lstm_weather.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "_3oPvg75wlId"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from keras.models import Sequential\n",
    "from keras.layers import LSTM,Dense ,Dropout,Bidirectional"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "T_W4bluqwoCE"
   },
   "outputs": [],
   "source": [
    "data = pd.read_csv(r\"april_to_june.csv\",sep=\",\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "4puAPlIyxO1J"
   },
   "outputs": [],
   "source": [
    "district = data[data.district == \"Adilabad\"]\n",
    "district = district.drop([\"dmcode\",\"district\",\"mandal\",\"wind_speed_max\",\"wind_speed_min\"],axis=1)\n",
    "district[\"temp\"] = (district[\"temp_min\"] + district[\"temp_max\"])/2\n",
    "district[\"humidity\"] = (district[\"humidity_min\"] + district[\"humidity_max\"])/2\n",
    "district = district.drop([\"temp_min\",\"temp_max\",\"humidity_min\",\"humidity_max\"],axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 195
    },
    "id": "o0K290nUxXQN",
    "outputId": "828d6576-b034-4712-8e39-a66a3eae2d53"
   },
   "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>date</th>\n",
       "      <th>rain</th>\n",
       "      <th>temp</th>\n",
       "      <th>humidity</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2020-04-01</td>\n",
       "      <td>0.0</td>\n",
       "      <td>29.80</td>\n",
       "      <td>57.35</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2020-04-02</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.15</td>\n",
       "      <td>55.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020-04-03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>31.60</td>\n",
       "      <td>46.40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2020-04-04</td>\n",
       "      <td>1.3</td>\n",
       "      <td>31.05</td>\n",
       "      <td>62.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2020-04-05</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.60</td>\n",
       "      <td>51.15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         date  rain   temp  humidity\n",
       "0  2020-04-01   0.0  29.80     57.35\n",
       "1  2020-04-02   0.0  30.15     55.40\n",
       "2  2020-04-03   0.0  31.60     46.40\n",
       "3  2020-04-04   1.3  31.05     62.30\n",
       "4  2020-04-05   0.0  30.60     51.15"
      ]
     },
     "execution_count": 4,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "district.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "79fIR9gGxbRh"
   },
   "outputs": [],
   "source": [
    "district = district.reset_index(drop=True)\n",
    "length = district.shape[0]\n",
    "train_size = int(length*7/10)\n",
    "training_set = district[:train_size]\n",
    "training_set = training_set.drop([\"date\"],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "-gDTmezbxh2T"
   },
   "outputs": [],
   "source": [
    "sc = MinMaxScaler(feature_range=(0,1))\n",
    "training_set_scaled = sc.fit_transform(training_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "7CCXVPdYyh-H",
    "outputId": "04dbb94e-e9ed-4b7d-a991-4446719cb269"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.        , 0.39880952, 0.52407034],\n",
       "       [0.        , 0.41964286, 0.50158547],\n",
       "       [0.        , 0.50595238, 0.39780917],\n",
       "       ...,\n",
       "       [0.10672854, 0.125     , 0.88036898],\n",
       "       [0.14269142, 0.1547619 , 0.83597579],\n",
       "       [0.00812065, 0.14583333, 0.8526953 ]])"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_set_scaled"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "IA0oOAiZxkoZ"
   },
   "outputs": [],
   "source": [
    "x_train = []\n",
    "y_train = []\n",
    "n_future = 6 # next 4 days temperature forecast\n",
    "n_past = 40 # Past 30 days \n",
    "for i in range(0,len(training_set_scaled)-n_past-n_future+1):\n",
    "    x_train.append(training_set_scaled[i : i + n_past , 0])     \n",
    "    y_train.append(training_set_scaled[i + n_past : i + n_past + n_future , 0 ])\n",
    "x_train , y_train = np.array(x_train), np.array(y_train)\n",
    "x_train = np.reshape(x_train, (x_train.shape[0] , x_train.shape[1], 1) )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "Ha1u3-R4xry6",
    "outputId": "bba7785c-f400-4f9c-fb5a-7ba4deb4d6f7"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/500\n",
      "71/71 [==============================] - 4s 62ms/step - loss: 0.0158 - acc: 0.1584\n",
      "Epoch 2/500\n",
      "71/71 [==============================] - 4s 63ms/step - loss: 0.0150 - acc: 0.1867\n",
      "Epoch 3/500\n",
      "71/71 [==============================] - 4s 63ms/step - loss: 0.0147 - acc: 0.1615\n",
      "Epoch 4/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0146 - acc: 0.1765\n",
      "Epoch 5/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0145 - acc: 0.1726\n",
      "Epoch 6/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0144 - acc: 0.1611\n",
      "Epoch 7/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0143 - acc: 0.1743\n",
      "Epoch 8/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0144 - acc: 0.2097\n",
      "Epoch 9/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0143 - acc: 0.1633\n",
      "Epoch 10/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0143 - acc: 0.1934\n",
      "Epoch 11/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0144 - acc: 0.1726\n",
      "Epoch 12/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0142 - acc: 0.1894\n",
      "Epoch 13/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0141 - acc: 0.1836\n",
      "Epoch 14/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0140 - acc: 0.1850\n",
      "Epoch 15/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0140 - acc: 0.2035\n",
      "Epoch 16/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0141 - acc: 0.1858\n",
      "Epoch 17/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0141 - acc: 0.1912\n",
      "Epoch 18/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0140 - acc: 0.1965\n",
      "Epoch 19/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0141 - acc: 0.1841\n",
      "Epoch 20/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0140 - acc: 0.1867\n",
      "Epoch 21/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0144 - acc: 0.1920\n",
      "Epoch 22/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0141 - acc: 0.1805\n",
      "Epoch 23/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0141 - acc: 0.1947\n",
      "Epoch 24/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0140 - acc: 0.1903\n",
      "Epoch 25/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0140 - acc: 0.1832\n",
      "Epoch 26/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0139 - acc: 0.1819\n",
      "Epoch 27/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0140 - acc: 0.1659\n",
      "Epoch 28/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0139 - acc: 0.1770\n",
      "Epoch 29/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0141 - acc: 0.2027\n",
      "Epoch 30/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0138 - acc: 0.1726\n",
      "Epoch 31/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0139 - acc: 0.2004\n",
      "Epoch 32/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0139 - acc: 0.1978\n",
      "Epoch 33/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0138 - acc: 0.1947\n",
      "Epoch 34/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0138 - acc: 0.1907\n",
      "Epoch 35/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0137 - acc: 0.1628\n",
      "Epoch 36/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0138 - acc: 0.2049\n",
      "Epoch 37/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0136 - acc: 0.1677\n",
      "Epoch 38/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0138 - acc: 0.1951\n",
      "Epoch 39/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0138 - acc: 0.1739\n",
      "Epoch 40/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0140 - acc: 0.1726\n",
      "Epoch 41/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0140 - acc: 0.1788\n",
      "Epoch 42/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0139 - acc: 0.1841\n",
