2371\n",
      "Epoch 666/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6940e-04 - accuracy: 0.2928 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 667/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4019e-04 - accuracy: 0.2847 - val_loss: 0.0040 - val_accuracy: 0.2680\n",
      "Epoch 668/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4406e-04 - accuracy: 0.2963 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 669/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2016e-04 - accuracy: 0.2847 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 670/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2103e-04 - accuracy: 0.2743 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 671/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1375e-04 - accuracy: 0.3252 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 672/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0469e-04 - accuracy: 0.3067 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 673/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0949e-04 - accuracy: 0.2905 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 674/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1411e-04 - accuracy: 0.2731 - val_loss: 0.0033 - val_accuracy: 0.3093\n",
      "Epoch 675/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2367e-04 - accuracy: 0.2940 - val_loss: 0.0034 - val_accuracy: 0.2887\n",
      "Epoch 676/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.6034e-04 - accuracy: 0.3102 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 677/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2753e-04 - accuracy: 0.2986 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 678/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1671e-04 - accuracy: 0.2755 - val_loss: 0.0040 - val_accuracy: 0.2165\n",
      "Epoch 679/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2154e-04 - accuracy: 0.3148 - val_loss: 0.0041 - val_accuracy: 0.2268\n",
      "Epoch 680/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0193e-04 - accuracy: 0.2743 - val_loss: 0.0035 - val_accuracy: 0.3093\n",
      "Epoch 681/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0751e-04 - accuracy: 0.2894 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 682/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0902e-04 - accuracy: 0.2870 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 683/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9393e-04 - accuracy: 0.3102 - val_loss: 0.0037 - val_accuracy: 0.2165\n",
      "Epoch 684/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0619e-04 - accuracy: 0.2928 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 685/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9963e-04 - accuracy: 0.3079 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 686/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0841e-04 - accuracy: 0.3090 - val_loss: 0.0040 - val_accuracy: 0.3402\n",
      "Epoch 687/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8833e-04 - accuracy: 0.2928 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 688/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.9121e-04 - accuracy: 0.2882 - val_loss: 0.0039 - val_accuracy: 0.3196\n",
      "Epoch 689/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9130e-04 - accuracy: 0.2998 - val_loss: 0.0038 - val_accuracy: 0.2577\n",
      "Epoch 690/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.1291e-04 - accuracy: 0.2836 - val_loss: 0.0039 - val_accuracy: 0.2165\n",
      "Epoch 691/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.9726e-04 - accuracy: 0.3032 - val_loss: 0.0037 - val_accuracy: 0.2474\n",
      "Epoch 692/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0953e-04 - accuracy: 0.3056 - val_loss: 0.0037 - val_accuracy: 0.2165\n",
      "Epoch 693/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.3948e-04 - accuracy: 0.3125 - val_loss: 0.0038 - val_accuracy: 0.2062\n",
      "Epoch 694/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0589e-04 - accuracy: 0.3183 - val_loss: 0.0039 - val_accuracy: 0.3402\n",
      "Epoch 695/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1748e-04 - accuracy: 0.2789 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 696/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1606e-04 - accuracy: 0.2998 - val_loss: 0.0035 - val_accuracy: 0.2887\n",
      "Epoch 697/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.0774e-04 - accuracy: 0.2697 - val_loss: 0.0040 - val_accuracy: 0.2990\n",
      "Epoch 698/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4166e-04 - accuracy: 0.2697 - val_loss: 0.0035 - val_accuracy: 0.3402\n",