      "Epoch 43/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0138 - acc: 0.1934\n",
      "Epoch 44/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0137 - acc: 0.1872\n",
      "Epoch 45/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0136 - acc: 0.1770\n",
      "Epoch 46/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0135 - acc: 0.1743\n",
      "Epoch 47/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0136 - acc: 0.1619\n",
      "Epoch 48/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0135 - acc: 0.2040\n",
      "Epoch 49/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0135 - acc: 0.1796\n",
      "Epoch 50/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0134 - acc: 0.1938\n",
      "Epoch 51/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0135 - acc: 0.1889\n",
      "Epoch 52/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0134 - acc: 0.1889\n",
      "Epoch 53/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0134 - acc: 0.1956\n",
      "Epoch 54/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0135 - acc: 0.2009\n",
      "Epoch 55/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0133 - acc: 0.1792\n",
      "Epoch 56/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0132 - acc: 0.1805\n",
      "Epoch 57/500\n",
      "71/71 [==============================] - 5s 64ms/step - loss: 0.0130 - acc: 0.1942\n",
      "Epoch 58/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0128 - acc: 0.1881\n",
      "Epoch 59/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0129 - acc: 0.1881\n",
      "Epoch 60/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0126 - acc: 0.2080\n",
      "Epoch 61/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0126 - acc: 0.1863\n",
      "Epoch 62/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0123 - acc: 0.1850\n",
      "Epoch 63/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0123 - acc: 0.2000\n",
      "Epoch 64/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0120 - acc: 0.2199\n",
      "Epoch 65/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0121 - acc: 0.2022\n",
      "Epoch 66/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0118 - acc: 0.1996\n",
      "Epoch 67/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0118 - acc: 0.2150\n",
      "Epoch 68/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0121 - acc: 0.1841\n",
      "Epoch 69/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0118 - acc: 0.1991\n",
      "Epoch 70/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0115 - acc: 0.2465\n",
      "Epoch 71/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0115 - acc: 0.2062\n",
      "Epoch 72/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0114 - acc: 0.2195\n",
      "Epoch 73/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0112 - acc: 0.1903\n",
      "Epoch 74/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0113 - acc: 0.2257\n",
      "Epoch 75/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0117 - acc: 0.2345\n",
      "Epoch 76/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0110 - acc: 0.2465\n",
      "Epoch 77/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0112 - acc: 0.2239\n",
      "Epoch 78/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0112 - acc: 0.2150\n",
      "Epoch 79/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0109 - acc: 0.2031\n",
      "Epoch 80/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0110 - acc: 0.2150\n",
      "Epoch 81/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0111 - acc: 0.2319\n",
      "Epoch 82/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0109 - acc: 0.2208\n",
      "Epoch 83/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0110 - acc: 0.2434\n",
      "Epoch 84/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0105 - acc: 0.2350\n",
      "Epoch 85/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0106 - acc: 0.2146\n",
      "Epoch 86/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0107 - acc: 0.2611\n",
      "Epoch 87/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0104 - acc: 0.2283\n",
      "Epoch 88/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0105 - acc: 0.2314\n",
      "Epoch 89/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0103 - acc: 0.2460\n",
      "Epoch 90/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0100 - acc: 0.2345\n",
      "Epoch 91/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0113 - acc: 0.2363\n",
      "Epoch 92/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0102 - acc: 0.2323\n",
      "Epoch 93/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0104 - acc: 0.2619\n",
      "Epoch 94/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0104 - acc: 0.2456\n",
      "Epoch 95/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0104 - acc: 0.2381\n",
      "Epoch 96/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0102 - acc: 0.2704\n",
      "Epoch 97/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0098 - acc: 0.2668\n",
      "Epoch 98/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0100 - acc: 0.2575\n",
      "Epoch 99/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0100 - acc: 0.2544\n",
      "Epoch 100/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0099 - acc: 0.2695\n",
      "Epoch 101/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0098 - acc: 0.2482\n",
      "Epoch 102/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0098 - acc: 0.2398\n",
      "Epoch 103/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0096 - acc: 0.2681\n",
      "Epoch 104/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0098 - acc: 0.2726\n",
      "Epoch 105/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0098 - acc: 0.2341\n",
      "Epoch 106/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0099 - acc: 0.2659\n",
      "Epoch 107/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0098 - acc: 0.2602\n",
      "Epoch 108/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0093 - acc: 0.2721\n",
      "Epoch 109/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0098 - acc: 0.2739\n",
      "Epoch 110/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0095 - acc: 0.2805\n",
      "Epoch 111/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0090 - acc: 0.2748\n",
      "Epoch 112/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0097 - acc: 0.2504\n",
      "Epoch 113/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0091 - acc: 0.2788\n",
      "Epoch 114/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0097 - acc: 0.2451\n",
      "Epoch 115/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0093 - acc: 0.2752\n",
      "Epoch 116/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0089 - acc: 0.2748\n",
      "Epoch 117/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0092 - acc: 0.2836\n",
      "Epoch 118/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0089 - acc: 0.2827\n",
      "Epoch 119/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0096 - acc: 0.2743\n",
      "Epoch 120/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0092 - acc: 0.2642\n",
      "Epoch 121/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0087 - acc: 0.2783\n",
      "Epoch 122/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0092 - acc: 0.2558\n",
      "Epoch 123/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0089 - acc: 0.2819\n",
      "Epoch 124/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0087 - acc: 0.2695\n",
      "Epoch 125/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0090 - acc: 0.2819\n",
      "Epoch 126/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0082 - acc: 0.2695\n",
      "Epoch 127/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0088 - acc: 0.2677\n",
      "Epoch 128/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0087 - acc: 0.2752\n",
      "Epoch 129/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0086 - acc: 0.2761\n",
      "Epoch 130/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0086 - acc: 0.2668\n",
      "Epoch 131/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0085 - acc: 0.2668\n",
      "Epoch 132/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0086 - acc: 0.2889\n",
      "Epoch 133/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0080 - acc: 0.2996\n",
      "Epoch 134/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0084 - acc: 0.2611\n",
      "Epoch 135/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0090 - acc: 0.2783\n",
      "Epoch 136/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0083 - acc: 0.2996\n",
      "Epoch 137/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0083 - acc: 0.2841\n",
      "Epoch 138/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0077 - acc: 0.2876\n",
      "Epoch 139/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0080 - acc: 0.2929\n",
      "Epoch 140/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0092 - acc: 0.2796\n",
      "Epoch 141/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0085 - acc: 0.2739\n",
      "Epoch 142/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0078 - acc: 0.2757\n",
      "Epoch 143/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0080 - acc: 0.3031\n",
      "Epoch 144/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0080 - acc: 0.2894\n",
      "Epoch 145/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0080 - acc: 0.2668\n",
      "Epoch 146/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0077 - acc: 0.2881\n",
      "Epoch 147/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0079 - acc: 0.3004\n",
      "Epoch 148/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0095 - acc: 0.2531\n",
      "Epoch 149/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0083 - acc: 0.3031\n",
      "Epoch 150/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.2646\n",
      "Epoch 151/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0085 - acc: 0.2681\n",
      "Epoch 152/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.2876\n",
      "Epoch 153/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0075 - acc: 0.2796\n",
      "Epoch 154/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0079 - acc: 0.2850\n",
      "Epoch 155/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0074 - acc: 0.2867\n",
      "Epoch 156/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.3018\n",
      "Epoch 157/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.2942\n",
      "Epoch 158/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0074 - acc: 0.2779\n",
      "Epoch 159/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0077 - acc: 0.2712\n",
      "Epoch 160/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0075 - acc: 0.2792\n",
      "Epoch 161/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0080 - acc: 0.2580\n",
      "Epoch 162/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.2845\n",
      "Epoch 163/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0067 - acc: 0.3124\n",
      "Epoch 164/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0071 - acc: 0.3106\n",
      "Epoch 165/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0073 - acc: 0.3049\n",
      "Epoch 166/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0076 - acc: 0.3071\n",
      "Epoch 167/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0070 - acc: 0.2770\n",
      "Epoch 168/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0068 - acc: 0.2903\n",
      "Epoch 169/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0081 - acc: 0.2827\n",
      "Epoch 170/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0074 - acc: 0.2889\n",
      "Epoch 171/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0076 - acc: 0.2735\n",
      "Epoch 172/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.2925\n",
      "Epoch 173/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0065 - acc: 0.3013\n",
      "Epoch 174/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0074 - acc: 0.2982\n",
      "Epoch 175/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0072 - acc: 0.2960\n",
      "Epoch 176/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0070 - acc: 0.3022\n",
      "Epoch 177/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0067 - acc: 0.2872\n",
      "Epoch 178/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.2659\n",
      "Epoch 179/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0069 - acc: 0.3106\n",
      "Epoch 180/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0067 - acc: 0.3000\n",
      "Epoch 181/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0070 - acc: 0.2898\n",
      "Epoch 182/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.2805\n",
      "Epoch 183/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0073 - acc: 0.2730\n",
      "Epoch 184/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0067 - acc: 0.2920\n",
      "Epoch 185/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0069 - acc: 0.2996\n",
      "Epoch 186/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0073 - acc: 0.3013\n",
      "Epoch 187/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0066 - acc: 0.2938\n",
      "Epoch 188/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0070 - acc: 0.2783\n",
      "Epoch 189/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0067 - acc: 0.2978\n",
      "Epoch 190/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0074 - acc: 0.2863\n",
      "Epoch 191/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0069 - acc: 0.2779\n",
      "Epoch 192/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.2973\n",
      "Epoch 193/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0064 - acc: 0.2916\n",
      "Epoch 194/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3217\n",
      "Epoch 195/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3111\n",
      "Epoch 196/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0065 - acc: 0.3062\n",
      "Epoch 197/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0062 - acc: 0.3128\n",
      "Epoch 198/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0084 - acc: 0.2752\n",
      "Epoch 199/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.3009\n",
      "Epoch 200/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0068 - acc: 0.2947\n",
      "Epoch 201/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.2752\n",
      "Epoch 202/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0092 - acc: 0.2854\n",
      "Epoch 203/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0071 - acc: 0.2898\n",
      "Epoch 204/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0068 - acc: 0.2836\n",
      "Epoch 205/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0064 - acc: 0.2912\n",
      "Epoch 206/500\n",
      "71/71 [==============================] - 5s 65ms/step - loss: 0.0065 - acc: 0.3035\n",
      "Epoch 207/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0066 - acc: 0.3199\n",
      "Epoch 208/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0070 - acc: 0.2920\n",
      "Epoch 209/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0080 - acc: 0.2907\n",
      "Epoch 210/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0067 - acc: 0.2827\n",
      "Epoch 211/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0080 - acc: 0.2973\n",
      "Epoch 212/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0070 - acc: 0.2991\n",
      "Epoch 213/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3248\n",
      "Epoch 214/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3062\n",
      "Epoch 215/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3257\n",
      "Epoch 216/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.2951\n",
      "Epoch 217/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0065 - acc: 0.3142\n",
      "Epoch 218/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3142\n",
      "Epoch 219/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0061 - acc: 0.2916\n",
      "Epoch 220/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3009\n",
      "Epoch 221/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3044\n",
      "Epoch 222/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3239\n",
      "Epoch 223/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0074 - acc: 0.3102\n",
      "Epoch 224/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.3053\n",
      "Epoch 225/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0058 - acc: 0.3035\n",
      "Epoch 226/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0059 - acc: 0.2982\n",