      "Epoch 699/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.1213e-04 - accuracy: 0.2859 - val_loss: 0.0040 - val_accuracy: 0.2062\n",
      "Epoch 700/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4022e-04 - accuracy: 0.3056 - val_loss: 0.0038 - val_accuracy: 0.2474\n",
      "Epoch 701/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1587e-04 - accuracy: 0.3021 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 702/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0036e-04 - accuracy: 0.2685 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 703/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9709e-04 - accuracy: 0.2824 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 704/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0159e-04 - accuracy: 0.3125 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 705/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7832e-04 - accuracy: 0.2812 - val_loss: 0.0039 - val_accuracy: 0.3196\n",
      "Epoch 706/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9265e-04 - accuracy: 0.2824 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 707/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0002e-04 - accuracy: 0.3021 - val_loss: 0.0038 - val_accuracy: 0.2371\n",
      "Epoch 708/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9017e-04 - accuracy: 0.2882 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 709/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1292e-04 - accuracy: 0.3090 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 710/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.5457e-04 - accuracy: 0.2847 - val_loss: 0.0039 - val_accuracy: 0.2680\n",
      "Epoch 711/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.2238e-04 - accuracy: 0.2917 - val_loss: 0.0031 - val_accuracy: 0.2268\n",
      "Epoch 712/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1636e-04 - accuracy: 0.2859 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 713/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8371e-04 - accuracy: 0.3044 - val_loss: 0.0043 - val_accuracy: 0.2990\n",
      "Epoch 714/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2664e-04 - accuracy: 0.2905 - val_loss: 0.0037 - val_accuracy: 0.2577\n",
      "Epoch 715/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 6.4815e-04 - accuracy: 0.2975 - val_loss: 0.0032 - val_accuracy: 0.3196\n",
      "Epoch 716/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 5.3835e-04 - accuracy: 0.2755 - val_loss: 0.0036 - val_accuracy: 0.3196\n",
      "Epoch 717/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0727e-04 - accuracy: 0.2755 - val_loss: 0.0034 - val_accuracy: 0.2268\n",
      "Epoch 718/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1398e-04 - accuracy: 0.2836 - val_loss: 0.0041 - val_accuracy: 0.3299\n",
      "Epoch 719/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4396e-04 - accuracy: 0.3137 - val_loss: 0.0035 - val_accuracy: 0.2165\n",
      "Epoch 720/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 4.7331e-04 - accuracy: 0.3067 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 721/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9090e-04 - accuracy: 0.3056 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 722/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8165e-04 - accuracy: 0.3380 - val_loss: 0.0041 - val_accuracy: 0.2990\n",
      "Epoch 723/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7372e-04 - accuracy: 0.2674 - val_loss: 0.0034 - val_accuracy: 0.2577\n",
      "Epoch 724/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7645e-04 - accuracy: 0.3252 - val_loss: 0.0037 - val_accuracy: 0.3299\n",
      "Epoch 725/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7543e-04 - accuracy: 0.3021 - val_loss: 0.0035 - val_accuracy: 0.2990\n",
      "Epoch 726/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9643e-04 - accuracy: 0.2674 - val_loss: 0.0039 - val_accuracy: 0.3196\n",
      "Epoch 727/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7572e-04 - accuracy: 0.2951 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 728/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8348e-04 - accuracy: 0.3021 - val_loss: 0.0038 - val_accuracy: 0.2784\n",
      "Epoch 729/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1241e-04 - accuracy: 0.2928 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 730/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9848e-04 - accuracy: 0.3275 - val_loss: 0.0040 - val_accuracy: 0.2165\n",