      "Epoch 227/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0064 - acc: 0.3035\n",
      "Epoch 228/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0065 - acc: 0.2903\n",
      "Epoch 229/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0060 - acc: 0.3013\n",
      "Epoch 230/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3049\n",
      "Epoch 231/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0069 - acc: 0.2969\n",
      "Epoch 232/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3159\n",
      "Epoch 233/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0075 - acc: 0.2858\n",
      "Epoch 234/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0063 - acc: 0.3124\n",
      "Epoch 235/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3204\n",
      "Epoch 236/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3031\n",
      "Epoch 237/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0060 - acc: 0.3265\n",
      "Epoch 238/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0059 - acc: 0.3137\n",
      "Epoch 239/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3190\n",
      "Epoch 240/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.3049\n",
      "Epoch 241/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3283\n",
      "Epoch 242/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3164\n",
      "Epoch 243/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3243\n",
      "Epoch 244/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0065 - acc: 0.3168\n",
      "Epoch 245/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3190\n",
      "Epoch 246/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0054 - acc: 0.3168\n",
      "Epoch 247/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0055 - acc: 0.3093\n",
      "Epoch 248/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3181\n",
      "Epoch 249/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3106\n",
      "Epoch 250/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3058\n",
      "Epoch 251/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3288\n",
      "Epoch 252/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0055 - acc: 0.3257\n",
      "Epoch 253/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0081 - acc: 0.3075\n",
      "Epoch 254/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0094 - acc: 0.2907\n",
      "Epoch 255/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.2920\n",
      "Epoch 256/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0063 - acc: 0.3190\n",
      "Epoch 257/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0068 - acc: 0.3062\n",
      "Epoch 258/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3146\n",
      "Epoch 259/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0061 - acc: 0.2991\n",
      "Epoch 260/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.3106\n",
      "Epoch 261/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0061 - acc: 0.3195\n",
      "Epoch 262/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0063 - acc: 0.3119\n",
      "Epoch 263/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0078 - acc: 0.2832\n",
      "Epoch 264/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0074 - acc: 0.3049\n",
      "Epoch 265/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3261\n",
      "Epoch 266/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0056 - acc: 0.3292\n",
      "Epoch 267/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0079 - acc: 0.2934\n",
      "Epoch 268/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0058 - acc: 0.3248\n",
      "Epoch 269/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0073 - acc: 0.3062\n",
      "Epoch 270/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3106\n",
      "Epoch 271/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3252\n",
      "Epoch 272/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0060 - acc: 0.3319\n",
      "Epoch 273/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0059 - acc: 0.3305\n",
      "Epoch 274/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0067 - acc: 0.3111\n",
      "Epoch 275/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.3279\n",
      "Epoch 276/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3487\n",
      "Epoch 277/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0069 - acc: 0.3292\n",
      "Epoch 278/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0063 - acc: 0.3142\n",
      "Epoch 279/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3385\n",
      "Epoch 280/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0061 - acc: 0.3208\n",
      "Epoch 281/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0061 - acc: 0.3049\n",
      "Epoch 282/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3164\n",
      "Epoch 283/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3394\n",
      "Epoch 284/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0059 - acc: 0.3142\n",
      "Epoch 285/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3243\n",
      "Epoch 286/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3093\n",
      "Epoch 287/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0071 - acc: 0.3128\n",
      "Epoch 288/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0062 - acc: 0.3173\n",
      "Epoch 289/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0059 - acc: 0.3212\n",
      "Epoch 290/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3447\n",
      "Epoch 291/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0058 - acc: 0.3195\n",
      "Epoch 292/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0058 - acc: 0.3341\n",
      "Epoch 293/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0060 - acc: 0.3235\n",
      "Epoch 294/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3398\n",
      "Epoch 295/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3190\n",
      "Epoch 296/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3212\n",
      "Epoch 297/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0068 - acc: 0.3296\n",
      "Epoch 298/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3323\n",
      "Epoch 299/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3243\n",
      "Epoch 300/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0060 - acc: 0.3248\n",
      "Epoch 301/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3327\n",
      "Epoch 302/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0057 - acc: 0.3460\n",
      "Epoch 303/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0053 - acc: 0.3195\n",
      "Epoch 304/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0055 - acc: 0.3385\n",
      "Epoch 305/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0054 - acc: 0.3407\n",
      "Epoch 306/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0058 - acc: 0.3381\n",
      "Epoch 307/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.3310\n",
      "Epoch 308/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0055 - acc: 0.3323\n",
      "Epoch 309/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3212\n",
      "Epoch 310/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0055 - acc: 0.3314\n",
      "Epoch 311/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3376\n",
      "Epoch 312/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0051 - acc: 0.3513\n",
      "Epoch 313/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0051 - acc: 0.3372\n",
      "Epoch 314/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0059 - acc: 0.3389\n",
      "Epoch 315/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0055 - acc: 0.3305\n",
      "Epoch 316/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0049 - acc: 0.3398\n",
      "Epoch 317/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0065 - acc: 0.3208\n",
      "Epoch 318/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0059 - acc: 0.3478\n",
      "Epoch 319/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0063 - acc: 0.3226\n",
      "Epoch 320/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3332\n",
      "Epoch 321/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0053 - acc: 0.3407\n",