      "Epoch 731/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6542e-04 - accuracy: 0.3067 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 732/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.7188e-04 - accuracy: 0.3009 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 733/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7070e-04 - accuracy: 0.2951 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 734/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.7179e-04 - accuracy: 0.3090 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 735/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7562e-04 - accuracy: 0.2917 - val_loss: 0.0037 - val_accuracy: 0.2680\n",
      "Epoch 736/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5154e-04 - accuracy: 0.3160 - val_loss: 0.0036 - val_accuracy: 0.2474\n",
      "Epoch 737/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9382e-04 - accuracy: 0.2975 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 738/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9332e-04 - accuracy: 0.3079 - val_loss: 0.0037 - val_accuracy: 0.3093\n",
      "Epoch 739/800\n",
      "864/864 [==============================] - 2s 2ms/step - loss: 4.3378e-04 - accuracy: 0.2928 - val_loss: 0.0034 - val_accuracy: 0.2680\n",
      "Epoch 740/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.2357e-04 - accuracy: 0.3079 - val_loss: 0.0039 - val_accuracy: 0.2990\n",
      "Epoch 741/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7575e-04 - accuracy: 0.2836 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 742/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5486e-04 - accuracy: 0.3148 - val_loss: 0.0038 - val_accuracy: 0.2990\n",
      "Epoch 743/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4457e-04 - accuracy: 0.2975 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 744/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4584e-04 - accuracy: 0.2963 - val_loss: 0.0041 - val_accuracy: 0.2990\n",
      "Epoch 745/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5798e-04 - accuracy: 0.3160 - val_loss: 0.0040 - val_accuracy: 0.2990\n",
      "Epoch 746/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8364e-04 - accuracy: 0.3067 - val_loss: 0.0041 - val_accuracy: 0.2887\n",
      "Epoch 747/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5867e-04 - accuracy: 0.3090 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 748/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8475e-04 - accuracy: 0.3090 - val_loss: 0.0040 - val_accuracy: 0.2887\n",
      "Epoch 749/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7988e-04 - accuracy: 0.3044 - val_loss: 0.0038 - val_accuracy: 0.1959\n",
      "Epoch 750/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6304e-04 - accuracy: 0.3241 - val_loss: 0.0041 - val_accuracy: 0.2887\n",
      "Epoch 751/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4861e-04 - accuracy: 0.3113 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 752/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.3538e-04 - accuracy: 0.3056 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 753/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.7113e-04 - accuracy: 0.2917 - val_loss: 0.0041 - val_accuracy: 0.1856\n",
      "Epoch 754/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7467e-04 - accuracy: 0.3194 - val_loss: 0.0039 - val_accuracy: 0.3093\n",
      "Epoch 755/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8871e-04 - accuracy: 0.3160 - val_loss: 0.0042 - val_accuracy: 0.2165\n",
      "Epoch 756/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7942e-04 - accuracy: 0.3044 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 757/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6150e-04 - accuracy: 0.2882 - val_loss: 0.0042 - val_accuracy: 0.2990\n",
      "Epoch 758/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4705e-04 - accuracy: 0.2917 - val_loss: 0.0041 - val_accuracy: 0.2990\n",
      "Epoch 759/800\n",
      "864/864 [==============================] - 3s 4ms/step - loss: 3.7312e-04 - accuracy: 0.3009 - val_loss: 0.0042 - val_accuracy: 0.2680\n",
      "Epoch 760/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6283e-04 - accuracy: 0.2928 - val_loss: 0.0037 - val_accuracy: 0.2990\n",
      "Epoch 761/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4973e-04 - accuracy: 0.3021 - val_loss: 0.0040 - val_accuracy: 0.2990\n",
      "Epoch 762/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.4726e-04 - accuracy: 0.3160 - val_loss: 0.0038 - val_accuracy: 0.3093\n",
      "Epoch 763/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8368e-04 - accuracy: 0.2986 - val_loss: 0.0038 - val_accuracy: 0.2887\n",
      "Epoch 764/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.8056e-04 - accuracy: 0.3009 - val_loss: 0.0041 - val_accuracy: 0.2887\n",
      "Epoch 765/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.6019e-04 - accuracy: 0.3079 - val_loss: 0.0038 - val_accuracy: 0.1134\n",
      "Epoch 766/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.7213e-04 - accuracy: 0.2963 - val_loss: 0.0035 - val_accuracy: 0.2680\n",
      "Epoch 767/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9116e-04 - accuracy: 0.3299 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 768/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5498e-04 - accuracy: 0.2674 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 769/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.9321e-04 - accuracy: 0.2685 - val_loss: 0.0037 - val_accuracy: 0.1753\n",
      "Epoch 770/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7549e-04 - accuracy: 0.2986 - val_loss: 0.0037 - val_accuracy: 0.2062\n",
      "Epoch 771/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8627e-04 - accuracy: 0.2986 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 772/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 4.3300e-04 - accuracy: 0.2755 - val_loss: 0.0042 - val_accuracy: 0.1856\n",
      "Epoch 773/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.4002e-04 - accuracy: 0.3125 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 774/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6895e-04 - accuracy: 0.2951 - val_loss: 0.0034 - val_accuracy: 0.1753\n",
      "Epoch 775/800\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4451e-04 - accuracy: 0.2824 - val_loss: 0.0035 - val_accuracy: 0.2784\n",
      "Epoch 776/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.3560e-04 - accuracy: 0.3241 - val_loss: 0.0037 - val_accuracy: 0.1546\n",
      "Epoch 777/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.3910e-04 - accuracy: 0.3009 - val_loss: 0.0036 - val_accuracy: 0.2990\n",
      "Epoch 778/800\n",
      "864/864 [==============================] - 3s 3ms/step - loss: 3.6528e-04 - accuracy: 0.2859 - val_loss: 0.0037 - val_accuracy: 0.1959\n",
      "Epoch 779/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4957e-04 - accuracy: 0.3252 - val_loss: 0.0038 - val_accuracy: 0.2165\n",
      "Epoch 780/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5301e-04 - accuracy: 0.3044 - val_loss: 0.0036 - val_accuracy: 0.2887\n",
      "Epoch 781/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7836e-04 - accuracy: 0.2928 - val_loss: 0.0039 - val_accuracy: 0.2784\n",
      "Epoch 782/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5592e-04 - accuracy: 0.2801 - val_loss: 0.0040 - val_accuracy: 0.2680\n",
      "Epoch 783/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5222e-04 - accuracy: 0.2940 - val_loss: 0.0037 - val_accuracy: 0.2062\n",
      "Epoch 784/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1597e-04 - accuracy: 0.2928 - val_loss: 0.0037 - val_accuracy: 0.2784\n",
      "Epoch 785/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.1289e-04 - accuracy: 0.3090 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 786/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0552e-04 - accuracy: 0.2975 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 787/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.5105e-04 - accuracy: 0.3275 - val_loss: 0.0037 - val_accuracy: 0.2887\n",
      "Epoch 788/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.3556e-04 - accuracy: 0.3079 - val_loss: 0.0037 - val_accuracy: 0.2165\n",
      "Epoch 789/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.3301e-04 - accuracy: 0.2975 - val_loss: 0.0038 - val_accuracy: 0.2680\n",
      "Epoch 790/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7840e-04 - accuracy: 0.3125 - val_loss: 0.0036 - val_accuracy: 0.2784\n",
      "Epoch 791/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.8872e-04 - accuracy: 0.3079 - val_loss: 0.0036 - val_accuracy: 0.3093\n",
      "Epoch 792/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 4.0456e-04 - accuracy: 0.2998 - val_loss: 0.0042 - val_accuracy: 0.2990\n",
      "Epoch 793/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.7175e-04 - accuracy: 0.3229 - val_loss: 0.0038 - val_accuracy: 0.2784\n",
      "Epoch 794/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.3592e-04 - accuracy: 0.3194 - val_loss: 0.0040 - val_accuracy: 0.3196\n",
      "Epoch 795/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.2713e-04 - accuracy: 0.3148 - val_loss: 0.0039 - val_accuracy: 0.2165\n",