      "Epoch 322/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3407\n",
      "Epoch 323/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0049 - acc: 0.3597\n",
      "Epoch 324/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0046 - acc: 0.3509\n",
      "Epoch 325/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0059 - acc: 0.3323\n",
      "Epoch 326/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0053 - acc: 0.3451\n",
      "Epoch 327/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0050 - acc: 0.3575\n",
      "Epoch 328/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0049 - acc: 0.3473\n",
      "Epoch 329/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3279\n",
      "Epoch 330/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0056 - acc: 0.3341\n",
      "Epoch 331/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0048 - acc: 0.3571\n",
      "Epoch 332/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0048 - acc: 0.3456\n",
      "Epoch 333/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0049 - acc: 0.3535\n",
      "Epoch 334/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0048 - acc: 0.3637\n",
      "Epoch 335/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0049 - acc: 0.3407\n",
      "Epoch 336/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0054 - acc: 0.3496\n",
      "Epoch 337/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0097 - acc: 0.2699\n",
      "Epoch 338/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0072 - acc: 0.3155\n",
      "Epoch 339/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0063 - acc: 0.3288\n",
      "Epoch 340/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3252\n",
      "Epoch 341/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0053 - acc: 0.3372\n",
      "Epoch 342/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0052 - acc: 0.3456\n",
      "Epoch 343/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3270\n",
      "Epoch 344/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3606\n",
      "Epoch 345/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3407\n",
      "Epoch 346/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0051 - acc: 0.3513\n",
      "Epoch 347/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3465\n",
      "Epoch 348/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3743\n",
      "Epoch 349/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0050 - acc: 0.3412\n",
      "Epoch 350/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0054 - acc: 0.3535\n",
      "Epoch 351/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0060 - acc: 0.3332\n",
      "Epoch 352/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0057 - acc: 0.3305\n",
      "Epoch 353/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3336\n",
      "Epoch 354/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3473\n",
      "Epoch 355/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0053 - acc: 0.3403\n",
      "Epoch 356/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0052 - acc: 0.3606\n",
      "Epoch 357/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0062 - acc: 0.3407\n",
      "Epoch 358/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0060 - acc: 0.3332\n",
      "Epoch 359/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0053 - acc: 0.3407\n",
      "Epoch 360/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0053 - acc: 0.3465\n",
      "Epoch 361/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0050 - acc: 0.3584\n",
      "Epoch 362/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3549\n",
      "Epoch 363/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3504\n",
      "Epoch 364/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0050 - acc: 0.3562\n",
      "Epoch 365/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0052 - acc: 0.3491\n",
      "Epoch 366/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0056 - acc: 0.3456\n",
      "Epoch 367/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0053 - acc: 0.3655\n",
      "Epoch 368/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0051 - acc: 0.3496\n",
      "Epoch 369/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0052 - acc: 0.3580\n",
      "Epoch 370/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0050 - acc: 0.3420\n",
      "Epoch 371/500\n",
      "71/71 [==============================] - 5s 66ms/step - loss: 0.0049 - acc: 0.3668\n",
      "Epoch 372/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0048 - acc: 0.3779\n",
      "Epoch 373/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0053 - acc: 0.3434\n",
      "Epoch 374/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0054 - acc: 0.3496\n",
      "Epoch 375/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3394\n",
      "Epoch 376/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0047 - acc: 0.3588\n",
      "Epoch 377/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0053 - acc: 0.3642\n",
      "Epoch 378/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0055 - acc: 0.3381\n",
      "Epoch 379/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0057 - acc: 0.3336\n",
      "Epoch 380/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0050 - acc: 0.3553\n",
      "Epoch 381/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3783\n",
      "Epoch 382/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0048 - acc: 0.3566\n",
      "Epoch 383/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3288\n",
      "Epoch 384/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0062 - acc: 0.3416\n",
      "Epoch 385/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3544\n",
      "Epoch 386/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0049 - acc: 0.3642\n",
      "Epoch 387/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0048 - acc: 0.3637\n",
      "Epoch 388/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0046 - acc: 0.3637\n",
      "Epoch 389/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0046 - acc: 0.3690\n",
      "Epoch 390/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0046 - acc: 0.3549\n",
      "Epoch 391/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0045 - acc: 0.3717\n",
      "Epoch 392/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0045 - acc: 0.3805\n",
      "Epoch 393/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0047 - acc: 0.3717\n",
      "Epoch 394/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3513\n",
      "Epoch 395/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3491\n",
      "Epoch 396/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0055 - acc: 0.3628\n",
      "Epoch 397/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0068 - acc: 0.3513\n",
      "Epoch 398/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0052 - acc: 0.3509\n",
      "Epoch 399/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0052 - acc: 0.3487\n",
      "Epoch 400/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0059 - acc: 0.3345\n",
      "Epoch 401/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0060 - acc: 0.3434\n",
      "Epoch 402/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3367\n",
      "Epoch 403/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0049 - acc: 0.3681\n",
      "Epoch 404/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0056 - acc: 0.3527\n",
      "Epoch 405/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0046 - acc: 0.3735\n",
      "Epoch 406/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0043 - acc: 0.3549\n",
      "Epoch 407/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0043 - acc: 0.3633\n",
      "Epoch 408/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0045 - acc: 0.3584\n",
      "Epoch 409/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0046 - acc: 0.3469\n",
      "Epoch 410/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0049 - acc: 0.3558\n",
      "Epoch 411/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0047 - acc: 0.3487\n",
      "Epoch 412/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0050 - acc: 0.3606\n",
      "Epoch 413/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0047 - acc: 0.3668\n",
      "Epoch 414/500\n",