      "Epoch 796/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.2316e-04 - accuracy: 0.3113 - val_loss: 0.0035 - val_accuracy: 0.2680\n",
      "Epoch 797/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.2486e-04 - accuracy: 0.3160 - val_loss: 0.0039 - val_accuracy: 0.2371\n",
      "Epoch 798/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4808e-04 - accuracy: 0.3229 - val_loss: 0.0039 - val_accuracy: 0.2887\n",
      "Epoch 799/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.6907e-04 - accuracy: 0.3067 - val_loss: 0.0039 - val_accuracy: 0.2371\n",
      "Epoch 800/800\n",
      "864/864 [==============================] - 2s 3ms/step - loss: 3.4088e-04 - accuracy: 0.3067 - val_loss: 0.0039 - val_accuracy: 0.3196\n"
     ]
    }
   ],
   "source": [
    "res = model.fit(X, y, epochs=800, batch_size=32, validation_split=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plotting the Accuracy and Loss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "visualize_training_results(res)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualizing the Predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validating the Model \n",
    "Plotting the difference between the Actual closing prices and the Predicted prices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted Prices:\n",
      " [[7400.789307966416], [7241.380163940355], [7106.899924937646], [7172.888593383611], [7127.823633456216], [7228.137184418529], [7248.061391475402], [7264.6668333025145], [7504.581478017058], [7683.578527556711]]\n",
      "\n",
      "Actual Prices:\n",
      " [[7397.796875], [7047.9169919999995], [7146.133789], [7218.371094], [7531.663573999999], [7463.105956999999], [7761.243651999999], [7569.629883], [7424.29248], [7321.988281]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,5))\n",
    "\n",
    "# Getting predictions by predicting from the last available X variable\n",
    "yhat = model.predict(X[-1].reshape(1, n_per_in, n_features)).tolist()[0]\n",
    "\n",
    "# Transforming values back to their normal prices\n",
    "yhat = scaler.inverse_transform(np.array(yhat).reshape(-1,1)).tolist()\n",
    "\n",
    "# Getting the actual values from the last available y variable which correspond to its respective X variable\n",
    "actual = scaler.inverse_transform(y[-1].reshape(-1,1))\n",
    "\n",
    "# Printing and plotting those predictions\n",
    "print(\"Predicted Prices:\\n\", yhat)\n",
    "plt.plot(yhat, label='Predicted')\n",
    "\n",
    "# Printing and plotting the actual values\n",
    "print(\"\\nActual Prices:\\n\", actual.tolist())\n",
    "plt.plot(actual.tolist(), label='Actual')\n",
    "\n",
    "plt.title(f\"Predicted vs Actual Closing Prices\")\n",
    "plt.ylabel(\"Price\")\n",
    "plt.legend()\n",
    "plt.savefig(\"BTC_validation.png\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Predicting/Forecasting the future prices"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                  Close\n",
      "2019-12-02  7750.527979\n",
      "2019-12-03  7993.556731\n",
      "2019-12-04  8048.777666\n",
      "2019-12-05  8293.133372\n",
      "2019-12-06  8278.055115\n",
      "2019-12-07  8168.779239\n",
      "2019-12-08  7833.326193\n",
      "2019-12-09  7327.463403\n",
      "2019-12-10  7053.108825\n",
      "2019-12-11  6831.816325\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Predicting off of y because it contains the most recent dates\n",
    "yhat = model.predict(np.array(df.tail(n_per_in)).reshape(1, n_per_in, n_features)).tolist()[0]\n",
    "\n",
    "# Transforming the predicted values back to their original prices\n",
    "yhat = scaler.inverse_transform(np.array(yhat).reshape(-1,1)).tolist()\n",
    "\n",
    "# Creating a DF of the predicted prices\n",
    "preds = pd.DataFrame(yhat, index=pd.date_range(start=df.index[-1], periods=len(yhat), freq=\"D\"), columns=df.columns)\n",
    "\n",
    "# Printing the predicted prices\n",
    "print(preds)\n",
    "\n",
    "# Number of periods back to visualize the actual values\n",
    "pers = 10\n",
    "\n",
    "# Transforming the actual values to their original price\n",
    "actual = pd.DataFrame(scaler.inverse_transform(df[[\"Close\"]].tail(pers)), index=df.Close.tail(pers).index, columns=df.columns).append(preds.head(1))\n",
    "\n",
    "# Plotting\n",
    "plt.figure(figsize=(16,6))\n",
    "plt.plot(actual, label=\"Actual Prices\")\n",
    "plt.plot(preds, label=\"Predicted Prices\")\n",
    "plt.ylabel(\"Price\")\n",
    "plt.xlabel(\"Dates\")\n",
    "plt.title(f\"Forecasting the next {len(yhat)} days\")\n",
    "plt.legend()\n",
    "plt.savefig(\"BTC_predictions.png\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}