      "71/71 [==============================] - 5s 67ms/step - loss: 0.0044 - acc: 0.3646\n",
      "Epoch 415/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0041 - acc: 0.3730\n",
      "Epoch 416/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3876\n",
      "Epoch 417/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3743\n",
      "Epoch 418/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3659\n",
      "Epoch 419/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3642\n",
      "Epoch 420/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0042 - acc: 0.3655\n",
      "Epoch 421/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0043 - acc: 0.3677\n",
      "Epoch 422/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0048 - acc: 0.3319\n",
      "Epoch 423/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0045 - acc: 0.3770\n",
      "Epoch 424/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0043 - acc: 0.3575\n",
      "Epoch 425/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0043 - acc: 0.3549\n",
      "Epoch 426/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0059 - acc: 0.3314\n",
      "Epoch 427/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0046 - acc: 0.3655\n",
      "Epoch 428/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3646\n",
      "Epoch 429/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0042 - acc: 0.3491\n",
      "Epoch 430/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0041 - acc: 0.3668\n",
      "Epoch 431/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0041 - acc: 0.3743\n",
      "Epoch 432/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0041 - acc: 0.3810\n",
      "Epoch 433/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3650\n",
      "Epoch 434/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0041 - acc: 0.3580\n",
      "Epoch 435/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0042 - acc: 0.3606\n",
      "Epoch 436/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0042 - acc: 0.3686\n",
      "Epoch 437/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0042 - acc: 0.3850\n",
      "Epoch 438/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0044 - acc: 0.3743\n",
      "Epoch 439/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0044 - acc: 0.3739\n",
      "Epoch 440/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0046 - acc: 0.3606\n",
      "Epoch 441/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0051 - acc: 0.3531\n",
      "Epoch 442/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0044 - acc: 0.3735\n",
      "Epoch 443/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3726\n",
      "Epoch 444/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0046 - acc: 0.3646\n",
      "Epoch 445/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0046 - acc: 0.3558\n",
      "Epoch 446/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0049 - acc: 0.3527\n",
      "Epoch 447/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3827\n",
      "Epoch 448/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0043 - acc: 0.3783\n",
      "Epoch 449/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3637\n",
      "Epoch 450/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3827\n",
      "Epoch 451/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3872\n",
      "Epoch 452/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3668\n",
      "Epoch 453/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0037 - acc: 0.3717\n",
      "Epoch 454/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0037 - acc: 0.3659\n",
      "Epoch 455/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3903\n",
      "Epoch 456/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3858\n",
      "Epoch 457/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0039 - acc: 0.3867\n",
      "Epoch 458/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3867\n",
      "Epoch 459/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0046 - acc: 0.3633\n",
      "Epoch 460/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0055 - acc: 0.3389\n",
      "Epoch 461/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0045 - acc: 0.3518\n",
      "Epoch 462/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3717\n",
      "Epoch 463/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0040 - acc: 0.3650\n",
      "Epoch 464/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0044 - acc: 0.3681\n",
      "Epoch 465/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3801\n",
      "Epoch 466/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3668\n",
      "Epoch 467/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0038 - acc: 0.3938\n",
      "Epoch 468/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3898\n",
      "Epoch 469/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0039 - acc: 0.3863\n",
      "Epoch 470/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0045 - acc: 0.3646\n",
      "Epoch 471/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3757\n",
      "Epoch 472/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0048 - acc: 0.3558\n",
      "Epoch 473/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0042 - acc: 0.3664\n",
      "Epoch 474/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0044 - acc: 0.3788\n",
      "Epoch 475/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3854\n",
      "Epoch 476/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0044 - acc: 0.3721\n",
      "Epoch 477/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3668\n",
      "Epoch 478/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0040 - acc: 0.3898\n",
      "Epoch 479/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0036 - acc: 0.3956\n",
      "Epoch 480/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3761\n",
      "Epoch 481/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3699\n",
      "Epoch 482/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0037 - acc: 0.3743\n",
      "Epoch 483/500\n",
      "71/71 [==============================] - 5s 71ms/step - loss: 0.0036 - acc: 0.3912\n",
      "Epoch 484/500\n",
      "71/71 [==============================] - 5s 71ms/step - loss: 0.0039 - acc: 0.4027\n",
      "Epoch 485/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0037 - acc: 0.3965\n",
      "Epoch 486/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0039 - acc: 0.3673\n",
      "Epoch 487/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0036 - acc: 0.4018\n",
      "Epoch 488/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0036 - acc: 0.3836\n",
      "Epoch 489/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3712\n",
      "Epoch 490/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0048 - acc: 0.3938\n",
      "Epoch 491/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0045 - acc: 0.3686\n",
      "Epoch 492/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0040 - acc: 0.3823\n",
      "Epoch 493/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3779\n",
      "Epoch 494/500\n",
      "71/71 [==============================] - 5s 68ms/step - loss: 0.0042 - acc: 0.3920\n",
      "Epoch 495/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0037 - acc: 0.3823\n",
      "Epoch 496/500\n",
      "71/71 [==============================] - 5s 70ms/step - loss: 0.0045 - acc: 0.3690\n",
      "Epoch 497/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3668\n",
      "Epoch 498/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3677\n",
      "Epoch 499/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0041 - acc: 0.3973\n",
      "Epoch 500/500\n",
      "71/71 [==============================] - 5s 69ms/step - loss: 0.0038 - acc: 0.3611\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x7ffb0818e5c0>"
      ]
     },
     "execution_count": 10,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "regressor = Sequential()\n",
    "regressor.add(Bidirectional(LSTM(units=30, return_sequences=True, input_shape = (x_train.shape[1],1) ) ))\n",
    "regressor.add(Dropout(0.2))\n",
    "regressor.add(LSTM(units= 30 , return_sequences=True))\n",
    "regressor.add(Dropout(0.2))\n",
    "regressor.add(LSTM(units= 30 , return_sequences=True))\n",
    "regressor.add(Dropout(0.2))\n",
    "regressor.add(LSTM(units= 30))\n",
    "regressor.add(Dropout(0.2))\n",
    "regressor.add(Dense(units = n_future,activation='linear'))\n",
    "regressor.compile(optimizer='adam', loss='mean_squared_error',metrics=['acc'])\n",
    "regressor.fit(x_train, y_train, epochs=500,batch_size=32 )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "k6HAOsNRxwiy"
   },
   "outputs": [],
   "source": [
    "testdataset = district[train_size+1:]\n",
    "testdataset = testdataset.drop([\"date\"],axis=1)\n",
    "testing = sc.transform(testdataset)\n",
    "testing = np.array(testing)\n",
    "testing = np.reshape(testing,(testing.shape[1],testing.shape[0],1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "qbjdntMU9g8k",
    "outputId": "68dabc60-de65-439d-9991-fc7c4273986b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.02830289 0.03721946 0.05675417 0.05975984 0.05839346 0.03808976]\n",
      " [0.02178546 0.05162836 0.06009634 0.05412932 0.05187536 0.12696737]\n",
      " [0.02791984 0.03319394 0.05865123 0.06084972 0.05578621 0.04237311]]\n"
     ]
    }
   ],
   "source": [
    "predicted_temperature = regressor.predict(testing)\n",
    "print(predicted_temperature)\n",
    "#predicted_temperature = sc.inverse_transform(predicted_temperature)\n",
    "predicted_temperature = np.reshape(predicted_temperature,(predicted_temperature.shape[1],predicted_temperature.shape[0]))\n",
    "predicted = sc.inverse_transform(predicted_temperature)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2ecPaxhe9wPD",
    "outputId": "9fab5f8c-1e49-4303-e777-4bc2ef55210e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adilabad (3294, 4)\n",
      "Adilabad (40, 4)\n",
      "Adilabad (40, 3)\n",
      "[[ 0.68315846 23.655895   16.499844  ]\n",
      " [ 3.744356   23.479536   25.286854  ]\n",
      " [ 1.3927945  23.77497    16.516207  ]\n",
      " [ 3.965298   23.79993    22.811487  ]\n",
      " [ 2.3627608  23.522583   16.39708   ]\n",
      " [ 4.631086   23.859846   13.989257  ]]\n",
      "Kumuram Bheem (2745, 4)\n",
      "Kumuram Bheem (40, 4)\n",
      "Kumuram Bheem (40, 3)\n",
      "[[ 0.7361548 24.031239  17.035828 ]\n",
      " [ 3.6107724 23.553925  32.150734 ]\n",
      " [ 2.2878335 24.012718  17.149565 ]\n",
      " [ 4.912265  24.053135  21.034851 ]\n",
      " [ 1.976992  23.541994  16.037872 ]\n",
      " [ 4.1978273 23.835129  13.999575 ]]\n",
      "Mancherial (3294, 4)\n",
      "Mancherial (40, 4)\n",
      "Mancherial (40, 3)\n",
      "[[ 1.056579  24.199738  17.199575 ]\n",
      " [ 3.7333002 23.618906  32.69194  ]\n",
      " [ 3.072333  24.353064  17.354576 ]\n",
      " [ 5.8042464 24.347557  24.216604 ]\n",
      " [ 1.7533089 23.612085  16.647343 ]\n",
      " [ 4.7435346 24.021172  16.356426 ]]\n",
      "Nirmal (3477, 4)\n",
      "Nirmal (40, 4)\n",
      "Nirmal (40, 3)\n",
      "[[ 1.0077344 24.142658  17.023916 ]\n",
      " [ 3.7956533 23.588068  32.28463  ]\n",
      " [ 3.5677953 24.547968  17.196085 ]\n",
      " [ 6.0531588 24.573692  26.719664 ]\n",
      " [ 1.8569427 23.627487  16.353521 ]\n",
      " [ 4.6139183 23.967943  16.019606 ]]\n",
      "Nizamabad (5307, 4)\n",
      "Nizamabad (40, 4)\n",
      "Nizamabad (40, 3)\n",
      "[[ 0.93930656 23.541676   16.548893  ]\n",
      " [ 4.0308585  23.61201    21.57368   ]\n",
      " [ 2.0299785  24.186913   16.788683  ]\n",
      " [ 4.7490034  24.182169   26.76177   ]\n",
      " [ 2.041098   23.536272   16.863173  ]\n",
      " [ 5.012731   23.971464   15.232627  ]]\n",
      "Jagtial (3294, 4)\n",
      "Jagtial (40, 4)\n",
      "Jagtial (40, 3)\n",
      "[[ 0.5771108 23.788141  16.814125 ]\n",
      " [ 3.236114  23.430119  29.037577 ]\n",
      " [ 3.5669274 24.617434  17.460787 ]\n",
      " [ 5.70266   24.604214  28.1888   ]\n",
      " [ 1.7520504 23.583237  16.626709 ]\n",
      " [ 4.654241  23.996765  15.857156 ]]\n",
      "Peddapalli (2562, 4)\n",
      "Peddapalli (40, 4)\n",
      "Peddapalli (40, 3)\n",
      "[[ 0.93301356 24.153286   17.010208  ]\n",
      " [ 3.6598465  23.629353   33.71005   ]\n",
      " [ 2.7643373  24.294176   17.145573  ]\n",
      " [ 5.463516   24.296953   24.679873  ]\n",
      " [ 1.8419745  23.602041   16.391068  ]\n",
      " [ 4.61306    23.957111   15.730934  ]]\n",
      "Jayashankar-Bhupalpally (2013, 4)\n",
      "Jayashankar-Bhupalpally (40, 4)\n",
      "Jayashankar-Bhupalpally (40, 3)\n",
      "[[ 0.77401936 23.936419   16.868101  ]\n",
      " [ 3.67884    23.510767   29.31032   ]\n",
      " [ 1.7395327  24.056492   16.530699  ]\n",
      " [ 4.1572766  24.008093   25.802912  ]\n",
      " [ 1.563266   23.344156   16.04002   ]\n",
      " [ 4.0747337  23.767815   13.782589  ]]\n",
      "Bhadradri-Kothagudem (4209, 4)\n",
      "Bhadradri-Kothagudem (40, 4)\n",
      "Bhadradri-Kothagudem (40, 3)\n",
      "[[ 1.4223769 24.18044   17.236563 ]\n",
      " [ 4.2213764 23.697842  28.048208 ]\n",
      " [ 3.1282856 24.56762   16.550192 ]\n",
      " [ 4.444854  24.583626  30.026905 ]\n",
      " [ 1.9349062 23.393732  15.864646 ]\n",
      " [ 3.8763885 23.726109  12.463489 ]]\n",
      "Mahabubabad (2928, 4)\n",
      "Mahabubabad (40, 4)\n",
      "Mahabubabad (40, 3)\n",
      "[[ 1.8004541 24.25863   17.45424  ]\n",
      " [ 4.7240424 23.916262  26.287462 ]\n",
      " [ 1.4974146 23.719418  16.18954  ]\n",
      " [ 3.752931  23.716934  20.394844 ]\n",
      " [ 1.7007581 23.422884  16.025    ]\n",
      " [ 4.1962366 23.828022  14.068075 ]]\n",
      "Warangal Rural (2928, 4)\n",
      "Warangal Rural (40, 4)\n",
      "Warangal Rural (40, 3)\n",
      "[[ 1.3307352 24.274511  17.428368 ]\n",
      " [ 4.3975377 23.778334  30.340733 ]\n",
      " [ 4.1449757 24.520557  16.037035 ]\n",
      " [ 7.058456  24.819962  24.244257 ]\n",
      " [ 1.9351088 23.493416  16.1069   ]\n",
      " [ 4.344166  23.838474  14.051855 ]]\n",
      "Warangal Urban (2013, 4)\n",
      "Warangal Urban (40, 4)\n",
      "Warangal Urban (40, 3)\n",
      "[[ 1.0074126 24.09589   17.053747 ]\n",
      " [ 4.002054  23.629112  29.80962  ]\n",
      " [ 2.884422  24.283371  16.815647 ]\n",
      " [ 5.3351717 24.230042  23.630543 ]\n",
      " [ 2.0659714 23.434418  15.910533 ]\n",
      " [ 4.143389  23.755293  13.011819 ]]\n",
      "Karimnagar (2928, 4)\n",
      "Karimnagar (40, 4)\n",
      "Karimnagar (40, 3)\n",
      "[[ 0.6033145 23.860088  16.767368 ]\n",
      " [ 3.462252  23.480158  30.958015 ]\n",
      " [ 2.890726  24.37686   17.01363  ]\n",
      " [ 5.4402695 24.315643  25.394466 ]\n",
      " [ 1.6131734 23.390646  16.068806 ]\n",
      " [ 4.2374635 23.82162   14.255573 ]]\n",
      "Rajanna-Siricilla (2379, 4)\n",
      "Rajanna-Siricilla (40, 4)\n",
      "Rajanna-Siricilla (40, 3)\n",
      "[[ 0.7261516 23.82029   16.800337 ]\n",
      " [ 3.4474623 23.370865  27.254377 ]\n",
      " [ 2.702733  24.419514  16.068546 ]\n",
      " [ 2.7124887 24.439054  31.034685 ]\n",
      " [ 3.2443051 23.572535  15.999    ]\n",
      " [ 4.34226   23.739046  11.905452 ]]\n",
      "Kamareddy (4026, 4)\n",
      "Kamareddy (40, 4)\n",
      "Kamareddy (40, 3)\n",
      "[[ 0.8515506 23.658638  16.421139 ]\n",
      " [ 3.8770862 23.500467  22.86796  ]\n",
      " [ 1.4823385 23.978354  17.071821 ]\n",
      " [ 4.5507445 24.054989  25.753538 ]\n",
      " [ 1.7722058 23.40611   16.099827 ]\n",
      " [ 4.304191  23.820234  14.124773 ]]\n",
      "Sangareddy (4758, 4)\n",
      "Sangareddy (40, 4)\n",
      "Sangareddy (40, 3)\n",
      "[[ 0.91493356 23.778215   16.642275  ]\n",
      " [ 4.033688   23.48115    25.130844  ]\n",
      " [ 1.5444993  23.360527   16.492325  ]\n",
      " [ 4.3185353  23.671534   14.807911  ]\n",
      " [ 1.3789057  23.271595   15.992072  ]\n",
      " [ 4.3182883  23.763704   13.779757  ]]\n",
      "Medak (3660, 4)\n",
      "Medak (40, 4)\n",
      "Medak (40, 3)\n",
      "[[ 1.3903561 24.141275  17.193592 ]\n",
      " [ 4.534441  23.78073   27.35993  ]\n",
      " [ 1.1187882 23.602957  16.46589  ]\n",
      " [ 3.8461816 23.661201  21.315413 ]\n",
      " [ 2.9861696 23.675936  16.526112 ]\n",
      " [ 4.85637   23.878662  13.333869 ]]\n",
      "Siddipet (4209, 4)\n",
      "Siddipet (40, 4)\n",
      "Siddipet (40, 3)\n",
      "[[ 1.046017  23.845026  16.577023 ]\n",
      " [ 4.0400143 23.558401  23.868437 ]\n",
      " [ 1.2432364 23.926287  16.372225 ]\n",
      " [ 1.9369423 23.62825   33.179157 ]\n",
      " [ 1.5613866 23.422533  16.510302 ]\n",
      " [ 4.658826  23.928719  16.163643 ]]\n",
      "Jangaon (2196, 4)\n",
      "Jangaon (40, 4)\n",
      "Jangaon (40, 3)\n",
      "[[ 1.4584653 24.328629  17.349072 ]\n",
      " [ 4.253416  23.7198    29.90138  ]\n",
      " [ 1.9547324 23.586906  15.680648 ]\n",
      " [ 3.7455378 23.630438  15.637184 ]\n",
      " [ 2.594207  23.501047  15.862741 ]\n",
      " [ 4.081021  23.726871  12.095589 ]]\n",
      "Yadadri-Bhongir (3111, 4)\n",
      "Yadadri-Bhongir (40, 4)\n",
      "Yadadri-Bhongir (40, 3)\n",
      "[[ 0.711339  23.874529  16.798416 ]\n",
      " [ 3.6391323 23.4905    28.705086 ]\n",
      " [ 1.9902619 23.706717  16.01632  ]\n",
      " [ 3.9623697 23.703983  17.260715 ]\n",
      " [ 2.279359  23.448622  15.920012 ]\n",
      " [ 4.2370944 23.748861  12.677756 ]]\n",
      "Medchal-Malkajgiri (2745, 4)\n",
      "Medchal-Malkajgiri (40, 4)\n",
      "Medchal-Malkajgiri (40, 3)\n",
      "[[ 1.0133084 23.724989  16.467138 ]\n",
      " [ 3.9910274 23.513569  22.323195 ]\n",
      " [ 2.6210282 23.645824  16.085281 ]\n",
      " [ 3.9432511 23.72989   12.530935 ]\n",
      " [ 1.9291228 23.437357  16.145065 ]\n",
      " [ 4.3215203 23.81207   13.860349 ]]\n",
      "Hyderabad (2928, 4)\n",
      "Hyderabad (40, 4)\n",
      "Hyderabad (40, 3)\n",
      "[[ 0.8586743 23.830454  16.487623 ]\n",
      " [ 3.685164  23.38715   25.987385 ]\n",
      " [ 1.5961719 23.3538    16.255392 ]\n",
      " [ 4.1013007 23.714663  13.741211 ]\n",
      " [ 1.7597553 23.352228  15.860466 ]\n",
      " [ 4.0773335 23.723263  13.373716 ]]\n",
      "Rangareddy (4941, 4)\n",
      "Rangareddy (40, 4)\n",
      "Rangareddy (40, 3)\n",
      "[[ 0.6925085 24.010746  17.04948  ]\n",
      " [ 3.1724226 23.3336    31.303427 ]\n",
      " [ 1.138516  23.684685  16.773514 ]\n",
      " [ 4.14369   23.660524  22.092293 ]\n",
      " [ 1.7616801 23.407341  16.221556 ]\n",
      " [ 4.505043  23.861311  14.83971  ]]\n",
      "Vikarabad (3294, 4)\n",
      "Vikarabad (40, 4)\n",
      "Vikarabad (40, 3)\n",
      "[[ 1.0546516 24.056751  16.819016 ]\n",
      " [ 4.094082  23.549091  28.834768 ]\n",
      " [ 1.298181  23.964235  16.227562 ]\n",
      " [ 3.1371863 23.64231   27.882298 ]\n",
      " [ 1.5394375 23.302893  15.874686 ]\n",
      " [ 4.1201105 23.731388  13.528141 ]]\n",
      "Mahabubnagar (2745, 4)\n",
      "Mahabubnagar (40, 4)\n",
      "Mahabubnagar (40, 3)\n",
      "[[ 1.1226413 23.736107  16.030891 ]\n",
      " [ 3.5371323 23.381174  20.726912 ]\n",
      " [ 1.3570325 23.46989   16.406923 ]\n",
      " [ 3.9443297 23.62821   17.13838  ]\n",
      " [ 2.2425706 23.518686  16.089462 ]\n",
      " [ 4.3752756 23.814886  13.388889 ]]\n",
      "Jogulamba-Gadwal (2196, 4)\n",
      "Jogulamba-Gadwal (40, 4)\n",
      "Jogulamba-Gadwal (40, 3)\n",
      "[[ 0.98433405 23.761826   16.55635   ]\n",
      " [ 4.1160126  23.584093   23.196892  ]\n",
      " [ 1.0680546  23.668941   16.456833  ]\n",
      " [ 3.7013197  23.622261   22.90296   ]\n",
      " [ 1.5763566  23.368921   16.166397  ]\n",
      " [ 4.3792963  23.838013   14.523872  ]]\n",
      "Wanaparthy (2562, 4)\n",
      "Wanaparthy (40, 4)\n",
      "Wanaparthy (40, 3)\n",
      "[[ 1.2671689 23.80532   16.191814 ]\n",
      " [ 3.9293928 23.502575  21.227057 ]\n",
      " [ 1.668878  23.815615  16.6874   ]\n",
      " [ 4.169959  23.85741   20.055603 ]\n",
      " [ 1.7377751 23.39094   16.273792 ]\n",
      " [ 4.3986645 23.834957  14.185115 ]]\n",
      "Nagarkurnool (3660, 4)\n",
      "Nagarkurnool (40, 4)\n",
      "Nagarkurnool (40, 3)\n",
      "[[ 1.9144295 24.367096  17.171738 ]\n",
      " [ 4.841078  23.879803  28.9744   ]\n",
      " [ 2.0865874 24.16568   17.22187  ]\n",
      " [ 4.6878457 24.068731  24.918312 ]\n",
      " [ 2.0517008 23.450201  16.052197 ]\n",
      " [ 4.1696715 23.781998  13.122973 ]]\n",
      "Nalgonda (5673, 4)\n",
      "Nalgonda (40, 4)\n",
      "Nalgonda (40, 3)\n",
      "[[ 0.94050115 23.9429     16.790438  ]\n",
      " [ 4.0708165  23.724497   27.96511   ]\n",
      " [ 0.95286393 23.700705   16.646662  ]\n",
      " [ 2.5781271  23.460941   29.812172  ]\n",
      " [ 1.7105452  23.395218   16.151169  ]\n",
      " [ 4.140561   23.818998   13.490006  ]]\n",
      "Suryapet (4209, 4)\n",
      "Suryapet (40, 4)\n",
      "Suryapet (40, 3)\n",
      "[[ 2.2593482 24.435574  17.415733 ]\n",
      " [ 5.2092237 23.930199  27.995459 ]\n",
      " [ 1.3388135 23.808771  16.477802 ]\n",
      " [ 4.325848  23.990509  23.522581 ]\n",
      " [ 1.7671437 23.400162  16.380453 ]\n",
      " [ 4.357932  23.849854  14.2806   ]]\n",
      "Khammam (3843, 4)\n",
      "Khammam (40, 4)\n",
      "Khammam (40, 3)\n",
      "[[ 1.6825901 24.356724  17.250772 ]\n",
      " [ 4.2256036 23.81683   30.994505 ]\n",
      " [ 2.6070771 24.21214   17.241413 ]\n",
      " [ 5.07947   24.129265  23.444656 ]\n",
      " [ 2.554459  23.880444  16.714613 ]\n",
      " [ 4.9482946 24.077263  16.646069 ]]\n",
      "Mulugu (1647, 4)\n",
      "Mulugu (40, 4)\n",
      "Mulugu (40, 3)\n",
      "[[ 1.46133   24.23115   17.030619 ]\n",
      " [ 4.1486506 23.750565  29.758389 ]\n",
      " [ 3.8650744 24.4889    16.337433 ]\n",
      " [ 6.6647134 24.735098  24.607618 ]\n",
      " [ 2.142881  23.602125  16.166386 ]\n",
      " [ 4.294223  23.868256  14.116781 ]]\n",
      "Narayanpet (2013, 4)\n",
      "Narayanpet (40, 4)\n",
      "Narayanpet (40, 3)\n",
      "[[ 1.0793675 23.643322  15.849202 ]\n",
      " [ 3.5081007 23.272484  20.022087 ]\n",
      " [ 3.198805  24.346813  15.216773 ]\n",
      " [ 4.567938  24.564722  23.461168 ]\n",
      " [ 1.9314054 23.391005  16.248207 ]\n",
      " [ 4.606793  23.808895  13.954055 ]]\n"
     ]
    }
   ],
   "source": [
    "to_predict = pd.read_csv(r\"july_to_sept.csv\",sep=\",\")\n",
    "predicted_list = []\n",
    "dis = to_predict.district.unique()\n",
    "for i in dis:\n",
    "    district = data[data.district == i]\n",
    "    district = district.drop([\"dmcode\",\"district\",\"mandal\",\"wind_speed_max\",\"wind_speed_min\"],axis=1)\n",
    "    district[\"temp\"] = (district[\"temp_min\"] + district[\"temp_max\"])/2\n",
    "    district[\"humidity\"] = (district[\"humidity_min\"] + district[\"humidity_max\"])/2\n",
    "    district = district.drop([\"temp_min\",\"temp_max\",\"humidity_min\",\"humidity_max\"],axis = 1)\n",
    "    district = district.reset_index(drop=True)\n",
    "    size = int(length*7/10)\n",
    "    print(i,district.shape)\n",
    "    testdataset = district[-40:]\n",
    "    print(i,testdataset.shape)\n",
    "    testdataset = testdataset.drop([\"date\"],axis=1)\n",
    "    print(i,testdataset.shape)\n",
    "    testing = sc.transform(testdataset)\n",
    "    testing = np.array(testing)\n",
    "    testing = np.reshape(testing,(testing.shape[1],testing.shape[0],1))\n",
    "    predicted_temperature = regressor.predict(testing)\n",
    "    \n",
    "    #predicted_temperature = sc.inverse_transform(predicted_temperature)\n",
    "    predicted_temperature = np.reshape(predicted_temperature,(predicted_temperature.shape[1],predicted_temperature.shape[0]))\n",
    "    predicted = sc.inverse_transform(predicted_temperature)\n",
    "    predicted_list.append((i,predicted[0][0],predicted[0][1],predicted[0][2]))\n",
    "    print(predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "oebFPUrY_i18"
   },
   "outputs": [],
   "source": [
    "dataset = []\n",
    "for i in range(len(predicted_list)):\n",
    "  predicted_list[i] = [predicted_list[i][0],predicted_list[i][1],predicted_list[i][2],predicted_list[i][3]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "4ow9H5dPDLIX"
   },
   "outputs": [],
   "source": [
    "dataset = pd.DataFrame(predicted_list,columns=[\"district\",\"rain\",\"temp\",\"humidity\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "qusC3D1cEA0Q"
   },
   "outputs": [],
   "source": [
    "dataset.to_csv(\"temp_rainfall.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "1af9AmJ6Eefp"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "lstm_weather.ipynb",
   "provenance": []
  },
  "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": "3.5.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}